7.
01 Heterogeneous Catalysis: Introduction
JW (Hans) Niemantsverdriet, Eindhoven University of Technology, Eindhoven, The Netherlands
R Schlogl, Fritz-Haber-Institut der MPG, Berlin, Germany
2013 Elsevier Ltd. All rights reserved.
7.01.1 Introduction
7.01.2 Heterogeneous Catalysis at the Molecular Level
7.01.3 Catalysts in Industry
7.01.4 Multiple Time and Length Scales in Catalysis
7.01.5 Scope and Content of Volume 7 on Heterogeneous Catalysis
7.01.5.1 Concepts, Mechanisms, and Kinetics
7.01.5.2 Supported Catalysts: Metals and Oxides
7.01.5.3 Solid Acids, Zeolites, and Mesoporous Materials
7.01.5.4 Sulfides, Carbon Polymorphs, Carbides, and Nitrides in Catalysis
7.01.5.5 Computational Modeling in Catalysis
7.01.5.6 Catalysis in Environmental and Energy Technology
References
7.01.1 Introduction
Catalysts accelerate chemical reactions without themselves
being consumed in the process. Consequently, catalysts are
enablers of reactions and as such the workhorses of the chem-
ical industry. An estimated 8590% of all chemical products
are manufactured in catalytic processes. These products are
transportation fuels, bulk and fine chemicals, or materials
such as polymers. Catalysts also play an important role in
environmental applications, such as the automotive exhaust
converter.
In heterogeneous catalysis, the catalyst is in a different
phase than the reactants. Normally it is a solid whose surface
is the seat of its catalytic activity, while the reactants are either
in the gas or the liquid phase. In homogeneous catalysis, on
the contrary, reactants and catalysts are in the same phase gas
or liquid while the catalyst is often an organometallic com-
plex. In biocatalysis, the catalyst is an enzyme.
Catalysts are nanomaterials, often in the form of nanosized
particles stabilized by some suitable support, or in the form of
micro- or mesoporous materials with high internal surface
areas, to optimize the fraction of material that can interact
with reacting molecules. The synthesis of catalysts and the
study of their structure and chemical composition, with preci-
sion close to atomic detail, are essential disciplines in catalysis
to understand catalytic behavior and to aid in optimization. To
overcome the inherent complexity of industrial catalysts, ide-
alized model systems such as single crystals, thin films, or
defined nanoparticles are commonly used in the clarification
of catalytic reaction mechanisms, while computational
methods have also become a fully accepted ingredient in the
research strategies of catalysis.
Heterogeneous catalysis forms the subject of this volume,
with the emphasis on materials and their surface reactivity.
Metals, oxides, zeolites, sulfides, carbons and carbides,
nitrides, phosphides, chlorides, and, in principle, any other
type of material with a surface that interacts with molecules
can serve as a catalyst. The purpose of this chapter is to
Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00701-4
1
1
3
3
4
4
5
5
5
5
6
6
introduce catalysis as a concept, and to provide a perspective
in context on the 21 chapters that constitute this important
volume on the industrially crucial application of inorganic
materials as heterogeneous catalysts. Several textbooks on
catalysis are available; a selection is listed in the references
section.110
7.01.2 Heterogeneous Catalysis at the
Molecular Level
A catalyst accelerates a chemical reaction, without itself being
consumed in the process. At the molecular level, a catalyst
binds molecules, breaks their bonds, lets the fragments react,
and then releases the product, after which the catalytic site is
available again for the next reaction cycle. A catalytic reaction is
essentially a cyclic series of elementary steps (bonding, reac-
tion, detachment), in which the catalyst participates; at the end
of the sequence, it leaves the reaction unaltered. This is the
cycle of catalysis, illustrated in Figure 1.
Figure 1 may nicely visualize the essence of catalytic action,
it does not explain why the reaction goes faster. To this end,
we need to consider the associated changes in potential energy
during the process, see Figure 2. To make the situation more
concrete, let us consider a simple, but realistic reaction, namely
the oxidation of CO. This reaction is catalyzed by many noble
metals, for example, palladium, platinum, rhodium, and even
gold (provided it is present in small nanoparticles and on the
right support), and it is important in automotive exhaust tech-
nology and in purification of hydrogen streams.
If there were no catalyst, CO and O2 would react upon
collision in the gas phase. A productive reaction event in the
gas phase would entail the breaking of the strong OdO bond
of molecular oxygen (498 kJ mol1), which forms the energy
barrier, or activation energy, for the reaction. Hence, a high
temperature of well over 1500 K would be needed to overcome
the high activation barrier for the homogeneous gas-phase
reaction between CO and O2.
1
2 Heterogeneous Catalysis: Introduction
A P
B filling of the antibonding orbitals of the O2 molecule, with the
Desorption
Catalyst
P
Catalyst
Adsorption
A B Reaction
Catalyst
Figure 1 A catalytic reaction follows a sequence of elementary steps,
in which reactant molecules bind to the catalyst, react on the catalyst,
and product molecules detach from the catalyst, leaving the latter
available for the next cycle. Adapted from Chorkendorff, I.;
Niemantsverdriet, J. W. Concepts of Modern Catalysis and Kinetics;
Wiley-VCH: Weinheim, 2007.
O2
CO CO2
+
Catalyst
Adsorption Reaction Desorption
E of the homogeneous gas-phase reaction, which translates to a
Adsorption Reaction Desorption
Reaction coordinate
Figure 2 Potential energy diagram of the heterogeneous catalytic CO
oxidation reaction. Upon adsorption, the O2 molecule dissociates, and the
surface is occupied by CO molecules and O atoms. In this example,
the rate will be determined by the surface reaction between adsorbed CO
and O (provided both are sufficiently available at the surface).
The barrier or activation energy for this elementary step is much lower
than that for the homogeneous gas-phase reaction between CO and
O2 molecules, illustrating why the catalytic reaction proceeds much
faster.
The catalytic reaction starts by bonding the reactants
CO and O2 to the catalyst. As this is a spontaneous reaction
for which the entropy change is negative (molecules lose
rotational and translational degrees of freedom upon adsorp-
tion), adsorption is by necessity an exothermic process.
(DG DH  T DS; DG should be negative for a spontaneous
reaction. Molecules lose entropy upon adsorption, and hence
the term TDS is positive. The only way to have DG < 0 is that
DH, the heat of adsorption, is negative: adsorption is an exo-
thermic process.) Hence the potential energy decreases upon
adsorption, see Figure 2. Moreover, the interaction of the O2
with the d-band of metals such as Rh, Pd, and Pt leads to the
result that it dissociates spontaneously. This step is driven
by the exothermicity between an adsorbed O2 and the two
O-atoms that bind more strongly to the metal surface than
the molecule does. The fact that the O2 molecule dissociates
spontaneously on the catalyst surface implies that the difficult
step of the homogenous gas phase reaction, OdO bond break-
ing, plays no role anymore. In heterogeneous catalysis, the
essential action of the catalyst lies almost always in the break-
ing of a strong bond, after which other bonds form, leading to
intermediates and products.
The overall rate of reaction is now determined by the
surface reaction between the adsorbed O-atom, and the CO.
Depending on the choice of metal and its surface structure,
the activation energy for this step is on the order of
50100 kJ mol1, implying that temperatures of a few hundred
Kelvin are sufficient to overcome this barrier. (As a crude
rule of thumb, 100 kJ mol1 activation energy corresponds
roughly to a reaction temperature of 400  50 K.) The cycle is
complete when the product CO2 molecule desorbs from the
surface, in an necessarily endothermic step. As CO2 is a
rather stable molecule, it interacts only weakly with the metal,
and the activation energy for desorption (which equals
the adsorption energy) is small and, therefore, kinetically
insignificant.
The potential energy diagram of Figure 2 illustrates that an
efficient catalyst brings reacting species together and offers
them an energetically favorable reaction path. The activation
energy of the catalytic reaction is significantly smaller than that
much larger reaction rate.
Note that the overall change in free energy for the catalytic
reaction equals that of the uncatalyzed reaction; catalysts do
not affect equilibrium constants, as they are properties of the
reactants and the products. Hence, catalysts change the kinetics
but NOT the thermodynamics.
Finally, a catalyst accelerates both the forward and the
reverse reactions. Hence, a good catalyst for a certain reaction
is also capable of catalyzing the reverse reaction, although, in
general, different conditions of temperature and pressure are
chosen.
Figure 2 sketches the potential energy profile for an effec-
tive combination of reaction and catalyst. It immediately sug-
gests what the success factors are. For example, if the catalyst
binds the reactants too strongly, the potential energy associ-
ated with this stage in the process will be low, and will add to
the activation energy of either the surface reaction or the
desorption. A similar consideration holds for the adsorbed
product, which should not bind too strongly to the surface;
otherwise, it will be retained. Hence, a good catalyst represents
a compromise: reactants and intermediates should bind to the
surface, but not too strongly and not too weakly. This concept
is also known as Sabatiers Principle.
A catalyst offers an alternative, energetically favorable
mechanism for the noncatalytic reaction, thus enabling pro-
cesses to be carried out at industrially feasible conditions of
pressure and temperature: the catalyst is therefore a facilitator
of chemical reactions.
 Heterogeneous Catalysis: Introduction 3

7.01.3 Catalysts in Industry Table 1 Examples of processes based on heterogeneous catalysis

Reaction/process Catalysts based on

Indeed, catalysts are indispensable for our society. An esti-
mated 8590% of all products of the chemical industry are Catalytic cracking of crude oil Zeolites
manufactured in catalytic processes. Some illustrative examples Hydrotreating of crude oil CoMo, NiMo, NiW
are mentioned: (sulfidic form)
Reforming of naphtha (to gasoline) Pt, PtRe, PtIr
More than 400 oil refineries worldwide use zeolite catalysts Alkylation H2SO4, HF, solid acids
to crack the long, complex, and diverse hydrocarbons in Polymerization of ethylene, propylene Cr Phillips catalysts,
crude oil into smaller fractions that are suitable for use as TiClx/MgCl2 Ziegler
transportation fuels. All hydrocarbon fractions are hydro- Natta catalysts
treated with sulfidic catalysts to remove sulfur, nitrogen, Ethylene epoxidation to ethylene oxide Ag
and metal impurities. Gasoline fractions are upgraded to Vinylchloride (ethylene Cl2) Cu (as chloride)
increased octane ratings with the help of reforming or Steam reforming of methane to CO H2 Ni
alkylation catalysts, usually based on noble metals and/or Water-gas shift reaction Fe (oxide), CuZnO
(CO H2O H2 CO2)
zeolites. Refineries are the largest users of catalysts.
Methanation (from CO and H2) Ni
Catalysts are used in the production of many base chemi-
Methanol synthesis (from synthesis gas) CuZnO
cals, such as hydrogen, ammonia, methanol, sulfuric acid, Dimethyl ether (from methanol) Solid acids
nitric acid, alcohols, ethylene oxide, ethylene glycol, ethyl FischerTropsch synthesis Co metal, Fe carbides
benzene, and many others. Also, polymers such as polyeth- (hydrocarbons)
ylene, polypropylene, polystyrene, and nylon are made Ammonia synthesis (N2 3H2 2NH3) Fe
using catalysts. Ammonia oxidation to NO and HNO3 PtRh
Many fine chemicals used in the pharmaceutical industry, Ammoxidation of propylene and BiMo, FeSb oxides
and specialties such as fragrances are produced in catalytic ammonia to acrylonitrile; oxidation of
processes. propylene to acrolein
Selective oxidation of butane to maleic Vanadium phosphates
Several applications of catalysts exist in environmental
anhydride (VPO catalysts)
pollution control, either by avoiding the formation of
Olefin metathesis W and Re oxides
hazardous compounds, or by degrading them in so-called Hydrogenation of vegetable oils Ni
end-of-pipe solutions. The three-way catalyst used to Aqueous phase reforming of polyalcohols Pt
remove CO, NO, and hydrocarbons from automotive and biomass components
exhausts is probably the best known application of catalysis Hydrodeoxygenation of biomass CoMo and NiMo sulfides,
by the public, but there are many others, such as the components zeolites
DeNOx technology to remove smog-forming nitrogen Sulfuric acid production V oxides
oxides from industrial exhaust streams, for example, from Total oxidation of CO and hydrocarbons Pt, Pd
power stations. (e.g., in car exhaust)
Combustion of volatile organic Various oxides, TiO2-based
Table 1 gives a number of important catalytic processes. compounds (VOCs) photocatalysts
Several applications of heterogeneous catalysts are discussed in Reduction of NOx (in exhaust) Rh, vanadium oxide
this volume, with particular emphasis on the surface chemistry
involved.
Industrial catalysts come in a variety of forms, varying from high mechanical strength to withstand the weight of the
loose particles to atomically dispersed species on a support. catalyst bed,
The latter can be a porous powder, such as silica or alumina high attrition resistance to mechanical wear.
with surface areas in the range of 200300 m2 g1, carbon,
with even higher surface areas up to 1000 m2 g1, zeolites, or To avoid pressure drops in catalyst beds, powders are
clays; or larger monolithic structures, such as the ceramic mono- often pressed into pellets, extrudates, or bodies with all kinds
liths used in automotive exhaust catalysts. The active materials of shapes, with dimensions on the order of millimeters to
can be applied on these supports by wet chemical impregnation centimeters, see also Figure 3.
methods, of which several varieties exist (Chapters 7.04 and
7.06). Whether a particular catalyst is a success in an industrial
application depends on many characteristics. Some of the most 7.01.4 Multiple Time and Length Scales in Catalysis
important considerations are listed below:
Catalysis as a phenomenon encompasses many length and
sufficient activity per unit of volume in the reactor, time scales, see Figure 3. At the atomic/molecular scale, mol-
high selectivity to the desired products at realistic conver- ecules bind to the atoms of the catalysts surface and react via
sion levels, several intermediates to products, which desorb and leave the
life time with respect to deactivation, surface available for subsequent reaction cycles (Figure 1).
regenerability, particularly when deactivation happens fast, The associated time scale of breaking and forming bonds is
reproducible preparation procedures, on the order of pico- to nanoseconds. The molecular descrip-
thermal stability against sintering, restructuring, and tion forms the basis for the potential energy considerations
volatilization, and the kinetic descriptions in Chapters 7.02 and 7.03. It is
4 Heterogeneous Catalysis: Introduction

10 mm

1 nm

Shaped catalyst
particles
Catalytic
1m
surface

Catalyst bed
in a reactor

1 mm
Catalytically active
particles on a support
Microscopic Mesoscopic Macroscopic
Figure 3 Length scales of interest in catalysis range from the subnanometer domain of the atomic and molecular level to the macroscopic domain
of industrial reactors. Characteristic time scales vary in parallel, from picoseconds in the microscopic regime, to minutes and longer in the
macroscopic domain. Adapted from Chorkendorff, I.; Niemantsverdriet, J. W. Concepts of Modern Catalysis and Kinetics; Wiley-VCH: Weinheim, 2007.

also the level at which much experimental surface science,
spectroscopy, and computational chemistry is performed.
Characteristic dimensions are in the subnanometer range,
and the relevant time domain is in the range of pico- to
nanoseconds. This is the domain where the inorganic chemis-
try of the catalytic surface manifests itself in atomic detail.
The next level is that of nanoparticles, which represent the
catalytically active phase, with typical sizes between 1 and
10 nm, and often supported on porous materials. For conve-
nience it is called the mesoscopic level here. Catalysts can also
be zeolites or mesoporous materials, again with characteristic
dimensions in the nano- to micrometer range. From the out-
side, catalysts are fine powders at this scale. Relevant properties
of the catalysts at this level are phase composition, morphol-
ogy, size, specific area, pore size, metalsupport interaction,
etc. This is the scale at which catalyst preparation and charac-
terization, and bench scale testing are important. The inorganic
chemistry at this level is that of collective properties, reflected
in the morphology of particles, sintering phenomena, porosity
in zeolites and supports, mechanical strength, stress and strain,
and structural stability.
Characteristic time scales are that of catalytic turnover cycles
and residence times of molecules on the surface, and can vary
from less than microseconds to more than seconds. When
mass transport phenomena such as the diffusion of molecules
in and out of pores become important, the characteristic time
scales can easily increase by an order of magnitude. Much
academic research and exploratory work occurs on this meso-
scopic scale.
As mentioned earlier, the fine powders of the mesoscopic
level can never be applied as such in large reactors, as catalyst
beds consisting of dense powders would give rise to unaccept-
able pressure drops. To avoid this, catalysts are shaped into
extrudates, pellets, or monoliths with sizes in the range of
several millimeters to several centimeters. Porosity, strength,
and attrition resistance are the properties of interest, and are
predominantly dealt with by catalyst manufacturers and reac-
tion engineers, and largely covered in patents.
Finally, we arrive at the level of catalyst beds of one to
several tens of meters, as they are applied in industrial reactors.
Although the intrinsic chemistry of molecules reacting at the
surface of nanoparticles is still at the heart of the industrial
process, the overall kinetic description of the process depends
heavily on extrinsic factors accounting for mass and heat trans-
port, and how efficiently the catalyst is utilized. Time scales are
again longer, and have to do with the movement of molecules
through catalyst beds.
Figure 3 summarizes the multiscale character of catalysis.
As a result, catalysis as a discipline is greatly multidisciplinary
in nature; many other science and engineering disciplines
such as inorganic, organic, physical, and analytical chemistry,
materials science, characterization, computational modeling
and simulation, and, of course, process engineering, are all
indispensable.
7.01.5 Scope and Content of Volume 7 on
Heterogeneous Catalysis
The purpose of this volume of Comprehensive Inorganic
Chemistry II is to give the reader a concise overview of hetero-
geneous catalysis from the perspective of inorganic materials
chemistry. Several experts from the field discuss the different
classes of materials and their catalytic properties. In addition,
there are short reviews of concepts, methodology, and indus-
trial applications. Together, the chapters give a good impres-
sion of heterogeneous catalysis.
7.01.5.1 Concepts, Mechanisms, and Kinetics
After having introduced basic concepts such as the catalytic
cycle and the potential energy changes associated with it,

Norbert Kruse and coworkers describe elementary reaction
steps such as adsorption, dissociation, and reaction and
desorption on well defined metal surfaces, along with some
of the typical spectroscopic and imaging methods used in
surface science, in Chapter 7.02. A catalytic reaction mecha-
nism is a sequence of elementary steps, and the reaction rate of
the entire process is the resultant of that of the individual steps.
In Chapter 7.03, Francesco Zaera discusses the kinetics of
mechanisms on the idealized surfaces of model systems.
These two chapters illustrate the power of surface science in
unraveling catalytic mechanisms and in providing kinetic and
thermodynamic data such as activation barriers and heats of
adsorption that lend themselves to comparison with the pre-
dictions of computational modeling, to be discussed in the
later chapters (Chapters 7.15, 7.16, 7.17, and 7.18).
7.01.5.2 Supported Catalysts: Metals and Oxides
Supported catalysts represent a substantial step up in complex-
ity as compared to the idealized surfaces of Chapters 7.02
and 7.03. In Chapter 7.04, John Regalbutto and coworkers
treat the synthesis of supported metal nanoparticles in admi-
rable detail. This implies consideration of the wet chemistry
involved in impregnation and, in particular, the nature of the
solidliquid interface where the catalyst precursors anchor or
deposit on the surface of the support. Knowledge of the pro-
cesses at this type of interface is of crucial importance for a
proper understanding of catalyst preparation, but is too often
ignored in, for example, the teaching of catalysis courses, or,
even more serious, in the synthesis of catalysts in many aca-
demic laboratories.
Seeing catalyst particles in all their morphologic detail is of
course the ideal of every researcher who works with supported
catalysts, and impressive progress has been made in imaging
over the past decade, not only of morphologies but also of the
accompanying chemical composition, sometimes in atomistic
detail. Abhaya Datye describes several imaging methods with
examples in Chapter 7.05.
Many industrial processes are based on oxidic catalysts (see
Table 1). Israel Wachs and Christopher Keturakis discuss sup-
ported oxides with the emphasis on monolayer systems
(Chapter 7.06). In passing they also introduce many of the
in situ characterization techniques used in the study of sup-
ported catalysts. Improved selectivity control in oxidation reac-
tions often requires fine-tuning of the surface reactivity, which
can be achieved by using mixed metal oxides, a subject
described along with many industrial examples by Gabriele
Centi and Siglinda Perathoner in Chapter 7.07. Heteropoly
compounds can be seen as large oxide molecules, and they
represent yet another form in which oxides can manifest them-
selves. Wataru Ueda and coworkers describe them in the con-
text of catalysis in Chapter 7.08. The subject forms a bridge to
the next chapters, as heteropoly compounds could also be
considered under the header of acids and bases.
7.01.5.3 Solid Acids, Zeolites, and Mesoporous Materials
Chapter 7.09 by Friederike Jentoft familiarizes the reader with
the concepts and conventions in the chemistry of solid acids
and bases. Inorganic solids are classified according to the
Heterogeneous Catalysis: Introduction 5
structural principles that give rise to acidic and basic properties,
and common methods for the characterization of acidic and
basic sites are explained. This chapter sets the stage for two
more specialized chapters on zeolitic materials.
Zeolites are crystalline porous aluminosilicates finding
wide-ranging applications in the petrochemical industry, in
catalytic processing, and in chemical synthesis based on their
high stability, tunable acidity and exceptional size- and shape
selectivity. In Chapter 7.10, Christian Hess describes recent
advances in the synthesis, characterization, and catalytic appli-
cations of ordered mesoporous and microporous materials
with heteroatoms, along with several catalytic applications.
Karin Moller and Thomas Bein (Chapter 7.11) review the
strategies for creating hierarchical micro- and mesoporous zeo-
lites that have been explored during the last decade, including
top-down leaching procedures and various bottom-up syn-
thetic approaches.
7.01.5.4 Sulfides, Carbon Polymorphs, Carbides, and
Nitrides in Catalysis
Oil refineries comprise some of the largest processes based on
heterogeneous catalysis. Chapter 7.12 briefly introduces the
refinery and zooms in on a family of processes called hydro-
treating. This includes hydrodesulfurization, hydrodenitro-
genation, hydrodeoxygenation, and demetallization of crude
oil feedstocks. Catalysts used for this purpose are mostly sul-
fides of molybdenum or tungsten, promoted with cobalt or
nickel. Wolfgang Bensch describes these sulfidic systems along
with their preparation routes, with particular attention to the
inorganic chemistry involved in the transformation of the
precursor oxides to the sulfides (Chapter 7.12).
Carbon finds its use in catalysis as a support and, although
less known, as a catalyst. Carbon can appear in many forms
polymorphs and some of these, such as coke (often graphitic/
graphene like deposits on the catalysts formed during
reactions of hydrocarbons or carbon monoxide) can also deac-
tivate a catalyst. Philippe Serp presents a comprehensive over-
view in Chapter 7.13.
Although there are no commercial applications yet, transi-
tion metal carbides and nitrides present many opportunities as
catalysts, and as electrode materials in for example, batteries
and supercapacitors. Catalytic applications may include those
using carbides in supports, and as cocatalysts together with
noble metals, or even as replacements for the latter, as Levi
Thompson and coworkers describe in Chapter 7.14.
7.01.5.5 Computational Modeling in Catalysis
Computational modeling based on density-functional theory
and Monte Carlo simulations have become generally accepted
ingredients of modern research in materials science and catal-
ysis, and have greatly contributed to the clarification of the
basic concepts underlying these disciplines. It is fair to say that
computational modeling has revolutionized the field by mak-
ing mechanistic details of surface reaction that can hardly be
studied experimentally accessible. Although many chapters in
this volume include results from computational approaches,
four chapters (Chapters 7.15, 7.16, 7.17, and 7.18) do so very
prominently.
6 Heterogeneous Catalysis: Introduction
In Chapter 7.15, Florian Goltl and Philippe Sautet intro-
duce the theoretical foundations of the quantum-mechanical
many-electron problem and expand it into the theorems and
approximations to exchange correlations that form the basis of
density functional theory. The applications discussed range
from probing the potential energy surface of surface reactions
to find ground-state geometries and transition states, to the
calculation of wave-function-related properties such as STM-
images or NMR-spectra, and even simulations of realistic
experimental conditions within the framework of ab initio
molecular dynamics calculations.
The enormous possibilities of density functional theory
calculations are further illustrated by Laszlo Guczi. (The editors
deeply regret the passing away of Professor Laszlo Guczi during
the completion of the CIC II series.) Arpad Molnar, and Detre
Teschner (Chapter 7.16). They discuss catalytic hydrogenation
reactions with small hydrocarbons to illustrate the principles,
followed by important industrial examples such as acetylene
and butadiene hydrogenation, methanation, methanol synthe-
sis, and asymmetric hydrogenations.
Another important, large-scale process based on hydroge-
nation is the ammonia synthesis presented by Anders Hellman
and colleagues in Chapter 7.17. Historically, catalytic ammo-
nia synthesis may rightfully be claimed to one of mankinds
greatest scientific achievements, as the socialeconomic impli-
cations of being able to produce fertilizers (and ammunition
for warfare) are enormous. Development of the ammonia syn-
thesis led by Fritz Haber went hand in hand with that of
continuous high-pressure processing (Carl Bosch). It also pro-
vided a systematic start of catalyst research, as thousands of
materials formulations were tested by Alwin Mittasch for their
catalytic potential. Chapter 7.17 describes how in recent years a
full understanding of the ammonia synthesis has evolved at the
atomic level through the combination of advanced experiments
with quantum-mechanical electronic structure calculations.
The section on computational modeling concludes with
Chapter 7.18 by Konstantin Neyman and coworkers on the
computational modeling of reacting systems. The authors pre-
sent several interesting cases and finish their contribution with
another hot topic: catalysis by gold nanoparticles.
7.01.5.6 Catalysis in Environmental and Energy Technology
The three-way catalyst for cleaning automotive exhaust is prob-
ably the best known example of a catalyst among the public. It
represents a marvelous example of robust catalytic technology
that works under a diverse range of conditions depending on
the particular combination of engine, type of fuel, and driving
style of individual drivers. Bob Farrauto and colleagues present
an illustrative overview of the many aspects that play a role
(Chapter 7.19).
Catalysis is a crucial asset in energy technology, a fact that
is not always appreciated by politicians and funding agencies.
In essence, catalysts are used in all transformation processes
that convert chemical energy sources into energy carriers that
store energy, and serve either as precursors to fuels, or as fuels
themselves. With all the new gas that has become available
recently, gas-to-liquids technology based on the time-tested
FischerTropsch synthesis is an attractive proposition. Jan van
de Loosdrecht and colleagues give a comprehensive overview
of the subject with ample attention to the inorganic chemistry
of the cobalt and iron catalysts used in FischerTropsch syn-
thesis in Chapter 7.20.
Biomass as a renewable feedstock for energy and chemicals
is the research focus of many industrial and academic labora-
tories in the world. Dmitry Murzin and Irina Simakova
(Chapter 7.21) describe the key processing technologies for
biomass utilization, such as gasification, pyrolysis, and depo-
lymerization by hydrolysis, and highlight the role of catalysts
in these applications. In chemical terms, cellulose, hemicellu-
loses, lignins, and extractives such as fatty acids and terpenes
form the type of molecules that need to be processed. Almost
the entire spectrum of catalytic reactions has to be invoked:
oxidation, hydrogenation, cracking, hydrodeoxygenation,
dehydration, ketonization, aldol condensation, and decarbox-
ylation. The authors discuss the basic principles for selecting
the catalysts for these reactions.
Eventually, the sun will be the ultimate provider of sustain-
able energy. Photocatalytic splitting of water would be the
ideal process for large-scale production of hydrogen. This can
be achieved by modifying metal-oxide-based photocatalysts
with suitable cocatalysts to provide active redox sites, as
described by Kazuhiko Maeda and Kazunari Domen in the
final contribution to this volume, Chapter 7.22.
We hope that many graduate students, postdoctoral
researchers as well as colleagues in academia and industry
will benefit from this valuable collection of chapters that
together provide a rather up-to-date summary of the field of
heterogeneous catalysis. We are very grateful to the authors as
well as the editors-in-chief for working with us to produce this
volume on heterogeneous catalysis in the highly reputed Com-
prehensive Inorganic Chemistry series.
References
1. Ertl, G., Knozinger, H., Schuth, F., Weitkamp, J., Eds.; In Handbook of
Heterogeneous Catalysis; Wiley-VCH: Weinheim, 2008.
2. Somorjai, G. A. Introduction to Surface Chemistry and Catalysis, 2nd ed.; Wiley:
New York, 2010.
3. Thomas, J. M.; Thomas, W. J. Principles and Practice of Heterogeneous Catalysis.
Wiley-VCH: Weinheim, 1997.
4. Boudart, M.; Djega-Mariadassou, G. Kinetics of Heterogeneous Catalytic Reactions.
Princeton University Press: Princeton, NJ, 1984.
5. Satterfield, C. N. Heterogeneous Catalysis in Industrial Practice. McGraw-Hill:
New York, 1991.
6. Bartholomew, C. H.; Farrauto, R. J. Fundamentals of Industrial Catalytic Processes.
Wiley: New York, 2006.
7. Ross, J. R. H. Heterogeneous Catalysis Fundamentals and Application. Elsevier:
Amsterdam, 2012.
8. Rothenberg, G. Catalysis. Wiley-VCH: Weinheim, 2008.
9. Cornils, B., Herrmann, W. A., Schlogl, R., Wong, C.-H., Eds.; In Catalysis from A to
Z; Wiley-VCH: Weinheim, 2003.
10. Chorkendorff, I.; Niemantsverdriet, J. W. Concepts of Modern Catalysis and
Kinetics. Wiley-VCH: Weinheim, 2007.
 7.02 Elementary Steps in Heterogeneous Catalysis
F Devred, P Dulgheru, and N Kruse, Universite Libre de Bruxelles, Brussels, Belgium
2013 Elsevier Ltd. All rights reserved.

7.02.1 Introduction 7
7.02.1.1 General Principles and Illustrative Examples 7
7.02.1.2 Adsorption 8
7.02.1.2.1 Physisorptionchemisorption 8
7.02.1.2.2 Sticking and coverage 10
7.02.1.2.3 Structure sensitivity 11
7.02.1.3 Diffusion 12
7.02.1.4 Reaction 13
7.02.1.5 Desorption 16
7.02.2 CO Adsorption and Oxidation 16
7.02.2.1 General Principles 16
7.02.2.2 CO Adsorption 16
7.02.2.3 CO Oxidation over Transition-Metal Surfaces 18
7.02.3 NO Adsorption, Decomposition, and Reaction 24
7.02.3.1 General Aspects 24
7.02.3.2 NO Adsorption and Dissociation 25
7.02.3.3 NO Oxidation and Reduction 28
7.02.3.3.1 NO reaction with oxygen 28
7.02.3.3.2 NOx Reduction 28
7.02.4 Methane Activation 29
7.02.4.1 General Principles 29
7.02.4.2 Single-Crystal Studies 29
7.02.4.3 Computational Studies: Geometry of Adsorption and Decomposition Mechanism 31
7.02.4.4 Real Catalyst Studies: Steam/Dry Steam Reforming 33
7.02.5 Conclusion 34
References 35

7.02.1 Introduction
7.02.1.1 General Principles and Illustrative Examples
There is a consensus among researchers worldwide that actual
mechanisms of catalysis at the atomic scale are still poorly
understood. A wealth of knowledge has been accumulated in
the last 20 years using surface science techniques to examine
well-defined single-crystal surfaces reacting with molecules at
overall low pressures. While some of these techniques such as
scanning probe microscopy or various photo-induced excita-
tion spectroscopies can be applied under ambient pressure
conditions or even in the liquid phase, it remains difficult to
transpose the results to real conditions. Moreover, real catalysts
are structurally complex and despite major achievements in
materials characterization we are far from controlling this
complexity on a truly atomistic level under conditions of an
ongoing catalytic reaction. The experimental difficulty encoun-
tered in basic catalysis research is therefore frequently para-
phrased by a complexity gap. The existence of pressure and
materials gaps between model experiments and practice is also
considered in this chapter.
Heterogeneous catalysis is, first of all, a kinetic phenome-
non which is usually described by splitting up the overall
process cycle into a series of elementary reaction steps.
For gas/solid systems, these elementary reaction steps can
be classified as
1. adsorption of gas species on the solid surface (physisorp-
tion or chemisorption),
2. diffusion of reactants on the surface of the solid,
3. reaction on the surface, and
4. desorption of the products.
In this process cycle we do not consider transport into or
out of the pores of a real catalyst.
We shall now briefly describe some of the key parameters
influencing adsorption and surface reactions from a viewpoint
of surface science (i.e., sticking coefficient, rate coefficients,
activation energies, structure dependence, etc.) which will ulti-
mately enable the reader to get a more profound understand-
ing on a truly atomistic level and set the scene for a mean-field
approach as frequently encountered in the research of catalytic
reaction mechanisms. This will lead to micro-kinetic rate
expressions following LangmuirHinshelwood,1 EleyRideal,2
or Marsvan Krevelen-type3 reaction mechanisms. After this
introduction which, due to its necessary brevity, is not meant
to provide a full account of the most determining papers
published over about five decades of research, several catalytic
reaction systems deemed to be important for other chapters

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00703-8 7

8 Elementary Steps in Heterogeneous Catalysis
(for instance Chapter 7.19) in this comprehensive dealing
with catalysis by metals and metal oxides will be reviewed
for their mechanistic details in a mean-field approach
(Figure 1).
7.02.1.2 Adsorption
7.02.1.2.1 Physisorptionchemisorption
Gassurface interactions can be classified according to vari-
ous aspects of gasdsurface bond formation. Physisorption is
weak and involves van der Waals forces between the gas
molecules and the solid surface. In the absence of short-
range interaction forces, physisorbed molecules largely con-
serve their spectroscopic characteristics which they have in
the gas phase. Therefore, they are never directly involved in
catalytic turnover.
Different from physisorption, chemisorption provides
strong interaction with the solid surface. Shortly after von
Laue4 (1914 Nobel Prize in Chemistry) described the atomic
arrangement of a bulk solid as an ordered lattice, Langmuir58
(1932 Nobel Prize in Chemistry) developed the concept of
atomic sites on surfaces: atoms or molecules striking the sur-
face may either bounce back or establish a bond to one or
several of these sites. As early as 1916, Langmuir distinguished
between strong, short-range forces during chemisorption on
Adsorption
Desorption
Dissociation Diffusion
Figure 1 Schematic interaction between gas molecules and a surface; the five elementary steps constitute the catalytic cycle: adsorption, dissociation,
diffusion, reaction, and desorption.
Figure 2 Principle of dissociative (activated) chemisorption for diatomic molecules. Reprinted from Ertl, G. J. Mol. Catal. Chem. 2002, 182(1), 516,
with permission, 2002 Elsevier Science.
substrates and long-range physisorption. Since then it became
clear that an understanding of chemisorption would afford a
detailed knowledge of the electronic structure of the system.9 It
is not, however, within the scope of this chapter to develop the
various fundamental aspects of chemisorption. Rather, we refer
to the excellent review by Freund.10
Bond breaking associated with chemisorption (dissociative
chemisorption) is one of the key processes for heterogeneous
catalysis. Its principle is sketched for a diatomic molecule in
Figure 2. The well-known Lennard-Jones potential diagram11
illustrates the progress of chemisorption along a one-dimensional
(1D) reaction coordinate (bottom right of Figure 2).
In reality, atoms and molecules move on a multidimen-
sional potential energy surface (PES). In the upper right part of
Figure 2, a simplified version of the PES is chosen to demon-
strate the minimum energy path leading from the gas phase
over the saddle point into the chemisorption state. The con-
tours demonstrate the variation of the potential energy as a
function of two pivotal coordinates: the distance (x) of the
diatomic molecule from the surface and the separation (y)
between the two atoms.
The height of the activation barrier () along the reaction
coordinate determines the dissociation probability. The pro-
cess encountered during dissociative chemisorption on the
surface may be summarized as follows:
Adsorption
Reaction Diffusion
y x
y
r
x
Ep
Ediss
A2
A2,ad
2Aad

While approaching the surface, the molecule is trapped in a
shallow potential trough, A2,ads (molecular physisorption).
Getting closer to the surface, the physisorbed molecule
might overcome the energy barrier for dissociation and fall
apart so as to fill the deep potential of the chemisorbed state.
The position of the crossing point between the potentials of
physisorption and chemisorption is the key parameter: it
decides on whether chemisorption is activated or not. Small
shifts in the position of either of the two curves can significantly
alter the size of the activation barrier and therefore influence the
resulting kinetics. Such small changes may be easily encountered
for increasing surface coverage. Quite generally, the potential
diagram in Figure 2 holds only for the limit of zero coverage.
For many simple adsorption systems, the crossover point lies
below the level of zero potential energy meaning that there is no
direct activation barrier for dissociation.
As mentioned above, the PES is multidimensional. In princi-
ple, all translational, vibrational, and rotational degrees of free-
dom of the molecule have to be considered for a complete
description of the molecules reactivity when it approaches the
surface for dissociation. Thus, for a diatomic molecule as shown
in Figure 2, all six degrees of freedom would have to be consid-
ered explicitly. In addition, lattice vibrational and surface electron
excitation would have to be included for quantitatively accurate
calculations. It appears, however, that an adiabatic theory in
BornOppenheimer approximation on the basis of a 6D model
excluding coupling to the lattice provides an approximately valid
or even reasonably good description of the dissociative chemi-
sorption of light molecules such as hydrogen. Possibly the first
studies in surface science which dealt with the nonactivated
dissociative chemisorption of H2 on Pd(100) theoretically in
6D adiabatic calculations12 as well as experimentally13,14 using
molecular beam techniques showed that the reaction probability
first decreases and then increases with increasing collision energy.
To explain the increasing reactivity at low energies, Gross et al.12
suggested a steering of the incoming molecules to favorable
reactive orientations. They also suggested that the absence of
rotational excitation under these conditions is responsible for
such steering. This was experimentally confirmed a little later.15
For more complicated PES topologies and heavier molecules
(N2) where the molecule undergoes a chemical interaction with
the metal surface so as to form one or multiple minima or
inflection points in the PES, the agreement of a 6D theory with
experiments is less satisfactory. This is most likely due to the
neglect of the lattice (and possibly electronhole pair formation)
in the theoretical model, thus freezing out any energy transfer to
or from the surface.16
Despite its limitations, the low-dimensional PES of Figure 2
is frequently used to distinguish between early and late acti-
vation barriers. If the molecule approaches the surface with high
translational energy along the path r, the location of the barrier
() may occur earlier as compared with a situation where the
molecule is translationally less excited. However, if the molecule
is too hot it may well move up the wall and leave the exit
channel in a vibrationally excited state. In the case of a late
barrier, the incoming molecule has a higher chance for sur-
mounting () while being vibrationally excited. Once having
crossed the barrier, the molecules will then possess a relatively
high translational energy. This picture is in complete agreement
Elementary Steps in Heterogeneous Catalysis 9
with the so-called Polanyi rules17 established for atom exchange
in simple atommolecule encounters in the gas phase.
In relation with the above discussion, it is instructive to
consider the experimental findings by Wintterlin et al.18 who
demonstrated the existence of nonthermal processes during
dissociation. Until then, there was a debate about the chemical
energy released during dissociation: does it dissipate into the
heat bath of the solid or does it transform partially into kinetic
energy, causing nonthermal motion of the ad particles across the
surface? The authors adsorbed O2 molecules on a Pt(111) sur-
face at 160 K (where no thermal random walk occurs). Using
scanning tunneling microscopy (STM) they followed the fate of
the O atoms on O2 dissociation. While they were first separated
by an average distance of two lattice sites (Figure 3), they moved
away from each other proving the existence of nonthermal,
transient effects. Using the travel distance of the hot atoms
until their full equilibration along with an assumption on the
mean velocity due to the releasing chemisorption energy, the
average relaxation time was estimated to be <1012 s. This must
mean that any collision-induced adsorption/reaction or desorp-
tion of products escapes a proper description by the transition-
state theory. Ertl (2007 Nobel Prize in Chemistry) provided a
detailed review on this subject.19
When turning to polyatomic molecules with many more
motional degrees of freedom than just six as in di-atomic mol-
ecules, 1D Lennard-Jones diagrams similar to Figure 2 are still
being used to describe the reaction. Following on from the
above discussion, this projects all information about the many
translational, vibrational, and rotational energetic coordinates
of a dissociating molecule onto a generic reaction coordinate.
The question of how energy is distributed while forming
the transition complex is circumvented this way. This may
be a reasonable approximation when reactivity scales with
total energy. However, the observation of vibrational mode-
specific chemistry may impose limitations to the approach.
(a)
2
3
7
7
2 2
(b)
Figure 3 STM images of Pt(111) recorded after O2 dissociation at
around 160 K: the black dots are oxygen atoms and the distances
between two atoms are indicated. Reprinted from Wintterlin, J.;
Schuster, R.; Ertl, G. Phys. Rev. Lett. 1996, 77(1), 123126, with
permission 1996, The American Physical Society.
10 Elementary Steps in Heterogeneous Catalysis
In Section 7.02.4, we shall come back to mode-specific and
even bond-selective chemistry for which the assumptions under-
lying statistical rate theories no longer apply.
7.02.1.2.2 Sticking and coverage
Traditionally, one considers sticking as the efficiency of a pro-
cess where a molecule approaching a surface is eventually
being trapped by the potential. For convenience, this efficiency
is expressed in kinetic terms by considering the fraction of
molecules getting adsorbed while impinging per unit of time
on the surface. We can then define the initial sticking coeffi-
cient S0 as the ratio of particles adsorbing versus those imping-
ing on the uncovered surface (0  S0  1). According to the
simple hit and stick model of Langmuir, S0 1. This is clearly
in conflict with the usual observation of most gases being
adsorbed with S0 values below, or even far below, 1. Dissocia-
tive nitrogen adsorption on fe single-crystal surfaces is an
example for very small S0 values. Carbon monoxide and nitric
oxide seem to stick fairly well (S0 close to 1) on many
transitions-metal surfaces, possibly under dissociation depend-
ing on the metal and the surface structure (see Figure 4).
The sticking coefficient can be measured experimentally by
exposing a clean surface at constant temperature (T ) to a gas at
well-defined pressure (p) for a given time (t). The amount of
gas retained by the surface is then compared to the total
amount of gas that hit it. Any suitable surface science technique
capable of titrating the amount of adsorbed species can then be
used to quantify the take-up and to provide numbers for the
sticking probability. King and Wells20 used a molecular beam
for surface bombardment along with background pressure
measurements by mass spectrometry to construct sticking
probability profiles, S versus surface coverage.
Once the sticking probability is known, we can use the
HertzKnudsen equation


J p=2pmkT 1=2 molecules m2 s1 [1]
for the flux of incident molecules (impingement rate) to deter-
mine the adsorption rate:

which may be either attractive or repulsive. For example, spe-
rad JS molecules m2 s1 [2]
where p is the gas pressure (N m2), m the mass of one mole-
cule (kg), and T the temperature (K).
1 Saturation
= kL.P/(1 + kL.P)
q Tad, 2>Tad, 1
0
P
Figure 4 Langmuir isotherm: evolution of the coverage as a function of
pressure. kL is the Langmuir constant that only depends on temperature.
According to the hit and stick model adopted by Langmuir,
the evolution of the coverage Y with the (equilibrium) pressure
can be written as
Y kL p=1 kL p [3]
where kL is the Langmuir constant (kL kad/kdes for first-order
adsorption/desorption).
Figure 4 shows the classical representation of the Langmuir
isotherm. The lower the temperature, the steeper the initial
slope (provided the temperature is chosen so as to guarantee
the occurrence of a dynamic adsorption equilibrium).
Figure 5 is a kinetic representation of the adsorption pro-
cess as a function of time. The initial slope at short times
defines the adsorption rate. A comparison with the impinge-
ment rate then allows determining absolute values of the
sticking probability. At long times a dynamic adsorption
desorption equilibrium is obtained (for Tad,2). For (molecular)
first-order adsorption/desorption, t is the mean lifetime before
thermal desorption as defined by the Frenkel equation
t t0 exp Edes =kT [4]
Thus, the activation energy for thermal desorption as well as
the preexponential factor t0 can be determined from variation
measurements of the temperature (in the sub-chapter of NO
surface reactivity we shall provide an example for such mea-
surements). However, if the temperature happens to be too
low, saturation of the surface will occur. The time dependence
of Y will then be largely determined by the details of S(Y) (for
molecular adsorption, the simple Langmuir model assumes
S S0(1  Y); for dissociative adsorption, we receive
S S0(1  y)2). While there are various shortcomings in the
assumptions adopted by Langmuir, the general trend of
adsorption isotherms always resembles that of Figure 4.
Probably one of the most obvious simplifications in
Langmuirs approach is the neglect of interactions between
adsorbed species while coverage increases. Clearly, if no inter-
actions occurred, saturation of the surface would be associated
with a (1  1) layer structure. The general observation is that
this is not the case due to the existence of interaction potentials
cies adsorbed in neighboring sites can repulsively interact via
their dipole moments. Such types of direct interactions are to
be distinguished from indirect interactions mediated through
1 Saturation
Tad, 1
Rads = Sf( )J Equilibrium
qeq
Tad, 2
q
Rads = S.J ~ S0J
0
t Time
Figure 5 Cartoon showing the dependence of coverage versus
temperature and time. The highest coverage is achieved at lower
temperatures.
 Elementary Steps in Heterogeneous Catalysis 11

the metal giving rise to either attraction or repulsion. As a
consequence, the heat of adsorption changes as a function of
the coverage. In most cases relevant to heterogeneous catalysis,
the heat of adsorption decreases due to repulsive interaction
potentials.
Another major assumption in Langmuirs model is that
molecules impinging on occupied sites will get reflected. The
reality is different though. Kisliuk21was probably the first to
change the Langmuir picture by assuming a kinetic scheme
with precursor states. Accordingly, incoming molecules hitting
an occupied site may become trapped with a certain probabil-
ity in an extrinsic precursor state. Subsequent diffusion across
the surface may eventually lead to chemisorption above empty
sites, thus increasing the measured sticking probability. The
kinetic scheme considered by Kisliuk actually contains the
various elemental rate processes including also intrinsic pre-
cursor adsorption above empty sites. The result is an expression
for the coverage dependence of the sticking probability leading
to positive deviations from the Langmuir model if the extrinsic
state desorbs with a probability lower than the summed prob-
ability of the intrinsic state either desorbing or converting into
the chemisorbed state. For more details we refer to the excellent
review of Freund.10
7.02.1.2.3 Structure sensitivity
Another shortcoming in the Langmuir model is the assumption
of an energetically homogeneous surface so as to ensure that all
adsorption sites provide the same sticking properties. While
such a situation may be encountered on well-defined single-
crystal surfaces, real catalyst particles are heterogeneous per se.
Modern surface science techniques can, however, fill this
materials morphology gap so as to provide realistic models of,
for example, metal nanoparticles on oxide support. In an
attempt to mimic the geometry of a real metal particle, the
apex of a field emitter tip (in the absence of a metal oxide
support) may be chosen.22
Figure 6 provides an example of a tip conditioned to form a
hemispherical morphology with a large number of crystallo-
graphically different facets becoming visible in real space with
high, sometimes atomic, resolution. A ball model provides a
guide to the eye in the same figure. Subjecting this particle to
an O2 environment, as in the present case, causes major shape
transformation from hemispherical to polyhedral.
While the overall shape transformation is driven by the
reduction of the surface free energy, the kinetics involved
depend on the facet-specific oxygen sticking and reaction prob-
abilities. In fact, the interaction leads to the formation of a
surface oxide whose local thickness (12 nm) depends on the
surface crystallography. Such plane dependence of the reactiv-
ity is a nice demonstration of the geometric factor in hetero-
geneous catalysis. It also demonstrates that catalysts may
undergo changes in the geometry as well as in the chemical
surface composition (in the present case, a RhO2 surface oxide
is formed)23 during the catalytic run-in.
Figure 7 provides another example for site and facet
distributions at the surface of a supported nanoparticle.
The 3D particle morphology here becomes visible through

-- --
(111) (111)

- -
(101) (101)

- -- - - --
(011) (317) (102) (011) -- (317)
- (137) - -
(111) (317) (111)

- -
(113) (137)
(001) - (001) -
- (113) (137)
(113) -
- (111)
(111) -
(317) (137)
(011) (011)
(317)

(101) (101)

a) (111) b) (111)

c) d)

Figure 6 A particle surface can undergo deep adsorption-induced reconstruction. The figure shows a field ion image of an Rh tip conditioned as a
nanoparticle either clean (a) or after exposure to O2 (b) at elevated temperatures (500 K). Mainly low-index {001} and {111} planes are present in the
reconstructed particle, whereas {011} and {113} planes are altered according to the missing-row type. The flattened corner parts of the polyhedral
particle reveal {137} symmetry. The morphological shape transformation is governed by the reduction of the surface free energy. Adapted from Visart de
Bocarme, T.; Kruse, N. Top. Catal. 2000, 14(1), 3542, 2000, Springer ScienceBusiness Media, LLC 2008.
12 Elementary Steps in Heterogeneous Catalysis

(a)

Step vacancy

A step
3
B step
2 nm
111
2
Step 010
Pt(S)- A step Kink
5(111)(100) 4 11-1

Island [110]
B step
1
3 C
Terrace

B 111/111
edge
1
111/100
edge
(b) A
Pt(S)-4(100)(111)

Figure 7 (a) Restored phase of a 6 nm Pt particle obtained by applying spherical aberration correction and through-focus exit wavefunction
restoration. (b) Best-fitting simulated phase. (c) 3D atomic model used to calculate the best-fitting phase in part (b). The large white arrow indicates the
direction of the electron beam. The inset overlapping parts (A, B) show the crystallographic details of the particle. Reprinted from Gontart, L. C.; Chang,
L. -Y.; Hetherington, C. J. D.; Kirkland, A. I.; Ozkaya, D.; Dunin-Borkowski, R. E. Angew. Chem. Int. Ed. 2007, 46(20), 36833685, with permission
2007 Wiley-VCH Verlag GmbH & Co.

high-resolution electron microscopy by applying exit-wave res-
toration methods.24 Exposing particle surfaces to various types
of gases in the mbar range while imaging with transmission
electron microscopy has enabled researchers to receive infor-
mation on 3D morphologies with lattice resolution of the
particles (see Giorgio et al.,25 for example). It is evident from
these few examples that the simple hit and stick model devel-
oped by Langmuir for an energetically homogeneous surface
does not apply to more realistic catalyst morphologies. The
complexity further increases once it comes to an active
exchange of adsorbed species between the support and the
catalyst particles.
7.02.1.3 Diffusion
Mobility in the adsorbed layer is a prerequisite to any surface
chemical reaction. It ensures that reactants find each other on
the surface of the catalyst. Surface diffusion has been described
extensively in the literature.2629
Conceptually, diffusion is the ability of an adparticle to
jump from one site to the other. Surface lifetimes of adsorbed
species are usually long compared to the characteristic time for
jumping. This applies to chemisorbed as well as physisorbed
species and allows large numbers of sites on the surface to be
visited before reaction or desorption occurs. Diffusion of phy-
sisorbed species actually ensures that sticking probabilities are
kept high under conditions of elevated coverages of the che-
misorbed layer. As discussed earlier in this chapter, this sce-
nario provides another reason for the occurrence of strong
deviations from the simple Langmuir model in which gaseous
species either stick to an empty surface site or are reflected from
an occupied one. Another scenario is depicted in Figure 8.
Here, a chemisorbed species diffuses from one empty site to
another before eventually becoming trapped in a low-
coordination site (a kink) offering higher binding energy
than terrace sites.
The diffusivity is usually parametrized via an Arrhenius
ansatz:
D D0 expEdif =RT [5]
where D0 is the prefactor and Edif the activation energy for
surface diffusion. D0 can be related to the jump length l via
D0 l2 n0 , where n0 can be written as

Kink
Figure 8 A molecule striking a stepped surface: for statistic reasons the landing site is a terrace atom. The resulting adatoms undergo random walk
across the surface so as to become trapped in a site of higher binding energy (in this case a kink).
n0 n expDSdif =k [6]
The activation energy for diffusion, Edif, can usually be
determined by plotting ln D as a function of 1/T since n
(attempt frequency), , and the change in the entropy of diffu-
sion DSdif may be assumed to be slightly dependent on the
temperature in many cases.
Figure 9 provides an example for the diffusion of a single
water molecule along the [010] trough of a TiO2 (110) surface.
Images were obtained by STM and reveal not only the impor-
tance of directional diffusion but also the site preferences due to
local deviations from the ideal TiO2 (subsurface) stoichiometry.
It is clear that the surface coverage must play an important role
in the diffusion of chemisorbed species. Directional transport is
encountered when local differences of the species coverages occur,
due to reaction for example. In the case of nonlinear reactions
and, in particular, oscillatory behavior at low pressures on single-
crystal surfaces, an area pioneered by Ertl and coworkers,30 the
coupling of reaction and diffusion far from equilibrium may lead
to spatiotemporal self-organization. Without diffusion, in the
absence of other coupling mechanisms, only an average behavior,
that is, a constant steady state, would be observable.
Finally, it should be noted that in the case of real catalysts,
the catalyst texture may impose stringent conditions on the
diffusivity of gas-phase reactants and products within the pore
structure that can have drastic consequences on reaction yields.
Dittmayer and Emig31 provided an excellent review of simul-
taneous heat and mass transfer in heterogeneous catalysis. It is
important to realize that transport limitation by diffusion may
also come into play when studying the kinetics of reactions
over single-crystal surfaces at higher gas pressures. Such effects
may become visible at pressures as low as 103 mbar as will be
discussed in context with studies of the CO oxidation.
7.02.1.4 Reaction
The surface chemical reaction involves the turnover from reac-
tants to products and must therefore be considered to be of
central importance to heterogeneous catalysis. As pointed out
by Stolze and Norskov,32 there exist a number of levels in
complexity to describe the surface chemical reaction. From
the viewpoint of a theorist, a description in full dynamics
applying quantum-chemical concepts would probably be
most rewarding. However, this approach is limited to a single
elementary step and will not be applicable to a multistep
Elementary Steps in Heterogeneous Catalysis 13
Terrasse
[10 -
1]
0]
[01
(a) (b)
Figure 9 Time lapse STM images showing a water molecule (red circle)
migrating on the surface. The final adsorption site marked with a yellow
arrow in (a) and the initial adsorption site (yellow arrow in (b)) have the
same appearance, indicative of a subsurface defect. Reprinted from
Aschauer, U.; He, Y.; Cheng, H.; Li, S. -C.; Diebold, U.; Selloni, A. J. Phys.
Chem. C 2009, 114(2), 12781284, with permission 2009, American
Chemical Society.
process involving several intermediates with different lifetimes
many orders of magnitude beyond the picosecond time range.
For the most simple step of a diatomic molecule (O2) disso-
ciating on a single-crystal surface (Ag(110)), see the topograph-
ical STM images in Figure 10. The dissociation process is
triggered by the application of a short electric field pulse
between the probing tip of the STM and the molecularly
adsorbed molecule.
Setting up the rate equations for a multistep process is a
problem in itself if the details of the interaction with the sur-
roundings of the reacting species are to be included. Under
certain conditions, effective values for the rate constant and
neighbor site occupation can be formulated to receive relatively
simple rate expressions.33 Neglecting the interaction with neigh-
boring species under conditions of elevated temperatures may
seem an oversimplification, but it has proved to be very power-
ful for treating many reaction systems. One of the advantages of
this approach is that a large amount of experimental data from
spectroscopic and microscopic observations under well-defined
experimental conditions is usually available to set up and test
the rate expressions. Important ingredients to such rate expres-
sions may also be furnished by theoretical calculations as, for
example, those using density functional theory (DFT).
The remarkable progress made in providing theoretical
insight into surface reactions by DFT calculations under
well-defined conditions will also be demonstrated in the
14 Elementary Steps in Heterogeneous Catalysis
(a) (b) (c)
-
[110]
[001]
(d) (e) (f)
(g) (h) (i)
Low High Low High
Figure 10 STM study of the dissociation of an O2(001) molecule
chemisorbed on Ag(110). The dissociation is triggered by a short electric
field pulse at 13 K. The left column shows topographical images ((a), (d),
and (g)); the center column shows atomically resolved topographical
images ((b), (e), and (h)). Schematic diagrams of the adsorption sites are
shown in the right column ((c), (f), and (i)). A positive voltage pulse on
O2(001) (panel a) led to the formation of either ((a)(c)) nearest-neighbor
oxygen atoms, ((d)(f)) next-nearest-neighbor oxygen atoms, or
((g)(i)) next-nearest-neighbor oxygen atoms. For details of the
experimental procedure along with the chemical functionalization of the
STM tip, consider the original paper. Reprinted from Hahn, J. R.; Ho,
W. J. Chem. Phys. 2005, 123(21), 214702214706, with permission,
2005 American Institute of Physics.
following sub-chapters for certain model systems. The pros-
pects of DFT to provide computer-based catalyst design by
calculating interaction energies of molecules and atoms with
metal surfaces so as to describe their reactivity were recently
discussed by Norskov.34
It is the mean-field approach described above that will form
the basis of the mechanistic considerations reviewed in this
chapter for a selected number of case studies. One may cate-
gorize the surface reaction according to whether one or all of
the species involved in the reaction step(s) are adsorbed on the
surface and to whether the catalyst surface remains inert or
participates by active (but reversible) exchange of atoms with
the reactants. In this spirit we can differentiate between three
types of scenarios named after their original protagonists:
LangmuirHinshelwood,
EleyRideal, and
Marsvan Krevelen.
In the LangmuirHinshelwood mechanistic model
(Figure 11), it is assumed that the reacting species are all
chemisorbed on the surface. Most of the surface catalytic pro-
cesses are considered to proceed this way. The reactants are
supposed to be thermally accommodated which is the case for
surface reactions with long time constants as compared to the
A B
A B A B A B
Figure 11 Schematic representation of the LangmuirHinshelwood
mechanism.
B B
A
A A A B
Figure 12 Schematic representation of the EleyRideal mechanism.
picosecond range (characteristic for thermal equilibration after
several hoppings).
As outlined above, a real catalyst is uniform neither in its
composition nor in its structure as is falsely suggested by Figure 11.
Furthermore, large deviations may be encountered to truly mean-
field conditions. Rather than forming A  B complexes by coop-
erative adsorption, nonmiscible adsorbate phases may form,
implying that the reaction occurs only at the phase boundaries.
While the basic reaction event may still be considered to be of the
LangmuirHinshelwood type, it is clear that the measured kinetics
may become rather involved. This is even the case for apparently
simple reaction systems, like the CO oxidation.
In the EleyRideal model, only one reactant is chemisorbed
on the surface while the other is not. In other words, if a gas-
phase molecule strikes a previously adsorbed molecule, the
collision might lead to reaction, with the product escaping
directly in the gas phase, as illustrated in Figure 12.
Once again, the representation is an oversimplification as
one may consider a B-type species as not fully accommodated
but rather existing in a vibrationally or rotationally excited
state when reacting with A. At the extreme, hot adspecies
which keep a substantial part of their incident translational
energy may be involved in this type of mechanism. It seems at
present that clear evidence for the occurrence of an EleyRideal
(ER) mechanism was only obtained in particular cases, as for
example in the HD formation from a Dad covered Cu(111)
during H2 impingement36 or the reaction of CO with atomic
oxygen on Pt(111).3739
To demonstrate the complexity of an apparently simple
LangmuirHinshelwood-type reaction, we provide a brief
account of water formation from hydrogen and oxygen on Pt
(111) and on Rh conditioned as a nanosized tip (resembling a
single metal grain; see Figure 13). Both reactants chemisorb
dissociatively on the surfaces of these metals. Using STM,
Sachs et al.40 demonstrated in low-pressure/low-temperature
studies with atomic-scale resolution the occurrence of reaction
sequences eventually giving rise to ring-like chemical waves of
OHad expanding into Oad-rich regions. Adsorbed water mole-
cules can diffuse through the region of high OHad concentration
so as to react with Oad according to Oad 2H2Oad ! 2OHad
from where H2Oad is produced via OHad Had ! H2Oad. It is
remarkable that this complicated scenario could be successfully
 Elementary Steps in Heterogeneous Catalysis 15

modeled in a mean-field approach reproducing the experimen-
tally observed length scale and wave front velocity.41
At temperatures allowing the desorption of water product
molecules, the rate-determining step seems to be the Oad Had !
OHad reaction.42,43 The reactive desorption of water may then
involve either Had OHad ! H2Og or 2OHad ! H2Og Oad.
For the same reaction on an Rh nanosized particle, low-
pressure studies at reaction temperatures between 500 and
550 K demonstrated oscillatory behavior in the field ion
microscope (FIM) (Figure 14).44 The observed spatiotemporal
pattern formation could be correlated with the reversible
surface oxidation of the metal. Rather than invoking a
standard reaction-diffusion mechanism, the oscillatory pat-
terns could be attributed to the different catalytic properties
of the various nanosized facets simultaneously exposed by the
metal particle. A mean-field theoretical approach with a

O OH OH H 2O
(a) (b) (c)

5 nm

65 nm (d) (e) (f)

O OH
Figure 13 Successive STM images (17  17 nm2) during exposure to 8  109 mbar H2 of an O-covered Pt (111) surface at 112 K. Dark dots in A mark
O-adatoms in the initial 2  2 structure. Bright features indicate the growing OHad patches. In C the area is mostly covered with ordered OHad domains,
where the bright fuzzy patches covered by H2O develop. DF (220  220 nm2) show larger-scale images at 112 K and 2  108 mbar H2. In D, bright
dots indicate small OH islands which exhibit enhanced concentrations in a ring expanding into the O-rich ring region, leaving H2O behind. Thin lines are
atomic steps. Reprinted from Sachs, C.; Hildebrand, M.; Volkening, S.; Wintterlin, J.; Ertl, G. Science 2001, 293(5535), 16351638, with permission,
2001, AAAS.

10 nm
011
111
012113
001

(a) 0s (b) 13 s (c) 36 s

011
111
012
113
001

(d) 0s (e) 13 s (f) 36 s

Figure 14 Series of FIM micrographs covering the complete oscillatory cycle and the corresponding time evolution of the subsurface oxygen
distribution on a logarithmic scale as obtained within a kinetic model of the field emitter tip. Starting from a surface in the quasi-metallic state (a and d),
an oxide layer invades the topmost plane and grows along the {011} facets forming a nanometric cross-like structure (b and e). The oxide front
spreads to finally cover the whole visible surface area (c and f). The temperature, electric field, and partial pressures of oxygen are as those in
Figure 6. For the subsurface site occupation, the white areas indicate a high site occupation value whereas the dark areas indicate a low site
occupation value. Reprinted from McEwen, J. -S.; Gaspard, P.; de Bocarme, T. V.; Kruse, N. Proc. Natl. Acad. Sci. 2009, 106(9), 30063010, with
permission, 2009 National Academy of Science of USA.
16 Elementary Steps in Heterogeneous Catalysis
A Olat O-Ogas
A A
OOOOO OOOOO O OOO O OOO OOOOO
Figure 15 Schematic representation of the Marsvan Krevelen
mechanism.
LangmuirHinshelwood-type mechanism including reversible
subsurface oxygen penetration enabled a nearly quantitative
description of patterns and oscillations periods.
In the case of metal oxide catalysts, a Marsvan Krevelen
mechanism may be encountered. In this case, the catalyst
surface is assumed to participate actively in the surface catalytic
process. A redox cycle on the catalyst has to be invoked so as to
make the overall process catalytic (with no consumption of the
active catalytic sites). The main features in this mechanistic
model are that the product of the reaction desorbs from the
surface with one or more constituents of the catalysts lattice.
Thus, the catalyst is no longer an inert substrate but rather a
reaction partner. The classical steps of the Marsvan Krevelen
mechanism for an oxidation reaction are shown in Figure 15.
The reactant A adsorbs on the surface and reacts with an
oxygen atom of the catalyst. The product desorbs and the
resulting O vacancy is rapidly refilled by an oxygen atom. The
details of the oxygen precursor molecule dissociation, after
adsorption from the gas phase, are frequently unclear and
may involve several steps. Moreover, the transient occurrence
of a vacancy may lead to a surface reconstruction and change in
the catalytic properties.
7.02.1.5 Desorption
Desorption is the final stage in the catalytic cycle. Usually the
desorption process is not rate determining; otherwise self-
poisoning of the catalytic process would occur. Product species
with full thermal accommodation to the surface temperature
exhibit a Boltzmann distribution of their translational, vibra-
tional, and rotational degrees of freedom and desorb with a
cosine angular distribution. Evidence for hyperthermal product
desorption with a cos Yn distribution (n up to 10) was, for
example, observed and summarized by Matsushima et al.45 for
the CO oxidation from various noble metal single-crystal sur-
faces. While a cosine distribution of CO2 product molecules
desorbing from Pd(111) was observed in early studies,46 strong
anisotropies in the spatial distribution of CO2 were reported47
for the strongly corrugated Pd(110) plane and explained on the
basis of local reaction site geometries considering Oad to be fixed
and COad to be mobile. If the transition state leading to CO2
formation is bent, considerable excitation of internal degrees of
freedom in the desorbing (linear) molecule may be encountered.
Indeed, Nakao et al.48 demonstrated the asymmetric stretch to be
more strongly excited than the other CO2 vibrational modes.
Most of the experimental data on the thermal desorption
kinetics of molecules from model surfaces as well as real cata-
lysts were provided by temperature-programmed heating. It
should be mentioned that in many cases studies of the desorp-
tion kinetics also provide information on the adsorption pro-
cess. If the adsorption process sketched in Figure 1 is fully
reversible and the same transition state accessible from either
side, while keeping detailed equilibrium with the surface, sta-
tistical models can be applied to interpret the data. Although
there are limits to such an approach, see above, we shall make
use of it in some examples of the following sections. A number
of reviews have appeared on the subject and are based on the
evaluation of temperature-programmed desorption (TPD) and
reaction spectroscopy data. Indeed, these techniques may be
regarded as methodical workhorses in describing the global
kinetics of desorption and catalytic reaction and will certainly
continue enjoying wide application in the community.
7.02.2 CO Adsorption and Oxidation
7.02.2.1 General Principles
A quick search through any scientific database using CO oxi-
dation as the keyword yields more than 50 000 titles, out of
which more than 30 000 were published in the last 10 years.
From this it is obvious that a comprehensive literature review
on this subject would be a challenge in itself. Even if we only
concentrate on the kinetic and mechanistic aspects of this
issue, it would be impossible to provide a complete picture.
Therefore, we shall select several aspects of CO adsorption and
reaction with oxygen which we think are instructive and useful
with regard to a fundamental understanding of elemental reac-
tion steps in heterogeneous catalysis. In particular, we shall
examine selected metal catalysts either supported or conditioned
as single crystals and then turn to some metal oxide systems. The
continuing interest in studies of CO oxidation to probe funda-
mental concepts in catalysis was recently highlighted in the
reviews by Freund et al.49and Royer et al.,50 the former discussing
general aspects at all levels of complexity and the latter providing
a summary of the achievements in studies with oxidic catalysts.
7.02.2.2 CO Adsorption
Carbon monoxide is isoster with N2 and has Lewis donating
(through the 5 s molecular orbital) as well as accepting proper-
ties (through the empty antibonding 2p* orbital). Its chemi-
sorption on transition-metal surfaces is the most intensively
studied of all adsorption systems.5153 Depending on the sur-
face coverage, CO can adsorb in hollow, bridge, or on-top sites.
On particle surfaces, step sites are first occupied due to a gain in
binding energy. Even if the molecules impinge on terrace sites,
the mobility of the adsorbed species usually provides access to
the steps. The underlying process is sketched in Figure 8. Ther-
mal desorption follows the reverse path (principle of the
microscopic reversibility), that is, the molecules are first
detached from the steps before desorbing from the terraces.
Molecular beam relaxation studies of CO on stepped Pt(111)
surfaces allowed determining the dynamics of the adsorption
desorption process at overall low coverages.54 The temperature-
dependent measurement of surface mean lifetimes then led to
adsorption energies (146 kJ mol1) largely determined by step-
site binding.
The chemisorption of carbon monoxide on nickel surfaces
is one of the most frequently studied gas/solid systems in
surface science. Stepped surfaces of single crystals or catalyst
particles are however also prone to the formation of adsorbed
subcarbonyls, Ni(CO)x(x 1  3).55 Quite surprisingly, such

Ni(CO)x species were found to form under surface science
conditions of rather low gas pressures. Later, it was demon-
strated that gaseous Ni(CO)4 was the final product of a reac-
tion sequence involving subcarbonyl intermediates.56 Detailed
kinetic research demonstrated a CO* compression layer to be a
prerequisite for Nisubcarbonyl formation. Furthermore,
pulse experiments allowed the measurement of the relaxation
times for adsorbed Ni(CO)x to reach either saturation or
steady-state conditions. In this way, the rate-determining step
of carbonylation could be identified as being associated with
Ni(CO)2 formation:
NiCO* kink CO* ! NiCO*2 kink [7]
The overall process is shown schematically in Figure 16
along with a video sequence. Accordingly, CO adsorption
takes place at terraces and steps (for the sake of clarity, only a
few CO* molecules are considered in an otherwise saturated
overlayer). Nidicarbonyl is formed by adding terrace-site CO*
to an occupied lowest-coordination kink site (Ni(CO)kink).
This is associated with NiNi bond breaking and liberation
of Ni(CO)2*. Rapid reaction to Ni(CO)3* in the terrace layer
and, finally, vaporization of Ni(CO)4 complete the reaction
process. The liberation of a Ni(CO)2* species usually repro-
duces a kink site, that is, the above reaction step occurs
repetitively until either an open chain of Ni atoms is used up
or inhibition comes into play as, for example, by carbon depo-
sition due to COkink decomposition. Clearly, this reaction
mechanism must lead to structural changes of the surface.
According to observations by Schmidt et al.,57 kink sites of a
Ni tip surface were indeed annihilated during the reaction with
CO. In particular, the hemispherical Ni tip morphology was
found to develop low-index planes of a polyhedron. The phe-
nomenon of subcarbonyl formation on stepped transition-
metal surfaces (without final vaporization) was also observed
on Co,58 Ru59 or Rh60 and others. Since Me(CO)x* are mobile
surface species, they are most probably involved in reconstruc-
tion processes during run-in of catalytic processes.
A frequently studied supported catalyst system is Rh/Al2O3.
Early infrared studies indicated the formation of gem
Figure 16 Model of Ni carbonylation: an adsorbed Nidicarbonyl
species is formed at a kink site; after diffusion into the terrace region
which is associated with NiNi bond breaking, Ni(CO)3,ad and, finally, Ni
(CO)4,g are formed; for clarity, only a few CO admolecules of the
incommensurate compression structure are shown. Reprinted from
Medvedev, V. K.; Borner, R.; Kruse, N., Surf. Sci. 1998, 401(1),
L371L374, with permission, 1998, Elsevier.
Elementary Steps in Heterogeneous Catalysis 17
dicarbonyl with the Rh atom in the 1 oxidation state. Basu
et al.61 suggested hydroxyl groups of the support surface to be
directly involved in the formation of RhI(CO)2. Strong evi-
dence for a considerable disturbance of the Rh coordination
number due to the interaction with CO was obtained in
extended x-ray absorption fine structure (EXAFS) studies.62
There, nanosized Rh particles were seen to dissolve, ultimately
leading to atomically dispersed RhI sites. It may therefore be
suggested that Rh(CO)2 species are produced on Rh particles
followed by spillover to the support so as to form
Al  O  RhI(CO)2 entities. Surface science work in the 1990s
provided a more refined picture of Rh gemdicarbonyl forma-
tion. For example, Evans et al.63 reported these species to be
formed on hydroxyl-free TiO2 (110) surfaces, presumably via
stabilizing the dispersed Rh by oxygen vacancies. A CO-
induced disruption of nanosized Rh particles on the same
support surface was also reported by Berko and Solymosi on
account of STM measurements.64The authors found size-
dependent effects, that is, Rh clusters of 12 nm size were
more readily disrupted than nanoparticles of 35 nm size
while no such process was observed for Rh particles larger
than 810 nm.
Combined STM, infrared reflection absorption spectroscopy
(IRAS), and XPS work of small and ultra-small Rh particles on
thin alumina films prepared via oxidation of a NiAl(110) single-
crystal surface provided further insight into the various reaction
scenarios.65,66 Experiments with isotopic 12CO and 13CO mix-
tures demonstrated small Rh clusters with nine atoms on aver-
age to bind CO (at 90 K) as Rh(CO)2 at oxide point defects.
While the formation of multiple-bonded CO seems to be
well documented for a number of noble metals, it is somewhat
less so for base transition metals. Magg et al.67yet reported a
striking example of di-and tri-carbonyl formation when
adsorbing CO at low temperatures on alumina-supported
V nanosized particles. A strong particlesupport interaction
was observed leading to partial incorporation into the support
and oxidation of the particles.
Another example of transition-metal

subcarbonyl forma-
tion is Co. Using a single 1010 oriented Co particle condi-
tioned as a field emitter tip, Kruse et al.68observed Co(CO)*2,3
formation by inspecting the dynamic interaction of gaseous
CO with 200 atomic sites of the (step-containing) (0001)
plane. They also observed CO dissociation which was in com-
petition with carbonyl formation. The time dependence of the
processes is shown in Figure 17. While chemisorbed CO
increases linearly with time in a Langmuirian-type adsorption
with sticking probabilities between 0.5 and 0.9 at overall low
coverages, higher-order processes are involved in both
Co(CO)*2 (and Co(CO)*3 , not shown) and carbon/oxygen
(carbidic and oxidic) formation at 450 K. In fact, the initial
slope, dln(Y)/dln tR 2, as observed for the latter, is in accor-
dance with a reaction scheme of CO adsorption first, that is,
CO * ! CO*, and dissociation subsequently, that is,
CO* * ! C* O*.
Infrared spectroscopy is commonly used to distinguish
between molecular and dissociative CO adsorption. Interest-
ingly, research with isolated clusters during the past years dem-
onstrated these to behave similarly to extended surfaces.69 The
transition from dissociative to molecular adsorption is fre-
quently considered to closely follow a diagonal line through
18 Elementary Steps in Heterogeneous Catalysis
COchem.
100
Co-carbides
50
Co-oxides
Co(CO)2
20
Ions per 1000 pulses
10
5 erations on specific model catalyst surfaces, with an emphasis
2 periodic table of transitions metals as defined by Broden
1 As a starting point, it is important to realize that indepen-
0.5
0.2
10-3 10-2
Time (s)
Figure 17 Time dependence of CO adsorption, Co(CO)2 formation, and
CO dissociation (Co-carbides/oxides) at low coverages (Y < 0.1)
Approximately 200 sites of the step-containing (0001) area of a Co field
emitter tip are probed. The ordinate scales linearly with the coverage Y.
Reprinted from Kruse, N.; Schweicher, J.; Bundhoo, A.; Frennet, A.; De
Bocarme, T. V. Top. Catal. 2008, 48(14), 145152, with permission,
Springer ScienceBusiness Media, LLC 2008.
the periodic system of the transition metals.70Accordingly, early
transition metals are supposed to bind CO dissociatively while
the later ones are incapable of breaking the triple bond in CO.
The transition from dissociative to molecular binding can be
understood from the fact that moving to the left in the periodic
table of the elements results in a rise of the Fermi level and of the
d orbital diffuseness. The observation of subcarbonyl formation
on early transition metals shows, however, that care is advised in
applying these simple concepts without considering specific
aspects such as CO coverages, surface temperature, particle
size, metalsupport interaction, and charge state.
7.02.2.3 CO Oxidation over Transition-Metal Surfaces
The reaction is exothermic (DH 298  282.6 kJ mol1) and prac-
tically irreversible up to 1500 K (DG 298  256.7 kJ mol1;
DS 298  86.5 J mol1K1). Using a catalyst enables the reaction
to occur through activation of the molecules. A Langmuir
Hinshelwood (LH)-type mechanism is usually considered to
describe the formation of CO2:
CO * ! CO*
O2 2* ! 2O*
CO* O* ! CO2 2*
Only the first step is truly elementary. O2 chemisorption is
dissociative and the CO2 product, once being formed from
chemisorbed CO molecules and O atoms, is liberated into the
gas phase due to its weak binding with the catalyst surface at
reaction temperatures. The elementary processes at an atomistic
level and the complex reaction dynamics in space and time have
always been and continue to be of interest for a fundamental
understanding of heterogeneous catalysis. A coincidence for
such research is the fact that the reaction occurs over a wide
range of 13 orders of magnitude of pressures.71
It is not our intention to provide here a full review of the
various achievements from fundamental studies at all levels of
complexity. We shall rather limit this section to basic consid-
on the insight gathered since the turn of the century in research
with metals located right of the dividing line through the
et al.70
dent of whether a supported metal catalyst or a single-crystal
surface is considered, a strong asymmetry is observed in the
adsorption of the reactants. While CO can be incorporated in a
complete oxygen overlayer structure (for the particular case of
the Ru(0001)(2  1)O phase, see below), the reverse is not
the case. Thus, CO2 formation may become inhibited by accu-
mulating CO* coverages. In order to maintain a steady CO2
production rate under flow conditions, the surface temperature
has to be increased so as to cause some of the CO* to desorb
(400450 K depending on the metal) and enable gaseous O2
to compete for the free adsorption sites. Note in this respect
that O2 molecules are at a disadvantage as compared to CO:
they need an ensemble of neighboring empty sites in order to
accomplish dissociative chemisorption while CO does not. The
(apparent) activation energy for the LH step leading to
adsorbed CO2 is coverage dependent on Pt, Pd, or Rh surfaces
and usually larger than the respective CO2,ad desorption
energy. Stable regimes of the reaction are shown schematically
in Figure 18. In region (I) the rate increases with increasing CO
pressure, while in region II further increasing CO pressures are
kinetically inhibiting and oxygen dissociative chemisorption
becomes rate limiting. This scenario is in agreement with the
frequent observation of a steady-state kinetic order dependence
of 1 with respect to O2 and of 1 with respect to CO. Region
III delimits multiple steady-state occurrences. Hysteresis can
now be found, that is, the CO2 yield depends on whether the
CO pressure is increased or decreased while moving from
region I to II and vice versa. The interplay of strong adsorption
and liberation of sites by reaction may actually give rise to
PCO2
I
[8] II
[9] III
[10]
PCO
Figure 18 Stable (I and II) and instable (III) regions in the isothermal
CO oxidation kinetics at constant O2 pressure.

complex behavior. Local variations in the reaction events will
cause the coverages to depend on time and space. As a conse-
quence, surface diffusion will be initiated and serve as a cou-
pling mechanism between different parts of the surface.
While the above considerations are general and, to some
extent, applicable to other reaction systems, they are also
important ingredients in the occurrence of spatio-temporal self-
organization. Under certain conditions even oscillatory reaction
behavior may be observed. The first observations of such rate
oscillations were made by Hugo and Wicke72,73 in studies of the
CO oxidation with supported Pt catalysts. Subsequent work by
Ertl and coworkers laid the scientific basis of nonlinear reaction
dynamics in surface reactions. Using mainly model systems
under surface science conditions, Ertl et al.30,74,75 established
the relevant mechanisms governing isothermal rate oscillations
and spatio-temporal pattern formation. It is not our intention to
review this work; however, it is felt that a chapter on elementary
steps in heterogeneous catalysis would be incomplete without
including the phenomenological description of at least one case
study: CO-oxidation on Pt (110) (see also Figure 19).
Among the three low-index planes only the close-packed
Pt(111) plane is stable in its (1  1) bulk termination while the
Pt (110) plane exposes atoms in lower coordination and
undergoes reconstruction into a (1  2) missing-row geometry.76
This reconstruction is influenced by adsorption: at critical cover-
ages adsorbate-induced phase transitions may be encountered.
An oscillatory cycle can be mounted by considering that:
1. CO coverages of YCO 0.2 cause the (1  2) missing rows
to reconstruct back to bulk-terminated (1  1)77 and
2. oxygen sticking probabilities on (1  2) are smaller
(sO2 0:3  0:4) than on (1  1) (sO2 0:5  0:6).30,78
The occurrence of temporal oscillations in the CO2 produc-
tion can then be rationalized by starting with a 1  2 recon-
structed (110) plane which, at properly chosen reactant
pressures, exhibits a low oxygen sticking probability and there-
fore low reactivity. Increasing the amounts of adsorbing CO
will lift the reconstruction so as to form 1  1 patches at critical
coverages. Due to the higher oxygen sticking probability on
these patches the CO2 production rate increases. The consump-
tion of excess COad will then cause YCO to fall short and the
surface to reconstruct back to the (1  2) geometry. The cycle is
now complete. The importance of reversible phase transitions
Pt (110)
CO covered Clean
- in addition to the LH scheme, was first employed by Bassett
(110)
(001)
11 12
Figure 19 Ball model illustrating the CO-induced 1  1 to 1  2 surface
phase transition of Pt(110). The different oxygen coefficients of the
two phases are responsible for rate oscillations during catalytic CO
oxidation. Reprinted from Imbihl, R.; Ertl, G. Chem. Rev. 1995, 95(3),
697733, with permission, 1995 American chemical society.
Elementary Steps in Heterogeneous Catalysis 19
is also implicitly indicated by the fact that the stable Pt(111)
surface does not show rate oscillations.
Time series under (isothermal) conditions in which the
particle pressure of O2 was kept constant while that of CO
was varied in small increments caused a transition to
chaos through a sequence of period doublings known as a
Feigenbaum scenario.79 In order to produce temporal varia-
tions of the integral reaction rate, the different surface regions
must be subject to a coupling mechanism. For the chosen
conditions, coupling by reaction diffusion and through the
gas phase (global coupling) must be considered (for supported
catalysts, isothermal conditions cannot be maintained and a
coupling via heat conductance may prevail). The coupling
via reactiondiffusion becomes manifest in the appearance
of spatio-temporal pattern formation as observed in photo-
emission electron microscopy (PEEM) experiments at low pres-
sures or with ellipsometry for surface imaging (EMSI) and
reflection anisotropy microscopy (RAM) at higher pressures
(up to atmospheric). If only diffusion coupling is present
typically target patterns and rotating spiral waves are observed.
Even for conditions under which the integral reaction rate is
constant and global coupling via the gas phase enforces syn-
chronization, spatio-temporal patterns remain observable. This
is shown in Figure 20 for elliptical target patterns on an oscil-
lating background. Due to the (110) surface geometry, diffusion
is anisotropic and causes the elliptic deformation; the pattern
propagation preferentially occurs in direction of the missing
rows since the CO diffusion is faster along the troughs than
perpendicular to them. The anisotropy of the surface may also
give rise to the occurrence of solitary pulses propagating perpen-
dicular to the missing row direction. The beauty of this experi-
mental work is brought to full effect by a detailed mathematical
description reproducing all types of pattern formation.
The important mechanistic ingredients so far taken into
account are those described by the (local) LH reaction scheme
(see above), surface diffusion (of mainly CO*), and surface
reconstruction. There is, however, at least one additional pro-
cess which may come into play, preferentially at higher pres-
sures and temperatures: (reversibly) oxygen subsurface
diffusion. Clearly, the accumulation of surface oxygen atoms
may lead to a penetration below the surface, especially on an
atomically open, that is, reconstructed surface, with the con-
sequence that the oxygen sticking probability decreases. This,
in turn, gives way to enhanced CO adsorption. An oscillation
cycle may now be constructed by assuming subsurface O atoms
to segregate back to the surface so as to produce CO2 by
reaction with CO* and to prepare a surface prone again to
enhanced oxygen dissociative chemisorption. Such a scenario,
and Imbihl81 to explain rate oscillations in the CO oxidation
over Pd (110). Subsurface oxygen can also occur on Pt(110)
under low pressures and at higher temperatures than those
adjusted in the experiments leading to Figure 20. Standing
waves in the form of stripes of rhombic cells observed under
these conditions were successfully described by theory once
subsurface oxygen was taken into account.82
The situation may even become more complex when mov-
ing to higher (atmospheric) pressures. Since oxygen subsurface
diffusion may be regarded a precursor step to oxidation, the
question may be asked if Pt oxides can be formed in a
20 Elementary Steps in Heterogeneous Catalysis

Figure 20 Sequence of PEEM images (200  300 mm2) recorded at intervals of 4.1 s (30 s between the last two images) for the CO oxidation on a
Pt(110) surface showing the evolution of target patterns; the bright parts are CO-covered and the dark ones O-covered, reflecting the different
work functions associated with these adsorbates; conditions: T 427 K, PO2 3.2  104 mbar, PCO 3.0  105 mbar. Reprinted from Jakubith,
S.; Rotermund, H. H.; Engel, W.; Von Oertzen, A.; Ertl, G. Phys. Rev. Lett. 1990, 65(24), 30133016, with permission, The American Physical
Society 1990.

reversible manner with time constants comparable to typical
oscillation periods. The answer is yes, they can, and this let
authors propose an oxide model to explain oscillatory
behavior.83 In fact, later on, the application of in situ x-ray
diffraction in studies with supported Pt catalysts provided
clear evidence of the reversible formation of PtO and
Pt3O4.84 Theoretical modeling of the transition to oxide and
the resulting CO oxidation kinetics showed good agreement
with the experimental observations.85 We will finally mention
that a reversible surface oxidation was also involved in the
kinetic oscillations of other systems as, for example, the pro-
duction of water from hydrogen and oxygen on Rh mentioned
in the introduction. In these studies, an Rh nanosized particle
was conditioned in the form of a tip and imaged with nano-
scale lateral resolution by FIM. A mean-field description was
used to successfully simulate the reaction behavior, in which
not only the reversible surface oxidation but also the fast
hydrogen diffusion played an important role.
Field emission techniques were also employed to study the
CO oxidation on Pt at low pressures. For example, Suchorski
et al. addressed the question of fluctuation-induced transitions
on the small facets of a Pt tip by field electron microscopy
(FEM).86,87 Measurements in the vicinity of the Pt(110) facet
showed fluctuations of the FEM current to be spatially well
correlated. On Pt(211), the spatial correlation decayed rapidly
when moving across surface steps terminating the facet. It must
be concluded that steps either slow down CO surface diffusion
or act as sites of enhanced chemical reaction.
As we have seen, a seemingly simple reaction may give rise
to complex behavior through the inclusion of cooperating
effects along with surface restructuring and chemical alter-
ations of the surface. A coherent picture of a variety of
phenomena has nevertheless been developed both experimen-
tally and theoretically for isothermal, low-pressure reaction
conditions. This is by far less so when moving to high pres-
sures. For example, Hendriksen et al.8890 used high-pressure
STM to study the CO oxidation on Pt(110) and Pd(100). The
authors claimed that a stable oxide film formed under O2 rich
conditions and caused higher activity than the pure metal.
Hysteresis effects were found as well, however, while the
behavior was clockwise on Pt(110); it was anticlockwise on
Pd(100). We recall that clockwise hysteresis would translate
into a picture where, while cycling the CO pressure from low to
high and back to low, the (highly reactive) oxide would form at
lower CO relative pressures than those necessary for its
removal. The Pd(100) sample also showed oscillatory behav-
ior, while Pt(110) did not under these high-pressure condi-
tions different from reports by others (see above). The
authors claimed that whereas the reaction on the metallic
surfaces followed LH kinetics, the oxides were (more) active
via the Marsvan Krevelen mechanism, in which CO molecules
withdraw O lattice atoms and oxygen subsequently refills the
vacancies thereby created. These results were in clear disagree-
ment with the traditional picture of the relevant mechanistic
steps and active phases under steady-state and oscillatory con-
ditions. Objection was therefore not a long time coming.
Gao9194 studied a number of single-crystal surfaces includ-
ing Pt(110) and Pd(100) from low to high pressure and
reported mass transport limitation by diffusion, the onset of
which depended on the specific conditions of reactor design,
sample size/texture, and total and relative pressures of reac-
tants. In particular, they defined three regimes of the reactivity
at high pressures. In the first regime, the reaction temperature
is low (T  575 K); the surface is CO-covered and the reaction

rate is limited by CO thermal desorption. This scenario is
similar to low-pressure conditions. The second regime is tran-
sient and may lead to high reaction rates. This regime is not
always easy to control as the exothermicity of the reaction may
cause local overheating. The surface is oxygen rich under these
conditions. In the third regime, the CO2 formation rate is
independent of the reaction temperature. At the high temper-
atures governing this regime, the reaction is either mass trans-
fer limited (imposed by CO diffusion) or limited by the
reduced reactivity of the oxidized surface. The authors con-
cluded that the CO oxidation displayed no discontinuities
when changing pressures from low to high, that is, no pressure
gap was observed for the reaction on these Pt and Pd single-
crystal surfaces. With respect to the conclusions reached by
Hendriksen et al.8890 they demonstrated that the transport
limitation along with the transient regime of exceedingly high
reactivity escaped the attention of these authors. For Pt(110),
applying very similar conditions of low temperature (425 K)
and high pO2 =pCO ratios, no oxidation of the surface was found;
hence, the Marsvan Krevelen mechanism claimed by Hen-
driksen et al. was not in operation in the measurements by
Gao and Goodman.
The different kinetic regimes of CO oxidation are demon-
strated in the Arrhenius plots of Figure 21. The highest (tran-
sient) reaction rates are obtained at the cross-over where the
CO inhibition regime at low temperatures meets the mass
transfer-limited regime at high temperatures. These high rates
are indicated by arrows for Pt (110) and Pd (100) (according to
Temperature (K)
900 800 700 600
103
CO2 TOF (molecule site-1 s-1)
102
101
Pt (110)
Pd (100)
100
Ru (0001)
101
1.0 1.2 1.4 1.6
1000/T (K-1)
Figure 21 Arrhenius plot of CO2 formation over Pt(110), Pd(100), and
Ru(0001) under stoichiometric reaction conditions (Ptotal 12 Torr,
O2/CO 1/2). Adapted from Gao, F.; Goodman, D. W. Phys. Chem. Chem.
Phys. 2012, 14, 6688; Gao, F.; Wang, Y.; Cai, Y.; Goodman, D. W. J. Phys.
Chem. C 2009, 113(1), 174181.
Elementary Steps in Heterogeneous Catalysis 21
Gao et al.93). Data for Ru (0001) are included in the same
figure and indicate that the high temperature regime in this
case is not mass transfer-limited but rather due to decreasing
mean lifetimes of CO molecules on a surface permeable for
adsorbed oxygen atoms. According to the arguments put
forward by Goodman et al.,95 the catalytically active Ru
(0001)(1  1)O phase is prone to oxygen diffusion into the
subsurface region.96 It should be mentioned here that the CO
oxidation on Ru model catalysts is the subject of a vigorous
debate in the literature. While it is not our intention to reveal
the details of this debate, we consider the Ru case as most
appropriate for an inspection, at various levels of complexity,
of the fairly complicated and interfering elemental processes
when moving from low to atmospheric pressure conditions of
the reaction. The various viewpoints in the ongoing debate
were published by its major protagonists.97,91
We first address the fundamental question of where the
energy for the bond rearrangement in the CO*  O* encounter
to form CO2 comes from and how fast this energy flows. Ru
metal has electronic and phonon degrees of freedom from
which energy can be transferred into the vibrational and elec-
tronic states of the adsorbed species. To distinguish between
the different transfer processes, that is, electron- versus
phonon-driven, ultra-short laser pulse excitation (pump)
was used along with a second pulse (probe) allowing the
time evolution of the reaction to be followed. For the experi-
ments a Ru(0001)(2  1)O surface was prepared and
exposed to CO up to saturation at 100 K. We note that under
UHV conditions this co-adsorption phase does not produce
CO2 by conventional heating; only CO is being thermally
desorbed this way. The two-pulse correlation measurement
allowed the unraveling of the timescales and mechanisms of
500
both desorption and reaction. It was found that the oxidation
of CO is caused by coupling to electrons, whereas the desorp-
O2/CO = 1/2
tion of CO is caused by coupling to phonons.98 The timescales
of the two competing processes differed by almost an order of
magnitude. While the ultra-fast response for the oxidation,
associated with 3 ps FWHM, was carried by hot electrons fol-
lowing excitation, the desorption took 20 ps implying cou-
pling to phonons. A quantitative analysis of the excitation
mechanism along with a theoretical description of the energy
transfer from the laser-excited surface to the adsorbed reactants
was subsequently provided on the basis of frictional coupling
between the electron and/or phonon heat bath to the harmon-
ically vibrating adsorbate.99 That the oxidation reaction was
electron-mediated and the desorption phono-mediated was
also verified in measurements with isotopically labeled oxygen.
The O-isotope effect finally demonstrated that the activation of
the RuO bond by electrons is the rate-determining step in the
oxidation reaction. For a more detailed discussion, we refer to
the sub-chapter on ultra-fast dynamics by Wolf et al.49
Most of the catalytic studies using Ru metal were so far con-
ducted with (0001) single crystals. As mentioned above, covering
1.8 2.0 2.2 a (0001) surface with a 0.5 monolayer (ML) of oxygen produces a
(2 1) phase rendering it completely inactive. Higher oxygen
coverages cannot be produced under net vacuum conditions.
Obviously, the strongly chemisorbed oxygen prohibits the reac-
tion with CO.100 Note that this behavior is very different from
other noble metal surfaces such as Pt, Pd, or Rh (see above). On
the other hand, when moving to near-atmospheric pressures a
22 Elementary Steps in Heterogeneous Catalysis

(1 1) phase corresponding to 1 ML oxygen can be produced, Obr O3f

as originally shown by Peden and Goodman.101 The respec-
tive Ru(0001)(1 1)O phase seems to be highly active;
however, doubts were cast as to the detailed interpretation of Rucus
the results102 since the CO binding energy on this surface is
quite low (<20 kJ mol1)92 to be reconciled with a Langmuir
Hinshelwood or even an ER mechanism.103,104 More recently,
[110] [001]
DFT calculations showed that minority step sites on Ru(0001)

(1 1)O may be responsible for the high CO oxidation [110]
activity.105 Careful kinetic measurements along with (a)
polarization-modulation infrared reflection absorption spectros- CO
copy (PM-IRAS) by Gao and Goodman91 arrived at the same
conclusion. Interestingly, the CO binding energy in these defect
sites (
68 kJ mol1) was found to be very similar to the calcu-
lated ones. Moreover, it seems that evidence was obtained for
carbonyl formation, Ru x (CO)y (where x and y may vary), which
should be associated with step sites for steric reasons. We note [110] [001]
that the low CO binding energy in such sites is most probably

influenced not only by the close proximity of co-adsorbed oxygen [110]
(b)
but also by adjacent CO bound to the same site. The effect would
be expected to increase with increasing y values. A rather compli- Figure 22 (a) Ball-and-stick model of the bridging O-terminated
cated infrared spectrum in the CO stretching region was also RuO2(110) surface. Large (green) balls represent O, and small (blue, red)
reported by Kim et al. in their studies of the CO oxidation over balls represent Ru atoms of RuO2(110). A highly active coordinatively
Ru(109).106 This surface is intrinsically stepped and demon- unsaturated Ru atom (red atom, cus) as well as bridge-bonded and
strates reactivity even under net vacuum conditions (which is threefold coordinated O atoms are indicated by arrows. (b) Ball-and-stick
model of the bridging O-terminated RuO2(110) surface on which
not the case for Ru(0001)) and sub-ambient temperatures.
additional CO is adsorbed. Reprinted from Over, H.; Kim, Y. D.;
The most prominent case for high CO oxidation activity of an Seitsonen, A. P.; Wendt, S.; Lundgren, E.; Schmid, M.; Varga, P.;
oxide surface is provided by RuO2 (110). Actually, a thin film of Morgante, A.; Ertl, G. Science 2000, 287(5457), 14741476, with
12 nm thickness grown epitaxially on Ru(0001) was shown by permission, 2000 by the American Association for the Advancement
Over et al.107 and Kim et al.108 to be most active. However, under of Science.
net oxidizing conditions and sufficiently high temperatures this
film may grow to form bulk RuO2 which is inactive.109,110 The
phase diagram of Ru in a CO/O2 atmosphere was calculated by
Reuter and Scheffler111 and shows that under realistic reaction The above discussion has shown that pure RuO2 is not
conditions the catalyst is bulk RuO2(110). This does not mean, active by itself. Our present understanding of the elemental
however, that the catalyst is bulk-terminated during the reaction steps associated with the activation process still seems to be
(for a model of the DFT-calculated perfect surface structure of rather scant. If, as argued by Gao and Goodman,91 the active
RuO2(110) under vacuum conditions, see Figure 22). Actually, it phase is a thin RuO2 film on an essentially metallic substrate,
is not, but rather breaks up under the influence of the reactants. one might think strong metal support interactions are
The reduction of RuO2 was also observed under mild tempera- involved. An interesting case study in this respect is the CO
tures in the mere presence of CO at low pressures.109 Gao and oxidation over monolayer FeOx films grown on top of Pt(111)
Goodman, on the basis of comparative kinetic and spectroscopic as performed by Freund et al.49,113118
studies with Ru(0001) and RuO2(110) thin films, arrived at Obviously, a combination of oxygen activation, Eley
the conclusion that defect-rich metallic Ru provides the catalyt- Rideal and Marsvan Krevelen mechanisms is in operation:
ically highly active phase under reducing and stoichiometric
conditions.91 Furthermore, according to the same authors, only
The oxygen interacts with the double-layer FeO by pulling
an iron atom up above the oxygen layer. This lowers the
under net-oxidizing conditions at elevated temperatures
work function at the interface locally so as to allow for an
(>475 K) is the RuO2 film stable and produces CO2 at a lower
electron transfer toward the oxygen accompanied by the
rate than metallic Pt and Pd catalysts under identical conditions.
formation of a transient O2 molecule, which dissociates
Studies by Rosenthal et al.112 with pure RuO2 crystals also
and results in higher oxygen coverage in the formation of a
showed the need for a reduction in order to establish CO oxida-
local OFeO trilayer (the STM picture in Figure 23 shows
tion activity. Obviously, the bulk oxide structure was retained
the existence of such a trilayer).
during CO oxidation at ambient pressures while reducing the
crystal surface of certain facets so as to give rise to a roughened
When the trilayer is exposed to CO, CO2 is formed via an
ER mechanism.
surface exhibiting a large excess of Ru over oxygen as determined
by energy-dispersive x-ray spectroscopy. Thus, a surface with
If the oxygen pressure is high enough, the oxygen vacancy
will be filled via a Marsvan Krevelen mechanism.
local variations in the Ru oxidation state was formed which,
according to the authors, may be regarded as most active. It was Strong metal support interaction as described in detail by
argued that such redox plasticity may also explain the kinetic Freund et al.119and also reported by Fu et al.120 may pave the
rate oscillations in the CO oxidation at ambient pressures. way for preparing heterogeneous catalysts. It seems that the

20 mbar O2, 450 K
c(22)
-C
-O
- Fe
50 nm50 nm
Figure 23 DFT calculations and structural models of the CO oxidation on FeO/Pt (111). The STM image shows experimental evidence of the FeO2x
trilayer formation. Reprinted from Freund, H. J.; Meijer, G.; Scheffler, M.; Schlogl, R.; Wolf, M. Angew. Chem. Int. Ed. 2011, 50(43), 1006410094, with
permission, Wiley-VCH Verlag GmbH & Co, 2011.
electron transfer between the thin oxide film and the substrate
becomes the key elementary step in such systems.
A somewhat unexpected case of high CO oxidation activity
was reported by Haruta et al.121in the early 1980s. Accordingly,
supported Au catalysts with finely dispersed gold nanoparticles
proved to be active far below room temperature. Before this
discovery, catalysis textbooks did not even refer to gold because
of its chemical inertness. This high activity identifies such
catalysts as true benchmark systems for CO oxidation studies.
It was shown that the catalytic performance of gold in CO
oxidation depends critically on the particle size.122 While ini-
tial reports showed that the highest catalytic activity was
observed for gold clusters of
3 nm in size, recent DFT simu- very accessible to gas molecules.
lations showed that gold clusters of eight gold atoms deposited
on MgO(100) are active in CO oxidation.123125
These theoretical findings were later confirmed experimen-
tally by the group of Hutchings who observed a series of sub-
nanometer bilayer clusters containing around ten gold atoms.
In the same publication, they showed that monolayer clusters
of three to four atoms or isolated gold atoms were essentially
inactive for CO oxidation.126
Mavrikakis et al. used DFT to model CO and O2 adsorption
on single-crystal surfaces. They concluded that the adsorption
energy of CO and O does not depend on the number of Au
layers if there are more than two layers.127 In the same study,
the authors suggested that CO chemisorption occurs on under-
coordinated gold atoms (steps and kinks) rather than on the
Elementary Steps in Heterogeneous Catalysis 23
O2-
flat 111 surface. This was confirmed in a field desorption study
by Chau et al.128 which demonstrated the formation of gem-Au
(CO)2 species mass spectrometrically. Such dicarbonyl species
can only form at step sites or kinks. Mavrikakis et al.127 in their
theoretical study ascribed the enhanced activity of smaller Au
particles to an increase in the concentration of steps and kinks
(as the size of the particle decreases). Freund et al.129 agreed
that not all atoms in the gold cluster have identical catalytic
properties and suggested that atoms at the border between the
nanoparticle and the support nanoparticles are more active
because while still being in contact with the support they are
There are a series of factors which have to be taken into
account when discussing the catalytic activity of nanoparticles.
Some of the most important factors are as follows:
1. structural dynamic instability (during chemical reactions,
small clusters shift between various energetically accessible
structural isomers, thus enhancing the rates for overcoming
reaction barriers);
2. quantum size effects (size-dependent characteristics of the
electronic spectra of small gold clusters);
3. charge transfer from the support to the clusters; and
4. impurity doping effects that allow control of the electronic
structure of small clusters allowing modification and con-
trol of the electronic structure, and consequently the chem-
ical reactivity, of small supported clusters.
24 Elementary Steps in Heterogeneous Catalysis
The influence of particle size and morphology on the mode
of adsorption and activation of O2 was analyzed separately,
and it was found that O2 dissociation is very sensitive to the
arrangement of the gold surface atoms. It was shown that
smaller gold particles supported on TiO2 (110) exhibited
stronger adsorption of atomic oxygen.123
Presently, there is still a debate in the literature on defining
the nature of the catalytically active oxygen species present
under working conditions. Some researchers have suggested
that superoxide and peroxide species are reaction intermediates
in CO oxidation over pure or iron-doped Au/TiO2 samples.130
Stiehl et al.131 showed that molecularly adsorbed oxygen can
be deposited on both gold single crystals and gold nanoparti-
cles supported on TiO2(110) at 77 K. More importantly,
they claimed that this (molecularly adsorbed) oxygen can
directly react with CO. This is in disagreement with the
recent results published by Behm et al.132 who suggested
that the active oxygen species is an atomic species formed
by adsorption, activation, and dissociation of molecular
oxygen on the previously formed active perimeter sites
(Figure 24).
According to Behm et al.133 the active site in CO oxidation
is the interface between the nanoparticles and the support.
Their results indicate two roles of the interface in the CO
oxidation reaction: (1) accumulate adsorbed CO and (2) acti-
vate surface lattice oxygen at the Au  TiO2 perimeter sites.
Some of the supports used in studies with Au-based cata-
lysts show a remarkable CO oxidation activity on their own.
Among these, cobalt and manganese oxides perform particu-
larly well.134 In the case of cobalt, the most active oxide is
spinel Co3O4 in which cobalt is present in two valence states
(2 and 3). Similarly, high activity was also reported for a
number of manganese oxides.135138
a be likewise considered before a complete picture can be drawn.
c
b
d 7.02.3
Higher T
Figure 24 Schematic description of the pathway for CO oxidation
on Au/TiO2 catalysts at 80  C, involving (a) CO adsorption on Au NPs,
(b) reaction with activated surface lattice oxygen species at the perimeter
of the Au  TiO2 interface (interface sites), and (c) replenishment of
these sites by dissociative adsorption of O2 at the perimeter sites. (d) At
higher temperatures (>80  C), migration of surface lattice oxygen and
surface oxygen vacancies also gives access to neighboring surface lattice
oxygen. During reaction under normal reaction conditions, only the
perimeter sites are involved. Reprinted from Widmann, D.; Behm, R. J.
Angew. Chem. Int. Ed. 50(43), 1024110245, with permission,
Wiley-VCH Verlag GmbH & Co, 2011.
With respect to spinel cobalt oxide Co3O4,139 most research
has assumed that CO adsorbs on Co3,140,141 while some others
stated that CO would be adsorbed on Co2 sites.142,143 While
Pollack et al.143 provided IR evidence for the occurrence of CO
adsorption on Co2 and also concluded that a specific ratio
between Co2 and Co3 is required to obtain high CO oxidation
activity, it seems that DFT calculations favor Co 3 144,145 as the
preferred adsorption site of CO although electron donation in
this case may lead to reduction of Co3 to Co2. The interaction
of CO with Ow (oxygen bonded to both Co2 and Co3 cations)
or Os (oxygen bonded to three Co3) is shown in Figure 25.
In the case of oxygen adsorption over oxides, the initial
state of the adsorbent remains undetermined since the concen-
tration of oxygen in the bulk and, particularly, in the sub-
surface layers of oxides must not necessarily be stoichiometric
and may depend on the temperature, pressure of oxygen, his-
tory of treatment, etc. The recently discovered high catalytic
activity of MnOx (a nanosized system having many features in
common with Mn5O8) seems to be a good example for such
nonstoichiometric oxide structures.138,146
Much remains to be learned of these oxide materials in
terms of structurereactivity relationships. There seems to be
some agreement, however, in that adsorbed O2 on oxides
undergoes a series of transformation steps during which the
solid acts as an electron donor50,147:
O2ads ! O2 ads ! O2 2 ads ! O ads ! O2 lattice [11]
Oxygen species may differ not only in charge, but also in
coordination, bond energy, etc. Which species are really avail-
able under catalytic conditions and especially which one of
them is the active oxidant, that is, the species which can inter-
act directly with the molecule to be oxidized, are questions of
great importance but encounter experimental difficulties for
verification. Although many authors agree that a Marsvan
Krevelen mechanism is involved in the CO turnover to CO2,
the mechanistic details of this process are not completely
solved at present. For the mechanistic suggestions on Co3O4
(see Xu et al.145 and Broqvist et al.144). Besides the occurrence
of charged species and specific local site requirements during
reaction, the formation of various types of carbonates144 must
NO Adsorption, Decomposition, and Reaction
7.02.3.1 General Aspects
Nitrogen oxides are generally produced at high temperatures in
combustion processes carried out in air. Their effects on
human health and environment are detrimental and have led
to stringent regulation of their emission.148 The most desirable
method for the removal of NO is its direct decomposition into
the elements because no reducing agent is required. According
to the thermodynamics NO is unstable and its decomposition
to N2 and O2 is favored at low temperatures149:
2NO ! N2 O2 DGR 0 86:7 kJ mol1 [12]
As the decomposition is kinetically very slow, a catalyst is
needed to accelerate the reaction. The most active catalyst for

(a)
Figure 25 (a) CO adsorption on a sublayer Co ion. The adsorption becomes sterically hindered because there is little space between the surface
oxygen ions where CO enters. (b) CO adsorption at a surface exposed Co3 ion. This is the only favorable site found for CO adsorption. Dark gray,
oxygen; light gray, Co3; black, Co2; white, carbon.
NO decomposition is Rh, which lowers the energy needed to
dissociate the NdO bond to about 10% of the energy required
to dissociate the same bond noncatalytically in gas phase.150
NO decomposition using catalysts was initially studied by
Jellinck151 as early as the beginning of the twentieth century,
but more detailed studies started emerging at the beginning of
the 1970s after the introduction of automotive exhaust after
treatment technologies (e.g., three-way catalysts).152,153 Even
so, little was known on the mechanism of the interaction
between NOx(NO NO2) and catalyst surfaces until Blakeley
and Somorjai154 pointed out that the electronic properties of
the step atoms on metal surfaces differ markedly from those
of the terrace atoms. The average work function decreases
with increasing step density due to the overall decrease in
the work function caused by the decreasing number of nearest
neighbors of surface atoms. This integral approach allowed
them to conclude that rough surfaces are more reactive for
breaking chemical bonds than flat surfaces.155,156 Nearly 10
years later, Banholzer et al.157 suggested that the reactivity of
rough surfaces (steps and kinks) should follow orbital sym-
metry rules (a concept in organic chemistry originally devel-
oped by Woodward and Hoffmann, Nobel Prize 1981). More
recent studies showed that this approach does not apply
generally.
There is common agreement that the elementary steps of
the catalyzed NO decomposition include NO molecular
adsorption via a precursor state (either extrinsic or intrinsic
depending on the site occupation), followed by dissociation.
Adsorbed nitrogen atoms can subsequently recombine and
desorb as N2. The situation for adsorbed oxygen atoms is
more involved. To cause associative desorption in form of O2
usually affords high temperatures for most noble metal sur-
faces. Instead, N* and O* atoms may react with co-adsorbed
molecular NO* so as to produce N2O and NO2, respectively. In
addition, to prevent self-poisoning of the catalyst by accumu-
lating O*, reducing agents such as H2 or CO may be used so as
to scavenge O* through H2O and CO2 formation. The ideal
decomposition cycle would include eqns [13][16]. To
account for NO2 and N2O formation, we have included eqns
[17] and [18] which, however, are not unique (other mecha-
nisms can be envisaged as well):
Elementary Steps in Heterogeneous Catalysis 25
(b)
2NO 2* ! 2NO* [13]
2NO* 2* ! 2N* 2O* [14]
N* N* ! N2 2* [15]
O* O* ! O2 2* [16]
O* NO* ! NO2 2* [17]
N* NO* ! N2 O 2* [18]
In the first part of this section we describe the molecular
and dissociative NO adsorption on noble metals, with an
emphasis on low Miller index planes of Pt,158,159 Rh,160,161
and Pd.162,163 We subsequently move on to structural effects in
the decomposition and finish by inspecting NO oxidation and
reduction. For a more detailed description of the interaction
between NOx and various metal surfaces, we refer to the
reviews of Brown et al.164, Bowker et al.,165 and Garin
et al.166 Similar to the sub-chapter on CO adsorption and
reaction, we point out that it is not our intention to provide a
full review here. Rather we focus on aspects in relation to
elemental reaction steps.
7.02.3.2 NO Adsorption and Dissociation
For a long time, the susceptibility of NO dissociation was
assumed to vary systematically with the position of the transi-
tion metal in the periodic table. Although there is a tendency
that the dividing line originally introduced by Broden70 (see
also Brown and King164) for CO dissociation is shifted to the
right for NO dissociation, it seems that only closed d-shell
metals are incapable of breaking the NO bond. Pt metal
plays an ambivalent role in this scenario: the atomically
dense (111) plane does not break the NO bond while stepped
planes may do, indicating that structural effects are most
important to be considered for this metal.
Similar to CO adsorption via the carbon atom, the NO
molecule binds through the nitrogen atom to the metal surface.
Scattering experiments with molecular beams using the Stark
effect and hexapole lens focusing demonstrated167,168 a very
strong orientation dependence of the interaction. In particular,
26 Elementary Steps in Heterogeneous Catalysis
molecules coming in with the O atom oriented toward the
surface seem to undergo steering (see also the Introduction),
that is, reorientation on the femtosecond timescale169 so as to
form a surface bond with the reactive N-end. Fitting to the
Kisliuk21 model indicates that not only the chemisorption
bond but also trapping in the precursor state are favored for
N-down orientations.
As mentioned previously, both CO and NO usually adsorb
with high sticking probabilities on transition-metal surfaces. In
the case of Pt, sticking of NO may be influenced by the presence
of steps. This was demonstrated by atom-probe measurements
using field desorption techniques. Probing about 140 atomic
sites of the (111) facet (containing steps of predominantly
(111) symmetry) of a Pt nanosized crystal conditioned as a tip,
sticking probabilities between 0.65 and 0.9 were measured at
overall low coverage.170 These values are in remarkable agree-
ment with those reported more recently by Fiorin et al.171 in
experiments with extended single-crystal surfaces. The probe-
hole measurements also provided insight into the thermal
desorption kinetics of the NO* molecules.170 By varying the
reaction time from 100 ms to 10 s, the characteristic times neces-
sary for an equilibration of the adsorbed layer could be mea-
sured (see Figure 26).
These times were identical with the mean lifetimes t before
thermal desorption in a first-order process. From the tempera-
ture dependence of the t values (see Section 7.02.1) the kinetics
of the desorption process were evaluated. An activation energy
of 139 kJ mol1 and a preexponential factor of 3.8  1015 s
were reported this way. These data are in very good agreement
with those obtained in molecular beam studies with stepped Pt
(111) 172,173 but seem to be larger than those reported more
recently from micro-calorimetric measurements.171
101
e
1 mbar on Pt
at
Ions/1000 pulses
tr
en
me
ng
pi
Im
0
10
101
103 102 101 100
tR (s)
Figure 26 Thermal desorption kinetics of NO by probing
140 surface not for Pt(111), as discussed above, and neither for 4d10 Pd).
atoms of the (111) plane with field pulses170 at p 1.3  107 mbar and
various surface temperatures. Mean lifetimes before thermal desorption
are obtained from the (1  1/e) levels of the measured signal intensities
at long reaction times; the line marked impingement rate shows the
behavior for the case all impinging NO molecules would have stuck to the
surface. Reprinted from Kruse, N.; Abend, G.; Block, J. H. J. Chem. Phys.
1988, 88(2), 13071312, with permission, The American Institute of
Physics 1988.
It became clear from all these studies that NO molecules must
not necessarily decompose at surface steps. The lifetime measure-
ments discussed here clearly demonstrate that steps act as trap-
ping sites without dissociation. Since terrace sites are more
abundant than steps in all these measurements, it must be con-
cluded that surface diffusion is in operation converting terrace
site NO* into step site NO* (precursor-mediated step site chem-
isorption). For reasons of the microscopic reversibility, step site
NO* returns to terraces before thermal desorption. Overall, the
rate data provided above are determined by step site adsorption.
The simplistic picture of molecular adsorption via the
N atom of the NO molecule, as developed here for stepped
Pt(111), becomes more complicated when moving to higher
coverages. Even closed d-shell metals such as Cu174177 and
Ag178,179 for which NO adsorption is entirely molecular seem
to produce N2O during mild heating, after adsorption at low
temperatures.180 This reaction phenomenon is most probably
related to the radical character of the molecule (single occupa-
tion of the 2p* molecular orbital). Accordingly, adsorbed NO
may dimerize so as to form (NO)2* followed by decomposi-
tion to N2O and O*. We also mention that there is at least one
report on (NO)2* dimer formation following NO exposure
of Pd(100) and Pd(111) at temperatures as low as 20 K.162
It should be emphasized that the formation of dimers follow-
ing NO adsorption on metals under low-temperature condi-
tions is entirely different from the (sub)carbonyl analog
encountered during adsorption of CO on stepped metal single
crystals. For the latter, similar to inorganic coordination
chemistry, two (or more) CO molecules bind to the same
step site (see Section 7.02.2.2). Step site-bound twin nitrosyls
were only rarely reported so far. Somewhat surprisingly,
the stepped surface of a Au nanosized crystal seems to present
such a case and also suggests such nitrosyls to provide a path-
way to N2O.181
Dimerization of the NO molecule in dense overlayers at
low temperatures may indicate a pathway to N2O; yet it does
526 K not prove per se that this is relevant under high-pressure
533 K
dynamic reaction conditions. In this context, more recent
545 K
room temperature NO adsorption studies using ambient pres-
552 K
sure x-ray photoelectron spectroscopy up to
561 K
568 K (111) indicated the formation of O* 182. The authors consid-
577 K ered either minority step sites to cause dissociation of NO* or a
592 K sequence of steps to take place on terraces during mild heating
602 K according to 2NO* ! (NO)2* ! N2O* ! O*. In any case, only
NO pressures higher than 106 mbar caused irreversible for-
mation of O* leading to the formation of NO* O* domains.
Interestingly, the formation of dimeric (NO)2* was recently
observed for NO-covered Rh(111) surfaces at
room temperature.183
Besides O* formation via dissociation of an (NO)2* inter-
mediate, the direct pathway via a NO* molecular precursor
has to be mainly considered for open d-shell noble metals (but
The dissociation is usually, similar to the CO case, considered
to involve a late transition state (see also Introduction) and to
be highly structure sensitive. Open surfaces with steps and
edges are generally more reactive than close-packed flat sur-
faces. The dissociation depends, however, on the surface cov-
erage: the requirement of an empty site next to adsorbed NO*
(or, possibly, several sites within the ensemble context184)

explains why the dissociation probability decreases with
increasing coverage. DFT calculations of the potential energy
profiles of the dissociation reaction suggest the existence of an
essentially linear relationship between the transition-state
energy for dissociation and the binding energy of the dissoci-
ated products.185188 Such a behavior is usually known as the
BrnstedEvansPolanyi (BEP) relation or linear free energy
relation. Gajdos et al.189 investigated the NO dissociation on a
number of (111) oriented transition-metal planes and con-
firmed the existence of such a BEP rule for these systems. They
suggested the linear relationship to arise from the similarity of
the transition states on all metals which turned out to be very
close to the final state geometries. NO dissociation on Pd, Pt,
and Cu displayed a modest endothermic heat of reaction.
Transitions-state energies for these systems were only weakly
positive. In agreement with experimental observations,
adsorbed N* was found to be unstable and, consequently,
underwent rapid associative desorption.
Limiting our considerations of structural effects to Rh,
Pt, and Pd metals, we point out that while any Rh surface is
capable of breaking the NO bond in NO*, only atomically
open planes can do so in the case of Pt and Pd. Starting from
the 1980s the adsorption and dissociation of NO on Rh sur-
faces were therefore the subject of numerous theoretical190193
and experimental studies.194210 With respect to Rh, moving
from the most dense (111) plane to (100) and stepped Rh
(331)/(511) causes an increase of the rate of dissociation. The
low dissociation barrier for Rh(100) and Rh(110) is associated
with the existence of a lying-down NO structure which acts as a
precursor for dissociation. The increased activity of a species
adsorbed in lying down geometry was previously discussed in
the review published by Brown and King164 (Figure 27).
Turning to the Pt case, NO dissociation was observed on
Pt (100) and (511) single-crystal surfaces for adsorption tem-
peratures higher than 400 K.211 The same process occurred on
Pt (410) for considerably lower temperatures (
100 K). This
suggests that Pt (410) with (100) terraces is much more reac-
tive in NO dissociation than Pt (100).212 A further study com-
pared the behavior of Pt(410) with that of the (210) plane and
concluded that the former was considerably more active than
the latter, although the geometry of the steps is identical on the
two surfaces and the step density is even higher on the Pt(210)
surface.213,214 Therefore, it seems that NO decomposition is
sensitive to not only the steps but also the arrangement of the
atoms in the steps.215
O O
236 29 2.012 028
124.631
N Rh
N majority of the surface shows the (2  2) structure, except for the upper
943 944
44 844
Rh Rh
20 A min1. (a) and (b) share a common defect,
Rh
TS1(100) TS2(100)
Figure 27 Optimized geometries of NO dissociation transition states on
the Rh(100) surface. Repirnted from Tian, K.; Tu, X. -Y.; Dai, S. -S., Surf.
Sci. 2007, 601(15), 31863195, with permission, Elsevier 2007.
Elementary Steps in Heterogeneous Catalysis 27
Pd metal, on the other hand, is thought to be the least active
in NO dissociation among the three metals considered here.
Both the Pd(111) and Pd(100) planes were found inactive
even at 550 K.216 Exposing a Pd (311) surface to NO at or
slightly above room temperature caused N2 and N2O
formation.216 Obviously, NO dissociation took place on this
plane, but was accompanied by subsequent oxygen diffusion
into the Pd bulk. Re-segregation to the surface and desorption
of O2 was reported to occur at around 800 K.
As indicated earlier, high NO coverages can alter the reaction
behavior on noble metal surfaces. If NO adsorption is per-
formed at high gas pressures (see also above), new adsorption
structures may be formed which are not encountered at low
dosing pressures and temperatures. The NO/Rh(111) adsorp-
tion system is such a case. According to Rider et al.217 a (3 3)
7NO adsorption structure forms in equilibrium with the gas
phase at elevated NO pressures of 0.04 mbar (see Figure 28).
In a recent publication, Ozensoy et al.218 suggested the
existence of a similar high-pressure surface structure for NO
adsorption on Pd(111).On the basis of PM-IRAS data, the
authors proposed a (3  3)7NO structure to form with a NO
coverage of 7/9 monolayer (ML). Their results were in disagree-
ment with those published by Vang et al.219 who attempted to
simulate these experiments but failed to find a 3  3 structure
on Pd(111).
0s 55 s
(a) (b)
110 s
(c)
Figure 28 STM images of Rh(111) in 0.04 mbar of NO at 25  C,
showing the phase transition between a (2  2) and the (3  3) structure.
The 100 A  100 A images were obtained at 55 s intervals. In (a), the
Rh
right-hand corner, which shows the (3  3) structure. In (b) and (c),
the domain boundary between the two phases moves across the
image at a rate of
Rh marked with a circle. The close-packed rows of both patterns are aligned,
and the maxima are separated by multiples of the RhRh distance. This
indicates that the bright spots correspond to NO molecules on similar
sites in both the (2  2) and the (3  3) structures. Reprinted from
Rider, K. B.; Hwang, K. S.; Salmeron, M.; Somorjai, G. A. Phys. Rev. Lett.
2001, 86(19), 43304333, with permission, The American Physical
Society 2001.
28 Elementary Steps in Heterogeneous Catalysis

7.02.3.3 NO Oxidation and Reduction Side reactions may produce (undesired) nitrous oxide and
ammonia:
7.02.3.3.1 NO reaction with oxygen
There is ongoing debate concerning the nature of the mecha- 2 NO H2 ! N2 O H2 O [26]
nism involving NO oxidation on a metallic surface. Some
2 NO 5 H2 ! 2 NH3 H2 O [27]
authors assume that the reaction obeys an ER220222 mecha-
nism in which O2 is dissociatively adsorbed on the metal The selectivity of Pt metal to promote N2 rather than NH3
surface so as to provide O* for the reaction with gaseous NO. or N2O is relatively poor and depends on the H2 partial pres-
If the reaction proceeds through this mechanism, increasing O2 sure, that is, on the relative coverages of the reactants and their
coverage should have a positive effect on the NO oxidation rate. dissociation products. To account for these reactions a set of
This is not the case. Some experimental223 observations and elementary steps can be considered in the following manner.
theoretical calculations159 therefore rule out the ER mechanism Both NO and H2 undergo dissociative adsorption, with molec-
and propose the reaction to proceed according to a Langmuir ular precursor states being relevant only for NO:
Hinshelwood224226 mechanism. Koop et al.227 proposed that
both mechanisms are equally important and modeled the NO NO * ! NO* [28]
oxidation by a combined LH and ER reaction scheme.
Igelsia and Weiss228 proposed the following sequence of ele- H2 2* ! 2H* [29]
mentary steps for NO oxidation: O2 adsorbs molecularly onto a
vacant site *, eqn [19] and dissociates via reaction with a vicinal NO* * ! N* O* [30]
site to form chemisorbed oxygen O*, eqn [20] or via reaction
with chemisorbed NO* to form NO2 and O* eqn [21]. Vacancies Product formation then occurs via a LangmuirHinshelwood
form via reaction of NO with O* to produceNO2. mechanism:
When O* and * are the most abundant surface intermedi- 2N* ! N2 2* [31]
ates and O2* species dissociate before they desorb, step 1
becomes the sole kinetically relevant step: O* 2H* ! H2 O 3* [32]

O2 * ! O2 * [19] N* 3H* ! NH3 4* [33]

O2 * * ! 2O* [20] Product molecules may have measurable surface lifetimes

depending on the actual reaction temperature. For reasons of
O2 * NO* ! O* NO2 * [21]
simplicity, they are assumed to desorb directly into the gas
NO O* ! NO2 * [22] phase. Furthermore, the hydrogenation of O* and N* usually
involves OH* and NHx* intermediate species which are not
2NO O2 v ! 2NO2 [23] considered in the above scheme either.
2O* ! O2 2*
v
[24] In the sub-chapter on CO oxidation over noble metal sur-
faces, we have referred to kinetic instabilities in the reaction
where the asterisk indicates catalytic active site and superscript behavior leading to rate oscillations under certain conditions.
v the vacant site. Such instabilities can also be observed in the catalytic reduc-
The oxidation of NO to NO2 has relevance in the discussion tion of NO. For example, rapid ignition phenomena (surface
of fast SCR reactions. explosions) were observed in studies with Pt(100) single-
crystal surfaces.231233 Assuming NO* dissociation to be rate
7.02.3.3.2 NOx Reduction limiting, an autocatalytic mechanism can be constructed232 on
We shall develop here some basic concepts in the NO reduction the basis of the above reaction steps. Neglecting ammonia
with hydrogen. Clearly, from the automotive catalysis point of formation which is a reasonable assumption for low H2 partial
view, the NO reduction with CO and NH3 are likewise impor- pressures and high temperatures, the reaction accelerates once
tant and are discussed in Chapter 7.17. We mention though that vacant sites are created (by desorption of N*, for example):
Srinivasan et al.229 have recently compiled a meticulous over-
view of the most important reaction mechanisms and elemen- NO* 2H* * ! H2 O 1=2 N2 4* [34]
tary steps involved in the NO reduction by CO. A detailed
According to eqn. [32], three sites are produced per reaction
review of NO reduction on various materials was also provided
cycle leading to fast emptying of the adsorbed layer. It must be
by Parvulescu and Granger.230
mentioned, however, that an alternative view to describe
The reduction of nitric oxide with H2 has attracted consid-
both this fast ignition and the oscillatory reaction behavior
erable attention in the past and continues to be an issue for
was developed234 on the basis of a combination of (1) a
the future. Much of the differences in the performance of Pt
reversible phase transition between the hexagonal (hex) and
and Rh catalysts can be attributed to the fact that the PtN
the bulk-truncated (1  1) phase, (2) the strong nonlinearity
bond is much weaker than the RhdN bond. We limit our
in the growth of the (1  1) phase leading to a power-law
considerations of elemental steps in the NO  H2 reaction to
dependence235 on the adsorbate coverage, and (3) the high
the Pt case.
reactivity on (1  1) as compared to hex. We also mention
Generally, NO reduction with hydrogen leads to the forma-
that an oscillatory behavior for the same reaction was
tion of nitrogen and water:
observed on the surface of a Pt nanosized crystal conditioned
2 NO 2 H2 ! N2 2 H 2 O [25] as a tip.236,237

7.02.4 Methane Activation
7.02.4.1 General Principles
Methane is the thermodynamically most stable alkane and is
most difficult to activate due to the strong covalent CH
bonds. Its conversion into more useful chemicals and fuels
may be considered a major challenge for heterogeneous catal-
ysis in the twenty-first century. Besides the problem of activa-
tion, the achievement of high selectivity in chemical reactions
must be regarded as a particularly important issue. We can
roughly distinguish between oxidative and nonoxidative con-
version of methane. The former category includes methane
coupling (CC bond formation in the presence of gaseous
oxygen)238240 and also syngas241,242 or oxygenates formation
(methanol, formaldehyde, for example),243247 with the possi-
bility of subsequent gasoline production (MTG), whereas the
latter category is targeted on either hydrogen production248
(thermodynamically unfavorable) C2 and C2 hydrocarbon
formation249,250 or aromatization.251,252 Syngas production by
steam reforming on Ni-based catalysts, CH4 H2O ! CO 3H2,
is currently the major application of methane conversion, but
both dry steam reforming,253,254 CH4 CO2 ! 2CO 2H2, and
partial oxidation,255258 CH4 O2 ! CO 2H2, likewise other hand, the surface temperature did not influence the trans-
enjoy interest. Through these examples of methane conversion,
scientists have focused their interest on the understanding of the
elementary steps of methane dissociation common to all of these
processes. This section is devoted to summarizing the achieve-
ments in this respect. Both from an experimental and theoretical
point of view, methane may be regarded as the prototype hydro-
carbon for studying CH bond activation. Rather than limiting
our review to mean-field considerations, we shall also briefly
mention achievements of quantum state-resolved reaction
dynamics. A recent review summarizes the detailed knowledge
obtained from such studies.259
Quite generally, the methane initial sticking probability S0
on single-crystal metal surfaces is very low and difficult to
measure with accuracy. We note in this context that the classi-
cal King and Wells method mentioned in the Introduction
cannot be used for determining S0 -values below
102. It which was considerably smaller for the (100) plane. The low
seems that except for reports of methane adsorption on
Ni(100),260 the S0-values are frequently several orders of mag-
nitude below this threshold for whatever the orientation of the
Ni single-crystal surface.261,262 Such small values of the sticking
probability can nevertheless be conveniently measured using
(supersonic) molecular beams along with time-of-flight analy-
sis262264 which, despite their rather complex setup, also allow
a separate control of the translational and vibrational energy of
the gas molecules, which is an important asset in the context of
discussions about which motional degrees of freedom are most
involved in the activation process.
Basically, two different interaction mechanisms can be envis-
aged: direct dissociation and molecular precursor-mediated dis-
sociation. There is experimental evidence for both mechanisms
to be possible although early surface science work with molec-
ular beams already clearly indicated the importance of direct
dissociation on impact.265,266 In principle, the distinction
between both mechanisms can be made experimentally by
changing the gas temperature relative to the surface temperature.
Nevertheless, there has been a considerable debate in the liter-
ature over the mechanism of methane dissociation and so a brief
Elementary Steps in Heterogeneous Catalysis 29
account of early and contemporary single-crystal work will be
provided first, followed by recent theoretical work using DFT,
before concluding with the implications that such fundamental
insight may have for catalysis.
7.02.4.2 Single-Crystal Studies
The first study employing molecular beam techniques was obvi-
ously published by Rettner et al.265 The authors demonstrated a
dramatic increase in the sticking probability by several orders of
magnitude when raising the translational energy of the methane
beam during its impact on a W(110) surface. To explain their
results, the authors claimed a translationally activated tunneling
mechanism to be in operation. This was questioned through
experiments with isotope-exchanged methane on W(211): Lo
and Ehrlich267 rather advocated the dissociation to be mediated
by vibrational excitation without tunneling. A little later, Lee
et al. performed molecular beam experiments to investigate the
interaction of methane with a Ni(111) surface.266 The dissocia-
tion probability in these studies was found to increase exponen-
tially by two orders of magnitude when changing the beam
translational energy Etrans from 50 to 71 kJ mol1. On the
lational energy dependence of the sticking probability. It was
clear from the results that methane underwent direct dissocia-
tion on collision and that the vibrational (Evib) and translational
energy (Etrans) were both effective in this process. Furthermore,
the model was advanced that the collision of the molecule with
the surface caused a deformation so as to create a transition state
favorable to dissociation.
This early molecular beam work sparked a number of stud-
ies using supersonic expansion along with laser excitation of
the beam. Before dealing with these, it is useful to recall some
of the kinetic (and partly conflicting) results obtained from
bulb experiments. Beebe et al.261 studied methane interaction
with the (111), (100), and (110) low-index Ni planes and
reported similar activation energies for the dissociation on Ni
(111) and Ni(110) (5060 kJ mol1), but an activation energy
reactivity of all these surfaces became particularly evident from
the initial sticking probabilities which were small, that is, 107
for Ni(110), and even below this value for Ni(111) and (110).
The authors interpreted these differences as proof of the struc-
ture sensitivity of the decomposition. While there was good
agreement for the CH4/Ni(111) system in the bulb studies by
Beebe et al.261 and the beam studies by Lee et al.,266 consider-
able discrepancies were reported by others using beam
techniques.260,268,269 When assessing the data one should con-
sider that the surface fluxes in molecular beam and bulb exper-
iments may be rather different: while in bulb measurements
the pressures may enter the mbar range, the equivalent pres-
sure in beam studies is usually several orders of magnitude
lower. Juurlink et al.270 probably reported the first state-
resolved study for methane dissociation on Ni. They found
the reactivity on Ni(100) to be significantly enhanced once
the methane molecule was excited to the v 1 level of its (triply
degenerate) antisymmetric CH stretching vibration (n3), rela-
tive to the vibrational ground state. On the other hand, Evib
and Etrans were similarly effective in activating the dissociation.
The same pattern of effectiveness was found after exciting the
30 Elementary Steps in Heterogeneous Catalysis
methane molecule into the two-quantum 2n3 state.271 A sepa-
rate study of rotational state effects showed that all rotational
vibrational sublevels of the n3 state have comparable reactivity
(see Figure 29).272 Interestingly, when Schmid et al.271 quan-
tified the reactivity of the v 1 excited n3 state on Ni(111), they
found Evib to be significantly more effective than Etrans, which
contradicts statistical considerations. A little later, the first clear
evidence for vibrational mode selectivity in methane activation
was reported by Beck et al.273 who studied the reactivity of
CH2D2 in which the CH stretch vibration was solely excited
either with two quanta or the CH and CD vibrations with
one quantum each. They found that the |2,0 > state led to
higher reactivity than the slightly more energetic |1,1 > state.
The authors interpreted this result assuming that no intra-
molecular vibrational redistribution (IVR) occurs during the
reactive moleculesurface encounter. Thus, the excitation of
only one localized bond activates reaction while keeping the
other bonds, either excited or not, as spectators. Following
these arguments, the |1,1 > state is less reactive because only
the single quantum-excited bond closest to the surface is
available for activation while the distant CH bonds are not.
A dramatic mode selectivity was reported by Maroni
et al.274 when comparing the symmetric n1 excitation with the
previously reported antisymmetric n3 fundamental and 2n3
overtone excitation on Ni(100). Despite the somewhat lower
internal energy of n1 as compared to n3, it was more reactive
100
10-1
Initial reaction probability (S0)
10-2
10-3 34 kJ mol-1
45 kJ mol-1
10 -4
65 kJ mol-1
10-5
10-6
10-7
0 25 50 75 100
Etrans (kJ mol-1)
Figure 29 State-resolved reaction probability (or sticking probability,
S0) curves for methane dissociation on Ni(111). Open circles represent
the n0 vibrational ground state, blue diamonds the n3 fundamental, and
red squares the 2n3 overtone asymmetrical stretch. Also included is the
3n4 bend second overtone in black solid circles. The horizontal offset of
the curves provides the vibrational efficacy. For example, n3 excitation
deposits 36 kJ mol1 of Evib in the methane molecule and reduces the
Etrans requirement for the dissociation by 45 kJ mol1. The quotient
DEtrans/Evib is a quantitative measure of the vibrational efficacy. Reprinted
from Juurlink, L. B. F.; Killelea, D. R.; Utz, A. L. Progr. Surf. Sci. 2009, 84
(34), 69134, with permission, Elsevier 2009.
and even came close to the 2n3 overtone reactivity with twice
the vibrational energy.
Current models describing the transition state of methane
dissociation require both CH stretching and bending excita-
tion. When comparing the n3 reaction probability with the 3n4
(three quanta of umbrella-type bending) on Ni(100) and
Ni(111), Juurlink et al.275 found that the CH asymmetric
stretching mode was significantly more efficient in promoting
dissociation on both surfaces whatever Etrans. This result is
remarkable because it demonstrates that reactivity does not
scale with vibrational energy which is about 30% higher in
the 3n4 (45 kJ mol1) than in the n3 state (36 kJ mol1). Once
again, IVR does not take place in the reaction complex,
presumably due to dynamical constraints.
It is interesting to note that the 2n3 efficacy to promote
dissociation is lower on Pt(111) than on Ni(111).276 Thus, the
barrier is later (the terminology of a late barrier refers to a
situation where the transition state in a dimensionality-reduced
PES is mainly reached by vibrational excitation see Section
7.02.1) with a more extended CH bond on Ni than on Pt.
Limitations in (or even absence of) IVR motivated Killelea
et al.277 to study bond selectivity in the methane decomposi-
tion over Ni(111). They selectively excited the n1 stretch vibra-
tion in trideuteromethane, CHD3, and determined the CD3 :
CHD2 product ratio after dissociation at 90 K by titrating the
surface methyl species with D atoms previously embedded in
Ni lattice interstitial sites.
In a subsequent TPD experiment they found a CD3 : CHD2
product ratio of >30:1 which contrasted with the 1:3 ratio of
an isoenergetic ensemble of CHD3 molecules whose vibrations
were statistically populated. The negligible influence of vibra-
tional energy redistribution between CH and CD stretch
modes in methane/surface interactions during sub-picosecond
collision times may be taken as an indication that vibrational
bond selectivity is the rule rather than the exception in hetero-
geneous reactions of small polyatomic molecules like meth-
ane. For a more detailed discussion of state-resolved studies in
general, and considerations of IVR in particular, we invite the
reader to refer to the review provided by Juurlink et al.259
Killelea et al.278 also studied the influence of the surface
temperature on methane activation. In their work on Ni(111),
the laser excitation prepared gas-phase methane in a single
rotational and vibrational quantum state with precisely
defined internal energy, while the molecular beam defined
translational energy. By accurately controlling methanes ener-
getic degrees of freedom, the experiments revealed the effects of
surface temperature on methanes dissociative chemisorption.
As the surface temperature increased from 90 to 475 K, the
125 150
reactivity likewise increased and the authors concluded that,
at higher surface temperature, surface phonon excitation
allows substrate atoms to sample lattice geometries with
more favorable sites for transition-state access.
Stepped or corrugated transition metals surfaces are ener-
getically more favorable to methane activation than atomically
flat ones, which is in excellent agreement with computational
studies that will be presented below. Bisson et al.279 showed
that the reaction probability on the anisotropic corrugated
Pt(110) 1  2 reconstructed surface (see Figure 30) was depen-
dent on the incident angle of the CH4 beam. By varying the
incidence of the methane beam from parallel to perpendicular

with respect to the missing rows, they showed that while the
methane beam stroked the surface perpendicularly (along the
more corrugated direction), the reaction probability signifi-
cantly increased. Furthermore, by taking into account the sha-
dowing effect (varying y angle), they were able to demonstrate
that the most reactive sites were the ridge atoms (green atoms
in Figure 30). This is in good agreement with what is claimed
in computational studies: less coordinated atoms are most
active for methane activation.
7.02.4.3 Computational Studies: Geometry of Adsorption
and Decomposition Mechanism
While approaching the surface, the methane molecule may
impact on different adsorption sites. Using DFT, Zhu et al.280
studied the influence of various binding sites for the
[110]
(q; j = 0)
(q; j = 90)
[001]
- activated hydrogen interacted both with the adatom and the
[110]
Figure 30 Pt (110) 1  2 surface model illustrating the two azimuthal
angles used for reactivity measurements: for parallel incidence, the
methane molecules impinge on the surface along the [110] direction.
For perpendicular incidence, the methane strikes the surface
perpendicular to the missing row. Surface atoms are differentiated from
the bulk ones using green, blue, and red colors. Reprinted from Bisson,
R.; Sacchi, M.; Beck, R. D. J. Chem. Phys. 132(9), with permission,
2010 The American Institute of Physics.
Hollow1 Hollow2 Bridge1
Figure 31 Schematic representation of the possible binding sites on Ni(100). Reprinted from Zhu, Y. A.; Dai, Y. C.; Chen, D.; Yuan, W. K. J. Mol. Catal.
Chem. 2007, 264(12), 299308, with permission, Elsevier 2006.
Elementary Steps in Heterogeneous Catalysis 31
adsorption of CH4, CH3, and H on a perfect Ni(100) surface.
Figure 31 presents the results for CH3.
After an investigation of the preferential adsorption sites for
CH4, CH3, and H, Zhu et al. modeled the first step in the
dissociation of CH4 leading to the formation of surface CH3
and atomic hydrogen. They found that whatever the orienta-
tion of the CH4 molecule approaching the surface (with one,
two, or three hydrogen atoms pointing to the nickel surface),
the energy barriers for dissociation were always the same. Thus,
the orientation of the methane molecule with respect to the
surface does not influence its reactivity. Figure 32 shows the
CH4 molecule approaching the surface with the 3CH bonds
pointing to the nickel atoms. The CH4 molecule adsorbs on
top of the Ni atoms. In the transition states, the active Ni atoms
are slightly displaced out of the surface so as to decrease the
coordination number and enhance the reactivity, and simulta-
neously the energy barrier for dissociation is lowered by the
formation of a CHNi three-atom encounter (Figure 32(b)).
The final state is characterized by hydrogen atom and methyl
groups on bridge sites.
While the majority of the efforts were focused on studying
flat surfaces, some consideration was also given to the influ-
ence of defects, like steps or kinks. AbildPedersen et al.281
carried out experimental and theoretical studies to investigate
the dissociation of methane on the terraces and steps of a
Ni(111) surface. They found the activation barrier of CdH
bond scission to decrease by 0.18 eV on steps with respect to
the flat Ni(111) surface. The same group282 also showed that
the stepped Ni(211) surface is more reactive than the flat Ni
(111) surface when studying methane activation on these sur-
faces in the presence of surface carbon and sulfur. In a some-
what different approach, Haroun et al.283 compared the
energies of methane dissociation on a flat Ni(111) surface
with those on the same surface modified by an additional Ni
adatom. They concluded from their calculations that when the
neighboring terrace site Ni atom, the energy barrier for meth-
ane dissociation decreased by 0.17 eV. The influence of the
surface metal coordination number was also studied for other
metals. For example, van Grootel et al.284 compared the
ninefold coordination of an Rh(111) plane with the
sevenfold coordination of an Rh(211) plane and found that
the barrier for CH4 dissociation significantly decreased when
reducing the coordination number. Thus, a corrugated surface
Bridge2 Atop1 Atop2
32 Elementary Steps in Heterogeneous Catalysis

1.0 was predicted to be more reactive for methane dissociation

than a planar one.
So far, computing was shown to provide an input in terms
0.8 of preferential adsorption sites and methane orientation with
respect to the surface. DFT calculations can go a step further
and simulate the overall reaction which consists of the stepwise
Energy (eV)

0.6
hydrogen abstraction from CHx :

0.4
CH4 2* ! CH3 * H* [35]
CH3 * * ! CH2 * H* [36]
0.2
CH2 * * ! CH* H* [37]

0.0 CH* * ! C* H* ([38]

0 1 2 3 4 5 6 7 8 285
Bunnik et al. used DFT to show that in the case of CH4
(a) Reaction coordinate ()
interaction with an Rh(111) surface, the first and last steps of the
reaction sequence are most difficult, the CH radical being the
most stable fragment. van Grootel et al.284 arrived at similar
conclusions. Moreover, these authors investigated the influence
of surface atomic steps on the methane decomposition by
1.673 inspecting also the Rh(211) surface as well as the edge of
48.363 atoms shared between two (111) facets of a nanorod model.
0.267 In all cases, CH* dissociation had the highest activation barrier.
However, the first step, CH4 2* ! CH3* H*, was rate con-
1.593
2.018 trolling because of the entropy loss associated with the adsorp-
Surface normal
tion process. Other groups investigated the combination of
structural effects as described above and chemical effects by
(b) surface alloying. One of the ideas behind such studies is to
selectively enable CH bond scission in CH4 while suppressing
Figure 32 Energy barrier and adsorption sites for methane dissociation
any further hydrogenolysis of adsorbed CH3 (Figure 33).
on Ni(100) with an indication of the initial state, transition state, and
final state geometries. Reprinted from Zhu, Y. A.; Dai, Y. C.; Chen, D.;
For example, Kokalj et al.286constructed a model system
Yuan, W. K. J. Mol. Catal. Chem. 2007, 264(12), 299308, with consisting of isolated Rh atoms on top of a barely reactive Cu
permission, Elsevier 2006. (111) surface. Considering the Rh adatom as the reaction center,

50 42

35
CH4(g)
0
26 27
41
50 40 57
60
Energy (kJ mol-1)

71 87
94
87 88
100 99
117
112 Rh-rod(111)-(111)
126
134
150 142
150 165

178 174
200
205
219
243
250
254 256
Rh(211)
270
CH4 CH3 + H CH3 CH2 + H CH2 CH + H CH C+H C + 4 H (inf)
300
Reaction coordinate
1
Figure 33 Potential energy diagram (energies in kJ mol ) for the dehydrogenation of CH4 to C and 4H on conjugated rhodium surfaces: (black)
periodic Rh(211) surface and (red) Rh-rod(111)_(111). Reprinted from van Grootel, P. W.; van Santen, R. A.; Hensen, E. J. M. J. Phys. Chem. C 115(26),
1302713034, with permission, American Chemical Society 2011.

the authors showed that the first dehydrogenation step to
methyl was favored whereas methylene formation from methyl
was hindered. It is felt that while such DFT studies may provide
interesting insight into hypothetical mechanistic scenarios, a
tailored large-scale design of such defect structures along with
a sustainable operation under harsh environmental reaction
conditions will hardly become feasible in the near future.
7.02.4.4 Real Catalyst Studies: Steam/Dry Steam Reforming
287
Xu et al. were probably the first to report on the kinetics and
mechanistic details of methane steam reforming on a Ni/
MgAl2O4 catalyst. Their calculations were based on the assump-
tion of a detailed reaction mechanism involving Langmuir
Hinshelwood steps. Without going into the details, one of the
major assumptions was the occurrence of COHx species in the
mechanistic sequence of steps. The proposed mechanism was
subject to controversy as it did not consider methane to decom-
pose in a (rate-determining) collision-induced process followed
by stepwise hydrogen abstraction to carbon.241
In a series of papers, Wei and Iglesia288293 investigated
the mechanisms of the CH4 reforming reactions with H2O on
various supported catalysts containing Ni, Rh, Pt, Ir, Ru, and
Pd metals. The authors found that the turnover rates were
proportional to the CH4 partial pressure, but independent
of the CO2 and H2O pressures, which led them to conclude
that the CH bond activation in methane is of sole kinetic
relevance. Their data indicated that co-reactant (CO2 or H2O)
activation and its kinetic coupling with CH4 activation and
decomposition via scavenging of chemisorbed CHx interme-
diates and surface carbon are fast steps and lead to metal
surfaces essentially being uncovered by reactive intermedi-
ates. In their kinetic schemes the authors therefore suggested
irreversible CH bond breaking in methane (rate-
determining) running into a sequence of (irreversible) steps
with hydrogen abstraction from CHx to form surface carbon,
C*. These CH bond activation steps are then followed by C*
and H* removal via either recombinative desorption, as for
H*, or reactiondesorption with chemisorbed O* derived
(111)
I (100)
III
II
IV
(100)
V (111)
(111)
(111) (100)
(a)
Figure 34 A 201-atom cuboctahedral Pt cluster (1.8 nm) consisting of eight (111) and six (100) facets saturated with chemisorbed O* atoms. There
are five distinct types of O* atoms (labeled IV) and four types of Pt atoms (labeled 14) on the facets. O* atoms chemisorb on threefold fcc sites
of the (111) (types II, IV, and V) and twofold bridge sites of the (100) (types I and III) facets. Saturation occurs at full coverage in which all of the sites
IV are occupied by O* atoms. Reprinted from Chin, Y. -H.; Buda, C.; Neurock, M.; Iglesia, E. J. Am. Chem. Soc. 2011, 133, 1595815978, with
permission, American Chemical Society 2011.
Elementary Steps in Heterogeneous Catalysis 33
from the reversible and quasi-equilibrated dissociative
adsorption of the CO2 or H2O co-reactants. The general
conclusion that CdH bond activation in methane is the
kinetically relevant step is also in good agreement with com-
putational studies as seen above.
Nevertheless, the scientific community was (and probably
still is) divided since support and promoter effects were not
considered in the kinetic scheme.294296 Other studies, such as
those reported by RostrupNielsen and Hansen on Ru-based
catalysts,297 concluded that while CO2 and H2O reforming
mechanisms may be similar, higher reaction rates occur for
H2O than for CO2 reforming. The authors proposed that the
replacement of H2O co-reactants with CO2 introduced a
kinetic bottleneck associated with CO2 activation, which
became the rate-determining step in CO2 reforming reactions.
As mentioned by Wei and Iglesia, a complex reaction
scheme aimed at reflecting the essential steps of methane
reforming needs a rigorous analysis of the reaction kinetics
before appropriate conclusions can be drawn. They also
emphasized the importance of kinetic isotope exchange mea-
surements which should be performed at CO conversions far
from equilibrium. Furthermore, transport effects have to be
avoided as they would mask true reaction kinetics. There is
no doubt that such is the case in their work. This also
becomes clear when examining the experimental and theo-
retical (DFT) data obtained by the same group for the
CH4  O2 reaction over Pt nanosized particles for varying
surface oxygen coverages.298 The authors were able to iden-
tify four regimes with distinct rate equations due to different
kinetically relevant steps, predominant adsorbed species, and
rate and equilibrium constants of the different elementary
steps. Transitions among regimes occurred as chemisorbed
oxygen (O*) coverages changed. To perform simulations, a
model of a 1.8 nm Pt cluster with 201 atoms in cuboctahe-
dral morphology was considered representative (Figure 34)
of their supported catalysts (Pt particle size from 1.8 to
8.5 nm).
In three of these kinetic regimes, CH bond activation was
identified as the sole kinetically relevant step, but as occurring
3
4
1
2
(b)
34 Elementary Steps in Heterogeneous Catalysis

Rate and
Steps Elementary reaction steps equilibrium
constant
1.1 CH4 + O* + O* CH3O* + OH* k1[O*-O*]
1.2 CH4 + O* + * CH3* + OH* k1[O*-*]
1.3 CH4 + * + * CH3* + H* k1[*-*]
2.1 O2 + * O 2* k2af, k2ar
2.2 O 2* + * 2O* k2bf, k2br
3 C* + O* CO* +* k3
4 CO* + O* CO2* +* k4f, k4r
5 2H* H2* + * KH2
6 H* + O* OH* +* KO*, H*
7 CO2* CO2 + * KCO2
8 2 OH* H 2O* + O* KOH*
9 OH* + H* H 2O* +* KH2O
10 CO* CO + * KCO

Kinetic regime and Kinetically relevant step Apparent

Rate equation Effective rate
MASIa activation energy
constantb (keff)
(kJ mol1)
Regime 1 (O* - O*) keff1(CH4) CH4 + O* + O* k1[O*-O*] 155 9
CH3O* + OH*
Regime 2 (O* - *) keff2(CH4) / (O2) CH4* + O* + * (k1[O*-*])/k2bf K2a 79
CH3* + OH*
Regime 3 (* - *) keff3(O2) O2 + 2* 2O* 0.5k2bf K2a <3 16

Regime 4 (* - *) keff4(CH4) CH4 + * + * k1[*-*] 1.66

CH3* + H* 1.77

Figure 35 Top: Reaction scheme with elementary steps for O2/CH4, H2O/CH4, and CO2/4 reactions; ! denotes an irreversible step and a
quasiequilibrated step; kif and kir are the forward and reverse rate coefficients of step i, respectively, and Ki is the equilibrium constant for step i. Bottom
scheme: rate equations, kinetically relevant steps, and apparent activation energies for a Pt/Al2O3 catalyst (8.5 nm Pt particle size); a and b refer to the
reaction scheme above. MASI, most abundant surface intermediate. Adapted from Chin, Y. -H.; Buda, C.; Neurock, M.; Iglesia, E. J. Am. Chem. Soc.
2011, 133, 1595815978.

at different active sites, that is, oxygenoxygen (O*O*),
oxygenoxygen vacancy (O**), and vacancyvacancy (**)
site pairs when O* coverages decreased due to diminishing O2/
CH4 reactant ratios. A fourth regime was considered to provide
a bridge to the kinetic behavior in the presence of H2O/ CO2
co-reactants, but in the absence of O2, neither H2O nor CO2 in
O2/CH4 influenced the turnover rates in any of the regimes,
that is, they do not modify the coverages of reactive interme-
diates and active site structures during CH4 reaction with O2,
H2O, and CO2 in all regimes.
The reaction scheme considered by the authors is shown
in Figure 35. On O* -saturated surface, O*O* site pairs
activate CH bonds in CH4 via homolytic hydrogen abstrac-
tion from ingradical-like CH3 groups weakly interacting with
the surface. In this regime, rates depend linearly on CH4
pressure but are independent of O2 pressure. O2 activation
steps are quasi-equilibrated under these conditions. Mea-
sured and DFT-derived CH bond activation barriers are
large. They are determined by the weak stabilization and
the relatively high entropy of the CH3 fragments. Much
higher turnover rates are observed as vacancies become
available.
Under these conditions, O** site pairs activate CH bonds
via concerted oxidative addition and H-abstraction in transi-
tion states effectively stabilized by CH3 interactions with the
vacancies. O2 activation now becomes irreversible, because fast
CH bond activation steps scavenge O* as it forms. Thus, O*
coverages are set by the prevalent O2/CH4 ratios rather than the
O2 pressures. As O2/CH4 ratios and O* coverages become
smaller, O2 activation on bare Pt becomes the sole kinetically
relevant step; turnover rates are now proportional to O2 pres-
sures and independent of CH4 pressure. In the absence of O2,
alternate weaker oxidants, such as H2O or CO2, lead to a
kinetic regime in which CH bond dissociation on ** pairs
limits CH4 conversion rates. Rates are of first order in CH4 and
independent of coreactant. In this case, turnover rates increase
with increasing dispersion, because low-coordination sites sta-
bilize the CH bond activation transition states more effec-
tively via stronger binding to CH3 and H fragments. These
findings and their mechanistic interpretations are consistent
with all rate and kinetic isotopic data and with theoretical
estimates of activation barriers and of particle-size effects on
transition states.
7.02.5 Conclusion
Starting from basic kinetic concepts and focusing on the ele-
mentary steps of adsorption on solid surfaces, diffusion of
adsorbed species, reaction, and desorption, we have developed

here state-of-the-art mechanistic evidence for seemingly simple
reactions on metal and metal oxide surfaces of (mainly) model
catalysts. In particular, we have provided a summary of the
progress made over the past two decades in the research of
carbon monoxide adsorption and oxidation as well as on nitric
oxide adsorption and reduction. Both reactions also play an
important role in air pollution control by catalysis and therefore
provide an introduction to other more specific chapters in this
comprehensive. This chapter concludes by inspecting a reaction
frequently deemed to be the Holy Grail in catalysis research:
methane activation. Although this chapter has been written by
experimentalists, a great deal of attention has been given to the
tremendous development of computational studies using DFT
to help develop mechanistic insight at the atomic scale.
CO oxidation is the most frequently studied reaction in
surface science and has been used extensively as a catalytic
test reaction to provide a fundamental understanding of het-
erogeneous catalysis. The same reaction is very often used to
illustrate the progress in the development of experimental
methods and theoretical calculations. Therefore, some room
is given in this chapter to the occurrence of nonlinear reaction
dynamics (leading to oscillatory behavior under certain condi-
tions). This way we hope to demonstrate the progress made
from the early concepts, which believed species move and react
on solid surfaces like checkers on a checkerboard (the Lang-
muir era), to cooperative reaction phenomena with character-
istic pattern formation as observed by high-resolution
microscopy during the ongoing reaction (the Ertl era).
CH4, the simplest alkane, is an excellent candidate for
studying the activation mechanism of a molecule during its
interaction with a solid surface. The concepts of vibrational
excitation of specific bonds as a means to control the dissoci-
ation and reaction behavior have been reviewed here. The
conversion of methane to more valuable products (methanol,
synthetic fuels, and so on) is the subject of continuous research
and will continue to be so in the future.
We have limited our review to reactions of carbon monox-
ide, nitric oxide, and methane. Despite the simplicity of these
molecules, it seems that we are still quite far from understand-
ing the mechanisms and kinetics of these reactions at all levels
of complexity. As we have witnessed over the past 20 years,
moving from surface science studies with structurally well-
defined surfaces at low gas pressures to catalytic studies with
real materials under atmospheric (or higher) gas pressures (to
fill materials and pressure gap) turned out not to be trouble
free. Many efforts were undertaken to adapt existing experi-
mental techniques or to develop new ones for truly in situ
catalytic research, and it is to be expected that this conceptual
approach will continue in the future by applying ever more
sophisticated methods. Interestingly, high-pressure research
does not always lead to a consistent picture of the elementary
reaction steps, nor does the communitys opinion always con-
verge in full agreement. CO oxidation is one such example and
has been developed here in more detail for this reason.
Nevertheless, tremendous progress has been made over the
past two decades. Considerable insight has been gained in
some cases methane activation, for example by combining
kinetic research with DFT calculations. Computational chem-
istry will remain a topic for many years to come, according
to our opinion. A concerted approach of theorists and
Elementary Steps in Heterogeneous Catalysis 35
experimentalists is expected to become more the rule than
the exception and will deepen our mechanistic understanding
of catalytic surface reactions.
References
1. Hinshelwood, C. Oxford university press, Oxford,1940.
2. Eley, D. D.; Rideal, E. K. Nature 1932, 130, 222.
3. Mars, P.; van Krevelen, D. W. Chem. Eng. Sci. 1954, 3(suppl 1), 4159.
4. Friedrich, W.; Knipping, P.; Laue, M. Sitzungsber. Bayer. Akad. Wiss, (Math.
Phys. Klasse) 1912, 303.
5. Langmuir, I. J. Am. Chem. Soc. 1918, 40(9), 13611403.
6. Langmuir, I. J. Am. Chem. Soc. 1916, 38(6), 11451156.
7. Langmuir, I. J. Am. Chem. Soc. 1915, 37(3), 417458.
8. Langmuir, I. J. Am. Chem. Soc. 1915, 37(5), 11391167.
9. Chorkendorff, I.; Niemantsverdriet, J. W. 2nd ed.; Concept of Modern Catalysis
and Kinetics Wiley-VCH: Weinheim, 2007.
10. Freund, H. J. In Handbook of Heterogeneous Catalysis; Ertl, G., Knozinger, H.,
Weitkamp, J., Eds.; 3. Wiley-VCH: Weinheim, 1997.
11. Jones, J. E. Proc. Roy. Soc. Lond. 1924, 106(738), 463477.
12. Gross, A.; Wilke, S.; Scheffler, M. Phys. Rev. Lett. 1995, 75(14), 27182721.
13. Rendulic, K. D.; Anger, G.; Winkler, A. Surf. Sci. 1989, 208(3), 404424.
14. Rettner, C. T.; Auerbach, D. J. Chem. Phys. Lett. 1996, 253(34), 236240.
15. Beutl, M.; Riedler, M.; Rendulic, K. D. Chem. Phys. Lett. 1995, 247(3),
249252.
16. Luntz, A. C. Surf. Sci. 2009, 603(1012), 15571563.
17. Polanyi, J. C. Science 1987, 236(4802), 680690.
18. Wintterlin, J.; Schuster, R.; Ertl, G. Phys. Rev. Lett. 1996, 77(1), 123126.
19. Ertl, G. In Handbook of Heterogeneous Catalysis; Ertl, G., Knozinger, H.,
Schuth, F., Weitkamp, J., Eds.; Wiley-VCH: Weinheim, 2008; vol. 3, p 1462.
20. King, D. A.; Wells, M. G. Proc. R. Soc. Lond. A Math. Phys. Sci. 1974,
339(1617), 245269.
21. Kisliuk, P. J. Phys. Chem. Solid 1957, 3(12), 95101.
22. Henry, C. R. Surf. Sci. Rep. 1998, 31(78), 231325.
23. Willinger, M.; Schlogl, R.; Visart de Bocarme T.; Kruse, N. Unpublished work.
24. Gontart, L. C.; Chang, L.-Y.; Hetherington, C. J. D.; Kirkland, A. I.; Ozkaya, D.;
Dunin-Borkowski, R. E. Angew. Chem. Int. Ed. 2007, 46(20), 36833685.
25. Giorgio, S.; Sao Joao, S.; Nitsche, S.; Chaudanson, D.; Sitja, G.; Henry, C. R.
Ultramicroscopy 2006, 106(6), 503507.
26. Bonifas, A. P.; McCreery, R. L. Nat. Nano. 2011, 5(8), 612617.
27. Balke, N.; Jesse, S.; Morozovska, A. N.; Eliseev, E.; Chung, D. W.; Kim, Y.;
Adamczyk, L.; Garcia, R. E.; Dudney, N.; Kalinin, S. V. Nat. Nano. 2010, 5(10),
749754.
28. Antczak, G.; Ehrlich, G. J. Chem. Phys. 2008, 129(12), 124702.
29. Antczak, G.; Ehrlich, G. Surf. Sci. Rep. 2007, 62(2), 3961.
30. Imbihl, R.; Ertl, G. Chem. Rev. 1995, 95(3), 697733.
31. Dittmeyer, R.; Emig, G.; Kapteijn, F.; Moulijn, J. A.; Karger, J. Kinetics and
Transport Processes. In Handbook of Heterogeneous Catalysis, Wiley-VCH Verlag
GmbH: Weinheim, 2008; pp 11891261.
32. Stoltze, P.; Norskov, J. K.; Freund, H. J.; van Santen, R. A.; Neurock, M.;
Boudart, M.; Mullins, C. B.; Tamura, K.; Naito, S.; Jonkers, G.; Ertl, G.
Elementary Steps and Mechanisms: Sections 5.1 5.2. In Handbook of
Heterogeneous Catalysis, Wiley-VCH Verlag GmbH: Weinheim, 2008
pp 9111051.
33. Zhdanov, V. P. Surf. Sci. Rep. 2002, 45(78), 231326.
34. Norskov, J. K.; Bligaard, T.; Rossmeisl, J.; Christensen, C. H. Nat. Chem. 2009,
1(1), 3746.
35. Hahn, J. R.; Ho, W. J. Chem. Phys. 2005, 123(21), 214702214706.
36. Rettner, C. T.; Auerbach, D. J. J. Chem. Phys. 1996, 104(7), 27322739.
37. Rettner, C. T.; Michelsen, H. A.; Auerbach, D. J.; Mullins, C. B. J. Chem. Phys.
1991, 94(11), 74997501.
38. Kori, M.; Halpern, B. L. Chem. Phys. Lett. 1984, 110(3), 223229.
39. Ree, J.; Kim, Y. H.; Shin, H. K. J. Chem. Phys. 1996, 104(2), 742757.
40. Sachs, C.; Hildebrand, M.; Volkening, S.; Wintterlin, J.; Ertl, G. J. Chem. Phys.
2002, 116(13), 57595773.
41. Sachs, C.; Hildebrand, M.; Volkening, S.; Wintterlin, J.; Ertl, G. Science 2001,
293(5535), 16351638.
42. Smith, J. J. N.; Palmer, R. L. J. Chem. Phys. 1972, 56(1), 1320.
43. Verheij, L. K.; Hugenschmidt, M. B.; Colln, L.; Poelsema, B.; Comsa, G. Chem.
Phys. Lett. 1990, 166(56), 523530.
36 Elementary Steps in Heterogeneous Catalysis
44. McEwen, J.-S.; Gaspard, P.; de Bocarme, T. V.; Kruse, N. Proc. Natl. Acad. Sci.
2009, 106(9), 30063010.
45. Matsushima, T.; Rzeznicka, I.; Ma, Y. Chem. Rec. 2005, 5(2), 8193.
46. Engel, T.; Ertl, G. J. Chem. Phys. 1978, 69(3), 12671281.
47. Matsushima, T. J. Chem. Phys. 1990, 93(2), 14641471.
48. Nakao, K.; Ito, S.-i.; Tomishige, K.; Kunimori, K. Chem. Phys. Lett. 2005,
410(13), 8689.
49. Freund, H. J.; Meijer, G.; Scheffler, M.; Schlogl, R.; Wolf, M. Angew. Chem. Int.
Ed. 2011, 50(43), 1006410094.
50. Royer, S.; Duprez, D. ChemCatChem. 2011, 3(1), 2465.
51. Somorjai, G. A. Introduction to Surface Chemistry and Catalysis Wiley: New York,
1994, p 667.
52. Engel, T.; Ertl, G. Adv. Catal. 1979, 28, 178.
53. Nieuwenhuys, B. E. Surf. Sci. 1983, 126(13), 307336.
54. Campbell, C. T.; Ertl, G.; Kuipers, H.; Segner, J. Surf. Sci. 1981, 107(1),
220236.
55. Liang, D. B.; Abend, G.; Block, J. H.; Kruse, N. Surf. Sci. 1983, 126(13),
392396.
56. Medvedev, V. K.; Borner, R.; Kruse, N. Surf. Sci. 1998, 401(1), L371L374.
57. Schmidt, W. A.; Block, J. H.; Becker, K. A. Surf. Sci. 1982, 122(3), 409421.
58. Kruse, N.; Schweicher, J.; Bundhoo, A.; Frennet, A.; Visart de Bocarme, T. Top.
Catal. 2008, 48(1), 145152.
59. Kruse, N. Surf. Sci. Lett. 1986, 178(13), A674.
60. Kruse, N.; Gaussmann, A. J. Catal. 1993, 144(2), 525543.
61. Basu, P.; Panayotov, D.; Yates, J. T. J. Am. Chem. Soc. 1988, 110(7),
20742081.
62. Vant Blik, H. F. J.; Van Zon, J. B. A. D.; Huizinga, T.; Vis, J. C.;
Koningsberger, D. C.; Prins, R. J. Am. Chem. Soc. 1985, 107(11), 31393147.
63. Evans, J.; Hayden, B.; Mosselmans, F.; Murray, A. Surf. Sci. 1994, 301(13),
6182.
64. Berko, A.; Solymosi, F. J. Catal. 1999, 183(1), 91101.
65. Frank, M.; Kuhnemuth, R.; Baumer, M.; Freund, H.-J. Surf. Sci. 2000, 454456,
968973.
66. Frank, M.; Kuhnemuth, R.; Baumer, M.; Freund, H. J. Surf. Sci. 1999, 427428,
288293.
67. Magg, N.; Giorgi, J. B.; Frank, M. M.; Immaraporn, B.; Schroeder, T.; Baumer, M.;
Freund, H.-J. J. Am. Chem. Soc. 2004, 126(11), 36163626.
68. Kruse, N.; Schweicher, J.; Bundhoo, A.; Frennet, A.; De Bocarme, T. V. Top. Catal.
2008, 48(14), 145152.
69. Fielicke, A.; Gruene, P.; Meijer, G.; Rayner, D. M. Surf. Sci. 2009, 603(1012),
14271433.
70. Broden, G.; Rhodin, T. N.; Brucker, C.; Benbow, R.; Hurych, Z. Surf. Sci. 1976,
59(2), 593611.
71. Somorjai, G. A.; Rioux, R. M. Catal. Today 2005, 100(34), 201215.
72. Hugo, P.; Wirges, H. P. ACS Sym. Ser. 1978, 65, 498511.
73. Schiefler, J.; Schwentker, H.; Wicke, E. Chemie Ingenieur Techniek 1973, 45(15),
990996.
74. Ertl, G. Angew. Chem. Int. Ed. 2008, 47(19), 35243535.
75. Ertl, G. In Handbook of Heterogeneous Catalysis; Ertl, G., Knozinger, H.,
Schuth, F., Weitkamp, J., Eds.; Wiley VCH: Weinheim, 2008; vol. 3, p 1492.
76. Tao, F.; Dag, S.; Wang, L.-W.; Liu, Z.; Butcher, D. R.; Bluhm, H.; Salmeron, M.;
Somorjai, G. A. Science 2010, 327(5967), 850853.
77. Gritsch, T.; Coulman, D.; Behm, R. J.; Ertl, G. Phys. Rev. Lett. 1989, 63(10),
10861089.
78. Freyer, N.; Kiskinova, M.; Pirug, G.; Bonzel, H. P. Surf. Sci. 1986, 166(1),
206220.
79. Eiswirth, M.; Krischer, K.; Ertl, G. Surf. Sci. 1988, 202(3), 565591.
80. Jakubith, S.; Rotermund, H. H.; Engel, W.; von Oertzen, A.; Ertl, G. Phys. Rev. Lett.
1990, 65(24), 30133016.
81. Bassett, M. R.; Imbihl, R. J. Chem. Phys. 1990, 93(1), 811821.
82. Von Oertzen, A.; Rotermund, H. H.; Mikhailov, A. S.; Ertl, G. J. Phys. Chem. B
2000, 104(14), 31553178.
83. Sales, B. C.; Turner, J. E.; Maple, M. B. Surf. Sci. 1982, 114(23), 381394.
84. Hartmann, N.; Imbihl, R.; Vogel, W. Catal. Lett. 1994, 28(24), 373381.
85. Carlsson, P. A.; Zhdanov, V. P.; Kasemo, B. Appl. Surf. Sci. 2005, 239(34),
424431.
86. Suchorski, Y.; Beben, J.; Imbihl, R. Surf. Sci. 2000, 454(1), 331336.
87. Suchorski, Y.; Beben, J.; James, E. W.; Evans, J. W.; Imbihl, R. Phys. Rev. Lett.
1999, 82(9), 19071910.
88. Hendriksen, B. L. M.; Bobaru, S. C.; Frenken, J. W. M. Surf. Sci. 2004, 552(13),
229242.
89. Hendriksen, B. L. M.; Bobaru, S. C.; Frenken, J. W. M. Top. Catal. 2005, 36(14),
4354.
90. Hendriksen, B. L. M.; Frenken, J. W. M. Phys. Rev. Lett. 2002, 89(4), 046101/
1046101/4.
91. Gao, F.; Goodman, D. W. Phys. Chem. Chem. Phys. 2012, 14, 6688.
92. Gao, F.; Wang, Y.; Cai, Y.; Goodman, D. W. Surf. Sci. 2009, 603(8), 11261134.
93. Gao, F.; Wang, Y.; Cai, Y.; Goodman, D. W. J. Phys. Chem. C 2009, 113(1),
174181.
94. Gao, F.; Wang, Y.; Goodman, D. W. J. Phys. Chem. C 2010, 114(9), 40364043.
95. Goodman, D. W.; Peden, C. H. F.; Chen, M. S. Surf. Sci. 2007, 601(23),
56635665.
96. Blume, R.; Niehus, H.; Conrad, H.; Bottcher, A.; Aballe, L.; Gregoratti, L.;
Barinov, A.; Kiskinova, M. J. Phys. Chem. B 2005, 109(29), 1405214058.
97. Over, H.; Muhler, M. Progr. Surf. Sci. 2003, 72(14), 317.
98. Budde, F.; Heinz, T. F.; Loy, M. M. T.; Misewich, J. A.; De Rougemont, F.;
Zacharias, H. Phys. Rev. Lett. 1991, 66(23), 30243027.
99. Funk, S.; Bonn, M.; Denzler, D. N.; Hess, C.; Wolf, M.; Ertl, G. J. Chem. Phys.
2000, 112(22), 98889897.
100. Stampfl, C.; Schwegmann, S.; Over, H.; Scheffler, M.; Ertl, G. Phys. Rev. Lett.
1996, 77(16), 33713374.
101. Peden, C. H. F.; Goodman, D. W. J. Phys. Chem. 1986, 90(7), 13601365.
102. Over, H.; Muhler, M.; Seitsonen, A. P. Surf. Sci. 2007, 601(23), 56595662.
103. Stampfl, C.; Scheffler, M. Phys. Rev. Lett. 1997, 78(8), 15001503.
104. Stampfl, C.; Scheffler, M. Surf. Sci. 1999, 433, 119126.
105. Sljivancanin, Z.; Hammer, B. Phys. Rev. B Condens. Matter 2010, 81(12), art. no.
121413.
106. Kim, Y. K.; Morgan, G. A.; Yates, J. T. J. Phys. Chem. C 2007, 111(8), 33663368.
107. Over, H.; Kim, Y. D.; Seitsonen, A. P.; Wendt, S.; Lundgren, E.; Schmid, M.;
Varga, P.; Morgante, A.; Ertl, G. Science 2000, 287(5457), 14741476.
108. Kim, Y. D.; Over, H.; Krabbes, G.; Ertl, G. Top. Catal. 2001, 14(14), 95100.
109. Assmann, J.; Narkhede, V.; Breuer, N. A.; Muhler, M.; Seitsonen, A. P.; Knapp, M.;
Crihan, D.; Farkas, A.; Mellau, G.; Over, H. J. Phys. Condens. Matter 2008, 20(18).
110. Over, H.; Balmes, O.; Lundgren, E. Catal. Today 2009, 145(34), 236242.
111. Reuter, K.; Scheffler, M. Appl. Phys. Mater. Sci. Process. 2004, 78(6), 793798.
112. Rosenthal, D.; Girgsdies, F.; Timpe, O.; Blume, R.; Weinberg, G.; Teschner, D.;
Schlogl, R. Zeitschrift fur Physikalische Chemie 2009, 223(12), 183207.
113. Martynova, Y.; Yang, B.; Yu, X.; Boscoboinik, J. A.; Shaikhutdinov, S.;
Freund, H. J. Catal. Lett. 2012, 142(6), 657663.
114. Qin, Z. H.; Lewandowski, M.; Sun, Y. N.; Shaikhutdinov, S.; Freund, H. J. J. Phys.
Condens. Matter 2009, 21(13).
115. Shaikhutdinov, S.; Freund, H. J. Annu. Rev. Phys. Chem. 2012, 63, 619633.
116. Sun, Y. N.; Giordano, L.; Goniakowski, J.; Lewandowski, M.; Qin, Z. H.;
Noguera, C.; Shaikhutdinov, S.; Pacchioni, G.; Freund, H. J. Angew. Chem. Int.
Ed. 2010, 49(26), 44184421.
117. Sun, Y. N.; Qin, Z. H.; Lewandowski, M.; Carrasco, E.; Sterrer, M.;
Shaikhutdinov, S.; Freund, H. J. J. Catal. 2009, 266(2), 359368.
118. Sun, Y. N.; Qin, Z. H.; Lewandowski, M.; Shaikhutdinov, S.; Freund, H. J. Surf.
Sci. 2009, 603(20), 30993103.
119. Qin, Z. H.; Lewandowski, M.; Sun, Y. N.; Shaikhutdinov, S.; Freund, H. J. J. Phys.
Chem. C 2008, 112(27), 1020910213.
120. Fu, Q.; Li, W. X.; Yao, Y.; Liu, H.; Su, H. Y.; Ma, D.; Gu, X. K.; Chen, L.; Wang, Z.;
Zhang, H.; Wang, B.; Bao, X. Science 2010, 328, 1141.
121. Haruta, M. Catal. Today 1997, 36(1), 153166.
122. Shaikhutdinov, S. K.; Meyer, R.; Naschitzki, M.; Baumer, M.; Freund, H. J. Catal.
Lett. 2003, 86(4), 211219.
123. Bondzie, V. A.; Parker, S. C.; Campbell, C. T. In Oxygen adsorption on well-
defined gold particles on TiO2(110), Papers from the 45th National Symposium of
the American Vacuum Society, Baltimore, Maryland, USA, AVS: Baltimore,
Maryland, USA, 1999; pp 17171720.
124. Yoon, B.; Hakkinen, H.; Landman, U.; Wurz, A. S.; Antonietti, J.-M.; Abbet, S. P.;
Judai, K.; Heiz, U. Science 2005, 307(5708), 403407.
125. Hakkinen, H.; Abbet, S.; Sanchez, A.; Heiz, U.; Landman, U. Angew. Chem. Int. Ed.
2003, 42(11), 12971300.
126. Herzing, A. A.; Kiely, C. J.; Carley, A. F.; Landon, P.; Hutchings, G. J. Science
2008, 321(5894), 13311335.
127. Mavrikakis, M.; Stoltze, P.; Nrskov, J. K. Catal. Lett. 2000, 64(2), 101106.
128. Chau, T. D.; Visart De Bocarme, T.; Kruse, N. Catal. Lett. 2004, 98(23), 8587.
129. Freund, H.-J.; Meijer, G.; Scheffler, M.; Schlogl, R.; Wolf, M. Angew. Chem. Int.
Ed. 2011, 50(43), 1006410094.
130. Carrettin, S.; Hao, Y.; Aguilar-Guerrero, V.; Gates, B. C.; Trasobares, S.;
Calvino, J. J.; Corma, A. Chem. Eur. J. 2007, 13(27), 77717779.
131. Stiehl, J. D.; Kim, T. S.; McClure, S. M.; Mullins, C. B. J. Am. Chem. Soc. 2004,
126(6), 16061607.
132. Widmann, D.; Behm, R. J. Angew. Chem. Int. Ed. 2011, 50(43), 1024110245.
133. Widmann, D.; Behm, R. J. Angew. Chem. Int. Ed. 2011, 50(43), 1024110245.

134. Krylov, O. V. Catalysis by Nonmetals: Rules for Catalyst Selection. Physical
Chemistry: A Series of Monographs Academic Press: New York, 1970; vol. 17,
p 283.
135. Katz, M. Adv. Catal. 1953, 5, 177216.
136. Zaki, M. I.; Hasan, M. A.; Pasupulety, L.; Kumari, K. Thermochim. Acta. 1998,
311(12), 97103.
137. Han, Y. F.; Chen, F.; Zhong, Z.; Ramesh, K.; Chen, L.; Widjaja, E. J. Phys. Chem.
B 2006, 110(48), 2445024456.
138. Iablokov, V.; Frey, K.; Geszti, O.; Kruse, N. Catal. Lett. 2010, 134(34), 210216.
139. Yu, R.; Hu, L. H.; Cheng, Z. Y.; Li, Y. D.; Ye, H. Q.; Zhu, J. Phys. Rev. Lett. 2010,
105(22).
140. Jansson, J.; Palmqvist, A. E. C.; Fridell, E.; Skoglundh, M.; Osterlund, L.;
Thormahlen, P.; Langer, V. J. Catal. 2002, 211(2), 387397.
141. Xie, X.; Li, Y.; Liu, Z.-Q.; Haruta, M.; Shen, W. Nature 2009, 458(7239),
746749.
142. Xie, Y.; Dong, F.; Heinbuch, S.; Rocca, J. J.; Bernstein, E. R. Phys. Chem. Chem.
Phys. 2010, 12(4), 947959.
143. Pollard, M. J.; Weinstock, B. A.; Bitterwolf, T. E.; Griffiths, P. R.; Piers
Newbery, A.; Paine Iii, J. B. J. Catal. 2008, 254(2), 218225.
144. Broqvist, P.; Panas, I.; Persson, H. J. Catal. 2002, 210, 198206.
145. Xu, X.-L.; Yang, E.; Li, J.-Q.; Li, Y.; Chen, W.-K. ChemCatChem 2009, 1(3),
384392.
146. Frey, K.; Iablokov, V.; Safran, G.; Osan, J.; Sajo, I.; Szukiewicz, R.; Chenakin, S.;
Kruse, N. J. Catal. 2012, 287, 3036.
147. Boreskov, G. K. Catal. Sci. Tech. 1982, 3, 39137.
148. Twigg, M. V. Platinum Metals Review 2011, 55(1), 4353.
149. Armor, J. N. Catal. Today 1995, 26, 99105.
150. Shelef, M.; Graham, G. W. Catal. Rev. 1994, 36(3), 433457.
151. Jellink, K. Z. Anorg. Allgem. Chem. 1906, 49, 229.
152. Shelef, M.; Kummer, J. T. Chem. Eng. Prog. Symp. Ser. 1971, 64.
153. Hightower, J. W.; van Leirsberg, D. A. In The Catalytic Chemistry of Nitrogen
Oxides; Kimilish, R. L., Larson, J. G., Eds.; Plenum Press: New York, 1975; p 63.
154. Blakely, D. W.; Somorjai, G. A. J. Catal. 1976, 42(2), 181196.
155. Levine, R. D.; Somorjai, G. A. Surf. Sci. 1990, 232(3), 407416.
156. Somorjai, G. A. J. Phys. Chem. 1990, 94(3), 10131023.
157. Banholzer, W. F.; Park, Y. O.; Mak, K. M.; Masel, R. I. Surf. Sci. 1983, 128(23),
176190.
158. Bartram, M. E.; Windham, R. G.; Koel, B. E. Surf. Sci. 1987, 184(12), 5774.
159. Getman, R. B.; Schneider, W. F. J. Phys. Chem. C 2006, 111(1), 389397.
160. Tian, K.; Tu, X.-Y.; Dai, S.-S. Surf. Sci. 2007, 601(15), 31863195.
161. Schwalke, U.; Parmeter, J. E.; Weinberg, W. H. Surf. Sci. 1986, 178(13),
625634.
162. Bertolo, M.; Jacobi, K. Surf. Sci. 1990, 226(3), 207220.
163. Hammer, B.; Norskov, J. K. Adv. Catal. 2000, 45, 71129.
164. Brown, W. A.; King, D. A. J. Phys. Chem. B 2000, 104(47), 11440.
165. Bowker, M. Chem. Soc. Rev. 2008, 37(10), 22042211.
166. Garin, F. Appl. Catal. Gen. 2001, 222(12), 183219.
167. Kleyn, A. W. Progr. Surf. Sci. 1997, 54(34), 407420.
168. Heinzmann, U.; Holloway, S.; Kleyn, A. W.; Palmer, R. E.; Snowdon, K. J. J. Phys.
Condens. Matter 1996, 8(19), 3245.
169. Lahaye, R. J. W. E.; Stolte, S.; Holloway, S.; Kleyn, A. W. J. Chem. Phys. 1996,
104(21), 83018311.
170. Kruse, N.; Abend, G.; Block, J. H. J. Chem. Phys. 1988, 88(2), 13071312.
171. Fiorin, V.; Borthwick, D.; King, D. A. Surf. Sci. 2009, 603(1012), 13601364.
172. Campbell, C. T.; Ertl, G.; Segner, J. Surf. Sci. 1982, 115(2), 309322.
173. Serri, J. A.; Bilotta, R. M.; Pritchard, D. E. J. Chem. Phys. 1982, 77(6),
29402946.
174. Bogicevic, A.; Hass, K. C. Surf. Sci. 2002, 506(12), L237L242.
175. Kim, C. M.; Yi, C. W.; Goodman, D. W. J. Phys. Chem. B 2002, 106(28),
70657068.
176. Brown, W. A.; Sharma, R. K.; King, D. A.; Haq, S. J. Phys. Chem. 1996, 100(30),
1255912568.
177. Dumas, P.; Suhren, M.; Chabal, Y. J.; Hirschmugl, C. J.; Williams, G. P. Surf. Sci.
1997, 371(23), 200212.
178. Brown, W. A.; Gardner, P.; King, D. A. J. Phys. Chem. 1995, 99(18), 70657074.
179. Brown, W. A.; Gardner, P.; Jigato, M. P.; King, D. A. J. Chem. Phys. 1995,
102(18), 72777280.
180. Centi, G.; Perathoner, S. Appl. Catal. Gen. 1995, 132(2), 179259.
181. Chau, T. D.; de Bocarme, T. V.; Kruse, N. Catal. Lett. 2004, 98(2), 8587.
182. Shimada, T.; Mun, B. S.; Nakai, I. F.; Banno, A.; Abe, H.; Iwasawa, Y.; Ohta, T.;
Kondoh, H. J. Phys. Chem. C 2010, 114(40), 1703017035.
183. Nakai, I.; Kondoh, H.; Shimada, T.; Nagasaka, M.; Yokota, R.; Amemiya, K.;
Orita, H.; Ohta, T. J. Phys. Chem. B 2006, 110(51), 2557825581.
Elementary Steps in Heterogeneous Catalysis 37
184. de Koster, A.; van Santen, R. A. Surf. Sci. 1990, 233(3), 366380.
185. Liu, Z.-P.; Hu, P. J. Chem. Phys. 2001, 114(19), 82448247.
186. Logadottir, A.; Rod, T. H.; Norskov, J. K.; Hammer, B.; Dahl, S.; Jacobsen, C. J. H.
J. Catal. 2001, 197(2), 229231.
187. Michaelides, A.; Liu, Z. P.; Zhang, C. J.; Alavi, A.; King, D. A.; Hu, P. J. Am. Chem.
Soc. 2003, 125(13), 37043705.
188. Bligaard, T.; Norskov, J. K.; Dahl, S.; Matthiesen, J.; Christensen, C. H.;
Sehested, J. J. Catal. 2004, 224(1), 206217.
189. Gajdos, M.; Hafner, J.; Eichler, A. J. Phys. Condens. Matter 2006, 18(1), 1340.
190. Ferre, D. C.; van Bavel, A. P.; Niemantsverdriet, J. W. Chemphyschem 2005, 6(3),
473480.
191. Gao, H. W.; Guo, L.; Deng, H. H.; Ge, L. Y.; Yan, T. X. Pol. J. Chem. 2009, 83(6),
11951201.
192. Hansen, E. W.; Neurock, M. Chem. Eng. Sci. 1999, 54(1516), 34113421.
193. Inderwildi, O. R.; Jenkins, S. J.; King, D. A. J. Am. Chem. Soc. 2008, 130(7),
22132220.
194. van Tol, M. F. H.; Nieuwenhuys, B. E. Appl. Surf. Sci. 1993, 67(14), 188197.
195. van Hardeveld, R. M.; van Santen, R. A.; Niemantsverdriet, J. W. J. Phys. Chem. B
1997, 101(40), 79017907.
196. van Hardeveld, R. M.; Hopstaken, M. J. P.; Lukkien, J. J.; Hilbers, P. A. J.;
Jansen, A. P. J.; van Santen, R. A.; Niemantsverdriet, J. W. Chem. Phys. Lett.
1999, 302(12), 98102.
197. van Hardeveld, R. M.; Borg, H. J.; Niemantsverdriet, J. W. J. Mol. Catal. Chem.
1998, 131(13), 199208.
198. van Bavel, A. P.; Hopstaken, M. J. P.; Curulla, D.; Niemanstsverdriet, J. W.;
Lukkien, J. J.; Hilbers, P. A. J. J. Chem. Phys. 2003, 119(1), 524532.
199. Siokou, A.; van Hardeveld, R. M.; Niemantsverdriet, J. W. Surf. Sci. 1998,
402404, 110114.
200. Root, T. W.; Schmidt, L. D.; Fisher, G. B. Surf. Sci. 1983, 134(1), 3045.
201. Root, T. W.; Fisher, G. B.; Schmidt, L. D. J. Chem. Phys. 1986, 85(8), 46794686.
202. Loffreda, D.; Simon, D.; Sautet, P. J. Chem. Phys. 1998, 108(15), 64476457.
203. Kose, R.; Brown, W. A.; King, D. A. J. Phys. Chem. B 1999, 103(41), 87228725.
204. Hopstaken, M. J. P.; van Gennip, W. J. H.; Niemantsverdriet, J. W. Surf. Sci.
1999, 433, 6973.
205. Hopstaken, M. J. P.; Niemantsverdriet, J. W. J. Phys. Chem. B 2000, 104(14),
30583066.
206. Hopstaken, M. J. P.; Niemantsverdriet, J. W. J. Chem. Phys. 2000, 113(13),
54575465.
207. Ho, P.; White, J. M. Surf. Sci. 1984, 137(1), 103116.
208. Hermse, C. G. M.; Frechard, F.; van Bavel, A. P.; Lukkien, J. J.;
Niemantsverdriet, J. W.; van Santen, R. A.; Jansen, A. P. J. J. Chem. Phys.
2003, 118(15), 70817089.
209. DeLouise, L. A.; Winograd, N. Surf. Sci. 1985, 159(1), 199213.
210. Bowker, M.; Guo, Q.; Joyner, R. Surf. Sci. 1991, 253(13), 3343.
211. Loffreda, D.; Simon, D.; Sautet, P. J. Catal. 2003, 213(2), 211225.
212. Amirnazmi, A.; Benson, J. E.; Boudart, M. J. Catal. 1973, 30, 5565.
213. Gohndrone, J. M.; Park, Y. O.; Masel, R. I. J. Catal. 1985, 95, 244.
214. Hatzikos, G. H.; Thomas, P. A.; Lang, J. F.; Gohndrone, J. M.; Masel, R. I. Abstr.
Am. Chem. Soc. 1986, 192, 85.
215. Ge, Q.; Neurock, M. J. Am. Chem. Soc. 2004, 126(5), 15511559.
216. Nakamura, I.; Fujitani, T.; Hamada, H. Surf. Sci. 2002, 514(13), 409413.
217. Rider, K. B.; Hwang, K. S.; Salmeron, M.; Somorjai, G. A. Phys. Rev. Lett. 2001,
86(19), 43304333.
218. Ozensoy, E.; Hess, C.; Loffreda, D.; Sautet, P.; Goodman, D. W. J. Phys. Chem. B
2005, 109(12), 54145417.
219. Vang, R. T.; Wang, J.-G.; Knudsen, J.; Schnadt, J.; Stensgaard, I.; Besenbacher, F.
J. Phys. Chem. B 2005, 109(30), 1426214265.
220. Chatterjee, D.; Deutschmann, O.; Warnatz, J. Faraday Discuss. 2001, 119, 371384.
221. Crocoll, M.; Kureti, S.; Weisweiler, W. J. Catal. 2005, 229(2), 480489.
222. Burch, R.; Watling, T. C. J. Catal. 1997, 169(1), 4554.
223. Mudiyanselage, K.; Yi, C.-W.; Szanyi, J. N. J. Phys. Chem. C 2009, 113(14),
57665776.
224. Mulla, S. S.; Chen, N.; Delgass, W. N.; Epling, W. S.; Ribeiro, F. H. Catal. Lett.
2005, 100(34), 267270.
225. Mulla, S. S.; Chen, N.; Cumaranatunge, L.; Blau, G. E.; Zemlyanov, D. Y.;
Delgass, W. N.; Epling, W. S.; Ribeiro, F. H. J. Catal. 2006, 241(2),
389399.
226. Olsson, L.; Persson, H.; Fridell, E.; Skoglundh, M.; Andersson, B. J. Phys. Chem.
B 2001, 105(29), 68956906.
227. Koop, J.; Deutschmann, O. Appl. Catal. Environ. 2009, 91(12), 4758.
228. Weiss, B. M.; Iglesia, E. J. Phys. Chem. C 2009, 113(30), 1333113340.
229. Srinivasan, A.; Depcik, C. ChemInform 2012, 43(33), 11.
230. Parvulescu, V. I.; Grange, P.; Delmon, B. Catal. Today 1998, 46(4), 233316.
38 Elementary Steps in Heterogeneous Catalysis

231. Siera, J.; Cobden, P.; Tanaka, K.; Nieuwenhuys, B. E. Catal. Lett. 1991, 10(5), 264. Libuda, J.; Freund, H. J. Surf. Sci. Rep. 2005, 57(78), 157298.
335342. 265. Rettner, C. T.; Pfnur, H. E.; Auerbach, D. J. Phys. Rev. Lett. 1985, 54, 2716.
232. Slinko, M.; Fink, T.; Loher, T.; Madden, H. H.; Lombardo, S. J.; Imbihl, R.; Ertl, G. 266. Lee, M. B.; Yang, Q. Y.; Ceyer, S. T. J. Chem. Phys. 1987, 87, 2724.
Surf. Sci. 1992, 264(12), 157170. 267. Lo, T.-C.; Ehrlich, G. Surf. Sci. 1987, 179(1), L19L25.
233. Cobden, P. D.; Siera, J.; Nieuwenhuys, B. E. In Oscillatory reduction of nitric 268. Schouten, F. C.; Kaleveld, E. W.; Bootsma, G. A. Surf. Sci. 1977, 63,
oxide with hydrogen over Pt(100), 38th National Symposium of the American 460474.
Vacuum Society, Seattle, Washington, USA, AVS: Seattle, Washington, USA, 269. Chorkendorff, I.; Alstrup, I.; Ullmann, S. Surf. Sci. 1990, 227(3), 291296.
1992, pp 24872494. 270. Juurlink, L. B. F.; McCabe, P. R.; Smith, R. R.; DiCologero, C. L.; Utz, A. L. Phys.
234. Gruyters, M.; Pasteur, A. T.; King, D. A. J. Chem. Soc. Faraday Trans. 1996, Rev. Lett. 1999, 83(4), 868871.
92(16), 29412949. 271. Schmid, M. P.; Maroni, P.; Beck, R. D.; Rizzo, T. R. J. Chem. Phys. 2002,
235. Hopkinson, A.; Bradley, J. M.; Guo, X.; King, D. A. Phys. Rev. Lett. 1993, 71(10), 117(19), 86038606.
15971600. 272. Juurlink, L. B.; Smith, R. R.; Utz, A. L. Faraday Discuss. 2000, 117, 147160.
236. Voss, C.; Kruse, N. Appl. Surf. Sci. 1995, 8788, 127133. 273. Beck, R. D.; Maroni, P.; Papageorgopoulos, D. C.; Dang, T. T.; Schmid, M. P.;
237. Voss, C.; Kruse, N. Appl. Surf. Sci. 1996, 9495, 186193. Rizzo, T. R. Science 2003, 302(5642), 98100.
238. Keller, G. E.; Bhasin, M. M. J. Catal. 1982, 73(1), 919. 274. Maroni, P.; Papageorgopoulos, D. C.; Sacchi, M.; Dang, T. T.; Beck, R. D.;
239. Hinse, W.; Baerns, M. Z. Chem. 1983, 107, 227. Rizzo, T. R. Phys. Rev. Lett. 2005, 94(24), 246104.
240. Lunsford, J. H. Catal. Today 2000, 63(24), 165174. 275. Juurlink, L. B. F.; Smith, R. R.; Killelea, D. R.; Utz, A. L. Phys. Rev. Lett. 2005,
241. Rostrup-Nielsen, J. R.; Sehested, J.; Norskov, J. K. Adv. Catal. 2002, 47, 65139. 94(20).
242. Rostrup-Nielsen, J. R. In Catalysis, Science and Technology; Anderson J. R., 276. Bisson, R.; Sacchi, M.; Dang, T. T.; Yoder, B.; Maroni, P.; Beck, R. D. J. Phys.
Boudart M., Eds; Springer-Verlag: Berlin, Heideberg, New York, Tokyo, 1984; Chem. A 2007, 111(49), 1267912683.
vol. 5. 277. Killelea, D. R.; Campbell, V. L.; Shuman, N. S.; Utz, A. L. Science 2008,
243. Haggin, J. Chem. Eng. News 1993, 71, 27. 319(5864), 790793.
244. Warren, B. K.; Oyama, S. T. Heterogeneous Hydrocarbon Oxidation. ACS: 278. Killelea, D. R.; Campbell, V. L.; Shuman, N. S.; Smith, R. R.; Utz, A. L. J. Phys.
Washington, 1996. Chem. C 2009, 113(48), 2061820622.
245. Crabtree, R. H. Chem. Rev. 1995, 95, 987. 279. Bisson, R.; Sacchi, M.; Beck, R. D. J. Chem. Phys. 2010, 132(9).
246. Periana, R. A.; Taube, D. J.; Gamble, S.; Taube, H.; Satoh, T.; Fuji, H. Science 280. Zhu, Y. A.; Dai, Y. C.; Chen, D.; Yuan, W. K. J. Mol. Catal. Chem. 2007,
1998, 280, 560. 264(12), 299308.
247. Raja, R.; Ratnasamy, P. Appl. Catal. Gen. 1997, 158(12), L7L15. 281. Abild-Pedersen, F.; Lytken, O.; Engbak, J.; Nielsen, G.; Chorkendorff, I.;
248. Bai, Z.; Chen, H.; Li, W.; Li, B. Int. J. Hydrogen Energ. 2006, 31(7), 899905. Norskov, J. K. Surf. Sci. 2005, 590(23), 127137.
249. Belgued, M.; Pareja, P.; Amariglio, A.; Amariglio, H. Nature 1991, 352(6338), 282. Abild-Pedersen, F.; Greeley, J.; Norskov, J. K. Catal. Lett. 2005, 105(12),
789790. 913.
250. Koerts, T.; Deelen, M. J. A. G.; van Santen, R. A. J. Catal. 1992, 138(1), 101114. 283. Haroun, M. F.; Moussounda, P. S.; Legare, P. J. Mol. Struct. (THEOCHEM) 2009,
251. Wang, L.; Tao, L.; Xie, M.; Tsu, J.; Huang, J.; Xu, Y. Catal. Lett. 1993, 21, 35. 903(13), 8388.
252. Xu, Y.; Bao, X.; Lin, L. J. Catal. 2003, 216(12), 386395. 284. van Grootel, P. W.; van Santen, R. A.; Hensen, E. J. M. J. Phys. Chem. 2011,
253. Bitter, J. H.; Seshan, K.; Lercher, J. A. Top. Catal. 2000, 10(34), 295305. 115(26), 1302713034.
254. Cui, Y. H.; Xu, H. Y.; Ge, Q. J.; Li, W. Z. J. Mol. Catal. Chem. 2006, 243(2), 285. Bunnik, B. S.; Kramer, G. J. J. Catal. 2006, 242(2), 309318.
226232. 286. Kokalj, A.; Bonini, N.; de Gironcoli, S.; Sbraccia, C.; Fratesi, G.; Baroni, S. J. Am.
255. Bruno, T.; Beretta, A.; Groppi, G.; Roderi, M.; Forzatti, P. Catal. Today 2005, Chem. Soc. 2006, 128(38), 1244812454.
99(12), 8998. 287. Xu, J.; Froment, G. F. Aiche J. 1989, 35(1), 97103.
256. de Lucas, A.; Valverde, J. L.; Canizares, P.; Rodriguez, L. Appl. Catal. Gen. 1999, 288. Wei, J.; Iglesia, E. Angew. Chem. Int. Ed. 2004, 43(28), 36853688.
184(1), 143152. 289. Wei, J.; Iglesia, E. Phys. Chem. Chem. Phys. 2004, 6(13), 37543759.
257. de Lucas, A.; Valverde, J. L.; Rodriguez, L.; Sanchez, P.; Garcia, M. T. Appl. Catal. 290. Wei, J.; Iglesia, E. J. Catal. 2004, 225(1), 116127.
Gen. 2000, 203(1), 8190. 291. Wei, J.; Iglesia, E. J. Catal. 2004, 224(2), 370383.
258. Knops-Gerrits, P. P.; Goddard Iii, W. A. J. Mol. Catal. Chem. 2001, 166(1), 292. Wei, J.; Iglesia, E. J. Phys. Chem. B 2004, 108(22), 72537262.
135145. 293. Wei, J.; Iglesia, E. J. Phys. Chem. B 2004, 108(13), 40944103.
259. Juurlink, L. B. F.; Killelea, D. R.; Utz, A. L. Progr. Surf. Sci. 2009, 84(34), 294. Cui, Y.; Zhang, H.; Xu, H.; Li, W. Appl. Catal. Gen. 2007, 318, 7988.
69134. 295. Wang, H. Y.; Ruckenstein, E. Appl. Catal. Gen. 2000, 204(1), 143152.
260. Hamza, A. V.; Madix, R. J. Surf. Sci. 1987, 179(1), 2546. 296. Matsui, N. O.; Anzai, K.; Akamatsu, N.; Nakagawa, K.; Ikenaga, N. O.; Suzuki, T.
261. Beebe, T. P., Jr.; Goodman, D. W.; Kay, B. D.; Yates, J. T., Jr. J. Chem. Phys. Appl. Catal. Gen. 1999, 179(1), 247256.
1987, 87(4), 23052315. 297. Rostrup-Nielsen, J. R.; Hansen, J. H. B. J. Catal. 1993, 144, 38.
262. Larsen, J.; Chorkendorff, I. Surf. Sci. Rep. 1999, 35(58), 163222. 298. Chin, Y.-H.; Buda, C.; Neurock, M.; Iglesia, E. J. Am. Chem. Soc. 2011, 133,
263. Kleyn, A. W. Chem. Phys. Solid Surf. 2003, 11, 79108. 1595815978.
 5.02 Adsorption Properties
Y Hattori and K Kaneko, Shinshu University, Nagano, Japan
T Ohba, Chiba University, Chiba, Japan
2013 Elsevier Ltd. All rights reserved.

5.02.1 Introduction 25
5.02.1.1 Nanoporous System 25
5.02.1.2 Vapor and Supercritical Gas 26
5.02.1.3 Adsorption, Absorption, and Sorption 27
5.02.2 GasSolid Interaction 28
5.02.2.1 Dispersion Interaction 28
5.02.2.2 MoleculePore Interaction 28
5.02.2.3 Contribution of Electrostatic Interaction 31
5.02.3 Vapor Adsorption 31
5.02.3.1 Adsorption Isotherm and Adsorption Mechanism 31
5.02.3.1.1 Type I adsorption isotherm 31
5.02.3.1.2 Type II and III adsorption isotherms 32
5.02.3.1.3 Type IV and V adsorption isotherms 33
5.02.3.2 Role of Modeling in Adsorption in Nanopores 34
5.02.4 Supercritical Gas Adsorption 35
5.02.5 New Developments in Adsorption 37
5.02.5.1 Adsorption of Water Vapor in Hydrophobic Pores 37
5.02.5.2 Quantum Effect in Physical Adsorption 39
5.02.5.3 Gate Adsorption 41
5.02.6 Conclusion 42
Acknowledgments 42
References 43

5.02.1 Introduction method, laser vaporization, and chemical vapor deposition.7,8

5.02.1.1 Nanoporous System
IUPAC classified pores into micropores, mesopores, and macro-
pores with the pore width w1 (see Table 1). Here, w is the slit
width of the slit pore and the diameter of the cylindrical pore, as
shown in Figure 1. The w is not defined by the internuclear
distance and it can be evaluated by the vapor adsorption tech-
nique. Although IUPAC does not define the term nanopore, the
nanopore should be defined as a pore whose width is less than
100 nm, embracing almost efficient pores according to the future
IUPAC recommendation. Representative microporous solids are
zeolites and activated carbons. Zeolites have micropores origi-
nating from the crystal structure. We can modify the preparation
method to add mesopores to zeolites because zeolites having
mesopores are better catalysts without intrapore diffusion sup-
pression, gathering a great affection.2 Activated carbon has an ill-
crystalline structure and so many activated carbons have a wide
pore width distribution. However, activated carbon fiber (ACF)3
and carbide-derived carbon4 have only uniform micropores.
Silica gel is a representative mesoporous solid, used widely as a
desiccant agent. These porous solids have a long history and have
been extensively applied to various technologies. As nanoporous
solids hold promise in the development of sustainable technol-
ogies, new nanoporous solids have been developed since
1990.5,6
Carbon nanotubes were found by transmission electron
microscopy and they were prepared by the arc discharge
Recently, highly pure single-wall carbon nanotubes (SWCNTs)9
and double-wall carbon nanotubes (DWCNTs)10 were prepared
and their unique adsorption properties were reported.11,12
As carbon nanotubes have unique physical properties, such as
high electronic conductivity, high thermal conductivity, and
high mechanical strength, their adsorption properties have
been actively studied for application to energy storage and gas
sensors. Mesoporous silicas that show sharp x-ray diffraction
patterns due to the ordered mesopore structure were developed
in Japan and the United States using the template of supra-
molecular assemblies of surfactant micelles.6,13,14 The novel
family of mesoporous silica has been extending year by
year.15,16 New nanoporous materials, the so-called porous coor-
dination polymers (PCPs) or metalorganic frameworks
(MOFs), appeared on the scientific stage shortly after the carbon
nanotube and mesoporous silica.1721 PCPs have a completely
different framework structure compared to other porous solids.
Zeolites, activated carbons, carbon nanotubes, and mesoporous
silicas have pore walls, whereas PCPs have pillars, not pore walls.
Consequently, PCPs offer the advantage of largepore (lp) vol-
ume and surface area. As PCPs have coordination bondings for
the linkage of molecular pillars, they have excellent designability
for the structure, giving rise to explosive research activity. Their
adsorption properties have attracted a great deal of attention.
The physical and chemical properties of zeolites, nano-
porous carbons including nanotubes, PCPs, and silicas includ-
ing mesoporous silicas are summarized in Table 2. As active

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00502-7 25

26 Adsorption Properties

studies on PCPs have produced new types such as thermally
stable covalent organic frameworks,21 the characteristics of
PCPs can be changed. Table 2 indicates the uniqueness of
porous carbons. They exhibit high electronic conductivity
and so they are applied to electrochemical devices such as
supercapacitors.22,23 Although zeolites have no super high
surface area, they have been widely applied because of their
unique properties. Mesopore-range zeolites are new porous
solids, exhibiting better adsorptivity, ion exchangeability, and
catalysis.24 Thus, both traditional and newly developed nano-
porous solids have been evolving, stimulating research on their
gas adsorption properties.
5.02.1.2 Vapor and Supercritical Gas
We must distinguish vapor and supercritical gas based on the
description of their gas adsorption behaviors. Adsorption
behaviors of vapor and supercritical gas are completely differ-
ent from each other, even though the nature of the gassolid
interaction is identical. Figure 2 shows a phase diagram of
substances. The line of coexistence of liquid and gas is
expressed by the curve OC in Figure 2. Both gas and liquid
are stable below the critical temperature Tc; the gas below Tc is
Table 1 Classification of pores
Micropore w < 2 nm
Mesopore 2 nm < w < 50 nm
Macropore w > 50 nm
named vapor, which has an intrinsic saturated vapor pressure
P0 that depends on the temperature. Here the temperature
dependence of P0 is expressed by the OC curve. Once the
temperature goes over Tc (T > Tc), there is no coexisting region
between gas and liquid. The state at T > Tc and pressure
P > critical pressure Pc is called supercritical fluid state. Mole-
cules in supercritical fluid form clusters with short half-lives
and the density of the supercritical fluid fluctuates remarkably
in space and time. Supercritical fluid has different characteris-
tics compared to both liquid and gas. Saturated vapor pressure
cannot be defined above Tc and the state above Tc is known as
supercritical gas, including supercritical fluid. As the sorption
phenomenon is the formation of a liquid-like molecular
assembly on a solid surface, as described later, vapor coexisting
with the liquid gives a predominant sorption phenomenon.
On the contrary, supercritical gas cannot induce a marked
sorption on solids.
There are many important supercritical gases whose Tc is
lower than ambient temperature. O2, N2, NO, H2, and CH4 are
representatives of supercritical gas. The Tc of CO2 is 304.2 K
and then CO2 at an ambient temperature (<Tc) is called a
subcritical gas being close to supercritical gas rather than
vapor. Table 3 summarizes the physical properties of impor-
tant gases.25,26 Here Tb is the boiling temperature. The units of
the dipole and quadrupole moments are (C m) and (C m2),
respectively. Although NO and CO have both the dipole
and quadrupole moments, their quadrupole moments of
the higher order are not shown because of their negligible
contribution to the gassolid interaction. These dipole and

Nanopore is not recommended by IUPAC, but it is often used for pores whose width is
less than 100 nm. Ultramicropore is often used for a pore whose width is less than P
0.7 nm, although it is not a term officially recommended by IUPAC.

Liquid
Solid C

Supercritical
gas
w O
w
Gas

(a) (b)
Tc T
Figure 1 Schematic diagram of representative pores: (a) slit pore and
(b) cylindrical pore. Figure 2 Phase diagram of substance.

Table 2 Characteristics of nanoporous solids

Zeolite Carbon PCP Silica

Electrical conductivity   
Thermal conductivity   
Thermal stability D
Antioxidation property  
Hydrophobicity  
Ion exchangeability   
Pore size Micropore Micro- and mesopore Micropore Mesopore
Uniform porosity D
Tunability of pore size D
High surface area (>1000 m2 g1) 
 Adsorption Properties 27

Table 3 Properties of important gases

Molecule Tb (K) Tc (K) Pc (MPa) sff (nm) eff/KB (K) Multipole moment Magnetism

H2 20.3 33.0 1.29 0.292 38.0 Quadrupole Diamag

2.1  1040
O2 90.2 154.6 5.04 0.338 126.3 Quadrupole Paramag
1.33  1040
N2 77.3 126.2 3.39 0.363 104.2 Quadrupole Diamag
4.90  1040
NO 121.4 180 6.48 0.347 119 Dipole Paramag
0.158  1030
CO 81.6 132.9 3.50 0.359 110 Dipole Diamag
0.112  1030
CO2 194.7 304.2 7.48 0.376 245.3 Quadrupole Diamag
14.9  1040
CH4 111.6 190.5 4.60 0.372 161.3 Octapole Diamag

Here Tb, Tc, and Pc are the boiling temperature, critical temperature, and critical pressure. The units of dipole and quadrupole moments are C m and C m2, respectively.
Although NO and CO have the quadrupole moment in addition to the dipole moment, their quadrupole moments are omitted. Diamag and Paramag denote diamagnetism
and paramagnetism, respectively.

quadrupole moments induce an electrostatic interaction with a
solid, which can contribute to about 10% of the whole inter-
action at a maximum. The electrostatic interaction can play an
important role in intermolecular orientation in the adsorbed
state. (diamag) and (paramag) denote diamagnetism and para-
magnetism, respectively. Almost all gases exhibit diamagne-
tism and therefore O2 and NO are quite unique in this point.
The paramagnetic property of molecules may affect their
adsorption on a solid remarkably. NO at an ambient temper-
ature is a supercritical gas. Consequently, NO molecules are
predominantly not physically adsorbed even on nanoporous
solids at room temperature. With a spinspin interaction, NO
molecules form the dimer, which is the vapor at room temper-
ature. Addition of iron oxide nanoparticles at the entrance of
carbon micropores can enhance the NO dimerization, leading
to the predominant physical adsorption of NO on the nano-
porous carbon; the adsorption amount of NO is over 30% of
the weight of the adsorbent.27,28 The dimer formation of O2
molecules in the pores of PCP at low temperatures is evidenced
by x-ray diffraction analysis.29 The dimer formation of O2 in
the pores of zeolite at low temperatures is also reported from
the magnetic susceptibility measurement.30 Thus, the magnetic
property of gas must be taken into account in adsorption on
nanoporous solids.31
5.02.1.3 Adsorption, Absorption, and Sorption
Brunauer describes the terminology of adsorption, absorption,
and sorption in his book32 thus: The molecules that disappear
from the gas phase either enter the inside of the solid, or
remain on the outside, attached to its surface. The former
phenomenon is called absorption, the latter adsorption.
Often the two occur simultaneously; the total uptake of the
gas is then designated by the term sorption. (The term sorption
was introduced by J. W. McBain in 1909). This definition of
adsorption, absorption, and sorption is still valid according to the
book by Rouquerol, Rouquerol, and Sing,33 although adsorp-
tion is more rigorously defined as enrichment of one or more
of the components in the region between two bulk phases.
Thus, the definition of adsorption, absorption, and sorption is too
phenomenological and unclear. One of the authors has pro-
posed the related terminologies of physical adsorption, chem-
isorption, absorption, and occlusion based on the structural
change in molecules and/or solids.34,35 These four interactions
are distinguished explicitly by the changes in the atomic struc-
ture of a molecule and/or a solid. No structural change occurs
in the molecule and the solid on physical adsorption. This is
because the main attractive interaction in physical adsorption
is the one of dispersion. On the other hand, chemisorption
induces an intensive interaction between the molecule and
the solid, changing the molecular structure because of the
chemical bond formation between the molecule and the
solid surface. The bulk solid does not change its structure on
chemisorption, although the solid surface does. The major
attractive interaction between a molecule and a solid is the
one of dispersion as in the case of absorption. Absorption of
molecules changes the solid structure but there is no molecular
structure change. Clay minerals often swell on contacting with
water vapor.36 This is a typical example of absorption. Gate
adsorption37 recently found in the PCPs is also a good example
of absorption, as described later. Also, absorption of molecules
in PCPs can be regarded as clathrate formation between the
molecules and PCP crystals.38 Occlusion follows structural
changes in the molecule and the solid. The well-known exam-
ple is the interaction between H2 and Pd solid; the H2 molecule
dissociates into hydrogen atoms, forming a new lattice consist-
ing of hydrogen and Pd atoms. Thus, occlusion changes the
structure of both molecules and solid. This classification can
avoid confusion in understanding an interaction of molecules
with a nanoporous solid. Here, sorption can be used for the
process of physical adsorption and absorption in which mol-
ecules do not change their structure. However, the term adsorp-
tion has often been used in the literature to mean sorption. In
this chapter too, adsorption is often used with the same mean-
ing as sorption. Of course, there can be a medium or a hybrid
interaction between two types of the interaction. In particular,
hydrogen bonding still presents an ambiguity in the classifica-
tion, because the hydrogen-bonding energy covers from several
kJ mol1 to 160 kJ mol1, being comparable to the dispersion
interaction to weak chemical bonding.39
28 Adsorption Properties

5.02.2 GasSolid Interaction Here, ssf and esf have the physical meaning corresponding
5.02.2.1 Dispersion Interaction to eqn [6]. ssf and esf are simply approximated by Lorentz
Berthelot rules as given by eqn [8].
The predominant attractive interaction of sorption is the disper-
sion interaction. The origin of the dispersion interaction can be p ssf sff sss
esf eff ess , [8]
understood using a quantum mechanical theory. We consider 2
two neutral molecules, (a) and (b), each of which has one These LJ parameters are available in the literature.25,26 The
electron, numbered 1 and 2. The coordinates of these electrons gassolid interaction is obtained by addition of the pair-
1 and 2 are assumed to be (x1, y1, z1) and (x2, y2, z2). When interaction of eqn [7]. The gassolid interaction Fsf (z) is
electrons 1 and 2 interact with each other on approaching expressed by the vertical distance z between a molecule and
the opposite molecule, the perturbation term H0 is expressed by the solid surface, as given by eqn [9].

s 9
s 3
sf sf
e2 Fsf z Asf  Bsf [9]
H 0 x1 x2 y1 y2  2z1 z2 [1] z z
r3
The attractive interaction between a molecule and a solid is
where r is the distance between the two molecules. The primary
still a short-range interaction. Figure 3 shows the relative
perturbation energy becomes zero and the secondary pertur-
change in the attractive term in eqn [9] with the distance. The
bation is suggestive. When the wave functions of the two
abscissa is normalized by ssf. The effective range of the disper-
molecules having electrons 1 and 2 are given by C1(1) and
sion interaction between a molecule and solid is four times the
C2(2), respectively, the wave function C of the system is
distance of ssf; the effective distance is 1.5 nm at best for a gas
molecule. Then, the nano-range structure in a solid is essen-
1
C p fC1 1 C2 2g [2] tially important for gas adsorption.
2
00
Thus, the second order perturbation energy E0 can be
obtained as 5.02.2.2 MoleculePore Interaction
00 e 2 2
4
PCPs have a pillar-frame structure, different from the wall-
E0 coefficient r r [3]
r6 1 2 frame structure of zeolite and nanoporous carbon. Here, gas
Here solid interactions are compared for both frame structures. We
introduce a neutral pillar consisting of 20 carbon atoms. A
r 2i C*i iri2 Ci i dti i 1, 2 [4] nitrogen molecule is placed on the center of the pillar at the
vertical distance z. The dispersive interaction of the molecule
If m2i e2 r 2i , mi is the mean value of a dipole moment when with the pillar is obtained by the LJ potential, providing the
the electron is continuously moving in the neutral molecule. potential minimum value. When the pillar number increases,
Then, the dispersion interaction Udisp is given by the potential becomes deeper. Figure 4 shows the potential
minimum change in the interaction with the carbon pillar
coefficientm21 m22 C6 number.40 The potential becomes rapidly deeper up to the
Udisp   6 [5]
r6 r
where C6 is the constant. This dispersion interaction is in
inverse proportion to the sixth power of the intermolecular 0
distance. As the mi stems from an instantaneous deviation
of an electron, a larger molecule has a stronger dispersion
interaction.
We need to take into account the repulsive interaction. The -200
whole intermolecular interaction can be approximated by the
pair potential Uff(r).
F sf /kBK



sff 12
sff 6
Uff r 4eff  [6]
r r -400
which is known as the Lennard-Jones (LJ) potential for a single
component gas. (f) indicates a fluid molecule. Here sff is the
size parameter at Uff 0 and eff is the potential depth. The LJ
potential is applicable to the interaction between neutral mol- -600
ecules and neutral solid surfaces. When the neutral molecule
(f) interacts with the ith surface atom having the mutual dis-
tance r, the moleculesolid atom interaction Csf (ri) is given by 1 2 3 4 5
z/ssf


ssf 12
ssf 6
Csf ri 4esf  [7] Figure 3 The relative change of the attractive interaction between a
r r molecule and solid surface with the mutual distance.
 Adsorption Properties 29

-500

Potential minimum (K)

-1000
Molecule

-1500

Carbon pillars 0 2 4 6 8 10 12
Pillar number
Figure 4 Relationship between the pillar number and interaction potential.

-200
Potential depth (kBK)

-400 1 4

3
-600

2
-800
-0.2 -0.1 0 0.1 0.2
Distance (nm)
Figure 5 Interaction potential profiles of N2 with two mutually parallel-oriented carbon pillars as a function of the interpillar distance of 2.0 (1), 2.2 (2),
2.6 (3), and 3.0 (4) in terms of ssf. Configuration of a molecule and two pillars is shown.

five pillar number; the moleculefive pillar interaction energy
is five times larger than the moleculesingle pillar interaction
energy. The potential minimum is 1707 KkB for the 12 pillar
number. Consequently, a pillar belt of an optimum pillar
number is necessary for production of PCP having enough
interaction strength and high surface area. If the molecule is
inserted by two pillar belts, the interaction is significantly
enhanced.
Here, we discuss the interaction of an N2 molecule with two
carbon pillars. Figure 5 shows the interaction potential profiles
for N2 with two carbon pillars as a function of the interpillar
distance of 2.03.0 in terms of ssf. Here, the effective distance
between two carbon pillars for the interpillar distance 2.0 is
only ssf. Therefore, the repulsion interaction between N2 and
two pillars is significant, giving the shallow minima. The inter-
pillar distance of two ssf provides the minimum value of
733 KkB. Thus, putting a molecule between two pillars lowers
the interaction potential minimum. The overlapping effect
of the interaction potential from both pillars is remarkable.
When the carbon pillars form a regular triangular prism struc-
ture (the triangle length 10sc and prism height 7sc), the
deepest potential in the prism is 1020 KkB, which is slightly
shallower than that of the graphite slit of w 0.7 (1200 KkB).
When pillars build a three-dimensional structure, the structure
leads to a considerably deep interaction potential minimum.
On the contrary, the complex pillar structure sacrifices the
lp volume and surface area. Consequently, the pillar-frame
structure is suitable for adsorbing vapor rather than supercrit-
ical gas.
How can we describe the moleculeslit pore interaction?
The slit pore is the extreme case of the two belt-model. How-
ever, an ordinary slit pore model consists of a slit space
between two bulk slabs whose thickness and breadth are infi-
nite. If the slit width between two surface atoms is H, the
moleculeslit pore interaction p(z) is given by the summation
of both solidmolecule interactions:
p z sf z sf H  z [10]
Here, the molecule is situated at the distance z from one
solid surface. It must be kept in mind that H (the so-called
physical width) is not w (experimentally measured pore
30 Adsorption Properties

width). The contribution of the effective thickness of the elec-
tron cloud of the solid surface must be subtracted. Kaneko et al.
showed the following relation of wide applicability for the
slit pore41:
w H  2z0  sff , z0 0:856 ssf
Here z0 is the distance of the closest approach. In the case of
N2 adsorption on graphite slit pores,
 Also, SWCNT has the interaction potential energy with a
eff
w H  0:24nm, 101:5 K, sff 0:3615 nm
kB
the simple relation given by eqn [12a] is obtained.
We can use the Steele 10-4-3 potential for graphite slit
pores.42 Figure 6 shows the potential profile of an interaction
between a graphite slit pore and a one-center nitrogen mole-
cule or a one-center hydrogen molecule. The interaction poten-
tial becomes deeper with a decrease in w. The potential
minimum for the slit pore of w 0.5 nm is much deeper than
that for the pore of w 1.6 nm, which is slightly deeper than
that of the flat surface. The potential minimum for w 0.5 nm
and N2 is 2020 KkB, whereas that for H2 is only 690 KkB,
twice that of thermal energy at room temperature. This is the
[11]
reason why H2 is not abundantly adsorbed in carbon micro-
pores at room temperature. The interaction profile of a mole-
cule confined with cylindrical pores such as SWCNTs gives a
deeper potential minimum than the slit pore.
[12a] molecule depending on the sign of the nanoscale curvature,
that is, the internal and external wall surfaces. The internal
surfacemolecule interaction provides a deeper potential min-
imum than the external surfacemolecule interaction. The dif-
ference between both depths is about 100 KkB for H2 and
SWCNT of the tube diameter of 3 nm; the potential depth in
the inner position is 420 KkB. SWCNTs form the bundle
structure of a hexagonal symmetry as shown in Figure 7.43

0 0

-500 -500 w = 1.6 nm

1.0 nm
0.7 nm
0.5 nm
-1000 -1000
f/kBK

f/kBK

w = 1.6 nm
1.0 nm
-1500 0.7 nm -1500

-2000 -2000
0.5 nm N2 H2

-2500 -2500
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8
z (nm) z (nm)
Figure 6 Molecular potential profiles of the slit-shaped pore of carbon materials as a function of pore width w for N2 and H2.

A B
Groove Interstitial 0

-400
F/kBK

A B
IP
-800 G

Quantum effective (77 K)

IC Classical

-1200
0.0 0.5 1.0 1.5 2.0 2.5 3.0
r (nm)
Figure 7 The bundle structure of SWCNT and the interaction potential profile of H2 with the bundle (the diameter of SWCNT 1.356 nm).

Figure 7 shows the interaction potential profile between H2
and the SWCNT bundle. The interstitial channel is surrounded
by three tube surfaces, giving the deepest potential minimum
even though all three tubes have the positive sign of the curva-
ture. The interstitial channel has the potential for adsorption of
supercritical H2. Nevertheless, the absolute capacity is too
small. As the interaction potential for carbon nanotubes can
be evaluated using analytical functions,44 carbon nanotubes
have explicit merit as a model system.
5.02.2.3 Contribution of Electrostatic Interaction
Electrostatic interaction must be taken into account in the
intermolecular interaction of a polar molecule. A water mole-
cule is a typical polar molecule. The electrostatic contribution
is added to the LJ potential, as expressed by eqn [12b]
Adsorption Properties 31
The amount of adsorption depends on the equilibrium
pressure of vapor, the adsorption temperature, vapor, and the
amount of solid adsorbent in the system. The adsorption is
measured at a constant temperature, and the relationship
between the adsorption amount and the equilibrium pressure
P is obtained. The equilibrium pressure and the adsorption
amount are expressed by P/P0 and the amount per unit
weight of the adsorbent, respectively. The adsorption
amount is expressed by the volume at STP (273.15 K
and 101.32 kPa) or weight, depending on whether the
manometric or gravimetric method is used. The adsorption
mechanism can be determined from a glance at the adsorption
isotherm. Accordingly, the shape of the adsorption isotherm
must be memorized before analyzing it in a routine way.

X 5.02.3.1 Adsorption Isotherm and Adsorption Mechanism
sff 12
sff 6 4 X 4
1 di dj
Fff r 4eff  [12b]
r r i j6i
4pe 0 rij
Here eff/kB 80.5 K and sff 0.312 nm for water. There are
many potential models for a water molecule. In the case of
TIP5P potential of a five-site model,45 the negatively charged
interaction sites are located symmetrically along the lone-pair
directions at an angle of 109.47 . A charge of 0.241e is placed
on each hydrogen site and charges of 0.241e are placed on
the lone-pair interaction sites. There is no charge on oxygen for
the TIP5P model. Thus, the electrostatic interaction can be
added to the dispersion interaction in the intermolecular inter-
action. The contribution of the charges of an adsorbed molecule
to the moleculesolid interaction can be evaluated with the
image potential approximation for a conducting solid.46 When
we treat dipolar and/or quadrupolar molecules such as SO2,
CO, N2, and CO2, other than water, the electrostatic interaction
is the key to determine both the intermolecular orientation
structure and the adsorbed structure on the wall surface.
5.02.3 Vapor Adsorption
Adsorption properties are measured quantitatively using static
and dynamic methods. The static method gives the so-called
adsorption isotherm, which provides vaporsolid interaction
information in an equilibrium state. Accordingly, the adsorp-
tion isotherm measurement is indispensable to the discussion
of the adsorption properties of a solid. The adsorption iso-
therm of vapor can be measured with manometric and gravi-
metric methods. Details of the adsorption measurement are
not described here.33 Adsorption of vapor can now be mea-
sured from the relative pressure P/P0 of 106 by using a high
quality vacuum system and precision pressure gauges. The low-
pressure vapor adsorption isotherm from P/P0 106 with
multi-measuring points is often called a high-resolution
adsorption isotherm. This is because the adsorption measure-
ments below P/P0 103 provide information on the fine
structure of subnanometer pores and adsorption
mechanisms.47 One of the authors showed the effectiveness
of the adsorption measurement from P/P0 109 for evalua-
tion of subnanometer pores.48
Adsorption of vapors on flat surfaces, mesoporous solids, and
micropores proceeds through inherent mechanisms of multi-
layer (layer-by-layer) adsorption, capillary condensation com-
bined with multilayer adsorption, and micropore filling,
respectively. The different adsorption mechanisms give the
inherent adsorption isotherms. IUPAC recommended six
types of vapor adsorption isotherms, as shown in Figure 8;
however, new recommendations will be made in the near
future. The type I, II, and IV isotherms stem from micropore
filling, multilayer adsorption, and capillary condensation with
multilayer adsorption, respectively. Isotherms of types IV are
discussed in the following sections. A special type VI isotherm
is not discussed in this chapter.
5.02.3.1.1 Type I adsorption isotherm
Micropores have a deep interaction potential well for vapor
molecules and therefore a predominant adsorption begins
from a very low P/P0 range below 103, depending on the
interaction potential depth. Adsorption is almost complete
below P/P0 101, leading to the sharp adsorption uptake
accompanying the plateau due to adsorption saturation. The
adsorption amount in the micropores is significantly larger
than that on the external surfaces at the higher P/P0 region.
Therefore, we can observe a type I isotherm for micropore
filling. In micropores, layer-by-layer adsorption on the pore
walls, which is explained later, does not proceed and almost all
pores are filled below the P/P0 corresponding to the monolayer
formation on the flat surface. This remarkably enhanced
adsorption, which starts even below P/P0 103, is called
micropore filling. When analyzing the stepwise mechanism
of micropore filling, an adsorption isotherm whose abscissa
is expressed by the logarithm of P/P0 is preferred. Figure 9
shows N2 adsorption isotherms of ZMS-5 at 77 K.49 In the
linear P/P0 expression, N2 adsorption increases vertically
at P/P0  0 and then the adsorption curve bends sharply,
becoming parallel to the horizontal axis. This adsorption
isotherm is representative of the type I isotherm. When we
express the adsorption isotherm in the log P/P0, a pre-
dominant adsorption starts below P/P0 103. The type I iso-
therm can be described by the DubininRadushkevich (DR)
equation.50,51
32 Adsorption Properties

I II III

Amount adsorbed IV V VI

Relative pressure P/P0

Figure 8 The six types of gas adsorption isotherms according to the IUPAC classification.

250

Amount adsorbed (103 m3 STP kg1)

250
Amount adsorbed (103 m3 STP kg1)

200 200

150 150

100 100

50 50

0 0
0 0.2 0.4 0.6 0.8 1 105 0.0001 0.001 0.01 0.1 1
(a) P/P0 (b) P/P0

Figure 9 The nitrogen adsorption and desorption isotherms of ZMS-5 at 77 K, (a) linear and (b) logarithmic scale.

(  )
W A 2 P0 aL and WL are the constant and the saturated adsorption
exp  , A RT ln [13]
W0 bE0 P amount corresponding to W0 in the DR equation, respectively.
Nevertheless, the Langmuir adsorption equation was devel-
Here W, W0, A, b, and E0 are the amount of adsorption at oped to describe chemisorption. In the case of micropore
the equilibrium pressure P, the micropore volume, the adsorp- filling, micropores have sufficiently strong interaction energy
tion potential, an affinity coefficient, and the characteristic to induce a Langmuir isotherm. Then, WL and aL have different
adsorption energy, respectively. bE0 gives the isosteric heat of meanings from the chemisorption phenomenon. The evalua-
adsorption at fractional filling f of e1 using the enthalpy of tion of the specific surface area of microporous solids must be
vaporization DHV as follows52: carried out carefully. An accurate determination of monolayers
in micropores is difficult because of the enhanced adsorption.
bE0 DHV qst, 1=e [14] Such micropore effects must be subtracted to obtain a true
The theoretical validity of the DR equation has been ques- value of the specific surface area.5456 The Brunauer, Emmett,
tionable owing to the non-transformation to Henrys law at and Teller (BET) surface area described later gives a highly
P ! 0; the contradiction was resolved by statistical mechanical overestimated value for small micropores.
study.53 Also, the Langmuir adsorption equation of eqn [15] is
effective for description of the type I isotherm, which has a
5.02.3.1.2 Type II and III adsorption isotherms
considerably steep uptake near P/P0 0,
The type II adsorption isotherm is the most familiar of exper-
imental isotherms. The multilayer adsorption of BET theory
W aL P was originally developed to interpret the type II isotherm.57
[15]
WL 1 aL P Hence, this adsorption isotherm is indicative of the multilayer

adsorption mechanism, suggesting the presence of nonporous
flat surfaces or macropores. The so-called BET plot is given
by eqn [16]
P=P0 1 c1 P
[16]
V 1  P=P0 Vm c Vm c P0
V and Vm are the adsorption amount at P/P0 and the mono-
layer capacity, respectively. c is the constant, which is approx-
imated by eqn [17].
c eE1 EL =RT [17]
where E1 and EL are heat of adsorption of the first layer (mono-
layer adsorption) and heat of liquefaction, respectively. The
plot of P/V(P0  P) against P/P0 gives a straight line for the
multilayer adsorption. The addition of the intercept and
slope provides Vm. The available P/P0 region is 0.050.35.
The BET surface area is derived from the above plot using the
occupied N2 molecular area of 0.162 nm2 in the case of N2
adsorption measurement at 77 K. The BET surface area can be
used for a common measure of the specific surface area of
nanoporous materials. Nevertheless, the BET surface area of
microporous materials whose width is more than twice the
probe molecular size (about 0.71.2 nm) is considerably over-
estimated; the surface area is overestimated by about 30% in
the slit micropores. Consequently, the subtracting pore effect
(SPE) method54,55 using the comparison plot58 is recom-
mended to evaluate the true specific surface area of the
wall-frame porous materials. As for the pillar-frame porous
materials, the applicability of the SPE method needs to be
examined. Type III adsorption isotherms are observed for
weakly interacting systems whose c is very small (c  1).
When E1 is almost equal to EL (c  1), the adsorption isotherm
is representative of type III. As the BET equation cannot
describe the adsorption isotherm above P/P0 0.4, the
FrenkelHalseyHill equation is used to describe the adsorp-
tion isotherm in the multilayer adsorption region.59
5.02.3.1.3 Type IV and V adsorption isotherms
The type IV adsorption isotherm comes from capillary conden-
sation after multilayer adsorption in mesopores. Vapor mole-
cules are adsorbed on the walls of mesoporous materials and
then the condensed adsorbed film is formed in the mesopores
due to the depression effect of the saturated vapor pressure of
confined vapor. The saturated vapor pressure-depression is
described by the Kelvin equation:
H1 H2
Amount adsorbed
Relative pressure P/P0
Figure 10 The IUPAC classification of hysteresis loops.
Adsorption Properties 33
P 2gVm cos y
ln  [18]
P0 rm RT
Here, the mean radius is the rm of curvature of the meniscus
of the condensate; g and Vm are the surface tension and molar
volume of the condensate, respectively. y is the contact angle of
the condensate against the pore wall, which is often approxi-
mated to zero (perfect wetting).33 The multilayer of the thick-
ness t is already formed on the pore wall before the capillary
condensation. A more rigorous discussion is carried out using
the modified Kelvin equation with (rm  t) instead of rm in eqn
[18]. The modified Kelvin equation has been used in evalua-
tion of the mesopore size distribution. The type IV and V
adsorption isotherms have an adsorption hysteresis consisting
of adsorption and desorption branches. Both adsorption iso-
therms are associated with capillary condensation. Capillary
condensation, however, does not necessarily give rise to
adsorption hysteresis. When the condensation pressure is
larger than the evaporation pressure, adsorption hysteresis
occurs. A test tube-type mesopore, that is, a cylindrical meso-
pore closed at one end that has a hemispherical shape does not
show adsorption hysteresis. This is because the meniscus is
hemispherical for both capillary condensation and capillary
evaporation and both the condensation and evaporation pres-
sures are identical. On the contrary, a cylindrical mesopore
open at both ends exhibits an explicit adsorption hysteresis.
The meniscus is cylindrical on adsorption, while it is hemi-
spherical on desorption. The rm of the cylindrical meniscus is
twice that of the spherical one, so that the condensation pres-
sure is larger than the evaporation pressure. Thus, the adsorp-
tion hysteresis depends on the geometrical structure of the
mesopores. This classical view has been confirmed for the
simple cylindrical pore in recent molecular simulation studies
by Sarkisov and Monson.60 However, these data are applicable
only to mesopores whose width is larger than ca. 10 nm.
de Boer distinguished five types of adsorption hysteresis
loops denoted by AE. The types A, B, and E are commonly
observed. Later, IUPAC recommended four representative
loops denoted by H1, H2, H3, and H4. The A, E, and B loops
correspond to H1, H2, and H4.1 Figure 10 shows H1-, H2-,
and H3-types. The H1(A)-type hysteresis loop indicates the
presence of tubular mesopores open at both ends, tubular
mesopores with slightly widened parts, and ink bottles. The
H2(E) type loop is indicative of the corpuscular systems having
ill-defined pore sizes and shapes. When the pores giving H1-
type hysteresis have mutual networks, evaporation depends on
H3
34 Adsorption Properties
the neighbor pore, showing the-H2 type hysteresis. The meso-
pores having ink bottles open at both ends and sheet-like
shapes give H3(B)-type loops.61
Mesoporous silicas having regular pore structures have
stimulated fundamental studies on adsorption in mesopores.
The adsorption hysteresis for cylindrical silica mesopores open
at both ends depends on the pore width and adsorption
temperature.15,6264 The hysteresis loop of the N2 adsorption
isotherm at 77 K disappears for pores of 3.24 nm. Also,
adsorption hysteresis is observed in N2 adsorption isotherms
on MCM41 (w 4.4 nm)65 below 72.6 K. In the case of small
mesopore systems, the vapormesopore wall interaction must
be taken into account in addition to the surface tension term,
which is predominant in capillary condensation. Saam and
Cole66 expressed the chemical potential Dms of molecules
in an open cylindrical mesopore on condensation by using
eqn [19].
asc gVm
Dms   3  [19]
Rp  t a approaches have provided new insights. In particular, adsorp-
Here, RP, a, and asc are the radius of mesopores, the effective
radius of multilayer-coated mesopores (a Rp  t), and the
constant, respectively. The first term corresponds to an average
moleculesolid interaction, which is evaluated by analysis
using the FrenkelHalseyHill equation. The second term
expresses the intermolecular interaction. As the chemical
potential of the desorption course is given by a slightly differ-
ent equation whose second term is 2gVm/a, the plausible
adsorption process can be predicted from the chemical poten-
tial with the progress of adsorption or desorption. The pres-
ence of critical pore width of disappearance of adsorption
hysteresis is associated with pore wall corrugation.67
The adsorption hysteresis on well-defined mesopores has
been actively studied experimentally and theoretically.65
Briefly speaking, adsorption and desorption in uniform,
straight mesopores without networks like the mesopores in
MCM41 can be basically understood by the classical interpre-
tation described above, although independent cylinder-
mesopores of small width that are open at both ends do not
give the adsorption hysteresis as discussed earlier. However,
strictly speaking, the multilayer adsorbed film on the meso-
pore walls is stabilized in the metastable state; delayed
capillary condensation occurs above the equilibrium conden-
sation pressure. The delayed capillary condensation occurs
spontaneously near the vapor-like spinodal stemming from
classical van der Waals theory.68 This case gives the H1-type
adsorption hysteresis.
However, adsorption hysteresis of cage-like pores that are
mutually connected through small necks cannot be sufficiently
explained using the classical capillary condensation theory.
According to the classical theory, the wide cage space having
necks (ink-bottle pores) is filled at the pressure corresponding
to the delayed condensation pressure and it remains filled
during desorption until the narrow neck empties first at a
lower pressure; that is, the evaporation of the condensates in
the cage is suppressed by the adsorbed layer in the narrow
neck, which is called pore blocking. When the pore blocking
mechanism is applicable, the desorption branch provides
information on the neck size. The concept of pore blocking
was suggested by McBain in 1935.69 Recent studies have
pointed out the importance of cavitation in evaporation from
the condensate in the cage space.7072 Figure 11 shows sche-
matically the adsorption hysteresis loops for cavitation and
pore blocking mechanisms. The neck diameter determines
the evaporation mechanism of condensates in the cage. The
evaporation occurs at the pressure corresponding to the equi-
librium meniscus in the pore neck for the pore blocking mech-
anism. On the other hand, the evaporation (cavitation)
pressure is not associated with the neck size, and provides no
information on neck size. When the neck size is smaller than
the critical width of wc, spontaneous local density fluctuation
in the narrow neck gives rise to the cavitation in the cage of
large volume. Figure 12 shows the molecular-scale picture on
cavitation. The cage becomes vacant before evaporation of
molecules in the neck pores. The wc is not clearly determined
yet; Ravikovitch and Neimark68 suggested ca. 4 nm for N2
adsorption on mesoporous silicas, which may be associated
with the liquid-like spinodal. Thus, capillary condensation is
not necessarily fully understood, although theoretical
tion in mesopores of connectability is an essential issue to be
elucidated.
5.02.3.2 Role of Modeling in Adsorption in Nanopores
Macroscopic adsorption isotherm measurements cannot
directly show the adsorption mechanism and the structure of
molecules in nanopores and, moreover, surface science tools
are not effective for elucidation of molecular assembly struc-
tures in nanopores due to shielding of electrons and low-
energy lights by the pore walls. The authors applied x-ray to
show the molecular assembly structure confined in carbon
nanopores for water, alcohol, SO2, and organic solution.7376
In the case of single-wall carbons, even infra-red light can
transmit and IR spectroscopy can be applied to unveil the
molecular motion in the nanopores.77 These experimental
studies on the structure of molecular assemblies in nanopores
using synchrotron x-ray have become popular recently. Thus,
structural understanding of adsorption of molecules in nano-
pores is becoming possible. Molecular simulation, being a set
of computer-based techniques derived from statistical mechan-
ics, is indispensable to the understanding of adsorption in
nanopores. This is because molecular simulation can predict
the properties of adsorbed molecules and molecular assembly
structures given a model of intra- and intermolecular inter-
action potentials. Equilibrium and nonequilibrium properties
can be predicted by averaging over many millions of molecular
configurations. The properties can be studied as a function of
external variables such as temperature and pressure. Conse-
quently, the adsorption isotherm is calculated for the model
pore system and the average snapshots can be compared with
the structural data obtained by, for example, x-ray diffraction.
Molecular simulation is able to provide a molecular level
understanding of adsorption in nanopores. Also, molecular
simulation can be applied to adsorption that cannot be exper-
imentally measured because of various restrictions. Thus,
modeling with molecular simulation is essentially important
in adsorption on nanoporous materials. Moreover, molecular
simulation is effective for supercritical gas adsorption as well as
vapor adsorption. However, this chapter cannot cover the
detailed explanation of molecular simulation.7880 The
 Adsorption Properties 35

(a)
w < wc

Amount adsorbed

0 0.2 0.4 0.6 0.8 1 P/P0

(b)
w > wc

0 0.2 0.4 0.6 0.8 1 P/P0

Figure 11 Schematic illustration of cavitation (a) and pore blocking (b) phenomena. wc is the neck diameter inducing cavitation (spontaneous
nucleation of a bubble in the condensate in the cage).

1 Adsorption
Amount adsorbed

0.8

0.6

0.4

0.2
0 Desorption
0 0.2 0.4 0.6 0.8 1
P/P0

Figure 12 The calculated isotherms and the molecular-scale pictures on cavitation in the neck pores. Figures are courtesy of Professor P. Monson.

methodology of density functional theory (DFT) is also not
given in this chapter although DFT studies are useful to predict
adsorption isotherms with shorter calculation time than ordi-
narily used grand canonical Monte Carlo simulation.
5.02.4 Supercritical Gas Adsorption
Little is understood about the fundamentals of supercritical gas
adsorption in spite of its importance; in particular, adsorption
storage of CH4 and H2, which is indispensable to the construc-
tion of a better sustainable society, is directly associated with
supercritical gas adsorption. This is because CH4 and H2 at
ambient temperature are supercritical gases (see Table 3).
One of the reasons for the poor understanding of supercritical
gas adsorption is the difficulty in determining the absolute
adsorption amount from the surface excess mass adsorption.
In the case of supercritical gas adsorption, the interaction
potential well depth is not enough to form the stable adsorbed
layer on the solid surface. Then, the application of high
36 Adsorption Properties
Adsorbed
density
Adsorbed potential
Distance from surface
Figure 13 The molecular density and interaction potential profiles of
adsorbed layers of supercritical gas.
Solid Adsorbed layer Bulk
Density
X posed a general equation of supercritical gas adsorption
r bulk
Y Z
0 L
Distance from surface
Figure 14 Relationship between the surface excess mass and absolute
adsorbed amount. Region (X) expresses the surface excess mass. The
sum of regions (X) and (Y) corresponds to the absolute adsorption. L is
thickness of adsorbed layer.
pressure is imperative to increase the surface excess adsorption
amount for supercritical gas. Figure 13 shows the adsorbed
layer of supercritical gas schematically. The molecular density
profile has no sharp peaks due to the weakly bound state.
Molecules in the gas phase are also populated near the surface,
since the bulk gas phase concentration is high due to the
application of high pressure. Therefore, we must take into
account the contribution by the bulk gas phase in adsorption.
Figure 14 shows the molecular density profile near the
surface. Region (X) denotes the surface excess mass and the
sum of regions (X) and (Y) corresponds to the absolute adsorp-
tion. Here rbulk and L are the bulk gas phase density and the
thickness of the adsorbed layer, respectively. In the case of
vapor adsorption, the surface excess mass is much larger than
the amount corresponding to the region (Y) even below ambi-
ent pressure. Therefore, the absolute adsorption can be approx-
imated by the surface excess mass adsorption. As both
gravimetric and manometric adsorption methods give only
the surface excess mass adsorption, we need to add the bulk
gas phase contribution within an adsorbed layer (the region
(Y)) to obtain the true adsorption amount (absolute
adsorption). The serious issue with respect to the evaluation
of the absolute adsorption is the difficulty in determining the
thickness of the adsorbed layer (or adsorbed layer volume
Vad).8185 There is no established method to determine L or
Vad, although Murata and Kaneko determined it, by the
analysis of the supercritical gas adsorption isotherm, to be
12 nm of L.8284
The relationship between the surface excess mass, nex, and
the absolute adsorption amount, nab, is given by
L
nab rad r dr rad r dr [20]
0 Vad
nab can be evaluated from eqn [21].
nex nab  rbulk Vad [21]
where rad is the density of the adsorbed layer. Here we can
directly measure nex and rbulk by using the adsorption measure-
ment. On the contrary, nab, Vad, L, and rad cannot be measured
directly. However, accurate information on the adsorption
mechanism needs an absolute adsorption isotherm. For exam-
ple, the isosteric heat of adsorption from the surface excess mass
adsorption isotherm gives the negative value.86 The adsorption
isotherm for supercritical gas is different from the vapor adsorp-
tion isotherm shown in Figure 8.87 Murata and Kaneko pro-
isotherms on the basis of classical DFT.82 This general equation
provides important parameters of the average fluidpore wall
and fluidfluid interaction energies from the analysis of the
adsorption isotherm. Also, it gives a simple classification of
supercritical gas adsorption isotherms depending on the relative
strength of the fluidfluid and fluidsurface interactions.83 We
introduce the compression factor zad of the adsorbed layer,
which can be defined by the average adsorbed layer density
hriad, as given by eqn [22],
P
zad [22]
hriad RT
where P is the pressure of the bulk gas phase.
If we transform the adsorption isotherms into the zad 
hriad relations, three types are obtained, as shown in Figure 15.
Figure 15 shows the horizontal line, the linear increase, and the
S-shaped increasing curve, which correspond to the Henry, virial,
and cooperative transition types, respectively. These three types
of zad versus hriad correspond to the three surface excess mass
adsorption isotherms shown in Figure 16. The virial and co-
operative transition types in the zad versus hriad are obtained
from Langmuir (L) and maximum (M) types of the surface excess
mass adsorption isotherms. The H-type guarantees the validity of
Henrys law, indicating the absence of fluidfluid interactions;
the adsorbed gas can be regarded as an ideal gas. When the
fluidsolid interaction is too weak, the H-type is observed.
The L (or virial) type needs the second virial coefficient to
describe the hriad, indicating the presence of a medium range
of the fluidfluid and fluidsolid interactions. When fluid mol-
ecules form a considerably thick adsorbed layer due to the strong
fluidsolid interaction, the repulsive interaction cannot be
neglected. This case gives the M-type or sharp uptake in the low
pressure. In the case of the surface excess mass adsorption iso-
therm of the M-type, the absolute adsorption amount can be

1 400
0.8
0.6
za
0.4
0.2
0 5 10 15 20 25
r (mol l1)
Figure 15 Three types of the compression factor of adsorbed layer, za,
versus adsorbed density.
M
Surface excess
L 5.02.5
H introduction of in situ structural analysis, computer simulation,
P developments in adsorption are shown.
Figure 16 Three types of adsorption isotherms of supercritical gas.
estimated using the analysis of the excess adsorption isotherm.
The detailed procedure is given in the literature.84,88 Figure 17
shows the surface excess mass adsorption and absolute adsorp-
tion isotherms of supercritical O2 on Cu-based PCP(ELM-11) at
196 K. This ELM-11 shows a unique gate adsorption behavior.
A marked decrease in adsorption is observed after gate adsorp-
tion in the surface excess mass adsorption isotherm (Figure 17
(diamond symbols)). However, a reasonable absolute adsorp-
tion isotherm in terms of the absolute adsorption is obtained by
the analysis, as shown in Figure 17 (circle symbols). Thus,
analysis of the surface excess mass adsorption isotherm needs a
basis different from the vapor adsorption isotherm.89
Although a general understanding of supercritical gas
adsorption on nanoporous materials is still not easy, better
adsorbents for storage of CH4 and H2 need to be developed.
Adsorbent design using molecular simulation, which provides
adsorption isotherms for a target molecule-adsorbent system,
is quite promising. Above all, the theoretical PCP structure
designing of better adsorbents for supercritical gases is really
preferable, because PCPs have splendid structural diversity and
Adsorption Properties 37
Absorbed amount of O2 (mg g1)
300
200
100
0
0 2 4 6 8
Fugacity (MPa)
Figure 17 Adsorption isotherms of O2 at 196 K in terms of surface
excess () and absolute amount (). The solid and open symbols denote
adsorption and desorption isotherms, respectively.
designability. A recent review by Getman et al.,80 for example,
has offered a valuable guideline for the synthesis of better
adsorbents for storage of CH4 and H2 based on molecular
simulation studies.
New Developments in Adsorption
Scientific understanding of adsorption on nanoporous mate-
rials has progressed as a result of continuous efforts including
and well-defined nanoporous materials. In particular, a funda-
mental understanding of adsorption processes in micropores
and small mesopores has been obtained. Here three novel
5.02.5.1 Adsorption of Water Vapor in Hydrophobic Pores
When we drop a water droplet on a hydrophobic surface, the
spherical droplet is formed; the more hydrophobic the surface,
the larger the contact angle. This contact behavior is reflected in
adsorption of water vapor on solid surfaces. Hydrophobic
surfaces and pores give the type III and type V adsorption
isotherms, respectively. Then water vapor adsorption isotherms
give a molecular measure of the hydrophobicity of a solid.
Unique water vapor adsorption has been observed on hydro-
phobic zeolites and hydrophobic microporous carbons.9092
Microporous carbons having just a small amount of surface
oxygen groups exhibit a hydrophobic nature for water vapor;
less oxidized surface states can be evidenced by x-ray photo-
electron spectroscopy. Water forms hemispherical droplets on
microporous carbon surfaces of such hydrophobicity. Then,
such carbon surfaces can be regarded as hydrophobic surfaces.
Figure 18 shows water vapor adsorption isotherms of hydro-
phobic microporous carbons whose widths are 0.6 and
1.5 nm. The water vapor adsorption isotherms of nonporous
carbon black and microporous carbon having many surface
oxygen groups (hydrophilic carbon) are shown in Figure 18
38 Adsorption Properties
Adsorbed amount of H2O
Hydrophilic carbon
0 0.2 0.4
P/P0
Figure 18 Adsorption isotherms of water vapor on hydrophilic carbon,
nanoporous carbon (pore width (w) 0.6 nm and w 1.1 nm), and
carbon black.
for comparison. Here, water vapor adsorption isotherms
of microporous carbons whose widths are in the range of
0.81.5 nm are almost similar to each other, although the
larger the pore width, the larger is the uptake P/P0. Although
nonporous carbon black has less than 10% of the surface area
of other microporous carbons, the adsorption amount of
nonporous carbon black is much smaller than 10% of the
adsorption amount of microporous carbons. The presence
of micropores is essentially important in the adsorption of
water vapor. We notice an unusual behavior in microporous
carbons whose pore width is less than 0.6 nm; a considerable
amount of adsorption begins even below P/P0 0.2, leading to
a gradual adsorption increase up to P/P0 0.5 and then a
plateau. A more striking behavior is observed in microporous
carbons of 1.2 nm; the adsorption amount is nil below
P/P0 0.4 and a marked adsorption begins above P/P0 0.8.
Furthermore, a pronounced adsorption hysteresis is observed,
while almost negligible hysteresis is observed in the pores
of 0.6 nm. Similar water vapor adsorption behavior is
observed on SWCNTs and DWCNTs.9396 Tao et al.95 found
the marked higher pressure shift of uptake pressure with van-
ishing subnanometer pores on high-temperature treatment
of DWCNTs.
How can we understand water vapor adsorption in hydro-
phobic carbon micropores? McBain et al. indicated already the
unique nature of water vapor adsorption on hydrophobic
microporous carbon in 1933.97 The water vapor adsorption
isotherm of type V cannot be interpreted by the capillary
condensation mechanism based on the Kelvin equation. Also,
water vapor is not adsorbed on the typical mesoporous carbon
of hydrophobicity below P/P0 0.9.98 The water vapor adsorp-
tion in hydrophobic micropores cannot be understood by the
capillary condensation mechanism. As the water adsorption
does not start from an extremely low P/P0, the water vapor
adsorption in the hydrophobic micropores cannot be inter-
preted by the micropore filling mechanism either.
Water adsorbed in hydrophobic carbon micropores at
ambient temperature is not ordinary liquid. Iiyama et al.73
and Bellissent-Funel et al.99 showed with the aid of x-ray
scattering that the structure of water adsorbed in hydrophobic
w = 1.1 nm
carbon micropores is rather close to that of the solid. Very
recently, Futamura et al. reported negative thermal expansion
of water and cubic ice formation in hydrophobic carbon
micropores by x-ray diffraction and small angle x-ray scattering
experiments over a wide temperature range of 20277 K.100
A similar conclusion is obtained for water adsorption on
SWCNT.96 Thus, the adsorbed state of water molecules in
w = 0.6 nm hydrophobic micropores is very unique, similar to other mol-
ecules adsorbed in micropores with the micropore filling
mechanism, as mentioned earlier. New structural mechanisms
Carbon black
on water vapor adsorption in hydrophobic micropores must
be introduced. The following study on adsorption hysteresis of
water is helpful to understand the adsorption mechanism.
0.6 0.8 1 Equilibration time for measuring adsorption points has a
great effect on the hysteresis loop of water vapor adsorption
isotherms.101 Varying the equilibration time from 5 min to
16 h for each measuring point shifts the adsorption branch,
not the desorption branch; the longer the equilibration time,
the narrower is the breadth of the hysteresis loop. The relation-
ship between the experimental equilibration time and the
breadth of the hysteresis loop suggests a necessary equilibra-
tion time of more than a thousand years for each measuring
point; the metastable state in the adsorption branch is really
stable compared with the thermodynamically stable state in
the desorption branch. This indicates that the metastable state
is quite close to the equilibrium state. Ohba and Kaneko
showed that the formation of pentamers or hexamers of
water molecules of about 0.50.6 nm in size induces enough
stabilization of water molecules in the hydrophobic carbon
micropores.102 As the energy difference between associated
clusters and solid-like equilibrium structures is estimated to
be less than a few kJ mol1, the metastable state on adsorption
is highly stable, as observed. The adsorption on the adsorption
branch can be associated with the growth of cluster size and the
increase of the cluster number.103 Strictly speaking, there are
three elementary structures of monolayers, associated clusters,
and uniformly hydrogen-bonded networks, which are formed
in the pores with the increase in P/P0. The route of the associ-
ated clusters to the uniformly hydrogen-bonded network is
metastable, giving the adsorption hysteresis loop.104 Figure 19
shows the changes in the stabilization energy for each model
adsorbed structure with the fractional filling for the carbon
slit pore of w 0.91.1 nm. In the course of adsorption,
adsorption begins on the route of the cluster formation and
transfers not to the monolayer route of the greatest stability,
but to the uniform layer route, because the transfer from the
cluster model to the monolayer one is kinetically forbidden
according to the molecular dynamics study; the adsorption
process is metastable in the fractional range of 0.40.8. On
the contrary, desorption takes the uniform layer model route
and transforms into the monolayer model at the fractional
filling of 0.8. When the filling decreases to 0.4, the monolayer
model must transform into the cluster one. However, this is
also kinetically forbidden. However, desorption behavior
below the fractional filling of 0.4 does not greatly affect the
adsorption isotherm. The structural difference between the
monolayer structure and the uniformly hydrogen-bonded net-
work in the small micropores of less than 0.7 nm width is only
slight. The transformation between the monolayer model and
 Adsorption Properties 39

Cluster Monolayer Uniform layer

Potential (kBK)
50
Stabilization energy (kJ mol1)

40
-20

30 FH H2
FH D2
Classical LJ
20 -40
0.3 0.4 0.5
Distance (nm)
10
Figure 20 The interaction potential profiles for quantum H2/quantum
D2 and classical H2 at 77 K.

0 0.2 0.4 0.6
Fractional filling
Figure 19 The change in the stabilization energy of each adsorbed
model structure with the fractional filling for the carbon slit pore of
w 1.1 nm. Clusters (triangle), monolayers (rectangle), and uniform
(circle) distribution structure.
cluster model is not kinetically forbidden. Then no hysteresis is
observed.
The above-mentioned adsorption mechanism may be noted
in the cluster-mediated filling mechanism. This mechanism can
be applied to water adsorption on hydrophobic silicalites90,105
and AlPO4-5.106108 In the case of AlPO4-5, perpendicular
uptake is observed around P/P0 0.2. Accordingly, the
hydrogen-bonding network between smaller clusters in the
cylindrical pores of hydrophobicity should be different from
that in the slit-shaped pores of microporous carbons. Still,
water adsorption on hydrophobic micropores has intriguing
issues,109 which many researchers have found challenging to
elucidate.
5.02.5.2 Quantum Effect in Physical Adsorption
Ordinary physical adsorption can be described in terms of
classical physics. However, a light molecule such as H2, D2,
and He has a meaningful fluctuation extent at low temperature.
Kaneko et al. pointed out that He in carbon micropores should
be larger than the classical size at 4.2 K.110,111 Moreh et al.
reported an anomaly of the kinetic energy of He in the carbon
micropores at 4.2 K.112 Tanaka et al. showed that micropore
filling of Ne at 27 K could not be interpreted using a classical
particle model.113 Beenakker et al. proposed a new concept of
quantum molecular sieving for hydrogen isotopes on a cylin-
drical pore using a simple model.114 The theoretical and sim-
ulation studies using the FeynmanHibbs (FH) effective
potential or the rigorous path integral method have been
actively applied to adsorption of quantum fluids.115118 The
applicability of the FH potential is evidenced for adsorption
above 40 K.
0.8 1 
2 00 2 0
VFH VLJ VLJ VLJ [23]
12mkB T V
The quadratic FH effective potential is given by eqn [23].
Here VLJ is the LJ potential function. The FH effective potential
can distinguish the intermolecular interaction of hydrogen iso-
topes, whereas the LJ potential cannot. Figure 20 shows the
interaction potential profiles for quantum H2, quantum D2,
and classical H2 at 77 K. Here the potential of quantum H2 and
D2 is expressed by the effective FH potential, while the interac-
tion potential of classical H2 (or D2) is expressed by the LJ
potential. The LJ potential gives the deepest potential mini-
mum and the equilibrium distance at the potential minimum
is the smallest. The depth and equilibrium distance obtained
from the FH potential depend on the mass of H2 and D2 and
therefore the lighter H2 provides the shallowest potential min-
imum and the largest equilibrium distance at the potential
minimum. If the temperature is lowered, the difference in the
potential depth and the equilibrium distance becomes larger;
that is, the quantum effect is more evident at a lower temper-
ature. The above interaction potential profiles indicate that
lighter isotope molecules have a larger molecular size and
they interact more weakly than heavier ones. This is the reason
for quantum molecular sieving. We can evaluate qualitatively
the fluctuation extent of a quantum molecule with de Broglie
wave length lt h/(2pmkT )1/2. Here m is the molecular mass.
The de Broglie wave lengths of H2 and D2 at 20 and 77 K are
0.27; 0.17 and 0.14; 0.11 nm, respectively. The difference in
the effective molecular sizes of D2 and H2 is only 0.03 nm at
77 K. The absolute difference is very small; nevertheless, it is
quite meaningful. The classical molecular size difference
between N2 and O2 is only 0.02 nm, although the quadrupole
moments of both molecules are different from each other. This
difference is applied to the separation of N2 and O2 with
molecular sieve carbon and zeolite for the industrial produc-
tion of pure N2 and O2 from air, for the benefit of human
society. The effective molecular size difference of hydrogen
isotopes is essentially important for adsorption in the subna-
noscale pores. Figure 21 shows snapshots of quantum H2 and
classical H2 at 40 K in the straight pores of AlPO4-5 whose
diameter is 0.73 nm. The adsorbed molecular numbers of
40 Adsorption Properties
classical H2 and quantum H2 are remarkably different regard-
less of the tiny molecular size difference. The adsorption dif-
ference between D2 and H2 has been experimentally shown for
SWCNH,119 SWCNT,120 ACF,121 AlPO4-5,122 and PCPs,123,124
using the adsorption isotherm measurement.
Quantum molecular sieving behavior on SWCNT is shown
here. Highly pure SWCNTs of 3.0 nm average diameter, pre-
pared by the CVD method and mutually isolated, were soni-
cated in toluene and dried after filtration. This drying
treatment after sonication in toluene gives fine bundles of
SWCNTs due to the capillary force on drying. This SWCNT
sample is denoted bund-SWCNT. On the other hand, the
caps were removed by oxidation treatment; thus-treated
SWCNT is denoted ox-SWCNT. The pore structural parameters
determined by N2 adsorption at 77 K are as follows.125,126
Surface area in m2 g1: SWCNT: 1040, ox-SWCNT: 1900, and
bund-SWCNT: 600; pore volume in ml g1: ox-SWCNT: 0.80,
bund-SWCNT: 0.27. As the geometrical surface area of an
empty SWCNT without caps is 2630 m2 g1, ox-SWCNT sam-
ples are considerably open (30%). The surface area of SWCNT
is close to half that of ox-SWCNT, indicative of the mutually
isolated state. TEM observation evidences the fine bundle for-
mation whose interstitial pore width is 0.52 nm on average.
The small surface area of the bund-SWCNT supports the
fine bundle structure consisting of three to four SWCNTs.
Figures 22 and 23 show H2 and D2 adsorption isotherms of
pore-controlled SWCNTs at 77 and 20 K, respectively. The
adsorption amounts of H2 and D2 on SWCNT at 77 K are
almost comparable to those on bund-SWCNT, although the
surface area of SWCNT is nearly twice that of bund-SWCNT.
This is because H2 and D2 are supercritical at 77 K and adsorp-
tion of H2 and D2 needs strong sites such as interstitial pores,
which the isolated SWCNT does not have. All samples have a
greater adsorption amount of D2 than H2. The interesting
point is that the difference between D2 and H2 for bund-
SWCNT is close to that for ox-SWCNT irrespective of the
great difference in their surface area. This indicates strongly
that the interstitial pores of 0.52 nm in bund-SWCNT are
effective for quantum molecular sieving. H2 is vapor at 20 K.
Therefore, adsorption of hydrogen isotopes on SWCNT
Quantum H2 Classical H2
Figure 21 The snapshots of quantum H2 and classical H2 at 40 K in the
straight pore of AlPO4-5.
samples is very different from that at higher temperature. The
adsorption amount of D2 is much larger than that of H2 for all
three samples. The adsorption amount of each sample corre-
sponds to the surface area. Hydrogen isotope vapor can be
adsorbed even on nonporous surfaces at 20 K. The intriguing
point is that even SWCNT without pores can show an explicit
quantum molecular sieving effect; the adsorption amount of
D2 is much larger than that of H2. This quantum molecular
sieving effect stems from the interaction potential wells at the
external tube walls; that is, the potential well, which is not
necessarily surrounded by physical matters, can induce the
remarkable quantum molecular sieving effect. The advantage
of the quantum molecular sieving effect by the potential well of
the surface has no diffusion restriction. Kumar and Bhatia127
first indicated the importance of the kinetic quantum molecu-
lar sieving effect theoretically. Very recently, the quantum
molecular sieving separation of D2 from the D2H2 mixture
was experimentally evidenced for nanoporous carbon and
zeolite,128 which should accelerate the practical application
10
Adsorption (mol kg1)
5
0
0 0.02 0.04 0.06 0.08 0.1
Pressure (MPa)
Figure 22 H2 and D2 adsorption isotherms of bundled and oxidized
SWCNTs at 77 K. H2, diamond; D2, circle; open symbol, nonoxidized;
solid symbol, oxidized.
60
Adsorption (mol kg1)
40
20
0 0.02 0.04 0.06 0.08
Pressure (MPa)
Figure 23 H2 and D2 adsorption isotherms of bundled and oxidized
SWCNTs at 20 K. H2, diamond; D2, circle; open symbol, nonoxidized;
solid symbol, oxidized.

of the quantum molecular sieving effect. The kinetic issue in
pore filling of quantum fluid is also developing a new adsorp-
tion science.129
5.02.5.3 Gate Adsorption
PCPs or porous MOFs display an extremely large range of
crystal structures. The combination of tunable porosity and
chemistry of the internal spaces has attracted great attention.
Above all, flexible PCPs exhibiting gate opening (transition
from nonporous phase to porous phase) and breathing (two
successive structural transitions) on gas adsorption have stim-
ulated research from fundamental and innovative perspectives.
Here, we focus gate adsorption in Cu-based PCP. Li and
Kaneko found a new type of adsorption on [Cu(bpy)
(H2O)2(BF4)2](bpy) crystals called gate adsorption in 2001.37
As the Cu-based PCP crystals have no effective pores from the
crystal structure and they change the structure for accepting gas
molecules above the threshold pressure, this crystal is named
latent porous crystal (LPC). The LPC has a layer-stacking
structure whose interlayer spacing can be controlled by the
anion, ligand, and central metal atom. Later, similar adsorp-
tion behaviors were observed for several PCPs having similar
layer-stacking structures.130135 Therefore, a more general
name of elastic layer-structured metalorganic framework
(ELM) is proposed for the family of PCPs; LPC is named
ELM-11 after the new nomenclature.135 The specification of
the ELM family is as follows. As Co2(2) and Ni2(3) in
addition to Cu2(1) and BF  
4 (1), CF3SO3 (2), CF3BF3 (OTf)
(3) form similar structures, the sort of metal ion (X) and
anion (Y) are noted by numbers corresponding to metal ion
and anion, which are shown in parenthesis, after ELM as ELM-
XY. Therefore, ELM-23 expresses [Co(bpy)2(OTf)2]. The
adsorption of gases on ELM-11 begins at the threshold pressure
vertically, reaching the saturation; on decreasing the gas pres-
sure, the adsorption amount does not change up to a lower
threshold pressure than that on adsorption, dropping to nil at
the pressure. Therefore, this adsorption is named gate adsorp-
tion/gate desorption, because the effective pores are produced
at the threshold pressure (adsorption gate pressure Pg,a), on
adsorption, while the effective pores are shut at the desorption
threshold pressure (desorption gate pressure Pg,d) on desorp-
tion. Figure 24 shows CO2 adsorption isotherms on ELM-11 at
different temperatures. As mentioned earlier, the adsorption
amount varies almost vertically at adsorption and desorption
gate pressures, giving rise to a rectangular adsorption hystere-
sis. When the temperature decreases, the rectangular hysteresis
loop shifts to a low-pressure side; Pg,a at 273 K is 37.5 kPa,
shifting to 11.8 kPa at 248 K. Similar gate behavior is observed
for other gases of CH4, N2, O2, and Ar on ELM-11.87,135137
The gate adsorption can be expressed by the clathrate forma-
tion equilibrium between ELM-11 and gas molecules, as
shown in eqn [24]; the adsorption gate pressure is an equilib-
rium pressure for the chemical equilibrium of the clathrate
formation.
ELM 11s Mg M:ELM0 s [24]
0
where M and [M:ELM ] denote a guest molecule and clathrate
compound.38,137 van der Waals and Platteeuw showed that the
Adsorption Properties 41
200
Amount of adsorbed CO2 (mg g1)
150
100
50
0
0 20 40 60 80 100
Pressure (kPa)
Figure 24 CO2 adsorption isotherms of ELM-11 at different
temperatures. Solid and open symbols indicate adsorption and
desorption, respectively, 248 K, ; 273 K, ; 298 K, .
ClapeyronClausius equation can be applied to binary
systems.138 The Pg is regarded as the vapor pressure of the
clathrate [M:ELM0 ](s). The volume change is approximated
by the volume of gas (zRT/P, z is the compression factor.
z 1 for an ideal gas). The ClapeyronClausius equation is
given by eqn [25].
d ln P DHf
[25]
d1=T R
Here DHf is the enthalpy charge of the clathrate formation.
The ln Pg,a versus 1/T relation gives the linear plot. The DHf is
27 kJ mol1 for CO2. The gate adsorption of CH4 is not critical
compared with CO2. The linear ClapeyronClausius relation
leads to DHf 13.0  1.2 kJ mol1. These values are reason-
able, because the interlayers are bound with the hydrogen
bonding, which has a stabilization of 10 kJ mol1. In the
case of CO2, no adsorption occurs even after 14 h at the CO2
pressure slightly lower than Pg,a. This fact supports the above
clathrate formation equilibrium.
Gate adsorption and desorption are considerably rapid
processes. Figure 25 shows the changes in the CO2 adsorption
amount after increasing the CO2 pressure from 20 to 98 kPa on
going over each gate pressure at 273 and 298 K. The adsorption
at 273 and 298 K almost finishes within 8 min. When the first-
order adsorption kinetics is assumed, the reverse values of the
first-order adsorption kinetic constants are 1.0 min at 273 K
and 0.40  0.3 min at 298 K. Thus, gate adsorption is a rapid
process.
The chemical potential difference between the adsorption
and desorption gate effects is given by kBTln(Pg,a/Pg,d). When
we assume that the lattice gate is bound by the hydrogen-
bonded spring, the energy difference of the both springs
for adsorption and desorption is expressed by [(1/2)ka,hddi2 
(1/2)kdhdai2]. Here hddi and hdai are the average displacements
on desorption and adsorption, respectively. hddi should be
close to hdai. Then, the average potential energy difference of
the lattice spring may be approximated by (1/2)(kd  kd)hdi2.
When the lattice gate is exposed to the gas, the potential energy
42 Adsorption Properties
160
Amount of adsorbed CO2 (mg g1)
140
120
100
80
60
40
20
0 10 20 30 40
Time (min)
Figure 25 The change in the CO2 adsorption with time after increasing
CO2 pressure from 20 to 98 kPa at different temperatures. , 273 K; ,
298 K.
of the lattice gate is balanced with the chemical potential of the
surrounding gas, leading to eqn [26].
 regions for the narrow-pore (np) form and the large-pore (lp)
1 Pg, a
ka  kd hdi2 kB T ln [26]
2 Pg, d
The assumption of hdi2 0.0050.01 nm provides Dk
(ka  kd) of 1.4  106  10 N m1 for CO2. The force con-
stants of OHO stretching and OHO bending of ice are
(1.61.9  10) and 9 N m1, respectively.139 Accordingly, the
above estimation of the lattice spring for gate behavior is
plausible, suggesting the expansion and shrinkage of the lat-
tices on gate adsorption and desorption, respectively. This
layered Cu-based PCP was studied with synchrotron x-ray
diffraction, infra-red spectroscopy, and EXAFS and systematic
adsorption experiments, showing that gate adsorption was
observed after dehydration. The original chemical formula
of [Cu(bpy)(H2O)2(BF4)2]bpy, whose crystal structure has
been published,140 changes into [Cu(bpy)(BF4)2]bpy (ELM-
11) after dehydration. [Cu(bpy)(H2O)2(BF4)2]bpy has a
hydrogen-based 3D structure, while [Cu(bpy)(BF4)2]bpy has
a complete 2D layer structure. Figure 26 shows the local struc-
ture change in ELM-11 around Cu ion sites on hydration. The
layers in ELM-11 are stacked against each other without effec-
tive pore spaces.141 The dehydration and hydration processes
are almost reversible, being very unique in PCP materials.142
Surprisingly, these layers open at the gate pressure. The inter-
layer distance increases from 0.458 to 0.578 nm by 0.0120 nm
(26%) on CO2 adsorption, accompanied with sliding of the
staggered stacking layers.141 This unusual expansion of the
crystal lattice recovers on desorption. Accordingly, ELM-11
can aspire CO2 without collapsing the lattices. The collective
motion of the lattices in order to adapt to the surrounding
atmosphere is really interesting. Kondo et al. evidenced an
essential role of surface adsorption inducing the collective
lattice change.143 There are few ELM family compounds that
exhibit gate adsorption phenomena, although more than 50
related crystals were synthesized and analyzed.135,137,144146
PCPs exhibiting breathing phenomenon such as MIL-53147
and MIL-88148are also important to develop efficient adsorp-
tion storage and separation processes as well as the ELM family
H2O
Figure 26 The local structure change in ELM-11 around Cu ion sites on
hydration.
of the gate effect. We must elucidate theoretically the mecha-
nism of the gate and breathing effects. Coudert et al.134,149,150
introduced the osmotic statistical ensemble that accounts for
50 the presence of a flexible adsorbent with variable unit cells,
leading to the osmotic potential involving the free energy of
the adsorbent phase, the adsorption isotherm, and molar vol-
ume of the pure gas as a function of pressure. They discussed
the adsorption-induced structural transition of the flexible PCP
with the osmotic potential using the Henry constant and the
saturated adsorption amount given by the experimental
adsorption isotherm. This theoretical analysis can provide the
phase diagram, which indicates the thermodynamic stability
one for the breathing effect. This osmotic potential analysis is
also effective in describing the gate effect. Watanabe et al.151,152
interpreted the gate adsorption phenomenon with grand free
energy calculation and grand canonical Monte Carlo simula-
tion using the model of LJ molecules. Their approach can
describe well the gate-opening and gate-closing pressures and
the adsorption isotherm.
These theoretical approaches can be applied to understand
the adsorption-induced structural transition for a target mole-
culePCP system. As PCPs showing the adsorption-induced
structure transition are promising to innovative and highly
efficient separation and storage system of valuable gases,135
the progress in the theoretical understanding should accelerate
the practical application of the PCPs to sustainable
technologies.
5.02.6 Conclusion
Adsorption on nanoporous materials has a great demand in
innovative eco-technologies. A fruitful bridging between mate-
rials chemistry and adsorption science and technology is
highly needed to accelerate the growth of eco-technologies.
Finally, the necessity of fundamental studies on gas adsorption
on nanoporous materials should be emphasized to promote
the construction of eco-societies. For a related chapter in this
Comprehensive, we refer to Chapter 9.34.
Acknowledgments
KK has been supported by Exotic Nanocarbons, Japan Regional
Innovation Strategy Program by the Excellence, JST. This work
is partially supported by Grant-in-Aid for Scientific Research
(A) (24241038) by JSPS. The authors wish to thank
Dr H. Tanaka for sending some figures of his work and for
the valuable discussions.

References
1. Sing, K. S. W.; Everett, P. H. R.; Haul, A. W.; Moscou, L.; Pierotti, R. A.;
Rouquerol, J.; Siemieniewska, T. Pure Appl. Chem. 1985, 57, 603619.
2. Tao, Y.; Kanoh, H.; Abrams, L.; Kaneko, K. Chem. Rev. 2006, 106, 896910.
3. Kaneko, K. J. Membr. Sci. 1994, 96, 5989.
4. Gogotsi, Y.; Dash, R. K.; Yushin, G.; Yildirim, T.; Laudisio, G.; Fischer, J. E. J. Am.
Chem. Soc. 2005, 127, 1600616007.
5. Iijima, S. Nature 1991, 354, 5658.
6. Yanagisawa, T.; Shimizu, T.; Kuroda, K.; Kato, C. Bull. Chem. Soc. Jpn. 1990, 63,
988992.
7. Oberlin, A.; Endo, M.; Koyama, T. J. Cryst. Growth 1976, 32, 335349.
8. Jorio, A., Dresselhaus, M. S., Dresselhaus, G., Eds.; Carbon Nanotubes:
Advanced Topics in the Synthesis, Structure, Properties and Application;
Springer: New York, 2008; Vol. 101.
9. Hata, K.; Futaba, D. N.; Mizuno, K.; Namai, T.; Yumura, M.; Iijima, S. Science
2004, 306, 13621364.
10. Endo, M.; Muramatsu, H.; Hayashi, T.; Kim, Y. A.; Terrones, M.;
Dresselhaus, M. S. Nature 2005, 433, 476.
11. Miyamoto, J.; Hattori, Y.; Noguchi, D.; Tanaka, H.; Ohba, T.; Utsumi, S.;
Kanoh, H.; Kim, Y. A.; Muramatsu, H.; Hayashi, T.; Endo, M.; Kaneko, K. J. Am.
Chem. Soc. 2006, 128, 1263612637.
12. Tao, Y.; Endo, M.; Inagaki, M.; Kaneko, K. J. Mater. Chem. 2011, 21, 313323.
13. Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature
1992, 359, 710712.
14. Inagaki, S.; Fukushima, Y.; Kuroda, K. J. Chem. Soc. Chem. Commun. 1993,
680682.
15. Thommes, M. In Nanoporous Materials: Science and Engineering; Lu, G. Q.,
Zhao, X. S., Eds.; Imperial College Press: London, 2004; Chapter 11.
16. Wan, Y.; Zhao, D. Chem. Rev. 2007, 107, 28212860.
17. Mori, W.; Sato, T.; Kato, C. N.; Takei, T.; Ohmura, T. Chem. Rec. 2005, 5,
336351.
18. Tanaka, D.; Kitagawa, S. Chem. Mater. 2008, 20, 922931.
19. Barton, T. J.; Bull, L. M.; Klemperer, W. G.; Loy, D. A.; McEnaney, B.; Misono, M.;
Monson, P. A.; Pez, G.; Scherer, G. W.; Vartuli, J. C.; Yaghi, O. M. Chem. Mater.
1999, 11, 26332656.
20. Salles, F.; Maurin, G.; Serre, C.; Llewellyn, P. L.; Knofel, C.; Choi, H. J.;
Filinchuk, Y.; Oliviero, L.; Vimont, A.; Long, J. R.; Feerey, G. J. Am. Chem. Soc.
2010, 132, 1378213788.
21. Dawson, R.; Cooper, A. I.; Adams, D. J. Prog. Polym. Sci. 2012, 37, 530563.
22. Chmiola, J.; Yushin, G.; Gogotsi, Y.; Portet, C.; Simon, P.; Taberna, P. L. Science
2006, 313, 17601763.
23. Zhang, L. L.; Zhao, X. S. Chem. Soc. Rev. 2009, 38, 25202531.
24. Na, K.; Jo, C.; Kim, J.; Cho, K.; Jung, J.; Seo, Y.; Messinger, R. J.; Chmelka, B. F.;
Ryoo, R. Science 2011, 333, 328332.
25. Hirschfelder, J. O.; Curtiss, C. F.; Bird, R. B. Molecular Theory of Gases and
Liquids; John Wiley and Sons: New York, 1964; Chapter 13.
26. Rigby, M.; Smith, E. B.; Wakeham, W. A.; Maitland, G. C. The Forces Between
Molecules; Oxford Science Publication, 1986; Chapter 1.
27. Kaneko, K. Langmuir 1987, 3, 357363.
28. Kaneko, K.; Fukuzaki, N.; Ozeki, S. J. Chem. Phys. 1987, 87, 776777.
29. Kitaura, R.; Kitagawa, S.; Kubota, Y.; Kobayashi, T.; Kindo, K.; Mita, Y.;
Matsuo, A.; Kobayashi, M.; Chang, H.-C.; Ozawa, T. C.; Suzuki, M.; Sakata, M.;
Takata, M. Science 2002, 298, 23582361.
30. Takaishi, T. J. Chem. Soc. Faraday Trans. 1997, 93, 12571260.
31. Kanoh, H.; Kaneko, K. J. Phys. Chem. 1995, 99, 57465748.
32. Brunauer, S. The Adsorptional Gases and Vapors; Oxford University Press,
1944; p 3.
33. Rouquerol, F.; Rouquerol, J.; Sing, K. Adsorption by Powders and Porous Solids:
Principles, Methodology and Applications; Academic Press: London, 1999.
34. Hanzawa, Y.; Kaneko, K. In Carbon Alloys; Yasuda, E., Inagaki, M., Kaneko, K.,
Endo, M., Tanabe, Y., Eds.; Elsevier: Amsterdam, 2003; Chapter 20.
35. Ustumi, S.; Kaneko, K. In Nanoporous Materials; Kanellopoulos, N., Ed.; CRC
Press, 2011; Chapter 10.
36. Pinnavaia, T. J. Science 1983, 220, 365371.
37. Li, D.; Kaneko, K. Chem. Phys. Lett. 2001, 335, 5056.
38. Noguchi, H.; Kondoh, A.; Hattori, Y.; Kanoh, H.; Kajiro, H.; Kaneko, K. J. Phys.
Chem. B 2005, 109, 1385113853.
39. Jeffrey, G. A. An Introduction to Hydrogen Bonding; Oxford University Press,
1997.
40. Ohba, T.; Kaneko, K. Unpublished results.
41. Kaneko, K.; Cracknell, R. F.; Nicholson, D. Langmuir 1994, 10, 46064609.
42. Steele, W. A. The Interaction of Gases with Solid Surfaces; Pergamon Press, 1974;
Chapter 2.
Adsorption Properties 43
43. Tanaka, H.; Fan, J.; Kanoh, H.; Yudasaka, M.; Iijima, S.; Kaneko, K. Mol. Simul.
2005, 31, 465474.
44. Murata, K.; Kaneko, K.; Steele, W. A.; Kokai, F.; Takahashi, K.; Kasuya, D.;
Hirahara, K.; Yudasaka, M.; Iijima, S. J. Phys. Chem. B 2001, 105,
1021010216.
45. Mahoney, M. W.; Jorgensen, W. L. J. Chem. Phys. 2000, 112, 89108922.
46. Wang, Z. W.; Kaneko, K. J. Phys. Chem. 1995, 99, 1671416721.
47. Kakei, K.; Ozeki, S.; Suzuki, T.; Kaneko, K. J. Chem. Soc. Faraday Trans. 1990,
86, 371376.
48. Sunaga, M.; Ohba, T.; Suzuki, T.; Kanoh, H.; Hagiwara, S.; Kaneko, K. J. Phys.
Chem. B 2004, 108, 1065110657.
49. Tao, Y.; Kanoh, H.; Kaneko, K. J. Am. Chem. Soc. 2003, 125, 60446045.
50. Bering, B. P.; Dubinin, M.; Serpinsky, V. V. J. Colloid Interface Sci. 1966, 21,
378393.
51. Kaneko, K.; Shimizu, K.; Suzuki, T. J. Chem. Phys. 1992, 97, 87058711.
52. Kawazoe, K.; Astakhov, V. A.; Kawai, T.; Eguchi, Y. Kagaku Kogaku 1971, 35,
10061012.
53. Chen, S. G.; Yang, R. T. Langmuir 1994, 10, 42444249.
54. Kaneko, K.; Ishii, C.; Kanoh, H.; Hanzawa, Y.; Setoyama, N.; Suzuki, T. Adv.
Colloid Interface Sci. 1998, 7677, 295320.
55. Setoyama, N.; Suzuki, T.; Kaneko, K. Carbon 1998, 36, 14591467.
56. Ohba, T.; Kaneko, K. J. Phys. Chem. B 2002, 106, 71717176.
57. Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309319.
58. Atkinson, D.; McLeod, A. I.; Sing, K. S. W. J. Chim. Phys. 1984,
81, 791794.
59. Adamson, A. W.; Gast, A. P. Physical Chemistry of Surfaces; John Wiley and
Sons: New York, 1997; p 628.
60. Sarkisov, L.; Monson, P. A. Langmuir 2001, 17, 76007604.
61. Linsen, B. G.; Van Der Heuvel, A. In Solid-Gas Interface; Flood, E. A., Ed.; Marcel
Dekker Inc: New York, 1967; Vol. 2, Chapter 35.
62. Branton, P. J.; Hall, P. G.; Sing, K. S. W.; Reichert, H.; Schuth, F.; Unger, K.
J. Chem. Soc. Faraday Trans. 1994, 90, 29652967.
63. Branton, P. J.; Kaneko, K.; Setoyama, N.; Sing, K. S. W.; Inagaki, S.;
Fukushima, Y. Langmuir 1996, 12, 599600.
64. Kruk, M.; Jaroniec, M. J. Phys. Chem. B 2002, 106, 47324739.
65. Morishige, K. Adsorption 2008, 14, 157163.
66. Cole, M. W.; Saam, W. F. Phys. Rev. Lett. 1974, 32, 985988.
67. Inoue, S.; Hanzawa, Y.; Kaneko, K. Langmuir 1998, 14, 30793081.
68. Ravikovitch, P. I.; Neimark, A. V. Langmuir 2002, 18, 98309837.
69. McBain, J. W. J. Am. Chem. Soc. 1935, 57, 699700.
70. Monson, P. A. In Characterization of Porous Solid VIII; Kaskel, S., Llewellyn, P.,
Rodriguez-Reinoso, F., Seaton, N., Eds.; Royal Society of Chemistry: Cambridge,
2009; pp 103110.
71. Morishige, K.; Tateishi, M.; Hirose, F.; Aramaki, K. Langmuir 2006, 22,
92209224.
72. Thommes, M.; Smarsly, B.; Groenewolt, M.; Ravikovitch, P. I.; Neimark, A. V.
Langmuir 2006, 22, 756764.
73. Iiyama, T.; Nishikawa, K.; Otowa, T.; Kaneko, K. J. Phys. Chem. 1995, 99,
1007510076.
74. Kaneko, K. Colloid Surf. A 1996, 109, 319333.
75. Iiyama, T.; Ohkubo, T.; Kaneko, K. In Recent Advances in Gas Separation by
Microporous Ceramic Membranes; Kanellopoulos, N. K., Ed.; Elsevier:
Amsterdam, 2000; pp 3566.
76. Tanaka, A.; Iiyama, T.; Ohba, T.; Ozeki, S.; Urita, K.; Fujimori, T.; Kanoh, H.;
Kaneko, K. J. Am. Chem. Soc. 2010, 132, 21122113.
77. Hashimoto, S.; Fujimori, T.; Tanaka, H.; Urita, K.; Ohba, T.; Kanoh, H.; Itoh, T.;
Asai, M.; Sakamoto, H.; Niimura, S.; Endo, M.; Rodoriguez-Reinoso, F.;
Kaneko, K. J. Am. Chem. Soc. 2011, 133, 20222024.
78. Nicholson, D.; Parsonage, N. G. Computer Simulation and the Statistical
Mechanics of Adsorption; Academic Press: London, 1982.
79. Gubbins, K. E., Quirke, N., Eds.; In Molecular Simulation and Industrial
Applications; Gordon and Breach Science Publishers: Amsterdam, 1996.
80. Getman, R. B.; Bae, Y. S.; Wilmer, C. E.; Snurr, R. Q. Chem. Rev. 2012, 112,
703723.
81. Murata, K.; El-Merraoui, M.; Kaneko, K. J. Chem. Phys. 2001, 114, 41964205.
82. Murata, K.; Kaneko, K. J. Phys. Chem. B 2001, 105, 84988503.
83. Murata, K.; Yudasaka, M.; Iijima, S.; El-Merraoui, M.; Kaneko, K. J. Appl. Phys.
2002, 91, 1022710229.
84. Murata, K.; Miyawaki, J.; Kaneko, K. Carbon 2002, 40, 425428.
85. Sircar, S. Ind. Eng. Chem. Res. 1999, 38, 36703682.
86. Salem, M. M. K.; Braeuer, P.; Szombathely, M. B.; Heuchel, M.; Harting, P.;
Quitzsch, K.; Jaroniec, M. Langmuir 1998, 14, 33763389.
87. Donohue, M. D.; Aranovich, G. L. Fluid Phase Equilib. 1999, 158160,
557563.
44 Adsorption Properties
88. Noguchi, H.; Kondo, A.; Hattori, Y.; Kajiro, H.; Kanoh, H.; Kaneko, K. J. Phys.
Chem. C 2007, 111, 248254.
89. Gumma, S.; Talu, O. Langmuir 2010, 26, 1701317023.
90. Desbiens, N.; Demachy, I.; Fuchs, A. H.; Kirsch-Rodeschini, H.; Soulard, M.;
Patarin, J. Angew. Chem. Int. Ed. 2005, 44, 53105313.
91. Miyawaki, J.; Kanda, T.; Kaneko, K. Langmuir 2001, 17, 664669.
92. Asai, M.; Ohba, T.; Iwanaga, T.; Kanoh, H.; Endo, M.; Campos-Delgado, J.;
Terrones, M.; Nakai, K.; Kaneko, K. J. Am. Chem. Soc. 2011, 133, 1488014883.
93. Bekyarova, E.; Hanzawa, Y.; Kaneko, K.; Silvestre-Albero, J.; Sepulveda-
Escribano, A.; Rodriguez-Reinoso, F.; Kasuya, D.; Yudasaka, M.; Iijima, S. Chem.
Phys. Lett. 2002, 366, 463468.
94. Wang, H. J.; Xi, X. K.; Kleinhammes, A.; Wu, Y. Science 2008, 322, 8083.
95. Tao, Y.; Muramatsu, H.; Endo, M.; Kaneko, K. J. Am. Chem. Soc. 2010, 132,
12141215.
96. Kyakuno, H.; Matsuda, K.; Yahiro, H.; Inami, Y.; Fukuoka, T.; Miyata, Y.;
Yanagi, K.; Maniwa, Y.; Kataura, H.; Saito, T.; Yumura, M.; Iijima, S. J. Chem.
Phys. 2011, 134, 244501244514.
97. McBain, J. W.; Porter, J. L.; Sessions, R. F. J. Am. Chem. Soc. 1933, 55,
22942304.
98. Hanzawa, Y.; Kaneko, K. Langmuir 1997, 13, 58025804.
99. Bellissent-Funel, M. C.; Sridi-Dorbez, R.; Bosio, L. J. Chem. Phys. 1996, 104,
1002310029.
100. Futamura, R.; Iiyama, T.; Hamasaki, A.; Ozeki, S. Phys. Chem. Chem. Phys. 2012,
14, 981986.
101. Nakamura, M.; Ohba, T.; Branton, P.; Kanoh, H.; Kaneko, K. Carbon 2010, 48,
305312.
102. Ohba, T.; Kanoh, H.; Kaneko, K. J. Am. Chem. Soc. 2004, 126, 15601562.
103. Ohba, T.; Kanoh, H.; Kaneko, K. Nano Lett. 2005, 5, 227230.
104. Ohba, T.; Kaneko, K. Langmuir 2011, 27, 76097613.
105. Beauvais, C.; Boutin, A.; Fuchs, A. H. C. R. Chimie 2005, 8, 485490.
106. Goldfarb, D.; Li, H. X.; Davis, M. E. J. Am. Chem. Soc. 1992, 114, 36903697.
107. Pillai, R. S.; Jasra, R. V. Langmuir 2010, 26, 17551764.
108. Floquet, N.; Coulomb, J. P.; Dufau, N.; Andre, G.; Kahn, R. Physica B 2004, 350,
265269.
109. Striolo, A. Ads. Sci. Tech. 2011, 29, 211258.
110. Kaneko, K.; Setoyama, N.; Suzuki, T. In Studies in Surface Science and Catalysis;
Characterization Porous Solids III; Rouquerol, J., Rodrguez-Reinoso, F.,
Sing, K. S. W., Unger, K. K., Eds.; Elsevier, 1994; Vol. 87, pp 593602.
111. Setoyama, N.; Kaneko, K. Adsorption 1995, 1, 165173.
112. Nemirovsky, D.; Moreh, R.; Andersen, K. H.; Mayers, J. J. Phys. Cond. Matter.
1999, 11, 66536660.
113. Tanaka, H.; El-Merraoui, M.; Kodaira, T.; Kaneko, K. Chem. Phys. Lett. 2002,
351, 417423.
114. Beenakker, J. J. M.; Borman, V. D.; Krylov, Y. Chem. Phys. Lett. 1995, 232,
379382.
115. Sese, L. M. Mol. Phys. 1995, 85, 931947.
116. Stan, G.; Cole, M. W. J. Low Temp. Phys. 1998, 110, 539544.
117. Wang, Q.; Johnson, J. K.; Broughton, J. Q. J. Chem. Phys. 1997, 107,
51085117.
118. Wang, Q.; Challa, S. R.; Sholl, D. S.; Johson, K. Phys. Rev. Lett. 1999, 82,
956959.
119. Tanaka, H.; Kanoh, H.; Yudasaka, H.; Iijima, S.; Kaneko, K. J. Am. Chem. Soc.
2005, 127, 75117516.
120. Noguchi, D.; Tanaka, H.; Fujimori, T.; Kagita, H.; Hattori, Y.; Honda, H.; Urita, K.;
Utsumi, S.; Wang, Z. M.; Ohba, T.; Kanoh, H.; Hata, K.; Kaneko, K. J. Phys.
Condens. Matter 2010, 22, 334207334221.
121. Hattori, Y.; Tanaka, H.; Okino, F.; Touhara, H.; Nakahigashi, Y.; Utsumi, S.;
Kanoh, H.; Kaneko, K. J. Phys. Chem. B 2006, 110, 97649767.
122. Tanaka, H.; Noguchi, D.; Yuzawa, A.; Kodaira, T.; Kanoh, H.; Kaneko, K. J. Low
Temp. Phys. 2009, 157, 352373.
123. Noguchi, D.; Tanaka, H.; Kondo, A.; Kajiro, H.; Noguchi, H.; Ohba, T.; Kanoh, H.;
Kaneko, K. J. Am. Chem. Soc. 2008, 130, 63676372.
124. Chen, B.; Zhao, X.; Putkham, A.; Hong, K.; Lobkovsky, E. B.; Hurtado, E. J.;
Fletcher, A. J.; Thomas, K. M. J. Am. Chem. Soc. 2008, 130, 64116423.
125. Yamamoto, M.; Itoh, T.; Sakamoto, H.; Fujimori, T.; Urita, K.; Hattori, Y.; Ohba, T.;
Kagita, H.; Kanoh, H.; Niimura, S.; Hata, K.; Takeuchi, K.; Endo, M.; Rodriguez-
Reinoso, F.; Kaneko, K. Adsorption 2011, 17, 643651.
126. Kagita, H.; Ohba, T.; Fujimori, T.; Hata, K.; Tanaka, H.; Taira, S.; Kanoh, H.;
Minami, D.; Hattori, Y.; Itoh, T.; Masu, H.; Endo, M.; Kaneko, K. J. Phys. Chem. C
2012, 116, 2091820922.
127. Kumar, A. V. A.; Bhatia, S. K. Phys. Rev. Lett. 2005, 95, 245901245904.
128. Niimura, S.; Fujimori, T.; Minami, D.; Hattori, Y.; Abrams, L.; Corbin, D. R.;
Hata, K.; Kaneko, K. J. Am. Chem. Soc. 2012, 134, 1848318486.
129. Kowalczyk, P.; Gauden, P. A.; Terzyk, A. P.; Furmaniak, S.; Kaneko, K. J. Phys.
Chem. C 2011, 115, 1810518110.
130. Seki, K. Phys. Chem. Chem. Phys. 2002, 4, 19681971.
131. Kitaura, R.; Seki, K.; Akiyama, G.; Kitagawa, S. Angew. Chem. Int. Ed. 2003, 42,
428431.
132. Bourrelly, S.; Llewellyn, P. L.; Serre, C.; Millange, F.; Loiseau, T.; Ferey, G. J. Am.
Chem. Soc. 2005, 127, 1351913521.
133. Serre, C.; Mellot-Draznieks, C.; Surble, S.; Audebrand, N.; Filinchuk, Y.; Ferey, G.
Science 2007, 315, 18281831.
134. Coudert, F.-X.; Boutin, A.; Jeffroy, M.; Mellot-Draznieks, C.; Fuchs, A. H. Chem.
Phys. Chem. 2011, 12, 247258.
135. Kajiro, H.; Kondo, A.; Kaneko, K.; Kanoh, H. Int. J. Mol. Sci. 2010, 11, 38033845.
136. Onishi, S.; Ohmori, T.; Ohkubo, T.; Noguchi, H.; Di, L.; Hanzawa, Y.; Kanoh, H.;
Kaneko, K. Appl. Surf. Sci. 2002, 196, 8188.
137. Kanoh, H.; Kondo, A.; Noguchi, H.; Kajiro, H.; Tohdoh, A.; Hattori, Y.; Xu, W. C.;
Inoue, M.; Sugiura, T.; Morita, K.; Tanaka, H.; Ohba, T.; Kaneko, K. J. Colloid
Interface Sci. 2009, 334, 17.
138. van der Waals, J. H.; Platteeuw, J. C. Adv. Chem. Phys. 1959, 2, 157.
139. Eisenberg, D.; Kauzmann, W. The Structure and Properties of Water; Clarendon
Press: Oxford, 1969; Chapter 2.
140. Blake, A. J.; Hill, S. J.; Hubberstey, P.; Li, W. S. J. Chem. Soc. Dalton Trans.
1997, 913914.
141. Kondo, A.; Noguchi, H.; Ohnishi, S.; Kajiro, H.; Tohdoh, A.; Hattori, Y.; Xu, W. C.;
Tanaka, H.; Kanoh, H.; Kaneko, K. Nano Lett. 2006, 6, 25812584.
142. Cheng, Y.; Kondo, A.; Noguchi, H.; Kajiro, H.; Urita, K.; Ohba, T.; Kaneko, K.;
Kanoh, H. Langmuir 2009, 25, 45104513.
143. Kondo, A.; Kojima, N.; Kajiro, H.; Noguchi, H.; Hattori, Y.; Okino, F.; Maeda, K.;
Ohba, T.; Kaneko, K.; Kanoh, H. J. Phys. Chem. C 2012, 116, 41574162.
144. Kondo, A.; Chinen, A.; Kajiro, H.; Nakagawa, T.; Kato, K.; Takata, M.; Hattori, Y.;
Okino, F.; Ohba, T.; Kaneko, K.; Kanoh, H. Chem. Eur. J. 2009, 15, 75497553.
145. Kondo, A.; Noguchi, H.; Carlucci, L.; Proserpio, D. M.; Ciani, G.; Kajiro, H.;
Ohba, T.; Kanoh, H.; Kaneko, K. J. Am. Chem. Soc. 2007, 129, 1236212363.
146. Kondo, A.; Kajiro, H.; Noguchi, H.; Carlucci, L.; Proserpio, D. M.; Ciani, G.;
Kato, K.; Takata, M. T.; Seki, H.; Sakamoto, M.; Hattori, Y.; Okino, F.; Maeda, K.;
Ohba, T.; Kaneko, K.; Kanoh, H. J. Am. Chem. Soc. 2011, 133, 1051210522.
147. Serre, C.; Millange, F.; Thouvenot, C.; Nogues, M.; Marsolier, G.; Luer, D.;
Ferey, G. J. Am. Chem. Soc. 2002, 124, 1351913526.
148. Mellot-Draznieks, C.; Serre, C.; Surble, S.; Audebrand, N.; Ferey, G. J. Am. Chem.
Soc. 2002, 127, 1627316278.
149. Coudert, F. X.; Jefforoy, M.; Fuchs, A. H.; Boutin, A.; Mellot-Draznieks, C. J. Am.
Chem. Soc. 2008, 130, 1429414302.
150. Coudert, F. X.; Mellot-Draznieks, C.; Fuchs, A. H.; Boutin, A. J. Am. Chem. Soc.
2009, 131, 1132911331.
151. Watanabe, S.; Sugiyama, H.; Adachi, H.; Tanaka, H.; Miyahara, M. J. Chem. Phys.
2009, 130, 164707164711.
152. Sugiyama, H.; Watanabe, S.; Tanaka, H.; Miyahara, M. Langmuir 2012, 28,
50935100.
 7.05 Imaging of Heterogeneous Catalysts
AK Datye, Center for Microengineered Materials, Albuquerque, NM, USA
2013 Elsevier Ltd. All rights reserved.

7.05.1 Introduction 103

7.05.1.1 Role of Imaging in Catalyst Characterization 103
7.05.1.2 From 2D to 3D Imaging Recent Developments 103
7.05.2 Transmission Electron Microscopy 104
7.05.2.1 Conventional Transmission Electron Microscopy 104
7.05.2.2 Weak-Beam Dark-Field Methods 105
7.05.2.3 Inelastic Scattering (EELS, EDS, and SAM) 105
7.05.2.4 High-Resolution TEM 108
7.05.2.5 Scanning Transmission Electron Microscopy 109
7.05.2.6 Electron Tomography 111
7.05.2.7 3D Imaging via Aberration-Corrected EM 113
7.05.3 Scanning Electron Microscopy 116
7.05.3.1 Operating Modes: SE and BSE Imaging 116
7.05.4 x-Ray-Based Methods 116
7.05.4.1 x-Ray Tomography 116
7.05.4.2 Small-Angle x-Ray Scattering and Grazing-Incidence SAXS 117
7.05.4.3 XRD and Grazing-Incidence XRD 118
7.05.4.4 Scanning Transmission x-Ray Microscopy (STXM) 118
7.05.5 Magnetic Resonance Imaging 122
7.05.6 Correlating 3D Images with Catalytic Performance 123
7.05.7 Conclusion 127
Acknowledgments 128
References 128

7.05.1 Introduction
7.05.1.1 Role of Imaging in Catalyst Characterization
As the old adage goes, seeing is believing. Consequently, imag-
ing plays a very central role in the characterization of
catalysts. As a first step, we start with a visual inspection of a
catalyst, whether it is dark, or light, colored, or white. When
necessary, we quantify our visual observations with UVVis
spectroscopy. To go deeper into the catalyst structure, we need
higher magnifications to see the internal structure of a hetero-
geneous catalyst. For this, we naturally turn to a microscope,
which can use visible light, electrons, x-rays, neutrons, or mag-
netic fields. The results from microscopic investigations are pre-
sented to us as an image, an inherently two-dimensional (2D)
representation. A microscope derives signals from a three-
dimensional (3D) object and renders them to us in the form
of a 2D image. In this process, we inevitably lose information on
the third dimension. The exception to this general statement is
scanning probe microscopy where image contrast arises from
the height of the object under study, a technique that is inher-
ently 3D. However, scanning probe techniques come with their
limitations, wherein resolution in the third dimension comes at
the price of decreased resolution in lateral dimensions due to
probe convolution effects, especially for 3D objects. The limita-
tions of conventional 2D imaging have led to the development
of numerous approaches to derive 3D information using these
same techniques. This chapter is devoted to understanding how
3D information can be derived from some of the most com-
monly applicable methods of catalyst characterization.
7.05.1.2 From 2D to 3D Imaging Recent Developments
Three-dimensional imaging is important because all real sam-
ples are 3D, yet most microscope images are 3D representa-
tions of a sample and it is not always possible to reconstruct a
3D model from these images. Each of the microscopy tech-
niques described in this chapter, except scanning probe micros-
copy, provides 2D images from 3D objects. There are various
approaches to obtain 3D images using these same microscopy
techniques. At the simplest level, we obtain 3D information as
we do in our visual world, by stereo microscopy. This is possi-
ble with both optical and scanning electron microscopy (SEM).
We can further derive 3D information by tilting and rotating a
sample and observing it along different directions. This is akin
to the multiple views we use to represent 3D objects, such as a
house, through plan views and projections. To obtain all the
information we need, we would need to tilt the object through
90 in two different directions. And unless the object is trans-
parent, we will miss information from the side away from the
direction of observation. Obtaining 3D images of the inside of
an object from a series of 2D x-ray images taken around a single
axis of rotation is the basis of the widely applicable methods
known as computed axial tomography. While the conven-
tional approach requires tilting through 180 along a single

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00715-4 103

104 Imaging of Heterogeneous Catalysts
axis of rotation, specimen stages in electron microscopes do
not permit the full degree of tilt needed. The development of
new stages and software algorithms for image analysis and
reconstruction has made this technique more widely applica-
ble to perform electron tomography of catalysts.1,2 More
recently, with the advent of aberration correctors in electron
microscopes, it has become possible to produce a sharply
focused electron probe (diameter 0.05 nm) having a high
convergence angle, which makes the depth of focus very
small (a few nanometers). Therefore, only a thin section of
the sample is in focus in a single image, and this provides a
novel method for depth sectioning and tomography3 in the
electron microscope.
While tilting an object can indeed help provide 3D informa-
tion from a 2D imaging technique, there are other approaches
that do not involve tilting. If the objects of interest are small,
such as nanoparticles, and they are randomly oriented, one
could derive 3D information from a single image since all
possible orientations of the object would be visible. This
method was proposed in 1984 as a means of studying the
nature of metal support interfaces and the 3D shape of nano-
particles by Datye and Long.4 Alternatively, if the image contrast
is sensitive to specimen thickness, we can obtain 3D informa-
tion from the object, as practiced in the weak-beam dark-field
(WBDF) technique.5 Finally, we can literally peel off layers of a
specimen using a focused ion beam, and the succession of
images and elemental maps can be used to reconstruct the 3D
information from the specimen of interest.6 Recently, informa-
tion from catalysts has also been obtained by scanning a
focused x-ray beam to obtain spatially resolved information.7
7.05.2 Transmission Electron Microscopy
7.05.2.1 Conventional Transmission Electron Microscopy
Transmission electron microscopy (TEM) is a powerful tool for
the characterization of materials. Due to the strong interaction
of electrons with matter, the thickness of the specimen to be
Electron beam
Object plane
Q
Objective lens
Focal plane
Contrast
aperture
Image plane
Figure 1 Schematic diagram of image formation in a TEM. The objective lens functions by combining the transmitted electron beams from the
specimen into a 2D diffraction pattern or an image, as indicated with results for an ensemble of ZnO nanocrystals. An aperture, outlined in the diffraction
plane, is used to generate amplitude contrast. Reprinted from Datye, A. K.; Hansen, P. L.; Helveg, S. Handbook of Heterogeneous Catalysis, 2nd ed.;
Wiley: Weinheim, 2008; Vol. 2; pp. 803833, with permission.
studied must be below a few hundreds of nanometers. The
technique is therefore particularly useful for the characteriza-
tion of thin films, nanoparticles, and also for the characteriza-
tion of the microstructure near a surface. Conventional
transmission electron microscopy (CTEM) refers to the tech-
nique where image contrast arises from the variations in the
amplitude of the transmitted electron wave, in contrast to
high-resolution imaging, where phase contrast dominates. In
the CTEM mode, contrast is generated by intercepting dif-
fracted beams through an objective aperture. Figure 1 shows
a schematic of a modern electron microscope and how the first
step is the formation of a diffraction pattern in the back focal
plane. A magnified image is formed from the diffraction
pattern through the use of post-specimen, such as the interme-
diate and projector lenses. The sample must be thin enough for
electrons to be transmitted. The transmitted electrons strike a
phosphor generating light, which can be seen by the observer.
The recording of images was traditionally done on film, but
more recently via CCD cameras close coupled to the phosphor.
The dynamic range of CCD cameras, their high sensitivity, and
the ability for digital signal acquisition and image processing
have supplanted the use of film for image recording.
Does image contrast give you 3D information? The answer is
yes, but only under carefully controlled conditions. If we go by
what appears dark or light in a CTEM image, this is determined
by whether the sample is oriented close to the diffraction condi-
tion. The small wavelengths of the electron beam (0.00251 nm
at 200 kV) compared to x-rays (Cu Ka 0.0154 nm) mean that
the diffraction occurs at very small angles. Another major differ-
ence compared with x-ray diffraction (XRD) is the dynamical
nature of electron diffraction, whereby a significant fraction of
electrons get diffracted even from very thin crystals. This consti-
tutes the primary mechanism for thickness contrast, since after
traveling through a distance referred to as the extinction distance
(x) the incident beam is now traveling along the direction of
the diffracted beam. With increasing distance traveled through
the sample, the beam intensity oscillates between the incident
and diffracted beam, a phenomenon that gives rise to thickness
(103)
(102)
(101)
(110)
(002)
(100)
Braggs law:
2d(hkl) sinQ(hkl) = nl
ZnO partikler
0.2 mm
 Imaging of Heterogeneous Catalysts 105

fringes, also termed pendellosung fringes since the intensity
oscillates like a pendulum. By counting the fringes, one can
determine the thickness of the specimen. To obtain a quantita-
tive relationship between contrast and sample thickness, it is
necessary to control the sample orientation and the deviation
from the Bragg condition, through a technique known as WBDF
microscopy, which has been used to infer the 3D shapes of
catalytic nanoparticles as described in the next section.
7.05.2.2 Weak-Beam Dark-Field Methods
When a sample is correctly oriented for WBDF method micros-
copy, every dark band corresponds to an increase in sample
thickness by one extinction distance. The method was first pro-
posed by Yacaman and Ocanaz as a means to derive 3D shape and
topographic information on nanoparticles.5 Later, the group of
Konrad Kayek at Innsbruck also used this approach for the study
of what they termed inverted catalysts, where the metal was
deposited on rock salt, then the amorphous support film was
deposited on top of the metal, and finally the underlying rock
salt was dissolved away.9,10 Figure 2 shows a WBDF image of a
model Rh on silica sample. This image shows how the 2D pro-
jected image can be completely erroneous for interpreting 3D
shapes. The particles with a square profile are not cubes; rather,
they are half octahedra. This work shows how the 3D shapes of
nanoparticles can be derived; however, the information was
derived from a model catalyst where the particles are all oriented
similarly since they were prepared by deposition on rock salt
crystals that impose a [100] orientation on the particles. It is also
possible to apply the same technique to conventional supported
catalysts, but the particles must be large enough so that they can be
tilted in the microscope.
7.05.2.3 Inelastic Scattering (EELS, EDS, and SAM)
While much of the contrast in a TEM relies on elastic scattering,
we get important information through the inelastic events,
namely x-ray fluorescence (through the technique of energy-
dispersive spectroscopy or EDS) and electron energy-loss
spectroscopy (EELS).12 High-energy electrons cause ionization
of the atoms as they travel through the specimen, and the excited
atom can relax by emission of x-rays or Auger electrons (AE). The
AE can also be detected in a specially modified UHV microscope.
Since both the secondary electrons (SE) and AE have limited
escape depths, the techniques are surface-sensitive and provide
useful 3D information about the catalyst. The scanning Auger
microscopy (SAM) technique was developed by John Cowley
and coworkers.13 They used an extensively modified VG
HB501S scanning electron microscope where the objective lens
region of the microscope column was modified to obtain high
collection efficiency in the SE and AE imaging. The samples were
biased to 500 V to further enhance the SE and AE collection.
The electron beam source was a high brightness cold field
emission gun with 100 keV beam energy. The microscope had
a completely UHV environment. All the interior parts of the
microscope, including gun, column, and the attached sample
preparation chamber, were evacuated to a vacuum of the order of
1011 mbar. The samples were transferred between the specimen
preparation chamber and the microscope column in UHV. This
totally eliminated the contamination from hydrocarbons and
ensures the true in situ observation of the surface morphology
and composition change.14
While the technique has lot of potential, it did not gain
wide popularity due to the difficulties of sample preparation,
low signal-to-noise ratio, and cost, with the result that no
commercial instrument is available today to perform such
studies. On the other hand, the EDS and EELS techniques
are implemented in almost every available microscope and
are routinely applied for the characterization of catalysts.
These techniques provide the ability to detect and visualize
coreshell morphologies in bimetallic clusters as applied to
supported AuPd catalysts (Figure 3) in the work of Graham
Hutchings, Chris Kiely, and coworkers.15 These materials

Bright field Dark field

oh 25 nm
RhAl2O3

oh
(a) (b)
pp

oh
hp
RhTiO2

oh
rp

hp
(c) (d)

Figure 2 Bright-field electron micrographs of oxide-supported Rh nanoparticles: (a) Rh on alumina (mean particle size, 7.8 nm) and (c) Rh on titania
(mean particle size, 13.3 nm). The corresponding (200) WBDF images are shown in (b) and (d). A few polyhedral nanocrystals are indicated:
half-octahedron (oh), rectangular pyramid (rp), pseudopentagonal pyramid (pp), and hexagonal pyramid (hp). Reprinted from Rupprechter, G.; Seeber,
G.; Goller, H.; Hayek, K. J. Catal. 1999, 186(1), 201213, with permission.
106 Imaging of Heterogeneous Catalysts
Au
AuPd /Al2O3
uncalcined
Au
AuPd /Al2O3
calcined
200 C
Au
AuPd /Al2O3
calcined
400 C
Au
AuPd /Al2O3
reduced
500 C
(a)
Figure 3 Montage of (a) Au La XEDS map, (b) Pd La XEDS map, and (c) RGB overlays [green Au; blue Pd] for a series of AuPd/Al2O3 samples
subjected to different thermal treatments. Row 1 dried at 120  C; row 2 calcined at 200  C; row 3 calcined at 400  C; row 4 calcined at
400  C then reduced at 500  C in H2. Reprinted from Herzing, A. A.; Watanabe, M.; Edwards, J. K.; Conte, M.; Tang, Z. R.; Hutchings, G. J.; Kiely, C. J.
Faraday Discuss. 2008, 138, 337351, with permission.
exhibit considerable potential for the direct production of
H2O2 from molecular H2 and O2 under mild reaction condi-
tions. Pd surface segregation is not predicted by the thermody-
namics of the AuPd system. Instead, the equilibrium phase
diagram predicts that an FCC solid solution of Au and Pd is
formed over the entire compositional range with special order-
ing compounds existing at the Au3Pd and AuPd3 composi-
tions. Furthermore, the surface energies of Au and Pd are
1.50 and 2.05 J m2, respectively, suggesting that if anything
were to migrate to the nanoparticle surface it should be the Au
component. In order to investigate the development of the
AuPd nanoparticle morphology more closely, they used scan-
ning transmission electron microscopy (STEM)XEDS spec-
trum image analysis on a systematic series of AuPd/Al2O3
samples that were subjected to different heat treatments:
namely (i) dried at 120  C, but left uncalcined, (ii) calcined
at 200  C, (iii) calcined at 400  C, and (iv) calcined at 400  C
then subsequently reduced at 500  C in H2. The typical AuPd
nanoparticle morphologies generated by these samples are
shown in the matrix of Au Ma, Pd La, and the RGB color
overlay maps presented in Figure 3. It is clear that the uncal-
cined sample (Figure 3, row 1) shows homogeneous AuPd
nanoalloy particles, whereas the samples calcined at 200  C
(Figure 3, row 2) and 400  C (Figure 3, row 3) show the
progressive development of a Pd-rich shell and Au-rich core
morphology. This characteristic coreshell morphology is pre-
sumably brought about by the preferential formation of PddO
bonds at the alloy surface since in this 200400  C tempera-
ture range palladium oxidizes more readily than gold. Interest-
ingly, it is the 400  C sample which is subsequently reduced in
H2 at 500  C, that the Pd-rich shell/Au-rich core morphology
becomes unfavorable, and the nanoalloy particles either revert
Pd RGB
20 nm
Pd RGB
30 nm
Pd RGB
20 nm
Pd RGB
20 nm
(b) (c)
to homogeneous mixtures or form an inverted coreshell with
gold enrichment on the surface and Pd in the interior. An
example of the latter inverted coreshell morphology is
shown in Figure 3 (row d). The application of the STEM
XEDS technique in a Cs-corrected AEM can clearly provide a
new level of insight into coreshell morphology development
in nanoalloy particles.
Another technique of interest for the study of catalysts is
EELS, which in addition to providing elemental analysis pro-
vides information similar to local electronic structure as in
extended x-ray absorption fine structure (EXAFS) and x-ray
absorption near-edge structure (XANES). EELS records the
energy lost by an incident electron as it excites a core electron
to unoccupied local states, leaving behind a core hole
(Figure 8). Analyzing the resulting loss spectrum of the trans-
mitted beam provides information on the unoccupied local
density of states, partitioned by site, chemical species, and
angular momentum.16 The site specificity is achieved by form-
ing a probe small enough to only excite electrons on a single
atomic column. The chemical specificity is obtained from the
uniqueness of the core-level binding energy. The angular
momentum of the final state is determined by the dipole
selection rules and the angular momentum of the initial
state. When the core hole is well screened, the shape of the
EELS spectrum reflects the ground state density of states.
In systems with strong corehole coupling, excitonic effects
not only can dramatically alter the shape of spectrum but
also provide useful local fingerprints of the formal charge and
coordination chemistry.
The other major application of EELS is spectroscopy, to
show the chemical nature of the specimen under study. In a
recent paper, Jin et al.17 demonstrated that this technique
 Imaging of Heterogeneous Catalysts 107

Carbide

2.1

Silica rich
region

40 nm 1 -1 50 nm 1 -1
(a) (b)

Figure 4 TEM images of the silica-containing Fe catalyst sample before and after carburization. (a) Precursor oxide which is amorphous as evident
from the diffuse diffraction pattern. (b) Sample after CO activation and FT synthesis for 24 h. The spotty diffraction ring in the inset corresponds to
a d spacing of 2.1 A. Reprinted from Jin, Y.; Xu, H. F.; Datye, A. K. Microsc. Microanal. 2006, 12, 124134, with permission.

could allow us to infer the nature of carbonaceous deposits on

the surface of Fe FT catalysts. Figure 4 shows a low magnifica- 1000 p
tion TEM image of the iron oxide precursor catalyst and the 285.8 eV s
295.7 eV
sample after carbiding in a syngas mixture sample. The parti- 800
cles with darker contrast are identified to be iron carbide18
based on their diffraction patterns. The electron diffraction a
600
pattern in Figure 4(b) shows a spotty ring corresponding to a
lattice spacing of 2.1 A, which is typical for iron carbide. b
400
Because the structure of iron carbide has a low symmetry,
many reflections from lattice planes with d values close to c
2.1 A are allowed. The w-carbide phase Fe5C2; ICDD card 200
d
file 20-0508 has prominent diffraction peaks at 2.06 and 1 eV
2.03 A. These carbide particles also possess the characteristic -0
amorphous surface layers we have observed previously in other
260 280 300 320 340 360
iron FT catalysts.18,19 From Figure 4(b), we can also clearly
Energy Loss (eV)
see the presence of another amorphous phase, a silica-rich
region having a morphology similar to that of the precursor Figure 5 Carbon K edge fine structures of iron carbide particles imaged
iron oxide in Figure 4(a), coexisting with the carbide phase. x- in Figure 4(b). (a) C K edge from a carbide particle with an amorphous
Ray energy-dispersive spectroscopy EDS measurements show oxide layer; (b) C K edge from a carbide particle with amorphous carbon
layer; (c) C K edge from the same carbide particle with the e-beam
that these amorphous regions in the catalyst sample have a
focused on the amorphous surface carbon layer; (d) C K edge of the
silica/iron ratio much higher than the average value for the amorphous carbon support film. Reprinted from Jin, Y.; Xu, H. F.; Datye,
entire sample; hence, the iron that was initially uniformly A. K. Microsc. Microanal. 2006, 12, 124134, with permission.
distributed in the precursor has segregated into the crystalline
iron carbide phase.
Figure 5 shows a set of carbon K edge spectra of different used in the Phillips microscope could not be focused to select
carbon species on the carbide particle surface imaged in only the surface layer, we obtained spectra from the surface
Figure 4(b). As seen in Figure 5, the relative intensity of p* layer by positioning the beam very close to the surface, so that
to s* line changes significantly among the various patterns the actual spatial resolution of the probe did not matter. The C
shown in this figure. Curves a and b are C K edges recorded K edge fine structure in pattern c is very close to that of amor-
from the bulk of two iron carbide particles, one with a surface phous carbon film shown in pattern d, which has a more
oxide layer and the other with an amorphous carbonaceous intense s* peak but less intense p* peak. A small edge shift,
surface layer. The pattern in curve a therefore represents the 1 eV, of the p* line is also observed for carbidic carbon
spectrum of carbon atoms in iron carbide, characterized by a compared with amorphous carbon support film shown in
strong p* line, whereas pattern b is a mixture of the K edge of pattern d. The EELS spectra show that the surface amorphous
carbidic carbon and amorphous surface carbon. Pattern c is the carbon is different from the amorphous carbon that is used as
pattern recorded with the beam focused near the surface layer a support film for TEM samples. The surface carbon is very
of the same carbide particle. Although the electron beam we reactive and constitutes the active carbon on FT catalysts.
108 Imaging of Heterogeneous Catalysts

7.05.2.4 High-Resolution TEM

(a)
The technique of high-resolution TEM (HRTEM) dates back to
the early 1970s when the first image showing directly inter-
pretable contrast at the 3 A level using axial illumination was
published by Sumio Iijima.20 The development of the high
voltage electron microscope at Cambridge University led to
further advancement of the technique and pushed the resolu-
tion to the 2 A level where you can start to resolve the lattice
structure of most transition metals. The Cambridge microscope
provided the first HRTEM images of nanoparticles of interest to
catalysis.21 HRTEM images do not directly provide 3D infor-
mation, but since thickness affects image contrast, we can infer
the sample thickness by comparison with calculated images of
a known structure. In 1983, Marks and Smith22 proposed the
profile imaging technique whereby one could directly image
the surface structure rather than rely on diffraction contrast as
in the reflection electron microscopy technique. These early
studies showed clean, sharp surface images and morphological
details of catalytic significance, such as the distribution of
surface steps, surface facets, and surface reconstructions. 2 nm
Detailed computer simulations confirmed that the electron
micrographs could be interpreted in terms of atomic columns.
Figure 6 shows images from this early study that established (b)
the validity of images of the near-surface region. Figure 7
shows an image of the surface reconstruction of the Au to
form a (2  1) surface. The atom positions were extracted by
comparisons with calculated images of postulated structures.
While nanoparticles are often perfect single crystals, under
certain growth conditions, one may find a large fraction of
particles to have internal defects. Such is the case of the icosa-
hedral nanoparticles reported by Wang et al.23 that were syn-
thesized by a gas-phase method. The bright-field TEM image of
the nanoparticles (Figure 8(a)) on carbon support film shows
uniform, well-dispersed particles with sizes of about 56 nm.
In the bright-field TEM image, we find lower contrast around
the edges than in the center, but as stated earlier, contrast does
not directly correlate with thickness in bright-field TEM. The
inset of part b of Figure 8 depicts a typical image intensity line
profile from a high-angle annular dark-field (HAADF) image of
one representative nanoparticle, showing lower intensity Figure 6 (a) High-resolution electron micrograph of a small slightly
around the edges of the nanoparticle. This indicates that the rounded, square-pyramidal particle of gold after tilting from the h100i
atomic columns near the particles surface either contain fewer epitaxial orientation to the h110i pole. Visible lattice spacings are 0.235
atoms or are of lower atomic number than in the center. It can and 0.204 nm. Reverse contrast image-atomic columns appear white.
also be seen that the lattice fringes in Figure 8(a) do not go (b) Enlargement from (a), with image simulation superimposed.
across the entire particle, but rather are only seen in patches. Note visibility of surface steps and faceting. Reprinted from Marks L. D.;
The absence of lattice fringes in certain parts of the image can Smith, D. J. Nature 1983, 303(5915), 316317, with permission.
be caused by atomic columns not being lined up along the
viewing direction or due to the limited resolution of the TEM. indicated by white arrows in part (a) of Figure 9. The intensi-
The authors of this work used the technique of reconstruction ties of edge columns (black arrows in part (a) of Figure 9) are
of electron exit waves from a through-focal series of images to much lower or even absent, proving that whole columns can
extend the resolution to the sub-Angstrom regime. In this be removed from these sites. The lattice fringes can be observed
manner, one can correct for the aberrations of the objective throughout the entire nanoparticle with sub-Angstrom resolu-
lens. Figure 9 shows the exit wave phase images of a represen- tion. Because of the special fringe structure in Figures 9(a) and
tative FePt nanoparticle that was reconstructed from a through- 10(a), we can infer that the structure corresponds to a multiply
focus series, each consisting of 20 images. In the reconstructed twinned particle (MTP) instead of a single crystal. The images
phase images, the intensity maxima mark atom columns that of part (a) of Figure 9 show well-resolved (111) and (220)
can be analyzed in terms of column positions and heights planes with spacing of 0.220 and 0.135 nm of the fcc FePt
(sample thickness). The low intensities indicate that the corre- structure (space group Fm3m, (a) 0.3816 nm). The 0.135 nm
sponding atomic columns are only marked partially filled, as (220) lattice spacing is below the 0.168 nm point resolution of

2 nm
Figure 7 Small particle of gold showing directly, at the atomic level, a
partially reconstructed (2  1) surface superstructure. Inset shows a
calculated image, corresponding to the so-called missing-row model,
which matches well with the experimental image in the region marked.
Reprinted from Marks L. D.; Smith, D. J. Nature 1983, 303(5915),
316317, with permission.
the instrument. The upper and lower parts of the phase are
structural projections of tetrahedra viewed along the h110i
directions, whereas the left and right parts correspond to
views along h111i directions. The complex fringes in the center
were determined as belonging to a projection along h112i.
A quantitative evaluation of strain relaxation processes is
possible by extraction of column position maps from the exper-
iments. Here, one must also consider the possibility that there
may be buckling of atoms in the atomic columns that might
cause an anisotropic smearing of the intensity profile of the
respective column. Within the resolution, such anisotropic
intensity was not observed, and hence, the authors could recover
the centers of such columns from the experiment by fitting
model functions to intensity maxima. The spacing between
shells n and n 1 in Figure 9(a) is plotted in Figure 11(a),
where n 0 marks the difference between the outermost and
the first shell, which is 0.231  0.004 nm. It indicates that a
smaller occupancy of the outermost layer results in a larger
spacing. It is most unexpected to see in Figure 11(a) that the
spacing between the shells decreases exponentially with shell
numbers. Figure 11(b) shows an idealized model of these find-
ings showing a Pt-enriched surface layer, which provides a nat-
ural explanation for the experimentally observed oxidation
resistance of the vapor-generated MTPs in comparison with
colloidally prepared FePt particles.23
Another example of HRTEM to derive 3D information
(Figure 12) is from the work of Gontard et al.24 These authors
used spherical aberration-corrected TEM to provide atomic-
resolution information about the local topologies of active sites
on commercial nanoparticles with greatly improved sensitivity.
They used two advanced TEM techniques, which are based on
recent developments in hardware (aberration correction) and
computation (exit wave function restoration), to examine a
powder of Pt nanoparticles on carbon black, which had been
heated to 900  C in a N2-rich atmosphere. Despite the presence
of a graphitic carbon support, terraces and steps around the edge
Imaging of Heterogeneous Catalysts 109
(a) (b)
260
240
220
200
180
160
140
120
100
80
60
0 2 4 6
nm
20 nm
(c)
2 nm
Figure 8 Microstructure and distribution of the FePt nanoparticles.
(a) Standard bright-field TEM of the nanoparticles on carbon support
films shows homogenously distributed particles with identical sizes of
about 56 nm. Different diffraction contrasts and shapes are due to the
complex structure nature of the nanoparticles. (b) Z-contrast STEM
image and image intensity line profile (inset) across one particle of FePt
nanoparticles, showing lower intensity around the edges of the
nanoparticle. (c) Standard HRTEM images of the nanoparticle showing
distinct lattice fringes and straight edges. Reprinted from Wang, R. M.;
Dmitrieva, O.; Farle, M.; Dumpich, G.; Acet, M.; Mejia-Rosales, S.;
Perez-Tijerina, E.; Yacaman, M. J.; Kisielowski, C. J. Phys. Chem. C 2009,
113(11), 43954400, with permission.
of the particle are visible directly with a sensitivity, visibility, and
spatial resolution that are improved greatly over those in a
conventional HRTEM image. The monatomic steps that are vis-
ible in Figure 2.4.7(a) are conventionally termed A or B type,
with either (100) or (111) microfacet risers that have different
catalytic properties. These results provide compelling evidence
that the outermost atomic layers consist of irregular islands of
atoms. Correlating catalytic behavior with images such as these
can help improve our understanding of the nature of working
catalysts (Figure 12).
7.05.2.5 Scanning Transmission Electron Microscopy
The basic concept of a STEM is relatively simple a high-energy
electron beam is focused down to a small spot and scanned
over a thin sample. Signals from scattered electrons and ion-
ized atoms are recorded as the beam is scanned across
the sample to build up a 2D map. Chemical and bonding
information along each projected atomic column can be
obtained by measuring the energy lost by transmitted electrons
110 Imaging of Heterogeneous Catalysts

(a) )
(111
(220)
2 nm
(b)
0.09 nm
Figure 9 Exit wave phase image (a), fast Fourier transform (b) of one
FePt nanoparticle from the sample imaged at low magnification in
Figure 8. In part (a), the white arrows mark partly occupied shells and the
dark arrows mark edge columns that are commonly missing. Part (b)
shows the circle that corresponds to information transfer at an
information limit of 0.090 nm. Reprinted from Wang, R. M.; Dmitrieva,
O.; Farle, M.; Dumpich, G.; Acet, M.; Mejia-Rosales, S.; Perez-Tijerina, E.;
Yacaman, M. J.; Kisielowski, C. J. Phys. Chem. C 2009, 113(11),
43954400, with permission.
to core and valence excitations at each point on the map.
The performance of the instrument is determined by how
small a spot the electron beam can be focused to and how
much current can be maintained in that beam. Recent
advances in the control of lens aberrations through the suc-
cessful development of multipole-based aberration correc-
tors2527 have enabled sub-Angstrom beam sizes and rapid,
atomic-resolution compositional imaging.28
Given that interatomic distances are all larger than 0.5 A,
one might well suspect, as did Dennis Gabor in 1948, that
resolution will have to stop here for a lack of objects.29
Although this lack of suitable test objects does make it difficult
to test the resolution of a microscope, there is no shortage of
problems that could be tackled with smaller beam sizes. Atoms
in amorphous materials or tilted crystals can appear much
closer when viewed in projection. Being able to resolve
arbitrarily short distances at multiple projections could lead
to direct structural solutions for general amorphous materials
or atomic-resolution reconstructions of 3D objects by electron
tomography. The number of resolvable projections increases
roughly as the square of the resolution, so even improvements
in resolution by a factor of two or three could make a dramatic
difference. Even when such small distances do not need to be
measured, a smaller probe size allows for higher contrast images
and improved detectability of single dopant atoms, enabling
studies of a wider range of dopant species and thicker, more

[1,17,14] [17,1,14]
III
[712] I [172]
[112] II
III
[11,13,14] [11,13,14]
[112]

(a) (b)

(c) 1 nm (d)

Figure 10 Icosahedral structures of the FePt nanoparticle in part (a) of Figure 9. (a) Diagram of the MT model of tetrahedrons from h112i orientation
to h110i and h111i orientations, (b) proposed the shell growth icosahedron model with 10 179 randomly distributed Fe and atoms. Red and blue circles
represent Fe and Pt atoms, respectively. (c) Simulated phase image from part (b). (d) Simulated phase image from the MT model relaxed by energy
minimization. Both simulations prove that the particle is an icosahedral MTP. Reprinted from Wang, R. M.; Dmitrieva, O.; Farle, M.; Dumpich, G.; Acet,
M.; Mejia-Rosales, S.; Perez-Tijerina, E.; Yacaman, M. J.; Kisielowski, C. J. Phys. Chem. C 2009, 113(11), 43954400, with permission.
 Imaging of Heterogeneous Catalysts 111

Experimental FePt data

Simulated CuPt data (a)
0.235
Unoccupied edge sites
100% Pt Fully occupied surface shell
0.230 50% occupied surface shell
90% Pt
d111(nm)

0.225 Pt

55% Pt Fe Pt
0.220 52 48

45% Pt
0.215 Fe Pt
80 20
0.210 Cu
-Fe
0.205
0 1 2 3 4 5 6 7 8 2 nm
111
(a) Shell (n) Step 010
Pt(S)- A Step
5(111)(100) 4 11-1
Island
B Step

3 Terrace

B 111/111
Edge
1
111/100
A Edge
(b)
Pt(S)-4(100)(111)

Step vacancy
Pt A step

(b) Fe
Figure 11 (a) Comparison of experimentally measured (FePt) and
simulated (CuPt) shell spacing. Element segregation can explain the
exponential lattice relaxation to create a Pt shell and a Fe (Cu)-rich core.
The simulations show that randomly distributed surface vacancies
contribute to a partial strain relaxation. Missing-edge atoms in the topmost
layer (n 0) are required (full pink dots) to match the experimental result.
(b) Proposed icosahedron 10 179 atom sites and Pt atoms that segregated
to the surface (pink). Edge atom columns are removed making some of the
(green) iron atoms of the second shell visible. Reprinted from Wang, R. M.;
Dmitrieva, O.; Farle, M.; Dumpich, G.; Acet, M.; Mejia-Rosales, S.; Perez-
Tijerina, E.; Yacaman, M. J.; Kisielowski, C. J. Phys. Chem. C 2009, 113(11),
43954400, with permission.
realistic substrates. Figure 13 shows a STEMHAADF image that
demonstrates the sub-Angstrom resolution that can now be
attained by this technique. This mode of imaging can be con-
sidered incoherent, almost completely eliminating diffraction
and phase contrast. The contrast is then, to a good approxima-
tion, monotonic with thickness and is also sensitive to changes
in composition; for a typical geometry and material, it is approx-
imately proportional to Z1.8, where Z is the atomic number.
These properties make the STEMHAADF signal ideally suited
for tomographic applications (as shown in the next section).
7.05.2.6 Electron Tomography
The rapid development of electron tomography, in particular the
introduction of novel tomographic imaging modes, has led to
3
B step
2
kink
[110]
1
(c)
Figure 12 (a) Restored phase of a 6-nm Pt particle obtained by
applying spherical aberration correction and through-focus exit wave
function restoration to a defocus series of 20 images acquired at 200 kV
with the coefficient of spherical aberration, CS, adjusted to 30 mm.
(b) Best-fitting simulated phase. (c) The 3D atomic model used to
calculate the best-fitting phase in part (b). The large white arrow indicates
the direction of the electron beam. The inset overlapping parts (a, b)
shows the crystallographic details of the particle. In parts
(b, c), 14 correspond to the same features on the surface of the particle.
The notation Pt(S)n(x,y,z)  (u,v,w) refers to the microfacets on which
n is the number of atoms in the terrace, (x,y,z) is the Miller index of
the terrace, and (u,v,w) is the Miller index of the step. Reprinted from
Gontard, L. C.; Chang, L. Y.; Hetherington, C. J. D.; Kirkland, A. I.; Ozkaya,
D.; Dunin-Borkowski, R. E. Angew. Chem. Int. Ed. 2007, 46(20),
36833685, with permission.
the visualization and analysis of 3D structural and chemical
information from materials at the nanometer level. Two recent
reviews provide an excellent overview of the possibilities in the
emerging field of electron tomography.30,31 As described by
Midgley and Dunin-Borkowski,30 the concept of tomography
112 Imaging of Heterogeneous Catalysts
originates in a 1917 paper on the projection of an object into a
lower-dimensional space. Nearly 50 years later, tomographic
x-ray scanning for 3D medical imaging was proposed. Since
then, the use and type of tomographic scanners for medical
imaging have proliferated. The application to electron tomogra-
phy requires that images of a single object are recorded about a
tilt axis. For high-resolution (1 nm3) 3D tomographic imaging
in TEM, images are recorded every one or two degrees about a tilt
axis, over as large a specimen tilt range as possible (Figure 13(a)).
When using CTEM samples and specimen holders, a tilt may be
reached beyond which the sample is too thick or shadowing
occurs owing to the holder, grid, or other parts of the specimen.
This tilt maximum leads to a missing wedge of information, as
shown in Figure 13(b). Such missing information can lead to
artifacts, and reconstructions can be elongated in the direction of
the missing wedge. The tilt range should therefore be as high as
possible. Recent work32 has suggested that 7580 should be
sufficient to reduce artifacts to a minimum and not lead to serious
errors when measuring object sizes or shapes. Ultrahigh-tilt
holders are now commercially available, including those in
which the use of a needle-shaped sample allows 360 rotation
within the pole piece gap of the objective lens, eliminating
missing-wedge artifacts.
Typically, the ensemble of images is then back-projected to
form a 3D reconstruction. The information required, and the
type of specimen examined, often dictates which of the many
imaging modes available in the TEM is used. The most com-
mon approaches involve either bright-field TEM or HAADF
imaging. In biological work and in noncrystalline inorganic
systems, the use of bright-field TEM is possible because mass-
thickness contrast satisfies the projection requirement that the
recorded signal should be a monotonic function of some phys-
ical property. Diffraction contrast and Fresnel fringes arise in
crystalline specimens which do not satisfy the projection
requirement and can lead to serious artifacts in reconstruc-
tions. The image signal seen in the scanning TEM (STEM),
using HAADF imaging, offers a more suitable alternative.
The resolution of a reconstructed tomogram is governed by
the number of images in the tilt series and by the tilt range over
which the series is recorded. The Crowther resolution is equal to
the angular increment of the tilt series multiplied by the size of
the object. Thus, as well as needing a fine angular increment, the
resolution scales with the size of the object studied. In practice, to
limit beam damage and to keep acquisition times to a sensible
(a)
Figure 13 (a) Illustration of two-stage tomography process with (left) acquisition of an ensemble of images (projections) about a single tilt axis and
(right) the back-projection of these images into 3D object space. (b) Representation in Fourier space of the ensemble of projections, indicating the
undersampling of high-spatial-frequency information and the missing wedge of information brought about by a restricted tilt range. y is the tilt
increment between successive images and a is the maximum tilt angle. Reprinted from Arslan, I.; Tong, J. R.; Midgley, P.A. Ultramicroscopy 2006,
106(1112), 9941000, with permission.
level, the image is recorded every 12 . With accurate, often
iterative, reconstruction techniques, the resolution is then
approximately 1/100 of the object diameter. Although
real-space back-projection methods are used routinely for tomo-
graphic reconstruction, it is useful to consider that each projec-
tion corresponds to a central slice in Fourier space. By recording a
tilt series about a single axis, low-frequency information is sam-
pled more finely than higher-frequency information. Thus, a
simple back-projection, in which all of the information from
the ensemble of images is smeared into 3D reconstruction space,
leads to a blurred version of the original object. Computer pro-
cessing of the images, via proper weighting in Fourier space,
helps minimize the problem.
Recent work on imaging nanoparticles within a porous silica
support is shown in Figure 14.30 The 3D reconstruction reveals
the distribution of particles on and within the porous framework
and the fractal nature of the internal surface. Figure 14 also
shows that STEM tomography can be used to relate the distribu-
tion of particles to the underlying surface curvature, showing in
this case the strong preference of the particles (red) for saddle-
shaped anchor points (blue). STEM tomography has also been
used to determine the real-space crystallography of mesoporous
structures. For example, in MCM-48 mesoporous silica, which
has a double-gyroid form, electron diffraction and 2D high-
resolution electron microscopy studies had concluded that an
additional pore system was present in the system. STEM tomog-
raphy was able to visualize this directly in three dimensions
and confirm the space group symmetry.34 Electron tomogra-
phy and image analysis were also combined recently to
characterize ordered mesoporous silica SBA-15.35 The mor-
phology of the mesopores with average diameter 6 nm is
analyzed in terms of cylinders having variable radii and cen-
ters that are statistically centered on the points of a distorted
hexagonal lattice. The variations in the mesopore centers and
radii add up and result in pore wall corrugation with ampli-
tude of 1.6 nm. The correlation length of the corrugation
along the pore axis was found to be 45 nm. The amplitude
of the corrugation compared well with the 1.9-nm thick
microporous corona obtained from XRD. It is clear that the
tomographic approach provides a detailed microscopic 3D
model of nanostructured materials that complement macro-
scopic measurements, such as physisorption and XRD.
Inelastic signals, such as those detected using EELS and
energy-dispersive x-ray spectroscopy, can be used to map
Missing wedge q
a
Data points Missing wedge
(b)

20 nm
Figure 14 Surface-rendered representation of a tomographic reconstruction of a heterogeneous catalyst based on disordered mesoporous silica
supporting bimetallic rutheniumplatinum nanoparticles. The surface has been color coded according to the Gaussian curvature of the surface, with blue
regions delineating saddle points. The nanoparticles (red) appear to prefer to anchor themselves at the (blue) saddle points. Reprinted from Midgley, P. A.;
Dunin-Borkowski, R. E. Nat. Mater. 2009, 8(4), 271280, with permission.
composition in two dimensions and, by extension, in three
dimensions. Energy-loss information can be recorded pixel by
pixel as a sequence of spectra (spectrum imaging) or by
choosing a particular energy loss (or a small width about that
loss, typically 510 eV) and forming an image using electrons
that have lost only those energies. The latter approach, known
as energy-filtered TEM (EFTEM), can be extended to record
images over an energy-loss series, which by analogy with the
EELS method is called image spectroscopy or EFTEM spectrum
imaging. By choosing a particular energy loss, elemental maps
can be recorded over a tilt series. 3D compositional informa-
tion can be extracted from such EFTEM series. More recently,
image spectroscopy was extended to volume spectroscopy by
recording a large energy series at every tilt angle.36 A low-loss
series of a nanocomposite composed of a multiwalled carbon
nanotube encased in nylon was recorded every 3 eV over a
wide range of tilts. The different plasmon excitation energies
of the nylon (22 eV) and the nanotube (27 eV) enabled the
two components of the composite to be distinguished. By
reconstructing tomograms at individual energy losses, it was
possible to identify a voxel or subvolume common to all of the
energy-loss tomograms and, by plotting the intensity of the
voxel as a function of energy loss, to extract spectral informa-
tion from within the tomogram. In conventional EELS, spectral
information is always projected through the structure, but now
it is possible to extract spectral information from a subvolume
without any projection artifact.
7.05.2.7 3D Imaging via Aberration-Corrected EM
Although the ability to focus an electron beam down to a spot
size of around 2 A with a simple, round magnetic lens is
Imaging of Heterogeneous Catalysts 113
remarkable, 2 A is still far from the fundamental performance
limit that could be possible for electron imaging.16 The typical
electron wavelength in a TEM is in the order of 0.02 A, yet the
spatial resolution of such a microscope is 12 A, between 50 and
100 times worse than the wavelength. Wavelength for wave-
length, conventional electron lenses have similar optical proper-
ties to the bottom of a beer bottle with respect to spherical and
chromatic aberrations. This is not the result of imperfect engi-
neering, but rather a consequence of Laplaces equation, where
the electric and magnetic potentials used to focus the electron
beam cannot take on arbitrary shapes in free space. The result is
that positive spherical and chromatic aberrations are unavoid-
able for static round lenses. This fundamental limit was first
identified by Otto Scherzer in 1936, who also suggested methods
to correct these aberrations by considering nonstatic, nonround
lenses or placing free charges in the path of the beam. Of these,
the most practical approach has proved to be abandoning cylin-
drical symmetry for some of the lenses and instead adding a
series of multipole optical elements that correct the leading
terms in an ever increasing power series expansion of the wave-
front distortions.37 It took nearly 50 years before the precision,
stability, and software control needed to implement these cor-
rector designs could be realized.26 Since the proof of concept was
established, corrector development has been rapid (and is still
continuing), with probe sizes close to 0.5 A now reached. Aber-
ration correctors have now removed spherical aberration (and
higher-order geometric aberrations) as limiting factors. Instead,
chromatic aberrations, combined with the finite energy spread of
the electron source, constitute the current limits on spatial reso-
lution. The resolution and aperture size can be increased by
building a chromatic aberration corrector, increasing the beam
voltage or reducing energy spread of the source. Improvements
114 Imaging of Heterogeneous Catalysts

of a factor of two appear possible, but not without cost: increas-
ing the beam voltage also increases the risk of radiation damage
so beam voltage is more likely to be decreased in future instru-
ments, only exacerbating the chromatic problems. Reducing the
source energy spread by adding a monochromator reduces the
available beam current. Cryogenically cooled sources can narrow
the source energy spread and improve gun brightness, but have
proved difficult to implement owing to the required mechanical
stability. Hence, we are currently at a limit of about 0.5 A in
probe resolution in the latest aberration-corrected microscopes
(see Figure 15).
There is another unexpected benefit from aberration correc-
tion. Because the Rayleigh diffraction limit is given by: 0.61l/y, in
order to take advantage of the reduced aberrations, the objective
aperture y must be made larger. This increased aperture size
allows a new technique for 3D imaging,3 similar to optical con-
focal microscopy. For a diffraction-limited system, in the absence
of aberrations, the depth of field is commonly defined [5] as l/y2.
The limited depth of focus will become even more relevant with
improvements in corrector design, because it is already compara-
ble to specimen thickness. Thus, for many samples, the variations
in height are far larger than the depth of field in a Cs-corrected
system. So this is not just a way to extract new information; it will
become essential to consider the reduced depth of focus in order
to interpret images from any nonflat sample. Similarly, for the
case of single atoms buried in an amorphous substrate, it has
been shown3 that the single atoms are not necessarily visible
unless the beam is focused within the sample. Thus, even where
3D information is not explicitly required, it may be essential
to use this technique in order to extract the correct distribution
of dopants or impurities. This also has implications for tomogra-
phy. If the depth of focus is too small, then the images at different
tilts are not simple projections and the approximations used to
derive the tomographic reconstruction are no longer valid, in
which case tomography will necessarily involve some form of
deconvolution.
By recording images at different focus values in an aberration-
corrected microscope, it is now possible to slice through the
TEM sample in a manner comparable to that used in confocal
optical microscopy. The work of Van Benthem et al.38 dem-
onstrates a volume resolution better than 0.1 nm  0.1 nm 
6.6 nm that enables the location of individual Hf atoms
with a precision of about 0.5 nm in depth. This was
the first report of the localization of individual atoms in
three dimensions by a direct imaging technique. Tomography
techniques based on tilt series provide depth resolution but
do not provide the necessary lateral atomic resolution to
image single atoms. Atom probe field ion microscopy39 is the

Ga 63 pm

Figure 15 Direct demonstration of sub-Angstrom resolution ADFSTEM imaging: the sample is [211]-oriented wurtzite GaN, showing a 0.63 nm
spacing between the projected Ga columns (yellow). The lighter N atoms (blue circles) are not visible in the image. The elastic signal is displayed in false
color to emphasize the Ga columns. Reprinted from Muller, D. A. Nat. Mater. 2009, 8(4), 263270, with permission.
 Imaging of Heterogeneous Catalysts 115

(a)
Output: none selected only other technique that allows single atom detection from a
Mode: image capture
3D structure but is more difficult to apply to insulating, nano-
Gate dielectric
(HfO2/SiO2 ) structured materials and cannot survey a broad area of a gate
Polycrystalline silicon
dielectric to image a low density of leakage sites. Figure 16(a)
shows a sketch view of such a gate dielectric interface structure.
High-resolution ADFSTEM imaging revealed single Hf atoms
within an amorphous SiO2 interlayer between the HfO2 film and
the Si substrate in which no channeling of the electron beam
S D
occurs. Figure 16(b) shows nine images extracted from a 41-
Silicon frame through-focal series recorded in h110i zone-axis orienta-
tion with a focal increment of 0.5 nm. At various positions
throughout the focal series, single Hf atoms come into focus in
(b) 35 frames each, corresponding to intervals of 1.52.5 nm in
f-2.0 nm f-1.5 nm f-1.0 nm depth (see white circles in Figure 16(b)). No evidence of beam
damage effects was observed in micrographs recorded from the
same areas immediately after the through-focal series acquisi-
tion. This technique is very promising for pinpointing the loca-
tion of promoter atoms on catalyst supports.
An example of this application is the recent work of Qiao
f-0.5 nm f+/-0.0 nm f+0.5 nm
et al.40 who imaged Pt on FeOx (Figure 17) and demonstrated
that these atomically dispersed species have very high reactivity
for CO oxidation even at room temperature. By focusing the
electron beam on the FeOx, they were able to identify that
the Pt atoms were located on the surface. Figure 17(a) clearly
f + 1.0 nm f + 1.5 nm f + 2.0 nm shows individual Pt atoms (marked by the white circles) uni-
formly dispersed on the surfaces of FeOx nanocrystals. The
sub-Angstrom resolution HAADF image in Figure 17(b)
reveals clearly that individual Pt atoms (white circles) occupy
exactly the positions of the Fe atoms. As plan-view HAADF
images represent the projection of atoms along the incident
Figure 16 (a) Sketch of a high-k gate dielectric device with S and D beam direction, surface atoms cannot be distinguished from
representing source and drain, respectively. Annular dark-field images the subsurface atoms in the HAADF images. However, whether
were acquired as a through-focal series from the Si/SiO2/HfO2 interface,
the Pt atoms are inside the FeOx nanocrystallites can be deter-
indicated by the red box in the sketch view. (b) Nine consecutive
micrographs extracted from a series of 41 images recorded at the focal
mined by changing the focus of the electron beam. By analyz-
values shown. One single Hf atom coming into and going out of focus is ing the images obtained sequentially with varying beam focus
marked by white circles. The micrographs are displayed using a nonlinear settings, the authors concluded that the observed Pt atoms
contrast scale. Reprinted from Van Benthem, K.; Lupini, A. R.; Kim, M.; were not located inside the individual FeOx nanocrystals. In
Baik, H. S.; Doh, S.; Lee, J. H.; Oxley, M. P.; Findlay, S. D.; Allen, L. J.; addition, by examining many HAADF images of sample A,
Luck, J. T.; Pennycook, S. J. Appl. Phys. Lett. 2005, 87(3), 13, with the authors estimated that the density of Pt single atoms
permission. was about 0.07 Pt atoms nm2, which is very close to the actual

4 nm 2 nm
(a) (b)

Figure 17 HAADFSTEM images of samples Pt/FeOx showing that Pt single atoms (white circles) are uniformly dispersed on the FeOx support (a) and
occupy exactly the positions of the Fe atoms (b). Reprinted from Qiao, B.; Wang, A.; Yang, X.; Allard, L.F.; Jiang, Z.; Cui, Y.; Liu, J.; Li, J.; Zhang, T.
Nat. Chem. 2011, 3, 634641, with permission.
116 Imaging of Heterogeneous Catalysts
Pt loading (about 0.09 Pt atoms nm2). It has long been pro-
posed that single atoms of dopants41 may help modify
the catalytic behavior of oxides. Images such as these and
correlation with catalytic behavior may help us prove the role
of such species in catalysis.
7.05.3 Scanning Electron Microscopy
7.05.3.1 Operating Modes: SE and BSE Imaging
In the technique of SEM, the image is obtained by scanning a
finely focused probe in a raster pattern on the specimen surface
in synchrony with writing the detector signal into a matrix to
form a digital image. High detector signals yield result in high
brightness in the image. The magnification is determined by
the ratio of the area scanned to the display area, unlike in a
TEM where the excitation of the post-specimen lenses deter-
mines the magnification. The most commonly used imaging
mode makes use of SE that are emitted when the sample is
excited by the electron beam. These are low-energy electrons
(0100 eV) making their collection very easy by means of an
appropriately biased scintillator. Due to the low energy, the
electron has a small mean free path ensuring that the signal
comes from the surface region of the specimen. The number of
electrons emitted depends markedly on the angle of the spec-
imen surface with respect to the electron beam and the detec-
tor. On the one hand, as the electron probe approaches a
vertical surface (with the surface normal orthogonal to the
beam direction), a large fraction of the SE can escape and be
collected. Hence, a vertical surface appears bright in an SEM
image. On the other hand, since only a small fraction of the SEs
can escape and be collected from a horizontal surface, it would
appear less bright. These differences in brightness therefore
reveal the sample topography. Hence, SEM images are inher-
ently 3D since they provide an intuitive sense of the third
dimension, as seen from Figure 18 of mesoporous silica
spheres obtained by an aerosol synthesis route.42
500 nm
Figure 18 Secondary electron image of a mesoporous silica sample
prepared via aerosol synthesis. The SE image is inherently 3D confirming
that these silica particles are spherical. Reprinted from Bore, M. T.;
Pham, H. N.; Switzer, E. E.; Ward, T. L.; Fukuoka, A.; Datye, A. K. J. Phys.
Chem. B 2005, 109(7), 28732880, with permission.
Another imaging mode within an SEM makes use of back-
scattered electrons (BSEs). These electrons have energies com-
parable to those in the incident beam (130 keV) and hence
have larger mean free paths. Therefore, the signal is less sensi-
tive to surface topography; instead, image intensity depends on
the composition of the specimen. When properly used, the BSE
images can provide excellent atomic number contrast. The
higher energies of the BSEs make them more difficult to collect
compared to the low-energy SEs, which can be collected effi-
ciently by applying a bias voltage to a grid in front of the
detector. For BSEs, a detector with a large solid angle is gener-
ally necessary, which can lead to constraints on the available
space in the sample chamber. However, modern in-lens SEMs
allow the detector to be placed within the bore of the lens,
enhancing the efficiency for BSE imaging. BSE imaging pro-
vides added 3D information by changing the beam excitation
voltage. Higher kilovolt electrons penetrate deeper and provide
information from subsurface regions. Figure 19 shows a set of
SEM images obtained from Au deposited in mesoporous silica
SBA-1543 after the use in the CO oxidation reaction (four cycles
each from room temperature to 400  C). The pictures on the
left are SE images (Figure 19(a) and (c)) and hence they reveal
the sample topography. The lower penetration depth of 2 kV
electrons provides much better rendering of the surface fea-
tures than the images recorded at 5 kV. The 7-nm pores of the
silica can be seen clearly in the 2 kV image but less so in the
5 kV image. Only the largest Au particles, which are located on
the silica surface, can be imaged in Figure 19(a) and (c). On
the other hand, the BSE images show the Au particles very
clearly due to the higher yield of BSEs from the Au compared
to the silica. Changing the accelerating voltage from 2 to 5 kV
reveals more Au particles, several of them elongated to con-
form to the pore shape. The higher penetration of the 5 kV
electrons allows us to see deeper into the sample. This confirms
that a significant fraction of the Au is present within the pore
structure. The elongated shapes of some of the Au particles
imply that the pore walls help confine the Au particles, as
they continue to grow. Ostwald ripening must be the domi-
nant sintering mechanism since particles continue to grow in
size as the catalyst is used for CO oxidation.
7.05.4 x-Ray-Based Methods
7.05.4.1 x-Ray Tomography
x-Ray tomography is a routine tool in many laboratories,
with desktop instruments achieving a 3D resolution of a
few micrometers. By using a synchrotron source, x-ray tomo-
grams can be produced with sub-100 nm resolution, and a
2D spatial resolution of 15 nm is possible using zone
plates.44 An x-ray approach based on diffractive imaging,
which involves recording a tilt series of coherent diffraction
patterns and using phase-retrieval methods to reconstruct
real-space tomograms, allows a transverse resolution of
10 nm to be achieved.45,46 However, this technique is
slow and still not at a resolution that has immediate impact
for the study of catalysts at the atomic scale; however, infor-
mation at the level of catalyst pellets can be obtained with
this technique.
 Imaging of Heterogeneous Catalysts 117

(a) (b)

273 nm 273 nm

(c) (d)

273 nm 273 nm

Figure 19 High-resolution SEM images of Au supported in SBA-15 mesoporous silica after use in the CO oxidation reaction. (a) SE image at 2 kV,
(b) BSE image at 2 kV, (c) SE image at 5 kV, and (d) BSE image at 5 kV. In the BSE images, more Au particles are seen at 5 kV (d) than at 2 kV (b).
The variation in voltage (2 and 5 kV) and the type of signal (SE vs. BSE) allow us to determine the location of the Au particles (on the surface or within the
pores). Reprinted from Bore, M. T.; Pham, H. N.; Switzer, E. E.; Ward, T. L.; Fukuoka, A.; Datye, A. K. J. Phys. Chem. B 2005, 109(7), 28732880,
with permission.

7.05.4.2 Small-Angle x-Ray Scattering and Grazing-

Incidence SAXS
Small-angle scattering can provide information on particle size
in catalysts, and in favorable cases, where the particles are
nearly monodisperse, it is also possible to derive particle size
distributions.47 But the technique is dependent on having suffi-
cient contrast between the catalyst and support, since the scatter-
ing is sensitive to each interface, the active phase and support, and
the support and the pore volume. Grazing-incidence SAXS
(GISAXS) can provide the same type of information as regular
small-angle x-ray scattering (SAXS), but in addition, it can give
depth profile information and particleparticle distances. Also,
the aspect ratio (height/diameter) of a metal cluster can be calcu-
lated from the GISAXS data. It is ideal for in situ studies since it is
very sensitive to surface species and there is less parasitic scattering
resulting from the substrate compared to a direct transmission
scattering experiment. This method also samples a large area,
yielding good statistics, which is always a problem with any
microscopy. It has been noted that clusters as small as 40 atoms
can be observed for the CPt clusters, and the anisotropy of the
aggregates can be clearly resolved as well.48 The technique has
become popular for studying quantum dots and clusters on, and
imbedded into, surfaces. Recent work by Renaud et al.49 demon-
strated the high sensitivity of the GISAXS technique to monitor
in real time the growth of Pd clusters on MgO(100) and of Co
on Au(111) during metal vapor deposition, and this work
has been extended to a quantitative study of the growth of Pd
particles on MgO.
Figure 20 shows the experimental setup for a GISAXS exper-
iment as described in the recent work of Winans et al.50 They
investigate the reactivity and thermal stability of platinum
clusters supported on Al2O3/SiO2/Si(100) as a function of
the thickness of the alumina film and presence of hydrogen.
Figure 21 shows a typical slice through the GISAXS data at two
different temperatures. The analysis of these curves leads to an
estimate of the height and width of the Pt clusters, which is
shown in Figure 22. These x-ray studies show that alumina
films grow as islands and that a low number of ALD cycles
produce a support with Al2O3 islands while leaving part of the
underlying SiO2/Si(100) surface exposed. With two-cycle ALD
118 Imaging of Heterogeneous Catalysts
qz
af
ki kf
ai
2q
Substrate
Figure 20 Principle of GISAXS. An incoming x-ray beam illuminates the
sample surface under a grazing incident angle ai, and the scattered
intensity is recorded at small angles (2y, af) on a plane perpendicular to
the sample surface. The sample can be rotated around its surface normal.
Reprinted from Winans, R. E.; Vajda, S.; Lee, B.; Riley, S. J.; Seifert, S.;
Tikhonov, G. Y.; Tomczyk, N. A. J. Phys. Chem. B 2004, 108(47),
1810518107, with permission.
I (a.u.)
1000
30 C
400 C
100
0.0 0.2 0.4 0.6 0.8
-1
qz ( )
Figure 21 Background subtracted vertical cuts of AGISAXS data of
Pt710 clusters on Al2O3 (2-cycle)/SiO2/Si(100) in vacuum. Reprinted
from Winans, R. E.; Vajda, S.; Lee, B.; Riley, S. J.; Seifert, S.; Tikhonov,
G. Y.; Tomczyk, N. A. J. Phys. Chem. B 2004, 108(47), 1810518107,
with permission.
films, holes in the alumina film have been observed.23 The
platinum clusters are much more mobile on silica than on
alumina and are likewise less stable on the alumina coated
silica, as seen in Figure 22. However, increasing the number
of ALD cycles leads to a surface that leads to extremely high
thermal stability of Pt710 clusters. The study shows how
GISAXS can be useful for monitoring the evolution of sup-
ported metal clusters. The authors have recently extended the
work also to powder supports.
7.05.4.3 XRD and Grazing-Incidence XRD
The conventional applications of XRD for studies of catalysts
involve the use of line broadening to infer average crystallite
size. The determination of 3D crystallite shape is more difficult
due to the high symmetry of the cubic lattice structure of most
catalytically relevant metals, which results in a multiplicity of
crystal planes. However, when the nanoparticles are located on
a single crystal substrate and oriented due to epitaxial growth,
it is possible to determine their shapes. The recent work by
Nolte et al.51 shows how synchrotron radiation can help pro-
vide microscopic insight into shape changes in nanoparticles
qy
during oxidation and reduction. The catalyst consisted of rho-
dium nanoparticles on magnesium oxide (001) substrates.
qx Wide-angle diffraction reciprocal-space mapping and GISAXS
were performed simultaneously. Figure 23 shows how the
diffraction from Rh nanoparticles can be utilized to determine
the shape of the Rh, which is characterized by the parameters
NP, NB, NT, and NE that are obtained from a Wulff analysis of
high-resolution in situ XRD, combined with TEM. The Rh NPs
were next exposed to 3  105 mbar O2 at 600 K (i.e., above the
oxygen chemical potential for Rh2O3 bulk oxide formation),
while the XRD pattern was simultaneously recorded. The
authors observed a distinct change in the XRD signal (see
difference map in Figure 24(b)) that essentially consists of an
intensity enhancement along the (001) rod and an intensity
loss along the (111) rods, which can also be observed in the
large area difference map in Figure 24(b).51 The best fit to this
x-ray intensity change (Figure 2(b)) results in an average NP
shape as characterized by NP 31  1, NT 16  1, NB 5  1,
and NE 7  1 (Figure 24(c)), with an unchanged average NP
diameter of 8.3 nm and a reduced average height of 4 nm.
These reconfigured Rh NPs now have a nanosized oxide skin
composed of an ultrathin hexagonal surface oxide layer. This
phenomenon was driven by the formation of an oxygenrho-
diumoxygen surface oxide at the rhodium nanofacets.
The oxygen-induced shape change of the Rh NPs was fully
reversible when the surface oxide is removed by CO exposure
(at 1  105 mbar). This example shows how XRD, especially
grazing-incidence XRD (GIXRD), can be used to determine
catalytically relevant 3D shape changes. One should, however,
1.0
recognize that the specific method is only applicable to ori-
ented nanoparticles that are reasonably monodisperse. The fact
that the work can be carried out at high pressures makes
this technique of potential interest to study the behavior of
model catalysts.
7.05.4.4 Scanning Transmission x-Ray
Microscopy (STXM)
The absorption of x-rays in catalysts is already well recognized
in the techniques of x-ray absorption spectroscopies, specifi-
cally EXAFS, NEXAFS (near-edge x-ray absorption fine struc-
ture), and XANES. But these techniques lack spatial resolution.
The advent of focusing optics permits the investigation of
spatially resolved signals from an x-ray beam. Drake et al.52
at Berkeley first reported in 2004 the design of a specially
fabricated cell for in situ characterization of catalysts and envi-
ronmental materials using soft x-ray absorption spectroscopy
and spectromicroscopy at photon energies above 250 eV. Lab-
on-a-chip technologies were used to fabricate the cell on a
glass wafer. The sample compartment is 1.0 mm in diameter
and has a gas path length of 0.8 mm to minimize x-ray absorp-
tion in the gas phase. The performance of the cell was tested by
acquiring Cu L-3-edge XANES data during the reduction and
oxidation of a silica-supported Cu catalyst. Oxidation state-
specific images of the catalyst clearly show the disappearance
of Cu(II) species during the exposure of the oxidized sample to
 Imaging of Heterogeneous Catalysts 119

13 13
400 C Horizontal
12 12

11 11

10 10
Horizontal
Rg ()

9 Vertical 9

8 8

7 7

6 6
Vertical
5 5
50 100 150 200 250 300 350 400 0 10 20 30
(a) Temperature (C) (b) Time (min)
Figure 22 Changes of the average vertical and horizontal radii of gyration Rg of Pt710 clusters supported on 2 ALD cycles Al2O3 film on SiO2/Si(100)
in the absence of H2. (a) Temperature dependence and (b) kinetics monitored after reaching 400  C. Reprinted from Winans, R. E.; Vajda, S.;
Lee, B.; Riley, S. J.; Seifert, S.; Tikhonov, G. Y.; Tomczyk, N. A. J. Phys. Chem. B 2004, 108(47), 1810518107, with permission.

(022)
NT
[001]

NB

[110] (202)

(111)
NE
(020)

[001]
NP
[110]
[110] [110] (200)

[110]

(a) (b)

Figure 23 (a) Schematic representation of the particle shape for fcc-based NPs. The lateral particle size is characterized by NP, its height by NT NB,
and the missing-edge atoms by NE. For a Wulff-shaped particle (without substrate), NT NB and NE NP  NT. (b) (110) plane reciprocal-space
map of clean Rh NPs on MgO(001) with an average lateral size of 8 nm. The white box indicates the area for high-resolution scans used for
the quantitative analysis. Reprinted from Nolte, P.; Stierle, A.; Jin-Phillipp, N. Y.; Kasper, N.; Schulli, T. U.; Dosch, H. Science 2008, 321(5896),
16541658, with permission.

4% CO in He while increasing the temperature from 308 to
473 K. Reoxidation restores the intensity of the image associ-
ated with Cu(II). Cu L-3-edge XANES spectra obtained from
stacks of STXM images show that with increasing temperature,
the Cu(II) peak intensity decreases as the Cu(I) peak intensity
increases.
More recently, with improved resolution, de Smit et al.53
describe how a highly focused beam of x-rays can be used to
acquire chemical maps and images of working heterogeneous
catalysts at nanometer spatial resolution. In this work, a beam
of monochromatic, low-energy x-rays (soft x-rays) was focused
to a spot size of 1020 nm, using a zone plate. The sample
under investigation was then scanned by moving the sample
relative to the sample and the absorption of the beam by the
sample was measured. This was repeated for x-rays of different
energies, so that absorption could be plotted as a function of
x-ray energy. Because the whole sample was scanned, the rela-
tionship between absorption and beam energy could be
 Diffraction map Difference map Diffraction map
clean O2 dissociation oxidized
400 -0.2
0.8 0.8 0.8
1021 0.025

2608 0.25
0.7 0.7 0.7
6658 0.475

(001)
4

Experiment
0.6 1.7 10 0.6 0.7 0.6
-0.4 -0.2 0 0.2 0.4 -0.4 -0.2 0 0.2 0.4 -0.4 -0.2 0 0.2 0.4
(110)

0.8 0.8 0.8

Fit
0.7 0.7 0.7

0.6 0.6 0.6

-0.4 -0.2 0 0.2 0.4 -0.4 -0.2 0 0.2 0.4 -0.4 -0.2 0 0.2 0.4

001
O2 dissociation 001

11 11
1 1
111 111
T = 600 K

100
CO oxidation
100
Clean Rh particle Oxidized Rh particle (core)
(a) (b) (c)

Figure 24 (a) (Top) (110) Diffraction map of clean Rh particles at 600 K. (Middle) Fitted diffraction map corresponding to the average particle shape given below. (b) (Top) Oxygen-induced signal change in
the (110) plane. (Middle) Simulated signal change for particles with increased (100) side facet area. (c) (Top) Experimental (110) diffraction map at 600 K and 2  105 mbar O2 pressure. (Middle) Fitted diffraction
map for particles under oxygen exposure. (Bottom) Best-fit core particle shape after oxidation. Reprinted from Nolte, P.; Stierle, A.; Jin-Phillipp, N. Y.; Kasper, N.; Schulli, T. U.; Dosch, H. Science
2008, 321(5896), 16541658, with permission.
 Imaging of Heterogeneous Catalysts 121

mapped out across the sample. By producing maps at charac-
teristic energies, an image of the spatial distribution of that
element was obtained. Alternatively, a plot of absorption
against x-ray energy at a given spot was used to provide an
x-ray absorption near-edge spectrum, which allowed the
deduction of the local composition of the material under
study. The STXM technique is ideal for studying heterogeneous
catalysts but has some practical problems associated with it.
Soft x-rays are readily absorbed by matter. This means that, for
STXM to work, the catalyst particles must be very small and the
distance traveled by the x-rays in the reactor cell the x-ray gas
path length must be short (less than 100 mm). The first STXM
study of Drake et al.52 of a catalyst was restricted to experi-
ments using diluted gases at temperatures up to only 533 K.
The report of de Smit et al.53 constructed a much improved
STXM cell, which has a gas path length of only 50 mm and
which can operate at 1 bar and at temperatures of up to 773 K.
The authors used their reactor cell to examine changes in a
catalyst for the FischerTropsch synthesis (Figure 25). Observ-
ing how the composition of a catalyst alters with changes in
reaction conditions and reaction time can reveal insight into
the factors controlling catalyst activity and stability. This study
demonstrates the potential of STXM for in situ chemical imag-
ing of catalysts at the nanometer scale. The spatial resolution of
the technique (15 nm) is impressive, but is still not high
enough to give a truly atomic-scale view of a catalysts structure.
Improvements in x-ray optics and imaging methods, however,
should allow higher spatial resolution, opening the way to a
deeper understanding of the structure and composition of
multicomponent catalysts and the changes they undergo
during a reaction.54 However, the technique suffers from the
limitation that the image and spectrum are averaged along the

Iron L2 and L3 edges

Oxygen K edge

a-Fe2O3 100%

(e) 525 530 535 540 545 550 555

50 nm

(d) 705 710 715 720 725 (a)

0
Fe Fe2SiO4 Fe3O4
Normalized absorption

Normalized absorption
25 25 50 100% 1
1 1

33 50 17 100% 2
2 2

50 nm

(f) 705 710 715 720 725 (b) (g) 525 530 535 540 545 550 555

Fec /FexCy Fe2SiO4 Fe3O4

20 75 100%
1 1
1

40 50 100%
2 2 2

50 nm

705 710 715 720 725 (c) 525 530 535 540 545 550 555
(h) x-Ray photon energy (eV) (i) x-Ray photon energy (eV)
a-Fe2O3 Fe3O4 Fe2SiO4
SiO2 Fec/FexCy

Figure 25 Chemical contour maps and x-ray absorption spectra of the catalyst material during the different stages of reaction. Chemical contour
maps (ac) of a 400 nm  3750 nm region and corresponding iron L2- and L3-edge (d, f, h) and oxygen K edge (e, g, i) x-ray absorption spectra, all
normalized to a maximum absorption of one. The contour lines have been removed for clarity. (a) Before treatment, at room temperature (25  C) in
helium; (b) after 2 h in H2 at 350  C; (c) after 4 h in synthesis gas at 250  C. Specific sampling regions and the corresponding x-ray absorption spectra
are indicated in the figures. The dotted lines in (d), (f), and (h) indicate the spectra fitted by a linear combination of reference spectra. The bar
graphs represent the calculated relative percentage contributions of the different iron phases at the sampling points. The arrows in (i) indicate the
observed shoulder feature. Reprinted from de Smit, E.; Swart, I.; Creemer, J. F.; Hoveling, G. H.; Gilles, M. K.; Tyliszczak, T.; Kooyman, P. J.;
Zandbergen, H. W.; Morin, C.; Weckhuysen, B. M.; de Groot, F. M. F. Nature 2008, 456(7219), 222239, with permission.
122 Imaging of Heterogeneous Catalysts
direction of the x-ray beam; hence, surface-sensitive 3D infor-
mation cannot be readily obtained, except at the edge of the
sample, as in other characterization methods, such as TEM.
7.05.5 Magnetic Resonance Imaging
Magnetic resonance (MR), in the form of solid-state nuclear
magnetic resonance (NMR) spectroscopy, is well established as
a research tool for investigations of the structures of solid
catalysts and molecular species adsorbed on them. However,
during the past decade, there has been increasing interest in
using magnetic resonance imaging (MRI) techniques to study,
in particular, flow fields inside reactors.55 These studies have
recently been extended to measurements of chemical conver-
sion within model reactor systems. The real power of MR
techniques is that by bringing together spectroscopy, diffusion,
microimaging, and flow imaging, they provide a noninvasive,
chemically specific measurement technique, which can charac-
terize a system over length scales ranging from the angstrom to
the centimeter scale. In a recent review,56 developments in MRI
pulse sequences are summarized and applications to investiga-
tions of both hydrodynamics and catalytic conversion within
catalysts and catalytic reactors are presented. Here, we focus on
the application of MRI for 3D imaging within millimeter-sized
catalyst bodies.
Quantitative data about the spatial distribution of the cat-
alyst precursor species during impregnation would be of great
interest to understand the preparation of catalysts. In a recent
study,57 the distribution of the Ni species within alumina was
analyzed by MRI imaging. Specifically, the [Ni(H2O)6]2 com-
plex within an alumina pellet was obtained via MRI imaging.
Figure 26 illustrates the 1D [Ni(H2O)6]2 concentration pro-
file inside an extrudate as a function of time after pore volume
impregnation. The impregnation solution contained 0.5 M Ni
(NO3)2. The T2-weighted images of the extrudate after impreg-
nation with this solution are also depicted in this figure. The
red regions indicate high Ni2 concentrations and the blue
1.2
[Ni(H2O)6]2+ (M)
0.8
0.4
0
-1.58 0 1.58
d (mm)
Figure 26 1D profiles of the [Ni(H2O)6]2 concentration as a function of the position inside the extrudates at certain times after impregnation,
together with the recorded T2-weighted images. Reprinted from Espinosa-Alonso, L.; Lysova, A. A.; De Peinder, P.; De Jong, K. P.; Koptyug, I. V.;
Weckhuysen, B. M. J. Am. Chem. Soc. 2009, 131(18), 65246534, with permission.
regions correspond to low Ni2 concentrations. At the edges
of the extrudate, 5 min after impregnation, the [Ni(H2O)6]2
concentration was higher than 0.5 M, while in the center, there
was no nickel present. During the pore volume impregnation,
three main phenomena occur: capillary flow of the solution
toward the core of the extrudate, diffusion of the metal ion in
the pores filled with water, and its adsorption on the walls of
the pores. The fact that [Ni(H2O)6]2 ions do not move toward
the core of the catalyst body together with the capillary flow of
water molecules and that a concentration gradient is observed
5 min after impregnation suggests that there are interactions of
this complex with the pore walls of the alumina extrudate.
Moreover, these interactions cannot be electrostatic since the
slightly acidic pH of the [Ni(H2O)6]2 solution below the pzc
of g-Al2O3 creates a positively charged alumina surface. These
interactions are responsible for the [Ni(H2O)6]2 concentra-
tion gradient, and, therefore, the transport of [Ni(H2O)6]2
toward the core takes place according to liquid-phase diffusion
and adsorption/desorption phenomena on the pore walls. A
uniform distribution of 0.53 M [Ni(H2O)6]2 was achieved
60 min after impregnation, which agreed with the concentra-
tion of the impregnation solution. This calculated concentra-
tion was within the range of the expected uncertainty (410%).
Moreover, the time needed to get a uniform distribution of [Ni
(H2O)6]2 was in line with liquid-phase diffusion of Ni(NO3)2
salt in bulk solution for a distance of 1.93 mm (radius of
the g-Al2O3 extrudate), which indicates that the adsorption
of [Ni(H2O)6]2 on alumina surface is limited.
In contrast, when [Ni(edtaHx)](2x) is used as the precur-
sor, the behavior of the Ni complex was quite different.
Figure 27 shows the T1-weighted images collected on an extru-
date at several points in time after its impregnation with solution
Niedta 1:1 pH 7 (complex [Ni(edta)]2). The blue external ring
represents high NMR signal intensity values, and as explained
earlier, it corresponds to the region where [Ni(edta)]2 is
present. The green core indicates the presence of pure water
(low NMR signal intensity). One hour after impregnation, [Ni
(edta)]2 still showed an egg-shell distribution inside the
5 min 15 min
-1.9
5 min
1.9
15 min 30 min 60 min
30 min
60 min
90 min
120 min 90 min 120 min
INMR (a.u.)

5 min 15 min 30 min 45 min 60 min
INMR (a.u.)
6 More specifically, the use of edta in a low edta:Ni2 molar ratio
60 min
5 acidic pH was used, with the edta ligand acting as a linker
INMR (105 a.u.)
4 are not easy to achieve in the preparation of Ni catalyst bodies,
5 min
15 min
3 restrictions. This example shows how MRI imaging shows poten-
30 min
45 min
2 5 min
60 min
1 7.05.6 Correlating 3D Images with Catalytic
0
1.58 0 1.58
Position (mm)
Figure 27 T1-weighted images after impregnation of an extrudate with
solution Ni EDTA 1:1 pH 7, and the corresponding 1D signal intensity
profiles as a function of the position inside the catalyst body. Reprinted
from Espinosa-Alonso, L.; Lysova, A. A.; De Peinder, P.; De Jong, K. P.;
Koptyug, I. V.; Weckhuysen, B. M. J. Am. Chem. Soc. 2009, 131(18),
65246534, with permission.
extrudate (blue external ring). It took 90120 min (images not
shown) to get a uniform distribution of this complex. This time
was expected according to a liquid-phase diffusion transport of
[Ni(edta)]2 when limited interactions between the complex
and the supports surface take place. Namely, it is assumed that
[Ni(edta)]2 has a smaller diffusion coefficient than [Ni
(H2O)6]2 since it is a bulkier complex. Moreover, electrostatic
interactions between the complex and alumina surface are not
very strong when the solution pH is close to the pzc of the
support. The 1D signal intensity profiles as a function of
the position inside the extrudate are also presented. Close to
the edges, the NMR signal intensity remains constant with time
after impregnation, since at these working conditions, relaxation
of water protons in the presence of [Ni(edta)]2 is fast enough.
On the other hand, in the core, where no [Ni(edta)]2 is present
after 5 min, water protons do not have time to relax, and their
signal is suppressed. As [Ni(edta)]2 diffuses to the core, water
protons in that region relax faster and ultimately have enough
time to relax, recovering the maximum signal. Since these 1D
profiles show NMR signal intensities and not concentrations,
no conservation of the area under the curves of Figure 27 is
expected.
MRI was also able to visualize the transport of Ni2 when two
different species, such as [Ni(H2O)6]2 and [Ni(edtaHx)](2x),
were present inside the catalyst body. After 2 h, the former was
present along the cross section of the extrudate, while the latter
remained in the outer rim of the catalyst body. This information
enabled the use of MRI in a quantitative manner for determining
the [Ni(H2O)6]2 concentration distribution in the presence of
[Ni(edtaHx)](2x). Additionally, MRI provided indirect infor-
mation on the interactions of different Ni2 complexes with each
other and with the surface of the support. These interactions,
Imaging of Heterogeneous Catalysts 123
which already occur in the first stage of catalyst preparation, are
directly responsible for the final metal-concentration distribu-
tions in the extrudate and for its final metal phase(s) and disper-
sion. The combination UVVis microspectroscopy and MRI
imaging facilitated the understanding of the transport and
adsorption of the different species on the surface of the support.
suggested the formation of Ni2 egg-shell distributions, when
between Ni2 and the alumina surface. These profiles, which
are interesting profiles for chemical reactions with diffusional
tial for better understanding of catalyst preparation.
Performance
The primary purpose of catalyst characterization is to learn
about the near-surface region of the specimen, since that is
where the catalytic processes occur. Hence, the emphasis of
this chapter has been on methods that help provide insight
into catalysis. Methods that provide only bulk information
were described in less detail in this review. Consequently, we
emphasized some of the new developments in surface-sensitive
imaging since they provide alternative ways to get 3D infor-
mation about catalytic materials. The ultimate goal of charac-
terization is not just to provide pretty pictures but also to
contribute to our understanding of catalytic phenomena.
Many of the characterization techniques operate only in a
vacuum environment; hence, it is difficult to perform charac-
terization on an operating catalyst. Photons are easier to get in
and out of a reactor than electrons, so naturally we see more
emphasis on x-ray and neutron methods for in situ studies.
Traditionally, these spectroscopies lacked spatial resolution,
but more recent improvements allow the resolution to
approach subnanometer levels when using x-rays. We present
below a few examples using x-rays and electrons where a direct
correlation was made between the observed 3D information
and catalytic reactivity.
An important issue is the oxidation state of the catalyst
during reaction conditions. In a recent report from the group
of Jeroen van Bokhoven,58 we learn how the state of the
catalyst can be monitored on the subsecond time scale.
Insights into the structural changes that occurred during igni-
tion were obtained with quick extended x-ray absorption fine
structure (QEXAFS). Figure 28 shows the spectra that were
recorded with a time resolution of 0.5 s. For clarity, not all
recorded spectra are shown. As soon as the ignition starts, the
white line increases in intensity. After only 9 s, the changes in the
spectra were complete. In this time period, the conversion of
carbon monoxide increased from 53% to 89% (Figure 28(b)).
The isosbestic points in Figure 28(a) indicate that there is
a direct conversion of reduced platinum to partially oxidized
platinum. The in situ HERFD XAS and QEXAFS data suggest
that oxidized platinum plays an active role in generating
high activity. As soon as the surface is sufficiently depleted of
carbon monoxide, oxygen reacts with the platinum surface,
with a simultaneous increase of rate of oxidation of carbon
124 Imaging of Heterogeneous Catalysts

1.5

1.2
Normalized
absorption

0.9

11 580 11 600 11 620

(a)
Energy (eV)

100

90

80

70
CO conversion %
60

50

40
470 471 472 473 474 475
(b) Temperature (K)
Figure 28 Pt L3-edge XANES of 2 wt% Pt/Al2O3 recorded in QEXAFS mode (a) taken at start of ignition (blue) and 1.5 s (red), 4 s (green), 6.5 s
(orange), and 9 s (black) after the start of ignition during heating at a rate of 2 K min1; (b) percentage conversion of carbon monoxide during ignition,
color coding of the squares correlates to the spectra in (a). Reprinted from Singh, J.; Alayon, E. M. C.; Tromp, M.; Safonova, O. V.; Glatzel, P.;
Nachtegaal, M.; Frahm, R.; van Bokhoven, J. A. Angew. Chem. Int. Ed. 2008, 47(48), 92609264, with permission.

monoxide. This decreases the concentration of carbon
monoxide in the gas phase and further depletes the surface
of carbon monoxide and subsequently increases the extent
of surface oxidation, which additionally enhances the rate
of reaction. The result is the autocatalytic enhancement in
conversion. Figure 29 shows a diagram that gives an over-
view of the oxidation of carbon monoxide over supported
platinum nanoparticles. Below the ignition temperature, the
surface is covered with carbon monoxide, and the reaction
rate is low and determined by the desorption of carbon
monoxide. At these temperatures, the platinum particles are
reduced with adsorbed carbon monoxide, even in an oxygen-
rich environment. As the temperature increases, some of the
carbon monoxide is converted and increasingly desorbs from
the platinum surface. At the ignition temperature, oxidic plat-
inum forms and the rate suddenly increases. At high activity
and temperatures above the ignition temperature, the particles
are largely oxidic, which is proposed to be required to generate
the highly active state of the catalyst (Figure 29).
While the electron tomography involves tilting of the sam-
ple to obtain views from multiple directions, we demonstrated
an alternate approach that involves the use of spherical catalyst
supports in which we can image the particles edge-on, and due
to their random orientations, we can now image all possible
 Imaging of Heterogeneous Catalysts 125

Alumina
Partially oxidized
platinum,
high rate of reaction
O O

C C

Alumina
472 K Partially desorbed
carbon monoxide
455 K

439 K
O O O O
422 K
C C C C
390 K
340 K
Alumina
308 K
Platinum with adsorbed
carbon monoxide,
11 550 11 560 11 570 11 580 11 590 11 600 low rate of reaction
Energy (eV)

Figure 29 Schematic diagram showing the state of the Pt catalyst surface at the light off phase during CO oxidation over 2 wt% Pt/Al2O3.
Reprinted from Singh, J.; Alayon, E. M. C.; Tromp, M.; Safonova, O. V.; Glatzel, P.; Nachtegaal, M.; Frahm, R.; van Bokhoven, J. A. Angew. Chem. Int. Ed.
2008, 47(48), 92609264, with permission.

reduction states, showing clearly how rough surfaces can be

2 nm
(111) Rh
(100)
SiO2
support
Figure 30 Profile view images of Rh/SiO2. Particles are nearly
cubo-octahedral and exhibit (111) and (100) facets. Reprinted from
Kalakkad, D.; Anderson, S. L.; Logan, A. D.; Pena, J.; Braunschweig, E. J.;
Peden, C. H. F.; Datye, A. K. J. Phys. Chem. 1993, 97(7), 14371444, with
permission.
orientations with respect to the support in a single TEM
image.4 This eliminates the need for sample tilting and allows
us to study the 3D shapes of metal particles very readily. The
edge-on views also permit correlations to be made between
surface structure and reactivity. Figure 30 shows a HRTEM
image of Rh/SiO2 after reduction in H2 at 500  C.59 The Rh
particles exhibit well-defined low index facets. Figure 31 shows
an image of Rh particles after oxidation and low-temperature
created on metal particles. These images show that the nature
of surface sites in metal particles is not a function of size alone.
In similarly sized particles, one can have very different surface
coordination leading to marked differences in activity. Further,
it is commonly thought that oxidation will lead to spreading
and wetting of the metal oxide on the support. However, we
find that noble metals, such as Pd or Rh, form oxide particles
that remain 3D in shape. A slight increase in size is caused by
swelling of the particle due to a greater mass and the lower
density of the oxide. It is the difference in specific volume of
the metal and metal oxide phases that gives rise to this restruc-
turing, leading to enhanced activity and altered selectivity.
The nature of Cu catalysts for methanol synthesis has been
of much interest. The work at the Topsoe laboratories60 has
reported the use of in situ TEM to study changes in the catalytic
surfaces after exposure to the gas phase. Time-resolved mea-
surements of the methanol synthesis reaction over a Cu/ZnO-
based catalyst reveal a transient methanol production that
depends on the pretreatment gas. Specifically, the methanol
production initially peaks after a pretreatment with an inter-
mediate mixture of H2 and CO (2080% H2 in CO). The
activity measurements were compared to environmental trans-
mission electron microscopy (ETEM) observations of Cu nano-
particles supported on ZnO during exposure to comparable
reaction conditions. The ETEM images reveal a gas-dependent
morphology of the Cu nanoparticles with a marked flattening
for a gas mixture of H2:CO 1:1. Figure 32 shows shape
changes in the Cu/ZnO catalyst that is used for methanol
126 Imaging of Heterogeneous Catalysts

synthesis as a function of gas atmosphere. The gas-dependent
morphology of the Cu nanoparticles provides a consistent
explanation of the observed coupling between the transient
methanol production and pretreatment conditions. These
examples demonstrate that the 3D imaging methods can pro-
vide important information for helping interpret catalytic
behavior in complex, multicomponent catalysts.
In recent years, there is increasing interest in the conversion of
biomass-derived reactants into fuels and chemicals. Catalysts for
the production of biorenewable chemicals must operate under
aqueous-phase conditions, generally at temperatures in excess of
473 K. Conventional oxide supports designed for gas-phase reac-
tions are not suitable for aqueous-phase reactions at these ele-
vated temperatures. For example, alumina undergoes a phase
2 nm
Rh
SiO2
support
Figure 31 HRTEM image of Rh/SiO2 after oxidationreduction cycling.
The catalysts show higher activity and altered selectivity due to the
rougher exposed surfaces compared to the annealed state, shown in
Figure 8.1.3. Reprinted from Kalakkad, D.; Anderson, S. L.; Logan, A. D.;
Pena, J.; Braunschweig, E. J.; Peden, C. H. F.; Datye, A. K. J. Phys. Chem.
1993, 97(7), 14371444, with permission.
change from g-Al2O3 to boehmite at 473 K with a consequent
loss of surface area. Likewise, mesoporous silica SBA-15 suffers
from collapse of the well-ordered mesoporous structure when
heated to 473 K in liquid water, resulting in loss of its surface area
and structural integrity. In a recent study, Pham et al.62 demon-
strated a simple and inexpensive approach for modifying the
surfaces of oxide supports so as to make them hydrothermally
stable in liquid water at 473 K. The approach involves the depo-
sition of a thin film of carbon, generally around 10 wt% carbon,
derived from simple sugars, such as sucrose. To confirm the
presence of carbon on the pore walls of SBA-15, the authors
acquired elemental carbon maps via EFTEM. Elemental carbon
maps of the SBA-15-based samples are shown in Figure 33,
where bright red regions indicate the location of carbon in the
EFTEM image at the carbon K edge. The uncoated SBA-15 particle
is dark in the map (b), indicating that carbon is not present. For
reference, the underlying carbon support film is visible in the
EFTEM map in Figure 1(b). The HRTEM image of carbon-coated
SBA-15 (c) is similar to that of uncoated SBA-15 (a) since carbon
does not produce any additional contrast in this image. How-
ever, the EFTEM image of coated SBA-15 (d) shows bright red
regions showing the location of the carbon within the pores, and
the dark regions within SBA-15 correspond to the silica walls.
These images confirm that the carbon is able to coat the internal
surfaces of the mesoporous silica SBA-15. From the EFTEM
image of coated SBA-15, it might appear that carbon has filled
the pores of SBA-15. However, N2 sorption isotherms obtained
from the coated samples are very similar to the uncoated SBA-15,
indicating a narrow pore size distribution. This example demon-
strates how the elemental analysis signals in a TEM can provide
information on 3D structures in heterogeneous catalysts.
The final example we present relates to the study of catalyst
deactivation. It is known that catalysts lose activity when they

Cu(200), d = 0.18 nm 2 nm Cu(111), 2 nm 2 nm

d = 0.21 nm

Cu(111), Cu(111),
d = 0.21 nm d = 0.21 nm
Cu(111),
d = 0.21 nm

ZnO(011), ZnO(012), ZnO(011),

d = 0.25 nm d = 0.19 nm d = 0.25 nm
(a) (c) (e)

(110) (110)
(100) (100)
(110)
(111) (111)
(100) (111)
(b) (d) (f)

Figure 32 In situ TEM images (a, c, and e) of a Cu/ZnO catalyst in various gas environments together with the corresponding Wulff constructions
of the Cu nanocrystals (b, d, and f). (a) The image was recorded at a pressure of 1.5 mbar of H2 at 220  C. The electron beam is parallel to the [011] zone
axis of copper. (c) Obtained in a gas mixture of H2 and H2O, H2:H2O 3:1 at a total pressure of 1.5 mbar at 220  C. (e) Obtained in a gas mixture of
H2 (95%) and CO (5%) at a total pressure of 5 mbar at 220  C. Reprinted from Hansen, P. L.; Wagner, J. B.; Helveg, S.; Rostrup-Nielsen, J. R.;
Clausen, B. S.; Topsoe, H. Science 2002, 295(5562), 20532055, with permission.

(a)
Carbon
support
SBA-15
Zero loss
100 nm image
(c)
CarbonSBA-15
Zero loss
50 nm image
Figure 33 HRTEM images of calcined (a) SBA-15 and (c) 10 wt% carbonSBA-15; elemental carbon maps of (b) SBA-15 and (d) 10 wt% carbonSBA-
15. The image shows how the carbon forms a conformal coating over the silica, helping to improve its hydrothermal stability. Reprinted from
Pham, H. N.; Anderson, A. E.; Johnson, R. L.; Schmidt-Rohr, K.; Datye, A. K. Angew. Chem. Int. Ed. 2012, 124, 1334013344, http://dx.doi.org/10.1002/
anie.201206675, with permission.
are used at elevated temperatures. The process of catalyst
sintering, growth of metal particles, is well documented but
the mechanisms of particle growth are still being debated.
The driving force for catalyst sintering is the increased surface
free energy of nanoparticles compared with the bulk metal
surfaces. Sintering is particularly acute in catalysts such as
automotive exhaust catalysts and catalysts used for steam
reforming of methane that operate at elevated temperatures.
Ostwald ripening, one of the mechanisms of sintering, is also
responsible for the dissolution of Pt nanoparticles and their
redistribution that occurs in fuel cell electrocatalysts. While
several models have been proposed to describe the kinetics of
catalyst sintering, they are often more useful to correlate
observed sintering rates than to describe the basic energetics
of the gasmetalsupport interface. Furthermore, most obser-
vations of industrial catalysts are performed before or after use
(ex situ), which make it difficult to infer the mechanisms
responsible for catalyst sintering. In this recent study,63 the
authors performed observations of catalyst sintering via in
situ TEM, which provides direct information about the mech-
anism of sintering. The observed rates of disappearance of
individual nanoparticles allowed them to extract energetic
parameters that help describe the evolution of the entire
assembly of nanoparticles. This method is generally applicable
to a broader range of catalyst systems and reaction environ-
ments and may thus provide a more fundamental approach for
understanding and predicting the long-term behavior of cata-
lytic materials. Figure 34 shows the disappearance of a 2.5 nm
Ni particle when heated at 750  C in a 1:1 H2:H2O gas atmo-
sphere at 3.6 mbar total pressure.
Imaging of Heterogeneous Catalysts 127
(b)
Carbon
support
No carbon
seen in SBA-15
Carbon K
energy loss
image
(d)
Carbon visible
on pore walls of
SBA-15
Carbon K
energy loss
image
7.05.7 Conclusion
The biggest advance in recent years has been the improvement
in spatial resolution of electron microscopes. In terms of beam
current and collection speeds in the electron microscope, there
is over a 100-fold difference between aberration-corrected and
conventional instruments. There is also the potential for per-
haps another 20-fold increase in beam current density from
investments in corrected optics and new source designs. The
same advances in optics and sources could probably improve
spatial resolution by another factor of two, or so (down to a
probe size of 0.25 A), at which point mechanical and elec-
tronic stabilities will dominate. Radiation damage is likely to
limit the signal and resolution for most materials and is already
driving future microscope development toward lower beam
voltages. Finally, there have been significant improvements in
detectors for collection of x-rays for elemental analysis, with
improved solid angle for collection of x-rays and the ability to
process a high flux of x-rays. Likewise, we have come a long
way toward low dose cameras with more efficient detection of
electrons that improve our ability to acquire images and to
perform energy-loss spectroscopy. The impact of all this new
instrumentation on our study of materials is that it is now to
resolve the location of all atoms in nanoparticles or amor-
phous materials and the imaging and chemical identification
of dopants, defects, and impurities with single atom resolu-
tion. The ability to record compositional and bonding maps in
three dimensions enables studies of microscopic inhomogene-
ities that may underlie the spatial variations in nanoscale
materials. The chemical environment of specific surface and
128 Imaging of Heterogeneous Catalysts
0.0 s 2.0 s
6.0 s 7.6 s
Figure 34 Time-lapsed TEM images of MgAl2O4-supported Ni nanoparticles at 750  C in 1:1 H2:H2O at a total pressure of 3.6 mbar. The scale bar
in the images is 5 nm and the times are indicated relative to the start of observation (which is within a few minutes of reaching the operating
temperature). The time dependence of the ripening of an individual nanoparticle was used to extract relevant energetic parameters, which could be used
to predict the evolution of the collection of nanoparticles that constitutes this catalyst. Reprinted from Challa, S. R.; Delariva, A. T.; Hansen, T. W.;
Helveg, S.; Sehested, J.; Hansen, P. L.; Garzon, F.; Datye, A. K. J. Am. Chem. Soc. 2011, 133(51), 2067220675, with permission.
bulk sites in catalyst particles can now be explored one atom
column at a time. The faster data acquisition opens new oppor-
tunities for 3D tomographic chemical imaging and real-time
STEM-based environmental microscopy in gas or liquid cells.
For heterogeneous catalysts, this makes electron microscopy
a universal and essential tool, and it is clearly being adopted as a
workhorse in industrial laboratories. It is no longer just a
research tool and will soon reach the production floor just as
we see electron microscopes as an integral part of the
manufacturing of semiconductor devices. The advances in reso-
lution make it possible to better compare results from other
techniques, such as x-ray analysis. Just as new tomographic
holders made it possible to reduce the missing wedge, improving
3D reconstruction of samples, we can now reliably assert that we
can see even single isolated atoms on oxide surfaces. As energy
resolution in a TEM approaches that available in x-ray absorp-
tion spectroscopy, we should see the application of electron loss
near-edge spectroscopy and extended electron energy-loss fine
structure. New sample cells based on Si MEMS devices improve
the stability for imaging at elevated temperatures, making it
possible to perform in situ studies. With windowed cells, it is
also now possible to image nucleation of nanoparticles in the
liquid phase and to study catalytic reactions at atmospheric
pressure in the electron microscope. We therefore conclude that
the future is bright for 3D imaging of catalytic materials and it
will lead to advances in our ability to synthesize catalysts with
atom specificity and low-temperature reactivity leading to
improved energy efficiency. For a related chapter in this Com-
prehensive, we refer to Chapter 7.04.
Acknowledgments
The author acknowledges financial support from grants from
the US National Science Foundation, from the US Department
of Energy, and from industry and national laboratories (Sandia
National Laboratories and Pacific Northwest National Labs).
The research conducted by the authors group makes use of the
facilities located in the Earth and Planetary Sciences
4.2 s
7.8 s
department at the University of New Mexico and at national
facilities such as the HTML at Oak Ridge National Lab and
EMSL at Pacific Northwest National Labs. The author also
thanks the authors whose work is reproduced in the figures
included in this chapter, each of whom is acknowledged in the
figure captions.
References
1. Koster, A. J.; Ziese, U.; Verkleij, A. J.; Janssen, A. H.; De Graaf, J.; Geus, J. W.; De
Jong, K. P. Stud. Surf. Sci. Catal. 2000, 130A, (International Congress on
Catalysis, 2000, Pt. A), 329334.
2. Midgley, P. A.; Weyland, M.; Thomas, J. M.; Johnson, B. F. G. Chem. Commun.
(Cambridge, U. K.) 2001, 10, 907908.
3. van Benthem, K.; Lupini, A. R.; Oxley, M. P.; Findlay, S. D.; Allen, L. J.;
Pennycook, S. J. Ultramicroscopy 2006, 106, (1112 SPEC. ISS.), 10621068.
4. Datye, A. K.; Long, N. J. Ultramicroscopy 1988, 25(3), 203208.
5. Yacaman, M. J.; Ocanaz, T. Phys. Status Solidi A: Appl. Res. 1977, 42(2),
571577.
6. Kotula, P. G.; Keenan, M. R.; Michael, J. R. Microsc. Microanal. 2006, 12(1),
3648.
7. de Smit, E.; Swart, I.; Creemer, J. F.; Karunakaran, C.; Bertwistle, D.;
Zandbergen, H. W.; de Groot, F. M. F.; Weckhuysen, B. M. Angew. Chem. Int. Ed.
2009, 48(20), 36323636.
8. Datye, A. K.; Hansen, P. L.; Helveg, S. In Handbook of Heterogeneous Catalysis,
2nd ed.; Wiley: Weinheim, 2008; Vol. 2, pp 803833.
9. Rupprechter, G.; Seeber, G.; Hayek, K.; Hofmeister, H. Phys. Status Solidi A: Appl.
Res. 1994, 146(1), 449459.
10. Hayek, K.; Fuchs, M.; KlAtzer, B.; Reichl, W.; Rupprechter, G. Top. Catal. 2000,
13(12), 5566.
11. Rupprechter, G.; Seeber, G.; Goller, H.; Hayek, K. J. Catal. 1999, 186(1),
201213.
12. Egerton, R. F. Electron Energy Loss Spectroscopy in the Electron Microscope, 2nd
ed.; Plenum Press: New York, 1996.
13. Huang, Y.; Cowley, J. M. Surf. Sci. 1995, 328(3), 277286.
14. Cowley, J. M.; Liu, J. Surf. Sci. 1993, 298(23), 456467.
15. Herzing, A. A.; Watanabe, M.; Edwards, J. K.; Conte, M.; Tang, Z. R.;
Hutchings, G. J.; Kiely, C. J. Faraday Discuss. 2008, 138, 337351.
16. Muller, D. A. Nat. Mater. 2009, 8(4), 263270.
17. Jin, Y.; Xu, H. F.; Datye, A. K. Microsc. Microanal. 2006, 12, 124134.
18. Shroff, M. D.; Kalakkad, D. S.; Coulter, K. E.; Koehler, S. D.; Harrington, M. S.;
Jackson, N. B.; Sault, A. G.; Datye, A. K. J. Catal. 1995, 156(2), 185207.
19. Jin, Y.; Datye, A. K. Stud. Surf. Sci. Catal. 1998, 119, (Natural Gas Conversion V),
209214.
 Imaging of Heterogeneous Catalysts 129

20. Iijima, S. J. Appl. Phys. 1971, 42(13), 5891. 44. Attwood, D.; Chao, W.; Anderson, E.; Liddle, J. A.; Harteneck, B.; Fischer, P.;
21. Marks, L. D.; Smith, D. J. J. Cryst. Growth 1981, 54(3), 425432. Schneider, G.; Le Gros, M.; Larabell, C. J. Biomed. Nanotechnol. 2006, 2(2),
22. Marks, L. D.; Smith, D. J. Nature 1983, 303(5915), 316317. 7578.
23. Wang, R. M.; Dmitrieva, O.; Farle, M.; Dumpich, G.; Acet, M.; Mejia-Rosales, S.; 45. Thibault, P.; Dierolf, M.; Menzel, A.; Bunk, O.; David, C.; Pfeiffer, F. Science 2008,
Perez-Tijerina, E.; Yacaman, M. J.; Kisielowski, C. J. Phys. Chem. C 2009, 321(5887), 379382.
113(11), 43954400. 46. Chapman, H. N. Science 2008, 321(5887), 352353.
24. Gontard, L. C.; Chang, L. Y.; Hetherington, C. J. D.; Kirkland, A. I.; 47. Rieker, T.; Hanprasopwattana, A.; Datye, A.; Hubbard, P. Langmuir 1999, 15(2),
Ozkaya, D.; Dunin-Borkowski, R. E. Angew. Chem. Int. Ed. 2007, 46(20), 638641.
36833685. 48. Naudon, A.; Babonneau, D.; Thiaudiere, D.; Lequien, S. Phys. B 2000, 283(13),
25. Haider, M.; Uhlemann, S.; Schwan, E.; Rose, H.; Kabius, B.; Urban, K. Nature 6974.
1998, 392(6678), 768769. 49. Renaud, G.; Lazzari, R.; Revenant, C.; Barbier, A.; Noblet, M.; Ulrich, O.; Leroy, F.;
26. Krivanek, O. L.; Dellby, N.; Lupini, A. R. Ultramicroscopy 1999, 78(14), 111. Jupille, J.; Borensztein, Y.; Henry, C. R.; Deville, J. P.; Scheurer, F.; Mane-
27. Krivanek, O. L.; Nellist, P. D.; Dellby, N.; Murfitt, M. F.; Szilagyi, Z. Ultramicroscopy Mane, J.; Fruchart, O. Science 2003, 300(5624), 14161419.
2003, 96(34), 229237. 50. Winans, R. E.; Vajda, S.; Lee, B.; Riley, S. J.; Seifert, S.; Tikhonov, G. Y.;
28. Muller, D. A.; Kourkoutis, L. F.; Murfitt, M.; Song, J. H.; Hwang, H. Y.; Silcox, J.; Tomczyk, N. A. J. Phys. Chem. B 2004, 108(47), 1810518107.
Dellby, N.; Krivanek, O. L. Science 2008, 319(5866), 10731076. 51. Nolte, P.; Stierle, A.; Jin-Phillipp, N. Y.; Kasper, N.; Schulli, T. U.; Dosch, H.
29. Gabor, D. The Electron Microscope; Chemical Publishing: New York, 1948. Science 2008, 321(5896), 16541658.
30. Midgley, P. A.; Dunin-Borkowski, R. E. Nat. Mater. 2009, 8(4), 271280. 52. Drake, I. J.; Liu, T. C. N.; Gilles, M.; Tyliszczak, T.; Kilcoyne, A. L. D.; Shuh, D. K.;
31. Friedrich, H.; de Jongh, P. E.; Verkleij, A. J.; de Jong, K. P. Chem. Rev. 2009, Mathies, R. A.; Bell, A. T. Rev. Sci. Instrum. 2004, 75(10), 32423247.
109(5), 16131629. 53. de Smit, E.; Swart, I.; Creemer, J. F.; Hoveling, G. H.; Gilles, M. K.; Tyliszczak, T.;
32. Kawase, N.; Kato, M.; Nishioka, H.; Jinnai, H. Ultramicroscopy 2007, 107(1), Kooyman, P. J.; Zandbergen, H. W.; Morin, C.; Weckhuysen, B. M.; de
815. Groot, F. M. F. Nature 2008, 456(7219), 222239.
33. Arslan, I.; Tong, J. R.; Midgley, P. A. Ultramicroscopy 2006, 106(1112), 54. Bell, A. T. Nature 2008, 456(7219), 185186.
9941000. 55. Sankey, M. H.; Holland, D. J.; Sederman, A. J.; Gladden, L. F. J. Magn. Reson.
34. Yates, T. J. V.; Thomas, J. M.; Fernandez, J. J.; Terasaki, O.; Ryoo, R.; 2009, 196(2), 142148.
Midgley, P. A. Chem. Phys. Lett. 2006, 418(46), 540543. 56. Gladden, L. F.; Mantle, M. D.; Sederman, A. J. Adv. Catal. 2006, 50, 175.
35. Gommes, C. J.; Friedrich, H.; Wolters, M.; de Jongh, P. E.; de Jong, K. P. Chem. 57. Espinosa-Alonso, L.; Lysova, A. A.; De Peinder, P.; De Jong, K. P.; Koptyug, I. V.;
Mater. 2009, 21(7), 13111317. Weckhuysen, B. M. J. Am. Chem. Soc. 2009, 131(18), 65246534.
36. Gass, M. H.; Koziol, K. K. K.; Windle, A. H.; Midgley, P. A. Nano Lett. 2006, 6(3), 58. Singh, J.; Alayon, E. M. C.; Tromp, M.; Safonova, O. V.; Glatzel, P.; Nachtegaal, M.;
376379. Frahm, R.; van Bokhoven, J. A. Angew. Chem. Int. Ed. 2008, 47(48),
37. Scherzer, O. J. Appl. Phys. 1949, 20(1), 2029. 92609264.
38. Van Benthem, K.; Lupini, A. R.; Kim, M.; Baik, H. S.; Doh, S.; Lee, J. H.; 59. Kalakkad, D.; Anderson, S. L.; Logan, A. D.; Pena, J.; Braunschweig, E. J.;
Oxley, M. P.; Findlay, S. D.; Allen, L. J.; Luck, J. T.; Pennycook, S. J. Appl. Phys. Peden, C. H. F.; Datye, A. K. J. Phys. Chem. 1993, 97(7), 14371444.
Lett. 2005, 87(3), 13. 60. Vesborg, P. C. K.; Chorkendorff, I.; Knudsen, I.; Balmes, O.; Nerlov, J.;
39. Miller, M. K. Atom-Probe Tomography: Analysis at the Atomic Level; Kluwer Molenbroek, A. M.; Clausen, B. S.; Helveg, S. J. Catal. 2009, 262(1), 6572.
Academic/Plenum: New York, 2000. 61. Hansen, P. L.; Wagner, J. B.; Helveg, S.; Rostrup-Nielsen, J. R.; Clausen, B. S.;
40. Qiao, B.; Wang, A.; Yang, X.; Allard, L. F.; Jiang, Z.; Cui, Y.; Liu, J.; Li, J.; Zhang, T. Topsoe, H. Science 2002, 295(5562), 20532055.
Nat. Chem. 2011, 3, 634641. 62. Pham, H. N.; Anderson, A. E.; Johnson, R. L.; Schmidt-Rohr, K.; Datye, A. K.
41. Pala, R. G. S.; Metiu, H. J. Phys. Chem. C 2007, 111(24), 86178622. Angew. Chem. Int. Ed. 2012, 124, 1334013344. http://dx.doi.org/10.1002/
42. Bore, M. T.; Ward, T. L.; Fukuoka, A.; Datye, A. K. Catal. Lett. 2004, 98(4), 167172. anie.201206675.
43. Bore, M. T.; Pham, H. N.; Switzer, E. E.; Ward, T. L.; Fukuoka, A.; Datye, A. K. 63. Challa, S. R.; Delariva, A. T.; Hansen, T. W.; Helveg, S.; Sehested, J.; Hansen, P. L.;
J. Phys. Chem. B 2005, 109(7), 28732880. Garzon, F.; Datye, A. K. J. Am. Chem. Soc. 2011, 133(51), 2067220675.
 2.15 Soft Chemistry Synthesis of Oxides
M Hayward, University of Oxford, Oxford, UK
2013 Elsevier Ltd. All rights reserved.

2.15.1 Introduction 417

2.15.2 Synthesis from Precursors 418
2.15.2.1 Co-Precipitation Precursor Routes 418
2.15.2.2 SolGel Precursor Routes 419
2.15.2.2.1 Inorganic polymer solgel precursors 419
2.15.2.2.1 Organic polymer solgel precursors 419
2.15.3 Synthesis in Solvents and Fluxes 420
2.15.3.1 Solvothermal Synthesis 420
2.15.3.1.1 Early transition-metal oxides 420
2.15.3.1.2 Complex manganese oxides 421
2.15.3.2 Flux Synthesis 422
2.15.4 Low-Temperature Topochemical Synthesis 422
2.15.4.1 Cation-Insertion Reactions 423
2.15.4.1.1 Intercalation into framework oxide phases 423
2.15.4.1.2 Intercalation into layered binary oxide phases 426
2.15.4.1.3 Intercalation into layered perovskite structures 427
2.15.4.2 Cation Deintercalation Reactions 429
2.15.4.2.1 a-NaFeO2-type phases 429
2.15.4.2.2 Lithium manganese oxides 432
2.15.4.2.2 Titanates 433
2.15.4.3 Anion Deintercalation Reactions 434
2.15.4.3.1 Cubic perovskite phases 434
2.15.4.3.2 RuddlesdenPopper phases 437
2.15.4.3.3 Reductive fusion of perovskite sheets 439
2.15.4.3.4 Structural selectivity 439
2.15.4.4 Anion-Insertion Reactions 440
2.15.4.4.1 Anion-deficient perovskites 440
2.15.4.4.2 RuddlesdenPopper phases 442
2.15.4.5 Redox-Neutral Topochemical Reactions 445
2.15.4.5.1 Cation-substitution reactions 445
2.15.4.5.2 Deintercalation 448
2.15.4.5.3 Redox-neutral intercalation 449
2.15.5 Conclusion 450
References 450

2.15.1 Introduction
Complex transition-metal oxides are of wide and enduring
interest due to the complex structural chemistry and broad
range of physical behavior these phases exhibit. The synthesis
of complex oxides has therefore been the subject of extensive
study (Chapters 2.01 and 4.03).
In contrast to gas-phase or solution-phase chemistry, where
the diffusion and mixing of reagents are relatively low-energy
processes, a characteristic feature of chemical reactions in the
solid state is the low mobility of chemical species. This is
especially true in the synthesis of metal oxides. As demon-
strated by the high lattice energies and melting points of
metal-oxide phases, there are strong bonding interactions be-
tween small, highly charged oxide ions and metal cations.
These interactions act to resist the motion of chemical species
within the extended lattices of metal oxides and thus present
large energetic barriers to the diffusion of reagents.
Typical ceramic synthesis strategies applied to the prepa-
ration of complex oxide phases involve heating mixtures of
binary metal oxides and/or metal oxy-acid salts (carbonates,
nitrates, etc.) at high temperature. Rapid reaction occurs ini-
tially at the contact point between component phases and then
proceeds by slow interdiffusion between reactant particles to
form product phases. Reactions of this form therefore require
ionic diffusion over length scales defined by the particle size of
reactants (typically 110 mm) in order to form homogeneous
products. While reaction rates can be increased by regular
grinding to expose fresh reagents and by compressing samples
to increase the degree of inter-grain contact, the intrinsically
poor ionic diffusion in metal-oxide phases means that overall
reaction rates are slow and reactions require multiple heating

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00219-9 417

418 Soft Chemistry Synthesis of Oxides
periods (typically measured in days) at elevated temperature
(T > 1000  C) to reach completion. However, even after pro-
longed heat treatment, the homogeneity of samples prepared
by ceramic processes can often be poor. For example, the
reaction between two component oxides AO and BO2 will
form not only the desired ABO3 product, but also A2BO4 or
AB2O5 depending on the local stoichiometry of the reaction
mixture. Thus extensive heating and grinding can be required
for samples to reach equilibrium, and the contamination
of samples with low levels of impurity phases is common.
In addition, the extensive physical processing of materials
required by ceramic synthesis (grinding, ball milling, and com-
pression) can also introduce impurities into samples. Thus,
it can be seen that ceramic routes are not optimal if high levels
of sample purity and homogeneity are required.
A second potentially undesirable feature of high-
temperature ceramic synthesis relates to product selection. As
noted above, there are considerable barriers to diffusion in
solid oxide systems. These barriers are in fact so large that
mass transport contributions account for the majority of the
activation energy of a typical solidsolid reaction. Given that
the motion of chemical species is a feature common to all
solidsolid reactions, it follows that the activation energies of
competing solidsolid reactions tend to be comparable. Thus
at the elevated synthesis temperatures required to overcome
these activation barriers, the products of these competing
reactions are all energetically accessible, and enter an equilib-
rium with each other resulting in the formation of the most
thermodynamically stable phase or mixture of phases. Thus we
can see that, as a consequence of the large energetic barriers to
diffusion in solids, solidsolid reactions are generally per-
formed under the influence of thermodynamic product selec-
tion. As a result, a large number of metastable product phases
are synthetically inaccessible by this route.
Motivated by the undesirable features of high-temperature
solid-state synthesis, there has been a concerted effort to
soften the reaction conditions (lower the reaction tempera-
tures) required for the synthesis of complex oxides by applying
a chimie douce approach.14
Strategies can be split into three major strands:
1. Methods that reduce the length scale over which solidsolid
diffusion must occur, by improving the mixing of reactant
species and/or reducing the particle size of reagent phases.
2. Methods which lower the barriers to diffusion by utilizing
an external medium or solvent to circumvent the need for
extensive solidsolid reaction.
3. Low-temperature topochemical reactions which exploit the
high mobility of particular components within extended
oxide phases, to bring about structure-conserving reactions
which form metastable phases.
2.15.2 Synthesis from Precursors
As noted above, ceramic synthesis routes involve reacting mix-
tures of binary metal oxides and metal oxo-acid salts (carbon-
ates and nitrates). Even after extensive mechanical processing,
these mixtures are extremely inhomogeneous on the length
scale of a typical mixed-oxide structural repeat unit (520 A),
and as a result extensive long-range diffusion of species is
required to form homogeneous products.
An obvious strategy for reducing reaction times and lower-
ing reaction temperatures is to prepare more homogeneous
mixtures of reagents. For example, if the homogeneity and
intermixing of a solution of metal cations could be trapped
in a solid precursor, so that metal cations were mixed on an
atomic length scale, much softer reaction conditions could be
used to prepare homogeneous materials. However, simply dry-
ing a solution of metal salts will not prepare such a precursor
due to the propensity of metal salts to crystallize into binary
phases rather than solid solutions. Thus, for example, a mixed
solution of metal nitrates will crystallize into a mixture of
binary nitrate crystals when the solvent is driven off. While
this mixture of small crystals may be more homogeneous than
a mechanically ground sample, it is still far from homogeneous
on the atomic length scale. More elaborate procedures are
therefore required to form suitable precursor materials.
2.15.2.1 Co-Precipitation Precursor Routes
Rapid co-precipitation of metal cations from solution offers
one method for preparing intimately mixed precursor phases.
In the simplest case, rapid precipitation leads to the formation
of very small particles of binary metal salts in suspension. If
these precipitated phases are sufficiently insoluble in the liquid
from which they were precipitated, the suspended particles will
not ripen and grow, and a finely divided mixture of reagent
phases can be separated by filtration for subsequent heat
treatment.
More homogeneous precursor phases can be prepared if
the reagent metal cations can be precipitated to form a solid-
solution salt. For example, it has been observed that many
binary carbonates of divalent metals adopt calcite-type struc-
tures related to that of CaCO3.5 The lattice parameters of these
carbonate phases show little variation with the identity of the
metal cation, suggesting a high degree of cation miscibility
between phases and further suggesting that materials contain-
ing cation solid solutions can be formed.6 Thus, for example, if
(NH4)2CO3 is added to a solution containing Ca(NO3)2 and
Mn(NO3)2, a Ca1xMnxCO3 solid-solution phase is precipi-
tated in which calcium and manganese cations are statistically
distributed over the metal cation sites of the calcite lattice and
are therefore homogeneously mixed on a unit cell length
scale.7 Decomposition of this mixed-metal carbonate readily
yields highly homogeneous calcium manganese oxide phases
as no long-range cation diffusion is required during the sinter-
ing process.
Similar procedures can be utilized to prepare mixed-metal
hydroxide precursor phases. Thus, for example, La0.5M0.5(OH)3
(M Al, Cr, Fe, Co) precursor phases precipitated from solution
can be readily transformed into the corresponding LaMO3 pe-
rovskite phases at low temperature (600700  C).8 In this case,
there is some choice of precipitating agent. Sodium hydroxide
and potassium hydroxide have been widely used as agents to
precipitate mixed-metal hydroxides. However, this usually re-
quires that the sodium and potassium cations are removed,
typically by washing in hot water, after precipitation is complete
and prior to thermal treatment. Therefore, the bases of nonmetal
cations such as ammonium or substituted variants are usually

preferred as these can be burned out of samples during the
sintering process leaving samples uncontaminated. Subsequent
sintering of co-precipitated precursor phases to yield highly
crystalline, homogeneous samples can be performed at much
lower temperatures than that would be required for ceramic
syntheses.
2.15.2.2 SolGel Precursor Routes
The preparation of homogeneous precursor phases containing
more than two or three different cations can be extremely
challenging via co-precipitation routes. However by utilizing
solgel chemistry, precursor phases which have homogeneous
distributions of large numbers of different cations can be read-
ily prepared.9,10 The term sol is generally used to refer to a
suspension or dispersion of colloidal particles. A gel is a
highly viscous fluid consisting of a polymer network which
contains large amounts of liquid within its structure. Thus,
precursor materials prepared by solgel chemistry consist of
intimately mixed arrays of coordinated metal cations or colloi-
dal particles, trapped within an extended polymer matrix, such
that when the precursor phase is dried, the homogeneous
mixture of metal cations present in solution/suspension is
retained in the resulting material.
Solgel precursor phases can be split into two broad
classes: inorganic polymeric materials and organic polymeric
materials.
2.15.2.2.1 Inorganic polymer solgel precursors
Solgel precursor phases based on inorganic polymers contain
continuous networks of MOMOM polymers, such that
the polymer chain is made from the trapped metal cations
which will ultimately form the product metal-oxide phase.
Such inorganic polymer solgel precursors can be prepared
by two principal routes: the hydrolysis of metal alkoxides and
the slow hydrolysis of strongly chelated metal cations.
2.15.2.2.1.1 Alkoxide hydrolysis
Metal alkoxides of the general formula M(OR)n, where R is an
alkyl group, are readily hydrolyzed as shown in reaction [1]
(Chapter 2.16). Hydrolysis can then be followed by dehydra-
tion or dealcholation to form condensed MOM links as
shown in reactions [2] and [3].11 Repeated hydrolysis steps
can be used to build up extended polymeric oxide networks:
MORn xH2 O ! MOHx OR nx xROH [1] adjust the acidity of the solution to pH 5, followed by gentle
M  OH HO  M ! M  O  M H2 O [2]
 M  OH RO  M ! M  O  M  ROH [3]
In order to prepare homogeneous precursor phases which
contain different types of metal cations, it is necessary to per-
form cross-hydrolysis reactions to incorporate all the metal
cations present in solution into the inorganic polymer chains.
This requires that the alkoxides of the different metal cations
present have similar hydrolysis rates, which can be a significant
obstacle to the formation of complex oxide phases by this
route. For example, rare earth alkoxides are much more readily
hydrolyzed than copper alkoxide complexes. Therefore, the
Soft Chemistry Synthesis of Oxides 419
introduction of water into a mixture of rare earth and copper
alkoxides would result in the preferential hydrolysis of the rare
earth alkoxides resulting in a highly inhomogeneous precursor
phase. The relative rates of alkoxide hydrolysis can be modified
by using substituted alkoxide ligands. Thus by using the ap-
propriate metal alkoxide complexes a wide range of mixed-
metal oxides can be prepared via this precursor route.
An alternative strategy is to prepare mixed-metal alkoxide
phases which can be hydrolyzed directly to form solgel pre-
cursor phases. For example, a mixture of Ba(OCH2CH2OCH3)2
and Ti(OCH2CH2OCH3)4 in 2-methoxyethanol (HOCH2-
CH2OCH3) leads to the formation of the mixed-metal alkoxide
BaTi(OCH2CH2OCH3)6.12 Hydrolysis with excess water read-
ily yields a solgel precursor phase which can be used to
prepare BaTiO3 at only 400  C, compared to the 1100  C
required for ceramic synthesis.13
2.15.2.2.1.2 Metal chelate gels
As noted above, metal complexes, particularly those in aque-
ous solution, are prone to hydrolysis and condensation poly-
merization. The addition of base to a solution of metal cations
can catalyze the rapid hydrolysis of the metal complexes pre-
sent. However, this generally results in the precipitation of
binary metal hydroxide particles, rather than the formation of
inorganic polymers. The rate of cation hydrolysis can be slo-
wed significantly if the metal centers are bound strongly to
chelating ligands, allowing a polymeric network to form. For
example, the complexation of Fe(III) cations with ethylenedia-
mine tetraacetic acid (EDTA) can lower the equilibrium con-
stant of hydrolysis by 22 orders of magnitude as shown in
reactions [4] and [5]14:
FeH2 O6 3 H2 O $ FeOHH2 O5 2 H3 O
[4]
K 103
   2
FeH2 O2 EDTA H2 O $ FeOHH2 OEDTA
H3 O [5]
25
K 10
Utilizing this approach, precursor phases which incor-
porate large numbers of different metal cations can be
prepared. For example, solgel precursors suitable for the for-
mation of the high-Tc superconducting copper oxide phase
Bi2Sr2CaCu2O8x can be readily prepared by adding EDTA
to an aqueous solution of the corresponding metal nitrates,
to achieve a cation:EDTA ratio of 1:1. Addition of ammonia to
heating to drive off water, leads to the formation of a glassy
solgel precursor material, which can then be thermally treated
to form the target Bi2Sr2CaCu2O8x phase.15
2.15.2.2.1 Organic polymer solgel precursors
A further strategy for the preparation of solgel precursor
phases is the use of organic polymers to form a highly
connected, extended, semi-rigid network containing a homo-
geneous distribution of metal cations immobilized within
coordination sites. The inorganicorganic hybrid precursor
can then be sintered into a homogeneous complex oxide
phase by burning out the organic framework.
420 Soft Chemistry Synthesis of Oxides
The most direct way of preparing such a precursor is the
addition of a water-soluble organic polymer, such as poly-
acrylic acid [CH2CH(COOH)]n or polyvinyl alcohol [CH2CH
(OH)]n, to a solution of metal cations.10 Complexation of the
metal cations by the acid or alcohol residues of the organic
polymer chain leads to crosslinking and immobilization of
metal cations within the precursor polymer network. Thus if
a 10% solution of polyacrylic acid is added to a mixed solution
of yttrium, barium, and copper cations, a viscous gel is formed
as the solution is concentrated by evaporation, which then sets
into an amorphous glassy precursor. Subsequent sintering of
this precursor yields YBa2Cu3Oz.16
In order to ensure that highly homogeneous precursor ma-
terials are prepared by this method, the polymer chains must
have a uniform complexing affinity for the different metal
cations present in solution. This requirement can be readily
achieved by the use of polymers with strong polydentate coor-
dination sites, such as EDTAethylene diamine polyamide,
that show a strong affinity for almost all metal cations.17
Organic polymer solgel precursors can also be prepared by
the in situ polymerization of organic ligands which are chelat-
ing metal centers. The most common chemical system used in
such processes is a combination of citric acid and ethylene
glycol, first described in the Pechini process.18 Citric acid is a
tridentate ligand which binds strongly to a wide variety of
metal cations. Reaction with ethylene glycol leads to a rapid
esterification reaction, which can drive an efficient condensa-
tion polymerization reaction of the citric acid units. Thus if an
aqueous solution of citric acid, ethylene glycol, and metal salts
is concentrated by heating to favor condensation polymeriza-
tion, a polyester network is formed trapping a homogeneous
array of coordinated metal cations. The resulting precursor
phase can be readily dried and sintered at low temperature to
form homogeneous complex oxide phases.
Citric acid and related polymerizable organic ligands form
strong bonds to a very wide variety of metal cations. This
makes the citrate gel precursor route suitable for the synthesis
of a large number of complex oxide phases which contain a
variety of different metal cations.19,20 This broad applicability
is particularly valuable for the synthesis of phases which con-
tain solid solutions of metal cations in which the physical
properties of the materials are decisively dependent on sample
homogeneity.
2.15.3 Synthesis in Solvents and Fluxes
The conventional ceramic synthesis of mixed-metal-oxide
phases requires cations to diffuse in the solid state over length
scales defined by the particle size of the reagent materials. The
tight binding of metal cations within extended oxide lattices
means that in general the rate of solid-state cation diffusion in
oxides is low, necessitating high reaction temperatures for
ceramic synthesis procedures. Rapid low-temperature synthesis
of complex oxide phases can be facilitated if the diffusion of
species occurs in a fluid medium, rather than in the solid state.
Thus by the use of molten fluxes or liquid solvents, in which
the diffusing species are soluble, the synthetic conditions for
oxide synthesis can be softened. An additional advantage of
such a synthetic approach is that the dissolution process
is reversible, facilitating the control of particle size and crystal
growth.
2.15.3.1 Solvothermal Synthesis
Solvothermal reactions, in which a solvent is used as a reaction
medium in a sealed apparatus above its boiling point, have
been utilized to prepare an extensive range of complex oxide
phases.21,22 Such reactions utilize the autogeneous pressure
which is generated under such conditions to modify the solva-
tion properties of solvents, and thus favor the dissolution of
insoluble oxide reagent phases. Taking water as an example,
the viscosity of this solvent decreases dramatically with increas-
ing temperature, facilitating the diffusion of dissolved spe-
cies.23 In addition, the ionic product and dielectric constant
of water are also strongly temperature dependent. These
changes have dramatic consequences on the solubility of
metal oxides in the temperature range typically employed for
synthesis. For example, the solubility of the rutile polymorph
of TiO2 increases by two orders of magnitude on heating to
300  C.24 Thus, the products of solvothermal synthesis routes
are often strongly dependent on the synthesis temperature and
pressure.
2.15.3.1.1 Early transition-metal oxides
Hydrothermal synthesis has been used widely to prepare di-
electric oxide phases containing early transition metals in their
group oxidation states (TiIV, NbV, and TaV). One of the most
extensively studied phases is the ferroelectric BaTiO3. Conven-
tional preparation of this phase via a ceramic synthesis route
involves the repeated heating of mixtures of BaCO3 and TiO2 at
1100  C.13 In contrast, the ferroelectric tetragonal polymorph
of BaTiO3 can be readily prepared by the hydrothermal reac-
tion of BaCl2, TiO2, and NaOH at 200  C.25 The much lower
reaction temperature employed clearly demonstrates the utility
of water as a solvent to facilitate the diffusion of reacting
species.
Interestingly, if Ba(OH)2 is used as the barium source in
this reaction rather than BaCl2, the cubic polymorph of BaTiO3
is prepared. It has been hypothesized that chloride ions facil-
itate the nucleation of BaTiO3 crystals in some manner. This
enhances the crystal size of the hydrothermal product suffi-
ciently to stabilize the structural transition to the low-
temperature tetragonal polymorph of BaTiO3.25 Such interac-
tions and effects demonstrate the sensitivity of hydrothermal
reactions to subtle influences from spectator species. This has
motivated many detailed studies of the mechanisms of this
class of reaction in order to utilize and control these features
of solvothermal chemistry.
Cation-substituted phases and solid solutions can also be
readily prepared by solvothermal methods. For example,
phases across the entire Ba1xSrxTiO3 compositional range
can be synthesized under the appropriate conditions.26 The
rapid diffusion and intimate mixing of reagents in solution
leads to products of high homogeneity with compositions
which are in good agreement with the ratio of reagents used.
Group 5 AMO3 (A Na, K; M Nb, Ta) perovskite phases
can also be readily prepared via solvothermal routes.2735 For
example, direct reaction between Nb2O5 and sodium hydroxide
solutions readily yields crystals of the cubic perovskite-phase

NaNbO3 with a cubic habit, at much lower temperatures than
that required for solid-state single-crystal growth. By lowering
the pH of the reaction, a metastable polymorph of NaNbO3
with the ilmenite structure is formed after 3 h.36 This further
emphasizes the sensitivity of solvothermal synthesis to the spe-
cific conditions employed. Indeed in this instance the reaction is
so sensitive to conditions that either raising the concentration of
NaOH in the reaction mixture (raising pH), or increasing the
reaction time, leads to the formation of the perovskite poly-
morph of NaNbO3 in increasing quantities.
The sensitivity of solvothermal synthesis to pH and other
reaction conditions can also be seen in the solvothermal prep-
aration of the pyrochlore-phase Pb2Ti2O6.37 Reaction of Pb(II)
acetate and titanium n-butoxide in aqueous KOH yields
Pb2Ti2O6, an anion-deficient pyrochlore phase in which the
O0 anion site of the A2B2O6O0 lattice is vacant. When the
concentration of potassium hydroxide in this synthesis is
high, products are formed in which the PbII and TiIV cations
are ordered on the A- and B-cation sites of the pyrochlore
lattice, respectively. However, synthesis at lower KOH concen-
trations yields products in which the cations are statistically
disordered over both cation sites demonstrating the influence
of pH on the synthesis. Heating either anion-deficient pyro-
chlore product above 500  C leads to a transformation to a
perovskite structure, demonstrating the ability of solvothermal
reactions to prepare metastable phases.
2.15.3.1.2 Complex manganese oxides
The solvothermal synthesis of complex manganese oxides has
been extensively studied due to the varied magnetic and elec-
tronic properties these phases exhibit38 (Chapter 4.03 and
4.11). The synthesis of complex manganese oxides under sol-
vothermal conditions is more complicated than that of the
early transition-metal oxides, due to the range of oxidation
states (MnIIMnIV) and nonstoichiometry (cation and anion
deficiency) exhibited by the manganate phases. In high-
temperature ceramic syntheses, the average manganese oxida-
tion state present in product phases is controlled via the partial
pressure of oxygen in equilibrium with the solid-state reaction.
In the analogous solvothermal syntheses, manganese oxida-
tion states are controlled by the comproportionation of a Mn
(II) salt and the Mn(VII) permanganate ion. Thus, for example,
the reaction of Ho(NO3)3, KMnO4, and MnCl2 in alkaline
solution readily yields the Mn(III) perovskite-phase HoMnO3
according to reaction [6].39 By utilizing this approach, crystals
of a wide range of REMnO3 (RE SmHo) cubic-type perov-
skite phases have been prepared3941:
5HoNO3 3 KMnO4 4MnCl2 22KOH
! 5HoMnO3 15KNO3 8KCl 11H2 O [6]
Reactions of the smaller rare earths (ErLy Y) at higher
temperature, utilizing Mn2O3 as a source of manganese,
yield hexagonal REMnO3 phases which adopt structures con-
sisting of sheets of apex-linked MnO5 trigonal bipyramids
(Figure 1).42 The high reaction temperatures of these syntheses
are required to decompose RE(OH)3 to the more reactive REO
(OH) as reflected by the variation in reaction temperature re-
quired for LuMnO3 (175  C) compared with YMnO3 (350  C).
Soft Chemistry Synthesis of Oxides 421
YMnO3
Figure 1 The layered structure of hexagonal YMnO3 consists of sheets
of apex-linked MnO5 trigonal bipyramids.
The decisive role that the average manganese oxidation state in
the reaction medium plays in the solvothermal synthesis of com-
plex manganese oxides can be demonstrated by tuning this
parameter. For example, if the MnII:MnO4 ratio is adjusted to
yield an average oxidation state of Mn3.5, REMn2O5 phases are
produced rather than Mn(III) perovskite phases. Thus, for
example, TbMn2O5 can be prepared according to reaction [7]41,43:
5TbNO3 3 3KMnO4 7MnCl2 26KOH
! 5TbMn2 O5 15KNO3 14KCl 13H2 O [7]
Following the same strategy of manganese oxidation state
tuning, the Mn(IV) hexagonal perovskite phases 2HBaMnO3
and 4HSrMnO3 can be readily prepared from a 4:3 ratio of
KMnO4:MnSO4 in the presence of basic solutions of BaCl2 and
SrSO4, respectively.44,45 In the case of 2HBaMnO3, this leads
to a much more rapid synthesis than the extended ceramic
process required to prepare this phase.46 By raising the tem-
perature during the synthesis of BaMnO3 to 420  C, and thus
the autogeneous pressure to 600 atm, the high-pressure 9R
polymorph of the material is prepared,44 demonstrating a
further parameter by which solvothermal synthesis reactions
can be controlled.
A wide range of mixed-valence MnIII/IV RE1xAxMnO3 perov-
skite phases can also be prepared via this method.45,4750 For
example by heating the appropriate mixture of MnII and MnIV
salts in a basic solution containing Pr(NO3)3 and Ca(NO3)2,
Pr1xCaxMnO3 phases with x 0.39, 0.46, 0.7, and 0.76 have
been prepared.51 This again demonstrates that the oxidation
state of the mixed-valent manganate product is directly defined
by the comproportionation of MnII and MnVII species.
The majority of manganese perovskite product phases pro-
duced by solvothermal routes appear to be directly compara-
ble, both structurally and in terms of magnetic behavior, to
analogous phases produced by the ceramic method. However,
there are some exceptions. For example, the ceramic synthesis
of the MnIII/IV-phase La0.5Ba0.5MnO3 yields a simple cubic
perovskite phase in which the La3 and Ba2 cations are sta-
tistically distributed over the A-cation site of the perovskite
lattice.52 However, the solvothermal preparation of
La0.5Ba0.5MnO3 from a 7:3:5:5 ratio of Mn2:Mn7:Ba2:La3
in basic solution yields a tetragonally distorted perovskite
phase in which the La3 and Ba2 cations are ordered into
layers.53,54 This A-cation-ordered phase can be prepared via
ceramic routes,52 but only in a two-step process in which an
A-cation-ordered, anion-deficient-phase LaBaMn2O5 is pre-
pared first and then topochemically oxidized as described in
Section 2.15.4.4.1.2, to form the metastable cation-ordered,
oxygen-stoichiometric material.
422 Soft Chemistry Synthesis of Oxides
2.15.3.2 Flux Synthesis
Metal salts with low melting points can also be used as fluid
media in which to perform low-temperature solid-state
reactions. An enormous variety of salt fluxes have been
utilized, with a particular emphasis of crystal growth from
these media.55 However, by far the most useful class of flux
for oxide synthesis are the hydroxides of group 1 and 2
metals.56 Simple binary hydroxides have melting points
ranging from 272 (CsOH) to 512  C (Sr(OH)2) and readily
form eutectic mixtures with melting points in the range
160325  C,57 and so can act as molten fluxes even at
modest reaction temperatures as demonstrated by the syn-
thesis of La2xMxCuO4 (M Na, K) from a NaOH/KOH flux
at 300  C.58
The extensive and adjustable acidbase chemistry of molten
hydroxides makes them among the best solvent systems for
oxide phases. When molten, the auto-ionization equilibrium
described in reaction [8] is established, in which H2O can be
considered an acid and O2 a base.
2OH $ H2 O O2 [8]
By analogy to the auto-ionization constant of water, Kw, the
dissociation constant of hydroxide melts, Kd, is defined as
Kd [H2O][O2], with pH2O log10[H2O] being a measure
of acidity in hydroxide fluxes analogous to pH in aqueous
systems. The dissociation constant Kd is observed to be strongly
dependent on temperature and on the identity of the metal
cations present, with cations of high charge density favoring
greater dissociation. Thus, the pH2O (acidity) of a hydroxide
flux can be controlled by selecting an appropriate reaction
temperature and metal counterions and then by adjusting the
level of hydration of the melt, by adding water or removing it
by evaporation.
This acidity adjustment is often decisive to the success of
syntheses as the solubility of metal cations in hydroxide fluxes
is strongly dependent on the pH2O of the melt. By carefully
controlling this parameter, initially to dissolve the reagent
cations and then to precipitate (crystallize) them as ternary
and quaternary phases, a wide range of complex oxides can
be prepared, many of which are metastable and unattainable
by ceramic methods.56
A particularly useful aspect of flux synthesis, with regard to
soft chemistry, is the preparation of highly oxidized phases
from metal hydroxide fluxes. As noted in Section 2.15.4.3.1,
at elevated temperatures complex oxides are in equilibrium
with the oxygen partial pressure in the atmosphere in contact
with their surfaces according to reaction [31]. Due to the
favorable entropy change on oxygen release, the oxygen con-
tent of product phases tends to decline as the synthesis tem-
perature rises. As a result, the preparation of complex metal
oxides containing transition metals in elevated oxidation states
can be extremely challenging if the reagent phases are highly
refractory and thus only reactive at high temperature.
The use of metal hydroxide fluxes can alleviate these prob-
lems in two ways. First, by acting as a fluid transport medium
and increasing the rate of cation diffusion, metal hydroxide
fluxes lower the required reaction temperature, favoring more
oxidized products. Second, hydroxide fluxes can react with
atmospheric oxygen to form peroxides (reaction [9]) and
superoxides (reaction [10]), which act as oxidizing agents
within the melt59:
1
2MOH O2 ! M2 O2 H2 O [9]
2
3
2MOH O2 ! 2MO2 H2 O [10]
2
As a result, molten hydroxide fluxes can be employed to
prepare metal oxides containing highly oxidized transition-
metal centers under ambient oxygen partial pressures. For
example, by heating a mixture of Nd2O3 and elemental iridium
at 550  C within a KOH flux, the Ir(VI)-phase Nd2K2IrO7 is
formed, which contains isolated IrVIO6 octahedra imbedded
within a neodymiumpotassium-oxide matrix.60 By compari-
son, the preparation of the Ir(VI) double perovskite-phase
Sr2CaIrO6 via a ceramic route requires an applied oxygen pres-
sure of 600 atm in order to stabilize the Ir6 cation.61 By
utilizing a strontium hydroxide flux, crystals of the highly
oxidized Rh(V) phases Sr3ARhO6 (A Li, Na) can be grown
from melts containing RhO2 and the corresponding AOH
hydroxide.62 The Sr3ARhO6 phases adopt K4CdCl6-type struc-
tures and represent the first Rh(V) oxides prepared.
Hydroxide fluxes can also be used to prepare oxide phases
containing highly oxidized 3d transition-metal centers. For
example, the Fe(IV)-containing RE0.5Sr1.5FeO4 (RE Nd, Sm,
Eu) phases can be readily prepared from the correspond-
ing lanthanide oxide, strontium hydroxide, and iron oxide
in a KOH flux.63 Similarly, the Ni(IV)-containing phases
Ba6Ni5O15 and BaNiO3 can be prepared from nickel oxide
and barium carbonate in a KOH flux.64,65 The ceramic synthe-
sis of complex oxide phases containing Fe(IV) or Ni(IV) centers
usually requires high oxygen pressures.66,67
Thus by employing hydroxide fluxes as strongly oxidizing
environments for low-temperature synthesis, an extensive
range of complex oxides containing highly oxidized transition-
metal centers can be prepared.56
2.15.4 Low-Temperature Topochemical Synthesis
Solid-state reactions performed at low temperature can utilize
large differences between the rates of solidsolid diffusion of
the different ions present in complex oxides, to prepare novel
metastable phases. For example, in certain complex oxides
particular classes of ion are observed to be significantly more
mobile, at a given temperature, than the remainder of the host
phase in which they reside. As a result, these highly mobile
species can be inserted into, or removed from, host phases
under conditions in which the remaining constituents of the
host lattice are effectively immobile. The resulting chemical
transformations are described as topochemical because they
conserve the basic topology of the parent complex oxide phase.
The term topotactic is often used interchangeably with
topochemical; however, it should be noted that topotactic
has a more crystallographic meaning, as it also requires the
orientation of crystallites in the product phase to be correlated
with those of the starting material. The conservation of struc-
ture observed in low-temperature topochemical reactions is in
stark contrast to conventional high-temperature ceramic syn-
thesis, in which there is generally no relation between the

structures of reagents and products. By utilizing this structural
conservation and carefully selecting the initial complex
oxide phase, topochemical reactions enable a degree of synthe-
sis planning and product targeting unavailable in high-
temperature synthesis routes.
The large difference in mobility between the mobile species
and the host lattice observed under soft chemical conditions
corresponds to a large difference in reactivity. This difference in
reactivity can be exploited to enable some kinetic control to be
exerted over the outcome of solid-state reactions. By perform-
ing reactions at temperatures where there is insufficient ther-
mal energy for the host lattice to respond to changes in the
composition or structure of the overall phase, solid-state re-
actions can be performed in which the products are those
which form fastest, rather than those which have the most
thermodynamically stable structures and/or compositions.
The resulting product phases are metastable and generally
synthetically inaccessible by conventional high-temperature
synthesis routes. Thus, it can be seen that in contrast to the
soft chemical synthesis routes described above, topochemical
syntheses do not attempt to overcome the intrinsically poor
rates of solidsolid diffusion in the reactions of complex ox-
ides. Instead, these synthetic strategies make use of the refrac-
tory nature of the complex oxide host lattices to set up a
situation in which one component of a complex oxide phase
is much more reactive than the remainder, to facilitate the
preparation of metastable complex oxide phases.
The low-temperature topochemical chemistry of complex
oxides described below is organized with respect to the identity
of the mobile species. It should be noted, however, that the
species added or removed from the parent oxide phase must
be electrically neutral to conserve the electroneutrality of the
product. Thus, the addition of a cation or removal of an anion
will reduce the host phase and conversely the addition of an
anion or removal of a cation will lead to a formal oxidation.
2.15.4.1 Cation-Insertion Reactions
In order to effectively insert cations into metal-oxide host
lattices in a topochemical manner, a number of conditions
need to be met.
1. There must be intercalation sites within the host framework
which can accommodate the inserted cations. These sites
can either be existing vacant sites within a close-packed
anion framework, or cation sites between weakly bound
structural layers.
2. The host lattice must contain reducible metal centers. The
insertion process is accompanied by a reduction of the host
lattice as shown in reaction [11]. Formally, a neutral species
is inserted into the metal-oxide host which auto-ionizes to
yield a cation, donating the resultant electron to the frame-
work, which it formally reduces:
A ! An ne [11]
3. There must be ample mobility of both the inserted cation
and the donated electron to allow penetration of the bulk
solid, not just the surface.
Soft Chemistry Synthesis of Oxides 423
2.15.4.1.1 Intercalation into framework oxide phases
Transition-metal cations in high oxidation states form strong
networks of metaloxygen covalent bonds in the binary oxides
they form. As a result, these phases tend to adopt structures
which optimize the geometry of these metaloxygen interac-
tions at the expense of more conventional nondirectional,
Coulombic, close-packing considerations. In addition, the
high cation charge of the transition-metal centers strongly
disfavors face-sharing and edge-sharing connectivity between
neighboring metaloxygen polyhedra, driving the prevalence
of corner-sharing MOx polyhedral networks. As a consequence
of these factors, the binary oxides of highly oxidized transition
metals often exhibit quite open structures which include large
numbers of vacant coordination sites. Intercalation of cations
into these sites allows the topochemical reduction of the host
phases to form metastable intercalation compounds.
2.15.4.1.1.1 ReO3-type hosts
Small electropositive cations can be readily intercalated into
dense oxide host phases which contain suitable cation vacan-
cies within their extended frameworks. ReO3 adopts a structure
consisting of a three-dimensional (3D) network of corner-
sharing ReO6 octahedra (Figure 2).5 This framework is
topologically identical to the TiO6 network in the mineral
perovskite, CaTiO3 (Figure 2). As such the structure of ReO3
contains large vacant cation sites which would be occupied by
the A-cations in ABO3 perovskite phases. These empty sites will
readily host inserted cations as demonstrated by the reaction
of ReO3 with lithium iodide which results in the formation of
Li0.2ReO3 as shown in reaction [12]68:
x
xLiI ReO3 ! Lix ReO3 I2 0 < x < 0:2 [12]
2
Using the more powerful reducing agent n-butyllithium,
greater levels of lithiation can be achieved, ultimately leading
ReO3 CaTiO3
BuLi
ReO3 Li2ReO3
Figure 2 The structures of ReO3 and CaTiO3 perovskite. Insertion of
lithium into ReO3 drives a cooperative rotation of the ReO6 polyhedra to
convert the 12-coordinate A-site into a pair of 6-coordinate sites
occupied by lithium.
424 Soft Chemistry Synthesis of Oxides
to the Re(IV)-phase Li2ReO3.69,70 Subsequent oxidation with
ethanol can partially delithiate the material, allowing the prep-
aration of LiReO3 and Li0.35ReO3. Neutron powder diffraction
studies have revealed that the lithium inserted into the ReO3
network is accommodated within octahedral sites within the
highly distorted ReO3 framework. The ReO6 octahedra un-
dergo a cooperative rotation around the h111i body diagonal
of the cubic unit cell, to convert the 12-coordinate cavity in the
ReO3 network into two edge-sharing octahedral interstitial
sites as shown in Figure 2.71 Formation of Li2ReO3 fills both
of these sites to form a structure consisting of chains of face-
sharing MO6 octahedra which share edges. Partial delithiation
to form LiReO3 results in an ordered arrangement of Re(V), Li
(I), and vacancies within the same anion framework.
Lithium intercalated into the homologous oxide WO3 is
accommodated in a different manner. Reaction of WO3 with
n-butyllithium allows the preparation of LixWO3 (0  x  0.67)
at ambient temperature.72 In the resulting materials, the WO3
host lattice undergoes a cooperative distortion which converts
75% of the large 12-coordinate A-cation sites into 4-coordinate
rectangular planar sites which are occupied by lithium
(Figure 3).68 Full occupation of these fourfold sites would
result in materials of composition Li0.75WO3 suggesting that
the low level of lithium insertion which can be achieved for
WO3, compared to ReO3, has a structural origin.
In contrast, NaxWO3 (0  x  0.9) phases prepared at
high temperature accommodate sodium ions within the 12-
coordinate cation sites of the host lattice with a minimum of
distortion to the WO3 network, consistent with the larger size
of the sodium cation.73 This undistorted lattice can accommo-
date sodium cations up to a composition of Na0.9WO3. Thus,
rather counter-intuitively, it is possible to insert a higher con-
centration of large sodium ions than small lithium ions into
the WO3 network, due to the large lattice distortion required
in the latter case.
In addition to metal cations, hydrogen can also be interca-
lated into metal-oxide lattices. Reaction at high temperature
leads to the removal of oxygen from the extended metal
oxygen lattice as described in detail in Section 2.15.4.3.
However, at low temperatures reductive intercalation reactions
are possible.
Insertion of hydrogen into ReO3 is achieved simply by
boiling in water, according to reaction [13].74,75 The resulting
gold lustrous material is thought to have a limiting composi-
tion of H0.4ReO3:
1 xReO3 xH2 O ! Hx ReO3 xHReO4 [13]
BuLi
WO3 Li0.75WO3
Figure 3 Reaction with BuLi inserts lithium into four-coordinate sites
within the structure of WO3.
Larger levels of hydrogen insertion are achieved using the
spill-over method.76 In this process, small particles of plati-
num are added to the sample. On exposure to hydrogen gas at
room temperature, these particles catalyze the cleavage of the
HH bond in H2 and so act as a source of atomic hydrogen,
which is then inserted into the ReO3 lattice. Using this tech-
nique, compositions in the range HxReO3 (0 < x < 1.36) have
been prepared. After auto-ionization, the hydrogen is incorpo-
rated into the structure of ReO3 as hydroxide ions, which drive
a significant cooperative tilting distortion of the Re(O,OH)6
octahedra.76 Nuclear magnetic resonance (NMR) studies at
low temperature suggest that the inserted protons are mobile
at temperatures above T > 235 K.77
Hydrogen can also be readily inserted into WO3 using
zinc in hydrochloric acid as a reducing agent according to reac-
tion [14].78,79 Compositions in the range HxWO3 (0 < x < 0.6)
can be prepared in this way, which can be considered as mixed-
oxide/hydroxide phases like HxReO3:
x x
WO3 Zn xHCl ! Hx WO3 ZnCl2 [14]
2 2
2.15.4.1.1.2 MO2 rutile-type hosts
The structure of the rutile polymorph of TiO2 is widely adopted
by the dioxides of other tetravalent transition metals. The
structure can be described as a hexagonally close-packed array
of oxide ions with the M4 transition-metal cations occupying
half of the octahedral coordination sites in this framework to
form an edge-sharing network of MO6 octahedra (Figure 4).5
While the rutile structure is based on close packing, it still
contains a large number of available coordination sites, into
which additional cations can be inserted. Most notably there
are still large numbers of vacant octahedral and tetrahedral
sites, which form chains parallel to the crystallographic c-axis
of the structure. Thus, it is no surprise that reaction with
TiO2 rutile
LiMoO2 Cation ordered NiAs
Figure 4 The rutile polymorph of TiO2. Lithiation of MoO2 inserts
lithium into octahedral sites to form a cation-ordered variant of the NiAs
structure.

n-butyllithium leads to the reductive intercalation of lithium
into MO2 (M Mo, Ru, Os, Ir) rutile phases.80 While initial
reports indicated that lithiation was possible up to composi-
tion of LixMO2 (x 1.31.5) for these host phases, in subse-
quent structural studies only limited compositions with x < 1
are reported. Detailed structural analysis of LiMoO281 and
Li0.9RuO282 reveals that the lithium in these materials is ac-
commodated exclusively within octahedral coordination sites,
and that at a composition of LiMO2 the resulting structure can
be related to a cation-ordered variant of the NiAs structure
shown in Figure 4.
Lithiation of rutile-type WO2 cannot be achieved using
chemical intercalating agents; however, lithium insertion is
possible electrochemically,80 suggesting that the chemical
agents employed are not sufficiently reducing. The rutile
phases of VO2, NbO2, and ReO2 however show no reactivity
either chemically or electrochemically to lithium intercalation.
It has been observed that the rutile phases which do undergo
lithium-insertion reactions are all metallic. This has prompted
the hypothesis that metallic conductivity is a necessary prereq-
uisite for reactivity, as a metallic host lattice can readily trans-
port the reducing electron and also efficiently screen the
polarizing influence of the Li cations. However, lithiation-
insertion reactions performed at temperatures above the
metalinsulator transition of VO2 (67  C) did not incorporate
lithium either, suggesting that metallic conductivity may be a
necessary but not sufficient condition of lithium intercalation
into rutile phases.
TiO2, MnO2, and CrO2 exhibit only modest lithiation.
Reaction with n-butyl lithium yields limiting compositions of
Li0.3TiO2, Li0.2MnO2, and Li0.8CrO2. It has been postulated
that the smaller size of these first-row transition-metal cations
leads to smaller, less flexible host structures which cannot
accommodate the lattice expansions and distortions which
occur on cation insertion.
2.15.4.1.1.3 AB2O4 spinel-type hosts
In contrast to the ReO3 and rutile host lattices, which contain
transition metals in high oxidation states, compounds which
adopt structures related to that of the mineral spinel, MgAl2O4,
typically contain metals in lower oxidation states consistent
with the increased degree of face-sharing between MOx coor-
dination polyhedra in this structure type.5
The spinel structure is rather complex and based on close
packing (Figure 5). Described in the face-centered cubic (FCC)
space group Fd3m,  the structure consists of a cubic close-
packed array of oxide ions with Al3 cations occupying 1/2 of
the 32 octahedral interstitial sites (16 per cell, assigned the
symmetry label 16d) and Mg2 cations occupying 1/8 of the
64 tetrahedral sites (eight per cell, symmetry 8a) in each unit
cell. Therefore, it follows that in each unit cell there are 16
empty octahedral sites (16c) and 56 empty tetrahedral sites
(which are split on symmetry grounds into two sets 8b and
48f) into which small cations can be inserted.
A large variety of oxide phases adopt spinel structures in-
cluding the binary oxides Mn3O4, Fe3O4, and Co3O4. Fe3O4
adopts the inverse spinel structure in which, due to crystal field
effects, FeIII centers are located on the 8a tetrahedral sites and a
1:1 mixture of FeII and FeIII centers is located on the 16d
octahedral sites. Reaction with n-butyllithium readily inserts
Soft Chemistry Synthesis of Oxides 425
MgAl2O4 spinel
16d octahedral sites 8a tetrahedral sites
unoccupied unoccupied 8b and 48f
16c octahedral sites tetrahedral sites
Figure 5 The occupied and unoccupied cation-coordination sites in the
MgAl2O4 spinel structure.
lithium into the spinel host phase, up to a limiting composi-
tion of Li1.3Fe3O4, while electrochemical insertion can add
further lithium to produce the Fe(II)-phase Li2Fe3O4.83
The lithium-insertion process in spinels is not entirely topo-
chemical. Lithium is inserted into the 16c octahedral coordi-
nation sites in Fe3O4. Associated with this insertion, there is a
migration of iron cations from the 8a tetrahedral sites to the
16c octahedral sites within the host lattice. As a result, mate-
rials of composition LiFe3O4 have a partially ordered rock salt
structure with iron cations on the 16d sites and a 1:1 disor-
dered array of lithium and iron cations on the 16c sites.83 Thus,
lithiation proceeds according to reaction [15]. It should be
noted that the [Fe2]16dO4 framework is retained throughout
the insertion process:
 3  h 2=3 i
Fe 8a Fe2 O4 xLi
  16dh i
! Lix Fe 2=3
Fe2 2=3 O4 [15]
16c 16d
Further lithiation to form phases of composition LixFe3O4
(1 < x < 2) proceeds via insertion of lithium ions into the 8b
and 48f tetrahedral coordination sites, to yield materials with
both octahedrally and tetrahedrally coordinated lithium cat-
ions.83 Lithium insertion into Mn3O484 and Co3O485 proceeds
via a similar mechanism.
Ternary oxides such as LiM2O4 (M Ti, V, Mn) also adopt
spinel structures in which the transition-metal cations are ac-
commodated within the 16d octahedral sites and the small
lithium cations in the 8a tetrahedral sites. The lithium-
insertion mechanism adopted by these phases is dependent
on the identity of the transition-metal ion. Lithiation of
LiTi2O4 follows a mechanism similar to that of the M3O4
(M Mn, Fe, Co) phases. Lithium insertion into the octahedral
426 Soft Chemistry Synthesis of Oxides
16c sites is accompanied by cation migration from the tetrahe-
dral 8a to the octahedral 16c coordination sites.86 As a result,
Li1xTi2O4 (x > 0) phases adopt cation-ordered rock salt-like
structures. In contrast, lithium insertion into LiMn2O4 occurs
without cation migration. Thus, Li2Mn2O4 contains lithium
cations in both 8a tetrahedral and 16c octahedral coordina-
tion sites84,87 despite the potential for unfavorable Columbic
interactions between lithium cations in these two coordina-
tion sites.
LiV2O4 exhibits a lithiation mechanism which is interme-
diate between that of LiTi2O4 and LiMn2O4. At low levels of
lithiation, cations are inserted into the 16d octahedral sites of
Li1xV2O4 (0 < x < 0.5) with no accompanying migration.88
However as the lithium content rises above Li1.5V2O4, cation
migration does occur to yield a cation-ordered rock salt struc-
ture for the V(III)-phase Li2V2O4.
The cation migration observed during the lithiation of spi-
nel phases can be attributed to structural features of the host
lattice. As noted above, cations within the ReO3 and rutile
structures are in high oxidation states and, as a result, are
tightly bound within octahedral coordination sites so the acti-
vation energy for cation migration is extremely high. In
contrast, the cations located within the tetrahedral sites of the
spinel structure are in lower oxidation states and are located
within coordination polyhedra which share faces with empty
interstitial sites. Consequently, the barriers to cation migration
are lower for the tetrahedral cations, and these cations can
move through the host lattice at room temperature. Thus, we
can see a general synthetic strategy for preparing metastable
materials by utilizing the facile migration of some of the cat-
ions within a structure, in combination with reductive interca-
lation, to induce cation order into extended oxide phases.
2.15.4.1.2 Intercalation into layered binary oxide phases
In addition to the framework oxide hosts described above, it is
also possible to reductively intercalate cations into layered oxide
host phases to facilitate the preparation of metastable phases.
Binary oxide phases of transition metals in high oxidation states
can adopt layered structures consisting of sheets of linked MOx
polyhedra. These sheets are then stacked within the material. As a
result, there is a strong network of covalent metaloxygen bonds
within each neutrally charged metal-oxide layer. However, the
only interlayer interactions present are weak van der Waals forces
which are destabilized by electrostatic repulsions between the
negatively charged oxide ions. As a result, the intercalation of
cationic species into these van der Waals gaps is favored.
2.15.4.1.2.1 a-MoO3
a-MoO3, which is typical of this type of host phase, has a
structure consisting of MoIVO6 octahedra which are linked
through edges and corners into charge neutral sheets, which
stack as shown in Figure 6.5 Reaction of a-MoO3 with sodium
dithionite (a reducing agent) leads to the reductive intercala-
tion of hydrated sodium ions according to reaction [16]89:
 
Na2 S2 O4 MoO3 yH2 O ! Na H2 On x MoO3
xSO2 [16]
Diffraction studies reveal that the hydrated sodium ions are
inserted topochemically into the van der Waals gap of a-MoO3
Na2S2O4/H2O
a-MoO3
[Na(H2O)n]xMoO3
Figure 6 Reaction with sodium dithionite reductively intercalates
hydrated sodium ions into a-MoO3.
as demonstrated by the large expansion in the interlayer spacing
on intercalation as shown in Figure 6. Lithium, potassium, ru-
bidium, magnesium, and calcium analogs can also be prepared,
either by cation exchange as described in Section 2.15.4.5.1.2, or
by electrochemical insertion from solutions of the corresponding
cations. Intercalation of unsolvated cations can be achieved via
electrochemical reduction reactions in suitable nonaqueous sol-
vents such as dimethoxyethane or dimethylsulfoxide.89
The layered nature of the structure of a-MoO3 means that
large cations can be inserted into this phase. By electrochemical
reduction, large organic cations such as methylammonium or
propylpyridinium can also be inserted to form layered inor-
ganic/organic hybrid phases.90 The reductive nature of the
insertion reactions is demonstrated by a change in the color
of the material from white to dark metallic blue, which is
accompanied by a large increase in electrical conductivity. All
intercalated phases react with strong oxidizing agents such as
MnO4 and NO2 to recover unintercalated a-MoO3.89
The hydrogen-insertion chemistry of a-MoO3 is extensive.
HxMoO3 (0 < x < 2) phases can be readily prepared by reaction
with Zn/HCl,91,92 or by electrochemical insertion.89 H2MoO3
has a deep green color, with less hydrogenated samples exhi-
biting deep red (1.55 < x < 1.7) and deep blue (0.93 < x < 0.28)
colors.91 Intercalated phases are good metallic conductors and
exhibit temperature-independent paramagnetism and no elec-
tron spin resonance (ESR) signal, consistent with the insertion
of the reducing electrons into the d-bands of the MoO3 host,
rather than residing in localized states.93,94
Detailed structural studies show that the insertion of hydro-
gen into a-MoO3 is topochemical. At low concentrations, the
inserted protons bind to oxide ions which bridge between
molybdenum centers within the metal-oxide layers.95 At higher
concentration, the protons are bound as OH2 units to the
terminal oxide ions on the outside of the metal-oxide sheets.96
The shift in binding site is attributed to changes in MoMo
bonding as the level of hydrogen in the phases increases. NMR
studies indicate that HxMoO3 phases with x  1.7 exhibit pro-
ton conductivity at room temperature,97 while phases with
lower hydrogen concentrations do not. This is consistent with
a proton-transport mechanism which involves hopping within
the van der Waals gap of the a-MoO3 host phase.

2.15.4.1.2.2 V2O5
V2O5 can be considered intermediate between the framework
oxides such as ReO3 and RuO2 and the layered oxides such as
a-MoO3. The structure of V2O5 can be considered either as
an array of square-based pyramidal VO5 units which share
edges and corners to form a layered 2D structure, or as a
three-dimensionally connected network of VO6 octahedra
(Figure 7).5 The different structural views depend on whether
the extended VO interlayer interaction is considered to be
within the coordination sphere of vanadium or not. Regardless
of the semantics of the bonding in V2O5, it is clear that the
extended VO contact has a significant influence on the inter-
calation chemistry of V2O5 as it prevents the delamination of
the structure and thus the intercalation of large cations. As a
result, V2O5 behaves much like the framework host structures
and only intercalates small cations.
Lithium can be intercalated into V2O5 either electrochemi-
cally or via reaction with n-butyllithium or lithium iodide.98
Phases in the range LixV2O5 (0 < x < 1) are crystalline, while
those with x > 1 become amorphous. The phase diagram of
LixV2O5 contains a number of different phases in which the Li
cations are intercalated into coordination sites within the
van der Waals gap of the host material, resulting in subtle
distortions to the host lattice.99 Despite these distortions, the
interlayer VO bond, which makes the vanadium centers
pseudo-octahedral, is retained. However, if V2O5 is reacted
with an excess of lithium iodide at room temperature, d-
LiV2O5 is formed. In this material the vanadium oxide sheets
shift relative to each other to accommodate the lithium cations
in more regular tetrahedral sites.100 This breaks the interlayer
VO bond. d-LiV2O5 has a rather small compositional stability
range and only forms at ambient temperature, suggesting that
it is a kinetically stabilized phase.
V2O5 can readily intercalate large amounts of hydrogen via
the spill-over route.101 Extended reaction at room temperature
results in the preparation of H3.8V2O5, a black material which
has an amorphous structure and contains vanadium in an
average oxidation state of V 3.1.102 Lower levels of hydroge-
nation (0 < x < 0.5) yield a series of green materials which are
isostructural with the V2O5 host phase and exhibit semicon-
ducting behavior.103
2.15.4.1.3 Intercalation into layered perovskite structures
There are a number of structure types which are commonly
adopted by complex oxide phases, which can be considered
as layered variants of the ABO3 cubic perovskite structure
(Chapter 2.03). Among the most common are the
An1BnO3n1 RuddlesdenPopper structures. As shown in
2D structure 3D structure
of V2O5 of V2O5
Figure 7 The 2D and 3D representations of the structure of V2O5.
Soft Chemistry Synthesis of Oxides 427
Figure 8, this family of structures can be thought of as a regular
stacked intergrowth of (ABO3) perovskite blocks and (AO) rock
salt layers. Thus, the value of n in the compositional formula
(AO)(ABO3)n An1BnO3n1 indicates the number of perov-
skite sheets which are stacked between adjacent AO rock salt
layers. Compounds which exhibit n 1 or n 2 structures are
common, while the materials with larger values of n are less so. It
can be seen in Figure 8 that RuddlesdenPopper structures with
n  2 exhibit two distinct A-cation-coordination sites: a 12-
coordinate site analogous to that in cubic perovskite phases
and a nine-coordinate site which resides within the rock salt
layers of the structure. Typically, alkaline earth (Ca, Sr, Ba) or
lanthanide cations can be accommodated within the A-cation
sites of RuddlesdenPopper phases, with ordering between the
9- and 12-coordinate sites possible in mixed cation systems
where the two A-cations have significantly different charges
and/or radii.
Layered perovskite phases which contain large alkali metals
(K, Rb, Cs) tend to adopt A0 An1BnO3n1 DionJacobson struc-
tures. In this structural family, the large monovalent A0 -cations
reside between the perovskite blocks in eight-coordinate
pseudo-cubic coordination sites (Figure 9). It should be
noted that the low formal charge on the A0 -cations means
that in order to charge-balance the A0 An1BnO3n1 composition
of DionJacobson phases, the B-cations must be tetravalent or
pentavalent transition metals. Such high oxidation states ap-
pear to provide a good electrochemical driving force for reduc-
tive cation-insertion reactions.
2.15.4.1.3.1 DionJacobson phases
While it may appear that there are no obvious vacant coordi-
nation sites suitable to host intercalated cations within the
DionJacobson family of structures, it has been demonstrated
that lithium can be inserted into A0 Ca2Nb3O10 (A0 K, Rb, Cs)
or A0 Sr2Nb3O10 phases by exposure to n-butyllithium, in a very
slow process that can take several weeks, or by electrochemical
reduction which is significantly faster.104,105 The structures of
the resulting LixA0 A2Nb3O10 (0 < x < 0.6) phases are not well
studied; however, it is clear that the only significant change
to x-ray diffraction data collected after intercalation is an ex-
pansion of the c-lattice parameter, suggesting that the lithium
cations are inserted, in a topochemical manner, into the
9 CN
12 CN
ABO3
n=
A2BO4
n=1
A3B2O7
n=2
Figure 8 The An1BnO3n1 RuddlesdenPopper phases.
428 Soft Chemistry Synthesis of Oxides
AAB2O7
n=2
AA2B3O10
n=3
Figure 9 The AAn1BnO3n1 DionJacobson phases.
A0 -cation layers which lie between the perovskite blocks in the
DionJacobson phases.106
On lithium intercalation, the LixA0 A2Nb3O10 materials
change from white to black, consistent with the reduction of
Nb(V). Magnetization data indicate that these phases are su-
perconductors at low temperature,104 with the highest Tc 7 K
observed for Li0.1CsSr2Nb3O10.107 Detailed studies suggest that
the materials are only superconductors over a small composi-
tional range (0.1  x  0.15) and that there is a correlation
between the in-plane lattice parameter of a LixA0 A2Nb3O10
phase and the maximum superconducting transition tempera-
ture observed for that material.
Intercalation of larger cations into DionJacobson phases
drives a structural change. On heating RbLaNb2O7 in the pres-
ence of a controlled amount of rubidium vapor, Rb2LaNb2O7
is formed.108 Associated with the insertion of rubidium, there
is a change in the stacking of the perovskite blocks from an
A0 AB2O7 DionJacobson arrangement in which the A0 rubid-
ium cations are located in eight-coordinate pseudo-cubic sites,
to an A3B2O7 RuddlesdenPopper-type stacking arrangement
in which the two rubidium cations are located in nine-
coordinate sites within the rock salt layer of the structure and
adjacent perovskite blocks have a staggered relationship to
each other (Figure 10). Similarly, exposure of CsLaNb2O7 to
cesium vapor induces an analogous reductive intercalation of
cesium to yield the n 2 RuddlesdenPopper-phase
Cs2LaNb2O7.109
A similar transformation is observed on insertion of sodium
into the n 3 DionJacobson-phase RbCa2Nb3O10.110 Heating
in sodium vapor yields RbNa0.83Ca2Nb3O10 which adopts a
structure in which adjacent perovskite blocks are arranged in a
staggered manner consistent with a RuddlesdenPopper-type
structure. The sodium and rubidium cations in the rock salt
layers of this phase are situated in highly distorted coordination
Rb
RbLaNb2O7 Rb2LaNb2O7
Figure 10 Reductive intercalation of rubidium into the DionJacobson-
phase RbLaNb2O7 drives a structural change to for the Ruddlesden
Popper-phase Rb2LaNb2O7.
sites, presumably due to the large mismatch in ionic radii be-
tween Na and Rb, which drives a significant buckling of the
perovskite framework.
The reactions above demonstrate that even though there are
no obvious intercalation sites for large cations within Dion
Jacobson structures, a small shift in the relative position of the
perovskite layers can create suitable sites. This change in the
perovskite stacking sequence, and the ready intercalation of
large cationic species, demonstrates the relatively weak binding
between the A0 -cations and the perovskite blocks in Dion
Jacobson-type structures and is somewhat reminiscent of the
layered oxides MoO3 and V2O5. Indeed, this weak interlayer
coupling in DionJacobson phases is often exploited in
their extensive cation-substitution chemistry described in
Section 2.15.4.5.1.3.
2.15.4.1.3.2 Cation-deficient RuddlesdenPopper Phases
Cation-deficient RuddlesdenPopper phases can be pre-
pared by exploiting the extensive cation-substitution chemistry
of materials which adopt DionJacobson-type structures
(Section 2.15.4.5.1.3). Thus, the reaction of RbCa2Nb3O10
with LiNO3 or NaNO3 yields the cation-deficient Ruddlesden
Popper phases ACa2Ta3O10 (A Li, Na) in which the small
alkali metal cations are accommodated within partially filled
tetrahedral coordination sites between the triple-layer perov-
skite blocks (Figure 11).111113 Reaction of LiCa2Ta3O10 with
n-butyllithium leads to the reductive intercalation of lithium,
to completely fill the remaining tetrahedral coordination sites
and form Li2Ca2Ta3O10, in which the lithium cations adopt an
anti-fluorite-type arrangement between the apical oxide ions of
the perovskite sheets (Figure 11).111 In contrast, the reaction of
NaCa2Ta3O10 with NaN3 inserts sodium into the oxide host and
drives a migration of sodium cations from tetrahedral to nine-
coordinate sites, so that the resulting phase, Na2Ca2Ta3O10,
adopts a conventional RuddlesdenPopper structure with so-
dium cations located within rock salt layers (Figure 11).112 The
intercalated ACa2Ta3O10 phases exhibit a deep blue color and
Pauli paramagnetic responses, consistent with metallic behavior.
Similar lithium intercalation reactions have been performed on

BuLi
Li2Ca2Ta3O10
Figure 11 Reductive intercalation of ACa2Ta3O10 (A Li, Na) yields phases with structures that depend on the size of the intercalated cation.
the cation-deficient n 2 RuddlesdenPopper phases LiLaNb2O7
and LiLaTa2O7.111,114
Cation-deficient RuddlesdenPopper phases can also be
prepared by direct, redox-neutral, cation exchange in which
one divalent cation replaces two monovalent cations, to create
a cation vacancy (Section 2.15.4.5.1.4). Thus, reaction of the
A-cation-ordered RuddlesdenPopper-phase NaLaTiO4 with
Ca(NO3)2 yields Na0.13Ca0.430.47TiO4 (where represents
a cation vacancy). On reaction with sodium vapor,
Na0.51Ca0.53LaTiO4, a phase with an average titanium oxida-
tion state of Ti 3.63, is formed.115
2.15.4.1.3.3 Aurivillius phases
Bi2An1BnO3n3 Aurivillius phases can be described as layered
intergrowths between (Bi2O2)2 blocks with fluorite-like struc-
tures and (An1BnO3n1) perovskite blocks. It has been dem-
onstrated that despite the apparently close-packed nature of
this structural family, lithium can be inserted into the Bi2O2
fluorite layers with concomitant reduction of the transition-
metal centers. Thus, the reaction of Bi4Ti3O10 with n-butyl-
lithium yields Li2Bi4Ti3O10. Neutron diffraction data indicate
that the lithium cations reside within pseudo-square-planar
coordination sites between the Bi2O2 blocks and perovskite
layers.116
2.15.4.2 Cation Deintercalation Reactions
The cation deintercalation chemistry of complex transition-
metal oxides which contain small monovalent cations (Li,
Na) has been widely studied as this is the chemical transfor-
mation which occurs during the charging of lithium-ion batte-
ries (Chapter 4.03). From the perspective of soft chemical
synthesis, low-temperature cation deintercalation reactions
allow the preparation of oxidized, metastable complex oxide
phases which exhibit metaloxygen frameworks which cannot
be prepared directly by ceramic routes.
Soft Chemistry Synthesis of Oxides 429
50%
occupied
NaN3
ACa2Ta3O10 Na2Ca2Ta3O10
In order to effectively deintercalate cations from metal-
oxide host lattices in a topochemical manner, a number of
conditions need to be met.
1. There must be suitable highly mobile cations present in the
parent phase and their mobility must be significantly
higher than that of the host lattice. This generally restricts
this type of reaction to phases which contain monovalent
cations, typically lithium, sodium, or silver.
2. The removal of cations must leave a kinetically stable phase.
Deintercalation of cations from a complex oxide phase can
lead to unscreened, unfavorable Columbic interactions be-
tween oxide ions in the host lattice. These interactions can
limit the level of cation deintercalation which can be achieved.
3. The deintercalation of cations formally oxidizes the host
phase as the metal cations are removed as neutral atoms.
Thus, the host phase must contain readily oxidizable metal
centers, and the oxidizing agent, or electrochemical driving
force, used must be of sufficient strength to bring about the
reaction.
2.15.4.2.1 a-NaFeO2-type phases
2.15.4.2.1.1 LiCoO2
LiCoO2 is a thermodynamically stable phase which is readily
prepared by heating a combination of Li2CoO3 and Co3O4
under flowing oxygen. The resulting Co(III) oxide adopts a
structure analogous to a-NaFeO2 which can be considered as
a cation-ordered variant of the NaCl rock salt structure in
which Co3 and Li cations are arranged into layers of edge-
sharing MO6 octahedra within the distorted FCC oxide ion
lattice as shown in Figure 12. The layered nature of this struc-
ture, and the large number of additional interstitial sites avail-
able to the lithium cations within each layer, lead to good
lithium-ion mobility. This high mobility can be utilized to
bring about the oxidative deintercalation of lithium ions
to form phases of composition Li1xCoO2.117 The degree of
deintercalation which can be achieved depends on the strength
of the oxidant used, as the deintercalation of lithium is
430 Soft Chemistry Synthesis of Oxides

accompanied by the oxidation of Co3 centers to Co4.118 This
is exemplified by the reactions with the halogens as shown in
reaction [17]:
2LiCoO2 xX2 ! 2Li1x CoO2 2xLiX [17]
x
LiCoO2 xH ! Li1x CoO2 xLi H2 [18]
2
Reaction with I2, Br2, or Cl2 yields Li1xCoO2 phases with
x 0.09, 0.53, and 0.69 respectively, broadly in line with the
electrochemical reduction potentials of the halogens.118
Alternatively, strong acids such as HCl or H2SO4 can be used
to deintercalate lithium from LiCoO2. Treating single crystals
of LiCoO2 with HCl for 7 days yields Li0.35CoO2 as shown in
reaction [18].119 In contrast, if powder samples of LiCoO2 are
treated with H2SO4, delithiation occurs via a combination of
disproportionation and proton-exchange processes as shown
in reactions [19] and [20], respectively118:
LiCoO2 2xH2 SO4 ! Li12x CoO22x xLi2 SO4
xCoSO4 2xH2 O [19]
The deintercalation of lithium from LiCoO2 is strictly topo-
chemical in the composition range 0.12  x  1. Removal of
lithium ions leads to vacancies within the planes of lithium
cations (Figure 12) which adopt a disordered arrangement
over most of the compositional range. Magnetic anomalies
provide some evidence for lithium vacancy ordering in com-
positions close to x 0.7; however, these ordered states appear
fragile and are easily disrupted by small compositional changes
and so are as yet uncharacterized.122,123 Compositions close to
x 0.5 however exhibit robust 1:1 ordering of lithium and
lithium vacancies which is associated with Co(III)/Co(IV)
charge order as described below.119,124
Complete removal of all the lithium ions from LiCoO2
drives a change to the structure, in which the [CoO2] planes
shift relative to each other such that the Co(IV)-phase CoO2
adopts the CdI2 structure as shown in Figure 12.120 At compo-
sitions close to x 0.12, a hybrid structure is adopted which
includes both the cubic and the hexagonal stacking of [CoO2]
planes.125,126

LiCoO2 0:5xH2 SO4 ! Li1x Hx CoO22x 2.15.4.2.1.2 NaCoO2

0:5xLi2 SO4
Complete deintercalation of lithium can be achieved elec-
trochemically.120,121 The resulting metastable Co(IV)-phase,
CoO2, is highly oxidizing, reacting rapidly with moisture
to form CoOOH according to eqn [21].120 However if mate-
rials are kept under appropriate conditions, a combination
of chemical and electrochemical oxidation allows the entire
Li1xCoO2 (0  x  1) composition range to be prepared:
4CoO2 2H2 O ! 4CoOOH O2
[20] Similar deintercalation reactions can be performed on the
analogous layered NaxCoO2 phases. However in this instance
the situation is more complex because there are four different
thermodynamically stable layered NaxCoO2 phases which can
be prepared by conventional high-temperature synthesis
routes: the a phase (0.9  x 1) which adopts the O3 structure
of NaFeO2, the a0 phase (x 0.75) which exhibits an O1 struc-
ture, the g phase (x  0.7) which adopts a P2 structure, and the
b phase (0.55  x  0.60) which adopts a P1 structure as shown
[21] in Figure 13.127,128

B B
A A
Lithium
C B
B Extraction
A
A B
C A
B B
A A
LiCoO2 CoO2

Figure 12 Extraction of lithium drives a change from the ABCABC cubic stacking present in LiCoO2 to ABAB hexagonal stacking in CoO2.

B A B
A B A

C B A
B A C
A A C
C B B
B B B
A A A
a-Nax CoO2 g-Nax CoO2 b-Nax CoO2
(0.9 < x <1) (x ~ 0.7) (0.55 < x < 0.66)
O3 P2 P1
Figure 13 The thermodynamically stable structures of NaxCoO2.

Oxidative deintercalation reactions performed on the a
phase of NaxCoO2, which like LiCoO2 adopts the a-NaFeO2
structure, reveal that the sodium phases are more easily oxi-
dized than the lithium-containing analogs. Thus oxidation of
Na0.92CoO2 with excess iodine or bromine in acetonitrile
yields Na0.5CoO2 and Na0.3CoO2, respectively, compared
with Li0.81CoO2 and Li0.47CoO2 as described above.118,128
Na0.3CoO2 prepared in this way retains the O3 structure of
the a-NaxCoO2 parent phase. In contrast, Na0.5CoO2 prepared
by the oxidative deintercalation of Na0.92CoO2 adopts the P1-
type structure of thermodynamically stable b-NaxCoO2
(Figure 13).128 Given that Na0.5CoO2 has a composition
which is very similar to the stability range of b-NaxCoO2, this
suggests that the change in [CoO2] layer stacking observed on
chemical oxidation of a-type Na0.92CoO2 to b-type Na0.5CoO2
is under thermodynamic control to some extent. Similar
changes to the stacking sequence of the [CoO2] sheets are ob-
served during the electrochemical oxidation of a-NaxCoO2.129
The intercalation/deintercalation chemistry of the double-
layer g-phase Na0.7CoO2 has been very extensively studied.
Sodium can be readily deintercalated from this phase using
bromine or iodine in a manner similar to that employed on
other NaxCoO2 compounds.130 Sodium concentrations as low
as Na0.34CoO2 are also achievable by employing NO2PF6 as an
oxidant. Additional sodium has been intercalated into g-
Na0.7CoO2 by gentle heating in tetrahydrofuran in which ele-
mental sodium and benzophenone had been dissolved. Thus
by utilizing both oxidative deintercalation and reductive inter-
calation 2-layer, g-type phases can be prepared over the com-
position range NaxCoO2 (0.34  x  0.89).130
In contrast to the NaxCoO2 phases derived from a-
Na0.92CoO2, products of the low-temperature intercalation/
deintercalation of g-Na0.7CoO2 are isostructural with respect
to the parent phase in that the [CoO2] layers retain their
AABBAA stacking sequence across the whole available compo-
sition range.130 This suggests that this stacking sequence is
kinetically stabilized under the reaction conditions employed.
This kinetic stability of the P2 structure of g-type NaxCoO2 is
also observed during electrochemical oxidation.129
One structural feature, which does change across the g-
NaxCoO2 compositional range, is the coordination site occu-
pied by the sodium ions. The sodium cations are located in
prismatic coordination sites within the 2P structure of g-Nax-
CoO2. There are two crystallographically distinct prismatic
coordination sites created by the stacking of the [CoO2] layers.
The distribution of sodium between these two sites changes
sharply as a function of sodium content, suggesting that the
configuration of the sodium layers is coupled to the complex
electronic configuration of the triangular CoO2 lattice.130
2.15.4.2.1.3 Physical properties of A1xCoO2
The extensive investigation of the cation deintercalation chem-
istry of AxCoO2 phases described above has been motivated by
the useful and exotic properties they exhibit. In addition to the
widespread utilization of LixCoO2 phases as cathode materials
in secondary lithium-ion batteries, AxCoO2 phases have re-
ceived significant attention because the cobalt centers in the
[CoO2] sheets within these materials are arranged into trian-
gular lattices. This unusual lattice symmetry, combined with
Soft Chemistry Synthesis of Oxides 431
the ability to change the electron count (oxidation state) of the
cobalt centers directly by simply changing the alkali metal
concentration,131,132 makes the AxCoO2-layered cobaltates an
ideal set of materials for the study of correlated electronic
interactions.
Taking LixCoO2 and the different polymorphs of NaxCoO2
as a whole, it has been observed that for phases with high alkali
metal concentrations, where Co3 is the predominant cobalt
oxidation state (x > 0.5), an unusual CurieWeiss metallic
state is observed.122,123,130,132,133 Samples exhibit metallic
conductivity (resistivities of the order of 1 mO cm) and simul-
taneously have magnetic responses which are indicative of
localized electronic behavior. In the specific case of NaxCoO2
(x  0.7), this combination of localized magnetic spins and
metallic conductivity is thought to be responsible for the
highly enhanced thermopower observed in the materials
(100 mV K1 at 300 K).134,135 Lowering the alkali metal con-
tent to x 0.5 induces cation order within the vacant sheets of
lithium and sodium cations.119,136 This is rather unexpected as
a 1:1 order of cations and vacancies appears to be incompatible
with the threefold symmetry of the [CoO2] sheets. Detailed
studies, particularly of g-type Na0.5CoO2, have revealed that
the observed cation order is assisted by partial Co(III)/Co(IV)
charge order which breaks the triangular symmetry of the
CoO2 planes and leads to insulating behavior at this composi-
tion.136 The interplay between the electronic configuration of
the [CoO2] sheets and the dispositions of the sodium cations
further confirms that these two features of the material are
strongly coupled.
As the alkali metal concentration declines further, the
CurieWeiss-like magnetic behavior diminishes, so that for
AxCoO2 phases with x < 0.5 temperature-independent paramag-
netism more typical of metallic systems is observed.123,130,133
However, the most striking physical behavior in the AxCoO2
system appears after a further piece of soft chemistry. If
two-layer g-type NaxCoO2 phases with x  0.3 are exposed to
water, they readily absorb water molecules into the partially
occupied sodium layers to form phases of approximate com-
position Na0.3CoO2 1.3H2O, with greatly expanded interlayer
separations as shown in Figure 14. The resulting material,
when cooled below 4.5 K, exhibits bulk superconductivity.137
Studies have shown that the presence of water in this phase
is critical for the appearance of superconductivity as even
partial dehydration to a composition of Na0.3CoO2 0.6H2O
is sufficient to suppress the superconducting state.138 However,
the exact role of water in inducing superconductivity into
Na0.3CoO2 is not clear. The large increase in separation be-
tween the [CoO2] layers in superconducting Na0.3CoO2
1.3H2O (9.8 A)137 compared with nonsuperconducting
Na0.3CoO20.6H2O (6.9 A)138 and Na0.3CoO2 (5.6 A),130 and
the observation that the superconducting critical temperature,
Tc, decreases under pressure,139 suggest the isolation of the
[CoO2] sheets, and the more 2D character of the superconduct-
ing phases is important. Preparation of hydrated phases with a
range of sodium contents reveals that superconductivity is only
supported over a small compositional range (NaxCoO2
1.3H2O (0.25  x  0.35)) with a maximum Tc at x  0.3 and
a dome like dependence of Tc on the [CoO2] layer electron
count.140
432 Soft Chemistry Synthesis of Oxides
H2O
5.6
g-Na0.3CoO2
Figure 14 g-Na0.3CoO2 intercalates water to form the superconducting phase g-Na0.3CoO21.3H2O.
2.15.4.2.1.4 AxNiO2
Analogous deintercalation reactions can be performed on lay-
ered AxNiO2 (A Li, Na) phases which also adopt a-NaFeO2-
type structures. Compared to the oxidations of the AxCoO2
materials, the oxidation of AxNiO2 phases requires more
powerful oxidizing agents, consistent with the greater electro-
negativity of nickel compared to cobalt. Partial delithiation
can be achieved by reaction with H2SO4 via a disproportion-
ation route to generate Li1xNiO2 and Ni2 as shown in
reaction [22]141:
LiNiO2 4yH ! 1  yLi12y=1y NiO2 yNi 2
2yLi 2yH2 O [22]
Complete delithiation of LiNiO2 to form NiO2 is possi-
ble electrochemically.142 From a structural perspective, the
delithiation of LiNiO2 is complicated by the possibility of
Li1zNi1zO2 nonstoichiometry which substitutes some of
the lithium centers with nickel ions. In the absence of this
nonstoichiometry, lithium deintercalation proceeds in a topo-
chemical manner until a low lithium concentration is
achieved. There is then a structural reorganization, like that
observed during the delithiation of LiCoO2, such that the
Ni(IV)-phase NiO2 adopts a hexagonal, CdI2-type structure.143
However in samples with appreciable nonstoichiometry
(Li1zNi1zO2 (z > 0.07)), the presence of nickel ions on inter-
layer coordination sites prevents the shift of [NiO2] sheets. As a
result, the oxidative deintercalation is topochemical over the
entire composition range. NiO2 prepared from such a sample
has a CdCl2-type structure in which the cubic stacking of the
oxide ions in LiNiO2 is retained.142 These structural complica-
tions and the highly oxidizing nature of AxNiO2 phases have to
date prevented a detailed study of their physical properties.
2.15.4.2.1.5 LiVO2
LiVO2 also adopts an a-NaFeO2-type ordered rock salt struc-
ture. Reaction with bromine in chloromethane leads to the
oxidative deintercalation of lithium and the formation of
cation-deficient Li1xVO2 phases over the compositional
range 0  x  1.144 In contrast to the cobaltates and nickelates
described above, the deintercalation reactions of LiVO2 are not
topochemical.145,146 In Li1xVO2 phases where x > 0.3, vana-
dium cations are observed to diffuse into octahedral sites made
vacant by the removal of lithium, to help stabilize the structure
of the cation-deficient phase. As a result, a significant degree of
disorder is introduced into the layered, ordered rock salt
9.8
g-Na0.3CoO2.1.3H2O
structure of the LiVO2 parent phase. For example, it is observed
that Li1xVO2 phases with x  0.7 have one-third of the vana-
dium cations located in the wrong cation sites. It is proposed
that vanadium cations diffuse between octahedral coordina-
tion sites via a network of face-sharing tetrahedral sites which
exist within the cubic close-packed lattice of the host phase.
The vanadium diffusion is further facilitated by the dispropor-
tionation of V4 and the preference of V5 for tetrahedral
coordination as shown in reactions [23] and [24]:
2V 4 octahedral ! V3 octahedral V5 tetrahedral [23]
V 3 octahedral V 5 tetrahedral ! 2V 4 octahedral [24]
The metastability of Li1xVO2 phases is exemplified by the
observation that if Li0.5VO2 is heated above 300  C, it trans-
forms to the spinel LiV2O4, with lithium cations located on the
8a tetrahedral sites and vanadium on 16d octahedral sites.147
2.15.4.2.2 Lithium manganese oxides
Manganese dioxide is known to adopt a number of different
polymorphs, the most thermodynamically stable of which is
the mineral pyrolusite (b-MnO2) which has the same structure
as the rutile polymorph of TiO2.5 Reaction of manganese
oxides with lithium leads to the formation of three principal
phases Li2MnO3, LiMn2O4, and LiMnO2 which adopt
different manganeseoxygen frameworks in order to accom-
modate the lithium cations in phases with average manganese
oxidations states of 4, 3.5, and 3, respectively.
All three lithium manganese oxide phases can be delithiated
in acidic solutions. On exposure to acid, Li2MnO3 decomposes
into known polymorphs of MnO2 by what is thought to be a
dissolution/precipitation mechanism.148 However, as described
below, the more reduced lithium manganese oxide phases un-
dergo oxidative lithium deintercalation reactions to form novel
MnO2 frameworks.
LiMn2O4 adopts a spinel-type structure in which lithium
cations are located on the 8c tetrahedral sites and manganese
cations accommodated within 16d octahedral sites (Figure 15).
Reaction with acid brings about the delithiation/disproportion-
ation reaction described in reaction [25] to yield a new poly-
morph of manganese dioxide, l-MnO2.149 The delithiation
reaction is topochemical such that l-MnO2 retains the
manganeseoxygen framework of the spinel parent phase
(Figure 15). The novel MnO2 framework has been observed to
exhibit complex magnetic frustration and good activity as a

LiMn2O4
ic
e m
ch on
t ro ati
ec id
H+ El ox
+ (p
H
l-MnO2
Figure 15 Delithiation of LiMn2O4 and LiMnO2 leads to the formation of new metastable polymorphs of MnO2.
water-oxidation catalyst.150,151 On heating l-MnO2 to 300  C, it
is observed to transform to b-MnO2, demonstrating the meta-
stability of the delithiated phase149:
2LiMn2 O4 4H ! 3MnO2 2Li Mn 2
2H2 O [25]
Orthorhombic LiMnO2 adopts an ordered rock salt struc-
ture consisting of zigzag sheets of edge-sharing MnO6 octa-
hedra, separated by octahedrally coordinated lithium ions
(Figure 15). In common with LiCoO2 (Section 2.15.4.2.1.1),
the delithiation reactions which occur between orthorhombic
LiMnO2 and acid depend on the physical form of the material.
On exposure to acid, a single crystal of LiMnO2 was observed
to undergo a delithiation reaction as described in reaction [26]
to form a phase of composition MnO2152:
2LiMnO2 2H ! MnO2 2Li H2 [26]
The extraction of lithium from LiMnO2 drives a structural
transition in which some of the manganese cations diffuse into
octahedral sites vacated by the lithium cations. As a result, the
MnO2 phase formed by acid delithiation of LiMnO2 has a rock
salt structure in which Mn4 cations are randomly distributed
over half the available octahedral sites (Figure 15).
In contrast, the reaction between powder samples of ortho-
rhombic LiMnO2 and acid proceed via a combination of dis-
proportionation and ion exchange as shown in reactions [27]
and [28].153 In this case, the deintercalation reaction drives a
migration of the manganese cations so that they adopt the
l-MnO2 network of delithiated LiMn2O4, which shares a com-
mon oxide ion lattice with the rock salt structure. Electrochem-
ical delithiation of orthorhombic LiMnO2 also results in the
formation of l-MnO2.154 The observation that both LiMn2O4
and orthorhombic LiMnO2 (under the correct conditions) are
Soft Chemistry Synthesis of Oxides 433
al LiMnO2
r)
de
o w H+ (Single crystal)
MnO2
(Disordered rocksalt)
delithiated to form l-MnO2 suggests that this phase has good
kinetic stability.
3LiMnO2 4yH ! Li32y Mn3y O62y yMn2
2yLi 2yH2 O [27]
LiMnO2 xH ! Li1x Hx MnO2 xH [28]
2.15.4.2.2 Titanates
There are three common polymorphs of titanium dioxide
rutile, anatase, and brookite which all adopt structures con-
sisting of connected TiO6 octahedra.5 It is possible to prepare
further metastable polymorphs of TiO2 by the low-temperature
oxidative deintercalation of cations from AxTiO2 phases.
2.15.4.2.2.1 TiO2(R)
Reaction of a 1:2 molar ratio of lithium metal and TiO2 at
1200  C yields Li0.5TiO2 which adopts a structure in which
TiO6 octahedra are arranged into a corner and edge-sharing
network analogous to that of the ramsdellite polymorph of
MnO2.155 The lithium cations are located in approximately
tetrahedral sites within this TiO2 network, as shown in
Figure 16. On exposure to air, lithium is topochemically dein-
tercalated, such that LixTiO2 phases with x 0.41, 0.16, and
0.14 are formed after a couple of days, 1 month, and 1 year,
respectively. Reaction with HCl enables the complete removal
of all the lithium from Li0.5TiO2 according to reaction [29]156:
Li0:5 TiO2 0:5HCl ! TiO2 0:5LiCl 0:25H2 [29]
As shown in Figure 16, the deintercalation of lithium from
Li0.5TiO2 is topochemical and thus enables the preparation of a
new metastable polymorph of TiO2 with the ramsdellite
434 Soft Chemistry Synthesis of Oxides
HCl
Li0.5TiO2
Figure 16 Oxidative deintercalation of lithium from the ramsdellite-phase Li0.5TiO2 yields a metastable polymorph of titanium dioxide, TiO2(R).
H+/H2O2
K0.25TiO2
Figure 17 Oxidative deintercalation of potassium from the hollandite-phase K0.25TiO2 yields a metastable polymorph of titanium dioxide, TiO2(H).
structure, referred to as TiO2(R). The metastability of the
TiO2(R) polymorph can be seen by observing that it transforms
to the brookite polymorph of TiO2 on heating above 370  C.157
2.15.4.2.2.2 TiO2(H)
K0.25TiO2 can be readily prepared by heating the appropriate
ratio of metallic potassium and titanium dioxide in a sealed
nickel tube at approximately 1000  C. The phase formed adopts
a structure analogous to that of the mineral BaMn8O16 hollan-
dite, which consists of double chains of edge-sharing TiO6 octa-
hedra which share edges to form wide tunnels occupied by
potassium ions as shown in Figure 17.158
Reaction between K0.25TiO2 and an acidic solution of hy-
drogen peroxide leads to the deintercalation of potassium in a
topochemical manner to yield a new metastable hollandite
polymorph of titanium dioxide, TiO2(H), as shown in Figure 17.
TiO2(H) has the lowest density of any of the TiO2 polymorphs
reported to date.158,159 On heating TiO2(H) above 410  C, a
structural transition occurs to form the more stable anatase poly-
morph of TiO2, emphasizing the metastability of the deinterca-
lated phase.
2.15.4.3 Anion Deintercalation Reactions
The topochemical deintercalation of oxide ions from complex
transition-metal oxides can be utilized to prepare a large num-
ber of metastable phases which contain transition-metal cen-
ters in unusual oxidation states and coordination geometries.
In order for anion extraction reactions to be effective, a number
of conditions must be met:
1. The mobility of the anion lattice must be much greater than
the cation lattice at the temperature of reaction, if anions
are to be removed from a host phase without disruption or
rearrangement of the cation framework.
TiO2(R)
TiO2(H)
2. The removal of anions must leave a kinetically stable phase.
The removal of anions from a complex oxide will tend to
lower the coordination number of the metal cations pre-
sent. The anions within complex oxides act to screen the
unfavorable Columbic repulsions between highly charged
metal cations; therefore, anion deintercalation is only pos-
sible if any unfavorable cationcation interactions in the
resulting phase are small. This generally requires the
transition-metal cations in host phases suitable for topo-
chemical reduction to initially reside in sites of high coor-
dination number (CN  5).
3. The deintercalation of oxide ions formally reduces the
transition-metal centers within the host phase. The reduc-
ing agent or method used must therefore have sufficient
reducing power at the temperature of reaction to bring
about the reduction.
2.15.4.3.1 Cubic perovskite phases
The anion deintercalation chemistry of phases which adopt
the cubic perovskite structure, or distorted variants of this
structure, has been widely studied due to the ability of phases
with this structure type to accommodate large numbers of
anion vacancies with minimal energetic penalties.160 This is
particularly true of host phases which contain late, first-row
transition-metal centers (Mn, Fe, Co, Ni) within their MO6
octahedral coordination sites.
For example, the reaction of the perovskite-phase CaMnO3
with hydrogen at 300  C yields the topochemically reduced
phase CaMnO2.5,161 in which the MnIVO6 octahedra of the
host phase have been converted into an ordered array of MnIIIO5
square pyramids in the reduced product (Figure 18).162 The
reduction proceeds by the reaction of hydrogen with CaMnO3
to form a reduced CaMnO3x phase and water vapor, according
to reaction [30]:

H2
300 C
CaMnO3 CaMnO2.5
Figure 18 The topochemical reduction of CaMnO3 yields CaMnO2.5.
CaMnO3 xH2 ! CaMnO3x xH2 O [30]
The reaction is spontaneous (DG < 0) due to the formation
of water. However, the reduction reaction is readily reversible
so the equilibrium position (level of anion deintercalation)
can be controlled by adjusting the partial pressure of water in
the reacting system allowing the controlled synthesis of phases
in the composition range CaMnO3x (0  x  0.5).163
The physical behavior of CaMnO3x phases reveals a feature
common to many topochemically reduced materials. On re-
duction, the coordination number of the transition-metal cat-
ion being reduced, in this case manganese, is lowered. As a
result, it is not only the electron count (oxidation state) of the
metal center which is changed on anion deintercalation, but
also the relative energies of the different partially filled metal
d-orbitals, as these are sensitive to the local symmetry of the
cation-coordination site. This change in the cation-coordination
number and d-orbital energy has a tendency to localize the
electrons added to the system on reduction at the low-coordinate
metal centers, often leading to insulating behavior in topo-
chemically reduced phases. This insulating behavior is often in
contrast with the behavior of analogous cation-doped phases.
For example, phases in the compositional series LaxCa1xMnO3
(0 x  1) adopt distorted perovskite structures164 and span the
manganese oxidation state ranging from Mn4 at x 0 to Mn3
at x 1, in an analogous manner to the anion-deficient series
CaMnO3x (0 x  0.5) which spans the same oxidation state
range. CaMnO3, the x 0 member of both series, is an antiferro-
magnetic insulator.165 Substitution of calcium with lanthanum
to form mixed-valent Mn3/4LaxCa1xMnO3 phases induces
metallic behavior and complex magnetism at low tempera-
ture.166 In contrast, topochemically reduced CaMnO3x phases
remain insulated across the whole composition range due to the
localization of electrons at five-coordinate MnIII centers.161
An alternate reduction method widely applied to transition-
metal perovskite phases utilizes metals as oxygen getters. At
temperatures where there is appreciable anion mobility, the
oxygen content of a complex metal oxide is in equilibrium with
the partial pressure of oxygen in contact with the surface of the
solid phase according to reaction [31]:
MOn $ MnOn2x xO2 [31]
The position of the equilibrium is a balance between the
lattice energy released on oxygen incorporation (oxidation of
Mnn2x to MOn) and the entropy increase of reduction, due to
the release of oxygen. This equilibrium can be driven to the
right by taking advantage of the favorable entropy change
Soft Chemistry Synthesis of Oxides 435
on reduction and simply raising the reaction temperature.
Alternatively, the reduction of the complex oxide phase can
be favored by artificially lowering the oxygen partial pressure
in the system, either by pumping the oxygen away or purging
with an inert atmosphere or by the use of a highly electropos-
itive metal such as titanium or zirconium. In this latter case, the
metals will react rapidly with the low partial pressure of oxygen
in the system gettering the oxygen away and thus driving the
reduction of the complex oxide phase. Thus if the cubic perov-
skite LaCoO3 is heated to 400  C in the presence of zirconium,
a deintercalation reaction occurs according to reaction [32], to
yield the Co(II)-phase La2Co2O5167:
LaCoO3 $ LaCoO32x xO2 Zr ! LaCoO2:5 ZrOy [32]
The oxygen removed from the complex oxide passes
through the gas phase. As a result, the substrate oxide and the
metal getter (Zr in this case) need to be in close proximity, but
do not need to be in contact, so the reduced product need not
be contaminated with ZrOy byproducts. In addition because
the substrate oxide and metal getter can be physically sepa-
rated, the getter can be heated to a higher temperature than the
substrate to facilitate rapid oxygen scavenging and a faster
reaction.
The reduced phase produced via getter reduction of
LaCoO3, La2Co2O5, adopts the brownmillerite structure. This
is one of the most common anion-deficient variants of the
perovskite structure and consists of sheets of CoIIO6 octahedra
stacked in an alternating sequence with sheets of CoIIO4 tetra-
hedra (Figure 19).167 The material is an electronic insulator
which adopts an antiferromagnetically ordered structure below
TN  300 K. In principle, it should be possible to tune the level
of reduction in getter reactions by adjusting the temperature of
the substrate phase to shift the position of the equilibrium in
reaction [32]. However, in practice this is extremely challeng-
ing so LaCoO3x phases of intermediate composition are
most easily prepared by the controlled partial oxidation of
La2Co2O5.168
The two reduction strategies discussed thus far, hydrogen
reduction and metal getters, operate at moderate temperatures.
In the case of hydrogen reduction, a reasonable amount of
thermal energy is required to overcome appreciable activation
barriers associated with the strength of the HH bond of the
hydrogen gas reductant. Likewise metal getter reactions must
operate at reasonably high temperatures due to the entropy-
driven nature of the oxygen-release process. As a result, these
two reduction techniques preclude the formation of the most
metastable phases because, at their temperature of operation
(T  300  C), the cations in the most metastable phases are
mobile, and therefore reduction reactions lead to the formation
of thermodynamic products typically binary oxide phases
rather than metastable topochemically reduced phases.
In order to address this problem, binary metal hydrides
such as NaH,169 CaH2,170 and LiH171 have been utilized as
solid-state reducing agents. The topochemical reduction of the
Ni(III) perovskite-phase LaNiO3, via a zirconium getter reaction
at 400  C, results in the formation of La2Ni2O5.172 This Ni(II)
phase adopts a unique structure consisting of stacked chequer-
board layers of NiIIO6 octahedra and NiIIO4 square planes.
Attempts to decrease the oxygen content of the LaNiO3x
436 Soft Chemistry Synthesis of Oxides
Zr getter
300 C
LaCoO3
Figure 19 The topochemical getter reduction of LaCoO3 with zirconium yields the brownmillerite-phase La2Co2O5.
LaNiO3
Figure 20 The topochemical reduction of LaNiO3 with NaH yields the infinite layer phase LaNiO2.
phase further by raising the temperature of the getter reaction
result in sample decomposition to La2O3 and elemental nickel.
However if an intimate mixture of LaNiO3 and NaH is heated
at 190  C under vacuum within a sealed glass ampoule, a
topochemical reduction of the substrate phase occurs to from
LaNiO2,169 driven by the formation of NaOH according to
reaction [33]:
LaNiO3 NaH ! LaNiO2 NaOH
The extremely low reaction temperature demonstrates that
there are no large activation barriers in the reduction process.
When combined with a large thermodynamic driving force
provided by the formation of NaOH, this allows the highly
metastable Ni(I)-phase LaNiO2 to be isolated (Tdecomposition of
LaNiO2  225  C). In order to act as an effective reducing
agent, sodium hydride must be in close physical contact
with the phase to be reduced. As a result, the reduced product
phases are contaminated with a mixture of NaOH byproducts
and unused NaH reducing agent. These can be readily re-
moved by reaction/dissolution in a weak acid such as
methanol.169
LaNiO2 adopts an infinite layer structure consisting of
sheets of apex-sharing NiIO4 square planes which are stacked
in an alternating sequence with sheets of La3 cations
(Figure 20). As a result, the phase is isoelectronic and iso-
structural with Sr1xCaxCuO2 which is considered to be the
parent structure of all the layered high-Tc superconducting
La2Co2O5
NaH
190 C
LaNiO2
cuprates.173 Despite the structural and electronic similarity of
LaNiO2 to the corresponding copper oxide phases, the nickel
compound does not exhibit superconductivity. This is attrib-
uted to insufficient mixing of the Ni 3d and oxygen 2p orbitals
in this material.174
In contrast to sodium hydride, which acts as a two-electron
reducing agent, formally converting H into H, calcium hydride
and lithium hydride act as one-electron reducing agents convert-
[33] ing H into H0 according to reactions [34] and [35]175,176:
Sr7 Mn4 O15 2CaH2 ! Sr7 Mn4 O13 2CaO 2H2 [34]
BaMnO3 2LiH ! BaMnO2 Li2 O H2 [35]
The oxide byproducts of these reactions can be readily
removed by dissolution in a weak acid such as NH4Cl in
methanol.170,171 However under some circumstances this is
not possible due to the acid sensitivity of the topochemically
reduced phase.177 The choice of reducing agent therefore
becomes a compromise between the synthetic simplicity of
hydrogen or getter reactions and the reducing power and
chemical compatibility of hydride reducing agents.
The utility of hydride reducing agents can be seen in the
preparation of mixed-valent MnII/III oxide phases. A very large
number of studies have investigated the reduction chemistry of
La1xAxMnO3 (A Ca, Sr, Ba) perovskite phases in order to
prepare anion-deficient materials with average manganese ox-
idation states between Mn(II) and Mn(III).178181 The majority

of these studies have utilized hydrogen as a reducing agent.
However, this synthetic route appears to be limited to prepar-
ing phases with a minimum average oxidation state of Mn2.5.
Thus, for example, reaction of La0.5A0.5MnO3 (A Ca, Sr) with
hydrogen at 600  C yields the respective La0.5A0.5MnO2.5 brown-
millerite phases182,183 and the analogous reduction of LaMnO3
yields LaMnO2.75.184,185 In contrast, reaction of La1xAxMnO3
(A Ca, Sr; 0.2  x  0.4) phases with NaH at 190  C yields the
corresponding La1xAxMnO2.5 brownmillerite phases with man-
ganese oxidation states which span the range 2.22.4.69,186
Topochemical reduction reactions can also be used to pre-
pare extended oxide phases which contain transition-metal
cations in highly unusual coordination environments.
SrFeO3d can be readily reduced to the Fe(III) brownmillerite-
phase Sr2Fe2O5 with a wide range of reducing agents. Further
reaction of Sr2Fe2O5 with CaH2 leads to an anion deintercala-
tion reaction which drives a large-scale reorganization of the
host anion lattice to yield the Fe(II)-phase SrFeO2.187,188 This
phase adopts an infinite layer structure like LaNiO2169
(Figure 21) in which Fe(II) centers are located within square-
planar FeIIO4 coordination sites.
The stabilization of a phase containing iron in square-
planar coordination sites is highly unusual. Square-planar
coordination sites are generally only occupied by transition-
metal centers which have d-electron counts which can provide
some electronic stabilization in D4h symmetry (e.g., d8 Ni2,
Pt2; d9 Cu2). However, the high-spin d6 electronic configu-
ration of the Fe(II) centers in SrFeO2, determined from neutron
diffraction data and Mossbauer spectra,187 provides no obvi-
ous electronic stabilization in a square-planar coordination
geometry, making their existence a bit of a puzzle. There is,
however, some evidence for a more complex stabilization of
the square-planar geometry by the Fe(II) centers as substitution
of greater than 30% of the iron centers with manganese or
cobalt leads to the formation of a different structure type on
reduction.189
SrFeO2 is an antiferromagnetic insulator with an ordering
temperature of 473 K.187 On the application of pressures
greater than 40 GPa, the material undergoes a spin transition
to a highly unusual S 1 intermediate spin state accompanied by
a change in physical behavior from an antiferromagnetic insula-
tor to a ferromagnetic conductor.190 This pressure-induced
CaH2
280 C
Sr2Fe2O5 SrFeO2
Figure 21 The topochemical reduction of Sr2Fe2O5 yields the infinite
layer phase SrFeO2.
Soft Chemistry Synthesis of Oxides 437
insulator to metal transition again highlights the large on-site
repulsion effects which typically localize the valence electrons in
topochemically reduced phases.
2.15.4.3.2 RuddlesdenPopper phases
Complex oxides which adopt layered RuddlesdenPopper-type
structures also readily exhibit oxygen nonstoichiometry and so
can act as good substrate phases for oxygen deintercalation
reactions. The reduction chemistry of An1BnO3n1 phases
where the transition-metal B-cation is manganese has been
particularly well studied, and is typical of the reduction chem-
istry of phases of this structure type.
The reaction of the n 1 RuddlesdenPopper-phase
Ca2MnO4 with hydrogen at 300  C readily yields the Mn(III)-
phase Ca2MnO3.5.161 On reduction oxide ions are deinterca-
lated from equatorial sites within the MnO2 sheets of the host
phase to form sheets of composition MnO1.5 which have an
identical anion-vacancy arrangement as that present in
CaMnO2.5 (Figure 22) with the manganese centers located
within MnIIIO5 square-based pyramids.191 Topochemical re-
duction of the structurally analogous phase Sr2MnO4d readily
forms Sr2MnO3.5, again by the removal of oxide ions from the
equatorial MnO2 sheets within the host phase. However, the
arrangement of anion vacancies in Sr2MnO3.5 is such that
the MnIIIO5 units formed on reduction are arranged into tetra-
mers which then share corners (Figure 22) in a strong contrast
to the anion-vacancy ordering present in Ca2MnO3.5.192
Utilizing the more powerful reducing agent CaH2, the n 1
phases LaSrMnO4 and LaBaMnO4 can be reduced to the re-
spective Mn(II) phases LaSrMnO3.5 and LaBaMnO3.5, consis-
tent with the greater low-temperature reducing powder of the
hydride reagent.193 Oxide ions are again deintercalated from
the equatorial sheets of the substrate phases, but in this in-
stance the arrangement of anion vacancies is disordered within
this plane.
The topochemical reduction of cobalt-based Ruddlesden
Popper phases can also be achieved. Reduction of the n 3
phase La4Co3O10 using a zirconium getter reaction readily
yields the Co(II)-phase La4Co3O9.194 During the reduction
process, anions are deintercalated from the central CoO2
sheet of the host phase to form an anion-deficient CoO layer
which adopts an anion-vacancy-ordered arrangement analo-
gous to that present in the mineral brownmillerite.
Reaction of the n 1 cobaltate LaSrCoO4 with hydrogen
readily yields Co(II)-phase LaSrCoO3.5 in which the anion va-
cancies are accommodated in a disordered arrangement within
equatorial CoO1.5 sheets.195 Reaction with the more powerful
reducing agent NaH yields the more oxygen-deficient-
phase LaSrCoO3.38 which has an average cobalt oxidation
state of Co1.76 indicating that approximately one-quarter of
the cobalt centers in the phase are monovalent.195
In contrast to the reaction with NaH, the reaction be-
tween LaSrCoO4 and CaH2 at slightly higher temperature
proceeds via a two-step process. In the first step, LaSrCoO4
is reduced to LaSrCoO3.35 according to reaction [36].
There then follows an anion-substitution reaction in
which one oxide ion is exchanged with two hydride ions,
to form an oxide-hydride-phase LaSrCoO3H0.7 according to
reaction [37]170,196:
438 Soft Chemistry Synthesis of Oxides

Ca2MnO3.5

A2MnO4 (A = Ca, Sr)

A2MnO4 (A = Ca, Sr)

Sr2MnO3.5

Figure 22 The reduction of A2MnO4 RuddlesdenPopper phases yields products with anion-vacancy-ordering schemes which are dependent on the
identity of the A-cation.

LaSrCoO4 0:65CaH2 ! LaSrCoO3:35 0:65CaO

0:375H2 [36]

LaSrCoO3:35 0:35CaH2 ! LaSrCoO3 H0:7 0:35CaO [37]

Computational studies have revealed that this highly meta-

stable oxide-hydride phase is stabilized by strong hybridization
between Co 3d, oxygen 2p, and hydrogen 1s orbitals.196 The
high-temperature antiferromagnetic order observed in LaSr-
CoO3H0.7 (TN 380 K) has also been attributed to this hybrid-
ization. Similar reactions with CaH2 to form oxide-hydride CaH2
phases have also been observed for PrSrCoO4, NdSrCoO4,197
and the n 2 phase Sr3Co2O7x.177 350 C
In contrast to the simple anion deintercalations ob-
served on the reduction of manganates and cobaltates, the
low-temperature reduction of RuddlesdenPopper phases
containing iron or nickel occurs in association with large-scale
rearrangements of the anion lattices of the host phases, result-
ing in reduced products which contain transition-metal centers
in square-planar coordination sites. The n 2 Ruddlesden
Sr3Fe2O6 Sr3Fe2O5
Popper-phase Sr3Fe2O7d is readily reduced to Sr3Fe2O6.198
This Fe(III) phase adopts an anion-deficient structure in which Figure 23 Low-temperature reduction of Sr3Fe2O6 drives a
the anion vacancies are located within the central bridging rearrangement of the anion lattice to yield Sr3Fe2O5 which contains
anion site of the n 2 RuddlesdenPopper lattice (Figure 23) corner-linked FeO4 square planes.

CaH2
350 C
La3Ni2O7 La3Ni2O6
Figure 24 Reduction of La3Ni2O7 drives a rearrangement of the anion
lattice to yield La3Ni2O6 containing sheets of corner-linked NiO4 square
planes.
such that the iron centers are located within square-based py-
ramidal coordination sites. Reaction with calcium hydride
leads to the deintercalation of anions from Sr3Fe2O6 to yield
the Fe(II)-phase Sr3Fe2O5199 in which the anion lattice has
undergone a reorganization, similar to that observed in the
reduction of Sr2Fe2O5 to SrFeO2187 (Section 2.15.4.3.1), result-
ing in a square-planar coordination for the iron centers in the
product phase (Figure 23).
Reduction of the n 2 nickelate La3Ni2O7 induces a differ-
ent anion reorganization. Reaction with CaH2 at 350  C yields
the mixed-valent Ni(I/II)-phase La3Ni2O6200 which adopts a
structure related to the T0 -structure of Nd2CuO4. During the
reduction of La3Ni2O7, the apical oxide ions, which reside
within rock salt layers of the host phase, migrate into coordi-
nation sites analogous to those adopted by fluoride ions in the
fluorite structure.5 As a result, the nickel cations in the reduced
product are located within apex-linked NiO4 square planes
(Figure 24) in a manner analogous to that of the infinite
layer phase LaNiO2, described in Section 2.15.4.3.1.
2.15.4.3.3 Reductive fusion of perovskite sheets
The anion deintercalation reactions described above are all
associated with the reduction of transition-metal cations.
However, there are a number of lanthanides, most notably
europium, which exhibit multiple stable oxidation states in
extended oxides. When present in transition-metal-oxide
phases which contain early transition metals in high oxidation
states, these lanthanide cations are reduced in preference to the
transition metals. As the transition-metal oxidation state is
unchanged during the reaction, the reduction of lanthanide
A-cations is associated with lattice rearrangements which pre-
serve the local coordination geometry of the oxophilic
transition-metal centers. Thus, for example, the reduction of
the n 3 DionJacobson-phase CaEuIII2Ti3O10 (prepared via
cation exchange as described in Section 2.15.4.5.1.3) with
hydrogen yields the perovskite-phase CaEuII2Ti3O9 according
to reaction [38]:
Soft Chemistry Synthesis of Oxides 439
H2
750 C
CaEu2Ti3O9
CaEu2Ti3O10
Figure 25 Reduction of the n 3 DionJacobson-phase CaEuIII2Ti3O10
leads to a fusion of adjacent TiO6 octahedra and the formation of an
A-cation-ordered perovskite-phase CaEuII2Ti3O9.
CaEu2 Ti3 O10 H2 ! CaEu2 Ti3 O9 H2 O [38]
The low-temperature nature of the oxygen deintercalation
reaction limits the rate of cation diffusion and thus the perov-
skite product retains the calciumeuropium cation order of the
original DionJacobson phase as shown in Figure 25. An
analogous reaction can be observed on reduction of Ca0.5EuIII-
TiO4, which forms the n 2 RuddlesdenPopper-phase
CaEuII2Ti2O7 when heated in hydrogen. In both of the reac-
tions described, the retention of the TiO6 coordination poly-
hedra leads to a fusion of separated perovskite sheets to form
either perovskite phases, or layered phases of higher order, on
oxygen deintercalation.
2.15.4.3.4 Structural selectivity
The variety of anion-vacancy-ordered structures adopted by
anion-deficient phases indicates that there are a number of
factors which direct the structural selectivity of topochemical
deintercalation reactions. For example, the anion-deficient pe-
rovskite phases Ca2Mn2O5, La2Co2O5, and La2Ni2O5 all adopt
different anion-vacancy-ordered structures in which the
transition-metal cations reside within coordination sites of
differing coordination number and geometry: Ca2Mn2O5
MnIIIO5 square-based pyramids162; La2Co2O5CoIIO6 octahe-
dra and CoIIO4 tetrahedra167; and La2Ni2O5NiIIO6 octahedra
and NiIIO4 square planes.172 In these three examples, the struc-
tures adopted by these A2B2O5 phases can be rationalized
by considering the respective coordination preferences of
the Mn3, Co2, and Ni2 cations as directed by the crystal
field stabilization provided to cations in coordination sites of
different geometry. However, the different anion-vacancy-
ordering schemes adopted by the two Mn(III), n 1
RuddlesdenPopper phases Ca2MnO3.5 and Sr2MnO3.5
(Figure 22) demonstrate that the coordination preferences of
the transition-metal cations are not the only factor to consider.
Instead, the anion-vacancy distributions in topochemically
reduced phases are best considered as being due to the coordi-
nation preferences of all the cations in the host phase. As a
result, changes to the identity of the A-cations within a host
phase can lead to striking changes to the anion-vacancy-
440 Soft Chemistry Synthesis of Oxides
ordered structure and oxygen stoichiometry of topochemically
reduced products. For example, the n 2 RuddlesdenPopper-
phase YSr2Mn2O7 can be readily reduced with sodium hydride
to yield the Mn(II)-phase YSr2Mn2O5.5 in which oxide ions
have been deintercalated from the equatorial MnO2 sheets of
the host phase.201 In contrast, hydrogen reduction of
NdSr2Mn2O7, which differs from YSr2Mn2O7 by the simple
aliovalent substitution of Nd3 for Y3, removes oxide ions
from the central bridging oxide site in the n 2 Ruddlesden
Popper lattice,202 demonstrating the structure directing role of
the spectator A-cations.
A more dramatic change can be observed as a function of A-
cation identity, in the reduction reactions of REBaCo2O5
phases. Phases of composition REBaM2O5 (RE lanthanide,
Y; M Mn, Fe, Co) adopt structures in which the strong pref-
erence of the large Ba2 cation for 12-coordination drives the
simultaneous ordering of both the A-cations and anion vacan-
cies to yield the anion-deficient cation-ordered perovskite
structure shown in Figure 26 (see Section 2.15.4.4.1.2).203
207
Reduction of YBaCo2O5 with sodium hydride leads to the
formation of YBaCo2O4.5. Deintercalation of oxide ions from
the equatorial anions sites of the host lattice yields a Co(II)
phase which contains cobalt centers located in four- and five-
coordinate sites as shown in Figure 26.208 In contrast, reduc-
tion of the isoelectronic and isostructural lanthanum-
substituted analog, LaBaCo2O5, with sodium hydride yields
LaBaCo2O4.25. This mixed-valent Co(I/II) phase adopts a dra-
matically different anion-vacancy-ordering pattern resulting in
a structure which contains CoIIO5 square-based pyramids,
CoIIO4 tetrahedra, and CoIO4 square planes.208 This dramatic
change in the reduction product of REBaCo2O5 phases on the
substitution of yttrium with lanthanum further demonstrates
the important structure directing role the A-cations play in
topochemical reduction reactions.
YBaCo2O4.5
REBaCo2O5
LaBaCo2O4.25
Figure 26 Topochemical reduction of REBaCoO5 results in product phases with anion-vacancy-ordering schemes directed by the identity of the RE
cations.
2.15.4.4 Anion-Insertion Reactions
In order to effectively insert anions into metal-oxide host lat-
tices in a topochemical manner, a number of conditions need
to be met.
1. There must be intercalation sites within the host framework
which can accommodate the inserted anions. These can be
anion vacancies which already exist within the close-
packed host frameworks, unoccupied interstitial anion
sites, or anion-insertion sites that can be formed by minor
rearrangements of the host oxide lattice.
2. The host lattice must contain oxidizable centers. The inser-
tion process is formally an oxidation in which a neutral
species is inserted into the host material (for example, an
oxygen atom); this species becomes reduced (to an O2
oxide ion for example), thus oxidizing the host phase. In
order for the insertion process to proceed, the oxidation
needs to be thermodynamically favorable under the soft
conditions applied. Therefore, the lattice/bond energy re-
leased during the oxidation must be larger than the ener-
getic cost of the electron transfer required to form the
cationanion pair.
3. There must be sufficient anion mobility to allow the
inserted species to diffuse into the interior of the host
material, rather than just remain at the surface. In addition,
there must be sufficient electronic mobility to enable the
motion of charge associated with the oxidation process.
2.15.4.4.1 Anion-deficient perovskites
The ABO3 cubic perovskite structure contains no obvious
sites into which additional anions can be inserted. When
phases which are over-stoichiometric in oxygen adopt cubic
perovskite structures, the elevated anion contents are
Co(II)
Co(II)
Co(II)
Co(I/II) Co(II) Co(I)

accommodated through cation deficiency. Thus, the over-
stoichiometric composition LaMnO3d is realized through lan-
thanum and manganese cation vacancies and is better written as
La1dMn1dO3.209 However as noted in Section 2.15.4.3.1,
phases which adopt the perovskite structure can accommodate
large numbers of anion vacancies without incurring significant
energetic penalties. As a result, the oxygen content of perovskite
phases is particularly sensitive to the preparative conditions
employed for their synthesis, and anion-deficient perovskite
phases are easily prepared. The anion deficiency can then be
readily tuned or removed by low-temperature topochemical
oxidation reactions.
2.15.4.4.1.1 Structural stabilization
The ABO3 cubic perovskite structure, and its distorted variants,
is almost ubiquitous in ternary transition-metal-oxide chemis-
try, due to the large number of cation combinations it can host.
This chemical diversity is attributable to the wide range of A:B
cation radius ratios which can be accommodated. As shown in
Figure 27, the large A-cations reside in 12-cordinate sites, while
the smaller transition-metal cations are located in six-
coordinate octahedral sites within the perovskite lattice. If the
unit cell is a perfect cube, it can be seen that the AO and BO
bond lengths must be in the ratio: (AO)/(2  (BO)) 1,
which is defined as the structural tolerance factor, t. This ap-
parently restrictive geometric requirement can be eased by a
series of cooperative twisting and tilting distortions of the BO6
octahedra, which lower the average AO bond length while
maintaining the BO bond length. This allows compositions
with small A-cations (t < 1) to adopt distorted perovskite struc-
tures, significantly extending the stability range of this structure
type. There are however a significant number of ABO3 compo-
sitions which have tolerance factors greater than 1, indicating
that the A-cation is too large for the 12-coordinate site of a
cubic perovskite framework. These phases tend to adopt hex-
agonal perovskite structures which incorporate face-sharing
links into the connectivity of BO6 octahedra, to accommodate
these larger A-cations. A small subset of these hexagonal pe-
rovskite phases can be induced to adopt cubic-type perovskite
structures by synthesizing them with anion-deficient composi-
tions. The ABO3x phases thus prepared will contain transition-
metal B-cations in lower oxidation states than present in
oxygen-stoichiometric ABO3 formulations. This B-cation re-
duction will tend to increase the average BO bond length,
B 2.15.4.4.1.2 Cation-ordered phases
A AO
O
BO
(AO)
ABO3 t= =1
2 (BO)
Cubic perovskite
Figure 27 The structural tolerance factor of ABO3 perovskite phases.
Soft Chemistry Synthesis of Oxides 441
reducing the tolerance factor and moving the material into the
cubic perovskite stability field. The anion-deficient phases can
then be oxidized to stoichiometric ABO3 compositions. If this
oxidation is performed at a suitably low temperature, there will
be insufficient cation mobility for the material to adopt the
thermodynamically stable hexagonal perovskite structure, and
the metastable, oxygen-stoichiometric, cubic phase will be ki-
netically trapped. Thus, for example, SrMnO3, which crystal-
lizes as a 4H hexagonal perovskite,210,211 can be converted into
a cubic perovskite-phase SrMnO2.72 by heating in air at high
temperature.210 Subsequent reaction in oxygen below 500  C
yields SrMnO3.00, which retains the cubic perovskite structure
of the anion-deficient phase.212
The oxidation of cubic SrMnO3x can also be achieved with
fluorine. SrMnO2.5,213 prepared in a manner analogous to that
described for CaMnO2.5 in Section 2.15.4.3.1, reacts readily
with XeF2 a convenient fluorine source to yield SrMnO2.5
xF0.5x, a mixed-valent Mn(III)/Mn(IV) phase as shown in reac-
tion [39].214 In addition to the simple fluorine insertion, the
highly oxidizing nature of fluorine leads to an accompanying
anion displacement reaction in which fluorine displaces oxy-
gen to yield SrMnO2.5xF0.5x rather than SrMnO2.5F0.5:
0:5 x x
SrMnO2:5 XeF2 ! SrMnO2:5x O0:5x O2
2 2
0:5 x
Xe [39]
2
SrFeO3x phases adopting cubic perovskite-derived struc-
tures can be prepared in a similar way. Reaction of the
brownmillerite-phase Sr2Fe2O5 at low temperature with oxy-
gen partial pressures of up to 500 atm allows phases of com-
position SrFeO2.5x (0 < x < 0.5) to be synthesized.215 These
materials exhibit a number of anion-vacancy-ordered interme-
diate phases which can be described by the compositional
series SrnFenO3n1 (n 2, 4, 8, 1).216 Oxidative insertion of
anions into SrFeO3x can also be performed electrochemically
by electrolyzing samples in KOH to yield highly crystalline
SrFeO3.217 The anion-deficient lattice of SrFeO3x phases also
enables fluorine insertion. Reaction of SrFeO2.88 with poly
(vinylidene fluoride), [CH2CF2]n, leads to an anion-insertion/
displacement reaction and the formation of SrFeO2F.218 In
contrast to the fluorination of Sr2Mn2O5, the fluorination of
SrFeO3x proceeds with the reduction of the iron oxidation
state from Fe3.76 to Fe3 indicating that the oxidizing nature
of Fe4 centers leads to a significant substitutive component to
the anion-insertion reaction.
Anion-deficient cubic perovskite phases are often prepared to
encourage cation ordering. In general, entropy strongly favors
the formation of cation-disordered perovskite phases. Thus, an
AA0 B2O6 double perovskite composition will tend to adopt a
structure with a statistically disordered arrangement of A- and
A0 -cations over the available 12-coordinate cation sites of the
perovskite framework. However, the introduction of anion
deficiency can lead to the synergic ordering of anion vacancies
and cations. As shown in Figure 28, LaBaMn2O5 adopts a
structure based on that of a cubic perovskite, in which the
anion vacancies order to yield two distinct A-cation sites: a
12-coordinate site occupied by Ba2 and a smaller 8-
coordinate site occupied by La3.52 The simultaneous ordering
442 Soft Chemistry Synthesis of Oxides
O2
220 C
LaBaMn2O5
Figure 28 Low-temperature oxidation of LaBaMn2O5 leads to the formation of the metastable A-cation-ordered perovskite-phase LaBaMn2O6 rather
than the more thermodynamically stable cation-disordered phase La0.5Ba0.5MnO3.
MnO2 MnO2
GaO2 GaO
MnO2 MnO2
GaO GaO
MnO2 MnO2
Ca2GaMnO5.5
Figure 29 Oxidation of A2GaMnO5 (A Ca, Sr) brownmillerite phases yields products with anion-vacancy distributions which depend on the identity of
the A-cations.
of anion vacancies and cations is driven by the differing coor-
dination preferences of the large barium and smaller lantha-
num cations. Subsequent low-temperature oxidation yields
metastable, cation-ordered LaBaMn2O6 which exhibits an ele-
vated magnetic ordering temperature (TN 335 K) compared
to the cation-disordered phase La0.5Ba0.5MnO3 (TN 270 K)
which is prepared directly at high temperature.52 Similar be-
havior is observed for LaBaCo2O5 which can be readily oxi-
dized to A-cation-ordered LaBaCo2O6.219
Anion deficiency can also be used to prepare B-cation-
ordered perovskite phases. As noted in Section 2.15.4.3.1, the
anion-deficient A2B2O5 brownmillerite structure is adopted
by a large number of anion-deficient perovskite phases.
The octahedral and tetrahedral coordination sites within
brownmillerite-type structures (Figure 29) can be used to
segregate and order different B-cations on the basis of their
stability in these two coordination geometries, to yield cation-
ordered A2BB0 O5 phases. Subsequent low-temperature anion-
insertion reactions can then be utilized to tune the oxygen
content of materials.
A widely studied group of cation-ordered brownmillerite
oxides are the A2BMnIIIO5 (A Ca, Sr; B Al, Ga) phases in
which cation ordering in these materials is particularly robust
due to the favorable location of the JahnTeller-active MnIII
cations within an octahedral coordination site. Oxidation of
these materials to MnIV-containing A2BMnO5x compositions
leads to the insertion of oxide ions into the vacant anion sites
LaBaMn2O6 La0.5Ba0.5MnO3
MnO2
GaO1.5
MnO2
GaO1.5
MnO2
A2GaMnO5 Sr2GaMnO5.5
which reside within the layers of apex-linked BO4 tetrahedra,
converting them to BO6 octahedra. However, the manner in
which these inserted anions are arranged depends on the size
of the A-cations.
When the A-cation is small, Ca2 for example, oxide ions
are inserted into alternate layers of tetrahedra, changing the
OTOT stacking sequence of octahedra (O) and tetrahedra (T)
in the brownmillerite host phase into an OOOTOOOT se-
quence. Thus, the oxidation of Ca2GaMnO5 to Ca2GaMnO5.4
yields a phase in which approximately half the Ga3 cations are
in octahedral coordination sites and half in tetrahedral coor-
dination sties (Figure 29).220,221 When the A-cations are larger,
oxidation follows a different pattern. Reaction of Sr2GaMnO5
with oxygen forms the MnIV-phase Sr2GaMnO5.5 in which the
additional anions have been inserted evenly into all the sheets
of tetrahedra within the brownmillerite structure, oxidizing the
anion-deficient GaO layers to a composition of GaO1.5
(Figure 29).222224 As a result, the gallium cations have an
average coordination number of 5. It is argued that the differ-
ences in oxidation behavior are driven by the differing ability
of the two A2BMnO5.5 structures to accommodate small A-
cations through cooperative twisting of polyhedra.225
2.15.4.4.2 RuddlesdenPopper phases
2.15.4.4.2.1 Oxygen-insertion reactions
In common with the perovskite phases described above, it is
possible to stabilize RuddlesdenPopper structures, particularly
 Soft Chemistry Synthesis of Oxides 443

Oxidation

La2MO4 La2MO4+x

Figure 30 Oxidation of La2MO4 (M Co, Ni, Cu) inserts additional

anions into tetrahedral coordination sites between the sheets of MO6
corner-sharing octahedra.
LaSrMnO4F
n 2 members, by the formation of anion-deficient phases. For
example, the reaction of a 3:2 molar ratio of SrCO3 and MnO2,
which would be expected to form the n 2 RuddlesdenPopper-
phase Sr3Mn2O7, actually results in the formation of Sr7Mn4O15 La1.2Sr1.8Mn2O7F2
and MnO2.226 However if prepared at high temperature
Figure 31 Fluorination of RuddlesdenPopper phases inserts fluoride
(1650  C) to favor anion deficiency, and then quenched to
ions into tetrahedral interstitial sites.
room temperature with dry ice, the n 2 RuddlesdenPopper-
phase Sr3Mn2O6.55 is formed, which can then be readily oxidized
to the MnIV-phase Sr3Mn2O7.227 Likewise Sr3Co2O7x can be
the perovskite blocks of this structural family, to form struc-
stabilized with respect to a mixture of competing phases, by
tures similar to those of the La2MO4d phases described above.
lowering the oxygen concentration of the material.228 In this
For example, reaction of the mixed-valent Mn3/4 n 2
instance, reaction with oxygen only facilitates oxidation to
RuddlesdenPopper-phase RE1.2Sr1.8Mn2O7 (RE La, Pr, Nd,
Sr3Co2O6.60,229 consistent with the increased difficulty in stabi-
Sm, Eu, Gd) with dilute fluorine gas or poly(vinylidene fluo-
lizing the highly oxidizing Co4 oxidation state.
ride) yields RE1.2Sr1.8Mn2O7F2.236,237 Fluoride ions are
In addition to the oxidative insertion of anions into vacant
inserted into four-coordinate sites within the rock salt layers
sites within anion-deficient RuddlesdenPopper phases, it is also
of the host phase, to form an arrangement reminiscent of the
possible to insert anions into interstitial sites within the rock salt
fluorite structure (Figure 31). Similar reactions with the n 1
layers of phases of this structure type. Thus, oxidation of La2MO4
phase LaSrMnO4 lead to the formation of LaSrMnO4F1.7 in
(M Co, Ni, Cu) yields phases of composition La2MO4d
which the analogous interstitial sites are only partially filled.236
(d 0.15) in which the additional oxide ions are inserted into
The level of intercalation is presumably limited in this case by
interstitial tetrahedral sites within the rock salt layers of these
the ability to stabilize the higher oxidation states of manga-
n 1 RuddlesdenPopper phases (Figure 30).230233 Oxidation
nese. If LaSrMnO4F1.7 is heated with the appropriate amount
of these phases can be achieved either by heating in oxygen gas or
of LaSrMnO4, the Mn(IV)-phase LaSrMnO4F can be formed via
electrochemically in a KOH electrolyte.
the thermal equilibration of the fluorine concentration. The
An unusual feature of these topochemical oxidations is
resulting oxyfluoride phase adopts a novel-staged structure in
that the anions are inserted into locations within the host
which the anion intercalation sites in alternate layers are either
phase which are not within the coordination spheres of the
full or empty (Figure 31).238 This staging behavior suggests
transition-metal cations being oxidized. As a result, the oxida-
that there is a large energy penalty associated with expanding
tion state of the transition metal is changed (raised) while the
the rock salt layers to make them suitable for anion intercala-
local coordination at that transition-metal site, and thus
tion and that this penalty is only worth paying if all the anion
the d-orbital energies of the metal center, remain largely
intercalation sites within a layer are filled, in a situation
unchanged. Modification of the electron count in this structur-
strongly reminiscent of some cation intercalation reactions of
ally independent, isolated manner is much more typical of
layered host phases. It should be noted that these fluorination
cation insertion or substitution, than manipulation of the
reactions must be performed with the rigorous exclusion of
anion lattice which tends to insert or remove anions from
water to prevent the formation of large quantities of binary
within the coordination spheres of the redox active transition-
metal fluoride impurities such as SrF2.
metal centers. Therefore, these remote anion insertions into
In an attempt to simplify the often elaborate experimental
rock salt layers of RuddlesdenPopper phases can be thought
setups required to handle and perform reactions with fluorine,
of as almost pure electronic doping processes.
a number of other fluorinating strategies have been investi-
gated. For example, binary metal fluorides of late transition
2.15.4.4.2.2 Fluorination reactions metals (MF2: M Ni, Cu, Zn, Ag) have been used to good effect
Extensive investigations of the soft fluorination chemistry of to fluorinate RuddlesdenPopper phases.239 For example,
RuddlesdenPopper phases234,235 have revealed that fluoride heating a 1:1 mixture of the n 2 RuddlesdenPopper-phase
ions can be oxidatively inserted into interstitial sites between Sr3Ru2O7 with CuF2 in air at temperatures up to 300  C yields
444 Soft Chemistry Synthesis of Oxides
the Ru(IV)-phase Sr3Ru2O7F2 and CuO.240 It is postulated
that the metal fluorides act as an in situ source of fluorine
through reaction with atmospheric oxygen according to reac-
tion [40],241 and that this fluorine then rapidly reacts with the
host phase to give the overall reaction [41].
1
MF2 O2 ! MO F2 [40]
2
1
Sr3 Ru2 O7 CuF2 O2 ! Sr3 Ru2 O7 F2 CuO [41]
2
It can be seen from these reaction schemes that the atmo-
spheric oxygen is acting as the oxidant and the metal fluoride
only acts as a fluorine source. The use of metal fluorides as
fluorinating agents not only eases the practical difficulties as-
sociated with handling fluorine, it is also observed to produce
lower levels of binary fluoride impurities than reactions with
elemental fluorine. However, it should be noted that samples
fluorinated in this manner almost inevitably become contam-
inated with the binary oxide byproducts of the reaction, which
are often hard to remove.
Soft fluorination chemistry has also been used extensively
to tune the electronic behavior of complex copper oxide phases
to induce superconductivity.241 The structural selectivity of the
anion-insertion/substitution reactions of these phases appears
to be directed by a strong energetic preference for the retention
of CuO4 square-planar motifs. Thus, for example, the fluorina-
tion of Sr2CuO3 to form Sr2CuO2F2x via a combination of
anion insertion and anion exchange leads to a large-scale rear-
rangement of the anion lattice. As a result, the majority of the
fluoride ions within the product phase are located in axial
coordination sites with a small minority located within inter-
layer interstitial sites (Figure 32).241243 Thus, the SrOSrO
CuOSrOSrO stacking of the host phase is converted into
SrFFxSrFCuO2SrFFxSrF in the product. While it is in-
trinsically difficult to determine the oxygenfluorine distribu-
tion in a mixed oxyfluoride by diffraction due to the similarity
of both the x-ray and neutron scattering lengths of the two
anions, Madelung energy calculations provide strong support
for the structural model presented. The observed large-scale
reorganization of the anion lattice also clearly demonstrates
that there is high anion mobility in the host phase.
F2
Sr2CuO2F2+x
Figure 32 Fluorination of A2CuO3 (A Ca, Sr) phases leads to a rearrangement of the anion lattice and the retention of sheets of apex-linked CuO4
square planes.
Fluorination of Ca2CuO3, which is isostructural and iso-
electronic with Sr2CuO3, leads to a subtly different struc-
ture.244 Reaction of Ca2CuO3 with fluorine gas yields
Ca2CuO2F2x in which the majority of the inserted fluoride
ions now reside within interlayer interstitial sites, with a small
minority of fluoride ions located in sites which bond axially to
the copper centers (Figure 32). The structure of Ca2CuO2F2x
is therefore more closely related to the T0 structure of Nd2CuO4
than that of the parent RuddlesdenPopper phase. The differ-
ent distribution of fluoride ions over the two coordination sites
in A2Cu2O2F2x to yield a RuddlesdenPopper phase when
A Sr and a T0 phase when A Ca is consistent with the small
size of Ca2 stabilizing the T0 structure.
2.15.4.4.2.3 Chlorine-insertion reactions
In general, chloride ions are too large to be inserted into vacant
oxide ion-coordination sites. However, chlorine can be
inserted into the interlayer region of RuddlesdenPopper
phases which contain large A-cations. As noted in
Section 2.15.4.1.3.1, exposure of A0 LaNb2O7 (A0 Rb, Cs)
DionJacobson phases to cesium or rubidium vapor leads to
the formation of A0 2LaNb2O7 (A0 Cs, Rb) RuddlesdenPop-
per phases via reductive intercalation.108,109 The A0 -cations in
these phase are sufficiently large to facilitate the oxidative
insertion of chlorine, to yield phases of composition (A0 2Cl)
LaNb2O7 (A0 Cs, Rb).109 The (A0 2Cl) layers thus formed
adopt a CsCl-type structure in which the A0 -cations are located
within eight-coordinate A0 O4Cl4 sites, facilitated by a relative
shift in the position of the LaNb2O7 perovskite blocks from a
centered stacking to a primitive stacking as shown in Figure 33.
2.15.4.4.2.4 Iodine-insertion reactions
The weak oxidizing potential of iodine means that in general it
does not intercalate into complex oxides in an oxidative manner.
It has been observed however that iodine can intercalate into
bismuth-containing oxides. Thus, for example, the reaction of the
high-temperature superconducting phase Bi2Sr2CaCu2O8x with
elemental iodine yields IBi2Sr2CaCu2O8x, in which iodine has
been inserted into the Bi2O2 layers of the host phase driving a
change in the stacking sequence of the phase from centered to
F2
A2CuO3 Ca2CuO3F2+x

primitive (Figure 34).245 Raman spectroscopy data collected
from intercalated samples are consistent with the presence of
I3 molecular anions, demonstrating the oxidative nature of the
insertion process.246 Analogous insertion reactions have been
observed to occur in AxBi9xO(27x)/2 (A Ba, Sr) phases, in
which the insertion of iodine leads to the partial oxidation of
bismuth.247,248
2.15.4.5 Redox-Neutral Topochemical Reactions
The topochemical reactions described in the previous sections
occur with concomitant oxidation or reduction of the complex
oxide host lattice. However, it is possible to insert, extract, or
exchange chemical species from complex oxide phases in
redox-neutral processes. In order for such redox-neutral pro-
cesses to proceed, a number of general conditions must be met:
1. The species to be intercalated and/or extracted must have
significant mobility within the host lattice at the reaction
temperature.
2. The product phase produced must be kinetically stable.
Thus, there must be suitable intercalation sites for insertion
Cl2
Room temperature
A2LaNb2O7
(A2Cl)LaNb2O7
Figure 33 Chlorination of A2LaNb2O7 (A Cs, Rb) leads to the insertion
of chloride ions into eight-coordinate cubic sites formed by a relative shift
in the position of the perovskite double layers.
I2
Bi2Sr2CaCu2O8 IBi2Sr2CaCu2O8
Figure 34 Reaction of Bi2Sr2CaCu2O8 with iodine oxidatively inserts
I3 ions into the Bi2O2 layers of the cuprate phase to yield
IBi2Sr2CaCu2O8.
Soft Chemistry Synthesis of Oxides 445
reactions, or minimal unfavorable-like ion interactions in
the resulting product phase in the case of deintercalation
reactions.
3. The redox-neutral reactions must be spontaneous (DG < 0).
Unlike the redox active reactions which are generally driven
by thermodynamically favorable electron-transfer pro-
cesses, redox-neutral reactions need to be driven by other
favorable reaction steps, such as increase in product lattice
energy (due to exchange of large cations for smaller cations
for example) or acid neutralization or the formation of
highly thermodynamically stable byproduct phases (e.g.,
water) in the case of deintercalation reactions.
2.15.4.5.1 Cation-substitution reactions
The majority of transition-metal cations in complex oxide
phases are tightly bound within oxide coordination polyhedra.
As a result, they are effectively immobile under soft chemical
conditions. In contrast, large, low-valent cations residing in
sites of low coordination number within the interlayer regions
of complex oxide structures can be readily exchanged at modest
temperatures. By utilizing the low-temperature mobility of this
class of cation, a wide range of topochemical reactions can be
performed.
2.15.4.5.1.1 NaFeO2-type phases
The high ionic mobility which facilitates the rapid extraction and
reintercalation of cations into phases which adopt a-NaFeO2-
type structures also enables numerous low-temperature cation-
exchange reactions to be performed.249 For example, reaction of
a-NaFeO2 with lithium chloride in a potassium chloride flux
brings about a topochemical cation exchange according to reac-
tion [42]. The metastable phase, LiFeO2, thus formed retains
the ordering of lithium and iron cations within the rock salt
lattice of the parent phase (the thermodynamically stable form
of LiFeO2 has a disordered rock salt structure) consistent with
the differential mobility of the monovalent cations (Na, Li)
compared to Fe3250:
NaFeO2 LiCl ! LiFeO2 NaCl [42]
However if a-NaFeO2-type phases of composition LiMO2
(M Cr, Co, Rh) are reacted with silver nitrate in a potassium
nitrate flux, the resulting cation-exchange reactions drive a change
to the structure of the host lattice so the product AgMO2 phases
adopt delafossite-type structures in which the inserted silver cat-
ions occupy linear coordination sites, as shown in Figure 35.251
AgNO3
LiCrO2 AgCrO2
Figure 35 Exchange of silver for lithium drives a change from an
ordered rock salt structure to a delafossite structure with silver ions in a
linear coordination geometry.
446 Soft Chemistry Synthesis of Oxides
The structural transition is presumably driven by the coordina-
tion preferences of the AgI cation.
Analogous cation-exchange reactions to form delafossite
phases are also observed when a-NaFeO2-type LiMO2 (M Cr,
Rh) phases are reacted with a mixture of palladium and palla-
dium dichloride, in a process which involves the comproportio-
nation of Pd0 and PdII according to reaction [43]251:
2LiMO2 Pd PdCl2 ! 2PdMO2 2LiCl [43]
It is also possible to substitute divalent cations into a-
NaFeO2-type phases in redox-neutral processes which exchange
two monovalent cations for a single divalent cation. For example,
the reaction of NaxCoO2 with anhydrous calcium nitrate readily
yields Cax/2CoO2 phases according to reaction [44]252,253:
x (A monovalent cation) which adopt structures consisting of
Nax CoO2 CaNO3 2 ! Cax=2 CoO2 xNaNO3 [44]
2
The cation-exchange reactions of NaxCoO2 phases with low
values of x are topochemical. Thus, g-Na0.5CoO2, which adopts
a P2-type structure as described in Section 2.15.4.2.1.2, forms g-
Ca0.25CoO2, and b-Na0.5CoO2, which has a P1-type structure,
forms b-Ca0.25CoO2 on cation exchange. However, NaxCoO2
phases with large values of x undergo structural changes similar
to those observed on sodium deintercalation. Thus, for example,
reaction of a-NaCoO2 with anhydrous calcium nitrate forms
Ca0.5CoO2 with the b-type P1 structure of the most thermody-
namically stable form of Na0.5CoO2 (see Section 2.15.4.2.1.2),
suggesting that the change in structure is driven by the declining
A-cation concentration in the AxCoO2 (A Na, Ca) phases rather
than a change in the identity of the A-cation.
Reaction between a-NaFeO2 and nickel nitrate or nickel
chloride leads to a further class of cation-exchange reactions,
in which the sodium and some of the iron cations are replaced
by nickel according to reaction [45] to yield a Ni1xFe2xO4
spinel-type phase254,255:
2NaFeO2 1 xNiCl2 ! Ni1x Fe22x=3 O4 2NaCl
2x
FeCl3 [45]
3
Detailed structural analysis reveals that despite the com-
mon oxide ion lattice shared by the ordered rock salt structure
of NaFeO2 and the spinel structure, there is significant cation
rearrangement associated with the exchange reaction as half
the iron centers are displaced from octahedral sites in NaFeO2
to tetrahedral sites in the Ni1xFe2(2x/3)O4 product.
Na2Ti3O7
Figure 36 The layered structures of Na2Ti3O7, Tl2Ti4O9, and Cs2Ti5O11.
2.15.4.5.1.2 Layered binary and ternary oxides
Layered binary oxide phases which contain monovalent cations
in between sheets of connected transition-metal coordination
polyhedra exhibit extensive cation-substitution chemistry. For
example, as noted in Section 2.15.4.1.1.2, the layered structure
of a-MoO3 enables a wide range of intercalation reactions to be
performed due to the high mobility of cations located between
the MoO3 layers. This high cation mobility also facilitates a
wide range of cation-exchange reactions, Thus, exposure of
[Na(H2O)2]0.25MoO3 to a solution containing potassium chlo-
ride leads to rapid and complete cation exchange and the forma-
tion of [K(H2O)n]xMoO3 phases.256
Layered structures are also observed for a wide range of
ternary oxides. For example, there is an extensive series of
ternary titanium oxides of composition A2TinO2n1
arrays of edge-sharing TiO6 octahedra separated by layers of
A-cations, as shown in Figure 36. The open structures adopted
by these phases enable a wide range of cation-substitution
chemistry to be performed.257,258 For example, the n 4
phase TlTi4O9 has been shown to undergo substitution reac-
tions with a very wide range of MCl metal monochlorides
(M Li, Na, K, Rb, Cs, Ag) to yield substituted phases accord-
ing to reaction [46]259:
Tl2 Ti4 O9 2MCl ! M2 Ti4 O9 2TlCl [46]
This family of phases has also been observed to react with
mineral acids to form proton-exchanged solid-acid phases. For
example, Na2Ti3O7, K2Ti4O9, and Cs2Ti5O11 all react with
hydrochloric acid to form the analogous H2TinO2n1
phases.260
2.15.4.5.1.3 DionJacobson phases
The large monovalent interlayer A0 -cations present in
A0 An1BnO3n1 DionJacobson phases are relatively mobile
under soft synthesis conditions, due to the low charge density
of the ions and the large sites of low coordination number they
occupy. As a result, these cations can be readily exchanged for
cations of smaller size in reactions which are driven by the
increased lattice energies of the substituted products. Thus, for
example, the reaction of the n 2 DionJacobson phases
A0 LaNb2O7 (A0 K, Rb, Cs) with fused LiNO3 or NaNO3 yields
A0 -cation-substituted phases (Figure 37) and A0 NO3 according
to reaction [47]261,262:
Tl2Ti4O9 Cs2Ti5O11

0 0
A LaNb2 O7 MNO3 ! MLaNb2 O7 A NO3 [47]
If the exchanged cations are small (Li, Na), their intro-
duction drives a structural change in which the perovskite
sheets shift from the primitive stacking of the DionJacobson
phases to the staggered stacking of the RuddlesdenPopper
structural series. This structural change leads to the formation
of small tetrahedral sites which can better accommodate the
small monovalent cations as shown in Figure 37. Analogous
cation-exchange reactions can be performed on DionJacobson
tantalates to yield a wide range of cation-substituted phases,263
including AgLaTa2O7 and AgCa2Ta3O10264 which exhibit
appreciable Ag cation mobility at high temperature.
The A0 -cations in DionJacobson phases can also be
replaced by divalent cations to yield cation-deficient phases.
Thus, reaction of RbCa2Nb3O10 with Sr(NO3)2 or Ba(NO3)2
leads to the formation of Rb0.55Sr0.230.24Ca2Nb3O10 and
Rb0.45Ba0.270.27Ca2Nb3O10, respectively,265 where represents
a vacancy on the A0 -cation site of the DionJacobson structure.
The large A0 -cations in DionJacobson phases can also
be readily exchanged for protons by treatment with mineral
acid, resulting in the formation of solid-acid phases.266269
For example, reaction of the n 3 A0 Ca2Nb3O10 or n 2
A0 LaNb2O7 (A K, Rb, Cs) phases with 6 M HCl readily yields
the solid-acid phases HCa2Nb3O10 and HLaNb2O7 respec-
tively according to reaction [48]266,269:
KCa2 Nb3 O10 HCl ! HCa2 Nb3 O10 KCl [48]
In addition to simple cation-exchange reactions, the
A0 -cations in DionJacobson phases can be replaced by
charged layers of other metal salts. Thus if the n 2
RuddlesdenPopper-phase RbLaNb2O7 is heated in the pres-
ence of CuCl2, an exchange reaction occurs to replace the large
Rb cations with a sheet of stoichiometry (CuCl) according to
reaction [49]270:
RbLaNb2 O7 CuCl2 ! CuClLaNb2 O7 RbCl [49]
The resulting phase consists of double blocks of corner-
sharing NbO6 octahedra stacked with sheets of edge-sharing
CuIIO2Cl4 octahedra as shown in Figure 38.270,271 By use of
RbO8 NaNO3 NaO4
50%
occupied
RbLaNb2O7 NaLaNb2O7
Figure 37 Substitution of sodium for rubidium leads to a shift of the
perovskite blocks to form a cation-deficient RuddlesdenPopper-type
phase.
Soft Chemistry Synthesis of Oxides 447
the appropriate dichloride, similar exchange reactions can be
performed to insert other (MCl) (M Cr, Mn, Fe, Co) layers
into DionJacobson phases in a directly analogous man-
ner.272274 The (MCl)LaMb2O7 phases thus formed can also
undergo cation substitution reactions. Reaction of (CuCl)
LaNb2O7 with n-butyllithium results in the insertion/exchange
of lithium and the excretion of elemental copper according to
reaction [50], to yield (LixCl)LaNb2O7, typically with x > 2.
Subsequent exposure to iodine can readily remove the dis-
placed copper and any excess lithium to yield (Li2Cl)
LaNb2O7 as shown in reaction [51]275:
CuClLaNb2 O7 xC4 H9 Li ! Lix ClLaNb2 O7
x
C8 H18 Cu [50]
2
x1
Lix ClLaNb2 O7 Cu I2
2
! Li2 ClLaNb2 O7 CuI x  2LiI [51]
The (Li2Cl) layer inserted between the perovskite blocks in
(Li2Cl)LaNb2O7 adopts an anti-fluorite structure in which the
lithium cations are located within LiO2Cl2 tetrahedra as shown
in Figure 38, and thus the exchange of two lithium cations for
each Cu2 cation can be considered topochemical.
2.15.4.5.1.4 RuddlesdenPopper phases
The majority of An1BnO3n1 RuddlesdenPopper phases
contain small, highly charged A-cations such as lanthanides
and/or alkaline earth metals which have high charge densities
limiting their mobility at low synthesis temperatures. However,
when the B-cations present in RuddlesdenPopper phases are
early transition metals in high oxidation states such as Ti4,
Nb5, or Ta5, monovalent A-cations can be incorporated into
RuddlesdenPopper phases. Under suitable conditions, these
monovalent cations can participate in substitution reactions in
which they are replaced by other, smaller, monovalent ions.276
Thus, for example, the reaction of the n 3 Ruddlesden
Popper-phase Na2La2Ti3O10 with LiNO3 or AgNO3 yields
Li2La2Ti3O10 and Ag2La2Ti3O10, respectively.277,278
The large monovalent A-cations in RuddlesdenPopper
phases can also be readily exchanged for protons to form
solid-acid phases.279285 For example, the n 3 phase
RbO8 CuO2Cl4 LiO2Cl2
CuCl2 BuLi
RbLaNb2O7 (CuCl)LaNb2O7 (Li2Cl)LaNb2O7
Figure 38 The large rubidium cations in RbLaNb2O7 can be exchanged
for (CuCl) salt units which can then undergo cation exchange in
which a Cu2 cation is replaced by two Li cations.
448 Soft Chemistry Synthesis of Oxides
A2La2Ti3O10 (A K, Rb) when reacted with nitric acid yields
proton-exchanged H2LaTi3O10 materials.261 Using these solid-
acid phases as precursors, subsequent reactions can replace the
incorporated protons with large monovalent cations to yield
phases which cannot be made directly. Thus, for example, if the
n 1 RuddlesdenPopper-phase NaLaTiO4 is first treated with
nitric acid to form HLaTiO4 and subsequently exposed to
KOH, the metastable, cation-substituted phase KLaTiO4 is
formed.285 It should be noted that the protonated intermediate
is an essential part of the reaction scheme as KLaTiO4 cannot be
made directly from NaLaTiO4 (by reaction with KNO3
for example) due to the loss of lattice energy on potassium
substitution. However, when KOH is reacted with HLaTiO4, the
formation of water due to the acid neutralization processes drives
the reaction forward as shown in reactions [52] and [53]:
NaLaTiO4 HNO3 ! HLaTiO4 NaNO3 [52]
HLaTiO4 KOH ! KLaTiO4 H2 O [53]
By substituting divalent cations for monovalent cations, it is
possible to convert An1BnO3n1 RuddlesdenPooper phases
into A0 An1BnO10 DionJacobson phases. Thus, the reaction of
K2Eu2Ti3O10 with Ca(NO3)2 or Sr(NO3)2 yields CaEu2Ti3O10
and SrEu2Ti3O10 according to reaction [54]286:
K2 Eu2 Ti3 O10 MII NO3 2 ! MII Eu2 Ti3 O10 2KNO3 [54]
In this instance the large Ca2 and Sr2 cations drive a shift
of the Eu2Ti3O10 perovskite blocks to adopt a DionJacobson
stacking sequence with large eight-coordinate A0 cation sites.
However, it is possible to exchange the monovalent cations
in RuddlesdenPopper phases for smaller divalent transition-
metal cations with retention of the RuddlesdenPopper frame-
work. For example, reaction of Na2La2Ti3O10 with a 1:1
mixture of CoCl2/KCl forms CoIILa2Ti3O10 according to reac-
tion [55], in which the small CoII cations occupy half the
tetrahedral interlayer cation sites as shown in Figure 39.287
Similar reactions can be performed to exchange CuII or ZnII
cations for sodium, in an analogous manner287:
CoCl2/KCl
50%
occupied
Na2La2Ti3O10 CoLa2Ti3O10
Figure 39 Reaction of the n 3 RuddlesdenPopper-phase
Na2La2Ti3O10 with CoCl2 in a KCl flux exchanges the nine-coordinate
sodium ions for tetrahedrally coordinated cobalt centers.
Na2 La2 Ti3 O10 CoCl2 ! CoII La2 Ti3 O10 2NaCl [55]
Charged metal-oxide layers can also be inserted into
RuddlesdenPopper phases via exchange reactions. If the
n 3 RuddlesdenPopper-phase K2La2Ti3O10 is reacted with
two equivalents of BiOCl, (Bi2O2)2 layers are inserted be-
tween the La2Ti3O10 perovskite sheets to yield the Aurivillius-
phase (Bi2O2)La2Ti3O10 according to reaction [56]288:
K2 La2 Ti3 O10 2BiOCl ! Bi2 O2 La2 Ti3 O10 2KCl [56]
By careful selection of the reacting salt, more elaborate
metal-oxide layers can be inserted. Thus, the reaction of
K2La2Ti3O10 with VOSO4 inserts (VO)2 into the host phase
with ejection of two potassium cations according to
reaction [57]288:
K2 La2 Ti3 O10 VOSO4 3H2 O
! VOLa2 Ti3 O10 K2 SO4 3H2 O [57]
Thus, it can be seen that when sufficiently mobile cations
are present, a wide range of substitution reactions can be
performed to prepare metastable phases.
2.15.4.5.2 Deintercalation
2.15.4.5.2.1 Dehydration reactions
The deintercalation of incorporated molecules is a common
feature of a wide range of hydrated oxide phases. However, the
dehydration reactions of solid-acid oxide phases can be uti-
lized to form new metaloxygen linkages between existing
MOn metaloxygen coordination polyhedra. For example, if
the solid-acid-phase HTiNbO5 is heated at 300  C, a dehydra-
tion reaction occurs in which water is formed and released
from the solid, leading to the synthesis of Ti2Nb2O9.289,290
The dehydration reaction drives the fusion of adjacent (Ti,
Nb)2O5 sheets via the formation of new corner-sharing links
between (Ti, Nb)O6 octahedra, resulting in a new 3D metal-
oxide network as shown in Figure 40. Similar dehydration
reactions are observed to occur when proton-exchanged tita-
nates in the H2TinO2n1 series are heated above 350  C, form-
ing a metastable polymorph of TiO2, referred to as TiO2(B),
although it should be noted that the dehydration reactions in
this instance are not topochemical.260,291,292
Following a similar dehydration strategy, layered double
hydroxide phases can be readily converted into dense
spinel oxides at low temperature. Thus by heating
Ni1/3Co1/3Al1/3(OH)2(CO3)x(NO3)ynH2O at 500  C, the
metastable spinel-phase NiCoAlO4 is formed.293 On heating
above 600  C, this phase decomposes by expelling NiO, dem-
onstrating the metastability of the spinel phase and further
HTiNbO5 Ti2Nb2O9
Figure 40 Dehydration of the layered solid-acid HTiNbO5 leads to the
formation of the 3D oxide-phase Ti2Nb2O9.

emphasizing the ability of low-temperature dehydration reac-
tions to prepare kinetically stabilized phases.
2.15.4.5.2.2 Layer extraction
Under suitable conditions, it is possible to extract charge neu-
tral salt units from extended oxide phases, thus generating
new extended lattices. For example, if the n 3 Ruddlesden
Popper-phase K2La2Ti3O10 is treated with two equivalents of
tetraphenylphosphonium bromide (PPh4Br), K2O is extracted
from the host phase according to reaction [58], to yield
La2Ti2O9 as shown in Figure 41.294 The potassium cations
are extracted from the rock salt layers of the host phase along
with oxide ions from within the Ti3O10 blocks:
K2 La2 Ti3 O10 2PPH4 Br ! La2 Ti3 O10 2KBr
PPh3 O PPh3 Ph2 [58]
An alternative route for the extraction of layers from ex-
tended oxide structures is to utilize the dehydration of solid-
acid phases described in Section 2.15.4.5.2.1. For example if
the n 3 solid-acid-phase H2La2Ti3O10 is heated to approxi-
mately 600  C, water is eliminated to yield La2Ti3O9.261 The
elimination of water is associated with a structural change in
which the TiO6 octahedra on either side of each rock salt sheet
fuse to form an A-cation-deficient perovskite structure, better
expressed as La2/3TiO3 (Figure 42). The large size of the La3
cations makes them effectively immobile during the dehydra-
tion process, and thus the lanthanum cations adopt a layered
arrangement within the A-cation sites of the perovskite prod-
uct, patterned by the cation arrangement of the Ruddlesden
Popper parent phase.
PPh4Br
50%
occupied
K2La2Ti3O10 La2Ti3O9
Figure 41 PPh4Br extracts K2O topochemically from the structure of
K2La2Ti3O10 to yield La2Ti3O9.
D Cu2O8 host lattice, so the intercalation is mildly oxidative in
H2La2Ti3O10 La2Ti3O9
Figure 42 Dehydration of H2La2Ti3O10 leads to the formation of the
cation-deficient perovskite La2/3TiO3 in which layered arrangement of the
La3 of the parent phase is retained.
Soft Chemistry Synthesis of Oxides 449
Thus, it can be seen that two different metastable phases of
identical composition, La2Ti3O9 (Figure 41) and La2/3TiO2
(Figure 42), can be formed by performing different sequences
of soft chemical reactions on the host K2La2Ti3O10 phase.
The analogous H2RE2Ti3O10 phases in which RE is a smaller
lanthanide (RE Nd, Sm, Eu, Gd, Dy) do not condense to
form perovskite structures on dehydration.261,280 Instead, the
RE2Ti3O9 phases formed retain layered structures in which
some of the small lanthanide cations have migrated into the
region between the perovskite blocks. On heating above
950  C, these layered dehydrated phases undergo transitions
to RE2Ti2O7 pyrochlore-type structures, demonstrating their
metastability.261
The n 2 RuddlesdenPopper-phase H2SrTa2O7 also un-
dergoes a dehydration/layer fusion reaction to form the
strontium-deficient perovskite-phase SrTa2O6, in which the
strontium cations are distributed randomly over the A-cation
sites of the perovskite lattice.281 On further heating, SrTa2O6
undergoes a structural reorganization to adopt a tetragonal
tungsten bronze structure, further demonstrating the metasta-
bility of cation-deficient perovskite phases.
2.15.4.5.3 Redox-neutral intercalation
2.15.4.5.3.1 Salt intercalation reactions
As described in Section 2.15.4.5.1.3, charged salt layers can be
substituted for metal cations, residing within layered complex
oxide host phases, to yield new hybrid materials. In analogous
processes, uncharged salt layers can be inserted into suitable
layered hosts in a redox-neutral manner. For example by
exploiting the weak interlayer bonding interactions within
the Bi2O2 layers of Aurivillius and related bismuth-containing
superconducting oxide materials, neutral salt fragments can be
inserted into Bi2Sr2Can1CunOy-type phases. Thus if the layered
n 2 phase Bi2Sr2CaCu2O8 is heated in the presence of silver
and iodine, AgI is readily intercalated within the Bi2O2 layers of
the cuprate phase to yield (AgI)Bi2Sr2CaCu2O8 according to
reaction [59]295:
1
Bi2 Sr2 CaCu2 O8 Ag I2 ! AgIBi2 Sr2 CaCu2 O8 [59]
2
The composition of the intercalated layer and the oxidation
state of silver have been confirmed by x-ray absorption mea-
surements, demonstrating that the intercalation reaction is
redox neutral with respect to the host lattice. Similar reactions
can be performed to insert HgI2 or HgBr2 into Bi2Sr2CaCu2O8
by heating the host phase with the appropriate mercury diha-
lide in the presence of iodine to act as a transport agent.296
However, x-ray absorption data collected from the resulting
(HgX2)0.5Bi2Sr2CaCu2O8 phases indicate that there is partial
charge transfer between the HgX2 guest layers and Bi2Sr2Ca-
this case.297
The intercalation reactions of n 3 Bi2Sr2Can1CunOy
phases are observed to proceed much more slowly than those
of n 2 and n 1 hosts. For example, Bi2Sr2Ca2Cu3O10 is
observed not to react with HgI2. However if Bi2Sr2Ca2Cu3O10
is first reacted with iodine to form IBi2Sr2Ca2Cu3O10 as
described in Section 2.15.4.4.2.4, reaction with AgI or HgI2
proceeds rapidly to yield (AgI)Bi2Sr2Ca2Cu3O10 or
(HgI2)0.5Bi2Sr2Ca2Cu3O10, respectively.298 Following such a
450 Soft Chemistry Synthesis of Oxides
two-step intercalation route also allows the intercalation of reactions in sequence, the directed synthesis of desired product
AuI3 into Bi2Sr2CaCu2O8. Thus if IBi2Sr2CaCu2O8 is heated phases can be undertaken in a much more controllable way
with a quarter of a mole equivalent of gold powder, than is possible via the high-temperature ceramic synthesis
(AuI3)0.25Bi2Sr2CaCu2O8 is formed.299 Raman spectroscopy routes traditionally used to prepare complex oxide phases.
indicates that AuI3 exists as a novel trigonal planar intercalated
molecule. Subsequent theoretical analysis suggests that AuIII is
unstable with respect to a strong JahnTeller distortion which
favors Y- or T-shaped molecules and that in fact the gold References
iodine species is AuI32 containing monovalent gold.300 1.
2.
2.15.4.5.3.2 Intercalation of organic bases 3.
Proton-exchanged complex oxide phases can act as Brnsted 4.
5.
solid acids facilitating the intercalation of organic bases to 6.
form organic/inorganic hybrid materials via an acid/base neu- 7.
tralization process. For example, if the proton-exchanged 8.
DionJacobson-phases HCa2Nan3NbnO3n1 (3  n  7) are 9. Livage, J.; Henry, M.; Sanchez, C. Prog. Solid State Chem. 1988, 18, 259.
10.
stirred with n-octylamine (CH3(CH2)7NH2) in hexane, the
11.
alkylamine is intercalated into the interlayer region of the 12.
oxide host phase, as demonstrated by a constant expansion
of the interlayer separation on intercalation, for all phases in 13.
the series.267 Similar base intercalation reactions can occur for
14.
a wide range of layered solid-acid phases, if the incorporated 15.
protons are sufficiently acidic. For example, the series of n 3
DionJacobson-phases HCa2xLaxNb3xTixO10 (0 < x < 2) have 16. Catania, P.; Hovnanian, N.; Cot, L. Mater. Res. Bull. 1990, 25, 1477.
Brnsted acidities which decrease with increasing x, such that 17.
the x 2 member of the series, HLa2Nb3Ti2O10, intercalates
18.
very strong bases only sparingly.279,301 19.
The stability of the inorganic/organic hybrid phases pre-
pared by the intercalation of long-chain alkylamines is thought 20.
to derive from favorable van der Waals forces which arise 21.
22.
from the efficient packing of the organic chains. However if 23.
short-chain or sterically bulky bases, which do not pack effi- 24.
ciently, are intercalated into layered solid acids, solvent can
penetrate the interlayer spaces, leading to the delamination of 25.
26.
the oxide host. Thus if HCa2Nb3O10 is treated with the bulky
amine tetra(n-butyl)ammonium hydroxide (TBA), sheets of 27.
TBAxH1xCa2Nb3O10 are exfoliated from the host phase and
suspended in solution.302,303 The exfoliated sheets have a net 28.
negative charge and can therefore be self-assembled onto a 29.
30.
positively charged surface, such as oxidized silicon treated 31.
with poly(diallyldimethylammonium chloride) (PDDA). The 32.
deposited monolayer, which is charge neutral, can then be
treated with PDDA and a further monolayer deposited on top 33.
of the first. This second layer need not be based on the same
34.
inorganic block as the first, allowing films with complex stack- 35.
ing sequences to be constructed. Thus by careful layer-by-layer 36.
synthesis, complex stacking sequences such as Ca2Nb3O10/
LaNb2O7/Sr2Nb3O10/LaNb2O7 (interleaved by organic bases) 37.
can be formed.303 This layer-by-layer deposition method al- 38.
lows the synthesis of films with complex stacking sequences
which could not be simply made by other synthesis routes. 39.
40.
2.15.5 Conclusion 41.
In summary, it can be seen that by applying a wide variety of 42.
low-temperature topochemical reactions, a large number of
43.
metastable complex oxide phases can be prepared, in which
the solid structure and metal oxidation states can be care- 44.
fully tuned and controlled. By applying such topochemical 45.
Gopalakrishnan, J. Chem. Mater. 1995, 7, 1265.
Schaak, R. E.; Mallouk, T. E. Chem. Mater. 2002, 14, 1455.
Stein, A.; Keller, S. W.; Mallouk, T. E. Science 1993, 259, 1558.
Disalvo, F. J. Science 1990, 247, 649.
Wells, A. F. Structural Inorganic Chemistry. Oxford University Press: Oxford, 1984.
Vidyasagar, K.; Gopalakrishnan, J.; Rao, C. N. R. Inorg. Chem. 1984, 23, 1206.
Horowitz, H. S.; Longo, J. M. Mater. Res. Bull. 1978, 13, 1359.
Vidyasagar, K.; Gopalakrishnan, J.; Rao, C. N. R. J. Solid State Chem. 1985, 58, 29.
Kakihana, M. J. Sol-Gel Sci. Techn. 1996, 6, 7.
Chandler, C. D.; Roger, C.; Hampdensmith, M. J. Chem. Rev. 1993, 93, 1205.
Campion, J. F.; Payne, D. A.; Chae, H. K.; Maurin, J. K.; Wilson, S. R. Inorg.
Chem. 1991, 30, 3244.
Kwei, G. H.; Lawson, A. C.; Billinge, S. J. L.; Cheong, S. W. J. Phys. Chem. 1993,
97, 2368.
Gustafson, R. L. J. Chem. Educ. 1960, 37, 603.
Fransaer, J.; Roos, J. R.; Delaey, L.; Vanderbiest, O.; Arkens, O.; Celis, J. P. J.
Appl. Phys. 1989, 65, 3277.
Maeda, S.; Tsurusaki, Y.; Tachiyama, Y.; Naka, K.; Ohki, A.; Ohgushi, T.;
Takeshita, T. J. Polym. Sci. Pol. Chem. 1994, 32, 1729.
Pechini M.P. (1967) Patent, U. S., Ed. USA, 3: 697.
Fjellvag, H.; Hansteen, O. H.; Tilset, B. G.; Olafsen, A.; Sakai, N.; Seim, H.
Thermochim. Acta 1995, 256, 75.
Sale, F. R.; Mahloojchi, F. Ceram. Int. 1988, 14, 229.
Walton, R. I. Chem. Soc. Rev. 2002, 31, 230.
Modeshia, D. R.; Walton, R. I. Chem. Soc. Rev. 2010, 39, 4303.
Rabenau, A. Angew. Chem. Int. Ed. 1985, 24, 1026.
Knauss, K. G.; Dibley, M. J.; Bourcier, W. L.; Shaw, H. F. Appl. Geochem. 2001,
16, 1115.
Dutta, P. K.; Gregg, J. R. Chem. Mater. 1992, 4, 843.
Wei, X.; Xu, G.; Ren, Z. H.; Wang, Y. G.; Shen, G.; Han, G. R. J. Cryst. Growth
2008, 310, 4132.
Ashbrook, S. E.; Le Polles, L.; Gautier, R.; Pickard, C. J.; Walton, R. I. Phys.
Chem. Chem. Phys. 2006, 8, 3423.
Goh, G. K. L.; Lange, F. F.; Haile, S. M.; Levi, C. G. J. Mater. Res. 2003, 18, 338.
Goh, G. K. L.; Levi, C. G.; Lange, F. F. J. Mater. Res. 2002, 17, 2852.
He, Y.; Zhu, Y. F.; Wu, N. Z. J. Solid State Chem. 2004, 177, 3868.
He, Y.; Zhu, Y. F.; Wu, N. Z. J. Solid State Chem. 2004, 177, 2985.
Liu, J. W.; Chen, G.; Li, Z. H.; Zhang, Z. G. Int. J. Hydrogen Energ. 2007, 32,
2269.
Santos, I.; Loureiro, L. H.; Silva, M. F. P.; Cavaleiro, A. M. V. Polyhedron 2002,
21, 2009.
Suchanek, W. L. Chem. Mater. 2004, 16, 1083.
Vasco, E.; Magrez, A.; Forro, L.; Setter, N. J. Phys. Chem. B 2005, 109, 14331.
Modeshia, D. R.; Darton, R. J.; Ashbrook, S. E.; Walton, R. I. Chem. Commun.
2009, 68.
Ju, J.; Wang, D. J.; Lin, J. H.; Li, G. B.; Chen, J.; You, L. P.; Liao, F. H.; Wu, N. Z.;
Huang, H. Z.; Yao, G. Q. Chem. Mater. 2003, 15, 3530.
Rao, C. N. R., Raveau, B., Eds.; In Colossal Magnetoresistance, Charge Ordering
and Related Properties of Manganese Oxides; World Scientific: Singapore, 1998.
Wang, Y. W.; Lu, X. Y.; Chen, Y.; Chi, F. L.; Feng, S. H.; Liu, X. Y. J. Solid State
Chem. 2005, 178, 1317.
Chen, Y.; Yuan, H. M.; Li, G. H.; Tian, G.; Feng, S. H. J. Cryst. Growth 2007, 305,
242.
Han, J. T.; Huang, Y. H.; Huang, W.; Goodenough, J. B. J. Am. Chem. Soc. 2006,
128, 14454.
Stampler, E. S.; Sheets, W. C.; Prellier, W.; Marks, T. J.; Poeppelmeier, K. R.
J. Mater. Chem. 2009, 19, 4375.
Chen, Y.; Yuan, H. M.; Tian, G.; Zhang, G. H.; Feng, S. H. J. Solid State Chem.
2007, 180, 1340.
Christen, A. N.; Ollivier, G. J. Solid State Chem. 1972, 4, 131.
Spooren, J.; Walton, R. I. J. Solid State Chem. 2005, 178, 1683.

46. Cussen, E. J.; Battle, P. D. Chem. Mater. 2000, 12, 831.
47. Li, C. P.; Li, T.; Wang, B.; Yan, H. J. Cryst. Growth 2006, 295, 137.
48. Liu, J. B.; Wang, H.; Zhu, M. K.; Wang, B.; Yan, H. Mater. Res. Bull. 2003,
38, 817.
49. Li, J. Q.; Sun, W. A.; Ao, W. Q.; Tang, J. N. J. Magn. Magn. Mater. 2006, 302, 463.
50. Urban, J. J.; Ouyang, L.; Jo, M. H.; Wang, D. S.; Park, H. Nano Lett. 2004,
4, 1547.
51. Chen, Y.; Yuan, H. M.; Tian, G.; Zhang, G. H.; Feng, S. H. J. Solid State Chem.
2007, 180, 167.
52. Millange, F.; Caignaert, V.; Domenges, B.; Raveau, B.; Suard, E. Chem. Mater.
1998, 10, 1974.
53. Spooren, J.; Rumplecker, A.; Millange, F.; Walton, R. I. Chem. Mater. 2003,
15, 1401.
54. Spooren, J.; Walton, R. I.; Millange, F. J. Mater. Chem. 2005, 15, 1542.
55. Laurent, Y. Rev. Chim. Miner. 1969, 6, 1145.
56. Mugavero, S. J.; Gemmill, W. R.; Roof, I. P.; zur Loye, H. C. J. Solid State Chem.
1950, 2009, 182.
57. Perry, D. L.; Phillips, S. L. Hand Book of Inorganic Compounds. CRC Press:
Orinda, CA, 1995.
58. Ham, W. K.; Holland, G. F.; Stacy, A. M. J. Am. Chem. Soc. 1988, 110, 5214.
59. Lux, H.; Kuhn, R.; Niedermaier, T. Z. Anorg. Allg. Chem. 1959, 298, 285.
60. Mugavero, S. J.; Smith, M. D.; Yoon, W. S.; zur Loye, H. C. Angew. Chem. Int. Ed.
2009, 48, 215.
61. Jung, D. Y.; Demazeau, G. J. Solid State Chem. 1995, 115, 447.
62. Reisner, B. A.; Stacy, A. M. J. Am. Chem. Soc. 1998, 120, 9682.
63. Zhao, Q. B.; Smith, M. D.; zur Loye, H. C. J. Chem. Crystallogr. 2011, 41, 674.
64. Campa, J. A.; Gutierrezpuebla, E.; Monge, M. A.; Rasines, I.; Ruizvalero, C. J.
Solid State Chem. 1994, 108, 230.
65. Dicarlo, J.; Yazdi, I.; Jacobson, A. J.; Navrotsky, A. J. Solid State Chem. 1994,
109, 223.
66. Dann, S. E.; Weller, M. T.; Currie, D. B.; Thomas, M. F.; Alrawwas, A. D. J. Mater.
Chem. 1993, 3, 1231.
67. Takeda, Y.; Shimada, M.; Kanamaru, F.; Koizumi, M.; Yamamoto, N. Chem. Lett.
1974, 107.
68. Cava, R. J.; Santoro, A.; Murphy, D. W.; Zahurak, S. M.; Roth, R. S. J. Solid State
Chem. 1983, 50, 121.
69. Cava, R. J.; Santoro, A.; Murphy, D. W.; Zahurak, S.; Roth, R. S. Solid State Ionics
1981, 5, 323.
70. Murphy, D. W.; Greenblatt, M.; Cava, R. J.; Zahurak, S. M. Solid State Ionics
1981, 5, 327.
71. Cava, R. J.; Santoro, A.; Murphy, D. W.; Zahurak, S.; Roth, R. S. J. Solid State
Chem. 1982, 42, 251.
72. Cheng, K. H.; Whittingham, M. S. Solid State Ionics 1980, 1, 151.
73. Straumanis, M. E. J. Am. Chem. Soc. 1949, 71, 679.
74. Kimizuka, N.; Akahane, T.; Matsumoto, S.; Yukino, K. Inorg. Chem. 1976,
15, 3178.
75. Horiuchi, S.; Kimizuka, N.; Yamamoto, A. Nature 1979, 279, 226.
76. Dickens, P. G.; Weller, M. T. J. Solid State Chem. 1983, 48, 407.
77. Weller, M. T.; Dickens, P. G. Solid State Ionics 1983, 910, 1081.
78. Dickens, P. G.; Moore, J. H.; Neild, D. J. J. Solid State Chem. 1973, 7, 241.
79. Wiseman, P. J.; Dickens, P. G. J. Solid State Chem. 1973, 6, 374.
80. Murphy, D. W.; Disalvo, F. J.; Carides, J. N.; Waszczak, J. V. Mater. Res. Bull.
1978, 13, 1395.
81. Cox, D. E.; Cava, R. J.; McWhan, D. B.; Murphy, D. W. J. Phys. Chem. Solids
1982, 43, 657.
82. Davidson, I. J.; Greedan, J. E. J. Solid State Chem. 1984, 51, 104.
83. Thackeray, M. M.; David, W. I. F.; Goodenough, J. B. Mater. Res. Bull. 1982,
17, 785.
84. Thackeray, M. M.; David, W. I. F.; Bruce, P. G.; Goodenough, J. B. Mater. Res.
Bull. 1983, 18, 461.
85. Thackeray, M. M.; Baker, S. D.; Adendorff, K. T.; Goodenough, J. B. Solid State
Ionics 1985, 17, 175.
86. Cava, R. J.; Murphy, D. W.; Zahurak, S.; Santoro, A.; Roth, R. S. J. Solid State
Chem. 1984, 53, 64.
87. David, W. I. F.; Thackeray, M. M.; Depicciotto, L. A.; Goodenough, J. B. J. Solid
State Chem. 1987, 67, 316.
88. Depicciotto, L. A.; Thackeray, M. M. Mater. Res. Bull. 1985, 20, 1409.
89. Schollhorn, R.; Kuhlmann, R.; Besenhard, J. O. Mater. Res. Bull. 1976, 11, 83.
90. Tagaya, H.; Ara, K.; Kadokawa, J.; Karasu, M.; Chiba, K. J. Mater. Chem. 1994,
4, 551.
91. Birtill, J. J.; Dickens, P. G. Mater. Res. Bull. 1978, 13, 311.
92. Glemser, O.; Lutz, G. Z. Anorg. Allg. Chem. 1951, 264, 17.
93. Schroder, F. A.; Weitzel, H. Z. Anorg. Allg. Chem. 1977, 435, 247.
Soft Chemistry Synthesis of Oxides 451
94. Tinet, D.; Canesson, P.; Estrade, H.; Fripiat, J. J. J. Phys. Chem. Solids 1980,
41, 583.
95. Dickens, P. G.; Birtill, J. J.; Wright, C. J. J. Solid State Chem. 1979, 28, 185.
96. Dickens, P. G.; Short, A. T.; Crouchbaker, S. Solid State Ionics 1988, 28, 1294.
97. Slade, R. C. T.; Halstead, T. K.; Dickens, P. G. J. Solid State Chem. 1980, 34, 183.
98. Dickens, P. G.; French, S. J.; Hight, A. T.; Pye, M. F. Mater. Res. Bull. 1979, 14, 1295.
99. Murphy, D. W.; Christian, P. A.; Disalvo, F. J.; Waszczak, J. V. Inorg. Chem.
1979, 18, 2800.
100. Cava, R. J.; Santoro, A.; Murphy, D. W.; Zahurak, S. M.; Fleming, R. M.;
Marsh, P.; Roth, R. S. J. Solid State Chem. 1986, 65, 63.
101. Dickens, P. G.; Chippindale, A. M.; Hibble, S. J.; Lancaster, P. Mater. Res. Bull.
1984, 19, 319.
102. Yoshikawa, A.; Yagisawa, K.; Shimoda, M. J. Mater. Sci. 1994, 29, 1319.
103. Shimoda, M.; Yoshikawa, A.; Yagisawa, K. J. Mater. Sci. 1994, 29, 478.
104. Takano, Y.; Takayanagi, S.; Ogawa, S.; Yamadaya, T.; Mori, N. Solid State
Commun. 1997, 103, 215.
105. Nagai, I.; Abe, Y.; Kato, M.; Koike, Y.; Kakihana, M. Physica C 2001, 357, 393.
106. Nagai, I.; Abe, Y.; Kato, M.; Koike, Y.; Kakihana, A. Solid State Ionics 2002, 151,
265.
107. Kato, M.; Inoue, A.; Nagai, I.; Kakihana, M.; Sleight, A. W.; Koike, Y. Physica C
2003, 388, 445.
108. Armstrong, A. R.; Anderson, P. A. Inorg. Chem. 1994, 33, 4366.
109. Choi, J.; Zhang, X.; Wiley, J. B. Inorg. Chem. 2009, 48, 4811.
110. Byeon, S. H.; Kim, H. J.; Kim, D. K.; Hur, N. H. Chem. Mater. 2003, 15, 383.
111. Toda, K.; Takahashi, M.; Teranishi, T.; Ye, Z. G.; Sato, M.; Hinatsu, Y. J. Mater.
Chem. 1999, 9, 799.
112. Toda, K.; Teranishi, T.; Takahashi, M.; Ye, Z. G.; Sato, M. Solid State Ionics 1998,
113, 501.
113. Toda, K.; Teranishi, T.; Ye, Z. G.; Sato, M.; Hinatsu, Y. Mater. Res. Bull. 1999,
34, 971.
114. Bohnke, C.; Bohnke, O.; Fourquet, J. L. J. Electrochem. Soc. 1997, 144, 1151.
115. McIntyre, R. A.; Falster, A. U.; Li, S. C.; Simmons, W. B.; OConnor, C. J.;
Wiley, J. B. J. Am. Chem. Soc. 1998, 120, 217.
116. Choy, J. H.; Kim, J. Y.; Chung, I. J. Phys. Chem. B 2001, 105, 7908.
117. Mizushima, K.; Jones, P. C.; Wiseman, P. J.; Goodenough, J. B. Mater. Res. Bull.
1980, 15, 783.
118. Gupta, R.; Manthiram, A. J. Solid State Chem. 1996, 121, 483.
119. Takahashi, Y.; Kijima, N.; Tokiwa, K.; Watanabe, T.; Akimoto, J. J. Phys. Condens.
Matter 2007, 19, 436202.
120. Amatucci, G. G.; Tarascon, J. M.; Klein, L. C. J. Electrochem. Soc. 1996, 143, 1114.
121. Motohashi, T.; Katsumata, Y.; Ono, T.; Kanno, R.; Karppinen, M.; Yamauchi, H.
Chem. Mater. 2007, 19, 5063.
122. Hertz, J. T.; Huang, Q.; McQueen, T.; Klimczuk, T.; Bos, J. W. G.; Viciu, L.;
Cava, R. J. Phys. Rev. B 2008, 77, 075119.
123. Motohashi, T.; Ono, T.; Sugimoto, Y.; Masubuchi, Y.; Kikkawa, S.; Kanno, R.;
Karppinen, M.; Yamauchi, H. Phys. Rev. B 2009, 80, 165114.
124. Reimers, J. N.; Dahn, J. R. J. Electrochem. Soc. 1992, 139, 2091.
125. Van der Ven, A.; Aydinol, M. K.; Ceder, G. J. Electrochem. Soc. 1998,
145, 2149.
126. Van der Ven, A.; Aydinol, M. K.; Ceder, G.; Kresse, G.; Hafner, J. Phys. Rev. B
1998, 58, 2975.
127. Fouassie, C.; Matejka, G.; Reau, J. M.; Hagenmul, P. J. Solid State Chem. 1973,
6, 532.
128. Viciu, L.; Bos, J. W. G.; Zandbergen, H. W.; Huang, Q.; Foo, M. L.; Ishiwata, S.;
Ramirez, A. P.; Lee, M.; Ong, N. P.; Cava, R. J. Phys. Rev. B 2006,
73, 174104.
129. Delmas, C.; Braconnier, J. J.; Fouassier, C.; Hagenmuller, P. Solid State Ionics
1981, 34, 165.
130. Huang, Q.; Foo, M. L.; Pascal, R. A.; Lynn, J. W.; Toby, B. H.; He, T.;
Zandbergen, H. W.; Cava, R. J. Phys. Rev. B 2004, 70, 184110.
131. Poltavets, V. V.; Croft, M.; Greenblatt, M. Phys. Rev. B 2006, 74, 125103.
132. Viciu, L.; Huang, Q.; Cava, R. J. Phys. Rev. B 2006, 73, 212107.
133. Foo, M. L.; Wang, Y. Y.; Watauchi, S.; Zandbergen, H. W.; He, T.; Cava, R. J.;
Ong, N. P. Phys. Rev. Lett. 2004, 92, 247001.
134. Terasaki, I.; Sasago, Y.; Uchinokura, K. Phys. Rev. B 1997, 56, 12685.
135. Wang, Y. Y.; Rogado, N. S.; Cava, R. J.; Ong, N. P. Nature 2003, 423, 425.
136. Williams, A. J.; Attfield, J. P.; Foo, M. L.; Viciu, L.; Cava, R. J. Phys. Rev. B 2006,
7, 134401.
137. Takada, K.; Sakurai, H.; Takayama-Muromachi, E.; Izumi, F.; Dilanian, R. A.;
Sasaki, T. Nature 2003, 422, 53.
138. Foo, M. L.; Schaak, R. E.; Miller, V. L.; Klimczuk, T.; Rogado, N. S.; Wang, Y.;
Lau, G. C.; Craley, C.; Zandbergen, H. W.; Ong, N. P.; Cava, R. J. Solid State
Commun. 2003, 127, 33.
452 Soft Chemistry Synthesis of Oxides
139. Lorenz, B.; Cmaidalka, J.; Meng, R. L.; Chu, C. W. Phys. Rev. B 2003, 68,
014512.
140. Schaak, R. E.; Klimczuk, T.; Foo, M. L.; Cava, R. J. Nature 2003, 424, 527.
141. Arai, H.; Tsuda, M.; Saito, K.; Hayashi, M.; Takei, K.; Sakurai, Y. J. Solid State
Chem. 2002, 163, 340.
142. Croguennec, L.; Pouillerie, C.; Delmas, C. Solid State Ionics 2000, 135, 259.
143. Tarascon, J. M.; Vaughan, G.; Chabre, Y.; Seguin, L.; Anne, M.; Strobel, P.;
Amatucci, G. J. Solid State Chem. 1999, 147, 410.
144. Vidyasagar, K.; Gopalakrishnan, J. J. Solid State Chem. 1982, 42, 217.
145. Depicciotto, L. A.; Thackeray, M. M.; David, W. I. F.; Bruce, P. G.;
Goodenough, J. B. Mater. Res. Bull. 1984, 19, 1497.
146. Thackeray, M. M.; Depicciotto, L. A.; David, W. I. F.; Bruce, P. G.;
Goodenough, J. B. J. Solid State Chem. 1987, 67, 285.
147. Depicciotto, L. A.; Thackeray, M. M. Mater. Res. Bull. 1985, 20, 187.
148. Shao-Horn, Y.; Ein-Eli, Y.; Robertson, A. D.; Averill, W. F.; Hackney, S. A.;
Howard, W. F. J. Electrochem. Soc. 1998, 145, 16.
149. Hunter, J. C. J. Solid State Chem. 1981, 39, 142.
150. Greedan, J. E.; Raju, N. P.; Wills, A. S.; Morin, C.; Shaw, S. M.; Reimers, J. N.
Chem. Mater. 1998, 10, 3058.
151. Robinson, D. M.; Go, Y. B.; Greenblatt, M.; Dismukes, G. C. J. Am. Chem. Soc.
2010, 132, 11467.
152. Akimoto, J.; Takahashi, Y.; Gotoh, Y.; Kawaguchi, K.; Dokko, K.; Uchida, I. Chem.
Mater. 2003, 15, 2984.
153. Tang, W. P.; Kanoh, H.; Ooi, K. J. Solid State Chem. 1999, 142, 19.
154. Gummow, R. J.; Liles, D. C.; Thackeray, M. M. Mater. Res. Bull. 1993, 28, 1249.
155. Akimoto, J.; Gotoh, Y.; Sohma, M.; Kawaguchi, K.; Oosawa, Y.; Takei, H. J. Solid
State Chem. 1994, 110, 150.
156. Akimoto, J.; Gotoh, Y.; Oosawa, Y.; Nonose, N.; Kumagai, T.; Aoki, K.; Takei, H. J.
Solid State Chem. 1994, 113, 27.
157. Takahashi, Y.; Kijima, N.; Akimoto, J. Chem. Mater. 2006, 18, 748.
158. Latroche, M.; Brohan, L.; Marchand, R.; Tournoux, M. J. Solid State Chem. 1989,
81, 78.
159. Watanabe, M.; Komatsu, Y.; Sasaki, T.; Fujiki, Y. J. Solid State Chem. 1991,
92, 80.
160. Anderson, M. T.; Vaughey, J. T.; Poeppelmeier, K. R. Chem. Mater. 1993, 5, 151.
161. Poeppelmeier, K. R.; Leonowicz, M. E.; Longo, J. M. J. Solid State Chem. 1982,
44, 89.
162. Poeppelmeier, K. R.; Leonowicz, M. E.; Scanlon, J. C.; Longo, J. M.; Yelon, W. B.
J. Solid State Chem. 1982, 45, 71.
163. Chiang, C. C. K.; Poeppelmeier, K. R. Mater. Lett. 1991, 12, 102.
164. Rormark, L.; Wiik, K.; Stolen, S.; Grande, T. J. Mater. Chem. 2002, 12, 1058.
165. MacChesney, J. B.; Williams, H. J.; Potter, J. F.; Sherwood, R. C. Phys. Rev.
1967, 164, 779.
166. Schiffer, P.; Ramirez, A. P.; Bao, W.; Cheong, S.-W. Phys. Rev. Lett. 1995,
75, 3336.
167. Hansteen, O. H.; Fjellvag, H.; Hauback, B. C. J. Solid State Chem. 1998,
141, 411.
168. Hansteen, O. H.; Fjellvag, H.; Hauback, B. C. J. Mater. Chem. 1998, 8, 2081.
169. Hayward, M. A.; Green, M. A.; Rosseinsky, M. J.; Sloan, J. J. Am. Chem. Soc.
1999, 121, 8843.
170. Hayward, M. A.; Cussen, E. J.; Claridge, J. B.; Bieringer, M.; Rosseinsky, M. J.;
Kiely, C. J.; Blundell, S. J.; Marshall, I. M.; Pratt, F. L. Science 2002, 295, 1882.
171. Adkin, J. J.; Hayward, M. A. Inorg. Chem. 2008, 47, 10959.
172. Alonso, J. A.; Martinez-Lope, M. J. J. Chem. Soc. Dalton Trans. 1995, 2819.
173. Siegrist, T.; Zahurak, S. M.; Murphy, D. W.; Roth, R. S. Nature 1988,
334, 231.
174. Lee, K. W.; Pickett, W. E. Phys. Rev. B 2004, 70, 045104.
175. OMalley, M.; Lockett, M. A.; Hayward, M. A. J. Solid State Chem. 2007,
180, 2851.
176. Hadermann, J.; Abakumov, A.; Adkin, J. J.; Hayward, M. A. J. Am. Chem. Soc.
2009, 131, 10598.
177. Helps, R. M.; Rees, N. H.; Hayward, M. A. Inorg. Chem. 2010, 49, 11062.
178. Suescun, L.; Dabrowski, B.; Mais, J.; Remsen, S.; Richardson, J. W.; ER, M.;
Jorgensen, J. D. Chem. Mater. 2008, 20, 1636.
179. Suescun, L.; Dabrowski, D.; Remsen, S.; Mais, J. J. Solid State Chem. 2009,
182, 187.
180. Mori, T.; Kamegashira, N. J. Alloy Compd. 2006, 408412, 1210.
181. Trukhanov, S. V.; Kasper, N. V.; Troyanchuk, I. O.; Tovar, M.; Szymczak, H.;
Barner, K. J. Solid State Chem. 2002, 169, 85.
182. Gonzalez-Calbet, J. M.; Herrero, E.; Rangavittal, N.; Alonso, J. M.; Martinez, J. L.;
Vallet-Regi, M. J. Solid State Chem. 1999, 148, 158.
183. Cortes-Gil, R.; Ruiz-Gonzalez, M. L.; Alonso, J. M.; Vallet-Regi, M.;
Hernando, M.; Gonzalez Calbet, J. M. Chem.-Eur. J. 2007, 13, 4246.
184. Hansteen, O. H.; Breard, Y.; Fjellvag, H.; Hauback, B. C. Solid State Sci. 2004,
6, 279.
185. Ruiz-Gonzalez, M. L.; Cortes-Gil, R.; Alonso, J. M.; Hernando, A.; Vallet-Regi, M.;
Gonzalez-Calbet, J. M. Chem. Mater. 2006, 18, 5756.
186. Parsons, T. G.; DHondt, H.; Hadermann, J.; Hayward, M. A. Chem. Mater. 2009,
21, 5527.
187. Tsujimoto, Y.; Tassel, C.; Hayashi, N.; Watanabe, T.; Kageyama, H.; Yoshimura, K.;
Takano, M.; Ceretti, M.; Ritter, C.; Paulus, W. Nature 2007, 450, 1062.
188. Hayward, M. A.; Rosseinsky, M. J. Nature 2007, 450, 960.
189. Seinberg, L.; Yamamoto, T.; Tassel, C.; Kobayashi, Y.; Hayashi, N.; Kitada, A.;
Sumida, Y.; Watanabe, T.; Nishi, M.; Ohoyama, K.; Yoshimura, K.; Takano, M.;
Paulus, W.; Kageyama, H. Inorg. Chem. 2011, 50, 3988.
190. Kawakami, T.; Tsujimoto, Y.; Kageyama, H.; Chen, X. Q.; Fu, C. L.; Tassel, C.;
Kitada, A.; Suto, S.; Hirama, K.; Sekiya, Y.; Makino, Y.; Okada, T.; Yagi, T.;
Hayashi, N.; Yoshimura, K.; Nasu, S.; Podloucky, R.; Takano, M. Nat. Chem.
2009, 1, 371.
191. Leonowicz, M. E.; Poeppelmeier, K. R.; Longo, J. M. J. Solid State Chem. 1985,
59, 71.
192. Gillie, L. J.; Wright, A. J.; Hadermann, J.; Van Tendeloo, G.; Greaves, C. J. Solid
State Chem. 2002, 167, 145.
193. Kitchen, H. J.; Saratovsky, I.; Hayward, M. A. Dalton Trans. 2010, 39, 6098.
194. Hansteen, O. H.; Fjellvag, H.; Hauback, B. C. J. Mater. Chem. 1998, 8, 2089.
195. Hayward, M. A.; Rosseinsky, M. J. Chem. Mater. 2000, 12, 2182.
196. Bridges, C. A.; Darling, G. R.; Hayward, M. A.; Rosseinsky, M. J. J. Am. Chem.
Soc. 2005, 127, 5996.
197. Bowman, A.; Claridge, J. B.; Rosseinsky, M. J. Chem. Mater. 2006, 18, 3046.
198. Dann, S. E.; Weller, M. T.; Currie, D. B. J. Solid State Chem. 1992, 97, 179.
199. Kageyama, H.; Watanabe, T.; Tsujimoto, Y.; Kitada, A.; Sumida, Y.; Kanamori, K.;
Yoshimura, K.; Hayashi, N.; Muranaka, S.; Takano, M.; Ceretti, M.; Paulus, W.;
Ritter, C.; Andre, G. Angew. Chem. 2008, 47, 5704.
200. Poltavets, V. V.; Lokshin, K. A.; Dikmen, S.; Croft, M.; Egami, T.; Greenblatt, M. J.
Am. Chem. Soc. 2006, 128, 9050.
201. Hayward, M. A. Chem. Mater. 2006, 18, 321.
202. Millburn, J. E.; Mitchell, J. F. Chem. Mater. 2001, 13, 1957.
203. Vogt, T.; Woodward, P. M.; Karen, P.; Hunter, B. A.; Henning, P.;
Moodenbaugh, A. R. Phys. Rev. Lett. 2000, 84, 2969.
204. Millange, F.; Suard, E.; Caignaert, V.; Raveau, B. Mater. Res. Bull. 1999, 34, 1.
205. Caignaert, V.; Millange, F.; Domenges, B.; Raveau, B. Chem. Mater. 1999, 11, 930.
206. Woodward, P. M.; Karen, P. Inorg. Chem. 2003, 42, 1121.
207. Karen, P.; Suard, E.; Fauth, F.; Woodward, P. M. Solid State Sci. 2004, 6, 1195.
208. Seddon, J.; Suard, E.; Hayward, M. A. J. Am. Chem. Soc. 2010, 132, 2802.
209. Tofield, B. C.; Scott, W. R. J. Solid State Chem. 1974, 10, 183.
210. Negas, T.; Roth, R. S. J. Solid State Chem. 1970, 1, 409.
211. Battle, P. D.; Gibb, T. C.; Jones, C. W. J. Solid State Chem. 1988, 74, 60.
212. Chmaissem, O.; Dabrowski, B.; Kolesnik, S.; Mais, J.; Brown, D. E.; Kruk, R.;
Prior, P.; Pyles, B.; Jorgensen, J. D. Phys. Rev. B 2001, 64, 134412.
213. Caignaert, V.; Nguyen, N.; Hervieu, M.; Raveau, B. Mater. Res. Bull. 1985, 20, 479.
214. Lobanov, M. V.; Abakumov, A. M.; Sidorova, A. V.; Rozova, M. G.;
DYachenko, O. G.; Antipov, E. V.; Hadermann, J.; Van Tendeloo, G. Solid State
Sci. 2002, 4, 19.
215. Takeda, Y.; Kanno, K.; Takada, T.; Yamamoto, O.; Takano, M.; Nakayama, N.;
Bando, Y. J. Solid State Chem. 1986, 63, 237.
216. Hodges, J. P.; Short, S.; Jorgensen, J. D.; Xiong, X.; Dabrowski, B.; Mini, S. M.;
Kimball, C. W. J. Solid State Chem. 2000, 151, 190.
217. Wattiaux, A.; Fournes, L.; Demourgues, A.; Bernaben, N.; Grenier, J. C.;
Pouchard, M. Solid State Commun. 1991, 77, 489.
218. Berry, F. J.; Ren, X. L.; Heap, R.; Slater, P.; Thomas, M. F. Solid State Commun.
2005, 134, 621.
219. Rautama, E. L.; Boullay, P.; Kundu, A. K.; Caignaert, V.; Pralong, V.;
Karppinen, M.; Raveau, B. Chem. Mater. 2008, 20, 2742.
220. Abakumov, A. M.; Rozova, M. G.; Pavlyuk, B. P.; Lobanov, M. V.; Antipov, E. V.;
Lebedev, O. I.; Van Tendeloo, G.; Sheptyakov, D. V.; Balagurov, A. M.; Bouree, F.
J. Solid State Chem. 2001, 158, 100.
221. Sheptyakov, D. V.; Abakumov, A. M.; Antipov, E. V.; Balagurov, A. M.;
Billinge, S. J. L.; Fischer, P.; Keller, L.; Lobanov, M. V.; Pavlyuk, B. P.;
Pomjakushin, V. Y.; Rozova, M. G. Appl. Phys. A-Mater. 2002, 74, S1734.
222. Abakumov, A. M.; Rozova, M. G.; Pavlyuk, B. P.; Lobanov, M. V.; Antipov, E. V.;
Lebedev, O. I.; Van Tendeloo, G.; Ignatchik, O. L.; Ovtchenkov, E. A.;
Koksharov, Y. A.; Vasilev, A. N. J. Solid State Chem. 2001, 160, 353.
223. Pomjakushin, V. Y.; Balagurov, A. M.; Elzhov, T. V.; Sheptyakov, D. V.;
Fischer, P.; Khomskii, D. I.; Yushankhai, V. Y.; Abakumov, A. M.; Rozova, M. G.;
Antipov, E. V.; Lobanov, M. V.; Billinge, S. J. L. Phys. Rev. B 2002,
66, 184412.

224. Abakumov, A. M.; Rozova, M. G.; Alekseeva, A. M.; Kovba, M. L.; Antipov, E. V.;
Lebedev, O. I.; Van Tendeloo, G. Solid State Sci. 2003, 5, 871.
225. Antipov, E. V.; Abakumov, A. M.; Istomin, S. Y. Inorg. Chem. 2008, 47, 8543.
226. Vente, J. F.; Kamenev, K. V.; Sokolov, D. A. Phys. Rev. B 2001, 64, 214403.
227. Mitchell, J. F.; Millburn, J. E.; Medarde, M.; Short, S.; Jorgensen, J. D.;
Fernandez-Diaz, M. T. J. Solid State Chem. 1998, 141, 599.
228. Dann, S. E.; Weller, M. T. J. Solid State Chem. 1995, 115, 499.
229. Hill, J. M.; Dabrowski, B.; Mitchell, J. F.; Jorgensen, J. D. Phys. Rev. B 2006, 74,
174417.
230. Girgsdies, F.; Schollhorn, R. Solid State Commun. 1994, 91, 111.
231. Jorgensen, J. D.; Dabrowski, B.; Pei, S.; Richards, D. R.; Hinks, D. G. Phys. Rev. B
1989, 40, 2187.
232. Jorgensen, J. D.; Dabrowski, B.; Pei, S. Y.; Hinks, D. G.; Soderholm, L.;
Morosin, B.; Schirber, J. E.; Venturini, E. L.; Ginley, D. S. Phys. Rev. B 1988, 38,
11337.
233. Nemudry, A.; Rudolf, P.; Schollhorn, R. Solid State Ionics 1998, 109, 213.
234. Greaves, C.; Francesconi, M. G. Curr. Opin. Solid State Mater. Sci. 1998, 3, 132.
235. McCabe, E. E.; Greaves, C. J. Fluorine Chem. 2007, 128, 448.
236. Aikens, L. D.; Gillie, L. J.; Li, R. K.; Greaves, C. J. Mater. Chem. 2002, 12, 264.
237. Sivakumar, T.; Wiley, J. B. Mater. Res. Bull. 2009, 44, 74.
238. Aikens, L. D.; Li, R. K.; Greaves, C. Chem. Commun. 2000, 2129.
239. Slater, P. R.; Hodges, J. P.; Francesconi, M. G.; Edwards, P. P.; Greaves, C.;
Gameson, I.; Slaski, M. Physica C 1995, 253, 16.
240. Li, R. K.; Greaves, C. Phys. Rev. B 2000, 62, 3811.
241. Francesconi, M. G.; Slater, P. R.; Hodges, J. P.; Greaves, C.; Edwards, P. P.; Al-
Mamouri, M.; Slaski, M. J. Solid State Chem. 1998, 135, 17.
242. Almamouri, M.; Edwards, P. P.; Greaves, C.; Slaski, M. Nature 1994, 369, 382.
243. Slater, P. R.; Edwards, P. P.; Greaves, C.; Gameson, I.; Francesconi, M. G.;
Hodges, J. P.; Almamouri, M.; Slaski, M. Physica C 1995, 241, 151.
244. Almamouri, M.; Edwards, P. P.; Greaves, C.; Slater, P. R.; Slaski, M. J. Mater.
Chem. 1995, 5, 913.
245. Xiang, X. D.; McKernan, S.; Vareka, W. A.; Zettl, A.; Corkill, J. L.; Barbee, T. W.;
Cohen, M. L. Nature 1990, 348, 145.
246. Huong, P. V.; Verma, A. L. Phys. Rev. B 1993, 48, 9869.
247. Scarfe, D. P.; Bhavaraju, S.; Jacobson, A. J. Chem. Commun. 1997, 313.
248. Scarfe, D. P.; Jacobson, A. J. Chem. Mater. 1997, 9, 3107.
249. England, W. A.; Goodenough, J. B.; Wiseman, P. J. J. Solid State Chem. 1983,
49, 289.
250. Shirane, T.; Kanno, R.; Kawamoto, Y.; Takeda, Y.; Takano, M.; Kamiyama, T.;
Izumi, F. Solid State Ionics 1995, 79, 227.
251. Shannon, R. D.; Rogers, D. B.; Prewitt, C. T. Inorg. Chem. 1971, 10, 713.
252. Cushing, B. L.; Falster, A. U.; Simmons, W. B.; Wiley, J. B. Chem. Commun.
1996, 2635.
253. Cushing, B. L.; Wiley, J. B. J. Solid State Chem. 1998, 141, 385.
254. Blesa, M. C.; Amador, U.; Moran, E.; Menendez, N.; Tornero, J. D.;
Rodriguezcarvajal, J. Solid State Ionics 1993, 635, 429.
255. Blesa, M. C.; Moran, E.; Amador, U.; Andersen, N. H. J. Solid State Chem. 1997,
129, 123.
256. Eda, K.; Suzuki, M.; Hatayama, F.; Sotani, N. J. Mater. Chem. 1997, 7, 821.
257. Izawa, H.; Kikkawa, S.; Koizumi, M. J. Phys. Chem. 1982, 86, 5023.
258. Marchand, R.; Brohan, L.; Tournoux, M. Mater. Res. Bull. 1980, 15, 1129.
259. Dion, M.; Piffard, Y.; Tournoux, M. J. Inorg. Nucl. Chem. 1978, 40, 917.
260. Feist, T. P.; Davies, P. K. J. Solid State Chem. 1992, 101, 275.
261. Gopalakrishnan, J.; Bhat, V. Inorg. Chem. 1987, 26, 4299.
262. Gopalakrishnan, J.; Bhat, V.; Raveau, B. Mater. Res. Bull. 1987, 22, 413.
263. Toda, K.; Sato, M. J. Mater. Chem. 1996, 6, 1067.
264. Toda, K.; Suzuki, T.; Sato, M. Solid State Ionics 1996, 93, 177.
Soft Chemistry Synthesis of Oxides 453
265. Mahler, C. H.; Cushing, B. L.; Lalena, J. N.; Wiley, J. B. Mater. Res. Bull. 1998,
33, 1581.
266. Jacobson, A. J.; Lewandowski, J. T.; Johnson, J. W. J. Less-Common Met. 1986,
116, 137.
267. Jacobson, A. J.; Johnson, J. W.; Lewandowski, J. T. Inorg. Chem. 1985,
24, 3727.
268. Uma, S.; Gopalakrishnan, J. Chem. Mater. 1994, 6, 907.
269. Palacin, M. R.; Lira, M.; Garcia, J. L.; Caldes, M. T.; CasanPastor, N.; Fuertes, A.;
GomezRomero, P. Mater. Res. Bull. 1996, 31, 217.
270. Kodenkandath, T. A.; Lalena, J. N.; Zhou, W. L. L.; Carpenter, E. E.;
Sangregorio, C.; Falster, A. U.; Simmons, W. B.; OConnor, C. J.; Wiley, J. B.
J. Am. Chem. Soc. 1999, 121, 10743.
271. Kodenkandath, T. A.; Kumbhar, A. S.; Zhou, W. L.; Wiley, J. B. Inorg. Chem.
2001, 40, 710.
272. Viciu, L.; Caruntu, G.; Royant, N.; Koenig, J.; Zhou, W. L. L.; Kodenkandath, T. A.;
Wiley, J. B. Inorg. Chem. 2002, 41, 3385.
273. Viciu, L.; Golub, V. O.; Wiley, J. B. J. Solid State Chem. 2003, 175, 88.
274. Viciu, L.; Koenig, J.; Spinu, L.; Zhou, W. L.; Wiley, J. B. Chem. Mater. 2003,
15, 1480.
275. Viciu, L.; Kodenkandath, T. A.; Wiley, J. B. J. Solid State Chem. 2007, 180, 583.
276. Gondrand, M.; Joubert, J. C. Rev. Chim. Miner. 1987, 24, 33.
277. Toda, K.; Watanabe, J.; Sato, M. Solid State Ionics 1996, 90, 15.
278. Toda, K.; Watanabe, J.; Sato, M. Mater. Res. Bull. 1996, 31, 1427.
279. Gopalakrishnan, J.; Uma, S.; Bhat, V. Chem. Mater. 1993, 5, 132.
280. Richard, M.; Brohan, L.; Tournoux, M. J. Solid State Chem. 1994, 112, 345.
281. Ollivier, P. J.; Mallouk, T. E. Chem. Mater. 1998, 10, 2585.
282. Bhuvanesh, N. S. P.; Crosnier-Lopez, M. P.; Duroy, H.; Fourquet, J. L. J. Mater.
Chem. 2000, 10, 1685.
283. Schaak, R. E.; Mallouk, T. E. J. Solid State Chem. 2000, 155, 46.
284. Crosnier-Lopez, M. P.; Le Berre, F.; Fourquet, J. L. J. Mater. Chem. 2001, 11, 1146.
285. Schaak, R. E.; Mallouk, T. E. J. Solid State Chem. 2001, 161, 225.
286. Schaak, R. E.; Mallouk, T. E. J. Am. Chem. Soc. 2000, 122, 2798.
287. Hyeon, K. A.; Byeon, S. H. Chem. Mater. 1999, 11, 352.
288. Gopalakrishnan, J.; Sivakumar, T.; Ramesha, K.; Thangadurai, V.;
Subbanna, G. N. J. Am. Chem. Soc. 2000, 122, 6237.
289. Rebbah, H.; Desgardin, G.; Raveau, B. Mater. Res. Bull. 1979, 14, 1125.
290. Rebbah, H.; Pannetier, J.; Raveau, B. J. Solid State Chem. 1982, 41, 57.
291. Feist, T. P.; Mocarski, S. J.; Davies, P. K.; Jacobson, A. J.; Lewandowski, J. T.
Solid State Ionics 1988, 28, 1338.
292. Tournoux, M.; Marchand, R.; Brohan, L. Prog. Solid State Chem. 1986, 17, 33.
293. Kobayashi, Y.; Ke, X.; Hata, H.; Schiffer, P.; Mallouk, T. E. Chem. Mater. 2008,
20, 2374.
294. Gonen, Z. S.; Paluchowski, D.; Zavalij, P.; Eichhorn, B. W.; Gopalakrishnan, J.
Inorg. Chem. 2006, 45, 8736.
295. Choy, J. H.; Kim, D. K.; Hwang, S. J.; Hwang, S. H.; Hur, N. H. Physica C 1994,
235, 1023.
296. Choy, J. H.; Park, N. G.; Hwang, S. J.; Kim, D. H.; Hur, N. H. J. Am. Chem. Soc.
1994, 116, 11564.
297. Choy, J. H.; Hwang, S. J.; Park, N. G. J. Am. Chem. Soc. 1997, 119, 1624.
298. Choy, J. H.; Kwon, S. J.; Hwang, S. J.; Kim, Y. I.; Jang, E. S. Int. J. Inorg. Mater.
2001, 3, 253.
299. Choy, J. H.; Kim, Y. I.; Hwang, S. J.; Huong, P. V. J. Phys. Chem. B 2000, 104, 7273.
300. Munzarova, M. L.; Hoffmann, R. J. Am. Chem. Soc. 2002, 124, 5542.
301. Uma, S.; Raju, A. R.; Gopalakrishnan, J. J. Mater. Chem. 1993, 3, 709.
302. Fang, M. M.; Kim, C. H.; Saupe, G. B.; Kim, H. N.; Waraksa, C. C.; Miwa, T.;
Fujishima, A.; Mallouk, T. E. Chem. Mater. 1999, 11, 1526.
303. Schaak, R. E.; Mallouk, T. E. Chem. Mater. 2000, 12, 2513.
 7.04 Metal Nanoparticle Synthesis
CJ Zhong, University of New York at Binghamton, Binghamton, NY, USA
JR Regalbuto, University of South Carolina, Columbia, SC, USA
2013 Elsevier Ltd. All rights reserved.

7.04.1 Metal Nanoparticle Synthesis in the Context of Catalysis 75

7.04.2 Metal Nanoparticle Synthesis in Solution and Deposition on Support Surfaces 76
7.04.2.1 Introduction 76
7.04.2.2 Synthesis of Metal Nanoparticle Catalysts with Different Surface Protection or Stabilization 77
7.04.2.2.1 Nanoparticles with organic molecular monolayer encapsulation 77
7.04.2.2.2 Nanoparticles synthesized using dendrimers as an encapsulating environment 81
7.04.2.2.3 Nanoparticles synthesized with polymeric encapsulation 82
7.04.2.2.4 Nanoparticles synthesized with other types of surface stabilization 82
7.04.2.3 Preparation of Catalysts from Metal Nanoparticles and Catalytic Activation 83
7.04.2.4 Examples of the Metal Nanoparticles in Heterogeneous Catalysis 84
7.04.2.5 Summary 87
7.04.3 Metal Nanoparticle Synthesis at the LiquidSolid Interface 87
7.04.3.1 Modes of Contacting Metal Precursor Solutions with Solid Supports 87
7.04.3.2 Single Metals 88
7.04.3.2.1 Dry and wet impregnation 88
7.04.3.2.2 Strong electrostatic adsorption and charge-enhanced dry impregnation 89
7.04.3.2.3 Depositionprecipitation 90
7.04.3.2.4 Grafting 93
7.04.3.3 Bimetallic and Multimetallic Catalysts 93
7.04.3.3.1 Coimpregnation 93
7.04.3.3.2 Co-SEA, co-CEDI, and sequential SEA 94
7.04.3.3.3 Reductionoxidation methods in preparing bimetallic catalysts 95
7.04.3.3.4 Codepositionprecipitation 97
7.04.4 Conclusion 97
Acknowledgments 99
References 99

7.04.1 Metal Nanoparticle Synthesis in the
Context of Catalysis
The design, synthesis, and characterization of metal nanopar-
ticles ranging from subnanometer to a few nanometers are now
in the center of nanotechnology in catalysis. This chapter
describes the two main ways in which supported metal nano-
particles are synthesized: in solution with three-dimensional
(3D) chemistry as J.A. Schwarz liked to discern,1 after which
they can be deposited onto catalyst supports, or by 2D chem-
istry involving deposition of metal precursors at the liquid
support interface, after which the precursors can be thermo-
chemically converted to metal particles. As presented in
Table 1, metal particles in catalysis can be delineated based
on size, from metal clusters of several atoms on the small end
to macroscale metals and metal oxides, such as Raney nickel or
iron oxide, on the large end. Neither polyatomic clusters nor
bulk metals and metal oxides are discussed in this chapter, as
they are treated in other volumes of the series.
Metal particles in catalysis can also be distinguished by the
type of support on which they reside. Planar materials, such as
single-crystal oxide facets or amorphous silica films grown on
silicon chips, make excellent model supports for studies of
catalysis via surface science. Practical applications necessitate
the use of porous, high-surface-area oxide or carbon supports,
such as silica, alumina, titania, and activated carbon. Nanopar-
ticle synthesis methods in this chapter are limited to those
methods which can be applied at scale (to synthesize deca- or
kilogram quantities of catalyst) to high-surface-area materials.
This eliminates the line-of-sight methods, such as sputtering,
chemical vapor deposition, and evaporation, commonly used
to synthesize nanoparticles over model planar substrates. While
it may currently have limits of scalability, atomic layer deposi-
tion is a promising synthesis method finding more application
over high-surface-area supports.4 Also notable, but not dis-
cussed, are clusters synthesized on model substrates by deposi-
tion of size-selected bare metal clusters2,3 and high-surface-area
catalysts derived from clusters.79 Metals ion-exchanged into
zeolites from solution5 or by solid-state ion exchange6 are also
properly clustered and so not included in this chapter.
Space limitations prevent the in-depth discussion of the
synthesis methods. For these, the reader is referred to a number
of excellent recent books1012 and review articles.1318 We
hope in the least to provide for researchers a synthesis road
map of the most common methods and to provide some
feeling for the advantages and disadvantages of each.

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00711-7 75

76 Metal Nanoparticle Synthesis

Table 1 Metal nanoparticle synthesis in the context of catalysis

Clusters Nanoparticles Macro metal/

310 atoms 0.510 nm metal oxides
Unsupported
3D chemistry - encapsulation

Organic monolayers Vols. 2, 4 and 5

Vol. 2 of this series
Dendrimers of this series
Polymers

Model supports
Chemical vapor deposition

Size-selected cluster
deposition [2,3]
Atomic layer deposition [4]
Sputtering
Evaporation

etc.

High SA supports
2D chemistry solidliquid interface
Zeolite ion exchange [5]
Solid state ion exchange [6]
Clusters [7]
Impregnation
Grafting
Electrostatic adsorption
Deposition-precipitation

10 nm
Reduction-oxidation

7.04.2 Metal Nanoparticle Synthesis in
Solution and Deposition on Support Surfaces
7.04.2.1 Introduction
For the preparation of nanocatalysts with controllable sizes,
shapes, and compositions, one important pathway is from the
as-synthesized metal nanoparticles in solution to supported
catalysts in heterogeneous catalysis (or direct use in homoge-
nous catalysis). Challenges for engineering the nanoparticles as
catalysts include the synthesis of nanoparticles with controlla-
ble nanoscale sizes and shapes and the prevention of the
intrinsic propensity of nanoscale aggregation. Aggregation of
nanoparticles leads to eventual loss of the nanoscale catalytic
activity in practical applications. The use of naked metal nano-
particle catalysts on supporting materials based on traditional
preparative methods has been well demonstrated for different
catalytic reactions.19,20 In the last decade, new approaches to
the synthesis of metal nanoparticles capped in molecular
monolayers, dendrimers, and polymers have been rapidly
emerging, in addition to traditional colloidal synthesis. While
some of the catalysts exploit the functional groups at the cap-
ping shell of the nanoparticles,21,22 most expose the surface
sites on the metal nanoparticles either after removing the cap-
ping layers19,20,23 or through open channels of the capping
layers.2426 The metal nanoparticles serve as building blocks
toward the design of catalytic materials by taking advantage of
the diverse attributes, including size monodispersity, shape
controllability, processability, solubility, stability, tunability,
dispersion capability, and nanoscale electronic and magnetic
properties.
There are different methods to stabilize nanoparticles
against aggregation. Steric stabilization is achieved by surround-
ing the metal center by layers of material that are sterically
bulky,27 such as polymers28 or surfactants.29 The stabilizers for
metal nanoparticles30 include polymers (e.g., poly(N-vinyl-2-
pyrrolidone (PVP)), dendrimers, ligands (e.g., R-dioctylcyclo-
hexyl-1-ethylamine), micelles, microemulsions, and surfactants
(e.g., fluorous strategies).31 Electrostatic (charge or inorganic)
stabilization method has been successfully utilized for synthesis
of metal nanoparticles on oxide supports.32 Strong electrostatic
adsorption (SEA) has been demonstrated to be highly effective
for preparing nanosized Pt catalysts on metal oxide supports in
aqueous solutions by controlling pH (see the next section). The
formation of a surface positive charge (OH2 below the oxide
point of zero charge (PZC)) on metal oxide supports leads to SEA
of negatively charged Pt precursors (e.g., PtCl62) on the metal
oxide surfaces. This adsorption creates an electrical double
layer,33 which results in a Coulombic repulsion force between
individual particles. Some nanocluster stabilizers combine both
electrostatic and steric stabilization,34,35 as in the case of the
polyoxoanion-stabilized nanoclusters. Interestingly, there have
recently been reports of nanoclusters that are claimed to be
stabilized by solvent molecules such as ethers and thioethers,
that is, in the absence of charge or steric stabilizers.29,36,37
In this section, some of the recent approaches to the
synthesis of metal nanoparticle catalysts derived from

encapsulation with monolayer, dendrimeric, and polymeric
structures will be described. Since the focus is the synthesis of
supported metal nanoparticles for various catalytic applica-
tions, effective approaches to supporting and activating the
capped metal nanoparticle catalysts will be discussed. In addi-
tion to the manipulation or removal of the capping layer
structures, the processing of atomic-scale structures of the
metal nanoparticle, especially for bimetallic/trimetallic nano-
particles, will be highlighted. Finally, selected examples of
the metal nanoparticle catalysts in catalytic reactions will be
highlighted, along with a discussion of challenges and oppor-
tunities in the exploration of metal nanoparticles in catalysis.
7.04.2.2 Synthesis of Metal Nanoparticle Catalysts with
Different Surface Protection or Stabilization
The use of capping agents in the synthesis is perhaps one of the
most effective strategies to control the size, shape, and compo-
sition of metal nanoparticles.13,14,3869 There are increasing
examples demonstrating that the metal nanoparticles synthe-
sized by this strategy display intriguing catalytic properties for a
number of chemical reactions. Table 2 summarizes some of
the examples, which is not meant to be exhaustive rather
illustrative. There are many reviews on the synthesis of metal
nanoparticles (e.g., Shiju and Guliants13 and Scholten et al.14).
In the following sections and subsections, selected exam-
ples will be highlighted to illustrate the synthesis of metal
nanoparticles with monolayer, dendrimeric, and polymeric
capping layer structures, the processing of the metal nan-
oparticles for catalytic activation, and some recent understand-
ing of the nanoscale structures of the metal nanoparticles in
catalytic reactions.
7.04.2.2.1 Nanoparticles with organic molecular
monolayer encapsulation
The synthesis, characterization, and application of organic
monolayer-protected nanoclusters or nanoparticles have been
extensively discussed in a number of chapters.20,7072 Among
many synthesis methods, wet chemical synthesis is very effec-
tive in producing monodispersed nanoparticles. The two-
phase synthesis reported first by Brust et al.73 is now widely
used for fabricating stable and soluble gold and many other
metal nanoparticles of a few nanometer core sizes. Briefly, this
protocol involves transfer of AuCl4 from aqueous to organic
solution by a phase-transfer reagent followed by reduction in
the presence of thiols, producing thiolate-capped gold nano-
particles of a few nanometer sizes, as shown in Figure 1.
Further modifications of the synthetic protocols or processing
of the produced nanoparticles have been shown to produce a
variety of chemically tunable nanoparticles in terms of size,
shape, and surface properties.74,75 Perhaps one of the most
important attributes of these nanoparticles is the reactivity of
both the metal nanoparticle and the capping molecules that
allows reengineering or assembly of nanoparticles. The cap-
ping molecules can be reconstituted to impart the desired
functional properties, which were demonstrated in a place-
exchange reaction first by Murrays group.76
This two-phase-based strategy has also been used for syn-
thesizing bimetallic nanoparticles.48,49,73,74 For example,
AuCl4 and PtCl62 were first transferred from aqueous
Metal Nanoparticle Synthesis 77
solution into toluene solution using a phase-transfer reagent
(tetraoctylammonium bromide (TOABr)). Thiols (e.g., decan-
ethiol, DT) or amine compounds (e.g., oleylamine, OAM)
were added to the organic solution as capping agents. An excess
of aqueous NaBH4 was slowly added for the reduction reac-
tion. The resulting DT/OAM-encapsulated AuPt nanoparticles
in toluene were collected by removing the solvent and cleaned
using ethanol. The control of the bimetallic composition in the
entire bimetallic composition range and the particle size with
high monodispersity (<0.5 nm) was achieved by manipulat-
ing the precursor feed ratio (Figure 2(a)).48 The as-synthesized
AumPt100-m nanoparticles with different compositions are
capped with thiol/amine monolayer shells. Varying the feeding
ratio of the metal precursors used in the synthesis controls the
bimetallic composition and sizes of the nanoparticles. These
particles are alloyed with nanocrystalline features, as shown by
the HAADF-STEM and EDX images for a sample of Pt51Au49
nanoparticle supported on carbon (Figure 2(b)). Bimetallic
nanoparticles in which the nanocrystal core consists of
one metal core and another metal shell (coreshell), such as
AuPt and PtAu, were also synthesized by seeded growth
method.77 Similar approach was used for synthesizing other
binary nanoparticles, such as AuAg,78 coreshell AuAg,79 and
Cu nanoparticles.80
One of the unique properties of metal nanoparticles is the
dramatic difference of their melting points from their bulk
counterparts. The melting point of nanoparticles is related to
that of its bulk metal81 in terms of melting temperatures of the
particle and the bulk solid, the radius of the particle, the
density of the solid and liquid, the surface energy of the solid
and liquid, and the heat of fusion. While the lowering of
melting point is inversely proportional to the particle size,
the surface-melting temperature could be even lower (e.g.,
140  C for 2 nm Au), and it could be influenced by the
capping-dominated surface tension. Under such surface-
melting and surface-tension effects, the driving force for the
coalescence of two surface-melting particles is the reduction in
free energy through a reduction in surface area, that is, increase
of size. The issue of how the capping/linking agents influence
the interparticle interactions and coalescence under various
thermal treatment conditions was in fact addressed from
both experimental and theoretical perspectives in several of
our reports.82 The size control of alkanethiolate
monolayer-capped Au nanocrystals is linked to the aggregative
nucleation and growth mechanism, which is evidenced by
both experimental and theoretical analysis of the nucleation
parameters.83
This strategy combines surface-melting and surface-tension
effects as part of the driving force for a controlled interparticle
coalescence for processing size and morphology of metal,
alloy, and coreshell nanoparticles in solutions (e.g., Au, Cu,
AuAg, AuPt, and Fe2O3 (or Fe3O4)Au)48,7780,82,84 under
mild elevation of temperatures (100200  C). In contrast to
Osterwald ripping known for small particle systems, the ther-
mally activated processing of a solution of presynthesized
nanoparticles involves coupling of the molecular capping and
reencapsulation to the interparticle coalescence. Figure 3 illus-
trates thermally activated aggregative growth of Au nanoparti-
cles from small to large sizes and from Cu nanoparticles to
Cu2S nanoparticles (left), and the formation of Fe2O3Au
78 Metal Nanoparticle Synthesis

Table 2 Selected examples of metal nanoparticles synthesized with different capping structures

Metal nanoparticles Synthesis method Catalytic reaction References

NPs with molecular monolayer encapsulation

Monometallic
Au b-D-Glucose reducing agent and capping agent Reduction of 4-nitrophenol in the 38
presence of NaBH4
Pd CTAB as both capping and structure-directing agents Hydrogenation of dienes into alkenes 39
Au DT as capping agent, NaBH4 as reducing agent MOR, ORR 48
Pt (no support) Oleic acid and oleylamine as reducing agent and capping ORR 40
agent
Bimetallic
MnPt (nanocubes) Benzyl ether or phenyl ether as solvent, oleic acid and ORR 41
oleylamine as reducing and capping agents
Pt3Zn (nanocrystals) Oleylamine and oleic acid as both stabilizers and solvents MOR 42
Pt3Co 1,2-Tetradecanediol as reducing agent, ORR 43
1-adamantanecarboxylic acid and oleylamine as capping
agent
PtM (M Co, Cr, and Fe) Reversed micelle method (NaBH4 as the reducing agent, ORR 44
AOT, 1-butanol, and cyclohexane as solvent and capping
agent)
PtFe (no support) Chemical reduction (oleylamine and oleic acid as capping ORR 45
agent, 1,2-hexadecanediol as reducing agent, dioctylether
as solvent)
AuPt DT, OAM/OAC as capping agent, NaBH4 as reducing agent, MOR, ORR 48
H2O, Tl as solvents
Trimetallic/multimetallic
AuAgPd (AuAg as Caped by CTAB (aq), HAuCl4, AgNO3, and H2PdCl4 as Sonogashira coupling reactions 46
core, Pd as shell) precursors
AuPdPt nanowires H2PdCl4, H2PtCl6, and HAuCl4 as precursors. Te nanowires EOR (ethanol electrooxidation) 47
as both reducing agent and sacrificial template
PtMM0 ( M, M0 Ni, Co, Fe, 1,2-hexadecanediol as reducing agent, oleic acid and ORR, MOR, EOR, and CO oxidation 49,50
V, Ir) oleylamine as reducing and capping agents (PROX)
NPs with dendrimer encapsulation
Monometallic
Pt, Rh (1 nm), or PAMAM dendrimers capped, supported on silica, such as Hydroalkoxylation, Hydrogenation 51
bimetallic nanoparticles SBA-15, MCF-17 dendrimer
Pt /SiO2 Generation 5 PAMAM dendrimer (G5-OH)-encapsulated Toluene hydrogenation and CO oxidation 52
Pd and Cu PAMAM and PPI dendrimer-encapsulated, NaBH4 as Hydrogenation 53
reducing agent (aq)
Cu, M and M/Cu bimetallic PAMAM, G2, G3, G4, and G6 and PPI Homogeneous catalysis in organic 54
NPs, M Ag, Au, Pt, Pd dendrimer-encapsulated, NaBH4 as reducing agent (aq) solvents, fluorous/organic biphasic
solvents and supercritical CO2
Bimetallic
AuPt PAMAM dendrimer-encapsulated, NaBH4 as reducing CO oxidation 55
agent; AuPt by Cu exchange method
AuAg Amine-terminated (G 3.0 and 5.0) and carboxyl-terminated Catalytic activities for reduction of 56
(G 3.5 and 5.5) poly(amidoamine) dendrimers in water; p-nitrophenol
NaBH4 as reducing agent
PdPt PAMAM dendrimer-encapsulated Hydrogenation in various solvents 57,58
NPs with polymer encapsulation
Monometallic
Pt Poly(vinylpyrrolidone)-capped, size range from 1.7 to Ethylene hydrogenation, ethane 59
7.1 nm hydrogenolysis
Pt PVP- and TTAB-capped High-temperature CO oxidation reaction 60
Pt Tetradecyltrimethylammonium bromide as a surface- Ethylene hydrogenation 61
stabilizing reagent. NaBH4
Pd Standard PVP-stabilized, ethanol as the reducing agent Suzuki reaction involving phenylboronic 62
acid and iodobenzene
Rh Capper by copolymers dissolved in [BMI][BF4] IL Benzene hydrogenation 63
Bimetallic
PtRu/Si Poly(vinylpyrrolidone)-capped CO oxidation 64
AuPd in TiO2 thin film Standard PVP-stabilized, H2 plasma reduction Methylacrylate hydrogenation 65
 Metal Nanoparticle Synthesis 79

Table 2 (Continued)

Metal nanoparticles Synthesis method Catalytic reaction References

Miscellaneous
PtAu Potential cycling regimes by modifying Pt nanoparticles ORR 66
with gold (Au) clusters
PtAu Potential cycling regimes effected by a submonolayer of Au MOR 67
clusters
PdPt Pt nanodendrites on a Pd core by reducing PtCl4 with ORR 68
L-ascorbic acid in the presence of Pd seeds
AuPtPd triple-layered core AuCl4, PdCl4, PtCl4, as precursors and PVP as the capping MOR 69
(not support) agent and structure-directing agent, ascorbic acid as
reducing agent

MOR, methanol oxidation reaction; ORR, oxygen reduction reaction; EOR, ethanol electrooxidation; CTAB, methanol oxidation reaction; DT, decanethiol; PVP, poly(vinylpyrrolidone);
OAM, oleylamine; OAC, oleic acid; PAMAM, poly(amido amine).

TOABr AuCl4- CA Reducing agent Au-2 nm NPs

Figure 1 Illustration of the two-phase synthesis of gold nanoparticles in terms of phase transfer and chemical reduction.

100

80
Pt % (in nanoparticles)

60

40

20

0
0 20 40 60 80 100
(a) Pt % (feeding) (b) 2 nm 2 nm

Figure 2 (a) Correlation between the Pt composition in nanoparticles and the synthetic feeding composition (reprinted with permission from
Wanjala, B.N.; Luo, J.; Loukrakpam, R.; Mott, D.; Njoki, P.N.; Fang, B.; Engelhard, M.; Naslund, H.R.; Wu, J.K.; Wang, L.; Malis, O.; Zhong, C.J. Chem.
Mater. 2010, 22, 42824294, Copyright 2010 American Chemical Society). (b) HAADF-STEM and EDX images for a sample of Pt51Au49 nanoparticle
supported on carbon (red dots, Pt; blue dots, Au).

nanoparticles from Au and Fe2O3 precursor nanoparticles and
from a mixture of Au and Cu nanoparticles to AuCu alloy nano-
particles (right). For example, using presynthesized gold nano-
particles (2 nm), the particle size growth is shown to be a function
of chain length of the capping alkanethiols.82 The average size of
the particles is dependent on the chain length of the alkanethiols
used in the thermal processing.82 The linear relation for the
average particle core size versus the chain length of alkanethiols
displays a slope of 0.2 nm/methylene. The increase in chain
length of the capping alkanethiolates leads to a gain in stabiliza-
tion energy due to interchain cohesive interactions, consistent
with the known linear correlation of cohesive energy of the
alkyl chains (1 kcal mol1 CH2). This chain length-dependent
stabilization energy contributes to a balance of alkanethiolate
desorption, core coalescence, and reencapsulation in the size
evolution.82 Using this approach with slight modifications,
gold-based alloy and other metal (e.g., copper) nanoparticles of
different sizes have also been prepared.80
In contrast to gold nanoparticles, iron oxide nanoparticles
do not exhibit a coalescence under such a mild temperature
elevation, which constitutes the basis for a thermally activated
hetero-interparticle coalescence to fabricate coreshell-structured
80 Metal Nanoparticle Synthesis
no
ry na
Una
Binary nano
Unary nano
Binary nano
Cu
Au
Fe2O3
CA
Figure 3 Illustrations of thermally activated aggregative growth of Au nanoparticles from small to large sizes and from Cu nanoparticles to Cu2S
nanoparticles (left) and the formation of Fe2O3Au nanoparticles from Au and Fe2O3 precursor nanoparticles and from a mixture of Au and Cu
nanoparticles to AuCu alloy nanoparticles (right).
(Fe-oxideAu) nanoparticles.84 In this case, the surface-melting
temperature of the precursor 2 nm-sized gold nanoparticles is
much lower than that for the 5 nm-sized iron oxide nanoparti-
cles. The interparticle heterocoalescence of gold-coated iron
oxides leads to the formation of coreshell nanostructures.
The processing environment is important for the stabilization
of the nanoparticles, which include the presence of organic ionic
materials such as TOABr in the solution and capping molecules
such as alkanethiols (e.g., decanethiols (DT)) for capping on
Au nanoparticles and those such as oleylamine and/or oleic
acid for capping on Fe-oxide nanoparticles. The presence of
molecularly capped iron oxide nanoparticles in the solution
created a unique thermal microenvironment for the coreshell
coalescence to produce monodispersed Fe-oxideAu nanoparti-
cles of controlled sizes.
More recently, this approach has been demonstrated for the
processing of Cu and AuCu alloy nanoparticles,80,85 demon-
strating an important concept for nanoscale aggregative
growth, that is, moleculesolid duality. The nanoscale duality
concept is illustrated with a binary system of Cu and Au
nanoclusters/nanoparticles in a solution or on a solid sub-
strate. The structural evolution at or near their melting temper-
atures is different from the evolution of unary nanocluster or
nanoparticle systems, for example, the evolution of thiolate-
capped Cu (<10) nanoclusters to Cu2S nanodiscs80 and of Au
(<2 nm) to larger-sized Au nanoparticles.82 The nanoscale
sizing, shaping, and alloying occur in a binary system at or
near the melting temperature for a binary system of thiolate-
capped Cu and Au nanoclusters/nanoparticles, forming well-
defined AuCu alloy nanostructures.
While a variety of supported Pt-group binary or ternary cata-
lysts have been prepared by traditional methods such as copre-
cipitation, depositionprecipitation (DP), ion exchange,
impregnation, successive reduction, and calcination,86,87 the
ability to control the size and composition is limited due to the
propensity of aggregation of metals at the nanoscale. Aggregation
of nanoparticles leads to eventual loss of the nanoscale catalytic
activity in practical applications. The preparation of nanoparti-
cles capped in monolayers, polymers, or dendrimers is rapidly
( At or near melting temperatures )
emerging, demonstrating remarkable parallels to supported
nanoparticle catalyst preparation.13,14,3869 The molecular
encapsulation-based synthesis and processing strategy involves
a sequence of three steps for the preparation of multimetallic
catalysts: (1) chemical synthesis of the metal nanocrystal cores
with molecular encapsulation, (2) assembly of the encapsulated
nanoparticles on support materials (e.g., carbon powders or
carbon nanotubes), and (3) thermal treatment of the supported
nanoparticles.4850 Figure 3 depicts the general concept and
some examples. Examples described in this chapter will include
gold (Au), goldplatinum (AuPt), platinumvanadiumiron
(PtVFe), and platinumnickeliron (PtNiFe) nanoparticles.
The bimetallic, trimetallic, and multimetallic nanoparticles
can be synthesized in a single-phase solution containing the
metal precursors, capping agents, and reducing agents in con-
trollable ratios. This approach is an expansion of an elegant
protocol previously reported for the synthesis of binary PtFe
nanoparticles with organic monolayer encapsulation.72 Based
on this method, we first reported the size- and composition-
controlled ternary nanoparticles with average sizes of a few
nanometers (Figure 4(a)).49 For example, the general synthesis
of PtVFe nanoparticles involved a combination of thermal
decomposition and reduction reactions of three metal precur-
sors, PtII(acac)2, VIVO(acac)2, and Fe0(CO)5, in controlled
molar ratios, forming oleylamine/oleic acid-capped PtVFe nano-
particles. The composition of the Pt0n1V0n2Fe0n3 nanoparticles,
where n1, n2, and n3 represent the atomic percentages of each
metal, is controlled by the feeding ratio of the metal precursors,
which is expressed in molar percentages (m1, m2, and m3). By
manipulating the relative concentrations of the three metal pre-
cursors, the coreshell-type nanoparticles in which the nano-
crystal PtVFe cores are encapsulated with a monolayer shell of
amines/acids were successfully synthesized. The average diame-
ters of the nanocrystal cores were well controlled between 1 and
3 nm with high monodispersity (Figure 4(b)). The ternary
composition is controlled by manipulating the feeding ratios
of the metal precursors used for the synthesis.
PtVFe nanoparticles capped with oleylamine/oleic acid
monolayers were synthesized in thermal decomposition and

reduction reactions by controlled molar ratios of three metal
precursors, PtII(acac)2, VIVO(acac)2, and Fe0(CO)5.49 This
method was also applied to the synthesis of PtNiFe nanoparti-
cles. The basic morphology of the observed nanoparticles is
largely characterized by the highly faceted nanocrystal feature,
which is observable by a close examination of the shapes of the
individual nanocrystals. The particles are highly monodis-
persed, with an average size of 1.9  0.3 nm. The fact that
the nanoparticles have well-defined interparticle spacing and
display domains of hexagonal ordering is indicative of the
encapsulation of the nanocrystal cores by organic monolayers.
While the sizes of the ternary nanoparticles varied slightly
depending on the actual composition, the data demonstrated
the ability to the size monodispersity. Nanoparticles with aver-
age diameters ranging from 1 to 3 nm have been obtained with
high monodispersity (0.2 nm).
This method was also used for the synthesis of other oley-
lamine/oleic acid-capped bi-/tri-/multimetallic nanoparticles,
for example, PtCo, PtNi, PtNiFe, PtNiCo/C, PtIrCo/C,
and PtVFe/C.49,50 This approach has been increasingly used
for the synthesis of nanoparticles for catalysts. Examples for
the oxygen reduction reaction include Pt3Co,43 Pt,40 PtFe,45
PtM (M Co, Cr, and Fe),44 and Pd containing trimetallic
nanoparticles.47 Nanoparticles with controllable shapes were
also studied for catalysts. Examples include Mn Pt nanocubes
for fuel cell,41 Pt3Zn nanocrystal methanol oxidation,42 and
Pd/a-Al2O3 for the hydrogenation of buta-1,3-diene.39
7.04.2.2.2 Nanoparticles synthesized using dendrimers as
an encapsulating environment
Another class of catalytic nanoparticles is derived by synthesis
of metal nanoparticles encapsulated in dendrimers (Figure 5).
The nanoparticles are prepared by sequestering metal ions
within dendrimers followed by chemical reduction to yield
Pt-
e.g., V- precursors
Fe-
Pt
V
Fe
Oleic amine : CH(CH)CH-CH(CH)NH
Synthesis Oleic acid : CH(CH)CH-CH(CH)COOH
(a)
Figure 4 (a) A schematic illustration of the synthesis of PtVFe nanoparticles by the polyol method. (b) TEM micrograph for a sample of ternary
Pt32V13Fe55 nanoparticles. Reprinted with permission from Luo, J.; Han, L.; Kariuki, N.; Wang, L.; Mott, D.; Zhong, C.J.; He, T. Chem. Mater. 2005, 17,
52825290, Copyright 2005 American Chemical Society.
Figure 5 A schematic illustration of the synthesis of metal nanoparticles encapsulated in a dendrimer structure.
Metal Nanoparticle Synthesis 81
the corresponding zerovalent metal nanoparticle. The size of
such particles depends on the number of metal ions initially
loaded into the dendrimer. As demonstrated recently by
Crooks and coworkers,31 such nanoparticles are prepared by
sequestering metal ions within dendrimers followed by chem-
ical reduction to yield metal nanoparticles. The nanoparticles
dispersed in the dendrimer exhibited remarkable catalytic
activities toward hydrogenation and carboncarbon coupling
reactions. Major attributes of the dendrimer encapsulation
include uniform structure, stabilization against agglomera-
tion, substantial fraction of active surface, access of small
molecules to the nanoparticles, and tunable periphery to
control solubility and interfacial linking. These have been
discussed in a recent review.54 The preparation by sequester-
ing metal ions within dendrimers followed by chemical reduc-
tion has been demonstrated to yield metal nanoparticles
(15 nm).31,53,54
The SiO2 supported metal catalysts were prepared using
dendrimer-encapsulated Pt nanoparticles as metal precursors.
Polyamidoamine (PAMAM) dendrimers were used to template
and stabilize Pt nanoparticles prepared in solution. These
dendrimer-encapsulated nanoparticles (DENs) were then
deposited onto a high-surface-area silica support.52 Bimetallic
dendrimer-stabilized nanoparticles were used to prepare sup-
ported PtAu catalysts within the bulk miscibility gap for this
binary system. Hydroxy-terminated generation 5 PAMAM den-
drimers were used to prepare Cu0 nanoparticles. The Cu0 NPs
were subsequently used to reduce K2PtCl4 and HAuCl4, prepar-
ing stabilized bimetallic PtAu NPs with a 1:1 stoichiometry. The
stabilized nanoparticles were adsorbed onto a high-surface-area
silica support.55
The synthesis and characterization of 13 nm diameter
bimetallic AuAg DENs are reported. Three different bimetallic
structures were examined: AuAg alloys synthesized by
20 nm
(b)
Support
82 Metal Nanoparticle Synthesis
co-complexation and subsequent reduction of dendrimer-
encapsulated Au3 and Ag and core/shell [Au](Ag) and
[AuAg alloy](Ag) structures (for structured materials, brackets
indicate the core metal and parentheses indicate the shell
metal) synthesized by a sequential loading method. Depend-
ing on the shell metal and its oxidation state, the AuAg nano-
particles can be extracted from the dendrimer into an organic
phase using different surfactants.54
Preparation of dendrimer-encapsulated metal nanoparticles
within amphiphilic dendrimers in an organic solvent is
reported. The important new result is that metal ion encapsula-
tion is driven by solubility differences between the metal salt in
solution and the dendrimer interior. This new approach makes
it possible to prepare relatively large dendrimer-encapsulated
metal nanoparticles.53 The synthesis, characterization, and
catalytic activity of bimetallic PdPt dendrimer-encapsulated
catalysts were reported. These materials are prepared by co-
complexing different ratios of palladium and platinum salts
with the interior tertiary amines of hydroxide-terminated
PAMAM dendrimers.57 Goldsilver binary nanoparticles were
prepared. These particles were stabilized by amine-terminated
(generation (G) 3.0 and 5.0) and carboxyl-terminated (G 3.5
and 5.5) PAMAM dendrimers in water. The mean diameters of
the particles were 710 nm for silver particles and 34 nm for
both gold and alloy particles, respectively.56
7.04.2.2.3 Nanoparticles synthesized with polymeric
encapsulation
The use of polymer-capped nanoparticles as catalysts is another
area of increasing interest. For example, using stepwise growth
method, PVP-stabilized Pd nanoparticles (37 nm) with vary-
ing particle size are prepared.88 Pt nanoparticles supported in a
nanoporous Al2O3 catalyst are prepared by reduction of
K2PtCl4 in the presence of Al2O3 and poly(acrylic acid) as
capping agent to protect against particle aggregation and pro-
duce various shapes. Polymer-mediated self-assembly of func-
tionalized Pd and SiO2 nanoparticles has been demonstrated
for the preparation of catalysts.89
A series of PVP-stabilized Pd nanoparticles with varying
particle size are prepared by using the stepwise growth reac-
tion. The PVPPd nanoparticles were synthesized by the reduc-
tion of the Pd ions with ethanol.62a The effect of Pd particle size
on the Suzuki reaction between phenylboronic acid and iodo-
benzene was investigated by the use of four Pd catalysts with
mean particle sizes of 3.0, 3.9, 5.2, and 6.6 nm, respectively.62b
Platinum nanoparticles in the size range of 1.77.1 nm were
produced by alcohol reduction methods. PVP was used to
stabilize the particles by capping them in aqueous solution.
Mesoporous SBA-15 silica with 9 nm pores was synthesized by a
hydrothermal process and used as a catalyst support.59 2D Rh/Pt
bimetallic nanoparticle arrays were also synthesized on a silicon
surface via the LangmuirBlodgett technique.64 The synthesis of
platinum nanoparticles capped with PVP and tetradecyl tributy-
lammonium bromide molecules was also demonstrated.60
Cubic, cuboctahedral, and porous Pt nanoparticles have
been prepared using tetradecyltrimethylammonium bromide
as a surface-stabilizing reagent. The morphology was con-
trolled by adjusting the reduction method. H2 production in
situ from NaBH4 has enabled the synthesis of uniform nano-
particles. By changing the pH value, which contributes to
controlling the reduction rate, shape evolution from cubocta-
hedra to cubes was observed. These nanoparticles, which are
electrostatically capped with alkylammonium ions and shape-
controlled without the aid of foreign metal ions, show superior
catalytic activity to nanoparticles prepared with a polymeric
stabilizing reagent and silver.61 Cube-shaped, multioctahedral-
shaped, nanowire or dendritic Pt nanoparticles were shown to
exhibit an enhanced electrocatalytic activity for the oxygen
reduction reaction.68,90
PdPt bimetallic nanodendrites consisting of a dense array
of Pt branch on a Pd core by reducing K2PtCl4 with L-ascorbic
acid in the presence of uniform Pd nanocrystal seeds in an
aqueous solution. The Pt branches supported on faceted Pd
nanocrystals exhibited relatively large surface areas and particu-
larly active facets toward the oxygen reduction reaction, the rate-
determining step in a proton-exchange membrane fuel cell. The
PdPt nanodendrites were two and a half times more active on
the basis of equivalent Pt mass for oxygen reduction than the
state-of-the-art Pt/C catalyst and five times more active than the
first-generation supportless Pt-black catalyst.68
7.04.2.2.4 Nanoparticles synthesized with other
types of surface stabilization
Electrocatalysts for oxygen reduction reaction were prepared by
modifying Pt nanoparticles with gold clusters. In situ x-ray
absorption near-edge spectroscopy and voltammetry data sug-
gest that the Au clusters confer stability by raising the Pt
oxidation potential.66,67 Rhodium nanoparticles stabilized by
the ionic copolymer in ionic liquids have been successfully
obtained. The nanoparticles showed unprecedented lifetime
and activity in arene hydrogenation under forcing conditions
(a temperature of 75  C and a hydrogen pressure of 40 bar)
with a total turnover of 20 000 (in five total recycles of 4000
total turnovers each) and a turnover frequency of 250 h1,
demonstrating that the combination of ionic liquids with
ionic liquid-like stabilizers is a pathway toward highly stable
and active nanoparticle catalysts.63
Ionic liquids provide a flexible liquid platform for catalysis
by transition metal nanoparticles. Indeed, ionic liquids can act
as the solvent, stabilizer, ligand, and support for metal nano-
particles. It is now possible to design metal nanoparticle cata-
lysts for specific applications, in particular, for the selective
hydrogenation of arenes under very mild reaction conditions.
In fact, it appears that soluble metal nanoparticles in ionic
liquids behave as single-site metal catalysts for the hydroge-
nation of alkenes, but as soluble-like heterogeneous catalysts
for the reduction of arenes or in FischerTropsch synthesis.
These nanoparticles may serve as simple reservoirs of soluble
monometallic catalysts for CC coupling reactions, such as the
Heck and Suzuki reactions.14
The recent study has shown that Pluronic F127 triblock
copolymer can assist in the formation of unique AuPdPt
triple-layered coreshell-structured nanoparticles consisting
of an Au core, a Pd inner layer, and a nanoporous Pt
outer shell.91 It was further demonstrated that such interesting
AuPdPt nanoparticles can be easily synthesized by using PVP
instead of Pluronic F127. PVP is a very typical capping and
structure-directing agent used for the synthesis of various
metallic nanostructures. The present synthetic route using a
PVP-based aqueous solution will greatly contribute to the

further design of multilayered metallic nanoarchitectures with
designed compositions and desired functions. The obtained
AuPdPt triple-layered nanoparticles displayed a drastically
enhanced catalytic activity for methanol oxidation, demon-
strating its promising potential as an effective catalyst.69
7.04.2.3 Preparation of Catalysts from Metal
Nanoparticles and Catalytic Activation
Molecularly mediated assembly (and/or dispersion) of the
capped metal or alloy nanoparticles on planar48 and carbon
nanomaterials (carbon black, e.g., Vulcan XC-72R)4850 and
nanotubes (CNT, 2060 nm diameters)92 exploits the inter-
particle and particlesupport interactions, such as van der
Waals, hydrogen, and covalent bonding. The loading of nano-
particles on carbon can be controlled effectively by manipulat-
ing the carbon-to-particle ratio and the solvent properties. The
carbon-supported nanoparticle catalysts can be activated by
thermal treatment in a tube furnace under controlled temper-
atures and atmosphere, including oxidative (in air), reductive
(in H2), and inert (N2) conditions. The activation and calcina-
tion involves (1) removal of the surface-capping agents and (2)
possible reconstitution of the nanocrystal composition or
change in morphology. The key is to control the size and
composition. The calcination treatment is well documented
for catalyst preparations, including surfactant-capped Pt/Ru,93
PtFe,45 coreshell Pt/Ru,94 and metal-carboxylate-linked Au95
nanoparticles. The surface mobility and adhesion of metals on
the support materials depends on surface composition and
surface interactions, as demonstrated for strong adhesion
support materials (e.g., TiO2).96
The thermochemical treatments of catalysts50d involved
removal of organic shells and calcination of the ternary nano-
particles, by heating the catalysts under oxygen (oxidative reac-
tion condition (ORC)) or nitrogen (minimum- or nonreactive
condition (MNC)) atmosphere to remove the capping molecules
followed by hydrogen atmosphere (reductive reaction condition
(RRC)) to calcine the alloy structures (Figure 6). For example,
the catalyst was heated at 260  C in N2 atmosphere for 30 min
Under N2
or low O2
Partially or fully oxidized
Under O2
A B C O
(LO: low O2%
HO: high O2%)
Temperature
T1
T0
MNC or ORC
Time
Figure 6 A schematic illustration of the thermal treatment protocols of the catalyst: minimum or nonreactive condition (T1) RRC (T2) and ORC
(T1, low or high oxygen) RRC (T2). Typically, low oxygen is 1.5% O2 and high oxygen is 15% O2. Reprinted with permission from Wanjala, B.; Fang, B.;
Loukrakpam, R.; Chen, Y.; Engelhard, M.; Luo, J.; Yin, J.; Yang, L.; Shan, S.; Zhong, C.J. ACS Catalysis, 2012, 2, 795806, Copyright 2012 American
Chemical Society.
Metal Nanoparticle Synthesis 83
for the MNC treatment or in 1.515% O2 (in N2) for 30 min for
the ORC treatment. The post-ORC or MNC catalysts were then
treated at 400  C in 15% H2 for 240 min for the RRC treatment.
The complete ORCRRC and MNCRRC treatment sequences
are illustrated in Figure 6, where T1 and T2 represent the temper-
atures for the first treatment (i.e., MNC or ORC) and the subse-
quent treatment (RRC). In the ORC treatment, different O2
concentrations were used, typically 1.5% or 15% O2, which is
labeled as LO (low O2) or HO (high O2). The effective removal
of the capping monolayers from the nanoparticles by the ther-
mal treatment process97 was supported by the absence of the
vibrational bands characteristic of the capping molecules in the
CH stretching region detected by F TIR and the absence of the
bands associated with sulfur species after the thermal treatment
detected XPS analysis. The average particle sizes after the calcina-
tion treatment were found to show a slight increase in compar-
ison with that before the treatment. The subtle increase in size
was found to be dependent on the calcination temperature. More
importantly, the alloying and structure depends on the oxida-
tivereductive treatment parameters.50d
The question whether the nanoparticles are multimetal-
lic in individual nanoparticles or in an ensemble of the
nanoparticles was addressed using transmission electron
microscopyenergy dispersive x-ray spectroscopy (TEM
EDX).98 For example, the atomic compositions for individ-
ually isolated PtVFe nanoparticles (Figure 7(a)) are found
to be almost identical, independent of the actual sizes, in
contrast to the results observed from the traditional synthe-
sis where large-sized particles are usually base metal-rich
and small particles are Pt-rich. Similar compositions were
found for other binary and ternary alloy catalysts, indicating
the composition uniformity in the individual nanoparticles.
Theoretically, thermally induced sintering or coagulation
occurs by either sintering of small particles into larger-sized
particles or agglomeration of small particles into coagulating
cluster.96 The driving force for sintering can be described by a
basic GibbsThompson relation, m(r)  m(1) 2gO/r, where g
represents surface free energy of the metal and O represents
bulk metal volume per atom. Recent model studies99
Under H2 (I)
Under H2 (II)
T2
RRC
84 Metal Nanoparticle Synthesis
13
Height (nm)
Atomic
composition (%)
Pt V Fe
10 (area) 34 16 50
11 12 10
11 (3 nm) 34 16 50
12 (3 nm) 33 15 52
13 (6 nm) 32 13 55
All area 32 19 49 30.00 nm
(a) (b)
Figure 7 (a) High-resolution TEMEDX data for the ternary nanoparticle catalyst (Pt33V14Fe53/C); (b) In situ AFM images (2  2 mm2) for DT-capped
Au2nm on mica before (b) and after (c) thermal treatment at 225  C for 30 min (image acquired after cooling to room temperature).
demonstrated that the dependence of this energy on particle
size is much stronger than the predicted. There are essentially
two limiting mechanisms96b: diffusion of intact particles fol-
lowed by coalescence and detachment of atoms from smaller
particles followed by diffusion across terraces and reattach-
ment to larger particles. Based on the model modified by
Campbell and coworkers,99 the particle size change depends
on DHsub (metals bulk sublimation enthalpy), Eadsupport (acti-
vation energy for adsorption on support), and Ediffsupport
(activation energy for diffusion on support). The rate for the
change of a particles radius (r) is a function of the apparent
activation energy Etot.99 The rate for the change of particle size
with temperature at constant g is increased with decreasing Etot,
whereas the rate at constant Etot is increased with increasing g.
The parameters such as g, Eadsupport, and Ediffsupport are related
to the functional groups of the capping or linking agents, the
hydrophobicity, or the binding properties of the substrate,
which can be manipulated by changing the functional groups
of the capping or interparticle-linking agents.
The change in size of individual nanoparticles (e.g.,
decanethiol-capped Au2-nm) cast on mica as a result of the ther-
mal treatment at 225  C can be detected by the change in height
data of the atomic force microscopy images (Figure 7(b)).
Before heating treatment, the average particle size determined
from the cross-sectional data (height) is consistent with the
expectation for the sum of the core diameter (2.0 nm) and
the shell molecular length times a factor of two (2.5 nm).
After the treatment, the relative position of nanoparticles
remains unchanged. The particles displayed a clear shrinkage in
size as a result of the removal of the capping shell molecules.
There is no apparent aggregation of the nanocrystal cores
because the particles are widely separated, and the adhesion
(adsorption) of nanocrystals to the surface of mica is quite
strong. The tendency of nanoparticles to aggregate depends on
the surface properties of the substrate and the interparticle
molecular interaction. The largely unchanged interparticle
spatial feature in the lateral plane demonstrates that the inter-
particle spatial property can be controlled even if there is a local
aggregation from the overlapped nanoparticles, indicating
the important role played by the combination of the initial
4 4
Height (nm)
0 0
0 1 2 0 1 2
Section (mm) Section (mm)
molecular capping/wiring and the subsequent adhesive interac-
tion of the particle to the substrate in the spatial fixation.
The experimental and theoretical results can be understood
based on comparisons of the relative activation energies
(Etot (mica) > Etot (HOPG)) and surface tensions (g(CH3)
(mica) < g(CO2H) (HOPG)). Based on the theoretical simu-
lation and the qualitative analysis, the initial aggregation of the
nanoparticles was considered to involve energetic contribu-
tions from both shellsubstrate interaction and Ausubstrate
interactions. The basic assumption in our assessment is that the
adsorption energy is mainly dictated by the interaction of the
monolayer-capped nanoparticle with the substrate, whereas
the diffusion energy is largely determined by the interaction
of gold nanocrystals with the substrate. These findings are
significant because it suggests that the manipulation of
the surface free energy, wettability, or adhesion can impact
the final size of the thermally processed nanoparticles on the
surface. An in-depth correlation of the thermally induced mor-
phological evolution of the nanoparticles with surface energy,
surface mobility, and molecular interactions will be useful for
exploring their applications in catalysis and electrocatalysis.
In addition to thermochemical treatments, other treatment
methods have also been demonstrated for the activation of
nanoparticle catalysts, including electrochemical,48 plasma,
photochemical, acid treatment,100 low-power sonication,59
UVozone treatment,60 and low T O2 plasma.101
7.04.2.4 Examples of the Metal Nanoparticles in
Heterogeneous Catalysis
There are increasing examples showing that the metal nanopar-
ticles prepared with capping layers function as highly active
catalysts for chemical reactions (see Table 1). This is largely
due to the well-defined size, shape, and composition for the
metal nanoparticles, in comparison with those prepared by
traditional methods. More importantly, it is the emergence of
the unique nanoscale properties because of the controllability
of the nanoscale parameters that impart enhanced catalytic
properties to the catalysts. Two recent examples are briefly
 Metal Nanoparticle Synthesis 85

discussed in the succeeding text to illustrate the significance of
the synthesis and processing methods discussed in this chapter.
The first example involves the study of binary and ternary
nanoalloy catalysts for oxygen reduction reaction in fuel cells.
For AuPt nanoparticle catalysts prepared by either molecular
monolayer encapsulation48 or dendrimer,52 one unique prop-
erty is the single-phase character and its variation different
from the bulk-scale counterparts. The phase and surface prop-
erties of supported AuPt nanoparticles have been demon-
strated to range from single-phase alloy and partial alloy to
phase-segregated structures depending on the preparation con-
ditions, bimetallic composition, and supporting materials.48
This conclusion is not only in contrast to the wide miscibility
gap well known for the bulk goldplatinum systems but also
shines a light on the importance of changes in physical and
chemical properties in which nanoscale multimetallic mate-
rials differ from their bulk counterparts. Synergistic catalytic
properties of the catalysts with different phase properties have
been demonstrated for methanol oxidation, formic acid oxida-
tion, and oxygen reduction.48,102 The surface properties due to
alloying and phase segregation as a function of the thermal
treatment temperature were further examined in terms of the
oxygen reduction activity in acidic electrolytes. In this case, the
mass activity was assessed in terms of Pt. For example, for
Au22Pt78/C treated under oxygen for 60 min followed by treat-
ment under hydrogen at temperatures ranging from 300 to
800  C (Figure 8(a)), the mass activity data showed an
increase of mass activity to a maximum at 400  C and further

O2 H2O

ORR activity structures thermal treatment

0.6 N2 (non oxdn.) O2 (partial oxdn.) O2 (full oxdn.)
H2
Mass activity (A/mgpt)

Support Support Support H2 H2

Alloyed or partially phase-segregated AunPt100-n/C
0.4
Size : size increases & Pt-specific surface area decreases
M-M: Pt-Pt bond decreases & enhanced heteroatomic alloying
M-O : Propensity of oxide decreases with increased alloying
0.2 Surface: increased enrichment of Pt & reduced Co/Ni oxides
0.6 1.5
MA (A/mgpt)

SA (mA/cm )
0.4 1.0
0.0
300 400 500 600 700 800 0.2 0.5

2
Treatment temperature (C) 0.0 0.0

(a) (b)

1.2 2.0

1.0
1.6
Specific activity (mA/cm2)
Mass activity (A/mgpt)

0.8
1.2
0.6
0.8
0.4

0.4
0.2

0.0 0.0
/C

i/C

C
o/

o/
Pt

iC
Pt

Pt

N
Pt

(c)

Figure 8 Electrocatalytic activity of binary and ternary nanoalloy catalysts for oxygen reduction reaction. (a) Mass activity at 0.858 V versus RHE
as a function of treatment temperature for Au22Pt78/C.48k Reproduced by permission of The Royal Society of Chemistry. (b) The ORR electrocatalytic
activity by the trimetallic catalysts as a function of the degree of oxidation in the ORC process, which are engineered by the controlled ORCRRC
treatment (top panel) and are correlated with a combination of metalmetal and metaloxygen coordination structures and their surface
redistributions (reprinted with permission from Wanjala, B; Fang, B; Loukrakpam, R; Chen, Y; Engelhard, M; Luo, J; Yin, J; Yang, L; Shan, S;
Zhong, C.J. ACS Catalysis, 2012, 2, 795806, Copyright 2012 American Chemical Society). (c) Comparison of ORR mass-specific activities at
0.900 V in 0.1 M HClO4. (Note that the mass-specific activity data were extracted directly from the RDE curves without IR-drop correction.)
86 Metal Nanoparticle Synthesis
decrease with increasing temperature. The surface treated
by the higher temperature showed Au enrichment on the sur-
face. The decrease of the activity with temperature is consistent
with the findings of the surface Au enrichment due to increased
phase segregation and the Ptcore Aushell formation.
For trimetallic PtNiCo nanoparticles for oxygen reduction,50
the metal coordination structures and their manipulation have
been shown to enhance the electrocatalytic activity for oxygen
reduction reaction. The basic hypothesis is that such atomic-
scale structure can be manipulated by oxidativereductive ther-
mal treatment to influence the binding site and strength of
molecular oxygen and oxygenated species on the nanoalloy
surface. The results have revealed remarkable increases in both
mass activity and specific activity for the catalysts processed
by the oxidativereductive treatment over those treated under
nonreactive or low-degree oxidative atmospheres before the
reductive treatment. A combination of the increase in the
heteroatomic alloying, the decrease in metal oxides, and
the enrichment of surface Pt by the oxidativereductive thermal
treatment has therefore been concluded to be responsible for the
enhanced electrocatalytic activity. The demonstration of this
new approach to manipulating the metal coordination struc-
tures forms the basis for an effective strategy in engineering
ternary nanoalloy catalysts and has provided new insights into
the role of such structures in the enhancement of the electro-
catalytic activity (Figure 8(b)).
A number of ternary catalysts have been demonstrated to
function as highly active oxygen reduction catalysts,49,50 includ-
ing PtNiCo/C, PtNiFe/C, PtIrCo/C, PtVFe/C, and PtVCo/C. The
catalysts not only showed enhanced catalytic activity over Pt/C
catalyst but also showed clear dependence on the nature of the
base metals alloyed into Pt-ternary nanoparticles. By comparing
the oxygen reduction mass-specific activities for a few Pt-ternary
catalysts with those of their Pt-binary counterparts, one impor-
tant finding is that the enhancement of the activity for the
ternary catalysts outperforms the binary ones. An example of
this finding is shown in Figure 8(c) comparing PtNiCo/C50a
with PtNi/C and PtCo/C.50b This finding demonstrates an
important role played by the detailed alloy composition in the
nanoparticle and the surface structures.
The ternary PtNiCo nanoalloy catalysts were also shown to
exhibit intriguing catalytic sites for CO oxidation reaction.103
How the supportnanoalloy interaction influences the cata-
lytic activity of the ternary nanoalloy particles on different
supports is an important fundamental question, which probed
by the oxidation reaction of carbon monoxide. The support
nanoalloy interaction is shown to influence structural and
chemical ordering in the nanoparticles, leading to support-
tunable active sites on the nanoalloys for oxygen activation in
the catalytic oxidation of carbon monoxide. The catalytic CO
conversion rates of the ternary PtNiCo/C catalyst are compared
with those of the binary PtNi/C and PtCo/C in both an as-
prepared state (Figure 9(a), i.e., the catalyst underwent
thermochemical treatment first under O2 and then H2 fol-
lowed by exposure to ambient air for more than 1 week) and
a re-activated state (Figure 9(b), i.e., the catalyst was re-
activated in situ under H2). For the as-prepared catalyst,
which is likely partially oxidized at the surface by air, PtNiCo
has an activity higher than PtCo and PtNi. PtCos activity is
slightly higher than Pt, whereas PtNi has almost no activity.
Interestingly, upon reactivation in situ by H2, which reduces
the particles back to alloy state, the activity for the ternary
catalyst shows an increase by factor of 23. PtCo shows the
largest increase in activity (by factor of 1020), though being
still lower than the ternary catalyst, especially at the high
reaction temperature. In contrast, PtNi shows practically
no change in activity, similar to Pt. Note that the observed
trends of catalytic activity remain unchanged even by taking
the differences in particle sizes into consideration, as reflected
by a comparison of the surface Pt-specific catalytic activities of
these catalysts.
Not only the degree of alloying but also the alloy composi-
tion have played important roles in the enhancement of the
catalytic activity, which can be understood by considering a
combination of several factors, including the propensity of oxi-
dation of the metals in the presence of oxygen, the reduction
potentials of the metals, the structure of the alloying, and most
importantly the surface activation of oxygen for the catalytic CO
oxidation reaction. The alloying of Pt with Co imparts the
catalyst with an enhanced capability for the activation of the
oxygen species, which explains the fact that PtCo/C after reacti-
vation has much higher catalytic activity than pure Pt/C. The
higher propensity of Co and the lower propensity of Ni to
surface oxidation or oxygenation are apparently taken advantage
of, as reflected by the relatively high stability of the as-prepared
ternary PtNiCo catalyst in ambient condition and the catalytic
activity higher than those of the bimetallic counterparts.
Oxygen activation is crucial for the CO oxidation, which
can occur on both the surface PtCo/Ni-active sites (type I)
of the nanoalloy particle and the anionic deficiency sites
(type II) of the support for the catalysts, thus substantially
influencing the overall reaction rate. For the reduced nanoal-
loys (Figure 9(b)), the differences among the catalytic activi-
ties of the three supported catalysts cannot be explained in
terms of type II sites since silica and carbon supports do not
have them. When the surface is covered by Co/Ni oxide species,
type I sites are basically blocked (Figure 9(b)). This is consis-
tent with the poor activities of all three oxidized catalysts, even
in the case of TiO2 support where type II sites are present. Thus,
if the fraction of surface base metals is significant, then the
blocking will be more serious, which explains the observation
of the propensity of deactivation of the carbon catalysts con-
taining higher percentages of base metals. The origin for the
differences among the catalysts on the three different supports
then can mainly come from type I sites. These sites and the
degree of CoPtNi bonds local disorder are very much influ-
enced by support nanoparticle. A strong interaction, as in the
case of TiO2, favors the formation of structurally ordered (i.e.,
with a low degree of local atomic positional disorder) but
chemically disordered (i.e., with a random distribution of the
base transition metals and Pt species with respect to each
other) nanoalloys. This leads to an enhanced population of
CoPtNi (type I) sites for effective oxygen activation. A
weaker supportnanoparticle interaction, on the other hand,
leaves more room for local structural distortions and rearran-
gement of the chemical species with respect to each other.
The nanoalloys could undergo structural evolution upon the
thermochemical treatment in which CoPtNi (type I) sites
rearrange to influence the catalytic activity, as in the case of
carbon-supported nanoalloys. For a weak SNI, as is the case
 Metal Nanoparticle Synthesis 87

20.00 20.00
CO oxidation rate (10-6 mol gpt-1 s-1)
Pt/C

CO oxidation rate (10-6 mol gpt-1 s-1)

Pt/C
PtCo/C
15.00 PtNi/C 15.00
PtCo/C
PtNi/C
PtNiCo/C
10.00 PtNiCo/C
10.00
5.00 5.00

0.80 0.80

0.40 0.40

0.00 0.00
40 60 80 40 60 80
(a) Temperature (C) Temperature (C)

CO2 CO2

O2 O2
d- 2- 2- 2- 2-
I 2- 2-
2- 2-

CO2 Nanoalloy CO2 Nanoalloy

O2 SNI O2 SNI
II Support II Support

Pt Co Ni CO Od- d- O2- 2-

(b)

Figure 9 (a) Comparisons of Pt mass-specific CO oxidation rates of Pt39Ni22Co39/C with those of Pt64Ni36/C, Pt45Co55/C, and Pt/C catalysts at two
different states: as-prepared and reactivated states. The Pt mass-specific CO oxidation rate is normalized against the metal loading on carbon
and the Pt composition in the nanoparticles. (b) Illustration of the proposed catalytic CO reaction on a supported nanoalloy catalyst in either reduced
nanoalloy or oxidized nanoalloy states, involving oxygen storage and release capacity where the transition base metals Ni/Co provide sites for the
activated O2 and Od species. While both cases may involve type II sites for O2 activation, the type I site for O2 activation exists only on reduced
nanoalloy surface and does not exist on the oxidized nanoalloy surface where the surface is completely blocked by the oxide species. Reprinted
with permission from Yang, L.; Shan, S.; Loukrakpam, R.; Petkov, V.; Ren, Y.; Wanjala, B.; Engelhard, M.; Luo, J.; Yin, J.; Chen, Y.; Zhong, C.J. J. Am.
Chem. Soc., 2012, 134, 1504815060, Copyright 2012 American Chemical Society.

with silica support, the chemical species in the ternary nanoal-
loys segregate, which diminishes the CoPtNi (type I) sites
and resulting in a decreased catalytic activity.
7.04.2.5 Summary
Various strategies have been developed for the synthesis of metal
and alloy nanoparticles capped with monolayer, dendrimeric,
and polymeric capping layers, and such metal nanoparticles
have been shown to be active for various catalytic or ele-
ctrocatalytic reactions upon different catalytic activations. In
addition to the nanoparticle composition, the atomic-scale
structuring and ordering of the metal nanoparticles have been
shown to play an important role in determining the catalytic
and electrocatalytic properties. Key to some of the highly active
catalytic properties of such catalysts is the ability to control
the size, composition, shape, and phase structures by nanoengi-
neered synthesis and processing of the metal or alloy
nanoparticles for the nanostructure formation, catalytic activa-
tion, and structural perfection. From the examples of bimetallic/
trimetallic nanoparticles highlighted in this chapter, it is evident
that the detailed nanostructures and the surface catalytic sites of
the as-synthesized metal and alloy nanoparticles can be further
engineered or perfected by postsynthesis thermochemical pro-
cessing for the enhancement of the catalytic activities or stabil-
ities. This attribute serves as an important strategy for the design
and preparation of metal and alloy nanoparticle catalysts.
7.04.3 Metal Nanoparticle Synthesis at the
LiquidSolid Interface
7.04.3.1 Modes of Contacting Metal Precursor Solutions
with Solid Supports
The most common methods to synthesize metal nanoparticles
on high-surface-area supports are summarized in Table 3. The
88 Metal Nanoparticle Synthesis

Table 3 Types of catalyst impregnation

Type of impregnation Distinguishing characteristics Advantages/disadvantages

Dry impregnation (DI)a
Wet impregnation (WI)
Grafting
Strong electrostatic
adsorption (SEA)b
Depositionprecipitation (DP)
a
Also known as pore filling and incipient wetness.
b
Also known as ion adsorption.
Use sufficient metal solution to fill pore volume of
catalyst support; adjust metal concentration for
desired weight loading
Amount of solution in excess of pore volume of support;
precursors which interact weakly with support are
washed/filtered away
Driving force exists for reaction of precursor with
support
Excess solution; pH held at optimal value for strong
precursor interaction with support
Excess solution; pH slowly and homogeneously
increased to precipitate precursor at support surface
A: simplest to employ; no filtering; metal content is fixedD:
strong precursorsupport interaction is not guaranteed
A: mixing is improvedD: filtering required; metal loading
must be measured, metal wasted if it does not strongly
interact with support
A: strong interaction of precursor with support is
assuredD: limited in applicability to certain systems
A: strong, monolayer adsorption of metal precursorD:
optimal pH must be determined and achieved during
preparation
A: high metal loadings are most easily achievedD: pH must
be altered during preparation

first, dry impregnation, is procedurally the simplest. A thick

paste is formed by contacting a support with just the amount of
solution needed to fill the supports pore volume. The solution
contains a known amount of metal, and since the paste is not
filtered, the metal loading can be precisely controlled. The
drawbacks from this method arise when metal precursors do
not interact strongly with the support surface. Metal complexes
that remain in solution can migrate significantly during drying.
Impregnations can be considered wet whenever an
amount of solution in excess of the support pore volume is
employed. Wet impregnation is as simple as contacting the
solution for a certain time and then recovering the solid by
filtration. The amount of metal retained by the solid must be
determined by analysis of either the solid or liquid. The extent
of metal retained by the support is a function of the precursor
support interaction, which may be chemical or physical (elec-
trostatic) in nature. In particular cases where metal precursors
do chemically react with surface functional groups, such as
silanols, the deposition method can be termed chemical graft-
ing or simple grafting. As WI is defined here, no grafting occurs
and no attention is paid to controlling the impregnation con-
ditions (varying pH, for example) to achieve any sort of pre-
cursorsurface interaction.
One way strong interactions can be created is via the electro-
static adsorption mechanism1518 illustrated in Figure 10. An
oxide surface contains terminal hydroxide groups that protonate
or deprotonate depending of the acidity of the impregnating
solution. The pH at which the hydroxide groups are neutral is
termed the PZC. Below this pH, the hydroxide groups protonate
and become positively charged, and the surface can adsorb
anionic metal complexes such as CPA. Above the PZC, the
hydroxide groups deprotonate and become negatively charged,
and cations such as PTA can be strongly adsorbed.
Another impregnation method often involving the manip-
ulation of solution pH is DP. The general idea is that by slowly
and homogeneously altering solution conditions, metal com-
plexes can be induced to precipitate in a controlled fashion on
the support surface as small particles, even at high loadings
over oxide and carbon supports.
These methods will now be reviewed in more detail as they
apply to single and bi- or multimetallic nanoparticle synthesis
Kads
pH<PZC + [PtCl6]2
OH2
K1
PZC OH
K2
Kads
pH>PZC O [(NH3)4Pt]+2
Figure 10 Schematic of electrostatic adsorption.
7.04.3.2 Single Metals
7.04.3.2.1 Dry and wet impregnation
Several of the impregnation modes are depicted in Figure 11.
Typical catalyst supports have pore volumes of about 1 cm3 g1.
The amount of acid or base (protons or hydroxides) in 1 cm3 of
metal precursor solution is almost always orders of magnitude
less than the number of surface hydroxide groups in 1 g of
support, and consequently, at dry impregnation conditions,
the support surface is essentially uncharged, which is to say,
the solution pH is buffered to the PZC of the support, and no
electrostatic interaction can occur.104 In the absence of other
precursorsurface interactions such as chemical grafting, metal
precursors remain in the liquid phase and can be transported to
the drying surface at the exterior of the paste.
Experiments with a range of solids confirm that most of the
moisture is removed during the period of constant drying rate
after impregnation in which water is transported to the external
surface.105107 The result is that substantial fraction of the dis-
solved precursor is deposited on the external edge of the support
bodies. Electron microscopy images of metal nanoparticles pre-
pared by dry impregnation are compared to those prepared by
SEA in the next section (Figure 14). Not only are the particles of
DI preparations larger in size (more poorly dispersed) and exist
 Metal Nanoparticle Synthesis 89

Dry impregnation Wet impregnation Strong elec. adsorption Charge enhanced

- pore filling - excess liquid - excess liquid Dry impregnation
- pH PZC - pH PZC - pH pHoptimal - pore filling
- poor dispersion - poor dispersion - good dispersion - pH pHoptimal
- wide size distr. - wide size distr. - narrow size distr. - good dispersion
- narrow size distr.

Figure 11 Depiction of several of the impregnation modes.

in a wide variation in particle sizes, but also dense agglomerates External surface
of particles often occur; these would likely be the particles which
form at the drying surface (Figure 12).
Several means are available to mitigate the transport of
metal during drying. The first is to adjust the drying rate. Van
den Berg and Rijnten105 presented results indicating that fast
drying brings about preferential deposition at the external
edge. Komiyama, Merrill, and Harnsberger109 mentioned that
rapid drying leads to deposition of nickel(II) chloride impreg-
1 2
nated in g-alumina near the center of the support bodies. External surface
A second method is to use precursors or additives to the
impregnating solution, which increase the viscosity of the
solution during drying, which essentially traps the metal in
place. Geus has recently written an excellent review of viscous
drying.108
Referring back to Figure 11, wet impregnation is named as
such since the volume of solution used is in excess of the pore
volume of the solid. This makes stirring easy and ensures a
homogeneous initial distribution of metal precursor in the 3 4
pore volume of the solid; however, the problem is that, like Figure 12 Movement of metal solution within support bodies during
DI, if there is no chemical interaction between the metal precur- drying. First, gas bubbles form within the liquid due to the capillary
sor and support, there will be little metal actually deposited on suction of the narrow pore mouths at the external surface (two pictures at
the support. Even at high ratios of solution to solid, the number the top). Next, the pores empty, which leads to transport of liquid along
of hydroxide groups on typical support surfaces is so large that narrow voids at the contact areas of the elementary particles of the
the surfaces do not appreciably charge (deprotonate or support to take over the transport of liquid to the external surface of the
protonate).104 The pH is still normally buffered to the PZC of support body (picture at the bottom).108
the support. Without any interaction, the only metal that
remains in the support after filtration is the amount of metal so as to achieve strong electrostatic attraction between precur-
left in the pore volume; the majority of the metal remains in the sor and surface.
solution phase and is wasted. And once again, drying causes the The SEA procedure is summarized in Figure 13 for a number
migration of metal precursors and large agglomerates form. of silica supports. The first step of the procedure is to determine
the PZC of the support by the pH shift measurements shown in
7.04.3.2.2 Strong electrostatic adsorption and charge- step 1. The acidic PZCs of these silicas (all about 4) indicate that
enhanced dry impregnation the surface tends to deprotonate and charge negatively, so the
The method known as SEA has been discussed in detail in a silicas are matched. The next step is to determine the pH of
number of recent book chapters.1518 SEA is normally per- strongest interaction with the uptake pH survey shown in
formed with an excess of solution as depicted in Figure 11. step 2. Next, higher quantities of material are prepared at this
This keeps pH shifts to a minimum and facilitates sampling optimal pH (about 10.5), and after drying, a mild reduction
the solution. Thus, SEA at high liquid/support ratios is a gives the ultrasmall nanoparticles (about 1.0  0.3 nm) seen
special case of WI in which the pH of the solution is controlled in step 3.
90 Metal Nanoparticle Synthesis

Kads 2. Uptake pH survey

pH<PZC
OH2+ [PtCl6 ]-2
1
K1
L90 3. Reduce to retain high dispersion
0.8 M7D

Pt adsorbed (mol m2)

PZC OH EH5
H2
VN3S
K2 0.6 FK300
Model
Kads 0.4
pH>PZC O- [(NH3)4Pt]+2
0.2

1. PZC determination 0
14
0 2 4 6 8 10 12 14
pH final
12

10 L90
M-7D
pH final

8
EH-5
VN-3S
6
FK300
4 Model
0.02 m
2

0
0 2 4 6 8 10 12 14 [(NH3)4Pt]2+/SiO2
pH initial

Figure 13 The strong electrostatic adsorption procedure: (1) PZCs are determined and the appropriate metal complex is selected (a cation for silica),
(2) uptakepH surveys are conducted to determine the pH of strongest interaction, and (3) catalysts are synthesized at this pH (10.5), and after
mild reduction, ultrasmall nanoparticles with a narrow size distribution result.

Typical SEA and dry impregnation preparations are con- is no evidence of agglomeration in any of the CEDI samples, as
trasted in Figure 14, from Jiao and Regalbuto.110 The SEA observed for the silica DI samples in Figure 15(b).
preparations give much smaller particle size with much smaller
standard deviations than the DI preparations. The agglomera-
tions of particles seen in Figure 14(b) are typical of DI prepa-
7.04.3.2.3 Depositionprecipitation
rations and are likely caused by the accumulation of metal in
The synthesis method of DP initially patented by IG Farben117
the solution phase during drying.
and further developed by Geus et al.118120 involves the addi-
Strong electrostatic adsorption has been demonstrated for
tion of a precipitating agent added homogeneously throughout
Pt/alumina,16 Pt/silica,111 other noble and base metals on
a solution generally consisting of a metallic salt precursor
amorphous110 and mesoporous silica,112 Pt113 and Pd114 on
in suspension with the catalyst support. The most common
carbon, and Pt on titania.115
precipitating agents used are sodium hydroxide121124 and
There is no reason other than experimental convenience
urea.121,125 The support surface then becomes a nucleating
that SEA is applied at high ratios of liquid to solid. In fact,
agent for precipitation of the metal hydroxide complexes in
incipient wetness represents the highest obtainable surface
the basic pH range when the solubility limit of the metal
loading for a particular solid. With enough acid or base
hydroxides is reached. In order to precipitate the metal hydrox-
(and these amounts are surprisingly large), final pH can be
ides on the support uniformly, it is imperative that a homog-
taken to the optimal value. We call this simple yet effective
enous hydroxide ion concentration exists in solution, which
variation charge-enhanced dry impregnation (CEDI); it is
will ensure the hydroxide precipitates on the support surface
depicted in Figure 11(d) and is SEA conducted at DI
instead of in solution. The method pioneered by Geus118120
pore-filling conditions.
uses the slow decomposition of urea at elevated temperature
CEDI has recently been demonstrated for improving metal
(323363 K)
dispersion and size distributions116 with three catalyst sup-
ports and two metal complexes. The STEM analysis from this CONH2 2 3H2 O ! 2NH 
4 CO2 2OH
study is shown in Figure 15. The average sizes of nanoparticles
prepared by CEDI are 1.5 nm for carbon, 1.8 for silica, and 2.9 to evenly disperse hydroxide ions, which limits the formation
for alumina and are much smaller particles than those pre- of pH gradients; this reduces nucleation in solution and
pared by DI, 10.3, 10, and 10 nm, respectively. The CEDI sizes increases metal hydroxide deposition on the support surface.
are not quite as small as those obtained from SEA but are close Deposition precipitation has many advantages over dry and
and at any rate are far better than those obtained with DI. This wet impregnation techniques. DP has a higher reproducibility
 Metal Nanoparticle Synthesis 91

20 nm 20 nm

(a) (b)

20 nm 20 nm

(c) (d)

Figure 14 Representative STEM images and corresponding particle distributions of reduced Ru and Ni catalyst samples: (a) SEA preparation, 3.0 wt%
Ru, particle size 1.2  0.21 nm; (b) DI preparation, 3.0 wt% Ru, particle size 13  4.7 nm; (c) SEA preparation, 1.6 wt% Ni, particle size
1.7  0.36 nm; and (d) DI preparation, 1.6 wt% Ni, particle size 3.2  1.7 nm. Reprinted from Jiao, L.; Regalbuto, J.R.; J. Catal. 2008, 260, 342350,
with permission from Elsevier.

with the nanoparticles synthesized showing a tighter size dis-
tribution even at higher metal loadings. Using urea as the
precipitating agent, up to 8% Au/TiO2 had an average particle
size of 1.8 nm126,127 as shown in Figure 16. The nucleation of
metal hydroxide ions at the surface also ensures the deposition
of metal particles stay on the surface,123 which can increase
thermal stability and metal support interactions.128 Recently,
much of the research using DP to synthesize nanoparticles has
been focused on Au, Ni, Co, and Pt and Pd systems128132 on
various metal oxide supports (TiO2, SiO2, and Al2O3) and,
more recently, carbon nanofiber supports for Ni and Fischer
Tropsch catalysts.132,133 Synthesis of gold nanoparticles using
DP of Au(OH)3 has been extensively studied in the literature
by Haruta and Louis122124,134 with much of the work being
focused on low-temperature CO oxidation. It has been
found that Pt and Au catalysts synthesized via DP show an
enhanced activity for CO oxidation compared to photodeposi-
tion and impregnation methods of preparation.131,135 Higher
gold loading with nanoparticles on the order of 13 nm has
been produced using both NaOH and urea depending on the
desired loading of gold.135
More recently, carbon nanofibers are being employed as
supports for DP synthesis of nanoparticles.132,133 The deposi-
tion of cobalt onto carbon nanofibers requires a basic pH to
produce Co3O4 on the carbon nanofiber surface that was
reduced at 350  C to give 8 nm particles that showed increased
reactivity toward the FischerTropsch reaction and shows
promise for future catalytic work.132
92 Metal Nanoparticle Synthesis

Carbon Silica Alumina

(a) (b) (c)

Dry
impregnation

50 nm 0.05 m 50 nm

(d) (e) (f)

Charge enhanced
dry impregnation

50 nm 50 nm 50 nm

Figure 15 STEM images of Pt catalysts after reduction at 200 C (a) Pt/C DI (10 nm), (b) Pt/C CEDI (1.8 nm), (c) Pt/SiO2 DI (7.5 nm), (d) Pt/SiO2 CEDI
(1.5 nm), (e) Pt/Al2O3 DI (11 nm), and (f) Pt/Al2O3 CEDI (2.9 nm).

50

40
Frequency (%)

30

20

20 nm
10

0
0.2
0.7
1.2
1.7
2.2
2.7
3.2
3.7
4.2
4.7
5.2
5.7
6.2
6.7
7.2
(a) Particle size (nm)

50

40
Frequency (%)

30

20

10

20 nm 0
0.2
0.7
1.2
1.7
2.2
2.7
3.2
3.7
4.2
4.7
5.2
5.7
6.2
6.7
7.2

(b) Particle size (nm)

Figure 16 TEM images and size histograms of (8%) gold particles of DP urea 16 h reduced under H2 at 120  C (a) and at 500  C (b).127 Reprinted with
permission of the American Chemical Society.

7.04.3.2.4 Grafting
Grafting is another method reported in the literature to synthesize
supported metal catalysts. In this case, the functional groups (e.g.,
hydroxide groups) of a solid support (generally, inorganic oxide,
e.g., silica and alumina) develop strong and stable covalent bonds
with a transition metal complex (e.g., metal alkoxides, metal
halides and oxyhalides, organometallics, and metal carbonyls)
serving as precursor. Consequently, a single-sited catalyst deposits
irreversibly and becomes immobilized on solid support while
maintaining the same (anchoring) or different (grafting) chemi-
cal surrounding as its parent precursor.136138
A schematic of the grafting technique applied to prepare
supported catalyst is depicted in Figure 17.136 Metal complexes
can graft onto support via single or multiple bonds.138 Cover-
age of metal obtained from grafting can be quite low depend-
ing on the number of hydroxide groups on the support surface
and reaction temperature.139 Hence, sequential grafting or
grafting cycle repeated several times can be employed to con-
trol metal loading on the surface from submonolayer to mul-
tilayer coverage.140,141 Supports are often pretreated thermally
to drive off physisorbed water and avoid unwanted reactions
occurring at the surface.141,142
Numerous studies over the last few years have used grafting of
organic groups to effectively modify or even add different useful
functionalities to different supports, particularly mesoporous
silica. Two routes are commonly applied to modify surface via
grafting: (1) postsynthesis grafting and (2) co-condenzation with
both having discrete set of advantages and disadvantages.143,144
Properties, for example, acid strength and wettability of support,
could be modified significantly via grafting to facilitate specified
catalytic application.139,145149 In another set of studies, it was
demonstrated that different oxide supports, for example, ZrO2,
TiO2, and SiO2, grafted with vanadyl tripropoxide produce cata-
lysts with improved redox properties.147 Organometallic grafting
has successfully been used to produce highly dispersed catalysts
with commendable catalytic activity for methane and propane
oxidation.150 Stability and catalytic activity of mesoporous mate-
rials can be improved remarkably by grafting with different metal
complexes.151 Catalysts synthesized via grafting showed promise
in base-catalyzed reactions and epoxidation of olefins.152,153
MoIV catalyst prepared via grafting was reported to show 11
times as high activity as that obtained via impregnation.154
Recently, in another study, Ni grafted on mesoporous silica
demonstrated highest activity in CO2 reforming of methane.155
Grafting of ionic liquids has been found to develop an array of
physical and chemical properties and is considered to have great
potential in a number of reactions.156
OH
Chemical/thermal
OH + Complex M Complex
OH treatment
OH
Metal complex
Support
Figure 17 Schematics of grafting of a representative metal complex onto inorganic support.
Metal Nanoparticle Synthesis 93
Unlike methods such as dry impregnation, synthesis via
grafting can prevent migration and sintering and exhibits
improved strength and stability. Nevertheless, applications of
grafting in industrial catalyst are still few and far between,
largely due to its stringent environment requirement and the
complexity associated with the synthesis of most of the com-
mercially unavailable metal precursors. Besides, in order to
achieve high metal loadings, grafting cycle needs to be repeated
several times, which often makes the process long and arduous.
In most cases, the interaction chemistry between solid and
liquid phases is not clearly understood. However, an increasing
number of cases related to grafting nanoparticle onto solid
surface are being studied presently, which may help improve
the understanding and reduce the complexity associated with
this highly potential catalyst synthesis technique.
7.04.3.3 Bimetallic and Multimetallic Catalysts
A recent bibliometric study of bimetallic catalyst synthesis by
one of the authors revealed that in over three quarters of the
published work, coimpregnation (co-DI) was used to deposit
the two metals. Certainly, the simplicity and precision of metal
loadings of this method are attractive. On the other hand,
when synthesizing a bimetallic catalyst, not only is high dis-
persion desirable, but most often, intimate contact of the two
metals is also desired. Much work on computational catalysis is
being conducted in the present day to predict optimal formu-
lations of bimetallic catalysts; much less work is going into
synthesizing the metal nanoparticles to actually achieve a high
degree of metal1metal2 interaction.
7.04.3.3.1 Coimpregnation
While dry impregnation with multiple metals is the most pop-
ular way to synthesize bimetallic catalysts at present, using this
method with bi- or multimetallic catalysts suffers from the
double disadvantage of the absence of precursorsupport
interactions and the absence of precursor1precursor2 interac-
tions. These disadvantages can be overcome in a small number
of co-DI systems, such as Pt/Re/carbon,157 where Pt tends to
form small particles on carbon by a reactive adsorption113 and
the oxidic Re precursor phase is mobile and can find the Pt
surface during catalyst pretreatment.157 In the majority of other
cases, however, the results appear as in Figure 18, which
resulted from the co-DI of Pd and Co precursors.158 The largest
particles are pure Pd, and the smaller particles are pure Co.
OH
O
M
O
OH
94 Metal Nanoparticle Synthesis

7.04.3.3.2 Co-SEA, co-CEDI, and sequential SEA
The exploitation of electrostatic adsorption for the preparation
of bimetallic catalysts is just beginning. As drawn in Figure 19,
co-SEA is hypothesized to form a well-mixed monolayer of pre-
cursors, which, after a low-temperature reduction, will assemble
into well-dispersed, homogeneously alloyed particles.
Recent results have confirmed this hypothesis with tetra-
ammine Pt and Pd precursors adsorbed by co-SEA over silica
and oxidized carbon supports of similar PZC (4).159 These
are compared to co-DI samples prepared from the same pre-
cursors and supports in Figure 20. Co-DI once again resulted
in large particles with many agglomerates; elemental mapping
by EDXS also revealed significant inhomogeneity of alloying.
In contrast, co-SEA yielded small, well-dispersed Pt/Pd bime-
tallic particles of average size of about 1.0 nm for both silica
and oxidized carbon supports. Elemental mapping confirmed
a high degree of Pt/Pd alloying by co-SEA.
Just as CEDI can dramatically improve the efficiency of dry
impregnation preparations, it appears that co-CEDI can
also improve the efficiency of co-DI preparations. The Pd/Co/
carbon materials shown in Figure 18 (for co-DI) were also
synthesized by co-CEDI, and the STEM results are shown in
Figure 21. Powder XRD and TPR158 were used to confirm the
presence of homogeneous PdCo alloys. Compared to the
co-DI sample, the co-CEDI particle size is much smaller, the
size distribution much tighter, and the degree of interaction
much more intimate.
Sequential or seq-SEA as depicted in Figure 19 (right)
involves an initial deposition of a core metal precursor, oxidiz-
ing the precursor to form the metal oxide, and then selectively
adsorbing the second metal precursor onto the first metal
oxide. After reduction, assuming the two metals do not blend
to a significant extent, the second metal will exist as a partial
shell on the first metal core. An initial set of data has been
collected for the Pd/Au/carbon system.160 While STEM images
requiring atomic resolution x-ray maps have not yet been
obtained, the degree of alloying has been inferred with powder
XRD and reactivity results.
In Figure 22, a co-DI-prepared 2.5 wt%Au/2.5 wt%Pd/
carbon (Darco) catalyst has been compared with Au/Pd/
carbon (Darco or Vulcan) catalysts prepared by selectively
adsorbing Au bis-ethylene diamine complexes onto Pd
nanoparticles.160 The coimpregnated catalyst contains large
metallic particles of an Au-rich phase (about 20 nm by XRD
line broadening) and exhibits low conversion of benzyl alco-
hol (2.5Au/2.5Pd co-DI sample on left). In contrast, the
SEA-prepared Au/Pd nanoparticles are below the limit of
XRD detection, and the best catalyst exhibits over ten times

0.1 m 0.05 m
0.05 0.1
0. 1 m

(a) (b) (c)

Figure 18 STEM images of (a) (5 wt% Co 5 wt% Pd)/VXC, (b) (5 wt% Co 2.5 wt% Pd)/VXC, and (c) (5 wt% Co 5 wt% Pd)/BP2000 catalyst
prepared by dry impregnation. Scale bar in a and c is 1000 A and in b is 500 A. The particle size and standard deviation in a, b, and c are 126  81,
124  85, and 134  90 A, respectively.

A
+
A
+ +
A B
+
B+ B+
A+ B+ B+ +
A A
+
A+ B+ B
+
B+
A+ B+ A+
A SEA, O2 B SEA, O2

SEA, dry AO2 AO2 AO2 BO2 BO2 BO2

+ + + +
B B A A
B
+
B
+
A+ A+
B+ B
+ A+ A+
+ + + + + + + +
A B A B A B A B
+ AO2 ++ AO2 ++ AO2 + + BO2 ++ BO2 ++ BO2 +

Low T, H2 Low T, H2 Low T, H2

B B B A A A
AB AB AB A A A B B B

Figure 19 Schematic of co-SEA (left) and sequential SEA (right).


Silica
Co-DI
50 nm
Co-SEA
10 nm
Figure 20 STEM images of Pt/Pd particles with silica and oxidized carbon. EDXS maps and line scans of individual particles have revealed that the
agglomerates are poorly alloyed and the co-SEA nanoparticles are homogeneously alloyed.
0.0 mm
0.02 0.0 mm
0.02
(a) (b)
Figure 21 STEM images of (a) 5 wt% Co 5 wt% Pd and (b) 5 wt%
Co 2.5 wt% Pd. Scale bar is 20 nm. The particle size and standard
deviation in a and b are 3.6  1.2 and 5.3  1.7 nm, respectively.
the activity of the impregnated catalyst with 25 times less Au
(the 0.1 wt%Au/2.5 wt%Pd/Darco catalyst). Sequential SEA
appears to be a highly effective way to deposit (sub)monolayer
of a second metal onto a first metal.
7.04.3.3.3 Reductionoxidation methods in preparing
bimetallic catalysts
The fine tuning of the bimetallic catalyst surface composition
rather than the bulk is an effective way to change its catalytic
properties. This can be achieved by surface redox reactions
where a second metal is reduced from solution and deposited
onto a primary or seed catalyst covering just a fraction of or
the entire seed metal surface. These redox reactions provide a
controlled method of depositing the secondary metal on the
surface of the primary metal, and not on the support
surface.161,162 One way of carrying this out is by direct redox
reaction due to the favorable difference of electrochemical
potentials between the seed metal oxidation and reduction of
Metal Nanoparticle Synthesis 95
Oxidized carbon
20 nm
20 nm
the secondary metal from solution.163,164 This is sometimes
referred to as galvanic displacement.165,166 Another method is
to have a reducing agent present, which reduces the secondary
metal precursor in the presence of the primary metal, referred
to as electroless deposition (ED). Such reducing agent can be in
solution161,162,167171 or preadsorbed on the surface of the
seed metal.163,172,173
Galvanic displacement can be a quick and easy method of
preparing a bimetallic catalyst. However, it is limited only to
pairings of the seed and secondary metals where displacement
is thermodynamically favored. It could not be implemented in
a bimetallic catalyst system where the choice of seed and
secondary metal is exchanged and thus thermodynamically
unfavorable.163,164 As illustrated in Figure 23, there should
be a net positive potential for the displacement process to
proceed spontaneously.
In the direct redox reaction method, the replacement of seed
metal with the secondary metal on the surface, through time, can
diminish the amount of available sites for the reaction to occur.
Exposed atoms of the primary metal, and enough of the second-
ary metal precursor in the solution, are requisite for the reaction
to proceed. Increasing coverage of the surface by the secondary
metal can result in kinetic limitations prohibiting deposition,
which can lead only up to monolayer loadings.164,174 However,
geometry and other simultaneous processes during deposition
can expose more seed metal to be available for displacement and
can create alloyed particles.165 Deposition of the additive metal
can be preferential such that the deposited atoms cover crystal
faces, where atoms have high coordination number.163 This can
be a problem if the desired catalytic activity is dependent on the
availability of the base metal faces.
In ED, a reducing agent is activated on the surface of the
primary metal. The secondary metal precursor catalytically
96 Metal Nanoparticle Synthesis

BnOH oxidation (40 ml-O2-10 bar, 140 C, 25 mg)

100
180 000 0.1Au/2.5Pd/Darco
90 0.1Au/2.5Pd/Vulcan
160 000

Benzyl alcohol conversion (%)

2.5Au/2.5Pd Co-DI
80
140 000
Blank
70 0.5Au/2.5Pd/Vulcan
120 000

60 0.5Au/2.5Pd/Darco
100 000 Darco G60
Intensity

80 000
50
2.5Pd
60 000
40
0.5Au 2.5Pd
40 000 30
0.1Au 2.5Pd
20 000 20
Co-DI
0 10

20 000 0
10 20 30 40 50 60 70 80 0 0.5 1 1.5
2 theta Time (h)

Figure 22 XRD patterns (left) of Darco co-DI and SEA-prepared catalysts and activity results (right) for benzyl alcohol oxidation for Au/Pd/carbon
catalysts.160

x+
M1(s) M1(aq) + xe- Eeq1
y+
M2(aq) + ye- M2(s) Eeq2
y+ x+
yM1(s) + xM2(aq) xM2(s) + yM1(aq) E>0

y+ x+
M1(s) M2(aq) M2(s) M1(aq)

Figure 23 Illustration of the direct redox reaction or galvanic displacement method.

(d)

(a) (b) (c)

y+
M1(s) M2(aq) RA RA(activated) M2(s)
(e)

Figure 24 Illustration of the ED method: A: immersion of parent catalyst into the ED bath, B: adsorption and activation of reducing agent, C: reduction
and deposition of secondary metal, D: catalytic deposition, E: autocatalytic deposition.

reduces as it comes in close proximity to these sites.168 It is also
known as electroless plating and has been used in industry as an
alternative to conventional electrochemical plating.167,171,175
Figure 24 illustrates the mechanism of secondary metal deposi-
tion onto the primary metal. The combined anodic and cathodic
electrochemical partial reactions occurring in the ED system are
similar to that of the direct redox reaction. The overall potential
of the reduction of the secondary metal precursor (cathodic) and
the oxidation of the reducing agent (anodic) has to be thermo-
dynamically favorable.
Upon initial deposition of the additive metal, the process
can proceed with two different deposition mechanisms: cata-
lytic, where the secondary metal deposits on activated sites on
the seed metal, or autocatalytic, where the deposited secondary
metal can be activated with exposure to the reducing agent.162
Codeposition of two metals can also be done by ED.169,170
In codeposition, the precursors of both the primary and the
secondary metal are present in the ED solution, which are
then simultaneously reduced onto the surface of the seed
catalyst. This method creates an alloyed surface of the catalyst,
and the principle can be extended to synthesis of multimetallic
catalysts.
Solution stability is a major requirement in the formulation
of electroless developer baths. At the condition at which the ED
reaction is carried out, the secondary metal should reduce only
in the presence of the seed metal and that there is no reduction
or thermal decomposition of the precursor in solution.161,167
Stabilizers or chelating agents like citrate and EDTA are some-
times added to enhance bath stability.167,168,171 The reducing
agent used in the bath should also be active at the conditions
where ED is carried out. A host of reducing agents, such as
NaBH4, dimethylamine borane, NaH2PO2, and hydrazine, are
widely used in ED.176182 Electroless developer baths using
these reducing agents typically have high pH as these have

high activity in basic conditions.167,171,175,183 Some reducing
agents like formic acid and ascorbic acid have been proven to
be active for ED in acidic conditions.169,170,184 An example of
the controlled deposition of various amount of Pt onto
carbon-supported Co is shown in Figure 25, from a submono-
layer in Figure 25(d) to many shell layers in Figure 25(a).
The appropriate pH of the ED bath is chosen such that the
secondary metal only deposits on the first metal so that there is
strong interaction between the two and a bimetallic catalyst is
synthesized. Conditions should be avoided where the second-
ary metal precursor becomes adsorbed on the support, which is
the case of SEA.16,167,168
ED is a reliable method of depositing low loadings of the
additive metal onto a catalyst surface while significantly mod-
ifying catalyst properties. It can also take advantage of the
anchoring properties of the primary metal to prevent agglom-
eration under catalytic conditions while providing high disper-
sion of the secondary metal.167,168
7.04.3.3.4 Codepositionprecipitation
Codepositionprecipitation (co-DP) follows the principle of
DP. Two (or more) metal precursors simultaneously were
mixed in aqueous solution with adjusting proper pH or adding
proper precipitating agent, such as urea, and then followed by
the addition of support materials, due to metalmetal and
metalsupport interactions, leading to form of smaller particle
size than that of catalyst prepared by conventional impregna-
tion methods and should show higher catalytic activity in
reactions.185192
(a) Pt3.5Co (23 wt% Pt) (b) Pt2.4Co (15.8 wt% Pt)
2 nm
2 nm
(c) Pt0.11Co (0.7 wt% Pt) (d) Pt0.07Co
2 nm
(0.46 wt% Pt)
2 nm
Figure 25 Z-contrast STEM images of PtCo bimetallic particles
prepared by ED of Pt on carbon-supported Co particles. The weight
loadings of Pt that were deposited and the resulting empirical formula of
the bimetallic particles are shown for each image. The gold-colored line in
(d) is the outline of the underling Co particle. Reprint from Beard, K.D.;
Borrelli, D.; Cramer, A.M.; Blom, D.; Van Zee, J.W.; Monnier, J.R.
ACS Nano. 2009, 3, 28412853, with permission from the American
Chemical Society.
Metal Nanoparticle Synthesis 97
There are several ways to perform co-DP, including (1)
co-DP by direct addition of a base (co-DP NaOH), (2) co-DP
with urea (co-DPU), (3) co-DP by sonolysis (co-DPS), and (4)
co-DP by ammonia evaporation (co-DPAE).193 Another
advantage of co-DP is that undesired ions and compounds
from metal precursors will not be introduced in the catalyst.
However, co-DP also presents some drawbacks. For instance,
in the co-DP NaOH method when the PZC of the support is
below 5, such as for SiO2 or oxidized carbon, the deposition of
gold nanoparticles is not effective.194,195
Co-DP has been widely applied for the development of sup-
ported bimetallic catalysts especially to gold bimetallic nanopar-
ticles for different reactions, such as CO oxidation,187,190
selective hydrogenation,191 water gas shift,196 the epoxidation
of propylene,197 methanol oxidation186,188,189 reactions involv-
ing halogens, and the oxidation of volatile organic compounds.
These references are summarized in Table 4, from which the
small particle sizes resulting from co-DP and the efficacy of
catalytic reactivity are readily apparent.
7.04.4 Conclusion
The nanoparticle synthesis methods reviewed in this chapter
can be compared against each other as shown in Table 5. For
single metals, each method can be rated for the simplicity of
the procedure and its scalability to large quantities (both
related to expense), reproducibility from both a macroscopic
level (composition) and microscopic level (tightness of nano-
particle size distribution), and efficacy of metal utilization or
dispersion (higher for smaller particles), as well as the ability to
achieve high metal loadings. The 3D, encapsulated-
then-deposited nanoparticle synthesis methods reviewed in
the first section of the chapter have been admittedly oversim-
plified into one category and compared to the 2D, surface
deposition methods of the second section.
Comparing the methods for each criterion (row by row),
the very simplest method to synthesize metal nanoparticles is
the essentially one-step method of dry impregnation, while the
most complex are the solution methods which require multi-
ple steps in solution followed by painstaking removal of the
scaffolds (although deposition of bare metal nanoparticles
from aqueous solutions is an exception to this). Wet impreg-
nation and SEA are a bit more complicated in that they involve
a filtering step after the adsorption step, and the pH control
necessitated by DP makes that method a bit more complicated
than SEA.
The methods of dry impregnation, wet impregnation, SEA,
and CEDI are as applicable to large quantities of catalyst sup-
port as they are to small quantities. Scalability is likely to be
more complex in the solution synthesis of encapsulated metal
particles in large batches and also in controlling pH for DP at
large scale. All methods in theory are reproducible, though
lower marks are given for dry impregnation and wet impreg-
nation as drying conditions can exhibit deleterious effects on
metal particle size and distribution in systems in which there is
no precursorsupport interaction. Microscopic reproducibility
is a measure of the tightness of particle size distributions, and
there is wide variation of rankings for this. 3D methods have
perhaps the tightest particle size distributions, at least before
98 Metal Nanoparticle Synthesis

Table 4 Examples of oxide-supported bimetallics prepared by co-DP

System Basic agent Support Particle size (nm) Reaction Catalytic activity References

AuCu, AuCo, AuRu Urea TiO2 3.8 a-pinene oxidation AuCu > AuCo > Cu > Au > AuRu 185

AuRu NH4OH Fe2O3 2.5 Methanol oxidation AuRu > (Au or Ru) 186

AuIr Urea TiO2 2.4 CO oxidation AuIr > (Au or Ir) 187

AuCu NaOH TiO2 2.4 Methanol oxidation AuCu > (Au or Cu) 188

CuNi Urea ZrO2 10 Methanol oxidation CuNi > (Cu or Ni) 189

AuPd NH4OH Al2O3 4 CO oxidation AuPd > (Au or Pd) 190

CoNi, NiCu, CuCo NH4OH SiO2 Furfuraldehydehydrogenation CuCo, NiCo > NiC 191

PdCoAu NaOH VXC72 5 O2 reduction PdCOAu > PdCo 192

Table 5 A comparison of methods for nanoparticle synthesis on high-surface-area supports

Single metals

3D Dry impregnation Wet impregnation Strong electrostatic Charge-enhanced dry Depositionprecipitation

(DI) (WI) adsorption (SEA) impregnation (CEDI) (DP)

Simplicity  
Scalability
Reproducibility
Macroscopic
Nanoparticle  
Efficacy
Size  
Loading 

Multimetals
3D DI WI co-SEA co-CEDI seq-SEA ED co-DP
Simplicity    
Scalability 
Reproducibility
Macroscopic
Nanoparticle  
Efficacy
Size  
Loading 
M1M2 contact  

the removal of the capping layers. Though not discussed in
detail in this article, 3D methods also afford the ability to
synthesize shape-selected nanoparticles featuring particular
surface facets such as cubes with (100) faces and tetrahedra
with (111) facets. Narrow particle size distributions of sup-
ported nanoparticles can also be achieved by SEA and DP.
Dry and wet impregnation methods frequently give poorly
dispersed particles with wide particle size distributions, includ-
ing large agglomerates of metal. Efficacy in terms of size refers
to metal utilization, or metal dispersion, with the smallest sizes
giving highest dispersion. For metal utilization over high-
surface-area supports, there is perhaps no more effective
method than SEA. 3D methods are also potentially very effec-
tive, and for high metal loadings, DP and CEDI are effective for
synthesizing relatively small nanoparticles. Finally, the best
methods simply in terms of depositing high metal loadings
are 3D, DI, CEDI, and DP.
For bi- or multimetallic catalysts, the simplicity, scalability,
and reproducibility of 3D, DI, WI, co-SEA, co-CEDI, and co-DP
are ranked the same as their single-metal versions. Co-SEA
leads to very highly dispersed and tightly size-distributed bime-
tallic particles at moderate loadings, while co-DP and co-CEDI
can yield reasonably small and tightly sized alloyed particles at
higher metal loadings.
Sequential SEA is more complicated than co-SEA and is
roughly as complicated as ED and DP. ED is eminently scalable,
while seq-SEA involves multistep filtration and reduction steps.
Given the precise metal1metal2 interaction during synthesis,
both seq-SEA and ED are very reproducible both from macro-
scopic (catalyst composition and metal distribution) and micro-
scopic (particle size distribution) viewpoints. Sequential SEA
and ED also give multimetallic particles of the smallest sizes
and with the highest degrees of intermetallic contact. SEA does
have limited core metal loading relative to most of the other
methods.
Summarizing the table on a column-by-column (method)
basis, 3D methods are complex to employ but have a potentially
high payoff with very well-dispersed, tightly size-distributed, and

even shape-controlled nanoparticles, with precise compositions
in the case of multimetallics, provided the scaffold can be
removed without sintering the particles. This is a key issue in
this field at present. Dry impregnation is a cheap, simple method
that very often gives poor and uneven metal dispersion and poor
M1M2 interaction in multimetallics as no provision is made to
achieve precursorsupport or metal1metal2 interaction. Wet
impregnation possesses no advantages over DI but has the addi-
tional disadvantages of requiring a filtration step and being lim-
ited to low metal loadings. SEA is more complex than DI, but the
payoff of an additional filtration step is the production of ultra-
small metal particles with perhaps the tightest size distributions
that can be achieved with 2D chemistry. Co-SEA for bimetallics is
as simple as single-metal SEA and, from the few results obtained to
date, provides well-dispersed, homogeneously alloyed multime-
tallic nanoparticles. Sequential SEA has more complexity yet and
is limited by the amount of shell metal that can be adsorbed in a
single sequence but may be the simplest synthesis of bimetallic
catalysts when high ratios of core to shell metal are desired. CEDI
is a hybrid method that combines the simplicity of dry impregna-
tion with electrostatic interactions achieved in SEA. The results for
CEDI obtained to date suggest that particle size is somewhat larger
than can be achieved with SEA conducted in excess liquid. While
ED requires the development of metastable metal precursor
reducing agent baths, the payoff for this complexity is the high
degree of control of shell metal loading onto the core metal
nanoparticles. DP is also a higher complexity method but can
give relatively small particles of single and multimetals at relatively
high metal loadings.
The synthesis of precisely controlled nanoparticle size, com-
position, and intermetallic contact in nanoparticles on high-
surface-area supports can be approached from either careful
removal of the scaffolds of nanoparticles derived from 3D chem-
istry or by 2D methods of SEA or deposition coupled with
ED. For related chapters in this Comprehensive, we refer to
Chapters 5.10 and 7.01.
Acknowledgments
The authors wish to acknowledge the efforts of their students in
the preparation of this chapter: Jin Luo, Shiyao Shan, and Jun
Yin, of the Department of Chemistry, State University of
New York at Binghamton, and Hye-Ran Cho, John Tengco,
Kerry O-Connell, Jadid Samad, and Shuo Cao of the Depart-
ment of Chemical Engineering, University of South Carolina.
CJZ thanks his collaborators for their contributions to different
parts of the work described here, and the financial support
from NSF (CBET-0709113, CHE-0848701, and CMMI
1100736) and DOE-BES Grants (DE-SC0006877).
References
1. Schwarz, J. A.; Contescu, C.; Contescu, A. Chem. Rev. 1995, 95, 477510.
2. Abbet, S.; Judai, K.; Klinger, L.; Heiz, U. Pure Appl. Chem. 2002, 74, 15271535.
3. Meyer, R.; Lei, Y.; Lee, S.; Vajda, S. In Model Systems in Catalysis; Rioux, R., Ed.;
Springer: New York, 2010; pp 345365.
4. Lu, J.; Lei, Y.; Elam, J. W. In Noble Metals; Su, Y.-H., Ed. InTech: Rijeka, 2012,
ISBN: 978-953-307-898-4.
Metal Nanoparticle Synthesis 99
5. Sherry, H. S. In Handbook of Zeolite Science and Technology; Auerbach, A.,
Carrado, K. A., Dutta, P. K., Eds.; Dekker: New York, 2003.
6. Price, G. L. In Catalyst Preparation: Science and Engineering; Regalbuto, J. R.,
Ed.; CRC Press: Florida, 2006.
7. Hermans, S. In Synthesis of Solid Catalysts; de Jong, K. P., Ed.; Wiley-VCH:
Weinheim, 2009; pp 153172.
8. Gates, B. C.; Guczi, L. In Metal Clusters in Catalysis; Knozinger, H., Ed.; Elsevier:
Amsterdam, 1986.
9. Gates, B. C. In Catalyst Preparation: Science and Engineering; Regalbuto, J. R.,
Ed.; CRC Press: Florida, 2006.
10. Ertl, G.; Knozinger, H.; Weitkamp, J. Preparation of Solid catalysts. Wiley-VCH:
Weinheim, 1999.
11. Regalbuto, J. R. Catalyst Preparation: Science and Engineering. CRC Press:
Florida, 2007.
12. de Jong, K. P. Synthesis of Solid Catalysts. Wiley-VCH: Weinheim, 2009.
13. Shiju, N. R.; Guliants, V. Appl. Catal. Gen. 2009, 356, 117.
14. Scholten, J. D.; Leal, B. C.; Dupont, J. ACS Catal. 2012, 2, 184200.
15. Regalbuto, J. R. In Surface and Nanomolecular Catalysis; Richards, R., Ed.; CRC
Press: Florida, 2006; pp 161193.
16. Regalbuto, J. R. In Catalyst Preparation: Science and Engineering;
Regalbuto, J. R., Ed.; Taylor & Francis: Boca Raton, 2007; pp 297318.
17. Regalbuto, J. R. In Synthesis of Solid Catalysts; de Jong, K. P., Ed.; Wiley-VCH:
Weinheim, 2009; pp 3358.
18. Regalbuto, J. R. In Silica and Silicates in Modern Catalysis; Halasz, I., Ed.;
Transworld Research Network: Singapore, 2010; pp 274345.
19. Aiken, J. D., III; Finke, R. G. J. Mol. Catal. A. 1999, 145, 144.
20. Klabunde, K. J.; Mulukutla, R. S. In Nanoscale Materials in Chemistry;
Klabunde, K. J., Ed.; John Wiley & Sons, Inc: New York, 2001; p 223.
21. Li, H.; Luk, Y.-Y.; Mrksich, M. Langmuir 1999, 15, 4957.
22. Ingram, R. S.; Murray, R. W. Langmuir 1998, 14, 4115.
23. Zhong, C. J.; Maye, M. M. Adv. Mater. 2001, 13, 1507.
24. (a) Qian, H. F.; Jiang, D. E.; Li, G.; Gayathri, C.; Das, A.; Gil, R. R.; Jin, R. C. J. Am.
Chem. Soc. 2012, 134, 1615916162; (b) Qian, H. F.; Zhu, M. Z.; Wu, Z. K.;
Jin, R. C. Acc. Chem. Res. 2012, 45, 14701479.
25. Philip, R.; Chantharasupawong, P.; Qian, H. F.; Jin, R. C.; Thomas, J. Nano Lett.
2012, 12, 46614667.
26. Cliffel, D. E.; Zamborini, F. P.; Gross, S. M.; Murray, R. W. Langmuir 2000, 16,
96999702.
27. Hunter, R. J. Foundations of Colloid Science, Vol. 1, Oxford University Press:
New York, 1987.
28. Napper, D. H. Polymeric Stabilization of Colloidal Dispersions. Academic Press:
London, 1983.
29. Bonnemann, H.; Braun, G.; Brijoux, W.; Brinkmann, R.; Schulze Tilling, A.;
Seevogel, K.; Siepen, K. J. Organomet. Chem. 1996, 520, 143162.
30. Astruc, D.; Lu, F.; Aranzaes, J. R. Angew. Chem. Int. Ed. 2005, 44, 78527872.
31. (a) Zhao, M.; Sun, L.; Crooks, R. M. J. Am. Chem. Soc. 1998, 120, 4877;
(b) Chechik, V.; Zhao, M.; Crooks, R. M. J. Am. Chem. Soc. 2000, 122, 1243;
(c) Yeung, L. K.; Crooks, R. M. Nano Lett. 2001, 1, 14; (d) Scott, R. W.;
Wilson, O. M.; Oh, S.-K.; Kenik, E. A.; Crooks, R. M. J. Am. Chem. Soc. 2004,
126, 15583; (e) Kim, Y.-G.; Ho, S.-K.; Crooks, R. M. Chem. Mater. 2004, 16,
167; (f) Scott, R. W. J.; Sivadinarayana, C.; Wilson, O. M.; Yan, Z.;
Goodman, D. W.; Crooks, R. M. J. Am. Chem. Soc. 2005, 127, 1380;
(g) Liu, J.-H.; Wang, A.-Q.; Shi, Y.-S.; Lin, H.-P.; Mou, C.-Y. J. Phys. Chem. B
2005, 109, 40.
32. (a) Job, N.; Lambert, S.; Chatenet, M.; Gommes, C. J.; Maillard, F.;
Berthon-Fabry, S.; Regalbuto, J. R.; Pirard, J.-P. Catal. Today. 2010, 150,
119127; (b) Lambert, S.; Job, N.; DSouza, L.; Pereira, M. F. R.; Pirard, R.;
Heinrichs, B.; Figueiredo, J. L.; Pirard, J.-P.; Regalbuto, J. R. J. Catal. 2009, 261,
2333; (c) Jiao, L.; Regalbuto, J. R. J. Catal. 2008, 260, 329341;
(d) DSouza, L.; Regalbuto, J. R.; Miller, J. T. J. Catal. 2008, 254, 157169.
33. Bard, A. J.; Faulkner, L. R. Electrochemical Methods. Wiley: New York, 1980.
34. Aiken, J. D., III; Lin, Y.; Finke, R. G. J. Mol. Catal. A: Chem. 1996, 114, 2951.
35. Lin, Y.; Finke, R. G. J. Am. Chem. Soc. 1994, 116, 83358353.
36. Reetz, M. T.; Lohmer, G. J. Chem. Soc., Chem. Commun. 1996, 19211922.
37. van Benthem, R. A. T. M.; Heimstra, H.; van Leeuwen, P. W. N. M.; Geus, J. W.;
Speckamp, W. N. Angew. Chem. Int. Ed Engl. 1995, 34, 547.
38. Liu, J.; Qin, G.; Raveendran, P.; Ikushima, Y. Chem. Eur. J. 2006, 12,
21312138.
39. Berhault, G.; Bisson, L.; Thomazeau, C.; Verdon, C.; Uzio, D. Appl. Catal. Gen.
2007, 327, 3243.
40. Wang, C.; Daimon, H.; Onodera, T.; Koda, T.; Sun, S. Angew. Chem. Int. Ed.
2008, 47, 35883591.
41. Kang, Y.; Murray, C. B. J. Am. Chem. Soc. 2010, 132, 7568.
100 Metal Nanoparticle Synthesis
42. Kang, Y. J.; Pyo, J. B.; Ye, X. C.; Gordon, T. R.; Murray, C. B. ACS Nano 2012, 6,
56425647.
43. Wang, C.; van der Vilet, D.; Chang, K. C.; You, H.; Strmcnik, D.; Schlueter, J. A.;
Markovic, N. M.; Stamenkovic, V. R. J. Phys. Chem. C 2009, 113, 19365.
44. Qian, Y.; Wen, W.; Adcock, P. A.; Jiang, Z.; Hakim, N.; Saha, M. S.; Mukerjee, S.
J. Phys. Chem. C 2008, 112, 11461157.
45. Chen, W.; Kim, J.; Sun, S.; Chen, S. J. Phys. Chem. C 2008, 112,
38913898.
46. Venkatesan, P.; Santhanalakshmi, J. Langmuir 2010, 26, 1222512229.
47. Zhu, C.; Guo, S.; Dong, S. J. Mater. Chem. 2012, 22, 1485114855.
48. (a) Maye, M. M.; Lou, Y.; Zhong, C. J. Langmuir 2000, 16, 7520; (b) Luo, J.;
Lou, Y.; Maye, M. M.; Zhong, C. J.; Hepel, M. Electrochem. Commun. 2001, 3,
172; (c) Lou, Y.; Maye, M. M.; Han, L.; Luo, J.; Zhong, C. J. Chem. Commun.
2001, 5, 473; (d) Zhong, C. J.; Maye, M. M. Adv. Mater. 2001, 13, 15071511;
(e) Luo, J.; Maye, M. M.; Lou, Y.; Han, L.; Hepel, M.; Zhong, C. J. Catalysis Today
2002, 77, 127138; (f) Luo, J.; Jones, V. W.; Maye, M. M.; Han, L.;
Kariuki, N. N.; Zhong, C. J. J. Am. Chem. Soc. 2002, 124, 1398813989;
(g) Luo, J.; Maye, M. M.; Petkov, V.; Kariuki, N. N.; Wang, L.; Njoki, P.; Mott, D.;
Lin, Y.; Zhong, C. J. Chem. Mater. 2005, 17, 30863091; (h) Luo, J.; Njoki, P.;
Lin, Y.; Mott, D.; Wang, L.; Zhong, C. J. Langmuir 2006, 22, 28922898;
(i) Luo, J.; Njoki, P.; Lin, Y.; Wang, L.; Mott, D.; Zhong, C. J. Electrochem. Comm.
2006, 8, 581587; (j) Wanjala, B. N.; Luo, J.; Loukrakpam, R.; Mott, D.;
Njoki, P. N.; Fang, B.; Engelhard, M.; Naslund, H. R.; Wu, J. K.; Wang, L.;
Malis, O.; Zhong, C. J. Chem. Mater. 2010, 22, 42824294; (k) Wanjala, Bridgid
N.; Luo, Jin; Fang, Bin; Mott, Derrick; Zhong, C. J. J. Mater. Chem. 2011, 21,
40124020; (l) Maye, M. M.; Kariuki, N. N.; Luo, J.; Han, L.; Njoki, P.; Wang, L.;
Lin, Y.; Naslunda, H. R.; Zhong, C. J. Gold Bull. 2004, 37, 217; (m) Hepel, M.;
Dela, I.; Hepel, T.; Luo, J.; Zhong, C. J. Electrochim. Acta 2007, 52, 55295547.
49. (a) Luo, J.; Han, L.; Kariuki, N.; Wang, L.; Mott, D.; Zhong, C. J.; He, T. Chem.
Mater. 2005, 17, 52825290; (b) Luo, J.; Kariuki, N.; Han, L.; Wang, L.;
Zhong, C. J.; He, T. Electrochim. Acta 2006, 51, 48214827; (c) Luo, J.;
Wang, L.; Mott, D.; Njoki, P.; Kariuki, N.; Zhong, C. J.; He, T. J. Mater. Chem.
2006, 16, 16651673; (d) Wanjala, B.; Fang, B.; Luo, J.; Chen, Y.; Yin, J.;
Engelhard, M.; Loukrakpam, R.; Zhong, C. J. J. Am. Chem. Soc. 2011, 133,
1271412727; (e) Wanjala, B. N.; Fang, B.; Petkov, V.; Zhu, P.; Loukrakpam, R.;
Chen, Y.; Luo, J.; Yin, J.; Yang, L.; Shan, S.; Shao, M.; Zhong, C. J. Chem. Mater.
2012, 24, 42834293.
50. (a) Wanjala, B.; Loukrakpam, R.; Luo, J.; Njoki, P. N.; Mott, D.; Shao, M.;
Protsailo, L.; Kawamura, T.; Zhong, C. J. J. Phys. Chem. C 2010, 114,
1758017590; (b) Loukrakpam, R.; Luo, J.; He, T.; Chen, Y.; Xu, Z.; Njoki, P. N.;
Wanjala, B. N.; Fang, B.; Mott, D.; Yin, J.; Klar, J.; Powell, B.; Zhong, C. J. J. Phys.
Chem. C 2011, 115, 16821694; (c) Loukrakpam, R.; Wanjala, B. N.; Yin, J.;
Fang, B.; Luo, J.; Shao, M.; Protsailo, L.; Kawamura, T.; Chen, Y.; Petkov, V.;
Zhong, C. J. ACS Catalysis 2011, 1, 562572; (d) Wanjala, B.; Fang, B.;
Loukrakpam, R.; Chen, Y.; Engelhard, M.; Luo, J.; Yin, J.; Yang, L.; Shan, S.;
Zhong, C. J. ACS Catalysis 2012, 2, 795806.
51. Gross, E.; Krier, J. M.; Heinke, L.; Somorjai, G. A. Top. Catal. 2012, 22, 111.
52. Lang, H.; May, R. A.; Iversen, B. L.; Chandler, B. D. J. Am. Chem. Soc. 2003, 125,
1483214836.
53. Niu, Y.; Crooks, R. M. Chem. Mater. 2003, 15, 3463.
54. Crooks, R. M.; Zhao, M.; Sun, L.; Chechik, V.; Yeung, L. K. Acc. Chem. Res.
2001, 34, 181.
55. Lang, H.; Maldonado, S.; Stevenson, K. J.; Chandler, B. D. J. Am. Chem. Soc.
2004, 126, 1294912956.
56. Endo, T.; Yoshimura, T.; Esumi, K. J. Colloid Interface Sci. 2005, 286,
602609.
57. Scott, R. W. J.; Datye, A. K.; Crooks, R. M. J. Am. Chem. Soc. 2003, 125,
37083709.
58. Crooks, R. M. J. Am. Chem. Soc. 2005, 127, 10151024.
59. Rioux, R. M.; Song, H.; Hoefelmeyer, J. D.; Yang, P.; Somorjai, G. A. J. Phys.
Chem. B 2005, 109, 21922202.
60. (a) Aliaga, C.; Park, J. Y.; Yamada, Y.; Lee, H. S.; Tsung, C. K.; Yang, P. D.;
Somorjai, G. A. J. Phys. Chem. C 2009, 113, 61506155; (b) Joo, S. H.;
Park, J. Y.; Tsung, C. K.; Yamada, Y.; Yang, P. D.; Somorjai, G. A. Nat. Mater.
2009, 8, 126131.
61. Lee, H.; Habas, S. E.; Kweskin, S.; Butcher, D.; Somorjai, G. A.; Yang, P. D.
Angew. Chem. Int. Ed. 2006, 45, 78247828.
62. (a) Narayanan, R.; El-Sayed, M. A. J. Am. Chem. Soc. 2003, 125, 8340; (b) Li, Y.;
Boone, E.; El-Sayed, M. A. Langmuir 2002, 18, 4921.
63. Mu, X.-D.; Meng, J.-Q.; Li, Z.-C.; Kou, Y. J. Am. Chem. Soc. 2005, 127,
96949695.
64. Park, J. Y.; Zhang, Y.; Grass, M.; Zhang, T.; Somorjai, G. A. Nano Lett. 2008, 8,
673677.
65. (a) He, J.-H.; Ichinose, I.; Kunitake, T.; Nakao, A.; Shiraishi, Y.; Toshima, N. J. Am.
Chem. Soc. 2003, 125, 11034; (b) Shiraishi, Y.; Ikenaga, D.; Toshima, N. Aust. J.
Chem. 2003, 56, 1025; (c) Sablong, R.; Schlotterbeck, U.; Vogt, D.; Mecking, S.
Adv. Synth. Catal. 2003, 345, 333.
66. Zhang, J.; Sasaki, K.; Sutter, E.; Adzic, R. R. Science 2007, 315, 220.
67. Adzic, R. R.; Zhang, J.; Sasaki, K.; Vukmirovic, M. B.; Shao, M.; Wang, J. X.;
Nilekar, A. U.; Mavrikakis, M.; Valerio, J. A.; Uribe, F. Topics in Catalysis 2007,
46, 249262.
68. Lim, B.; Jiang, M.; Camargo, P. H. C.; Cho, E. C.; Tao, J.; Lu, X. M.; Zhu, Y. M.;
Xia, Y. N. Science 2009, 324, 1302.
69. Wang, L.; Yamauchi, Y. Chem. Mater. 2011, 23, 24572465.
70. Templeton, A. C.; Wuelfing, W. P.; Murray, R. W. Acc. Chem. Res. 2000, 33, 27
and references therein.
71. Caruso, F. Adv. Mater. 2001, 13, 11 and references therein.
72. (a) Sun, S.; Murray, C. B.; Weller, D.; Folks, L.; Moser, A. Science 2000, 287,
1989; (b) Sun, S. H.; Fullerton, E. E.; Weller, D.; Murray, C. B. IEEE Trans. Magn.
2001, 37, 1239.
73. Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. J. Chem. Soc.
Chem. Commun. 1994, 801.
74. (a) Hostetler, M. J.; Zhong, C. J.; Yen, B. K. H.; Anderegg, J.; Gross, S. M.;
Evans, N. D.; Porter, M. D.; Murray, R. W. J. Am. Chem. Soc. 1998, 120, 9396;
(b) Hostetler, M. J.; Wingate, J. E.; Zhong, C. J.; Harris, J. E.; Vachet, R. W.;
Clark, M. R.; Londono, J. D.; Green, S. J.; Stokes, J. J.; Wignall, G. D.;
Glish, G. L.; Porter, M. D.; Evans, N. D.; Murray, R. W. Langmuir 1998, 14, 17.
75. Hostetler, M. J.; Zhong, C. J.; Yen, B. K. H.; Anderegg, J.; Gross, S. M.;
Evans, N. D.; Porter, M. D.; Murray, R. W. J. Am. Chem. Soc. 1998, 120, 9396.
76. Hostetler, M. J.; Templeton, A. C.; Murray, R. W. Langmuir 1999, 15, 3782.
77. Luo, J.; Wang, L. Y.; Mott, D.; Njoki, P.; Lin, Y.; He, T.; Xu, Z.; Wanjana, B.;
I-Im Lim, S.; Zhong, C. J. Adv. Mater. 2008, 20, 43424347.
78. Kariuki, N. N.; Luo, J.; Maye, M. M.; Hassan, S. A.; Menard, T.; Naslund, H. R.; Lin, Y.;
Wang, C.; Engelhard, M. H.; Zhong, C. J. Langmuir 2004, 20, 1124011246.
79. Nishimura, S.; Anh, T. N. D.; Mott, D.; Ebitani, K.; Maenosono, S. J. Phys. Chem.
C 2012, 116, 45114516.
80. (a) Mott, D.; Galkowski, J.; Wang, L.; Luo, J.; Zhong, C. J. Langmuir 2007, 23,
57405745; (b) Mott, D.; Yin, J.; Engelhard, M.; Loukrakpam, R.; Chang, P.;
Miller, G.; Bae, I.-T.; Chandra Das, N.; Wang, C.; Luo, J.; Zhong, C. J. Chem.
Mater. 2010, 22, 261271.
81. (a) Buffat, P. H.; Borel, J. P. Phys. Rev. A 1976, 13, 2287; (b) Borel, J. P. Surf.
Sci. 1981, 106, 1; (c) Lewis, L. L.; Pablo, J.; Barrat, J.-L. Phys. Rev. B. 1997, 56,
2248.
82. (a) Zhong, C. J.; Zheng, W. X.; Leibowitz, F. L.; Eichelberger, H. H. Chem.
Commun. 1999, 13, 1211; (b) Maye, M. M.; Zheng, W.; Leibowitz, F. L.;
Ly, N. K.; Zhong, C. J. Langmuir 2000, 16, 490; (c) Schadt, M. J.; Cheung, W.;
Luo, J.; Zhong, C. J. Chem. Mater. 2006, 18, 5147; (d) Maye, M. M.; Zhong, C. J.
J. Mater. Chem. 2000, 10, 1895; (e) Luo, J.; Jones, V. W.; Maye, M. M.; Han, L.;
Kariuki, N. N.; Zhong, C. J. J. Am. Chem. Soc. 2002, 124, 13988; (f) Luo, J.;
Jones, V. W.; Han, L.; Maye, M. M.; Kariuki, N.; Zhong, C. J. J. Phys. Chem.
2004, 108, 9669.
83. Shields, S. P.; Richards, V. N.; Buhro, W. E. Chem. Mater. 2010, 22, 32123225.
84. (a) Park, H.-Y.; Schadt, M. J.; Wang, L. Y.; Lim, S.; Njoki, P. N.; Kim, S. H.;
Jang, M.-Y.; Luo, J.; Zhong, C. J. Langmuir 2007, 23, 9050; (b) Wang, L. Y.;
Park, H.-Y.; I-Im Lim, S.; Schadt, M. J.; Mott, D.; Luo, J.; Wang, X.; Zhong, C. J.
J. Mater. Chem. 2008, 18, 26292635; (c) Wang, L. Y.; Miller, D.; Fan, Q.;
Luo, J.; Schadt, M.; Rendeng, Q.; Wang, G. R.; Wang, J.; Kowach, G. R.;
Zhong, C. J. J. Phys. Chem. C 2008, 112, 24482455.
85. (a) Yin, J.; Hu, P.; Wanjala, B.; Malis, O.; Zhong, C. J. Chem Comm 2011, 47,
98859887; (b) Yin, J.; Shan, S.; Yang, L.; Mott, D.; Ng, M. S.; Luo, J.; Malis, O.;
Petkov, V.; Engelhard, M.; Zhong, C. J. Chem. Mater. 2012. ASAP.
86. Klabunde, K. J. Nanoscale Materials in Chemistry. Wiley: New York, 2001.
87. (a) Paulus, U. A.; Wokaun, A.; Scherer, G. G.; Schmidt, T. J.; Stamenkovic, V.;
Radmilovic, V.; Markovic, N. M.; Ross, P. N. J. Phys. Chem. B 2002, 106, 4181;
(b) Antolini, E. Mater. Chem. Phys. 2003, 78, 563; (c) Rao, C. R. K.; Trivedi, D. C.
Coord. Chem. Rev. 2005, 249, 613; (d) Yang, H.; Vogel, W.; Lamy, C.; Alonso-
Vante, N. J. Phys. Chem. B 2004, 108, 11024; (e) Waszczuk, P.; Lu, G. Q.;
Wieckowski, A.; Lu, C.; Rice, C.; Masel, R. I. Electrochim. Acta 2002, 47, 3637;
(f) Park, K. W.; Choi, J. H.; Kwon, B. K.; Lee, S. A.; Sung, Y. E.; Ha, H. Y.;
Hong, S. A.; Kim, H.; Wieckowski, A. J. Phys. Chem. B 2002, 106, 1869;
(g) Schmidt, T. J.; Gasteiger, H. A.; Behm, R. J. Electrochem. Commun. 1999,
1, 1; (h) Raja, R.; Khimyak, T.; Thomas, J. M.; Hermans, S.; Johnson, B. F. G.
Angew. Chem. Int. Edn 2001, 40, 4638; (i) Jones, F. E.; Milne, S. B.; Gurau, B.;
Smotkin, E. S.; Stock, S. R.; Lukehart, C. M. J. Nanosci. Nanotechnol. 2002,
2, 81; (j) Dickinson, A. J.; Carrette, L. P. L.; Collins, J. A.; Friedrich, K. A.;
Stimming, U. Electrochim. Acta 2002, 47, 3733.

88. Yoo, J. W.; Hathcock, D.; El-Sayed, M. A. J. Phys. Chem. A 2002, 106, 2049.
89. Galow, T. H.; Drechsler, U.; Hanson, J. A.; Rotello, V. M. Chem. Commun. 2002,
1076.
90. (a) Wang, C.; Daimon, H.; Lee, Y.; Kim, J.; Sun, S. J. Am. Chem. Soc. 2007, 129,
6974; (b) Chen, J.; Lim, B.; Lee, E. P.; Xia, Y. Nano Today 2009, 4, 81; (c) Yuan, Q.;
Zhou, Z.; Zhuang, J.; Wang, X. Chem. Commun. 2010, 46, 1491; (d) Xia, B. Y.;
Wang, J. N.; Wang, X. X. J. Phys. Chem. C 2009, 113, 18115; (e) Zhou, H.;
Zhou, W.; Adzic, R. R.; Wong, S. S. J. Phys. Chem. C 2009, 113, 5460.
91. Wang, L.; Yamauchi, Y. J. Am. Chem. Soc. 2010, 132, 13636.
92. Han, L.; Wu, W.; Kirk, F. L.; Luo, J.; Maye, M. M.; Kariuki, N.; Lin, Y.; Wang, C.;
Zhong, C. J. Langmuir 2004, 20, 6019.
93. Paulus, U. A.; Endruschat, U.; Feldmeyer, G. J.; Schmidt, T. J.; Bonnemann, H.;
Behm, R. J. J. Catal. 2000, 195, 383.
94. Nashner, M. S.; Frenkel, A. I.; Somerville, D.; Hills, C. W.; Shapley, J. R.;
Nuzzo, R. G. J. Am. Chem. Soc. 1998, 120, 8093.
95. Wuelfing, W. P.; Zamborini, F. P.; Templeton, A. C.; Wen, X. G.; Yoon, H.;
Murray, R. W. Chem. Mater. 2001, 13, 87.
96. (a) Nakaso, K.; Shimada, M.; Okuyama, K.; Deppert, K. J. Aerosol Sci.
2002, 33, 1061; (b) Mitchell, C. E. J.; Howard, A.; Carney, M.; Egdell, R. G. Surf.
Sci. 2001, 490, 196; (c) Weber, A. P.; Friedlander, S. K. J. Aerosol Sci. 1997,
28, 179.
97. Luo, J.; Maye, M. M.; Han, L.; Kariuki, N.; Jones, V. W.; Lin, Y.; Engelhard, M. H.;
Zhong, C. J. Langmuir 2004, 20, 42544260.
98. He, T.; Kreidler, E.; Xiong, L.; Luo, J.; Zhong, C. J. J. Electrochem. Soc. 2006,
153, A16371643.
99. Campbell, C. T.; Parker, S. C.; Starr, D. E. Science 2002, 298, 811.
100. (a) Liu, Z.; Reed, D.; Kwon, G.; Shamsuzzoha, M.; Nikles, D. E. J. Phys. Chem. C
2007, 111, 1422314229; (b) Sun, X. C.; Huang, Y. H.; Nikles, D. E. Int.
J. Nanotechnol. 2004, 1, 328346.
101. Nazarpoor, Z.; Ma, S. G.; Fanson, P. T.; Alexeev, O. S.; Amiridis, M. D. J. Catal.
2012, 290, 2636.
102. Petkov, V.; Wanjala, B. N.; Loukrakpam, R.; Luo, J.; Yang, L.; Zhong, C. J.;
Shastri, S. Nano Lett. 2012, 12, 42894299.
103. Yang, L.; Shan, S.; Loukrakpam, R.; Petkov, V.; Ren, Y.; Wanjala, B.;
Engelhard, M.; Luo, J.; Yin, J.; Chen, Y.; Zhong, C. J. J. Am. Chem. Soc. 2012,
134, 1504815060.
104. Park, J.; Regalbuto, J. R. J. Colloid Interface Sci. 1995, 175, 239252.
105. van den Berg, G. H.; Rijnten, H. T. In Preparation of Catalysts II: Scientific Bases
for the Preparation of Heterogeneous Catalysts; Delmon, B., Jacobs, P. A.,
Poncelet, G., Eds.; Elsevier Scientific Publishing Company: Amsterdam, 1976;
pp 265278.
106. Coulson, J. M.; Richardson, J. F. Chemical Engineering. Pergamon Press: Oxford,
1980.
107. Boon, A.Q.M. Thesis, Utrecht University, 1990.
108. Geus, J. W. In Catalyst Preparation: Science and Engineering, CRC Press: Florida,
2007; pp 341374.
109. Komiyama, M. Catal. Rev.: Sci. Eng. 1985, 27, 341372.
110. Jiao, L.; Regalbuto, J. R. J. Catal. 2008, 260, 342350.
111. Miller, J. T.; Schreier, M.; Kropf, A. J.; Regalbuto, J. R. J. Catal. 2004, 225,
203212.
112. Jiao, L.; Regalbuto, J. R. J. Catal. 2008, 260, 329341.
113. Hao, X.; Barnes, S.; Regalbuto, J. R. J. Catal. 2011, 279, 4865.
114. Kratzer, E., Regalbuto, J.R. Manuscript in Preparation. MS Thesis, University of
Illinois, Chicago, 2012.
115. DSouza, L.; Schreier, M.; Regalbuto, J.R. Manuscript in Preparation.
116. Zhu, X.; CHo, H. R.; Pasupong, M.; Regalbuto, J. R. ACS Catal. in press
117. Stowenert, F.; Farben, I. German Patent, 740,634, 1943.
118. Hermans, L. A. M.; Geus, J. W. In Preparation of Catalysts II; Delmon, B.,
Jacobs, P. A., Poncelet, G., Eds.; Elsevier: Amsterdam, 1979; p 113.
119. Hermans, L. A. M.; Geus, J. W. In Studies in Surface Science and Catalysis;
Delmon, B., Grange, P., Jacobs, P., Poncelet, G., Eds.; Elsevier: Amsterdam,
1979; pp 113130.
120. vanDillen, J.A.; Geus, J.W.; Hermans, L.A.; van der Meijden, J. In Proceedings
of the 6th International Congress on Catalysis; Wells, P.B.; Tompkins, F.C.;
Bond, G.C., Eds.; London, 1976.
121. Burattin, P.; Che, M.; Louis, C. J. Phys. Chem. B 1997, 101, 70607074.
122. Zanella, R.; Louis, C.; Giorgio, S.; Touroude, R. J. Catal. 2004, 223,
328339.
123. Bond, G. C.; Louis, C.; Thompson, D. T. In Catalysis by Gold, Imperial College
Press: London, 2006.
124. Nguyen, D. L.; Umbarkar, S.; Dongare, M. K.; Lancelot, C.; Girardon, J. S.;
Dujardin, C.; Granger, P. Catal. Commun. 2012, 26, 225230.
125. de Jong, K. P. Curr. Opin. Solid State Mater. Sci. 1999, 4, 5562.
Metal Nanoparticle Synthesis 101
126. Bamwenda, G. R.; Tsubota, S.; Nakamura, T.; Haruta, M. Catal. Lett. 1997, 44,
8387.
127. Zanella, R.; Giorgio, S.; Henry, C. R.; Louis, C. J. Phys. Chem. B 2002, 106,
76347642.
128. Montes, M.; Soupart, J.; Saedeleer, M. J. Chem. Soc., Faraday Trans. 1984, 80,
3209.
129. Chytil, S.; Glomm, W. R.; Kvande, I.; Zhao, T.; Walmsley, J. C.; Blekkan, E. A.
Top. Catal. 2007, 45, 9399.
130. Chytil, S.; Glomm, W. R.; Blekkan, E. A. Catal. Today 2009, 147, 217223.
131. Qian, K.; Fang, J.; Huang, W. X.; He, B.; Jiang, Z. Q.; Ma, Y. S.; Wei, S. Q.
J. Mol. Catal. A: Chem. 2010, 320, 97105.
132. Bezemer, G. L.; Radstake, P. B.; Koot, V.; van Dillen, A. J.; Geus, J. W.;
de Jong, K. P. J. Catal. 2006, 237, 291302.
133. van der Lee, M. K.; van Dillen, A. J.; Bitter, J. H.; de Jong, K. P. J. Am. Chem. Soc.
2005, 127, 1357313582.
134. Haruta, M. Gold Bull. 2004, 37, 2736.
135. Wolf, A.; Schuth, F. Appl. Catal. Gen. 2002, 226, 113.
136. Dal Santo, V.; Liguori, F.; Pirovano, C.; Guidotti, M. Molecules 2010, 15,
38293856.
137. Fraile, J. M.; Garca, J. I.; Mayoral, J. A.; Pires, E. In Heterogenized Homogeneous
Catalysts for Fine Chemicals Production; Bianchini, D. J. C.-H. C., van
Leeuwen, P. W. N. M., Eds.; Springer: New York, 2010; pp 65123.
138. Louis C, C. M. In Preparation of Solid Catalysts; Ertl G, K. H., Weitkamp, J., Eds.;
Wiley-VCH Verlag GmbH: Weinheim, 1999; pp 341354.
139. Chen, C. C.; Cheng, S. F.; Jang, L. Y. Microporous Mesoporous Mater. 2008,
109, 258270.
140. Hamilton, N.; Wolfram, T.; Muller, G. T.; Havecker, M.; Krohnert, J.; Carrero, C.;
Schomacker, R.; Trunschke, A.; Schlogl, R. Catal. Sci. Technol. 2012, 2,
13461359.
141. Montini, T.; Rogatis, L. D.; Gombac, V.; Cargnello, M.; Fornasiero, P. In Stabilized
Metal Nanoparticles Embedded into Porous Oxides: A Challenging Approach for
Robust Catalysts; Schwartz, A. I., Prescott, W. V., Eds.; Nova Science Publishers,
Inc: New York, 2008; pp 71123.
142. Sheldon, R. A. Fine Chemicals through Heterogeneous Catalysis. Wiley-VCH
Verlag GmbH: Weinheim, 2001.
143. Gaslain, F.; Lebeau, B.; Fernandez-Martin, C.; Babonneau, F. In Ordered Porous
Solids: Recent Advances and Prospects; Valtchev, V., Mintova, S., Tsapatsis, M.,
Eds.; Elsevier: Oxford, 2009.
144. Jaroniec, M.; Olkhovyk, O. In Environmental Applications of Nanomaterials:
Synthesis, Sorbents and Sensors; Cao, G., Fryxell, G. E., Eds.; Imperial College
Press: London, 2007; pp 185190.
145. Capel-Sanchez, M. C.; Barrio, L.; Campos-Martin, J. M.; Fierro, J. L. G. J. Colloid
Interface Sci. 2004, 277, 146153.
146. del Pozo, C.; Corma, A.; Iglesias, M.; Sanchez, F. Green Chem. 2011, 13,
24712481.
147. Iengo, P.; Di Serio, M.; Sorrentino, A.; Solinas, V.; Santacesaria, E. Appl. Catal.
Gen. 1998, 167, 85101.
148. Mbaraka, I. K.; Shanks, B. H. J. Catal. 2005, 229, 365373.
149. Melero, J. A.; Bautista, L. F.; Morales, G.; Iglesias, J.; Sanchez-Vazquez, R. Chem.
Eng. J. 2010, 161, 323331.
150. Wegener, S. L.; Marks, T. J.; Stair, P. C. Acc. Chem. Res. 2012, 45, 206214.
151. Xiao, F. S. Top. Catal. 2005, 35, 924.
152. Kao, H. M.; Liao, C. H.; Palani, A.; Liao, Y. C. Microporous Mesoporous Mater.
2008, 113, 212223.
153. Wang, X. G.; Tseng, Y. H.; Chan, J. C. C.; Cheng, S. F. Microporous Mesoporous
Mater. 2005, 85, 241251.
154. Iwasawa, Y. Adv. Catal. 1987, 35, 187264.
155. Quek, X. Y.; Liu, D. P.; Cheo, W. N. E.; Wang, H.; Chen, Y.; Yang, Y. H. Appl.
Catal. Environ. 2010, 95, 374382.
156. Han, L.; Choi, H. J.; Choi, S. J.; Liu, B. Y.; Park, D. W. Green Chem. 2011, 13,
10231028.
157. Phillips, J.; Weigle, J.; Herskowitz, M.; Kogan, S. Appl. Catal. Gen. 1998, 173,
273287.
158. D Souza, L., Regalbuto, J.R. Manuscript in Preparation.
159. Cho, H.R., Regalbuto, J.R. Manuscript in Preparation.
160. Barnes, S., Cho, H.R., Regalbuto, J.R., Shaw, G., Hutchings, G. Manuscript in
Preparation.
161. Rebelli, J.; Detwiler, M.; Ma, S.; Williams, C. T.; Monnier, J. R. J. Catal. 2010,
270, 224233.
162. Schaal, M. T.; Pickerell, A. C.; Williams, C. T.; Monnier, J. R. J. Catal. 2008, 254,
131143.
163. Micheaud, C.; Marecot, P.; Guerin, M.; Barbier, J. Appl. Catal. Gen. 1998, 171,
229239.
102 Metal Nanoparticle Synthesis
164. Podlovchenko, B. I.; Zhumaev, U. E.; Maksimov, Y. M. J. Electroanal. Chem.
2011, 651, 3037.
165. Lee, C.-L.; Tseng, C.-M. J. Phys. Chem. C 2008, 112, 1334213345.
166. Lee, C.-L.; Tseng, C.-M.; Wu, R.-B.; Wu, C.-C.; Syu, S.-C. Electrochim. Acta
2009, 54, 55445547.
167. Beard, K. D.; Schaal, M. T.; Van Zee, J. W.; Monnier, J. R. Appl. Catal. Environ.
2007, 72, 262271.
168. Beard, K. D.; Van Zee, J. W.; Monnier, J. R. Appl. Catal. Environ. 2009, 88, 185193.
169. Mustain, W. E.; Hyea, K.; Osborn, T.; Kohl, P. A. Isr. J. Chem. 2008, 48,
251257.
170. Mustain, W. E.; Kim, H.; Narayanan, V.; Osborn, T.; Kohl, P. A. J. Fuel Cell Sci.
Technol. 2010, 7, 041013041017.
171. Ohno, I. In Modern Electroplating John Wiley & Sons, Inc: Hoboken, NJ, 2010
pp 477482.
172. Barbier, J.; Marecot, P.; Del Angel, G.; Bosch, P.; Boitiaux, J. P.; Didillon, B.;
Dominguez, J. M.; Schiftef, I.; Espmosa, G. Appl. Catal. Gen. 1994, 116,
179186.
173. Menezo, J. C.; Hoang, L. C.; Montassier, C.; Barbier, J. React. Kinet. Catal. Lett.
1992, 46, 16.
174. Vasilic, R.; Viyannalage, L. T.; Dimitrov, N. J. Electrochem. Soc. 2006, 153,
C648C655.
175. Okinaka, Y.; Wolowodiuk, C. In Electroless Plating: Fundamentals and
Applications; Mallory, G. O., Hajdu, J. B., Eds.; Noyes Publications/William
Andrew Publishing LLC: New York, 1990; pp 421440.
176. Beard, K. D.; Borrelli, D.; Cramer, A. M.; Blom, D.; Van Zee, J. W.; Monnier, J. R.
ACS Nano 2009, 3, 28412853.
177. Chien-Liang, L.; Chun-Ming, T.; Chen-Chung, W. J. Chin. Inst. Eng. 2010, 33,
105109.
178. Ohashi, M.; Beard, K. D.; Ma, S.; Blom, D. A.; St-Pierre, J.; Van Zee, J. W.;
Monnier, J. R. Electrochim. Acta 2010, 55, 73767384.
179. Ohashi, M.; Monnier, J.R.; Van Zee, J.W. 10th Polymer Electrolyte Fuel Cell
Symposium, PEFC 10218th ECS Meeting, 10 October 201015 October 2010,
Las Vegas, NV, USA.
180. Rebelli, J.; Rodriguez, A. A.; Ma, S.; Williams, C. T.; Monnier, J. R. Catal. Today
2011, 160, 170178.
181. Schaal, M.; Williams, C; Monnier, J.; Pickerell, A.; Hoang, T. 05AIChE: 2005
AIChE Annual Meeting and Fall Showcase, 30 October 20054 November 2005,
Cincinnati, OH, USA.
182. Schaal, M. T.; Metcalf, A. Y.; Montoya, J. H.; Wilkinson, J. P.; Stork, C. C.;
Williams, C. T.; Monnier, J. R. Catal. Today 2007, 123, 142150.
183. Okinaka, Y. In Electroless Plating: Fundamentals and Applications; Mallory, G. O.,
Hadju, J. B., Eds.; Noyes Publications/William Andrew Publishing LLC:
New York, 1990; pp 401420.
184. Valenzuela, K.; Raghavan, S.; Deymier, P. A.; Hoying, J. J. Nanosci. Nanotechnol.
2008, 8, 34163421.
185. Ajaikumar, S.; Ahlkvist, J.; Larsson, W.; Shchukarev, A.; Leino, A. R.; Kordas, K.;
Mikkola, J. P. Appl. Catal. Gen. 2011, 392, 1118.
186. Chang, F.-W.; Roselin, L. S.; Ou, T.-C. Appl. Catal. Gen. 2008, 334, 147155.
187. Gomez-Cortes, A.; Daz, G.; Zanella, R.; Ramrez, H.; Santiago, P.; Saniger, J. M.
J. Phys. Chem. C 2009, 113, 97109720.
188. Ou, T.-C.; Chang, F.-W.; Roselin, L. S. J. Mol. Catal. A: Chem. 2008, 293, 816.
189. Perez-Hernandez, R.; Mondragon Galicia, G.; Mendoza Anaya, D.; Palacios, J.;
Angeles-Chavez, C.; Arenas-Alatorre, J. Int. J. Hydrogen Energy 2008, 33,
45694576.
190. Qian, K.; Huang, W. Catal. Today 2011, 164, 320324.
191. Reddy, B. M.; Reddy, G. K.; Rao, K. N.; Khan, A.; Ganesh, I. J. Mol. Catal. A:
Chem. 2007, 265, 276282.
192. Wei, Y.-C.; Chen, T.-Y.; Liu, C.-W.; Chan, T.-S.; Lee, J.-F.; Lee, C.-H.; Lin, T.-L.;
Wang, K.-W. Catal. Sci. Technol. 2012, 2, 16541664.
193. Wachs, I. E. AIChE J. 2009, 55, 284.
194. Haruta, M. Chem. Rec. 2003, 3, 7587.
195. Wolf, A.; Schuth, F. Appl. Catal. Gen. 2002, 226, 113.
196. Gamboa-Rosales, N. K.; Ayastuy, J. L.; Gonzalez-Marcos, M. P.; Gutierrez-
Ortiz, M. A. Int. J. Hydrogen Energy 2012, 37, 70057016.
197. Liu, Y. W.; Yu, H.; Zhang, X. M.; Suo, J. S. Acta Phys.-Chim. Sin. 2010, 26,
15851592.
 5.03 MetalOrganic Frameworks
S Noro, Hokkaido University, Sapporo, Japan
2013 Elsevier Ltd. All rights reserved.

5.03.1 Introduction 45
5.03.2 Parts of Coordination Frameworks 47
5.03.2.1 Metal Ions 47
5.03.2.2 Secondary Building Units 47
5.03.2.3 Organic Ligands 47
5.03.2.4 Metalloligands 48
5.03.2.5 Inorganic Counteranions 48
5.03.2.6 Guest Molecules 49
5.03.3 Synthetic Techniques 49
5.03.3.1 Solvothermal Methods 49
5.03.3.2 Microwave Syntheses 49
5.03.3.3 Post-Synthetic Modification 49
5.03.3.4 Methods for Controlling Morphologies of PCP Crystals 51
5.03.3.5 Techniques for Obtaining PCP Thin Films 51
5.03.3.6 Other Techniques 52
5.03.4 Representative Topologies of Frameworks 52
5.03.4.1 CPL Structures 52
5.03.4.2 Jungle-Gym-Type Structures 53
5.03.4.3 Zeolitic Imidazolate Frameworks 53
5.03.4.4 Coordination Polymers with an Interdigitated Structure 54
5.03.4.5 PCPs Based on One-Dimensional Frameworks 54
5.03.5 Porous Functionalities 54
5.03.5.1 Storage 54
5.03.5.1.1 Methane 54
5.03.5.1.2 Carbon dioxide 56
5.03.5.1.3 Hydrogen 56
5.03.5.1.4 Water 57
5.03.5.1.5 Acetylene 57
5.03.5.2 Separation 58
5.03.5.2.1 Carbon dioxide 58
5.03.5.2.2 Acetylene 58
5.03.5.2.3 Oxygen 60
5.03.5.3 Catalysis 61
5.03.5.3.1 Lewis acid catalysis 61
5.03.5.3.2 Brnsted acid catalysis 61
5.03.5.3.3 Base catalysis 61
5.03.5.4 Reaction Vessels in Polymer Syntheses 62
5.03.5.5 Framework Dynamics 62
5.03.5.6 Framework Stability 63
5.03.5.7 Combined Functions of Porosities and Magnetic Properties 63
5.03.5.8 Conducting PCPs 64
5.03.5.9 Other Properties 67
5.03.6 Composite PCPs 67
5.03.7 Conclusion 68
References 68

5.03.1 Introduction discovery of novel phenomena in them through their character-

ization and processing; there is also great commercial interest in
Porous compounds have attracted much attention from chem- their application in separation, storage, and heterogeneous
ists, physicists, and materials scientists because of scientific catalysis. Until the mid-1990s, there were two types of porous
interest in the creation of nanometer-sized spaces and the materials: inorganic zeolites and carbon-based materials such as

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00503-9 45

46 MetalOrganic Frameworks
activated carbons and carbon nanotubes. Recently, another type,
namely porous materials consisting of metal ions and organic
ligands, was developed that is beyond the scope of the former
two classes of porous materials. Porous materials constructed
from metal ions and organic ligands are now of great interest
because of their advantages in terms of storage, separation, and
catalysis, each of which can be rationally controlled by using their
characteristics, namely high designability, structural diversity and
regularity, and flexibility. There are three types of such porous
materials, as shown in Figure 1: the first type, called porous
coordination polymers (PCPs) or metalorganic frameworks
(MOFs), has polymeric frameworks with regular pores126; the
second type comprises discrete clusters (metalorganic polyhe-
dra, MOPs) with tailorable internal space2731; and the third is
formed by the assembly of discrete coordination complexes with-
out internal space via weaker intermolecular interactions (van der
Waals, electrostatic, and hydrogen-bonding interactions) than
coordination bonds.3238 The third type is defined as porous
assemblies of coordination complexes (PACs).38 In this section,
we focus on PCPs and MOFs.
During the past decade, coordination compounds with infi-
nite structures have been intensively studied. In particular, com-
pounds with frameworks constructed from metal ions and
organic bridging ligands form a family of polymers that are
known as coordination polymers or metalorganic frameworks have been also crystallographically characterized.46,47 The key
(in this section, we use the term coordination polymers). The
expression coordination polymer first appeared early in the
1960s, and by the mid-1960s this family of compounds had
already been reviewed.39 In an effort to open up this area of
chemistry fully, versatile synthetic approaches involving assem-
blies of molecular building blocks (metal ions, bridging ligands,
counteranions, and guest molecules) in solution have been pur-
sued to achieve the formation of target structures. The key to
success is the design of molecular building blocks that direct
the desired architectural, chemical, and physical properties to
result in solid-state materials. Coordination polymers are highly
crystalline materials, and chemists need to obtain good-quality
single crystals to understand their structure and properties. How-
ever, in general, such polymers are insoluble in solvents and
retain their original framework. Therefore, the most common
MOF or PCP
(a)
MOP
Metal ion
Organic ligand (b)
Guest molecule
PAC
(c)
Figure 1 Three types of porous materials obtained from metal ions and
organic ligands in the presence of guest molecules.
crystallization method, recrystallization, is not available. In a
surprisingly short period, the structural chemistry of coordina-
tion polymers has reached a mature level because of the applica-
tion of useful crystallization methods, such as slow diffusion, as
well as hydrothermal and solvothermal methods. A survey of
recent research work shows an extraordinary increase in the
number of published articles. The next challenge in this area is
to develop coordination polymers with chemical and physical
functionalities by modifying the coordination framework.
Unique functionalities have already begun to be found in
coordination polymers. In particular, the porous properties of
coordination polymers have been intensively studied.
In the history of coordination frameworks, Prussian blue
compounds4042 and Hofmann clathrates and their derivatives,
which are built of CN linkages between square-planar or tetra-
hedral tetracyanometallate(II) units and octahedral metal(II)
units coordinated by complementary ligands,4244 are early-
stage compounds. Indeed, their small cavities determined by
cyanide anions intrinsically limit their use to the accommoda-
tion of small molecules and cannot be applied for a variety of
guest species. In 1959, a coordination framework of {[Cu
(adiponitrile)2] NO3}n, in which organic molecules were used
as bridging ligands, was first synthesized and structurally
characterized.45 After that, Ag(I) coordination frameworks
for successful formation of a crystalline coordination framework
is to use M(I) ions (M Cu and Ag) because of their isotropic
electron configurations (d10). Hereafter, M(I) ions have been
used to provide various coordination frameworks possessing
longer bridging nitrogenous ligands than cyanide. However,
due to the structural instability after removal of guest solvents
from as-synthesized compounds, those frameworks are totally
collapsed and their porosities are no longer accessible by guest
species.46,48,49
Permanent porosity is the most significant property for
coordination polymers. The recent explosion of the research
field of coordination polymers definitely relies on the applica-
tions of the permanent porosity, in which various gaseous
molecules are accommodated. Permanent porosity based on
robust framework structures of porous coordination polymers
was for the first time confirmed by gassolid-phase sorption
experiments.50,51 This demonstration of a porous function of
coordination polymers was certainly a milestone in the
research field and triggered an expansion to the syntheses of
many other PCPs/MOFs that demonstrated functional porous
properties (anion exchange, catalytic properties, and adsorp-
tion).126 The following three features of PCPs provide pivotal
advantages: (1) designability, (2) regularity, and (3) flexibility.
(1) Designability: the key to success in obtaining highly func-
tional materials is the design of the desired architectural,
chemical, and physical properties of the resulting solid-
state compounds. One can take advantage of this in the
design of PCPs because the formation reactions mostly
occur under mild conditions, and the choice of a certain
combination of discrete building units leads with a high
probability to a desired extended network.
(2) Regularity: regular pore distribution in a porous solid is
important for adsorption because when the size of a pore
is comparable with that of a guest molecule, then the

periodic potential from the pore wall can influence the form
and orientation of the adsorbed guest molecule. A regular
pore distribution can be readily achieved for coordination
polymers as well as for inorganic porous materials. The
pores of coordination polymers have a regular periodical
structure because of their crystalline form, which affords a
periodic potential on their channel surface. The structural
relationships between adsorbed guest molecules and host
frameworks (e.g., (1) position of the guest molecules in the
pores, (2) the assembled structure of the guest molecules in
the pores, and (3) the perturbation that the guest molecules
receive from the pore walls) are key subjects for understand-
ing the adsorption behavior and the physical or chemical
properties of adsorbed guest molecules in the pores. In
addition, molecules confined in a uniform restricted nano-
space form molecular assemblies and provide unique prop-
erties that are not achieved in the bulk state.
(3) Flexibility: recent reports on the dynamic properties of
PCPs show that they are much more flexible than generally
believed. Dynamic pores can form a type of soft frame-
work with bistability, whose two states change back and
forth. A system can exist in either of these two states for
different parameters of an external field. The structural
rearrangement of molecules proceeds from the closed
phase to the open phase in response to a guest molecule.
Some PCPs have such flexibility, and they can be devel-
oped as a special class of materials, such as highly selective
gas sensors or gas-separation compounds, that cannot be
obtained using a rigid porous material. Dynamic structural
transformations based on these flexible frameworks are
among the most interesting phenomena and are presum-
ably a characteristic of coordination polymers, the so-
called third-generation PCPs.19
Coordination polymers and MOFs are among the most
prolific research areas of inorganic chemistry and crystal engi-
neering in the last 15 years. The diversity in both focus and
scientific basis of the researchers involved has led to numerous
terminology suggestions and practices for this class of com-
pounds and of several subgroups within them; additionally, a
disquieting number of acronyms are also in use for these
materials. Therefore, the International Union of Pure and
Applied Chemistry (IUPAC) division of Inorganic Chemistry
has initiated a project on Co-ordination polymers and metal
organic frameworks: terminology and nomenclature guidelines,
and a summary of the work of the task group so far and the
interactions with scientists in the area has been recently
reported.52 A coordination polymer is a coordination com-
pound with bridging ligands giving extended one-, two-, or
three-dimensional structures, while an MOF is a coordination
polymer with organic ligands and an open framework with
potential voids. Two tentative hierarchies were proposed on
the basis of the discussion, although these should not yet be
regarded as final propositions from the IUPAC project.
The realization that PCPs particularly could be designed and
synthesized in a rational way from molecular building blocks
led to the emergence of a discipline called reticular chemistry.2
A set of symbols, which is called Reticular Chemistry Structure
Resource (RCSR) symbols, have been developed and used to
describe the topology.53
MetalOrganic Frameworks 47
A large number of PCPs showing diverse and unprece-
dented crystal structures and unique and applicable porous
functionalities have been reported to date. Here, we introduce
PCP chemistry with a central focus on recent research progress.
5.03.2 Parts of Coordination Frameworks
5.03.2.1 Metal Ions
Early in the study of PCPs, transition metal ions such as Cu(I), Cu
(II), Ag(I), Zn(II), Cd(II), and Co(II) were employed despite their
high weights, high costs, and/or low abundances on Earth. Since
PCPs became applicable porous materials in place of zeolites and
porous carbon materials, it has become necessary to drive down
the cost of PCP materials. Hence, the target metal ions are now
those with small weights, low costs, and/or high abundances,
that is, Al(III), Mg(II), Ca(II), Fe(II), Na(I), and K(I). MgMOF-
74 [Mg(dot)]n (dot4 2,5-dioxidotetraphthalate) has a rigid
framework, composed of one-dimensional hexagonal channels
with a nominal diameter of approximately 12 A running parallel
to the dot ligands. The desolvated material has unsaturated
open Mg(II) sites (decorated on the edges of the hexagonal pore
channels) with an open-pore structure of large surface area
(>1000 m2 g1).54 The crystallization of the C8 symmetrical
polyoxygenated ligand, g-cyclodextrin (CD), in aqueous
alcohol in the presence of the alkali metal salt, KOH, affords a
robust PCP [K(OH)(CD)]n with a surface area of approximately
1200 m2 g1.55 A common feature of polyoxygenated ligands
that bind alkali and alkaline earth metal cations is the presence
of OCCO repeating units with gauche conformations defining
their CC bonds. This geometry orients the hard oxygen donors
in a manner that allows them to interact with hard group IA and
IIA metal cations and is undoubtedly responsible for the strong
dipoledipole interactions that lead cooperatively to large associ-
ation constants for complexes incorporating a number of them.
5.03.2.2 Secondary Building Units
Polynuclear clusters constructed from two or more metal ions
and carboxylate units (so-called secondary building units
(SBUs)) can have special coordination numbers and geome-
tries. SBUs are conceptual units that are not employed in the
synthesis as distinct molecular building blocks. However, spe-
cific SBUs can be generated in situ under the correct reaction
conditions.2 Because the metal ions are locked into their posi-
tions by the carboxylate units, SBUs are sufficiently rigid to
produce extended frameworks with a high structural stability.
In SBUs with terminal guests, the reactivity of the metal site can
be studied through the removal of these guests, which creates
coordination sites.
5.03.2.3 Organic Ligands
Charge-transfer (CT) interactions play a significant role in
developing novel photoactive materials. One available method
to introduce CT interactions between a host framework and a
guest molecule is to use redox-active ligands, because such
ligands forming the pore walls of PCPs have enough space
to attract a guest molecule at a nearby site. For example,
48 MetalOrganic Frameworks

frameworks, the resulting PCPs contain a counteranion or

cation in the pores for charge compensation, and therefore,
bpy
the pores are significantly blocked against guest species. To
Zn retain an electrostatic interaction site that has enough space
to attract guest molecules, a charge-separated neutral organic
ligand, in which a pyridinium cation and a carboxylate
anion act as a guest-accessible site and a coordination site, respec-
tively, was designed.62 The two-dimensional PCP {[Zn2(1,4-
bdc)2(cpb)] 2DMF H2O}n (1,4-bdc2 1,4-benzenedicarboxy-
(a) TCNQ
late, Hcpb 1-(4-carboxyphenyl)-4,40 -bipyridinium, DMF
N,N-dimethylformamide) possesses straight two-dimensional
channels, in which the pyridinium part of the cpb ligand is
exposed to the pore surface. This PCP adsorbs MeOH with a
large isosteric heat of adsorption (5095 kJ mol1) compared
with those of alkali-metal-exchanged molecular sieves, implying
that the pyridinium cationic surface participates in the strong
a c adsorption of MeOH.
a
5.03.2.4 Metalloligands
(b) b (c) b
Coordinatively unsaturated metal centers (UMCs) can provide
Figure 2 (a) Coordination environment of Zn(II) ion and (b) top view
functionalities as chromic sensors and flexible frameworks,
and (c) side view of {[Zn(TCNQ)(4,40 -bpy)] 6MeOH}n. Reproduced from
because the ligation and release of the UMCs guests often
Shimomura, S.; Matsuda, R.; Tsujino, T.; Kawamura, T.; Kitagawa, S. J.
Am. Chem. Soc. 2006, 128, 1641616417, with permission of the influence the coordination geometry or ligand field. Moreover,
copyright holders. a variety of metal ions can act as active centers in catalyzed
reactions. A combination of UMC characteristics and tradi-
tional porous properties (shape and size selectivity) can
{[Zn(TCNQ)(4,40 -bpy)] 6MeOH}n (TCNQ2 7,7,8,8-tetra-
be used to create desired, highly efficient, and tailor-made
cyano-p-quinodimethane dianion, 4,40 -bpy 4,40 -bipyridine)
functional materials. Recently, a new synthetic method was
has a three-dimensional pillared-layer framework with a redox-
proposed for PCPs with UMCs using metalloligands, that is, a
active ligand, TCNQ2.56 The Zn(II) ions are linked by TCNQ2
two-step self-assembly.21 First, a metalloligand is synthesized
dianions to give a two-dimensional corrugated layer. The 4,40 -
that acts as a framework linker and a source of coordinatively
bpy ligands act as pillars, with the Zn(II) ions in the adjacent
UMCs. Second, the metalloligand is added to another metal
layers linked to form a three-dimensional pillared-layer structure
ion, which acts as a nodal unit in a framework. Consequently,
(Figure 2). This compound possesses two-dimensional channels
two types of metal centers coexist in the framework, and a
with cross-sections of 3.4  5.9 A2 between the layers. As a result,
larger space around the metal ion in the channel wall can
a dense array of strong donor sites (TCNQ2) is formed on the
be obtained, which is significant for an attack by a guest
pore surface of the framework, providing a highly electron-rich
molecule. The key point of this method is a partial separation
surface for the guest molecules. When {[Zn(TCNQ)(4,40 -bpy)]
of the metal functionalities: the framework node and the
6MeOH}n is immersed in benzene, the guest MeOH can be
coordinatively UMCs.
exchanged with benzene to form {[Zn(TCNQ)(4,40 -bpy)] 2ben-
Metalloligands are also employed for the insertion of metal
zene}n within a period of 10 s, producing a compound that
ions that predominantly afford discrete coordination complexes
is red in color versus the yellow of the parent compound.
in coordination polymer frameworks. Ir(III) complexes such as
The crystal structure of {[Zn(TCNQ)(4,40 -bpy)] 2benzene}n
[Ir(ppy)3] (ppy 2-phenylpyridine) are highly efficient phos-
indicates that the benzene guests interact with the most nega-
phorescent molecules because of the facile intersystem crossing
tively charged carbon atom of the TCNQ2 dianions via
to populate the triple metal-to-ligand charge-transfer (3MLCT)
CH C-type hydrogen bonds, showing the existence of an
states, and have been extensively used in light-emitting devices.
electrostatic and/or CT interaction between the TCNQ2 dia-
However, PCPs containing only Ir(III) ions have been seldom
nions and the benzene guests. For other guests (e.g., toluene,
observed because of the substitution inertness of Ir(III) com-
ethylbenzene, anisole, benzonitrile, and nitrobenzene), changes
plexes. Recently, Ir(III) ions were successfully inserted into PCPs
to colors different from that of {[Zn(TCNQ)(4,40 -bpy)]
through metalloligands.63 [Ir(ppy)3] derivatives containing three
6MeOH}n, which was similar to that observed for benzene,
carboxylic acid groups were first synthesized as metalloligands
were seen; the exceptions were anisole and nitrobenzene,
(H3L1 and H3L2; see Figure 3(a)). They were then reacted with Zn
where the occlusion of anisole and nitrobenzene forms light-
(II) salts to form two-dimensional bilayer PCPs (Figure 3(b) and
yellow and dark-brown crystals, respectively. Such a difference
3(c)), one of which shows permanent porosity and oxygen sens-
in color is associated with the electron-accepting characteristics
ing via efficient and reversible luminescence quenching.
of the guest molecules.
Electrostatic interaction sites afford a stronger confinement
5.03.2.5 Inorganic Counteranions
effect for guest molecules than that in the case of a dispersion-
type interaction. Several charged frameworks have been syn- Charge is an important factor in the rational construction of
thesized to enhance the hostguest interactions.5761 In these PCPs. Because most transition metal ions exist with a positive

COOH
COOH
Ir Ir
N N
3 3
H3L1 H3L2
(a)
(b)
Figure 3 (a) Ir(III)-containing metalloligands, H3L1 and H3L2.
(b) Connectivity of the metalloligand L23 and (c) two-dimensional
bilayer structure of {[Zn4(m4-O)(L2)2] 6DMF H2O} n. Reproduced from
Xie, Z.; Ma, L.; deKrafft, K. E.; Jin, A.; Lin, W. J. Am. Chem. Soc. 2010,
132, 922923, with permission of the copyright holders.
charge in the framework, an anionic source must be included
to neutralize the overall charge. Many organic N-containing
heterocyclic ligands are neutral and, therefore, counteranions
are included in the crystals. Frequently used inorganic anions
are ClO4, BF4, NO3, NCS, PF6, NO2, SiF62, CN,
CF3SO3, SO42, N3, and halides, which are introduced
along with metal ions from the corresponding metal, sodium,
and potassium salts. These inorganic anions exist as: (1) free
guests, (2) hydrogen-bond acceptors, and (3) coordination
units. An important characteristic of inorganic anions is their
ability to act as hydrogen-bond acceptor sites through their
O atoms.
5.03.2.6 Guest Molecules
A method of tweaking the porous properties simply by
replacing a molecular cartridge with a different cartridge
has been developed in a new class of PCPs.64,65 The three-
dimensional PCP, {[Zn3I6(tpt)2(triphenylene)] x(nitrobenzene)
y(MeOH)}n (tpt 2,4,6-tris(4-pyridyl)triazine), is composed of
two interpenetrating networks, in which the pores are surrounded
by aromatic bricks.66 These bricks consist of alternatively layered
tpt and triphenylene. The former forms an infinite three-
dimensional network via coordination to ZnI2, whereas the
latter is involved in the three-dimensional network without
forming any covalent or coordination bonds with other com-
ponents. Therefore, the noncovalently intercalated tripheny-
lene can be replaced with functionalized triphenylenes
without causing any change in the porous structure, and inter-
calated triphenylenes are regarded as the cartridges of the func-
tional groups. For example, one of the triphenylenes directs a
phenol group toward the interior of the pores. The resulting
PCP selectively adsorbs alcohol guests such as propan-2-ol.
MetalOrganic Frameworks 49
HC(O)NR2 + H2O HCO2H + HNR2
R = Me, Et
Scheme 1 Hydrolysis reaction of DMF and DEF.
5.03.3 Synthetic Techniques
5.03.3.1 Solvothermal Methods
Solvothermal synthesis in DMF or diethylformamide (DEF) solu-
tion is an effective method for obtaining good-quality single
crystals of PCPs with carboxylate-type and imidazolate-type
ligands. DMF and DEF readily undergo a small degree of hydro-
lysis in air, forming formic acid and HNR2 (R Me and Et), as
shown in Scheme 1. The resulting amines can promote the
deprotonation of carboxylic acids and imidazoles and generate
carboxylates and imidazolates that have the ability to bridge
metal centers.
5.03.3.2 Microwave Syntheses
Microwave irradiation has been developed as a novel heating
technique for the solvothermal synthesis of organic or inorganic
solid-state materials. In particular, this method has been shown
to provide an efficient way to synthesize purely inorganic porous
materials with short crystallization times, facile morphology
control, narrow particle-size distributions, phase selectivity,
and efficient evaluation of process parameters. Even under
ambient conditions, microwave heating of a mixture of starting
materials results in an efficient reaction to provide interesting
PCPs. The first successful example of a microwave synthesis for
PCPs was achieved in the case of {[Cr3FO(1,3,5-btc)2(H2O)3]
xH2O}n (1,3,5-btc3 1,3,5-benzenetricarboxylate, MIL-100),
which has a hierarchical pore system (microporous 59 A
and mesoporous 2530 A) with a very high specific surface
area (3100 m2 g1).67,68 The x-ray diffraction (XRD) pattern
of a sample obtained after microwave irradiation (reaction
temperature 493 K) for a period of 4 h agrees well with the
pattern of MIL-100 synthesized for a period of 4 days at 493 K
using conventional solvothermal heating. The crystal yield
from microwave synthesis was 44%, which is similar to the result
of 45% from the conventional synthesis after 4 days. The phys-
ical and textural properties are also very similar to those of
crystals synthesized using conventional solvothermal heating.
5.03.3.3 Post-Synthetic Modification
Post-synthetic modification of the pore surface without
destruction of the framework structure after the formation of
PCPs is the next challenging task for obtaining adequate space
for specific guests.69 Four types of post-synthetic modifications
are known to regulate pore environments: (1) modification of
organic ligands, (2) introduction of metal-containing units,
(3) incorporation of neutral organic molecules, and (4) cation
exchanges.
Novel strategies have been developed to introduce a catalytic
metal center by post-synthetic modification of the porous frame-
work. A chiral bridging ligand, L3, with orthogonal functional
groups, as shown in Figure 4, bridges the Cd(II) centers of the
50 MetalOrganic Frameworks
Cl
Cl
(a)
(b)
Figure 4 (a) Axially chiral bridging ligand, (R)-L3. (b) Schematic representation and space-filling model of the three-dimensional framework of
{[Cd3Cl6(L3)8] 4DMF 6MeOH 3H2O}n. Reproduced from Wu, C.-D.; Hu, A.; Zhang, L.; Lin, W. J. Am. Chem. Soc. 2005, 127, 89408941, with permission
of the copyright holders.
one-dimensional [Cd(m-Cl)2]n units to form an extended porous
network in {[Cd3Cl6(L3)8] 4DMF 6MeOH 3H2O}n with very
large chiral channels with an area of 16  18 A2.70 The channel
walls of this PCP have free OH sites acting as secondary func-
tional groups that can further interact with other metal ions.
Indeed, Ti(OiPr)4 can react with the chiral dihydroxy groups in
the L3 parts of this PCP, producing an active heterogeneous
catalyst for the addition of ZnEt2 to aromatic aldehydes to
produce chiral secondary alcohols. The performance of this het-
erogeneous catalyst rivals that of the homogeneous analogue
under similar conditions. The three-dimensional PCP [Zn4O
(1,4-bdc)3]n (MOF-5) can incorporate photoactive Cr(CO)3
groups onto the benzene rings of 1,4-bdc2 ligands in an Z6
fashion after construction of the framework.71 The obtained
PCP, [Zn4O((Z6-1,4-bdc)Cr(CO)3)3]n, shows substitution
of a single CO ligand per metal by either N2 or H2 under
gentle photolysis conditions, forming [Zn4O((Z6-1,4-bdc)Cr
Primary
N functional
group
OH Secondary
functional
OH group
Primary
N functional
group
(R)L3
1.8 nm
(CO)2(X))3]n (X N2 or H2). These PCPs display a remarkable
stability compared with their analogues [(C6H5Me)Cr(CO)2(X)]
(X N2 or H2) because the Cr(0) centers are pinned to the walls of
the framework, which prevents reactions involving two or
more Cr(0) atoms and enables the irreversible evacuation of CO.
{(Me2NH2)2[Zn8O(ad)4(bpdc)6]}n (ad adeninate,
bpdc biphenyldicarboxylate) is a permanent PCP with
2
infinite Zn(II)ad columnar building units that are intercon-
nected via bpdc2 linkers, resulting in a three-dimensional
structure with large one-dimensional channels.72 This PCP
is anionic with dimethylammonium cations residing in
its pores. A series of organic cations (tetramethylammonium,
tetraethylammonium, and tetrabutylammonium) were intro-
duced via cation exchange into the pores to systematically
decrease the pore volume and surface area of the material.73
Such modifications can be used to tune systematically the CO2
adsorption capacity of this material.
 MetalOrganic Frameworks 51

5.03.3.4 Methods for Controlling Morphologies
of PCP Crystals
A key challenge is the precise control of crystal growth of PCPs,
which is required to integrate novel functions into the mate-
rials without changing the characteristic features of the PCP
crystal itself. Functionalized organic polymers have been
widely used as substances for the preparation of metal-based
nanoparticles. The selection of appropriate protective polymers
leads not only to various colloid morphologies by controlling
the growth and handling the agglomeration process, but also
to attractive new properties based on organicinorganic
hybridizations at the nanometer scale. The control of crystal
growth of the PCP [Cu2(pzdc)2(pyz)]n (pzdc2 pyrazine-2,3-
dicarboxylate, pyz pyrazine, CPL-1 (CPL coordination
pillared layer)) has been achieved by addition of the organic
polymer poly(vinylsulfonic acid, sodium salt) (PVSA), which
has the ability to complex and stabilize Cu(II) by electrostatic
interaction.74 The addition of PVSA to the reaction mixture
significantly affects the nucleation process, which determines
the crystal size of this PCP in the range of 1100 mm. PVSA not
only controls the crystal size and shape but also prompts the
preferential orientation of the plate crystals, which results in
alignment of the channel direction in the bulk powder state. N2
adsorption measurements show typical micropore adsorption
isotherms that are independent of the crystal size. However, the
adsorption speed of the samples shows a size-dependent feature
because of changes in the diffusion length of N2 in the one-
dimensional channels.
A simple but straightforward method using capping reagents
(modulators) that perturb the framework extension of PCPs
has been developed.75,76 Initial work shows that addition of
p-perfluoromethylbenzenecarboxylic acid as a modulator
yields stable size-selected MOF-5 colloids.75 A coordination
modulation method (Figure 5), in which acetic acid is used
to influence directly the coordination equilibria, can be used to
control the crystal growth of nanosized PCP crystals of [Cu2(1,4-
ndc)2 (dabco)]n (1,4-ndc2 1,4-naphthalenedicarboxylate,
dabco 1,4-diazabicyclo[2.2.2]octane).76 The competitive inter-
action between the coordination mode used to construct the
framework and the acetatecopper interaction plays a crucial
role in determining the reaction rate and crystal morphology.
The mechanism of anisotropic crystal growth is clearly explained
by oriented attachment, in which nanorods are formed. This
method also affords perfect framework regularity in PCP nano-
crystals, contributing to an increase in the adsorbed amounts for
N2 and CO2 gases.
5.03.3.5 Techniques for Obtaining PCP Thin Films
Most of the research on the properties and applications of PCPs
deal with microcrystalline powder samples. In contrast to the
overall very rapid progress in PCP research, PCP thin films are
relatively little developed. The fabrication of novel functional
devices ranging from sensing to smart membranes and micro-
reactors will require the fabrication of PCP materials that pos-
sess the desired properties in a variety of sizes, shapes, and
orientations including smooth, nanosized crystalline multi-
layers and thick, homogeneous, and defect-free films in the
micron-size regime on the appropriate substrate. Recently,
SURMOF (surface-grown metalorganic framework) was
introduced and defined specifically to address layer-by-layer
grown, crystalline, oriented, and smooth multilayers and/or
thin films of PCPs attached to surfaces.77 The SURMOF of the
three-dimensional framework of [Zn2(1,4-bdc)2(4,40 -bpy)]n
was fabricated by the step-by-step synthesis on an Au substrate
covered with a self-assembled monolayer from 4,4-pyridyl-
benzenemethanethiol.78 The microcrystalline powder sample
of this PCP consists of two interpenetrating, pillared, paddle-
wheel-type networks.79 In contrast, the obtained SURMOF
showed an XRD pattern expected for an oriented and nonin-
terpenetrated PCP layer (Figure 6). That is, the formation of
the second, interpenetrated framework can be suppressed by
using liquid-phase epitaxy on an organic template.

Conventional method

Crystal growth

Nucleation

Metal ion
+
Bulk crystal

Coordination modulation method

Organic linker

+
Modulator

Crystal growth
by selective modulation
Competitive interactions Nanocube Nanorod Nanosheet
Modulated nucleation

Figure 5 Coordination-modulation method for fabricating PCP nanocrystals. Modulated nucleation is induced by the addition of modulators with the
same functionality as organic linkers to impede the coordination interaction between metal ions and organic linkers. Selective coordination modulation is
one of the coordination modes (illustrated as blue shells) for constructing the framework, which then leads to anisotropic growth of the PCP crystals.
Reproduced from Tsuruoka, T.; Furukawa, S.; Takashima, Y.; Yoshida, K.; Isoda, S.; Kitagawa, S. Angew. Chem., Int. Ed. 2009, 48, 47394743, with
permission of the copyright holders.
52 MetalOrganic Frameworks
SURMOF
experimentel
Calculated
no interpenetration
/
001
002
Calculated MOF-508a
5 10 15
2q ()
(a)
Figure 6 (a) Out-of-plane XRD patterns. (b) Proposed model structure of the SURMOF, solvent-free, noninterpenetrated pseudomorph of
[Zn2(1,4-bdc)2(4,40 -bpy)]n, grown along the [001] direction. Reproduced from Zacher, D.; Schmid, R.; Woll, C.; Fischer, R. A. Angew. Chem. Int.
Ed. 2011, 50, 176199, with permission of the copyright holders.
5.03.3.6 Other Techniques
Solvent-free synthesis of PCPs is of interest from the viewpoint
of green chemistry. The quantitative and rapid synthesis of the
PCP [Cu(ina)2]n (ina isonicotinate) was performed by
grinding together copper(II) acetate and isonicotinic acid for
a few minutes and then heating to remove the water and acetic
acid byproducts.80 The obtained compound has the character-
istic odor of acetic acid, released as a byproduct. The fully
desolvated compound is obtained by heating the mixture to
473 K for a period of 3 h, and shows a similar XRD pattern to
that simulated from the single-crystal data for the empty frame-
work host that was synthesized in the solvothermal reaction
between 4-cyanopyridine and CuCl2 in a mixture of water and
ethanol after heating to 423 K under autogenous pressure for a
period of 48 h.81,82 This solvent-free method is quick and gives
a quantitative yield without the need for solvents or external
heating. Clearly, it can present higher efficiency in terms of mate-
rials, energy, and time compared with solvothermal methods.
The in situ slow hydrolysis of precursor ligands can result in less
soluble phases that are not accessible directly from their corre-
sponding hydrolyzed ligands by virtue of the presence of a large
excess of metal ions under solvothermal conditions.83 For exam-
ple, Cd(II) and Zn(II) coordination polymers that show second-
order, nonlinear optical properties can be obtained by reaction of
the metal salts with cyanopyridine and with pyridinecarboxalde-
hyde and its derivatives, whose cyano, carboxaldehyde, and ester
substituent groups slowly hydrolyze to form the corresponding
carboxylic acid.84 The reaction of NaN3, 3-cyanopyridine, and
ZnCl2 or CdCl2 under solvothermal conditions yields [Zn(OH)
(3-ptz)]n or [CdN3(3-ptz)]n (3-ptz 5-(3-pyridyl)tetrazolate).85
For each reaction, the infrared (IR) spectrum indicates the absence
of the cyano group, which is consistent with a [2 3] cycloaddi-
tion between the cyano group and the azide anion.
The nucleation process is the most important step in deter-
mining the crystal phase of the system, because an alteration of
the conditions often leads to several different crystal structures
for the same initial chemical system, so-called polymorphism.
Therefore, an understanding of the nucleation process of PCPs
provides considerable insight into the synthetic procedure of
PCP materials as they are obtained as crystalline materials. A
methodology for systematically controlling the connectivity
H
C
N
O
Zn
20
(b)
between metal clusters in PCPs was recently established by
controlling the nucleation process.86 Two polymorphs of
[Zn2(1,4-bdc)2(dabco)]n have been reported so far; one is a
tetragonal framework, in which two-dimensional square-grid
sheets constructed by Zn(II) paddlewheel units and 1,4-bdc2
are connected by dabco pillar ligands,87 and the other is
a trigonal framework with two-dimensional Kagome nets
instead as layers.88 The tetragonal framework is obtained with
a long reaction time at high temperature (393 K) to initiate the
nucleation and the solvent-mediated transformation process,
whereas the trigonal framework is formed by a shorter reaction
time at 393 K followed by nucleation at room temperature
(298 K), because of the large difference in solubility of the
oligomers at 393 and 298 K.
5.03.4 Representative Topologies of Frameworks
5.03.4.1 CPL Structures
The structures of CPLs, [Cu2(pzdc)2(P)]n (P pillar ligands),
consist of two-dimensional layers [Cu(pzdc)]n and various
bridging ligands as pillars (Figure 7). The CPLs are synthesized
by a simple procedure; a mixture of EtOH and H2O solution of
Na2pzdc and the pillar ligand (P) is slowly added to the H2O
solution containing Cu(II) ion at room temperature in air.
Stirring for 1 day provides the desired compounds in good
yield. [Cu2(pzdc)2(pyz)]n is the prototype CPL (CPL-1).89 A
neutral two-dimensional layer of [Cu(pzdc)]n is formed in the
ac plane. These layers are linked by pillar pyz ligands, forming a
three-dimensional pillared-layer structure with one-dimensional
channels (4  6 A2). Because a large number of atoms are densely
packed in the two-dimensional layer, guest molecules cannot
pass through this layer. The pillar ligands are aligned closely
along the a-axis, in which the space between the nearest pillar
ligands is less than 0.5 A, based on the van der Waals radii.
Therefore, the translational motion of guest molecules along the
b- and c-axes is forbidden. That is, guest molecules can only pass
along the a-axis. As a result, CPL series have one-dimensional
channels suitable for the formation of low-dimensional assem-
blies of small guest molecules.9092 The CPL derivatives,
[Cu2(pzdc)2(P)]n (P 4,40 -bpy (CPL-2), 2,7-diazapyrene
 MetalOrganic Frameworks 53

N COO- N

Cu2+
-
N COO

Dypyridyl pillar ligand (L)

2D sheets 3D porous framework

[Cu(pzdc)]n [Cu2(pzdc)2L]n

Figure 7 Schematic view of the structure of CPL. Reproduced from Sakamoto, H.; Kitaura, R.; Matsuda, R.; Kitagawa, S.; Kubota, Y.; Takata, M. Chem.
Lett. 2010, 39, 218219, with permission of the copyright holders.

(CPL-3), 4,40 -azobispyridine (azpy, CPL-4), 1,2-bis(4-pyridyl)

ethylene (bpethe, CPL-5), and 3,6-bis(4-pyridyl)-1,2,4,5-
tetrazine (bpt, CPL-11)), have also been crystallographically char-
acterized.9398 Because the crystals of these CPLs are very small,
the crystal structures were determined by the Rietveld analysis of
synchrotron x-ray powder diffraction patterns measured on the
BL02B2 beamline at SPring-8. The two-dimensional layers paral-
lel to the ac plane are constructed from Cu(II) ions and pzdc2
ligands with the same connectivity as that of CPL-1. They are
further connected by pillar ligands, forming CPL structures with
different channel sizes (8 A  6 A (CPL-2), 8 A  3 A (CPL-3), and
10 A  6 A (CPL-4, CPL-5, and CPL-11)). A CPL with another
pillar ligand, [Cu2(pzdc)2(4-pia)]n (4-pia N-(4-pyridyl)isoni-
cotinamide, CPL-6), was also synthesized and its structural
similarity was checked briefly by the normal powder XRD
measurement.89 These results strongly indicate that the pore
dimension and surface functionalities can be controlled simply
by changing the pillar ligands.

5.03.4.2 Jungle-Gym-Type Structures

The a-Po-type framework, whose appearance is like a jungle
gym, is a candidate for the construction of exceptionally rigid
and highly porous structures with three-dimensional inter-
secting channels. A systematic design of the pore size and
functionality has been achieved in three-dimensional PCPs
[Zn4O(L)]n (IRMOFs, L a variety of anionic dicarboxylate
ligands),99 in which an oxide-centered Zn4O tetrahedron is
edge-bridged by six carboxylates to give an octahedron-shaped
SBU that reticulates into a three-dimensional porous network
(Figure 8). Other a-Po-type PCPs have been systematically
synthesized in [M2(L)2(L0 )]n (M Zn(II) and Cu(II), L a
variety of anionic dicarboxylate ligands, and L0 neutral
nitrogen-containing heterocyclic ligands) with two types of
organic bridging ligands.87,100102 The prototype compound,
[Zn2(1,4-bdc)2(dabco)]n, is composed of dinuclear Zn2 units
with a paddlewheel structure that are bridged by 1,4-bdc2
dianions to form a distorted two-dimensional square grid
[Zn2(1,4-bdc)2]n.87 The axial sites of the Zn2 paddlewheels
are occupied by dabco molecules, which act as pillars to extend
the two-dimensional layers into a three-dimensional structure
with a compressed primitive cubic (a-Po) net. The framework
has large, three-dimensional interconnected voids. The wide,
open channels (7.5  7.5 A2) running along the c-axis are
interlinked by smaller windows (4 A) oriented along the
Figure 8 Single crystal x-ray structures of IRMOFs. Reproduced from
Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; OKeeffe, M.;
Yaghi, O. M. Science 2002, 295, 469472, with permission of the
copyright holders.
a- and b-axes, which are still large enough for the passage of
small gas molecules. In the case of compounds containing
dabco pillar ligands, noninterpenetrated frameworks are
always observed because the dabco ligand is too short to
allow the insertion of another framework into the windows
that form between the two-dimensional layers. On the other
hand, compounds with longer 4,40 -bpy pillar ligands than
dabco have easily formed interpenetrated frameworks.
5.03.4.3 Zeolitic Imidazolate Frameworks
Zeolite networks are the most significant topologies for
microporous materials. Three new zeolite-type PCPs, {[Zn
(mim)2] 2H2O}n, {[Zn(eim)2] 2H2O}n, and {[Zn(eim/mim)2]
54 MetalOrganic Frameworks
1.25H2O}n (mim 2-methylimidazolate and eim 2-ethyli-
midazolate), have been synthesized by the liquid-phase diffusion
method.103 Imidazolate (im) bridges make an MimM angle
close to 145 , which is coincident with the SiOSi angle, which
is the preferred configuration that is commonly found in many
zeolites. Therefore, under the right conditions, it is possible to
prepare microporous zeolite-type frameworks using imidazolate-
type ligands. Although all of these compounds have infinite
three-dimensional frameworks, their topologies are different
from each other (zeolitic sodalite (SOD, 4264), analcime (ANA,
426282), and rho (RHO, 4363) topologies, respectively). By taking
advantage of simple modifications of imidazolates at a coordina-
tively unimportant position, the orientation of adjacent metal
coordination polyhedra can be tuned. Since the above report
was published,103 a large number of porous metalimidazolate
frameworks, known as zeolitic imidazolate frameworks (ZIFs),
have been systematically synthesized and characterized.104106 A
study of the gas adsorption and thermal and chemical stabilities of
two prototypical members, [Zn(mim)2]n (ZIF-8) and [Zn
(phim)2]n (phim benzimidazolate, ZIF-11), demonstrated
their permanent porosity (Langmuir specific surface area-
1810 m2 g1), high thermal stability (up to 823 K), and remark-
able chemical resistance to boiling alkaline water and organic
solvents.104
5.03.4.4 Coordination Polymers with an
Interdigitated Structure
The two-dimensional layer-type PCPs constructed from M(II)
ions, V-shaped dicarboxylates, and bipyridyl derivatives are avail-
able for the systematic design of different types of two-
dimensional layers.107111 In the frameworks, the ribbon struc-
tures of M(II) and dicarboxylate are linked by bipyridyl derivatives
to form two-dimensional layers and each layer is interdigitated to
create a three-dimensional structure, called a coordination poly-
mer with an interdigitated structure (CID).112119 The structure of
the prototype CID [Zn(1,3-bdc)(4,40 -bpy)]n (CID-1) is shown in
Figure 9. Here, replacement of substituent dicarboxylate groups
can modulate the pore shape and affinity. For example, CID-5
[Zn(5-NO2-1,3-bdc)(4,40 -bpy)]n (5-NO2-1,3-bdc2 5-nitro-1,
3-benzenedicarboxylate) and CID-6 [Zn(5-MeO-1,3-bdc)(4,40 -
bpy)]n (5-MeO-1,3-bdc2 5-methoxy-1,3-benzenedicarboxy-
late) show different adsorption properties: gate-opening-type
adsorption and a type I adsorption characteristic of permanent
porosity, respectively.118 Furthermore, the ligand-based solid
solution approach affords the solid solutions formulated as [Zn
(5-NO2-1,3-bdc)1x(5-MeO-1,3-bdc)x(4,40 -bpy)]n (CID-5/6).
Thereby, the gate-opening pressure is precisely tuned. In this
family of compounds, the channels are surrounded by bipyridyl
derivatives (typically 4,40 -bpy). Therefore, the use of other, longer
bipyridyl derivatives can create a larger pore volume in the struc-
tures and affords another strategy by which to modify the
chemical environment of the pore interiors. CID-21 [Zn(1,3-
bdc)(bpb)]n (bpb 1,4-bis(4-pyridyl)benzene), CID-22 [Zn(1,
3-bdc)(bpt)]n (bpt 3,6-bis(4-pyridyl)-1,2,4,5-tetrazine), and
CID-23 [Zn(1,3-bdc)(bpa)]n (bpa 1,4-bis(4-pyridyl)acetylene)
all have similar void volumes.119 However, CID-21 and CID-22
have relatively rigid porous structures, whereas CID-23 has greater
flexibility and exhibits a high selectivity for benzene over cyclo-
hexane at 298 K.
5.03.4.5 PCPs Based on One-Dimensional Frameworks
Porous materials based on one-dimensional frameworks are
limited to date. The typical one-dimensional porous materials
are carbon nanotubes, which have two types of pores: one is
inside the tubes and the other is in the interstices of the
nanotube bundle.120
One-dimensional PCPs [M2(bza)4(L)]n (M Rh(II) and Cu
(II); bza benzoate; L pyz, 2-methylpyrazine (2-mpyz),
2-ethyl-3-methylpyrazine) are compounds famous as single-
crystal adsorbents.121123 The component chains adopt a per-
fectly linear geometry with the chain skeleton bridged by the
pyz ligands in the axial direction of the well-known paddle-
wheel type of dinuclear metal benzoate. Each one-dimensional
chain aggregates by pp stacking of the aromatic rings of the
ligands. Although the original framework has no channels,
there are empty cages with dimensions of 9 A  4 A  3 A,
with narrow windows of approximately 1 A at their four cor-
ners, which are formed by the benzene rings of bza moieties
arranged parallel to the chain vector. On cooling samples in
the presence of guest molecules, the crystal undergoes a phase
transition to a new structure, generating one-dimensional
channels. Because their single crystals retain their crystallinity
and shape after guest adsorption, the exact aggregate structures
of various accommodated guests by conventional single-crystal
XRD analysis can be determined. For example, [Rh2(bza)4
(2-mpyz)]n captures O2 trimers into one-dimensional chan-
nels over a wide range of temperatures, from room temperature
down to 10 K.122
5.03.5 Porous Functionalities
5.03.5.1 Storage
5.03.5.1.1 Methane
Methane, CH4, stands out among various alternative fuels
when its profusion and availability are considered. However,
the lack of an effective, economic, and safe on-board storage
system is one of the major technical barriers preventing CH4-
fuelled automobiles from competing with traditional vehicles.
To promote the vehicular application of methane, the US
Department of Energy (DOE) has set a target for CH4 storage
of 180 cm3(STP) cm3 (standard temperature and pressure
equivalent volume of CH4 per volume of the adsorbent mate-
rial) at 3.5 MPa, near to ambient temperature, with the energy
density of adsorbed natural gas being comparable to that of
compressed natural gas used in current practice. The ideal
material for achieving high CH4 adsorption per mass or vol-
ume of adsorbent would have the following properties: a high
specific surface area, a high free volume, a low adsorbent
framework density, and a strong energetic interaction with
the adsorbed CH4; however, these are not all necessarily com-
patible. For example, increasing the free volume may create
very wide pores, but the heat of adsorption for molecules in the
centers of such wide pores may be close to zero, and so this
space would be wasted. On the other hand, the heat of adsorp-
tion is higher for smaller pores, but the capacity of such pores
may be smaller. Therefore, the pores of CH4 adsorbents must
have an appropriate size and surface area. A computational
study has indicated that aromatic rings in PCPs can improve
 MetalOrganic Frameworks 55

(a)

(b)

(c)

Figure 9 (a) One-dimensional ribbon structure, (b) two-dimensional layer structure, and (c) three-dimensional assembled structure of CID-1
[Zn(1,3-bdc)(4,40 -bpy)]n.
56 MetalOrganic Frameworks

the CH4 binding energy, enhancing both the uptake and the
heat of adsorption.124 Several PCPs have been screened for CH4
storage, but none has reached the DOE target.99,125127 Recently,
the three-dimensional PCP [Cu2(adip)]n (adip4 5,50 -(9,
10-anthracenediyl)di-isophthalate) with cuboctahedral nano-
scopic cages (Figure 10(a)) exhibited a CH4 adsorption capacity
that is higher than the DOE target for CH4 storage (Figure 10(b)
and 10(c)).128 The Langmuir specific surface area and specific
pore volume are 2176 m2 g1 and 0.87 cm3 g1, respectively,
estimated from N2 adsorption measurements at 77 K. High-
pressure CH4 adsorption studies show that this PCP has an
absolute CH4 adsorption capacity of 230 cm3(STP) cm3
(28% higher than the DOE target of 180 cm3(STP) cm3 at
ambient temperature), and a heat of adsorption of around
30 kJ mol1: both record high values among those reported
(a)
for CH4 storage materials. The high isosteric heat of adsorption
supports the predictions that adding more aromatic rings to the 500
ligand and incorporating nanoscopic cages into the framework
125 K
can lead to a higher affinity for CH4.124 150 K
400
200 K
5.03.5.1.2 Carbon dioxide

Uptake (v(STP)/v)
Removal of carbon dioxide, CO2, from the flue exhausts of 240 K
300 270 K
power plants, currently a major source of emissions, is com- 280 K
monly accomplished by chilling and pressurizing the exhaust 290 K
gas or by passing the fumes through a fluidized bed of aqueous 200
amine solution, neither of which is cost effective.129 Other
methods based on the chemisorption of CO2 on oxide surfaces
or adsorption within porous silicates, carbon, and membranes 100
have been pursued as means of CO2 uptake.130 The room-
temperature CO2 adsorption isotherms of some PCPs ranging
0
in specific surface area from 345 to 4508 m2 g1 have been
0 10 20 30 40 50
measured to examine a range of structural and porous
attributes.125,131 The voluminous space enclosed by [Zn4O (b) P (bar)
(btb)2]n (btb3 1,3,5-benzenetribenzoate, MOF-177) with a
specific surface area of 4508 m2 g1 results in a CO2 capacity 500
of 33.5 mmol g1 at 3.5 MPa, which is far greater than that of 125 K
any other reported porous material, and exceeds standard mate- 150 K
400
rials (a carbon sample of MAXSORB shows an uptake of 200 K
Uptake (v(STP)/v)

25 mmol g1 at 3.5 MPa),132 having 150% of their capacity.131 240 K

The three-dimensional PCP [Cr(OH)(1,4-bdc)]n (MIL-53) 300 270 K 280 K
shows unprecedented CO2 adsorption properties.133 Desolvated
MIL-53 exhibits a CO2 uptake of around 8.5 mmol g1 at 290 K
2.0 MPa and 304 K. The adsorption of CO2 on hydrated 200
MIL-53 shows very little uptake at pressures up to 1.0 MPa,
while a single and distinct uptake of  7.7 mmol g1 occurs in
100
the 1.21.8 MPa pressure range. This is not the case for other
CO2 adsorbents, such as zeolites, where H2O often blocks spe-
cific sites. It has been shown previously that MIL-53 is stable at 0
high humidity, and thus, a process in which a significant 0 10 20 30 40 50
amount of H2O is present may not diminish the adsorption of (c) P (bar)
CO2. Thus, a pre-adsorber would not be required in a pressure-
swing adsorption-type process, and process design could be Figure 10 (a) Three-dimensional framework of [Cu2(adip)]n viewed as
a space-filling model. (b and c) High-pressure CH4 adsorption isotherms
simplified.
at various temperatures ((b) excess adsorption and (c) absolute
adsorption). Reproduced from Ma, S.; Sun, S.; Simmons, J. M.; Collier,
5.03.5.1.3 Hydrogen C. D.; Yuan, D.; Zhou, H.-C. J. Am. Chem. Soc. 2008, 130, 10121016,
Hydrogen, H2, is one of the leading candidates as an energy with permission of the copyright holders.
carrier of the future because of its high energy content and clean
burning, potentially renewable nature. A particularly daunting
challenge hampering its use in transportation, however, is the MOF-177 [Zn4O(btb)2]n is an excellent material to serve
development of a safe and practical storage system. PCPs are as a benchmark H2 adsorber because of the following prop-
perhaps the most promising candidates for H2 storage. erties: (1) high gravimetric and volumetric uptake capacity

(surface excess weight uptake of 75 mg g1 and total (absolute)
uptake of 110 mg g1 at 70 bar, 77 K), leading to ease of high
accuracy measurements; (2) simple and highly reproducible
synthesis; and (3) crystalline structure that is well characterized
in terms of atomic connectivity and chemical composition.134
In addition to the large storage space, implementation of
the large enthalpy of adsorption for H2 is a key issue in build-
ing practical storage systems. To achieve this, creation of co-
ordinatively UMCs and doping of PCPs with alkali metals have
been proposed and performed. Creation of PCPs with high
charge density is another way to achieve a high enthalpy of
adsorption with a wide loading amount of H2. Azulenes con-
stitute dipolar aromatic systems and, therefore, azulenes con-
tained in a PCP could contribute to the building up of
Coulombic fields required for the polarization and polarized
binding of H2. It was calculated that the H2 molecules over an
azulene ring system would have higher interaction energy in
comparison with simple benzene and isomeric naphthalene
systems because of the internal charge separation of the
azulene ring system.135,136 Based on this viewpoint, a PCP
bridged by internally polarized angular 1,3-azulenedicarboxylate
ligands was synthesized.137 [Zn4O(L)3]n (MOF-646; L 1,3-
azulenedicarboxylate) has the same chemical formula as other
IRMOFs and small pore-size distribution (<6 A). The Zn4O unit
is a slightly distorted tetrahedron and two DMF molecules are
coordinated to one of the four Zn(II) ions in the unit. The
permanent porosity of the activated MOF-646 was demonstrated
by measuring the adsorption of Ar gas at 87 K. In conjunction
with the polarized pore wall and the small pore-size distribution
of MOF-646, the H2 adsorption isotherms on the activated mate-
rials were recorded. Despite the relatively high crystal density
(1.19 g cm3), activated MOF-646 shows considerable H2 uptake
of up to 17.5 mg g1 at 77 K and 1 bar. This gravimetric uptake
outperforms PCPs with Zn4O SBUs except for IRMOF-11 ([Zn4O
(4,5,9,10-tetrahydropyrene-2,7-dicarboxylate)3]n), whose nets
are interwoven.138 More importantly, the H2 uptake in the volu-
metric unit (20.9 g l1) approaches that of high-performance
PCPs having Cu(II) open metal sites (2224 g l1 for HKUST-1,
PCN-14, and SNU-5)139141 measured under similar conditions.
Furthermore, the qst curve is nearly flat (6.87.8 kJ mol1)
throughout the loading range examined, which is in sharp con-
trast to many PCPs, especially those possessing open metal sites.
From these results, it is presumed that the internally polarized
azulene backbone is advantageous to stabilize H2 molecules in
the framework.
Unfortunately, at room temperature, the storage capacity of
H2 is low because of the weak van der Waals forces between the
adsorbate and the adsorbent. At present, the highest value of
H2 storage capacity is shown by the three-dimensional PCP
[Be12(OH)12(1,3,5-btb)]n (2.3 wt% (total uptake) at 298 K
and 95 bar).142 Promising results of higher H2 uptake at
room temperature have been obtained by Yangs group and
have been ascribed to the so-called spillover effect.143,144 The
composites formed by a physical mixture of MOF-5, a catalyst
containing 5 wt% Pt supported on activated carbon (AC/Pt)
and, or without, sucrose (bridging material), display a H2
uptake of 1.6 and 3 wt% for the unbridged and bridged
composites, respectively, at room temperature. However,
another group carefully re-examined the H2 uptake of these
composites at room temperature and showed that no
enhanced H2 uptake in the composites was found compared
MetalOrganic Frameworks 57
with the initial MOF-5.145 Similar trends have been observed
in carbonaceous systems.146,147 Hence, the spillover effect
remains in doubt.
5.03.5.1.4 Water
The applications of heat transformation through the use of
water adsorption, such as thermally driven adsorption heat
pumps (AHPs) and desiccant cooling systems (DCSs), have
attracted attention as promising energy-saving materials for
efficient air conditioning. The energy efficiency and working
humidity range of AHP and DCS strongly depend on the
properties of the water adsorbent. Therefore, it is important
to prepare adsorbents showing a high degree of water adsorp-
tion and to control their adsorption pressure. Currently, silica
gels or zeolites are used as adsorbents for AHP and DCS.
New MIL-100-type PCPs using several counteranions were
synthesized to achieve this purpose. MIL-100, {[Cr3FO
(1,3,5-btc)2] xH2O}n (x  28), has two types of giant
cages with inner diameter sizes of 25 and 30 A and an
extremely large specific surface area (the Langmuir surface
area 3100 m2 g1).67 Interestingly, the free counteranions
are accommodated in the pores; therefore, they can be replaced
with other counteranions using different inorganic acids in the
synthesis instead of hydrofluoric acid. The anion-exchanged
PCPs, [Cr3ClO(1,3,5-btc)2]n and [Cr3(SO4)0.5O(1,3,5-btc)2]n,
obtained using hydrochloric acid and sulfuric acid, respec-
tively, show XRD patterns similar to the reported MIL-100. It
is worth noting that an exceptionally large amount of water
(over 0.5 g g1) can be adsorbed and desorbed on each PCP in
the middle relative pressure range from 0.22 to 0.58. The
durability test, which involves recycling the water adsorption
and desorption processes, indicates that the amount of adsorp-
tion does not change even after 2000 adsorption/desorption
cycles. The water uptake pressure can be tuned by the counter-
anions incorporated in the structure, because each anion has a
different hydration energy.148
5.03.5.1.5 Acetylene
Acetylene, C2H2, is a key starting material for the synthesis of
various organic compounds, such as 1,4-butanediol, which is
widely used in the preparation of polyurethane and polyester
plastics. One issue associated with C2H2 is its storage. Unlike
other gases, C2H2 cannot be stored in a steel cylinder under
high pressure, because it becomes highly explosive when
compressed above 0.2 MPa at room temperature, even in the
absence of oxygen. The current method of storing C2H2
involves dissolution of the gas in acetone or DMF placed in a
steel cylinder along with an absorbent to prevent polymeriza-
tion. However, C2H2 stored by such a method contains volatile
solvents as impurities, and the degree of its purity is not very
high. Hence, high-density storage of highly pure C2H2 at low
pressure by a nonvolatile solid matrix could be a key research
subject for the preparation of advanced organic materials.
The C2H2 adsorption properties were thoroughly
investigated using six isostructural PCPs [M2(L)2(dabco)]n
(M Zn(II) and Cu(II), L 1,4-bdc2, 1,4-ndc2, and
9,10-anthracenedicarboxylate (adc2)).149 The C2H2 ad-
sorption isotherms show that the adsorption amounts for
[Zn2(adc)2(dabco)]n and [Zn2(1,4-ndc)2 (dabco)]n at 298 K
and 1 atm (101 and 106 cm3(STP) g1, respectively) are higher
than those for other conventional porous materials such as
58 MetalOrganic Frameworks
carbon molecular sieves (45 cm3(STP) g1),150 mesoporous
silica SBA-15 (36 cm3(STP) g1),151 and type-A zeolite
(98 cm3(STP) g1).152 In particular, the adsorption amount
for [Zn2(adc)2(dabco)]n is 84 cm3(STP) g1 even at 323 K.
The storage density of C2H2 in [Zn2(adc)2(dabco)]n at 298 K
is 0.135 g ml1, which is equivalent to the density of C2H2 at
13 MPa at 298 K and is 65 times larger than the value of the
compression limit for the safe storage of C2H2 at room
temperature, 0.20 MPa: 0.002 g ml1. The greater uptake
capacity of [Zn2(adc)2(dabco)]n under ambient conditions
is most likely attributed not only to the micropore effect but
also to C2H2anthracene interactions, because the polariz-
ability of anthracene is known to be a 25.4  1024 A3,
which is much larger than that of benzene or naphthalene.
A series of isostructural PCPs [M2(dot)]n (M Co(II), Mn
(II), Mg(II), and Zn(II)) with coordinatively unsaturated metal
sites was examined for their differential recognition of C2H2
molecules and realized to be high C2H2 storage materials153;
of these, Co(II) has the strongest interactions with C2H2 and a
very high enthalpy of adsorption (50.1 kJ mol1) at low cov-
erage. Such a strong interaction makes [Co2(dot)]n the highest
volumetric C2H2 storage material reported so far, with a capac-
ity of up to 230 cm3(STP) cm3 at 295 K and 1 atm. This high
capacity maximized the safe C2H2 storage density of adsorbed
C2H2 in bulk material (up to 0.27 g cm3). High-resolution
neutron powder diffraction studies on the 0.54 C2D2 per Co
(II)-loaded [Co2(dot)]n sample confirm that the open Co(II)
sites are the first preferential C2H2 storage sites (the distance
between a C2D2 molecule and the closest Co(II) atom is about
2.65 A) that make a major contribution to the high volumetric
C2H2 storage capacity. Two PCPs, {[Cu2(ebtc)(H2O)2] the acidic hydrogen atoms at both ends of C2H2 (pKa 25).
8H2O DMF DMSO}n (ebtc4 1,10 -ethynebenzene-3,30 ,5,
0
5 -tetracarboxylate) and HKUST-1, with coordinatively unsatu-
rated Cu(II) sites, also exhibit very high C2H2 storage of 160
and 201 cm3(STP) g1, respectively, at 295 K and 1 atm.154,155
5.03.5.2 Separation
5.03.5.2.1 Carbon dioxide
For the practical separation application of PCPs in CO2 capture,
the evaluation of gas mixtures containing CO2 is essential. The
breakthrough experiment and gas chromatographic separation
are simple and straightforward in the evaluation of separation.
Some breakthrough experiments on CO2 separation in PCPs
have been reported; these gave qualitative evaluations in most
cases. For example, several ZIFs are feasible for the separation of
CO2 from other gases (CO, CH4, and N2) and have longer
retention times for CO2 than BPL carbon under the same
conditions.106,156,157 Gas chromatography is an alternative
experimental method used for the evaluation of gas separation
in an adsorbent. Variable-temperature pulse-gas chromatogra-
phy experiments were performed with a complex gas mixture
(N2, CH4, CO2, and C2H2) for the PCPs {A[Cu3(m3-OH)(m3-4-
carboxypyrazolato)3]}n (A NH4, Et3NH), which is based on
an anionic three-dimensional framework built up of trinuclear
Cu3(m3-OH) clusters and m3-4-carboxypyrazolato bridges, to
examine these PCPs possible utility for gas-separation
purposes.158 The results show that these PCPs have significant
interactions with CO2 and C2H2, whereas their interactions with
N2 and CH4 are negligible. At low temperatures (T < 286 K),
C2H2 is more tightly retained by these PCPs framework than
is CO2. At higher temperatures (T > 286 K), the reverse situation
is found, with C2H2 being eluted before CO2. This is the first
example to show the utility of PCPs for the separation process,
thus highlighting the importance of the adsorption temperature
in the efficiency of the separation.
5.03.5.2.2 Acetylene
As mentioned in Section 5.03.5.1, high-density storage of
C2H2 gas is an important issue. Another issue in C2H2 produc-
tion and use is its separation from other gases, which includes
CO2, a common impurity present in many industrial pro-
cesses. Zeolites and activated carbons are known to be capable
of accommodating large amounts of both gases. No relevant
difference in capacity has been observed, because these mole-
cules are similar to one another in equilibrium adsorption
parameters, related physicochemical properties, and molecular
size and shape. A marked difference in the C2H2 and CO2
adsorption isotherms is observed in CPL-1.91,159 The adsorp-
tion isotherms of C2H2 show a steep rise in the very low pressure
region and reach saturation (42 cm3(STP) g1), whereas those
of CO2 show a gradual adsorption (Figure 11(a)11(c)). The
degree of saturation of C2H2 corresponds to one molecule per
unit pore. The maximum ratio of adsorbed C2H2 relative to that
of CO2 is 26.0 at 1.1 kPa and 270 K, indicating that CPL-1
accommodates C2H2 in preference to CO2. The isosteric heat
at the fractional filling ratio of 0.2 of C2H2, 42.5 kJ mol1, is
higher than that of CO2, 31.9 kJ mol1. This result indicates a
higher enhancement in the interaction of C2H2 with CPL-1 than
that of CO2. Such a strong interaction with C2H2 is attributed to
Such an acidic property of C2H2 makes it possible to interact
with the oxygen atoms of the carboxylate parts, which was
checked directly using synchrotron powder XRD measurements
on CPL-1 with adsorbed C2H2 (Figure 11(d)). The density of
adsorbed C2H2 was estimated to be 0.434 g cm3. This density
is equivalent to that of an extrapolated state of C2H2 of 41 MPa
at room temperature, and is 200 times larger than the value of
the compression limit for the safe use of C2H2 at room temper-
ature and 0.20 MPa (0.002 g cm3). Selective and high-density
C2H2 adsorption is observed in other PCPs, [M(HCOO)2]n
(M Mg(II) and Mn(II)).160 The amount of adsorbed C2H2 at
298 K is above 60 cm3(STP) g1 for [Mg(HCOO)2]n and
50 cm3(STP) g1 for [Mn(HCOO)2]n, whereas the uptake of
other gases, including CO2, at 298 K decreases substantially,
especially at low pressure. The x-ray crystal structure of C2H2-
adsorbed [M(HCOO)2]n shows that C2H2 molecules are held in
the zigzag channels of the frameworks through only van der
Waals interactions, in contrast to the case with CPL-1. Therefore,
such discrimination is presumably due to the slightly larger size
of C2H2 (5.5 A) relative to CO2 (5.3 A), which provides the
former with more effective van der Waals interactions with the
framework walls.
C2H2 is known to form metalp complexes with Ag(I) and
Cu(I) with low coordination numbers. The monovalent coin-
age metals Ag(I) and Cu(I) are attractive UMC candidates as
soft Lewis acids. They can adopt diverse coordination geome-
tries, such as linear, trigonal planar, and tetrahedral. In the
two- and three-coordinated forms, Ag(I) and Cu(I) retain
large unoccupied spaces that are readily accessible to other soft
 MetalOrganic Frameworks 59

1.4

1.2

(molecules per unit pore)

Amount adsorbed
1.0

0.8

0.6

0.4

0.2

0.0
0 20 40 60 80 100
(a) Pressure (kPa)

1.4 1.4

1.2 1.2
(molecules per unit pore)

(molecules per unit pore)

Amount adsorbed

Amount adsorbed
1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0 20 40 60 80 100 0 20 40 60 80 100
(b) Pressure (kPa) (c) Pressure (kPa)

b
b

(d)
a a

Figure 11 (ac) Adsorption isotherms for C2H2 (red) and CO2 (blue) on CPL-1 at (a) 270 K, (b) 300 K, and (c) 310 K. (d) Crystal structure of
CPL-1 with C2H2 at 170 K. Reproduced from Matsuda, R.; Kitaura, R.; Kitagawa, S.; Kubota, Y.; Belosludov, R. V.; Kobayashi, T. C.; Sakamoto, H.;
Chiba, T.; Takata, M.; Kawazoe, Y.; Mita, Y. Nature 2005, 436, 238241, with permission of the copyright holders.

Lewis base molecules. Indeed, Ag(I)- or Cu(I)-exchanged
zeolites have shown high activity toward unsaturated
hydrocarbons.161 However, immobilizing Ag(I) or Cu(I) UMC
on the pore surface of PCP is very difficult, because Ag(I) and Cu
(I) tend to have energetically stable tetrahedral coordination
geometry rather than linear or trigonal planar geometries and
their PCPs are very unstable and sensitive to light, heat, air, and
moisture, among others. The concept of a binary metalazolate
framework can be used to solve these problems. Because of their
strong metalazolate coordination bonds, binary metalazolate
frameworks have demonstrated very high thermal and chemical
stabilities. More importantly, the coordination numbers of the
monovalent metal centers are always equal to, or lower than, the
number of nitrogens in the ligand, because each nitrogen donor
can only coordinate to one metal center. Three-dimensional
PCPs, a-[M2(bpz)]n (M Ag(I) and Cu(I), bpz2 3,30 ,5,50 -
tetramethyl-4,40 -bipyrazolate), show reversible type I isotherms
for C2H2 and CO2 at 298 K, indicating that no chemical adsorp-
tion process occurs during the measurements.162,163 At 1 atm,
the C2H2 uptakes are 44 and 58 cm3(STP) g1 for M Ag(I) and
Cu(I), respectively, and the CO2 uptakes are 30 and 39 cm3(STP)
g1, corresponding to C2H2/CO2 uptake ratios of 1.46 and
1.48. These PCPs exhibit very large isosteric heat differences
between C2H2 and CO2. Regarding the electronegativity of the
60 MetalOrganic Frameworks

framework components, the hydrogen-bonding ability of the bpy

pore surface of a-[M2(bpz)]n is much weaker than that of
metalcarboxylate PCPs. Enlightened by the crystal structures of
arene inclusion forms, it is concluded that the weak metalp
interactions formed between coordinatively UMCs and C2H2
molecules are likely to be responsible for the discrimination
Zn TCNQ-TCNQ
effect of C2H2 over CO2.

5.03.5.2.3 Oxygen
Generally, it is difficult to achieve selectivity between mole-
(a) (b)
cules of similar shapes. In particular, focusing on oxygen, O2,
as a target adsorbate molecule presents some difficulties
because N2 and Ar have a similar molecular shape or size,
respectively. In addition, CO2 is a good adsorbate for micro-
porous and mesoporous materials because of its small kinetic
diameter and high affinity for the adsorbent. It is, therefore,
typically preferred over other guests, and there are only a few
examples in which the adsorbent shows no uptake of CO2 even
though it possesses porosity for O2. The selective adsorption of
O2 over CO2 is one of the most crucial challenges for any 5.05
porous material. [Zn(TCNQTCNQ)(4,40 -bpy)]n has a flexible
three-dimensional structure constructed from Zn(II) centers
and two ligands, 4,40 -bpy and the dimeric TCNQ dianion
(c)
(Figure 12(a)12(c)), and shows a closed/open structural
transformation accompanying the adsorption/desorption
NO
process.164 The electron-rich pore surfaces derived from the
Amount adsorbed (cm3 g-1)

dimeric TCNQ dianion should have a high affinity for
electron-accepting molecules. This tendency can be observed
in the case of gas molecules. The adsorption isotherms of
several types of gas molecules indicate that this PCP shows
the adsorption characteristic of gate-type adsorption behavior
only in the cases of O2 and NO.165 The saturation-adsorbed
amounts of O2 (268 cm3(STP) g1) and NO (322 cm3(STP)
g1) correspond to 7.5 and 9 molecules per formula, respec-
tively. In contrast, no sharp uptake above one molecule per
formula is observed with any of the other gas molecules (N2,
CO, CO2, C2H2, or Ar). To examine the origin of the selective
adsorption, IR and Raman spectra were measured under an O2
atmosphere. The IR spectra indicate that the electron density
on the TCNQ dianion decreases as O2 is adsorbed, implying a
CT interaction between the framework and O2 molecule. The
Raman spectra show a large red shift of the sharp n(OO)
band by  100 cm1, that is, a decrease in the bond order of
the adsorbed O2, which is commonly caused by electron accep-
tance into the anti-bonding p*-orbitals. Hence, the observed
n(OO) band suggests that adsorbed O2 molecules have a
partial negative charge and that there is no redox reaction
with one-electron transfer. This is direct evidence for the exis-
tence of a CT interaction between the framework and the O2
molecules. Therefore, it is concluded that this specific adsorp-
tion selectivity can be ascribed to the closed/open structural
transformation triggered by the CT interaction.
Other PCPs also can achieve the selective adsorption of O2
over other gases. The three-dimensional Yb(III) PCP, {[Yb4
(4,40 ,400 -S-triazine-2,4,6-triyl-tribenzoate)8/3(SO4)2(m4-H2O)] primary structure, featuring one-dimensional channels with
3H2O 10DMSO}n (PCN-17), based on a novel square-planar
Yb4(m4-H2O) SBU, possesses exceptionally high thermal stabil-
ity up to 753 K while maintaining permanent porosity (BET
surface area: 820 m2 g1).166 The pore size is around 3.5 A,
which results in the selective adsorption of O2 over N2 and
300
O2
200
CO2
100 C2H2
Ar
N2
CO
0
0 0.2 0.4 0.6 0.8 1
(d) Relative pressure, PIP0
Figure 12 (a) Coordination environment of the Zn(II) ion, (b) TCNQ
dimer (green) connected to four one-dimensional chains of Zn(II) and
4,40 -bpy (gray), and (c) benzene arranged in the cage of the undulating
channel in {[Zn(TCNQTCNQ)(4,40 -bpy)] 1.5benzene}n. (d) Adsorption
isotherms of several gas molecules (N2, O2, and CO at 77 K; Ar at 87 K;
NO at 121 K; CO2 and C2H2 at 193 K) on [Zn(TCNQTCNQ)(4,40 -bpy)]n.
Reproduced from Shimomura, S.; Horike, S.; Matsuda, R.; Kitagawa, S. J.
Am. Chem. Soc. 2007, 129, 1099010991; Shimomura, S.; Higuchi, M.;
Matsuda, R.; Yoneda, K.; Hijikata, Y.; Kubota, Y.; Mita, Y.; Kim, J.; Takata,
M.; Kitagawa, S. Nat. Chem. 2010, 2, 633637, with permission of the
copyright holders.
CO using a molecular sieving effect. Two three-dimensional
PCPs of formulas [Cu(BDTri)(DMF)]n and [Cu(BDTri)(DEF)]n
(H2BDTri 1,4-benzenedi(1H-1,2,3-triazole)) have the same
the bridging DMF or DEF molecules pointing into the cavity.167
The O2 adsorption isotherms show a two-step adsorption
behavior associated with a permanent microporosity and a
pore-opening process. In the case of N2 adsorption, only [Cu
(BDTri)(DMF)]n exhibits a two-step adsorption isotherm,

whereas [Cu(BDTri)(DEF)]n does not present any pore opening.
The selectivity of O2 over N2 reaches 5.9:1 (at 168 mbar) and
13.5:1 (at 200 mbar) for [Cu(BDTri)(DMF)]n and [Cu(BDTri)
(DEF)]n, respectively, the latter value representing the highest
value yet observed for a microporous material under similar
conditions.
5.03.5.3 Catalysis
It is well known that porous materials are useful for heteroge-
neous catalysis. In contrast to conventional microporous mate-
rials such as zeolites and activated carbons, PCPs have pore
surfaces that can be rationally designed and functionalized by
their constituents. Hence, functionalized PCPs bearing specific
interaction sites in/on micropores have been envisaged for
useful catalytic applications in organic and polymer syntheses,
offering several advantages: (1) reaction acceleration, (2) reg-
ulated and controlled reactions (stereo- and regiocontrols),
and (3) substrate specificity.168 All of these properties are
similar to those of enzyme reactions in biological systems.
5.03.5.3.1 Lewis acid catalysis
The two-dimensional PCP {[Cd(4,40 -bpy)2(H2O)2] 2NO3
4H2O}n is the first example to show catalytic properties for
cyanosilylation of aldehydes.169 Experimental data in the case
of cyanosilylation of imines, which is also performed on the
same compound, lead to the conclusion that the present het-
erogeneous reaction should involve the selective activation
of the imino nitrogen by the weak Lewis acid Cd(II) center.170
The active sites exist mainly around the surface of the crystal.
In this PCP, the NO3 anions exist in a coordination-free state.
This situation contributes to an increase in the Lewis acidity of
the Cd(II) centers.
5.03.5.3.2 Brnsted acid catalysis
It is very difficult to introduce Brnsted acid sites into PCP frame-
works because candidate low pKa moieties are deprotonated by
the unavoidable presence of Lewis basic ligands. An alternative
approach to the synthesis of PCPs with Brnsted acid sites is to
introduce them after the construction of the framework. A sus-
pension of the amino acid-based PCP [Ni(L-asp)(4,40 -bpy)0.5]n
(L-asp2 L-aspartate), which has a three-dimensional porous
HOOC
O O
P
O OH
HOOC
(R)-H4L4
Figure 13 Structure of chiral organic ligands, (a) (R)-H4L4, with both carboxylic acid and phosphoric acid sites, and (b) (R)-HL5.
MetalOrganic Frameworks 61
framework (3.8 4.7 A2) consisting of chiral [Ni(L-asp)]n layers
and 4,40 -bpy pillars,171 in anhydrous Et2O upon treatment with
HCl leads to the isolation of the protonated phase {[Ni(L-asp)
(4,40 -bpy)0.5] 0.9HCl 0.5MeOH}n.172 Powder XRD confirms the
crystallinity of the protonated phase, with only minor differences
observed upon comparison with the parent material, confirming
comparable framework connectivities. The test reaction exam-
ined was the methanolysis of rac-propylene oxide expected to
yield 2-methoxy-1-propanol and 1-methoxy-2-propanol. The
protonated phase is active for this conversion with catalyst activity
limited by its effectively nonporous nature. The derivative, {[Cu
(L-asp)(bpethe)0.5] HCl H2O}n, exhibits a significant increase in
catalytic activity compared with the former, as anticipated upon
increasing the pore dimensions.
The direct incorporation of Brnsted acid sites into PCP
frameworks succeeded in using organic ligand (R)-L4 with both
carboxylic acid and phosphoric acid sites (Figure 13(a)).173 The
resulting PCP {[Cu2((R)-L4)(H2O)2] 21DMF 12H2O}n has a
three-dimensional network of the hexagonal PtS topology with
protonated phosphoric acid sites and an enormous void
space, comprising 80.8% of the total volume. In the Brnsted
acid-catalyzed asymmetric FriedelCrafts reactions between
indole and imines, this PCP gives the major enantiomers of the
opposite chirality to those produced by the corresponding
homogeneous catalyst. This flip in product handedness results
from the chiral environment of the channels, similar to enzy-
matic catalysis in which the product stereoselectivity is deter-
mined by the enzyme pocket.
5.03.5.3.3 Base catalysis
The two-dimensional homochiral PCP {[Zn3O(L5)6]
2H3O 12H2O}n has been synthesized using the enantiopure
chiral bridging ligand L5, as illustrated in Figure 13(b).174
In the crystal, the trinuclear Zn3O units are bridged by L5,
generating two-dimensional layers consisting of large edge-
sharing chair-shaped hexagons with a trinuclear unit at each
corner. The two-dimensional layers stack to form large
chiral one-dimensional triangular channels with a side length
of 13.4 A. There are six pyridyl groups per trinuclear unit in
the structure. Interestingly, three of them accidentally
remain in the channel without any interactions with the frame-
work and, therefore, provide unique opportunities in base
catalysis. The reaction of an ester with a large excess of racemic
COOH
O O
HOOC NH
O
N
COOH
(R)-HL5
62 MetalOrganic Frameworks
1-phenyl-2-propanol in the presence of the homochiral reac-
tion field or its enantiomorphic form produces the correspond-
ing esters with  8% enantiomeric excess.
Reaction of Cd(NO3)2 4H2O with a tridentate amide ligand
affords the three-dimensional PCP {[Cd(NO3)2(4-btapa)2] nomenon. The gate pressure depends on the adsorbate, which
6H2O 2DMF}n (4-btapa 1,3,5-benzenetricarboxylic acid tris
[N-(4-pyridyl)amide]).175 The amide groups occur on the sur-
face of channels with dimensions of 7.3  4.7 A2. The highly
ordered amide groups in the channels enable selective hetero-
geneous Lewis base catalytic reactions. A Knoevenagel conden-
sation reaction is catalyzed by this PCP in good selectivity and
yield. The selectivity depends on the relationship between the
size of the reactants and the pore window of the PCP. This solid
catalyst maintains its crystalline framework after the reaction
and is easily recycled.
5.03.5.4 Reaction Vessels in Polymer Syntheses
Considering the attractive features of the channels of PCPs, the
utilization of PCPs should be of key importance for the creation
of unique nanosized reaction vessels. In particular, the use of the
PCP channels as a vessel for polymerization is an attractive idea
for many reasons, and would not only allow multilevel control
of polymerization (e.g., control of stereochemistry, regiochem-
istry, molecular weight, and helicity), but would also provide
well-defined nanostructures and nanohybrids to enable the
fabrication of new-generation materials.22
Three-dimensional PCPs, [M2(1,4-bdc)2(dabco)]n (M Zn
(II) and Cu(II)), exhibit the topotactic selective radical polymer-
ization of divinylbenzenes (DVBs) inside their one-dimensional
channels.176 Polymerization of p-DVB in the channels of the Zn
(II) PCP affords the linear polymer, in which only one vinyl
group of p-DVB is selectively polymerized, while hardly any
undesired kink structures (e.g., branching and/or bonding
through both vinyl groups) are formed. Linear polymers bearing
pendant vinyl groups in their backbone are attractive in view of
the potential reactivity of the olefinic double bonds and for
application as polymer precursors of functional materials. In
contrast, no trace of polymeric product was observed in the Cu
(II) PCP having almost the same structure as the Zn(II) PCP.
PCPs containing Zn(II) often show similar flexible and dynamic
properties upon guest inclusion. Therefore, the flexibility of the
Zn(II) PCP allows adsorption of the relatively large p-DVB mol-
ecules by expansion of the host structure, such that the mono-
mer molecules are arranged adjacent to each other, which is
essential for polymerization.
The structural diversity of PCPs makes them feasible as
templates for preparing nanoporous carbon materials. Nano-
porous carbons were prepared by polymerizing and then car-
bonizing a carbon precursor (furfuryl alcohol) accommodated
in a MOF-5 template.177,178
5.03.5.5 Framework Dynamics
{[Cu(BF4)2(4,40 -bpy)(H2O)2] 4,40 -bpy}n forms noninterpene-
trated two-dimensional layers with grids of 7.7  11.6 A2.179
The dehydration of this compound on vacuum heating affords
another two-dimensional framework, [Cu(BF4)2(4,40 -bpy)2]n.180
Micropores for the accommodation of guest molecules are
not observed in this dehydrated compound because the
neighboring layers obstruct the pores. However, the adsorp-
tion of N2, Ar, CH4, and CO2 suddenly begins at a definite
relative pressure, despite there being almost zero adsorption
below this threshold pressure; this is known as a gate phe-
implies the possibility of separating mixed gases in this com-
pound. Structural analysis of the compound adsorbing CO2
gas shows that the interlayer distance changes from 0.46 to
0.68 nm on CO2 adsorption. Although details of the interlayer
interactions have not yet been reported, weak interlayer inter-
actions (e.g., hydrogen bonds, pp stacking interactions, and
van der Waals interactions) may contribute to the dynamic
behavior of the porous framework. Recently, a two-
dimensional framework analogous to [Cu(BF4)2(4,40 -bpy)2]n
was synthesized, in which the PCP [Cu(OTf)2(4,40 -bpy)2]n
(OTf trifluoromethanesulfonate) has two-dimensional
square-grid sheets with axially coordinated OTf anions.181
The stacking of these sheets creates free space (17.9%). The N2
adsorption isotherm at 77 K shows that the adsorption branch
is characterized by a definite double step, and a marked hys-
teresis is present. The mass of N2 adsorbed at the first uptake is
in good agreement with the void volume of [Cu(OTf)2(4,40 -
bpy)2]n calculated from the crystal structure. The second
uptake may originate from an increase in the pore volume
on expansion of the interlayer regions, which was verified by
synchrotron XRD measurements. Compared with [Cu
(BF4)2(4,40 -bpy)2]n, which shows a single-step isotherm, [Cu
(OTf)2(4,40 -bpy)2]n exhibits the double-step isotherm derived
from the presence of open pores because of the pillaring role
of the OTf anions that fix neighboring sheets at an appropri-
ate position and prevent the clogging of the pores.
{[Cu(dhbc)2(4,40 -bpy)] H2O}n (dhbc 2,5-dihydroxy-
benzoate) is a flexible and dynamic PCP based on
interdigitation.100 The Cu(II) ions are connected by 4,40 -bpy
ligands to produce one-dimensional linear chains and are
linked by dhbc ligands to give a two-dimensional sheet
motif. This motif has interlocking ridges and hollows con-
structed by the dhbc benzene planes in an upright fashion.
The distance of 3.4 A between the planes of the nearest-
neighbor dhbc ligands indicates the presence of pp stacking
interactions, and the motifs are mutually interdigitated to cre-
ate one-dimensional channels along the a-axis with a cross-
sectional area of 3.6  4.2 A2. This PCP undergoes a marked
crystal transformation triggered by the desorption of included
H2O and guest adsorption. A detailed structural investigation
by synchrotron powder XRD measurements showed a change
in cell parameters on dehydration. This change corresponds to
a cell volume contraction of 27%, and is attributed to a glide
motion of the two p stack ring moieties (dhbc), which results
in a decrease in the channel cross-sectional area. Characteristic
hysteretic adsorption isotherms that have different gate-
opening and gate-closing pressures for various gases (e.g.,
CO2, CH4, O2, and N2) are observed in the dehydrated com-
pound. This may be attributed to expansion and contraction of
the crystal structure triggered by gas adsorption and desorp-
tion, respectively.
Two types of framework motions responding to guest
comings and goings can also be introduced into PCPs.
The two-dimensional PCP {[Cu2(tci)(OH)(H2O)3] 1.5H2O}n
(H3tci tris(2-carboxyethyl)isocyanurate) undergoes a reversible

single-crystal-to-single-crystal transformation to a three-
dimensional framework of [Cu2(tci)(OH)]n upon dehydra-
tion and rehydration, associated with the mutual sliding of
the two-dimensional sheets and contracting of the spaces
between the sheets.182 This structural bistability also results
in drastic changes in the optical and magnetic properties.
An unprecedented selective adsorption phenomenon has
been achieved by a flexible PCP, CPL-11.95 CPL-11 adsorbs
large guest molecules at high temperature but does not adsorb
any guest molecules at low temperature. Furthermore, cooling
the framework leads to a decrease in the amount adsorbed,
which is the opposite of the tendency in usual porous mate-
rials. Such adsorption behaviors are strongly associated with
the degree of structural uniformity. In CPL-11, the lattice strain
increases as temperature decreases, which was confirmed by
the measurements of the full width at half maximum of the
XRD peaks.
5.03.5.6 Framework Stability
Generally, PCPs based on neutral pyridine-type bridging
ligands are thermally less stable than those based on
carboxylate-based ligands. On the other hand, many PCPs
with carboxylate-based bridging ligands are very sensitive to
atmospheric water and therefore decompose to form nonpor-
ous compounds. For example, MOF-5 easily decomposes upon
exposure to air.183 Consequently, the preparation, handling,
and utilization of such materials require moisture-free condi-
tions that limit their industrial applications. To overcome this
weak point, imidazolate-bridged PCPs (ZIFs) are used that
exhibit high thermal stability and remarkable chemical resis-
tance in boiling alkaline water and other solvents.104,106 Two
plausible explanations can be proposed: (1) the hydrophobic
pore and surface structure of ZIFs prevent water molecules
from attacking ZnN4 units and decomposing the framework
and (2) the bond between imidazolate and Zn(II) is among the
most stable for N-donor ligands.184
5.03.5.7 Combined Functions of Porosities and
Magnetic Properties
Porous coordination polymer magnets (PCPMs) are optimal
compounds for achieving multifunctionality responding to
external stimuli, such as a magnetic field and guest adsorption/
desorption, because coordination polymers have metal ions as
magnetic centers and have flexible characteristics inducing mag-
netic perturbations. In general, PCPs need long bridging ligands,
while a strong magnetic interaction is achieved using short
bridging ligands. Such a difference in structural demands
makes it difficult for coordination polymers simultaneously to
have a high porosity and a strong magnetic interaction.
The use of bridging ligands with a magnetic center is an
effective method to overcome the above problem. Bridging
ligands with a magnetic center have the characteristics of
the bridging of metal centers by coordination bonds and also
enhancement of the magnetic interactions between the metal
centers. {[Cu3(ptmtc)2(py)6(H2O)(EtOH)2] 6H2O 11EtOH}n
(H3ptmtc 4,40 ,400 -tricarboxydodecachlorotriphenyl)methyl rad-
ical, py pyridine) is a two-dimensional PCPM with one-
dimensional channels oriented along the [00 1] and [10 0]
MetalOrganic Frameworks 63
directions, which incorporate guest H2O and EtOH molecules.185
The ptmtc3 bridging ligand has the organic radical part
(a methyl radical, S 1/2) centered in the molecular structure,
which enables long-range ferrimagnetic ordering below 2 K.
Interestingly, after removal of all of the guest molecules, the
resulting desolvated compound, [Cu3(ptmtc)2(py)6]n, loses its
crystallinity and behaves as a paramagnet in the measured tem-
perature range. The crystallinity and magnetic properties are
recovered when the desolvated compound is exposed to MeOH
or EtOH vapor.
Metalloligands, which are metal complexes containing two or
more Lewis base sites that are able to bridge with other metal
ions, can also act as building blocks for developing PCPMs. The
hexacyanometallate and octacyanometallate anions, [M(CN)6]n
and [M(CN)8]n, respectively, are suitable ligands because they
provide a relatively long (10 A) inner pore, and form mono-
dentate cyanide bridges for effective magnetic interactions and
structural flexibility. The three-dimensional PCPM ferromagnet,
{[Ni(dipn)]2[Ni(dipn)(H2O)][Fe(CN6)]2 11H2O}n (dipn N,
N-di(3-aminopropyl)amine), shows a significant magnetic
conversion associated with a reversible gaseous H2O-induced
crystalline-to-amorphous-like phase transformation.186 This
compound has one-dimensional rectangular channels with
dimensions of 3.5 8.7 A2 and a solvent-accessible void of
29%. A long-range ferromagnetic ordering (Tc 8.5 K) is
observed in the crystalline phase, {[Ni(dipn)]2[Ni(dipn)(H2O)]
[Fe(CN6)]2 11H2O}n with H2O guests. Once the H2O guests
are partially removed from the channels, the resulting desolvated
compound, {[Ni(dipn)]2[Ni(dipn)(H2O)][Fe(CN6)]2 H2O}n,
shows an amorphous-like XRD pattern and exhibits an almost
paramagnetic behavior, which indicates that the ferromagnetic
interaction is weakened, and the magnetic domain is moderately
fragmented while retaining short-range ordering upon dehy-
dration. A change in magnetic properties corresponding to the
capture and release of coordinated and/or hydrogen-bonded
guests has been reported in other PCPMs with [M(CN)6]n or
[M(CN)8]n.187189
Multicarboxylate ligands are used in the synthesis of
PCPMs, although it is very difficult to control the coordination
structures. [Mn3(HCOO)6]n, a three-dimensional flexible
porous diamondoid framework based on Mn(II)-centered
MnMn4 tetrahedral nodes, exhibits a wide spectrum of guest-
inclusion behavior and long-range magnetic ordering with
guest-modulated critical temperature (Tc 4.89.7 K).190,191
Spin-crossover (SCO) centers represent a convenient phys-
ically addressable molecular-scale magnetic switch. Much cur-
rent attention is directed toward the coordinative link of SCO
sites, motivated largely by the supposition that a physical
connection of this type increases cooperative communication
between SCO centers, which would be applicable in develop-
ing bistable materials for binary switching applications.
Recent work has shown that such SCO coordination polymers
can incorporate an additional level of functionality associated
with their often porous nature. Guest adsorption/desorption
in such systems, and the associated perturbation of the Fe(II)
coordination environment and framework structure, provides
a unique avenue for investigating SCO phenomena in the
solid state.192,193 The PCP {[Fe2(NCS)4(azpy)4] EtOH}n dis-
plays a broad half-SCO transition that depends on guest
inclusion.192 Desorption of the unbound EtOH guests from
64 MetalOrganic Frameworks

this interpenetrated compound leads to a reversible single-

Pt
crystal-to-single-crystal transformation in which the one- Fe
dimensional channels partially collapse. In contrast to the
EtOH-loaded compound, the desolvated compound retains its
high-spin (HS) state down to a low temperature, a feature that
may reflect a weakening of the ligand field with distortion of the
coordination geometries following guest removal. Adsorption
of other alcohol guests switches back on the SCO sites, with c
subtle differences in behavior attributed to the steric influence of b
the guest on the local Fe(II) coordination geometry. a
A new approach to guest-responsive SCO was recently (a) (b)
adopted using Hofmann-type three-dimensional SCO PCPs, 1-CS2 1-HS
[Fe(pyz)M(II)(CN)4]n (M(II) Ni, Pd, Pt), as a platform.194
These PCPs display cooperative magnetic and chromatic +CS2
thermal- and light-induced spin transitions in the region of
room temperature. In particular, [Fe(pyz)Pt(CN)4]n displays a
first-order spin transition with an approximately 25-K wide
hysteresis (critical temperature: Tc# 285 K and Tc" 309 K).
The crystal structure of the dihydrated form {[Fe(pyz)Pt
+DT
(CN)4] 2H2O}n (Figure 14(a) and 14(b)) shows the presence
of two guest-interactive sites, one between the pyz-bridges (site -CS2 -bz
A) and the other between the four-coordinate Pt centers (site -DT
B). Furthermore, the pyz-bridges are provided with rotational
freedom in the framework. These features should be important
for the guest-responsive framework.
The HS and low-spin (LS) states of [Fe(pyz)Pt(CN)4]n are
reversibly induced with coupling of guest adsorption.195 The
adsorption isotherm of benzene using the guest-free LS form +bz
shows a steep rise at a very low P/P0 value (0.05) with
(c) 1-LS 1-bz
saturation accompanied by a color change. The benzene-
including form {[Fe(pyz)Pt(CN)4] benzene}n is paramagnetic Figure 14 (a) The basic cavity structure and (b) projection of the three-
at all temperatures. Interestingly, a complete and relatively dimensional framework of {[Fe(pyz)Pt(CN)4] 2H2O}n. (c) Schematic
rapid conversion from the LS state to the HS state is induced chemical and thermal memory process in [Fe(pyz)Pt(CN)4]n. Reproduced
by the adsorption of benzene at P/P0 0.19, which was mon- from Ohba, M.; Yoneda, K.; Agust, G.; Munoz, M. C.; Gaspar, A. B.; Real,
itored directly using a sample holder specially designed to J. A.; Yamasaki, M.; Ando, H.; Nakao, Y.; Sakaki, S.; Kitagawa, S. Angew.
Chem., Int. Ed. 2009, 48, 47674771, with permission of the copyright
introduce vapor into the SQUID magnetometer. After desorp-
holders.
tion of benzene, the system does not recover the initial LS state
over a period of months after releasing benzene within the
bistable temperature region. This memory function retains
of Pt(II) to Pt(IV) and reduction of the dihalogen to the corre-
information about adsorption of the guest molecules in the
sponding halide to give [Fe(pyz)Pt(CN)4(X)p]n (X Cl (p 1),
form of the spin state, magnetism, color, and structure, infor-
Br (p 1), I (0  p  1)).197 In the particular case of iodine,
mation that can be erased by desorption and cooling
chemisorption takes place from aqueous solution or vapor with
(Figure 14(c)). Other six- or five-membered aromatic mole-
strong stabilization of the LS state and is accompanied by a
cules, such as pyrazine, pyridine, thiophene, pyrrole, and
remarkable change of color from yellow to black-violet, which
furan, or solvents, such as methanol, ethanol, propanol, and
could be useful for sequestering and sensing iodine.
tetrahydrofuran, display a similar behavior to that shown by
benzene at room temperature. In contrast, the guest-free HS
form promptly adsorbs one molecule of CS2 at P/P0 < 0.1 and
5.03.5.8 Conducting PCPs
simultaneously changes to the LS state leading to {[Fe(pyz)Pt
(CN)4] CS2}n. {[Fe(pyz)Pt(CN)4] CS2}n maintains the LS state A few PCPs show relatively high electrical conductivity. A PCP
without a spin transition in the temperature range 2330 K. with high electron conductivity would be applicable to a porous
The framework retains the LS state after desorption of CS2 at electrode for batteries, fuel cells, and capacitors, among others.
298 K as a result of the memory effect. In the case of the CS2 Because PCPs have potentially large surface areas, these com-
guest, it interacted with sites A and B and exceptionally stabi- pounds have high potential as such electrodes with high power
lized the LS state, while other guests stabilized the HS state density. The three-dimensional, metal-based mixed-valence
because of steric hindrance that prevents further contraction of PCP {[Rh4(acam)8I] 6H2O}n (Hacam acetamide) forms a dia-
the clathrates framework. The LS state was also stabilized by mondoid network, in which the paddlewheel-type [Rh2(acam)4]
CH3CN on [Fe(pyz)Ni(CN)4]n.196 units are linked by m4-iodide.198 The electrical conductivity of
Furthermore, [Fe(pyz)Ni(CN)4]n shows the chemisorptive this compound varies by 105 orders of magnitude during
uptake of dihalogen molecules involving associative oxidation dehydrationrehydration cycles of the interstitial H2O molecules.

The isotropic electron spin resonance (ESR) and the moderate
conductivity of the hydrated compound (1.4  103 S cm1 at
room temperature) indicate that the odd electrons are
hopping over the Rh2 units without experiencing much of a
barrier, whereas those in the partially dehydrated compound
({[Rh4(acam)8I] 1.9H2O}n) are localized in some of the Rh2
sites, resulting in an anisotropic ESR spectrum and low conduc-
tivity (order of 108 S cm1). Another type of conducting PCP,
[Cu2(pdt)2]n (pdt2 2,3-pyrazinedithiolate), was synthesized
by the reaction of CuI with the metalloligand {Na[Cu(pdt)2] rather than through the framework.
2H2O}, in which the Cu(I) ion acts as an electron donor and the
[Cu(III)(pdt)2] unit as an electron acceptor.199 In this PCP,
nitrogen atoms of the pyrazine moiety of the [Cu(pdt)2] unit
coordinate to Cu ions, forming a three-dimensional infinite
tetragonal lattice with one-dimensional vacant space of
3.4 3.4 A2. The electronic states are assigned to be [CuII-
CuII(pdt)2]n through analysis of the bond distances and coordi-
nation geometry. This PCP shows relatively high electrical
conductivity (6 104 Scm1 at 300 K). The temperature depen-
dence of the electrical conductivity shows semiconducting behav-
ior, and the activation energy (Ea) was determined to be 0.19 eV.
The cause of such a relatively high electrical conductivity is prob-
ably the charge bistability between [CuICuIII(pdt)2]n and
[CuIICuII(pdt)2]n. After this discovery, an isostructural PCP
[CuNi(pdt)2]n was reported that shows a type I N2 adsorption
isotherm, indicating microporosity (BET and Langmuir surface
areas are 385 and 423 m2 g1, respectively).200 The conductivity
of this compound is not exceptionally high: it was measured as
1  108 Scm1 at room temperature (Ea 0.49 eV). However,
the conductivity was enhanced through partial oxidation of the
framework. Treating a film of this evacuated PCP with a stream of
I2 vapor at 323 K resulted in a 104-fold increase in conductivity to
approximately 1  104 Scm1 (Ea 0.18 eV). Exposure of the
I2-doped PCP film to air for 12 h resulted in a drop in conduc-
tivity to 5 105 Scm1, suggesting that the doping can be
reversed to some extent. Because the conductivity of I2 itself is
not particularly high (1  109 to 1 106 Scm1, depending on
the crystallographic axis) and the amount of I2 involved in the
doping is very small, the conduction must occur primarily
through the PCP framework rather than through I2.
The electrical conduction of the above-mentioned PCPs
occurs through the frameworks. In contrast, electrical conduc-
tion through guest molecules included in the pores occurs in the
three-dimensional PCP {[Zn3(DL-lac)2(pybz)2] 2.5DMF}n, in
which DL-lac2 and pybz are lactate and 4-pyridylbenzoate
anions, respectively.201 The one-dimensional infinite pillars of
[Zn3(DL-lac)2]n are bridged in a square fashion by the pybz
spacers to form a porous framework with the dimensions
11.2 10.2 A2, in which the DMF molecules are housed. The
N2 adsorption isotherm of the desolvated compound at 77 K
displays type I behavior, demonstrating the permanent porosity
(BET and Langmuir surface areas are 763 and 919 m2 g1,
respectively). Interestingly, I2 molecules were successfully
loaded into this PCP by suspending the desolvated crystals in a
solution of I2 in cyclohexane. The resulting I2-included PCP
{[Zn3(DL-lac)2(pybz)2] 3I2}n should have confined and
aligned I2 guests in each pore, taking the calculated density of
I2 in the pore. The potential intermolecular interactions between
I2 and p-electron pybz walls are important, as they allow a
single path for I2 molecules to access and be restricted within
MetalOrganic Frameworks 65
well-regulated narrow limits within the channels, inducing
n ! s* CT. Electrical conductivity measurements on single
crystals of {[Zn3(DL-lac)2(pybz)2] 3I2}n show sk and s? values
of 3.4  103 and 1.7  104 S cm1, respectively. This is
significantly greater than the value for I2 itself. An anisotropy
factor of 21 implies an ordered arrangement mode of I2 mole-
cules and a preferred conductivity along the I2 chain in the
channels. Given the high concentration of I2, a considerable
portion of the conductivity is likely to occur through the I2
The synthesis of a novel proton conductor is one of the most
urgent subjects from the viewpoint of developing energy and
energy conservation technologies, including photovoltaics,
hydrogen storage, and fuel cells. In general, proton-conducting
materials need (1) proton carriers such as H3O or H given by
acid or OH groups and (2) proton-conducting pathways com-
posed of hydrogen-bonding networks. There are three types of
rational design to introduce proton carriers into PCPs. The
simplest method is to introduce them directly as counterions
such as NH4, H3O, and HSO4 into the pores of frameworks
(type I). The second is to put acid groups on frameworks, the
protons being provided from them (type II). The third method
is to incorporate acidic molecules into voids (type III). In an
early study for proton-conductive PCPs, the two-dimensional
PCP [Cu(H2dtoa)]n (H2dtoa2 dithiooxamide anion) with
pores with a diameter of 6 A exhibited a very high proton
conductivity (106 S cm1) at a relative humidity (RH) of 75%
and 300 K.202 Furthermore, the proton conductivity increased
by four orders of magnitude for a RH up to 100%. The mecha-
nism of proton conductivity seems to be similar to that of the
proton exchange membrane Nafion, which contains many clus-
ters of H2O molecules in the pores of the polymer. This PCP also
includes many H2O guests in its pores, which is the cause of the
high proton conductivity. The proton conductivity of the deriv-
ative [Cu(R2dtoa)]n (R C2H4OH) is two orders of magnitude
higher than that of the [Cu(Et2dtoa)]n (Et C2H5) derivative,
which is derived from the presence of OH groups.203
A high proton conductivity has been achieved in the
two-dimensional PCP {(NH4)2[Zn2(ox)3] adp 3H2O}n (ox2
oxalate, adp adipic acid),204 in which a combination of two of
the above concepts is adopted by introducing NH4 ions using
the anionic framework (type I) and putting carboxyl end groups
of adipic acid into a honeycomb-shaped void (type III). This PCP
forms typical oxalate-bridged two-dimensional anionic honey-
comb sheets of [Zn2(ox)3]n (Figure 15(a) and 15(b)).205 The
presence of an extended two-dimensional hydrogen-bonding
network and the additional protons from the carboxylic acid
and ammonium ions forming this network indicate potential
as a high proton-conducting material (Figure 15(c)). Indeed,
this PCP shows a superprotonic conductivity of 8  103 S cm1
at 298 K under 98% RH conditions (Figure 15(d)), comparable
to organic polymers such as Nafion, which is in practical use in
fuel cells.206,207 This proton conductivity is expected to include
several processes: (1) proton transport of additional protons
such as H3O by reforming hydrogen bonds between H3O
and H2O molecules (Grotthuss mechanism) and (2) direct
diffusion of additional protons with H2O or NH3 molecules
(vehicle mechanism).208,209
Although the two-dimensional anionic honeycomb
layer [Zn2(ox)3]n is magnetically inactive, the replacement of
66 MetalOrganic Frameworks

OX
adp

Zn

a
b b

(a) c c (b)

adp

O(9)
O(10)
O(10)

N(1)

b
(c)

0.8
log (sT ) (S cm-1 K)

Ea = 0.63 eV
0.6

0.4

0.2

0
3.2 3.25 3.3 3.35 3.4
(d) 1000T -1 (K-1)
Figure 15 (a) Honeycomb layer structure and (b) perspective view along the b-axis of {(NH4)2[Zn2(ox)3] adp 3H2O}n. (c) Hydrogen-bond
arrangements of COOH, H2O, and NH4 in the interlayer. (d) Arrhenius plots of the proton conductivity of {(NH4)2[Zn2(ox)3] adp 3H2O}n under 98% RH
conditions. Reproduced from Sadakiyo, M.; Yamada, T.; Kitagawa, H. J. Am. Chem. Soc. 2009, 131, 99069907, with permission of the copyright
holders.

Zn(II) ions by magnetically active metal ions could achieve the
formation of PCPs with a variety of magnetic orderings. In
particular, since ferromagnetic ordering in {[N(n-C4H9)4][M
(II)Cr(III)(ox)3]}n was first recognized,205,210 oxalate-bridged
bimetallic compounds of the {(A)[M(II)M0 (III)(ox)3]}n type
have been extensively studied as novel molecular magnets.
Introduction of tri(3-hydroxypropyl)ammonium ions [NH
(prol)3] (prol 3-hydroxypropyl) as a cation (A) to the
oxalate-bridged anionic bimetallic layers affords the PCPs
{[NH(prol)3][MCr(ox)3] xH2O}n (M Mn(II), Fe(II), Co
(II)), which exhibit coexisting ferromagnetism and proton
conduction.211 In these PCPs, [NH(prol)3] ions perform
two roles: as a source of protons and acidic parts (hydroxyl),
which correspond to the above-mentioned concepts, type I and
type III, respectively. Magnetic studies demonstrate ferromag-
netic ordering with transition temperature of 5.5 K for the

MnCr PCP, 9.0 K for the FeCr PCP, and 10.0 K for the CoCr
PCP. These PCPs show proton conduction of 1.2  1010 to
4.4  1010 S cm1 under 40% RH and 1  104 S cm1
under 75% RH at 298 K. On the basis of H2O adsorption/
desorption isotherms, the conduction under 40% RH is medi-
ated through the hydrogen-bonding network formed by the
bimetallic layer, [NH(prol)3] ions, and H2O molecules.
Under 75% RH, additional H2O molecules are associated
with the high proton conduction. This is the first example of
a PCP system showing the coexistence of ferromagnetism and
high proton conduction.
Metal phosphonates are other appealing candidates for
proton conduction. The phosphonate group possesses three
oxygen atoms; this means that ligation of metal ions into
multidimensional frameworks is possible while oxygen atoms
may still be available to act as further hydrogen-bond accep-
tors. These sites could serve to anchor carrier molecules or
directly transfer protons as part of a conduction pathway. The
novel PCP, {[Zn3(L)(H2O)2] 2H2O}n (L 1,3,5-benzenetri-
phosphonate), has a layered structure in which the phospho-
nate and Zn(II) ions do not saturate each others coordination
spheres.212 The result is an interlayer region that is rich in
phosphonate oxygen atoms and Zn-ligated H2O molecules.
These groups serve to anchor free H2O molecules into ordered
chains. The proton conductivity was determined to be
3.5  105 S cm1 at 298 K under 98% RH. The H2O molecules
serve as the chief source of protons (type III concept).
An Arrhenius plot gives a low activation energy of 0.17 eV,
corroborating the solid-state NMR data, which show exchange
between all deuterium sites in the D2O analogue, even at 253 K.
The above-mentioned proton-conductive PCPs have H2O
molecules as proton carriers. Anhydrous proton-conducting
solids that are able to operate at high temperature (above
373 K) are required in fuel-cell technology. Heterocyclic
organic molecules such as imidazole (im) or triazole (tz)
have attracted considerable attention for this purpose because
they are nonvolatile molecules with high boiling points and
they can transfer protons via the Grotthuss-type mechanism
using their amphoteric characters. Based on this strategy,
two types of PCPs with amphoteric heterocyclic organic
guests have been synthesized toward anhydrous proton con-
ductors. The first type comprises imidazole-containing
PCPs, {[Al(OH)(1,4-ndc)] 0.6im}n and {[Al(OH)(1,4-bdc)] connections at the interface between crystals, where the mod-
1.3im}n.213 Guest-free [Al(OH)(1,4-ndc)]n shows a proton
conductivity lower than 1013 S cm1, indicative of negligible
proton conductivity for this apohost framework. After installa-
tion of im, the proton conductivity of {[Al(OH)(1,4-ndc)] reported.217220 The first successful synthesis of coreshell-type
0.6im}n is 2.2  105 S cm1 at 393 K. Evidence from the isotope
effect indicates that the proton conductivity is mainly contributed
by the Grotthuss-type mechanism. In contrast, the proton con-
ductivity of {[Al(OH)(1,4-bdc)] 1.3im}n is 1.0 107 S cm1 at
393 K, about two orders of magnitude lower than that of {[Al
(OH)(1,4-ndc)] 0.6im}n, despite the higher amount of loaded
im guests. This may also be due to the difference in dynamic
motion of the im guest, which is based on the interaction between
the im guest and the PCP host. The hydrophilic pore surface of
{[Al(OH)(1,4-bdc)] 1.3im}n results in a strong interaction with
even half the amount of adsorbed im guests and significantly
decelerates their mobility, resulting in a poor proton-transfer
rate. Because of the hydrophobic and flat pore surface of
MetalOrganic Frameworks 67
{[Al(OH)(1,4-ndc)] 0.6im}n, adsorbed im can move more freely
in it than in {[Al(OH)(1,4-bdc)] 1.3im}n or in the bulk phase,
and eventually higher proton conductivity is observed. The
second type is the triazole-containing PCP, b-{[Na3(L)] x(tz)}n
(L3 2,4,6-trihydroxy-1,3,5-benzenetrisulfonate).214 This PCP
crystallizes in a honeycomb structure with one-dimensional
pores lined with sulfonate groups. The tz-loaded PCPs are much
better proton conductors than their unloaded, hydrated counter-
parts, and indeed function above 373 K. The data confirm a
dependence of proton conductivity on tz loading. At 423 K, the
conductivity measured for x 0.30, 0.45, and 0.60 tz-loaded
PCPs reaches roughly 2  104, 5  104, and 4  104 S cm1,
respectively. Neither pure tz molecules nor pure PCP show pro-
ton conductivity greater than 106 S cm1.
5.03.5.9 Other Properties
If we can combine ferroelectrically polarizable guest molecules
with PCPMs, then we can obtain new types of multiferroic
materials, where ferroelectricity and ferromagnetism coexist.
Such multifunctional materials have recently aroused increasing
interest from the viewpoint of the development of new materials
with very large magnetoelectric effects. {[Mn3(HCOO)6]
EtOH}n is considered to be the first ferroelectric PCPM.215
Because the desolvated compound [Mn3(HCOO)6]n shows a
very small and almost temperature-independent dielectric con-
stant, irrespective of the direction of the electric field,216 the
observed ferroelectric property is considered to come mainly
from the guest EtOH molecules.
5.03.6 Composite PCPs
Functionalization of PCP outer surfaces is a great challenge,
but it is a promising methodology not only for modification of
the porous properties but also for the addition of a new func-
tion to the PCP without changing the characteristic features of
the PCP crystal itself, resulting in multifunctional PCPs. One
way to decorate the crystal surfaces of a PCP is to hybridize the
core PCP crystal with a different shell crystal by epitaxial
growth at the single-crystal level, thus creating coreshell PCP
heteroepitaxial crystals. Such a lattice match promises pore
ified crystal structure should influence the mobility and diffu-
sion of adsorbates. A few pioneering synthetic studies on the
hybridization of extended coordination structures have been
composite PCP single crystals was performed using the
[M2(dicarboxylate)2(N ligand)]n PCP.221 The surfaces of the
core PCP crystal ([Zn2(1,4-ndc)2(dabco)]n) support the growth
of a single-shell crystal ([Cu2(1,4-ndc)2(dabco)]n), which was
otherwise obtained only as a microcrystalline powder in the
bulk. In this hybrid crystal, all surfaces of the colorless core PCP
crystal are covered by the greenish shell PCP crystal. In contrast,
the anisotropic hybridization of two PCP frameworks into a
single crystal was achieved, in which BAB-type block PCP crystals
with the core PCP crystal [Zn2(1,4-ndc)2(dabco)]n between
the second crystals [Zn2(1,4-ndc)2(dpndi)]n (dpndi N,N0 -di
(4-pyridyl)-1,4,5,8-naphthalenetetracarboxydiimide) were sel-
ectively formed.222 The pillar ligands with different lengths
68 MetalOrganic Frameworks
A A
B B
Figure 16 View of two types of pores, A and B, in H3AM12O40@HKUST-
1 (A Si, Ge, P, and As; M W and Mo). Reproduced from Sun, C.-Y.;
Liu, S.-X.; Liang, D.-D.; Zhao, K.-Z.; Ren, Y.-H.; Su, Z. M. J. Am. Chem.
Soc. 2009, 131, 18831888, with permission of the copyright holders.
(dabco and dpndi) were used for the crystal hybridization,
which prohibited secondary growth at the specific surfaces of
the core crystal because of the lattice mismatch.
The numerous applications of polyoxometallates (POMs)
in acid and oxidation catalysis have recently attracted great
attention.223,224 However, their applications are limited by
their low specific surface area, low stability under catalytic
conditions, and high solubility in aqueous solution. One of
the strategies to overcome these drawbacks involves their
encapsulation within porous solid matrices. Recently, this
strategy was applied to a few selected PCPs with large pores.
An interesting case concerns the direct synthesis of HKUST-1
loaded with POM moieties with a Langmuir surface area close
to 460 m2 g1 (Figure 16).225,226 The representative acid cata-
lytic performance was assessed through the hydrolysis of esters
in excess water, which showed high catalytic activity and can be
used repeatedly without activity loss. Moreover, catalytic selec-
tivity, which is dependent on the molecular size of substrates,
and substrate accessibility for the pore surface were observed.
Encapsulation of the Keggin-type POM H3PMo12O40 within
the framework of a mesoporous MIL-100(Fe) was achieved by
direct hydrothermal synthesis in the absence of fluorine.227
The obtained composite H3PMo12O40@MIL-100(Fe) exhibits
a pore volume of 0.373 cm3 g1 and a BET surface area close to
1000 m2 g1, indicating that small gas molecules can easily
diffuse inside the cavities despite the presence of POMs. More-
over, this composite is stable in aqueous solution with no
POM leaching even after more than 2 months. In addition,
no exchange of the POMs by tetrabutylammonium perchlorate
in organic media was observed.
5.03.7 Conclusion
As a new class of porous materials, PCPs will continue to attract
interest and inquiry among both academia and industry. They
have shown not only very diverse structural topologies but also
great promise because of their storage, separation, catalysis,
magnetic, and conductive properties, among others. Their
own particular ability to tune precisely pore size and shape
and pore surface characteristics allows researchers to fabricate
desired porous frameworks with excellent porous properties.
In addition, flexible PCPs could become materials with the
potential to produce properties that are not observed in con-
ventional materials. The next target in this research area is to
achieve ultra-precise control of PCPs properties using more
multicomponent systems, including solid solutions, to inte-
grate more than two types of porous functionalities (separation
and storage, separation and catalysis, and storage and catalysis)
into one PCP framework, and to work toward practical use
of flexible PCPs showing gate-opening adsorption. I believe
that, in the near future, PCPs will be recognized as truly
useful porous materials, alongside zeolites and activated car-
bons. For related chapters in this Comprehensive, we refer to
Chapters 4.05, 5.02, 5.04, 5.05, 5.10, 8.01, 9.18, and 9.37.
References
1. Kitagawa, S.; Kitaura, R.; Noro, S. Angew. Chem. Int. Ed. 2004, 43, 23342375.
2. Yaghi, O. M.; OKeeffe, M.; Ockwig, N. W.; Chae, H. K.; Eddaoudi, M.; Kim, J.
Nature 2003, 423, 705714.
3. Janiak, C. Dalton Trans. 2003, 27812804.
4. Bradshaw, D.; Claridge, J. B.; Cussen, E. J.; Prior, T. J.; Rosseinsky, M. J. Acc.
Chem. Res. 2005, 38, 273282.
5. Ferey, G. Chem. Soc. Rev. 2008, 37, 191214.
6. Moulton, B.; Zaworotko, M. J. Chem. Rev. 2001, 101, 16291658.
7. Kesanli, B.; Lin, W. B. Coord. Chem. Rev. 2003, 246, 305326.
8. Kepert, C. J. Chem. Commun. 2006, 695700.
9. Kawano, M.; Fujita, M. Coord. Chem. Rev. 2007, 251, 25922605.
10. Noro, S.; Kitagawa, S.; Akutagawa, T.; Nakamura, T. Prog. Polym. Sci. 2009, 34,
240279.
11. Noro, S. Phys. Chem. Chem. Phys. 2010, 12, 25192531.
12. Rao, C. N. R.; Natarajan, S.; Vaidhyanathan, R. Angew. Chem. Int. Ed. 2004, 43,
14661496.
13. Hagrman, P. J.; Hagrman, D.; Zubieta, J. Angew. Chem. Int. Ed. 1999, 38,
26382684.
14. Nagaoka, M.; Ohta, Y.; Hitomi, H. Coord. Chem. Rev. 2007, 251, 25222536.
15. Morris, R. E.; Wheatley, P. S. Angew. Chem. Int. Ed. 2008, 47, 49664981.
16. Blake, A. J.; Champness, N. R.; Hubberstey, R.; Li, W.-S.; Withersby, M. A.;
Schroder, M. Coord. Chem. Rev. 1999, 183, 117138.
17. Batten, S. R.; Robson, R. Angew. Chem. Int. Ed. 1998, 37, 14601494.
18. Hosseini, M. W. Acc. Chem. Res. 2005, 38, 313323.
19. Kitagawa, S.; Kondo, M. Bull. Chem. Soc. Jpn. 1998, 71, 17391753.
20. Uemura, K.; Kitagawa, S. Chem. Soc. Rev. 2005, 34, 109119.
21. Kitagawa, S.; Noro, S.; Nakamura, T. Chem. Commun. 2006, 701707.
22. Uemura, T.; Horike, S.; Kitagawa, S. Chem. Asian J. 2006, 1, 3644.
23. Kitagawa, S.; Matsuda, R. Coord. Chem. Rev. 2007, 251, 24902509.
24. Maji, T. K.; Kitagawa, S. Pure Appl. Chem. 2007, 79, 21552177.
25. Uemura, T.; Yanai, N.; Kitagawa, S. Chem. Soc. Rev. 2009, 38, 12281236.
26. Horike, S.; Shimomura, S.; Kitagawa, S. Nat. Chem. 2009, 1, 695704.
27. Tranchemontagne, D. J.; Ni, Z.; OKeeffe, M.; Yaghi, O. M. Angew. Chem. Int. Ed.
2008, 47, 51365147.
28. Fujita, M.; Tominaga, M.; Hori, A.; Therrien, B. Acc. Chem. Res. 2005, 38, 369378.
29. Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N. Acc. Chem. Res. 2005,
38, 349358.
30. Seidel, S. R.; Stang, P. J. Acc. Chem. Res. 2002, 35, 972983.
31. Dobrzanska, L.; Lloyd, G. O.; Raubenheimer, H. G.; Barbour, L. J. J. Am. Chem.
Soc. 2005, 127, 1313413135.
32. Allison, S. A.; Barrer, R. M. J. Chem. Soc. A 1969, 17171723.
33. Soldatov, D. V.; Enright, G. D.; Ripmeester, J. A. Cryst. Growth Des. 2004, 4,
11851194.
34. Manakov, A. Y.; Lipkowski, J.; Pielaszek, J. J. Incl. Phenom. 1999, 35, 531548.
35. Soldatov, D. V.; Ripmeester, J. A.; Shergina, S. I.; Sokolov, I. E.; Zanina, A. S.;
Gromilov, S. A.; Dyadin, Y. A. J. Am. Chem. Soc. 1999, 121, 41794188.
36. Dalrymple, S. A.; Shimizu, G. K. H. J. Am. Chem. Soc. 2007, 129, 1211412116.
37. Takamizawa, S.; Akatsuka, T.; Ueda, T. Angew. Chem. Int. Ed. 2008, 47,
16891692.
38. Noro, S.; Ohba, T.; Fukuhara, K.; Takahashi, Y.; Akutagawa, T.; Nakamura, T.
Dalton Trans. 2011, 40, 22682274.

39. Bailar, J. C., Jr. Prep. Inorg. React. 1964, 1, 127.
40. Wilde, R. E.; Ghosh, S. N.; Marshall, B. J. Inorg. Chem. 1970, 9, 25122516.
41. Buser, H. J.; Schwarzenbach, D.; Petter, W.; Ludi, A. Inorg. Chem. 1977, 16,
27042710.
42. Dunbar, K. R.; Heintz, R. A. Prog. Inorg. Chem. 1997, 45, 283373.
43. Atwood, J. L.; Davies, J. E. D.; MacNicol, D. D. Inclusion Compounds; Academic
Press: London, 1984; Vol. 1.
44. Atwood, J. L.; Davies, J. E. D.; MacNicol, D. D. Inclusion Compounds; Oxford
University Press: Oxford, 1991; Vol. 5.
45. Kinoshita, Y.; Matsubara, I.; Higuchi, T.; Saito, Y. Bull. Chem. Soc. Jpn. 1959,
32, 12211226.
46. Gardner, G. B.; Venkataraman, D.; Moore, J. S.; Lee, S. Nature 1995, 374,
79279795.
47. Venkataraman, D.; Gardner, G. B.; Lee, S.; Moore, J. S. J. Am. Chem. Soc. 1995,
117, 1160011601.
48. Hoskins, B. F.; Robson, R. J. Am. Chem. Soc. 1989, 111, 59625964.
49. Kitagawa, S.; Munakata, M.; Tanimura, T. Inorg. Chem. 1992, 31, 17141717.
50. Kondo, M.; Yoshitomi, T.; Seki, K.; Matsuzaka, H.; Kitagawa, S. Angew. Chem. Int.
Ed. 1997, 36, 17251727.
51. Li, H.; Eddaoudi, M.; Groy, T. L.; Yaghi, O. M. J. Am. Chem. Soc. 1998, 120,
85718572.
52. Batten, S. R.; Champness, N. R.; Chen, X.-M.; Garcia-Martinez, J.; Kitagawa, S.;
Ohrstrom, L.; OKeeffe, M.; Suh, M. P.; Reedijk, J. CrystEngComm 2012, 14,
30013004.
53. OKeeffe, M.; Peskov, M. A.; Ramsden, S. J.; Yaghi, O. M. Acc. Chem. Res. 2008,
41, 17821789.
54. Caskey, S. R.; Wong-Foy, A. G.; Matzger, A. J. J. Am. Chem. Soc. 2008, 130,
1087010871.
55. Forgan, R. S.; Smaldone, R. A.; Gassensmith, J. J.; Furukawa, H.; Cordes, D. B.;
Li, Q.; Wilmer, C. E.; Botros, Y. Y.; Snurr, R. Q.; Slawin, A. M. Z.; Stoddart, J. F.
J. Am. Chem. Soc. 2012, 134, 406417.
56. Shimomura, S.; Matsuda, R.; Tsujino, T.; Kawamura, T.; Kitagawa, S. J. Am.
Chem. Soc. 2006, 128, 1641616417.
57. Yang, G.; Raptis, R. G. Chem. Commun. 2004, 20582059.
58. Liu, S. X.; Xie, L. H.; Gao, B.; Zhang, C. D.; Sun, C. Y.; Li, D. H.; Su, Z. M. Chem.
Commun. 2005, 50235025.
59. Olenev, A. V.; Baranov, A. I.; Shevelkov, A. V.; Popovkin, B. A. Eur. J. Inorg.
Chem. 2002, 547553.
60. Lin, Z. J.; Wragg, D. S.; Morris, R. E. Chem. Commun. 2006, 20212023.
61. Oliver, S.; Kuperman, A.; Lough, A.; Ozin, G. A. J. Mater. Chem. 1997, 7,
807812.
62. Higuchi, M.; Tanaka, D.; Horike, S.; Sakamoto, H.; Nakamura, K.; Takashima, Y.;
Hijikata, Y.; Yanai, N.; Kim, J.; Kato, K.; Kubota, Y.; Takata, M.; Kitagawa, S. J. Am.
Chem. Soc. 2009, 131, 1033610337.
63. Xie, Z.; Ma, L.; deKrafft, K. E.; Jin, A.; Lin, W. J. Am. Chem. Soc. 2010, 132,
922923.
64. Kawano, M.; Kawamichi, T.; Haneda, T.; Kojima, T.; Fujita, M. J. Am. Chem. Soc.
2007, 129, 1541815419.
65. Zaworotko, M. J. Nature 2008, 451, 410411.
66. Ohmori, O.; Kawano, M.; Fujita, M. Angew. Chem. Int. Ed. 2005, 44,
19621964.
67. Ferey, G.; Serre, C.; Mellot-Draznieks, C.; Millange, F.; Surble, S.; Doutour, J.;
Margiolaki, I. Angew. Chem. Int. Ed. 2004, 43, 62966301.
68. Jhung, S. H.; Lee, J.-H.; Chang, J.-S. Bull. Korean Chem. Soc. 2005, 26,
880881.
69. Cohen, S. M. Chem. Rev. 2012, 112, 9701000.
70. Wu, C.-D.; Hu, A.; Zhang, L.; Lin, W. J. Am. Chem. Soc. 2005, 127, 89408941.
71. Kaye, S. S.; Long, J. R. J. Am. Chem. Soc. 2008, 130, 806807.
72. An, J.; Geib, S. J.; Rosi, N. L. J. Am. Chem. Soc. 2009, 131, 83768377.
73. An, J.; Rosi, N. L. J. Am. Chem. Soc. 2010, 132, 55785579.
74. Uemura, T.; Hoshino, Y.; Kitagawa, S.; Yoshida, K.; Isoda, S. Chem. Mater. 2006,
18, 992995.
75. Hermes, S.; Witte, T.; Hikov, T.; Zacher, D.; Bahnmuller, S.; Langstein, G.;
Huber, K.; Fischer, R. A. J. Am. Chem. Soc. 2007, 129, 53245325.
76. Tsuruoka, T.; Furukawa, S.; Takashima, Y.; Yoshida, K.; Isoda, S.; Kitagawa, S.
Angew. Chem. Int. Ed. 2009, 48, 47394743.
77. Zacher, D.; Schmid, R.; Woll, C.; Fischer, R. A. Angew. Chem. Int. Ed. 2011, 50,
176199.
78. Shekhah, O.; Wang, H.; Paradinas, M.; Ocal, C.; Schupbach, B.; Terfort, A.;
Zacher, D.; Fischer, R. A.; Woll, C. Nat. Mater. 2009, 8, 481484.
79. Chen, B.; Liang, C.; Yang, J.; Contreras, D. S.; Clancy, Y. L.; Lobkovsky, E. B.;
Yaghi, O. M.; Dai, S. Angew. Chem. Int. Ed. 2006, 45, 13901393.
80. Pichon, A.; Lazuen-Garay, A.; James, S. L. CrystEngComm 2006, 8, 211214.
MetalOrganic Frameworks 69
81. Liu, Y.-H.; Lu, Y.-L.; Tsai, H.-L.; Wang, J.-C.; Lu, K. L. J. Solid State Chem.
2001, 158, 315319.
82. Lu, J. Y.; Babb, A. M. Chem. Commun. 2002, 13401341.
83. Chen, X.-M.; Tong, M.-L. Acc. Chem. Res. 2007, 40, 162170.
84. Evans, O. R.; Lin, W. Acc. Chem. Res. 2002, 35, 511522.
85. Xiong, R.-G.; Xue, X.; Zhao, H.; You, X.-Z.; Abrahams, B. F.; Xue, Z. Angew.
Chem. Int. Ed. 2002, 41, 38003803.
86. Kondo, M.; Takashima, Y.; Seo, J.; Kitagawa, S.; Furukawa, S. CrystEngComm
2010, 12, 23502353.
87. Dybtsev, D. N.; Chun, H.; Kim, K. Angew. Chem. Int. Ed. 2004, 43, 50335036.
88. Chun, H.; Moon, J. Inorg. Chem. 2007, 46, 43714373.
89. Kondo, M.; Okubo, T.; Asami, A.; Noro, S.; Yoshitomi, T.; Kitagawa, S.; Ishii, T.;
Matsuzaka, H.; Seki, K. Angew. Chem. Int. Ed. 1999, 38, 140143.
90. Kitaura, R.; Kitagawa, S.; Kubota, Y.; Kobayashi, T. C.; Kindo, K.; Mita, Y.;
Matsuo, A.; Kobayashi, M.; Chang, H.-C.; Ozawa, T. C.; Suzuki, M.; Sakata, M.;
Takata, M. Science 2002, 298, 23582361.
91. Matsuda, R.; Kitaura, R.; Kitagawa, S.; Kubota, Y.; Belosludov, R. V.;
Kobayashi, T. C.; Sakamoto, H.; Chiba, T.; Takata, M.; Kawazoe, Y.; Mita, Y.
Nature 2005, 436, 238241.
92. Kitaura, R.; Matsuda, R.; Kubota, Y.; Kitagawa, S.; Takata, M.; Kobayashi, T. C.;
Suzuki, M. J. Phys. Chem. B 2005, 109, 2337823385.
93. Matsuda, R.; Kitaura, R.; Kitagawa, S.; Kubota, Y.; Kobayashi, C.; Horike, S.;
Takata, M. J. Am. Chem. Soc. 2004, 126, 1406314070.
94. Sakamoto, H.; Kitaura, R.; Matsuda, R.; Kitagawa, S.; Kubota, Y.; Takata, M. Chem.
Lett. 2010, 39, 218219.
95. Matsuda, R.; Tsujino, T.; Sato, H.; Kubota, Y.; Morishige, K.; Takata, M.;
Kitagawa, S. Chem. Sci. 2010, 1, 315321.
96. Li, D.; Kaneko, K. J. Phys. Chem. B 2000, 104, 89408945.
97. Uemura, T.; Kitaura, R.; Ohta, Y.; Nagaoka, M.; Kitagawa, S. Angew. Chem. Int. Ed.
2006, 45, 41124116.
98. Matsuda, R.; Kitaura, R.; Kubota, Y.; Kobayashi, T. C.; Takata, M.; Kitagawa, S.
Micropor. Mesopor. Mater. 2010, 129, 296303.
99. Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; OKeeffe, M.; Yaghi, O. M.
Science 2002, 295, 469472.
100. Kitaura, R.; Seki, K.; Akiyama, G.; Kitagawa, S. Angew. Chem. Int. Ed. 2003, 42,
428431.
101. Ma, B.-Q.; Mulfort, K. L.; Hupp, J. T. Inorg. Chem. 2005, 44, 49124914.
102. Chen, B.; Ma, S.; Hurtado, E. J.; Lobkovsky, E. B.; Zhou, H.-C. Inorg. Chem.
2007, 46, 84908492.
103. Huang, X.-C.; Lin, Y.-Y.; Zhang, J.-P.; Chen, X.-M. Angew. Chem. Int. Ed. 2005,
45, 15571559.
104. Park, K. S.; Ni, Z.; Cote, A. P.; Choi, J. T.; Huang, R.; Uribe-Romo, F. J.;
Chae, H. K.; OKeeffe, M.; Yaghi, O. M. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
1018610191.
105. Hayashi, H.; Cote, A. P.; Furukawa, Y.; OKeeffe, M.; Yaghi, O. M. Nat. Mater.
2007, 6, 501506.
106. Banerjee, R.; Phan, A.; Wang, B.; Knobler, C.; Furukawa, H.; OKeeffe, M.;
Yaghi, O. M. Science 2008, 319, 939943.
107. Tao, J.; Chen, X. M.; Huang, R. B.; Zheng, L. S. J. Solid State Chem. 2003, 170,
130134.
108. Ma, C. B.; Chen, C. N.; Liu, Q. T.; Liao, D. H.; Li, L.; Sun, L. C. New J. Chem.
2003, 27, 890894.
109. Zeng, M. H.; Feng, X. L.; Chen, X. M. Dalton Trans. 2004, 22172223.
110. Wen, Y. H.; Cheng, H. K.; Feng, Y. L.; Zhang, H.; Li, Z. H.; Yao, Y. G. Inorg. Chim.
Acta 2005, 358, 33473354.
111. Tian, G.; Zhu, G. S.; Fang, Q. R.; Guo, X. D.; Xue, M.; Sun, J. Y.; Qiu, S. L. J. Mol.
Struct. 2006, 787, 4549.
112. Horike, S.; Tanaka, D.; Nakagawa, K.; Kitagawa, S. Chem. Commun. 2007,
33953397.
113. Tanaka, D.; Nakagawa, K.; Higuchi, M.; Horike, S.; Kubota, Y.; Kobayashi, T. C.;
Takata, M.; Kitagawa, S. Angew. Chem. Int. Ed. 2008, 47, 39143918.
114. Nakagawa, K.; Tanaka, D.; Horike, S.; Shimomura, S.; Higuchi, M.; Kitagawa, S.
Chem. Commun. 2010, 46, 42584260.
115. Tanaka, D.; Henke, A.; Albrecht, K.; Moeller, M.; Nakagawa, K.; Kitagawa, S.;
Groll, J. Nat. Chem. 2010, 2, 410416.
116. Fukushima, T.; Horike, S.; Inubushi, Y.; Nakagawa, K.; Kubota, Y.; Takata, M.;
Kitagawa, S. Angew. Chem. Int. Ed. 2010, 49, 48204824.
117. Sato, H.; Matsuda, R.; Sugimoto, K.; Takata, M.; Kitagawa, S. Nat. Mater. 2010, 9,
661666.
118. Horike, S.; Inubushi, Y.; Hori, T.; Fukushima, T.; Kitagawa, S. Chem. Sci. 2012,
3, 116120.
119. Hijikata, Y.; Horike, S.; Sugimoto, M.; Sato, H.; Matsuda, R.; Kitagawa, S. Chem.
Eur. J. 2011, 17, 51385144.
70 MetalOrganic Frameworks
120. Miyamoto, J.; Hattori, Y.; Noguchi, D.; Tanaka, H.; Ohba, T.; Utsumi, S.;
Kanoh, H.; Kim, Y. A.; Muramatsu, H.; Hayashi, T.; Endo, M.; Kaneko, K. J. Am.
Chem. Soc. 2006, 128, 1263612637.
121. Takamizawa, S.; Nakata, E.; Yokoyama, H.; Mochizuki, K.; Mori, W. Angew. Chem.
Int. Ed. 2003, 42, 43314334.
122. Takamizawa, S.; Nakata, E.; Akatsuka, T.; Kachi-Terajima, C.; Miyake, R. J. Am.
Chem. Soc. 2008, 130, 1788217892.
123. Takamizawa, S.; Nakata, T.; Akatsuka, T.; Miyake, R.; Kakizaki, Y.; Takeuchi, H.;
Maruta, G.; Takeda, S. J. Am. Chem. Soc. 2010, 132, 37833792.
124. Duren, T.; Sarkisov, L.; Yaghi, O. M.; Snurr, R. Q. Langmuir 2004, 20,
26832689.
125. Bourrelly, S.; Llewellyn, P. L.; Serre, C.; Millange, F.; Loiseau, T.; Ferey, G. J. Am.
Chem. Soc. 2005, 127, 1351913521.
126. Noro, S.; Kitagawa, S.; Kondo, M.; Seki, K. Angew. Chem. Int. Ed. 2000, 39,
20822084.
127. Seki, K. Chem. Commun. 2001, 14961497.
128. Ma, S.; Sun, S.; Simmons, J. M.; Collier, C. D.; Yuan, D.; Zhou, H.-C. J. Am.
Chem. Soc. 2008, 130, 10121016.
129. Johnson, J. Chem. Eng. News 2004, 82, 3642.
130. Yong, Z.; Mata, V.; Rodrigues, A. E. Sep. Purif. Technol. 2002, 26, 195205.
131. Millward, A. R.; Yaghi, O. M. J. Am. Chem. Soc. 2005, 127, 1799817999.
132. Himeno, S.; Komatsu, T.; Fujita, S. J. Chem. Eng. Data 2005, 50, 369376.
133. Llewellyn, P. L.; Bourrelly, S.; Serre, C.; Filinchuk, Y.; Ferey, G. Angew. Chem. Int.
Ed. 2006, 45, 77517754.
134. Furukawa, H.; Miller, M. A.; Yaghi, O. M. J. Mater. Chem. 2007, 17, 31973204.
135. Hubner, O.; Gloss, A.; Fichtner, M.; Klopper, W. J. Phys. Chem. A 2004, 108,
30193023.
136. Wong, M.; Kuiken, B. E. V.; Buda, C.; Dunietz, B. D. J. Phys. Chem. C 2009, 113,
1257112579.
137. Barman, S.; Furukawa, H.; Blacque, O.; Venkatesan, K.; Yaghi, O. M.; Berke, H.
Chem. Commun. 2010, 46, 79817983.
138. Rowsell, J. L. C.; Millward, A. R.; Park, K. S.; Yaghi, O. M. J. Am. Chem. Soc.
2004, 126, 56665667.
139. Rowsell, J. L. C.; Yaghi, O. M. J. Am. Chem. Soc. 2006, 128, 13041315.
140. Lee, Y.-G.; Moon, H. R.; Suh, P. Angew. Chem. Int. Ed. 2008, 47, 77417745.
141. Ma, S.; Simmons, J. M.; Sun, D.; Yuan, D.; Zhou, H.-C. Inorg. Chem. 2009, 48,
52635268.
142. Sumida, K.; Hill, M. R.; Horike, S.; Dailly, A.; Long, J. R. J. Am. Chem. Soc. 2009,
131, 1512015121.
143. Li, Y.; Yang, R. T. J. Am. Chem. Soc. 2006, 128, 726727.
144. Li, Y.; Yang, R. T. J. Am. Chem. Soc. 2006, 128, 81368137.
145. Campesi, R.; Cuevas, F.; Latroche, M.; Hirscher, M. Phys. Chem. Chem. Phys.
2010, 12, 1045710459.
146. Contescu, C. I.; Brown, C. M.; Liu, Y.; Bhat, V. V.; Gallego, N. C. J. Phys. Chem. C
2009, 113, 58865890.
147. Stadie, N. O.; Purewal, J. J.; Ahn, C. C.; Fultz, B. Langmuir 2010, 26,
1548115485.
148. Akiyama, G.; Matsuda, R.; Kitagawa, S. Chem. Lett. 2010, 39, 360361.
149. Tanaka, D.; Higuchi, M.; Horike, S.; Matsuda, R.; Kinoshita, Y.; Yanai, N.;
Kitagawa, S. Chem. Asian J. 2008, 3, 13431349.
150. Reid, C. R.; Thomas, K. M. Langmuir 1999, 15, 32063218.
151. Newalkar, B. L.; Choudary, N. V.; Kumar, P.; Komarneni, S.; Bhat, T. S. G. Chem.
Mater. 2002, 14, 304309.
152. Breck, D. W.; Eversole, W. G.; Milton, R. M.; Reed, T. B.; Thomas, T. L. J. Am.
Chem. Soc. 1956, 78, 59635972.
153. Xiang, S.; Zhou, W.; Zhang, Z.; Green, M. A.; Liu, Y.; Chen, B. Angew. Chem. Int.
Ed. 2010, 49, 46154618.
154. Hu, Y.; Xiang, S.; Zhang, W.; Zhang, Z.; Wang, L.; Bai, J.; Chen, B. Chem.
Commun. 2009, 75517553.
155. Xiang, S.; Zhou, W.; Gallegos, J. M.; Liu, Y.; Chen, B. J. Am. Chem. Soc. 2009,
131, 1241512419.
156. Banerjee, R.; Furukawa, H.; Britt, D.; Knobler, C.; OKeeffe, M.; Yaghi, O. M. J. Am.
Chem. Soc. 2009, 131, 38753877.
157. Wang, B.; Cote, A. P.; Furukawa, H.; OKeeffe, M.; Yaghi, O. M. Nature 2008, 453,
207211.
158. Quartapelle Procopio, E.; Linares, F.; Montoro, C.; Colombo, V.; Maspero, A.;
Barea, E.; Navarro, J. A. R. Angew. Chem. Int. Ed. 2010, 49, 73087311.
159. Ferey, G. Nature 2005, 436, 187188.
160. Samsonenko, D. G.; Kim, H.; Sun, Y.; Kim, G.-H.; Lee, H.-S.; Kim, K. Chem. Asian
J. 2007, 2, 484488.
161. Kargol, M.; Zajac, J.; Jones, D. J.; Steriotis, Th.; Roziere, J.; Vitse, P. Chem.
Mater. 2004, 16, 39113918.
162. Zhang, J.-P.; Horike, S.; Kitagawa, S. Angew. Chem. Int. Ed. 2007, 46, 889892.
163. Zhang, J.-P.; Kitagawa, S. J. Am. Chem. Soc. 2008, 130, 907917.
164. Shimomura, S.; Horike, S.; Matsuda, R.; Kitagawa, S. J. Am. Chem. Soc. 2007,
129, 1099010991.
165. Shimomura, S.; Higuchi, M.; Matsuda, R.; Yoneda, K.; Hijikata, Y.; Kubota, Y.;
Mita, Y.; Kim, J.; Takata, M.; Kitagawa, S. Nat. Chem. 2010, 2, 633637.
166. Ma, S.; Wang, X.-S.; Yuan, D.; Zhou, H.-C. Angew. Chem. Int. Ed. 2008, 47,
41304133.
167. Demessence, A.; Long, J. R. Chem. Eur. J. 2010, 16, 59025908.
168. Farrusseng, D.; Aguado, S.; Pinel, C. Angew. Chem. Int. Ed. 2009, 48,
75027513.
169. Fujita, M.; Kwon, Y. J.; Washizu, S.; Ogura, K. J. Am. Chem. Soc. 1994, 116,
11511152.
170. Ohmori, O.; Fujita, M. Chem. Commun. 2004, 4, 15861587.
171. Vaidhyanathan, R.; Bradshaw, D.; Rebilly, J.-N.; Barrio, J. P.; Gould, J. A.;
Berry, N. G.; Rosseinsky, M. J. Angew. Chem. Int. Ed. 2006, 45, 64956499.
172. Ingleson, M. J.; Barrio, J. P.; Bacsa, J.; Dickinson, C.; Park, H.; Rosseinsky, M. J.
Chem. Commun. 2008, 12871289.
173. Zheng, M.; Liu, Y.; Wang, C.; Liu, S.; Lin, W. Chem. Sci. 2012, 3, 26232627.
174. Seo, J. S.; Whang, D.; Lee, H.; Jun, S. I.; Oh, J.; Jeon, Y. J.; Kim, K. Nature 2000,
404, 982986.
175. Hasegawa, S.; Horike, S.; Matsuda, R.; Furukawa, S.; Mochizuki, K.; Kinoshita, Y.;
Kitagawa, S. J. Am. Chem. Soc. 2007, 129, 26072614.
176. Uemura, T.; Hiramatsu, D.; Kubota, Y.; Takata, M.; Kitagawa, S. Angew. Chem. Int.
Ed. 2007, 46, 49874990.
177. Liu, B.; Shioyama, H.; Akita, T.; Xu, Q. J. Am. Chem. Soc. 2008, 130,
53905391.
178. Liu, B.; Shioyama, H.; Jiang, H.; Zhang, X.; Xu, Q. Carbon 2010, 48,
456463.
179. Blake, A. J.; Hill, S. J.; Hubberstey, P.; Li, W.-S. J. Chem. Soc. Dalton Trans.
1997, 913914.
180. Kondo, A.; Noguchi, H.; Ohnishi, S.; Kajiro, H.; Tohdoh, A.; Hattori, Y.; Xu, W.-C.;
Tanaka, H.; Kanoh, H.; Kaneko, K. Nano Lett. 2006, 6, 25812584.
181. Kondo, A.; Noguchi, H.; Carlucci, L.; Proserpio, D. M.; Ciani, G.; Kajiro, H.;
Ohba, T.; Kanoh, H.; Kaneko, K. J. Am. Chem. Soc. 2007, 129, 1236212363.
182. Ghosh, S. K.; Kaneko, W.; Kiriya, D.; Ohba, M.; Kitagawa, S. Angew. Chem. Int.
Ed. 2008, 47, 88438847.
183. Kaye, S. S.; Dailly, A.; Yaghi, O. M.; Long, J. R. J. Am. Chem. Soc. 2007, 129,
1417614177.
184. Sundberg, R. J.; Martin, R. B. Chem. Rev. 1974, 74, 471517.
185. Maspoch, D.; Ruiz-Molina, D.; Wurst, K.; Domingo, N.; Cavallini, M.;
Biscarini, F.; Tejada, J.; Rovira, C.; Veciana, J. Nat. Mater. 2003, 2, 190195.
186. Yanai, N.; Kaneko, W.; Yoneda, K.; Ohba, M.; Kitagawa, S. J. Am. Chem. Soc.
2007, 129, 34963497.
187. Ohkoshi, S.; Tsunobuchi, T.; Takahashi, H.; Hozumi, T.; Shiro, M.; Hashimoto, K.
J. Am. Chem. Soc. 2007, 129, 30843085.
188. Milon, J.; Daniel, M. C.; Kaiba, A.; Guionneau, P.; Brandes, S.; Sutter, J. P. J. Am.
Chem. Soc. 2007, 129, 1387213878.
189. Ohkoshi, S.; Arai, K.; Sato, Y.; Hashimoto, K. Nat. Mater. 2004, 3, 857861.
190. Wang, Z. M.; Zhang, B.; Fujiwara, H.; Kobayashi, H.; Kurmoo, M. Chem.
Commun. 2004, 416417.
191. Wang, Z.; Zhang, Y.; Liu, T.; Kurmoo, M.; Gao, S. Adv. Funct. Mater. 2007, 17,
15231536.
192. Halder, G. J.; Kepert, C. J.; Moubaraki, B.; Murray, K. S.; Cashion, J. D. Science
2002, 298, 17621765.
193. Neville, S. M.; Halder, G. J.; Chapman, K. W.; Duriska, M. B.; Southon, P. D.;
Cashion, J. D.; Letard, J.-F.; Moubaraki, B.; Murray, K. S.; Kepert, C. J. J. Am.
Chem. Soc. 2008, 130, 28692876.
194. Niel, V.; Martinez-Agudo, J. M.; Munoz, M. C.; Gaspar, A. B.; Real, J. A. Inorg.
Chem. 2001, 40, 38383839.
195. Ohba, M.; Yoneda, K.; Agust, G.; Munoz, M. C.; Gaspar, A. B.; Real, J. A.;
Yamasaki, M.; Ando, H.; Nakao, Y.; Sakaki, S.; Kitagawa, S. Angew. Chem. Int. Ed.
2009, 48, 47674771.
196. Southon, P. D.; Liu, L.; Fellows, E. A.; Price, D. J.; Halder, G. J.; Chapman, K. W.;
Moubaraki, B.; Murray, K. S.; Letard, J. F.; Kepert, C. J. J. Am. Chem. Soc. 2009,
131, 1099811009.
197. Agust, G.; Ohtani, R.; Yoneda, K.; Gaspar, A. B.; Ohba, M.; Sanchez-Royo, J. F.;
Munoz, M. C.; Kitagawa, S.; Real, J. J. A. Angew. Chem. Int. Ed. 2009, 48,
89448947.
198. Fuma, Y.; Ebihara, M.; Kutsumizu, S.; Kawamura, T. J. Am. Chem. Soc. 2004,
126, 1223812239.
199. Takaishi, S.; Hosoda, M.; Kajiwara, T.; Miyasaka, H.; Yamashita, M.;
Nakanishi, Y.; Kitagawa, Y.; Yamaguchi, K.; Kobayashi, A.; Kitagawa, H. Inorg.
Chem. 2009, 48, 90489050.

200. Kobayashi, Y.; Jacobs, B.; Allendorf, M. D.; Long, J. R. Chem. Mater. 2010, 22,
41204122.
201. Zeng, M.-H.; Wang, Q.-X.; Tan, Y.-X.; Hu, S.; Zhao, H.-X.; Long, L.-S.;
Kurmoo, M. J. Am. Chem. Soc. 2010, 132, 25612563.
202. Kitagawa, H.; Nagao, Y.; Fujishima, M.; Ikeda, R.; Kanda, S. Inorg. Chem.
Commun. 2003, 6, 346348.
203. Nagao, Y.; Kubo, T.; Nakasuji, K.; Ikeda, R.; Kojima, T.; Kitagawa, H. Synth. Met.
2005, 154, 8992.
204. Sadakiyo, M.; Yamada, T.; Kitagawa, H. J. Am. Chem. Soc. 2009, 131,
99069907.
205. Tamaki, H.; Zhong, Z. J.; Matsumoto, N.; Kita, S.; Koikawa, M.; Achiwa, N.;
Hashimoto, Y.; Okawa, H. J. Am. Chem. Soc. 1992, 114, 69746979.
206. Slade, R. C. T.; Hardwick, A.; Dickens, P. G. Solid State Ion. 1983, 910,
10931098.
207. Alberti, G.; Casciola, M. Solid State Ion. 2001, 145, 316.
208. Agmon, N. Chem. Phys. Lett. 1995, 244, 456462.
209. Kreuer, K. D.; Rabenau, A.; Weppner, W. Angew. Chem. Int. Ed. 1982, 21, 208.
210. Zhong, Z. J.; Matsumoto, N.; Okawa, H.; Kida, S. Chem. Lett. 1990, 8790.
211. Okawa, H.; Shigematsu, A.; Sadakiyo, M.; Miyagawa, T.; Yoneda, K.; Ohba, M.;
Kitagawa, H. J. Am. Chem. Soc. 2009, 131, 1351613522.
212. Taylor, J. M.; Mah, R. K.; Moudrakovski, I. L.; Ratcliffe, C. I.; Vaidhyanathan, R.;
Shimizu, G. K. H. J. Am. Chem. Soc. 2010, 132, 1405514057.
213. Bureekaew, S.; Horike, S.; Higuchi, M.; Mizuno, M.; Kawamura, T.; Tanaka, D.;
Yanai, N.; Kitagawa, S. Nat. Mater. 2009, 8, 831836.
214. Hurd, J. A.; Vaidhyanathan, R.; Thangadurai, V.; Ratcliffe, C. I.;
Moudrakovski, I. L.; Shimizu, G. K. H. Nat. Chem. 2009, 1, 705710.
MetalOrganic Frameworks 71
215. Cui, H.; Wang, Z.; Takahashi, K.; Okano, Y.; Kobayashi, H.; Kobayashi, A. J. Am.
Chem. Soc. 2006, 128, 1507415075.
216. Cui, H.; Takahashi, K.; Okano, Y.; Kobayashi, H.; Wang, Z.; Kobayashi, A. Angew.
Chem. Int. Ed. 2005, 44, 65086512.
217. MacDonald, J. C.; Dorrestein, P. C.; Pilley, M. M.; Foote, M. M.; Lundburg, J. L.;
Henning, R. W.; Schultz, A. J.; Manson, J. L. J. Am. Chem. Soc. 2000, 122,
1169211702.
218. Noveron, J. C.; Lah, M. S.; Sesto, R. E. D.; Arif, A. M.; Miller, J. S.; Stang, P. J.
J. Am. Chem. Soc. 2002, 124, 66136625.
219. Ferlay, S.; Hosseini, W. Chem. Commun. 2004, 788789.
220. Bres, E. F.; Ferlay, S.; Dechambenoit, P.; Leroux, H.; Hosseini, W.; Reybtjens, S.
J. Mater. Chem. 2007, 17, 15591562.
221. Furukawa, S.; Hirai, K.; Nakagawa, K.; Takashima, Y.; Matsuda, R.; Tsurukoka, T.;
Kondo, M.; Haruki, R.; Tanaka, D.; Sakamoto, H.; Shimomura, S.; Sakata, O.;
Kitagawa, S. Angew. Chem. Int. Ed. 2009, 48, 17661770.
222. Furukawa, S.; Hirai, K.; Takashima, Y.; Nakagawa, K.; Kondo, M.; Tsuruoka, T.;
Sakata, O.; Kitagawa, S. Chem. Commun. 2009, 50975099.
223. Muller, A. S. R. Coord. Chem. Rev. 2003, 245, 153166.
224. Long, D.-L.; Tsunashima, R.; Cronin, L. Angew. Chem. Int. Ed. 2010, 49,
17361758.
225. Sun, C.-Y.; Liu, S.-X.; Liang, D.-D.; Zhao, K.-Z.; Ren, Y.-H.; Su, Z. M. J. Am.
Chem. Soc. 2009, 131, 18831888.
226. Hundal, G.; Hwang, Y. K.; Chang, J.-S. Polyhedron 2009, 28, 24502458.
227. Canioni, R.; Roch-Marchal, C.; Secheresse, F.; Horcajada, P.; Serre, C.; Hardi-
Dan, M.; Ferey, G.; Greneche, J.-M.; Lefebvre, F.; Chang, J.-S.; Hwang, Y.-K.;
Lebedev, O.; Turner, S.; Tendeloo, G. V. J. Mater. Chem. 2011, 21, 12261233.
 1.33 Precursor Chemistry Main Group Metal Oxides
D Pugh, L Bloor, S Moniz, and CJ Carmalt, University College London, London, UK
2013 Elsevier Ltd. All rights reserved.

1.33.1 Introduction 1021

1.33.2 Precursors for the Growth of Group 13 Oxides 1022
1.33.2.1 Introduction 1022
1.33.2.2 Aluminum Oxide Precursors 1022
1.33.2.3 Gallium Oxide Precursors 1025
1.33.2.4 Indium Oxide Precursors 1027
1.33.2.5 Thallium Oxide Precursors 1028
1.33.2.6 Ternary Group 13 Oxide Precursors 1029
1.33.3 Precursors for the Growth of Group 14 Oxides 1030
1.33.3.1 Introduction 1030
1.33.3.2 Germanium Oxide Precursors 1030
1.33.3.3 Tin Oxide Precursors 1031
1.33.3.4 Lead Oxide Precursors 1033
1.33.3.5 Precursors to Ternary Group 14 Oxides 1036
1.33.4 Precursors for the Growth of Group 15 and 16 Oxides 1036
1.33.4.1 Introduction 1036
1.33.4.2 Antimony Oxide Precursors 1037
1.33.4.3 Bismuth Oxide Precursors 1039
1.33.4.4 Tellurium Oxide Precursors 1042
1.33.4.5 Ternary Group 15 Oxide Precursors 1044
1.33.5 Conclusion 1048
Note added in proof 1048
References 1048

Nomenclature LPCVD Low-pressure CVD

AACVD Aerosol-assisted CVD Mes Mesityl, 2,4,6-trimethylphenyl
Acac Acetylacetonate MOCVD Metal-organic CVD
AFM Atomic force microscopy o-tol ortho-toluene
ALD Atomic layer deposition PECVD Plasma-enhanced CVD
APCVD Atmospheric pressure CVD p-tol para-toluene
Bdk b-diketonate RBS Rutherford backscattering spectroscopy
Bzac Benzoylacetonate SEM Scanning electron microscopy
CVD Chemical vapor deposition TCO Transparent conducting oxide
Dipp 2,6-diisopropylphenyl TEM Transmission electron microscopy
DMAP 4-dimethylaminopyridine tfac 1,1,1-trifluoroacetylacetonate
EDX analysis Energy dispersive X-ray analysis TGA Thermogravimetric analysis
fod 6,6,7,7,8,8,8-heptafuoro-2,2-dimethyl-3,5- thd 2,2,6,6-tetramethylheptane-3,5-dionate
octanedione XPS X-ray photoelectron spectroscopy
hfac 1,1,1,5,5,5-hexafluoroacetylacetonate XRD X-ray diffraction
ITO tin-doped indium oxide

input these are called single-source or molecular precursors

1.33.1 Introduction
In recent years, due to their wide range of applications, there
has been much interest in the production of main-group metal
oxide thin films. A variety of well-defined molecular com-
pounds exist, which contain a preformed MdO bond
(M Al, Ga, In, Tl, Ge, Sn, Pb, Sb, Bi, and Te) that can be
used in the deposition process without an additional oxygen
(for precursors to other materials, see Chapters 1.31, and
1.32). Single-source precursors are ideal for producing main-
group metal oxide thin films because they provide a clean and
simple route to these materials that eliminates the need for a
mixture of precursors which can often be toxic and/or expen-
sive, as well as involving complicated gas-phase reaction dy-
namics which can result in the formation of nonstoichiometric
films. Main-group metal oxides find application as gas sensors,

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00134-0 1021

1022 Precursor Chemistry Main Group Metal Oxides
transparent conducting oxides (TCOs), ferroelectrics, and in op-
toelectronics. Typical precursors for the deposition of thin films
of these metal oxides include metal alkoxides and b-diketonates,
although a range of compounds have been employed.
A variety of methods to deposit these films exist including
chemical vapor deposition (CVD), atomic layer deposition
(ALD), solgel, and spin coating methods. The solgel tech-
nique is a quick and simple method for obtaining thin films.
Sols are either dip- or spin-coated followed by an annealing
stage to obtain metal oxide thin films. CVD is a more sophis-
ticated technique than other methods that allows for larger
area of growth, superior conformal coverage, and stoichiomet-
ric films. Due to a large variety of variables and different types
of CVD, the properties of the thin films can be fine-tuned.
Generally, it is important to have volatile precursors for use
in CVD and the reaction chemistry can be complex. These
include two stages: solid/solution-based precursorgas-phase/
aerosol precursor reactions and gas-phase/aerosolsubstrate
reactions. Research into the reaction mechanism chemistry
involved in CVD has been reviewed by van Mol et al.1 ALD is
a thin-film growth method utilizing alternating, self-limiting
chemical reactions between gaseous precursors and a solid
surface to deposit materials in an atomic layer-by-layer fashion.
This method can produce films with exquisite control over
thickness and composition, and allows precise coatings to be
applied on all exposed surfaces of nanoporous substrates such
as powders or mesoporous membranes.2 If the self-limitation
is successful, the films will be uniform over large areas and
have a thickness that is proportional to the number of com-
pleted cycles, thus giving great control over the final properties
of the film.3
We present here an overview of recent developments in the
formation of main-group metal oxide films from precursors
and also some mention of nonmetal and metalloid oxides of
the main-group elements. A number of reviews have covered
some areas of main-group metal oxides and these are refer-
enced in the appropriate sections.
1.33.2 Precursors for the Growth of Group 13 Oxides
1.33.2.1 Introduction
Thin films of the binary group 13 oxides have varying and
interesting properties, making them extremely useful materials.
From boron at the top of the group, which forms a non-
conducting oxide, to thallium at the bottom, oxides of which
are key components of high-temperature superconductors, the
group 13 oxides have a fascinating range of chemistry (see
Chapter 1.10 for further examples of the chemistry of group
13). This section of the chapter is subdivided into four parts, one
for each metallic member of group 13. The section focuses on the
chemistry of precursors to the thin films although mention is
made of techniques (such as ALD), which tend to utilize simple,
commercially available precursors in combination with a second-
ary oxygen source. Additionally, although boron is a nonmetal
(and thus technically falls outside the scope of this chapter), there
have only been a couple of reports of the deposition of B2O3 and
they are briefly discussed here.
Boron oxide has a low heat expansion and a high refractive
index, making it useful in the ceramics industry when
combined with other oxide materials. However, as a pure
material it readily absorbs atmospheric water to form boric
acid. Moon et al. used triisopropylborate as a single-source
precursor to deposit thin films of B2O3 by low-pressure CVD
(LPCVD).4 Deposition occurred onto a Si(100) substrate at
temperatures ranging from 500 to 650  C and at a pressure of
1.2 mTorr. Growth rates were highest at temperatures between
550 and 650  C and it was also found that the crystallinity of
the resulting films was highest at 650  C. Water contact angle
measurements were taken at 550, 600, and 650  C with the
films deposited at lower temperatures having a lower contact
angle when measurements were originally taken. However,
over the course of the experiment, the water droplet hydro-
lyzed the surface of the B2O3 film to form the extremely hy-
drophilic B(OH)3, resulting in the contact angle reducing
below 10  C in all cases.
Putkonen and Niinisto used BBr3 and water in an ALD
process to deposit amorphous B2O3 films on Si (100) and
soda glass substrates at temperatures ranging between 20 and
400  C.5 The X-ray diffraction (XRD) pattern of the as-deposited
B2O3 films was consistent with that of boric acid, but by
depositing a small protecting layer of Al2O3 over the B2O3,
no XRD spectrum was observed, consistent with the formation
of amorphous B2O3. These two syntheses were the only
reported depositions of pure B2O3 thin films, most other re-
ports of boron-containing films concentrate on either other
materials (e.g., BN) or multimetallic films which fall outside
the scope of this chapter.
1.33.2.2 Aluminum Oxide Precursors
Aluminum oxide forms a number of crystalline polymorphs,
with the most stable a-phase complemented by six metastable
phases. A comprehensive description of the phases and their
solid-state structures can be found in a review by Wu et al.6 Thin
films of Al2O3 have been deposited by a range of methods, with
a variety of different precursors. However, the precursor chem-
istry itself is extremely limited, with all the Al-containing pre-
cursors used being commercially available. A small amount of
synthetic work was undertaken in some cases, which will be
highlighted at the relevant point, but an overview of the synthe-
sis of single-source precursors will also be included for consis-
tency with the rest of the chapter.
A large volume of work has been carried out using ALD as a
technique to deposit thin films of Al2O3. While it is a dual-
source method, a volatile Al-containing precursor is required
and a brief survey of the Al-containing compounds used will be
given. Although it is not necessary for ALD precursors to be
single-source, some ALD reports have utilized compounds that
can be considered as single-source precursors; where this
occurs, it will be discussed with the relevant class of single-
source precursor. What follows is a brief overview of the types
of Al-containing compounds that can only be used in dual-
source methods.
The most commonly utilized Al-containing precursors are
aluminum alkyls, AlR3, (R Me, Et, etc.), with over 20 papers
using them in combination with an oxygen source (commonly
O2, O3, or H2O).7,8 However, any oxygen-containing molecule
may be used as an oxygen source, as evidenced by the work of
Lee and Jeon who used a nitrous oxide plasma as the oxygen

source with AlMe3.9 The Al-containing precursor may be any
volatile Al-containing molecule. Katamreddy and coworkers
used [Al(NR2)3] (R Me, Et) in combination with both water
and O2 to deposit thin films of Al2O3.10,11 However, the ulti-
mate in light, volatile, Al-containing precursors is alane, AlH3.
Although unstable at room temperature, complexing alane to a
Lewis base results in the formation of thermally stable adducts.
Jeong et al. used dimethylethylamine as the Lewis base, with
[AlH3(NMe2Et)], a suitably volatile Al-containing precursor for
the deposition of Al2O3.12,13
The category of single-source aluminum precursors can be
divided into two main areas: simple monodentate alkoxides
and b-diketonates. There are many commercially available
aluminum alkoxides to choose from, with the most commonly
used one being [Al(OiPr)3]. This was used by Raisanen and
coworkers in the ALD of Al2O3. Interestingly, the alkoxide
served as both a precursor and an oxygen source, with AlCl3
being the second precursor used. This methodology avoided
the use of highly oxidizing oxygen sources (such as O2 or
H2O), which can react with certain substrates, notably silicon,
forming an unwanted surface oxide layer in the initial stages of
deposition.14 The precursor [Al(OiPr)3] was also used to form
an Al2O3 sol and dip coating onto a TiAl alloy. The alkoxide
was dissolved in water before acidification with HNO3 formed
the Al2O3 sol. Between each dip, the substrate was dried in air,
then dried in a vacuum oven for 40 min at 75  C. Annealing at
1000  C afforded crystalline films of g-Al2O3.15
The CVD of [Al(OiPr)3] has been carried out by several
groups. The majority of work has been reported since 2000
when Niska et al. used Al2O3 as one layer of a dielectric
material, which was deposited onto a NiCoCrAlY alloy (in
the form of gas turbines for jet engines) using an LPCVD
process. The precursor was volatilized at 0.9 Torr and carried
into the deposition chamber using a small gas flow. The sub-
strate was heated between 700 and 1100  C, with the desired
a-phase of Al2O3 only forming at temperatures above 1000  C.
The best deposition conditions were found when O2 was used
as the carrier gas; using argon resulted in a higher level of
carbon contamination of the film.16
In 2007, both Gleizes et al. and Huntz et al. reported the use
of [Al(OiPr)3] as a precursor to thin films of Al2O3. The former
found that in a single-source LPCVD process at 110  C and
5-Torr pressure, Al(O)OH was deposited on silicon and stain-
less steel at 350  C but as the temperature was increased,
increasing amounts of Al2O3 were formed. The transition to
pure amorphous Al2O3 occurred at 415  C but crystallization
to the g-phase of Al2O3 did not occur until 700  C.17 A more
detailed investigation into the mechanism of formation was
carried out by Huntz et al., who discovered that 710  C was the
temperature at which crystallization to the g-phase of Al2O3
occurred. Conversion of the metastable g-phase to the thermo-
dynamically stable a-phase was observed to occur at 900  C by
XRD, but the onset of conversion was predicted to occur
around 810  C.18
By changing from isopropyl groups to sec-butyl groups on
the alkoxide ligand, another commonly used precursor can be
made, namely [Al(OsecBu)3]. Whereas [Al(OiPr)3] has mostly
been utilized after 2000, the sec-butyl analog has mostly been
used before 2000, and thus falls outside the scope of this
chapter. However, in 2001, Kuo et al. used [Al(OsecBu)3] to
Precursor Chemistry Main Group Metal Oxides 1023
deposit thin films of Al2O3 on silicon and glass by an LPCVD
process at 1 Torr. Heating the precursor to 138  C with a small
flow of argon as the carrier gas led to the deposition of amor-
phous Al2O3 films at substrate temperatures between 300 and
500  C.19 Six years later, Jing et al. used the same precursor to
synthesize an Al2O3 sol in water prior to spin coating onto a
stainless steel substrate. After the final cycle of spin coating, the
films were annealed in air at 1000  C, forming the a-phase of
Al2O3.20 An adduct of [Al(OsecBu)3] with AlEt3 was used by
Parsons and coworkers to deposit Al2O3 by LPCVD. The pre-
cursor was heated to 150  C and carried into the reactor by a
stream of argon carrier gas, with a small amount of O2 as
oxidant. The substrate (silicon) was heated to 400  C, which
resulted in the deposition of amorphous Al2O3 and the film
composition was confirmed by X-ray photoelectron spectros-
copy (XPS).21
The only other tris(alkoxide) of aluminum which has been
used as a precursor is [Al(mmp)3] (mmp 1-methoxy-
2-methyl-2-propoxide). This was used in the ALD of Al2O3
with water as co-precursor and the substrate (silicon) was
heated to temperatures between 250 and 450  C. Above
350  C, the water precursor was not used owing to decompo-
sition of the Al precursor to Al2O3. As with all Al2O3 films
grown at low temperature the film was amorphous, but XPS
analysis of the film confirmed that Al2O3 had been grown.22
Aluminum mono(alkoxides) have not been used as fre-
quently as the tris(alkoxides) but a few reports have appeared
using compounds of formulation [R2Al(OR0 )]2 (1, R Me,
R0 iPr; 2, R R0 Et) as precursors to Al2O3 thin films. The
group of Kim published the use of [Me2Al(OiPr)]2 (1) as a
precursor to aluminum oxide thin films via ALD.23 This was
conveniently synthesized in one step from commercially
available AlMe3 by reacting the trialkylaluminum compound
with 1 equiv. of alcohol (eqn [1]). The silicon substrate was
heated to 150  C and alternating pulses of precursor and water
were used, affording uniform thin films of amorphous Al2O3.
Auger electron spectroscopy confirmed the Al:O ratio of 2:324:
R R R
O
2 AlR3 + 2 ROH Al
Al
-2 RH
O 1
R R
R
1: R = Me; R = iPr; 2: R = R = Et
Guidi et al. deposited Al2O3 on a brass substrate at 380  C
from 1 using an LPCVD (2 Torr) process. Compound 1 was
carried into the reactor by a nitrogen flow and mixed with an
O2/H2O reactant gas prior to deposition. Powder XRD (PXRD)
of the resulting films indicated that they were amorphous, as is
expected from such a low deposition temperature. Rutherford
backscattering spectroscopy (RBS) indicated that the films
formed were pure Al2O3.25 Similarly, the group of Rogers depos-
ited aluminum oxide on silicon at temperatures of 417659  C
in an LPCVD process at ultra-high vacuum using 1 (0.2 mTorr).
The films grown were all amorphous, but interestingly the films
grown at lower temperatures had lower levels of carbon contam-
ination than the films grown at higher temperatures as measured
by XPS.26
A related compound [Et2Al(OEt)]2 (2) was used in an un-
usual CVD technique, electron-cyclotron-resonance-assisted
1024 Precursor Chemistry Main Group Metal Oxides
CVD. This process is carried out at extremely low pressure
(1 nTorr) with deposition occurring on both silicon and
GaN. The key difference to other LPCVD processes is the for-
mation of plasma created by microwave energy which assisted
the deposition. XPS of the resulting film indicated that the ratio
of Al:O was 1:1.48, very close to the expected 1:1.5. There was
no detectable carbon contamination, ascribed to the addition
of O2 to the plasma. As expected for films deposited at low
temperatures, they were amorphous.27
The use of aluminum b-diketonates is largely restricted to
just one compound, namely [Al(acac)3] (3). This is by far the
most widely used b-diketonate-based precursor to thin films
of Al2O3 and, since 2000, CVD-based techniques have been
the primary method used. However, Ortiz et al. used a
pyrosol process to deposit Al2O3 from a solution of 3 in water/
methanol (3:1). Depositions onto glass and quartz were con-
ducted at 480 and 510  C resulting in amorphous oxygen-rich
films, as verified by RBS.28 The group of Falcony used spray
pyrolysis to deposit Al2O3 from a DMF solution of 3 onto silicon
substrates heated to temperatures between 450 and 650  C. The
solution was ultrasonically vaporized before being sprayed onto
the substrate with air as a carrier gas. The resulting films were
amorphous and oxygen-rich, but the addition of a small amount
of water vapor into the carrier gas improved the film stoichiom-
etry to the desired Al2O3.29
Although it is commercially available, [Al(acac)3] (3) was
synthesized by Singh and Shivashankar prior to its use in
LPCVD experiments (eqn [2]). Depositions were carried out
on silicon, stainless steel, and TiN-coated cemented carbide at
temperatures ranging from 500 to 600  C using compound 3.
The precursor (3) was sublimed at 120  C and carried into the
reactor with a flow of argon. The films were slightly black as-
deposited, indicating carbon contamination (confirmed by XPS
data). This is somewhat in contrast to other groups who used 3
as a precursor where transparent, oxygen-rich alumina films
were deposited. Another oddity was the observation of polycrys-
talline Al2O3 at deposition temperatures as low as 500  C,
something which had not previously been observed. The au-
thors ascribed this to the carbonaceous nature of the films,
providing nucleation sites for the growth of alumina
microcrystals30:
R
R deposition temperature, all the films were amorphous but
O O O
O
R transparent alumina.36
O
Al(NO3)3 + 3 Al
R R 3 HNO3 O O R 2
O of the two categories above, there are a few other precursors
R
R
3: R = Me, acac; 4: R = CF3, hfac
Pradhan et al. also used compound 3 in an LPCVD process to
form thin films of Al2O3. Depositions were carried out at 50 Torr
on a silicon substrate heated to temperatures between 350 and
950  C. Amorphous films were obtained from temperatures
below 550  C, whereas films grown between 550 and 750  C
showed the onset of crystallinity within a mainly amorphous
film. Between 750 and 950  C, the film was mainly crystalline,
but indexing of the XRD pattern was only possible at 950  C,
where the metastable k-phase of alumina was observed.31
Atmospheric pressure CVD (APCVD) was also carried out
using 3, with the precursor being sublimed at 140  C and
carried to the stainless steel substrates by a flow of air at AP.
The reactor temperature was increased from 147 to 597  C with
the onset of deposition occurring about 327  C. This was noted
by a CO2 absorption band in the IR spectrum of the exhaust
gas, which grew in intensity until ca. 447  C whereafter it
remained constant. This indicated that the optimum substrate
temperature for deposition to occur was about 450  C, but
even at temperatures as high as 500  C no crystallinity was
observed in the films. Annealing under vacuum at 800, 970,
and 1100  C slowly increased the crystallinity of the film, with
the g-phase being observed. At the highest temperature, con-
version to the a-phase of alumina was observed but delamina-
tion of the film also occurred.32 Parsons and coworkers used
supercritical CO2 to dissolve compounds 3 and [Al(hfac)3] (4)
and deposit Al2O3 onto silicon substrates. The deposition tem-
perature was only 150  C for 3 (a slightly higher deposition
temperature of 310  C was used for 4) and IR spectroscopy of
the film showed residual absorption peaks for the b-diketonate
ligand. However, annealing in air at 600  C greatly reduced the
residual ligand peaks in the IR spectrum.33
Laser-assisted CVD was carried out by Ito et al. who used
both yttria-stabilized zirconia and AlN as substrates, preheated
from room temperature to 800  C. Compound 3 was sublimed
at 170  C and carried into the deposition chamber by a flow of
argon, where it was mixed with a small amount of O2 and
irradiated with a laser. At 400  C substrate temperature,
crystallinity was observed: the degree of crystallinity (and the
phase of alumina obtained) depended upon the power of the
Nd:YAG laser irradiation, with a 238-W beam affording crys-
talline a-phase Al2O3. The Nd:YAG laser irradiation gave an
effective deposition temperature of 960  C, still well below the
usual temperature at which a-Al2O3 is observed.34 Switching
to a diode laser resulted in the lowest temperature at which
a-Al2O3 was observed dropping to 655  C, significantly lower
than any other method used to date.35
The only other b-diketonate ligand utilized to deposit alu-
mina films is ethyl acetoacetate. Yamaguchi et al. dissolved
[Al(OsecBu)3] in iPrOH, water, and ethyl acetoacetate
(1:20:1:1 ratio) and dip-coated the solution onto polymer
substrates. The films were dried at 50  C then analyzed, with
some films needing plasma irradiation to improve the adher-
ence of the film to the substrate. As is expected from such a low
While the majority of Al-containing precursors fall into one
that warrant mention but are not classed as either alkoxides or
b-diketonates. The AlMe3/O2 combination described above for
ALD has also been used in dual-source CVD methods. Ogita
et al. used a catalytic CVD process to deposit alumina thin
films on silicon substrates heated to 1000  C.37 The AlMe3
was carried into the reactor using a stream of nitrogen gas
and passed through a 0.2-mm-diameter tungsten filament
(the catalyst) heated to 1000  C prior to mixing with oxygen
and depositing alumina on the substrate. It was found that the
film thickness was strongly dependent on the catalyzer
temperature, with thicker films being formed at higher catalyst
temperatures. Unusually, however, the films were amorphous:
normally films deposited at 1000  C show crystallinity with a
metastable alumina phase. A later study using mass

spectrometry to analyze the exhaust gases elucidated that the
catalyst allowed aluminum to separate from the AlMe3 pre-
cursor, prior to combining with O2 to form AlO and depositing
alumina onto the substrate. It was also discovered that high
substrate temperatures were not necessary, with deposition
occurring on a substrate which was not heated at all (room
temperature).38 The same precursor combination was also
used by Szymanski et al. to deposit Al2O3 films on silicon
substrates at ambient conditions. The method utilized was
plasma-enhanced CVD (PECVD), with argon serving as the
carrier gas for the AlMe3, which was stored in a bubbler
at 1  C. A quadrupole mass spectrometer was used to analyze
the emission from the plasma, with O, H, CO, Al, Ar, AlO, and
O2 being the prominent products detected. IR spectroscopy of
the resulting films confirmed the formation of alumina.39
Finally, aluminum tris(diisopropylcarbamate) [Al(O2CNi
Pr2)3] (5) was used as a single-source precursor for the deposi-
tion of alumina by CVD and solvothermal methods. Lamb and
coworkers synthesized the precursor by saturating a benzene
solution of diisopropylamine with CO2 before introducing
AlCl3. Continuous bubbling of CO2 through the reaction mixture
eventually yielded a colorless solution with a white precipitate,
which was filtered and discarded. The solution was evaporated to
dryness, affording precursor 5 in a 71% yield (eqn [3])40:
i
Pr2N
O precursors to gallium oxide in the style of the aluminum sec-
O tion (Section 1.33.2.2), readers are directed to the review in
AlCl3 + HNiPr2(xs) + CO2(xs) Al
O
NiPr2
O O
O covering the main classes of precursor to gallium oxide.
i
Pr2N 5 Single-source precursors to thin films of Ga2O3 can be
Compound 5 was initially used in LPCVD experiments
depositing onto silicon substrates at 450550  C. It was
thought that the octahedral coordination environment at
aluminum in 5 and the clean decomposition pathway of the
carbamate ligand (Scheme 1) would lead to the clean forma-
tion of alumina films at low temperatures.
Films deposited from 5 at substrate temperatures of 500  C
were amorphous, but carrying out the deposition at 550  C
resulted in the detection of one peak in the XRD spectrum. This
corresponded to the 1 1 3  phase of a-Al2O3, which is not
commonly detected until >1000  C. Annealing the amor-
phous films at 850  C led to the formation of d-Al2O3. Com-
pound 5 was also used in the solvothermal synthesis of
alumina. The precursor was dissolved in benzene and a silicon
wafer was immersed in the solution, which was then heated to
iPr N
2
O O C NiPr2
O O
O i
Al N Pr2 Al
O O
O O
i
Pr2N 5 O C N
Scheme 1 Decomposition pathway of compound 5.
Precursor Chemistry Main Group Metal Oxides 1025
temperatures between 65 and 105  C in an autoclave. Clean,
carbon-free films of amorphous alumina were obtained at all
temperatures as confirmed by XPS data.41
1.33.2.3 Gallium Oxide Precursors
Gallium oxide exists solely as the Ga(III) oxide, Ga2O3. Thin
films of Ga2O3 are electrically insulating at room temperature,
but become semiconducting above 450  C with their electrical
resistivity varying according to the concentration of reducing
gases (e.g., CO and EtOH) in the atmosphere.42,43 When heated
above 900  C, the electrical resistivity of the film becomes sen-
sitive to the concentration of O2 in the atmosphere,44,45 thus
gallium oxide thin films can function as a gas sensor for both
reducing and oxidizing gases depending on temperature.
Gallium oxide has many crystalline forms, but the most com-
monly encountered is the monoclinic form, often referred to as
the b-form. This is usually the result of annealing films at 900  C
or above, although crystalline monoclinic films have been grown
at temperatures as low as 600  C.46,47 A rhombohedral form is
sometimes observed at lower temperatures (500  C), although
most films grown at this temperature are amorphous, with con-
version to the monoclinic form occurring around 700  C.48
A comprehensive review on single-source precursors to gal-
lium oxide covering the literature to the end of 2010 was
recently published.49 For an in-depth analysis of single-source
3
Bloor et al.49. What follows in this section is a brief overview
divided into two broad classes: b-diketonates and alkoxides.
The b-diketonates are all of the form [Ga(bdk)3] (69) where
bdk b-diketonate (eqn [4]). These compounds are air and
moisture stable, owing to their nature as a six-coordinate octa-
hedral species with a Ga:O ratio of 1:6. The stability of the
homoleptic b-diketonates is evidenced from attempts to syn-
thesize complexes of the type [GaX(bdk)2] (bdk b-diketonate;
X Cl, H, Me): no evidence for the formation of the hetero-
leptic bis(b-diketonate) complexes was observed, with the
thermodynamically stable tris(b-diketonates) being the only
products isolated.50 Some of the simpler diketonates, for ex-
ample, [Ga(acac)3] (6), are commercially available, albeit
expensive, and especially so when the synthesis of the diketo-
nates from GaCl3 is straightforward (eqn [4]).33,51,52 Although
potential precursors of the type [GaMe2(acac)] have been syn-
thesized, they have not been used to deposit Ga2O3:
NiPr2
iPr i
Pr
O C N
CHMe
H
H
H
1026 Precursor Chemistry Main Group Metal Oxides

R
Although these exist as either dimers or tetramers depending
R
M O
R
on the steric bulk of the alkoxide, they can be split into mono-
O O O
GaCl3 + 3 3 MCl Ga
O mers by coordination of a Lewis base (Scheme 2).54 This is
O
R R O
O
R
4 important for techniques such as CVD, which rely on the pre-
M = Li, Na R
cursors having good volatility.
R
Donor-functionalized alkoxides incorporate Lewis bases,
6: R = Me, acac; 7: R = tBu, thd
8: R = Ph, dbm; 9: R = CF3, hfac which enable the formation of monomeric complexes (e.g.,
1719, 22, and 23). Many different ligand architectures are
Gallium alkoxides can be further divided into two sub- plausible, but the most widely used have been ethanol-based
categories, namely, monodentate alkoxides and donor- ligands with either NMe2 or OMe Lewis bases in the
functionalized alkoxides. The former consist of simple gallium 2-position. These ligands can be attached to gallium in various
alkoxides such as methoxide and iso-propoxide (1014), ways, as summarized in Scheme 3. While it is possible to
which can easily be synthesized in a number of ways.53 obtain a wide range of ligands with varying chain lengths,

RO
OR
Me Ga
Me H 6 ROH RO
N RO R OR RO R OR
O Ga O Ga
Ga Ga
Ga OR O RO O OR
RO
[Ga(NMe2)3]2 RO R RO OR R
OR
Ga OR 13: R = iPr
10: R = CH(CF3)2
11: R = CMe2CF3 OR 14: R = tBu
12: R = CMe(CF3)2

Scheme 2 Synthesis of gallium alkoxides.

GaMe3 Me L L
O O
Me Ga O Ga Me Me Ga
L O
x HO L Me L

15: L = OMe 17: L = OMe

x = 1 or excess 18: L = NMe2
16: L = NMe2

GaCl3 CF3
Me2N CF3
2 NaH O
F3C CF3 Cl Ga
2 -2 NaCl O
NMe2 Me2N CF3
HO -2H2
CF3
19

Cl L
O 20: L = OMe
Cl Ga O Ga Cl 21: L = NMe2
L Cl
x=1
[Cl3-xGa(NMe2)x]2
L
x=2 O 22: L = OMe
Cl Ga 23: L = NMe2
NMe2 O
2x HO L
x=3
RO L
O 24: L = OMe
RO Ga O Ga OR 25: L = NMe2
L OR

Scheme 3 Gallium donor-functionalized alkoxides incorporating Lewis bases.


chain substituents, and Lewis base substituents, for conve-
nience only two ligands will be used as representative exam-
ples: HOCH2CH2NMe2 and HOCH2CH2OMe.
The reaction of donor-functionalized alkoxides with GaMe3
affords either pure gallium mono-alkoxide products or a mix of
gallium mono- and bis-alkoxide products, depending on the
ratio of ligand to GaMe3. A 1:1 ratio of reactants results in a
mono-alkoxide (15, 16), whereas an excess of alkoxide affords
a mixture (17, 18). The formation of gallium tris-alkoxides by
this method has not been observed.55
Reacting the sodium salt of a trifluoromethylated donor-
functionalized alkoxide directly with GaCl3 has been observed
to afford an air-stable gallium bis-alkoxide complex 19.56
However, this method is specific to the fluorinated alkoxide
noted in Scheme 2. Attempts at making it general to all donor-
functionalized alkoxides led to difficulties in controlling the
stoichiometry of the reaction and a mixture of mono- and bis-
alkoxides was isolated which could not be easily purified.
The formation and isolation of gallium mono-, bis-, and
tris-alkoxides are possible using the amine elimination route
(Scheme 2); indeed, all have been synthesized and used as
precursors.57,58 However, only the bis-alkoxides (22, 23) are
actually monomers, with the mono- and tris-alkoxides forming
dimeric oxygen-bridged species. While this is unsurprising for
the mono-alkoxides (20, 21), where coordinative saturation of
the gallium center is not achieved through the use of only one
extra Lewis base, it is surprising for the tris-alkoxides (24, 25).
Although crystallographic evidence for the formation of a
dimer has not been obtained, indirect evidence from tech-
niques such as mass spectrometry and IR spectroscopy indi-
cates that the tris-alkoxide complexes are dimeric in nature
with a Ga2O2 ring at the center of the molecule. However,
one recent article suggests that the formation of monomeric
tris-alkoxides is possible, starting from the gallium mono-
alkoxide dihydride dimer 26 and reacting with 2 equiv. of
alcohol (eqn [5]).59 While the single-crystal X-ray dataset was
merely sufficient to show the atom connectivity, it confirmed
the monomeric nature of complex 27 as a five-coordinate
species with one dangling Lewis base. The 1H NMR spectrum
of 27 was broad, indicating probable fluxionality on the NMR
timescale, with the dangling amine exchanging with other
donor-functionalized ligands attached to the gallium center:
H NMe2 Me2N
O 2 HOCH2CH2NMe2 O
H Ga Ga H Me2N
O O Ga
Me2N H Me2N
O
26 27
Since 2000, ALD has not been used very often to deposit
films of Ga2O3. Most of the gallium-containing precursors that
have been used are derivatives of GaMe3 with either alkoxide or
amide groups replacing one of the methyl groups. Lee et al.
used [GaMe2(OiPr)]2 as a precursor (along with H2O as oxygen
source) to deposit gallium oxide using both ALD and CVD.60
The films were stoichiometric Ga2O3 but were found to be
amorphous, as expected from the low deposition temperature.
The compound [GaMe2(NH2)]2 has also been used as an ALD
Precursor Chemistry Main Group Metal Oxides 1027
precursor, in combination with an oxygen plasma,61 to deposit
thin films of Ga2O3 on Si(100) substrates.62,63 The as-deposited
films were found to be smooth (by atomic force microscopy
(AFM) analysis) but amorphous. However, after annealing at
700  C, the films had started to crystallize and it was possible
to obtain an XRD pattern which corresponded to b-Ga2O3.
ALD growth of Ga2O3 films was also achieved using [Ga
(NMe2)3]2 and water at substrate temperatures between 150
and 300  C.64 The as-deposited films were amorphous but
upon annealing between 700 and 900  C under a nitrogen
atmosphere crystallized to b-Ga2O3 films.
1.33.2.4 Indium Oxide Precursors
Although the chemistry of indium is not dominated by the 3
state to the same extent as the lighter members of group 13, with
a few commercially available In(I) compounds, thin films of
indium oxide still exist solely as the In(III) oxide, In2O3. These
films of In2O3 belong to the TCO family of materials, which, as
the name suggests, are transparent to visible light and conduct
electricity. The TCO properties of In2O3 can be improved in
various ways, but the most well known one is by doping with
tin; indium tin oxide (ITO) is the industry standard for TCO
materials (Section 1.33.2.6). This gives In2O3 film applications
in areas such as solar cell devices and optoelectronics.
The main crystalline form of indium oxide, and the one
which is most commonly encountered, is the cubic form.
Crystalline films form at much lower temperatures than for
gallium oxide, with the onset of crystallization usually occur-
ring around 300  C. However, substrate temperatures of up to
600  C have been used, with 450  C proving to be most effi-
cient for indium donor-functionalized alkoxide complexes.
As is the case for gallium oxide precursors, a comprehensive
review covering the literature to the end of 2010 regarding
single-source precursors to In2O3 was published.49 Readers are
directed to this review for an in-depth analysis of single-source
precursors to indium oxide (in the style of Section 1.33.2.2),
including each individual precursor, in addition to deposition
conditions and techniques used. What follows in this section is a
general overview of the area of precursors to thin films of In2O3.
Similarly to gallium oxide, single-source precursors to In2O3
can generally be divided into two categories: b-diketonates
and alkoxides. Fewer b-diketonates have been used for indium
oxide than for gallium, and those that have been employed
turned out to have remarkable thermal stability. For example,
5
[In(acac)3] (28) does not sublime until 180  C.65 Attempts to
reduce the thermal stability of the homoleptic complexes by
synthesizing compounds, of the type [In(bdk)2Cl] (bdk b-
diketonate), were foiled by the thermodynamic stability of the
tris(b-diketonates) 28 and 29 which were isolated preferentially
from all reactions (eqn [6])50:
R
R R
O
2 Li(acac) O O
In
InCl3 or
O O 6
2 Na(thd)
-2 LiCl R O R
or
-2 NaCl R
28: R = Me, acac; 29: R = tBu, thd
1028 Precursor Chemistry Main Group Metal Oxides
2In(HMDS)3 RO
+ In
6ROH RO
HMDS = 1,1,1,3,3,3-hexamethyldisilazide
InMe3 Me
+ O
Me
L
HO
Scheme 4 Synthesis of indium alkoxides.
Indium alkoxides can also be subdivided into monodentate
alkoxides and donor-functionalized alkoxides. As for the gal-
lium case, donor-functionalized alkoxides will be represented
by two generic examples: 2-dimethylaminoethanol and 2-meth-
oxyethanol. The use of alkoxides in general has been extremely
limited when compared to the case for gallium. This may in part
be due to the fact that homoleptic monodentate indium alkox-
ides are not commercially available, unlike their gallium conge-
ners. Another issue is that many literature routes to homoleptic
indium alkoxides have been proven to form indium cluster
compounds with a central oxo anion, instead of well-defined
alkoxide complexes.66 However, perhaps the main drawback
is that the chemistry used to synthesize gallium alkoxides
(both monodentate and donor-functionalized) is not always
applicable to indium, especially bis-alkoxide complexes of
indium. Nonetheless, the amine elimination route does lead
to tris-alkoxide complexes of indium starting from indium tris-
hexamethyldisilazide (Scheme 4: 3033).67
Whereas gallium mono- and bis-donor-functionalized
alkoxides can be made by reacting GaMe3 with a donor-
functionalized alcohol, only mono-alkoxides of indium (34,
35) can be synthesized by reacting InMe3 with alcohols.68 The
reason for this is not clear, but it has not yet proven possible
to synthesize indium bis-alkoxide complexes of donor-
functionalized alkoxides from the starting material InMe3.
Bis-alkoxide complexes of gallium can be easily synthesized
by employing the amine elimination route with a gallium
bis-dimethylamido chloride starting material. However, for
indium, the use of the amine elimination route has proven
problematic because of the paucity of suitable indium precur-
sors. The compound [In(NMe2)2Cl]2 is thermally unstable and
cannot be used as a starting material. Attempts to bulk up the
amide ligand in order to prevent decomposition centered
around the use of the hexamethyldisilazide (HMDS) ligand.
However, when 2 equiv. of LiHMDS were reacted with InCl3,
an unusual methyl migration from the silylamide group to
indium occurred, forming the dimeric species [In(Cl)Me
(HMDS)]2 (36, see eqn [7])69:
Me R2
Cl
N
2InCl3 + 4LiHMDS In In
Cl N Me 7
R2
36: R = SiMe3
Much like the case for gallium, there have been relatively
few reports since 2000 concerning the ALD of In2O3. A slightly
R OR 30: R = tBu
O In
31: R = CMe2Et
O OR 32: R = CMeEt2
R
33: R = CMe2iPr
L
34: L = OMe
In In Me
O 35: L = NMe2
L Me
more unusual precursor to In2O3 thin films is the In(I) com-
pound [InCp].70 This can easily be made by reacting InCl with
LiCp, but it is also commercially available. It was initially used
by Elam et al. in the ALD of In2O3, with O3 as oxidant and was
chosen because they found that indium b-diketonate com-
plexes were not good precursors. However, Nilsen et al. suc-
cessfully used 28 as an ALD precursor with O3 or H2O as the
oxygen source to deposit In2O3 onto Si(100) and glass.71
At substrate temperatures below 75  C, the films were amor-
phous, but films grown at 200  C and above were crystalline,
with XRD analysis revealing that the expected cubic phase of
In2O3 had formed.
1.33.2.5 Thallium Oxide Precursors
While the chemistry of the lighter group 13 elements is dom-
inated by the 3 oxidation state, with this being especially true
of the oxides, thallium is rather different. It forms a range of
compounds, including oxides, in both the 1 and 3 oxida-
tion states. Thallous oxide (Tl2O) is a hygroscopic, volatile
substance, thin films of which have limited usefulness because
they readily hydrolyze to thallium hydroxide (TlOH), eventu-
ally absorbing atmospheric CO2 to form thallium carbonate,
Tl2CO3.72
Thin films of thallium(III) oxide, Tl2O3, have received little
attention, perhaps unsurprisingly given the toxicity of thallium
compounds. However, the films do have potential applications
on account of their high optical transparency and low electrical
resistivity (70 mO cm). A prototype photovoltaic cell was even
built using Tl2O3 as a protective layer over an electrode of
n-silicon, using photoelectrodeposition of an aqueous solution
of [TlOAc].73 Currently, however, the most important use of
thallium oxide is as a component in high-temperature super-
conductors: the material Tl2Ba2Ca2Cu3O10 has a Tc of 127 K.74
Although these materials have been deposited using CVD
methods (featuring thallium precursors such as [Tl(thd)]75)
they are multimetallic oxides; thus, they fall outside the scope
of this chapter and will not be covered further.
There are very few published routes to thin films of Tl2O3,
which do not utilize commercially available starting materials.
The most commonly used synthetic precursor is cyclopentadie-
nyl thallium [TlCp] (37), an air- and moisture-stable yellow
solid which can be easily synthesized (Scheme 5), even though
it is slightly light sensitive and slowly decomposes upon expo-
sure to light. Although the initially reported synthesis reliably
affords high yields of the product,76 a later modification from

Tl2(SO4) R Tl(OEt)
+ +
NaOH Hexane
H H R H
Tl
37: R = H
38: R = Me
Scheme 5 Synthesis of [TlCp].
TlMe3 Me Me
Tl 39
O O O O
Tl2(CO3)
L depending on the ligand used in the coordination polymer.
Tl
O O 40: L = diglyme
O O
41: L = tetraglyme
F 3C CF3
F3C CF3
Scheme 6 Synthesis of thallium b-dketonate complexes.
Zanella et al., which avoids the use of water, has also
appeared.77 The former method has been used by both Berry
et al.72 and Daolio et al.78 to synthesize [TlCp] as a precursor to
thin films of Tl2O3.
The main advantage to the use of [TlCp] is its volatility at
low temperatures. Both Daolio et al. and Zanella et al. used
LPCVD to deposit Tl2O3. The precursor was sufficiently volatile
such that, at a pressure of 60 mbar, it was necessary to heat
[TlCp] to only 6070  C in order to get it into the gas phase.
Berry et al. used an APCVD process, which required heating the
precursor to a temperature of 125130  C in order to volatilize
it, but both methods reliably afforded thin films of Tl2O3 with
O2 as an oxygen source.
It is possible to synthesize an analog of [TlCp] by adding a
methyl group to the cyclopentadienyl ring (Cp0 ). The com-
pound [TlCp0 ] (38) was reported by Zanella et al. starting
from thallium (I) ethoxide (Scheme 5), affording a product
which was much more volatile than [TlCp]. However, one
significant drawback to this compound is the fact that it is
much less stable toward O2 than the nonmethylated analog;
thus, a different oxygen source has to be found in order for it to
be a viable CVD precursor.
The b-diketonate complexes [Tl(acac)], which is commer-
cially available, and [TlMe2(acac)] (39) which is not, but is
easily synthesized by reacting TlMe3 with 1 equiv. of Hacac
(Scheme 6), are both potential single-source precursors.
However, when used by Berry et al. oxygen gas was used in
the APCVD procedure as a secondary oxygen source. Likewise,
the compounds [Tl(hfac)(diglyme)] (40) and [Tl(hfac)(tetra-
glyme)] (41) can potentially be used as single-source precur-
sors, but Malandrino et al. used O2 as a secondary oxygen
source. The fluorinated b-diketonate complexes were synthe-
sized by reacting Tl2CO3 with Hhfac in the presence of either
diglyme or tetraglyme (Scheme 6).79
The temperatures required to volatilize the nonfluorinated
b-diketonate derivatives were slightly higher than for [TlCp],
being 135140  C for 39 and 140145  C for [Tl(acac)].
Precursor Chemistry Main Group Metal Oxides 1029
However, the fluorinated b-diketonate complexes were found
to have a much higher volatility, with a melting point of
2527  C for 41 and 3942  C for 40. Compound 41 was
used as a precursor in LPCVD experiments with argon as the
carrier gas and O2 as a secondary oxygen source, depositing
crystalline Tl2O3 films on glass at a substrate temperature of
430  C. Both nonfluorinated b-diketonates were also tried as
APCVD precursors, depositing crystalline Tl2O3 onto a MgO
substrate at 400  C.
Tl(I) coordination polymers have also been used as pre-
cursors to thin films of Tl2O3.80 Reacting TlNO3 with potas-
sium 3-nitrophenoxide (or 2,4-dinitrophenoxide) led to the
isolation of crystalline air-stable white powders, which were
observed to be three-dimensional (3D) polymers from single-
crystal XRD studies. These were calcined at 600  C, affording
crystalline Tl2O3, with the nanostructure of the product
1.33.2.6 Ternary Group 13 Oxide Precursors
Some of the key applications of group 13 oxide films actually
require the use of impurity-doped metal oxides or ternary
metal oxides. This section presents a brief overview of the
recent developments involving precursor routes to thin films
of these types of materials incorporating group 13 metal
oxides. For most optoelectronic devices such as flat panel
displays, it is essential to use a transparent electrode consisting
of a thin film of a TCO semiconductor. The development of
TCO films and their applications have been extensively
reviewed including a panoramic review,81 surveys of current
TCOs,82,83 and discussions on the future of TCOs.84,85 At
present, In2O3:Sn (ITO) films deposited via magnetron sput-
tering deposition methods are in practical use for most thin-
film transparent electrode applications. However, indium is a
scarce metal and the price of indium has risen steadily; hence,
researchers are searching for indium-free TCOs. As of now,
over 10 000 articles have been published on ITO alone and
what follows in this section is a brief overview of ITO and other
TCOs; for more detailed information, readers are directed to
the reviews outlined above. Alternative TCOs include SnO2:F,
In2O3:Ga, and ZnO:Al.
Various methods including CVD,86,87 sputtering,88 dip-
coating,89 solgel,90 spray pyrolysis,91 and electrostatic spray-
assisted vapor deposition92 have been used to deposit ITO
films. The precursor mix employed is usually InCl3 and SnCl4-
5H2O or [In(NO3)3] and hydrated stannic chloride and for
solution-based routes the precursors are dissolved in ethanol
or acetylacetone.93 However, Ni et al. used a precursor mix of
[In(thd)3] and [Sn(acac)2] to deposit ITO films via LPCVD and
Mottern et al. used indium(III) acetate and [Sn(OtBu)4] in
2-methoxyethanol to grow ITO films by spin coating.94 A
novel heterobimetallic single-source precursor, indium tin tri-
tert-butoxide, [Sn(m2-OtBu)3In] (42), was recently utilized by
Aksu et al. to prepare tin-rich mesoporous ITO films.95 Com-
pound 42 was synthesized from the reaction between InBr and
Na(OtBu)3Sn, in the presence of N,N,N0 ,N0 -tetramethylethyle-
nediamine and after heating at 100  C for 24 h. The films were
grown using solutions of 42 by spin- and dip-coating tech-
niques and were found to have a high and durable electrical
conductivity and transparency. Many of the traditional routes
1030 Precursor Chemistry Main Group Metal Oxides
to ITO have resulting problems with phase segregation due to
the limited solubility of Sn in In2O3; however, this single-
source approach allows for high control over the In/Sn molar
ratio. Other dopants for In2O3-based TCOs include Sn, Ge,
Mo, Ti, Zr, Hf, Nb, Ta, W, Te, F, and Zn,96 all of which
produced films with very good optical and electrical properties.
tBu
O tin oxide thin films are doped with other elements to improve
100 C
InBr + Na(OtBu)3Sn Sn In
24 h
O O
t t
Bu Bu
42
Gallium-doped indium oxide (GaxIn2xO3) is the reported
parent phase of a promising, relatively unexplored TCO
family of materials. Indeed, gallium-doped indium oxide has
been shown to be an exceptional material for TCO applications
with transmittance that exceeds all other oxides.97 The GaxIn2
xO3 (x 01.1) were grown by LPCVD using the volatile meta-
lorganic precursors [In(thd)3] (29) and [Ga(thd)3] (7).97 Films
grown on quartz from 7 and 29 were smooth (root mean
square roughness 5065 A) with a homogeneously Ga-
substituted, cubic In2O3 microstructure. More recently,
Knapp et al. utilized the aerosol-assisted CVD (AACVD) reac-
tion of GaMe3, InMe3, and 6 equiv. of the donor-
functionalized alcohol, HOCH2CH2OMe, in toluene to de-
posit colorless, transparent galliumindiumoxide films at a
range of temperatures (350450  C).98 The optimum growth
temperature was found to be 450  C, which produced trans-
parent films with a composition of Ga0.6In1.4O3 and glancing-
angle PXRD showed that a gallium-substituted cubic In2O3
lattice was adopted. Recent developments investigating poten-
tial replacements for ITO for optoelectronic devices include
zincindiumtin oxide, as well as aluminum-doped and
gallium-doped zinc oxide.85,82
1.33.3 Precursors for the Growth of Group 14 Oxides
1.33.3.1 Introduction
Group 14 oxides have a vast range of technological applica-
tions and are used widely in everyday life from silicon dioxide,
or silica, which is widely used in electronics manufacturing, to
Ge(OiPr)4 + (C16H33)N(CH3)3Br
Ge(OEt)4 + H2O
Scheme 7 Route to preparing germanium oxide sol from germanium alkoxides.
tin dioxide that is commonly used for gas sensors, as well as
TCO applications. These oxides have been deposited using a
variety of methods that can alter the final properties of the
films. In most cases, commercially available precursors have
been employed, for example, tin halides and organotin pre-
cursors for SnO2, with the addition of an oxygen source (usu-
ally H2O, H2O2, O2, and O3); however, there are a few reported
examples of more sophisticated precursors. More commonly,
certain properties of the film to make them more suitable for
application purposes. Examples of compounds of group 14 can
8
be found in Chapters 1.11, 1.18, and 1.19.
This section of the chapter has been subdivided into three
parts, comprising germanium oxide, tin oxide, and lead oxide.
Although germanium is a metalloid rather than metal, it has
been included since there are no extensive reviews available
and reports on the deposition of germanium oxide are very
limited. Silicon, also a metalloid, has not been included,
although a large volume of work has been reported on the
formation of silicon oxide films, much of it has been exten-
sively reviewed previously. These include a number of brief
reviews,99,100 comprehensive reviews,101 and book chap-
ters,102,103 on the deposition of silicon oxide films and readers
are directed to these references for further information.
1.33.3.2 Germanium Oxide Precursors
Germanium oxide is of interest in both its amorphous form,
GeOx, and its crystalline form, GeO2, for its optical and electrical
properties.104 Germanium oxide is a high k interlayer dielectric
material that enables deep submicron technology for the con-
tinued improvement of integrated circuits.105 It also has high
carrier mobilities, which makes it particularly interesting for
high-frequency applications.106,107 Thin films of germanium
oxide can be produced in several ways including pulsed laser
deposition (PLD),108 thermal evaporation,109 magnetron sput-
tering,104 solgel methods,110 photochemical vapor deposi-
tion,111 electron beam evaporation,112 and CVD.113 Most of
these techniques use germanium oxide as the precursor to
form the thin films.
There are very few reports on the deposition of germanium
oxide thin films since 2000 and those that have been described
utilized solgel, photochemical vapor deposition, or CVD to
grow the films. Phani et al. report the use of germanium
isoproproxide, [Ge(OiPr)4], to deposit thin films of germa-
nium oxide via the solgel method, as shown in Scheme 7.114
acac
2-methoxyethanol
700/800 C
GeO2 sol GeO2 film
Spin coating
EtOH

The resulting sol is then used to spin coat on Si(100) sub-
strates followed by annealing at 800  C. The films deposited
were shown to be polycrystalline GeO2 in the hexagonal phase
by PXRD. Jang et al. also reported the use of solgel to deposit
germanium oxide thin films using tetraethoxyorthogermanate,
[Ge(OEt)4], as a starting material, as shown in Scheme 7.115
The resulting sol was also spin coated onto Si(100) substrates
but annealed at 700  C. The GeO2 films deposited were, how-
ever, found to be amorphous by XRD analysis.
Photochemical vapor deposition has been utilized to obtain
germanium oxide thin films.111 Tetraethoxyorthogermanate was
again used as the precursor to deposit the films onto silica sub-
strates which were then radiated with Ar2*, Kr2*, and Xe2*
excimer lamps to produce germanium oxide thin films. Khan
et al. used LPCVD to deposit germanium oxide thin films using
solid germanium and oxygen as the metal and the oxygen
source, respectively.113 The germanium was heated to vaporize
the element, which was then transported to the chamber, under
pressure. On reaction with the oxygen source, GeO2 thin films
were then deposited on Si(100) heated substrates.
In order to enhance the optoelectronic properties, various
elements have been doped into GeO2 including Cu, Ag, and
Au.110,114 Moreover, germanium is much more commonly
used with other elements to produce oxide thin films. For
example, GeO2SiO2 is a commonly produced material and
used for optoelectronic devices,116 whereas TeO2GeO2K2O
is used for optical wave guides and acousto-optic devices.112
GeOPbO is also a material that has received a lot of attention
due to photonic applications, since it presents low absorption
in the visible and IR regions, small cutoff phonon energy, and
high linear and nonlinear refractive index.117,118 Overall, ger-
manium oxide is not a material that receives much attention
and there is very little in the way of precursor chemistry to this
oxide; however, when incorporated with other elements, the
resulting materials have useful properties and are used in op-
toelectronic and photonic applications.
1.33.3.3 Tin Oxide Precursors
Tin oxide can be found in many forms including a-SnO,
b-SnO, SnO2, Sn2O3, and Sn3O4. The most common form is
tin dioxide or stannic oxide, SnO2, which is also the form that
tin oxide is found as a mineral (cassiterite). Tin dioxide has a
tetragonal rutile structure and thin films are often found to be
amorphous or polycrystalline but can also be epitaxial. As a
thin film, tin oxide has a wide bandgap of 3.7 eV at room
temperature and is an n-type metallic oxide semiconductor119
with high transparency (85%),120 and very good electrical
conductivity and chemical stability.121 Tin oxide is often
doped with fluorine122 or antimony123 to improve conductiv-
ity and additionally is commonly used to dope indium oxide
films to make ITO,124 which is one of the most widely used
TCOs in industry (Section 1.33.2.6). The morphology and
surface state of the SnO2 thin films can play an important
part in the final characteristics of the film and so a precise
control of the microstructure is required.125 This can be
obtained by carefully selecting the techniques used to deposit
the thin films as well as the choice of precursor.
Currently, tin dioxide is a material of considerable interest
for a number of applications. The material is useful in its
Precursor Chemistry Main Group Metal Oxides 1031
amorphous and crystalline form for preparing semiconductor
gas sensors,126 ion-sensitive field-effect transistors,127 dye-
sensitized solar cells,128 solar cells with active layers from amor-
phous silicon129 or IIVI compounds,130 solar-selective reflectors
for photovoltaics,131 catalysts for CO132 and CH4133 oxidation,
electrochemical oxidants of organic species,134 and lithium ion
batteries.135 Also high on the list are coatings for architectural
glass,136 and rigid137 and flexible138 electro-optical devices.
For SnO2 to be suitable for these applications, thin films are
generally required. The formation of SnO2 thin films has been
achieved via radiofrequency magnetron sputtering,138 although
there are many other techniques that have also been utilized,
which use SnO2 as a starting material as well as a number of
techniques that exploit the use of precursors. These techniques
including mid-frequency dual magnetron sputtering,136 bipolar
pulsed magnetron sputtering,139 direct current magnetron sput-
tering,140 dual ion beam sputtering,141 electron beam evapora-
tion,142 dual ion beam-assisted electron beam evaporation,124
PLD,143 atmospheric environment laser deposition,144 rheotax-
ial growth and thermal oxidation,145 spray pyrolysis,146 ultra-
sonic spray pyrolysis,147 solgel dip coating,148 deposition
from aqueous solutions,149 ALD,150 and a variety of CVD
techniques including AP,151 low pressure (LP),152 ion beam
induced,153 plasma enhanced (PE),154 laser-induced,155 and
photo-induced.156 This section gives an overview of research
that employs precursor reactions to form SnO2 thin films.
The most commonly used starting materials to precursors to
SnO2 are anhydrous or hydrated tin halide compounds which
are usually refluxed with an alcohol to form tin alkoxide com-
pounds. Thus, SnCl2157 or SnCl22H2O158 was refluxed with
ethanol,159 iso-propanol,148 or tert-butanol160 and the hydrated
product was used as a solgel for spin161 or dip coating.162 A
base can be added to speed up the precipitation of the tin
alkoxide by the addition of ammonia,163 triethanolamine,164
or ammonium hydroxide.158 Brezesinski et al. also reported on
an evaporation-induced self-assembly process that incorporated
the solgel/dip-coating method using SnCl4 in ethanol and the
amphiphilic block-copolymer template (H(CH2CH2CH2(CH)
CH2CH3)89(OCH2CH2)79OH).165 The tin alkoxide was then
dissolved in a solvent, such as ethanol,159 water,158 or acetyla-
cetone166 to prepare the sols. Tin tartrate (from hydrous SnCl2
and tartaric acid) in ethanol167 and tin 2-ethylhexanoate, [Sn
(O2CCH(C2H5)C4H9)2], in hexane168 as a single-source precur-
sor have also been reported.167 Films have been deposited on
quartz,158 silica,166 glass,169 and patterned surfaces157 which are
then annealed in air between 300 and 900  C to crystallize the
resulting SnO2 films.
The pyrosol process is a method for film deposition on a
heated surface, based on precursor solutions passing through a
spray nozzle142 or by an aerosol produced by ultrasonic atom-
ization.147 It is a low-cost process that produces high-quality
films for TCO applications. Precursor solutions are sprayed or
atomized via a carrier gas onto a heated substrate to obtain
uniform thin films. Precursors commonly in use are readily
available compounds, typically tin halide compounds. The Sn-
containing precursors, such as SnCl4,170 SnCl22H2O,171 and
SnCl45H2O,172 were dissolved in solvents including distilled
water,147 ethanol,142 methanol,171 and iso-propanol.173 Okuya
et al. reported the use of tetra-n-butyltin(IV), SnnBu4, dissolved
in H2O2174 and Aguilar-Leyra et al. utilized di-n-butyltin(IV)
1032 Precursor Chemistry Main Group Metal Oxides
diacetate in ethanol175 as alternative nonhalide precursors. The
precursor solutions were atomized and transported through a
nozzle using a carrier gas at flow rates of 35 ml min1. The
precursor solutions were sprayed at the heated substrates,
which include glass and quartz, at temperatures of
200500  C. The resulting films were found to be of uniform
thickness and crystalline rutile SnO2 had been grown.
The majority of precursors for ALD and plasma-assisted
deposition of SnO2 films are tin halides. ALD was used by
Sundqvist et al. to deposit SnO2 using SnCl4 with the oxygen
sources H2O176 and H2O2.177 Alternatively, SnI4 has been used
in ALD along with O2 as the oxygen source.177,178 Elam et al.
deposited SnO2 using an ALD process of tetrakis(dimethyla-
mino) tin, [Sn(NMe2)4], with H2O2.179 Dibutyl tin acetate has
also been used as a precursor in plasma-assisted ALD.180 Films
of SnO2 were deposited on glass, silica (100), or a-Al2O3 (012)
substrates via ALD at temperatures of 50750  C. Those films
grown below 350  C were mostly deposited on glass and were
amorphous. In contrast, films deposited on a-Al2O3 (012)
at higher temperatures (typically 600  C) formed epitaxial
rutile SnO2.
CVD of SnO2 films has been investigated by many groups
and a range of precursors, most of which are commercially
available, have been utilized to deposit the films. Van Mol
et al. reported vapor pressures of commonly used tin precur-
sors, which is important information when attempting to
choose a volatile precursor necessary for a CVD process.181
The tin precursors investigated included monobutyltin
trichloride [SnBuCl3], tin tetrachloride SnCl4, monomethyltin
trichloride [SnMeCl3], dimethyltin dichloride [SnMe2Cl2],
trimethyltin chloride [SnMe3Cl], and tetramethyltin [SnMe4].
The enthalpy of evaporation determined for the various tin
precursors is given in Table 1 and shows that SnCl4 and
[SnMe4] have the lowest DHvap pressures at 37.8 and 33.7 kJ
mol1, respectively.
There are many reports on the use of tin halide compounds
to deposit SnO2 thin films via CVD. Since 2000, there have been
reports using SnCl4,182 SnCl22H2O,183,184 SnCl45H2O,153,
185187
and SnI4.176,188 All of the CVD methods require an
oxygen source, such as O2154,176,182,183,185188 and air,184,186 as
carrier gases or the addition of ethanol in AACVD.184 Other
CVD techniques used with these tin halide precursors include
APCVD,182,183,186 LPCVD (using pressures of 98 Torr),176 and
PECVD (using RF powers of 60 W154 and 100150 W185). Glass
Table 1 Enthalpy of evaporation of organotin and tin halide
precursors181
Precursor DHvap (kJ mol1) Temp. range of
measurements (C)
SnBuCl3 (l) 58.8  0.8 70145
SnCl4 (l) 37.8  0.4 040
SnMeCl3 (s) 61  2 2040
SnMeCl3 (l) 46.1  0.8 6090
SnMe2Cl2 (s) 77.6  0.8 60100
SnMe2Cl2 (l) 53.5  0.6 100130
SnMe3Cl (s) 61  2 020
SnMe3Cl (l) 49.0  0.3 4085
SnMe4 (l) 33.7  0.2 025
substrates were mainly utilized as well as porous alumina154 and
a-Al2O3 (012)176 and the substrate temperatures ranged from
200 to 700  C.
Organotin compounds, such as tetramethyltin, have also
been used to deposit thin films of tin oxide. Lancock et al. used
tetramethyltin, SnMe4, to deposit tin oxide by laser-assisted
CVD using O2 as an oxygen source and depositing onto alu-
mina substrates at room temperature, 150, and 300  C.155 The
same precursor, SnMe4, was utilized by Tamura et al. where
SnO2 films were grown via photochemical CVD using an O2
(containing O3) carrier gas as the oxygen source.156 The films
were then irradiated using a mercury lamp and exposed to UV
radiation. Supothina used tin 2-ethylhexanoate as a precursor
to deposit SnO2 films via photochemical CVD,189 whereas
Huang et al. used dibutyltin di-acetate as a precursor with O2
as a carrier gas via PECVD with RF powers of 0.41.5 kW on Si/
SiO2 wafers.190
In general, tin alkoxide precursors have been widely used
and they do have the advantage of being single-source and do
not contain any SnX (X halide) or SndC bonds, thus elim-
inating contamination from these elements. AP191 and
LP191,192 CVD (pressures of 15 Torr) have been utilized to
deposit thin films of SnO2 onto glass heated at 250400  C
using [Sn(OtBu)4] as a precursor. Generally, films deposited
below 400  C were found to be amorphous. Those deposited
above 400  C were analyzed as polycrystalline rutile SnO2.
However, when using the a-Al2O3 substrates, epitaxial SnO2
was observed.188 The majority of precursors to SnO2 using the
CVD process were commercially bought; however, there are a
few reports of more sophisticated precursors, which are de-
scribed below.
There are a few precursors that have been used by other
methods than those described above that are worth mention-
ing. Liquid flow deposition188,193 has been used to deposit
SnCl45H2O dissolved in HCl at a flow rate of between 0.3
and 10 ml min1 at 80  C onto silica wafers. Lee et al. used a
precursor solution consisting of SnCl4 dissolved in acetonitrile
as an ink to print precursor material onto SiO2/Si substrates.194
The resulting films were then annealed at 500  C producing
polycrystalline rutile SnO2 films. Tin dioxide films have also
been formed from SnS films. Glass substrates were dipped
successively into solutions of Na2S and SnCl2 which produced
SnS thin films.195 These films were then annealed in air at
400  C oxidizing the films to a mixture of Sn2O3/SnO2.
Electrodeposition has been used in several ways to deposit
tin oxide thin films: SnCl4 which is reacted with a viscous
solution of poly(ethylene oxide) and deposited onto Si/SiO2
wafers196 and an acidic solution of SnSO4 with prior O2 bub-
bling has been deposited onto ITO glass197 by electrode-
position. Additionally, SnCl22H2O dissolved in HCl/KCl
mixture which was bubbled with O2 prior to film growth via
electrodeposition.198 Butyl-rhodamine B was added as a
structure-directing agent to electrodeposit a series of SnO2
rutile, nanocrystalline thin films with different morphologies
and porous structure onto glass and titanium substrates.
There are few reports in the literature on the synthesis of
precursors to tin oxide. The majority of the precursors used are
simple and commercially available compounds that have been
mentioned previously. Chi et al.199 have reported the synthesis
and CVD of a range of bis(b-diketonate)tin(II) compounds, as
 Precursor Chemistry Main Group Metal Oxides 1033

shown in eqn [9]. Compounds 4348 exist as monomers and
the structure of 46 was determined via single-crystal XRD,
which showed that the molecule is monomeric in the solid
state with the central tin atom adopting a distorted trigonal
bipyramid geometry.
R H
R O
Na
O O O
SnCl2 + 2 Sn
R R
-2 NaCl O
O 10
R
R 9
43: R = R = Me, acac
44: R = R = CF3, hfac
45: R = Me, R = CF3, tfac
46: R = C3F7, R = tBu, fod
47: R = CF3, R = Ph, btfac
48: R = CF3, R = 2-thienyl, ttfac
Compounds 4345 were selected for use as precursors for
LPCVD since they can be readily evaporated at ambient tem-
peratures under reduced pressure. The precursors were
heated to around 35  C before being allowed to enter the depo-
sition chamber either under dynamic evacuation (system pres-
sure ca. 0.1 Torr) or with a carrier gas of O2 (system pressure ca.
100 Torr). The deposition temperatures ranged over 300700  C
at 50  C intervals. Films were deposited on glass substrates and
the results of the LPCVD experiments are summarized in Table 2.
The films deposited using compounds 4345, as single-source
precursors (i.e., without the addition of a carrier gas), resulted in
the formation of polycrystalline tin films. The deposition of Sn
rather than tin dioxide was thought to be the result of the SndO
bonds breaking during deposition, which is also consistent with
mass spectroscopy fragmentation. Only when using compounds
4345 in the presence of an O2 carrier gas did the LPCVD result
in the deposition of tin dioxide films and these were crystalline
above 400  C for 44 and 450  C for 43 and 45. The resulting
SnO2 films consist of continuous, round grains and the grain
size increases with increasing temperature as observed by SEM
and AFM.
Talaty et al. synthesized an alternative tin dioxide precursor,
cis-[SnCl4{OC(H)OC2H5}2] (49),200 as outlined in eqn [10].
Compound 49 was prepared via the reaction of SnCl4 and
2 equiv. of ethyl formate in hexane resulting in the formation
of colorless crystals in a good yield (70%). Single-crystal XRD
of 49 confirmed the monomeric nature of the complex as a
six-coordinate species with a distorted octahedral geometry.
The ethyl formate group bonds via the oxygen in the cis
position.
Cl
Cl Cl
O H Sn H
SnCl4 + 2
O O
O
O Cl O
49
The in situ reaction, shown in eqn [10], was explored using
APCVD with nitrogen as a carrier gas by Talaty et al., where the
reaction took place in the gas phase and was compared to
depositions using compound 49 as a single-source precursor to
thin films of tin dioxide. Films were deposited onto glass sub-
strates heated to 560  C for optimum film growth. Tin dioxide
films grown by both routes were found to have very few differ-
ences and so the in situ method was utilized for further studies
for its ease of use. The tin dioxide films were also doped with
2,2,2-trifluoroethyl trifluoroacetate to make SnO2:F films using
varying amounts of the precursor (Section 1.33.3.5). The growth
rate was explored in order to optimize the film thickness for TCO
applications, which needs to be in the region of 400010 000 A.
It was found that a flow rate of 0.81.2 l min1 was needed to
obtain a thickness in that range. SnO2 and SnO2:F were analyzed
by XRD and SEM analysis. Both films formed crystalline films
and had similar morphology; however, the SnO2:F films have a
slightly larger grain size which is thought to account for lower
resistivity and high mobility.
Tin dioxide thin films are very useful materials for a variety
of applications as they are conductive and visibly transparent
thus perfect for use as TCOs. Precursor and substrate choice,
as well as the technique utilized, can be important for fine
tuning the final properties of the film. As described above,
techniques frequently used include solgel, spray pyrolysis,
and CVD. Tin halide compounds and an oxygen source are the
most commonly used precursors as they are cheap, easy to
handle, and readily available. Organotin and tin alkoxides are
also utilized but are less frequently reported. More sophisticated,
synthetically designed precursors are seldom observed in the
literature.
1.33.3.4 Lead Oxide Precursors
Lead oxide can be found in many forms including a-PbO
(litharge) and b-PbO (massicot), a-PbO2 (scrutinyite) and
b-PbO2 (plattnerite), and Pb3O4 (minium).201 Litharge,

Table 2 Summary of CVD conditions and results for compounds 4345

Precursor Carrier gas Deposition temp ( C) Deposition rate (A min1) Film characterization XRD

43 600700 750900 Polycrystalline Sn

O2 350600 180600 Amorphous (<400  C) polycrystalline SnO2 (>450  C)
44 500600 120200 Polycrystalline Sn
O2 300600 20150 Amorphous (<350  C) polycrystalline SnO2 (>400  C)
45 500600  C 180300 Amorphous
O2 300600  C 50300 Amorphous (<400  C) polycrystalline SnO2 (>450  C)
1034 Precursor Chemistry Main Group Metal Oxides
plattnerite, and minium have tetragonal crystal phases, whereas
massicot and scrutinyite have orthorhombic structures. With lead
oxide, PbO, both modifications have very low electrical conduc-
tivity and possess interesting semiconducting and photoconduct-
ing properties. Therefore, PbO has potential applications of
industrial importance in imaging devices, electrophotography,
electroradiography, and for laser technology.202 Thin films of
lead monoxide are of interest as antireflection coatings in the
production of detectors,203 and in the fabrication of semiconduc-
tor gas sensors.204 Lead dioxide is also of interest as PbO2 anodes
have received much attention because of their structural, mor-
phological, optical, and mechanical properties and their potential
applications in wastewater treatment,205,206 ozone generation,207
analytical sensors,208 electro-winning of metals,209 and battery
electrodes.210 The substantial growth in new avenues such as
telecommunication networks, electronics, computers, and so
forth, has raised new challenges for PbO2 batteries. PbO2 elec-
trodes are also applied in industrial processes such as an energy-
conversion process, recycling, and environmental treatment.211 It
is well known that PbO2 exhibits excellent chemical stability,
high conductivity, large overpotential, and chemical inertness
for electrolysis in an acid medium.
Lead oxide and lead dioxide thin films have been obtained
by a variety of techniques, such as thermal evaporation,212
exposing lead films to RF oxygen plasma or oxygen,213 laser-
assisted deposition,204 atomic layer epitaxy,214 ALD,215 and
metal-organic CVD (MOCVD).216 Most of these techniques
use preformed lead oxide to deposit thin films; however, ALD
and MOCVD use precursor chemistry to deposit thin films,
which are explored in this section.
Harjuoja et al. report the use of a variety of lead-based
precursors to deposit lead oxide thin films using ALD.215
Three different precursors are used, including lead diethyl-
dithiocarbamate [Pb(S2CNMe2)2] (50), lead 2,2,6,6-tetra-
methyl-3,5-heptadione [Pb(thd)2] (51), and tetraphenyllead,
PbPh4, along with an ozone oxygen source. Films were depos-
ited using compounds 50, 51, and PbPh4 on silicon(100) sub-
strates at temperatures of 150400  C and pressures between
1.5 and 2.25 Torr using nitrogen as a carrier gas. On using
compound 50 as the precursor, only PbS2 thin films were
deposited, whereas 51 and PbPh4 resulted in the formation
of polycrystalline PbO2 thin films. Lee et al. reported the ALD
of lead oxide films using alternating cycles of lead bis(3-N,N-
dimethyl-2-methyl-2-propanoxide) [Pb(dmamp)2] (52) and
H2O depositing on Ir/IrO2/SiO2/Si substrates.217 Compound
52 was heated to 126  C and allowed to pass into the reaction
chamber where the substrates were heated to 200  C and the
system was at 0.50.8 Torr. Cycles were subsequently purged
with argon before H2O was passed through the system as the
oxygen source. Films deposited from 52 and H2O were found
to be a mixture of polycrystalline PbO and Pb metal.
CVD is the most common way of depositing lead oxide thin
films and typical precursors used are 51,218,219 PbEt4,220 and
[PbEt(OCH2C(CH3)3)].218 Various CVD techniques have been
employed for the deposition, most commonly LPCVD;
PbCl2 + 1/2 O2
Cl2
Scheme 8 Formation of PbO from lead dichloride.
however, the use of laser, plasma, and liquid delivery CVD
has also been reported. For LPCVD, the precursors were heated
to elevate into the vapor phase and transported with a carrier
gas and an additional O2 source. The lead oxide films were
grown on a range of substrates, including silicon(100),220
SiO2-coated silicon,220 and Pt-coated Si/SiO2 at temperatures
of 350700  C.218 The lead oxide thin films formed were gen-
erally polycrystalline PbO, but a mixture of the litharge and
massicot forms, as well as rarer occurrences of scrutinyite and
Pb metal, was also observed.
Dai et al.221 reported the use of PbCl2 as a precursor using
LPCVD, as shown Scheme 8. A temperature of 600800  C was
used to get PbCl2 into the gas phase and argon and nitrogen were
used as carrier gases. The oxygen source comes from an O2/H2O
mixture and films were deposited on silicon(100) wafers at sub-
strate temperatures of 500700  C with reactor pressures of
0.2 Torr. Polycrystalline PbO thin films were obtained and the
mechanism for its formation from PbCl2 was thought to follow
the reactions shown in Scheme 8. Lead oxide films were also
deposited via PECVD using [PbEt4]/O2222 and by laser-assisted
CVD of compounds 51223 and [PbEt(OCH2C(CH3)3)].224 In
addition, liquid delivery MOCVD has been exploited using 51,
dissolved in ethylcyclohexane, with an O2 source.225 The lead
oxide films were deposited on the substrates, Pt/SiO2/Si and
Ir/IrO2/SiO2/Si.
The typical precursors used to form lead oxide thin films are
51 and PbEt4. However, there are a few reports describing more
sophisticated lead oxide precursors, which are summarized
below.
Lead b-diketonate compounds have been used as precursors
to lead oxide thin films and a range of compounds were devel-
oped.216 The synthetic routes to these lead b-diketonate com-
pounds using a variety of b-diketonate ligands is shown in eqn
[11].216,226 Krisyuk et al. used thermogravimetric analysis (TGA)
to investigate the decomposition of compounds 51 and 5358,
and the order of volatility was (obtained from the data on the
saturated vapor pressure in the sublimation region) found to be
57 > 51 > 53 > 56 > 54 > 58. Compounds 51, 54, 56, and 57
were chosen as precursors for use in LPCVD. Depositions using
compounds 51, 54, 56, and 57 were carried out at temperatures
of 350600  C using silica(001) and silica(111) substrates in the
presence of an additional O2 source. Compounds 51 and 54
afforded polycrystalline PbO films, whereas 56 and 57 resulted
in the formation of lead thin films, especially above 500  C
which is above the melting temperature of lead.
Mass spectrometry was used to investigate the reaction
mechanism of the deposition of compound 51 to PbO. Com-
pound 51 was found to be monomeric in the gas phase and the
study showed that lead oxide was formed by two parallel
routes, as shown in Scheme 9. The first stage of the thermal
decomposition of 51 is thought to involve the opening of the
chelate rings of thermally excited molecules.227
Hedinger et al. reported the use of a lead alkoxide precursor
to deposit lead oxide thin films.228 The trinuclear lead alkoxide
[Pb3(H3tdci)3]5H2O (59) (tdci 1,3,5-trideoxy-1,3,5-tris
PbO PbCl2 + H2O
2 HCl
 Precursor Chemistry Main Group Metal Oxides 1035

R1 R2

O O O O
PbL2 + 2 Pb
2 LH O O
R1 R2

L = CH3CO2 or NO3 R1 R2
51: R1= R2= tBu
(thd) 11
53: R1= R2= CH3 (acac)
54: R1= C(CH3)3, R2 = C(CH3)2OCH3 (zis)
55: R1= CH3, R2 = CF3 (tfac)
56: R1= C(CH3)3, R2 = CF3 (pta)
57: R1= R2= CF3 (hfac)
58: R1= R2= C6F11 (dhfac)

PbO(s) + CH3COC(CH3)3(g) + C(CH3)3C(O)CH=CH(CH3)3(g) + CH2=C(CH3)CH=CO(g)

Route 1

[Pb(thd)2]
51
Route 2
PbO(s) + Hthd(g) + (CH3)3CC CC(O)C(CH3)3(g)

Scheme 9 Decomposition routes of 51 to lead oxide.

OH OH
OH N
H2O, NEt2 12
+ Pb(NO)3 [Pb3(H3tdci)3]5H2O
N N 60 C
59
tdci

(dimethylamino)-cis-inositol) was synthesized via the reaction N

of Htdci with lead nitrate in water, as shown in eqn [12].
Single-crystal XRD confirmed that 59 was trinuclear and
showed that the compound adopts a sandwich-like structure N
where the three lead atoms are encapsulated by two triply N
deprotonated ligand molecules (Figure 1). Each lead atom in
O
59 is coordinated to four oxygen and two nitrogen atoms of the O O
two tdci ligands, forming one 6-membered and two 5-membered
Pb
chelate rings, and all six deprotonated oxygen atoms of the Pb
two tdci ligands act as bridges. The two cyclohexane residues
O Pb
together with the 12 methyl groups bound to the nitrogen
atoms are directed to the outside, while the polar alkoxo groups
O
together with the metal cations are hidden in the inside of the
O
molecule. The rather rigid and compact (almost spherical)
structure with its hydrophobic surface, together with the lack N
of charge, generates a volatile compound that is soluble in N
organic solvents such as dichloromethane or acetone but insol-
uble in water.
Compound 59 was used as a precursor to deposit lead oxide N
thin films via LPCVD. The precursor was heated to 260  C and
the chamber was under a pressure of 16 Torr. Depositions were Figure 1 Structure of compound 59. Hydrogen atoms omitted for clarity.
carried out at 300600  C onto a range of substrates including
stainless steel and copper, and silver-, gold-, or copper-coated
silicon quartz and silicon substrates. Helium was used as a substrates formed polycrystalline PbO thin films containing
carrier gas and O2 was added to the gas phase mixture. Thin mixtures of the litharge and massicot phase as well as various
films of PbO deposited from 59 onto quartz and silicon were metal/lead type compounds from reaction of the precursor
amorphous. However, those films deposited on the metal with the substrate.
1036 Precursor Chemistry Main Group Metal Oxides
Lead oxide thin films are most commonly deposited in the
PbO form with mixtures of the a-tetragonal litharge phase and
the b-orthorhombic massicot phase. The most common pre-
cursors in use are [Pb(thd)2] and [PbEt4] with an O2 oxygen
source; however, other precursors do exist.
1.33.3.5 Precursors to Ternary Group 14 Oxides
Tin dioxide is commonly doped with other elements to
modify the qualities of the film, such as to improve sensi-
tivity, selectivity, and stability as well as conductivity and
optical transmittance. Fluorine and antimony are frequently
used as dopants in SnO2 thin films and there is extensive
literature on fluorine-doped tin dioxide (SnO2:F; FTO) thin
films due to their use as TCOs. FTO is the second widely
used TCO material after ITO, mainly in solar cells where
high temperatures are required for device fabrication. There
has been extensive research in this area and a comprehen-
sive review was published in 2005.229 In most cases, pre-
cursors for the formation of thin films of FTO have
included low-cost tin halides or organotin compounds
with the addition of a fluorine source, typically NH4F and
HF. There are few reports of more sophisticated precursors,
although Cachet et al. used the compound [Sn(bdk)2
(OtAm)F] (60, bdk acac; 61, bdk thd) to deposit FTO
films via spin coating and spray pyrolysis techniques.122
R RF sputtering,257 ion-beam sputtering,258 solgel,259 laser
R
O
O
F [PbEt(OCH2C(CH3)3)], [Zr(thd)4], [Zr(OiPr)4], [Zr(OtBu)4],
Sn
OR
O
O
R
R 16 Oxides
60: R = Me, R = tAm
61: R = tBu, R = tAm
Antimony is also doped into tin oxide thin films, which
can change the electronic properties of the films.123,230 As
with SnO2 and SnO2:F, common precursors to SnO2:Sb
include tin halides: anhydrous SnCl2,231 as well as hydrated
SnCl22H2O232 and SnCl44H2O,233 with SbCl3 being the
common source of antimony. These precursors are usually
dissolved with the aid of an acid for the tin precursor and a
base for the antimony halide before being utilized in the
deposition of SnO2:Sb films by solgel methods,158 thermal
evaporation,234 or spray pyrolysis.235 There are reports of
other precursors to deposit thin films of SnO2:Sb by the
solgel method, such as using SnC2O4 as the tin precursor,
with SbCl3 and citric acid and triethanolamine as the
solvent236; SnCl2H2O and Sb2O3237; [Sn(OBu)4] and
[Sb(OBu)3].238 Epifani et al. reported the use of SnF2 and
SbF3 to deposit thin films of SnO2:Sb by thermal evapora-
tion using distilled water as the solvent.239 There is also a
report of the use of tin 2-ethylhexanote and a Symetrix Sb
solution precursor for the deposition of SnO2:Sb by thermal
evaporation and an excimer laser-assisted metal-organic
deposition method.240 Kololuoma et al. used spray pyroly-
sis to deposit SnO2:Sb thin films using the alkoxide pre-
cursors [Sn(OiPr)4] and [Sb(OiPr)3] with methacrylic
acid.241 However, there are no reports of more complex
precursors and none incorporating a SnOSb fragment
for use as a single-source precursor.
Numerous other elements have been doped into tin oxide
thin films to attempt to improve the electronic properties,
including Ag,175 Au,242 Ca,243 Ce244 and CeO2,245 Co,246
Cu,247 Er,248 Fe,246 GeO,70 Li,249 Ni,239 Os,240 Pd,250 Pt,251 and
Ta.154 Various properties of the films are modified on addition of
these dopants including variation of film morphologies. Of these
materials TCO properties have been improved (Ag, Ce, Er, Li, Pd,
Pt, and Ta) and others expand the range of gases that can be
detected for gas sensor applications: SnO2:Ag-propane; SnO2:
Ca-ethanol; SnO2:CeO2-H2S and CH4; SnO2:Pd-H2252; SnO2:
Pd-H2S253 and O2.254,255
Lead oxide ternary or higher materials are currently of
high interest. Fabrication of lead-based ferroelectric thin
films, such as PbTiO3, (Pb,La,)TiO3 (PLT), and Pb(Zr,Ti)
O3 (PZT), has been extensively investigated with the aim of
applying them to IR sensors, nonvolatile random access
memories (ferroelectric random access memory (FeRAM)),
and dynamic random access memories.217 A good review
on PZT ceramics is reported in the book section Piezoelectric
PZT Ceramics.256 Various deposition processes, such as
ablation,260 and MOCVD,261,262 have been employed to
fabricate high-quality films. The precursors which are typi-
cally used in the deposition of the films include [Pb(thd)2],
[Ti(thd)2(OiPr)2], [Ti(OEt)4], and [Ti(OiPr)4].218
1.33.4 Precursors for the Growth of Group 15 and
1.33.4.1 Introduction
Thin films of the binary group 15 and 16 oxides have limited
properties in comparison to the corresponding group 13
and 14 oxides (see Chapters 1.12, 1.20, and 1.21 for exam-
ples of compounds from these group). For example, antimony
oxide finds use in catalysis and gas sensor devices and bismuth
oxide is of interest in high Tc superconductors and potentially
for fuel cells. There appear to be no reports on the growth of
arsenic oxide films, although As2O3 has been observed as a
minor impurity phase at the surface of arsenic-containing
semiconductor films, such as gallium arsenide (GaAs). This
section of the chapter has been subdivided into three parts,
which include antimony oxide, bismuth oxide, and tellurium
oxide. The focus is on the chemistry of precursors to the thin
films, although mention is made of techniques, such as
APCVD and ALD, which tend to utilize simple, commercially
available precursors in combination with a secondary oxygen
source. Some discussion of compounds that have been synthe-
sized as potential precursors to these materials is given, since
film deposition using single-source precursors for these ele-
ment oxides is rather limited.
 Precursor Chemistry Main Group Metal Oxides 1037

1.33.4.2 Antimony Oxide Precursors ammonia gas in ethanol solvent; however, TGA analysis
revealed it to be a poor candidate for CVD of Sb2O3 films
The chemistry of antimony oxide is almost entirely dominated
due to premature decomposition:
by the largely stable antimony 3 oxidation state (Sb2O3),
of which there are two distinct polymorphic forms: the high- [Sb(NMe2)3] + 3 ROH Sb
temperature cubic a-Sb2O3 (>570  C) and the lower-temperature -3 HNMe2 OR
RO
b-Sb2O3. However, other antimony oxides exist, including anti- OR
mony pentoxide (Sb2O5) and diantimony tetroxide (Sb2O4). The 62; R = C6H3Me2-2,6
applications of antimony oxide coatings are rather limited to
13
catalysis, opacifying agents, and gas sensor devices. CVD studies
detailing the growth of antimony oxide films are rather scarce, Molloy et al. also synthesized a series of tris(N,N-dialkylcar-
with only three or four examples found in the literature. The bamato)antimony(III) complexes, [Sb(O2CNR2)3] (64, R Me;
formation of antimony oxide from single-source precursors is 65, R Et; 66, R iPr).266 Compound 64 was prepared via a
also limited; so an overview of compounds synthesized as poten- carbonyl insertion into the SbNMe2 bond of trisdimethylami-
tial precursors is also given in this section. doantimony [Sb(NMe2)3] using dry ice (CO2) vapors, as shown
The first study concerning the deposition of antimony oxide in Scheme 10. In contrast, 65 and 66 (the ethyl and di-isopropyl
films via CVD was carried out by Badawy in 1990,263 via a CVD- derivatives), were prepared via the reaction of SbCl3 and the
spray technique consisting of antimony pentachloride in an corresponding amine (diethylamine or diisopropylamine) in
ethyl acetate solution which was sprayed onto glass substrates toluene followed by bubbling through CO2 gas (Scheme 10).
at 400  C to yield Sb2O5 films. Similarly, amorphous Sb2O3 Single-crystal XRD showed that 64 exists as a weakly bound
films were deposited via the APCVD reaction of antimony pen- dimer, whereas its diethyl and diisopropyl analogs were found
tachloride and ethyl acetate on glass substrates and gas sensors at to be monomeric (Figure 2). Compound 64 was then
550 and 600  C.264 The films were identified as Sb2O3 from XPS employed as a single-source precursor to Sb2O3. LPCVD of
measurements and displayed a notable response to reducing 64 produced highly crystalline cubic Sb2O3 films at a substrate
gases, such as ethanol, at an operating temperature of 400  C. temperature of 380  C.
Much work regarding precursor synthesis and deposition of
antimony oxide films was carried out by Molloy and coworkers
from 2002 onward, beginning with a report of the first struc- R = Me
Sb(NR2)3 + 3CO2 [Sb(O2CNR2)3]
turally characterized monomeric antimony(III) alkoxides [Sb
(OC6H3Me2)3] (62) and [Sb(OEt)5.NH3] (63).265 Compound 64, R = Me
62 is soluble in most solvents and was synthesized via an 65, R = Et
+ 3 CO2 66, R = iPr
alcoholamine exchange using tris(dimethylamido) anti-
mony(III), [Sb(NMe2)3], and 2,6-dimethylphenol in hexane,
as shown in eqn [13]. TGA studies of 62 revealed a clean one- R = Et or iPr
SbCl3 + 6R2NH Sb(NR2)3 + 3 R2NH2Cl
step weight loss starting at 75  C and completing by 500  C
with a residual mass consistent with the formation of Sb2O5. Scheme 10 Synthesis of tris(N,N-dialkylcarbamato)antimony(III)
Compound 63 was synthesized via a reaction of SbCl5 and complexes.

N
N O

O O
O
O Sb
N
O
O
N Sb O

O O O
O
N N

Figure 2 Structure of compound 64. Hydrogen atoms omitted for clarity.

1038 Precursor Chemistry Main Group Metal Oxides

Molloy et al. also deposited cubic Sb2O3 on glass sub- R

strates between 500 and 700  C via the APCVD of [Sb O O O
(OEt)3] (67) and [Sb(OnBu)3] (68).267 Initially, a secondary Sb(OEt)3 + 3 Sb
R R -3 EtOH
oxygen source was not used; however, its use led to deposi- O 14
3
tion of oxygen-rich Sb6O13 using 67 as a precursor at a depo- R
sition temperature of 700  C. The use of both 67 and 68 70: R = R = tBu, thd;
resulted in the deposition of antimony metal as an impurity -2 EtOH
71: R = tBu, R = C3F7, fod
phase in the films. Using a dual-source approach, the same
group deposited metallic antimony and antimony oxide Attempts at synthesizing [Sb(hfac)3] from 67 yielded an
thin films using the metalorganic precursors 64, 67, 68, and unexpected dimeric 3,4-dihydro-2H-pyran product, whereby
[Sb(OCH2CMe3)3] (69) on glass substrates via AP- and LP- the two hfac ligands combined to bond directly to a
CVD.268 The majority of the films deposited were mixed [Sb(OEt)] moiety, possessing the empirical formula [Sb(OEt)
phase, some of which contained antimony metal along with (hfac)2] (75) (Figure 3).
antimony oxide despite the use of O2 as an oxidizing gas The ligand exchange reaction of [Sb(OEt)3] with 1 or 2
during deposition. When no oxidizing gas was used, metallic molar equivalents of the appropriate diketone yielded the
antimony was deposited. Pure sentarmontite Sb2O3 was de- heteroleptic complexes [Sb(OEt)2(fod)] (76) and [Sb(OEt)
posited using 67 at 150300  C; films deposited above (thd)2] (77), respectively, whereas refluxing [Sb(OEt)5.NH3]
400  C led to mixed-phase films. Overall, it was observed with 1 equiv. of Hthd in toluene yielded [Sb(OEt)4(thd)] (78;
that the use of different precursors had little effect on the eqns [15][17]). The structures of 70, 71, and 77 were reported
resulting films; however, the use of 64 and 69 resulted in as monomeric, and the antimony(V) compound 78 contains
deposition of pure sentarmonite films at much higher sub- the antimony atom in an octahedral coordination geometry.
strate temperatures compared to using compound 67. TGA studies revealed that although the homoleptic complexes
Homoleptic antimony b-diketonates, [Sb(thd)3] (70) and 70 and 71 were volatile, they displayed premature weight
[Sb(fod)3] (71), and heteroleptic diketonates, [Sb(OEt)4(thd)] losses due to hydrolysis in air. APCVD using 70 and AACVD
(72), [Sb(OEt)(thd)2] (73), and [Sb(OEt)2(fod)] (74), were using 71 and 75 all proved unsuccessful in depositing anti-
also synthesized as potential precursors to antimony mony oxide or antimony-containing films.
oxide.269 Compounds 70 and 71 were synthesized via a ligand
exchange using [Sb(OEt)3] and 3 molar equivalents of the free -2 EtOH
Sb(OEt)3 + 2 Hthd Sb(OEt)(thd)2
diketone in either hexane or toluene (eqn [14]): 15
77

Sb
O

O
O
O O
O
O

O
Sb

Figure 3 Structure of compound 75. Hydrogen atoms omitted for clarity.


-EtOH
Sb(OEt)3 + Hfod Sb(OEt)2(fod)
16
76
-EtOH
Sb(OEt)5.NH3 + 2 Hthd Sb(OEt)4(thd)
-NH3 17
78
Tetraphenylantimony phenylacetylacetonate [Ph4Sb{OC
(Me)CPhC(Me)O}] and tetraphenylantimony thiobutylacety-
lacetonate [Ph4Sb{OC(Me)CSC(Me)O}] were synthesized by
the reaction of pentaphenylantimony with phenylacetylace-
tone or thiobutylacetylacetone in toluene under reflux.270
These colorless complexes are stable in air and soluble in
most solvents, and contain the central antimony atom in a
distorted octahedral environment. These complexes were not
investigated for deposition of antimony oxide films.
Thin films of Sb2Ox were deposited via the ALD reaction of
triisopropylantimony SbiPr3 and O3 on silicon wafers at 200
and 300  C.271 XPS studies indicated the presence of the penta-
valent oxide Sb2O5 at the surface of the film, with elemental
antimony found underneath during depth profiling. In contrast,
the ALD reaction of SbiPr3 and SbMe3 using H2O as a co-
reactant resulted in no film growth. Bachmann et al. used [Sb
(NMe2)3] in conjunction with O3 to form Sb2O5 films,272 in
order to investigate the suitability of this precursor to deposit
antimony oxides and sulfides (requiring H2S as a co-reactant)
via ALD. Amorphous Sb2O5 films were deposited between 120
and 270  C on silicon substrates and porous alumina mem-
branes, resulting in Sb2O5 nanotubes after dissolution of the
membrane matrix. Elemental analysis using atomic absorption
spectroscopy suggested that the composition of the films was
Sb2O5, which was also supported by mass spectrometry studies.
There are very few reports detailing the synthesis of anti-
mony oxide coatings via solgel methods. However, Bohra
investigated the synthesis of some novel ketoximate-modified
antimony(III) alkoxides for use as solgel precursors to Sb2O3
and Sb2O4.273 Hydrolysis of [Sb(OiPr)3] followed by annealing at
500  C yielded sentarmonite Sb2O3 exclusively. However, the
reaction of [Sb(OiPr)3] with a range of oximes was said to form
monomeric complexes of the type [Sb(OiPr)3n{ONC(CH3)
R}n] which yielded Sb2O3 and Sb2O4 after solgel processing
at 500  C.
1.33.4.3 Bismuth Oxide Precursors
The chemistry of bismuth oxide is almost entirely dominated
by the largely stable bismuth 3 oxidation state (Bi2O3), of
which there are a number of distinct crystallographic poly-
morphs (a, b, g, d, and o) that have been reported. The low-
temperature a- and high-temperature d-phases are stable while
the others are metastable. Sub-stoichiometric phases including
Bi2O2.33, Bi2O2.75, and BiO have been characterized and are
commonly found as impurity phases in samples of Bi2O3.
Interest in the synthesis of thin films of bismuth oxide and
bismuth oxide-containing materials grew during the late 1980s
with the discovery of a family of high Tc superconductors
Bi2Sr2CanCun1O2n6x which contain two BiO layers sand-
wiched between SrO, CuO2, and Ca layers.274 The drive to
develop cheap, renewable energy sources led to the discovery
that the metastable d-Bi2O3 phase possessed the highest oxide
Precursor Chemistry Main Group Metal Oxides 1039
ion conductivity of any material,275 making it suitable as a
possible electrolyte in solid oxide fuel cells. Furthermore, the
electronic bandgap of Bi2O3 is 23.2 eV, which has resulted in
research into its use as a photocatalyst for degradation of
pollutant dyes276 and water splitting.277 Recently, phases of
bismuth oxide have also seen use in dielectric,278 ferroelec-
tric,279 multiferroic,280 and gas-sensing materials.281
There are many techniques that have been reported to
deposit thin films of bismuth oxide. These include CVD
LPCVD, AACVD, direct-liquid-injection (DLI) CVD, APCVD
and ALD, sputtering, solgel processes, spray pyrolysis, dip
coating, screen printing, and spin coating. This section gives
an overview of the film deposition experiments that have
utilized single-source precursors and a discussion of the well-
known dual-source CVD routes to bismuth oxide films. It will
become apparent that although there are a wide variety of
methods to deposit such films and a similarly large collection
of well-defined molecular compounds to choose from, the
number of suitable single-source precursors is in fact extremely
limited, and the reasons for this are discussed. However, vola-
tile single-source precursors containing a preformed BiO
bond would be ideal for deposition of bismuth oxide, as this
would provide a much simpler route to obtain such materials,
eliminating any possibility of complex side reactions in the gas
phase together with the reduction in the use of oxygen gas at
high substrate temperatures.
There are numerous examples of dual-source CVD routes to
bismuth oxide, most of which require a bismuth source and
oxygen gas or plasma as the reactive gas. Bismuth halides to-
gether with oxygen gas have been shown to deposit amorphous
bismuth oxide films on glass substrates using APCVD at a variety
of substrate temperatures.282 As predicted, the use of BiI3 as the
bismuth source led to an impurity BiOI phase observed in the
XRD pattern at substrate temperatures between 600 and 800  C.
Similarly, Schuisky and Harsta283 demonstrated the use of BiI3
and oxygen to deposit oxygen-deficient bismuth oxide
(Bi2O2.33) and b-Bi2O3 films on SrTiO3(100) and MgO(100)
substrates via LPCVD. Only at substrate temperatures above
600  C was the deposition of iodine-free bismuth oxide films
achieved. The solid precursor, BiI3, was maintained at 200  C
during deposition. Kim et al. used the extremely pyrophoric
alkylbismuth liquid precursor, trimethylbismuth BiMe3, to de-
posit 1D nanostructured a-Bi2O3 films at 450  C using an oxy-
gen reactive gas.284 Due to its high volatility and pyrophoricity,
the precursor was maintained at 5  C during deposition.
As a consequence of its low air and moisture sensitivity,
triphenylbismuth, BiPh3, is the most commonly used dual-
source CVD precursor to bismuth oxide films.285 Similarly,
[Bi(o-tol)3] and [Bi(p-tol)3] have been shown to be useful
bismuth precursors and their gas-phase transport mechanisms
have been studied extensively.286 Tondello et al. deposited
phase-selective Bi2O3 on SiO2 and Al2O3 substrates at 420  C
using BiPh3 and oxygen gas via LPCVD.287 At a system pressure
of 6 mbar and at low gas flow rates, the b-phase was deposited,
whereas mixed a- and b-phase films were grown at higher flow
rates. The crystalline phase obtained depends strongly on the
oxygen flow rate and on the nature of the substrate. Recently,
Caussat288 used a fluidized-bed MOCVD procedure at AP to
deposit mixed-phase bismuth oxide films on the surface of
alumina particles.
1040 Precursor Chemistry Main Group Metal Oxides
There are numerous examples of potential single-source
precursors that have been described in some detail by Mehring
in his comprehensive review in 2007.289 For an in-depth anal-
ysis of single-source precursors to bismuth oxide, readers are
directed to that review. What follows below is a brief overview
covering the main classes of precursors to bismuth oxide. The
basic requirements for incorporation of oxygen into a mole-
cule using sterically bulky ligands to inhibit oligomerization
have led to the development of three main types of potential
single-source precursors to bismuth oxide:
1. Alkoxides, [M(OR)3], R alkyl, aryl, donor-functionalized,
CX3 (X halogen).
2. b-Diketonates, [M(R1COCHCOR2)3], R alkyl, aryl, CX3
(X halogen).
3. Carboxylates, [M(O2CR)3], R alkyl, aryl, CX3
(X halogen).
The first synthesis of bismuth alkoxides was reported in
1960 by Mehrotra and Rai290; however, structural characteri-
zation of a bismuth alkoxide was only reported in 1990 by
Buhro with the synthesis of the polymeric bismuth methox-
yethoxide [Bi(m-OCH2CH2OMe)2(OCH2CH2OMe)]1 (79) via
alcoholysis of bismuth amide [Bi(NMe2)3].291 A further study
by Hubert-Pfalzgraf and coworkers in 1991292 investigated the
synthesis of bismuth alkoxides first via the alcoholysis of the
less sensitive bismuth silylamide [Bi{N(SiMe3)2}3], and
second from the metathesis reaction between bismuth halides
and sodium alkoxides. Compound 79 was found to be poly-
meric and not volatile, whereas [Bi(OtBu)3] (80) was volatile,
but at the time predicted to be polymeric in the solid state. Kou
et al. recently prepared several bismuth aryl alkoxides via the
alcoholysis of [Bi{N(SiMe3)2}3],293 whereas Knispel et al. syn-
thesized a range of bismuth allyl oxides via alcoholysis of 80
and also via salt metathesis reactions starting from BiCl3 and
sodium allylates.294 Highly fluorinated alkoxides of the type
[Bi(OCH(CF3)2)3] were synthesized via the reaction of BiCl3
with the in situ generated NaOCH(CF3)2295,296; however, these
compounds decompose upon heating to form BiOF. The reac-
tion of BiPh3 and pentafluorophenol (C6F5OH) in refluxing
toluene also yielded the dimeric bismuth pentafluorophenox-
ide complex [Bi(OC6F5)3.C7H8)]2. The various synthetic routes
to bismuth(III) alkoxides are summarized in Scheme 11.
Alkoxides of bismuth appear to exhibit poor solubility,
poor volatility, and high air and moisture sensitivity. Jones
and coworkers297 synthesized the donor-functionalized mono-
meric bismuth alkoxide [Bi(mmp)3], (81; Figure 4) via the
BiCl3 + 3 MOR [Bi(OR)3]
- 3 MCl
M = Li, Na, K
R = Me, Et, iPr, iBu (80), OCPh3, CH2CH2OMe (79)
[Bi(OR)3] + 3 ROH [Bi(OR)3]
- 3 ROH
[BiPh3] + 3 C6F5OH [Bi(OC6F5)3]
- 3 C6H6
Scheme 11 Common synthetic routes to bismuth(III) alkoxides.
reaction of [Bi{N(SiMe3)2}3] and 3 molar equivalents of the
free alcohol in diethyl ether. Although the synthesis was first
reported in 1993 by Hermann,298 the molecular structure and
use in thin film deposition were only confirmed in 2001.
Compound 81 is particularly volatile, and when dissolved in
heptane was shown to deposit bismuth oxide films in conjunc-
tion with oxygen on silicon substrates at 250550  C. Jones and
coworkers described one deposition using 81 as a single-source
precursor, whereby no reactive gas was added to obtain a film.
The resulting bismuth oxide films were low in carbon contam-
ination and growth rates appeared to be constant over a wide
temperature range. The decomposition properties of 81 suggest
that it has excellent compatibility for use as a co-precursor with
strontium and tantalum alkoxides for the deposition of ferro-
electric strontium bismuth tantalate (SBT) films.
Bismuth tert-butoxide, [Bi(OtBu)3] (80), has generally been
regarded by researchers as an unsuitable precursor to bismuth
oxide owing to its high air and moisture sensitivity. However,
Blackman299 and coworkers have recently deposited phase-
selective Bi2O3 films on glass substrates via LPCVD of 80
without a reactive gas. Substrate temperatures of 425500  C
were used, and the system pressure and carrier gas flow rates
were varied. This study showed that the volatility of this pre-
cursor is intermediate between [Bi(thd)3] (82) and 81, and its
decomposition pathway suggests clean formation of Bi2O3,
possibly as a result of a b-hydride elimination process. The
main advantage of using 80 for film growth via CVD is its
ability to be used in conventional MOCVD apparatus and
evaporators; therefore, its use is not confined to DLI-CVD.
The films produced in this study were used to photo-oxidize
water using sacrificial reagents in order to highlight the poten-
tial of Bi2O3 for photocatalytic water splitting.
The range of structurally characterized bismuth(III)
b-diketonates found in the literature that do not contain bulky
aromatic substituents are relatively low in comparison to the
alkoxides, and hence only one such molecule has been used as a
CVD precursor to Bi2O3 films. Bismuth tetramethylheptanedio-
nate ([Bi(thd)3] 82), despite containing an excess of oxygen
within the coordination sphere of bismuth, has been used in
conjunction with oxygen to deposit amorphous bismuth oxide
films on silicon substrates between 240 and 435  C.300
Although compound 82 possesses a lower decomposition tem-
perature in comparison to bismuth halides and alkylbismuth
compounds, it has poor thermal stability, undergoing prema-
ture weight loss according to many TGA studies which limits its
use to DLI-CVD apparatus.286 The first reported synthesis of a
[Bi(NR2)3] + 3 ROH
- 3 HNR2
R = Me, SiMe3
 Precursor Chemistry Main Group Metal Oxides 1041

O Bi
O

O
O

Figure 4 Structure of compound 81. Hydrogen atoms omitted for clarity.

BiCl3 + 3 MR [BiR3] [Bi(NR2)3] + 3 RH

- 3 MCl - 3 HNR2
M = Li, Na, K R = thd (82)
R = acac, thd (82), fod (83), hfac (84) R = Me, SiMe3

[BiR3]
BiPh3 + 3 RH
- 3 C6H6
R = fod (83), hfac (84)

, hexane
BiPh3 + 2 RH [BiPhR2(L)]
- 2 C6H6 + L
R = hfac
L = MeCN, DMSO, THF, etc.
Scheme 12 Various synthetic routes to bismuth(III) diketonates.

bismuth(III) diketonate was compound 82, published by
Hubert-Pflazgraf in 1991.292 The compound was synthesized
via a salt metathesis reaction of BiCl3 and Na(thd) and the
volatile product (82) was purified by sublimation (110  C,
104 mmHg), although attempts to grow single crystals via
sublimation led to decomposition.
Brooks et al.301 synthesized a number of volatile bismuth
diketonates intended for MOCVD application. The reaction of
the free ligand with BiPh3 in refluxing toluene yielded [Bi(fod)3]
(83) and [Bi(hfac)3] (84). It was reported that the same syn-
thetic procedure when using the thd ligand yielded no reaction
due to its lower acidity, and, the reaction with HTfac produced a
mixed phenyl/diketonate complex. The authors utilized 83 in
a hot-wall LPCVD reactor to produce bismuth films which
deposited on the walls of the reactor and on pyrex substrates.
The reactor temperature was maintained at 400450  C using a
quartz-halogen lamp and 83 was heated to 90  C. No carrier or
oxidizing gas was used. Analysis of the resulting brown-colored
film revealed that it consisted of bismuth 73%, carbon 15%, and
oxygen 12%, with carbon levels increasing toward the hottest
area of the reaction chamber; however, all the films had surpris-
ingly low fluorine contamination (below detector limit). The
crystal structure of 84 was reported by Dikarev et al. in 2005 as
monomeric and it was described as a useful reagent in the
synthesis of heterobimetallic bismuth-transition-metal com-
plexes.302 The same group synthesized a range of phenyl bis-
muth b-diketonate complexes of the type [BiPh(hfac)2(L)] from
refluxing BiPh3 with 2 equiv. of hexafluoroacetylacetone, but
were only able to grow single crystals with the addition of
coordinating solvents including MeCN, Me2CO, THF, DMA,
1042 Precursor Chemistry Main Group Metal Oxides
DMSO, and PhCN (L).303 The synthetic routes to bismuth b-
diketonates are summarized in Scheme 12.
The synthesis and decomposition study of a number of bis-
muth(III) carboxylates were carried out by Pisarevsky to find
suitable precursors to bismuth-containing superconductors.304
Bismuth acetate [Bi(O2CCH3)3] (85) was prepared via reaction
of bismuth oxide Bi2O3 with acetic acid in the presence of acetic
anhydride. Detailed experimental procedures were not given;
however, it was reported that 85 can further react with a variety
of carboxylic acids to yield other bismuth carboxylates. Indeed,
bismuth(III) 2,2-dimethylpropanoate, [{Bi(O2CCMe3)3}4] (86),
was synthesized via a ligand exchange reaction between 86 and
2,2-dimethylpropanoic acid and purified via sublimation in a
sealed ampoule possessing a temperature gradient between 120
and 160  C. The structure of compound 86 was shown to be
composed of isolated tetrameric [Bi(O2CCMe3)3]4 groups.305
The solvent-free synthesis of a range of bismuth(III) carboxylates
was published by Andrews in 2006.306 The bismuth tri-
substituted carboxylates were synthesized via refluxing BiPh3 in
toluene with a host of carboxylic acids possessing different pKa
values as well as in solvent-free conditions. Reaction of BiPh3 with
2-picolinic acid resulted in the formation of only the di-
substituted tetrameric complex [PhBi(2-(C5H4N)CO2)2]4 via
both routes. In contrast, reaction with p-nitrobenzoic acid yielded
the tri-substituted complex via reflux in toluene and the di-
substituted complex [PhBi(p-O2NC6H4CO2)2]n (87) under
solvent-free conditions. The structure of compound 87 was
found to be polymeric, as shown in Figure 5.
The volatility of the bismuth carboxylate complexes is
rather low in comparison to the alkoxide and the diketonate
N
Bi
N which has two polymorphic forms: orthorhombic b-TeO2 (tell-
N Bi N
N Bi N
Bi
N potential uses in numerous optical applications. Most thin-
N tellurium(IV) alkoxides, and these are discussed later.
Figure 5 Structure of compound 87. Hydrogen atoms omitted for
clarity.
compounds, and any vaporization is usually accompanied by
some decomposition, which has made these complexes virtu-
ally unusable for most CVD processes. There appear to be no
references in the literature to the use of bismuth carboxylates as
single- or dual-source precursors to bismuth oxide films;
however, their use in binary and ternary oxide materials
synthesized via solution deposition techniques is more
commonplace.
There are few examples of the growth of bismuth oxide
films by ALD, owing to the general trend of poor thermal
stabilities of bismuth precursors which such a deposition pro-
cess seeks to avoid. The ALD reaction of BiPh3 and O3 resulted
in deposition of dark and patchy films, possibly composed of
metallic bismuth,307 and the use of H2O2 (instead of O3) also
proved unsuccessful.308 Vehkamaki et al.309 synthesized a
range of bismuth(III) amido and thioamido complexes to
assess their suitability for the growth of ferroelectric BiTiO
films in conjunction with suitable titanium precursors. The
bismuth silylamide [Bi{N(SiMe3)2}3] was chosen for further
study due to its high volatility and thermal stability, resulting
in deposition of BiOx films using H2O as a reactant. However,
poor reproducibility in forming the binary bismuth oxide
phase was encountered. The same group reported the syntheses
and structures of a number of bismuth(III) alkoxides as
potential ALD precursors for bismuth oxide.310 The use of
[Bi(OtBu)3] (80) has been suggested for the ALD of bismuth
oxide but proved unsuccessful in this methodology due
to decomposition of the precursor. The structures of 80 and
[Bi(OCiPr3)3] were solved by single-crystal XRD; however, only
the alkoxide [Bi(OCMe2iPr)3], in conjunction with an H2O co-
reactant, was suitable for ALD of bismuth oxide films due to its
comparatively higher thermal stability. This precursor melts at
40  C and an evaporator temperature of 85  C was employed.
Films grown at 150  C were amorphous, whereas those grown
above 200  C contained an unidentified Bi2O3 phase. On
crystallization from toluene single-crystal XRD showed that
dimeric [Bi2O(OCMe2iPr)4]2 (88) was formed (Figure 6).
1.33.4.4 Tellurium Oxide Precursors
The common stable oxide of tellurium is tellurium(IV) oxide,
urite) and tetragonal a-TeO2 (paratellurite), with the a form
converting to the b form under high-pressure conditions or at
high temperatures. The main use of tellurium oxide is found in
acousto-optical materials, as the acousto-optical figure of merit
for single-crystal paratellurite is 1.2  1015 s3 g1 in the visible
region of the electromagnetic spectrum one of the largest for
transparent materials. This value, along with its stability, dura-
bility, wide bandgap, and high dielectric constant, leads to
film deposition studies of tellurium oxide have made use of
CVD studies of tellurium oxide are extremely scarce. There
appear to be only two reports in the literature of the synthesis
of TeO2 films via CVD methods, both by Moiseev et al., and
employing the same organotellurium precursors, dimethyltel-
lurium TeMe2 and diethyltellurium TeEt2, for the formation of
ZnOTeO glasses as refractory materials in optical fibers.311,312
Deposition of crystalline TeO2 and Zn2Te3O8 powders was

O
O
O
Bi
O
Bi
O
Figure 6 Structure of compound 88. Hydrogen atoms omitted for clarity.
carried out between 420 and 650  C, and further heat treat-
ment led to the formation of glasses.
The solgel chemistry of TeO2 has, in comparison, been
studied in more detail. Hodgson et al.313 synthesized films of
TeO2 via a solgel process incorporating tellurium(IV) iso-
propoxide, [Te(OiPr)4] (89), in isopropanol, followed by ad-
dition of 1,3-propanediol and nitric acid as a precursor
solution. After dip coating the samples in the solution, the
films were annealed at 80600  C in air; films heated below
280  C were amorphous and those heated at 340  C contained
mixtures of metallic tellurium, a-TeO2, and b-TeO2 from XRD
analysis. In contrast, films heated at 450  C consisted purely of
paratellurite a-TeO2. It was noted that solutions consisting of
pure 89 in isopropanol were extremely sensitive to hydrolysis
in air, forming a white precipitate within seconds of exposure.
Coste et al.314 also used compound 89 in an isopropanol/citric
acid solution to form crystalline TeO2 powders via the solgel
method. Heat treatment at 250  C yielded metallic tellurium;
at 330370  C, mixtures of metallic tellurium, a-TeO2, and
b-TeO2 were formed, and at 400  C pure a-TeO2 was formed.
Compound 89 was also utilized in conjunction with stabilizers
such as acetylacetone, 1,3-propandiol, and acetic acid to con-
trol precursor hydrolysis, before dip coating glass substrates in
the solution followed by drying.315 Acetylacetone and acetic
acid were found to be unsuitable as stabilizers as they form
insoluble reaction products, notable metallic tellurium. The
solution of 89 in 1,3-propanediol effectively suppressed hy-
drolysis of the precursor and the resulting films were com-
posed of crystalline a-TeO2 after annealing at 380  C.
Solutions of tellurium ethoxide [Te(OEt)4] in 1,2-
propanediol and ethylene glycol have also been demonstrated
as suitable solgel precursors to films of TeO2TiO2 in con-
junction with [Ti(OiPr)4] as the titanium precursor.316 Films
on glass substrates were synthesized by dip coating the glass
slides in the sol followed by drying, then heating between 350
Precursor Chemistry Main Group Metal Oxides 1043
O
Bi
O
O
Bi
O
O
and 600  C. Mixtures of TiO2, a-TeO2, TiTe3O8, and metallic
tellurium were present at films annealed at 500 and 600  C,
and those heated at 350  C consisted of metallic tellurium
only. The bandgap for the amorphous film heated at 450  C
was calculated as 3.29 eV.
Ikeda et al.317 synthesized BaOTeO2 films via a solgel
route using an air-stable tellurium(VI) precursor. The precursor
was synthesized via the reaction of orthotelluric acid (H6TeO6)
and ethanol in a sealed silica tube at 150  C for 12 h, followed
by reflux and subsequent distillation to obtain an ethanolic
solution of tellurium(VI) ethoxide. The solution was added to
aqueous barium acetate and spin coated on glass and Pt sub-
strates to yield BaOTeO2 films after annealing. Only films
heated at 550  C were crystalline and contained XRD peaks
ascribed to TeO2, BaTe2O6, and BaTe2O5.
Problems that were encountered when using extremely
moisture-sensitive tellurium alkoxides were overcome by Wei
et al. via a nonhydrolytic solgel process to synthesize thin films
of TeO2.318 The group used tellurium(IV) 1,2-propylene digly-
collate [Te(O2C3H6)2], synthesized via the reaction of tellurium
dioxide and 1,2-dipropylene glycol in the presence of catalytic
amounts of p-toluene sulfonic acid under reflux conditions.
Films were obtained via spin coating on glass substrates fol-
lowed by postdeposition annealing. Heating at 450  C led to
crystallization of a-TeO2, whereas heating at 300  C led to me-
tallic tellurium; those heated at 350 and 400  C were composed
of mixed a- and b-TeO2 and tellurium metal. Lower heating
temperatures resulted in amorphous films.
The syntheses of a range of volatile homoleptic monoden-
tate tellurium(IV) alkoxides have already been discussed by
Bradley319 in his comprehensive review of metal alkoxides
and are not covered in detail here. However, more recently a
range of donor-functionalized tellurium(IV) alkoxides were
synthesized by Fleischer and coworkers320 via the same alco-
holysis procedure in benzene, using 89 as a starting material
1044 Precursor Chemistry Main Group Metal Oxides
and utilizing equimolar amounts of donor-functionalized al-
cohols as ligands. The alkoxides [Te(OiPr)3(OCH2CH2NMe2)],
[Te(OiPr)3(OCH2CH2OMe)], [Te(OiPr)3(OCH2CH2SMe)],
and [Te(OCH2CH2NMe2)4] were synthesized via this route.
Related dichloridotellurium(IV) alkoxides and trichloridotel-
lurium(IV) alkoxides were synthesized via the reaction of TeCl4
and the related silylated alcohol Me3SiOCH2CH2XMen (X N,
n 2; X O or S, n 1). The trichloridotellurium(IV) alkox-
ides, [TeCl3OCH2CH2NMe2] (90) and [TeCl3OCH2CH2OMe]
(91), and the dichloridotellurium alkoxide, [TeCl2(OCH2-
CH2OMe)2] (92), were synthesized in this manner (eqn [18]).
TeCl4 + n Me3SiOR [TeCl4-n(OR)n]
-n Me3SiCl
90; n = 1; R = CH2CH2NMe2
91; n = 1; R = CH2CH2OMe
92; n = 2; R = CH2CH2OMe
18
The volatilities and decomposition products of the com-
pounds 9092 were not investigated; however, the investigators
noted that the donor atom L in the alkoxide ligand facilitates a
TeL donoracceptor interaction which stabilizes the chlorido-
tellurium compounds and reduces the reactivity of the tellurium
center toward nucleophiles. Similarly, binary mixtures of
[Te(OMe)4], [Te(OiPr)4], [Te(OMe)3Cl], and [Te(OiPr)3Cl]
could be produced from stoichiometric amounts of TeCl4 and
[Te(OiPr)4] via ligand exchange and redistribution.321
1.33.4.5 Ternary Group 15 Oxide Precursors
A broad range of applications are possible for bismuth oxide-
based ternary and higher oxide materials, particularly in the
fields of ferroelectrics, multiferroics, photocatalysts, and
semiconductor-based materials. As a result, the drive to incor-
porate bismuth and other metal atoms within the same mole-
cule for use as potential single-source precursors to advanced
oxide materials has accelerated within the past two decades.
Somewhat surprisingly, although the list of potential single-
source precursors to some of these materials is fairly extensive,
those utilized for the deposition of heterobimetallic oxide thin
films are virtually nonexistent. There are, however, numerous
examples of such single-source molecular precursors decom-
posing to the desired metal oxide powders upon thermal treat-
ment, inside and outside of oxidizing conditions, but there are
no examples of heterobimetallic thin films incorporating bis-
muth that were grown from these precursors. Most of the
heterobimetallic bismuth-transition-metal complexes and
some bismuth-main-group metal compounds have been de-
scribed in detail in a comprehensive review of bismuth oxide
precursors published in 2007.289 What follows below is a brief
overview covering the main classes of complexes incorporating
bimetallic species of bismuth.
The bismuthtitanium alkoxide [BiTi2(m3-O)(m-OiPr)4
i
(O Pr)5] (93) was synthesized from the reaction of
titanium isopropoxide [Ti(OiPr)4] and bismuth oxoisopropox-
ide, which was formed in situ by controlled microhydrolysis
(in a toluene/water solution) and subsequent alcoholysis of [Bi
(OtBu)3].322 The structure of 93 contains the bismuth atom in
a four-coordinate environment with a stereochemically active
lone pair with isopropoxide bridging groups between the metal
O
O
O
Ti
O*
O
Bi O
O
Ti
O O
O
Figure 7 Structure of compound 93. Hydrogen atoms omitted for
clarity.
centers, leading to a six-coordinate titanium center in a distorted
octahedral coordination (Figure 7). The mixed-metal ethoxides,
[BiTi2O(OEt)9] and [Bi4Ti3O4(OEt)16], were also prepared using
[Ti(OEt)4] by varying the hydrolysis ratio. Pure bismuth titanate
Bi4Ti3O12 powder was obtained via hydrolysispolycondensation
of compound [BiTi2O(OEt)9] at 450  C.
The bifunctional salicylate ligand, (O2CC6H4-2-OH), was
used to synthesize the bimetallic bismuth-transition-metal spe-
cies [Bi2M2(sal)4(Hsal)4(OR)4] (94, M Nb; 95, M Ta),
[Bi2Ti3(sal)8(Hsal)2] (96), and [Bi2Ti4(OiPr)(sal)10(Hsal)] (97)
using in situ formed bismuth salicylate [Bi(HSal)3] (98) and its
reaction with metal alkoxides [M(OCH2CH3)5] (M Nb, Ta) or
[Ti(OiPr)4]. Thermal treatment of compounds 9497 at 500  C
in air resulted in the formation of Bi4Ti3O12 and Bi5Nb3O15, as
well as the expected oxides BiMO4 (M Nb, Ta) and Bi2Ti4O11.
A range of similar bismuth-transition-metal species but with
varying numbers of salicylate and alkoxide ligands have also
been prepared as potential single-source precursors to ferroelec-
tric bismuth-containing oxides.323325 The Lewis acidity of the
bismuth atom in 98 allows for easy formation of bimetallic
bismuth-transition-metal species via its reaction with metal
b-diketonates.326 The complexes [Bi2(Hsal)6 M(acac)3] (99,
M Al; 100, M V; 101, M Cr; 102, M Fe; 103, M Co)
were synthesized under mild conditions and all contain a 2:1
bismuth to metal ratio. Upon thermal treatment of 99 and 100
at 750  C, crystalline heterometallic oxides Bi2Al4O9 and
Bi4V2O11 were formed. Nanoparticulate BiVO4 was the major
crystalline product formed after thermal treatment of [BiV(O)
(Hsal)(sal)(salen*).CH2Cl2] (104, salen* ethylenebis(3-
methoxysalicylimine)). Compound 104 was formed via reac-
tion of 98 and [V(O)(salen*)], with the resultant molecule
possessing a 1:1 ratio of metal atoms.327
 Precursor Chemistry Main Group Metal Oxides 1045

A range of stable complexes of the type [Bi2M(b-diketo- (OSiMe2tBu)]2 (116) and [(CO)3FeBi3(OtBu)4{OCO(OtBu)}]2
nate)8] (105, M Mn; 106, M Fe; 107, M Co; 108, M Ni; (117).332 Compound 116 undergoes thermal decomposition in
109, M Cu; 110, M Zn) were synthesized using [Bi(hfac)3] air to yield [Bi4(SiO4)3] and Bi2Fe4O9, as shown by PXRD analysis
and either [M(hfac)2] or finely divided metals.302 These reac- and TGA of the residual compound. In contrast, compound 117
tions were carried out in a sealed ampoule which was warmed decomposes in a helium atmosphere to afford BiFeO3 with a
at 7090  C to yield crystals of the desired product at the colder minimal Bi25FeO39 impurity phase.
end of the ampoule after a few days. The structures of 105110 Antimony is commonly used as a dopant in numerous
simply contain two [Bi(hfac)3] units that sandwich a [M metal oxides, including SnO2 as described in Section 1.33.5.3,
(hfac)2] molecule and are held together via Lewis acidbase in order to achieve improved electronic properties in the
interactions between one oxygen molecule of [Bi(hfac)3] and
the transition-metal(II) center. It is not yet known whether the
complexes decompose to the desired mixed-metal oxides. A
somewhat unexpected paddle-wheel structure was obtained
for the bimetallic complexes [BiM(O2CCF3)4] (111, M Rh;
112 Ru) which contain a metalbismuth single bond (Figure 8)
and may be isolated via the equimolar reaction of
[Bi2(O2CCF3)4] and [M2(O2CCF3)4] in a sealed ampoule.328
Thermal decomposition studies of these complexes have not O O
been carried out; so it is not clear whether they may be suitable
O
molecular precursors to mixed-metal oxides.
A range of bismuthmolybdenum complexes were prepared Rh O
primarily using bismuth(III) alkoxides.329,330 The MoO3/Bi2O3 O
materials are industrially important catalysts for the difficult
Bi
Standard Oil of Ohio (SOHIO) process involving oxidation of
propene to acrolein. Roggan et al. reacted [Cp0 2MoH2] O
(Cp0 MeC5H4) with the bismuth alkoxides, 80 and [{Bi O
{OCH(CF3)2}3(thf)}2], to yield [Cp0 2Mo{Bi(OtBu)2}2] (113,
see Figure 9) and [Cp2Mo(Bi{OCH(CF3)2}2)2] (114), respec- O
tively. Compounds 113 and 114 were formed via alcohol elim-
ination and CdH bond activation.331 Further addition of
[Cp0 2MoH2] afforded [(Cp0 2Mo)2{Bi2}] (115) via additional
alcohol elimination and intramolecular CdH bond activation.
More recently, Wojcik et al. reacted Fe2(CO)9 with the bismuth
alkoxides, [Bi(OSiMe2tBu)3] and [Bi(OtBu)3], which resulted in Figure 8 Structure of compound 111. Hydrogen atoms omitted for
the formation of bismuthiron carbonyl clusters [(CO)4FeBi clarity.

Mo

O
O
Bi Bi

O
O

Figure 9 Structure of compound 113. Hydrogen atoms omitted for clarity.

1046 Precursor Chemistry Main Group Metal Oxides
resulting materials. However, reports of heterobimetallic
complexes containing antimony are limited. Hanna et al.333
have synthesized soluble antimony(III)/molybdenum(VI)
calix[5]arene complexes as potential catalysts for the SOHIO
process. The bimetallic compound [Sb2Mo4O11{tBuC5(H)}2]
was synthesized via reaction of [MoO2(OtBu)2] and [Sb
{tBuC5(H)2}] in dimethoxyethane and heated to 95  C. Recently,
Molloy334 has synthesized the tungstenantimony complex
[Sb2W(NBut)(m2-NBut)(m3-NBut)2Cl3(m2-Cl)] via the reaction of
[Li2W(NtBu)4] and 2 equiv. of SbCl3 in toluene. Trace amounts
of water in the reaction mixture were thought to be responsi-
ble for formation of [Sb2W3Cl7(m2-O)(m3-O)2(NBut)3(NHBut)
(m2-NBut)2]. It should be noted that the aforementioned bime-
tallic compounds have yet to be converted to mixed-metal oxides.
Bismuth silicates Bi12SiO20, Bi2SiO5, and Bi4Si3O12 have
been investigated with regard to potential uses as optical and
dielectric and even ferroelectric materials. These materials are
easily synthesized via mixing of Bi2O3 and SiO2 in the correct
molar ratios; however, thin-film growth relies on the availabil-
ity of precursors that either are soluble in common solvents or
can be easily transported into the gas phase for vapor deposi-
tion. Much work has focused upon deposition of bismuth
oxide onto silicon or silicon dioxide substrates, as the post-
deposition annealing process allows facile interdiffusion be-
tween the oxide layers to form bismuth silicates quite
effectively. There are, however, some single-source approaches
to depositing these materials, mainly via ALD, which may be
advantageous if silicon-containing substrates are not used.
Massiani et al. reported the synthesis of the poorly soluble
monomeric [Bi(OSiPh3)3(thf)3] complex via the alcoholamine
exchange of [Bi{N(SiMe3)2}3] with Ph3SiOH.292 Bismuth silox-
ides, of the type [Bi(OSiR3)3] (118, R Me; 119, R Et, 120,
R iPr), were synthesized via a salt metathesis route using bis-
muth halides and alkali metal/silanols, or via alcoholysis of
80.335,336 The possible synthetic routes to bismuth siloxides are
summarized in Scheme 13. The salt metathesis route is generally
disfavored, however, due to a tendency to form sodium
bismuth oxo clusters.337 Compounds 118 and 120 are highly
air-sensitive solids, whereas 119 is liquid at room temperature
but all are very soluble in common organic solvents. Compound
118 is trimeric in the solid state, as a result of weak intermole-
cular BiO interactions; however, the sterically hindered 120
possesses a monomeric structure. TGA studies of 118 revealed a
rapid weight loss shortly after melting at 80  C with a residual
mass suggesting formation of Bi2O3; however, PXRD analysis of
the residue instead indicated formation of Bi12SiO20. Com-
pound 120 also decomposed to silicate Bi12SiO20 along with
Bi4Si3O12 and a-Bi2O3, although with a weight loss of 77% this
indicated partial evaporation of the precursor, suggesting
- 3 MCl
BiCl3 + 3 MOSiR3 [Bi(OSiR3)3]
M = Li, Na, K
R = Me, Et, iPr
- 3 ROH
[Bi(OR)3] + 3 R3SiOH [Bi(OSiR3)3]
R = tBu (80)
Scheme 13 Synthetic routes to bismuth(III) siloxides.
potential suitability for CVD experiments. TGA experiments
using liquid 119 were not reproducible. It was suggested that
the poor thermal stability of these precursors (none could be
purified via sublimation) could hamper their use as CVD
precursors; however, their good solubility may allow use in
solgel and liquid delivery CVD systems. Complexes, of the
type [Bi{OSi(OtBu)2R}3] (R OtBu, Ph), were also synthesized
via the salt metathesis and alcoholysis routes but their structures
were not reported.338
Mansfield et al. synthesized dimeric [Bi(OSitBuPh2)3]2 (121)
in low yields via reaction in the melt of BiPh3 and tBuPh2SiOH
over 7 days, which also resulted in some decomposition to
metallic bismuth.339 Alcoholysis of 80 with tBuPh2SiOH
resulted in a better yield of 121. The formation of a dimer for
121 is linked to the weak bismuthp-arene interactions [3.340
(7) A] present in the molecule. TGA studies of 121 showed that
the compound has a high thermal stability with a single-step
weight loss occurring at 250  C; however, thermolysis in argon
resulted in identification of elemental bismuth as the only crys-
talline species present. Perhaps the most intriguing bismuth
siloxide compound is that of [Bi(OSiMe2tBu)3] (122), formed
via alcoholysis of 80.340 The molecular structure of 122 is mo-
nomeric, with the central bismuth adopting a trigonal pyrami-
dal geometry. It can be sublimed at relatively mild conditions
(150  C, 102 Torr) and is soluble in most organic solvents,
suggesting a potential use in CVD experiments. The most inter-
esting aspect concerning 122 is the product obtained via its
hydrolysis when dissolved in benzene, which yielded the bis-
muth oxo cluster [Bi22O26(OSiMe2tBu)14] (123), containing a
[Bi22O26]14 core surrounded by a hydrophobic layer of 14
t
BuMe2SiO ligands. The core structure symmetry of 123
(Figure 10) was said to be closely related to that of b-Bi2O3,
which may explain the formation of b-Bi2O3 upon thermolysis
of the cluster at 350  C. The formation of Bi12SiO20 from 123
occurred at higher decomposition temperatures.
There are very few reports describing the growth of bismuth
silicate thin films via CVD. Kim et al.284 used BiPh3 in the
presence of oxygen gas to deposit bismuth silicate films
(Bi4Si3O12) on silicon substrates at 750  C. The SiO2 layer of
the substrate was converted to bismuth silicate through reac-
tion with the bismuth precursor and oxygen gas with no silicon
precursor required. Postdeposition heat treatment of the films
led to an increase in crystallinity and had a similar effect upon
capacitance, highlighting their potential use either as buffer
layers for growth of other oxide layers on silicon or as capaci-
tance layers in RAM devices. Similarly, Kang et al.300 noticed
reaction of the silicon substrate with the CVD-deposited bis-
muth oxide film to yield crystalline Bi4Si3O12 after annealing at
750  C. Nagao341 grew crystalline epitaxial Bi12SiO20 films on
- 3 HNR2
[Bi(NR2)3] + 3 R3SiOH
R = Ph
R = Me, SiMe3
in the melt
BiPh3 + 3 R3SiOH
160 C
R = tBuPh2 (121)

Bi12SiO20 substrates via CVD using bismuth metal (main-
tained at 800  C) and either gaseous tetramethoxy- or
tetraethoxy-silane as the silicon source. Oxygen gas was used
as a co-reactant and the substrate temperature was maintained
at 750  C. The ratio of bismuth to silicon observed in the films
varied greatly as a function of precursor delivery rate. Clearly,
the disadvantage of using this CVD method stems from the
Bi
O (TEOS) as precursors and acetic acid and 2-methoxyethanol as
Si
C of secondary g-Bi2O3 phase present. Klebanskii et al.345 also
Figure 10 Molecular structure of the bismuth oxo-cluster [Bi22O26-
(OSiMe2tBu)14] 123. Hydrogen atoms omitted for clarity.
Si
Figure 11 Structure of the ALD precursor 124. Hydrogen atoms omitted for clarity.
Precursor Chemistry Main Group Metal Oxides 1047
high temperature required for transportation of the bismuth
source.
Harjuoja et al. demonstrated the use of the silyl bismuth
precursor [Bi(CH2SiMe3)3] (124) in conjunction with O3 as an
ALD precursor for deposition of bismuth silicate Bi2SiO5
films.342 The molecular structure of 124 consists of a bismuth
atom sitting at the top of a trigonal pyramid (Figure 11); TGA
studies revealed that 124 was volatile with complete evapora-
tion taking place by 200  C in one step. Films were deposited
from 124 and annealed at 600800  C in oxygen and nitrogen
to obtain crystalline samples but this also resulted in some
Bi12SiO20 as a minor phase. Samples annealed at 1000  C in
nitrogen were composed of Bi4Si3O12, and it is believed that as
the annealing temperature increases, the higher the chance of
bismuth evaporation or diffusion into the substrate. The resul-
tant films were found to be electrically insulating. The same
group later introduced BiPh3/O3 ALD cycles into the film
deposition process in order to control the bismuth content in
the BiSiO films.307 The use of the silylamido complex [Bi{N
(SiMe3)2}3] in conjunction with O3 using ALD was also pre-
dicted to result in silicates.309
Thin films of bismuth silicate (Bi4Si3O12) were synthesized
by Armelao et al.343 by dipping silica substrates in ethanolic
solutions of 82 followed by heat treatment in air at 700  C. The
high moisture sensitivity of 82 restricted the dip coating to be
carried out in a N2 atmosphere glovebox. The formation of
bismuth silicates was due to reaction of bismuth oxide with
the silica substrate at these high temperatures. Annealing at
temperatures between 600 and 700  C resulted in the formation
of Bi2SiO5, whereas between 300 and 500  C only bismuth
oxide films were produced. Jain et al.344 employed a dual-source
approach to synthesize Bi12SiO20 films on quartz and platinum
substrates using bismuth nitrate and tetraethyl orthosilicate
solvents. Films were spin coated, then annealed at 750  C which
resulted in crystalline Bi12SiO20 films, albeit with large amounts
used the same bismuth and silicon precursors together with
acetylacetone as a stabilizer and complexing agent to synthesize
Si
Bi
Si
1048 Precursor Chemistry Main Group Metal Oxides
Bi12SiO20 films via a solgel process. To obtain thicker films,
each layer was annealed at 400  C then re-coated via spin-coat-
ing, after which heat treatment at 650  C for 1 h resulted in
crystalline films. More recently, Veber et al.346 synthesized thin
films of Bi12SiO20 films using bismuth nitrate and TEOS as
precursors with ethanolamine as a stabilizer via spin coating
onto Pt/TiO2/SiO2/Si, sapphire, and MgAl2O4 substrates. For-
mation of metastable Bi2SiO5 was observed following heat
treatment at 400  C, with complete transformation to the de-
sired sillenite phase Bi12SiO20 at 700  C. The effect of varying
the solvent used in the solgel process was also investigated.347
A spray deposition technique was utilized by Ortiz and
coworkers348 to deposit bismuth oxide films on quartz and
silicon substrates, which were subsequently annealed in air for
2 h, enabling conversion to bismuth silicates. Bismuth acetate
and bismuth chloride dissolved in methyl alcohol were used as
bismuth precursors. Crystalline Bi2SiO5 was obtained via
annealing at 600  C and Bi4Si3O12 was observed at 750  C,
indicating diffusion between the bismuth oxide and silicon
dioxide layers. Bandgaps for the as-deposited bismuth oxide
films were calculated as 3.78 eV, and increased with conversion
to the silicate phase, up to 4.89 eV for Bi4Si3O12.
1.33.5 Conclusion
The previous sections clearly demonstrate how research has
progressed on precursor development for the deposition of
thin films of main-group metal oxides. Thin films of binary
main-group metal oxides exhibit a range of interesting proper-
ties and are used in many applications, as described in
Sections 1.33.21.33.4. An advantage of using binary com-
pounds is the relative ease of controlling the chemical compo-
sition in film depositions, compared to using ternary
compounds and multicomponent oxides. However, some of
the key applications of these films actually require the use of
impurity-doped metal oxides or ternary/higher metal oxides.
For example, multicomponent semiconductor oxides mainly
composed of elements such as indium, zinc, tin, or gallium are
a very promising new class of materials for application in
transparent electronics, multifunctional sensors, and other
electronic applications. As more flexible devices are required
to incorporate into a range of applications, for example, flexi-
ble displays for e-readers and organic light-emitting diodes
(OLEDs), flexible barrier layers such as TCOs will be in de-
mand. Currently, TCOs are widely used for rigid devices but
the technology used to coat rigid substrates (i.e., glass) is not
always applicable for coating flexible substrates. Solution-
based processes are ideal for coating flexible substrates and
here the development of soluble single-source precursors will
play a crucial role.
For many deposition techniques, such as CVD, the precur-
sors utilized, besides appropriate volatility, should also exhibit
stable vapor pressures, and cleanly decompose to the desired
product at useful substrate temperatures. Precursor design rep-
resents a significant challenge which will be important in the
future of thin-film production: there is a need to create not
only compounds with preformed MdO bonds, but also com-
pounds that will undergo clean decomposition under the de-
position conditions. Another consideration is if the precursors
are to be employed in solution-based processes, then they
should be highly soluble. However, these precursors should
also be easily synthesized in high-yielding reactions to be
suitable for scale-up in industrial processes. Careful consider-
ation of both the nature of the oxygen-containing group(s) and
the other ligands attached to the metal center should lead to
exciting reports of new and optimized single-source precursors
appearing in the near future.
Note added in proof
Group 13 beta-ketoiminate Compounds have been reported as
precursors to gallium oxide.349 Beta-ketoiminate gallium hy-
dride compounds, [Ga(L-1)H] and [Ga(L-2)H2] (L-1 R{NC
(Me)CHC(Me)O}2, R (CH2)2; L-2 Me2N(CH2)3NC(Me)
CHC(Me)O) have been used as single-source precursors for
the deposition of Ga2O3 by AACVD with toluene as the solvent.
The quality of the films varied according to the precursor used,
with the complex [Ga(L-1)H] giving by far the best quality films.
Although the films were amorphous as deposited, they could
be annealed at 1000  C to form crystalline Ga2O3. A range of
molecular gallium and indium siloxide precursors have
also been reported.350 Compounds, of the type [Me2M(OR)]2
(M In, R SiEt3 or OSi(OtBu)3, M Ga, R SiEt3 or Si
(OtBu)3), were prepared via the reaction of InMe3 or GaMe3
with the corresponding silanols. Alternatively, [In[OSi
(OtBu)3)] was isolated from the reaction of [In{N(SiMe3)2}3]
with an excess amount of (tBuO)3SiOH. The suitability of 15 to
serve as molecular precursors for low-temperature synthesis of
amorphous indium and gallium oxide for electronic applica-
tions was investigated and, with the exception of [In[OSi
(OtBu)3)] which form indium silicate, yielded indium or gal-
lium oxide. Driess and co-workers have also further developed
a range of bimetallic indium(III)-tin(II) alkoixdes which have
been used for the formation of ITO films.351 Thermal treat-
ment of [(tBuO)Sn(m-(OBu)2InX2] (X OtBu, N(SiMe3)2)
resulted in the formation of amorphous tin-rich ITO and crys-
talline ITO with traces of the cassiterite phase of tin dioxide.
References
1. van Mol, A. M. B.; Chae, Y.; McDaniel, A. H.; Allendorf, M. D. Thin Solid Films
2006, 502, 7278.
2. Elam, J. W.; Routkevitch, D.; Mardilovich, P. P.; George, S. M. Chem. Mater.
2003, 15, 35073517.
3. Sundqvist, J.; Tarre, A.; Rosental, A.; Harsta, A. Chem. Vap. Deposition 2003, 9,
2125.
4. Moon, O. M.; Kang, B.-C.; Lee, S.-B.; Boo, J.-H. Thin Solid Films 2004,
464465, 164169.
5. Putkonen, M.; Niinisto, L. Thin Solid Films 2006, 514, 145149.
6. Wu, Q. H.; Fortunelli, A.; Granozzi, G. International Reviews in Physical Chemistry
2009, 28, 517576.
7. Niskanen, A.; Arstila, K.; Ritala, M.; Leskela, M. J. Electrochem. Soc. 2005, 152,
F90F93.
8. Potts, S. E.; Keuning, W.; Langereis, E.; Dingemans, G.; van de Sanden, M. C. M.;
Kessels, W. M. M. J. Electrochem. Soc. 2010, 157, P66P74.
9. Lee, S.; Jeon, H. Electronic Mater. Lett. 2007, 3, 1721.
10. Katamreddy, R.; Inman, R.; Jursich, G.; Soulet, A.; Takoudis, C. J. Electrochem.
Soc. 2006, 153, C701C706.
11. Katamreddy, R.; Inman, R.; Jursich, G.; Soulet, A.; Takoudis, C. J. Mater. Res.
2007, 22, 34553464.
12. Jeong, C. W.; Lee, B. I.; Joo, S. K. Mater. Sci. Eng. C 2001, 16, 5964.

13. Jeong, C. W.; Lee, J. S.; Joo, S. K. Jpn. J. Appl. Phys. 2001, 40, 285289.
14. Raisanen, P. I.; Ritala, M.; Leskela, M. J. Mater. Chem. 2002, 12, 14151418.
15. Zhang, X. J.; Li, Q.; Zhao, S. Y.; Gao, C. X.; Wang, L.; Zhang, J. Appl. Surf. Sci.
2008, 255, 18601864.
16. Niska, R. H.; Constant, A. P.; Witt, T.; Gregory, O. J. J. Vac. Sci. Tech. A 2000, 18,
16531658.
17. Huntz, A. M.; Andrieux, M.; Vahlas, C.; Sovar, M. M.; Samelor, D.; Gleizes, A. N.
J. Electrochem. Soc. 2007, 154, P63P68.
18. Gleizes, A. N.; Vahlas, C.; Sovar, M. M.; Samelor, D.; Lafont, M. C. Chem. Vap.
Deposition 2007, 13, 2329.
19. Kuo, D. H.; Cheung, B. Y.; Wu, R. J. Thin Solid Films 2001, 398, 3540.
20. Jing, C. B.; Xu, X. G.; Hou, J. X. J. SolGel Sci. Tech. 2007, 43, 321327.
21. Klein, T. M.; Niu, D.; Epling, W. S.; Li, W.; Maher, D. M.; Hobbs, C. C.; Hegde, R. I.;
Baumvol, I. J. R.; Parsons, G. N. Appl. Phys. Lett. 1999, 75, 40014003.
22. Min, Y. S.; Cho, Y. J.; Hwang, C. S. Chem. Mater. 2005, 17, 626631.
23. Cho, W.; Sung, K.; An, K. S.; Lee, S. S.; Chung, T. M.; Kim, Y. J.Vac. Sci. Tech. A
2003, 21, 13661370.
24. An, K. S.; Cho, W. T.; Sung, K. H.; Lee, S. S.; Kim, Y. Bull. Kor. Chem. Soc. 2003,
24, 16591663.
25. Guidi, F.; Moretti, G.; Carta, G.; Natali, M.; Rossetto, G.; Pierino, Z.; Salmaso, G.;
Rigato, V. Electrochim. Acta 2005, 50, 46094614.
26. Schmidt, B. W.; Sweet, W. J.; Bierschenk, E. J.; Gren, C. K.; Hanusa, T. P.;
Rogers, B. R. J. Vac. Sci. Tech. A 2010, 28, 238243.
27. Uesugi, T.; Kachi, T.; Sugimoto, M.; Matsuyama, T.; Mizue, C.; Hashizume, T.
J. Appl. Phys. 2008, 104, 016103.
28. Ortiz, A.; Alonso, J. C.; Pankov, V.; Huanosta, A.; Andrade, E. Thin Solid Films
2000, 368, 7479.
29. Aguilar-Frutis, M.; Garcia, M.; Falcony, C.; Plesch, G.; Jimenez-Sandoval, S. Thin
Solid Films 2001, 389, 200206.
30. Singh, M. P.; Shivashankar, S. A. Surf. Coat. Tech. 2002, 161, 135143.
31. Pradhan, S. K.; Reucroft, P. J.; Ko, Y. K. Surf. Coat. Tech. 2004, 176, 382384.
32. Pflitsch, C.; Muhsin, A.; Bergmann, U.; Atakan, B. Surf. Coat. Tech. 2006, 201,
7381.
33. Peng, Q.; Hojo, D.; Park, K. J.; Parsons, G. N. Thin Solid Films 2008, 516,
49975003.
34. Kadokura, H.; Ito, A.; Kimura, T.; Goto, T. Surf. Coat. Tech. 2010, 204,
23022306.
35. You, Y.; Ito, A.; Tu, R.; Goto, T. Appl. Surf. Sci. 2010, 256, 39063911.
36. Yamaguchi, N.; Tadanaga, K.; Matsuda, A.; Minami, T.; Tatsumisago, M. Surf.
Coat. Tech. 2006, 201, 36533657.
37. Ogita, Y. I.; Iehara, S.; Tomita, T. Thin Solid Films 2003, 430, 161164.
38. Ogita, Y. I.; Tomita, T. Thin Solid Films 2006, 501, 3942.
39. Szymanski, S. F.; Seman, M. T.; Wolden, C. A. Surf. Coat. Tech. 2007, 201,
89918997.
40. Duan, X. F.; Tran, N. H.; Roberts, N. K.; Lamb, R. N. Thin Solid Films 2009, 517,
67266730.
41. Duan, X. F.; Tran, N. H.; Roberts, N. K.; Lamb, R. N. Thin Solid Films 2010, 518,
42904293.
42. Fleischer, M.; Meixner, H. Sens. Actuators B 1993, 13, 259263.
43. Liu, Z.; Yamazaki, T.; Shen, Y.; Kikuta, T.; Nakatani, N.; Li, Y. Sens. Actuators B
2008, 129, 666670.
44. Fleischer, M.; Meixner, H. Sens. Actuators B 1991, 5, 115119.
45. Fleischer, M.; Meixner, H. Sens. Actuators B 1991, 4, 437441.
46. Basharat, S.; Betchley, W.; Carmalt, C. J.; Barnett, S.; Tocher, D. A.; Davies, H. O.
Organometallics 2007, 26, 403407.
47. Kokubun, Y.; Miura, K.; Endo, F.; Nakagomi, S. Appl. Phys. Lett. 2007, 90, 031912.
48. Mishra, S.; Daniele, S.; Petit, S.; Jeanneau, E.; Rolland, M. Dalton Trans. 2009,
25692577.
49. Bloor, L.; Carmalt, C. J.; Pugh, D. Coord. Chem. Rev. 2011, 255, 12931318.
50. Pugh, D.; Bloor, L. G.; Sathasivam, S.; Parkin, I. P.; Carmalt, C. J. Eur. J. Inorg.
Chem. 2011, 19531960.
51. Ballarin, B.; Battiston, G. A.; Benetollo, F.; Gerbasi, R.; Porchia, M.; Favretto, D.;
Traldi, P. Inorg. Chim. Acta 1994, 217, 7178.
52. Ortiz, A.; Alonso, J. C.; Andrade, E.; Urbiola, C. J. Electrochem. Soc. 2001, 148,
F26F29.
53. Carmalt, C. J.; King, S. J. Coord. Chem. Rev. 2006, 250, 682709.
54. Valet, M.; Hoffman, D. M. Chem. Mater. 2001, 13, 21352143.
55. Basharat, S.; Carmalt, C. J.; King, S. J.; Peters, E. S.; Tocher, D. A. Dalton Trans.
2004, 34753480.
56. Chi, Y.; Chou, T.-Y.; Wang, Y.-J.; Huang, S.-F.; Carty, A. J.; Scoles, L.;
Udachin, K. A.; Peng, S.-M.; Lee, G.-H. Organometallics 2004, 23, 95103.
57. Basharat, S.; Knapp, C. E.; Carmalt, C. J.; Barnett, S. A.; Tocher, D. A.
New J. Chem. 2008, 32, 15131518.
Precursor Chemistry Main Group Metal Oxides 1049
58. Basharat, S.; Carmalt, C. J.; Binions, R.; Palgrave, R.; Parkin, I. P. Dalton Trans.
2008, 591595.
59. Pugh, D.; Bloor, L.; Parkin, I. P.; Carmalt, C. J. Chem. Eur. J. 2012, 18,
60796087.
60. Lee, H.; Kim, K.; Woo, J.-J.; Jun, D.-J.; Park, Y.; Kim, Y. ECS Trans. 2009, 25,
587592.
61. Liu, G. X.; Shan, F. K.; Park, J. J.; Lee, W. J.; Lee, G. H.; Kim, I. S.; Shin, B. C.;
Yoon, S. G. J. Electroceram. 2006, 17, 145149.
62. Shan, F. K.; Liu, G. X.; Lee, W. J.; Lee, G. H.; Kim, I. S.; Shin, B. C. Integr.
Ferroelectr. 2006, 80, 197206.
63. Liu, G. X.; Shan, F. K.; Lee, W. J.; Shin, B. C.; Kim, S. C.; Kim, H. S.; Cho, C. R.
Integr. Ferroelectr. 2010, 94, 1120.
64. Dezelah, C. L.; Niinisto, J.; Arstila, K.; Niinisto, L.; Winter, C. H. Chem. Mater.
2006, 18, 471475.
65. Maruyama, T.; Fukui, K. J. J. Appl. Phys. 1991, 70, 38483851.
66. Bradley, D. C.; Chudzynska, H.; Frigo, D. M.; Hursthouse, M. B.; Mazid, M. A. J.
J. Chem. Soc., Chem. Commun. 1988, 12581259.
67. Suh, S.; Hoffman, D. M. J. Am. Chem. Soc. 2000, 122, 93969404.
68. Basharat, S.; Carmalt, C. J.; Barnett, S. A.; Tocher, D. A.; Davies, H. O. Inorg.
Chem. 2007, 46, 94739480.
69. Veith, M.; Hill, S.; Huch, V. Eur. J. Inorg. Chem. 1999, 13431350.
70. Elam, J. W.; Martinson, A. B. F.; Pellin, M. J.; Hupp, J. T. Chem. Mater. 2006, 18,
35713578.
71. Nilsen, O.; Balasundaraprabhu, R.; Monakhov, E. V.; Muthukumarasamy, N.;
Fjellvag, H.; Svensson, B. G. Thin Solid Films 2009, 517, 63206322.
72. Berry, A. D.; Holm, R. T.; Mowery, R. L.; Turner, R. H.; Fatemi, M. Chem. Mater.
1991, 3, 7277.
73. Switzer, J. A. J. Electrochem. Soc. 1986, 133, 722728.
74. Speller, S. C. Mater. Sci. Technol. 2003, 19, 269282.
75. Zhang, K.; Boyd, E. P.; Kwak, B. S.; Wright, A. C.; Erbil, A. Appl. Phys. Lett. 1989,
55, 12581260.
76. Nielson, A. J.; Rickard, C. E. F.; Smith, J. M. Inorg. Synth. 1986, 24, 9799.
77. Zanella, P.; Rossetto, G.; Sitran, S.; Ferro, D. Eur. J. Solid State Inorg. Chem.
1992, 29, 181194.
78. Daolio, S.; Ajo, D.; Rossetto, G.; Aschieri, C.; Armelao, L. Org. Mass. Spectrom.
1992, 27, 11711175.
79. Malandrino, G.; Borzi, A. M.; Castelli, F.; Fragala, I. L.; Dastru, W.; Gobetto, R.;
Rossi, P.; Dapporto, P. Dalton Trans. 2003, 00, 369374.
80. Mohammadi, M.; Morsali, A. Mater. Lett. 2009, 63, 23492351.
81. Granqvist, C. G. Solar Energy Mater. Solar Cells 2007, 91, 15291598.
82. Liu, H.; Izyumskaya, N.; Ozgur, U.; Morkoc, H. Superlattice. Microst. 2010, 48,
458484.
83. Exarhos, G. J.; Zhou, X.-D. Thin Solid Films 2007, 515, 70257052.
84. Minami, T.; Miyata, T. Thin Solid Films 2008, 517, 14741477.
85. Hoel, C. A.; Mason, T. O.; Gaillard, J.-F.; Poeppelmeier, K. Chem. Mater. 2010,
22, 35693579.
86. Liu, J.; Hains, A. W.; Wang, L.; Marks, T. J. Thin Solid Films 2010, 518,
36943699.
87. Ni, J.; Wang, L.; Yang, Y.; Yan, H.; Jin, S.; Marks, T. J. Inorg. Chem. 2005, 44,
60716076.
88. Milde, F.; Brucker, Deus, C.; Kammer, M.; Neumann, H. Proc. 4th Int. Conf. on
Coatings on Glass, ICCG, 2002; pp 419424.
89. Korosi, L.; Papp, S.; Dekany, I. Thin Solid Films 2011, 519, 31133118.
90. Liu, J.; Wu, D.; Zeng, S. J. Mater. Process. Technol. 2009, 209, 39433948.
91. Ait Aouaj, M.; Diaz, R.; Belayachi, A.; Rueda, F.; Abd-Lefdil, M. Mater. Res. Bull
2009, 44, 14581461.
92. Chandrasekhar, R.; Choy, K. L. Thin Solid Films 2001, 398399, 5964.
93. Yuan, J.; Ren, L. Micro Nano Lett. 2010, 5, 5357.
94. Mottern, M. L.; Tyholdt, F.; Ulyashin, A.; van Helvoort, A. T. J.; Verweij, H.;
Bredesen, R. Thin Solid Films 2007, 515, 39183926.
95. Aksu, Y.; Frasca, S.; Wollenberger, U.; Driess, M.; Thomas, A. Chem. Mater.
2011, 23, 17981804.
96. Minami, T. Thin Solid Films 2008, 516, 58225828.
97. Wang, A. C.; Edleman, N. L.; Babcock, J. R.; Marks, T. J.; Lane, M. A.;
Brazis, P. R.; Kannewurf, C. R. J. Mater. Res. 2002, 17, 3155.
98. Knapp, C. E.; Hyett, G.; Parkin, I. P.; Carmalt, C. J. Chem. Mater. 2011, 23,
17191726.
99. Revesz, A. G.; Hughes, H. L. J. Non-Crystallin. Sol. 2003, 328, 4863.
100. Barranco, A.; Yubero, F.; Cotrino, J.; Espinos, J. P.; Benteza, J.; Rojasa, T. C.;
Allain, J.; Girardeau, T.; Riviere, J. P.; Gonzalez-Elipe, A. R. Thin Solid Films
2001, 396, 915.
101. Murarka, S. P. Encyclopedia of Materials: Science and Technology, 2008;
pp 114.
1050 Precursor Chemistry Main Group Metal Oxides
102. Foggiato, J. In Handbook of Thin Film Deposition Processes and Techniques,
Seshan K., Ed.; 2nd ed.; William Andrew Publishing: Norwich, NY, 2001;
pp 111150.
103. Eranen, S. In Handbook of Silicon Based MEMS Materials and Technologies,
Lindroos, V., Tilli, M., Lehto, A., Motooka, T., Eds.; 1st ed.; William Andrew
Publishing Place: Oxford, UK, 2010; pp 137148.
104. Lange, T.; Njoroge, W.; Weis, H.; Beckers, M.; Wuttig, M. Thin Solid Films 2000;
365, 8289.
105. da Silva, D. M.; Kassab, L. R. P.; Martinelli, J. R.; de Araujo, C. B. J. Non-Cryst.
Sol. 2010, 356, 26022605.
106. Prasad, P. N. Nanophotonics. Wiley: New York, 2004.
107. Novotny, L.; Hecht, B. Principles of Nano-Optics. Cambridge University Press:
Cambridge, 2007.
108. Mei, Y. F.; Siu, G. G.; Huang, X. H.; Cheah, K. W.; Dong, Z. G.; Fang, L.;
Sheng, M. R.; Wu, X. L.; Bao, X. M. Phys. Lett. A 2004, 331, 248251.
109. Lorenz, H.; Zhao, Q.; Turner, S.; Lebedev, O. I.; Van Tendeloo, G.; Klotzer, B.;
Rameshan, C.; Penner, S. Mater. Chem. Phys. 2010, 122, 623629.
110. Serezhkina, S. V.; Tyavlovskaya, E. A.; Shevchenko, G. P.; Rakhmanov, S. K.
J. Non-Cryst. Sol. 2005, 351, 3540.
111. Maezono, Y.; Yangita, K.; Nishi, J.-I.; Miyano, A.; Yokoyani, A.; Kurosawa, K.;
Hishinuma, N.; Matsuno, H. J. Phys. IV France 2001, 11, 811.
112. El-Hagary, M.; Emam-Ismail, M.; Shaaban, E. R.; Shaltout, I. J. Alloy. Comp.
2009, 485, 519523.
113. Khan, M. A.; Hogan, T. P.; Shanker, B. J. Raman Spectrosc. 2008, 39, 893900.
114. Phani, A. R.; Di Claudio, D.; Passacantando, M.; Santucci, S. J. Non-Cryst. Sol.
2007, 353, 692696.
115. Jang, J. H.; Koo, J.; Bae, B.-S. J. Am. Ceram. Soc. 2000, 83, 13561360.
116. Rana, M. M.; Butler, D. P. Thin Solid Films 2006, 514, 355360.
117. Gomez, L. A.; dos Santos, F. E. P.; Gomes, A. S. L.; de Araujo, C. B.;
Kassab, L. R. P.; Hora, W. C. Appl. Phys. Lett. 2008, 92, 141916.
118. Naranjo, L. P.; de Araujo, C. B.; Malta, O. L.; Santa Cruz, P. A.; Kassab, L. R. P.
Appl. Phys. Lett. 2005, 87, 241914.
119. Korotkov, R. Y.; Gupta, R.; Ricou, P.; Smith, R.; Silverman, G. Thin Solid Films
2008, 516, 47204727.
120. Chopra, K. L.; Major, S.; Pandya, D. K. Thin Solid Films 1983, 102, 146.
121. Korotkov, R. Y.; Ricou, P.; Farran, A. J. E. Thin Solid Films 2006, 502, 7987.
122. Cachet, H.; Gamard, A.; Campet, G.; Jousseaume, B.; Toupance, T. Thin Solid
Films 2001, 388, 4149.
123. Egdell, R. G.; Flavell, W. R.; Tavener, P. J. SolGel Sci. Techn. 1984, 51,
345354.
124. Bae, J. W.; Kim, J. S.; Yeom, G. Y. Nucl. Instrum. Meth. B 2001, 178, 311314.
125. Choi, G.; Ryu, H.; Seo, Y.; Lee, W.; Hong, K.; Shin, D.; Park, J.; Akbar, S. A.
J. Korean Phys. Soc. 2003, 43, L967.
126. Hoefer, U.; Frank, J.; Fleischer, M. Sens. Actuators B 2001, 78, 6.
127. Chin, Y. L.; Chou, J. C.; Sun, T. P.; Liao, H. K.; Chung, W. Y.; Hsiung, S. Y. Sens.
Actuators B 2001, 75, 36.
128. Tennakone, K.; Bandara, J.; Bandaranayake, P. K. M.; Kumara, G. R. A.; Konno, A.
Jpn. J. Appl. Phys. 2001, 40, L732.
129. Imamori, K.; Masuda, A.; Matsumura, H. Thin Solid Films 2001, 395, 147.
130. Veluchamy, P.; Tsuji, M.; Nishio, T.; Aramoto, T.; Higuchi, H.; Kumazawa, S.;
Shibutani, S.; Nakajima, J.; Arita, T.; Ohyama, H.; Hanafusa, A.; Hibino, T.;
Omura, K. Sol. Energ. Mat. Sol. C. 2001, 67, 179.
131. Mwamburi, M.; Wackelgard, E.; Roos, A. Thin Solid Films 2000, 374, 1.
132. AkIn, A. N.; Kilaz, G.; Isli, A. I.; Onsan, Z. I. Chem. Eng. Sci. 2001, 56, 881888.
133. Wang, X.; Xie, Y.-C. Appl. Catal. B: Environ. 2001, 35, 8594.
134. Amjoud, M. B.; Maury, F.; Soukane, S.; Duverneuil, P. Surf. Coat.Technol. 1998,
100101, 169172.
135. Kim, Y. I.; Lee, W. H.; Moon, H. S.; Ji, K. S.; Seong, S. H.; Park, J. W. J. Power.
Sources 2001, 101, 253258.
136. Vergohl, M.; Malkomes, N.; Matthee, T.; Brauer, G.; Richter, U.; Nickol, F. W.;
Bruch, J. Thin Solid Films 2001, 392, 258264.
137. Park, G.-S.; Yang, G.-M. Thin Solid Films 2000, 365, 711.
138. Hao, X.; Ma, J.; Zhang, D.; Xu, X.; Yang, Y.; Ma, H.; Ai, S. Appl. Phys. A: Mater
2002, 75, 397399.
139. Krause, U.; List, M.; Fuchs, H. Thin Solid Films 2001, 392, 196200.
140. Martel, A.; Caballero-Briones, F.; Bartolo-Perez, P.; Iribarren, A.; Castro-
RodrIguez, R.; Zapata-Navarro, A.; Pena, J. L. Surf. Coat.Technol. 2001, 148,
103109.
141. Choe, Y.-S. Sensor. Actuat. B: Chem. 2001, 77, 200208.
142. Shamala, K.; Murthy, L.; Narasimha Rao, K. B. Mater. Sci. 2004, 27, 295301.
143. Radecka, M.; Przewoznik, J.; Zakrzewska, K. Thin Solid Films 2001, 391,
247254.
144. Lawrence, J.; Lubrani, P.; Li, L. Surf. Coat.Technol. 2001, 137, 235240.
145. El Khakani, M. A.; Dolbec, R.; Serventi, A. M.; Horrillo, M. C.; Trudeau, M.;
Saint-Jacques, R. G.; Rickerby, D. G.; Sayago, I. Sens. Actuators B 2001, 77,
383388.
146. Patil, P. S.; Kawar, R. K.; Seth, T.; Amalnerkar, D. P.; Chigare, P. S. Ceram. Int.
2003, 29, 725734.
147. Jadsadapattarakul, D.; Euvananont, C.; Thanachayanont, C.; Nukeaw, J.;
Sooknoi, T. Ceram. Int. 2008, 34, 10511054.
148. Lee, S. C.; Lee, J. H.; Kim, Y. H. J. Korean Electrochem. Soc. 2000, 3, 15.
149. Acciarri, M.; Canevali, C.; Mari, C. M.; Mattoni, M.; Ruffo, R.; Scotti, R.;
Morazzoni, F.; Barreca, D.; Armelao, L.; Tondello, E.; Bontempi, E.; Depero, L. E.
Chem. Mater. 2003, 15, 26462650.
150. Lee, W.; Choi, Y.; Hong, K.; Kim, N. H.; Park, Y.; Park, J. J. Korean Phys. Soc.
2005, 46, L756.
151. de Graaf, A.; van Deelen, J.; Poodt, P.; van Mol, T.; Spee, K.; Grob, F.; Kuypers, A.
Energy Procedia 2010, 2, 4148.
152. Lee, S. W.; Kim, Y. W.; Chen, H. Appl. Phys. Lett. 2001, 78, 350.
153. Jimenez, V. M.; Espinos, J. P.; Caballero, A.; Contreras, L.; Fernandez, A.;
Justo, A.; Gonzalez-Elipe, A. R. Thin Solid Films 1999, 353, 113123.
154. Ansari, S. G.; Ansari, Z. A.; Wahab, R.; Kim, Y.-S.; Khang, G.; Shin, H.-S. Biosens.
Bioelectron. 2008, 23, 18381842.
155. Lancok, J.; Santoni, A.; Penza, M.; Loreti, S.; Menicucci, I.; Minarini, C.;
Jelinek, M. Surf. Coat.Technol. 2005, 200, 10571060.
156. Tamura, S.; Ishida, T.; Magara, H.; Mihara, T.; Mochizuki, S.; Tatsuta, T. Appl.
Surf. Sci. 2001, 169170, 425427.
157. Tadanaga, K.; Fujii, T.; Matsuda, A.; Minami, T.; Tatsumisago, M. Ceram. Int.
2004, 30, 18151817.
158. Rizzato, A. P.; Santilli, C. V.; Pulcinelli, S. H.; Craievich, A. F. J. Appl. Crystallogr.
2003, 36, 736739.
159. Fu, Y.; Jin, Z.; Liu, Z.; Li, W. J. Eur. Ceram. Soc. 2007, 27, 22232228.
160. Luyo, C.; Fabregas, I.; Reyes, L.; Sols, J. L.; Rodrguez, J.; Estrada, W.;
Candal, R. J. Thin Solid Films 2007, 516, 2533.
161. Majumder, S. Mater. Sci. - Poland 2009, 27, 123.
162. Rizzato, A. P.; Santilli, C. V.; Pulcinelli, S. H. J. SolGel Sci. Techn. 2000, 19,
811816.
163. Guo, L.; Liu, X.; He, Z.; Gao, J.; Yang, J.; Guo, T. J. SolGel Sci. Techn. 2008, 47,
239242.
164. Lee, S.-C.; Lee, J.-H.; Oh, T.-S.; Kim, Y.-H. Sol. Energ. Mat. Sol. C 2003, 75,
481487.
165. Brezesinski, T.; Fischer, A.; Iimura, K. I.; Sanchez, C.; Grosso, D.; Antonietti, M.;
Smarsly, B. M. Adv. Funct. Mater. 2006, 16, 14331440.
166. Tadanaga, K.; Owan, T.; Morinaga, J.; Urbanek, S.; Minami, T. J. SolGel Sci.
Techn. 2000, 19, 791794.
167. Dal Santos, M. A.; Antunes, A. C.; Ribeiro, C.; Borges, C. P. F.; Antunes, S. R. M.;
Zara, A. J.; Pianaro, S. A. Mater. Lett. 2003, 57, 43784381.
168. Kim, H.; Park, H.-H.; Chang, H. J.; Jeon, H.; Park, H.-H. Thin Solid Films 2008,
517, 10721076.
169. Gu, Z.; Liang, P.; Liu, X.; Zhang, W.; Le, Y. J. SolGel Sci. Techn. 2000, 18,
159166.
170. Amlouk, A.; Boubaker, K.; Amlouk, M. J. Alloy. Compd. 2010, 490, 602604.
171. Elangovan, E.; Singha, M. P.; Dharmaprakash, M. S.; Ramamurthi, K. J.
J. Optoelectron. Adv. M. 2004, 6, 197.
172. Lopes, A.; Nunes, P.; Vilarinho, P. M.; Monteiro, R. C. C.; Fortunato, E.;
Martins, R. Key Eng. Mat. 2002, 230232, 357.
173. Murakami, K.; Nakajima, K.; Kaneko, S. Thin Solid Films 2007, 515, 86328636.
174. Okuya, M.; Kaneko, S.; Hiroshima, K.; Yagi, I.; Murakami, K. J. Eur. Ceram. Soc.
2001, 21, 20992102.
175. Aguilar-Leyva, J.; Maldonado, A.; Olvera, M.de la L. Mater. Charact. 2007, 58,
740744.
176. Sundqvist, J.; Lu, J.; Ottosson, M.; Harsta, A. Thin Solid Films 2006, 514,
6368.
177. Lu, J.; Sundqvist, J.; Ottosson, M.; Tarre, A.; Rosental, A.; Aarik, J.; Harsta, A.
J. Cryst. Growth 2004, 260, 191200.
178. Rosental, A.; Tarre, A.; Gerst, A.; Uustare, T.; Sammelselg, V. Sensor. Actuat. B:
Chem. 2001, 77, 297300.
179. Elam, J. W.; Baker, D. A.; Hryn, A. J.; Martinson, A. B. F.; Pellin, M. J.; Hupp, J. T.
J. Vac. Sci. Technol. A 2008, 26, 244252.
180. Choi, G.; Satyanarayana, L.; Park, J. Appl. Surf. Sci. 2006, 252, 78787883.
181. van Mol, A. M. B.; Driessen, J. P. A. M.; Linden, J. L.; de Croon, M. H. J. M.;
Spee, C. I. M. A.; Schouten, J. C. Chem. Vap. Deposition 2001, 7, 101104.
182. Koutsogianni, A.; Tsamakis, D. J. Phys. IV, France 2001, 11, Pr3Pr377.
183. Karsloglu, R.; Alp, A.; Akbulut, H. J. J. Coat. Technol. Res. 2010, 7, 503510.
184. Tucic, A.; Marinkovic, Z. V.; Mancic, L.; Cilense, M.; Milosevic, O. J. Mater.
Process. Tech. 2003, 143144, 4145.

185. Ansari, S. G.; Dar, M. A.; Ansari, Z. A.; Seo, H. K.; Kim, Y.-S.; Al-Hajry, A.;
Shin, H.-S. Sci. Adv. Mat. 2009, 1, 254261.
186. Papadopoulos, N. D.; Tsakiridis, P. E.; Hristoforou, E. J. Optoelectron. Adv. M.
2005, 7, 2693.
187. Ansari, S. G.; Dar, M. A.; Dhage, M. S.; Kim, Y. S.; Ansari, Z. A.; Al-Hajry, A.;
Shin, H. S. Rev. Sci. Instrum. 2009, 80, 045112.
188. Sundqvist, J.; Ottosson, M.; Harsta, A. Chem. Vapor Deposition 2004, 10, 7782.
189. Supothina, S. Sens. Actuators B 2003, 93, 526530.
190. Huang, H.; Tan, O. K.; Lee, Y. C.; Tse, M. S. J. Cryst. Growth 2006, 288, 7074.
191. Houlton, D. J.; Jones, A. C.; Haycock, P. W.; Williams, E. W.; Bull, J.;
Critchlow, G. W. Chem. Vapor Deposition 1995, 1, 2628.
192. Brown, J. R.; Cheney, M. T.; Haycock, P. W.; Houlton, D. J.; Jones, A. C.;
Williams, E. W. J. Electrochem. Soc. 1997, 144, 295.
193. Supothina, S.; de Guire, M. R.; Heuer, A. H. J. Am. Ceram. Soc. 2003, 86,
20742081.
194. Lee, D. H.; Chang, Y. J.; Stickle, W.; Chang, C. H. Electrochem. Solid St. 2007,
10, K51K54.
195. Mukhopadhyay, A. K.; Mitra, P.; Chatterjee, A. P.; Maiti, H. S. Ceram. Int. 2000,
26, 123132.
196. Leon, N.; Melendez, A.; Figueroa, G.; Ramos, I.; Pinto, N. J. P. SPIE Proc. Ser.
2007, 6591, 65910G.
197. Vequizo, J. J. M.; Wang, J.; Ichimura, M. Jpn. J. Appl. Phys. 2010, 49, 125502.
198. Yang, J.; Bai, S. L.; Luo, R. X.; Chen, A. F.; Lin, Y.; Zhang, J. B. Chinese Chem.
Lett. 2010, 21, 15051508.
199. Chi, K.; Lin, C.; Lu, U.; Liao, J. J. Chin. Chem. Soc. 2000, 47, 425.
200. Talaty, N. N.; Beck, K.; Citeau, H.; Kirschbaum, K.; Giolando, D. M. Z. Anorg. Allg.
Chem. 2009, 635, 5363.
201. Soderquist, R.; Dickens, B. J. Phys. Chem. Solids 1967, 28, 823826.
202. Trinquier, G.; Hoffmann, R. J. Phys. Chem. 1984, 88, 6696.
203. Tuncheva, V. D.; Baleva, M. I. Proc. NATO ASI Ser. 1995, 3, 103.
204. Baleva, M. I.; Tuncheva, V. D. J. Mater. Sci. Lett. 1994, 13, 3.
205. Polarco, A. M.; Palmas, S.; Renoldi, F.; Mascia, M. J. Appl. Electrochem. 1999,
29, 147.
206. Johnson, D. C.; Feng, J.; Houk, L. L. Electrochim. Acta 2000, 46, 323.
207. Stucki, S.; Theis, G.; Kolz, R.; Devantay, H.; Christen, H. J. J. Electrochem. Soc.
1985, 132, 367.
208. Velayuthan, D.; Noel, M. Tantala 1992, 39, 481.
209. Beck, F. Electrochemistry of Lead. Academic Press: New York, 1979.
210. Grugeon-Dewaele, S.; Laruelle, L.; Torcheux, L.; Tarascon, J.-M.; Delahaye-
Vidal, A. J. Electrochem. Soc. 1998, 145, 3358.
211. Kuhn, A. T. The Electrochemistry of Lead. Academic Press: London, 1979.
212. Light, T. B.; Eldridge, J. M.; Matthews, J. W.; Greiner, J. H. J. Appl. Phys. 1975,
46, 1489.
213. Kramarenko, N. L.; Miloslavskii, V. K.; Naboikin, Y. V. Opt. Spektrosk. 1968, 24, 971.
214. Hubert-Pfalzgraf, L. Appl. Organometal. Chem. 1992, 6, 627.
215. Harjuoja, J.; Putkonen, M.; Niinisto, L. Thin Solid Films 2006, 497, 7782.
216. Krisyuk, V. V.; Turgambaeva, A. E.; Igumenov, I. K. Chem. Vapor Deposition
1998, 4, 4346.
217. Lee, H. J.; Kim, G. H.; Lee, K.; Hwang, C. S. Electrochem. Solid St. 2007, 10,
G89G92.
218. Tokita, K.; Okada, F. Ferroelectrics 1998, 21, 12.
219. Hendricks, W. C.; Desu, S. B.; Tsai, C. Y. Mater. Res. Sym. Proc. 1993, 335, 215.
220. Madsen, L. D.; Weaver, L. J. Am. Ceram. Soc. 1998, 81, 988996.
221. Dai, G.; Xu, J. J. Mater. Sci. Lett. 1998, 17, 969.
222. Lee, W. G.; Woo, S. I. J. Mater. Chem. 1997, 32, 815.
223. Watanabe, A.; Tsuchiya, T.; Imai, Y. Thin Solid Films 2002, 419, 7681.
224. Tokita, K.; Okada, F. J. Appl. Phys. 1996, 80, 7073.
225. Zhao, J. S.; Sim, J. S.; Lee, H. J.; Park, D.-Y.; Hwang, C. S. J. Electrochem. Soc.
2005, 152, C277C282.
226. Malik, M. A.; OBrien, P.; Motevalli, M.; Jones, A. C.; Leedham, T. Polyhedron
1999, 18, 16411646.
227. Turgambaeva, A. E. Ph.D. Thesis, Institute of Inorganic Chemistry, Novosibirsk,
1996.
228. Hedinger, R.; Kradolfer, T.; Hegetschweiler, K.; Worle, M.; Dahmen, K.-H. Chem.
Vapor Depos. 1999, 5, 2935.
229. Cachet, H.; Tsuyoshi, N.; Henri, G. Fluorinated Materials for Energy Conversion.
Elsevier Science: Amsterdam, 2005; pp 513534.
230. Kim, K. H.; Lee, S. W.; Shin, D. W.; Park, C. G. J. Am. Ceram. Soc. 1994, 77,
915921.
231. Elangovan, E.; Ramamurthi, K. Appl. Surf. Sci. 2005, 249, 183196.
232. Buntem, R.; Kraisingdecha, P.; Sadee, W. Adv. Mat. Res. 2008, 5557, 513.
233. Kololuoma, T.; Rantala, J. T. Electronik. Lett. 2000, 38, 172.
234. Rodrigues, E. C. P. E.; Olivi, P. J. Phys. Chem. Solids 2003, 64, 11051112.
Precursor Chemistry Main Group Metal Oxides 1051
235. Elangovan, E.; Ramamurthi, K. Cryst. Res. Technol. 2003, 38, 779784.
236. Liu, T. J.; Jin, Z. G.; Feng, L. R.; Wang, T. Appl. Surf. Sci. 2008, 254, 65476553.
237. Bernardi, M. I. B.; Barrado, C. M.; Soledade, L. E. B.; Leite, E. R.; Longo, E. J.
J. Mater. Sci.: Mater. Electronics 2002, 13, 403.
238. Wang, L.; Bi, W. Y.; Zhang, X. D.; Yu, X. G.; Bao, H. Q.; Zhao, H. F. Key Eng. Mat.
2007, 336338, 1799.
239. Epifani, M.; Alvisi, M.; Mirenghi, L.; Leo, G.; Siciliano, P.; Vasanelli, L. J. Am.
Ceram. Soc. 2001, 84, 4854.
240. Tsuchiya, T.; Kato, F.; Nakajima, T.; Igarashi, K.; Kumagai, T. Appl. Surf. Sci.
2009, 255, 98089812.
241. Kololuoma, T. P Soc. Photo-Opt. Ins. 2002, 4804, 106.
242. Ramgir, N. S.; Hwang, Y. K.; Jhung, S. H.; Kim, H.-K.; Hwang, J.-S.; Mulla, I. S.;
Chang, J.-S. Appl. Surf. Sci. 2006, 252, 42984305.
243. Cheng, H.; Lu, Z. Solid State Sci. 2008, 10, 10421048.
244. Pinheiro, M. A. L.; Pineiz, T. F.; de Morais, E. A.; Scalvi, L. V. A.; Saeki, M. J.;
Cavalheiro, A. A. Thin Solid Films 2008, 517, 976981.
245. Mihaiu, S.; Marta, L.; Zaharescu, M. J. Eur. Ceram. Soc. 2007, 27, 551555.
246. Korotcenkov, G.; Brinzari, V.; Boris, I. J. Mater. Sci. 2008, 43, 27612770.
247. Ghimbeu, C. M.; van Landschoot, R. C.; Schoonman, J.; Lumbreras, M. J. Eur.
Ceram. Soc. 2007, 27, 207213.
248. Ravaro, L. P.; Morais, E. A.; Scalvi, L. V. A.; Li, M. S. Ceramica 2007, 53, 187.
249. Joseph, D. P.; Renugambal, P.; Saravanan, M.; Raja, S. P.; Venkateswaran, C.
Thin Solid Films 2009, 517, 61296136.
250. Buono-Core, G. E.; Cabello, G. A.; Espinoza, H.; Klahn, A. H.; Tejos, M.; Hill, R. H.
J. Chil. Chem. Soc. 2006, 51, 950.
251. Hubner, M.; Koziej, D.; Bauer, M.; Barsan, N.; Kvashnina, K.; Rossell, M. D.;
Weimar, U.; Grunwaldt, J.-D. Angew. Chem. Int. Ed. 2011, 50, 28412844.
252. Kida, T.; Kurachi, H.; Yuasa, M.; Shimanoe, K.; Yamazoe, N. Adv. Mat. Res. Vols.
2008, 4750, 1506.
253. Gaidi, M.; Chenevier, B.; Labeau, M.; Hazemann, J. L. Sens. Actuators B 2006,
120, 313315.
254. Morazzoni, F.; Canevali, C.; Chiodini, N.; Mari, C.; Ruffo, R.; Scotti, R.; Armelao, L.;
Tondello, E.; Depero, L. E.; Bontempi, E. Chem. Mater. 2001, 13, 43554361.
255. Morazzoni, F.; Canevali, C.; Chiodini, N.; Mari, C.; Ruffo, R.; Scotti, R.;
Armelao, L.; Tondello, E.; Depero, L. E.; Bontempi, E. Mater. Sci. Eng. C 2001,
15, 167169.
256. Bauer, S.; Bauer, F. In Piezoelectricity: Evolution and Future of a Technology;
Heywang, W., Lubitz, K., Wersing, W., Eds.; Springer Series in Materials Science,
Vol. 114, Springer: New York, 2008; pp 89130.
257. Leng, W.; Yang, C.; Ji, H.; Zhang, J.; Tang, J.; Chen, H. J. Mater. Sci: Mater.
Electron. 2006, 17, 1041.
258. Kanno, I.; Kamada, T.; Hayashi, S.; Kitagawa, M.; Hirao, T. Jpn. J. Appl. Phys.
1993, 32, L950.
259. Wang, D. G.; Chen, C. Z.; Ma, J.; Lui, T. H. Appl. Surf. Sci. 2008, 255, 1637.
260. Tanaka, H.; Tabata, H.; Kawai, T.; Kawai, S. Jpn. J. Appl. Phys. 1994, 33, L451.
261. Foster, C. M.; Li, Z.; Buckett, M.; Miller, D.; Baldo, P. M.; Rehn, L. E.; Bai, G. R.;
Guo, D.; You, H.; Merkle, K. L. J. Appl. Phys. 1995, 78, 2607.
262. Sakashita, Y.; Ono, T.; Segawa, H.; Tominaga, K.; Okada, M. J. Appl. Phys. 1991,
69, 83528359.
263. Badawy, W. A. Thin Solid Films 1990, 186, 5972.
264. Binions, R.; Carmalt, C. J.; Parkin, I. P. Polyhedron 2006, 25, 30323038.
265. Horley, G. A.; Mahon, M. F.; Molloy, K. C.; Venter, M. M.; Haycock, P. W.;
Myers, C. P. Inorg. Chem. 2002, 41, 16521657.
266. Horley, G. A.; Mahon, M. F.; Molloy, K. C.; Haycock, P. W.; Myers, C. P. Inorg.
Chem. 2002, 41, 50525058.
267. Haycock, P. W.; Horley, G. A.; Molloy, K. C.; Myers, C. P.; Rushworth, S. A.;
Smith, L. M. Chem. Vap. Deposition 2001, 7, 191193.
268. Myers, C. P.; Haycock, P. W.; Pichot, M.; Horley, G. A.; Molloy, K. C.;
Rushworth, S. A.; Smith, L. M. Chem. Vap. Deposition 2004, 10, 3544.
269. Horley, G. A.; Mahon, M. F.; Mazhar, M.; Molloy, K. C.; Haycock, P. W.;
Myers, C. P. Dalton Trans. 2002, 44164421.
270. Sharutin, V. V.; Pakusina, A. P.; Egorova, I. V.; Sharutina, O. K.; Fukin, G. K.
Russian J. Coord. Chem. 2008, 34, 259263.
271. Kalkofen, B.; Matichyn, S.; Burte, E. Electrochem. Soc. Trans. 2009, 19(1),
105114.
272. Yang, R. B.; Bachmann, J.; Reiche, M.; Gerlach, J. W.; Gosele, U.; Nielsch, K.
Chem. Mater. 2009, 21, 25862588.
273. Yadav, N.; Nagar, M.; Bohra, R. J. SolGel Sci. Techn. 2010, 54, 119128.
274. Michel, C.; Hervieu, M.; Borel, M. M.; Grandin, A.; Deslandes, F.; Provost, J.;
Raveau, B. Z. Phys. B Condens. Matter 1987, 68, 421423.
275. Azad, A. M.; Larose, S.; Akbar, S. A. J. Mater. Sci. 1994, 29, 41354151.
276. Weidong, H.; Wei, Q.; Xiaohong, W.; Xianbo, D.; Long, C.; Zhaohua, J. Thin Solid
Films 2007, 515, 53625365.
1052 Precursor Chemistry Main Group Metal Oxides
277. Gurunathan, K. Int. J. Hydrogen Energy 2004, 29, 933940.
278. Ren, W.; Trolier-McKinstry, S.; Randall, C. A.; Shrout, T. R. J. Appl. Phys. 2001,
89, 767774.
279. Park, B. H.; Kang, B. S.; Bu, S. D.; Noh, T. W.; Lee, J.; Jo, W. Nature 1999, 401,
682684.
280. Wang, J.; Neaton, J. B.; Zheng, H.; Nagarajan, V.; Ogale, S. B.; Liu, B.;
Viehland, D.; Vaithyanathan, V.; Schlom, D. G.; Waghmare, U. V.; Spaldin, N. A.;
Rabe, K. M.; Wuttig, M.; Ramesh, R. Science 2003, 299, 17191722.
281. Sarala Devi, G. Sens. Actuators B 1999, 56, 98105.
282. Takeyama, T. J. Phys. Chem. Solids 2004, 65, 13491352.
283. Schuisky, M.; Harsta, A. Chem. Vap. Deposition 1996, 2, 235238.
284. Kim, H.; Myung, J.; Shim, S. Solid State Communications 2006, 137, 196198.
285. Bedoya, C.; Condorelli, G. G.; Anastasi, G.; Baeri, A.; Scerra, F.; Fragala, I. L.;
Lisoni, J. G.; Wouters, D. Chem. Mater. 2004, 16, 31763183.
286. Bedoya, C.; Condorelli, G. G.; Finocchiaro, S. T.; Di Mauro, A.; Fragala, I. L.;
Cattaneo, L.; Carella, S. Chem. Vap. Deposition 2005, 11, 261268.
287. Bandoli, G.; Barreca, D.; Brescacin, E.; Rizzi, G. A.; Tondello, E. Chem. Vap.
Deposition 1996, 2, 238242.
288. Reuge, N.; Dexpert-Ghys, J.; Caussat, B. Chem. Vap. Deposition 2010, 16,
123126.
289. Mehring, M. Coord. Chem. Rev. 2007, 251, 9741006.
290. Mehrotra, R. C.; Rai, A. K. Indian J. Chem. 1966, 4, 537.
291. Matchett, M.; Chiang, M.; Buhro, W. Inorg. Chem. 1990, 29, 358360.
292. Massiani, M.-C.; Papiernik, R.; Hubert-Pfalzgraf, L. G.; Daran, J.-C. Polyhedron
1991, 10, 437445.
293. Kou, X.; Wang, X.; Mendoza-Espinosa, D.; Zakharov, L. N.; Rheingold, A. L.;
Watson, W. H.; Brien, K.; Jayarathna, L. K.; Hanna, T. Inorg. Chem. 2009, 48,
1100211016.
294. Knispel, C.; Limberg, C.; Ziemer, B. Inorg. Chem. 2010, 49, 43134318.
295. Jones, C.; Burkart, M.; Whitmire, K. Angew. Chem. Int. Ed Engl. 1992, 31,
451452.
296. Whitmire, K. H.; Hoppe, S.; Sydora, O.; Jolas, J. L.; Jones, C. M. Inorg. Chem.
2000, 39, 8597.
297. Williams, P. A.; Jones, A. C.; Crosbie, M. J.; Wright, P. J.; Bickley, J. F.;
Steiner, A.; Davies, H. O.; Leedham, T. J.; Critchlow, G. W. Chem. Vap. Deposition
2001, 7, 205209.
298. Hermann, W. A.; Huber, N. W.; Anwander, R.; Priermeier, T. Eur. J. Inorg. Chem.
1993, 126, 11271130.
299. Moniz, S. J. A.; Blackman, C. S.; Carmalt, C. J.; Hyett, G. J. Mater. Chem. 2010,
20, 78817886.
300. Kang, S. Thin Solid Films 2004, 468, 7983.
301. Brooks, K. C.; Turnipseed, S. B.; Barkley, R. M.; Sievers, R. E.; Tulchinsky, J. V.;
Kaloyerosj, A. E. Chem. Mater. 1992, 912916.
302. Dikarev, E. V.; Zhang, H.; Li, B. J. Am. Chem. Soc. 2005, 127, 61566157.
303. Stavila, V.; Dikarev, E. V. J. Organomet. Chem. 2009, 694, 29562964.
304. Pisarevsky, A. In AIP Conference Proceedings; Kao, Y. H., Kwok, H. S., Kaloyeros,
A. E., Eds.; Vol. 251; Springer: Buffalo, New York, 1992; p 231.
305. Troyanov, S. I.; Pisarevsky, A. P. J. J. Chem. Soc., Chem. Commun. 1993, 1,
335336.
306. Andrews, P. C.; Deacon, G. B.; Junk, P. C.; Kumar, I.; Silberstein, M. Dalton
Trans. 2006, 48524858.
307. Harjuoja, J.; Vayrynen, S.; Putkonen, M.; Niinisto, L.; Rauhala, E. J. Cryst. Growth
2006, 286, 376383.
308. Schuisky, M.; Kukli, K.; Ritala, M.; Harsta, A.; Leskela, M. Chem. Vap. Deposition
2000, 6, 139145.
309. Vehkamaki, M.; Hatanpaa, T.; Ritala, M.; Leskela, M. J. Mater. Chem. 2004, 14,
31913197.
310. Hatanpaa, T.; Vehkamaki, M.; Ritala, M.; Leskela, M. Dalton Trans. 2010, 39,
32193226.
311. Moiseev, A. N.; Chilyasov, A. V.; Dorofeev, V. V.; Vostrukhin, O. A.; Dianov, E. M.;
Plotnichenko, B. G.; Koltashev, V. V.; Novgorod, N. J. Optoelectr. Adv. Mater.
2005, 7, 18751879.
312. Moiseev, A. N.; Dorofeev, V. V.; Chilyasov, A. V.; Kutin, A. M.; Pimenov, V. G.;
Plotnichenko, V. G.; Koltashev, V. V. Inorg. Mater. 2007, 43, 675681.
313. Hodgson, S.; Weng, L. J. SolGel Sci. Techn. 2000, 18, 145158.
314. Coste, S.; Lecomte, A.; Thomas, P.; Merle-Mejean, T.; Champarnaud-
Mesjard, J. C. J. SolGel Sci. Techn. 2006, 41, 7986.
315. Hodgson, S. J. Non-Cryst. Solids 2000, 276, 195200.
316. Weng, L. Opt. Mater. 2002, 19, 313317.
317. Ikeda, H.; Fujino, S.; Kajiwara, T. J. Am. Ceram. Soc. 2009, 92, 26192622.
318. Wei, H. Y.; Lin, J.; Huang, W. H.; Feng, Z. B.; Li, D. W. Mater. Sci. Eng. B 2009,
164, 5159.
319. Bradley, D. C.; Mehrotra, R. C.; Rothwell, I. P.; Singh, A. Alkoxo and Aryloxo
Derivatives of Metals. 2nd ed.; Academic Press, 2001; pp 7374.
320. Fleischer, H.; Dienes, Y.; Schollmeyer, D. Eur. J. Inorg. Chem. 2002, 2002,
20732081.
321. Fleischer, H.; Schollmeyer, D. Inorg. Chem. 2001, 40, 324328.
322. Parola, S.; Papiernik, R.; Hubert-Pfalzgraf, L. G.; Jagner, S.; Hkansson, M. Dalton
Trans. 1997, 2, 46314636.
323. Thurston, J. H.; Whitmire, K. H. Inorg. Chem. 2002, 41, 41944205.
324. Thurston, J. H.; Kumar, A.; Hofmann, C.; Whitmire, K. H. Inorg. Chem. 2004, 43,
84278436.
325. Thurston, J. H.; Whitmire, K. H. Inorg. Chem. 2003, 42, 20142023.
326. Thurston, J. H.; Trahan, D.; Ould-Ely, T.; Whitmire, K. H. Inorg. Chem. 2004, 43,
32994305.
327. Thurston, J. H.; Ely, T. O.; Trahan, D.; Whitmire, K. H. Chem. Mater. 2003, 15,
44074416.
328. Dikarev, E. V.; Gray, T. G.; Li, B. Angew. Chem. Int. Ed Engl. 2005, 44,
17211724.
329. Hunger, M.; Limberg, C.; Kircher, P. Organometallics 2000, 19, 10441050.
330. Limberg, C.; Hunger, M.; Habicht, W.; Kaifer, E. Inorg. Chem. 2002, 41,
33593365.
331. Roggan, S.; Limberg, C.; Ziemer, B.; Brandt, M. Angew. Chem. Int. Ed Engl.
2004, 43, 28462849.
332. Wojcik, K.; Ruffer, T.; Lang, H.; Auer, A. A.; Mehring, M. J. Organomet. Chem.
2011, 696, 16471651.
333. Mendoza-Espinosa, D.; Hanna, T. A. Inorg. Chem. 2009, 48, 74527456.
334. Choujaa, H.; Johnson, A. L.; Kociok-Kohn, G.; Molloy, K. C. Polyhedron 2010,
29, 16071611.
335. Bergfeld, M.; Schmidbaur, H. Z. Anorg. Allg. Chem. 1968, 363, 8488.
336. Paalasmaa, S.; Mansfeld, D.; Schurmann, M.; Mehring, M. Z. Anorg. Allg. Chem.
2005, 631, 24332438.
337. Mehring, M.; Mansfeld, D.; Paalasmaa, S.; Schurmann, M. Chem. Eur. J. 2006,
12, 17671781.
338. Terry, K.; Su, K.; Dontilley, T.; Rheingold, A. Polyhedron 1998, 17,
891897.
339. Mansfeld, D.; Mehring, M.; Schurmann, M. Z. Anorg. Allg. Chem. 2004, 630,
17951797.
340. Mansfeld, D.; Mehring, M.; Schurmann, M. Angew. Chem. Int. Ed Engl. 2004, 44,
245249.
341. Nagao, Y.; Mimura, Y. Jap. J. Appl. Phys 1985, 24, L954L955.
342. Harjuoja, J.; Hatanpaa, T.; Vehkamaki, M.; Vayrynen, S.; Putkonen, M.;
Niinisto, L.; Ritala, M.; Leskela, M.; Rauhala, E. Chem. Vap. Deposition 2005, 11,
362367.
343. Armelao, L.; Colombo, P.; Fabrizio, M. J. SolGel Sci. Techn. 1998, 217,
213217.
344. Jain, M.; Tripathi, A.; Goel, T.; Pillai, P. J. Mater. Sci. Lett. 1999, 18, 479481.
345. Klebanskii, E. O.; Kudzin, A. Y.; Pasalskii, V. M.; Plyaka, S. N.; Sadovskaya, L. Y.;
Sokolyanskii, G. K. Phys. Solid State 1999, 41, 913915.
346. Veber, A.; Kunej, S.; Suvorov, D. Ceramics Int. 2010, 36, 245250.
347. Veber, A.; Kunej, S.; Cerc Korosec, R.; Suvorov, D. J. J. Eur. Ceram. Soc. 2010,
30, 24752480.
348. Ricofuentes, O.; Sanchezaguilera, E.; Velasquez, C.; Ortegaalvarado, R.;
Alonso, J.; Ortiz, A. Thin Solid Films 2005, 478, 96102.
349. Pugh, D.; Marchand, P.; Parkin, I. P.; Carmalt, C. J. Inorg. Chem. 2012, 51,
63856395.
350. Samedov, K.; Aksu, Y.; Driess, M. ChemPlusChem 2012, 77, 663674.
351. Samedov, K.; Aksu, Y.; Driess, M. Chem. Mater. 2012, 24, 20782090.
 2.02 Chemistry of Polar Transition-Metal Oxides
PS Halasyamani, University of Houston, Houston, TX, USA
2013 Elsevier Ltd. All rights reserved.

2.02.1 Introduction 41
2.02.2 Polar Oxide Materials Structure Types 42
2.02.2.1 Perovskite and Perovskite-Related Materials 42
2.02.2.1.1 BaTiO3 43
2.02.2.1.2 KNbO3 43
2.02.2.1.3 PbTiO3 43
2.02.2.1.4 LiNbO3 and LiTaO3 43
2.02.2.2 Aurivillius Phases 44
2.02.2.2.1 m 1: (Bi2O2)(BO4) 44
2.02.2.2.2 m 2: (Bi2O2)(AB2O7) 45
2.02.2.2.3 m 3: (Bi2O2)(A2B3O10) 45
2.02.2.2.4 m 4: (Bi2O2)(A3B4O13) 46
2.02.2.2.5 m 5: (Bi2O2)(A4B5O16) 46
2.02.2.3 Tetragonal Tungsten Bronzes 46
2.02.2.3.1 Barium sodium niobate 47
2.02.2.3.2 Strontium barium niobate 47
2.02.2.3.3 Potassium lithium niobate 47
2.02.2.3.4 Potassium sodium strontium barium niobate 47
2.02.2.3.5 Lead barium niobate 47
2.02.3 Multiferroic Materials 48
2.02.3.1 Perovskites 48
2.02.3.2 Hexagonal manganites 48
2.02.3.3 Metal halide and oxy-halide systems 48
2.02.4 Strategies on Designing Polar Oxide Materials 48
2.02.4.1 Metal oxyfluoride systems 50
2.02.4.2 Salt-Inclusion Solids 51
2.02.4.3 Borates 52
2.02.5 Functional Properties Associated with Polar Oxide Materials 52
2.02.5.1 Second-Harmonic Generation 54
2.02.5.1.1 Measurement of SHG and data interpretation 54
2.02.5.2 Piezoelectricity 55
2.02.5.2.1 Sample preparation and measurements 55
2.02.5.2.2 Direct piezoelectric measurements 55
2.02.5.2.3 Converse piezoelectric measurements 56
2.02.5.3 Pyroelectricity 56
2.02.5.3.1 Sample preparation and measurement 57
2.02.5.3.2 Pyroelectric current 57
2.02.5.3.3 Pyroelectric charge 57
2.02.5.4 Ferroelectricity 58
2.02.5.4.1 Sample preparation and hysteresis loop 58
2.02.6 Conclusion 59
Acknowledgments 59
References 59

2.02.1 Introduction state, however, a material is considered polar only if it is deter-

Polar oxides, that is, those that exhibit a dipole moment, are of
current interest attributable to their fascinating structural chem-
istry, as well as their uses in a variety of important technologies
such as pyroelectric sensors, ferroelectric memories, and imag-
ing.14 The concept of polarity is straightforward to understand
in simple molecules such as H2O, NH3, and CO, as the dipole
moment is toward the most electronegative atom. In the solid
mined to be in one of ten polar crystal classes: 1 (C1), 2 (C2), 3
(C3), 4 (C4), 6 (C6), m (Cs), mm2 (C2v), 3m (C3v), 4mm (C4v), or
6mm (C6v).5 In addition, specific polar directions have been
defined for each of the aforementioned crystal classes. In crystal
classes mm2, 3, 3m, 4, 4mm, 6, and 6mm the polar direction is
defined to be along the [001] direction, whereas for crystal class 2
(b-axis unique) the polar direction is along the [010] direction.
With crystal classes 1 and m, no polar direction is defined;

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00202-3 41

42 Chemistry of Polar Transition-Metal Oxides

Noncentrosymmetric crystal classes

Optical activity (circular dichroism)

LiIO3

Enantiomorphic Polar (pyroelectric)

432 4 (S4)
(O) 422 (D4)
m (Cs)
222 (D2) 1 (C1)
2 (C2) KTiOPO4
a-SiO2 622 (D6) 3 (C3) Ba2NaNb5O15
4 (C4)
32 (D3) 6 (C6) mm2
23 (T) (C2v)
LiNbO3
42m (D2) b-BaB2O4
BiFeO3
6 (C3h) 43m (Td) 3m (C3v)
CO(NH2)2 (urea) 4mm (C4v)
(NH4)H2PO4(ADP) 62m (D3h) 6mm (C6v) BaTiO3

Second-harmonic generation, piezoelectric

Figure 1 Symmetry-dependent property relationships between the noncentrosymmetric crystal classes are shown along with representative materials.
Note that polar crystal classes are a subset of noncentrosymmetric crystal classes. Adapted from Halasyamani, P. S.; Poeppelmeier, K. R. Chem. Mater.,
1998, 10, 2753. Copyright 1998 American Chemical Society.

however, as stated in the International Tables for Crystallography

Perovskite structure
Volume A, Table 10.2.1.2, All directions other than those in the
column Nonpolar direction are polar. Thus for crystal class 1
this is any direction, as there are no defined nonpolar directions,
whereas for crystal class m (unique axis b) the polar directions are
any directions other than the [010] defined to be nonpolar.
Before we delve into the details regarding polar oxides, it is
instructive to briefly discuss polarity with respect to noncentro-
symmetry as a whole. Polarity is a subset of noncentrosymmetry,
that is, polarity is just one of the symmetry-dependent non-
centrosymmetric (NCS) properties. Other symmetry-dependent
properties include piezoelectricity, circular dichroism, and chiral-
ity. The relationships between NCS crystal classes and symmetry-
dependent property are shown in Figure 1. It is of interest to note Figure 2 Ball-and-stick diagram of the cubic perovskite structure,
that materials that are polar also have the correct symmetry for ABO3. Yellow spheres A cations, blue spheres B cations, and red
spheres O2.
second-harmonic generation (SHG) and piezoelectricity. Polar
oxides of note include BaTiO3 (4mm),6 LiNbO3 (3m),7,8
BiFeO3 (3m),913 and KTiOPO4 (mm2).14 Interest in polarity 2.02.2 Polar Oxide Materials Structure Types
and polar materials stems from two technologically important 2.02.2.1 Perovskite and Perovskite-Related Materials
functional properties pyroelectricity and ferroelectricity. For
either to manifest, the material in question must be polar. With The largest group of oxide materials that are polar can be found
pyroelectricity, a change in the magnitude of the polarization in a perovskite, or perovskite-related, structure (see Chapter
(dipole moment) with respect to temperature is observed,2 2.01). The cubic, ABO3, perovskite structure is shown in Figure 2.
whereas with ferroelectricity the macroscopic polarization may The larger A cation is 12-coordinate, whereas the B cation is in
be reversed in the presence of an external electric field. Thus, all an octahedral coordination environment. Essentially, the perov-
ferroelectrics are pyroelectrics, but the converse is not always true. skite structure may be described as a three-dimensional network
This chapter is divided into three sections. The first describes of corner-shared BO6 octahedra. Alternatively, it may also be
polar oxides in well-known structure-types, for example, perov- regarded as a cubic close-packed arrangement of A and O ions,
skite and perovskite-related oxides, Aurivillius phases, and tetrag- with the B cations filling the interstitial octahedral positions.
onal tungsten bronzes (TTBs). Within each of these structure A tolerance factor, t, has been defined by Goldschmidt15 for the
types, specific oxide materials are described and discussed. perovskite structure:
Also, multiferroic materials are discussed. In the second section, RA RO
strategies on creating new polar oxide materials are discussed. In t p 
2RB RO
addition, a list of recently synthesized polar oxides is given. The
final section deals with properties associated with polar oxides, where RA, RB, and RO are the ionic radii for the A, B, and O ions
as well as their characterization. respectively. An ideal cubic perovskite is observed when

t 1.0. A stable perovskite structure is generally observed
for 0.85 < t < 1.05. For t < 1.0 BO6 octahedral rotations are
observed, whereas for t > 1.0 the B cation is displaced from
the center of its oxide octahedron resulting in local polarity.
There are a number of polar oxide materials that exhibit pe-
rovskite or perovskite-related structures, so only a few will be
described.
2.02.2.1.1 BaTiO3
A polar perovskite oxide that has generated the most interest and
has been among the most thoroughly investigated is BaTiO3.6
Originally determined to be ferroelectric by von Hippel et al.16
and Wul and Goldman,17 the room-temperature structure
consists of corner-shared TiO6 octahedra that are separated by
Ba2 cations (see Figure 3).6 In the room-temperature structure,
the Ti4 cation is not in the center of its oxide octahedron;
rather, it is displaced toward a corner resulting in one short,
one long, and four normal TiO bonds. This displacement
creates a local dipole moment in each TiO6 octahedron, and as
all of the Ti4 displacements are parallel, macroscopic polariza-
tion toward the c-axis direction is observed. At room tempera-
ture, BaTiO3 is found in the polar tetragonal space group P4mm
(crystal class 4mm), with a spontaneous polarization of
26 mC cm2.3 As the temperature is lowered to below 0  C, a
phase change occurs. Below 0  C, BaTiO3 is still polar but the
polarization direction has changed. The Ti4 cation, instead of
being shifted toward a corner of its oxide octahedron, is now
displaced toward an edge resulting in two short, two long, and
two normal TiO bonds. The symmetry of BaTiO3 changes to
the polar orthorhombic crystal class mm2, and macroscopic
polarization is directed toward the [11 0]. At 90  C, another
polar ! polar phase change occurs, as the symmetry changes
from orthorhombic mm2 to rhombohedral 3m. The displace-
ment of the Ti4 cation is toward a face, resulting in three short
and three long TiO bonds. The macroscopic polar direction
has also changed, and is toward the [11 1]. Finally, the cubic
perovskite structure, in crystal class m3m (nonpolar and
centrosymmetric), is observed in BaTiO3 above 120  C the
Curie temperature. Above this temperature, all of the TiO
bonds are the same length, and the polarity, local and macro-
scopic, no longer exists.
BaTio3
c
b
a
Figure 3 Ball-and-stick representation of room temperature BaTiO3.
Green spheres Ba2, blue spheres Ti4, and red spheres O2.
Note that the Ti4 cations are displaced toward a corner of their oxide
octahedron resulting in local polarity, and one short, one long, and four
normal TiO bonds. The local polar moments are aligned toward the
c-axis producing a macroscopic polar material.
Chemistry of Polar Transition-Metal Oxides 43
2.02.2.1.2 KNbO3
Potassium niobate, KNbO3, was first discovered in 1949 as a
perovskite-type ferroelectric.18 The ferroelectric properties of
KNbO3 are very similar to BaTiO3, as KNbO3 exhibits a similar
sequence of phase transitions. At room temperature KNbO3 is
orthorhombic, space group Amm2 (crystal class mm2), and the
material undergoes a polar ! nonpolar centrosymmetric phase
transition at 435  C, that is, the Curie temperature.19 The sym-
metry of the polar phase below 435  C is tetragonal, and a phase
change to a polar orthorhombic structure, the room-temperature
structure, is observed at 225  C.2022 Finally, a polar ortho-
rhombic to polar rhombohedral phase change is observed at
10  C.2022 As with BaTiO3, all of the phase transitions are
first order. The spontaneous polarization magnitude of KNbO3
at room temperature is 30 mC cm2, similar to that of BaTiO3.19
2.02.2.1.3 PbTiO3
Lead titanate, PbTiO3, is a perovskite-type ferroelectric that is
also similar to BaTiO3.23,24 The room-temperature structure of
PbTiO3 is tetragonal, space group P4mm (crystal class 4mm),
but the material only undergoes one phase change, its Curie
temperature, at 490  C. Above this temperature, PbTiO3 has
the centrosymmetric cubic perovskite structure. Similar to
BaTiO3, in PbTiO3 the Ti4 cation is displaced toward a corner
of its oxide octahedron, resulting in one long, one short, and
four normal TiO bonds. The displacement of the Ti4 cation
is larger in PbTiO3 compared with BaTiO3 that is thought to
result in the substantially greater spontaneous polarization of
57 mC cm2.25 Recall that the spontaneous polarization for
BaTiO3 is 26 mC cm2.
2.02.2.1.4 LiNbO3 and LiTaO3
Although LiNbO3 and LiTaO3 do not exhibit the perovskite
structure, the materials do contain oxide octahedra. The struc-
tures have been determined with high precision through x-ray
and neutron diffraction, and were shown to be in the polar
trigonal space group R3c (crystal class 3m).7,8 The materials are
isostructural and exhibit face-sharing and alternating LiO6 and
MO6 (M Nb5 or Ta5) octahedra (see Figure 4). The direc-
tion of face sharing is along the polar c-axis of the trigonal cell.
In LiMO3 the Nb5 and Ta5 cations are substantially shifted
LiNbO3
c
b a
Figure 4 Ball-and-stick representation of room temperature LiNbO3.
Yellow spheres Li, blue spheres Nb5, and red spheres O2.
Note that the Nb5 cations are displaced toward a face of their oxide
octahedron resulting in local polarity, and three short and three long
NbO bonds. The local polar moments are aligned toward the c-axis
producing a macroscopic polar material.
44 Chemistry of Polar Transition-Metal Oxides
toward a face of their oxide octahedron, unlike the corner-
directed displacement of the Ti4 cation in BaTiO3 and
PbTiO3, resulting in three short and three long MO bonds.
The cation distortions result in large room-temperature spon-
taneous polarizations of 70 and 50 mC cm2 for LiNbO3 and
LiTaO3, respectively.3 In addition, the Curie temperatures
are also high, 1200  C and 620  C for LiNbO3 and LiTaO3,
respectively.3
2.02.2.2 Aurivillius Phases
The family of Aurivillius phases2628 may be formulated as
(Bi2O2)(Am1BmO3m1) where A is a 12-fold coordinated cat-
ion, for example, Sr2, Ba2, Bi3, etc., and B is a cation in an
octahedral coordination environment, for example, Ti4,
Nb5, W6, etc. The Aurivillius phases are structurally de-
scribed as being built up from alternating layers of (Bi2O2)2
cations and perovskite-like anion blocks (Am1BmO3m1)2,
with m being an integer corresponding to the number of
corner-shared octahedra forming the perovskite blocks. The
idealized structure, along the c-axis, for the m 13 phases,
Bi2WO6,29 Bi3TiNbO9,30 and Bi4Ti3O12,31 respectively, is
shown in Figure 5. The polarity observed in some Aurivillius
phases emanates structurally from both the (Bi2O2)2 layer
(Bi2O2)(Am-1BmO3m+1)
Bi
3+
Bn+ (Bn+ = Ti4+,Nb5+ or W6+)
O2
m=1 m=2 m=3
Figure 5 Idealized representations of the m 1, 2, and 3 Aurivillius
phases.
b
a a
Figure 6 Representation of the m 1 Aurivillius phase: (Bi2O2)(BO4).
Green spheres Bi3, blue spheres B cation (Mo6 or W6), and red
spheres O2. Note that there is a single (m 1) octahedral block
between the (Bi2O2)2 layers.
and the perovskite blocks. In the idealized structures (see
Figure 5), the Bi3 cations in the (Bi2O2)2 layer have equal
BiO bonds. In addition, the BO bonds in the perovskite
blocks are equal. These idealized structures, however, do not
represent actual crystal structures. Representations of actual
crystal structures for m 13 are shown in Figures 68. As
seen, a number of distortions are observed. In the (Bi2O2)2
layer, a stereo-active lone pair on Bi3 is observed and results
in an asymmetric polar coordination environment. This asym-
metric coordination environment causes structural strain in
the perovskite blocks and results in octahedral rotation.
In addition to the octahedral rotations in the perovskite blocks,
an intra-octahedral displacement of the d0 transition metal is
observed. These displacements render the perovskite blocks
polar, and thus the entire Aurivillius phase is structurally
polar. Polar Aurivillius phases that are structurally well deter-
mined are listed in Table 1. For clarity, the stoichiometries are
also written using the Aurivillius formulation.
2.02.2.2.1 m 1: (Bi2O2)(BO4)
Two polar m 1 Aurivillius phases have been reported and
structurally characterized Bi2MoO6 and Bi2WO6. For the
m 1 Aurivillius phase member, the (Bi2O2)(Am1BmO3m1)
formulation is simplified to (Bi2O2)(BO4), or in the case of the
reported materials, (Bi2O2)(MoO4) and (Bi2O2)(WO4). Thus,
both structures may be described as having a (Bi2O2)2 layer
separated by corner-shared MO6 (M Mo6 or W6) octahe-
dra (see Figure 6). The structure of Bi2MoO6 was correctly
determined in 1984 by Theobald et al.32 Through neutron
diffraction measurements, they determined the space group
to be Pca21 (crystal class mm2), with the Mo6 cations dis-
placed toward an edge of its oxide octahedron. Bi2WO6 was
c
b
Figure 7 Representation of the m 2 Aurivillius phase: (Bi2O2)
(AB2O7). Green spheres  Bi3, blue spheres  B cation, and red
spheres  O2. Note that there is a double (m 2) octahedral block
between the (Bi2O2)2 layers.
 Chemistry of Polar Transition-Metal Oxides 45

originally reported in 1969 by Wolfe et al.33 The structure was
re-determined correctly in 1991 by Rae et al. using a commen-
surately modulated model.29 In addition, Rae et al. used bond
valence concepts34 to explain the driving for the atomic dis-
placements and subsequently the ferroelectric behavior.
2.02.2.2.2 m 2: (Bi2O2)(AB2O7)
The polar m 2 Aurivillius phases include Bi3TiNbO9,30
ABi2M2O9 (A Ca2, Sr2, Ba2, Pb2; M Nb5, Ta5),3541
Bi2.5Na0.5Nb2O9,42 and the A cation-deficient Bi2W2O9.43 All
of the materials contain a (Bi2O2)2 layer separated by a dou-
ble octahedral perovskite block (see Figure 7). The first three
materials listed are found in space group A21am (crystal class
2mm), with the spontaneous polarization directed toward the
a-axis direction. Shimakawa et al. have demonstrated that both
the Bi3 and the d0 transition-metal cations, Ti4, Nb5, and
Ta5, contribute to the polarization. The spontaneous polari-
zation has been measured at room temperature for only SrBi2-
Ta2O9 (Ps 5.8 mC cm2).3 A slightly different situation is
observed for Bi2W2O9.43 In this m 2 Aurivillius phase, there
is no A cation in the perovskite block, and thus the material is
deemed to be cation deficient. Bi2W2O9 is found in space
group Pna21 with a macroscopic polarization directed toward
the c-axis. Interestingly, the local polarizations associated with
the W6 cation cancel they are pointed in opposite direc-
tions. Thus, the macroscopic polarization is wholly attribut-
able to the asymmetric coordination of the Bi3 cation.

2.02.2.2.3 m 3: (Bi2O2)(A2B3O10)
The reported polar m 3 Aurivillius phases are Bi4Ti3O1231,44
b and Bi2.5Na1.5Nb3O12.45 Both materials contain a (Bi2O2)2
layer separated by triple octahedral perovskite-like blocks (see
Figure 8). Bi4Ti3O12 was originally reported in 1971 by Dorian
a et al.44 As with Bi2WO6, the structure was re-refined correctly in
1990 by Rae et al. using a commensurately modulated
model.31 Rae et al. also used bond valence concepts to explain
Figure 8 Representation of the m 3 Aurivillius phase: (Bi2O2) the driving force for the cation displacements and subsequent
(A2B3O10). Green spheres Bi3, blue spheres B cation, and red ferroelectric behavior. Unlike Bi2WO6, the spontaneous polar-
spheres O2. Note that there is a triple (m 3) octahedral block ization has been measured for Bi4Ti3O12. Ferroelectric mea-
between the (Bi2O2)2 layers. surements indicate a spontaneous polarization at room

Table 1 Polar Aurivillius phases: (Bi2O2)(Am1BmO3m1)

Compound Space group Crystal class Reference

m 1: (Bi2O2)(BO4)
Bi2WO6 (Bi2O2)(WO4) P21ab 2mm 29
Bi2MoO6 (Bi2O2)(MoO4) Pca21 2mm 32
m 2: (Bi2O2)(AB2O7)
Bi3TiNbO9 (Bi2O2)(BiTiNbO7) A21am 2mm 30
ABi2M2O9 (Bi2O2) (AM2O7) (A Ca, Sr, Ba, Pb; M Nb, Ta) A21am 2mm 3541
Bi2.5Na0.5Nb2O9 (Bi2O2)(Bi0.5Na0.5Nb2O7) A21am 2mm 42,45
(Bi2O2)(W2O7) Pna21 mm2 43
m 3: (Bi2O2)(A2B3O10)
Bi4Ti3O12 (Bi2O2)(Bi2Ti3O12) B1a1 m 31
Bi2.5Na1.5Nb3O12 (Bi2O2)(Bi0.5Na1.5Nb3O10) Aba2 mm2 45

m 4: (Bi2O2)(A3B4O13)
ABi4Ti4O15 (AxBi2xO2)(AyBi2yTi4O13) (A Ca, Sr, Ba, Pb) A21am 2mm 48
Bi5Ti3FeO15 (Bi2O2)(Bi3Ti3FeO13) A21am 2mm 49,237
SrBi4Ti4O15 (Bi2O2)(SrBi2Ti4O13) A21am 2mm 49
Bi2.5Na2.5Nb4O15 (Bi2O2)(Bi0.5Na2.5Nb4O13) A21am 2mm 45
m 5: (Bi2O2)(A4B5O16)
A2Bi4Ti5O18 (AxBi2xO2)(AyBi2yTi5O16) (A Ca, Sr, Ba, Pb) B2cb 2mm 50
46 Chemistry of Polar Transition-Metal Oxides
temperature of 3050 mC cm2.46,47 With Bi2.5Na1.5Nb3O12
(Bi2O2)(Bi0.5Na1.5Nb3O10), a slightly different situation is ob-
served. Although the (Bi2O2)2 layers are separated by triple
octahedral perovskite-like blocks, the A cations in the
perovskite-like blocks are statistically disordered between Bi3
and Na. In addition, Bi2.5Na1.5Nb3O12 is found in the polar
orthorhombic space group B2cb (crystal class 2mm), with the
polarization direction pointed toward the a-axis. No spontane-
ous polarization data have been reported for this material.
2.02.2.2.4 m 4: (Bi2O2)(A3B4O13)
With the polar m 4 Aurivillius phases, ABi4Ti4O15 (A Ca,
Sr, Ba, and Pb),48 Bi5Ti3FeO15,49 SrBi4Ti4O15,49 and
Bi2.5Na2.5Nb4O1545 have been reported. All of the materials
contain a (Bi2O2)2 layer separated by a quadruple octahedral
perovskite-like block (see Figure 9). In ABi4Ti4O15, statistical
cation disorder occurs in both the Bi2O2 layer and the
perovskite-like block. Specifically, Ca, Sr, Ba, or Pb is statisti-
cally disordered with Bi3 in the Bi2O2 layer, as well as the Bi3
x(B2O5) (A Na, K, Sr,
in the perovskite-like block. All four materials are found in
space group A21am with spontaneous polarization values rang-
ing between 21 and 34 mC cm2.48 Statistical disorder is also
observed in Bi5Ti3FeO15, SrBi4Ti4O15, and Bi2.5Na2.5Nb4O15.
In Bi5Ti3FeO15, the disorder is observed in the B cation in
the perovskite-like block, that is, between Ti4 and Fe3,
whereas in SrBi4Ti4O15 and Bi2.5Na2.5Nb4O15 the disorder is
observed in the A cation of the perovskite block, that is, Bi3/
Sr2 and Bi3/Na, respectively. For Bi5Ti3FeO15, SrBi4Ti4O15,
and Bi2.5Na2.5Nb4O15, no spontaneous polarization data have
been reported.
c c
a b
Figure 9 Representation of the m 4 Aurivillius phase:
(Bi2O2)(A3B4O13). Green spheres Bi3, blue spheres B cation, and
red spheres O2. Note that there is a quadruple (m 4) octahedral
block between the (Bi2O2)2 layers.
2.02.2.2.5 m 5: (Bi2O2)(A4B5O16)
The only polar m 5 Aurivillius phase family that has been
reported and structurally characterized is A2Bi4Ti5O18 (A Ca,
Sr, Ba, and Pb).50 All of the materials contain a (Bi2O2)2 layer
separated by a quintiple octahedral perovskite-like block (see
Figure 10). Similar to ABi4Ti4O15 (A Ca, Sr, Ba, and Pb),
there is statistical cation disorder with the Bi3 cation in both
the (Bi2O2)2 layer and the perovskite-like block. All four
materials are found in the polar space group B2cb (crystal
class 2mm); however, spontaneous polarization data have
not been reported.
2.02.2.3 Tetragonal Tungsten Bronzes
The TTB structure consists of a corner-sharing framework of
MO6 octahedra that contain interstitial sites where a host
of metal cations may reside. The bronze family may be divided
into three groups: the nonstoichiometric, highly colored
bronzes, compounds of the type (AO)
Ba, or Pb; B Nb, Ta, or W), and materials with a general
formula (A1)2(A2)4C4(M1)2(M2)8O30.51 The focus of this sec-
tion is on the latter type of materials, as these have the most
relevance with respect to polar oxides. The archetypal structure
for the (A1)2(A2)4C4(M1)2(M2)8O30 stoichiometry is shown
in Figure 11. The A, C, and M sites may have full or partial
occupancy depending on the cation. The A sites contain alkali,
alkaline earth, or Pb2 cations, whereas for the vast majority of
TTBs the M sites are fully occupied by Nb5 or Ta5. The C
sites, if they are occupied at all, are necessarily small given the
restricted coordination environment. The observed polarity in
the TTBs may be attributed to displacements of the M cation
b
Figure 10 Representation of the m 5 Aurivillius phase:
(Bi2O2)(A4B5O16). Green spheres Bi3, blue spheres B cation, and
red spheres O2. Note that there is a quintuple (m 5) octahedral
block between the (Bi2O2)2 layers.

A1
M2
M1
A2
C (KLN).63 At room temperature, KLN is found in the polar
b tetragonal space group P4bm, with the A1 and A2 sites fully
a
Figure 11 Polyhedral representation of a layer of the archetypal
(A1)2(A2)4C4(M1)2(M2)8O30 tetragonal tungsten bronze structure.
from the center of its octahedron, that is, second-order Jahn
Teller (SOJT) effects. We will be discussing SOJT effects in more
detail later in the chapter. TTBs are termed incompletely filled,
filled, and completely filled, depending on the extent to which
the A, C, and M sites are occupied. With respect to symmetry,
TTBs are usually found in the polar tetragonal space group
P4bm (crystal class 4mm); however, lower polar symmetries
are possible.5154 Some of the well-known and characterized
polar TTB oxides will be discussed.
2.02.2.3.1 Barium sodium niobate
Ba2NaNb5O15 (BNN) was first reported in 1967 by Geusic
et al.55 Also termed the banana phase, BNN has the filled
TTB structure where the A1 and A2 sites are completely occu-
pied by Na and Ba2 respectively, and the M sites are filled by
Nb5. The C sites are empty. At room temperature BNN is
found in the polar orthorhombic space group Cmm2 (crystal
class mm2), with an experimentally determined spontaneous
polarization of 40 mC cm2.56 Other polar oxides that exhibit
the BNN structure include Pb2KNb5O15 (PKN),57 Sr2KNb5O15
(KSN),58 and nonstoichiometric phases such as La0.1Li0.1Na-
Ba1.8Nb5O1559 and K0.8Na0.2Ba2Nb5O15, as well as solid
solutions (Ba1xSrx)2NaNb5O15, (Ba1xSrx)2LiNb5O15, and
(Ba1xMgx)2NaNb5O15.60 In all of the BNN and BNN-related
phases, the polarity is attributable to intra-octahedral displace-
ments of the Nb5 cation from the center of its oxide octahe-
dron. These displacements render the NbO6 octahedron
locally polar, and additionally the local polarity constructively
adds resulting in a macroscopically polar material.
2.02.2.3.2 Strontium barium niobate
The solid solution Sr1xBaxNb2O6 (SBN) where 0.20 < x < 0.80
has the TTB structure and is found in the polar tetragonal space
group P4bm.61 In SBN, the A1 site is partially occupied by Sr2,
whereas the larger A2 site has Sr2 and Ba2 statistically disor-
dered. As with BNN, the M sites are fully occupied by Nb5 and
the C sites are empty. The Sr2/Ba2 ratio has been found over
a wide range, and for different applications SBN crystals with
Chemistry of Polar Transition-Metal Oxides 47
different Sr2/Ba2 ratios have been grown. In fact, by chang-
ing the Sr2/Ba2 ratio the ferroelectric Curie temperature may
be varied from 60 to 250  C. For Sr2/Ba2 75/25, a sponta-
neous polarization of 32 mC cm2 has been measured, with the
polarization along the c-axis direction in the polar tetragonal
space group P4bm.62 As with the BNN and BNN-related mate-
rials, the local and macroscopic polarity is attributable to
displacements of the Nb5 cation.
2.02.2.3.3 Potassium lithium niobate
A completely filled TTB structure is observed for K3Li2Nb5O15
occupied by K and the M sites fully occupied by Nb5. Unlike
the BNN and SBN materials, in KLN the C sites are fully occu-
pied by Li cations. It is reported in KLN that the d31 nonlinear
optical (NLO) coefficient is equal to that of LiNbO3.63 Crystals
of the solid solution, K3Li2(Nb1xTax)Nb5O15 (KLTN) have
been grown by the Kyropoulos technique, where some of the
Nb5 is replaced by Ta5.64 It has been shown that for x  0.50
the structure is tetragonal, whereas for x > 0.50 the structure is
orthorhombic.64 In addition to KLTN, Sr4NaLiNb10O30 (SNLN)
also exhibits the completely filled TTB structure.65 A melting
point of 1380  C has been reported and colorless single crystals
of SNLN were grown by the Czochralski method.65 In addition,
a spontaneous polarization of 1520 mC cm2 has been mea-
sured at room temperature.65
2.02.2.3.4 Potassium sodium strontium barium niobate
A family of polar oxides with the stoichiometric range
(Na1xKx)0.4(SrxBa1x)0.8Nb2O6 (0.5 < x < 0.75, 0.30 <
y < 0.90) with the TTB structure has been synthesized. The
materials are found in the polar tetragonal space group
P4bm. For these materials, the A sites are fully occupied with
Na and K statistically disordered on the A1 site, and Sr2 and
Ba2 statistically disordered on the A2 site. The C sites are
empty and the M sites are fully occupied by Nb5. The family
has been termed KNSBN by Xu and Chen,66 and BSKNN by
Neurgaonkar et al.67 Large single crystals of KNSBN/BSKNN
have been grown by the Czochralski method. For x 0.5
and y 0.75, (Na0.5K0.5)0.4(Sr0.75Ba0.25)0.8Nb2O6, a KNSBN/
BSKNN, crystal has been shown to have a spontaneous polar-
ization of 35 mC cm2.67
2.02.2.3.5 Lead barium niobate
The solid solution PbxBa1xNb2O6 with 0 < x < 1 has also been
shown to exhibit the TTB structure.68 An orthorhombic struc-
ture is observed for 0.53 > x > 0.20, whereas a tetragonal
structure is found for x > 0.63. For 0.53 < x < 0.63, a two-
phase mixture is found. However, the crystal structure of
Pb0.7Ba0.3Nb2O6 (x 0.7) has been reported in the polar or-
thorhombic space group Cm2m (crystal class m2m).69 In the
proposed model, one of the A2 sites is fully occupied by Pb2
cations, whereas the other A2 and all of the A1 sites are statis-
tically disordered between Ba2 and Pb2. As with BNN, SBN,
and KNSBN/BSKNN, the C sites are empty and the M sites are
fully occupied by Nb5. A crystal of the x 0.37 phase,
Pb0.37Ba0.63Nb2O6, has also been grown and is reported to be
in the polar crystal class 4mm.60
48 Chemistry of Polar Transition-Metal Oxides
2.02.3 Multiferroic Materials
One area where polar oxide materials have seen a revival is
with multifunctional, specifically multiferroic materials. (see
Chapter 4.11)7072 With multiferroics, the definition given by
Schmid will be used a material is considered multiferroic if at
least two primary ferroic properties occur in the same mate-
rial.73 For our purposes, this would be magnetic ordering, anti-
, ferri-, or ferromagnetism, and ferroelectricity. The physics of
these materials is beyond the scope of this chapter, but struc-
tural polarity is required. A few families of multiferroics have
been described, and we will briefly discuss each of these as well
as their polar attributes.
2.02.3.1 Perovskites
With perovskite multiferroic materials, the most extensively
studied material is BiFeO3.7478 This material is reported
to be ferroelectric, ferroelastic, and weakly ferromagnetic. The
material exhibits trigonal symmetry, crystallizing in the polar
crystal class 3m. The structure of BiFeO3 consists of BiO6 and
FeO6 octahedra that are corner, edge, and face shared. The Fe3
cation is effectively undisplaced, bonded to six oxygen atoms
with nearly equal FeO distances. The main structural distor-
tion is from the Bi3 cation, which exhibits a lone pair.
The lone pair results in an asymmetric coordination environ-
ment, with three short BiO bonds (2.34 A) and three long
BiO bonds (2.56 A). Thus, the BiO6 octahedra are locally
polar. It is this local polarity, attributable to the lone pair that is
responsible for the observed ferroelectric behavior. In other
words, the polarization associated with the lone-pair cation is
being switched. It is interesting to note that polarization
reversibility in lone-pair cations has only been observed with
sixth period elements, that is, Tl, Pb2, and Bi3. Ferroelectric
behavior involving fourth and fifth period lone-pair cations,
that is, Se4, Sn2, Te4, and I5, where the polarization on
these cations is reversed, has never been observed. It should be
noted that LiH3(SeO3)2 is ferroelectric, but the ferroelectric
behavior is attributable to hydrogen bonding and not the
Se4 cation.79
2.02.3.2 Hexagonal manganites
These materials have the RMnO3 (R Sc or small rare earth
cation) stoichiometry80 and have been erroneously referred to
as hexagonal perovskites. It is emphasized that the compounds
do not exhibit the perovskite structure. The Mn3 cations are
not octahedrally coordinated; rather the cation is surrounded
by five oxide anions in a trigonal bipyramidal coordination
environment. Also the R cations are not 12-coordinate, as
would be the case in a perovskite, but are in sevenfold coordi-
nation. The materials are multiferroic, with antiferromagnetic
and ferroelectric properties.81 The nature of the polarity and
therefore the ferroelectric behavior were only recently de-
scribed.82 Careful structural studies indicated that although
the dipole moments are attributable to the RO bonds and
not the MnO bonds, the R-cations are not directly responsible
for the ferroelectric behavior. The noncentrosymmetry is at-
tributable to the tilting of the MnO5 polyhedra, which in
conjunction with the dipole moments in the RO bonds
results in ferroelectric behavior. Thus, the ferroelectric behavior
in these materials is termed improper83 and occurs by a much
different mechanism than BaTiO3 or even BiFeO3.
2.02.3.3 Metal halide and oxy-halide systems
Multiferroic behavior has also been observed in metal halide
materials, specifically BaMF4-type compounds (M Mg, Mn,
Fe, Co, Ni, and Zn)8486 and boracites, M3B7O13X (M Cr,
Mn, Fe, Co, Ni, or Cu; X Cl, Br, or I).8789 The former consists
of puckered layers of corner-shared MF6 octahedra that are
separated by Ba2 cations. The materials are isostructural,
crystallizing in the orthorhombic crystal class mm2. Anti-
ferromagnetic ordering has been observed for M Mn, Fe,
Co, and Ni, but only BaCoF4 and BaNiF4 are ferroelectric.
The transition metal in the BaMF4 materials is in a nearly
regular octahedral environment. The origin of the ferroelectric-
ity has been recently elucidated and occurs through the rota-
tion of the MF6 (M Co or Ni) octahedra and polar
displacements of the Ba2 cation.90 It is suggested that the
reason ferroelectric behavior occurs in the Co and Ni phases
and not the Mn and Fe analogs is attributable to geometric
constraints and size effects.90 Boracite-type compounds have
the stoichiometry M3B7O13X (M Cr, Mn, Fe, Co, Ni, or Cu;
X Cl, Br, or I) and exhibit three-dimensional crystal struc-
tures. The structures consist of linked MO4X, BO4, and BO3
polyhedra. The materials are isostructural crystallizing in
the trigonal crystal class 3m. Complex ferromagnetic order-
ing below 100 K has been observed in Mn3B7O13Cl and
Ni3B7O13Cl.91,92 Ferroelectric hysteresis loops have also been
measured for these compounds as well,88,93 indicating polari-
zation reversibility. What is unclear is the origin of the ferro-
electric behavior. The polarity in the boracite structure is likely
attributable to the MO4X and BO3 polyhedra. In these
materials, the BO3 polyhedra are not strictly planar, but are
found as a trigonal pyramid with boron at the apex. Thus, both
cations are in locally acentric and polar coordination environ-
ments, whereas the BO4 tetrahedra are only acentric. Thus,
polarization reversal, that is, ferroelectricity, must occur be-
cause of the MO4X and/or BO3 polyhedra. To date, however,
no mechanism has been confirmed.
2.02.4 Strategies on Designing Polar Oxide Materials
In the past decade or so, a number of strategies have been put
forth whose aim is to design new polar oxide materials (see
Chapter 4.01). In one manner or another, these strategies
involve crystal engineering.94 One question that needs to be
addressed is why are there so few polar oxides with respect to
all known oxide materials? Less than 10% of all oxides are
polar. This would suggest that in the vast majority of oxides,
the building blocks of the structure contain a center of sym-
metry, that is, the building blocks are undistorted or regular
metal-oxide polyhedra. In addition, these regular polyhedra
are related by inversion symmetry. Thus, in order to design
polar oxide materials, two challenges must be successfully
addressed. First, the building blocks of the structure must
necessarily be intrinsically polar. Either the metal oxide poly-
hedra needs to be intrinsically polar, for example, square py-
ramidal, or there must be a displacement that requires or forces

C3[1 1 1]
C2[1 1 0]
Figure 12 Out-of-center distortions of octahedrally coordinated d0 cations along the local C2 [110], C3 [111], or C4 [001] direction. Reproduced with
permission from Halasyamani, P. S. Chem. Mater. 2004, 16, 3586. Copyright 2004 American Chemical Society.
the metal cation not to be on an inversion center. If local
centricity is observed, macroscopic centricity occurs. Second,
these building blocks must be connected or related in the
structure by noninversion-type polar symmetry elements. In
other words, it is not sufficient to only have polar oxide poly-
hedra; these polyhedra must be related by polar symmetry
elements. One manner in which new polar oxide materials
may be designed is to use cations susceptible to SOJT distor-
tions (see Chapter 4.08).95101 These cations are octahedrally
coordinated d0 transition metals, Ti4, Nb5, W6, etc., and
cations with nonbonded electron pairs, or lone pairs, Sn2,
Se4, Pb2, etc. With the octahedrally coordinated d0
transition-metal cations, SOJT effects are observed when the
empty d-orbitals of the metal mix with the filled p-orbitals
of oxygen. In extended structures, this mixing results in a
host of nearly degenerate electronic configurations that can
be removed through the spontaneous distortion of the d0
transition-metal cation. This cation distortion can occur to-
ward an edge (local C2 direction), face (local C3 direction),
or corner (local C4 direction), or between these special di-
rections (see Figure 12). The distortion results in unequal MO
bond distances, and creates an MO6 octahedron that is locally
polar. SOJT distortions are observed in well-known polar ma-
terials such as BaTiO3, LiNbO3, Bi2WO6, and Ba2NaNb5O15. In
these materials, the d0 cation displacement is toward a corner,
BaTiO3 and Ba2NaNb5O15, or face, LiNbO3, resulting in MO6
(M Ti4 or Nb5) octahedra that are locally polar. With the
lone-pair cations, the original work of Sidgwick and Powell102
followed by the valence shell electron pair repulsion (VSEPR)
theory of Gillespie and Nyholm103 attempted to rationalize the
asymmetric coordination geometry. It was, however, Orgel104
who explained the structural distortion and polarization
through the mixing of the metal s- and p-orbitals. This tradi-
tional view of metal cation sp orbital mixing has been shown
to be incomplete. A number of researchers have shown that the
interaction of the s- and p-orbitals of the metal cation with the
oxide anion p-states is critical for lone-pair formation.77,105111
H2O,143,149 an additional lone-pair cation,
Regardless of how the lone pair is created, its structural conse-
quences are profound (see Figure 13). The lone pair pushes the
oxide ligands toward one side of the metal cation resulting in a
locally polar coordination environment. The lone-pair cation
coordination environment may be considered as pre-
distorted,112 as these cations are almost always found in a locally
polar environment.
Utilizing SOJT-distorted cations to design and synthesize new
polar oxide materials has been done by Halasyamani,54,113130
Chemistry of Polar Transition-Metal Oxides 49
C4[0 0 1]
Mao,131143 and others. Halasyamani et al. have synthesized
a variety of new polar oxide materials that contain SOJT-
distorted cations octahedrally coordinated d0 transition
metals and/or lone-pair cations. Table 2 lists these polar ox-
ides along with their space group and crystal class. By using
SOJT cations, they were able to increase the incidence of
polarity to nearly 50% in any new material. That is, when
SOJT cations are incorporated into any new compound, the
material crystallizes in a polar space group nearly 50% of the
time. They have also demonstrated that when a d0 transition
metal oxide octahedron is linked to a cation that exhibits a
lone pair, the d0 cation is displaced away from the oxide ligand
that links the two polyhedra. Thus, the lone-pair polyhedra
serve to reinforce the SOJT distortion of the d0 cation.112
Additionally, with the octahedrally coordinated d0 cations, a
continuous symmetry measure approach144146 has been used
to quantify the magnitude and direction of the distortion.147
He was able to divide the d0 transition metals into three
categories: strong (Mo6 and V5), moderate (W6, Ti4,
Nb5, and Ta5), and weak (Zr4 and Hf4) distorters (see
Figure 14). The magnitude of the distortion scales with the
electronegativity of the cation,148 that is, the more electroneg-
ative the cation, the stronger the distortion (on average).
In addition, the preferred direction of the distortion for each
d0 cation was examined and discussed. With respect to direc-
tional preferences, for V5 distortions toward a vertex or edge
are common. Interestingly, for V5 face-directed distortions
are never observed. For Mo6 and Hf4, only edge- and face-
directed distortions are observed, whereas with the other four
cations, W6, Ti4, Nb5, and Ta5, the three directions, ver-
tex, edge, and face, are observed in similar proportions. Mao
et al. have exploited SOJT distortions in the iodate polyhe-
dron, (IO3), in the synthesis of several polar oxides
(see Table 2). In addition to the polar iodate group, these
polar materials contain an octahedrally coordinated d0
metal, Nb5BaNbO(IO3)5,138 V5NaVO2(IO3)2(H2O) and
LaVO2(IO3)4
Pb2Ln3Pb3(IO3)13(m3-O) (Ln LaNd),137 or neither, K2M
(IO3)4(H2O)2 (M Mn, Co, Zn, and Mg) and BaPd
(IO3)4.150,151 Iodates have also been used by Albrecht-Schmitt
et al. to synthesize polar materials, including some new polar
uranium compounds.152158 Norquist et al. have developed a
novel strategy to design and synthesize new polar oxides by
using an SOJT distorted cation, Mo6, in combination with
chiral organic amines.159 As previously stated, the first chal-
lenge in designing polar oxides is to use inherently polar
50 Chemistry of Polar Transition-Metal Oxides

Table 2 Recently synthesized polar oxide materials containing

second-order JahnTeller distorted cations

Compound Space Crystal Reference

group class

Bi2TeO5 Abm2 mm2 113,238

TeSeO4 I2 2 114,239
Na2TeW2O9 Ia m 115
BaTeM2O9 (M Mo6 or W6) P21 2 117
A2TeW3O12 (A Rb or Cs) P31c 3m 116
Cs2I4O11 P63 6 118
Ba6M2M8O30 (M Ti4, Zr4, or P4bm 4mm 54
Hf4; M Nb5 or Ta5)
Sr6Ti2Ta8O30 Pba2 mm2 54
Pb6Ti2Nb8O30 Cm2m m2m 54
Na2Te3Mo3O16 I2 2 119 Figure 13 ORTEP (50% probability ellipsoids) diagram for lone-pair
TlSeVO5, TlTeVO5 Pna21 mm2 121 cation polyhedra. Reproduced by permission of The Royal Society of
(NH4)2Te2WO8 P21 2 120,123 Chemistry: http://dx.doi.org/10.1039/b110420a.
ANbW2O9 (A K or Rb) Cmm2 mm2 122
KTaW2O9 Cmm2 mm2 122
ASe2V3O12 (A Rb or Tl) P63 6 240
Li2Ti(IO3)6, Na2Ti(IO3)6 P63 6 124,125 <r2> (2)
Pb3SeO5 Cmc21 mm2 126
Rb2(MoO3)3(SeO3) P63 6 127
Mo6+
Tl2(MoO3)3(SeO3) P31c 3m 127
A3V5O14 (A K, Rb, or Tl) P31m 3m 128
0.15 V5+
Te2O(PO4)2 Cc m 129
VOSe2O5 P4cc 4mm 130
Nd4CuTe5O15Cl3 I2 2 131
Cd7(Te7O17)Cl8 Pca21 mm2 132
Ag2Mo3Te3O16 I2 2 139
H2MnMoSeO7 Pmc21 mm2 241
[R-C7H16N2][Mo8O26] P21 2 159 0.10
[S-C7H16N2][Mo8O26] P21 2 159
AMoO3(IO3) (A Rb or Cs) Pna21 mm2 140,152
A[(VO)2(IO3)3O2] (A NH4, Ima2 mm2 153
Rb, Cs) W6+
RE(MoO2)(IO3)4(OH) [RE Nd, P2 2 154 Ti4+
Nb5+
Sm, Eu] 0.05
Cs[(UO2)3(HIO6)(OH)(O) Cc m 155 Ta5+
(H2O)].1.5H2O
Na2[UO2(IO3)4(H2O)] C2 2 158
Ln3Pb3(IO3)13(m3-O) (Ln LaNd) R3c 3m 137 Zr4+ Hf4+
BaNbO(IO3)5 Cc m 138
NaVO2(IO3)2(H2O) P21 2 143
LaVO2(IO3)4.H2O P21 2 149 0.00
K2M(IO3)4(H2O)2 (M Mn, Co, I2 2 150 Figure 14 Average magnitude of the off-center distortion for individual
Zn, Mg) octahedrally coordinated d0 transition metal cation. Reproduced with
BaPd(IO3)4 P1 1 151 permission from Ok, K. M.; Halasyamani, P. S.; Casanova, D.; Llunell, M.;
Cs6[(UO2)4(W5O21)(OH)2(H2O)2] I4cm 4mm 156 Alemany, P.; Alvarez, S. Chem. Mater. 2006, 18, 3176. Copyright 2006
Cs2[UO2(VO2)2(PO4)2].0.59H2O Cmc21 mm2 157 American Chemical Society.

polyhedra. In using the SOJT-distorted Mo6 cation, Norquist
et al. synthesized materials where the d0 cation is substantially
displaced from the center of its oxide octahedra. Thus, each of
the MoO6 octahedra is inherently polar. The second challenge,
ensuring that the octahedral are not related by inversion cen-
ters, was successfully addressed by using single enantiomer-
templating agents. An example of this strategy involves the
synthesis of [(R)-C7H16N2][Mo8O26] and [(S)-C7H16N2]
[Mo8O26].159 These materials were synthesized as pure polar
compounds by using reaction gels in which a single enantio-
mer of 3-aminoquinuclidine was present, either as the
[(R)-C7H16N2]2 or as [(S)-C7H16N2]2 cation. By using single
enantiomer species, the cancellation of any local d0 cation
distortion through extra-framework inversion centers is pro-
hibited, since the structure would need to contain both S and
R cations. Norquist et al. are able to control the presence or
absence of each enantiomer.
2.02.4.1 Metal oxyfluoride systems
Poeppelmeier et al. have developed a strategy for designing
and synthesizing polar materials using octahedrally

O2
1.697(2)
F1 2.059(1) Mo F3
F4 1.709(2)
2.114(1)
F2
O1
1.760(1)
Nb
F4
F1
F3
F2
2.072(1)
F5
Figure 15 ORTEP (50% probability ellipsoids) for [MoO2F4]2, [VOF5]2, [NbOF5]2, and [TaOF5]2 octahedra. Note that the cation is distorted
toward the oxygen ligand(s). Reproduced with permission from Welk, M. E.; Norquist, A. J.; Stern, C. L.; Poeppelmeier, K. R. Inorg. Chem. 2000, 39,
3946. Copyright 2000 American Chemical Society.
coordinated d0 transition metal oxidefluoride anions (see
Chapter 2.07). Specifically, these are anions of the type
[MOxF6x]2 (x 1, M V5, Nb5, Ta5; x 2, M Mo6,
W6) (see Figure 15).160170 Similar to the pure d0 oxide
systems discussed earlier, the metal cation in the center of the
oxyfluoride octahedra spontaneously displaces toward a corner
(x 1) or edge (x 2) to form short MO bonds. This sponta-
neous distortion is inherent to the oxyfluoride anion and is a
result of metaldp oxygenpp orbital interactions. For exam-
ple, in anions such as [NbOF5]2 the Nb5 cation distorts
toward the oxide ligand, resulting in a short NbO bond and
a long trans-NbF bond. This distortion is analogous to those
observed in KTiOPO4 (KTP)14 and BaTiO3.6 There are also two
challenges in synthesizing NCS polar materials based on
[MOF5]2 (M V5, Nb5, Ta5) or [MO2F4]2 (M Mo6,
W6) anionic octahedra. The first is to prevent crystallographic
disorder between the oxide and fluoride ligands. Crystallo-
graphic disorder between the oxide and fluoride ligands can
impose a center of symmetry on the d0 cation, rendering the
structure centrosymmetric. The second challenge is to have
these crystallographically ordered anions arranged in an NCS
manner with respect to each other. By closely examining all of
the intra-octahedral distortions, Poeppelmeier et al. were able
to overcome the first challenge and successfully order the oxide
and fluoride ligands.163,166168 The second challenge was ac-
complished by using the [NbOF5]2 anions in a hybrid
inorganicorganic compound as well as more recently in a
solid-state material.170 The researchers were able to success-
fully meet both challenges by not only examining the primary
Chemistry of Polar Transition-Metal Oxides 51
O1
1.596(2)
F3 V F5
F2
O1 F4
2.084(2)
F1
O1
1.782(2)
Ta
F4
F1
F3
F2
2.063(2)
F5
distortion, the spontaneous displacement of the d0 transition
metal toward the oxide ligand, but also focusing on the sec-
ondary distortion, the interaction of the ordered [MOxF6x]2
anion with the extended crystal structure. With all of the or-
dered anions, an uneven amount of residual negative charge is
observed on the ligands. They demonstrated that coordination
within the structure is directed by the most anionic ligands.
With the [NbOF5]2 and [TaOF5]2 anion, the most negative
charges are found on the oxide and trans-fluoride ligands; thus,
coordination is directed in a trans-fashion. Interestingly, this
type of trans-directing is also observed in the [WO2F4]2 anion;
however, both the [MoO2F4]2 and [VOF5]2 anions are cis-
directors.166 By investigating and understanding these direc-
tional effects, both at the local, primary level and at the more
macroscopic, secondary level, Poeppelmeier et al. have been
able to design and synthesize NCS polar materials by aligning
the ordered [MOxF6x]2 anions in an acentric polar manner.
2.02.4.2 Salt-Inclusion Solids
Another novel strategy for synthesizing and designing NCS
polar materials has been described by Hwu et al.171174 They
have focused on salt inclusion solids whose framework con-
sists of mixed ionic and covalent sublattices. They have utilized
a combination of acentric polar polyanions, such as those
found in silicates, along with first-order JahnTeller cations,
along with chlorine-centered acentric secondary building units
(SBUs). We will describe each of these groups in more detail.
52 Chemistry of Polar Transition-Metal Oxides

With the acentric polar polyanions, moieties such as P2O74, 2.02.4.3 Borates
Si2O76, and V2O74 are used. Acting in a cooperative manner
In addition to the aforementioned strategies and materials,
with these anions are the first-order JahnTeller-distorted
polar borates have received a great deal of attention.175177
cations, for example, Mn3 (d4) and Cu2 (d9). Attributable
The fundamental idea behind using borates to design and
to the first-order JahnTeller effect, the coordination of these
synthesize polar materials is the inclusion of the [BO3]3
cations is inherently asymmetric. One of the most novel
anion group in the structure. It has been shown that this
features of Hwus strategy is the use of anion-based acentric
group is most often observed with 1 (C1) site symmetry.178
SBUs. Specifically, the acentric SBU ClA6nMn (A Cs, Ba;
In addition, delocalized p-type bonds are observed perpendic-
M Mn, Cu; n 1, 2) is utilized (see Figure 16). This SBU has
ular to the BO3 plane that, when coupled with MO6 (M d0
a templating effect on the framework, resulting in the observed
transition metal) octahedral, often result in large nonlinear
NCS structure. Hwu et al. have successfully used these ideas to
susceptibilities. As with the other systems discussed, the ori-
synthesize a variety of NCS salt-inclusion solids including
entation of the [BO3]3 anion and their density in the unit
Ba2Mn(Si2O7)Cl, Cs2Cu7(P2O7)4 6CsCl, Cs2Cu5(P2O7)3 3CsCl,
cell profoundly influences the NLO properties. Large non-
Ba6Mn4Si12O34Cl, and Ba6Fe5Si11O34 Cl3.171174
linear susceptibilities are thought to occur when a large num-
ber of these borate group are aligned in a crystal structure.176
(a) (b) (c) (d)
The first polar borate to gain widespread attention was b-
(i) BaB2O4 (BBO).179 BBO is found in both the centrosymmet-
ric, R-3c, high-temperature a-phase, and the NCS polar, R3c,
low-temperature b-phase. A reversible phase transition is ob-
served around 925  C. In most instances, b-BaB2O4 crystals
Mn
o
are grown by the top-seeded solution growth (tssg) method
CI
below 925  C that suppresses the formation of the centro-
Ba
symmetric a-phase. In BBO, the BO3 groups form a B3O6 ring
whose plane is perpendicular to the polar axis. The rings
(ii) themselves are slightly misaligned reducing their maximum
theoretically possible NLO susceptibility. In addition, BBO
has an exceptional optical transmission range, 1903500 nm,
as well as a high damage threshold, 5 GW cm2. In addition
c
to BBO, LiB3O5 (LBO),180 CsB3O5 (CBO),181 CsLiB6O10
a b (CLBO),182 and BiB3O6 (BIBO)183 are attractive materials
for efficient SHG; however, only BBO, LBO, and BIBO are
polar. Keszler et al. have also synthesized a host of borate
Mn materials,175,177,184 some of which are polar. The polar borates
CI
include pyroborates such as CsMO(B2O5) (M Nb5 or
Ba
Ta5)185 and BaCuB2O5,186 and LiNaB4O7187 that contain
o B4O9 groups. Mao et al. and Pan et al. have also used borate
(iii) groups to design and synthesize new polar oxide materials.
Recently, Mao et al. have used cyclic B3O8 groups to create
c new polar oxides, namely, RbGeB3O7, Rb2GeB4O9, and
a b Rb4Ge3B6O17.188 Pan et al. have synthesized polar borates
that incorporate corner-shared BO3 groups, that is, B2O5 di-
mers. Specifically, they have grown large crystals of two con-
gruently melting polar borates, Li6CuB4O10189 and
(a) (b) (c) Bi2ZnOB2O6.190193 In addition to being highly stable, both
materials have large transmission windows. Pan et al. have also
synthesized other polar borates, namely, Na3La2(BO3)3,194
Na5B2P3O13,195 Bi2Cu5B4O14,196 Pb2[B5O9]Br,197
198 199
Na2B5O8(OH), and Na2B4O12H10. The polar borates are
Figure 16 Figure depicting (i) coordination surrounding chlorine in (a)
listed in Table 3.
NaCl, (b) Cs2Cu7(P2O7)46CsCl, (c) Ba2Mn(Si2O7)Cl, and (d)
Ba6Mn4Si12O34Cl3; (ii) a slab of the (Ba2Mn)Cl lattice in Ba2MnSi2O7Cl
showing the origin of the polar lattice alternating short and long MnCl 2.02.5 Functional Properties Associated with Polar
linkages; (iii) a cage view of the Si2O7 unit residing in the center of the
Oxide Materials
anti-ReO3 type (Ba2Mn)Cl lattice in Ba2MnSi2O7Cl (left) with the Si2O7
unit comprising two corner-shared SiO4 tetrahedra that are eclipsed (right);
(iv) coordination around chlorine in (a) NaCl, (b) Cs2Cu7(P2O7)46CsCl, In Section 2.02.1, we briefly discussed some of the properties
and (c) Ba2MnSi2O7Cl. Reproduced with permission from Mo, X.; associated with polar oxide materials. In this section, we de-
Hwu, S.-J. Inorg. Chem. 2003, 42, 3978; Mo, X.; Ferguson, E.; Hwu, scribe these properties in more detail, as well as discuss their
S.-J. Inorg. Chem. 2005, 44, 3121. Copyright 2003 and 2005 measurements. We will focus on the characterization of bulk
American Chemical Society. materials, as opposed to thin films or single crystals, since with
 Chemistry of Polar Transition-Metal Oxides 53

Table 3 Polar borates Curie.206 They observed that some materials become elec-
trically polarized when subjected to a mechanical force.
Compound Space Crystal Reference Soon after, the converse effect was discovered wherein the
group class
application of a voltage resulted in a macroscopic strain. In
b-BaB2O4 (BBO) R3c 3m 179 1910, Voigt published a standard reference detailing the
LiB3O5 (LBO) Pna21 mm2 180 electromechanical relationships in piezoelectric mate-
BiB3O6 (BIBO) C2 2 183 rials.207 A thorough review of the early history of piezoelec-
Li2B4O7 I41cd 4mm 242 tricity can be found in Cadys seminal book.200 Thus with
A2B4O7 (A Sr2 or Pb2) Pmn21 mm2 243 piezoelectricity, two effects are possible: direct and con-
ACa4(BO3)3O (A Y3 or Cm m 244,245 verse. Both direct and converse effects are used in a variety
Gd3) of applications. The direct effect results in generator action
NaBe2BO3F2 C2 2 246,247
the piezoelectric material converts mechanical energy to
KNbOB2O5 Pbn21 mm2 248
electrical energy. This generator action is used in solid-state
CsMO(B2O5) (M Nb5 or Pmn21 mm2 185
Ta5) batteries, sensing devices, and fuel lighting applications.
BaCuB2O5 C2 2 186 The converse effect results in motor action the piezoelec-
LiNaB4O7 Fdd2 mm2 187 tric material converts electrical energy to mechanical en-
RbGeB3O7 Pna21 mm2 188 ergy. This motor action is used in ultrasonic and acoustic
Rb2GeB4O9 P21 2 188 applications, micromotor devices, and electromechanical
Rb4Ge3B6O17 Cc m 188 transducers. Measurements on bulk materials utilizing
Li6CuB4O10 P1 1 189 both direct and converse piezoelectric techniques will be
Bi2ZnOB2O6 Pba2 mm2 190193 described.
Na3La2(BO3)3 Amm2 mm2 194
Na5B2P3O13 C2 2 195
Pyroelectricity
The pyroelectric effect may be defined as the change in
Bi2Cu5B4O14 P1 1 196
Pb2[B5O9]Br Pnn2 mm2 197 spontaneous polarization, Ps, as a function of tempera-
Na2B5O8(OH) Pna21 mm2 198 ture.208 The symmetry requirements for pyroelectricity are
Na2B4O12H10 Pna21 mm2 199 far more restrictive compared with SHG and piezoelectric-
ity. To exhibit a spontaneous polarization, the material in
question must crystallize in one of ten polar crystal classes
(1, 2, 3, 4, 6, m, mm2, 3m, 4mm, or 6mm). Thus, polarity
is required for pyroelectric behavior. Determining the py-
the latter large (>5 mm) crystals are necessary. The techniques
roelectric coefficient may be done two ways measuring
described can, however, be used on single crystals. In addition
either the pyroelectric current or the pyroelectric charge.209
to having large single crystals, in many cases these crystals
Both techniques will be described.
must be cut and polished to expose specific crystallographic
faces. With new and even well-known materials, growing, cut- Ferroelectricity
A ferroelectric is formally defined as a pyroelectric material
ting, and polishing such crystals are exceptionally difficult
that has a reversible, or switchable, polarization.3 Thus
and remain an ongoing challenge. In this section, the charac-
for a material to be ferroelectric, the compound must
terization of second-harmonic generating, piezoelectric, pyro-
be polar, that is, it must possess a permanent dipole mo-
electric, and ferroelectric properties in bulk NCS materials is
ment, and must be capable of having this moment reversed
described. These phenomena have been discussed extensively
in the presence of an applied voltage. The former occurs if
in the literature2,3,200203 so only a brief description of each
the material crystallizes in one of ten polar crystal classes
phenomena will be given here.
(1, 2, 3, 4, 6, m, mm2, 3m, 4mm, or 6mm). Determining
Second-harmonic generation the latter is more involved. Polarization reversal, or ferro-
SHG, or frequency doubling, is defined as the conver- electric hysteresis, may be measured through a Sawyer
sion of a specific wavelength of light into half its original, Tower circuit.210 Additionally, because of the relatively
that is, l1 ! (1/2)l1, or with respect to frequency o, large voltages needed for polarization reversal, the material
o1 ! 2o1. The first report of SHG was by Franken et al. in under investigation must be insulating. Another feature that
1961,204 who reported SHG on a crystal of a-SiO2 using a is observed in some, but not all, ferroelectric materials is a
ruby laser. Following this experimental result, a classic dielectric anomaly at the Curie temperature. A maximum
paper by Bloembergen and coworkers was published that in the dielectric constant is often observed at the Curie
provided a theoretical foundation for the origin of the NLO temperature. This temperature indicates a phase change
susceptibility.205 For several years following Frankens ex- to a centrosymmetric nonpolar, that is, nonferroelectric,
perimental result, large single crystals were required to often termed paraelectric, structure. We will describe mea-
measure SHG. Kurtz and Perry published, in 1968, a sem- surement techniques that allow one to determine ferroelec-
inal paper that described a technique whereby SHG could tric hysteresis curves.
be measured from polycrystalline samples.201 It is this tech-
nique that we will describe in more detail. This section is divided into four parts. Each part describes a
Piezoelectricity specific NCS property, the history of the phenomena, and
Piezoelectricity, derived from the Greek piezen, meaning provides details on the measurement as well as an interpreta-
to press, was discovered in 1880 by Jacques and Pierre tion of the resulting data.
54 Chemistry of Polar Transition-Metal Oxides
Light box
Beamsplitter
Nd-YAG laser
1064 nm
Oscilloscope
Figure 17 Schematic diagram of a modified Kurtz-powder laser second-harmonic generation measurement system. Reproduced by permission of The
Royal Society of Chemistry: http://dx.doi.org/10.1039/b110420a.
2.02.5.1 Second-Harmonic Generation
SHG, or frequency doubling, is defined as the conversion of a
specific wavelength of light into half its original, that is,
l1 ! (1/2)l1, or with respect to frequency o, o1 ! 2o1 (see
Chapter 8.20). It was not until the invention of the laser in
1960 by Maiman211 that sizeable NLO effects, such as SHG,
could be observed. Attributable to these optical fields, the
induced polarization, P, in the material can be written as a
power series: P wE dE2
where w is the linear electric
susceptibility, with the higher-order terms resulting in non-
linear effects such as SHG. These nonlinear effects are described
by expanding the polarization:
Pi wijk Ej wijk Ej Ek wijkl Ej Ek El
where wij is the electric susceptibility, with the second-order
nonlinear coefficient described as wijk. Third-order terms, wijkl,
give rise to third-harmonic generation. Only in an NCS envi-
ronment is wijk 6 0. Mathematically, wijk is a third-rank tensor,
and in experimental SHG measurements is replaced by dijk,
where 2dijk wijk.
After the discovery of SHG in 1961,204 large crystals, on the
order of several mm, were required to investigate the phenom-
enon. A technique, described in 1968, allowed the determina-
tion of the SHG efficiency on polycrystalline samples. It is this
technique that is described in more detail. At its most basic, the
powder SHG technique requires very little instrumentation.
Additionally, all of the instrumentation is commercially avail-
able. A typical setup for powder SHG measurements is shown
schematically in Figure 17. A low-energy laser, pulsed or con-
tinuous, is needed. Usually, a commercially available NdYAG
laser (1064 nm output) is used,212 since any SHG will appear
in the visible at 532 nm (green), and thus the experimentalist is
literally able to see the SHG. The sample, a polycrystalline
powder, is placed in a fused silica tube a capillary tube
or nuclear magnetic resonance (NMR) tube can be used. For
Sample
holder
Focusing mirror
532 nm narrow
bandpass filter
Photomultiplier tube
Trigger
a quick and dirty measurement that addresses the simple
question of whether the material is SHG active or not, only
1050 mg of sample is required. If more quantitative SHG
information is needed, a larger amount of sample, around
1 g, is required. For more quantitative measurements, a photo-
multiplier tube (PMT) connected to an oscilloscope is also
necessary. Using the PMT and oscilloscope allows the user to
collect SHG data on a standard, usually a-SiO2 or urea, and
compare the results with newly synthesized materials. The
entire system laser, PMT, power supplies, optics, and oscillo-
scope must be placed on a flat surface. A laser table is ideal,
since the various optical pieces can be attached to the table, but
is not required. The total footprint of the system is 20  60 ; thus,
only a relatively small flat area is needed.
2.02.5.1.1 Measurement of SHG and data interpretation
Once the laser, PMT, and optics are aligned, collecting the SHG
is reasonably straightforward. As previously stated, a small
amount of the powder to be tested (1050 mg) is placed in a
fused silica tube. The SHG efficiency of the new material may be
compared with standard materials. SHG properties were first
measured on a-SiO2; thus, the material is defined to have an
efficiency of 1.0 (dimensionless). Most SHG efficiencies
are reported with respect to a-SiO2; however, other materials
can be used. BaTiO3 and urea both have SHG efficiencies
of 400  SiO2, whereas LiNbO3 has an efficiency of
600  SiO2.201,213 Once the SHG of a polycrystalline sample
of a-SiO2 has been measured, it is very straightforward to
roughly determine the SHG efficiency of any new material. If
more accurate SHG information is required, additional experi-
ments become necessary. Typically, a larger amount of material
is necessary, on the order of 1 g, and the powder needs to be
sieved into particle sizes ranging between 20 and 120 mm. This
sieving may be done using commercially available sieves. Mea-
suring the SHG as a function of particle size, 20120 mm, has
two advantages. First, the SHG efficiency is determined more
accurately. Second, type 1 phase-matching information may be

determined.201 Type 1 phase matching, or index matching,
occurs when the phase velocity of the fundamental radiation
(1064 nm) equals the second harmonic (532 nm). If type 1
phase matching occurs, the SHG efficiency will increase with
the particle size and plateau at a maximum value. If phase
matching does not occur, the SHG efficiency will reach a max-
imum value and then decrease, as the particle size increases.
Phase-matching (LiNbO3) and nonphase-matching (a-SiO2)
curves are shown in the diagrams below. Note that the curves
are drawn to guide the eye, and are not a fit to the data.
These figures also clearly illustrate that accurate SHG effi-
ciencies can only be determined by measuring similar particle
size ranges. For example, if the SHG of a-SiO2 is measured with
particles >90 mm and the unknown material is measured at a
smaller particle size, the SHG efficiency of the unknown mate-
rial would be overestimated. Thus, it is critical that the SHG of
SiO2 and the unknown material be measured at the sample
particle size range, that is, 4563 mm. Once the phase-matching
capabilities of the material are known, the bulk NLO suscepti-
bility, hdeffi, can be estimated. The value of hdeffi for phase-
matchable, PM, and non-phase-matchable, NPM, materials are:
 1=2
I2o A
hdeff iPM  7:98  102
I2o LiNbO 3
 1=2
I2o A
hdeff iNPM  0:3048
I SiO2
2o
The SHG efficiency of the unknown compound (A) is either
compared with LiNbO3SHG efficiency of 600  SiO2 or SiO2
depending on the phase-matching behavior of the unknown
compound. The unit for hdeffi is picometer per volt (pm V1).
2.02.5.2 Piezoelectricity
The piezoelectric phenomena occur as both the direct and
converse effect.200 With the direct effect, an external stress,
sjk, results in a change in polarization, Pi. The direct effect is
mathematically described as Pi dijksjk, where dijk (i, j, k 1,
2, 3) is the piezoelectric charge coefficient, given in Coulombs
per Newton (C N1). With the converse effect, an applied field,
Ei, results in a strain, ejk. The converse effect is mathematically
described as ejk dijkEi, where dijk is the piezoelectric strain
coefficient, given in meter per volt (m V1). With both effects,
dijk is a third-rank tensor. It is important to note that the units
for dijk, when measuring direct or converse effects, are
1.6
1.5
LiNbO3
SHG (arbitary units)
1.4
1.3
1.2
1.1
1.0
0.9
20 40 60 80
Particle size (mm)
Phase-matchable
Chemistry of Polar Transition-Metal Oxides 55
equivalent, that is, 1 C N1 1 m V1. Often the piezoelectric
equation is written as Pi dijsj (i 1, 2, 3; j 1, 2, . . . 6), where
dij is the contracted notation for dijk.214 It is important to note
that dij does not transform as a second-rank tensor. The piezo-
electric constants, both charge and strain, given as dij, are
usually reported as one or more terms, d33, d31, and/or d15.
With d33 the induced polarization, strain, is parallel to the
applied stress, electric field, whereas with d31 and d15 the
induced polarizations, strains, are perpendicular to the applied
stresses, electric fields. Another important variable with respect
to piezoelectric devices is the electromechanical coupling fac-
tor, k. This factor describes the efficiency in the conversion of
mechanical energy to electrical energy, the direct effect, or the
conversion of electrical energy to mechanical energy, the con-
verse effect. Generally, large k values are desirable for efficient
energy conversion; however, the coupling term does not take
into account dielectric or mechanical losses.
2.02.5.2.1 Sample preparation and measurements
Unlike the SHG measurement where a loose polycrystalline
powder can be used, a well-sintered ceramic is necessary for
bulk piezoelectric measurements. For the measurements de-
scribed herein, the dimensions of the sintered disk are a diam-
eter of 0.5 in. and a thickness of 0.5 mm. In addition, the
ceramic must undergo electrical poling. With the poling tech-
nique, electrodes are applied to both sides of the sample
usually silver or gold has been sputtered or mechanically ap-
plied. The poling process takes place above room temperature
(100300  C), with an applied voltage (10002000 V) for
2030 min. After poling, the material has a response similar
to a single crystal, where the entire ceramic acts as a single unit.
Although poling will not align 100% of the crystallites, the
extent of alignment is sufficient to measure the piezoelectric
response. It is interesting to note that it was only in 1949 that
poling was discovered to be critical in changing a seemingly
inert ceramic into an electromechanically active material.215
Before this time, the assumption was that the individual crys-
tallites in a ceramic would effectively cancel, rendering the
material useless for industrial applications.
2.02.5.2.2 Direct piezoelectric measurements
The direct piezoelectric effect occurs when a mechanical force
on a material results in a change in polarization. The resultant
piezoelectric charge constant, d, is a third-rank tensor, dijk, and
is measured in C N1, or pC N1. Often this tensor notation is
reduced to matrix notation,214 and the dijk terms become dij,
0.15
0.14 SiO2
SHG (arbitary units)
0.13
0.12
0.11
0.10
0.09
0.08
0.07
100 125 20 40 60 80 100 125
Particle size (mm)
Nonphase-matchable
56 Chemistry of Polar Transition-Metal Oxides
High voltage amplifier
High voltage interface
Drive Return
Computer
Figure 18 Experimental system to measure converse piezoelectric effects. Reproduced by permission of The Royal Society of Chemistry: http://dx.doi.
org/10.1039/b110420a.
-55.55
-55.60
-55.65
Displacement (mm)
-55.70
-55.75
-55.80
-55.85
-55.90
-55.95
-800 -600 -400 -200 0 200 400
E (V)
Figure 19 Representative butterfly loop observed in converse
piezoelectric measurements.
with i 1, 2, 3; j 1, 2, . . ., 6. As stated earlier, dij does not
transform as a second-rank tensor. For both the direct as well as
the converse effects, the d33 value of the material is usually
reported. The subscripts specifically denote that what is being
measured is a polarization parallel to the direction of the
mechanical force. Lateral, d31, and shear, d15, polarizations
may also be determined, but these measurements are usually
done on specifically cut single crystals. To measure the direct
piezoelectric effect, a static or a quasistatic method is used.
Although this method is less precise than the resonance
method,216 the ease of use and availability of instrumentation
make the static method preferable. The static method employs
a Berlincourt d33 meter, for which a number of commercial
systems are available.217 The instrument is very straightforward
to use. A known force is applied to the poled ceramic, as well as
to a standard piezoelectric. Comparing the resultant electric
signals allows one to determine the d33 of the sample. The
value of d33 is simply read off the meter. These meters can
measure d33 charge constants within a few percent, with a
range from 20 to 2000 pC N1.
Optical sensor
Optical
fiber probe
Micropositioner
Bulk sample
Copper plate
Insulating acrylic
2.02.5.2.3 Converse piezoelectric measurements
The d33 strain constant may be determined on bulk samples
through converse piezoelectric measurements. As noted earlier,
the strain constant is expressed in units of m V1 that are
equivalent to C N1. Converse piezoelectric measurements
are more experimentally difficult compared with the direct
measurements, but provide greater accuracy. The converse
measurements use an optical technique in order to measure
the small strains in the sample caused by the application of
a voltage. The experimental system is composed of a high-
voltage amplifier and interface as well as an optical sensor
and probe. The optical probe remains stationary and is approx-
imately 1 mm above the sample (see Figure 18).
In the native state, or zero voltage, a finite amount of light is
collected in reflection from the sample. When the voltage is
600 800 applied, the sample undergoes a macroscopic strain and the
amount of light collected by the optical sensor changes. The
change in collected light is converted to a displacement change,
in m V1. Mathematically, d33 may then be calculated through:
d33 S/E Dt/V, where S is the strain, E is the field strength
(in V m1), Dt is the change in thickness, and V is the applied
voltage. A plot of strain versus electric field produces the com-
monly observed butterfly curves, similar to the one shown in
Figure 19. The equation also indicates that d33 depends on the
change in thickness of the sample, not the initial sample thick-
ness. The magnitude of the piezoelectric response can vary
greatly among oxide materials. For example, ZnO, LiNbO3,
and LiTaO3 have d33 values of around 10 pC N1, whereas
BaTiO3 and lead zirconate titanate (PZT) have corresponding
d33 coefficients of approximately 190 and 100600 pC N1 for
various PZT compositions.209
2.02.5.3 Pyroelectricity
The pyroelectric effect is defined as the change in spontaneous
polarization, Ps, as a function of temperature. The pyroelectric
coefficient, p, is mathematically defined as
dPs
p
dT

with spontaneous polarization, Ps, and temperature, T. The
coefficient, p, is a vector and is described in units of
mC m2 K1. Surprisingly, the effect has been known for over
2400 years, with the first account attributed to the Greek phi-
losopher Theophrastus. He described a stone, lyngourion
probably tourmaline that was capable of attracting straw
and pieces of wood. A thorough and comprehensive descrip-
tion and history of the phenomenon may be found in Langs
seminal text and recent papers.1,2,208 Brewster, in 1824, was the
first researcher to use the term pyroelectricity.218 Interestingly,
the material Brewster investigated, Rochelle salt, was studied
nearly a century later by Valasek in his discovery of ferroelec-
tricity.219,220 The pyroelectric effect was mainly an academic
curiosity until 1938, when Y. Ta suggested that tourmaline
crystals could be used as an infrared sensor.221 After this pub-
lication, and with the onset of the Second World War, investi-
gation into pyroelectricity grew rapidly and remains an active
area of current research among chemists, material scientists,
and engineers. Currently, pyroelectrics are mainly used for
thermal detectors. Pyroelectric devices respond to changes in
temperature and therefore can be used to detect and observe
stationary or moving objects. A few of the applications for
pyroelectric detectors include burglar alarms, pollution moni-
tors, and the measurement of thermal properties of materials.
2.02.5.3.1 Sample preparation and measurement
The sample preparation for a bulk pyroelectric measurement
is very similar to what is required for a bulk piezoelectric
measurement. A well-sintered ceramic disk that has been elec-
trically poled is required. Determining the pyroelectric coeffi-
cient may be divided into two groups the measurement of
the pyroelectric current and the measurement of the charge.209
We will describe measurement techniques for both groups. In
addition, the pyroelectric effect can be subdivided into primary
and secondary effects. The primary effect is observed when the
material is rigidly clamped under a constant strain to prevent
any thermal expansion or contraction. Secondary effects occur
when the material is permitted to deform, that is, the material
is under constant stress. Thermal expansion results in a strain
that changes the spontaneous polarization, attributable to the
piezoelectric effect. Thus, the secondary pyroelectric effect in-
cludes contributions caused by piezoelectricity. Exclusively
measuring the pyroelectric coefficient under constant strain is
Ground
High voltage amplifier
Drive Return
Tc
P
High voltage interface
Temp (c)
Computer
Precision tester
General purpose interface bus
(GPIB, optional)
Figure 20 Schematic of the static method for determining the pyroelectric coefficient. Reproduced by permission of The Royal Society of Chemistry:
http://dx.doi.org/10.1039/b110420a.
Chemistry of Polar Transition-Metal Oxides 57
experimentally very difficult. What is usually experimentally
measured is the total pyroelectric effect exhibited by the mate-
rial the sum of the primary and the secondary effects.
2.02.5.3.2 Pyroelectric current
The most straightforward technique to measure the pyroelec-
tric current is the direct method.222 With the direct method, the
pyroelectric material is heated uniformly at a constant rate, that
is, DT/Dt 12  C min1. The pyroelectric coefficient is deter-
mined by measuring the pyroelectric current, given by i(T ) A
(DT/Dt)p, where A is the sample area. Thus, a plot of p(T ) over
a wide range of temperature can be easily obtained. More
experimentally, complicated methods may also be used to
determine the pyroelectric current. These include radiation
heating223 and capacitive charging.224 These methods are
more accurate compared with the direct method, but experi-
mentally more complicated.
2.02.5.3.3 Pyroelectric charge
The original method for measuring the pyroelectric charge was
developed in 1915.225 This technique, known as the static
method, determines the charge, that is, polarization, of the
material as a function of temperature. The technique works
very well at discrete temperatures. The static method was im-
proved upon by Glass, with an integration technique that
allowed for larger changes in temperature.226 The pyroelectric
coefficient may be obtained by graphical differentiation.
A schematic of the static method is shown in Figure 20. If
the material under investigation is ferroelectric, that is, the
polarization is reversible, the pyroelectric coefficient may be
determined by measuring the temperature dependence of
the remanent polarization. The remanent polarization is deter-
mined through a SawyerTower loop (see Section 2.02.4.4).210
As with the piezoelectric technique discussed earlier, graphical
differentiation is used.
Clearly, ferroelectric and nonferroelectric pyroelectrics are
possible, and the pyroelectric coefficient varies widely between
the two groups. The pyroelectric coefficients for ferroelectrics
such as BaTiO3, LiNbO3, and LiTaO3 are 200, 83, and
176 mC m2 K1 respectively, whereas for nonferroelectrics
such as ZnO, tourmaline, and CdS the corresponding values
are 9.4, 4.0, and 4.0 mC m2 K1, respectively.2
Temperature
controller
Furnace
58 Chemistry of Polar Transition-Metal Oxides

Ground
Computer
High voltage amplifier
Copper bar

High voltage interface Bulk sample

Precision LC
Copper plate
Drive Return
High voltage test fixture

Figure 21 Experimental Sawyer-Tower circuit for measuring Ferroelectric hysteresis. Reproduced by permission of The Royal Society of Chemistry:
http://dx.doi.org/10.1039/b110420a.

2.02.5.4 Ferroelectricity
A ferroelectric is formally defined as a pyroelectric material that
has a reversible, or switchable, polarization. Ferroelectricity +P (mC cm-2)
was discovered in ca. 1920 by Valasek219,220 in Rochelle salt
(NaKC4H4O6 4H2O) a material that was known at the time +Ps
for its piezoelectric and pyroelectric properties. For years after
+Pr
this discovery, ferroelectricity was viewed simply as a scientific
curiosity, and was thought to occur only rarely in materials. In
1935, the first family of ferroelectrics was discovered in
-E +E (V)
KH2PO4 and related materials.227,228 The first nonhydrogen-
Ec
bonded ferroelectric, BaTiO3, was subsequently discovered in
ca. 1945 by Wul and Goldman in the Soviet Union and von
Hippels group in the United States.16,17 Until this discovery it
was assumed that hydrogen bonding was necessary for ferro- -Pr
electricity to occur. The fact that oxides could exhibit ferroelec-
-Ps
tric behavior ushered in a new era, and soon thereafter a
number of ferroelectric oxides were discovered. A thorough -P
and rigorous discussion that encompasses all aspects of ferro-
electricity may be found in the comprehensive text by Lines
and Glass.3

Figure 22 Ferroelectric hysteresis loop (polarization vs. applied

2.02.5.4.1 Sample preparation and hysteresis loop
Similar to piezoelectric measurements, for ferroelectric mea-
surements a well-sintered and dense (>95%) ceramic disk that
has been electrically poled is necessary. The circuit design for
measuring ferroelectric hysteresis curves was published in
1930.210 Since that time there have been a few reports modi-
fying the original design,229,230 but the overall concept has not
changed in over 70 years. At its most basic, a linear capacitor is
placed in series with the sample. An AC or DC voltage is then
applied. The voltage measured across the capacitor is equiva-
lent to the polarization of the sample (see Figure 21). The
measurement of spontaneous polarizations on the order of
550 mC cm2 in bulk samples requires voltages in excess of
1000 V. The circuit is used to measure a ferroelectric hystersis
loop, that is, the materials switchability in the presence of
an applied voltage (see Figure 22). The hysteresis loop is
determined by measuring the polarization of the material, in
mC cm2, as a function of applied voltage, V. The full details of
a ferroelectric hystersis loop have been extensively dis-
cussed,231 so only a brief description will be given here. Several
points of the loop are of interest, the spontaneous (Ps) and
remanent (Pr) polarizations, the coercive field (Ec), and the
general shape of the loop. In a ferroelectric material, when all
voltage). Reproduced by permission of The Royal Society of Chemistry:
http://dx.doi.org/10.1039/b110420a.
of the dipole moments are aligned, the material is considered
saturated since an increase in applied voltage will not increase
the polarization. The linear extrapolation of the curve back to
the polarization axis represents the spontaneous polarization
(Ps). As the applied voltage is reduced from its maximum
positive value to zero, some dipole moments will remain
aligned, and a remanent polarization (Pr) is observed. As
the applied voltage spans the range from its maximum positive
to negative values, Ps and Pr, will be observed. Structurally
all of the dipole moments have switched from the positive to
the negative the up and down arrows in Figure 22. This is the
switchability alluded to earlier. Additional information that
can be obtained from a ferroelectric hystersis loop includes the
coercive field (Ec) and the shape of the loop. The coercive field
is the magnitude of the external applied voltage necessary to
remove all the polarization in the material. The Ec as well as the
shape of the loop, that is, the squareness or sharpness, are
sample preparation dependent and are influenced by grain size
and homogeneity.231 It is important to know that the claim of

ferroelectricity has been made for a number of materials based
on the observation of a closed hysteresis loop.78,232235 An
excellent article by J. F. Scott has recently appeared that de-
scribes this false ferroelectricity.236 In this chapter, he de-
scribes how some reported ferroelectric loops are simply from
lossy dielectrics and have nothing to do with polarization
reversal. As previously stated, ferroelectrics may be divided
into two groups, hydrogen bonding and nonhydrogen bond-
ing. The spontaneous polarization, Ps, values vary greatly be-
tween the two groups. With KH2PO4 and related materials, Ps
values range from 4.0 to 6.0 mC cm2, whereas with oxides
such as BaTiO3 and LiNbO3 the corresponding values are 26
and 71 mC cm2, respectively.3
2.02.6 Conclusion
This chapter described polar inorganic materials their struc-
ture types, design, and functional properties. As seen, polar
materials represent a vast and growing class of materials. In
addition, a number of design strategies have been put forth
that enables the synthetic chemist to create new and exciting
polar materials. The inherent polarization associated with
these materials also results in a host of technologically impor-
tant properties including ferroelectricity (reversible polariza-
tion), pyroelectricity (polarization change as a function of
temperature), and multi-ferroic behavior (simultaneous mag-
netic ordering and ferroelectric behavior). Clearly, the poten-
tial for discovering new and technologically relevant polar
materials is very bright!
Acknowledgments
The Robert A. Welch Foundation (Grant E-1457), the Texas
Center for Superconductivity, and the NSF (DMR-0652150)
are thanked for support.
References
1. Lang, S. B.; Das-Gupta, D. K. In Nalwa, H. S., Ed.; Handbook of Advanced
Electronic and Photonic Materials and Devices; Academic Press: San Francisco,
2001; Vol. 4, p 1.
2. Lang, S. B. Phys. Today 2005, 58, 31.
3. Lines, M. E.; Glass, A. M. Principles and Applications of Ferroelectrics and
Related Materials. Oxford University Press: Oxford, 1991.
4. Waser, R.; Bottger, U.; Tiedke, S. Polar Oxides: Properties, Characterization, and
Imaging. Wiley-VCH: Weinheim, 2005.
5. Hahn, T. International Tables for Crystallography. D. Reidel Publishing Company:
Dordrecht, 1983.
6. Frazer, B. C.; Danner, H. R.; Pepinsky, R. Phys. Rev. 1955, 100, 745.
7. Abrahams, S. C.; Hamilton, W. C.; Reddy, J. M. J. Phys. Chem. Solids 1966, 27,
1013.
8. Abrahams, S. C.; Levinstein, H. J.; Reddy, J. M. J. Phys. Chem. Solids 1966, 27,
1019.
9. Zaslavskii, A. I.; Tutov, A. G. Dokl. Akad. Nauk SSSR 1960, 135, 815.
10. Tomashpolskii, Y. Y.; Venevtsev, Y. N.; Zhdanov, G. S. Dokl. Akad. Nauk SSSR
1963, 153, 1313.
11. Tomashpolskii, Y. Y.; Venevtsev, Y. N.; Zhdanov, G. S. Kristallografiya 1967, 12,
252.
12. Michel, C.; Moreau, J. M.; Achenbach, G. D.; Gerson, R.; James, W. J. Solid State
Commun. 1969, 7, 701.
13. Moreau, J. M.; Michel, C.; Gerson, R.; James, W. J. J. Phys. Chem. Solids 1971,
32, 1315.
Chemistry of Polar Transition-Metal Oxides 59
14. Tordjman, I.; Masse, R.; Guitel, J. C. Z. Kristallogr. Kristallgeom. Kristallphys.
Kristallchem. 1974, 139, 103.
15. Goldschmidt, V. M. Geochem. Vert. Elem. 1927, 7, 8.
16. von Hippel, A.; Breckenridge, R. G.; Chesley, F. G.; Tisza, L. Ind. Eng. Chem.
1946, 38, 1097.
17. Wul, B. M.; Goldman, I. M. Compt. Rend. Acad. Sci. URSS 1945, 49, 139.
18. Matthias, B. T.; Wood, E. A.; Holden, A. N. Phys. Rev. 1949, 76, 175.
19. Shirane, G.; Newnham, R. E.; Pepinsky, R. Phys. Rev. 1954, 96, 581.
20. Shirane, G.; Danner, H.; Pavlovic, A.; Pepinsky, R. Phys. Rev. 1954, 93, 672.
21. Shirane, G.; Pepinsky, R.; Frazer, B. C. Phys. Rev. 1955, 97, 1179.
22. Shirane, G.; Pepinsky, R.; Frazer, B. C. Acta Cryst. 1956, 9, 131.
23. Shirane, G.; Hoshino, S. J. Phys. Soc. Jpn. 1951, 6, 265.
24. Shirane, G.; Suzuki, K.; Takeda, A. J. Phys. Soc. Jpn. 1952, 7, 12.
25. Remeika, J. P.; Glass, A. M. Mater. Res. Bull. 1970, 5, 37.
26. Aurivillius, B. Arkiv. foer Kemi. 1949, 1, 499.
27. Aurivillius, B. Arkiv. foer Kemi. 1949, 1, 463.
28. Aurivillius, B. Arkiv. foer Kemi. 1950, 2, 519.
29. Rae, A. D.; Thompson, J. G.; Withers, R. L. Acta Cryst. 1991, B47, 870.
30. Thompson, J. G.; Rae, A. D.; Withers, R. L.; Craig, D. C. Acta Cryst. 1991,
B47, 174.
31. Rae, A. D.; Thompson, J. G.; Withers, R. L.; Willis, A. C. Acta Crystallogr. 1990,
B46, 474.
32. Theobald, F. R.; Laarif, A.; Hewat, A. W. Ferroelectrics 1984, 56, 219.
33. Wolfe, R. W.; Newnahm, R. E.; Kay, M. I. Solid State Commun. 1969, 7, 1797.
34. Brown, I. D. The Chemical Bond in Inorganic Chemistry: The Bond Valence
Model, 1st ed.; Oxford University Press: Oxford, 2002.
35. Srikanth, V.; Idink, H.; White, W. B.; Subbarao, E. C.; Rajagopal, H.; Sequeira, A.
Acta Cryst. 1996, B52, 432.
36. Ismunandar, ; Hunter, B. A.; Kennedy, B. J. Solid State Ionics 1998, 112, 281.
37. Kim, J.-S.; Cheon, C.-I.; Lee, C. H.; Choo, W.-G. J. Kor. Ceram. Soc. 2001, 38, 962.
38. Shimakawa, Y.; Kubo, Y.; Nakagawa, Y.; Kamiyama, T.; Asano, H.; Izumi, F. Appl.
Phys. Lett. 1999, 74, 1904.
39. Shimakawa, Y.; Kubo, Y.; Nakagawa, Y.; Goto, S.; Kamiyama, T.; Asano, H.;
Izumi, F. Phys. Rev. B 2000, 61, 6559.
40. Shimakawa, Y.; Kubo, Y.; Tauchi, Y.; Kamiyama, T.; Asano, H.; Izumi, F. Appl.
Phys. Lett. 2000, 77, 2749.
41. Shimakawa, Y.; Imai, H.; Kimura, H.; Kimura, S.; Kubo, Y.; Nishibori, E.;
Takata, M.; Sakata, M.; Kato, K.; Hiroi, Z. Phys. Rev. B 2002, 66, 144110.
42. Borg, S.; Svensson, G. J. Solid State Chem. 2001, 156, 160.
43. Champarnaud-Mesjard, J. C.; Frit, B.; Wanatabe, A. J. Mater. Chem. 1999, 9,
1319.
44. Dorrian, J. F.; Newnham, R. E.; Smith, D. K. Ferroelectrics 1971, 3, 17.
45. Borg, S.; Svennson, G.; Bovin, J. O. J. Solid State Chem. 2002, 167, 86.
46. Cummins, S. E.; Cross, L. E. J. Appl. Phys. 1968, 39, 2268.
47. Cross, L. E.; Pohanka, R. C. Mater. Res. Bull. 1971, 6, 939.
48. Kennedy, B. J.; Zhou, Q.; Ismunandar, ; Kubota, Y.; Kato, K. J. Solid State Chem.
2008, 181, 1377.
49. Hervoches, C. H.; Snedden, A.; Riggs, R.; Kilcoyne, S. H.; Manuel, P.;
Lightfoot, P. J. Solid State Chem. 2002, 164, 280.
50. Ismunandar, ; Kamiyama, T.; Hoshikawa, A.; Zhou, Q.; Kennedy, B. J.; Kubota, Y.;
Kato, K. J. Solid State Chem. 2004, 177, 4188.
51. Jamieson, P. B.; Abrahams, S. C. Acta Cryst. 1968, B24, 984.
52. Jamieson, P. B.; Abrahams, S. C.; Bernstein, J. L. J. Chem. Phys. 1968, 48,
5048.
53. Jamieson, P. B.; Abrahams, S. C.; Bernstein, J. L. J. Chem. Phys. 1969, 50,
4352.
54. Chi, E. O.; Gandini, A.; Ok, K. M.; Zhang, L.; Halasyamani, P. S. Chem. Mater.
2004, 16, 3616.
55. Guesic, J. E.; Levingstein, H. J.; Rubin, J. J.; Singh, S.; Van Uitert, L. G. Appl.
Phys. Lett. 1967, 11, 269.
56. Wemple, S. H.; Didomenico, M.; Camlibel, I. Appl. Phys. Lett. 1968, 12, 209.
57. Nakano, J.; Yamada, T. J. Appl. Phys. 1975, 46, 2361.
58. Giess, E. A.; Burns, G.; OKane, D. F.; Smith, A. W. Appl. Phys. Lett. 1967, 11, 233.
59. Masuda, Y.; Wada, M. Ferroelectrics 1974, 8, 515.
60. Xu, Y. Ferroelectric Materials and Their Applications; Elsevier: New York, NY,
1991.
61. Carruthers, J. R.; Grasso, M. J. Electrochem. Soc. 1970, 117, 1426.
62. Glass, A. M. J. Appl. Phys. 1969, 40, 4699.
63. van Uitert, L. G.; Singh, S.; Levinstein, H. J.; Geusic, J. E.; Bonner, W. A. Appl.
Phys. Lett. 1967, 11, 161.
64. Fukuda, T.; Hirano, H.; Koide, S. J. Cryst. Growth 1970, 6, 293.
65. Yano, T.; Ohta, T.; Watanabe, A. Jpn. J. Appl. Phys. 1970, 9, 1008.
66. Xu, Y.; Chen, H.; Cross, L. E. Ferroelectr. Lett. Sect. 1984, 2, 189.
60 Chemistry of Polar Transition-Metal Oxides
67. Neurgaonkar, R. R.; Cory, W. K.; Oliver, J. R.; Clark, W. W., III; Wood, G. L.;
Miller, M. J.; Sharp, E. J. J. Cryst. Growth 1987, 84, 629.
68. Fracomber, M. H. Acta Cryst. 1960, 13, 131.
69. Brusset, H.; Gillier-Pandraud, H.; Mahe, R. Bull. Soc. Chim. Fr. 1972, 926.
70. Spaldin, N. A.; Fiebig, M. Science 2005, 309, 391.
71. Fiebig, M. J. Phys. D: Appl. Phys. 2005, 38, R123.
72. Khomskii, D. I. J. Magn. Magn. Mater. 2006, 306, 1.
73. Schmid, H. Magnetoelectric Interaction Phenomena in Crystals; Kluwer:
Dordrecht, 2004; p. 1.
74. Kiselev, S. V.; Ozerov, R. P.; Zhdanov, G. S. Dokl. Akad. Nauk SSSR 1962, 145,
1255.
75. Skinner, S. M. IEEE Trans. Parts, Mater. Packaging 1970, 6, 68.
76. Kubel, F.; Schmid, H. Acta Crystallogr. 1990, B46, 698.
77. Seshadri, R.; Hill, N. A. Chem. Mater. 2001, 13, 2892.
78. Pradhan, A. K.; Zhang, K.; Hunter, D.; Dadson, J. B.; Loutts, G. B.;
Bhattacharya, P.; Katiyar, R.; Zhang, J.; Sellmyer, D. J.; Roy, U. N.; Cui, Y.;
Burger, A. J. Appl. Phys. 2005, 97, 093903.
79. Vedam, K.; Okaya, Y.; Pepinsky, R. Phys. Rev. 1960, 119, 1252.
80. Bertaut, E. F.; Mercier, M. Phys. Lett. 1963, 5, 27.
81. Sugie, H.; Iwata, N.; Kohn, K. J. Phys. Soc. Jpn. 2002, 71, 1558.
82. Aken, v.B.; Palstra, T. T. M.; Filippetti, A.; Spaldin, N. A. Nat. Mater. 2004, 3, 164.
83. Fennie, C. J.; Rabe, K. M. Phys. Rev. B 2005, 72, 100103.
84. Eibschutz, M.; Guggenheim, H. J. Solid State Commun. 1968, 6, 737.
85. Fox, D. L.; Tilley, D. R.; Scott, J. F.; Guggenheim, H. J. Phys. Rev. B 1980, 21,
2926.
86. Kim, S.-W.; Chang, H. Y.; Halasyamani, P. S. J. Am. Chem. Soc. 2010, 132,
17684.
87. Ito, T.; Morimoto, N.; Sadanaga, R. Acta Crystallogr. 1951, 4, 310.
88. Ascher, E.; Schmid, H.; Tar, D. Solid State Commun. 1964, 2, 45.
89. Schmid, H.; Rieder, H.; Ascher, E. Solid State Commun. 1965, 3, 327.
90. Ederer, C.; Spaldin, N. A. Phys. Rev. B 2006, 74, 024102/1.
91. Crottaz, O.; Rivera, J. P.; Revaz, B.; Schmid, H. Ferroelectrics 1997, 204, 125.
92. Mao, S.-Y.; Schmid, H.; Triscone, G.; Muller, J. J. Magn. Magn. Mater. 1999,
195, 65.
93. Crottaz, O.; Rivera, J.-P.; Schmid, H. J. Kor. Phys. Soc. 1998, 32, S1261.
94. Schmidt, G. M. Pure Appl. Chem. 1971, 27, 647.
95. Opik, U.; Pryce, M. H. L. Proc. R. Soc. (London) 1957, A238, 425.
96. Bader, R. F. W. Mol. Phys. 1960, 3, 137.
97. Bader, R. F. W. Can. J. Chem. 1962, 40, 1164.
98. Pearson, R. G. J. Am. Chem. Soc. 1969, 91, 4947.
99. Pearson, R. G. J. Mol. Struct. (Theochem.) 1983, 103, 25.
100. Kunz, M.; Brown, I. D. J. Solid State Chem. 1995, 115, 395.
101. Goodenough, J. B. Annu. Rev. Mater. Sci. 1998, 28, 1.
102. Sidgwick, N. V.; Powell, H. M. Proc. R. Soc. Lond. Ser. A 1940, 176, 153.
103. Gillespie, R. J.; Nyholm, R. S. Q. Rev. Chem. Soc. 1957, 11, 339.
104. Orgel, L. E. J. Chem. Soc. 1959, 3815.
105. Lefebvre, I.; Lannoo, M.; Allan, G.; Ibanez, A.; Fourcade, J.; Jumas, J. C. Phys.
Rev. Lett. 1987, 59, 2471.
106. Lefebvre, I.; Szymanski, M. A.; Olivier-Fourcade, J.; Jumas, J. C. Phys. Rev. B
1998, 58, 1896.
107. Watson, G. W.; Parker, S. C. J. Phys. Chem. B 1999, 103, 1258.
108. Watson, G. W.; Parker, S. C.; Kresse, G. Phys. Rev. B 1999, 59, 8481.
109. Waghmare, U. V.; Spaldin, N. A.; Kandpal, H. C.; Seshadri, R. Phys. Rev. B 2003,
67, 12511.
110. Stoltzfus, M. W.; Woodward, P.; Seshadri, R.; Park, J.-H.; Bursten, B. Inorg.
Chem. 2007, 46, 3839.
111. Payne, D. J.; Egdell, R. G.; Walsh, A.; Watson, G. W.; Guo, J.; Glans, P. A.;
Learmonth, T.; Smith, K. E. Phys. Rev. Lett. 2006, 96, 157403/1.
112. Halasyamani, P. S. Chem. Mater. 2004, 16, 3586.
113. Ok, K. M.; Bhuvanesh, N. S. P.; Halasyamani, P. S. Inorg. Chem. 2001,
40, 1978.
114. Porter, Y.; Bhuvanesh, N. S. P.; Halasyamani, P. S. Inorg. Chem. 2001, 40, 1172.
115. Goodey, J.; Broussard, J.; Halasyamani, P. S. Chem. Mater. 2002, 14, 3174.
116. Goodey, J.; Ok, K. M.; Broussard, J.; Hofmann, C.; Escobedo, F. V.;
Halasyamani, P. S. J. Solid State Chem. 2003, 175, 3.
117. Ra, H.-S.; Ok, K. M.; Halasyamani, P. S. J. Am. Chem. Soc. 2003, 125, 7764.
118. Ok, K. M.; Halasyamani, P. S. Angew. Chem. 2004, 43, 5489.
119. Chi, E. O.; Ok, K. M.; Porter, Y.; Halasyamani, P. S. Chem. Mater. 2006, 18, 2070.
120. Kim, J.-H.; Baek, J.; Halasyamani, P. S. Chem. Mater. 2007, 19, 5637.
121. Sivakumar, T.; Chang, H. Y.; Baek, J.; Halasyamani, P. S. Chem. Mater. 2007,
19, 4710.
122. Chang, H. Y.; Sivakumar, T.; Ok, K. M.; Halasyamani, P. S. Inorg. Chem. 2008,
47, 8511.
123. Kim, J.-H.; Baek, J.; Halasyamani, P. S. Chem. Mater. 2008, 20, 3542.
124. Chang, H. Y.; Kim, S.-H.; Halasyamani, P. S.; Ok, K. M. J. Am. Chem. Soc. 2009,
131, 2426.
125. Chang, H. Y.; Kim, S.-H.; Ok, K. M.; Halasyamani, P. S. J. Am. Chem. Soc. 2009,
131, 6865.
126. Kim, S.-H.; Yeon, J.; Halasyamani, P. S. Chem. Mater. 2009, 21, 5335.
127. Chang, H. Y.; Kim, S.-H.; Halasyamani, P. S. Chem. Mater. 2010, 22,
32413250.
128. Yeon, J.; Kim, S.-H.; Halasyamani, P. S. Inorg. Chem. 2010, 49, 6986.
129. Kim, M. K.; Kim, S.-H.; Chang, H. Y.; Halasyamani, P. S.; Ok, K. M. Inorg. Chem.
2010, 49, 7028.
130. Kim, S.-H.; Melot, B. C.; Seshadri, R.; Green, M. A.; Sefat, A. S.; Mandrus, D.;
Halasyamani, P. S. Chem. Mater. 2010, 22, 5074.
131. Shen, Y. L.; Mao, J. G. Inorg. Chem. 2005, 44, 5328.
132. Jiang, H.-L.; Mao, J. G. Inorg. Chem. 2006, 45, 717.
133. Jiang, H.-L.; Huang, S.-P.; Fan, Y.; Mao, J.-G.; Cheng, W.-D. Chem. Eur. J.
2008, 14, 1972.
134. Jiang, H.-L.; Mao, J.-G. J. Solid State Chem. 2008, 181, 345.
135. Kong, F.; Jiang, H.-L.; Hu, T.; Mao, J.-G. Inorg. Chem. 2008, 47, 10611.
136. Mao, J.-G.; Jiang, H.-L.; Kong, F. Inorg. Chem. 2008, 47, 8498.
137. Hu, T.; Qin, L.; Kong, F.; Zhou, Y.; Mao, J.-G. Inorg. Chem. 2009, 48, 2193.
138. Sun, C.-F.; Hu, C.-L.; Xu, X.; Ling, J.-B.; Hu, T.; Kong, F.; Long, X.-F.; Mao, J.-G.
J. Am. Chem. Soc. 2009, 131, 9486.
139. Zhou, Y.; Hu, C.-L.; Hu, T.; Kong, F.; Mao, J.-G. Dalton Trans. 2009, 5747.
140. Hu, C.-L.; Mao, J.-G. J. Phys. Condens. Matter 2010, 22, 155801/1.
141. Li, P.-X.; Hu, C.-L.; Xu, X.; Wang, R.-Y.; Sun, C.-F.; Mao, J.-G. Inorg. Chem.
2010, 49, 4599.
142. Sun, C.-F.; Hu, C.-L.; Xu, X.; Mao, J.-G. Inorg. Chem. 2010, 49, 9581.
143. Yang, B.-P.; Hu, C.-L.; Xu, X.; Sun, C.-F.; Zhang, J.-H.; Mao, J.-G. Chem. Mater.
2010, 22, 1545.
144. Alvarez, S.; Avnir, D.; Llunell, M.; Pinsky, M. New J. Chem. 2002, 26, 996.
145. Llunell, M.; Casanova, D.; Cirera, J.; Bofill, J. M.; Alemany, P.; Alvarez, S.;
Pinsky, M.; Avnir, D. Shape, 1.1b ed.; University of Barcelona: Barcelona, Spain,
2004.
146. Alvarez, S.; Alemany, P.; Avnir, D. Chem. Soc. Rev. 2005, 34, 313.
147. Ok, K. M.; Halasyamani, P. S.; Casanova, D.; Llunell, M.; Alemany, P.; Alvarez, S.
Chem. Mater. 2006, 18, 3176.
148. Eng, H. W.; Barnes, P. W.; Auer, B. M.; Woodward, P. M. J. Solid State Chem.
2003, 175, 94.
149. Sun, C.-F.; Hu, T.; Xu, X.; Mao, J.-G. Dalton Trans. 2010, 39, 7960.
150. Li, P.-X.; Hu, C.-L.; Xu, X.; Wang, R.-Y.; Sun, C.-F.; Mao, J.-G. Inorg. Chem.
2010, 49, 4599.
151. Sun, C.-F.; Hu, C.-L.; Xu, X.; Mao, J.-G. Inorg. Chem. 2010, 49, 9581.
152. Sykora, R. E.; Ok, K. M.; Halasyamani, P. S.; Albrecht-Schmitt, T. E. J. Am. Chem.
Soc. 2002, 124, 1951.
153. Sykora, R. E.; Ok, K. M.; Halasyamani, P. S.; Wells, D. M.; Albrecht-Schmitt, T. E.
Chem. Mater. 2002, 14, 2741.
154. Shehee, T. C.; Sykora, R. E.; Ok, K. M.; Halasyamani, P. S.; Albrecht-Schmitt, T. E.
Inorg. Chem. 2003, 42, 457.
155. Sullens, T. A.; Jensen, R. A.; Shvareva, T. Y.; Albrecht-Schmitt, T. E. J. Am. Chem.
Soc. 2004, 126, 2676.
156. Sykora, R. E.; Albrecht-Schmitt, T. E. J. Solid State Chem. 2004, 177, 3729.
157. Shvareva, T. Y.; Beitz, J. V.; Duin, E. C.; Albrecht-Schmitt, T. E. Chem. Mater.
2005, 17, 6219.
158. Bray, T. H.; Beitz, J. V.; Bean, A. C.; Yu, Y.; Albrecht-Schmitt, T. E. Inorg. Chem.
2006, 45, 8251.
159. Veltman, T. R.; Stover, A. K.; Sarjeant, A. N.; Ok, K. M.; Halasyamani, P. S.;
Norquist, A. J. Inorg. Chem. 2006, 45, 5529.
160. Halasyamani, P.; Heier, K. R.; Willis, M. J.; Stern, C. L.; Poeppelmeier, K. R. Z.
Anorg. Allg. Chem. 1996, 622, 479.
161. Halasyamani, P.; Willis, M. J.; Stern, C. L.; Poeppelmeier, K. R. Inorg. Chem.
1996, 35, 1367.
162. Halasyamani, P. S.; Poeppelmeier, K. R. Chem. Mater. 1998, 10, 2753.
163. Heier, K. R.; Norquist, A. J.; Wilson, C. G.; Stern, C. L.; Poeppelmeier, K. R. Inorg.
Chem. 1998, 37, 76.
164. Norquist, A. J.; Heier, K. R.; Stern, C. L.; Poeppelmeier, K. R. Inorg. Chem. 1998,
37, 6495.
165. Heier, K. R.; Norquist, A. J.; Halasyamani, P. S.; Duarte, A.; Stern, C. L.;
Poeppelmeier, K. R. Inorg. Chem. 1999, 38, 762.
166. Welk, M. E.; Norquist, A. J.; Stern, C. L.; Poeppelmeier, K. R. Inorg. Chem. 2000,
39, 3946.
167. Welk, M. E.; Norquist, A. J.; Stern, C. L.; Poeppelmeier, K. R. Inorg. Chem. 2001,
40, 5479.
 Chemistry of Polar Transition-Metal Oxides 61

168. Welk, M. E.; Norquist, A. J.; Arnold, F. P.; Stern, C. L.; Poeppelmeier, K. R. Inorg. 206. Curie, J.; Curie, P. Bull. Soc. Min de France 1880, 3, 90.
Chem. 2002, 41, 5119. 207. Voigt, V. G. Lehrbuch der Kristallphysik. B. G. Teubner: Leipzig, Berlin, 1910.
169. Izumi, H. K.; Kirsch, J. E.; Stern, C. L.; Poeppelmeier, K. R. Inorg. Chem. 2005, 208. Lang, S. B. Sourcebook of Pyroelectricity. Gordon & Breach Science: London, 1974.
44, 884. 209. Landolt, H., Ed.; Numerical Values and Functions from the Natural Sciences
170. Marvel, M. R.; Lesage, J.; Baek, J.; Halasyamani, P. S.; Stern, C. L.; and Technology (New Series), Group 3: Crystal and Solid State Physics;
Poeppelmeier, K. R. J. Am. Chem. Soc. 2007, 129, 13963. Springer Verlag: Berlin, 1979; Vol. 11.
171. Huang, Q.; Hwu, S.-J. Inorg. Chem. 2003, 42, 655. 210. Sawyer, C. B.; Tower, C. H. Phys. Rev. 1930, 35, 269.
172. Mo, X.; Hwu, S.-J. Inorg. Chem. 2003, 42, 3978. 211. Maiman, T. Br. Commun. Electron. 1960, 7, 674.
173. Mo, X.; Ferguson, E.; Hwu, S.-J. Inorg. Chem. 2005, 44, 3121. 212. http://www.continuumlasers.com; http://www.cohr.com/lasers.
174. Queen, W. L.; West, J. P.; Hwu, S.-J.; VanDerveer Donald, G.; Zarzyczny 213. Dougherty, J. P.; Kurtz, S. K. J. Appl. Cryst. 1976, 9, 145.
Matthew, C.; Pavlick Ryan, A. Angew. Chem. Int. Ed. 2008, 47, 3791. 214. Nye, J. F. Physical Properties of Crystals. Oxford University Press: Oxford, 1957.
175. Keszler, D. A. Curr. Opin. Solid State Mater. Sci. 1996, 1, 204. 215. Gray, R. B. US Patent # 2,486,560, 1949.
176. Becker, P. Adv. Mater. 1998, 10, 979. 216. Newnham, R. E. Properties of Materials: Anisotropy, Symmetry, Structure. Oxford
177. Keszler, D. A. Curr. Opin. Solid State Mater. Sci. 1999, 4, 155. University Press: Oxford, 2004.
178. Zobetz, E. Z. Krist. 1982, 160, 81. 217. http://www.sensortech.ca; http://www.americanpiezo.com.
179. Chen, C.; Wu, B.; Jiang, A.; You, G. Sci. Sin. (Ser. B) 1985, 28, 235. 218. Brewster, D. Edinburgh J. Sci. 1824, 1, 208.
180. Chen, C. T.; Wu, Y. C.; Jiang, A.; Wu, B. C.; You, G.; Li, R. K.; Lin, S. J. J. Opt. 219. Valasek, J. Phys. Rev. 1920, 15, 537.
Soc. Am. B 1989, 6, 616. 220. Valasek, J. Phys. Rev. 1921, 17, 475.
181. Wu, Y. C.; Sasaki, T.; Nakai, S.; Yokotani, A.; Tang, H.; Chen, C. T. Appl. Phys. 221. Ta, Y. Compt. Rendus lAcad. Sci. 1938, 207, 1042.
Lett. 1993, 62, 2614. 222. Byer, R. L.; Roundy, B. Ferroelectrics 1972, 3, 333.
182. Mori, Y.; Kuroda, I.; Najakima, S.; Sasaki, T.; Nakai, S. Appl. Phys. Lett. 1995, 223. Chynoweth, A. G. J. Appl. Phys. 1956, 27, 78.
67, 1818. 224. Lang, S. B.; Steckel, F. Rev. Sci. Instrum. 1965, 36, 929.
183. Hellwig, H.; Liebertz, J.; Bohaty, L. Solid State Commun. 1999, 109, 249. 225. Ackermann, W. Annln. Physik. 1915, 351, 197.
184. Keszler, D. A.; Akella, A.; Schaffers, K. I.; Alekel, T., III Mater. Res. Soc. Symp. 226. Glass, A. M. J. Appl. Phys. 1969, 40, 4699.
Proc. 1994, 329, 15. 227. Busch, G. Helv. Phys. Acta 1937, 10, 261.
185. Akella, A.; Keszler, D. A. J. Solid State Chem. 1995, 120, 74. 228. Busch, G.; Scherrer, P. Naturwissenschaften 1935, 23, 737.
186. Smith, R. W.; Keszler, D. A. J. Solid State Chem. 1997, 129, 184. 229. Dias, E. D.; Pragasam, R.; Murthy, V. R. K.; Viswanathan, B. Rev. Sci. Instrum.
187. Maczka, M.; Waskowska, A.; Majchrowski, A.; Zmija, J.; Hanuza, J.; 1994, 65, 3025.
Peterson, G. A.; Keszler, D. A. J. Solid State Chem. 2007, 180, 410. 230. Dawber, M.; Farnan, I.; Scott, J. F. Am. J. Phys. 2003, 71, 819.
188. Zhang, J.-H.; Hu, C.-L.; Xu, X.; Kong, F.; Mao, J. G. Inorg. Chem. 2011, 50, 231. Haertling, G. H. J. Am. Ceram. Soc. 1999, 82, 797.
1973. 232. Lee, J.-S.; Kang, B. S.; Jia, Q. X. Appl. Phys. Lett. 2007, 91, 142901.
189. Pan, S.; Smit, J. P.; Watkins, B.; Marvel, M. R.; Stern, C. L.; Poeppelmeier, K. R. 233. Park, J. Y.; Park, J. H.; Jeong, Y. K.; Jang, H. M. Appl. Phys. Lett. 2007, 91,
J. Am. Chem. Soc. 2006, 128, 11631. 152903.
190. Li, F.; Hou, X.; Pan, S.; Wang, X. Chem. Mater. 2009, 21, 2846. 234. Gu, Z.-G.; Zhou, X.-H.; Jin, Y.-B.; Xiong, R.-G.; Zuo, J.-L.; You, X.-Z. Inorg.
191. Li, F.; Pan, S.; Hou, X.; Yao, J. Cryst. Growth Des. 2009, 2009, 4091. Chem. 2007, 46, 5462.
192. Li, F.; Pan, S.; Hou, X.; Zhou, Z. J. Cryst. Growth 2010, 312, 2083. 235. Fu, D.-W.; Song, Y.-M.; Wang, G.-X.; Ye, Q.; Xiong, R.-G.; Akutagawa, T.;
193. Li, F.; Pan, S.; J. Cryst. Growth, in press. Nakamura, T.; Chan, P. W. H.; Huang, S. D. J. Am. Chem. Soc. 2007, 129, 5346.
194. Zhang, G.; Wu, Y.; Fu, P.; Wang, G.; Pan, S.; Chen, C. Chem. Lett. 2001, 456. 236. Scott, J. F. J. Phys. Condens. Matter 2008, 20, 021001.
195. Li, Z.; Wu, Y.; Fu, P.; Pan, S.; Lin, Z.; Chen, C. J. Cryst. Growth 2003, 255, 119. 237. Ko, T.-Y.; Jun, C.-H.; Lee, J.-S. Kor. J. Ceram. 1999, 5, 341.
196. Pan, S.; Smit, J. P.; Marvel, M. R.; Stampler, E. S.; Haag, J. M.; Baek, J.; 238. Mercurio, D.; El Farissi, M.; Frit, B.; Goursat, P. Mater. Chem. Phys. 1983,
Halasyamani, P. S.; Poeppelmeier, K. R. J. Solid State Chem. 2008, 181, 2087. 9, 467.
197. Wang, Y.; Pan, S. J. Synth. Cryst. 2009, 38, 87. 239. Porter, Y.; Ok, K. M.; Bhuvanesh, N. S. P.; Halasyamani, P. S. Chem. Mater.
198. Wang, Y.; Pan, S.; Tian, X.; Zhou, Z.; Liu, G.; Wang, J.; Jia, D. Inorg. Chem. 2009, 2001, 13, 1910.
48, 7800. 240. Chang, H. Y.; Kim, S.-H.; Ok, K. M.; Halasyamani, P. S. Chem. Mater. 2009, 21,
199. Wang, Y.; Pan, S.; Hou, X.; Liu, G.; Wang, J.; Jia, D. Solid State Sci. 2010, 1654.
12, 1726. 241. Zhang, S.-Y.; Hu, C.-L.; Sun, C.-F.; Mao, J.-G. Inorg. Chem. 2010, 49, 11627.
200. Cady, W. G. Piezoelectricity; An Introduction to the Theory and Applications of 242. Krogh-Moe, J. Acta Crystallogr. 1962, 15, 190.
Electromechanical Phenomena in Crystals. Dover: New York, 1964. 243. Block, S.; Perloff, A.; Weir, C. E. Acta Crystallogr. 1964, 17, 314.
201. Kurtz, S. K.; Perry, T. T. J. Appl. Phys. 1968, 39, 3798. 244. Aka, G.; Kahn-Haradi, A.; Vivien, D.; Benitez, J.-M.; Salin, F.; Godard, J. Eur. J.
202. Jaffe, B.; Cook, W. R.; Jaffe, H. Piezoelectric Ceramics. Academic Press: London, Solid State Inorg. Chem. 1996, 33, 727.
1971. 245. Iwai, M.; Kobayashi, T.; Furuya, H.; Mori, Y.; Sasaki, T. Jpn. J. Appl. Phys. 1997,
203. Schwartz, R.; Ballato, J.; Haerling, G. Materials Engineering; Marcel Dekker: 36, L276.
New York, 2004; p. 207. 246. Mei, L.; Wang, Y.; Chen, C. Mater. Res. Bull. 1994, 29, 81.
204. Franken, P. A.; Hill, A. E.; Peters, C. W.; Wienrich, G. Phys. Rev. Lett. 1961, 7, 118. 247. Mei, L.; Wang, Y.; Chen, C. T.; Wu, B. J. Appl. Phys. 1993, 74, 7014.
205. Armstrong, J. A.; Bloembergen, N.; Ducuing, J.; Pershan, P. S. Phys. Rev. 1962, 248. Nicholls, J. F. H.; Chai, B. H. T.; Corker, D. L.; Calabrese, T. C.; Henderson, B.
127, 1918. Proc. SPIE Int. Soc. Opt. Eng. 1993, 54, 1863.
 7.07 Mixed-Metal Oxides
G Centi and S Perathoner, University of Messina, Messina, Italy
2013 Elsevier Ltd. All rights reserved.

7.07.1 Introduction 153

7.07.1.1 Uses of Mixed-Metal Oxides in Catalysis, Characteristics, and Open Problems 153
7.07.1.1.1 The role of the synthesis procedure and of the activation processes 154
7.07.1.1.2 Multifunctional behavior of mixed-metal oxide catalysts 160
7.07.1.1.3 Open fundamental questions in the reactivity of oxide catalysts 162
7.07.1.1.4 Open questions from an industrial perspective 164
7.07.2 Bulk and Surface Structure of Metal Oxides 165
7.07.2.1 Long- and Short-Order Ordering: The Role of Microstructure 166
7.07.2.2 The Role of Strains and Defects 167
7.07.2.3 Oxide Surfaces 168
7.07.3 Low-Dimensional Metal Oxide Nanostructures 169
7.07.3.1 Metal Oxides Confined within a Host Ordered Matrix 170
7.07.3.2 Ordered Assemblies of Metal Oxide
Nanostructures 171
7.07.3.2.1 Thin oxide films 172
7.07.3.3 Reactivity of Low-Dimensional Metal Oxides 172
7.07.3.4 Opportunities to Design Advanced Catalytic Materials 173
7.07.4 Catalytic Properties of Mixed-Metal Oxides 175
7.07.4.1 Surface Acidity and Basicity of Metal Oxides 178
7.07.4.2 Modification of the Surface Reactivity by Chemisorbed Species 179
7.07.5 Conclusion 181
References 181

These aspects will be specifically discussed in other chapters

7.07.1 Introduction
(see Chapters 7.04, 7.05, 7.16 and Volume 5, respectively).
7.07.1.1 Uses of Mixed-Metal Oxides in Catalysis, The use of supported metal oxide monolayer type catalysts, of
Characteristics, and Open Problems heteropoly compounds, metal oxide solid acids, and photoca-
talysts will be discussed in other chapters of this volume (see
Catalysis is one of the key technologies for the development
Chapters 7.06, 7.09, and 7.22, respectively).
of society, economy, and industrial production, with about
Oxide-based catalysts are widely used in chemical processes
a quarter of GNP (gross national production) and over 90%
in several applications ranging from selective oxidation to total
of the chemical industry involving products made using
oxidation, hydrogenation/dehydrogenation, acido-base, and
catalysts.1 They include food, fuels, polymers, textiles, pharma/
environmental applications (NOx conversion, etc.). More
agrochemicals, and a large variety of other industrial and con-
than one-third of the worldwide catalyst production is based
sumer products. In addition, catalysis plays a fundamental role
on oxide-type catalysts, apart from their use as support for
for the clean and sustainable use and production of energy, and in
other active components. Selective oxidation can be consid-
many pollution and global warming control technologies. It also
ered the most typical example of the use of oxide-type materials
plays an important role for the preparation of new materials, both
as heterogeneous catalysts6,7 and a forefront area in the
organic and inorganic, for example, carbon nanotubes, advanced
challenge for a more sustainable chemical industry.8,9 Around
materials for production, storage, conversion of energy, etc.1,2
one-quarter of the main organic chemicals are produced by
Metal oxides are used in over three-quarter of these appli-
selective oxidation processes.
cations,35 either directly as catalysts or as catalytic support
Table 2 reports selected examples of material systems relevant
(both for dispersed metals or metal oxides), even if often the
in selective oxidation, their active phase and related structural
support plays also a co-catalytic role, for example, in bifunc-
features, together with a list of typical dopants used in industrial
tional catalysts. The main classes of application of mixed-metal
compositions for these catalysts, with the main role of these
oxides in catalytic industrial processes are listed in Table 1,
dopants.
showing the very large range of applications in which oxide
Although the typical composition of mixed oxide catalysts
catalysts are used. This table does not include the use of oxides
used in industrial processes of selective oxidation may be com-
and mixed oxides as
plex, with up to 20 elements, research has typically identified a
1. the support for dispersed metals, in a large range of appli- main active phase and the role of the principal dopants: tune the
cations from refinery to fine chemical synthesis and surface activity to control surface reactivity, improve the redox
2. catalysts with an ordered micro- or mesoporous structure or structural properties, or promote the activity/selectivity.
(zeolites, mesoporous oxides, etc.). The active phases are characterized from key active elements

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00718-X 153

154 Mixed-Metal Oxides

Table 1 Examples of application of metal oxides in catalytic industrial processes

Processes Examples of industrial applications

Selective oxidation and ammoxidation of olefins, Propene oxidation to acrolein and ammoxidation to acrylonitrile on bi-molybdate-based
alkanes, and alkylaromatics catalysts
Methanol oxidation to formaldehyde over iron-molybdate-based catalysts
Acrolein oxidation to acrylic acid on V/Mo/W-oxide
n-Butane oxidation to maleic anhydride over vanadium phosphate catalysts
Oxidative dehydrogenation of light alkanes over V- and Cr-based mixed oxides
o-Xylene oxidation to phthalic anhydride over V2O5/TiO2
Total oxidation of pollutants Combustion of volatile organic compounds on metal oxides
Catalytic wet oxidation of pollutants in aqueous waste streams
Photocatalytic removal of pollutants in liquid or gaseous streams on TiO2
Selective reduction of pollutants Selective catalytic reduction of NOx on V/TiO2-based catalysts
Selective reduction of SO2 (Claus process) on TiO2
Olefin metathesis Conversion of ethene and 2-butene to propene on WO3/SiO2
Alkane isomerization n-Butane isomerization to isobutane on sulfated zirconia
Water gas shift (WGS) H2 production process: WGS over Fe/Cr oxide high-temperature shift (HTS) and Cu/Zn/Al
oxide low-temperature shift (LTS) catalysts
Alkylation Alkylation of phenol with methanol on MgO
Dehydration/condensation Dehydration of 1-hexylethanol and of iso-butyraldehyde on ZrO2
Transesterification Production of biodiesel from vegetable oils using solid catalysts (Ca/Al2O3)
Esterification Esterification of ethylene oxide with alcohols on hydrotalcite catalysts

Table 2 Examples of material systems relevant in selective oxidation, their active phase and related structural features, and typical dopants used
in industrial compositions for these catalysts, with the main role of these dopants

Material Type of reaction Active phase Structure Typical dopants and

systems their main role

Bi-molybdate Propylene to acrolein g-Bi2MoO6 [Bi2O2]2 layers alternating with layers of (K,Cs)acidity
or acrylonitrile corner-sharing MoO6 octahedra (Ni,Mg,Mn)structural
(Fe,Cr)redox
Fe-molybdate Methanol to Fe2(MoO4)3 Isolated MoO4 tetrahedra and FeO6 octahedra, (MoO3)stability/activity
formaldehyde which share corners (K,Cs,P)acidity
(Bi,Te)redox
Vanadium n-Butane to maleic (VO)2P2O7 Pairs of edge-sharing pseudo-octahedrally (VV-phosphate on
phosphate anhydride coordinated vanadium ions which are isolated by the surface)activity/
pyrophosphate groups selectivity
MoVTeNb Propane to acrylic acid Mo7.8 V1.2NbTe0.94O28.9 M1 phase (orthorombic)
oxides or acrylonitrile Sharing corner MO6
Octahedra (M Mo, V) assembled into layers in
the ab-plane while hexagonal and
heptagonal channels are formed along the c-axis
V-antimonate Propane to VSbO4 Rutile (Sb2Ox)stability/activity
acrylonitrile (W,Sn)structural
(Te)redox

(typically, V, Mo, Te, and Sb) and a second element whose role is
to create the necessary bulk and surface structure to generate the
surface geometrical arrangement necessary for the catalytic reac-
tion, as well as to provide the structural stability to maintain
the catalytic behavior over long time. Figure 1 reports the
crystallographic structures for the active phases g-Bi2MoO6,
(VO)2P2O7, and Mo7.8V1.2NbTe0.94O28.9 (M1 phase, see later).10
These are model structures for the active catalysts. Although
the bulk structure well corresponds to these crystallographic
structures, the surface may be different, either due to relaxation
phenomena as discussed later or due to the presence of other
surface phases. This is well exemplified in the case of
(VO)2P2O7, the active catalyst and phase in n-butane selective
oxidation to maleic anhydride (see later).6,7 Although the cat-
alyst is well crystalline and x-ray diffraction (XRD) shows the
presence of only (VO)2P2O7 phase, high-resolution transmis-
sion electron microscopy (HRTEM) images of (VO)2P2O7 ac-
tive catalyst show the presence of surface VV-phosphate species
as well as of amorphous regions.11
7.07.1.1.1 The role of the synthesis procedure and of
the activation processes
The catalytic selective conversion of hydrocarbons, due to the
need of controlled multifunctional properties as described
later, requires demanding characteristics in the material. For
this reason, specific catalyst compositions are necessary, which
 Mixed-Metal Oxides 155

b a
2+
Bi2O2 layer
c

2
MoO4 layer b
(distorted MoO6
octahedra)
c
c a (VO)2P2O7
-Bi2MoO6
Mo7.8V1.2NbTe0.94O28.9(M1 phase)

Mo/V
c V P O
Mo
Nb
Te
O

Figure 1 Crystallographic structures for the active phases g-Bi2MoO6, (VO)2P2O7, and Mo7.8 V1.2NbTe0.94O28.9 (M1 phase). Adapted from (a)
Kongmark, C.; Martis, V.; Pirovano, C.; Lofberg, A.; van Beek, W.; Sankar, G.; Rubbens, A.; Cristola, S.; Vannier, R. -N;. Bordes-Richard, E. Catal. Today
2010, 157, 257262; (b) Kuhlmann, U. Phonon and Spin Dynamics of (VO)2P2O7. Ph.D. Thesis, Technischen Universitat, Berlin, 2005; (c) Grasselli, R. K.;
Lugmair, C. G.; Volpe Jr., A. F. Top. Catal. 2011, 54, 595604, respectively.

Amorphous region

Surface Vv-phosphate

2 nm 2 nm

Figure 2 (Left) Triangular diagram of composition for the Mo/V/Nb(Te or Sb)/O system, indicating the stoichiometry of compounds which form, and
the area of existence for systems active and selective in alkanes oxidation and ammoxidation. (Right) Detail of the left figure. Adapted from Cavani, F.;
Centi, G.; Marion, P. In Metal Oxide Catalysis, Vol. 2; Jackson, S. D.; Hargreaves, J. S., Eds.; Wiley-VCH: Weinheim 2009; pp 771818.

should be related to the specific type of catalytic reaction
investigated. This concept is illustrated in the analysis of the
MoVTeSb(Nb)-oxide catalysts.12 The elements constitut-
ing this catalyst are the basis for several crystalline structures,
as illustrated in Figure 2, which summarizes the main mixed
oxides that can be obtained by combination of these elements.
The figure shows the various bi-component systems in a
triangular diagram composition, as well as the areas which
include tri- and multicomponent systems, exhibiting superior
performance in ethane oxidation, and in propane oxidation
and ammoxidation. Many of the reported structures are re-
lated. For example, orthorhombic Mo5 x(V,Nb)xO14 solid
solutions are isostructural with y-Mo5O14 and with several bi-
component Sb/Mo, Nb/Mo, and Te/Mo mixed oxides. The M1
phase is structurally derived from the latter oxides, and the M2
phase is related to other Mo-leaner Te/Mo and Sb/Mo mixed
oxides. Oxides with similar structure may result in a quite
different catalytic behavior; however, selective catalysts may
156 Mixed-Metal Oxides
often be reasonably good in the selective oxidation of different
substrates, due to their multifunctional characteristics (see
later). It may be noted in Figure 2 that a very limited phase
space designates all good catalyst systems for selective oxida-
tion and ammoxidation. It is worth mentioning that the ma-
jority of selective oxidation catalysts are bulk mixed oxides,
with few exceptions of oxides supported on other oxides
(such as vanadium oxide on TiO2) or metal nanoparticles
supported on oxides (for example, Au/CeO2).
Not only the occurrence of a specific surface structure and
phase composition (active phase) but often a synergic cooper-
ation with the other phases present are necessary. The nano-
structure at the interface between these phases thus becomes a
key element for the definition of an active catalyst. As a
consequence, the synthesis procedures required to arrive at
useful oxidation catalysts play a critical role, because they
determine the effective phase purity, their morphology and
nanostructure, and the interface with other eventually present
phases. The book Synthesis of Solid Catalysts13 provides many
examples of recent progress made in mixed oxide catalysts
preparation as well as related basic principles and tools, but
only two examples are discussed here:14
1. how in situ analytics are the necessary tool to correctly
understand the often complex preparation and phase
chemistry, as well as surface and compositional dynamics
during the catalytic reaction;
2. the importance of synthesis procedure for mixed oxide
catalysts and how performance is an extrinsic property of
materials.
The first example regards the vanadium phosphate (VPO)
catalysts, which is the first industrially applied case of mixed
oxide catalysts selective in the demanding reaction of alkane
oxidation (n-butane to maleic anhydride, a 14th electron
oxidation).6,7 Many well-characterized, crystalline VPO phases
Mo5-x(V,Nb)xO14 solid
solutions derived from Mo
Orthorh. structures
orthorhombic q-Mo5O14
Mo0.93V0.07O2.8 Sb2Mo20O5s, Nb0.09Mo0.91O2.8
Mo0.88V0.12O2.94 Sb2Mo10O31, TeMo3O16
TeMo4O13
VMo3O11 Sb2Mo6O19, Te2Mo6O20
Sb4Mo10O31
V0.33Mo0.67O2
Orthorhombic
Hex Hexag. structures
Ethane to ethylene
or acetic acid Nb2Mo3O11
VMoO5, V0.95Mo0.97O5 Active and selective in:
Propane oxidation to acrylic acid
V6Mo4O25, V9Mo6O40 propane ammoxid. to acrylonitrile
Ethane oxidation to ethylene/acetic acid
V2MoO8 Te2MoO7 1. Mo7.5V1.5TeNbO29
Benzene to
maleic anh.
V TeVO4, VNbO5, VNbO4, VSbO4
Figure 3 (Left) simplified scheme of the procedure of preparation of VOHPO4 1/2H2O in aqueous and organic medium, and relative SEM micrograph
images. Structure of the hemihydrate VOHPO41/2H2O precursor phase. (Right) Scheme of the evolution of the phases in vanadium phosphate catalyst
during activation. Structure of the (VO)2P2O7 active phase. Adapted from Dummer, N. F.; Bartley, J. K.; Hutchings, G. J. Adv. Catal. 2011, 54, 189247.
have been identified, whose structure and catalytic properties
have been well documented. Some of the most widely studied
are given as follows15:
1. V5 vanadyl orthophosphate (a-, b-, g-, d-, e-, and
o-VOPO4, and VOPO42H2O);
2. V4 hydrogen phosphates (VOHPO44H2O, VOHPO41/
2H2O, VO(H2PO4)2);
3. vanadyl pyrophosphate ((VO)2P2O7);
4. vanadyl pyrophosphate (VO(PO3)2).
Of these phases, VOHPO41/2H2O (vanadyl hydrogen
phosphate hemihydrate) is the precursor phase (during the
synthesis procedure) for the active phase (vanadyl pyrophos-
phate) which is obtained through a topochemical transforma-
tion, for example, retaining the morphological characteristics
of the precursor phase. Figure 3 (top) shows how the mor-
phology of the obtained VOHPO41/2H2O depends on the
procedure of preparation (in aqueous and organic medium;
other routes also exist).15 The procedure in the organic me-
dium using an alcohol such as isobutanol (both as reaction
medium and as reducing agent for V2O5) leads to a layered-
type VOHPO41/2H2O which gives, after transformation to
vanadyl pyrosphate, a more active/selective catalyst. To note
the structural analogies between the precursor and active
phases, a main reason explaining why the characteristics of
the latter could be in part controlled from the synthesis proce-
dure of the precursor phase.
The structure of the VPO catalysts is thus dependent on a
number of factors, among which the most important are the
(1) reaction medium of the synthesis procedure (aqueous or
organic medium) and type of reducent used to prepare the
precursor phase from V5 source, (2) P/V stoichiometry, and
(3) thermal treatment procedure (time, temperature, and at-
mosphere). These factors determine the type and relative
amounts of the crystalline phases present, but the complete
80% Mo 80% Mo
20% V 20% Te(sb) + Nb
70% Mo
5 1
2 1
4
3
6C% Mo 60% Mo
4C% V 40% Te(Sb) + Nb
Orthorhombic (M1) (pseudo)-hexagonal (M2)
(active and selective in propane (amm)ox (active and selective in propane (amm)ox
(isostr. Te2M23O57, Te2M10O32,
Sb2M23O58; M = Mo, V, Nb)
2. Mo6V2SbOx,Mo6V2TeOx
3. Mo6V3Te1Ox
4. Mo1.0V0.44Te0.10Ox
5. Mo1.0V0.25Te0.11Nb0.12Ox
(Te or Sb) + Nb

identification requires the combined use of different tech-
niques, from XRD to infrared/Raman, HRTEM, and other
methodologies (solid-state NMR, etc.). Often, characterization
is limited to the starting (fresh) catalyst, but this slowly trans-
forms during the catalytic reaction (sometimes, up to 1 month
of activation procedure is necessary), leading typically to the
presence of only one crystalline phase (vanadyl pyrophos-
phate) in the most active catalysts, even if the presence of
some amorphous and V5 phases could also be detected by
the HRTEM analysis (Figure 4). The most active (VO)2P2O7
samples have a preferential exposition of the (100) face (bc
plane in Figure 1) as well as the presence of some stacking
disorder between these planes, which increases the Lewis acid-
ity of the surface-exposed V4 ions, an important aspect for the
activation of n-butane molecule.16 The presence of some sur-
face V5 ions is instead important for the oxygen insertion
mechanism, but an excessive amount, as well as the presence
of V5-phosphate nanocrystals, leads to overoxidation with
formation of carbon oxides and a lowering of the selectivity.
Some slight excess of phosphorus above the P/V 1.0 ratio
(typically 1.031.05 in the best catalysts) is necessary to avoid
the presence of unreacted V2O5 nanoparticles on the surface
(lowering the selectivity) and to maintain a stable catalytic
behavior. However, a larger excess is detrimental, leading to
the formation of metaphosphate species. During activation,
the precursor phase topochemically transforms to the active
phase, but oxidation to V5 phases can also occur in the
presence of gaseous O2. Figure 3 (bottom) schematically
shows this evolution, reporting only two phases for V5-
phosphate, but different polymorphic V5-phosphate phases
are possible and form, depending on the conditions of pre-
paration, as introduced earlier. These polymorphic phases can
1. , 2 h
V2O5 + HCI
2. H3PO4
V2O5 + H3PO4 + alcohol
V4+
V5+
1 3
2
VO(HPO4)0.5H2O
Precursor
1: Oxydehydration
2: Topotactic transformation
Figure 4 HRTEM images of (VO)2P2O7 catalyst. Courtesy of FHI-MPG. Adapted from Wang, D. Characterization of catalysts by TEM. Presented at 36th
Erice Crystallographic Course on Electron Crystallography. Erice, Italy, Jun 920, 2004.
Mixed-Metal Oxides 157
transform to the active phase in the presence of reducing
agents, which include the hydrocarbons present in the reaction
medium during the catalytic reaction. However, the different
V5-phosphate phases have a different rate of reduction and
reducibility. A combination of in situ techniques is necessary to
understand this complex phase chemistry and its dynamic of
transformation during catalytic reaction. This aspect was dem-
onstrated by Schlogl, Hutchings, and coworkers in a seminal
paper,17 where it was evidenced how the dynamics of VOPO4
transformation during the catalyst activation during n-butane
oxidation depends on the chemical potential of hydrocarbon
species. The catalytic behavior at steady-state conditions
depends on the ability of the starting (fresh) precatalyst to
transform gradually into a surface microstructure producing
site-isolated active species, which in turn depends on the
chemical potential of the reagents during the activation stage.
The evolution of phosphate coverage during catalyst activation
was monitored by surface-sensitive x-ray photoelectron spec-
troscopy (XPS),18 showing that mobile phosphate units are
present at the reaction conditions. HRTEM (Figure 4) also
provides detailed indications about the terminating surface
structure of the catalysts. By combining these and many other
indications, Schlogl suggested the wiring diagram reported in
Figure 5 for the mode of operation of the VPO system during
steady-state operations.14 This structural dynamics may be
considered an extension and refinement of the idea of living
active surface earlier proposed for these catalysts.6
VPO is a bulk precatalyst that is not identical to the pro-
jections of the bulk VPO structure to its terminations. The
surface chemistry has to be linked to the effect of gas-phase
controls to the solid-state chemistry which in turn is related
to the real structure (nanosize, defects, and compositional
VOHPO41/2H2O
10 m
VOHPO41/2H2O
, 16 h
VOHPO41/2H2O
10 m
4
VOPO4 VOPO4
5
(VO)2P2O7
Catalyst
c
3, 5: Reduction V5+ to V4+
4: Isovalence transformation (VO)2P2O7
b
158 Mixed-Metal Oxides

Hemihydrate
VPO large
Equilibration
support
Butane Oxygen
VPO nanostructured

Metastable VOPO4
nanostructured Activation

Reduction
Restructuring
Restructuring

VOPO4 large
Passivation
passivated
Supported
VxOy/H3PO4 H3PO4

Hydrolysis Activation

Block Nano VxOy5+ C-H activation

H2O Nano VxOy3+ sel. oxidation

adsorbates
Products

MA Inhibition
Regeneration CO2

Nano VxOy5+ Re-oxidation

Hydrothermal

VxOy large Re-combination

VOPO4
Exposed

Combustion

H 2O
CO2

Figure 5 Wiring diagram of the mode of operation of the VPO system under steady-state operations. The central shaded part represents the catalytic
operation, the outer parts highlight the complexity of the structural dynamics creating a steady-state abundance of isolated active sites. Reproduced with
permission from Schlogl, R. In Modern Heterogeneous Oxidation Catalysis; Mizuno, N., Ed.; Wiley-VCH: Weinheim 2009; pp 142, Chapter 1.

fluctuations). The central shaded part in Figure 5 describes the
catalytic operation of active sites. These are considered isolated
sites of a cluster of vanadium oxo species, but where the
surrounding pending phosphate group acts as co-catalysts
and helps in controlling mobility of surface adspecies.6 Al-
though due to space limits it is possible to provide only the
flavor of this complex phase and surface chemistry, it clearly
emerges from the presented discussion the determining role of
a precise control in the chemistry of preparation and activation
to produce highly active/selective catalysts, which in turn de-
pend on the possibility to use many in situ characterization
techniques to determine the exact nature at the nanoscale level
of the active mixed-metal oxide catalyst.
The second example regards the synthesis chemistry of
phase-pure M1MoVTeNb oxide catalysts for propane ox-
idation to acrylic acid, another industrially relevant case of
light alkane selective oxidation. The M1 phase, whose surface
crystalline structure is shown in Figure 1, is the active phase in
the complex MoVTeNb oxide material system (Figure 2)
for the propane (amm)oxidation. Two main preparation
methods the dry-up method and the hydrothermal synthesis
method have been used to synthesize these catalysts.12 The
 Mixed-Metal Oxides 159

dry-up method involves mixing aqueous slurries of metal oxide

precursors followed by a gradual evaporation of the combined
aqueous slurry. Solvent evaporation leads to nucleation and
growth of precursor metal oxide phases, which require further
heat treatment to obtain active catalysts. In addition to the
dominant M1 and M2 phases, impurity phases, for example,
MoO3, Mo6V9O40, and Mo3Nb2O11, are also present. Purifica-
tion by posttreatment with nitric acid or H2O2 gives access to
phase-pure M1. However, the chemical composition of M1 is
also affected by this treatment.
In the M1 phase (ICSD 55097), corner-linked MO6
(M Mo, V) octahedrons are assembled forming six- and,
seven-membered rings, partially accommodating TeO units.
Niobium has been postulated to be exclusively located in the
center of a MO7 pentagonal bipyramidal unit sharing edges
with the surrounding octahedrons.19,20 The (001) planes are
congruently stacked along the [001] direction, forming a
bronze-like structure similar to the structure of Cs0.7(Nb2.7
W2.3)O14. The M2 phase differs from M1 by the absence of
the pentagonal bipyramidal unit, and the seven-membered
ring.19 The formulas of the refined unit cells have been deter-
mined to be Mo7.8 V1.2Te0.937NbO28.9 for M1 and Mo4.31
V1.36Te1.81Nb0.33O19.81 for M2, respectively.19 Symbiosis be-
tween the two phases has been suggested to be responsible
for maximum yield of acrylic acid or acrylonitrile in the oxida-
tion or ammoxidation of propane, respectively.21 Selective
oxidation of propane has been assigned to M1. M2, on its
own, does not activate propane, but selectively oxidizes the
propylene intermediate.
The preferable method for catalyst preparation is the appli-
cation of hydrothermal conditions, because they allow the
synthesis of single-phase complex metal oxides, such as binary
MoVO, ternary MoVTe, and quaternary MoVTeNb mixed ox-
ides, which have various crystal structures and performances.12
Selective synthesis of phase-pure M1 requires a very precise
control of preparation parameters, such as temperature, metal
salt concentration, metal stoichiometry, batch size, pressure,
and redox potential of the synthesis mixture. The latter is
determined by pH, solvent, organic additives, and the nature
of metal salt precursors. Moreover, different characteristics of a
specific autoclave, for example, wall material, volume, and
heating/cooling regime, have a substantial influence on the
crystallization behavior. Schlogl and coworkers22 showed
how the precise control of all these aspects (summarized in
Figure 6) is necessary to prepare phase-pure samples and
that application of identical reaction parameters does not
necessarily mean that identical materials with comparable
catalytic properties will be produced when the reaction is
performed in different autoclaves. The availability of in situ
techniques is required to monitor the hydrothermal reaction
inside the autoclave and to generate quantitative kinetic
information.
In situ Raman spectroscopy allows insight into molecular
processes under hydrothermal conditions during synthesis of
complex nanostructured MoVTeNb oxides (Figure 7).23 basal plane of the cylindrical M1 particles (Figure 8).25 Termi-
Based on the knowledge acquired, the synthesis can be more
efficiently directed toward the desired product with improved
functionality. It is possible, for example, not only have reliable
procedures for controlled synthesis of phase-pure MoVTeNb
mixed oxides with M1 structure, but also produce tunable
(NH4)6Mo7O24 (52 mmol Mo/100 ml H2O)
T = 353 K
VOSO4(s) Te(OH) 6(s)
(13 mmol V) (12 mmol Te)
Cool to 313 K
(NH4)[NbO(C2O4)2(H2O)2]3H2O
(6.5 mmol Nb/50 ml H2O)
Initial reaction mixture
Molar ratio Mo/V/Te/Nb = 1/0.25/0.23/0.12
Hydrothermal synthesis
(A1 or A2)
383448 K, 24144 h
Washing, filtration, drying (air, 353 K, 16 h)
Hydrothermal
Precursor post treatment
T = 773 K, p = 20 MPa
atmosphere: Ar
Activation
(calcination in air, 498548 K, 1 h)
heat treatment in Ar, 873923 K, 2 h
MoVTeNb mixed oxide catalyst
Figure 6 Hydrothermal synthesis of MoVTeNb mixed oxides.
Reproduced with permission from Celaya Sanfiz, A.; Hansen, T. W.;
Girgsdies, F.; Timpe, O.; Rodel, E.; Ressler, T.; Trunschke, A.; Schlogl, R.
Top. Catal. 2008, 50, 1932.
crystal dimensions, with this aspect also being critical to
study the structure sensitivity of selective oxidation of propane
to acrylic acid.24
The availability of phase-pure M1, with tunable crystal di-
mensions, also allows analyzing the structure-activity/
selectivity relationships. STEM of a single M1 crystal viewed
in the projection along <001 > revealed that the crystal peri-
odicity is broken alongside the thinnest sections through the
mesh of polyhedra of the M1 structure, resulting in similar
metal-oxo arrangements on the lateral surface and on the
nation is achieved by fracturing basic structural units along
weak features. Only limited local relaxation may occur, such as
conversion of terminal MdO bonds into M^O bonds. The
exact fracture line may vary between individual basic structural
units with their compositional fluctuation (phase width). The
160 Mixed-Metal Oxides

Mo M5O14 intermediate
773
+
Nb
+V 788

Solution, amorphous 809 [(TeO3)2Mo12O36]4-

* *
or semi-crystalline solid 835

T > 120 C 952

+Te Target phase [Mo36O112]8- 975
930 952 918
968
980
1050 988 644
950
448 K
985
912 600
1010

910 -1 )
m
313 K
sh ift (c Te(OH)6
Raman
VOSO4
(Mo,Nb,V)5O14
(a) M1 (b)

Figure 7 (a) Representation of the assembly of structural motifs during formation of the target-phase M1 and the by-product (Mo,Nb,V)5O14. Nb
yellow, Mo blue, V/Mo pale blue, and Te red. (b) Raman in situ spectra taken during hydrothermal synthesis of MoVTeNb oxide. Adapted from
Sanchez Sanchez, M.; Girgsdies, F.; Jastak, M.; Kube, P.; Schlogl, R.; Trunschke, A. Angew. Chem. Int. Ed. 2012, 51, 71947197.

10 nm
(a) (b)

Figure 8 (a) Color mode (based on the contrast variation) HR TEM image of one M1 particle along the {001} direction. The perimeter of the M1 particle
is highlighted. (b) Termination model of the M1 phase. Adapted from Zhang, W.; Trunschke, A.; Schlogl, R.; Su, D. Angew. Chem. Int. Ed. 2010, 49,
60846089.

resulting termination will be substantially different from an
assumed termination along low-indexed planes of basic struc-
tural units. Details of the synthesis kinetics leading to a distri-
bution of cation positions and hence to a fluctuation of bond
strengths will control the exact surface morphology. The termi-
nation mode is consistent with a perfectly ordered bulk structure
that gives, however, no indication of the surface structure.
By analyzing infrared spectroscopy, microcalorimetry, and in
situ photoelectron spectroscopy characterizations the change of
the phase-pure M1 during the catalytic conversion of propane to
acrylonitrile using different feed compositions,26 it is possible to
relate the surface composition of the M1 phase (which differs
significantly from the bulk) to the catalytic performances.
Acrylic acid formation correlates with surface depletion in
Mo6 and enrichment in V5 sites. In the presence of steam in
the feed, the active ensemble for acrylic acid formation appears
to consist of V5 oxo-species in close vicinity to Te4 sites with
an element ratio Te/V 1.4. The active sites are formed under
propane oxidation conditions and are embedded in a thin layer
enriched in V, Te, and Nb on the surface of the structural stable
self-supporting M1 phase.27
Also this second example evidences the need to develop very
precise and controlled methods for the preparation of mixed-
metal oxides, being typically possible to produce different
phases which may substantially alter the catalytic performances,
in either negative or positive way. In situ methodologies to
control the synthesis procedure are necessary, as well as a com-
bination of different characterization methods to check the
nature and nanostructure of the mixed oxides produced. In
addition, a change of the phases is typically present during
the catalytic reaction, depending on the composition of
the feed. Therefore an in-depth in situ characterization of the
catalysts also during the catalytic reaction and the activation
processes, which are needed to produce the most active
catalysts, is necessary.
7.07.1.1.2 Multifunctional behavior of mixed-metal
oxide catalysts
The large use of oxides as catalysts is related to their multi-
functional characteristics, making them suitable in a broad
range of applications, also to synthetize very complex mole-
cules whose preparation would otherwise require many steps.
 Mixed-Metal Oxides 161

From this perspective, oxide-type catalysts are key enabling
elements for the future of chemistry and industrial chemical
production, because they allow to realize, in one single step,
complex syntheses.28 This concept is one of the pillars for a
sustainable or green chemistry approach, making it possible
to reduce raw material, energy consumption, and the waste
produced.8,28
An example is the synthesis of maleic anhydride from n-
butane over vanadyl pyrophosphate [(VO)2P2O7] catalyst.15,16
This chemical is produced worldwide in several million tons
per year. The industrial process operates with high selectivities
(up to over 80% at low conversion) without desorption of any
reaction intermediate from the catalyst surface. The reaction is
a 14 e oxidation, with abstraction of 8 H atoms from butane
molecule and insertion of 3 O atoms, plus four other oxygen
atoms necessary for desorbing as water molecules the H atoms
abstracted from n-butane. The selective transformation
requires different catalytic functionalities (active sites) for (1)
abstracting paraffinic H atoms or (2) allylic H atoms, (3)
coordinating olefinic-type intermediates p-allyl complex, (4)
inserting O into allylic position and (5) 1,4 dienic position,
etc.6,16,29 Figure 9 reports the reaction network in n-butane
oxidation on (VO)2P2O7 and a scheme of the structure of the
catalyst (see also Figure 1).
In order to avoid desorption of the reaction intermediates
and have a selective transformation, it is necessary to have a
concerted mechanism, with the turnover frequency of each
elementary step in this complex transformation synchronized
with the other steps in the pathway of transformation.6,16 In
fact, when the rate of the catalyst of oxygen insertion is de-
creased, or alternatively the rate of the first step of n-butane
conversion to the adsorbed olefinic-type first intermediate is

Carboxylate (acetate)
2-butoxide 2-butanone
R Acetic
CH3COOH acid
n-butane O O
O
O

O Methyl vinyl
Butenes O ketone
but-1-ene-3-oxide

+
but-3-ene-1-oxide
O

Butadiene

CHO Crotonaldehyde

Dihydrofuran
O

O O COX
Furan
O
O O Butirrolactone

HOOC

Maleic anhydride HOOC COOH O O

H 2O
O O Maleic acid
O

On the surface of the catalyst

(VO)2P2O7

Figure 9 Reaction network in n-butane oxidation on (VO)2P2O7 and a scheme of the structure of the catalyst.
162 Mixed-Metal Oxides
increased, alkenes (butenes, butadiene) start to be observed in
the selective oxidation of n-butane and in these conditions
there is a drop in the selectivity.30
7.07.1.1.3 Open fundamental questions in the
reactivity of oxide catalysts
The observations made in the previous sections introduce an
interesting fundamental aspect of the surface catalytic reactivity
of mixed oxides catalysts in complex multistep reactions which
are yet to be investigated in depth. A volcano-type dependence
of the activity as a function of the catalyst properties (e.g., MdO
bond strength) is the basis of the concept of catalysis (Sabatier
principle), because the optimum behavior is the compromise
between the rate and strength of chemisorption. This is still the
basis of the key molecular principle of catalysis,31 and, in a
revised and modern approach based on theoretical modeling,
is also the way to describe catalytic reactions at surfaces and the
dependence on the catalyst characteristics.32,33 However, the
underlying concept is that of the rate-limiting step, for example,
of the slow step in a catalytic surface reaction determines the
overall reaction rate. In microkinetics, overall rate expressions
are deduced from the rates of elementary rate constants within a
molecular mechanistic scheme of the reaction and the identifi-
cation of the rate-limiting step.31 In n-butane selective oxidation
on vanadyl pyrophosphate, the rate-limiting step is the alkane
activation in n-butane selective oxidation and thus it is assumed
that all the consecutive steps are not fast enough to influence the
rate-limiting step. While this approach is correct to describe the
rate of conversion, the selectivity cannot properly be discussed,
because in these complex reactions on mixed oxides the selec-
tivity depends mainly on the steps after the rate-limiting one and
thus formally not relevant from the microkinetic point of view.
Using a simplified microkinetic approach based on the concept
of a concerted mechanism which considers the steps after the
rate-limiting one,34 the selectivity passes through a maximum as
a function of the rate of the oxygen insertion, because the
increase in the concentration of adsorbed intermediates after
the rate-limiting step (concentration assumed to be negligible
and not changing in the conventional kinetic approaches based
on the rate-limiting step) modifies the relative rates of the
elementary steps in a complex surface pathway. The microki-
netic model is in agreement with the experimental data.34
The same issue is presented to describe the behavior ob-
served in the selective conversion of propene to acrolein as a
function of the degree of surface reduction of CuO (Figure 10).
This is a classical experiment that was the basis of the develop-
ment of industrial catalysts for the synthesis of acrolein and
acrylonitrile from propene. The effect was explained in terms
of the need of site isolation35 of surface oxygen sites to avoid
consecutive oxidation of the product of reaction, but it is a
kinetic problem which cannot be correctly described with the
current (micro)kinetic approaches, without considering the sur-
face dynamics of the active sites and of the adsorbed species.6
Therefore, there is the need to develop more advanced
microkinetic approaches for complex catalytic transformations
over mixed-metal oxides. This is a major challenge from the
fundamental point of view and is also relevant for a better
design of the industrial catalysts.
Mixed oxides are thus characterized from the presence of a
multifunctional catalytic reactivity, and this surface chemistry
100
Relative propene conversion (%)
3CO + 3H2O Propene oxidation on CuO
80
Acrolein
60 + H2O
40
20
0
0 20 40 60 80 100
Surface reduction (%)
Figure 10 Propene oxidation to acrolein as a function of the degree of
surface reduction of CuO. In the inset are Monte Carlo models of the
distribution of sites capable of oxygen insertion (blue circles) with
respect to reduced sites (open circles). Adapted from (a) Grasselli, R. K.
Top. Catal. 2001, 15, 93101. (b) Grasselli, R. K. Top. Catal. 2002, 21,
7988.
richness is a key feature of their use in a large variety of
applications as well as in multistep catalytic reactions and to
realize sustainable processes; on the other hand, this surface
chemistry richness is also still a limit for the understanding of
the reaction mechanisms of the catalytic reactions. Different
from the surface-science approach of metals and supported
metals, models of heterogeneous catalysis based on well-defined
oxides were studied only recently and they are able to catch only
part of the complexity of the real systems,36,37 in particular, to
describe the multiple functional properties of mixed-metal ox-
ides (acidbase, redox, electron transfer and transport, chemi-
sorption by s- and p-bonding of hydrocarbons, O-insertion and
H-abstraction, etc.), the dynamics of surface species, and recon-
struction of the active sites.3840
The need for a better understanding and modeling of the
complex surface dynamics of catalytic reactions over mixed-
metal oxides is also expected, which involves both a continu-
ous surface reconstruction of the active sites associated also to
the direct incorporation of lattice oxygen in most of the rele-
vant reactions and the modification of the surface reactivity
associated to chemisorption (see also later).6 It is thus neces-
sary to develop in situ (operando) methodologies to character-
ize these surface phenomena, able to discriminate between
active and spectator (or slowly reacting) surface species.
Figure 11 summarizes this concept showing how it is necessary
to pass from a localized catalysis at a single ensemble site
(Figure 11(a), outlining the reaction mechanism for the con-
version of propene to acrolein on bismuth-molybdate catalysts)
to new, more complex models of the surface reactivity at oxide
surfaces (indicated as living active surface; Figure 11(b)).6
Although several evidences are already present in the literature
on the aspects summarized in Figure 11(a) and 11(b), the
comprehensive model of the surface reactivity which includes
these aspects in a surface microkinetic model of general validity
for catalysis at mixed-metal oxides is still far to be realized. This
is a great challenge for future studies (see also previous discus-
sion regarding the structural dynamics of the catalysts, with
reference to VPO catalysts (Section 7.07.1.1.1) and Figure 5).
 Mixed-Metal Oxides 163

CHO

O2
O O
O O O O
CHO M1 M2 M1 M2
O O O O O O

M2 M2

O O
O O O O
O2 M1 M2 M1 M2
O OO O OO
M2 M2

(a)
Long-living adspecies
possible role as co-catalysts

Influence on mobility of
O2 adspecies and reactivity

R1
O O O R Adspecies (long
O1 living, spectator,
M1 M1
Mobile reagents, etc.)
adspecies
OS R
H R2 R3
O
O O
M O
M M
M
O
Possible hopping
Microdomains influence of adsorbed
surface properties of intermediate
Bulk oxygen and
nearlying regions electron transfer

Long-range electron
(b) delocalization

Figure 11 (a) Catalysis at a single ensemble site on mixed-metal oxides (simplified reaction mechanism for the conversion of propene to acrolein on
bismuth-molybdate catalysts). (b) New models of the surface reactivity at oxide surfaces (living active surface). Adapted from Centi, G.; Cavani, F.;
Trifiro F. Selective Oxidation by Heterogeneous Catalysis; Kluwer Academic/Plenum (Springer): New York, 2000.

The control of the catalyst multifunctionality requires
also the ability to control their nanoarchitecture, for example,
the three-dimensional (3D) spatial arrangement around the
sites of adsorption. The active site is not the only relevant
aspect of catalysis. The environment around the active site
orients or assists the coordination of the reactants, may induce
sterical constrains on the transition state, and influences short-
range transport (nanoscale level).41 Therefore, it plays a critical
role in determining the reactivity and selectivity in multiple
pathways of transformation. In addition, there are indications
pointing out that the dynamics of adsorbed species, for exam-
ple, their mobility during the catalytic processes which is
also an important factor in determining the catalytic
performances in complex surface reaction, is influenced from
the nanoarchitecture.6,41
These observations have relevant consequences to bridge
the gap between model and real catalysts.42,43 In fact, in addi-
tion to the issues of pressure and material gap, the complexity
gap exists. Goodman,44 over 15 years ago, pointed out that
despite the successes in modeling catalysts with single crystals,
there is a clear need to develop models with higher levels of
complexity. Three-dimensional nanoarchitecture is one of the
critical aspects to take into account in bridging model to real
catalysts. In addition, understanding the relationship between
catalyst 3D nanoarchitecture and reactivity is a fundamental
element for the design of new-generation catalysts. On the
164 Mixed-Metal Oxides
other hand, oxides at the nanoscale level have properties quite
different from those of the same materials with larger crystal
size.45 About 20 years ago, researchers began to prepare model
systems of truly heterogeneous catalysts based on thin oxide
layers grown on single crystals of metal, but recently it was
realized that these systems may behave quite differently from
their corresponding bulk oxides (see also discussion in
Section 7.07.1.1.1).46 For example, in thin oxide layers, there
is a spontaneous charge transfer from the metal support to an
adsorbed species through the thin insulating layer (or vice
versa). On the other hand, this type of phenomena (electron
tunneling, work function changes, defects engineering, and so
forth), which are typical of semiconductor physics, may be
applied to design novel catalysts and catalytic functionalities.
A significant progress has been made recently in understanding
the atomic structure of low-dimensional metal oxides,47 open-
ing up new possibilities for heterogeneous catalysts, also in
terms of bridging the gap between the model systems for
surface-science studies and more technical catalysts.48
Most of the studies on real nanostructured oxides are based
on materials not having a well-defined 3D structure (both on
short- and long-range), being composed of irregularly shaped
nanocrystals. These materials are polycrystalline, and show
several nano-interfaces, which stabilize microstrains, oxygen va-
cancies, or metal ions in unusual coordination states. A 3D
environment for adsorption/transformation may significantly
modify the adsorption of reactants and stabilization of transition
state complexes, a well-known concept in enzymes, typically not
considered for solid catalysts. Often, the surface is far from clean,
and instead there are evidences that these surface impurities,
including carbon-based deposits, can be relevant to the catalytic
behavior. Surface science instead typically uses well-ordered,
planar-type, clean metal oxide surfaces prepared in ultrahigh
vacuum. A well-defined surface order and stoichiometry are
necessary to identify structurereactivity relations.
Therefore, even if mixed-metal oxide catalysts found a large
range of applications, there is still a wide space for their further
optimization based on a fundamental understanding and the use
of their multifunctional behavior to develop novel catalytic mul-
tistep reactions in a single passage over a solid catalyst surface.
The understanding of the reaction mechanisms and the control
of selectivity in catalysis by oxides are a relevant passage toward
this more general scope. This understanding requires, however,
an integrated effort between theory, modeling, surface character-
ization, and reactivity testing. There is a definitive trend in going
from metal to metal oxide surface science. Mainly relatively
simple catalytic transformations such as CO or NH3 oxidation
and model catalysts have been addressed up to now, although
some more complex reactions are to be addressed.
Metal oxide catalysts show a number of additional charac-
teristics quite relevant for catalysis with respect to metals,
which indicates the intrinsic higher complexity of arriving to
a full understanding of the surface catalytic chemistry. Some of
these relevant aspects of the metal oxide catalytic chemistry6
are listed as follows:
1. easy stabilization of defect structures;
2. pronounced surface relaxation;
3. surface heterogeneity (different oxygen atoms, acid and
redox sites, possible multiple cations, etc.);
4. reducibility and easy reconstruction (aggregated defects,
shear planes, etc.);
5. easy bulk diffusion (cation, oxygen, etc.) which is derived
from the presence of an open structures; often this diffusion
is anisotropic and different for the different crystal planes;
6. easy intergrowth structures;
7. anisotropic electron transfer.
Therefore, the metal oxide catalytic chemistry is characterized
from a high degree of complexity. In bridging the gap between
surface science and applied catalysis, there is a need to combine
the surface-science bottom-up approach with the top-down
approach of applied catalysis. The bottom-up approach of sur-
face science consists in modeling from (almost) first principles
and step-by-step increase in complexity, whereas the top-down
approach of applied catalysis uses complex (industrial and
model) materials and identifies key aspects in the reactivity
and surface structure. In order to bridge this gap, not only the
materials and pressure gap but also the complexity gap should
be considered.
7.07.1.1.4 Open questions from an industrial perspective
Mixed-metal oxides are used industrially in various types of
catalytic industrial processes (Table 1), but selective oxidation
is the main area of application. A number of reviews have
discussed recent development in catalytic industrial processes
of selective oxidation.8,9,4951 In addition to classical parame-
ters to metric industrial processes (yield, selectivity, spacetime
yield, and stability), the impact on environment and society
has been progressively becoming important, for example, sus-
tainability of the processes is today a major element to measure
process performances, in addition to parameters which have
gained new relevance (energy intensity, greenhouse gas emis-
sions, etc.). All these parameters are part of the general sustain-
ability criterion. The use of mixed-metal oxide catalysts is a key
element to achieve these objectives, and a series of examples of
industrial processes is given by Cavani et al.1
Even for already well-established industrial processes,
which have been progressively improved over the years, there
is still need for improvements, which in most of the cases
correspond to the need to develop novel or improved mixed-
metal oxide catalysts. Table 3 reports examples of relevant
industrial processes using mixed-metal oxides, current results
in terms of conversion and selectivity, main achievements
made over the last years, and targets of improvements still nec-
essary, which combine a better sustainability and economics.
In fact, for example, a better selectivity means a lower consump-
tion of raw materials and higher amount of more valuable
products, and also lower environmental impact, lower energy
costs in separation and processing, etc. Most of these processes
were developed at an industrial scale around 196070. Even if a
continuous improvement over the years has been achieved, the
selectivity has still to be improved. Table 3 shows that there
are many industrial targets which require an improvement in
catalysts and thus in their design.
The potential economic impact of developing more selec-
tive catalysts, based on a better understanding and design of
mixed-metal oxides, has to be noted. A moderate selectivity
improvement in the currently used catalysts for the reactions
listed in Table 3 would correspond to a significant saving in
 Mixed-Metal Oxides 165

resources, energy, and a large potential economic impact. This
concept is remarked in Table 4, reporting the potential saving
in resources and the economic impact to achieve the selectivity
targets indicated in Table 3. It should be mentioned that this
potential saving and impact is calculated in the hypothesis that
all the actual worldwide production of these products (year
2011) may be realized with the improved selectivity indicated
in Table 3. Even assuming that today the required new catalysts
are available, it will take several years before the older ones are
substituted; nevertheless, Table 4 indicates the major impact
which can have an increase in selectivity.
7.07.2 Bulk and Surface Structure of Metal Oxides
Metal oxides are inherently complex systems, due to the possi-
ble easy deviation from the stoichiometry, with formation in
some cases of also structural defective phases, the easy accom-
modation of dopants, the formation of intergrowth phases,
and the presence of surface structures which may be signifi-
cantly different from those obtained by ideal cutting of the
crystal. In most of the catalytic reactions involving mixed-
metal oxides, there are one or more redox stages, which often
involve the incorporation of the lattice oxide which in steady-
state operations is then renewed through the adsorption and
dissociation of gaseous dioxygen. This redox cycle is known as
Marsvan-Krevelen mechanism (Figure 12(a)).52
In metal oxides, the catalyst is thus often directly involved
in the catalytic cycle, but depending on the rate of release of
lattice oxygen and reincorporation by dioxygen dissociation
under steady-state catalytic operations (Figure 12(b)), the sur-
face of the catalyst may not be fully oxidized. When extended
defects form, shear planes or other types of more localized
reconstructions in the oxide occur (Figure 12(c)).
These phenomena are easier in mixed-metal oxides with
respect to single metal oxides. In addition, the rate of reduction
and reoxidation of the metal oxide depend on the concentra-
tions of hydrocarbon/oxygen reactants and the reaction tem-
perature, besides other structural aspects (size and morphology
of the metal oxide particles, presence of other oxides, phases,
dopant or metal nanoparticles which can influence these redox
rates, etc.); the same catalyst may be effectively different during
catalytic operations depending on these aspects and chemical
potential (e.g., composition) of the reactants. This chemical
potential changes along the reactor axial profile (in fixed bed
reactor operations), and, in addition, also the temperature

Table 3 Examples of industrial processes using mixed-metal oxides, improvement achieved, and future targets

Industrial process Catalysta Conversion, Achievements Future targets

selectivity (%)

Methanol to FeMo oxides 99, 9495 Increase selectivity, better heat integration Increase selectivity to >98%,
formaldehyde (adiabatic postreactor), higher productivity structured catalysts for better
(higher methanol inlet concentration) temperature profile, increase
catalyst lifetime
Propene to acrylic BiMo 9295, 8489 Increase selectivity and productivity and Increase selectivity to >95%, use
acid (two steps) oxides MoV energy integration propane as raw material, single
oxides step-process
Propene to BiMo oxides 9799, 7583 Increase selectivity and productivity, fluid Increase selectivity to >90%, use
acrylonitrile bed reactors propane as raw material, lower NH3
oxidation
n-Butane to maleic (VO)2P2O7 7585, 6573 Replace earlier process from benzene, Increase selectivity to >90%, increase
anhydride increase selectivity and productivity productivity
o-Xylene to V-oxide supported >99, 8083 Replace earlier process from naphthalene, Increase selectivity to >92%, lower
phthalic on TiO2 increase selectivity and productivity formation by-products (phthalide),
anhydride better heat removal for higher
productivity

Elaborated from Cavani and Teles.8a

a
Main active phase.

Table 4 Feedstock saving potential and economic potential of selectivity improvements

Industrial process Assumed selectivity Capacity Potential saving of raw material (Mt yr1) Economic potential
increase (%) (Mt yr1) saving (M)
Feed CO2 equivalent

Methanol to formaldehyde 3 19 0.61 0.84 338

Propene to acrylic acid (two steps) 7 5 0.20 0.64 626
Propene to acrylonitrile 9 6 0.43 1.34 1085
n-Butane to maleic anhydride 18 2 0.21 0.65 769
o-Xylene to phthalic anhydride 10 5 0.29 1.76 518
Total 5.22 3327
166 Mixed-Metal Oxides
(H2O)
O2 Oxidized catal.
(lattice O)
M2m+ M1n+
Reduced
catalyst
Oxide catalyst
(a)
O2- O- O22-
- -
O2- M(n-1) Mn+ O2- Mn+
O2- O2- O2-
O2- Mn+ O2- M(n-1) Mn+
(b)
Figure 12 (a) Redox cycle in selective oxidation on metal oxides, known as Marsvan-Krevelen mechanism. (b) Mechanism of incorporation of oxygen
in metal oxides. (c) Generation of point defects during incorporation of lattice oxygen in organic molecules, and mechanism of compensation of point
defects by change of linkage by coordination polyhedra.
axial profile along the catalytic bed is not uniform, further
influencing the catalyst effective surface state. For this reason,
one of the options currently explored in industrial reactors is to
use graded catalytic bed, with the composition of the catalyst
changing along the catalytic bed. A better understanding of
these aspects would allow one to put on more rational bases
to the optimization of the catalyst bed in industrial reactors.
It is thus not surprising as to how the surface catalytic
chemistry of mixed-metal oxides may be still considered a
challenge, particularly for more complex and realistic catalytic
reactions.
While the bulk properties of simple binary oxides are well
understood and there are excellent reviews and books discussing
the thermodynamics, structure, and its nonstoichiometric as-
pects (particularly important for oxides), spectroscopy, as well
as transport and other relevant properties, rather little is known
about the surfaces of oxides, even the most simple ones during
catalytic reactions. Some significant progress in this direction
has been made recently.53,54 These progresses in the characteri-
zation of the surface structure of oxides also open new possibil-
ities for the understanding from the theoretical point of view,
even if still more attention has to be given to a number of
simplifications such as the use of ordered O vacancies to
model reconstructions, avoiding the inclusion in modeling of
other molecules such as water or surface impurities, the effect of
the bulk in inducing surface reconstructions, the effect of gas
phase chemical potential in surface reconstruction, etc.
Transition metal oxides exist in many different crystallo-
graphic forms. There does not appear to be a simple generali-
zation that relates the structure to the stoichiometry or the
position in the periodic table, although in general the ionic
radii of transition metals are smaller than that of O2 and thus
Oxygenate H
H
C C H
H
C
O
Hydrocarbon
O2
Generation of
point defect
O2-
-
O2-
Compensation by change
of linkage of coordination
polyhedra
O2- More extended
clustering
(C) of defects
the oxygen ions are usually close-packed with the smaller metal
ions situated in the octahedral and tetrahedral holes among the
oxygen ions. Oxides commonly studied as catalytic materials
(oxides of Ti, V, Fe, Mo, Cu, Co, Zn, etc.) belong to the structural
classes of corundum, rocksalt, wurtzite, spinel, perovskite, rutile,
and layer structure, but often the positions of the ions may not
be at the ideal positions of the highest symmetry.
Many transition metal ions possess multiple stable oxida-
tion states. This is a relevant aspect for their reactivity (e.g., in
V, Mo, and W oxides) and it also introduces the possibility of
different stoichiometries. Often in catalytic relevant oxides, the
structure differs only slightly from each other making possible
the easy conversion of one oxide to another of adjacent stoi-
chiometry by oxidation or reduction, and/or the stabilization
by topological effects of a phase with respect to others. These
aspects are important to understand, for example, the surface
chemistry of monolayer type oxides supported over other
oxides (see Chapter 7.06), such as vanadium oxide supported
on TiO2, and industrial catalysts for selective oxidation
(o-xylene to phthalic anhydride) and for the catalytic selective
reduction of NOx with ammonia in the presence of O2. The
easy oxidation and reduction as well as the existence of cations
of different oxidation states in the intermediate oxides are
important factors for these oxides for their use in selective
oxidation catalysis.
7.07.2.1 Long- and Short-Order Ordering: The Role
of Microstructure
In mixed-metal oxides, often more than one phase is present,
and in this case microstructures may develop. Often the con-
cept of microstructures has not been properly addressed in

catalysis by mixed-metal oxides, while it has been more exten-
sively addressed in the ceramic and metallurgic literature.55
The microstructures may develop when two or more phases
grow simultaneously. If the growth is ideally unconstrained, a
mixture of crystals of the two phases forms which resembles
that of a mechanical mixture. However, the growth is typically
not unconstrained, and the presence of the other species may
change the crystal habits of the products and this could have a
profound effect on the catalytic behavior.
When the size of the phase regions is so small to be unde-
tectable by XRD, it may appear to be present a solid solution.
On the contrary very small micrograins are present. More
complex microstructures will result if solid-state phase trans-
formations occur. Even true solid solution may unmix to form
micrograin regions (which are thermodynamically more sta-
ble) of difficult detection, but which creates a lot of interface
regions with clear consequences on the catalytic behavior.
Often a true solid solution in the fresh catalyst transforms on
the surface layer to a microstructure due to the progressive
transformation during the catalytic reaction. The cited VPO
catalyst for n-butane oxidation has quite a long (around 1
month or more) activation time in industrial reactors and
part of the transformation occurring during this activation
time is related to the formation of these surface microstruc-
tures, although the same remains XRD crystalline. 56
The microstructures developed in materials are also im-
portant in understanding the chemistry of the preparation of
mixed-metal oxides. For example, crystallization close to phase
boundaries may result in different phase microstructures. Thus,
apparently minor changes in the preparation may occur in
apparently similar materials, but with rather different catalytic
performances. Patent literature of mixed-metal oxides for catalytic
applications is rather rich of similar unexplained phenomena.
The formation of nonstoichiometric phase is another im-
portant aspect in catalysis by mixed-metal oxides and also is
often critical to understand the in situ evolution of the catalysts
during the catalytic reaction. Sometimes nonstoichiometric
phases formed during the process of doping do not indicate
a change in the structure but the presence of low levels of
dopant elements localized on the surface to block reactivity
on some sites (e.g., doping with alkaline metals to neutralize
surface acid sites) or modify surface reactivity (favoring oxygen
dissociation for example). During the reaction and/or thermal
treatments, these surface dopants may diffuse in the bulk and
are stabilized in nonstoichiometric positions of the oxide. In
this case, the dopant does not yield a new phase, but a single
phase which persists over a significant range of composition.
For higher concentrations of the dopant, instead, formation of
new phases (even if not XRD detectable) may occur. In non-
stoichiometric phases, the metal to nonmetal ratio varies,
while the total metal to nonmetal ratio remains constant in
solid solutions.
In several compounds, it is possible to replace one of the
atom types present, either partially or completely, while main-
taining a fixed metal to nonmetal ratio. These materials are
defined (substitutional) solid solution, and the structure re-
mains almost unchanged, but with a variation of one or more
of the cell dimensions. This is also the procedure to check the
presence of a solid solution, but monitoring whether there is a
linear correlation between cell dimension and amount of
Mixed-Metal Oxides 167
dopant. All the industrial catalysts cited in Table 3, for in-
stance, are multicomponent catalysts, with up to over 20 ele-
ments in catalysts for acrylonitrile synthesis, for example. Part
of these elements form solid solutions and/or nonstoichio-
metric phases, and these aspects are very relevant in under-
standing the catalytic chemistry of these multicomponent
catalysts.12,5759
In the solid solutions, the atoms which are involved in the
substitution are often randomly distributed, but above a cer-
tain atomic ratio ordering may occur, also favored during
annealing and/or catalytic reaction. Even when the departure
from stoichiometry remains small, point defect overstructures
(clustering of point defects) may form. The microstructure of
these phases can be considered to consist of small units of a
second structure coherently intergrown within a parent matrix.
These small units are often termed microdomains. Large order-
ing of point defects leads to crystallographic shear (CS) phases,
for example, when oxygen atoms are removed and the structure
collapses along certain directions to produce planar faults
without change in the metal coordination polyhedra. CS
phases are important in several metal oxides relevant for cata-
lytic applications such as WO3, MoO3, Nb2O5, and TiO2. The
CS planes formed during catalytic reactions may be more or
less disordered, depending on the conditions of reactions, and
the microstructures of such phases will thus depend on the
history of the sample.
Ordering of twin planes may give rise to chemical twinning
(CT) of planar boundary, thus forming complex superstructures;
when multiple phases are present (as typically occurring in
mixed-metal oxide catalysts), intergrowth phases may also lead
to complex superstructures. When these superstructures are ex-
tended, their detection by current advanced characterization
methods might be relatively easy, but often this is limited to
local effects (strains) confined to the surface, which detection is
difficult, and even more challenging is the correlation with the
catalytic activity.
7.07.2.2 The Role of Strains and Defects
A number of recent papers have addressed the role of lattice
strains on the catalytic behavior. Chuang et al.60 have discussed
the influence of lattice strain on the ceriazirconia oxides
which are the typical active support for a three-way catalyst
(TWC). In particular, they showed how the doping with Al ions
(alumina is the base support on which the ceriazirconia is
deposited) inhibits the crystallite growth of ceriazirconia (a
main factor for deactivation) due to inhibition of the mecha-
nism of strain aggregation. Wang et al.51 showed that in the
selective hydrogenolysis of glycerol to propylene glycol (on
important reaction for the valorization of glycerol waste pro-
duced during the synthesis of biodiesel by transesterification)
on CuZnO composite catalysts, the activity may be related
to the presence of microstrain in supported Cu nanoparticles
strongly interacting with ZnO. This paper of Wang et al.61
extends to a different reaction, the original concept presented
by Schlogl et al.62 that, in methanol synthesis, catalyst
(copperzinc oxides are the base components for the indus-
trial catalysts used in methanol synthesis from syngas) imper-
fections in the Cu lattice arising from twinning, fault defects,
and strain may be related to increased activity in methanol
168 Mixed-Metal Oxides
synthesis via watergas shift reaction of CO, CO2, and H2. The
modalities of preparation have a strong influence on these
effects. Schlogl et al.63 showed how the bulk structure of copper
nanoparticles can be tailored during the preparation and it en-
ables control of microstrain and particle size properties.
The formation and superstructure of defects (strains) in
mixed-metal oxide depend on the type and crystal shape of
the other phases present, due to topological and interfacial
effects.64 However, also the dimensions and shape of the par-
ticles have a role. Chen et al.65 showed how in iron oxide
nanowires and nanobelts different types of long-range-
ordering structures of O vacancies are present, depending on
the nanostructure. Due to interfacial effects, the growth of
oxide nanostructures is strongly influenced by the substrate
which may control the growth directions,66 and in turn the
catalytic properties and formation of strains. There are other
aspects related to defect chemistry of solids (e.g., stacking faults
and dislocations) which were not discussed here to avoid
divergence of the discussion. Many properties of the samples
(such as ionic and electronic conductivity) also related to the
catalytic behavior may be influenced by this defect chemistry of
solids. However, this short overview has already pointed out
how the real catalyst structure may be far from the ideal (crys-
tallographic) structure and how the complexity of the problem
is exponentially increased when multicomponent (real) cata-
lysts are analyzed, which furthermore are subjected to signifi-
cant changes during the catalytic reaction. Nevertheless, this is
the challenge to be addressed to put on more rational bases for
the design of the next-generation mixed-metal oxide catalysts.
7.07.2.3 Oxide Surfaces
The composition, size, and shape in particular can be con-
trolled uniformly for each and virtually every nanocrystal
using bottom-up chemistry techniques today.67 However,
often the catalytic behavior does not depend on the presence
of well-defined facets,67 but particularly in mixed-metal oxides
it relies, in a far more complex way, on the dynamics during
reaction and the presence of microstructures (the definition
used here includes all the range of phenomena from localized
to extended defects, intergrown and nanograin interfaces, etc.).
[110]
-
[110]
Max. relaxation-11.9%
[001]
(a)
Figure 13 (a) Surface structure TiO2(110) rutile, and (b) modification of rutile VSbO4 due to growing onto Sb2O4. (a) Adapted from Charlton, G.;
Hoowes, P. B.; Nicklin, C. L.; Steadman, P.; Taylor, J. S. G.; Muryn, C. A.; Harte, S. P.; Mercer, J.; McGrath, R.; Norman, D. Phys. Rev. Lett. 1997, 78,
495498. (b) Adapted from Centi, G.; Perathoner, S.; Trifiro, F. Appl. Catal. A: Gen. 1997, 157, 143172 and Centi, G.; Trifiro, F. Catal. Rev. Sci. Eng.
1986, 28, 165184.
The understanding of these effects is also important to de-
velop multifunctional materials which are the basis to design
more complex and sustainable catalytic reactions, as introduced
before. As an example, it may be cited the recent finding68 that a
Zn-in-Cu near-surface alloy covered by a thin wetting layer of
interfacial Zn(ox) is the most active state of an inverse CuZn
catalyst in hydrogen production by methanol steam reforming,
another industrially relevant reaction. The bifunctional action of
the mixed Cu(Zn)O/Zn(ox) surface allows for selective dehy-
drogenation of methanol to formaldehyde and for optimized
water activation, thus providing the required source of oxygen
for the total oxidation of HCHO to CO2.
These bulk effects clearly reflect on the surface, particularly
for small particles, and in turn on the effective nature of the
catalytic active sites present during the reaction. It is known
that even for simple single oxides such as TiO2,53 the surface
cannot be simply predicted from the cutting of the crystal
structure and saturation of the surface vacancies. In TiO2, it
was determined by surface XRD the presence of structural re-
laxations (magnitudes range 00.27 A), leading to rumpling
of the Ti planes (Figure 13(a)).69 Similar surface reconstruc-
tions may be induced from the growing onto other oxides.
Figure 13(b) shows the model of how the surface of rutile
VSbO4 is modified in a similar way to that reported for TiO2,
due to the formation of a trirutile-like over-structure.70 This
modification of the surface structure of VSbO4 was suggested to
be responsible for enhanced catalytic properties in propene
and propane ammoxidation.70 Similarly, as discussed also
later, ultrathin oxide films (as typically used in surface-science
studies) show relaxation phenomena strongly induced from
the substrate and they are thus not representative of the situa-
tion present in catalytically active oxide nanoparticles.
This is a general issue in catalysis by metal oxides and there
is a strong modification of the oxide surface induced from the
particle size and interfacial effects with other nanoparticles. For
example, the role of ZnO in methanol synthesis catalysts, to
modify by interface of active Cu nanoparticles, was discussed
before. In agreement, it was shown recently that the micro-
structure of the ZnO/Al2O3 component influences the catalytic
activity of a Cu/ZnO/Al2O3 industrial catalyst.71 ZnO crystal-
lizes in the wurtzite structure. The oxygen sublattice forms
hexagonally close-packed layers with hcp-type stacking, and
de
da
(b)

the Zn ions are located in tetrahedral interstices; the tetrahedral
units are edge and corner sharing. At the surface, the minimi-
zation of the surface free energy is what drives the energetics.
When crystal is cut normal to a randomly chosen direction,
there is an overwhelming probability that the resulting surface
corresponds to a polar termination and is highly unstable.72
Polar oxide surfaces, such as ZnO, are subject to complex
stabilization processes that ultimately determine their physical
and chemical properties. When the surface is terminating with
Zn, many small, triangular islands and holes form; these
islands have a height of half a unit cell, and each step edge
contains only oxygen atoms. This provides excess negative
charge on the Zn-terminated surface and stabilizes the surface
for a wide range of oxygen chemical potentials.73 On the
oxygen terminating surface, a different stabilization mecha-
nism is operational. The surface is flatter; the islands are
hexagonal; and most step edges are one unit cell high and
contain both Zn and O atoms.74 However, as pointed out by
Goniakowski et al.,72 the rules for surfaces polarity compensa-
tion are lifted in the case of ultrathin films and nanostructures.
Therefore, in small nanoparticles, particularly when many
interfaces are present which could induce different mecha-
nisms of surface relaxation, the surface termination and recon-
struction would be different from those present in larger
crystals and even in the latter may be quite different from the
ideal situation obtaining by cutting the crystals along some
crystallographic planes.
A further aspect to consider is that the adsorbate may in-
duce surface reconstruction and changes in the surface reactiv-
ity. Limiting to ZnO as an example, there are ample
demonstrations which indicate that the presence and nature
of adsorbate influence the stability of the terminating surface
and induce a reconstruction. For example, for the clean
OZnO surface a new reconstruction has been observed exper-
imentally, a (1 3) O-vacancy structure which is very reactive,
in particular, toward water forming a fully hydroxylated sur-
face, HOZnO.75 The question is how to translate these results
from surface science to catalytically active materials and their
surface reactivity. However, significant progress has been made
recently in this directly. Sautet et al.,76 for example, showed
how the surface reactivity of alumina depends greatly on the
surface coverage of water. A low OH coverage, resulting from
pretreatment at 700  C, generates metastable three-coordinate
AlIII sites and reactive (AlIII,O) Lewis acidbase pairs on
g-alumina. Water plays a dual role by increasing the basicity
of oxygen atoms without affecting the Lewis acidity of the AlIII
sites and by stabilizing the metastable (110) termination,
where these sites exist.
In mixed-metal oxides, it is thus very challenging to bridge
the gap between model ideal systems for surface and theoret-
ical studies and applied catalyst. It worth mentioning the effort
to develop a single-crystal flow reactor77 to allow studying the
catalytic reactivity of the same materials studied by surface
chemistry methodologies, but previous discussion has shown
that the surface of these mixed-metal oxides may be quite
different from those practically used. There are also additional
aspects recently discovered. For example, it was also shown
that in these ultrathin oxide layers it may also be possible for
a spontaneous charge transfer from the metal support to an
adsorbed species through the thin insulating layer (or vice
Mixed-Metal Oxides 169
versa).46 Thus, the mechanism of adsorption itself may be
quite different. It is thus necessary to develop specific mixed-
metal oxide materials which on the one hand show a more
ordered situation with respect to practically used catalysts, and
on the other hand overcome the limits shown from the use of
ultrathin metal oxides typically used in surface-science studies.
Among the approaches proposed to bridge this gap is the use of
an ordered array of 1D nanostructures48 or the use of micro-
crystallines oxides.78 Recently, the use of metal-free (oxidized)
nanocarbon (sp2) materials was also proposed to mimic the
active sites present in metal oxides. In some reactions of selec-
tive oxidation, they show good catalytic performances, but the
absence of polyvalent metal sites with complex electronic and
spin structures allows for facile and in-depth theoretical
modeling and thus a better understanding of the reaction
mechanism. In particular, Frank et al.79 discussed the reaction
mechanism of selective C3H4O oxidation, and the importance
of nucleophilic and electrophilic oxygen species, the prismatic
edges as well as the basal planes of graphite, and of protons
delivered either by carboxy groups or by water. The authors
identify substantial similarities between the metal oxide and
the carbon-catalyzed reaction, indicating that this could be a
promising approach to better understand the mechanism of
the reaction and surface catalytic chemistry of metal oxides.
Well-defined single-site catalysts prepared by surface organo-
metallic chemistry (SOMC) on oxides80 is also another
valuable approach for a better understanding of the catalytic
chemistry of metal oxides. However, often the limit is the
difficulty in the passage from single reactions, as may be
investigated in these approaches, to the complex multistep
catalytic chemistry which is the peculiar and still unique char-
acteristics of mixed-metal oxides, as discussed in the previous
sections.
Significant development has also been made in recent
years in the synthesis of well-defined metal oxide materials,
for example, shape-controlled nanocrystals for catalytic
applications,81 (multi)layered films by atomic layer deposition
(ALD),82 high-quality metal oxide nanoparticle with con-
trolled defects,83 and other synthesis methodologies. Some
useful indications on these aspects are reported in recent
books on selective nanocatalysts and nanoscience,2 and metal
oxide nanomaterials.84,85 Nevertheless, it is still a challenge to
apply these advanced synthesis methods in the preparation of
mixed-metal oxides used in the industrially relevant catalytic
reactions, cited in Table 3, for example.
7.07.3 Low-Dimensional Metal Oxide Nanostructures
Building a defined nanoarchitecture in oxide-type materials
further extends the concept of nanocatalysis,84,86 for example,
when the electrons are confined, and physical and chemical
properties are not scalable from the bulk properties. In
clusters/metal particles in the nanometer length scale, it is
known that the presence of quantum size effects and structural
fluxionality (capability of small clusters to exhibit several struc-
tural forms of comparable energy) markedly influence the
reactivity.86 It is also known that the confinement of small
metal oxide particles within a host ordered nanoporous struc-
ture (e.g., silica micro- and mesoporous materials8789) may
170 Mixed-Metal Oxides
change their characteristics and reactivity. Oxide nanoparticles
introduced within an ordered mesopore structure90,91 act as
discrete units. As the particle size is scaled down to a few
nanometers, and stabilized within an ordered micro- or meso-
porous structure, the microstructure could be significantly
altered, changing the surface-to-volume ratio and particle-to-
support interaction.41,92 The misfit between the oxide particle
and the host ordered porous material could stabilize defects,
create unusual coordination states, and significantly alter the
metal oxide properties and reactivity.90,91
7.07.3.1 Metal Oxides Confined within a Host Ordered
Matrix
Vanadium-oxide nanoparticles confined with the ordered
structure of a microporous material (silicalite with MFI struc-
ture) show enhanced activity and selectivity in propane oxida-
tive dehydrogenation (ODH), but the limit is the need for very
low loading of the vanadium to have high selectivities (up to
above 80% for V-oxide nanoparticles below 1 nm, in contrast
to below about 20% for larger particles, above 45 nm).93
Figure 14 shows this concept reporting how the decrease of
the loading of vanadium in silicalite-1 microporous material
leads to an increase in the specific activity (per atom of V;
turnover number) and selectivity in propane ODH. Confine-
ment within the host zeolite is necessary to avoid the otherwise
fast sintering of the vanadium-oxide nanoparticles during the
reaction. The total catalyst productivity (per volume of cata-
lyst), however, is low for practical applications, but this was
among the first examples showing how extremely small metal
oxide nanoparticles may have completely different catalytic
performances with respect to larger particles.
A further example is shown by the behavior of iron-
molybdate-based catalysts for the selective side-chain oxida-
tion of substituted toluene to the corresponding substituted
Moles s-1 V-1atoms
0.04
+ O2
0.035 CO2 + H2O
0.03
0.025
0.02
100 200
Figure 14 Specific activity per V atom and selectivity to propene in the propane oxidative dehydrogenation at 500  C on V-oxide nanoparticles within
silicalite-1 (MFI) microporous materials. In the inset is a model of V-oxide nanoparticles active in propane oxidative dehydrogenation. Adapted from
Centi, G.; Perathoner, S.; Vazzana, F. In Catalysis by Unique Ion Structures in Solid Matrices NATO Science Series II: Mathematics, Physics and
Chemistry; Bell, A., Centi, G., Wichterlowa, B., Eds.; Chapter 11, Kluwer/Academic Press: New York, 2001; pp 165186.
benzaldehyde. Reported in Figure 15 is an example for the
challenging reaction of conversion of 3-fluoro toluene to 3-F
benzaldehyde, a valuable product for industrial production of
fine chemicals.94 The presence of the F atom in meta position
has a strong inhibition effect on the selectivity/activity and this
reaction shows low selectivity in both liquid- and gas-phase
selective oxidation. Figure 15 compares the selectivity to benz-
aldehyde as a function of the conversion on a series of bulk
iron-molybdate (pure or doped with Bi and Sb to improve the
performances) and with respect to the behavior shown by
small iron-molybdate nanoparticles inserted in a microporous
material with MFI structure. The microporous material in this
case is a boralite (analogous to ZSM-5 but having B instead of
Al ions) which has been deboronated to create a defective
silicalite-1 (MFI structure) hosting microporous material for
the small iron-molybdate nanoparticles.94 Confinement within
the host ordered matrix leads to enhanced catalytic properties,
and Raman data confirmed that this may be associated with a
deformed iron-molybdate structure induced from the hosting
matrix.94 However, from a practical perspective, the stabilization
within the hosting matrix induces problems of diffusional lim-
itations which negatively affect the catalytic behavior, besides
the issue remarked earlier of the low productivity per volume of
the catalyst.91
These data introduce the concept which is known for many
years that low-dimensional metal oxide nanostructures may
have interesting catalytic properties. In recent years, several stud-
ies have been focused on metal oxide particles within a host
zeolite (or mesoporous material) to change their properties.95
Metal oxide nanoparticles within host zeolites are attractive to
develop novel photocatalysts,96 to improve the shape-selective
alkylation of toluene with propylene to produce p-cymene,97 to
develop catalysts for the disproportionation of toluene98 and
the alkylation of benzene,99 methane total oxidation100 and
diphenyl carbonate synthesis,101 and other applications.
Selectivity to propylene (%)
+ H2O T = 500 C
80
60
40
20
300 400 500 600
SiO2/V2O3 ratio
Increases the V loading

100
Selectivity of 3 fluoro-benzaldehyde (%)
80
60
40
FeMo/Boralite + Mo
20 Bulk FeSbMooxide
Bulk FeBiMooxide
Bulk FeMooxide
0
0 20
Conversion of 3-fluoro-toluene, %
Figure 15 Selectivity as a function of conversion in the selective oxidation of 3 F-toluene to 3 F-benzaldehyde in bulk pure and doped iron-molybdate
catalysts and in iron-molybdate nanoparticles stabilized within a hosting MFI silicalite (derived from deboronation of a boralite). In the inset is a model of
iron-molybdate nanoparticles active in 3 F-toluene selective oxidation. Adapted from (a) Centi, G.; Perathoner, S. Top. Catal. 2001, 15, 145152; (b)
Centi, G.; Perathoner, S.; Tonini, S. Top. Catal. 2000, 11/12, 195204; (c) Centi, G.; Perathoner, S.; Tonini, S. Catal. Today 2000, 61, 211221.
However, productivity may be often still low for industrial
applications, and the issue is to preserve the possibility of en-
hanced performances of low-dimensional metal oxides nano-
particles, without the need of a hosting structure that avoids
their sintering. A possible approach to solve this issue, also
opening new perspectives, is the synthesis of the metal oxide
particles with defined low-dimensional characteristics such as
nano-rows, -tubes, etc., which improve stability against sintering
with respect to irregular nanoparticles, due to the presence of
a specific nanoarchitecture.
This approach has a number of other possible advantages
with respect to the stabilization of metal oxide nanoparticles
within an ordered host matrix. The first regards the fact that
often the catalytic cycle involves the exchange of electrons
between the reacting molecule and the active centers. This is a
critical aspect, for example, in selective oxidation reactions
where it is needed to delocalize these electrons far from the
adsorbed intermediate.6 This long-range transport is also an
important aspect for the fast regeneration of the active site, for
example, for turnover. Decreasing the size of oxide particles,
therefore, has a positive effect in terms of creation of reactive
sites, but a negative effect in terms of long-range transport
properties which also play a relevant role. This is the reason
why 1D-type nanostructures such as nanowires and nanotubes
are often preferable for oxides with respect to 0D-type nanos-
tructures, for example, spherical nanoparticles below few
nanometers. Nesper and coworkers102 have discussed in detail
how oxidic nanotubes and nanorods represent anisotropic
modules for a future nanotechnology, although with focus
on the synthesis aspects. They clearly show how these aniso-
tropic materials differ from those of the corresponding bulk
material and those of isotropic nanoparticles.
The second limit of oxide nanoparticles within ordered
nanoporous matrices is related to the fact that due to the
wetting characteristics of silica, small oxide nanoparticles, for
Mixed-Metal Oxides 171
Mo
Fe
Mo
Mo
Fe
Mo
Fe
CH3 CHO
400 C/air
F FeMo-oxide F
based catal.
40 60 80 100
example, below about 2 nm, could be stabilized only for very
low loading of the oxide, as pointed out before. A third prob-
lem is related to the slow back-desorption of the products of
reaction, when they form on metal oxide nanoparticles within
a host ordered porous silica matrix. For example, in toluene
oxidation to benzaldehyde over FeMo oxide nanoparticles
stabilized within a silicalite matrix, the slow rate of reoxidation
of the reduced FeMo oxide, due to the low nanoparticle size,
increases the presence of reduced molybdenum sites, which,
interacting with the carbonyl group of benzaldehyde, slow
down the desorption and enhance the rate of the consecutive
oxidation.94c
7.07.3.2 Ordered Assemblies of Metal Oxide
Nanostructures
The points discussed in the previous sections strengthen the
interest for catalysis toward the development of ordered assem-
blies of 1D nanostructures for oxide materials, for example,
metal oxide catalysts in which the 3D macrostructure is con-
stituted by an ordered assembling of regular 1D structures with
nanometric size. Note that this type of structure is significantly
different from that of metal oxide supported over other metal
oxides, such as monolayer-type V2Ox/TiO2 materials (see
Chapter 7.06).
These ordered array materials find interest not only in
catalysis, but also in several other applications, from optical
materials, sensors, low-k materials, ionic conductors, photonic
crystals, to bio-mimetic materials.103 However, with respect to
these applications, catalysis requires additional specific char-
acteristics, such as the presence of a thermally stable nanostruc-
ture, the minimization of grain boundaries where side
reactions may occur, and the presence of a porous structure
which guarantees a high surface area coupled to low heat and
mass transfer limitations. An ordered assembly of 1D
172 Mixed-Metal Oxides
nanostructures for oxide materials could, in principle, meet
these different requirements.
Ordered mesoporous oxides also offer interesting opportu-
nities as novel materials for catalytic applications. They can be
prepared using different methodologies suitable in principle
for scaling-up (even if costs are still high), such as by block
copolymer-templated,104 micelle-templated,105 inverted opals,
colloidal templating or double templating procedures,106 and
cooperative self-assembly methodologies.107 Wan et al.108
have recently discussed the development of ordered nonsilic-
eous mesoporous materials based on surfactant assembly and
confined space growth. The materials prepared by these meth-
odologies include not only mesoporous metal oxides, but also
polymers and carbons with open framework structures, and
single-crystal metal, metal oxide and carbon nanoarrays with
replicated mesostructures. The same authors109 also discussed
the hostguest chemistry, and soft-template versus hard-
template approaches in the synthesis of nonsiliceous meso-
porous materials. Other reviews have recently addressed the
synthesis of low-dimensional, ordered, nanostructured metal
oxides. Gao and Li110 discussed the synthesis of nanostruc-
tured and low-dimensional transition metal oxides such as
ZnO, TiO2, and V2O5. Zhu et al.111 analyzed the synthesis,
property, and structural characterization of perovskite oxide
nanowires, materials with a broad spectrum of applications
going from catalysis to switchable polarization, piezoelectric-
ity, pyroelectricity, and nonlinear dielectric behavior. It is well
demonstrated that, at nanoscale, perovskite oxide materials
exhibit a pronounced size effect manifesting itself in a signifi-
cant deviation of the properties of low-dimensional structures
from the bulk and film counterparts. Ostrikov et al.112
reviewed instead the self-assembled low-dimensional nanoma-
terials via low-temperature plasma processing, with reference
in particular to ZnO and ultrahigh-aspect ratio Si nanowires.
It should be noted, however, that in most of the cases, these
materials were not prepared specifically for catalytic applica-
tions, and often they lack the necessary thermal stability, crys-
talline long-range order, and in particular the 1D-type
nanostructure characteristics which could differentiate the
properties of nanooxides from those of the corresponding
bulk material. An important aspect to discuss, already
remarked before, regards the effective nanostructure of these
materials and the presence of strains and defects, which were
not usually analyzed in detail, but have a marked influence on
the properties.
Cao and Wu113 discussed the strain effects in low-
dimensional transition metal oxides, evidencing how many
of the functional properties of nanostructured metal oxides
originate from intrinsic coupling between lattice deformation
and nanoscale electronic and magnetic ordering. Lattice strain
thus has a profound influence on the electronic, optical,
and magnetic properties of these materials, as well as their
reactivity. In epitaxial thin films, biaxial strains are developed
as a result of lattice mismatch between the film and the sub-
strate. By choosing different substrates, a wide range of strains
can be established at discrete values that allow for exploration
of new phase spaces, enhancement of order parameters, crea-
tion of complicated domain textures, and stabilization of new
phases. On the other hand, continuous tuning of uniaxial
strain is possible in nano/microwires, where a variety of
phase transitions and their dynamics could be probed at
the single or few-domain scale. There is thus a basic difference
in the formation mechanisms (and possibility of tuning)
of strains in thin films (2D dimensionality) and nanowires,
tubes, etc.
7.07.3.2.1 Thin oxide films
A new area of growing interest in relation to these aspects,
relevant for catalysis, is the defect nanostructure of thin oxide
films forming over metal nanoparticles. Mittendorfer114 dis-
cussed low-dimensional surface oxides in the oxidation of Rh
particles, an important catalytic element. The oxidation of
Rh particles leads to the formation of ultrathin oxide films
and stripes adsorbed on the metallic surface. These structures
display unique electronic and structural properties. Netzer
et al.115 reviewed the properties of low-dimensional oxide
nanostructures on metals, showing the presence of a large
number of undercoordinated atoms. The interaction with the
metal substrate provides constraints on their structure and
morphology, yielding elastic strain and/or uncompensated
charge. These factors modify significantly the physical and
chemical properties of the nanophases as compared to bulk
oxides. However, a systematic understanding is still missing on
how to tune the metal oxide particles using the concepts pre-
sented below and thus how to put on rational bases for
the design of advanced mixed-metal oxide catalysts with low
dimensionality, although it is clear that this is the challenge to
develop next-generation catalytic materials.
7.07.3.3 Reactivity of Low-Dimensional Metal Oxides
There are increasing demonstrations which indicate that di-
mensionality of oxides significantly influences their catalytic
behavior. An elegant demonstration was given by Ueda
et al.116 recently. Schlogl and coworkers117 showed how 1D
vanadium-oxide nanostructures, induced by nucleation over
carbon nanotubes, have different characteristics from bulk
vanadium oxide and how they are selective in n-butane oxida-
tion to maleic anhydride, while bulk vanadium oxide produces
only carbon oxides. Wang et al.118 showed that gold nanopar-
ticles supported on b-MnO2 having nanorods or conventional
particle shapes have different catalytic behavior in the liquid-
phase aerobic oxidation of benzyl alcohol. The enhanced cat-
alytic activity of the Au/MnO2-nanorod catalyst was attributed
to the beneficial presence of a higher amount of oxidized gold
species and surface oxygen vacancies resulting from the strong
interaction between Au and the well-defined reactive surface of
MnO2 nanorods. Rajeswari et al.119 showed that Pt deposited
on WO3 nanorods shows enhanced properties in methanol
oxidation. Zhong et al.120 reported that gold deposited on
a-Fe2O3 nanorods exhibited higher catalytic activity in CO
oxidation than an Au/a-Fe2O3 nanoparticle catalyst. Tungsten-
oxide deposited over titania nanotubes shows enhanced
catalytic properties in the oxidation of dibenzothiophene.121
LaSrCuO4 nanowires, prepared using carbon nanotubes
as template, show higher specific activity in CO oxidation with
respect to LaSrCuO4 nanoparticles prepared by a conventional
route.122a TiO2 nanotube-supported Cu species show higher
activity in the reduction of NO by ammonia in the presence

of O2 in comparison with similar 2 wt.% Cu/TiO2 catalysts
prepared using TiO2 nanoparticle supports.122b
Ueda and coworkers123a have recently reviewed the synthe-
sis and catalytic applications of metal oxide nanotube. The
typical synthesis method discussed for these materials is
the hard template method, a technique in which the templates
are covered with metal oxides and the templates were removed
to form metal oxide nanotubes. However, this methodology is
not suited for the preparation of catalysts in a larger scale, due
to the many steps necessary and the cost. However, the review
shows how carbon nanofibers (CNFs) may be used as tem-
plate, allowing to reduce the costs. A variety of oxide nano-
tubes such as thick, thin, and helical nanotubes reflecting the
shapes of CNFs could be synthesized. Ueda et al.123b also
showed how new crystalline solid materials of Mo3VOx having
low-dimensional structures can be synthesized in a single-
crystalline form using a hydrothermal method, as a result of
self-organization between negatively charged pentagonal units
and VO2 cations. Miyauchi and Tokudome124 reviewed in-
stead the synthesis and applications of titanium oxide nano-
tube thin films, prepared in particular by using hydrothermally
grown titanate nanotubes. A detailed discussion of the nanoarch-
itecture and reactivity of titania catalytic materials was carried out
by Centi and Perathoner in two reviews, the first dedicated to
quasi- 1D nanostructures125a and the second to bidimensional
nanostructured films.125b The application of ordered array of
vertically aligned titania nanotubes for catalytic use in solar
cells was also reported recently.126 The synthesis of metal oxide
nanotubes was also discussed by Lee et al.127 classifying the
fabrication methods into two approaches: (1) a directed method
using nanotemplates (solgel method and ALD) and (2) a non-
directed method (anodization). The latter may be better classi-
fied as a self-assembling method in the presence of a strong
electrical field.
These examples evidence the large scientific interest on 1D-
type nanostructures for metal oxide catalysts and how they
often show a higher catalytic performance related to not only
a higher external surface area, but a change in the intrinsic
reactivity properties as well. However, more solid bases to
rationalize these results and design ad-hoc nanostructures to
improve the catalytic performances are still missing.
7.07.3.4 Opportunities to Design Advanced
Catalytic Materials
Previous sections have shown not only the great poten-
tial interest of low-dimensional metal oxides, but also the
need for further studies aimed at improved understanding
toward the
tailored control of the nanoarchitecture of these materials,
besides their nanomorphology;
precise control of nanostructure, for example, local structure
and composition, presence of defects and strains, etc.;
detailed analysis of the interfaces between nanoparticles,
-grains, etc. and the effect of these interfaces on the surface
relaxation phenomena and structure;
analysis of the charge, heat and mass transfer phenomena in
these nanomaterials, and how these effects are influenced by
the nanoarchitecture, -morphology and -structure; and
Mixed-Metal Oxides 173
study of the stability of these materials during catalytic reac-
tion, and investigation of the relationship between nanodi-
mension and reactivity.
In recent years, the research focused mainly on the devel-
opment of novel methodologies to synthesize oxide nano-
wires, -tubes, and -rods.102,128131 Most of these studies were
focused on the preparation of 1D type materials of zinc, tin,
vanadium, or molybdenum-oxide, while fewer were related to
ternary oxides.132 A review on the synthesis, characterizations,
properties, and applications of 1D metal oxide nanotubes,
nanowires, nanoribbons, and nanorods has been recently
published.133 Most of the work refers to applications for elec-
tronics, sensors, and solar cells. One of the few exceptions is
the work of Pan et al.134 on CeO2 nanorods, nanotubes, nano-
wires, and nanocubes for the oxidation of CO at low temper-
atures. It was shown that CeO2 nanotubes have the best
performance due to the large BET surface area and the presence
of inner (inside nanotube) surface, while CeO2 nanorods show
a better stability and preferentially exposed (110) crystal
planes. This work thus mainly reports a phenomenological
analysis of the relationship between nanostructure and cata-
lytic behavior. In general, very limited attention has been given
in the literature to the analysis of nanostructurefunctional
properties relationships. The use as catalysts was often men-
tioned, but not specifically investigated. These materials are
used as catalysts only because they are novel, with limited
effort to clearly discriminate when they offer real advantages
or when the effect is only apparent, for example, due to a
higher surface area. A higher activity does not necessarily
imply a better catalyst, because, for example, when the
activity is higher due to a double surface area, the cost of
preparation is four times higher and the cost-effectiveness is
worse. There are many other aspects to consider (stability,
scaling-up, etc.), but in general it should be analyzed in detail
when the use of new methods to synthesize nanostructured
materials offers real advantages for the preparation of novel
catalysts.
In nanocrystalline materials, the electrons are confined in
regions having one, two, or three dimensions, when the rela-
tive dimension is comparable with the De Broglie wavelength
of free electron which, for oxides, is a value around 510 nm.
The nanostructures of metal crystals having the z-direction
below this critical value (thin film, layer structure, and quan-
tum well) are defined as 2D nanostructures. When the dimen-
sion, both in the x- and z-direction, is below this critical value
(linear chain structure and quantum wire) the nanostructures
are defined as 1D, and when the y-direction is also below this
threshold (cluster, colloid, nanocrystal, and quantum dot) the
nanostructures are defined as 0D. There is thus a progressive
quantum confinement, which modifies the corresponding di-
agram of the density of states. The term 1D nanostructures
comprises different types of nanoordered materials, such as
nanorods, -wires, -coils, -fibers, -pillars (or -columns), and
-tubes. Note that, often, the diameter of nanowires is above
the threshold indicated earlier, but due to the high aspect ratio
(length-to-diameter ratio), these materials are usually consid-
ered 1D-like nanostructures. In nanotubes, the wall thickness
is instead the relevant parameter and this value is often below
about 10 nm.
174 Mixed-Metal Oxides
These 1D nanostructures may then form a 3D macrostruc-
ture in which these units may have a random orientation or a
specific alignment in one single direction, to form quasi-2D
nanostructures. In the latter case, there are thus strongly aniso-
tropic characteristics which are of interest not only for appli-
cations in electronics and solar cells, but also for catalysis and
photo- or electrocatalysis.
One-dimensional-type nanostructures may have preferen-
tial paths for transport electrons (aspects important in photo-
and electrocatalysis) and the presence of ordered arrays reduces
grain boundaries, eventually minimizing side reactions which
may occur in unordered oxide nanoparticles. It is possible to
realize high surface areas together with a reduced presence of
microporosity, which can be negative for selectivity in various
catalytic reactions. In a 1D-nanostructure, specific crystalline
planes can be maximized. There are thus various motivations by
which the intrinsic reactivity of metal oxide catalysts depends on
the specific nanomorphology of crystals. As commented above,
oxide nanotubes are thus intrinsically different from nanowires
and analogous nanostructures, in terms of quantum confine-
ment, charge transport, etc., but this difference is not adequately
recognized in the literature.
These low-dimensional oxides are interesting not only for
their possible intrinsic different reactivity (with respect to con-
ventionally prepared oxides), but also for additional possibilities
illustrated in Figure 16(b) reporting a conceptual example of a
nano-architectured surface formed by an ordered array of metal
Conventional
catalysts
(b1) 200 nm
(a)
New concept
200600 nm
2080 nm Nanoreactor
(b)
Surface
active sites
Nanoconfined
liquid
Anchored
organometallic
complexes
Homogeneous
complexes
Bio-catalyst
(c) Oxide nanotubes
Figure 16 Concept of nanoengineering of oxide catalytic surface in
terms of nanoreactor array, some of the possibilities offered by this
concept (in particular in terms of realizing multifunctional catalysts for
cascade reactions in nanoconfined liquids) and an FESEM image of an
array of TiO2 nanotubes produced by anodic oxidation of Ti foils. See text
for details. Adapted from Centi, G.; Perathoner, S. In: Catalysis, Vol. 21;
Spivey, J. J., Ed.; RSC: Cambridge, 2009; pp 82130.
oxide nanotubes which can be viewed as an ensemble of nanor-
eactors. The field emission scanning electron microscopy
(FESEM) image in Figure 16(b1) shows an example of the TiO2
nanostructures obtained in the case of anodic oxidation of tita-
nium foils,135 to remark the concept that these materials could be
effectively prepared. Figure 16(a) shows the situation present in
conventional oxide catalysts, whereas Figure 16(c) shows the
concept of nano-confinement of liquids discussed below.
Some of the possibilities offered from this nanoreactor
structure are also outlined in Figure 16 and may be summa-
rized as follows:
Nano-confinement. It is known that nematic liquids within
mesoporous SBA-15 host materials (a silica mesoporous
materials formed by an array of ordered channels with size
around 410 nm) evidence a change in the phase transition,
when confined within the mesoporous cavities.136 Dosseh
et al.137 studying the properties of cyclohexane and benzene
confined in mesoporous silica materials (MCM-41 and SBA-
15) also evidence a change in the physical properties of the
liquid related to the influence of the inner surfaces of meso-
porous silica. Other authors have further demonstrated the
influence of confinement on the adsorption and properties
of fluids within an ordered mesoporous material.138,139
Fajula and coworkers140 showed that an enthalpic excess is
present during the adsorption process, due to a wall effect
(confinement) in mesoporous materials, and that this con-
finement effect influences the catalytic behavior. Chen
et al.141 showed the presence of confinement effects during
the catalytic reaction over mesoporous silica. The role of
confinement effects in catalysis by microporous materials
was reviewed recently by Sastre and Corma.142 There are
thus many evidences that a mesoporous channel may induce
a modification in the properties of hosted fluids, due to the
electrostatic field generated by the channel walls and other
motivations (for example, it was demonstrated recently that
wall curvature influence the molecular orientation of the
transition states143). The effects are different from those
related to geometrical constraints induced on complexes
anchored in mesoporous channels, or shape-selectivity ef-
fects present in zeolites. Nano-confinement within a host
oxide nanotube may thus modify the reactivity. However,
there are a number of other possible effects which change
the reactivity. As shown in Figure 16, each oxide nanotube
may be viewed as a single nanoreactor. There is an increasing
research interest on exploring the benefits of nanoreactor
vessels to enhance catalytic reactivity. Ouali and Caminade144
discussed the influence of the confinement of the catalyst
inside the dendrimers on the catalyst activities and selectiv-
ities. He et al.145 showed that Pd-confined catalysts possess
higher turnover frequencies in toluene oxidation. Bao and
coworkers146 showed that FeN nanoparticles within carbon
nanotubes exhibit a five to seven times higher activity than
unconfined nanoparticles. Bao and his coworkers were pio-
neering the research to show how the reactivity properties of
nanoparticles are modified by the confinement within carbon
nanotubes.147 Confined CeO2 particles with titanate nano-
tube show enhanced resistance to deactivation.148 Therefore,
even if data are far from systematic, there are growing evi-
dences to show that confinement of reactants and

nanoparticles within an ordered mesoporous matrix can be
a novel approach to catalysis. A bidimensional array of
oxide nanotubes as that schematically shown in Figure 16
(b) shows some advantages over mesoporous silica. In fact,
in mesoporous silica the elongation factor of the channels,
for example, the ratio between the length and the diameter
is typically a factor with a value of at least 100 or higher,
while it is lower in a bidimensional oxide nanotube array
as schematically shown in Figure 16. Furthermore, it is
difficult to realize a mesoporous silica film with the chan-
nels vertically aligned with respect to the surface. Both
these aspects are relevant for full accessibility of the inner
surface, for example, to eliminate diffusion limitations. By
tuning the characteristics of the metal oxide and its surface
hydrophilicity, the local electrostatic field within the nano-
tube and the adsorption properties may be changed to
optimize the performances, and realize a selective adsorp-
tion of nonpolar reactants in a polar medium, for example.
There are thus multiple possibilities to realize novel cata-
lysts with tailored reactivity.
3D geometrical architectures of the active centers. While in
enzymes and homogeneous catalysis there are many evi-
dences showing how the supramolecular structure is
influencing the reactivity of an active site,149,150 the role of
3D architecture around the active sites in solid catalysts is
much less investigated and understood. Figure 16(c) shows
that oxide nanotubes can host metal nanoparticles, an-
chored as metallo-organic complexes, and enzymes as well,
to develop a multifunctional and highly efficient nanoreac-
tor, easily accessible by reactants. Interesting recent results
have been reported for cascade catalytic reactions in poly-
meric nanoreactors.151 Catalysts based on metal oxide nano-
tube arrays could offer new possibilities to realize robust
and recyclable materials for cascade catalytic reactions. In
a recent review152 catalytic reactions within a confined nano-
space (molecular capsules, zeolites, and micelles) have been
discussed, with reference in particular to self-assembled nano-
capsules based on noncovalent bonds and used as nanoreac-
tors for various types of organic and metal-catalyzed reactions.
Integration of homogeneous, heterogeneous, and bio-catalytic el-
ements. Figure 16(c) shows how the array of metal oxide
nanotubes offers interesting possibilities to integrate in a
defined nanospace organometallic complexes, metal parti-
cles, other active species (Brnsted or Lewis sites, transition
metal ions coordination sites, etc.), and bio-catalysts.
Enzymes immobilized within nanoreactors could be better
reused and show enhanced reactivity.153 A TiO2 nanotube
array was shown to be a very effective substrate to host
horseradish peroxidase for preparing biosensing devices154
and immobilize glucose oxidase for electrochemical
sensors.155 Porphyrin-sensitized TiO2 was studied as a
photoanode material for a new cell that converts light energy
into electricity.156 These examples show how bidimensional
nanostructured metal oxide films such as those schemati-
cally represented in Figure 16(b) offer various opportunities
to host homo-, hetero-, and bio-catalysts in a single nanor-
eactor to realize cascade catalytic reactions157 and to develop
a number of other interesting applications, from novel elec-
trodes for biofuel cells to photoanodes for conversion of
light energy into electricity.
Mixed-Metal Oxides 175
These examples show how the concept of bidimensional
nanostructured metal oxide film presented in Figure 16 offers
several opportunities to develop novel catalysts. For practical
applications, a major problem is related to the fact that if a thin
film on a flat surface is used, the amount of catalyst is quite
low. The challenge is thus to translate this concept on a 3D
high geometrical area material (foams, monoliths, etc.) over
which the nanostructured metal oxide film is then created to
realize a hierarchical-organized catalyst. Another solution is to
create a multistack of thin substrates over which the nanostruc-
tured metal oxide film is present, in order to create a multilayer
microstructured reactor. Integration of nanoreactor concepts
using low-dimensional metal oxides into microreactors for
process intensification is another great challenge to improve
sustainability of the chemical industrial processes.
7.07.4 Catalytic Properties of Mixed-Metal Oxides
The key features of metal oxides, which determine their surface
catalytic properties, are the following: (1) coordination envi-
ronment of the surface atoms, (2) redox properties of the
oxide, which are also related to the oxidation state of the
surface atoms, and (3) the acido-base properties. These prop-
erties are usually interrelated, although the relationship be-
tween acido-base and redox characteristics depends on the
specific metal oxides, since cations are Lewis acids while lattice
oxygen anions are basic and hydroxyl groups could be either
acidic (Brnsted site) or basic. The acidic character of the
cations depends on their positive charge and size, while
the basic character of the lattice oxygen anions depends on
the ionic character of the metaloxygen bonds. Chemisorption
further changes these properties, and the temperature of reac-
tion further induces changes in the surface properties, for
example, inducing a progressive dehydroxylation with forma-
tion of surface-exposed cations (Lewis sites). It is thus conse-
quent that only true operando spectroscopy studies, for
example, during catalytic reaction, may provide correct infor-
mation on the surface reactivity, although often it was
remarked that in these conditions the species detected, being
the most abundant, are also probably those less reactive. Some
advances have been made recently in this direction. An exam-
ple is given by the recent work of Daturi and coworkers158 that
combines kinetic and Fourier transform infrared (FTIR) oper-
ando steady-state isotopic transient kinetic analysis (SSITKA)
approaches to determine the active sites, intermediate/specta-
tor species, and reaction mechanism of methanol oxidation
over nanostructured Au/CeO2 and Au/TiO2 catalysts. These
studies show that the reactivity of a clean oxide surface
may be significantly different from that during reaction.
Often the reaction rate is determined from the presence of
strongly adsorbed species which block the surface reactivity,
and it is necessary to have a threshold reaction temperature to
desorb these species inhibiting the reactive sites. An example is
the formation of surface carboxylate and carbonate-like
species.
A second aspect to remark in discussing the catalytic prop-
erties of metal oxides is that often the same substrate may be
converted to different products depending on the oxide.
The classical example is propene which may be converted to
176 Mixed-Metal Oxides

1,5-hexadiene by oxidative dimerization on BiZnO catalysts,
to acrolein by selective oxidation on BiMoO catalysts, to
acrylic acid on VMoO catalysts, to benzene on BiSbO
catalysts, to acetic acid on VTiO catalysts, to propene oxide
on TeWO catalysts, and to CO2 on CuCrO catalysts.
There are two main general concepts6 which are used to
elucidate the reasons for the differences in the surface
chemistry between the mixed-metal oxides:
1. The combination of acido-base and redox properties of the oxide
surface determines the mechanism of transformation. An exam-
ple is given in Scheme 1. On a basic oxide, unlike an acidic
oxide where a carbocation may form by proton transfer, the
alkene (p-bonded to Lewis acid sites) is susceptible to attack
on the a-hydrogen from the neighboring lattice oxygen. The
allyl species that forms can be side-on or end-on bonded,
depending on the nature of the metal. The latter is an
intermediate to a heavier product, whereas the former
may undergo nucleophilic attack to give rise to a ketone
or aldehyde, depending on the stereochemistry of oxygen
attack, a function of the nature of the adsorbed species. The
alkene may also interact directly with surface acid Brnsted
sites, which add to the p-bond system to form an alkoxide
intermediate that, by further hydrogen abstraction, trans-
forms to a ketone, but differ from those obtained by nucle-
ophilic oxygen attack. These surface intermediate species
can be in equilibrium or not, depending on the character-
istics of the oxide. This description of the reaction mecha-
nism applies to alkene oxidation, but evidences the general
case of the reactivity on metal oxides and the possibility of
multiple pathways of transformation, which relative rates
in turn depend on the specific character of the multiple
types of sites present on metal oxide surfaces. The problem
of the selectivity in mixed oxides is how to control these
possible multiple pathways of reaction. This critical issue
applies to different types of reactions catalyzed by metal
oxides, not only selective oxidation. As an example, the
reaction network in the selective catalytic reduction of NO
with ammonia (DeNOx) over supported copper-oxide cat-
alysts is shown (Scheme 2).159 It is an interesting example,
because it shows how even for an apparently simple reac-
tion it is possible to have different pathways of transforma-
tion, which relative rates in turn depend on the specific
characteristics of the catalysts (type of support and disper-
sion of copper) and the feed composition, reaction temper-
ature, etc. Ammonia is preferentially adsorbed forming
copperamino complexes, but when gaseous oxygen
is present, heterolytic splitting occurs with formation of
CuNH2 and OH groups. The amido species further reacts
with oxygen to form N species and H2O, the former mi-
grating on the surface to recombine with another N species
to form N2. NO also coordinates on Cu2 ions forming a

H H H H
C C H C C H
R C R + C
H H
H Reaction by
H H H H carbocation mechanism
Acidic -
Me O Me O Me O Me oxide Me O Me O Me O Me

H H
C C H
Oxidation to R C
H H Basic H H aldehyde H
C C H oxide C C H
R C R C
H H Me O Me O Me O Me
H H
Side-on bond
Me O Me O Me O Me Me O Me O Me O Me R H
C C H
Oxidation to ketone H C
H
O Me O Me O Me O M
R H
C
Oxidative
C R-CH = (CH2)4 = CH-R
H coupling

H C H End-on bond
Me O Me O Me O Me

Attack by
H H Bronsted
C C H acid site H H H O
R C
H C C H Oxidation
R C R-C-CH2-CH3
H H H H to ketone
Me O Me O Me O Me Me O Me O Me O Me

Scheme 1 Schematic representation of different modalities of alkene transformation depending on the acido-base and redox properties of a metal oxide.
Adapted from Centi, G.; Cavani, F.; Trifiro F. Selective Oxidation by Heterogeneous Catalysis; Kluwer Academic/Plenum (Springer): New York, 2000.

O2 NH2
NH3 H +O
Cu O Cu
-H2O
Cu-O
NO
NO
NO
H+
Cu O
NH3
Scheme 2 Surface reaction network in the conversion of NO and ammonia in the presence of O2 over a supported copper-oxide catalyst. Adapted from
(a) Centi, G.; Perathoner, S. Catal. Rev. -Sci. Eng. 1998, 40, 175208. (b) Centi, G.; Perathoner, S. Appl. Catal. A Gen. 1995, 132, 179259.
mononitrosyl species, but chemisorption of ammonia dis-
places coordinated NO. However, NO readily reacts with
lattice or gaseous oxygen to form NO2 and NO3 adspecies,
rather more stable than the mononitrosyl species. The
nitrate-like species, for example, dissociates forming NO
and O2 only at temperatures above 400450  C. The forma-
tion of such a species thus leads to an inhibition of the
reactivity of Cu2 ions toward the heterolytic dissociative
activation of ammonia and thus to an inhibition of the rate
of the direct oxidation of ammonia to N2. NO2- and NO3-
like adspecies may react with chemisorbed ammonium
ions. NO2 reaction with ammonium ions forms N2 H2O,
whereas the nitrate species, via ammonium nitrate for-
mation, produces mainly N2O. The activity in ammonia
heterolytic dissociation of the copper ions is thus inhibited
by the formation of strongly bound NO2 and NO3, but the
oxidized nitrogen oxides may be transformed to N2 or par-
tially to N2O by reaction with surface ammonium ions.
Amide-like species (NH2) produced by the heterolytic dis-
sociation of ammonia may also react with NO to form a
nitroso-amide-like intermediate which also further trans-
forms to N2 H2O. This pathway of transformation is faster
and more efficient, but in stationary conditions the reaction
of NO with surface Cu2 ions to form NO2/NO3 species is
faster than ammonia heterolytic dissociation. In fact, when
heterolytically dissociated ammonia species are preformed
over the catalyst surface during a pretreatment in the absence
of NO, an initial overshoot in the transient reactivity is pre-
sent, but then the rate of reaction progressively decreases to
reach that observed in steady-state conditions. Ammonia can
thus be activated dissociatively either at copper sites with
a Lewis acid character or at Brnsted sites, when the
stronger adsorption of nitrogen oxides limits adsorption of
ammonia as such over Cu2 ions. In the presence of NO, the
prevailing pathway thus follows the second route, but in
nonstationary catalytic tests the first route dominates leading
to an enhanced transient surface reactivity. Notwithstanding
the very different types of reactions discussed in Scheme 2
with respect to Scheme 1, the same general aspects regarding
the surface reactivity of metal oxide are observed, evidencing
how these aspects are a common feature of these catalysts.
Mixed-Metal Oxides 177
NH
O N2 + H2O
(N)
NO NH2
H O
Cu O N2 + H2O
NO NH4+ NO2- NH4+
Cu- O O
Cu- O
O2 NO3- NH4+
N2O + H2O
Cu - O
Their understanding, however, is still far from being
satisfactory.
2. The nature of oxygen species on the catalyst determines the kind
of selectivity. This aspect applies to reactions in the presence
of oxygen, and in particular to the selective oxidation of
hydrocarbons. In the process of oxygen incorporation into
the oxide structure (see Figure 12(b)) various types of
electrophilic-activated oxygen species form before being
incorporated as structural (lattice) oxygen of the oxide.
The latter has a nucleophilic character and gives rise to a
different type of attack on an adsorbed hydrocarbon and
thus to different types of products. In the process of incor-
poration of catalyst structural oxygen into the organic mole-
cule, a point defect forms (Figure 12(c)), which can be
compensated by a change in the linkage of the coordination
polyhedra. If the latter process is not a rapid one, and the
replenishment of surface oxygen vacancies through oxygen
bulk diffusion is also slow, the population of surface electro-
philic oxygen species increases with a lowering of the selec-
tivity to partial oxidation products. The selectivity is thus a
function of the surface geometry of active sites, the redox and
oxygen transport properties of the catalyst, and the reaction
conditions which modify the rates of all these processes.
These concepts are known for about one to two decades,
but nevertheless, a more generalized approach to control the
surface selectivity in metal oxide-catalyzed reactions is still
missing. Even on a more restricted but still general question,
such as the relation between characteristics of low-dimensional
metal oxides and selectivity, data are mainly phenomenologi-
cal. It is thus necessary to address on more fundamental bases,
the theory of catalysis by mixed-metal oxide catalysts. A num-
ber of recent reviews have addressed the relation between
selectivity and metal oxide catalyst characteristics. Mann
et al.160 have analyzed the properties of metal oxide clusters
in nontraditional oxidation states, in particular reviewing how
oxygen vacancies on metal oxide surfaces create active sites for
catalytic processes, ranging from selective oxidation of alkenes
to photocatalytic decomposition of water. These authors try to
associate the structures and electronic properties of small metal
oxide clusters to chemical reactivity, in group 6 metal oxides
178 Mixed-Metal Oxides
(Mo and W). Hutchings and coworkers15 have reviewed the
use of VPO materials as selective oxidation catalysts, and in
particular the strategies to achieve improved performances by
addition of secondary metal ions to the catalyst, and optimi-
zation of the catalyst precursor formation. Marks, Star, and
Wegener161 reviewed the design strategies for the molecular-
level synthesis of supported catalysts (metal or oxide catalytic
centers constructed on an underlying solid phase), in particu-
lar, in terms of advanced characterization methodologies for
these materials and new synthetic approaches to tailor sup-
ported catalyst structures. Although this contribution, as the
other one mentioned above, shows that it is possible today to
synthetize defined surface species using a molecularly oriented
approach, the catalyst design is still not based on robust
modeling of the bulk and surface properties required in metal
oxides to maximize the selectivity. Other recent reviews have
discussed the characteristics of selective oxidation reactions,50
the selective oxidation of n-butane over vanadium phospho-
rous oxide,15 the ammoxidation of propylene and propane to
acrylonitrile,57 the propane selective oxidation to acrylic
acid,162,163 and CdH activation of alkanes in selective oxida-
tion reactions on solid oxide catalysts.164 Haber165 has com-
prehensively discussed 5 years of experimental and theoretical
advances on vanadium-oxide catalysts and the contribution to
the understanding of the mechanism of chemical reactions at
metal oxide surfaces. The aspects related to reconstruction of
the vanadium oxides during the catalytic reaction, for example,
when exposed to gas phase of sufficient redox potential reduc-
tion, are also discussed. Hess166 also reviewed the use of nano-
structured vanadium oxide in selective oxidation reactions, and
in particular vanadium-oxide catalysts supported on mesopor-
ous silica SBA-15 as model systems. Guliant and coworkers167
also reviewed the effects of size, shape, structure, and compo-
sition of catalyst nanoparticles (metal and metal oxides) on the
mechanisms of conversion of hydrocarbons in a number of
reactions (alkylation, dehydrogenation, hydrogenation, and
selective oxidation), concluding that fundamental under-
standing of surface molecular structurereactivity relationships
of these systems remains highly limited. Thus, there is still a
large interest on research of metal oxide catalysts, and also a
common agreement on the need to make a step forward in
their understanding.
7.07.4.1 Surface Acidity and Basicity of Metal Oxides
The surface acidity and basicity of metal oxides is another
relevant property for the catalytic reactivity of these
materials.168 Acidbase properties of solids are usually dis-
cussed in terms of Lewis and Brnsted definitions of acidity
and basicity, but in some cases also in terms of hard and soft
acid and bases (HSAB) theory to solidgas interfaces.
When the exposed planes in a metal oxide crystal are
ideally cut, metaloxygen bonds (either covalent or ionic)
have been broken, leaving exposed on the surface metal or
nonmetal centers coordinatively unsaturated. To stabilize the
surface, reconstruction phenomena as well as reaction with
molecules from the environment (e.g., water and CO2) occur.
This limits the number of coordinatively unsaturated centers
and causes the formation of new surface species such as
hydroxy-groups and surface carbonates. Some unsaturated
centers at the surface can remain or be generated by desorp-
tion of adsorbed water (and CO2) during heating. As a con-
sequence of these phenomena, the surface of solid oxides can
be constituted by
exposed coordinatively unsaturated cationic centers, poten-
tially acting as Lewis acid sites;
exposed oxide species, potentially acting as basic sites;
exposed hydroxy-groups, arising from water dissociative
adsorption, potentially acting as Brnsted acid sites, or,
alternatively, as basic sites;
other surface species arising from the reactivity of the surface
with the environment, such as carbonate species, or other
adsorbed species (including those which may form during
the catalytic reaction).
When chains (e.g., CrO3) or layers (e.g., V2O5 and MoO3)
are present, the cutting of the crystal along these planes causes
the breaking of van der Waals interactions (e.g., for the 100 and
010 faces of a-MoO3169). However, also in these cases coordi-
natively unsaturated atoms can be present, although limited to
some of the exposed planes (e.g., the nonbasal planes for
layered structures such as the V2O5170 and the 001 face of
a-MoO3169) and to defects.
The density (amount of sites per gram of the solid or per
unit surface area) and strength of these acido-base sites are
both relevant parameters determining the catalytic behavior.
A number of techniques for the characterization of these as-
pects have been discussed by Busca,168 while Glazneva et al.171
discussed some complementary aspects, with focus on in situ
measurements such as SSITKA for H/D exchange. The prepara-
tion of superacidic metal oxides and their catalytic behavior
were reviewed by Arata172 and the use of nanosheets as highly
active solid acid catalysts by Domen and coworkers.173 Macht
and Iglesia174 instead discussed the catalytic consequences of
size and composition in oxide nanostructures, showing how
both the redox and acidbase properties of dispersed oxide
nanostructures change markedly as their local structure and
electronic properties vary with domain size. These changes
give rise to catalytic behavior, site structures, and reaction
chemistries often unavailable on bulk crystalline oxides. Turn-
over rates for redox and acid catalysis vary as oxide domains
evolve from isolated monomers to 2D oligomers, and ulti-
mately into clusters with bulk-like properties. These reactivity
changes reflect the ability of oxide domains to accept or
redistribute electron density in kinetically relevant reduction
steps, in the formation of temporary acid sites via reductive
processes, and in the stabilization of cationic transition states.
Reduction steps are favored by low-lying empty orbitals
prevalent in larger clusters, which also favor electron delocal-
ization, stable anions, and strong Brnsted acidity. Isomeriza-
tion of xylenes and alkanes, elimination reactions of alkanols,
and oxidation of alkanes to alkenes on V, Mo, Nb, and W
oxide domains were used by Macht and Iglesia174 to demon-
strate the remarkable catalytic diversity made available by
changes in domain size. The reactive and disordered nature
of small catalytic domains introduces significant challenges in
their synthesis and their structural and mechanistic character-
ization, which require in situ probes and detailed kinetic
analyses. The local structure and electronic properties of
these materials must be probed during catalysis and their

catalytic function must be related to specific kinetically rele-
vant steps.
7.07.4.2 Modification of the Surface Reactivity by
Chemisorbed Species
In the previous sections it was already remarked that the concept
of clean catalyst surface, for example, of reaction of a molecule
at single specific active sites without considering the modifica-
tion of the surface reactivity induced by the presence of other
coadsorbates (reactants, intermediates, and spectator species),
may lead to a not correct description of the real working catalyst
surface and reaction mechanism, especially when complex, mul-
tistep reactions (typical of metal oxide catalysts) are considered.
It is thus necessary to consider the role of chemisorbed species in
the modification of the surface reactivity.159,175
An example of the application of this concept is presented
in analyzing the catalytic behavior of VSbO4 Sb2O4 catalyst
(Sb:V 2) in the ammoxidation of propane and propene to
acrylonitrile, a relevant industrial chemical (Table 3). These
catalysts are selective in the formation of acrylonitrile from
both the C3 alkane and the alkene. The structural model of
VSbO4 is shown in Figure 17.176 Propane and propene, due to
their different chemical properties (the alkene has allylic hy-
drogens and a p-bond system different from the alkane), inter-
act differently with the catalyst, and in particular the alkene
may more easily and strongly chemisorb on the surface. The
comparison of their conversion on VSb oxides is thus a useful
approach to understand how the catalytic properties of the
catalyst are influenced by the chemisorption of the reactants
on the catalyst surface, because this catalytic system is one of
the few being active and selective in both reactions under
comparable reaction conditions. Propane and propene
ammoxidation using the same catalyst and reaction tempera-
ture are compared in Figure 9, which reports the dependence
of the selectivity to acrylonitrile on the conversion of the
hydrocarbon at 480  C. Acrylonitrile can be further oxidized
70
60
ACN from propene
50
Selectivity (%)
40
ACN from
propane
30
20
C3H6 from
10 propane
0
0 20 40 60 80
C3 conversion (%)
Figure 17 Comparison of the selectivity dependence on acrylonitrile in propane and propene ammoxidation at 480  C on a VSbO4 Sb2O4 catalysts
(Sb:V 2). Adapted from Centi, G.; Perathoner, S. Int. J. Mol. Sci. 2001, 2, 183196. Model of VSbO4. Adapted from Seitza, H.; Germana, E.; Juana, A.;
Irigoyenb, B. Appl. Surf. Sci. 2012, 258, 36173623.
Mixed-Metal Oxides 179
to give carbon oxides and thus it is reasonable that at high
conversion of the hydrocarbon its selectivity decreases, while
that of carbon oxides increases. However, the data in Figure 17
point out the contradiction that on the same catalyst, and at
the same reaction temperature, acrylonitrile converts faster to
carbon oxides when formed from propane than when formed
from propene, even though propene is the reaction intermedi-
ate from propane.70 In fact, the maximum in acrylonitrile
formation from propane is observed at much lower conver-
sions than in the case of propene and as a consequence the
maximum yield is significantly lower. The apparent kinetic
paradox can be explained only considering the different mod-
ifications of the surface reactivity when feeding alkane or al-
kene. This is a more general phenomenon not only restricted to
the case of VSb oxides.
Another example of how the reactant induces a change in
the surface properties was observed studying the effect of the
change in the oxygen to alkane ratio in the feed on the selec-
tivity to propene from propane on the (VO)2P2O7 catalyst.70a
This is the industrial catalyst for n-butane conversion to maleic
anhydride (see Figure 17), but under the same conditions gives
only carbon oxides in the oxidation of propane, for example,
by decreasing the alkane chain of only one carbon atom. By
decreasing the O2/propane inlet ratio the selectivity to the
reaction intermediate propene increases linearly from about
zero to 20% when propane concentration is 2% and from zero
to about 70% when the propane initial concentration is 32%.
The marked difference in the trend between the two cases de-
rives from the different formation of alkene product, which is
about ten times larger using the higher inlet concentration of
propane in the feed. This indicates that when the formation of
propene is higher, the rate of its consecutive oxidation to
carbon oxides also decreases and thus the selectivity increases.
The formation of propene (reaction intermediate) thus induces
a self-modification of the surface reactivity. This in confirmed
from the analysis of the kinetics of reaction and the analysis of
the transient reactivity as well as spectroscopic studies.
Propane Propene Acrylonitrile (ACN)
O2, NH3
O2
C3H8 C3H6 CH2 = CH-CN
-H2O -H2O
CO, CO2, H2O
[-110]
[110]
[001]
V Sb O
100
180 Mixed-Metal Oxides
Both (VO)2P2O7 and VSbO4 catalysts are characterized
from the presence of coordinative unsaturated V4 O surface
sites which act as strong Lewis acid sites and which play a
relevant role in the mechanism of alkane activation. Alkenes
through their p-bond system can chemisorb on these sites
forming relatively stable chemisorbed species, although they
may be considered spectator species, because they are not
directly involved in the mechanism of further selective oxida-
tion of these alkene intermediates. Oxygen also strongly chem-
isorbs on the surface Lewis acid sites forming relatively thermal
stable species. When the surface concentration of the interme-
diate alkenes in alkane oxidation is as high as to limit the
amount of chemisorbed oxygen, due to this competitive chem-
isorption, it is thus possible to control the population of oxy-
gen adspecies by this mechanism. This not only explains the
considerable promotion of selectivity to partial oxidation
products by increasing alkane inlet concentration but also
explains the apparent contradiction of the different rates of
acrylonitrile consecutive oxidation when formed from propene
rather than from propane (Figure 17).
The chemisorption of reactants or intermediates may influ-
ence not only the rate of the selective versus unselective oxida-
tion, but also the rate of competitive pathways of transformation,
when multiple pathways of surface transformation are possible.
The chemisorption of reactants or intermediates may influence
the rates of these multiple pathways of transformation which are
characterized by a different intrinsic selectivity. This concept is
exemplified by studying the behavior of (VO)2P2O7 and VSbO4
catalysts in propane ammoxidation to acrylonitrile. On the latter
catalyst, using low concentration of ammonia in the feed (am-
monia is a reactant necessary to synthetize acrylonitrile), there is
a strong decrease of carbon oxide formation. In parallel there is
an increase in the selectivity to the intermediate propene and not
to acrylonitrile, which forms only above some threshold in
ammonia concentration. At low levels, ammonia thus does not
promote the formation of the N-containing product, but only
strongly depresses the unselective conversion to carbon oxides of
the intermediate propene.
FTIR data177 show that ammonia interacts with the Brnsted
acid sites of the catalyst forming ammonium ions, the stability of
O
OH, H2O
O O O
NH
-H
C OH, H2O
O O
N -NH3
CN Acrylonitrile
(gas phase)
Scheme 3 Reaction schemes of acrylate (a) and acrylonitrile (b) oxidation to carbon oxides.
which, however, is weak and evacuation at temperatures of about
200  C is enough for their nearly complete removal. This sug-
gests that at the typical reaction temperatures for the catalytic
tests of propane ammoxidation (400500  C) the interaction of
ammonia with these surface Brnsted acid sites may be consid-
ered negligible. Ammonia remains instead strongly chemisorbed
as such on surface Lewis acid sites, up to evacuation temperatures
of about 400  C. However, in the presence of water vapor the
situation is different and it was observed that ammonium ions
form at high temperature from ammonia coordinated on Lewis
acid sites. Water is one of the main products of reaction in
hydrocarbon oxidation and thus all considerations on the stabil-
ity of surface species must take into account its presence. Fur-
thermore, desorption of ammonia adspecies is an equilibrium
reaction with ammonia in the gas phase. When NH3 is present in
the gas phase, Brnsted acid sites are nearly completely converted
to ammonium ions even at high temperature. Recent density
functional theory (DFT) calculations176 are in good agreement
with these conclusions. It is thus reasonable that during the
catalytic reaction of propane conversion on VSb oxide in the
presence of ammonia, the larger part of the surface Brnsted acid
sites is transformed to ammonium ions and consequently the
reactivity of these Brnsted acid sites is inhibited.
Brnsted acid sites catalyze various side reactions. In particu-
lar, it has been observed177 that acrylic acid forms stable and
relatively inert acrylate species with VSb oxide catalyst up to
high temperatures of evacuation. In the presence of Brnsted
acid sites and water, however, this species transforms to the
corresponding free acid form weakly bound to the surface and
easily susceptible to consecutive decarboxylation and further
total oxidation (Scheme 3(a)). The strong interaction of acrylate
with the surface explains why even when formed from propane
and propene (as detected by FTIR spectroscopy) acrylic acid is
detected only in traces as a product of C3 hydrocarbon oxidation
on VSb oxide. The interaction with the surface of the corre-
sponding species formed in the presence of ammonia (acryloni-
trile) is weaker and thus the desorption is easier (Scheme 3(b)).
Acrylonitrile may also be transformed to the corresponding amide
by the action of Brnsted acid sites and water (Scheme 3(b)),
which more easily further transforms to the free acid and then
H
COx
O
H
OH, H2O
COx

finally to carbon oxides. The self-inhibition of ammonia on the
activity of Brnsted acid sites converting them to ammonium
ions, however, inhibits this pathway.
This example well illustrates the complexity of the surface
mechanisms in reactions of industrial relevance over mixed-
metal oxides, and the need to consider this complexity for the
design of more selective catalysts. It is necessary to develop new
models of the surface reactivity which include aspects such as
(1) the role of chemisorbed species on the surface reactivity,
(2) the presence of multiple pathways of reaction, (3) the
dynamics of catalyst reconstruction, and (4) the mobility of
surface adspecies, etc. This is the challenge to understand the
surface reactivity at oxide surfaces.
7.07.5 Conclusion
Mixed-metal oxides are an important class of catalytic mate-
rials, with applications in a wide range of sectors, but particu-
larly in the selective oxidation industrial processes to produce a
large variety of relevant chemicals. Their key reactivity feature is
related to the multifunctional behavior, for example, the pres-
ence of specific surface ensemble sites having all the acido-
base, redox, and electron transfer sites necessary to perform
selectively multi-electron, -proton, and -oxygen transfer reac-
tions. Selectivity is the critical characteristic, because the reac-
tion products have typically an intrinsic reactivity higher than
the reactant and multifunctionality means also the possibility
of multiple pathways of transformations. Therefore, the mixed
oxide catalyst must not only possess the requested active sites,
which should operate in a concerted/synchronized way to
transform the reactant to the product, but also avoid other
(unselective) transformations that the reactant/product inter-
mediates/final product may undergo. There are thus very de-
manding characteristics required in a mixed oxide to operate
selectively in these catalytic reactions, in addition to other
characteristics (pore structure, mechanical properties, stability,
shape, etc.) necessary for optimal and long-term operations in
industrial reactors.
In selective oxidation reactions, mixed oxides operate typi-
cally through a Marsvan-Krevelen mechanism, where the ox-
ygen involved in selective oxygen insertion or in hydrogen
abstraction from the hydrocarbon substrate is lattice oxygen.
Gaseous dioxygen is instead reduced through a different mech-
anism involving different active sites. Inequivalent activation
of oxygen into nucleophilic and electrophilic oxygen species is
an important characteristic of mixed-metal oxides for selective
oxidation reactions, but the effectiveness of this process re-
quires solid-state chemistry with controlled kinetics at low
temperatures. To achieve this aspect and to avoid the forma-
tion of extensive defects which may lead to reconstruction of
the active sites, the catalyst nanostructure and the interface
between different (nano)phases (often operating in synergy/
symbiosis) are critical to optimize the performances. A clear
consequence is the presence of a structural dynamic which is
the key feature necessary to understand and to control, but
which depends on the reaction environment. An active mixed
oxide catalyst contains adaptive sites, which adjust their struc-
ture according to local chemical potential and guarantee
Mixed-Metal Oxides 181
selective operation. Understanding this structural dynamics
during the catalytic reaction is the challenge which requires
not only the combination of various in situ characterization
techniques, but also the ability of a precise control of the
nanostructure, morphology (preferential exposition of specific
crystal planes), and phase purity during the synthesis proce-
dure. Often the fresh material after synthesis is a precatalyst, for
example, a precursor leading to the effective active phase in the
reaction environment. Nanostructuring is thus a prerequisite
to reach a time domain compatible with chemical transforma-
tion. It is also necessary to understand how to maintain for a
long time, often in a dynamic process, this surface active struc-
ture formed during reaction, avoiding phase segregation, trans-
formation, or other effects leading to a progressive catalyst
deactivation.
Surface termination is often different from the ideal crystal
structure obtained by cutting the crystal, and often spectator
species or simply physically blocking unselective sites are also
present. Strongly chemisorbed adspecies (e.g., with turnover of
transformation significantly slower than that of the reaction
intermediates), formed during the catalytic reaction, also play
a role in determining the surface living active state of the
catalyst.
A rational design of mixed oxides requires considering all
these aspects, from the role of nanostructure in facilitating
redox processes, the formation of specific surface structures,
and the defect chemistry and kinetics, to the material dynamics
during the catalytic reaction, passing through the detailed un-
derstanding of the complex chemistry during the synthesis
procedure, and the activation processes. With the spectacular
development in the last decade of new characterization (in situ)
tools, including theoretical modeling, a new era in catalysis,
based on a knowledge-driven design, has started. New possi-
bilities are offered for the understanding of the catalytic chem-
istry of mixed-metal oxides, which is highly challenging due to
their complex characteristics outlined here, but which is also
the key for the design of new sustainable chemical processes.
Design of a controlled 3D (nano)architecture, ad-hoc strategy
for an optimal microkinetic sequence of active sites, and how
to force this sequence in a substrate transformation, bridging
the complexity gap, atomic-level monitoring of active site dy-
namics during the catalytic reaction, coupling the catalytic
chemistry of substrate (hydrocarbon) transformation to mate-
rial dynamic, and synthesis of ad-hoc mixed oxide surface
nanostructures by controlled assembling of nano-units are
some of the fascinating challenges discussed here, which
make the area of mixed-metal oxide catalysis a driver for sci-
entific developments with key relevance for a sustainable fu-
ture and industrial progress.
References
1. Cavani, F.; Centi, G.; Perathoner, S.; Trifiro, F. Sustainable Industrial Chemistry.
Wiley-VCH: Weinheim, 2009.
2. Zecchina, A.; Bordiga, S.; Groppo, E. Selective Nanocatalysts and Nanoscience:
Concepts for Heterogeneous and Homogeneous Catalysis. Wiley-VCH:
Weinheim, 2009.
3. Bartholomew, C. H.; Farrauto, R. J. Fundamentals of Industrial Catalytic
Processes, 2nd ed.; John Wiley and Sons: New York, 2006.
182 Mixed-Metal Oxides
4. Fierro, J. L. G. Metal Oxides. Chemistry and Applications. CRC Press: Boca Raton,
FL, 2006.
5. Jackson, S. D.; Hargreaves, J. S. J. Metal Oxide Catalysis (Two Volumes). Wiley-
VCH: Weinheim, 2009.
6. Centi, G.; Cavani, F.; Trifiro, F. Selective Oxidation by Heterogeneous Catalysis.
Kluwer Academic/Plenum (Springer): New York, 2000.
7. Hodnett, B. K. Heterogeneous Catalytic Oxidation. John Wiley and Sons:
New York, 2000.
8. (a) Cavani, F.; Teles, J. H. ChemSusChem 2009, 2, 508534; (b) Cavani, F.
Catal. Today 2010, 157, 815.
9. Cavani, F.; Centi, G. Kirk-Othmer Encyclopedia of Chemical Technology. Wiley
online Library, 2011; pp 161.
10. (a) Kongmark, C.; Martis, V.; Pirovano, C.; Lofberg, A.; van Beek, W.; Sankar, G.;
Rubbens, A.; Cristola, S.; Vannier, R.-N; Bordes-Richard, E. Catal. Today 2010,
157, 257262; (b) Kuhlmann, U. Phonon and Spin Dynamics of (VO)2P2O7. Ph.D.
Thesis, Technischen Universitat, Berlin, 2005; (c) Grasselli, R. K.; Lugmair, C. G.;
Volpe Jr., A. F. Top. Catal. 2011, 54, 595604.
11. Wang, D. Characterization of Catalysts by TEM, Presented at 36th Erice
Crystallographic Course on Electron Crystallography, Erice, Italy, June 920,
2004, 2004.
12. Cavani, F.; Centi, G.; Marion, P. In Metal Oxide Catalysis; Jackson, S. D.,
Hargreaves, J. S., Eds.; Wiley-VCH: Weinheim, 2009; vol. 2, pp 771818.
13. de Jong, K. P. Synthesis of Solid Catalysts. Wiley-VCH: Weinheim, 2009.
14. Schlogl, R. In Modern Heterogeneous Oxidation Catalysis; Mizuno, N., Ed.;
Chapter 1, Wiley-VCH: Weinheim, 2009, pp 142.
15. Dummer, N. F.; Bartley, J. K.; Hutchings, G. J. Adv. Catal. 2011, 54, 189247.
16. Centi, G.; Trifiro, F.; Ebner, J. R.; Franchetti, V. M. Chem. Rev. 1988, 88, 5580.
17. Conte, M.; Budroni, G.; Bartley, J. K.; Taylor, S. H.; Carley, A. F.; Schmidt, A.;
Murphy, D. M.; Girgsdies, F.; Ressler, T.; Schlogl, R.; Hutchings, G. J. Science
2006, 313, 12701273.
18. Salmeron, M.; Schlogl, R. Surf. Sci. Rep. 2008, 63, 169199.
19. DeSanto, P.; Buttrey, D. J.; Grasselli, R. K.; Lugmair, C. G.; Volpe, A. F.;
Toby, B. H.; Vogt, T. Z. Kristallogr. 2004, 219, 152165.
20. Murayama, H.; Vitry, D.; Ueda, W.; Fuchs, G.; Anne, M.; Dubois, J. L. Appl. Catal.
A: Gen. 2007, 318, 137142.
21. Grasselli, R. K.; Buttrey, D. J.; DeSanto, P., Jr.; Burrington, J. D.; Lugmair, C. G.;
Volpe, A. F., Jr.; Weingand, T. Catal. Today 2004, 9192, 251258.
22. Celaya Sanfiz, A.; Hansen, T. W.; Girgsdies, F.; Timpe, O.; Rodel, E.; Ressler, T.;
Trunschke, A.; Schlogl, R. Top. Catal. 2008, 50, 1932.
23. Sanchez Sanchez, M.; Girgsdies, F.; Jastak, M.; Kube, P.; Schlogl, R.;
Trunschke, A. Angew. Chem. Int. Ed. 2012, 51, 71947197.
24. Kolenko, Y. V.; Zhang, W.; Zhang, W.; dAlnoncourt, R. N.; Girgsdies, F.;
Hansen, T. W.; Wolfram, T.; Schlogl, R.; Trunschke, A. ChemCatChem 2011, 3,
15971606.
25. Zhang, W.; Trunschke, A.; Schlogl, R.; Su, D. Angew. Chem. Int. Ed. 2010, 49,
60846089.
26. Havecker, M.; Wrabetz, S.; Krohnert, J.; Csepei, L.-I.; Naumann dAlnoncourt, R.;
Kolenko, Y. V.; Girgsdies, F.; Schlogl, R.; Trunschke, A. J. Catal. 2012, 285,
4860.
27. Sanfiz, A. C.; Hansen, T. W.; Teschner, D.; Schnorch, P.; Girgsdies, F.;
Trunschke, A.; Schlogl, R.; Looi, M. H.; Abd Hamid, S. B. J. Phys. Chem. C 2010,
114, 19121921.
28. Centi, G.; Perathoner, S. Catal. Today 2003, 77, 287297.
29. Centi, G.; Perathoner, S. Int. J. Mol. Sci. 2001, 2, 183196.
30. Centi, G.; Fornasari, G.; Trifiro, F. J. Catal. 1984, 89, 4451.
31. van Santen, R. A. In Novel Concepts in Catalysis and Chemical Reactors:
Improving the Efficiency for the Future; Cybulski, A., Moulijn, J. A.,
Stankiewicz, A., Eds.; Wiley-VCH: Weinheim, 2010.
32. Norskov, J. K.; Bligaard, T.; Hvolbaek, B.; Abild-Pedersen, F.; Chorkendorff, I.;
Christensen, C. H. Chem. Soc. Rev. 2008, 37, 21632171.
33. (a) Norskov, J. K.; Bligaard, T.; Rossmeisl, J.; Christensen, C. H. Nat. Chem.
2009, 1, 3746; (b) Norskov, J. K.; Abild-Pedersen, F. Nature 2009, 461,
12231225.
34. Centi, G.; Trifiro, F. Chem. Eng. Sci. 1990, 45, 25892596.
35. (a) Grasselli, R. K. Top. Catal. 2001, 15, 93101; (b) Grasselli, R. K. Top. Catal.
2002, 21, 7988.
36. Freund, H.-J. Chem. Eur. J. 2010, 16, 93849397.
37. Nilius, N.; Risse, T.; Schauermann, S.; Shaikhutdinov, S.; Sterrer, M.;
Freund, H.-J. Top. Catal. 2011, 54, 412.
38. Schlogl, R.; Knop-Gericke, A.; Havecker, M.; Wild, U.; Frickel, D.; Ressler, T.;
Jentoft, R. E.; Wienold, J.; Mestl, G.; Blume, A. Top. Catal. 2001, 15, 219228.
39. Havecker, M.; Knop-Gericke, A.; Bluhm, H.; Kleimenov, E.; Mayer, R. W.; Fait, M.;
Schlogl, R. Appl. Surf. Sci. 2004, 230, 272282.
40. Ozkan, U. S.; Watson, R. B. Catal. Today 2005, 100, 101114.
41. Centi, G.; Perathoner, S. Coord. Chem. Rev. 2011, 255, 14801498.
42. Freund, H.-J.; Ernst, N.; Risse, T.; Hamann, H.; Rupprechter, G. Phys. Status
Solidi A Appl. Res. 2001, 187, 257274.
43. Ressler, T.; Kniep, B. L.; Kasatkin, I.; Schlogl, R. Angew. Chem. Int. Ed. 2005, 44,
47044707.
44. Goodman, D. W. Chem. Rev. 1995, 95, 523536.
45. Centi, G.; Perathoner, S. Eur. J. Inorg. Chem. 2009, 26, 38513878.
46. Giordano, L.; Pacchioni, G. Acc. Chem. Res. 2011, 44, 12441252.
47. Sierka, M. Prog. Surf. Sci. 2010, 85, 398434.
48. Centi, G.; Passalacqua, R.; Perathoner, S.; Su, D. S.; Weinberg, G.; Schlogl, R.
Phys. Chem. Chem. Phys. 2007, 9, 49304938.
49. Mizuno, N.; Kamata, K.; Yamaguchi, K. Top. Organomet. Chem. 2011, 37,
127160.
50. Schlogl, R.; Hess, C. RSC Nanosci. Nanotech. 2011, 19, 355397.
51. Cavani, F. J. Chem. Tech. Biotech. 2010, 85, 11751183.
52. Mars, P.; van Krevelen, D. W. Chem. Eng. Sci. 1954, 3, 4159.
53. Diebold, U.; Li, S.-C.; Schmid, M. Annu. Rev. Phys. Chem. 2010, 61,
129148.
54. Lewandowski, M.; Groot, I. M. N.; Shaikhutdinov, S.; Freund, H.-J. Catal. Today
2012, 181, 5255.
55. Recnik, A.; Bernik, S.; Daneu, N. J. Mater. Sci. 2012, 47, 16551668.
56. Cavani, F.; Luciani, S.; Esposti, E. D.; Cortelli, C.; Leanza, R. Chem. Eur. J. 2010,
16, 16461655.
57. Grasselli, R. K. RSC Nanosci. Nanotech. 2011, 19, 96140.
58. Centi, G. Diffus. Defect Data-Solid State Data A: Defect Diffus. Forum 2001, 191,
Soft Chemistry Leading to Novel Materials, 1734.
59. Courtine, P.; Bordes, E. Stud. Surf. Sci. Catal. 1997, 110, 177184.
60. Chuang, C.-C.; Hsiang, H.-I.; Yen, F.-S. Ceram. Trans. 2010, 210, 199205.
61. Wang, S.; Zhang, Y.; Liu, H. Chem. Asian J. 2010, 5, 11001111.
62. Kasatkin, I.; Kurr, P.; Kniep, B.; Trunschke, A.; Schlogl, R. Angew. Chem. Int. Ed.
2007, 46, 73247327.
63. Kniep, B. L.; Ressler, T.; Rabis, A.; Girgsdies, F.; Baenitz, M.; Steglich, F.;
Schlogl, R. Angew. Chem. Int. Ed. 2004, 43, 112115.
64. Schoiswohl, J.; Sock, M.; Chen, Q.; Thornton, G.; Kresse, G.; Ramsey, M. G.;
Surnev, S.; Netzer, F. P. Top. Catal. 2007, 46, 137149.
65. Chen, Z.; Cvelbar, U.; Mozetic, M.; He, J.; Sunkara, M. K. Chem. Mater. 2008, 20,
32243228.
66. Wang, N.; Cai, Y.; Zhang, R. Q. Mater. Sci. Eng. R Rep. 2008, R60, 151.
67. Zhou, K.; Li, Y. Angew. Chem. Int. Ed. 2012, 51, 602613.
68. Rameshan, C.; Stadlmayr, W.; Penner, S.; Lorenz, H.; Memmel, N.; Haevecker, M.;
Blume, R.; Teschner, D.; Rocha, T.; Zemlyanov, D.; Knop-Gericke, A.; Schlogl, R.;
Klotzer, B. Angew. Chem. Int. Ed. 2012, 51, 30023006.
69. Charlton, G.; Hoowes, P. B.; Nicklin, C. L.; Steadman, P.; Taylor, J. S. G.;
Muryn, C. A.; Harte, S. P.; Mercer, J.; McGrath, R.; Norman, D. Phys. Rev. Lett.
1997, 78, 495498.
70. (a) Centi, G.; Perathoner, S.; Trifiro, F. Appl. Catal. A: Gen. 1997, 157, 143172;
(b) Centi, G.; Trifiro, F. Catal. Rev. Sci. Eng. 1986, 28, 165184.
71. Behrens, M.; Furche, A.; Kasatkin, I.; Trunschke, A.; Busser, W.; Muhler, M.;
Kniep, B.; Fischer, R.; Schlogl, R. ChemCatChem 2010, 2, 816818.
72. Goniakowski, J.; Finocchi, F.; Noguera, C. Rep. Prog. Phys. 2008, 71,
016501016555.
73. Kresse, G.; Dulub, O.; Diebold, U. Phys. Rev. B 2003, 68, 245409-1/15.
74. Staemmler, V.; Fink, K.; Meyer, B.; Marx, D.; Kunat, M. Phys. Rev. Lett. 2003, 90,
106102106104.
75. Woll, C. Prog. Surf. Sci. 2007, 82, 55120.
76. Wischert, R.; Coperet, C.; Delbecq, F.; Sautet, P. Angew. Chem. Int. Ed. 2011, 50,
32023205.
77. Kuhrs, C.; Swoboda, M.; Weiss, W. Top. Catal. 2001, 15, 1318.
78. Zecchina, A.; Spoto, G.; Scarano, D.; Bordiga, S. In Handbook of Heterogeneous
Catalysis, 2nd ed.; Ertl, G., Ed.; Wiley-VCH: Weinheim, 2008; pp 13521362.
79. Frank, B.; Blume, R.; Rinaldi, A.; Trunschke, A.; Schlogl, R. Angew. Chem. Int. Ed.
2011, 50, 1022610230.
80. Basset, J.-M.; Baudouin, A.; Bayard, F.; Candy, J.-P.; Coperet, C.; De
Mallmann, A.; Godard, G.; Kuntz, E.; Lefebvre, F.; Lucas, C. In Modern Surface
Organometallic Chemistry; Basset, J.-M., Ed.; Wiley-VIC: Weinheim, 2009;
pp 75135.
81. Lee, H.; Kim, C.; Yang, S.; Han, J. W.; Kim, J. Catal. Surv. Asia 2012, 16, 1427.
82. Stair, P. C. Top. Catal. 2012, 55, 9398.
83. Li, G. In Metal Oxide Nanostructures and Their Applications; Umar, A., Hahn, Y.-B.,
Eds.; American Scientific Publishers: CA, 2010; pp 161210.
84. Sattler, K. D. Handbook of Nanophysics: Nanotubes and Nanowires. CRC Press
(Taylor and Francis Group): Boca Raton, FL, 2011.

85. Wu, J.; Cao, J.; Han, W.-Q.; Janotti, A.; Kim, H.-C. Functional Metal Oxide
Nanostructures. Springer: Heidelberg, 2012.
86. Abbet, S.; Heiz, U. In Encyclopedia of Nanoscience and Nanotechnoloy;
Nalwa, H. S., Ed.; American Scientific Publishers: Valencia, CA, 2004; vol. 6,
pp 161177.
87. Corma, A. Chem. Rev. 1997, 97, 23732419.
88. Schuth, F.; Schmidt, W. Adv. Mater. 2002, 14, 629638.
89. Zones, S. I. Micropor. Mesopor. Mater. 2011, 144, 18.
90. Schuth, F.; Wingen, A.; Sauer, J. Micropor. Mesopor. Mater. 2001, 4445,
465476.
91. Centi, G.; Perathoner, S.; Vazzana, F. In Catalysis by Unique Ion Structures in
Solid Matrices NATO Science Series II: Mathematics, Physics and Chemistry;
Bell, A., Centi, G., Wichterlowa, B., Eds.; Chapter 11, Kluwer/Academic Press:
New York, 2001; pp 165186.
92. Jayadevan, K. P.; Tseng, T. Y. In Encyclopedia of Nanoscience and
Nanotechnoloy; Nalwa, H. S., Ed.; American Scientific Publishers: Valencia, CA,
2004; vol. 8, pp 333376.
93. (a) Centi, G.; Perathoner, S.; Trifiro, F.; Aboukais, A.; Aissi, C. F.; Guelton, M.
J. Phys. Chem. 1992, 96, 26172629; (b) Bellussi, G.; Centi, G.; Perathoner, S.;
Trifiro, F. In Selective Oxidation Catalysts ACS Symposium Series; Oyama, T.,
Hightower, J., Eds.; American Chemical Society: Washington, 1992;.
94. (a) Centi, G.; Perathoner, S. Top. Catal. 2001, 15, 145152; (b) Centi, G.;
Perathoner, S.; Tonini, S. Top. Catal. 2000, 11/12, 195204; (c) Centi, G.;
Perathoner, S.; Tonini, S. Catal. Today 2000, 61, 211221.
95. Chen, Ji. In Modern Inorganic Synthetic Chemistry; Xu, R., Pang, W., Huo, Q.,
Eds.; Elsevier: Amsterdam, 2011; pp 405428.
96. Corma, A.; Garcia, H. Chem. Commun. 2004, 13, 14431459.
97. He, J.; Zhang, J.; Li, J.; Zhu, Z. Adv. Mater. Res. 2012, 396398, 739744.
98. Mitra, B.; Kunzru, D. Catal. Lett. 2011, 141, 15691579.
99. Oveisi, H.; Anand, C.; Mano, A.; Al-Deyab, S. S.; Kalita, P.; Beitollahi, A.; Vinu, A.
J. Mater. Chem. 2010, 20, 1012010129.
100. Asedegbega-Nieto, E.; Diaz, E.; Vega, A.; Ordonez, S. Catal. Today 2010, 157,
425431.
101. Su, K.; Li, Z.; Cheng, B.; Ren, Y.; Yu, L.; Wang, F. Kinet. Catal. 2010, 51,
359363.
102. Patzke, G. R.; Krumeich, F.; Nesper, R. Angew. Chem. Int. Ed. 2002, 41,
24462461.
103. Maekawa, H. Materia (Nihon Kinzoku Gakkai) 2006, 464472 (part 2), 45(part 1),
359370.
104. Soler-Illia, G. J. A. A.; Crepaldi, E. L.; Grosso, D.; Sanchez, C. Curr. Opin. Coll.
Int. Sci. 2003, 8, 109126.
105. Di Renzo, F.; Galarneau, A.; Trens, P.; Fajula, F.; Di Renzo, F.; Galarneau, A.;
Trens, P.; Fajula, F. In Handbook of Porous Solids; Schuth, F., Sing, K. S. W.,
Weitkamp, J., Eds.; Wiley-VCH: Weinheim, 2002; pp 13111394.
106. Iskandar, F.; Mikrajuddin, F.; Okuyama, K. In Encyclopedia of Nanoscience and
Nanotechnoloy; Nalwa, H. S., Ed.; 8, American Scientific Publishers: Valencia, CA,
2004; pp 259270.
107. de Soler-Illia, G. J.; Sanchez, C.; Lebeau, B.; Patarin, J. Chem. Rev. 2002, 102,
40934138.
108. Wan, Y.; Yang, H.; Zhao, D. In Biomolecular Catalysis ACS Symposium Series; ;
Jungbae, K., Seong, H. K., Ping, W., Eds.; American Chemical Society:
Washington, DC, 2008; vol. 986, pp 248.
109. Wan, Y.; Yang, H.; Zhao, D. Acc. Chem. Res. 2006, 39, 423432.
110. Gao, W.; Li, Z. In Physical Properties and Applications of Polymer Nanocomposites;
Tjong, S. C., Mai, Y.-W., Eds.; Woodhead: Cambridge, 2010; pp 723742.
111. (a) Zhu, X.; Liu, Z.; Ming, N. J. Nanosci. Nanotechnol. 2010, 10, 41094123;
(b) Zhu, X.; Liu, Z.; Ming, N. J. Mater. Chem. 2010, 20, 40154030.
112. Ostrikov, K.; Levchenko, I.; Xu, S.; Huang, S. Y.; Cheng, Q. J. Long, J.; Xu, M.
Thin Solid Films 2008, 516, 66096615.
113. Cao, J.; Wu, J. Mater. Sci. Eng. R Rep. 2011, R71(24), 3552.
114. Mittendorfer, F. J. Phys. Condens. Matter 2010, 22, 393001/1393001/11.
115. (a) Netzer, F. P.; Allegretti, F.; Surnev, S. J. Vac. Sci. Tech. B Microelectron.
Nanometer Struct. -Process. Meas. Phenom. 2010, 28, 116; (b) Netzer, F. P.
Surf. Sci. 2010, 604, 485489.
116. Ueda, W.; Sadakane, M.; Ogihara, H. Catal. Today 2008, 132, 28.
117. Chen, X.-W.; Zhu, Z.; Haevecker, M.; Su, D.; Schlogl, R. Mater. Res. Bull 2007,
42, 354361.
118. Wang, L.-C.; Liu, Y.-M.; Chen, M.; Cao, Y.; He, H.-Y.; Fan, K.-N. J. Phys. Chem.
C 2008, 112, 69816987.
119. Rajeswari, J.; Viswanathan, B.; Varadarajan, T. K. Mater. Chem. Phys. 2007, 106,
168174.
120. Zhong, Z.; Ho, J.; Teo, J.; Shen, S.; Gedanken, A. Chem. Mater. 2007, 19,
47764782.
Mixed-Metal Oxides 183
121. Cortes-Jacome, M. A.; Morales, M.; Angeles, Chavez C.; Ramirez-Verduzco, L. F.;
Lopez-Salinas, E.; Toledo-Antonio, J. A. Chem. Mater. 2007, 19, 66056614.
122. (a) Teng, F.; Gaugeu, B.; Liang, S.; Zhu, Y. Appl. Catal. A: Gen. 2007, 328,
156162; (b) Nian, J.-N.; Chen, S.-A.; Tsai, C.-C.; Teng, H. J. Phys. Chem. B
2006, 110, 2581725824.
123. (a) Ogihara, H.; Sadakane, M.; Ueda, W. Top. Appl. Phys. 2010, 117,
147158; (b) Ueda, W.; Sadakane, M.; Ogihara, H. Catal. Today 2008,
132, 28.
124. Miyauchi, M.; Tokudome, H. Top. Appl. Phys. 2010, 117, 4557.
125. (a) Centi, G.; Perathoner, S. In Catalysis; Spivey, J. J., Ed.; RSC: Cambridge,
2007; vol. 20, pp 367394; (b) Centi, G.; Perathoner, S. In Catalysis; Spivey, J. J.,
Ed.; RSC: Cambridge, 2009; vol. 21, pp 82130.
126. Centi, G.; Perathoner, S. In Nanoporous Materials for Energy and the
Environment; Rios, G., Centi, G., Kanellopoulos, N., Eds.; Pan Stanford
Publishing: Singapore, 2012; pp 257282.
127. Lee, M.; Kim, T. W.; Bae, C.; Shin, H.; Kim, J. JOM 2010, 62, 4449.
128. Cao, G.; Limmer, S. J. In Encyclopedia of Nanoscience and Nanotechnoloy;
Nalwa, H. S., Ed.; 8, American Scientific Publishers: Valencia, CA, 2004;
pp 377396.
129. Sharma, S.; Li, H.; Chandrasekaran, H.; Mani, R. C.; Sunkara, M. K.
In Encyclopedia of Nanoscience and Nanotechnoloy; Nalwa, H. S., Ed.; 10,
American Scientific Publishers: Valencia, CA, 2004; pp 327346.
130. (a) Shang, H. M.; Wang, Y.; Cao, G. Key Eng. Mat. 2007, Trans Tech Publications
Ltd, 336338, 21222127; (b) Wang, Y.; Cao, G. Chem. Mater. 2006, 18,
27872804.
131. Wang, X.; Sun, X.; Xu, J.; Li, Y. In Nanowires and Nanobelts; Wang, Z. L., Ed.; 2,
Kluwer Academic Publishers: Norwell, MA, 2003; pp 173188.
132. Mao, Y.; Park, T.-J.; Wong, S. S. Chem. Commun. 2005, 57215735.
133. Li, Y.; Yang, X.-Y.; Feng, Y.; Yuan, Z.-Y.; Su, B.-L. Crit. Rev. Solid State Mater.
Sci. 2012, 37, 174.
134. Pan, C.; Zhang, D.; Shi, L.; Fang, J. Eur. J. Inorg. Chem. 2008, 24292436.
135. Perathoner, S.; Passalacqua, R.; Centi, G.; Su, D. S.; Weinberg, G. Catal. Today
2007, 122, 313.
136. Frunza, L.; Frunza, S.; Schonhals, A.; Bentrup, U.; Fricke, R.; Pitsch, I.;
Kosslick, H. Stud. Surf. Sci. Catal. 2002, 142B, 13231330.
137. Dosseh, G.; Xia, Y.; Alba-Simionesco, C. J. Phys. Chem. B 2003, 107,
64456453.
138. Rother, G.; Woywod, D.; Schoen, M.; Findenegg, G. J. Chem. Phys. 2004, 120,
1186411873.
139. Floquet, N.; Coulomb, J. P.; Andre, G.; Kahn, R. Stud. Surf. Sci. Catal. 2007, 160,
375382.
140. (a) Trens, P.; Tanchoux, N.; Galarneau, A.; Fajula, F. J. Phys. Chem. B 2005, 109,
1641516420; (b) Pariente, S.; Trens, P.; Fajula, F.; Di Renzo, F.; Tanchoux, N.
Appl. Catal. A: Gen. 2006, 307, 5157.
141. Chen, W.-H.; Huang, S.-J.; Zhao, Q.; Lin, H.-P.; Mou, C.-Y.; Liu, S.-B. Top. Catal.
2009, 52, 211.
142. Sastre, G.; Corma, A. J. Mol. Catal. A: Chem. 2009, 305, 37.
143. Kidder, M. K.; Buchanan, A. C., III J. Phys. Chem. C 2008, 112, 30273031.
144. Ouali, A.; Caminade, A.-M. In Dendrimers: Towards Catalytic, Material and
Biomedical Uses; Caminade, A.-M., Cedric-Olivier, T., Regis, L.,
Armelle Ouali, B. D.-N., Eds.; John Wiley and Sons: New York, 2011;
pp 183195.
145. He, C.; Zhang, F.; Yue, L.; Shang, X.; Chen, J.; Hao, Z. Appl. Catal. B: Environ.
2012, 111112, 4657.
146. Yang, Z.; Guo, S.; Pan, X.; Wang, J.; Bao, X. Energy Environ. Sci. 2011, 4,
45004503.
147. Pan, X.; Bao, X. Chem. Commun. 2008, 62716281.
148. Chen, X.; Wang, H.; Wu, Z.; Liu, Y.; Weng, X. J. Phys. Chem. C 2011, 115,
1747917484.
149. Ballester, P.; Vidal-Ferran, A.; van Leeuwen, P. W. N. M. Adv. Catal. 2011, 54,
63126.
150. Meeuwissen, J.; Reek, J. N. H. Nat. Chem. 2010, 2, 615621.
151. Gall, B.; Bortenschlager, M.; Nuyken, O.; Weberskirch, R. Macromol. Chem. Phys.
2008, 209, 11521159.
152. Koblenz, T. S.; Wassenaar, J.; Reek, J. N. H. Chem. Soc. Rev. 2008, 37,
247262.
153. Neumann, T.; Haupt, B.; Ballauff, M. Macromol. Biosci. 2004, 4, 1320.
154. Xiao, P.; Garcia, B. B.; Guo, Q.; Liu, D.; Cao, G. Electrochem. Commun. 2007, 9,
24412447.
155. Bao, S.-J.; Li, C. M.; Zang, J.-F.; Cui, X.-Q.; Qiao, Y.; Guo, J. Adv. Funct. Mater.
2008, 18, 591599.
156. Brune, A.; Jeong, G.; Liddell, P. A.; Sotomura, T.; Moore, T. A.; Moore, A. L.;
Gust, D. Langmuir 2004, 20, 83668371.
184 Mixed-Metal Oxides
157. Kieboom, T. In Catalysis for Renewables: From Feedstock to Energy
Production; Centi, G., van Santen, R. A., Eds.; Wiley-VCH: Weinheim, 2007;
pp 273298.
158. (a) Rousseau, S.; Marie, O.; Bazin, P.; Daturi, M.; Verdier, S.; Harle, V. J. Am.
Chem. Soc. 2010, 132, 1083210841; (b) Bazin, P.; Thomas, S.; Marie, O.;
Daturi, M. Catal. Today 2012, 182, 311.
159. (a) Centi, G.; Perathoner, S. Catal. Rev. -Sci. Eng. 1998, 40, 175208;
(b) Centi, G.; Perathoner, S. Appl. Catal. A: Gen. 1995, 132, 179259.
160. Mann, J. E.; Mayhall, N. J.; Jarrold, C. C. Chem. Phys. Lett. 2012, 525526, 112.
161. Wegener, S. L.; Marks, T. J.; Stair, P. C. Acc. Chem. Res. 2012, 45, 206214.
162. Trunschke, A. RSC Nanosci. Nanotech. 2011, 19, 5695.
163. Lopez Nieto, J. M.; Solsona, B.; Ivars, F.; Concepcion, P.; Melo, F. In Advances in
Chemical Engineering; San Jose, M. J., Ed.; Research Signpost: Kerala, 2010;
pp 149174.
164. Lercher, J. A.; Naraschewski, F. N. RSC Nanosci. Nanotech. 2011, 19, 532.
165. Haber, J. Catal. Today 2009, 142, 100113.
166. Hess, C. Chemphyschem 2009, 10, 319326.
167. Shiju, N. R.; Guliants, V. V. Appl. Catal. A: Gen. 2009, 356, 117.
168. Busca, G. In Metal Oxides; Fierro, J. L. G., Ed.; CRC (Taylor and Francis): Boca
Raton, FL, 2006; pp 247318.
169. Papakondylis, A. A.; Sautet, P. J. Phys. Chem. 1996, 100, 1068110688.
170. Busca, G.; Ramis, G.; Lorenzelli, V. J. Mol. Catal. 1989, 50, 231240.
171. Glazneva, T. S.; Kotsarenko, N. S.; Paukshtis, E. A. Kinet. Catal. 2008, 49, 859867.
172. Arata, K. In Metal Oxide Catalysis; Jackson, S., Hargreaves, J. S., Eds.;
Wiley-VCH: Weinheim, 2009; vol. 2, pp 665704.
173. Takagaki, A.; Tagusagawa, C.; Hayashi, S.; Hara, M.; Domen, K. Energy Environ.
Sci. 2010, 3, 8293.
174. Macht, J.; Iglesia, E. Phys. Chem. Chem. Phys. 2008, 10, 53315343.
175. Centi, G.; Perathoner, S. Catal. Today 1998, 41, 457469.
176. Seitza, H.; Germana, E.; Juana, A.; Irigoyenb, B. Appl. Surf. Sci. 2012, 258,
36173623.
177. (a) Centi, G.; Marchi, F.; Perathoner, S. Appl. Catal. A: Gen. 1997, 149, 225244;
(b) Centi, G.; Marchi, F.; Perathoner, S. J. Chem. Soc. Faraday Trans. 1996, 92,
51515159.
 5.07 Porous Metals and Metal Oxides
SL Suib, The University of Connecticut, Storrs, CT, USA
2013 Elsevier Ltd. All rights reserved.

5.07.1 Review 151

5.07.2 Porosity 151
5.07.3 Synthesis of Porous Metals 151
5.07.3.1 Macroporous Metals 154
5.07.4 Applications of Porous Metals 154
5.07.5 Synthesis of Porous Metal Oxides 156
5.07.5.1 Microporous Materials 156
5.07.5.2 Mesoporous Materials 159
5.07.5.3 Macroporous Materials 164
5.07.6 Applications of Porous Metal Oxides 164
5.07.7 Conclusion 166
References 167

5.07.1 Review materials. Various isotherms are used to describe various

This chapter consists of five major sections. Section 5.07.2
concerns porosity, methods used to study porosity, and general
principles. Section 5.07.3 concerns synthesis of porous metals.
This section concerns various newly reported methods used
to make porous metal materials. Section 5.07.4 includes
information on new applications of porous metals. In addition,
properties that are needed for such applications are discussed.
Section 5.07.5 involves synthesis of porous metal oxide
materials. The focus is primarily on transition-metal oxide
materials. Section 5.07.6 concerns new applications of porous
metal oxide systems. Articles from the last 5 years are reviewed
in order to summarize the most recent work that is available
in the literature.
5.07.2 Porosity
Porosity is an important property of various inorganic mate-
rials. The International Union of Pure and Applied Chemistry
(IUPAC) definition of the three major types of porous mate-
rials is based on relative size of pores. These three types of
porous materials include micropores (<2 nm), mesopores
(250 nm), and macropores (>50 nm) as well as sometimes
combinations of these.1 Porosity is often referred to as void
volume or void fraction. The void fraction is a measure of the
empty or void space in a material. Void fraction varies from
0 to 1 and is defined as the fraction of the void volume over
total volume. An example of a high void fraction is 48% for
zeolite Y, a very porous material.
Surface area is another important property of inorganic
materials. Surface area is related to porosity as well as other
factors such as particle size and particle shape. Measurements
of porosity and surface area can be done with a variety of
methods. Electron microscopy measurements can be used to
directly observe pores but can be limited due to being able to
study a limited number of particles and other factors such as
buried pores. Gas adsorption is often used to study porous
pore-size materials.2 Langmuir and BrunauerEmmettTeller
(BET) models are useful in understanding surface area, site
occupancy, porosity, and other phenomena.
Types I, IV, and V adsorption isotherms are associated with
porous materials.2 Type I isotherms are typical of microporous
materials. DeBoer defined five types of hysteresis loops found
in isotherms of mesoporous materials, known as A, B, C, D,
and E where type A represents cylindrical pores, type B repre-
sents either slit-shaped pore or spaces between parallel plates,
type C is indicative of a mixture of wedge-shaped and tapered
pores open at both ends, type D is like type C but has narrow
necks at either one or both ends, and type E represents ink
bottle or bottleneck pores.
The desorption branch of the isotherm is often used to
measure pore-size distributions. Typically, nitrogen gas is
used in these measurements although other inert gases can
also be used. There are certainly other methods that are used
to study porous materials including gravimetric methods, mer-
cury porosimetry, density measurements, pressure jump, and
other methods.
5.07.3 Synthesis of Porous Metals
In the last 5 years there has been a tremendous emphasis on
nanosize particles. Some of these materials are porous and
some are nonporous. The general notion that high surface
area will enhance catalytic activity and therefore small nano-
size particles are needed is often correct but this depends on the
nature of the reaction. Structure-sensitive (demanding) reac-
tions are ones where rates may depend on either the particle
size or the types of surface atoms, whereas structure-insensitive
(facile) reactions do not. In the same vein, porous material
may enhance the rate of a catalytic reaction or may be inactive
depending on the nature of the reaction. For example, shape-
selective catalytic reactions over porous zeolite materials can be
enhanced when reactants are small enough to enter pores
where active sites are located. These are just a few examples

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00508-8 151

152 Porous Metals and Metal Oxides
showing the importance of pores, control of pore-size distri-
butions, and physical properties of inorganic materials.
The control of properties often results from the controlled
synthesis of a material. Control of pore sizes and pore-size
distributions is a difficult area of research. Very often, such
pore-size distributions are wide and there are several maxima
in the distributions. One method to control porosity is to use
templates. Methods that avoid templates are continually being
sought since such templates are often expensive and can be
difficult to recover. Monolithic porous Ni materials have been
prepared using a nontemplate route.3 For example, NiZn oxa-
lates have been used as precursors to porous NiO, which is then
reduced in hydrogen to form porous Ni metal. A similar
approach was used to prepare porous Fe by heating ZnFe2O4
in hydrogen with high flow rates of hydrogen. The pore dimen-
sions of such materials range from 1 to 5 mm but have narrow
pore-size distributions. This approach is a clever one and may be
extended to other metals. An extension of this work to other
precursors such as La4Ni3O10 with formation of NiO and con-
comitant reduction to Ni porous metal has also been done.4
Besides selective dissolution of mixed metal oxides fol-
lowed by reduction, dealloying by removal of one or more
metals from an alloy has recently become a major technique
for making porous metals.5 One example is the dealloying of Si
by etching of Si from PtSi alloys. In this way the amount of Pt
can be limited and controlled pore sizes can be obtained. The
resultant Pt porous films are excellent catalysts for methanol
electro-oxidation.6 Selective etching of AgAu alloys has also
been done and yields porous Au particles that are active in the
gas-phase oxidation of benzyl alcohol. Low levels of remain-
ing Ag may be important in the catalytic activity but this
was not confirmed. Such catalysts were active under ambient
conditions.7 The presence of small amounts of a second metal
would appear to be important in other applications such as the
electrooxidation of methanol.8 In this case, dealloying of Au
Ag alloys to produce porous Au may profit from small amounts
of remaining Ag. Intentional addition of Pt by surface deposi-
tion on porous Au significantly enhanced the electrooxidation
of methanol as indicated by increased anodic peak intensity.
Preparation of novel alloy materials with selective dealloying
would appear to be an area that is currently growing and may
afford new unique porous systems that cannot be produced
with other methods.
Another recent method that has been used to make porous
metals is atomic layer deposition (ALD). ALD allows deposi-
tion of atomic layers of various materials and is quite versatile
in terms of resultant structures and compositions. One exam-
ple is the deposition of Pt onto porous metals such as Ag, Pd,
Ru, and Au.9 These materials were tested in solid oxide fuel
cells as cathodes. The ALD method is in its infancy and one
expects that many new porous materials will be prepared by
clever design of materials at the atomic level. Figure 1 is a
drawing of processes that occur during the ALD process. Dur-
ing the first step a clean surface is exposed to a precursor.
Typically this is an organometallic small molecule with a lim-
ited number of different atoms. A multilayer will be present, so
there needs to be a purge step that then allows deposition of an
atomic layer. Then a second precursor can be added, followed
by another purge step. The purge gas is typically an inert gas.
Some potential drawbacks of the ALD method that provide
Substrate Substrate
Purge gas
2nd precursor
Substrate Purge gas Substrate
Figure 1 Diagram of processes occurring during atomic layer
deposition.
room for exploration include the expense of precursors, the
relatively few types of reactant precursors currently available,
multiple phase or composition materials, and the difficulty in
characterizing the resultant structures. Nevertheless, ALD pro-
vides a relatively simple way to produce porous thin films and
potentially highly reactive systems that cannot readily be
formed with other methods.
Aerogels can lead to extremely porous materials. Reactions
between nanoparticles in aerogels are controlled by high sur-
face to volume ratios and high compactness.10 Interpenetrating
networks of resorcinolformaldehyde and metal oxides of Fe,
Co, Ni, Sn, Cu, Cr, Ti, Hf, Y, and Dy nanoparticles were
produced using aerogel methods. These are one-pot processes
that use acidic gels of hydrated metal ions. Acidity drives the
gelation of the resorcinolformaldehyde. Supercritical fluid
drying using CO2 produces porous aerogels, although drying
at ambient pressure produces compact xerogel materials.
Macroporous materials consisting of xerogel walls can be
formed by using a polyurea coating to cause crosslinking of
the metal oxide nanoparticles that have limited stability and
are porous to 200  C. The metal oxide particles are then pyro-
lyzed up to 800  C and produce monoliths of metal such as Fe,
Co, Ni, Sn, and Cu. Carbothermal reactions can also occur and
predominate for metals such as Cr, Ti, and Hf leading to metal
carbide materials. The innate compactness of the xerogels and
the induced skeletal compactness in the aerogels are both
important factors in determining the resultant porosity. Inter-
estingly, attempts to make such systems by high-temperature
mixing do not lead to any of these phases. This approach that
involves generation of resorcinolformaldehyde metal oxide
networks is intriguing and suggests that similar procedures
may be used to make a variety of novel porous materials.
The use of organic materials that can be evolved during
synthesis to generate porosity is an approach that has been
used for many years. Combinations of synthetic methods on
the other hand are becoming more common in order to con-
trol resultant physical and chemical properties. Mechano-
chemical methods using a planetary mill (high-speed ball
mill) have been used to prepare porous copper materials.11
In this synthesis, cellulose/copper blends were used as precur-
sors. Surface properties were studied using surface potential
microscopic mode of a scanning probe microscope. The cellu-
lose material became amorphous during this treatment as
shown by XRD experiments. The copper on the other hand
did not change in terms of structural properties. Various

methods were used to remove the cellulose including dissolution
using enzymes as well as high-temperature treatment. The
resultant Cu material has a surface area of 18 m2 g1 with
fragments of Cu that are connected at different junctions.
This leads to a porous material due to decreased cohesion
of the Cu particles. In many cases, various metals aggregate
and sinter due to strong cohesive forces that lead to
decreased porosity and low-surface-area materials. The use
of other organic agents would be an area to investigate,
especially if such organic moieties can be used to tailor the
pore structure of such systems.
Porous Al is an excellent material for sound absorption.
Porous Al can be prepared by using a spacer method where
Al powder and space holding NaCl particles are sintered.12
Pore size and porosity can both be controlled by varying the
ratios of these constituents. Pore sizes of 212300 and 610
700 mm have been realized in these preparations. Porosities
from 85 to 95% can be obtained with this method. The motion
of air depends on porosity, so it is reasonable to associate
porosity with sound absorption capability. Mechanical stabil-
ity and corrosion resistance are important in these applications
and are provided by porous Al. In situ studies of the develop-
ment of these pores would be of interest.
Cu metal and an Al space holder can be deformed together
by bulk forming.13 In such composites, the Al space holder is
then removed. Cold extrusion methods and chemical leaching
are used to make porous Cu materials in this process. The Cu
material is of high strength and can be formed as a honeycomb
structure. Lotus-type Cu systems, which have oriented pores
parallel to the solidification direction, are also formed depend-
ing on conditions. The honeycomb Cu material and the Cu
lotus-type material are robust materials with higher specific
yield strengths than those made using conventional methods.
The initial steps of this method are similar to the dealloying
methods described above.
Another approach that uses a different type of spacer is the
unidirectional solidification of gas dissolved in a melt.14 Porous
metals can be produced with this method and long cylindrical
pores that are aligned in one direction can be obtained. This
method involves the unidirectional solidification using pressur-
ized hydrogen gas method followed by thermal decomposition.
When the hydrogen gas becomes insoluble in the metal, then
the H2 is evolved and pores are formed. In the conventional
platinum group metals (PGMs), the hydrogen pressurized in a
high-pressure chamber is used as the dissolving gas. Since there
are safety concerns when dealing with hydrogen, another
approach involving the thermal treatment of gas-evolving pre-
cursors was also developed. In this case, gas-evolving species
Top down
Scale
m Bulk systems
Films
Lithography
nm Quantum dots, size effects
Figure 2 The top-down and bottom-up approaches to synthesis of materials.
Porous Metals and Metal Oxides 153
such as ammonia borane were added to the molten metal.
Lotus-type metals can be produced in this way. Both porous
copper and aluminum were produced with these methods.
Surfaces of all these materials can be modified using a
variety of methods. A friction-surface-modifying and rolling
method was used on Al foams to modify the surface of the
particles.15 Such surfaces are considerably smoother and more
dense than those of unprocessed foams of Al. The number and
morphology of the residual pores of the treated surfaces are
controlled by the friction surface modification procedure. An
additional rolling step appears to create the smoothest surface.
The treated Al foam samples had the highest average bonding
strengths, yield strengths, and toughness values of both treated
and untreated samples. The final rolling process after the fric-
tion surface modification was helpful in enhancing the bond-
ing character of the Al foams primarily by smoothing the
surfaces. Control of surface bonding is important in many
applications, so this rolling process may be exploited in other
types of porous metals to control the bonding and morphol-
ogy of such surfaces.
Pattern-forming instabilities, theoretical modeling, and a
review of the preparation of porous metals using dealloying
types of methods are available.16 This chapter suggests that
kinetics are important in controlling the tunability of porous
metals. A general model of porosity evolution has been pro-
posed that involves competitive kinetic processes of electro-
chemical dissolution, surface diffusion, and capillary action.
Evolution of porosity was demonstrated with a Ag/Au alloy.
During initial stages of pore development, the less noble com-
ponent of such binary alloys is dissolved. Atomic-scale kinetic
Monte Carlo simulations were used to develop this model. The
dissolution occurs in a layer-by-layer way, with Au atoms accu-
mulating at step edges. The Au atoms cover surface mound-like
areas and there are passivating processes that also occur leading
to equilibrated adatoms. These Au-coated mounds are under-
cut and some new gold atoms are used in the creation of new
mounds rather than add to already occurring mounds. As this
process developed, pores are formed. A coreshell microstruc-
ture is formed that has a base alloy core of Ag and a noble
metal skin of Au. Secondary coarsening processes occur that
coarsen these microstructures and lead to enrichment of the
noble metal Au component in these porous metals.
An excellent review regarding porous Pt concerns synthesis as
well as potential applications.17 There is a discussion of both
top-down and bottom-up strategies. Chemical, photochemical,
sonochemical, and other activation processes are discussed.
Figure 2 shows the two different strategies of making materials
from a top-down versus a bottom-up approach. There are various
Bottom up
Scale
Nanosize materials nm
Wires
Biological materials
Atoms 0.1 nm
154 Porous Metals and Metal Oxides
methods that can be used in either the top-down or the bottom-
up approach and different types of materials can result depend-
ing on the scale and size of the particles. Thin films are a major
focus of this chapter. The major application discussed is the
preparation of Pt-based porous films for electrochemical
applications.
Porous metals such as titanium have a number of useful
properties in dental and orthopedic applications. Powder met-
allurgical syntheses of titanium alloys with Sn or Nb lead to
porous foams that mimic bone. A vacuum furnace is used to
sinter the alloys. Initial powders were ball milled. Densities as
low as 0.4 g cc1 were observed. This is a unique way to make
such foams.18
5.07.3.1 Macroporous Metals
Sodium acrylate-stabilized gold nanoparticles have been used to
adsorb irreversibly onto crosslinked poly(acrylamide) beads.19
Large (>0.5 mm) macroporous gold beads have been prepared
in this way by using porous polymers as scaffolds. An oil-in-
water-in-oil (O/W/O) sedimentation polymerization method
was used to synthesize the porous, emulsion-templated poly
(acrylamide) scaffold. This clever approach and design lead to
the control of porosity and chemical composition of these
polymers over a very wide range. Changing the emulsion com-
position and the type of monomer gives excellent control of
porosity. The method has been extended to produce hierarchi-
cally porous inorganic materials silica and metal oxides using
this O/W/O method. An aqueous gold nanoparticle solution
was used to make these emulsion-templated gold materials by
soaking the poly(acrylamide) beads in solution. The gold nano-
particles are irreversibly adsorbed onto the poly(acrylamide)
surface. When this happens the particles do not aggregate. Fine-
tuning of this method has led to generation of single site iso-
lated nanosize particles. Microscopy studies have shown that
these gold beads have an open porous structure, which consists
of emulsion-templated cells on the order of 10 mm in diameter
connected by windows that are on the order of 5 mm.
This method leads to very large pore structure materials.19
The use of templated gold particles has several advantages such
as the production of more pure materials than would be
obtained using solution-based gold salt precursors. Such gold
salt precursors also form replicas of the poly(acrylamide) struc-
ture but these are not as well ordered. Further control of
porosity was obtained by changing the immersion time and
the ratio of gold particles to polymer. The distribution of
nanoparticles in the composite was controlled in this way.
The resultant materials were hollow gold beads made by the
emulsion template method. Such structures were only
observed when gold particle precursors were used rather than
with the gold salt precursors. This formation of macroporous
gold particles may be controlled by the rate of adsorption of
the nanoparticles onto the surface of the polymer, which is
much greater than the rate of diffusion of the nanoparticles
throughout the interconnected pores. The method appears to
be general and can be extended to other metals such as
palladium.19 In addition, hybrid materials such as mixed pal-
ladium platinum systems should be able to be produced. This
method is simple, versatile, and potentially could be optimized
to control porosity at other levels like in the mesopore regime.
5.07.4 Applications of Porous Metals
This section summarizes some recent applications of porous
metals. This is not meant to be all inclusive but rather to empha-
size the most current applications that have been published.
Orthopedic applications of porous metals are continually
important in the field of biomedicine. Titanium and titanium
alloys are often used in orthopedics and dental applications
since they have excellent mechanical properties, chemical sta-
bility, and they are biocompatible. There are several inherent
limitations of these materials such as their low porosity, high
modulus of elasticity, and they are often not easily deposited as
coatings. A new commercial Trabecular Titanium (which
was developed in order to overcome these barriers) has been
developed.20 Stem cell studies have been carried out with
Trabecular Titanium and this material has provided an
enhanced understanding of the mechanism of bone regenera-
tion at the molecular scale and is often used as a reference
material to compare to other materials.
Various methods are used to study bone growth. The exact
mechanism of bone intergrowth into Ti-based implants and
how they function is not known. Generally and traditionally,
complicated time-consuming procedures that include tissue
processing, embedding, sectioning, polishing, imaging, and
image analyses are used. Micro-computed tomography has
recently been used to study how tissues respond to porous
titanium implants.21 This method is a three-dimensional
(3D) imaging technique that can be used to quantify the tissue
response to porous metals. Cylindrical implants were
implanted into rabbits for a 6-week period and a comparison
was made using traditional and micro-computed tomography
methods. The results obtained from micro-computed tomog-
raphy were very comparable to results obtained from the tra-
ditional histological methods. Tomography data show that
bone was present in the pores that were connected with open-
ings that are larger than 9.4 mm. The size of the interconnection
had little impact on bone density for pore openings that were
greater than 28.2 mm for these porous titanium foams. Such
tomographic methods of analysis seem promising in terms of
obtaining a 3D image of the pore structure for bone growth.
While porous implants are known to promote cell adhesion
and have low elastic modulus, pores can lead to a decreased
usage lifetime due to fatigue. Studies of fatigue of porous
metals are rare.22 High-cycle rotating bending fatigue
responses of porous NiTi alloys made by laser-engineered net
shaping (LENS) has recently been reported22. Mechanical
properties and features of the microstructure were studied. If
porosity in these NiTi alloys is on the order of 10%, then
rotating bending fatigue results show that the fatigue failure
stress of 106 cycles was as high as 54%. Such fatigue analyses
show that surface porosity is extremely important in the fatigue
failure of these porous NiTi materials.
The use of porous metal foam in orthopedics is a relatively
recent development. Such foams are more porous than tradi-
tional coatings. Contact of the implant with bone is crucial in
such applications. Bone ingrowth and interlocking mecha-
nisms rely on specific surfaces and porosity. An in vivo study
of ingrowth into Ti implants covered with a novel Ti foam
coating has been done.23 The Ti foam coating has 50% in
volume of interconnected pores and a considerably higher

surface area as compared to dense-phase Ti. Rat tibia were used
for these implant studies. Two weeks after implantation, the
amount of bone in the implants was studied using several
techniques such as x-ray microtomography and backscattered
electron imaging. The pores in the Ti foam showed that 97.7%
of the bone was filled. The bone implant area of contact was
significantly increased as compared to controls of the dense-
phase Ti. There are several advantages of these coated Ti
implants such as increased long-term bone formation, excel-
lent biocompatibility, and potential benefits as regards firm
anchoring of such surgical implants.
LENS methods can also be used to produce porous load-
bearing implants that have controlled microstructures and mac-
rostructures. For tissue regeneration after implantation, there
may need to be specific implants for different patients. There
are many challenges in preparing and using load-bearing
implants such as for hip prosthesis applications.24 Specific com-
mercial materials such as Tritanium (Stryker, Mahwah, NJ),
Regenerex (Biomet, Warsaw, IN), Stiktite (Smith and
Nephew, Memphis, TN), and Trabecular Metal (Zimmer,
Warsaw, IN) are all available now for use in orthopedic
surgery.25 Most implants used today are made by using tradi-
tional materials and methods such as sintered beads, fibers, and
plasma spray. All of these methods have inherent limitations.
The new porous metals have been recently developed in order to
enhance biocompatibility of such systems. All these materials
look like cancellous bone and have open-cell structures. This in
turn leads to unique properties such as high volumetric poros-
ities of 6080%, low elastic moduli, and high frictional charac-
teristics. Such highly porous materials and their advantages have
recently been reviewed.25 A similar review with applications in
orthopedic surgery is also available.52
Another area for applications of porous materials involves
battery materials. The stability of microstructures in battery
materials is critical as regards long-term usage. Rechargeable
lithium ion batteries rely on the stability of the structure of a
support material during charge and discharge. A diagram of a
lithium ion secondary battery is given in Figure 3. A unique
concept of the reversible formation of porosity in nanowires of
germanium was observed by in situ transmission electron
Electrons
Anode Separator Cathode
Li+ Li+ Li+ Li+
Li+ Li+ Li+ Li+
Li+ Li+ Li+ Li+
C6 + xLi+ + xe- LixC6 LiCoO2 xLi+ xe- Li1-xCoO2 heat storage is produced when porosities of such metals is greater
Figure 3 Diagram of a lithium ion rechargeable battery.
Porous Metals and Metal Oxides 155
microscopy (TEM).27 During lithium insertion, the Ge went
through two phase transformations: an intermediate LixGe
phase was first formed and then a Li15Ge4 phase was formed.
The latter phase is quite interesting in that the more Li that can
be incorporated the higher should be the specific capacitance
of such materials. Nanopores were observed to form during the
delithiation step due to aggregation of vacancies produced by
extraction of lithium. This process is very similar to the deal-
loying mechanisms that are prevalent in many of the studies
discussed above. A dense nanowire segment of crystalline
Li15Ge4 was formed with a porous sponge-like segment com-
posed of interconnected ligaments of amorphous Ge also pre-
sent. These porous nanowires show fast lithiation/delithiation
rates and excellent mechanical stability. The porous structure
of these materials enhances mass transfer and increases the rate
of lithium diffusion. This study suggests that Ge can develop a
reversible porous network structure and is therefore a promis-
ing anode material for lithium ion batteries that have excellent
energy capacity, rate performance, and cycle stability.
Another area for use of porous metals is ceramics. Cellular
solids are highly porous materials that have many different
morphologies and forms. Some of these forms include foams,
replicated porous or additive structures, honeycombs, lattice
structures, porous wires, interconnected fibers, and hollow
spheres. Pores in these materials can be totally open, partially
open, or closed. The porosity of these systems is one factor that
can determine potential engineering applications. Some possi-
ble applications of porous ceramics include auto exhaust cata-
lysts, filters for liquids and solid particulates in gas streams,
burners, biomedical devices, lightweight and stiff structures
such as for fishing poles and golf clubs, materials that can
absorb impact, devices that dampen sound, and materials used
for heat management. This particular chapter concerns ceramics
and porous metals that have >50 vol.% porosity.28
Magnetic properties can also be influenced by porosity.
Orientation of pores may be ordered magnetically. Lotus-type
porous nickel and cobalt were studied and magnetic data show
that the saturation magnetization of porous metals varies lin-
early with the porosity and appears to be independent of the
pore growth direction.29 The magnetization data are aniso-
tropic for Ni systems. The magnetization perpendicular to the
pore growth direction is always smaller than magnetization
that is parallel to the pore growth direction. The anisotropic
formation of pores is responsible for the observed magnetic
anisotropy. The dependence of magnetization on porosity
under an applied magnetic field follows Archies law. These
data are consistent with anisotropic electrical conductivities for
porous nickel in the parallel and perpendicular directions. This
chapter leads to the conclusion that there are direct relation-
ships between porosity and magnetization. Perhaps pores can
be filled with other reagents to markedly change the magnetic
character of such materials.
In fact, intentional addition of materials to pores has been
done in order to control and enhance thermal conductivity in
porous Cu and Al.30 Saturating the pores with phase-change
materials such as H2O, octadecane, Li2O3, and NaCl is done in
order to increase thermal conductivity. High capacity for latent
than 0.9. The optimum porosity corresponds to an increase in
the heat transfer coefficient. Heat transfer enhancement and the
156 Porous Metals and Metal Oxides
role of metal foams have been investigated.31 A nonequilibrium
heat transfer model was used which probed coupled heat con-
duction and natural convection and phase transitions and liquid
zones. Validation of the model was done by carrying out exper-
iments. Major observations are that heat conduction rate is
increased significantly by using metal foams. The enhanced
heat conduction rate is due to high thermal conductivities, and
suppression of natural convection due to the large flow resistance
in metal foams. When metal foams are embedded into phase-
change materials, overall heat transfer performance is improved.
These data suggest that enhancement of heat conduction out-
does the natural convection loss. Different metal foam samples
can have enhanced heat transfer rates when metal foams with
smaller porosities and larger pore densities are used. Predictive
studies have been done that further our understanding of hollow
sphere structures of porous metals.32 These systems are excellent
damping materials, have high stiffness, are energy absorbing,
and are thermal insulating. Such properties are enhanced by
using metallic hollow sphere structural materials.
5.07.5 Synthesis of Porous Metal Oxides
Porous transition-metal oxides consist of micropores (<2 nm),
mesopores (250 nm), and macropores (>50 nm) as well as
sometimes combinations of these.1 Considerable interest in
the control of pore sizes and pore-size distributions of such
materials has been a focus for quite some time. The control of
particle size in particular in the nanometer regime in the syn-
thesis of nanosize metal oxides is also currently being
pursued.33 Nanosize materials can have markedly different
properties compared to similar compositions that are bulky
(1 mm and above).34
Morphologies are potentially an important property of such
porous transition-metal oxide systems. Various morphologies
often are controlled by synthetic methods, synthetic conditions,
the nature of the materials, and by clever design. Examples of a
few different morphologies are given in Figure 4. Figure 4(a)
shows that TEM experiments can paint a somewhat different
picture compared to the scanning electron micrograph (SEM) of
the same material in Figure 4(b). Figure 4(c) shows porous
particles that have a spherical shape. The phase of the materials
in Figure 4(a)4(c) is the same as that of gamma MnO2. Many
times at the nanosize it is difficult to obtain nonaggregated
particles such as the situation of CeO2 systems of Figure 4(d).
0.2 mm
(a) (b) (c)
Figure 4 Various morphologies of porous transition-metal oxides: (a) TEM image of g-MnO2 synthesized for 15 min using batch UV light method;
(b) SEM image of g-MnO2 synthesized for 15 min using batch UV light method; (c) SEM image of g-MnO2 synthesized for 30 min using batch UV light
method; (d) SEM image of CeO2 synthesized at 1 ml min1 using continuous-flow UV light method.
These are some of the issues in this field as regards preparation
of single morphologies and controlled shapes.
Control of morphologies of porous transition-metal oxides
such as hollow spheres, rods, helices, spirals, and many other
shapes has been a major focus of researchers over at least the
last 10 years.35 Some representative porous metal oxides are
shown in Figure 5.
Such control comes from specific synthetic methods such as
use of templates, structure directors, surfactants, coreshell,
self-assembly, epitaxial growth, size reduction, capping agents,
solgel, and other methods.36 Morphologies can be controlled
by compositions including dopants. The parameters controlled
during syntheses include temperature, light, pH, point of zero
charge, stirring, high pressure, and others that are also clearly
important.37 While there is substantial research on incorpora-
tion of metal ions into nonmetal oxide materials, little of this
subject is discussed here in contrast to metal oxide materials.
5.07.5.1 Microporous Materials
There are several reviews of microporous and mesoporous
transition-metal oxide (MTMO) materials. Sadakane and Ueda
focus primarily on mesoporous materials, although there is some
discussion of microporous materials.38 Synthetic methods, char-
acterization of these systems, and applications are all covered.
Some of the more common microporous transition-metal oxides
are based on manganese oxide materials. A historical account of
the development of octahedral molecular sieve (OMS) and octa-
hedral layer (OL) materials based on porous mixed-valent man-
ganese oxides is available.39 Some unique properties of such
systems include excellent semiconductivity on the order of
doped Si in certain cases and controllable porosity. Both funda-
mental and applied properties of such systems are discussed. Such
OMS and OL materials have been used in studies of fuel cells,
catalysis, adsorption, sensors, batteries, and other applications.
A review of mesostructured nonsilicate oxides that have
well-defined structures on the order of 250 nm focuses on
the role of such materials in areas such as catalysis, energy
conversion, and biotechnology.40 Various methodologies
that use solgel processing combined with organic-directed
assembly have been used to make different mesostructured
oxides. The type of inorganic precursor is a major factor in
the synthesis of such materials. The goal of this research was to
develop general synthetic methods to produce nanosize,
hydrophilic, inorganic precursors that can be used with organic
(d)
 Porous Metals and Metal Oxides 157

Mn Mn Mn

Al Al Al
Al
Ti(2)
Ti(2) Ti(2)
Ti(2)

Ti(1) Mn Mn Mn
Ti(1) Ti(1) Mn

Ti(1) Ti(1) Al Al Al
Ti(1)

Ti(2) Ti(2)
Ti(2)
Ti(2)
Mn Mn Mn
Mn
Ti(1)
Ti(1) Ti(1)

Al Al Al
Al
c
Ti(1)
Ti(1)
Ti(1)

Ti(2) x
x b
Ti(2) Ti(2) Mn Mn Mn y
Ti(2)
z

(b) (c) (d)

(a)

y x

z
z x

(e) (f) y (g)

Figure 5 Some representative porous metal oxides: (a) Bi2Ti4O11, (b) cryptomelane, KMn8O16, (c) lithiophorite, MnO2(Al,Li)(OH)2, (d) brucite, Mg
(OH)2, (e) goethite FeOOH, (f) rutile, TiO2, G, sanmartinite ZnWO4.

co-assembly methods to generate stable materials. Organic- molecules, much like zeolitic systems. Other materials such as
acid-modified transition-metal oxide solgel precursors were [Cd3(UO2)6(PO3CH2CO2)6(H2O)13]6H2O were reported
used in these syntheses. that form rhombohedral channel structures with hydrated
Many of the materials that are reported here are mixed ions such as Cd(II) inside the channels, again reminiscent of
titania transition-metal oxide materials. Specific uses that zeolites. Selective adsorption of gases was observed for the
were investigated were optical and catalytic applications. cubic cobalt-based material, which shows excellent CO2
Structure-directing agents were used along with the solgel to uptake of 40 m2 g1 at 273 K, and no uptake at all of N2 at
form various mesostructures. One important parameter in the 77 K. Extensions of these fascinating materials to other com-
syntheses is control of pH. An interesting observation was that positions are anticipated.
to obtain nonsilicate oxides the most suitable solgel process Solgel methods have been used to produce various porous
depended on the desired product. In order to obtain thin films materials. Solgel syntheses were used to produce a novel class
that have crystalline frameworks, metal chlorides in alcoholic of alumina/silicatransition-metal-based systems.43 These sys-
solutions appear to be desirable. For multicomponent mesos- tems have type I adsorption and desorption isotherms with
tructured metal oxides or in the synthesis of large volumes of pores mainly in the range of 1020 A. The size and shape of the
material, a synthetic method was used which consists of a alkylamine template used in these preparations were used to
solution of acetic acid, aqueous HCl, ethanol, and metal alkox- control pore sizes, surface areas, and the pore volumes. These
ides. In order to obtain glasslike hybrids at room temperature solgel aluminosilicate systems were doped with transition
strongly binding organic additives such as trifluoroacetic acid metals with excellent control of porosity. Further control of
were used to obtain optical hybrid materials. These systems these systems was obtained by varying the Si/Al ratios, the
ranged from 2 to 50 nm in size. The selective leaching experi- transition-metal content, the amount of water used to control
ments discussed in the porous metal section have also been the rates of hydrolysis as well as how these materials were
used to make porous transition-metal oxide materials.3,4,41 dried. These mixed aluminosilicate transition-metal materials
Microporous materials of cubic and rhombohedral hetero- were characterized by a variety of methods including N2 phy-
bimetallic networks have been synthesized from uranium, tran- sisorption, solid-state magic angle spinning nuclear magnetic
sition metals, and phosphonoacetate.42 These unique synthetic resonance (MASNMR) methods, thermogravimetric analyses
methods have led to microporous materials. Several heterobi- (TGAs), magnetic measurements, and TEM. A major focus of
metallic U(VI)/M(II) (M Mn, Co, Cd) carboxyphosphonates the characterization work was to elucidate the nature of the
were synthesized in this work. Compositions such as relationship between the transition metal and the silica alu-
M2[(UO2)6(PO3CH2CO2)3O3(OH)(H2O)2]16H2O (M Mn mina matrix. Such combinations of transition-metal oxides,
(II), Co(II), and Cd(II)) have been produced that have cubic alumina, and silica are well known. Many of these systems
3D network structures that have large pores of about 16 A have been prepared using solgel methods. The solid-state
in diameter. These pores are filled with co-crystallized water MASNMR experiments used Hahn-echo and T1 studies and
158 Porous Metals and Metal Oxides
suggest that less than 4% of the transition metal is incorpo-
rated into the matrix. Like in many porous systems, use of
excess transition metal causes migration of the excess ions to
the pore surfaces and can lead to alterations in the pore struc-
ture. Tin dopants were some of the systems studied here. The
tin silica materials were used as catalysts in the BaeyerVilliger
oxidation reaction and such systems have been shown to have
high selectivity and conversion rates. Control of the pore struc-
ture and catalytic activity of these and related materials is a
focus of future research. While this study does not represent a
pure transition-metal oxide or lanthanide oxide set of mate-
rials, the importance of transition-metal ion doping into other
systems is clear. Such hybrid systems are the focus of a variety
of synthetic efforts and there are many more examples of such
combined materials. The same concepts that are used in the
synthesis of these hybrid systems are exploited in pure metal
oxide porous materials.
Another general concept that uses metal oxide materials to
create porous systems involves pillaring of layered materials.
Very often a layered material like a clay is used as a substrate
and metal oxide pillars are used to create high-surface-area
pillared clay or other pillared materials. The charge density of
the layers is often important in the pillaring process as regards
the number of pillars and types of pillars that can be introduced.
Porous clay heterostructures have been intercalated with silica
and with mixed silica titania pillars.44 The clay materials used in
this study were natural montmorillonite samples. The abun-
dance of a variety of natural clay materials such as montmoril-
lonite, hectorite, and others often leads to use of such substrates,
although in certain limited cases synthetic clays can be made
and used. The porous clay heterostructure materials reported
here were modified with copper or iron ions by using an ion-
exchange method. As with the above aluminosilicate transition-
metal oxide system43, numerous characterization methods were
used such as chemical analyses, structural x-ray powder diffrac-
tion (XRD) methods, surface area studies using BET methods,
determination of coordination and aggregation of transition-
metal species via ultravioletvisible (UVVis) diffuse reflectance
spectroscopy (DRS), electron paramagnetic resonance (EPR),
surface functional group and acidity measurements with Fourier
transform infrared (FTIR), and ammonia temperature-
programmed desorption (TPD) studies. The titanium ions intro-
duced into the silica pillars was present primarily as separate
cations. When the titanium loadings were high, some polymeric
titanium oxide species were observed. Clearly the amount of
transition-metal ion dopant can be limited in these and other
systems and this is a major factor in the preparation of any of
these systems. By adding titanium into the silica pillars, there was
a significant increase in the surface acidity of these porous clay
heterostructures materials. Transition metals such as Cu and Fe
when deposited on the surface of the porous clay heterostruc-
tures were mostly formed as isolated cations and oligomeric
metal oxide species. Transition-metal-modified Ti-containing
porous clay heterostructures were active and selective catalysts
for the decomposition of NOx. The Cu- and Fe-modified catalysts
were studied over a temperature range of 350550 and 300
500  C, respectively. These porous clay heterostructure catalysts
were somewhat deactivated by water and by SO2. Natural mont-
morillonite was used as the support structure and provided ion-
exchange sites and functional groups for anchoring of different
moieties. Both Lewis and Brnsted acid sites were observed by
titrating these acidic sites in these materials by using bases. The
acidity increased substantially by incorporation of titanium into
the porous clay heterostructures. Another role of the titanium
ions was to influence the type of transition-metal surface species
formed during the deposition process of copper and iron ions.
Transition-metal ions deposited on the surface of the porous clay
heterostructure supports were determined to be isolated cations
and small metal oxide clusters. When a series of Ti-modified
porous clay heterostructures with varying titanium content
were made, the role of iron and copper ions in the clusters was
more significant than in the clay which was only intercalated
with pure silica pillars.
Porous clay heterostructure samples with titanium addi-
tionally modified with transition metals were found to be
significantly more active than the catalysts based on porous
clay heterostructures that were intercalated with pillars of pure
silica. Ammonia oxidation is a side reaction which occurs
particularly when high loading Cu-doped catalysts were used
at temperatures up to 500  C.
Titanium was also incorporated into g-MnO2 OMSs that
had hollow sphere morphologies.45 A one-step synthesis was
used to make these materials. Titanium cations were used as
structure-directing agents to prepare the hollow spherical struc-
tures. The initial stages of the reaction are critical as regards the
formation of the hollow spheres. A variety of characterization
methods were used to study these systems such as XRD, N2
adsorption, SEM, energy dispersive x-ray (EDX) analysis,
Raman spectroscopy, TEM, x-ray photoelectron spectroscopy
(XPS), and TGA. The incorporation of titanium in these sys-
tems into the framework of the manganese oxide porous mate-
rials occurs in isolated sites. Separate phases of titania such as
anatase, rutile, or brookite are not observed. Such systems were
studied as electrodes in lithium air batteries of medium size.
These systems showed very high specific capacities of up to
2.3 Ah g1. The titanium-doped g-MnO2 OMS materials were
also used for the catalytic oxidation of toluene by using molec-
ular oxygen as the oxidant at atmospheric pressure. A key to
incorporation of dopants into the frameworks of such mate-
rials is introduction of the dopant at the initial stages of syn-
thesis. The specific method used to prepare g-MnO2 OMSs
doped with titanium was a co-precipitation method. In these
syntheses, titanium cations were used as structure-directing
agents. Characterization studies provide evidence that the tita-
nium dopants were introduced into the framework of the
g-MnO2 OMSs. The toluene oxidation experiments at atmo-
spheric pressure showed high conversion. Localized Ostwald
ripening was proposed as a mechanism of self-transformation
of the titanium containing g-MnO2 OMSs. This research pro-
vides a general, novel, and simple pathway for the synthesis of
hollow spheres with strong electrocatalytic and oxidative cata-
lytic activities in Li-air batteries and chemical oxidation reac-
tions, respectively, which may offer new material platforms for
batteries and chemical reactions.
Growth of oriented nanosize crystalline materials is impor-
tant for optimization of physical and chemical properties and
has been a major goal of many research groups for quite some
time. Various applications include oriented magnetic mate-
rials, optoelectronics, energy, and other areas. Arrays of nano-
size particles and of specific crystalline faces can be important
in catalytic reactions, absorption of light, strength of materials,
in sensors, and many other applications. The control of shape

is very difficult. Control of monomodal size distributions
especially over a large size scale is also very complicated. Con-
trol of growth planes is difficult and is usually not a process
available for all materials. Most of the reported methods that
have led to controlled structures require complicated proce-
dures, surfactants, and/or templates, or expensive instrumen-
tation and methods to produce such nanoscale arrays. The
research reported here46 exploited the multitude of available
porous manganese oxide structures in order to grow titania in
an oriented heteroepitaxial manner. A nanoscale shell-fiber
superstructure (NSFS) of rutile TiO2 was produced heteroepi-
taxially on porous manganese oxide (cryptomelane-like MnO2,
also known as OMS-2) in an inexpensive single-step reaction,
under mild conditions, to yield unique materials. Removal of
the core of porous manganese oxide by acid treatment gave rise
to novel hollow NSFS of titania. The versatility of this method
is unprecedented and has great potential for growing titania as
well as other metal oxides on various porous manganese oxide
(and other) materials that have very different properties such
as crystallographic structural parameters, nanosizes, and
shapes. The method can likely be extended to a variety of
other metal oxide cores and other metal oxide coatings. This
general synthetic method which involved heteroepitaxial
growth on tunnel-structure substrates led to the production of
shape-controlled hollow structures of nanosize particles, super-
structures, and thin films that had antenna-like particles ema-
nating from the surface as small as 20 nm. Extensions of this
method could involve all the known tunnel structure materials
such as 3  3 to 1  1, 3  3 to 2  2, 2  2 to 1  infinite layered
structures, or 3  2 to 2  1, 4  3 to 3  2, 5  4 to 4  3, via
heteroepitaxy by using different metal oxide coatings such as
materials that have the rutile structure such as CrO2, IrO2, and
GeO2. Other possible substrates like the 2  2 structure of aka-
ganeite [FeO (OH)], porous ankangite Ba(Ti,V)8O16, and others
could also be investigated. Control of the growth of nanosize
structures by heteroepitaxial growth of transition-metal porous
oxides is an area of potential significant growth for the produc-
tion of nanosize superstructures.
5.07.5.2 Mesoporous Materials
Mesoporous materials with controlled pore sizes and pore-size
distributions can be obtained for some systems such as silicon
and titanium-based oxide materials.47 However, control of
pore-size distributions to make single size pores and to system-
atically control such pore sizes are difficult, especially with
transition-metal oxide systems.48 Control of the structure of
the material is also an issue. Many systems have both micro-
pores and mesopores and pore interconnectivity is of interest
with these materials.49 Enhanced mass transport for catalytic
reactions might be realized by fine-tuning the porosity of such
systems.50 Incorporation of biomolecules larger than the
Table 1 Thermodynamic parameters of surfactant (S) transition-metal (M) mesopore syntheses
DGm DGorg DGI
SS interaction determines High Lewis acidity unsaturated
mesostructure formed coordination H (hydrolysis ratio H  1),
(lamellar, hexagonal, cubic) condensation hindering molecules
(carboxyl, amine, ethylene glycol)
Porous Metals and Metal Oxides 159
micropore regime also might be done using well-ordered crys-
talline mesoporous materials.51
Most studies of MTMO materials have focused on groups I
IV including Y, Ti, Hf, Zr, V, Nb, Ta, Cr, Mo, and W.52 These
have low-angle x-ray diffraction peaks and type IV isotherms.
These syntheses have focused on use of water or water plus a
base or urea with various amine- and carboxyl-containing
surfactants (S). There are either strong Coulombic interactions
(S,I; SI; SXI; SXI) or strong ligandmetal interac-
tions (I:S < 2, very thin walls), and such systems have limited
thermal stability and amorphous walls, where I inorganic
species and X is a mediator.53 Exceptions of systems that are
relatively stable and can be controlled are Ti and nonmetallic Si
systems.47 Many of the transition-metal syntheses primarily
work with F127 surfactant.54 Such syntheses are open to air
and various aging times and environmental conditions can
influence the porosity of these materials.
Water content is a critical parameter with porous transition-
metal oxides.55 Water competes with ethoxy and other alkoxy
groups for coordination to the metal. When excess water is
used, total removal of water is very difficult. Since most syn-
theses are open to the air the water content is very difficult to
control. On the other hand, water is essential for reaction.
When the number of water molecules per metal atom
(H) is >1 then phase separation and nonporous oxides
result.56 When H is <1, ordered mesoporous materials are
formed when the metal has empty t2g orbitals.56 These mate-
rials obtain water from the environment during synthesis.
When H is 1, strong surfactant/transition-metal interactions
occur with weak surfactantsurfactant interactions and there is
no reaction.56 Thermodynamic interactions in such syntheses
and factors influencing each term are given in Table 1:
DGm DGorg DGI DGinter DGsol [1]
In eqn [1], DGm is the formation energy of the mesostruc-
tured material, DGorg is the surfactantsurfactant interaction, DGI
is the metalmetal interaction, DGinter is the surfactantmetal
interaction, and DGsol is the solvent interaction.57 The objective
of a new synthetic strategy we have used (vide infra) concerns
minimization of the last two terms, DGinter and DGsol in order to
make well-ordered MTMO materials. Filled t2g orbitals such as
in systems containing Mn, Fe, Co, and others are difficult to
make with the above methods since charge-transfer reactions
do not occur.57
An extension41 of the nontemplate approach described
above in the formation of porous metals30 has been used to
make porous metals and coatings. ZnMn2O4 having the haus-
mannite structure was heated in hydrogen to form the MnO
rock salt structure. The resultant material contains substantial
Zn2; however, heating at high temperature reduces the zinc
ions that are removed in the gaseous state. Mesoporous MnO
crystals that have 3060 nm pores are produced. This is a
DGinter DGsol
Strong SM interaction at interface Unknown and
(Coulombic, covalent bonding, unpredictable
hydrogen bonding)
160 Porous Metals and Metal Oxides
topochemical synthetic approach that results in unique mate-
rials. This is an excellent balanced example of novel synthesis
with proper characterization including XRD, electron diffrac-
tion, chemical analyses, pore-size distributions, and TEM to
unravel these complicated systems.
By far the most extensive study of porous materials in the past
10 years has been in the area of mesoporous materials. It is
impossible to cover this whole area even for the past 5 years,
but some highlights are given here. Antonellis group has pio-
neered research in the synthesis of mesoporous materials in the
early transition-metal series in particular Ti, Nb, Ta, and others.58
For example, porous transition-metal oxides of Ti, Nb, and Ta
have been synthesized that have pore sizes of 1230 A by using a
ligand-assisted amine templating method. Several characteriza-
tion methods were used to fully characterize these materials
including adsorption, desorption, XRD, FTIR spectroscopy,
high-resolution TEM, SEM, TGA, differential scanning calorime-
try, and elemental analysis. These materials were used in catalytic
experiments such as the isomerization of 1-hexene and results
were compared to commercially available zeolites such as HY-
zeolite and ion-exchange resins such as Amberlyst 15. Mesopor-
ous tantalum oxide materials with a Hammett acidity function of
Ho  8.2 gave the best catalytic results. These materials were
shown to have 19.8 mmol g1 acid sites via acidity measure-
ments. Conversions of 1-hexene to trans and cis 2-hexene isomers
up to 96% were achieved in 4 h with ratios of trans to cis isomers
of up to 3.7 being observed after 6 h. These catalytic data were
believed to be very close to the theoretical ratio of 3.37 as
determined by thermodynamic equilibrium. The catalytic data
were influenced by the specific size of the templates used in the
synthesis. In addition, Ti and Nb materials were not active in this
reaction. Further acidity measurements were done by using TPD
studies of ammonia in order to explain the much higher activity
of the tantalum-based materials. These TPD results showed that
the Ta materials had a much greater concentration of Brnsted
sites in the active Ho acidity range for this isomerization reaction
than the Ti- or Nb-based materials. Other factors are important
in the high catalytic activity of these materials besides acidity
such as high BET surface area, increased concentration of effec-
tive Brnsted acid sites, and the necessity of optimal pore sizes
for this isomerization reaction. These data clearly show that the
control of acid sites and the fine-tuning of the porosity of these
mesoporous materials create a confinement effect for the effi-
cient heterogeneous catalytic 1-hexene isomerization using these
tantalum oxide-based mesoporous materials. These tantalum
oxide catalysts showed higher activities and selectivities than
sulfated mesoporous Ti, Amberlyst 15, HY zeolite, or H-ZSM-5
zeolite. For all of the systems studied with different pore sizes
of mesoporous Nb and Ta oxides, the C12H2SO4 templated
mesoporous Ta oxide systems (dodecylamine was used as a
template to form the mesoporous material which was then
sulfated with sulfuric acid) showed both the highest activity
and selectivity due to its high BET surface area of 292 m2 g1,
an optimal pore size of 18.2 A, and increased concentration of
active Brnsted acid sites believed to be located on the surface of
the mesoporous channels. The acid sites in these materials are
due to the sulfate groups and are therefore not oxide based. The
isomerization of 1-hexene is catalyzed over weak acid sites, so
these materials are much weaker acids than materials like zeo-
lites. This is a classic example of fine control of synthesis
conditions to make unique materials that show excellent cata-
lytic activity via clever design and full characterization of systems
before and after catalytic reactions.
A related study reviews the catalytic applications of MTMO
materials such as Ti-, Nb-, and Ta-based systems in a variety of
reactions.59 Another approach to enhance catalytic activity of
mesoporous materials is to dope small amounts of transition
metals into such materials. Considerable research has been
done in silica-based MCM-41 materials along these lines but
far less involving transition-metal oxide mesoporous systems.
In these Ti-, Nb-, and Ta-based oxide materials catalytic activity
was readily modified by doping with other elements, or by
using reducing conditions such as reduction with hydrogen,
or by treatment with various acid solutions in order to extend
applications of these materials in catalysis. Modification of
existing MTMOs has been done as well as characterization of
these newly modified mesoporous structures, with an empha-
sis on studying their respective surface properties. Some of the
types of catalytic reactions that have been studied include
activation of dinitrogen, benzylation, alkylation, and isomeri-
zation reactions. A major role of confinement effects and con-
trol of porosity are very important in such reactions since
catalytic reactions are well known to be influenced by diffu-
sion, mass transport, and the sizes of pores. By fine-tuning the
chemical and physical properties of such systems, the catalytic
properties can be optimized. Key factors include the composi-
tion, pore-size distributions, surface areas, and other proper-
ties. The importance of narrow and uniform pore sizes of
MTMO materials of the type Ti, Nb, or Ta and the ability to
control oxidation states of such systems that can have multiple
valences are all critical. Novel catalysts with high surface acidity
and variable oxidation states have been prepared in this work
in order to enhance catalytic activity and selectivity. A key focus
of Antonellis group has been the pursuit of catalysts to activate
dinitrogen to produce ammonia, a very important industrial
reaction.59 This reaction is difficult to catalyze under mild
conditions and few active catalysts are known to activate dini-
trogen. The acidity can be tailored in these mesoporous sys-
tems to control the types of acid groups, the number of acid
groups, and their strengths. Since the pores of such mesopor-
ous materials are wide with respect to reactants, mass transfer
limitations are minimized. Bulky products are also not inhib-
ited from leaving the pores after reaction, which can happen
with smaller-pore molecular sieves. A common theme in much
of todays research in catalysis and synthesis of molecular sieve
materials is control of synthetic intermediates and catalytic
intermediates by confinement effects. These mesoporous mate-
rials have also been proposed to be used in several other areas
besides catalysis including gas sensors, purification of air, dye-
sensitized photovoltaic solar cells, and lithium ion batteries.
Variable oxidations states and controlled acid sites are often
common requirements for many of these applications. Some
key shortcomings of such systems include long-term thermal,
chemical, and structural instabilities. These and most other
mesoporous systems suffer from these drawbacks and have
prevented the use of these types of materials in commercial
endeavors. In addition, there are other factors that need special
consideration such as the cost of reagents and the final product
as well as the environmental friendliness of synthesis condi-
tions and in the final use of such materials.

Coatings on the surfaces of transition-metal oxide mesopor-
ous materials can lead to ordered MTMOs. Silica layers can be
used for strengthening these materials that typically have
amorphous walls.60 There are advantages of such coatings as
regards control of crystallization and control of mass transfer.
Other methods in contrast to surface silica (silicone) coating
include carbon back filling to help stabilize mesoporous mate-
rials. By careful control of conditions, the crystallinity of the
resultant crystalline MTMO product can be influenced.
The majority of recent articles on mesoporous materials
have involved synthesis of first-row transition-metal systems
with a predominance of articles on titanium-based systems.
This may be due to the initial studies with silica-based systems
and similarities of group IV elements. Certainly for silicon- and
titanium-based mesoporous materials, stability and porosity
are more controllable than most transition-metal oxide meso-
porous systems.
The patterning of nanosize metal oxides is of great impor-
tance for numerous applications such as catalysis, semiconduc-
tors, and biomaterials. The formation of ordered nanosize
dimple arrays on titanium, tungsten, and zirconium oxides
during electropolishing has been extended in recent research
in contrast to previous systems that have focused on the use of
tantalum oxides.61 Electrochemical patterning is not common
and this reported process may allow extensions to numerous
other transition-metal oxide surfaces. These dimpled surfaces
of titanium, tungsten, and zirconium oxides have been charac-
terized with SEM, TEM, atomic force microscopy, XPS, and
electrochemical methods. Electrochemical conditions were
optimized for each of these materials. Electrochemical condi-
tions for dimple growth on titanium and tungsten oxides are
similar to those of tantalum oxide. However, formation on
zirconium oxide requires a different type of electrolyte. A key
factor in such syntheses is electropolishing and conditions for
electropolishing need to be optimized. The authors believe that
these methods can be used on any type of transition-metal
oxide surface. This electropolishing and deposition procedure
process is very robust when homogeneous surfaces are used.
However, on surfaces that are inhomogeneous such as those
that have spatial differences in chemical composition, the
resultant materials are not so stable. Examples of inhomoge-
neous surfaces are those that result from acid or base treatment
after chemical etching. One way around problems of instability
involves the preparation of a thin film of the metal on the
dimpled surfaces, such as deposition of platinum. Dimple
formation with stable products was shown to occur for tanta-
lum, titanium, tungsten, and zirconium oxide surfaces. Further
applications may be possible such as the use of such systems in
biomedical applications if materials like titanium alloys can be
used as substrates. It may be possible to specifically coat certain
parts of a surface to make nanosize patterns if one can control
the surface chemistry and selectively coat specific chemical
species in localized areas on the surface. A specific example
for specific coatings of localized areas on surfaces would be
nickel impurities on tungsten surfaces. In this case, the nickel
oxide surface species do not readily dissolve. Dissolution is
necessary for the optimal conditions for dimple formation on
tungsten. When such nickel oxide impurities in tungsten sur-
faces were studied with this method, a clear perturbation of the
dimple pattern was seen around the nickel oxide islands in the
Porous Metals and Metal Oxides 161
tungsten substrate. Such systems with heterogeneous surface
compositions are ideal for the selective pattern formation in
specific areas on a surface. This process is self-limiting and
results in uniformly shallow features. These deformities in
turn can also be used as templates that can be used for specific
pattern formation. In certain cases like surfaces that can be
used for biocompatibility, high aspect ratios are needed
where 50-nm lengths scales are needed. Other applications
could be preparations of surfaces that can be used as templates
for the formation of active catalyst particles. This method
appears to be very general and most surfaces could be used for
preparing specific surfaces that have unique porosity. Such sur-
faces do need to be conducting, so there are limitations to using
metal of semiconducting materials. Potential applications are
for sensor array production, combinatorial syntheses, templated
growth of nanostructures, growth of novel electronic materials,
composite materials, new optoelectronics, catalytic systems,
photovoltaics, biomaterials, and other materials.
We have seen in many of the above reports that MTMOs are
of great interest and potential as catalyst supports, optical
materials, shape-selective catalysts, supports for slow release
of drugs, photocatalysts and photocatalytic supports, sensors,
and other materials. Stable crystalline mesoporous oxides are
rare and are often obtained by thermally induced crystalliza-
tion or by segregating the nanosize crystals with an amorphous
phase.62 A direct novel method was recently developed to
produce crystalline mesoporous frameworks by using the
spontaneous growth and assembly of transition-metal oxide
nanocrystals. Materials such as the rutile structure TiO2, fluo-
rite structure CeO2, cassiterite SnO2, and the anatase structure
SnxTi1xO2 were made by oxidative hydrolysis and condensa-
tion in the presence of anionic surfactants. Parameters such as
synthesis time, surfactants that have different chain lengths,
concentrations of oxidants such as hydrogen peroxide, and
synthesis temperatures are optimized to make desirable mate-
rials. The compositions and morphologies of these mesopor-
ous materials were studied using x-ray diffraction, adsorption
of nitrogen, TEM, selected area electron diffraction, SEM, and
XPS methods. The surfactants act like templates in these syn-
theses. The rutile structure TiO2 and fluorite structure CeO2
mesoporous supports were decorated with gold nanosize par-
ticles and their catalytic activity for CO oxidation and the
watergas shift (WGS) reaction were investigated. The Au
nanosize particles on the crystalline mesoporous transition-
metal titanium and cerium oxides have considerably higher
activity and stability than similar nanosize Au particles sup-
ported on commercial titania and ceria supports in both CO
oxidation and WGS reactions.
Many syntheses these days involve the use of microwave
radiation to heat precursors in the formation of mesoporous
materials. Microwave radiation provides a unique route for
heating materials, typically from the inside out rather than
outside in for thermal heating. Various microwave apparatus
are available such as modified microwave ovens. Both contin-
uous and pulsed microwave syntheses are currently being done
in many laboratories. Combinatorial microwave studies can be
done with some systems that have robotic arms that can intro-
duce a variety of samples into microwave chambers. Either
several similar materials can be introduced and treated with
different microwave conditions such as power and time of
162 Porous Metals and Metal Oxides
radiation or different samples can be treated under similar
conditions in order to optimize the syntheses of final products.
Microwave apparatus are now economical and apparatus for
scale-up is available. The microwave-assisted combustion syn-
thesis of porous nanocrystalline titania and carbon-coated tita-
nia materials has been done with dextrose used as a template.63
As in many microwave studies, comparable syntheses are done
with similar heating rates using thermal methods and proper-
ties of the thermal and microwave methods are then com-
pared. The metal to dextrose ratio is a key parameter in these
microwave syntheses of mesoporous titania and carbon-coated
titania systems. These samples were characterized using numer-
ous methods such as SEM and TEM, EDX analyses to determine
compositions, XRD, and x-ray mapping analyses. This is a
general and environmentally friendly synthetic method. A nat-
ural polymer, dextrose, was used to create these mesoporous
materials. Extensions to other metal oxide systems such as
ZrO2, Al2O3, and SiO2 were done.
Chromium oxide-based mesoporous materials seem to be
limited with few studies reported.64 Highly ordered mesopor-
ous Cr2O3 materials have been reported that have high surface
areas and narrow pore-size distributions. A vacuum-assisted
impregnation method was used to prepare such materials.
Both SBA-15 and KIT-6, which are mesoporous silica materials,
were used as templates to prepare 2D hexagonal and 3D cubic
Cr2O3 mesoporous materials by a replica method. These sys-
tems have enhanced properties as gas sensors and as lithium
ion batteries in comparison to Cr2O3 bulk materials. Very often
a last resort for making such systems appears to be replica
methods.64 Another example involves porous single crystals
of Co3O4 and Cr2O3.65 SBA-15 and KIT-6 were also used as
templates to make these mesoporous cobalt oxide and chro-
mium oxide systems.65 Specimens were characterized by using
x-ray diffraction and high-resolution TEM methods at various
stages of the crystal growth to optimize the syntheses of these
materials. Porosity of these materials was studied using adsorp-
tion methods using nitrogen as a probe. The synthetic method
appears to involve a mechanism initiated by the formation of
multiple seeds. The pore diameters of the silica templates
SBA-15 and KIT-6 are very important in this synthetic method.
Negative replicas were used and the nanorod and nanowire
building blocks control the resultant structure of the mesoporous
metal oxides. The mesopore system, regarded as a nanoreactor,
had a significant influence on the sizes and morphologies of the
mesoporous metal oxides, which were highly dependent on the
silica templates and the conditions of the syntheses.
Iron oxide and sulfide mesoporous materials are more
common. The syntheses of these mesoporous materials involve
the combination of polyamine with magnetic semiconductors.
Solvothermal synthetic methods are now also more commonly
used to make porous metal oxides. A mixed solvent system was
used here composed of triethylenetetramine (TETA) and deio-
nized water. Ordered nanoribbons of [Fe18S25](TETAH)14 were
produced in this solvothermal preparation. The volume ratio of
TETA and water needs to be optimized in order to produce the
[Fe18S25](TETAH)14 nanoribbons.66 These [Fe18S25](TETAH)14
nanoribbons can be thermally decomposed in argon or air and
used as precursors for the production of Fe7S8 nanowires and
porous a-Fe2O3 nanorods. Novel hybrid nanostructured mate-
rials are produced in this way. The novel hybrid materials can be
thermally decomposed to make these transition-metal chacol-
genide nanomaterials and porous transition-metal oxides. This
method is an excellent example of a thermal decomposition
method that can be used to prepare novel nanoribbon mate-
rials. Since controlled synthesis of transition-metal chalcogenide
nanomaterials and porous transition-metal oxides can be done
with this method, extensions to other materials such as selenides
or phosphides might be possible.
Mixed metal oxide mesoporous systems are known. A recent
study reports that ordered mesoporous NiFe2O4 thin films hav-
ing the spinel structure can be synthesized via co-assembly of
hydrated ferric nitrate and nickel chloride with an amphiphilic
diblock copolymer.67 Highly crystalline NiFe2O4 samples are
produced after calcination at 600  C. A cubic network is formed
that has pores with diameters of about 16 nm. This synthetic
method can be used to produce other ferrites such as CoFe2O4,
CuFe2O4, MgFe2O4, and ZnFe2O4. Self-assembly processes are
important in the synthesis of these mesoporous spinel struc-
tures. SEM and high-resolution electron microscopy, grazing
incidence small-angle x-ray scattering (GISAXS), time-of-flight
secondary ion mass spectrometry (TOF-SIMS), XPS, UVvis, and
Raman spectroscopy methods were used to study these spinel
systems. Electrochemical methods were used to study the redox
properties of these systems. The nanocrystalline films have excel-
lent porosity. The mesoporous NiFe2O4 electrodes show good
storage of lithium ions at short charging times due to ready
pseudocapacitance, which is charge transfer at the surface
atoms, which are present in substantial quantities. Optimizing
these materials as lithium ion battery materials and capacitor
applications appears to be a realistic goal. Other examples of
mixed metal oxide spinel-like systems using other synthesis
methods can yield ZnFe2O4 systems. These mesoporous nano-
crystalline ZnFe2O4 materials have a timber-like superstructure.
The synthesis involved the thermal decomposition of zinc fer-
rioxalate precursors that were made by mixing iron sulfates and
sodium oxalate without any other additives. The resulting super-
structures of ZnFe2O4 superstructures also showed good thermal
stability as determined by TGAs and differential scanning calo-
rimetry methods. In addition, the magnetic properties of this
material were unique, leading to superparamagnetic particles at
room temperature. Such porous nanocrystalline ZnFe2O4 super-
structures may be useful in magnetic devices. This method
was extended to other mixed metal spinel systems of various
compositions such as NiFe2O4 and CoFe2O4 that also have
timber-like morphologies and could likely be extended to
other materials such as MnFe2O4 and ZnFe2O4.
Several articles on cobalt oxide mesoporous materials
are available. A microwave-assisted reflux route was used to
synthesize layered nanostructures of Co hydrotalcite-like
(layered double hydroxide, LDH) materials that have spherical
shapes.68 Morphologies determined by SEM methods show
images that are self-assemblies of nanosheets with uniform 5-
nm thickness. The LDH Co hydrotalcite-like materials have
high surface areas and large pore volumes of 176 m2 g1 and
1.13 m3 g1, respectively. These cobalt hydrotalcite-like mate-
rials can be transformed into the spinel structure material
Co3O4, which retains a uniform spherical shape. The surface
areas and pore volumes of the Co3O4 are on the order of
61 m2 g1 and 0.5 cm3 g1, respectively. Both the LDH and
Co3O4 materials show electrocatalytic activity and stability

for the electrochemical reduction of oxygen, which is impor-
tant in batteries, fuel cells, and electrocatalytic processes. The
high surface areas of these catalysts are critical for generating
high rates of electrochemical reduction of oxygen. These mate-
rials have potential not only as electrocatalytic materials but
also as substitute catalysts for noble metal catalyst supports.
This method is an efficient synthetic route that may also be
used to produce other MTMOs such as manganese oxide,
nickel oxide, and others.
ZnO mesoporous systems have also been prepared. ZnO and
doped ZnO mesoporous self-aggregated nanoclusters of both
pure and transition-metal (manganese, nickel, and cobalt)-
doped ZnO have been synthesized by refluxing acetate precur-
sors in diethylene glycol (DEG) media.69 The porous spherical
nanoclusters consist of numerous stable nanocrystals. These
systems are well defined and have the hexagonal wurtzite ZnS
structure. Dopants such as manganese, nickel, and cobalt ions
were successfully doped into the ZnO structure and these
transition-metal ions are uniformly distributed throughout the
samples. There is a 3D intersection network of pores that is
mostly mesoporous as determined by nitrogen adsorption and
desorption experiments. The mesopores of these ZnO systems
are irregular in shape and are randomly distributed throughout
the nanoclusters. These materials are active photocatalysts for
the degradation of organic molecules under radiation with UV
light. Degradation of the dye species methylene blue was
achieved with the ZnO mesoporous materials.
Several second-row transition-metal oxide materials have
been discussed throughout this contribution. Most of these
materials are in the early part of the periodic table.
Zirconium-based materials are often studied since they are
similar to titanium-based systems which have received consid-
erable attention. Zirconium oxide mesostructured hexagonally
ordered surfactant composites have been made using zirco-
nium sulfate ions as inorganic precursors.70 On the basis of
the mesostructured zirconium sulfate surfactant composites,
two different mesoporous analogs of MCM-41 have been syn-
thesized. Zirconium oxide-sulfate and zirconium oxo phos-
phate have been synthesized with a postsynthetic treatment
being used for preparation of the oxo phosphate materials.
Numerous methods were used to characterize these samples
including x-ray diffraction, nitrogen adsorption, extended x-ray
absorption fine structure, and TEM methods. These zirconia
mesoporous systems have hexagonal mesostructures with uni-
formly sized pores and amorphous pore walls. The thermal
stabilities of these two different materials are excellent, up to
500  C. There is complete removal of surfactant during this
thermal treatment. This exceptional thermal stability as com-
pared to other porous transition-metal oxides prepared by
similar surfactant controlled synthesis is unique. Extensions
to mixed zirconium cerium oxide materials have been done
recently.71 Mesoporous phases of ceria (CeO2) and ceriazirconia
solid solutions of compositions up to 50% of Zr have been
synthesized using a hard template or a replica method. The
structures of the metal oxide replicas correspond to the silica
precursors KIT-6 or SBA-15. When such replicas are used, uni-
form and homogeneous pore-size distributions are produced.
Such open porous structures allow efficient interactions with
gaseous species via enhanced diffusivity and control of pore
size. The final products were analyzed with several methods.
Porous Metals and Metal Oxides 163
The porous structure was confirmed by BET and TEM experi-
ments. Materials prepared using KIT-6 templates have some-
what higher active surface areas (125 m2 g1) by about 5% than
the SBA-15 templates. Raman and XRD data suggest the forma-
tion of solid solutions having cubic structures unless the
amount of zirconium increases to >20% where tetragonal
structures are formed. The mixed ceriazirconia solid solutions
have excellent oxygen storage capacity. Even when only 5% Zr is
added to the ceria, there is a 30-fold enhancement in oxygen
storage capacity. This may be due to substituted zirconium in
the lattice of ceria. There may be other explanations why the
oxygen storage capacity is very high for low amounts of zirco-
nia, such as enhanced mobility of the oxygen ions or enhanced
oxygen vacancies in the structure.
Further replica methods have been used to make a variety of
mesoporous metal oxide systems. Crystalline mesoporous
metal oxides have been known for quite some time, but their
catalytic properties have rarely been explored. A series of crys-
talline 3D mesoporous metal oxides such as CeO2, Co3O4,
Cr2O3, CuO, Fe2O3, b-MnO2, Mn2O3, Mn3O4, NiO, and
NiCoMnO4 has been synthesized using mesoporous KIT-6 of
silica as a hard template.72 Such ordered mesoporous metal
oxides have highly crystalline walls in contrast to the many
amorphous-walled MCM-41 materials. These systems were
characterized with several methods such as XRD, TEM, and
N2 adsorption studies. These systems were also studied in
catalytic oxidation of CO. Most of these mesoporous materials
showed considerably higher catalytic activity than their bulk
counterparts. Mesoporous NiO, Co3O4, and b-MnO2 showed
excellent catalytic oxidation of CO at 0  C. The b-MnO2 system
is of interest due to the low cost, low toxicity, and high activity.
Using silica-based mesoporous materials as replicas is also
well known with a variety of metal oxide systems being possi-
ble. Many of the mesoporous silicas such as SBA-15, KIT-6,
FDU-12, and SBA-16 have been used in replica syntheses of
porous crystalline transition-metal oxides. Many are men-
tioned above but there are others, including Cr2O3, In2O3,
and WO3, which have been synthesized using the mesoporous
silicas as hard templates.73a Several synthetic methods have
been developed to produce such materials. Potential applica-
tions include catalysis, lithium ion rechargeable batteries,
adsorption, and gas sensors. This chapter reviews different
porous crystalline metal oxides. Recent studies have shown
that these types of materials can be used in many applications
such as photo-assisted catalysis. Ordered mesoporous rutile
TiO2 and anatase TiO2 samples have been prepared with meso-
porous silica SBA-15 as a hard template.73b Titanium nitrate
and titanium chloride solutions were used as precursors. The
rutile material prepared from nitrate solutions is crystalline
and has small amounts of Si contaminants (Si/Ti ratio of
0.031). The mesoporous TiO2 anatase material prepared from
chloride solutions is composed of nanosize crystals and has a
Si/Ti ratio of 0.18. The crystallinity of these materials and the
amount of silica contaminant depend on synthetic conditions
in particular, the type of titanium precursor as well as the
temperature of the synthesis. These mesoporous crystalline
TiO2 systems were characterized with a variety of methods
such as XRD, high-resolution TEM, and adsorption and
desorption with nitrogen. When SBA-15 was used as a hard
template, mesoporous TiO2 having the rutile structure was
164 Porous Metals and Metal Oxides
formed. A higher lithium ion insertion capability was observed
for the SBA-15 template than with mesoporous TiO2 templated
with KIT-6-perhaps due to a larger surface area for the SBA-15
templated material. The mesoporous anatase TiO2/SiO2 com-
posite has a higher photoactivity than bulk TiO2 or TiO2
loaded SBA-15 in the bleaching methylene blue.
5.07.5.3 Macroporous Materials
Macroporous materials have received limited attention in the
open literature. Organicinorganic hybrid microstructures
have been made by the use of chitosan biopolymers to coor-
dinate vanadium, tungsten, and molybdenum metallic
species.74 The control of mineralization was important in this
system. These materials were dried using supercritical methods
and allowed dispersion of the gel to remain. This led to forma-
tion of macroporous materials that have surface areas from 253
to 278 m2 g1. These materials were used for catalytic oxidation
of cinamylalcohol in air at low temperature. High activity was
believed to be due to the redox species being entangled within
the fibrillar network of the polysaccharide aerogels. The super-
critical drying process led to formation of highly porous mate-
rials with tunable chemical properties.
A review of general methods for the preparation of mesopor-
ous and macroporous structured single and mixed metal oxides
has been summarized.75 Hydrothermal methods are one of the
major ways to make such mesoporous systems that have high
surface areas and high thermal stability. A self-rising approach
is discussed that affords macroporous transition-metal oxides
that have long-range order, low filled density, and high yield
of product.
Copper and cobalt-doped carbon aerogels can be obtained
from the polycondensation of resorcinol and formaldehyde.76
These materials were studied for NO reduction. Interactions of
NO and carbon were studied with temperature-programmed
surface reaction (TPSR) conditions and by in situ FTIR adsorp-
tion studies. Metal-doped carbons showed higher activity for
NO conversion on the order of 80% at 450  C than free metal
carbon systems. The detected products are N2, CO, and CO2.
The conversion into N2 is much lower than total NO conver-
sion, suggesting that some nitrogen compounds are still
located on the carbon surface. The metal species present in
carbon aerogels are active for the dissociation of NO and for
the transfer of oxygen from the oxidized sites on the mesopor-
ous catalysts to the carbon matrix. The reaction mechanism
depends on reaction parameters such as the temperature and
contact times. Surface complexes of type C(N), C(NO), and C
(O2) are observed by FTIR methods. A high NO conversion rate
requires metal cations that catalyze the decomposition of NO
into N2, CO, and CO2 as final products. Micropores are initial
reaction sites that lead to the opening and creation of other
pores. The ability to undergo redox cycling promotes the best
catalytic performance of copper-doped carbon aerogels. Car-
bon aerogels doped with transition metals are very active for
NO reduction in the absence of oxygen. The high activity and
selectivity of the Cu-doped catalyst for reduction of NO into N2
depend on the capacity for undergoing progressive cycles of
NO oxidation and reduction with carbon. The carbon matrix
also has some limited catalytic activity. The reaction mecha-
nism was proposed to be dependent on temperature and
involves formation of surface species such as C(N), C(NO),
and C(O2). NO reduction and carbon matrix gasification occur
simultaneously.
Electrochromic materials consisting of porous transition-
metal oxide thin films have opened up new possibilities in the
construction of working devices.77 Tungsten oxides are some of
the most promising materials that are used in electrochromic
systems. The synthesis of mixed mesoporous and macroporous
WO3 thin films has been done by using tungstic acid as a
precursor and organically modified silane as a templating
agent. Nanosize crystalline monoclinic tungsten oxide having
crystallite sizes of about 16 nm was made by calcining films at
500  C. After this calcination, the combined mesoporous/
macroporous structure is retained. An electrochromic device
was made of the mesoporous/macroporous WO3 thin film
that shows an asymmetric response. The electrochromic device
was characterized with micro-Raman spectroscopy, SEM, TEM,
and XRD. Micro-Raman and XRD experiments were used to
study intercalation and de-intercalation of lithium as a function
of the applied voltage. The mesoporous/macroporous WO3 film
is monoclinic before lithium intercalation. After intercalation
the structure changes to higher symmetry and a tetragonal
phase. After complete de-intercalation these phase transitions
were reversed and the monoclinic WO3 is reformed. Optical
studies show that the mesoporous/macroporous monoclinic
WO3 thin film can be used in electrochromic devices that
show a high color contrast during coloration and bleaching.
Macroporous transition-metal oxides having various compo-
sitions such as CuO, NiO, CoO, Mn2O3, Cr2O3, and ZnO were
synthesized and used as catalysts for several different organic
transformations.78 These reactions include N-formylation, N-
acylation, O-acylation, and FriedelCraft acylation under
solvent-free conditions in good yields. Five macroporous
transition-metal oxides and zinc oxide were used as efficient
environmentally friendly heterogeneous catalysts for these
organic reactions. The reaction conditions involve a simple,
solvent-free protocol. The process is easy to operate, has a low
cost for the catalyst, uses mild, clean reaction conditions, has high
yields, and is a green method. Due to the presence of large
internal surface area and pore volume, porous metal oxides
allow control over the diffusion of reactants and products into
and out of the porous medium. The large macroporosity allows
formation of large catalytic intermediates. The type of metal oxide
and type of reaction being catalyzed dictate the amount of catalyst
that needs to be used. In the four reactions of N-formylation, N-
acylation, O-acylation, and FriedelCraft acylation, a reduction in
the yield of product was observed when the amount of catalyst
was increased. Increasing the amount of catalyst beyond an opti-
mal value leads to aggregation of the clusters. That in turn leads to
a reduction in the effective internal surface area and pore volume
of the catalyst. This of course leads to a large decrease in catalytic
activity. Many of the catalysts were active in most of the reactions.
Only macroporous CoO and Cr2O3 materials showed significant
activity for FriedelCraft acylation.
5.07.6 Applications of Porous Metal Oxides
Several applications have already been mentioned throughout
this chapter; besides the more common adsorption, catalytic,

and sensor applications, some more unique or recent applica-
tions are discussed in this section. A template-assisted solution-
phase synthesis was developed to prepare 2-line ferrihydrate
(2LFh) as a partially dehydroxylated material that has accessible
mesopores.79 The readily accessible and reactive surface-defect
sites lead to use of this material for elimination of acetaldehyde
(CH3CHO), which is a noxious gas found in both indoor and
outdoor atmospheres, at faster rates than over several iron
oxides. This research exploited the highly desirable physico-
chemical characteristics of ferrihydrate for air purification.
Both the mesoporous structure and the physico-chemical char-
acteristics of the ferrihydrate phase allowed high activity at room
temperature because surface water was minimized. This work
may provide insight for design of highly active cheap
ferrihydrate-based composite materials for rapid removal and
decomposition of different types of noxious gases present in the
atmosphere under ambient conditions.
Highly ordered mesoporous NiO systems have been made
by a nanocasting method that uses mesoporous silica KIT-6 as
the hard template.80 These mesoporous NiO particles were
characterized by several methods including small-angle x-ray
diffraction, nitrogen adsorption/desorption, and TEM experi-
ments. The mesoporous NiO has an ordered symmetric mesos-
tructure and a high surface area of 96 m2 g1. These mesoporous
NiO materials were fabricated as anodes for lithium ion batte-
ries. Faster lithium intercalation kinetics were observed for this
mesoporous material than bulk NiO. A specific capacity of
680 mAh g1 at 0.1  C for this mesoporous NiO material was
stable for at least after 50 cycles. Commercial bulk NiO had a
lower specific capacity of 188 mAh g1. The discharge capacity
of the mesoporous NiO was 515 mAh g1, which is significantly
high for applications such as high-power lithium ion batteries.
The synthesis of self-assembled surfactantmesophase tem-
plated films can be accomplished with a solgel route.81 The
formation of metal oxide films using this method has become
a highly active area of research. The process parameters need to
be carefully designed in order to successfully synthesize these
materials. Control of pore orientation can be done via confine-
ment or by surface modification. Pore alignment inside a 1D
template can be done within the 200-nm channel diameters of
an anodic alumina membrane. The self-assembled nanosize
particles need to be smaller than the diameter and are typically
310 nm. When the size of the confinement is similar to that of
the micelle, new nanostructure morphologies are observed.
Pluronic P123 surfactant in silica mesophases were self-
assembled inside anodic alumina membranes that have pore
diameters that are varied between <20 and 80 nm. Coiled
channels and discrete cage structures were found at small
membrane sizes. Stacked doughnut, single helices, or double
helices that are chiral can be prepared.
Nanoscale morphologies are important building blocks to
control during synthesis and by complete characterization after
synthesis. Many techniques have been used for patterning of
such materials such as ink jet printing, light, micromolding, dip-
pen lithography, and other methods. Thin-film inorganic metal
oxide mesoporous patterns are becoming more common these
days. UV radiation to remove the surfactant has been used;
ozone can be used to treat surfaces via chemical etching, electron
beam lithography, mechanical polishing, and other methods.
Masks are important for selective deposition and etching. Some
Porous Metals and Metal Oxides 165
of the systems that have been produced include mesoporous
silica and titania arrays that are on the order of 250 nm.
These confined mesostructures can be used as templates for
electrochemical deposition of metals and metal oxides. Prepa-
ration of self-assembled inorganic thin films with different
morphologies at various scales is a goal of researchers. New
materials such as hybrid inorganic/organic solar cells, novel
valve materials, and others have been produced using such
strategies. Metal wires, semiconducting organic polymers, and
a host of other systems can be prepared by using such ordered
materials and removing templates.
Continuous-flow syntheses are becoming more common.
1D transition-metal oxide nanostructures and their integration
into hierarchical structures is a difficult task. Nonthermal
continuous-flow methods were used for the preparation of
microspheres of g-MnO2 and CeO2.82 Microspheres has been
developed. Surface areas of these g-MnO2 materials were 240,
98, and 87 m2 g1 and for CeO2 microspheres were 1 m2 g1.
Characterization of the materials was done with XRD, IR,
elemental analyses, nitrogen adsorption, SEM, TEM, and
TGAs. The g-MnO2 systems were active in the catalytic oxida-
tion of a-methyl styrene. This study developed a new route for
preparation of crystalline transition-metal and lanthanum
oxide nanostructures under nonthermal and atmospheric con-
ditions. Extensions to other transition-metal oxide systems
such as ZnO and CoOOH may be possible.
Water oxidation to oxygen can provide the electrons needed
for the reduction of carbon dioxide by sunlight to liquid trans-
portation fuels by artificial photosynthetic processes. This is
important for solar photocatalysis and the generation of clean
fuels. The catalyst needs to have a low overpotential, excellent
reaction kinetics, and be inexpensive and simple. The high
photon flux at high solar intensity is also an issue to be dealt
with. The catalyst needs to be stable under the demanding
water oxidation conditions. Cobalt oxide and manganese
oxide materials have provided impetus and hope that such
systems might be achievable.83 Methods that involve nanos-
tructuring of the surface and incorporation into supporting
silica scaffolds have produced active and stable water oxidation
catalysts. Operation under mild conditions of temperature and
pH is also a requirement. These manganese and cobalt oxide
catalysts are oxygen-evolving materials that are cheap, operate
under mild conditions, have low overpotentials, and have
decent rates. Systems of most importance are nanosize Co3O4
and Mn oxide mixed-valent species. Mesoporous silica scaf-
folds allow these catalysts to have very high densities of surface
metal sites per projected area. The mesoporous silica gives
stability for dispersion of these small clusters. Turnover fre-
quencies per projected area that can keep up with the high
photon flux at high solar intensity are available with such
catalysts. Coreshell geometries may be important for such
devices. A key factor is that such catalysts need to catalyze
multielectron transfers in such reactions. Other competing
systems include cobalt phosphate solutions, mixed calcium
manganese oxide clusters, and polyoxometallate species.
Photo-assisted catalytic oxidation of water to oxygen is a
complicated process and difficult to drive. Visible light can be
used to drive O2 evolution with aqueous suspensions of Mn
oxide/KIT-6 catalysts.84 These materials mimic homogeneous
catalysts and biological systems. Mass spectrometric analysis
166 Porous Metals and Metal Oxides
was used to determine oxygen concentrations. Sensitizers such as
[Ru(bpy)3]Cl26H2O were used. Micron-sized Mn oxide particles
of MnO2, Mn2O3, or Mn3O4 were studied. Mildly acidic condi-
tions are needed since the sensitizer can decompose under very
acidic conditions. Long-term experiments with Mn oxide/KIT-6
catalysts show oxygen yields that are stable in time. x-Ray absorp-
tion near-edge spectroscopy experiments were done to show that
the oxidation of the manganese remained stable throughout the
catalytic oxidation. These experiments are done at room temper-
ature and at a pH of 5.8. The high surface area of the silica
support is important for dispersing the manganese oxide clusters.
Surface restructuring may be important during the process.
Fuel cell applications are another area where microporous,
mesoporous, and macroporous materials could be exploited.
Microbial fuel cells with multianodes/cathodes with 12 anodes
and 12 cathodes were developed to harvest electric power during
treatment of domestic wastewater.85 The power density of these
fuel cell systems increased from 300 to 380 mW m2 in the
range of the organic loading rates of 0.190.66 kg m3 day1.
The microbial fuel cells achieved 80% of contaminant removal
at the hydraulic retention time of 20 h. The contaminant
removal decreased to 66% at a hydraulic retention time of 5 h.
Metal-doped microporous and mesoporous manganese dioxide
cathodes were developed to replace the costly platinum cath-
odes and exhibited higher power density than Pt-based systems.
Cu-doped manganese oxide (cryptomelane structure) cathodes
produced 465 mW m2 and Co-doped manganese oxide cath-
odes produced 500 mW m2. Cathode fouling of the precipita-
tion of calcium and sodium led to a decrease in the power
density from 400 to 150 mW m2. There was an increase in
internal resistance from 175 to 225 O in these microbial fuel
cells. In addition, these microbial fuel cells were used to study
power generation and contaminant removal at the pilot-scale
level by treating real wastewater in a continuous mode. The
performance of the pilot-scale systems was evaluated in terms
of several parameters such as power density, removal of nutri-
ents, solids removal, and cathode fouling. The increase in the
electrode numbers in these microbial fuel cells led to higher
power generation, which is promising for integrating multiple
electrodes to enhance power generation in such fuel cell sys-
tems. Contaminant removal efficiencies varied with hydraulic
retention times. The microbial fuel cells had a removal of 30%
nitrogen due to the nitrogen uptake by biomass in these sys-
tems, but was not effective for removal of phosphorous. Finally,
cathode fouling was found to occur after several weeks of oper-
ation. This fouling occurred on the cathodes and was caused by
the deposition of calcium and sodium.
Biodiesel production from methanol and oil generally uses
homogeneous catalytic processes. Homogeneous catalysts are
difficult to separate from reactants and products. In addition,
in these transesterification reactions soaps can form which can
impede the process. Therefore, the search for solid-state mate-
rials with high catalytic activities and with no leaching into the
reaction medium is a key step in the reduction of cost in the
production of biodiesel fuel. A high biodiesel yield (>95%) in
less than 5 min under mild reaction conditions (<100  C) on a
ZnO/La2O2CO3 heterogeneous catalyst, showing no catalyst
leaching into the reaction medium, has been reported.86
ZnO/La2O2CO3 catalysts were prepared by a co-precipitation
method and characterized by several techniques including
XRD, TGAs, and TEM. The desired fatty acid methyl ester
(FAME) product has yields that differ with different amounts
of catalyst that are used. Catalyst loadings of 1 wt% catalyst can
be used in the reaction to obtain 95% FAME yield under mild
reaction conditions. The catalysts are active after storage of the
catalyst in argon for a month. No catalyst leaching into the
products was observed based on elemental analyses. The Zn/La
oxide catalyst has a higher reaction rate than the homogeneous
KOH catalysts when activated with microwave irradiation.
These results suggest that industrial applications of the synthe-
sized ZnO/La2O2CO3 might be possible for rapid biodiesel
production. Microwave treatment allows very short reaction
times for biodiesel production.
Binary copper microporous and mesoporous manganese
oxides were prepared with novel redox methods and their
catalytic activity for CO oxidation at ambient temperature
was evaluated.87 The catalytic activity of these mixed pore
materials was found to be high. The catalytic data were com-
pared to similar catalytic with commercial Hopcalite catalysts
and the activities are excellent for the binary copper-doped
manganese oxide catalysts. The most active catalysts were able
to completely oxidize CO to CO2 at ambient temperature.
A decrease in catalytic activity in time is most likely due to
carbon dioxide retention. The catalysts were also deactivated
by moisture but expelling water at moderate temperatures via
heating easily restored their catalytic activity. The catalysts were
characterized by several methods such as BET surface area and
pore-size distribution experiments, field-emission SEM, TEM,
EDX analyses, XPS, TPD, and XRD. The optimum copper load-
ing was shown to be about 9% of the manganese content.
These copper manganese oxides were successfully synthesized
using a novel redox method and their excellent catalytic activity
for CO oxidation at ambient temperature was evaluated. The
redox method, which involves titration of various oxidation
states and measurement of the resultant oxidation states by
potentiometric titrations and x-ray absorption spectroscopy,
allowed for the synthesis of amorphous catalysts with high
surface areas and high catalytic performance. Very often amor-
phous materials of these types of compositions are more active
than crystalline materials since surface areas are typically
higher. The synthesized catalysts have several potential appli-
cations in areas such as respiratory protection systems. Opti-
mization of these catalysts was done by compositional
variation and by doping. Many of these latter examples have
involved combined microporous and mesoporous materials.
If micropores and mesopores are well connected, this can
markedly enhance mass transport of reagent and products lead-
ing to faster rates of reaction. Many of these examples also
suggest that like biological systems mixed valency is important.
The high activity of many amorphous systems which are more
readily characterized these days and more active than crystalline
counterparts in photo-assisted water oxidation and small-
molecule oxidation may be due to the similarity or closeness
of amorphous systems to homogeneous-based catalysts.
5.07.7 Conclusion
This chapter focuses on the major developments in porous
metal and metal oxide systems that have been published in

the last 10 years. Such contributions point to a very active field
of research that continues to be developed. Several different
applications have been discussed including preparation of new
adsorbents, catalysts, sensors, and optoelectronic and magnetic
materials as well as others. As new characterization methods
and new developments in existing methods advance, even bet-
ter resolution of nanosize materials will be achievable. More and
more studies of advanced energy materials such as batteries, fuel
cells, and supercapacitors are regularly appearing and will con-
tinue. Emphases on green and safe methods of synthesis, green
products, enhanced atom efficiency, and minimal cost of pro-
duction are also extremely relevant to human endeavors and
will continue to control these new developments. Synthesis of
new materials is a fundamental step in the advancement of
technology and will be a key driver in our quests for new
compounds, composites, and devices for applications in energy,
health and pharmaceuticals, and materials. For related chapters
in this Comprehensive, we refer to Chapters 5.01 and 7.04.
References
1. Naik, B.; Ghosh, N. N. Recent Pat. Nanotechnol. 2009, 3, 213224.
2. Lowell, S. Introduction to Powder Surface Area. John Wiley and Sons: New York,
1979.
3. Toberer, E. S.; Joshi, A.; Seshadri, R. Chem. Mater. 2005, 17, 21422147.
4. Toberer, E. S.; Seshadri, R. Chem. Commun. 2006, 31593165.
5. Jung, H. Y.; Kim, D. H.; Chun, H. K.; Kim, S. H.; Lim, C. S.; Byun, J. Y.; Jung, Y. J.
Adv. Energy Mater. 2011, 1, 11261132.
6. Yu, J.; Zhou, X.; Zhang, C. Materials Lett. 2012, 69, 1315.
7. Han, D.; Xu, T.; Su, J.; Xu, X.; Ding, Y. ChemCatChem 2010, 2, 383386.
8. Zhang, J.; Liu, P.; Ma, H.; Ding, Y. J. Phys. Chem. C 2007, 111, 1038210388.
9. Shim, J. H.; Jiang, X.; Bent, S. F.; Prinz, F. B. J. Electrochem. Soc. 2010, 157,
B793B797.
10. Leventis, N.; Chandrasekaran, N.; Sadekar, A. G.; Mulik, S.; Sotiriou-Leventis, C.
J. Mater. Chem. 2010, 20, 74567471.
11. Miyamoto, H.; Ueda, C.; Yamane, C.; Ago, M.; Okajima, K.; Endo, T.; Hattori, M.
Trans. Mater. Res. Soc. Jpn. 2009, 34, 719722.
12. Hakamada, M.; Watanabe, H.; Kuromura, T.; Asao, Y.; Chen, Y.; Kusuda, H.;
Mabuchi, M. In Porous Metals and Metallic Foams, MetFoam 2007, Proceedings
of the 5th International Conference on Porous Metals and Metallic Foams,
Montreal, QC, Canada, Sept. 57, 2007; Lefebvre, L. P., Banhart, J., Dunand, D. C.,
Eds.; DEStech Publications: Lancaster, 2008; pp 359362.
13. Utsonomiya, H.; Koh, H.; Miyamoto, J.; Sakai, T. Adv. Eng. Mater. 2008, 10,
826829.
14. Nakajima, H.; Ide, T.; Kim, S. Y. Mater. Sci. Forum 2009, 620622 (Eco-Materials
Processing and Design X), 785790.
15. Kwon, Y. J.; Shigematsu, I.; Saito, N. Mater. Trans. 2009, 50, 192196.
16. Erlebacher, J.; Seshadri, R. MRS Bull. 2009, 34, 561566.
17. Kloke, A.; von Stetten, F.; Zengerle, R.; Kerzenmacher, S. Adv. Mater. 2011, 23,
49765008.
18. Nouri, A.; Chen, X. B.; Hodgson, P. D.; Wen, C. E. Adv. Mater. Res. 2007, 1517,
7176.
19. Zhang, H.; Cooper, A. I. J. Mater. Chem. 2005, 15, 21572159.
20. Sollazzo, V.; Palmieri, A.; Girardi, A.; Farinella, F.; Carinci, F. J. Biotechnol.
Biomater. 2011, 1, 102.
21. Baril, E.; Lefebvre, L. P.; Hacking, S. A. J. Mater. Sci. Mater. Med. 2011, 22,
13211332.
22. Bernard, S.; Balla, V. K.; Bose, S.; Bandyopadhyay, A. Mater. Sci. Eng. C: Mater.
Biol. Appl. 2011, 31, 815820.
23. Wazen, R. M.; Lefebvre, L. P.; Baril, E.; Nanci, A. J. Biomed. Mater. Res. B Appl.
Biomater. 2010, 94, 6471.
24. Krishna, B. JOM 2008, 60, 4548.
25. Levine, B. Adv. Eng. Mater. 2008, 10, 788792.
26. Levine, B. R.; Fabi, D. W. Materialwiss. Werkst. 2010, 41, 10021010.
27. Liu, X. H.; Huang, S.; Picraux, S. T.; Li, J.; Zhu, T.; Huang, J. Y. Nano Lett. 2011,
11, 39913997.
28. Colombo, P.; Degischer, H. P. Mater. Sci. Technol. 2010, 26, 11451158.
Porous Metals and Metal Oxides 167
29. Onishi, H.; Hyun, S. K.; Nakajima, H.; Mitani, S.; Takanashi, K.; Yakushiji, K.
J. Appl. Phys. 2008, 103, 093539/1093539/5.
30. Shiina, Y.; Inagaki, T. Int. J. Heat Mass Transf. 2005, 48, 373383.
31. Tian, Y.; Zhao, C. Y. Energy 2011, 36, 55395546.
32. Fiedler, T.; Belova, I. V.; Murch, G. E.; Ochsner, A. Adv. Comp. Mater. Model.
2011, 73110.
33. (a) Sanchez-Sanchez, C. M.; Solla-Gullon, J.; Vidal-Iglesias, F. J.; Aldaz, A.;
Montiel, V.; Herrero, E. J. Am. Chem. Soc. 2010, 132, 56225624;
(b) Igarashi, A.; Ichikawa, N.; Sato, S.; Takahashi, R.; Sodesawa, T. Appl. Catal. A
Gen. 2006, 300, 5057.
34. (a) Lee, H.; You, S.; Pikhitsa, P. V.; Kim, J.; Kwon, S.; Woo, C. G.; Choi, M. Nano
Lett. 2011, 11, 119124; (b) Ozdemir, T.; Sandal, D.; Culha, M.; Sanyal, A.;
Atay, N. Z.; Bucak, S. Nanotechnology 2010, 21, 125603/1125603/6.
35. (a) Masuda, Y. In Bio-Inspired Materials Synthesis; Gao, Y., Ed.; Research
Signpost: Kerala, 2010; p 35; (b) Asahina, S.; Takami, S.; Otsuka, T.; Adschiri, T.;
Terasaki, O. ChemCatChem 2011, 3, 10381044.
36. (a) Shang, M.; Wang, W.; Yin, W.; Ren, J.; Sun, S.; Zhang, L. Chem. Eur. J. 2010,
16, 1141211419; (b) Sankaranarayanan, S. K. R. S.; Ramanathan, S. J. Phys.
Chem. C 2010, 114, 66316639.
37. (a) Yue, C.; Qiu, L.; Trudeau, M.; Antonelli, D. Inorg. Chem. 2007, 46, 50845092;
(b) Nikonova, O. A.; Nedelec, J. M.; Kessler, V. G.; Seisenbaeva, G. A. Langmuir
2011, 27, 1162211628; (c) Wang, D.; Ma, Z.; Dai, S.; Liu, J.; Nie, Z.;
Engelhard, M. H.; Huo, Q.; Wang, C.; Kou, R. J. Phys. Chem. C 2008, 112,
1349913509.
38. Sadakane, M.; Ueda, W. In Porous Materials; Bruce, D. W., OHare, D., Walton, R. I.,
Eds.; Wiley: Chichester, 2011; pp 147215.
39. Suib, S. L. Acc. Chem. Res. 2008, 41, 479487.
40. Boettcher, S. W.; Fan, J.; Tsung, C. K.; Shi, Q.; Stucky, G. D. Acc. Chem. Res.
2007, 40, 784792.
41. Toberer, E. S.; Lofvander, J. P.; Seshadri, R. Chem. Mater. 2006, 18, 10471052.
42. Alsobrook, A. N.; Hauser, B. G.; Hupp, J. T.; Alekseev, E. V.; Depmeier, W.;
Albrecht-Schmitt, T. E. Chem. Commun. 2010, 46, 91679169.
43. Shpeizer, B. G.; Bakhmoutov, V. I.; Zhang, P.; Prosvirin, A. V.; Dunbar, K. R.;
Thommes, M.; Clearfield, A. Colloids Surf. A Physicochem. Eng. Asp. 2010, 357,
105115.
44. Chmielarz, L.; Piwowarska, Z.; Kustrowski, P.; Gil, B.; Adamski, A.; Dudek, B.;
Michalik, M. Appl. Catal. B Environ. 2009, 91, 449459.
45. Jin, L.; Xu, L.; Morein, C.; Chen, C. H.; Lai, M.; Dharmarathna, S.; Dobley, A.;
Suib, S. L. Adv. Funct. Mater. 2010, 20, 33733382.
46. Chen, C. H.; Jin, L.; Espinal, A. E.; Firliet, B. T.; Xu, L.; Aindow, M.; Joesten, R.;
Suib, S. L. Small 2010, 6, 988992.
47. (a) Zimny, K.; Ghanbaja, J.; Carteret, C.; Stebe, M. J.; Blin, J. L. New J. Chem.
2010, 34, 21132117; (b) Liu, D.; Quek, X. Y.; Cheo, W. N. E.; Lau, R.; Borgna, A.;
Yang, Y. J. Catal. 2009, 266, 380390; (c) Balas, F.; Manzano, M.; Horcajada, P.;
Vallet-Regi, M. J. Am. Chem. Soc. 2006, 128, 81168117; (d) Wu, K. C. W.;
Yamauchi, Y.; Hong, C. Y.; Yang, Y. H.; Liang, Y. H.; Funatsu, T.; Tsunoda, M.
Chem. Commun. 2011, 47, 52325234; (e) Oveisi, H.; Jiang, X.; Nemoto, Y.;
Beitollahi, A.; Yamauchi, Y. Micropor. Mesopor. Mater. 2011, 139, 3844.
48. (a) Pazzuconi, G.; Bassi, G.; Millini, R.; Perego, C.; Perego, G. Eur. Pat. Appl.
1996, EP 691305 A1 19960110; (b) Li, C. C.; Yin, X. M.; Wang, T. H.; Zeng, H. C.
Chem. Mater. 2009, 21, 49844992; (c) Antonelli, D. M.; Nakahira, A.; Ying, J. Y.
Inorg. Chem. 1996, 35, 31263136.
49. (a) Balzer, C.; Wildhage, T.; Braxmeier, S.; Reichenauer, G.; Olivier, J. P. Langmuir
2011, 27, 25532560; (b) Takai, A.; Doi, Y.; Yamauchi, Y.; Kuroda, K. J. Phys.
Chem. C 2010, 114, 75867593.
50. (a) Dacquin, J. P.; Lee, A. F.; Pirez, C.; Wilson, K. Chem. Commun. 2012, 48,
212214; (b) Zhou, W.; Sun, F.; Pan, K.; Tian, G.; Jiang, B.; Ren, Z.; Tian, C.;
Fu, H. Adv. Funct. Mater. 2011, 21, 19221930.
51. (a) Sun, X.; Zhao, Y.; Lin, V. S. Y.; Slowing, I. I.; Trewyn, B. G. J. Am. Chem. Soc.
2011, 133, 1855418557; (b) Kim, S. H.; Huang, Y.; Sawatdeenarunat, C.;
Sung, S.; Lin, V. S. Y. J. Mater. Chem. 2011, 21, 1210312109.
52. Sadakane, M.; Ueda, W. In Porous Materials; Bruce, D. W., Ohare, D., Walton, W. I.,
Eds.; John Wiley and Sons: West Sussex, 2011; pp 147215.
53. (a) Soler-Illia, G. J. A. A.; Crepaldi, E. L.; Grosso, D.; Sanchez, C. Curr. Opin.
Colloid Interface Sci. 2003, 8, 109126; (b) Brinker, C. J.; Lu, Y.; Sellinger, A.;
Fan, H. Adv. Mater. 1999, 11(7).
54. (a) Wang, L.; Yamauchi, Y. Chem. Mater. 2011, 23, 24572465;
(b) Hartono, S. B.; Qiao, S. Z.; Liu, J.; Jack, K.; Ladewig, B. P.; Hao, Z.; Lu, G. Q. M.
J. Phys. Chem. C 2010, 114, 83538362.
55. (a) Katou, T.; Lu, D.; Kondo, J. N.; Domen, K. J. Mater. Chem. 2002, 12,
14801483; (b) Boettcher, S. W.; Fan, J.; Tsung, C.; Shi, Q.; Stucky, G. D. Acc.
Chem. Res. 2007, 40, 784792.
56. Soler-Illia, G. J. A. A.; Sanchez, C. New J. Chem. 2000, 24, 493499.
168 Porous Metals and Metal Oxides
57. Grosso, G.; Cagnol, F.; Soler-Illia, G. J. A. A.; Crepaldi, E. L.; Amenitsch, H. Adv.
Funct. Mater. 2004, 14, 4.
58. Rao, Y.; Kang, J.; Trudeau, M.; Antonelli, D. M. J. Catal. 2009, 266, 18.
59. Rao, Y.; Antonelli, D. M. J. Mater. Chem. 2009, 19, 19371944.
60. Shirokura, N.; Nakajima, K.; Nakabayashi, A.; Lu, D.; Hara, M.; Domen, K.;
Tatsumi, T.; Kondon, J. N. Chem. Commun. 2006, 20, 2194.
61. Singh, S.; Barden, W. R. T.; Kruse, P. ACS Nano 2008, 2, 24532464.
62. Idakiev, V.; Tabakova, T.; Tenchev, K.; Yuan, Z. Y.; Ren, T. Z.; Vantomme, A.;
Su, B. L. J. Mater. Sci. 2009, 44, 66376643.
63. Nadagouda, M. N.; Varma, R. S. Smart Mater. Struct. 2006, 15, 12601265.
64. Liu, H.; Du, X.; Xing, X.; Wang, G.; Qiao, S. Z. Chem. Commun. 2012, 48,
865867.
65. Li, L.; Zhu, Z.; Yao, X.; Lu, G.; Yan, Z. Micropor. Mesopor. Mater. 2008, 112,
621626.
66. Zang, Z.-A.; Yao, H. B.; Zhou, Y. X.; Yao, W. T.; Yu, S. H. Chem. Mater. 2008, 20,
47494755.
67. Haetge, J.; Suchomski, C.; Brezesinski, T. Inorg. Chem. 2010, 49,
1161911626.
68. Ding, Y.; Xu, L.; Chen, C.; Shen, X.; Suib, S. L. J. Phys. Chem. C 2008, 112,
81778183.
69. Barick, K. C.; Singh, S.; Aslam, M.; Bahadur, D. Micropor. Mesopor. Mater. 2010,
134, 195202.
70. Ciesla, U.; Froeba, M.; Stucky, G.; Schueth, F. Chem. Mater. 1999, 11, 227234.
71. Abdollahzadeh-Ghom, S.; Zamani, C.; Andreu, T.; Epifani, M.; Morante, J. R. Appl.
Catal. B Environ. 2011, 108109, 3238.
72. Ren, Yu. Catal. Lett. 2009, 131, 146154.
73. (a) Yue, W.; Zhou, W. Prog. Nat. Sci. 2008, 18, 13291338; (b) Yue, W.;
Randorn, C.; Attidekou, P. S.; Su, Z.; Irvine, J. T. S.; Zhou, W. Adv. Funct. Mater.
2009, 19, 28262833.
74. Kadib, A.; Primo, A.; Molvinger, K.; Bousmina, M.; Brunel, D. Chem. Eur. J. 2011,
17, 79407946 S7940/1S7940/7.
75. Liu, Q.; Chen, F.; Song, W. ECS Trans. 2009, 2009(19), 18.
76. Catalao, R. A.; Maldonado-Hodar, F. J.; Fernandes, A.; Henriques, C.;
Ribeiro, M. F. Appl. Catal. B Environ. 2009, 88, 135141.
77. Balaji, S.; Djaoued, Y.; Albert, A. S.; Ferguson, R. Z.; Bruning, R.; Su, B. L. J. Mater.
Sci. 2009, 44, 66086616.
78. Thakuria, H.; Borah, B. M.; Das, G. J. Mol. Catal. A Chem. 2007, 274, 110.
79. Mathew, T.; Suzuki, K.; Nagai, Y.; Nonaka, T.; Ikuta, Y.; Takahashi, N.; Suzuki, N.;
Shinjoh, H. Chem. Eur. J. 2011, 17, 10921095 S1092/1S1092/9.
80. Liu, H.; Wang, G.; Liu, J.; Qiao, S.; Ahn, H. J. Mater. Chem. 2011, 21, 30463052.
81. Brinker, C. J.; Dunphy, D. R. Curr. Opin. Colloid Interface Sci. 2006, 11, 126132.
82. Kingondu, C. K.; Opembe, N. N.; Genuino, H. C.; Garces, H. F.; Njagi, E. C.; Iyer, A.;
Huang, H.; Dharmarathna, S.; Suib, S. L. J. Phys. Chem. C 2011, 115, 2327323282.
83. Jiao, F.; Frei, H. Energy Environ. Sci. 2010, 3, 10181027.
84. Jiao, F.; Frei, H. Chem. Commun. 2010, 46, 29202922.
85. Li, X.; Hu, B.; Suib, S.; Lei, Y.; Li, B. Biochem. Eng. J. 2011, 54, 1015.
86. Jin, L.; Zhang, Y.; Dombrowski, J. P.; Chen, C. H.; Pravatas, A.; Xu, L.; Perkins, C.;
Suib, S. L. Appl. Catal. B Environ. 2011, 103, 200205.
87. Njagi, E. C.; Chen, C. H.; Genuino, H.; Galindo, H.; Huang, H.; Suib, S. L. Appl.
Catal. B Environ. 2010, 99, 103110.
 2.01 Transition-Metal Perovskites
AM Abakumov, University of Antwerp, Antwerp, Belgium
AA Tsirlin, Max Planck Institute for Chemical Physics of Solids, Dresden, Germany
EV Antipov, Lomonosov Moscow State University, Moscow, Russia
2013 Elsevier Ltd. All rights reserved.

2.01.1 Introduction 1
2.01.2 Cooperative Octahedral Tilting Distortions in Perovskites 2
2.01.3 Lone-Pair-Driven Instabilities in Perovskites 4
2.01.4 Electronic Instabilities in Perovskites 6
2.01.4.1 Polar Distortions in Perovskites Based on d0 B Cations 6
2.01.4.2 JahnTeller Distortions and Orbital Ordering 9
2.01.4.3 Charge Instabilities 14
2.01.4.4 Instabilities Related to the Spin State and Magnetism 16
2.01.5 Ordering Phenomena at Cation and Anion Sublattices of Perovskites 16
2.01.5.1 B-Site Ordering 16
2.01.5.2 Noncooperative Octahedral Tilting in 1:1 Ordered A2BB0 O6 Perovskites and A2BB0 F6 Elpasolites 19
2.01.5.3 A-Site Ordering 21
2.01.5.4 Ordering of Anion Vacancies in Perovskites 22
2.01.5.5 Coupled Cation and Anion-Vacancy Ordering 24
2.01.6 Aperiodic Order in Perovskites 26
2.01.7 Mixed-Anion Perovskites: Oxynitrides and Oxyfluorides 29
2.01.8 Conclusion 32
References 34

2.01.1 Introduction
The perovskite crystal structure is one of the most important
inorganic structures, thanks to its high relevance in a variety of
modern technologies, including high-critical temperature (Tc)
superconductors, colossal magnetoresistant and other magnetic
materials, mixed electronic and ionic conductors for applica-
tions in fuel cells and oxygen separation membranes, materials
for heterogeneous catalysis in oil industry, ferroelectrics, and
microwave ceramics. A salient feature of the perovskite structure
is its high flexibility toward variations in the chemical compo-
sition, and the ability to accommodate a wide range of elements.
Thousands of perovskite-type compounds have been already
prepared, and more will be discovered in the near future.
The mineral perovskite CaTiO3 was discovered by
G. Rose in 1839 in the Ural Mountains and named after the
Russian count Lev Perovskii. The ideal perovskite ABX3 struc-
ture (Figure 1) has cubic symmetry. It is conveniently repre-
sented as a three-dimensional (3D) isotropic framework of
corner-sharing BX6 octahedra that form cubooctahedral cavi-
ties accommodating the large A cations. In the cubic perovskite
structure the AX and BX interatomic distances (2dBX dAX)
are congruent, although significant deviations from this ideal
ratio can also be tolerated at the cost of geometrical distortions.
The most typical distortion, cooperative tilts of the BX6 octahe-
dra, is described in the first part of this chapter.
This chapter largely focuses on perovskites inheriting the
arrangement of the A and B sites from the parent ABX3 struc-
ture. Both A and B positions can be occupied by single cations,
or by different chemical species having distinct formal charges,
ionic radii, and coordination numbers. The mixing of different
cations at the same site is one of the main approaches to a
deliberate modification of chemical and physical properties of
a perovskite material. Ordered cation arrangements in perov-
skites and the interplay between the cation ordering and octa-
hedral tilting are also discussed, with particular attention to the
novel noncooperative tilting distortion mode of the perovskite
framework.
We devote significant attention to anion-deficient perov-
skites ABX3d showing a wide range of anion nonstoichiometry
(up to d 1). In anion-deficient perovskites, the robust anion-
vacancy ordering is favored for B cations preferring the stable
local environment with the coordination number less than 6
(e.g., tetragonal pyramid, square, tetrahedron, or even dumb-
bell) according to peculiarities of the electronic structure. The
ordering of anion vacancies is often coupled to the A and/or
B-cation ordering, thus triggering the formation of various
structures with reduced dimensionality. However, if the anion
deficiency is present in a perovskite with the lone-pair A cation,
such as Pb2 and Bi3, a new class of perovskites with crystal-
lographic shear (CS) planes emerges. Although CS planes were
generally considered forbidden structural elements in perov-
skites, this type of ordered planar defects has been recently
found in the anion-deficient transition-metal perovskites with
the Pb2 and Bi3 cations, as discussed further in this chapter.
Recently, a very unusual square coordination was found for the
Fe2 cation in the SrFeO2 compound prepared by a soft-
chemistry route at low temperatures. This finding puts forward
the new approach of the kinetic rather than thermodynamic
stabilization of new perovskite-type structures. It may trigger
new studies and substantially expand the list of anion-deficient
perovskites with the unusual coordination environment of

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00201-1 1

2 Transition-Metal Perovskites
-A -B -X
Figure 1 The crystal structure of perovskite ABX3.
transition-metal cations. For the review of the recent progress in
this field, the reader is referred to Chapter 2.15.
The last part of this chapter deals with mixed-anion perov-
skites, including oxyfluorides and oxynitrides. In these com-
pounds, the kinetic stabilization is extremely important.
Moreover, the formation of perovskites with the ordered or
disordered anion patterns can be controlled by a proper selec-
tion of the synthesis route.
The perovskite family can be greatly expanded by preparing
layered and/or intergrowth structures. The perovskite structure
can be seen as the cubic close-packed structure of the mixed
cationanion AX3 layers alternating along the threefold axes
(body diagonal) of the cubic unit cell, with of the octahedral
cavities formed exclusively by the X anions. These cavities are
filled by the small B cations. For large A cations, different
packings of the (AX3) layers can occur, thus increasing the
number of possible structures enormously, because a number
of such packings is infinite. These structures are known as hex-
agonal polytypes of the perovskite. They feature the BX6 octa-
hedra sharing both corners and faces, in contrast to the cubic
perovskite showing the corner-sharing connections only. This
broad field is, however, beyond the scope of the present chapter.
The lattice parameters of certain perovskites can match those
of other simple inorganic structures, such as rock salt or calcium
difluoride. This match facilitates the formation of so-called
intergrowth structures, where slabs of different structures alter-
nate along one of the crystallographic directions. The interest
in the perovskite-based intergrowth structures has drastically
increased during the last 30 years, mainly inspired by the
discovery of superconductivity in Cu-based complex oxides.
The intergrowth structures form different homologous series of
inorganic compounds. For example, the RuddlesdenPopper
series with the general formula An1BnO3n1 includes a perov-
skite slab of variable thickness sandwiched between the rock-salt
bilayered blocks. Perovskite-related intergrowth phases are
described in detail in Chapter 2.03.
Due to space limitation we are not able to review all the
aspects of crystallography and crystal chemistry of the perovskite-
related materials. Perovskite-related structures, which are out
of scope of this chapter, are comprehensively reviewed by
Mitchell.1
Recent discoveries of new structure types and synthesis
routes, as well as new exciting properties resulting from various
electronic instabilities, suggest that the perovskite structure is
indeed an inexhaustible source of new materials. The main
goal of the present chapter is to give an outline of recent
research on transition-metal perovskites, with a focus on their
fascinating and diverse crystal chemistry that underlies the
variable and technologically important chemical and physical
properties.
2.01.2 Cooperative Octahedral Tilting Distortions in
Perovskites
The 3D framework of corner-sharing octahedra in the perov-
skite structure is a very flexible construction. The ability of this
framework to sustain distortions due to rotations of the octa-
hedral units, or their deformations, or both together, without
losing the corner-sharing connectivity, allows the perovskite
structure to accommodate a very wide range of cations with
diverse size, formal charge, and electronic properties. Crystal-
lographic aspects of the distortions have been widely discussed
in the literature starting from pioneering works by Glazer2 and
Megaw,3 and developed nowadays toward a rigorous group
symmetry analysis for different types of distortions and their
combinations. The crystallographic concepts and the specific
language behind the distorted perovskites are important for
understanding the structures, chemical bonding, and physical
properties of these compounds and they are briefly summa-
rized here.
The most common type of perovskite structure distortion
can be intuitively understood from simple geometrical consid-
erations. In the ideal cubic ABX3 perovskite, the BX and AX
bond lengths must obey the d(AX) 2d(BX) p relation that
also can be expressed as t (RX RA)/[ 2(RX RB)], t 1,
where RA, RB, and RX are ionic radii of A, B, and X, respectively.
The tolerance factor t was first introduced by Goldshmidt,4
who calculated the optimal size of the A cation by considering
the structure as a close packing of atomic spheres. However,
if the ionic radius of the A cation is smaller than optimal
(t < 1), the perovskite structure adjusts the first coordination
sphere of this cation at the cost of a deformation in the octa-
hedral framework. In a first approximation, this deformation
can be viewed as a rotation (or tilt) of rigid BO6 octahedra
around one, two, or all three unit cell axes of the parent cubic
perovskite structure (or around the fourfold symmetry axes of
the BX6 octahedron). The deformation of the BX6 octahedra is
assumed to be small in comparison with the magnitude of
atomic displacements introduced by the rotation.
The octahedra do not behave independently, because they
are linked into a framework by common corners. The require-
ment to maintain the corner-sharing connectivity restricts the
amount of possible tilting patterns. In order to enumerate such
patterns (or tilt systems), Glazer introduced a special notation,
with the rotations of the octahedra represented by a symbol
consisting of three positions associated to three Cartesian axes
[100], [010], and [001]. Every position contains a letter a, b, or c
and a superscript symbol 0, , or . Equal letters in different
positions denote equal tilt angles around the corresponding
axes. The superscript symbol 0 indicates the absence of the
tilt. The superscript symbol stands for an in-phase tilt of
two neighboring octahedra along the tilt axis, whereas the
symbol  stands for an out-of-phase tilt (Figure 2). Thus, the
undistorted cubic perovskite is described by the a0a0a0 tilt

bp bp
ap ap
a0a0c+ a0a0c
0 0 0 0 
Figure 2 In-phase (a a c ) and out-of-phase (a a c )
single-component tilt systems.
NdAlO3 GdFeO3
aaa a+bb
  
Figure 3 The crystal structures of NdAlO3 (a a a tilt system) and
GdFeO3 (abb tilt system).
system. Most common tilt systems occurring among the
perovskite-like compounds are aaa (NdAlO3 structure
type) and abb (GdFeO3 structure type) (Figure 3).
The octahedral tilt is, of course, impossible within the cubic
perovskite unit cell with the Pm3m space symmetry. Therefore,
the tilting distortion leads to an increase in the unit cell volume
and a symmetry reduction. Glazer defined 23 possible tilt
systems in perovskites and determined their space symmetries.
A similar analysis was independently performed by Ale-
ksandrov et al.5 Their ideas have been later expanded by
Woodward.6 Howard and Stokes performed a group-theory
analysis of the tilt systems in perovskites, thus reducing the
amount of unique tilt systems from 23 to 15.7 Considering
each tilt system as a linear combination of six basic tilt com-
ponents, ab0b0, b0ab0, b0b0a, ab0b0, b0ab0, and b0b0a,
Howard and Stokes found that the coefficients of this linear
combination span a reducible representation of the Pm3m
space group (the atoms A, B, and X are at 12 , 12 , 12 (Wyckoff
position 1b), 0,0,0 (1a), and 12, 0, 0 (3d), respectively), consist-
ing of two irreducible representations (irreps) M3 and R4
associated with the in-phase and out-of-phase basic tilts, re-
spectively (here, the notations of Miller and Love are used8).
Transition-Metal Perovskites 3
GdFeO3 NdAlO3
a+bb aaa
Figure 4 A-site coordination environment in the abb (GdFeO3) and
aaa (NdAlO3) tilt systems. The A cations and oxygen atoms are
shown as yellow and blue spheres, respectively. Short, medium, and long
bonds are shown in brown, green, and cyan colors, respectively. The
bond lengths for GdFeO3 are: 2.2652.382 A (short), 2.822 A (medium),
and 3.1923.481 A (long). The bond lengths for NdAlO3 are: 2.386 A
(short), 2.659 A (medium), and 2.936 A (long).
This way, a list of subgroups of Pm3m for these irreps is
obtained. The list of the 15 tilt systems, corresponding space
groups, perovskite subcell-to-supercell transformation matri-
ces, and the atomic positions in the supercell are given in
Appendix. Note that the tilt systems featuring a combination
of in-phase and antiphase tilts around the same axis (which
results in unequal tilt magnitudes of the octahedra along this
axis) are not included in this list. The symmetry analysis in-
cluding such tilt systems has been performed by Aleksandrov
and Bartolome.9
The driving forces stabilizing the tilting distortion of the
perovskite structure can be understood from a simple ionic
model. Although in the transition-metal perovskites the
bonds between the transition-metal cations and the X anion
are far from being purely ionic and have always a significant
degree of covalency, the bonds between electropositive A cat-
ions (alkali, alkali-earth, and rare-earth metals) and the X
anion can be considered as substantially ionic. The tilting
distortion does not noticeably affect the first coordination
sphere of the B cations, but alters significantly the first coordi-
nation sphere of the A cations. Some anions move closer to the
A cation, while others move away from it, out of the first
coordination sphere. This way, 12 equivalent AX interatomic
distances in the AX12 cuboctahedron are split into several
groups.
In the distorted perovskites with the most common tilt
systems aaa and abb (Figure 4), the A-site coordination
environments are represented by three short six medium
three long AX bonds and six short two medium four long
AX bonds, respectively, where the long AX separations can
be considered as basically nonbonding. Woodward estimated
the lattice energy of the perovskites with different tilt systems
by taking into account Coulomb attraction (Madelung energy)
and noncoulombic short-range repulsion.10 The lowest energy
was found for the aaa and abb tilt systems that are most
frequently observed among perovskites. The aaa system
maximizes both Coulomb and repulsive interactions because
of three short AX distances that cannot be relaxed due to a
4 Transition-Metal Perovskites
displacement of the A cation in a high-symmetry 6a site of the
R3c space group (see Appendix). Thus, this tilt system should
be the most favorable for highly charged A cations, such as
trivalent rare-earth metals (large Coulomb interaction) and
moderately small tilting distortion corresponding to
0.975 < t < 1.
At t < 0.975, the abb tilt system becomes more stable,
because the local A-site symmetry (site 4c in Pnma space group)
allows the repulsive forces to be relieved by a displacement of
the A cation. Covalency of the AX bonding seems to be of
secondary importance for the competition between the tilt
systems. Woodward qualitatively demonstrated that the short-
ening of the AX interatomic distances increases the AX
orbital overlap, while the modified symmetry of the coordina-
tion environment of the A cation may significantly reduce the
participation of certain orbitals of the A cation in the covalent
bonding with the X anion.10 In the aaa tilt system, three
closest anions form a triangle around the A cation, and the
orbitals of the A cation, which are orthogonal to the plane of
this triangle, would have almost zero overlap with the
p-orbitals of the anions, thus being nearly excluded from
covalent AX interaction. For the optimal AX orbital overlap,
the abb tilt system is more favorable because of six anions
effectively coordinating the A cation. This could be a possible
reason why the GdFeO3-type structure is preferred for relatively
electronegative A cations, such as Ca2, showing a large cova-
lency of the AX bonds.11 Ab initio calculations for LaTiO3
demonstrated that the GdFeO3-type distortion is partly driven
by the covalent bonding due to s-overlap between occupied
oxygen p states and empty A-cation d states.12,13 The effect of
the BX covalent bonding on the relative energies of different
tilt systems is quite weak and rather concerns the energy dif-
ference between the distorted structure and its cubic a0a0a0
prototype. For the transition-metal perovskites, the p-bonding
along the BXB chain is an additional factor stabilizing the
undistorted cubic structure. The effect of this stabilization de-
pends on the spatial extent of the d-orbitals of the transition
metal, which is crucial for an effective p-overlap with the
p-orbitals of the anions. It is also sensitive to the filling of the
p* antibonding band predominantly formed by t2g orbitals of
the transition metal.
The ionic model behind the octahedral tilting distortion in
perovskites can be expressed quantitatively using a bond
valence sum (BVS) concept. For a given MXn coordination
polyhedron, the sum of n individual bond valences
s er0 r =B (r is the BX distance, and r0 and B are empirical
tabulated constants) should match the formal oxidation state
of the cation.1416 The algorithm of predicting the perovskite
structure distortion, as implemented in the SPuDS program,17
is based on the calculation of unit cell dimensions and vari-
able atomic coordinates from the BX bond distances, octa-
hedral tilt angle, and the positional parameters of the A
cations.18 The BX distances for the rigid BX6 octahedra and
the optimum tilt angle are calculated using BVS taking into
account CN 6 for the B cation, CN 6 for the X anion
(2B 4A), and CN 12 for the A cation. The optimized struc-
ture corresponds to a minimal difference between the calcu-
lated BVSs and the formal valence of each ion. The relative
structure stability is measured by the global instability index
(GII), expressed as
s
XN 
2
Vi, f  Vi, c
GII i1
N
where Vi,f and Vi,c are the formal valences and calculated BVSs
for the ion i, respectively, and N is a number of cations in the
asymmetric unit.19 The SPuDS method is proven to be accurate
for the ABX3 perovskites and can be also extended to the
A2BB0 X6 ordered double perovskites (see Section 2.01.5.1).20
Many transition-metal perovskites reveal an intricate inter-
play of the tilting distortions and deformations of the BX6
octahedra. Since deformations of the octahedra originate
from electronic effects on the B site, such combined distortions
are considered in Section 2.01.4.2.
2.01.3 Lone-Pair-Driven Instabilities in Perovskites
Another type of geometrical instabilities in perovskites is a distor-
tion driven by a lone-pair cation. The cations having an ns2 lone
pair are abundant, yet only Pb2 and Bi3 form ABX3 perovskites
with a transition metal on the B site. The typical effect of the lone-
pair cation is the displacement from the ideal position at the
center of the cubooctahedron, and the ensuing reduction in the
coordination number from 12 to 78. The BVS is maintained by
shortened AO bonds, whereas the local coordination environ-
ment of the A cation becomes strongly asymmetric. This feature
distinguishes the lone-pair-driven distortions from the conven-
tional tilting distortions discussed above (Section 2.01.2). While
the small A cation without the lone pair remains nearly at the
center of the cubooctahedral cage, the lone-pair cation shows a
sizable displacement and a strong tendency for the asymmetric
environment. Generally, the lone-pair-driven distortions are not
directly coupled to the octahedral tilting and may either precede
or follow them. The lone-pair-driven distortions are also present
in those perovskite structures that are not prone to other geomet-
rical instabilities.
The basic features of the lone-pair-driven distortions are as
follows:
(i) at ambient pressure and room temperature, sizable dis-
placements of the lone-pair A cations are generally
observed;
(ii) the lone-pair distortion may be superimposed on a tilting
distortion, and generally sets in/disappears upon dispa-
rate phase transitions; and
(iii) high temperature/high pressure tend to suppress the lone-
pair distortions.
Intuitively, the asymmetric local environment can be un-
derstood as the formation of a void accommodating the 6s2
lone pair. The microscopic picture is, however, quite different.
The 6s2 states of the lone-pair cations are not inactive, but
rather mix with p states of the ligand, and it is this mixing
that causes the shortening of the bonds and the A-cation
displacement.21,22 Instead of being tucked into the void, the
6s2 electrons strongly hybridize with the ligand states and gain
energy by creating short bonds to the ligands. An important
implication of this mechanism is the remarkable sensitivity
of the sterical activity to the electronic structure of the ligand.
For example, the lone-pair-driven distortions are very

pronounced in oxides, but strongly suppressed in sulfides,
where the 3p states of sulfur only weakly mix with the cation
6s2 states.23 Since oxygen is the typical ligand in transition-
metal perovskites, strong geometrical distortions and low-
symmetry structures are generally observed.
Combining the lone-pair A cation with most of the
transition-metal B cations in the perovskite structure is a chal-
lenging task due to the metastable nature of the target com-
pounds. The low thermodynamic stability, which is likely
caused by the strong geometrical distortion, is mitigated under
high pressure, where structures with higher symmetries are
formed. Indeed, most Pb- and Bi-based perovskites can only
be prepared by solid-state synthesis under high pressure.2428
Interest in such compounds is largely motivated by the polar
nature of the lone-pair-driven instabilities. The implications of
this effect are discussed later in this section.
The lone-pair-driven distortions are most easy to follow in
perovskites with a negligible size mismatch between the A and B
cations (t = 1). For systems with the tolerance factor close to
unity, tilting distortions do not occur, and the effect of the lone-
pair cation is most tangible. The displacement of the A cations
inevitably changes the local environment of the B cation that
can further induce its own instability. For example, PbTiO3
(t 1.022)29,30 is a tetragonally distorted perovskite (c/
a 1.064), with both Pb and Ti atoms displaced along [001].
Titanium forms one short bond to oxygen along the c axis due to
the second-order JahnTeller (SOJT) effect typical for d0 cations
(see Section 2.01.4.1). The tetragonal distortion is enhanced in
PbVO3 (t 1.039, c/a 1.229)24,25 and BiCoO3 (t 0.974, c/
a 1.267),27 where V4 and Co3 favor the square-pyramidal
coordination and basically shed the sixth oxygen atom. This
way, the conventional perovskite framework of corner-sharing
octahedra transforms into layers of square pyramids, thus ren-
dering the structure 2D (Figure 5). Indeed, PbVO3 reveals low-
dimensional magnetism,31,32 which is highly unusual for perov-
skites. A similar layered structure is formed in BiFeO3 under
compressive strain.3335
The smaller Bi atoms induce a size mismatch for most of the
3d transition-metal cations and, therefore, cause tilting distor-
tions. This size mismatch is accommodated by the conventional
GdFeO3-type abb tilt that is observed in BiCrO3 (t 0.931),36
BiMnO3 (t 0.914),37,38 and BiFeO3 (t 0.915)3941 at elevated
temperatures, where the Bi displacements are small or even
absent. At low temperatures, the lone-pair-driven displacements
Cell volume
1013%
volume
collapse
Pressure
Figure 5 Sketch of the pressure-induced volume collapse in BiCoO3.
The low-pressure tetragonal phase transforms into the high-pressure
orthorhombic (GdFeO3-type) phase with lower volume.
Transition-Metal Perovskites 5
arise and form an additional superstructure. The polarity of the
structure depends on whether the Bi displacements are arranged
in a parallel (ferroelectric) or an antiparallel (antiferroelectric)
manner (Figure 6). Both ferroelectric and antiferroelectric struc-
tures can be stabilized. BiScO3 (t 0.876)28 and BiCrO3
(t 0.931)36 are antiferroelectric (space group C2/c), BiFeO3 is
ferroelectric (space group R3c),42 whereas BiMnO3 demonstrates
a subtle balance of ferroelectric and antiferroelectric instabilities.
The peculiar behavior of this compound is related to the coexis-
tence of the lone-pair distortion and orbital degrees of freedom
of Mn3 (see Section 2.01.4.2). The structure determined from
neutron powder diffraction is centrosymmetric,4345 but local
studies based on transmission electron microscopy and pair
distribution function analysis indicate the lower C2 symmetry
of the long-range structure and even signatures of the short-range
order with P2 symmetry.46,47 These weak distortions induce the
ferroelectric response of BiMnO3.37,48 Additionally, a series
of intricate oxygen-excess49,50 and oxygen-deficient51 phases
BiMnO3d has been reported.
In contrast to Bi-based compounds, Pb-containing perov-
skites with low tolerance factors show the simultaneous onset
of the octahedral tilting and ordered lone-pair-driven distortions.
Both PbZrO3 (t 0.967)5254 and PbHfO3 (t 0.969)55 undergo
a single phase transition from the high-temperature cubic phase
to the low-temperature orthorhombic structure featuring the
ab0a tilting distortion and the antiferroelectric displacement
pattern of the Pb atoms (Figure 6, right). This superstructure is
driven by a combination of the lone-pair effect and size mis-
match, along with a weak tendency of the B cation to displace
from the center of the octahedron.56,57 A similar superstructure
has been discovered in BiFeO3-based solid solutions upon dop-
ing with La or Nd on the A site or with Mn on the B site.5861
While most of the Pb- and Bi-containing perovskites show
clear signatures of the lone-pair-driven distortions, the robust
local environment of the B cation can also suppress the sym-
metry reduction and mask the effect of the lone pair. For exam-
ple, PbCrO3 (t 1.025) is metrically cubic due to the tendency
of Cr4 perovskites to maintain the cubic framework and
the metallic behavior. However, the Pb atoms do not reside
in the high-symmetry position, and randomly displace to form
the typical asymmetric local environment. The displacements
are short-range ordered in a complex superstructure.26 A similar
scenario has been proposed for cubic PbFeO2F, although no
signatures of ordered Pb displacements are reported.62
Similar to the tilting distortions, lone-pair-driven displace-
ments of the A cation are suppressed at high temperatures or
under hydrostatic pressure. At high temperatures, many of the
lone-pair transition-metal perovskites undergo phase transitions
toward cubic (PbTiO3, PbZrO3, PbHfO3) or GdFeO3-type
(BiFeO3, BiMnO3, BiCrO3) structures. The local A cation displace-
ments are, however, preserved even above the transition
temperature.53 Many of the Pb-containing perovskites are unsta-
ble at high temperatures and ambient pressure, but their lone-
pair-driven distortions can be suppressed under high pressure,
thus stabilizing the structure. Perovskites with a marginal size
mismatch (PbVO3, PbCrO3) transform into cubic phases under
a pressure of about 2 GPa, and show fascinating examples of
about 10% volume collapse,63 presumably accompanied by the
metallization.25 An even larger volume collapse of 13% in
BiCoO3 is supplemented with a spin-state transition for Co3
6 Transition-Metal Perovskites
PbTiO3
Figure 6 Crystal structures of perovskite ferroelectrics. Arrows denote ferroelectric displacements of A cations in PbTiO3 and BiFeO3, as well as
antiferroelectric displacements in PbZrO3.
P4mm (T) Amm2 (O)
Figure 7 B-cation displacements in different ferroelectric phases of BaTiO3 and KNbO3. The right panel shows a sketch of pressuretemperature
phase diagram with C denoting the paraelectric cubic phase.
resulting in the GdFeO3-type structure at high pressures
(Figure 5).64 Other Bi-based compounds, such as BiMnO3 and
BiFeO3, lack the spectacular volume collapse effect and rather
transform into the GdFeO3-type structure above approximately
10 GPa, while at lower pressure a welter of complex superstruc-
tures is observed.6568 Upon further compression, some of the
lone-pair perovskites reveal intriguing new phenomena, such as
the metallization of BiFeO3 above 47 GPa69 and the likely re-
entrance of polar distortions (ferroelectricity) in PbTiO3 above
45 GPa.7072 Although details of these transformations have not
been elucidated, the structural changes under very high pressures
are likely unrelated to the lone-pair cations and reflect peculiarities
of the B cations or general electronic effects in the perovskite
family.69,73
The lone-pair cations are a handy tool for introducing local
distortions and polarity into the perovskite structure. For exam-
ple, in PbTiO3 the off-center displacement of Ti4 is reinforced
by the polar displacement of Pb2. The applications of this
approach were boosted by an interest in multiferroics, the
systems combining electric and magnetic long-range orders.74
Simple microscopic arguments show that ferroelectricity and
magnetism are unlikely to coexist, because polar displacements
are typical for nonmagnetic d0 B cations only, while the cations
with the partially filled d shell are usually reluctant to displace,
or undergo nonpolar JahnTeller distortions (see Section
2.01.4.2).75,76 A natural solution is to use the lone-pair A cation
for inducing polarity and the magnetic B cation for introducing
the magnetism.75,76 The main obstacles of this approach are
antiferroelectric and/or antiferromagnetic orders that do not
produce any net polarization/magnetization, and low break-
down voltages that prevent the switching of the polarization
BiFeO3 PbZrO3
Pressure
C
T
O
R
R3m (R) Temperature
by an electric field. For example, tetragonally distorted perov-
skites PbVO3 and BiCoO3 are pyroelectrics with an outstanding
electric polarization, which is impaired by very high coercive
fields exceeding the breakdown voltage.25,27
Mixed perovskites open further opportunities for combining
magnetism and ferroelectricity. For example, the rock-salt-ordered
double-perovskite Bi2MnNiO6 presents a rare example of a ferro-
electric ferromagnet.77 While this compound can only be prepared
under high pressure, an adjustment of the B-cation sublattice may
facilitate the synthesis of Bi-containing perovskites at ambient
pressure. The compounds Bi2Mn4/3Ni2/3O6, BiTi3/8Fe2/8Ni3/8O3,
and BiTi3/8Fe2/8Mg3/8O3 were successfully isolated by a conven-
tional solid-state synthesis.78,79 Electronic properties of these com-
pounds are fairly complex due to the local order of the B-site
cations and the formation of incommensurately modulated
structures.80 Despite the appealing properties, the application of
Bi2MnNiO6 is impeded by a relatively low Curie temperature of
140 K. By contrast, the antiferromagnetic ferroelectric BiFeO3
demonstrates a sizable magnetoelectric coupling at room temper-
ature and is actively studied with respect to possible applications.
The lack of the net magnetization is mitigated by the preparation
of thin films and the exchange bias effect. Further information on
this material can be found in the comprehensive review article.42
2.01.4 Electronic Instabilities in Perovskites
2.01.4.1 Polar Distortions in Perovskites Based on d0 B
Cations
The B cations with an empty d shell (so-called d0 cations) are
prone to polar distortions driven by a cation displacement
from the center of the BO6 octahedron.81 This effect

originates from the mixing between the cation d (empty) and
ligand p (filled) states, and entails the formation of short BX
bonds. Such displacements are of polar nature, while their
long-range order and interplay with other types of distortions
determine whether the resulting structure is ferroelectric or
antiferroelectric. The displacements of the d0 cations are often
termed SOJT, because the leading energy term favoring the
distortion is quadratic in the displacement coordinate, in con-
trast to the linear term for systems with partially filled d
shells.82,83 The conventional JahnTeller effect is, however, dif-
ferent from its second-order counterpart. The former involves
a change in the ligand positions, while the latter causes a dis-
placement of the B cation inside the rigid BX6 octahedra.
In contrast to the conventional JahnTeller distortion, the dis-
placements of the d0 cations are usually of polar nature.83
Polar distortions driven by the B cation can be summarized
as follows:
(i) Similar to lone-pair-driven distortions and octahedral
tilts, the B-cation displacements are eliminated at high
temperatures and/or under high pressure. However, they
do not always coexist with other types of distortions, and
may be even suppressed by octahedral tilts.
(ii) The tendency to polar displacement is not an exclusive
property of the d0 B cations. It rather depends on details
of electronic structure and, under certain conditions, may
be also relevant for non-d0 cations.
(iii) The B-cation polar displacements are unusually robust,
and can introduce polarity even in systems with large
cation disorder.
The origin of polar distortions in perovskites based on d0 B
cations has been addressed in numerous microscopic studies.
It is now well established that the strong hybridization be-
tween the cation d and ligand p states is responsible for the
effect.8486 However, the reasons why this effect is missing in
non-d0 perovskites are less clear. A possible explanation is
based on a competition between partially filled and empty d
orbitals favoring the nonpolar and polar local environments,
respectively.87 Another conjecture puts forward the crucial role
of Hunds coupling in destabilizing the polar distortions (i.e.,
the d electrons should be idle to produce the maximum spin,
instead of forming covalent bonds with oxygen).76 While the
general microscopic model has not been developed yet, com-
putational studies arrive at an intriguing prediction of B-site-
induced ferroelectricity in non-d0 Mn4 (CaMnO3, SrMnO3,
BaMnO3)83,88,89 and Cr3 (LaCrO3)90 perovskites under a
strong tensile strain. Additionally, ferroelectricity is experimen-
tally observed in the Sr1xBaxMnO3 solid solutions at x  0.5.91
Here, the empty eg states of the d3 Cr3 and Mn4 should
be responsible for the SOJT. However, the bulk of the experi-
mental information still concerns perovskites based on d0
cations.
Superimposing octahedral tilting and polar B-cation dis-
placements in perovskites creates numerous possible distorted
structures. The distortion associated with off-center displace-
ments of the B cations conforms to the irreducible representa-
tion G 4 of the parent Pm3m space group, which can be
combined with the M
3 and R4 irreps imposing the in-phase
and out-of-phase octahedral tilting. The symmetries of the
ensuing distorted structures were analyzed by Aleksandrov
and Bartolome9 and further refined by Stokes et al.92 This
Transition-Metal Perovskites 7
way, six isotropy subgroups for the polar structures without
octahedral tilts and 40 isotropy subgroups for the polar struc-
tures combined with the octahedral tilts have been identified.
Small A cations induce the GdFeO3-type distortion and
generally prevent polar instabilities at the B site, as in ortho-
rhombic CaTiO3 (t 0.946)93,94 that shows a robust tilting
pattern of GdFeO3 type up to at least 60 GPa.95 However,
CdTiO3 with a similar size mismatch (t 0.924) is ferro-
electric (space group Pna21) because of the strong CdO
hybridization.9698 Larger A cations eliminate tilting distor-
tions, and allow for a sequence of structural transformations
primarily driven by the B-site cation.55,56 Both BaTiO3
(t 1.061) and KNbO3 (t 1.091) undergo phase transitions
from the high-temperature cubic to the tetragonal (P4mm),
orthorhombic (Amm2), and eventually to the low-temperature
rhombohedral (R3m) phases. The last three phases are ferro-
electric, with Ti or Nb atoms shifted toward a vertex (<[001]>
direction), edge (<[110]> direction), and face (<[111]>
direction) of the BO6 octahedron, respectively (Figure 7).99,100
The nature of phase transitions in BaTiO3 and KNbO3 has
long been a matter of debate. On the one hand, the trans-
formations can be understood as displacive transitions driven
by a softening of a specific phonon mode.101,102 This phonon
mode entails certain atomic displacements, and goes down in
frequency (softens) as the transition temperature is
approached from above. Once the frequency of the mode
reaches zero, the respective atomic displacements freeze and
form the new structure. An alternative mechanism assumes
that the displaced atoms develop ordered domains above the
transition temperature, and the formation of the new structure
is triggered by the growth of such domains (orderdisorder
mechanism).103 Since experimental data reveal signatures of
both mechanisms,104106 the ferroelectric transitions in perov-
skites are presently ascribed to a special class combining the
displacive and orderdisorder components.107,108
The d0 perovskites with a negligible size mismatch demon-
strate even more complex behavior. In NaNbO3 (t 0.973) and
AgNbO3 (t 0.987), about six different polymorphs with differ-
ent combinations of Nb5 displacements and tilting distortions,
both ferroelectric and antiferroelectric, are reported; yet only a
meager understanding of the respective structures and their sta-
bility has been obtained so far (see also Section 2.01.6).109114
Compared to NaNbO3 and AgNbO3, SrTiO3 with negligible
size mismatch (t 1.001) is seemingly more simple and thor-
oughly studied. It features both ferroelectric (PbTiO3-like) and
antiferroelectric soft modes that approach zero frequency at
low temperatures.115,116 In contrast to BaTiO3 that undergoes
ferroelectric transition at 393 K, SrTiO3 remains cubic down to
105 K. This remarkable stability of the cubic phase should be
traced back to the negligible size mismatch rendering any kind
of distortion unfavorable. Below 105 K, SrTiO3 forms the te-
tragonal antiferroelectric phase (I4/mcm) with marginal rota-
tions of the TiO6 octahedra and without any notable
displacements of the Ti atoms.117 The ferroelectric instability
should be operative at even lower temperatures but, surpris-
ingly, it does not set in. The suppression of ferroelectricity in
SrTiO3 is explained by quantum fluctuations that induce the
tunneling between the competing ground states and are similar
in effect to thermal fluctuations that are present in other ferro-
electrics, yet remain operative even at zero temperature.118
Therefore, SrTiO3 is often termed quantum paraelectric.119
8 Transition-Metal Perovskites
The low-temperature quantum paraelectric state of SrTiO3
is highly fragile. For example, elastic strain induces the ferro-
electric transition.120 Alternatively, ferroelectricity can be
triggered by isotope substitution, with SrTi18O3 undergoing
a ferroelectric transition at 24 K.121 Details of the low-
temperature phase are, however, obscure because of the severe
phase separation. Both rhombohedral (BaTiO3-type) and or-
thorhombic (Ima2 symmetry) structures have been proposed
as plausible candidates.122125
Similar to high temperature, high pressure suppresses ferro-
electric instabilities in Ti4- and Nb5-based perovskites.
For example, BaTiO3 undergoes the same sequence of phase
transitions (rhombohedraltetragonalorthorhombiccubic)
upon heating and upon compression (Figure 7, right).126,127
However, the origin of pressure-induced transitions in ferroelec-
trics is not obvious. Unlike tilting distortions or lone-pair dis-
tortions, polar displacements are not a size effect and might be
present even in the compressed environment. The polar distor-
tions are suppressed under pressure because of the rapidly in-
creasing short-range repulsion that disfavors the formation of
local dipoles. By contrast, the metalligand hybridization gen-
erally increases with pressure, and may even lead to the reen-
trance of ferroelectricity at very high pressures: for example,
PbTiO3 likely becomes ferroelectric above 45 GPa,71,73 while
BaTiO3 is predicted to be ferroelectric above 140 GPa.128 Both
effects are assigned to a new mechanism involving the hybridi-
zation between Ti 3d and semicore O 2s states.73
The d0 cations show different tendency to polar distortions.
While Ti4 and Nb5 are likely to displace and form polar
structures, larger cations (Zr4, Hf 4, and Ta5) prefer to stay
in the center of the oxygen octahedron. For example, both
BaZrO3 (t 1.003) and KTaO3 are cubic and paraelectric
down to lowest temperatures, as contrasted with ferroelectric
BaTiO3 and KNbO3. Although they feature soft modes that
could trigger the distortion, such instabilities are removed by
quantum fluctuations rendering BaZrO3 and KTaO3 quantum
paraelectrics or so-called incipient ferroelectrics, similar to
SrTi16O3.129,130
The weak tendency of Zr4 and Ta5 to polar displace-
ments is attributed to the larger cation size and electronic
effects. Zr4 is notably larger than Ti4 and gains less energy
by forming a short bond with oxygen. The ionic radii of Nb5
and Ta5 are very similar, yet their 4d and 5d states, respec-
tively, have different energies compared to the oxygen 2p
bands. The larger energy difference prevents Ta5 from forming
covalent bonds with oxygen, thus eliminating polar distortions
in tantalates.56,57,131 Similar to the lone-pair distortions, one
might expect the influence of ligand, that is, the enhancement
of polar distortions in sulfides because of the higher energy of
the S 3p states. However, BaZrS3 (t 0.948) and SrZrS3
(t 0.899), rare examples of perovskite-type sulfides, adopt
the GdFeO3-type structure with nearly nondisplaced Zr
atoms.132134
Empirical analysis of more than 750 structures with the d0
transition-metal cations in octahedral coordination results in
the following trend for the magnitude of the off-center
displacements81:
Mo6  V 5 >> W6  Ti4  Nb5 > Ta5 >> Zr4 polar displacements, while the B cations remain only weakly
 Hf 4
This trend is attributed to the reduction in the effective
electronegativities of these cations.135
Solid solutions based on the d0 B cations are among the
most studied perovskite-type materials. The tendency of Ti4
and Nb5 to the polar displacement is so strong that the
polarity sustains the dilution of the B site with nonpolar
cations. The resulting solid solutions are known as ferroelectric
relaxors. In contrast to conventional ferroelectrics, the relaxors
do not undergo an abrupt ferroelectric transition, but gradually
develop electric polarization, as evidenced by the broad max-
imum in the temperature dependence of their permittivity.
These systems also show appealing piezoelectric properties
with a sizable elastic response to the applied electric field.136
The properties of ferroelectric relaxors stem from their pe-
culiar microstructure driven by dissimilar coordination prefer-
ences of the two B cations. While the polar cation is displaced
from the center of the oxygen octahedron, the nonpolar cat-
ion stays in the symmetric environment and prevents the sym-
metry reduction in the bulk, thus creating a strong tendency for
phase separation. For example, the PbZrxTi1xO3 (PZT) series
demonstrates two ferroelectric phases, P4mm (PbTiO3-like) at
small x and R3c (BaTiO3-like) at large x, separated by a so-
called morphotropic phase boundary around x 0.48. The
compositions close to the phase boundary reveal relaxor prop-
erties, and form an intermediate, monoclinic phase (possibly,
several different monoclinic phases) that generally coexists
with one of the simpler ferroelectric phases.136142
Solid solutions based on two B cations with different oxi-
dation states and the lone-pair Pb2 cation in the A position
are another large group of ferroelectric relaxors. For example,
in PbMg1/3Nb2/3O3 (PMN) and PbZn1/3Nb2/3O3 (PZN) the
ferroelectric displacements of Nb5 trigger the formation of
polar nanoregions that grow in size and eventually induce
the bulk ferroelectric response once the percolation threshold
is reached. However, the emergence of ferroelectricity is not the
bulk structural transition but rather a subtle effect of the
microstructure.143145 To make things even more complicated,
PMN and PZN are often mixed with PbTiO3 (PT), in order to
adjust the dielectric properties. Such PMNPT or PZNPT
compositions are excessively complex but still amenable to a
microscopic study, and specific effects related to the ferroelec-
tric phonon modes have been discovered.136
Finally, the polar d0 cation can be mixed with nonpolar
magnetic atoms to produce magnetic ferroelectric relaxors. This
approach could be promising to obtain new multiferroics,
because the coexisting electric and magnetic orders are induced
by the different cations on the B site, while ferroelectricity can
be reinforced by the lone-pair Pb2 in the A position. The
resulting solid solutions, such as Pb(Fe1/2Nb1/2)O3 (PFN)
and Pb(Fe1/2Ta1/2)O3 (PFT), indeed show interesting magne-
toelectric effects, although they do not resolve the ultimate
problem of finding a ferroelectric ferromagnet with the strong
coupling between the magnetic and electric orders.146,147
The dilution on the B site impairs the magnetism and creates
strong tendency toward spin freezing.148150 In contrast to the
magnetism, the ferroelectricity is additionally supported by the A
cation and, consequently, more robust. However, it is basically
the lone-pair effect of Pb2 that creates electric polarization and
displaced.151 Therefore, in magnetic relaxors both polar and

nonpolar d0 cations show similar behavior: for example, Pb
(Fe1/2Nb1/2)O3 strongly resembles Pb(Fe1/2Ta1/2)O3, although
KNbO3 and KTaO3 are notably different.
The lone pairs on the A cations have an indirect effect on
the stability of the B-cation ordering. Ab initio calculations
demonstrate that the A(B1/3B0 2/3)O3 (B Zn, Mg, B0 Nb, Ta)
solid solutions are more prone to disorder for A Pb than for
A Ba.152 This effect can be explained by the additional PbO
bonding (see Section 2.01.3) that stabilizes the otherwise
underbonded oxygen atoms in the BOB0 links and favors
the homogeneous distribution of the B and B0 cations. This
effect is impossible for A Ba, because Ba2 adopts the sym-
metric 12-fold coordination.
2.01.4.2 JahnTeller Distortions and Orbital Ordering
In molecules, any degeneracy of potential energy surfaces leads
to a geometrical distortion known as the JahnTeller effect.
Similar distortions take place in high-symmetry crystal struc-
tures featuring transition-metal cations with orbitally degenerate
electronic configurations. However, two important differences
between molecules and crystals should be emphasized. First, in
crystals the polyhedra of transition-metal cations are connected,
and the distortions are always cooperative.153,154 While local
distortions can still be understood as the JahnTeller effect for
the single transition-metal cation, the structural transformation
in a crystal is more precisely termed orbital ordering, because
different transition-metal sites may have either same (ferro-type
order) or different (antiferro-type order) orbital configurations.
The second distinctive feature of the JahnTeller effect in
crystals is the competition between the orbitally ordered insu-
lating state and the itinerant (metallic) state that usually pre-
serves the orbital degeneracy and lacks any structural
distortions. Therefore, the JahnTeller distortions are only ob-
served in insulating systems, while metallic perovskites will
usually retain the high symmetry even if the orbital degeneracy
is present.
The orbital-ordering transitions in perovskites conform to
the following trends:
(i) The formation of the orbitally ordered or orbitally degen-
erate (metallic) ground state depends on the electronic
configuration of the transition-metal ion as well as on the
specific electronic structure of the compound in general.
(ii) Ordered orbital patterns (JahnTeller distortions) are sta-
bilized by the gain in the magnetic energy and by the extra
lattice energy (formation of covalent bonds).
(iii) Numerous consecutive orbital transitions accompanied
by the respective structural changes are possible in the
same compound.
The preference to the insulating orbitally ordered state or
the metallic orbitally degenerate state is basically determined
by the transition-metal ion and its oxidation state and not by
electronic configuration alone. For example, Ti3 (3d1) is
prone to the orbital-ordering transitions, while V 4 (3d1)
forms metallic perovskites unless the orbital ordering is stabi-
lized by the lone-pair distortion, as in PbVO3.31,32 The same
vanadium atom in the 3 oxidation state (3d2) demonstrates
an intricate orbital physics, but Cr4 (also 3d2) still forms
metallic perovskites. However, the Cr2 perovskites (3d4) are
Transition-Metal Perovskites 9
again likely to undergo the orbital ordering instead of being
metallic. Although there is no general rule for specific orbital
configurations, each oxidation state of each transition metal
can be considered as prone to orbital ordering or resistant to
orbital ordering in a rather general fashion.
The orbital-ordering transitions are intimately coupled to
the magnetism. In insulators, magnetic interactions are usually
of the superexchange type and depend on the specific orbitals
that are amenable to overlaps and hybridization. Therefore, the
orbital state is of crucial importance for the magnetism.
Whether or not this relationship is mutual has long remained
a matter of debate. An early theoretical concept claimed that
the specific orbital pattern is stabilized solely by the gain in the
magnetic energy; hence, the orbital ordering is essentially a
magnetic phenomenon.155 Therefore, the temperatures of the
orbital and magnetic transitions should be comparable or even
coincide, which is not true for many of the experimentally
studied systems, including perovskites. As the temperature of
the orbital-ordering transition exceeds the available magnetic
energy, further aspects have to be considered. In this case, the
covalency, which is basically the formation of short bonds
featuring strong metalligand hybridization, is an additional
contribution that favors the orbital ordering. The relevant
mechanism largely depends on the specific electronic configu-
ration: in systems with partially filled eg orbitals, the covalency
is of crucial importance, because the eg states strongly hybridize
with the ligand states. By contrast, the orbital-ordering transi-
tions in systems with the partially filled t2g orbitals demon-
strate a stronger coupling to the magnetism and weaker
contribution of covalency.156,157
In perovskite compounds, the distortion of the octahedral
local environment can follow two symmetry modes (Figure 8):
the Q2 mode implies that two opposite ligands shift toward the
metal, two opposite ligands shift away from the metal, and the
remaining two ligands are intact.154,158 The Q3 mode entails
four ligands shifting in and two ligands shifting out, or the other
way around. In simple ABX3 compounds, the Q2 modes are
generally observed.154 Each of the modes (or their combination
in a low-symmetry structure) stabilizes a specific orbital occu-
pation on the transition-metal site. To distinguish between dif-
ferent patterns, the notation originally developed for the
antiferromagnetic structures on the cubic lattice is used.
Thus, the A-type orbital order involves the same orbitals in the
plane and different orbitals in the adjacent planes. The C-type
orbital order is exactly the opposite (alternating orbitals in
the plane and the same orbitals along the interplane direction),
X X
B B
Q2 mode
Q3 mode
Figure 8 Q2 and Q3 normal modes of JahnTeller-driven distortions of
the BX6 octahedra.
10 Transition-Metal Perovskites
while the G-type order means that the orbital configurations are
alternating along all three directions. In perovskites, the C- and
G-types of the orbital order are most common and sometimes
also termed d- and a-types following an alternative notation.
However, the same notation for the magnetic and orbital orders
is preferable, because it facilitates the comparison between these
phenomena. Since GoodenoughKanamori rules for the mag-
netic exchange predict antiferromagnetic/ferromagnetic ex-
change for orbitals of the same/different symmetry, the type of
the orbital order is generally different from the magnetic pattern
(e.g., C-type orbital order may impose antiferromagnetic order
of either G-type or A-type, but not of C-type).
The magnitude of the JahnTeller distortion can be quanti-
fied by the octahedral distortion parameter Dd defined as
X observed for JahnTeller inactive B cations, although the mag-
Dd 1=6 dn  d=d2
n16
where dn is an individual BX bond length and d is an average
BX bond length.159 The cooperative behavior of the Jahn
Teller distortions in the perovskite framework can be intui-
tively understood from BVS arguments.16 Assume, for example,
that the Q2 distortion is confined to the xy plane. In the
ferrodistortive pattern, all short bonds are aligned with x,
while all long bonds are directed along y (Figure 9, left).154
Alternatively, an antiferrodistortive pattern can be formed,
with the BX6 octahedra rotated by 90 around z upon a trans-
lation along <100 >p, so that the short and long bonds switch
their directions from x to y and vice versa (Figure 9, right). In
the first pattern (Figure 9, left), half of the X anions (X1) are
linked to the B cations by two short BX distances, whereas
another half (X2) forms two long BX bonds. For the X1 type
of anions, BVS will exceed their formal valence due to two
short BX bonds, so these anions will be overbonded. Con-
trarily, the underbonding will be observed for the X2 anions
that form two long BX bonds. The antiferrodistortive pattern
in the right panel of Figure 9 should be much more favorable,
because each X atom is involved into one short and one long
BX bond, and the formal valence in maintained. The
antiferrodistortive patterns are indeed most typical for
transition-metal perovskites with a single type of the B cation,
as will be shown below.
x
X1
y axes Z and R defined in a Cartesian coordinate system XYZ,
X2 X2
X1
Ferrodistortive Antiferrodistortive
Figure 9 Sketch of the ferrodistortive (left) and antiferrodistortive
(right) orbital orders in the perovskite framework. The ferrodistortive
pattern involves the overbonded X1 anions along with the underbonded
X2 anions, whereas the antiferrodistortive pattern results in the uniform
and optimal anion bonding.
Based on a similar BVS argument, one expects that the
cooperative Q3 tetragonal distortion of the BX6 octahedra
with long BX distances along the same axis will be uncom-
mon for the transition-metal perovskites with the single type of
the B cation and X anion. The only example of such tetrago-
nally deformed structure (P4/mmm space symmetry) is
La0.5Ba0.5CoO3, where the JahnTeller effect occurs below
T  180 K due to the presumed intermediate-spin t2g5eg1 elec-
tronic configuration of Co3 (see also Section 2.01.4.4).160 The
octahedral distortion parameter in La0.5Ba0.5CoO3
Dd 5.34  107 is several orders of magnitude smaller than
Dd typically observed for the Q2 mode (e.g., Dd 3.31  103
for LaMnO3 at room temperature161).
The local distortions of the BO6 octahedra can also be
nitude of such distortion is rather small: for example, compare
Dd 2.15  106 for LaFeO3 162 with 3.31  103 for
LaMnO3.161 The amplitudes of the Q2 and Q3 deformation
modes vary systematically with the ionic radii of the A cations,
as demonstrated for RFeO3 with a rare-earth R3.163165 The
distortion reaches its maximum at an intermediate size of
R Gd mainly due to the nonlinear variation in the Q3
mode. This effect indicates the intrinsic coupling between the
octahedral tilting distortion and the deformations of the BO6
octahedra.
The JahnTeller distortions can also be considered in the
framework of symmetry transformations. The C- and G-types
of the orbital order are associated with M
2 and R3 irreps of
the Pm3m space group, respectively. Combining these irreps
166
with the M
3 and R4 irreps responsible for the octahedral tilting
distortion, one defines the set of isotropy subgroups for perov-
skites with the JahnTeller deformation superimposed onto
the octahedral tilts. The list of space groups, perovskite
subcell-to-supercell transformation matrices, and the atomic
positions in the supercell for the d- and a-type structures in
the 15 tilt systems is provided by Carpenter and Howard.165
The list for the rare Q3-type structures can be found also in Ref.
166 and generated with the programs ISOTROPY and
ISODISTORT.167
The quantitative treatment of the ABO3 perovskites with a
combination of the octahedral deformation and tilt was pro-
posed by Tamazyan et al.168 Their approach enables one to
evaluate the tilt angles of the deformed octahedra (up to the

1 point symmetry) from crystallographic data. The octahedral
tilt system is described by two angles and y with respect to the
where R lies in the XY plane and d measures the angle between
the R and Y axes. The tilt system notations are given as
(y,,d,G0), where the or  signs correspond to in-phase
and out-of-phase tilts, respectively, and G0 is a point symmetry
group of the distorted BX6 octahedron. Although such nota-
tions are quite useful for the analysis of groupsubgroup re-
lations, they are not as easy for intuitive understanding as the
conventional Glazernotations.
Among the specific orbitally ordered perovskites, we first
consider the eg systems featuring a single (although by far not
simple) orbital-ordering transition. The model compound of
this type is LaMnO3 with the d4 Mn3 cation. Three out of four
d electrons occupy the t2g states, while the fourth electron can
take either of the eg orbitals in the orbitally degenerate

octahedral local environment. The orbital degeneracy is lifted
by the Q2 mode that increases two MnO distances, decreases
two other MnO distances, and consequently stabilizes the
d3z2 r2 orbital state. However, the orientation of the local z-axis
varies from one Mn atom to another, thereby creating the
orbital order of C-type and the antiferromagnetic order of A-
type (Figure 10).169,170
The orbital ordering in LaMnO3 (t 0.942) coexists with a
conventional GdFeO3-type tilting distortion that is driven by
size effects and largely independent from electronic effects on
the Mn site. The isosymmetric transition to the orbitally or-
dered state around 750 K does not alter the orthorhombic
symmetry of the structure, as imposed by the tilting distortion.
The tilting distortion is removed at higher temperatures only,
upon the orthorhombic-to-rhombohedral transition at
1050 K.171 The orbitally disordered orthorhombic phase be-
tween 750 and 1050 K is rather unusual, because it lacks the
long-range orbital order, while the local JahnTeller distor-
tions of the MnO6 octahedra are preserved.161,171,172 There-
fore, the 750 K transition is a melting of the orbital order. The
nature of this transition is illustrated by the volume collapse,
which is reminiscent of the volume collapse upon melting
of ice.173
The robust nature of the local JahnTeller distortions
clearly identifies their nonmagnetic origin. Indeed, the avail-
able magnetic energy is rather low, with the Neel temperature
of 140 K, five times smaller than the orbital-ordering temper-
ature of 750 K. The orbital ordering in LaMnO3 is largely
stabilized by the covalency (more generally, lattice effects)
and prescribes the magnetism.174 However, the order parame-
ter for the orbital transition, as measured by resonant x-ray
scattering, reveals an additional effect at the magnetic transi-
tion and confirms the mutual influence of the spin and orbital
degrees of freedom.169,175
The pressure evolution of LaMnO3 strongly resembles the
transformations of this material upon heating. The long-range
orbital order is likely destroyed around 7 GPa, but a large
fraction of the distorted MnO6 octahedra survives up to
32 GPa, and prevents LaMnO3 from becoming a metal.176180
This further confirms the strong preference Mn3 has in lifting
the orbital degeneracy and stabilizing the JahnTeller distor-
tion by the lattice effects (covalency) rather than by magnetism
alone.
Other Mn3 perovskites are rather similar to LaMnO3 with
respect to the orbital ordering, although the smaller size of
rare earths additionally stabilizes the GdFeO3-type tilting
Figure 10 Left panel: orbitally ordered structure of LaMnO3. Thin orange lines denote longer MnO bonds (local z axis). Right panel: lifting of the
orbital degeneracy for the d4 electronic configuration upon the JahnTeller distortion.
Transition-Metal Perovskites 11
distortion and the orbital order.159,181183 For example,
PrMnO3 (t 0.930) reveals the orbital ordering at 1050 K, in
contrast to 750 K in LaMnO3.184 The compounds with other
rare earths are orbitally ordered in the whole temperature range
studied, while even smaller trivalent cations (e.g., R Y, Lu)
eventually destabilize the perovskite framework, which can be
formed by soft-chemistry methods185 or under high pressures
only.186190 BiMnO3 is, however, different from the rest of the
RMnO3 family because of the larger size and lone-pair-driven
sterical activity of Bi3 (see Section 2.01.3). In BiMnO3, the
lone-pair distortion sets in at 770 K upon the orthorhombic-
to-monoclinic transition, while the subsequent monoclinic-
to-monoclinic transformation at 474 K is identified as the
onset of G-type orbital ordering that prescribes ferromagnetic
behavior of the compound.44
Another example of the orbitally ordered eg perovskite is
KCuF3, where Cu2 (3d9) encounters the same problem of
arranging an odd number of electrons on two eg orbitals.
Similar to LaMnO3, the orbital degeneracy is lifted by the Q2
mode, but the relevant half-filled orbital is now of dx2 y2 type
and lies in the plane with shorter CuF bonds, in contrast to
the d3z2 r2 orbital along the longest MnO bonds in LaMnO3
with only one eg electron. The half-filled orbitals on different
sites are colloquially termed dx2 z2 or dy2 z2 , because CuF
bonds along the z direction are always short, while the second
direction of the short CuF bonds is variable (Figure 11).
Moreover, the orbital order along the z direction (crystallo-
graphic c-axis) is not well defined, and polymorphs with either
C-type (so-called d-polymorph) or G-type (a-polymorph) or-
bital patterns can be formed.191,192 Both polymorphs are te-
tragonal and demonstrate A-type antiferromagnetic ordering
with distinct Neel temperatures of 22 and 38 K, respectively.193
According to the ionic radii, KCuF3 (t 1.01) should lack
any tilting distortions and represent the cleanest example of
the orbital-ordering transition in perovskites. However, the
formation of different polymorphs strongly complicates the
microstructure and results in abundant stacking faults.191
This leads to controversy regarding the possible orthorhombic
distortion, either static194 or dynamic.195,196 The cooperative
JahnTeller distortion is also observed in RbCuF3197 as well as
in the triclinic NaCuF3 (t 0.918)197 and AgCuF3 (t 0.961)198
that show additional tilting distortions due to the size mis-
match. However, the polymorphism induced by the different
orbital-ordering patterns has not been reported.
Similar to LaMnO3, the orbital ordering in KCuF3 is largely
driven by the effects of the CuF covalency,199 with the orbital-
dx2 y2
d3z2 r 2
Regular Distorted
octahedron octahedron
12 Transition-Metal Perovskites

d-Polymorph a-Polymorph
(C-type orbital order) (G-type orbital order)
Figure 11 Variable orbital ordering in KCuF3. Thin orange lines denote long CuF bonds. The d- and a-polymorphs represent different stacking of
ordered layers along the z axis.

ordering temperature of 750 K well exceeding the magnetic

transition temperatures (22 or 38 K) and even the energy
scale of the leading magnetic coupling (about 350 K).200 How-
ever, the magnetic transition seems to affect the orbital
order,201,202 which strongly reminds us of similar observations
for LaMnO3,169 yet remaining a poorly understood effect and a
matter of long-standing debate regarding an additional low-
temperature structural transformation that possibly accom-
panies the magnetic transition. Interestingly, the scenario of
the orbital ordering in KCuF3 is fully reproduced in the 4d
perovskite KAgF3,203205 where the enhanced mixing between
the metal d and ligand p states would otherwise stabilize the
metallic state that eliminates any orbital transitions in most 4d
and 5d perovskites.
The LaMnO3 and KCuF3 families feature the single orbital- Tetragonal polymorph Monoclinic polymorph
ordering transition. However, subsequent orbital transitions in Figure 12 Different patterns of orbital order in KCrF3. Thin orange lines
the eg systems are also possible, as demonstrated by KCrF3 with denote longer CrF bonds. The tetragonal polymorph (left panel) features
the Cr2 (3d4) B cation. At first glance, this compound should G-type orbital order. The monoclinic polymorph (right panel) reveals two
be very similar to LaMnO3 that entails Mn 3 with the same 3d4 nonequivalent Cr sites with different local distortions.
electronic configuration. The high orbital-ordering tempera-
ture of 973 K (compared to 750 K in LaMnO3) and the low
temperature of antiferromagnetic ordering (82 K), as well as
the volume collapse and the Q2-mode local distortion, resem- The RVO3 oxides reveal intricate orbital physics for most of
ble LaMnO3 indeed.206 However, the local environment of Cr the R3 cations (R LaLu, Y, and Sc). Here, V3 with the
atoms is different and features four long as well as two short electronic configuration 3d2 can show different arrangements
CrF bonds instead of four short and two long MnO bonds of two electrons over three t2g orbitals.164,209 The high-
in LaMnO3 (Figure 12). The orbital order is of G-type (tetrag- temperature polymorphs are orthorhombic because of the
onal polymorph) down to 250 K, where a peculiar tetragonal- conventional GdFeO3-type distortion. In LaVO3, the C-type
to-monoclinic transformation accompanied by a change in antiferromagnetic ordering around 140 K slightly precedes
the orbital-ordering pattern takes place.207,208 The low- the orthorhombic-to-monoclinic transition (Tt), which is ob-
temperature orbital state features two nonequivalent Cr posi- served two degrees below the magnetic transition (TN) and
tions with different local environment. The origin of this corresponds to the G-type orbital ordering.210,211 In LaVO3,
additional structural and orbital transformation has not been the tiny separation between the orbital and magnetic transi-
elucidated. tions is largely coincidental. This separation is increased for
The orbital transitions in t2g perovskites show a stronger smaller rare-earth cations: for example, in CeVO3 TN and Tt
coupling to the magnetic order, because the t2g orbitals only differ by 12 K.212
weakly overlap with the ligand orbitals; hence, the gain in the In RVO3 compounds with smaller R3 cations, the low-
lattice energy upon the JahnTeller distortion is small, and the temperature behavior is drastically changed. For example,
orbital order is largely stabilized by the magnetism. The lattice YVO3 first transforms into the monoclinic phase with the
distortions are much weaker than in the eg systems. G-type orbital order (below 200 K) and undergoes the C-type

Yb Y Dy La
RVO3
Temperature
G-type OO
C-type SO
G-type SO
G-type OO
C-type OO
Ionic radius (rR3+)
Figure 13 Sketches of phase diagrams for the RVO3 (left) and RTiO3 (right) perovskite families (OO orbital order; SO spin order).
antiferromagnetic ordering (at 116 K), similar to LaVO3, while
further cooling leads to the transition from G-type to C-type
orbital order and, consequently, from C-type to G-type mag-
netic order at 77 K (Figure 13, left). Therefore, the transition
between different orbital states is accompanied by a change in
the magnetic structure, thus demonstrating the mutual influence
of the orbital and magnetic degrees of freedom.213215 The low-
temperature structure of YVO3 is orthorhombic. This way, the
crystallographic symmetry is first reduced from orthorhombic to
monoclinic around 200 K and further increased back to ortho-
rhombic at 77 K. The behavior of DyVO3 is even more perplexing
because of the third spin and orbital transition that eventually
leads to the G ! C ! G sequence for the orbital order, C ! G ! C
sequence for the magnetic order, and orthorhombic ! mono-
clinic! orthorhombic ! monoclinic trend for the evolution of
symmetry upon cooling (Figure 13, left).216
LaVO3 and YVO3 represent two different orbital scenarios
for larger and smaller R3 cations, respectively.164,215,217 The
rare earths with intermediate size may either show three con-
secutive transitions, as in DyVO3, or separate into the LaVO3-
like and YVO3-like phases. For example, below 130 K SmVO3
develops a mixture of the orthorhombic phase (orbital C-type
and magnetic G-type) and monoclinic phase (orbital G-type
and magnetic C-type).218,219 In YbVO3,220 LuVO3,221 and
ScVO3 (high-pressure modification)222 with the smallest R3
cations still forming the perovskite structure, the C-type orbital
order seems to prevail, whereas magnetic structures become
noncollinear.
The variable structural and magnetic behavior of the RVO3
perovskites opens abroad and meagerly explored field for mi-
croscopic interpretation. Intuitively, the observations correlate
with the ionic radius that controls the magnitude of the tilting
distortion and, consequently, VOV angles as well as the
bandwidth.223,224 However, a crucial ingredient of the micro-
scopic picture is the presence of local JahnTeller distortions
(slightly asymmetric VO6 octahedra) at all temperatures, even
above the orbital-ordering transitions.225 In contrast to the eg
systems, such local distortions are weak and only slightly favor
the specific orbital occupation. Therefore, above 150200 K
thermal fluctuations are sufficiently strong to induce equal
occupations of all three t2g orbitals (this phenomenon is gen-
erally termed orbital fluctuations). The low-temperature or-
bital ordering is intimately related to the magnetism, and
may even appear as a consequence of the magnetic order that
sets in prior to the structural transformation, as in LaVO3 or
Transition-Metal Perovskites 13
Yb Y Gd Sm La
RTiO3
Metal
Temperature
Insulator
G-type SO,
Ferromagnetic
insulator insulator
Ionic radius (rR3+)
LuVO3. The fragile nature of the orbital order in RVO3 is
corroborated by its strong sensitivity to pressure and external
magnetic field.216,226,227
The subtle effects of orbital order are also observed in the
family of RTiO3 perovskites (R3 LaLu, Y), where the inter-
play between orbitals and magnetism is additionally compli-
cated by the proximity of the metallic state.164,228 In contrast to
RVO3, the titanates reveal single orbital transitions that coin-
cide with the magnetic transitions, yet involve only weak
structural changes. Additionally, the orbital ordering may be
accompanied by a change in the conductivity regime from the
metallic behavior in the high-temperature phase to the insu-
lating behavior at low temperatures. In LaTiO3, the orbitally
disordered phase is metallic, while the orbital ordering at
132 K is concomitant with a magnetic and metalinsulator
transition to the G-type antiferromagnetic insulator.229233
YTiO3 is insulating both above and below the orbital-ordering
and ferromagnetic transition at 27 K.234236 These two distinct
regimes are systematically observed in the RTiO3 perovskites
with the large (LaSm) and small (GdYb) R3 cations, re-
spectively (Figure 13, right).164,228,237
The microscopic description of the RVO3 and RTiO3 perov-
skite families is remarkably similar, although early studies
extensively speculated on the lack of orbital order and the
peculiar orbital-liquid ground state of titanates.234,238,239
However, later experimental work confirmed the formation
of the orbitally ordered ground states in all RTiO3 compounds,
and even identified the orbital configurations throughout the
whole family.13,236 While the magnetism itself is induced by
this orbital ordering, the variable magnetic structures strongly
depend on the bandwidth, which is in turn determined by
the R3 cation size and the magnitude of the GdFeO3-type
distortion.12,13,240 The overall scenario is, therefore, strongly
reminiscent of the RVO3 family. An unusual feature of titanates
is the possibility to observe distinct orbital excitations related to
the changes in the orbital configuration of the Ti atoms.241,242
Orbital instabilities are typical for perovskites based on 3d
transition metals. In 4d and 5d compounds, the large spatial
extension of the d orbitals strongly favors electron delocaliza-
tion and results in the abundance of metallic states. Structural
transformations may still be present, but originate solely from
geometrical effects. For example, the SrMO3 compounds with
M Mo4, Tc4, Ru4, and Rh4 reveal the Pnma  Imma  I4/
mcm  Pm3m sequence of structural phase transitions upon
heating (the GdFeO3-type Pnma phase is, however, missing in
14 Transition-Metal Perovskites
SrMoO3).243246 The electronic properties of these compounds
are quite diverse. SrMoO3 and SrRhO3 are simple paramagnetic
metals, whereas SrRuO3 is a ferromagnetic metal and SrTcO3 is
an antiferromagnetic insulator with remarkably high Neel tem-
perature.247250
Electronic instabilities in 4d and 5d perovskites are only
possible for systems with a single half-filled orbital, such as the
aforementioned Ag2 fluorides, or in systems featuring an
additional instability competing with the cubic/metallic state.
For example, PbRuO3 is a conventional metallic perovskite
above 90 K and a paramagnetic semiconductor below this
temperature. In contrast to SrRuO3, the transformation pro-
ceeds from the Pnma space group (GdFeO3-type distortion) at
high temperatures to its supergroup (Imma) at low tempera-
tures. Therefore, the conventional groupsubgroup relation is
violated, because the transition has electronic origin (metal-to-
insulator) and coexists with the lone-pair-driven distortion due
to Pb2.251,252 The role of the lone-pair cation is illustrated by
CaRuO3. This compound remains metallic down to the lowest
temperatures, despite the smaller size of Ca2 and larger or-
thorhombic distortion/smaller bandwidth.253
2.01.4.3 Charge Instabilities
The orbital-ordering transitions considered in the previous
section are one of the routes to stabilize insulating magneti-
cally ordered phases in transition-metal perovskites. More gen-
erally, the competition between the metallic and insulating
phases in a system with partially filled d bands is determined
by the MottHubbard mechanism. This mechanism considers
the interplay between electron transfers from one transition-
metal site to another, and the strong Coulomb repulsion
experienced by the d electrons on each site. The transfers are
quantified by hopping parameters (t) related to the bandwidth,
while the strong repulsion (generally termed strong electronic
correlations) is represented by an effective potential U. At
t U, the electrons prefer to be localized (integer occupation
at each site), and a magnetic insulator is formed. By contrast,
t
U implies the metallic regime. In the ABX3 perovskites, the
largest hopping parameters are observed for BXB angles
close to 180 . The lower angles reduce the hoppings and sup-
press electron transfers. Therefore, cubic perovskites are more
likely to be metallic, while tilting distortions will rather favor
the insulating regime.254
The Hubbard model involves only transition-metal d states
and basically neglects the ligands, or considers them implicitly
as small contributions to the d states. Although the latter
approximation holds for many systems, the ligand p states
may be instrumental in forming the low-temperautre insulat-
ing state of a transition-metal compound. The energy required
for transferring an electron from the p orbital of the ligand to
the d orbital of the transition metal is known as the charge-
transfer energy (or charge-transfer gap) D. This energy corre-
lates with the metalligand hybridization; the stronger the
hybridization, the smaller the D. If D < U, the lowest-lying
electronic excitations are related to the pd charge transfer
instead of the dd excitations, so that D instead of U becomes
the relevant parameter competing with the electron hoppings t.
This D < U regime, also known as the charge-transfer regime as
opposed to the Mott regime at D > U, is relevant for the
transition metals in high oxidation states and even for lower
oxidation states of late 3d elements, such as Cu and Ni.
Most electronic instabilities originate from the general ten-
dency of transition-metal compounds to form insulating states
rather than remain metallic. The orbital instabilities related to
the orbital-ordering transitions are typical for Mott insulators
(D > U). The charge-transfer systems (D < U) are prone to
charge instabilities that involve a partial transfer of charge
between different transition-metal sites.255 Because the charge
transfer is mediated by ligand orbitals, the small charge-
transfer gap is an essential prerequisite of this mechanism.
The charge transfer is typical for systems with noninteger
band filling (i.e., fractional oxidation state of the transition
metal), although materials with integer filling (integer oxida-
tion state) also demonstrate several examples of charge-
ordering transitions. Since the charge transfer leads to different
oxidation states of the transition-metal atoms, distinct local
environments and significant structural changes are generally
observed.
In simple ABX3 perovskites with an integer oxidation state
of the transition metal, charge instabilities are observed in two
compound families: AFe4O3 (A2 Ca, Sr, Ba) and RNi3O3
(R3 LaLu, Y, Tl, Bi). The Fe4 (3d4) cation in AFeO3 re-
sembles Mn3 (also 3d4), but the crystal and electronic struc-
tures of the respective compounds are very different. The strong
FeO hybridization renders the partially filled eg band itiner-
ant, while the narrower t2g bands form localized magnetic
moments. This scenario is observed in metallic SrFeO3
(t 1.021), where localized magnetic moments arising from
the t2g electrons interact via the itinerant eg bands and produce
helical magnetic order below 130 K.256258 BaFeO3 (t 1.082)
is also cubic and, presumably, shows similar behavior, al-
though the helical magnetic order is easily flipped toward
ferromagnetism in the applied magnetic field.259 In CaFeO3
(t 0.966), the GdFeO3-type distortion reduces the width of
the eg bands and triggers the metalinsulator transition at
290 K. In contrast to manganites, the orbital degeneracy (one
electron per two eg orbitals) is lifted by a charge instability:
Fe4 disproportionates into partially reduced and partially
oxidized iron cations that can be considered as Fe3 (3d5)
and Fe5 (3d3), both lacking orbital degeneracy.260 The two
Fe sites reveal a notably different local environment, and reflect
the symmetry lowering from orthorhombic to monoclinic
(Figure 14, left).261 The helical magnetic structures in metallic
SrFeO3 and insulating CaFeO3 are surprisingly similar and
even have similar origin.258
The structures and physics of the RNiO3 compounds are
more intricate. The general trend of the reduced metallicity
with the reduction in the cation size is retained: while
LaNiO3 (t 0.946) is rhombohedral and metallic, smaller
R3 cations (PrLu, Y, also Tl3) form orthorhombic
GdFeO3-type structures and trigger charge-ordering/metalin-
sulator transitions.164 The charge-ordering temperature sys-
tematically increases with the reduction in the ionic radius of
R3, and exceeds 500 K for the smallest R cations, such as Y and
Lu. Similar to CaFeO3, the magnetic ordering accompanies the
charge-ordering transition for R Pr and Nd, but for smaller
R3 cations (R GdLu, Y) the magnetic transition takes place
well below the charge-ordering transition (100200 K vs.
400600 K, respectively, see Figure 15).164 The latter systems

show clear structural signatures of the charge ordering, with
monoclinic crystal structures and two distinct Ni sites below
the transition temperature (Figure 15, left).262265 In RNiO3
with R Pr, Nd, and Sm, the formation of nonequivalent Ni
sites is only detected by local techniques, such as Raman scatter-
ing and x-ray absorption spectroscopy,266268 while bulk non-
resonant diffraction methods evidence the nondistorted
orthorhombic crystal structure in the whole temperature
range.262,269 Therefore, the charge ordering in PrNiO3,
NdNiO3, and SmNiO3 leads to very subtle structural changes
and may even be of local rather than bulk nature.
In contrast to CaFeO3, the charge ordering in nickelates is
only partial. The nonequivalent Ni sites have the oxidation
states of Ni3d and Ni3d with d  0.35 in YNiO3,263 although
in this case the very notion of the oxidation state is somewhat
ambiguous because of the strong hybridization between the Ni
3d and O 2p states. The incomplete charge separation can be
formally ascribed to a charge-density-wave instability that is
frequently observed in metals, yet the microscopic origin of the
transitions in nickelates is quite different. While charge-density
waves arise from a specific topology of the Fermi surface (nest-
ing instabilities), charge instabilities in perovskites are related
to the strongly correlated nature of Ni3 and other transition-
metal cations.255 Such transitions do not involve the orbital
ordering, however.270272
CaFeO3 BiNiO3
Figure 14 Examples of checkerboard charge order on the B site
(CaFeO3, left) and A site (BiNiO3, right) of the perovskite structure.
Temperature
Figure 15 Left panel: charge-ordered monoclinic structure of YNiO3 with two nonequivalent Ni sites. Right panel: sketch of the phase diagram showing
the transition from the metallic phase for large rare earths (e.g., LaNiO3) to the insulating charge-ordered phase for small rare earths (e.g., YNiO3).
Transition-Metal Perovskites 15
The ANiO3 nickelates can also be prepared with Tl and Bi as
A cations. Unlike Y and rare earths, Tl and Bi feature a lone-pair
(Tl1, Bi3) as well as s and p valence states interacting with the
3d electrons of Ni. While Tl3Ni3O3 (t 0.887) is rather
similar to SmNiO3,273 BiNiO3 does not show the anticipated
combination of Bi3 and Ni3, and forms Ni2 along with the
mixed-valence Bi3/Bi5 state instead (Figure 14, right).274,275
Therefore, the charge ordering is observed for the A site instead
of the B site in other nickelates. External pressure induces an
insulator-to-metal/charge-transfer transition toward the more
conventional Bi3Ni3O3, which is cubic and metallic.276,277
Mixed-valence systems with fractional oxidation states of
the transition metals are even more prone to charge instabil-
ities. However, the charge ordering will only occur in itinerant
systems with large hopping parameters. For example, the R1/
3Sr2/3FeO3 solid solutions with the average oxidation state of
Fe3.67 reveal the charge ordering concomitant with the metal
insulator transition for R La, Pr, and Nd. The transition tem-
perature decreases upon the reduction in the ionic radius of
R3. The solid solutions based on smaller R3 cations (Sm and
Gd) do not undergo the charge ordering, because even at high
temperatures the respective compounds are insulating.278281
The effect is conventionally explained in terms of the enhanced
GdFeO3-type distortion and reduced bandwidth for smaller
R3 cations. The GdFeO3-type distortion impedes electron
hoppings and eventually leads to electron localization on ran-
dom transition-metal sites, thus eliminating the charge
instability.
Peculiar examples of charge-ordering transitions are found
in mixed-valence manganites that were extensively studied
because of the colossal magnetoresistance (CMR) effect. The
perovskite manganites can probably compete with ferroelectric
relaxors (see Section 2.01.4.1) in terms of both popularity and
complexity. A comprehensive review of CMR perovskites lies
far beyond the scope of the present work and can be found
elsewhere.282,283 Here, we only mention that charge instabil-
ities in manganites are intimately related to the magnetism.
Metallic states are additionally stabilized by the double-
exchange mechanism that results in strong ferromagnetic in-
teractions mediated by the itinerant electrons. Such interac-
tions are impossible in the charge-ordered insulating state
that would, however, favor conventional superexchange
Lu Y Sm Nd La
RNiO3 Paramagnetic
metal
Charge-ordered
insulator
(paramagnetic)
Antiferromagnetic
insulator
Ionic radius (rR3+)
16 Transition-Metal Perovskites
couplings typical for insulators. The charge-ordered states are
generally formed in manganites with Mn3.5, while other oxi-
dation states induce ferromagnetic metallic states or a nano-
scale phase separation toward the ferromagnetic and
antiferromagnetic or spin-glass phases with the itinerant and
localized electrons, respectively.284 Unlike Ni3- or Fe4-based
perovskites, the mixed-valence manganites are also prone to
the orbital ordering that takes place for the orbitally degenerate
Mn3 formed upon the 2Mn3.5 ! Mn3 Mn4 charge-
ordering process.285 A special feature of the charge- and
orbital-ordering transitions in the mixed-valence manganites
is their proximity to the ferromagnetic metallic state that is
stable in high magnetic fields or even in zero field above the
charge- and orbital-ordering transitions.286,287
2.01.4.4 Instabilities Related to the Spin State and
Magnetism
Magnetism is an important ingredient in most of the electronic
instabilities observed in perovskites. The orbital-ordering tran-
sitions are frequently mediated by magnetic interactions (see
Section 2.01.4.2), while the charge ordering is typically
accompanied or followed by the onset of the magnetic order
(see Section 2.01.4.3). However, the role of magnetism may be
even more explicit. For example, Cr4 (3d2) perovskites are
metallic systems that basically avoid both charge and orbital
ordering. This contrasts with the RVO3 compounds showing
peculiar orbital physics of V3, which is also a 3d2 cation. The
metallic nature of ACr4O3 arises from the strong CrO hy-
bridization that increases the electron hoppings (t as compared
to D and U, see Section 2.01.4.3) and undermines the orbital
ordering.
At room temperature, SrCrO3 (t 1.003) is cubic, while
CaCrO3 (t 0.949) shows the conventional GdFeO3-type dis-
tortion. Both perovskites are metallic at room temperature and
remain metallic below the magnetic transitions taking place
around 40 and 90 K for SrCrO3 and CaCrO3, respectively.288290
The transitions are accompanied by structural changes, espe-
cially in SrCrO3 that separates into the cubic and tetragonal
phases, with only the latter being magnetically ordered.290,291
Although the structural changes in ACrO3 are sometimes under-
stood in terms of the orbital ordering,292 a complete orbital
order typical for RVO3 and RMnO3 is clearly impossible, be-
cause the electrons remain largely itinerant.289,290 Since no sig-
natures of the charge ordering are observed either, magnetism is
the only remaining and, therefore, the most likely origin of
structural instabilities. The experimental C-type antiferromag-
netic structure can be understood in terms of both itinerant
(magnetic instability of electronic gas)293 and insulating (local-
ized magnetic moments)292,294 scenarios, while the actual phys-
ics of ACrO3 is believed to be intermediate between the purely
metallic and insulating regimes.290,295
Cobalt oxides form a special group of perovskite-type com-
pounds. The electronic structure of Co3 is basically interme-
diate between TiMn, forming Mott insulators (U < D), and
Fe4 or Ni3, forming charge-transfer insulators (U > D). This
intermediate regime may allow for both orbital and charge
instabilities, but neither of the two is in fact operative in Co-
based perovskites. The structures of cobaltites are mainly de-
termined by the flexibility of the Co3 (3d6) spin state. Six d
Lu Y La
RCoO3
Paramagnetic
metal
Temperature
Paramagnetic
insulator
Nonmagnetic
insulator
(LS-Co)
Ionic radius (rR3+)
Figure 16 Sketch of the phase diagram for RCoO3 perovskites. While
the low-temperature nonmagnetic and insulating phase features low-spin
(LS) Co3, the increase in temperature leads to a spin-state transition
toward the paramagnetic insulating phase and, eventually, the
paramagnetic metallic phase.
electrons may fully occupy t2g orbitals (t2g6), thus leading to
the nonmagnetic low-spin (LS) state, or occupy both t2g and eg
orbitals (t2g4eg2), thus forming the magnetic high-spin (HS)
state. The intermediate t2g5eg1 configuration is known as the
intermediate-spin (IS) state, but its existence is a matter of
heavy and long-standing debate. The variable spin state of
Co3 is best illustrated by rhombohedral LaCoO3. At low
temperatures, this compound is a nonmagnetic band insulator
(possibly, with a weak monoclinic distortion296) because of
the LS Co3 and the complete filling of all t2g bands. Above
100 K, the insulator-to-metal transition and anomalous ther-
mal expansion signify the occupation of the eg levels.297 Dif-
ferent scenarios suggest that either all Co atoms transform into
the IS state,297300 or a fraction of Co atoms attain the
HS configuration (therefore, the mixed LSHS system is
formed).301305
Different spin states of Co3 are expected to show different
ionic radii, with the HS ion being about 0.065 A larger than
its LS counterpart because the eg electrons in the HS ion
experience stronger repulsion from oxygen atoms.297 This
structural feature results in a peculiar thermal expansion in
all R3Co3O3 perovskites and a counterintuitive suppression
of the metallic state under pressure. Indeed, the contraction of
the CoO6 octahedra increases the t2geg crystal-field splitting
and stabilizes the nonmagnetic LS state that results in the
insulating behavior. In LaCoO3, the temperature of the
metalinsulator transition is systematically increased under
pressure.306,307 The A-site substitution leads to a similar effect,
because small R3 cations induce the strong GdFeO3-type or-
thorhombic distortion and reduce the width of eg bands
(Figure 16).308,309
2.01.5 Ordering Phenomena at Cation and Anion
Sublattices of Perovskites
2.01.5.1 B-Site Ordering
Different atomic species can simultaneously occupy the octa-
hedral B sites in perovskites. The B cations of different sizes and
charges are either disordered or form an ordered arrangement

of the B and B0 cations with the idealized stoichiometries
A2BB0 X6 and A3BB0 2X9 (termed as 1:1 and 1:2 ordering). In
the overwhelming majority of the ordered perovskites, the B
and B0 cations are separated into (111) layers alternating along
the [111]p direction, thus leading to an increased translation
along [111]p by a factor of 2 and 3 for the 1:1 and 1:2 ordering,
respectively (Figure 17).
The A2BB0 X6 family, also known as double perovskites
(X O) and elpasolites (X F), embraces the largest number
of the B-site ordered perovskites. The undistorted A2BB0 X6
structure has the a 2ap cubic unit cell with the Fm3m space
symmetry (Figure 18). The B and B0 cations form a structural
motif that follows the arrangement of the Na and Cl atoms in
the rock-salt structure. Such an order is therefore termed the
rock-salt ordering. A large number of the A2BB0 O6 perovskites
were analyzed, in order to understand the influence of the
B-cation ionic radii and formal charges on the relative stability
of the ordered and disordered structures. The charge difference
Dq between the B and B0 cations has a major effect on the
cation ordering.310,311 When Dq > 2, the B and B0 cations are
usually ordered (e.g., Ba2MnWO6312), whereas Dq < 2 leads to
disordered structures (e.g., Nd2FeMnO6313). In the intermedi-
ate case of Dq 2, the size difference DIR plays a role. Gener-
ally, DIR > 0.2 A promotes the rock-salt ordering. At high Dq,
the rock-salt ordering is clearly preferential, because it places
the highly charged B0 cations at maximal distances from each
other, thus diminishing the electrostatic repulsion. In the rock-
salt-ordered structure, the B and B0 cations are placed at trans
BX6
BX6
[111]p
A2BBX6 A2BB2X9
Figure 17 Patterns of the BX6 and B0 X6 octahedra in the B-site ordered
A2BB0 X6 (1:1) and A3BB0 2X9 (1:2) perovskites. The cubic 2ap and
hexagonal ap2  ap2  ap3 supercells are outlined.
BX6
BX6
Figure 18 Left: the crystal structure of the A2BB0 X6 double perovskites. The B0 cation is smaller and has a higher charge; the B cation is larger and has
a smaller charge. Right: displacements of the X anions optimizing BVSs for both the B and B0 cations. The arrows indicate the displacements from
the ideal perovskite positions.
Transition-Metal Perovskites 17
positions with respect to the anions, while the anions are
allowed by symmetry to shift along the BB0 line (Figure 18).
Thus, all anions have similar coordination environment and
the BX and B0 X bonds can be relaxed, optimizing the BVS for
the anions. In the disordered structure, a part of the anions is
surrounded by the cations of the same type. The anions having
only highly charged B cations as nearest neighbors will be
overbonded, while the anions bonded to the B cations with
lower charge will be underbonded.
The strongly simplified picture of the purely ionic bonding
fails to explain all the peculiarities of B-cation ordering in
perovskites. For example, the Sb(V)-containing double perov-
skites demonstrate a strong tendency toward the rock-salt or-
dering, even at Dq 2 and DIR < 0.2 A. In A2BSbO6 (A Ca, Sr,
B Fe, Cr), the B3/Sb5 cations are ordered at least partially in
spite of the small DIR 0.035 A (B Fe) and 0.005 A
(B Cr).314317Mizoguchi et al. attributed this order to the
strong covalency of the Sb5O bonding and a tendency to
avoid the linear Sb5OSb5 chains, where the oxygen atom
forms two strong s-bonds using only one out of three
p orbitals.318 The other two orbitals cannot be involved in
the p-bonding because of the fully occupied d-shell of the
Sb5 cations.
The 1:1 ordering with the B and B0 cations forming separate
(001)p layers is very rare in the A2BB0 O6 perovskites. It is
observed in the compounds where the B cation demonstrates
a pronounced JahnTeller effect (e.g., Cu2) and with Dq 2
between the B and B0 cations (R2CuSnO6 (A La, Pr, Nd, Sm),
La2xSrxCuSnO6, La2CuZrO6)319321 (Figure 19). The Q3-type
JahnTeller distortion (see Section 2.01.4.2) resulting in the
strong apical elongation of the CuO6 octahedra is the necessary
CuO6
SnO6
Figure 19 Layered B-site ordering in the La2CuSnO6 structure.
BX6
BX6
18 Transition-Metal Perovskites

but insufficient prerequisite of the layered ordering. The charge their coordination environment by such displacements.
difference Dq 2 is another important condition, because at They appear to be overbonded because of the links to two
higher Dq the dominating electrostatic repulsion causes the highly charged B0 cations simultaneously. Off-center displace-
rock-salt ordering even for Q3-type deformation of the CuO6 ments of the B0 cations along [111]p are required to increase
octahedra, as observed in A2CuMO6 (A Sr, Ba, M W, the B0 O separation (Figure 20). Such displacements are pro-
Te).322324 On the contrary, Dq < 2 leads to a disordered struc- moted by SOJT if the B0 cation has d0 electronic configuration
ture, where the JahnTeller effect manifests itself as a Q2 mode (see Section 2.01.4.1). This explains why the 1:2 ordering is
(CaRMnSnO6 (R La, Pr, Nd, SmDy) and LaAMnSnO6 most frequently observed for B0 Nb5, Ta5. Except for the
(A Sr, Ba)).325,326 The B0 cation of larger size than the B A32B2B0 25O9 family, examples of the 1:2 ordered perov-
cation allows for a tilting distortion in the B0 O6 octahedral skites have been found also among the A32B6B0 23O9
layer and for the buckling of the (B0 O2) plane. This distortion (Ba3TeBi2O9332) and A33B1B0 24O9 (La3LiTi2O9333) com-
is necessary to shift the oxygen atoms of the (B0 O2) plane pounds. A deviation from the perfect 1:2 B-cation ratio may
toward the A cations. Otherwise, the A cations are under- also result in a partial 1:2 ordering, found in Ba3(Li0.76Nb0.24)
bonded due to an increased separation between the (BO2) (Nb0.63W0.3Li0.07)2O9334 and Sr3(Li0.724Nb0.276)(Nb0.612W0.3
and (B0 O2) planes, as caused by elongation of the apical BO Li0.088)2O9.335 The 1:2 B-cation ordering itself results in a
distance and the shift of the A cations toward negatively trigonal a ap2, c ap3 supercell and the P3m1 space sym-
charged (BO2) planes. metry (Figure 17). Most often, it is coupled with the abb
As in the ABX3 perovskites, the octahedral framework of GdFeO3-type octahedral tilting distortion that decreases sym-
double perovskites and elpasolites can undergo the tilting metry down to P21/c and leads to the a ap6, b ap2,
distortion. The tolerance factor, calculated using the average c 3ap2, b  125 supercell.
ionic radii of the B and A cations (if more than one A cation is The 1:3 ordering in the Ba4BB0 35O12 (B Li, Na, B0 Sb,
present), can also be used for predicting the stability of the Bi) perovskites does not have layered character. Instead, these
AA0 BB0 X6 structure toward the tilting distortion: compounds adopt a body-centered perovskite structure with
a 2ap, where the B cations occupy the positions at the cor-
RX 1 =2 RA RA0 ners and in the center of the unit cell (Figure 21).336,337 Linear
t p
2RX 1 =2 RB RB0 B0 5OB0 5 chains are present in these structures along all
three crystallographic directions. This is unusual for the Bi5-
For the A2BB0 F6 elpasolites, the broad stability range of for and Sb5-containing perovskites. Indeed, the KB5O3 (B Bi,
the undistorted Fm3m structures at 0.88  t 1.0 was Sb) compounds do not crystallize in the perovskite structure,
proposed,327,328 indicating that the fluorides are apparently but adopt a cubic structure consisting of corner-sharing B2O10
more tolerant to the mismatch of the AX and BX bonds
than oxides.
The rock-salt cation ordering is described by the R BX6
1 irrep of
the Pm3m space group, which, in combination with the M 3
and R 4 irreps of the octahedral tilts, produces 12 possible
structures, where the tilting distortion is superimposed on the
rock-salt cation ordering.329 Note that the number of tilt sys- BX6
tems is reduced in comparison with that for the ABX3 perov-
skites, because different tilt systems with the same space group
are reduced to the most general tilt system for this space group.
Similar symmetry analysis has been performed by Aleksandrov
and Misjul.330 It resulted in a larger list of possible structures,
because the in-phase and out-of-phase rotations around the
same axis were simultaneously taken into account. The 12 tilt
systems can be subdivided into two groups, where the A cat- Figure 20 Off-center displacements (marked by arrows) of the B0
ions occupy either single or multiple crystallographic sites.20 cations in the B-site ordered A3BB0 2X9 (1:2) perovskites.
The overwhelming majority of the double perovskites has a
single A site, whereas only a few crystallize with the tilt systems
having multiple A sites stabilized by a drastic difference be-
BX6
tween the A and A0 cations (see Section 2.01.5.3).
The 1:2 B-cation ordering is mostly confined to the
A32B2B0 25O9 family, where A2 Ca, Sr, Ba, Na0.5La0.5,
BX6
B2 Mg, Ca, Sr, Mn, Fe, Co, Ni, Cu, Zn, B0 5 Nb, Ta, Ru
(see Ref. 331 and references therein). As in the 1:1 ordered
double perovskites, the size and charge difference are impor-
tant for the 1:2 ordering. Oxygen atoms bridging the B and B0
(111)p layers are shifted from the larger B cation with a smaller
charge toward the smaller B0 cation with a higher charge, thus
optimizing the BOB0 bond lengths. However, the oxygen
atoms positioned between the two B0 layers cannot optimize Figure 21 The crystal structure of the A4BB0 3X12 (1:3) perovskites.

dimers (two BO6 octahedra sharing common edge) instead.338
The 1:3-ordered cubic structure is not observed for B0 Nb, Ta.
This effect is attributed to a difference in polarizing power of
these cations and that of Sb5 and Bi5 having a noble gas
electronic configuration.
2.01.5.2 Noncooperative Octahedral Tilting in 1:1 Ordered
A2BB0 O6 Perovskites and A2BB0 F6 Elpasolites
The group-theory analysis of the octahedral tilt systems and their
combinations with other types of ordering in perovskites pro-
vides a comprehensive list of possible structures. However, this
analysis is valid only if the BO6 octahedra maintain their regu-
larity, or at least the distortion of the octahedra is much smaller
than the tilt magnitude, and corner-sharing connectivity is
strictly preserved.7 If any of these conditions is violated, new
perovskite structures lying beyond the predictions of the group-
theory analysis emerge. Such violation can occur if a part of the
octahedra do not behave as rigid units. In the 1:1 ordered
A2BB0 O6 double perovskites and A2BB0 F6 elpasolites with a
significant difference between the ionic radii of the B and B0
cations, the octahedra around the small B0 cations are nearly
rigid. However, the B0 X6 octahedra are not connected to each
other and remain separated by the B cations. If the octahedral
coordination is not typical for these large B cations, they readily
increase their coordination number to 7 or 8 due to a rotation of
one or two neighboring B0 X6 octahedra by 45o, which is the
maximum possible angle for the octahedral rotation. The rota-
tion of the B0 X6 octahedra replaces the corner-sharing connec-
tion of the B and B0 polyhedra by the edge-sharing connection,
thus increasing the coordination number of the B cation by 1
per each rotated B0 X6 octahedron (Figure 22). As a result of the
significant transformation of the B-cation polyhedra, the coop-
erative corner-linked topology of the perovskite framework is
lost, and the rotations of the B0 X6 octahedra are not directly
linked to each other. When this occurs, entirely new perovskite
crystal structures emerge. Such structures tend to be much more
complex than those normally adopted by perovskites. The non-
cooperative octahedral tilting is likely to appear if the size dif-
ference between B and B0 cations is as large as DIR  0.75 A for
elpasolites and DIR  0.5 A for double perovskites, while the
tolerance factor is relatively small (t 0.890.93).339 The com-
pounds exhibiting the noncooperative octahedral tilting include
BX6
45 BX7 45 BX8
45
(a) (b)
Figure 22 A formation of the BX7 (a) and BX8 (b) polyhedra as a result
of rotation of one or two B0 X6 octahedra over 45 . Reprinted with
permission from Abakumov, A. M.; King, G.; Laurinavichute, V. K.;
Rozova, M. G.; Woodward, P. M.; Antipov, E. V. Inorg. Chem. 2009, 48,
93369344. Copyright 2012 American Chemical Society.
Transition-Metal Perovskites 19
Rb2KCrF6, Rb2KGaF6, K3AlF6, K3MoO3F3, Sr3WO6, Sr3TeO6,
Ba3TeO6, and A11B4O24 (A Ca, Sr; B Re, Os, where some of
the A sites are vacant).339346
In the A-site-deficient Sr11Re4O24 (Sr1.750.25)SrReO6,
B Sr2, B0 Re6.5, -cation vacancy, half of the small ReO6
octahedra are rotated by 45 relative to their positions in the
ideal double-perovskite structure around either the a or b axis
of the tetragonal unit cell, thus resulting in an increased coor-
dination number of the Sr atoms at the B position from 6 to
8 (Figure 23).345 In the room temperature a-K3AlF6 K2KAlF6
structure, four out of ten of the AlF6 octahedra undergo large
45 rotations (Figure 23). One-fifth of the AlF6 octahedra are
rotated about the c cubic perovskite subcell axis, while 1/10 are
rotated about the a cubic perovskite subcell axis, and 1/10
are rotated about the b cubic perovskite subcell axis. These
rotations increase the coordination numbers of the K cations
on the B sites so that 2/5 of the B sites are seven-coordinate and
3/5 of the B sites are eight-coordinate. This pattern of octahe-
dral tilting results in a structure with Z 80 and a unit cell with
a volume of over 12 000 A3.339 The large octahedral rotations
also affect the coordination of the A cations. A rotation of one
B0 X6 octahedron decreases the coordination number of the
A cation by 1, and shortens one of the remaining 11 AX
distances (Figure 24).
Although there is no direct link between the rotation di-
rections of the B0 X6 octahedra, the well-established ordering of
the octahedral units suggests that an indirect interaction exists.
Assuming only rotations over a fixed 45 angle around one of
the ap, bp, or cp axes, several possibilities can be considered,
resulting in a coordination environment of the B cations with
CN 6, 7, and 8 (Figure 25). The B cation can be surrounded
by six B0 X6 octahedra keeping their orientations in the parent
perovskite structure, with fourfold octahedral symmetry axes
parallel to the cubic axes (Figure 25(a)); two B0 X6 octahedra
above and below the drawing plane are not shown for clarity.
The B cation has an octahedral environment with d(BX) > d
(B0 X). The rotation of a single B0 X6 octahedron over 45
transforms the BX6 octahedron into the BX7 distorted pentag-
onal bipyramid (Figure 25(b)). The increase in the coordina-
tion number up to CN 8 occurs by the rotation of two B0 X6
octahedra in cis positions with respect to the B cation over 45
around two perpendicular axes, which can be either noncopla-
nar (Figure 25(c)) or coplanar (Figure 25(d)). In all three
cases of CN 7 and 8, a displacement of the B cations from
their ideal positions is possible, resulting in an optimization of
the interatomic distances between the B cations and X anions
that belong to the unrotated B0 X6 octahedra. Two more cases
should also be considered, which seem to be energetically less
favorable. The rotation of two cis-positioned octahedra around
the same axis would result in an abnormally short XX separa-
tion and a strong electrostatic repulsion (Figure 25(e)). The
rotation of two octahedra in trans positions (Figure 25(f)) is
not compatible with the displacements of the B cations due to
the formation of centrosymmetric coordination. Shifting the B
cation toward the edge of the rotated octahedron introduces
very long distances to the X atoms of a second rotated octahe-
dron at the trans position and does not significantly increase
short distances to the X atoms of the nonrotated B0 X6 octa-
hedra. The same conclusion is applicable to the rotation of
octahedra in trans positions around two different axes. Thus,
20 Transition-Metal Perovskites

KF8
AlF6

KF7
ReO6

SrO8

Sr11Re4O24 K3AlF6

Figure 23 The crystal structures of Sr11Re4O24 and K3AlF6 with noncooperative tilting distortion of the octahedral framework.

ap
B X BX6
B
0 0
bp
A B

0 0 0 0

(b) c
(a) 0

(a)

0 0
B X
45
0 b 0 b
B A
(c) c (d) a

c
0

(b)
Figure 24 Coordination environment of the A cation in the cubic B-site
ordered A2BB0 X6 structure (a) and its modification by a rotation of one B0 X6
octahedron over 45 (b). 1/8 of the F-centered cubic unit cell is shown.
Reprinted with permission from Abakumov, A. M.; King, G.; Laurinavichute,
V. K.; Rozova, M. G.; Woodward, P. M.; Antipov, E. V. Inorg. Chem. 2009,
48, 93369344. Copyright 2012 American Chemical Society.
the rotation modes of the octahedral units are mediated by a
possibility of relaxation of the BX bonds due to off-center
displacement of the B cations and the anionanion repulsion
between the neighboring B0 X6 octahedra. It should be noted
that BX9 polyhedra can also be constructed by a rotation of
three cis-B0 X6 octahedra around different axes, for example, but
CN 9 has not been observed in such structures yet.
0 c 0 0
(e) c (f) c
Figure 25 Different modes of the B0 X6 octahedra rotations resulting
in CN 6 (a), 7 (b), and 8 (cf) for the B cations. The octahedra above
and below the drawing plane are not shown for clarity. Arrows mark
directions of displacements of the B cations. The double-headed arrow
indicates short XX separation. Reprinted with permission from
Abakumov, A. M.; King, G.; Laurinavichute, V. K.; Rozova, M. G.;
Woodward, P. M.; Antipov, E. V. Inorg. Chem. 2009, 48, 9336 -9344.
Copyright 2012 American Chemical Society.
Upon heating, the amplitude of the octahedral rotations does
not decrease continuously and rather maintains the 45 value.
A step-like decrease in the fraction of the rotated B0 X6 octahedra
results in a series of first-order phase transitions between several

polymorphs. The fraction of rotated octahedra decreases from 2/5
in a-K3AlF6 to 1/5 in b-K3AlF6, 1/6 in g-K3AlF6, and 0 in d-K3AlF6,
which adopts the cubic Fm3m double-perovskite structure. This
discontinuous nature of the phase transitions is drastically differ-
ent from octahedral tilting distortions in conventional perov-
skites, where the corner-sharing links between the octahedral
units force them to behave cooperatively.347
2.01.5.3 A-Site Ordering
A-cation ordering is by far less common than the B-cation
ordering in perovskites. This effect can be attributed to the
smaller charge difference between the A cations entering the
perovskite structure. Indeed, the charge difference for the B
cations can be as high as 6 (alkali cations and Re7 in
A2BReO6, A Sr, Ba, B Li, Na),348,349 whereas it does not
exceed 3 for the A cations (Na2/3Th41/3TiO3).350 In contrast
to the B-site ordering, the examples of the rock-salt type
A-cation ordering along the (111)p perovskite planes are scarce
compared to the ordering along the (001)p planes (layered
ordering). The long-range rock-salt A-cation ordering was
reported for the mixed perovskite fluorides NaBaLiNiF6 and
NaBa1/21/2Fe2F6 (-cation vacancy) only.351,352 The reasons
of the prevailing layered ordering have been discussed in detail
by King, Knapp, and Woodward using bond valence arguments
for the oxygen atoms.311,353 Every oxygen atom in the perov-
skite structure is situated in an octahedron comprising two B
cations and four A cations. Due to the rock-salt ordering of the
A cations, the trans octahedral vertices will be occupied by A
cations of the same type (Figure 26(a)). Thus, the oxygen
atoms retain the centrosymmetric environment and have no
possibility to displace from the highly charged A cation toward
the cation of smaller charge, as required for the optimization of
their bond valences. For the B-site rock-salt ordering, the op-
posite vertices are occupied by different cations making this
relaxation possible. However, the layered ordering only par-
tially improves the ability of the structure to relax. In the
AA0 B2X6 ordered perovskite, the anions of the (BX2) layers are
surrounded by two A and two A0 cations in cis positions and
can be displaced toward the neighboring (A0 X) layer with the A0
cation of higher charge and/or smaller size (Figure 26(b)).
Thus, 2/3 of the anions can optimize their coordination.
A A
A MgO6
A
(a) (b)
Figure 26 Rock salt (a) and layered (b) arrangement of the A and A0
cations in the AA0 B2X6 ordered perovskites.
Transition-Metal Perovskites 21
However, 1/6 of the anions in the (AX) layers and 1/6 of the
anions in the (A0 X) layers are still surrounded by four cations of
the same type (A or A0 , respectively), thus leading to bonding
instability. This instability can be relieved if the layered A-site
ordering is combined with cooperative antiferrodistortive dis-
placements of the transition metal d0 B cations due to the SOJT
effect (see Section 2.01.4.1). Such displacements proceed
toward the anions of the (AX) layer, where the A cation has
smaller charge, thus compensating the underbonding of the
anions in this layer and relieving the overbonding of the an-
ions in the (A0 X) layer. Thus, the factors promoting the layered
A-site ordering are primarily the charge difference of the A and
A0 cations and the SOJT-driven off-center displacements of the
d0 B cations. The layered A-site ordering is most frequently
observed if the place of the low-charged A cation is taken by
Li or even a cation vacancy, as in the perovskite families A1/
3xBO3 (A La, Ce, Pr, Nd, B Nb, Ta),
354360
A2/3xTiO3
(A (La,Sr), Ce, Pr, Nd), 361364
and (La2/3xLi3x)TiO3.365369
The layered A-site ordering is governed by the X3 irrep of the
parent Pm3m space group that gives P4/mmm as the space sym-
metry of the ordered structure without any octahedral tilts.361
Although the layered A-site ordering and octahedral tilting dis-
tortion separately retain the inversion center in the perovskite
structure, their combination leads to acentric structures for the
aac, ab0c, abc, and abc tilt systems. There are 12 tilt
systems compatible with the layered A-site ordering.
The off-center displacements of the d0 B cations stabilizing
the layered A-site ordering are compatible with the rock-salt
ordering pattern if two different cations are present at the B site.
In such double-ordered AA0 BB0 O6 perovskites, two kinds of or-
dering synergistically support each other.353 All double-ordered
perovskites are formed with B0 W6 (A Na, K, A0 LaHo,
B Mg, Ni, Co, Mn)353,370373 (Figure 27). Only the W6 cat-
ions are subjected to the off-center displacements, whereas the B
cations retain a fairly symmetric octahedral coordination.
The synergetic character of two orderings and the off-center dis-
placements is best illustrated by the suppression of the off-center
displacements in the rock-salt ordered A2BWO6 perovskites
(A Ca, Sr, Ba, B Fe, Mn, Mg, Ca)374376 without the layered
A-site ordering and the absence of the layered A-site ordering in
NaLaTi2O6 without the rock-salt B-cation ordering.377
La
Na
WO6
Figure 27 The crystal structure of NaLaMgWO6 with the layered A-site
and rock-salt B-site ordering. Yellow arrows indicate off-center
displacements of the W6 cations toward the (NaO) layers.
22 Transition-Metal Perovskites
CuO4
Ca
TiO6
Figure 28 The crystal structure of the A-site ordered CaCu3Ti4O12.
Besides the electrostatically driven A-site ordering, a signif-
icant difference in electronic structures of the A and A0 cations
can also promote their ordered arrangement. In the AA0 3B4O12
perovskites (A Na, Ca, Sr, Pb, Bi, La, Nd, A0 Cu2 or Mn3,
B Ti, V, Mn, Fe, Cr, Ru, Ga, Ge, Sb, Sn, Ta),378390 the A0
positions are taken exclusively by the cations showing a strong
tendency toward the first-order JahnTeller effect (e.g., Cu2 or
Mn3, see Section 2.01.4.2). These cations should normally
reside at the B-type positions of the perovskite structure, be-
cause they are too small for the large 12-coordinated A-site
cavity. However, the specific anion environment, which is
suitable for accommodating Cu2 or Mn3 at the A0 position,
is created by the aaa octahedral tilt system with large tilt
magnitude (e.g., the TiOTi bond angle is only 140.7 in
CaCu3Ti4O12378) (Figure 28). The tilt creates the square-planar
anion coordination for the A0 cations, which is additionally
stabilized by the intrinsic JahnTeller effect. The larger A cat-
ions are located in the icosahedral 12-coordinated cavities. The
close positioning of the A0 and p B transition-metal cations,
which are separated only by ap/ 3  2.15 A, allows for a strong
electronic interaction. In LaCu3Fe4O12, this interaction mani-
fests itself as an isostructural phase transition with the charge
transfer between the Cu and Fe cations.387 The transition from
the high-temperature LaCu32Fe43.75O12 phase to the low-
temperature LaCu33Fe43O12 phase is accompanied by
converting a paramagnetic metal into an antiferromagnetic
insulator and a unit cell expansion of 1%.
The A-site ordering is possible even without the charge
difference and the additional energy gain due to JahnTeller
effect, but solely due to octahedral tilting distortion that creates
crystallographically distinct A positions. In CaFeTi2O6, the rare
tilt system aac renders the ordering of the A-cations Ca2
and Fe2 in separate chains.391 The Fe atoms adopt the coor-
dination number 4, with the FeO4 tetrahedra and squares
alternating along the chain. The coordination number of the
Ca atoms can be represented as 6 4 (Figure 29).
Besides the A-site ordering in stoichiometric perovskites,
fascinating examples of the A-site ordering combined with
the ordering of anion vacancies are known. Such compounds
are discussed in more detail in Section 2.01.5.5.
2.01.5.4 Ordering of Anion Vacancies in Perovskites
Anion nonstoichiometry is intrinsic in transition-metal perov-
skites due to a variable oxidation state of the transition-metal
Figure 29 The crystal structure of the A-site ordered CaFeTi2O6.
The TiO6 octahedra are blue. The Fe cations at 50% of the A sites are
situated in brown FeO4 squares and tetrahedra. The other 50% of the A
sites are occupied by the Ca cations (yellow).
cations and their ability to adopt the coordination numbers
from 4 to 6 (or even 2, as for Cu). When the concentration
of anion vacancies is large enough for their mutual interaction,
the structure gains additional energy due to the anion-vacancy
ordering. The anion-deficient perovskites adopt the composi-
tions ranging between ABX3 and ABX2. A striking example of
structural changes driven by anion ordering in perovskites is
given by the SrFeO3d compounds (Figure 30). The increase in
the concentration of oxygen vacancies in the cubic SrFeO3 pe-
rovskite first decreases the coordination number of the Fe cat-
ions, and at d 0.125 and 0.25 the ordered patterns of oxygen
atoms and vacancies arise, resulting in the Sr8Fe8O23 and
Sr4Fe4O11 compounds with ordered FeO5 pyramids and FeO6
octahedra.392 Further reduction leads to the Sr2Fe2O5 (d 0.5)
brownmillerite with alternating layers of the FeO4 tetrahedra
and FeO6 octahedra.393 The d5 electronic configuration of the
Fe3 cations in Sr2Fe2O5 favors the brownmillerite structure
with two distinct coordination numbers of 4 and 6 for the Fe
cations because of the same crystal-field energy for the tetrahe-
dral and octahedral environments. A reduction of Sr8Fe8O23
with CaH2 produces the SrFeO2 (d 1) compound with the
tetragonally distorted perovskite structure.394 This so-called
infinite layer-type structure consists of layers of corner-sharing
FeO4 squares with the Sr cations in between the layers. This
structure can be derived from the parent perovskite structure
by removing all apical oxygen atoms from the FeO6 octahedra.
A similar sequence of anion-deficient phases is observed in
the SrMnO system. Starting from cubic SrMnO3, the reduc-
tion leads subsequently to the Sr7Mn7O19 and Sr5Mn5O13
phases with the gradually increasing fraction of the MnO5
tetragonal pyramids at the cost of the MnO6 octahedra.395397
At the final stage of the reduction, the Sr2Mn2O5 compound is
obtained, but the vacancy ordering pattern in this compound is
significantly different from the brownmillerite-type Sr2Fe2O5
(Figure 31). Instead of the alternating octahedral and
 Transition-Metal Perovskites 23

FeO6
one or two apical octahedral vertices, thus increasing the en-
ergy splitting of the eg orbitals and minimizing the free energy.
FeO
eO5 Interestingly, the BO5 pyramids are always connected in such a
FeO4
Fe
way that the apical vertex of one pyramid forms an equatorial
vertex of another pyramid, thus promoting the elongation of
the remaining apical BO distance and the shortening of the
equatorial BO distances, as typical for the JahnTeller cations.
Indeed, the Cu-based (La,Sr)2Cu2O5 anion-deficient perov-
Sr8Fe8O23
skites also adopt the Sr2Mn2O5-type structure.401,402 However,
even at nearly the same A2B2O5 oxygen stoichiometry, which
favors the purely tetragonal-pyramidal coordination, other or-
FeO5
dering patterns are also known and demonstrate the partial
disproportionation of the coordination environment of the
Sr2Fe2O5 transition-metal cation, maintaining the (BO1.5) composition
FeO4 of the layer. In La8xSrxCu8O20d (1.28  x  1.92),403 one half
of the Cu cations retains the tetragonal-pyramidal oxygen co-
Sr4Fe4O11
ordination, one-quarter of the Cu cations possess an octahe-
dral environment, whereas for the remaining quarter two
SrFeO2
oxygen atoms are missing, resulting in the square-planar coor-
Figure 30 The crystal structures of Sr8Fe8O23, Sr4Fe4O11, Sr2Fe2O5, dination (Figure 31).
and SrFeO2. Generally, anions can be removed from the <100>p or
< 110>p rows or {100}p planes of the ABX3 perovskite
structure.404 This results in a transformation of the BX6 octa-
hedra into BX5 tetragonal pyramids and/or BX4 squares for the
anions removed along the < 100>p rows or {100}p planes, or
into the BX4 tetrahedra if the anions are removed along the
<110>p rows. The most abundant family of the anion-
deficient perovskites with half of the B cations in tetrahedral
coordination is the brownmillerite-type compounds A2BB0 O5.
MnO5
Viewed along one of the cubic perovskite subcell axes, the
brownmillerite structure can be represented as the ordered
alternation of the (AO), (BO2), and (B0 O) layers according to
the BO2AOB0 OAOBO2 sequence. The anion vacancies
Sr2Mn2O5 reside in the (B0 O) layers resulting in infinite chains of corner-
sharing B0 O4 tetrahedra. Both B and B0 positions can be taken
CuO6 by a 3d transition-metal cation, if this cation demonstrates an
CuO4
affinity to both octahedral and tetrahedral coordinations
(A2B2O5, A Ca, Sr, B Fe3, Co3),405410 or the cation
CuO5 adopts two oxidation states favoring the octahedral and tetrahe-
dral coordinations, respectively (La2xAxMn2O5, B Mn(3d),
B0 Mn2).411,412 However, more often, the transition-metal
cation occupies the octahedral site, whereas the tetrahedral site
is preferentially taken by the group III cations, such as Al3 and
Ga3. Such a combination of the tetrahedral and octahedral
layers, as in the brownmillerite structure, provides an excellent
host for the octahedrally coordinated JahnTeller active
transition-metal B cations.413 The BO6 octahedra are apically
elongated in brownmillerites, thus giving rise to four shorter
in-plane BO bonds and two longer apical BO bonds
(Figure 32). However, not only the JahnTeller effect dictates
La8xSrxCu8O20 such apical elongation of the BO6 octahedra. The short metal
oxygen bonds for the B0 cations in the tetrahedra also require
Figure 31 Ordering patterns of oxygen vacancies in the crystal this apical elongation in order to keep an appropriate separation
structures of Sr2Mn2O5 and La8xSrxCu8O20d. between the octahedral and tetrahedral layers and reasonable
distances between the A cations and neighboring oxygen atoms.
tetrahedral layers, the Sr2Mn2O5 structure consists of the MnO5 A comparison of the perovskite NdFeO3414 and brownmillerite
tetragonal pyramids linked by common corners.398 This illus- Ca2FeAlO5415 (Figure 32) shows that the decrease of the bond
trates that the exact pattern of the anion-vacancy ordering is lengths for the B0 cation requires a shift of the apical oxygen atom
determined, to a large extent, by the electronic configuration away from the B cation, even for JahnTeller inactive cations,
and crystal chemistry of the transition-metal cations.399,400 such as Fe3. In the Fe-containing brownmillerites, the difference
JahnTeller cations, such as Mn3 and Cu2, prefer to lose between the equatorial and apical FeO bond lengths is quite
24 Transition-Metal Perovskites
2.002.02
2.372.61
B bic GdFeO3-type perovskite structure with the Mn and Ga
A disorder was also observed for the solid solutions La2
2.22
B lerites, both corresponding to the A(B,B0 )O2.5 formula, but the
Oeq 2.492.57
1.94
Oap
2.36
1.71
B structure is formed, consisting of quasi-2D blocks of the corner-
Figure 32 Fragments of the NdFeO3 perovskite (a) and Ca2AlFeO5
brownmillerite (b) structures with the most relevant bond lengths.
Hadermann, J.; Abakumov, A. M.; DHondt, H.; Kalyuzhnaya, A. S.;
Rozova, M. G.; Markina, M. M.; Mikheev, M. G.; Tristan, N.; Klingeler, R.;
Buchnerd, B.; Antipov, E. V. J. Mater. Chem. 2007, 17, 692698
Reproduced by permission of The Royal Society of Chemistry.
remarkable: d(FeOeq) 1.9421.954 A, d(FeOap) 2.217 A
in Ca2FeAlO5415; d(FeOeq) 1.9541.959 A, d(FeOap)
2.128 A in Ca2FeGaO5.416 If B is a JahnTeller active cation, the
coinciding geometric and electronic requirements should pro-
mote a formation of the brownmillerite structure. Indeed, for
B Mn3 a family of A2MnB0 O5 brownmillerites is known for
A Ca, Sr, and B0 Al, Ga.417423 A similar family exists
for B Cu2: LnACuGaO5 (Ln La, Pr, Nd, A Ca, Sr).424426
2.01.5.5 Coupled Cation and Anion-Vacancy Ordering
The A2BB0 O5 brownmillerites with the JahnTeller B cation
demonstrate the example of the B-site ordering promoted by
the ordering of the anion vacancies. The main factor forcing the
B and B0 cations to occupy the positions in separate layers is the
different coordination numbers of the B and B0 cations in
complex oxides. According to Section 2.01.5.1, the layered
ordering of the B cations is extremely rare in stoichiometric
perovskites, and certain criteria for the charge and size differ-
ence must be fulfilled to stabilize this ordering pattern. In most
of the brownmillerites, the B and B0 cations have the same
formal charge of 3 or their charges differ by not more
than 1, which is obviously insufficient for the layered ordering.
Nevertheless, a complete layered ordering of the B and B0
cations is observed in many brownmillerites, especially those
containing the JahnTeller cation at the B position. The Mn
and Ga cations in the A2MnGaO5 brownmillerites are
completely ordered,417420 but this ordering is destroyed by
increasing the coordination number of the B0 Ga cations up
to 6, resulting in disordered perovskite phases. The
Ln2MnGaO6 (Ln La, Nd) compounds have the orthorhom-
cations randomly occupying the octahedral sites.427 Similar
xSrxMnGaO6 (x 0.7).
428,429
A variation of the Mn/Ga ratio
also prevents the formation of the layered brownmillerite
structure: the solid solutions SrMn1xGaxO3d (0 x  0.33)
have a disordered cubic perovskite structure.430
A coupling between the B-cation and anion-vacancy ordering
is particularly well illustrated by a comparison of the perovskites
with the same anion deficiency, but different B/B0 ratio. In the
A3BB0 2O7.5 (A Na, Ba, B Mn, Ln, Y, B0 Al, Ga, V)431,432
perovskites, the oxygen content is the same as in the brownmil-
amount of the tetrahedrally coordinated B0 cations is twice the
amount of the octahedrally coordinated B cations. The layered
ordering of B0 O4 tetrahedra and BO6 octahedra, where each
oxygen atom is shared between two coordination polyhedra
around the B and/or B0 cations, is impossible for the
A3BB0 2O7.5 compounds, because the oxygen content is not com-
patible with the amount of the tetrahedrally coordinated cations
at such linkage of the polyhedra. Thus, another perovskite-like
sharing B0 O4 tetrahedra and BO6 octahedra, ordered in a rock-
salt manner (Figure 33). The blocks are linked together by the
B0 2O7 dimers that comprise two B0 O4 tetrahedra sharing a com-
mon corner and having one dangling oxygen vertex, which is
linked to only one B0 cation. Thus, half of the B0 O4 tetrahedra
have such dangling oxygen vertices. In the similar perovskite-
like Sr10Sc4Ga6O25 structure with fully ordered ScO6 octahedra
and GaO4 tetrahedra, the amount of tetrahedra also exceeds the
amount of octahedra resulting in 1/3 of GaO4 tetrahedra that
possess the dangling oxygen vertex (Figure 33).433
Topotactic reduction of nearly oxygen-stoichiometric pe-
rovskites may result in the anion-vacancy ordering coupled
with the charge ordering for the transition-metal cation in the
mixed-valence state. The reduction of (La,A)MnO3d (A Ca,
Sr, Ba) with alkali-metal hydrides generally leads to the (La,
A)2Mn2O5 brownmillerites with the Mn cations having a
mixed valence between 2 and 3.411,412 The tetrahedrally
coordinated positions in these brownmillerites are taken by
the Mn2 cations, which are isoelectronic with the Fe3 d5
cations and thus have no additional crystal-field stabilization
of the octahedral coordination. However, in contrast to the
previous example of the A(B,B0 )O2.5 compounds with the var-
iable B/B0 ratio strongly linked to the oxygen-vacancy ordering
pattern, such a link is noticeably weaker if the B and B0 cations
differ only in their formal oxidation states. In La4Mn4O11, the
Mn3/Mn2 occur in the 1:1 ratio, but the structure is built
of three octahedral (O) layers and one brownmillerite-type
tetrahedral (T) layer alternating in the OOOTOOOT sequence
that matches the overall oxygen content, but does not imply
the complete Mn3/Mn2 charge ordering.434
A large variety of anion-vacancy ordering patterns can be
obtained in a combination with the ordering of different
chemical species in the A sublattice of the (A,A0 )BO3d perov-
skites. The primary driving force for such ordering is a

GaO4
ErO6
Ba3ErGa2O7.5
ScO6
GaO4
Sr10Sc4Ga6O25
Figure 33 The crystal structures of Ba3ErGa2O7.5 and Sr10Sc4Ga6O25.
difference between size and coordination number of the A and
A0 cations. In the YBa2Cu3O6d perovskites, a repetition period
along one of the cubic perovskite axes is tripled due to the layer
sequence CuOdBaOCuO2YCuO2BaOCuOd (-oxygen A-site ordering in the stoichiometric BaLaMn2O6 perovskite.
vacancy) (Figure 34). Smaller Y3 cations (r 1.015 A, CN 8)
residing in the anion-deficient (Y) layers are surrounded by a
distorted cube of the oxygen atoms with dav(YO) 2.41 A.
Larger Ba2 cations (r 1.42 A, CN 8) always occupy the po- such as brownmillerite, which interleaves with the A2A0 Fe3O8-
sitions in the complete (BaO) layers. Although the coordination
number at d  0 is the same for the Y and Ba cations due to the
almost complete absence of the oxygen atoms in the (CuOd)
layers, the average distance dav(BaO) 2.85 A for the Ba cat-
ions significantly exceeds that for the Y cations.435 Indeed, the
increase in d adds extra oxygen atoms exclusively to the (CuOd)
layer, thus completing the coordination environment of the Ba
cations up to 10 in YBa2Cu3O7, whereas the Y cations retain
their coordination number of 8 (Figure 34).
One can compare two homologous series AnBn1B0 O3n1
and An1A0 BnO3n1, where divalent alkali-earth and trivalent
rare-earth cations are located in the A sublattice. The AnBn
0
1B O3n1 series was already mentioned above. It is the brown-
millerite homologous series where the perovskite blocks with
the average thickness of n  1 octahedra are separated by the
anion-deficient (B0 O) (-oxygen vacancy) layers, thus result-
ing in a larger amount of the octahedral (O) layers compared
to the tetrahedral (T) layers, unlike in the brownmillerite struc-
ture with the O/T ratio of 1. In the An1A0 BnO3n1 compounds,
due to smaller size of A0 cations in comparison with that of the
A cations, the anion vacancies are concentrated in the A0
Transition-Metal Perovskites 25
Cu 2 CuO
BaO
CuO2
Y
CuO2
BaO
Cu 2 CuO
YBa2Cu3O6 YBa2Cu3O7
Figure 34 The crystal structures of YBa2Cu3O6 and YBa2Cu3O7.
layers, so that the B cations in the adjacent (BO2) layers acquire
tetragonal-pyramidal coordination. Thus, the layer sequences
for both series can be written as
An Bn1 B0 O3n1 : B0 Oh  AO  BO2  AO  BO2n2  AO
 B0 Oh
A n1 A 0 Bn O3n1 : A 0 h  BO2  AOn1  BO2  A 0 h
In BaRFe2O5436438 and Ba2RFe3O8d,439442 the cations A0
(Y) and cations A (Ba) have CNs of 8 and 12, respectively
(Figure 35). An important influence of the anion-vacancy or-
dering on the A-site ordering can be shown using an example of
the BaLaMn2O6d (d 0, 1) compounds.443445 The
BaLaMn2O6 compound with a complete anion sublattice dem-
onstrates no order of the Ba2 and La3 cations. However, the
synthesis of the anion-deficient BaLaMn2O5 results in the lay-
ered ordering of the anion vacancies and A cations, as found in
BaYFe2O5. Subsequent soft oxidation allows preservation of the
The structural fragments combining both A-site and anion-
vacancy ordering are sufficiently robust to even form inter-
growths with the blocks of other anion-deficient perovskites,
type blocks in the Ca2Ba2Nd2Fe6O15.6 compound.446 How-
ever, one can expect that the ordered stacking of layers in the
homologs of both An1A0 BnO3n1 and AnBn1B0 O3n1 series will
become less stable with increasing n, that is, with increasing the
distance between the anion-deficient layers and decreasing the
amount of oxygen vacancies. Due to repulsive forces, positively
charged vacancies tend to be more homogeneously distributed
in the structure, thus impeding the formation of ordered layer
sequences. The ordering in the n 4, 5, 6 An1A0 BnO3n1 com-
pounds is only achieved by a concomitant order of two B
cations with the stable CN 5 and 6, as observed in the
mixed oxides (AA0 Cu2O5)(ABO3)n2 (B Ti, Sn).447449 Start-
ing from a certain n value, it becomes more energetically prof-
itable to increase the coordination number of the cations in the
anion-deficient layers rather than to increase the thickness of
the perovskite block between the anion-deficient layers.
Although a configuration with the layered ordering of the A
and A0 cations with the anion vacancies located at the (A0 )
(-oxygen vacancy) layers seems to be most favorable, the
ordering of the A and A0 cations within the same layer can
occur if the anion vacancies are ordered in the adjacent (BO2d)
26 Transition-Metal Perovskites
FeO2
BaO
FeO2
Y Sr3/4Y1/4O
FeO2
BaO
BaYFe2O5 FeO2
Ba2YFe3O8
Figure 35 The members of the An1A0 BnO3n1 homologous series
BaYFe2O5 (n 2) and Ba2YFe3O8 (n 3).
layers. The introduction of extra oxygen into the anion-
deficient Sr2Co2O5 perovskite by a heterovalent replacement
of Sr2 by R3 results in the anion-deficient cobaltites (Sr,R)
CoO2.5d: Sr0.7Y0.3CoO2.62,450 Sr0.7Dy0.3CoO2.62,451 Sr0.66
Y0.33CoO2.79,452 SrxR1xCoO3y, R SmYb, Y, 0.6 < x  0.9.453
These compounds, as well as the Sr0.73Y0.27Mn0.67Ga0.33O2.63
manganite,454 adopt a tetragonal structure with a  2ap, c  4ap
and contain layers of tilted BO6 octahedra that alternate with the
oxygen-deficient layers according to the BO2(Sr,R)O
BO12d(Sr,R)OBO2 sequence (Figure 36). The (BO12d) tilting and/or deformation of the octahedra. The incommensurate
layer consists of clusters of four BO4 tetrahedra, where each
tetrahedron shares two oxygen atoms with the adjacent tetrahe-
dra. An extra oxygen atom, located between four tetrahedral
clusters, expands the coordination of a fraction of the B cations
in these layers up to CN 4 1 (four shorter and one longer
BO separations). The Sr2 and R3 cations are ordered on a
square mesh in such a way that each R3 cation is surrounded by
eight Sr2 cations (for the idealized Sr/R 3/1 ratio).
2.01.6 Aperiodic Order in Perovskites
Aperiodic structures possess long-range ordering and, at the
same time, lack the translational symmetry. This can be imag-
ined as a periodic perturbation superimposed onto the parent
basic structure. It means that neighboring unit cells of the basic
structure are no longer equivalent, but deviation from the basic
structure can be described by a periodic function with the
period related to the periodicity of the basic structure by an
irrational number. Then no common supercell can be found to
fit both periodicities or, at least, such a supercell will be too
large. The extra periodicity manifests itself on the diffraction
patterns by the appearance of extra reflections (satellites) at the
positions that require more than three reciprocal basis vectors
to be indexed with integer numbers. The extra reciprocal vec-
tors are the modulation vectors. For such incommensurately
modulated structures the periodicity, which is lost in 3D space,
is restored when embedding the structure into a higher dimen-
sional space, and the structure can be handled as a periodic one
at the cost of increasing the dimensionality.455
CoO1.25
Sr
Sr3/4Y1/4O
CoO2
Y
CoO1.25
Figure 36 The crystal structure of Sr3YCo4O10.5.
The incommensurate order often arises from coexisting dis-
tortion modes that interact with each other and produce a mod-
ulation wave. In perovskites, such distortion modes can be
associated with dipolar interactions due to atomic displacements
and/or elastic deformations, for example, due to the octahedral
ordering in brownmillerites provides an example of such compet-
ing modes: dipoledipole interaction between the tetrahedral
chains tends to minimize the free energy, but the most favorable
arrangement with the shortest separation between the chains
causes deformations of the octahedra in the neighboring layers,
thus increasing the elastic strain. The atomic displacements within
the tetrahedral (B0 O) layers of the brownmillerite structure can
be described as a consequence of a cooperative rotation of
the B0 O4 tetrahedra. The tetrahedra with a different sense of
rotation clockwise and counterclockwise alternate along the
direction of the chain propagation (Figure 37). The chains can a-
dopt two mirror-related configurations, which are arbitrarily
called left (L) and right (R) and have opposite directions of
the tetrahedral rotations. The transformation of the L chain into
the R one and vice versa does not noticeably influence the first
coordination spheres for the cations and does not change the
interatomic distances, thus leading to the similar configurational
energy for the sets of L or R chains. As a result, these chains should
have similar probability of formation. However, a cooperative
rotation of the tetrahedra results in a net shift of the oxygen
atoms of the chains with respect to the cations causing antiparallel
dipoles to be associated with the L and R chains (Figure 37). The
tetrahedral chain ordering is driven by the minimization of free
energy due to the interaction between these antiparallel dipoles,
which is most effective when the chains are at the shortest dis-
tances from each other, that is, located in the same layer in the L
RLR order. This ordering results in a commensurately modu-
lated structure with the modulation vector q [0,0, 1/2]. Such
structures are observed in Sr2MnGaO5,456 Sr2CuGaO5,457
Ba2In2O5,458 and La1.5Sr0.5Mn2O5.411

B O
L R
Figure 37 Left (L) and right configurations of the tetrahedral chains
in the (B0 O) brownmillerite layer. The curved arrows denote the rotation
of the B0 O4 tetrahedra with respect to their ideal perovskite positions.
The red and green arrows indicate the dipole moments of the B0 O bonds
and total net dipole moments of the chains.
A factor competing with the dipolar interaction between the
tetrahedral chains arises from cooperative displacements of the
oxygen atoms connecting the tetrahedral and octahedral layers.
When all tetrahedral chains in the layer are of the same con-
figuration, these oxygen atoms are displaced in-phase and
induce the gliding of the octahedral layer with respect to the
tetrahedral one. However, the LRLR ordering results in an
out-of-phase displacement of the bridging atoms for the chains
of different configurations, thus causing an elastic deformation
of the octahedral layers. The ordering schemes of the tetrahe-
dral chains depend on the parameters that determine the
dipoledipole interaction: the degree of the tetrahedral rota-
tion and the separation between the neighboring tetrahedral
layers.423 As a result of this competition, the compromising
long period LR ordering patterns arise. In Ca2Co2xAlxO5
(x  0.75) with q [0,0, 5/8] and q [0,0, 2/3] these periodic-
ities correspond to the chain sequences LRLLRLRR and
LRLLRL, respectively.459
Besides the brownmillerites with the rational values of the q-
vector components, the incommensurately modulated brown-
millerites are also observed (Ca2Al2O5 (q [0, 0, 0.595]),460 the
high-temperature form of Ca2Fe2O5 (q [0, 0, 0.588])461).
A large separation between the tetrahedral layers (2ap)
weakens the interaction between the tetrahedral chains in
the neighboring layers, thus making the disorder and polytypism
possible. In Sr2Fe2O5, the perfect LRLR order is established
within the tetrahedral layers, but different stacking sequences of
the layers are energetically very close and realized simulta-
neously on the local scale462 (Figure 38). In La2xAxMn2O5,
different polymorphs were isolated as single phase materials.412
A competition between the dipolar interaction and elastic
deformation also underlies the incommensurability in the
transition-metal perovskites with the lone-pair A cations. In
Bi0.75La0.25FeO358 and Pb2CoWO6,463 the dipolar interactions
originate mainly from lone-pair-driven displacement of the A
cations, which form short covalent bonds with a part of the
oxygen atoms (see Section 2.01.3). These displacements are
ordered in an antiferroelectric manner. In both cases, the
octahedral tilts can be decomposed into the part which is
Transition-Metal Perovskites 27
nearly nonmodulated, and the part, which is modulated sig-
nificantly. In Bi0.75La0.25FeO3, the modulated tilt is almost
completely suppressed in the regions where the in-phase
A-cation off-center displacement is locally violated by an out-
of-phase displacement that is accompanied by a noticeable oc-
tahedral deformation (Figure 39). Thus, the modulation arises
from the coupling between one of the tilt components and polar
displacements of the A cations through the lattice strain.
The progressing tilting distortion of the perovskite octahe-
dral framework upon cooling may lead to intermediate incom-
mensurately modulated (or long-period commensurately
modulated) phases. Due to the gradual evolution of one of
the tilt components, the octahedral tilt sequences may be more
complex than the generally considered in-phase or out-of-
phase tilts. In the room-temperature structure of NaNbO3,
the primary aac0 tilting distortion is accompanied by an
alternating oioi sequence of the in-phase (i) and out-of-phase
(o) octahedral tilts. More generally, this structure can be con-
sidered as the d 1/4 point in the series of commensurately
modulated phases with the modulation vector q [1/2,1/2,d,]
(d 1/4, 1/6, 1/12). These modulation vectors correspond to
the in-phase and out-of-phase tilt sequences (oi)2, (ooi)2, and
(ooooii)2 arising from a weak coupling between the rotations in
the neighboring layers111,464 (Figure 40).
The local interruption of the out-of-phase tilt of the aac0
system by an in-phase tilt produces a twin boundary in the
pattern of the tilted octahedra (Figure 41). Such twin bound-
aries can be arranged in a 2D checkerboard pattern or a 1D
stripe pattern coupled with the separation of the A cations into
domains of different compositions in the Nd2/3xLi3xTiO3
perovskites465,466 and AA0 BB0 O6 (A Na, K, A0 La, Nd,
B Mg, Mn, B0 W) double perovskites.467470 For example,
Nd7/12Li1/4TiO3 separates into domains with the Nd1/2Li1/2
TiO3 and Nd2/3TiO3 compositions, thus resulting in a
14ap  28ap  2ap supercell for the commensurate approxi-
mant. The AA0 BB0 O6 double perovskites are supposed to split
into domains with an A/A0 ratio close to 1:1 and the domains
with the A13xA0 1xBB0 O6 composition, in order to maintain
the charge balance. The commensurate supercells can be as
large as 10ap  10ap  2ap for KLaMnWO6 and 14ap  14ap 
2ap for NaNdMgWO6. The octahedral tilt twin boundaries
help to alleviate the strain at the boundaries of the domains
with different chemical compositions.
The incommensurability in perovskites can also be caused
by a necessity to gradually vary the chemical composition in
order to adopt the oxygen deficiency due to heterovalent re-
placement. In the perovskites with the lone-pair A cations, the
large oxygen deficiency does not manifest itself as ordered
or disordered point vacancies, but rather appears as a set
of equidistantly spaced parallel planar defects with the crystal-
lographic properties similar to those of crystallographic shear
(CS) planes in ReO3-type oxides.471473 The ReO3-type struc-
ture is based on the perovskite octahedral framework where all
A sites are vacant. Formally, the CS plane can be generated
from a cut of the basic structure along a certain crystallographic
plane. Then one part of the structure is displaced with respect
to the other part over a vector, which is a fraction of the lattice
translation of the basic structure (Figure 42). As a result of such
displacement, the octahedra along the cut plane share edges
instead of corners, thus eliminating part of the oxygen atoms
and reducing the overall oxygen content.
28 Transition-Metal Perovskites

b
L R
b
R
L R R

b L L
R L

R
R R

c R
(a) R R

L
L
b
L
R
R

R c
(d)
c R
(b)
c
(c)

Figure 38 Different polymorphs locally observed in Sr2Fe2O5 brownmillerite. Reprinted with permission from DHondt, H.; Abakumov, A. M.;
Hadermann, J.; Kalyuzhnaya, A. S.; Rozova, M. G.; Antipov, E. V.; Van Tendeloo, G. Chem. Mater. 2008, 20, 71887194. Copyright 2012 American
Chemical Society.

+ + + + + + + +
+ + + + + + + +

Figure 39 Overview of the incommensurately modulated Bi0.75La0.25FeO3 crystal structure. The Bi atoms are shown as large yellow spheres, the La
atoms are not shown for clarity, and the Fe atoms are inside the octahedra. The plus and minus signs mark the displacement directions of the Bi atoms in
the BiO chains running along the a-axis in two adjacent (BiO) layers along the b-axis (first and second rows, respectively). The in-phase displacement
(both or both ) prevails in the structure. Note the antiphase displacements at the region where the   sequence is violated by the   
fragment. Reprinted with permission from Rusakov, D. A.; Abakumov, A. M.; Yamaura, K.; Belik, A. A.; Van Tendeloo, G.; Takayama-Muromachi, E.
Chem. Mater. 2011, 23, 285292. Copyright 2012 American Chemical Society.

The simplest CS plane in perovskites with the filled A
positions can be derived from the parent perovskite structure
through a shear operation474,475 (Figure 43). Assuming that
the oxygen vacancies in the perovskite structure are arranged
along the (101)p plane, one can apply the R0 1/2[110]p
displacement vector to this plane in order to eliminate the
oxygen vacancies. This transforms the corner-sharing BO6 octa-
hedra of the perovskite structure into the edge-sharing BO5
tetragonal pyramids along the interfaces. Such shear operation
creates six-sided tunnels with two columns of the lone-pair A
cations inside that are displaced over 1/2ap relative to each
other. In the structure generated this way, the distance between
two A cations in the six-sided tunnels should be as short as 1/2
[110]p  2.8 A. An extra displacement component R1 e[001]
(e  1/31/2) should be applied to the interface in order to
achieve the reasonable AA distance of 3.6 A. This extra dis-
placement does not change the connectivity scheme of the
polyhedra and is defined as a relaxation part. The resulting
translational 1/2[110] (101)p interfaces can be considered as
CS planes. The six-sided tunnels along the interfaces are occu-
pied exclusively by the lone-pair A cations. The coordination
environment of the A cations in the tunnels is quite

i o
i
o o i
NbO6
o o
o o i i
o
i i o
o o
Figure 40 The (oi)2, (ooi)2, and (ooooii)2 polymorphs of NaNbO3.
asymmetric (three shorter AO distances (2.282.43 A) and
three longer ones (2.632.70 A)) and needs to be completed
by the localized lone-pair domain.
Combining the (101)p, (100), and (001)p-oriented parts of
the interfaces, the CS planes can adopt any arbitrary orientation
according to the required oxygen content.476 The oxygen content
can also be changed by changing the thickness of the perovskite
blocks between the interfaces. These, together with atomic dis-
placements due to lattice relaxation at the interfaces, result in an
incommensurability of the perovskites modulated by the CS
planes. The modulation vector then describes the orientation
and spacing between the interfaces, and its components are
related to the chemical composition of the compounds. For the
(Pb,Bi)1xFe1xO3y perovskite-based solid solutions, the com-
position can be expressed through the components of the q
[a,0,g] modulation vector as Pb6g2aBi17gaFe1gaO33ga with metal cations, such as Ti4, Zr4, Ta5, Nb5, Mo4, 5, and
a homogeneity range varying from Pb0.857Bi0.094Fe1.049O2.572 to
Pb0.409Bi0.567Fe1.025O2.796. The CS planes in these structures are
directed nearly along the (509)p perovskite plane, and the orien-
tation does not significantly change with the composition
(Figure 44). The compositional variation occurs due to a change
in the distance between the interfaces, which, expressed as an
average amount of the FeO6 octahedra, varies from 4.2 to
12.1.477 Other known anion-deficient perovskites modulated
by the CS planes contain Fe, Mn, Ti, and Sn as the B cations
and Pb and Bi as the lone-pair A cations.478484
2.01.7 Mixed-Anion Perovskites: Oxynitrides and
Oxyfluorides
In comparison with the large variety of cations, which are
amenable to occupying the A or B positions in the perovskite
Transition-Metal Perovskites 29
o
o i
is much smaller. The anions in the perovskite structure form a
close-packed arrangement together with the A cations forming
the mixed (AO3) close-packed layers stacked in the ABCABC
(or ccc) sequence. This requires the size of the anions to be
close to the size of the A cations. Only the fluoride F anion
(r 1.30 A) and nitride N3 anion (r 1.50 A) are able to
satisfy the geometrical requirements and can substitute
oxide anion O2 (r 1.40 A) in the perovskite structure.485
Other halogenide and chalcogenide anions are too large to be
incorporated into the perovskite close-packed layers. They usu-
ally form layered structures separating quasi-2D perovskite
blocks.
The O2 ! N3 replacement requires a compensation of
excessive negative charge. Thus, the B-sublattice of oxonitride
perovskites usually accommodates highly charged transition-
W5, 6.486489 Another factor, which severely limits the variety
of transition-metal cations forming perovskite oxynitrides, is
the most commonly used preparation route of these com-
pounds via the ammonolysis reaction: a presynthesized oxide
with the necessary cationic ratio or a mixture of oxides and/or
carbonates with the required stoichiometry is fired in flowing
NH3 at temperatures of 7001000  C.485,490 At these tempera-
tures and at atmospheric pressure, NH3 is thermodynamically
unstable with respect to the N2 H2 mixture, thus creating
strongly reducing conditions at the ammonolysis reaction.
Therefore, if metal oxides are prone to a partial or a complete
reduction, an oxynitride cannot be prepared by the ammono-
lysis reaction. Although NH3 decomposes at high temperature,
it serves as a nitridizing agent anyway due to high NH3 flow
through the reaction zone. Thus, the ammonolysis reaction
occurs at highly nonequilibrium conditions. The N2 itself can-
not be an effective nitridizing agent because of the high activa-
structure, the amount of anions accommodated in perovskites tion energy required to break the NN triple bond.
30 Transition-Metal Perovskites

Ochahedral tilt WO6

[100]p MgO6
twin boundary

[010]p

+ + +

Out-of-phase tilt () In-phase tilt (+)

WO6 rotates clockwise WO6 rotates counterclockwise
MgO6 rotates counterclockwise MgO6 rotates counterclockwise

ap

12ap

[010]p

[100]
0.5 nm

Figure 41 Top: octahedral tilting in NaLaMgWO6. The and symbols indicate out-of-phase and in-phase octahedral tilting, respectively.
The A cations are omitted for clarity. Bottom: HRTEM image of the modulated structure showing the domains with 12ap periodicity. Reprinted with
permission from Garca-Martn, S.; Urones-Garrote, E.; Knapp, M. C.; King, G.; Woodward, P. M. J. Am. Chem. Soc. 2008, 130, 1502815037 and Garca-
Martn, S.; King, G.; Urones-Garrote, E.; Nenert, G.; Woodward, P. M. Chem. Mater. 2011, 23, 163170. Copyright 2012 American Chemical Society.

Due to the lower electronegativity of nitrogen (3.04) in
comparison with oxygen (3.44), the BN bond is expected to
be more covalent than the BO bond. The change in the
covalency has a direct affect on the optical properties of the
d0 transition-metal oxynitrides. Since the energy of the 2p N
states forming the top of the valence band increases, the energy
difference between the valence band and p* conduction band
formed mainly by empty d-states of the transition metal
decreases,491 thus narrowing the bandgap492 and shifting the
optical absorption band from ultraviolet (for perovskite ox-
ides) to the visible light region for perovskite oxynitrides.
Oxynitrides appear to be intensively colored, in contrast to
their oxide counterparts. Such compounds are even suggested
as potential pigments, because their color can be tuned by
adjusting the bandgap through changing the O/N ratio.493
Another consequence, which is also attributed to the high
covalency of the BN bond, is the shrinkage of the BX average
interatomic distance upon the O2 ! N3 replacement, even in
spite of the larger ionic radius of N3 compared to O2. In
BaTaO2N, the average TaN distance is 1.994 A and an average
TaO distance is 2.091 A, as measured locally by extended
x-ray absorption fine structure (EXAFS) spectroscopy.494
Most of the perovskite oxynitrides crystallize either in the
cubic perovskite structure or in its distorted variants due to the
cooperative octahedral tilting according to the tolerance factor.
Long-range ordering of the oxide and nitride anions generally
is not observed. However, short-range correlations in the ar-
rangement of oxygen and nitrogen exist. In the TaO4N2
 Transition-Metal Perovskites 31

octahedra of cubic BaTaO2N, the occupation of cis-positions by
two N atoms is energetically more favorable in comparison
with the trans-positions, presumably to stabilize the dp
p-interaction between the d0 B cation and the nitride
anion.492,495,496 In SrBO2N (B Nb, Ta), the O and N anions
are partially ordered into alternating O and O0.5N0.5 layers, in
agreement with the preferential cis-configuration of the BO4N2
(102)
octahedra.497 The preferential cis-configuration of the TaO4N2
octahedra in ATaO2N (A Sr, Ba) creates an acentric coordina-
tion environment for the Ta cations, thus promoting its polar
off-center displacement. However, due to disorder in the anion
sublattice, perovskite oxynitrides do not demonstrate cooper-
ative ordering of the dipoles, but rather behave as relaxors
(see Section 2.01.4.1) with the unusually high dielectric per-
mittivity (e.g., 4900 in BaTaO2N).498
Oxyfluoride perovskites can be prepared by different routes
depending mainly on the chemical nature of the A cations. The
oxyfluorides comprising alkali-earth A cations are prepared
through low-temperature topotactic fluorination. A route via
a high-temperature solid-state reaction at ambient pressure is
inapplicable because of the high thermodynamic stability of
alkali-earth metal difluorides, especially BaF2.499 The fluorina-
tion reaction can occur as a fluorine insertion into the

[010]

(a) (b)

(102) (c) 2.5 nm (d)

(509)p

Figure 44 Top: HAADF-STEM images of Pb0.857Bi0.094Fe1.049O2.572 (a),

Pb0.648Bi0.311Fe1.041O2.676 (b), Pb0.543Bi0.423Fe1.034O2.729 (c), and
- Pb0.441Bi0.532Fe1.028O2.780 (d) showing nearly constant orientation of the
R = 1/2[101]
crystallographic shear planes and increasing distance between the
Figure 42 Generation of the 1/2[101](102)p crystallographic shear plane planes with increasing oxygen content. Bottom: the commensurate
in the ReO3-type structure: a cut of the parent structure along the (102) approximant of the Pb0.64Bi0.32Fe1.04O2.675 crystal structure. The
lattice plane and application of the R 1/2[101]p displacement vector. crystallographic shear planes are marked by dashed lines.

- Shear vector Relaxation part

(101)p

R0 = 1/2 [110]p R1 = 1/3 [001]p

ap
cp d(A-A) ~ 2.8 d(A-A) ~ 3.6
Figure 43 Application of the shear operation to the perovskite structure. The anion vacancies along the (101)p plane (indicated by black squares)
are eliminated due to displacement over the R0 1/2[110]p shear vector, transforming the BO6 octahedra (blue) into FeO5 pyramids (orange).
The relaxation displacement over the R1 1/3[001]p vector increases the separation between the A cations (indicated in yellow). The lone-pair domain
on the A cations are shown schematically.
32 Transition-Metal Perovskites
preformed anion-deficient perovskite at the temperatures of
200400  C, where fluorinating agents such as a N2/F2 gaseous
mixture,500 NF3,501 or XeF2502 serve as a source of fluorine:
ABO3x xF ! ABO3x Fx
This process has an oxidative character and increases the
formal oxidation state of the transition-metal cations. In the
preformed perovskites with a complete anion sublattice, fluo-
rine can be inserted via the anion-exchange reaction, where
oxygen is partially replaced with an equal amount of fluorine:
ABO3 xF ! ABO3x Fx xO
The anion-exchange reaction decreases the formal oxida-
tion state of the transition-metal cations and requires non-
oxidizing or reducing fluorinating agents, such as binary
fluorides MF2 (M Zn, Cu, Ni)503 or polyvinylidene
fluoride.504 Usually, the anion exchange occurs at higher tem-
perature than the anion insertion, but from a practical point of
view a complete separation of these reactions is difficult:
ABO3x x yF ! ABO3xy Fx Fy yO Sr2MnGaO5d and fluorinated Sr2MnGaO5xF1x brownmillerites.
Perovskite oxyfluorides with the alkali-metal A cations can
be prepared via direct solid-state reaction of the transition-
metal oxide with the alkali-metal fluoride (ANbO2F, A Na,
K, K3MoO3F3)341,505 or by using a high-temperature high-
pressure technique (KTiO2F).506 The high-pressure high-
temperature synthesis was also successfully utilized for the
preparation of the perovskite oxyfluorides containing lone-
pair A cations (PbBO2F, B Sc, Fe).507,508
Fluorine insertion into anion-deficient perovskites with
ordered anion vacancies suppresses the structural distor-
tions caused by the anion-vacancy ordering. Fluorinated brown-
millerites LaSrCuGaO5xF2xy (A Ca, Sr)509 and Sr2MnGaO5
xF1x (x 0.22, 0.46, 0.61)
510,511
adopt a simple tetragonal
a ap, c 2cp perovskite structure without the octahedral tilting,
while the doubling of the translation along the c-axis is induced
by the preserved layered ordering of the B and B0 cations. The
fluorinated brownmillerite LaSrCoFeO5F lacks the cation order-
ing at the B site and has a rhombohedrally distorted perovskite
structure due to the aaa octahedral tilt.512 If the JahnTeller
active cations, such as Cu2 or Mn3, are present in the structure
of the parent anion-deficient oxide, a completion of their coor-
dination environment to the octahedron by fluorine insertion
causes the stretching of the unit cell along one of the perovskite
subcell axes due to the Q3 distortion mode of the B(O,F)6
octahedra. Fluorination of the anion-deficient La6.5Sr1.5-
Cu8O19.65 cuprate513 and Sr2Mn2O5 manganite514 leads to oxy-
fluorides with a simple tetragonally distorted perovskite
structure with a ap, c cp, c > a (a 3.807 A, c 3.989 A for the ensuing physical properties. Intensive investigations of pe-
Sr2Mn2O5). The fluorine insertion into YBa2Cu3O6.11 (Y-123)
cuprate increases the c-parameter of the tetragonal a ap, c 3cp
unit cell from 11.815 to 13 A due to the formation of the
CuO4F2 octahedra, and induces superconductivity with
Tc 94 K.515,516 However, the layered ordering of the B
cations of different sizes can promote the opposite direction of
the Q3 distortion mode. In the Sr2MnGaO4.78F1.22 structure,
the MnO6 octahedra are characterized by two short apical
2.50
d(MnO)ap (O)
2.40
2.30
2.20
d ()
2.10
2.00
d(MnO)eq (O) d(MnO)eq (F)
1.90
d(MnO)ap (F)
1.80
3.00 3.20 3.40 3.60 3.80 4.00
VMn
Figure 45 Equatorial and apical MnO distances as a function of the
formal Mn valence (VMn) in the MnO6 octahedra of the oxygenated
MnOap distances of 1.876 A and four long equatorial
MnOeq distances of 1.928 A interpreted as an apically
compressed JahnTeller distortion, in contrast to the apically
elongated distortion described above510 (Figure 45). One can
speculate that such apically compressed JahnTeller distortion
in this specific case is more favorable for the optimal SrO
bonding.
Perovskite oxyfluorides containing JahnTeller inactive
transition-metal cations usually adopt the cubic perovskite
structure without the long-range O/F ordering. However,
according to the Mossbauer spectroscopy, on a local scale cis-
configuration of the FeO4F2 octahedra slightly prevails over the
trans-configuration in BaFeO2F and PbFeO2F,61,517 similar to
the local O/N ordering in the perovskite oxynitrides. In both
PbBO2F, B Sc, Fe the Pb cations are shifted from the center of
the perovskite cube along <110>, but these shifts occur in a
random way and do not produce polar structures.61,507
2.01.8 Conclusion
No other structure type of complex transition-metal oxides can
be compared with the perovskite in its variety of chemical
compositions, richness of crystal chemistry, and diversity of
rovskites during at least the last 50 years provide a solid back-
ground for understanding the crystal chemistry and physics of
transition-metal compounds. Plenty of experimental data sup-
ported by sound theoretical understanding render perovskites a
kind of toolbox for the design of new materials. Implementing
the available knowledge, one can look for a proper tool (com-
position, structure distortion, ordering of atomic species and
vacancies, appropriate electronic configuration and exchange
 Transition-Metal Perovskites 33

interactions, etc.) or their combinations, and use these instru- account parallel evolution of the synthesis methods, we can
ments to deliberately create a material with desired functional- expect that the perovskite-based complex transition-metal ox-
ity, improved performance, or novel physics. Taking into ides will remain a source of new inspiring discoveries.

Appendix
Crystallographic Data on the 15 Tilt Systems in ABX3
Perovskites

Tilt system Transformation matrix Unit cell Space group Atomic coordinates and Wyckoff positions
and origin shift parameters
0 1
a0a0a0 1 0 0 a b c ap Pm3m A: 12,12,12 (1b)
@0 1 0A (#221) B: 0,0,0 (1a)
0 0 1 0, 0, 0 O: 12,0,0 (3d)
0 1
a0a0c 1 1 0 a b  2ap P4/mbm A: 0,12,12 (2c)
@1 1 0A c  ap (# 127) B: 0,0,0 (2a)
0 0 1 0, 0, 0 O: 0,0,12 (2b); x,x 12,0, x  14 (4g)
0 1
a0bb 2 0 0 a b  2ap I4/mmm A: 0,0,0 (2a); 0,0,12 (2b); 0,12,0 (4c)
@0 2 0A c  2ap (# 139) B: 14,14,14 (8f)
0 0 2 1, 1, 1 O: 0,y,z, y  z  14 (16n); x,x,0, x  14 (8h)
0 14 4 4
aaa 2 0 0 a b c  2ap Im3 A: 0,0,0 (2a); 0,12,12 (6b)
@0 2 0A (# 204) B: 14,14,14 (8c)
0 0 2 1, 1, 1 O: 0,y,z, y  z  14 (24g)
0 14 4 4
abc 2 0 0 a  2ap Immm A: 0,0,0 (2a); 0,12,12 (2b); 12,12,0 (2c); 12,0,12 (2d)
@0 2 0A b  2ap (# 71) B: 14,14,14 (8k)
0 0 2 1, 1, 1 c  2ap O: 0,y,z, y  z  14 (8l); x,0,z, x  z  14 (8m); x,y,0, x  y  14 (8n)
0 14 4 4
a0a0c 1 1 0 a b  2ap I4/mcm A: 0,12,14 (4b)
@1 1 0A c  2ap (# 140) B: 0,0,0 (4c)
0 0 2 0, 0, 0 O: 0,0,14 (4a); x,x 12,0, x  14 (8h)
0 1
a0bb 0 1 1 a  2ap Imma A: 0,14,z, z  12 (4e)
@2 0 0A b  2ap (# 74) B: 0,0,0 (4a)
0 1 1 0, 0, 0 c  2ap O: 0,14,z, z  0 (4e); 14,y,14 y  12 (8g)
0 1
aaa 1 1 0 a b  2ap R3c A: 0,0,14 (6a)
@0 1 1A c  23ap (# 167) B: 0,0,0 (6b)
2 2 2 0, 0, 0 g 120o O: x,0,14, x  12 (18e)
0 1
a0bc 0 1 1 a  2ap I2/m A: x,0,z, x  14, z  34 (4i);
@2 0 0A b  2ap (# 12) B: 14,14,14 (4e)O: x,0,z, x  z  14 (4i); 0,y,0, y  14 (4g); 12,y,0, y  14 (4h)
0 1 1 1, 1, 1 c  2ap
4 4 4
b 6 90o
   0 1
a b b 2 0 0 a  2ap I2/a A: 14,y,0, y  12 (4e)
@0 1 1A b  2ap (# 15) B: 0,0,0 (4a)
0 1 1 0, 0, 0 c  2ap O: 14,y,0, y  0 (4e); x,y,z, x  0, y  z  14 (8f)
b 6 90o
0 1
abc 2 0 0 a  2ap I1 A: x,y,z, x  14, y  12, z  0; B: 0,0,0; 12,0,0
@0 0 1A b  2ap (# 2) O: x,y,z, x  14, y  z  0; x,y,z, x  0, y  z  14;
0 1 1 0, 0, 0 c  2ap x,y,z, x  0, y  34, z  14
a 6 b 6 g 6 90o
0 1
a0bc 2 0 0 a  2ap Cmcm A: 0,y,14, y  0 (4c); 0,y,14, y  12 (4c)
@0 2 0A b  2ap (# 63) B: 14,14,0 (8d)
0 0 2 1, 1, 0 c  2ap O: 0,y,z, y  14, z  0 (8f); x,0,0, x  14 (8e); x,y,14, x  y  14 (8g)
0 14 4
abb 1 0 1 a  2ap Pnma A: x,14,z, x  12, z  0 (4c)
@0 2 0A b  2ap (# 62) B: 0,0,0 (4a)
1 0 1 0, 0, 0 c  2ap O: x,14,z, x  z  0 (4c); x,y,z, x  14, y  0, z  14 (8d)
0 1
abc 1 0 1 a  2ap P21/m A: x,14,z, x  0, z  12 (2e); x,14,z, x  12, z  0 (2e)
@0 2 0A b  2ap (# 11) B: 0,0,0 (2a); 12,0,12 (2d)
1 0 1 0, 0, 0 c  2ap O: x,14,z, x  0, z  0 (2e); x,14,z, x  z  12 (2e);
b 6 90o x,y,z, x  14, y  0, z  14 (4f); x,y,z, x 34, y  0, z  14 (4f)
0 1
aac 2 0 0 a b  2ap P42/nmc A: 34,14,34 (2a); 34,14,14 (2b); 14,14,z, z  34 (4d) ; B: 0,0,0 (8e)
@0 2 0A c  2ap (# 137, origin O: x,-x,14, x  0 (8f); 14,y,z, y  0, z  12 (8g);
4,y,z, y  z  0 (8g)
0 0 2 0, 0, 0 choice 2) 1
34 Transition-Metal Perovskites
References
1. Mitchell, R. H. Perovskites: Modern and Ancient. Almaz Press Inc: Thunder Bay,
Canada, 2002.
2. Glazer, A. M. Acta Crystallogr. 1972, B28, 33843392.
3. Megaw, H. D. Crystal Structures A Working Approach. W. B. Saunders:
Philadelphia, 1973.
4. Goldschmidt, V. M. Naturwissenschaften 1926, 14, 477485.
5. Aleksandrov, K. S. Kristallografiya 1976, 21, 249255.
6. Woodward, P. M. Acta Crystallogr. 1997, B53, 3243.
7. Howard, C. J.; Stokes, H. T. Acta Crystallogr. 1998, B54, 782789.
8. Miller, S. C.; Love, W. F. Tables of Irreducible Representations of Space Groups
and Corepresentations of Magnetic Space Groups. Pruett: Boulder, CO, 1967.
9. Aleksandrov, K. S.; Bartolome, J. Phase Transit. 2001, 74, 255335.
10. Woodward, P. M. Acta Crystallogr. 1997, B53, 4466.
11. Goodenough, J. B. Prog. Solid State Chem. 1971, 5, 145.
12. Pavarini, E.; Yamasaki, A.; Nuss, J.; Andersen, O. K. New J. Phys. 2005, 7, 188.
13. Pavarini, E.; Biermann, S.; Poteryaev, A.; Lichtenstein, A. I.; Georges, A.;
Andersen, O. K. Phys. Rev. Lett. 2004, 92, 176403.
14. Brown, I. D. Chem. Soc. Rev. 1978, 7, 359376.
15. Brese, N. E.; OKeeffe, M. Acta Crystallogr. 1991, B47, 192197.
16. Brown, I. D.; Altermatt, D. Acta Crystallogr. 1985, B41, 244247.
17. http://www.unf.edu/michael.lufaso/spuds/index.html.
18. Lufaso, M. W.; Woodward, P. M. Acta Crystallogr. 2001, B57, 725738.
19. Salinas-Sanchez, A.; Garcia-Munoz, J. L.; Rodriguez-Carvajal, J.; Saez-Puche, R.;
Martinez, J. L. J. Solid State Chem. 1992, 100, 201211.
20. Lufaso, M. W.; Barnes, P. W.; Woodward, P. M. Acta Crystallogr. 2006, B62,
397410.
21. Watson, G. W.; Parker, S. C.; Kresse, G. Phys. Rev. B 1999, 59, 84818486.
22. Watson, G. W.; Parker, S. C. J. Phys. Chem. B 1999, 103, 12581262.
23. Walsh, A.; Watson, G. W. J. Solid State Chem. 2005, 178, 14221428.
24. Shpanchenko, R. V.; Chernaya, V. V.; Tsirlin, A. A.; Chizhov, P. S.;
Sklovsky, D. E.; Antipov, E. V.; Khlybov, E. P.; Pomjakushin, V.; Balagurov, A. M.;
Medvedeva, J. E.; Kaul, E. E.; Geibel, C. Chem. Mater. 2004, 16, 32673273.
25. Belik, A. A.; Azuma, M.; Saito, T.; Shimakawa, Y.; Takano, M. Chem. Mater. 2005,
17, 269273.
26. Arevalo-Lopez, A.M.; Alario-Franco, M.A. J. Solid State Chem. 2007, 180, 61. Belik, A. A.; Abakumov, A. M.; Tsirlin, A. A.; Hadermann, J.; Kim, J.; Van
32713279.
27. Belik, A. A.; Iikubo, S.; Kodama, K.; Igawa, N.; Shamoto, S.; Niitaka, S.;
Azuma, M.; Shimakawa, Y.; Takano, M.; Izumi, F.; Takayama-Muromiachi, E.
Chem. Mater. 2006, 18, 798803.
28. Belik, A. A.; Iikubo, S.; Kodama, K.; Igawa, N.; Shamoto, S.; Maie, M.; Nagai, T.;
Matsui, Y.; Stefanovich, S. Yu.; Lazoryak, B. I.; Takayama-Muromachi, E. J. Am.
Chem. Soc. 2006, 128, 706707.
29. Glazer, A. M.; Mabud, S. A. Acta Crystallogr. 1978, B34, 10651070.
30. Kuroiwa, Y.; Aoyagi, S.; Sawada, A.; Harada, J.; Nishibori, E.; Takata, M.;
Sakata, M. Phys. Rev. Lett. 2001, 87, 217601.
31. Tsirlin, A. A.; Belik, A. A.; Shpanchenko, R. V.; Antipov, E. V.; Takayama-
Muromachi, E.; Rosner, H. Phys. Rev. B 2008, 77, 092402.
32. Oka, K.; Yamada, I.; Azuma, M.; Takeshita, S.; Satoh, K. H.; Koda, A.; Kadono, R.;
Takano, M.; Shimakawa, Y. Inorg. Chem. 2008, 47, 73557359.
33. Zeches, R. J.; Rossel, M. D.; Zhang, J. X.; Hatt, A. J.; He, Q.; Yang, C.-H.;
Kumar, A.; Wang, C. H.; Melville, A.; Adamo, C.; Sheng, G.; Chu, Y.-H.;
Ihlefeld, J. F.; Erni, R.; Ederer, C.; Gopalan, V.; Chen, L. Q.; Schlom, D. G.;
Spaldin, N. A.; Martin, L. W.; Ramesh, R. Science 2009, 326, 977980.
34. Bea, H.; Dupe, B.; Fusil, S.; Mattana, R.; Jacquet, E.; Warot-Fonrose, B.;
Wilhelm, F.; Rogalev, A.; Petit, S.; Cros, V.; Anane, A.; Petroff, F.;
Bouzehouane, K.; Geneste, G.; Dkhil, B.; Lisenkov, S.; Ponomareva, I.;
Bellaiche, L.; Bibes, M.; Barthelemy, A. Phys. Rev. Lett. 2009, 102, 217603.
35. Hatt, A. J.; Spaldin, N. A.; Ederer, C. Phys. Rev. B 2010, 81, 054109.
36. Belik, A. A.; Iikubo, S.; Kodama, K.; Igawa, N.; Shamoto, S.; Takayama-
Muromachi, E. Chem. Mater. 2008, 20, 37653769.
37. Kimura, T.; Kawamoto, S.; Yamada, I.; Azuma, M.; Takano, M.; Tokura, Y. Phys.
Rev. B 2003, 67, 180401(R).
38. Montanari, E.; Calestani, G.; Migliori, A.; Dapiaggi, M.; Bolzoni, F.; Cabassi, R.;
Gilioli, E. Chem. Mater. 2005, 17, 64576467.
39. Palai, R.; Katiyar, R. S.; Schmid, H.; Tissot, P.; Clark, S. J.; Robertson, J.;
Redfern, S. A. T.; Catalan, G.; Scott, J. F. Phys. Rev. B 2008, 77, 014110.
40. Arnold, D. C.; Knight, K. S.; Morrison, F. D.; Lightfoot, P. Phys. Rev. Lett. 2009,
102, 027602.
41. Arnold, D. C.; Knight, K. S.; Catalan, G.; Redfern, S. A. T.; Scott, J. F.;
Lightfoot, P.; Morrison, F. D. Adv. Funct. Mater. 2010, 20, 21162123.
42. Catalan, G.; Scott, J. F. Adv. Mater. 2009, 21, 24632485.
43. Montanari, E.; Calestani, G.; Righi, L.; Gilioli, E.; Bolzoni, F.; Knight, K. S.;
Radaelli, P. G. Phys. Rev. B 2007, 75, 220101(R).
44. Belik, A. A.; Iikubo, S.; Yokosawa, T.; Kodama, K.; Igawa, N.; Shamoto, S.;
Azuma, M.; Takano, M.; Kimoto, K.; Matsui, Y.; Takayama-Muromachi, E. J. Am.
Chem. Soc. 2007, 129, 971977.
45. Baettig, P.; Seshadri, R.; Spaldin, N. A. J. Am. Chem. Soc. 2007, 129,
98549855.
46. Kodama, K.; Iikubo, S.; Shamoto, S.; Belik, A. A.; Takayama-Muromachi, E.
J. Phys. Soc. Jpn. 2007, 76, 124605.
47. Yokosawa, T.; Belik, A. A.; Asaka, T.; Kimoto, K.; Takayama-Muromachi, E.;
Matsui, Y. Phys. Rev. B 2008, 77, 024111.
48. Moreira des Santos, A.; Parashar, S.; Raju, A. R.; Zhao, Y. S.; Cheetham, A. K.;
Rao, C. N. R. Solid State Commun. 2002, 122, 4952.
49. Belik, A. A.; Kolodiazhnyi, T.; Kosuda, K.; Takayama-Muromachi, E. J. Mater.
Chem. 2009, 19, 15931600.
50. Belik, A. A.; Kodama, K.; Igawa, N.; Shamoto, S.; Kosuda, K.; Takayama-
Muromachi, E. J. Am. Chem. Soc. 2010, 132, 81378144.
51. Belik, A. A.; Matsushita, Y.; Tanaka, M.; Takayama-Muromachi, E. Inorg. Chem.
2011, 50, 76857689.
52. Corker, D. L.; Glazer, A. M.; Dec, J.; Roleder, K.; Whatmore, R. W. Acta
Crystallogr. 1997, B53, 135142.
53. Yamasaki, K.; Soejima, Y.; Fischer, K. F. Acta Crystallogr. 1998, B54, 524530.
54. Teslic, S.; Egami, T. Acta Crystallogr. 1998, B54, 750765.
55. Corker, D. L.; Glazer, A. M.; Kaminsky, W.; Whatmore, R. W.; Dec, J.; Roleder, K.
Acta Crystallogr. 1998, B54, 1828.
56. Ghosez, Ph.; Cockayne, E.; Waghmare, U. V.; Rabe, K. M. Phys. Rev. B 1999, 60,
836843.
57. Ghita, M.; Fornari, M.; Singh, D. J.; Halilov, S. V. Phys. Rev. B 2005, 72, 054114.
58. Rusakov, D. A.; Abakumov, A. M.; Yamaura, K.; Belik, A. A.; Van Tendeloo, G.;
Takayama-Muromachi, E. Chem. Mater. 2011, 23, 285292.
59. Levin, I.; Tucker, M. G.; Wu, H.; Provenzano, V.; Dennis, C. L.; Karimi, S.;
Comyn, T.; Stevenson, T.; Smith, R. I.; Reaney, I. M. Chem. Mater. 2011, 23,
21662175.
60. Mandal, P.; Sundaresan, A.; Rao, C. N. R.; Iyo, A.; Shirage, P. M.; Tanaka, Y.;
Simon, Ch.; Pralong, V.; Lebedev, O. I.; Caignaert, V.; Raveau, B. Phys. Rev. B
2010, 82, 100416(R).
Tendeloo, G.; Takayama-Muromachi, E. Chem. Mater. 2011, 23, 45054514.
62. Inaguma, Y.; Greneche, J.-M.; Crosnier-Lopez, M.-P.; Katsumata, T.; Calage, Y.;
Fourquet, J.-L. Chem. Mater. 2005, 17, 13861390.
63. Xiao, W.; Tan, D.; Xiong, X.; Liu, J.; Xu, J. Proc. Natl. Acad. Sci. 2010, 107,
1402614029.
64. Oka, K.; Azuma, M.; Chen, W.; Yusa, H.; Belik, A. A.; Takayama-Muromachi, E.;
Mizumaki, M.; Ishimatsu, N.; Hiraoka, N.; Tsujimoto, M.; Tucker, M. G.;
Attfield, J. P.; Shimakawa, Y. J. Am. Chem. Soc. 2010, 132, 94389443.
65. Haumont, R.; Bouvier, P.; Pashkin, A.; Rabia, K.; Frank, S.; Dkhil, B.;
Crichton, W. A.; Kuntscher, C.; Kriesel, J. Phys. Rev. B 2009, 79, 184110.
66. Belik, A. A.; Yusa, H.; Hirao, N.; Ohishi, Y.; Takayama-Muromachi, E. Chem.
Mater. 2009, 21, 34003405.
67. Belik, A. A.; Yusa, H.; Hirao, N.; Ohishi, Y.; Takayama-Muromachi, E. Inorg.
Chem. 2009, 48, 10001004.
68. Guennou, M.; Bouvier, P.; Chen, G. S.; Dkhil, B.; Haumont, R.; Garbarino, G.;
Kreisel, J. Phys. Rev. B 2011, 84, 174107.
69. Gavriliuk, A. G.; Struzhkin, V. V.; Lyubutin, I. S.; Ovchinnikov, S. G.; Hu, M. Y.;
Chow, P. Phys. Rev. B 2008, 77, 155112.
70. Ahart, M.; Somayazulu, M.; Cohen, R. E.; Ganesh, P.; Dera, P.; Mo, H.;
Hemley, R. J.; Ren, Y.; Liermann, P.; Wu, Z. Nature 2008, 451, 545549.
71. Janolin, P.-E.; Bouvier, P.; Kreisel, J.; Thomas, P. A.; Kornev, I. A.; Bellaiche, L.;
Crichton, W.; Hanfland, M.; Dkhil, B. Phys. Rev. Lett. 2008, 101, 237601.
72. Jaouen, N.; Dhaussy, A. C.; Itie, J. P.; Rogalev, A.; Marinel, S.; Joly, Y. Phys. Rev.
B 2007, 75, 224115.
73. Kornev, I. A.; Bellaiche, L.; Bouvier, P.; Janolin, P.-E.; Dkhil, B.; Kreisel, J. Phys.
Rev. Lett. 2005, 95, 196804.
74. Wang, K. F.; Liu, J.-M.; Ren, Z. F. Adv. Phys. 2009, 58, 321448.
75. Hill, N. A. J. Phys. Chem. 2000, 104, 66946709.
76. Khomskii, D. I. J. Magn. Magn. Mater. 2006, 306, 18.
77. Azuma, M.; Takata, K.; Saito, T.; Ishiwata, S.; Shimakawa, Y.; Takano, M. J. Am.
Chem. Soc. 2005, 127, 88898892.
78. Hughes, H.; Allix, M. M. B.; Bridges, C. A.; Claridge, J. B.; Kuang, X.; Niu, H.;
Taylor, S.; Song, W.; Rosseinsky, M. J. J. Am. Chem. Soc. 2005, 127,
1379013791.
79. Bridges, C. A.; Allix, M.; Suchomel, M. R.; Kuang, X.; Sterianou, I.; Sinclair, D. C.;
Rosseinsky, M. J. Angew. Chem. Int. Ed. 2007, 46, 87858789.

80. Claridge, J. B.; Hughes, H.; Bridges, C. A.; Allix, M.; Suchomel, M. R.; Niu, H.;
Kuang, X.; Rosseinsky, M. J.; Bellido, N.; Grebille, D.; Perez, O.; Simon, C.;
Pelloquin, D.; Blundell, S. J.; Lancaster, T.; Baker, P. J.; Pratt, F. L.;
Halasyamani, P. S. J. Am. Chem. Soc. 2009, 131, 1400014017.
81. Ok, K. M.; Halasyamani, P. S.; Casanova, D.; Llunell, M.; Alemany, P.; Alvarez, S.
Chem. Mater. 2006, 18, 31763183.
82. Pearson, R. G. J. Mol. Struct. (Theochem) 1983, 103, 2534.
83. Rondinelli, J. M.; Eidelson, A. S.; Spaldin, N. A. Phys. Rev. B 2009, 79, 205119.
84. Cohen, R. E.; Krakauer, H. Phys. Rev. B 1990, 42, 64166423.
85. Cohen, R. E. Nature 1992, 358, 136138.
86. Posternak, M.; Resta, R.; Baldereschi, A. Phys. Rev. B 1994, 50, 89118914.
87. Filippetti, A.; Hill, N. A. Phys. Rev. B 2002, 65, 195120.
88. Bhattacharjee, S.; Bousquet, E.; Ghosez, P. Phys. Rev. Lett. 2009, 102,
117602.
89. Lee, J. H.; Rabe, K. M. Phys. Rev. Lett. 2010, 104, 207204.
90. Ederer, C.; Harris, T.; Kovacik, R. Phys. Rev. B 2011, 83, 054110.
91. Sakai, H.; Fujioka, J.; Fukuda, T.; Okuyama, D.; Hashizume, D.; Kagawa, F.;
Nakao, H.; Murakami, Y.; Arima, T.; Baron, A. Q. R.; Taguchi, Y.; Tokura, Y. Phys.
Rev. Lett. 2011, 107, 137601.
92. Stokes, H. T.; Kisi, E. H.; Hatch, D. M.; Howard, C. J. Acta Crystallogr. 2002, B58,
934938.
93. Kennedy, B. J.; Howard, C. J.; Chakoumakos, B. C. J. Phys. Condens. Matter
1999, 11, 14791488.
94. Ali, R.; Yashima, M. J. Solid State Chem. 2005, 178, 28672872.
95. Guennou, M.; Bouvier, P.; Krikler, B.; Kreisel, J.; Haumont, R.; Garbarino, G.
Phys. Rev. B 2010, 82, 134101.
96. Kennedy, B. J.; Zhou, Q.; Avdeev, M. J. Solid State Chem. 2011, 184,
29872993.
97. Moriwake, H.; Kuwabara, A.; Fisher, C. A. J.; Taniguchi, H.; Itoh, M.; Tanaka, I.
Phys. Rev. B 2011, 84, 104114.
98. Taniguchi, H.; Soon, H. P.; Shimizu, T.; Moriwake, H.; Shan, Y. J.; Itoh, M. Phys.
Rev. B 2011, 84, 174106.
99. Hewat, A. W. J. Phys. C 1973, 6, 25592572.
100. Kwei, G. H.; Lawson, A. C.; Billinge, S. J. L.; Cheong, S.-W. J. Phys. Chem. 1993,
97, 23682377.
101. Cochran, W. Phys. Rev. Lett. 1969, 3, 412414.
102. Zhong, W.; Vanderbilt, D.; Rabe, K. M. Phys. Rev. Lett. 1994, 73, 18611864.
103. Comes, P. R.; Lambert, M.; Guinier, A. Acta Crystallogr. 1970, A26, 244254.
104. Shirane, G. Rev. Mod. Phys. 1974, 46, 437449.
105. Zalar, B.; Laguta, V. V.; Blinc, R. Phys. Rev. Lett. 2003, 90, 037601.
106. Pasciak, M.; Welberry, R. Zeitschrift fur Kristallographie 2011, 226, 113125. 153. Gehring, G. A.; Gehring, K. A. Rep. Prog. Phys. 1975, 38, 189.
107. Stern, E. A. Phys. Rev. Lett. 2004, 93, 037601.
108. Bussmann-Holder, A.; Beige, H.; VolkI, G. Phys. Rev. B 2009, 79, 184111.
109. Mishra, S. K.; Choudhury, N.; Chaplot, S. L.; Krishna, P. S. R.; Mittal, R. Phys.
Rev. B 2007, 76, 024110.
110. Johnston, K. E.; Tang, C. C.; Parker, J. E.; Knight, K. S.; Lightfoot, P.;
Ashbrook, S. E. J. Am. Chem. Soc. 2010, 132, 87328746.
111. Mishra, S. K.; Mittal, R.; Pomjakushin, V. Yu.; Chaplot, S. L. Phys. Rev. B 2011,
83, 134105.
112. Sciau, Ph.; Kania, A.; Dkhil, B.; Suard, E.; Ratuszna, A. J. Phys. Condens. Matter
2004, 16, 27952810.
113. Levin, I.; Kryazman, V.; Woicik, J. C.; Karapetrova, J.; Proffen, T.; Tucker, M. G.;
Reaney, I. M. Phys. Rev. B 2009, 79, 104113.
114. Yashima, M.; Matsuyama, S.; Sano, R.; Itoh, M.; Tsuda, K.; Fu, D. Chem. Mater.
2011, 23, 16431645.
115. Shirane, G.; Yamada, Y. Phys. Rev. 1969, 177, 858863.
116. Yamada, Y.; Shirane, G. J. Phys. Soc. Jpn. 1969, 26, 396403.
117. Cowley, R. A. Philos. Trans. R. Soc. Lond. A 1996, 354, 27992814.
118. Zhong, W.; Vanderbilt, D. Phys. Rev. B 1996, 53, 50475050.
119. Muller, K. A.; Burkard, H. Phys. Rev. B 1979, 19, 35933602.
120. Uwe, H.; Sakudo, T. Phys. Rev. B 1976, 13, 271286.
121. Itoh, M.; Wang, R.; Inaguma, Y.; Yamaguchi, T.; Shan, Y.-J.; Nakamura, T. Phys.
Rev. Lett. 1999, 82, 35403543.
122. Blinc, R.; Zalar, B.; Laguta, V. V.; Itoh, M. Phys. Rev. Lett. 2005, 94, 147601.
123. Shigenari, T.; Abe, K.; Takemoto, T.; Sanaka, O.; Akaike, T.; Sakai, Y.; Wang, R.;
Itoh, M. Phys. Rev. B 2006, 74, 174121.
124. Taniguchi, H.; Itoh, M.; Yagi, T. Phys. Rev. Lett. 2007, 99, 017602.
125. Bartkowiak, M.; Kearley, G. J.; Yethiraj, M.; Mulders, A. M. Phys. Rev. B 2011,
83, 064102.
126. Ishidate, T.; Abe, S.; Takahashi, H.; Mri, N. Phys. Rev. Lett. 1997, 78,
23972400.
127. Iniguez, J.; Vanderbilt, D. Phys. Rev. Lett. 2002, 89, 115503.
128. Bousquet, E.; Ghosez, P. Phys. Rev. B 2006, 74, 180101(R).
Transition-Metal Perovskites 35
129. Akbarzadeh, A. R.; Kornev, I.; Malibert, C.; Bellaiche, L.; Kiat, J. M. Phys. Rev. B
2005, 72, 205104.
130. Akbarzadeh, A. R.; Bellaiche, L.; Leung, K.; Iniguez, J.; Vanderbilt, D. Phys. Rev. B
2004, 70, 054103.
131. Fabricius, G.; Peltzer y Blanca, E. L.; Rodriguez, C. O.; Ayala, A. P.; de la Presa, P.;
Garca, A. L. Phys. Rev. B 1997, 55, 164168.
132. Lelieveld, R.; IJdo, D. J. W. Acta Crystallogr. 1980, B36, 22232226.
133. Lee, C.-S.; Kleinke, K. M.; Kleinke, H. Solid State Sci. 2005, 7, 10491054.
134. Bennett, J. W.; Grinberg, I.; Rappe, A. M. Phys. Rev. B 2009, 79, 235115.
135. Eng, H. W.; Barnes, P. W.; Auer, B. M.; Woodward, P. M. J. Solid State Chem.
2003, 175, 94109.
136. Hirota, K.; Wakimoto, S.; Cox, D. E. J. Phys. Soc. Jpn. 2006, 75, 111006.
137. Frantti, J. J. Phys. Chem. B 2008, 112, 65216535.
138. Kornev, I. A.; Bellaiche, L.; Janolin, P.-E.; Dkhil, B.; Suard, E. Phys. Rev. Lett.
2006, 97, 157601.
139. Yokota, H.; Zhang, N.; Taylor, A. E.; Thomas, P. A.; Glazer, A. M. Phys. Rev. B
2009, 80, 104109.
140. Phelan, D.; Long, X.; Xie, Y.; Ye, Z.-G.; Glazer, A. M.; Yokota, H.; Thomas, P. A.;
Gehring, P. M. Phys. Rev. Lett. 2010, 105, 207601.
141. Hinterstein, M.; Rouquette, J.; Haines, J.; Papet, Ph.; Knapp, M.; Glaum, J.;
Fuess, H. Phys. Rev. Lett. 2011, 107, 077602.
142. Schierholz, R.; Fuess, H. Phys. Rev. B 2011, 84, 064122.
143. Blinc, R.; Laguta, V.; Zalar, B. Phys. Rev. Lett. 2003, 91, 247601.
144. Jeong, I.-K.; Darling, T. W.; Lee, J. K.; Proffen, Th.; Heffner, R. H.; Park, J. S.;
Hong, K. S.; Dmowski, W.; Egami, T. Phys. Rev. Lett. 2005, 94, 147602.
145. Cowley, R. A.; Gvasaliya, S. N.; Lushnikov, S. G.; Roessli, B.; Rotaru, G. M. Adv.
Phys. 2011, 60, 229327.
146. Lente, M. H.; Guerra, J. D. S.; de Souza, G. K. S.; Fraygola, B. M.;
Raigoza, C. F. V.; Garcia, D.; Eiras, J. A. Phys. Rev. B 2008, 78, 054109.
147. Martinez, R.; Palai, R.; Huhtinen, H.; Liu, J.; Scott, J. F.; Katiyar, R. S. Phys. Rev. B
2010, 82, 134104.
148. Falqui, A.; Lampis, N.; Geddo-Lehmann, A.; Pinna, G. J. Phys. Chem. B 2005,
109, 2296722970.
149. Rotaru, G.; Roessli, B.; Amato, A.; Gvasaliya, S. N.; Mudry, C.; Lushnikov, S. G.;
Shaplygina, T. A. Phys. Rev. B 2009, 79, 184430.
150. Kleemann, W.; Shvartsman, V. V.; Borisov, P.; Kania, A. Phys. Rev. Lett. 2010,
105, 257202.
151. Jeong, I.-K.; Ahn, J. S.; Kim, B. G.; Yoon, S.; Singh, S. P.; Pandey, D. Phys. Rev.
B 2011, 83, 064108.
152. Burton, B. P. J. Phys. Chem. Solids 2000, 61, 327333.
154. Lufaso, M. W.; Woodward, P. M. Acta Crystallogr. 2004, B60, 1020.
155. Kugel, K. I.; Khomskii, D. I. Sov. Phys. Usp. 1982, 25, 231256.
156. Mostovoy, M. V.; Khomskii, D. I. Phys. Rev. Lett. 2004, 92, 167201.
157. van den Brink, J. New J. Phys. 2004, 6, 201.
158. Goodenough, J. B. Annu. Rev. Mater. Sci. 1998, 28, 127.
159. Alonso, J. A.; Martnez-Lope, M. J.; Casais, M. T.; Fernandez-Daz, M. T. Inorg.
Chem. 2000, 39, 917923.
160. Fauth, F.; Suard, E.; Caignaert, V. Phys. Rev. B 2001, 65, 060401(R).
161. Rodrguez-Carvajal, J.; Hennion, M.; Moussa, F.; Moudden, A. H.; Pinsard, L.;
Revcolevschi, A. Phys. Rev. B 1998, 57, R3189R3192.
162. Marezio, M.; Dernier, P. D. Mater. Res. Bull 1971, 6, 2329.
163. Zhou, J.-S.; Goodenough, J. B. Phys. Rev. Lett. 2005, 94, 065501.
164. Goodenough, J. B.; Zhou, J.-S. J. Mater. Chem. 2007, 17, 23942405.
165. Zhou, J.-S.; Goodenough, J. B. Phys. Rev. B 2008, 77, 132104.
166. Carpenter, M. A.; Howard, C. J. Acta Crystallogr. 2009, B65, 134146.
167. Stokes, H. T.; Hatch, D. M.; Campbell, B. J. (2007). ISOTROPY, http://stokes.byu.
edu/isotropy.html; ISODISTORT, http://stokes.byu.edu/cgi-bin/iso/isodistort.cgi.
168. Tamazyan, R.; van Smaalen, S. Acta Crystallogr. 2007, B63, 190200.
169. Murakami, Y.; Hill, J. P.; Gibbs, D.; Blume, M.; Koyama, I.; Tanaka, M.;
Kawata, H.; Arima, T.; Tokura, Y.; Hirota, K.; Endoh, Y. Phys. Rev. Lett. 1998, 81,
582585.
170. Moussa, F.; Hennion, M.; Rodriguez-Carvajal, J.; Moudden, H.; Pinsard, L.;
Revcolevschi, A. Phys. Rev. B 1996, 54, 1514915155.
171. Qiu, X.; Proffen, Th.; Mitchell, J. F.; Billinge, S. J. L. Phys. Rev. Lett. 2005, 94,
177203.
172. Sanchez, M. C.; Subas, G.; Garca, J.; Blasco, J. Phys. Rev. Lett. 2003, 90,
045503.
173. Chatterji, T.; Fauth, F.; Ouladdiaf, B.; Mandal, P.; Ghosh, B. Phys. Rev. B 2003,
68, 052406.
174. Pavarini, E.; Koch, E. Phys. Rev. Lett. 2010, 104, 086402.
175. Wall, S.; Prabkhakaran, D.; Boothroyd, A. T.; Cavalleri, A. Phys. Rev. Lett. 2009,
103, 097402.
36 Transition-Metal Perovskites
176. Loa, I.; Adler, P.; Grzechnik, A.; Syassen, K.; Schwarz, U.; Hanfland, M.;
Rozenberg, G. Kh.; Gorodetsky, P.; Pasternak, M. P. Phys. Rev. Lett. 2001, 87,
125501.
177. Pinsard-Gaudart, L.; Rodrguez-Carvajal, J.; Daoud-Aladine, A.; Goncharenko, I.;
Medarde, M.; Smith, R. I.; Revcolevschi, A. Phys. Rev. B 2001, 64, 064426.
178. Trimarchi, G.; Binggeli, N. Phys. Rev. B 2005, 71, 035101.
179. Ramos, A. Y.; Tolentino, H. C. N.; Souza-Neto, N. M.; Itie, J.-P.; Morales, L.;
Caneiro, A. Phys. Rev. B 2007, 75, 052103.
180. Baldini, M.; Struzhkin, V. V.; Goncharov, A. F.; Postorino, P.; Mao, W. L. Phys.
Rev. Lett. 2011, 106, 066402.
181. Tachibana, M.; Shimoyama, T.; Kawaji, H.; Atake, T.; Takayama-Muromachi, E.
Phys. Rev. B 2007, 75, 144425.
182. Mori, T.; Kamegashira, N.; Aoki, K.; Shishido, T.; Fukuda, T. Mater. Lett. 2002,
54, 238243.
183. Mori, T.; Aoki, K.; Kamegashira, N.; Shishido, T.; Fukuda, T. Mater. Lett. 2000,
42, 387389.
184. Sanchez, D.; Alonso, J. A.; Martnez-Lope, M. J. J. Chem. Soc. Dalton Trans.
2002, 44224425.
185. Munoz, A.; Casais, M. T.; Alonso, J. A.; Martnez-Lope, M. J.; Martnez, J. L.;
Fernandez-Daz, M. T. Inorg. Chem. 2001, 40, 10201028.
186. Huang, Y. H.; Fjellvag, H.; Karppinen, M.; Hauback, B. C.; Yamauchi, H.;
Goodenough, J. B. Chem. Mater. 2006, 18, 21302134.
187. Zhou, J.-S.; Goodenough, J. B.; Gallardo-Amores, J. M.; Moran, E.;
Alario-Franco, M. A.; Caudillo, R. Phys. Rev. B 2006, 74, 014422.
188. Ye, F.; Lorenz, B.; Huang, Q.; Wang, Y. Q.; Sun, Y. Y.; Chu, C. W.; Fernandez-
Baca, J. A.; Dai Pengcheng, ; Mook, H. A. Phys. Rev. B 2007, 76, 060402(R).
189. Okamoto, H.; Imamura, N.; Hauback, B. C.; Karppinen, M.; Yamauchi, H.;
Fjellvag, H. Solid State Commun. 2008, 146, 152156.
190. Uusi-Esko, K.; Malm, J.; Imamura, N.; Yamauchi, H.; Karppinen, M. Mater. Chem.
Phys. 2008, 112, 10291034.
191. Tsukuda, N.; Okazaki, A. J. Phys. Soc. Jpn. 1972, 33, 10881099.
192. Buttner, R. H.; Maslen, E. N.; Spadaccini, N. Acta Crystallogr. 1990, B46,
131138.
193. Hutchings, M. T.; Samuelsen, E. J.; Shirane, G.; Hirakawa, K. Phys. Rev. 1969,
188, 919923.
194. Hidaka, M.; Eguchi, T.; Yamada, I. J. Phys. Soc. Jpn. 1998, 67, 24882494.
195. Eremin, M. V.; Zakharov, D. V.; Krug von Nidda, H.-A.; Eremina, R. M.;
Shuvaev, A.; Pimenov, A.; Ghigna, P.; Deisenhofer, J.; Loidl, A. Phys. Rev. Lett.
2008, 101, 147601.
196. Deisenhofer, J.; Leonov, I.; Eremin, M. V.; Kant, Ch.; Ghigna, P.; Mayr, F.;
Iglamov, V. V.; Anisimov, V. I.; van der Marel, D. Phys. Rev. Lett. 2008, 101,
157406.
197. Kaiser, V.; Otto, M.; Binder, F.; Babel, D. Z. Anorg. Allg. Chem. 1990, 585,
93104.
198. Tong, J.; Lee, C.; Whangbo, M.-H.; Kremer, R. K.; Simon, A.; Kohler, J. Solid
State Sci. 2010, 12, 680684.
199. Pavarini, E.; Koch, E.; Lichtenstein, A. I. Phys. Rev. Lett. 2008, 101, 266405.
200. Lake, B.; Tennant, D. A.; Frost, C. D.; Nagler, S. E. Nat. Mater. 2005, 4, 329334.
201. Paolasini, L.; Caciuffo, R.; Sollier, A.; Ghigna, P.; Altarelli, M. Phys. Rev. Lett.
2002, 88, 106403.
202. Caciuffo, R.; Paolasini, L.; Sollier, A.; Ghigna, P.; Pavarini, E.; van den Brink, J.;
Altarelli, M. Phys. Rev. B 2002, 65, 174425.
203. Odenthal, R.-H.; Hoppe, R. Monat. Chem. 1971, 102, 13401350.
204. Mazej, Z.; Goreshnik, E.; Jaglicic, Z.; Gawe, B.; asocha, W.; Grzybowska, D.;
Jaron, T.; Kurzydowski, D.; Malinowski, P.; Kozminski, W.; Szydowska, J.;
Peszczynski, P.; Grochala, W. Cryst. Eng. Comm. 2009, 11, 17021710.
205. Zhang, X.; Zhang, G.; Jia, T.; Guo, Y.; Zeng, Z.; Lin, H. Q. Phys. Lett. A 2011, 375,
24562459.
206. Margadonna, S.; Karotsis, G. J. Mater. Chem. 2007, 17, 20132020.
207. Margadonna, S.; Karotsis, G. J. Am. Chem. Soc. 2006, 128, 1643616437.
208. Xiao, Y.; Su, Y.; Li, H.-F.; Kumar, C. M. N.; Mittal, R.; Persson, J.; Senyshyn, A.;
Gross, K.; Brueckel, Th. Phys. Rev. B 2010, 82, 094437.
209. Miyasaka, S.; Okimoto, Y.; Iwama, M.; Tokura, Y. Phys. Rev. B 2003, 68, 100406
(R).
210. Zhou, J.-S.; Ren, Y.; Yan, J.-Q.; Mitchell, J. F.; Goodenough, J. B. Phys. Rev. Lett.
2008, 100, 046401.
211. Tung, L. D.; Ivanov, A.; Schefer, J.; Lees, M. R.; Balakrishnan, G.; Paul, D. McK.
Phys. Rev. B 2008, 78, 054416.
212. Munoz, A.; Alonso, J. A.; Casais, M. T.; Martnez-Lope, M. J.; Martnez, J. L.;
Fernandez-Daz, M. T. Phys. Rev. B 2003, 68, 144429.
213. Noguchi, M.; Nakazawa, A.; Oka, S.; Arima, T.; Wakabayashi, Y.; Nakao, H.;
Murakami, Y. Phys. Rev. B 2000, 62, R9271R9274.
214. Blake, G. R.; Palstra, T. T. M.; Ren, Y.; Nugroho, A. A.; Menovsky, A. A. Phys. Rev.
Lett. 2001, 87, 245501.
215. Reehuis, M.; Ulrich, C.; Pattison, P.; Ouladdiaf, B.; Rheinstadter, M. C.; Ohl, M.;
Regnault, L. P.; Miyasaka, M.; Tokura, Y.; Keimer, B. Phys. Rev. B 2006, 73,
094440.
216. Miyasaka, S.; Yasue, T.; Fujioka, J.; Yamasaki, Y.; Okimoto, Y.; Kumai, R.;
Arima, T.; Tokura, Y. Phys. Rev. Lett. 2007, 99, 217201.
217. Reehuis, M.; Ulrich, C.; Prokes, K.; Matas, S.; Fujioka, J.; Miyasaka, S.;
Tokura, Y.; Keimer, B. Phys. Rev. B 2011, 83, 064404.
218. Sage, M. H.; Blake, G. R.; Nieuwenhuys, G. J.; Palstra, T. T. M. Phys. Rev. Lett.
2006, 96, 036401.
219. Sage, M. H.; Blake, G. R.; Marquina, C.; Palstra, T. T. M. Phys. Rev. B 2007, 76,
195102.
220. Munoz, A.; Alonso, J. A.; Casais, M. T.; Martnez-Lope, M. J.; Martnez, J. L.;
Fernandez-Daz, M. T. J. Mater. Chem. 2003, 13, 12341240.
221. Munoz, A.; Alonso, J. A.; Casais, M. T.; Martnez-Lope, M. J.; Martnez, J. L.;
Fernandez-Daz, M. T. Chem. Mater. 2004, 16, 15441550.
222. Castillo-Martnez, E.; Bieringer, M.; Shafi, S. P.; Cranswick, L. M. D.;
Alario-Franco, M.A. J. Am. Chem. Soc. 2011, 133, 85528563.
223. Mizokawa, T.; Khomskii, D. I.; Sawatzky, G. A. Phys. Rev. B 1999, 60, 73097313.
224. Horsch, P.; Oles, A. M.; Feiner, L. F.; Khaliullin, G. Phys. Rev. Lett. 2008, 100,
167205.
225. De Raychaudhury, M.; Pavarini, E.; Andersen, O. K. Phys. Rev. Lett. 2007, 99,
126402.
226. Bizen, D.; Nakatsuka, K.; Murata, T.; Nakao, H.; Murakami, Y.; Miyasaka, S.;
Tokura, Y. Phys. Rev. B 2008, 78, 224104.
227. Zhou, J.-S.; Goodenough, J. B.; Yan, J.-Q.; Cheng, J.-G.; Matsubayashi, K.;
Uwatoko, Y.; Ren, Y. Phys. Rev. B 2009, 80, 224422.
228. Mochizuki, M.; Imada, M. New J. Phys. 2004, 6, 154.
229. Arao, M.; Inoue, Y.; Koyama, Y. J. Phys. Chem. Solids 2002, 63, 995997.
230. Kiyama, T.; Itoh, M. Phys. Rev. Lett. 2003, 91, 167202.
231. Cwik, M.; Lorenz, T.; Baier, J.; Muller, R.; Andre, G.; Bouree, F.; Lichtenberg, F.;
Freimuth, A.; Schmitz, R.; Muller-Hartmann, E.; Braden, M. Phys. Rev. B 2003,
68, 060401(R).
232. Hemberger, J.; Krug von Nidda, H.-A.; Fritsch, V.; Deisenhofer, J.; Lobina, S.;
Rudolf, T.; Lunkenheimer, P.; Lichtenberg, F.; Loidl, A.; Bruns, D.; Buchner, B.
Phys. Rev. Lett. 2003, 91, 066403.
233. Haverkort, M. W.; Hu, Z.; Tanaka, A.; Ghiringhelli, G.; Roth, H.; Cwik, M.;
Lorenz, T.; Schuler-Langeheine, C.; Streltsov, S. V.; Mylnikova, A. S.;
Anisimov, V. I.; de Nadai, C.; Brookes, N. B.; Hsieh, H. H.; Lin, H.-J.; Chen, C. T.;
Mizokawa, T.; Taguchi, Y.; Tokura, Y.; Khomskii, D. I.; Tjeng, L. H. Phys. Rev. Lett.
2005, 94, 056401.
234. Ulrich, C.; Khaliullin, G.; Okamoto, S.; Reehuis, M.; Ivanov, A.; He, H.;
Taguchi, Y.; Tokura, Y.; Keimer, B. Phys. Rev. Lett. 2002, 89, 167202.
235. Nakao, H.; Wakabayashi, Y.; Kiyama, T.; Murakami, Y.; Zimmerman, M. V.; Hill, J. P.;
Gibbs, D.; Ishihara, S.; Taguchi, Y.; Tokura, Y. Phys. Rev. B 2002, 66, 184419.
236. Iga, F.; Tsubota, M.; Sawada, M.; Huang, H. B.; Kura, S.; Takemura, M.; Yaji, K.;
Nagira, M.; Kimura, A.; Jo, T.; Takabatake, T.; Namatame, H.; Taniguchi, M. Phys.
Rev. Lett. 2004, 93, 257207.
237. Komarek, A. C.; Roth, H.; Cwik, M.; Stein, W.-D.; Baier, J.; Kreiner, M.; Bouree, F.;
Lorenz, T.; Braden, M. Phys. Rev. B 2007, 75, 224402.
238. Keimer, B.; Casa, D.; Ivanov, A.; Lynn, J. W.; Zimmermann, M. V.; Hill, J. P.;
Gibbs, D.; Taguchi, Y.; Tokura, Y. Phys. Rev. Lett. 2000, 85, 39463949.
239. Khaliullin, G.; Maekawa, S. Phys. Rev. Lett. 2000, 85, 39503953.
240. Mochizuki, M.; Imada, M. Phys. Rev. Lett. 2003, 91, 167203.
241. Ulrich, C.; Gossling, A.; Gruninger, M.; Guennou, M.; Roth, H.; Cwik, M.;
Lorenz, T.; Khaliullin, G.; Keimer, B. Phys. Rev. Lett. 2006, 97, 157401.
242. Ulrich, C.; Ament, L. J. P.; Ghiringhelli, G.; Braicovich, L.; Moretti Sala, M.;
Pezzotta, N.; Schmitt, T.; Khaliullin, G.; van den Brink, J.; Roth, H.; Lorenz, T.;
Keimer, B. Phys. Rev. Lett. 2009, 103, 107205.
243. Macquart, R. B.; Kennedy, B. J.; Avdeev, M. J. Solid State Chem. 2010, 183,
249254.
244. Thorogood, G. J.; Avdeev, M.; Carter, M. L.; Kennedy, B. J.; Ting, J.;
Wallwork, K. S. Dalton Trans. 2011, 40, 72287233.
245. Kennedy, B. J.; Hunter, B. A.; Hester, J. R. Phys. Rev. B 2002, 65, 224103.
246. Kennedy, B. J.; Yamaura, K.; Takayama-Muromachi, E. J. Phys. Chem. Solids
2004, 65, 10651069.
247. Nagai, I.; Shirakawa, N.; Ikeda, S.; Iwasaki, R.; Nishimura, H.; Kosaka, M. Appl.
Phys. Lett. 2005, 87, 024105.
248. Rodriguez, E. E.; Poineau, F.; Llobet, A.; Kennedy, B. J.; Avdeev, M.;
Thorogood, G. J.; Carter, M. L.; Seshadri, R.; Singh, D. J.; Cheetham, A. K. Phys.
Rev. Lett. 2011, 106, 067201.

249. Allen, P. B.; Berger, H.; Chauvet, O.; Forro, L.; Jarlborg, T.; Junod, A.; Revaz, B.;
Santi, G. Phys. Rev. B 1996, 53, 43934398.
250. Yamaura, K.; Takayama-Muromachi, E. Phys. Rev. B 2001, 64, 224424.
251. Kimber, S. A. J.; Rodgers, J. A.; Wu, H.; Murray, C. A.; Argyriou, D. N.;
Fitch, A. N.; Khomskii, D. I.; Attfield, J. P. Phys. Rev. Lett. 2009, 102, 046409.
252. Cheng, J.-G.; Zhou, J.-S.; Goodenough, J. B. Phys. Rev. B 2009, 80, 174426.
253. Bouchard, R. J.; Gillson, J. L. Mater. Res. Bull. 1972, 7, 873878.
254. Goodenough, J. B.; Zhou, J.-S. Chem. Mater. 1998, 10, 29802993.
255. Mazin, I. I.; Khomskii, D. I.; Lengsdorf, R.; Alonso, J. A.; Marhsall, W. G.;
Ibberson, R. M.; Podlesnyak, A.; Martnez-Lope, M. J.; Abd-Elmeguid, M. M.
Phys. Rev. Lett. 2007, 98, 176406.
256. Takeda, T.; Yamaguchi, Y.; Watanabe, H. J. Phys. Soc. Jpn. 1972, 33, 967969.
257. Lebon, A.; Adler, P.; Bernhard, C.; Boris, A. V.; Pimenov, A. V.; Maljuk, A.;
Lin, C. T.; Ulrich, C.; Keimer, B. Phys. Rev. Lett. 2004, 92, 037202.
258. Mostovoy, M. Phys. Rev. Lett. 2005, 94, 137205.
259. Hayashi, N.; Yamamoto, T.; Kageyama, H.; Nishi, M.; Watanabe, Y.; Kawakami, T.;
Matsushita, Y.; Fujimori, A.; Takano, M. Angew. Chem. Int. Ed. 2011, 50,
1254712550.
260. Bocquet, A. E.; Fujimori, A.; Mizokawa, T.; Saitoh, T.; Namatame, H.; Suga, S.;
Kimizuka, N.; Takeda, Y.; Takano, M. Phys. Rev. B 1992, 45, 15611570.
261. Woodward, P. M.; Cox, D. E.; Moshopoulou, E.; Sleight, A. W.; Morimoto, S.
Phys. Rev. B 2000, 62, 844855.
262. Alonso, J. A.; Martnez-Lope, M. J.; Casais, M. T.; Aranda, M. A. G.; Fernandez-
Daz, M. T. J. Am. Chem. Soc. 1999, 121, 47544762.
263. Alonso, J. A.; Garca-Munoz, J. L.; Fernandez-Daz, M. T.; Aranda, M. A. G.; 297. Radaelli, P. G.; Cheong, S.-W. Phys. Rev. B 2002, 66, 094408.
Martnez-Lope, M. J.; Casais, M. T. Phys. Rev. Lett. 1999, 82, 38713874.
264. Alonso, J. A.; Martnez-Lope, M. J.; Casais, M. T.; Garca-Munoz, J. L.;
Fernandez-Daz, M. T.; Aranda, M. A. G. Phys. Rev. B 2001, 64, 094102.
265. Alonso, J. A.; Martnez-Lope, M. J.; Demazeau, G.; Fernandez-Daz, M. T.;
Presniakov, I. A.; Rusakov, V. S.; Gubaidulina, T. V.; Sobolev, A. V. Dalton Trans.
2008, 65846592.
266. Zaghrioui, M.; Bulou, A.; Lacorre, P.; Laffez, P. Phys. Rev. B 2001, 64,
081102(R).
267. Piamonteze, C.; Tolentino, H. C. N.; Ramos, A. Y.; Massa, N. E.; Alonso, J. A.;
Martnez-Lope, M. J.; Casais, M. T. Phys. Rev. B 2005, 71, 012104.
268. Acosta Alejandro, M.; Mustre de Leon, J.; Medarde, M.; Lacorre, Ph.; Konder, K.;
Montano, P. A. Phys. Rev. B 2008, 77, 085107.
269. Henry, P. F.; Weller, M. T.; Wilson, C. C. Chem. Mater. 2002, 14, 41044110.
270. Staub, U.; Meijer, G. I.; Fauth, F.; Allenspach, R.; Bednorz, J. G.; Karpinski, J.;
Kazakov, S. M.; Paolasini, L.; dAcapito, F. Phys. Rev. Lett. 2002, 88, 126402.
271. Scagnoli, V.; Staub, U.; Janousch, M.; Mulders, A. M.; Shi, M.; Meijer, G. I.;
Rosenkranz, S.; Wilkins, S. B.; Paolasini, L.; Karpinski, J.; Kazakov, S. M.;
Lovesey, S. W. Phys. Rev. B 2005, 72, 155111.
272. Scagnoli, V.; Staub, U.; Mulders, A. M.; Janousch, M.; Meijer, G. I.; Hammerl, G.;
Tonnerre, J. M.; Stojic, N. Phys. Rev. B 2006, 73, 100409(R).
273. Kim, S.-J.; Demazeau, G.; Alonso, J. A.; Choy, J.-H. J. Mater. Chem. 2001, 11,
487492.
274. Ishiwata, S.; Azuma, M.; Takano, M.; Nishibori, E.; Takata, M.; Sakata, M.; Kato, K.
J. Mater. Chem. 2002, 12, 37333737.
275. Carlsson, S. J. E.; Azuma, M.; Shimakawa, Y.; Takano, M.; Hewat, A.; Attfield, J. P.
J. Solid State Chem. 2008, 181, 611615.
276. Ishiwata, S.; Azuma, M.; Hanawa, M.; Moritomo, Y.; Ohishi, Y.; Kato, K.;
Takata, M.; Nishibori, E.; Sakata, M.; Terasaki, I.; Takano, M. Phys. Rev. B 2005,
72, 045104.
277. Azuma, M.; Carlsson, S.; Rodgers, J.; Tucker, M. G.; Tsujimoto, M.; Ishiwata, S.;
Isoda, S.; Shimakawa, Y.; Takano, P.; Attfield, J. P. J. Am. Chem. Soc. 2007,
129, 1443314436.
278. Li, J. Q.; Matsui, Y.; Park, S. K.; Tokura, Y. Phys. Rev. Lett. 1997, 79,
297300.
279. Park, S. K.; Ishikawa, T.; Tokura, Y.; Li, J. Q.; Matsui, Y. Phys. Rev. B 1999, 60,
1078810795.
280. McQueeney, R. J.; Ma, J.; Chang, S.; Yan, J.-Q.; Hehlen, M.; Trouw, F. Phys. Rev.
Lett. 2007, 98, 126402.
281. Herrero-Martin, J.; Subias, G.; Garcia, J.; Blasco, J.; Sanchez, M. C. Phys. Rev. B
2009, 79, 045121.
282. Rao, C. N. R.; Arulraj, A.; Santosh, P. N.; Cheetham, A. K. Chem. Mater. 1998, 10,
27142722.
283. Van Tendeloo, G.; Lebedev, O. I.; Hervieu, M.; Raveau, B. Rep. Prog. Phys. 2004,
67, 13151365.
284. Uehara, M.; Mori, S.; Chen, C. H.; Cheong, S.-W. Nature 1999, 399, 560563.
285. Radaelli, P. G.; Cox, D. E.; Marezio, M.; Cheong, S.-W. Phys. Rev. B 1997, 55,
30153023.
Transition-Metal Perovskites 37
286. Okuda, T.; Kimura, T.; Kuwahara, H.; Tomioka, Y.; Asamitsu, A.; Okimoto, Y.;
Saitoh, E.; Tokura, Y. Mater. Sci. Eng. 1999, B63, 163170.
287. Tokura, Y.; Nagaosa, N. Science 2000, 288, 462468.
288. Zhou, J.-S.; Jin, C.-Q.; Long, Y.-W.; Yang, L.-X.; Goodenough, J. B. Phys. Rev.
Lett. 2006, 96, 046408.
289. Bhobe, P. A.; Chainani, A.; Taguchi, M.; Eguchi, R.; Matsunami, M.; Ohtsuki, T.;
Ishizaka, K.; Okawa, M.; Oura, M.; Senba, Y.; Ohashi, H.; Isobe, M.; Ueda, Y.;
Shin, S. Phys. Rev. B 2011, 83, 165132.
290. Komarek, A. C.; Moller, T.; Isobe, M.; Drees, Y.; Ulbrich, H.; Azuma, M.;
Fernandez-Daz, M. T.; Senyshyn, A.; Hoelzel, M.; Andre, G.; Ueda, Y.;
Gruninger, M.; Braden, M. Phys. Rev. B 2011, 84, 125114.
291. Ortega-San-Martin, L.; Williams, A. J.; Rodgers, J.; Attfield, J. P.; Heymann, G.;
Huppertz, H. Phys. Rev. Lett. 2007, 99, 255701.
292. Komarek, A. C.; Streltsov, S. V.; Isobe, M.; Moller, T.; Hoelzel, M.; Senyshyn, A.;
Trots, D.; Fernandez-Daz, M. T.; Hansen, T.; Gotou, H.; Yagi, T.; Ueda, Y.;
Anisimov, V. I.; Gruninger, M.; Khomskii, D. I.; Braden, M. Phys. Rev. Lett. 2008,
101, 167204.
293. Qian, Y.; Wang, G.; Li, Z.; Jin, C. Q.; Fang, Z. New J. Phys. 2011, 13,
053002.
294. Lee, K.-W.; Pickett, W. E. Phys. Rev. B 2009, 80, 125133.
295. Streltsov, S. V.; Korotin, M. A.; Anisimov, V. I.; Khomskii, D. I. Phys. Rev. B 2008,
78, 054425.
296. Maris, G.; Ren, Y.; Volotchaev, V.; Zobel, C.; Lorenz, T.; Palstra, T. T. M. Phys.
Rev. B 2003, 67, 224423.
298. Korotin, M. A.; Ezhov, S. Yu.; Solovyev, I. V.; Anisimov, V. I.; Khomskii, D. I.;
Sawatzky, G. A. Phys. Rev. B 1996, 54, 53095316.
299. Vanko, G.; Rueff, J.-P.; Mattila, A.; Nemeth, Z.; Shukla, A. Phys. Rev. B 2006, 73,
024424.
300. Klie, R. F.; Zheng, J. C.; Zhu, Y.; Varela, M.; Wu, J.; Leighton, C. Phys. Rev. Lett.
2007, 99, 047203.
301. Haverkort, M. W.; Hu, Z.; Cezar, J. C.; Burnus, T.; Hartmann, H.; Reuther, M.;
Zobel, C.; Lorenz, T.; Tanaka, A.; Brookes, N. B.; Hsieh, H. H.; Lin, H.-J.;
Chen, C. T.; Tjeng, L. H. Phys. Rev. Lett. 2006, 97, 176405.
302. Podlesnyak, A.; Streule, S.; Mesot, J.; Medarde, M.; Pomjakushina, E.;
Conder, K.; Tanaka, A.; Haverkort, M. W.; Khomskii, D. I. Phys. Rev. Lett. 2006,
97, 247208.
303. Hoch, M. J. R.; Nellutla, S.; van Tol, J.; Choi, E. S.; Lu, J.; Zheng, H.;
Mitchell, J. F. Phys. Rev. B 2009, 79, 214421.
304. Hozoi, L.; Birkenheuer, U.; Stoll, H.; Fulde, P. New J. Phys. 2009, 11, 023023.
305. Hsu, H.; Blaha, P.; Wentzcovitch, R. M.; Leighton, C. Phys. Rev. B 2010, 82,
100406(R).
306. Vogt, T.; Hriljac, J. A.; Hyatt, N. C.; Woodward, P. Phys. Rev. B 2003, 67,
140401(R).
307. Kozlenko, D. P.; Golosova, N. O.; Jirak, Z.; Dubrovinsky, L. S.; Savenko, B. N.;
Tucker, M. G.; Le Godec, Y.; Glazkov, V. P. Phys. Rev. B 2007, 75, 064422.
308. Yan, J.-Q.; Zhou, J.-S.; Goodenough, J. B. Phys. Rev. B 2004, 69, 134409.
309. Tachibana, M.; Yoshida, T.; Kawaji, H.; Atake, T.; Takayama-Muromachi, E. Phys.
Rev. B 2008, 77, 094402.
310. Anderson, M. T.; Greenwood, K. B.; Taylor, G. A.; Poeppelmeier, K. R. Prog. Solid
State Chem. 1993, 22, 197233.
311. King, G.; Woodward, P. M. J. Mater. Chem. 2010, 20, 57855796.
312. Azad, A. K.; Ivanov, S. A.; Eriksson, S.-G.; Eriksen, J.; Rundlof, H.; Mathieu, R.;
Svedlindh, P. Mater. Res. Bull. 2001, 36, 22152228.
313. Troyanchuk, I. O.; Bushinskii, M. V.; Szymczak, H.; Baran, M.; Barner, K. J. Magn.
Magn. Mater. 2007, 312, 470475.
314. Cussen, E. J.; Vente, J. F.; Battle, P. D.; Gibb, T. C. J. Mater. Chem. 1997,
7(3), 459463.
315. Faik, A.; Igartua, J. M.; Iturbe-Zabalo, E.; Cuello, G. J. J. Mol. Struct. 2010, 963,
145152.
316. Retuerto, M.; Alonso, J. M.; Garca-Hernandez, M.; Martnez-Lope, M. J. Solid
State Commun. 2006, 139, 1922.
317. Faik, A.; Igartua, J. M.; Gateshki, M.; Cuello, G. J. J. Solid State Chem. 2009,
182(7), 17171725.
318. Mizoguchi, H.; Woodward, P. M.; Byeon, S.-H.; Parise, J. B. J. Am. Chem. Soc.
2004, 126, 31753184.
319. Anderson, M. T.; Poeppelmeier, K. R. Chem. Mater. 1991, 3, 476482.
320. Anderson, M. T.; Poeppelmeier, K. R.; Gramsch, S. A.; Burdett, J. K. J. Solid State
Chem. 1993, 102, 164174.
321. Azuma, M.; Kaimori, S.; Takano, M. Chem. Mater. 1998, 10, 31243130.
322. Iwanaga, D.; Inaguma, Y.; Itoh, M. J. Solid State Chem. 1999, 147, 291295.
323. Gateshki, M.; Igartua, J. M. J. Phys. Condens. Matter 2003, 15, 67496757.
38 Transition-Metal Perovskites
324. Gateshki, M.; Igartua, J. M.; Hernandez-Bocanegra, E. J. Phys. Condens. Matter
2003, 15, 61996217.
325. Abakumov, A. M.; Rossell, M. D.; Seryakov, S. A.; Rozova, M. G.; Markina, M. M.;
Van Tendeloo, G.; Antipov, E. V. J. Mater. Chem. 2005, 15, 48994905.
326. Yang, T.; Perkisas, T.; Hadermann, J.; Croft, M.; Ignatov, A.; Van Tendeloo, G.;
Greenblatt, M. J. Mater. Chem. 2011, 21, 199205.
327. Babel, D.; Haegele, R.; Pausewang, G.; Wall, F. Mater. Res. Bull. 1973, 8,
13711382.
328. Flerov, I. N.; Gorev, M. V.; Aleksandrov, K. S.; Tressaud, A.; Grannec, J.;
Couzi, M. Mater. Sci. Eng. 1998, 24, 81151.
329. Howard, C. J.; Kennedy, B. J.; Woodward, P. M. Acta Crystallogr. 2003, B59,
463471.
330. Aleksandrov, K. S.; Misjul, S. V. Sov. Phys. Crystallogr. 1981, 26, 612618.
331. Davies, P. K.; Wu, H.; Borisevich, A. Y.; Molodetsky, I. E.; Farber, L. Annu. Rev.
Mater. Res. 2008, 38, 369401.
332. Park, J.-H.; Woodward, P. M. Int. J. Inorg. Mater. 2000, 2, 153166.
333. Borisevich, A. Y.; Davies, P. K. J. Solid State Chem. 2003, 170, 198201.
334. Wu, H.; Davies, P. K. J. Solid State Chem. 2004, 177, 34693478.
335. Wu, H.; Davies, P. K. J. Solid State Chem. 2004, 177, 43054315.
336. Subramanian, M. A. J. Solid State Chem. 1994, 111, 134140.
337. Alonso, J. A.; Mzayek, E.; Rasines, I. Mater. Res. Bull. 1987, 22, 6974.
338. Nguyen, T. N.; Giaquinta, D. M.; Davis, W. M.; Zur Loye, H. C. Chem. Mater.
1993, 5, 12731276.
339. Abakumov, A. M.; King, G.; Laurinavichute, V. K.; Rozova, M. G.;
Woodward, P. M.; Antipov, E. V. Inorg. Chem. 2009, 48, 93369344.
340. Zuniga, F. J.; Tressaud, A.; Darriet, J. J. Solid State Chem. 2006, 179,
36073614.
341. Withers, R. L.; Welberry, T. R.; Brink, F. J.; Noren, L. J. Solid State Chem. 2003,
170, 211220.
342. King, G.; Abakumov, A. M.; Hadermann, J.; Alekseeva, A. M.; Rozova, M. G.;
Perkisas, T.; Woodward, P. M.; Van Tendeloo, G.; Antipov, E. V. Inorg. Chem.
2010, 49, 60586095.
343. Stoger, B.; Weil, M.; Zobetz, E. Z. Kristallogr. 2010, 225, 125138.
344. Jeitschko, W.; Mons, H. A.; Rodewald, U. C.; Moller, M. H. Z. Naturforsch. 1998,
53, 3136.
345. Bramnik, K. G.; Miehe, G.; Ehrenberg, H.; Fuess, H.; Abakumov, A. M.;
Shpanchenko, R. V.; Pomjakushin, V. Yu.; Balagurov, A. M. J. Solid State Chem.
2000, 149, 4955.
346. Tomaszewska, A.; Muller-Buschbaum, H. Z. Anorg. Allg. Chem. 1993, 619,
17381742.
347. King, G.; Abakumov, A. M.; Woodward, P. M.; Llobet, A.; Tsirlin, A. A.; Batuk, D.;
Antipov, E. V. Inorg. Chem. 2011, 50, 77927801.
348. Bharathy, M.; zur Loye, H. C. J. Solid State Chem. 2008, 181, 27892795.
349. Macquart, R. B.; Zhou, Q.; Kennedy, B. J. J. Solid State Chem. 2009, 182,
16911693.
350. Mitchell, R. H.; Chakhmouradian, A. R. J. Solid State Chem. 1998, 138,
307312.
351. Ducau, M.; Suh, K. S.; Senegas, J.; Darriet, J. Mater. Res. Bull. 1992, 27,
11151123.
352. Darriet, J.; Mayorga, S. G.; Tressaud, A. Eur. J. Solid State Inorg. Chem. 1990,
27, 783790.
353. Knapp, M. C.; Woodward, P. M. J. Solid State Chem. 2006, 179, 10651074.
354. Dilanian, R. A.; Yamamoto, A.; Izumi, F.; Kamiyama, T. Mol. Cryst. Liquid Cryst.
2000, 341, 10291034.
355. Kennedy, B. J.; Howard, C. J.; Kubota, Y.; Kato, K. J. Solid State Chem. 2004,
177, 45524556.
356. Abakumov, A. M.; Shpanchenko, R. V.; Antipov, E. V. Mater. Res. Bull. 1995, 30,
97103.
357. Zhou, Q.; Saines, P. J.; Sharma, N.; Ting, J.; Kennedy, B. J.; Zhang, Z.;
Withers, R. L.; Wallwork, K. S. Chem. Mater. 2008, 20, 66666676.
358. Ruiz-Trejo, E.; Gomez Garca, J. F. ECS Trans., Electrochem. Soc. 2008, 15,
99110.
359. Zhang, Z.; Howard, C. J.; Kennedy, B. J.; Knight, K. S.; Zhou, Q. J. Solid State
Chem. 2007, 180, 18461851.
360. Roudeau, S.; Weill, F.; Pechev, S.; Bassat, J. M.; Grenier, J. C. Comptes Rendus
Chimie 2008, 11, 734740.
361. Howard, C. J.; Zhang, Z. J. Phys. Condens. Matter 2003, 15, 45434553.
362. Yoshii, K. J. Alloys Compd. 2000, 305, 7275.
363. Yoshii, K. J. Solid State Chem. 2000, 149, 354359.
364. Sefat, A. S.; Amow, G.; Wu M,; Botton, G. A.; Greedan, J. E. J. Solid State Chem.
2005, 178, 10081016.
365. Inaguma, Y.; Katsumata, T.; Itoh, M.; Morii, Y.; Tsurui, T. Solid State Ion. 2006,
177, 30373044.
366. Garca-Martn, S.; Alario-Franco, M. A.; Ehrenberg, H.; Rodrguez-Carvajal, J.;
Amador, U. J. Am. Chem. Soc. 2004, 126, 35873596.
367. Sanz, J.; Varez, A.; Alonso, J. A.; Fernandez, M. T. J. Solid State Chem. 2004,
177, 11571164.
368. Inaguma, Y.; Katsumata, T.; Itoh, M.; Morii, Y. J. Solid State Chem. 2002, 166,
6772.
369. Yashima, M.; Itoh, M.; Inaguma, Y.; Morii, Y. J. Am. Chem. Soc. 2005, 127,
34913495.
370. Lopez, M. L.; Veiga, M. L.; Pico, C. J. Mater. Chem. 1994, 4, 547550.
371. Arillo, M. A.; Gomez, J.; Lopez, M. L.; Pico, C.; Veiga, M. L. J. Mater. Chem.
1997, 7, 801806.
372. King, G.; Thimmaiah, S.; Dwivedi, A.; Woodward, P. M. Chem. Mater. 2007, 19,
64516458.
373. King, G.; Wayman, L. M.; Woodward, P. M. J. Solid State Chem. 2009, 182,
13191325.
374. Azad, A. K.; Eriksson, S. G.; Mellergard, A.; Ivanov, S. A.; Eriksen, J.; Rundlof, H.
Mater. Res. Bull. 2002, 37, 17971813.
375. Azad, A. K.; Ivanov, S. A.; Eriksson, S.-G.; Eriksen, J.; Rundlof, H.; Mathieu, R.;
Svedlindh, P. Mater. Res. Bull. 2001, 36, 24852496.
376. Gateshki, M.; Igartua, J. M. J. Phys. Condens. Matter 2004, 16, 66396649.
377. Mitchell, R. H.; Chakhmouradian, A. R.; Woodward, P. M. Phys. Chem. Miner.
2000, 27, 583589.
378. Bochu, B.; Deschizeaux, M. N.; Joubert, J. C.; Collomb, A.; Chenavas, J.;
Marezio, M. J. Solid State Chem. 1979, 29, 291298.
379. Subramanian, M. A.; Sleight, A. W. Solid State Sci. 2002, 4, 347351.
380. Prodi, A.; Gilioli, E.; Gauzzi, A.; Licci, F.; Marezio, M.; Bolzoni, F.; Huang, Q.;
Santoro, A.; Lynn, J. W. Nat. Mater. 2004, 3, 4852.
381. Yamada, I.; Takata, K.; Hayashi, N.; Shinohara, S.; Azuma, M.; Mori, S.; Muranaka, S.;
Shimakawa, Y.; Takano, M. Angew. Chem. Int. Ed. 2008, 47, 70327035.
382. Shiraki, H.; Saito, T.; Shimakawa, Y. Chem. Mater. 2008, 20, 70777080.
383. Ebbinghaus, S. G.; Weidenkaff, A.; Cava, R. J. J. Solid State Chem. 2002, 167,
126136.
384. Byeon, S.-H.; Lee, S.-S.; Parise, J. B.; Woodward, P. M.; Hur, N. H. Chem. Mater.
2004, 16, 36973701.
385. Byeon, S.-H.; Lee, S.-S.; Parise, J. B.; Woodward, P. M.; Hur, N. H. Chem. Mater.
2005, 17, 35523557.
386. Ozaki, Y.; Ghedira, M.; Chenavas, J.; Joubert, J. C.; Marezio, M. Acta Crystallogr.
B 1977, 33, 36153617.
387. Long, Y. W.; Hayashi, N.; Saito, T.; Azuma, M.; Muranaka, S.l; Shimakawa, Y.
Nature 2009, 458, 6063.
388. Mezzadri, F.; Calestani, G.; Calicchio, M.; Gilioli, E.; Bolzoni, F.; Cabassi, R.;
Marezio, M.; Migliori, A. Phys. Rev. B 2009, 79, 100106(R).
389. Kadyrova, N. I.; Tyutyunnik, A. P.; Zubkov, V. G.; Zakharova, G. S.; Volkov, V. L.;
Dyachkova, T. V.; Zainulin, Yu.G. Zhurnal Neorganicheskoi Khimii 2003, 48,
19411946.
390. Byeon, S. H.; Lufaso, M. W.; Parise, J. B.; Woodward, P. M.; Hansen, T. Chem.
Mater. 2003, 15, 37983804.
391. Leinenweber, K.; Parise, J. J. Solid State Chem. 1995, 114, 277281.
392. Hodges, J. P.; Short, S.; Jorgensen, J. D.; Xiong, X.; Dabrowski, B.; Mini, S. M.;
Kimball, C. W. J. Solid State Chem. 2000, 151, 190209.
393. Greaves, C.; Jacobson, A. J.; Tofield, B. C.; Fender, B. E. F. Acta Crystallogr. B
1975, 31, 641646.
394. Tsujimoto, Y.; Tassel, C.; Hayashi, N.; Watanabe, T.; Kageyama, H.;
Yoshimura, K.; Takano, M.; Ceretti, M.; Ritter, C.; Paulus, W. Nature 2007, 450,
10621065.
395. Suescun, L.; Chmaissem, O.; Mais, J.; Dabrowski, B.; Jorgensen, J. D. J. Solid
State Chem. 2007, 180, 16981707.
396. Suescun, L.; Dabrowski, B.; Mais, J.; Remsen, S.; Richardson, J. W.; Maxey, E. R.,
Jr.; Jorgensen, J. D. Chem. Mater. 2008, 20, 16361645.
397. Suescun, L.; Dabrowski, B. Acta Crystallogr. 2008, B64, 177186.
398. Caignaert, V.; Nguyen, N.; Hervieu, M.; Raveau, B. Mater. Res. Bull. 1985, 20,
479484.
399. Stlen, S.; Bakken, E.; Mohn, C. E. Phys. Chem. Chem. Phys. 2006, 8,
429447.
400. Hadermann, J.; Van Tendeloo, G.; Abakumov, A. M. Acta Crystallogr. 2005, A61,
7792.
401. Otzschi, K.; Ueda, Y. J. Solid State Chem. 1993, 107, 149158.
402. Hiroi, Z. J. Solid State Chem. 1996, 123, 223235.
403. Er-Rakho, L.; Michel, C.; Raveau, B. J. Solid State Chem. 1988, 73, 514519.
404. Anderson, M. T.; Vaughey, J. T.; Poeppelmeier, K. R. Chem. Mater. 1993, 5,
151165.
405. Redhammer, G. J.; Tippelt, G.; Roth, G.; Amthauer, G. Am. Mineral. 2004, 89,
405420.

406. Schmidt, M.; Campbell, S. J. J. Solid State Chem. 2001, 156, 292304.
407. de la Calle, C.; Aguadero, A.; Alonso, J. A.; Fernandez-Daz, M. T. Solid State Sci.
2008, 10, 19241935.
408. Munoz, A.; de la Calle, C.; Alonso, J. A.; Botta, P. M.; Pardo, V.; Baldomir, D.;
Rivas, J. Phys. Rev. B 2008, 78, 054404.
409. Battle, P. D.; Gibb, T. C.; Lightfoot, P. J. Solid State Chem. 1988, 76, 334339.
410. Harrison, W. T. A.; Lee, T. H.; Yang, Y. L.; Scarfe, D. P.; Liu, L. M.; Jacobson, A. J.
Mater. Res. Bull. 1995, 30, 621630.
411. Casey, P. S.; Barker, D.; Hayward, M. A. J. Solid State Chem. 2006, 179,
13751382.
412. Parsons, T. G.; DHondt, H.; Hadermann, J.; Hayward, M. A. Chem. Mater. 2009,
21, 55275538.
413. Abakumov, A. M.; Rozova, M. G.; Antipov, E. V. Russ. Chem. Rev. 2004, 73,
847860.
414. Streltsov, V. A.; Ishizawa, N. Acta Crystallogr. B 1999, 55, 17.
415. Colville, A. A.; Geller, S. Acta Crystallogr. B 1971, 27, 23112315.
416. Arpe, R.; Muller-Buschbaum, Hk.; von Schenck, R. V. Z. Anorg. Allg. Chem.
1974, 410, 97103.
417. Abakumov, A. M.; Rozova, M. G.; Pavlyuk, B. Ph.; Lobanov, M. V.; Antipov, E. V.;
Lebedev, O. I.; Van Tendeloo, G.; Sheptyakov, D. V.; Balagurov, A. M.; Bouree, F.
J. Solid State Chem. 2001, 158, 100111.
418. Abakumov, A. M.; Rozova, M. G.; Pavlyuk, B. Ph.; Lobanov, M. V.; Antipov, E. V.;
Lebedev, O. I.; Van Tendeloo, G.; Ignatchik, O. L.; Ovtchenkov, E. A.;
Koksharov, Yu.A.; Vasilev, A. N. J. Solid State Chem. 2001, 160, 353361.
419. Wright, A. J.; Palmer, H. M.; Anderson, P. A.; Greaves, C. J. Mater. Chem. 2001,
11, 13241326.
420. Wright, A. J.; Palmer, H. M.; Anderson, P. A.; Greaves, C. J. Mater. Chem. 2002,
12, 978982.
421. Battle, P. D.; Bell, A. M. T.; Blundell, S. J.; Coldea, A. I.; Gallon, D. J.; Pratt, F. L.;
Rosseinsky, M. J.; Steer, C. A. J. Solid State Chem. 2002, 167, 188195.
422. Abakumov, A. M.; Kalyuzhnaya, A. S.; Rozova, M. G.; Antipov, E. V.;
Hadermann, J.; Van Tendeloo, G. Solid State Sci. 2005, 7, 801811.
423. Hadermann, J.; Abakumov, A. M.; DHondt, H.; Kalyuzhnaya, A. S.; Rozova, M. G.;
Markina, M. M.; Mikheev, M. G.; Tristan, N.; Klingeler, R.; Buchnerd, B.;
Antipov, E. V. J. Mater. Chem. 2007, 17, 692698.
424. Luzikova, A. V.; Kharlanov, A. L.; Antipov, E. V.; Mueller-Buschbaum, HK. Z.
Anorg. Allg. Chem. 1994, 620, 326328.
425. Vaughey, J. T.; Wiley, J. B.; Poeppelmeier, K. R. Z. Anorg. Allg. Chem. 1991, 598,
327338.
426. Roth, G.; Adelmann, P.; Knitter, R.; Massing, S.; Wolf, Th. J. Solid State Chem.
1992, 99, 376387.
427. Cussen, E. J.; Rosseinsky, M. J.; Battle, P. D.; Burley, J. C.; Spring, L. E.;
Vente, J. F.; Blundell, S. J.; Coldea, A. I.; Singleton, J. J. Am. Chem. Soc. 2001,
123, 11111122.
428. Coldea, A. I.; Blundell, S. J.; Marshall, I. M.; Steer, C. A.; Singleton, J.; Pratt, F. L.;
Noailles, L. D.; Rosseinsky, M. J.; Spring, L. E.; Battle, P. D. Phys. Rev. B 2001,
65, B054402.
429. Battle, P. D.; Blundell, S. J.; Claridge, J. B.; Coldea, A. I.; Cussen, E. J.;
Noailles, L. D.; Rosseinsky, M. J.; Singleton, J.; Sloan, J. Chem. Mater. 2002,
14, 425434.
430. Caspi, E. N.; Avdeev, M.; Short, S.; Jorgensen, J. D.; Dabrowski, B.;
Chmaissem, O.; Mais, J.; Kolesnik, S. J. Solid State Chem. 2004, 177,
14561470.
431. Abakumov, A. M.; Shpanchenko, R. V.; Lebedev, O. I.; Van Tendeloo, G.;
Amelinckx, S.; Antipov, E. V. Acta Crystallogr. A 1999, 55, 828839.
432. Ben Yahia, H.; Gaudin, E.; Boulahya, K.; Darriet, J. Solid State Sci. 2011, 13,
12351241.
433. Chernov, S. V.; Dobrovolsky, Yu.A.; Ya, Istomin S.; Antipov, E. V.; Grins, J.;
Svensson, G.; Tarakina, N.; Abakumov, A. M.; Van Tendeloo, G.; Eriksson, S. G.;
Rahman, S. M. H. Inorg. Chem. 2012, 51, 10941103.
434. Ruiz-Gonzalez, M. L.; Cortes-Gil, R.; Alonso, J. M.; Hernando, A.; Vallet-Reg, M.; 475. Abakumov, A. M.; Hadermann, J.; Bals, S.; Nikolaev, I. V.; Antipov, E. V.; Van
Gonzalez-Calbet, J. M. Chem. Mater. 2006, 18, 57565763.
435. Hewat, A. W.; Capponi, J. J.; Chaillout, C.; Marezio, M.; Hewat, E. A. Solid State
Commun. 1987, 64, 301307.
436. Karen, P.; Woodward, P. M.; Linden, J.; Vogt, T.; Studer, A.; Fischer, P. Phys. Rev.
B 2001, 64, 214405.
437. Woodward, P. M.; Karen, P. Inorg. Chem. 2003, 42, 11211129.
438. Woodward, P. M.; Suard, E.; Karen, P. J. Am. Chem. Soc. 2003, 125,
88898899.
439. Karen, P.; Kjekshus, A.; Huang, Q.; Karen, V. L.; Lynn, J. W.; Rosov, N.;
Sora, I. N.; Santoro, A. J. Solid State Chem. 2003, 174, 8795.
440. El Massalami, M.; Elzubair, A.; Ibrahim, H. M.; Rizgalla, M. A. Physica C 1991,
183, 143148.
Transition-Metal Perovskites 39
441. Huang, Q.; Karen, P.; Karen, V. L.; Kjekshus, A.; Lynn, J. W.; Mighell, A. D.;
Rosov, N.; Santoro, A. Phys. Rev. B 1992, 45, 96119619.
442. Karen, P.; Kjekshus, A.; Huang, Q.; Lynn, J. W.; Rosov, N.; Sora, I. N.;
Karen, V. L.; Mighell, A. D.; Santoro, A. J. Solid State Chem. 1998, 136, 2133.
443. Caignaert, V.; Millange, F.; Domenges, B.; Raveau, B.; Suard, E. Chem. Mater.
1999, 11, 930938.
444. Millange, F.; Caignaert, V.; Domenges, B.; Raveau, B.; Suard, E. Chem. Mater.
1998, 10, 19741983.
445. Trukhanov, S. V.; Troyanchuk, I. O.; Hervieu, M.; Szymczak, H.; Barner, K. Phys.
Rev. B 2002, 66, 184424.
446. Tenailleau, C.; Allix, M.; Claridge, J. B.; Hervieu, M.; Thomas, M. F.; Hirst, J. P.;
Rosseinsky, M. J. J. Am. Chem. Soc. 2008, 130, 75707583.
447. Gormezano, A.; Weller, M. T. J. Mater. Chem. 1993, 3, 771772.
448. Anderson, M. T.; Poeppelmeier, K. R.; Zhang, J. P.; Fan, H. J.; Marks, L. D. Chem.
Mater. 1992, 4, 13051313.
449. Otzschi, K. D.; Vander Griend, D. A.; Poeppelmeier, K. R.; Sinkler, W.;
Marks, L. D.; Mason, T. O. Chem. Mater. 1998, 10, 25792581.
450. Istomin, S. Ya; Grins, J.; Svensson, G.; Drozhzhin, O. A.; Kozhevnikov, V. L.;
Antipov, E. V.; Attfield, J. P. Chem. Mater. 2003, 15, 40124020.
451. Istomin, S. Ya; Drozhzhin, O. A.; Svensson, G.; Antipov, E. V. Solid State Sci.
2004, 6, 539546.
452. Withers, R. L.; James, M.; Goossens, D. J. J. Solid State Chem. 2003, 174,
198208.
453. James, M.; Cassidy, D.; Goossens, D. J.; Withers, R. L. J. Solid State Chem.
2004, 177, 18861895.
454. Gillie, L. J.; Palmer, H. M.; Wright, A. J.; Hadermann, J.; Van Tendeloo, G.;
Greaves, C. J. Phys. Chem. Solids 2004, 65, 8793.
455. van Smaalen, S. Incommensurate Crystallography. Oxford University Press:
New York, 2007.
456. Abakumov, A. M.; Alekseeva, A. M.; Rozova, M. G.; Antipov, E. V.; Lebedev, O. I.;
Van Tendeloo, G. J. Solid State Chem. 2003, 174, 319328.
457. Ruiz-Gonzalez, M. L.; Prieto, C.; Alonso, J.; Ramrez-Castellanos, J.; Gonzalez-
Calbet, J. M. Chem. Mater. 2002, 14, 20552062.
458. Berastegui, P.; Hull, S.; Garca-Garca, F. J.; Eriksson, S.-G. J. Solid State Chem.
2002, 164, 119130.
459. Lambert, S.; Leligny, H.; Grebille, D.; Pelloquin, D.; Raveau, B. Chem. Mater.
2002, 14, 18181826.
460. Lazic, B.; Kruger, H.; Kahlenberg, V.; Konzett, J.; Kaindl, R. Acta Crystallogr. B
2008, 64, 417425.
461. Kruger, H.; Kahlenberg, V. Acta Crystallogr. B 2005, 61, 656662.
462. DHondt, H.; Abakumov, A. M.; Hadermann, J.; Kalyuzhnaya, A. S.; Rozova, M. G.;
Antipov, E. V.; Van Tendeloo, G. Chem. Mater. 2008, 20, 71887194.
463. Baldinozzi, G.; Calvarin, G.; Sciau, Ph.; Grebille, D.; Suard, E. Acta Crystallogr. B
2000, 56, 570576.
464. Tomeno, I.; Tsunoda, Y.; Oka, K.; Matsuura, M.; Nishi, M. Phys. Rev. B 2009, 80,
104101.
465. Guiton, B. S.; Davies, P. K. Nat. Mater. 2007, 6, 586591.
466. Guiton, ; Wu, H.; Davies, P. K. Chem. Mater. 2008, 20, 28602862.
467. Garca-Martn, S.; Urones-Garrote, E.; Knapp, M. C.; King, G.; Woodward, P. M.
J. Am. Chem. Soc. 2008, 130, 1502815037.
468. Garca-Martn, S.; King, G.; Urones-Garrote, E.; Nenert, G.; Woodward, P. M.
Chem. Mater. 2011, 23, 163170.
469. Licurse, M. W.; Davies, P. K. Appl. Phys. Lett. 2010, 97, 123101.
470. King, G.; Garca-Martn, S.; Woodward, P. M. Acta Crystallogr. B 2009, 65,
676683.
471. Magneli, A. Acta Chem. Scand. 1948, 2, 501517.
472. Kihlborg, L. Ark. Kemi. 1963, 21, 443460.
473. Wadsley, A. D. Rev. Pure Appl. Chem. 1955, 5, 165193.
474. Bougerol, C.; Gorius, M. F.; Grey, I. E. J. Solid State Chem. 2002, 169,
131138.
Tendeloo, G. Angew. Chem. Int. Ed. 2007, 45, 66976700.
476. Abakumov, A. M.; Hadermann, J.; Van Tendeloo, G.; Antipov, E. V. J. Am. Ceram.
Soc. 2008, 91, 18071813.
477. Abakumov, A. M.; Batuk, D.; Hadermann, J.; Rozova, M. G.; Sheptyakov, D. V.;
Tsirlin, A. A.; Niermann, D.; Waschkowski, F.; Hemberger, J.; Van Tendeloo, G.;
Antipov, E. V. Chem. Mater. 2011, 23, 255265.
478. Lepoittevin, C.; Hadermann, J.; Malo, S.; Perez, O.; Van Tendeloo, G.; Hervieu, M.
Inorg. Chem. 2009, 48, 82578262.
479. Malo, S.; Lepoittevin, C.; Perez, O.; Hebert, S.; Van Tendeloo, G.; Hervieu, M.
Chem. Mater. 2010, 22, 17881797.
480. Raynova-Schwarten, V.; Massa, W.; Babel, D. Z. Anorg. Allg. Chem. 1997, 623,
10481054.
40 Transition-Metal Perovskites
481. Tzvetkov, P.; Petrova, N.; Kovacheva, D. J. Alloys Compd. 2009, 485, 862866.
482. Nikolaev, I. V.; DHondt, H.; Abakumov, A. M.; Hadermann, J.; Balagurov, A. M.;
Bobrikov, I. A.; Sheptyakov, D. V.; Pomjakushin, V. Yu.; Pokholok, K. V.;
Filimonov, D. S.; Van Tendeloo, G.; Antipov, E. V. Phys. Rev. B 2008, 78,
024426.
483. Hadermann, J.; Abakumov, A. M.; Perkisas, T.; DHondt, H.; Tan, H.; Verbeeck, J.;
Filonenko, V.; Antipov, E. V.; Van Tendeloo, G. J. Solid State Chem. 2010, 183,
21902195; Abakumov, A. M.; Hadermann, J.; Batuk, M.; DHondt, H.;
Tyablikov, O. A.; Rozova, M. G.; Pokholok, K. V.; Filimonov, D. S.;
Sheptyakov, D. V.; Tsirlin, A. A.; Niermann, D.; Hemberger, J.; Van Tendeloo, G.;
Antipov, E. V. Inorg. Chem. 2010, 49, 95089516.
484. Korneychik, O. E.; Batuk, M.; Abakumov, A. M.; Hadermann, J.; Rozova, M. G.;
Sheptyakov, D. V.; Pokholok, K. V.; Filimonov, D. S.; Antipov, E. V. J. Solid State
Chem., 2011, 184, 31503157.
485. Ebbinghaus, S. G.; Abicht, H.-P.; Dronskowski, R.; Muller, T.; Reller, A.;
Weidenkaff, A. Prog. Solid State Chem. 2009, 37, 173205.
486. Pors, F.; Marchand, R.; Laurent, Y.; Batcher, P.; Roult, G. Mater. Res. Bull. 1988,
23, 14471450.
487. Marchand, R.; Pors, F.; Laurent, Y. Ann. Chim. Fr. 1991, 16, 553560.
488. Fawcett, I. D.; Ramanujachary, K. V.; Greenblatt, M. Mater. Res. Bull. 1997, 32,
15651570.
489. Clarke, S. J.; Guinot, B. P.; Michie, C. W.; Calmont, M. J. C.; Rosseinsky, M. J.
Chem. Mater. 2002, 14, 288294.
490. Kim, Y.-I.; Woodward, P. M. J. Solid State Chem. 2007, 180, 32243233.
491. Goodenough, J.-B. Rep. Prog. Phys. 2004, 67, 19151993.
492. Fang, C. M.; de Wijs, G. A.; Orhan, E.; de With, G.; de Groot, R. A.; Hintzen, H. T.;
Marchand, R. J. Phys. Chem. Solid 2003, 64, 281286.
493. Jansen, M.; Letschert, H. P. Nature 2000, 404, 980982.
494. Ravel, B.; Kim, Y.-I.; Woodward, P. M.; Fang, C. M. Phys. Rev. B. 2006, 73,
184121.
495. Wolff, H.; Dronskowski, R. J. Comput. Chem. 2008, 29, 22602267.
496. Page, K.; Stoltzfus, M. W.; Kim, Y.-I.; Proffen, T.; Woodward, P. M.;
Cheetham, A. K.; Seshadri, R. Chem. Mater. 2007, 19, 40374042.
497. Yang, M.; Oro-Sole, J.; Rodgers, J. A.; Belen Jorge, A.; Fuertes, A.; Attfield, J. P.
Nat. Chem. 2011, 3, 4752.
498. Kim, Y.-I.; Woodward, P. M.; Baba-Kishi, K. Z.; Tai, C. W. Chem. Mater. 2004,
16, 12671276.
499. Clemens, O.; Haberkorn, R.; Slater, P. R.; Beck, H. P. Solid State Sci. 2010, 12,
14551463.
500. Al-Mamouri, M.; Edwards, P. P.; Greaves, C.; Slaski, M. Nature 1994, 369,
382384.
501. Perrin, C.; Pena, O.; Sergent, M.; Fonteneau, G.; Lucas, J. Europhys. Lett. 1990,
11, 181187.
502. Ardashnikova, E. I.; Lubarsky, S. V.; Denisenko, D. I.; Shpanchenko, R. V.;
Antipov, E. V.; Van Tendeloo, G. Physica C 1995, 253, 259265.
503. Lightfoot, P.; Pei, S.; Jorgensen, J. D.; Tang, X.-X.; Manthiram, A.;
Goodenough, J. B. Physica C 1990, 169, 1522.
504. Slater, P. R. J. Fluorine Chem. 2002, 117, 4345.
505. Rudorff, W.; Krug, D. Z. Anorg. Allg. Chem. 1964, 329, 211217.
506. Chamberland, B. L. Mater. Res. Bull. 1971, 6, 311316.
507. Katsumata, T.; Nakashima, M.; Umemoto, H.; Inaguma, Y. J. Solid State Chem.
2008, 181, 27372740.
508. Troyanchuk, I. O.; Kasper, N. V.; Mantytskaya, O. S.; Shapovalova, E. F. Mater.
Res. Bull. 1995, 30, 421425.
509. Hadermann, J.; Van Tendeloo, G.; Abakumov, A. M.; Pavlyuk, B. Ph.;
Rozova, M. G.; Antipov, E. V. Int. J. Inorg. Mater. 2000, 2, 493502.
510. Alekseeva, A. M.; Abakumov, A. M.; Rozova, M. G.; Antipov, E. V.; Hadermann, J.
J. Solid State Chem. 2004, 177, 731738.
511. Antipov, E. V.; Abakumov, A. M.; Alekseeva, A. M.; Rozova, M. G.; Hadermann, J.;
Lebedev, O. I.; Van Tendeloo, G. Phys. Stat. Sol. (a) 2004, 201, 14031409.
512. El Shinawi, H.; Marco, J. F.; Berry, F. J.; Greaves, C. J. Mater. Chem. 2010, 20,
32533259.
513. Abakumov, A. M.; Hadermann, J.; Rozova, M. G.; Pavlyuk, B. Ph.; Antipov, E. V.;
Lebedev, O. I.; Van Tendeloo, G. J. Solid State Chem. 2000, 149, 189196.
514. Lobanov, M. V.; Abakumov, A. M.; Sidorova, A. V.; Rozova, M. G.;
Dyachenko, O. G.; Antipov, E. V.; Hadermann, J.; Van Tendeloo, G. Solid State
Sci. 2002, 4, 1922.
515. Shpanchenko, R. V.; Rozova, M. G.; Abakumov, A. M.; Ardashnikova, E. I.;
Kovba, M. L.; Putilin, S. N.; Antipov, E. V.; Lebedev, O. I.; Van Tendeloo, G.
Physica C 1997, 280, 272280.
516. vanTendeloo, G.; Lebedev, O. I.; Shpanchenko, R. V.; Antipov, E. V. J. Electron
Microsc. 1997, 1, 2331.
517. Berry, F. J.; Coomer, F. C.; Hancock, C.; Helgason, O.; Moore, E. A.; Slater, P. R.;
Wright, A. J.; Thomas, M. F. J. Solid State Chem. 2011, 184, 13611366.
 4.15 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
D Neagu and JTS Irvine, University of St Andrews, St Andrews, UK
2013 Elsevier Ltd. All rights reserved.

4.15.1 Nonstoichiometry and Doping 397

4.15.2 The Impact of Defect Chemistry on Structure 398
4.15.2.1 The Ideal Perovskite Structure 398
4.15.2.2 Accommodation of Cation Substitutions 398
4.15.2.3 Accommodation of Deficiency 399
4.15.2.4 Accommodation of Oxygen Excess 401
4.15.3 The Effect of Defect Chemistry on Properties 401
4.15.3.1 Sintering and Microstructure 401
4.15.3.2 Unit Cell Size 401
4.15.3.3 The Extent of Reduction 402
4.15.3.3.1 Meaning and importance of the extent of reduction 402
4.15.3.3.2 Reduction in titanates 403
4.15.3.4 Electronic Conductivity 407
4.15.3.4.1 Electronic conduction in the perovskite lattice 407
4.15.3.4.2 Factors that decide the conductivity of titanates 407
4.15.3.5 B-Site Cation Segregation from A-Site-Deficient Perovskites 410
4.15.3.5.1 Reasons behind the phenomenon 410
4.15.3.5.2 Different manifestations of the B-site segregation 411
4.15.3.5.3 Factors controlling B-site segregation 412
4.15.4 Conclusion 414
References 414

4.15.1 Nonstoichiometry and Doping
Perovskite oxide materials possess the general stoichiometry
ABO3 in which conventionally A denotes the larger cation and
B the smaller cation. Several factors such as the variety and
number of defects that can be accommodated by the perov-
skite structure, their formation and migration energies, and the
equilibrium between them and the response of this equilib-
rium to external conditions (temperature, oxygen partial
pressure, etc.) give rise to a diverse and valuable defect
chemistry.1,2 The nature and concentrations of defects and
thus defect chemistry can be significantly manipulated by
doping (Figure 1(a)). This can be achieved by substitution
of parent cations with cations of similar size but different
valence (Figure 1(b)). Bearing in mind the electro-neutrality
condition, different mechanisms for charge compensation
can be envisaged, as shown schematically in Figure 1(a) for
a general perovskite ABO3, together with corresponding
examples derived by doping the archetype perovskite,
Sr2Ti4O32.
Doping the A-site with a cation of higher charge than Sr2,
such as La3, can be compensated for by introducing extra oxygen
beyond the ABO3 stoichiometry (hence the terminology oxygen
excess) as in LaxSr1xTiO3x/2; by creating A-site vacancies, as in
LaxSr13x/2TiO3 (named A-site deficient); or by reducing the
oxidation state of some Ti4 ions, LaxSr1xTix3Ti1x4O3 when
Ti3 can be stabilized, for example, by preparing the sample
in highly reducing conditions. Oxygen ion vacancies or
interstitials can be generated by substitution of Ti ions with
cations Mm of lower or higher charge respectively, producing
compounds of stoichiometry SrMxTi1xO3(4m)(x/2). Com-
pounds with simultaneous A-site deficiency and oxygen defi-
ciency, LaxSr13x/2MyTi1yO3(4m)(y/2) (A, O-site deficient),
can be produced, depending on the nature of M and if the overall
deficiency is small enough. When doping both A- and B-sites it is
also possible to balance charge between sites, thus creating A-site
deficient oxygen stoichiometric compounds such as the ones
belonging to the series La0.4(4m)xSr0.4(4m)xMxTi1xO3.
When exposed to reducing conditions (for example, in H2-
containing atmospheres) at sufficiently high temperatures (usu-
ally above 900  C), oxygen ions can be stripped from the lattice,
essentially doping electrons and oxygen vacancies:
0 1
2TixTi OxO $ 2Ti Ti V
O O2 [1]
2
It is worth noting that depending on the stoichiometry
resulting after doping with respect to ABO3, the formulations
fall into two broad categories: sub-stoichiometric or super-
stoichiometric, as indicated in Figure 1(a). The distinct fea-
tures exhibited by these classes will be outlined on numerous
occasions during the following subsections.
This chapter focuses on the effect that some of the above-
mentioned extrinsic defects have on several aspects including
the structure, formation, and conduction of charge carriers,
and finally discusses unique phenomena that may arise from
the interplay between defects and structure.

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00421-6 397

398 Perovskite Defect Chemistry as Exemplified by Strontium Titanate

4.15.2 The Impact of Defect Chemistry on Structure
4.15.2.1 The Ideal Perovskite Structure
SrTiO3 exhibits the ideal perovskite structure in which sixfold
coordinated Ti4 ions occupy the corners of a cubic cell. The
12-fold coordinated Sr2 ions are located in the center of the
cube while oxide ions sit on the edges, halfway between Ti4
ions (Figure 2(a)). Alternatively, the structure can be visual-
ized as being composed of a network of corner-shared TiO6
octahedra, in which the large cation, Sr2, occupies the re-
sulting cuboctahedral cavities (Figure 2(b)). Depending on
the nature of the point defects introduced, the perovskite struc-
ture can accommodate them through various mechanisms
discussed in the following subsections.
4.15.2.2 Accommodation of Cation Substitutions
The mechanism through which cation substitution-type de-
fects are accommodated by the perovskite lattice depends
mainly on the size of the dopant relative to the host cation
(see Figure 1(b)). This is often discussed in relation to the so-
called perovskite tolerance factor, t:
rA rO
t p : [2]
2rB rO
The tolerance factor is derived from the geometry of the
perovskite cell (see Figure 2(a)) and is a measure of the mis-
match between the (AO) and (BO) bond lengths that are
calculated as the sums of empirical ionic radii (rA, rB, and rO)

DionJacobson
1.4 O2
Aurivillius

An+1BnO3n+1 Sr3Ti2O7 1.2 Sr2+

Ce4+

AnBnO3n+2 LaxSr1-xTiO3+x/2 La3+

Ionic radius ()
1.0
A-site
Super-stoichiometric
ABO3 SrTiO3 2+
Sub-stoichiometric
0.8 Ni Fe3+,Mn3+
B-site

Zn2+
ABO3g SrTiO3-d 0.6
Cu2+
A1-aBO3 LaxSr1-3x /2TiO3 Ga3+
0.4 Ti4+

A0.6BO3 La0.4+(4-m)xSr0.4-(4-m)xMxTi1-xO3 2 4 6 8 10 12
Coordination number
A1-aBO3-g LaxSr1-3x/2MyTi1-yO3-(4-m)y/2
(a) (b)

Figure 1 (a) Possible nonstoichiometry in perovskites and corresponding examples derived from doped SrTiO31,3 and (b) plot showing that the ionic
radii as a function of coordination number for some cations would fit the A- and B-sites of SrTiO3 (ionic radii from the work of Shannon4).

r A + rO
O
Sr

rB + rO
Ti

(a) (b)

Figure 2 The perovskite structure viewed from different perspectives: (a) emphasizes the unit cell and the coordination of cations and (b) outlines the
network of corner-sharing TiO6 octahedra.
 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
obtained from X-ray diffraction at room temperature and at-
mospheric pressure (some ionic radii plotted in Figure 1(b)4).
When multiple cations occupy the same site, the average of
their ionic radii is used. For SrTiO3 the size of Sr2 and Ti4 is
of just the right ratio to create a cubic lattice and, in conse-
quence, t 1. For t > 1, the A-site cation is too large for the
cuboctahedral cavity; therefore, the TiO6 octahedra need to
stretch slightly in order to accommodate the A-site ion.5,6
Because of this, Ti is not pinned any more in the symmetry
center of the octahedra and can move, potentially giving rise to
ferroelectricity, as in the case of BaTiO3 (rBa2 > rSr2 ) depicted
in Figure 3(a),5,7 When size mismatch is even more significant,
hexagonal polytypes are produced.8
Situations when t < 1 are encountered when substituting
some Sr2 by smaller cations such as La3 or Ce4/3, and/or
replacing some Ti4 by larger cations such as Fe3, Zn2, etc. In
this case, the A-site cation is undersized for the cuboctahedral
cavity and thus it cannot make contact with all 12 oxide ions.
The structure adjusts by undergoing cooperative rotations of
the octahedra while maintaining both the regularity and the
corner sharing, a process that brings some of the oxide ions in
touch with the small A-site cation. This is known as octahedral
tilting and it has been the subject of many studies over the
years.5,915 Octahedra in adjacent layers can tilt in the same
direction (in-phase tilting see Figure 3(b)) or in the oppo-
site direction (out-of-phase tilting see Figure 3(c)), and with
respect to all three crystallographic directions (although some
combinations are restricted, see the work of Howard and
Stokes14). There are two main consequences arising from oc-
tahedral tilting, aside from the overall decrease in symmetry
that accompanies the process. First and most important, the
angle between TiOTi bonds will drop from 180 as it is in
the ideal structure to (180  f). f usually increases with t, but
is typically not larger than 15 in titanates. This change in the
(TiO) interactions (orbital overlap, inter-atomic distances,
etc.) is expected to have a direct impact on many properties
including the width of the conduction band, the band gap, or
the strength of the (TiO) bond itself.8,16,17 Second, tilting
leads to a reduction in the coordination number of the A-site
cation from 12 to 810, creating an asymmetric environment,
which may alter the diffusion pathways of oxide ions and cause
t>1
Ba
Ti
(a) (b)
Figure 3 Accommodation of cation substitution in perovskites: when t > 1, the BO6 octahedra are slightly stretched, allowing for displacement of the
B-site ion from its symmetry position (a). When t < 1, octahedra undergo a cooperative tilting that, with respect to the adjacent layer, can occur in
the same direction which is known as in-phase tilting (b), or in the opposite direction and is called out-of-phase tilting (c).
399
further splitting of the degeneracy.18 It is worth noting that, for
most perovskites, tilting gradually decreases with increasing
temperature,13,1924 such that at sufficiently high temperatures
(5001500  C) some even become cubic. This is possibly due
to the fact that the (AO) bonds expand more than the (BO)
bonds;8,25 thus, if the perovskite starts with t < 1, upon heating
the numerator in eqn [2] will increase faster that the denomi-
nator, meaning that the tolerance factor can approach unity.
Although this description of tilting as an effect originat-
ing essentially from steric causes explains many observed
phenomena,26 there have been indications that increased co-
valent interactions between cations and anions would actually
be the driving force behind tilting.5,27 Either way, the tilting
phenomenon is what makes the perovskite structure so easy to
adapt and thus to be able to incorporate a wide variety of
substitutions. Generally, the perovskite can accommodate
cations by tilting such that 0.78 < t < 1. Below this value, the
ilmenite structure becomes more stable.5
The substitution level can also play an important role.
When in small enough number, the dopants are randomly
distributed on the host sites. However, if the doping level can
lead to the formation of ordered structures then these struc-
tures will be preferred over the random ones. As a result, a large
number of perovskites exhibiting A- or B-site ordering are
known.12,15,28
4.15.2.3 Accommodation of Deficiency
Sub-stoichiometry (or deficiency) is accommodated in a simi-
lar manner to cation substitutions. This is probably because
vacancies might simply behave as dopants of different size
compared to the host cation. When A-site and/or oxygen va-
cancies are present, the perovskite preserves its corner-shared
octahedra network and adjusts to the different size of the
vacancies by tilting.
The accommodation of A-site vacancies is illustrated in
Figure 4 for the series LaxSr13x/2TiO3. Until x  0.4, which
corresponds to an A-site deficiency of a 0.2, the vacancies
are randomly distributed in the perovskite, occasionally asso-
ciating into pairs (see Figure 4(b)).20,29,30 Although the sub-
stitution fraction is considerable (x 0.4), the perovskite is
t<1
(c)
400 Perovskite Defect Chemistry as Exemplified by Strontium Titanate

SrTiO3
Pm-3m
Sr Sr Sr

Ti Ti

Sr Sr Sr

Ti Ti

Sr Sr Sr
(a)

LaxSr1-3x/2TiO3 LanSrn-4TinO3n+2
(x = 0.4) I4/mcm (n = 12)

Sr Sr Sr
Shearing

Ti Ti

Sr Sr Sr

5 nm
Ti Ti

Sr Sr Sr
(b) (e)

LaxSr1-3x/2NiyTi1-yO3-y LaxSr1-3x/2TiO3 La4Srn-4TinO3n+2

(x = 0.4, y = 0.1) I2/a (x = 0.6) Cmmm (n = 5) P1c1

.7
15
2 nm

(d)

16
(c)

(f)

La4Srn-4TinO3n+2
Viewing direction (n = 4) P21

(a), (c), (d),

(b)
(e), (f), (g)
x z

y y

(g)

Figure 4 Accommodation of vacancies and oxygen-excess defects in perovskites: (a) the defect free SrTiO3, (b) randomly distributed A-site
vacancies,20,30 (c) cation-vacancy ordering on the A-site,30,31 (d) randomly distributed A, O-site vacancies,32 (e) randomly distributed oxygen-excess
intergrowths for low La-doping levels,3,33 (f) ordered oxygen-excess intergrowths form distinct crystallographic shears,3,34 and (g) the end-member of
the solid solution La4Srn4TinO3n2, La2Ti2O7.34,35
 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
only slightly distorted from ideal (Figure 4(b)), probably due
to the similar size of the dopant and the host (Figure 1(a)). As
soon as more A-site vacancies are inserted into the structure by
increasing La doping, x, they start to associate on a larger scale,
eventually reaching a clear repetitive arrangement for x 0.6
(Figure 4(c)). The A-site species order in such a way that fully
occupied planes alternate with planes that are only 1/3 oc-
cupied. The octahedra are strongly tilted only around the x axis,
which is perpendicular to the direction of cation-vacancy or-
dering. As it will be shown in the next subsections, ordering
plays a key role in the variation of properties.
Substituting Ni on the B-site of the x 0.4 member of
LaxSr13x/2TiO3, which is accompanied by generation of vacancies,
LaxSr13x/2NiyTi1yO3y, does not lead to major changes in the
perovskite lattice, nor is any ordering observed. However, tilting
increases noticeably with respect to all three axes (Figure 4(d)).
4.15.2.4 Accommodation of Oxygen Excess
The means through which additional oxygen beyond the nor-
mal ABO3 stoichiometry can be accommodated by the perov-
skite structure has been subject to many debates. While it is
difficult to imagine that oxygens could be accommodated in
interstitials of a cubic close-packed arrangement, compensa-
tion through cation vacancies has also been ruled out recently
for the strontium titanate system.3
The accommodation of excess oxygen is based on a distinct
mechanism, which allows the perovskite structure to alternate
with intergrowths belonging to other crystal structures. Some
of the most important structures containing such intergrowths
are listed in Figure 1(a).
RuddlesdenPopper phases,36,37 An1BnO3n1, consist of
rock-salt intergrowths between n-layer thick perovskite blocks
and are of interest particularly in the area of high-temperature
superconductors.38 The Aurivillius phases39 are made of
[Bi2O2]2 layers alternating with perovskite blocks.40,41 In.
DionJacobson phases perovskite blocks intergrow with
halides.42,43
The AnBnO3n2 (n  4) series has been subject to many
studies (3,34,44) which indicate that the structure of these com-
pounds consists of {110} perovskite slabs joined by crystallo-
graphic shearing along the cubic [001] direction. This structure
can be better visualized by examining the member n 4,
La2Ti2O7, shown in Figure 4(g). La2Ti2O7 has a layered struc-
ture consisting of four-layer-thick perovskite blocks that
are offset from one another at the crystallographic shear
(Figure 12(b)). Thus, the connectivity between the corners of
the octahedra is broken along the shears where the excess
oxygen is actually located. Notice also that the octahedra
are tilted within the perovskite blocks and additionally become
increasingly distorted as they approach the crystallographic
shear. Moreover, the A-site cations in the perovskite layers
adjacent to the crystallographic shear are significantly dis-
placed from their ideal lattice position toward the shear.
By considering a solid solution between the cubic SrTiO3
and the layered La2Ti2O7, (La4Srn4TinO3n2, n  4), it has
been shown possible to decrease the frequency of crystallo-
graphic shears by increasing the value of n.3 Increasing the
value of n from 4 to 5 leads to an increase in the number of
perovskite layers composing the perovskite blocks (compare
401
Figure 4(g) to Figure 4(f)), but the structure still retains
its layered appearance. As n is further increased, the shearing
is broken, the extended defects becoming more and more
sporadic (Figure 10(b)) until n  12, where even though
oxygen excess is still present, these intergrowths are randomly
distributed throughout the structure in such a way that the
overall symmetry resembles the cubic SrTiO3 end-member
(see Figure 4(e)).3 The possibility of controlling the extent of
ordering in this system allows one to control the functionality
of these materials, an aspect that will be further addressed in
the following subsections. It is also worth noting that the series
LaxSr1xTiO3x/2 has been shown to be equivalent to
La4Srn4TinO3n2.34
4.15.3 The Effect of Defect Chemistry on Properties
4.15.3.1 Sintering and Microstructure
Different titanate defect classes yield very distinct microstruc-
tures when processed under similar conditions. When prepared
by solid-state routes, the A-site-deficient titanates form at
lower temperatures between 1350 and 1450  C, while oxygen-
excess compositions generally require temperatures around
14501600  C. Even so, the A-site-deficient titanates grow larger
grains compared to the oxygen-excess ones. This is more obvi-
ous when trying to prepare porous structures from these differ-
ent classes. Figure 5 shows that the A-site-deficient perovskite
sinters much better, yielding large grains. Both phase formation
and grain growth are likely to be enhanced by the presence of
the A-site vacancies, which represent available hopping sites for
cations, thus promoting cation diffusion.4547
Oxygen nonstoichiometry seems to correlate with the as-
pect of the grains: oxygen-deficient samples show terrace-like
features (Figure 18(d)) that become more sporadic upon
approaching oxygen stoichiometric compositions (Figure 18
(b)) and tend to fade away for the oxygen-excess composition
(Figure 5(b)).32 Oxygen-excess titanates are also softer com-
pared to the A-site-deficient ones, probably due to the presence
of the intergrowths, which have been reported to be prone to
cleavage.34
4.15.3.2 Unit Cell Size
It is important to realize that the size of the unit cell and/or the
symmetry or the actual cell will change upon introduction of
defects such as cation substitutions, vacancies, etc. This will not
be addressed in detail here; instead, a few representative exam-
ples will be presented.
Depending on how the defects are actually incorporated in
the perovskite, the lattice will react accordingly. In Figure 6(a),
the pseudo-cubic cell parameters for the primitive perovskite
cell (or reduced cell parameters) were plotted against dopant
content (La) in two representative series: the oxygen-excess
LaxSr1xTiO3x/2 and the A-site-deficient LaxSr13x/2TiO3. The
plot shows linear fits to the experimental data taken from the
work of Hashimoto et al.48 and Howard et al.,20 respectively.
While one would expect the cell parameter to decrease in size
slightly as the smaller La3 substitutes for the larger Sr2, this is
not the case for the oxygen-excess series. This is most likely due
to the oxygen-excess intergrowths where the packing is not as
402 Perovskite Defect Chemistry as Exemplified by Strontium Titanate

x:0.000 m
St Andrews SEI 5.0 kV 3000 1 mm
WD 5.3mm St Andrews SEI 5.0 kV 3000 1 mm
WD 6.1mm
(a) (b)

Figure 5 Microstructures of the fracture of porous samples belonging to two different titanate classes: (a) the A-site-deficient La0.52Sr0.28Ni0.06Ti0.94O3
sintered with 20% glassy carbon as pore-former at 1390  C for 6 h; (b) the oxygen-excess La0.33Sr0.67Ni0.06Ti0.94O3.09 sintered with 20% glassy carbon
as pore-former at 1400  C for 6 h.32

Ti3+, x (at/uc)
3.92 0.00 0.04 0.08 0.12 0.16 0.20
Lax Sr1-x TiO3+x/2
Pseudocubic cell parameter, a ()

3.902 4+
a+ = 3.905 + 0.015x La0.4Sr0.4Ti1-xTi3+
x O3-d (5) 0.5
3.91 3.900
a(d ) or a(x) (4)
Linear fit 0.4
Cell parameter, a ()

Cell expansion (%)

3.898 Cell expansion (3)
3.90 Linear fit
0.3
3.896 (2)

3.894 0.2
3.89
Lax Sr1-3x/2 TiO3
3.892
a = 3.8891 + 0.139 d 0.1
a = 3.905 - 0.045x
3.88 a = 3.8891 + 0.0695 x
3.890 (1)
0.0
0.0 0.2 0.4 0.6 3.888
0.00 0.02 0.04 0.06 0.08 0.10
x, La Oxygen deficiency, d (at/uc)
(a) (b)

Figure 6 Pseudo-cubic cell parameter as a function of defect concentration: (a) substitution of Sr2 by La3 in the oxygen-excess LaxSr1xTiO3x/2
(linear fit on data from the work of Hashimoto48) and in the A-site deficient LaxSr13x/2TiO3 (linear fit on data from the work of Howard20); (b) the cell
parameter against oxygen deficiency for different samples: (1) sintered in air, (2) pre-reduced at 1200  C, (3) pre-reduced at 1300  C, (4) the coarse
powder reduced at 1300  C; and (5) the fine powder reduced at 1300  C.49

compact as in the perovskite layers. On the other hand, the
A-site-deficient titanate does show a negative slope, which is in
fact quite steep. Considering that La3 is only slightly smaller
than Sr2, it seems possible to admit that the A-ion vacancies
are smaller, at most equal, but not larger than the dopant. If the
size of the vacancy would be larger, then when their concentra-
tion becomes significant, one should observe a change in the
slope pointing to the fact that the larger sizes of the vacancies are
compensating for the size of the dopant by increasing the size of
the unit cell. Additional proof toward the assumption that
rV00  rLaSr comes from the fact that stronger tilting is observed
Sr
in these systems as the number of vacancies increases, which
correlates to the presence of undersized species on the A-site.
The cell parameter is also sensitive to changes in the oxida-
tion state of the B-site cation, which occur, for example, during
reduction. This is because as the B-ions are reduced, their oxi-
dation state decreases; hence, their size increases. Figure 6(b)
shows that a linear correlation can be found between the cell
parameter and the Ti3 content for a La0.4Sr0.4TiO3 sample after
reduction. The slope of this line seems to be linked to the change
in size of the B-ion, as pointed out elsewhere.49
4.15.3.3 The Extent of Reduction
4.15.3.3.1 Meaning and importance of the extent
of reduction
The reduction of perovskites is generally carried out by heating
the material in environments having low oxygen partial pres-
sures (pO2). The temperatures employed are usually between
800 and 1000  C, but can be as low as 500600  C or as high
 Perovskite Defect Chemistry as Exemplified by Strontium Titanate 403

0.0 1.0
0
(1) s at pO2 = 10-18 atm
-2
-0.5 s 0.8
pO2 -4 (3)
-6 (2)
log (s/S cm1)v

log (s/S cm1)

-1.0 0.6

log (pO2 /atm)

-8

-10
0.4
-1.5 -12

-14
0.2
-2.0 -16

-18
0.0
-2.5 -20
0.0 0.2 0.4 0.6 0.8 1.0 0 5 10 15 20
Time (h)
Time (h)

e e e
e
e e

Figure 7 The conductivity of a 60% dense La0.4Sr0.4TiO3 pellet measured on reduction in 5%H2/Ar at 900  C: (1) upon starting reduction (notice the
lowering of the pO2) the conductivity shows a jump of two orders of magnitude within minutes, which implies that a thin layer of reduced material covers
the grains and acts as conductive path for the flow of electrons; (2) as the reduction proceeds toward the core of the grain, it becomes slower; (3) finally,
all the bulk has been reduced and the conductivity does not increase further, reaching equilibrium.49

as 1400  C. Various pO2 values can be achieved by using inert
gases (Ar and N2), purely reducing gases (H2), mixtures of
these two, or other methods. During the reduction process,
some oxide ions are stripped from the perovskite lattice, leav-
ing behind vacancies, and a corresponding number of elec-
trons (B-site cations are reduced):
0 1
2BxB OxO $ 2BB V
O O2
2
The decrease in oxygen stoichiometry that occurs upon
reduction, from ABO3g to ABO3gd, denoted by d, represents
the oxygen ions lost or equivalently the number of oxide ion
vacancies created in the process, per formula unit of perovskite.
It is often referred to as the extent of the reduction, oxygen
deficiency on reduction, or even simply oxygen content. It is
worth noting that from the definition of d and eqn [3], it
0 
follows that [VO ] d and BB 2d. Because the reduction
process occurs with weight loss, d values can be easily obtained
by thermogravimetry.
The variable d plays an important role in understanding
the structureproperty relationship in perovskites as it
correlates with many of the properties, including electronic
conductivity,47,4952 catalytic activity,53 transition temperature
of oxide superconductors,54,55 or magnetic ordering.56,57
4.15.3.3.2 Reduction in titanates
Reduction in titanates has often been reported to proceed
slowly.49,5860 Further insight into how this process occurs on
a microscopic level and also about the timescale required to
reach an equilibrium value for the extent of reduction can
be obtained from conductivity measurements. The reduction
process can be monitored in this way because the extent of
the reduction is proportional to the concentration of charge
carriers generated through reduction, which in turn is propor-
tional to the conductivity (this relationship will be extensively
addressed in Section 4.15.3.4). An example of such an experi-
ment in which the conductivity of a porous sample of an A-site-
deficient perovskite is monitored on reduction is presented
[3] in Figure 7.
Upon switching from air to a reducing atmosphere, the
conductivity of the sample increases by two orders of magni-
tude. This suggests that the surface reaction between the gas
phase and the grains is very fast at this temperature (900  C),
and a thin film of reduced material will cover the surface of the
grains rapidly, as shown in Figure 7(1). This thin film of
reduced and thus conductive material provides a current path
for the electrons, accounting for the initial rapid increase of the
conductivity. Moreover, this also implies that the grain bound-
aries are subject to the same fast reduction on atmosphere
change, acting as conductive paths, and not as blocking fea-
tures. As the reduction advances toward the core of the grain,
the fraction of material that is reduced increases and conse-
quently the conductivity increases as well. However, the in-
crease in conductivity is considerably slower at this stage. If, in
the first stage, the conductivity increased by two orders of
magnitude within minutes, the conductivity will hardly reach
an additional order of magnitude by the end of the process,
after a total of 20 h. These observations point to the fact that
the removal and transport of oxide ions in the bulk of these
titanates represent the rate-determining step in the reduction
process, limiting the rate and the extent of the reduction. It is
therefore worth identifying the key factors that influence it.
404 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
Previously, it has been pointed out that valuable informa-
tion on how vacancies form and migrate in the perovskite
lattice can be acquired by looking at oxide ion conduction
studies in some perovskites, such as doped lanthanum gallates,
for which a large volume of both experimental and simulation
data is available.24,6168 The diffusion of oxide ions occurs by
hopping of oxide ions between available sites which are effec-
tively the oxide ion vacancies. The diffusion pathway of the
oxide ion through an undistorted perovskite is schematically
depicted in Figure 8, and the factors controlling it can be
synthetically expressed as follows65:
 
DO2  V O a2 eDHf DHm DHa =RT
Equation [4] shows that diffusion is a thermally activated
process proportional to the concentration of mobile vacancies,
[VO
], and the square of the cell parameter, a. Indeed, by
increasing the reduction temperature higher d values can be
obtained; Figure 9. Diffusion also depends on three main
processes and their associated thermal effects: formation
(DHf), migration (DHm), and association (DHa) of vacancies.
The way these factors come into play can be better understood
by examining Figure 8. When the oxide ion starts moving from
its position, it will need to break the bond shared with one of
Asite
V o
O2
Ti4+
(a)
Figure 8 Moving oxide ions (or oxygen vacancies) in undistorted perovskites. (a) Model emphasizing the curved diffusion pathway47 and
(b) space-filling model showing the real size of the ions (after Slater et al.24).
0.08
1300 C
Oxygen deficiency, d(at/uc)
0.06
1200 C
0.04
1100 C
0.02
0.00
60
Figure 9 The dependence of reduction on temperature and internal surface area of the samples. Reproduced from Neagu, D.; Irvine, J. T. S. Chem.
Mater. 2010, 22(17), 50425053, with permission.
the adjacent B cations. The AO bonds are significantly longer
than the BO bonds and thus weaker and are not expected to
play an important role at this stage. When the BO bond
breaks, the corresponding B cation will have its coordination
number decreased from 6 to 5. Most of the cations on the B-site
tend to form strong bonds with oxygen and strongly prefer
sixfold coordination. As a result, the energy required for the
formation of the vacancy, DHf, is high. Ti is a typical example
of such a cation. Although this stability of the BO bond is
detrimental for achieving high extents of reduction or promot-
ing oxide ion mobility, it also contributes to the remarkable
stability of the titanates. On the contrary, other elements such
[4] as Mn, Co, and Ga form cations that are known to be stable in
coordination numbers lower than 6 as well1 and have been
successfully employed in the design of materials with good
electronic and ionic conduction.69,70 Once the BO bond has
been broken, the oxide ion will have to migrate through the
triangle described by the two A-site cations and the B-site
cation, while maintaining constant distance with respect to
the B-site cation it is still bonded to. As a consequence, the
trajectory will be curved, as depicted in Figure 8. Previous
studies have shown that this is probably the most energy
intensive step in the migration of the oxide ion.24,62,65,67
Among the factors that come into play during this stage, the
O
A A
B
(b)
La0.4 Sr0.4 TiO3-d
70 80 90 100
rr (%)
 Perovskite Defect Chemistry as Exemplified by Strontium Titanate 405

most important are the nature (electronic structure, size, etc.)
of the A-site cations and the distortions of the lattice, which can
prevent the oxide ion from taking the simple path indicated in
Figure 8. Several studies suggest that the highest mobility for
vacancy hopping should be expected in lattices exhibiting high
symmetry where repulsing effects are minimal.62,66,67 Usually,
the migration energy is smaller compared to the vacancy for-
mation energy.65
At low concentrations, the interaction between defects is
negligible. However, when in sufficiently large numbers, de-
fectdefect interactions can become a dominant factor in the
energy requirements for ionic movement. These interactions
usually manifest themselves as associations between defects of
the same type and/or between defects of different nature. As-
sociations can be local or extended over significant distances,
in either cases potentially leading to an increase of the term
DHa and, consequently, the number of mobile vacancies be-
comes smaller than the stoichiometric one. Indeed, when ox-
ygen vacancy ordering occurs, ionic conductivity significantly
diminishes; therefore, a random distribution of these vacancies
throughout the crystal is to be desired. When the A-site ion is
partially substituted by another ion, oxygen vacancy trapping
can occur, mostly due to size mismatch between the host and
the substituting ion.71 Where the size of the host and dopant
are similar, for example when substituting Sr2 (1.44 A) for
La3 (1.36 A) in lanthanum gallates, DHa has been calculated
to be zero.71
Reduction is also negatively affected by the ordering of de-
fects, with examples being shown in Figure 10 for two distinct
classes of titanates: the A-site-deficient series LaxSr13x/2TiO3
and the oxygen-excess series La4Srn4TinO3n2 (or equivalently
LaxSr1xTiO3x/2). In the case of the A-site-deficient system, La
doping and the subsequent increase of A-site vacancies greatly
promote reduction up to a doping level of x  0.4, after which
the extent of the reduction falls dramatically. The observed posi-
tion for the maximum in reducibility coincides well with the
borderline between a disordered A-site and an extensive cation-
vacancy ordering on an A-site in this series (Figure 10(a)). As
explained in Section 4.15.2.3, the x 0 member of the series
LaxSr13x/2TiO3 starts as cubic, followed by a tetragonal distor-
tion due to slight out-of-phase tilting for 0 < x  0.4. In this
compositional domain the A-site vacancies are randomly distrib-
uted, but may occasionally associate in pairs (Figure 4(d)).
Above x  0.4, the A-site vacancies start ordering in at least one
dimension, forming a layered structure with vacant sites concen-
trated in alternate layers20,72,73 (Figure 4(c)). The higher value of
d for x < 0.4 can be explained in several ways. It is likely that the
00
large number of V Sr formed through doping limits the number
of intrinsic Schottky defects, which in turn facilitates the removal
of lattice oxygen by the reducing gas and the associated genera-
tion of free electrons (this mechanism will be discussed in the
subsection on conductivity). Another possible explanation
might be related to the way oxide ions travel through the bulk.
As explained above, when migrating, the oxide ion must pass
through the triangle described by the two A-site ions and the
B-site ion (see Figure 8). However, in this case, a large number of
A-sites are vacant and therefore the oxide ion might not be
required to pass through a triangle-shaped bottleneck.
In the case of the oxygen-excess series, a similar correlation
is observed between reduction and ordering. As explained
earlier, the reduction process can also be monitored by mea-
suring the electronic conductivity. Such an experiment is pre-
sented for the La4Srn4TinO3n2 series on mild reduction in
Figure 10(b). As discussed in Section 4.15.2.3, for small values
of n (n < 12) the structure resembles the end-member La2Ti2O7
exhibiting a layered structure of the intergrowths where the
extra oxygen is located (Figure 10(b)(3)). When n increases,
extended intergrowths become sporadic (Figure 10(b)(2)) and

A-site deficiency (1-x/2) (1) (2) (3)

0.00 0.05 0.10 0.15 0.20 0.25 0.30

0.035
LaxSr1-3x/2TiO3-d
0.030

4 nm 5 nm 4 nm
Oxygen deficiency, d

0.025
0
0.020 (2) (3)
-1.0
0.015
log[s (S cm1)]

0.010
(1) -2.0
Local Extended Layered
-3.0 defects defects defects
0.005
(n > 12) (n 12) (n < 12)
Disordered A-site Ordered A-site
0.000 -4.0

0.0 0.1 0.2 0.3 0.4 0.5 0.6

-5.0
x (La) 0 0.1 0.2 0.3 0.4 0.5
d

Figure 10 The impact of ordering on reduction: (a) the A-site-deficient LaxSr13x/2TiO3 series reduced at 1000  C in dry 5%H2/Ar. (1) corresponds to
the cubic structure of the x 0 member, followed by a compositional domain x  0.4, where a slight tetragonal distortion show in (2) is observed.
Around x > 0.4 cation-vacancy ordering on the A-site starts occurring;20,72,73 (b) the conductivity against excess oxygen for the LaxSr1xTiO3x/2 (or
equivalently La4Srn4TinO3n2) series, on mild reduction. For n < 12, the intergrowths where excess oxygen is located are ordered in layers shown in
inset (1). As n increases, the intergrowths become sporadic inset (2) and finally randomly distributed around n  12 inset (3), where a maxim in the
electronic conductivity, hence in reducibility is observed. Reproduced from Ruiz-Morales, J. C.; Canales-Vazquez, J.; Savaniu, C.; Marrero-Lopez, D.;
Zhou, W.; Irvine, J. T. S. Nature 2006, 439(7076), 568571, with permission.33
406 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
finally randomly distributed within the structure for n  12
(Figure 10(b)(1)). The transition point between the extended
ordering of the intergrowths and the structure where they are
randomly distributed is marked by an increase of several orders
of magnitude in conductivity, and hence in the extent of the
reduction.33 The breaking of ordering was of significant im-
portance for introducing functionality in this system, which
with additional minor improvements can be successfully used
as an anode in a solid oxide fuel cell running on methane.33
The extent of the reduction of these systems can be further
controlled by considering B-site doping. Therefore, the com-
positions that mark the borderline between ordered and
disordered structures have been brought forward for doping.
For the LaxSr1xTiO3x/2 A-site-deficient system, this is
La0.4Sr0.4TiO3, and for the oxygen excess, La4Srn4TinO3n2,
this is La4Sr8Ti12O38. The extent of the reduction achieved
following the substitution of some Ti4 (6% for the A-site
deficient system and 1/128% for the oxygen excess) with the
indicated dopants is plotted in Figure 11.32,47,74 The order in
which the dopants are listed corresponds to their atomic num-
ber. The variation of the extent of the reduction with different
dopants is strikingly different for the two defect chemistries
considered. The first aspect to notice is that the undoped
oxygen-excess sample reduces more at this temperature
(900  C) than the A-site-deficient sample. This was somewhat
expected as the oxygen-excess sample has an innate feature that
allows for good mobility of oxide ions which reside in the
oxygen-excess intergrowths.
Bearing in mind that oxide ions (or vacancies) can move only
on the edges of the TiO6 octahedra, as shown in Figure 8(b),
Figure 12 compares the path an oxide ion vacancy would take
to travel from the bulk to the surface either in the perovskite-
like slabs or near the intergrowth region. As pointed out in the
0.10
0.08
Extent of reduction, d
0.06
0.04
0.02
0.00
Mg2+ Al3+ Sc3+ Ti4+ Mn3+ Fe3+ Ni2+ Cu2+ Zn2+ Ga3+
Figure 11 The extent of the reduction achieved at 900  C after doping the oxygen-excess titanate La4Sr8Ti12O38 (trace 1) and the A-site-deficient
titanate La0.4Sr0.4TiO3 (trace 2) with the indicated B-site dopants (8% and 6%, respectively).32,47,74
subsection describing the structure, at the oxygen-excess inter-
growths the TiO6 octahedra are significantly stretched and the
A-site ions are moved from their symmetry position toward the
intergrowth. It is apparent by comparing pathway (1) to path-
way (2) in Figure 12(a) that the oxide ion vacancy traveling
near the intergrowth will migrate faster and easier mainly
because of two reasons: the TiO bonds are stretched and
therefore weaker, and the A-site cations do not hinder the
movement of the oxide ions because they are situated further
away from the diffusion pathway (compare to Figure 8 as
well). Indeed, recent impedance measurements have pointed
out that the oxygen-excess structures are similar to yttrium-
stabilized zirconium (a very good oxide ion conductor) in
terms of oxide ion mobility, whereas the A-site-deficient com-
positions are situated at the opposite pole, and resemble more
the inflexible SrTiO3 structure.75 A second aspect that arises
from the examination of Figure 11 is the effect of the dopant in
the two different defect chemistries. None of the dopants
employed seemed to be able to effectively enhance the extent
of the reduction of the oxygen-excess composition, many of
them having a rather negative impact. On the contrary, all the
cations, apart from Al3 increase the extent of the reduction of
the A-site-deficient titanate. This is probably related to the way
the dopant is compensated. In both cases, the lower charge of
the dopants (notice that all dopants actually have lower charge
than Ti4) will be compensated for by oxygen vacancies.
This means that in the A-site-deficient titanate a number of
available hopping sites will be created, which will help bulk
diffusion and thus the removal of oxide ions. On the other
hand, in the oxygen-excess stoichiometry, creating vacancies
means decreasing the number of the excess oxygen ions, which
is in fact the species responsible for the good reducibility of this
class. However, the effect of the dopants resides not only in the
(1) Oxygen-excess titanate
(2) A,O-site-deficient titanate
Undoped
oxygen excess
Undoped
A-site deficient
B-site dopant, Mm+
 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
Oxygen-excess
Pathway (2)
intergrowth
Pathway (1)
(a)
Figure 12 Proposed pathways for oxide ion vacancy diffusion in the perovskite slabs and in the layer adjacent to the oxygen-excess intergrowths.
(a) The two proposed pathways seen along the x axis; (b) the oxygen-excess intergrowth layer seen from the side, along the y axis the offset between
the perovskite slabs is evident.
changes they impose on stoichiometry and the availability of
charge carriers, but also in their ability to interfere with the
diffusion process itself. Small ions with high charge and thus
high-attraction fields such as Al3 will tend to pin oxide ions
into strong bonds, impairing diffusion in both classes of tita-
nates. Al3 in particular also favors sixfold oxygen coordina-
tion, effectively trapping excess oxygen ions. Cations such as
Mn3, Fe3, and Ni2 that are known to be able to adopt
flexible coordination numbers and thus allow for an easier
flow of oxide ions around them greatly benefit reduction
when doped in the A-site-deficient titanate, and also help
alleviate the effect of the decreased concentration of charge
carriers in the oxygen-excess composition, maintaining the
extent of the reduction more or less at the same level as in
the undoped sample.
It is worth noting that doping performed in the fashion
described here is an extremely powerful method for improving
the extent of reduction. For comparison, consider the A-site-
deficient composition La0.4Sr0.4TiO3. After a reduction at
1000  C this composition reaches a d  0.036 (Figure 10),
while after reducing it at 900  C, d  0.018 (Figure 11). By
doping as little as 6% Zn or Ga on the B-site, the material
will show the same d  0.036, at the lower temperature of
900  C, rather than 1000  C.
4.15.3.4 Electronic Conductivity
4.15.3.4.1 Electronic conduction in the perovskite lattice
Generally, electronic conduction in perovskites occurs through
the B-site network when these sites are occupied by cations
capable of adopting multiple oxidation states (Ti4/Ti3,
Nb5/Nb4, Mn4/Mn3, etc.). The BB distance is long in
perovskites (4 A), so direct overlap of the cation orbitals is
unlikely.18 Thus, conduction can be visualized either as elec-
tron hopping from B(n1) cations to Bn cations via the
oxygen bridges or through the conduction band formed by
the overlap of Ti t2g orbitals with oxygen p orbitals.8,18 The
407
z
z
y x
Perovskite
slabs
(b)
conductivity is proportional to the concentration of (free)
charge carriers, their charge Z, and their mobility:
h 0i
se e Zeme [5]
In titanates, the concentration of charge carriers depends on
the concentration of Ti3 ions in the sample. If the electrons
are not localized (trapped) on the cations, and thus are free to
take part in conduction, then [e0 ] [Ti3]. This is usually the
case in the systems where only Ti3/Ti4 occupy the B-site, as
will be shown in the next subsection (Figure 15(a)). The
mobility of the electrons depends mainly on the impact of
the crystal composition and structure on the TiOTi bonds,
and the microstructure of the sample (for polycrystalline sam-
ples). Factors including tilting, size of the unit cell, and the
nature of the A-ion can change the degree of overlap between
the t2g Ti orbitals and the oxygen p orbitals, some of which give
rise to the conduction band.8,17,18 The microstructure of the
sample is very important and should always be considered
when comparing conductivity values.
4.15.3.4.2 Factors that decide the conductivity of titanates
Because the conductivity of titanates depends primarily on the
concentration of Ti3 ions, it will be greatly influenced by the
conditions in which the sample has been prepared and/or
processed. There are several means of changing the Ti3 con-
tent in a titanate, and all gravitate around the defect chemistry
of the starting composition and the way this changes when
exposed to different environments usually characterized by the
oxygen partial pressure and temperature.
The first step in which the Ti3 content can be controlled is
the synthesis stage. When prepared in air, the vast majority of the
Ti ions are at 4, and dopants are usually compensated for by
oxygen defects (oxygen excess or deficiency). For example, when
synthesized in air the oxygen-excess series discussed previously
can be written as Lax3Sr1x2Ti1e4Tie3O3x/2e/22 and it
will show semiconducting-type behavior with conductivities
408 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
around 103 S cm1 when measured in air at 900  C, owing to
the very dilute electronic conductivity ([Ti3] e). These low
conductivity values are typical for all titanates in oxidized
form. An important aspect to notice is that very high sintering
temperatures (in air) can act as slightly reducing, because of
the significant increase in the entropic factor. Because of this
effect, oxide ions can be lost from the lattice and electrons are
subsequently generated (e). As a direct consequence, samples
sintered at high temperatures, say above 1500  C, may show
slightly different structure (notice that the oxygen excess de-
creases as well) and conductivity compared to samples sintered
at 13501450  C. When sintered under reducing conditions, a
greater number of Ti3 ions can be generated and hence,
correspondingly more oxide ions can be removed. Typically
the perovskite will tend to become oxygen stoichiometric, so e
will tend to become equal to x. If the reducing conditions are
even stronger, the titanate could even step into the oxygen-
deficient domain for e > x. After being sintered under highly
reducing conditions, at 1650  C, the above-mentioned series
shows metallic behavior when measured at 1000  C under
reducing conditions (to preserve the reduced states) and values
of conductivity increase from 100 to 400 S cm1 when x in-
creases from 0.1 to 0.4.76 As it can be noticed, the difference in
the conductivity values lies on the order of five orders of
magnitude and directly reflects the distinct defect chemistry
of the same series depending on the conditions employed
during phase formation.
The conductivity of the phases sintered in oxidizing condi-
tion can be increased by subsequently reducing them. There are
several ways to conduct such reduction experiments and each of
these can provide complementary information about the reduc-
tion process in titanates. The main parameters that can be mea-
sured are the conductivity (s) and the extent of the reduction (d)
and the most important variables are the oxygen partial pressure
(pO2), the temperature at which the reduction occurs (TR), the
100.0
99.8
Weight loss (%)
99.6
La Sr0.4TiO3-d
99.4 0.4
Coarse powder 0.035 m2 g-1
Fine powder 0.384 m2 g-1
99.2
Temperature profile
0 200 400 600
Reduction time, tR (min)
Figure 13 Thermogravimetric analysis for two La0.4Sr0.4TiO3 powders performed in 5%H2/Ar at various isothermal steps. Reproduced from Neagu, D.;
Irvine, J. T. S. Chem. Mater. 2010, 22(17), 50425053, with permission.49
temperature at which the conductivity is measured (TM), and the
duration of reduction (tR). Three experiments that can make use
of these parameters are presented below.
By measuring s or d as a function of reduction time while
keeping all the other variables fixed, information about the
kinetics of the reduction process can be obtained. Such an
experiment is shown in Figure 7 and has already been dis-
cussed in Section 4.15.3.3.2. Measurements of d(tR) can be
done by using thermogravimetric analysis, where the weight
loss of the sample in 5%H2/Ar is recorded against time at
isothermal steps (Figure 13). Notice that the shape of the
curves in both experiments is the same, as it describes essen-
tially the same process: the removal of oxygens in d(tR) and the
subsequent generation of electrons in s(tR). The exponential-
type curve can be interpreted as arising from two diffusion
processes, one which is fast and is responsible for a small
fraction of the total weight loss (surface reduction) and the
other one which is slow and accounts for most of the weight
loss observed (bulk reduction).49,77 This is consistent with the
qualitative view of the reduction process presented in
Section 4.15.3.3.2. It is apparent from Figures 9 and 13 that
as the specific surface area of the samples used increases, the
reduction kinetics are improved, most likely due to the higher
contact area between the gas phase and the solid. It is worth
noting that this slow reduction kinetics is improved by doping
of cations that usually cause an increase of d as well (see
Section 4.15.3.5.2.2 and the work of Neagu and Irvine47).
Another type of experiment that can be performed is that in
which s is measured as a function of pO2. This gives informa-
tion about the nature of the charge carriers in relation to the
defect chemistry, across a wide range of pO2 values. Two dis-
tinct behaviors can be identified be examining Figure 14(a).
For the low pO2 domain the conductivity increases with de-
creasing pO2, which is known as n-type behavior; in this
regime, the charge carriers are electrons. For the higher pO2
1400
1200
1000
Temperature (C)
800
600
400
200
0
800 1000 1200 1400
 Perovskite Defect Chemistry as Exemplified by Strontium Titanate 409

-25.00 -20.00 -15.00 -10.00 -5.00 0.00

0.00 1.0
Ga
-0.50 A-site-deficient Sr0.4La0.4TiO3 Fe
Slater J Mater. Chem 1997 (7) 24952489 0.5 Al
-1.00
Mn
Ti
-1.50 0.0
Mg

log (s) (S cm-1)

log (conductivity)

Sc
-2.00
Slope = -0.210 -0.5
Slope = 0.201
-2.50
Slope = -0.246
-1.0
-3.00

m = -1/6
-3.50 -1.5
m = -1/4
5% Bsite mg @ 835 C
-4.00 Undoped @ 930 C
-2.0
-20 -18 -16 -14 -12 -10
-4.50
(a) log (pO2)
(b) log (pO2)

Figure 14 pO2 dependencies for various doping scenarios of SrTiO3: (a) undoped, Mg-doped and La-doped SrTiO3. Reproduced from Irvine, J. In
Perovskite Oxide for Solid Oxide Fuel Cells; Ishihara, T., Ed.; Springer: London, 2009, with permission.1 (b) oxygen-excess La4Sr8Ti11MO38g.
Reproduced from Miller, D. N.; Irvine, J. T. S. J. Power Sources 2011, 196(17), 73237327, with permission.74

domain, the conductivity increases with increasing pO2, which
is known as p-type behavior; in this regime, the charge carriers
are holes. The p-type character of the undoped composition is
probably due to the intrinsic oxygen vacancies being filled with
oxygen at higher pO2, which in turn generates holes1:
1 
V x
O O2 $ OO 2h
2
Similarly, some of the extrinsic oxygen vacancies generated
by Mg2 doping in SrTiO3 can be filled with ambient oxygen at
higher pO2 values, showing p-type behavior.
The n-type behavior has already been shown to be due to
the removal of oxide ions at low pO2 and the subsequent
generation of electrons:
0 1
OxO $ 2e V 
O O2
2
The high n-type conductivity observed at low pO2 for the
A-site-deficient sample might be related to the diminished
number of intrinsic Schottky defects due to the large number
of A-site vacancies introduced by doping:
A xA OxO 
 AO V  00
! O VA [8]
This in turn will push eqn [7] to the right, facilitating the
removal of lattice oxygen by the reducing gas.1
The slopes of the dependencies log(s) versus log(pO2) re-
flect the defect equilibrium and the charge compensation of
the defects, but this will not be detailed here, as numerous
examples can be found in the literature.2,48,58,59
Another experiment that can be envisaged attempts to corre-
late s and d with the reduction temperature, and also correlate
between them, in the following manner. Samples are reduced in
5%H2/Ar at increasingly higher temperatures TR, and their ex-
tent of the reduction is measured while performing the reduc-
tion itself. The conductivity of the samples with known extent of
reduction is then measured at TM  880  C in the same gas. It is
important to note that thermogravimetric analysis shows that,
upon cooling, the extent of the reduction achieved at high
temperatures is frozen-in, if the pO2 is low enough. Therefore,
the concentration of charge carriers should not vary significantly
between room temperature and TM, but only their mobility.
Finally, the conductivity is plotted against its corresponding
extent of reduction. Such a plot is presented in Figure 15 for
the A-site-deficient La0.4Sr0.4TiO3. A clear linear correlation can
be observed between s and d, as expected. This is because
s [Ti3]eme and [Ti3] 2d, which follows that s kd. The
[6] fact that the electronic conductivity increases with the extent of
the reduction is an important conclusion and has led us to
propose it as a method for enhancing electronic conductivity
in titanates through increasing their extent of reduction.47 How-
ever, an interesting effect occurs when doping the B-site with a
cation capable of enhancing the extent of reduction. Although d
increases with the dopant, the conductivity does not follow the
same trend, but reaches a maximum, after which it declines. This
effect is depicted in Figure 15. We attribute this effect to the fact
[7] that the dopant cannot support electronic conduction and,
therefore, as its concentration on the B-site increases, the perco-
lation between the Ti ions decreases, effectively decreasing the
mobility of the charge carriers. Notice, however, that for only
6% Ga doping the conductivity has doubled, reflecting a dou-
bling in the extent of the reduction.
Another important observation is that according to Figure 11,
the d value at 900  C for La4Sr8Ti12O38g (La0.33Sr0.67TiO3g0 )
is double compared to La0.4Sr0.4TiO3, but Figures 14 and 16
show almost equal conductivities when equilibrated at the same
temperature. This might be attributed to several causes. It is
possible that the lower than expected conductivity of the
oxygen-excess composition is due to its typical microstructure
exhibiting a greater number of grain boundaries (see Figure 5).
On the other hand, the randomly distributed intergrowths
might disrupt conduction network throughout the grain (at the
intergrowths the corner sharing between octahedra is broken),
leading to an overall decrease in electron mobility. In fact,
recent studies on the RuddlesdenPopper series Srn1TinO3n1
have pointed out that the electron density around the rock-salt
intergrowths is very small and has a negative impact on the
electronic conduction of the structure.78
A different behavior in mobility with respect to temperature
is observed for the two classes of perovskites as well. Figure 16
shows the conductivity as a function of temperature. As
410 Perovskite Defect Chemistry as Exemplified by Strontium Titanate

50
45 0.060
La Sr Ti41+2dTi32+dO 50
0.4 0.4 3d

Electronic conductivity, s (S cm-1)

40

Oxygen deficiency, d (at/uc)

35 r 63% 45 0.055
r
Conductivity, s (S cm-1)

30 880 C
s 540 d 0.050
25 40

20
35 0.045
15 3+
(La0.4Sr0.4)(Ga3x+Ti41+-x-2dTi2d)O3-x/2-d
10
30 0.040
5 Experimental points s
Linear fit d
0 25 0.035
0.00 0.02 0.04 0.06 0.08
0.00 0.03 0.06 0.09 0.12 0.15
(a) Extent of reduction, d (b) x (Ga)
Figure 15 The correlation between conductivity and the concentration of charge carriers when (a) only Ti occupies the B-site and (b) when Ti is
partially substituted with a cation that does not support electronic conduction. (a) The data were obtained by reducing 63% dense pellets at
temperatures between 1000 and 1300  C and subsequently measuring their extent of reduction and conductivity (at 880  C) see the work of Neagu and
Irvine49; (b) conductivity (at 880  C) and extent of reduction were measured after a reduction at 1000  C on a series of Ga-doped samples (composition
specified on the plot) see the work of Neagu and Irvine.47

14

12 La0.2Sr0.7TiO3
Electronic conductivity (S)

10

2
La0.3Sr0.7TiO3+g

0
300 450 600 750 900 1050
Temperature (K)
Figure 16 Conductivity against temperature for an A-site-deficient and an oxygen-excess sample measured after being equilibrated in 5%H2/Ar, at
900  C. Reproduced from Tsekouras, G.; Irvine, J. T. S. J. Mater. Chem. 2011, 21, 93679376, with permission.77

explained above, oxygen content is usually frozen-in after 4.15.3.5 B-Site Cation Segregation from A-Site-Deficient
equilibrating the sample and while cooling it, if of course, the Perovskites
pO2 is kept reducing.49 It follows that the concentration of the
4.15.3.5.1 Reasons behind the phenomenon
charge carriers is constant with temperature, and thus the plot
Recently it has been suggested that when increasing the extent
shown in Figure 16 is actually a plot of the mobility as a
of the reduction of an A-site-deficient perovskite past a limit
function of temperature. Both the A-site deficient and the
(dlimit), the overall concentration of vacancies per unit cell
oxygen-excess sample show a transition from a semiconduct- 00
increases to the point where VSr and [VO ] are likely to start
ing to a metallic behavior, but the transition occurs at a higher 49,71
associating. Locally, the unit cells in which this association
temperature (500  C) and is much broader for the oxygen-
occurs will be effectively lacking the A-site ion and at least one
excess sample.
 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
Vo Ti3+
4+ O2-
(La,Sr)0.8TiO3-d Ti
H2, t > 1000 C
VA
(La,Sr)0.8TiO3-d limit
(La,Sr)0.8+xTiO3-d TinO2n-1
Figure 17 B-site cation segregation from A-site-deficient perovskites
under reducing conditions. When reduced past a limit, A-site-deficient
perovskites can be forced to form precipitates containing B-site cations
due to the impossibility of the lattice to accommodate such a high
number of defects at once. The inset shows an SEM image of such a
phenomenon arising when reducing the A-site-deficient perovskite
La0.4Sr0.4TiO3.32
oxygen ion, as depicted in Figure 17. At this stage, the unit cell
only consists of B-site cations and the remaining oxide ions
and it will be energetically favorable for them to rearrange
their packing and thus precipitate as oxides or metal, depen-
ding how easy they are to reduce. This will lead to the decrease
of the number of vacancies, equivalent to an increased occu-
pancy on the A- and O-sites ((La,Sr)0.8xTiO3d0 , d0 < dlimit, see
Figure 17), thus making the lattice more stable. The microstruc-
ture resulting from such a phenomenon and first observed in the
A-site-deficient, oxygen stoichiometric La0.4Sr0.4TiO3 is shown
in the inset of Figure 17. By substituting Ti with elements
exhibiting different properties, interesting and potentially useful
effects can be produced and additionally more insight into this
phenomenon can be achieved.32
4.15.3.5.2 Different manifestations of the B-site segregation
4.15.3.5.2.1 Driving the formation of nanoparticles from
perovskite frameworks
When the substituting cations are easily reduced to metals,
as is the case for dopants such as Ni2 or Fe3, metallic nano-
particles containing the dopant precipitate on the parent pe-
rovskite grains (see Figure 18). When reduction cannot
411
proceed all the way to the metallic state, as for example in the
undoped or Mn-doped samples, oxides precipitate instead
(Figure 18(c)). The first aspect that stands out by comparing
the phenomenon in different systems is the size and morphol-
ogy of the precipitates. The exsolutions from Fe-doped samples
are generally bulky and have well-defined facets. On the con-
trary, the exsolutions originating from Ni-doped samples are
always very fine (usually below 50 nm) and therefore features
of their surface are hard to distinguish. The precipitates arising
from the Mn-doped sample are considerably larger than
the ones in the Fe-doped systems ( 1 mm) and are very similar
to the ones observed for the undoped sample. This Mn system
is also the only one in which we observe nanoparticle growth
at the grain boundaries too, and not only on the surface of
the grains.
The way in which the dopant is compensated is also im-
portant for many reasons. Compensating doping by balancing
the charge on the A-site as in La0.42xSr0.42xNixTi1xO3 rather
than by oxygen vacancies as in La0.4Sr0.4NixTi1xO3x allows
for a wider substitution range and also produces better dis-
persed exsolutions (compare Figure 18(b) with Figure 18(d)).
The latter effect is partially due to the higher extents of re-
duction achievable in the sample starting as an oxygen stoi-
chiometric one. Another reason might be related to the
location where precipitates form. Interestingly, the exsolutions
do not seem to form randomly, but at specific locations on the
surface of the grains, that is, at the edge of the terraces that
inherently form when preparing these A-site-deficient perov-
skites. This is more easy to see in the compositions starting as
oxygen deficient, for which the separation between the terraces
is significant (Figure 18(d)). By increasing initial oxygen stoi-
chiometry we have decreased the length of the terraces, thus
increasing the number of crystallographic features per surface
area from which the exsolutions can grow. As a result, the
oxygen stoichiometric compositions have a more even distri-
bution of precipitates throughout the structure compared to
the ones starting as oxygen deficient (compare Figure 18). The
fact that the exsolutions form at these particular sites suggests
that those might be the B-site-terminated layers which are also
A-site deficient.
It is worth noting that the exsolution process does not alter
the parent perovskite framework and it is most likely a surface
effect, as observed in the case of the reduced Ga-doped samples
(see next subsection). In terms of structure, we always see an
increase in symmetry after the reduction and subsequent re-
lease of precipitates. It is also interesting to note that the
produced nanoparticles do not tend to cluster after prolonged
exposure to wet reducing gases at 900  C (100 h), or during
their formation at temperatures as high as 1000  C. This sug-
gests the formed exsolutions might be pinned to the parent
framework, which would act in favor of preserving the overall
microstructure.
4.15.3.5.2.2 The coreshell structure of the reduced
Ga-doped, A-site-deficient titanates
An interesting effect arises when Ti is replaced by some Ga in
the series La0.4Sr0.4GaxTi1xO3x/2. As other A,O-site-deficient
perovskites this is prone to segregate the B-site dopant, Ga.
However, because Ga3 can reduce to Ga which is able to
form volatile species,79,80 no precipitates are observed, but
412 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
La0.4+xSr0.4-xFexTi1-xO3 (x = 0.09, dry, 930 C, 20 h)
100 nm
St Andrews SEI 5.0 kV X1900 10 mm WD 4.7 mm
(a)
La0.4Sr0.4MnxTi1-xO3-x/2 (x = 0.06, dry, 1000 C, 20 h)
1 m
(c)
Figure 18 Exsolutions formed in different systems after reduction performed in dry or wet (3%H2O) 5%H2/Ar, at various temperatures (specific
parameters indicated in brackets).
weight loss corresponding to some Ga loss is recorded instead,
under reducing conditions.47 Several experiments47 point to
the idea that Ga is only depleted from a thin layer covering the
grains, and not from the bulk of the samples, as depicted in
Figure 19(a). The Ga-depleted layer should consist of an
A-site-deficient titanate (La0.5Sr0.5)1a0 TiO3d0 with a0 becom-
ing smaller as Ga doping increases (1  a0 )  (0.8/1  x).
By monitoring the conductivity of these reduced samples
during a redox cycle, an interesting and potentially useful effect
can be observed. During the first stage (the oxidation), the
conductivity is expected to drop by around two orders of
magnitude due to the oxidation of Ti3 to Ti4. While this is
clearly the case for the undoped compositions, the drop is
less and less significant as the Ga doping level increases (see
Figure 19). The fact that, for 10% Ga doping, the conductivity
does not decrease below 2 S cm1 even after 2 h of oxidation is
quite remarkable. It seems likely that the Ga-depleted layer acts
as a barrier that prevents the oxidation of the bulk, which in
turn can still support electronic conduction. The fact that this
effect is more pronounced as Ga doping increases can be
explained by the oxygen exchange properties of the Ga-
depleted layer. As explained earlier, the titanate composing
this layer will have an a0 decreasing with initial Ga doping, x.
On the other hand, Figure 10(a) shows that decreasing the
A-site occupancy of a titanate significantly diminishes its oxy-
gen exchange capability. This means that once the oxygen
La0.4+2xSr0.4-2xNixTi1-xO3 (x = 0.06, dry, 930 C, 20 h)
100 nm
100 nm
1 mm
St Andrews SEI 5.0 kV X10,000 WD 5.1 mm
(b) La0.4Sr0.4NixTi1-xO3-x/2
( x = 0.03, dry, 930 C, 20 h then wet, 900 C, 100 h)
100 nm
St Andrews SEI 5.0 kV X20,000 1 mm WD 5.2 mm
(d)
vacancies in this layer are filled, it will be more and more
difficult to migrate oxygen through it as x increases. Finally,
when the samples are reduced again, the samples having
higher Ga doping are able recover more than half of their
equilibrium conductivity within minutes, as most of the grain
is still reduced. In this case however, the Ga-depleted layer does
not seem to play an important role probably because the
reduction of surface layers in A-site-deficient perovskites occurs
inherently fast, as shown in Figure 7.
4.15.3.5.3 Factors controlling B-site segregation
The most important factors that control the formation of the
precipitates seem to be the A-site deficiency, the nature of the
substituting ion, and the extent of the reduction. Indeed, we
do not observe exsolutions from A-site stoichiometric, oxygen-
excess compositions such as La0.3Sr0.7Ti0.94Ni0.06O3.09, but we
do observe them for other A-site-deficient compositions with
the same B-site stoichiometry. Further, we do not observe Ga
loss under reducing conditions from Ga-doped stoichiometric
compounds such as La1/3Sr2/3Ga1/3Ti2/3O3, even though the
amount of Ga is considerably higher in the latter. The number
of A-site vacancies per formula unit (a) we typically employed
were 0.20, but a 0.10 can also yield good results. Higher
values of a usually lead to difficulties in producing pure
phase compositions and could cause an ordering of the A-site
that typically hinders reduction. The benefit of using Ti as the
 Perovskite Defect Chemistry as Exemplified by Strontium Titanate 413

La0.4Sr0.4GaxTi1-xO3-x/2
2.0 Oxidation

1.5

1.0
x = 0.10

log(s/S cm-1)
0.5
x = 0.05

Reduction
0.0 1 S cm-1

x = 0.02
0.5
(La,Sr)1-a TiO3-d
La0.4Sr0.4GaxTi1-xO3-x/2-d 1.0
La0.4Sr0.4GaxTi1-xO3-x/2-d x = 0.00
e- 1.5

e- 2.0
0.0 0.5 1.0 1.5 2.0
Time (h)
(a) (b)

20 Reduction

La0.4Sr0.4GaxTi1xO3x/2d x = 0.10

15
Conductivity, s (S cm-1)

x = 0.05

x = 0.02
10
x = 0.00

0 5 10 15 20 25
(c) Time (h)
Figure 19 The coreshell effect: (a) after reduction of the oxidized (as-prepared) phase, a thin layer of Ga-depleted titanate will cover the grains, while the bulk
remains unchanged; (b) this layer acts as a barrier which prevents oxidation of the core which is still able to transport charge; (c) because the bulk has not been
fully oxidized, upon reduction it will recover half of the final conductivity within minutes. Reproduced from Neagu, D.; Irvine, J. T. S. Chem. Mater. 2011, 23(6),
16071617, with permission.

main cation in the B-site sublattice resides in the fact that it can
form a robust network of octahedra that will not easily collapse
on reduction owing to its strong preference for sixfold coordi-
nation. The other B-site cation(s) (M), on the contrary, should
be able to easily support multiple coordination numbers in
order to accommodate oxygen vacancies around them. This
will effectively place the dopant M in the unit cells that are
prone to become deficient and thus form the exsolutions. As
discussed previously, these types of cations also promote the
creation of vacancies through reduction, and are also the type
of cations that show catalytic activity as well, and are of interest
to precipitate. Several other reasons for which the dopant will
exsolve predominantly can be envisaged: the size mismatch
between the dopant and the host which upon precipitating
might alleviate local distortions; the fact that most dopants
we used in this study, such as Ni2 or Fe3, are much easier
to reduce to metals than Ti4 so that the metallic phase will
spontaneously separate from the oxide phase; in cases such as
Ga, in which the dopant can form volatile species, this can act
as an additional driving force.
The oxygen deficiency achieved on reduction is also a key
parameter that dictates the formation of the exsolutions, as it
encompasses multiple aspects of the reduction process (oxygen
vacancies created during reduction and number of reduced
B-site ions). We have found that generally when the perov-
skites show d around or below a value of 0.05, the exsolutions
do not form at all or are sporadic (Figure 20). Such is the case
with samples that inherently do not reduce to great extents (for
example, the undoped composition and the Cu-doped com-
position), or that cannot reach that value due to the pO2 not
being low enough, for example when the reducing gas contains
some steam (3%H2O in 5%H2/Ar). Because the removal of
414 Perovskite Defect Chemistry as Exemplified by Strontium Titanate

La0.4Sr0.4MxTi1-xO3-x(4-m)/2-d

0.14
Fe3+ (dry, 1000 C)

Frequently occurring
0.12
Ni2+ (dry, 1000 C)

Oxygen lost during reduction, d

0.10

Exsolutions
Fe3+ (dry, 900 C)
0.08 Mn3+ (dry, 1000 C)

Cu2+ (dry, 1000 C)

0.06
d ~ 0.05
0.04 Exsolutions
No or few

Fe3+ (wet, 1000 C)

0.02

Fe3+ (wet, 900 C)

0.00
0.00 0.03 0.06 0.09
m+
x (B-site dopant, M )
Figure 20 Domains of the extent of the reduction in which the exsolution phenomenon was observed in La0.4Sr0.4MxTi1xO3x(4m)/2 systems. The
plot shows the extent of the reduction (d) as a function of the stoichiometry (x) of dopant Mm after 20 h of reduction in the conditions specified in
brackets (dry/wet refers to the absence or the presence of 3%H2O in the reducing gas 5%H2/Ar, respectively).

oxygen is a thermally activated process, one can achieve higher 8. Goodenough, J., Ed.; In Localized to Itinerant Electronic Transition in Perovskite
d values by performing the reduction at higher temperatures Oxides; Springer: Berlin, New York, 2001.
9. Glazer, A. M. Acta Crystallogr. B 1972, 28, 33843392.
(albeit not practical). In this manner, the undoped composi-
10. Glazer, A. M. Acta Crystallogr. A 1975, 31, 756762.
tion can reach d > 0.04 by reducing at temperatures in excess of 11. Woodward, P. M. Acta Crystallogr. B 1997, 53(1), 3243.
1000  C and thus show exsolutions.49 12. Woodward, P. M. Acta Crystallogr. B 1997, 53(1), 4466.
13. Howard, C. J.; Stokes, H. T. Acta Crystallogr. A 2004, 61, 93111.
14. Howard, C. J.; Stokes, H. T. Acta Crystallogr. B 1998, 54(6), 782789.
4.15.4 Conclusion 15. Lufaso, M.; Woodward, P. Prediction of the Crystal Structures of Perovskites Using
the Software Program SPuDS. Chemistry Faculty Publications, 2001.
16. Hui, S.; Petric, A. J. Electrochem. Soc. 2002, 149(1), J1.
In summary, we have shown here that the defect chemistry of 17. Eng, H. W.; Barnes, P. W.; Auer, B. M.; Woodward, P. M. J. Solid State Chem.
SrTiO3 arising from different extrinsic defects is extremely di- 2003, 175(1), 94109.
verse. We also emphasized the importance of relating this to 18. Wolfram, T.; Ellialtoglu, S. Electronic and Optical Properties of D-Band
the structure and to probe structural factors underpinning the Perovskites. Cambridge University Press: New York, 2006.
19. Carpenter, M. A.; Howard, C. J.; Knight, K. S.; Zhang, Z. J. Phys. Condens. Matter
described phenomena. It is when the interaction between
2006, 18(48), 1072510749.
structure and defects comes into play that unique and poten- 20. Howard, C.; Lumpkin, G.; Smith, R.; Zhang, Z. J. Solid State Chem. 2004, 177(8),
tially useful consequences such as enhanced oxide ion mobility 27262732.
in oxygen-excess composition or B-site cation segregation in 21. Howard, C. J.; Kennedy, B. J. J. Phys. Condens. Matter 1999, 11, 32293236.
A-site-deficient titanates are produced. For a related chapter in 22. Kennedy, B. J.; Howard, C. J.; Thorogood, G. J.; Mestre, M. A. T.; Hester, J. R.
J. Solid State Chem. 2000, 155(2), 455457.
this Comprehensive, we refer to Chapter 2.01. 23. Kennedy, B. J.; Howard, C. J.; Chakoumakos, B. C. J. Phys. Condens. Matter
1999, 11, 14791488.
24. Slater, P. R.; Irvine, J. T. S.; Ishihara, T.; Takita, Y. J. Solid State Chem. 1998,
References 139(1), 135143.
25. Dabrowski, B.; Chmaissem, O.; Mais, J.; Kolesnik, S.; Jorgensen, J. D.; Short, S.
1. Irvine, J. In Perovskite Oxide for Solid Oxide Fuel Cells; Ishihara, T., Ed.; Springer: J. Solid State Chem. 2003, 170(1), 154164.
London, 2009. 26. Attfield, J. P. Cryst. Eng. 2002, 5(34), 427438.
2. Moos, R.; Hardtl, K. H. J. Am. Ceram. Soc. 1997, 80(10), 25492562. 27. Garcia-Fernandez, P.; Aramburu, J. A.; Barriuso, M. T.; Moreno, M. J. Phys. Chem.
3. Canales-Vazquez, J.; Smith, M. J.; Irvine, J. T. S.; Zhou, W. Adv. Funct. Mater. Lett. 2010, 1(3), 647651.
2005, 15(6), 10001008. 28. Howard, C. J.; Kennedy, B. J.; Woodward, P. M. Acta Crystallogr. B 2003, 59(4),
4. Shannon, R. D. Acta Crystallogr. A 1976, 32(5), 751767. 463471.
5. Megaw, H. D. Ferroelectricity in Crystals. Methuen & Co, Ltd: London, 1957. 29. Smith, K. L.; Lumpkin, G. R.; Blackford, M. G.; Colella, M.; Zaluzec, N. J. J. Appl.
6. Allan, N. L.; Dayer, M. J.; Kulp, D. T.; Mackrodt, W. C. J. Mater. Chem. 1991, 1(6), Phys. 2008, 103(8), 083531.
10351039. 30. Battle, P.; Bennett, J. E.; Sloan, J.; Tilley, R. J. D.; Vente, J. F. J. Solid State Chem.
7. West, A. R. Basic Solid State Chemistry, 2nd ed.; Wiley: New York, 1999. 2000, 149(2), 360369.
 Perovskite Defect Chemistry as Exemplified by Strontium Titanate
31. Zhang, Z.; Howard, C. J.; Knight, K. S.; Lumpkin, G. R. Acta Crystallogr. B 2006,
62(1), 6067.
32. Neagu, D.; Irvine, J. T. S. Manuscript in preparation
33. Ruiz-Morales, J. C.; Canales-Vazquez, J.; Savaniu, C.; Marrero-Lopez, D.;
Zhou, W.; Irvine, J. T. S. Nature 2006, 439(7076), 568571.
34. Bowden, M. E.; Jefferson, D. A.; Brown, I. W. M. J. Solid State Chem. 1995,
117(1), 8896.
35. Harvey, E. J.; Ashbrook, S. E.; Lumpkin, G. R.; Redfern, S. A. T. J. Mater. Chem.
2006, 16(48), 46654674.
36. Ruddlesden, S. N.; Popper, P. Acta Crystallogr. 1957, 10, 538539.
37. Ruddlesden, S. N.; Popper, P. Acta Crystallogr. 1958, 11, 5455.
38. Bednorz, J. G.; Muller, K. A. Z. Phys. B: Condens. Matter 1986, 64, 189193.
39. Aurivillius, B. Ark. Kemi 1949, 1, 499.
40. Kusainova, A. M.; Stefanovich, S. Y.; Irvine, J. T. S.; Lightfoot, P. J. Mater. Chem.
2002, 12(12), 34133418.
41. Kusainova, A. M.; Zhou, W.; Irvine, J. T. S.; Lightfoot, P. J. Solid State Chem.
2002, 166(1), 148157.
42. Dion, M.; Ganne, M.; Tournoux, M. Mater. Res. Bull 1981, 16(11),
14291435.
43. Jacobson, A. J.; Johnson, J. W.; Lewandowski, J. T. Inorg. Chem. 1985, 24(23),
37273729.
44. Williams, T.; Schmalle, H.; Reller, A.; Lichtenberg, F.; Widmer, D.; Bednorz, G.
J. Solid State Chem. 1991, 93(2), 534548.
45. Amaral, L.; Senos, A. M. R.; Vilarinho, P. M. Mater. Res. Bull 2009, 44(2),
263270.
46. Fu, Q. X.; Mi, S. B.; Wessel, E.; Tietz, F. J. Eur. Ceram. Soc. 2008, 28(4), 811820.
47. Neagu, D.; Irvine, J. T. S. Chem. Mater. 2011, 23(6), 16071617.
48. Hashimoto, S.; Kindermann, L.; Poulsen, F.; Mogensen, M. J. Alloys Compd.
2005, 397(12), 245249.
49. Neagu, D.; Irvine, J. T. S. Chem. Mater. 2010, 22(17), 50425053.
50. Sgaard, M.; Vang Hendriksen, P.; Mogensen, M. J. Solid State Chem. 2007,
180(4), 14891503.
51. Vashook, V.; Vasylechko, L.; Ullmann, H.; Guth, U. J. Alloys Compd. 2002,
340(12), 263269.
52. Hui, S.; Petric, A. Mater. Res. Bull 2002, 37(7), 12151231.
53. Dai, H.; He, H.; Li, P.; Gao, L.; Au, C.-T. Catal. Today 2004, 90(34), 231244.
54. Liang, R.; Nakamura, T.; Kawaji, H.; Itoh, M.; Nakamura, T. Physica C 1990,
170(34), 307314.
55. Kishio, K.; Shimoyama, J.-ichi; Hasegawa, T.; Kitazawa, K.; Fueki, K. Jpn. J. Appl.
Phys. 1987, 26, L1228L1230.
56. Coey, J. M. D.; Viret, M.; von Molnar, S. Adv. Phys. 1999, 48, 167293.
57. Dabrowski, B.; Dybzinski, R.; Bukowski, Z.; Chmaissem, O.; Jorgensen, J. J. Solid
State Chem. 1999, 146(2), 448457.
58. Slater, P. R.; Fagg, D. P.; Irvine, J. T. S. J. Mater. Chem. 1997, 7(12), 24952498.
415
59. Blennow, P.; Hagen, A.; Hansen, K.; Wallenberg, L.; Mogensen, M. Solid State Ion.
2008, 179(3536), 20472058.
60. Hui, S.; Petric, A. J. Eur. Ceram. Soc. 2002, 22(910), 16731681.
61. Ishihara, T.; Matsuda, H.; Azmi bin Bustam, M.; Takita, Y. Solid State Ion. 1996,
8688(1), 197201.
62. Ishihara, T. In Perovskite Oxide for Solid Oxide Fuel Cells; Ishihara, T., Ed.;
Springer: London, 2009.
63. Iwahara, H. In Perovskite Oxide for Solid Oxide Fuel Cells; Ishihara, T., Ed.;
Springer: London, 2009.
64. Kharton, V. V.; Marques, F. M. B.; Atkinson, A. Solid State Ion. 2004, 174(14),
135149.
65. Kilner, J. A.; Berenov, A.; Rossiny, J. In Perovskite Oxide for Solid Oxide Fuel Cells;
Ishihara, T., Ed.; Springer: London, 2009.
66. Mogensen, M.; Lybye, D.; Bonanos, N.; Hendriksen, P. V.; Poulsen, F. W. Solid
State Ion. 2004, 174(14), 279286.
67. Yashima, M. In Perovskite Oxide for Solid Oxide Fuel Cells; Ishihara, T., Ed.;
Springer: London, 2009.
68. Yashima, M. Solid State Ion. 2008, 179(2126), 797803.
69. Tao, S.; Irvine, J. T. S. Nat. Mater. 2003, 2(5), 320323.
70. Escudero, M. J.; Irvine, J. T. S.; Daza, L. J. Power Sources 2009, 192(1), 4350.
71. Islam, M. S.; Davies, R. A. J. Mater. Chem. 2004, 14(1), 86.
72. Kendrick, E.; Islam, M. S.; Slater, P. R. Solid State Ion. 2005, 176(3940),
29752978.
73. Slater, P. R.; Irvine, J. T. S. Solid State Ion. 1999, 124(12), 6172.
74. Miller, D. N.; Irvine, J. T. S. J. Power Sources 2011, 196(17), 73237327.
75. Marina, O. A.; Canfield, N. L.; Stevenson, J. W. Solid State Ion. 2002, 149(12),
2128.
76. McColm, T. D.; Irvine, J. T. S. Solid State Ion. 2002, 152153, 615623.
77. Tsekouras, G.; Irvine, J. T. S. J. Mater. Chem. 2011, 21, 93679376.
78. Sirikanda, N.; Matsumoto, H.; Ishihara, T. Solid State Ion. 2010, 181(57),
315321.
79. Butt, D. P.; Park, Y.; Taylor, T. N. J. Nucl. Mater. 1999, 264(12), 7177.
80. Stanislowski, M.; Seeling, U.; Peck, D.-H.; Woo, S.-K.; Singheiser, L.; Hilpert, K.
Solid State Ion. 2005, 176(3536), 25232533.
Relevant Websites
http://www.eastchem.ac.uk EaSTCHEM: Professor John TS Irvine, EPSRC Senior
Fellow.
http://chemistry.st-and.ac.uk EaSTCHEM, University of St. Andrews: Prof. John Irvine
Group.
 2.03 Ruddlesden-Popper Phases and Derivatives: Homologous Series of
Transition Metal Oxides
B Raveau, Universite de Caen, Caen, France
2013 Elsevier Ltd. All rights reserved.

2.03.1 Introduction 63
2.03.2 Intergrowth of Stoichiometric Perovskite Layers with Single RS Layers 64
2.03.2.1 The n 1 Members 64
2.03.2.2 The n 2 Members 65
2.03.2.3 Other n Members 65
2.03.3 Oxygen-Deficient RP Phases 65
2.03.3.1 The n 1 Members 65
2.03.3.2 The n 2 Members 66
2.03.4 Intergrowth of Perovskite Layers with Multiple RS-Like Layers: The Thallium and Bismuth RP Derivatives 68
2.03.4.1 Intergrowths Involving Double RS-Like Layers: [R. S.]2.[Pe]n Intergrowths 69
2.03.4.2 Intergrowths Involving Triple RS Layers, [RS]3-[Pe]n 71
2.03.4.2.1 The n 1 members 71
2.03.4.2.2 The n 2 members 75
2.03.4.2.3 The n 3 and n 4 members 76
2.03.4.3 Intergrowths Involving Larger RS Layers 77
2.03.5 Creation of Ordered Oxygen Vacancies in Perovskite and RS Layers: The Lead and Mercury RP Derivatives 77
2.03.5.1 The Lead Cuprates 78
2.03.5.2 Mercury Cuprates 79
2.03.6 Insertion of Additional Fluorite-Type Layers between Pyramidal Copper Layers in Cuprates 83
2.03.7 Substitution of Carbonate Layers for RS Layers in RP-Type Phases: Copper Oxycarbonates 85
2.03.8 Shearing Perpendicular to RP Layers: The Collapsed Layered Oxides 88
2.03.9 The Sr4Fe6O13-Type Structure 91
2.03.10 Tubular Oxides 91
2.03.11 Conclusion 99
References 99

2.03.1 Introduction
Transition metal oxides1 have been studied for more than 60
years for their wide range of fascinating chemical and physical
properties, and it is only during the past 20 years that it was
realized that they could have a great potential as functional
materials. The originality of these materials stems from the
ability of the transition element to accommodate at least two
oxidation sates, leading to a mixed valence. The carriers, holes or
electrons, may be delocalized in the structure, leading to excep-
tional transport properties, but they may be also trapped in the
framework, leading to charge ordering phenomena and to par-
ticular magnetic and magneto transport properties. The first and
second series of transition elements all exhibit this mixed
valence peculiarity, which cause the corresponding oxides to
be characterized by complex nonstoichiometry phenomena
SrO. In these structures, the intergrowth
that have been investigated and described by many authors.
Among those systems, the oxides of the group Fe, Mn, Co, and
Cu occupy a particular place because they can adopt several
different coordinations, from octahedral, square, or trigonal
pyramidal, to tetrahedral or square planar , in addition to their
ability to exhibit several oxidation states: Fe2, Fe3, Fe4;
Mn2, Mn3, Mn4; Co2, Co3, Co4; and Cu, Cu2, and
Cu3. As a consequence, the metal-oxygen framework
associated with these elements is more flexible and susceptible
to large deviations from an ideal close-packed structure, gener-
ating a new physics that requires accurate knowledge of the
crystal chemistry of these oxides. This type of chemistry allowed
the discovery of high-temperature superconductivity in layered
cuprate2 in 1986, the colossal magnetoresistance in the manga-
nite perovskites3,4 in 1995, and, more recently, the great poten-
tial for thermoelectricity in cobaltites. This chapter focuses on
recent progress in the knowledge of the crystal chemistry of
layered oxides of the Ruddlesden and Popper (RP) family and
derivatives containing Mn, Fe, Co, or Cu.
First discovered for the titanates by Ruddlesden and
Popper,5 the tetragonal structure of these compounds
(Figure 1) consists of an intergrowth of single SrO rock salt
(RS) layers with multiple perovskite layers, according to the
formulation (SrTiO3)n
phenomenon consists of the alternate stacking of two different
structures, the perovskite (Pe) and the RS structures, which
both exhibit one identical single SrO plane (001) characteristic
of the RS structure. Thus, a single RS SrO layer is built of two
single SrO planes. Each SrO plane belongs to both the perov-
skite and the RS layer and consequently is counted as 1/2 in the
single RS layer and 1/2 in the perovskite layer, leading to the
generic formula given above. In a more general way, the

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00203-5 63

64 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
(a) n=1 (b)
Figure 1 Structure of the Ruddlesden and Popper (RP) oxides (SrMO3)n
SrO (a) n 1, (b) n 2, (c) n 3.
intergrowth structures (ATiO3)n.(AO)m, also labeled
[Pe]n[RS]m, are built up of n perovskite subcells and m AO RS
layers, that is, (m 1)AO planes; two extreme AO planes being
shared with the perovskite layers and counted as 1/2.
The exploration of oxide systems involving Cu, Co, Mn,
and Fe has shown that such structural types are observed for
these different elements but are limited to the first members of
the series, that is, namely n 1, 2. However, the crystal chem-
istry of these oxides seems to be more complex than for tita-
nates because of the appearance of oxygen deficiency in
connection with the JahnTeller properties of these cations.
Moreover, by introducing additional elements such as Tl, Bi, or
Hg, the thickness of the RS layer can be increased, leading to
numerous derivatives.
All of these oxides with various thicknesses of the perovskite
and RS layers exhibit various oxygen deficiencies, various
orderings of anionic vacancies, insertions of fluorite layers,
and substitution of carbonate layers for RS layers, leading to
the large homologous series that are described herein.
The various thicknesses of the layers, which may vary from
n 1 and m 1 to n 1 and to m 4, have an important
impact on the dimensionality of the structure, from two-
dimensional (2D) for low n values to 3D for n 1 (perov-
skite), upon the physical properties of such materials, namely
magnetism (e.g., colossal magnetoresistance) and transport
properties (e.g., high-Tc superconductivity). In the same way,
the nature of the cations of the different layers and the intro-
duction of vacancies or excess oxygen in the different layers,
which may involve charge transfers, have a great impact upon
the physical properties of these materials and play a major role
for the tuning of such properties.
2.03.2 Intergrowth of Stoichiometric Perovskite
Layers with Single RS Layers
2.03.2.1 The n 1 Members
The n 1 members of this homologous series, also called
K2NiF4-type compounds, with the generic formula A2MO4, con-
sist of single perovskite layers intergrown with single RS layers
(Figure 1(a)). In the case of manganese, they are synthesized for
a pure tetravalent state of this element, as in Ca2MnO468 and
Sr2MnO4.9 In these two compounds, the symmetry depends on
n=2 (c) n=3
the size of the A-site cation, that is, Sr2MnO4 exhibits the undis-
torted tetragonal I4/mmm ap  12.5 A cell, whereas for
Ca2MnO4, a superstructure
p corresponding to the I41/a symmetry
with  a  ap 2 and c 2  12.1 A is observed. The latter is due
to a distortion of the manganeseoxygen framework. One inter-
esting feature of these manganese oxides is their ability to adopt
the mixed valence Mn3/Mn4 by partial substitution of the
divalent alkaline earth cation by a trivalent lanthanide cation.
This is the case for Sr2xLnxMnO4, with Ln La, Pr, and Nd,1014
synthesized for x ranging from 0.50 to 1.2, all showing the
undistorted tetragonal cell a  ap, c  1213 A. Similarly,
mixed valent manganites Ca2xLnxMnO4 containing calcium
15
were synthesized for p Ln Pr, Nd, Sm, Eu, and Gd, with
x  0.50 and the ap 2  12 tetragonal cell.
A large series of cobaltites Ln2xSrxCoO4 exhibit the n 1
RP-type structure. The end-member Sr2CoO4, which formally
contains Co4, exhibits the undistorted I4/mmm structure,16,17
whereas the other end-member, La2CoO4, exhibits an ortho-
rhombic cell (Cmca) at room temperature, with tetragonal
symmetry (I4/mmm) observed only above 410 K.1820 For
intermediate compositions, in the range 0  x  1.40, the
Ln2xSrxCoO4 cobaltites mainly crystallize in the tetragonal
I4/mmm structure.2125 The range of solid solution in these
cobaltites depends on the size of the A-site cations. An upper
boundary of the solid solution was found at x 1.4 for Ln La,
x 1.3 for Ln Nd, and x 1.2 for Ln Gd.22 The lower solid
solution limit generally increases as the ionic radius of the
rare-earth cation becomes smaller, leading, for example, to
x 1.2 for Ln Dy.26 An important characteristic of the
Ln2xSrxCoO4 oxides is their ability to exhibit structural transi-
tions versus temperature, similar to La2CoO4. This is the case for
La1.7Sr0.3CoO4, which shows a similar transition to La2CoO4
at 227 K,27 and for La1.5Sr0.5CoO4, which shows a transition
from the room temperature tetragonal structure I4/mmm to an
orthorhombic symmetry (F4/mmm), the F4/mmm symmetry
at low temperature.28 The systematic studies of the LnSrCoO4
cobaltites, with Ln La, Ce, Pr, Nd, Eu, Gd, and Tb, have shown
that they all exhibit the I4/mmm structure at room temperature
with a decrease of their cell parameters as the size of Ln3
decreases.29 In contrast to cobaltites and manganites, the num-
ber of ferrites that exhibit the K2NiF4 structure is limited to the
oxides Sr2FeO4,30 Sr2xLaxFeO4 with 0.60  x  1,31,32
SrNdFeO4,33 CaLaFeO4,34 and CaNdFeO4.35 All of the oxides
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
containing strontium crystallize in the undistorted I4/mmm
tetragonal ap  12 Apstructure, whereas the calcium-based Sm2SrCo2O7,52,53 which contains only trivalent cobalt and has
phases exhibit the ap 2  12 tetragonal cell.
Because of the difficulty in stabilizing the trivalent state
of copper at ambient pressure, there are very few oxygen
stoichiometric cuprates that exhibit the K2NiF4 structure
when they are synthesized under ambient oxygen pressure.
36,37
La2CuO p4is onepof
them, crystallizing in an orthorhombic
cell ap 2  ap 2  13 , resulting from a monoclinic distor-
tion in the basal plane (001) of the classical tetragonal struc-
ture ap  13 A. The mixed valent Cu2/Cu3 tetragonal oxides
were stabilized by substituting alkaline earth cations for lan-
thanum in La2CuO4, according to the formula La2xAxCuO4,
with A Ca, Sr, and Ba.38,39 Nevertheless, it is quite remarkable
that, at ambient oxygen pressure, the stabilization of the per-
fectly stoichiometric O4 oxides is observed only for x 0:20,
leading to high-temperature superconductivity.2 For higher x
values, it is shown further that these structures become oxygen
deficient, due to the JahnTeller effect of copper. Only high
oxygen pressure synthesis allows stabilization of the O4 stoi-
chiometry, as shown, for instance, for LaSrCuO4.40 The stoi-
chiometric La2xAxCuO4 cuprates with 0.05 < x  0.20 exhibit,
like most of the other n 1 members of the family, the undis-
torted I4/mmm structure with a  ap and c  13 A.
2.03.2.2 The n 2 Members
These oxides, which correspond to the intergrowth of double
perovskite layers with single RS layers (Figure 1(b)), should
exhibit the stoichiometric composition A3M2O7. Such oxygen
stoichiometric oxides can only be synthesized for manganese,
iron, and cobalt. In the case of copper, the tendency of this
element to accommodate a lower oxidation state Cu2 and its
JahnTeller effect make an ordered oxygen-deficient n 2-RP-
type structure more preferable, as shown further (see Section
2.03.3.2).
The n 2-RP oxides with the tetragonal ap  20 A structure
have been synthesized for Sr3Mn2O7 containing Mn4 only41
and for the mixed valent Mn3/Mn4 oxides Sr2xLa1xMn2O742
and Sr2xNd1xMn2O7,43 with x ranging from 0 to 0.60.
Ca3Mn2O7 also belongs to pthis family
p of RP phases, but with
an orthorhombic cell ap 2  ap 2  19 .44 Moreover, Ca-
rich n 2-RP phases with the generic formula Ca2xLa1xMn2O7
were synthesized for 0  x  0.60.45,46 All of them crystallize in
the tetragonal ap  20 A structure.
In the case of ferrites, the presence of a lanthanide is neces-
sary to stabilize the O7 stoichiometry under a normal oxygen
pressure, as shown for the oxides SrLn2Fe2O7,4750 which
exhibit the tetragonal ap  20 A p structure
for Ln La, Nd,
Gd, and Dy and the tetragonal ap 2  20 cell for Ln Eu,
Tb. Nevertheless, the pure Fe(IV) oxide Sr3Fe2O7 can be syn-
thesized under high oxygen pressure,51 with the undistorted
ap  20 A tetragonal structure. In contrast to manganites, no
n 2-RP ferrite containing calcium seems to be known to date.
Very few n 2-RP cobaltites with the O7 stoichiometry have
been synthesized. Particularly, the pure ideal member Sr3Co2O7
was not observed. In contrast, this structural type exists with a
large deviation from the oxygen stoichiometry Sr3Co2O7x, as
described further (see Section 2.03.3.2). The stoichiometric O7
n 2-intergrowth can only be synthesized by introducing a large
65
amount of lanthanide into the structure. This is the case of
the tetragonal I4/mmm ap  20 A structure. The cobaltite
Sm2BaCo2O7 also belongs to this RP family, but its symmetry is
orthorhombic ap  ap  20 A instead of tetragonal.53,54 The
oxide Gd2SrCo2O7, was also found to exhibit a n 2-RP-type
structure, but its tetragonal ap  20 A structure displays a differ-
ent space group P42/mmm, instead of the ideal one (I4/mmm),
due to small distortions of the oxygen sublattice.55
2.03.2.3 Other n Members
The stabilization of superior n members, with n  3, keeping
the O10 stoichiometry is quite difficult, whatever the nature
of the transition element M Mn, Fe, Co, Cu. The manganite
Ca4Mn3O1056 is one of the rare
p examples.
p It exhibits an ortho-
rhombic Pbca symmetry ap 2  ap 2  27 (Figure 1(c)).
The phase Sr4Mn3O1057,58 does not belong to this structural
family.
2.03.3 Oxygen-Deficient RP Phases
Due to their ability to accommodate various coordinations,
smaller than six, the four elements Fe, Co, Mn, Cu are suscep-
tible to induce large oxygen deficiencies in the RP phases,
similarly to their behavior in the perovskite structure. This
tendency is reinforced for copper due to its pronounced
JahnTeller effect.
2.03.3.1 The n 1 Members
Due to its strong JahnTeller effect, Cu2 can accommodate at
least two coordinations (pyramidal and square planar) besides
its octahedral coordination, favoring a tendency to the ordering
of the oxygen vacancies in the RP structure. This is the case
of the cuprates Sr2CuO3 and Ca2CuO3,59,60 which crystallize
with an orthorhombic symmetry (a  3. 9 A  ap, b  3. 5 A  ap,
c 13 A). In fact, the structure of these oxides (Figure 2) is closely
related to that of the n 1 RP members. Like the latter, the
structure consists of single RS AO (A Ca, Sr) layers, but differ-
ently from the K2NiF4 structure, the AO layers are intercon-
nected through single layers of CuO4 square planar groups
instead of single layers of CuO6 octahedra. In fact, this structure
can be derived from the stoichiometric La2CuO4 structure by
elimination of oxygen rows in the perovskite layers in an ordered
way, resulting in the formation of [CuO2]/ rows of corner-
sharing CuO4 groups. This ordered creation of anionic vacancies
involves a compression of the AO layers along one direction,
leading to a decrease of the b parameter from 3.9 to 3.5 A. Thus,
the Sr2CuO3 structure can be described as an ordered oxygen-
deficient n 1 RP phase. The oxygen-deficient cuprates
La2xAxCuO4x/2d, synthesized for A Ca, Sr, Ba38,39 show a
crystal chemistry mainly dominated by the ability of copper to
adopt the pyramidal and square planar coordination. The homo-
geneity range of these phases depends on the nature of the A-site
cation, with 0  x  0.20 for A Ca, Ba and 0  x 1.33 or A Sr.
The symmetry of the RP-cell and its parameters depend on the
substitution rate and on the oxygen deficiency, that is, on the
oxygen pressure and temperature used during the synthesis.
66 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

This is illustrated by the oxides

p Lap
61
AxCuO4x/2d, which
2x
exhibit an orthorhombic ap 2  ap 2  13 cell for x  0.10
(see Section 2.03.2.1), and a tetragonal ap  13 A subcell for
0.25  x  1.34. Quite remarkably, for the highest x values
(1  x  1.34), these oxides are easily synthesized with a high
oxygen deficiency, corresponding to the formula La2xAx
CuO4x/2 (d 0), and, for these compositions, electron micros-
copy observations show superstructure reflections with a variable
periodicity in the (a, b) plane of the tetragonal subcell, that is,
c b
a
a  n  aT, depending on the substitution rate (Table 1). Such
superstructures are, in fact, due to various orderings of the oxygen
vacancies in the equatorial (a, b) plane of the tetragonal structure,
and can be strongly influenced by the methods of synthesis
(temperature, oxygen pressure).
In contrast to copper, cobalt and iron do not exhibit
oxygen-deficient n 1 RP phases, suggesting that the JahnTeller
effect of copper is crucial for the generation of oxygen defi-
ciency in single perovskite layers. In the case of manganese,
several phases involving an oxygen deficiency with respect to
the O4 n 1 RP structure have been synthesized: the oxides
Sr2MnO3.84 and Sr2MnO3.562 and the manganites LaSrMnO4d
with 0.30  0.50.63 Both manganites exhibit a tetragonal
ap  13 A cell, the anionic vacancies being distributed at ran-
dom in the equatorial plane of the MnO6 octahedra of the
ideal structure (Figure 1(a)).

2.03.3.2 The n 2 Members

!
Figure 2 View of the structure of Sr2CuO3 along b showing the rows of
CuO4 corner-shared groups (Cu blue spheres) interleaved with SrO
layers (Sr yellow spheres).
As for the n 1 intergrowths, copper is the most favorable
candidate for the generation of oxygen vacancies, and for their
ordering in the structure. The La-rich cuprates La2SrCu2O6
and La2CaCu2O664 represent the limit n 2 member of these
oxygen-deficient RP-families, which can be formulated as
La2xA1xCu2O6x/2d (Table 2), and show a narrow
homogeneity range, 0  x  0.14 for A Sr and 0  x  0.10 for
Ca. The tetragonal ap  20 A structure of the end-member
La2ACu2O6 (A Ca, Sr), in which copper is only divalent
(Figure 3), is closely related to that of the n 2 member
of the A3Mn2O7 intergrowth (Figure 1(b)). It consists of an
intergrowth of single AO-type layers with double layers of
corner-sharing CuO5 pyramids. In fact, this structure is directly
derived from the A3M2O7 structure by an ordered elimination
of the apical oxygen atoms located in the median plane of the
double perovskite layers. Quite remarkably, the La3 and A2
!
cations are preferentially ordered in the form of layers along c :
3
that is, the La cations sit mainly in the SrO-type layers
(7090% occupation), whereas the divalent cations are prefer-
entially distributed inside the double layers of CuO5 pyramids
at the level of the oxygen vacancies. The oxygen content in

Table 1 Crystallographic data of the cuprates La2xSrxCuO4x/2d, and values of n observed by electron diffraction for oxides of range III

Range x d Composition a(A) b(A) c(A)

I 0 0 La2CuO4 5.366 5.402 13.149

0.08 0.030 (1) La1.92Sr0.08CuO3.99 5.368 5.368 13.200
II 0.25 0.060 (4) La1.75Sr0.25CuO3.935 3.775 13.247
0.333 0.119 (4) La1.666Sr0.333CuO3.95 3.776 13.250
0.50 0.100 (4) La1.50Sr0.50CuO3.85 3.773 13.204
0.666 0.092 (4) La1.33Sr0.666CuO3.758 3.775 13.150
0.880 0.088 (4) La1.12Sr0.88CuO3.648 3.773 13.073
III 1.00 0.0 LaSrCuO3.5 3.767 13.002
1.28 0.0 La0.72Sr1.28CuO3.36 3.761 12.922
1.34 0.0 La0.66Sr1.34CuO3.33 3.759 12.907
Composition x n
LaSrCuO3.5 1.00 1,4.5
La0.88Sr1.12CuO3.44 1.12 1,4.5,5
La0.80Sr1.20CuO3.40 1.20 5
La0.72Sr1.28CuO3.36 1.28 1,4.6,5,5.3,5.4
La0.66Sr1.34CuO3.33 1.34 1,4,5,5.6,6
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 67

these oxides depends on the method of synthesis and espe-
cially on the oxygen pressure and on the annealing tempera-
ture. For instance, annealing of La2xA1xCu2O6x/2d under
an oxygen pressure of 1 bar allows intercalation of oxygen up
to d 0.29 for A Sr and up to d 0.08 for A Ca. This oxygen
uptake is of capital importance, since it induces the mixed
valence Cu3/Cu2 in these compounds, necessary for the
appearance of superconductivity at high temperature. In this
way, a critical Tc of 60 K could be reached in this system, using
oxygen pressures of about 20 bars.65
Sr-rich cuprates, Ln2xSr1xCu2O6x/2d, with
0.70  x  0.9 for Ln La, Sm, Eu, and Gd,66,67 and with x 1
for Ln Nd68,69 have been isolated. The oxygen content in
these oxides is smaller than O6 and their crystallographic
characteristics (Table 3) are very similar to those of La2Sr-
Cu2O6, but differ from the latter by a tripling of one of the
parameters, leading to an orthorhombic cell ap  3ap  20 A.
This structural type is illustrated by the cuprate Sr6Nd3Cu6O17
(NdSr2Cu2O5.66) on Figure 4. It derives from the stoichio-
metric intergrowth A3M2O7 by creating oxygen deficiency in
the double perovskite layers in an ordered way, so that double
YBa2Cu3O7-type layers are formed. Thus, Sr6Nd3Cu6O17 can
be described as an ordered oxygen-deficient n 2 RP phase
built up of double 123 perovskite layers stacked with single
RS SrO-type layers. The distribution of the strontium and
neodymium cations in this structure is remarkable and is

Table 2 Crystallographic data and composition for the oxides Table 3 Crystallographic parameters of the cuprates
La2xA1xCu2O6x/2d Ln2xSr1xCu2O6x/2d

Composition d Cell parameters Ln Composition x a(A) b(A) c(A)

a(A) c(A) Nd NdSr2Cu2O5.66 (or 1 3.755 11.488 20.098

Sr6Nd3Cu6O17)
(a) Quenched oxides (in air) NdSr2Cu2O5.76 1 3.770 11.438 20.094
La2SrCu2O6 0 3.865 19.887 Sm Sm1.30Sr1.70Cu2O5.65 0.70 3.746 11.416 20.068
La1.9Sr1.1Cu2O5.97 0.02 3.863 19.963 Sm1.10Sr1.90Cu2O5.55 0.90 3.747 11.439 20.058
La1.86Sr1.14Cu2O5.97 0.04 3.959 19.956 Eu Eu1.40Sr1.60Cu2O5.70 0.60 3.742 11.325 20.052
La1.9Ca1.1Cu2O5.97 0.02 3.825 19.404 Eu1.30Sr1.70Cu2O5.65 0.70 3.744 11.337 20.047
(b) Annealed oxides (in O2) Eu1.10Sr1.90Cu2O5.55 0.90 3.742 11.380 20.020
La2SrCu2O6.2 0.20 3.865 20.065 Gd Gd1.30Sr1.70Cu2O5.65 0.70 3.738 11.350 20.045
La1.86Sr1.14Cu2O6.22 0.29 3.868 20.051 Gd1.10Sr1.90Cu2O5.55 0.90 3.738 11.385 20.015
La1.9Ca1.1Cu2O6.02 0.08 3.825 19.404 La La1xSr2xCu2O5.5d 0.050.15 3.80 11.48 20.23

a
a
b
c
c

Figure!3 View of the 0212 structure of La2ACu2O6 (A Ca,Sr),
along b showing the double layers of corner-shared CuO5 pyramids,
intergrown with SrO layers (Sr yellow spheres).
!
Figure 4 View of the structure of Sr6Nd3Cu6O17 along a showing 123-
type layers intergrown with rock salt-type layers. The rock salt layers are
occupied by Sr2 (yellow spheres) and Sr2/Nd3 (blue spheres),
whereas Nd3 (green spheres) sit between CuO5 pyramids. Note that the
CuO4 groups between the CuO5 pyramids can be described as strongly
distorted tetrahedra.
68 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

certainly an important factor for its stabilization. One indeed be described as the intergrowth of single RS-SrO layers with
observes that the Sr2 cations sit preferentially in the RS layers, double ordered oxygen-deficient perovskite layers closely related
whereas the smaller Nd3 cations are preferentially located to the 123 structure of YBa2Cu3O7. In fact, in the double
between the CuO5 pyramids, instead of Y3 in the 123 layers. perovskite layers, rows of CoO6 octahedra are sandwiched
Note that such a cationic distribution may induce strains that between two double rows of CoO5 pyramids, so that they can
exist at the interface of the SrO-type layers and 123-type be described as 123 oxygenated layers. Orderings of the oxygen
layers. It results in a strong distortion and a tilting of the vacancies that are more complex, distributed in a semi-ordered
CuO4 square groups inside the 123 layers, which can be way within the double perovskite layers, were observed for
described as flattened tetrahedra. Smaller oxygen deficiency lower oxygen contents, such as in Sr3Co2O5.6478 and in
with respect to the ideal composition O5.66 is also observed cerium-substituted phases like Sr2.75Ce0.25Co2O5.9.79
(Table 2). The structure of these other Sr-rich phases, with The oxygen deficiency in n 2 RP ferrites is mainly limited
oxygen contents intermediate between O5.66 and O6, has to the two series of oxides Sr3Fe2O7d with 0  d  0.480 and
been studied by neutron diffraction for La1.1Sr1.9Cu2O5.83,70 Sr2.7La0.3Fe2O7d with d  0.20.81 All of these oxides exhibit
NdSr2Cu2O5.7669, and Nd1.4Sr1.6Cu2O5.79.71 The oxygen distri- the undistorted I4/mmm structure with a  ap and c  20 A.
bution in their structure can be interpreted as the intimate The oxygen vacancies are located on the apical sites of the
intergrowth of two extreme oxygen-deficient n 2-type RP FeO6 octahedra inside the double perovskite layers, forming
derivatives, corresponding to Sr6Nd3Cu6O17 and to a hypo- FeO5 pyramids displayed in a similar manner to LaSrCu2O6
thetical compound, Sr2LnCu2O6,61 respectively. In fact, the (Figure 3), so that the occupancy factor of these sites can be
ideal structure of the hypothetical Sr2LnCu2O6 end-member lowered down to 64% for Sr3FeO6.64 and to 84% for
is also built up of ordered oxygen-deficient double perovskite Sr2.7La0.2Fe2O6.84.
layers closely related to the 123 structure, but containing only
CuO5 pyramids instead of CuO4 and CuO5 polyhedra.
Manganites also exhibit an oxygen deficiency with respect to
2.03.4 Intergrowth of Perovskite Layers with
the n 2 member Sr3Mn2O7. This oxygen nonstoichiometry
Multiple RS-Like Layers: The Thallium and Bismuth
is also characterized by an ordering of the anionic vacancies.
RP Derivatives
However, the latter is different from that observed for the cup-
rates, as illustrated for the oxides Sr3Mn2O6.11 and
The introduction of trivalent thallium and trivalent bismuth
Sr3Mn2O6.19.72 For instance, in the structure of Sr3Mn2O6.11,
into the RS layers of the RP phases described above is possible
one observes that the anionic vacancies do not sit on the apical
in many cases, leading to the formation of either double
sites of the double perovskite layers, but are distributed in a
RS-like or triple RS-like layers. It will be seen further that, due
semi-ordered way in the basal plane
p of the MnO6 octahedra. It to the smaller size of Tl3, or to the 6s2 lone pair of Bi3, the
results in a tetragonal cell 2ap 2  20 , with the P4/mbm
so-formed layers exhibit a structural arrangement significantly
symmetry, where several oxygen sites are only partially occu-
different from the ideal RS layers, involving either a splitting of
pied, in a preferential way.
the Tl or O sites inside the layers with a disordered distribution
The oxygen-deficient cobaltites A3Co2O7d with A Sr, Ln
of the atoms over these sites; an incommensurability of the
exhibit a similar behavior to the cuprates. This is, for instance,
structure in the case of bismuth phases; or a possible oxygen
the case of the cobaltite Sr2Y0.8Ca0.2Co2O6,73,74 which crystal-
lizes in the tetragonal I4/mmm space group, ap  19 A, with a
similar structure to La2SrCu2O6 (Figure 3), built up of double a
b
oxygen-deficient ordered layers of corner-sharing CoO5 pyra-
mids intergrown with SrO RS-type layers. Note that, similarly
c
to the cuprate, an ordering of the Sr2 and of the Y3/Ca2
species takes place in the form of alternative layers stacked
!
along c , the smaller cations Y3/Ca2 sitting inside the double
pyramidal cobalt layers (Figure 3). In the same way, the oxides
Sr3Co2O7d, with 0.84  d  1 adopt the undistorted RP tetrag-
onal structure ap  20 A, as illustrated for the cobaltite
Sr3Co2O6.13,75 which consists also of the intergrowth of
double pyramidal cobalt layers and single SrO-RS layers, the
anionic vacancies between the pyramidal layers being filled
at 13% only by the additional oxygen with respect to O6.
Like the cuprates, the n 2 RP cobaltites can exhibit an oxygen
content smaller than O6. This was observed for the oxides
Sr3Co2O7d with 1  d  1.30,76,77 which crystallize in the
orthorhombic Immm symmetry with a ap, b 3ap, and
c 20A. The ordering of the oxygen vacancies is rather similar
!
to that observed for the Sr-rich cuprates with oxygen content Figure 5 View of the structure of Sr3Co2O5.82 along a showing
smaller than O6. They differ from the cuprates by the fact that oxygenated 123-type layers intergrown with single SrO rock salt layers
the CuO4 groups are replaced by CoO6 octahedra. This is exem- (Sr yellow spheres). Note that the MnO6 octahedra between the MnO5
plified by the structure of Sr3Co2O5.8277 (Figure 5), which can pyramids can be oxygen deficient.
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
deficiency, especially in the case of thallium phases, or oxygen
excess in the case of the bismuth phases. These complex struc-
tural phenomena have a great impact on the physical proper-
ties of these materials, as, for example, on superconductivity
and magnetism.
2.03.4.1 Intergrowths Involving Double RS-Like Layers:
[R. S.]2.[Pe]n Intergrowths
The first n 1 member of this series, called 1201, corresponds
to the ideal formula ASr2MO5 or ABa2MO5, with A Tl or Bi/Sr.
The ideal structure, which consists of double RS layers inter-
grown with single perovskite layers (Figure 6), was obtained for
TlSr2CoO582 and TlSr2CuO583,84 with a tetragonal cell
ap  8.9 A. In fact, this structure is very flexible, so that an excess
of strontium can be introduced in the RS layers, and thallium
can be replaced partly by bismuth, or lanthanide cations, or
even lead, and, simultaneously, oxygen vacancies can appear.
This is the case of the cuprates Tl1xPrxSr2yPryCuO5d,85
Tl0.5Pb0.5Sr2CuO5d,86 and TlBa2CuO5d87 and of the cobalt-
ites Sr2.6Tl0.4CoO5d88 and Bi0.4Sr2.6Co1.1O4.9,89 which all
exhibit the same tetragonal cell ap  8.9 A. In those oxides,
the oxygen vacancies are located in the RS layers and at the
apical sites of the octahedra. Moreover, the oxygen and thallium
positions in the TlO layers are often split in order to allow
correct TlO distances. Such a structure also involves a partial
ordering of the cations in the structure, leading to a multiple
cell. This is the case for the ferrite Bi0.4Sr2.5FeO5,90 which
a
b
Figure 6 View of ! the tetragonal 1201 structure of TlSr2CoO5 and
TlSr2CuO5 along b , showing single perovskite layers of CoO6 (CuO6)
octahedra intergrown with double RS-like layers. The latter are built up of
one TlO layer (Tl blue spheres, O red spheres) sandwiched between
two SrO layers (Sr yellow spheres).
69
p
exhibits a tetragonal structure ap 2  18 (Figure 7). In this
structure, one [010] SrO row alternates with one mixed (Bi/
! !
Sr)1O row along a in the median RS plane. Along c , one SrO
row alternates with one mixed row, leading to a doubling of the
c parameter. Moreover, in the SrO rows, the oxygen sites are
split with respect to their ideal positions, with one half-
occupancy factor. Similarly, the manganite Bi0.4Sr2.6MnO4.8
exhibitspa quadrupling of the c parameter, that is,
a  ap 2, c  36 .91 Its 1201 structure (Figure 8) shows a
similar alternation of the SrO rows and mixed (Bi/Sr)1O rows
!
within each RS plane, but the ordering of these rows along c is
different, according to the sequence SrSrBiSrBiSrSrSr.
The second n 2 member of the series, called 1212, which
would correspond to the ideal formula TlA3M2O8, with A Sr,
Ba, is rarely obtained with the O8 stoichiometry, due to redox
conditions that do not favor the coexistence of Tl3 with the
appropriate valence states of the transition elements. The fer-
rite Tl1xPbxSr3Fe2O8 with 0  x  0.592 is the only representa-
tive of the fully oxygenated 1212 structure (Figure 9). This
tetragonal ap  12.8 A structure can be described as a regular
intergrowth of double perovskite layers containing only Sr2
cations, with double RS layers, built up of one TlO layer
sandwiched between two SrO layers. The RS layers are dis-
torted, particularly the Tl and O positions, which are split
and 1/4 occupied in order to satisfy correct TlO distances.
Similarly to the 0212 structure of La2SrCu2O6, the ordered
oxygen-deficient 1212 structure is obtained for TlBa2Ca
Cu2O7,9395 TlSr2CaCu2O79698 and for TlBa2LnCu2O7 with
Ln Y, Nd,99 with the same tetragonal ap  12.8 structure
b
a
Figure!7 View of the 1201 orthorhombic structure of Bi0.4Sr2.5Fe1.1O5
along b showing the ordering of the SrO rows (Sr yellow spheres,
O red spheres) and mixed (Bi, Sr)1O rows (Bi/Sr blue spheres).
70 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

b b
a a

c c

Figure 8 View of the structure

! of the 1201 orthorhombic manganite Figure 9 View of!the tetragonal 1212 structure of the ferrite
Bi0.4Sr2.5Mn0.5O5x along b showing the complex ordering of the SrO TlSr3Fe2O8 along a . The Tl3 cations (blue spheres) and Sr2 cations
rows (Sr yellow spheres, O red spheres) and of the (Bi,Sr)1O rows (yellow spheres), form with oxygen (red spheres) double rock salt-type
(Bi/Sr blue spheres). layers. Note that the positions of Tl and O in the RS layers are split.

(Figure 10) built up of double layers of corner-shared CuO5
pyramids intergrown with distorted double-RS layers TlBaO2.
Similarly to the 0212 structure, the 1212 structure is charac-
terized by an ordering of the three types of cations Tl, Ba, or Sr,
Ca, or Ln in the form of layers, that is, the Tl3 cations sit in the
median plane of the RS layers, between two Sr2 (or Ba2)
layers, whereas the Ca2 or Ln3 cations are located within the
ordered oxygen-deficient double perovskite layers, that is,
between the CuO5 pyramids, leading to the following ideal
!
sequence along c : Sr-Tl-Sr-Ca-Sr-Tl-Sr. Quite remarkably, this
structure can exhibit a significant oxygen deficiency with
respect to the O7 composition and a thallium excess, accord-
ing to the formulas TlSr2Ca1xTlxCu2O7d9699 and
TlBa2Ca1xTlxCu2O7d.100,101 One observes that, for an excess
of Tl (0 < x < 0.20), the calcium sites are partially occupied by
the latter, and like for all of the Tl cuprates, the Tl sites are split
in the TlO layers, whereas oxygen vacancies (0 < d < 0.5) are
clearly seen in the RS layers. This 1212-type structure was also
extended to various rare-earth substitutions, such
as TlBa2LnCu2O7,102 TlSr2Ca1xLnxCu2O7,103,104 and (Tl,
Bi)1Sr2(Ca,Y)1Cu2O7d.105,106 Pure lead 1212 cuprates were
also synthesized with the compositions Pb0.5Sr2.5Y1xCax
Cu2O7d107,108 and Pb0.5Sr2.50Y1xCaxCu2O7d.109 Both oxides
exhibit a wide homogeneity range, 0  x  0.60, with super-
conducting properties for x  0.50. Note also that, for these
oxides, several superstructures are observed, which can be
interpreted in terms of ordering of lead, strontium, and anionic
vacancies.
The n  3 members with double RS-like layers are only syn-
thesized for thallium cuprates. Moreover, no fully oxygenated
phase involving CuO6 octahedra is obtained. TlBa2Ca2Cu3O9
called 1223110,111 represents the n 3 member of the series.
Its tetragonal structure (Figure 11(a)) is characterized by the
cell ap  16 A. It consists of triple oxygen-deficient perovskite
layers of copper polyhedra intergrown with double RS layers.
Each triple copper layer derives from the perovskite by remov-
ing the apical oxygen of the CuO6 octahedra in the n 3 mem-
ber, so that it can be described as one CuO2 layer of corner-
sharing CuO4 square planar groups sandwiched between
two layers of CuO5 pyramids. The double RS layers are similar
to those described for all of the above thallium cuprates, with
one median TlO layer sandwiched between two BaO layers.
Like the other members of the series, this 120-K superconductor
shows small deviations from the oxygen stoichiometry,
together with a possible excess of Tl on the calcium site,
according to the formula Tl1xCa2xBa2Cu3O9d, with x  0.1
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
b fluorite cages formed by the copper layers. The ordered sequence
a of the cations along c is, as expected, TlBaCaCaCaCa
c
Figure 10 View of ! the tetragonal 1212 structure of the cuprate
TlSr2CaCu2O7 along a showing double layers of CuO5 pyramids
interleaved with Ca2 cations (green sphere) and double rock salt
layers built up of one TlO layer (Tl pink spheres, O red spheres)
sandwiched between two SrO layers (Sr yellow spheres). Note
the splitting of the Tl sites.
and d  0.10.4.112114 Again, in that structure, the Ba2, Ca2,
!
and Tl3 cations are ordered in the form of layers along c
according to the sequence TlBaCaCaBaTlBaCaCa factor of 20%.
BaTl, calcium sitting in the fluorite cages formed by the
CuO4 square planar groups and the CuO5 pyramids within
the ordered oxygen-deficient perovskite layers. The other
n-superior members are more difficult to synthesize as pure
phases. The 1234 110-K superconductor, with the com-
position TlBa2Ca3Cu4O11,115,116 is tetragonal, with
a  ap and c 19 A. It can be deduced from the 1223 n 3
member by inserting one additional CaCuO2 layer into the
ordered oxygen-deficient perovskite layer. It results in quadru-
ple copper layers built up of two layers of CuO4 square planar
planes sandwiched between two copper pyramidal layers
(Figure 11(b)). As a consequence, the cations of the RS layers
!
and of the fluorite cages are ordered along c according to
the sequence TlBaCaCaCaBaTlBaCaCaCaBaTl. a  ap and c  23 A, and for Tl2Sr2CuO6,127 which exhibits similar
The 1245 105-K superconductor, with the composition
TlBa2Ca4Cu5O13,117,118 is tetragonal, with a  ap and c 22 A.
In this structure, the quintuple copper layer consists of three
layers of CuO4 square planar groups sandwiched between two
pyramidal copper layers and all of the calcium cations sit in the
71
!
BaTlBaCaCaCaCaBaTl.
Double RS-type layers may coexist with simple RS layers in
the same structure leading to the formation of multiple inter-
growths. This complex ordering is rather rare: it has been
observed for one structural family, corresponding to the generic
formulation TlA4M2O9 with M Co, Fe and A Sr, La and called
the 1201-0201 intergrowth. The ferrite TlSr4Fe2O9119,120 and the
cobaltite TlSr3LaCo2O9121 are the most representative of these
double intergrowths. Their tetragonal structure (Figure 12) con-
sists of the stacking of the two types of RS layers single SrO-type
and double TlSrO2-type with single perovskite layers, leading
to the cell parameters a ap and c  30.5 A. In the ferrite
TlSr4Fe2O9 (Figure 12(a)) and in the cobaltite TlSr3LaCo2O9
(Figure 12(b)), the double RS [SrTlO2]/ layers exhibit an iden-
!
tical stacking sequence along c with respect to the [MO2]/ layers
formed by the basal plane of the MO6 octahedra (M Fe, Co):
MO2SrOTlOSrOMO2SrOSrOMO2. In the first series of
ferrites, the Tl and O sites are split with respect to the ideal
positions (Figure 12(a)), allowing correct TlO distances to be
obtained. In the second series of cobaltites (Figure 12(b)), a
deviation from the thallium and oxygen stoichiometry is
observed, according to the formula Tl1xSr3LaCo2O9d, with
x  0.1 and d  0.20.4. The oxygen vacancies are located in the
TlO layers (10% vacancies) and at the apical site of the octahe-
dra (20% vacancies). The possibility of substituting lead for
thallium in the ferrite has been observed in the whole homoge-
neity range of the solid solution Pb1xTlxSr4Fe2O9, with
0  x  1.119 The structure of the limit compound,122124 which
exhibits a lead deficiency according to the formula Pb1xSr4Fe2O9
(x  0.10.2) (Figure 12(c)), is essentially the same as that of
TlSr4Fe2O9, concerning the stacking sequence of the RS-like layers
and copperoxygen layers. The only difference deals with the
coordination of iron, which is better described as FeO5 pyramids
than FeO6 octahedra, due to the fact that one apical oxygen of the
latter polyhedra belonging to the PbO layers has moved to a
!
larger distance along c (2.6 A). In this structure, the oxygen sites
of the SrO layers are split into triple rows, and the Pb sites are
also split with respect to their ideal positions, with an occupancy
2.03.4.2 Intergrowths Involving Triple RS Layers,
[RS]3-[Pe]n
2.03.4.2.1 The n 1 members
The first n 1 member of this family, called 2201, is represented
by the ideal formula Tl2A2MO6 or Bi2A2MO6 with A Ba, Sr, and
M Cu, Mn, Co, Fe. In fact, the crystal chemistry of these oxides is
strongly dependent on the nature of the cation forming the triple
RS layer, Tl3 or Bi3, especially concerning the oxygen non-
stoichiometry and the symmetry and commensurability or
incommensurability of the structure. The most simple structure
was first observed for Tl2Ba2CuO6,125,126 with a tetragonal cell,
parameters. This tetragonal structure (Figure 13) consists of single
perovskite layers of corner-shared CuO6 octahedra stacked with
triple RS-type layers built up of double TlO layers sandwiched
between single BaO or SrO layers. In fact, detailed structural
studies showed that these tetragonal phases can be thallium
72 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

b
b a

c c

(a)
(b)

Figure 11 View of the tetragonal structures of (a) the 1223 TlBa2Ca2Cu3O9 cuprate and (b) the 1234 TlBa2Ca3Cu4O11 cuprate showing triple and
quadruple copper layers, respectively, intergrown with double rock salt layers. Ba, Ca, Tl, and O and correspond to blue, green, pink, and red spheres,
respectively.

deficient, with a copper excess sitting in the RS layers, according to
the formula Tl2xCu1xBa2CuO6.128,129pMoreover,
pa second form
with an orthorhombic symmetry ap 2  ap 2  23 was
also observed and was shown to be Tl deficient and containing
a slight excess of oxygen within the RS-type layers, according to
the formula Tl1xBa2CuO6d, with x ranging from 0 to 0.18
and 0d0:30,130132 whereas excess oxygen was also shown
to exist in the tetragonal form of Tl2Ba2CuO6d, with
0d0:4.133,134 This oxygen nonstoichiometry, although it
appears to be small, has a great impact upon the superconducting
properties of this oxide, since it was shown that the critical
temperature of this phase could be increased dramatically from
81 K up to 98 K by annealing the material at low temperature in a
reducing atmosphere.135 In contrast to copper, no n 1 member
containing thallium was observed for manganites, cobaltites,
and ferrites. The n 1 member of the 2201 bismuth homolo-
gous series is observed for the four transition elements M Mn,
Fe, Co, and Cu. However, the oxygen stoichiometry is not the
expected ideal O6 composition, but contains generally excess
oxygen. This is the case of the 2201 manganite, Bi2Sr2MnO6.5
(Bi8Sr8Mn4O26),136,137 for which the orthorhombic struc-
ture was solved in different space groups (Imma, Pnma, or
Pn21a), pwith
a superstructure
p along one direction, that is,
a 4ap 2, b ap 2, and c 23:6 . In such a structure
(Figure 14), the triple RS-like layers built up of double bismuth
oxygen layers sandwiched between two SrO layers are waving
!
along a , containing an excess oxygen with respect to the Bi2O2
composition. This excess oxygen, together with the 6s2 lone pair
of Bi3, are at the origin of this waving and this implies a
quadrupling of the a parameter. As a consequence, the single
perovskite layers of MnO6 polyhedra are also waving along this
direction. Moreover, a displacement of the apical oxygen of
several MnO6 octahedra is induced due to this waving, so that
the structure of the perovskite layers can better be described as
formed of MnO6 octahedra and MnO5 pyramids, with the alter-
nation either of two octahedra with two pyramids (Figure 14(a))
!
or of three octahedra with one pyramid (Figure 14(b)) along a .
The substitution of lead for bismuth in this manganite allows
the oxide BiPbSr2MnO6 to be synthesized.138,139 Contrary to
the pure bismuth manganite, this phase does not show any
!
superstructure
p p a , that is, orthorhombic (Amaa), with
along
a  ap 2, b ap 2, and c 24 . In fact, the latter structure
is similar to that observed for the orthorhombic form of Tl2Ba2
CuO6, with a splitting of the bismuth and oxygen positions in the
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
a
b a
c
c
(a)
(b)
b
a structure is also observed for the cobalt oxides. In this system,
c oxygen-deficient cobaltite BiSr3CoO6d with d  0.20143 and
(c)
Figure 12 View of!the tetragonal structures of the double 12010201
intergrowths along a of (a) TlSr4Fe2O9, (b) TlSr3LaCo2O9, and (c)
Pb1xSr4Fe2O9. Tl or Pb, Sr, or La, and O are represented by pink,
yellow, and red spheres, respectively. Note the splitting of the Tl and O
sites in the RS layers for TlSr4Fe2O9, and of the Pb and O sites in
Pb1xSr4Fe2O9, in contrast to TlSr3LaCo2O9.
Bi2O2 layers (Figure 15). It is also observed for the ferrite
BiPbSr2xFeO6d, with 0 < x < 0.1 and 0 < d < 0.30,140142
which also shows additional oxygen in the bismuth-oxygen
layers with respect to the ideal O6 stoichiometry. The 2201
73
b a
c
Figure
!
13 View of the 2201 tetragonal structure of Tl2Ba2CuO6 along
a . Tl, Ba, and O are represented by pink, blue, and red spheres,
respectively. Note that the oxygen positions in the TlO layers are split.
two classes of 2201 compounds can be distinguished, the
the cobaltite Bi2Sr2 CoO6d with d  0.25.144,145 The oxygen-
deficient
pperovskite
pBiSr3CoO5.8 exhibits a monoclinic cell with
a  ap 2, b  ap 2, c  23 , and g 90 :5, showing
numerous distortions of the CoO6 octahedra, which are
!
strongly elongated along the c direction, so that the structure
can better be described as CoO5 pyramids whose apical corners
!
are directed along two opposite directions along c alternately
(Figure 16). The single perovskite layers and triple RS layers
!
are stacked along c according to the sequence: CoO2SrO
(BiSr)O(Bi, Sr)OSrOCoO2. The structure of the over-
oxygenated 2201 cobaltite Bi2Sr2CoO6.25 (or Bi8Sr8Co4O25)
(Figure 17) is closely related to that of Bi2Sr2MnO6.25. The
double BiO layers are sandwiched between SrO layers and
wave in the structure similar to the manganite, whereas the
single perovskite layers have their octahedra extremely elon-
!
gated along c so that they are better described as CoO5 pyra-
mids, forming quadruple ribbons, with apical corners pointed
along opposite directions alternately. This phase exhibits simi-
p
lar parameters
p to the manganese homologous, a 4ap 2,
b  ap 2, and c 24 , but a different space group I2mb. In
fact, the Bi2Sr2CoO6d cobaltites exhibit both commensurate
and incommensurate structures due to the presence of the 6s2
lone pair of Bi3 and also due to oxygen excess. Moreover, the
incommensurability can be modified by partly substituting
lead for bismuth. For instance, in Bi2Sr2CoO6d, an incommen-
surate structural modulation with the modulation vector
q 0.24b* c* is observed,145,146 and the polar group
74 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

b
a

(a)
b
a

(b)
!
Figure 14 The 2201 structure of Bi2Sr2MnO6.5 view along b . The single perovskite layers can be described either as the alternation of two MnO6
octahedra with two MnO5 pyramids (a), or of three octahedra with one pyramid (b). Note that, in the second description, the apical apex of the pyramid is
split over four positions. The Bi, Sr, and O atoms are represented by blue, yellow, and red spheres, respectively.

refinement leads to the superspace group F2mm (0b1)00s. In
this structural type, the modulation of the structure is mainly
related to the configuration of the Bi and O arrangements
(Figure 18). There are two regions, modulated (M zone) and
nonmodulated (I zone), in the BiO layers. The origin of the
modulation phenomena lies in the interactions within and
between the BiO layers. The oxygen disorder observed in the
M region leads to different possible configurations for the BiO
layers. Of the two possible configurations, honeycomb and
double chain configurations of the Bi and O atoms
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 75

a a
b b

c c

!
Figure 15 The 2201 structure of BiPbSr2MnO6 (view along a ) can be
described, like Tl2Ba2CuO6 as single octahedral perovskite layers
intergrown with triple BiPbSrO3 layers. The Bi/Pb, Sr, and O atoms are
represented by blue, yellow, and red spheres, respectively.
!
Figure 16 The 2201 structure of BiSr3CoO6d: view along a . The
single perovskite layers are better described on the basis of CoO5
(Figure 18), the latter is the preferred configuration.145 pyramids. The Bi, Sr, and O atoms of the triple rock salt layers are
The substitution of Pb for Bi, according to the formula represented by blue, yellow, and red spheres, respectively.
Bi2xPbxSr2CoO6, leads to significant change in the structure.147
Although for x  0.16, one observes structural modulations with
this composition have been sometimes observed as shown for
the vector q 0.245b* c* and the space group F2mm, the
Tl1.8Ca0.9Ba2Cu2O7.9,160 Tl2Ca0.9Ba1.96Cu2O8.17,161 Tl2.16Ca0.72
highly substituted phase, x 0.8, is no longer modulated. The
Ba2Cu2O8,162 and Tl2Ba2CaCu2O8x.163 The structure of these
modulation is now replaced by a new disordered atomic
oxides, called 2212 (Figure 19), is characterized by a tetragonal
sequence that can be described as the simultaneous presence of
cell with a ap and c 29 A. Double-ordered oxygen-deficient
two structural configurations compatible with symmetry group
perovskite layers, containing calcium ions, similar to those
Pnan.147 The nonmodulated structure is the average structure of
observed for the 1212 thallium cuprates, are stacked with triple
the modulated oxide (F2mm). Interestingly, the honeycomb !
RS layers along c . The latter are built up of double TlO layers
arrangement of the BiO layer is observed in this highly
sandwiched between single RS layers, like in the 2201 prototype,
substituted phase, which was considered as a possible hypothe- !
Tl2Ba2CuO6. Thus, the cations are ordered along c in the form
sis in the pristine phase, containing only bismuth. The structure
of layers according to the sequence CaCuO2BaOTlOTlO
of the 22-K superconductor, Bi2Sr2CuO6d,148 is very similar
BaOCuO2CaCuO2BaOTlOTlOBaOCuO2Ca. The
to that of the cobalt and manganese phases, withpthe same p deviation from the stoichiometry results from vacancies on
orthorhombic symmetry of the subcell ap 2  ap 2
the oxygen sites of the RS layers, on the thallium sites, or from
24:4 ,148151 and apmodulated
psuperstructure, corresponding the partial occupancy of Ca sites by Tl.
to the parameters ap 2  5ap 2  24:4 .148,152,153 Its struc-
The n 2 members containing double BiO layers are syn-
ture is characterized similarly to the cobaltites by a waving of
thesized for ferrites and cuprates. The 2212 bismuth ferrites are
the BiO layers and involves either an excess or a deficiency
represented by the generic formula Bi2xSr3xFe2O9d.164166
of oxygen, depending on the synthesis conditions.
This
p structure
p is modulated with an orthorhombic
ap 2  ap 2  31:7 subcell and a modulation vector,
2.03.4.2.2 The n 2 members whichpallowspalso
the structure to be solved in a multiple cell
The number of n 2 members containing double TlO layers is 5ap 2  ap 2  31:7 . The waving of the BiO layers in the
quite limited. The cuprate Tl2xBa2CaCu2O8 is the only repre- structure changes with d values ranging from 0.50 to 2, and the
sentative of this family with 0  x  0.2.154159 Deviations from coordination of iron can also vary involving either only FeO6
76 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

b octahedra (Figure 20(a)) or FeO5 pyramids and FeO6 octahedra

a (Figure 20(b)). In contrast to ferrites, the bismuth cuprates
exhibit ordered oxygen-deficient double perovskite layers simi-
c lar to those observed for the 2212 Tl-cuprates. Their ideal
formula, Bi2Sr2CaCu2O8, is principally modified by the possi-
bility of an oxygen excess in the triple RS-like layers, but
cationic deficiency can also appear in those layers. Thus, this
structure has been described by many pauthors on the basis of an
orthorhombic subcell with a  ap 2, b  ap , and c  30:7
for various compositions ranging from Bi2Sr2CaCu2O8d167170
to Bi2xSr1yCa1zCu2O8d171173 or to Bi2x(Sr,Ca)3yCu2
O8d,174178 and finally to lead-substituted compounds
Bi2xPbx(Sr,Ca)3Cu2O8d.179,180 Thus, the structural principle
of this 2212 cuprate is the same as that described for the Tl-
phase (Figure 19), except that the structure is incommensurate,
!
that is, incommensurate satellites are observed along a in the
169,181
electron diffraction patterns. As a consequence, differently
from the Tl-2212 homologous, the BiO layers that are waving
in the structure can accommodate excess oxygen, which is dis-
tributed in an incommensurate way. Such a property, most
probably connected to the 6s2 lone pair stereoactivity of Bi3,
is of capital importance because it governs the superconducting
properties of this material, with optimal Tc around 90 K that
can be tuned by oxygen over or under doping.

2.03.4.2.3 The n 3 and n 4 members

Figure 17 !The 2201 structure of Bi8Sr8Co4O25 (Bi2Sr2CoO6.25):
view along b . Note that the CoO5 pyramids form quadruple ribbons
with apical corners directed in opposite directions alternately. The Bi, Sr,
and O atoms of the triple rock salt layers are represented by blue, yellow,
and red spheres, respectively.
The n 3 and n 4 members, involving triple RS-like layers,
are limited to the cuprates and ferrites. In the case of ferrites,
the stabilization of these members requires the presence of
bismuth oxygen layers partly substituted by lead. The ferrite
Bi3.5Pb0.5Sr2Fe3O12182 represents the n 3 member of this
series. This 2223 tetragonal structure (Figure 21(a)), with
a  ap and c 40 A, can be described as built up of triple octa-
hedral perovskite layers intergrown with triple distorted RS
layers. The RS-type layers consist of double (Bi3/4Pb1/4)O
layers sandwiched between two (Sr2/3Bi1/3)O layers. Note the
ordered distribution of the Bi3 and Pb2 cations with respect
to the Sr2 cations in the triple RS layers. Similarly, the perov-
skite cages are preferentially occupied by Bi3 (Bi2/3Sr1/3). The

(a)

(b)
Lone pair Bi OB1 OB2

Figure 18 Two possible configurations of the BiO layer in Bi2Sr2CoO6d: (a) honeycomb and (b) double-chain. OB1a and OB1b sites are alternately
occupied. BiO interatomic distances <2.6 A are represented by solid lines.
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
b Ba2Ca2Cu3O10d, x and d indicating possible deviations from
a the stoichiometry due to the conditions of synthesis.184188 This
c
Figure!19 View of the tetragonal 2212 structure of Tl2Ba2CaCu2O8
along a . The Ba, Tl, Ca, and O atoms are represented by blue, pink,
yellow, and red spheres, respectively. Note that the O positions of the
TlO layers are split.
n 4 member of the ferrite series corresponds to the formula
Bi4.15Pb0.45Sr2Fe4O15.183 This 2234 tetragonal structure
(Figure 21(b)) consists of quadruple octahedral perovskite
layers intergrown with triple distorted RS layers, leading to the
cell parameters, a  ap and c 48 A. Here again, the distribution
of the three sorts of cations, Bi3, Pb2, and Sr2, is semi-
ordered (Figure 21(b)). Nevertheless, this semi-ordering is
significantly different from that observed in the n 3 member,
although the composition of the double (Bi2/3Pb1/3)O layers is
close to that observed for the latter.
The cuprates, in contrast to the ferrites, are synthesized for
both types of cations in the triple RS layers, bismuth and
thallium. The important difference of copper with respect to
iron deals with its JahnTeller effect, which makes it that layers
of corner-shared CuO4 square planar groups can be formed
instead of layers of FeO6 octahedra, significantly decreasing the
thickness of the oxygen-ordered triple perovskite layers. As a
consequence, the c parameter of the crystallographic cell of the
cuprates is significantly reduced with respect to the ferrites.
Moreover, concomitantly with the ordering of the anionic
vacancies, an ordering of the alkaline earth cations in the struc-
ture appears in the form of layers. The n 3 member of the
cuprate series, called 2223, is represented by the formula Tl2x
77
125-K superconductor, characterized by a tetragonal cell with
a ap and c 36 A, has its structure (Figure 22) directly derived
from the ferrite by elimination of the apical oxygen atoms of the
octahedra, located inside the triple perovskite layers (Figure 21
(a)). The result is that each triple-ordered oxygen-deficient
perovskite layer consists of one layer of CuO4 square planar
groups sandwiched between two layers of CuO5 pyramids.
Like in the 2201 and 2212 thallium cuprates, the triple RS
layers consist of double TlO planes sandwiched between two
BaO planes. In the same way, vacancies can appear on the Tl
sites, and Ca sites can also be partly occupied by Tl, whereas
oxygen excess can be introduced into the structure. The ordering
!
of the cationic layers stacking along c obeys the sequence Ba
TlTlBaCaCaBaTlTlBaCaCaBa. Similarly to thal-
lium, bismuth allows the n 3 member to be generated, but it
requires the presence of strontium instead of barium, and a
partial substitution of lead for bismuth is often necessary to
stabilize the structure of this 110-K superconductor, according
to the ideal composition Bi2xPbxSr2Ca2Cu3O10.189192 The
structure of this phase, described either in a tetragonal cell
with a  ap p c  37p
and A, or in an orthorhombic symmetry
with a  ap 2, b  ap 2, and c  37 , is close to that of
the thallium 2223 phase (Figure 22(a)), with again a possible
modulation of the structure along one direction involving a
waving of the Bi2O2 layers and a possible oxygen overstoichio-
metry. The n 4 member of the cuprate series was only synthe-
sized in the case of thallium. This 2234 phase, with the ideal
composition Tl2Ba2Ca3Cu4O12, is difficult to synthesize as a
pure phase.193,194 The tetragonal structure (Figure 22(b)) of
this 104-K superconductor with ap  42 A, consists of double
layers of CuO4 groups sandwiched between two layers of CuO5
pyramids, intergrown with triple RS-type layers. No n 4 mem-
ber was isolated for bismuth cuprates.
2.03.4.3 Intergrowths Involving Larger RS Layers
Intergrowths involving larger RS layers are extremely rare and
difficult to synthesize. The ferrite Bi2Sr4Fe2O10195 seems to be
the only example of structure (Figure 23) that consists of
quadruple RS-type layers intergrown with double octahedral
perovskite layers. In this tetragonal structure, with a  ap and
c  18.5 A, triple RS-type layers containing bismuth and stron-
tium are sandwiched between two pure SrO layers, whereas the
cages in the double perovskite layers are mainly occupied by
strontium (80%).
2.03.5 Creation of Ordered Oxygen Vacancies in
Perovskite and RS Layers: The Lead and Mercury RP
Derivatives
The possibility of replacing either single perovskite layers or
medium AO planes in double RS layers by single layers of
cations can be favored by introducing cations with linear
dumbbell coordination, such as univalent copper or divalent
mercury. Such a behavior is observed in lead cuprates and in
mercury cuprates.
78 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
c
a
b a
b
c
(a)
(b)
Figure 20 Structures of the 2212 bismuth ferrites Bi2xSr3xFe2O9d built up of (a) double octahedral layers in Bi2Sr2.6Fe2O11 or (b) of octahedra and
pyramids in Bi10Sr15Fe10O46. The Bi, Sr, and O atoms of the triple RS layers are represented by blue, yellow, and red spheres, respectively.
2.03.5.1 The Lead Cuprates
The most simple compound deals with the 30-K superconductor
Pb2Sr2xLaxCu2O6d,196,197 which is characterized by an aver-
age oxidation state of copper smaller than 2. This compound
crystallizes
p eitherpin the orthorhombic symmetry, with
a  ap 2, b  ap 2, and c 12:6 or in the tetragonal sym-
metry with a  ap and c 12.6A. Its structure (Figure 24) can be
described as the intergrowth of single RS layers (Sr, La, Pb)O
with single layers of CuIO2 dumbbells and of CuO6 octahedra.
In fact, this structure can be deduced from the 0201 structure of
La2CuO4 (Figure 1(a)) by removing rows of oxygen atoms in
one perovskite layer out of two, in an ordered way. This leads to
two sorts of copper layers, that is, non oxygen-deficient perov-
skite layers, like in La2CuO4, and ordered oxygen-
deficient layers of CuO2 dumbbells (Figure 24). Such a
distribution of the oxygen vacancies implies a copper dispropor-
tionation, with univalent copper in the CuO2 dumbbells and
mixed valent copper Cu(II)Cu(III) in the superconductive
octahedral layers. It is quite remarkable that the Pb2 cations
and the La2/Sr2 cations are ordered, forming (Sr, La)O and
PbO layers that are stacked according to the sequence (Sr, La)
!
OPbOPbO(Sr, La)O(Sr, La)O along c . In fact, one
observes that the PbO layers border the oxygen-deficient
[CuIO2]/ layers. This suggests that the 6s2 lone pairs of Pb2
are directed toward the anionic vacancies of these layers, thus
stabilizing such an oxygen-deficient structure. Thus, this struc-
ture can be considered as an ordered 0201 oxygen-deficient
ordered phase.
The 68-K superconductor Pb2Sr2Ln1xCaxCu3O8 with p
Ln La,p Y
198,199
is also orthorhombic with a  ap 2,
b  ap 2, and c  15:9 . Its structure (Figure 25) consists
of double pyramidal copper layers interleaved with Y3/Ca2
cations, similar to those observed in the 0212 structure of
La2CaCu2O6. These ordered oxygen-deficient perovskite layers
are also stacked with single (Sr, Pb)O RS layers similar to the
LaO layers of the 0212 structure. However, they differ from
this structure by the fact that the single RS layers are connected
through layers of Cu(I) in linear twofold coordination. In fact,
this structure is directly derived from a double intergrowth
structure with the hypothetical formula Pb2Sr2YCu3O10
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 79

b (Figure 25), which consists of an intergrowth of the 0201,

a La2CuO4-type structure, with the 0212, La2CaCu2O6-type
structure. This again implies a disproportionation of Cu(II)
into Cu(I) and Cu(III), with an ordering of these species, that
b c is, Cu(I) in dumbbell coordination, and Cu(II)Cu(III) mixed
a
valence in the pyramidal layers. This disproportionation is
crucial for tuning the superconducting properties of this mate-
rial. As a consequence, the Tc of such a phase could indeed be
c
increased up to 80 K, by bismuth substitution in the isotypic
cuprate Pb2xBixSr2Y1yCayCu3O8.200
The cuprate PbBaYSrCu3O8d201,202 exhibits a tetragonal
structure (Figure 26), with a  ap and c 27.45 A. The latter
compound, called 0223, consists of triple-ordered oxygen-
deficient perovskite layers intergrown with single RS-type
layers. The oxygen-deficient triple perovskite layers are them-
selves built up of two pyramidal copper layers interleaved with
Ln3 (Y3 or La3) cations and connected to a layer of CuIO2
dumbbells. Each single RS layer consists of one (Ba, Sr)O
plane shared with the pyramidal copper layers and one PbO
plane. Most probably, the 6s2 lone pair of Pb2 is directed
toward the Cu (I) layer, tending to sit close to the oxygen
vacancy in this oxygen-deficient layer. Thus, this structure is
directly derived from the hypothetical 0223 intergrowth with
the ideal formula (Ba, Sr)2PbYCu3O10, which would corre-
spond to an intergrowth of triple perovskite layers with single
RS layers, by an appropriate ordering of the Pb2, Ba2/Sr2,
!
and Y3 cations along c and a concomitant ordering of the
oxygen vacancies in the perovskite layers. Here, again, the role
of the lead lone pair is important to stabilize the oxygen
deficiency and ordering, and has an impact upon the dis-
proportion of divalent copper into trivalent and univalent
copper, allowing the superconductivity to be generated in the
pyramidal copper layers.201,203

2.03.5.2 Mercury Cuprates

(a)
(b)
Figure 21 Tetragonal structures of (a) 2223 Bi3.5Pb0.5Sr2Fe3O12,
where Bi/Pb, Bi/Sr, and O atoms are represented by blue, yellow, red
spheres, respectively; and (b) 2234 Bi4.15Pb0.45Sr2Fe4O15, where Bi/Pb,
Bi, Sr/Bi, Sr, and O atoms are represented by blue, green, pink, yellow,
and red spheres, respectively.
The structures of the mercury cuprates are directly derived from
those of the thallium cuprates containing single TlO layers by
replacing the TlO layers by Hg layers. This is possible because
of the ability of mercury to adopt the linear two fold coordi-
nation HgO2. Thus the 96-K superconductor HgBa2CuO4204
exhibits a tetragonal cell with a  ap and c  9.5 A, characteristic
of the 1201 structure (Figure 27(a)). It consists of single
octahedral perovskite layers interconnected through single
layers of HgO2 dumbbells, and the structure is susceptible to
accommodate additional oxygen at the level of the mercury
layers, leading to the formula HgBa2CuO4d.205,206 An excess
of copper sitting on the mercury sites is also observed.
HgBa2CaCu2O6d represents the second member of the series.
This 1212 tetragonal structure (Figure 27(b)), with a  ap and
c  12.7 A, consists of double pyramidal copper layers inter-
connected through single layers of HgO2 dumbbells accepting
an excess oxygen up to d 0.4 at the level of the Hg layers.207
209
The 1223 compound Hg1xBa2Ca2Cu3O8d, which
exhibits the highest Tc of 133 K, is also tetragonal, with a ap
and c 15. 8 A. Its structure (Figure 27(c)) is similar to the
homologous Tl phase, involving single layers of CuO4 groups
80 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

b b
a a

c c

(a) (b)
!
Figure 22 View of the tetragonal structures of (a) 2223 Tl2Ba2Ca2Cu3O10 and (b) 2234 Tl2Ba2Ca3Cu4O12 along a . The Ba, Tl, Ca, and O atoms are
represented by blue, pink, green, and red spheres, respectively. Note the splitting of the O positions in the TlO layers in the 2223 cuprate.

sandwiched between two layers of CuO5 pyramids,210214 and Hg1xBa2Ca3Cu4O10d (Figure 27(d)), containing quadruple
can accept rather large deviations from the stoichiometry, that copper layers; Hg1xBa2Ca4Cu5O12d, containing quintuple
is, Hg deficiency (x up to 0.3) and oxygen excess (d up to 0.4). copper layers; and Hg1xBa2Ca5Cu6O14d, containing sextuple
The superior members, such as the 1234 cuprate, copper layers, can also be obtained, all with the same
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 81

b a b
a

c c

Figure 24 Structure of Pb2SrLaCu2O6 built up of single octahedral

perovskite layers intergrown with single (Sr,Pb)O RS layers and single
layers of CuIO2 sticks. The Sr, Pb, Cu, and O atoms are represented by
yellow, green, turquoise, and red spheres, respectively.

Figure!23 View of the tetragonal structure of the ferrite Bi2Sr4Fe2O10
along a , containing quadruple rock salt layers. The Bi/Sr, Sr,
and O atoms are represented by blue, yellow, and red spheres,
respectively.
tetragonal symmetry and with the same a ap parameter and c
values of 19, 22, and 25 A, respectively.215,216
This structural family is remarkable in its ability to accept
additional oxygen in the mercury layers, so that it can be stabi-
lized by the substitution of other cations, such as bismuth, lead,
chromium, molybdenum, tungsten, niobium, manganese, and
ruthenium for mercury,217,218 as summarized in Table 4. The
homologous series of mercurystrontium-based cuprates does
not exist. However Sr-based 1201 and 1212 mercury cuprates
can be stabilized by a partial substitution of various cations for
mercury: Bi, Pb, Pr, Mo, Cr, Tl, Ce, and Re.217,219 Many of these
oxides are superconductors, especially those containing lantha-
nides (La, Nd, Pr), with Tc ranging from 27 to 100 K. This is
exemplified by the 1201 Hg0.4Pr0.6Sr2CuO4d and the 1212
Hg0.4Pr0.6Sr2Ca1xPrxCu2O6d phases,220 which show a partic-
ular ordering of the mercury and praseodymium cations on the
mercury sites. In those structures (Figures 28 and 29), the Pr3
and Hg2 cations form rows, where one Pr3 alternates with one
Hg2 cation. It results in the structure of the 1201 phase
(Figure 28), characterized by a doubling of one cell parameter,
that is, an orthorhombic symmetry with a  2ap, b  ap and
c  c1201. This ordering of the mercury and praseodymium in
such rows is also observed for the 1212 structure (Figure 29),
but there exist two polymorphs, depending on the relative
!
positions of these rows along c : they correspond to the super-
cells ap  2ap  c1212 (Figure 29(a)) and ap  2ap  2c1212
(Figure 29(b)).
82 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

a b
b a

c
c

Figure 25 Structure of the cuprate Pb2Sr2LaCu3O8 built up of double

pyramidal copper layers intergrown with single (Sr,Pb)O RS layers and
single layers of CuIO2 sticks. The Pb, Sr, Cu(I), and O atoms are
represented by blue, yellow, turquoise, and red spheres, respectively.
Note that the O of the PbO layers are split, making difficult the
representation of the linear OCuO bonds.

Figure!26 View of the 0223 tetragonal structure of PbBaYSrCu3O8

The simultaneous introduction of LaO and Hg layers in
the same structure is also possible, leading to a complex
intergrowth, HgBa2La2Cu2O8.221 This structure (Figure 30)
consists of an intergrowth of the 1201 structure of
HgBa2CuO4 and of the 0201 structure of La2CuO4, previously
described. Another complex intergrowth was also synthesized
by association of thallium to mercury. This is exemplified
by the 50-K superconductor, HgTl2Ba4Cu2O10d,222 whose
structure (Figure 31) consists of an intergrowth of the
2201 structure of Tl2Ba2CuO6 and of the 1201 structure
of HgBa2CuO4.
along a showing triple copper layers intergrown with single (Pb,Ba)O RS
layers. The Ba/Sr, Pb, Y, Cu, and O atoms are represented by yellow, blue,
green, turquoise, and red spheres, respectively. Note that the O positions
of the PbO layers are split.
In contrast to thallium, mercury, due to its two fold coordi-
nation, should not be able to exhibit double mercury layers.
Nevertheless, one cuprate involving double Hg layers has been
synthesized. The oxide Hg1.5(Cu,Pr)0.5Ba2PrCu2O8d223 has a
structure (Figure 32) closely related to that of the 2212 structure
of Tl2Ba2CaCu2O8d. It derives from the latter by substituting Hg
for Tl and Pr for Ca in this cuprate.
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 83

2.03.6 Insertion of Additional Fluorite-Type Layers
between Pyramidal Copper Layers in Cuprates
Concerning the cubic coordination of calcium or yttrium,
located between the double pyramidal copper layers of the
1212 and 2212 cuprates, it appears that this layer of edge- 1222 structure of the cuprate TlBa2Eu1.5Ce0.5Cu2O9
sharing CaO8 cubes (Figure 33(a)) can be described as a single
fluorite-type layer. This suggests the possibility of replacing such
a layer by double or triple fluorite-type layers (Figure 33(b) and
33(c)). In order to generate such cuprates, the introduction of a
lanthanide with a higher valence than 3 is necessary. Ce4 is a
good candidate, bearing in mind that it has a great ability to form
the fluorite CeO2-type structure. Three classes of cuprates, with
double fluorite-type layers, directly derived from the single inter-
growths described above, can be distinguished.
The first class is represented by the cuprate Nd2.64Sr0.82Ce0.54
Cu2O8d.224226 This oxide is often called T* phase, due to the
parameters of its tetragonal cell, a ap and c 12.5 A similar to ple, in the bismuth cuprates Bi2Sr2xLn2yCeyCu2O10 with
those of La2CuO4. Its structure (Figure 34) is directly derived from
that of the 0212 phase, La2CaCu2O6, by replacing the single
layers of calcium fluorite type with double layers of fluorite type
containing Nd3/Ce4 cations. As a consequence, it can be
described as the representative member of the 0222 family,
which consists of double pyramidal copper layers interleaved
with double (Nd, Ce)2O4 fluorite-type layers and intergrown
with single (Sr,Nd)O RS layers. Thus, the cations are ordered in
!
the form of layers perpendicular to c according to the sequence
(Sr, Nd)OCuO2(Nd, Ce)O2(Nd, Ce)CuO2(Sr, Nd)O. The cuprates involving triple fluorite-type layers are rarely
The second class can be described as the 1222 family,
which is derived from a 1212 structure, such as TlBa2Ca
Cu2O7, by replacing the single calcium layers by double
fluorite-type layers, containing Ln3 and Ce4 cations. This is
the case of the cuprates Tl1xA2y(Ln,Ce)2Cu2O9, with A Sr,
Ba and Ln Pr, or Nd or Eu,227,228 of the oxides (Pb0.5Sr0.5)
Sr2(Nd2xCex)Cu2O9,229 and of the cuprates Pb0.5A0.5
(Sr2xEux)(Eu2xCex)Cu2O9.230,231 All of these cuprates are
tetragonal, with a  ap and c  30 A. As exemplified from the
(Figure 35), double RS layers built up of single TlO layers
sandwiched between two BaO layers are stacked with double
fluorite layers Eu1.5Ce0.5O4 sandwiched between two pyra-
!
midal layers along c , leading to the stacking sequence TlO
BaOCuO2(Eu, Ce)1O2(Eu, Ce)CuO2BaOTlOBaO
CuO2(Nd, Ce)O2(Nd, Ce)CuO2BaOTlO.
The third class, called 2222, derives from the 2112 family
of cuprates such as Tl2Ba2CaCu2O8 or Bi2Sr2CaCu2O8, by
replacing the single calcium layers by double (Ln, Ce)2O4
fluorite layers. Thesepcompounds
pexhibit either an orthorhom-
bic cell, with a  ap 2, b  ap 2 and c  18 , as, for exam-
Ln Nd, Eu,232 or a tetragonal cell with a  ap and c  18 A, as,
for example, in the oxides Tl2Ba2Ln2xCexCu2O10.233 Their struc-
ture (Figure 36) consists of double Bi2O2 or double Tl2O2
layers sandwiched between single BaO or SrO layers
forming triple RS layers. The latter layers are stacked with double
fluorite-type layers Ln2xCexO4 sandwiched between two
pyramidal copper layers leading to the following ordered
!
sequence of the layers along c : CuO2BaOTlOTlOBaO
CuO2(Nd, Ce)O2(Nd, Ce)CuO2.
observed. One example is shown by the series (Bi0.5Cu0.5)
Sr2(Ln,Ce)3Cu2O10, with Ln Nd, Gd.234 These oxides exhibit
a tetragonal cell, close to that observed for the 2222 Tl-
cuprates, that is, a  ap and c  17 A. However, the structure of

b b
a a

c
c

(a) (b)

Figure 27 (Continued)
84 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

b
b a
a

c
c

(c)
(d)
Figure 27 View of the!tetragonal structures of (a) 1201 HgBa2CuO4, (b) 1212 HgBa2CaCu2O6, (c) 1223 HgBa2Ca2Cu3O8, and (d) 1234
HgBa2Ca3Cu4O10 along a . The Ba, Ca, and Hg atoms are represented by yellow, pink, and green spheres, respectively.

this phase exhibits a different stacking of the layers, as illus-
trated for the oxide (Bi0.5Cu0.5)Sr2(Nd1.5Ce1.5)Cu2O10
(Figure 37). In fact, the latter structure is directly derived
from that of the Tl-1212 cuprate by replacing the single Ca
layers between the CuO5 pyramids with triple fluorite layers
Nd1.5Ce1.5O6. In other words, this structure represents the
1232 family, which consists of the intergrowth of double
RS-like layers TlSrO2 with double pyramidal copper layers
interleaved with triple fluorite-type layers. As a consequence,
!
the stacking sequence of the layers along c can be described as
(Cu, Bi)OSrOCuO2(Nd, Ce)O2(Nd, Ce)O2(Nd, Ce)
CuO2SrO(Cu, Bi)O.
Besides these single intergrowths, which are all deduced
from the parent 0212, 1212, and 2212 structures, the
possibility of multiple intergrowths is also to be considered,
though it is rare. One example is given by the oxides
Pb2Sr2LnCeCu3O10,235,236
p whichpexhibit
an orthorhombic
symmetry with a  ap 2, b  ap 2 and c  37 . The com-
plex structure of this phase (Figure 38) consists of the inter-
growth of single RS layers, with layers of CuO2 dumbbells and
with double pyramidal copper layers interleaved with double
fluorite layers. The double fluorite-type layers are occupied by
Ln and Ce according to the formula LnCeO4, the single layers
of CuO2 dumbbells are, in fact, single-ordered oxygen-
deficient perovskite layers occupied by univalent copper,
whereas the single RS layers consist of one SrO plane stacked
with one PbO plane. In fact, this structure derives from the
parent complex 02010212 intergrowth described above for
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 85

Table 4 Crystallographic parameters of the substituted HgBa Pb2Sr2Ca1xYxCu3O8 (Figure 25), by simply replacing in this
cuprates structure one single fluorite (Ca,Y)O2 layer by one double
fluorite (Ln,Ce)2O4 layer. Thus, this structure can be described
Composition Unit cell parameters
as a 02010222 multiple intergrowth. It is worth pointing out
a(A) b(A) that an excess oxygen can be intercalated in this structure at the
level of the Cu(I) layers. Although they do not belong strictly to
Hg0.8V0.2Ba2CuO4.3 3.8863 9.338 the RP family, the oxides (Ln,Ce)2(Ba,Ln)2Cu3O10d237240
Hg0.8Mo0.2Ba2CuO4.4 3.8819 9.378 can be considered as the end-members of this family, where
Hg0.9Mo0.1Ba2CuO4.2 3.8748 9.435 the RS layers have disappeared. These cuprates, which exhibit a
Hg0.8W0.2Ba2CuO4.4 3.8713 9.416 tetragonal cell, with a  ap and c  29 A, can be described as
Hg0.9W0.1Ba2CuO4.2 3.8766 9.450
derived from the ordered oxygen-deficient 123 structure by
Hg0.8Cr0.2Ba2CuO4.1 3.9261 9.306
Hg0.8Mn0.2Ba2CuO4.2 3.890 9.343
replacing the single (Ca,Y)O2 fluorite layers by double (Ln,
Hg0.8V0.2Ba2CaCu2O6.3 3.8692 12.500 Ce)2O4 fluorite layers (Figure 39). Thus, this structure consists
Hg0.8Mo0.2Ba2CaCu2O6.4 3.8602 12.576 of triple-ordered oxygen-deficient perovskite layers similar to
Hg0.8W0.2Ba2CaCu2O0.64 3.8607 12.661 those of the 123 cuprate YBa2Cu3O7, which can accept
Hg0.9W0.1Ba2CaCu2O6.2 3.8635 12.685 additional oxygen at the level of the median copper layer,
Hg0.8Ti0.2Ba2CaCu2O6.2 3.8565 12.851 forming either CuO5 pyramids or CuO6 octahedra, instead of
Hg0.8Cr0.2Ba2CaCu2O6.1 3.8753 12.488 CuO4 square planar groups. Those triple copper layers are
Hg0.8Bi0.2Ba2Ca2Cu3O8d 3.850 15.813 interleaved with double (Ln,Ce)2O4 fluorite layers instead of
Hg0.66Pb0.33Ba2Ca2Cu3O8d 3.849 15.860 single YO2 fluorite layers in the 123 structure. Thus, this
Hg0.8V0.2Ba2Ca2Cu3O8.3 3.8629 15.652
cuprate can be described as a 223 cuprate derived from the
Hg0.8Mo0.2Ba2Ca2Cu3O8.4 3.8537 15.731
Hg0.6Ti0.4Ba2Ca2Cu3O8.4 3.8507 15.707
123 phase.
Hg0.8Cr0.2Ba2Ca2Cu3O8.1 3.8631 15.625

Hg0.5Pr0.5O 2.03.7 Substitution of Carbonate Layers for RS

Layers in RP-Type Phases: Copper Oxycarbonates
SrO
CuO2 Considering the 1201 structure of the Tl or Hg cuprates, it is
easy to understand that the intermediate TlO or Hg layer
SrO belonging to the double RS layer can be replaced by a single
c
Hg0.5Pr0.5O layer of CO3 groups. This is the case with the oxycarbonate
Sr2CuO2CO3,241243
p which exhibits a tetragonal structure, with
a a  ap , or ap 2 and c  7:5 , (Figure 40), corresponding
Figure 28 Structure of the 1201 Hg cuprate Hg0.4Pr0.6Sr2CuO4d: to the intergrowth of single perovskite layers with single car-
ordering of the mercury (green spheres)!and praseodymium (blue bonate layers, and, thus, belongs to the 1201 family. Then,
spheres) in the form of rows parallel to a . The Sr atoms are represented several intergrowths corresponding to the coexistence of
by yellow spheres. Red dots correspond to additional oxygen sites. carbonate layers and RS-type layers in the same structure can

Hg0.5Pr0.5O Hg0.5Pr0.5O

SrO SrO
CuO2 CuO2
A1xPrx A1xPrx

CuO2 CuO2
SrO SrO

(a) Hg Pr (b) Hg Pr

Figure 29 Structure of the two forms of the 1212

!
cuprate Hg0.4Pr0.6Sr2Ca1x Pr Cu2O6: ordering of the mercury (green spheres), praseodymium (blue
! x
spheres) atoms in the form of rows parallel to a and stacked differently along c , leading to (a) the ap  2ap c superstructure and (b) the ap  2ap  2c
superstructure. The Sr and Ca/Pr atoms are represented by yellow and pink spheres, respectively. Red dots correspond to additional oxygen sites.
86 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

HgO
1201
(Ba, La)O CuO2
CuO2 BaO
(Ba, La)O Hg0.75(Cu,Pr)0.25O1d/2
0201
(Ba, La)O Hg0.75(Cu,Pr)0.25O1d/2
CuO2
BaO
(Ba, La)O CuO2
HgO Pr

Figure 32 Structure of the double Hg-layered mercury cuprate

Figure 30 Structure of the double 12010201 intergrowth
Hg1.5(Cu,Pr)0.5Ba2PrCu2O8. The Hg, Ba, Pr, and O atoms are represented
HgBa2La2Cu2O8. The Hg, Ba/La, and O atoms are represented by green, by green, yellow, pink, and red spheres, respectively.
yellow, and red spheres, respectively.

described as an intergrowth of 1201 Sr2CuO2CO3 carbonate

TlO
TlO
BaO
CuO2
BaO
1201 HgO
BaO
CuO2
2201
Figure 31 Structure of the 12012201 intergrowth HgTl2Ba4Cu2O10. ence between the CO3 groups and the Tl3 cations along c , the
The Hg, Tl, Ba, and O atoms are represented by green, pink, yellow, and
red spheres, respectively.
be synthesized. This is the case for the superconductors
Tl0.5Pb0.5Sr4Cu2O7CO3244 and Bi2Sr4Cu2CO3O8.245 The struc-
ture of the Tl phase is tetragonal, with a  ap and c 16.5 A,
(Figure 41(a)). It contains single octahedral perovskite layers
interconnected through single carbonate and double RS
TlSrO2 layers alternately. Thus, it can be described as an
intergrowth of two 1201 structures with the composition
Sr2CuO2CO3 and Tl0.5Pb0.5Sr2CuO5 respectively. The struc- as shown, for example, for the compounds Tl2/3Ln1/3Sr4Cu2
Bi2 Sr4Cu2CO
ture ofp p3O8 (Figure 41(b)) is orthorhombic with, CO3O7, Tl0.8 V0.2Sr4Cu2CO3O7, Hg0.5Pb0.5Sr4xBaxCu2CO3O7,
a  ap 2, b  ap 2, and c  39:5 , and contains also sin-
gle perovskite layers, but, in the latter, single carbonate layers
alternate with double Bi2SrO3 RS layers, so that it can be
layers with 2201 Bi2Sr2CuO6 layers.
Bearing in mind that the thickness difference between the
TlO and CO3 layers is not too high, it has been possible to
generate mixed layers where the two species coexist. The result-
ing structures can be described as the result of a transversal
shearing with respect to the layers in the 1201 Tlcuprates.
On application of a shearing to the 1201 structure of
!
Tl0.5Pb0.5Sr4Cu2CO3O7, along c either in the (100) plane or
in the (110) plane, every n octahedra allows two series of super-
conducting oxycarbonates to be generated. This is illustrated by
the oxycarbonate TlBa2Sr2Cu2O7CO3.246 The orthorhombic
structure of this phase, with a  ap, b  8ap, and c  16.7 A
!
(Figure 42), corresponds to a shifting by c =2 of the 1201
blocks in the (100) plane every four octahedra. Thus, this struc-
ture consists of single perovskite undulating layers BaCuO3 of
corner-shared octahedra, interconnected through mixed thal-
liumcarbonate layers, (TlO)4(CO)4. Note that these mixed
layers are flat, whereas, in contrast, the single perovskite layers
wave in the structure in order to accommodate the size differ-
!
!
first ones being located in the smaller space c , whereas the larger
Tl3 cations sit in the larger space. This oxycarbonate represents
the prototype of a series of superconducting oxycarbonates
called (100)-collapsed oxycarbonates, which exhibit various
periodicities of the shearing, as exemplified for TlSr4x
BaxCu2CO3O7, where the shearing appears every three octahe-
dra; or for Tl1xHgxBa2Sr2Cu2CO3O7, which shows a shearing
every four octahedra; or more complex sequences (see for a
review Raveau217,218).
!
The application of a crystallographic shear along c in the
(110) plane of the 1201 oxycarbonate generates a series of
cuprate superconductors, called (110)-collapsed oxycarbonates,
and HgBa2Sr2Cu2CO3O7.217,218,247,248 All of these oxycar-
bonates exhibit a tetragonal subcell a ap and c 16.7 A,
but their actual symmetry is orthorhombic and involves a
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 87

(a)

(b)

(c)

Figure 33 Single (a), double (b), and triple (c) fluorite-type layers built up of CaO8 edge-sharing cubes.

modulation of the structure or a superstructure along one direc-
tion, depending on the nature of the cations. In the Tloxycar-
bonates, the (110) shearing plane appears every five octahedra,
with a periodic waving of the octahedral copper layers that
allows the accommodation of the CO3 group and Tl3 cations
in the small and large interspaces, respectively (Figure 43(a)). In
the mercury oxycarbonates, the structure (Figure 43(b)) is char-
acterized by a more complex periodicity of the (110) shear plane
that takes places every five and six octahedra, alternately.
The application of a crystallographic shear plane to the com-
plex 12012201 structure of the oxycarbonate Bi2Sr4Cu2CO3O8
does not allow continuous CuO2 layers to be maintained due to
the size difference between the thickness of the CO3 layers and
the double Bi2O2 layers. Nevertheless, a collapsed structure can
be formed, which is not superconductive. This is the case of the
collapsed oxycarbonate, Bi15Sr29Cu12(CO3)7O56,249 which p
exhibits a monoclinic symmetry, with a  18:8 , b  ap 2,
c  39:5, and b  111 7. The structure of this phase (Figure 44)
88 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

a b
a

c
c

Figure 34 View of the structure

!
of the tetragonal T* cuprate 0222
Nd2.6Sr0.8Ce0.5Cu2O8 along a , showing double fluorite layers of edge-
sharing (Nd, Ce)1O8 cubes interleaved between two layers of CuO5
pyramids. The Nd/Sr atoms of the RS layers are represented by yellow
spheres.

Figure 35 View of the 1222 tetragonal

is obtained from a shearing in the (010) plane every seven
octahedra from the 12012201 structure. Thus, this structure
consists of 12012201 mixed carbonate ribbons that are seven
!
octahedra wide, shifted along c with respect to each other.
Although they do not contain any RS layer, the superconductive
oxycarbonates (CaCuO2)m(Ba2CuO2CO3)n250,251 can be con-
sidered as the end-members of this series, like Sr2CuO2CO3,
where the RS layers have disappeared. Their structure
(Figure 45) can also be described as derived from the 123
structure of YBa2Cu3O7 by replacing the CuO4 square planar
groups by carbonate groups in the triple copper layers.
2.03.8 Shearing Perpendicular to RP Layers:
The Collapsed Layered Oxides
Due to its great flexibility, which originates from the particular
chemistry of the Bi3 6s2 lone pair, the system BiSrCuO
offers a very rich crystal chemistry, allowing generation of
layered structures that are more complex by applying a shear-
ing mechanism to the 2201 Bi2Sr2CuO6 and 2212-type
Bi2Sr2CaCu2O8-type structures. By extension, closely related
structures can also be obtained in the systems BiBaCuO
and BiBaSrCuO.
!
structure of the
TlBa2Eu1.5Ce0.5Cu2O9 cuprate along a , showing double layers
Eu1.5Ce0.5O4 interleaved between two pyramidal copper layers,
intergrown with double RS TlBaO2 layers. The Tl, Ba, and O atoms are
represented by blue, yellow, and red spheres, respectively.
The application of a shearing mechanism transversally to the
RS and perovskite layers of the 2201Bi2Sr2CuO6-type structure
leads to the generation of a large family of layered cuprates
called collapsed or stair-like cuprates, with the generic formu-
lation (Bi2A2CuO6)n2(Bi4xA4Cu2xO12x/2), where A Sr,Ba.
Bi14xSr7Ba7Cu7xO42x/2252 and Bi17Sr16Cu7O49x253 represent
the n 7 and n 8 members of this family, respectively. The n 7
member Bi15Sr7Ba7Cu6O42.5 p crystallizes in a monoclinic cell
with a  20:1 , b  ap 2, c  25:7 , and b 111 . Its
structure (Figure 46(a)) corresponds to a periodic rupture
of the infinite layers of the 2201 oxide every seven octahedra,
forming 2201-type ribbons that are seven polyhedra wide. At the
junction between the ribbons, the polyhedra are not connected
to each other, so that these oxides cannot be classified as real
shear structures, and, for this reason, are called collapsed. It is
also quite remarkable that the RS-type blocks are not separ-
ated from one ribbon to the other and form a stair-like config-
uration. The n 8-member Bi17Sr16Cu7O49x, which exhibits an
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 89

b a b
a

c
c

Figure 36 View of the orthorhombic 2222 structure of the

Bi2Sr1.7Nd1.8Ce0.5Cu2O10 cuprate along [110]. One observes triple rock
salt layers Bi2SrO3, stacked with double fluorite layers (Nd, Ce)2O4
interleaved between two pyramidal copper layers. The Bi, Sr/Nd, and O
atoms are represented by blue, yellow, and red spheres, respectively.

Figure 37 View of the tetragonal 1232 structure of the cuprate

p !
(Bi0.5Cu0.5)Sr2(Nd1.5Ce1.5) Cu2O10d along a . The latter consists of
orthorhombic cell (a  ap 2, b 23:2 , and c 24:4 ),
shows similar ribbons (Figure 46(b)), that are eight polyhedra triple fluorite layers Nd1.5Ce1.5O6 sandwiched between two pyramidal
! copper layers, stacked with double (Bi, Cu)1SrO2 rock salt layers. The
wide and shifted with respect to each other along the c direction
Bi/Cu, Sr, and O atoms of the RS layers are represented by green,
of the 2201 Bi2Sr2CuO6d structure. Longitudinal intergrowths yellow, and red spheres, respectively.
of these different members are also observed, resulting from the
variation of the thickness of the 2201-type ribbons. This is,
for instance, the case of the n 6.5 member, which consists of
the regular intergrowth of the n 6 and n 7 members.252 consist of identical slices but parallel to the (010) plane of the
More complex shearing operations can be applied to the 2201 phase.
2201, Bi2Sr2CuO6-type structure, resulting in layered struc- Thus, from these structural observations, we can distin-
tures closely related to the 2201 family. This is the case of the guish, by analogy with the crystallographic shear structures,
cuprate Bi6Ba4Cu2O15,254 whichp exhibits
a monoclinic symme- two families of structures that differ by the nature of their
try, with a 12:2 , b ap 2, c 27 , and b 93:3 . crystallographic collapsing planes (CCP): the {010}2201 col-
This structure (Figure 47) can be deduced from the 2201 lapsed phases represented by the oxides Bi17Sr16Cu7O49 (n 8)
structure by a double shearing mechanism and, for this reason, and Bi15(Sr,Ba)14Cu6O42 (n 7), and the {012}2201 double-
is called a double-collapsed
 p2201-type
 structure. A first transla- collapsed phases represented by Bi6Ba4Cu2O15 (n 5).
tion along c2201 of ap 1 2=2 every five octahedral blocks The strategy of the shearing mechanism can be also applied
leads to the hypothetical n 5 member described above for the to the 2212-Bi2Sr2CaCu2O8-type structure. The cuprate
single 2201-collapsed family. Then, a second p translation
per- Bi16Sr28Cu17O69d255 illustrates this viewpoint. This oxide p
formed in the latter structure along b2201 of ap 2 leads to this crystallizes in a monoclinic cell, with a  48:4 , b  ap 2,
double-collapsed 2201 phase. In fact, this structure consists of c  16:3 , and b  110 . Its structure (Figure 48) derives from
2201-type slices parallel to the (012) plane of the 2201 struc- that of the 2212 bismuth cuprate by a shearing mechanism
ture, whereas the hypothetical n 5-collapsed structure would applied transversally to the perovskite/RS layers, involving a
90 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

b b
a a

c c

Figure 38 View of the orthorhombic structure of Pb2Sr2PrCeCu3O10 Figure 39 View of the tetragonal
!
223 structure of
along, b which corresponds to a double 02010222 intergrowth. The Eu1.3Ba1.3Ce1.3Cu3O10 along a showing triple 123-type copper layers
Pb, Sr, Cu, and O atoms are represented by blue, yellow, turquoise, intergrown with double fluorite (Eu, Ce)2O4 layers. The Ba/Eu atoms are
and red spheres, respectively. Note that the oxygen sites of the CuIO2 represented by yellow spheres.
groups are split.

double translation. As a result, this structure exhibits 2212
!
ribbons running along c , which are 9 Bi atoms wide, and shifted
by 5.3 A with respect to each other. The crystallographic shear
plane, with respect to the 2212 structure, corresponds to the
{011}2212 plane. However, the junction between two 2212
slices requires additional CuO6 and CuO5 polyhedra that exhibit
a different arrangement from that of the 2212 structure. As a
consequence, this structure cannot be described as a simple
shear structure but as the n 9 member of the collapsed 2212
family. Note that the copper polyhedra form [CuO2]17 segments
that are sandwiched between two [BiO]9 segments. Moreover,
through the junction between two ribbons, each [BiO]9 segment
is connected to a [SrO]26 segment on one side and to a [CuO2]17
segment on the other side. The ideal formula of this compound
is Bi18Sr26Cu17O70. In fact, its actual structure is more complex
than schematized in Figure 48, due to a waving of the [BiO]/
ribbons similar to the one observed in the incommensurate
layers of the 2201 and 2212 bismuth cuprates.
The application of the shearing mechanism to the 2212
modulated-type structure of Bi2Sr3Fe2O9 allows also collapsed
structures to be generated. A few examples are given by the
ferrites Bi12Sr18Fe10O52256,257 and Bi14Sr21Fe12O61,258 which
can be described as the m 4 and m 5 members of the generic
family [Bi2Sr3Fe2O9]m[Bi4Sr6Fe2O16], respectively. Indeed, the
monoclinic structure of pthe m 4 members Bi12Sr18Fe10O52,
with a  16:5 , b ap 2, c 30 , and b 91 4, consists
of 2212-type slices parallel to (011)2212. The latter are four
FeO6 (or FeO5) polyhedra thick and interconnected through
[Bi4xSr6Fe2xO16] slices that are two FeO6 octahedra thick.
The detailed structure of this double-collapsed ferrite shows
that the arrangement of the isolated blocks of six bismuth ions
is very similar to that of the ordered regions of the modulated
structure of the parent phase Bi2Sr3Fe2O9d, whereas the dis-
ordered bismuth regions are missing with respect to the latter.
The m 5 member, Bi14Sr21Fe12O61, exhibits a monoclinic cell
closely related to that of the m 4 member, with
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
a and in membranes for oxygen separation,267273 and for the
b solid solution (Sr4yCay)(Co4xFex)O13d.274
c Sr4Fep6O13 is characterized
Figure
!
40 View of the structure of the oxycarbonate Sr2CuO2CO3 along
a showing the triangular CO3 groups (green) connected with the
distorted and tilted CuO6 octahedra. The Sr atoms are represented by
yellow spheres.
p
a  16:6 , b  ap 2, c  35:3 , and b  90 6. Its closely
related structure (Figure 49) consists of 2212-type ribbons
that are five FeO5/FeO6 polyhedra thick. Two successive rib-
bons are shifted with respect to each other and are intercon-
nected by means of a [Bi4Sr6Fe2O16] slice. Note that the iron
polyhedra can be described either in terms of elongated FeO6
octahedra or FeO5 pyramids. The existence of one abnormally
long apical FeO distance makes octahedra better described
as FeO5 pyramids. It is quite remarkable that, in the 2212
ribbons, two successive FeO2 planes are strongly interconnected
through this apical oxygen, whereas in the cuprate this oxygen
is missing.
2.03.9 The Sr4Fe6O13-Type Structure
The strontium ferrite Sr4Fe6O13, first investigated as a normal
commensurate structure,259262 was shown to exhibit oxygen
deficiency or excess according to the formula Sr4Fe6O13d, with
various commensurate and incommensurate modulated struc-
tures depending on the oxygen content.263266 Moreover, the
possibility of substitution of cobalt for iron was shown for the
oxides Sr4Fe6xCoxO13d extensively studied for their mixed
conduction as potential materials in electrochemical processes
91
The average structure of the stoichiometric compound
p by an orthorhombic subcell
2ap 2  19  ap 2 closely related to that of the RP-phases
and more particularly to the 1201- and 2201-type structures
(Figure 50(a)). In fact, the structure resolution of one of the
incommensurate modulated structures of Sr4Fe6O13d, refined
in the superspace formalism from single crystal x-ray data,266
shows that the geometry of the iron polyhedra and their
arrangement are fundamentally different from those described
for the average structure. This modulated structure, character-
ized by the wave vector q 0.46a* c*, can be described as the
stacking of two complex layers. The first one is a [Sr2Fe2O6]/
perovskite single slab, similar to that observed in the 0201,
1201, and 2201 structures, built up of one [FeO2]/ layer
sandwiched between two [SrO]/ layers. The second one is a
[Sr2Fe4O7d]/ slab built up from a double iron [Fe4O5d]/
layer sandwiched between the two [SrO]/ layers.
In the single perovskite layers, the FeO6 octahedra are strongly
elongated, with four equatorial distances ranging from 1.92 to
2.02 A, and two apical distances comprised between 2.08 and
2.25 A. Quite remarkably, these perovskite slabs are not flat, but
!
are waving along a in the structure, according to the modulation
that takes place. In the [Sr2Fe4O7d]/ slab, the important feature
concerns the environment of iron in the [Fe4O5d]/ layer, which
dictates the modulation of the structure similar to the 6s2 lone
pair of the Bi3 cation in the 2201 structure. Indeed, the double
[Fe4O5d
!
]/ layers are strongly undulated, with main modulation
!
along b , and out-of-phase displacements along c . It results in an
alternation of convex and concave zones within the [Fe4O5d]/
double layers that are correlated to condensed and diluted Fe
regions, respectively (Figure 50(b)). Owing to the expansion of
the layer, the diluted Fe region can accommodate extra oxygen,
explaining the oxygen nonstoichiometry in this phase. Note the
analogy with the bismuth 2201 phases, which can accept extra
oxygen in a rather similar manner. It is quite remarkable that
these extra oxygen atoms are located in the median plane of the
[Fe4O5d]/ double layer and belong to two adjacent [FeO1o]
layers. The result is that the iron environment in those
[Fe4O5d]/ layers is very different from that in the perovskite
layers, that is, triple, forming trigonal bipyramids (TBP), tetrago-
nal pyramids (TP), and monocapped tetrahedral (MT). The
close relationships between this structural type and all the
other RP-type derivatives suggest that many intergrowths
remain to be discovered in this field, especially dealing with
the alternation of BiO layers and Fe4O5d layers. The oxide
Fe2Bi0.7Sr2.3Fe2O9.5d/2 (Bi4Sr10Fe24O56)275 confirms this
statement. The commensurately modulated structure of this
phase (Figure 51), can p be described as a threefold superstruc-
ture with a 3ap 2, b ap , and c 26 . It is closely
related to the 2212 structure, that is, consists of double
perovskite layers intergrown with (Bi,Sr)Fe2O3.5d/2 layers
related to the RS-type blocks.
2.03.10 Tubular Oxides
Although they are not layered structures, the tubular cuprates
and cobaltites can be described as directly derived from the
92 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

b
a

(a)
(b)
b
a
Figure 41 Structures of (a) the tetragonal oxycarbonate Tl0.5Pb0.5Sr4Cu2CO3O7 and (b) the orthorhombic oxycarbonate Bi2Sr4Cu2CO3O8. The CuO6
octahedra (turquoise) and CO3 groups (green) show splitting of some of their apices and distortions. More particularly, in the TlPb phase, the apical
corners shared by the CO3 groups and the CuO6 octahedra are split. In the Bi phase, the CuO6 octahedra are distorted and the CO3 groups show a
splitting of their equatorial oxygen atom. The Tl/Pb, Sr, Bi, and O atoms are represented by pink, yellow, blue, and red spheres, respectively.

2201-type Bi2Sr2CuO6 or Bi2Sr2CoO6 oxides, by introducing
pillars transversally to the single octahedral layers of the 2201
structure. The tubular cuprates form a large family of oxides with
the generic formulation (Bi2xSr2xCuO6)n.(Sr8x0 Cu6O16y). A
large number of them exhibit orthorhombic symmetry,276279
whereas one member, n 2, is also observed for the ortho-
rhombic cobalt oxide Bi42xSr122xCo8O28d.280,281 The cell
parameters of these p different
n members
 can p
be summarized
 forming double perovskite chains running along a . It results in
as follows: a  2ap 2 and b  2 p n 1ap 2=2 ap for
n-even members and b n 1ap 2=2 ap for n-uneven
 p 
members, and c  2 ap 3=2ap 2  24 . Pure orthorhom-
bic phases have been obtained for n 47 in the case of cup-
rates. The structure of these oxides (Figure 52) can be described
as built up of 2201-type!
slices that are n CuO6 (or CoO6)
octahedra wide along b , interconnected through an original
Sr8M6O16y layer (M Cu or Co) oriented at 90 , that is,
parallel to (010). The latter is closely related to the perovskite,
!
!
large tubes running along a , stuffed with 2201-type Bi2O2
ribbons that are n Bi polyhedra wide. This is exemplified by
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 93

Four octahedra c/2 Translation

Figure 42 Structure of the oxycarbonate TlBa2Sr2Cu2CO3O7 obtained by transverse shearing from the 1201 Tl0.5Pb0.5Sr4Cu2CO3O7 oxycarbonate
every four CuO6 octahedra. The Tl, Ba/Sr, and O atoms are represented by pink, yellow, and red spheres, respectively.

5ap2 11ap2

(a) [110]p (b) [110]p

Figure 43 Schematic structures of the (110) collapsed oxycarbonates (a) (Tl,M)Sr4Cu2CO3O7 and (b) HgBa2Sr2Cu2CO3O7. The Tl or Hg, Sr, or Ba, and
O atoms are represented by green, yellow, and red spheres, respectively.

S2CC - 2201 ribbon

Figure 44 Structural model of the collapsed oxycarbonate Bi15Sr29Cu12(CO3)7O56. The Bi, Sr, and O atoms are represented by blue, yellow, and red
spheres, respectively.
94 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

Ba2CuCO3O2

Ba2CuCO3O2 (CaCuO2)2

CaCuO2

(a) m=1 n=1

(b) m=2 n=1

Ba2CuCO3O2
Ba2CuCO3O2

Ba2CuCO3O2

(CaCuO2)3
CaCuO2

(c) m=3 n=1 (d) m=1 n=2

Figure 45 Idealized structures of the copper oxycarbonates (CaCuO2)m (Ba2CuO2CO3)n. The Ba and Ca atoms are represented by yellow and pink
spheres, respectively.

[110]p
[110]p
Bi
Bi Sr
Sr Cu
Cu

(a) 2201 collapsed : n = 7 (b) 2201 collapsed : n = 8

Figure 46 Structural models of the 2201-collapsed layered cuprates (Bi2A2CuO6)n2(Bi4xA4Cu2xO12x/2): (a) n 7-member Bi15Sr7Ba7Cu6O42.5, (b)
n 8-member Bi17Sr16Cu7O49. The Bi, Sr, or Ba and O atoms are represented by blue, yellow, and red spheres, respectively.

the n 2 member cobaltite Bi4Sr12Co8O30, whose structure
(Figure 52(a)) is built up of double ribbons of CoO6 octahedra
of the perovskite type forming practically square tubes stuffed
with 2 2 BiO rows and by the n 4 member cuprate Bi4Sr8-
Cu5O20 (Figure 52(b)), whose double and quadruple perov-
skite ribbons form tubes with a rectangular section stuffed by
2 4 BiO rows.
Besides the orthorhombic tubular cuprates, several cuprates
exhibit a closely related structure but with a different shape of the
tubes hosting the BiO ribbons. These oxides, which exhibit the
monoclinic symmetry, can be synthesized by slightly increasing
the strontium content or by Li doping.281,282 This is the case of
the n 7 members of the (Bi2xSr2xCuO6)n (Sr8xCu6O16)
family that exhibit
p a second form with this monoclinic sym-
metry a  ap 2, b 24:7 , c 24:7 , and a 96 . The
detailed structure of this phase was not completely determined,
but a structural model (Figure 53) could be established from
HREM observations.282,283 In this model, 2201-type slices that
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 95

are n 7 octahedra (or 7BiO rows) large are observed, similar to

2201 double collapsed : n = 5
Figure 47 Structural model of the n 5-double-collapsed cuprate
Bi6Ba4Cu2O15. The Bi, Ba, and O atoms are represented by blue, yellow,
and red spheres, respectively.
the orthorhombic form. The main difference with the ortho-
rhombic structure deals with the nature of the transversal
[Sr8Cu6O16]/ slices, which do not exhibit corner-sharing CuO6
octahedra but form double rows of corner-sharing CuO4 tetra-
!
hedra and CuO5 bypiramids running along a . The structure of
this monoclinic phase was described as derived from the ortho-
rhombic phase by a mechanism of shifting of the [Bi,O]/ layers,
resulting in a double intergrowth of orthorhombic and mono-
clinic members along b and c.281 In fact, the HREM study of
the Li-doped tubular cuprates283 showed that any structure
orthorhombic, monoclinic, or mixed consists of a stacking of
regular blocks, which, at their intersection, can be stacked verti-
cally, forming an orthorhombic variant, or shifted over one unit,
creating a monoclinic variant.

a
a2212
c
c2212

2212 ribbon

2212 collapsed : n = 9
Figure 48 Schematic structure of the 2212 double-collapsed cuprate Bi16Sr28Cu17O69. The Bi, Sr, and O atoms are represented by blue, yellow, and
red spheres, respectively.

c
b
!
Figure 49 View of the monoclinic collapsed structure of Bi14Sr21Fe12O61 along b , built up of 2212 ribbons that are five FeO5/FeO6 polyhedra thick. The
Bi, Bi/Sr, Sr, and O atoms are represented by blue, pink, yellow, and red spheres, respectively.
96 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

(a)

Concave
region

TP Convex
region
TP

TBP Concave
region
TBP

MT Convex
TP region
MT

(b) b

Figure 50 Structure of Sr4Fe6O13: (a) average orthorhombic structure. (b) Drawing of one [FeO1o]/ layer showing the different possible
environments of iron, that is, tetragonal pyramids (TP), trigonal bipyramids (TBP), and monocaped tetrahedrals (MT).
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides 97

!
Figure 51 View of the orthorhombic commensurate structure of Bi4Sr10Fe24O56 along b showing double octahedral perovskite layers intergrown with
(Bi, Sr)Fe2O3.5 layers related to RS blocks. Bi and Sr atoms are represented by blue and yellow spheres, respectively.

(a) bb
a
Figure 52 (Continued)
98 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides

(b) b
a
Figure 52 Orthorhombic structure of (a) the cobaltite Bi4Sr12Co8O30 and (b) the cuprate Bi4Sr8Cu5O20, n 2 and n 4 members of the tubular family
(Bi2xSr2xMO6)n.(Sr8x0 Cu6O16y), respectively. The Bi, Sr, and O atoms are represented by blue, yellow, and red spheres, respectively.

Figure 53 Idealized structure of the monoclinic form of the n 7 member of the tubular family (Bi2xSr2xCuO6)n
(Sr8x0 Cu6O16y). The Bi, Sr, and O
atoms are represented by blue, yellow, and red spheres, respectively.
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
2.03.11 Conclusion
In transition metal oxides of the Fe, Mn, Co, and Cu series, the
RP phases and derivatives form a large family of closely related
oxides, corresponding to the intergrowth of two different layers
with the perovskite and the rock salt structures, respectively.
These compounds are remarkable by their ability to accommo-
date anionic vacancies with order and disorder phenomena,
but also other structural types of layers such as fluorite type and
carbonate layers. Moreover, the great flexibility of these
structures allows shearing mechanisms to be generated, lead-
ing to collapsed layered and tubular oxides. These materials
play a crucial role in the physics of strongly electron-correlated
systems such as high Tc superconductors and colossal
magnetoresistance materials. For related chapters in this Com-
prehensive, we refer to Chapters 2.01 and 4.15.
References
1. Rao, C. N. R.; Raveau, B. Transition Metal Oxides. Wiley-VCH: New York, 1998.
2. Bednorz, J. G.; Muller, K. A. Z. Phys. B 1986, 64, 189193.
3. Rao, C. N. R.; Raveau, B. Colossal Magnetoresistance, Charge Ordering and
Related Properties of Manganese Oxides. World Scientific: Singapore, 1998.
4. Tokura, Y., Ed.; In Colossal Magnetoresistive Oxide; Gordon Breach Science:
New York, 1999.
5. Ruddlesden, S. N.; Popper, P. Acta Cryst. 1957, 10, 538540; Ruddlesden, S. N.;
Popper, P. Acta Cryst. 1958, 11, 5455.
6. Leonowicz, M. E.; Poeppelmeier, K.; Longo, J. M. J. Solid State Chem. 1985, 59,
7180.
7. Fawcett, I. D.; Sunstrom, J. E.; Greenblatt, M.; Croft, M.; Ramanjachary, K. V.
Chem. Mater. 1998, 10, 36433651.
8. Tezuka, K.; Inamura, M.; Hinatsu, Y. J. Solid State Chem. 1999, 145, 705710.
9. Bouloux, J. C.; Soubeyroux, J. L.; Le Flem, G.; Hagenmuller, P. J. Solid State
Chem. 1981, 38, 3439.
10. Bouloux, J. C.; Soubeyroux, J. L.; Daoudi, A.; Le Flem, F. Mater. Res. Bull. 1981,
16, 855860.
11. Hong, C. S.; Chi, E. O.; Kim, W. S.; Hur, N. H.; Lee, K. W.; Lee, C. H. Chem. Mater.
2001, 13, 945951.
12. Ki, R. K.; Greaves, C. J. Solid State Chem. 2000, 153, 3440.
13. Senf, D.; Ruetler, P.; Braden, M.; Friedt, O.; Bruns, D.; Cousson, A.; Bouree, F.;
Revcolevschi, A. Phys. Rev. B 2005, 71, 0244251-10244251-8.
14. Sander, K.; Lehman, U.; Mueller-Buschbaum, H. Z. Anorg. Allg. Chem. 1981,
480, 153156.
15. Daoudi, A.; Le Flem, G. J. Solid State Chem. 1972, 5, 5761.
16. Wang, X. L.; Takayama-Muromachi, E. Phys. Rev. B 2005, 72, 064401064407.
17. Wang, X. L.; Sakurai, H.; Takayama-Muromachi, E. J. Appl. Phys. 2005,
10M5190110M51903.
18. Kajitani, T.; Hosoya, S.; Hiraga, K.; Fukuda, T. J. Phys. Soc. Jpn. 1990, 59,
562570.
19. Ang, R.; Sun, Y. P.; Zhu, X. B.; Song, W. H. J. Phys. D: Appl. Phys. 2008, 41,
215009215016.
20. Yamada, K.; Matsuda, M.; Endoh, Y.; Keimer, B.; Birgeneau, R. J.; Onodera, S.;
Mizusaki, J.; Matsuura, T.; Shirane, G. Phys. Rev. B 1989, 39, 23362343.
21. Chichev, A. V.; Dlouha, M.; Vratislav, S.; Knzek, K.; Hejtmanek, J.; Marysko, M.;
Veverka, M.; Jirak, Z.; Golosova, N. O.; Kozlenko, D. P.; Savenko, B. N. Phys. Rev.
B 2006, 74, 134414134421.
22. Sanchez-Andujar, M.; Senars-Rodrguez, M. A. Solid State Sci 2004, 6, 2127.
23. Zaliznyak, I. A.; Hill, J. P.; Tranquada, J. M.; Erwin, R.; Moritomo, Y. Phys. Rev.
Lett. 2000, 85, 43534356.
24. Mastsuma, T.; Tabuchi, J.; Mizusaki, J.; Yamauchi, S.; Fueki, K. J. Phys. Chem.
Solids 1988, 49, 14031408.
25. Moritimo, Y.; Higashi, K.; Matsuda, K.; Nakamura, A. Phys. Rev. B 1997, 55,
R14725R14728.
26. James, M.; Tedesco, A.; Cassidy, D.; Colella, M.; Smythe, P. J. J. Alloys Compd.
2006, 419, 201207.
27. Zalinyak, I. A.; Hill, J. P.; Tranquada, J. M.; Erwin, R.; Moritomo, Y. Phys. Rev.
Lett. 2000, 85, 43534356.
99
28. Sathe, V. G.; Pimpale, A. V.; SIruguri, V.; Paranjpe, S. K. J. Phys. Condens. Matter
1996, 8, 38893896.
29. Ang, R.; Sun, Y. P.; Luo, X.; Hao, C. Y.; Song, W. H. J. Phys. D: Appl. Phys.
2008, 41, 045404.1045404.6.
30. Dan, S. E.; Weller, M. T.; Currie, D. B. J. Solid State Chem. 1991, 92, 237240.
31. Soubeyroux, J. L.; Courbin, P.; Fournes, L.; Fruchart, D.; Le Flem, G. J. Solid
State Chem. 1980, 313320.
32. Jenning, A. J.; Skinner, S. J.; Helgasm, O. J. Solid State Chem. 2003, 175,
207217.
33. Joubert, J. C.; Collomb, A.; Elmaleh, D.; Le Flem, G.; Daoudi, A.; Ollivier, G.
J. Solid State Chem. 1970, 2, 343346.
34. Nguyen, T. D.; Vlasse, M.; Perrin, M.; Le Flem, G. J. Solid State Chem. 1980, 32,
18.
35. Liu, X.; Lu, X. J.; Chen, X.; Lu, G. L. J. Electroceramics 2008, 21, 487490.
36. Longo, J. M.; Raccah, P. M. J. Solid State Chem. 1973, 6, 526531.
37. Grande, B.; Mueller-Buschbaum, H. K.; Schweitzer, M. Z. Anorg. Allg. Chem.
1977, 428, 120124.
38. Michel, C.; Raveau, B. Revue de Chimie Minerale 1984, 21, 407425.
39. Nguyen, N.; Choisnet, J.; Hervieu, M.; Raveau, B. J. Solid State Chem. 1981, 39,
120127.
40. Goodenough, J. B.; Demazeau, G.; Pouchard, M.; Hagenmuller, B. J. Solid State
Chem. 1973, 8, 325330.
41. Mitchell, J. F.; Millburn, J. E.; Medarl, M.; Short, S.; Jorgensen, J. D.; Fernandez-
Dias, M. T. J. Solid State Chem. 1998, 141, 599603.
42. Kubota, M.; Fujioka, H.; Hirota, K.; Ohoyama, K.; Moritomo, Y.; Yoshizawa, H.;
Endoh
, Y. J. Phys. Soc. Jpn. 2000, 69, 16061609.
43. Battle, P. D.; Green, M. A.; Laskey, N. S.; Millburn, J. E.; Radaelli, P. G.;
Rosseinsky, M. J.; Sullivan, S. P.; Vente, J. F. Phys. Rev. B 1996, 54,
1596715977; Battle, P. D.; Green, M. A.; Laskey, N. S.; Millburn, J. E.;
Radaelli, P. G.; Rosseinsky, M. J.; Sullivan, S. P.; Vente, J. F. Chem. Mater. 1997,
9, 32153221.
44. Lobanov, M. V.; Greenblah, M.; Caspi, E. N.; Jorgensen, J. D.; Sheptyakov, D. V.;
Toby, B. H.; Botez, C. E.; Stephens, P. W. J. Phys. Condens. Matter 2004, 16,
53395346.
45. Green, M. A.; Neuman, D. A. Chem. Mater. 2000, 12, 9097.
46. Zhu, J. L.; Yu, R. C.; Li, Fy; Jin, C. Q. Phys. Status Solidi 2002, 194, 151166.
47. Brisi, C.; Rolando, P. AH Della Acad. Nat. Dei Lincei, Cl. Sci. Fis. Mat. Nat. 1971,
50, 3336.
48. Sharma, I. B.; Singh, D.; Magotra, S. K. J. Alloys Compd. 1998, 269, 1316.
49. Dofrenik, M.; Grolic, L. J. Cryst Growth 1973, 21, 305306.
50. Samaras, D.; Collomb, A.; Joubert, J. C. J. Solid State Chem. 1973, 7, 337348.
51. Dan, S. E.; Weller, M. T.; Currie, D. E.; Thomas, M. F.; Al Rawwas, A. D. J. Mater.
Chem. 1993, 3, 12311237.
52. Siwen, L.; Yufang, R. Mater. Res. Bull. 1994, 29, 9931000.
53. Siwen, L.; Yufang, R. J. Solid State Chem. 1995, 114, 286288.
54. Gillie, L. J.; Hadermann, J.; Hervieu, M.; Maignan, A.; Martin, C. Chem. Mater.
2008, 20, 62316237.
55. Hickey, P. J.; Knee, C. S.; Henry, P. F.; Weller, M. T. Phys. Rev. B 2007, 75,
024113024119.
56. Battle, P. D.; Green, M. A.; Lago, J.; Millburn, J. E.; Rosseinsky, M. J.; Vente, J. F.
Chem. Mater. 1998, 10, 658664.
57. Russel, H. J.; Goodman, P.; Bulcock, S. R.; Mareh, R. H.; Kennedy, S. J.;
White, T. J.; Lincoln, F. J.; Murray, K. S. Aust. J. Chem. 1996, 49, 205217.
58. Floros, N.; Hervieu, M.; Van Tendeloo, G.; Michel, C.; Maignan, A.; Raveau, B.
Solid State Sci 2000, 2, 19.
59. Kriegel, R.; Feltz, A.; Walz, L.; Simon, A.; Mattausch, H. J. Z. Anorg. Allg. Chem.
1992, 617, 99104.
60. Teske, C. L.; Muller-Buschbaum, H. Z. Anorg. Allg. Chem. 1969, 371,
325332; Teske, C. L.; Muller-Buschbaum, H. Z. Anorg. Allg. Chem. 1970,
379, 234241.
61. Raveau, B.; Michel, C.; Hervieu, M.; Groult, D. Crystal Chemistry of High-Tc
Superconductivity Copper Oxides. Springer Verlag: Berlin, 1991;; Hadermann, J.;
Perez, O.; Creon, N.; Michel, C.; Hervieu, M. J. Mater. Chem. 2007, 17,
23442350.
62. Kriegel, R.; Bormann, H.; Feltz, A. Z. Naturforsch. B 1993, 48, 1518; Gillie, L. J.;
Wright, A. J.; Hadermann, J.; Van Tendeloo, G.; Greaves, C. J. Solid State Chem.
2002, 167, 145151.
63. Kitchen, H. J.; Saratovsky, I.; Hayward, M. A. Dalton Trans. 2010, 39,
60986105.
64. Nguyen, N.; Er-Rakho, L.; Michel, C.; Choisnet, J.; Raveau, B. Mater. Res. Bull.
1980, 15, 891897.
100 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
65. Cava, R. J.; Battlog, B.; VanDover, R. B.; Krajewski, J. J.; Wasczak, J. W.;
Fleming, R. M.; Peck, W. F.; Rupp, L. W.; Marsh, P.; James, A. C.;
Schneemeryer, L. F. Nature 1990, 345, 602604.
66. Nguyen, N.; Choisnet, J.; Raveau, B. Mater. Res. Bull. 1982, 17, 567573.
67. De Leeuw, D. M.; Mutsaers, C. A.; Geelen, G. P.; Langeyers, C. J. Solid State
Chem. 1989, 80, 276285.
68. Caignaert, V.; Retoux, R.; Michel, C.; Hervieu, M.; Raveau, B. J. Solid State Chem.
1991, 91, 4146.
69. Caignaert, V.; Retoux, R.; Michel, C.; Hervieu, M.; Raveau, B. Physica C 1990,
167, 483490.
70. Steadman, R. A.; De Leeuw, D. M.; Geelen, G. P.; Fikkee, E. Physica C 1989,
162164, 542543.
71. Grasmeder, J. R.; Weller, M. T. J. Solid State Chem. 1990, 85, 8899.
72. Gillie, L. J.; Wright, A. J.; Hadermann, J.; Van Tendeloo, G.; Greaves, C. J. Solid
State Chem. 2003, 175, 188196.
73. Yamaura, K.; Huang, Q.; Cava, R. J. J. Solid State Chem. 1999, 146,
277286.
74. Gillie, L. J.; Hadermann, J.; Hervieu, M.; Maignan, A.; Martin, C. Chem. Mater.
2008, 20, 62316237.
75. Viciu, L.; Zandbergen, H. W.; Xu, Q.; Huang, Q.; Lee, M.; Cava, R. J. J. Solid State
Chem. 2006, 179, 500511.
76. Dann, S. E.; Weller, M. T.; Currie, D. B. J. Solid State Chem. 1992, 97, 179185.
77. Dann, S. E.; Weller, M. T. J. Solid State Chem. 1995, 115, 499507.
78. Hill, J. M.; Dabrowski, B.; Mitchell, J. F.; Jorgensen, J. D. Phys. Rev. B 2006, 74,
174417174425.
79. Demont, A.; Hebert, S.; Pelloquin, D.; Maignan, A. J. Solid State Chem. 2008,
181, 13141320.
80. Dan, S. E.; Weller, M. T.; Currie, D. B. J. Solid State Chem. 1992, 97, 179185.
81. Prado, F.; Armstrong, T.; Caneiro, A.; Manthiram, A. J. Electrochem. Soc. 2001,
148, 714.
82. Doumerc, J. P.; Coutanceau, M.; Demourgues, A.; Elkaim, E.; Grenier, J. C.;
Pouchard
, M. J. Mater. Chem. 2001, 11, 7885.
83. Ganguli, A. K.; Subramanian, M. A. J. Solid State Chem. 1991, 93, 250255.
84. Matheis, D. P.; Snyder, R. L. Powder Diffr. 1990, 5, 824.
85. Bourgault, D.; Martin, C.; Michel, C.; Hervieu, M.; Provost, J.; Raveau, B. J. Solid
State Chem. 1989, 78, 326328.
86. Martin, C.; Bourgault, D.; Michel, C.; Provost, J.; Hervieu, M.; Raveau, B. Eur. J.
Inorg. Solid State Chem. 1989, 26, 16.
87. Liong, J. K.; Zhang, Y. L.; Huang, J. Q.; Xie, S. S.; Che, G. C.; Cheng, X. Science
in China 1989, 32, 826835.
88. Maignan, A.; Pelloquin, D.; Martin, C.; Hervieu, M.; Raveau, B. J. Mater. Chem.
2002, 12, 10091016.
89. Masset, A. C.; Pelloquin, D.; Martin, C.; Hervieu, M.; Raveau, B. J. Inorg. Mater.
2000, 2, 687699.
90. Allix, M.; Pelloquin, D.; Studer, F.; Nguyen, N.; Wahl, A.; Maignan, A.; Raveau, B.
J. Solid State Chem. 2002, 167, 4858.
91. Pelloquin, D.; Maignan, A.; Hervieu, M.; Michel, C.; Raveau, B. J. Solid State
Chem. 2000, 151, 210219.
92. Daniel, P.; Barbey, L.; Nguyen, N.; Ducourt, A.; Groult, D.; Raveau, B. J. Phys.
Chem. Solids 1994, 55, 795802.
93. Hervieu, M.; Maignan, A.; Martin, C.; Michel, C.; Provost, J.; Raveau, B. J. Solid
State Chem. 1988, 75, 212215.
94. Parkin, S.; Lee, V. Y.; Nazzal, A. I.; Savoy, R.; Huang, T. C.; Gurman, G.;
Beyeres, R. Phys. Rev. B 1988, 38, 65316537.
95. Chen, X.; Chen, Z. F.; Tang, Y. J. Solid State Chem. 1988, 75, 212215.
96. Shimakawa, Y.; Jorgensen, J. D.; Shaked, H.; Hitterman, R. L.; Kondo, T.;
Manako, T.; Kubo, Y. Phys. Rev. B 1995, 51, 568575.
97. Martin, C.; Maignan, A.; Labbe, P.; Chardon, J.; Hegtmanck, J.; Raveau, B. Chem.
Mater. 1995, 7, 14141418.
98. Shimakava, Y.; Jorgensen, J. D.; Hunter, B. A.; Shaked, H.; Hitterman, R. L.;
Kubo, Y.; Kondo, T.; Manako, T.; Takashi, H.; Mori, N. Physica C 1995, 253,
7178.
99. Martin, C.; Bourgault, D.; Michel, C.; Hervieu, M.; Raveau, B. Mod. Phys. Lett. B
1989, 3, 9399.
100. Liang, J. K.; Zhang, Y. L.; Huang, J. Q.; Xie, S. S.; Che, C. C.; Chen, X. R.;
Ni, Y. M.; Zhen, D. N. Z. Phys. B 1988, 73, 913.
101. Morosin, B.; Venturini, E. L.; Dunn, R. G.; Newcoma Provincio, P.; Missert, N.;
Padilla, R. R. Physica C 1998, 302, 119129.
102. Ganguli, A. K.; Nagarajan, R.; Nanjun, K. S.; Rao, C. N. R. Mater. Res. Bull. 1989,
24, 103109.
103. Rao, C. N. R.; Ganguli, A. K.; Vijarayaghavan, R. Phys. Rev. B 1989, 40,
25652567.
104. Virajayaraghavan, R.; Ganguli, A. K.; Sarma, D.; Sood, A. K.; Chandrabhas, N.;
Rao, C. N. R. Supercond. Sci. Technol. 1989, 2, 195201.
105. Li, S.; Greenblatt, M. Physica C 1989, 157, 365369.
106. Huang, Y. T.; Liu, R. S.; Wang, N. W.; Wu, P. T. Jpn. J. Appl. Phys. 1989, 28,
L1514L1517.
107. Rouillon, T.; Provost, J.; Hervieu, M.; Groult, D.; Michel, C.; Raveau, B. Physica C
1989, 159, 201209.
108. Rouillon, T.; Provost, J.; Hervieu, M.; Groult, D.; Michel, C.; Raveau, B. J. Solid
State Chem. 1990, 84, 375385.
109. Rouillon, T.; Maignan, A.; Hervieu, M.; Groult, D.; Michel, C.; Raveau, B. Physica
C 1990, 171, 713.
110. Martin, C.; Michel, C.; Maignan, A.; Hervieu, M.; Raveau, B. C. R. Acad. Sci.
1988, 307, 2732.
111. Parkin, S.; Lee, V. Y.; Nazzal, A. I.; Savoy, R.; Beyers, R.; Laplaca, S. Phys. Rev.
Lett. 1988, 61, 750753.
112. Morosin, B.; Ginley, D. S.; Schirber, J. E.; Venturini, E. L. Physica C 1988, 156,
587591.
113. Gao, Y.; Li, R.; Coppens, P.; Sleight, A. W. Acta Cryst. A 1989, 45, 1114.
114. Morosin, B.; Venturini, E. L.; Ginley, D. S. Physica C 1991, 183, 9098.
115. Ihara, H.; Sugise, R.; Hirabayshi, M.; Terada, N.; Jo, M.; Hayashi, K.; Negishi, A.;
Tokumoto, M.; Kimura, Y.; Shimomura, T. Nature 1988, 334, 510511.
116. Ogborne, D. M.; Weller, M. T. Physica C 1994, 230, 153158.
117. Sugise, R.; Hirabayashi, M.; Terada, N.; Jo, M.; Shimomura, T.; Ihara, H. Physica
C 1989, 157, 131134.
118. Weller, M. T.; Pack, M. J.; Knee, C. S.; Ogborne, D. M.; Gormezano, A. Physica C
1997, 282, 849850.
119. Daniel, P.; Barbey, L.; Groult, D.; Nguyen, N.; Van Tendeloo, G.; Raveau, B. Inorg.
Chem. 1994, 31, 235244.
120. Seguelong, T.; Maestro, P.; Grenier, J. C.; Fournes, L.; Pouchard, M. Physica B
1995, 215, 427434.
121. Courjault, S.; Doumerc, J. P.; Dordor, P.; Grenier, J. C.; Hagenmuller, P.;
Pouchard, M. Z. Anorg. Allg. Chem. 2002, 628, 20572062.
122. Lucas, S.; Groult, D.; Nguyen, N.; Hervieu, M.; Raveau, B. J. Solid State Chem.
1993, 102, 2030.
123. Caignaert, V.; Daniel, P.; Nguyen, N.; Ducouret, A.; Groult, D.; Raveau, B. J. Solid
State Chem. 1994, 112, 126131.
124. Fries, T.; Steudner, C.; Schlichenmeyer, M.; Kemmler-Sack, S.; Nissel, T.;
Huebner, R. P. J. Solid State Chem. 1994, 109, 8896.
125. Torardi, C. C.; Subramanian, M. A.; Calabrese, J. C.; Gopalakrishnan, J.; Mc
Carron, E. M.; Morrissey, K. J.; Askew, T. R.; Flippen, R. B.; Chowdry, U.;
Sleight, A. W. Phys. Rev. B 1988, 38, 225231.
126. Schimakawa, Y.; Kubo, Y.; Nakabayashi, Y.; Manako, T.; Higarashi, H. Physica C
1988, 156, 97102.
127. Martin, C.; Maignan, A.; Huve, M.; Michel, C.; Hervieu, M.; Raveau, B. Eur. J.
Solid State Chem. 1993, 30, 718.
128. Shimakawa, Y.; Kubo, Y.; Manako, T.; Igarashi, H.; Izumi, F.; Asano, H. Phys. Rev.
B 1990, 42, 1016510170.
129. Mogilevski, R.; Mitchell, J. F.; Hinks, D. G.; Argyriou, D. N.; Jorgensen, J. M.
Physica C 1995, 250, 1524.
130. Aranda, M. A.; Sinclair, D. C.; Attfield, J. P.; Mc Kenzie, A. Phys. Rev. B 1995, 51,
1274712753.
131. Wagner, J. L.; Chmaissem, O.; Jorgensen, J. D.; Hinks, D. G.; Radaelli, P. G.;
Hunter, B. A.; Jensen, W. R. Physica C 1997, 277, 170182.
132. Pederzolli, D. R.; Witscheck, G. M.; Attfield, J. P. Chem. Commun. 1997, 1997,
435436.
133. Shimakawa, Y.; Jorgensen, J. D.; Manako, T.; Kubo, Y. Phys. Rev. B 1994, 50,
1603316039.
134. Manako, T.; Kubo, Y.; Shimakawa, Y. Phys. Rev. B 1992, 46, 1101911024.
135. Maignan, A.; Martin, C.; Huve, M.; Provost, J.; Hervieu, M.; Michel, C.; Raveau, B.
Physica C 1990, 170, 350360.
136. Tarascon, J. M.; Le Page, Y.; Mc Kinnon, W. R.; Tselepsis, E.; Barboux, P.;
Bagley, B. G.; Ramesh, R. Phys. Rev. B 1990, 41, 44894501.
137. Jirak, Z.; Pollert, E.; Sedmibusky, D.; Dlouha, M.; Vratislav, S. Physica C 1992,
196, 6872.
138. Tarascon, J. M.; Le Page, Y.; Mc Kinnon, W. R.; Ramesh, R.; Eibschutz, M.;
Tselepsis, E.; Wang, E.; Hull, G. W. Physica C 1990, 167, 2034.
139. Levy, D.; Fu, W. T.; Ijdo, D.; Catti, M. Solid State Commun. 1994, 92, 659663.
140. Fries, T.; Mayer von Kuerthy, C.; Ehman, A.; Wischert, W.; Kemmler-Sack, S.
J. Less Common Met. 1990, 159, 337341.
141. Stuesser, N.; Sonntag, R.; Hohlwein, D.; Zeiske, T.; Hoser, A.; Kemmler-Sack, S.;
Fries, T.; Vogt, T. Z. Phys. B 1991, 83, 165170.
142. Toebbens, D. M.; Stusser, N.; Knorr, K.; Mayer, H. M.; Lampert, G. Mater. Sci.
Forum 2001, 378, 288293.
 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
143. Pelloquin, D.; Masset, A. C.; Maignan, A.; Hervieu, M.; Michel, C.; Raveau, B.
J. Solid State Chem. 1999, 148, 108118.
144. Tarascon, J. M.; Miceli, P. F.; Barboux, P.; Hwang, D. M.; Hull, G. W.; Giroud, M.;
Greene, L. H.; LePage, Y.; McKinnon, W. R.; Tselepis, E.; Pleizier, G.;
Eibschutz, M.; Neumann, D. A.; Rhyne, J. J. Phys. Rev. B 1989, 39,
1158711598.
145. Jakubowicz, N.; Grebille, D.; Leligny, H.; Evain, M. J. Phys. Condens. Matter
1999, 11, 39974008.
146. Tarascon, J. M.; LePage, Y.; McKinnon, W. R.; Ramesh, R.; Eibschutz, M.;
Tselepis, E.; Wang, E.; Hull, G. W. Physica C 1990, 167, 2034.
147. Jakubowicz, N.; Grebille, D.; Hervieu, M.; Leligny, M.; Raveau, B. J. Phys.
Condens. Matter 2000, 12, 53715389.
148. Michel, C.; Hervieu, M.; Borel, M. M.; Grandin, A.; Deslandes, F.; Provost, J.;
Raveau, B. Z. Phys. B 1987, 68, 421423.
149. Von Schnering, H. G.; Walz, L.; Schwartz, M.; Becker, W.; Hartweg, M.; Popp, T.;
Hettich, B.; Mueller, B.; Kaempf, G. Angew. Chem. 1988, 100, 604607.
150. Torardi, C. C.; Subramanian, M. A.; Calabrese, J. C.; Gopalakrishnan, J.; Mc
Carron, E. M.; Morrissey, K. J.; Askew, T. R.; Flippen, R. B.; Chowdry, U.;
Sleight, A. W. Phys. Rev. B 1988, 38, 225231.
151. Leligny, H.; Durcok, P.; Labbe, P.; Ledesert, M.; Raveau, B. Acta Cryst. B 1992,
48, 407418.
152. Onoda, M.; Sato, M. Solid State Commun. 1988, 67, 799804.
153. Beskrovniji, A. I.; Durcok, S.; Hejtmanek, J.; Jirak, Z.; Pollert, E.; Shelkova, I. G.
Physica C 1994, 222, 375385.
154. Subramanian, M. A.; Calabrese, J. C.; Torardi, C. C.; Gopalakrishnan, J.;
Askew, T. R.; Flippen, R. B.; Morrissey, K. J.; Chowdry, U.; Sleight, A. W. Nature
1988, 332, 420422.
155. Maignan, A.; Michel, C.; Hervieu, M.; Martin, C.; Groult, D.; Raveau, B. Mod.
Phys. Lett. B 1988, 2, 681687.
156. Tang, Y.; Lin, E.; Zhu, W.; Zhang, Y.; Zeng, X. Mod. Phys. Lett. B 1988, 2,
799804.
157. Cox, D. E.; Torardi, C.; Subramanian, M. A.; Gopalakrishnan, J.; Sleight, A. W.
Phys. Rev. B 1988, 38, 66246630.
158. Ogborne, D. M.; Weller, M. T.; Lanchester, P. C. Physica C 1992, 200,
207214.
159. Michel, C.; Martin, C.; Hervieu, M.; Maignan, A.; Provost, J.; Huve, M.; Raveau, B.
J. Solid State Chem. 1992, 96, 271286.
160. Sequeira, A.; Rajagopal, H.; Gopalakrishnan, I. K.; Sastry, P. V.; Phatak, G. M.;
Yaklmi, J. V.; Tyer, R. M. Physica C 1988, 156, 599603.
161. Heruat, A. W.; Brynestad, J.; Monk, H. A.; Specht, E. D. Physica C 1988, 152,
438444.
162. Morosin, B.; Ginley, D. S.; Venturini, E. L.; Baughman, R. S.; Tigges, C. P.
Physica C 1991, 172, 413422.
163. Johansson, L. G.; Strom, C.; Erikson, S. G.; Bryntse, I. Physica C 1994, 220,
295302.
164. Le Page, Y.; Mc Kinnon, W. R.; Tarascon, J. M.; Barboux, P. Phys. Rev. B 1989,
40, 68106816.
165. Perez, O.; Leligny, H.; Grebille, D.; Labbe, P.; Groult, D.; Raveau, B. J. Phys.
Condens. Matter 1995, 7, 1000310014.
166. Perez, O.; Leligny, H.; Grebille, D.; Greneche, J. M.; Labbe, P.; Groult, D.;
Raveau, B. Phys. Rev. B 1997, 55, 12361246.
167. Sequeira, A.; Rajagopal, H.; Yakhmi, J. V. Physica C 1988, 157, 515519.
168. Gao, Y.; Coppens, P.; Moodenbaugh, A. R. Acta Cryst. A 1993, 49, 141148.
169. Hervieu, M.; Michel, C.; Domenges, B.; Laligant, Y.; Lebail, A.; Ferey, G.;
Raveau, B. Mod. Phys. Lett. B 1988, 2, 491500.
170. Tarascon, J. M.; Le Page, Y.; Barboux, P.; Bagley, B. G.; Greene, L. H.; Mc
Kinnon, W. R.; Hull, G. W. Phys. Rev. B 1988, 37, 93829389.
171. Levin, A. A.; Smolin, X.; Sheplev, F. J. Phys. Condens. Matter 1994, 6,
35393551.
172. Sampathkumaran, E. V.; Suzuki, A.; Kohn, K.; Shibuya, T.; Thokada, A.;
Ishikawa, M. J. Appl. Phys. 1988, 27, 584586.
173. Gao, Y.; Lee, P.; Coppens, P.; Subramanian, M.; Sleight, A. W. Science 1988,
241, 954956.
174. Imai, K.; Nakai, I.; Kawashima, T.; Sueno, S.; Ono, A. Jpn. J. Appl. Phys. 1988,
27, 16611664.
175. Petricek, V.; Gao, Y.; Lee, P.; Coppens, P. Phys. Rev. B 1990, 42, 387392.
176. Sunshine, A.; Siegrist, T.; Schneemeyer, L. F.; Murphy, D. W.; Cava, R. J.;
Batlogg, B. Phys. Rev. B 1988, 38, 893897.
177. Gladyshevskii, D. E.; Flukiger, R. Acta Cryst. B 1996, 52, 3853.
178. Yamamoto, A.; Onada, A.; Takayama Muromachi, E.; Izumi, F. Phys. Rev. B
Condens. Matter 1990, 42, 42284239.
179. Sequerra, A.; Yakhmi, J. V.; Iyer, R. M.; Rajagopal, H.; Sastry, P. V. Physica C
1991, 173, 267273.
101
180. Gladyshevskii, R.; Musolino, N.; Flukiger, R. Phys. Rev. B 2004, 70, 184522-
1184522-8.
181. Hervieu, M.; Domenges, B.; Michel, C.; Raveau, B. Mod. Phys. Lett. 1988, 2,
835839.
182. Mayer Von Kuerty, G.; Fries, T.; Ehmann, A.; Loesch, S.; Schlichenmeier, M.;
Kemmler-Sack, S.; Badel, K.; Koelle, D.; Huebener, R. P. J. Less Common Met.
1989, 153, 4348.
183. Fries, T.; Mayer von Kuerty, G.; Ehmann, A.; Kemmler-Sack, S. J. Less Common
Met. 1989, 154, 17.
184. Hervieu, M.; Michel, C.; Maignan, A.; Martin, C.; Raveau, B. J. Solid State Chem.
1988, 74, 428432.
185. Otto, H. H.; Zetterer, T.; Renk, K. F. Naturwissen 1988, 75, 509510.
186. Chaplot, S. L. Phys. Rev. B 1992, 45, 48854891.
187. Ogborne, D. M.; Weller, M. T.; Lanchester, P. C. Physica C 1992, 200, 167174.
188. Sinclair, D. C.; Aranda, M. A.; Attfield, J. P.; Rodriguez Carvajal, J. Physica C
1994, 225, 307316.
189. Watanabe, T.; Fujii, T.; Matsuda, A. J. Phys. Soc. Jpn. 2003, 72, 29242929.
190. Shamrai, V. F.; Mikhailova, A. B.; Mitin, A. V. Crys. Rep. 2009, 54, 584590.
191. Zhu, X.; Feng, S.; Zhang, J.; Lu, G.; Chen, K.; Wu, K.; Gan, Z. Mod. Phys. Lett.
1989, 3, 707717.
192. Bellingeri, E.; Grasso, C.; Gladyshevkii, R. E.; Dhalli, M.; Fluekiger, R. Physica C
2000, 329, 267278.
193. Hervieu, M.; Maignan, A.; Martin, C.; Michel, C.; Provost, J.; Raveau, B. Mod.
Phys. Lett. 1988, 2, 11031109.
194. Ogborne, D. M.; Weller, M. T. Physica C 1992, 201, 5357; Ogborne, D. M.;
Weller, M. T. Physica C 1994, 223, 283290.
195. MayervonKuerty, G.; Fries, T.; Ehman, A.; Kemmler-Sack, S. J. Less Common
Met. 1989, 155, 1923.
196. Zandbergen, H. W.; Fu, W. T.; Van Ruitenbeck, J. M.; de Jongh, L. J.; Van
Tendeloo, G.; Amelinckx, S. Physica C 1989, 159, 8186.
197. Benshchop, F. J.; Helmoldt, R. B.; Maasbaut, W. J. Physica C 1994, 227,
129142.
198. Cava, R. J.; Battlog, B.; Krajewski, J. J.; Rupp, L. W.; Schneemeyer, L. F.;
Siegrist, T.; Van Dover, R. B.; Marshi, P.; Peck, W. F.; Callagher, I. K.;
Glarusm, S. H.; Marshall, J. H.; Farrow, R. C.; Waszczak, J. V.; Hull, R.; Trevor, P.
Nature 1988, 336, 211214.
199. Xue, J. S.; Greedans, J. E.; Marics, M. J. Solid State Chem. 1993, 102, 501518.
200. Retoux, R.; Michel, C.; Hervieu, M.; Raveau, B. Mod. Phys. Lett. B 1989, 3,
591595.
201. Rouillon, T.; Retoux, R.; Groult, D.; Michel, C.; Hervieu, M.; Provost, J.;
Raveau, B. J. Solid State Chem. 1989, 78, 322325.
202. Rouillon, T.; Caignaert, V.; Hervieu, M.; Michel, C.; Groult, D.; Raveau, B. J. Solid
State Chem. 1992, 97, 5664.
203. Tokiwa, A.; Nagoshi, M.; Oku, T.; Kobayashi, N.; Kikuchi, M.; Hiraga, K.;
Syono, Y. Physica C 1990, 168, 285290.
204. Putilin, S. N.; Antipov, E. V.; Marezio, M. Nature 1993, 362, 226228.
205. Huang, Q.; Lynn, J. W.; Xiong, Q.; Chu, C. W. Phys. Rev. B 1995, 52, 462470.
206. Aksenov, V. L.; Balagurov, A. M.; Savenko, B. N.; Sheptyakov, D. V.;
Glazkov, V. P.; Somenkov, V. A.; Shilstein, F.; Antipov, E. V. Physica C 1997,
275, 8792.
207. Radaelli, P. G.; Wagner, J. L.; Hunter, B. A.; Beno, M. A.; Knapp, G. S.;
Jorgensen, J. D. Physica C 1993, 216, 2935.
208. Huang, Q.; Lynn, J. W.; Meng, R. L.; Chu, C. W. Physica C 1993, 218, 356364.
209. Finger, L. W.; Hazen, R. M.; Down, R. T.; Meng, R. L.; Chu, C. W. Physica C
1994, 226, 216221.
210. Hunter, B.; Jorgensen, J. D.; Wagner, J. L.; Radaelli, P. G.; Hinks, D. G.;
Shaked, H.; Hitterman, R. L.; Von Dreele, R. B. Physica C 1994, 221, 110.
211. Wagner, J. L.; Hunter, B. A.; Hinks, D. G.; Jorgensen, J. D. Phys. Rev. B 1995, 51,
1540715414.
212. Armstrong, A. R.; David, W.; Gameson, I.; Edwards, P. P.; Capponi, J. J.;
Bordet, P.; Marezio, M. Phys. Rev. B 1995, 52, 1555115557.
213. Knopin, E. M.; Antipov, E. V.; Capponi, J. J.; Bordet, P.; Chaillout, C.; de
Brion, S.; Marezio, M.; Bobylev, A. P.; Van Tendeloo, G. Physica C 1995, 243,
222232.
214. Chmaissem, O.; Huang, Q.; Antipov, E. V.; Putilin, S. N.; Marezio, M.;
Loureiro, S. M.; Capponi, J. J.; Tholence, J. L.; Santoro, A. Physica C 1993, 217,
265272.
215. Huang, Q.; Chmaissem, O.; Capponi, J. J.; Chaillout, C.; Marezio, M.;
Tholence, J. L.; Santoro, A. Physica C 1994, 227, 19.
216. Parantham, M.; Chakounakos, B. C. CSUEB 1998, 26, 133152.
217. Raveau, B.; Michel, C.; Hervieu, M.; Maignan, A. J. Mater. Chem. 1995, 5, 803815.
218. Raveau, B.; Michel, C.; Hervieu, M.; Maignan, A. C. R. Acad. Sec. 1996, 322,
609620.
102 Ruddlesden-Popper Phases and Derivatives: Homologous Series of Transition Metal Oxides
219. Raveau, B.; Michel, C.; Hervieu, M. In Coherence in High Temperature
Superconductors; Deutscher, G., Revcolevschi, A., Eds.; World Scientific:
Singapore, 1995; pp 4268.
220. Hervieu, M.; Van Tendeloo, G.; Maignan, A.; Michel, C.; Goutenoire, F.;
Raveau, B. Physica C 1993, 216, 264272.
221. Huve, M.; Martin, C.; Van Tendeloo, G.; Maignan, A.; Michel, C.; Hervieu, M.;
Raveau, B. Solid State Commun. 1994, 90, 3740.
222. Martin, C.; Huve, M.; Van Tendeloo, G.; Maignan, A.; Michel, C.; Hervieu, M.;
Raveau, B. Physica C 1993, 212, 274278.
223. Martin, C.; Herve, M.; Van Tendeloo, G.; Goutenoire, F.; Michel, C.; Maignan, A.;
Hervieu, M.; Raveau, B. Solid State Commun. 1995, 1, 5356.
224. Akimitsu, J.; Suzuki, M.; Watanabe, M.; Sawa, H. Jpn. J. Appl. Phys. 1988, 27,
L1859L1860.
225. Sawa, H.; Suzuki, S.; Watanabe, M.; Akimitsu, J.; Matsubara, H.; Watabe, H.;
Uchida, S.; Kokusho, K.; Asano, H.; Izumi, F.; Takayama-Muromachi, E. Nature
1989, 337, 347348.
226. Izumi, F.; Takayama-Muromachi, E.; Fujimori, A.; Akimitsu, J.; Sawa, H. Physica
C 1989, 158, 440448.
227. Liu, R. S.; Hervieu, M.; Michel, C.; Maignan, A.; Martin, C.; Raveau, B.;
Edwards, P. P. Physica C 1992, 197, 131135.
228. Martin, C.; Bourgault, D.; Hervieu, M.; Michel, C.; Provost, J.; Raveau, B. Mod.
Phys. Lett. B 1989, 3, 9931000.
229. Gambardella, M. T.; Domenges, B.; Rouillon, T.; Hervieu, M.; Raveau, B. Eur. J.
Solid State Inorg. Chem. 1992, 29, 333345.
230. Maeda, T.; Sakuyamak, K.; Sakai, N.; Yamauchi, H.; Tanaka, S. Physica C 1991,
177, 337344.
231. Min, R. J.; Liang, J. K.; Wang, C.; Wu, F.; Chen, H.; Dong, C.; Rao, G. C. J. Phys.
Condens. Matter 1995, 7, 59755982.
232. Mironov, A. V.; Coppens, P.; Khasanova, N. R.; Antipov, E. V.; Petrick, V. J. Solid
State Chem. 1994, 109, 7482.
233. Tokura, Y.; Arima, T.; Takagi, H.; Uchida, S.; Ishigaki, T.; Asano, H.; Beyers, R.;
Nazzal, A. I.; Lacorre, P.; Torrance, J. B. Nature 1989, 342, 890893.
234. Chen, X.; Ding, Z.; Qian, Y.; Chen, Z.; Gao, L. Phys. Rev. B 1993, 48, 97999802;
Chen, X.; Ding, Z.; Qian, Y.; Chen, Z.; Gao, L. Physica C 1993, 216, 458462.
235. Rouillon, T.; Hervieu, M.; Domenges, B.; Raveau, B. J. Solid State Chem. 1993,
103, 6374.
236. Rouillon, T.; Groult, D.; Hervieu, M.; Raveau, B. Physica C 1990, 167, 107111.
237. Izumi, F.; Kito, H.; Sawa, H.; Akimitsu, J.; Asano, H. Physica C 1989, 160, 235242.
238. Sawa, H.; Obara, K.; Akimitsu, J.; Matsui, Y.; Horiuchi, S. J. Phys. Soc. Jpn.
1989, 58, 22522255.
239. Idemento, Y.; Hayakama, Y.; Koura, N.; Richardson, J. W.; Loong, C. K. Physica C
2000, 329, 2936.
240. Wada, T.; Ichinose, A.; Yaegashi, Y.; Yamauchi, H.; Tomaka, S. Phys. Rev. B
1990, 41, 19841989.
241. Nakata, H.; Akimitsu, J.; Katanos, S.; Minami, T.; Ogita, N.; Udagawa, M. Physica
C 1995, 255, 157166.
242. Narenda-Babu, T. G.; Fish, D. J.; Greaves, C. J. Mater. Chem. 1991, 1, 677679.
243. Nakata, H.; Okajima, H.; Yokoo, T.; Yamashita, A.; Akimitsu, J.; Katano, S. Physica
C 1996, 263, 344348.
244. Huve, M.; Michel, C.; Maignan, A.; Hervieu, M.; Martin, C.; Raveau, B. Physica C
1993, 205, 219224.
245. Pelloquin, D.; Caldes, M.; Maignan, A.; Michel, C.; Hervieu, M.; Raveau, B.
Physica C 1993, 208, 121129.
246. Goutenoire, F.; Hervieu, M.; Maignan, A.; Michel, C.; Martin, C.; Raveau, B.
Physica C 1993, 210, 359366.
247. Maignan, A.; Pelloquin, D.; Malo, S.; Michel, C.; Hervieu, M.; Raveau, B. Physica
C 1995, 249, 220233; Maignan, A.; Pelloquin, D.; Malo, S.; Michel, C.;
Hervieu, M.; Raveau, B. Physica C 1995, 246, 110.
248. Huve, M.; van Tendeloo, G.; Hervieu, M.; Maignan, A.; Raveau, B. Physica C
1994, 231, 1520.
249. Hervieu, M.; Caldes, M. T.; Pelloquin, D.; Michel, C.; Cabrera, S.; Raveau, B.
J. Mater. Chem. 1996, 6, 175181.
250. Allen, J. L.; Mercey, B.; Prellier, W.; Hamet, J. F.; Hervieu, M.; Raveau, B. Physica
C 1995, 241, 158166.
251. Hervieu, M.; Mercey, B.; Prellier, W.; Allen, J.; Hamet, J. F.; Raveau, B. J. Mater.
Chem. 1996, 6, 165173.
252. Hervieu, M.; Michel, C.; Pham, A. Q.; Raveau, B. J. Solid State Chem. 1993, 104,
289301.
253. Ikeda, Y.; Hito, H.; Shimomura, S.; Ono, Y.; Ineba, K.; Hiroi, Z.; Takano, M.
Physica C 1989, 159, 93104.
254. Hervieu, M.; Michel, C.; Caldes, M. T.; Pham, A. Q.; Raveau, B. J. Solid State
Chem. 1993, 107, 117126.
255. Hervieu, M.; Caldes, M. T.; Cabrera, S.; Michel, C.; Pelloquin, D.; Raveau, B.
J. Solid State Chem. 1995, 119, 169175.
256. Hervieu, M.; Perez, O.; Groult, D.; Grebille, D.; Leligny, H.; Raveau, B. J. Solid
State Chem. 1997, 129, 214222.
257. Perez, O.; Leligny, H.; Baldinozzi, G.; Grebille, D.; Hervieu, M.; Labbe, Ph.;
Groult, D. Phys. Rev. B 1997, 56, 56625672.
258. Allix, M.; Perez, O.; Pelloquin, D.; Hervieu, M.; Raveau, B. J. Solid State Chem.
2004, 177, 31873196.
259. Kanamaru, F.; Shimada, M.; Koizumi, M. J. Phys. Chem. Solids 1972, 33,
11691171.
260. Yoshiasa, A.; Ueno, K.; Kanamura, F.; Horuschi, H. Mater. Res. Bull. 1996, 21,
175181.
261. Fjellvag, E.; Hauback, B. C.; Bredesen, R. J. Mater. Chem. 1997, 7, 24152419.
262. Mitchell, B. J.; Richardson, J. W.; Murphy, C. D.; Balechandran, B. M.;
Hodges, J. P.; Jorgensen, J. D. J. Eur. Ceram. Soc. 2002, 22, 661671.
263. Rossell, M. D.; Abakumov, A. M.; Van Tendeloo, G.; Lomakov, M.; Istomin, S.;
Antipov, E. Chem. Mater. 2005, 17, 47174726.
264. Mellene, B.; Retoux, R.; Lepoittevin, C.; Hervieu, M.; Raveau, B. Chem. Mater.
2004, 16, 50065013.
265. Rossell, M. D.; Abakumov, A. M.; Van Tendeloo, G.; Pardo, J. A.; Santiso, J.
Chem. Mater. 2004, 16, 25782584.
266. Perez, O.; Mellene, B.; Retoux, R.; Raveau, B.; Hervieu, M. Solid State Sci. 2006,
8, 431443.
267. Steele, B. C. H. Mater. Sci. Eng. 1992, 13, 7987.
268. Minh, N. Q. J. Am. Ceram. Soc. 1993, 76, 563588.
269. Kendall, K. R.; Navas, C.; Thomas, J. K.; Loye, H. C. Solid State Ion. 1995, 82,
215223.
270. Balachandran, B. M.; Park, J. H.; Segre, C. U. Solid State Ion. 1996, 83,
6571.
271. Ma, B.; Balachandran, B. M.; Park, J. H.; Segre, C. U. J. Electrochem. Soc. 1996,
143, 17361744.
272. Kim, S.; Yang, Y. L.; Christoffersen, R.; Jacobson, A. J. Solid State Ion. 1998,
109, 187196.
273. Patrakeev, M. V.; Mitberg, E. B.; Leonidov, I. A.; Kozhevnikov, V. L. Solid State
Ion. 2001, 139, 325330.
274. Gugilla, S.; Armstrong, T.; Manthiram, A. J. Solid State Chem. 1999, 145,
260266.
275. Grebille, D.; Lepoittevin, C.; Malo, S.; Perez, O.; Nguyen, N.; Hervieu, M. J. Solid
State Chem. 2006, 179, 38493859.
276. Fuertes, A.; Miratvilles, C.; Gonzales-Calvet, J.; Vallet-Regi, M.; Obradors, X.;
Rodriguez-Carjaval, J. Physica C 1989, 157, 525530.
277. Caldes, M. T.; Navarro, J. M.; Perez, F.; Carrera, M.; Foncuberta, J.; Casan
Pastor, M.; Rodriguez-Carvajal, J.; Gonzales-Calbet, J. M.; Vallet-Regi, M.;
Garcia, A.; Fuentes, A. Chem. Mater. 1991, 3, 844852.
278. Caldes, M. T.; Hervieu, M.; Fuertes, A.; Raveau, B. J. Solid State Chem. 1992, 98,
301311.
279. Matsui, Y.; Takekawa, S.; Kishio, K.; Umezono, A.; Nakamura, S.; Tsuruta, C.;
Ibe, K. Mater. Trans. JIM 1990, 31, 595601.
280. Pelloquin, D.; Masset, A. C.; Maignan, A.; Michel, C.; Hervieu, M.; Raveau, B.
Chem. Mater. 1999, 11, 8489.
281. Perez, O.; Masset, A. C.; Leligny, H.; Baldinozzi, G.; Pelloquin, D.; Dutheil, M.
Acta Cryst. B 2002, 58, 191197.
282. Caldes, M. T.; Hervieu, M.; Fuertes, A.; Raveau, B. J. Solid State Chem. 1992, 98,
4858.
283. Van Tendeloo, G.; Domenges, B.; Caldes, M. T.; Hervieu, M.; Raveau, B. J. Solid
State Chem. 1994, 112, 161169.
 7.10 Ordered Mesoporous and Microporous Materials with Heteroatoms
C Hess, Technische Universitat Darmstadt, Darmstadt, Germany
2013 Elsevier Ltd. All rights reserved.
7.10.1 Introduction
7.10.2 Stability and Controlled Surface Functionality
7.10.3 Introduction of Heteroatoms Strategies
7.10.4 Catalytic Applications
7.10.5 Conclusion
Acknowledgment
References
7.10.1 Introduction
Since the discovery of zeolites in 1938, porous materials have
been widely studied. Presently, a large variety of porous solids
exists, with pore systems covering the entire nanometer scale.
Because of their high surface area, porous solids are used in
industry for separation and catalytic applications. According to
the International Union of Pure and Applied Chemistry, the
pores of the solids are classified on the basis of their pore
diameter as microporous (<2 nm), mesoporous (250 nm),
and macroporous (>50 nm). Zeolites are prominent members
of the first class; they are crystalline inorganic materials (mostly
alumosilicates) with a well-defined microporous structure,
providing unique catalytic properties for oil refining or petro-
chemical processes.13 Figure 1 shows industrially important
zeolite structures.4 The catalytic properties of zeolites can be
changed by introducing heteroatoms, for example, transition-
metal ions such as Cu and V, for applications in pollution
abatement or selective oxidation reactions.36
While favorable regarding various aspects of shape
selectivity,7 micropores limit the applications of zeolites
because of restricted molecular transfer and accessibility to
larger molecules.8 In zeolite synthesis, a single molecule or
ion serves as template; however, the use of templates based
on self-assembled aggregates of molecules or ions allows for
materials with enlarged pores. Using such an approach, meso-
porous solids with an ordered pore structure and a narrow
distribution of pore size were discovered in 1992 and referred
to as Mobil Composition of Matter (MCM).9 In the ensuing
period, many other mesoporous silica materials were discov-
ered using a variety of surfactants and synthesis conditions, for
example, the class of Santa Barbara (SBA) materials. Well-
known mesoporous silicas are MCM-41 (hexagonal),9,10
MCM-48 (cubic),9,10 hexagonal mesoporous silica (HMS;
wormlike),11,12 and SBA-15 (hexagonal).13,14 As an example,
Figure 2 shows transmission electron microscopy images of
silica SBA-15 in a cross-sectional view (left) and a top view
(right), highlighting the ordered arrangement of pores.15 It is
interesting that the walls of SBA-15 contain microporosity,
which is tunable to an extent by the synthesis conditions,
connecting neighboring mesopores.1618
Whereas zeolites are crystalline materials, ordered meso-
porous solids are not necessarily ordered at the atomic level.
In fact, pure mesoporous silicas possess amorphous walls.
Other ordered mesoporous solids such as Al2O3, TiO2,
Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00732-4
231
231
234
238
241
242
242
Nb2O5, Ta2O5, WO3, and SnO2 have amorphous walls,
which contain nanocrystalline domains.1922 Some of these
non-siliceous materials may be directly used as catalysts or
catalyst supports.23,24 In contrast, ordered mesoporous silicas
are mostly modified for catalytic applications, either by the
deposition of active species on the inner surface of the meso-
porous matrix or by the incorporation of active sites into the
walls.2527 The introduction of heteroatoms enables their
application in acid catalysis, acid/base reactions for fine che-
micals, and redox catalysis.27 The benefits of using such
ordered mesoporous solids in catalysis are (i) their high surface
areas, enabling a large number of active sites per gram of
catalyst and (ii) their relatively large pore sizes, which enhance
mass transport inside the pores and give accessibility to larger
molecules. In fact, diffusion in mesopores was found to be
several orders of magnitude faster as that in micropores.28
To overcome the limitations of zeolites while maintaining
their highly desirable catalytic properties, several strategies
have been followed to combine micropores and mesopores
into one material.7,29,30
There are many reviews on ordered mesoporous and micro-
porous solids covering a variety of aspects, such as synthesis,
functionalization, and catalytic applications.16,15,2527,3153
The structure of this chapter is as follows. First, modifications
of ordered mesoporous and microporous materials relevant for
applications in catalysis are discussed. Then, strategies for the
introduction of heteroatoms into mesoporous and microporous
solids are outlined. Particular attention is given to the prepara-
tion and characterization of solids containing isolated or highly
dispersed heteroatoms, as well as their application in gas-phase
oxidation reactions. Hybrid framework compounds containing
both inorganic and organic materials as integral parts of the
network are not covered.54 In addition, the discussion is
restricted to powder materials. Thus, ordered nanoporous struc-
tures in membranes prepared by, for example, anodization or
track etching are not discussed.55,56
7.10.2 Stability and Controlled Surface Functionality
The synthesis of ordered mesoporous and microporous solids
has been reviewed in detail previously; see, for example, the
work of Schuth et al.57 and Cejka et al.58 As an example, for
the template-assisted synthesis of ordered mesoporous mate-
rials, Figure 3 depicts the possible steps during the formation
231
232 Ordered Mesoporous and Microporous Materials with Heteroatoms

II
III
II II

I
I
III

(a) (b) (c)

(d) (e) (f)

Figure 1 Industrially important zeolite structures (a) LTA; (b) X and Y (FAU topology); (c) ZSM-5 (MFI topology); (d) MOR; (e) FER; (f) BEA. The
symbols I, I0 , II, II0 , III, and III0 indicate exchangeable cation sites in LTA and FAU. The lines represent the O atoms and the corners Si or Al atoms.
Reproduced from Smeets, P. J.; Woertink, J. S.; Sels, B. F.; Solomon, E. I.; Schoonheydt, R. A. Inorg. Chem. 2010, 49, 35733583, with permission from
the American Chemical Society.

100 nm 100 nm

Figure 2 TEM images of mesoporous silica SBA-15 in a cross-sectional view (left) and top view (right). Reproduced from Hess, C. Highly Dispersed
Vanadium-Oxide Catalysts. In Nanostructured catalysts: Selective Oxidations; Hess, C., Schlogl, R., Eds.; RSC: Cambridge, 2011, Chapter 12;
pp 299325, with permission from the Royal Society of Chemistry.

of silica MCM-41.10 In the following, the discussion will be
focused on the (post-synthesis) modifications of ordered meso-
porous and microporous materials relevant for their use as
catalysts. To this end, important aspects constitute their hydro-
thermal, thermal, and mechanical stability, as well as changes in
the surface functionality to either facilitate the introduction of
active sites or to remove nonselective active sites.
Hydrothermal stability is of particular importance regard-
ing catalytic applications. As compared to mesoporous solids,
zeolites possess excellent hydrothermal stability.25 The hydro-
thermal properties of important zeolite types such as ZSM-5
(MFI), mordenite (MOR), Y in dealuminated modification
(FAU), and beta (BEA) have been reported.59 Dense-structured
zeolites ZSM-5 and MOR were demonstrated to be relatively
 Ordered Mesoporous and Microporous Materials with Heteroatoms
Hexagonal
array
Surfactant micelle Micellar rod
Figure 3 Possible mechanistic pathways for the formation of MCM-41 (1) liquid crystal phase initiated and (2) silicate anion initiated.
Reproduced from Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T. W.; Olson, D. H.; Sheppard, E. W.;
McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 1083410843, with permission from the American Chemical Society.
stable under treatment for 72 h at temperatures up to 513 K,
losing a maximum of 20% of their crystallinity. In contrast,
open-structured zeolites BEA and FAU show stronger degrada-
tion, with up to 80% loss of crystallinity. Besides, it is known
that framework aluminum can be removed from the zeolite
lattice in the presence of steam, forming extra-framework alumi-
num species.8,6062 With increasing Al content, zeolites become
more affected by steaming.63 Dealumination by steaming under
well-defined conditions may generate mesoporosity in the crys-
tallites, enhancing the accessibility of larger molecules. The ben-
eficial effects of such a post-synthesis modification on catalytic
activity were demonstrated, for example, for the cracking of gas
oil over Y zeolite.64 An improvement in catalytic properties as
induced by the presence of extra-framework aluminum species
has also been observed after the introduction of rare-earth cat-
ions such as La3.65,66 An increase in the hydrothermal stability
by the addition of La was reported also for Cu or Co exchanged
ZSM-5 used for NO reduction of hydrocarbons.67,68 Very
recently, the hydrothermal stability of Fe/BEA zeolites was
enhanced by ion-exchange of rare-earth metals with radii of
1.051.15 A.69
Dealumination may also be employed to remove non-
selective sites on the external surface of zeolites by treatment
with bulky acids (e.g., oxalic acid) or complexing agents (e.g.,
EDTA), by selective coking, or by chemical vapor deposition
(CVD) of silica.53 Modification using CVD has been applied to
medium-pore zeolites to control the size of the pore openings
besides decreasing the number of nonselective external
sites.70,71 Post-synthesis modification of the internal surface of
pure silica MFI zeolite can be achieved by using methanol or
larger alcohols, such as butanol and hexanol.7275 Reducing the
hydrophilicity of Ti-containing micro- and mesoporous mate-
rials by silylation was shown to improve the selectivity of olefin
epoxidation by preventing the ring opening of the epoxide.76,77
The thermal, hydrothermal, and mechanical stabilities of
ordered mesoporous materials such as MCM-41, MCM-48,
HMS, FSM-16 (folded-sheet mesoporous material),78,79 PCH
(porous clay heterostructure),80 and SBA-15 have been the sub-
ject of a number of studies.8187 Table 1 summarizes the results
on the thermal and hydrothermal stability obtained by Cassiers
et al.82 Thermal stability depends strongly on the wall thickness
233
Silicate Calcination
MCM-41
Silicate
and the silica precursor; however, hydrothermal stability is
influenced by the wall thickness, the degree of polymerization,
as well as the pore structure. Among the ordered mesoporous
solids studied, SBA-15 shows the highest hydrothermal stability,
followed by MCM-48 and PCH. Regarding mechanical stability,
only slight differences between ordered mesoporous materials
were observed: while SBA-15 showed minor degradation at
296 MPa, essentially, all materials collapsed at a pelletizing
pressure of 450 MPa.82 In comparison, zeolites exhibit superior
mechanical stability due to their large porosity and the presence
of a stabilizing crystal stucture.88,89
The limited hydrothermal stability of ordered mesoporous
materials results from the hydrolysis of SiOSi linkages within
the relatively thin amorphous silica walls. To further improve the
hydrothermal stability, various approaches have been pursued,
including post-synthesis modifications by silylation,9092 surface
coating with zeolite nanoclusters,93,94 introduction of additional
Si87 or heteroatoms such as Al,9597 Ti,98,99 and V,92 formation of
internal porous plugs,100 repeated pH adjustment,101 and/or
addition of inorganic and organic salts during the hydrothermal
synthesis,102104 or the incorporation of heteroatoms into the
framework, such as Al,105107 Ti,105,108,109 and V.110
It has been shown that the modification of the silica surface
by complete111 or partial92 hydrophobization using silylation
is an effective means to increase the hydrothermal stability
of ordered mesoporous solids by suppressing the attack of
water. Interestingly, the modification by silylation can be
extended by introducing a catalytic functionality into a previ-
ously hydrophobized mesoporous material. This has been
demonstrated for the preparation of vanadium containing
MCM-48 catalysts.92,112 First, the isolated silanol groups of
MCM-48 are reacted with dimethyldichlorosilane. As a result,
the surface is composed of 20% bidentate species ((SiO)2Si
(CH3)2, which are completely inert toward further reaction)
and 80% monodentate species (SiOSi(CH3)2Cl). Hydrolysis
leads to the formation of SiOSi(CH3)2OH containing hy-
droxy groups, which allow for secondary anchoring of a vana-
dium precursor.
Besides the concentration of silanol groups, the structural
stability of mesoporous silicas is related to the wall thickness
(see Table 1) and the local order of the amorphous silica walls.
234 Ordered Mesoporous and Microporous Materials with Heteroatoms

Table 1 Effect of thermal and hydrothermal treatments on the surface area and primary volume of mesoporous materials

Wall Surface area (m2 g1) Pore volume (cm3 g1)

thickness
(nm)
550  C 850  C 400  C, 100  C, 100 vol%, 550  C 850  C 400  C, 100  C, 100 vol%,
30 vol%, relative humidity, 30 vol%, relative humidity,
steam, 120 h 16 h steam, 120 h 16 h
c
MCM-41 (T)a 0.97 1128 403b 1019 145 0.95 0.26b 0.47
c
MCM-41(FS)a 1.10 1027 795 864 106 0.92 0.53 0.58
a a
MCM-48 (T)a 0.93 1433 108b 1318 197 1.14 0.93
a
MCM-48 (FS)a 0.94 1319 1094 1130 168 1.22 0.74 0.94
a a
HMS 1.07 1021 213b 830 228 0.81 0.48
a
FSM-16 0.99 1172 476 789 67 0.78 0.20 0.36
d d
KIT-1 1.07 1059 967 974 0.88 0.68 0.78
d
PCH 899 441 728 335 0.50 0.25 0.43
SBA-15 2.97 632 446b 500 281 0.63 0.48b 0.55 0.47
a
Prepared from TEOS (T) or fumed silica (FS) as a silica source.
b
Data from 750  C.
c
No capillary condensation step was observed in the N2 isotherm.
d
Not measured.
Source: Taguchi, A.; Schuth, F. Mesopor. Micropor. Mater. 2005, 77, 145; Cassiers, K.; Linssen, T.; Mathieu, M.; Benjelloun, M.; Schrijnemakers, K.; Van Der Voort; P.; Cool, P.;
Vansant, E. F. Chem. Mater. 2002, 14, 23172324, with permission from Elsevier.

To this end, an increased stability has been demonstrated for OH

thicker-walled mesoporous silicas prepared by using long crystal- Si OH Si O
Si(CH2)3NH3+Cl-
lization times,113 high-temperature calcination,114,115 or hydro- Si OH Si O
OH 1. (H3CO)3Si(CH2)3NH2 Si O Si(CH2)3NH3+Cl-
thermal restructuring of as-synthesized materials.116118 In Si
comparison to other ordered silicas, SBA-15 exhibits a high intrin- Si OH 2. HCl Si O
O
sic hydrothermal stability, which was found to be further Si OH Si O
Si(CH2)3NH3+Cl-
improved by the presence of internal porous plugs100 or hetero- Si OH Si O
atoms such as Al, Ti, and V.98,119121 However, regarding catalytic Figure 4 Schematic representation of post-synthetic surface
applications, it would be desirable to develop integrated synthesis functionalization to increase the stability of mesoporous silica SBA-15.
routes, which allow for a stability increase of SBA-15 on the one See text for details. Reproduced from Herbert, R.; Wang, D.;
hand and for controlled introduction of catalytic functionality on Schomacker, R.; Schlogl, R.; Hess, C. ChemPhysChem, 2009, 10,
the other. Recently, we were able to show that the hydrothermal, 23302233, with permission from John Wiley and Sons.
thermal, and mechanical stabilities of SBA-15 can be significantly
increased by post-synthesis surface functionalization, avoiding
pore blockage or heteroatom incorporation,87 followed by functionalized with heteroatoms such as Al126 and Ti.127129
the introduction of vanadium, as discussed in more detail Regarding Ti-containing PMOs, it has been demonstrated that
below.45 Surface functionalization consists of grafting using 3- the hydrophobicity was directly correlated with their catalytic
aminopropyltrimethoxysilane (APTMS) and the subsequent reac- performance in epoxidation. Recently, internal hydrophobization
tion of the amino groups with HCl in aqueous solution, as shown by the incorporation of organic groups within the silica frame-
schematically in Figure 4. Further steps involve an ion-exchange work was combined with external hydrophobization using silyla-
process and the removal of the organic moieties by calcination.45 tion to further improve the catalytic properties of Ti-containing
The increase in stability may be explained by a smoothening of materials for epoxidation reactions.130
the rough surface of SBA-15 by the introduction of additional Si Alternatively, the polarity of the external and internal sur-
atoms (from APTMS), resulting in a corona with higher density. faces can be controlled by stepwise post-synthesis functionali-
The above examples have shown that the incorporation of zation. Following this approach, first, the external surface of
organic groups by the grafting of, for example, alkylchloro- or the material is modified while the surfactant is still inside the
aminoalkylalkoxysilanes provides an effective means for property pores. Then, the surfactant is removed from the porous matrix
modification of ordered mesoporous solids. Alternatively, peri- and the internal surface can be functionalized.130,131
odic mesoporous organosilicas (PMOs) have been prepared by
the co-condensation of silica- and organosilica precursors, as well
as the direct condensation of bissilanes. The preparation and 7.10.3 Introduction of Heteroatoms Strategies
characterization of PMOs have been the subject of a number of
reviews; see, for example, the work of Van Der Voort et al.,44 To provide porous solids with specific catalytic properties, het-
Hoffmann et al.,122,123 Hunks and Ozin,124 and Slowing et al.125 eroatoms are introduced either in the walls or on the internal
For catalytic applications, ethane-based PMOs have been surface of the porous matrix. For the introduction of heteroatoms
 Ordered Mesoporous and Microporous Materials with Heteroatoms 235

in ordered mesoporous and microporous materials, a variety of
different strategies have been developed.27,132 By far the most
frequently applied strategies are based on direct synthesis and
post-synthetic modification, as illustrated in Figure 5 for meso-
porous (silica) materials.27 Other more recent approaches
include the use of PMOs44 or functional templates.43 There are
a number of transition metals, such as Ti, V, Mo, Fe, and Cu,
which are of major interest as heteroatoms in porous solids for
selective oxidation catalysis.5,25,27,133,134 In the following discus-
sion, emphasis will be put on the preparation, characterization,
and catalytic properties of vanadium and molybdenum contain-
ing zeolite- and ordered mesoporous silica-based catalysts for
selective gas-phase oxidation reactions.
Regarding microporous solids such as zeolites, the introduc-
tion of heteroatoms, for example, vanadium, can be achieved by
direct (hydrothermal) synthesis135138 and post-synthetic mod-
ification, including solid-state reaction,139,140 wet impregna-
tion,140142 CVD,143 ion exchange,144 and approaches via
dealuminated zeolites.145 To change the properties of zeolites,
it is common to incorporate heteroatoms by substituting metal
ions in the crystalline framework (isomorphic substitution),
leading to the formation of defined sites.132 On the other hand,
ordered mesoporous silicas possess amorphous walls. Thus,
while heteroatoms have been incorporated in the framework
by direct synthesis,146 the substitution of framework silicon
atoms may result in the formation of a mixture of sites with
different local environments. With respect to ordered meso-
porous silica materials, it has turned out to be more attractive
to use their large internal surface area for anchoring metals
or metal-oxide species, most frequently via surface silanol
groups. To this end, for the introduction of foreign atoms such
as vanadium in ordered mesoporous silica materials, a variety of
post-synthetic approaches have been used, including wet impreg-
nation,147150 incipient wetness impregnation,151,152 CVD,153
molecular designed dispersion (MDD),154156 template-ion
exchange,157 grafting/ion exchange,45,152,158,159 and atomic
layer deposition (ALD).160
An important parameter in the post-synthesis modification
of ordered mesoporous silicas is the OH density, as silanol
groups represent the anchoring site for many metal or metal-
oxide species, as well as silanes commonly used as a platform for
further functionalization. Furthermore, the OH density might
influence surface properties such as the hydrophobicity but also
catalytic properties, as it determines the podality of grafted metal
or metal-oxide centers. Previously, a fully hydroxylated silica
surface of conventional silica was reported to have an OH density
of 4.25.7 OH nm2.162,163 This value was claimed to be inde-
pendent of the origin and characteristics of the sample, that is,
the specific surface area, type of pores, pore size distribution. As
compared to conventional silica, ordered mesoporous silicas
such as MCM-41, MCM-48, and SBA-15 have been shown to
exhibit lower density values, ranging between 1.4 and 3.7
OH nm2.164,165 Such behavior may be explained by their
higher order, resulting in a more condensed surface with less
silanol groups. Using FT-IR spectroscopy, the OH density in
MCM-48 was determined as 1.41.9 OH nm2.164 To this end,
for increasing pore diameter (and, thus, decreasing curvature),
an increase in the silanol density was observed.164 On the basis
of solid-state NMR studies on ordered mesoporous silicas, sila-
nol density values of 2.53 and 3.7 OH nm2 were obtained
for MCM-41 and SBA-15, respectively.165,166 For MCM-41 and
SBA-15 pretreated at 200  C in vacuum, lower values of 1.7 and
1.0 OH nm2 were determined by NMR quantification of the
amount of toluene formed by reaction with Mg(CH2Ph)2.167 For
conventional silica, such density values were obtained only at
higher pretreatment temperatures, that is, 500  C or higher.162
Besides the value of the OH density, the local arrangement of
OH groups might have an influence on surface properties such

Direct synthesis or post-synthesis

Ion
Substitution M+ exchange
grafting L Silylation Enzyme
M E encapsulation
L Nano-
Immobilization particle
Non-silica material
Organicinorganic
Surface coating hybrid framework

Mesoporous materials High surface area (1000 m2 g-1)

(MCM-41, -48, SBA-15, etc.) narrow pore size distribution
thermal stability

Figure 5 Schematic sketch of the various methods for the functionalization of mesoporous materials. Reproduced from Taguchi, A.; Schuth, F.
Mesopor. Micropor. Mater. 2005, 77, 145, with permission from Elsevier.
236 Ordered Mesoporous and Microporous Materials with Heteroatoms
as hydrophilicity and roughness. In fact, an inhomogeneous
distribution of surface OH groups has been reported for MCM-
41.81,168 Based on EPR data using TEMPO probes on the surfac-
tant head groups, it was shown that, in a hexagonal section of
MCM-41, the flat sides exhibit a higher degree of hydrophilicity
than the more condensed zones around the pore corners, cov-
ering nearly 5 A at each side of the corner, independent of the
size of the pore. In addition, comparing the nature of the inner
surface among HMSs, SBA-15 was found to exhibit a significant
amount of geminal Si(OH)2 groups in contrast to MCM-41.166
The Si(OH)2 groups have been associated with surface defects,
indicating a more irregular, that is, rougher, inner surface of
SBA-15 as compared to that of MCM-41. Such a material may
contain surface OH groups close enough for interaction,
although attached to more distant silicon atoms.
The introduction of vanadium in ordered mesopor-
ous45,169,170 and microporous169173 materials has been cov-
ered in previous reviews. In the following, besides discussing
more recent synthetic strategies, advances in the characteriza-
tion of the introduced functionality will be highlighted. The
incorporation of heteroatoms in microporous solids by
substituting metal ions in the crystalline framework offers a
possibility for the formation of defined active sites. There is
agreement in the literature that only small amounts of
vanadium species (1 wt%) can be incorporated into the
framework using direct hydrothermal synthesis.69,172 As an
alternative, a two-step post-synthesis strategy has been
proposed145,142 and applied to the preparation of vanadium
containing dealuminated BEA zeolite with a content of V
(exclusively in the framework position) of 2.05 wt% (see
Figure 6).145 In the first step of this approach, BEA zeolite is
dealuminated by treatment with nitric acid solution. This
results in the formation of vacant T sites, which are occupied
with vanadium from an aqueous solution of NH4VO3 in the
second step. Detailed characterization by a combination of IR,
x-ray absorption, diffuse reflectance UVVis, 51V NMR, and
photoluminescence spectroscopy172,173 has revealed the pres-
ence of isolated distorted tetrahedral V5 framework species
possessing one vanadyl V]O bond. In particular, by the use of
photoluminescence spectroscopy, three vanadium species with
different degrees of distortion (a < b < g) were identified,
changing from a symmetry close to Td (a) to C3v (g)
HNO3
AI
b b
Si Si
Si
H O
O O
AI
O O
O Si
Si Si
Figure 6 Two-step synthesis method for the incorporation of vanadium in microporous materials such as zeolite SiBEA. Reproduced from Dzwigaj, S.;
Che, M. Catal. Today, 2011, 169, 232241, with permission from Elsevier.
symmetry.174,175 For the a species, a hydroxylated
(SiO)2(HO)V]O site and for the b and g species, non-
hydroxylated (SiO)3V]O sites were proposed. A higher degree
of homogeneity of framework sites may be expected by apply-
ing the above strategy to ZSM-5 zeolites, exhibiting a substan-
tial chemical uniformity of sites.173 On the other hand,
vanadium containing ZSM-5 zeolites with uniform sites have
been prepared via the reaction of VOCl3 vapor with acidic OH
groups in H-ZSM-5.154,176 The structure of the grafted V5 oxo
species was investigated in detail using Raman, IR, NMR, and
x-ray absorption spectroscopy, revealing the presence of (VO2)
monomers and, starting at V/Al > 0.5, also V2O42 dimers.176
Vanadium containing ordered mesoporous silicas is mostly
prepared by post-synthesis modification via silanol groups as
anchoring sites.147161 It has been demonstrated that the large
surface area provided by ordered mesoporous silicas enables the
introduction of a high density of dispersed vanadium centers of
up to 2.3 V per square nanometer.15 With respect to catalytic
applications, it would be highly desirable to develop strategies
allowing for the preparation of materials with both increased
(hydro)thermal stability and a large number of uniform active
sites. As discussed in the following, such strategies have been
employed for the synthesis of vanadium containing MCM-
4892,112 and SBA-15 catalysts.45,87 Regarding vanadium contain-
ing MCM-48, first, silylation using dimethyldichlorosilane leads
to a hydrophobization and, thus, stability increase of the meso-
porous matrix (see Section 7.10.2). After hydrolysis, part of the
surface is composed of SiOSi(CH3)2OH. Residual hydroxy
groups are then used for the grafting of vanadium oxide surface
species by MDD of VO(acac)2. Using this approach, MCM-48-
supported vanadium oxide catalysts with a vanadium density of
0.67 V per square nanometer were reported, which contain dis-
persed tetrahedral vanadium oxide surface species with one
vanadyl V]O bond. Recently, we have demonstrated that
SBA-15-supported vanadium oxide with increased stability and
a large density of dispersed vanadium oxide species (2.3 V per
square nanometer) are accessible by controlled surface func-
tionalization, including grafting and ion exchange steps.87,45
To illustrate this strategy, Figure 7 shows a series of visible
(532 nm) Raman spectra obtained during the synthesis of
2.7 wt% V/SBA-15 (0.7 V per square nanometer).159 As the
first step, an organic framework is formed by the reaction of
NH4VO3
V
b
H
H Si Si Si
O
O O
O
H V
H
O
O O Si
Si Si
H

APTMS with silanol groups, followed by protonation of the
amino groups using aqueous HCl solution (see Figures 4 and
7(b)). Next, vanadium is introduced in the mesoporous matrix
by anion exchange of chloride by decavanadate (see Figure 7(c)).
As a result, the intact decavanadate precursor is electrostati-
cally bound within the pores in a prearranged geometry (see
Figure 7(d)). This is considered to be a crucial pre-step regarding
the formation of homogeneously distributed surface vanadium
oxide species by the thermolysis of decavanadate accompanied
by removal of the organic moieties. During dehydration, SBA-15-
supported vanadium oxide (see Figure 7(e)) transforms into a
distorted tetrahedral (VO4) structure containing one short V]O
bond (see Figure 7(f)).
For characterization of the structure and dispersion of surface
vanadium oxide species in ordered mesoporous silicas, various
techniques such as IR, Raman, UVVis, 51V NMR, EPR, x-ray
absorption, and x-ray photoelectron spectroscopy (XPS) have
been applied. Visible Raman spectroscopy is known to be a
very sensitive indicator for the presence of (micro)crystalline
V2O5 (see Figure 7(g)). Thus, monitoring the onset of V2O5
formation in a loading-dependent series of samples allows for
an estimate of the maximum dispersion of surface vanadium
oxide species. Using the grafting/ion exchange strategy outlined
above, the maximum experimental dispersion of SBA-15-
supported vanadium oxide species is reached at about 7.2 wt%
V (2.3 V per square nanometer).45 Previously, a comparable
level of dispersion has been achieved on conventional silicas by
ALD177 or multiple incipient wetness impregnation.178 The high
degree of dispersion underlines the presence of a homogeneous
707
g 487 532
f 478
Raman intensity (a.u.)
520 707
e
939 991
d 1040
592
c 826
b 480 940
a 485
400 500 600 700 800 900
Raman shift (cm-1)
Figure 7 Raman spectra during the synthesis of 2.7 wt% V/SBA-15
(0.7 V per square nanometer) showing the low-frequency region of (a)
SBA-15, (b) APTMS-SBA-15, (c) butyl ammonium decavanadate (deca),
(d) deca-SBA-15, (e) 2.7 wt% V/SBA-15 under ambient conditions, (f)
dehydrated 2.7 wt% V/SBA-15, and (g) crystalline V2O5. The spectra are
offset for clarity. Reproduced from Hess, C.; Wild, U.; Schlogl, R.
Micropor. Mesopor. Mater. 2006, 95, 339349, with permission from
Elsevier.
Ordered Mesoporous and Microporous Materials with Heteroatoms 237
distribution of surface vanadium oxide within the pores of the
mesoporous matrix. For further characterization of the disper-
sion of vanadium containing mesoporous SBA-15 catalysts, XPS
was employed, partly in combination with visible Raman spec-
troscopy and UVVis spectroscopy.152,179181 Recently, by using
XPS, it has been shown that the vanadium oxide dispersion is
influenced by the synthesis method.152 In particular, grafting/
ion exchange leads to less oligomerized species, that is, a higher
dispersion than incipient wetness impregnation. Figure 8 illus-
trates the possible structural motifs for (dehydrated) surface
vanadium oxide species supported on mesoporous silica con-
taining a distorted tetrahedral (VO4) structure with one short
vanadyl V]O bond. Recent findings from IR,182 UVVis, and
x-ray absorption spectroscopy (EXAFS, NEXAFS)181,182 have pro-
vided evidence for the presence of VOV containing vanadium
oxide species such as dimers (n 1) and/or oligomers (n > 1).
Detailed analysis of visible Raman spectra reveals two vanadyl
bands at 1027 and 1040 cm1, which have been assigned to
hydroxylated and non-hydroxylated vanadium species,
respectively.152 In the case of SBA-15-supported dispersed
vanadium oxide, an intensity ratio I1027/I1040 of 0.8 was
observed, indicating the presence of smaller oligomers (e.g.,
n 1, 2, 3). While the presence of monomeric species cannot
be excluded, also, circular structures, for example, tetramers as
well as linkages of vanadium centers via two VOV bridges, are
feasible. In particular, the tetrameric ring structure has been
considered a perfect active site model, as it can supply the four
necessary redox equivalents (electrons) for the activation
of molecular oxygen in a full catalytic cycle, for example, for
the oxidative dehydrogenation (ODH) of propane,185 assuming
that vanadium sites change their redox state by only one
998 formal unit.186
0.1 Besides vanadium, molybdenum is an important component
of many catalysts for selective gas-phase oxidation reactions. The
1040 incorporation of large molybdenum ions in microporous solids
is hindered by their rigid framework structure, as in the case of
ZSM-5.187 On the other hand, using post-synthetic modification,
998 1021 Mo/H-ZSM-5 catalysts were prepared by slurry or incipient wet-
ness impregnation of H-ZSM-5 with ammonium heptamolyb-
date (AHM), followed by treatment in air188191 or by solid-
state reaction between MoO3 and H-ZSM-5.192,193 The latter
strategy was reported to yield dimeric tetrahedrally coordinated
960 991 molybdenum oxides (Mo2O5)2 species within the zeolite cavi-
ties, interacting with two cation-exchange sites. Molybdenum
has been introduced in mesoporous materials using direct syn-
1040
thesis, for example, by the co-condensation of tetraethylorthosi-
977 licate and AHM,187,194 Mo(OiPr)5,195 or oxide-peroxide
O O
1000 1100
V V
HO O OH
O O
Si Si
n
Figure 8 Structural motifs of surface vanadium oxide supported by
mesoporous silica such as SBA-15. Reproduced from Chlosta, R.;
Tzolova-Muller, G.; Schlogl, R.; Hess, C. Catal. Sci. Technol. 2011, 1,
11751181, with permission from the Royal Society of Chemistry.
238 Ordered Mesoporous and Microporous Materials with Heteroatoms
species,196 as well as various post-synthetic strategies, such as
impregnation using AHM,197199 thermal spreading of
MoO3,200 sonochemical decomposition of Mo(CO)6 in organic
solvents,201 or grafting/ion exchange using AHM.202 As an exam-
ple, for catalyst samples prepared by direct synthesis, silica
MCM-41 containing up to 4 wt% Mo was characterized by
XRD, UVVis, photoluminescence, and x-ray absorption
spectroscopy.187 Three different types of framework molybde-
num oxide species were identified: highly dispersed tetra-
coordinated monomeric molybdate, tetra-coordinated dimeric
and/or oligomeric molybdate, as well as hexa-coordinated poly-
meric molybdate; the monomeric species was reported to be
dominant at low loadings (1 wt% Mo), whereas the latter spe-
cies was observed only at higher loadings (4 wt% Mo). Using
a strategy based on the thermal spreading of MoO3, molyb-
denum oxide was supported on mesoporous silicas MCM-41,
MCM-48, and SBA-15.203 Such catalysts showed an increase in
the dispersion in the order of SBA-15 < MCM-48< MCM-41.203
Post-synthetic impregnation has been a common strategy to
introduce molybdenum in silica SBA-15.197199 At low loadings
(1.0 Mo nm2), dispersed-surface molybdenum oxide species
have been observed,197 whereas, in some cases, also, the presence
of crystalline MoO3 has been reported.195 The detailed structure
of silica-supported (dispersed) molybdenum oxide has been a
matter of discussion in the literature. Important aspects of this
discussion concern (i) the connectivity of the molybdenum oxide
centers related to the question as to whether monomeric or
oligomeric centers or a mixture of both are present and (ii) details
of the monomeric/oligomeric structure related to the question as
to whether the molybdenum oxide species have a square pyrami-
dal structure (monooxo) or a tetrahedral structure (monooxo or
dioxo).202 Recently, the grafting/ion exchange approach outlined
above has been successfully applied to the preparation of molyb-
denum containing silica SBA-15.202 This strategy opens the pos-
sibility to anchor high loadings (3.5 Mo nm2) of exclusively
dispersed molybdenum oxide on the surface of the mesoporous
matrix. Using Raman, UVVis, IR, and x-ray absorption spectros-
copy, new insights into the structure of the dispersed molybde-
num oxide were obtained204,205 as (i) both isolated and
connected, that is, MoOMo-containing, dispersed-surface
molybdenum oxide species were identified even at low loadings
(1.0 Mo nm2) and (ii) tetrahedral and octahedral units with a
ratio of about 3:1 were observed. Current activities focus on the
structural dynamics of dispersed molybdenum oxide (monooxo
vs. dioxo), as discussed later (see Section 7.10.4).
Other strategies for the introduction of heteroatoms
such as vanadium and molybdenum in ordered mesoporous
silicas include the thermolysis of single-source molecular
precursors.206,207 This approach is based on the grafting of
oxygen-rich metal siloxide complexes, for example, OV[OSi
(OtBu)3]3, onto a silica surface and subsequent low-temperature
calcination to remove the remaining hydrocarbon groups.
Another strategy with the potential to produce single-site cata-
lysts is the incorporation of heteroatoms in the framework of
PMOs. As an example, vanadium containing ethane-PMOs were
successfully prepared by the direct condensation of 1,4-bis-
(triethoxysilyl)ethane (BTEE) in the presence of vanadyl
sulfate.208 While the above discussion focused on the introduc-
tion of atoms and highly dispersed oxidic species, it is worth
mentioning that strategies have also been developed for the
encapsulation of nano-sized clusters, for example, molybdenum
oxide clusters inside mesoporous silica matrices.209
Ordered transition metal oxides containing vanadium or
molybdenum either as bulk constituent or heteroatoms are
promising materials for catalytic applications because of their
high surface areas, pore structure, and flexibility regarding surface
composition and wall structure (amorphous, crystalline).210,211
Besides, their internal surfaces allow for further post-synthetic
modification as described above. A controlled removal of the
surfactant is often essential to avoid collapse of the mesoporous
matrix. Thermally stable mesoporous titania (350 m2 g1) was
prepared by evaporation-induced self-assembly (EISA), followed
by ammonia treatment to increase thermal stability.212 Subse-
quently, using MDD, vanadium oxide was deposited with a
maximum density of 0.7 V per square nanometer. A variety of
mesostructured binary and mixed metal oxides such as V2O5,
MoO3, VPO, VMg oxide, and VNb oxide have been prepared
by self-assembly strategies.210 However, only a few thermally
stable materials have been reported so far, for example, mesopor-
ous VMg oxide213 and VNb oxide.214
7.10.4 Catalytic Applications
A major challenge in catalysis is the rational synthesis, that is,
the design of active sites allowing for controlled activity and
selectivity. Ordered mesoporous and microporous solids with
their uniform channels and potential to vary various properties
(in part) independently offer a promising materials base to
achieve this goal. For the same reasons, both types of materials
can serve as excellent model systems. In the following, exam-
ples will be given for the catalytic application of vanadium and
molybdenum containing ordered mesoporous and micropo-
rous solids in selective gas-phase oxidation reactions. These
include the selective oxidation of methanol to formaldehyde
(FA), the selective oxidation of methane to FA, and the ODH
of propane.
The selective oxidation of methanol to FA has been com-
monly used as a probe reaction to study the types of active sites
present (redox, basic, and acidic). Redox sites are responsible
for the formation of FA, methyl formate, and dimethoxy meth-
ane, whereas acidic and basic sites give rise to dimethyl ether
and carbon oxides. Table 2 summarizes the literature results on
the catalytic performance of vanadium containing ordered
mesoporous and microporous materials in the selective oxida-
tion of methanol to FA. Catalysts were prepared by direct
synthesis (direct) and post-synthetic grafting/ion exchange
(ion-ex), dealumination/impregnation (post), and MDD. The
catalysts possess vanadium sites either incorporated in the
framework, as in the case of the microporous materials
VAPO-5, VSiBEA, and V-silicalite, or as supported vanadium
oxide surface species, as in the case of mesoporous samples.
With respect to FA selectivities, some differences are observed.
Whereas the pure silica catalyst V-silicalite behaves largely as
supported systems showing rather high FA selectivity, the other
microporous materials (containing foreign atoms, e.g., Al)
show only moderate FA selectivities, which increase as the
vanadium content is increased.216 In the case of VSiBEA, the
increase in FA selectivity has been associated with the presence
of pseudo-tetrahedral hydroxylated (SiO)2(OH)V]O species,
 Ordered Mesoporous and Microporous Materials with Heteroatoms 239

Table 2 Performance of vanadium containing mesoporous and microporous catalysts in the selective oxidation of methanol to formaldehyde
(for details, see text)

Porous matrix Prep. V (mmol g1) V (nm2) T ( C) CH3OH/O2/He m (g) X(MeOH) S(FA) Y(FA) References
(ml min1) (%) (%) (%)
b
APO-5 Direct 200 6.6/3.3/23.1c 0.2 10.2 69.2 7.1 132
SBA-15 Ion-ex. 0.45 0.6 230 3/7/1990 0.2 3.2 96.3 3.1 215
d
VSiBEA Post 0.4 250 0.04 2.4 56 1.3 216
d
VSiBEA Post 0.92 250 0.04 5.4 63 3.4 216
MCM-48 mod.a MDD 1.2 0.67 300 3.2/6.4/70.4 0.2 48 21 10.1 112
MCM-48 MDD 1.4 0.75 300 3.2/6.4/70.4 0.2 24 60 14.4 154
SBA-15 Ion-ex. 1.4 2.3 300 3/7/1990 0.2 7.8 95.8 7.5 215
MCM-41 Direct 0.22 350 5/5/90e 47.7 90.3 43.1 144
SBA-15 Ion-ex. 0.45 0.6 350 3/7/1990 0.2 15.7 94 14.8 217
MCM-48 mod.a MDD 1.2 0.67 350 3.2/6.4/70.4 0.2 80 44 35.2 112
MCM-48 MDD 1.4 0.75 350 3.2/6.4/70.4 0.2 38 85 32.3 154
SBA-15 Ion-ex. 1.4 2.3 350 3/7/1990 0.2 39 93.3 36.4 217
V-silicalite Direct 0.17 380 3/6.5/40.5 1545  103 12.9 87 11.2 218
MCM-48 MDD 0.2 0.1 400 3.2/6.4/70.4 0.2 31 98.1 30.4 154
MCM-48 MDD 0.4 0.2 400 3.2/6.4/70.4 0.2 60.9 94.7 57.7 154
MCM-48 MDD 0.7 0.31 400 3.2/6.4/70.4 0.2 75 92.8 69.6 154
SBA-15 Ion-ex. 0.45 0.6 400 3/7/1990 0.2 49 91.3 44.7 215
MCM-48 mod.a MDD 1.2 0.67 400 3.2/6.4/70.4 0.2 84 58 48.7 195
MCM-48 MDD 1.4 0.75 400 3.2/6.4/70.4 0.2 55 85 46.8 154
SBA-15 Ion-ex. 1.4 2.3 400 3/7/1990 0.2 86.7 83.2 72.1 215
a
Silica MCM-48 was modified via reaction with dimethyldichlorosilane. Subsequently, VO(acac)2 was grafted on the silylated surface. Finally, the material was calcined at 450  C.
b
The V/Al/P ratio was reported as 1/20.5/20.1.
c
Ar was used as an inert gas instead of He.
d
A CH3OH/O2 molar ratio of 2 and a total flow rate of 40 ml min1, but no flow rates for methanol and oxygen were reported.
e
No total flow rate was reported.

exhibiting moderate nucleophilicity.216 For microporous VSi-
BEA containing vanadium exclusively in the framework posi-
tions, a maximum V content of 2.05 wt% was achieved; for
higher V contents (4.7 wt%), the amount of (SiO)2(OH)V]O
species increased; however, also, extra-framework V with octa-
hedral coordination was observed. Mesoporous silicas with
their high specific surface areas allow for grafting a large num-
ber of active sites onto the surface. In fact, the use of mesopor-
ous catalysts can give higher FA yields as compared to catalysts
based on conventional silicas for the same vanadium loading.
As discussed previously,15,45 this behavior can be attributed to
the higher dispersion of vanadium oxide species on mesopor-
ous supports. Regarding catalyst stability, it should be men-
tioned that, particularly at higher loadings, supported catalysts
show a significant loss of vanadium during methanol oxida-
tion. Thus, while mesoporous materials allow for the accom-
modation of significantly larger densities of active sites, these
supported sites are less stable than the framework positions in
microporous catalysts.
The presence of titanium influences the reactivity behavior of
vanadium containing ordered mesoporous and microporous
catalysts, for example, in the selective oxidation of alcohols. In
the case of vanadium-substituted aluminophosphate molecular
sieve VAPO-5, the addition of titanium (VTAPO-5; Ti:V 0.015/
0.022) strongly reduces the FA selectivity and leads to the for-
mation of significant amounts of methyl formate and acetalde-
hyde, thus, opening new reaction pathways.132 To increase the
vanadium oxide dispersion, mesoporous MCM-41 was modi-
fied with TiO2 as the anchoring phase.219 Using the selective
oxidation of ethanol as the probe reaction, the presence of
titanium has been shown to markedly increase the activity for
acetaldehyde formation and to change the product distribution.
Thus, at higher temperatures (>550 K), acetic acid was observed
for catalysts containing a TiO2 phase. This behavior has been
explained by further oxidation of acetaldehyde over very active
redox centers present on the VxOyTiO2 phase.219
Silica-supported vanadium and molybdenum oxides
have been reported to be among the most active and selective
catalysts for the partial oxidation of methane toward
oxygenates.220 A few studies have addressed the catalytic per-
formance of vanadium and molybdenum containing ordered
mesoporous materials in the selective oxidation of methane to
FA.146,147,206,221223 In general, mesoporous silica-based cata-
lysts showed a superior performance in methane oxidation
compared to catalysts using conventional silica as the support
material. This behavior was attributed to the presence of highly
dispersed and isolated vanadium and molybdenum species
favoring FA formation. For MCM-41-supported vanadium
oxide catalysts (2.8 wt% V) prepared by impregnation, an FA
productivity of 2255 g kg1 h1 at 622  C and a GHSV of
280 000 l kg h1 were reported.146 The presence of steam was
found to be beneficial for catalytic performance. The activity
was attributed to surface VVOx species carrying one vanadyl V]
O bond and at least one VOH group. For catalysts based on
SBA-15 as support (3.0 wt% V), an FA productivity of
2790 g kg1 h1 at 625  C and a GHSV of 144 000 l kg h1
were observed.221 Improvements in catalytic performance
were reported for catalysts containing site-isolated vanadium
oxide species prepared by either the thermolysis of grafted
molecular vanadium oxide precursors onto SBA-15
240 Ordered Mesoporous and Microporous Materials with Heteroatoms
(0.47 V per square nanometer)206 or the co-condensation of
monomeric vanadium VO2(OH) 2 and TEOS in the presence of
C16TMABr (1.9 wt%).222 For the former catalyst, an FA pro-
ductivity of 5837 g kg1 h1 at 625  C and a GHSV of
403 000 l kg h1 were observed. The latter catalyst exhibited
an FA productivity of 4621 g kg1 h1 at 598  C and a GHSV
of 545 000 l kg h1. Molybdenum containing SBA-1 catalysts
prepared by direct synthesis (8.8 wt% Mo) and post-synthetic
impregnation (9.0 wt% Mo) showed a significantly better cat-
alytic performance compared to molybdenum oxide impreg-
nated on conventional silica (9.0 wt% Mo).223 For the SBA-1
catalyst containing framework molybdenum, the highest FA
yield was observed.
In the case of the catalytic oxidation of ethane over vana-
dium containing silica catalysts (0.6 V per square nanome-
ter) based on conventional silica, MCM-41, and silicalite,
ethylene was observed as the main product of selective
oxidation.224 Based on EPR analysis after reaction, the catalysts
based on conventional silica, MCM-41, and silicalite were
reported to contain large aggregates, mainly small aggregated
species, and mainly isolated (framework) species, respectively.
Vanadium containing MCM-41 catalysts showed the highest
selectivities for ODH of 50% at 500600  C. While gas-phase
contributions to the yield cannot be ruled out, this study
suggests that catalysts accommodating small vanadium oxide
aggregates may exhibit high efficiencies for the ODH of ethane.
Similar conclusions were drawn from reactivity studies over
vanadium containing MCM-41 catalysts (3.6 wt% V) prepared
by direct hydrothermal synthesis and a post-synthetic
approach.157 Framework vanadium showed only poor selec-
tivity for the ODH of ethane, whereas for supported vanadium
oxide species, selectivities toward C2H4 of 50% (550  C)
were reported. As discussed above, vanadium containing
ZSM-5 zeolites can be prepared by the sublimation of VOCl3
onto H-ZSM-5.176 Activities in ethane ODH at 400  C were
found to be significantly higher as compared to impregnated
ZSM-5 catalysts. The activity increase was attributed to the
presence of stable highly dispersed species, such as V2O42
dimers. Recently, the use of SBA-15-supported molybdenum
oxide for the selective oxidation of ethane was examined.199
Larger molybdenum oxide aggregates and molybdenum oxide
nanoparticles were found to be more active toward ethane
activation than highly dispersed molybdenum oxide species.
Acetaldehyde selectivity was highest for highly dispersed spe-
cies, whereas larger aggregates and nanoparticles favored FA
formation. For ethane oxidation toward FA, yields of 30% were
reported at 590  C and 51% conversion.199 The formation of
FA was postulated to result from the consecutive oxidation of
the primary product ethylene.
Currently, olefins are mainly obtained via steam cracking,
fluid catalytic cracking (FCC), and catalytic dehydrogenation.
However, increasing demand of olefins has intensified research
activities on alternative processes, such as ODH. By now, various
studies have addressed the catalytic performance of vanadium
containing ordered mesoporous45,81,148,149,157,168,225232 and
microporous81,168,233235 materials in propane ODH. For
microporous materials with vanadium incorporated into the
framework of silicalite (V-silicalite; with MFI structure) and
aluminophosphates ALPO4-5 (VAPO-5; with AFI structure),
relatively high propene selectivities were achieved233,234; for
example, at 470  C, V-silicalite showed 80% propene selectivity
at 20% conversion. Despite these relatively high selectivities, the
overall propene productivity is quite low due to the small
number of vanadium atoms that can be incorporated in the
framework (1 wt% V). To this end, using dealumination/
impregnation, microporous VSiBEA catalysts with higher
vanadium framework content (up to 1.44 wt% V) have been
prepared and tested in propane ODH, showing a propene selec-
tivity of 60% at 8% conversion and at 470  C.235 The catalytic
activity is attributed to the presence of tetrahedral vanadium
species containing a V]O bond and two [(SiO)2(HO)V]O]
or three [(SiO)3V]O] VOSi linkages. Regarding mesoporous
materials, catalysts with vanadium incorporated into the frame-
work walls and supported catalysts have been prepared. For
MCM-based materials, both types of catalysts exhibit very sim-
ilar catalytic behavior in propane ODH; for example, vanadium
containing MCM-41 catalysts (0.30.6 wt% V) tested at 550  C
showed 65% (55%) propene selectivity at 10% (20%)
conversion.225,226 Using N2O (instead of O2) as the oxidant,
significantly higher selectivities for propene were achieved over
MCM-41-supported vanadium oxide catalysts.229 For vanadium
oxide supported on SBA-15 (2.7 wt% V; 0.7 V per square nano-
meter), a propene selectivity of 50% was observed at 500  C
and at 10% conversion (see Figure 9).232 As discussed in
Section 7.10.2, surface functionalization of mesoporous silica
SBA-15 increases the resistance toward mechanical, thermal,
and hydrothermal damage. As a result, the catalytic performance
may be improved.87 Figure 9 shows the catalytic performance in
propane ODH at 500  C over SBA-15-supported vanadium
oxide catalysts (2.7 wt% V; 0.7 V per square nanometer), pre-
pared by incipient wetness impregnation and by the grafting/ion
exchange procedure outlined in Section 7.10.3. Both types of
catalysts were tested before and after pressure treatment at
752 MPa and showed a comparable catalytic performance
before pressure treatment. However, after pressure treatment,
80
75
70
SC3H6 [%]
65
60
55
50
45
0 2 4 6 8 10 12
XC3H8 [%]
Figure 9 Influence of surface functionalization on catalytic
performance in propane ODH at 500  C over SBA-15-supported
vanadium oxide (2.7 wt% V, 0.7 V per square nanometer) before
pressure treatment: ( ) grafting/ion-exchange, () incipient wetness
impregnation; and after pressure treatment at 752 MPa with BET area
correction: () grafting/ion-exchange, () incipient wetness
impregnation. Reproduced from Herbert, R.; Wang, D.; Schomacker, R.;
Schlogl, R.; Hess, C. ChemPhysChem 2009, 10, 23302233, with
permission from John Wiley and Sons.
 Ordered Mesoporous and Microporous Materials with Heteroatoms
the beneficial effect of surface functionalization on catalytic
performance is clearly observed, for example, at a conversion
of 34% (see Figure 9). In this case, the structure of the sup-
ported vanadium oxide species has been studied in great detail
and found to consist of hydroxylated oligomers and also mono-
mers with a distorted tetrahedral (VO4) structure containing
one short V]O bond (see Section 7.10.3).152,182184 The high-
est propene yield (24%) reported in the literature for vana-
dium containing mesoporous and microporous materials was
achieved over SBA-15-supported vanadium oxide catalysts
(2.8 wt% V; 0.69 V per square nanometer) at 600  C;150 it
should be mentioned, however, that at this temperature, the
contribution of homogeneous gas-phase processes cannot be
ruled out.
Molybdenum containing mesoporous silicas are inactive
in propane activation. Recently, the reactivity behavior of
SBA-15-supported molybdenum oxide catalysts prepared by
grafting/ion exchange202,236 and impregnation237 was studied
in the selective oxidation of propene. Grafting/ion exchange
allows for the preparation of catalysts containing highly dis-
persed molybdenum oxide species; for example, at 450  C,
such a catalyst with a molybdenum content of 5.9 wt% Mo
(0.8 Mo nm2) shows propionaldehyde and acrolein selectiv-
ities of 53% and 10% at 3% conversion, respectively.238 In the
case of a higher-loaded catalyst (13 wt% Mo) prepared by
impregnation, the presence of hexagonal MoO3 was observed
and found to be more active and selective in propene oxidation
to acrolein than reference a-MoO3.237
A few studies addressed the selective oxidation of butane
using vanadium containing mesoporous catalysts.239242
Vanadium containing MCM-41 catalysts (2 wt% V) prepared
by direct hydrothermal synthesis were tested in the ODH of
isobutane at 500  C.239 At 18% conversion, isobutene selectiv-
ities of 50% were observed, besides the formation of metha-
crolein (8%). In comparison, an MCM-41-supported
vanadium oxide prepared by a post-synthetic approach
(1.7 wt% V) showed significantly lower isobutene selectivities
of 13% at 10% conversion. Using SBA-15-supported vana-
dium oxide catalysts prepared by incipient wetness impregna-
tion, the best catalytic performance in n-butane ODH at 520  C
was obtained for a catalyst with a V content of 8.96 wt%
(1.69 V per square nanometer), showing a total dehydrogena-
tion selectivity of 27.3% at 56.1% conversion.240 Mesoporous
MCM-41 and SBA-15 were also employed for the preparation
of supported VPO catalysts for n-butane oxidation.241,242 At
400  C, maleic anhydride yields of 47% and 44% were
reported for MCM-41- and SBA-15-based catalysts,
respectively.241 It should be mentioned, however, that at the
high VPO loadings used (45%), the VPO entities were present
as nanorods inside the channels, as well as large crystallites on
the external surface.
As discussed above, catalysts based on heteroatoms contain-
ing ordered mesoporous and microporous materials offer sev-
eral advantages, such as high specific surface areas, defined pore
structures, large flexibility regarding functionalization, and the
potential to accommodate isolated and highly dispersed active
sites. In situ spectroscopy under working conditions (operando
spectroscopy) is an important tool to gain insight into the
structure of the active sites of mesoporous and microporous
catalysts.15 This is illustrated by Figure 10, which shows recent
241
results on the selective oxidation of propylene at 500  C using
silica SBA-15-supported molybdenum oxide catalysts (10.1 wt%
Mo, 2.0 Mo nm2).236 The 514-nm Raman spectra recorded
under nonreactive (a, c) and reactive (b) conditions are charac-
terized by two molybdyl stretching bands at 967 and 991 cm1
assigned to tetrahedral dimolybdyl (dioxo) and square pyrami-
dal (monooxo) molybdenum oxide species, respectively. Thus,
Raman spectroscopy reveals that the molybdenum oxide species
present under reaction conditions possess both dioxo and
monooxo structures. More detailed analysis shows that the
structure of SBA-15-supported molybdenum oxide catalysts
depends on temperature, gas-phase composition, and molybde-
num oxide loading.243 Such knowledge will be crucial for the
rational synthesis of high-surface-area solids containing active
sites in a stable configuration to resist the harsh reaction condi-
tions of selective gas-phase oxidation catalysis.
7.10.5 Conclusion
Rapid development in the chemistry of porous materials enables
the rational synthesis of compounds with specific catalytic func-
tions. As illustrated in the above discussion, controlled modifi-
cation of ordered mesoporous and microporous solids offers the
possibility to increase the (hydro)thermal and mechanical sta-
bility on the one hand and to introduce catalytic functionalities
in the framework or onto the surface of the porous matrix on the
other. In favorable cases, both modifications can be combined
within one strategy, as demonstrated for the preparation of
vanadium containing mesoporous silica catalysts. A variety of
strategies based on the substitution of framework elements, use
of hybrid materials (e.g., PMOs), or anchoring of catalytic func-
tionalities are available that can produce isolated and highly
dispersed uniform catalytic sites. The detailed structure as
well as the synthesis mechanism of such sites is nowadays
accessible using state-of-the-art spectroscopic characterization,
as illustrated in some detail for two case studies dealing with
the controlled introduction of vanadium heteroatoms into
zeolites and mesoporous silica matrixes. Regarding vanadium
containing BEA zeolite, a two-step post-synthesis strategy is
discussed, which is based on the dealumination of BEA zeolite
using nitric acid solution, resulting in the formation of vacant
T sites and their occupation with vanadium from an aqueous
solution of NH4VO3. For the controlled synthesis of mesopor-
ous silica SBA-15-supported vanadium oxide, a multistep
approach has been developed, including functionalization of
the silica surface by grafting and further reaction of aminopro-
pyltrialkoxysilanes, anion exchange of the vanadium precur-
sor, for example, decavanadate, and thermal decomposition
of the precursor material, yielding the final supported vana-
dium oxide catalyst.
As shown in the literature for a number of cases, the pres-
ence of isolated and highly dispersed sites may be beneficial for
catalytic applications, for example, gas-phase selective oxida-
tions such as methane conversion to FA or the ODH of pro-
pane. In addition, controlled surface functionalization allows
for an increase in resistance toward mechanical, thermal, and
hydrothermal damage, which may result in superior catalytic
performance. The rational synthesis of improved mesoporous
and microporous catalysts requires profound knowledge on
242 Ordered Mesoporous and Microporous Materials with Heteroatoms

1.0 a 991
b 990
c
991
0.8

Intensity (norm.)
967
0.6 968
967
0.4

0.2

0.0

700 800 900 1000 1100 1200

Raman shift [cm-1]

a b c m/z
1E-7 18
44
56
1E-8 58
60
Intensity

1E-9 72
74

1E-10

1E-11

1E-12

0 1 2 3 4 5 6
Time [h]
Figure 10 Top: in situ Raman spectra during propylene selective oxidation using SBA-15-supported molybdenum oxide (10.1 wt% Mo, 2.0 Mo nm2)
at 500  C for (a) C3H6/O2/N2 0/7/93, (b) C3H6/O2/N2 3/7/90, (c) C3H6/O2/N2 0/7/93 at a total flow rate of 90 ml min1. Spectra were normalized
to the peak at 990 cm1. Bottom: corresponding mass traces. Reproduced from Thielemann, J. P.; Hess, C. J. Catal. 2012, 288, 124126, with
permission from Elsevier.

structureactivity relations. To this end, the use of in situ spec-
troscopy under working conditions is an important tool to gain
insight or, ultimately, define the structure of the active sites.
Future developments in the field of ordered mesoporous and
microporous materials with heteroatoms will continue to ben-
efit from interdisciplinary efforts of porous materials chemistry
with related fields such as physical chemistry, chemical engi-
neering, and inorganic chemistry. For related chapters in this
Comprehensive, we refer to Chapters 4.05, 5.05, 5.06, 5.07,
6.08, and 7.11.
Acknowledgment
It is a pleasure to thank my former graduate students working
on nanoporous materials, Rita Chlosta and Jorg Thielemann.
References
1. Corma, A. Chem. Rev. 1995, 95, 559614.
2. Davis, M. E. Nature 2002, 417, 813821.
3. Corma, A. J. Catal. 2003, 216, 298312.
4. Smeets, P. J.; Woertink, J. S.; Sels, B. F.; Solomon, E. I.; Schoonheydt, R. A.
Inorg. Chem. 2010, 49, 35733583.
5. Tatsumi, T. In Zeolites and Catalysis; Cejka, J., Corma, A., Zones, S., Eds.; Wiley-
VCH: Weinheim, 2010, Chapter 23, pp 713743.
6. Centi, G.; Perathoner, S. In Zeolites and Catalysis; Cejka, J., Corma, A., Zones, S.,
Eds.; Wiley-VCH: Weinheim, 2010, Chapter 24, pp 745774.
7. Marcilly, C. R. Top. Catal. 2000, 13, 357366.
8. van Donk, S.; Janssen, A. H.; Bitter, J. H.; de Jong, K. P. Catal. Rev. 2003, 45,
297319.
9. Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartulli, J. C.; Beck, J. S. Nature
1992, 359, 710712.
10. Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.;
Schmitt, K. D.; Chu, C. T. W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.;
Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114,
1083410843.
11. Tanev, P. T.; Pinnavaia, T. J. Science 1995, 267, 865867.
12. Tanev, P. T.; Chibwe, M.; Pinnavaia, T. J. Nature 1994, 368, 321323.
13. Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.; Chmelka, B. F.;
Stucky, G. D. Science 1998, 279, 548552.
14. Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. J. Am. Chem. Soc.
1998, 120, 60246032.
15. Hess, C. In Nanostructured Catalysts: Selective Oxidations; Hess, C., Schlogl, R.,
Eds.; RSC: Cambridge, 2011, Chapter 12, pp 299325.
16. Galarneau, A.; Cambon, H.; Renzo, F. D.; Ryoo, R.; Choi, M.; Fajula, F. New J.
Chem. 2003, 27, 7379.
 Ordered Mesoporous and Microporous Materials with Heteroatoms
17. Ryoo, R.; Ko, C. H.; Kruk, M.; Antochshuk, V.; Jaroniec, M. J. Phys. Chem. B
2000, 104, 1146511471.
18. Liu, Z.; Terasaki, O.; Ohsuna, T.; Hiraga, K.; Shin, H. J.; Ryoo, R. ChemPhysChem
2001, 2, 229231.
19. Bagshaw, S. A.; Pinnavaia, T. J. Angew. Chem. Int. Ed. 1996, 35, 11021105.
20. Yang, P.; Zhao, D.; Margolese, D. I.; Chmelka, B. F.; Stucky, G. D. Nature 1998,
396, 152155.
21. Yang, P.; Zhao, D.; Margolese, D. I.; Chmelka, B. F.; Stucky, G. D. Chem. Mater.
1999, 11, 28132826.
22. Lee, B.; Lu, D.; Kondo, J. N.; Domen, K. Chem. Commun. 2001, 21182119.
23. He, X.; Antonelli, D. Angew. Chem. Int. Ed. 2002, 41, 214229.
24. Rao, Y.; Antonelli, D. M. J. Mater. Chem. 2009, 19, 19371944.
25. Corma, A. Chem. Rev. 1997, 97, 23732419.
26. Trong On, D.; Desplantier-Giscard, D.; Danumah, C.; Kaliaguine, S. Appl. Catal. A:
Gen. 2003, 253, 545602.
27. Taguchi, A.; Schuth, F. Micropor. Mesopor. Mater. 2005, 77, 145.
28. Weisz, P. B. Chemtech 1973, 3, 498505.
29. Meynen, V.; Cool, P.; Vansant, E. F. Micropor. Mesopor. Mater. 2007, 104,
2638.
30. Perez-Ramirez, R. J.; Christensen, C. H.; Egeblad, K.; Christensen, C. H.;
Groen, J. C. Chem. Soc. Rev. 2008, 37, 25302542.
31. Ciesla, U.; Schuth, F. Micropor. Mesopor. Mater. 1999, 27, 131149.
32. Tuel, A. Micropor. Mesopor. Mater. 1999, 27, 151169.
33. Thomas, J. M. Angew. Chem. Int. Ed. 1999, 38, 35883628.
34. Anwander, R. Chem. Mater. 2001, 13, 44194438.
35. Schuth, F. Chem. Mater. 2001, 13, 31843195.
36. Liu, Y.; Pinnavaia, T. J. J. Mater. Chem. 2002, 12, 31793190.
37. Wight, A. P.; Davis, M. E. Chem. Rev. 2002, 102, 35893613.
38. De Vos, D. E.; Dams, M.; Sels, B. F.; Jacobs, P. A. Chem. Rev. 2002, 102,
36153640.
39. Soler-Illia, G. J.; de, A. A.; Sanchez, C.; Leveau, B.; Patarin, J. Chem. Rev. 2002,
102, 40934138.
40. Wingen, A.; Kleitz, F.; Schuth, F. In Basic Principles in Applied Catalysis;
Baerns, M., Ed.; Springer: Berlin, 2003 p 281.
41. Di Renzo, F.; Galarneau, A.; Trens, P.; Fajula, F. In Handbook of Porous Solids;
Schuth, F., Sing, K. S. W., Weitkamp, J., Eds.; Wiley-VCH: Weinheim, 2002;
Vol. 3, Chapter 4, pp 13111395.
42. Angloher, A.; Bein, T. In Molecular Sieves: From Basic Research to Industrial
Applications Studies in Surface Science and Catalysis; Cejka, J., Zilkova, N.,
Nachtigall, P., Eds.; Elsevier: Amsterdam, 2005; Vol. 158, pp 20172026, Part B.
43. Vinu, A.; Hossain, K. Z.; Ariga, K. J. Nanosci. Nanotechnol. 2005, 5, 347371.
44. Van Der Voort, P.; Vercaemst, C.; Schaubroeck, D.; Verpoort, F. Phys. Chem.
Chem. Phys. 2008, 10, 347360.
45. Hess, C. ChemPhysChem 2009, 10, 319326.
46. Slowing, I. I.; Vivero-Escoto, J. L.; Trewyn, B. G.; Lin, V. S.-Y. J. Mater. Chem.
2010, 20, 78898174.
47. Wu, Z.; Zhao, D. Chem. Commun. 2011, 47, 33323338.
48. Strohmaier, K. G. In Zeolites and Catalysis; Cejka, J., Corma, A., Zones, S., Eds.;
Wiley-VCH: Weinheim, 2010 Chapter 2; pp 5786.
49. Wilson, S. T. In From Zeolites to Porous MOF Materials The 40th Anniversary of
IZC; Xu, R., Gao, Z., Chen, J., Yan, W., Eds.; Elsevier B.V: SSSC, 2007
Part A; p 3.
50. Kuhl, G. H. In Catalysis and Zeolites Fundamentals and Applications;
Weitkamp, J., Puppe, L., Eds.; Springer: Berlin, 1999 Chapter 3, p 81.
51. Szostak, R. In Introduction to Zeolite Science and Practice Studies in Surface
Science and Catalysis; Bekkum, H., Jacobs, P. A., Flanigen, E., Jansen, J. C.,
Eds.; Elsevier: Amsterdam, 2001; Vol. 137; Chapter 6, p 261.
52. Chen, C.-Y.; Zones, S. I. In Zeolites and Catalysis, Synthesis, Reactions and
Applications; Cejka, J., Corma, A., Zones, S., Eds.; Wiley-VCH: Weinheim, 2010
Chapter 6; p 155.
53. Martnez, C.; Corma, A. Coord. Chem. Rev. 2011, 255, 15581580.
54. Cheetham, A. K.; Rao, C. N. R.; Fellera, R. K. Chem. Commun. 2006, 47804795.
55. Schmid, G. J. Mater. Chem. 2002, 12, 12311238.
56. Apel, P. Radiat. Meas. 2001, 34, 559566.
57. Schuth, F., Sing, K. S. W., Weitkamp, J., Eds.; In Handbook of Porous Solids;
Wiley-VCH: Weinheim, 2002.
58. Cejka, J., Corma, A., Zones, S., Eds.; In Zeolites and Catalysis; Wiley-VCH:
Weinheim, 2010.
59. Lutz, W.; Toufar, H.; Kurzhals, R.; Suckow, M. Adsorption 2005, 11, 405413.
60. Beyerlein, R. A.; Choi-Feng, C.; Hall, J. B.; Huggins, B. J.; Ray, G. J. Top. Catal.
1997, 4, 2741.
61. Triantafillidis, C. S.; Vlessidis, A. G.; Evmiridis, N. P. Ind. Eng. Chem. Res. 2000,
39, 307319.
243
62. Triantafillidis, C. S.; Vlessidis, A. G.; Nalbandian, L.; Evmiridis, N. P. Micropor.
Mesopor. Mater. 2001, 47, 369388.
63. Breck, D. W.; Flanigen, E. M. In Molecular Sieves; Barrer, R. M., Ed.; Society of
the Chemical Industry: London, 1968 pp 4761.
64. Corma, A. Stud. Surf. Sci. Catal. 1989, 49, 49.
65. Carajal, R.; Chu, P.-J.; Lunsford, J. H. J. Catal. 1990, 125, 123131.
66. Sievers, C.; Onda, A.; Guzman, A.; Otillinger, K. S.; Olindo, R.; Lercher, J. A.
J. Phys. Chem. C 2007, 111, 210218.
67. Budi, P.; Howe, R. F. Catal. Today 1997, 38, 175179.
68. Yan, J. Y.; Sachtler, W. M. H.; Kung, H. H. Catal. Today 1997, 33, 279290.
69. Iwasaki, M.; Shinjoh, H. Chem. Commun. 2011, 47, 39663968.
70. Liu, H.; Li, Y.; Shen, W.; Bao, X.; Xu, Y. Catal. Today 2004, 9395, 6573.
71. Tominaga, K.; Maruoka, S.; Gotoh, M.; Katada, N.; Niwa, M. Micropor. Mesopor.
Mater. 2009, 117, 523529.
72. Bosacek, V.; Klik, R.; Genoni, F.; Spano, G.; Rivetti, F.; Figueras, F. Magn. Reson.
Chem. 1999, 37, S135S141.
73. Bosacek, V. J. Phys. Chem. 1993, 97, 1073210737.
74. Cheng, C.-H.; Bae, T.-H.; McCool, B. A.; Chance, R. R.; Nair, S.; Jones, C. W.
J. Phys. Chem. C 2008, 112, 35433551.
75. Artioli, G.; Lamberti, C.; Marra, G. L. Acta Crystallogr. B 2000, 56, 210.
76. Corma, A.; Domine, M. E.; Gaona, J. A.; Navarro, M. T.; Rey, F.; Valencia, S.
In Zeolites and Mesoporous Materials at the Dawn of the 21st Century. Studies in
Surface Science and Catalysis; Proceedings of the 13th International Zeolite
Conference, Montpellier, France, 813 July, 2001; Galarneau, A., Di Renzo, F.,
Fajula, F., Eds.; Elsevier: Amsterdam, 2001; Vol. 135.
77. Corma, A.; Jorda, J. L.; Navarro, M. T.; Perez-Pariente, J.; Rey, F.; Tsuji, J.
In Nanoporous Materials II. Studies in Surface Science and Catalysis;
Proceedings of the Conference on Access in Nanoporous Materials, Banff,
Alberta, Canada, 2000; Sayari, A., Jaroniec, M., Eds.; Elsevier: Amsterdam, 2001;
Vol. 129, p 169.
78. Inagaki, S.; Fukushima, Y.; Kuroda, K. J. Chem. Soc. Chem. Commun. 1993,
680682.
79. Inagaki, S.; Koiwai, A.; Suzuki, N.; Fukushima, Y.; Kuroda, K. Bull. Chem. Soc.
Jpn. 1996, 69, 14491457.
80. Galarneau, A.; Barodawalla, A.; Pinnavaia, T. J. Nature 1995, 374, 529531.
81. Galarneau, A.; Desplantier-Giscard, D.; di Renzo, F.; Fajula, F. Catal. Today 2001,
68, 191200.
82. Cassiers, K.; Linssen, T.; Mathieu, M.; Benjelloun, M.; Schrijnemakers, K.; Van
Der Voort, P.; Cool, P.; Vansant, E. F. Chem. Mater. 2002, 14, 23172324.
83. Igarashi, N.; Koyano, K. A.; Tanaka, Y.; Nakata, S.; Hashimoto, K.; Tatsumi, T.
Micropor. Mesopor. Mater. 2003, 59, 4352.
84. Hartmann, M.; Vinu, A. Langmuir 2002, 18, 80108016.
85. Borodko, Y.; Ager, J. W. I.; Marti, G. E.; Song, H.; Niesz, K.; Somorjai, G. A.
J. Phys. Chem. 2005, 109, 1738617390.
86. Chytil, S.; Haugland, L.; Blekkan, E. A. Micropor. Mesopor. Mater. 2008, 111,
134142.
87. Herbert, R.; Wang, D.; Schomacker, R.; Schlogl, R.; Hess, C. ChemPhysChem
2009, 10, 23302333.
88. Bosacek, V.; Dubinin, M. M.; Kadlets, O.; Murdmaa, K. O.; Navratil, V. Dokl.
Chem. 1967, 174, 305.
89. Gregg, S. J.; Langford, J. F. J. Chem. Soc. Faraday Trans. 1977, 73, 747759.
90. Kisler, J. M.; Gee, M. L.; Stevens, G. W.; OConnor, A. J. Chem. Mater. 2003, 15,
619624.
91. Zhao, X. S.; Lu, G. Q. J. Phys. Chem. B 1998, 102, 15561561.
92. Van Der Voort, P.; Baltes, M.; Vansant, E. F. J. Phys. Chem. B 1999, 103,
1010210108.
93. Trong On, D.; Kaliaguine, S. J. Am. Chem. Soc. 2003, 125, 618619.
94. Trong On, D.; Kaliaguine, S. J. Am. Chem. Soc. 2004, 126, 1432414325.
95. Mokaya, R.; Jones, W. Chem. Commun. 1997, 21852186.
96. Mokaya, R. Angew. Chem. Int. Ed. 1999, 38, 29302934.
97. ONeil, A. S.; Mokaya, R.; Poliakoff, M. J. Am. Chem. Soc. 2002, 124,
1063610637.
98. Wu, S.; Han, Y.; Zou, Y.-C.; Song, J.-W.; Zhao, L.; Di, Y.; Liu, S.-Z.; Xiao, F.-S.
Chem. Mater. 2004, 16, 486492.
99. Morey, M. S.; OBrien, S.; Schwarz, S.; Stucky, G. D. Chem. Mater. 2000, 12,
898911.
100. Celer, B.; Kruk, M.; Zuzek, Y.; Jaroniek, M. J. Mater. Chem. 2006, 16,
28242833.
101. Ryoo, R.; Kim, J. M. J. Chem. Soc. Chem. Commun. 1995, 711712.
102. Ryoo, R.; Kim, J. M.; Ko, C. H.; Shin, C. H. J. Phys. Chem. B 1996, 100,
1771817721.
103. Ryoo, R.; Jun, S. J. Phys. Chem. B 1997, 101, 317320.
104. Das, D.; Tsai, C.; Cheng, S. Chem. Commun. 1999, 473474.
244 Ordered Mesoporous and Microporous Materials with Heteroatoms
105. Han, Y.; Xiao, F.-S.; Wu, S.; Sun, Y.; Meng, X.; Li, D.; Lin, S. J. Phys. Chem. B
2001, 105, 79637966.
106. Shen, S.-C.; Kawi, S. J. Phys. Chem. B 1999, 103, 88708876.
107. Mokaya, R. J. Phys. Chem. B 2000, 104, 82798286.
108. Xiao, F.-S.; Han, Y.; Yu, Y.; Meng, X.; Yang, M.; Wu, S. J. Am. Chem. Soc. 2002,
124, 888889.
109. Zhang, W.-H.; Lu, J.; Han, B.; Li, M.; Xiu, J.; Ying, P.; Li, C. Chem. Mater. 2002,
14, 34133421.
110. Lim, S.; Haller, G. L. J. Phys. Chem. B 2002, 106, 84378448.
111. Koyano, K. A.; Tatsumi, T.; Tanaka, Y.; Nakata, S. J. Phys. Chem. B 1997, 101,
94369440.
112. Van Der Voort, P.; Baltes, M.; Vansant, E. F. Catal. Today 2001, 68, 119128.
113. Mokaya, R. J. Phys. Chem. B 1999, 103, 1020410208.
114. Kruk, M.; Jaroniec, M.; Sayari, A. Micropor. Mesopor. Mater. 1999, 27,
217229.
115. Kim, J. M.; Jun, S.; Ryoo, R. J. Phys. Chem. B 1999, 103, 62006205.
116. Chen, L.; Horiuchi, T.; Mori, T.; Maeda, K. J. Phys. Chem. B 1999, 103,
12161222.
117. Huo, Q.; Margolese, D. I.; Stucky, G. D. Chem. Mater. 1996, 8, 11471160.
118. Kruk, M.; Celer, E. B.; Jaroniec, M. Chem. Mater. 2004, 16, 698707.
119. Luan, Z.; Maes, E. M.; van der Heide, P. A. W.; Zhao, D.; Czernuszewicz, R. S.;
Kevan, L. Chem. Mater. 1999, 11, 36803686.
120. Luan, Z.; Hartmann, M.; Zhao, D.; Zhou, W.; Kevan, L. Chem. Mater. 1999, 11,
16211627.
121. Luan, Z.; Bae, J. Y.; Kevan, C. Chem. Mater. 2000, 12, 32023207.
122. Hoffmann, F.; Cornelius, M.; Morell, J.; Froba, M. Angew. Chem. Int. Ed. 2006,
45, 32163251.
123. Hoffmann, F.; Cornelius, M.; Morell, J.; Froba, M. J. Nanosci. Nanotechnol.
2006, 6, 265288.
124. Hunks, W. J.; Ozin, G. A. J. Mater. Chem. 2005, 15, 37163724.
125. Slowing, A.; Vivero-Escoto, J. L.; Trewyn, B. G.; Lin, S.-Y. J. Chem. Mater. 2010,
20, 79247937.
126. Yang, Q.; Li, Y.; Zhang, L.; Yang, H.; Liu, J.; Li, C. J. Phys. Chem. B 2004, 108,
79347937.
127. Kapoor, M. P.; Sinha, A. K.; Seelan, S.; Inagaki, S.; Tsubota, S.; Yoshid, H.;
Haruta, M. Chem. Commun. 2002, 29022903.
128. Cho, W.; Park, J.-W.; Ha, C.-S. Mater. Lett. 2004, 58, 35513554.
129. Melero, J. A.; Iglesias, J.; Arsuaga, J. M.; Sainz-Pardo, J.; de Frutos, P.;
Blazquez, S. J. Mater. Chem. 2007, 17, 377385.
130. Park, M.; Komarneni, S. Micropor. Mesopor. Mater. 1998, 25, 7580.
131. de Juan, F.; Ruiz-Hitzky, E. Adv. Mater. 2000, 12, 430432.
132. Xu, R.; Pang, W.; Yu, J.; Huo, Q.; Chen, J. In Chemistry of Zeolites and Related
Porous Materials: Synthesis and Structure John Wiley & Sons: Singapore, 2007
Chapter 6, pp 345395.
133. Ratnasamy, P.; Srinivas, D. Catal. Today 2009, 141, 311.
134. Viswanathan, B.; Jacob, B. Catal. Rev. 2005, 47, 182.
135. Reddy, K. M.; Moudrakovski, I.; Sayari, A. Chem. Commun. 1994, 14911492.
136. Sen, T.; Rajamohanan, P. R.; Ganapathy, S.; Sivasanker, S. J. Catal. 1996, 163,
354364.
137. Sen, T.; Ramaswamy, V.; Ganapathy, S.; Rajamohanan, P. R.; Sivasanker, S. I.
J. Phys. Chem. 1996, 100, 38093817.
138. Centi, G.; Perathoner, S.; Trifiro, F.; Aboukais, A.; Aissi, C. F.; Guelton, M. P.
J. Phys. Chem. 1992, 96, 26172629.
139. Zhang, S. G.; Higashimoto, S.; Yamashita, H.; Anpo, M. J. Phys. Chem. B 1998,
102, 55905594.
140. Higashimoto, S.; Matsuoka, M.; Yamashita, H.; Anpo, M.; Kitao, O.; Hidaka, H.;
Che, M.; Giamello, E. J. Phys. Chem. B 2000, 104, 1028810292.
141. Catana, G.; Rao, R. R.; Weckhuysen, B. M.; Van Der Voort, P.; Vansant, E.;
Schoonheydt, R. A. J. Phys. Chem. 1998, 102, 80058012.
142. Kim, G. J.; Cho, D. S.; Kim, L. H.; Ko, W. S.; Kim, J. H.; Shoji, H. Catal. Lett.
1995, 31, 91102.
143. Whittington, B. I.; Anderson, J. R. J. Phys. Chem. 1991, 95, 33063310.
144. Komorek, J.; Romotowski, T.; Serwicka, E. M.; Mastikhin, V. M. Zeolites 1994,
14, 629634.
145. Dzwigaj, S.; Peltre, M. J.; Massiani, P.; Davidson, A.; Che, M.; Sen, T.;
Sivasanker, S. J. Chem. Soc. Chem. Commun. 1998, 8788.
146. Lim, S.; Haller, G. L. Appl. Catal. A 1999, 188, 277286.
147. Luan, Z.; Meloni, P. A.; Czernuszewicz, R. S.; Kevan, L. J. Phys. Chem. B 1997,
101, 90469051.
148. Berndt, H.; Martin, A.; Bruckner, A.; Schreier, E.; Muller, D.; Kosslick, H.;
Wolf, G.-U.; Lucke, B. J. Catal. 2000, 191, 384400.
149. Fornes, V.; Lopez, C.; Lopez, H. H.; Martinez, A. Appl. Catal. A 2003, 249,
345354.
150. Liu, Y.-M.; Cao, Y.; Yi, N.; Feng, W.-L.; Dai, W.-L.; Yan, S.-R.; He, H.-Y.;
Fan, K.-N. J. Catal. 2004, 224, 417428.
151. Santamara-Gonzalez, J.; Luque-Zambrana, J.; Merida-Robles, J.;
Maireles-Torres, P.; Rodrguez-Castellon, E.; Jimenez-Lopez, A. Catal. Lett.
2000, 68, 6773.
152. Chlosta, R.; Tzolova-Muller, G.; Schlogl, R.; Hess, C. Catal. Sci. Technol. 2011,
1, 11751181.
153. Bronkema, J. L.; Bell, A. T. J. Phys. Chem. C 2007, 111, 420430.
154. Van Der Voort, P.; Morey, M.; Stucky, G. D.; Mathieu, M.; Vansant, E. F. J. Phys.
Chem. B 1998, 102, 585590.
155. Baltes, M.; Cassiers, K.; Van Der Voort, P.; Weckhuysen, B. M.;
Schoonheydt, R. A.; Vansant, E. F. J. Catal. 2001, 197, 160171.
156. Kustrowski, P.; Segura, Y.; Chmielarz, L.; Surman, J.; Dziembaj, R.; Cool, P.;
Vansant, E. F. Catal. Today 2006, 114, 307313.
157. Zhang, Q.; Wang, Y.; Ohishi, Y.; Shishido, T.; Takehira, K. J. Catal. 2001, 202,
308318.
158. Hess, C.; Hoefelmeyer, J. D.; Tilley, T. D. J. Phys. Chem. B 2004, 108,
97039709.
159. Hess, C.; Wild, U.; Schlogl, R. Micropor. Mesopor. Mater. 2006, 95, 339349.
160. Herrera, J. E.; Kwak, J. H.; Hu, J. Z.; Wang, Y.; Peden, C. H. F. Top. Catal. 2006,
39, 245255.
161. Segura, Y.; Chmielarz, L.; Kstrowski, P.; Cool, P.; Dziembaj, R.; Vasant, E. F.
J. Phys. Chem. B 2006, 110, 948955.
162. Zhuravlev, L. T. Langmuir 1987, 3, 316318.
163. Zhuravlev, L. T. Colloids Surf. A: Physicochem. Eng. Asp. 1993, 74, 7190.
164. Widenmeyer, M.; Anwander, R. Chem. Mater. 2002, 14, 18271831.
165. Zhao, X. S.; Lu, G. Q.; Whittaker, A. K.; Millar, G. J.; Zhu, H. Y. J. Phys. Chem. B
1997, 101, 65256531.
166. Shenderovich, I. G.; Buntkowsky, G.; Schreiber, A.; Gedat, E.; Sharif, S.;
Albrecht, J.; Golubev, N. S.; Findenegg, G. H.; Limbach, H.-H. J. Phys. Chem. B
2003, 107, 1192411939.
167. Jarapatrakorn, J.; Tilley, T. D. J. Am. Chem. Soc. 2002, 124, 83808388.
168. Ottaviani, M. F.; Galarneau, A.; Desplantiers-Giscard, D.; Di Renzo, F.; Fajula, F.
Micropor. Mesopor. Mater. 2001, 4445, 18.
169. Lopez Nieto, J. M. Top. Catal. 2001, 15, 189194.
170. Lopez Nieto, J. M. Top. Catal. 2006, 41, 315.
171. Bellussi, G.; Rigutto, M. S. Stud. Surf. Sci. Catal. 1994, 85, 77213.
172. Dzwigaj, S. Curr. Opin. Solid State Mater. Sci. 2003, 7, 461470.
173. Dzwigaj, S.; Che, M. Catal. Today 2011, 169, 232241.
174. Dzwigaj, S.; Matsuoka, M.; Anpo, M.; Che, M. J. Phys. Chem. B 2000, 104,
60126020.
175. Dzwigaj, S.; Matsuoka, M.; Franck, R.; Anpo, M.; Che, M. J. Phys. Chem. B 1998,
102, 63096312.
176. Lacheen, H. S.; Iglesia, E. J. Phys. Chem. B 2006, 110, 54625472.
177. Keranen, J.; Guimond, C.; Iiskola, E.; Auroux, A.; Niinisto, L. J. Phys. Chem. B
2003, 107, 1077310784.
178. Gao, X.; Bare, S. R.; Weckhuysen, B. M.; Wachs, I. E. J. Phys. Chem. B 1998,
102, 1084210852.
179. Hess, C.; Schlogl, R. Chem. Phys. Lett. 2006, 432, 139145.
180. Hess, C.; Tzolova-Muller, G.; Herbert, R. J. Phys. Chem. C 2007, 111,
94719479.
181. Hess, C. J. Catal. 2007, 248, 120123.
182. Venkov, T. V.; Hess, C.; Jentoft, F. C. Langmuir 2007, 23, 17681777.
183. Walter, A.; Herbert, R.; Hess, C.; Ressler, T. Chem. Central J. 2010, 4, 3.
184. Cavalleri, M.; Hermann, K.; Knop-Gericke, A.; Havecker, M.; Herbert, R.; Hess, C.;
Oestereich, A.; Dobler, J.; Schlogl, R. J. Catal. 2009, 262, 215223.
185. Schlogl, R.; Hess, C. In Nanostructured Catalysts: Selective Oxidations; Hess, C.,
Schlogl, R., Eds.; RSC: Cambridge, 2011 Chapter 15; pp 355397.
186. Dinse, A.; Ozarowski, A.; Hess, C.; Schomacker, R.; Dinse, K.-P. J. Phys. Chem. C
2008, 112, 1766417671.
187. Higashimoto, S.; Hu, R.; Tsumura, Y.; Iino, K.; Matsuoka, M.; Yamashita, H.;
Shul, Y. G.; Che, M.; Anpo, M. J. Catal. 2005, 235, 272278.
188. Wang, L.; Tao, L.; Xie, M.; Xu, G. Catal. Lett. 1993, 21, 3541.
189. Solymosi, F.; Erdohelyi, A.; Szoke, A. Catal. Lett. 1995, 32, 4353.
190. Wang, D.; Lunsford, J. H.; Rosynek, M. P. Top. Catal. 1996, 3, 289297.
191. Liu, S.; Dong, Q.; Ohnishi, R.; Ichikawa, M. J. Chem. Soc. Chem. Commun.
1997, 14551456.
192. Borry, R. W.; Kim, Y. H.; Huffsmith, A.; Reimer, J. A.; Iglesia, E. J. Phys. Chem. B
1999, 103, 57875796.
193. Ding, W.; Li, S.; Meitzner, G. D.; Iglesia, E. J. Phys. Chem. B 2001, 105,
506613.
194. Melero, J. A.; Iglesias, J.; Arsuaga, J. M.; Sainz-Pardo, J.; de Frutos, P.;
Blazquez, S. Appl. Catal. A 2007, 331, 8494.
 Ordered Mesoporous and Microporous Materials with Heteroatoms
195. Arnold, U.; da Cruz, R. S.; Mandelli, D.; Schuchardt, U. J. Mol. Catal. A 2001,
165, 149158.
196. Briot, E.; Piquemal, J.-Y.; Bregeault, J.-M. New J. Chem. 2002, 26, 14431447.
197. Huang, Z.; Bensch, W.; Sigle, W.; van Aken, P. A.; Kienle, L.; Vitoya, T.;
Modrow, H.; Ressler, T. J. Mater. Sci. 2008, 43, 244253.
198. Dhar, G. M.; Kumaran, G. M.; Kumar, M.; Rawat, K. S.; Sharma, L. D.; Raju, B. D.;
Rao, K. S. R. Catal. Today 2005, 99, 309314.
199. Lou, Y.; Wang, H.; Zhang, Q.; Wang, Y. J. Catal. 2007, 247, 245255.
200. Sampieri, A.; Pronier, S.; Blanchard, J.; Breysse, M.; Brunet, S.; Fajerwerg, K.;
Louis, C.; Perot, G. Catal. Today 2005, 537, 107108.
201. Vradman, L.; Landau, M. V.; Herskowitz, M.; Ezersky, V.; Talianker, M.;
Nikitenko, S.; Koltypin, Y.; Gedanken, A. Stud. Surf. Sci. Catal. 2003, 146, 721.
202. Thielemann, J. P.; Weinberg, G.; Hess, C. ChemCatChem 2011, 3, 18141821.
203. Topka, P.; Balcar, H.; Rathousky, J.; Zilkova, N.; Verpoort, F.; Cejka, J. Micropor.
Mesopor. Mater. 2006, 96, 4454.
204. Thielemann, J. P.; Krohnert, J.; Hess, C. J. Phys. Chem. C 2010, 114, 17092.
205. Thielemann, J. P.; Ressler, T.; Walter, A.; Tzolova-Muller, G.; Hess, C. Appl. Catal.
A 2011, 399, 2834.
206. Ruddy, D. A.; Ohler, N. L.; Bell, A. T.; Tilley, T. D. J. Catal. 2006, 238, 277285.
207. Jarupatrakorn, J.; Coles, M. P.; Tilley, T. D. Chem. Mater. 2005, 17,
18181828.
208. Shylesh, S.; Singh, A. P. Micropor. Mesopor. Mater. 2006, 94, 127138.
209. Hornebecq, V.; Mastai, Y.; Antonietti, M.; Polarz, S. Chem. Mater. 2003, 15,
35863593.
210. Carreon, M. A.; Guliants, V. V. Eur. J. Inorg. Chem. 2005, 2743.
211. Carreon, M. A.; Guliants, V. V. In Nanostructured Catalysts: Selective
Oxidations; Hess, C., Schlogl, R., Eds.; RSC: Cambridge, 2011 Chapter 6.1; pp
141168.
212. Segura, Y.; Chmielarz, L.; Kustrowski, P.; Cool, P.; Dziembaj, R.; Vansant, E. F.
J. Phys. Chem. B 2006, 110, 948955.
213. Chao, Z.-S.; Ruckenstein, E. Langmuir 2002, 18, 734743.
214. He, X.; Lo, A. Y. H.; Trudeau, M.; Schurko, R. W.; Antonelli, D. Inorg. Chem.
2003, 42, 335347.
215. Hess, C.; Drake, I. J.; Hoefelmeyer, J. D.; Tilley, T. D.; Bell, A. T. Catal. Lett. 2005,
105, 18.
216. Trejda, M.; Ziolek, M.; Millot, Y.; Chalupka, K.; Che, M.; Dzwigaj, S. J. Catal.
2011, 281, 169176.
217. Hess, C. Surf. Sci. 2006, 600, 36953701.
218. Wang, C. B.; Deo, G.; Wachs, I. E. J. Catal. 1998, 178, 640648.
219. Kwak, J. H.; Herrera, J. E.; Hu, J. Z.; Wang, Y.; Peden, C. H. F. Appl. Catal. A
2006, 300, 109119.
220. Herman, R. G.; Sun, Q.; Shi, C.; Klier, K.; Wang, C.-B.; Hu, H.; Wachs, I. E.;
Bhasin, M. M. Catal. Today 1997, 37, 114.
245
221. Lin, B.; Wang, X.; Guo, Q.; Yang, W.; Zhang, Q.; Wang, Y. Chem. Lett. 2003,
32, 860.
222. Hguyen, L. D.; Loridant, S.; Launay, H.; Pigamo, A.; Dubois, J. L.; Millet, J. M. M.
J. Catal. 2006, 237, 3848.
223. Dai, L.-X.; Teng, Y. H.; Tabata, K.; Suzuki, E.; Tatsumi, T. Micropor. Mesopor.
Mater. 2001, 4445, 573580.
224. Kucherov, A. V.; Ivanov, A. V.; Kucherova, T. N.; Nissenbaum, V. D.; Kustov, L. M.
Catal. Today 2003, 81, 297305.
225. Solsona, B.; Blasco, T.; Nieto, J. M. L.; Pena, M. L.; Rey, F.; Vidal-Moya, A.
J. Catal. 2001, 203, 443452.
226. Pena, M. L.; Dejoz, A.; Fornes, V.; Rey, F.; Vazquez, M. I.; Lopez Nieto, J. M. Appl.
Catal. A 2001, 209, 155164.
227. Kilos, B.; Nowak, I.; Ziolek, M.; Tuel, A.; Volta, J. C. Stud. Surf. Sci. Catal. 2005,
158, 14611468.
228. Bruckner, A.; Rybarczyk, P.; Kosslick, H.; Wolf, G.-U.; Baerns, M. Stud. Surf. Sci.
Catal. 2002, 142, 11411148.
229. Kondratenko, E. V.; Cherian, M.; Baerns, M.; Su, D.; Schlogl, R.; Wang, X.;
Wachs, I. E. J. Catal. 2005, 234, 131142.
230. Grune, P.; Wolfram, T.; Pelzer, K.; Schlogl, R.; Trunschke, A. Catal. Today 2010,
157, 137142.
231. Ovsitser, O.; Cherian, M.; Kondratenko, E. V. J. Phys. Chem. C 2007, 111,
85948602.
232. Dinse, A.; Khennache, S.; Frank, B.; Hess, C.; Herbert, R.; Wrabetz, S.; Schlogl, R.;
Schomacker, R. J. Mol. Catal. A 2009, 307, 4350.
233. Centi, G.; Trifiro, F. Appl. Catal. A 1996, 143, 316.
234. Blasco, T.; Concepcion, P.; Lopez Nieto, J. M.; Perez-Pariente, J. Appl. Catal. A
1996, 143, 1728.
235. Dzwigaj, S.; Gressel, I.; Grzybowska, B.; Samson, K. Catal. Today 2006, 114,
237241.
236. Thielemann, J. P.; Hess, C. J. Catal. 2012, 288, 124126.
237. Ressler, T.; Walter, A.; Huang, Z.-D.; Bensch, W. J. Catal. 2008, 254,
170179.
238. Thielemann, J.P.; Hess, C. Unpublished results.
239. Liu, W.; Lai, S. Y.; Dai, H.; Wang, S.; Sun, H.; Au, C. T. Catal. Lett. 2007, 113,
147154.
240. Li, X. K.; Ji, W. J.; Zhao, J.; Zhang, Z.; Au, C. T. Appl. Catal. A 2006,
306, 816.
241. Li, X. K.; Ji, W. J.; Zhao, J.; Zhang, Z.; Au, C. T. Appl. Catal. A 2006, 244,
265272.
242. Nie, W. Y.; Wang, Z. Y.; Ji, W. J.; Chen, Y.; Au, C. T. J. Catal. 2006, 238,
232241.
243. Thielemann, J. P.; Hess, C. ChemPhysChem, http://dx.doi.org/10.1002/
cphc.201200648.
 5.06 Mesoporous Silica
T Kimura, National Institute of Advanced Industrial Science and Technology (AIST), Nagoya, Japan
K Kuroda, Waseda University, Tokyo, Japan
2013 Elsevier Ltd. All rights reserved.
5.06.1 Introduction
5.06.2 Preparation of Mesoporous Materials Using Surfactants
5.06.3 Structural Characterization of Mesoporous Silica
5.06.4 Morphological Control of Periodic Mesoporous Silica
5.06.5 Applications Using Periodic Mesoporous Silica
5.06.6 Conclusion
References
5.06.1 Introduction
Porous materials are categorized on the basis of pore size. Meso-
porous materials possess a large number of pores ranging from 2
to 50 nm in size. Periodic mesostructures of various composi-
tions, including inorganic oxides and metal phosphates, are
successfully replicated by using amphiphilic organic molecules
possessing a self-assembling property. It is essential for the
preparation of periodic mesoporous materials to generate the
interactions between soluble inorganic species and amphiphilic
organic molecules.1,2 Amphiphilic organic molecules, whose
hydrophilic headgroups are interacted with inorganic species,
are self-assembled with gradual condensation of inorganic spe-
cies, being called as cooperative organization for the formation
of periodic mesostructures.3,4 Materials with uniform meso-
pores can then be prepared after the removal of surfactant
assemblies. Such periodic mesoporous materials have been
expected as reaction vessels showing high selectivity for rela-
tively large organic molecules and thus increasingly and widely
investigated.
Through the investigations on the reactions between sur-
factants and layered polysilicate, the formation of mesostruc-
tured silica and importance of surfactant assemblies on the
formation of uniform mesopores were disclosed.5,6 Mobil
researchers reported the simple hydrothermal synthesis of
silica-based mesoporous materials reflecting the liquid-crystal
phases of surfactants.7,8 Usefulness of the self-assembling abil-
ity of surfactants for the mesostructural control of materials has
been indicated by beautiful transmission electron microscopic
(TEM) images showing periodic mesopore arrangements.
Representative TEM images of two-dimensional (2D) hexago-
nal mesoporous silica (MCM-41) with different pore diame-
ters are shown in Figure 1. Different periodic mesostructures
(cubic MCM-48, lamellar MCM-50, etc.) in silica were also
reported,919 and then, the surfactant-assisted synthesis of
materials and their applications has been surveyed all over
the world, and there have been a huge number of papers
describing compositional and morphological controls of
periodic mesoporous materials, which has resulted in many
strong expectations on practical applications not only due
to porous features for adsorbents and catalysts but also due
to electronic, magnetic, and optical properties for advanced
devices.
Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00507-6
133
133
136
142
144
147
149
5.06.2 Preparation of Mesoporous Materials
Using Surfactants
Initially, the preparative methods of mesoporous materials
using surfactants were mainly focused on silica-based
materials.20 Periodic mesoporous materials can generally be
prepared by the addition of an inorganic source to solutions
containing surfactants. When the kind of surfactants is appro-
priately selected as cationic, anionic, or nonionic one, the
interactions between inorganic and organic species are gener-
ated in solutions. Mesoporous silicas with a wide variety of
periodicities have been synthesized, and each material is
named by using three alphabetical codes, such as FSM, MCM,
and SBA. Representative periodic mesoporous silicas are sum-
marized in Figure 2. Actual synthetic procedures should be
referred to the contents of the original papers listed in this
section.
MCM-41, MCM-48, and MCM-50 are prepared using cat-
ionic alkyltrimethylammonium (CnTMA) surfactants.7,8 All
the mesostructures reflect the liquid-crystal phases of CnTMA
surfactants in the concentrated solutions. They are synthesized
by mixing silica and alumina sources with aqueous solutions
of surfactants under basic conditions. In addition, cage-type
mesoporous silicas are generally obtained using cationic sur-
factants with large headgroups, such as alkyltriethylammo-
nium (CnTEA) and gemini-type diammonium surfactants,
under acidic conditions.10,11 Formation of the mesostructures
is interpreted by using packing parameters (g V/a0l), which is
calculated by using the total volume of surfactant chains (V)
and kinetic length of hydrophobic alkyl chain (l) and effective
area of hydrophilic headgroup (a0) of the surfactant molecules.
Thus, judging from the molecular structures of surfactants
used, the structural ordering of formed mesoporous silica is
predictable to some extent. The validity of the aforementioned
explanation using the packing parameters is demonstrated for
some mesoporous silicas having sufficient structural informa-
tion. However, it is well known that the mesostructure varies
according to not only the packing parameter of surfactant but
also the reaction temperature, reaction time, pH of precursor
solution, and hydrothermal posttreatment. Accordingly, it is
crucial for the precise control of formed mesostructures to
take an account of the packing parameters affected due to the
presence of inorganic species interacting with hydrophilic
133
134 Mesoporous Silica

headgroups of surfactant geometrically. In addition, the size
of the hydrophilic part is easily varied according to the con-
densation degree of silicate species, possibly leading to the
variation of resultant mesostructure. Phase transformation
between mesostructured silicas is a good example of the valid-
ity to consider the packing parameters containing attached
inorganic parts.3,21
Neutral and nonionic surfactants, such as alkylamine (CnA)
and poly(oxyethylene) alkyl ether (CnEOm), are also useful
for obtaining mesostructured materials through hydrogen bond-
ing between silicate species and the headgroups of the sur-
factants.1219 Mesostructural orderings are very low in the
beginning, but the ordering is improved according to the syn-
thetic conditions selected appropriately for controlling the inter-
actions from hydrogen bonding to ionic ones. Nonionic
10 nm
(a) (b)
(c) (d)
Figure 1 TEM images of MCM-41 with pore diameters of (a) 2.0 nm,
(b) 4.0 nm, (c) 6.5 nm, and (d) 10 nm. Reproduced from Figure 2 in Beck,
J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt,
K. D.; Chu, C. T.-W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.;
Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834.
surfactants are very important because they are easily extracted
in solvents, recycled, and reused. In particular, CnEOm and
poly(oxyethylene)-block-poly(oxypropylene)-block-poly(oxyethy-
lene) triblock copolymer (EOnPOmEOn) are low cost, non-
toxic, and biodegradable. In those cases, because of the week
interactions between silicate species and the EO moieties, thick
silica walls surround mesopores, and therefore, mesoporous
silica, possessing enough thermal, hydrothermal, and mechanical
stabilities, is obtained. As the most special case, MSU-G synthe-
sized using gemini-type alkyl amines is ultrastable due to
extremely high condensation degree in the silicate frameworks
(Q4/Q3  7),16 being related to the relatively low specific
surface areas. As shown in Figure 3, vesicular MSU-G has an
intermediate mesostructure between lamellar (La) and lyotropic
L3 phase and thus would be formed by phase transition from the
La phase. The mesopore size of calcined MSU-G is controlled by
changing the alkyl chain length of the surfactants. However, it was
reported that surfactant assembly-mediated porous silica was
already replicated by using lyotropic L3 phase of hexadecylpyr-
idinium chloride (C16PyCl).22 Interestingly, continuous pore size
control is possible in a wide range of 635 nm by changing
the volume fraction of solvent (water) without the removal of
C16PyCl (see Figure 3(c)).
Considering the interactions between inorganic and
organic species, it is impossible to synthesize mesoporous
silica using anionic surfactants because both silicate species
and surfactants are charged as anionic under both acidic and
basic conditions. Nevertheless, a synthetic strategy of periodic
mesoporous silica even using anionic surfactants has also been
developed by using another interaction of silicate species with
amino and ammonium groups in organosilanes as extra silica
sources.23,24 Because the possibility to use general anionic
surfactants is proposed for the preparation of periodic meso-
porous silica (AMS-n), amino acid-type surfactants are also
used, and the method is occasionally directed to the creation
of silica with modulated and chiral mesopore arrangements.25
Although it is explained that chiral and/or helical mesopores
are formed on the basis of the molecular structures with

Lamellar
Disorder
p6mm (KSW-L)
MCM-50 KSW-1
FSM-16 AMS-3
AMS-5
MCM-41
Ia-3d Wormhole
SBA-3 SBA-15
c2mm AMS-6
KSW-2 Im-3m KIT-6 HMS KIT-1
MCM-48
SBA-8 SBA-16 MSU-n
FDU-5
FDU-1 P63/mmc
P42/mnm
Fd-3m SBA-2 AMS-1
AMS-9
AMS-8 SBA-11
P4/mmm Pmmm FDU-2 Pm-3n
FDU-13 Fm-3m
FDU-11
Pn-3m AMS-2
SBA-12 KIT-5
SBA-6
AMS-10 FDU-12
SBA-1

Figure 2 Mesostructures and abbreviations of periodic mesoporous silicas.


d d
d
L phase Intermediate
(a)
TEOS
H2O
(b) Gemini L MSU-G mesostructure
Figure 3 (a) Phase transition from La to L3 phases of surfactant and (b) related formation mechanism of mesostructured precursor of MCU-G ((a and
b) reproduced from Figure 1 in Kim, S. S.; Zhang, W.; Pinnavaia, T. J. Science 1998, 282, 1302) with (c) schematic structure at the interface of
silicate frameworks (reproduced from Figure 1 in McGrath, K. M.; Dabbs, D. M.; Yao, N.; Aksay, I. A.; Gruner, S. M. Science 1995, 269, 1242).
asymmetric carbons in headgroups of anionic surfactants, it is
proposed that cylindrical mesopores are finally twisted by
transition from 2D hexagonal to helical structures in order to
minimize the surface free energy. As a new formation mecha-
nism of chiral and/or helical mesopores, the curvature of the
helical mesopores is suggested to be defined by the increase of
the energy to organize the curved surfaces.26 On the basis of the
same route to obtain AMS-type mesoporous silica, 2D square
(p4mm) symmetry is also organized using anionic DNA with a
double-helix structure.27
The preparation of 2D hexagonal (MCM-41, FSM-16) and
3D cubic mesoporous silicas (MCM-48)79 was the first to
discuss the relation between molecular structures of surfactants
and mesostructures in various ways.10,11 It was reported that
3D hexagonal (P63/mcm) mesoporous silica (IBN-9) with a
tricontinuous mesopore network was obtained using a unique
surfactant, such as (S)-(1-tetradecylcarbomyl-2-phenylethyl)
dimethyethylammonium bromide.28 Silica pore walls in the
mesostructure, which follow a hexagonal minimal surface, are
analyzed mathematically as a mesophase between bicontinu-
ous mesopore network of MCM-48 and 2-D hexagonal meso-
pore arrangement of MCM-41.
The synthetic methods from layered polysilicates (mainly
kanemite composed of single silicate sheets) are exceptionally
interesting,5,6,9,2933 and unusual mesoporous silica named as
KSW-2 can be synthesized using common CnTMA surfactants.32
Square-shape mesopores in the mesostructure (c2mm) has never
been found in periodic mesoporous silicas formed through
cooperative organization of surfactant molecules attaching to
silicate species. As shown in Figure 4, formed mesostructures
from layered polysilicates are simply understood by illustrating
possible surfactant assemblies in 2D limited spaces provided
from layered polysilicates.34 Lamellar, rodlike, and spherical
assemblies of surfactants are considered to be present in the
limited space. In general, lamellar phases are preferentially
formed by restriction of the sheet structure. However, it is
Mesoporous Silica 135
d
L3 phase
Water
Silicate
Silicate
Water
(c)
possible to accommodate rodlike micelles in mesostructures
(FSM-16 and KSW-2) formed under the conditions that allow
fragmentation and bending of silicate sheets, respectively. It is
quite difficult to wrap spherical micelles with flat silicate sheets,
leading to the formation of disordered mesopore arrangements.
Structural design from different layered polysilicates3537 has
also been developed so far with incorporation of metal species
in the silicate frameworks.3840 The periodicity of the original
layered polysilicates can be retained in the silicate frameworks by
surface modification of a mesostructured precursor of KSW-2-
type silica using organosilanes with appropriate molecular
structures.41 Previously synthesized mesoporous silicas are con-
structed by amorphous frameworks, being strongly related that
metal species doped in the mesoporous silicas do not generally
show high catalytic properties. Accordingly, KSW-2-based meso-
porous silica having periodic species in the silicate frameworks
has been investigated and expected as high-performance
catalysts.42
Facile pore size control of mesoporous silica is one of the
significant features of the synthetic method using surfactants,
which is quite important for applications of mesoporous silicas
as catalysts and catalytic supports. The pore diameter is variable
from 1.5 to 10 nm by changing alkyl chain length of surfac-
tants with auxiliary organic additives.7,8 In the synthesis using
EOmPOnEOm, it is reported that the pore diameter of periodic
mesoporous silica can be expanded up to 30 nm.18,19 Foamlike
mesoporous silica (the average pore diameter; 30 nm), named
as mesocellular foam (MCF), is also obtained from diluted
acidic aqueous solutions of (EO)20(PO)70(EO)20 (Pluronic
P123) containing trimethylbenzene (TMB) as an organic
additive.43 Without the addition of organic additives, silica
with relatively uniform pores (100 nm) can be prepared in
a buffer solution (pH 5, 35  C) containing Pluronic P123.44
Siliceous vesicles are regularly deposited, being similar to the
formation mechanism of diatoms (biosilica) under the mild
acidic conditions. This kind of researches on the formation of
136 Mesoporous Silica
Possible assemblies
C16316
Lamellar
2D hexagonal
C16516
2D Hexagonal
(Lamellar)
Surface curvature
C16TMA
Lamellar
cubic (la-3d)
2D Hexagonal
(Intralayer
C16TEA
Cubic (Pm-3n)
C1631
3D Hexagonal
(P63Immc)
Formation route of ordered phases
Formation route of disordered phases
Figure 4 Formation mechanisms of mesostructured materials derived by the reactions between layered polysilicates and alkylammonium-type
surfactants. Reproduced from Scheme 2 in Kimura, T.; Itoh, D.; Shigeno, T.; Kuroda, K. Bull. Chem. Soc. Jpn. 2004, 77, 585.
biosilica and the morphological variation is also important for
further understanding biomineralization in the presence of
peptides as structure-directing agents.4549 Otherwise, poly
(isoprene)-block-poly(oxy ethylene) (PI-b-PEO)50 and ionic
poly(ethyl ethylene)-block-poly(styrene sulfonate)51,52 are uti-
lized for the synthesis of silica with extremely large pores. In
another way, spherical mesopores are preferentially formed
using poly(styrene)-block-poly(oxyethylene) (PS-b-PEO).5357
Although inorganicorganic composite spheres are well orga-
nized, the spherical pores are isolated and/or connected
through micropores after complete removal of PS-b-PEO by
calcination. Accordingly, appropriate applications, excluding
catalysts and adsorbents, should be found on the basis of the
unique porous structures.
In the synthesis of periodic mesoporous silica using PI-b-
PEO, disassembly of the mesostructures was also investigated
as one of the interesting methods to fabricate new nanobuilding
units, as shown in Figure 5, which is beyond the common
discussion that nanosized composites of silicate species and
PI-b-PEO molecules are assembled for the formation of
mesostructures.58 Advanced synthetic methods of porous silica
using a wide variety of organic compounds have been proposed
for further structural controls. In the syntheses using common
triblock copolymers, the copolymers behave as amphiphilic
molecules because they consist of hydrophobic and hydrophilic
units. Then, such copolymers whose hydrophilic moieties are
interacted with silicate species are cooperatively self-assembled
with condensation of silicate species, affording precursors for
periodic mesoporous silica. In contrast, the utilization of hydro-
philichydrophilic diblock copolymers, such as poly
(ethylene oxide)-block-poly(methacrylic acid) (PEO-b-PMAA)
and poly(ethylene oxide)-block-poly(acrylic acid) (PEO-b-PAA),
Within 2D Limited space
Lamellar 2D Hexagonal
(FSM-16)
Transformation
(Fragmentation)
condenzation) Bending
(Less condensed)
2D Orthorhombic
(KSW-2)
Disordered phase
was proposed as an ecofriendly synthetic route.59 In the pres-
ence of oligochitosan lactate as a weak acid source, PEO-b-PMAA
and PEO-b-PAA were self-assembled and dissociated reversibly
by changing pH only. Accordingly, they can be recovered and
reused for the synthesis of mesoporous silica. Organic molecules
for the preparation of mesoporous silica have not been limited
to surfactants; supramolecular assemblies induced by pp inter-
action of organic molecules are useful for obtaining periodic
mesoporous silica.60,61 The synthetic method is very important
for the arrangement and alignment controls of functional
organic molecules in silica matrix, rather than for the formation
of periodic structures.
The synthesis using oil-in-water emulsions stabilized by
surfactants led to the formation of hierarchically porous struc-
tures with pores larger than 50 nm that were surrounded by
surfactant-templated mesopores.62,63 Hollow nanospheres of
silica are also prepared using poly(diisopropylamino)ethyl
methacrylate-block-poly(dimethylamino)ethyl methacrylate
(PDPA-b-PDMA).64 Poly(N-isopropylacrylamide) (PNIPAAm)
was useful for the formation of micrometer-scaled silica hol-
low spheres, and ordered mesopores can be introduced in the
silica walls using C16TMABr.65 In this case, the hydrophilicity
and hydrophobicity of the product can easily be controlled by
the presence of thermosensitive PNIPAAm.
5.06.3 Structural Characterization
of Mesoporous Silica
x-Ray diffraction (XRD) measurements, TEM observations,
nitrogen adsorption measurements, etc., are useful for struc-
tural characterization of periodic mesoporous materials.

Figure 5 Assembly and disassembly of inorganicorganic hydride mesostructures. Reproduced from Figure 1 in Warren, S. C.; Disalvo, F. J.; Wiesner,
U.; Nat. Mater. 2007, 6, 156.
Structural information of inorganic frameworks can be col-
lected by using several characterization techniques. Typical
analytical results by using XRD, TEM, and N2 sorption are
introduced here.
In the typical XRD patterns of periodic mesoporous mate-
rials, diffraction peaks related to the presence of mesostructural
orderings are observed in low 2y regions, while such diffraction
peaks are not observed in the scattering angles that are used for
characterization of conventional crystal structures. Such diffrac-
tion peaks are generally broadened with lowering in the mesos-
tructural orderings, and then, only one diffraction peak is
observed in low 2y regions for materials named as disordered
and wormhole-like mesoporous materials.12,13,17 However,
even when diffraction peaks seem to be broad, such materials
should be evaluated carefully, because a possibility to contain
several diffraction peaks located very closely in low 2y regions
cannot be excluded. Representative XRD patterns of periodic
mesoporous silicas are shown in Figures 6 and 7. Depending
on each mesostructural ordering, corresponding XRD patterns
are observed in low scattering angles.9,11,30,32,66,67 Indexes due
to each mesostructure are shown in the figure, which would be
helpful for structural assessment.
It is quite important for the determination of mesostructures
to utilize electron diffraction (ED) patterns. TEM is powerful for
direct observation of periodic mesostructures. Overlapping and
tilting of specimens may lead to misunderstanding of periodic
mesostructures, so it is necessary to observe a thin part of
samples and confirm the symmetry of ED patterns. The diffi-
culty to interpret TEM images comes from the dimensions of
periodic mesostructures, such as lamellar, 2D hexagonal, and
3D cubic ones. In each mesostructural ordering, representative
TEM images are shown in Figures 811. Striped patterns are
observed for lamellar materials and the repeated distances
should be consistent with d-spacings confirmed by XRD. Tilting
Mesoporous Silica 137
of the specimens would occasionally lead to disappearance of
striped patterns. Both striped patterns and honeycomb images
are observed for 2D hexagonal mesoporous materials, such as
MCM-41, FSM-16, and SBA-15. The distances of the striped
patterns are changed according to the directions of electron
beam, which should be paid attention for their understanding.
In addition, the presence of micropores is confirmed in the
silicate frameworks of SBA-15 by high-magnification TEM
observation.68 Mesostructures with 3D periodicities are more
complicated for the structural evaluation, and then, it should
be required to observe them from three different directions
at least.6972 It is recommended to observe such materials
while referring to typical TEM images that are observed for
each mesostructure.
Considering the accumulation of cage-type mesopores,
simple understanding of 3D mesostructures has been tried by
using polyhedra as spherical cage-type mesopores with differ-
ent sizes.73,74 Because cubic close-packed (ccp, Fm-3m) and
hexagonal close-packed (hcp, P63/mmc) structures have the
same packing density, the difference in the mesostructural
orderings between Fm-3m and P63/mmc is only the stacking
direction of cage-type mesopores. They are just constructed by
a b c a b c packing along with [111] and a b a b a b packing
along with [001], respectively. Other periodicities of cage-type
mesoporous silicas are also understood by using accumulation
of several polyhedra. Im-3m (body-centered cubic, bcc), Fm-3m,
and P63/mmc structures are close-packed by one kind of poly-
hedron. Pm-3n and Fd-3m structures are composed of two
kinds of polyhedra, while P42/mmm structure is constructed
by three kinds of polyhedra. A method to understand the
earlier mentioned periodic structures has been proposed
by illustration using 12- (512), 14- (51262), 15- (51263), and
16-hedra (51264). As identified by coloring in Figure 12,
the Pm-3n structure is constructed by four 12-hedra and six
138 Mesoporous Silica

Lamellar FSM-16 KSW-2

(10)

(11)
(001)
Intensity (a.u.)

Intensity (a.u.)
Intensity (a.u.)
(002)

(22)
(20)
(11)

(31)
(003)

(20)

(21)
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
2q / 2q/ 2q/

SBA-1 SBA-2 MCM-48

(211)
(210)

(002)
Intensity (a.u.)

Intensity (a.u.)
Intensity (a.u.)

(101)
(200)

(211)

(101)

(220)
(112)

(332)
(420)
(110)

(332)
(103)

(400)
(321)
(211)

(431)
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
2q/ 2q/ 2q/

Figure 6 Low-angle XRD patterns of periodic mesoporous silica that can be prepared using alkyltrimethylammonium surfactants. (SBA-1, MCM-48)
Reproduced from Figure 1 in Kruk, M.; Jaroniec, M.; Ryoo, R.; Kim, J. M. Chem. Mater. 1999, 11, 2568; (SBA-2) Reproduced from Figure 5 in Huo, Q.;
Margolese, D. I.; Stucky, G. D. Chem. Mater. 1996, 8, 1147.

(Fm-3m) (Im-3m) (p6mm)

(111)

(011)

(10)
Intensity (a.u.)
Intensity (a.u.)

Intensity (a.u.)
(200)

(331, 420)
(220)

(002)
(400)

(112)
(022)
(422)

(013)

(11)
(20)

(21)

0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
2q / 2q / 2q/

Figure 7 Low-angle XRD patterns of periodic mesoporous silica that can be prepared using EOmPOnEOm-type triblock copolymers. Reproduced from
Figure 2 in Kleitz, F.; Kim, T.-W.; Ryoo, R. Langmuir 2006, 22, 440.

[100] [110]

100 nm 100 nm

Figure 8 Representative TEM images of 2D hexagonal (p6mm) mesoporous silica (SBA-15) along with [100] and [110] planes.
 Mesoporous Silica 139

100 nm 004
[100] [110] 100 nm
220

330
004
004

040 004 040

50 nm 50 nm [100] [110]

[111] [210] 220 [111]

022

404 240

004
100 nm
50 nm
440
50 nm
Figure 11 Representative TEM images of cubic (Ia-3d) mesoporous
Figure 9 Representative TEM images of cubic (Pm-3n) mesoporous
silica along with [hkl] planes. Reproduced from Figure 1 in Sakamoto, Y.;
silica (SBA-6) along with several [hkl] planes. Reproduced from Figure 2
Kim, T.-W.; Ryoo, R.; Terasaki, O. Angew. Chem., Int. Ed. 2004,
in Sakamoto, Y.; Kaneda, M.; Terasaki, O.; Zhao, D. Y.; Kim, J. M.;
43, 5231.
Stucky, G.; Shin, H. J.; Ryoo, R. Nature 2000, 408, 449.

phases of Pm-3n and Fd-3m structures. On the basis of the

[100] [110] detailed TEM observations of Pm-3n and Fd-3m mesoporous
silicas, an additional method to illustrate the defects in cage-
type mesoporous silica has been proposed by using unique
polyhedra, such as 13- and 15-hedra.74 In addition, it is
reported that the stacking faults and pore connections in
cage-type mesoporous silica (FUD-12) can be observed directly

220 by TEM.75 Thus, it is expected that the characterization tech-
002
niques will be further developed by using electron beams.
020 002
Scanning electron microscopy (SEM) is used for observa-
50 nm 50 nm tion of morphologies of materials, but recent development of
the instruments for high-resolution SEM observation leads to
[111]
direct observation of mesopore arrangements.76 In addition to
the periodic arrangements of mesopores that can be observed
by TEM, their connectivity can also be observed by high-
resolution SEM. As shown in Figure 13, the SEM image of 2D
hexagonal SBA-15 shows that a part of 1D mesopores are

connected at the surfaces of particles,77 which reduces the
022
effective use of mesopores. In order to fully utilize all meso-
pores, further development of synthetic methods controlling

50 nm
202
the connected parts of the mesopores may be needed. In this
sense, synthesis from layered polysilicates would be considered
Figure 10 Representative TEM images of cubic (Im-3m) mesoporous as an ideal reaction system33 that does not contain the forma-
silica (SBA-16) along with several [hkl] planes. Reproduced from Figure 4 tion of new siloxane networks at the entrances of mesopores.
in Sakamoto, Y.; Kaneda, M.; Terasaki, O.; Zhao, D. Y.; Kim, J. M.; In addition to 2D hexagonal MCM-41 and SBA-15, porous
Stucky, G.; Shin, H. J.; Ryoo, R. Nature 2000, 408, 449. structures of cubic (Ia-3d) mesoporous silica (KIT-6) were
also observed by high-magnification SEM that provided the
14-hedra, and the 12-hedra are located at the position of a bcc images shown in Figure 14.78 The presence of bicontinuous
structure. The Fd-3m structure is also constructed by 12- and mesopores is proved clearly; the size of pores connected
14-hedra, but similar sheets A, B, and C of only 12-hedra are between individual bicontinuous mesopore networks is varied
alternatively accumulated with another sheet a, b, and g com- with the synthesis temperature.
posed of 12- and 14-hedra, such as A a B b C g packing along N2 adsorptiondesorption measurement is a general
with [111]. Such accumulations of cage-type mesopores do method to evaluate porosity, such as size of mesopores. Spe-
not occur so ideally, and then, it is difficult to obtain pure cific surface area and pore volume are calculated using
140 Mesoporous Silica
(a) (b)
A(B, C)
[110] [110]
[112] [112]
[111] [111]
(c) (d)
Figure 12 (a) Pm-3n73 and (b) Fd-3m structures illustrated by polyhedra and (c, d) the sheet structures composed of polyhedra and (e) of the
accumulated Fd-3m. Reproduced from Figures 2 and 3 in Han, L.; Sakamoto, Y.; Che, S.; Terasaki, O. Chem. Mater. 2009, 21, 223.
20 nm
Figure 13 High-resolution SEM image of 2D hexagonal (p6mm)
mesoporous silica (SBA-15). Reproduced from Figure 1d in Che, S.;
Lund, K.; Tatsumi, T.; Iijima, S.; Joo, S. H.; Ryoo, R.; Terasaki, O. Angew.
Chem., Int. Ed. 2003, 42, 2182.
adsorption data and pore size distribution and shape of meso-
pores are derived on the basis of the adsorptiondesorption
isotherm. Typical N2 sorption data of SBA-15 and SBA-16 are
shown in Figures 15. N2 molecules are initially adsorbed over
the surfaces of samples, and a multilayered state is obtained
with the increase in the relative pressure (P/P0) of N2. Capillary
condensation of N2 molecules then occurs at certain P/P0
values depending on pore size. Adsorption and desorption
behaviors in pores smaller than 4 nm are almost consistent,
while hysteresis loops are observed for materials with meso-
pores larger than 4 nm. When mesopores are tubular (e.g., 1D
cylinders and wormhole-like mesopores), the adsorption iso-
therm due to capillary condensation was almost parallel to
(,)
A
C
[111]
A
[112]
[110]
(e)
hysteresis behavior in desorption.79 In cage-type mesopores,
spherical mesopores and pore windows to connect the meso-
pores are present in the porous structures, and then, the
desorption behavior is very different from those observed for
materials with cylindrical mesopores. Sudden decrease due to
pore windows smaller than spherical mesopores is observed in
the desorption blanch.80
It is necessary to decide the periodic structures of mesopor-
ous materials totally. For SBA-15, the 2D hexagonal structure is
verified by XRD and the presence of cylindrical mesopores is
confirmed by TEM and N2 sorption measurement.18,19 High-
resolution SEM observation gives us further information
whether mesopores are open or not. In addition, the presence
of micropores inside silicate frameworks is viewed in high-
magnification TEM images.68 The presence of such micropores
was supposed to be present on the basis of N2 sorption data at
low P/P0 regions. For example, such micropores in the silicate
frameworks of SBA-15 disappeared with the increase in tem-
peratures during calcination.79,81 In cage-type mesoporous sil-
ica (FDU-12, SBA-16), pore windows are reduced and finally
closed by elevating calcination temperatures, leading to the
transformation into silica containing regularly arranged and
isolated spherical pores.82 In the case of SBA-16, the presence
of the ordered mesostructures is confirmed with shrinkage
of the ordered mesostructure by XRD even when SBA-16
is calcined at higher temperatures, as shown in Figure 16.
In contrast, the N2 adsorptiondesorption isotherm of
SBA-16 calcined above 900  C showed a nonporous behavior,
though adequate adsorption of N2 with H2 hysteresis
 Mesoporous Silica 141

50 nm 50 nm
(a) (b)

Figure 14 High-magnification SEM images of cubic (Ia-3d) mesoporous silica (KIT-6) with bicontinuous mesopore networks prepared at (a) 40  C and
(b) 100  C. Reproduced from Figure 2 in Tuysuz, H.; Lehmann, C. W.; Bongard, H.; Tesche, B.; Schmidt, R.; Schuth, F. J. Am. Chem. Soc. 2008,
130, 11510.

1000 500
Amount adsorbed (cm3 g1)

Amount adsorbed (cm3 g1)

800 400 Desorption
Desorption

600 300
Adsorption
Adsorption
400 200

200 100

0 0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(a) P/P0 (b) P/P0

Figure 15 N2 adsorptiondesorption isotherms of 2D hexagonal SBA-15 and cubic (Im-3m) SBA-16. (b) Reproduced from Figure 4d in Kleitz, F.;
Czuryszkiewicz, T.; Solovyov, L. A.; Linden, M. Chem. Mater. 2006, 18, 5070.

110

200
211 220
310
Amount adsorbed (cm3 STP g1)

200 550 C 400 550 C

211 220 700 C
310
Log(intensity)

300 800 C
200 700 C
211 220
310
110
200
200 800 C
211 220
310
100 900 C
200
900 C 950 C
211
310 0
0.0 0.2 0.4 0.6 0.8 1.0
950 C
Relative pressure

0.4 0.6 0.8 1.0 1.2 1.4 1.6

q (nm1)

Figure 16 Low-angle XRD patterns and N2 adsorptiondesorption isotherms of SBA-16 calcined at different temperatures. Reproduced from Figure 3
in Kruk, M.; Hui, C. M. J. Am. Chem. Soc. 2008, 130, 1528.
142 Mesoporous Silica

is observed for SBA-16 calcined up to 800  C. The results reveal
that cage-type mesopores are isolated completely by calcina-
tion at 950  C.
5.06.4 Morphological Control of Periodic
Mesoporous Silica
The development of morphological control, including fabrica-
tion of thin films, is very important for extending the applica-
tions of mesoporous silica. Precise control of hydrolysis and
condensation of tetramethoxysilane (TMOS) with the self-
assembly of CnTMA molecules are the first example to engineer
transparent mesoporous silica films,83 and afterward, many
papers on mesoporous silica films have been reported. Clear
precursor solutions, which are prepared under acidic conditions,
are spin-coated and dip-coated, easily affording thin films of
highly ordered mesoporous silica.84,85 The fabrication process
Substrate
Substrate
of thin films by dip coating is shown in Figure 17. Because the
self-assembly of surfactant molecules is induced during evapora-
tion of solvents (evaporation-induced self-assembly, EISA),
highly ordered silicasurfactant composite films can be obtained
when silicate species are not condensed so excessively.85 Hydro-
thermal synthesis of mesoporous silica films was also reported at
the almost same time.86 When a substrate is placed at the inter-
face of acidic precursor solutions containing both CnTMA sur-
factant and silicate species, CnTMAsilica composites are
deposited at the surface of the substrate. In comparison with
the aforementioned solgel process, the hydrothermal fabrication
process of thin film is also illustrated in Figure 17. The
most distinguishing point of the hydrothermal process is a pos-
sibility to obtain self-standing mesoporous silica films at the
airwater interface.87
The synthetic method of mesoporous silica films through
solgel process is highly versatile. Inorganicorganic composite
coatings, such as nacre, are possible by molecular design of
surfactants.88 By using surfactants whose hydrophobic tail con-
taining a diacetylene unit, a chromatic variation was observed
by UV irradiation to the nanocomposite films.90 Mesoporous
silicas that are morphologically controlled by various methods
are summarized in Figure 18. Elastomeric relief stamp com-
posed of poly(dimethyl siloxane) (PDMS) was placed at the

(Self-assembly)
substrate surface, and then, a clear precursor solution was
poured inside the relief by capillary force, leading to the facile
(Evaporation of solvent) formation of micropatterned precursor films of mesoporous
silica.90,91 Spherical nanoparticles of mesoporous silica can be
Surfactant Silicate species
prepared by spray drying of clear precursor solution.92 Direct
(Precursor solution) formation of nanocomposite spheres containing fluorescent
organic molecules, organic metal nanoparticles, and polymers
(a) (Precursor solution) (b) Closed vessel also occurred during the aerosol-assisted self-assembly process.
Figure 17 General synthetic routes of mesostructured silica films by Micropatterning of ordered mesoporous silica was also extended
(a) solgel and (b) hydrothermal methods. by micropen lithography and ink-jet printing.93 Micropatterning

5 m 5 m 5 m
(a) (b) (c)

1 mm

50 nm 3 mm 200 m
(d) (e) (f) (g)

Figure 18 (ac) SEM images of micropatterned mesoporous silica by using designed PDMS stumps,90,91 (d) TEM image of spherical mesoporous silica
particles prepared by spray drying,92 and optical micrographs of mesoporous silica patterned by (e) micropen lithography and, (f) ink-jet printing 93 and
(g) by UV irradiation.94 Reproduced from (a,b) Figure 2 in Trau, M.; Yao, N.; Kim, E.; Xia, Y.; Whitesides, G. M.; Aksay, I. A. Nature 1997, 390, 674. (c)
Figure 2 in Yang, P.; Deng, T.; Zhao, D.; Feng, P.; Pine, D.; Chmelka, B. F.; Whitesides, G. M.; Stucky, G. D. Science 1998, 282, 2244. (d) Figure 2b in Lu, Y.;
Fan, H.; Stump, A.; Ward, T. L.; Rieker, T.; Brinker, C. J. Nature 1999, 398, 223. (e,f) Figures 2a and 3b in Fan, H.; Lu, Y.; Stump, A.; Reed, S. T.; Baer, T.;
Schunk, R.; Perez-Luna, V.; Lopez, G. P.; Brinker, C. J. Nature 2000, 405, 56. (g) Figure 2B in Doshi, D. A.; Huesing, N. K.; Lu, M.; Fan, H.; Lu, Y.; Simmons-
Potter, K.; Potter, B. G., Jr.; Hurd, A. J.; Brinker, C. J. Science 2000, 290, 107.

of mesoporous silica was possible by area-selective UV
irradiation.94 When UV light was irradiated over mesostructured
silica film through photomask, silicate frameworks in unmasked
regions were condensed by UV irradiation.94 Because silicate
frameworks in the masked region are not condensed, the masked
region is selectively leached for micropatterning. After micropat-
terning of mesostructured films of (EO)106(PO)70 (EO)106
(Pluronic F127) containing an organic acid by lithography, the
micropatterning can be replicated with mesoporous silica by
infusion of silicon alkoxide using supercritical CO2 and the selec-
tive condensation.95 Freestanding mesoporous silica films are
obtained by slow evaporation of solvent (water) from suspen-
sions of nanocrystalline cellulose that organizes a tunable chiral
nematic structures.96
Alignment control of mesopores is also very interesting for
a variety of applications. As a main purpose, 1D mesopores in
2D hexagonal mesoporous silica are aligned over substrates.
As rodlike micelles of surfactants are oriented to flow direction,
it has been reported that 1D mesopores of 2D hexagonal
mesoporous silica are oriented to flow direction during
coating.97,98 However, the entire alignment of 1D mesopores
strongly depends on the coating conditions. Accordingly, it is
demanded for precise alignment of 1D mesopores. The align-
ment control in strong magnetic fields was also reported,99 but
the direction of 1D mesopores over the entire films has not
been controlled. The process by using powerful magnetic field
may not be a final goal for alignment control of mesoporous
silica. Although there is a possibility to orient 1D mesopores
over a particular crystal plane of a silicon substrate,100 the use
of polymer-coated substrates is proposed for the alignment
control of 1D mesopores. A fabrication process of fully aligned
mesoporous silica film by using polyimide is shown in
Figure 19.101 Polyimide is coated over a glass substrate and
the polymer chains are aligned to the same direction by a
rubbing treatment. When a mesostructured precursor film
was prepared over the polyimide-coated substrate, it was
O O
N N
O O n
Polyimido coating
Substrate
Coating Rubbing
10 m Not aligned
Figure 19 Alignment control of tubular mesopores by using polyimide-coated substrate. Reproduced from Figures 2a, 2b, 7, 8b in Miyata, H.;
Kuroda, K. Chem. Mater. 1999, 11, 1609.
Mesoporous Silica 143
found that 1D mesopores were fully aligned perpendicular to
the rubbing direction. The orientation of the mesopores can be
proved by in-plane and out-of-plane XRD. The interaction
between the part of alkyl chains and polyimide chains is
important for the uniaxial alignment. It is reported that single
crystal-like 3D hexagonal (P63/mmc) mesoporous silica films
are formed by phase transition over a polyimide-coated
substrate.102 Photo alignment of 1D mesopores is also
reported.103 Monolayer of poly(vinyl alcohol) with photosen-
sitive azobenzene side chains was constructed by the Lang-
muirBlodgett method over a substrate, and the azo groups
are oriented perpendicular to the substrate by UV irradiation.
Silica with 1D mesopores aligned parallel to the polymer-
coated substrate is obtained after covering with poly(di-n-
hexylsilane).
Vertical alignment of 1D mesopores is required to easily
locate nanomaterials in a hexagonal order. Although the syn-
thesis of mesoporous silica was carried out in strong magnetic
fields, a part of 1D mesopores were vertically aligned to a low
extent. The synthesis inside trenches is not expected for full
alignment of 1D mesopores because the planes parallel to a
substrate must be present in the trenches. The development of
the strategies to synthesize vertically aligned mesoporous sil-
icas is summarized in Figure 20. When mesoporous silica is
synthesized inside columnar pores of alumina membrane, 1D
mesopores are oriented parallel to the columnar pores (per-
pendicular to the substrate).104 However, the alignment
and orientation of 1D mesopores are almost defined according
to the relation between diameter of the 1D cylinders of porous
anodic alumina (PAA) membranes and size of 1D
mesopores.105 Accordingly, to change the size of 1D meso-
pores within the confined space is not so simple. It was
reported that surfactant-assisted mesoporous silica was synthe-
sized inside the cylindrical pores of polycarbonate (PC)
membranes.106 It was found that 1D mesopores were aligned
inside cylinders of PC membranes through the introduction of
Coating 20 nm
Aligned 10 m
144 Mesoporous Silica

50 nm
50 nm 200 nm
(a) (b) (c)

20 nm 20 nm 50 nm
(d) (e) (f)

Figure 20 (a) Top and (b) side view by SEM observations of silica with aligned tubular mesopores in PAA membrane (reproduced from Figure 3b, 3c in
Yamaguchi, A.; Uejo, F.; Yoda, T.; Uchida, T.; Tanamura, Y.; Yamashita, T.; Teramae, N. Nat. Mater. 2004, 3, 337), (c) TEM image of silica with tubular
mesopores aligned by shear flow in porous polycarbonate membrane (reproduced from Figure 2c in Yamauchi, Y.; Suzuki, N.; Kimura, T. Chem.
Commun. 2009, 5689), (d) top and (e) side view by TEM observations of silica with electrochemically aligned tubular mesopores (reproduced from
Figure 1c, 1d in Walcarius, A.; Sibottier, E.; Etienne, M.; Ghanbaja, J. Nat. Mater. 2007, 6, 602), and (f) silica with vertically aligned mesopores that are
grown epitaxially over cubic mesoporous titania film (reproduced from Figures 1c, 2a in Richman, E. K.; Brezesinski, T.; Tolbert, S. H. Nat. Mater. 2008,
7, 712).

clear precursor solution by aspiration.107 It is considered that
the alignment control of 1D mesopores is achieved in PC
membrane by shear flow because the mesostructured precursor
interacts with PC surfaces weakly rather than PAA surfaces.
Vertically aligned 1D mesopores are fabricated through
electrochemically assisted self-assembly,108 but the method
can only be adapted over conductive substrates. Although
perfect synthetic methods to align 1D mesopores vertically
have not been reported so far, a synthetic method developed
on the basis of the concept using epitaxial growth would be
one of the most flexible ones for obtaining vertically oriented
2D hexagonal mesoporous silica films.109 When a cubic meso-
porous titania film is synthesized over a substrate, a hexagonal
arrangement of mesopores is expected to be exposed on the
film surface. Then, 1D mesopores of a mesoporous silica film
can be vertically oriented when the lattice constants are con-
sistent with those of the titania film. A PAA membrane with
designed conical spaces was engineered electrochemically,
which was useful for standing of 1D mesopores because spher-
ical mesopores can be formed inside the conical spaces.110
Although rodlike micelles are strongly interacted with the sur-
faces of substrates and then reclined parallel to substrates,
discotic lyotropic liquid crystal of nonsurfactant supramolecu-
lar assemblies is useful for achieving vertical alignment of 1D
mesopores,111 which will be very helpful if the concept is
widely applicable for the synthesis of mesoporous silica films.
5.06.5 Applications Using Periodic Mesoporous
Silica
There have been enormous papers concerning the incorpora-
tion of different metal species in silicate frameworks that are
quite effective for the generation of unique functions.112,113
Functions coming from mesoporous structures themselves have
rarely been reported so far. Toward applications of ordered
mesoporous silica to catalysts, there have been several methods
to introduce metal species inside mesopores. In addition to imp-
regnation and ion exchange of metal species, metal complexes
are fixed at the surfaces of the silicate frameworks 114 and con-
structed inside ordered mesopores.115,116 It is also possible to
encapsulate metal species into micelles during the synthesis
of mesostructured precursors.117 As another case, metal nano-
particles are coated with ordered mesoporous silica, affording
coreshell mesoporous silica nanoparticles.118
Much attention was initially paid to the possibilities of
periodic mesoporous materials for applications in catalytic
reactions dealing with large organic molecules that cannot be
handled inside micropores as found for zeolites.112 Because
there have often been problems on low catalytic activities due
to amorphous silica-based frameworks, crystallization of silica-
based frameworks has strongly been demanded. On the other
hand, the presence of uniform mesopores is quite attractive as
particular nanovessels involving catalytic reactions. Here, sev-
eral reactions are described as successful examples due to shape
selectivity of uniform mesopores.
The synthesis of fine chemicals was firstly reported using
acidity of Al-FSM-16 in 1995.119 As shown in Figure 21, this
reaction is directed to the formation of a bulky organic mole-
cule (meso-tetraporphyrin) and optimal pore diameter for the
reaction was found in mesopore region. In the enrichment of
Taxol (a very small amount of Taxol is recovered as an anti-
tumor agent from the stem bark of western yew), in addition to
the shape selectivity by different mesopores of FSM-16, the
affinities of functional groups in Taxol to both silicate surfaces
and solvents are important for the separation.120 Although

H O
H NH
N Al-FSM-16 p-chloranil
+
CH2Cl2 (RT) (reflux)
Figure 21 Synthetic scheme of meso-tetraporphyrin.119
4.5
Reaction rate constant (ms1 g-cat1)
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.0
1.5 2.5 3.5
Pore diameter (nm)
Figure 22 Variation of reaction rate constant relating acetalization of
cyclohexanone with methanol over siliceous MCM-41. Reproduced from
Figure 1 in Iwamoto, M.; Tanaka, Y.; Sawamura, N.; Namba, S. J. Am.
Chem. Soc. 2003, 125, 13032.
periodic mesoporous silica does not show noticeable acidity
without incorporating other metal units, such as Al into silicate
frameworks, several reactions proceed over siliceous mesopore
surfaces. As shown in Figure 22, in acetalization of cyclohexa-
none with methanol over siliceous MCM-41, the reaction
rate constant varies depending on the mesopore size.121 Con-
sidering the small molecular size of cyclohexanone, this result
cannot be explained only by shape selectivity. Accordingly,
future elucidation of the peculiar phenomenon provided by
mesospaces is expected.
Inclusion chemistry in uniform mesopores has increa-
singly been developed since the discovery of ordered meso-
porous silica.122 Polymer syntheses in uniform mesopores
are a typical example. Conductive polyaniline filaments and
carbon wires were fabricated in 1D mesopores of MCM-41 by
oxidation and radical polymerization with subsequent pyrolysis,
respectively.123,124 Polymerization of ethylene in 2D hexagonal
mesoporous silica fiber containing a titanocene catalyst led to
the formation of crystalline linear polyethylene nanofibers.125
In these cases, the direction of mesochannels in mesoporous
silica fibers should be carefully characterized. Device appli-
cations using mesoporous silica films for photoelectric conver-
sion are summarized in Figure 23. Energy transfer from a
Mesoporous Silica 145
N
N HN
semiconducting polymer depending on the orientation of poly
[2-methoxy-5-(20 -ethylhexyloxy)-1,4-phenylene vinylene]
(MEH-PPV) was investigated by using 1D mesopores.126 The
orientation of polymer chains is controlled by using a fully
aligned mesoporous silica film, and then, amplified spontane-
ous emission is reduced 20 times lower than that in random
polymer chains.127
When micropatterned mesostructured silica was fabricated
over a mesoporous silica film showing low reflectance, the
mesostructured silica film worked as waveguides, and then,
mirrorless lasing was observed by doping rhodamine 6G in
the mesostructured silica film.128 It was reported that the sep-
aration of biomolecules was possible through oriented 1D
mesopores in PAA membranes.104 Although rodlike mesopor-
ous silica fabricated inside columnar pores was detached from
PAA surfaces during the removal of surfactant molecules, there
4.5
was a possibility for complete separation of biomolecules
using mesoporous membranes in nanometer scale. Otherwise,
there have been a large number of papers describing encapsu-
lation of intelligent biomolecules inside uniform and periodic
mesopores.129 Pluronic P123-assisted mesoporous silica films
were aligned by dip coating and used as proton-conductive
films.98 For example, a gate effect was observed at the isoelec-
tric point of silica when hydrochloric acid was poured into the
film; proton was not transported below the isoelectric point.
Modification of silicate surfaces with organic functional
groups has widely been extended for applications using meso-
porous silica. When mercapto groups are anchored by silyla-
tion of MCM-41, toxic heavy metal ions, such as mercury, are
selectively adsorbed over the organic groups and removed
from the aqueous solutions.130 Partial organic modification
only at the entrance of mesopores is also possible by direct
modification of mesostructured silica.131,132 Mesostructured
precursor of MCM-41 is modified with an organosilane com-
pound possessing coumarin group is at the end of the organic
chain. Photosensitive mesoporous silica using reversible
dimerization of coumarin groups is obtained (open or closed
at the entrance of mesopores) after the extraction of surfactant
molecules accommodated inside the mesopores. The sche-
matic illustration of the reaction is shown in Figure 24.
Although this finding is actually an epoch-making study for
drug delivery system (DDS) applications, appropriate and real-
istic stimulation method should be discussed for practical
applications in human body because dimerization of couma-
rin molecules requires UV irradiation. In the synthesis of chiral
mesoporous silica prepared using anionic surfactants, in
146 Mesoporous Silica

1.6 ||

1.4

1.2
Absorbance (OD)

1
O
0.8
Excitation
0.6
n
MeO
0.4

0.2

0 n = 1.43

400 450 500 550 600 650 n1.15 To detector

n = 3.5
(a) Wavelength (nm) (b)

V
100
Mesoporous film
Photoresist
Conductance (nS)

10

1.0

HCl or KCl

0.1
108 106 104 102 100
(c) (d) [H+] (M)
Figure 23 Device applications using mesoporous silica films due to (a) optical property of encapsulated semiconducting polymer (reprodued from
Figure 1a, 1c in Martini, I. B.; Craig, I. M.; Molenkamp, W. C.; Miyata, H.; Tolbert, S. H.; Schwartz, B. J. Nat. Nanotechnol. 2007, 2, 647), (b) mirrorless
lasing from mesostructured waveguides (reproduced from Figure 5 in Yang, P.; Wirnsberger, G.; Huang, H. C.; Cordero, S. R.; McGehee, M. D.; Scott, B.;
Deng, T.; Whitesides, G. M.; Chmelka, B. F.; Buratto, S. K.; Stucky, G. D. Science 2000, 287, 465), and (c) proton conductivity (reproduced from
Figure 2a, 2c in Fan, R.; Huh, S.; Yan, R.; Arnold, J.; Yang, P. Nat. Mater. 2008, 7, 303).

2.62 nm

1.3 nm Photo irradiation

>310 nm

250260 nm

Coumarin
dimer
Release Storage
Figure 24 Conceptual scheme of photo-switched drug release by coumarin-modified MCM-41. Reproduced from Figure 15 in Mal, N. K.; Fujiwara, M.;
Tanaka, Y.; Taguchi, T.; Matsukata, M. Chem. Mater. 2003, 15, 3385.

+
N Na
N
+ OOC R
HN O
OOC R
+
HN O
N
OOC R O
HN
N+
Extraction
CMS
Figure 25 Mesoporous silica with the helical arrangement of surface organic groups and corresponding packing of functional organic molecules.
Reproduced from Figure 1 in Qiu, H.; Inoue, Y.; Che, S. Angew. Chem., Int. Ed. 2009, 48, 3069.
addition to simple silica sources, such as silicon alkoxides,
organosilanes containing ammonium groups must be utilized
to generate interactions between anionic surfactant molecules
and frameworks.25 Although chiral mesoporous silica is
obtained after complete removal of organic fractions by calci-
nation, the ammonium groups can remain at the surfaces of
the silicate frameworks by extraction of surfactant molecules
only. As shown in Figure 25, when several kinds of anionic
functional molecules were adsorbed over the surfaces, the
chiral conformation of polymer chains, chiral stacking of por-
phyrin compound, and recognition of DNA chirality were
observed over the organic group remaining chiral mesoporous
silica,133 revealing a possibility of the imprinting of chiral
arrangement at the silicate surfaces.
A wide variety of particle morphologies have been reported as
the synthesis of ordered mesoporous silica.134138 It is possible to
synthesize mesoporous silica nanoparticles by controlling con-
densation rate of silicate species and different strength of interac-
tions between silicate species and surfactant molecules.139143 The
fabrication of mesoporous silica nanoparticles is a key technology
for realizing smooth diffusion of reactants and products during
catalytic reactions. Mesoporous silica nanoparticles with a high
density of amino groups144 and coreshell-type mesoporous sil-
ica nanoparticles were also prepared by controlled procedures.145
The synthetic method in cylindrical pores of PC membranes is
interesting to obtain mesoporous silica rods because PC mem-
branes and surfactant molecules can be eliminated at the same
time by calcination.106,107 There have been many attempts
to fabricate mesoporous silica using fibrous bacteria and sponge-
like polyurethane for incorporation of macrospaces inside
Mesoporous Silica 147
N+
N+ Na ooc
ooc
PPAS ooc
+ ooc
N
N+ ooc
N+
TPPS
Loading
B-DNA
L-CMS R-CMS
Slower loading Faster loading
materials.146,147 The macrospaces are helpful for effective diffu-
sion in the research fields of catalysts and adsorbents.
DDS systems using mesoporous silica nanoparticles have
been increasingly investigated by utilizing organic modification
techniques.148 Photo-controlled reversible release of guest mol-
ecules was reported using selective anchoring of organic
groups.131,132 DNA and chemicals were successfully delivered
into plants using organically modified mesoporous silica
nanoparticles.149 The entrances of mesopores can be capped
with Au nanoparticles. The schematic reaction process is shown
in Figure 26. Mesoporous silica nanoparticles pass through cell
membranes by covering with DNA, transporting DNA and che-
micals. Visualization of mesoporous silica nanoparticles in plant
cells was achieved by labeling with fluorescent dye molecules.149
Likewise, such visualization methods were widely applied in
various areas investigating cell viability extended by cell-directed
assembly of silicalipid nanocomposites,150 diffusion of mole-
cules inside mesopores,151 diffusion of molecules during cata-
lytic reactions,152 distribution of organic functional groups by
silylation,153,154 and so on.
5.06.6 Conclusion
It is no doubt that a family of surfactant-assisted mesoporous
silica is extremely useful for applications because of fine tuning
of mesopores with compositional and morphological varia-
tions. Mesoporous silica-based materials applicable for cata-
lytic reactions can also be prepared through incorporation of
148 Mesoporous Silica
Gold Gold
Type-II Type-IV
Plasmid DNA
Type-II
Gold
Type-II
Gold
Type-IV
Cell membrane
Figure 26 Use of organically modified mesoporous silica nanospheres for plant cell internalization. Reproduced from Figure 1a in Torney, F.; Trewyn,
B. G.; Lin, V. S.-Y.; Wang, K. Nat. Nanotechnol. 2007, 2, 295.
other metal units into silicate frameworks and fabrication of
metal nanoparticles inside mesopores.
Extension of applications utilizing mesoporous materials
can be expected, which is strongly related to further progress
of synthetic methods of mesoporous materials. Crystallization
of silica-based mesoporous materials is being realized steadily,
as for the cases of mesoporous zeolites by hard templating in a
replicated mesoporous carbon155 and by tiling of zeolite
nanocrystals.156 Another type of mesoporous zeolites prepared
in the presence of organosilane with long alkyl chain is also
known.157 Further design of the molecular structure of orga-
nosilane compounds leads to the successful synthesis of zeolite
nanosheets.158 Nevertheless, silicate frameworks basically lack
in rich functions; thus, combination of silica-based mesopor-
ous materials with organic functionalities will be extremely
promising hereafter. A number of periodic mesoporous silica
synthesized from only organosilanes with long alkyl chain
have been reported, and the molecular structures of both
= Fluorescein
= -oestradiol
Gold
Type-II Type-III
Freed
plasmid
Type-IV
Gold
inorganic units and organic groups can be designed.159,160 In
fact, the applications of mesoporous silica modified with orga-
nosilane coupling agents have been developed by using uni-
form mesopores and features arising from designable organic
functional groups. A rapid extension of studies on hybrid
mesoporous materials (periodic mesoporous organosilica,
PMO) prepared from organically bridged silane compounds
(bridged silsesquioxanes)161165 also supports the significance
of the use of organically modified mesoporous silica. A wide
variety of organic groups can be embedded164,165 in the frame-
works, and conjugated organic groups are arranged in the
hybrid frameworks regularly under the appropriate synthetic
conditions.166,167 In PMO-type materials, uniform mesopores
that are defined by the exact size of surfactant assemblies are
utilized effectively. Artificial photosynthesis systems con-
structed inside mesopores of PMO is a typical example.168
When ordered mesoporous materials are chosen for ap-
plied studies, the following points should be considered:

understanding of functions directed to applications, the
essential/required role of silica-based frameworks, full utiliza-
tion of uniform mesospaces (if applicable), selection of appro-
priate morphologies, and so on. These key factors should
be further investigated to understand the total feature of meso-
porous silica. For related chapters in this Comprehensive,
we refer to Chapters 5.10 and 7.10.
References
1. Huo, Q.; Margolese, D. I.; Ciesra, U.; Feng, P.; Gler, T. E.; Sieger, P.; Leon, R.;
Petroff, P. M.; Schuth, F.; Stucky, G. D. Nature 1994, 368, 317.
2. Huo, Q.; Margolese, D. I.; Ciesla, U.; Demuth, D. G.; Feng, P.; Gier, T. E.;
Sieger, P.; Firouzi, A.; Chmelka, B. F.; Schuth, F.; Stucky, G. D. Chem. Mater.
1994, 6, 1176.
3. Monnier, A.; Schuth, F.; Huo, Q.; Kumar, D.; Margolese, D.; Maxwell, S.;
Stucky, G. D.; Krishnamurty, M.; Petroff, P.; Firouzi, A.; Janicke, M.;
Chmelka, B. F. Science 1993, 261, 1299.
4. Firouzi, A.; Kumar, D.; Bull, L. M.; Besier, T.; Sieger, P.; Huo, Q.; Walker, S. A.;
Zasadzinki, J. A.; Glinka, C.; Nicol, J.; Margolese, D.; Stucky, G. D.; Chmelka, B. F.
Science 1995, 267, 1138.
5. Yanagisawa, T.; Shimizu, T.; Kuroda, K.; Kato, C. Bull. Chem. Soc. Jpn. 1990,
63, 988.
6. Yanagisawa, T.; Shimizu, T.; Kuroda, K.; Kato, C. Bull. Chem. Soc. Jpn. 1990, 63,
1535.
7. Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature
1992, 359, 710.
8. Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.;
Schmitt, K. D.; Chu, C. T.-W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.;
Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834.
9. Inagaki, S.; Fukushima, Y.; Kuroda, K. J. Chem. Soc. Chem. Commun. 1993, 680.
10. Huo, Q.; Leon, R.; Petroff, P. M.; Stucky, G. D. Science 1995, 268, 1324.
11. Huo, Q.; Margolese, D. I.; Stucky, G. D. Chem. Mater. 1996, 8, 1147.
12. Tanev, P. T.; Chibwe, M.; Pinnavaia, T. J. Nature 1994, 368, 321.
13. Tanev, P. T.; Pinnavaia, T. J. Science 1995, 267, 865.
14. Attard, G. S.; Glyde, J. C.; Goltner, G. Nature 1995, 378, 366.
15. Tanev, P. T.; Pinnavaia, T. J. Science 1996, 271, 1267.
16. Kim, S. S.; Zhang, W.; Pinnavaia, T. J. Science 1998, 282, 1302.
17. Bagshaw, S. A.; Prouzet, E.; Pinnavaia, T. J. Science 1995, 269, 1242.
18. Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.; Chmelka, B. F.;
Stucky, G. D. Science 1998, 279, 548.
19. Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. J. Am. Chem. Soc.
1998, 120, 6024.
20. Wan, Y.; Zhao, D. Chem. Rev. 2007, 107, 2821.
21. Fyfe, C. A.; Fu, G. J. Am. Chem. Soc. 1995, 117, 9709.
22. McGrath, K. M.; Dabbs, D. M.; Yao, N.; Aksay, I. A.; Gruner, S. M. Science 1997,
277, 552.
23. Yokoi, T.; Yoshitake, H.; Tatsumi, T. Chem. Mater. 2003, 15, 4536.
24. Che, S.; Garcia-Bennett, A. E.; Yokoi, T.; Sakamoto, K.; Kunieda, H.; Terasaki, O.;
Tatsumi, T. Nat. Mater. 2003, 2, 801.
25. Che, S.; Liu, Z.; Ohsuna, T.; Sakamoto, K.; Terasaki, O.; Tatsumi, T. Nature 2004,
429, 281.
26. Yang, S.; Zhao, L.; Yu, C.; Zhou, X.; Tang, J.; Yuan, P.; Chen, D.; Zhao, D. J. Am.
Chem. Soc. 2006, 128, 10460.
27. Jin, C.; Han, L.; Che, S. Angew. Chem. Int. Ed. 2009, 48, 9268.
28. Han, Y.; Zhang, D.; Chng, L. L.; Sun, J.; Zhao, L.; Zou, X.; Ying, J. Y. Nat. Chem.
2009, 1, 123.
29. Inagaki, S.; Koiwai, A.; Suzuki, N.; Fukushima, Y.; Kuroda, K. Bull. Chem. Soc.
Jpn. 1996, 69, 1449.
30. Kimura, T.; Itoh, D.; Okazaki, N.; Kaneda, M.; Sakamoto, Y.; Terasaki, O.;
Sugahara, Y.; Kuroda, K. Langmuir 2000, 16, 7624.
31. Kimura, T.; Itoh, D.; Shigeno, T.; Kuroda, K. Langmuir 2002, 18, 9574.
32. Kimura, T.; Kamata, T.; Fuziwara, M.; Takano, Y.; Kaneda, M.; Sakamoto, Y.;
Terasaki, O.; Sugahara, Y.; Kuroda, K. Angew. Chem. Int. Ed. 2000, 39, 3855.
33. Kimura, T.; Kuroda, K. Adv. Funct. Mater. 2009, 19, 511.
34. Kimura, T.; Itoh, D.; Shigeno, T.; Kuroda, K. Bull. Chem. Soc. Jpn. 2004, 77, 585.
35. Kato, M.; Shigeno, T.; Kimura, T.; Kuroda, K. Chem. Mater. 2004, 16, 3224.
36. Kato, M.; Shigeno, T.; Kimura, T.; Kuroda, K. Chem. Mater. 2005, 17, 6416.
37. Tamura, H.; Mochizuki, D.; Kimura, T.; Kuroda, K. Chem. Lett. 2007, 36, 444.
Mesoporous Silica 149
38. Shigeno, T.; Inoue, K.; Kimura, T.; Katada, N.; Niwa, M.; Kuroda, K. J. Mater.
Chem. 2003, 13, 883.
39. Kitayama, Y.; Asano, H.; Kodama, T.; Abe, J.; Tsuchiya, Y. J. Porous Mater. 1998,
5, 139.
40. Kimura, T.; Suzuki, M.; Ikeda, T.; Kato, K.; Maeda, M.; Tomura, S. Microporous
Mesoporous Mater. 2006, 95, 146.
41. Kimura, T.; Tamura, H.; Tezuka, M.; Mochizuki, D.; Shigeno, T.; Ohsuna, T.;
Kuroda, K. J. Am. Chem. Soc. 2008, 130, 201.
42. Kimura, T.; Huang, S.; Fukuoka, A.; Kuroda, K. J. Mater. Chem. 2009, 19,
3859.
43. Schmidt-Winkel, P.; Lukens, W. W., Jr.; Zhao, D.; Yang, P.; Chmelka, B. F.;
Stucky, G. D. J. Am. Chem. Soc. 1999, 121, 254.
44. Wang, H.; Zhou, X.; Yu, M.; Wang, Y.; Han, L.; Zhang, J.; Yuan, P.;
Auchterlonie, G.; Zou, J.; Yu, C. J. Am. Chem. Soc. 2006, 128, 15992.
45. Kroger, N.; Deutzmann, R.; Sumper, M. Science 1999, 286, 1129.
46. Sumper, M. Science 2002, 295, 2430.
47. Kessel, S.; Thomas, A.; Borner, H. G. Angew. Chem. Int. Ed. 2007, 46, 9023.
48. Pouget, E.; Dujardin, E.; Cavalier, A.; Moreac, A.; Valery, C.; Marchi-Artzner, V.;
Weiss, T.; Renault, A.; Paternostre, M.; Artzner, F. Nat. Mater. 2007, 6, 434.
49. Ehrlich, H.; Deutzmann, R.; Brunner, E.; Cappellini, E.; Koon, H.; Solazzo, C.;
Yang, Y.; Ashford, D.; Thomas-Oates, J.; Lubeck, M.; Baessmann, C.;
Langrock, T.; Hoffmann, R.; Worheide, G.; Reitner, J.; Simon, P.; Tsurkan, M.;
Ereskovsky, A. V.; Kurek, D.; Bazhenov, V. V.; Hunoldt, S.; Mertig, M.;
Vyalikh, D. V.; Molodtsov, S. L.; Kummer, K.; Worch, H.; Smetacek, V.;
Collins, M. J. Nat. Chem. 2010, 2, 1084.
50. Templin, M.; Franck, A.; Chesne, A. D.; Leike, H.; Zhang, Y.; Ulrich, R.;
Schadler, V.; Wiesner, U. Science 1997, 278, 1795.
51. Goltner, C. G.; Henke, S.; Weissenberger, M. C.; Antonietti, M. Angew. Chem. Int.
Ed. 1998, 37, 613.
52. Kramer, E.; Forster, S.; Goltner, C.; Antonietti, M. Langmuir 1998, 14, 2027.
53. Yu, K.; Hurd, A. J.; Eisenberg, A.; Brinker, C. J. Langmuir 2001, 17, 7961.
54. Smarsly, B.; Xomeritakis, G.; Yu, K.; Liu, N.; Fan, H.; Assink, R. A.; Drewien, C. A.;
Ruland, W.; Brinker, C. J. Langmuir 2003, 19, 7295.
55. Yu, K.; Smarsly, B.; Brinker, C. J. Adv. Funct. Mater. 2003, 13, 47.
56. Deng, Y.; Yu, T.; Wan, Y.; Shi, Y.; Meng, Y.; Gu, D.; Zhang, L.; Huang, Y.; Liu, C.;
Wu, X.; Zhao, D. J. Am. Chem. Soc. 2007, 129, 1690.
57. Bloch, E.; Llewellyn, P. L.; Phan, T.; Bertin, D.; Hornebecq, V. Chem. Mater.
2009, 21, 48.
58. Warren, S. C.; Disalvo, F. J.; Wiesner, U. Nat. Mater. 2007, 6, 156.
59. Baccile, N.; Reboul, J.; Blanc, B.; Coq, B.; Lacroix-Desmazes, P.; In, M.;
Gerardin, C. Angew. Chem. Int. Ed. 2008, 47, 8433.
60. Atluri, R.; Hedin, N.; Garcia-Bennett, A. E. J. Am. Chem. Soc. 2009, 131, 3189.
61. Lintang, H. O.; Kinbara, K.; Tanaka, K.; Yamashita, T.; Aida, T. Angew. Chem. Int.
Ed. 2010, 49, 4241.
62. Schacht, S.; Huo, Q.; Voigt-Martin, I. G.; Stucky, G. D.; Schuth, F. Science 1996,
273, 768.
63. Imhof, A.; Pine, D. J. Nature 1997, 389, 948.
64. Yuan, J.-J.; Mykhaylyk, O. O.; Ryan, A. J.; Armes, S. P. J. Am. Chem. Soc. 2007,
129, 1717.
65. Fu, Q.; Rao, G. V. R.; Ward, T. L.; Lu, Y.; Lopez, G. P. Langmuir 2007, 23, 170.
66. Kruk, M.; Jaroniec, M.; Ryoo, R.; Kim, J. M. Chem. Mater. 1999, 11, 2568.
67. Kleitz, F.; Kim, T.-W.; Ryoo, R. Langmuir 2006, 22, 440.
68. Fan, J.; Yu, C.; Wang, L.; Tu, B.; Zhao, D.; Sakamoto, Y.; Terasaki, O. J. Am.
Chem. Soc. 2001, 123, 12113.
69. Sakamoto, Y.; Kaneda, M.; Terasaki, O.; Zhao, D. Y.; Kim, J. M.; Stucky, G.;
Shin, H. J.; Ryoo, R. Nature 2000, 408, 449.
70. Kaneda, M.; Tsubakiyama, T.; Carlsson, A.; Sakamoto, Y.; Ohsuna, T.; Terasaki, O.
J. Phys. Chem. B 2002, 106, 1256.
71. Sakamoto, Y.; Kim, T.-W.; Ryoo, R.; Terasaki, O. Angew. Chem. Int. Ed. 2004, 43,
5231.
72. Sakamoto, Y.; Daz, I.; Terasaki, O.; Zhao, D.; Perez-Pariente, J.; Kim, J. M.;
Stucky, G. D. J. Phys. Chem. B 2002, 106, 3118.
73. Sakamoto, Y.; Han, L.; Che, S.; Terasaki, O. Chem. Mater. 2009, 21, 223.
74. Han, L.; Sakamoto, Y.; Che, S.; Terasaki, O. Chem. Eur. J. 2009, 15, 2818.
75. Ersen, O.; Parmentier, J.; Solovyov, L. A.; Drillon, M.; Pham-Huu, C.;
Werckmann, J.; Schultz, P. J. Am. Chem. Soc. 2008, 130, 16800.
76. Miyata, H.; Kuroda, K. Adv. Mater. 1999, 11, 857.
77. Che, S.; Lund, K.; Tatsumi, T.; Iijima, S.; Joo, S. H.; Ryoo, R.; Terasaki, O. Angew.
Chem. Int. Ed. 2003, 42, 2182.
78. Tuysuz, H.; Lehmann, C. W.; Bongard, H.; Tesche, B.; Schmidt, R.; Schuth, F.
J. Am. Chem. Soc. 2008, 130, 11510.
79. Ryoo, R.; Ko, C. H.; Kruk, M.; Antochshuk, V.; Jaroniec, M. J. Phys. Chem. B
2000, 104, 11465.
150 Mesoporous Silica
80. Kleitz, F.; Czuryszkiewicz, T.; Solovyov, L. A.; Linden, M. Chem. Mater. 2006, 18,
5070.
81. Miyazawa, K.; Inagaki, S. Chem. Commun. 2000, 2121.
82. Kruk, M.; Hui, C. M. J. Am. Chem. Soc. 2008, 130, 1528.
83. Ogawa, M. J. Am. Chem. Soc. 1994, 116, 7941.
84. Ogawa, M. Chem. Commun. 1996, 1149.
85. Lu, Y.; Ganguli, R.; Drewien, C. A.; Anderson, M. T.; Brinker, C. J.;
Gong, W.; Guo, Y.; Soyez, H.; Dunn, B.; Huang, M. H.; Zink, J. I. Nature 1997,
389, 364.
86. Yang, H.; Kuperman, A.; Coombs, N.; Mamiche-Afara, S.; Ozin, G. A. Nature
1996, 379, 703.
87. Yang, H.; Coombs, N.; Sokolov, I.; Ozin, G. A. Nature 1996, 381, 589.
88. Sellinger, A.; Weiss, P. M.; Nguyen, A.; Lu, Y.; Assink, R. A.; Gong, W.;
Brinker, C. J. Nature 1998, 394, 256.
89. Lu, Y.; Yang, Y.; Sellinger, A.; Lu, M.; Huang, J.; Fan, H.; Haddad, R.; Lopez, G.;
Burns, A. R.; Sasaki, D. Y.; Shelnutt, J.; Brinker, C. J. Nature 2001, 410, 913.
90. Trau, M.; Yao, N.; Kim, E.; Xia, Y.; Whitesides, G. M.; Aksay, I. A. Nature 1997,
390, 674.
91. Yang, P.; Deng, T.; Zhao, D.; Feng, P.; Pine, D.; Chmelka, B. F.; Whitesides, G. M.;
Stucky, G. D. Science 1998, 282, 2244.
92. Lu, Y.; Fan, H.; Stump, A.; Ward, T. L.; Rieker, T.; Brinker, C. J. Nature 1999, 398,
223.
93. Fan, H.; Lu, Y.; Stump, A.; Reed, S. T.; Baer, T.; Schunk, R.; Perez-Luna, V.;
Lopez, G. P.; Brinker, C. J. Nature 2000, 405, 56.
94. Doshi, D. A.; Huesing, N. K.; Lu, M.; Fan, H.; Lu, Y.; Simmons-Potter, K.;
Potter, B. G., Jr.; Hurd, A. J.; Brinker, C. J. Science 2000, 290, 107.
95. Pai, R. A.; Humayun, R.; Schulberg, M. T.; Sengupta, A.; Sun, J.-N.; Watkins, J. J.
Science 2004, 303, 507.
96. Shopsowitz, K. E.; Qi, H.; Hamad, W. Y.; MacLachlan, M. J. Nature 2010,
468, 422.
97. Su, B.; Lu, X.; Lu, Q. J. Am. Chem. Soc. 2008, 130, 14356.
98. Fan, R.; Huh, S.; Yan, R.; Arnold, J.; Yang, P. Nat. Mater. 2008, 7, 303.
99. Tolbert, S. H.; Firouzi, A.; Stucky, G. D.; Chmelka, B. F. Science 1997, 278, 264.
100. Miyata, H.; Kuroda, K. J. Am. Chem. Soc. 1999, 121, 7618.
101. Miyata, H.; Kuroda, K. Chem. Mater. 1999, 11, 1609.
102. Miyata, H.; Suzuki, T.; Fukuoka, A.; Sawada, T.; Watanabe, M.; Noma, T.;
Takada, K.; Mukaide, T.; Kuroda, K. Nat. Mater. 2004, 3, 651.
103. Kawashima, Y.; Nakagawa, M.; Seki, T.; Ichimura, K. Chem. Mater. 2002, 14,
2842.
104. Yamaguchi, A.; Uejo, F.; Yoda, T.; Uchida, T.; Tanamura, Y.; Yamashita, T.;
Teramae, N. Nat. Mater. 2004, 3, 337.
105. Wu, Y.; Cheng, G.; Katsov, K.; Sides, S. W.; Wang, J.; Tang, J.; Fredrickson, G. H.;
Moskovits, M.; Stucky, G. D. Nat. Mater. 2004, 3, 816.
106. Liang, Z.; Susha, A. S. Chem. Eur. J. 2004, 10, 4910.
107. Yamauchi, Y.; Suzuki, N.; Kimura, T. Chem. Commun. 2009, 5689.
108. Walcarius, A.; Sibottier, E.; Etienne, M.; Ghanbaja, J. Nat. Mater. 2007, 6, 602.
109. Richman, E. K.; Brezesinski, T.; Tolbert, S. H. Nat. Mater. 2008, 7, 712.
110. Yamauchi, Y.; Nagaura, T.; Ishikawa, A.; Chikyow, T.; Inoue, S. J. Am. Chem. Soc.
2008, 130, 10165.
111. Hara, M.; Nagano, S.; Seki, T. J. Am. Chem. Soc. 2010, 132, 13654.
112. Corma, A. Chem. Rev. 1997, 97, 2373.
113. Ying, J. Y.; Mehnert, C. P.; Wong, M. S. Angew. Chem. Int. Ed. 1999, 38, 56.
114. Maschmeyer, T.; Rey, F.; Sanker, G.; Thomas, J. M. Nature 1995, 378, 159.
115. Yamamoto, T.; Shido, T.; Inagaki, S.; Fukushima, Y.; Ichikawa, M. J. Am. Chem.
Soc. 1996, 118, 5810.
116. Zhou, W.; Thomas, J. M.; Shephard, D. S.; Johnson, B. F. G.; Ozkaya, D.;
Maschmeyer, T.; Bell, R. G.; Ge, Q. Science 1998, 280, 705.
117. Fan, H.; Yang, K.; Boye, D. M.; Sigmon, T.; Malloy, K. J.; Xu, H.; Lopez, G. P.;
Brinker, C. J. Science 2004, 304, 567.
118. Joo, S. H.; Park, J. Y.; Tsung, C.-K.; Yamada, Y.; Yang, P.; Somorjai, G. A. Nat.
Mater. 2009, 8, 126.
119. Shinoda, T.; Izumi, Y.; Onaka, M. J. Chem. Soc. Chem. Commun. 1995, 1801.
120. Hata, H.; Saeki, S.; Kimura, T.; Sugahara, Y.; Kuroda, K. Chem. Mater. 1999, 11,
1110.
121. Iwamoto, M.; Tanaka, Y.; Sawamura, N.; Namba, S. J. Am. Chem. Soc. 2003, 125,
13032.
122. Moller, K.; Bein, T. Chem. Mater. 1998, 10, 2950.
123. Wu, C.-G.; Bein, T. Science 1994, 264, 1757.
124. Wu, C.-G.; Bein, T. Science 1994, 266, 1013.
125. Kageyama, K.; Tamazawa, J.; Aida, T. Science 1999, 285, 2113.
126. Nguyen, T.-Q.; Wu, J.; Doan, V.; Schwartz, B. J.; Tolbert, S. H. Science 2000,
288, 652.
127. Martini, I. B.; Craig, I. M.; Molenkamp, W. C.; Miyata, H.; Tolbert, S. H.;
Schwartz, B. J. Nat. Nanotechnol. 2007, 2, 647.
128. Yang, P.; Wirnsberger, G.; Huang, H. C.; Cordero, S. R.; McGehee, M. D.;
Scott, B.; Deng, T.; Whitesides, G. M.; Chmelka, B. F.; Buratto, S. K.; Stucky, G. D.
Science 2000, 287, 465.
129. Hudson, S.; Cooney, J.; Magner, E. Angew. Chem. Int. Ed. 2008, 47, 8582.
130. Feng, X.; Fryxell, G. E.; Wang, L.-Q.; Kim, A. Y.; Liu, J.; Kemner, K. M. Science
1997, 276, 923.
131. Mal, N. K.; Fujiwara, M.; Tanaka, Y. Nature 2003, 421, 350.
132. Mal, N. K.; Fujiwara, M.; Tanaka, Y.; Taguchi, T.; Matsukata, M. Chem. Mater.
2003, 15, 3385.
133. Qiu, H.; Inoue, Y.; Che, S. Angew. Chem. Int. Ed. 2009, 48, 3069.
134. Lin, H.-P.; Mou, C.-Y. Science 1996, 273, 765.
135. Yang, H.; Coombs, N.; Ozin, G. A. Nature 1997, 386, 692.
136. Kim, J. M.; Kim, S. K.; Ryoo, R. Chem. Commun. 1998, 259.
137. Che, S.; Sakamoto, Y.; Terasaki, O.; Tatsumi, T. Chem. Mater. 2001, 13, 2237.
138. Miyasaka, K.; Han, L.; Che, S.; Terasaki, O. Angew. Chem. Int. Ed. 2006, 45,
6516.
139. Shio, S.; Kimura, A.; Yamaguchi, M.; Yoshida, K.; Kuroda, K. Chem. Commun.
1998, 2461.
140. Ikari, K.; Suzuki, K.; Imai, H. Langmuir 2004, 20, 11504.
141. Suzuki, K.; Ikari, K.; Imai, H. J. Am. Chem. Soc. 2004, 126, 462.
142. Ikari, K.; Suzuki, K.; Imai, H. Langmuir 2006, 22, 802.
143. Garcia-Bennett, A. E.; Lund, K.; Terasaki, O. Angew. Chem. Int. Ed. 2006,
45, 2434.
144. Niu, D.; Ma, Z.; Li, Y.; Shi, J. J. Am. Chem. Soc. 2010, 132, 15144.
145. Suteewong, T.; Sai, H.; Cohen, R.; Wang, S.; Bradbury, M.; Baird, B.;
Gruner, S. M.; Wiesner, U. J. Am. Chem. Soc. 2011, 133, 172.
146. Davis, S. A.; Burkett, S. L.; Mendelson, N. H.; Mann, S. Nature 1997, 385, 420.
147. Xue, C.; Wang, J.; Tu, B.; Zhao, D. Chem. Mater. 2010, 22, 494.
148. Vallet-Reg, M.; Balas, F.; Arcos, D. Angew. Chem. Int. Ed. 2007, 46, 7548.
149. Torney, F.; Trewyn, B. G.; Lin, V. S.-Y.; Wang, K. Nat. Nanotechnol. 2007, 2, 295.
150. Baca, H. K.; Ashley, C.; Carnes, E.; Lopez, D.; Flemming, J.; Dunphy, D.;
Singh, S.; Chen, Z.; Liu, N.; Fan, H.; Lopez, G. P.; Brozik, S. M.; Werner-
Washburne, M.; Brinker, C. J. Science 2006, 313, 337.
151. Zurner, A.; Kirstein, J.; Doblinger, M.; Brauchle, C.; Bein, T. Nature 2007,
450, 705.
152. Cremer, G. D.; Roeffaers, M. B. J.; Bartholomeeusen, E.; Lin, K.; Dedecker, P.;
Pescarmona, P. P.; Jacobs, P. A.; Vos, D. E. D.; Hofkens, J.; Sels, B. F. Angew.
Chem. Int. Ed. 2010, 49, 908.
153. Gartmann, N.; Bruhwiler, D. Angew. Chem. Int. Ed. 2009, 48, 6354.
154. Fang, Y.; Hu, H. J. Am. Chem. Soc. 2006, 128, 10636.
155. Kremer, S. P. B.; Kirschhock, C. E. A.; Aerts, A.; Villani, K.; Martens, J. A.;
Lebedev, O. I.; Tendeloo, G. V. Adv. Mater. 2003, 15, 1705.
156. Choi, M.; Cho, H. S.; Srivastava, R.; Vankatesan, C.; Choi, D.-H.; Ryoo, R. Nat.
Mater. 2006, 5, 718.
157. Cohen, B.; Sanchez, F.; Douhal, A. J. Am. Chem. Soc. 2010, 132, 5507.
158. Choi, M.; Na, K.; Kim, J.; Sakamoto, Y.; Terasaki, O.; Ryoo, R. Nature 2009,
461, 246.
159. Shimojima, A.; Kuroda, K. Chem. Rec. 2006, 6, 53.
160. Sakamoto, S.; Shimojima, A.; Miyasaka, K.; Ruan, J.; Terasaki, O.; Kuroda, K.
J. Am. Chem. Soc. 2009, 131, 9634.
161. Inagaki, S.; Guan, S.; Fukushima, Y.; Ohsuna, T.; Terasaki, O. J. Am. Chem. Soc.
1999, 121, 9611.
162. Asefa, T.; MacLachlan, M. J.; Coombs, N.; Ozin, G. A. Nature 1999, 402, 867.
163. Melde, B. J.; Holland, B. T.; Blanford, C. F.; Stein, A. Chem. Mater. 1999, 11,
3302.
164. Kapoor, M. P.; Inagaki, S. Bull. Chem. Soc. Jpn. 2006, 79, 1463.
165. Hoffmann, F.; Cornelius, M.; Morell, J.; Froba, M. Angew. Chem. Int. Ed. 2006,
45, 3216.
166. Inagaki, S.; Guan, S.; Ohsuna, T.; Terasaki, O. Nature 2002, 416, 304.
167. Fujita, S.; Inagaki, S. Chem. Mater. 2008, 20, 891.
168. Inagaki, S.; Ohtani, O.; Goto, Y.; Okamoto, K.; Ikai, M.; Yamanaka, K.; Tani, T.;
Okada, T. Angew. Chem. Int. Ed. 2009, 48, 4042.
 7.11 Mesoporous Zeolitic Materials
K Moller and T Bein, University of Munich (LMU), Munich, Germany
2013 Elsevier Ltd. All rights reserved.

7.11.1 Introduction 247

7.11.2 Postsynthetic Strategies: Top-Down Synthesis of Mesoporous Materials Through
Leaching of Preformed Zeolites 248
7.11.2.1 Mesopore Formation Through Dealumination by Acid Exposure 249
7.11.2.2 Mesopore Formation Through Desilication by Exposure to Bases 249
7.11.2.3 Desilication Combined with Surfactant-Supported Mesopore Formation 257
7.11.3 Bottom-Up Synthesis of Hierarchical Zeolites by Dual-Templating Methods 258
7.11.3.1 Hard Templating 259
7.11.3.2 Soft Templating with Polymers and Molecular Additives 261
7.11.3.3 Silylated Alkyl Chains as Mesostructuring Templates 264
7.11.3.3.1 Cationic organosilane surfactants 264
7.11.3.3.2 Seed silanization with neutral alkylamino siloxanes 268
7.11.3.3.3 One-step reactions with alkyl-containing silylating agents 270
7.11.4 Bottom-Up Synthesis of Hierarchical Zeolites with Single Templates 271
7.11.4.1 Bifunctional Structure-Directing Agents 271
7.11.4.2 Mesopore Formation Through Nanoparticle Assembly with Monofunctional Templates 274
7.11.4.2.1 Nanoparticle aggregation using single templates 275
7.11.4.2.2 Aggregation of nanoparticles by seeds or gel pretreatment 278
7.11.4.2.3 Steam-assisted conversion and other specialized methods 279
7.11.4.2.4 Nanoparticle aggregation from concentrated gels 280
7.11.5 Conclusion 282
Reference 282

7.11.1 Introduction
Classical zeolites are microporous crystalline three-dimensional
(3D) aluminosilicate networks of low density featuring well-
defined pores and cavities on the nanometer scale and large
surface areas. Highly regarded as multifunctional molecular
sieves and catalysts, they are considered indispensable for the
chemical industry based on their high thermal and chemical
stabilities and their optional strong acidity. The value of zeolites
for industrial processing has been appreciated already about
50 years ago when systematic studies on synthetic zeolites
started in academia in the group of Richard Barrer and in the
industry by researchers including Robert Milton, Donald Breck,
George Kerr, and Edith Flanigen. In addition to their large-scale
production as ion exchangers for detergents, zeolites have
played an ever-more important role in petroleum processing
and they are the basis for many of the reactions performed by
heterogeneous catalysis today. This tremendous success story is
based on the specific crystalline structure and composition of
zeolites endowing them not only with ion-exchange properties,
but also with molecular sieving functions exploited in gas sep-
aration and liquid- or gas-phase catalytic processing, strong
adsorption capabilities used in waste treatments, and tunable
Brnsted acidity, just to name their most prominent features.
Most large-scale alkylation, aromatization, isomerization, crack-
ing, or condensation reactions are performed with zeolite cata-
lysts, thus showing their enormous economic impact.
Nevertheless, certain limitations presented by the micropo-
rous nature of zeolites have also challenged researchers for
decades. Small-, medium-, and large-pore zeolites are character-
ized by their 8, 10, and 12 or more TO4 elements (with
T mainly Si, Al, and P, and to a smaller extent Ge, Ga, Ti, and
V), defining accessible pores of effective dimensions of about 0.4,
0.55, and 0.74 nm, thus allowing the entry and escape of a large
number of small-sized reagents, but excluding the processing of
bulkier molecules. Furthermore, extended porous networks
impose diffusional limitations on the reacting molecules, reduc-
ing the mass transport to and away from the reactive sites in the
pore system and opening time windows for unwanted secondary
reactions and coke formation. A whole new range of chemical
processing options would become available if these limitations
were overcome while simultaneously preserving the desirable
properties of the well-established zeolites.
In the last decade, much effort has therefore been undertaken
to implement a new property into zeolitic materials: adding the
dimension of mesoporosity to their inherently microporous
network. The resulting hierarchical zeolites will feature new
pathways for faster molecular diffusion and will allow access of
previously excluded larger molecules to the catalytically active
zeolite surface. This goal can be approached in different ways: (1)
by trying to synthesize zeolites with larger pores or even hierar-
chical pore systems directly, (2) by trying to use pre-made meso-
porous amorphous materials and transform their existing walls
into zeolites, (3) by using low-dimensional zeolite sheets or
subcrystalline zeolitic precursors and arrange them into meso-
porous materials, or (4) by introducing the secondary pore
system of larger dimensions by combining the zeolite synthesis
with an additional mesostructuring process.

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00716-6 247

248 Mesoporous Zeolitic Materials
The drive to synthesize zeolites with ever-larger crystalline
pores is an ongoing research effort that often relies on using
increasingly complex organic structure-directing agents
(OSDAs). However, discoveries of structures with pores
defined by more than 12 T-atoms were mainly made with
zeolitic alumino-phosphates, with VPI-5 (framework type
VFI; zeolites are classified according to their unique topology
by a 3-letter code indicating different framework types. Mem-
bers of a framework type are distinguished by their materials
composition, carrying individual names) synthesized in the
group of Davis being a prominent member with a pore open-
ing of 18 T-elements, an 18-membered ring (18R) in the late
1980s.1 Today, the record is held by the group of Corma,
who discovered the 28R silicogermanate ITQ-43, featuring
crystalline micro- as well as mesopores in the same structure
with the largest pore dimension being 2.4 nm.2 Unfortunately,
the discovery of such extra-large-pore zeolites is not straight-
forward and these structures are usually not as stable as
common aluminosilicates. Presently, over 200 different zeolite
framework types (differing only in topology, not accounting
for additional variations in composition or morphology)
of aluminosilicate as well as metal-substituted zeolites are
recognized.3 The efforts of synthesizing extra-large-pore
zeolites have been reviewed by Corma and coworkers.4
Much excitement was caused by the discovery of the meso-
porous M41S materials in 1992 by Mobil researchers. Shifting
the templating strategy away from the classical small OSDAs
used in zeolite synthesis to charged or uncharged supramolec-
ular assemblies created a whole new family of porous struc-
tures that were readily made through chemical processes based
on solgel chemistry.5 Pore sizes far beyond 2 nm became
accessible and could be controlled in size and topology in a
systematic manner, but the silica (or related oxide) walls lacked
the crystalline nature of zeolites. This important feature had
consequences with respect to a lower thermal and chemical
stability as well as the lack of strong acidity. Therefore, the new
mesoporous materials did not fulfill the requirements for
many catalytic applications. Hence, much work, initiated by
the group of Pinnavaia, has been devoted to implementing
zeolitic building blocks into the mesoporous walls to render
them more stable and catalytically active (zeolitization) while
simultaneously retaining their mesoporous structure.6 How-
ever, the introduction of zeolite templates or pre-made zeolite
seeds together with surfactant templates into the reaction mix-
ture often resulted in phase separation and the formation of
composite structures. Although intensive research in this area
has resulted in materials of much higher catalytic activity than
the purely amorphous mesoporous structures, the strong acid-
ity and catalytic activity of zeolites are still unsurpassed.7
Another route to form mesostructured zeolites with acces-
sible crystalline surfaces might be envisioned via the assembly
of 2D zeolitic sheets or small zeolite precursor entities such as
zeotiles or zeogrids. The recently discovered MWW zeolite
family has shown that layered precursors either can form the
3D MCM-22 zeolite or can be transformed into related lamel-
lar or pillared derivatives featuring additional larger pores.
Several new, layered precursors of other zeolite structures
have been found recently, thus opening up possibilities for
being restructured into mesoporous zeolite assemblies.8
While these 2D zeolites are truly crystalline in the sense of
showing x-ray diffraction, however broad, in the wide-angle
region, zeogrids and zeotiles lack the presence of an extended
lattice. They are, for example, small MFI-type precursor units
that can be assembled via supramolecular templating resulting
in mesostructures featuring only small-angle diffraction.9
These mesoporous zeolitic precursor assemblies can thus be
envisioned as an intermediate between amorphous mesopor-
ous aluminosilicates and hierarchical zeolites embracing
micro- as well as mesoporosity.
In this chapter, we restrict our discussion to truly mesoporous
zeolitic materials, excluding micro-/mesoporous composites. We
focus our attention on recent developments in synthetic
approaches for combining hierarchical features with micropo-
rous zeolites. Two strictly distinct methods are highlighted: the
postsynthetic destructive (leaching) treatments performed on
pre-made zeolites and the constructive methods developed for
directly creating hierarchical zeolites. The latter include a number
of different strategies. Hard scaffolds are used to leave a direct
imprint of the preformed template after its removal by calcina-
tion, forming mostly intrazeolite mesopores. Soft-templating
methods include the use of polymers, building a separate phase
from the zeolite gel as well as silylating agents that form protec-
tive layers around zeolite precursor units, thus generating inter-
particle secondary pores. We describe the new approach of using
bifunctional templates that enable zeolite growth in restricted
domains and simultaneously induce the formation of more or
less ordered secondary pores. The above-mentioned constructive
methods use dual-templating or multifunctional templating
approaches. Alternatively, mono-templating routes rely only on
micropore-directing agents; here the mesostructures result from
special gel treatments or gel-conversion methods. In many cases,
the development of these synthetic approaches is directed
toward an improvement of the catalytic properties of zeolites,
and often model reactions are used to evaluate catalytic benefits
generated by the new morphologies. We occasionally mention
results of such test reactions; additionally, an excellent review on
the impact of hierarchical zeolites on catalysis is available.10
Furthermore, the journal Catalysis Today has dedicated a special
volume (Vol. 168) to this topic.
7.11.2 Postsynthetic Strategies: Top-Down
Synthesis of Mesoporous Materials Through
Leaching of Preformed Zeolites
Creating mesopores through leaching procedures principally
constitutes a destructive method, intended to etch intercon-
nected pores directly into the single-crystal bodies. Starting
from pre-made zeolites, usually micron-sized crystals are
exposed to corrosive media in form of acids or bases to dissolve
part of their microporous framework, thus resulting in intra-
zeolite mesopores. Alumina is predominantly extracted with
strong acids such as HCl, HNO3, or HF or mild acids such as
oxalic acid, or with complexing agents. Desilication is usually
performed with NaOH or recently also with large structure-
directing alkyl ammonium cations. Inherently, the nature of
the zeolite as well as the nature of the corrosive medium will
determine the degree of dissolution, while the resulting meso-
pore geometry is not easily controlled and three-dimensionally
connected pore networks are not necessarily obtained. The

duration of the treatment, temperature of the reaction, as well
as preceding treatments performed on the zeolites will all have
an influence on the degree of framework extraction, the size,
penetration depth, and the location of the resulting pores.
Usually, the Si/Al ratio will change and consequently the num-
ber of acid sites and possibly the acid strength of the remaining
sites. Some fraction of microporosity is usually lost when
compared to the parent material. The roughening of the crystal
faces typically leads to an increase of the external surface area
that is decorated with large numbers of exposed silanol groups.
Regarding the large range of possible reaction parameters,
leaching can offer great flexibility for creating custom-made
reactive zeolite surfaces. Processing appears to be straightfor-
ward since the starting material is typically pre-made micron-
sized zeolite material. However, as stated above, the destructive
character of this method generally leads to a substantial loss of
starting material. Table 1 gives an overview of the applied
leaching protocols and the catalytic test reactions performed
on the resulting materials, when available.
7.11.2.1 Mesopore Formation Through Dealumination
by Acid Exposure
Probably the oldest method for creating mesopores has been
practiced since the early 1980s, especially with zeolite Y
(framework type FAU).52 Steaming at elevated temperatures
hydrolyzes aluminum-containing sites from lattice positions,
thus forming amorphous debris inside and outside of the
zeolite framework, leaving behind more or less open secondary
pores. The concomitant self-healing of the generated defect
sites through condensation reactions finally enhances the ther-
mal stability of the product, appropriately called ultrastable Y
or short USY. This treatment is still used to implement meso-
pores into zeolite Y,53 but an additional treatment of mild acid
washing is often added in order to remove the newly created
extra-framework alumina deposits.
A recent excellent overview about synthesis strategies for
mesopore formation in zeolites by Chal et al. recalls the differ-
ent strategies for selective removal of alumina from zeolites.54
One can distinguish dealumination induced by simple calcina-
tion, by steaming in combination with a mild acid wash, as well
as the immediate extraction using strong acids. These methods
were also applied for zeolites other than faujasites, such as
mordenite, beta, or ferrierite. Foremost, dealumination is used
to reduce the number of acid sites in zeolites, which are made
responsible for carbon deposition and shortening of the catalyst
lifetime. For instance, mordenite (MOR) was especially treated
with mild acetic acid to carefully dissolve alumina from the
framework without drastically influencing the micropore struc-
ture. The average micropore size was largely unaffected even
with an extended exposure to the acid while the zeolite acidity
decreased as intended. This effect was made responsible for the
observed increase in lifetime during cumene alkylation; how-
ever, 40-nm secondary pores were formed simultaneously.55 It is
therefore often difficult to identify a clear causeeffect relation-
ship for the observed improvements in catalytic performance.
The mesopore formation through dealumination is very much
dependent upon the type of zeolite structure, as shown by
Gonzales et al. Studying beta (BEA), ZSM-5 (MFI), and morde-
nite, only the latter experienced a major increase in pore volume
Mesoporous Zeolitic Materials 249
due to 40-nm pores created upon conventional or microwave
heating with 6 M HCl at 100  C for 15 min.56 Some dealuminat-
ing agents create mesopores of a different kind: Qin e al. exposed
Y zeolites to ammonium hexafluoridosilicate (AHFS), which
resulted only in a mesoporous amorphous silicate layer around
the crystal surface, leaving the overall internal integrity of the
crystallites unaffected.57 By contrast, when this dealumination
process was performed after a treatment with NaOH, evenly
distributed mesopores of about 30 nm were obtained, thus
increasing the catalyst reactivity toward cumene conversion.
The authors speculate that silanol nests were created during the
base treatment that allowed access of the AHFS reagent, resulting
in a mesoporous zeolite Y with homogeneous silica/alumina
distribution (see Figure 1). The comparably small external sur-
face area of 61 m2 g1 created here (parent material: 18 m2 g1)
was markedly increased by more complex desilication proce-
dures, as will be described later.
The exploitation of defect sites for framework extraction was
performed via a very different route by Valtchev et al.58 They
used 238U ion-beam irradiation on oriented single crystals of
ZSM-5 to create parallel, 1D defect zones of different density.
Subsequent chemical attack with diluted HCl uniformly
removed Si/Al from the framework, thus creating diffusion
pathways of 50 nm without deteriorating the microporous envi-
ronment. This model system nicely shows the dominating effect
of mesopores for enhancing the chemical diffusion, as the over-
all catalytic conversion of m-xylene was markedly improved
while the selectivity of the original zeolite was maintained.
7.11.2.2 Mesopore Formation Through Desilication
by Exposure to Bases
The removal of silica species (desilication) has become a dom-
inant way to create mesopores in zeolites by leaching. In this
case, mesopores are usually formed without affecting the acid-
ity of the respective zeolite, an important aspect for most
catalytic reactions. Especially the group of Perez-Ramirez has
studied the desilication on various zeolites in depth and has
brought this process close to industrial realization.59 Mecha-
nistic aspects of desilication were reviewed in early 2011.60 We
will highlight major achievements in this area and will com-
plement the existing overview with the latest developments.
A compilation of experimental studies on desilication is given
in Table 1, documenting the specific type of reaction, the
zeolite used, and the catalytic test reaction investigated, when
available. It includes publications where desilication is supple-
mented by a template-assisted transformation of the zeolite
debris as described in Section 7.11.2.3.
Apart from the early patent literature by Young in 1967,
who observed increased adsorption of olefins in base-treated
mordenite, it took about 30 years before the existence of meso-
pore formation upon leaching was documented by nitrogen
sorption.60 Early studies had claimed that sodium carbonate
and sodium hydroxide solutions would not deteriorate the
x-ray crystallinity during desilication in ZSM-5 and silicalite
(MFI). These findings were supported by Matsukata and his
group, but they documented that about 40% of the silicon was
dissolved from the original zeolite crystals, thus leaving strong
indentations in the form of surface roughening on the zeolite
250 Mesoporous Zeolitic Materials

Table 1 Overview of synthetic procedures developed for top-down processing (leaching) (loosely grouped according to zeolite phases)

Year and Method Zeolite structure Si/Al ratio Mesopore Catalysis performed (special
reference (type, pore before size (nm) findings)
dimension) treatment

Desilication
2000 11
0.2 M NaOH, 80  C, 300 min ZSM-5 (MFI, 3D) 39 4 Cumene cracking
200412 0.11 M NaOH, 2575  C, 0.2510 h ZSM-5 37 10 (MFI)
BEA (3D) 122 6 (BEA)
MOR (1D) 45
FER (2D) 28
200513 0.2 M NaOH, 65/85  C, 30 min ZSM-5 151000 1020 Optimum Si/Al range 2050
200514 0.2 M NaOH, 65  C, 1530 min Al zoned ZSM-5 2050 4 Pore-directing role of Al
201015 0.12.0 M NaOH, 90  C, 300 min ZSM-5 38 4
200916 1 M, 0.5 M NaOH, 85 , 100  C, 1 h ZSM-5 40 n.a. Liquid-phase transformation of
ZSM-12 (MTW, 1D) 40 alpha-pinene;
MCM-22 (MWW, 40 MCM-22 not stable
2D)
200917,18 0.2 M NaOH mixed with TPAOH or TBAOH, ZSM-5 42 310 Hierarchy factor
65  C, 30 min or 1 M TPAOH, 65  C, 8 h Pore-growth moderators
Alkylation of benzene
200919 0.2 M NaOH, 65  C, microwave heating at ZSM-5 40 10
150 W for 35 min
201020 (a) 0.2 M NaOH, 65  C, 30 min ZSM-5 2550 8 Isomerization of o-xylene
(b) 0.1 HCl, 70  C, 6 h
201021 (a) 0.21 M NaAlOH, 6580  C, 0.53 h ZSM-5 1540 5 Higher yield of ZSM-5, up to 95%
(b) 0.13 M HCl, 6580  C, 3 h
201122 NaOH, HCl treatment, and SH ZSM-5 26 10 Hydrolysis of nitrophenylbutyrate
functionalization for lipase anchoring
201123 (a) 0.11.8 M NaOH, 65  C, 30 min ZSM-5 101000 10 Alkylation of toluene
(b) 0.020.1 M HCl, 65  C, 6 h
201124b (a) 0.11.5 M NaOH, 70 , 1 h H-ZSM-5-Ru 26 3.87.2 FischerTropsch
(b) RuCl3 impregnation and reduction in H2
201125 NaOH Al(OH)4 or Ga(OH)4, TMA or Silicalite-1 (MFI, 3D) 1320 Pore-directing agents
TPA
201126 Template-containing zeolite aggregates ZSM-5 280 Etherification of propylene glycol
0.21 M NaOH, 70  C, 2060 min ZSM-12
BEA
200827 0.2 M NaOH, 2565  C, 1060 min BEA 35 24 Alkylation of benzene (only 10%
conversion compared to parent
beta)
200928 Partial detemplation by calcination at 230, BEA 710 Pyrolysis of LDPE
330 or 550  C
0.2 M NaOH, 65  C, 30 min
200929 O.1 M TMAOH, 5100  C, 30360 min BEA 35 1020
201030 0.05, 0.1 M NaOH, RT, 15 min USY (FAU, 3D) 28 3/30 Hydrocracking of n-hexadecane
on Pt-loaded samples
201231 H4EDTA, Na2H2EDTA, citric acid, HCl, Y, USY 2.430 210 Alkylation of benzyl alcohol
TPAOH, NaOH Pyrolysis of LDPE
201232 0.2 M NaOH TPA, TEA, TMA or a variety USY 15385 6 Continuous-flow reactor
of surfactants, 65  C, 30 min BEA 220 4
200933 0.5 M NaOH, 80  C 3 h (dealumination by FER (2D) 29 >100 Pyrolysis of LDPE
steaming or oxalic acid unsuccessful)
201034 (a) 1 M NaAlO2, 6080  C, 0.56 h FER n.a. Pyrolysis of LDPE
(b) 3 M HCl, 80  C, 3 h Pd dispersion
(c) 0.2 M NaOH, 60  C, 0.5 h
201035 0.20.4 M NaOH, 75  C, 2 h SSZ-13 (CHA, 3D) 14 210 Styrene oligomerization, MTO
201036 0.2 M NaOH, 65  C, 30 min SSZ-35 (STF, 1D) 33 n.a. p-Xylene alkylation
201037 0.050.4 M NaOH, 4565  C, 30 min ITQ-4 (IFR, 1D) 32 36 Pyrolysis of LDPE
201138a (a) 0.2 M NaOH, 65  C, 30 min ZSM-22 (TON, 1D) 42 n.a. Pt-modified: n-octane
(b) 1 M HCl, 65  C, 360 min hydroisomerization
201239 0.4 or 0.6 M NaOH, 80  C, 15120 min NU-10 (TON, 1D) 33 2020 nm o-Xylene isomerization
200640 0.050.4 M NaOH, 3595  C, 30150 min ZSM-12 (MTW, 1D) 31500 1520

(Continued)
 Mesoporous Zeolitic Materials 251

Table 1 (Continued)

Year and Method Zeolite structure Si/Al ratio Mesopore Catalysis performed (special
reference (type, pore before size (nm) findings)
dimension) treatment

200741 0.2 M NaOH, 65  C, 30 min MOR (1D) 2030 10 Alkylation of benzene

200842 0.051 M NaOH, 4585  C, 15300 min AST (0D) 31 5 Pyrolysis of HDPE
201124d 0.2 M NaOH, 2 h GaZSM-11 Ga/Si 94 13 Pt dispersion, aromatization of
(MEL, 3D) propane
201143 0.2 M NaOH, 65  C, 0.25 h TUN (3D) 20 n.a. Toluene alkylation

Surfactant-assisted desilication
200544 (a) 0.33.0 M NaOH, RT, 30 min CTAB, MOR/MCM-41 18 2.5 25 Hydrodealkylation of heavy
30 min aromatics
(b) 100  C 24 h
(c) pH ! 8.5, 100  C, 24 h
200645 (a) 0.45 M NaOH CTAB, RT, 60 min USY, ZSM-5 -> 40, 40 2.5 Cumene conversion
(b) 100  C, 24 h MCM-41
201246 (a) 0.10.65 M NaOH CTAB, RT, 2 h Ga,Al-ZSM-5/MCM- Ga,Al/ 3 Aromatization of propane
(b) 100  C, 24 h 41 Si 9.8
(c) pH adjustment
(d) 100  C 24 h
201147 (a) 2.0 M NaOH, RT, 30 min ZSM-5/MCM-41 50 4 Cracking of triisopropylbenzene
(b) CTAB, 50  C, 1 h BEA/MCM-41 25 3.6
(c) 110  C, 24 h ZSM-5/SAPO-5/
(d) Reduction of pH and add SAPO-5 MCM-41
(d) 110  C, 24 h, SAPO-5 added at (c)
201048 (a) 0.09 M TMAOH CTAB, RT, 30 min Y/MCM-41 16 4
(b) 150  C, 24 h
201149c (a) 0.41.2 M NaOH, RT MOR/MCM-41 48 3 Alkylation of naphtalene
(b) CTAB 100110  C, 24 h FER/ MCM-41 27 3 1-Butene isomerization
BEA/MCM-41 42 3
201250 (a) 0.37 M NH4OH C8 to C16TAB 20 min Y (also: MFI, FAU, 15, 2.4 2.54.5 Catalytic cracking of vacuum
(b) 150  C, 10 h (also: NaOH, Na2CO3, MOR) gas oil
TPAOH) FCC
201251 CTAB NaOH, 100  C, 24 h ZSM-5 100, 50, 30 3, 1030

FAU

NaOH
PLUS
AHFS AHFS
0.05
dV/d log(D)

dV/d log(D)

0.05

4 10 100 20 nm 10 30 100 20 nm
Pore diameter/nm Pore diameter/nm

Figure 1 Acid treatment with ammonium hexafluoridosilicate (AHFS) of faujasite creates a mesoporous silica shell (left), while intrazeolite mesopores
form after consecutive application of NaOH and AHFS (right). Reprinted from Qin, Z.; Shen, B.; Gao, X.; Lin, F.; Wang, B.; Xu, C. J. Catal. 2011, 278,
266275. Copyright (2011), with permission from Elsevier.

morphology as seen in scanning electron microscopy (SEM)
micrographs.11 They also observed the positive implications of
mesopore formation during cumene cracking.61 More system-
atic investigations of the details of silica depletion in silica-rich
commercial zeolites such as MFI, BEA, MOR, and FER started
in 2004 with the work by Groen et al.12 They studied the
influence of NaOH concentration, reaction temperature and
time on pore formation by comparing sorption data of
untreated and treated samples. All zeolites showed an increase
in mesopore surface area after base exposure, but the reaction
conditions had to be adjusted for each type of zeolite. Simul-
taneously it was observed that this process was accompanied
by a substantial loss of micropore volume.
The group of Perez-Ramirez soon recognized that the Si/Al
ratio of the respective zeolites plays a major role in the success-
ful pore formation and that the amount and location of alu-
minum in the lattice can direct the extraction of silica.13 An
optimal Si/Al range of 2550 for desilication was identified by
comparing ZSM-5 samples of increasing aluminum content. It
was suggested that in aluminum-rich samples extraction of
silica is reduced through a shielding effect of the negatively
charged AlO4 tetrahedra, while in silica-rich samples this
shielding is obviously less abundant. Consequently, uncon-
trolled excessive Si extraction can lead to wide pore-size distri-
butions and extremely low yields through massive dissolution
in the latter case. The overall acidity did not change drastically
252 Mesoporous Zeolitic Materials

in samples of intermediate Si/Al ratio, indicating a preservation
of the aluminum content; therefore, a reinsertion of the Al
species close to the pore surface was proposed. Because of its
strong influence the aluminum was labeled a pore-directing
agent (PDA). The major impact of the aluminum distribution
was nicely visualized with SEMenergy-dispersive x-ray spec-
troscopy (EDX) and 3D transmission electron microscopy
(TEM) micrographs on Al-zoned ZSM-5. A comparison with
untreated ZSM-5 (sample SZ-nt) shows the striking introduc-
tion of internal mesopores through alkaline treatment
while the aluminum-rich outer surface is much less affected
(see Figure 2, sample SZ-at).14 Favorable reaction conditions
for desilication of ZSM-5 with medium Si/Al range were
identified: a treatment in 0.2 M NaOH was used at 65  C
for 30 min.
Recently, it was shown by the Perez-Ramirez group that
ZSM-5 mesoporous zeolites, desilicated according to this
method, retained their intrinsic structure and acidity even
when pelletized with an attapulgite clay, commonly used as a
binder in industrial packaging of catalyst powders.62
A systematic study by Zhao et al. addressing the influence of
the NaOH concentration on leaching of ZSM-5 with a Si/Al
ratio of 38 showed the inverse correlation of increasing meso-
pore formation and loss of micropore volume and crystallinity.
A maximum concentration of 0.4 M NaOH was recommended
for alkali leaching when performed under more drastic condi-
tions at 90  C for 300 min.15 The less aggressive standard
base treatments of the Perez-Ramirez group can be substan-
tially shortened to 34 min by using microwave heating to
create mesoporous ZSM-5.19

Vads/cm3 STP g1
1 m

400
SZ-at30

300
SZ-at15

200

SZ-nt

p/p0/-
(a) (b) 0.2 0.4 0.6 0.8

100 nm

(c) (d)

Figure 2 Desilicated small ZSM-5 zeolites (400700 nm, SZ) through alkaline leaching. (a) SEM micrograph of nontreated sample SZ-nt. (b) Nitrogen-
sorption isotherms of SZ-nt and alkaline treated samples for 15 and 30 min (SZ-at15 and SZ-at30). (c) 3D-TEM virtual cross sections through the
reconstruction of an alkaline-treated crystal. (d) Surface rendering of a section of an alkaline-treated crystal; zeolite material red, porosity green.
Reprinted with permission from Groen, J. C.; Bach, T.; Ziese, U.; Paulaime-van Donk, A. M.; de Jong, K. P.; Moulijn, J. A.; Perez-Ramirez, J. J. Am. Chem.
Soc. 2005, 127, 1079210793. Copyright (2005) American Chemical Society.

Sodium aluminate solution was studied as another reagent
for desilication of ZSM-5.21 It turned out that Si leaching
occurred to a much lower degree when compared to NaOH,
most likely because aluminum hydroxide deposits at the zeo-
lite surface hindered a more severe extraction. When these
deposits were removed by a secondary acid wash with HCl or
oxalic acid, a comparable mesopore surface area as in the
NaOH-leached reference sample was generated. An advantage
of this procedure was seen in the greatly reduced loss of zeolite
material, which additionally led to a smaller mesopore size of
about 5 nm. A shielding function was also established for other
metal hydroxides or tetraalkylammonium cations when stud-
ied with the silica-rich silicalite-1. Usually, zeolites with a high
Si/Al ratio experience massive dissolution and a gain in surface
area is only obtained from surface roughening and/or particle
deagglomeration when the standard 0.2 M NaOH solutions
are used for desilication, while mesopore formation is hardly
present. However, the addition of low concentrations of
Al(OH)4, Ga(OH)4, and tetramethyl- or tetrapropyl- ammo-
nium (TMA or TPA) to the alkaline solution served well in
protecting the surface of silicalite-1 and hence these reagents
were labeled PDAs.25 Now, mesopores of about 10 nm were
generated with higher yields of the mesoporous zeolites. It is
assumed that the metal hydroxides served as protectors
through partial reinsertion into the zeolite framework or by
forming an amorphous oxide layer. The structure-directing
agents (SDAs) on the other hand seem to act as a shield
through their high adsorption affinity toward the zeolite sur-
face, with TPA having the strongest effect.
ZSM-5 with lower Si/Al ratio can also itself serve as an
internal PDA source by generating amorphous alumina-rich
debris on the mesopore and outer zeolite surface when NaOH
Morphology and
crystal characteristics
- No/little intergrowths/defects
- Si/Al = 2050 (required)
- Few intergrowths/defects
- Si/Al = 2050 (preferred)
- Many intergrowths/defects
- Si/Al less important
Internal defects
Figure 3 Overview over different desilication mechanisms depending on the morphology of the parent compound. Reprinted from Svelle, S.; Sommer,
L.; Barbera, K.; Vennestrom, P. N. R.; Olsbye, U.; Lillerud, K. P.; Bordiga, S.; Pan, Y.-H.; Beato, P. Catal. Today 2011, 168, 3847. Copyright (2012), with
permission from Elsevier.
Mesoporous Zeolitic Materials 253
is used. In a very detailed study, Verboekend et al. screened the
optimum conditions for generating mesopore surface area in
alumina-rich as well as in silica-rich ZSM-5, outside of the
established optimum Si/Al range.23 By applying an additional
treatment with HCl on alumina-rich ZSM-5 or using PDAs on
high-silica samples, they could extend the overall accessible
Si/Al range for desilication from 2050 to 121000.
It was documented by Svelle et al. that not only composi-
tion but also morphology might have a tremendous effect on
the desilication mechanism.63 Comparing single crystals and
aggregated ZSM-5 particles after exposure to identical desilica-
tion procedures led the authors to assume two major routes of
mesopore formation: one being controlled by the Al-directed
dissolution and the other mainly dominated by crystal bound-
aries. Fused particles with grain boundaries and defects were
much more susceptible to etching. A summary of the different
morphologies formed with varying compositions is given in
Figure 3, which is mainly derived from studies on ZSM-5.
However, the authors assert that these findings should also
be relevant for other framework structures.
Recent publications envision new areas of application for
mesoporous ZSM-5 by adding additional functionality to
10-nm-sized mesopores. Consecutive desilication and surface
silanization with mercapto- or aminopropyltrimethoxysilane
allowed for the anchoring of a lipase-type enzyme (approxi-
mate size 4 nm) within the interconnected mesopores of the
zeolite, even at a higher content than in the 1D pores of simi-
larly functionalized mesoporous silica (MCM-41). The thiol-
functionalized ZSM-5 was the most active biocatalyst for the
hydrolysis of p-nitrophenyl butyrate, possibly due to enzyme
stabilization by formation of disulfide bridges with the func-
tionalized zeolite support.22 Other new applications include
Desilication
characteristics
OH
- Mesopores from AI distribution
OH
- Mesopores from AI distribution
and intergrowths/defect removal
OH - Mesopores predominantly from
intergrowths/defect removal
External surface
254 Mesoporous Zeolitic Materials
the ion exchange of mesoporous ZSM-5 with large cobalt
molybdenum polyoxometalates ions ([H6CoMo6O24]3) to
create bifunctional catalysts featuring close proximity of hydro-
genation sites (CoMo6) and catalytic cracking sites (Brnsted
sites) for an effective performance in hydrodesulfurization.64
The studies on desilication with NaOH have been mostly
focused on the industrially important zeolite ZSM-5, but some
reports also address other structures. Significant advantages
are envisioned with more restricted, 1D zeolites. Laboratory-
synthesized large mordenite crystals with Si/Al ratios 20 and 30
were desilicated and it was demonstrated that, in contrast to
dealumination procedures, mesoporosity could be introduced
while preserving the Brnsted acidity. The positive effect of
additional pores in these 1D zeolites was reflected in improved
catalytic performance in the alkylation of benzene. Addition-
ally, less coke formation was observed.41 Wei et al. performed a
similar detailed desilication study on 1D ZSM-12 (MTW) pre-
pared with a Si/Al range between 31 and 500; they found the
concentration of NaOH to be most important for the pore
formation.40 They also confirmed the key role of aluminum
in the leaching process for this structure since removal of
silicon was again most efficient in the optimum Si/Al range
between 31 and 58. It was further noted that when the SDA was
present in the micropores, it effectively prevented silicon
extraction. All leached samples showed increased diffusivity
for large hydrocarbon molecules.
Other network structures are also susceptible to desilica-
tion: even clathrasils, 0D aluminosilicates that are usually not
catalytically active, could be fragmented into nanoclusters with
a highly increased intercrystalline pore surface that showed
some activity in the degradation of high-density polyethylene
(HDPE).42 Ferrierite (FER, 2D) is a zeolite that was found to be
highly resistant to demetallation by steaming and acid or base
treatment. Only when using harsh conditions with 0.5 M
NaOH it is possible to increase the mesopore area three to
four times through platelet deaggregation and perforation
with undefined pores; however, the yield decreased down to
about 40% and micropore volume was nearly half.33 A multi-
step demetallation was specially designed to overcome the
restrictions in ferrierite.34 To this end, sodium aluminate was
used in a first step to desilicate part of the zeolite. However,
most of the Si species remained in the solid because precipi-
tated aluminum hydroxide hindered their dissolution and
removal. Hence, an HCl wash was added to remove alumina
debris, and in a third step a mild NaOH wash finally dissolved
the silica nanoparticles. A spray-drying impregnation with Pd
proved to produce a fine dispersion of metal particles in the
final ferrierite sample.
A two-step procedure was also necessary for mesopore for-
mation in the 1D zeolite ZSM-22 (TON).38 This structure poses
a challenge for demetallation because the crystals exhibit a large
fraction of the aluminum content on their outer surface. Thin-
ning and defragmentation of the rod-like crystals occurred
upon base treatment, accompanied by alumina deposition on
the outer surface and almost complete blocking of the micro-
pores. However, a supplemental acid wash regenerated nearly
90% of the original pore volume. This latter problem seems to
be absent in another structure of the same family. When NU-10
(TON) was severely desilicated with more concentrated NaOH
at a higher temperature, the microporosity was not affected
while the mesoporosity increased steadily with time. The
authors believe that the reason can be found in the different
aluminum distribution.39 SSZ-35 (also a 1D zeolite, STF fam-
ily) was either dealuminated or desilicated in order to study the
changes in selectivity toward p-xylene alkylation. It proved to be
amenable to either treatment and in both cases showed an even
greater selectivity to 2,5-dimethyl cumene as compared to the
parent sample, however for different reasons. Dealumination
removed mainly surface aluminum and decreased the surface
acidity responsible for secondary alkylation reactions, while
desilication was presumably efficient due to a shorter contact
time of reactants caused by increased diffusivity.36 ITQ-4 (IFR),
another 1D zeolite, showed also mesoporosity after NaOH
treatment.37
A careful study of the desilication on H-SSZ-13 with CHA
topology (3D) was performed by Sommer et al.35 They
included Fourier transform infrared (FTIR) spectroscopy with
CO as the probe molecule for acid site determination as well as
confocal fluorescence microscopy to follow the access of a
styrene oligomer to the newly created mesopores to assess the
progress of desilication. Contrary to findings with other zeolite
structures, they observed a remarkable decrease of surface area
accompanied by a loss of micropore volume and Brnsted site
density. This was explained by a partial damage of the crystal
structure, likely caused by the large number of four rings in
CHA. The high tension in these small rings had previously
been associated with a low stability of framework aluminum.
Overall, the introduction of mesopores was insufficient to
offset the detrimental changes in surface properties brought
about by desilication, which therefore turned out to be coun-
terproductive for the methanol-to-olefin (MTO) conversion.
Similar observations are made for the desilication of zeolite
beta (see below within this chapter).
The relatively new high-silica structure TUN with ten-ring
pores was nearly unaffected by a dealumination treatment even
with highly concentrated 6 M HNO3, but it formed mesopores
upon desilication with NaOH.43 Here it was found that the
desilication procedure changes the type of acid sites by a reduc-
tion of Brnsted acid sites and a concomitant increase in Lewis
acid sites, which was held responsible for the decrease in
selectivity during toluene alkylation.
Commercial steamed and acid-leached zeolite Y was ana-
lyzed by de Jong et al. using 3D tomography combined with a
special imaging analysis in order to identify the type and
morphology of the created pores, information that cannot be
obtained from sorption experiments in this depth.65 Nitrogen
sorption experiments of this sample indicated the existence of
constricted pores, seen by cavitation effects as a sharp peak
when the pore-size distribution was evaluated from the desorp-
tion branch (see Figure 4, top). Additional mesopores of
2030 nm are evident from both branches. While TEM micro-
graphs indicate mesopores running throughout the crystal,
individual pores can be recognized in a virtual slice of the
reconstructed volume of this crystal (see Figure 4, middle).
Using their advanced imaging analysis enabled the authors
to differentiate between open, interconnected pores and
two types of blocked mesopores, either being closed and
only accessible via micropore diffusion or constricted, hav-
ing a diameter of less than 4 nm, the latter amounting to
about 20% (see Figure 4). These zeolites were additionally
 Mesoporous Zeolitic Materials 255

1.8

Quantity adsorbed (cm3 g1 STP)

320
1.6
1.4 280

240
0.5
200

dV/d log D (cm3 g1)

0.4 160

120
0.3 0 0.2 0.4 0.6 0.8 1.0
Relative pressure PP01
0.2

0.1

0
1 10 100
Pore diameter (nm)

100 nm 100 nm
(a) (b)

467 467
z (nm)

z (nm)

0 0
502 502
y( y(
nm nm
) 177 ) 177
0 0 0 0
x (nm) x (nm)
Figure 4 (Top) N2 physisorption analysis (inset) of acid leached HY30. Pore-size distribution derived from the desorption branch (green) shows
cavitation effects at 4 nm, assigned to closed pores. (Middle) (a) TEM micrograph and (b) a virtual slice reconstructed by electron tomography depicts
channel-like mesopores distributed throughout the crystal. (Bottom) Volume-rendered 3D representations of the open mesopores in green (left)
and the closed mesopores in red (right). Adapted with permission from Zecevic, J.; Gommes, C. J.; Friedrich, H.; de Jong, P. E.; de Jong, K. P.
Angew. Chem. Int. Ed. 2012, 51, 42134217. Copyright (2012) John Wiley and Sons.

base-leached, resulting in a zeolite with trimodal porosity.
Analysis with electron tomography revealed that the original
large mesopores had increased slightly to about 30 nm diam-
eter while a new pore system with about 3 nm pores was visible
throughout the zeolite. Large external surface areas of 339 or
443 m2 g1 were achieved with prolonged treatment, however
accompanied by an increasing loss of micropore volume down
to only 0.06 ml g1.30 A comprehensive study on zeolite Y and
ultrastable USY steamed to different degrees and partially acid-
treated was recently published by Verboekend et al.31 Various
256 Mesoporous Zeolitic Materials
combinations of acid (H4EDTA, Na2H2EDTA, citric acid, HCl)
and NaOH treatments with or without PDAs tetrapropylammo-
nium hydroxide (TPAOH) established reaction conditions for
Y zeolites of Si/Al ranges between 2.4 and 30 that are appropri-
ate to largely retain the zeolite crystallinity (microporosity)
and still generate mesopore surface areas that are very high
(>200300 m2 g1) and can even reach 500 m2 g1 (in USY).
Despite these promising results there is evidence indicating
that simple basic leaching is not necessarily beneficial for the
catalytic behavior of all zeolite types. Using zeolite beta (BEA)
with a Si/Al ratio of 122, far out of the optimal range, had led to
extensive dissolution in previous studies.12 Thus, micron-sized
zeolite beta was intentionally prepared in fluoride medium in
order to obtain a crystal morphology with few defects and a
presumably high chemical stability.27 Nevertheless, when the
standard procedure using 0.2 M NaOH was applied even for a
short time and at lower temperature, dissolution and loss of
microporosity occurred again, generating increasing mesopore
surface areas with proportionally decreasing microporosity.
When this desilicated zeolite beta was tested in the alkylation of
benzene to ethylbenzene, the conversion was dramatically
reduced to 10% of the parent zeolite. Obviously, the crystallinity
and tetrahedrally coordinated aluminum were both strongly
reduced and the Brnsted acidity was nearly lost. It was specu-
lated that the lattice destruction in this case is due to the lower
stability of aluminum species in the highly acidic framework of
zeolite beta compared to MOR or ZSM-5. To overcome this
limitation, Perez-Ramirez introduced the concept of partial
detemplation prior to desilication, relying on the known effect
of retarded dissolution of template-containing zeolites.28 Starting
from the hypothesis that calcination at lower temperature directs
the content as well as location of the remaining micropore-
directing template, silicon extraction was performed after differ-
ent stages of template removal in zeolite beta. As a result, nitrogen
sorption data now revealed a controlled increase in mesopore
surface area, while the loss of microporosity was reduced. The
mesoporous beta samples were tested in the pyrolysis of low-
density polyethylene (LDPE), a large-scale process applied for
recycling plastic materials. It is an acid-catalyzed reaction that is
diffusion limited in virgin zeolite beta due to the long, branched
hydrocarbons involved. In mesoporous beta, a significant
decrease of the breakdown temperature of LDPE was observed
concomitant to the increase in mesopore surface area. However,
the best-performing sample had lost nearly half of its micropore
volume, indicating that this process is strongly controlled by the
accessibility of the external surface of the zeolite domains.
Enhanced diffusion and a relative increase in acid sites through
desilication together might result in this positive trend in
reactivity.
The protective function of the micropore template during
leaching was also exploited by van Iaak et al. They used this
concept for decoupling mesopore formation from changes in
acidity on the zeolites ZSM-5, ZSM-12, and beta that consisted of
agglomerates of small crystallites, thus possessing a large external
surface area.26 Uncalcined zeolites were treated with NaOH for
different times. Some of the small individual crystallites were
supposedly displaced from the aggregated particles under these
conditions, thereby forming interparticle mesopores, thus leav-
ing the overall Si/Al ratio and original micropore volume of the
remaining particles unaffected. Holm et al. introduced a different
desilication agent in order to ease the structural degeneration
observed especially in zeolite beta with NaOH.29 They heated
calcined zeolite beta in the comparatively mild organic base
tetramethylammonium hydroxide (TMAOH), thus preventing
severe dissolution or major reduction of micropore volume.
Instead of exploiting the shielding function of alkylammo-
nium templates when situated within the micropores, bulky
organic bases such as TPAOH or TBAOH whose sizes exceed the
diameter of the micropore openings of ZSM-5 were used by
Perez-Ramirez et al. as pore-growth moderators.17 Used in com-
bination with NaOH at a fixed overall OH concentration of 0.2 M
and a variable TPA(TBA)/OH ratio they were able to steer the
mesopore formation and their diameter in a controlled fashion by
blocking the micropore entrances. This shielding effect was
not noticeable when the alkyl ammonium ion was smaller than
the zeolite micropore diameter, for example, when TMA was
employed. Using the pore-growth moderators, the desilicating
action was limited to the surface of the respective zeolite, thus
diminishing the loss of micropore volume. In order to find a
descriptor for the optimum desilication process independent of
zeolite structure, a hierarchy factor (HF) was introduced. It is
based on sorption data and is the product of the newly created
relative mesopore surface area (Smeso/SBET) and the relative micro-
pore volume (Vmicro/Vtotal) that is still retained after desilication.
A high HF indicates that a large mesopore volume is obtained
without severe reduction in micropore volume. The authors could
correlate the HF with the catalytic activity of mesoporous ZSM-5
in the alkylation of benzene to ethylbenzene. Due to the use of the
large organic bases, the silica dissolution was slowed down sub-
stantially, allowing a better control of the process. When only
TPAOH was used as a pore generator, the optimum leaching
conditions were 1 M TPAOH at 65  C for 8 h.18
The use of PDAs is especially beneficial for zeolites that easily
undergo amorphization under common leaching procedures.
For this reason, Verboekend et al. studied a comprehensive
group of organic bases in addition to anionic, nonionic, and
cationic surfactants as additives in NaOH solution for the
desilication of commercial USY and beta.32 Best results were
obtained with large cationic PDAs containing 1020 C atoms,
such as TPA or the surfactant cetyltrimethylammonium, CTA.
A high mesopore surface area of 578 m2 g1 was obtained for
USY with TPA while mostly preserving the microporosity
(0.28 ml g1, down from 0.33 ml g1 in the parent zeolite)
with a final yield of 56%. Very good results were also obtained
using surfactant molecules, especially CTA for high-silica zeo-
lite beta (Si/Al 220), however caused by a different mechanism.
Here, the usually observed massive amorphization was mark-
edly reduced and the crystallinity was maintained at about 85%.
This surfactant-assisted recrystallization process of zeolite debris
is described in detail in Section 7.11.2.3. The authors also
introduced a continuously operable reactor for desilication
allowing a 100-fold increase in productivity as compared to
the commonly applied batch reactors (see Figure 5).
In this context, another soft but inexpensive desilicating
medium was presented by Vennestrom et al. in the form of
guanidinium hydroxide.66 It is less aggressive than NaOH and
thus requires a 1.6 M solution and reaction for over 5 h at
65  C to create substantial mesopores of about 13 nm in
H-ZSM-5. Compared to a treatment in NaOH, a much higher
yield and no loss of microporosity were observed.

Zeolite Base and PDA
suspension solution
Micro-reactor
Product 6 cm
3
High-shear
disperser
Figure 5 Continuous-flow reactor for the synthesis of mesoporous zeolites through desilication. Adapted with permission from Verboekend, D.; Vile,
G.; Perez-Ramirez, J. Cryst. Growth Des. 2012, 12, 31233132. Copyright (2012) American Chemical Society.
Additional studies with desilicated ZSM-5 specifically
addressed the effect of Brnsted and Lewis acid site distributions
on the catalytic performance. For example, Fernandez et al.
observed an increased conversion with leached mesoporous
ZSM-5 samples in o-xylene isomerization, however at lower selec-
tivity compared to the microporous parent compound.20 The
lower selectivity was attributed to an apparent redistribution of
the aluminum species, thus generating a high number of Brnsted
sites near the mesopore mouths and numerous Lewis acid sites on
the external surface. The latter was assumed to have a detrimental
effect toward p-xylene selectivity. These undesired acid sites could
be eliminated by a consecutive mild acid wash, which regenerated
high selectivity and also reduced coke formation.
Mesoporous zeolites prepared by leaching procedures are
frequently used to support highly dispersed metal or metal
oxide particles such as ZnO/ZSM-5, Ru/ZSM-5 or Ru/BEA, or Pt
on GaZSM-11 (MEL).24 To cite a specific example, Kang et al.
supported Ru nanoparticles on desilicated mesoporous acidic
ZSM-5 for the coupling of FischerTropsch reactions with subse-
quent hydrocracking activity, resulting in an enhanced selectivity
toward C5C11 hydrocarbons.24b In this case, the mesopores
are favorable for improved mass transfer of intermediates, thus
preventing excessive cracking reactions to light hydrocarbons.
In summary, basic leaching procedures represent popular
methods for the formation of mesoporous zeolites via desilica-
tion. They are being viewed as scalable once the best procedure
is established for a specific zeolite structure and composition.
However, although the initial limitations regarding the acces-
sible Si/Al range have now been overcome, remaining chal-
lenges regarding multistep reactions and the precise selection
of processing conditions have to be addressed to retain zeolitic
yield, especially for structures more sensitive than ZSM-5
within the ideal Si/Al ratio between 25 and 50.
7.11.2.3 Desilication Combined with
Surfactant-Supported Mesopore Formation
As described above, base leaching creates, as a side effect,
dissolved silica/alumina species that sometimes markedly
decrease the yield of mesoporous zeolite. In some instances,
Mesoporous Zeolitic Materials 257
Inlets
Outlet
Stirrer
Reaction
chamber
Static
blades
this zeolite debris is purposely created in order to function as a
precursor for mesostructured aluminosilicates to ultimately
result in zeolite/mesoporous oxide composites. This is sup-
posed to create a combination of highly acidic, microporous
zeolite nanocrystals with a material having a mesoporous pore
architecture for accelerated molecular transport. Even though
often described as recrystallization, we like to refer to this
process as transformation as long as an amorphous meso-
porous portion is present to form a composite, and refer
to recrystallization only when a true recrystallization into a
mesoporous zeolite phase has occurred.
In this regard, a partial zeolite transformation was under-
taken by Wang et al. by dispersing pre-made mordenite zeolites
at room temperature in a solution containing varying concen-
trations of NaOH to which cetyltrimethylammonium bromide
(CTAB) was later added. A subsequent hydrothermal conver-
sion at 100  C was at some point interrupted for lowering the
pH and continued for further mesoporous restructuring into a
MOR/MCM-41 composite. This composite showed the envi-
sioned improved catalytic activity in the hydrodealkylation of
heavy aromatics (C10 ) when compared to physical mixtures
of zeolite and MCM-41. This effect was tentatively attributed to
zeolite-like subspecies in the mesopore walls that had out-
lasted the zeolite dissolution.44 Following related procedures,
either solely mesoporous materials were obtained or compos-
ite micro/mesopore structures showing simultaneously small
and wide-angle x-ray diffraction depending on the zeolite
structure and severity of the reaction conditions. Ordered
mesostructured aluminosilicates were, for example, generated
in the group of Pinnavaia et al. starting from ZSM-5 or USY,
showing unusual acidity and stability when compared to the
common amorphous mesoporous MCM-materials.45 Com-
posites were observed by Qi et al.67 who treated large pre-
formed beta or mordenite single crystals this way. The partial
transformation was also performed with Ga,Al-ZSM-546 and
extended for building ZSM-5/SAPO-5/MCM-41 composites.47
These complex materials showed a very high activity for the
catalytic cracking of bulky 1,3,5-triisopropylbenzene (TIPB).
Much research along these lines (applying a two-step treat-
ment of base leaching and subsequent mesostructuring with
258 Mesoporous Zeolitic Materials

CTAB) has been performed by Ivanova et al., who carefully adding a dealumination step with citric acid before pore restruc-
studied the impact of NaOH concentration on the respective turing. Further, mesoporous ultrastable USY was made by
zeolites, including dealuminated mordenite and H-BEA or steaming the template-containing zeolite. These mesostruc-
H-FER zeolites.49 They found that depending on the severity tured USY zeolites have already been scaled up to the multi-
of desilication different composites are formed. With mild kilogram scale followed by formulation. The resulting materials
desilication, intrazeolite mesopores are formed together with showed increased fluid catalytic cracking (FCC) conversion for
an amorphous mesoporous layer on the outside of the crystals. gasoline and light cycle oil with less coke formation.
Medium concentrated bases create coexisting phases of morde- Similarly, recrystallized zeolite Y was made by Chal et al.
nite and MCM-41, while strong bases induce a complete rear- with TPAOH as base, following the patented recipe by Garcia-
rangement into purely mesoporous noncrystalline material. Martinez.48 A one-step recrystallization of base-leached ZSM-5
All transformed samples showed an increase in catalytic activ- fragments with CTAB was studied by Yoo et al. by carefully
ity in the alkylation of naphthalene with cyclohexene when tuning the amount of NaOH and CTAB and compared to base-
compared to the pure zeolite parent, where the reaction is leached samples without using a surfactant.51 The latter gener-
mostly restricted to the outer surface. However, best perfor- ally resulted in smaller surface areas, lower micropore volume,
mance was observed only when major parts of the micropo- and less yield. In the template-recrystallized materials, nearly
rosity were still preserved. all of the zeolite crystallinity and microporosity were retained.
A related synthetic route was explored by Garcia-Martinez Overall, a trimodular material with zeolitic micropores in
et al., who focused on intrazeolite recrystallization through addition to micelle-generated 3-nm mesopores and base-
restricted local bond breaking by mild leaching rather leached 1030-nm pores was obtained.
than dissolving most of the framework and transforming
it into a mesoporous amorphous phase. They recently dem-
onstrated this process with commercial zeolite Y, and also 7.11.3 Bottom-Up Synthesis of Hierarchical
applied it to other structures such as MOR and MFI as disclosed Zeolites by Dual-Templating Methods
in earlier patents.50 For example, using an ammonium hydrox-
ide/CTAB surfactant mixture at a relatively low pH between 9 In the previous section we have described how a secondary
and 11 in a one-step hydrothermal treatment at 150  C, they pore system can be implemented in pre-made, often com-
could effectively restrict the Si dissolution by a proposed mercial, zeolites via a posttreatment of partial dissolution, a
shielding function of the template. Nearly 100% of the zeolite process where the extent, size, and morphology of the newly
was recovered with retained Si/Al ratio and featuring a narrow formed mesopore network are not always easily controlled and
mesopore size distribution. Cationic surfactants with different where the overall yield is often reduced. We will now describe
chain lengths even allowed for tuning the mesopore size alternative procedures working from the opposite viewpoint,
between 2.5 and 4.5 nm, and some of the samples showed where a hierarchical pore system is created directly during the
hexagonal mesopore ordering as indicated by small-angle x- zeolite formation process with help of a dual-templating sys-
ray scattering. A very high mesopore surface of over 700 m2 g1 tem. This approach requires a synthetic setup that includes the
was reached under these conditions with zeolite Y, while usual template being responsible for the micropore structure of
micropore volume was still significant (0.21 ml g1 vs. the zeolite and a second template that is structure directing for
0.3 ml g1 in the parent material). The authors suggest that the mesopore architecture of the hierarchical material. These
an in situ surfactant-assisted crystal rearrangement takes place, secondary templates can be very different in nature. We will
where the negatively charged defect sites created by the crystal distinguish somewhat arbitrarily hard and soft (including
dissolution are neutralized by electrostatic interactions with the polymer and small-molecular reagents) templates based on
positively charged surfactant micelles (see Figure 6). Mesopor- their physical nature and interaction with the zeolite gel.
ous zeolite Y with a low Si/Al ratio of 2.4 was also prepared by Hard templates are simply added to the zeolite gel and exist

400
Volume adsorbed @ STP (cc g-1)

300

200
3.5
)
1

3.0
Diff. pore volume (cc g

2.5
2.0
100
1.5
1.0
0.5
0.0
0
20 40 60 80 100 120 (a) (b) (c) (d)
Pore width ()
0.0 0.2 0.4 0.6 0.8 1.0
P/P0

Figure 6 Mesoporous zeolite Y obtained by restructuring with CTAB. (Left) Nitrogen sorption on parent (black) and treated (blue) zeolite Y. (Right)
Schematic of the proposed zeolite mesopore formation process by Garcia-Martinez et al.: (a) original zeolite Y, (b) SiOSi bond opening/reconstruction
in basic media, (c) crystal rearrangement to accommodate the surfactant micelles, and (d) removal of the template to expose the mesoporosity
introduced. Reproduced by permission of The Royal Society of Chemistry, rearranged from Garcia-Martinez, J.; Johnson, M.; Valla, J.; Li, K.; Ying, J. Y.
Catal. Sci. Technol. 2012, 2, 987994.

as a physical mixture with the gel. Either they will retain their
physical shape and direct the zeolite growth within the con-
fined spaces created between these scaffolds or they will be
embedded as isolated units inside the growing zeolite phase,
leaving their mesoporous imprints behind when removed by
calcination. Soft templates such as positively charged or neu-
tral polymers can influence the crystallization process in situ by
directly interacting with the negatively charged zeolite frame-
work via electrostatic interactions or through covalent bonds
when silylating alkylammonium spacers are used. The direct
synthesis of mesoporous zeolites through secondary templat-
ing could be advantageous with respect to catalytic applica-
tions, since well-defined zeolites with high microporosity can
be synthesized at high overall yields.
7.11.3.1 Hard Templating
Hard templates can come in a variety of forms such as resin
beads, metal oxide particles, plant materials, and aerogels. To
date, most of the work was done with carbon scaffolds of
different shapes and sizes. These templates are generally simply
combined with the reaction gels, making this route very versa-
tile and applicable to many types of zeolites including metal-
lophosphates. Much work on mesoporous zeolites obtained
via the carbon-templating route has been reported by the
groups of Jacobsen and Christensen et al. and is described in
several reviews.68 The carbon templates used in this process
mostly consist of commercial carbon black pearls (BP2000 or
BP700) with an average diameter of 12 or 18 nm, respectively.
Usually, this carbon template is first infiltrated with an etha-
nolic, micropore-template-containing solution and second
with the silica solution, separated by a drying step. Zeolite
crystallization is then frequently performed with the dried
powder using a steam-assisted conversion (SAC), where the
impregnated zeolite precursor mixture and liquid water are
kept apart within the autoclave in order to avoid a phase
separation between the carbon phase and the zeolite gel.
Depending upon the amount of precursor gel used for imbibi-
tion, aggregates assembled from 20 to 30 nm nanoparticles of
ZSM-5 or zeolite beta were obtained, with interparticle pores
of 17 nm (ZSM-5) or >50 nm (beta) when the amount of gel
was just sufficient to fill the interstitial spaces between the
carbon pearls, a process described as confined-space synthesis.69
By contrast, when an excess amount of gel was used that embed-
ded the carbon particles, single-crystal zeolites were formed, as
proven by selected area diffraction, with intraparticle mesopor-
ous imprints of about 550 nm (see Figure 7(a)).70
Most research along these lines has focused on ZSM-5, but
a variety of other zeolite structures, including the MFI mem-
bers silicalite-1,71 TS-1,72 MTW (ZSM-12, Si/Al 20100, with
mesopores of about 50 nm),73 and AEL (the silicoalumino-
phosphate SAPO-11)74 were similarly prepared. Regarding
catalytic activity, zeolites ZSM-12 and beta with 40-nm
pores generated from carbon black showed an almost four-
fold increase in n-heptane hydro-isomerization activity.75 A
comparison of different carbon template materials used in the
silicalite-1 synthesis including carbon pearls, aggregated
carbon powders, or carbon fibers established that best results
were achieved with the carbon black pearls, which resulted in
highly tortuous cylindrical pores of about 20 nm.71 Fibers
Mesoporous Zeolitic Materials 259
and carbon aggregates favored the formation of inhomoge-
neous pores with an inkbottle shape. Uniform mesopore
distribution and particle morphology were also generated
with carbon black in TS-1 when the synthesis was performed
via steam-assisted conversion (SAC) of the dried synthesis
gel.76 Samples showed an enhanced catalytic activity for the
phenol hydroxylation to catechol and hydrochinone with
selectivity similar to the conventional TS-1.
Using the fluoride route (in order to create larger crystals with
less defects) with carbon black pearls, MFI, MEL, CHA (AlPO-
34), and AFI (AlPO-5) zeolites were prepared, resulting in meso-
pores of about 30 nm in all these cases.77 Moreover, mesoporous
single-crystal BEA materials were also obtained by this route.
Ideally, the mesopore volume as well as the size of the
secondary pores should be tunable. One strategy to fine-tune
these pores includes a combination of carbon-templating and
desilication. Treating ZSM-5 with 2030-nm mesopores, previ-
ously made by carbon templating, with increasing amounts of
base in a second reaction step, Holm et al. could create a hierar-
chical structure with newly created smaller pores of varying sizes
and doubling the mesopore surface area up to 234 m2 g1,
leaving the microporosity unaffected (0.11 ml g1).78 A one-
step fine-tuning of the pore size between 10 and 20 nm was
induced by increasing the amount of carbon black (from 10 to
40 wt%) in the synthesis gels for ZSM-5 or TS-1 and using
microwave conversion. Here, similar mesopore surface areas as
above were obtained; however, the increasing pore size was
accompanied by decreasing microporosity down to only
0.06 ml g1 or less. Nevertheless, the catalytic activity for the
epoxidation of cyclic olefins was still increased, which was attrib-
uted to the increase in mesopore surface area.79
Single-crystal-like faceted ZSM-5 zeolites were also made
with an inexpensive, in situ generated carbon matrix: impreg-
nation of the silica precursor gel with sugar and subsequent
pyrolysis in Ar resulted in a silica/carbon composite, which was
then treated with the template-containing basic solution in a
second impregnation step. A simple variation in sugar concen-
tration helped to control the size of the resulting mesopores
between about 10 and 20 nm (see Figure 7(c)).80 TS-1 was
made similarly either using carbonization of sucrose or by a
one-step method using a direct dry-gel conversion of caramel-
impregnated zeolite gels.81 These samples showed increased
activity in the epoxidation of cyclohexene and the oxidation of
benzothiophene.
Carbon-pearl-templated mesoporous silicalite and ZSM-5
zeolites were used as substrates for the introduction of nano-
particles, for instance, with nano-metal (Pt), -alloy (PtSn), or
-carbide (Mo2C) particles.82 Recently, similarly made Pd- or
Mo-containing ZSM-5 showed a better metal dispersion and
thus also catalytic activity in hydro-conversion of n-octane and
in methane dehydro-aromatization, respectively, compared to
conventional ZSM-5. The higher resistance to deactivation by
coke formation was explained by the trapping of coke in
the mesopores, leaving the active sites still accessible.83 A one-
step generation of metal oxide particles in mesoporous silicalite-
1 was proposed by Jiang et al. who used TiO2ZrO2-decorated
carbon particles as hard templates that were further exam-
ined for the dehydrogenation of ethylbenzene.84 MoZSM-5
obtained with disordered carbon rods or ordered carbon
mesoporous molecular sieve templates (CMK) (see below)
260 Mesoporous Zeolitic Materials

P/P0

(a) 0 0.2 0.4 0.6 0.8 1 (b)

250

Amount adsorbed (cm STP g )

1
Meso-ZSM-5
200 ZSM-5

3
150

Pore size of meso-ZSM-5 (nm)

100
0.4 0.6 0.8 1 0 20 40 60 80

(cm g nm )
distribution
1
4

Pore size
12
3
50 8

3 1
2
4 1
0 0
0
mm 10 20

100 nm 100 nm

(c) (d) (e)

1 m 200 nm
100 nm

Figure 7 Zeolites made with hard-templating strategies: (a) single crystal ZSM-5 made with carbon pearls BP2000 (adapted with permission from
Jacobsen, C. J. H.; Madsen, C.; Houzvicka, J.; Schmidt, I.; Carlsson, A. J. Am. Chem. Soc. 2000, 122, 71167117. Copyright (2000) American Chemical
Society), (b) monolithic ZSM-5 made with monolithic carbon aerogels (adapted with permission from Tao, Y.; Kanoh, H.; Kaneko, K. J. Am. Chem. Soc.
2003, 125, 60446045. Copyright (2003) American Chemical Society), (c) aggregated ZSM-5 nanoparticles made through sugar addition (adapted with
permission from Kustova, M.; Egeblad, K.; Zhu, K.; Christensen, C. H. Chem. Mater. 2007, 19, 29152917. Copyright (2007) American Chemical
Society), (d) meso/macroporous silicalite by CaCO3 inclusion (adapted with permission from Zhu, H.; Liu, Z.; Wang, Y.; Kong, D.; Yuan, X.; Xie, Z. Chem.
Mater. 2008, 20, 11341139. Copyright (2008) American Chemical Society), (e) ordered BEA generated from a periodic 3DOm carbon replica (adapted
with permission from Chen, H.; Wydra, J.; Zhang, X.; Lee, P.-S.; Wang, Z.; Fan, W.; Tsapatsis, M. J. Am. Chem. Soc. 2011, 133, 1239012393. Copyright
(2011) American Chemical Society).

showed similar conversion in the nonoxidative aromatization
of methane when compared to common ZSM-5, but provided
higher yields in aromatics and a longer lifetime.85
Carbon aerogels (CAs) can be made by the catalytic poly-
merization of resorcinolformaldehyde gels. By drying under
supercritical conditions with CO2 and subsequent pyrolysis,
aerogels can be transformed into vitreous black mesoporous
monoliths consisting of uniform aggregated carbon nanoparti-
cles. The group of Kaneko has used these materials as alternative
carbon template for zeolite synthesis.86 These CAs allowed them
to mold millimeter-sized monolithic ZSM-5 with a narrow
mesopore distribution of around 11 nm by impregnation with
the zeolite gel (see Figure 7(b)). Large mesopore volumes of
1.37 ml g1 were reached when this scaffold was used in the
synthesis of zeolite Y. Zeolite A (LTA) was not easily made in
carbonized aerogels, but could be synthesized with 15-nm pores
when the resorcinolformaldehyde aerogels (RFAs) were used
in their native form. These uncarbonized aerogels are removed
under milder conditions, are more easily tuned in pore size, and
offer a more flexible utilization in zeolite synthesis due to more
favorable electrostatic interactions. By contrast, CAs can achieve
a more narrow pore-size distribution. Kanekos group recently
presented a comparison of RFAs and CAs made with controlled
hydrophilicity.87 Beautiful ZSM-5 single crystals with varying
mesopore sizes between 15 and 40 nm were formed from dif-
ferent CAs made simply with increasing amounts of the sodium
carbonate catalyst for the polymerization of the resorcinol gel.88
Carbon black pearls acting as templates frequently lead
to multiphase systems containing mesoporous zeolite single
crystals in addition to nonmesoporous zeolite nanoparticles,
and hence may show relatively broad mesopore-size dis-
tributions. In order to achieve greater structural control, a
higher-priced version of carbon has been used as template in
the form of multiwalled carbon nanotubes. As a result, straight
mesopore channels of uniform size similar to the carbon
nanotubes (about 20 nm) are formed that completely pene-
trate single crystals of silicalite-1.89 On the other hand, slit-like
2025-nm-wide pores have recently been introduced into

silicalite-1 crystals using graphene oxide sheets as template,
which were believed to stimulate nucleation along the phase
boundaries due to their hydrophilic surface.90
Other carbon-containing template scaffolds include 20
30-mm long, 2060 nm wide carbon nanotubes grown on SiC
wafers or on particles, prepared via ferrocene chemical vapor
deposition (CVD) and subsequent pyrolysis. When zeolite layers
were crystallized on these decorated substrates, the carbon nano-
tubes were found to penetrate the zeolite layers and to form
mesopores after calcination.91 Silica/carbon composites made
by simultaneous pyrolysis of silica gel and carbon-containing
gases were used as multifunctional precursors in the synthesis of
ZSM-5.92 The mesoporosity could be controlled by adjusting the
carbon content.
Great efforts have been made to form zeolites with ordered
mesopores. The preparation of ordered scaffolds made of car-
bon or partially carbonized compounds on siliceous substrates
constitutes a whole field on its own, and only most recent
results are discussed here. Principally, a twofold replication
process is necessary when colloidal silica imprinted carbons
(CIC93 or 3DOm94) are used, that is, first replicating packed
silica spheres by carbon to form an inverse opal, which is then
used to replicate a zeolite network composed of nanoparticles by
confined growth. The groups of Stein and Tsapatsis have beauti-
fully orchestrated the synthesis of organized LTA, FAU, BEA,
silicalite, and LTL phases this way by using three-dimensionally
ordered mesoporous (3DOm) carbon with pore sizes on the
meso- or macroscale, depending on the size of the silica
spheres and the sequence of gel infiltration (see Figure 7(e)).95
Here, multiple hydrothermal synthesis steps starting from a
seeded 3DOm gel were performed in the inverse carbon opal,
until the void volume of the porous carbon structure was filled
with zeolite nanocrystals in order to limit extraneous bulk zeolite
formation. These highly ordered zeolite phases with tunable
37-nm mesopore systems are thought to be good model
systems for studying mass transfer limitations and might be
used as building blocks for zeolite films.
On the other hand, pre-made ordered mesoporous MCM-
41, SBA-15, or KIT-6 silica can serve as matrix for carbon repli-
cation (CMK materials), where the silica phases remain in the
composites and are used as digestible silica source in the zeoli-
tization process.96 A detailed investigation was recently reported
by Cho et al., using various CMK carbon materials that were
infiltrated with MFI precursor gels and processed by SAC in
order to reduce the separation between the carbonaceous and
the silicate phase.97 They showed that the conditions for the
steam conversion and the template pore size have to be carefully
tuned in order to achieve a crystalline zeolite array. Highly
ordered silicalite-1 was only obtained with a 10-nm pore carbon
scaffold and using 85% humidity during SAC conversion.
A higher humidity leads to migration of the zeolite precursor
onto the outside of the carbon scaffold forming bulk zeolite.
Carbon templates with smaller pore diameters were supposedly
too narrow to allow zeolite nucleation. Recently, an in situ
formed tri-component SiC matrix was prepared by co-assembly
of phenolformaldehyde resin, tetraethyl orthosilicate (TEOS),
and the surfactant F123 and partially converted into silicalite-1
or ZSM-5.98
Examples for hard templating with ingredients other than
carbon materials are not as abundant. Close-packed arrays of
Mesoporous Zeolitic Materials 261
monodisperse polystyrene pearls (PSs) leave macroporous
imprints of about 250 nm in a silica/silicalite-1 composite.99
Millimeter-sized self-bonded spherical silicalite-1 and beta par-
ticles were generated when porous anion-exchange resin beads
of similar dimensions were used as sacrificial compartments
for the zeolite gel.100 These beads were composed of zeolite
nanoparticles and amorphous silica residues, displaying hier-
archical interparticle pores of about 40 nm. A side product of
non-resin-bonded nanozeolites was always observed. Eighty-
nanometer-sized poly(methyl methacrylate) (PMMA) spheres
could be assembled into micrometer aggregates (presumably
through electrostatic interactions) when they were mixed with
a ZSM-5 zeolite precursor gel.101 Hydrothermal conversion of
these composite spheres and subsequent calcination resulted in
similarly sized beads composed of small MFI particles arranged
in single-particle appearance showing large surface areas and
about 13 nm additional mesopores. The small size of the meso-
pores is believed to result from shrinkage of the 80-nm PMMA
spheres during the high-temperature zeolite conversion.
Inorganic particles such as MgO102 or CaCO3103 can be
used as alternative hard scaffolds. This way meso- to macro-
sized pores were recently introduced into silicalite-1 by the
addition of 50100-nm hydrophilic CaCO3 particles (see
Figure 7). It could be shown that the surface properties of
these particles are responsible for the mesopore formation:
when fatty-acid-treated hydrophobic CaCO3 particles were
used, conventional silicalite-1 particles resulted. A mild acid
wash was then used for dissolution of the metal oxide tem-
plates, adding slightly to the mesopore growth. Bio-inspired
templating on the macroscale, using for instance carbonized
seeds, shells or fibers104 or wood cells, palm stems, silica-rich
plants, or diatoms for preparing zeolitic replicas with micro
mesomacropores, was recently reviewed by Valtchev and
Mintova and is not described here.105
7.11.3.2 Soft Templating with Polymers and
Molecular Additives
In contrast to using a scaffold that exhibits a predefined mor-
phology on the nanoscale, there is a different class of templates
that may vary in shape and functionality when added to a
zeolite gel. Flexible polymers of different chemical composi-
tion, molar mass, and surface charges can act as inexpensive
mesoporous templates that may create internal pores through
inclusion, or interstitial mesopores in zeolite nanoassemblies
by functioning as a flocculating agent. The application of cat-
ionic polymers as a versatile alternative to hard templating was
studied by Xiao et al. in order to overcome the interface incom-
patibility of uncharged templates and the negatively charged
aluminosilicate gels under basic conditions, a problem that is
frequently encountered in carbon templating. They studied
polydiallyldimethylammonium chloride (PDDA) for the syn-
thesis of zeolite beta and obtained polycrystalline hierarchical
particles sized about 300 nm.106 The mesopore dimensions
were similar to the estimated polymer size of about 540 nm
when measured with nitrogen sorption and were supposedly
created by a direct templating function of the polymer. These
zeolites showed, in comparison to zeolite beta made without
polymer addition, a higher conversion, selectivity, and stability
in the alkylation of benzene, and after Pd loading a superior
262 Mesoporous Zeolitic Materials
hydrogenation activity with pyrene.107 By increasing the poly-
mer content in the synthesis gel it was possible to increase the
mesopore volume (up to 0.41 ml g1 in beta) with pore sizes
between 18 and 25 nm, not only in zeolite beta but also in
ZSM-5.108 In the case of zeolite beta, adding a 3-day hydrother-
mal pretreatment of the gel in order to facilitate zeolite nucle-
ation before adding the polymer generated even higher
mesopore volumes of up to 0.84 ml g1, and a mesopore surface
area of 185 m2 g1 was reached.109 In this case, 1020-nm-sized
nanozeolites acted as additional spacers within the polycrystal-
line aggregates. The authors suggest that the high charge density
of the cationic polymer is responsible for the aggregation into
filterable mesoporous zeolites as cationic surfactants such as
CTAB produced only ordinary zeolites. They extended their stud-
ies to large single-crystal-shaped mesoporous zeolite X. In this
case, a preformed FAU precursor solution was enriched with
additional sodium aluminate and mixed with PDDA prior to
the thermal treatment. The resulting mesoporous FAU showed a
superior exchange rate for calcium.110
Very recently, Xiaos group designed a high-molecular-
weight cationic amphiphilic copolymer that forms large micelles
with a small curvature able to direct mesopore growth with
controllable orientations. They used methyl iodide-treated
polystyrene-co-4-polyvinylpyridine (MW of 105) as admixture
to the zeolite precursor gel and were able to induce the second-
ary pore formation along the b-orientation in ZSM-5. A self-
assembly of the polymer with the zeolite crystal along the b-
direction is energetically favorable according to density func-
tional theory (DFT) calculations and is probably due to the
location of the straight ten-ring channels in ZSM-5. The 1
2 mm-sized single crystals showed 1355-nm unidirectional
mesopores and were called ZSM-5-OM (see Figure 8). The
oriented diffusion pathways enabled a higher catalytic conver-
sion for bulky molecules even when compared to randomly
oriented mesoporous ZSM-5 prepared with PDDA.111
m n
N N+
5200 1.0kV 30.0k SE 1.00 m
Figure 8 Oriented mesopores formed in ZSM-5 using the copolymer C-PSt-co-P4VP. SEM and TEM micrographs showing the mesopores in
b-direction. Adapted with permission from Liu, F.; Willhammar, T.; Wang, L.; Zhu, L.; Sun, Q.; Meng, X.; Carrillo-Cabrera, W.; Zou, X.; Xiao, F.-S. J. Am.
Chem. Soc. 2012, 134, 45574560. Copyright (2012) American Chemical Society.
We have studied the formation of zeolite beta starting from
very dense gels with varying amounts of PDDA of different
molecular weights in a one-step reaction.112 Under these con-
ditions, the polymer acts predominantly as a flocculating agent
for zeolite nanocrystals. In the absence of the polymer, highly
crystalline nanozeolite beta of 40130 nm was obtained as
colloidal solution that had to be retrieved via centrifugation.
When these nanosized powders were analyzed with nitrogen
sorption, interparticle pores of about 20 nm were observed.
On the other hand, polymer-flocculated nanosized zeolite beta
resulted in stable, macroscopic aggregates consisting of about
50500-nm primary zeolite beta particles that could simply be
filtered. Here, mesopore volumes of up to 0.89 ml g1 were
reached without sacrificing the micropore volume of the zeo-
lite phase. The secondary (mesopore and macropore) dimen-
sions could be tuned between 40 and more than 360 nm
(measured by mercury intrusion) by increasing the polymer
content in the gel, due to an increasing primary particle size.
The polymer PDDA was also used to coat polystyrene spheres
or macroporous silica that electrostatically attracted pre-made
ZSM-5 or beta nanozeolites, thus establishing a macroscopic
molding process.113
Polymers have been used before to stimulate the formation
of zeolite aggregates in multiple-step processing. In this case,
pre-made nanozeolites were used in the process of polymer-
induced colloid aggregation (PICA). Here, a colloidal solution
of nanozeolites is mixed with an acidified urea solution and
flocculated upon addition of formaldehyde via instantaneous
polymerization. After calcination zeolite macrospheres of sev-
eral microns in diameter were retrieved, with mesopores
between 15 and 70 nm, depending upon the primary nano-
zeolite size.114 The morphology of the aggregates could be
tuned by adjusting the acid concentration in the mixture:
homogenously distributed nanozeolites were produced at
low HCl concentrations, while higher HCl concentrations
[010]
o
I
CH3
20 nm

led to irregular structures.115 Solid mesoporous spheres of
zeolite beta, ZSM-5, and silicalite-1 were processed this way,
showing benefits in the dehydration of fructose to furfural
and for adsorption rates of proteins, suggesting applications
in the chromatographic separation of biomolecules. The
aggregation of 50100-nm nanozeolite ZSM-5 into mesopor-
ous micron-sized particles was also observed upon addition
of an inexpensive reagent consisting of untreated sugar cane
bagasse.116
Self-organizing charged surfactants and noncharged larger
block copolymers present challenges when used for the syn-
thesis of mesoporous zeolites. Phase separation of zeolites and
amorphous mesoporous materials can easily occur. Micellar
curvature and charge balance with zeolite building blocks are
assumed to be critical factors. However, Zhu et al. found that
the adequate preparation of the zeolite building blocks via
kinetic control enables mesostructuring with zeolite CTAB
without resulting in a phase separation.117 They performed a
careful study of the gel compositions and the necessary aging
conditions to prepare the appropriate degree of oligomeriza-
tion of the zeolite subunits. Aging of clear precursor solutions
for 23 days at 100  C before addition of CTAB gave the best
size match between CTAB micelles and zeolite precursor enti-
ties. Subsequently, a second hydrothermal treatment provided
the self-assembly into ZSM-5 nanocrystal aggregates with 23-
nm mesopores. Shorter aging times were insufficient to
produce the necessary nanometer-sized subnanocrystal-type
zeolite seeds and an amorphous mesophase formed as a sec-
ond phase. This process can probably be also described as a
flocculation mechanism because the aggregation into larger
particles did not occur without the addition of CTAB. An
alternative route to prevent phase separation is restricting the
mass transport in the zeolite precursor gel by using a SAC
conversion. This way it was possible to grow mesoporous
nanozeolite beta aggregates under the confining presence of
varying concentrations of CTAB in the reaction gel.118
Other polymers such as the triblock copolymer F127 stim-
ulated the spontaneous self-assembly of about 100-nm-sized
nanocrystalline MFI particles into 35-mm spheres without
further pretreatment when it was directly added into the zeolite
gel used for hydrothermal conversion. The resulting mesopore
volume of 0.08 ml g1 could be slightly increased to about
0.11 ml g1 upon co-addition of styrene.119 A detailed study
of the related system of TS-1 found that macrosphere aggrega-
tion occurred only when the synthesis gel (containing F127 or
P123 block-copolymer) was stepwise heated to the final tem-
perature of 170  C; otherwise, nanozeolite suspensions were
formed. The role of the polymer was proposed to be twofold,
serving first as an inhibitor of fast silica polymerization and
second for providing a confined space to accelerate the aggre-
gation of nanoparticles, thus leaving about 10-nm interparticle
pores with an overall mesopore volume of 0.1 ml g1. These
mesoporous particles showed an activity similar to conven-
tional TS-1 zeolites for the oxidation of pyridine, but the
mesoporosity was beneficial for the oxidation of bulky reac-
tants such as 3-picoline and cyclohexene.120
Using the SAC process, block copolymer F123-enriched,
dried synthesis gels for TS-1 and ZSM-5 with different Si/Al
ratios resulted in 700-nm zeolite particles with a pore size of
2530 nm and with a much higher mesopore volume of about
Mesoporous Zeolitic Materials 263
0.5 ml g1. A lower Al content led to higher crystallinity, hydro-
thermal stability, acidity, and better selectivity toward 2,4-di-tert-
butylphenol in the solvent-free tert-butylation of phenol.121 A
similar procedure was further used to create hollow ZSM-5
spheres with a mesoporous zeolite shell of about 120 nm by
relying on the inhomogeneous Al distribution in these particles.
A mild desilication with a sodium carbonate solution selectively
removed the Al-depleted interior leaving behind a single-crystal-
like shell with penetrating mesopores.122 The triblock copolymer
F127 in combination with cerium salts and SAC conversion
found application for manufacturing Ce-loaded mesoporous
ZSM-5 with different Si/Al ratios and cerium content.123 The
dry-gel conversion in combination with polymers was also
effectively used in the synthesis of SAPO-34 (CHA). Inclusion
of either polyethylene oxide (PE) or propylene oxide (PO) trig-
gered the formation of trimodal monoliths with pores at the
micro-/meso-/macroscale featuring different morphologies that
were highly dependent on the polymer concentration. These
hierarchical materials showed good performance in the MTO
conversion.124
A one-step generation of mesoporous zeolites was achieved
with an ethanolic solution of polyvinylbutyral polymer (PVB)
that was combined with TEOS and the remaining reagents
for zeolite beta or ZSM-11 and converted hydrothermally
after evaporation of alcohol.125 Calcination revealed highly
crystalline 200-nm zeolite beta of single-crystal habit with
intraparticle pores of about 30 nm and an external surface
area of 114 m2 g1, or 400 nm polycrystalline particles with
interparticle pores of about 10100 nm and an external surface
area of 130 m2 g1 for ZSM-11 (see Figure 9). These zeolites
showed an extended resistance toward deactivation in the
cracking of trimethylbenzene.
Since mesoporous zeolites find increasing applications in
catalysis and scalability might become an issue, there is a desire
to investigate alternative, inexpensive, and environmentally
friendly sacrificial templates. Exploiting their hydrophilic
nature, carbohydrates in the form of cellulose, starch,126 even
starch-derived bread from Walmart,127 or sugars are promis-
ing candidates and have been successfully used for generating
mesoporous zeolites, mainly of the MFI family. Soluble starch
and sodium carboxymethylcellulose (CMC) helped to generate
mesoporous ZSM-5, TS-1, and silicalite single crystals of about
300 nm, most of them however with low microporosity.128
A variety of saccharides (starch, sucrose, cellulose, and glucose)
was also studied for the aluminophosphate zeotypes ALPO-5
(AFI) and ALPO-11 (AEL), and glucose worked best for
pore generation of about 1020 nm. The microporosity was
unfortunately not reported.129 Glucose was also used for hier-
archical titanium-containing aluminophosphate TiAPO-5
(10-nm mesopores), however, showing a lower micro-
porosity than the parent zeolite, as frequently observed. The
products were used as adsorbers for the enrichment of phos-
phopeptides from complex mixtures prior to mass spectro-
scopic analysis.130
A different approach of molecular templating uses amino
acids (lysine) in an aqueous zeolite synthesis solution contain-
ing octane to modulate close-packed colloidal silica spheres
that are dried and converted via SAC into 0.53-mm MFI zeolite
aggregates established from nanoparticles with interstitial pores
of about 815 nm, depending upon the Si/Al ratio.131
264 Mesoporous Zeolitic Materials

250

0.3

nm )
1
3 1
dv/(d log D)(cm g
200 0.2

Volume adsorbed (cm g , STP)

0.1

1
150

3
0.0
1 10 100
Pore diameter (nm)

100

50

50 nm 20 nm 0

250

0.3

nm )
1
3 1
0.2
200

dv/(d log D)(cm g

Volume adsorbed (cm g , STP)
0.1

1
150 0.0

3
1 10 100
Pore diameter (nm)

100

50

0
100 nm 100 nm 0 0.2 0.4 0.6 0.8 1
Relative pressure (P/P0)

Figure 9 One-step generation of mesoporous zeolites beta and ZSM-11 using the polymer polyvinylbutyral. TEM and nitrogen sorption data of zeolite
beta (top) and zeolite ZSM-11 (bottom). Reprinted with permission from Zhu, H.; Liu, Z.; Kong, D.; Wang, Y.; Xie, Z. J. Phys. Chem. C 2008, 112,
1725717264. Copyright (2008) American Chemical Society.

A covalent anchoring of the porogen to zeolitic precursors
was intended when the group of Pinnavaia explored silylated
polymers in the synthesis of ZSM-5.132 The primary goal was to
create mesopores of medium size (below 10 nm) that would
be beneficial for catalytic gas-oil cracking. For this purpose,
they used polyethylene imine or polypropylene oxide diamine
that was silylated with glycidoxypropyl trimethoxysilane. The
hydrolyzed silyl-groups along the backbone of the polymer
were anticipated to interact with zeolite nucleation processes
while the polymer could act as spacer at the same time.
The dimensions of the specific polymers dictated an intra-
particle mesopore diameter between 2 and 5 nm. Further,
alkoxysilylated nonionic surfactant (poly(oxyethylene ether)
TES-C16EO10) was found beneficial for the synthesis of
mesoporous silicalite-1, which could not be obtained with a
silylated cationic surfactant.133
systems, siloxy-functionalized alkyl chains seemed promising
as secondary templates. The possibility to control the mesopore
dimensions by predetermining the length and composition of
the alkyl chains could open up a large field for custom-made
hierarchical zeolites. Hence, a number of different alkylated
siloxanes with increasing complexities were used by several
research groups starting around 2006 after the groundbreaking
work of the groups of Ryoo and Serrano. We will discuss these
approaches along the structures of the applied templates in
the following manner: (1) amphiphilic silylated surfactants,
(2) neutral alkyl amino siloxanes, and (3) simple alkylated
siloxanes. The molecular structures of these porogens are given
in Scheme 1 and an overview of their applications in zeolite
synthesis is listed in Table 2, combined with catalytic test reac-
tions, at the end of Section 7.11.3.3.2.

7.11.3.3.1 Cationic organosilane surfactants

7.11.3.3 Silylated Alkyl Chains as Mesostructuring
Templates
Much effort has been extended during the last few years to exploit
supramolecular templating on the basis of common surfactants
such as CTAB and its analogs for the immediate generation
of crystalline mesoporous zeolites, relying on their success in
the synthesis of periodic amorphous mesoporous silicates. How-
ever, when used as pure surfactants in a common zeolite conver-
sion, often a phase separation was observed that produced
zeolite domains entangled with amorphous mesostructured
regions. In order to gain more synthetic control over these
The two research groups of Ryoo and Serrano almost simulta-
neously presented the idea that mesoporosity can be forced
into a zeolite system by modifying the growth process through
interactions with an organic agent that carries a reactive silane
functional residue. The silane head group works essentially as a
growth inhibitor during zeolite formation while the alkyl chain
dictates the mesopore dimensions through the formation of
micellar aggregates. Ryoos group focused on the rational
design of functionalized amphiphilic surfactants. They mainly
used (3-(trimethoxysilyl propyl))hexadecyl dimethyl ammo-
nium chloride (TPHAC). This molecule contains a long hydro-
phobic alkyl chain resembling the cationic surfactants used in
 Mesoporous Zeolitic Materials 265

(H3CO)3Si CH3, C3H7, C8H17 M, P, OTMS

(H3CO)3Si NH2 APTMS

H
(H3CO)3Si N
TMPED
NH2

(H3CO)3Si N DATMS

PHAPTMS
(H3CO)3Si N
H

Cl
(H3CO)3Si N+
C12H25, C16H33, C18H37 TPDAC, TPHAC, TPOAC

R
(H3CO)3Si N
TPAV (Z-6032)
NR2

R: H or

Scheme 1 Overview over silylating agents. MTMS, PTMS, OTMS: methyl-, propyl-, and octyl-trimethoxysilane; APTMS:
3-aminopropyltrimethoxysilane; TMPED: N-[3-(trimethoxysilyl)propyl]ethylenediamine; DATMS: (N,N-diethyl-3-aminopropyl)trimethoxysilane;
PHAPTMS: phenylaminopropyl trimethoxysilane; TPDAC, TPHAC, TPOAC: [3-(trimethoxysilyl)propyl]dodecyl,-hexadecyl,-octadecyl-
dimethylammonium chloride; TPAV: reaction product of vinylbenzyl chloride and ethylene diaminopropyltrimethoxysilane, named N-(vinylbenzyl)-
2-aminoethyl-3-aminopropyl trimethoxysilane by the authors (DOW CORNING Z-6032).

the synthesis of ordered mesoporous silica materials and a
hydrophilic quaternary ammonium head group carrying a
hydrolyzable siloxy group for direct bond formation with the
zeolite precursor units. Both fragments are bridged via a SiC
bond that can withstand the alkaline conditions used in com-
mon hydrothermal zeolite conversions. Adding this mesoscale
template to the usual zeolite reaction mixture directed the
MFI growth into highly crystalline nanozeolite assemblies
with disordered but uniformly tailored mesopores of about
27 nm, depending upon the respective alkyl chain length
ranging from C12 to C18 and the reaction temperature.134
Besides ZSM-5, they were able to synthesize single-crystal-like
LTA following this route, displaying a narrow mesopore-size
distribution around 10 nm when using TPHAC (Figure 10). A
comparison of the above mesoporous ZSM-5 with Al-MCM-41
and protozeolitic MCM materials (assembled from zeolite
seeds) with respect to the catalytic activity in the MTO reaction
confirmed the superior acidity of these mesoporous zeolites;
however, the activity of mesoporous and common ZSM-5 was
nearly identical due to the small size of reactants. By contrast,
mesoporous ZSM-5 was beneficial in activity and selectivity
when bulkier molecules were tested. A good resistance against
deactivation in cracking, esterification, or isomerization reac-
tions with large reactants was also observed.155
A number of studies has exploited the above synthesis route
using TPHAC or TPOAC ([3-(trimethoxysilyl)propyl]hexadecyl
or -octadecyl-dimethylammonium chloride, containing C16 or
C18 fragments, respectively, see Scheme 1 above for silylating
agents) to create mesoporous acidic ZSM-5 zeolites in order to
evaluate them in catalytic test reactions, such as benzene alkyl-
ation or selective hydroxylation.156 Furthermore, mesoporous
ZSM-5 was recently explored in the catalytic fast pyrolysis of
biomass (maplewood, also glucose and furan) for direct pro-
duction of aromatic hydrocarbons as additive to transportation
fuels, a process that relies on medium-pore-size zeolites. The
zeolites performed best with a Si/Al ratio of 30; however, their
mesoporosity had little effect on the catalytic activity. In fact,
slightly more coke was formed and the product distribution was
shifted to heavier alkylated monoaromatics.157 Shetty et al.
found a high activity for bulky molecules with similar zeolites;
however, the activity was nearly lost when a dealuminating step
removed the aluminum from external surface sites. Since the
266 Mesoporous Zeolitic Materials

Table 2 Overview of hierarchical zeolites structured with silylating agents (loosely grouped according to zeolite phases)

Year and Method (template, Zeolite structure Particle size Mesopore size Catalysis
reference reaction conditions) (aggregate size/
primary particle size)

Cationic organosilane surfactants single-step reactions

2006134 TPDAC 0.41 mm aggregates C12 ! C16: MTO, jasmine aldehyde synthesis
TPHAC MFI 2.17.4 nm
TPOAC LTA
2011135 TPOAC ZSM-5, Fe-ZSM-5 <1 mm aggregates 2.5 nm Hydroisomerization, benzene
oxidation
2010136 TPOAC TS-1 0.3 mm/5 nm 2.53.5 nm Oxidation/hydroxylation reactions
2009137 TPHAC functionalization Pd MFI 3 nm Sonogashira coupling
complexation BEA 2 nm
LTA 10 nm
2009138a TPHAC LTA 12 mm Tunable Conversion of MeOH
TPHAC P123 610 nm
2011139 TPHAC LTA 0.21 mm aggregates Tunable NMR: water diffusion
Tdi(PA) BEA 58 nm
2011140 TPOAB PdY 0.71.5 mm 18 nm Hydrodesulfurization of
dimethyldibenzylthiophene
2012141 TPHAC X 9 mm 7 nm
2006142a TPHAC AlPO-5, AlPO-11, SAPO-5, >1 mm/2050 nm 3 17 nm Liquid-phase oxidation of tetralin
2010142b Co-SAPO-5, CrSAPO-5
2010143a TPHAC SAPO-5 Aggregates of 100-nm 3.8 nm broad Alkylation of benzene,
platelets cyclohexylbenzene
2009144 TPHAC SOD >1 mm aggregates 2.5 8 nm Knoevenagel, ClaisenSchmidt
condensation

Seed silanization multiple-step reactions

2006145 PHAPTMS (three-step) ZSM-5 0.30.4 mm/40-60 nm 4.5 nm Low-density polyethylene (LDPE)
2009146 BEA 16 nm cracking
2010147
2009148 PHAPTMS (three-step) TS-1 0.10.15 mm 15 nm Olefin oxidation
aggregates
2011149 PHAPTMS alcohols ZSM-5 0.10.15 mm 2 up to LDPE cracking
(three-step) aggregates 60 nm
2012150 TPAV (three-step) ZSM-5 0.20.4 mm/20 nm 3 nm MTO, benzylation reactions
aggregates
2011151 APTMS ZSM-5 0.6 mm/25 nm Alkylation of benzene
TMPED aggregates
DATMS (two-step)
2008152 MTES ZSM-5 0.5 mm/20 nm Alkylation, condensation reactions
PTES aggregates
OTES (one-step)
2012153 PHAPTMS (two-step) LTA 0.71 mm aggregates 3 nm Fast Mg2 exchange
2012154 TDPA (one-step) SAPO-11 46 mm 5 nm Isomerization of n-octane

catalysis with small molecules was unaffected by this treatment
they concluded that the catalytic conversion of larger molecules
occurs mainly at the pore mouth.158 Xin observed that meso-
porous FeZSM-5 was almost four times more active in the
hydroxylation of benzene than an optimized steam-calcined
FeZSM-5.159
A careful study of the synthesis mechanism of mesoporous
ZSM-5 with TPOAC was performed by Koekkoek et al. by
varying concentrations of the mesostructuring template as
well as temperature and crystallization times. A low Si/TPOAC
ratio implying high template concentrations was found to sub-
stantially hinder crystallization. The authors observed an ini-
tially rapid formation of ultrasmall microporous crystalline
domains embedded in an amorphous mesoporous matrix, and
a slow transformation into long-range crystalline areas at the
expense of the mesoporous fraction. Depending upon the Si/
TPOAC ratio, temperature, and reaction time, the crystallinity
could be improved at the expense of mesoporosity. Nearly fully
crystalline mesoporous samples were obtained with a Si/TPOAC
ratio of 240 during hydrothermal conversion, where the pres-
ence of Fe3 accelerated the crystallization. These mesoporous
FeZSM-5 materials showed a better catalytic longevity and out-
performed common FeZSM-5 in the hydroxylation of
benzene.135 SAC processing delivered similar results with
slightly less crystallinity, but when different mesostructured
FeZSM-5 materials (made by using carbon black templating,

500 nm
Figure 10 SEM and nitrogen sorption data of mesoporous LTA zeolite synthesized using [(CH3O)3SiC3H6N(CH3)2C16H33]Cl (TPHAC). Adapted by
permission from Macmillan Publishers Ltd: (Nature Materials) from Choi, M.; Cho, H. S.; Srivastava, R.; Venkatesan, C.; Choi, D.-H.; Ryoo, R. Nat. Mater.
2006, 5, 718723. Copyright (2006).
desilication, or TPOAC templating routes) were compared in
the same catalytic reaction, the latter showed the best per-
formance. This was attributed to the high interconnectivity
of micro- and mesopores that organize the zeolite into a
matrix with very small domain sizes, allowing an efficient
mass transfer.160
In some instances, it was found that the catalytic behavior
of zeolites is negatively affected by the introduction of meso-
pores. Cheneviere et al. used TPOAC in the synthesis of TS-1, a
titanium-containing MFI zeolite that is especially active in
oxidation reactions with H2O2.136 Its performance is based
upon isolated tetrahedrally coordinated titania species and
the hydrophobic character of the zeolite framework. After
ensuring that mesoporous templating did not change the coor-
dination of the active Ti centers in the zeolite lattice, conven-
tional TS-1 and its mesoporous counterpart were compared in
several oxidation/hydroxylation reactions typical for the appli-
cation of this zeolite. It turned out that the mesoporous TS-1
was nearly unreactive in the hydroxylation of phenol, a reac-
tion that is known to be very sensitive to catalyst structure,
crystal size, and hydrophobicity. This unexpected result was
attributed to the extremely small TS-1 crystalline domains
(<5 nm) that are evidently covered with silanol groups to a
much higher extent than the conventional TS-1 zeolites. The
resulting higher hydrophilic character presumably hinders the
access of phenol to the active sites in the aqueous solution.
The considerably increased outer surface area and the large
mesopore sizes generated with the organosilane surfactant
route are suitable for successive reactions in order to introduce
additional functionality to the zeolite surface. Even though the
attainable mesopore surface area is often markedly smaller
than in the amorphous MCM or SBA materials, surface
Mesoporous Zeolitic Materials 267
Distribution (ml g1 nm1)
0.6
300
0.4
Ca2+-form
0.2
Adsorbed amount (ml g1)
0
0 10 20 30 40 50
200
Pore size (nm)
Na+-form
100
0
0 0.2 0.4 0.6 0.8 1.0
Relative pressure (P/P0)
functionalization is very attractive on zeolitic materials because
they are generally more stable and contain additional acidity.
The group of Ryoo realized this idea on mesoporous MFI,
BEA, and LTA that were synthesized following their organosi-
lane surfactant route as described above. For example, after
calcination the reaction with oxoiminopropyl triethoxysilane
established effective surface functionalization. Subsequent
complexation of palladium resulted in a platform for the
Sonogashira coupling reaction (see Figure 11).137
The group of Ryoo also extended their organosilane
surfactant-directed synthesis to mesoporous aluminopho-
sphates such as AlPO-5, SAPO-5, CoSAPO-5, CrAPO-5, and
AlPO-11.142 The template TPHAC survived the synthesis tem-
perature of 200  C and directed the growth of 2050-nm
nanocrystals into aggregated multimicron-sized particles. The
synthesis conditions for SAPO-5 with TPHAC were studied by
Danilina et al.143 When compared to the reference material
SAPO-5 made without a second porogen, the TPHAC-
templated SAPO-5 showed a lower BET surface area as well as
a markedly reduced micropore volume. Nevertheless, the
mesoporous SAPO-5 was much more active in the sterically
demanding alkylation reactions of benzene and cyclohexyl-
benzene with benzyl alcohol than the solely microporous
counterpart. It was not possible to increase the mesopore
volume by increasing the TPHAC concentration. Some addi-
tional amorphous material was always obtained, which the
authors believe is partially responsible for the catalytic conver-
sion. Hard templating with carbon black or soft templating
with amphiphilic surfactants resulted in inferior materials.
TPHAC was further successfully used to create large acces-
sible surface areas composed of an aluminum-rich sodalite. Its
inner surface area is inaccessible for small molecules due to the
268 Mesoporous Zeolitic Materials

Zeolite
framework
Grafting of dichloro--oxoiminato-Pd on mesopore walls

O Pd(CH3CN)2Cl2 O
Porous Porous NH O
Mesopore O Si NH O O Si
materials CHCl3, 65 C materials pd
O O
Cl Cl
O O O O O O O O O O O O O O O O O O
Si Si Si AI Si Si Si AI Si Si Si Si AI Si Si Si AI Si
- - - + - Na +
OH OH OH Na+ OH OH OH + O O O O O O
Na -RO Na
Si Si 1 mol% Pd catalyst, Cul
-
RO
Si
OR
-
R R R Cl + R1 R1
R R
H2O : DMA, Na2CO3
+ +
Toluene Na
-
Si
+
Si
Na OH OH OH Na OH OH OH O O O Na O O O
- - reflux - -
O
Ai
O
Si Si Si
O O O
AI Si Si Si
O O O O O
AI Si Si Si AI Si Si Si
O O O O O O O O
Sonogashira reaction

Figure 11 Mesoporous zeolites synthesized with TPOAC and successively functionalized for palladium grafting.137

small pore opening of 0.28 nm. However, a tenfold increase of comparable to that of mesoporous zeolite Y (130 m2 g1).
the outer surface area as compared to common sodalite and its This synthesis is especially noteworthy because the high alumi-
stronger basic character compared to CsX zeolites made this num content (Si/Al ratio 1.01.5) is prohibitive for a synthesis
usually unremarkable zeolite a promising candidate for basic route via desilication, and this report is the first one document-
catalysis such as the Knoevenagel or ClaisenSchmidt conden- ing the synthesis of mesoporous zeolite X. The complex archi-
sation reactions.144 tecture observed in these large micron-sized particles is not
A follow-up study by the group of Ryoo on the mesoporous commonly observed using TPHAC, which often tends to result
zeolite LTA verified that the mesopores were indeed templated in unstructured nanoparticle aggregates.
by TPHAC and did not simply form through nanozeolite
assembly. When the concentration of the meso-template was 7.11.3.3.2 Seed silanization with neutral
increased, a pore extension from 6 to 10 nm was observed. As alkylamino siloxanes
the micellar diameter of TPHAC is smaller than the obtained In the year 2006, the research group of Serrano presented the
mesopore diameters, it is believed that this template acts as synthesis of mesoporous zeolite ZSM-5 and beta, almost simulta-
its own pore extension agent by incorporating excess molecules neous with the group of Ryoo. Their work relies on the interrupted
into the covalently bonded bilayer that decorates the mesopore zeolite growth through seed silanization using commercially
boundaries. A pore extension to 24 nm was even possible available uncharged organosilanes as direct additives to the syn-
upon enrichment with the triblock copolymer EO20PO70EO20 thesis gels.145 This method consists of a three-step procedure:
(P123). Xe-nuclear magnetic resonance (NMR) studies revealed (1) zeolite seeds are first crystallized by refluxing the zeolite gel
that the diffusion process in mesoporous LTA occurred more at 90  C, (2) the seeds are then functionalized under reflux with
rapidly (200-fold increase) as compared to common LTA.138 phenylaminopropyl trimethoxysilane (PHAPTMS), and (3) are
Moreover, pulsed-field gradient NMR spectroscopy estab- finally crystallized under hydrothermal conditions. Mesoporous
lished an increase in the intracrystalline diffusivities and a nanoaggregates are obtained after centrifugation and calcination
decrease in the transport resistance for water diffusion in LTA. (see Figure 13). The precrystallization into seeds is necessary for
Similarly, the molecular diffusion of water in mesoporous the silylating agent to form a passivating layer around the proto-
beta, made with varying concentrations of the siloxane tem- zeolitic seeds and no crystallization occurs when leaving this step
plate (CH3O)3SiC3H6N(CH3)2C3H6N(CH3)2C3H6 (as cited out.146 The resulting MFI materials obtained with the methods of
in the reference, here abbreviated Tdi(PA)) that resulted in Ryoo134 and Serrano are almost identical with respect to total
58-nm mesopores, was found to be increased threefold.139 surface area (about 590 m2 g1), mesopore surface area (about
Few reports deal with the synthesis of mesoporous Y, a large 350 m2 g1), and mesopore size (about 5 nm) and both lead to
pore zeolite that was found to be highly active in the hydro- the aggregation of very small, about 10 nm, fully crystalline nano-
desulfurization reaction of bulky dimethyldibenzothiophene zeolites. Many of these nanozeolites show the same lattice orien-
when loaded with Pd particles. This zeolite was synthesized tation indicating some intergrowth. This might also explain the
with the help of TPOAB ([3-(trimethoxysilyl)propyl]octadecyl- very high microporosity of these materials, which is often reduced
dimethylammonium bromide) and contained mesopores of when other methods such as leaching or polymer-induced aggre-
about 18 nm and a mesopore surface area of 142 m2 g1, gation are used, as described earlier. By varying the precrystalliza-
about twice than that of zeolite Y made without the mesoscale tion conditions, concentration of the organosilane, or even the
template. On the other hand, Pd-loaded mesoporous zeolites tail group in the latter (from a phenyl to a smaller isobutyl group),
beta and ZSM-5 were much less effective, which was attributed it was possible to influence the textural properties of zeolite beta,
by the authors to their restrictive micropore size limiting diffu- mordenite, or ZSM-5.146,161
sion of the reactants.140 The aluminum-rich mesoporous coun- Marked changes in the total surface area and slight shifts of
terpart of zeolite Y (zeolite X) was constructed very recently by the mesopore dimensions were especially noted in zeolite beta.
using TPHAC. In this case, well-defined nanosheets formed A very high surface area of 857 m2 g1 (mesopore surface
structured assemblies with narrow intracrystalline mesopores 145 m2 g1) was gained with higher concentrations of
of about 7 nm and a globular appearance featuring additional PHAPTMS, enabling a maximum conversion in the cracking
macropores (Figure 12).141 The mesoporous surface area is of the bulky LDPE. The conversion of this reaction is highly
 Mesoporous Zeolitic Materials 269

(a) (d) 350

300

250

Vads(cm3 g1 STP)
200

150
10 m 100

(b) (e) 50

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
P/P0

0.07

0.06

2 m 0.05

dV/dd (cm3 nm1 g1)

(c) (f) 0.04

0.03

0.02

0.01

0.00
2 3 4 5 6 7 8 9 10 11 12 13 14 15
400 nm dp (nm)

Figure 12 Formation of mesoporous aluminum-rich FAU zeolite X via silylation with [3-(trimethoxysilyl propyl)]hexadecyl dimethyl ammonium
chloride (TPHAC). SEM of reference zeolite X (ac) and mesoporous zeolite X (df). The corresponding nitrogen sorption isotherms and pore-size
distributions of reference (black) and silylated X zeolite (open) materials are also shown. Adapted with permission from Inayat, A.; Knoke, I.; Spiecker, E.;
Schwieger, W. Angew. Chem. Int. Ed. 2012, 51, 19621965. Copyright (2012) John Wiley and Sons.

350
ZSM-5 (PHAPTMS) ZSM-5 (PHAPTMS)
300
STP)

250
-1

200
Vads (cm g
3

150

100

50 ZSM-5 Beta
ZSM-5 (PHAPTMS) Beta (PHAPTMS)

200 nm 20 nm 0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(a)
P/P0 P/P0
Beta (PHAPTMS) Beta (PHAPTMS) 0.32 ZSM-5 Beta
dVads/d log(DP) (cm3/g-A)

ZSM-5 (PHAPTMS) Beta (PHAPTMS)

0.28
0.24
0.20
0.16
0.12
0.08
0.04

200 nm 20 nm 0.00
2 4 6 8 10 12 2 4 6 8 10 12
Pore diameter (nm) Pore diameter (nm)
(b)

Figure 13 Aggregated nanozeolites obtained through seed silanization; zeolite ZSM-5 (top left) and zeolite beta (bottom left) and corresponding
nitrogen sorption data (right, a) of parent zeolites (black) and mesoporous zeolites (light) including pore size distributions (b). Adapted with permission
from Serrano, D. P.; Aguado, J.; Escola, J. M.; Rodriguez, J. M.; Peral, A. Chem. Mater. 2006, 18, 24622464. Copyright (2006) American Chemical
Society.

influenced by the amount of the accessible external surface
area.147 This phenomenon was also observed when
microwave-prepared TS-1 zeolites were made with varying
amounts of PHAPTMS and studied in olefin epoxidation
reactions.148 The hierarchical features (larger mesopore
volumes and pore sizes) were further enhanced upon addition
of auxiliary alcohols such as 2-propanol or methanol.149
According to the authors, this leads to a decrease in gel viscos-
ity, thus increasing the silanization of the nanozeolites and
additionally allows for an alkoxylation of the surface,
270 Mesoporous Zeolitic Materials

enhancing the protective layer structure (see Figure 14). It was ZSM-5.150 Similar morphologies are obtained as those
also noted that these hierarchical zeolites were sensitive toward reported by Serrano, with a mesopore size of 3 nm that did
the calcination conditions, in contrast to common ZSM-5. A not change upon increasing the degree of silylation. However,
two-step calcination, first in nitrogen at 400  C followed by air the mesopore volume did scale with silylation, up to a three-
at 550  C, stabilized the aluminum distribution and conse- fold amount when compared to the untreated sample (external
quently showed a better catalytic cracking activity.162 Hierarchi- surface area up to 209 m2 g1, from 39 m2 g1).
cal zeolite beta and ZSM-5 made by seed silanization found Three different commercial organosilanes with increasing
promising applications in biodiesel production from microal- complexity have been used by Guo et al. for the synthesis
gae oil163 and even under the drastic conditions (400  C of ZSM-5. They were added either directly to the zeolite gel or
1.5 bar) of the Ni-catalyzed hydroreforming of degradation to a solution containing preformed seeds according to the
products from LDPE.164 method of Serrano et al.; the zeolite synthesis was thus
Instead of using preformed zeolite seeds Xue et al. used performed in a two- or three-step crystallization procedure. 3-
native fumed silica for the synthesis of hierarchical LTA. In this Aminopropyltrimethoxysilane (APTMS), N-[3-(trimethoxysilyl)
two-step procedure, they immediately functionalized the silica propyl]ethylenediamine (TMPED), and (N,N-diethyl-3-amino-
with PHAPTMS without any pretreatment and then used it in propyl)trimethoxysilane (DATMS) were used as additives. The
the hydrothermal synthesis.153 According to the authors, the major finding was that a precrystallization of the zeolite gel is not
rationale for this approach was the formation of nano- necessary and that aggregated nanocrystals formed with pores of
supercages within the zeolite structure. They argued that this mainly 2.8 nm under both conditions. The highest surface area
direct procedure made the zeolite formation independent of the of 606 m2 g1 was obtained with the smallest silylating agent
different pretreatment conditions that otherwise influenced the APTMS as compared to 379 m2 g1 for the unsilylated parent
seed aggregation. With their new route, Xue et al. obtained material. These zeolites were used as catalysts for benzene alkyl-
nearly micron-sized LTA crystals with tunable surface areas as ation and showed slightly higher activity and selectivity for eth-
well as increasing mesopore volumes reflecting an increasing ylbenzene than the parent compound.151
silanization degree of the silica precursor when the silylating
agent was offered in higher concentrations. The pore size
remained unchanged at about 3 nm, but the mesopore volume 7.11.3.3.3 One-step reactions with alkyl-containing
could be tuned up to 0.27 ml g1 and the external surface area silylating agents
up to 360 m2 g1 (with loss in microporosity). Following The decoration of zeolite precursor particles with hydrophobic
adsorption studies of molecules with increasing diameters, spacers via siloxane bridges was usually performed with bulky
including the large Jacobsen salen complex (1.54  1.2 nm), (or long-chain) reagents in multistep reactions as discussed in
the authors concluded that the mesopores are of inkbottle-type the preceding sections. It seems, however, that similar results
shape with 3-nm cages that are connected by channels of regarding the mesoporosity and surface areas can even be
0.71.2 nm diameter, affording a 170 times faster diffusion obtained in single-step reactions involving small, commercially
rate for the exchange of calcium against magnesium ions. available alkyltriethoxysilanes. Hence, when simple alkoxysi-
Seed silanization was recently reported with a new com- lanes (methyl-, propyl-, and octyl-triethoxysilane (MTES, PTES,
mercial silylating agent, named N-(vinylbenzyl)-2-aminoethyl- and OTES)) were combined directly with the synthesis solution
3-aminopropyl trimethoxysilane (TPAV), for the synthesis of of ZSM-5, filterable mesoporous MFI materials with very high

N N
HO

N
HO

HO

OH

(a) Silanization HO
OH (c) Silanizationalkoxylation
HO agent interactions
ZSM-5
N

Protozeolitic units nanocrystal

N

O
OH
HO

OH
HO

H
HO

OH

N N
CH3
OH
N CH2 N
HO OH H2C Grafted silanization
N agent (PHAP)
HO CH2
N
N
O
HO

O
OH

HO

O
OH

H N
OH

OH Structure directing
O agent (TPA*)
O
(b) Silanization and HO
N

alkoxylation
N

OH

O
OH

H
OH

N N

Figure 14 Schematic diagram showing the seed silanization without (a) and with incorporation of alkoxy moieties by alkoxylation when the silanization
proceeds in the presence of alcohols (b) as well as the interaction between PHAPTMS and n-butoxy grafted species in close vicinity on the surface of
protozeolitic units (c). Adapted from Serrano, D. P.; Aguado, J.; Escola, J. M.; Peral, A.; Morales, G.; Abella, E. Catal. Today 2011, 168, 8695. Copyright
(2011) with permission from Elsevier.

total surface areas (up to 526 m2 g1) and external surface areas
(262 m2 g1) were obtained that were comparable or even better
than those discussed above.152 The morphology is derived from
about 20-nm nanocrystals that are aggregated into spherical
particles of up to 500 nm, showing no loss in micropore volume
with respect to untreated ZSM-5. The concentration of the silanes
as well as their hydrophobicity (chain length) could be used to
modulate the size of the aggregated particles. By contrast, the
pore size was unaffected and centered around 5 nm. Best sam-
ples were obtained with a 10% addition of PTES to the synthesis
mixture. These MFI zeolites showed also highest conversions in
alkylation and condensation reactions involving larger mole-
cules, while the activity with small substrates was similar for all
materials, including the unmodified ZSM-5.
A new concept for creating interrupted growth in silicoalumi-
nophosphates has just been presented.154 In order to increase the
isomerization of large n-alkanes to dibranched products it was
argued that especially 1D microporous zeotypes could benefit
from additional large-pore reaction spaces. SAPO-11 was chosen
because of its known isomerization activity, and its one-step
synthesis was modified by adding an alkylphosphonic acid (tet-
radecylphosphonic acid, TDPA) as a mesopore mediator. Large
46-mm pseudo-spherical aggregates were obtained with or with-
out a modifier; however, only with the TDPA modifier the
authors observed a markedly increased surface area. The TDPA
modifier had apparently donated the P atoms as framework-
terminating elements and generated additional intracrystalline
mesopores of 5 nm by virtue of its long alkyl residues (see
Figure 15). Changes in mass-transfer ability were probed with
2,4 dimethylhexane, which showed dramatically enhanced diffu-
sion in the mesoporous SAPO-11. When decorated with Pt nano-
particles, SAPO-11 provided the anticipated higher dibranching
activity in the isomerization of n-octane.
7.11.4 Bottom-Up Synthesis of Hierarchical
Zeolites with Single Templates
The bottom-up synthesis of mesoporous zeolites focuses to a
large part on the independent action of two SDAs generating
10 nm
Figure 15 TEM and nitrogen sorption measurements of mesoporous H-SAPO-11-HI made in a one-step reaction with addition of tetradecylphosphonic
acid as secondary template. Adapted from Fan, Y.; Xiao, H.; Shi, G.; Liu, H.; Bao, X. J. Catal. 2012, 285, 251259. Copyright (2012) with permission
from Elsevier.
Mesoporous Zeolitic Materials 271
zeolitic micropores on the one hand and mesopores on the
other hand. In this chapter, we discuss synthetic strategies that
rely on the sole presence of only one template. Under these
conditions, the introduction of a mesoscale pore system has to
be achieved either by a change in reaction conditions or by
endowing the template with a dual-scale functionality. In the
former case, the synthesis can be driven to result in nanozeolites,
rendering the secondary template obsolete by the ability of the
nanoparticles to pack into assemblies with self-generated inter-
stitial mesopores. Intrinsically short diffusion distances and
high external surface areas come as a bonus. The second
approach is elegantly realized when using multifunctional tem-
plating agents that are able to direct the zeolite growth on the
micro- as well as the mesoscale simultaneously in a synergistic
way. This approach stimulates the formation of morphologies
with seemingly exfoliated zeolite surfaces similar to 2D zeolites.
Two-dimensional or layered zeolites, consisting of sheets
with only one-unit-cell thickness, have recently attracted increas-
ing attention. They either are formed from lamellar precursors
or are synthesized directly with the sole aid of common zeolite
templates. One of the best-known members of this group is
MCM-22 belonging to the MWW family. To date, the discovery
of 2D zeolites has mostly relied on serendipity. Often they are
intermediates on the way toward formation of microporous
lattices and they might collapse upon calcination. Recent pro-
gress in the preparation and rearrangement of layered zeolites
through pillaring or delamination has been reviewed8 and here
we focus on the synthesis of thin-walled zeolite architectures by
using specially designed bifunctional templates.
7.11.4.1 Bifunctional Structure-Directing Agents
A new avenue for a controlled synthesis of seemingly exfoliated
zeolite architectures has been introduced in the last few years
by the research group of Ryoo based upon a templating route
using individual extended bifunctional templates. These are
able to act as SDAs on two different length scales, where only
one part of the template stimulates zeolite growth as opposed
to the whole molecule. The Ryoo group found that alkylchain-
modulated multiquaternary ammonium ion templates were
300
0.8
Volume adsorbed (cm3 g1)
250
dV/d log D (cm3 g1)
0.6
0.4
200
0.2 H-SAPO-11-HI
150
0.0
1 10 100
Pore diameter (nm)
100
50 H-SAPO-11
0
0.0 0.2 0.4 0.6 0.8 1.0
Relative pressure (P/P0)
272 Mesoporous Zeolitic Materials

suitable for structuring zeolites into two-dimensionally
ordered extended sheets. This is achieved by providing in one
part of the template micropore-templating ammonium cations
for the nucleation of the zeolitic entities that are spaced
through short alkyl linkers. These cations are then either asym-
metrically linked to or symmetrically embraced by long hydro-
phobic chains directing the mesopore formation. This basic
concept has been used to create a variety of complex bifunc-
tional templates that allowed the synthesis of disordered 2D
zeolite sheets, even with predetermined wall thickness, as well
as highly ordered 3D structured sheet assemblies. This work
was discussed in recent reviews and highlights165 and will be
detailed and updated below.
Using these newly developed linear gemini-type diquaternary
templates with hexamethylene linkers such as C22H45
N(CH3)2C6H12N(CH3)2C6H13 (also labeled C22-66, for
a generalized formula of these asymmetric templates see
Figure 16) enabled the Ryoo group to create very thin zeolite
nanosheets consisting of only three pentasil layers (about
2 nm) of MFI (see also Figure 17(a)). These 2D sheets were
effectively held apart from adjacent sheets with a spacing of
2.8 nm controlled by the length of the surfactant chains.166
Calcination removed these soft spacers and the structural
order was lost but since the multilamellar nanosheets were
macroscopically intergrown with each other like a house of
cards they functioned as a kind of solid spacer themselves and
maintained some mesoporosity with about 2-nm pore diame-
ter. Unilamellar sheets could also be formed when the sodium
concentration in the gel was reduced. Here, an overall smaller
size of the single sheets caused their irregular assembly that
stabilized mesopores of about 15 nm upon calcination. Both
morphologies were more active in catalytic performance with
bulky molecules and showed a better lifetime for methanol to
gasoline (MTG) conversion than common MFI materials. The
modular arrangement of these multifunctional templates
allowed for a systematic variation of the interlamellar spacing
by changing the length of the alkyl chains. Shortening the
surfactant tail from C22 down to C12 shifted the low-angle x-
ray diffraction peak in the uncalcined samples accordingly.167
Additionally, pillaring of the nanosheets with TEOS in a post-
synthesis treatment before calcination proved to be very effec-
tive in retaining a large surface area (615 m2 g1 as compared to
505 m2 g1 without pillaring). However, changing the hexam-
ethylene head groups from C22-6-6 to C22-6-0, C22-6-3, or C22-6-
8 disturbed the perfect match with the MFI crystal lattice and
either prolonged the synthesis time or resulted in less ordering.
As the C6 linkage seemed to be most suitable for nanosheet
formation, the concept was developed further for directing
the thickness of the sheets by inserting additional C6 frag-
ments with alkylammonium moieties into the asymmetric
templates.168 By offering not only two but three or four C6
linker groups (n 2, 3, 4), the number of nanosheets increased
accordingly from three to five to seven parallelly aligned
zeolite planes (see Figure 17). Two adjacent ammonium
groups are nevertheless a requirement for zeolite structur-
ing. Using only one C6 alkylammonium headgroup led to
mesoporous silica instead of ZSM-5 (see Figure 17(d)).
MFI nanosheets synthesized with the hexadecyl analog

Asymmetric bifunctional templates

CH3 CH3
+ + m = 6,8,10,12,16,18,22
CmH2m+1 N CiH2i N CkH2k+1 i = 3,6,8
n = 1,2,3,4
CH3 CH3 n1 k = 3,4,6

Symmetric bifunctional templates

CH3 CH3 CH3

+ + +
CmH2m+1 N (C6H12)N (C6H12) N CmH2m+1 N3
CH3 CH3 CH3

CH3 CH3 CH3 CH3

+ + + +
CmH2m+1 N (C6H12)N (C6H12)N (C6H12) N CmH2m+1 N4
CH3 CH3 CH3 CH3

CH3 CH3 CH3 CH3

+ + + +
CmH2m+1 N (C6H12)N CH2 CH2 N (C6H12) N CmH2m+1 N4-phe
(also N6-diphe, N8-triphe)
CH3 CH3 CH3 CH3

Figure 16 Generalized structures of asymmetric as well as symmetric gemini-type templates studied by the Ryoo group. The suffix Nx accounts for the
number of ammonium cations in the whole surfactant. The asymmetric units contain two different surfactant tails as end groups, as opposed to the
symmetric templates.
 Mesoporous Zeolitic Materials 273

(a) 3 (b)
5
3 3 CH3 CH3
3 5 + +

3 5 C22H45 N C6H12 N C6H13

3 5 CH3 CH3 n1

(a)
10 nm 5
10 nm

(c) (d)

10 nm 10 nm 500 nm

Figure 17 TEM (left) and SEM (right) micrographs of MFI nanosheets assembled using a bifunctional template with increasing number of ammonium
cation centers n. (a, a) n 2, (b) n 3, (c) n 4, (d) n 1; as in the generalized bifunctional template formula (top right). Values in (ac) indicate the
number of MFI sheets formed with varying n. Adapted with permission from Park, W.; Yu, D.; Na, K.; Jelfs, K. E.; Slater, B.; Sakamoto, Y.; Ryoo, R. Chem.
Mater. 2011, 23, 51315137. Copyright (2011) American Chemical Society.

(a)

]
01
[1
c
a
MFI

500 nm 10 nm

(b)

Figure 18 Hexagonally ordered zeolites formed with symmetrical gemini-type surfactants: (a) SEM, TEM and graphical depiction of the ordered
mesoporous MFI-type zeolites, (b) examples for symmetric bifunctional templates.165b

[C16H33N(CH3)2C6H12N(CH3)2C6H13] showed a 18-N3-18 module (C18H37N(CH3)2C6H12N(CH3)2

good catalytic activity and catalyst lifetime in the Beckmann
rearrangement of cyclohexanone oxime to e-caprolactam.169
Extrapolating the bifunctional templating route to yet a
new combination of meso- and micro-templating modules
led to the first ever examples of truly hexagonally ordered
mesoporous zeolites.170 Here, Na et al. used a symmetric
C6H12N(CH3)2C18H37(Br)3) where the long surfactant
units direct a cylindrical arrangement of thin, 1.7-nm crystal-
line MFI-type walls (see Figure 18). Similar structures were
found with the corresponding 22-N4-22 porogen rendering
tremendous surface areas of over 1000 m2 g1 and mesopores
of about 3.5 nm. Three low-angle diffraction peaks in the x-ray
274 Mesoporous Zeolitic Materials
diffractogram confirmed the long-range ordering of the meso-
pores. Adding one, two, or three bridging phenyl groups to the
center part of the bifunctional SDA (Nx-phe) initiated crystal-
lization of a different zeolite. Now, mesostructurally disor-
dered but extremely small beta particles with very large
surface areas between 780 and even 940 m2 g1 and about
4-nm mesopores were formed. As observed before, the thick-
ness of the zeolite domains was directed by the number of
micropore-structuring modules, in this case the ammonium
units bridged by phenyl rings. Owing to their large external
surface area, high catalytic conversions in acid-catalyzed reac-
tions were accomplished with these structures.
Using the same mesostructuring symmetric templates but
different reaction conditions allowed for the formation of a
macromesomicro-morphology of zeolite beta.171 Macro-
pores contained in diatomaceous earth (used here as silica
source) were retained to some extent when the hydrothermal
conversion was terminated early enough. Alternatively, 3.7-
and 30-nm mesopores were obtained when a sodium-free
synthesis was performed with TEOS.
Slight changes in the gel composition for the MFI zeolite
containing the 18-N3-18 template created nanosheets of only a
single-pore thickness of 1.5 nm, which is smaller than one
unit cell.172 Larger versions of the symmetric template with N4
or N5 ammonium groups increased the thickness to two- and
three-pore layers, thus suggesting that the respective two outer
ammonium cations are located at the pore entrances while the
inner cations template the zeolite wall. Symmetric gemini sur-
factants with only two ammonium groups of formula
CmH2m1N(CH3)2C6H12N(CH3)2CmH2m1 with tail authors were able to generate beautiful orthogonally interwoven
compositions of m 4, 6, or 8 resulted in MFI particles of
different sizes and structures, while m 10 generated only
mesoporous materials.173 Samples with m 4 consisted of
>300 nm purely microporous MFI zeolites. Sponge-like aggre-
gated nanoparticles of 1020 nm were observed with m 6,
resulting in a broad distribution of mesopores around 20 nm.
Using m 8 finally gave small needle-shaped thin flakes that
formed slit-like pores of around 2 nm. In these three samples,
the BET surface areas increased from 360 m2 g1 as commonly
observed for bulk ZSM-5 up to 580 m2 g1 in the latter sample.
These samples were analyzed in detail with respect to hydro-
thermal stability and acidic properties, which are important
parameters for possible future applications in cracking reac-
tions or as automotive exhaust catalysts. It was found that the
amount of the external acid sites increased with decreasing size
of the particles, that these sites were less acidic than the internal
acid sites but strong enough to catalyze the MTG reaction, and
that their steam stability exceeded that of the internal acid sites.
It is obvious that this bifunctional modular templating
strategy developed by the Ryoo group constitutes a promising
toolbox for generating structurally well-defined mesoporous
zeolite materials featuring extremely short diffusion paths to
their internal active sites and large accessible external surface
areas. This work also points to the potential of generating
highly anisotropic morphologies of other zeolite families fol-
lowing a similar rational strategy.
A recent example of a mesoporous zeolite that was made
with a similar bifunctional template is SSZ-13 with CHA topol-
ogy. As described in Section 7.11.2.2, desilication of CHA-type
crystals turned out to be detrimental for these types of zeolites
because of their low framework aluminum stability. However,
when mesoporous CHA was synthesized following this
bottom-up approach by using the micropore template N,N,
N-trimethyl-1-adamantanammonium hydroxide (TMAdOH)
and the multifunctional surfactant C22H45N(CH3)2
(CH2)4N(CH3)2C4H9Br2 (C22-4-4) simultaneously (molar
ratio of 2:1) large aggregates of chabazite nanocrystals were
formed instead of uniform single crystals when the surfactant
was omitted.174 The specific surfactant was chosen from a
number of similar structures based on molecular dynamics
simulations and was experimentally proven to be indeed the
best candidate regarding the resulting zeolite crystallinity. The
amount of costly bifunctional template was reduced this way,
but it did still function as a structure-specific spacer that limits the
crystalline domains to nanoparticles. Micropore volumes of orig-
inal and mesoporous SSZ-13 were similar (0.18 and 0.21 ml g1,
respectively) while the surface area was markedly increased
and a substantial amount of mesopore volume was acquired
(0.21 ml g1) in the latter. These mesoporous CHA zeolites
showed an improved stability in the MTO conversion as op-
posed to those posttreated with NaOH during desilication.35
Very recently it was shown that only 2-nm-thin nanosheets of
MFI can even be arranged in a house-of-cards morphology
when synthesized with the sole use of a common monofunc-
tional template (see also next Section 7.11.4.2) such as tetrabu-
tylphosphonium (TBPOH) or tetrabutylammonium hydroxide
(TBAOH) in a single-step reaction.175 The group of Tsapatsis
exploited the tendency of spatial branching through epitaxial
and rotational intergrowth occurring in MFI nanocrystals or
between the related MFI/MEL zeolite structures and these
nanosheets over a range of conditions and Si/Al ratios (see
Figure 19). Thirty to forty percent of the acid sites are now
located on the outer surface area and are accessible to larger
molecules that are otherwise unable to enter the micropores of
ZSM-5. As a result, the rate constant of the liquid-phase alkyl-
ation of mesitylene by benzyl alcohol increased by more than
two orders of magnitude when compared to common ZSM-5.
7.11.4.2 Mesopore Formation Through Nanoparticle
Assembly with Monofunctional Templates
Most concepts for the preparation of hierarchical zeolites
described so far either rely on pre-made crystalline zeolites
that are exposed to secondary reactions (e.g., partial dissolu-
tions of various forms performed with conventional zeolite
particles or assisted aggregation of premade nanozeolites) or
use secondary templates (hard, polymer, or alkoxy silanes) in
addition to the common zeolite SDAs. Only the intriguing
concept of bifunctional templates described in the last chapter
shows that mesoporosity might be imparted into a zeolite
structure directly during the synthesis using a single structure-
directing additive. In this case, however, the templating agents
have to be carefully designed and may have to be specially
optimized for each zeolite structure, thus imposing the need
for adaptation and special synthetic protocols. They also
impose restrictions regarding the reaction conditions due to a
more limited thermal stability.
When aiming to create hierarchical zeolites as easily and
economically as possible, it would be highly advantageous if
only a single structure-directing additive of simple structure
was needed, requiring a minimum number of reaction steps.

(a) (b)
200 nm 20 nm
Figure 19 Intersecting lamellar zeolite nanosheets created through repetitive branching in a single-step reaction with tetrabutyl ammonium hydroxide
monofunctional template; (a) and (b) TEM images, (c) experimental (i) and simulated XRD patterns, assuming either an MFI/MEL (ii) or solely MFI
(iii) orthogonal intergrowth. Reproduced from Zhang, X.; Liu, D.; Xu, D.; Asahina, S.; Cychosz, K. A.; Agrawal, K. V.; Al Wahedi, Y.; Bhan, A.; Al Hashimi,
S.; Terasaki, O.; Thommes, M.; Tsapatsis, M. Science 2012, 336, 16841687, with permission from AAAS.
Recently, this challenge was addressed by several research
groups, and recent publications focus on synthesis strategies
directed at the aggregation of nanozeolites into stable micron-
sized zeolite assemblies that inherently contain interparticle
mesopores. Several important aspects are combined this way:
costs are reduced by using mainly well-established templates, a
network of accessible secondary mesopores is created as a
result of the aggregation process, and large external surface
areas as well as short internal micropore diffusion pathways
intrinsic to nanozeolites are created. Different approaches can
be taken for assembling nanozeolites. Sometimes a specific
molecular template is necessary or the reaction includes special
pretreatment procedures, it may be conducted under uncom-
mon conditions, the synthesis may require seeds, or it is per-
formed with concentrated gels. These different strategies
relying on a single micropore-forming template will be dis-
cussed in the following; they are summarized in Table 3.
7.11.4.2.1 Nanoparticle aggregation using single templates
For mesostructuring zeolite beta, Choi et al. applied a novel
cyclic diammonium cation (CDM1) in place of the usual tetra-
ethylammonium hydroxide SDA (see Figure 20 below for
structural details).176 Using this template, instantaneous gela-
tion and formation of a mesoporous gel was observed that
could be transformed into crystalline zeolite beta nanoparticles
(1015 nm) by maintaining the mesoscale morphology, a pro-
cess termed pseudomorphic crystallization. These nanocrys-
tals were disorderly arranged to form mesopores with a broad
size distribution around 20 nm (mesopore volume about
0.9 ml g1). The unique molecular structure of the CDM1
template featuring a rigid phenyl moiety for bridging the
ammonium cations was proposed to direct the polymerization
of the silicates in a controlled way. Derivatives of this cyclic
diammonium cation facilitated the synthesis of other highly
crystalline nanozeolite aggregates of MTW and MFI structure
(see Figure 20) with similar broad pore diameters around
20 nm.177 It seemed that the rigid phenyl linkers were appro-
priate to template 12-membered zeolite pores in BEA and
MTW while MFI preferred flexible hexamethylene linkages.
Mesoporous Zeolitic Materials 275
(c)
(iii)
Intensity (a.u.)
(ii)
Experiment
(i)
10 20 30 40 50
2q ()
This concept was recently extended to piperidine- and
imidazole-based geminal dicationic liquids for the synthesis
of zeolite beta by Kore et al.178 They synthesized a whole array
of different cyclic as well as linear templates that were used in
their bromide form. It was found that cyclic templates were
preferable for the nucleation of nanosized beta, especially the
phenyl-linked piperidine derivative DCIL4; however, other
ionic liquids assisted in the synthesis of zeolite beta as well
(see Scheme 2). The authors were able to vary the Si/Al ratio to
a great extent (17 to 1) and obtained much smaller intercrystal-
line mesopores centered around 5 nm under reaction conditions
similar to those used by the Ryoo group. External surface areas
between 90 and 290 m2 g1 could be obtained (total pore vol-
umes 0.38 up to 0.65), and the resulting zeolite materials were
active catalysts for transesterification and acylation reactions.
Sometimes it is only necessary to find the appropriate com-
bination of slight changes of the reaction conditions and a
well-chosen, but common, template to obtain micron-sized
particles that consist of numerous smaller crystals creating a
secondary pore architecture by their specific packaging. These
secondary pores are especially advantageous in 1D zeolite
structures that are naturally limited in their diffusional prop-
erties. One example is the medium-pore zeolite AlPO4-11. By
using di-n-propylamine (DPA), a template frequently applied
in these systems, but without the common HF addition, it was
found that large 1020-mm spheres of AlPO4-11 are formed.181
While the outer surface of these spheres was decorated with the
typical bar-shaped micron-long crystals in an interwoven man-
ner, the interior consisted of very small 50-nm crystallites,
being responsible for 7-nm mesopores. An increase of the
water content in the synthesis gel changed the mesoporous
interior into a dense system. The overall surface area was more
than doubled in the mesoporous sample (to about 250 m2 g1)
and comparable to SAPO4-11 made via seed silylation (and the
identical SDA) as discussed in Section 7.11.3.3.3.154 Similar
diamines with different hydrocarbon chain lengths ranging
from C2 to C10 were used in the synthesis of pentasil zeolites.179
Adding ethylene diamine (EDA) or hexane-1,6-diamine (HDA)
to a freshly precipitated aluminosilicate rendered ZSM-5, a sys-
tem that can be obtained with a large number of templates. In
276 Mesoporous Zeolitic Materials

Table 3 Overview of mesoporous zeolites generated with a single monofunctional template

Year and Method (template, reaction Zeolite structure Particle size (aggregate size/ Mesopore Catalysis
reference conditions) primary particle size) size

Template-assisted aggregation and intergrowth

2009176 CDM1 BEA 1015 nm aggregates 20 nm Isopropylation of
2009177 CDM2, CDM3, DM363 MTW naphtalene
MFI
2011178 DCIL4, DCIL5, DCIL6, DCIL9 BEA 15 nm aggregates 5 nm Transesterification,
acylation
2010179 EDA, HAD, DAOT, DADC ZSM-5 58 mm/50100 nm Conv. of phenol w.
ZSM-11 tert-butanol
2011180 MEA, DEA, TEA ZSM-5 1020 mm Conv. of phenol w.
tert-butanol
2012175 TBPOH or TBAOH Intergrown ZSM-5 100200 nm 27 nm Liquid-phase alkylation
of mesitylene
2009181 DPA, no HF, low H2O conc. AlPO4-11 1020 mm/50 nm 7 nm
2012182b MCHA MeAPO4s (Si, Co, Mg, Zn, 10 mm 1020 nm
Mn, Sn, Cr, Fe, V, Zr)
CoAPO4-5

Specialized methods
2006183 Age under reflux 90  C ZSM-5 0.5 mm 1020 nm
2009184a/ TUD-1 TPAOH, SAC ZSM-5 >1 mm 10 nm Aldol condensation
2010185
2010186 Compressed CO2 emulsions ZSM-5, Sil-1 0.2 mm 1040 nm
2010187 HDA Sil-1 seeds ZSM-5 0.20.3 mm 23 nm Alkylation of phenol w.
isopropanol
2012188 TPAOH Sil-1 seeds ZSM-5 >1 mm 1030 nm
2010189 Complex pretreatment of ZSM-5 Nanosized 440 nm
2011190 clear solutions BEA 140 nm 1220 nm
2011191 60-nm nanozeolites, slow TS-1 >50 mm/6080 nm 10 nm
evaporation MMM
2011192a/ Macroporous support, dry ZSM-5, BEA, TS-1, Zr-Sil-1 MMM Epoxidation of styrene
2012193a gel conv. w. glycerol
2011194b TEAOH, SAC BEA >10 mm/20 nm 13 nm
2012195 Dense gel fusion BEA >1 mm 1535
H-BEA tunable
35140
2012196 Concentrated gels LTL 0.4 mm/10 nm 4 nm Functionalization, Hg
capture
2009197 Kaolin precursor SAPO4-5 48 mm 3040 nm MTO
2012198 Ionothermal synthesis with SAPO4-5 12 nm
choline chloride

DPA, di-n-propylamine; EDA, ethylenediamine; HAD, hexane diamine; DAOT, 1,8 diamine octane; DADC, 1,10 diaminodecane; TBPOH, tetrabutylphosphonium hydroxide; TBAOH,
tetrabutylammonium hydroxide; MEA, monoethanolamine; DEA, diethanolamine; TEA, triethanolamine; MCHA, n-methyldicyclohexylamine; MTO, methanol to olefin; MMM,
micromesomacro.

this case, very similar morphologies as in the AlPO4-11 system
were found, that is, 58-mm sized spheres, consisting of 50100-
nm nanocrystals that constitute 0.2 ml g1 mesopore volume
and surface areas of around 350 m2 g1. Larger diamines (1,8-
diaminooctane and 1,10-diaminodecane (DAOT and DADC))
favored the structure of ZSM-11, a 3D ten-ring zeolite such as
ZSM-5 but with a single pore size of 0.53  0.54 nm in all
directions. When compared in the catalytic liquid alkylation of
phenol with tert-butanol, HDA-ZSM-5 was the superior catalyst,
featuring the highest mesopore volume and almost threefold
conversion when compared to a commercial ZSM-5. The same
research group also used commercial alkanolamines instead of
diamines, such as mono-, di-, and tri-ethanolamines (MEA,
DEA, and TEA) at extremely low concentrations (60 SiO2:0.38
SDA), and observed the formation of similar 1520-mm
hierarchical ZSM-5 spheres.180 Well-shaped but relatively
densely assembled nanocrystals giving rise to slit-like mesopores
appeared, with an additional mesopore volume of up to
0.11 ml g1 and surface areas around 300 m2 g1 when DEA
was used (see Figure 21).
The aluminophosphate AlPO-5 is comparable to ZSM-5
with respect to its flexibility in the templating behavior. The
catalytic importance of this structure rests on its ability to
implement a number of metal framework elements, thus incor-
porating acidity and redox-active species. In order to implement
mesoporosity, detailed studies were undertaken by Manjon-
Sanz et al. on metal-doped MeAPOS containing Si, Co, Mg,
Zn, Mn, Sn(II), Sn(IV), Cr, Fe, V, or Zr using N-methyldicyclo-
hexylamine (MCHA) as the SDA. Again, self-assembly into large
spherical agglomerates occurred with interparticle mesopores
 Mesoporous Zeolitic Materials 277

BEA-CDM1 1600
BEA-CDM2

Adsorbed amount (ml g1, STP)

1400
Cl Cl Cl Cl

Distribution
BEA-CDM1

+ 1200
N N +
+ BEA, MTW
N N + BEA-conv

1000 0 20 40 60 80 100
Pore diameter (nm)

500 nm 800
CDM1 CDM2
2
600
BEA-CDM
+
Br 1
N Br + Br MFI BEA-CDM
N N + N 400
+ Br
200 BEA-conv

DM363 CDM3 0
0.0 0.2 0.4 0.6 0.8 1.0

20 nm Relative pressure, P/P0

Figure 20 Zeolite beta made with cyclic bifunctional templates. Reproduced by permission of The Royal Society of Chemistry, rearranged from Na, K.;
Choi, M.; Ryoo, R. J. Mater. Chem. 2009, 19, 67136719. Left, bifunctional templates used. Middle, electron micrographs of the product obtained with
CDM1. Right, nitrogen sorption isotherms of different zeolite beta products and the resulting pore size distributions (inset).

Cl Cl
N N
+ N + + N + + N +
N N N + +
PhH2C CH2Ph N N
Br Br
Br Br Br Br

DCIL4 DCIL5 DCIL6 DCIL9

178
Scheme 2 Cyclic templates that were used successfully by Kore et al. for the synthesis of mesostructured zeolite beta.

150 SiO2/Al2O3 = 30

135
Volume (cm3 g1)

c
120

105 b

90
a
75
0.0 0.2 0.4 0.6 0.8 1.0
P/P0 (U) 10.0 m

Figure 21 SEM of ZSM-5 produced with the single template ethanol amine MEA and corresponding nitrogen isotherms of samples made with (a) MEA,
(b) DEA, (c) TEA. Adapted from Chen, L.; Zhu, S. Y.; Wang, H. M.; Wang, Y. M. Solid State Sci. 2011, 13, 20242029. Copyright (2011) with permission
from Elsevier Masson.
278 Mesoporous Zeolitic Materials

ranging between about 10 and 20 nm, depending on the spe-
cific doping.182 They established a direct relation between the
reaction conditions and the mesoporosity created: the shortest
reaction time of 2 h at 175  C providing crystalline CoAPO-5
showed the highest mesopore volume (0.1 ml g1) and smal-
lest particle (48 nm) as well as mesopore size (9.4 nm).
7.11.4.2.2 Aggregation of nanoparticles by seeds
or gel pretreatment
Adding zeolite seeds to a precursor solution is well established
as a means to promote the rate of crystallization, in addition to
favoring the crystallization of desired phase-pure zeolites.
When the synthesis gel for ZSM-5 was enriched with silicalite-
1 seeds and the same template HDA was used as discussed in
the preceding Section 7.11.4.2.1.179 a markedly shortened
crystallization time from over 3 days for the reference sample
down to 824 h was observed at 160  C.187 Furthermore, it
was observed that the seeds initiated the growth and aggrega-
tion of nanozeolites; now instead of forming micron-sized
spheres, much smaller 200300-nm clusters were obtained
with the nanoparticles well aligned along the crystal c-axis
(see Figure 22(a)). The highly crystalline ZSM-5 nanorods
that are partially intergrown along the same zone axis establish
interstitial mesopores of 23 nm with a mesopore volume of
0.17 ml g1 and 370 m2 g1 surface area. An increased yield of
p-isopropyl phenol was observed in the alkylation reaction
of phenol with isopropanol when compared to an unseeded
ZSM-5 sample. This beneficial behavior was explained by the
much shorter diffusion paths in the assembled nanorods.
Another interesting study documents that seeding can also
be a means to save costly templating agents.188 Here, large
aggregates of nanoparticles in the micrometer range were pre-
pared in a seed-supported synthesis that requires only very
small amounts of the common ZSM-5 template TPAOH. A
seed gel of silicalite-1 was prepared at 100  C for 24 h
containing a SiO2:TPAOH ratio of 1:0.35. This seed gel was
added to a template-free aluminosilicate solution, amounting
to 1.5% up to 10% of the silica content, hence effectively
diluting the SiO2:TPAOH ratio down to only 0.005. Highly
crystalline mesoporous aggregates of 50100-nm nanoparti-
cles were formed after 24 h at 180  C with about 0.2 ml g1
mesopore volume and pore sizes of about 1030 nm, covering
an Si/Al range between 50 and 600.
The addition of seeds can also be a valuable method in a
synthesis where otherwise competitive phases may lead to
unacceptable impurities. On the other hand, a disadvantage
of seed-assisted growth is the need for preparing appropriate
seeds reproducibly ahead of time. Alternatively, the growth of
nanoparticles can be favored by different aging reactions per-
formed with the synthesis gels before they are hydrothermally
converted. One strategy involves pre-aggregation into a nearly
amorphous material consisting of 500-nm spheres through
refluxing a ZSM-5 reaction gel at 90  C for 30 h.183 The con-
version of these spheres into crystalline ZSM-5 follows in a
second step through conventional hydrothermal treatment at
100  C for 1 day without visibly affecting the overall morphol-
ogy (see Figure 22(b)). Very large surface areas (over
704 m2 g1) and total pore volumes (1.13 ml g1) were
obtained on the fully crystalline samples showing mesopores
between 10 and 20 nm. Using a more elaborate precondition-
ing step, aggregates of ZSM-5 with smaller primary (57 nm)
and secondary particles (about 200 nm) were synthesized by
Larson et al. (see Figure 22(c)).189 Thus, starting from highly
diluted clear solutions containing TPAOH as well as TPABr,
reaction conditions were systematically changed. The solvent
was increasingly evaporated at 65  C, removing 25% up to
70%, while adjusting the pH and the remaining gel was then
hydrothermally converted at different temperatures (between
140 and 165  C) and times (1275 h) in order to influence the
primary particle size. The final product still had to be retrieved

500 nm
20 nm
100 nm
50 nm
(a) (b) (c)

Figure 22 Mesoporous zeolites through aggregation of nanoparticles, (a) c-axis aligned ZSM-5 nanorods templated with 1,6 hexylenediamine and
seeded with silicalite-1 (adapted from Wang, D.; Li, X.; Liu, Z.; Zhang, Y.; Xie, Z.; Tang, Y. J. Colloid Interface Sci. 2010, 350, 290294. Copyright (2010)
with permission from Elsevier); (b) SEM of TPAOH templated ZSM-5, preconditioned by refluxing the precursor gel (adapted from Holland, B. T.;
Microporous Mesoporous Mater. 2006, 89, 291299. Copyright (2006) with permission from Elsevier); (c) TEM of TPAOH templated ZSM-5
preconditioned by acidification and evaporation of the precursor gel (adapted from Petushkov, A.; Yoon, S.; Larsen, S. C. Microporous Mesoporous
Mater. 2010, 137, 92100. Copyright (2010) with permission from Elsevier).

via centrifugation and showed mesopores between 4 and
40 nm with a total pore volume of about 0.5 ml g1. The
authors suggest that the excess templating molecules associate
with the negatively charged small silica particles, thereby
screening these particles during aggregation and thus being
responsible not only for the crystallization of the zeolite but
also for the formation of the mesopores. This protocol was also
applied to the synthesis of zeolite beta at about 60% yield of
mesoporous aggregates approximately 140 nm in size consist-
ing of 1640-nm nanocrystals when using the more concen-
trated solutions. A mesopore volume of about 0.5 ml g1 was
obtained with 1220-nm mesopores.190
7.11.4.2.3 Steam-assisted conversion and other specialized
methods
The steam-assisted conversion (SAC) of zeolite precursor gels
has already been mentioned in previous sections, as it has been
found to be helpful under synthesis conditions that require
noncharged (block copolymer) or hydrophobic (carbon) tem-
plating. This method is also sometimes used in common zeo-
lite synthesis as it permits the use of less template or enables
higher Si/Al ratios that would be difficult to reach under usual
hydrothermal conversion conditions. The essential feature of
this procedure is a drying step of the synthesis gel, and subse-
quent conversion in a steam atmosphere. This implies that the
local concentration of reaction partners is maximized and their
mobility in the reaction medium is reduced. These are ideal
conditions to favor nucleation and aggregation of nanozeolites
if the steam conditions are chosen correctly. This important
aspect was pointed out by Matsukata et al. who applied SAC for
the synthesis of a number of nonhierarchical zeolite
structures.199 We have recently used the SAC process to prepare
completely crystalline mesoporous beta zeolites using only the
common TEAOH template. It is important to note that the
amount of water has to be chosen correctly in order to avoid
an incomplete conversion.194 For zeolite beta we have found
that at least a twofold concentration of the amount of water
necessary for establishing a saturated steam is needed in the
autoclave to achieve a 100% conversion into highly mesopor-
ous zeolite aggregates. In this case, very homogeneous meso-
pores with the same narrow size distribution around 13 nm
were determined from either the adsorption or the desorption
branch of the nitrogen sorption isotherm of samples that were
prepared in just 26 h at 180  C. We could show that the
hierarchical zeolite beta aggregates are formed via (initially indi-
vidual) nanozeolites that function as bricks and upon contin-
ued steam treatment convert into a mesoporous zeolite beta
network with interconnected pores, without the need of a sec-
ondary template. Large surface areas of up to 763 m2 g1 and
mesopore volumina over 0.6 ml g1 were obtained due to the
small size (about 20 nm) of the individual intergrown bricks.
This process was applicable to a Si/Al range between 10 and 33.
The SAC procedure was also successfully applied for the
formation of mesoporous ZSM-5 as well as TS-1 through nano-
particle assembly resulting in single-crystal-like mesostructures
following a more complex synthesis protocol. The authors
used a recipe that was originally developed in the group of
Maschmayer et al.200 to generate a 3D mesoporous alumino-
silicate network termed TUD-1, which was formed with the aid
of triethanolamine (TEA). Zhou et al. added the MFI template
Mesoporous Zeolitic Materials 279
TPAOH to the TUD-1 composition and steam-converted the
dry gels at high yield (micropore volume 0.16 ml g1, surface
area 447 m2 g1) into the MFI structure featuring 10 nm
mesopores.184 Compared to common ZSM-5, these materials
showed an increased conversion for the aldol condensation
reaction. A careful follow-up crystallization study highlighted
the importance of the gel aging temperature and the steam
conditions.185 Drying the gel at temperatures higher than
80  C affected the mesoporosity negatively, presumably due
to extensive SiOAl condensation reactions leading to rigid
amorphous aluminosilicates. Similar studies based on the
TUD-1 precursor were aimed at performing the same reaction
without TEA.201 The TPAOH-templated gels were slowly dried
over 7 days at room temperature and treated via SAC at 130  C.
Composite materials with ZSM-5 crystals (150250 nm)
embedded in an amorphous matrix were obtained after 15 h,
featuring rather high micropore (0.13 ml g1) as well as meso-
pore (0.7 ml g1) volumina and a high overall surface area of
580 m2 g1.
A process related to the SAC procedure is a quasi-solid-
state reaction performed by Chen et al.192 In an effort to
synthesize macromesomicro (MMM) zeolite (ZSM-5, TS-
1, and beta) materials, the authors first fabricated a macro-
porous amorphous titanosilicate (intended for TS-1) by the
spontaneous reaction of titanium isopropoxide and tetra-
methyl orthosilicate (TMOS) in water at high pH. The dry
scaffold was then impregnated with a solution containing
TEOS and TPAOH, dried and reacted at 130  C in a glycerol
medium. The formation of TS-1 nanoparticles consumed the
amorphous matrix and resulted in the (disordered) aggrega-
tion of 200-nm-sized TS-1 crystals with 4.7-nm interparticle
spaces arranged around 1-mm macropores (see Figure 23).
This hierarchical TS-1 material showed nearly twice the sur-
face area (580 m2 g1) compared to TS-1 nanoparticles pre-
pared as reference and performed better in the epoxidation of
styrene and 2,4,6-trimethylstyrene. The same authors also
applied this promising strategy to the multimodal synthesis
of Zr-doped silicalite-1 and zeolite beta.193
Mesoporous aluminophosphates were recently obtained
by another unusual protocol based on an ionothermal
synthesis.198 Here, a eutectic solvent made of pentaerythritol
and choline chloride allowed the authors to perform the syn-
thesis of SAPO-5 at ambient pressure through microwave treat-
ment within short times (1560 min). The choline salt acted as
both the solvent and the micropore templating agent at the
same time. The authors adopted an explanation for the occur-
rence of the observed mesopores (around 12 nm with a broad
size distribution) from an earlier study on metal-containing
SAPOs.202 The latter had been prepared by common mono-
templated hydrothermal synthesis and mesopores were created
unintentionally. It was argued that the additional pores were
probably formed through evaporation of water and by carbo-
naceous species from template decomposition during the cal-
cination process, in other words by a process resembling the
steaming of aluminosilicates at higher temperatures.
Gaseous nanobubbles originating from compressed CO2
emulsions or foams were used as indirect mesopore templating
agents in the synthesis of MFI zeolites. Sun and coworkers used
traditional zeolite synthesis gels for silicalite-1 and ZSM-5 and
pressurized the autoclaves with CO2 to between 8 and 15 MPa
280 Mesoporous Zeolitic Materials
(a)
5 m
(c)
500 nm
Figure 23 Macromesomicroporous zeolite TS-1 made from a macroporous amorphous titanosilicate scaffold via SAC in a glycerol medium.
SEM (a and b) and TEM (c and d; bright field and dark field) of MMM-TS-1 after a 3-day conversion at 130  C. Adapted with permission
from Chen, L.-H.; Li, X.-Y.; Tian, G.; Li, Y.; Rooke, J. C.; Zhu, G.-S.; Qiu, S.-L.; Yang, X.-Y.; Su, B.-L. Angew. Chem. Int. Ed. 2011, 50, 1115611161.
Copyright (2011) John Wiley and Sons.
prior to thermal treatment.186 This method efficiently reduces
the crystallization time from 96 h down to 48 h at 150  C
(attributed to carbonate species), and in addition creates intra-
crystalline pores with a broad size distribution between 10 and
40 nm. Higher CO2 pressure increased the mesopore volume
while gels pressurized with nitrogen showed morphologies
similar to common ZSM-5 zeolites; the carbonate species
thus seem to be directly involved in the pore formation.
Moreover, evaporation processes were exploited for the
preparation of MMM TS-1 films. In this case, 6080-nm zeolite
particles were first prepared from clear solutions, purified, and
then processed by slow evaporation from a 4 wt% TS-1 syn-
thesis solution. This way 50-mm-thick films with intercrystal-
line mesopores of 10 nm and 15-mm macropores created
from convective evaporation of the solvents with a total pore
volume of about 0.4 ml g1 were established. Nonthermal
plasma removal of the template sustained the mesoporous
morphology of the films.191
It is well known in zeolite synthesis that the source of the
precursor can have a dominant influence on the growth pro-
cess and that it can even lead to the crystallization of unin-
tended phases. This phenomenon can be exploited to direct the
zeolite growth into a hierarchical morphology. In this context,
the layered natural mineral metakaolin was used by Zhu et al.
as an alternative Si/Al source (vs. pseudo-boehmite and silica
sol) for the synthesis of SAPO-5.197 Without using an addi-
tional mesopore template, the metakaolin induced the growth
of 60-nm thin sheets that developed in time into 48-mm
spheres of nearly perpendicularly oriented interwoven planes
with pores of 3040 nm. The authors explain this self-
organized growth by the specific structure of their precursor,
where two silica layers form a confined space at the nanometer
(b)
500 nm
(d)
500 nm
scale filled with the alumina source that initiates the crystal
growth along a preferred orientation. The resulting SAPO-5
spheres were highly stable against severe steaming treatments
and showed a much higher MTO activity while retaining the
selectivity, when compared to SAPO-5 of similar macroscopic
size and Brnsted acidity. These striking direct comparisons
convincingly show the importance of diffusion for reactions on
zeolites by reducing the internal diffusion paths down to the
nanoscale.
7.11.4.2.4 Nanoparticle aggregation from
concentrated gels
Using the SAC procedure, the reaction gels are first completely
dried and then converted by introducing only small amounts
of water in the form of saturated steam. It can be viewed as the
most concentrated gel synthesis possible and, as described
above, it has been successfully applied for creating mesoporous
zeolites. Often SAC was employed to prevent a phase separa-
tion between the hydrophilic zeolite gels and hydrophobic
mesoporogens. Since highly concentrated gels stimulate the
massive formation of nuclei that subsequently grow into
numerous nanozeolites, the mesoporous materials usually
were assembled from agglomerated nanoparticles. We have
shown that excellent mesoporosity was obtained for zeolite
beta without any addition of a secondary template when the
SAC method was used.194b However, since this process is not
easily scalable, we aimed to simplify it even further. To this
end, we prepared highly concentrated gels by using similar
amounts of water as had been used for the SAC reaction, but
we converted these gels by a common hydrothermal reaction
in an autoclave.195 Using a SiO2:H2O ratio of only about 1:6
and reaction temperatures between 150 and 170  C, we could
 Mesoporous Zeolitic Materials 281

show that highly mesoporous zeolite beta can be formed
within very short reaction times between 4 and 6 h.
This mono-templated (with TPAOH) synthesis results in
translucent gels of crystalline zeolite beta nanoparticles with
nearly 100% yield. Depending on the subsequent processing
paths of these converted gels it is possible to obtain very
different results: when the gel is diluted with water it disinte-
grates into a stable colloidal solution of nanozeolites, which
can be purified and retrieved via centrifugation. However,
when the gel is dried and directly calcined, the nanozeolites
are fused together into mesoporous aggregates (see scheme in
Figure 24). This nanofusion process results in a material that
can be purified via simple filtration and that is stable against
prolonged ultrasonic treatment. Here, nanozeolite beta parti-
cles are likely partially condensed, thus creating interstitial
mesopores. The nanoparticle size can be easily changed by
prolonging the reaction time, resulting in tunable mesopores
between about 15 and 35 nm and creating external surface
areas between about 240 and 150 m2 g1 (mesopore volume
0.40.8 ml g1, see Figure 24). Acidic zeolite beta can be made
similarly by using a sodium-free synthesis gel, thus eliminating
the secondary processing step of ion exchange. Slightly larger
mesopores between 35 and 140 nm are then obtained due to
larger primary particle formation. Hence, by avoiding the SAC
procedure it is now possible to exploit the whole volume of the
reactor since an extra inset for separating the dry gel from the
liquid water is not needed. Extraordinary spacetime yields of
up to 600 g Ld1 can be produced this way.
The idea of using nanozeolites as building blocks for meso-
porous materials has stimulated a recent study by Yin et al.196
Highly concentrated gels led to the self-assembly of 10-nm
nanozeolite L particles into 400-nm aggregates featuring

Nanozeolite
suspension
Hydrothermal
conversion
Concentrated Nanozeolite
gel beta

Nano-fusion

Nitrogen sorption measurement Mercury intrusion measurement

500 3.0
Log differential instruction (ml g1)

6h 35 nm
450 6h
23 h 2.5 23 h
400 24 nm
72 h 72 h
350
Volume (ml g1)

2.0
300
250 1.5
200 15 nm

150 1.0
100
0.5
50
0 0.0
0.0 0.2 0.4 0.6 0.8 1.0
0.01 0.1 1
Relative pressure (P/P0) Pore diameter (mm)

6h 23 h 72 h

20 nm 20 nm 20 nm

Figure 24 Mesoporous zeolite beta made by the nanofusion process.195 (Top) Schematic representation of the process yielding either nanozeolite
suspensions or mesoporous aggregates. (Middle) Nitrogen sorption and corresponding mesopore size distributions after different reaction times at
170  C. (Bottom) TEM micrographs of the corresponding samples.
282 Mesoporous Zeolitic Materials
mesopores of about 4 nm. In a model reaction for utilizing
these additional pores, the samples were used as carriers for
magnetic iron oxide particles for separation and for functiona-
lization with a fluorescent rhodaminehydrazine receptor for
simultaneous selective mercury capture and detection.
Since nanozeolite synthesis schemes have been developed
for a number of zeolites, including zeolite beta, ZSM-5, LTA,
FAU, and LTL,203 we believe that the above dense-gel approach
might be very promising for creating mesoporous zeolites with a
significant number of additional structures. Nanozeolites are
usually synthesized from clear solutions in order to facilitate
the formation of stable colloidal suspensions which can be used
for further processing. By contrast, dense gels provide a means to
stimulate massive nucleation and in situ aggregation of the
resulting nanoparticles, thus instantaneously creating mesopor-
ous zeolites at high yields. As described above, the stabilization
of these aggregates can be achieved in a convenient fusion step.
7.11.5 Conclusion
As discussed throughout this chapter, numerous examples
have now established the great potential of mesoporous zeo-
lites that feature both catalytically active micropores and addi-
tional diffusional pathways of larger dimensions. Already
having a high impact on industrial catalytic processing, with
additional mesoporosity zeolites can be envisioned for a
greatly extended range of applications. New fields concerning
the conversion of bulky molecules can be explored, and even
traditional applications can benefit from the morphological
changes brought about with mesopores. For instance, a recent
spectroscopic investigation confirmed that the important
Brnsted acidity remains unaffected after introducing meso-
porosity into ZSM-5 since the Brnsted centers are primarily
located within the microporous body.204 Furthermore, the
effective shortening of the diffusional pathways in hierarchical
zeolites through the presence of smaller zeolitic domains was
shown to have a tremendous effect also on classical catalytic
applications of zeolites. For instance, in the MTO conversion
on mesoporous SAPO-5 it was shown that not only the overall
reaction rates improved but also the selectivity was enhanced
because secondary reactions were suppressed.197 Enhancing
mass transport to and away from the active sites may also be
the reason for the diminished coke formation and prolonged
catalyst lifetime observed in many test reactions. However,
most importantly, incorporating mesopores into zeolitic mate-
rials inherently increases their external or accessible surface
area for larger molecules, thus opening completely new fields
of application for zeolites, such as the processing of heavier
feedstock and the synthesis or conversion of complex organic
molecules. The interested reader is referred to the extensive
recent review by Holm et al. summarizing the implications of
mesopores on diverse catalytic reactions such as alkylation,
isomerization, cracking, or condensation reactions.10 Consid-
ering that even small changes in catalytic efficiency in tradi-
tional applications may have a substantial impact on the
economic outcome in industrial processing, there is an enor-
mous potential yet to be discovered for reactions that up to
now were excluded from zeolite-catalyzed conversions.
In addition to a careful evaluation of the impact of meso-
porosity on catalytic behavior, it is similarly important to con-
sider the significant changes brought upon the zeolite material
by changing its morphology. Moreover, the tuning of meso-
pores should be possible to adjust the zeolitic material for the
specific catalytic demands. It has become apparent that top-
down leaching methods such as desilication and dealumination
are relatively straightforward, but they inherently lead to partial
destruction and changes in chemical composition, both of
which have to be considered when adapting these methods for
industrial applications. On the other hand, in developing
bottom-up strategies the application of more complex synthesis
steps or high-cost templating agents might outweigh the benefits
gained in catalytic applications. The assembly of nanozeolites
appears to be a promising avenue toward creating mesoporous
zeolites with a minimum of reagents and processing steps. To
bring mesoporous zeolites into industrial use, it is clearly impor-
tant to find synthetic approaches that are as simple and eco-
nomical as possible and are easily adaptable to industrial
processing. For related chapters in this Comprehensive, we
refer to Chapters 4.01 and 5.06.
Reference
1. Davis, M. E.; Saldarriaga, C.; Montes, C.; Garces, J.; Crowdert, C. Nature 1988,
331, 698699.
2. Jiang, J.; Jorda, J. L.; Yu, J.; Baumes, L. A.; Mugnaioli, E.; Diaz-Cabanas, M. J.;
Kolb, U.; Corma, A. Science 2011, 333, 11311134.
3. http://www.iza-structure.org/databases/.
4. Jiang, J.; Yu, J.; Corma, A. Angew. Chem. Int. Ed. 2010, 49, 31203145.
5. Zhao, D.; Wan, Y. In Introduction to Zeolite Science and Practice, Studies in
Surface Science and Catalysis; 3rd revised ed.; Cejka, J., van Bekkum, H.,
Corma, A., Schuth, F., Eds.; Elsevier Science: Amsterdam, 2007; 168; pp 241300.
6. Liu, Y.; Pinnavaia, T. J. J. Mater. Chem. 2002, 12, 31793190.
7. Mintova, S.; Cejka, J. In Introduction to Zeolite Science and Practice, Studies in
Surface Science and Catalysis; 3rd revised ed.; Cejka, J., van Bekkum, H.,
Corma, A., Schuth, F., Eds.; Elsevier Science: Amsterdam, 2007; Vol. 168,
pp 301326.
8. Roth, W. J.; Cejka, J. Catal. Sci. Technol. 2011, 1, 4353.
9. Martens, J. A.; Thybaut, J. W.; Denayer, J. F. M.; Sree, S. P.; Aerts, A.;
Reyniers, M.-F.; Van Speybroeck, V.; Waroquier, M.; Buekenhoudt, A.;
Vankelecom, I.; Buijs, W.; Persoons, J.; Baron, G. V.; Bals, S.; Van Tendeloo, G.;
Marin, G. B.; Jacobs, P. A.; Kirschhock, C. E. A. Catal. Today 2011, 168, 1727.
10. Holm, M. S.; Taarning, E.; Egeblad, K.; Christensen, C. H. Catal. Today 2011,
168, 316.
11. Ogura, M.; Shinomiya, S.-Y.; Tateno, J.; Nara, Y.; Kikuchi, E.; Matsukata, M.
Chem. Lett. 2000, 882883.
12. Groen, J. C.; Peffer, L. A. A.; Moulijn, J. A.; Perez-Ramirez, J. Microporous
Mesoporous Mater. 2004, 69, 2934.
13. Groen, J. C.; Peffer, L. A. A.; Moulijn, J. A.; Perez-Ramirez, J. Chem. Eur. J. 2005,
11, 49834994.
14. (a) Groen, J. C.; Bach, T.; Ziese, U.; Paulaime-van Donk, A. M.; de Jong, K. P.;
Moulijn, J. A.; Perez-Ramirez, J. J. Am. Chem. Soc. 2005, 127, 1079210793;
(b) Groen, J. C.; Moulijn, J. A.; Perez-Ramirez, J. J. Mater. Chem. 2006, 16,
21212131.
15. Zhao, L.; Xu, C.; Gao, S.; Shen, B. J. Mater. Sci. 2010, 45, 54065411.
16. Mokrzycki, .; Sulikowski, B.; Olejniczak, Z. Catal. Lett. 2009, 127, 296.
17. Perez-Ramirez, J.; Verboekend, D.; Bonilla, A.; Abello, S. Adv. Funct. Mater.
2009, 19, 39723979.
18. Abello, S.; Bonilla, A.; Perez-Ramrez, J. Appl. Catal. A Gen. 2009, 364, 191198.
19. Abello, S.; Perez-Ramirez, J. Phys. Chem. Chem. Phys. 2009, 11, 29592963.
20. Fernandez, C.; Stan, I.; Gilson, J.-P.; Thomas, K.; Vicente, A.; Bonilla, A.;
Perez-Ramirez, J. Chem. Eur. J. 2010, 16, 62246233.
21. Caicedo-Realpe, R.; Perez-Ramirez, J. Microporous Mesoporous Mater. 2010,
128, 91100.
22. (a) Mitchell, S.; Bonilla, A.; Perez-Ramirez, J. Mater. Chem. Phys. 2011, 127,
278284; (b) Mitchell, S.; Perez-Ramrez, J. Catal. Today 2011, 168, 2837.

23. Verboekend, D.; Mitchell, S.; Milina, M.; Groen, J. C.; Perez-Ramirez, J. J. Phys.
Chem. C 2011, 115, 1419314203.
24. (a) Cheng, X.-w.; Meng, Q.-y.; Chen, J.-y.; Long, Y.-c. Microporous Mesoporous
Mater. 2012, 153, 198203; (b) Kang, J.; Cheng, K.; Zhang, L.; Zhang, Q.;
Ding, J.; Hua, W.; Lou, Y.; Zhai, Q.; Wang, Y. Angew. Chem. Int. Ed. 2011, 50,
52005203; (c) Cheng, K.; Kang, J.; Huang, S.; You, Z.; Zhang, Q.; Ding, J.;
Hua, W.; Lou, Y.; Deng, W.; Wang, Y. ACS Catal. 2012, 2, 441449;
(d) Akhtar, M. N.; Al-Yassir, N.; Al-Khattaf, S.; Cejka, J. Catal. Today 2012,
179, 6172.
25. Verboekend, D.; Perez-Ramrez, J. Chem. Eur. J. 2011, 17, 11371147.
26. van laak, A. N. C.; Zhang, L.; Parvulescu, A. N.; Bruijnincx, P. C. A.;
Weckhuysen, B. M.; de Jong, K. P.; de Jongh, P. E. Catal. Today 2011, 168,
4856.
27. Groen, J. C.; Abello, S.; Villaescusa, L. A.; Perez-Ramirez, J. Microporous
Mesoporous Mater. 2008, 114, 93102.
28. Perez-Ramirez, J.; Abello, S.; Bonilla, A.; Groen, J. C. Adv. Funct. Mater. 2009,
19, 164172.
29. Holm, M. S.; Hansen, M. K.; Christensen, C. H. Eur. J. Inorg. Chem. 2009,
11941198.
30. de Jong, K. P.; Zecevic, J.; Friedrich, H.; de Jong, P. E.; Bulut, M.; van Donk, S.;
Kenmogne, R.; Finiels, A.; Hulea, V.; Fajula, F. Angew. Chem. Int. Ed. 2010, 49,
1007410078.
31. Verboekend, D.; Vile, G.; Perez-Ramirez, J. Adv. Funct. Mater. 2012, 22,
916928.
32. Verboekend, D.; Vile, G.; Perez-Ramirez, J. Cryst. Growth Des. 2012, 12,
31233132.
33. Bonilla, A.; Baudouin, D.; Perez-Ramrez, J. J. Catal. 2009, 265, 170180.
34. Verboekend, D.; Caicedo-Realpe, R.; Bonilla, A.; Santiago, M.; Perez-Ramirez, J.
Chem. Mater. 2010, 22, 46794689.
35. Sommer, L.; Mores, D.; Svelle, S.; Stocker, M.; Weckhuysen, B. M.; Olsbye, U.
Microporous Mesoporous Mater. 2010, 132, 384394.
36. Musilova-Pavlackova, Z.; Zones, S. I.; Cejka, J. Top. Catal. 2010, 53, 273282.
37. Verboekend, D.; Groen, J. C.; Perez-Ramirez, J. Adv. Funct. Mater. 2010, 20,
14411450.
38. (a) Verboekend, D.; Chabaneix, A. M.; Thomas, K.; Gilson, J.-P.; Perez-
Ramirez, J. CrystEngComm 2011, 13, 34083416; (b) Verboekend, D.;
Thomas, K.; Milina, M.; Mitchell, S.; Perez-Ramirez, J.; Gilson, J.-P. Catal. Sci.
Technol. 2011, 1, 13311335.
39. Sa Couto, C.; Matias, P.; Santos, E. T.; Fernandes, A.; Graca, I.; Lopes, J. M.;
Ribeiro, M. F. Eur. J. Inorg. Chem. 2012, 26, 41904199.
40. Wei, X.; Smirniotis, P. G. Microporous Mesoporous Mater. 2006, 97, 97106.
41. Groen, J. C.; Sano, T.; Moulijn, J. A.; Perez-Ramrez, J. J. Catal. 2007, 251,
2127.
42. Perez-Ramrez, J.; Abello, S.; Villaescusa, L. A.; Bonilla, A. Angew. Chem. Int. Ed.
2008, 47, 79137917.
43. Kubu, M.; Zilkova, N.; Cejka, J. Catal. Today 2011, 168, 6370.
44. Wang, S.; Dou, T.; Li, Y.; Zhang, Y.; Li, X.; Yan, Z. Catal. Commun. 2005,
6, 8791.
45. Wang, H.; Liu, Y.; Pinnavaia, T. J. J. Phys. Chem. B 2006, 110, 45244526.
46. Al-Yassir, N.; Akhtar, M. N.; Ogunronbi, K.; Al-Khattaf, S. J. Mol. Catal. A Chem.
2012, 360, 115.
47. Qi, J.; Zhao, T.; Xu, X.; Li, F.; Sun, G. J. Porous Mater. 2011, 18,
6981.
48. Chal, R.; Cacciaguerra, T.; van Donk, S.; Garardin, C. Chem. Commun. 2010, 46,
78407842.
49. (a) Ivanova, I. I.; Kuznetsov, A. S.; Yuschenko, V. V.; Knyazeva, E. E. Pure Appl.
Chem. 2004, 76, 16471658; (b) Ordomskii, V. V.; Monakhova, Y. V.;
Knyazeva, E. E.; Nesterenko, N. S.; Ivanova, I. I. Russ. J. Phys. Chem. A 2009, 83,
10121017; (c) Ivanova, I. I.; Kuznetsov, A. S.; Knyazeva, E. E.; Fajula, F.;
Thibault-Starzyk, F.; Fernandez, C.; Gilson, J.-P. Catal. Today 2011, 168,
133139; (d) Khitev, Y. P.; Kolyagin, Y. G.; Ivanova, I. I.; Ponomareva, O. A.;
Thibault-Starzyk, F.; Gilson, J.-P.; Fernandez, C.; Fajula, F. Microporous
Mesoporous Mater. 2011, 146, 201207.
50. Garcia-Martinez, J.; Johnson, M.; Valla, J.; Li, K.; Ying, J. Y. Catal. Sci. Technol.
2012, 2, 987994.
51. Yoo, W. C.; Zhang, X.; Tsapatsis, M.; Stein, A. Microporous Mesoporous Mater.
2012, 149, 147157.
52. van Donk, S.; Janssen, A. H.; Bitter, J. H.; de Jong, K. P. Catal. Rev. Sci. Eng.
2003, 45, 297319.
53. Lutz, W.; Enke, D.; Einicke, W.-D.; Taschner, D.; Kurzhals, R. Z. Anorg. Allg.
Chem. 2010, 636, 25322534.
54. Chal, R.; Gerardin, C.; Bulut, M.; van Donk, S. ChemCatChem 2011, 3, 6781.
55. Chung, K.-H. Microporous Mesoporous Mater. 2008, 111, 544550.
Mesoporous Zeolitic Materials 283
56. Gonzalez, M. D.; Cesteros, Y.; Salagre, P. Microporous Mesoporous Mater.
2011, 144, 162170.
57. Qin, Z.; Shen, B.; Gao, X.; Lin, F.; Wang, B.; Xu, C. J. Catal. 2011, 278, 266275.
58. Valtchev, V.; Balanzat, E.; Mavrodinova, V.; Diaz, I.; El Fallah, J.; Goupil, J.-M.
J. Am. Chem. Soc. 2011, 133, 1895018956.
59. Perez-Ramirez, J.; Mitchell, S.; Verboekend, D.; Milina, M.; Michels, N.-L.;
Krumeich, F.; Marti, N.; Erdmann, M. ChemCatChem 2011, 3, 17311734.
60. Verboekend, D.; Perez-Ramirez, J. Catal. Sci. Technol. 2011, 1, 879890.
61. Ogura, M.; Shinomiya, S.-y.; Tateno, J.; Nara, Y.; Nomura, M.; Kikuchi, E.;
Matsukata, M. Appl. Catal. A 2001, 219, 3343.
62. Michels, N.-L.; Mitchell, S.; Milina, M.; Kunze, K.; Krumeich, F.; Marone, F.;
Erdmann, M.; Marti, N.; Perez-Ramirez, J. Adv. Funct. Mater. 2012, 22,
25092518.
63. Svelle, S.; Sommer, L.; Barbera, K.; Vennestrom, P. N. R.; Olsbye, U.;
Lillerud, K. P.; Bordiga, S.; Pan, Y.-H.; Beato, P. Catal. Today 2011, 168, 3847.
64. Hojholt, K. T.; Vennestrom, P. N. R.; Tiruvalam, R.; Beato, P. Chem. Commun.
2011, 47, 1286412866.
65. Zecevic, J.; Gommes, C. J.; Friedrich, H.; de Jongh, P. E.; de Jong, K. P. Angew.
Chem. Int. Ed. 2012, 51, 42134217.
66. Vennestrom, P. N. R.; Grill, M.; Kustova, M.; Egeblad, K.; Lundegaard, L. F.;
Joensen, F.; Christensen, C. H.; Beato, P. Catal. Today 2011, 168, 7179.
67. (a) Qi, J.; Zhao, T.; Xu, X.; Li, F.; Sun, G.; Miao, C.; Wang, H. Catal. Commun.
2009, 10, 15231528; (b) Qi, J.; Zhao, T. B.; Li, F. Y.; Sun, G. D.; Xu, X.;
Miao, C.; Wang, H. W.; Zhang, X. M. J. Porous Mater. 2010, 17, 177184.
68. (a) Egeblad, K.; Christensen, C. H.; Kustova, M.; Christensen, C. H. Chem. Mater.
2008, 20, 946960; (b) Perez-Ramirez, J.; Christensen, C. H.; Egeblad, K.;
Christensen, C. H.; Groen, J. C. Chem. Soc. Rev. 2008, 37, 25302542.
69. (a) Jacobsen, C. J. H.; Madsen, C.; Janssens, T. V. W.; Jakobsen, H. J.;
Skibsted, J. Microporous Mesoporous Mater. 2000, 39, 393401; (b) Schmidt, I.;
Madsen, C.; Jacobsen, C. J. H. Inorg. Chem. 2000, 39, 22792283.
70. Jacobsen, C. J. H.; Madsen, C.; Houzvicka, J.; Schmidt, I.; Carlsson, A. J. Am.
Chem. Soc. 2000, 122, 71167117.
71. Janssen, A. H.; Schmidt, I.; Jacobsen, C. J. H.; Koster, A. J.; de Jong, K. P.
Microporous Mesoporous Mater. 2003, 65, 5975.
72. Schmidt, I.; Krogh, A.; Wienberg, K.; Carlsson, A.; Brorson, M.; Jacobsen, C. J. H.
Chem. Commun. 2000, 21572158.
73. Wei, X.; Smirniotis, P. G. Microporous Mesoporous Mater. 2006, 89, 170178.
74. Liang, J.; Wang, F. Adv. Mater. Res. 2011, 306307, 15761579.
75. Moushey, D. L.; Smirniotis, P. G. Catal. Lett. 2009, 129, 2025.
76. Xin, H.; Zhao, J.; Xu, S.; Li, J.; Zhang, W.; Guo, X.; Hensen, E. J. M.; Yang, Q.;
Li, C. J. Phys. Chem. C 2010, 114, 65536559.
77. Egeblad, K.; Kustova, M.; Klitgaard, S. K.; Zhu, K.; Christensen, C. H.
Microporous Mesoporous Mater. 2007, 101, 214223.
78. Holm, M. S.; Egeblad, K.; Vennestrom, P. N. R.; Hartmann, C. G.; Kustova, M.;
Christensen, C. H. Eur. J. Inorg. Chem. 2008, 51855189.
79. (a) Ok, D.-Y.; Jiang, N.; Prasetyanto, E. A.; Jin, H.; Park, S.-E. Microporous
Mesoporous Mater. 2011, 141, 27; (b) Koo, J.-B.; Jiang, N.;
Saravanamurugan, S.; Bejblova, M.; Musilova, Z.; Cejka, J.; Park, S.-E. J. Catal.
2010, 276, 327334.
80. Kustova, M.; Egeblad, K.; Zhu, K.; Christensen, C. H. Chem. Mater. 2007, 19,
29152917.
81. (a) Wang, X.; Li, G.; Wang, W.; Jin, C.; Chen, Y. Microporous Mesoporous Mater.
2011, 142, 494502; (b) Wang, W.; Li, G.; Li, W.; Liu, L. Chem. Commun. 2011,
47, 35293531.
82. Christensen, C. H.; Schmidt, I.; Carlsson, A.; Johannsen, K.; Herbst, K. J. Am.
Chem. Soc. 2005, 127, 80988102.
83. (a) Martinez, A.; Peris, E.; Derewinski, M.; Burkat-Dulak, A. Catal. Today 2011,
169, 7584; (b) Martinez, A.; Arribas, M. A.; Derewinski, M.; Burkatdulak, A. Appl.
Catal. A: Gen. 2010, 379, 188197.
84. Jiang, N.; Han, D.-S.; Park, S.-E. Catal. Today 2009, 141, 344348.
85. Liu, H.; Yang, S.; Hu, J.; Shang, F.; Li, Z.; Xu, C.; Guan, J.; Kan, Q. Fuel Process.
Technol. 2012, 96, 195202.
86. (a) Tao, Y.; Kanoh, H.; Kaneko, K. J. Am. Chem. Soc. 2003, 125, 60446045;
(b) Tao, Y.; Kanoh, H.; Kaneko, K. J. Phys. Chem. B 2003, 107,
1097410976; (c) Tao, Y.; Kanoh, H.; Abrams, L.; Kaneko, K. Chem. Rev. 2006,
106, 896910.
87. (a) Tao, Y.; Hattori, Y.; Matumoto, A.; Kanoh, H.; Kaneko, K. J. Phys. Chem. B
2005, 109, 194199; (b) Tao, Y.; Kanoh, H.; Kaneko, K. Langmuir 2005, 21,
504507.
88. Fang, Y.; Hu, H.; Chen, G. Microporous Mesoporous Mater. 2008, 113,
481489.
89. Schmidt, I.; Boisen, A.; Gustavsson, E.; Stahl, K.; Pehrson, S.; Dahl, S.;
Carlsson, A.; Jacobsen, C. J. H. Chem. Mater. 2001, 13, 44164418.
284 Mesoporous Zeolitic Materials
90. Li, D.; Qiu, L.; Wang, K.; Zeng, Y.; Li, D.; Williams, T.; Huang, Y.; Tsapatsis, M.;
Wang, H. Chem. Commun. 2012, 48, 22492251.
91. Gu, L. J.; Ma, D.; Yao, S.; Wang, C.; Shen, W.; Bao, X. Chem. Commun. 2010,
46, 17331735.
92. Wattanakit, C.; Warakulwit, C.; Pantu, P.; Sunpetch, B.; Charoenpanich, M.;
Limtrakul, J. Can. J. Chem. Eng. 2012, 90, 873880.
93. Kim, S.-S.; Shah, J.; Pinnavaia, T. J. Chem. Mater. 2003, 15, 16641668.
94. (a) Fan, W.; Snyder, M. A.; Kumar, S.; Lee, P.-S.; Yoo, W. C.; McCormick, A. V.;
Lee Penn, R.; Stein, A.; Tsapatsis, M. Nat. Mater. 2008, 7, 984991;
(b) Tsapatsis, M. Prepr. Am. Chem. Soc. Div. Pet. Chem. 2011, 56, 59.
95. (a) Chen, H.; Wydra, J.; Zhang, X.; Lee, P.-S.; Wang, Z.; Fan, W.; Tsapatsis, M.
J. Am. Chem. Soc. 2011, 133, 1239012393; (b) Yoo, W. C.; Kumar, S.;
Penn, R. L.; Tsapatsis, M.; Stein, A. J. Am. Chem. Soc. 2009, 131,
1237712383.
96. (a) Fang, Y.; Hu, H. J. Am. Chem. Soc. 2006, 128, 1063610637;
(b) Sakthivel, A.; Huang, S.-J.; Chen, W.-H.; Lan, Z.-H.; Chen, K.-H.; Kim, T.-W.;
Ryoo, R.; Chiang, A. S. T.; Liu, S.-B. Chem. Mater. 2004, 16, 31683175.
97. Cho, H. S.; Ryoo, R. Microporous Mesoporous Mater. 2012, 151, 107112.
98. Xue, C.; Zhang, F.; Wu, L.; Zhao, D. Microporous Mesoporous Mater. 2012, 151,
495500.
99. Holland, B. T.; Abrams, L.; Stein, A. J. Am. Chem. Soc. 1999, 121,
43084309.
100. (a) Tosheva, L.; Valtchev, V.; Sterte, J. Microporous Mesoporous Mater. 2000,
3536, 621629; (b) Tosheva, L.; Mihailova, B.; Valtchev, V.; Sterte, J.
Microporous Mesoporous Mater. 2001, 48, 3137.
101. Zhao, J.; Zhou, J.; Chen, Y.; He, Q.; Ruan, M.; Guo, L.; Shi, J.; Chen, H. J. Mater.
Chem. 2009, 19, 76147616.
102. Carrillo, A. I.; Linares, N.; Martinez, J. G. In NSTI Nanotech, Nanotechnology
Conference and Trade Show, Technical Proceedings, Boston, MA, USA, Jun 15,
2008; Vol. 1, pp 638641.
103. Zhu, H.; Liu, Z.; Wang, Y.; Kong, D.; Yuan, X.; Xie, Z. Chem. Mater. 2008, 20,
11341139.
104. Hernandez-Ramirez, O.; Al-Nasri, S. K.; Holmes, S. M. J. Mater. Chem. 2011, 21,
1652916534.
105. Valtchev, V.; Mintova, S. In Ordered Porous Solids; Valtchev, V., Mintova, S.,
Tsapatsis, M., Eds.; Elsevier: Oxford, 2009; pp 477499.
106. Xiao, F.-S.; Wang, L.; Yin, C.; Lin, K.; Di, Y.; Li, J.; Xu, R.; Su, D. S.; Schlogl, R.;
Yokoi, T.; Tatsumi, T. Angew. Chem. Int. Ed. 2006, 45, 30903093.
107. Tang, T.; Yin, C.; Wang, L.; Ji, Y.; Xiao, F.-S. J. Catal. 2007, 249, 111115.
108. Wang, L.; Zhang, Z.; Yin, C.; Shan, Z.; Xiao, F.-S. Microporous Mesoporous
Mater. 2010, 131, 5867.
109. Song, J.; Ren, L.; Yin, C.; Ji, Y.; Wu, Z.; Li, J.; Xiao, F.-S. J. Phys. Chem. C 2008,
112, 86098613.
110. Liu, S.; Cao, X.; Li, L.; Li, C.; Ji, Y.; Xiao, F.-S. Colloids Surf. A Physicochem.
Eng. Asp. 2008, 318, 269274.
111. Liu, F.; Willhammar, T.; Wang, L.; Zhu, L.; Sun, Q.; Meng, X.; Carrillo-
Cabrera, W.; Zou, X.; Xiao, F.-S. J. Am. Chem. Soc. 2012, 134, 45574560.
112. Moller, K.; Yilmaz, B.; Muller, U.; Bein, T. Chem. Mater. 2011, 23, 43014310.
113. (a) Ling, X.; Zong, R. L.; Qiu, B. K. Adv. Mater. Res. 2011, 217218, 368371;
(b) Lei, Q.; Zhao, T.; Li, F.; Wang, Y. F.; Hou, L. J. Porous Mater. 2008, 15,
643646; (c) Lei, Q.; Zhao, T.; Li, F.; Zhang, L.; Wang, Y. Chem. Commun. 2006,
17691771.
114. Kang, Y.; Shan, W.; Wu, J.; Zhang, Y.; Wang, X.; Yang, W.; Tang, Y. Chem. Mater.
2006, 18, 18611866.
115. Shi, Y.; Li, X.; Hu, J.; Lu, J.; Ma, Y.; Zhang, Y.; Tang, Y. J. Mater. Chem. 2011,
21, 1622316230.
116. Ocampo, F.; Cunha, J. A.; de Lima Santos, M. R.; Tessonnier, J. P.;
Pereira, M. M.; Louis, B. Appl. Catal., A 2010, 390, 102109.
117. Zhu, Y.; Hua, Z.; Zhou, J.; Wang, L.; Zhao, J.; Gong, Y.; Wu, W.; Ruan, M.; Shi, J.
Chem. Eur. J. 2011, 17, 1461814627.
118. Sakthivel, A.; Iida, A.; Komura, K.; Sugi, Y.; Chary, K. V. R. Microporous
Mesoporous Mater. 2009, 119, 322330.
119. Hua, J.; Han, Y. Chem. Mater. 2009, 21, 23442348.
120. Wang, Z.; Xu, L.; Jiang, J.-g.; Liu, Y.; He, M.; Wu, P. Microporous Mesoporous
Mater. 2012, 156, 106114.
121. (a) Song, Y.; Hua, Z.; Zhu, Y.; Zhou, J.; Zhou, X.; Liu, Z.; Shi, J. J. Mater. Chem.
2012, 22, 33273329; (b) Zhou, J.; Hua, Z.; Liu, Z.; Wu, W.; Zhu, Y.; Shi, J. ACS
Catal. 2011, 1, 287291.
122. Zhou, J.; Hua, Z.; Wu, W.; Liu, Z.; Zhu, Y.; Chen, Y.; Shi, J. Dalton Trans. 2011,
40, 1266712669.
123. Neumann, G. T.; Hicks, J. C. Top. Catal. 2012, 55, 196208.
124. Yang, H.; Liu, Z.; Gao, H.; Xie, Z.; Mater, J. J. Mater. Chem. 2010, 20,
32273231.
125. Zhu, H.; Liu, Z.; Kong, D.; Wang, Y.; Xie, Z. J. Phys. Chem. C 2008, 112,
1725717264.
126. Liu, Y.; Zhang, W.; Liu, Z.; Xu, S.; Wang, Y.; Xie, Z.; Han, X.; Bao, X. J. Phys.
Chem. C 2008, 112, 1537515381.
127. Wang, L.; Yin, C.; Shan, Z.; Liu, S.; Du, Y.; Xiao, F.-S. Colloids Surf.
A Physicochem. Eng. Asp. 2009, 340, 126130.
128. Tao, H.; Li, C.; Ren, J.; Wang, Y.; Lu, G. J. Solid State Chem. 2011, 184,
18201827.
129. Yang, X.; Lu, T.; Chen, C.; Zhou, L.; Wang, F.; Su, Y.; Xu, J. Microporous
Mesoporous Mater. 2011, 144, 176182.
130. Xu, B.; Zhou, L.; Wang, F.; Qin, H.; Zhu, J.; Zou, H. Chem. Commun. 2012, 48,
18021804.
131. Li, C. L.; Wang, Y. Q.; Shi, B. F.; Ren, J. W.; Liu, X. H.; Wang, Y. G.; Guo, Y.;
Guo, Y. L.; Lu, G. Z. Microporous Mesoporous Mater. 2009, 117, 104110.
132. (a) Park, D. H.; Kim, S. S.; Wang, H.; Pinnavaia, T. J.; Papapetrou, M. C.;
Lappas, A. A.; Triantafyllidis, K. S. Angew. Chem. Int. Ed. 2009, 48,
76457648; (b) Wang, H.; Pinnavaia, T. J. Angew. Chem. Int. Ed. 2006, 45,
76037606.
133. Mukti, R. R.; Hirahara, H.; Sugawara, A.; Shimojima, A.; Okubo, T. Langmuir
2010, 26, 27312735.
134. Choi, M.; Cho, H. S.; Srivastava, R.; Venkatesan, C.; Choi, D.-H.; Ryoo, R. Nat.
Mater. 2006, 5, 718723.
135. Koekkoek, A. J. J.; Tempelman, C. H. L.; Degirmenci, V.; Guo, M.; Feng, Z.; Li, C.;
Hensen, E. J. M. Catal. Today 2011, 168, 96111.
136. Cheneviere, Y.; Chieux, F.; Caps, V.; Tuel, A. J. Catal. 2010, 269, 161168.
137. Lee, D. H.; Choi, M.; Yu, B. W.; Ryoo, R. Chem. Commun. 2009, 7476.
138. (a) Cho, K.; Cho, H. S.; de Menorval, L.-C.; Ryoo, R. Chem. Mater. 2009, 21,
56645673; (b) Mehlhorn, D.; Valiullin, R.; Kaerger, J.; Cho, K.; Ryoo, R.
ChemPhysChem 2012, 13, 14951499.
139. Valiullin, R.; Karger, J.; Cho, K.; Choi, M.; Ryoo, R. Microporous Mesoporous
Mater. 2011, 142, 236244.
140. Fu, W.; Zhang, L.; Tang, T.; Ke, Q.; Wang, S.; Hu, J.; Fang, G.; Li, J.; Xiao, F.-S.
J. Am. Chem. Soc. 2011, 133, 1534615349.
141. Inayat, A.; Knoke, I.; Spiecker, E.; Schwieger, W. Angew. Chem. Int. Ed. 2012, 51,
19621965.
142. (a) Choi, M.; Srivastava, R.; Ryoo, R. Chem. Commun. 2006, 43804382;
(b) Kim, J.; Bhattacharjee, S.; Jeong, K.-E.; Jeong, S.-Y.; Choi, M.; Ryoo, R.;
Ahn, W.-S. New J. Chem. 2010, 34, 29712978.
143. (a) Danilina, N.; Castelanelli, S. A.; Troussard, E.; van Bokhoven, J. A. Catal.
Today 2011, 168, 8085; (b) Danilina, N.; Krumeich, F.; van Bokhoven, J. A.
J. Catal. 2010, 272, 3743.
144. Shanbhag, G. V.; Choi, M.; Kim, J.; Ryoo, R. J. Catal. 2009, 264, 8892.
145. Serrano, D. P.; Aguado, J.; Escola, J. M.; Rodriguez, J. M.; Peral, A. Chem. Mater.
2006, 18, 24622464.
146. Serrano, D. P.; Aguado, J.; Morales, G.; Rodriguez, J. M.; Peral, A.; Thommes, M.;
Epping, J. D.; Chmelka, B. F. Chem. Mater. 2009, 21, 641654.
147. Serrano, D. P.; Aguado, J.; Escola, J. M.; Rodriguez, J. M.; Peral, A. J. Catal.
2010, 276, 152160.
148. Serrano, D.; Sanz, R.; Pizarro, P.; Moreno, I. Chem. Commun. 2009, 14071409.
149. Serrano, D. P.; Aguado, J.; Escola, J. M.; Peral, A.; Morales, G.; Abella, E. Catal.
Today 2011, 168, 8695.
150. Xue, Z.; Zhang, T.; Ma, J.; Miao, H.; Fan, W.; Zhang, Y.; Li, R. Microporous
Mesoporous Mater. 2012, 151, 271276.
151. Guo, Y.-P.; Wang, H.-J.; Guo, Y.-J.; Guo, L.-H.; Chu, L.-F.; Guo, C.-X. Chem.
Eng. J. 2011, 166, 391400.
152. Srivastava, R.; Iwasa, N.; Fujita, S.-I.; Arai, M. Chem. Eur. J. 2008, 14,
95079511.
153. Xue, Z.; Ma, J.; Hao, W.; Bai, X.; Kang, Y.; Liu, J.; Li, R. J. Mater. Chem. 2012,
22, 25322538.
154. Fan, Y.; Xiao, H.; Shi, G.; Liu, H.; Bao, X. J. Catal. 2012, 285, 251259.
155. Srivastava, R.; Choi, M.; Ryoo, R. Chem. Commun. 2006, 44894491.
156. (a) Zhang, H.; Zhu, X.; Gao, Y.; Zhu, K. Adv. Mater. Res. 2011, 311313,
540544; (b) Rana, B. S.; Singh, B.; Kumar, R.; Verma, D.; Bhunia, M. K.;
Bhaumik, A.; Sinha, A. K. J. Mater. Chem. 2010, 20, 85758581.
157. Foster, A. J.; Jae, J.; Cheng, Y.-T.; Huber, G. W.; Lobo, R. F. Appl. Catal., A 2012,
423424, 154161.
158. Shetti, V. N.; Kim, J.; Srivastava, R.; Choi, M.; Ryoo, R. J. Catal. 2008, 254,
296303.
159. Xin, H. C.; Koekkoek, A.; Yang, Q. H.; van Santen, R.; Li, C.; Hensen, E. J. M.
Chem. Commun. 2009, 75907592.
160. (a) Koekkoek, A. J. J.; Degirmenci, V.; Hensen, E. J. M. J. Mater. Chem. 2011, 21,
92799289; (b) Koekkoek, A. J. J.; Xin, H.-C.; Yang, Q.-H.; Li, C.;
Hensen, E. J. M. Microporous Mesoporous Mater. 2011, 145, 172181.

161. (a) Aguado, J.; Serrano, D. P.; Rodriguez, J. M. Microporous Mesoporous Mater.
2008, 115, 504513; (b) Aguado, J.; Serrano, D. P.; Escola, J. M.; Peral, A.
J. Anal. Appl. Pyrol. 2009, 85, 352358.
162. Serrano, D. P.; Garcia, R. A.; Linares, M.; Gil, B. Catal. Today 2012, 179, 91101.
163. Carrero, A.; Vicente, G.; Rodriguez, R.; Linares, M.; del Peso, G. L. Catal. Today
2011, 167, 148153.
164. Escola, J. M.; Aguado, J.; Serrano, D. P.; Briones, L.; Diaz de Tuesta, J. L.;
Calvo, R.; Fernandez, E. Energy Fuels 2012, 26, 31873195.
165. (a) Hua, Z. L.; Zhou, J.; Shi, J. L. Chem. Commun. 2011, 47, 1053610547;
(b) Moller, K.; Bein, T. Science 2011, 333, 297298.
166. Choi, M.; Na, K.; Kim, J.; Sakamoto, Y.; Terasaki, O.; Ryoo, R. Nature 2009, 461,
246249.
167. Na, K.; Choi, M.; Park, W.; Sakamoto, Y.; Terasaki, O.; Ryoo, R. J. Am. Chem. Soc.
2010, 132, 41694177.
168. Park, W.; Yu, D.; Na, K.; Jelfs, K. E.; Slater, B.; Sakamoto, Y.; Ryoo, R. Chem.
Mater. 2011, 23, 51315137.
169. Kim, J.; Park, W.; Ryoo, R. ACS Catal. 2011, 1, 337341.
170. Na, K.; Jo, C.; Kim, J.; Cho, K.; Jung, J.; Seo, Y.; Messinger, R. J.; Chmelka, B. F.;
Ryoo, R. Science 2011, 333, 328332.
171. Cho, K.; Na, K.; Kim, J.; Terasaki, O.; Ryoo, R. Chem. Mater. 2012, 24,
27332738.
172. Jung, J.; Jo, C.; Cho, K.; Ryoo, R. J. Mater. Chem. 2012, 22, 46374640.
173. Kim, K.; Ryoo, R.; Jang, H.-D.; Choi, M. J. Catal. 2012, 288, 115123.
174. Wu, L.; Degirmenci, V.; Magusin, P. C. M. M.; Szyja, B. M.; Hensen, E. J. M.
Chem. Commun. 2012, 48, 94929494.
175. Zhang, X.; Liu, D.; Xu, D.; Asahina, S.; Cychosz, K. A.; Agrawal, K. V.;
Al Wahedi, Y.; Bhan, A.; Al Hashimi, S.; Terasaki, O.; Thommes, M.; Tsapatsis, M.
Science 2012, 336, 16841687.
176. Choi, M.; Na, K.; Ryoo, R. Chem. Commun. 2009, 28452847.
177. Na, K.; Choi, M.; Ryoo, R. J. Mater. Chem. 2009, 19, 67136719.
178. Kore, R.; Satpati, B.; Srivastava, R. Chem. Eur. J. 2011, 17, 1436014365.
179. Chen, L.; Zhu, S. Y.; Wang, Y. M.; He, M.-Y. New J. Chem. 2010, 34,
23282334.
180. Chen, L.; Zhu, S. Y.; Wang, H. M.; Wang, Y. M. Solid State Sci. 2011, 13,
20242029.
181. Wang, Q.; Chen, G.; Xu, S. Microporous Mesoporous Mater. 2009, 119,
315321.
182. (a) Manjon-Sanz, A.; Sanchez-Sanchez, M.; Munoz-Gomez, P.; Garca, R.;
Sastre, E. Microporous Mesoporous Mater. 2010, 131, 331341; (b) Manjon-
Sanz, A.; Sanchez-Sanchez, M.; Sastre, E. Catal. Today 2012, 179, 102114.
183. Holland, B. T. Microporous Mesoporous Mater. 2006, 89, 291299.
184. (a) Zhou, J.; Hua, Z.; Shi, J.; He, Q.; Guo, L.; Ruan, M. Chem. Eur. J. 2009, 15,
1294912954; (b) Zhou, J.; Hua, Z.; Cui, X.; Ye, Z.; Cui, F.; Shi, J. Chem.
Commun. 2010, 46, 49944996.
Mesoporous Zeolitic Materials 285
185. Zhou, J.; Hua, Z.; Zhao, J.; Gao, Z.; Zeng, S.; Shi, J. J. Mater. Chem. 2010, 20,
67646771.
186. Sun, J.; Wang, C.; Wang, L.; Liang, L.; Liu, X.; Chen, L.; Xiao, F.-S. Catal. Today
2010, 158, 273278.
187. Wang, D.; Li, X.; Liu, Z.; Zhang, Y.; Xie, Z.; Tang, Y. J. Colloid Interface Sci. 2010,
350, 290294.
188. Xue, T.; Chen, L.; Wang, Y. M.; He, M.-Y. Microporous Mesoporous Mater.
2012, 156, 97105.
189. Petushkov, A.; Yoon, S.; Larsen, S. C. Microporous Mesoporous Mater. 2010,
137, 92100.
190. Petushkov, A.; Merilis, G.; Larsen, S. C. Microporous Mesoporous Mater. 2011,
143, 97103.
191. Lakiss, L.; Rivallan, M.; Goupil, J.-M.; El Fallah, J.; Mintova, S. Catal. Today
2011, 168, 112117.
192. (a) Chen, L.-H.; Li, X.-Y.; Tian, G.; Li, Y.; Rooke, J. C.; Zhu, G.-S.; Qiu, S.-L.;
Yang, X.-Y.; Su, B.-L. Angew. Chem. Int. Ed. 2011, 50, 1115611161;
(b) Yang, X.-Y.; Tian, G.; Chen, L.-H.; Li, Y.; Rooke, J. C.; Wei, Y.-X.; Liu, Z.-M.;
Deng, Z.; Van Tendeloo, G.; Su, B.-L. Chem. Eur. J. 2011, 17, 1498714995.
193. (a) Chen, L.-H.; Xu, S.-T.; Li, X.-Y.; Tian, G.; Li, Y.; Rooke, J. C.; Zhu, G.-S.;
Qiu, S.-L.; Wei, Y.-X.; Yang, X.-Y.; Liu, Z.-M.; Su, B.-L. J. Colloid Interface Sci.
2012, 377, 368374; (b) Chen, L.-H.; Li, X.-Y.; Tian, G.; Li, Y.; Tan, H.-Y.;
Van Tendeloo, G.; Zhu, G.-S.; Qiu, S.-L.; Yang, X.-Y.; Su, B.-L. ChemSusChem
2011, 4, 14521456.
194. (a) Majano, G.; Mintova, S.; Ovsitser, O.; Mihailova, B.; Bein, T. Microporous
Mesoporous Mater. 2005, 80, 227235; (b) Moller, K.; Yilmaz, B.;
Jacubinas, R. M.; Muller, U.; Bein, T. J. Am. Chem. Soc. 2011, 133, 52845295.
195. Moller, K.; Yilmaz, B.; Muller, U.; Bein, T. Chem.-Eur. J. 2012, 18, 76717674.
196. Yin, M.; Li, Z.; Liu, Z.; Yang, X.; Ren, J. ACS Appl. Mater. Interfaces 2012, 4,
431437.
197. Zhu, J.; Cui, Y.; Wang, Y.; Wei, F. Chem. Commun. 2009, 32823284.
198. Zhao, X.; Wang, H.; Kang, C.; Sun, Z.; Li, G.; Wang, X. Microporous Mesoporous
Mater. 2012, 151, 501505.
199. Matsukata, M.; Ogura, M.; Osaki, T.; Rao, P. R. H. P.; Nomura, M.; Kikuchi, E.
Top. Catal. 1999, 9, 7792.
200. Jansen, J. C.; Shan, Z.; Marchese, L.; Zhou, W.; Puil, N.v.d.; Maschmeyer, T.
Chem. Commun. 2001, 713714.
201. (a) Wang, J.; Yue, W.; Zhou, W.; Coppens, M.-O. Microporous Mesoporous
Mater. 2009, 120, 1928; (b) Wang, J.; Groen, J. C.; Yue, W.; Zhou, W.;
Coppens, M.-O. Chem. Commun. 2007, 46534655.
202. Murthy, K. V. V. S. B. S. R.; Kulkarni, S. J.; Khaja Masthan, S. Microporous
Mesoporous Mater. 2001, 43, 201209.
203. Tosheva, L.; Valtchev, V. P. Chem. Mater. 2005, 17, 24942513.
204. Tzoulaki, D.; Jentys, A.; Perez-Ramirez, J.; Egeblad, K.; Lercher, J. A. Catal. Today
2012, 198, 311.
 5.05 Zeolites
C Martnez and A Corma, Universidad Politecnica de Valencia, Valencia, Spain
2013 Elsevier Ltd. All rights reserved.

5.05.1 Introduction 103

5.05.2 Zeolite Structure 103
5.05.3 Zeolite Synthesis Mechanisms 109
5.05.4 Main Factors Affecting the Synthesis Chemistry 110
5.05.4.1 Mineralizing Agents and Their Role as Inorganic Structure Directiong Agents (SDAs) 110
5.05.4.2 Isomorphic Substitution in Zeolites: Structure-Directing Effect of Heteroatoms in Framework Positions 111
5.05.4.3 Organic SDAs 113
5.05.4.4 Solvents 115
5.05.5 New Trends in Zeolite Synthesis 116
5.05.5.1 Phosphorus-Containing SDAs 116
5.05.5.2 Zeolite Synthesis in Ionothermal Media 117
5.05.5.3 Synthesis of Extra-Large Pore Zeolites 117
5.05.5.4 Synthesis of Multipore Zeolites 117
5.05.5.5 Synthesis of Organic-Free Zeolites 118
5.05.5.6 Synthesis of Zeolite Layers 119
5.05.5.7 Application of HT Techniques and Data Mining to the Discovery of New Structures 120
5.05.5.8 Other Approaches 120
5.05.6 Zeolite Properties 122
5.05.6.1 Adsorption and Separation 122
5.05.6.2 Ion Exchange 123
5.05.6.3 Catalytic Applications 124
5.05.6.3.1 Reactions catalyzed by acid and basic sites 125
5.05.6.3.2 Redox reactions 126
5.05.6.3.3 Zeolite-based multifunctional catalysts 126
5.05.7 Conclusion 126
Acknowledgment 127
References 127

5.05.1 Introduction
Zeolite materials are crystalline inorganic microporous solids
formed by TO4 tetrahedra (with T Si, Al, P, Ge, B, among others),
whose structures enclose channels and/or cavities of molecular
dimensions (0.31.5 nm diameter). When heated rapidly, they
release water and appear to boil, and this behavior inspired the
Swedish mineralogist A.F. Cronstedt, who discovered these min-
erals in 1756, to name them zeolites, which derives from the
Greek words Zeo and Lithos, and can be translated as stones
that boil.
These silicate-based minerals can be found in nature, commonly
in volcanic areas, alkaline desert lake sediments and soils, or marine
sediments. There are about 45 natural zeolites, and although some of
them occur in large amounts, their limited structures and composi-
tions, together with the presence of impurities, reduce their commer-
cial application mainly to animal feed, pet litter, or water purification,
although other uses have also been described.16 Therefore, most of
the zeolites employed in industrial processes nowadays, related with
adsorption, ion exchange, and catalysis, are synthetic materials and
highly pure, with well-defined composition and pore dimensions.
According to the International Zeolite Association,7 at present there
are 197 different structures enclosed in their database.
Zeolites are present in our everyday life, being widely employed as
sorbents, ion exchangers in detergents, or as industrial catalysts for
upgrading and/or production of liquid fuels and intermediates for the
petrochemical, chemical, or pharmaceutical industry. Besides their
exceptional physicochemical properties and high functionality, zeo-
lites have an additional advantage: they are environmentally friendly,
safe, and sustainable, and this is a key driving force for their increasing
implementation as green alternatives to, for example, chlorine in
swimming pools, to polyphosphates in detergents, or to mineral
acids, such as hydrochloric, hydrofluoric, or sulfuric acid, in many
catalytic industrial processes.
This section reviews the synthesis and main properties of
zeolites and gives a brief overview on their main applications,
some of which are extensively covered in the other sections of
this chapter.
5.05.2 Zeolite Structure
Zeolite materials are inorganic crystalline microporous mate-
rials, constructed from TO4 tetrahedra, with T being the atom
in tetrahedral position (Si, Al, but also P, Ge, Ti, Ga, B, Zn, etc.).

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00506-4 103

104 Zeolites

The adjacent tetrahedra are connected by shared bridging
oxygens, so the resulting framework has an overall O/T ratio
of 2. Thus, zeolites and zeolite-like materials belong to the
group of the tectosilicates, and may be distinguished from
denser phases according to their framework density (FD),
defined as the number of T-atoms per 1000 A3. The minimum
FD for dense structures, such as quartz or feldspars, ranges
from 20 to 22, whereas zeolites are characterized by FD values
below 21, usually below 19. As we will see later in more
detail, the FD is directly related to the pore volume, but
does not necessarily reflect the size of the pore openings.
However, an early analysis of 70 framework structures carried
out by Brunner and Meier in 19898 already showed that when
the frameworks are grouped together as a function of the
smallest ring contained (Figure 1(a)), the minimum FDSi
(the FD calculated for an idealized SiO2 composition) for
each of the groups decreases with the smallest ring. The same
correlation was found9 when considering the 191 structures
accepted by the International Zeolite Association (IZA) in
2010 (Figure 1(b)).
From a strict structural point of view, the connected zeolite
framework is formed by corner-sharing aluminate [AlO4] and
silicate [SiO4] tetrahedra. Pure silicate frameworks (SiO2) are
neutral, but the substitution of Si4 with Al3 in framework
positions generates a negative charge, which has to be com-
pensated by charge-balancing cations in extra-framework posi-
tions, located within the cages and/or pores of the structure.
These cations may be both, organic or inorganic, and the
original ones present in the as-synthesized sample can be
exchanged by others, more convenient, conferring the zeolite
cation-exchange capacity, as well as other specific properties
that will be presented further on. Neutral guest molecules, such
as water, may also be present inside the micropores, in some
cases hydrating the cations mentioned before.

26 26

24 H184-3
24 H178-32
H178-34
22

22
Framework density/T atoms per 1000 3

Framework density/T atoms per 1000 3

20

20
18 AET
H178-43
CFI MTN
DON
SFN
18 16 NPO
SFH
ETR UTL
IWS
VFI
14
16 OSO FAU
TSC
H191-89 VNI

12 ITQ-33
14 H91-1 H191-97

10

12
8
RWY
H191-104
10
6

3 3+ 4 4+ 5 5+ 6 3 3+ 4 4+ 5 5+ 6
(a) Size of smallest rings (b) Size of smallest rings
Figure 1 Distribution of framework density (FD) versus size of smallest ring in the framework. The FD values are taken from the Atlas of Zeolite
Framework Types.10 CLO, RON, WEN, LIT, PAR, and CHI are excluded due to their interrupted structure. Framework types: (a) dense framework;
zeolite; hypothetical. (b) normal zeolite; mineral; real ELP structure; hypothetical ELP structure. ELP extra-large pore. Dotted lines
indicate the linearity for real ELP zeolites and hypothetical ELP zeolites. Parallelogram indicates the gap between zeolite and dense framework.
Reproduced from Jiang, J.; Yu, J.; Corma, A., Angew. Chem. Int. Ed. 2010, 49 (18), 31203145, with permission from John Wiley and Sons.

AlOAl bonds have not been observed in zeolites. This
fact, known as the Lowenstein rule, is based on the lower
stability of negatively charged clusters, as compared to isolated
negative charges.11 Thus, Si/Al ratios will always be 1.
A simplified empirical formula for a zeolite is the following:
Mn
x=n Alx Si1x O2 yX [1]
with x ranging from 0 to 0.5, Mn being the inorganic or
organic charge-balancing cation, and X the neutral guest mol-
ecules (H2O or others) which may be occluded in the pores.
Other elements, besides Si and Al, can be introduced in
tetrahedral positions of the framework. When this occurs, the
resultant materials are known as zeotypes, which include crys-
talline microporous aluminophosphates (AlPOs), silicoalumi-
nophosphates (SAPOs), or titanosilicates, among others.
Structural chemistry of zeolites has been recently surveyed by
Wright and Lozinska.12
An updated list of all zeolite-type materials is given by the
Structure Commission of the IZA,7,10 including interrupted
frameworks and intergrown materials. Each different structure
topology is defined by a three-letter code assigned by the IZA
and accepted by the International Union of Pure and Applied
Chemistry (IUPAC), the Framework Type Code, which usually
derives from the name of the type materials. It is based on the
connectivity of the tetrahedral atoms of the framework in the
highest symmetry possible, and is independent of possible
differences in distribution of the T-atoms or chemical composi-
tion that can lead to small changes in unit cell size and symme-
try. A detailed protocol for the nomenclature of structural and
compositional characteristics of ordered microporous and
mesoporous materials, following the IUPAC recommendations,
was published in 2001.13
The framework tetrahedra are considered as the primary
building blocks, which can be organized in secondary building
units (SBUs), containing up to 16T-atoms. The condition
for considering a finite group of T-atoms as an SBU is that it
is present in at least two different tetrahedral framework topol-
ogies. Some of the most abundant SBUs are represented in
Figure 2.
These SBUs are stacked together forming cages or channels
which will finally build up the three-dimensional (3D) struc-
ture, characteristic of a given zeolite. See, as an example, three
possible structures containing the sodalite cage in Figure 3.
Although some frameworks, such as the pure silica clathrasil,
contain no channels but only cages with maximum rings of six
T-atoms (six-member ring 6MR), most of them have channels
defined by at least eight T-atoms (8R).15 In this way, zeolites
may be classified according to the size of the channels, defined
by the number of T-atoms conforming the channel window or
ring, into small (8R), medium (10R), large (12R), and extra-
large (14R or larger). These channels will belong to a channel
system, which can be mono-(1D), di-dimensional (2D), or tri-
dimensional, and that can be straight, sinusoidal, or follow a
zigzag pattern. The channel system can be independent, with no
direct access from one channel system to another, or it can be
connected with a second channel system. Again, the reader is
directed to the Atlas of Zeolite Frameworks for more detailed
information, nomenclature, and specific examples.10
Zeolites 105
4 6 8
(71) (51) (24)
42 51
(23) (25)
Figure 2 Secondary building units and their symbols (the number in
parenthesis represents frequency of occurrence). Reproduced from
Baerlocher, C.; McCusker, L. B.; Olson, D. H. Atlas of Zeolite Framework
Types, 6th ed.; Elsevier: Amsterdam, 2007, with premission from Elsevier.
Despite the relatively low number of frameworks synthe-
sized up to date, more than 1 million thermodynamically
feasible zeolite structures are included in the database
published by Foster and Treacy.16 Some of the structures pre-
dicted there by computational simulation have already been
obtained, such as the extra-large 18-R pore ITQ-33.17 Other
databases with hypothetical zeolite or zeotype frameworks are
also available.1823
The presence of channel intersections (2D or 3D channel
systems) is highly convenient in many catalytic applications of
zeolite-based materials, as it favors diffusion of reactants and
products involved in the process. Thus, 3D microporous zeolites
such as Y and Beta (FAU and BEA structures, respectively) have
been widely employed since the very early commercial zeolite
applications. In these two examples, the channel system is tridi-
mensional, but the pore size is unique, for example, 12R. When
the interconnected pores are of different sizes the zeolite is
considered as multidimensional, and in this case the concept
of molecular traffic control (MTC) can be applied. This special
type of shape selectivity was first described by Derouane24 and
refers to a situation in which reactant molecules preferentially
enter the catalyst through a given channel system, while the
products diffuse out by the other , thereby preventing the occur-
rence of important counterdiffusion limitations in the catalytic
conversions. Thus, in zeolites with interconnected pores of
different diameter, molecules of different sizes can diffuse pref-
erentially through one channel or the other (Figure 4).
Some multidimensional zeolites are Ferrierite25 (FER, 8  10
MR), Mordenite26 (MOR, 8  12 MR), SSZ-26 and SSZ-33 (con-
taining major structural units of CON, 12  12  10 MR),27 CIT-1
(CON, 12  12  10 MR),28 essentially the pure polymorph B of
SSZ-33, or IM-12/ITQ-15 (UTL, 12 14 MR).25,29,30 More
recent examples are zeolite ITQ-22 (IWW, 8  10  12 MR),31
MCM-68 (MSE 12  10  10 MR),32 ITQ-33 (10  18 MR),17
ITQ-39 (10  12 MR),33,34 ITQ-40 (15  16  16 MR),35 ITQ-44
(18  12  12 MR),36 or ITQ-43 (12  12  28 MR).37 More
106 Zeolites

Sodalite

Zeolite A

Silicon or
Oxygen aluminum

Sodalite
cage

Faujasite
Figure 3 Schematic of zeolite frameworks. Construction of a sodalite cage and of three different framework topologies containing this building unit.
Reproduced from Davis, M. E. Ind. Eng. Chem. Res. 1991, 30, 16751683, with permission from ACS Publications.

P1o P2o ii

i
P1o P1.2 P1o

i
P1.2 i

o
P1

o o
(a) (b) P1 P2

Figure 4 Illustration of the molecular control effect. (a) Zeolite with single type of channel. (b) Schematic representation of zeolite ZSM-5:
(i) near-circular and sinusoidal channels, (ii) elliptical straight channels. Reproduced from Derouane, E. G. J. Catal. 1980, 65, 486489, with permission
of Elsevier.

details on this type of zeolites and their properties are given in
Section 5.05.5.4.
Traditionally, zeolites grow into tridimensional crystals, in
sizes and shapes that can greatly differ, depending not only on
the structure, but also on the composition and/or synthesis
conditions, as we will see in the next section. The need for
increasing the accessibility to the microporous structure, espe-
cially for catalytic applications, has directed research not only
to the reduction of crystallite size, but also to alternative solu-
tions, such as the permanent separation of zeolite layers when
starting from layered zeolite precursors, or, more recently, to
the direct synthesis of zeolite layers.
The formation of layered crystalline structures has been
described by several authors in early stages of the synthesis
process of Beta,38 ZSM-48,39 and other zeolites.4042 In these
cases, the layers are unstable precursors or intermediates, which
further evolve within the synthesis media resulting in the final
zeolite,43 although no real evidences have been found for the
direct transformation of these 2D structures into 3D crystals. In
fact, according to Tuel,39 the layered structures were completely
degraded before the zeolite products started to form.
However, other zeolite precursor materials crystallize into
stable 2D layers,4446 which are separated by hydrogen bonding
and/or organic molecules located in the interlamellar region,
instead of being covalently connected in the third dimension by
SiO or AlO bonds.43,4749 They can be converted into the
corresponding tridimensional zeolite structures after topotactic
condensation at high temperatures (>450 C).50 However, in
 Zeolites 107

these cases, it has also been possible to separate the 2D sheets,
obtaining single zeolite layers with thickness ranging between 0.6
and 2.0 nm, by expanding and/or exfoliating the layers by the
application of vigorous stirring or ultrasonic treatment to the
suspension. The expansion of the layers can be achieved by
the so-called swelling procedure, wherein the organic molecules
used during the synthesis are substituted by others, bulkier,
decreasing in this way the interlayer interactions and facilitating
their separation (Figure 5).
This methodology allows obtaining the so-called delami-
nated zeolites, which on the one hand maintain short-range
order and high thermal stability, but on the other present
highly increased external surface areas, that can go up to
1000 m2 g1, conferring them very interesting properties as
catalysts, as carriers or supports, and potential application as
zeolitic membranes.51 In this way, and starting from different
layered precursors, zeolites ITQ-252,53 (Figure 4), ITQ-6,54,55
ITQ-18,56 and ITQ-2057 have been prepared. Furthermore,

Swelling Delamination

MWW Precursor Swollen MWW ITQ-2

HMI HTMA+

OH OH OH OH OH OH
12 MR

Hexagonal
prism 10 MR

OH OH OH OH OH OH

ITQ-2 Layer structure

25 A
12

12 A

13 A

(a) (b)

OH OH OH OH OH OH OH OH OH OH

OH OH OH OH OH OH OH OH OH OH
12 MR

Hexagonal
10 MR
prism
(c)

Figure 5 Preparation and structural characterization of ITQ-2.

108 Zeolites

a recent publication describes the use of the single layers to overlapping of reflections that greatly limit this technique for
prepare hybrid organicinorganic materials, which combine zeolite structure resolution. Electron diffraction technique,
micro- and mesopores.58 based on low-dose high-resolution electron microscopy, is a
Besides calcination or delamination, the layered precursors recently developed approach for zeolite structure determina-
can be converted into pillared zeolites, in a similar way as tion, which has extended the range of crystals to be studied to
layered clays were started to be pillared in the late 1970s.59,60 nanocrystals.81 Recently, PXRD and transmission electron
These pillared materials, with a permanently expanded inter- microscopy (TEM) have been combined to solve structures as
lamellar space, are obtained by a first expansion or swelling of complex as TNU-9.82 The implementation of charge flipping or
the layers, in a similar way as for delaminated zeolites, fol- structure envelope calculations to the former techniques led to
lowed, in a second step, by a treatment for insertion of a the resolution of IM-5 (Figure 6), ITQ-33, and ITQ-37.83
soluble precursor, usually a liquid alkoxide, which will trans-
form into pillars by calcination. The first pillared zeolite-based
material MCM-3661 was obtained by treatment with tetraethy- Table 1 Layered precursors and materials formed starting
lorthosilicate (TEOS) of the swollen MCM-22P, followed by from them
hydrolysis and calcination. MCM-36 contains interlamellar Precursor Treatment Final Reference
mesopores in addition to the layer microporosity. material
The layered precursors known so far are summarized in
Table 1, as well as the different zeolitic materials that can be MWW(P) (MWW) Calcination MCM-22 62
obtained by calcination, delamination, or pillaring. Pillarization MCM-36 63
Recently, mono- and bilayered zeolite materials have been Delamination ITQ-2 52
obtained by direct synthesis and they are described later.79 PREFER (FER) Calcination Ferrierite 64
Pillarization ITQ-36 65
The determination of the framework structure of new zeo-
Delamination ITQ-6 54
lites is a challenge from the crystallographic point of view, as NU-6(1) (NSI) Calcination NU-6(2) 66
well as an issue of major importance, not only because of their Delamination ITQ-18 56
possible applications, but also regarding the zeolite chemistry PREITQ-19 (CDO) Calcination ITQ-19 57
itself, elucidating or directing new synthetic paths.15,80 A his- Delamination ITQ-20 57
torical perspective on the impact of structure analysis on zeo- SSZ-25 ASa (MWW) Calcination SSZ-25 Cb 67
lite science has been recently given by Gies.81 When single ERB-1 AS (MWW) Calcination ERB-1 C 68
crystals are available, crystal structure analysis is fairly direct.80 EU-19 (CAS-NSI Calcination EU-20b 69
However, this is rarely the case when it comes into zeolites, intergrowth)
since they are usually obtained as polycrystalline powders, RUB-15 (SOD) Calcination Silica SOD 70
RUB-18 (RWR) Calcination RUB-24 71
with crystals too small (in the micron range or smaller) to
RUB-39 (RRO) Calcination RUB-41 72
allow the use of single-crystal methods. MCM-65 AS (CDO) Calcination MCM-65 C 73
In the last decade, there has been an important develop- PLS-1 (CDO) Calcination CDS-1 74
ment of direct and indirect methods to solve zeolite framework UZM-13 (CDO) Calcination UZM-25 75
structures of increasing complexity. Moreover, other structural UZM-17 (CDO) Calcination UZM-25 75
issues, such as the detailed crystallographic geometry or the UZM-19 (CDO) Calcination UZM-25 75
location of isomorphically substituted framework atoms or HLS (SOD) Calcination Ga-SOD 76
extra-framework species, can also be obtained.80 Powder [F, Tet-A]-AlPO-1 (AFO) Calcination AlPO-41 77
x-ray diffraction (PXRD) is the most common method used SAPO-34 prephase Calcination SAPO-34 78
for characterizing a zeolite powder. However, the diffraction a
AS, As-synthesized.
pattern obtained in this way is the superposition of all the b
C, Calcined.
individual diffraction patterns of randomly oriented single Adapted from Strohmaier, K. G. In Zeolites and Catalysis, Synthesis, Reactions and
crystallites present in the powder sample, and gives, therefore, Applications; Cejka, J.; Corma, A.; Zones, S., Eds.; Wiley-VCH: Weinheim, 2010;
the average crystal structure of the bulk sample. This leads to pp 5786, with permission of John Wiley and Sons.

z
x y
Figure 6 Projection of the framework of zeolite IM-5 (IMF) seen along [100]. Reproduced from Gies, H.; Marler, B. In Zeolites And Ordered
Porous Solids: Fundamentals and Applications, Martinez, C.; Perez-Pariente, J., Eds.; Editorial Universitat Politecnica de Valencia: Valencia, 2011;
pp 6790.
 Zeolites 109

Although the combination of high-resolution transmission previously formed directly from the liquid phase. The further
electron microscopy (HRTEM), electron diffraction, and x-ray incorporation of the species in solution to the nuclei will
powder diffraction (XRPD) data has shown to be useful for the enable their growth into the final zeolite crystals. The solid
solution of complicated zeolitic structures,84 ab initio structure phases, if present, will only play the role of nutrient reservoir.
solution by direct methods may be faster and more reproducible. The heterogeneous mechanism proposes the reorganization of
In this way, automated electron diffraction tomography,85,86 an initially amorphous solid phase into the final crystalline
based on pure electron diffraction data, has recently allowed zeolite structure (solidsolid transformations), and nucleation
the resolution of a new extra-large-pore zeolite, ITQ-43,37 and at the liquidsolid hydrogel interface. Figure 7 shows a sche-
the highly complex structure of the new intergrown aluminosil- matic representation of these two mechanisms.
icate ITQ-3933 has been solved by means of rotation electron
diffraction tomography. The procedure is similar to the one Table 2 Summary of principal proposals for zeolite synthesis
used for single-crystal x-ray diffraction but on crystals million mechanisms
times smaller.34
Despite the large number of already existing zeolite frame- Author(s) Main futures of mechanism
work topologies and compositions, it is not always possible to 87
Barrer Condensation polymerization of polygonal and
find the ideal structure for a certain application. Thus, the
polyhedral anions
tailor-made synthesis would be the key objective within zeolite
Flanigen and Breck95 Linkage of polyhedra (formed by M-assisted
research. However, the prediction of new zeolite structures arrangement of anions): crystal growth
based on accumulated knowledge and understanding of the mainly in the solid phase
zeolite formation process is still to be achieved, due to the Kerr96 Crystal growth from solution species
complex synthesis mechanisms taking place. This issue is Zhdanov97 Solid $ liquid solubility equilibrium, nuclei
discussed in the next section. from condensation reactions, crystal growth
from solution
Derouane, Synthesis A: liquid-phase ion transportation
5.05.3 Zeolite Synthesis Mechanisms Detremmeric, Synthesis B: solid hydrogel phase
Gabelica and transformation
Blom98,99
The first synthetic zeolites were prepared in the 1950s under
Chang and Bell100 Embryonic clathrate TPA-silicate units,
hydrothermal conditions, in basic aqueous media and using ordered into nuclei through OH-mediated
inorganic cations as charge-balancing species.8789 Since then, SiOSi cleavage recombination
many approaches have been made which have enabled the Burkett and Pre-organized inorganicorganic composites,
discovery of a large number of new zeolite framework types, Davis101,102 nucleation through aggregation, crystal
and detailed historical perspectives on the evolution of zeolite growth layer by layer
synthesis have been written by Flanigen90 and some years later by Leuven group103,104 Oligomers ! precursor trimer (33 Si) ! 
Cundy.43,48 Still, many syntheses nowadays take place under 12 ! nanoslabs, growth by aggregation
hydrothermal conditions, following the next sequence of
Adapted from Cundy, C. S.; Cox, P. A. Microporous Mesoporous Mater. 2005, 82, 178,
steps:43 (1) mixing of the amorphous reactants containing silica
with permission of Elsevier.
and alumina with the cation source; (2) heating of the aqueous
reaction mixture, usually at a temperature ranging from 70 to
200  C, in a Teflon-lined stainless steel autoclave; (3) evolution
of the reaction mixture until crystallization of the final zeolite
phase after a first induction period; and (4) recovery of the zeolite
crystals by filtration or centrifugation, and washing and drying.
Although the zeolite synthesis procedure is rather simple,
the synthesis itself is a highly complex process, and a large
number of factors can strongly influence the type of zeolite
formed, its crystal size and morphology, or its composition.
Many empirical evidences relate those parameters, for example, (a)

type of mineralizing agent, solvent, structure-directing agents

(SDAs) or the presence of other heteroatoms, with the final
structure obtained,9193 but various aspects are yet not well
understood from the mechanistic point of view, and are still
a matter of controversy.43,48,94 Table 2 summarizes the main
mechanisms proposed.
At present, two extreme alternatives of nucleation and crys-
tal growth are considered,92,94 which can be related to the (b)
synthesis reaction media: the homogeneous mechanism, Figure 7 (a) Zeolite formation through gel dissolution and solution-
applied to clear solutions (dispersed low-density gels), and mediated crystallization. (b) Zeolite formation by in situ rearrangement
the heterogeneous mechanism. The former involves the diffu- of the gel. Reproduced from Guth, J. L.; Kessler, H. In Catalysis and
sion of soluble silicate, aluminate, and/or aluminosilicates Zeolites; Weitkamp, J.; Puppe, L., Eds.; Springer: Berlin, 1999; p 1, with
species from the liquid phase to the nuclei, which have been permission from Springer.
110 Zeolites
When it comes to real synthesis processes, the frontier
between these two proposals is not so clear. In fact, clear
solutions are in many cases sols and the colloidal phase present
will play a determinant role. A recent paper by Nikolakis
reviews the advances achieved in the period 200005 in
the understanding of interactions in zeolite colloidal
suspensions,107 according to experimental determination of
z-potential, theoretical and experimental identification of
zeolite crystal surface structure, and theoretical models.
As summarized by Tsapatsis et al. for the formation of
silicalite-1,108 the colloidally stable nanoparticles present in
the sol may participate in the synthesis process in several
ways: (1) as a source of nutrients during crystal growth,
(2) directly contributing to the crystal growth by addition to
the forming crystal, and (3) by addition to the growing crys-
tal, but possessing already the silicalite-1 crystal structure.
From their work it was concluded that crystal growth in this
case takes place by nanoparticles aggregation, and TEM stud-
ies revealed that the aggregates already presented oriented
nanocrystalline domains. There are different theories on
whether the zeolite structure is already present in the clear
sol nanoparticles or if it is formed only after attachment of the
amorphous nanoparticle to the growing crystal.104,109,110 In
one of the latest proposals for silicalite-1 crystallization,109
it is concluded that although oligomeric silicate species
and nanoparticles of different crystallization degrees are all
present in the clear sol, only the highly crystalline nanoparti-
cles participate in the prenucleation phase forming small
6-nm units which take part in the subsequent nucleation
and aggregative growth of colloidal silicalite-1 crystals
(Figure 8).
Zeolite crystallization process can be followed at the molec-
ular scale by in situ spectroscopic tools, such as nuclear mag-
netic resonance (NMR), electron paramagnetic resonance
(EPR), infrared (IR), Raman, x-ray absorption near-edge struc-
ture (XANES), or extended x-ray absorption fine structure
(EXAFS), which contribute to the understanding of the pro-
cesses that govern the crystallization of zeolites. A tutorial
review has been recently published.111 Instead of elucidating
a unique mechanism, the authors conclude that the informa-
tion obtained through these in situ techniques reveals the
variety and complexity of crystallization mechanisms.
A more theoretical insight into nucleation and crystal
growth of zeolites and, in general, any crystalline material has
been given recently by Cubillas and Anderson.112
Prenucleation
Figure 8 Scheme of the crystallization mechanism proposed by Martens et al. Particles with poorly developed framework connectivity are depicted
in light gray; zones with zeolite-like framework connectivity are depicted in dark gray; the zigzag arrows indicate an aggregation event. Reproduced
from Aerts, A.; Haouas, M.; Caremans, T. P.; Follens, L. R. A.; van Erp, T. S.; Taulelle, F.; Vermant, J.; Martens, J. A.; Kirschhock, C. E. A. Chem.-Eur. J.,
2010, 16 (9), 27642774, with permission from John Wiley and Sons.
5.05.4 Main Factors Affecting the Synthesis
Chemistry
5.05.4.1 Mineralizing Agents and Their Role as
Inorganic Structure Directiong Agents (SDAs)
Barrer and Milton can be considered as the founders of zeolite
synthesis science due to their contributions in the 1940s.48
Barrer synthesized the first zeolite which was not known as a
natural mineral. This zeolite, with a KFI structure, was obtained
as two different variants, P and Q.113,114 In 1949, Milton
synthesized zeolites A, B (now known as NaP), and C (hydro-
xysodalite) together with a crystalline impurity designated
X.89,115 One year later, the pure zeolite X was obtained.
All these first synthetic zeolites, such as zeolites A (LTA)88 or
X (FAU),89 zeolite P,113,114 sodalite,116 or chabazite (CHA),117
were aluminum-rich materials. The synthesis was performed in
basic media, and the high pH was obtained by means of alkali-
metal hydroxides. In these cases, the hydroxyl anion acts as a
mineralizing agent, increasing the solubility of silicon and
aluminum precursors, facilitating the interaction of the reac-
tants and leading to nucleation and growth of the zeolite
crystals.118 Moreover, there is an optimum hydroxide concen-
tration for every case, at which the species in solution and the
solid species in the gel are in equilibrium. At concentrations
below the optimum, solubility of the precursors would not be
sufficient, whereas at concentrations too high crystallization
will not take place due to the high stability of the dissolved
species.47 Concerning the alkali cations, they have a charge-
balancing role, and the higher the cation concentration, the
lower the Si/Al ratio obtained, with a minimum value limited
to 1. However, it has been found that the alkali metals may also
act as SDAs. A nice example was presented by Dutta et al.119 By
varying the cation in the series Na, K, and Cs maintaining
the rest of the synthesis conditions, they obtained zeolite A,
chabazite and edingtonite, respectively, all of them containing
the 4R SBU in their structure.
Besides the extended use of hydroxide ions as mobilizing
agents, fluoride anions can also be used as mineralizing agents
in zeolite synthesis, as was first described by Flanigen and
Patton120 for the preparation of a pure silica MFI (silicalite-1).
Guth and Kessler thoroughly studied the synthesis of micropo-
rous materials in fluoride media,121 as recently reviewed by
Caullet et al.,122 and those early results evidenced important
differences as compared to the OH route, such as the possibil-
ity of solubilizing the species at lower pH values (in the range of
Nucleation Crystal growth
 Zeolites 111

68) and the formation of larger crystallites with very low lattice
defects.92,93,123,124 The lower pH of the synthesis media enables
the incorporation of new elements in the framework, which are
less soluble in alkaline media. Moreover, it increases the number
of organic molecules to be potentially used as SDAs, as many of
them are unstable at the high pH values obtained in the pres-
ence of the alkaline hydroxides. The formation of large crystals
has a clear impact on the possibility of structure determination
by direct methods, and the absence of defects and high Si/Al
ratios result in zeolites with a highly hydrophobic character,
and, therefore, with a completely different behavior in their
possible applications (adsorption and catalysis).125 Finally, the
syntheses in fluoride media are performed in the absence of
alkaline cations.
The use of highly concentrated fluoride synthesis media has
led to the discovery of new pure silica materials with highly open
framework.127132 Initially, the formation of these low-FD zeo-
lites was assigned to the low dilution of the synthesis media (low
H2O/SiO2 ratio),127 but later studies have shown that the deter-
minant factor is the concentration of fluoride anions.133
The F anions tend to be located in small cages within the
SiO2 framework (Figure 9), compensating the positive charge
of organic cations used as SDAs, and allowing in this way the
synthesis of pure silica materials with no need for trivalent
elements to be present. More recently, it has been found that
the F ions can be preferentially found in small [435261] and
[4354] cages in ITQ-4 and SSZ-23, respectively134 as well as
in D4R [46] cages in ITQ-7,132 ITQ-17,135 ITQ-21,136
ITQ-13,137 ITQ-22,31 and ITQ-24.138 This fact, which can be
considered as a true directing effect of the F anions,139 was
also supported by theoretical calculations140 and magic angle
O D4R units could be used to synthesize frameworks with lower
N+ F FD, which might result in structures with high micropore
F F
O volume and extra-large pores.9
(a) (b) (c)
Figure 9 Three types of F ion environments found in siliceous
zeolites: (a) as part of an ion pair, (b) in the center of a small cage far
from any Si atoms, and (c) coordinated to a Si atom to form part of a
pentacoordinated SiO4F unit. Reproduced from Attfield, M. P.;
Catlow, C. R. A.; Sokol, A. A. Chem. Mater. 2001, 13, 47084713, with
permission of ACS publishers.
spinning (MAS) and cross-polarization-magic angle spinning
(CPMAS) NMR characterization results.141,142
5.05.4.2 Isomorphic Substitution in Zeolites:
Structure-Directing Effect of Heteroatoms in Framework
Positions
Although the presence of F directs to frameworks containing
D4R units, as discussed in the former section, and all-silica
zeolites have been successfully synthesized following this
route, the introduction of germanium as a tetrahedrally coor-
dinated framework element has also an important structure-
stabilizing effect toward those SBUs, and D4R SBUs were
observed to be formed during the synthesis of germanates
such as ASU-7.143 The isomorphous substitution of Si by tet-
ravalent Ge retains the neutral charged zeolite lattice, but
increases the average TO distance and decreases the TOT
angle, relaxing the structure and increasing its flexibility.9 The
higher thermodynamic stability of D4R when Si4 is isomor-
phically substituted by Ge4 and the special preference of Ge
for occupying positions in the D4R, even in small amounts,
has been confirmed by theoretical calculations and evidenced
in the synthesis of ITQ-17.144 ITQ-17, the C polymorph of
zeolite beta, contains D4R in the structure and was never
synthesized before as a pure silica or silicogermanate structure.
The results given in Table 3 show that in the presence of
different organic SDAs different zeolite structures were formed
in pure silica media. However, ITQ-17 was always obtained,135
provided that Ge was present, indicating the strong directing
effect of Ge toward structures with D4R. Since this discovery,
many new zeolite structures containing D4R have been syn-
thesized, as will be shown.
At this point, and considering the predictions enclosed in
the theoretical paper of Brunner and Meier,8 it appeared that
the directing effect of Ge to synthesize structures containing
O
O In fact, some interesting zeolites, most of them with multi-
dimensional large- or extra-large-pore systems, have been
discovered in the last decade by the incorporation of germa-
nium to highly concentrated fluoride synthesis media in com-
bination with the application of high-throughput (HT)
experimentation. In this sense, ITQ-32, a bidirectional zeolite
with interconnected 8R and 12R channels, can be prep-
ared as a nearly pure silica zeolite and as aluminosilicate.145

Table 3 Synthesis conditions and zeolitic structures obtained by using different SDAs with and without germanium in the crystallization gel

SDA H2O (w) Si/Ge T ( C) Time (h) Ge zeolite Si zeolite

BQ 8 110 150 15120 Polymorph C ITQ-4
BD 824 0.530 135175 15120 Polymorph C ITQ-4
m-XydQ2 8 5 150 1696 Polymorph C Beta
M4BQ2 7.515 210 175 2496 Polymorph C Beta,ZSM-12
M4BD2 15 220 175 2496 Polymorph C ZSM-12
CyHMP 7.25 5 135175 1596 Polymorph C ZSM-12
MCyHMP 7.25 5 135175 1596 Polymorph C Beta
TEA 7.5 2 140 96 Polymorph C Beta

Adapted from Corma, A.; Navarro, M. T.; Rey, F.; Rius, J.; Valencia, S. Angew. Chem. Int. Ed. 2001, 40 (12), 22772280, with permission of John Wiley and Sons.
112 Zeolites
Zeolite ITQ-33 is synthesized as a silicoaluminogermanate,17
and its structure, already predicted by Foster and Treacy as
thermodynamically feasible by computational simulation,
exhibits straight large-pore channels with circular openings of
18R along the c axis, interconnected by a bidirectional system
of 10R channels. All the former structures present double
4R units, but frameworks with D3R have also been synthesized
as germanosilicates, such as the extra-large-pore zeolites ITQ-
4436 (18  12  12-R), shown in Figure 10, and ITQ-40,146
formed by 15  16  16-membered channels, and with the
lowest FD of all existing four-coordinated oxide frameworks
up to date, 10.1 T-atoms per 1000 A3. The first chiral zeolite
with a single gyroidal channel, ITQ-37,147 has also been
obtained following this synthesis route. This germanosilicate
presents extra-large 30R windows and pore-size dimension
within the mesoporous range.
Very recently, the first mesomicroporous hierarchical zeo-
lite (ITQ-43) has been synthesized.37 This structure contains
mesopores of 2.2 nm in one direction, and micropores of
0.68  0.61 and 0.78  0.57 nm in the other two directions.
Schemes of the pore topology and dimensions, as well as a
representation of the global structure, are given in Figure 11.
Elements different from Si or Al in tetrahedral framework
positions will present different TOT angles, TO bond
lengths, and different charges, and could therefore stabilize
other SBUs favoring the synthesis of new zeolites. Table 4148
shows the range of bond angle and bond length variations for
structures where part of the Si atoms is isomorphically
substituted by other elements, di-, tri-, or tetravalent. Thus, in
the same way that germanium in tetrahedral framework posi-
tions favors D3R and D4R by affording less stressed OTO
angles as compared with pure silica structures,149,150 other
heteroatoms such as Be2,151,152 Zn2,153155 Mg2,156 or
Li,157,158 have been successfully incorporated in framework
positions favoring the formation of new structures, most of
T2
T1
T4
T3
(b)
(a)
(c)
Figure 10 The structure of ITQ-44. (a) Basic cluster of ITQ-44, indicating the location of the four independent T-atoms. (b) Connection of the basic units,
forming D4R cages, as the origin of the basic sheets of the structure. The bold arrow indicates the connection. (c) View of the structure along [001], showing
the 18R channels. (d) ITQ-33: condensation of the basic sheets through the 3R, forming the 10R. (e) ITQ-44: connection of the 3R by bridging O atoms,
forming the D3R and the 12R; the ellipses highlight the different connections in both zeolites. Oxygen atoms have been omitted for clarity. Reproduced
from Jiang, J.; Jorda, J. L.; Diaz-Cabanas, M. J.; Yu, J.; Corma, A. Angew. Chem., Int. Ed. 2010, 49, 4986, with permission of John Wiley and Sons.
them containing spiro-5 units (two 3R sharing one of their
vertices). The beryllium containing OSB-1 presents an extra-
large 14R channel system, whereas the microporous zincosili-
cates discovered so far, for example, VPI-7, VPI-9, VPI-10, and
RUB-17, are small-pore materials with 8R and 9R.159 The
isomorphic substitution of Mg2 results in the synthesis of
crystalline dense phases.
The trivalent boron and gallium have also been introduced
in framework positions. The inclusion of boron changes the
phase selectivity as compared to aluminum, and the substitu-
tion of Al with B, maintaining the rest of the synthesis condi-
tions, has directed to the formation of different zeolite
structures, as in the case of SSZ-31 and SSZ-33.160 The use of
Ga3 has also led to new structures with no aluminosilicate
analogs, such as TsG-1 (CGS)161 or ECR-34 (ETR). The latter,
discovered by Strohmaier and Vaughan,162 presents extra-large
18-MR pores. The framework of these two gallosilicates, both
synthesized in the presence of K, can be built from the same
SBU (Figure 12) which has not been observed, as far as we
know, in any other aluminosilicate structure. This clearly indi-
cates a structure-directing effect of gallium for this particular
unit.
Lithium-containing zeotypes have also been synthesized,
despite the higher charge generated by the incorporation of
a monovalent element in the framework. However, these
lithosilicates, RUB-23157 and RUB-29,158 containing strain-
free 3R, have not been approved yet by the Structure
Commission of the IZA.
Other elements have been isomorphically substituted in
known zeolitic frameworks, not with structure-directing aims,
but with the purpose of modifying the physicochemical prop-
erties of the final materials, and therefore their potential appli-
cation. Thus, Ga and Fe can easily substitute Al,163 resulting in
materials with lower intrinsic acid strength of the zeotypes in
their acid form. This has been proved to be useful in the
(d)
(e)
 Zeolites 113

19.6

21.9 9.7 4.9

7.0

(a)

c
b [100]

c a
a [010] b [001]
(b)

Figure 11 Zeolite ITQ-43: (a) the 28R channel of zeolite ITQ-43 (oxygens omitted for clarity, except those in the silanol groups); (b) structure viewed
along the three main crystallographic axes. Oxygens omitted for clarity, except those in the silanol groups. Reproduced from Jiang, J.; Jorda, J. L.;
Yu, J.; Baumes, L. A.; Mugnaioli, E.; Diaz-Cabanas, M. J.; Kolb, U.; Corma, A. Science 2011, 333(6046), 11311134, with permission of The American
Association for the Advancement of Science (AAAS).

Table 4 Some changes in substitution for Si in a tetrahedral lattice
with oxygen
Substituting atom TO bond length (A) SiOT bond angle (average)
T Si (no change) 1.6 155  20
T Al  1.7 135  10
T Ge  1.74 145  15
T B  1.35 125  10
T SiF  1.76 120
Adapted from Zones, S. I. Microporous Mesoporous Mat. 2011, 144(13), 18, with
permission from Elsevier.
catalytic isomerization of light n-olefins (n-C4 to i-C4 ),
where higher acidity favors undesired side reactions.164,165
Tetravalent titanium and tin have also been successfully
substituted in framework positions. The first titanosilicate,
TS-1, was reported by researchers from Eni S.p.A.166 It presents
a medium-pore MFI structure and is an active and selective
oxidation catalyst for conversion of small molecules in the
presence of H2O2 as oxidation agent. Some years later, the
first large-pore Ti-Beta zeotype was described, able to oxidize
reactants of larger size.167 Ti-YNU-1168 and Ti-ITQ-17169 are
other large-pore titanosilicates synthesized in the last decade.
Finally, tin located in tetrahedrally coordinated framework
positions presents Lewis-type acidity. The presence of these
controlled, highly dispersed isolated Lewis acid sites confers
the zeolitic material with exceptional catalytic properties. Thus,
Sn-Beta has been described as an effective catalyst for oxidation
reactions,170 selective reduction processes,171 or transforma-
tion of biomass.172174
5.05.4.3 Organic SDAs
The need for increasing the Si/Al ratio of the final materials led
Barrer and Denny87 to the first use of organic cations in a
zeolite synthesis, and the high-silica ZSM-5 and beta zeolites
were obtained.175,176 Since then, the use of organic com-
pounds as SDAs has been decisive for the discovery of new
zeolites with different topologies, especially for the synthesis of
AlPOs or high-silica materials, as revised recently by Burton
and Zones.177
The main characteristics of an organic structure-directing
agent (OSDA) which will determine its suitability for the syn-
thesis of a certain zeolite structure are its polarity (hydrophobic/
hydrophilic character), its size, charge, shape, and rigidity.9
Concerning the polarity of the OSDA, a compromise has to be
achieved, as the organic molecule must be soluble in water, but
it should not form complexes with the solvent, at least to a large
extent. In the case of quaternary ammonium compounds used
as OSDAs, Zones et al. concluded that an adequate C/N ratio
ranged between 11 and 16.178 These molecules had the right
polarity for being used as OSDAs in the synthesis of high-silica
zeolites. The size to charge ratio of the OSDA will determine the
114 Zeolites
ETR
Figure 12 Chain connected open hexagonal prisms the secondary building unit found only in the gallosilicate ETR and CGS frameworks.
final framework charge density, or, in other words, the TIV/TIII
ratio of the framework, although the presence of F adds an
additional parameter to the extent of isomorphic substitution of
Si4 with TII or TIII elements, as we have seen in a previous
section. Concerning size and shape, a true templating effect is
rather unusual, and only one case can be cited, which is the
perfect match between the specific triquaternary amine (triquat,
C18H36N3) used for the synthesis of ZSM-18 and the zeolites
micropore volume.179,180. The role of the OSDA as a true SDA is
not frequent either, and only a few cases have been described,
such as in the synthesis of SSZ-24181 and in the synthesis of
hexagonal faujasite.182185 In the rest of the cases, the organic
molecules employed in the synthesis of zeolites have rather a
space-filling role, as could be the case of the 22 possible mole-
cules that can be used to synthesize ZSM-5. A more detailed
description on the role of OSDA in zeolite synthesis has been
given by Jiang and Corma.9
Besides charge balancing and determining the framework
TIV/TIII ratio, their packing into the cages and channels can
increase the thermodynamic stability of the organic-framework
system when the number of van der Waals interactions
between the organic molecule and the void space are maxi-
mized. Concerning this stabilization effect, Gies et al.186,187
reached the following conclusions, which are valid not only
for clathrasils, but also for other pure silica zeolites: the organic
guest molecule must be stable under the synthesis conditions;
it should have sufficient room within the cage and it should fit
the inner surface of the cage with the least deformation but
with as many van der Waals contacts as possible; the organic
molecule should have only weak tendency to form complexes
with the solvent; and more rigid molecules will tend to form
clathrasils more easily than flexible molecules, and the ten-
dency to form a clathrasil will increase with the basicity or
polarizability of the guest molecule.
The evolution of zeolite A [LTA] toward compositions with
decreasing Al content thanks to the introduction of organic
molecules is an instructive example, as previously pointed out
by Burton and Zones.188 Traditionally this zeolite was
synthesized in the presence of alkaline cations with a Si/Al
ratio of 1. The incorporation of tri-methylammonium (TMA)
increased the Si/Al ratio to values of 387,189,190 and UOP LLC
researchers synthesized an LTA structure with Si/Al ratio up to
9 (UZM-9)191 by combination of TMA with the bulkier tetra-
ethylammonium cations. Recently an LTA Ge-containing zeo-
lite with a (Si Ge)/Al ratio of 80 and even a pure silica LTA
zeolite (ITQ-29) has been synthesized.192 This was achieved by
using a rigid multicyclic quaternary ammonium compound as
SDA, able to self-assembly in solution forming a supramolec-
ular complex, which perfectly fits within the alpha cage of the
CGS
LTA structure (Figure 13). The pure silica LTA zeolite (ITQ-29)
was produced by adding TMA cations to the synthesis media
which were able to stabilize the sodalite units and to balance
the charge excess due to the fluoride anions within the D4R,
together with the bulkier OSDA. One of the most interesting
applications of this Al-free LTA, being highly stable and hydro-
phobic, has been shown to be gas separation in the presence of
H2O or other polar molecules, and shape-selective cracking.192
The above method has been the first example wherein a double
self-assembling process was demonstrated between two
OSDAs, and between OSDA and silica, within the synthesis
gel, as well as in the final zeolite.192
The group of Perez-Pariente has studied the combination of
small and bulky OSDAs in order to modify the location of the
tetrahedral Al, and therefore the accessibility of the active sites
associated to these Al, in ferrierites. They observed that both
SDAs, TMA, and the bulkier benzyl-1-methylpyrrolidium
(bmp), participated actively in the crystallization of the
zeolite.193197 By addition of a third template and variation
of the relative proportions of OSDAs, it was possible to modify
the distribution of the bridging hydroxyls located in the 10MR
channels and in the ferrierite cage, accessible only through
8MR windows. This tailored active site distribution has been
evidenced, not only by specific characterization techniques,
but also by a different catalytic behavior in isomerization of
m-xylene,196 conversion of n-butene,197 or carbonylation of
dimethyl ether with carbon monoxide.198 More details on the
cooperative structure-directing effects in the synthesis of crys-
talline molecular sieves when using more than one template
can be found in a recent review.199
The cooperative templating is also the basis for the charge-
density mismatch concept, developed by UOP researchers.200
They have shown that the use of a low-charge density large
organocation and TMA, with a high-density charge, enables the
synthesis of zeolites with lower Si/Al ratios. The final purpose
of multitemplating is substitution of expensive and compli-
cated OSDAs for synthesizing new zeolitic materials with large
pores or intersecting channels.
One of the main disadvantages of the use of OSDA is their
considerable impact on the synthesis procedure cost, as they
are usually expensive compounds, and because the most com-
mon way of removing these organic compounds is by combus-
tion. An interesting approach for nondestructive recovery of
the OSDA was presented by Lee et al.201,202 They propose the
use of ketal-containing SDA, which can easily be disassembled
and reassembled and therefore removed from the microporous
channel system without the need for calcination (Figure 14).
The SDA molecule, stable at the high pH values used during the
synthesis, is chemically cleaved into fragments when
 Zeolites 115

Self-assembling

Time
Temperature

Figure 13 Self-assembling SDA used for synthesis of ITQ-29 (high-silica LTA). Reproduced from Corma, A.; Rey, F.; Rius, J.; Sabater, M. J.; Valencia, S.
Nature 2004, 431, 287290, with permission from Nature Publishing.

H+, H2O

PFA PFA
SDA SDA SDA SDA
PFA PFA
PFA
SDA SDA
Extraction Cleavage
PFA
PFA PFA reaction
SDA SDA SDA SDA
PFA PFA

O
O O O O OH
Inside pore: + NH2 +
+ + +
N N N OH

SDA PFA

Figure 14 Schematic diagram of ZSM-5 containing both SDA and PFA and how they are removed. Reprinted with permission from Lee, H.; Zones, S. I.;
Davis, M. E. J. Phys. Chem. B, 2005, 109(6), 21872191. Copyright (2003) American Chemical Society.

decreasing the pH, being extracted in this way from the micro-
porous structure. In a last step, it can be recombined to form
the original SDA. Following this general methodology, the
authors have described the synthesis of ZSM-5, VPI-8, and
ZSM-12.
What can be concluded from the former sections, related to
the use of SDAs in zeolite synthesis, is that their exact role is not
yet fully elucidated. Although in some particular cases the
effect of the OSDA is very clear, these organic molecules are,
in general, not as specific as expected. The same OSDA can lead
to different zeolite structures depending on the synthesis
conditions,49 and certain structures can be synthesized with a
considerable number of different OSDAs. Moreover, it has
been shown that the presence of certain inorganic cations
(alkalines, Zn, B, Ga, and Ge) in the synthesis gel, or the use
of fluoride anions as silica mobilizing agents, can also direct
certain structural features, and will therefore promote specific
structures.
5.05.4.4 Solvents
Most of the zeolite materials are synthesized in aqueous media.
The main advantages of using water are its nontoxicity, low cost,
and capacity to dissolve the precursors used. The substitution of
water by other solvents is described below, but when consider-
ing aqueous synthesis media, the water content can play a
decisive role in the crystallization process and the final structure
obtained. Conventional zeolite synthesis uses H2O:SiO2 ratios
above 25, obtaining in this way homogeneous synthesis gels
that are easy to handle. Decreasing this ratio to values below 10
in the presence of fluoride has allowed not only to make the all-
silica form of known zeolites, but also to discover new low- FD
structures.9,47 Furthermore, the water concentration can modify
in some extent the structure-directing ability of the SDA.127
Thus, very different pore architectures can be obtained in the
presence of the same SDA, with a clear influence of the
H2O:SiO2 ratio on the FD of the final phase obtained. The effect
116 Zeolites
is illustrated in Figure 15, showing the trend to form less dense
phases at lower H2O concentration.
The extreme case of decreasing the water content in the
zeolite synthesis media is the dry-gel conversion (DGC)
procedure,203 which was first described by Xu et al. in 1990.204
They reported the synthesis of an MFI zeolite by contacting a dry
aluminosilicate gel with vapors of water and volatile amines.
This DGC technique was further developed during that decade,
leading to the synthesis of zeolites with new structures and
compositions.203 Two different DGC methods have been
described. In the vapor-phase transport synthesis (VPT)204206
the dry gel does not contain the SDAs, but the organic com-
pounds are mixed with water and placed at the bottom of the
autoclave. The amorphous powder obtained by drying the gel is
exposed to the steam and the SDA vapors when temperature is
increased to the synthesis values, and transformation of the gel
leads to the crystalline phase. The second method is applied
when nonvolatile SDAs are used. In this case, the SDA is added
to the aluminosilicate gel before the drying step, and only steam
is supplied as a vapor. This is the so-called steam-assisted con-
version (SAC).207 In any case, both procedures take place in the
presence of water, although the synthesis media cannot be
considered as an aqueous phase. Figure 16 compares both
methodologies.
The main advantage of the DGC procedure is the non gen-
eration of a mother liquor. However, recovery of the organics for
recycling, manipulation of the dry paste, and homogenization
of the temperature profile during the synthesis may be serious
drawbacks for a large-scale commercial application.208
Besides water, other solvents have been described,209 such
as alcohols, amines, or ionic liquids, which are described sep-
arately in Section 5.05.5.2. Bibby and Dale, in 1985, described
the first nonaqueous (solvothermal) zeolite synthesis in ethyl-
ene glycol or propanol media, obtaining the all-silica form of
sodalite,210 although a complete anhydrous reaction media
was not ensured. Other sodalite-related structures were synthe-
sized in ethanolamine and ethylenediamine media. The use of
ITQ-9, 1D 10 MR
H2O/SiO2 = 3.75
Figure 15 Changes in the selectivity of DMABO to pure silica phases at 150  C as a function of the degree of dilution of the crystallizing mixture.
Reproduced from Camblor, M. A.; Villaescusa, L. A.; Diaz-Cabanas, M. J. Top. Catal. 1999, 9, 5976, with permission from Springer.
nonaqueous solvent also provided a method for growing
uniform, well-defined large crystals of molecular sieves with
different compositions (pure silica, aluminosilicates, and
AlPO).211 These large crystals, in the range of 0.45.0 mm,
facilitate structure refining and the study of adsorption and
diffusion properties for zeolites and zeotypes. Other reaction
media, such as mixtures of aqueous and immiscible organic
phases, liquid ammonia, aqueous ammonia, or, more recently,
hexanol combined with a soluble organic co-solvent, have
also been described for synthesizing different zeolite analogs.
A more detailed overview is given by Strohmaier.47
5.05.5 New Trends in Zeolite Synthesis
5.05.5.1 Phosphorus-Containing SDAs
Most of the organic compounds employed as OSDA, as
described in Section 5.05.4.3, are quaternary ammonium
salts, susceptible to Hoffman degradation at high pH and
temperature. One way of approaching the OSDA stability
problem is the synthesis in fluoride media, at nearly neutral
Gel Gel-containing SDA
H2O + SDA H2O
(a) (b)
Figure 16 Illustration of two DGC methodologies: (a) VPT method and
(b) SAC method. Adapted from Matsukata, M.; Ogura, M.; Osaki, T.;
Hari Prasad Rao, P.; Nomura, M.; Kikuchi, E. Top. Catal. 1999, 9(1), 7792,
with permission from Springer.
ITQ-3, 2D 8 MR SSZ-31, 1D 12 MR
7.5 15
N*
DMABO+
 Zeolites 117

R4 R3 R2 R1 vessels. Although most ionic liquids are stable under ionother-

R2 R1 mal conditions, some may suffer decomposition at high tem-
R3 N R5 N N peratures or in the presence of fluoride ions.217 It has been
NR7 NR11 observed that the decomposition products may also act as
N P N P N P
SDAs, leading to specific zeolite structures. A detailed descrip-
R4 R6
N N N tion of this synthesis approach has been recently presented by
R5 R6 Morris.221
R7 R8 R9 R10
Base P1 Base P2

5.05.5.3 Synthesis of Extra-Large Pore Zeolites

R2 R18 R17
R1
R3 N NR19 N defined by more than 12 T-atoms were metalphosphates, such
R4 N P N P N P
R5 N N N R14
R6 R13
N P N considered as true zeotypes. Other microporous solids with
R7 R12
R8 N manates, but they all present a limited thermostability, which
R11
R9 R10
Base P4
Figure 17 Examples of phosphazene structures. Reproduced from
Simancas, R.; Dari, D.; Velamazan, N.; Navarro, M. T.; Cantin, A.; Jorda, J.
L.; Sastre, G.; Corma, A.; Rey, F. Science 2010, 330, 1219, with permission
of The American Association for the Advancement of Science (AAAS).
pH, described above. A recent alternative proposal has been the
use of phosphonium (instead of ammonium)-based OSDA,
more thermally stable, which has led to the discovery of new
structures such as ITQ-27212 and ITQ-34.213 Besides phospho-
nium cations, phosphazene derivatives (Figure 17) have also
been used as SDA; in this way the Boggsite zeolite, known up to
now only in its natural form, has been produced214 and named
as ITQ-47. Its structure presents a bidirectional microporous
channel system formed by connected 10R and 12R pores.
Moreover, it has been recently shown that the presence of P
due to the OSDA may stabilize the zeolite structure toward
hydrothermal treatments equivalent to the severe regeneration
conditions during catalytic cracking process.215
5.05.5.2 Zeolite Synthesis in Ionothermal Media
In Section 5.05.4.4, we have seen that alcohols or amines can
be used as alternative solvents to water for zeolite synthesis.
Another approach is the synthesis of zeolites and other micro-
porous materials in ionothermal media,216221 developed by
the group of Morris. The ionic liquids or the eutectic mixtures
used here act not only as a solvent, but also as an SDA. The low
solubility of silica precursors in these synthesis media limited
the initial microporous materials obtained to AlPOs. However,
the procedure was improved by adding bromidehydroxide,
enabling in this way the synthesis of pure silica zeolites with
TON and MFI structures. An additional advantage of ionic
liquids as compared to water is their low vapor pressures, lead-
ing to the possibility of increasing the synthesis temperature
while maintaining ambient pressure and working in open
The first microporous crystalline materials with channels
R16
as the 18R AlPO4 VPI-5,222 the 20R JDF-20 AlPO,223 or the 3D
N
R15 20R Cloverite,130 a gallophosphate. Structures with up to 24R
channels have been described,9 but most of them present penta-
and/or hexa-coordinated atoms, and can, therefore, not be
extra-large pores are phosphites, organic phosphates, and ger-
limits their practical application.9
The first extra-large-pore zeolite was UTD-1, a high-silica
material with 1D 14R channels.224226 Other 14R structures are
CIT-5,227 the beryllosilicate OSB-1,152 and the borosilicates
SSZ-53 and SSZ-59228 discovered by Chevron. Strohmaier
and Vaughan reported an 18  8  8R gallosilicate, ECR-34,
with a pore diameter of 1.08 nm for the 18R pore. A series of
extra-large-pore silicogermanates have been discovered in the
last decade by the group of Corma (Section 5.05.4.2). A thor-
ough revision of the influence of different synthesis parame-
ters on the formation of extra-large-pore zeolites led to the
application of the following strategies9: (1) the use of large
and rigid 3D OSDAs with the required polarity to produce
materials with larger micropore volume and a multidimen-
sional system of channels; (2) the use of concentrated gels
and fluoride media, increasing in this way the probabilities
of obtaining structures with lower FD; (3) the synthesis of
frameworks containing 3R and 4R, and more specifically
D4R, which are known to favor structures with low FD; and
(4) HT synthesis techniques to facilitate the exploration of
wider synthesis spaces. Table 5 shows the extra-large-pore
zeolite structures known so far and the OSDA used for their
synthesis.
Besides catalytic applications, extra-large-pore zeolites may
find use as sensors or in photonics, or in the manufacture of
improved microelectronic devices. The latter is justified by the
low dielectric constant of these large-pore materials, related to
their porosity.230
5.05.5.4 Synthesis of Multipore Zeolites
We have already seen in Section 5.05.2 that multipore zeolites,
with interconnected pores of different sizes, are very interesting
from the perspective of their catalytic application. Besides the
shape selectivity and molecular-sieving properties, inherent to
their microporous channel systems, they may present MTC
applied to reactants or products of different sizes, whose diffu-
sion will be favored through the different-size pores in differ-
ent degrees. Some multidimensional zeolites are Ferrierite25
(FER, 8  10 MR), Mordenite26 (MOR, 8  12 MR), SSZ-26
118 Zeolites

Table 5 Extra-large-pore zeolites

+
HN N Co N+ N
N + N
N +
F
1 2 3 4 5 6 7

+ H H
+N N N +N N+ +N
+ 3 N+
8 9 10 11 H H
12

Structure code Material Year Channel dimension Framework atoms Framework density Template Reference
(T-atoms/nm3)

VFI VPI-5 1988 1D 18R Al, P 14.5 1,2 222

AET AlPO-8 1990 1D 14R Al, P 18.2 2 229
-CLO Cloverite 1991 3D 20R Ga, P 11.1 3 130
DON UTD-1 1996 1D 14R Si 17.1 4 224226
CFI CIT-5 1997 1D 14R Si 16.8 5 227
FH SSZ-53 2003 1D 14R B, Si 16.5 6a 228
SFN SSZ-59 2003 1D 14R B, Si 16.6 7a 228
OSO OSB-1 2001 3D 14  8  8R Be, Si 13.3 K 152
ETR ECR-34 2003 3D 18  8  8R Ga(Al), Si 15.4 8 and Na, K 162
UTL IM-12 or ITQ-15 2004 2D 14  12R Ge, Si 15.6 9 (for IM-12) 29,30
10 (for ITQ-15)
ITQ-33 2006 3D 18  10  10R Ge, Si 12.3 11 17
ITQ-37 2009 3D 30R Ge, Si 10.3 12 83
ITQ-40 2010 3D 15  16  16R Ge, Si 10.1 35
ITQ-44 2010 3D 18  12  12R Ge, Si 11.7 36
ITQ-43 2011 3D 28  2  12R Ge, Si 37
a
Only one example with this SDA.
Adapted from Jiang, J.; Yu, J.; Corma, A. Angew. Chem. Int. Ed. 2010, 49(18), 31203145, with permission from John Wiley and Sons.

and SSZ-33 (containing major structural units of CON,
12  12  10 MR),27 CIT-1 (CON, 12  12  10 MR),28 the
pure polymorph B of SSZ-33, IM-12/ITQ-15 (UTL, 12  14
MR),25,29,30 ITQ-22 (IWW, 8  10  12 MR),31 MCM-68
(MSE 12  10  10 MR),32 ITQ-33 (10  18 MR),17 ITQ-39
(10  12 MR),33,34 ITQ-40 (15  16  16 MR),35 ITQ-44
(18  12  12 MR),36 or ITQ-43 (12  12  28 MR).37
Some recent multipore zeolites have shown very interesting
catalytic properties. Thus, ITQ-22, whose structure is formed by
a system of connected 8-, 10-, and 12-MR channels,31 has shown
exceptional high selectivity to the desired products in the cata-
lytic alkylation of benzene with light olefins.231 The specific
topology of this material enhanced the rapid diffusion of these
products, avoiding their further conversion by consecutive
undesired reactions. ITQ-33, with intersecting 10R and 18R
pores, presents a previously unreported simultaneous selectivity
to propylene and to hydrocarbons in the diesel range in catalytic
cracking of a vacuum gasoil (VGO).232 The diesel molecules
formed by cracking of the bulky feed compounds present in
the VGO diffuse out the zeolite structure through the 18R chan-
nels, whereas the linear or monobranched aliphatics belonging
to the gasoline fraction will be able to enter the 10R pores and
suffer secondary re-cracking reactions to the higher-value
propylene. ITQ-39, presenting a 3D intersecting 12R and 10R
pore system,34 is highly selective to mono-alkylated products in
the alkylation of benzene to ethylbenzene and cumene,33 and
shows very interesting results as acid alkylation catalyst in the
conversion of low-value naphtha fractions into diesel by
alkylation.34
5.05.5.5 Synthesis of Organic-Free Zeolites
We have seen in the former sections that the incorporation of
OSDA has been a key factor in the development of zeolite
research, enabling the formation of high-silica zeolites and
new structures. Besides charge balancing and determining
the framework TIV/TIII ratio, their packing into the cages and
channels can increase the thermodynamic stability of the
organic-framework system when the number of van der Waals
interactions between the organic molecule and the void space
are maximized. Although some OSDAs are inexpensive, com-
mercially available compounds, others are complex organic
molecules, especially designed for synthesizing a specific struc-
ture. Their elimination from the final solid is usually performed
by high-temperature combustion, so they are not recovered.
Thus, the use of OSDA considerably increases the final
 Zeolites 119

production cost of the zeolite.148 Some approaches proposed to
decrease the cost due to OSDA have already been discussed in
Section 5.05.4.3, such as the combustion-free methodology or
the charge-density mismatch concept.200 However, the final
objective, and still a main challenge in zeolite science, is to
modify the synthesis procedure in order to obtain the zeolite
material in the absence of OSDA. This has been achieved
in some cases, such as ZSM-34,233,234 ZSM-12,235 ECR-1,236
ZSM-5,237 RUB-13,238,239 and beta,240,241 by seed-assisted syn-
thesis. The case of OSDA-free beta zeolite must be highlighted,
due to its very large commercial significance. Usually, the seeds
are prepared by conventional hydrothermal processes in the
presence of the corresponding OSDA, but Kamimura et al.
recently presented a process for the environmentally friendly
production of high-quality zeolite beta in a completely OSDA-
free Naaluminosilicate system.241
5.05.5.6 Synthesis of Zeolite Layers
Zeolite layers were initially obtained by means of a two-step
procedure, starting from laminar zeolite precursors. This led to
different delaminated and pillared zeolites as described in
Section 5.05.3, such as delaminated ITQ-2,52,53 ITQ-6,54,55
ITQ-18,56 and ITQ-20,57 or pillared material MCM-36.61 The
resultant materials showed very high accessible surface areas
and remarkable advantages toward conventional zeolites when
used as catalysts for a large number of catalyzed reactions
where bulky reactants were involved.58,242247
Single layers of zeolite ZSM-5 of 2-nm thickness have been
recently obtained by direct synthesis by Ryoo et al.79 They have
used large organic molecules with diquaternary ammonium
groups located at the end of the chain, such as C22H45
N1(CH3)2C6H12N1(CH3)2C6H13, composed of a long-
chain alkyl group (C22) and two quaternary ammonium groups
spaced by a C6 alkyl linkage. The cationic groups act as SDAs to
crystallize the ZSM-5 monolayers, whereas the long hydropho-
bic chain limits their growth along the b direction, as shown
in Figure 18. The group of Ryoo has also been able to pillar
the ordered multilamellar mesostructure with silica pillars,
as in pillared clays, stable even after surfactant removal by
calcination, with controlled mesopore diameters according to
the surfactant tail lengths. Moreover, the pillars were highly
likely to be built with an MFI structure, as deduced from Ar
adsorption analysis and TEM results.248 In a recent paper, the
group of Ryoo presents a new approach to hierarchical
zeolites,249 obtained by growing thin zeolitic (microporous)
walls into hexagonally ordered or disordered mesoporous
structures. Surfactant aggregates generate the mesopores,
whereas the cationic surfactant head groups direct the crystal-
lization of the microporous zeolite walls. The use of different
surfactants will have a direct influence on wall thicknesses,
framework topologies, and mesopore sizes. The molecular
sieves have been shown to be highly active acid catalysts for
conversion of bulky molecular substrates, as compared with
conventional zeolites and ordered mesoporous amorphous
materials.
Another application of the ZSM-5 single layers is the prepa-
ration of surfactant-modified MFI-zeolite nanosheets (SMZNs),
which can be used as anion exchangers for removal of
nitrates.250 The SMZN material, easily regenerable, presents
high anionic exchange capacity and removal efficiency as com-
pared to classical SMZ prepared from clinoptilolite.

Surfactant location in MFI layer

b
[001] c
b Regular
a stacking
a
along
b-axis
Zigzag
channel

(b) MFI layer with wide ac plane Multilamellar MFI

Straight
channel

b
c
Random
[010]
a stacking

c
a

(c) MFI layer with narrow ac plane Unilamellar MFI

(a)

Figure 18 Crystallization of MFI nanosheets. Reproduced from Choi, M.; Na, K.; Kim, J.; Sakamoto, Y.; Terasaki, O.; Ryoo, R. Nature 2009, 461,
246249, with permission of Nature Publishing Group.
120 Zeolites
5.05.5.7 Application of HT Techniques and Data Mining
to the Discovery of New Structures
The former sections have evidenced the complexity of the
zeolite synthesis process, and the large number of parameters
which may play a decisive role in the final structure obtained.
Exploring large regions within a certain synthesis phase dia-
gram by means of traditional laboratory synthesis, for exam-
ple, modifying one parameter at the time, is usually a
considerable effort as well as a time-consuming process, with
no guaranty of finally obtaining the desired crystalline zeolite
material, or even any zeolite. An alternative approach devel-
oped during the last decade is the application of automated
combinatorial methods to the hydrothermal synthesis of zeo-
lites. The first study, performed by researchers from Sintef,
already proved the large improvement in efficiency achieved
by applying these strategies.251 A multiautoclave system was
designed (Figure 19), which allowed the parallel synthesis of
up to 1000 combinations in one experiment. The design of a
fully automated PXRD device, essential for a rapid characteri-
zation of the phases obtained, was at that time still in progress.
Since then, the so-called HT approach, based on a rational
design of experiments, has allowed the discovery of new zinc
phosphates252 and of new zeolites with open structures such as
ITQ-21,136,253,254 ITQ-30,255 or ITQ-24,138 in combination
with HT characterization and data-mining techniques. The
latter is necessary in order to identify each individual zeolite
Figure 19 View of the multiautoclave showing the mode of stacking
of the Teflon blocks and one of the alternative designs using Teflon
inserts. Reproduced from Akporiaye, D. E.; Dahl, I. M.; Karlsson, A.;
Wendelbo, R. Angew. Chem. Int. Ed. 1998, 37, 609, with permission of
John Wiley and Sons.
structure in a synthesis product that contains several crystalline
structures and still amorphous material256 (Figure 20, enclos-
ing the data set used in the discovery program of ITQ-33).256
UOP researchers have used the HT methods to study the
combination of commercially available templates with the aim
of reducing the cost of potential new zeolites,257 and discov-
ered in this way zeolites UZM-4, UZM-8, UZM-15, UZM-17, or
UZM-22, among others.258,259
The application of the combinatorial synthesis approach
has led to the discovery of many of the interesting zeolites
already described in the previous sections, such as ITQ-32,145
ITQ-33,17 ITQ-37,83 ITQ-40,35 ITQ-44,36 or ITQ-43.37
Nowadays, most of the HT experimentation applied to
zeolite synthesis combines the automated parallel synthesis
of the solids, their parallel physicochemical characterization,
in some cases fast sequential testing of some of their most
interesting properties, and the use of data-mining techniques
to maximize the information acquired.260263 It has proved to
be a powerful technique in zeolite synthesis research, provided
that clear and founded starting hypothesis is made, and it is
expected to be increasingly implemented in most laboratories
working in this area. It has to be remarked that the use of HT
strategies has to be based on a scientific approach which will be
now facilitated by the production of a larger number of
results,264266 but it should never be considered as the substi-
tute for a well-planned strategy.251,267,268 Besides the discovery
of new materials or achieving a better understanding of the
zeolite crystallization mechanisms, it allows the exploration of
synthesis conditions in nonconventional ranges, or finding
alternative and/or improved synthesis methods with reduced
costs for existing zeolites.
5.05.5.8 Other Approaches
Important advances have been achieved recently in zeolite syn-
thesis, mainly focused on cost reduction of the production pro-
cess of existing zeolites and on facilitating the industrial
application of new zeolite materials. Some approaches have
already been discussed in the previous sections, such as the low
effluent or dry hydrothermal synthesis (Section 5.05.4.4), the use
of degradable SDAs and pore-filling agents (Section 5.05.4.3), or
the synthesis OSDA-free zeolites (Section 5.05.5.5). The micro-
wave zeolite synthesis269,270 has been proposed as an energy-
saving and synthesis-time-saving procedure, thanks to the use of
microwave heating. For more information, the reader is directed
to a recent revision by Yu.91
The conversion of kaolin clay into type A zeolites was
already claimed by researchers of Minerals & Chemicals
Philipp Corporation in the early 1960s.271 This approach,
although not new, has been further developed by Engelhard,
and is highly interesting from an applied point of view. Thus,
in 1985, the in situ technique for the production of fluid
catalytic cracking (FCC) catalysts was described for the first
time.272 This process is one of the two general manufacturing
processes of FCC catalysts, together with the incorporation
technique, and it is characterized by the in situ growth of
the zeolite component within preformed kaolin-based micro-
spheres. The advantages of this method are the possibility of a
larger proportion of zeolite within the final catalyst particle, an
intimate contact of zeolite and matrix, and better accessibility
 Zeolites 121

Automatic
10 20 30 2Theta identification of
crystallographic
Si/Ge = 2 Si/Ge = 4 Si/Ge = 15 Si/Ge = 30 phases from
powder x-ray
B/TIV B/TIV B/TIV B/TIV
diffraction
0 0.005 0.02 0.05 0 0.005 0.02 0.05 0 0.005 0.02 0.05 0 0.005 0.02 0.05
5
13.3
OH/TIV = 0.1 H2O/TIV 21.7

30

5
13.3
OH/TIV = 0.3 H2O/TIV
21.7
30

5
13.3
OH/TIV = 0.5 H2O/TIV
21.7
30

Amorphous ITQ-24 SSZ-31 ITQ-24-ITQ-17 Lam-ITQ-24

ITQ-22-ITQ-24
ITQ-22 Lamellar EU-1 ITQ-22-ITQ-24-ITQ-33 ITQ-33-ITQ-24

Figure 20 The automatic analysis of the 192 sets of x-ray diffraction data in the discovery program of zeolite ITQ-33 is expected to identify the different
crystallographic phases present in each sample, allowing the creation of the corresponding phase diagram. This contains eight different phases and
numerous mixtures from which the presence of a new material has to be established. Reproduced from Baumes, L. A.; Moliner, M.; Corma, A. Chem.
Eur. J. 2009, 15(17), 42584269, with permission of John Wiley and Sons.

to the active sites. The method has been further improved for
processing petroleum residues.273275 A similar approach,
assigned to Lummus Technology, is the transformation of an
appropriate binder, mesoporous silicaalumina, into a zeolite
via a secondary synthesis.276,277 This methodology, known as
zeolite-bound zeolites, has been demonstrated for some of the
most demanded zeolites in the catalysts and adsorbents indus-
try, for example, beta, ZSM-5, mordenite, X, Y, and A.278
Finally, an interesting option is the low-cost production of
zeolites using waste fly ash or biomass ash as silica and/or
alumina sources. Waste ashes are byproducts generated in
large amounts. Considering only fly ash, the current annual
production worldwide is estimated around 500 million tons,
which is 7580% of the total ash produced. Thus, different
alternatives are being studied for recycling this product.279
Their conversion into low-grade zeolites is one of them.
These ashes, although differing in composition according to
their origin and combustion conditions,280,281 are all rich in
silica and alumina, which are present as crystalline and/or
amorphous phases, and may also contain Na. Thus, their
122 Zeolites

composition is similar to some volcanic material, known to Table 6 Zeolite consumption and supply, on anhydrous basis
originate the crystallization of natural zeolites, and this was the
main reason for starting research on zeolite formation from fly Zeolite 2001 2004 2004 share (%) Growth
consumption/ (%)
ash.279 The general procedure for synthesizing zeolites starting Volume Value
supply (kta)
with these precursors is by dissolution of the ash phase with
alkaline solutions (mainly NaOH or KOH) and the consequent Synthetic zeolites consumption
precipitation of the crystalline zeolite material; however, two Detergent builder 1188 1325 78.2 56 11.5
different methods have been described,279 the direct synthesis, Catalysts 184 241 14.2 27 30.9
where fly ash is directly treated hydrothermally with an alka- Adsorbents/ 92 126 7.4 17 36.9
line solution, and the nondirect synthesis, which includes a desiccants
first step where silica and alumina are extracted from the ash Total 1464 1692 15.6
Synthetic zeolites 1926 2029 5.3
with a hot alkaline solution, and these extracts are used as the
capacity
starting solution for obtaining the zeolite. Following these
Natural zeolites >3100 >2500 19.3
procedures, single phases of zeolites A, X, Y, P1, and ZSM-5 consumption
have been described. However, some important drawbacks Natural zeolites 3293 2557 22.3
have to be considered. Dissolution of the Si and Al precursors productions
and crystallization of the zeolite take place simultaneously, so
the formation of zeolite crystals on the ash particles can limit Adapted from Vermeiren, W.; Gilson, J.-P. Top. Catal. 2009, 52, 31, with permission
further reaction, inhibiting the complete conversion of ash of Springer.
into the crystalline zeolite. Moreover, fly ash is usually pro-
cessed without prior purification, and remaining coal and
impurities such as Fe2O3 may contaminate the final zeolitic Besides that, the cation-exhange properties of zeolites can also be
product. useful in terms of catalytic applications. Indeed, the substitution
of the original cations by protons or transition metals results
in the introduction of active sites for acid-catalyzed or redox
5.05.6 Zeolite Properties processes.
Besides Si and Al, we have seen in the former sections
One of the main characteristics of zeolites is their pore topol- that other elements (di-, tri-, or tetravalent) can also be intro-
ogy, with channels and/or cages of molecular dimensions that duced in tetrahedral coordination within the crystalline
will allow discrimination among molecules of different sizes framework of zeolites. This possibility paves the way for
and shapes. This effect is responsible for one of their first introducing potential catalytic active sites in framework
applications, that is, the use as molecular sieves in separation position of the zeolites. Then, the combination of cation-
processes. Zeolites are commercially applied in permanent exchange properties and the possibility of introducing a variety
gas and hydrocarbon separation processes, as we will see in of framework cations that can act as Lewis acids enlarges
Section 5.05.6.1. Moreover, the intracrystalline porosity is also the possibilities of zeolites for catalysis, as will be seen in
decisive for catalytic applications by introducing shape- Section 5.05.6.3.
selectivity effects for diffusion of reactant and products as Table 6 presents the production of zeolites for different
well as for transition state selection. The above-named shape applications. It can be seen there that the largest growth occurs
selectivity can play a decisive role in the observed rate of reaction for uses as catalysts and adsorbents/desiccants. It is also inter-
and product selectivity, as will be shown in Section 5.05.6.3 esting to remark the decrease in the consumption/production
dedicated to this topic. of natural in favor of synthetic zeolites.
Due to the microporous channel systems, zeolites present
very high surface areas and pore volumes, which confer these
materials a high adsorption capacity. This is the basis of their
5.05.6.1 Adsorption and Separation
use as adsorbents, and in molecular separation processes.
Adsorption effects are also important in catalytic applications, Although the microporous structure of zeolites and the term
and, for instance, modification of the hydrophilic/hydrophobic molecular sieve could lead us to conclude that separation in
properties allows discrimination among reactants or reactants zeolites takes place mainly by size exclusion, most of the
and products with different polarities. industrial separation processes take advantage of the large
The presence of tetrahedrally coordinated Al in the zeolite adsorption capacity of these materials. Thus, preferential
framework results in a negative charge, which has to be com- adsorption of polar molecules by Al-rich zeolites is the basis
pensated by cations in extra-framework positions, located, for most commercial gas separation. The first industrial
therefore, inside the channels and/or cavities of the zeolite adsorption processes appeared in the mid-1960s and in the
structure. We have seen in the former sections that in the case 1970s for O2 production, H2 and air purification, and drying
of aluminum-rich zeolites, usually prepared in alkaline media gaseous streams. Historically, natural zeolites have been pro-
and in the absence of OSDAs, these charge-balancing ions are posed for these applications, because of their abundance and
alkaline cations, generally Na, that can be easily replaced low cost.283 However, the presence of impurities, lack of uni-
by other cations by simple ion-exchange procedures. The formity, and the need for processing the natural zeolite-based
ion-exchange property of zeolites is the basis for their use in adsorbents can be limiting factors for their use, and synthetic
detergents, for water softening, as we will see in Section 5.05.6.2. zeolites are now mainly employed.

The success and increasing number of applications of zeo-
lites as adsorbents are due to the following zeolite properties284:
a large number of structures being commercially available;
versatility of commercial zeolites according to their adsorp-
tion characteristics that can be tailored by controlling frame-
work (Si/Al ratio) and extra-framework composition (use of
different cations) as well as by postsynthesis modifications,
with the final result being a fine tuning of adsorption capac-
ity and differentiation;
conformation procedures such as extrusion and bead-
making processes are easily applicable to zeolites, facilitating
their use in packed-bed systems;
high structural stability; and
safety and environmental considerations.
Table 7 shows the main industrial adsorbents based on zeo-
lites and some application examples. There it can be seen that
most of the processes employ Al-rich zeolites, with the exception
of some desulfurization and hydrocarbon separation processes
performed by zeolite HY with SiO2/Al2O3 ratio above 20.
Molecular separations by adsorption have been proved to
be advantageous over other processes such as distillation,
which is energetically more demanding. An interesting over-
view on zeolite-based adsorbents and their main properties has
been given recently by Barret and Stephenson.284
From a practical point of view, different adsorption processes
such as the pressure swing adsorption (PSA), the temperature
swing adsorption (TSA), and the vacuum swing adsorption
(VSA) can be used. More information regarding the above pro-
cesses can be found in the surveys by Yang285 and Kerry.286
Noncryogenic air separation is one of the oldest zeolite-
based adsorption technologies, known since the 1960s, able to
provide oxygen purity above 95%. The adsorbent used is a Li
exchanged X zeolite. An adsorption process is also used for air
pretreatment before its cryogenic separation, in order to
remove moisture, CO2, traces of hydrocarbons, NOx, or SOx.
Purification of H2 produced by steam reforming, purification
of CO2 or He gases, and treatment of flue gas or exhaust gas
streams from fossil-fuel power plants and fuelgas combustion
processes to remove atmospheric pollutants, such as CO2, SOx,
Table 7 Commercial zeolite-based adsorbents and processes
Zeolite type Commercial products SiO2/Al2O3
A 3A 2 K > 70%
4A 2 Na > 99%
5A 2 Ca > 70%
X 13X 22.5
LiX 22.5
AgX 22.5
BaX
CaX 2.52.0
Y NaY 35
HY 201000
Mordenite Small/large port 5 Na
Clinoptilolite TSM-140/ 6 Mixed (Na, K, Ca, Mg)
CS400
a
HC accounts for hydrocarbons.
Adapted from Barret, P. A.; Stephenson, N. A. In Zeolites and Ordered Porous Solids: Fundamentals and Applications; Martinez, C.; Perez-Pariente, J., Eds., Editorial Universitat
Politecnica de Valencia: Valencia, 2011; pp 6790.
Zeolites 123
NOx, H2S, and NH3, are other known applications. Finally,
hydrocarbon streams produced in refinery, petrochemical, and
organic synthesis processes usually require separation of ole-
fins from paraffins, n-paraffins from iso-paraffins, and cyclic
paraffins from aromatics or different aromatic isomers
(xylenes). Some of the processes commercialized for such
applications are SorbexTM and ParexTM, both developed by
UOP, or Petrofin, by BOC.285
Besides those equilibrium and kinetic separations, based on
different capacities and diffusion rates of the adsorbates, gas
separation can be carried out by size exclusion of one gas
within a mixture of more gases by the zeolite pores. This type
of steric separation is performed usually by zeolite membranes,
which are specially suited when processing high flux streams,
as they offer high efficiency with minimum pressure drops.287
These membranes consist of thin zeolite films on ceramic
(mainly alumina) or metal supports. The advantages of zeolite
membranes are temperature stability and solvent resistance
(especially when compared to polymeric membranes), com-
bined with their molecular sieve function. Thus, zeolite mem-
branes can principally discriminate the components of gaseous
or liquid mixtures based on their molecular size. However, this
molecular sieving principle requires the membrane to be abso-
lutely free of pinholes or cracks.
Despite the high consolidation degree of existing zeolite-
based adsorption processes, there is always room for new
zeolite structures with improved separation efficiency. Some
recently synthesized zeolites such as ITQ-12288 and ITQ-32289
have been proposed for propane/propylene and trans-2-butene
and 1-butene separation (Figure 21), and LTA with different
SiO2/Al2O3 ratios has shown to be effective for the separation
of CO2 from methane streams.290
5.05.6.2 Ion Exchange
The presence of tetrahedrally coordinated Al in the zeolite
framework results in a negative charge, which has to be com-
pensated by cations in extra-framework positions, located
inside the channels and/or cavities of the zeolite structure.
Cations Application examples
EtOH dehydration, drying
Desiccant, medical O2, other PSA, natural gas processing
H2PSA, medical O2, VPSA O2, natural gas processing
>99% Na Air prepurification
>88% Li VPSA O2, medical O2
>80% Ag CO, H2 abatement
Ba, K p-Xylene recovery
>70% Ca H2PSA, N2O removal, O2 VPSA
>99% Na Desulfurization,a HC separation
H Desulfurization, HC separation
Desulfurization, HC separation
N2O removal
124 Zeolites

0.0012 which at temperatures close to ambient presents a large hydra-

298 K a tion shell, and hinders its entrance through these small open-
ings. Still, the group of Dyer has reported Mg- and Li-A samples
0.0009 where 10 of the 12 Na cations from the pseudo-unit cell were
replaced.292,293 However, when Mg removal is required, zeolite
Q (mol g-1)

X, with larger pores (0.75 nm), is the most adequate zeolite.

0.0006 When both cations are present, mixtures of the different zeo-
lites are used.
Naturally occurring zeolites are rarely pure and are there-
0.0003 fore often unsuitable for commercial applications. Also for
b
detergents, synthetic zeolite grades are preferred, and have
c been developed to fulfill the role of the product builder.
0.0000
0 100 200 300 400 500 600
t (min) 5.05.6.3 Catalytic Applications

0.0012 Catalysis is a key tool for chemical sustainability. Improved cata-

lytic processes will make a more efficient use of natural resources,
333 K a reduce the amount of byproducts formed, and eliminate contam-
0.0009 inant effluents, besides lowering energy requirements. Thus, cat-
b alytic processes are desirable as they accomplish not only
Q (mol g-1)

economical objectives, but also environmental protection. In

0.0006
0.0003 d ration and producing less salt and waste.296 Among the heteroge-
c neous catalysts, molecular sieves are the most widely used
0.0000
0 400 800 1200 1600 2000
t (min)
Figure 21 Adsorption kinetics of C4 hydrocarbons at 298 and 333 K
and 30.4 kPa in ITQ-32. trans-2-Butene (a), 1-butene (b), butane (c), and
cis-2-butene (d). Reproduced from Palomino, M.; Cantin, A.; Corma, A.;
Leiva, S.; Rey, F.; Valencia, S. Chem. Commun.(Cambridge, U. K.) 2007,
12331235, with permission of The Royal Society of Chemistry.
We have seen in the former sections that in the case of
aluminum-rich zeolites, usually prepared in alkaline media
and in the absence of OSDAs, these charge-balancing ions are
alkaline cations, generally Na, that can be easily replaced by
other cations by simple ion-exchange procedures. The increas-
ing concern on eutrophication issues led to an intensive search
for replacing the phosphates as detergent builders, and zeolites
emerged as particularly promising materials, not only due to
their ion-exchange capacity but also due to their safety and
environmentally benign nature. The main function of a builder
is to soften the wash water by removing the components
responsible for water hardness, present as calcium and magne-
sium compounds. Additionally, in the case of detergent pow-
ders, builders can also absorb liquid components of the
formulation (such as surfactants) in order to maintain dry,
free-flowing powder characteristics.291 Thus, zeolites used in
detergent are characterized by a high aluminum content (Si/Al
ratio of 1 or virtually 1), which in turn results in a maximum
cation-exchange capacity. The zeolites available for detergents
today are zeolite A, zeolite P, and zeolite X, with significantly
different crystalline structures. Depending on the pore size, the
zeolite will selectively remove Ca, Mg, or both. Zeolites A and
P, with small pores and windows (0.42 and 0.3 nm, respec-
tively), are able to remove Ca, but are less suitable for Mg,
fact, more than 90% of all industrial chemicals are being pro-
duced by catalytic processes.294,295 Focusing on heterogeneous
catalysts, they provide additional advantages of facilitating sepa-
materials in commercial applications.297,298 As can be deduced
from Table 6 (Section 5.05.6.1), the zeolite production directed
to catalytic uses is below 20% of the total zeolite market (although
with an increasing growth trend), with the rest being directed to
detergents (70%) and adsorbents (10%).282,299 However, cata-
lytic applications account for the largest market value, especially
in the oil-refining sector. Most of the current large-scale commer-
cial processes using zeolite-based catalysts are in the petroleum
refining and petrochemical industry,300 but their implementation
in the chemical industry, environmental applications, non-
conventional conversion processes of gas,299 oil, and coal, bio-
mass transformation,301 and car exhaust treatment302 is
increasing rapidly.303
As we have been learning from the previous sections,
besides being ecological and safe, zeolite materials combine a
well-defined microporous structure, and pore sizes in the range
of molecular dimensions, with a flexible chemical composition
and an outstanding thermal and hydrothermal stability.298 The
channels, channel intersections, or cages of the zeolite frame-
work, where the reaction will take place, can be considered as
nanoreactors. Thus, the reaction will be affected not only by
the active site, but also by the shape, size, windows, and
interconnections of pores or cages where the active site is
located. One of the first consequences of molecules diffusing
through the zeolite micropores is the so-called confinement
effect, due to electronic interactions of the guest molecule with
the zeolite host when the size of the molecules approaches that
of the pores. Among other effects, these interactions result in
an increased reactivity of the confined reactant.298 The shape-
selectivity effects are also directly related to the zeolite struc-
ture, and are key for some important commercial applications
of zeolites as heterogeneous catalysts. Indeed, the dimension
of the zeolite pores is responsible for the reactants and product
shape selectivity, or mass transport discrimination among

reactants and products of different sizes due to a true molecular
sieve effect.304 Transition-state selectivity occurs when the
geometry of the pore stabilizes one transition state among
several possible ones. An overview on shape selectivity applied
to petroleum and petrochemical applications has been given
some years ago by Degnan.305
The advantages of shape selectivity are lost when processing
bulkier reactants. This has directed research to possible routes
for increasing accessibility to the active sites, and for reducing
diffusional problems, which may be directly related to catalyst
life. Some approaches are the synthesis of extra-large-pore
zeolites, which has already been discussed in Section 5.05.5.3,
the synthesis of nanocrystalline zeolites, zeolite delamination
(Section 5.05.5.6), and postsynthesis modification of zeolites
for generation of additional mesoporosity.306310
5.05.6.3.1 Reactions catalyzed by acid and basic sites
Brnsted acid sites are generated on the surfaces of zeolites
when Si4 is isomorphically replaced by a trivalent metal cat-
ion such as Al3. This substitution creates a negative charge in
the lattice that can be compensated by a proton. From a struc-
tural point of view, the Brnsted acid site in a zeolite can be
seen as a resonance hybrid of structures I and II (Figure 22)
where structure I is a fully bridged oxygen with a weakly
bonded proton and structure II is a silanol group with a weak
Lewis acid interaction of the hydroxyl oxygen with an Al.
Some of the most important commercial processes, such as
FCC, hydrocracking, aromatic alkylation, isomerization of short-
chain alkanes and C8 alkylaromatics, or disproportionation of
toluene, among others, are based on acid zeolite catalysts.303 The
particular case of the FCC application, which converts heavy
fractions containing bulky hydrocarbons into lighter com-
pounds within the ranges of diesel, gasoline, or liquefied petro-
leum gases (LPGs, mainly propylene and butenes), represents a
good example of the constant evolution of a zeolite-based cata-
lyst with the aim of adapting it to variation of the process
specification.282,311,312 In the early 1960s, zeolite Y, just patented
by Union Carbide,313 substituted amorphous silicaalumina as
the active component of catalytic cracking catalysts, and nowa-
days it is still being employed. The introduction of zeolite Y as a
component in FCC catalysts was an important breakthrough
concerning the evolution of this process, as it was more active
and more selective to gasoline than its predecessor. However, this
3D large-pore zeolite was originally synthesized with Si/Al con-
tents as low as 3 and, when in the acid form (HY), it presents a
high acid site density, which results in low-quality products, low
hydrothermal stability, and fast catalyst deactivation. This led
research to improvement of stability and selectivity of Y zeolite.
To do that, a series of modifications were introduced with time,
which are summarized below:
H H
O O
Si Al Si Al+
I II
Figure 22 Brnsted acid site in a zeolite seen as a resonance hybrid
of structures I and II.
Zeolites 125
Introduction of rare earths (RE La, Ce) by ion exchange. In
this way, the zeolite was stabilized toward dealumination,
and total acidity was reduced, improving the quality of the
products.
Ultrastabilization of zeolite Y, which consists of controlled
processes of dealumination of the zeolites framework with
migration of the silicon atoms into defect sites and generation
of additional mesoporosity. These treatments increase the
framework Si/Al content of the zeolite, stabilizing its structure
toward undesired (uncontrolled) dealumination during the
catalyst regeneration process that occurs at high temperatures
and in the presence of steam. Accessibility of the reactant
molecules to the active sites is also increased, because of the
mesopores generated in the zeolite crystallites.
Combination of the former by incorporation of RE into the
ultrastabilized Y (USY) zeolite.
Removal of RE: because of the present high cost of RE, catalyst
manufacturers are searching for new ways to improve stability
avoiding the use of RE or using them in smaller amounts.
Many attempts have been performed for replacing zeolite Y
as main active component of the FCC catalyst by other new
zeolites, especially by others with larger pores, such as ZSM-
20,314,315 Beta,316 ITQ-21,136,317 or ITQ-33.318 However, zeo-
lite cost is too high compared with Y zeolite and differences in
the catalytic performance do not compensate for the extra
manufacturing cost.
Besides improving zeolite Y, other zeolites have been incor-
porated as specific additives into the final formulation. This is the
case of medium-pore zeolite ZSM-5, which selectively cracks
olefins and some alkanes, belonging to the gasoline fraction,
into light olefins of the LPG range. Although the total yield to
gasoline is reduced in this way, the removal of the linear olefins
and some alkanes, while concentrating the aromatics, improves
the octane of the gasoline fraction while increasing propylene
yield. Other medium-pore zeolites studied as FCC additives are
MCM-22,319 ITQ-13,137 IM-5,320 and TNU-9.321 Large-pore
BETA316,322 and ITQ-7323 have also been proposed for increasing
mainly the selectivity to butenes, isobutane or isobutene, and
amylenes.
The use of acid zeolites is not limited to refining and pet-
rochemistry; in the last decades, an important effort has been
made to introduce these heterogeneous acid materials in the
production of chemicals and fine chemicals. For more infor-
mation on the subject, the reader is directed to several recent
revisions.303,324,325
Basic sites may also be generated within the pores of
zeolites by modifying the nature of extra-framework cations.
Thus, alkaline-exchanged zeolites present (weak) basicity and
are able to abstract protons in organic molecules with pKa of
10.7. The basicity of the framework oxygens may be increased
by partial substitution of Si by Ge.326 In this way, it will be
possible to take advantage of zeolite properties, also in base
catalysis. The basic sites in zeolites are of Lewis type and
correspond to framework oxygens, with the basicity of a
given oxygen being related to the density of negative charge.
Concerning heterogeneous basic process, they account for only
8% of the total industrial processes catalyzed by bases and, as
far as we know, none of them is performed with basic zeolite
catalysts.327,328 An important hurdle for the industrial
126 Zeolites
application of basic zeolites relies on the fact that the currently
used catalysts have to compete with the inexpensive NaOH and
KOH. Their low cost and the easy processing of the residues
formed reduce the possibility of zeolite application to cases
where special-selectivity effects will be needed.
5.05.6.3.2 Redox reactions
Several liquid-phase oxidation processes are catalyzed by solu-
ble oxometallic compounds. Their substitution by zeolite-
based oxidation catalyst presents several advantages, beyond.
the benefits of working with a solid catalyst. The zeolite struc-
ture contributes with steric and confinement effects, and the
incorporation of the desired metal atom in the framework (Ti,
Fe, and Sn) results in well-dispersed active sites, generally
stable toward leaching. Moreover, their tunable adsorption
properties allow varying the hydrophobicityhydrophilicity
of the zeolite, enabling discrimination among reactants with
different polarities, which is a very useful tool when oxidizing
organic hydrocarbons in the presence of aqueous H2O2 as the
oxidation agent.326
Titanium-silicalite (TS-1, developed by Eni S.p.A.166) is, at
present, the most employed zeolite-based oxidation
catalyst.329 This Al-free MFI zeolite, which has the titanium
atoms in tetrahedral framework positions, is commercially
used for olefin epoxidation, hydroxylation of phenols and
ammoximation of ketones, or oxidation of sulfur compounds
and ethers, using hydrogen peroxide as the oxidant. More
details and additional examples of zeolite-based redox pro-
cesses can be found in303 and in a recent review by Tatsumi.330
Ti-containing large-pore zeolites have also been described as
oxidation catalysts for conversion of bulkier molecules that
cannot diffuse through the medium pores of TS-1. Some exam-
ples are Ti-Beta,167,331 Ti-ZSM-12,332 Ti-ITQ-7,333 and Ti-MCM-
22.168,334 Table 8 shows, in the case of Ti-Beta, the importance
of zeolite polarity for oxidation processes involving H2O2.
Selective redox catalyst can also be obtained by introducing
metals in extra-framework positions. This is the case of copper
(also cobalt or iron)-containing zeolites used for selective cat-
alytic reduction (SCR) of nitrogen oxides present in emissions
from stationary (thermal power plants) and mobile (vehicle
emissions) sources. Cu-ZSM-5 has been the most described
Table 8 Influence of Al content and zeolite polarity on activity and
selectivity of Ti-Beta for epoxidation of 1-hexene with H2O2 using
methanol as solvent
Zeolite Chemical TOF (mol/ Epoxide
characteristics composition mol Ti h) selectivity (%)
Si/Al TiO2
ratio (wt%)
Framework 300 4.7 20.8 25.9
Al defects
No Al but defects 1 2.5 28.6 75.4
present
No Al, no defects 1 2.5 32.2 96.4
(F synthesis)
Adapted from Camblor, M. A.; Corma, A.; Lightfoot, P.; Villaescusa, L. A.; Wright, P. A.
Angew. Chem. Int. Ed. Engl. 1997, 36, 26592661, with permission of John Wiley
and Sons.
zeolite,335 although other structures have also been claimed
such as ZSM-11336 or ZSM-12,337 among others. However,
they all share an important drawback concerning commercial
application of zeolite-based SCR catalysts, their reduced long-
term hydrothermal stability.338340 This fact has directed
researchers to the use of more stable small-pore zeolites (8R),
such as SSZ-13 and SAPO-34. In these cases, Co is the preferred
metal when NH3 is used as the reducing agent. Avoiding
agglomeration of the metal during the catalyst use is a key
objective in this type of catalysts. More information on SCR
with zeolites can be found in recent reviews.302,303,341
5.05.6.3.3 Zeolite-based multifunctional catalysts
It is possible to combine the intrinsic active functions of
zeolite-based catalysts, for example, Brnsted acidity, with
other catalytic functions to prepare bifunctional catalysts.
For instance, the incorporation of metals such as Ni, Pd, or
Pt on acid zeolites introduces an additional hydrogenation
dehydrogenation function to the acidity of the zeolite that
allows preparing bifunctional acid/hydrogenatingdehydro-
genating catalysts useful for a series of processes such as
short-chain alkane isomerization or hydrogenation and ring
opening of polyaromatic compounds. Metal sulfides of groups
VIIIB (Mo, W) and VIB (Co, Ni) on acid USY zeolite are
extensively applied in commercial hydrocracking operation.
These catalysts, besides reducing the content of sulfur and
nitrogen of the feed processed by the metal sulfides, convert
the original heavy hydrocarbons into higher-value compounds
within the diesel or gasoline fractions. As we saw in the former
section, it is possible to tune the activity of zeolite Y by con-
trolled ultrastabilization. In this way, the higher the framework
Si/Al ratio, the higher the selectivity to diesel. More details on
the processes using multifunctional zeolite-based catalysts can
be found elsewhere.303,326,342
5.05.7 Conclusion
In the previous sections, we have summarized different aspects
of zeolite chemistry that involve from synthesis to characteri-
zation and final applications.
Although many zeolite structures and compositions have
been prepared, the possibility to synthesize tailor-made zeo-
lites still remains a challenge, not only from the point of view
of their porous structure, but also considering the distribution
and location of their active sites. The hydrothermal stability of
large- and extra-large-pore zeolites is also a key factor, espe-
cially regarding their potential commercial application in fields
such as catalytic cracking or hydrocracking. The field with the
largest expansion, and also with the largest environmental and
economical impact, is the use of zeolites as heterogeneous
catalysts. In this sense, all zeolite structures, that is, small,
medium, large, and extra-large zeolites, can offer interesting
opportunities. For instance, small-pore zeolites, besides their
potential applications for CO2, N2, and CH4 separation, are
promising materials for NOx reduction in car exhaust.
Medium-pore zeolites are useful in many catalytic processes
in the fields of oil refining, petrochemistry, and chemicals and
fine chemicals. Small variations in pore dimensions can still
improve those processes. Nevertheless, any new structure should

be stable enough and with a cost low enough to be competitive.
The same applies for large- and extra-large-pore zeolites, for
which stability and cost are the determinant factors. One should
also consider zeolites with a hierarchical system of pores (med-
ium  large, medium  extra-large, large  extra-large) that can
introduce new selectivity issues. Finally, zeolites in the form of
single or double layers have already demonstrated their possi-
bilities for converting very large molecules.246,247,343346 Again,
to achieve the synthesis with inexpensive or recyclable OSDA, or
even template-free, will further boost the use of these lamellar
zeolites in industry. For related chapters in this Comprehensive,
we refer to Chapters 4.05, 5.06, and 7.10.
Acknowledgment
The authors acknowledge financial support by Consolider-Ingenio
2010 (proyecto MULTICAT) and Subprograma de apoyo a Centros
y Universidades de Excelencia Severo Ochoa (proyecto SEV 2012
0267) from Ministerio de Economay Competitividad.
References
1. Virta, R. L. Zeolites (natural). U.S. Geological Survey, Mineral Commodity Summaries.
http://minerals.usgs.gov/minerals/pubs/commodity/zeolites/mcs-2011-zeoli.pdf
(accessed 2011).
2. Misaelides, P. Microporous Mesoporous Mater. 2011, 144(1), 1518.
3. Apreutesei, R. E.; Catrinescu, C.; Teodosiu, C. Environ. Eng. Manage. J. 2008,
(Copyright (C) 2011 American Chemical Society (ACS). All Rights Reserved.), 7,
149161.
4. Colella, C. Stud. Surf. Sci. Catal. 2007, 168, 9991035.
5. Roque-Malherbe, R. Applications of Natural Zeolites in Pollution Abatement and
Industry. Academic Press: New York, 2001; pp 495522.
6. Pavelic, K.; Subotic, B.; Colic, M. Stud. Surf. Sci. Catal. 2001, 135, 52515258.
7. International Zeolite Asociation. http://www.iza-online.org/ (accessed 2011).
8. Brunner, G. O.; Meier, W. M. Nature 1989, 337, 146.
9. Jiang, J.; Yu, J.; Corma, A. Angew. Chem. Int. Ed. 2010, 49(18), 31203145.
10. Baerlocher, C.; McCusker, L. B.; Olson, D. H. Atlas of Zeolite Framework Types,
6th ed.; Elsevier B.V: Amsterdam, 2007.
11. Lowenstein, W. Am. Mineral. 1954, 39, 92.
12. Wright, P. A.; Lozinska, M. In Zeolites and Ordered Porous Solids: Fundamentals
and Applications; Martinez, C., Perez-Pariente, J., Eds.; Editorial Universitat
Politecnica de Valencia: Valencia, 2011; pp 6790.
13. McCusker, L. B.; Liebau, F.; Engelhardt, G. Pure Appl. Chem. 2001, 73(2),
(IUPAC Recommendations 2001), 381394.
14. Davis, M. E. Ind. Eng. Chem. Res. 1991, 30, 16751683.
15. McCusker, L. B.; Baerlocher, C. In Studies in Surface Science and Catalysis;
van Bekkum, H., Flanige, E. M., Jacobs, P. A., Jansen, J. C., Eds.; Elsevier:
Amsterdam, 2001; Vol. 137, pp 3767.
16. Foster, M. D.; Treacy, M. M. J. http://www.hypotheticalzeolites.net.
17. Corma, A.; Diaz-Cabanas, M. J.; Jorda, J. L.; Martinez, C.; Moliner, M. Nature
2006, 443, 842845.
18. Earl, D. J.; Deem, M. W. In Abstracts of Papers, 237th ACS National Meeting, Salt
Lake City, UT, March 2226, 2009; pp COMP-067.
19. Bell, R. G.; Foster, M. D.; Simperler, A.; Klinowski, J. Stud. Surf. Sci. Catal. 2004,
154B, 12221229.
20. Earl, D. J.; Deem, M. W. Ind. Eng. Chem. Res. 2006, 45, 54495454.
21. Li, J.; Yu, J.; Sun, J.; Dong, X.; Li, Y.; Wang, Z.; Wang, S.; Xu, R. Stud. Surf. Sci.
Catal. 2007, 170A, 168176.
22. Li, Y.; Yu, J.; Wang, Z.; Zhang, J.; Guo, M.; Xu, R. Chem. Mater. 2005, 17,
43994405.
23. Wells, S. A.; Foster, M. D.; Treacy, M. M. J. Microporous Mesoporous Mater.
2006, 93, 151157.
24. Derouane, E. G. J. Catal. 1980, 65, 486489.
25. Vaughan, P. A. Acta Cryst. 1966, 21, 983.
26. Meier, W. M. Z. Kristallogr. 1961, 115, 439.
27. Lobo, R. F.; Pan, M.; Chan, I.; Li, H. X.; Medrud, R. C.; Zones, S. I.; Crozier, P. A.;
Davis, M. E. Science 1993, 262, 15431546.
28. Lobo, R. F.; Davis, M. E. J. Am. Chem. Soc. 1995, 117(13), 37663779.
Zeolites 127
29. Paillaud, J. L.; Harbuzaru, B.; Patarin, J.; Bats, N. Science 2004, 304, 990.
30. Corma, A.; Diaz-Cabanas, M. J.; Rey, F.; Nicolopoulus, S.; Boulahya, K. Chem.
Commun. 2004, 13561357.
31. Corma, A.; Rey, F.; Valencia, S.; Jorda, J. L.; Rius, J. Nat. Mater. 2003, 2,
493497.
32. Dorset, D. L.; Weston, S. C.; Dhingra, S. S. J. Phys. Chem. B 2006, 110(5),
20452050.
33. Moliner, M.; Gonzalez, J.; Portilla, M. T.; Willhammar, T.; Rey, F.; Llopis, F. J.;
Zou, X.; Corma, A. J. Am. Chem. Soc. 2011, 133(24), 94979505.
34. Willhammar, T.; Sun, J.; Wan, W.; Oleynikov, P.; Zhang, D.; Zou, X.; Moliner, M.;
Gonzalez, J.; Martnez, C.; Rey, F.; Corma, A. Nat. Chem. 2012, 4, 188194.
35. Corma, A.; Diaz-Cabanas, M. J.; Jiang, J.; Afeworki, M.; Dorset, D. L.;
Soled, S. L.; Strohmaier, K. G. Proc. Natl. Acad. Sci. U. S. A. 2010, 16, 6.
36. Jiang, J.; Jorda, J. L.; Diaz-Cabanas, M. J.; Yu, J.; Corma, A. Angew. Chem. Int.
Ed. 2010, 49, 4986.
37. Jiang, J.; Jorda, J. L.; Yu, J.; Baumes, L. A.; Mugnaioli, E.; Diaz-Cabanas, M. J.;
Kolb, U.; Corma, A. Science 2011, 333(6046), 11311134.
38. Lohse, U.; Altrichter, B.; Fricke, R.; Pilz, W.; Schreier, E.; Garkisch, C.; Jancke, K.
J. Chem. Soc. Faraday Trans. 1997, 93, 505.
39. Tuel, A. Chem. Mater. 1865, 1999, 11.
40. Oberhagemann, U.; Bayat, P.; Marler, B.; Gies, H.; Rius, J. Angew. Chem. Int. Ed.
1996, 35, 2869.
41. Vortmann, S.; Rius, J.; Siegmann, S.; Gies, H. J. Phys. Chem. B 1997, 101,
1292.
42. Vortmann, S.; Rius, J.; Marler, B.; Gies, H. Eur. J. Mineral. 1999, 11, 125.
43. Cundy, C. S.; Cox, P. A. Microporous Mesoporous Mater. 2005, 82, 178.
44. Roth, W. J. Stud. Surf. Sci. Catal. 2007, 168, 221239.
45. Roth, W. J.; Cejka, J. Catal. Sci. Technol. 2011, 1, 4353.
46. Roth, W. J.; Dorset, D. L. Microporous Mesoporous Mater. 2011, 142, 3236.
47. Strohmaier, K. G. In Zeolites and Catalysis, Synthesis, Reactions and
Applications; Cejka, J., Corma, A., Zones, S., Eds.; Wiley-VCH Weinheim, 2010;
pp 5786.
48. Cundy, C. S.; Cox, P. A. Chem. Rev. 2003, 103, 663701.
49. Corma, A. In Proceedings 14th International Zeolite Conference, Stud. Surf. Sci.
Catal.; Steen, E. V., Claeys, I. M., Eds.; Elsevier B.V: Amsterdam, 2004 falta el
volumen; pp 2540.
50. Gies, H. Chem. Mater. 2009, 21, 43394346.
51. Jeong, H.-K.; Nair, S.; Vogt, T.; Dickinson, L. C.; Tsapatsis, M. Nat. Mater. 2003,
2(1), 5358.
52. Corma, A.; Fornes, V.; Pergher, S. B.; Maesen, T.; Buglass, J. G. Nature 1998,
396(6709), 353356.
53. Corma, A.; Fornes, V.; Guil, J. M.; Pergher, S.; Maesen, T. L. M.; Buglass, J. G.
Microporous Mesoporous Mater. 2000, 38(23), 301309.
54. Corma, A.; Diaz, U.; Domine, M. E.; Fornes, V. Angew. Chem. Int. Ed. 2000,
39(8), 14991501.
55. Corma, A.; Diaz, U.; Domine, M. E.; Fornes, V. J. Am. Chem. Soc. 2000, 122(12),
28042809.
56. Corma, A.; Fornes, V.; Diaz, U. Chem. Commun. 2001, 24, 26422643.
57. Corma Canos, A.; Daz Morales, U.; Fornes Segui, V. Microporous Crystalline
Zeolite Material (ITQ-20) and Production Method Thereof. US7008611, 2006.
58. Corma, A.; Daz, U.; Garca, T.; Sastre, G.; Velty, A. J. Am. Chem. Soc. 2010, 132,
1501115021.
59. Vaughan, D. E. W.; Lussier, J.; Magee, J. S. US4176090, 1979.
60. Pinnavaia, T. J. Science 1983, 220, 365371.
61. Landis, M. E.; Aufdembrink, B. A.; Chu, P.; Johnson, I. D.; Kirker, G. W.;
Rubin, M. K. J. Am. Chem. Soc. 1991, 113, 31893190.
62. Rubin, M. K.; Chu, P. US4954325, 1990.
63. Kresge, C. T.; Roth, W. J.; Simmons, K. G.; Vartuli, J. C. US5229341, 1993.
64. Schreyeck, L.; Caullet, P.; Mougenel, J. C.; Guth, J. L.; Marler, B. Microporous
Mater. 1996, 6(56), 259271.
65. Chica, A.; Corma, A.; Fornes, V.; Daz, U. US6555090, 2003.
66. Zanardi, S.; Alberti, A.; Cruciani, G.; Corma, A.; Fornes, V.; Brunelli, M. Angew.
Chem. Int. Ed. 2004, 43, 49334937.
67. Zones, S. I.; Hwang, S.-J.; Davis, M. E. Chem. Eur. J. 2001, 7, 19902001.
68. Millini, R.; Perego, G.; Parker, W. O., Jr.; Bellussi, G.; Carluccio, L. Microporous
Mater. 1995, 4(23), 221230.
69. Marler, B.; Camblor, M. A.; Gies, H. Microporous Mesoporous Mater. 2006, 90,
87101.
70. Moteki, T.; Chaikittisilp, W.; Shimojima, A.; Okubo, T. J. Am. Chem. Soc. 2008,
130(47), 1578015781.
71. Marler, B.; Stroter, N.; Gies, H. Microporous Mesoporous Mater. 2005, 83(13),
201211.
72. Wang, Y. X.; Gies, H.; Lin, J. H. Chem. Mater. 2007, 19(17), 41814188.
128 Zeolites
73. Dhingra, S.; Kresge, C. T.; Casmer, S. G. US6869587, 2005.
74. Ikeda, T.; Akiyama, Y.; Oumi, Y.; Kawai, A.; Mizukami, F. Angew. Chem. Int. Ed.
2004, 43(37), 48924896.
75. Lewis, G. J.; Knight, L. M.; Miller, M. A.; Wilson, S. T. US2005/0065016, 2005.
76. Kiyozumi, Y.; Ikeda, T.; Hasegawa, Y.; Nagase, T.; Mizukami, F. Chem. Lett. 2006,
35(6), 672.
77. Wheatley, P. S.; Morris, R. E. J. Mater. Chem. 2006, 16, 10351037.
78. Vistad, . B.; Akporiaye, D. E.; Lillerud, K. P. J. Phys. Chem. B 2001, 105(50),
1243712447.
79. Choi, M.; Na, K.; Kim, J.; Sakamoto, Y.; Terasaki, O.; Ryoo, R. Nature 2009,
461, 246249.
80. McCusker, L. B.; Baerlocher, C. In Studies in Surface Science and Catalysis;
Antoine Gedeon, P. M., Florence, B., Eds.; Elsevier: Amsterdam, 2008; Vol. 174
(Part A), pp 312.
81. Gies, H.; Marler, B. In Zeolites and Ordered Porous Solids: Fundamentals and
Applications; Martinez, C., Perez-Pariente, J., Eds.; Editorial Universitat
Politecnica de Valencia: Valencia, 2011; pp 6790.
82. Hong, S. B.; Min, H.-K.; Shin, C.-H.; Cox, P. A.; Warrender, S. J.; Wright, P. A.
J. Am. Chem. Soc. 2007, 129, 1087010885.
83. Sun, J.; Bonneau, C.; Cantin, A.; Corma, A.; Diaz-Cabanas, M. J.; Moliner, M.;
Zhang, D.; Li, M.; Zou, X. Nature 2009, 458, 11541157.
84. Baerlocher, C.; Gramm, F.; Massuger, L.; McCusker, L. B.; He, Z. B.;
Hovmoller, S.; Zou, X. D. Science 2007, 315, 1113.
85. Kolb, U.; Gorelik, T.; Kubel, C.; Otten, M. T.; Hubert, D. Ultramicroscopy 2007,
107, 507.
86. Liu, Y.; Pinnavaia, T. J. J. Mater. Chem. 2002, 12, 3179.
87. Barrer, R. M.; Denny, P. J. J. Chem. Soc. 1961, 971982.
88. Reed, T. B.; Breck, D. W. J. Am. Chem. Soc. 1956, 78, 59725977.
89. Milton, R. M. US2882244, 1959.
90. Flanigen, E. M. In Studies in Surface Science and Catalysis; van Bekkum, H.,
Flanige, E. M., Jacobs, P. A., Jansen, J. C., Eds.; Elsevier: Amsterdam, 2001;
Vol. 137, pp 1135.
91. Yu, J. In Studies in Surface Science and Catalysis; Cejka, J., van Bekkum, H.,
Corma, A., Schuth, F., Eds.; Elsevier: Amsterdam, 2007; Vol. 168, pp
39103.
92. Corma, A. In Studies in Surface Science and Catalysis; van Steen, E., Claeys, M.,
Callanan, L. H., Eds.; Elsevier: Amsterdam, 2004; Vol. 154 (Part A), pp 2540.
93. Moliner, M. In Zeolites and Ordered Porous Solids: Fundamentals and
Applications; Martinez, C., Perez-Pariente, J., Eds.; Editorial Universitat
Politecnica de Valencia: Valencia, 2011; pp 3766.
94. Serrano, D. P.; van, G. R. J. Mater. Chem. 2001, 11, 23912407.
95. Flanigen, E. M.; Breck, D. W. In 137th Meeting of the ACS, Division of Inorganic
Chemistry, Cleveland (OH), 1960.
96. Kerr, G. T. J. Phys. Chem. 1966, 70, 1047.
97. Zhdanov, S. P. In Adv. Chem. Ser.; Flanigen, E. M., Sand, L. B., Eds.; ACS
Publishers: Washington, DC, 1971; Vol. 101, pp 2043.
98. Derouane, E. G.; Detremmerie, S.; Gabelica, Z.; Blom, N. Appl. Catal. 1981,
1, 201.
99. Gabelica, Z.; Blom, N.; Derouane, E. G. Appl. Catal. 1983, 5, 227.
100. Chang, C. D.; Bell, A. T. Catal. Lett. 1991, 8, 305.
101. Burkett, S. L.; Davis, M. E. Chem. Mater. 1995, 7, 920.
102. Burkett, S. L.; Davis, M. E. J. Phys. Chem. 1994, 98, 4647.
103. Ravishankar, R.; Kirschhock, C.; Schoeman, B. J.; Vanoppen, P.; Grobet, P. J.;
Storck, S.; Maier, W. F.; Martens, J. A.; Schryver, F. C. D.; Jacobs, P. A. J. Phys.
Chem. B 1998, 102, 2633.
104. Kirschhock, C. E. A.; Buschmann, V.; Kremer, S.; Ravishankar, R.;
Houssin, C. J. Y.; Mojet, B. L.; van Santen, R. A.; Grobet, P. J.; Jacobs, P. A.;
Martens, J. A. Angew. Chem. Int. Ed. 2001, 40(14), 26372640.
105. Guth, J. L.; Caullet, P. J. Chim. Phys. 1986, 83, 155.
106. Guth, J. L.; Kessler, H. In Catalysis and Zeolites; Weitkamp, J., Puppe, L., Eds.;
Springer: Berlin, 1999; p 1.
107. Vladimiros, N. Curr. Opin. Colloid Interface Sci. 2005, 10(56), 203210.
108. Davis, T. M.; Drews, T. O.; Ramanan, H.; He, C.; Dong, J.; Schnablegger, H.;
Katsoulakis, M. A.; Kokkoli, E.; McCormick, A. V.; Penn, R. L.; Tsapatsis, M. Nat.
Mater. 2006, 5(5), 400408.
109. Aerts, A.; Haouas, M.; Caremans, T. P.; Follens, L. R. A.; van Erp, T. S.;
Taulelle, F.; Vermant, J.; Martens, J. A.; Kirschhock, C. E. A. Chem.-Eur. J. 2010,
16(9), 27642774.
110. de Moor, P.-P. E. A.; Beelen, T. P. M.; Komanschek, B. U.; Beck, L. W.;
Wagner, P.; Davis, M. E.; van Santen, R. A. Chem.-Eur. J. 1999, 5(7),
20832088.
111. Aerts, A.; Kirschhock, C. E. A.; Martens, J. A. Chem. Soc. Rev. 2010, 39,
46264642.
112. Cubillas, P.; Anderson, M. W. In Zeolites and Catalysis, Synthesis, Reactions and
Applications; Cejka, J., Corma, A., Zones, S., Eds.; Wiley-VCH: Weinhem, 2010;
pp 87106.
113. Barrer, R. M., J. Chem. Soc. 1948, 127132.
114. Barrer, R. M.; Hinds, L.; White, E. A. J. Chem. Soc. 1953, 14661475.
115. Milton, R. M. ACS Symp. Ser. 1989, 398, 1.
116. Pauling, L. Z. Kristallogr. 1930, 74, 213225.
117. Dent, L. S.; Smith, J. V. Nature 1958, 181, 17941796.
118. Davis, M. E.; Lobo, R. F. Chem. Mater. 1992, 4, 756768.
119. Dutta, P. K.; Puri, M. M.; Shieh, D. C. Mater. Res. Soc. Symp. Proc. 1988,
Vol. 111, p 101.
120. Flanigen, E. M.; Patton, R. L. US4073864, 1978.
121. Guth, J. L.; Kessler, H.; Higel, J. M.; Lamblin, J. M.; Patarin, J.; Seive, A.;
Chezcau, J. M.; Wey, R. ACS Symp. Ser. 1989, 398, 176.
122. Caullet, P.; Paillaud, J. L.; Simon-Masseron, A.; Soulard, M.; Patarin, J. C. R.
Chim. 2005, 8, 245266.
123. Zones, S. I.; Darton, R. J.; Morris, R.; Hwang, S.-J. J. Phys. Chem. B 2005, 109,
652661.
124. Zones, S. I.; Hwang, S.-J.; Elomari, S.; Ogino, I.; Davis, M. E.; Burton, A. W. C. R.
Chim. 2005, 8, 267282.
125. Blasco, T.; Camblor, M. A.; Corma, A.; Esteve, P.; Guil, J. M.; Martinez, A.;
Perdigon-Melon, J. A.; Valencia, S. J. Phys. Chem. B 1998, 102, 7588.
126. Attfield, M. P.; Catlow, C. R. A.; Sokol, A. A. Chem. Mater. 2001, 13,
47084713.
127. Camblor, M. A.; Villaescusa, L. A.; Diaz-Cabanas, M. J. Top. Catal. 1999, 9,
5976.
128. Guth, J. L.; Kessler, H.; Bourgogne, M.; Wey, R.; Szabo, G. F. P. FR2567868,
1986.
129. Kessler, H.; Patarin, J.; Schott-Darie, C. Stud. Surf. Sci. Catal. 1994, 85,
75113.
130. Estermann, M.; McCusker, L. B.; Baerlocher, C.; Merrouche, A.; Kessler, H. Nature
1991, 352, 320.
131. Camblor, M. A.; Corma, A.; Lightfoot, P.; Villaescusa, L. A.; Wright, P. A. Angew.
Chem. Int. Ed. Engl. 1997, 36, 26592661.
132. Villaescusa, L. A.; Barrett, P. A.; Camblor, M. A. Angew. Chem. Int. Ed. 1999, 38,
19972000.
133. Leiva, S.; Sabater, M. J.; Valencia, S.; Sastre, G.; Fornes, V.; Rey, F.; Corma, A.
C. R. Chim. 2005, 8(34), 369378.
134. Camblor, M. A.; Diaz-Cabanas, M.-J.; Perez-Pariente, J.; Teat, S. J.; Clegg, W.;
Shannon, I. J.; Lightfoot, P.; Wright, P. A.; Morris, R. E. Angew. Chem. Int. Ed.
1998, 37, 21222126.
135. Corma, A.; Navarro, M. T.; Rey, F.; Rius, J.; Valencia, S. Angew. Chem. Int. Ed.
2001, 40, 22772280.
136. Corma, A.; Diaz-Cabanas, M. J.; Martinez-Triguero, J.; Rey, F.; Rius, J. Nature
2002, 418, 514517.
137. Castaneda, R.; Corma, A.; Fornes, V.; Martinez-Triguero, J.; Valencia, S. J. Catal.
2006, 238, 7987.
138. Castaneda, R.; Corma, A.; Fornes, V.; Rey, F.; Rius, J. J. Am. Chem. Soc. 2003,
125(26), 78207821.
139. Guth, J. L.; Kessler, H.; Caullet, P.; Hazm, J.; Merrouche, A.; Patarin, J. 9th Int.
Zeol. Conf., Montreal. Butterworth-Heinemann: Montreal, 1993; p 73.
140. Toby, B. H.; Khosrovani, N.; Dartt, C. B.; Davis, M. E.; Parise, J. B. Microporous
Mesoporous Mater. 2000, 39, 77.
141. Koller, H.; Wolker, A.; Eckert, H.; Panz, C.; Behrens, P. Angew. Chem. Int. Ed.
1997, 36, 2823.
142. Koller, H.; Wolker, A.; Villaescusa, L. A.; Daz-Cabanas, M. J.; Valencia, S.;
Camblor, M. A. J. Am. Chem. Soc. 1999, 121, 3368.
143. Li, H.; Yaghi, O. M. J. Am. Chem. Soc. 1998, 120(40), 1056910570.
144. Blasco, T.; Corma, A.; Daz-Cabanas, M. J.; Rey, F.; Vidal-Moya, J. A.; Zicovich-
Wilson, C. M. J. Phys. Chem. B 2002, 106(10), 26342642.
145. Cantin, A.; Corma, A.; Leiva, S.; Rey, F.; Rius, J.; Valencia, S. J. Am. Chem. Soc.
2005, 127, 1156011561.
146. Corma, A.; Daz-Cabanas, M. J.; Jiang, J.; Afeworki, M.; Dorset, D. L.;
Soled, S. L.; Strohmaier, K. G. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 13997.
147. Sun, J. L.; Bonneau, C.; Cantn, A.; Corma, A.; Daz-Cabanas, M. J.; Moliner, M.;
Zhang, D.; Li, M.; Zou, X. D. Nature 2009, 458, 1154.
148. Zones, S. I. Microporous Mesoporous Mater. 2011, 144(13), 18.
149. Blasco, T.; Corma, A.; Diaz-Cabanas, M. J.; Rey, F.; Vidal-Moya, J. A.; Zicovich-
Wilson, C. M. J. Phys. Chem. B 2002, 106, 26342642.
150. Medina, M. E.; Platero-Prats, A. E.; Snejko, N.; Rojas, A.; Monge, A.; Gandara, F.;
Gutierrez-Puebla, E.; Camblor, M. A. Adv. Mater. 2011.
151. Ueda, S.; Koizumi, M.; Baerlocher, C.; McCusker, L. B.; Meier, W. M. In 7th Int.
Zeolite Conf., Tokyo, 1986, pp Poster Paper 3C-3.

152. Cheetham, A.; Fjellvag, H.; Gier, E.; Kongshaug, K. O.; Lillerud, K. P.;
Stucky, G. D. In Studies in Surface Science and Catalysis; Galarneau, A.,
Fajula, F., Di Renzo, F., Vedrine, J., Eds.; Elsevier: Amsterdam, 2001; Vol. 135,
p 158.
153. Annen, M. J.; Davis, M. E.; Higgins, J. B.; Schlenker, J. L. J. J. Chem. Soc. Chem.
Commun. 1991, 1175.
154. McCusker, L.; Grosse-Kunstleve, R. W.; Baerlocher, C.; Yoshikawa, M.;
Davis, M. E. Microporous Mater. 1996, 6, 295.
155. Rohrig, C.; Gies, H. Angew. Chem. Int. Ed. 1995, 34, 63.
156. Hazen, R. M.; Yang, H.; Finger, L. W.; Fursenko, B. A. Am. Mineral. 1999, 84,
987.
157. Park, S. H.; Daniels, P.; Gies, H. Microporous Mesoporous Mater. 2000, 37,
129143.
158. Park, S.-H.; Parise, J. B.; Gies, H.; Liu, H.; Grey, C. P.; Toby, B. H. J. Am. Chem.
Soc. 2000, 122, 1102311024.
159. Davis, M. E. Nature 2002, 417, 813.
160. Valerio, G.; Plevert, J.; Goursot, A.; Renzo, F. Phys. Chem. Chem. Phys. 2000,
2(5), 10911094.
161. Krutskaya, T. M.; Koleyhev, A. M.; Morozkova, V. E.; Berger, A. S. Russ. J. Inorg.
Chem. 1985, 30, 438442.
162. Strohmaier, K. G.; Vaughan, D. E. W. J. Am. Chem. Soc. 2003, 125,
1603516039.
163. Aiello, R.; Nagy, J. B.; Giordano, G.; Katovic, A.; Testa, F. C. R. Chim. 2005, 8,
321329.
164. Asensi, M. A.; Corma, A.; Martinez, A.; Derewinski, M.; Krysciak, J.;
Tamhankar, S. S. Appl. Catal. A 1998, 174, 163175.
165. van Donk, S.; Bitter, J. H.; de Jong, K. P. Appl. Catal. Gen. 2001, 212(12),
97116.
166. Taramasso, M.; Perego, C.; Notari, B. US4410501, 1983.
167. Camblor, M. A.; Corma, A.; Perez-Pariente, J. Zeolites 1993, 13, 8287.
168. Fan, W.; Wu, P.; Namba, S.; Tatsumi, T. Angew. Chem. Int. Ed. 2003, 43, 236.
169. Moliner, M.; Serna, P.; Cantin, A.; Sastre, G.; Daz-Cabanas, M. J.; Corma, A.
J. Phys. Chem. C 2008, 112, 19547.
170. Corma, A.; Nemeth, L. T.; Renz, M.; Valencia, S. Nature 2001, 412,
423425.
171. Corma, A.; Domine, M. E.; Nemeth, L.; Valencia, S. J. Am. Chem. Soc. 2002, 124,
31943195.
172. Moliner, M.; Roman-Leshkov, Y.; Davis, M. E. Proc. Natl. Acad. Sci. U. S. A.
2010, 107(14), 61646168.
173. Roman-Leshkov, Y.; Moliner, M.; Labinger, J. A.; Davis, M. E. Angew. Chem. Int.
Ed. 2010, 49(47), 89548957.
174. Holm, M. S.; Saravanamurugan, S.; Taarning, E. Science 2010, 328, 602.
175. Wadlinger, R. L.; Kerr, G. T.; Rosinski, E. J. US3308069, 1967.
176. Argauer, R.; Landolt, G. R. US3702886, 1972.
177. Burton, A. W.; Zones, S. I. In Introduction to Zeolite Science and Practice Stud.
Surf. Sci. Catal.; Cejka, J., Bekkum, H. V., Corma, A., Schuth, F., Eds.; 168;
Elsevier B.V: Amsterdam, 2007; pp 137179.
178. Burton, A. W.; Zones, S. I. Stud. Surf. Sci. Catal. 2007, 168, 137179.
179. Lawton, S. L.; Rohrbaugh, W. J. Science 1990, 247, 1319.
180. Hong, S. B.; Cho, H. M.; Davis, M. E. J. Phys. Chem. 1993, 97(8), 16221628.
181. Nordstand, R. A. V.; Santilli, D. S.; Zones, S. I. ACS Symp. Ser. 1988, 368, 236.
182. Arhancet, J. P.; Davis, M. E. Chem. Mater. 1991, 3, 567.
183. Anderson, M. N.; Pachis, K. S.; Prebin, F.; Carr, S. W.; Terasaki, O.; Ohsuna, T.;
Alfredson, V. J. Chem. Soc. Chem. Commun. 1991, 1660.
184. Delprato, F.; Delmotte, J. L.; Guth, J. L.; Huve, L. Zeolites 1990, 10, 546.
185. Annen, M. J.; Young, D.; Arhancet, J. P.; Davis, M. E.; Schramm, S. Zeolites
1991, 11, 98.
186. Gies, H. Inclusion Compounds. Academic Press: London, 1995; Vol. 5, p 1.
187. Gies, H.; Marler, B. Zeolites 1992, 12, 4249.
188. Burton, A. W.; Zones, S. I. In Introduction to Zeolite Science and Practice;
Cejka, J., Bekkum, H. V., Corma, A., Schuth, F., Eds.; 168, Elsevier B.V.:
Amsterdam, 2007; pp 137179.
189. Aiello, R.; Barrer, R. M. J. Chem. Soc. A 1970, 1470.
190. Kerr, G. T. Inorg. Chem. 1966, 5, 1537.
191. Moscoso, J.; Lewis, G.; Gisselquist, J.; Miller, M.; Rohde, L. US6713041, 2004
192. Corma, A.; Rey, F.; Rius, J.; Sabater, M. J.; Valencia, S. Nature 2004, 431,
287290.
193. Pinar, P. A. B.; Perez, P. J.; Gomez-Hortigueela, S. L. Method for Preparation of an
Aluminosilicate with Ferrierite Structure from Gels Containing Tetramethyl
Ammonium and Benzylmethylpyrrolidine, and Uses Thereof. WO2008/116958,
2008.
194. Gomez-Hortigueela, L.; Pinar, A. B.; Cora, F.; Perez-Pariente, J. Chem. Commun.
2010, 46, 20732075.
Zeolites 129
195. Pinar, A. B.; Gomez-Hortiguela, L.; Perez-Pariente, J. Chem. Mater. 2007,
19, 56175626.
196. Pinar, A. B.; Marquez-Alvarez, C.; Grande-Casas, M.; Perez-Pariente, J. J. Catal.
2009, 263, 258265.
197. Marquez-Alvarez, C.; Pinar, A. B.; Garcia, R.; Grande-Casas, M.; Perez-Pariente, J.
Top. Catal. 2009, 52, 12811291.
198. Roman-Leshkov, Y.; Moliner, M.; Davis, M. E. J. Phys. Chem. C 2011,
115, 1096.
199. Perez-Pariente, J.; Garca, R.; Gomez-Horiguela, L.; Pinar, A. B. In Zeolites and
Catalysis, Synthesis, Reactions and Applications; Cejka, J., Corma, A., Zones, S.,
Eds.; Wiley-VCH: Weinhem, 2010; pp 107130.
200. Lewis, G.; Miller, M.; Moscoso, J.; Wilson, B.; Knight, L.; Wilson, S. In 14th Int.
Zeol. Conf, Cape Town, South Africa, 2004; p 188.
201. Lee, H.; Zones, S. I.; Davis, M. E. J. Phys. Chem. B 2005, 109, 21872191.
202. Lee, H.; Zones, S. I.; Davis, M. E. Nature 2003, 425, 385388.
203. Matsukata, M.; Ogura, M.; Osaki, T.; Hari Prasad Rao, P.; Nomura, M.; Kikuchi, E.
Top. Catal 1999, 9(1), 7792.
204. Xu, W.; Dong, J.; Li, J.; Li, W.; Wu, F. J. Chem. Soc. Chem. Commun. 1990,
755.
205. Kim, M. H.; Li, H. X.; Davis, M. E. Microporous Mater. 1993, 1, 191.
206. Matsukata, M.; Nishiyama, N.; Ueyama, N. Microporous Mater. 1993, 1, 219.
207. Hari Prasad Rao, P. R.; Leon Y Leon, C. A.; Ueyama, K.; Matsukata, M.
Microporous Mesoporous Mater. 1998, 21, 305.
208. Schmidt, F. Appl. Catal. Gen. 2001, 221(12), 1521.
209. Morris, R. E.; Weigel, S. J. Chem. Soc. Rev. 1997, 26, 309318.
210. Bibby, D. M.; Dale, M. P. Nature 1985, 317, 157.
211. Kuperman, A.; Nadimi, S.; Oliver, S.; Ozin, G. A.; Garces, J. M.; Olken, M. M.
Nature 1993, 365(6443), 239242.
212. Dorset, D. L.; Kennedy, G. J.; Strohmaier, K. G.; Diaz-Cabanas, M. J.; Rey, F.;
Corma, A. J. Am. Chem. Soc. 2006, 128, 88628867.
213. Corma, A.; Diaz-Cabanas, M. J.; Jorda, J. L.; Rey, F.; Sastre, G.; Strohmaier, K. G.
J. Am. Chem. Soc. 2008, 130, 1648216483.
214. Simancas, R.; Dari, D.; Velamazan, N.; Navarro, M. T.; Cantin, A.; Jorda, J. L.;
Sastre, G.; Corma, A.; Rey, F. Science 2010, 330, 1219.
215. Dari, D.; Velamazan, N.; Simancas, R.; Martnez-Triguero, J.; Martnez, C.;
Navarro, M. T.; Cantn, A.; Jorda, J. L.; Corma, A.; Rey, F. 16th IZC Sorrento
(Italy), July 49, 2010; Colella, C.; Aprea, P.; Gennaro, B. d.; Liguori, B., Eds.;
Book of Abstracts; A. De Frede: Napoli, 2010; p 16.
216. Drulie, E. A.; Wragg, D. S.; Parnham, E. R.; Wheatley, P. S.; Slawin, A. M. Z.;
Warren, J. E.; Morris, R. E. Angew. Chem. Int. Ed. 2007, 46, 78397843.
217. Parnham, E. R.; Morris, R. E. Chem. Mater. 2006, 18, 48824887.
218. Parnham, E. R.; Morris, R. E. J. Am. Chem. Soc. 2006, 128, 22042205.
219. Parnham, E. R.; Morris, R. E. Acc. Chem. Res. 2007, 40, 10051013.
220. Cooper, E. R.; Andrews, C. D.; Wheatley, P. S.; Webb, P. B.; Wormald, P.;
Morris, R. E. Nature 2004, 430, 10121016.
221. Morris, R. E. In Zeolites and Catalysis, Synthesis, Reactions and Applications;
Cejka, J., Corma, A., Zones, S., Eds.; Wiley-VCH: New Jersey, 2010; pp 87106.
222. Davis, M. E.; Saldarriaga, C.; Montes, C.; Garces, J.; Crowder, C. Nature 1988,
331, 698.
223. Huo, Q. S.; Xu, R. R.; Li, S. G.; Ma, Z. G.; Thomas, J. M.; Jones, R. H.;
Chippindale, A. M. J. Chem. Soc. Chem. Commun. 1992, 875876.
224. Freyhardt, C. C.; Tsapatsis, M.; Lobo, R. F.; Balkus, K. J., Jr.; Davis, M. E. Nature
1996, 381, 295.
225. Lobo, R. F.; Tsapatsis, M.; Freyhardt, C. C.; Khodabandeh, S.; Wagner, P.;
Chen, C.-Y.; Balkus, K. J., Jr.; Zones, S. I.; Davis, M. E. J. Am. Chem. Soc. 1997,
119, 84748484.
226. Wessels, T.; Baerlocher, C.; McCusker, L. B.; Creyghton, E. J. J. Am. Chem. Soc.
1999, 121, 62426247.
227. Wagner, P.; Yoshikawa, M.; Lovallo, M.; Tsuji, K.; Taspatsis, M.; Davis, M. E.
Chem. Commun. 1997, 21792180.
228. Burton, A.; Elomari, S.; Chen, C.-Y.; Medrud, R. C.; Chan, I. Y.; Bull, L. M.;
Kibby, C.; Harris, T. V.; Zones, S. I.; Vittoratos, E. S. Chem. Eur. J. 2003,
9, 57375748.
229. Vogt, E. T. C.; Richardson, J. W. J. Solid State Chem. 1990, 87, 469.
230. Schaefer, H.-E. Nanoscience: The Science of the Small in Physics, Engineering,
Chemistry, Biology and Medicine. Springer: Berlin Heidelberg, 2010.
231. Corma, A.; Llopis, F. J.; Martnez, C.; Sastre, G.; Valencia, S. J. Catal. 2009,
268(1), 917.
232. Corma, A.; Daz-Cabanas, M. J.; Jorda, J. L.; Martinez, C.; Moliner, M. Nature
2006, 443, 842.
233. Zhang, L.; Yang, C.; Meng, X.; Xie, B.; Wang, L.; Ren, L.; Ma, S.; Xiao, F.-S.
Chem. Mater. 2010, 22, 8.
234. Wu, Z.; Song, J.; Ji, Y.; Ren, L.; Xiao, F.-S. Chem. Mater. 2008, 20(2), 3.
130 Zeolites
235. Iyoki, K.; Kamimura, Y.; Itabashi, K.; Shimojima, A.; Okubo, T. Chem. Lett. 2010,
39(7), 730.
236. Song, J.; Dai, L.; Ji, Y.; Xiao, F.-S. Chem. Mater. 2006, 18, 2775.
237. Majano, G.; Darwiche, A.; Mintova, S.; Valtchev, V. Ind. Eng. Chem. Res. 2009,
48, 7084.
238. Yokoi, T.; Yoshioka, M.; Imai, H.; Tatsumi, T. Angew. Chem. 2009,
121, 10068.
239. Yokoi, T.; Yoshioka, M.; Imai, H.; Tatsumi, T. Angew. Chem. Int. Ed. 2009,
48, 9884.
240. Xie, B.; Song, J.; Ren, L.; Ji, Y.; Li, J.; Xiao, F.-S. Chem. Mater. 2008, 20(14), 3.
241. Kamimura, Y.; Chaikittisilp, W.; Itabashi, K.; Shimojima, A.; Okubo, T. Chem.
Asian J. 2010, 5, 10.
242. Gonzalez-Arellano, C.; Corma, A.; Iglesias, M.; Sanchez, F. Adv. Synth. Catal.
2004, 346(11), 13161328.
243. Climent, M. J.; Corma, A.; Velty, A. Appl. Catal. Gen. 2004, 263(2), 155161.
244. Rodriguez, I.; Climent, M. J.; Iborra, S.; Fornes, V.; Corma, A. J. Catal. 2000,
192(2), 441447.
245. Corma, A.; Diaz, U.; Domine, M. E.; Fornes, V. Chem. Commun. 2000, 2,
137138.
246. Corma, A.; Diaz, U.; Fornes, V.; Guil, J. M.; Martinez-Triguero, J.; Creyghton, E. J.
J. Catal. 2000, 191, 218224.
247. Corma, A.; Garcia, H.; Miralles, J. Microporous Mesoporous Mater. 2001,
43, 161169.
248. Na, K.; Choi, M.; Park, W.; Sakamoto, Y.; Terasaki, O.; Ryoo, R. J. Am. Chem. Soc.
2010, 132(12), 41694177.
249. Na, K.; Jo, C.; Kim, J.; Cho, K.; Jung, J.; Seo, Y.; Messinger, R. J.; Chmelka, B. F.;
Ryoo, R. Science 2011, 333(6040), 328332.
250. Schick, J.; Daou, T. J.; Caullet, P.; Paillaud, J. L.; Patarin, J.; Mangold-
Callarec, C. Chem. Commun. 2011, 47(3), 902904.
251. Akporiaye, D. E.; Dahl, I. M.; Karlsson, A.; Wendelbo, R. Angew. Chem. Int. Ed.
1998, 37, 609.
252. Song, J.; Yu, J.; Li, G.; Li, Y.; Wang, Y.; Xu, R. Chem. Commun. 2002,
17201721.
253. Corma, A.; Diaz-Cabanas, M. J.; Rey, F. Chem. Commun. 2003, 10501051.
254. Blasco, T.; Corma, A.; Diaz-Cabanas, M. J.; Rey, F.; Rius, J.; Sastre, G.;
Vidal-Moya, J. A. J. Am. Chem. Soc. 2004, 126, 1341413423.
255. Corma, A.; Diaz-Cabanas, M. J.; Moliner, M.; Martinez, C. J. Catal. 2006,
241, 312318.
256. Baumes, L. A.; Moliner, M.; Corma, A. Chem. Eur. J. 2009, 15(17),
42584269.
257. Knight, L. M.; Lewis, G. J. In Studies in Surface Science and Catalysis;
van Steen, E., Claeys, I. M., Callanan, L. H., Eds.; Elsevier: Amsterdam, 2005;
Vol. 154 (Part A), pp 170179.
258. Moscoso, J. G.; Lewis, G. J.; Gisselquist, J. L.; Miller, M. A.; Knight, L. M.
US6713041, 2002.
259. Miller, M. A.; Moscoso, J. G.; Koster, S. C.; Gatter, M. G.; Lewis, G. J. In Studies
in Surface Science and Catalysis; Xu, R., Gao, Z., Chen, J., Yan, W. Eds.; Elsevier:
Amsterdam, 2007; Vol. 170 (Part A), pp 487492.
260. Klanner, C.; Farrusseng, D.; Baumes, L.; Lengliz, M.; Mirodatos, C.; Schuth, F.
Angew. Chem. Int. Ed. 2004, 43, 5347.
261. Baumes, L. A.; Blansche, A.; Serna, P.; Tchougang, A.; Lachiche, N.; Collet, P.;
Corma, A. Mater. Manuf. Processes 2009, 24, 282292.
262. Baumes, L. A.; Gaudin, R.; Serna, P.; Nicoloyannis, N.; Corma, A. Comb. Chem.
High Throughput Screening 2008, 11, 266282.
263. Corma, A.; Moliner, M.; Serra, J. M.; Serna, P.; Diaz-Cabanas, M. J.;
Baumes, L. A. Chem. Mater. 2006, 18, 32873296.
264. Bergh, H. S. In High-Throughput Screening in Chemical Catalysis; Hagmeyer, A.,
Strasser, P., Volpe, A. F., Eds.; Wiley-VCH: Weinheim, 2004; pp 6388.
265. Baumes, L. A.; Serna, P.; Corma, A. Appl. Catal. Gen. 2010, 381(12), 12.
266. Serna, P.; Baumes, L. A.; Moliner, M.; Corma, A. J. Catal. 2008, 258,
2534.
267. Newsam, J. M.; Bein, T.; Klein, J.; Maier, W. F.; Stichert, W. Microporous
Mesoporous Mater. 2001, 48, 355365.
268. Pescarmona, P. P.; Rops, J. T.; Waal, J. C. V. D.; Jansen, J. C.; Maschmeyer, T.
J. Mol. Chem. A 2002, 182183.
269. Tompsett, G. A.; Conner, W. C.; Yngvesson, K. S. Chem. Phys. Chem. 2006,
7, 296319.
270. Gharibeh, A. In 21st NAM, San Francisco (US), 2009.
271. Haden, W.L. Jr.; Dzierzanowski, F. J. Method for Making Synthetic Zeolitic
Material. US2992068, 1961.
272. Brown, S. M.; Durante, V. A.; Reagan, W. J.; Speronello, B. K. Fluid Catalytic
Cracking Catalyst Comprising Microspheres Containing more than About 40
Percent by Weight Y-faujasite and Methods for Making. US4493902, 1985.
273. Madon, R.; Harris, D. H.; Xu, M.; Stockwell, D.; Lerner, B.; Dodwell, G. W. Fluid
Catalytic Cracking Catalysts, for Nickel- and Vanadium-Containing Feeds,
Prepared from Kaolin-Derived Microspheres with Dispersible Boehmite and
Y-faujasite Zeolites. US20030130107A1, 2003.
274. Madon, R. J.; Harris, D. H.; Xu, M.; Stockwell, D. M.; Lerner, B.; Dodwell, G. W.
Fluid Catalytic Cracking Catalysts Prepared from Kaolin-Derived Microspheres
with Dispersible Boehmite and Y-faujasite Zeolites. WO2003054114A2,
2003.
275. Madon, R.; Harris, D. H.; Xu, M.; Stockwell, D.; Lerner, B.; Dodwell, G. W.
Zeolite-Based Fluid Cracking Catalysts for Petroleum Residues with Alumina
Phases for Passivation of Nickel and Vanadium Contaminants.
US20030089640A1, 2003.
276. Dautzenberg, F. M.; Mukherjee, M. Chem. Eng. Sci. 2001, 56(2), 17.
277. Gaffney, A. M. In 21st NAM, San Francisco (US), 2009.
278. van der Puil, N.; Overbeek, R.; Murrell, L.; Koegler, J. H.; Yeh, P.; Chang, Y. F.;
Khonsari, A.; Dautzenberg, F. In Abstracts of Papers, 220th ACS National
Meeting, Washington, DC, 2024 August, 2000, COLL-108.
279. Ahmaruzzaman, M. Prog. Energy Combust. Sci. 2010, 36, 327363.
280. Fernandes, A. A.; Frajndlich, E. U.; Riella, H. G. Mater. Sci. Forum 2005,
498499, 676680.
281. Querol, X.; Umana, J. C.; Plana, F.; Alastuey, A.; Lopez-Soler, A.; Medinaceli, A.;
Valero, A.; Domingo, M. J.; Garcia-Rojo, E. Fuel 2001, 80, 857865.
282. Vermeiren, W.; Gilson, J.-P. Top. Catal. 2009, 52, 31.
283. Ackley, M. W.; Rege, S. U.; Saxena, H. Microporous Mesoporous Mater. 2003,
61(13), 2542.
284. Barret, P. A.; Stephenson, N. A. In Zeolites and Ordered Porous Solids:
Fundamentals and Applications; Martinez, C., Perez-Pariente, J., Eds.; Editorial
Universitat Politecnica de Valencia: Valencia, 2011; pp 6790.
285. Yang, R. T. Adsorbents: Fundamentals and Applications. John Wiley & Sons:
New Jersey, 2003.
286. Kerry, F. G. Industrial Gas Handbook: Gas Separation and Purification; CRC
Press, Taylor & Francis Group: Boca Raton, FL, 2007; Vol. 1, pp 273287.
287. Caro, J.; Noack, M.; Kolsch, P.; Schafer, R. Microporous Mesoporous Mater.
2000, 38(1), 324.
288. Gutierrez-Sevillano, J. J.; Dubbeldam, D.; Rey, F.; Valencia, S.; Palomino, M.;
Martin-Calvo, A.; Calero, S. J. Phys. Chem. C 2010, 114, 1490714914.
289. Palomino, M.; Cantin, A.; Corma, A.; Leiva, S.; Rey, F.; Valencia, S. Chem.
Commun. 2007, 12331235.
290. Palomino, M.; Corma, A.; Rey, F.; Valencia, S. Langmuir 2010, 26,
19101917.
291. http://www.zeolites.eu/.
292. Dyer, A.; Enamy, H. Zeolites 1981, 1(1), 710.
293. Allen, S.; Carr, S.; Chapple, A.; Dyer, A.; Heywood, B. Phys. Chem. Chem. Phys.
2002, 4(11), 24092415.
294. Armor, J. N. Catal. Today 2010, 163(1), 39.
295. Perego, C.; Carati, A., In Zeolites: From Model Materials to Industrial Catalysts;
Cejka, J.; Perez-Pariente, J.; Roth, W. J., Eds.; Transworld Research Network:
Trivandrum, 2008; pp 357389.
296. Anastas, P. T.; Kirchhoff, M. M.; Williamson, T. C. Appl. Catal. A 2001, 221, 11.
297. Corma, A. Chem. Rev. 1997, 97, 23732419.
298. Corma, A. Chem. Rev. 1995, 95, 559614.
299. Yilmaz, B.; Muller, U. Top. Catal. 2009, 52, 888895.
300. Di Renzo, F.; Fajula, F. In Zeolites and Ordered Mesoporous Materials: Progress
and Prospects; Cejka, J., Bekkum, H. V., Eds.; Studies in Surface Science and
Catalysis; Elsevier: Amsterdam, 2005; pp 112.
301. Rinaldi, R.; Schutz, F. Energy Environ. Sci. 2009, 2, 17.
302. Forzatti, P.; Lietti, L.; Tronconi, E. In Catalysis for Sustainable Energy Production;
Barbaro, P., Bianchini, C., Eds.; Wiley-VCH: Weinhem, 2009; pp 393438.
303. Martinez, C.; Corma, A. Coord. Chem. Rev. 2011, 255, 15581580.
304. Weisz, P. B. Pure Appl. Chem. 1980, 52, 2091.
305. Degnan, T. F., Jr. J. Catal. 2003, 216, 15.
306. Madsen, C.; Jacobsen, C. J. H. Chem. Commun. 1999, 673.
307. Jacobsen, C. J. H.; Madsen, C.; Janssens, T. V. W.; Jakobsen, H. J.; Skibsted, J.
Microporous Mesoporous Mater. 2000, 39, 393.
308. Van Donk, S.; Janssen, A. H.; Bitter, J. H.; Jong, K. P. D. Catal. Rev. Sci. Eng.
2003, 45, 297319.
309. Egeblad, K.; Christensen, C. H.; Kustova, M.; Christensen, C. H. Chem. Mater.
2008, 20, 946960.
310. Perez-Ramirez, J.; Christensen, C. H.; Egeblad, K.; Christensen, C. H.;
Groen, J. C. Chem. Soc. Rev. 2008, 37(11), 25302542.
311. Corma, A.; Martnez, A. In Zeolites and Ordered Mesoporous Materials: Progress
and Prospects; Cejka, J., Bekkum, H. V., Eds.; Studies in Surface Science and
Catalysis; Elsevier: Amsterdam, 2005; p 337.

312. Rigutto, M. In Zeolites and Catalysis, Synthesis, Reactions and Applications;
Cejka, J., Corma, A., Zones, S., Eds.; Wiley-VCH: Weinhem, 2010; pp 547584.
313. Breck, D. W. US3130007, 1964.
314. Dwyer, J.; Millward, D.; OMalley, P. J.; Araya, A.; Corma, A.; Fornes, V.;
Martinez, A. J. Chem. Soc. Faraday Trans. 1990, 86, 10011005.
315. Haas, A. Microporous Mesoporous Mater. 1999, 28(2), 325333.
316. Bonetto, L.; Camblor, M. A.; Corma, A.; Perez-Pariente, J. Appl. Catal. A 1992,
82, 3750.
317. Corma Canos, A.; Diaz Cabanas, M. J.; Martinez-Triguero, J.; Rey Garcia, F. Use of
Zeolite ITQ-21 in the Catalytic Cracking of Petroleum Fractions. EP1445297 2004.
318. Moliner, M.; Diaz-Cabanas, M. J.; Fornes, V.; Martinez, C.; Corma, A. J. Catal.
2008, 254, 101109.
319. Corma, A.; Martinez-Triguero, J. J. Catal. 1997, 165, 102120.
320. Corma, A.; Martinez-Triguero, J.; Valencia, S.; Benazzi, E.; Lacombe, S. J. Catal.
2002, 206, 125133.
321. Corma, A.; Martnez, C.; Martnez-Triguero, J.; Valencia, S. In Medium Pore
Zeolites as FCC Additives: Effect of an Additional Channel System on
Shape-Selectivity, 16th IZC Sorrento (Italy); Colella, C., Aprea, P., Gennaro, B. D.,
Liguori, B., Eds.; 2010; p 200.
322. Xie, A. Petrochem. Technol. 2002, 31(9), 691.
323. Corma, A.; Martinez-Triguero, J.; Martinez, C. J. Catal. 2001, 197, 151159.
324. Derouane, E. G., Robert, S. M., Eds.; In Catalysis for Fine Chemicals:
Microporous and Mesoporous Solid Catalysts. John Wiley and Sons: Chichester,
2006.
325. Climent, M. J.; Corma, A.; Iborra, S. In Zeolites and Catalysis, Synthesis,
Reactions and Applications; Cejka, J., Corma, A., Zones, S., Eds.; Wiley-VCH:
Weinheim, 2010; pp 775826.
326. Corma, A. J. Catal. 2003, 216(12), 298312.
327. Weitkamp, J.; Hunger, M.; Rymsa, U. Microporous Mesoporous Mater. 2001,
48, 16.
328. Tanabe, K.; Holderich, W. F. Appl. Catal. Gen. 1999, 181(2), 399434.
329. Peeters, A.; Majano, G.; Steenackers, B.; Philippaerts, A.; Dusselier, M.;
Geboers, J.; Dijkmans, J.; Vyver, S. V. D.; Sels, B. F.; Vos, D. E. D. In Zeolites and
Ordered Porous Solids: Fundamentals and Applications; Martinez, C.,
Perez-Pariente, J., Eds.; Editorial Universitat Politecnica de Valencia: Valencia,
2011; pp 6790.
330. Tatsumi, T. In Zeolites and Ordered Porous Solids: Fundamentals and
Applications; Martinez, C., Perez-Pariente, J., Eds.; Editorial Universitat
Politecnica de Valencia: Valencia, 2011; pp 6790.
Zeolites 131
331. Corma, A.; Camblor, M. A.; Perez, P. J.; Valencia, S.; Costantini, M.; Gilbert, L.
Process for Manufacturing Zeolites Ti-b, and the Zeolites Obtained and Their Use.
FR2730722A1, 1996.
332. Tuel, A. Zeolites 1995, 15(3), 236242.
333. Diaz-Cabanas, M.-J.; Villaescusa, L. A.; Camblor, M. A. Chem. Commun. 2000,
761762.
334. Peng, W.; Tatsumi, T.; Komatsu, T.; Yashima, T. Chem. Lett. 2000, 29(7),
774775.
335. Hidenori, Y.; Iwamoto, M. Catal. Catal. 2003, 45(1), 2628.
336. Price, G. L.; Kanazirev, V. US5583081, 1996.
337. Kawaga, S.; Teraoka, Y. US5078981, 1992.
338. Baik, J. H.; Yim, S. D.; Nam, I.-S.; Mok, Y. S.; Lee, J.-H.; Cho, B. K.; Oh, S. H.
Top. Catal. 2004, 30/31, 3741.
339. Balle, P.; Geiger, B.; Klukowski, D.; Pignatelli, M.; Wohnrau, S.; Menzel, M.;
Zirkwa, I.; Brunklaus, G.; Kureti, S. Appl. Catal. B 2009, 91, 587595.
340. Fedeyko, J. M.; Chen, H.-Y.; Ballinger, T. H.; Weigert, E. C.; Chang, H.-L.;
Cox, J. P.; Andersen, P. J. Soc. Automot. Eng., [Spec. Publ.] SP 2009, SP-2254,
367373.
341. Centi, G.; Perathoner, S. In Zeolites and Catalysis, Synthesis, Reactions and
Applications; Cejka, J., Corma, A., Zones, S., Eds.; Wiley-VCH: Weinhem, 2010;
pp 745774.
342. Corma, A.; Martinez, A. Stud. Surf. Sci. Catal. 2005, 157, 337366.
343. Corma, A.; Fornes, V.; Guil, J. M.; Pergher, S.; Maesen, T. L. M.; Buglass, J. G.
Microporous Mesoporous Mater. 2000, 38(23), 301309.
344. Corma, A.; Martinez, A.; Martinez-Soria, V. J. Catal. 2001, 200(2), 259269.
345. Corma, A.; Botella, P.; Mitchell, C. Chem. Commun. 2004, (17),
20082010.
346. Martinez, A.; Prieto, G. J. Catal. 2007, 245(2), 470476.
Relevant Websites
http://minerals.usgs.gov/minerals/pubs/commodity/zeolites/mcs-2011-zeoli.pdf U.S.
Geological Survey (USGS).
http://www.iza-online.org/ International Zeolite Association (IZA).
http://www.hypotheticalzeolites.net MD Foster and MMJ Treacy, A Database of
Hypothetical Zeolite Structures.
http://www.zeolites.eu/ The European Zeolites Producers Association (EUZEPA).
 5.10 Zeolite Nanoparticles
S Mintova, Universite de Caen, Caen, France
E-P Ng, Universiti Sains Malaysia, Penang, Malaysia
2013 Elsevier Ltd. All rights reserved.

5.10.1 Introduction 286

5.10.2 Syntheses of Zeolite Nanoparticles 286
5.10.2.1 Precursor Suspensions 286
5.10.2.1.1 Templated synthesis approach 286
5.10.2.1.2 Template-free synthesis approach 289
5.10.2.1.3 Seed-induced synthesis 290
5.10.2.1.4 Multistep synthesis approach 291
5.10.2.1.5 Ionothermal synthesis 291
5.10.2.2 Types of Heating 292
5.10.2.2.1 Microwave-assisted synthesis 292
5.10.2.2.2 Microchannel-assisted synthesis 293
5.10.2.3 Confined Space Synthesis 293
5.10.2.3.1 Reverse microemulsion synthesis 294
5.10.2.4 Other Methods 295
5.10.2.4.1 Direct-conversion synthesis approach 295
5.10.2.4.2 Centrifugation-assisted grinding 295
5.10.2.4.3 Laser-induced fragmentation method 295
5.10.2.5 Separation of Zeolite Nanocrystals 296
5.10.3 Applications of Zeolite Nanoparticles 297
5.10.3.1 Zeolite Membranes 297
5.10.3.2 Optical and Other Devices 297
5.10.3.3 Biological and Medical Applications 299
5.10.4 Conclusion 300
Acknowledgment 300
References 301

Abbreviations MEL Mobil number eleven

C Degree Celsius MnAlPO-5 Manganoaluminophosphate number five
AEI Aluminophosphate number eighteen MOR Mordenite
AFI Aluminophosphate number five MTW Mobil twelve
AlPO-5 Aluminophosphate number five nm Nanometer
AlPO-18 Aluminophosphate number eighteen NMR Nuclear magnetic resonance
BEA Beta OFF Offretite
CHA Chabazite RCF Relative centrifugal force
CTAB Cetyltrimethylammonium bromide rpm Rotation per minute
DLS Dynamic light scattering SAPO-5 Silicoaluminophosphate number five
EMT Elf Mulhouse two SAPO-34 Silicoaluminophosphate number thirty four
FAU Faujasite SDA Structure-directing agent
GIS Gismondine SOD Sodalite
HRTEM High-resolution transmission electron TEOS Tetraethyl orthosilicate
microscopy TS-1 Titanosilicate number one
IL Ionic liquid wt.% Weight percent
IR Infrared XRD X-ray diffraction
LTA Linde type A ZSM-5 Zeolite socony material number five
MFI Mobil number five

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00512-X 285

286 Zeolite Nanoparticles

conditions directing the crystal size, that is, the nutrient pool
5.10.1 Introduction
is limited, and thus after the exhausting a building component
the growth process would stop. Hence, the formation of zeolite
Nanoscale science and engineering provide unique under-
nanocrystals requires conditions that favor the nucleation over
standing and control of matter and mostly on a fundamental
crystal growth.
level. In particular, the nanosized particles have been fascinat-
The main approaches applied for the synthesis of nanosized
ing the world of science due to their unique properties and use
molecular sieves are (1) synthesis from clear precursor suspen-
in diverse fields including catalysis, photography, photonics,
sions in the presence of organic template, (2) low-temperature
electronics, labeling, imaging, sensing, etc.13 Inorganic nano-
syntheses from highly alkaline organic-template-free hydro-
particles, in particular zeolite materials, are the focus of many
gels, and (3) other methods including ionothermal, seed-
researchers due to their diverse framework-type structures con-
induced, confined space synthesis, etc.
taining one-, two-, or three-dimensional channel systems
In order to prepare zeolite nanocrystals, the type of the
(pores) whose dimension is on the order of the molecular
precursor suspensions (Section 5.10.2.1), type of heating
size.4 Hence, these materials with nanosized dimensions have
(Section 5.10.2.2), confined space effect (Section 5.10.2.3),
been considered in wide-ranging applications such as photon-
and other physicochemical parameters (Section 5.10.2.4) have
ics, sensors, electronic and optical detection systems, therapeu-
to be considered. Once the zeolite nanoparticles are prepared
tics, diagnostics, photovoltaics, and catalysis.
with desired size, structure, and morphology, the emphasis is on
In addition to the conventional zeolites, their nanosized
their stabilization in colloidal suspensions, to prevent further
counterparts with a size in the range of 51000 nm have
agglomeration and coalescence (Section 5.10.2.5).
attracted considerable attention during the last two de-
cades.533 Although the chemical composition and the
framework-type structure of zeolites are important, even 5.10.2.1 Precursor Suspensions
more vital are the size and shape of the nanoparticles which
5.10.2.1.1 Templated synthesis approach
determine their surface/colloidal properties. Different mor-
The templated synthesis approach is considered as the most
phologies and sizes of the zeolite nanocrystals can result
common method for preparation of zeolite nanocrystals. The
from fine-tuning of the synthesis parameters such as initial
zeolite nanocrystals are synthesized via hydrothermal treat-
gel composition, type of precursor materials, heating time
ment of clear aqueous suspensions at moderate temperatures
and type, and temperature of preparation and postsynthesis
(30120  C). Clear precursor suspensions have been used to
treatment in order to alter the nucleation and crystal growth
synthesize several nanocrystalline materials with faujasite
processes.2024 The reduction in particle size from micrometers
(FAU)-, Mobil number five (MFI)-, Mobil number eleven
to the nanometer scale leads to substantial changes in their
(MEL)-, sodalite (SOD)-, gismondine (GIS)-, Linde type A
properties and thus different performances even in traditional
(LTA)-, Beta (BEA)-, aluminophosphate number eighteen
applications are expected. Therefore, the reduction in the size
(AEI)-, and chabazite (CHA)-type framework structures.524
of zeolite crystals to the range of some unit cells is expected to
Organic templates (or structure-directing agents, SDAs) in
provide materials with completely new properties. Moreover,
most of these syntheses play the structure-directing role and
the significance of these nanosized materials is mainly related
provide the high alkalinity and high super-saturation level
with the emerging area of applications that goes far beyond
which is needed for crystallization of nanocrystals. The organic
traditional separation and catalytic processes. The possibility to
templates with different size, shape, and hydrophilicity influ-
obtain stable colloidal suspensions of microporous particles
ence the position of tetrahedrally coordinated Al and Si atoms in
capable of processing on different surfaces by rapid techniques
the zeolite framework, and lead to the formation of crystals with
is also of great importance for their advanced applications.
various type structures. On the other hand, several amines can
The development of versatile methods for the preparation
direct the synthesis of the same structure, while one type organic
of zeolite nanocrystals with defined structure, size, stability,
amine can also template many different structures. For example,
and morphology has attracted significant attention during the
tetraethylammonium cations (TEA) can template nanosized
last two decades. Some of the recent developments in the
zeolites such as BEA, ZSM-5 (MFI), AlPO-5 (AFI), SAPO-5
synthesis, characterization, and application of zeolite nano-
(AFI), AlPO-18 (AEI), and SAPO-34 (CHA).
crystals have been summarized in the review papers.57,2527
The organic templates mostly used for preparation of zeo-
This chapter aims to give a vivid look on the use of various
lite nanocrystals are summarized in Table 1. In addition to the
techniques for synthesis of zeolite nanocrystals with controlled
mostly used tetraalkylammonium (TAA) cations, some amines
size, stability, morphology, and possibility for increased crys-
are also used as co-templates or SDAs.
talline yield and scale-up processes. The most recent applica-
In general, within a given range of molecular compositions,
tions of the zeolite nanocrystals different from conventional
an increase of the SDA induces a rather narrow particle-size
zeolites are presented.
distribution of smaller crystals by hindering the agglomeration
of the precursor and crystalline particles. However, alkali cat-
ions used as counterbalancing ions for the crystalline structure
5.10.2 Syntheses of Zeolite Nanoparticles cannot be used that liberally as their influence in the crystalline
phase obtained is far greater due to the different interactions
The zeolite syntheses are performed in closed reacting systems, during the early stages of crystallization.20 Normally, the large
where the high super-saturation leads to spontaneous nucle- amount of organic templates is used not only to control the
ation and controlled crystallization process. Upon such crystal size but also to obtain stable crystalline suspensions.

Table 1 Organic templates used for the preparation of nanosized
zeolites
Templates Crystalline phase
Tetramethyl ammonium, TMA LTA, FAU, SOD, GIS, OFF
Tetraethyl ammonium, TEA AEI, AFI, CHA, BEA
Tetrapropyl ammonium, TPA MFI, AFI
Tetrabutyl ammonium, TBA AFI, MEL
Isopropyl amine, i-Pr2NH AEL
1-Ethyl-2,3-dimethylimidazolium, edmim AFI
4,40 -Trimethylenebis(N-methyl,N- BEA
benzylpiperidinium), TMP2
The as-prepared precursor suspensions with excess of organic
templates are treated under mild conditions (syntheses temper-
ature in most of the cases do not exceed 120  C) so that the
nucleation occurs more readily than crystal growth.9,11,20,2932
The high super-saturation is also important to provide concen-
trated precursor mixtures containing soluble species as depoly-
merized and reactive as possible. This is generally accomplished
through lowering the synthesis temperature that favors nucle-
ation over crystal growth. On the other hand, these conditions
result in slow nucleation and crystal growth and, thus, a
prolonged synthesis time and reduced crystalline yields are ob-
served in most of the syntheses carried out in highly super-
saturated precursor suspensions.
Besides, the hydrothermal process in clear precursor sus-
pensions can be systematically modified in order to have more
control over the size of the nanocrystals by changing the reac-
tion parameters such as template concentration/type, temper-
ature, aging, heating time, and source of reactants (silica,
alumina, or phosphate). The ratio between the organic tem-
plates and inorganic reactants allows additional control of
crystalline size; that is, the high ratio results in the formation
of more nuclei that subsequently result in the formation of
zeolite nanocrystals.33 Besides, the choice of template also
strongly influences the size of crystals. For instance, in the
synthesis of AlPO-n materials, amines usually produce
micrometer-sized crystals, whereas quaternary ammonium
salts lead to the formation of AlPO-n nanocrystals.5,34,35
Synthesis of zeolite nanocrystals is generally accomplished
through lowering the synthesis temperature and prolonging
the synthesis time. An example of how the synthesis tempera-
ture and treatment time affect the crystallite size of silicalite-1 is
given Table 2. As can be seen, the size of the crystals increased
from 20 to 40 nm by systematically increasing the synthesis
time at a constant synthesis temperature (60  C). On the other
hand, the crystal size increased more significantly at higher
temperatures (7080  C), that is, from 20 to 80 nm.6
Furthermore, aging under ambient condition has a pro-
nounced effect on the subsequent crystallization process. Dur-
ing aging, structural rearrangement in the precursor suspensions
occurs, which leads to the formation of zeolite nuclei. As a
result, the crystal size, induction period, and crystallization
time decrease upon lengthening of the aging process.
Another parameter that plays a significant role in the syn-
theses of zeolite nanoparticles is the type of heating applied to
the clear precursor suspensions. The synthesis of zeolite nano-
crystals can be performed in conventional air-driven ovens,
through reflux, and in microwave ovens. The general trend in
Zeolite Nanoparticles 287
Table 2 Synthesis of silicalite-1 nanocrystals
Zeolite T ( C) Time (h) Crystal size (nm)
Silicalite-1a 60 240 20
60 288 25
60 360 40
70 240 58
80 96 80
a
Precursor suspension: 9 TPAOH:0.16 NaOH:25 Si:495 H2O:100 EtOH.
the syntheses of zeolite crystals using the three heating systems
is shown in Figure 1. Generally, microwave irradiation pro-
vides the fastest crystallization rate and the smallest crystal size
in comparison to those prepared using reflux and conventional
heating. On the other hand, the synthesis through microwave
irradiation, reflux, and hydrothermal treatment under opti-
mized conditions renders particles with desired morphologies,
high purity, and colloidal stability. An example of how the size
of zeolites can be changed by replacing the slow conventional
heating with fast and short heating under microwave irradiation
is shown in Figure 1. Zeolite nanocrystals with spheroidal,
square, plate, orthorhombic crystals, and spheres can be synthe-
sized via optimizing the crystallization conditions.
Precursors such as alkoxysilanes and metal alcoholates pro-
vide reactants in molecular form which favor the synthesis of
microporous materials. The uses of different sources of reac-
tants (silica, alumina, and phosphate) play a significant role in
the nucleation and crystallization processes, and therefore they
influence the growth and morphology of the final material
(Figure 2). The silica source can influence different aspects of
zeolite crystallization and it leads to changes in the properties
of the final product. Some of the important parameters de-
scribing the process of zeolite crystallization such as the nucle-
ation and crystallization rates depend on the dissolution of the
silica precursors. The fragile silicate intermediates released dur-
ing the process of the silica source dissolution play an impor-
tant role in the zeolite formation as well. Besides the effects on
the formation of a particular zeolite, the silica source can
influence the particle size and shape of the crystals.3638 The
impurities introduced by the silica in the starting system can
also affect the properties of the zeolite. The substitution of
tetraethyl orthosilicate (TEOS) by colloidal silica in the synthe-
sis of nanosized silicalite-1 was found to prolong the duration
of the nucleation period.21 The nucleation rate is faster for
silicalite-1 when TEOS is used as a silica source compared to
amorphous silica (colloidal or fume silica). Depending on the
silica source employed, the size of the silicalite-1 nanocrystals
increases in the following order TEOS (15 nm) ! Cab-O-Sil
(25 nm) ! Ludox LS 30 (50 nm).
Several microporous materials have been prepared in the
form of colloidal suspensions with narrow particle-size
distribution.39,40 The synthesis of zeolite crystals with equal
particle radius requires a homogeneous distribution of the
viable nuclei in the system. Therefore, the homogeneity of
the starting system and simultaneity of the events leading to
the formation of precursor gel particles and their transforma-
tion into crystalline zeolitic material is of primary importance.
In order to obtain such homogeneous starting systems, abun-
dant amounts of TAA hydroxides and water are employed. On
288 Zeolite Nanoparticles

Conventional heating

Hydrothermal

Size (nm)
Magn 50 m
Reflux 301x lar417

Beta zeolite
Microwave heating
Microwave

Time

Min Days

Reflux heating

Figure 1 Trends in the change of zeolite dimension and morphology using different methods of heating (microwave, reflux, and conventional heating).

(a) (b) (c)

Figure 2 SEM micrographs of silicalite-1 nanocrystals synthesized using (a) TEOS, (b) Ludox LS-30, and (c) Cab-O-Sil as silica source.
Scale bar: 500 nm. From Fig. 6 in Mintova, S.; V. Valtchev, V. Micropor. Mesopor. Mater. 2002, 55, 171.

parameters used for characterization of zeolite nanocrystals

are diameter, colloidal stability, crystallinity, porosity, surface
Intensity (a.u.)

100 nm
5 10 15 20 25 30
2 (degree)
Size (nm)
Figure 3 Characterization of LTL zeolite nanocrystals with XRD,
HRTEM, and DLS. Modified from Fig. 2c in Wong, J. -T.; Ng, E.-P.; Adam,
F. J. Am. Cer. Soc. 2012, 95, 805808.
the other hand, the content of alkaline cations is very limited.
All these factors together with the careful choice of the reac-
tants allow the stabilization of clear starting mixtures where
only discrete gel particles are present.38,41,42
Once the zeolite crystals are prepared with the desired size
and shape, the emphasis is on isolating a nonagglomerated
form of the particles and also on their characterization mainly
in liquid form as well as nanosized powder. Common
chemistry, etc. Very often, the standard techniques used for
characterization of inorganic crystalline materials cannot pro-
vide conclusive information on the crystallization process of
35
zeolite nanocrystals. Therefore, the primary particles (diameter
smaller than 10 nm), their aggregation, and further transfor-
mation into zeolite crystallites in most cases can be character-
ized by complementary techniques such as spectroscopy
(infrared (IR), Raman, and nuclear magnetic resonance
(NMR)), small-angle x-ray diffraction (XRD), high-resolution
transmission electron microscopy (HRTEM), and dynamic
light scattering (DLS).524,2933 In Figure 3, DLS, HRTEM,
and XRD data for Linde Type L (LTL)-type zeolite nanocrystals
are shown.
The templating concept for preparation of zeolite nanocrys-
tals is further developed for other organic additives based on
metal complexes, which allow to form the zeolite and at the
same time to introduce desired metals with possible applica-
tions in catalysis.43,44 Metalamine complexes [M(NH3)4]2
are used as templates for zeolite nanocrystals as they carry a
high positive charge density and interact with the anionic

silicate species. Besides, they have different shapes (square,
planar, or linear), which are not common for classic quater-
nary ammonium templates. The use of metal complexes was
recently exemplified in the preparation of Edingtonite (EDI)-
type zeolite nanocrystals using [M(NH3)4]2, where M is Cu,
Pd, Pt. It is observed that the square-planar Pd and Pt amine
complexes having the same geometry reinforce their role as
templates and also co-template the formation of EDI- and
FAU-type zeolites.43 The Pd and Pt complexes also lead to a
fast nucleation in the precursor aluminosilicate suspensions,
resulting in an exceptionally small crystalline particle with size
below 20 nm. The copperamine complexes act as templates at
temperatures lower than 100  C, thus avoiding the decompo-
sition of thermally unstable complexes.
In summary, the templated synthesis approach offers sev-
eral advantages, which include the (1) possibility to control
overall particle size, (2) colloidal stability, (3) morphology,
and (4) surface reactivity. Nevertheless, this approach is not
easily scalable due to the use of a large amount of expensive
and in some cases toxic organic templates. Besides, the crystal-
line yield from the templated clear precursor suspensions is
usually very low (<10%).
5.10.2.1.2 Template-free synthesis approach
The synthesis of zeolite nanocrystals from template-free pre-
cursor suspensions is highly desired since it opens an alterna-
tive route for preparation of nanosized materials. This
approach, however, could be applied only in the synthesis of
low-silica zeolites crystallizing from highly reactive precursor
gels without employing organic SDAs.
In the template-free synthesis, the alkali metal cations such
as Na and K act as SDAs. In these precursor suspensions, the
type of cations strongly affects the nucleation and crystalliza-
tion process. For example, a potassium aluminosilicate zeolite
(zeolite L) is formed when K is used in abundant amount,
while the high concentration of Na provokes the formation of
zeolite Y.
(a) (b)
(d) (e)
Figure 4 HRTEM images of low-silica zeolite nanocrystals prepared without organic templates (a) LTA, (b) FAU, (c) LTL, (d) EMT, (e) MFI, and (f) SOD.
Scale bar: 100 nm. (c) From Hu, Y.; Liu, C.; Zhang, Y.; Ren, N.; Tang, Y. Micropor. Mesopor. Mater. 2009, 119, 306314.
Zeolite Nanoparticles 289
Typically, the synthesis of zeolite nanocrystals in template-
free suspensions is performed at a relatively low temperature
(<100  C).31,4552 After completion of the crystallization reac-
tion, the nanocrystals with diverse particle-size distribution
and high yield (>80%) are obtained. A substantial reduction
in the synthesis time combined with full conversion of the
initial amorphous system into zeolite nanomaterial under
low temperature has been developed for several zeolite struc-
tures including LTL, FAU, EMT, SOD, MFI, and LTA
(Figure 4).31,4552
Various silica sources with different specific surface areas,
impurities, and abilities to dissolve in alkaline mixtures are
employed. The vigorous mixing of the alkaline silicate and
aluminate solutions produced a precursor gel where all com-
ponents are expected to be homogeneously distributed. Usu-
ally the lowest possible temperature for a particular zeolite is
employed in order to favor the nucleation over the growth and
thus smaller crystals to be obtained. This example illustrates
the utility of the clear solutions in the investigation of different
populations of precursor particles involved in the crystalliza-
tion process. Contrary to the conventional gel systems, where a
large diversity of (alumino) silicates species is usually present,
the initially clear solutions contain a limited number of well-
defined discrete amorphous precursor particles.
The T (Si, Al) elements are introduced in the reaction mix-
ture as easily dissolvable sources. Most often, amorphous hy-
droxides, hydrous oxides, or related solids are employed. These
solids may be introduced in different physical states, for in-
stance, precipitated gels, ground glasses, volcanic ashes, colloi-
dal suspensions, or fumed silicas. Primary reactants are also
used to prepare such solids. Among them, the ones most often
used are alkaline silicate solutions, halides, silicon fluoro com-
plexes or alkoxides, aluminum salts, or aluminate solutions.
The silica content in the zeolite framework determines the
basic characteristics of the zeolite, namely, thermal stability,
hydrophilichydrophobic properties, amount and distribution
of the active sites, ion-exchange properties, etc. Pure siliceous
(c)
(f)
290 Zeolite Nanoparticles
zeolitic materials have a neutral framework. Isomorphous sub-
stitution of Si4 with Al3 introduces a negative charge in the
zeolite framework, which is compensated by the cations located
in the channels and voids of the structure. The level of this
substitution (the Si/Al ratio) determines the density of the active
sites and the ion-exchange capacity of the particular zeolite.
Numerous ways for reducing the mean particle size of
zeolite with LTA-type structure, keeping the economic impera-
tives in mind, have been presented.53 A study of the effect of
type of heating, postsynthesis treatment, and ultrasonication
on the size reduction in LTA crystals has been explored. It was
found that the application of ultrasonication did not lead to a
decrease in particle size, while the crystal morphology has
changed only. However, the microwave radiation is leading
to a narrow particle-size distribution of the zeolite with smaller
mean diameter. It is apparent that the rate of gel dissolution,
the distribution of the germ nuclei, and the crystallization time
play a significant role in determining the particle-size distribu-
tion and size of the crystals. Agitation promotes the formation
of larger crystals, although the type of agitation, namely, stir-
ring versus ultrasonication leads to completely different mor-
phologies. Aging leads to a substantial decrease in the spread of
particle size, which can be seen as a consequence of the relatively
low energy of nucleation. The template-free approach is opti-
mized to obtain nanosized LTA-type crystals.20 The high alka-
linity (high Na2O and/or low H2O) of the precursor suspension
is able to reduce the crystallization time from 2 weeks to 3 days.
Another zeolite, NaX, with controlled particle size from 20 to
800 nm at 60  C, is synthesized from template-free precursor
suspensions.45 It is shown that the silica source and the hydro-
thermal conditions including crystallization temperature and
agitation govern the crystallization rate of the zeolite nanocrys-
tals. Additionally, the crystallization temperature has a pro-
nounced effect on the ultimate zeolite crystal size. Highly
crystalline zeolites A and X were obtained within 3 and
21 days, respectively.46,52 Both zeolites have nanosized particles;
however, the morphological appearance differs substantially.
Single zeolite A crystals (80250 nm) were obtained, whereas
zeolite X crystallized in the form of spherical aggregates
(200300 nm) built of very small 2040 nm crystallites. The
relatively rapid transformation of the amorphous aluminosili-
cate species into zeolite under room temperature conditions was
achieved by a fine-tuning of all parameters affecting the crystal-
lization kinetics. It is also found that the postsynthesis ultrasonic
treatment has disintegrated the loosely attached particles and
provided a product with a relatively narrower particle-size dis-
tribution. The fraction of the silica converted into zeolite nano-
crystals of A and X was 75% and 83%, respectively.
In conclusion, the synthesis of zeolites from organic-
template-free precursors is an alternative approach for prepa-
ration of nanosized crystals. This method is environment
friendly since the synthesis does not employ any harmful
organic templates. Moreover, the absence of organic template
implies that no high-temperature calcination step is required
for opening up the pore system for the intended applications,
and thus this method is considered beneficial for both the
environment and the scaling-up through reducing the produc-
tion cost and reducing the use of harmful chemicals. However,
stable colloidal suspensions of these nanocrystals can be
prepared under additional treatment.
5.10.2.1.3 Seed-induced synthesis
The use of preformed zeolite seeds as nucleation centers in the
synthesis has a long history.54,55 Ideally, the advantage of this
method is that the nucleation of the desired phase is stimulated
and the growth process results in nanosized zeolites with a
narrow particle size and desired morphology. This method is
used to suppress the formation of undesired phases and enable
the intergrowth of different crystalline phases.56,57 The main
importance of this approach is the complete elimination of
organic templates with calcined or noncalcined seed crystals
and the formation of zeolite nanoparticles with high yields.
Successful preparation of MFI and BEA nanoparticles using
seeding method has been reported by several groups.48,5860 In
all cases, zeolite seeds (0.110 wt%) were used as nuclei for
further crystal growth. For example, Al-rich zeolite BEA with Si/
Al ratio as low as 3.9 is synthesized starting from the precursor
suspension without organic SDA, but in the presence of zeolite
seeds. In some cases, the zeolite seeds were partially dissolved
and the amorphous entities nucleate and grow with the help of
these preformed seeds. The product yields are above 80% and
the resulting nanocrystals do not require calcination to open
the zeolite porosity, and, thus, the aggregation between zeolite
particles is avoided during the postsynthesis treatment. Be-
sides, the Mobil Twelve (MTW)61 and Mordenite (MOR)62
nanocrystals have also been made at moderate temperature
(160180  C) using the seed-induced approach.
In some cases, the zeolite nanocrystals synthesized by seed-
ing of an organic-template-free initial gel are aggregated in
comparison to the particles synthesized from clear suspension
rich in organic templates. The high aggregation level is most
probably due to the nanoseeds, which are polycrystalline ag-
gregates built of much smaller crystallites. Secondary growth of
closely situated nanocrystallites leads to the formation of com-
plex aggregates larger in size in respect to the seeding particles
(Figure 5). Consequently, a part of the product does not be-
have as colloidal matter, although the size of the individual
crystals is in the nanometric range.
Besides, functionalization of the zeolite seeds to obtain
zeolite nanocrystals with enhanced surface area and porosity
has been developed.63,64 This method is based on perturbing
(a)
Single seed Growth of
500 nm
growth aggregrate
20 nm
(b)
(c) (d)
100 nm
200 nm
Figure 5 SEM micrograph of (a) MFI-type nanocrystals used for
seeding of initial gels, (b) a sketch of a seed aggregate, and (c, d) the
result of the secondary growth of a monocrystalline particle and a
polycrystalline aggregate exemplified by TEM images, respectively.
Modified from Majano, G.; Darwiche, A.; Mintova, S.; Valtchev, V. Ind. &
Engineer. Chem. Res. 2009, 48, 7084.

the growth of the zeolite crystals by functionalization with the
hydrophobic organosilane group in order to hinder and pre-
vent their further agglomeration, according to the following
steps: (1) formation of the zeolite nuclei in synthesis gel during
the precrystallization step, (2) functionalization of the zeolite
seeds by reaction with organosilanes, which form a protective
organic barrier against aggregation, and (3) crystallization to
complete the zeolitization of the functionalized seeds. In ad-
dition, the organosilane can also influence the stability of tetra-
hedrally coordinated Al species during the nucleation process,
and thereby impact the crystalline process of nanozeolites.65 The
nanoproduct obtained by this method consists of ultra-small
zeolite nanocrystals (<50 nm) having an additional porosity in
the meso-/macropore regions generated by the presence of the
silanization agents (Figure 6). This synthesis approach also
opens new ways for the development of environmentally
friendly synthesis procedure for preparation of Al-rich zeolite
Beta and ZSM-5 without organic template.
5.10.2.1.4 Multistep synthesis approach
Multistep templated synthesis is another method for prepara-
tion of nanosized zeolites with high crystalline yield.6668 The
basis of this concept is the reuse of nonreacted chemicals
separated from the crystalline suspensions and subjected to
further crystallization with or without addition of chemicals.
The reuse of clear precursor suspensions for the multistep
synthesis approach is rational since (1) a low amount of che-
micals are consumed and (2) only zeolite nanocrystals are
recovered by centrifugation, so that very small zeolite crystals
(10 nm or less) are still present in the clear suspension. This
approach is applied for the preparation of colloidal NaY and
silicalite-1 nanocrystals.66,67 The cumulative product yield is
increased from 4% to 44% after several synthesis cycles. Fi-
nally, the crystallization process is terminated since the
Amorphous Functionalized
nanoparticles with organosilanes
Figure 6 Schematic diagram of the synthesis of zeolite nanocrystals with enhanced textural properties synthesized from organo-functionalized seeds.
From Serrano, D. P.; Aguado, J.; Escola, J. M.; Rodriguez, J. M.; Peral, A. Chem. Mater. 2006, 18, 2462.
1 2
H3PO4 H 2O
Multi
2 1 vs. single 1 3
synthesis
25 g
Experimental
AI(OiPr)3
Multi
1
3
Single
TEAOH
(a)
Figure 7 Total consumption of chemical reagents for the synthesis of nanocrystalline AlPO-18 by single- and multistep synthesis approaches in
(a) laboratory (experimental) and (b) industry (calculated) scales production. From Fig. 5 in Ng, E.-P.; Delmotte, L.; Mintova, S. Green Chem. 2008, 10,
10431048.
Zeolite Nanoparticles 291
precursor suspension is discarded after several cycles when
the silica and alumina nutrients are insufficient in the precur-
sor suspensions for further syntheses.
It is very important to note that the zeolite nanocrystals
prepared from several synthesis cycles have the same proper-
ties, that is, chemical composition, colloidal stability, particle
size, morphology, and porosity.
This approach is also considered environmentally benign
since the nonreacted chemicals are reused and thus the ap-
proach is considered beneficial for both the environment and
the scaling-up through reducing the production cost and dis-
posal of chemical waste. Besides, the conventional chemical
process for preparation of nanosized microporous molecular
sieves, which is based on the use of large amounts of toxic
reactants (amine/ammonium salt and phosphoric acid) and
volatile solvents (ethanol and methanol), is optimized by
applying the multistep synthesis approach.
The multistep synthesis approach is also applied for prepa-
ration of nanosized aluminophosphates.68 Unlike the alumino-
silicate system, the multistep synthesis can proceed many times
with a minimal chemical compensation of the reacting mixture
after recovering the crystalline nanoparticles from each step.
Thus, almost complete consumption of the organic templates
and inorganic species without disposing harmful reagents to the
environments, and making possible the scaling-up process of
nanocrystalline zeotype materials, is reported (Figure 7).
5.10.2.1.5 Ionothermal synthesis
Ionic liquids (ILs) as green solvents have shown great promise
as an attractive alternative or replacement to conventional
volatile organic solvents, attributed to the following distinct
features, that is, negligible vapor pressure, better thermal sta-
bility, tunable hydrophilibility/hydrophobicity, ease of recir-
culation, and manipulation.69
Crystallization of Zeolite
amorphous particles nanocrystals
1 3
H 2O
H3PO4
Multi
vs. single
4 1 synthesis
1 13
1 ton
Estimated
AI(OiPr)3
Multi
1
5
Single
TEAOH
(b)
292 Zeolite Nanoparticles

IL
AIO4 PO4+ IL O

IL IL Mn2O3
O
T T = AI, P, Mn

IL IL IL
O O

IL PO4+ IL
IL IL H H
IL H
AIO4
C
IL H C
H
AIO4 IL
IL IL H
H
C C
N
C
H
IL MnO2 H N C
IL IL AIO4 IL H
IL PO4+ IL IL
C
H
IL H H
IL IL

Homogenization Dissolution Formation of intermediates

5 min 5h 25 h
Mixing Induction

IL IL
IL IL
IL IL H H IL H H H
IL H H
H
H
IL H
H C
C
H
IL H
H C
C
H IL H
H C
N
C
H

IL
N H IL H C C
N
C
H
IL H C C C
H

IL
H

H
C C
N
C

C
IL H N C IL H N C
H
H
H IL C
H IL C
H
C
H
IL H
H IL H
H

IL IL H
H
IL IL
IL IL
IL Liberation of
Crystal growth at Additional crystal growth discrete
Nucleation of MnAIPO-5
intermediate surface from surface to core nanocrystals
50 h
70 h 80 h 90 h

Crystallization

Figure 8 Schematic illustration of the crystallization pathway of MnAlPO-5 nanocrystals under ionothermal conditions. From Scheme 1 in Ng, E.-P.;
Itani, L.; Sekhon, S. S.; Mintova, S. Chem. Eur. J. 2010, 16, 1289012897.

Recently, the ILs are applied for the preparation of zeotype

nanocrystals and inorganicorganic hybrids materials.7072 (a) (b)
The so-called ionothermal technique uses an IL or eutectic
mixture as reaction solvent and as SDA. The most important
features of this strategy are the use of green IL instead of
harmful quaternary ammonium hydroxides, and the synthesis
can be performed in an open vessel instead of using autoclave
due to the low vapor pressure of ILs. Since then, many studies
have been reported on ionothermal synthesis of metal oxide
nanoparticles.7377 The synthesis of nanocrystalline zeolite (c) (d)
materials in ILs, however, is still a new field, which has
emerged over the last several years. A novel hexagonal AlPO
molecular sieve with a crystal thickness of 170 nm was first
prepared in the presence of HF.78 Later, discrete nanosized
MnAlPO-5 crystals using 1-ethyl-2,3-dimethylimidazolium
bromide ([edmim]Br) as IL media were synthesized.79,80 Com-
pared with conventional organic templates, the ILs are more
environmentally benign and recyclable.81 An example of the Figure 9 Zeolite A crystals synthesized under (a) conventional and (b)
microwave heating, and AlPO-18 nanoparticles prepared using (c)
crystallization pathway of MnAlPO nanocrystals is presented
conventional and (d) microwave treatments. Scale bar: 1 mm.
schematically in Figure 8. This approach can contribute to
green and sustainable chemical synthesis and production of
zeolite nanocrystals by improving product yield and so pro- crystallization processes coexist, resulting very often in a
ducing less waste, and avoiding environment pollution. broad particle-size distribution of products.82,83 By contrast,
the microwave heating causes internal heating of the precursor
suspension and very often a significant change of the kinetics
5.10.2.2 Types of Heating
and selectivity of ongoing reactions is observed. Mainly, lower
5.10.2.2.1 Microwave-assisted synthesis temperatures and shorter time for crystallization of the nano-
As compared to the conventional techniques, the microwave- sized zeolites are observed in the systems subjected to micro-
assisted hydrothermal method provides an efficient way for the wave heating compared to those under conventional heating.
synthesis of various zeolite nanocrystals and rational control The energy transfer to the precursors is achieved through the
on their particle-size distribution, yield, phase purity, and interaction of the microwaves with water or other compounds
morphology. In the conventional heating reactors (auto- with high dielectric constant or large dipole moments.83
claves), the time needed for heating the zeolite synthesis sus- Various zeolite nanoparticles such as AEI,34 AFI,84 CHA,85
pensions to a specified temperature is long due to the slow LTA,52,86 MFI,86 EMT,31 and BEA86 have been prepared by
heating rate. During this heating period, nucleation and microwave-assisted hydrothermal treatment (Figure 9).

Among them, a rapid synthesis of silicalite-1, BEA, ZSM-5, LTL,
and LTA nanocrystals by using microwave heating is
achieved.86 A comparison of conventional and microwave hy-
drothermally synthesized zeolite A nanocrystals indicates that
the conventional crystals are larger and highly nonuniform in
shape with secondary crystals apparently growing from pri-
mary crystals. By contrast, the particle-size distribution of
microwave-synthesized zeolites is visually smaller and more
uniform than the conventionally heated zeolites.
Synthesis of zeotype nanocrystals (AlPO-18, SAPO-34)
with narrow particle-size distribution under microwave radia-
tion has been reported.34,85 The chemical compositions of the
starting sols and the microwave hydrothermal synthesis con-
ditions exhibit synergic effect and lead to the formation of
nanosized crystals. The microwave synthesis is also reported
to produce zeolite nanocrystals with unique morphology.34 It
was also shown that the AlPO-18 nanoparticles are growing as
thin elongated hexagons, which are different from the conven-
tional square plate-shaped AlPO-18 crystals.
The effective synthesis of ultra-small (615 nm) crystals of
the large-pore zeolite EMT-type zeolite with very high yield
from template-free colloidal precursors at low temperature
(30  C) for 4 min under microwave irradiation and for 36 h
under conventional heating is reported (Figure 10).31 From an
environmental perspective, the synthesis of EMT zeolite is
extremely attractive as the nanocrystals can be easily synthe-
sized at very high yield at near-ambient temperature without
using any organic templates. This is suggesting that scale-up of
an energy-efficient synthesis would be easily feasible. These
nanoscale EMT materials offer exciting opportunities for both
fundamental study and potential industrial applications.
5.10.2.2.2 Microchannel-assisted synthesis
Similarly, fast heating microchannel reactors are used to syn-
thesize zeolite nanocrystals with a small mean particle size and
(a) (b)
10 nm 500 nm
(c) (d)
20 nm 500 nm
Figure 10 HRTEM pictures of EMT nanocrystals synthesized in (a, b)
conventional oven and (c, d) microwave oven. From Fig. 4 in Ng, E.-P.;
Chateigner, D.; Bein, T.; Valtchev, V.; Mintova, S. Science 2012, 335, 70.
Zeolite Nanoparticles 293
narrow particle-size distribution. Unlike normal conventional
heating, the microreactor with continuous operation has fast
heat and mass-transfer features, which aid in the synthesis of
uniform nanoparticles.87 It was shown that the crystallization
rate of the zeolite nanocrystals conducted in the segmented
flow microreactor is comparable to that in the microwave-
heated reactor.8891
A schematic diagram of the microfluidic system for the
synthesis of zeolite nanocrystals in microchannel reactors is
shown in Figure 11. The application of microchannel reactors
for production of zeolite A and silicalite-1 nanocrystals is
reported.8891 These reactors are suitable for continuous syn-
thesis of nanosized zeolites and show great advantages for
controlling both the reaction conditions and the crystal prop-
erties. However, one of the main drawbacks is that the micro-
channel reactor always suffers from channel blockage due to
the viscous aluminosilicate precursor suspensions commonly
used for zeolite synthesis.
5.10.2.3 Confined Space Synthesis
The confined space synthesis method is applied for prepara-
tion of highly dispersible zeolite nanocrystals.92 This method
involves the crystallization of the zeolite inside a matrix with
controlled porosity. Thus, the crystals cannot grow larger than
the pores of the used matrix.
Confined space synthesis has several advantages in compar-
ison with other methods. First, it is possible to predict the
maximal crystal size of the resulting zeolites by choosing an
appropriate porous matrix. Second, the crystallization of zeo-
lite nanocrystals in confined space proceeds until complete
consumption of the precursor suspension. Third, the zeolite
crystals can be isolated simply by calcinations at a sufficiently
high temperature to ensure a complete combustion of the inert
matrix. Moreover, the recovery of the zeolite is simple since
no filtration (high-speed centrifugation) for the nanocrystals
is involved which is a needed step in the syntheses of small
crystals. Finally, this method is highly reproducible and can
be used to prepare a wide range of zeolites and possibly
other materials.
To date, a variety of space confiners such as carbon blacks,9295
carbon nanotubes,96,97 starch,98 gelling polymer,99,100 and poly-
mer spheres101 have been utilized to confine the crystallization
of zeolite nanocrystals. A summary of the confiner, zeolite
nanocrystals synthesized within this matrix with sizes in the
range of 10300 nm is displayed in Table 3.
Heating bath
Liquid
paraffin
Nanozeolites
Liquid
paraffin
Synthesis Nanodroplets
solution
Figure 11 Synthesis of zeolite nanocrystals in microchannel fluidic
system.
294 Zeolite Nanoparticles

Table 3 Zeolite nanoparticles synthesized in confined space

Confiner Zeolite nanoparticles Size (nm)

Carbon black Na-ZSM-5, Silicalite-1, Beta, NaY, NaA, Sodalite, LTL 1330, 5563, 1015, 100, 230240, 3750, 30
Carbon nanotubes NaZSM-5, NaY 2030
Starch NaY 50100
Polymer hydrogels NaA, NaX 20180, 10100
Polymer spheres ZSM-5, Beta, TS-1, MORa, Aa, Xa, Ya, La 300, 90, 220
a
Crystallite size not reported.

Hydrothermal
Heating treatment

Precursor gel with Precursor gel trapped Zeolite nucleation

gelling polymer inside hydrogel and growth

Cooling down

Purification
and re-dispersion

Colloidal zeolite suspension Zeolite nanocrystals in

gelling polymer solution
Figure 12 Schematic representation of template-free zeolite synthesis using thermoreversible polymer hydrogels.

A slightly modified approach involves the application of
thermoreversible polymer hydrogel as confined space for limita-
tion of crystal growth of low-silica zeolites (Figure 12).99 In
particular, the polymer hydrogels are of interest as they are con-
sidered a soft space confinement additive. The three-dimensional
pores of the polymer hydrogels can be adjusted and can serve as
microreactors or nanoreactors for controlling the zeolite growth.
Unlike natural starch polymer and other carbon confine spacers,
the thermoreversible gelling polymers can be readily removed
after the synthesis by simple washing and the zeolite nanocrystals
obtained are readily redispersed in various solvents.
Besides, polymer spheres have been used in the preparation
of zeolite nanoparticles. These confined voids can be simply
adjusted by the solid content and diameter of the polymer
spheres. The adjustable confined voids formed by polymer
spheres potentially serve as micro- or nanoreactors for control
growth of zeolite nanocrystals. So far, ZSM-5, A, X, Y, L, MOR,
Beta, and TS-1 nanocrystals have been synthesized using this
approach.101
5.10.2.3.1 Reverse microemulsion synthesis
The reverse microemulsion or water-in-oil microemulsion syn-
thesis approach consists of aqueous domains (termed reverse
micelles) dispersed in a continuous oil phase solvent, so that a
variety of reactants can be introduced into the nanometer-sized
aqueous domains (Figure 13). By the reaction confined within
the reverse micelles, nanostructure materials with controllable
particle size, shape, and higher colloidal stability are obtained.
Recently, reverse microemulsions have been extensively
applied in the synthesis of nanocrystalline solids.102107
Various parameters including initial synthesis composition,
microemulsion composition, surfactant identity, and electro-
lyte that affect the morphology and size of silicalite-1 crystals
are investigated.108 It was observed that the formation of zeolite
nanocrystals is mainly dependent on the type of surfactants.
Microemulsions formed with a straight-chained surfactant
(cetyltrimethylammonium bromide (CTAB)) lead to amor-
phous silica, whereas microemulsions formed with a branched
surfactant lead to the formation of silicalite-1 nanocrystals.
The morphology of silicalite-1 nanocrystals is well defined.
These results also indicate that the silicalite-1 nanoparticles do
not nucleate in the microemulsion, but rather nucleate and
grow heterogeneously after amorphous silica particles in the
microemulsion phase are formed.
Zeolite nanoparticles with the MFI-type structure using
a nonionic emulsion system are synthesized.108 The results

Mixing and
heating
Zeolite precursor
Emulsion system
Hydrophilic
Oil phase
Figure 13 Schematic illustration of zeolite formation process in the microemulsion system.
reveal that the emulsion system allows rapid crystallization of
ZSM-5 in comparison with the conventional hydrothermal
synthesis. Zeolite A nanocrystals with narrow particle-size
distribution (100120 nm) and sphere-shaped morphology
are prepared in the presence of a cationic microemulsion.109
The microemulsion system has been proven to be able to
accelerate the crystallization process of the LTA nanoparticles.
However, an increase in the concentration of the surfactant
and co-surfactant results in a decrease in the crystallinity of
the LTA-type zeolite.
In spite of the desirable characteristics of zeolite nanoparti-
cles, reverse microemulsions have the drawback of relatively
low crystalline yield.110 In addition, this method uses excess of
solvents to stabilize the microemulsion system and compli-
cates the recycling of the solvents. The high consumption of
solvents not only causes detrimental effects to human health
but also requires significant waste disposal. This is a very
important issue when scale-up of this approach is considered.
5.10.2.4 Other Methods
5.10.2.4.1 Direct-conversion synthesis approach
Natural rocks such as bentonite, clinoptilolite, stilbite, erionite,
and kaolinite are minerals applied as precursors for preparation
of zeolite nanocrystals. Prior to the synthesis, the minerals are
treated with acid to remove impurities before dissolving the
aluminosilicate residue in concentrated alkaline solutions. The
conversion of natural bentonite to zeolite Y nanocrystals has
been reported.111 Also nanosized zeolite A is successfully
obtained from natural clinoptilolite via direct conversion.112
The direct conversion demonstrated to be simple, and the pro-
duction cost is low in comparison to the conventional templat-
ing approach. This method is also used for preparing zeolite BEA
nanocrystals starting from FAU zeolite under hydrothermal
treatment with and without SDA.113 The crystallization rate of
BEA zeolite is significantly enhanced and also demonstrated the
possibility of synthesizing other template-free zeolite nanocrys-
tals. Nevertheless, the zeolite nanocrystals derived from natural
minerals rocks have more impurities and therefore they are not
useful for some applications including semiconductor industry,
medical, and pharmaceutical applications.
Zeolite Nanoparticles 295
Crystallization
Aging Purification
Zeolite
nanocrystals
Magnify
Zeolite
head nanocrystal
Hydrophobic
tail
5.10.2.4.2 Centrifugation-assisted grinding
The centrifugation-assisted grinding method is considered ef-
fortless and convenient, and permits the preparation of zeolites
with tunable morphology and fairly narrow size distribution.
This method is applied in the production of MFI and FAU
nanocrystalline solids without organic templates.114 In a typi-
cal procedure, a known amount of micrometer-sized zeolites
dispersed in water are mechanically treated to smaller crystals.
The nanocrystals with various sizes are separated using centri-
fugation, and a very high crystalline yield (above 85%) is
achieved. However, the nanocrystals might face low mechani-
cal stability and some of the zeolite nanocrystals are destroyed
during the grinding treatment. This approach is improved by
combining a bead mill grinding and recrystallization, and
demonstrated for the preparation of nanozeolite A free of
organic templates.115 Conventional grinding methods such as
ball milling and planetary ball milling downsize the zeolites
but the destruction of the outer zeolite framework may cause
pore blocking, and this impedes the desirable properties of the
zeolite.116 Therefore, a milder grinding method is required to
prevent the degradation of crystallinity, for example, amorphi-
zation and/or formation of dislocations, by the use of small
beads (30500 mm in diameter). The damaged part as a result
of bead milling treatment, however, can be subjected to recrys-
tallization and repaired using a dilute aluminosilicate solution
(Figure 14). Under these conditions, the poorly crystalline parts
of the milled zeolite particles are more easily dissolved than the
crystalline parts, and tend to be recrystallized into zeolites.
5.10.2.4.3 Laser-induced fragmentation method
Laser-induced fragmentation of micron-sized into nanosized
particles has been developed to reduce the dimensions of
colloidal metal particles.117119 Recently, this approach has
been applied in the preparation of LTA nanocrystals.120 The
laser-induced fracture method is valuable because it can pro-
duce nanoparticles directly and rapidly from many commer-
cially available zeolites. In addition, no organic SDAs are
involved in the synthesis, thus avoiding the calcination proce-
dure. However, the appropriate wavelength and laser energy
density have to be used in order to reduce the possible damage
to the zeolite crystalline structure during the fragmentation.
296 Zeolite Nanoparticles
Milling/grinding
34 m
50200 nm
Micronsized LTA
Crystalline
phase
Figure 14 Schematic illustration describing the fabrication process of zeolite nanocrystals by bead milling and postmilling recrystallization.
5.10.2.5 Separation of Zeolite Nanocrystals
Special attention is paid to the stability of zeolite nanocrystals
during purification process. The colloidal suspensions contain-
ing nanocrystalline zeolites are purified by (1) high-speed
centrifugation, (2) membrane filtration, or (3) coagulation.
Typically, after the hydrothermal treatment of the precursor
suspension, the crystalline product is subjected to a series of
high-speed centrifugation steps (for instance, 1 h at 47.800 g
relative centrifugal force (RCF) or 25 000 rpm) and subse-
quently redispersed in water under sonication (12 h in ice
bath) until the pH of the crystalline purified suspension reaches
about 7.59. Alternatively, colloidal suspensions containing ze-
olite nanoparticles are purified by dialysis using a dialysis tubing
device with different sizes. The advantage of this method is that
no additional agglomeration of the particles is provoked as it is
observed often during the centrifugation treatment. In addition,
the purified and stabilized aqueous suspensions of zeolite nano-
particles keep their original solid concentrations. At pH of the
colloidal suspensions in the range of 910, the concentration of
solid particles can be varied from 1 to 10 wt%. However, during
the prolonged dialysis, the unreacted organic template present
in the colloidal suspension can be removed almost completely
and then the zeolite nanoparticles will become unstable.
Therefore, postsynthesis treatment has to be considered leading
to disintegration of the particles and keeping as discrete in the
colloidal suspension.
Coagulation is the third process used to precipitate out the
zeolite nanoparticles of colloidal suspensions. The coagulation
of the zeolite nanoparticles can be carried out by mixing two
oppositely charged compounds. Both compounds may be par-
tially or completely precipitated. This approach also can be
accomplished with boiling of the zeolite suspensions. This
reduces the charge on the particles and ultimately they settle
down to form a precipitate where the particles stay discrete and
do not agglomerate irreversibly.
Different electrolytes are used for coagulation of zeolite
nanocrystals; these electrolytes have different coagulation
values, and the smaller the coagulation value of the electrolyte,
the larger is its coagulating power. The coagulation of zeolite
nanoparticles is governed by two factors: (1) ions carrying
charge opposite to that of the zeolite particles are effective in
bringing about coagulation and (2) coagulation power of an
electrolyte is directly proportional to the valency of its ions.
The zeolite nanoparticles can be coagulated by mutual
precipitation, repeated dialysis, and heating. This is performed
Recrystallization
50200 nm
Amorphous
phase
Highly crystalline
nanosized LTA
High
Rate of flocculation
30 0 +30
Zeta potential, mV
Figure 15 Variation of zeta potential values of zeolite suspensions.
by adding the electrolyte and by choosing a suitable solvent.
For instance, the as-prepared zeolite nanoparticles (BEA and
MFI) can be treated with an aqueous solution of diluted am-
monium chloride, and subsequently washed with ethanol to
prevent further irreversible agglomeration.121
The stability of the zeolite nanocrystals in different solvents
is of significant importance for their applications. The stability
is usually governed by intermolecular interactions such as
electrostatic, van der Waals, and London forces, between the
particles and the solvents.122 The stability of colloidal suspen-
sions is determined by measuring the zeta potential values at
constant pH and constant solid concentration (Figure 15).
This refers to the electrostatic potential generated by the accu-
mulation of ions at the surface of a colloidal particle that is
organized into an electrical double layer, consisting of a stern
layer and a diffuse layer. Flocculates and/or aggregates are
formed if the coulomb interactions are lower than the van
der Waals forces between the particles, and finally the sedi-
mentation of particles also causes a substantial increase in the
zeta potential value.
The stability of zeolite nanoparticles is described by their
resistance to coalescence and aggregation. Two or more small
particles fusing together and form a single large particle that is
described as coalescence. The essential feature of the coales-
cence is the fact that the total surface area is reduced. Aggrega-
tion is a process by which small particles clump together
(aggregates) but do not fuse into new entities. There is no
 Zeolite Nanoparticles 297

RO

RO
OH
H OR

RO
OH

RO
OR
H
O O

OH
H Esterification RO OR RO OR
O OH OH
with alcohols

O
RO

OR
OH

OR

OR
R
OH

OH

OR
OH
O

H
OH
H
O

O
H

O
O

OH

H
OH

OH

RO
H

RO
OH O

RO
OH

R
OH

O
RO
RO
OH OR

OH
OR

OH
HO
OR RO
HO RO

OR
RO
HO
HO

OR
OR

OR
Aggregation of zeolite nanoparticles Stabilization of zeolite nanoparticles
due to hydrophilic interactions due of hydrophobic surfaces
Figure 16 Effect of chemical posttreatment on the stability of zeolite nanocrystals.

reduction of the surface area, although certain surface sites may
be blocked at the points at which the smaller particles touch.
The term coagulation is also used to describe the process of
aggregation; colloids stable against coagulation and coales-
cence are considered kinetically stable. The stability of the
colloids is described by the extent of which small particles
remain uniformly distributed through the sample. The stability
of zeolite nanoparticles depends on their size, solid concentra-
tion, surface charge, and type of solvent.
The effect of surface chemical treatment on the degree of
aggregation of colloidal particles is important. An esterification
step of the surface of zeolite nanoparticles can be performed
after purification in alcoholic media, which has resulted in the
formation of estersil and hydrophobization of the particles
according to the following scheme:
SiOH + ROH SiOR + H2O
Zeolite surface Alcohol Zeolite covered
by estersil
This procedure strongly disaggregates and stabilizes the par-
ticles in alcohols such as ethanol, iso-propanol, and n-butanol at
certain pH of the suspensions (Figure 16). The posttreated
particles are between 10 and 150 nm, while the nontreated
particles are between 200 and 500 nm. A redispersion of the
posttreated particles in toluene or benzyl alcohol leads to a
completely clear solution, indicating the presence of very small
particles. This indicates that the aggregation level depends not
only on the effect of silanol groups at the surface of particles, but
also on the type of solvents. The disaggregationreaggregation
process of the esterified particles is a reversible process and
depends on the surface chemistry of the crystals.28
5.10.3 Applications of Zeolite Nanoparticles
The possible green mass production of zeolite nanocrystals
provides excellent opportunities for applications in catalysis,
adsorption and separations involving larger molecules, and for
designing thin films, membranes, and nanoscale devices.
5.10.3.1 Zeolite Membranes
Over the last two decades, the development of zeolite-based
membranes has attracted considerable research efforts. The
supported membrane based on zeolite nanocrystals was first
reported in the 1990s. Since then, many types of zeolite
membranes (LTA, MFI, LTL, SOD, BEA, AFI, and MWW) de-
rived from zeolite nanoparticles have been prepared.19,123130
Zeolite membranes are used for gasvapor and liquidliquid
separation. Both groups are important from industrial and
environmental points of view. For instance, the CO2 separation
from different gas mixtures is among the most serious environ-
mental problems nowadays. Consequently, numerous studies
have been performed to extract hydrogen from H2/N2, H2/O2,
H2/CH4, and multicomponent gas mixtures. Liquidliquid
separation on zeolite membranes is a promising alternative to
such an energy-intensive process as the distillation. Different
alcoholwater (methanol, ethanol, propanol, butanol, etc.) mix-
tures have been subjected to separation by zeolite membranes.
Additionally, sodalitepolyimide membranes with good hy-
drogen permeability and selectivity over nitrogen at the same
testing temperature have been prepared.128 In addition, crystals
with controlled orientation have been shown to exhibit new
properties, since their porous framework is capable of orienting
and restricting rotation of the adsorbed guest molecules.126 A
significant improvement in the separation of p-/o-xylene mixture
in c-oriented MFI membrane compared with the conventionally
calcined membranes has been reported.130 Also, the oriented
zeolite membrane prepared via secondary growth gave excellent
cumulative O2 permeance.131 A number of comprehensive
reviews devoted to the preparation and characterization of
zeolite-based membranes have been published.123,132137
The great expectations related to zeolites as separation
media are based on their selectivity, long-term stability at
high temperature, resistance to harsh environments, resistance
to high-pressure drop, inertness of microbiological degrada-
tion, and easy cleanability and catalytic activation.
The preparation of zeolite membranes, in particular on a
large scale, remains a challenge. The major disadvantages of
zeolite membranes are poor processability and mechanical
stability, and very often poor fluxes and technical difficulties
related to sealing at high temperature. Substantially higher
costs of zeolite membranes are also an obstacle to their large
industrial utilization. Obviously, the processing methods will
have to be further developed in order to ensure an industrial
application of supported zeolite membranes.
5.10.3.2 Optical and Other Devices
There has been a considerable amount of work toward the
development of other applications for zeolite nanocrystals
298 Zeolite Nanoparticles
such as optical devices, films with low dielectric constant, and
gas/liquid sensors. A comprehensive knowledge for prepara-
tion of zeolite crystals with desired shape, size, morphology,
and crystalline structure and their subsequent assembly in
thin-to-thick films is of great help in the design of these novel
applications. Moreover, the mechanical and thermal stability
and physicochemical properties of the zeolites assembled in
the films are of significant importance.27
Many fabrication methods are used in the production of
zeolite-based chemical sensors. Factors that must be considered
when selecting the production technique include their expense,
purity, porosity, reliability, and reproducibility. Common tech-
niques for preparation of sensors include (1) screen printing,
(2) solgel techniques, (3) dip and spin coatings, and (4) direct
growth with and without preseeding of the substrates. Indeed,
the completely controlled zeolite structures, fine-tuned chemical
composition, and ion-exchange capacity make them attractive
as true shape-selective compounds.138 Considerable progress
has already been reached in this field by assembling zeolites in
thin films either by controlled attachment on self-assembled
monolayers followed by growth or by different patterning
techniques.
The zeolite-based chemical sensors are divided into two
groups depending on the respective role of the molecular
sieves. The zeolite can act as a main functional element,
which is the case for sensor principles relying directly on con-
ductive, adsorptive, or catalytic properties of one specific type
molecular sieve and its interaction with the analytes.
The second group includes devices where the zeolites are sup-
plementary or secondary elements.
High sensitivity, good reversibility, and long life of zeolite-
based sensors were demonstrated for detection of hydrocar-
bons and water at low concentrations. Such materials are of
interest in biochemical and sensing systems.19,84,139 Besides,
optical sensors based on acrylamide photopolymer doped with
zeolite nanoparticles (BEA, LTA, and AEI) have also been de-
veloped.140,141 These holographic sensors have shown a higher
dynamic range, a higher ultimate refractive index modulation,
and a lower level of shrinkage.142 The ultimate modulation of
the optical refractive index (n1) is caused by polymerization of
the monomer and its conversion into polymer, density
4
Intensity 5 mW cm2
Refractive index modulation 103
3
2 0.6
1 1.2 wt. % 40 nm
0 0.0
50
(a) Time, s
Figure 17 (a) Refractive index modulation of films with thickness of 40 mm doped with 1.2 wt% of zeolite nanoparticles with a size of 40 and 60 nm;
(b) Angular selectivity curves for nondoped (gray) and doped (black) photopolymer with nanocrystals with a size of 60 nm.
variation due to concentration-driven monomer diffusion
from dark to bright fringe areas, density variation due to
concentration-driven short/mobile polymer chain diffusion
from bright to dark fringe areas, and spatial patterning of the
zeolite nanosized particles. The modulation of the optical
refractive index is observable only if the refractive index of
the nanoparticles is significantly different from that of the
photopolymer matrix. The results shown in Figure 17 reveal
that an increase in the refractive index modulation is observed
in the nanocomposite layers doped with MFI nanoparticles. In
addition to the improved dynamic range, a lower level of
shrinkage in the zeolite-doped photopolymer was observed
(Figure 17). This is of vital importance for the zeolite-doped
photopolymer layers in applications such as holographic data
storage. The level of shrinkage in the photosensitive nanocom-
posites is studied by recording holographic transmission
slanted gratings in samples containing different concentrations
of zeolite nanoparticles. The shift in the position of the angular
selectivity curve due to shrinkage in nondoped and doped
photopolymers is shown in Figure 17. A substantial decrease
from 1.07% to 0.1% after adding the zeolite nanoparticles to
the photopolymer is demonstrated. Moreover, dependence of
the level of shrinkage on the film thickness of the layers was
observed, that is, the percentage shrinkage increases as the
sample thickness decreases. In all cases, the shrinkage of the
layers containing zeolite nanoparticles was significantly lower
than in the pure photopolymer samples.
Many organic and inorganic polymer materials have been
considered potential candidates for low-k dielectrics. The or-
ganic polymers (e.g., highly fluorinated alkane derivatives), for
example, could have k below 2.2, but they suffer from low
thermal stability and thermal conductivity.143 For porous
inorganic-based materials (e.g., solgel silica), they offer a tun-
able k-value, but its low mechanical strength, wide pore-size
distribution, and hydrophilicity have been cited as concerns.143
The pure siliceous zeolite films, however, possess many distin-
guished features, such as high mechanical strength, good heat
conductivity, and small uniform pore size (<3 nm), which
make them a very promising candidate as low-k dielectric for
future-generation microprocessors. Thin films with a k-value in
the range of 1.82.1 are prepared by spin coating or secondary
Laser off
Doped
1.0 Nondoped
0.8
Signal (a.u.)
0.1% 1.07 %
0.4
1.2 wt. % 60 nm 0.2
0 wt. %
100 150 1.0 0.5 0.0 0.5 1.0
(b) Angle (degrees)

growth of pure-silica MFI and MEL nanocrystals.144,145 Porogen
(such as g-cyclodextrin) is incorporated in the amorphous silica
phase between nanocrystals to reduce k-value to the ultra-low-k
range (<2.0).146 Additionally, the silica zeolite films can be
silylated to increase the hydrophobicity.
The preparation of antireflection coating on solar glass
support using zeolite nanocrystals has been reported.147,148
The antireflection coatings are very thin and the existence of
4070-nm nanoparticles in the coating is not a serious prob-
lem as long as it does not affect the transparency. A very strong
film for solar glasses is prepared using crystalline microporous
aluminosilicate nanoparticles.
Attempts to build an artificial antenna by enclosing dyes
inside a microporous material have been demonstrated.149151
Nanosized LTL crystals are used as a host for different lumines-
cent dyes organized in the one-dimensional channels of
the zeolite.
The reduction in sensitivity of high energetic materials has
been an essential subject of interest since the discovery of
energetic materials and it continues to be important nowadays.
Desensitizing these compounds is of great importance not only
to enhance their safe use, but also to make them less dangerous
in fields such as production of standards for analytical
purposes and the detection of explosive device. In this regard,
MFI-type nanosized zeolite with three-dimensional channel
network is used as a host for stabilizing a high-density energetic
material (Fox-7).152,153 The immobilized energetic material in
the zeolite nanocrystals showed high stability (100  C above
the explosive temperature), which opens up new possibilities
for preparation of safe standards and safe manipulations.
Novel nanostructured functional materials for applications
in the area of ultraviolet (UV) filtering, sensing, and molecular
switching by encapsulation of photophysical and photochem-
ical active guests in the zeolite nanocrystals have been pre-
pared.154156 The nanosized zeolites stabilized in suspensions
and thin films are promising for the development of host/guest
ultrafast switching systems since on a picosecond (1012) time-
scale a functional guest molecule may behave in a zeolite host
like in free solution.
In addition to the incorporation of organic guest molecules,
the stabilization of metal (Me Cu, Pd, Pt, etc.) and semicon-
ducting (SC CdS, ZnS, PbS, etc.) clusters in nanosized zeo-
lites has attracted considerable attention.157,158 On the one
hand, the size and shape of the Me and SC clusters are deter-
mined by the dimensions of the pores of the zeolites, and, on
the other hand, once they are hosted they do not agglomerate
and keep their unique properties. The metal and semiconduc-
tor clusters stabilized in zeolites are important for the develop-
ment of new materials with remarkable catalytic, optical, or
electronic properties.159 In addition, the Me clusters stabilized
in zeolites are proven to be good catalysts for disproportion-
ation of ethylbenzene,160 photochemical/thermal cleavage of
water to H2 and O2,161 photo-oxygen production from
water,162 photodimerization of alkanes,163 and selective reduc-
tion of NO by ethylene164 and ethanol.165
5.10.3.3 Biological and Medical Applications
During the last few decades, zeolite nanocrystals have been
considered for medical use due to their properties and stability
Zeolite Nanoparticles 299
in biological environment.166 The vast diversity of the zeolite
nanoparticles makes them very interesting as selective sorbent,
antioxidant, UV, and antimicrobial and antiviral protection.
Based on their unique properties, they are applied in processes
including (1) absorption of unpleasant odors, (2) skin cleaning,
(3) protection from harmful emission and microorganisms, (4)
protection from rheumatic problems, (5) favoring of healing
process in skin wounds, and (6) regulating of the pH of skin.
Zeolite nanocrystals have been investigated as drug carriers,
positive magnetic resonance imaging (MRI) agents, enzyme-
immobilizing carriers, etc. Several diagnostic pharmaceuticals
for MRI based on zeolite nanocrystals have been investigated.
Zeolite nanoparticles modified with gadolinium(III), Gd3,
are used as effective positive contrast agents.167 The zeolite
nanocrystals function by shortening the proton relaxation
times of water and tissue, which results in image brightening.
The immobilization of biomolecules onto the surface of
zeolite nanocrystals has been the focus of intense activity
in biotechnology and biomedicine. Especially, amine groups
in the proteins can be bound onto the surface of zeolite
particles and can be stabilized electrostatically. In addition,
zeolite nanocrystals show great promise as substrates for
immobilizing proteins due to their small size, sufficient func-
tional groups for further grafting or attachment, special surface
hydrophilichydrophobic microregion distribution, stable col-
loidal properties, and high level of adsorption capacity.
Nanozeolites are used as good enzyme-immobilizing car-
riers for biosensing and as substances in the enrichment and
identification of low-abundance peptides/proteins.168,169 For
example, LTL nanocrystals have shown to be a promising
trypsin-immobilizing carriers and have been patterned in poly
(methyl methacrylate) (PMMA) microfluidic channels for fabri-
cating an enzyme microreactor.170 The resulting trypsin micro-
reactor not only achieves the efficient digestion of proteins at a
low concentration in a very short reaction time but also shows
good stability and universality for protein identification. There-
fore, the LTL has been shown to have a great potential in
automated high-throughput analysis using a parallel-channel
microchip platform for proteomic analysis integrated with the
developed separation and identification procedures.
Nanosized zeolites are also used in the purification of pro-
tein on the basis of electrostatic or hydrophobic interaction
between the sought-after biomolecules and zeolites. A success-
ful separation of the selenoprotein-P (Se-P) in mouse plasma
through their chelation with Co2-exchanged nanozeolite/
diatomite composites has been reported.171 It was found
that, after immobilizing transition-metal ions in zeolite crys-
tals, the transition-metal-ion immobilized zeolites would se-
lectively adsorb the histidine-rich domains in the targeted
protein molecules.
Nanocrystalline zeolites have been modified and used as
unique magnetic zeolites. The zeolite nanocrystals combined
with superparamagnetic magnetite (Fe3O4) nanoparticles
demonstrated an excellent adsorption separation of enzymes
and a good biocatalytic performance.172 Zeolite nanocompo-
sites consisting of magnetite and FAU zeolite are used to store
and release substantial amounts of doxorubicin, an anticancer
antibiotic belonging to the tetracycline group. In this form of
drug delivery, an external or internal magnetic field can be used
to direct drug delivery particles to the proximity of the tumor
300 Zeolite Nanoparticles

Separation and catalysis

Storage materials: heat,
hydrogen, methane
Membranes: gas and liquids
Sensors: acoustic,
electrochemical, optical
Films: dielectric layers,
anticorrosion and antireflective
coatings
Hosts for clusters and organics:
photoconductors, semiconductors,
supercapacitors
Biology and medicine: selective
sorbent, antioxidant, antimicrobial
and antiviral protection, drug carrier

Figure 18 Advanced applications of zeolite nanocrystals.

cells, thus enabling significant reduction of the necessary dose
of medication and minimizing the side effects.173
Moreover, recent papers showed that nanozeolite particles
possessed marvelous adsorption and immobilization cap-
ability for biomolecules due to their unique surface prop-
erty,169,170 which make them promising for delivery and
stabilization of protein medicine. Nanosized FAU zeolite had
also been applied to follow endosomal acidification and pro-
teolysis for the investigation of the endocytosing mechanisms
of human peripheral dendritic cells owing to their high ad-
sorption capacity for various biomolecules.174 It was reported
that low concentration of zeolite particles (13 mg ml1)
showed nontoxicity in exposure to the cells for 3 days. In
summary, the toxicities of various zeolite nanoparticles were
studied and the results indicate that toxicities of nanozeolites
are relative to their size, composition, and shape. The spherical
pure-silica nanozeolites display a nontoxic effect, but the
aluminum-containing zeolite nanoparticles show a dose-
dependent toxicity. More importantly, the toxic nanozeolites
seem to induce cell necrosis rather than cell apoptosis by
demolishing the cell membranes.175,176 Thus, the spherical
pure-silica zeolites nanoparticles are potentially useful for
medical purposes.
In conclusion, the design and synthesis of nanosized zeo-
lites with diverse compositions, structures, and their further
processing to macroscopic constructs are important for enlarg-
ing their technological perspectives and also they are interest-
ing from a fundamental point of view. The reduction of particle
size from the micrometer to the nanometer scale leads to
substantial changes in the properties of the materials, which
have an impact on the performance of the zeolite nanoparticles
even in the traditional applications. The possibility to obtain
stable suspensions of zeolite nanoparticles able to be processed
on different surfaces by rapid techniques is also of great im-
portance for the advanced application of colloidal molecular
sieves.
Some of the important advanced applications of zeolite
nanoparticles are presented in Figure 18.
5.10.4 Conclusion
This chapter gives a general view on the different approaches
for preparation of zeolite nanoparticles. The emphasis is on the
various approaches for preparation of zeolite nanocrystals
including templated synthesis, template-free, ionothermal,
seed-induced, and confined space synthesis. Other methods
including microemulsion, direct conversion, and multistep
crystallization of zeolite nanoparticles are presented. Depend-
ing on the conditions of synthesis, discrete nanoparticles can
be obtained with variable morphologies and particle sizes.
Most of the zeolite nanocrystals for basic research and used
in advanced applications are prepared using the templated
synthesis approach. This approach results in zeolite nanocrys-
tals with monomodal particle-size distribution and high sta-
bility in suspensions. However, low crystalline yield and high
consumption of organic template are the main concerns.
Therefore, the zeolite nanocrystals are considered in areas
where very small amounts are required including preparation
of coatings for sensor application, optics, drug delivery sys-
tems, solar cells, medicine, etc. Additionally, the organic-free
systems provide nanocrystals with high yield; however, the
very fine colloidal particles are not stable and tend to agglom-
erate with time. Additional modification of their surfaces is
required to prevent further agglomeration and enable further
applications.
The studies on the toxicity of zeolite nanocrystals with
different sizes, compositions, and shapes are expected to fur-
ther direct their applications in biomedicine and pharmaceu-
tical industry. For related chapters in this Comprehensive, we
refer to Chapters 4.05, 5.05, 5.06, 7.04, and 7.10.
Acknowledgment
The financial support from MEET INTEREG EC project is
acknowledged.

References
1. Binns, C. Introduction to Nanoscience and Nanotechnology. John Wiley & Sons:
New York, 2010.
2. Ramsden, J. Applied Nanotechnology: The Conversion of Research Results to
Products. Elsevier Inc.: New York, 2009.
3. Mathur, S.; Singh, M. Nanostructural Materials and Nanotechnology. John Wiley
& Sons: New York, 2007.
4. Barrer, D. W. Zeolite Molecular Sieves. Wiley-Interscience: New York, 1974.
5. Tosheva, L.; Valtchev, V. Chem. Mater. 2005, 17, 2494.
6. Larsen, S. C. J. Phys. Chem. C 2007, 111, 18464.
7. Lew, C. M.; Cai, R.; Yan, Y. Acc. Chem. Res. 2010, 43, 210.
8. Kumar, S.; Wang, Z.; Penn, R. L.; Tsapatsis, M. J. Am. Chem. Soc. 2008, 130,
17284.
9. Yoo, W. C.; Kumar, S.; Penn, R. L.; Tsapatsis, M.; Stein, A. J. Am. Chem. Soc.
2009, 131, 12377.
10. Lee, P.-S.; Zhang, X.; Stoeger, J. A.; Malek, A.; Fan, W.; Kumar, S.; Yoo, W. C.;
Hashimi, S. A.; Penn, R. L.; Stein, A.; Tsapatsis, M. J. Am. Chem. Soc. 2011, 133,
493.
11. Hould, N. D.; Kumar, S.; Tsapatsis, M.; Nikolakis, V.; Lobo, R. F. Langmuir 2010,
26, 1260.
12. Huang, L.; Wang, Z.; Sun, J.; Miao, L.; Li, Q.; Yan, Y.; Zhao, D. J. Am. Chem. Soc.
2000, 122, 3530.
13. Wang, H.; Wang, Z.; Huang, L.; Mitra, A.; Yan, Y. Langmuir 2001, 17, 2572.
14. Lew, C. M.; Liu, Y.; Day, B.; Kloster, G. M.; Tiznado, H.; Sun, M.; Zaera, F.;
Wang, J.; Yan, Y. Langmuir 2009, 25, 5039.
15. Li, Q.; Creaser, D.; Sterte, J. Stud. Surf. Sci. Catal. 2001, 135, 140.
16. Wang, Z.; Hedlund, J.; Sterte, J. Micropor. Mesopor. Mater. 2002, 52, 191.
17. Kobler, J.; Abrevaya, H.; Mintova, S.; Bein, T. J. Phys. Chem. C 2008, 112,
14274.
18. Moller, K.; Yilmaz, B.; Jacubinas, R. M.; Muller, U.; Bein, T. J. Am. Chem. Soc.
2011, 133, 5284.
19. Biemmi, E.; Bein, T. Langmuir 2008, 24, 11196.
20. Valtchev, V. P.; Tosheva, L.; Bozhilov, K. N. Langmuir 2005, 21, 10724.
21. Li, Q.; Mihailova, B.; Creaser, D.; Sterte, J. Micropor. Mesopor. Mater. 2001, 43,
51.
22. Mintova, S.; Valtchev, V.; Micropor, V. Micropor. Mesopor. Mater. 2002, 52,
171.
23. Holzl, M.; Mintova, S.; Bein, T. Stud. Surf. Sci. Catal. 2005, 158, 11.
24. Mintova, S.; Fieres, B.; Bein, T. Stud. Surf. Sci. Catal. 2002, 142, 223.
25. Mintova, S.; Valtchev, V. Microporous Molecular Sieves as Colloids. In Zeolites:
From Model Materials to Industrial Catalysts; Cejka, J., Perez-Pariente, J.,
Roth, W. J., Eds.; Research Signpost: India, 2007; p 63.
26. Mintova, S. Collect. Czech. Chem. Commun. 2003, 68, 2032.
27. Bein, T.; Mintova, S. Advanced Applications of Zeolites. In Zeolites and Ordered
Mesoporous Materials, Progress and Prospects; Cejka, J., van Bekkum, H., Eds.;
Elsevier: Amsterdam, 2005; p 263.
28. Larlus, O.; Mintova, S.; Wilson, S. T.; Willis, R. R.; Abrevaya, H.; Bein, T.
Micropor. Mesopor. Mater. 2011, 142, 17.
29. Mintova, S.; Olson, N.; Valtchev, V.; Bein, T. Science 1999, 283, 958.
30. Mintova, S.; Olson, N.; Bein, T. Angew. Chem. Int. Ed. 1999, 38, 3201.
31. Ng, E.-P.; Chateigner, D.; Bein, T.; Valtchev, V.; Mintova, S. Science 2012, 335,
70.
32. Knagge, K.; Johnson, M.; Grassian, V. H.; Larsen, S. C. Langmuir 2006, 22,
11077.
33. Tokay, B.; Karvan, O.; Senatalar, A. E. Micropor. Mesopor. Mater. 2010, 131, 230.
34. van Heyden, H.; Mintova, S.; Bein, T. J. Mater. Chem. 2006, 16, 514.
35. Wilson, S. T.; Lok, B. M.; Messina, C. A.; Cannan, T. R.; Flanigen, E. M. J. Am.
Chem. Soc. 1982, 104, 1146.
36. Sharma, P.; Rajaram, P.; Tomar, R. J. Colloid. Interf. Sci. 2008, 325, 547.
37. Chu, C. T. W.; Kuehl, G. H.; Lago, R. M.; Chang, D. D. J. Catal. 1985, 93, 451.
38. de Moor, P. P. E. A.; Beelen, T. P. M.; van Santen, R. A. J. Phys. Chem. B 1999,
103, 1639.
39. Persson, A. E.; Shoeman, B. J.; Sterte, J.; Ottersted, J. E. Zeolites 1994, 14, 557.
40. Schoeman, B. J. Micropor. Mesopor. Mater. 1998, 22, 9.
41. Burkett, S. L.; Davis, M. E. Chem. Mater. 1995, 7, 920.
42. Dokter, W. H.; Garderen, H. F.; Beelen, T. P. M.; Santen, R. A.; Bras, W. Angew.
Chem. Int. Ed. 1995, 34, 73.
43. Kecht, J.; Mintova, S.; Bein, T. Chem. Mater. 2007, 19, 1203.
44. Kecht, J.; Mintova, S.; Bein, T. Langmuir 2008, 24, 4310.
45. Zhan, B. Z.; White, M. A.; Lumsden, M.; Neuhaus, J. M.; Robertson, K. N.;
Cameron, T. S.; Gharghouri, M. Chem. Mater. 2002, 14, 3636.
Zeolite Nanoparticles 301
46. Valtchev, V. P.; Bozhilov, K. N. J. Phys. Chem. B 2004, 108, 15587.
47. Fan, W.; Morozumi, K.; Kimura, R.; Toshiyuki, T.; Yokoi, T.; Okubo, T. Langmuir
2008, 24, 6952.
48. Ren, N.; Bronic, J.; Subotic, B.; Lv, X.-C.; Yang, Z.-J.; Tang, Y. Micropor.
Mesopor. Mater. 2011, 139, 197.
49. Larlus, O.; Tosheva, L.; Holzl, M.; Mintova, S.; Metzger, T.; Valtchev, V. P. Stud.
Surf. Sci. Catal. 2005, 158, 367.
50. Meng, X.; Zhang, Y.; Meng, M.; Pang, W. In Proceedings of the 9th International
Zeolite Conference, Montreal 1992; von Ballmoos, R., et al. Eds.; Butterworth-
Heinemann, London, 1993; p 297.
51. Yao, J.; Wang, H.; Ratinac, K. R.; Ringer, S. P. Chem. Mater. 2006, 18, 1394.
52. Jawor, A.; Jeong, B. H.; Hoek, E. M. V. J. Nanopart. Res. 2009, 11, 1795.
53. Brar, T.; France, P.; Smirniotis, P. G. Ind. Eng. Chem. Res. 2001, 40, 1133.
54. Vaughan, D. E. W. U.S. Patent 4,534,947 A 19850813, 1985
55. Majano, G.; Darwiche, A.; Mintova, S.; Valtchev, V. Ind. & Engineer. Chem. Res.
2009, 48, 7084.
56. Majano, G.; Delmotte, L.; Valtchev, V.; Mintova, S. Chem. Mater. 2009, 21, 4184.
57. Hincapie, B. O.; Garces, L. J.; Zhang, Q.; Sacco, A.; Suib, S. L. Micropor.
Mesopor. Mater. 2004, 67, 19.
58. Serrano, D. P.; Aguado, J.; Escola, J. M.; Rodriguez, J. M.; Peral, A. Chem. Mater.
2006, 18, 2462.
59. Serrano, D. P.; Aguado, J.; Rodrguez, J. M.; Peral, A. Stud. Surf. Sci. Catal.
2007, 170, 282.
60. Hu, Y.; Zhang, Y.; Tang, Y. Chem. Commun. 2010, 46, 3875.
61. Iyoki, K.; Kamimura, Y.; Itabashi, K.; Shimojima, A.; Okubo, T. Chem. Lett. 2010,
39, 730.
62. Hincapie, B. O.; Garces, L. J.; Zhang, Q.; Sacco, A.; Suib, S. L. Micropor.
Mesopor. Mater. 2004, 67, 19.
63. Serrano, D. P.; Aguado, J.; Escola, J. M.; Rodriguez, J. M.; Peral, A. Chem. Mater.
2006, 18, 2462.
64. Serrano, D. P.; Aguado, J.; Rodrguez, J. M.; Peral, A. Stud. Surf. Sci. Catal.
2007, 170, 282.
65. Hu, Y.; Zhang, Y.; Tang, Y. Chem. Commun. 2010, 46, 3875.
66. Song, W.; Grassian, V. H.; Larsen, S. C. Chem. Commun. 2005, 295.
67. Larlus, O.; Mintova, S.; Bein, T. Micropor. Mesopor. Mater. 2006, 96, 405.
68. Ng, E.-P.; Delmotte, L.; Mintova, S. Green Chem. 2008, 10, 1043.
69. Welton, T. Chem. Rev. 1999, 99, 2071.
70. Cooper, E. R.; Andrews, C. D.; Wheatley, P. S.; Webb, P. B.; Wormald, P.;
Morris, R. E. Nature 2004, 430, 1012.
71. Parnham, E. R.; Morris, R. E. Acc. Chem. Res. 2007, 40, 1005.
72. Harrison, W. T. A. Inorg. Chem. Commun. 2007, 10, 833.
73. Cho, S. D.; Park, H. J. Colloid Interface Sci. 2011, 357, 46.
74. Lian, J.; Ma, J.; Duan, X.; Kim, T.; Li, H.; Zheng, W. Chem. Commun. 2010,
2650.
75. Yin, S.; Luo, Z.; Xia, J.; Li, H. J. Phys. Chem. Solid 2010, 71, 1785.
76. Liu, X.; Zhao, J.; Sun, Y.; Song, K.; Yu, Y.; Du, C.; Kong, X.; Zhang, H. Chem.
Commun. 2009, 6628.
77. Hayakawa, Y.; Nonoguchi, Y.; Wu, H.-P.; Diau, E. W.-G.; Nakashima, T.; Kawai, T.
J. Mater. Chem. 2011, 21, 8849.
78. Zhang, X.; Chen, H.; Zheng, M.; Liu, J.; Cao, J. Chem. Lett. 2007, 36, 1498.
79. Ng, E.-P.; Sekhon, S. S.; Mintova, S. Chem. Commun. 2009, 1661.
80. Ng, E.-P.; Itani, L.; Sekhon, S. S.; Mintova, S. Chem. Eur. J. 2010, 16, 12890.
81. Han, L.; Wang, Y.; Li, C. AIChE J. 2008, 54, 280.
82. Slangen, P. M.; Jansen, J. C.; van Bekkum, H. Micropor. Mater. 1997, 9, 259.
83. Tompsett, G. A.; Conner, W. C.; Yngvesson, K. S. Chemphyschem 2006, 7, 296.
84. Mintova, S.; Mo, S.; Bein, T. Chem. Mater. 1998, 10, 4030.
85. Lin, S.; Li, J.; Sharma, R. P.; Yu, J.; Xu, R. Top. Catal. 2010, 53, 1304.
86. Hu, Y.; Liu, C.; Zhang, Y.; Ren, N.; Tang, Y. Micropor. Mesopor. Mater. 2009,
119, 306.
87. Jahnisch, K.; Hessel, V.; Lowe, H.; Baerns, M. Angew. Chem. Int. Ed. 2004, 43, 406.
88. Pan, Y.; Yao, J.; Zhang, L.; Xu, N. Ind. Eng. Chem. Res. 2009, 48, 8471.
89. Ju, J.; Zeng, C.; Zhang, L.; Xu, N. Chem. Eng. J. 2006, 116, 115.
90. Pan, Y.; Ju, M.; Yao, J.; Zhang, L.; Xu, N. Chem. Commun. 2009, 7233.
91. Pan, Y.-C.; Yao, J.-F.; Zhang, L.-X.; Ju, J.-X.; Wang, H.-T.; Xu, N.-P. Chem. Eng.
Technol. 2009, 32, 732.
92. Madsen, C.; Jacobsen, C. J. H. Chem. Commun. 1999, 673.
93. Persson, A. E.; Shoeman, B. J.; Sterte, J.; Ottersted, J. E. Zeolites 1994, 14, 557.
94. Schmidt, I.; Madsen, C.; Jacobsen, C. J. H. Inorg. Chem. 2000, 39, 2279.
95. Jacobsen, C. J. H.; Madsen, C.; Janssens, T. V. W.; Jakobsen, H. J.; Skibsted, J.
Micropor. Mesopor. Mater. 2000, 39, 393.
96. Huu, C. P.; Wine, G.; Tessonnier, J. P.; Ledoux, M. J.; Rigolet, S.; Marichal, C.
Carbon 2004, 42, 1941.
302 Zeolite Nanoparticles
97. Tang, K.; Wang, Y. G.; Song, L. J.; Duan, L. H.; Zhang, X. T.; Sun, Z. L. Mater.
Lett. 2006, 60, 2158.
98. Wang, B.; Ma, H. Z.; Shi, Q. Z. Chin. Chem. Lett. 2002, 13, 385.
99. Wang, H. T.; Holmberg, B. A.; Yan, Y. S. J. Am. Chem. Soc. 2003,
125, 9928.
100. Li, D.; Ratinac, K. R.; Ringer, S. P.; Wang, H. Micropor. Mesopor. Mater. 2008,
116, 416.
101. Yang, X.; Feng, Y.; Tian, G.; Du, Y.; Ge, X.; Di, Y.; Zhang, Y.; Sun, B.; Xiao, F.-S.
Angew. Chem. Int. Ed. 2005, 44, 2563.
102. Carr, C. S.; Shantz, D. F. Micropor. Mesopor. Mater. 2005, 85, 284.
103. Axnanda, S.; Shantz, D. F. Micropor. Mesopor. Mater. 2005, 84, 236.
104. Zhang, J.; Yan, W.; Ding, H.; Liu, Y.; Tang, K.; Yu, J.; Xu, R. Stud. Surf. Sci. Catal.
2007, 170, 475.
105. Lin, J.-C.; Yates, M. Z. Langmuir 2005, 21, 2117.
106. Carr, C. S.; Shantz, D. F. Chem. Mater. 2005, 17, 6192.
107. Naskar, M. K.; Kundu, D.; Chatterjee, M. Mater. Lett. 2011, 65, 436.
108. Zhang, Y.; Jin, C. J. Solid State Chem. 2011, 184, 1.
109. Zhang, J.; Yan, W.; Ding, H.; Liu, Y.; Tang, K.; Yu, J.; Xu, R. Stud. Surf. Sci. Catal.
2007, 170, 282.
110. Reyes, P. Y.; Espinosa, J. A.; Trevino, M. E.; Saade, H.; Lopez, R. G. J. Nanomater.
2010, 1.
111. Faghihian, H.; Godazandeha, N. J. Porous Mater. 2007, 170, 282.
112. Kamali, M.; Vaezifar, S.; Kolahduzan, H.; Malekpour, A.; Abdi, M. R. Powder
Technol. 2009, 189, 52.
113. Honda, K.; Yashiki, A.; Itakura, M.; Ide, Y.; Sadakane, M.; Sano, T. Micropor.
Mesopor. Mater. 2011, 142, 161.
114. Kong, C.; Tsuru, T. Chem. Eng. Process 2010, 49, 809.
115. Wakihara, T.; Ichikawa, R.; Tatami, J.; Endo, A.; Yoshida, K.; Sasaki, Y.;
Komeya, K.; Meguro, T. Cryst. Growth Des. 2011, 11, 955.
116. Kharitonov, A. S.; Fenelonov, V. B.; Voskresenskaya, T. P.; Rudina, N. A.;
Molchanov, V. V.; Plyasova, L. M.; Panov, G. I. Zeolites 1995, 15, 253.
117. Mafune, F.; Kohno, J. Y.; Takeda, Y. J. Phys. Chem. B 2001, 106, 7575.
118. Mafune, F.; Kohno, J. Y.; Takeda, Y. J. Phys. Chem. B 2002, 106, 8555.
119. Kawasaki, M.; Masuda, K. J. Phys. Chem. B 2005, 109, 9379.
120. Nichols, W. T.; Kodaira, T.; Sasaki, Y.; Shimizu, Y.; Sasaki, T.; Koshizaki, N. J.
Phys. Chem. B 2006, 110, 83.
121. Kobler, J.; Abrevaya, H.; Mintova, S.; Bein, T. J. Phys. Chem. C 2008, 112, 14274.
122. Maurer, T.; Kraushaar-Czarnetzki, B. Helv. Chim. Acta 2001, 84, 2550.
123. Bein, T. Chem. Mater. 1996, 8, 1636.
124. Lai, Z.; Bonila, G.; Diaz, I.; Nery, J. G.; Sujaoti, K.; Amat, M. A.; Kokkoli, E.;
Terasaki, O.; Thompson, R. W.; Tsapatsis, M.; Vlachos, D. G. Science 2003, 300,
456.
125. Yoo, W. C.; Kumar, S.; Penn, R. L.; Tsapatsis, M.; Stein, A. J. Am. Chem. Soc.
2009, 131, 12377.
126. Chen, Y.; Zhu, G.; Peng, Y.; Yao, X.; Qiu, S. Micropor. Mesopor. Mater. 2009,
124, 8.
127. Ghoroghchian, F.; Aghabozorg, H.; Farhadi, F.; Kazemian, H. Chem. Eng.
Technol. 2010, 33, 2066.
128. Li, D.; Zhu, H. Y.; Ratinac, K. R.; Ringer, S. P.; Wang, H. Micropor. Mesopor.
Mater. 2009, 126, 14.
129. Holmberg, B. A.; Wang, X.; Yan, Y. J. Membrane. Sci. 2008, 320, 86.
130. Choi, J.; Jeong, H.-K.; Snyder, M. A.; Stoeger, J. A.; Masel, R. I.; Tsapatsis, M.
Science 2009, 325, 590.
131. Kuzniatsova, T. A.; Mottern, M. L.; Chiu, W. V.; Kim, Y.; Dutta, P. K.; Verweij, H.
Adv. Funct. Mater. 2008, 18, 952.
132. Tavolaro, A.; Drioli, E. Adv. Mater. 1999, 11, 975.
133. Caro, J.; Noack, M.; Kolsch, P.; Schafer, R. Micropor. Mesopor. Mater. 2000, 38, 3.
134. Chiang, A. S. T.; Chao, K. J. Phys. Chem. Solids 2001, 62, 1899.
135. Noack, M.; Kolsch, P.; Schafer, R.; Toussaint, P.; Caro, J. Chem. Eng. Technol.
2002, 25, 221.
136. Coronas, J.; Santamaria, J. Top. Catal. 2004, 29, 29.
137. Nair, S.; Tsapatsis, M. In Handbook of Zeolite Science and Technology;
Auerbach, S. M., Carrado, K. A., Dutta, P. K., Eds.; Marcel Dekker: New York,
2003; p 869.
138. Urbitzondo, M. A.; Pina, M. P.; Santamaria, J. Gas sensing with silicon-based
nanoporous solids. In Ordered Porous Materials; Valtchev, V., Mintova, S.,
Tsapatsis, M., Eds.; Elsevier: Amsterdam, 2009; p 387.
139. Mintova, S.; Bein, T. Micropor. Mesopor. Mater. 2001, 50, 159.
140. Leite, E.; Naydenova, I.; Mintova, S.; Leclercq, L.; Toal, V. Appl. Optics. 2010, 49,
3652.
141. Leite, E.; Naydenova, I.; Pandey, N.; Babeva, T.; Majano, G.; Mintova, S.; Toal, V.
J. Opt. Pure Appl. Opt. 2009, 11, 1.
142. Naydenova, I.; Sherif, H.; Mintova, S.; Martin, S.; Toal, V. Proc. SPIE 2006,
6252, 1.
143. Miller, R. D. Science 1999, 286, 421.
144. Wang, Z.; Wang, H.; Mitra, A.; Huang, L.; Yan, Y. Adv. Mater. 2001, 13, 746.
145. Wang, Z.; Wang, H.; Mitra, A.; Huang, L.; Yan, Y. Adv. Mater. 2001, 13, 746.
146. Li, S.; Li, Z.; Yan, Y. Adv. Mater. 2003, 15, 1528.
147. Chen, C.-H.; Li, S.-Y.; Chiang, A. S. T.; Wu, A. T.; Sun, Y. S. Solar Energ. Mater.
Solar Cell 2011, 95, 1694.
148. Chiang, A. S. T.; Wong, L. J.; Li, S. Y.; Cheng, S. L.; Lee, C. C.; Chen, K. L.;
Chen, S. M.; Lee, Y. J. Stud. Surf. Sci. Catal. 2007, 170, 1583.
149. Minkowski, C.; Pansu, R.; Takano, M.; Calzaferri, G. Adv. Funct. Mater. 2006,
16, 273.
150. Calzaferri, G.; Huber, S.; Maas, H.; Minkowski, C. Angew. Angew. Chem. Int. Ed.
Engl. 2003, 42, 3732.
151. Calzaferri, G.; Pauchard, M.; Maas, H.; Huber, S.; Khatyr, A.; Schaafsma, T. J.
Mater. Chem. 2002, 12, 1.
152. Majano, G.; Mintova, S.; Bein, T.; Klapotke, T. M. Adv. Mater. 2006, 18, 2440.
153. Majano, G.; Mintova, S.; Bein, T.; Klapotke, T. M. J. Phys. Chem. 2007, 111, 6694.
154. Mintova, S.; De Waele, V.; Schmidhammer, U.; Riedle, E.; Bein, T. Angew. Chem.
Int. Ed. 2003, 42, 1611.
155. Mintova, S.; De Waele, V.; Holzl, M.; Schmidhammer, U.; Mihailova, B.; Riedle, E.;
Bein, T. J. Phys. Chem. A 2004, 108, 10640.
156. Meinershagen, J. L.; Bein, T. Adv. Mater. 2001, 13, 208.
157. Kecht, J.; Tahri, Z.; De Waele, V.; Mostafavi, M.; Mintova, S.; Bein, T. Chem.
Mater. 2006, 18, 3373.
158. Yordanov, I.; Knoerr, R.; De Waele, V.; Bazin, P.; Thomas, S.; Rivallan, M.;
Lakiss, L.; Metzger, T. H.; Mintova, S. J. Phys. Chem. C 2010, 114, 20974.
159. Tahri, Z.; Luchez, F.; Yordanov, I.; Poizat, O.; Moissette, A.; Valtchev, V.;
Mintova, S.; Mostafavi, M.; De Waele, V. Res. Chem. Intermed. 2009, 35, 379.
160. Baba, T.; Ono, Y. Zeolites 1987, 7, 292.
161. Jacobs, P. A.; Uytterhoeven, J. B.; Beyer, H. K. J. Chem. Soc. Chem. Commun.
1977, 128.
162. Calzaferri, G.; Hugues, S.; Hugentobler, T.; Sulzberger, B. J. Photochem. 1984,
26, 108.
163. Ozin, G. A.; Hugues, F. J. Phys. Chem. 1982, 86, 5174.
164. Sato, S.; Yu-u, Y.; Yahiro, H.; Mizuno, N.; Iwamoto, M. Appl. Catal. 1991, 70, L1.
165. Aoyama, N.; Yoshida, K.; Abe, A.; Miyadera, T. Catal. Lett. 1997, 43, 249.
166. Danilczuk, M.; Dlugopolska, K.; Ruman, T.; Pogocki, D. Mini Rev. Med. Chem.
2008, 8, 1407.
167. Bresinska, I.; Balkus, K. J. J. Phys. Chem. 1994, 98, 12989.
168. Zhang, Y. H.; Wang, X. Y.; Shan, W.; Wu, B. Y.; Fan, H. Z.; Yu, X. J.; Tang, Y.;
Yang, P. Y. Angew. Chem. Int. 2005, 117, 621.
169. Yu, T.; Zhang, Y. H.; You, C. P.; Zhuang, J. H.; Wang, B.; Liu, B. H.; Kang, Y. J.;
Tang, Y. Chem. Eur. J. 2006, 12, 1137.
170. Zhang, Y.; Liu, Y.; Kong, J.; Yang, P.; Tang, Y.; Liu, B. Small 2006, 2, 1170.
171. Xu, F.; Wang, Y. J.; Wang, X. D.; Zhang, Y. H.; Tang, Y.; Yang, P. Y. Adv. Mater.
2003, 15, 1751.
172. Shan, W.; Yu, T.; Wang, B.; Hu, J.; Zhang, Y.; Wang, X.; Tang, Y. Chem. Mater.
2006, 18, 3169.
173. Arruebo, M.; Pacheco, R. F.; Irusta, S.; Arbiol, J.; Ibarra, M. R.; Santamaria, J.
Nanotechnol. 2006, 17, 4057.
174. Andersson, L. I. M.; Eriksson, H. Scand. J. Immunol. 2007, 66, 52.
175. Kihara, T.; Zhang, Y.; Hu, Y.; Mao, Q.; Tang, Y.; Miyake, J. J. Biosci. Bioeng.
2011, 111, 725.
176. Bhattacharyaa, K.; Nahaa, P. C.; Naydenova, I.; Mintova, S.; Byrne, H. J. Toxicol.
Lett. 2012, 215, 151.
 5.01 Diffraction and Spectroscopy of Porous Solids
W Schmidt, F Schuth, and C Weidenthaler, MPI fur Kohlenforschung, Mulheim, Germany
2013 Elsevier Ltd. All rights reserved.

5.01.1 Introduction 1
5.01.2 Materials Covered 2
5.01.2.1 Ordered Porous Materials 2
5.01.2.1.1 Zeolites and zeotypes 2
5.01.2.1.2 Metal organic frameworks 2
5.01.2.1.3 Ordered mesoporous materials 3
5.01.2.1.4 Photonic crystals 3
5.01.2.2 Disordered Porous Material 3
5.01.3 Scattering and Diffraction Methods 3
5.01.3.1 Crystal Structures 4
5.01.3.2 Principles of XRD 4
5.01.3.3 Single-Crystal Diffraction 7
5.01.3.3.1 Structure solution from single-crystal data 7
5.01.3.3.2 The Patterson method 7
5.01.3.4 Powder Diffraction 7
5.01.3.4.1 Structure solution from powder diffraction data 8
5.01.3.4.2 In situ diffraction 8
5.01.3.5 Electron Diffraction 9
5.01.3.6 Small-Angle Scattering 11
5.01.3.6.1 Small-angle x-ray scattering 12
5.01.3.6.2 Approximation methods 15
5.01.3.6.3 Scattering on photonic crystals 16
5.01.4 Spectroscopies 17
5.01.4.1 IR Spectroscopy 18
5.01.4.1.1 Analysis of surface properties 19
5.01.4.1.2 Analysis of adsorbate/pore orientation 19
5.01.4.1.3 Analysis of diffusion processes 20
5.01.4.2 NMR Spectroscopy 20
5.01.4.2.1 Analysis of motion in porous materials on the molecular scale 21
5.01.4.2.2 Pulsed field gradient NMR spectroscopy 21
129
5.01.4.2.3 Xe NMR spectroscopy 21
5.01.4.3 Interference Microscopy 22
5.01.5 Conclusion 22
References 23

5.01.1 Introduction nonporous counterparts with respect to analysis using probe

Basically any type of porous solid can and routinely is
characterized by diffraction techniques or different types of
spectroscopy. While in many respects the analysis of porous
solids does not differ from that of nonporous materials, there
are some features that deserve special consideration.
The crucial feature of porous solids is the void space which
is an integral part of the solid itself, in contrast to the inter-
particle voids formed by packing of particles. Thus, specific
methods for the characterization of porous solids probe this
internal void space and its relation to the solid material. This
means that the spectroscopies will typically use probe mole-
cules exploring this void space or the surface exposed in the
pores. The big advantage of porous solids compared to their
molecules is the typically rather high surface area. This means
that the signal-to-noise (S/N) ratio is mostly much better than
for nonporous materials, as much more probe molecules inter-
acting with the surface contribute to the signal.
Diffraction techniques applied to (crystalline) porous
solids will often be faced with the problem of large unit cells,
since in order to accommodate sufficiently big pores in
a structure, the unit cell needs to exceed a certain size. The
diffraction fingerprint of a porous material thus often contains
reflections at very low angles. Such specific features will be in
the focus of this chapter. For the more generic methods for
the characterization of solids, the readers are referred to the
many textbooks available on the diffraction or the spectros-
copy of solids.

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00501-5 1

2 Diffraction and Spectroscopy of Porous Solids

5.01.2 Materials Covered
Porous materials can be classified in materials with ordered
and disordered pore systems, and within each of these classes,
one can discriminate according to pore size. On the final level,
different chemical compositions provide the basis for a more
fine-grained classification (Figure 1).1
While there is a multitude of different porous solids, this
chapter is restricted to such solids that are technologically
important and/or of high current scientific interest on the
one hand, and for which there are special considerations with
respect to characterization by diffraction or spectroscopy on
the other hand.
5.01.2.1 Ordered Porous Materials
Ordered porous materials have one-, two-, or three-dimensionally
ordered pore systems. Physically speaking, they can be considered
as having a basis and a lattice, which together provide for the
macroscopic order. Diffraction with radiation of an appropriate
wavelength is the key technique for the analysis of such ordered
materials. In the ideal case, diffraction will provide information
on the position of each atom in the unit cell, if the order reaches
down to the atomic scale, but there are quite a number of ordered
porous materials, in which the pores are periodic in space, but
which are disordered on the atomic scale. Also for such ordered
materials, diffraction can provide structural information, but only
on the typical length scale of the periodic repeat unit, that is, the
pore system.
5.01.2.1.1 Zeolites and zeotypes
Zeolites are technologically very important materials that find
applications in catalysis, sorption, and ion exchange, to name
the most important ones. Structurally, they are characterized
by corner-sharing, fully tetrahedral frameworks in which the
linking of the tetrahedra occurs in such a way that pores are
formed. In order for the pore systems to be accessible for guest
species, the openings in the frameworks should be formed by at
least eight tetrahedral units. The classical zeolites are alumosili-
cates with a minimum Si:Al ratio of 1. As silicon and aluminum
have almost the same electron count, they are indistinguishable
by conventional x-ray diffraction (XRD), while neutron diffrac-
tion can in principle provide information on aluminum siting
in zeolites. Nowadays, many compositions beyond silicon and
aluminum as framework constituents are known; the biggest
class, beyond the classical zeolites, are probably aluminum
phosphates. A survey of known structure types is given on the
web page of the International Zeolite Association.2
5.01.2.1.2 Metal organic frameworks
Metal organic frameworks (MOFs), sometimes also called
porous coordination polymers,3,4 are a class of materials that
are built up of both inorganic and organic moieties. They can be
considered as being formed in a modular fashion from vertices
and linkers. The vertices typically consist of oxo-clusters of main
group or transition metals; the linkers are polyfunctional
organic molecules that connect the vertices. One of the most
well-known MOFs is probably the MOF-5.5 It is prototypical of
many of these materials and consists of tetranuclear Zn4O clus-
ters which are linked to a very open, cubic structure by capping
1,4-benzenedicarboxylic acid linkers.
The analysis of MOFs by diffraction often presents substan-
tial challenges, as the unit cells are typically very large, contain-
ing hundreds or thousands of atoms. Moreover, the pore
system is often filled with disordered guest species. In addition,
these frameworks are often highly flexible, which can lead to
structural changes upon guest adsorption.

Porous solids

Structural porosity Textural porosity

Microporous Mesoporous Macroporous

Ordered Disordered Ordered Disordered Ordered Disordered

Zeolites Active Anodic aluminas Oxide gels Colloidal crystals Foams

carbons Sintered metals
Zeotypes Carbons (hard or Carbon gels Extruded monoliths
Tetrahedral/ Oxide gels soft templated) ... Ceramics
Compacted
octahedral ... Oxides (hard or powders Compacted
frameworks soft templated) ... powders
...
Octahedral ...
molecular
sieves
MOFs

...

Figure 1 Classification of porous solids. Reproduced from Schuth, F. Chem. Ing. Tech. 2010, 82, 769.

5.01.2.1.3 Ordered mesoporous materials
In contrast to the two classes discussed above, ordered meso-
porous materials are not periodic on the atomic length scale,
but on the scale of several nanometers.6 On the atomic scale,
the materials are mostly disordered, although in some cases the
mesostructure is formed from nanoparticles, which are crystal-
line on the atomic scale, so that there are two length scales on
which there is order.
The synthesis of such materials proceeds via templating,
using soft templates, such as surfactants that can form liquid
crystalline phases. Alternatively, ordered mesoporous mate-
rials, which were obtained by soft-templating, can in turn be
used as hard templates to replicate their pore system with a
different chemical composition. As hard templating allows
heat treatment at high temperatures, which is not possible for
soft-templated materials without decomposing the organic
templates, it is typically easier to achieve crystalline wall struc-
tures for hard-templated ordered mesoporous solids.
The diffraction analysis of ordered mesoporous materials
presents challenges in different respects: the repeat distances
are typically rather long, on the order of several to several tens
of nanometers. For XRD, this means that the first reflections
(which are incidentally often the high-intensity reflections) are
close to the primary beam. Therefore, it can be necessary to use
dedicated small angle diffractometers for the analysis of
ordered mesoporous materials. Moreover, typically there are a
very limited number of reflections that can be observed on a
laboratory diffractometer, and thus unambiguous determina-
tion of the mesostructure by XRD alone is often impossible.
5.01.2.1.4 Photonic crystals
Photonic crystals are macroporous materials with interesting
properties, especially optical properties.7 They are regular
arrangements of matter on the scale corresponding roughly to
the wavelength of visible light, that is, on the order of several
100 nm. Due to the fact that photonic crystals are periodic on
the same length scale as light, they influence the propagation
of light. XRD does not provide substantial information on the
structure of photonic crystals. However, neutron diffraction is a
suitable technique. Further, light diffraction could be used for
the analysis of photonic crystals. However, as visible light and
the photonic crystal interact strongly with each other, multiple
scattering occurs which makes structure elucidation very
difficult.
5.01.2.2 Disordered Porous Material
In many of the technologically relevant porous solids, the pore
systems are disordered, even if there might be crystallographi-
cally ordered domains in the solid fraction of the material.
These disordered porous solids include oxide gels (silica gel
probably being the best-known example), carbons, compacted
powders, extrudates, etc. In such materials, as the pores are not
periodically arranged, no Bragg diffraction, which would allow
us to conclude on the porosity, occurs. Thus, no distinct reflec-
tions indicative of the pore systems are observed, if electromag-
netic radiation is scattered from such samples. Nevertheless,
small-angle scattering (SAS), both of x-rays (SAXS) and
neutrons (SANS), can provide information on the porosity of
Diffraction and Spectroscopy of Porous Solids 3
such materials, as the contrast between matter and void influ-
ences the propagation of radiation in the materials. SAXS and
SANS are therefore valuable techniques for the characterization
of porous materials and are discussed in this chapter.
5.01.3 Scattering and Diffraction Methods
Scattering of waves is a common phenomenon that is generally
observed if waves meet objects that are of comparable sizes as
the wavelengths of the waves. Nautic waves are scattered by
larger rocks, radio waves by larger objects such as houses, and
visible light waves are scattered by microscopically small water
droplets in mist. The results of the scattering are new wave-
fronts that propagate from the objects that are met by the
incident waves. x-Rays and neutron radiation are electromag-
netic waves that can be also scattered. If x-ray or neutron
wavefronts meet atoms, the electrons (x-ray) or nuclei (neu-
tron radiation) of the atoms resonate with the radiation and
emit secondary waves into all directions. The scattered waves
that are of interest here have the same wavelength as the
incident beam (elastic scattering). The intensity of the scattered
radiation, that is, the square of the amplitude of the scattered
wave, depends on specific properties of the atoms that are
exposed to the radiation. As each electron interacts with
x-rays from the primary beam, the scattering intensity scales
with the number of electrons in the atoms, that is, with the
atomic numbers. For neutrons, the scattering intensity depends
on the coherent neutron scattering lengths of atomic nuclei.
The intensities that are scattered by individual atoms are thus
element specific. Scattered waves that are emitted by two indi-
vidual atoms that are close to each other superimpose. The
generated interference pattern depends on the intensity of the
scattered waves and on the distance and the orientation of the
two atoms relative to each other and to the incident wavefront.
An interference pattern that is caused by scattering from atoms
in a crystal structure is thus determined by the types of atoms
(elements) and the mutual distances and the orientations of
the atoms relative to each other.
In crystalline materials, atoms are located periodically on
defined positions. Thus, specific interatomic distances are
found more often whereas others will not be found at all.
The interatomic distance distribution is discontinuous for a
crystalline material and has discrete maxima. This is reflected
by discrete interference maxima in the interference patterns,
denoted as diffraction patterns. Structure determination of
crystalline matter by analysis of the diffraction patterns is the
subject of single-crystal and powder diffraction studies.
In contrast to diffraction on crystalline materials at higher
angles, scattering at small angles is not determined by the
arrangement of atoms within a crystal structure of a material
but by the overall shape and size of particles that are investi-
gated. For homogeneous particles, electron densities and/or
coherent neutron scattering length densities are considered to
be constant throughout the particle and scattering of x-rays or
neutrons is basically determined by sets of distances that are
observed for distinct particle shapes. The histogram of all
possible distances within a given particle is denoted as the
pair distance distribution function (PDDF). The maximum
distance in a PDDF corresponds to the maximum diameter of
4 Diffraction and Spectroscopy of Porous Solids
a given particle. PDDFs are thus characteristic for specific par-
ticle shapes. The relationship between the interference pattern
caused by scattering on particles and the PDDF will be
discussed below.
Crystal structure on the one hand and size and shape of
particles on the other hand affect diffracted or scattered intensi-
ties at different angular ranges of scattering curves. Longer dis-
tances, such as those relevant for scattering from particles, are
reflected by scattering intensities in the vicinity of the incident
beam and are subject of SAXS or SANS analyses. Shorter dis-
tances, representing atomic arrangements within crystalline
materials, cause diffraction patterns at larger angles and are
subject to single-crystal and powder diffraction studies using
either x-ray or neutron radiation. However, for materials with
large unit cell scattering due to particles and diffraction from the
crystal structure may be observed in the same angular range.
XRD or scattering experiments can be performed either on
smaller laboratory x-ray devices, for example, using sealed x-ray
tubes or rotating anode x-ray sources, or at synchrotron radia-
tion facilities. The respective neutron diffraction or scattering
experiments have to be performed at nuclear reactors or spall-
ation sources.
5.01.3.1 Crystal Structures
Crystal structures of solids are formed by periodic arrange-
ments of atoms, ions, or molecules. The periodic order can
occur in two dimensions as in graphene or in three dimensions
as in graphite as illustrated in Figure 2.
(a)
Figure 2 (a) Two-dimensional (2D) ordered arrangement of carbon atoms in graphene and (b) 3D arrangement of carbon atoms in graphite.
Na a a
Cl
(a) (b)
Figure 3 (a) Lattice row, (b) lattice plane, and (c) unit cell of the NaCl structure.
Each crystal structure can be described in terms of a small
repetition unit, the so-called unit cell. The unit cell is a struc-
ture unit that contains all information about the entire crystal
structure, such as size of the repetition unit (unit cell parame-
ters), symmetry elements (summarized by the space group),
and atomic positions (defined by coordinates xi, yi, zi). Trans-
lation of the unit cell in three dimensions generates the whole
crystal structure. As an example, Figure 3 shows how the
simple unit cell of the NaCl structure can be generated step
by step. The periodic arrangement of Na and Cl ions in one
direction is represented by a lattice row (Figure 3(a)). The
shortest repetition unit in this direction is defined as the length
from one atom to the next identical atom. The length of that
repetition unit is denoted as lattice parameter a. Continuing
the periodic arrangement in the direction of the second lattice
parameter b results a two-dimensional (2D) (ab) lattice plane
(Figure 3(b)). Finally, Figure 3(c) shows the unit cell of NaCl
after adding alternating Na and Cl ions also in the direction
of the third lattice parameter c. The shown unit cell contains all
information to build an infinite crystal.
Unit cells are defined by the three unit cell parameters a, b,
and c and by the angles a, b, and g which the edges of the unit
cell form with each other. Atomic positions within unit cells
are usually given as fractionate coordinates x, y, z of the three
unit cell parameters.
5.01.3.2 Principles of XRD
Almost exactly 100 years ago, Max von Laue and his coworkers
Walter Friedrich und Paul Knipping discovered the diffraction
a
b
(b)
b
c
b
a
(c)

of x-rays by crystals. They confirmed the wave character of
x-rays by diffraction of electromagnetic radiation by a crys-
tal810 and they were the first describing the interference effect
caused by the interaction of x-rays with the ordered arrange-
ment of atoms in ZnS, NaCl, and PbS structures.
From the curious arrangement of spots observed on a
photographic plate, father and son Bragg developed shortly
afterward a rule that later on became famous as Braggs
law.11,12 In the year 1913, W.H. Bragg and W.L. Bragg pub-
lished a series of papers on the theory of diffraction of x-rays by
crystals and on the crystal structures of rock salt, diamond, and
other compounds.1316
The spots that were observed on the photographic plates
were the result of constructive interference of x-rays that were
scattered by the periodic arrangements of the atoms in the
crystal structures. They are referred to as reflections. The con-
ditions that have to be fulfilled for reflections to occur are quite
simple and the related equations are closely associated with the
names of von Laue and Bragg. The relationship between x-rays,
crystals, and diffraction phenomena has been revealed by von
Laue and his coworkers Knipping and Friedrich.9,10 Laue
derived equations that describe the interaction of incident
waves via three basis vectors and the angles that they form
with the incident beam. All three Laue equations are simulta-
neously fulfilled for the case of positive interference of the
scattered waves. Alternatively, father and son Bragg derived
an equation that describes the diffraction of x-rays by lattice
planes. The so-called Bragg equation (Braggs law) can be
explained by the scheme shown in Figure 4.
Crystallographic planes are described by the Miller indices
(hkl) and the interplanar spacing d. Incoming waves that are in
phase with each other are reflected from each of the planes.
Regarding Snells law, the angle y of the incident beam and the
angle of reflection are identical. For constructive interference, it
is necessary that the waves that are reflected from the different
planes are in phase. This requirement is only fulfilled for cases
where the path difference between waves that are scattered
from different planes is nl.
As can be derived from Figure 4, the path difference
between waves reflected from neighboring planes is 2d siny.
If the path difference is an integer, n, of the wavelength l, the
phase difference is zero and the two waves are in phase. Max-
imum amplitude or constructive interference is obtained only
for specific angles yn (see Bragg equation, eqn [1]).
2dhkl sin yn nl [1]
Usually, the wavelength l is known and yn can be measured
(determination of the position of a reflection). With this infor-
mation, the interplanar spacing d can be determined. To assign
1
2 Miller indices
O
q q (hkl)
d
(hkl)
(hkl)
Figure 4 Diffraction of x-rays on lattice planes.
Diffraction and Spectroscopy of Porous Solids 5
interplanar spacings to certain lattice planes, the respective Miller
indices are usually used as suffix to d as shown in eqn [1].
In a diffraction experiment, incoming x-rays interact
with the electrons within atoms of a crystal. If collisions
between photons and electrons lead to a partial loss of energy,
this process is denoted as incoherent scattering (Compton
scattering). Incoherent scattering contributes to the noise/
background of a diffraction pattern. However, in most cases,
the major fraction of the incident electromagnetic radiation
interacts with the electrons of the atoms in a different way. The
electrons start to oscillate at the same frequency as the incident
x-ray wave and in turn become the origin of a new wave with
the same wavelength as the incoming radiation. The interac-
tion can be considered as a perfectly elastic collision or a
coherent scattering process (Thomson scattering).17 Each
atom becomes the origin of secondary waves that propagate
spherically from their origin.
As already discussed, the intensity of the scattered wave
depends strongly on the number of electrons in a given atom.
The scattering powers of individual atoms are expressed by the
so-called atomic form factors, fj, which scale with the atomic
numbers of elements. At a diffraction angle of zero, the atomic
form factor corresponds exactly to the atomic numbers. How-
ever, secondary waves that are emitted from electrons within
an atom are exactly in phase only in direction of the incident
beam, that is, at diffraction angle zero. At higher diffraction
angle, waves that originate from different positions in a given
atom are slightly off phase, which results in a decrease of the
atomic form factors, fj, with increasing diffraction angle. This
has to be considered in structure analysis.
In a diffraction experiment, all atoms in a given periodic
structure serve as origins of scattered waves. However, because
of the periodicity and the symmetry of the crystal structure, the
number of atoms that have to be considered as scattering
centers is reduced to the ones in the unit cell. Secondary
waves originating from all atoms in the unit cell superimpose.
Most diffraction experiments are performed with a constant
monochromatic x-ray wavelength. Thus, l is fixed in the Bragg
equation and only dhkl and y remain as variables that mutually
determine each other. As illustrated in Figure 4, the lattice planes
of a crystal structure are defined by the Miller indices (hkl). The
distances, dhkl, between lattice planes depend on the arrange-
ment of atoms within the unit cell which can be described by
the symmetry relation and the size of the unit cell. In turn, the
dhkl values define the diffraction angle at which given (hkl)
reflection are to be observed (see the Bragg equation, l is
fixed). The amplitude and the phase of the diffraction from a
given set of (hkl) lattice planes depend on the positions (xj, yj, zj)
of all N atoms with atomic form factors fj and can be calculated
via summation of the contributions from all atoms (j 1 to N):
X
N   
Fhkl fj exp 2pi hxj kyj lzj [2]
j1
Fhkl is denoted as the structure factor, its amplitude is its
modulus |Fhkl|, and its phase is denoted as ahkl. Constructive
interference of diffracted waves is only observed if the Bragg
equation is fulfilled. Diffraction intensities Ihkl are proportional
to the square of the structure factors |Fhkl|2, that is, the product
of the structure factor and its complex conjugate.
6 Diffraction and Spectroscopy of Porous Solids
In addition, measured diffraction intensities also depended
on the Lorentz factor, L, the polarization factor, P, and the
intensity I0 of the incident beam. The polarization factor P
1 cos 2 2y
P [3]
2 Synchrotron Radiation Facility (ESRF) in Grenoble enabled
is a function of the scattering angle and is independent of the
scattering experiment. The Lorentz factor depends on the dif-
fraction angle and the experimental setup.18 Diffraction inten-
sities thus are basically determined by
Ihkl I0 Fhkl
2
LP [4]
For a given structure model, structure factors can be calcu-
lated with eqn [2], which can be also written in a somewhat
more convenient way as
X tion, the different types of radiation sources should be consid-
Fhkl fj eidj [5]
j about the crystalline material and a combination of both can
with dj being the phase differences that calculate as dj 2p
(hxj kyj lzj). Calculation of structure factors Fhkl for structure
models thus permits calculation of theoretical diffraction pat-
terns for these structure models. Comparison of measured and
calculated diffraction patterns then allows a successive refine-
ment of a given structure model. For the structure models, both
amplitudes and phases can be calculated explicitly.
If, on the contrary, one now tries calculating structure fac-
tors Fhkl directly from measured intensities, one encounters the
problem that the amplitudes |Fhkl| can be derived from
  r tion angle. Neutron diffraction can distinguish between iso-
  Ihkl
Fhkl  [6]
  I0 LP
however, the exact phases remain unknown. This is known as
the phase problem in crystal structure determination. As Ihkl is
proportional to the square of the structure
 factor Fhkl, the phase
p
information is lost as Ihkl  Fhkl .
Ways of dealing with that problem will be described below.
At this point, one should point out that the angle at which
diffraction occurs and therefore the position of reflections is
only determined by the size of the unit cell and the symmetry
of the crystal structure. The intensity of the individual reflec-
tions and the intensity distribution are mainly determined
by the positions the individual atoms occupy within the unit
cell and their atomic form factors. The scattered intensity at
a defined angle is the result of the combination of waves
scattered from different atoms in the same direction. However,
measured diffraction patterns do not contain any phase
information. For the determination of the crystal structure,
single-crystal techniques are the method of choice. However,
single crystals comprise only a minute fraction of materials
even from a very small sample and the crystal might not be
representative for the bulk material. Single crystals also do not
allow representative analyses of microstructural properties for
a polycrystalline sample with respect to stacking faults, dislo-
cation densities, size/strain effects, or the quantification of
phase mixtures.
Single-crystal or powder diffraction techniques fail if the
porous material is microcrystalline.
If the crystals are too small, the intensities of the diffraction
peaks will be very weak and structure solution becomes
impossible. For smaller crystals, synchrotron radiation with a
high flux might be an option. For minute crystals with sizes in
the range of a few micrometers (12 mm), techniques have
been developed for analyzing the average crystal structure. As
an example, the Laue microdiffraction setup at the European
data collection from a single crystal with a spot size of about
1  1 mm2.19 The crystal structure of a very complex MOF, MIL-
110, could thus be precisely determined by a combination of
direct methods on the microdiffraction data, solid-state
nuclear magnetic resonance (NMR) spectroscopy, and molec-
ular modeling.19
For diffraction experiments not only x-rays but also neu-
trons or electrons can be used. Depending on the material
properties and on the individual questions related to diffrac-
ered. x-Ray and neutron data contain different information
be a powerful tool for solving structure problems.
While x-rays are interacting with the electrons of the atoms,
neutrons are diffracted from the nuclei of atoms. The interac-
tion of x-ray photons with the electrons leads to higher inten-
sities of the diffracted x-rays for crystalline materials that
contain atoms with high atomic numbers. However, as
pointed out above the atomic form factor and the scattering
intensity of an atom decrease with the scattering angle.
Neutrons, on the other hand, interact with the nuclei and
the diffracted intensity is therefore not affected by the atomic
number and the scattering power is independent of the diffrac-
topes of one element. For example, the different scattering
lengths of hydrogen (scattering length b  3.7423 fm) and
deuterium (b 6.674 fm) influence the diffracted intensities
and this can be used to distinguish both. Neutrons also enable
the localization of light elements that are basically invisible
for XRD. From single-crystal data, hydrogen with carbon atoms
as neighbors can be determined about 12 times more accu-
rately with neutrons than with x-rays.20
Neighboring elements in the periodic table are hardly dis-
tinguishable with x-rays because the number of electrons is not
significantly different. However, they may have significantly
different scattering lengths for neutron radiation, which may
allow the evaluation of the site distribution of elements with
similar atomic number such as Al/Si in zeolite structures. The
crystal structure of analcime, NaAlSi2O6H2O, has been inves-
tigated with respect to Al/Si disorder from single-crystal neu-
tron diffraction data.21 Hughes and Weller studied the Si/Al
distribution on T-sites for Cs4[Al4Si20O48] by neutron powder
diffraction. The combination of 29Si magic angle spinning
NMR experiments and structure refinement strongly suggests
that aluminum is preferentially located on one of the three
tetrahedral sites.22 Neutron diffraction also enables interpreta-
tion of atomic displacement parameters more precisely than
XRD because neutron diffraction does not need to consider the
polarization factor. For neutrons, the scattering intensity does
not decrease with diffraction angle. This results in better
defined reflections also at higher angles. As the high-angle
region contains valuable information on smaller interatomic
distances, more structural details become accessible. Finally,
diffraction experiments can be also performed with electrons.
 Diffraction and Spectroscopy of Porous Solids 7

Electrons are charged particles but due to their dual character as
particles and waves they interact with periodically arranged
atoms in a very similar manner as x-rays and neutrons. The
major advantage of electron diffraction is the fact that struc-
tural studies can be performed even for nanoscopic crystallites.
For choosing the appropriate diffraction technique, the prop-
erties of the sample and of x-ray and neutron radiation should
be considered (Table 1).
5.01.3.3 Single-Crystal Diffraction
Single-crystal diffraction provides detailed information about
the internal arrangement of atoms of crystalline compounds.
This includes information on the sizes of unit cells, symmetry
elements, bond lengths, and bond angles. If the scattering
properties of different atoms differ significantly, the popula-
tion of different atoms on one crystallographic site (ordering)
can be determined. However, single-crystal diffraction requires
the availability of single crystals of sufficient size and quality.
For conventional laboratory diffractometers, the ideal size of
the single crystal should be between 30 and 300 mm.
Nowadays, the solution of crystal structures from single-
crystal data is usually a very straightforward procedure with
software packages that can be easily used. Single-crystal neu-
tron diffraction requires large crystals of millimeter size. There-
fore, this method is less commonly used.
5.01.3.3.1 Structure solution from single-crystal data
Direct methods are based on the use of the relationship
between intensities and phases of reflections. In 1985, the
Nobel Prize for chemistry was awarded to Jerome Karle and
Herbert A. Hauptman for their contribution to the develop-
ment of direct methods for the solution of crystal structures. A
crystal structure can be described as electron density distribu-
tion r(r). The electron density can be thought as superposition
of density waves parallel to the lattice planes.23 However, the
electron density determines the diffraction pattern of a crystal.
The amplitude and the phase of each scattered wave depend on
the crystal structure. The information about phase and ampli-
tude lies in the structure factor Fhkl. As discussed above, the
intensity of the scattered radiation is proportional to the square
of the amplitude |Fhkl|2. Direct methods, that is, the direct
structure solution from measured diffraction patterns via inverse
Fourier transformations, suffer from the phase problem. The
phases of diffracted waves cannot be directly measured. How-
ever, the method allows ab initio structure determination from
measured amplitudes. The phase and the amplitude of a wave
are generally independent but they can be related via consider-
ation of certain properties of the electron density function.24 The
electron density is always positive and is composed of discrete
atoms. Phases are then assigned to a small subset of data. From
this first approximation, electron density maps are calculated
and successively refined.
5.01.3.3.2 The Patterson method
As pointed out above, Fhkl is not accessible from measured
intensities. Therefore, Patterson proposed using |Fhkl|2 for a
Fourier transformation instead of Fhkl. The Patterson function
thus uses squared amplitudes that are proportional to the
diffraction intensities. The Patterson function as Fourier trans-
formation calculates as follows:25
 
1 XXX 2
P u; v; w Fhkl  cos 2phu kv lw [7]
Vc
Phase information is not required. The function is evaluated
in real space and spans a 3D grid of points u, v, w with size and
shape of the unit cell. However, no direct structure information
is obtained. The Patterson function can be considered to be the
sum of appearances of a given structure by viewing it from each
atom in turn.26 The maxima of the Patterson function corre-
spond to a map of interatomic vectors between atoms in the
structure under investigation. As no phase information is avail-
able, the origins of all vectors lie on the origin of the unit cell.
The aim of the Patterson synthesis is then reconstruction of the
crystal structure from this set of interatomic vectors. If present,
heavy elements within a structure may serve as the starting
points for the reconstruction. Their location is easier and after
assignment of phases to the structure factor, the correctness of
the assignment can be verified for a subset of atoms.
5.01.3.4 Powder Diffraction
In addition to the information on crystal structures, powder
diffraction data provide a variety of useful information about
solids. The composition of phase mixtures can be analyzed
both qualitatively and quantitatively. Further, fractions of
amorphous solid in a crystalline sample can be determined
quantitatively. Microstructure properties, such as size, strain,
and stacking faults or other types of dislocations, can be
deduced from the evaluation of diffraction profiles. Further-
more, the effect of temperature, pressure, and gas atmosphere
on phase mixtures, different polymorphs, or on crystallization
conditions can be investigated best by powder methods.

Table 1 Comparison of different radiations for crystal structure determination

x-Rays Neutrons Electrons

Small sample amounts required
Elements with very similar atomic number and
isotopes cannot be distinguished
Light elements are difficult to analyze
Determination of atomic structure
Large sample amounts required
Elements with very similar atomic number and
isotopes can be often distinguished
Light elements are good to analyze
Determination of magnetic and atomic structure
Minute sample amounts and nanoscopic
crystals required
Elements with very similar atomic number and
isotopes cannot be distinguished
Light elements are difficult to analyze
Determination of atomic structure
8 Diffraction and Spectroscopy of Porous Solids
5.01.3.4.1 Structure solution from powder diffraction data
If single crystals of sufficient size and quality are available,
single-crystal XRD is the approved method for solving crystal
structures. Before the late 1980s, structure determination was
very complicated if not impossible for materials comprising
crystals too small to perform single-crystal analysis. A tremen-
dous step forward in deduction of structural information from
powder data was achieved when Rietveld introduced a method
for the structure analysis from powder data by curve-fitting
procedures.27,28 The big advantage of this method was that it
allowed overcoming the problem of substantial overlapping of
diffraction peaks that was often encountered. Integrated inten-
sities had to be assigned to several reflections, which resulted in
situations in which the number of variables was substantially
lower than the number of observed independent data points.
The systems were strongly underdetermined. In the Rietveld
method, the intensity information of each data point in a
diffraction pattern contributes to the structure refinement.
In this way, the number of independent observables was
increased substantially and crystal structure refinement became
feasible. From a known or assumed crystal structure, a powder
diffraction pattern is generated by calculating intensities yic for
each individual data point j by the following equation:
X
k2
yjc yjb Gjk Ik [8]
kk1
with Ik as the intensity of Bragg reflection k, yjb as intensity of the
background at point j, and Gjk as the peak profile function. This
calculated XRD pattern is then compared to a measured diffrac-
tion profile and the differences between the observed and the
calculated profile are minimized by least-squares refinements.
The model is stepwise adjusted until the best fit of calculated
and observed patterns is obtained. The minimization is per-
formed over all measured data points of the pattern, not only
for integrated Bragg reflections. The sum of the squared differ-
ences multiplied by a weight factor is minimized as
X transformation. Random phases are assigned to the structure
R wi yio  yic 2 [9]
i
with yio being the observed intensity at position i, yic being the
calculated intensity at position i, and wi being the weight
assigned to each intensity.
The method, initially developed for neutron diffraction
data, nowadays is widely used for XRD data as well. A very
comprehensive monograph by R.A. Young provides an excel-
lent introduction to the Rietveld method and its application in
modern structure refinement.29 However, the Rietveld method
requires a structure model for the refinement and direct solu-
tion of structures is not possible with this approach. In the early
1980s, ab initio determination of structures from powder data
became feasible but only a few structures were solved by ab
initio methods30,31 or direct methods32 at that time. Structure
solution from powder data was not a prevalent method. In the
1990s, the solution of crystal structures from powder data was
driven by the development of methodologies and computer
technology. The problem which has to be dealt with is the fact
that the 3D reciprocal space is projected onto a 1D powder
pattern. The consequence is a loss of information. Thus, the
diffraction lines are partially or completely superimposed. Out
of this, individual reflections cannot be indexed unambiguously
and the determination of the integrated intensities of the indi-
vidual reflections is not possible. The overlap of reflections,
texture effects, and intrinsic line broadening makes structure
solution from powder data a very complicated endeavor if not
impossible.
The techniques developed for the solution of structures
from powder data are basically based on two approaches. The
first one is in analogy to the treatment of single-crystal data.
First, the intensities of the Bragg reflections are extracted from
the diffraction peaks and then these intensities are used in
direct methods or Patterson methods combined with Fourier
analysis. If reflections are intensively overlapping or the struc-
tures are very complex, the structure solution may fail misera-
bly. The second approach is based on direct-space models.33
Trial structures are generated in direct space and their calculated
powder patterns are compared with experimental patterns.
The structure corresponding to the lowest figure-of-merit
value (R-factor) represents the structure with a global mini-
mum on the hypersurface. The methods used for global opti-
mization can be the Monte Carlo method, simulated
annealing, or genetic algorithms. An excellent overview of the
structure determination from powder data is given by David
and coworkers.34
Several program packages have been developed enabling
the solution of crystal structures from powder data. Just as
one example, the program package EXPO2009 has been devel-
oped for structure solution in direct and reciprocal space.35 The
charge flipping method is very powerful to solve crystal struc-
tures without any knowledge on the chemical composition or
the symmetry of a given material. Originally, the method had
been introduced by Oszlanyi and Suto for structure solution
from single-crystal diffraction data.36 The method uses infor-
mation from reciprocal space contained in the diffraction data
(structure factors) and from direct space reflecting the crystal
structure (electron density).37 Real-space charge density and
reciprocal space structure factors are related by discrete Fourier
factors obtained from the diffraction data. By performing a
Fourier transformation with these structure factors, an electron
density map is calculated in real space. In a next step, the sign of
density for all values below a certain threshold is changed
(flipped). From this new electron density map, new structure
factors are calculated. The calculated data are then combined
with the experimental data for the next iteration step until a
satisfying structure solution is obtained. Baerlocher et al. adapted
this algorithm to solve crystal structures from powder diffraction
data of zeolites with strongly overlapping reflections.38
5.01.3.4.2 In situ diffraction
A significant number of studies using in situ XRD methods on
porous materials were published over the last decades. Of very
fundamental interest are in situ studies on nucleation and
growth of solids in order to understand the formation pro-
cesses of these materials and on catalytic reactions with the aim
to investigate a solid catalyst under reaction conditions. One
apparent problem is the construction of in situ cells that can be
operated under realistic conditions. The chambers have to
resist high temperatures, pressures, and/or higher vacuum, as
well as reactive gas atmosphere. If crystallization processes are

studied, the cell has to be chemically resistant to the reaction
solutions, which are often rather aggressive at elevated temper-
atures. The high brilliance of synchrotron sources and the high
energies of neutron radiation allow the design of furnaces that
are fabricated of appropriate materials. For this reason, most of
the data published were collected at synchrotron facilities but
there are also examples for the use of conventional laboratory
diffractometers.39
The in situ monitoring of phase transformations, segregation,
and melting processes at elevated temperatures provides useful
information on the behavior of porous materials under
nonambient conditions. The investigation of reactions under
hydrothermal conditions gives first insights into the reaction
mechanisms and the quantitative changes of phase compositions
during synthesis.
The first energy-dispersive in situ x-ray experiments that
were performed during hydrothermal synthesis of sodalite
and zeolite A and the requirements for a suitable reaction cell
were reported by Munn et al.40 The authors studied mixtures of
kaolinite and NaOH at increasing temperatures up to 110  C
and revealed the transformation of kaolinite to sodalite under
hydrothermal conditions. As another example, the synthesis of
open-framework gallium oxyfluorophosphates, potential cata-
lysts, was monitored by in situ energy-dispersive XRD. The
formation of transient intermediate phases that are unstable
under ambient conditions was observed. The knowledge about
the presence of such intermediate phases and the mechanism
of conversion substantially help to understand the kinetics of
the reaction system.41
Norby et al. performed several in situ synchrotron studies
monitoring processes during the hydrothermal crystallization
of inorganic materials, especially zeolites and aluminopho-
sphates.4246 With the results from the in situ investigations,
the nucleation and crystallization pathways of zeolites have
been modeled.
Studies on the formation of mesoporous silica MCM-41
reveal the timescale of crystallization.47 A tubular reactor
setup was mounted on a conventional laboratory diffractome-
ter. For the first time, the kinetics of formation of the hexagonal
surfactantsilicate mesophase could be monitored with the
setup shown in Figure 5. Already after 90 s the main reflections
of the hexagonal mesophase were detected, showing the very
rapid formation of the hexagonal micellar structure (Figure 6).
= Peristaltic pump
Teflon cell for
in situ XRD
measurements
Air Polypropylene tube
Aqueous
reactants
0.335 m
Static mixer(s), Air Reaction
ultrasonic solution
mixing
TEOS
Figure 5 Sketch of a tubular reactor setup. Reproduced from Linden,
M.; Schunk, S. A.; Schuth, F. Angew. Chem. Int. Ed. 1998, 37, 821823,
with permission from Wiley.
Diffraction and Spectroscopy of Porous Solids 9
Recently, Innocenzi et al. performed simultaneous in situ
and time-resolved grazing incident SAXS data in combination
with wide-angle x-ray scattering (WAXS) to study the concur-
rent formation of micelles and nanoscopic salt particles that
serve as templates for thin films with hierarchical pore
systems.48 From the studies, the kinetics of formation of
micelles and salt templates of different shape and length scale
could be deduced. As the formation and organization of the
salt crystals and micelles runs on a timescale of only seconds,
rapid data acquisition was mandatory, which asked for an
experimental setup using synchrotron radiation.
Another very interesting area where in situ studies are
highly valuable are studies on the behavior of porous materials
during the adsorption of guest molecules. This was the topic of
numerous publications of the last decade. As an example, in
situ diffraction studies allow monitoring of breathing effects of
MOFs during the exposure to organic molecules in different
liquid phases. Details on hostguest and guestguest interac-
tions could be revealed by the structural evaluation of in situ
synchrotron radiation data.49 For the incorporation of differ-
ent guest molecules into the porous structure of MIL-53, sig-
nificant changes of the unit cell volume associated with
substantial changes of the symmetry of the crystal structure
were observed as shown in Figure 7.
Matsuda et al. report on in situ synchrotron radiation stud-
ies on porous coordination polymers during the adsorption of
different guest molecules.50 To increase the effective interac-
tion between adsorbed benzene molecules and the pore walls,
a large contraction of the cell takes place as illustrated in
Figure 8. The shape-responsive fitting properties of flexible
frameworks are discussed to be suitable for molecular recogni-
tion or selective guest inclusion.
5.01.3.5 Electron Diffraction
Transmission electron microscopy (TEM) is the appropriate
method to gain information about the electrostatic potential
distribution within a specimen. Electrons pass through a sample
(100)
1000
hk0 d/nm
100 4.43
800
110 2.57
200 2.23
600
I
(110)
(200) d
400
c
b
200
a
0
1 2 3 4 5
2q/
Figure 6 In situ XRD patterns of MCM-41 as measured in a tubular
reactor at different reaction times: (a) no TEOS added tR 0 s, and after
reaction times tR of (b) 15 s, (c) 90 s, and (d) 180 s. Reproduced from
Linden, M.; Schunk, S. A.; Schuth, F. Angew. Chem. Int. Ed. 1998, 37,
821823, with permission from Wiley.
10 Diffraction and Spectroscopy of Porous Solids

D Acetonitrile
Lutidine
1.0 Butan-1-ol
a Methanol b
d
DEF
Imcm Ethanol
0.9 Toluene a
V = k sin [2 arctan (d/D)] Nitrobenzene
m-Xylene
d
Ethyl acetate

0.8 DEF
Chloroform

THF (a)
0.7 D
Quinone
DMF
C 2/c Pyridine
Pnam
d/D

0.6 DMSO
Propan-2-ol

0.5 b
Watermethanol a
Water intermediate phase
0.4 d

0.3

1000 1100 1200 1300 1400 1500 1600

(b)
V/3 D

Figure 7 Comparison of the changes of the unit cell volume versus d/D (lozenge diagonals in inset at top left) for metalorganic framework MIL-53 due
to incorporation of various guest molecules (left). During swelling of the structure, changes of the structure symmetry were detected. Structures of
MIL-53(Fe) (right) loaded with H2O (top) and pyridine (bottom). Reproduced from Millange, F.; Serre, , C.; Guillou, N.; Ferey, G.; Walton, R. I. Angew.
Chem. Int. Ed. 2008, 47, 41004105, with permission from Wiley.

Electron beam
Sample

Microscopy
Shape-responsive SAED
fitting

Rectangular shape pore Z shaped pore Micrograph

FT of 2D image
Figure 8 Pore structure of a coordination polymer host and the
structural changes upon adsorption of benzene in the pores. Reproduced SAFD (repeat for
from Matsuda, R.; Kitaura, R.; Kitagawa, S.; Kubota, Y.; Kobayashi, T. C.; different zone axes)
Horike, S.; Takata, M. J. Am. Chem. Soc. 2004, 126, 1406314070, with Processing
Combine all reflections
permission from The American Chemical Society.
intensity and phase
Complete dataset of
amplitudes and phases
by elastic and inelastic scattering processes. Information about FT with 3D symmetry
the crystal structure is deduced from elastically scattered elec-
trons. Interactions between electron and matter are strong and a Electron density map
highly coherent and bright electron beam can be used to obtain (structure)
electron diffraction patterns.51 The beam is focused on a small
Figure 9 Sketch of methods used in electron crystallography.
part of the sample and the diffraction pattern is obtained by Reproduced from Anderson, M. W.; Ohsuna, T.; Sakamoto, Y.; Liu, Z.;
insertion of an aperture in the first intermediate imaging plane Carlsson, A.; Terasaki, O. Chem. Commun. 2004, 907916, with
as shown in Figure 9 (selected area electron diffraction, SAED). permission from Marcel Dekker.
Lens aberrations limit the smallest area selected to about
0.1 mm. The diffracted beam can be refocused to form an
image which is the electron micrograph. (0.005 nm at 60 kV electrons compared to 0.15 nm for
Crystal structures can be solved from the electron micro- x-rays) resulting in much more reflections collected in the
graph, which is a hologram, constructed from the diffracted same angular diffraction range. In addition, electron diffrac-
beams and carries all information about the amplitude and the tion can be combined with energy-dispersive x-ray spectros-
relative phases.52 Compared to XRD experiments, the wave- copy for the analysis of the chemical composition of a sample.
length of electrons used for electron diffraction is much shorter The electronic structure and bonding is accessible through

electron energy loss spectroscopy experiments. Carlsson et al.
were able to determine the atomic positions of an Fe6On cluster
inside the sodalite cage of zeolite Y (Figure 10).53 From high-
resolution micrographs and SAED, the atomic positions could
be reconstructed.
Very important for a successful analysis is the careful prep-
aration of ultrathin sections of the crystals of comparable
thickness. The collected images were Fourier transformed and
corrected for the effect of the contrast transfer function of the
objective lens. The accuracy is not as high as in XRD, but the
accuracy is good enough to deduce the position of the cluster
and the general shape of the framework.53
Recently, novel methods for structure solution from powder
diffraction data have been established. Thus, the structures of
complex 3D zeolite structures could be solved. The structure of
SSZ-48 with 12-membered ring openings has been solved by
direct methods from electron diffraction data collected from a
0.05  0.25  10 mm crystal.54 It also has been shown that the
combination of powder diffraction and electron diffraction data
is a powerful tool to solve large, complex zeolite structures. The
unit cell parameters and space group assignment are in agree-
ment with the synchrotron powder x-ray data. The crystal struc-
ture, determined by electron diffraction refinement, was
confirmed by high-resolution TEM (HRTEM) imaging and was
also confirmed by Rietveld refinement of the powder synchro-
tron diffraction data.
The group of McCusker and Baerlocher developed
methods to solve complex structures by a combination of
high-resolution powder diffraction with HRTEM.5557 The over-
lap of reflections, and therewith the intensity determination of
the individual Bragg reflections contributing to measured peaks,
restricts the solution of structures from powder diffraction data.
Extensive peak overlap is problematic for structures with large
unit cells such as zeolites. Once the intensities of overlapping
reflections can be determined, the same methods as for single-
crystal structure solution (e.g., direct methods) can be applied to
the powder data. The program FOCUS55 has been devolved to
solve zeolite crystal structures. It uses the combination of pow-
der diffraction data and chemical information of crystals, such
as bond distances and/or connectivity, to determine crystal
structures of zeolites. For zeolites, the bond angles and the
bond lengths between framework atoms are usually known
Si
Fe
Figure 10 Position of the iron atoms of the Fe6On cluster inside the
sodalite cage as reconstructed from electron microscopy. Reproduced
from Carlsson, A.; Karlsson, G.; Bovin, J. O.; Oku, T.; Okamoto, Y.;
Ohnishi, N.; Terasaki, O. Eur. Phys. J. D 1999, 9, 623626, with
permission from EDP Sciences.
Diffraction and Spectroscopy of Porous Solids 11
and can be used as constraints. The algorithm combines exhaus-
tive framework search with Fourier cycling. Reflection intensities
are extracted from powder data while overlapping reflections are
equipartitioned and the starting phases are assigned randomly
to structure factors. From this starting model, an electron density
map is calculated and from this a possible topology is con-
structed. From this model, a new set of phases is calculated. If
the original phases and the new phases differ, the new phases
are used to generate a new electron density map. This Fourier
cycling is repeated until the phases converge. The combination
of information obtained from HRTEM images with FOCUS
enables the solution of complex structures of materials such as
TNU-9 with 24 symmetrically independent T atoms in the final
structure.57 The large unit cell with low symmetry C2/m and unit
cell parameters of a 28.2219 A, b 20.0123 A and
c 19.4926 A, b 92.33 was determined by SAED. From
high-quality HRTEM images, the similarity with the mobil five
framework topology was concluded. All information from
HRTEM was used as input to FOCUS.57
Corma and coworkers solved the crystal structure of the
hierarchical mesomicroporous zeolite ITQ-43 by means of dif-
fraction tomography.58 The unit cell parameters and the space
group could be determined from powder diffraction data. How-
ever, the large unit cell with a 26.090 A, b 41.866 A, and
c 12.836 A and the low symmetry leads to extensive peak over-
lap and a low resolution of individual reflections (Figure 11). As
larger single crystals were not available, the crystal structure could
not be solved from XRD data. Finally, the complex structure was
solved from electron diffraction data collected by automatic
diffraction tomography as shown in Figure 11. Electron diffrac-
tion patterns of crystals down to nanometer size are collected by
combination of a quasi-parallel beam allowing nanodiffraction
and scanning transmission electron microscopy for imaging. The
advantage of this approach is that 2D diffraction patterns are
collected and, by successive rotation of the sample, 3D diffrac-
tion data can be reconstructed. The method also prevents beam
damage of the sample and the area of diffraction is visible in real
space.59,60
These examples show that combinations of electron micros-
copy, electron diffraction, and XRD as well as newly developed
structure solution strategies for the individual methods are
powerful tools to reveal complex structures of porous solids
even for crystallites of nanoscopic sizes.
5.01.3.6 Small-Angle Scattering
SAS is observed if plane monochromatic waves of x-ray or
neutron radiation pass through matter with inhomogeneous
density distribution. The waves are then scattered and form
intensity patterns that are angle-dependent and characteristic
for the average shape of the particles or pores. SAS is sensitive
to density inhomogeneities on nanoscopic dimensions, for
example, domains of higher density (e.g., particles) or, alter-
natively, domains of lower density in a given matter (e.g.,
pores). This technique thus provides information about nano-
scopic particles and/or nanoscopic pores. SAXS is the result of
the interaction of x-rays with electrons within a given volume
of matter. Low-angle neutron scattering (SANS) is caused
mainly by interaction of neutron radiation with the nuclei of
atoms within a volume. The basic information that can be
12 Diffraction and Spectroscopy of Porous Solids

Intensity (arbitrary units)

19.6

9.7 4.9
21.9

7.0

0 5 10 15 20 25 30 35 40 45 50
2q ()

Figure 11 XRD pattern (left) and pore opening (right) of the ITQ-43 structure as formed by 28 Si(Al) atoms. Reproduced from Jiang, J.; Jorda, J. L.;
Yu, J., Baumes, L. A.; Mugnaioli, E.; Diaz-Cabanas, M. J.; Kolb, U.; Corma, A. Science 2011, 333, 11311134, with permission from American
Association for the Advancement of Science.

derived from both methods is quite similar if not identical.61 In

the following, SAS will be illustrated briefly for both methods.
In a typical SAS experiment, the incident beam passes
through a sample and the scattered intensity is recorded as a
function of the scattering angle 2y as illustrated in Figure 12.
In SAS, the measured intensities are commonly plotted in
2D plots as a function of the scattering vector denoted either as
q or s, with:
4p 2
q sin y or s sin y
l l
The scattering vector q is occasionally also referred to as
momentum transfer. The observed scattering intensity distri-
bution along the scattering angle 2y is the result of interference
of radiation that is scattered by atoms within the sample. If
pores exist in a material, the density within the pores is differ-
ent from that in the surrounding solid for which a homoge-
neous density is assumed. The pores are characterized by their
dimensions d. The scattered intensity that is observed in scat-
tering experiments then depends on the size d of the pores and
the difference of electron density (Dre) between solid and pore
for SAXS or the difference in coherent neutron scattering length
density (Drb) for SANS. In contrast to SAXS, scattered intensi-
ties in SANS are also sensitive to magnetic moments of mate-
rials. For empty pores, the density difference is characterized by
the average electron density of the atoms within the solid
(SAXS) or the average scattering length density (SANS). If the
pores are filled, the differences in electron densities or scatter-
ing lengths between the solid and the matter within the pores
are decisive.
SAS is generally sensitive to density inhomogeneities that
are caused either by nanoscopic pores and/or particles with
dimensions ranging from 1 nm to more than 100 nm. Accord-
ingly, SAS can be used to investigate pore characteristics in the
whole nanopore size range. However, as pointed out above,
SAS is only sensitive to density differences, irrespective of
the origin of the density differences. Scattering curves that
are caused by particles or by pores of identical shape and size
thus cannot be discerned, provided the respective relative
2q
m
nt bea
cide Sample
In
[10]
Figure 12 Schematic setup of SAS experiments.
differences in electron densities or scattering length densities
are of the same magnitude (Babinets principle).
In addition to the scattering due to density differences
within a material, background scattering is observed. For
x-rays, this is caused mainly by radiation that is scattered
elastically into all directions by the irradiated atoms. This
scattering background is superimposed on scattering (interfer-
ences) due to the presence of particles or pores within matter.
Background scattering is generally not angle-dependent and
has to be subtracted from scattering curves. Scattering of
x-rays by gas molecules from ambient air causes additional
background intensity that can be reduced by using evacuated
beam paths.
In the following, SAS will be illustrated for the case of small
angle x-ray scattering. SANS can provide similar information.
Both methods have been reviewed extensively in well-established
textbooks.6164
5.01.3.6.1 Small-angle x-ray scattering
All electrons in a solid object can be considered as sources of
scattered x-rays and thus contribute to the total scattered inten-
sity. In most cases, all electrons in the object contribute equally
to the total scattering intensity. The intensity, Ie, scattered by
each electron can be calculated by the Thomson formula:
1 1 cos 2 2y
Ie I0  7:90  1026  [11]
a2 2
 Diffraction and Spectroscopy of Porous Solids 13

with I0 being the incident intensity of the x-rays, 2y being the
scattering angle, and a being the distance of the object to the
point of registration.65 The electrons are origins of secondary x-
ray waves that interfere with each other. At larger 2y angles, no
scattering is observed due to destructive interference. At smaller
scattering angles, the phase differences become smaller and
waves will enforce each other due to constructive interference
with the maximum intensity in the direction of zero scattering
angle where all waves are in phase. An observed interference
pattern is a function of the distance distribution of the elec-
trons in a given object. The latter is described by a PDDF p(r). It
is defined by the geometry of the object, that is, the geometry of
particles or of pores, and can be considered as the number
distribution (histogram) of distances within a given object
weighted by the electron density differences.66,67 The scattering
curve, that is, the scattered intensity versus the scattering angle
or the scattering vector q, calculates as the Fourier transform of
p(r) according to
Z 1
sin qr
Iq 4p pr dr [12]
0 l
Consequently, the PDDF can be calculated via inverse Fou-
rier transformation from the scattering curve as
Z 1
1 The total scattering intensity curve has the same shape as
pr 2 4p Iq qr sin qr dq [13]
2p 0 that of a given individual object just multiplied with the total
This procedure allows determination of basic parameters
for a scattering object (particle or pore). From the PDDF, the
radius of gyration, Rg, can be calculated as
Z or pores) are present in the sample.
pr r 2 dr
Rg Z
2
[14]
2 pr dr
and the forward scattering, that is, the scattering intensity at
q 0, calculates as
Z it is clear that the minima are shifted to lower q with increasing
Iq 0 4p pr dr [15]
The maximum dimensions of a given object and the shape
of p(r) are characteristic for certain geometries of the objects.
This is illustrated in Figure 13 for a sphere and a cylinder with
L
R. The maximum distance for the spherical object is found
at r 2R with R being the radius of the sphere. For the cylin-
drical object, the maximum distance corresponds to the length
L of the cylinder and from end of the initial maximum, the
diameter Dc of the cylinder may be estimated.
All electrons in a solid particle will contribute to the scat-
tering of x-rays. The form factor F(q) of one particle then is
defined as the ratio of the total scattered amplitude and the
radiation scattered by one electron. The scattered intensity then
calculates as
Iq Ie F2 q [16]
The square of the form factor, F2(q), is characteristic of spe-
cific shapes of the objects as shown in Figure 14. The pro-
nounced minima are q positions at which F2(q) is zero. It
should be noted at this point that the terms form factor and
structure factor are used differently in the fields of classical
crystallography and SAS, as we will see later. If the objects have
identical shapes and sizes, the scattered intensity of all objects are
identical and the total intensities are the sum of identical F2(q) if
the distances of the object to each other are long. Thus, the
scattering intensity by N identical particles or pores in a given
volume calculates as
Iq Ie qNF 2 q [17]
number of all objects. Thus, if measured scattering curves show
discrete minima as shown in Figure 14 for different objects,
one can assume that relatively monodisperse objects (particles
The functions used for the calculation of the particle form
factors F(q), as shown in Figure 14, are listed in Table 2.
Polydispersity is reflected by less pronounced minima, if
not the complete absence of minima for very polydisperse
particles or pores. This is illustrated in Figure 15 for cylindrical
objects with different cylinder radii R. From the shown graphs,
radii of the cylinders. It is also clear that even though the radii
vary only from 5.0 to 6.5 nm, the minima of the total F2(q)
become very shallow and that the oscillations of minima that
are still present get quite complex. Fitting of convolutions of
calculated F2(q) for assumed size distributions to measured
data allows determination of particle size distributions or
pore size distributions for materials under investigation.

Sphere
2R q/nm-1
0.1 1 10
p(r)

1.0E + 00
Hollow cylinder
2R Solid cylinder
Solid sphere
1.0E - 01

1.0E - 02
Cylinder Dc
F 2(q)

1.0E - 03
p(r)

Dc 1.0E - 04
L L
1.0E - 05

r 1.0E - 06
Figure 13 Pair distance distribution functions, p(r), of spherical (top) Figure 14 Form factors F2(q) of spheres, cylinders, and hollow
and cylindrical (bottom) objects. cylinders with radii of 5 nm (for cylinders R Dc/2).
14 Diffraction and Spectroscopy of Porous Solids

For systems that consist of individual objects that are well If distances between objects are of the same magnitude as
separated and statistically distributed, the scattering intensities the sizes of the objects, the presence of these closeby objects
are basically determined by the particle form factors (see causes density variations that are reflected by distances that are
eqn [17]). This situation is illustrated in Figure 16 (left). Also similar to those that can be found within the objects. Scattering
shown is the so-called pair correlation function for that situa- due to the presence of nearby particles thus causes interferences
tion. The pair correlation function, g(r), basically describes the that are reflected by the same q range as scattering from the
probability of finding another particle at a certain distance r. particles themselves. At higher concentrations objects get closer
The distances between particles are rather large and the pair to each other (Figure 16, middle). As a consequence, the pair
correlation function has the value 1 which means that every correlation function differs from g(r) = 1 and fluctuations of
distance has the same probability to be found, except for g(r) indicate the presence of nearby objects at average distances
distances at which particles would meet each other. At dis- that correspond to the maxima of the fluctuations.
tances below that distance, R, obviously no other particle will Density fluctuations due to the statistical presence of
be found. objects at given distances, as described by the pair correlation
function g(r) in real space, are reflected in the scattered inten-
sities and can be described in reciprocal space by a function
Table 2 Particle form factors F(q) for different objects, with R being
the outer radius of the sphere and cylinders, J1 a first-order Bessel called the structure factor S(q). The close relation between g(r)
function, and  the factor that defines the shell thickness of the hollow and S(q) is illustrated in Figure 17.
cylinder as fraction of R108 The presence of nearby objects thus affects the scattered
intensities and I(q) can be calculated as:
Shape F(q)
Iq F2 q  Sq [18]
3 sin qR  qR cos qR
Sphere
qR3
2J1 qR
Observed scattering intensities are no longer directly
Cylinder proportional to F2(q) but still related to it. This is illustrated
h qR
i
Hollow cylinder 2 qRJ1 qR 2 qRJ21 qR in Figure 18 for block copolymer micelles in water.67 As can be
qR 1
seen easily, the measured intensities I(q) versus q differ signi-
ficantly from the square of the form factor F2(q). The reason for
that deviation from F2(q) is the high concentration of micelles,
q/nm-1 which is reflected by the structure factor S(q). The observed
0.1 1 10 intensities I(q) are obtained by multiplying the particle form
1.0E + 01
R = 5.0 nm factor F2(q) by the structure factor S(q).
R = 5.5 nm
R = 6.0 nm 1.0E + 00
R = 6.5 nm
Total 1.0E - 01 S(q)
g(r)
1.0E - 02
F 2(q)

1.0E - 03 1 1

1.0E - 04
r q
1.0E - 05
Figure 17 Pair correlation function g(r) and structure factor S(q) of
1.0E - 06 spherical particles.
Figure 15 F2(q) of solid cylinders with different radii (R Dc/2).
I(q) (arbitrary units) and S(q)

2
F2(q)

S(q)
1

I(q)

g(r) 0
1
0 1
q (nm-1)
R r r r
Figure 18 SAXS curve of 5% block copolymer P94 in water. Circles
Figure 16 Distances of particles in a dilute system (left), a system with correspond to desmeared measured data of I(q), rhombs to particle form
high concentration of objects (middle), and objects in a periodically factor F2(q), and triangles to structure factor S(q). Reproduced from
ordered array (right). Under the sketches pair correlation functions, g(r), Orthaber, D.; Bergmann, A.; Glatter, O. J. Appl. Cryst. 2000, 33, 218225,
for the individual situations are shown. with permission from International Union of Crystallography.
 Diffraction and Spectroscopy of Porous Solids 15

100
I(q)
S(q)
F 2(q)

100

110
200

210
300
220
310
400

200

210

400
220
310
110

300
0.5 1 1.5 2 2.5 0.5 1 1.5 2 2.5 0.5 1 1.5 2 2.5
q (nm-1) q (nm-1) q (nm-1)
Figure 19 Convolution of particle form factor and structure factor for SBA-15 silica with lattice parameter a 12.5 nm and pore radius Rpore 4 nm.

Finally, if the objects get aligned in ordered arrays (Figure 16, 1

right) in one, two, or three dimensions, the distances in between Guinier Bragg
objects are well defined and discrete peaks appear in the pair 0.1
correlation function. In that case, the structure factor S(q) has Porod
sharp maxima reflecting the spatial arrangement of the scatter-
I(q)
0.01
ing objects. In crystallography, the positions of such maxima qRg < 1 q-4
would be referred to as lattice points of a crystal structure. 0.001
For mesoporous materials, such arrangements of pores can be
found in ordered mesoporous silica or carbon materials. MCM- 0.0001
41 and SBA-15, two well-investigated mesoporous silica mate-
0.01 0.1
rials, have pores that are hexagonally arranged in two dimen- q
sions. The pore systems can be considered as hexagonal packing
of cylindrical pores. SBA-16, another mesoporous silica mate- Figure 20 Different ranges of a scattering curve.
rial, contains interconnected spherical pores that are arranged as
3D close packing with body-centered cubic symmetry of the
applied basically to all scattering curves will be described in the
pore arrangement. The scattering curves of such materials
following section.
show sharp maxima denoted as Bragg peaks. They are caused
again by convolution of the particle form factor F2(q) with the
structure factor S(q) according to eqn [18]. 5.01.3.6.2 Approximation methods
For a 2D representation of the scattering curve, the structure Scattering curves contain different information depending on
factor can be considered as lines with S(q) 1 for positions which range it is deduced from. The shape of form factors at
where Bragg peaks are allowed and zero for any other position. low q can be approximated by Gaussian curves, the curvatures
Convolution of the particle form factor with the structure of which depend on the overall size of particles or pores in a
factor results in Bragg peaks at q positions for which the struc- given volume, that is,
ture factor is nonzero; the intensities I(q) of the peaks are
defined by the values of F2(q) at these q positions. Iq  F2 q !
This is illustrated in Figure 19 for SBA-15 with a hexagonal R2g q2
 I0 exp in the vicinity of q 0 [19]
packing of cylindrical pores. The particle form factor depends 3
strongly on the lattice parameter, a, and on the radius of the
cylindrical pores. For clarity, the reflections are shown with a In linear form, this equation can be written as
relatively narrow peak width. In reality, the peaks are broader
R2g q2
and therefore the peaks with low intensity will be hardly visible ln Iq  ln I0  [20]
in a real SAXS (or low-angle XRD) pattern. 3
Fitting of calculated peak intensities to measured ones by Thus, a plot of ln(I0) versus q2 (Guinier plot) should result in
adjusting the pore radius for a given lattice constant then a straight line and the radius of gyration, Rg, can be determined
allows calculation of pore diameters for the silica materials from the slope of that line. This method had been proposed by
under investigation from the SAXS patterns. Guinier and the range to which this approximation can be
As illustrated above, SAXS is a valuable tool for the investi- applied, that is, the q range for which qRg < 1 applies, is often
gation of porous solids with either nonordered pores or with referred to as the Guinier range (see Figure 20).
ordered pore arrangements. However, it helps the data evalu- The radius of gyration is related to the radii that are
ation if some basic pre-information on the pore geometry and observed in bodies of given shapes. It is occasionally called
arrangement is available. More general methods that can be the first moment of a scattering profile. The radius of gyration
16 Diffraction and Spectroscopy of Porous Solids

does not contain direct information on the dimensions of
particles or pores. However, once the shape of the particles or
pores is known, Rg allows an easy access to structural parame-
ters. Some exemplary relations between radii Ri and Rg are
listed in Table 3.61
The Guinier approximation is restricted to the initial part of
a scattering curve close to q 0. Another approximation that
can be made holds for the final slope of a scattering curve and
has been described by Porod. For very large q, the scattered
intensity decays with KP/q4 (Figure 20), where the Porod con-
stant KP is proportional to the surface of the material under
investigation. From the slope of the scattering curve in the
Porod range, the surface area of particles (or pores) per unit
volume of sample can be calculated from
As p q4 Iq As p KP
1  or 1 
V Qq V Qq
with being the volume fraction of the particles. Q(q) is the
mean square fluctuation of electron density in the volume and
is called the invariant. It is proportional to the total scattered
energy and calculates as
Z 1
Qq Iq q2 dq [22]
0
Thus, Q(q) represents the integral of the function I(q) q2
versus q, as shown in Figure 21.
As scattering curves are neither measured down to q 0 nor
up to q 1, the scattering curve has to be extrapolated to q 0
and 1 for the integration. The Porod approximation thus
allows calculation of surface areas of given materials without
any knowledge about the particle or the pore geometry.
It should be noted that the approximation that scattering
Table 3 Examples of radii of gyration for selected bodies61
Geometry Formula
Sphere of radius R R2g 35 R2
Spherical shell with radii R1 > R2 R2g 35 R31 R32
curves decays with q4 is only valid in the higher q range
where a low S/N ratio might be encountered (noisy data). At
lower q, the scattering intensity decays with q4 only for parti-
cles with spherical geometry. If objects with other geometries
are investigated, the slopes within the Porod range may differ
significantly, from q4 for spherical objects, via q2 for disks, to
q1 for rod-like objects.
The final range that is shown in Figure 20 reflects dimen-
sions that are in the range of interatomic distances of crystal-
line materials. The Bragg reflections that are observed in that
range are caused by diffraction on atoms from crystal struc-
tures. Occasionally, scattering in this angular range is referred
to as WAXS. However, the synonymous term XRD is more
commonly used.
As an example for SAXS investigations on a porous solid,
[21] Figure 22 shows SAXS curves of monolithic carbon aerogels
containing both micro- and mesopores.68 The contributions of
both types of pores to the total scattering curve are highlighted
by the dotted and dashed lines. Salinger et al. have investigated
the specific surface areas of the carbons by SAXS and by nitro-
gen adsorption at 77 K. Differences between the surface areas
as determined by the two methods have been assigned to
micropores that are inaccessible for nitrogen but visible with
SAXS. Activation of the carbon aerogels with CO2 at elevated
temperatures opened the inaccessible pores. For activated sam-
ples that had been thoroughly activated, specific surface areas
measured by SAXS and by nitrogen adsorption were alike.
This example shows one of the strengths of SAXS investiga-
tion on porous materials: it is sensitive not only to pores that
are accessible to the surface but also to inaccessible pores
within a solid and thus probes all pores in a given material.
Pore size distributions as calculated from SAXS data for
nanoporous methyl silsesquioxane films are shown in Figure 23
(left).69 Huang et al. used a hard sphere model to describe the
pores and found log-normal pore size distributions to fit the
measured data best. With these assumptions, pore size distribu-
tions have been calculated as shown in Figure 23 (right).
It is obvious that the pores in the film get larger with
R5 R5
increasing porosity. Simultaneously, the pore size distributions
1 2

Ellipsoid with semiaxes a, b, and c R2g a

2
b2 c2 get broader. The results are reported to be in good agreement
5
with data from field emission scanning electron microscopy
2
B2 C2
Prism with edges A, B, and C R2g A 12 studies. The interested reader will have no problems finding
a2 b2 h2
Elliptical cylinder with semiaxes a and b and height h R2g 4 12
many further studies in the literature. The above explanations
R21 R22 h2 and examples should serve as starting points and should allow
Hollow circular cylinder with radii R1 > R2 and height h R2g 2 12
the reader to find a somewhat easier start into the field.

5.01.3.6.3 Scattering on photonic crystals

As pointed out above, the properties of photonic crystals con-
sisting of opal or inverse opal structures7 are not easily acces-
sible with conventional diffraction or scattering techniques.
Q(q) = 0 I(q).q2dq Because of the large dimensions of the spherical particles of
some hundred nanometers, they scatter only in regions that are
I (q)q2

very close to the primary beam. However, recently a few reports

have been published that show that diffraction is nevertheless
possible with neutron radiation.70,71 2D diffraction patterns as
Q(q) shown in Figure 24 reveal details of structures of an opal and
an inverse opal. From the reflection positions, the lattice
parameters of these materials have been calculated to be 360
q
(polystyrene opal) and 269 nm (TiO2 inverse opal). The
Figure 21 Graphical representation of the invariant Q(q). shown measurements could be performed on a neutron
 Diffraction and Spectroscopy of Porous Solids 17

1000 C 200/45
C 1500/30

ds/dW/(cm2 g-1 srad-1)

100

Sharp bend
10

1 I(q) = Am/(1 + 1.414/3*Rm2*q2)2

~ q-4

.1 Mesoporosity Microporosity

.1 1 10
Scattering vector q (nm-1)
Figure 22 SAXS curves of carbon aerogels with micro- and mesopores. Reproduced from Salinger, R.; Reichenauer, G.; Fricke, J. Stud. Surf. Sci.
Catal. 2000, 128, 381390, with permission from Elsevier.

107 50%
40%
50%
0.24
30%
106
25% 40%
0.20
5 15%
10
Scattering intensity

10% 30%
0.16
4
10 5%
N(r)

25%
0.12
3
10
15%
0.08
2
10
10%
0.04
1
10
5%
0.00
100
10-2 10-1 10-0 0 50 100 150
q (-1 ) Pore radius ()
Figure 23 SAXS curves and pore size distributions as calculated from these curves for thin nanoporous methyl silsesquioxane films with porosities
ranging from 5% to 50%. Reproduced from Huang, E.; Toney, M. F.; Volksen, W.; Mecerreyes, D.; Brock, P.; Kim, H. C.; Hawker, C. J.; Hedrick, J. L.; Lee,
V. Y.; Magbitang, T.; Miller, R. D. Appl. Phys. Lett. 2002, 81, 22322234, with permission from American Institute of Physics.

scattering instrument with a sample to detector distance of d spacing from the peak position, that is, the wavelength at
21.5 m using a wavelength of 0.58 nm. peak maximum. The changes of the position of the Bragg peak
The existence of symmetry forbidden reflections has been have been assigned to structural rearrangement of the PS
assumed to be caused from scattering contributions from the spheres during drying of the opal film.
opal surfaces to the diffraction patterns. Transmission spectros-
copy with ultraviolet/visible-near infrared (UV/vis-NIR) on
such samples also exhibit a Bragg peak as depicted in Figure 25. 5.01.4 Spectroscopies
It shows the evolution of the spectra during drying of an opal
with time.72 As can be seen, the Bragg peak shifts from which As already discussed in the introduction, most spectroscopies
information about the formation process of the opal film can are generic and not specific for porous solids. There is ample
be deduced. From the peak position, interparticle distances can literature available on the general aspects of all spectroscopic
be deduced using the Bragg equation by calculating the techniques and therefore these techniques will not be covered
18 Diffraction and Spectroscopy of Porous Solids

(a) (b)

220
202
y

022


022
n x

202
220
(c)

Figure 24 Neutron diffraction data from (a) polystyrene opals and (b) TiO2 inverse opals obtained from polystyrene opals as hard templates. In the
setup image (c) symmetry forbidden reflections are marked with circles. Reproduced from Marlow, F.; Muldarisnur, M.; Sharifi, P.; Zabel, H. Phys. Rev.
B 2011, 84, 073401, with permission from American Physical Society.

3.2 and a fraction of the light is accordingly absorbed. The fraction

Final absorbed can be calculated by the LambertBeer law
2.8 tm = 15 h

2.4 lg I0 =I ecd [23]

2.0
with I0 and I being the intensity of the incident and transmitted
-log(T )

1.6 Intermediate Initial light, e the extinction coefficient, c the concentration, and d the
7, 22, 62 min tm = 2 min
1.2 path length of the light in the absorbing medium.
This equation allows us to correlate the attenuation of the
0.8
light with the concentration of the absorbing species, and can
0.4 thus be used for quantitative analysis. However, in most cases,
IR spectroscopy of solids is used qualitatively, as the control of
0.0
the path length is often difficult, and thus the errors in the
300 600 900 1200 1500 1800 2100
analysis can be substantial.
Wavelength (nm) The vibrations excited in a solid typically do not give highly
Figure 25 Transmission UV/vis-NIR spectra recorded on a polystyrene specific information, but are more of a fingerprint character.
opal film during its drying. Reproduced from Marlow, F.; Popa, I. Moreover, the IR spectra of a bulk and a corresponding porous
ChemPhysChem 2008, 9, 15411547, with permission from Wiley. solid are very similar, as localized vibrations do not differ
substantially, and collective motions of larger groups of
atoms that could be indicative of the pore structure are typi-
here. Instead, in the following we will highlight specific aspects cally outside the spectral range (mid IR) accessible in most
of selected spectroscopic techniques that are especially useful laboratories. In addition, the interpretation of collective
for the analysis of porous solids and provide specific informa- modes is very complicated, and thus the IR analysis of the
tion on the porosity and the surface properties of porous solid itself does not provide useful information with respect
solids. Specifically excluded from this treatment are all to the pore system of porous solids.
methods for the in situ study of catalysts under working con- This is different, though, if probe molecules are used. Probe
ditions, because this would go substantially beyond the scope molecules can enter the pore system and interact with the
of such a chapter. The reader is referred to a complete recent internal surfaces of the materials. If this interaction leads to a
issue of Chemical Society Reviews, containing many contribu- change in the vibrational spectra, this change can be used to
tions on this issue.73 analyze the surface properties in the pore space of the solid.
For this type of analysis, it is obviously necessary for the
probe molecules to enter the pore system of the porous solid.
5.01.4.1 IR Spectroscopy
Therefore, the conventional KBr pellet technique used in the IR
IR spectroscopy is a well-established, routine technique for the characterization of porous solids is unsuitable, as it can block
analysis of solid, liquid, and gaseous samples. With IR spec- the pore openings. Therefore, a self-supporting wafer of the
troscopy, molecular vibrations are excited by the incident light, porous solid under study needs to be prepared, which has to be

very thin in order to let a sufficient amount of IR radiation pass
without too strong scattering losses. However, preparation of
such thin samples is not always possible because of the
mechanical properties of the solids, or the particle sizes are in
such a range that scattering is too strong. In such cases, analysis
in a diffuse reflectance setup is a viable alternative. In this mode,
the sample is placed in a sample cup and IR light is directed to
the flat surface of the sample. The scattered light is collected
typically with ellipsoidal mirrors and directed to the detector. In
the range of absorption frequencies of the solid, the scattered
light is attenuated, as in the transmission case, and thus similar
information as in a transmission experiment can be extracted.
An absorbance-equivalent spectrum is obtained using the
so-called KubelkaMunk function. However, this is only directly
comparable with a conventional spectrum recorded in transmis-
sion mode, if the scattering coefficient is wavelength indepen-
dent, which is not guaranteed in the IR region. Thus, diffuse
reflectance spectra are typically only interpreted in a qualitative
manner.
Probe molecules are interesting in analyzing the type of
surface sites in porous solids, especially the type of acidic
sites. In addition, IR spectroscopy on single crystals can be
used for the analysis of the orientation of molecules in the
pore system of porous solids, and for the study of the diffusion
behavior of guest molecules in pore systems. These latter two
fields have recently been covered in a very useful review by
Stavitski and Weckhuysen.74
5.01.4.1.1 Analysis of surface properties
In many applications, the acidic properties of the internal
surfaces of porous solids are important. Many techniques
have been suggested for such an analysis,75 but IR spectroscopy
of adsorbed pyridine is probably the one most often used, as it
allows the discrimination between Bronsted and Lewis acidic
sites. Pyridine has characteristic bands, if protonated to a pyr-
idinium ion (around 1545 cm1) or if it interacts with Lewis
acidic sites (around 1455 cm1). There are also other charac-
teristic bands that are, however, less specific than these two
bands (Figure 26). As stated above, it is rather difficult to
precisely quantify the concentration of acid sites by IR spec-
troscopy of adsorbed pyridine, as the thickness of the self-
B+L L
L band. These shifts of the CO-absorption band can be used in
Absorbance
B B
1700 1600 1500 1400
Wavenumber (cm1)
Figure 26 IR spectrum of adsorbed pyridine on a ZSM-5 zeolite with a
Si/Al ratio of 20. Sample was activated at 550  C. The letters indicate
bands characteristic for Bronsted (B) and Lewis (L) acidic sites. For
analysis of the acidity, typically the bands at around 1545 and 1455 cm1
are used. Reproduced from Schuth, F. Habilitationsschrift, Mainz, 1995.
Diffraction and Spectroscopy of Porous Solids 19
supporting wafer cannot easily be measured reliably. More-
over, the extinction coefficients (here, typically integral coeffi-
cients over a range of wavenumbers are used) vary by about a
factor of three in different publications7679. Own data lie in
the middle of the range reported in the iterature, with
1.11 cm mmol1 (15521514 cm1) and 1.47 cm mmol1
(14731445 cm1).80 For practical purposes more important,
however, is the fact that the ratio of the integral extinction
coefficients for the Lewis and Bronsted acid sites is reported
to be around 1.3 in these publications, with the exception of
the study by Datka et al.80 where this ratio is about 1.45. The
ratio between Lewis and Bronsted acid sites, which is catalyti-
cally relevant, can thus relatively reliably be determined by IR
spectroscopy.
Another method to study the acidity of porous solids relies
on the use of typically weakly basic probe molecules which
do not lead to full proton transfer, but interact with the acidic
OH groups and perturb their absorption. Partial transfer of the
acidic proton to the basic probe molecules leads to a red-shift of
the OH-absorption band (typically found for the free OH groups
above 3600 cm1). There is a whole range of different probe
molecules reported, with more or less pronounced advantages
and disadvantages for each of them. A good survey of the anal-
ysis of acid sites using such probe molecules is given by Zecchina
and coworkers.81 Simple, two-atomic molecules are also good
probe molecules for the surface and the pore system of porous
solids. The IR spectroscopy of such molecules has recently been
comprehensively reviewed, and a good survey of the state of the
art can be found in the publication by Lamberti et al.82
Especially CO is a valuable molecule. Interaction with a
metal site leads to a change in the strength of the CO bond
of the molecule, and this change gives information on the
adsorption site. Depending on the nature of the interaction,
the frequency of the free CO-molecule (2143 cm1) can be
either blue or red shifted. A blue shift occurs, if the CO interacts
with cations, where either the electric field causes a polarization,
or the interaction of the CO lone pair with an acceptor site
causes a strengthening of the CO bond and thus a blue shift.
Generally, the stronger the blue shift, the stronger is the Lewis
acidity of the acceptor site. On the other hand, if the CO is
bound to a transition metal with available d-electrons, these
interact with an antibonding CO p*-orbital, leading to a weak-
ening of the CO bond and thus a red shift of the IR absorption
many different ways to analyze the surface properties of the
solids of porous materials, such as zeolites, gels, or MOFs, and
many application examples are discussed by Lamberti et al.83
5.01.4.1.2 Analysis of adsorbate/pore orientation
If pore systems in a porous material are oriented with respect to
the particle, the pore orientation can be analyzed with polar-
ized IR radiation, using probe molecules with a snug fit in the
pore system. Such a snug fit is possible in microporous mate-
rials, where sufficiently mobile probe molecules are available,
whose dimensions are in the same range as the dimensions of
the pore systems. If then the transition dipole moments for
certain absorption bands are known, the orientation of the
probe molecules can be analyzed. The crystal of the material
to be analyzed provides the reference system.
20 Diffraction and Spectroscopy of Porous Solids

However, as the crystals are typically small and individual 1.2

crystals need to be analyzed in order to have a unique reference
system, the analysis or such adsorbate orientation requires the 1
use of IR microspectrometers (mostly known as IR micro-
scopes). Resolution in principle is diffraction limited, and 0.8

Absorbance
thus, depending on the exact wavelength, a spatial resolution
of a few micrometers is expected. However, in practice such 0.6

good resolutions are hardly reached, as the S/N ratio decreases

0.4
substantially, if the aperture for selection of the analysis area 0
becomes too small. Analysis of adsorbate orientation until
0.2
recently was thus confined to crystalline porous materials, for 90
which crystals of several tens of micrometers size are available. 0
However, it has recently elegantly been shown that the bril- 3300 3200 3100 3000 2900 2800 2700
liance of IR synchrotron sources allows imaging of features
close to the diffraction limit,75 and thus, the previous limits 1.2
could be overcome in the future.
The first of such studies has been reported by Schuth83. The 1
general procedure for such experiments is described in this
paper, and the procedure has been used in a similar way also 0.8

Absorbance
in later publications: Single crystals of the molecular sieve (in
the case studied silicalite I) were loaded with p-xylene, for 0.6
which the IR spectrum is well known and the transition dipoles 0
0.4
are well described. IR spectra of individual, loaded crystals
were then recorded with polarized IR radiation with different
0.2 90
polarization. A number of bands with strongly differing inten-
sity, depending on the polarization of the incident light, were
0
identified (Figure 27). From the analysis of the polarization 3300 3200 3100 3000 2900 2800 2700
dependent band intensities, it could be concluded that the
Wavenumber (cm1)
p-xylene molecules align perpendicular to the long axis of the
silicalite I crystals, at least in the center of the crystal. No Figure 27 Polarized IR spectra of silicalite I crystals loaded with
polarization dependence was observed at the crystal edges, p-xylene. The angle 0 indicates polarization of the IR radiation
which can be rationalized by the silicalite I intergrowth struc- perpendicular to the long crystal axis. Crystals are oriented with the (010)
axis parallel to the IR beam. Top: spectrum of middle of the crystal and
ture, using current knowledge of this intergrowth structure.75
bottom: spectrum at the crystal edge. Reproduced from Stavitski, E.;
In the following, a number of other systems have been ana-
Weckhuysen, B. M. Chem. Soc. Rev. 2010, 39, 4615.
lyzed using polarized IR microspectrometry.8487 However, as
the technique is rather complex and suffers from the boundary
with a typical problem of microporous crystals: they are often
condition of requiring well-developed, sufficiently large crys-
cracked, and the cracks provide very fast transfer paths for
tals, it has not been used extensively.
molecules which can thus reach the internal regions of the
crystals much more quickly than expected for crystals of a
5.01.4.1.3 Analysis of diffusion processes
given size. While the initial studies have been performed with
IR spectroscopy can also be used to monitor the uptake of guest
conventional detectors, the advent of spatially resolving focal
molecules into the pore system of a porous solid. This works
plane array detectors has made it possible to push the limits of
for polycrystalline wafers or powders as well as for single
the technique. This allowed the analysis of the uptake of alkanes
crystals. For such experiments, one detects the increase in the
into MOFs, such as in the CuBTC material, for which also the
intensity of an absorption band with time after exposing
presence of surface barriers for uptake into the pore system was
the porous solid to the adsorptive. This curve is then fitted
detected.92 IR microspectroscopy is not only useful for the anal-
with an appropriate rate law, and by this procedure the diffu-
ysis of just the transport processes in porous solids, but also the
sivities can be calculated.
coupling of transport and reaction can be studied, as exempli-
Pioneering work in this field comes from the group of
fied for many cases and reviewed by the group of Weckhuysen.75
Hellmut Karge, and was carried out on polycrystalline
powders.88,89 A pressure jump of the adsorptive was induced,
and the uptake of the gas was followed. The diffusivities calcu-
5.01.4.2 NMR Spectroscopy
lated from such experiments were in reasonable agreement with
such obtained from pulsed field gradient NMR (PFG NMR) As with IR spectroscopy, the analysis of the porous solid by
experiments, which was not the case for many other methods NMR spectroscopy often does not reveal substantial differences
to study diffusion processes. Later on, this method was extended between a bulk material and a porous material, with the excep-
to IR microspectroscopy on single crystals by the groups of tion that the high surface area may lead to a strong contribu-
Lercher90 and Schuth91. The calculated diffusion coefficients tion of surface species to the spectra. However, there are several
were found to be one order of magnitude higher than diffusion NMR techniques that give information on the pore space in
coefficients reported in the literature. This could be associated porous solids, such as 129Xe NMR spectroscopy or PFG NMR

spectroscopy. These techniques will be briefly discussed in the
following. NMR tomography is a very interesting technique
that is in rapid development, but it has not reached the reso-
lution required to analyze structural porosity. Nevertheless, it
allows us to detect flow through particle beds. However, as
such structures are not porous solids in the strict sense, the
reader is referred to a recent review on this topic.93
5.01.4.2.1 Analysis of motion in porous materials
on the molecular scale
There are a number of NMR techniques suitable for the analysis
of motion on the molecular scale, such as spin-echo 2H NMR
line-shape analysis94, or spinspin or spinlattice NMR relax-
ation time measurements.95 A recent, very useful review cover-
ing different techniques with application examples is given by
Geppi et al.96
In 2H NMR line-shape analysis, static NMR spectra of
deuterium-containing guest molecules are analyzed at variable
temperature. In the absence of motions on the NMR timescale
(103106 Hz97), powder samples containing deuterium nuclei
give a typical Pake doublet, due to the different orientations of
the nuclei with respect to the field which are not averaged by
dynamical processes. However, if there is fast dynamics of the
deuterium-containing molecules, a liquid-state like spectrum
results, with a single narrow line. Intermediate situations result
in distortions of the Pake patterns, which reflect the character-
istics of the motion of the molecules. By simulating the pat-
terns, details of the dynamics of the guest species in the pore
systems of porous materials can be elucidated, including rates
of molecular motion. Typically, variable temperature experi-
ments are carried out, and the change in dynamic behavior is
recorded over a range of temperatures. This technique has
typically been used to study the dynamics of hydrocarbons in
zeolites, and a prototypical example of such a study can be
found in the study of Sato et al.97
Relaxation times are the characteristic times needed for the
spin population to return to the equilibrium distribution after
application of the radio-frequency pulse in an NMR experi-
ment. Relaxation can occur as the so-called spinlattice relax-
ation (also called longitudinal relaxation, characteristic time
T1), which probes the fast-to-intermediate dynamics of a sys-
tem of the mid-kHz to GHz scale, while the spinspin relaxa-
tion (transverse, T2) probes slower processes. In typical
experiments, the relaxation times are determined and assigned
to certain dynamical processes. Mitchell and Strange98 have
analyzed the dynamics of naphtalene in porous silica gels
with different pore sizes. By analysis of the relaxation behavior,
they concluded that in sufficiently large pores a core of naph-
talene with a structure corresponding to the solids exists, with a
disordered region close to the surface. At sufficiently small pore
sizes (4 nm), the bulk-like structure was not formed any more,
and the presence of a disordered structure was suggested.
There are also more advanced NMR techniques that allow
the elucidation of details of molecular motion in porous solids.
The group of Chmelka has published a series of papers on the
hopping dynamics or aromatic molecules in faujasite type
zeolites, studied by 2D 13C exchange NMR experiments.99,100
This method allows model-free determination of the geometry
and timescale of reorientation dynamics.
Diffraction and Spectroscopy of Porous Solids 21
5.01.4.2.2 Pulsed field gradient NMR spectroscopy
While it is normally important to have a highly homogeneous
magnetic field in NMR spectroscopy, PFG NMR spectroscopy
makes use of well-defined field inhomogeneities during an
NMR experiment. For the analysis of diffusion in porous
solids, this technique has been developed largely by the groups
of Pfeifer and Karger; the early work is reviewed in a paper from
1987,101 a recent review from the group of Karger is available
as well.102
The basis of the method is a spin-echo experiment, in which
the sample is excited first with a p/2 pulse and then with a
p-pulse. This leads to a spin-echo signal, the intensity of which
is proportional to the total magnetization. However, if short
field gradient pulses are superimposed over the constant mag-
netic field after the two radio-frequency pulses, the spin-echo
intensity is reduced, if molecules have moved away from their
original position during the time between the two field gradi-
ent pulses, as then they experience a different local magnetic
field which changes their Larmor frequency (Figure 28). This
leads to an attenuation of the spin-echo signal according to103
Z
Ct, gdg P r; t cos gdgr dr [24]
with C being the attenuation of the signal, P(r,t) the mean
propagator, which is the probability that during the time inter-
val t between the two field pulses a molecule in the sample has
been shifted over a distance r in the direction of the field
gradient, g the gyromagnetic ratio which is characteristic for
the nucleus under study, d the length of the field pulse, and gr
the field gradient. For a homogeneous system, the mean prop-
agator is given by
 2
r
P r; t 4pDt 1=2 exp [25]
4Dt
with D being the diffusion coefficient. Using this equation,
eqn [2] becomes
 
Ct, gdg exp g2 d2 g2 Dt [26]
so that the diffusion coefficient is easily obtained from a plot of
ln C over (gdg)2. Mean displacements of molecules on the
timescale of PFG NMR experiments are on the order of micro-
meters, and thus typically the intraparticle diffusion in porous
solids is measured in such experiments. However, depending
on the nature of the system under study, also interparticle
diffusion can be analyzed by PFG NMR spectroscopy. PFG
NMR is most often used for diffusion studies of single compo-
nent systems, but it is also a very powerful method to deter-
mine self-diffusion coefficients even in multicomponent
systems, as it is sensitive to a particular diffusing species.104
129
5.01.4.2.3 Xe NMR spectroscopy
Xenon is a very suitable probe for the pore space of porous
solids, due to its good NMR receptivity and inertness. Its appli-
cation for the study of porous materials has been reviewed by
Bonardet et al.105 The key property of xenon for probing the
pore space of porous solids is the large polarizability of the
atom, which is very sensitive to interactions with the environ-
ment. Since the chemical shift range of xenon in NMR
22 Diffraction and Spectroscopy of Porous Solids
)
te
or ce
B magnetic field
na
co spa
di
r(
r=0
B0
B0 + gr
d t
p/2
RF pulses
pulse
Signal Free induction
decay (FID)
0 2t
Figure 28 Principle of a PFG NMR experiment. Reproduced from Chmelik, C.; Karger, J. Chem. Soc. Rev. 2010, 39, 4864, with permission.
spectroscopy is very wide (7500 ppm), even small changes in
the environment can be easily detected.
Chemical shifts of xenon in porous solids are related to the
mean free path of xenon, and thus to the pore sizes and the
pore topologies of the materials under study, with pronounced
effects observed for microporous materials. If the xenon con-
centrations are extrapolated to zero (the chemical shift is
strongly loading dependent), the chemical shift of xenon
adsorbed into the pores increases with decreasing pore size
and more restricted diffusion. However, these correlations are
difficult to quantify, and thus the NMR spectra of adsorbed Xe
are mostly interpreted qualitatively.
In a recent application of Xe NMR spectroscopy, the transi-
tion between an open and more closed form of the MOF MIL-
53 was studied.106 During the adsorption of Xe, at low pressure
only one signal is observed in the NMR spectrum, characteristic
for the large pore form of MIL-53. Upon pressure increase, a
second, broad signal appears, which indicates a transformation
to the narrow pore form. However, at room temperature this
transformation is not completed up to xenon pressures of
1.34  105 Pa, as the intensity of the line characteristic of the
large pore form never completely vanishes.
5.01.4.3 Interference Microscopy
While interference microscopy is not a spectroscopy technique
in the strict sense of the word, it is nevertheless an interesting,
rather recent development for the study of porous solids.
The method relies on the change of refractive index upon
adsorption of a guest species into porous solids. If the refractive
index is recorded in a spatially and temporally resolving man-
ner, time-dependent relative concentration profiles (absolute
Nuclear
spin
Rotates about B
with w = g B
Time
d d 0.110 ms
t 1 ms 1 s
p
S(g = 0)
pulse
S(g 0)
Time
determination of concentrations is not possible with this tech-
nique) can be recorded during an uptake experiment.103
This technique has been largely pioneered by Kargers group,
and several interesting application studies have been published.
Figure 29 shows, for instance, the uptake of methanol in a
ferrierite crystal after a pressure step from 10 to 0 mbar. One
can clearly observe the development of the desorption profiles
over the approximately 2 h over which desorption occurs. Fit-
ting of the desorption profiles with appropriate transport equa-
tions gives numerical values for diffusivities.
A highly interesting recent study107 using interference
microscopy gave unprecedented insight into the nature of the
surface barriers, which are frequently observed for microporous
materials. Studying the uptake of different hydrocarbons by the
MOF Zn(tbip), it was discovered that the surface barrier is not a
homogeneous barrier, but that most of the pores are completely
blocked, while only a minor fraction gave unobstructed access
to the internal pore system of the material. While the nature of
the blockage is still unclear, this work provides a sound basis for
a better understanding of the factors which hinder access of
molecules to the pore system of microporous solids.
5.01.5 Conclusion
This chapter has highlighted several techniques that are partic-
ularly useful for the analysis of porous solids. In many respects,
the characterization of porous solids by diffraction and spec-
troscopy does not differ from the analysis of the nonporous
analogs. However, some techniques probe particular features
related to the porosity of the materials, and these were the
focus of this chapter. With advances in spatial and temporal

Methanol in FER: desorption 10 0 mbar
Concentration profiles (IFM)
Calculation using D(c) and a(c)
1.0
Norm. concentration, Cnorm
0.8
0.6
0.4
0.2
0.0
0 10 20 30 40
y (m)
Figure 29 Relative concentration profiles of methanol in a ferrierite
crystal over as detected by interference microscopy after a pressure jump
from 10 to 0 mbar. Dots: experimental, lines calculated. Reproduced
from Chmelik, C.; Karger, J. Chem. Soc. Rev. 2010, 39, 4864, with
permission.
resolution of these and other techniques, in the future even
more detailed insight will be possible into the characteristic
properties of porous solids. For a related chapter in this Com-
prehensive, we refer to Chapter 4.01.
References
1. Schuth, F. Chem. Ing. Tech. 2010, 82, 769.
2. www.iza-structure.org (accessed 1 March 2012).
3. Ockwig, N. W.; Delgado-Friedrichs, O.; OKeeffe, M.; Yaghi, O. M. Acc. Chem.
Res. 2005, 38, 176.
4. Kitagawa, S.; Kitaura, R.; Noro, S. Angew. Chem. Int. Ed. 2004, 43, 2334.
5. Li, H.; Eddaoudi, M.; OKeeffe, M.; Yaghi, O. M. Nature 1999, 402, 276.
6. Taguchi, A.; Schuth, F. Micropor. Mesopor. Mater. 2005, 77, 1.
7. Marlow, F.; Muldarisnur; Sharifi, P.; Brinkmann, R.; Mendive, C. Angew. Chem.
Int. Ed. 2009, 48, 6212.
8. von Laue, M.; Friedrich, W.; Knipping, P. Sitzungsber. Bayer. Akad., Math. Klasse
1912, 303322.
9. Friedrich, W.; Knipping, P.; von Laue, M. Ann. Phys.-Berlin 1913, 41, 971988.
10. von Laue, M. Ann. Phys.-Berlin 1913, 41, 9891002.
11. Bragg, W. H. Nature 1913, 90, 219.
12. Bragg, W. H. Nature 1913, 90, 360361.
13. Bragg, W. H.; Bragg, W. L. Proc. R. Soc. Lond., A-Math. Phys. 1913, 88, 428438.
14. Bragg, W. H. Proc. R. Soc. Lond., A-Math. Phys. 1913, 89, 246248.
15. Bragg, W. L. Proc. R. Soc. Lond., A-Math. Phys. 1913, 89, 248277.
16. Bragg, W. H.; Bragg, W. L. Proc. R. Soc. Lond., A-Math. Phys. 1913, 89, 277291.
17. Jenkins, R.; Snyder, R. L. Introduction to X-ray Powder Diffractometry. Wiley:
New York, 1996.
18. Stout, G. H.; Jensen, L. H. X-ray Structure Determination. Macmillan: New York,
1968.
19. Volkringer, C.; Popov, D.; Loiseau, T.; Guillou, N.; Ferey, G.; Haouad, M.;
Taulelle, F.; Mellot-Dratzinieks, C.; Burghammer, M.; Riekel, C. Nat. Mater. 2007,
6, 760764.
20. Wilson, C. C. Z. Kristallogr. 2005, 220, 385398.
21. Ferraris, G.; Jones, D. W.; Yerkess, J. Z. Kristallogr. 1972, 135, 240252.
22. Hughes, R. W.; Weller, M. T. Micro. Mesopo. Mater. 2002, 51, 189196.
23. Schenk, H. An Introduction to Direct Methods: The Most Important Phase
Relationships and Their Application in Solving the Phase Problem. International
Union of Crystallography, University College Cardiff Press: Cardiff, 1984.
24. Viterbo, D. In Fundamentals of CrystallographyIUCr Texts on Crystallography;
Giacovazzo, C., Monaco, H. L., Viterbo, D., Scordari, F., Gilli, G., Zanotti, G.,
Catti, M., Eds.; Oxford University Press: Oxford, 1992; Vol. 2.
25. Patterson, A. L. Phys. Rev. 1934, 46, 372376.
Diffraction and Spectroscopy of Porous Solids 23
26. Glusker, J. P.; Trueblood, K. N. Crystal Structure Analysis A Primer. Oxford
University Press: New York, Oxford, 1985.
27. Rietveld, H. M. Acta Cryst. 1967, 22, 151.
28. Rietveld, H. M. J. Appl. Cryst. 1969, 2, 65.
0s 29. Young, R. A. In IUCr Monographs on Crystallography; Oxford University Press:
Oxford, 1993; Vol. 5.
30 s 30. Berg, J. E.; Werner, P. E. Z. Kristallogr. 1977, 145, 310320.
31. Attfield, J. P.; Sleight, A. W.; Cheetham, A. K. Nature 1986, 322, 620622.
90 s
32. Nolang, B. I.; Tergenius, L. E. Acta Chem. Scand. A 1980, 34, 311312.
210 s 33. Harris, K. D. M.; Tremayne, M.; Lightfoot, P.; Bruce, P. G. J. Am. Chem. Soc.
1994, 116, 35433547.
400 s 34. David, W. I. F.; Shankland, K.; McCusker, L. B.; Baerlocher, C. In IUCr
600 s Monographs on Crystallography; Oxford University Press: Oxford, 2002; Vol. 13.
900 s 35. Altomare, A.; Camalli, M.; Cuocci, C.; Giacovazzo, C.; Moliterni, A.; Rizzi, R.
1350 s J. Appl. Cryst. 2009, 42, 11971202.
1950 s 36. Oszlanyi, G.; Sutto, A. Acta Crystallogr. 2004, A60, 134.
3450 s
7350 s 37. Gies, H. Science 2007, 315, 10871088.
38. Baerlocher, C.; McCusker, L.; Palatinus, L. Z. Kristallog. 2007, 222, 4753.
50 39. Walton, R. I.; OHare, D. Chem. Commun. 2000, 22832291.
40. Munn, J.; Barnes, P.; Hausermann, D.; Axon, S. A.; Klinowski, J. Phase Transit
Part B 1992, 39, 129134.
41. Walton, R. I.; Loiseau, T.; OHare, D.; Ferey, G. Chem. Mater. 1999, 11,
32013209.
42. Norby, P.; Christensen, A. N.; Hanson, J. C. Stud. Surf. Sci. Catal. 1994, 84,
179186.
43. Christensen, A. N.; Norby, P.; Hanson, J. C. Acta Chem. Scand. 1997, 51, 249258.
44. Norby, P. J. Am. Chem. Soc. 1997, 119, 52155221.
45. Norby, P.; Hanson, J. C. Catal. Today 1998, 39, 301309.
46. Norby, P. Curr. Opin. Colloid. Int. Sci. 2006, 11, 118125.
47. Linden, M.; Schunk, S. A.; Schuth, F. Angew. Chem. Int. Ed. 1998, 37, 821823.
48. Malfatti, L.; Marongiu, D.; Amenitsch, H.; Innocenzi, P. J. Phys. Chem. C 2011,
115, 1270212707.
49. Millange, F.; Serre, C.; Guillou, N.; Ferey, G.; Walton, R. I. Angew. Chem. Int. Ed.
2008, 47, 41004105.
50. Matsuda, R.; Kitaura, R.; Kitagawa, S.; Kubota, Y.; Kobayashi, T. C.; Horike, S.;
Takata, M. J. Am. Chem. Soc. 2004, 126, 1406314070.
51. Terasaki, O.; Ohsuna, T. In Handbook of Zeolite Science and Technology;
Auerbach, S. M., Carrado, K. A., Dutta, P. K., Eds.; Marcel Dekker: New York, 2003.
52. Anderson, M. W.; Ohsuna, T.; Sakamoto, Y.; Liu, Z.; Carlsson, A.; Terasaki, O.
Chem. Commun. 2004, 907916.
53. Carlsson, A.; Karlsson, G.; Bovin, J. O.; Oku, T.; Okamoto, Y.; Ohnishi, N.;
Terasaki, O. Eur. Phys. J. D 1999, 9, 623626.
54. Terasaki, O.; Ohsuna, T.; Inagaki, S. Catal. Surv. Jpn. 2000, 4, 99106.
55. Grosse-Kunstleve, R. W.; McCusker, L. B.; Baerlocher, C. J. Appl. Cryst. 1997,
30, 985995.
56. Grosse-Kunstleve, R. W.; McCusker, L. B.; Baerlocher, C. J. Appl. Cryst. 1999,
32, 536542.
57. Gramm, F.; Baerlocher, C.; McCusker, L. B.; Warrender, S. J.; Wright, P. A.;
Han, B.; Hong, S. B.; Liu, Z.; Ohsuna, T.; Terasaki, O. Nature 2006, 444, 7981.
58. Jiang, J.; Jorda, J. L.; Yu, J.; Baumes, L. A.; Mugnaioli, E.; Diaz-Cabanas, M. J.;
Kolb, U.; Corma, A. Science 2011, 333, 11311134.
59. Kolb, U.; Gorelik, T.; Kubel, C.; Otten, M. T.; Hubert, D. Ultramicroscopy 2007,
107, 507513.
60. Kolb, U.; Gorelik, T.; Otten, M. T. Ultramicroscopy 2008, 108, 763772.
61. Feigin, L. A.; Svergun, D. I. In Structure Analysis by Small-Angle X-ray and
Neutron Scattering; Taylor, G. W., Ed.; Plenum Press: New York, 1987.
62. Guinier, A.; Fournet, G. Small-Angle Scattering of X-rays. Wiley: London, 1955.
63. Glatter, O.; Kratky, O. Small Angle X-ray Scattering. Academic Press: London,
1982.
64. Brumberger, H. In NATO ASI Series, Series C: Mathematics and Physical
Sciences; Kluwer: Dordrecht, 1995; Vol. 451.
65. Porod, G. In Small Angle X-ray Scattering; Glatter, O., Kratky, O., Eds.; Academic
Press: London, 1982; pp 1751.
66. Kratky, O. In Small Angle X-ray Scattering; Glatter, O., Kratky, O., Eds.; Academic
Press: London, 1982; pp 313.
67. Orthaber, D.; Bergmann, A.; Glatter, O. J. Appl. Cryst. 2000, 33, 218225.
68. Salinger, R.; Reichenauer, G.; Fricke, J. Stud. Surf. Sci. Catal. 2000, 128,
381390.
69. Huang, E.; Toney, M. F.; Volksen, W.; Mecerreyes, D.; Brock, P.; Kim, H. C.;
Hawker, C. J.; Hedrick, J. L.; Lee, V. Y.; Magbitang, T.; Miller, R. D. Appl. Phys.
Lett. 2002, 81, 22322234.
70. Grigoriev, S. V.; Napolskii, K. S.; Grigoryeva, N. A.; Vasilieva, A. V.;
Mistonov, A. A.; Chernyshov, D. Yu.; Petukhov, A. V.; Belov, D. V.; Eliseev, A. A.;
24 Diffraction and Spectroscopy of Porous Solids

Lukashin, A. V.; Tretyakov, Yu.D.; Sinitskii, A. S.; Eckerlebe, H. Phys. Rev. B 91. Demuth, D.; Kallus, S.; Schuth, F. Stud. Surf. Sci. Catal. 1994, 84, 1223.
2009, 79, 045123. 92. Chmelik, C.; Hibbe, F.; Tzoulaki, D.; Heinke, L.; Caro, J.; Li, J.; Karger, J.
71. Marlow, F.; Muldarisnur, M.; Sharifi, P.; Zabel, H. Phys. Rev. B 2011, 84, 073401. Micropor. Mesopor. Mater. 2010, 129, 340.
72. Marlow, F.; Popa, I. ChemPhysChem 2008, 9, 15411547. 93. Gladden, L. F.; Mantle, L. D.; Sederman, A. J. Adv. Catal. 2006, 50, 1.
73. Chem. Soc. Rev. 2010, 39(12), 45435067. 94. Gladden, L. F. Chem. Eng. Sci. 1994, 49, 3339.
74. Stavitski, E.; Weckhuysen, B. M. Chem. Soc. Rev. 2010, 39, 4615. 95. Tycko, R. Nuclear Magnetic Resonance Probes of Molecular Dynamics. Kluwer:
75. Karge, H. G. In Handbook of Heterogeneous Catalysis; Ertl, G., Knozinger, H., Boston, MA, 1994.
Schuth, F., Weitkamp, J., Ed.; Wiley-VCH: Weinheim, 2008; Vol. 2, Chapter 3.2.4, 96. Geppi, M.; Borsacchi, S.; Mollica, G.; Veracini, C. A. Appl. Spectr. Rev. 2009,
p 1097. 44, 1.
76. Hughes, T. R.; White, H. M. J. Phys. Chem. 1967, 71, 2192. 97. Sato, T.; Kunomori, K.; Hayashi, S. Phys. Chem. Chem. Phys. 1999, 1, 3839.
77. Stock, T.; Dombrowski, D.; Hoffmann, J.; Furwert, J. Z. Phys. Chem. Leipzig 98. Mitchell, J.; Strange, J. H. Mol. Phys. 2004, 102, 1995.
1984, 265, 551. 99. Schaefer, D. J.; Favre, D. E.; Wilhelm, M.; Weigel, S. J.; Chmelka, B. F. J. Am.
78. Emeis, C. A. J. Catal. 1993, 141, 347. Chem. Soc. 1997, 119, 9252.
79. Datka, J.; Turek, A. M.; Jehng, J. M.; Wachs, I. E. J. Catal. 1992, 135, 186. 100. Geil, B.; Isfort, O.; Boddenberg, B.; Favre, D. E.; Chmelka, B. F.; Fujara, F.
80. Schuth, F. Habilitationsschrift, Mainz, 1995. J. Chem. Phys. 2002, 116, 2184.
81. Zecchina, A.; Spoto, G.; Bordiga, S. Phys. Chem. Chem. Phys. 2005, 7, 1627. 101. Karger, J.; Pfeifer, H. Zeolites 1987, 7, 90.
82. Lamberti, C.; Zecchina, A.; Groppo, E.; Bordiga, S. Chem. Soc. Rev. 2010, 39, 4951. 102. Chmelik, C.; Karger, J. Chem. Soc. Rev. 2010, 39, 4864.
83. Schuth, F. J. Phys. Chem. 1992, 96, 7493. 103. Karger, J. In Handbook of Heterogeneous Catalysis; Ertl, G., Knozinger, H.,
84. Schuth, F.; Demuth, D.; Zibrowius, B.; Kornatowski, J.; Finger, G. J. Am. Chem. Schuth, F., Weitkamp, J., Ed.; Wiley-VCH: Weinheim, 2008; Vol. 3, Chapter 6.2,
Soc. 1994, 116, 1090. p 1721.
85. Muller, G.; Budis, J.; Eder-Mirth, G.; Kornatowski, J.; Lercher, J. A. J. Mol. Struct. 104. Hong, U.; Karger, J.; Pfeifer, H. J. Am. Chem. Soc. 1991, 113, 4812.
1997, 410411, 173. 105. Bonardet, J. L.; Fraissard, J.; Gedeon, A.; Springuel-Huet, A. M. Catal. Rev.-Sci.
86. Jacobs, W. P. J. H.; Demuth, D. G.; Schunk, S. A.; Schuth, F. Micropor. Mater. Eng. 1999, 41, 115.
1997, 10, 95. 106. Springuel-Huet, A. M.; Nossov, A.; Adem, Z.; Guenneau, F.;
87. Popescu, S. C.; Thomson, S.; Howe, R. F. Phys. Chem. Chem. Phys. 2001, 3, 111. Volkringer, C.; Loiseau, T.; Ferey, G.; Gedeon, A. J. Am. Chem. Soc.
88. Karge, H. G.; Klose, K. Ber. Bunsenges. Physikal. Chem. 1979, 79, 454. 2010, 132, 11599.
89. Karge, H. G.; Niessen, W. Catal. Today 1991, 8, 451. 107. Hibbe, F.; Chmelik, F.; Heinke, L.; Prahmanik, S.; Li, J.; Ruthven, D. M.;
90. Muller, G.; Narbeshuber, T.; Mirth, G.; Lercher, J. A. J. Phys. Chem. 1994, Tzoulaki, D.; Karger, J. J. Am. Chem. Soc. 2011, 133, 2804.
98, 7436. 108. Oster, G.; Riley, D. P. Acta Cryst. 1952, 5, 272276.
 7.20 FischerTropsch Synthesis: Catalysts and Chemistry
J van de Loosdrecht, Sasol Technology Pty (Ltd), Sasolburg, South Africa; Eindhoven University of Technology, Eindhoven,
The Netherlands
FG Botes, Sasol Technology Pty (Ltd), Sasolburg, South Africa
IM Ciobica, Eindhoven University of Technology, Eindhoven, The Netherlands
A Ferreira, P Gibson, DJ Moodley, AM Saib, and JL Visagie, Sasol Technology Pty (Ltd), Sasolburg, South Africa
CJ Weststrate and JW (Hans) Niemantsverdriet, Eindhoven University of Technology, Eindhoven, The Netherlands
2013 Elsevier Ltd. All rights reserved.

7.20.1 Introduction: Processes, Catalysts, and Recent History 526

7.20.1.1 Anything-to-liquids Technology: Syngas Production, FTS, and Product Workup 526
7.20.1.2 FTS, the Product Distribution 526
7.20.1.3 FischerTropsch Catalysts and Modes of Operation 527
7.20.1.4 Historical Development of the FTS 529
7.20.2 Iron-Based FTS Catalysts 531
7.20.2.1 Introduction 531
7.20.2.2 Commercial Applications 531
7.20.2.3 Iron FischerTropsch Catalyst Preparation 531
7.20.2.3.1 Fusion 532
7.20.2.3.2 Precipitation 533
7.20.2.3.3 Improving iron FischerTropsch catalyst precursors by promoters 534
7.20.2.3.4 Activation and reduction procedures 534
7.20.2.4 Selectivity Manipulation of Iron Catalysts 535
7.20.2.5 Catalyst Stability During FTS 535
7.20.2.6 Spent Catalyst Management 536
7.20.3 Cobalt-Based FTS Catalysts 537
7.20.3.1 Introduction 537
7.20.3.2 Composition of Cobalt Catalysts 537
7.20.3.3 Preparation of Cobalt FischerTropsch Catalysts 539
7.20.3.3.1 Precipitation 539
7.20.3.3.2 Preparation methods involving pre-shaped supports 540
7.20.3.3.3 Calcination 541
7.20.3.3.4 Reduction 541
7.20.3.4 Cobalt Catalyst FischerTropsch Performance 541
7.20.3.5 Deactivation and Regeneration of Cobalt FischerTropsch Catalysts 544
7.20.4 Mechanisms and Kinetics of FTS Over Iron and Cobalt Catalysts 546
7.20.4.1 Introduction 546
7.20.4.2 Surface Science Studies and Model Reactions 547
7.20.4.2.1 Adsorption of CO and hydrogen on model surfaces 547
7.20.4.2.2 CO bond scission 548
7.20.4.2.3 Hydrogenation and the stability of C1Hx species 549
7.20.4.3 DFT Modeling 549
7.20.4.3.1 CO Dissociation 549
7.20.4.3.2 C H reactions 550
7.20.4.3.3 Chain growth 551
7.20.4.4 Macrokinetic Observations and Models 551
7.20.4.4.1 General observations regarding kinetics 551
7.20.4.4.2 Simple macrokinetic models 552
7.20.4.4.3 Selectivity modeling 552
7.20.4.5 Mechanistic and Kinetic Implications 553
7.20.5 Conclusion 554
References 554

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00729-4 525

526 FischerTropsch Synthesis: Catalysts and Chemistry
7.20.1 Introduction: Processes, Catalysts,
and Recent History
The FischerTropsch synthesis (FTS) represents technology
from the 1920s1,2 that has continuously been revived to pro-
vide synthetic hydrocarbon fuels and chemicals from initially
coal, later natural gas, and nowadays also biomass. Virtually
any source of (hydro)carbon feedstock can be converted to a
mixture of synthesis gas, or syngas (CO and H2), which is in
fact a key intermediate on which theoretically the entire chem-
ical industry could be based. FTS stands for the reaction(s) of
synthesis gas to predominantly straight-chain hydrocarbons,
which can be paraffins from CH4 to waxes (CnH2n2 with n
from 1 to over 100), olefins from ethylene to much longer
molecules (CnH2n, with n  2), and to a lesser extent oxygen-
ated products such as alcohols. It produces as main by-
products water and/or carbon dioxide, that is, due to the
water-gas shift (WGS) reaction. Being a highly exothermic
reaction, it generates large amounts of heat. The process is
represented by the simplified reaction equations
FTS : CO 2H2 ! CH2  H2 O  165 kJ mol1 [1]
WGS : CO H2 OH2 CO2  42 kJ mol1 [2]
Reaction [1] represents in essence a polymerization, imply-
ing that the product will be a mixture of hydrocarbons with
a distribution in molecular weights. Selectivity and control
thereof are therefore of key importance in FTS technology.
FischerTropsch technology represents a subject of inten-
sive research both in industry and in academia. Many excellent
reviews are available.39
In this chapter, we first describe the general aspects of the
technology in which the FTS features, then the more chemical
aspects of the process in relation to the iron and cobalt catalysts
that are used in practical applications, and finally mechanistic
insight, on the basis of kinetics, surface science, and computa-
tional modeling.
7.20.1.1 Anything-to-liquids Technology: Syngas
Production, FTS, and Product Workup
The overall process from original carbon source for the syngas
to the FTS product is named after the feedstock employed,
hence the terminology coal-to-liquids (CTL), gas-to-liquids
(GTL) and biomass-to-liquids (BTL), collectively known as
XTL (anything-to-liquids).
In all instances, the carbon source is first converted to
synthesis gas (or syngas for short), which is a mixture of CO
and H2. Solid feedstocks such as coal or biomass are gasified,
usually noncatalytically, by partial oxidation with oxygen (sup-
plying the heat for the endothermic gasification reactions) and
reaction with steam (which acts as a gasification agent, hydro-
gen source, and coolant).
When the starting material is natural gas, it can also be
adiabatically reformed in the presence of oxygen and steam.
There are different embodiments of this approach, such as
autothermal reforming (ATR), noncatalytic partial oxidation
(POX), and catalytic partial oxidation (CPOX), but in essence
the chemistry of all is the same and very similar to that of coal
gasification. Alternatively, heat can be supplied externally, in
which case the gas is reformed only with steam and/or CO2,
but no oxygen is added. Examples of this approach include
steam reforming (where mainly steam is added), dry reforming
(where mainly CO2 is added), and heat exchange reforming
(where process heat is supplied to the reformer tubes). Typical
reforming catalysts are based on nickel as the active metal.10
The second step in the XTL process is to catalytically convert
the syngas to a range of hydrocarbons via the FT synthesis,
which mainly yields linear alkanes and 1-alkenes, and which
will be the main subject of this chapter hereafter.
The third and last step is usually the workup of the
hydrocarbons to final products, which are typically fuels, but
optionally also chemicals. A popular application at present is
to target the production of long chain waxes in the FT synthe-
sis, followed by hydrocracking to middle distillate range
components, such as diesel (C9C22) and jet fuel (C9C15).
Hydrocracking catalysts are bifunctional in nature, with either
a noble metal (e.g., Pt) or sulfided base metals (e.g., Ni/W or
Co/Mo) as the hydrogenation function on a catalytically active
acidic support, such as a silicaalumina. We refer to the liter-
ature for further information on this subject.11,12
7.20.1.2 FTS, the Product Distribution
At the chemistry level, the FT synthesis is both a CO hydroge-
nation reaction and a polymerization reaction. The former is
reflected by the fact that the CO bond must be broken and
new CH bonds formed. Additionally, CC bonds must be
formed in order to effect hydrocarbon chain growth. Since the
product carbon number distribution approximately follows a
statistical function called the AndersonSchulzFlory relation-
ship, it is widely accepted that chain growth occurs one carbon
atom at a time via a polymerization mechanism. Proposals for
the monomer of chain growth, which is produced in situ, have
included adsorbed CO, an enol species and a CHx species,7,1315
and will be discussed further in the section on mechanism and
kinetics.
The competition between chain growth (yielding a surface
intermediate with one higher carbon number) and chain ter-
mination (yielding a desorbed final product) is determined by
the probability for growth, called the a-value. A higher a-value
will result in longer hydrocarbons and thus a heavier product
spectrum (Figure 1). If a is independent of carbon number, the
scheme presented in Figure 2 applies and the total amount of
carbon contained in products with n carbon atoms (namely
Cn) can be formulated on a relative basis:
C1 11  a
C2 21  aa
C3 31  aa2
Cn n1  aan1
The total amount of carbon in the product spectrum then
forms a convergent infinite sum with an analytical solution:
X
1 X
1
1
Cn n1  aan1
1 1
1a
 FischerTropsch Synthesis: Catalysts and Chemistry 527

HTFT LTFT
100
90
CH4 C5+

Carbon atom selectivity (%)

80
70 C20+
waxes
60 C2C4
C5C11
50 gasoline

40 C9C22
30 diesel-
distillates
20
10
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Chain-growth probability, a

Figure 1 Hydrocarbon product spectrum that is produced during FischerTropsch synthesis for varying chain growth probability a. High-temperature
FischerTropsch technology (HTFT) corresponds approximately to 0.70 < a < 0.75, and low-temperature FischerTropsch (LTFT) to about
0.85 < a < 0.95.

Products are primary FT products and which are formed subsequently by

C1 C2 C3 secondary reactions. For example, olefins and alcohols can
1-a 1-a 1-a undergo a range of secondary reactions, such as hydrogenation,
a a a double bond isomerization, skeletal isomerization, and con-
C* C1* C 2* C 3* ..
version to heavier compounds.4,16
Intermediates

Figure 2 Carbon chain growth and termination scheme for the

derivation of the AndersonSchulzFlory equation, with a the chain
growth probability factor, Cn (n 1, 2, 3, . . .) the final products with n
carbon atoms, and Cn* the intermediates with n carbon atoms.
This means that the selectivity toward products with n
carbon atoms on a carbon atom basis, namely Sn, can be
expressed as follows:
Cn
Sn X1 n1  a2 an1
1
Cn
After converting this equation to the logarithmic domain
and rearranging, it is found that
 
Sn 1  a2
ln n ln a ln [3]
n a
As a result, a plot of ln(Sn/n) versus carbon number (n) gives
rise to a straight line with a slope equal to ln(a). However,
deviations in the actual FT product spectrum from the ideal
AndersonSchulzFlory distribution are usually observed.4,14,16,17
These include a higher methane and a lower C2 selectivity than
predicted by the equation. There is also an increase in the chain
growth probability factor and concomitant decrease in the
olefin/paraffin ratio with hydrocarbon chain length. In addi-
tion to linear alkanes and 1-alkenes, a variety of other products
are also formed, including branched aliphatic compounds,
alcohols, aldehydes, ketones, acids, and (at sufficiently high
operating temperatures) even aromatics. This alludes to the
complexity of the reaction and many unresolved issues remain
regarding the reaction mechanism. A further complicating
factor is that it is not always clear which of these compounds
7.20.1.3 FischerTropsch Catalysts and Modes of Operation
Metals known to catalyze the FT reaction mainly include iron,
cobalt, ruthenium, and nickel.6 Ruthenium is a scarce and
expensive metal, whereas nickel only forms methane at reac-
tion temperatures sufficiently high to suppress nickel carbonyl
formation (note that methanation is the reverse reaction of
methane reforming, for which nickel-based catalysts are com-
monly used10). As a result, only iron- and cobalt-based FT
catalysts have found commercial application.1820 Iron gener-
ally produces more olefins and oxygenates than cobalt (i.e., a
less hydrogenated product spectrum), which may be related to
the lower hydrogenating ability of iron. While cobalt is active
in the metallic state,19 iron catalysts change under Fischer
Tropsch conditions to a complex mixture of iron carbides
and oxides.20,21
Byproducts of the FTS originate from the way oxygen from
CO is removed. With cobalt catalysts, essentially all oxygen
from CO dissociation (typically around 99%) is discarded as
water. Iron catalysts differ in this respect, as a significant por-
tion of the oxygen is also discarded as CO2. The latter is often
visualized as a separate, consecutive reaction, namely the WGS.
Stoichiometrically, the overall process can be represented by
reactions [1] and [2], which we repeat here:
FTS : CO 2H2 ! CH2  H2 O  165 kJ mol1 [1]
WGS : CO H2 OH2 CO2  42 kJ mol1 [2]
The net rate of hydrogen conversion divided by CO conver-
sion (sometimes referred to as the usage ratio) is extremely
important for the gas loop design around an FT reactor. In the
one extreme, where virtually no WGS takes place, the usage
528 FischerTropsch Synthesis: Catalysts and Chemistry

ratio is only determined by the FT reaction (with a high selec-
tivity to long hydrocarbons and a low selectivity to methane)
and assumes a value of around 2. In the other extreme, where
almost all water is shifted to CO2, the usage ratio can approach
a value of 0.5. The low propensity of cobalt catalysts for the
WGS makes them the preferred catalysts for GTL application,
since the H2/CO ratio of syngas derived from natural gas is
already close to or above the usage ratio. Any additional WGS
will result in an excess of hydrogen that would not be fully
consumed by the FT reaction, even if CO is converted to
extinction. Conversely, the WGS is more facile over iron
catalysts, yielding a much lower overall usage ratio that is better
suited to syngas feeds with a low H2/CO ratio, such as that
generally obtained from coal gasifiers.
Table 1 presents the current commercial application of the
FTS. There are two important aspects to note from this table.
First, the worldwide FTS capacity is expected to reach a total of
just over 400 000 barrels per day by 2013 (1 barrel 159 l),
which is very small compared to the total crude oil production
of around 8085 million barrels per day. Second, the FTS has
been applied in a variety of forms, which determines the type
of reactor employed. As indicated in Figure 3, these reactors

Table 1 FischerTropsch synthesis, current commercial plants and plants under construction

Company Location Carbon feedstock Catalyst type Reactor type Start-up date Approximate plant
capacity (barrels
per day)

Sasol Sasolburg, South Initially coal, currently Fused Fe/K HTFT circulating 1955 to
Africa natural gas fluidized bed 1985
Precipitated Fe/K LTFT multitubular 1955
fixed bed 5000
Precipitated Fe/K LTFT slurry phase 1993
(spray dried)
Sasol Secunda, South Mostly coal, now Fused Fe/K HTFT circulating 19801999 160 000
Africa supplemented by fluidized bed
natural gas HTFT SAS reactora 1995
Shell Bintulu, Malaysia Natural gas Co/SiO2 LTFT multitubular 1992 14 500
Co/TiO2 fixed bed
PetroSA Mosselbay, South Natural gas Fused Fe/K HTFT circulating 1993 22 000
Africa fluidized bed
(Sasol technology)
Sasol-QP (Oryx) Ras Laffan, Qatar Natural gas Co/Al2O3 LTFT slurry phase 2007 34 000
Shell (Pearl) Ras Laffan, Qatar Natural gas Co/TiO2 LTFT multitubular 2011 140 000
fixed bed
Chevron-Sasol Escravos, Nigeria Natural gas Co/Al2O3 LTFT slurry phase 2013 34 000
a
SAS: Sasol Advanced Synthol, fixed fluidized bed.

Moving bed Stationary bed reactors

1200 m particles 2005000 m particles

Multitubular
LTFT, 200250 C
fixed bed
3-phase system:
6000 barrels
gasliquidsolid
per day
a = 0.850.95
Products: Slurry bubble
wax, diesel, naphta column, Microchannel reactor
Catalysts: filled with wax ~ 2001000 b/d (for assembly)
supported cobalt or 24 000 barrels Microchannel process
technology module
precipitated iron per day
Boiling heat transfer

HTFT, 320350 C High heat flux

10 times higher heat flux
FT

than conventional reactors

2-phase:gassolid
a = 0.700.75
Products:
petrol and chemicals
Catalysts:
fused iron, K-promoted
Circulating
fluid bed Fixed fluid bed
7000 barrels per day 20 000 barrels per day

Figure 3 Overview of FischerTropsch technology with reactors (figure microchannel reactor: courtesy of the Oxford Catalysts Group).

can broadly be classified in two classes, namely two-phase or
three-phase reactors, and moving or stationary catalyst bed
reactors.2224
The high-temperature FischerTropsch (HTFT) synthesis
process is characterized by operating temperatures of about
320350  C and the products are essentially only in the gas
phase under reaction conditions, giving rise to a gassolid
system without any bulk liquid phase. Originally, this process
was operated in circulating fluidized bed reactors and more
recently in fixed fluidized bed reactors. Cobalt catalysts would
essentially only produce methane at these temperatures, mak-
ing alkali-promoted iron catalysts the only option for this
application. Due to the mechanical demand that these moving
bed reactors place on the catalyst, particle strength is an impor-
tant consideration; consequently, only fused bulk iron catalysts
have been employed commercially. The light product spectrum
is best suited to the production of gasoline, but the high
selectivity toward linear 1-olefins and (to a lesser extent) oxy-
genates allows for the extraction of chemicals from the product
slate. These include monomers such as ethylene and propyl-
ene, co-monomers such as 1-hexene and 1-octene, and sol-
vents (e.g., propanol, butanol, methyl ethyl ketone (MEK),
and acetaldehyde).25
The low-temperature FischerTropsch (LTFT) synthesis is
operated between 200 and 250  C.26,27 Both cobalt and iron
catalysts are suitable for this application, although cobalt cat-
alysts would typically be used toward the lower half of the
quoted temperature range. The heavy product spectrum
extends well into the domain of waxes, which are liquid
under reaction conditions. The presence of a bulk liquid phase
gives rise to a three-phase gasliquidsolid system. Originally,
only fixed-bed reactors operating in a trickle bed mode were
employed for this synthesis. In order to limit the pressure
drop over the stationary catalyst bed, catalyst particle sizes
must be in the millimeter range, which brings about signifi-
cant intra-particle diffusion limitations. This not only limits
catalyst utilization, but also adversely affects product selectiv-
ities due to the differences in diffusion rates between hydro-
gen and CO that causes higher H2/CO ratios toward the
center of the particles. The highly exothermic nature of
the FT reaction causes axial and radial temperature profiles
in the catalyst bed.
More recently, slurry bubble-column reactors have been
developed to overcome some of these drawbacks. Syngas is
bubbled through a suspension of fine catalyst particles in the
liquid product phase. The catalyst particle sizes are usually less
than about 100 mm, which is sufficiently small to prevent intra-
particle diffusion limitations, while the well-mixed liquid phase
ensures virtual isothermal operation of the reactor. There are,
however, certain technical challenges associated with FTS slurry
reactors. A prerequisite of a slurry process is the development of
an efficient solidliquid separation step to remove product wax
from the reactor. It is extremely important to ensure the
mechanical integrity of the catalyst to limit the extent of break-
up and attrition in the moving bed environment.
Of late there have been some new reactor developments for
FTS application, but none of these have been commercially
applied yet. Microchannel reactors can support very high heat
and mass transfer rates and thereby address the problems of
traditional fixed-bed reactors, while the stationary bed circum-
vents the challenges of slurry reactors. This approach shows
FischerTropsch Synthesis: Catalysts and Chemistry 529
promise, especially with respect to the small-scale application
of a few hundred or a few thousand barrels per day production
capacity, and some relatively new commercial companies are
actively pursuing this technology.28
Structured reactors (monolith type reactors) for FT applica-
tion have also attracted the attention of mainly academia,
although there has been some limited interest from commer-
cial companies as well.29 In this approach, the active FT metal
(e.g., cobalt) is coated onto a large structure with a specific
geometry, which is then inserted into a reactor tube.
The LTFT synthesis is ideally suited for the production of
high-quality middle distillates (diesel and jet fuel) after hydro-
cracking of the long chain waxes. In addition, the heavy prod-
uct spectrum provides chemical opportunities in the form of
speciality waxes and base oils. The naphtha from the process is
also a high-quality feedstock for naphtha steam crackers that
produce mainly ethylene, but also some propylene.
7.20.1.4 Historical Development of the FTS
Historically, the first syngas conversion results were pub-
lished by Sabatier30 in 1902 where it was shown that a mix-
ture of carbon monoxide and hydrogen could be converted
into methane over nickel and cobalt catalysts. In the 1920s,
Franz Fischer and Hans Tropsch took this process a step
further and showed that syngas could be converted into a
mixture of higher hydrocarbons that could be used as petrol
or diesel (i.e., FTS).1,2 In their first patent,31 they described
the production of higher hydrocarbons using iron- and
cobalt-based catalysts operated at atmospheric pressure and
at temperatures below 300  C. Further research in Germany
led to improved versions of this process. The first commercial
plant started in 1936. Several others followed and provided
Germany and Japan with synthetic fuel during the Second
World War. These plants used mainly cobalt catalysts
supported on kieselguhr (i.e., silica-based supports) and pro-
moted by magnesia and thoria, in fixed-bed reactors. Further,
China had FTS plants in the 1940 through 1960s, all based on
cobalt catalysts.32
After the war, the German FT technology came in the hands
of the Allied Forces. Many scientists and engineers who con-
tributed to the German developments were interrogated and
the entire FischerTropsch technology was extensively investi-
gated at the US Bureau of Mines, which resulted in new two-
phase HTFT technology. The classical textbook by Storch,
Golumbic, and Anderson originates from this period.3 Small
plants were built in the US and operated in the 1950s.
Large-scale FTS developments mainly occurred in South
Africa.33 Sasol started an FTS plant in 1955 based on HTFT to
make petrol and on LTFT to produce wax. Both HTFT and LTFT
used iron-based catalysts. The HTFT technology formed the
basis for the large expansion of Sasol in the late 1970s/early
1980s when Sasol 2 and 3 were built in Secunda (see Table 1).
The main reasons for this expansion were the oil crises in the
1970s, which led to a significant increase in the crude oil price
(see Figure 4).
These oil crises also initiated renewed interest in FTS from
other companies like BP, ExxonMobil, Gulf, Shell, and Statoil,
which was mainly based on cobalt FTS catalysts.34,35 In the last
20 years, this has led to new commercial GTL plants by PetroSA
(South Africa; 1993), Shell (Malaysia; 1992), Sasol-Qatar
530 FischerTropsch Synthesis: Catalysts and Chemistry

450 100

400 Patents 90

350 80

Number of articles/patents
Articles

US crude oil price ($)

70
300 US crude oil price
60
250
50
200
40
150 Articles Oil price 30
100 20
50 Patents 10

0 0
1970 1975 1980 1985 1990 1995 2000 2005 2010
Publication year
Figure 4 Patents and articles per year compared with the crude oil price (figure inspired by de Smit and Weckhuysen21).

FTS

ATR
ASU

Figure 5 Photo of the Sasol-QP Oryx GTL plant in Qatar, showing the air separation units (ASUs), the auto-thermal reformers (ATRs), and the
FischerTropsch synthesis (FTS) slurry reactors. The product work-up section located behind the FTS reactors is not visible (photo courtesy of Sasol).

Petroleum (Qatar; 2007; see Figure 5), Shell (Qatar; 2011),
and Sasol Chevron (Nigeria; under construction start-up
2013). An overview of the current commercial operations
using FTS technology is shown in Table 1.
The investment decision to build the Sasol-Qatar Petroleum
Oryx-GTL plant was taken in 2003 when the oil price was $25/
barrel. The facility was built at a cost of $1 billion. Currently, a
yearly profit is generated of about $500 million.36 Sasols
much larger Secunda CTL facility is generating currently
about $2 billion profit annually.36 Shells Pearl plant (both
the FTS production 140 000 barrels per day and the
upstream natural gas condensates 120 000 barrels per day
together) was built at a cost of $20 billion,37 and Shell
announced to make annually $4 billion cash when Pearl is at
full production with the oil price at $70/barrel. It is clear that
new GTL/CTL facilities require large capital investments, and
are heavily dependent on the prevailing crude oil price. How-
ever, over the long term these large-scale GTL/CTL facilities do
make economic sense.
A new challenge to the GTL/CTL technology is global warm-
ing and the emission of CO2. Fuel products from GTL facilities
have a similar environmental footprint compared to crude
oil-derived fuels. However, products from CTL facilities have
a much larger CO2 impact, which is immediately clear from the
overall stoichiometric equations [1] and [2]. In the hypothet-
ical limit of using a carbon feedstock which does not contain
any hydrogen, one CO2 molecule is formed for every carbon
atom that ends up in a hydrocarbon. A large portion of the
CO2 produced in CTL plants is removed and concentrated, and
is therefore ideally suited for capturing, that is, capture ready.
At the same time new focus on products from biomass can
stimulate interest in FTS further, as biomass can be used as a
carbon source for syngas generation. Recently, Oxford Cata-
lysts has demonstrated their FTS technology using syngas made
from wood chips.28
Other opportunities for GTL applications in the future
might be the use of associated natural gas in small-scale plants
(<1000 barrels per day, as Oxford Catalysts and CompactGTL
are pursuing), as well as the use of shale gas in large-scale
facilities (as pursued by Sasol).
From an academic point of view, renewed interest in FTS
was clearly observed in two main waves (see Figure 4). The first
wave occurred in the late 1970s, while the second one started
around 1995 and is still gaining momentum. The latter
 FischerTropsch Synthesis: Catalysts and Chemistry 531

observation is also matched by an increase in patenting activ- 7.20.2.2 Commercial Applications

ity. Of course, this revived activity is related to the increase in
Sasol is a leader in the field when it comes to commercializing
the crude oil price, although the present academic interest is
iron-catalyzed FTS processes. In the early 1950s, Sasol com-
certainly also inspired by the notion that crude oil resources are
mercialized Fe-catalyzed FTS based on the Ruhrchemie process
limited. Typical topics in FischerTropsch research with a high
to produce a variety of synthetic petroleum products using the
academic interest are catalyst preparation methods, deactiva-
Arge Tubular Fixed bed reactors (see Table 1 and Figure 3).
tion studies, and mechanistic and kinetic studies, in which
Later on, they developed the slurry-bed reactor and this reactor
sophisticated tools such as in situ catalyst characterization,
together with the Arge reactors are used to produce high-
surface science, molecular modeling, and transient kinetic
molecular-weight hydrocarbons for the wax industry.38 In the
studies are the common ingredients. The increased interest
late 1950s, Sasol also commercialized a circulating fluidized
from both the academic as well as the commercial world has
bed reactor at their Sasolburg facilities in which fused iron
created excellent scientific interactions and discussions, which
catalyst is fluidized at high temperatures to produce lighter
enabled further progress on this exciting topic of FTS. Many
hydrocarbons ideally suited for producing fuel and chemical
questions are still outstanding, such as the state of the catalyt-
feedstock. In the late 1970s/early 1980s Sasols Secunda plant
ically active surface under reaction conditions, and the reaction
was built using this circulating fluidized bed technology, which
mechanism in terms of elementary steps.
was replaced in the late 1990s by the improved fixed fluidized
bed technology. Subsequent to Sasols successful commercial-
ization of iron-catalyzed FTS, South Africas national oil
7.20.2 Iron-Based FTS Catalysts
company (PetroSA) commercialized a GTL facility using
7.20.2.1 Introduction previous-generation high-temperature FTS technology (Sasol
licensed technology). This technology is based on a fused
The iron-catalyzed FTS process is, along with ammonia synthe-
iron catalyst operated in a fluidized bed reactor at high tem-
sis, one of the most studied systems in the field of heteroge-
perature (330350  C). In 2010, it was still recognized as one
neous catalysis. The reason for this is possibly the fact that the
of the worlds largest GTL refineries, producing about 22 000
application of the process is so versatile. Not only can iron FTS
barrels per day of high-quality FTS-derived fuels.
produce a light hydrocarbon product stream ideal for the fuel
Rentech, based in Colorado, USA has long been investing in
and chemical industry, it can also produce heavier hydrocar-
iron-based FTS research. Rentech demonstrated their iron-
bons (C35) suited for the waxes market. Iron is also a cheap
based FTS technology in their Product Demonstration Unit
raw material when compared to its cobalt counterpart (cobalt
(PDU) in the middle of 2008. The PDU produces approxi-
is on average 250 times more expensive than iron raw mate-
mately ten barrels per day of ultra-clean diesel, aviation fuels,
rials) and it has been commercially applied since the late 1950s
and naphtha.39
by Sasol38 (Table 1). Iron is believed to be more tolerant of
Synfuels China has recently emerged as an important player
poisons, for example, sulfur in synthesis gas than cobalt. It is
in the FischerTropsch industry.32 Their development of a so-
also known to be responsive to selectivity manipulation by the
called high-temperature slurry-phase technology (HTSFTP)
addition of promoters and a variation of typical process param-
and associated iron-based catalyst is novel to the industry. The
eters, for example, temperature, pressure, and H2/CO ratio. The
integrated technology promises improvements in process ther-
disadvantage, however, is the fact that iron FTS catalysts deac-
mal efficiency and a highly active catalyst. This technology has
tivate rather quickly (activity or selectivity loss) and this will be
been demonstrated in a 4000 barrels per day semi-commercial
discussed in more detail later in this section. As already men-
CTL facility owned by the YiTai Coal Liquefaction Company in
tioned the iron FTS process can be manipulated to produce a
Xue JiaWan, Erdos, Inner Mongolia.
range of carbon number distributions with the final product
stream depending mainly on the temperature applied during
FTS. At lower temperatures, for example, 220250  C the chain
7.20.2.3 Iron FischerTropsch Catalyst Preparation
growth probability (a) of the catalyst is approximately 0.94
indicating that the bulk of the products will consist of hydro- There are several preparation methods available in the litera-
carbons longer than C21. In the case of higher temperatures, for ture for the synthesis of Fe FTS catalysts, like precipitation and
example, 320350  C, the chain growth probability decreases fusion. Iron catalysts prepared commercially are actually iron
to 0.7 and even lower with the main products being light oxides, hydroxides, or oxy-hydroxides, which undergo an acti-
hydrocarbons utilized for the production of transportation vation step such as reduction or pre-treatment in syngas prior
fuel and chemical feedstocks. Figure 1 shows the carbon to FTS. General requirements for the catalysts are, among
chain length as a function of chain growth probability. The others, selectivity (low for methane; high for the targeted
influence of promoters on selectivity will be discussed later in hydrocarbon fraction), activity and stability, and mechanical
this section. robustness. The operation conditions, for example, high or low
Although there are many advantages with regard to iron- temperature (HTFT and LTFT), and the type of reactor
catalyzed FTS, the transformations of the iron catalyst during employed put specific demands on the catalyst synthesis
activation and FTS are rather complex and still not fully under- procedure. For a tubular fixed-bed-type reactor, minimization
stood. During catalyst preparation, iron oxides (e.g., hematite of mass transfer limitations is an important consideration, and
(Fe2O3) and magnetite (Fe3O4)) are produced and these are here catalyst strength is less important than catalyst shape
transformed to either a-Fe or iron carbides during activation and form. For a fluidized bed reactor, however, catalyst
depending on the conditions. strength as well as particle size and density are very important.40
532 FischerTropsch Synthesis: Catalysts and Chemistry

Table 2 summarizes the typical preparation methods used for
the various applications.
Key factors when choosing an iron source are cost and
availability. Although most of the iron oxides, hydroxides,
and oxy-hydroxides are readily available in nature, precursors
for iron catalysts are rather chemical grade raw materials.41
This is done to ensure that impurities that can influence
the catalyst are either removed or carefully controlled. It
is typical for commercial manufacturers of iron catalysts to
produce chemical grade iron(III)nitrate from sufficiently
pure scrap iron on site as part of the preparation. Large-scale
fusion preparation methods use iron ores or mill scale from
steel mills. Complex preparation methodologies that involve
novel chemicals and/or intricate transformations are rarely
commercially viable when compared to the tried and trusted
methods of precipitation and fusion. The gains from such
novel preparations must be truly unique to justify the addi-
tional expense.
7.20.2.3.1 Fusion
Fusion produces oxidic iron particles of low surface area, high
density, and high strength, which are ideally suited for appli-
cation in circulating fluidized bed reactors (Figure 3). During
the fusion process, the iron oxide raw material together with
the promoters are fed into an arc furnace where it is subjected
to temperatures above 1000  C. After fusion, the molten mate-
rial is cast into flat bars (ingots) and cooled. These ingots are
milled to a specified particle-size range to ensure optimum
fluidization (Figure 6).
The disadvantage of fusion is the fact that the inorganic
impurities, for example, silica and alumina oxides present in
the raw mill scale starting material, form inclusions during
cooling of the ingot. The alkali promoters migrate and bind
during cooling to these inclusions, which negates the promo-
tion effect. Figure 7(a) is a scanning electron microscopy
(SEM) image of a fused ingot showing clearly the inclusions
and with scanning electron microscopy energy-dispersive X-ray

Table 2 Catalyst preparation methods used for high and low temperature iron-based FischerTropsch processes

Reactor Important catalyst properties Raw material Synthesis method

HTFT
Circulating or fixed fluidized bed reactors, Low surface area (<10 g m2), high density, Mill scale Fusion followed by crushing and
320350  C high strength milling
LTFT
Tubular fixed bed reactor, 220250  C High surface area, sufficient strength Fe(NO3)3 and silica Precipitation followed by
source extrusion/shaping
Slurry bed reactors, 220250  C High surface area, small particles Fe(NO3)3 and silica Precipitation followed by spray
(50250 mm) source drying and calcination

Promoters

Cooling Size
Mill scale Oxidation Arc furnace reduction Reduction
step
(milling)

Figure 6 Catalyst preparation diagram for the HTFT fused iron catalyst. Mill scale means iron metal pieces from the steel industry.

Inclusion

Fe
Si
Alkali 1
Alkali 2

(b)
80 mm
(a)

Figure 7 (a) SEM image of a HTFT Fe catalyst cast ingot, showing inclusions of silica and alkali and (b) SEM EDX mapping on one of these inclusions.
 FischerTropsch Synthesis: Catalysts and Chemistry 533

spectroscopy (SEM EDX)(Figure 7(b)) one can see the pres-
ence of the alkali in these inclusions.
The development of fixed fluidized bed reactors diminished
the need for particles of high mechanical strength, and permit-
ted the use of catalysts with lower density and higher surface
area. One of the prospective routes to alternative HT FTS
catalyst precursors is the precipitation of high-density, low-
surface-area iron oxides, hydroxides, and (oxy)hydroxides
from solutions of iron(III) salts followed by calcination.
7.20.2.3.2 Precipitation
Precipitation of iron(III)oxides from iron(III)nitrate solutions
was one of the first methods reported in the literature for the
preparation of iron FTS catalysts.42 In the late 1930s, Ruhrch-
emie developed a large-scale preparation based on pre-
cipitation. In this procedure, the iron(III) salt is reacted with
a base to form an iron(III) oxide(oxy)hydroxide precipitate.
By variation in process conditions, for example, pH, precipita-
tion rate, and temperature, catalyst properties such as surface
area and crystallite size can be controlled. Figure 8 illustrates
the decrease of crystallinity versus surface area for the various
iron oxides known as precursors for Fe FTS catalysts.
After precipitation, the slurry is filtered and washed to
remove all the salts (e.g., NH4NO3) from the filter cake (see
Figure 9). The latter is then reslurried and impregnated with
structural promoters like Si, Al, etc. The application of chemical
promoters in the iron catalyst is discussed in more detail later in
this section. Next, the slurry is spray-dried to yield spherical
particles that are suited to slurry-bed and fixed fluidized-bed
reactors. The final step in the catalyst preparation is calcination.
This high-temperature treatment removes volatile impurities
such as water and NOx and increases the strength of the catalyst
particles. Figure 10 shows spherical particles obtained from
spray drying. For fixed-bed reactor applications, the impreg-
nated filter cake is extruded and dried at about 150  C.
The methodology described above is commercially applied
by Sasol for the synthesis of their slurry-bed reactor (SBR) and
tubular fixed-bed reactor catalyst precursors. Rentech uses a
similar method to synthesize their Fe FTS catalyst precursor.39
As mentioned above, Sasol developed in-house precipitation
methodology to synthesize a suitable Fe-HTFTS catalyst that
offers many advantages over the fused Fe HTFT catalyst.43 Some
of these include improved promoter distribution, increased
strength, and spherical particles which improve fluidization
of the catalyst. By replacing fusion with precipitation, the
negating effect of the alkali promoters could be eliminated
owing to the purity of the starting iron(III)salt.
Hematite: 12 27 m2 g-1
Magnetite: 4 100 m2 g-1
Degree
of
crystallinity
Goethite: 8 200 m2 g-1
Ferrihydrite: 100 700 m2 g-1
Figure 8 Degree of crystallinity is decreasing with an increased surface
area for various iron oxides.

Filtration/
Dissolution Precipitation washing
HNO , Base
Water
(aq) (aq)

Iron

Re-slurry and
impregnation

Salt
(aq)

OR

Spray drying Extrusion

Calcination Drying

Slurry bed Fixed bed

catalyst catalyst

Figure 9 Catalyst preparation diagram for the slurry-bed and fixed-bed precipitated iron LTFT catalyst.
534 FischerTropsch Synthesis: Catalysts and Chemistry
Figure 10 Spray-dried iron LTFT slurry-bed catalyst particles.
7.20.2.3.3 Improving iron FischerTropsch catalyst
precursors by promoters
FTS processes catalyzed by unmodified and unpromoted
iron catalysts suffer from poor selectivity, low activity, and
sintering, but the addition of structural and chemical pro-
moters addresses most of these issues.
Structural promoters, for example, Si, Al, and Mg, may sup-
press sintering, stabilize the active phase, and improve mechan-
ical strength. The addition of alumina and silica typically
increases the stability of hematite under FTS conditions.44 In
general, it is observed that in the presence of a structural pro-
moter such as silica the surface area of the iron oxide remains
high even after calcination at relatively high temperatures.
A potential disadvantage, however, of adding structural
promoters is that the activation, for example, reduction of the
iron oxide, becomes more difficult due to the formation of iron
silicates or aluminates. For this reason, chemical promoters
such as Cu or Ag are added during catalyst synthesis to increase
the rate of reduction, most likely due to hydrogen spillover
from the Cu surface to the iron oxide surface. Apart from
increasing the rate of reduction, chemical promoters are
known to (i) enhance nucleation of iron intermediates which
leads to higher surface areas, (ii) increase the number/type of
CO adsorption sites, (iii) stabilize selected phases, and (iv)
influence the rate of secondary reactions.45
Addition of alkali metals (e.g., potassium) to LTFT iron oxide
catalyst precursors is known to enhance the chain growth prob-
ability (increased C5 selectivity), to diminish methane forma-
tion, and inhibit secondary hydrogenation reactions, leading to
higher olefin to paraffin ratios.46 In a similar way, but perhaps
not as effective, alkali earth metals have been shown to increase
alpha values and to suppress methane formation.
7.20.2.3.4 Activation and reduction procedures
Iron oxides and (oxy)hydroxides are inactive for FTS and must
be activated to render an active catalyst. Depending on the
application HTFT or LTFT and iron oxide precursor, activa-
tion is performed in hydrogen, carbon monoxide, or synthesis
gas. The optimum activation conditions are influenced by the
type and quantity of chemical and structural promoters.
The stability and composition of the final activated iron phase
determine the performance of the catalyst under Fischer
Tropsch conditions. Because of this, the activation procedure
has a strong effect on selectivity and activity of the catalyst.
For activation in hydrogen, the extent of reduction of iron is
governed by the role of water removed during the activation.
The degree of reduction is governed by the equation47:
pH2 O
DG DG RT ln [4]
pH2
where G is the free energy change for the reduction under the
conditions employed, G is the standard free energy change
for the reduction reaction, R is the gas constant, T is the
temperature, and p is the partial pressure of the gases indicated.
The equation implies that the rate at which water is
removed from the reactor plays a critical role: the faster the
water is removed, the faster the reduction process proceeds,
and the higher the degree of reduction is. Rewriting exp-
ression [4] to include the equilibrium constant in the form of
the ratio (pH2O/pH2)eq at equilibrium, enables one to estimate
the degree of reduction that can be obtained:
"    #
pH2 O pH2 O
DG nRT ln [5]
pH2 pH2 eq
As the equilibrium ratio for reduction from Fe2O3 to FeO is
0.7, and for FeO to a-Fe is 0.1, the theoretical degree of reduc-
tion would be 50% at 10% water in the gas phase. Hence, for
high degrees of reduction the water content should be well
below 1%.47,48 Structural promoters such as silica and alumina
increase the resistance against reduction.
To fully understand activation, it is necessary to understand
the possible phase transformations that can occur. x-Ray dif-
fraction (XRD) and Mossbauer spectroscopy are ideal tech-
niques to distinguish between the different iron phases that
can arise, while small particle effects, which so often limit the
information content of XRD, are generally absent in the unsup-
ported iron FT catalysts.4951 Activation with CO present in the
gas typically leads to a mixture of metallic iron (a-Fe), iron
carbides (general formula FexCy), and magnetite (Fe3O4). The
relative quantities are a function of the reducing gas, the gas
hourly space velocity, and the temperature.
It is generally believed that carbides such as Hagg-carbide
(w-Fe5C2) are the active phase for FTS.52 The exact nature of the
surface carbidic species is still a subject of debate. The stabilities of
different bulk iron carbides have been reported to be in decreas-
ing order: e0 -Fe2.2C> e-Fe2C > w-Fe5C2 > y-Fe3C.53 Depending on
the type of catalyst (e.g., fused or precipitated), different iron
carbides were found to be characteristic for each type after acti-
vation. However, the possibility that a-Fe plays a role during FTS
cannot be ignored.21 Typically during activation, precipitated
catalyst precursors, for example, hematite (Fe2O3), are converted
to magnetite (Fe3O4) irrespective of the activation gas used.
However, after this transformation the final iron phase will
depend on the activation gas used, for example, a-Fe in the case
of hydrogen or Hagg-carbide (w-Fe5C2) in the case of CO or
synthesis gas. In a study by Herranz et al., it was found that the
activation of hematite using CO resulted in mainly cementite
(y-Fe3C) while activation in synthesis gas yielded Hagg carbide
(w-Fe5C2) (Table 3).54

The fused magnetite HTFT catalyst is reduced with hydrogen
at temperatures between 350 and 450  C and high linear flow
rates to avoid re-oxidation by water, as explained above. During
reduction, oxygen atoms are removed from the lattice leading to
an increase in surface area from <1 g m2 to 58 g m2 55.
The extent of reduction under these conditions was measured at
about 80% (a-Fe). Under CO or synthesis gas, virtually no
reduction of the nonporous magnetite was observed.
In the case of precipitated iron oxide catalyst precursors for
LTFT catalysts, the activation is usually done under much milder
conditions than for fused catalysts. These catalysts are more
amorphous with a high pore volume and surface area and the
oxide crystallites can sinter under too harsh activation condi-
tions. It is important to note that the success of activation of
precipitated iron catalyst precursors is coupled to FTS activity,
stability, and selectivity. This is dependent not only on the type
of reduction gas but also on process conditions, for example,
temperature and pressure. From the literature, it seems as
though activation under CO yields the optimally activated pre-
cipitated iron catalyst for FTS synthesis, as these gave the best
syngas conversion and lowest methane selectivity when com-
pared to catalysts activated with H2 or synthesis gas.56 However,
the final catalyst also had a relatively high WGS activity.
7.20.2.4 Selectivity Manipulation of Iron Catalysts
A key advantage of iron-catalyzed FTS is the fact that the
selectivity of the process can be manipulated, either by process
conditions (less responsive) or by catalyst composition (more
responsive). Lowering the temperature shifts the selectivity
Table 3 Names of the various iron phases
a-Fe2O3 Hematite
a-FeOOH Goethite
Fe3O4 Magnetite
FeO Wustite
w-Fe5C2 Hagg carbide
y-Fe3C Cementite
Ln (X )
Alpha 1
(LT FT) = 0.80
Alpha (HT FT) = 0.75
0 5 10 15 20
Carbon number
Figure 11 AndersonSchulzFlory distribution of hydrocarbons formed over an LTFT catalyst (typical Ruhrchemie catalyst) and a fused HTFT catalyst.
FischerTropsch Synthesis: Catalysts and Chemistry 535
from lighter to heavier hydrocarbons. Although the Anderson
SchultzFlory (ASF) distribution curve (see Section 7.20.1 and
Figure 2) gives a good indication of expected selectivities, alkali
promotion of iron catalysts leads to selectivities that tend to
deviate from ASF and are characterized by two alpha values
(see Figure 11).
In attempts to increase the selectivity toward valuable base
chemicals, mixed-metal oxides and/or multicomponent metals
are typically incorporated in precipitated iron catalysts. Addi-
tion of manganese to a typical Ruhrchemie catalyst increases
the selectivity toward alpha-olefins at low-temperature FTS
conditions (230  C and 20 bar total pressure).57 Another
example is the Fe/Zn/Mn/Cu/K/SiO2 catalyst for direct conver-
sion of synthesis gas to chemicals.58 The development of such a
technology is known as ChemFT, and has as primary focus to
shift the selectivity toward alcohols. Table 4 compares the
selectivities of the various iron FTS technologies (HTFT, LTFT,
and ChemFT) and illustrates that the alcohol selectivity in
ChemFT is much higher than that of a typical Ruhrchemie
catalyst under similar LTFT conditions.58
7.20.2.5 Catalyst Stability During FTS
Stability is a key characteristic of a successful catalyst. The ideal
FTS catalyst should maintain constant activity and a correspond-
ing stable selectivity during time on stream. Commercial reac-
tors and product workup sections are designed for a very narrow
set of optimum process conditions. The catalyst must perform
within these design constraints for as long as possible. This
determines the useful catalyst life. Unfortunately, iron catalysts
show considerable loss of performance over time. During recent
years, the focus of research has shifted from improving catalyst
activity to increasing the lifetime of the catalyst.
Deactivation of iron FTS catalysts is usually attributed to the
following factors:
(i) free carbon formation, leading to catalyst fouling,
(ii) activity loss due to transformation of the phase, for exam-
ple, oxidation,
LT FT
HT FT
Alpha 2 (LT FT) = 0.94
25 30 35 40 45 50
536 FischerTropsch Synthesis: Catalysts and Chemistry

(iii) mechanical break-up of the catalyst, so-called competition model.59 After the adsorption and dis-
(iv) deposition of poisons in the synthesis gas on the catalysts sociation of CO and H2, three reactions are possible:
surface, and
(i) C* iron ! carbides
(v) sintering.
(ii) C* xH* ! CHx*
Buildup of free carbon is one of the major causes of (iii) C* yC* ! inactive carbon
deactivation in HTFT. It leads to a decrease in density and
The first reaction describes the formation of iron carbides
strength of catalyst particles and results in catalyst bed expan-
from the reduced a-Fe under FTS conditions. The dissociated
sion, particle break-up, and carry-over of fine catalyst material
carbon (C*) can either react with dissociated hydrogen atoms
into downstream processes.25 Figure 12 shows an SEM image,
(H*) to yield hydrocarbons or react with another carbon atom
along with the distribution of elements of a spent catalyst
(C*) to yield inactive/ so-called free carbon.59 This type of
retrieved from a commercial fixed fluidized bed reactor. Cata-
deactivation can be suppressed by chemical promoters. In
lyst break-up and fines formation are easily recognized. As
recent years, Sasol developed another propriety catalyst involv-
carbon is dispersed through the bulk of the particle, break-up
ing the addition of chromium to reduce the amount of free
may lead to exposure of new active surfaces, thus helping to
carbon formed during FTS.60
maintain activity. At the same time, the carbon that is lost in
The formation of free carbon is less pronounced in the case
the form of fines is rich in alkali and thus removes some of the
of the precipitated LTFT iron catalysts. The main deactivation
chemical promoter, which degrades the selectivity.
mechanisms in this case are sintering and oxidation of the
Much has been discussed regarding the origin of the free
active phase. Interconversion of different carbides may lead
carbon in the catalyst. A plausible explanation is given by the
to a stoichiometric excess of carbon which in turn leads to
weakening of catalyst particles. Figure 13 shows a deactivation
curve for a typical Ruhrchemie catalyst under low-temperature
Table 4 Selectivity comparison between LTFT, HTFT, and FTS conditions. Samples of this catalyst taken from the reactor
ChemFT25,58 usually contain mixtures of highly dispersed magnetite and
iron carbide (both containing around 2 nm particles).61 The
Product Fe HTFT Fe LTFT Fe ChemFT highly dispersed magnetite particles can either react in synthe-
sis gas to the required iron carbide, or they can agglomerate or
CH4 (%) 8.0 3.0 18.0
sinter into larger inactive particles (about 40 nm). The larger
C2C4 olefins (%) 24.0 4.0 21.0
C2C4 paraffin (%) 6.0 4.5 17.7 magnetite particles can agglomerate further to yield large glob-
C5C6 (%) 16.0 7.0 13.3 ules (around 400 nm). Surprisingly, agglomeration or sinter-
C7350  C (middle distillate 36.0 26.5 20.5 ing of highly dispersed iron carbide into less active or inactive
product) iron carbide particles of about 20 nm has also been observed
350  C (wax products) 5.0 51.0 0.0 (Figure 14).
Oxygenates as alcohols (%) 2.8 3.8 8.3
Oxygenates as acids ketones 2.2 0.2 0.9
% breakdown (C5C12 cut) 7.20.2.6 Spent Catalyst Management
% total paraffins 13.0 29.0 49.6
% total olefins 70.0 64.0 37.8 Regeneration of spent iron FTS catalysts is difficult, due to the
% aromatics 5.0 0.0 0.0 sintering of the particles during FTS. Successful regeneration
% oxygenates 12.0 7.0 12.5 requires redispersion of the sintered phase, and this cannot
easily be achieved. Reactivation by re-reduction is possible, but
the activity of the reactivated catalyst is lower because the

2.0
Relative activity ratio

1.5

1.0

0.5

0.0
-Fe -C -Si 0 50 100 150 200 250
Time on line (h)
Figure 12 Scanning electron microscopy (SEM) image of a spent,
fused Fe HTFT catalyst; color coding: red, iron; yellow, carbon; and green, Figure 13 Relative activity versus time on stream for a precipitated
silicon. Ruhrchemie-type iron LTFT catalyst.

H2 + CO
(Fe3O4)HD (FeO)HD
-H2O
Sintering
Sintering
(Fe3O4)LP (Fe3O4)G
50 nm
HD highly dispersed phase 2 nm
LP larger particles 20 nm
G globules 400 nm
Figure 14 Deactivation mechanism for a typical Ruhrchemie iron catalyst under low-temperature FischerTropsch synthesis conditions.
original surface area cannot be recovered. Multiple reactivation
steps are therefore not viable. Spent HTFT catalysts may in
principle be recycled to make new catalysts. However, using
the material in the fusion process has a high energy cost associ-
ated with it, due in part to its high carbon content. Iron is a
cheap material, and there is little economic incentive for recov-
ering it. Therefore, spent catalysts have usually been landfilled.
Currently, awareness of the environmental impact of such pro-
cedures is growing, and reclamation of metal even iron from
spent catalysts, for example, by acid dissolution is more and
more seen as a social responsibility of the industry to reduce the
impact of commercial processes on the environment.
7.20.3 Cobalt-Based FTS Catalysts
7.20.3.1 Introduction
Cobalt as an FTS catalyst was already claimed by Fischer and
Tropsch in their original patent of 1925.31 The commercializa-
tion of the FTS by Germany and Japan in the period 193845
relied fully on cobalt catalysts. Only after World War II did the
focus shift to the use of iron catalysts for FTS applications.
Since the oil crises of the 1970s the interest in cobalt-based
FTS catalysts reappeared, which has resulted in numerous sci-
entific papers and patents (see Figure 4). Many companies
showed interest in cobalt FTS, for example, BP, ConocoPhilips,
Gulf, ExxonMobil, IFP, Johnson Matthey, Sasol, Shell, Statoil,
and Syntroleum. Almost all focused on wax production, fol-
lowed by hydrotreating to produce diesel. This is also the
application that will receive most attention in this section.
Cobalt FTS catalysts are exclusively utilized in low-
temperature synthesis or LTFT, and are applied in fixed-bed,
slurry-phase, and micro-channel FTS reactors. Catalyst design
needs to be adjusted to the targeted reactor as well as the applied
FTS conditions. Important for catalyst design are the composi-
tion, method of preparation, activity and selectivity behavior,
deactivation and regeneration, and mechanical integrity.
FischerTropsch Synthesis: Catalysts and Chemistry 537
H2 + CO
CnHm
FTS
CO CO
(Fe)HD (FexCy)HD
-CO2 -CO2
H2O Oxidation
Sintering
(Fe2C)LP
50 nm
200 nm
Cobalt FTS catalysts are currently commercially applied by
Sasol/QP in the Oryx GTL plant, Qatar (Co/Al2O3), in a slurry-
phase reactor, and by Shell in the SMDS plant in Bintulu,
Malaysia, as well as in the Pearl plant, Qatar (both Co/Mn/
TiO2), in a fixed-bed reactor. Figure 15 shows the catalyst that
is used in slurry-phase application.
Exciting academic and industrial research in the last 20
years has increased the fundamental knowledge of cobalt FTS
catalysts substantially on topics like the nature of the active
site, impact of crystallite size on activity and selectivity, and
deactivation mechanisms, owing to the application of surface
science techniques, model catalysts, in situ analyses at relevant
industrial conditions, and molecular modeling.8,19,27,6267 The
literature of the last 20 years shows that quite a wide variety of
cobalt catalyst compositions prepared by numerous methods
can result in academically and industrially relevant cobalt-
based FTS catalytic systems.
7.20.3.2 Composition of Cobalt Catalysts
Modern cobalt catalysts are similar to the ones prepared by
Fischer and Tropsch in the sense that they consist of promoted
cobalt on a metal oxide support. An inspection of the literature
and patents on this topic reveals the following general charac-
teristics, with almost all companies with FTS catalysts having a
similar formulation for them18,35,68 (Table 5):
(a) Cobalt as the FTS active metal (typically 1030 wt%)
(b) A second metal (usually noble) as a reduction promoter
(0.051 wt%)
(c) A structural oxidic promoter (e.g., Zr, Si, and La) (110 wt%)
(d) A refractory oxidic support (most likely modified)
Cobalt is expensive and to maximize its use, it needs be well
dispersed on the support. Since cobalt metal is considered the
active phase, it is imperative that there is a high density of
cobalt metal sites available. The number of cobalt surface
sites is a function of particle size and morphology, extent of
538 FischerTropsch Synthesis: Catalysts and Chemistry

Cobalt catalyst

~ 60 m
Wax

0.1 m

-4
5 10 m
Support particles
10 m

Cobalt

Slurry phase reactor Support

-8
1 10 m 1 10-7 m
A cobalt nanoparticle Cobalt and support

Structural promoter

Reduction promoter

Cobalt

Cobalt
Cobalt

Modified support

Figure 15 Cobalt catalysts for application in a slurry-phase reactor, and schematical composition of a typical cobalt-based FischerTropsch catalyst.

Table 5 Examples of catalyst formulations, as patented by several industrial FischerTropsch synthesis companies

Company Composition Reference Preparation route

Co (wt%) 2nd metal Structural promoter Support

Shell 20 MnO (Co/Mn 12) TiO2 WO 199700231 Coprecipitation

ExxonMobil 12 Re (1 wt%) Al2O3 (6 wt%) TiO2 US 5268344 Impregnation
Syntroleum 20 Ru (0.1 wt%) La (1 wt%), SiO2 (0.110.6 Si/nm2) Al2O3 WO 2005058493 Impregnation
BP 10 Al2O3 (0.5 wt% Al) ZnO WO 19913400 Impregnation
Sasol 20 Pt (0.05 wt%) SiO2 (0.8 Si/nm2) Al2O3 US 7365040B2 Impregnation

reduction, and particle stability.68 It is preferred to have a fairly
high extent of reduction (>60%), but it should also be noted
that the cobalt is further reduced during the FTS reaction. An
optimum cobalt particle size of just above 810 nm is pre-
ferred as particles below those have shown to have a lower
turnover frequency (TOF).69 Additionally, very small particles
(46 nm) could be more prone to sintering and also may prove
very difficult to reduce due to an increased metal-support
interaction. It is important that there is a minimum amount
of cobalt-support compounds as these are reducible at very
high temperatures and are inactive for the FTS reaction.70 The
two most common phases of metallic cobalt in supported
cobalt FTS catalysts are face-centered cubic (fcc) and hexago-
nally close-packed (hcp), which often coexist.19,71 It has been
reported that for cobalt particles less than 40 nm, the predom-
inant phase should be fcc.72 The mode of activation, addition
of promoters, and support may influence the relative amounts
of the phases.19 Some authors have reported that the hcp phase
is more active for FTS.73
Nanometer-sized cobalt particles when supported on tradi-
tional oxidic carriers like silica, alumina, and titania are diffi-
cult to reduce due to strong interactions with the support.
Therefore, catalysts are often promoted with a second metal
(e.g., Ru, Pt, or Re) which leads to improved reducibility of the
cobalt oxide particles; the increase in amount of active sites
results in higher activity compared to un-promoted catalysts.
The more facile cobalt reduction is attributed to faster hydro-
gen activation in the presence of promoter metals and

subsequent spillover of hydrogen to cobalt oxides and reduc-
tion of cobalt species.19,68 In many cases, the promotion with
noble metals leads to a smaller average size of either cobalt
oxide or cobalt metal particles. Promotion with noble metal
may also play a role during the decomposition of the cobalt
precursors and can lead to crystallization of smaller cobalt
oxide particles. This increased dispersion is most likely due to
a higher rate of nucleation, enabled by the promoter.68
Promoter metals such as Ru have also been claimed to lead
to the formation of bimetallic particles and alloys. This influ-
ences catalyst activity and selectivity, may inhibit deactivation
by keeping the surface clean, and allows easier regeneration of
the cobalt surface.74 The metal promoter is usually present at
levels of 0.10.5 wt%. At these low concentrations reduction is
efficiently promoted, and the hydrocarbon selectivity is hardly
negatively affected.
Structural promoters affect the formation and stability of the
active phase of a catalyst material. For Co/silica catalysts, it has
been shown that promotion with Zr results in a decreased cobalt
silica interaction, which in turn leads to a higher degree of cobalt
reduction and increase in the metallic atoms on the surface.19,75
Zr promotion of cobalt/alumina catalysts has been claimed to
prevent formation of cobalt aluminate.76 Incorporation of ele-
ments such as B77 and Ni78 increases the stability of cobalt
catalysts by suppressing carbon formation. Irreducible oxides
such as MnO and CeO2 may also slow down cobalt sintering.63
A wide range of promoters has been studied; the reader is referred
to a detailed review by Morales and Weckhuysen.63
The support provides mechanical strength and thermal sta-
bility to the cobalt crystallites, while facilitating high cobalt
dispersion. The properties of the support are an important
factor. For alumina, high purity, low acidity, and relatively
high surface area (150250 m2 g1) are required, according to
patents from the 1980s.7981 More recently, however, alumina-
based supports of relatively low surface area (50 m2 g1), such
as Ni-promoted a-Al2O3, have been reported to have a positive
effect on both mechanical strength and C5 selectivity.82 The
pore size of the support can also influence the size of the cobalt
crystallites, as shown by Saib et al. for SiO2-supported
catalysts.83 Van Steen and Claeys reported that the desired
pore size of the support for the optimum cobalt crystallite
size should be around 1216 nm.61
The support needs to be robust under FTS conditions, imply-
ing that it should be able to cope with the presence of several
bars of steam that occur at high conversion levels. Van Berge
et al. found that an unprotected alumina-supported cobalt FTS
catalyst was susceptible to hydrothermal attack during realistic
FTS conditions, which resulted in contamination of product wax
with ultra-fine, cobalt-rich particulates.23,84,85 This problem was
solved by pre-coating the support with silica as structural pro-
moter. TiO2 seems to be the support of choice for both Exxon
and Shell based on the most recent patents (Table 5). An advan-
tage of TiO2 is that it has a high hydrothermal stability and can
withstand high water partial pressures. The rutile/anatase ratio
can be tailored, which influences the surface area and mechan-
ical properties.
Supported cobalt catalysts should also be resistant to attrition
especially if applied in a slurry bubble-column environment.
Wei et al.86 noted that the attrition resistance of supported cobalt
catalysts follows the sequence: Co/Al2O3 > Co/SiO2 > Co/TiO2.
FischerTropsch Synthesis: Catalysts and Chemistry 539
There has also been work conducted on less conventional
supports such as MCM-41, SBA-16, and carbon nanofibers,
nanotubes, and spheres.19,69,87 These studies are mainly aca-
demic in nature but further fundamental understanding of
cobalt FTS catalysts considerably. Carbon supports interact
weakly with cobalt and allow for a high degree of cobalt reduc-
tion, thus enabling the study of cobalt particle-size effects.69,87
7.20.3.3 Preparation of Cobalt FischerTropsch Catalysts
The preparation of cobalt FTS catalysts aims to achieve the
optimal crystallite size distribution in a particle that is optimal
for its application in a specific fixed-bed, bubble-column, or
microchannel reactor. As the optimum size range of catalyst
particles for the different reactor types varies (see Figure 3),
preparation methods and equipment depend on the targeted
reactor application. Important considerations for choosing a
particular method of preparation and starting components are
to minimize poisons (e.g., Na, S, Cl) in the catalyst and the
type of waste streams resulting from the chosen method.
A number of procedures for preparing cobalt FT-catalyst
precursor exist:
coprecipitation of cobalt, promoters, and support, followed
by catalyst particle shaping. In a variation of this method,
the support is added just before particle shaping,
precipitation or impregnation of cobalt and promoters
onto pre-shaped support particles, and
impregnation of cobalt (oxide or metal) particles onto pre-
shaped supports.
7.20.3.3.1 Precipitation
Most of the initial FTS-catalysts (e.g., Co/ThO2/kieselguhr)
were made by coprecipitation.88 This method has been applied
for some of the modern cobalt catalysts as well, for example,
for Co/Mn catalysts,89 Co/Mg/SiO2 and Co/ZnO2.90
Catalyst preparation based on coprecipitation usually con-
sists of three steps: precipitation, washing and drying, and
shaping. Selection of chemicals is of course an important con-
sideration in view of the associated waste streams.
Chemical precipitation of the cobalt, promoter, and sup-
port by a precipitation agent can be done batchwise or contin-
uously at constant pH. The cobalt precipitates as a hydroxide,
which can exist as green a-Co(OH)2 or pink b-Co(OH)2 poly-
morphs. The former is metastable and readily transforms into
the stable b-phase. Crystallite size and composition of the
precipitate are controlled by temperature, precipitation agent,
precursor salts, structure directing or organic hydrolysis
reagents, aging time, and reaction atmosphere (air or N2).
Using Na2CO3 or KOH as precipitation agents in the prepara-
tion of Co/SiO2 catalysts would lead to cobalt silicate
formation.91 To prevent formation of the inactive cobalt sili-
cate, the silica is added after the precipitation.
Filtration and washing of the precipitate is required to
remove excess chemicals. Even low levels of alkali metals and
halogens left in the washed precipitate can severely degrade the
catalysts performance.
Shaping of the catalyst precursor depends on the reactor
application. For bubble beds, the precipitate is usually reslurried
and spray-dried to obtain the required particle-size distribution.
540 FischerTropsch Synthesis: Catalysts and Chemistry

For fixed-bed reactors, the precipitate is extruded or pelletized.
Addition of acids to the washed and dried precipitate is done to
improve the final catalysts particle strength.92
7.20.3.3.2 Preparation methods involving pre-shaped
supports
Support morphology and characteristics play an important role
in optimizing the preparation of cobalt catalysts on pre-shaped
support particles. As the aim is to get a desired amount of
cobalt crystallites onto the support and maintain a crystallite
size of around 810 nm, the following support characteristics
need consideration:
The support pore volume dictates how much cobalt precursor
can be added per impregnation. As shown in Table 6, 30 g of
metallic cobalt per 100 g of support occupies only 0.03 ml g1
of support material, but when using Co(NO3)26H2O as the
precursor, a pore volume of 0.79 ml g1 is required.
The time required to get a homogeneous cobalt distribution
throughout a support particle during impregnation depends
on particle geometry (diffusion path length and pore size),
viscosity of the suspension, interaction between solution and
support surface (contact angle and surface tensions), and
diffusion coefficients.9395 Figure 16 shows the effect of dif-
fusion on the different cobalt distributions observed on small
alumina particles for incipient wetness impregnation com-
pared to slurry-phase impregnation, as obtained from SEM/
EDX line scans.96,97
For fixed-bed catalysts, eggshell-type cobalt distributions
are sometimes preferred to overcome pore diffusion limita-
tions on performance and selectivity. Concentrating the
cobalt in the outer layers of the support is, among others,
achieved by adding viscosity enhancers or using cobalt salt
melts for impregnation.98
For deposition precipitation onto pre-shaped supports, the
same parameters that determine the time required during
impregnation to get a homogenous distribution (e.g., par-
ticle geometry, diffusion path length, pore size, viscosity of
the suspension, interaction between solution and support
surface, contact angle and surface tensions, and diffusion
coefficients) are important.99

Table 6 Pore volume requirements for different cobalt components

Cobalt compound Molar mass (g mol1) Cobalt mass fraction (%) Density (g cm3) Pore volume required for a loading of 30 g
of Co per 100 g of support (ml)

Co 59 100 8.9 3.4

CoO 75 0.79 6.4 5.9
Co3O4 241 0.73 6.1 6.7
Co2O3 166 0.71 5.2 8.2
CoOOH 92 0.64 5.0 9.4
Co(OH)2 93 0.64 3.6 13.2
Co(NO3)2 183 0.32 2.5 37.7
Co(NO3)2 6H2O 291 0.20 1.9 79.0
CoCl26H2O 237 0.25 1.9 62.5

60
60

50 50

40 40
Mass% Co3O4
Mass% Co3O4

30 30

20 20

10 10

0 0
-10 0 10 20 30 40 50 60 70 80 -10 0 10 20 30 40 50 60 70
Distance from edge of catalyst particle (mm) Distance from edge of the catalyst particle (mm)
Figure 16 Macroscopic cobalt crystallite distribution, as measured by scanning electron microscopy (SEM) line scans for: (a) incipient wetness
impregnation followed by immediate fast drying and (b) slurry-phase impregnation allowing 3 h for the cobalt to slowly disperse throughout the alumina
particle, followed by fast drying.

In addition, diffusion differences between the precipitation
agent and cobalt will impact the final cobalt distribution
during precipitation. When depositing bulky cobalt
hydroxide crystallites or cobalt metal particles100 onto pre-
shaped support particles, the important parameters are pore
size, particle diameter, and bulkiness of the precipitate.
During the drying of the catalyst precursor, the same
parameters as highlighted above for impregnation and
deposition should be taken into account to prevent the
cobalt precursor (if not chemically fixed to the support)
from migrating out of the particle again. In addition, heat
transfer coefficients, evaporation enthalpies, and particle
outer surface area need consideration for optimizing the
drying phase of the cobalt catalyst preparation.
Each preparation method needs to be optimized carefully,
and one cannot assume that the optimum procedures for one
type of support and support particle shape will be the same for
all other supports and support particle shapes.
7.20.3.3.3 Calcination
Usually, drying is not fully accomplished and therefore the first
stages of calcination actually complete the drying phase. To
maintain the cobalt distribution achieved by impregnation or
precipitation during drying and calcination, the cobalt compo-
nent mobility must be hindered. One way of achieving this is
to ensure that the cobalt component stays in a viscous or solid
form. For catalysts obtained by impregnation from cobalt
nitrate solutions, this implies that during calcination the com-
bination of heating rate and air flow must be such that water
and NOx are immediately removed.101 As the mobility of the
cobalt phase can be minimized by fast calcination, heat flow
into the system is also important as both the drying and nitrate
decomposition are endothermic.
Performing calcination under different atmospheres pro-
vides a way to affect the dispersion of the cobalt phase. NO
addition during calcination leads to the formation of a less
mobile cobalt hydroxyl nitrate.102 Using H2 or CO as decom-
position medium at temperatures below those where reduction
takes place also gives catalysts with good cobalt dispersions.103
Adding organic additives during impregnation is another
method to influence cobalt nitrate decomposition. Oxidation
of the additive is an exothermic process, which provides heat
for the endothermic nitrate decomposition, and thus acceler-
ates its decomposition.
The transmission electron microscopy (TEM) images in
Figure 17 illustrate how cobalt distributions change when
different calcination conditions are applied.
7.20.3.3.4 Reduction
Cobalt catalysts are usually reduced in hydrogen or a diluted
hydrogen atmosphere. Examples of CO reductions are also
found, but carbon formation on the cobalt crystallites should
be avoided. Reduction of cobalt oxide to cobalt metal occurs in
two exothermic steps:
Co3 O4 H2 ! 3CoO H2 O [6]
CoO H2 ! Co H2 O [7]
FischerTropsch Synthesis: Catalysts and Chemistry 541
For optimal reduction, care must be taken to optimize heat
transfer, minimize hydrogen diffusion and mass transfer limi-
tations, and to remove water effectively. The latter benefits
from high hydrogen space velocities and application of low
heating rates.104 Figure 18 shows the impact of the water
partial pressure during reduction of a 30 g Co/0.075 g
Pt/100 g alumina catalyst on the starting FTS performance,
indicating that low water content should be targeted for max-
imum activity, in agreement with the thermodynamics of
reduction as expressed in eqn [5]. Hence, the hydrogen stream
used for reduction must be as dry as possible. For small catalyst
particles as used in slurry-bed reactors, fluidized bed reduction
reactors are preferred and the above requirements are easily
met. For reductions in fixed-bed reactors, more care must be
taken to overcome the limitations especially toward the reduc-
tion reactor outlet.
The maximum temperature required for reduction of cobalt
catalysts depends on the level of reduction promoter present
(Pt, Ru, Pd, etc.), the presence of other promoters (e.g., alkali
metals make reduction more difficult), the support, the sup-
port modifiers, and the catalyst precursor used in the prepara-
tion. Reduction temperatures that are too high can cause
sintering and loss of cobalt metal surface area. In the case of
Co/SiO2 catalysts, cobalt silicate formation has been reported
for temperatures higher than 350  C.
Catalyst performance depends critically on the reduction
procedure.105 Application of reductionoxidationreduction
(ROR) cycles has been reported to improve the FTS perfor-
mance of cobalt catalysts by up to 30%.106 Some of the reasons
given in the literature for this improved performance from
ROR treatment are: (1) rougher (more steps on the surface)
cobalt crystallites, (2) higher degree of reduction, and (3) re-
dispersion of the cobalt on the support surface.
7.20.3.4 Cobalt Catalyst FischerTropsch Performance
Both activity and selectivity are of course important parameters
for cobalt-based FTS catalysts. High activity is important for
slurry-phase catalysts, while for fixed-bed catalyst the heat
removal capacity needs to be balanced with the activity of the
catalyst. From a selectivity point of view, a low methane selec-
tivity is normally desired, combined with a high C5 selectivity
or a high chain growth probability (a). Determining the intrin-
sic catalytic performance of cobalt FTS catalysts is not a
straightforward exercise, as it is influenced by the choice of
reactor and conditions. Khodakov et al.19 summarize a num-
ber of issues and choices related to the testing of FTS catalysts,
for example: (i) reactor choice: fixed-bed, slurry-bed (or con-
tinuous stirred-tank reactor, CSTR), or high-throughput reac-
tors, (ii) hydrodynamics, (iii) heat transfer and hot spots, (iv)
intra particle and external mass-transfer limitations, and (v)
atmospheric or elevated pressure. As testing catalysts under
different FTS conditions (H2/CO ratio, T, and P) will result in
different catalyst performances35,107 and therefore possibly
selection of different catalysts, it is important in the early stages
of research to understand the long-term scaling-up view, with
respect to reactor choice and FTS conditions. Taking the same
cobalt catalyst and testing it in different manners can result in
very different catalytic activity behavior. Figure 19(a) clearly
shows that the FTS activity is influenced by the water partial
542 FischerTropsch Synthesis: Catalysts and Chemistry

0.5 m 0.5 m

(a) (b)

0.5 m 0.5 m

(c) (d)

Figure 17 Cobalt crystallite distributions as measured with STEM for cobalt alumina catalysts calcined in different manners. (a) Cobalt oxide
microglobical formation of a 30 g Co/100 g alumina catalysts using a heating rate of 1  C min1 and an air space velocity of 1 m3n per kg Co(NO3)2.6H2O
per hour. (b) Cobalt oxide distribution of a 30 g Co/100 g alumina catalysts using optimized heating rate and air space velocity to ensure optimum
calcination. (c) Cobalt oxide distribution on a 30 g Co/100 g alumina catalysts using carbon coated alumina , using the same heating rate and air flow rate
as in (a). (d) Cobalt oxide distribution on 30 g Co/100 g alumina catalysts using the same heating rate and flow rate as (a) but with 1% NO in He as
calcination atmosphere.

pressure applied during the test, which possibly impacts factors
such as sintering and carbon deposition, as well as surface and
active site reconstruction.65 The selectivity behavior of cobalt
catalysts is also strongly influenced by parameters such as
temperature, hydrogen and carbon monoxide partial pressures,
and conversion. Comparing catalysts tested at different condi-
tions should thus be done with care.
Figure 19(a) clearly shows that cobalt catalysts are more
active in fixed-bed than in slurry-bed reactors. However, cobalt
catalysts in slurry-phase reactors are normally applied at tem-
peratures around 230  C, while in fixed-bed reactors they are
normally used at temperatures around 210  C. The productiv-
ity per gram of catalyst is therefore higher in slurry-phase
reactors (Figure 19(b)).
For heterogeneous catalysts the activity often increases with
smaller particle size, as the metal surface area increases. This
was confirmed by Iglesia108 who showed that the activity of
cobalt catalysts is directly proportional to the amount of cobalt
metal surface in the catalyst. The TOF or the reaction rate per
unit of cobalt surface area was stable over the range of cobalt
particles that was investigated (9200 nm). FTS over cobalt
catalysts was therefore regarded to be structure insensitive.
Thereafter, a number of authors have investigated the effect
of cobalt metal particle size on the intrinsic activity of sup-
ported cobalt catalysts for smaller cobalt particles, that is, well
below 10 nm.69,87,109112 Bian et al.110 using Co/SiO2 con-
firmed Iglesias results for samples with cobalt particles
between 11 and 29 nm, but Barbier et al.,109 Bezemer et al.,69
Martinez and Prieto,111 Coville and coworkers87 all showed
that the TOF was stable for catalysts with cobalt particles above
810 nm, while it decreased sharply for catalyst with smaller
particles. Only Borg et al.113 reported no sensitivity for the
 FischerTropsch Synthesis: Catalysts and Chemistry 543

activity of cobalt particles with sizes down to 3 nm. The
relationship between TOF (or reaction rate per unit cobalt
surface area) and cobalt particle size for above-mentioned
publications is summarized in Figure 20. As the TOF numbers
were obtained under different FTS conditions (i.e., tempera-
ture and partial pressures), they were normalized to enable
comparison of trends in the different papers and therefore
expressed in arbitrary units. It is clear that for catalysts with
cobalt crystals above 10 nm the TOF is structure insensitive,
while there is a sharp decrease in activity for particles smaller
than 810 nm.
1.0
Relative FT activity
As mentioned above, the C5 selectivity depends strongly
on the FTS process conditions. For a constant set of conditions,
Iglesia114 found no particle-size effect on selectivity for parti-
cles between 9 and 100 nm. A few years later, however, Barbier
et al.109 reported a strong dependency of the chain growth
probability, a, on particle size. Increasing the cobalt diameter
from 4.5 to 9.5 nm caused the a-value to increase from 0.74 to
0.87 (at 170  C, 1 bar). Bian et al.110 showed a similar, though
less pronounced, trend with the a-value increasing from 0.85
to 0.89 when the particle size increased from 11 to 29 nm (at
200  C, 10 bar). Bezemer et al.69 reported a very clear particle-
size effect on selectivity at atmospheric pressure, with a meth-
ane selectivity that was stable for particles larger than 6 nm, but
increased sharply for smaller particles (220  C, 1 bar).
However, the reported data at high pressure (35 bar and
210  C) clearly show that the C5 selectivity still increases with
increasing particle size up to 15 nm. Xiong et al.,87 Prieto
et al.,112 and Borg et al.113 all confirmed the general trend on

an increasing C5 selectivity with increasing particle size

0.9 extending beyond 10 nm, and up to 20 nm.
Little fundamental understanding has been offered to
explain this particle-size effect on both activity and selectivity.
0.8 Interestingly, the effect of size on activity is very pronounced
for particles smaller than 10 nm, while the impact on selectiv-
ity seems to be more gradual and does not level off above
0.7
10 nm. The particle-size effects cannot, as previously suggested,
10 30 50 70 90 110 130 150 170
be explained by the oxidation of the smallest particles. Bezemer
PH 0 (mbar)
2
et al.69 showed with x-ray absorption near-edge structure
Figure 18 Impact of water partial pressure during reduction of a 30 g (XANES) measurements that oxidation did not occur. This is
Co/0.075 g Pt/100 g alumina catalyst on its initial FischerTropsch in line with extensively reported research that cobalt oxidation
synthesis performance. during FTS does not occur for cobalt particles larger than

1.6
Productivity (mole CO converted/g cat/s)

1.4

1.2
Activity (au)

0.8

0.6

0.4

0.2

0
0 200 400 600 800 1000 0 200 400 600 800 1000
Time (h) Time (h)

Figure 19 (a) Three FischerTropsch synthesis runs with same Co/Al2O3 catalyst, as tested at 20 bar, 230  C, and H2/CO 2. Red solid: fixed bed,
PH2 O inlet 0; PH2 O outlet 3.0 bar. Red open: fixed bed with water co-feeding; PH2 O inlet 2.6, PH2 O outlet 4.0 bar. Blue solid: slurry bed;
PH2 O inlet 0; PH2 O outlet 4.5 bar; all the catalyst is exposed to outlet water partial pressure in a CSTR laboratory slurry reactor. (b) Two Fischer
Tropsch synthesis runs with same Co/Al2O3 catalyst, as tested at 20 bar, 60% conversion, with the fixed bed run at 210  C (red circles) and the
slurry phase run at 230  C (blue triangles).
544 FischerTropsch Synthesis: Catalysts and Chemistry

2 nm.65 The particle-size effect can also not be explained by
sintering as this was not observed by Bezemer et al.69 and by
the Coville group.87 It is clear that the Co particle-size effect in
FTS extends beyond the classical impact of size, which derives
from the fraction and type of surface atoms as a function of
crystallite size and normally does not extend beyond 4 nm
particles. It was suggested69 that the optimum combination
of active sites for the different elemental reactions of FTS (i.e.,
CO dissociation, hydrogenation, and insertion) requires rela-
tively large cobalt particles, possibly combined with a CO-
induced surface reconstruction. This might be related to the
presence or absence of the so-called B5 site,65 which has been
speculated to be the most active site for CO dissociation, and
needs a certain particle size to be present in high abundance.
Another speculation is that the particle-size effect might be
related to specific bonding modes of CO, such as the bridge-
bonded CO coordination; this mode is believed to be favored
on large particles and held responsible for an increased CO
dissociation rate, which would lead to an increased reaction
rate.87,109 Based on in situ Fourier transform infrared
spectrometry results, Prieto et al.112 proposed an enhancement
TOF (au)
of CodSiO2 sites in small particles, caused by flattening of
the crystals during FTS, and possibly due to carbon-induced
surface reconstruction. Another study showed that carbon and
oxygen atoms, originating from dissociated CO, were very
strongly bonded on small particles, possibly blocking active
sites for further CO dissociation.117
As the impact of the particle size is different for activity from
for selectivity, the fundamental explanation of these effects
seems to involve more than one kinetically relevant mechanis-
tic step in the FTS process. Understanding this requires more
fundamental research.
7.20.3.5 Deactivation and Regeneration of Cobalt
FischerTropsch Catalysts
Catalyst stability is crucial for the economics of cobalt FTS,
in addition to other important factors such as high activity,
selectivity, and mechanical strength. Understanding catalyst
deactivation is essential for improving catalyst stability and for
developing effective regeneration procedures. Figure 21 shows a
typical deactivation profile for Co FTS catalysts under

Coville (2011)
Fischer (2010)
Pietro (2009)
Borg (2008)
Martinez (2007)
De Jong 1 bar (2006)
De Jong 35 bar (2006)
Bian (2003)
Barbier (2001)
Iglesia (1997)

0 5 10 15 20 25 30
Co particle size (nm)

Figure 20 The TOF or FischerTropsch synthesis rate per unit surface area, as a function of the cobalt metal particle size.69,109,110,112116 The TOF had
to be scaled due to variations in process conditions, and is therefore reported in arbitrary units.

1.0

0.8
RIAF

0.6

0.4

0.2

0.0
0 10 20 30 40 50 60
Time on line (days)

Figure 21 Normalized activity stability for a Co/Pt/Al2O3 catalyst during realistic FischerTropsch synthesis in a laboratory scale micro-slurry
reactor at fixed CO conversion (230  C, 20 bar, H2 CO conversion of 5070%, feed gas composition of 5060 vol.% H2 and 3040 vol.% CO). Adapted
from van de Loosdrecht, J.; Bazhinimaev, B.; Dalmon, J. A.; Niemantsverdriet, J. W.; Tsybulya, S. V.; Saib, A. M.; van Berge, P. J.; Visagie, J. L.
Catal. Today 2007, 123, 293302.

commercially relevant FTS conditions.118 Deactivation is initially
stronger, after which it starts leveling off. Similar deactivation
profiles have been reported by other laboratories.18,62,119,120
Fundamental studies on catalyst deactivation essentially
involve understanding differences in the fresh and spent cata-
lysts with respect to the active site. The B5 sites on metallic
cobalt are currently considered as active sites for the FTS.65
Hence, changes to the number or nature of these sites will
contribute to deactivation. However, due to the complexity of
the FTS process and the lack of suitable techniques to charac-
terize any particular site on a surface, most studies on deacti-
vation are necessarily limited to observables such as changes
in metallic cobalt surface area. This in itself is no trivial matter,
as cobalt catalysts are sensitive to their environment, and spent
FT catalysts are embedded in wax (which actually protects
them from being exposed to the air). As is so often the case
in fundamental studies of catalysts, a combined approach
using real catalyst and model systems, advanced in situ and
ex situ characterization, combined with molecular modeling
has given detailed insight into FTS catalyst deactivation.65
The main deactivation mechanisms of cobalt FTS catalysts
and its active sites, as proposed in the literature, are: (1) oxi-
dation, (2) mixed metal-support interaction, (3) carbon depo-
sition and carburization, (4) sintering, (5) poisoning, and (6)
surface reconstruction.19,62,64,65,70,71,108,114,118,121133
Over the last 15 years, oxidation of cobalt by the product
water was seen as the major deactivation mechanism in the
open literature.65 However, many recent publications have dis-
proved this.62,64,65,71,118,134Following an in-depth study on oxi-
dation using model systems, molecular modeling, surface
thermodynamic calculations, and an industrial catalyst tested
under commercially relevant conditions, the key finding is that
oxidation is crystallite size and condition dependent, that is,
under realistic FTS conditions (H2O/H2 0.53). Co crystallites
with diameters larger than about 2 nm will not undergo oxida-
tion. In fact, from XANES analyses of spent Co catalysts from an
extended FTS run, it was found that a further reduction took
place during FTS118 and has been confirmed by others.135137
Further, the formation of metal-support compounds such
as cobalt aluminate have been considered as a deactivation
mechanism.129,138 Although thermodynamically favorable,
this reaction needs CoO formation as an intermediate, which
does not take place under realistic FTS conditions. Indeed, a
recent study showed that the minor cobalt aluminate formed
during FTS originates from unreduced CoO present in the fresh
catalyst and not from Co metal.65,122 This leaves sintering and
carbon deposition as the major contributors to Co FTS catalyst
deactivation.
Carbon deposition on an FTS catalyst that is covered by
growing hydrocarbons and is entirely embedded in product
wax represents a real challenge. Nevertheless, deleterious car-
bon arising from CO or FT products can have a wide range of
negative effects on Co FTS catalysts. We mention pore block-
age, resulting in mass transfer limitations, formation of bulk or
surface carbides, and blockage or alteration of active sites.121
Pore blockage by long hydrocarbon products resulting in dif-
fusion limitations of the reactants CO and H2 has been men-
tioned as a deactivation mechanism from the onset of the
discovery of the FTS.139 A hydrogen treatment or solvent
wash of the spent catalyst resulted in a partial recuperation of
FischerTropsch Synthesis: Catalysts and Chemistry 545
1.8
Hydrogen resistant carbon (wt%)
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0 20 40 60 80 100
Time on stream (days)
Figure 22 Build up of polymeric carbon on a cobalt slurry-phase
FischerTropsch synthesis catalyst as a function of time on stream.121
FT activity. Although detailed studies are unknown to us, pore
blocking is generally accepted as a deactivation mechanism.
Recently, strong evidence has been found for the accumula-
tion of a rather inactive polymeric carbon on the metallic surface
of cobalt resulting in catalyst deactivation (Figure 22).121 Tech-
niques involved in this work were temperature-programmed
hydrogenation, low-energy ion scattering, and energy-filtered
TEM. Molecular modeling suggested that the polymeric carbon
might be a form of graphene. In general, there is good agree-
ment that bulk cobalt carbide is metastable and will not be
present in substantial quantities during FTS for cobalt, although
Karaca et al.71 observed small quantities of Co2C in their in situ
XRD experiments (which, interestingly, was entirely absent for
the first 9 h on stream, but appeared in the measurements after
10 h; exposure to pure CO made the signal grow further).
Although characterization of spent cobalt FT catalysts run for
several months in a slurry bubble column did not show bulk
cobalt carbide formation,118,121 a possible role of subsurface
carbon cannot be ruled out. To date, our knowledge on subsur-
face carbon comes from molecular modeling. Calculations on
carbon clusters by Zonneville et al.140 indicate that carbon in
subsurface positions affects the CO dissociation rate and may
therefore affect the FTS activity as well. More work is needed to
ascertain the impact of subsurface carbon.
Carbon deposition can be decreased by adding promoters.
Examples from the literature include ruthenium74 and
boron.77,141,142
Sintering is a thermodynamically driven process whereby
smaller, more unstable particles grow to form larger, more
stable particles that are lower in surface energy. Sintering as a
deactivation mechanism is easier to investigate than carbon
deposition, with TEM and XRD being the most common tech-
niques used. Due to this in general there is good agreement in
the literature on the importance of sintering as a deactivation
mechanism of cobalt FTS catalysts.62,65,70,71,121,134,137,143145
Key factors that affect the rate of sintering are the reaction
temperature and the partial pressure of water: an increase in
either of these results in enhanced sintering. The choice of
support also plays a key role. Alumina is considered to provide
more stability against sintering than silica does, due to the
improved metal support interaction in the former. Both
546 FischerTropsch Synthesis: Catalysts and Chemistry
particle migration and Ostwald ripening are expected to be
important, but further studies are desirable to gain full under-
standing on the contribution of each.
Poisoning of cobalt catalysts by S, NH3, HCN, Hg, and Cl is
well known and is an issue for cobalt-based FTS62,146149 espe-
cially for coal-to-liquid applications.65 Sulfur is a strong, irre-
versible poison with a large adsorption energy. Due to its size
and electronic effect, sulfur will also poison adjacent cobalt
sites. Once adsorbed, sulfur is difficult to remove and will
accumulate with time.146 Sulfur poisoning can however rela-
tively easily be prevented by cleaning the synthesis gas feed
properly, for example, by using zinc-oxide or lead-oxide guard
beds. Poisoning of cobalt-based FTS catalysts by means of
nitrogen-containing compounds such as NH3 and HCN is a
known effect and postulated to arise from competitive
adsorption.150 The impact of N-compounds is less severe
than that of S-compounds and can be reversed by mild hydro-
gen treatment. Nevertheless, reducing their level to parts per
billion is recommended.150
Surface reconstruction is a thermodynamically driven pro-
cess which results in a lowering of the surface energy and
therefore can contribute to catalyst deactivation. Using molec-
ular modeling, Ciobica et al.124 showed that atomic carbon
from dissociated CO can cause a reconstruction of the Co fcc
(111) surface to a Co (100)-like structure, followed by a clock
reconstruction. This surface is less active and could therefore
contribute to deactivation. As the reconstruction is accompa-
nied by a change in surface density, it could, ironically, also
assist in the formation of more reactive sites, as proposed by
Wilson and de Groot.151 This is a complex phenomenon that
needs further investigation.
Methods to reverse deactivation and regenerate deactivated
Co FTS catalysts have been around since the early days of
Fischer and Trospch.3 The most common methods reported in
the open literature are treatment of the deactivated catalyst
in hydrogen or in steam, applying oxidationreduction cycles,
and combinations of these.65,152 With carbon and sinte-
ring being the major deactivation mechanisms during FTS,
oxidationreduction is considered to be the most robust and
preferred method to regenerate spent Co FT catalysts.65 By
careful control of the oxidation step, deleterious carbon is
removed at temperatures above 250  C. The oxidation step
is also key for the redispersion of cobalt (see Figure 23). The
20 nm
Figure 23 Co particles supported on a flat SiO2 support during different stages of an oxidationreduction cycle. Cobalt particles in the metallic
state before oxidation (left). Hollow oxide particles, formed upon oxidation (middle). Reduction of the hollow particles, which break up into several
smaller metallic particles (right).65,153
mechanism of redispersion of cobalt has been proposed to be a
two-step process, that is, (1) oxidation to form hollow spheres
by the Kirkendall effect and (2) multinucleation of Co3O4
during reduction to produce smaller crystallites.65,153
Poisons such as sulfur are removed by oxidation with steam
and air to sulfates, followed by washing them out. However,
phases originating from strong metal-support interaction are
very difficult to reverse and their formation should be
prevented.154
Substantial progress has been made toward understanding
deactivation and regeneration of Co FTS catalysts, but more spe-
cific knowledge, for example, on deactivation by carbonaceous
species and sintering mechanisms, is definitely necessary. With
currently available in situ characterization, synchrotron-based
techniques, and molecular modeling, we expect major advance-
ments in the coming years.
7.20.4 Mechanisms and Kinetics of FTS Over Iron
and Cobalt Catalysts
7.20.4.1 Introduction
The complicated nature of the FTS is among others reflected by
its complex product spectrum, consisting of methane, C2
olefins and paraffins (linear and branched), oxygenates
(mainly alcohols, but also aldehydes and ketones), and even
aromatics (at sufficiently high operating temperatures). Three
classes of mechanisms have been proposed, each assuming a
different monomer for chain propagation.
The carbide mechanism (Table 7) was first formulated by
Fischer and Tropsch in 1926, which proposes that CO dissoci-
ates before the carbon atom is partially hydrogenated to a CHx
species.2 These CHx species combine by the addition of one
monomer at a time to effect hydrocarbon chain growth. The
growing intermediate can then terminate in different ways
before leaving the catalyst surface, giving rise to an alkane, an
alkene, or an oxygenate. A number of variations have been
considered within the basic carbide mechanism, arising from,
for instance, whether CO dissociation occurs unassisted or via
interaction with hydrogen, and the number of hydrogen atoms
in the monomer (CH or CH2). The enol mechanism (Table 7),
proposed by Storch in the 1950s, assumes that CO is partially
hydrogenated to a CHOH species (oxymethylene) which then
20 nm
20 nm

Table 7 FischerTropsch Reaction Mechanisms
Carbide mechanism Enol mechanism
Initiation CO ! C O CO 2H ! CHOH
O 2H ! H2O
C xH ! CHx
Propagation R CHx ! RCHx RCOH CHOH ! RCCOH H2O
RCHx (2  x)H ! RCH2 RCCOH H ! RCH2COH
Termination RCH2 H ! RCH3 RCH2COH 4H ! RCH2CH3 H2O
RCH2CH2H ! RCH CH2
R CO ! oxygenates RCH2COH nH ! oxygenates
NB : R H, alkyl(CH3, CH3CH2, CH3CH2CH2, . . .).
acts as the monomer.3 Chain growth occurs via a condensation
reaction with water elimination. Again, different termination
routes determine what final product molecule is formed. In the
1970s, Pichler and Schultz introduced the CO insertion mech-
anism (Table 7), which assumes a similar initiation step to the
carbide mechanism.155 The difference resides in the way that
chain growth is proposed to occur, namely by direct CO inser-
tion into the growing intermediate followed by hydrogenation
to remove the oxygen atom.
Even though the FTS has been known since the 1920s, it is
evident from the foregoing that its mechanism is still a matter
of debate. In order to progress the understanding of the mech-
anism, a multidisciplinary approach is required. Subsequently,
three such disciplines (model surface science experiments,
density functional theory (DFT) calculations, and macrokinetic
studies) will be briefly discussed with particular emphasis on
their implications for mechanistic and kinetic understanding.
7.20.4.2 Surface Science Studies and Model Reactions
A surface science approach is very powerful to study elemen-
tary reaction steps in isolation. Conceptually, it is very close to
the approach taken by DFT calculations: take a well-defined
surface, that is, a single-crystal surface of the material you want
to study and use an ultrahigh vacuum so that the adsorbate of
interest can be introduced with high purity and with a high
accuracy, down to a sub-monolayer coverage. The model sys-
tem can then be studied with many different sophisticated
analysis techniques that give information on the atomic level.
The number of studies with cobalt and iron single-crystal
surfaces related to the FTS is relatively small. Iron carbide
single-crystal work is not available, as far as we know. Studies
on nickel and rhodium crystals are more numerous, as they are
easier to use and because of the fact that both metals are used
as a catalyst for a number of reactions. In this section, we
discuss the most relevant surface science findings on cobalt,
with a focus on studies where a single elementary step was
studied in isolation.
7.20.4.2.1 Adsorption of CO and hydrogen on model
surfaces
One of the simplest experiments one can do is to study the
interaction of a surface with CO and H2, the reactants in the FT
reaction. In a typical experiment, a clean close-packed Co
surface is exposed to increasing amounts of CO at a sample
FischerTropsch Synthesis: Catalysts and Chemistry 547
CO insertion mechanism
CO ! C O
O 2H ! H2OH
RCH CO ! RCHCO
RCHCO H ! RCH2CH H2O
RCH 2H ! RCH3
RCHCO nH ! oxygenates
temperature of 180 K.156 Such a low-temperature experiment
in ultrahigh vacuum (UHV) conditions is equivalent to incre-
asing the CO pressure in a room temperature experiment.157
Initially CO adsorbs, with a sticking coefficient of 0.7, on top
sites up to a coverage of 0.33 ML (monolayer), with an adsorp-
tion energy of 115 kJ mol1.156 This translates into a desorp-
tion temperature around 400 K. Upon further dosing, the CO
coverage increases, until a saturation coverage of  0.65 ML is
reached. Increasing the CO coverage beyond 0.33 ML leads to
complex overlayers where CO occupies bridge and threefold
sites as well as top sites. The downward shift of the CO desorp-
tion temperature for coverages beyond 0.33 ML is mainly
caused by repulsive interactions between CO molecules rather
than by the difference in adsorption site. This has important
implications for the interpretation of vibrational spectra on
supported catalyst particles, where occupation of both top
and bridge/threefold sites is typically detected. Occupation of
bridge and threefold sites can simply be caused by a high CO
coverage on the facets of the particle rather than by the pres-
ence of special sites.
Hydrogen/deuterium adsorption on a close-packed Co
surface was studied in a similar fashion: hydrogen adsorbs at
180 K with a low sticking coefficient, up to a coverage of  0.5
ML hydrogen atoms, with an adsorption energy of 33 kJ mol1
(per H atom). Recombinative desorption occurs between 300
and 400 K.158 Hydrogen desorbs at lower temperature from
more open surfaces, around 300 K, indicating a weaker adsorp-
tion onto those surfaces. The sticking coefficient on the other
hand is much higher than on close-packed surfaces: 0.76
for an open surface compared to 0.05 on a close-packed
surface.159,160 This enhanced hydrogen sticking is commonly
observed on the more open crystal planes of different metal
surfaces.
When CO is dosed at 180 K on hydrogen-covered surfaces,
CO partially replaces the hydrogen, and only 50% of the initial
coverage remains on the surface. The remaining hydrogen is
less strongly bound due to the CO, and as a result the desorp-
tion peak shifts downward by 100 K. CO, on the other hand,
is only mildly influenced by the presence of hydrogen, and any
influence of hydrogen is only seen at low temperatures.158
These experiments show that repulsive interactions exist
between hydrogen and CO, which adds to the barrier to form
hydrogenated HxCO species. Other adsorbates such as sulfur,
oxygen, and carbon give rise to a similar downward shift of the
hydrogen desorption temperature, and those species also
(partly) block the surface for hydrogen adsorption.158
548 FischerTropsch Synthesis: Catalysts and Chemistry
7.20.4.2.2 CO bond scission
A key step in any FT mechanism is the cleavage of the CO
bond. The close-packed surface of cobalt, Co (0001), is unable
to cleave the CO bond: CO desorbs as a molecule with the CO
156
bond
 intact. Some open surfaces of cobalt, such as Co
1012 161,162 and Co 1120 ,163 and cobalt foils are capable
of cleaving the CO molecule directly: after a CO thermal
desorption experiment161,162 or a prolonged exposure to CO
at elevated temperature,163 carbon and oxygen are found to be
left on the surface. CO dissociation stops when enough carbon
and oxygen has built up to block all the active sites for disso-
ciation. UHV studies on 2-nm Co particles on alumina show
that CO dissociates during a CO thermal desorption experi-
ment, which can be explained by the high defect density on
such small nanoparticles.164 These studies did not report exact
temperatures or activation barriers for CO dissociation, but in
all cases a typical reaction temperature in the order of 400 K
can be deduced. In short, surface science shows that direct CO
dissociation is possible, but not on the (most abundant) close-
packed surface.
As direct CO dissociation is not possible on the close-packed
surfaces of cobalt, one might consider if the CO molecule is
(partly) hydrogenated before the CO bond breaks. When CO
and hydrogen are co-adsorbed onto a close-packed surface at
low temperature, the molecules just desorb upon heating, with-
out reaction.158,160 An alternative experimental approach is to
study the decomposition of (partly) hydrogenated CO mole-
cules such as methanol and formaldehyde. Methanol adsorbs as
methoxy (H3CO) when dosed at 165 K. During heating this
methoxy species is stable up to 300 K, after which it decom-
poses to CO and H2.165,166 Experimentally, this is seen as a
single step, indicating that the first dehydrogenation is rate-
limiting. The intermediate species, H2CO and HCO, decompose
O1s O1s, TP-XPS
hv = 650 eV hv = 650 eV
OCH2CH3
Photo-emission intensity (a.u.)
250 K
Oad
370 K
534 532 530 528 532 530
Binding energy (eV)
Figure 24 Ethanol decomposition on a close-packed cobalt surface, followed with temperature programmed (TP) synchrotron x-ray photoelectron
spectroscopy (XPS). At low temperature, the ethoxy intermediate is found. During heating ethoxy decomposes around 350 K, yielding atomic oxygen
and C2Hx (acetylene) on the surface, showing clear evidence for CO bond cleavage. Adapted from Weststrate, C. J.; Gericke, H. J.; Verhoeven, M.;
Ciobica, I. M.; Saib, A. M.; Niemantsverdriet, J. W. J. Phys. Chem. Lett. 2010, 1, 17671770.
rapidly and are not observed in significant concentrations on
the surface. This is in line with the finding that formaldehyde
(H2CO) decomposes with 100% selectivity to CO and H2
between 100 and 200 K.160 The experiments show that partially
hydrogenated species such as HCO and H2CO are very unstable,
and the barrier to decompose via dehydrogenation is obviously
much smaller than that of (Hx)CO bond cleavage. Similar
experiments have been reported on 2-nm cobalt particles sup-
ported on alumina. In those experiments, 60% of the methanol
that was present dehydrogenated, while 40% underwent CO
bond cleavage.164 In this experiment, it was not clear whether
the CO bond scission was assisted by the presence of hydrogen,
as CO (produced by methanol dehydrogenation in the metha-
nol experiment) also dissociates on those particles in the
absence of hydrogen.
In the CO insertion mechanism (Table 7), a CO molecule is
inserted into a CxHy, after which the CO bond has to be
cleaved to generate a Cx1Hz intermediate that can insert
another CO molecule. In other words, the CO molecule is
chemically modified by insertion of an alkyl group on the
C-end of the molecule. Experiments using ethanol on a close-
packed Co surface gave information about the effect of alkyl
modification on the CO bond cleavage.166 Figure 24 shows
the result of such an experiment: ethanol adsorbs as an ethoxy
species at 160 K. This ethoxy species decomposes around
350 K, via an acetaldehyde intermediate, with an activation
barrier of 70 kJ mol1. The products of this dissociation
step are atomic O and a C2Hx species, demonstrating CO
bond cleavage. This means that alkyl insertion in the CO
molecule facilitates CO bond scission. For the CO insertion
mechanism, it implies that after the CO is inserted the CO
bond can be readily broken and a Cx1Hy species is formed that
can undergo further chain growth.
C1s, TP-XPS C1s (high res.)
250 K
hv = 380 eV hv = 321 eV
OCH2CH3
Photo-emission intensity (a.u.)
300 K
250 K
350 K
C2H2
370 K
528 286 285 284 283 287 286 285 284 283 282
Binding energy Binding energy (eV)

7.20.4.2.3 Hydrogenation and the stability of C1Hx species
C1Hx species are important ingredients in a carbide mechanism
as they are the monomeric species responsible for chain initi-
ation and growth. Surface science experiments can give infor-
mation about the stability of the different C1Hx species.
Regarding cobalt there is only one article that addresses this
question directly, using CHxCly intermediates on a cobalt
foil.167 The authors mention that the behavior on nickel foil
was essentially the same. In an experiment on a close-packed
nickel surface, a molecular beam was used to dissociate meth-
ane at low surface temperature, which generates CH3 species
(H) on the surface.168 Heating of this CH3 layer showed
transformation of CH3 to CH around 200 K, and no sign of
CH2 was found. On close-packed Pt reported in the same
study, a very similar trend was seen: CH3 decomposes around
250 K, yielding solely CH, which decomposes around
500 K.168 These surface science results demonstrate that the
CH2 species, which is typically seen as the monomer for
chain growth, is particularly unstable, which means that its
concentration under equilibrium concentrations will be
much lower than that of CH3 and CH.
Generally speaking, surface science studies on cobalt and
iron are scarce in comparison to those on nickel and the more
noble metals. Studies on iron foils and single crystals,169176
although very interesting from the point of view of surface
chemistry, are even further removed from the reality of
FischerTropsch reactions than cobalt, as iron FTS catalysts
are essentially carbides. Surface science studies on iron carbides
are, to the best of our knowledge, not available.
7.20.4.3 DFT Modeling
Molecular modeling by DFT offers a relatively new way to
understand reaction mechanisms at the molecular level.177,178
Adsorption configurations along with their energetics as well as
transition states can be modeled, and thus adsorption energies,
activation energies, and heats of reactions can be obtained. In
addition, entropy changes over elementary reactions enable one
Transition state
Adsorbed CO
1.14 eV
(exp 110 kJ mol-1)
2.30 eV
-0.34 eV
-1.16 eV
Extent of reaction
Figure 25 Energy diagram for the dissociation of carbon monoxide on the (100) surface of iron, showing the exothermicity of the dissociation, and the
effect of repulsion between carbon and oxygen atoms that are adsorbed onto adjacent sites. Energies are given in electron volt (eV);
1 eV  96.5 kJ mol1). Adapted from Bromfield, T. C.; Ferre, D. C.; Niemantsverdriet, J. W. ChemPhysChem 2005, 6, 254260.
FischerTropsch Synthesis: Catalysts and Chemistry 549
to estimate pre-exponential factors, although these are generally
less often calculated than enthalpies.
Validation of calculations has to be sought by comparing
calculated adsorption energies and activation barriers or vibra-
tional frequencies from stable adsorption states with experi-
mental values, obtained from surface science experiments with
single crystals. However, relatively little experimental data are
available, partly also because surface science studies are by
necessity often performed in vacuum, whereas FTS reaction
steps occur at higher pressures.
As this is an area of research that is emerging rapidly, we
intend to present some examples of how DFT modeling is used
in mechanistic studies. Although DFT results refer to tempera-
tures of zero K and pressure, the results give valuable insight
into the energetic of the underlying surface chemistry. It is not
our intention to give a full review here, as it is still too early for
conclusive statements on the FTS reaction mechanism.
7.20.4.3.1 CO Dissociation
Conversion of CO and H2 into CxHy H2O necessarily implies
that the CO bond has to broken. The question is now if this
happens before or after reaction with H-atoms. As an example, we
show a computational study of direct CO dissociation on the
square (100) surface of bcc-iron in Figure 25.179,180 On this
surface, the CO molecule is known to adsorb in a tilted
geometry.171 In the transition state for dissociation, the CO
bond elongates, and the energy rises by 1.14 eV (109 kJ mol1),
which is the activation energy for dissociation. This value com-
pares well with the experimentally measured activation energy of
110 kJ mol1 reported by Bernasek and coworkers.171 When the
bond breaks, the C and O atoms each end up in a fourfold hollow
site of the Fe(100) surface. However, the two atoms significantly
repel each other, implying that the total energy decreases substan-
tially when the two atoms move apart, as shown in the last
structure of Figure 25. This series of calculations demonstrates
that the direct dissociation of a CO molecule on this (100) surface
is very well possible under reaction conditions (a barrier of
110 kJ mol1 corresponds roughly to a reaction temperature of
Dissociated CO
Dissociated CO
repulsion relieved
0.82 eV
550 FischerTropsch Synthesis: Catalysts and Chemistry

400450 K, whereas the FTS reaction temperature is at least
475 K), provided the space is available for the dissociation
products to move apart. Hence, this elementary step needs a so-
called ensemble of iron atoms, which is determined by the
condition that the C and the O atom can end up on next nearest
neighbor sites, that is, at distances at least equal to 2 times the
lattice constant.
Table 8 compares the activation energy for dissociation on a
number of iron surfaces with increasing reactivity.181,182 The
trend is that dissociation becomes easier when the surface be-
comes more reactive. Note that the (110) surface is unlikely to play
a role in CO activation, as the barrier is forbiddingly high.181183
Several authors have proposed that CO dissociation becomes
easier when the CO first reacts with hydrogen.183,185,186 The
reasoning is evident, as the CO bond in an HCO or COH
fragment is expected to be weaker than in the CO molecule.
However, forming the HCO or COH fragment also costs
energy.186 It appears that on surfaces of low reactivity, such as
Co(0001), Fe(110), or the (100) surface of the Hagg carbide, H-
assisted dissociation indeed leads to a lower activation barrier
than direct CO dissociation does.183,185 On more reactive sur-
faces, and notably on surfaces containing steps, the direct
dissociation is favored, and often with a lower barrier than for
the reaction H CO.186 Nevertheless, it is good to realize that
on certain surfaces the H-assisted pathway may be an alternative
that gives less reactive facets the chance to play a role in FTS.
7.20.4.3.2 C H reactions
Many studies have addressed the formation of CHx fragments
all the way to methane on several surfaces. Particularly inter-
esting is the comparison made by the Nrskov group.187 Their
calculations are based on a full set of DFT calculations on the
fcc (211) step of ruthenium surfaces, from which they esti-
mated the adsorbate energies on several other transition and
group IB surfaces. According to the Sabatier Principle,188 the
optimum pathway for a reaction is that in which the interme-
diates adsorb at the catalyst surface in a moderate way, that is,
not too strongly and not too weakly. The set of profiles in
Figure 26 illustrates that metals such as Ni, Rh, and Co are
close to ideal methanation catalysts, but that (the stepped
surfaces of) metals such as Ru, Fe, and W bind the intermediate
species too strongly. On the other extreme are Au and Ag,
where formation of intermediates is strongly endothermic
and therefore unfeasible. Metals such as Cu, Pt, and Pd are

Table 8 Activation energies for direct CO dissociation

Iron Characteristics CO dissociation activation Reference

surface energy

(110) Flat, close packed, least reactive surface 149 kJ mol1 Sorescu181
(100) Flat, somewhat more open, and thus more 103110 kJ mol1 Bromfield et al.,179 Scheijen et al.,184 and Sorescu181
reactive than (110)
(310) Stepped surface with narrow terraces 7087 kJ mol1 Sorescu181
(710) Stepped surface with broader terraces 6486 kJ mol1 Sorescu181

5 Ag

Au

Cu
Energy (eV)

0 CO + 3H2
Pd, Pt

CH4 + H2O
Ni, Rh
Co
-5
Ru
Fe

W
)

O*

O*

H*

2*

3*

4 (g)

4 (g)

4 (g)
O(g

+C

H
+C
+C

*+

+C

+C

CH
+C

H*

+C

+C
+C

H*
2 (g)

+6

H*

H*

g) +
2 (g)

+5

H*

H*
2 (g)

+4

+3
3H

O*

2 O(
+2
3H

*+
O*
3H

O*

O*

H
OH
O*

Figure 26 Density functional calculation for the energy profile of the methanation reaction on the group VIII and I-B metals, showing that nickel,
rhodium, and cobalt fall close to the optimum profile for the reaction, while iron and tungsten form too strong bonds with the carbon, oxygen, and the
CHx intermediates. Adapted from Jones, G.; Bligaard, T.; Abild-Pedersen, F.; Norskov, J. K. J. Phys. Condens. Matter 2008, 20.

incapable of breaking the CO bond, but can carry out CO
hydrogenation to methanol.
Not shown in the figure are the barriers for the individual
steps. As shown by several authors, activation energies for
hydrogenation of adsorbed CHx (x 03) species on uncorru-
gated (atomically flat) surfaces are in the range of 50
100 kJ mol1 and therefore not expected to be problematic
for the reaction mechanism.189194
7.20.4.3.3 Chain growth
Next comes the question how the hydrocarbon chains grow.
This has been and still is a matter of scientific debate. The
conventional view is that chains grow by a reaction between
an alkyl and a CH2 species, for example, CH3 CH2 to C2H5,
and C2H5 CH2 to C3H7. Termination to an alkane would
then occur by hydrogenation of the alkyl, and olefins would
form by b-H abstraction from the alkyl fragment. This view
originates from the work of Biloen et al., who, however, pro-
posed this mechanism with some reservation, and with a judi-
cious discussion of the assumptions involved.195,196 Ciobica
et al. found that reactions between CxHy fragments that contain
less hydrogen are energetically more favorable, making chain
growth via reactions of the type CH2 CH and CHCH2 CH
more likely.197 Nevertheless, the class of mechanisms, based
on CO dissociation, formation of CHx species, and incorpora-
tion of CHx in a growing chain, albeit with a rich variety in the
details, comes close to the original proposal of Fischer and
Tropsch, entitled the carbide mechanism (see Table 7).
Another class of mechanisms considers CO insertion as the
step leading to chain growth (see Table 7). This line of thought
also has a long history, the archetype being the PichlerSchulz
mechanism.155 The most detailed mechanism has been pre-
sented by Saeys and coworkers,198 who proposed that CO
inserts in an adsorbed CH2, which then further hydrogenates
to a fragment in which the CO bond breaks, leading to a C2
intermediate in which the next CO inserts. As discussed in the
section on surface science, experimental proof exists that CO
bond breaking in adsorbed species derived from ethanol, such
as ethoxy, is a facile step, even on the least reactive surface of
cobalt, that is, the close-packed (0001) surface.166
At the time of writing, medio 2011, the authors believed
that DFT is a highly valuable tool for getting insight into
reaction mechanisms at the level of elementary steps. It is,
however, much too early to draw definitive conclusions on
which mechanism is prevalent in the FTS. Further, it should
be acknowledged that not all catalysts and conditions can be
captured under one dominating mechanism, and it is even not
at all certain that this would be the case on one catalyst. For
example, the initiation by CO dissociation might simulta-
neously occur via direct dissociation on highly reactive parts
of a cobalt particle, and by H-assistance on facets of moderate
reactivity. We also point to the differences between cobalt and
iron. Whereas the former is believed to operate as a metal, the
latter is active as a carbide, in which the intrinsically high
reactivity of the iron atoms is considerably decreased by carbon
neighbors. The DFT literature has suggested mechanisms vary-
ing from Mars-van Krevelen-type reactions199 (in which carbon
atoms from the lattice become the CHx species for initiation
and chain growth) to H-assisted CO dissociation and
CO insertion as the major ingredients for hydrocarbon
FischerTropsch Synthesis: Catalysts and Chemistry 551
formation.200 It is obvious that mechanistic understanding
would greatly benefit from more surface science work, but
unfortunately the possibilities for this branch of physical
chemistry research are limited for reactions that by necessity
require high pressure conditions.
7.20.4.4 Macrokinetic Observations and Models
In order to derive a macrokinetic model, a scheme of elemen-
tary reaction steps is usually assumed to represent the
mechanistic pathway of the reaction. Normally, a number of
simplifying assumptions are made regarding the catalyst sur-
face and the adsorption of species on it, for example, that only
one type of adsorption site is considered, and that these sites
are homogeneously distributed over the catalyst surface, that
species only adsorb in a monolayer, and that adsorbed species
do not interact apart from the involved chemical reactions. It is
usually further assumed that most reaction steps are suffi-
ciently fast to reach equilibrium, but that one or more rate-
determining steps exist that are relatively slow and control the
overall rate of reaction.188 These assumptions allow for the
derivation of simple, manageable kinetic expressions, such as
those presented in Table 9.
7.20.4.4.1 General observations regarding kinetics
Macrokinetic models are mainly developed for use in process
modeling, yet their mechanistic importance stems from the
fact that they capture the overall behavior of the FT synthesis.
Considering eqn [3] from Table 9 as an example, it is seen that
the reaction rate is predicted to increase with the square root of
the hydrogen partial pressure (everything else being constant).
Therefore, it is said that the reaction order of hydrogen is 0.5 in
this kinetic model. The variation in reaction rate with CO
partial pressure is more complex, since the overall reaction
order is not constant. At very low CO partial pressures or at
high reaction temperature, KCOPCO is much smaller than 1 and
the denominator term can be ignored, yielding an overall CO
reaction order of 1. In the other extreme, at very high CO
partial pressures or at low temperatures where KCOPCO is much
larger than 1, the constant term in the denominator can be
ignored and the overall CO reaction order strives to 1.
Table 9 Kinetic expressions for the rate of the FischerTropsch
synthesis, rFT
Equation number Rate expression
PH PCO
1 rFT A PCO K2H P
2 O H2 O
PH0:5 PCO
2 rFT A PCO K2H P
2 O H2 O
PH0:5 PCO
3 rFT A 2
1KCO PCO 2
PH2 PCO
4 rFT A
1KCO PCO 2
PH0:75 PCO
0:5
5 rFT A  2
2
0:5 P0:25 K P0:5 P0:25
1KC=OH PCO H O CO H
PH0:75 PCO
0:5
2 2
6 rFT A 2
1KCO PCO
0:5 2
K stands for equilibrium constant, P for partial pressure, and A is an effective rate
constant.
552 FischerTropsch Synthesis: Catalysts and Chemistry
Similarly, it can be said for eqn [1] that the reaction order in
hydrogen is constant at 1, while the overall reaction order in
CO varies from 0 to 1 as the conversion increases from low to
high values.
There are distinct differences between the chemical reaction
kinetics over iron and cobalt FTS catalysts. The reaction rate
increases roughly linearly with pressure over iron catalysts at
constant temperature and H2/CO ratio,201 but only to an order
of around 0.5 for cobalt catalysts.202 Furthermore, CO has a
strong inhibiting influence over cobalt catalysts to the extent
that it has a significantly negative reaction order under com-
mercially relevant reaction conditions. To the contrary, CO has
a positive effect on the rate over iron catalysts in the normal
operating regime and it has been estimated that CO will only
start to negatively influence the kinetics above a partial pres-
sure of around 11 bar.203
The influence of water on the reaction kinetics has proven
to be a highly controversial topic. Historically, it was firmly
believed that water (and possibly CO2) inhibited the reaction
rate over iron catalysts via competitive adsorption, but more
recently it has been shown that there is no convincing evidence
for this notion.203 The results of water co-feeding studies over
cobalt catalysts have been inconsistent, since some have
reported a positive and some a negative influence of water on
the reaction rate, while others have found no effect at all. At
least for alumina-supported cobalt catalysts, it appears as
though water in the range of 16 bar has no significant influ-
ence on the overall rate of CO conversion, but that it does
decrease the methane selectivity.204
7.20.4.4.2 Simple macrokinetic models
For the derivation of macrokinetic models, a key question is
whether or not there is a rate-determining step involved in the
formation of the monomer of chain growth. If monomer
formation is relatively facile, then the rate of CO conversion
and the product distribution obtained are intimately linked
and must be modeled together. This typically yields a complex,
implicit type model that requires an advanced numerical rou-
tine to solve. However, in order to keep these models manage-
able, questionable simplifying assumptions are often made, for
example, the assumption by Yang et al.205 that the monomer is
in thermodynamic equilibrium with the gas phase concentra-
tions of CO, H2, and water. Furthermore, such models require
a large number of parameters that are inevitably highly cross-
correlated, implying that it is virtually impossible to accurately
estimate their values. If, however, there is a rate-determining
step in the formation of the monomer, the overall reaction rate
can be decoupled from the product distribution. In such a case,
the overall rate of CO conversion is determined by the CO
hydrogenation reaction (monomer formation), while the
product distribution is determined by the polymerization
part of the FT reaction (i.e., the competition between chain
growth and desorption). The steady state isotopic transient
kinetic analysis study of van Dijk206 has indeed provided
microkinetic support for the notion that there is a rate-
determining step in the formation of the monomer.
The approach followed during FTS kinetic studies by Botes
et al.202,207 has been to consider different reaction schemes and
rate-determining steps in the formation of the monomer to
obtain a variety of explicit rate expressions. A systematic
experimental approach was then followed to eliminate non-
applicable models until ultimately a robust kinetic expression
remained as the preferred rate equation. A further notable
feature of the FTS kinetic studies was to operate the reaction
at a baseline condition, where the catalyst is known to be quite
stable, for most of the run. Changes to other conditions were
only made for short intervals sufficient to allow for hydrody-
namic steady state inside the reactor, but insufficient to effect
changes in the intrinsic catalyst behavior. The importance of
this approach is related to the fact that FT catalysts, especially
those based on iron, readily respond to changes in operating
conditions. It is imperative to avoid reversible and irreversible
changes in the catalyst when chemical reaction kinetics is
studied.
Originally within Sasol, the rate equation [1] in Table 9, by
Anderson201 was used to describe the FTS over iron. Following
a systematic in-house study focusing on the reaction order of
hydrogen, its exponent was later reduced to a value of 0.5,
yielding eqn [2] (Table 9). The most recent study on iron-FT
kinetics has considered the implications of the historic rate
equations, specifically the single order denominator which
implies that hydrogen reacts directly from the gas phase. By
applying a second order denominator to obtain a more appro-
priate LangmuirHinshelwoodHougenWatson-type equa-
tion where both CO and H2 first absorb onto the surface
before reaction, and also including a constant term in the
denominator to provide for the possibility of vacant sites, it
could in fact be shown that there is no statistical justification
for including a water term in the kinetic model.207 Experiments
were designed to conclusively show that eqn [3] (Table 9) is
more accurate than the foregoing expressions. This study
highlighted the fact that the historic perception regarding the
effect of water on iron-FT kinetics was self-specified by the old
models, but never tested. Furthermore, the CO order of unity is
consistent with a mechanism where CO interacts with a hydro-
gen atom before being dissociated.
The most recent cobalt kinetic study involved the derivation
of several rate equations to cover various reaction schemes of
CO hydrogenation.202 Models assuming hydrogen-assisted CO
dissociation, such as eqn [4] that was originally proposed by
Yates and Satterfield,208 generally described the measured data
poorly and could be eliminated early on in the study. Figure 27
illustrates that eqn [4] underestimates the reaction rate at low
CO partial pressure, but overestimates it at higher CO partial
pressures. Ultimately, after further work to distinguish between
those models where CO first dissociates before it is hydroge-
nated, eqn [5] (Table 9) was the only rate expression that could
not be eliminated. Therefore, it was selected as the most appro-
priate kinetic model. It should be noted that this expression can
be very closely approximated by eqn [6], which contains one
model parameter less and is thus preferably used from a practi-
cal perspective. Figure 27 shows that the preferred equation [6]
does not suffer from the same systematic errors as eqn [4], since
it is reasonably accurate across a range of CO partial pressures.
7.20.4.4.3 Selectivity modeling
The vast number of components in the FT product slate does
not allow for the prediction of individual product selectivities;
consequently, the product spectrum is rather represented by a
product characterization model with a limited number of

1.5
rFT (measured) / rFT (predicted)
1
0.5
0 1
Figure 27 Performance of rival kinetic expressions as a function of CO partial pressure for a data series where the CO flow rate into a slurry reactor was
varied while that of H2 was maintained constant.
parameters. After correlating these parameters with process
conditions, an explicit selectivity model can be obtained. The
simplest model for describing selectivity is the Anderson
SchulzFlory distribution, with the chain growth probability
(a-value) as the only parameter. Approaches have been pro-
posed to account for deviations from the ideal distribution.
The double-a model by Donnelly et al.209 assumes that two
types of catalytic sites or two types of mechanisms simulta-
neously form the observed product spectrum. However, the
three model parameters have a high degree of covariance
when estimated from experimental data. Furthermore, neither
the C1 and C2 selectivities, nor the olefin content of the prod-
uct spectrum can be predicted. Some of these limitations have
been addressed by considering a chain length-dependent
desorption model, which assumes that termination by desorp-
tion becomes increasingly more difficult as the chain length
increases.17 It has also been reported that the chain length
effects in the FT product spectrum, in particular the positive
deviation of methane and the negative deviation of ethylene,
can be explained by symmetry effects as accounted for by the
single event kinetic theory.210 Some have ascribed the chain
length-dependent deviations to secondary olefin reactions, but
many concerns remain over this approach.203 These include
the observation that secondary olefin reactions are much less
facile (almost negligible) over iron catalysts compared to
cobalt catalysts, yet the bend in the AndersonSchulzFlory
graph is much more pronounced with iron catalysts.
7.20.4.5 Mechanistic and Kinetic Implications
Despite the large number of kinetic and mechanistic studies on
FTS, there is still substantial uncertainty regarding the most
relevant steps in the reaction pathway(s). A variety of elementary
reaction steps have been proven to be realistically possible under
typical FTS conditions, while very few steps could be eliminated
FischerTropsch Synthesis: Catalysts and Chemistry 553
Equation 4
PH2PCO
rFT = A
(1 + KCOPCO)2
Equation 6
P 0H.75P 0.5
CO
2
rFT = A
(1 + KCOP 0.5
CO
)2
2 3 4 5 6 7 8
CO partial pressure (bar)
with certainty. For example, and as described before, DFT calcu-
lations have shown that the coupling reactions of CHx fractions
are reasonably facile, which lends support to the steps of chain
growth as proposed by the carbide mechanism. However, DFT
calculations have also shown that each of the steps required for
the CO insertion mechanism is energetically feasible, while the
surface science approach has demonstrated experimentally that
the scission of the CO bond of a CO molecule into which an
alkyl group has been inserted is a facile reaction, even on a low
reactivity cobalt surface.166
The current inability to clearly discriminate between rival
mechanisms partly stems from the overlap between the pro-
posed reaction pathways. For instance, the two most popular
mechanisms are not mutually exclusive with respect to each
other. The carbide mechanism has to assume direct CO inser-
tion as a termination step at least in order to explain the
observed formation of oxygenates in the FTS. On the other
hand, the initiation step in the CO insertion mechanism is
similar as for the carbide mechanism. The most plausible con-
clusion currently is therefore that a variety of reaction pathways
simultaneously contribute to the overall synthesis. The question
still remains though whether one pathway is dominant over the
rest and individually determines the bulk of the observed behav-
ior of the system, or whether two or more parallel pathways
have similar contributions to the overall kinetics. The answer to
this question may not even be absolute, as it may depend on
what aspect of the reaction is of interest. Hypothetically speak-
ing, if the carbide mechanism predominates in the formation of
olefins and paraffins (the main products of the synthesis), it
may well be accurate to describe the overall rate of syngas
conversion by only considering this mechanism. However,
even in such an event, CO insertion cannot be ignored if the
object is to model oxygenate formation.
Despite all the foregoing uncertainty, some consistencies
have also emerged from the studies performed in the various
554 FischerTropsch Synthesis: Catalysts and Chemistry
disciplines. For example, the results of DFT calculations suggest
that unassisted CO dissociation readily occurs on the more
open (high reactivity) metal surfaces, while hydrogen-assisted
CO dissociation would be required on the close packed (low
reactivity) surfaces.179,181,183,185,186 As described in the section
on surface science above, unassisted CO dissociation is not
easy on close packed surfaces, while it is facile over open
surfaces. It has further been found that HxCO species are
more inclined to dehydrogenate than to undergo CO bond
cleavage on cobalt surfaces. During a macrokinetic study on an
actual cobalt-FT catalyst, models assuming hydrogen-assisted
CO dissociation failed comprehensively, while the preferred
model based on unassisted CO dissociation could describe the
experimental data over a range of commercially relevant con-
ditions. Together all these findings suggest that, in the case of
the cobalt-based FT synthesis, the main pathway in the conver-
sion of CO to a CHx species proceeds via unassisted cleavage of
the CO bond.
To the contrary, it is known that the carbiding of iron cata-
lysts substantially decreases the reactivity of iron surfaces. There-
fore, one may well expect hydrogen-assisted CO dissociation to
predominate over iron-FT catalysts,199 which are in the carbided
state under actual synthesis conditions. The most preferred
macrokinetic model for iron is indeed consistent with a CO
dissociation step that occurs via interaction with hydrogen.
A further consistency that is steadily emerging relates to the
most likely nature of the species responsible for chain propaga-
tion in terms of the carbide mechanism. Originally, it was
believed (not necessarily based on strong evidence) that these
species are quite saturated with hydrogen, that is, that the grow-
ing intermediate is a CH3CH2  CH2 species, while the mono-
mer being added is a CH2 species.195 DFT calculations have
shown that reactions between intermediates that are leaner in
hydrogen (e.g., CHCH2 and CH) are energetically more favor-
able than reactions between more hydrogen-saturated
species.197 In line with this, it has been found during surface
science experiments the coupling of two CH species to from
acetylene is facile over nickel catalysts.168,211Further support is
provided by the steady state isotopic transient kinetic analysis
(SSITKA) study of Govender,212 who concluded from HD
switching experiments that the C2H species is the only abundant
C2 intermediate on the fully carbided, working iron-FT catalyst
surface. Therefore, even though the carbide mechanism as a
whole cannot be discarded as a prominent reaction pathway
for the FTS, it seems unlikely that it proceeds in the form that
was originally proposed. This has particular significance for the
termination toward olefins, since it implies that a hydrogen
abstraction is not required (possibly even a hydrogen addition).
7.20.5 Conclusion
The FTS represents proven technology, which has secured its
position in modern energy technology. Originally used to
convert coal into liquid fuels, nowadays the emphasis is on
monetizing natural gas, by converting it to diesel fuel, waxes,
and naphtha. It is expected that some 500 000 barrels of fuel
per day will be produced using FischerTropsch technology by
2013. Although small in comparison to the 85 million barrels
of crude oil that are produced daily, the 0.5 million daily
barrels of synfuels is undoubtedly significant, particularly
locally where the production takes place.
The technology has much potential for wider use, for exam-
ple, in emerging economies, or at a smaller scale in the utili-
zation of biomass. Interesting applications of FTS have been
proposed for conversion of remote natural gas at off-shore oil
production locations.
Both GTL and BTL can be important tools in strategies
aimed at reduction of CO2 emissions. CTL technology is clearly
disadvantaged here, and will in the future have to be combined
with CO2 sequestration technology. The rapid increase in dis-
coveries of shale gas (in, for example, USA and Canada) can
also provide the GTL industry with a significant boost.
Although proven technology, FTS continues to pose chal-
lenges from an industrial perspective. Stability improvement of
the catalysts is an important aspect, but also selectivity
improvement would be very advantageous. Economically one
would like to have the highest possible C 5 and the lowest
possible CH4 selectivity, because recycling of CH4 means that
the carbon atoms involved have to go through the expensive
syngas generation more than once, with the associated effi-
ciency losses. Syngas production is the most expensive part of
a GTL plant; it accounts for 4060% of the capital investments.
Increased research efforts on reducing the costs of syngas pro-
duction will make XTL projects even more viable. Although
new XTL facilities require large capital investments and are
dependent on the price ratio of crude oil to natural gas, in
the long term they are expected to be economically successful.
From a more academic perspective, understanding the
mechanism of the FTS has been and will be a challenge. It is
more and more realized that mechanisms may differ with
conditions and catalysts. It is highly unlikely that one unique
mechanism can account for all different forms of FTS. Molec-
ular modeling represents a very important tool for getting
mechanistic insight, but the problem is that experimental val-
idation of its predictions at the level of elementary steps is very
difficult to achieve, as the opportunities for relevant surface
science experiments are limited. Mechanistic studies aimed at
describing FTS selectivity from first principles are in their
infancy and have a long way to go before accurate predictions
can be expected.
Describing the physical/chemical state of the catalysts
under reaction conditions is another field where significant
progress has been booked, but major advances would still be
very welcome. The advent of in situ imaging tools in combina-
tion with realistic catalysts,213 as well as the use of planar
model catalysts in simulated environments,153 has proven
promising and will almost certainly lead to improved insight
in the relation between catalyst properties on the nanoscale
and performance in the reaction.
The FTS is therefore expected to remain an inspiring source
of industrial and academic research for many years to come.
For a related chapter in this Comprehensive, we refer to
Chapter 7.01
References
1. Fischer, F.; Tropsch, H. Brennst. Chem. 1923, 4, 276285.
2. Fischer, F.; Tropsch, H. Brennst. Chem. 1926, 7, 97104.

3. Storch, H.; Golumbic, N.; Anderson, R. B. The FischerTropsch and Related
Syntheses. Wiley: New York, 1951.
4. van der Laan, G. P.; Beenackers, A. A. C. M. Catal. Rev. 1999, 41, 255318.
5. Kolbel, H.; Ralek, M. Catal. Rev. 1980, 21, 225274.
6. Dry, M. E.; Hoogendoorn, J. C. Catal. Rev. 1981, 23, 265278.
7. Roferdepoorter, C. K. Chem. Rev. 1981, 81, 447474.
8. Iglesia, E., Reyes, S. C., Madon, R. J., Soled, S. L., Eds.; Advances in Catalysis;
1993; Vol. 39, pp 221302.
9. Steynberg, A. P.; Dry, M. E. FischerTropsch Technology; Elsevier: Amsterdam,
2004; Vol. 152.
10. Rostrup-Nielsen, J. R., Sehested, J., Norskov, J. K., Eds.; Advances in Catalysis;
2002; Vol. 47, pp 65139.
11. Dancuart, L. P.; de Haan, R.; de Klerk, A. In Studies in Surface Science and
Catalysis; Steynberg, A. P., Dry, M. E., Eds.; 152, Elsevier, 2004; pp 482532.
12. De Klerk, A. FischerTropsch Refining. Wiley-VCH: Weinheim, 2011.
13. Claeys, M.; van Steen, E. FischerTropsch Technology; Elsevier: Amsterdam,
2004; Vol. 152.
14. Schulz, H.; Vansteen, E.; Claeys, M. In Natural Gas Conversion II 1994; Vol. 81,
pp 455460.
15. Claeys, M.; Van Steen, E. In FischerTropsch Technology; Studies in Surface
Science: An Catalysis; Steynberg, A. P., Dry, M. E., Eds.; Elsevier: Amsterdam,
2004; Vol. 152, Chapter 8.
16. Botes, F. G.; Govender, N. S. Energy Fuel 2007, 21, 30953101.
17. Botes, F. G. Energy Fuel 2007, 21, 13791389.
18. Davis, B. H. Ind. Eng. Chem. Res. 2007, 46, 89388945.
19. Khodakov, A. Y.; Chu, W.; Fongarland, P. Chem. Rev. 2007, 107, 16921744.
20. Dry, M. E. Catal. Lett. 1991, 7, 241251.
21. de Smit, E.; Weckhuysen, B. M. Chem. Soc. Rev. 2008, 37, 27582781.
22. Steynberg, A. P.; Dry, M. E.; Davis, M. E.; Davis, B. H.; Breman, B. B. Stud. Surf.
Sci. Catal. 2004, 152, 64195.
23. Van Berge, P. J.; Van De Loosdrecht, J.; Caricato, E. A.; Barradas, S. Process for
Producing Hydrocarbons from a Synthesis Gas, and Catalysts Therefore. Patent
WO 9942214A1, 1999.
24. Davis, B. H. Catal. Today 2002, 71, 249300.
25. Steynberg, A. P.; Espinoza, R. L.; Jager, B.; Vosloo, A. C. Appl. Catal. A: Gen.
1999, 186, 4154.
26. Espinoza, R. L.; Steynberg, A. P.; Jager, B.; Vosloo, A. C. Appl. Catal. A: Gen.
1999, 186, 1326.
27. Geerlings, J. J. C.; Wilson, J. H.; Kramer, G. J.; Kuipers, H.; Hoek, A.;
Huisman, H. M. Appl. Catal. A: Gen. 1999, 186, 2740.
28. Deshmukh, S. R.; Tonkovich, A. L. Y.; McDaniel, J. S.; Schrader, L. D.;
Burton, C. D.; Jarosch, K. T.; Simpson, A. M.; Kilanowski, D. R.; LeViness, S.
Biofuels 2011, 2, 315324.
29. Visconti, C. G.; Tronconi, E.; Lietti, L.; Groppi, G.; Forzatti, P.; Cristiani, C.;
Zennaro, R.; Rossini, S. Appl. Catal. A: Gen. 2009, 370, 93101.
30. Sabatier, P.; Senderens, J. B. Hebd. Seances Acad. Sci. 1902, 134, 514.
31. Fischer, F.; Tropsch, H. Patent DE 484 337, 1925.
32. Liu, Z.; Shi, S.; Li, Y. Chem. Eng. Sci. 2010, 65, 1217.
33. Dry, M. E. Catal. Today 2002, 71, 227241.
34. Sie, S. T. Rev. Chem. Eng. 1998, 14, 109157.
35. Oukaci, R.; Singleton, A. H.; Goodwin, J. G. Appl. Catal. A: Gen. 1999, 186,
129144.
36. Sasol Financial Report JulyDecember 2010. http://www.sasol.com.
37. Brown, A. Pearl GTL Presentation, XTL Summit, June 2011.
38. Steynberg, A. P. In Studies in Surface Science and Catalysis; Steynberg, A. P.,
Dry, M. E., Eds.; 152, Elsevier: Amsterdam, 2004; pp 163.
39. Penning, R. In New Developments in Synthetic Fuels, CTL/GTL Conference 2010,
Brisbane, Australia.
40. Davis, B. H. Catal. Today 2003, 84, 8398.
41. Schwertmann, U.; Cornell, R. M. Iron Oxides in the Laboratory, Preparation and
Characterization. Wiley-VCH: Weinheim, 2000.
42. Smith, D. F.; Hawk, C. O.; Golden, P. L. J. Am. Chem. Soc. 1930, 52,
32213232.
43. Bromfield, T. C.; Botes, F. G.; Visagie, R.; Espinoza, R.; Gibson, P.; Van Lawson,
K. H. Hydrocarbon Synthesis Catalyst and Process. US Patent 6844370, 2002.
44. Hayakawa, H.; Tanaka, H.; Fujimoto, K. Appl. Catal. A: Gen. 2006, 310,
2430.
45. OBrien, R. J.; Xu, L. G.; Spicer, R. L.; Bao, S. Q.; Milburn, D. R.; Davis, B. H.
Catal. Today 1997, 36, 325334.
46. Luo, M. S.; OBrien, R. J.; Bao, S. Q.; Davis, B. H. Appl. Catal. A: Gen. 2003, 239,
111120.
47. Niemantsverdriet, J. W. Spectroscopy in Catalysis; An Introduction, 3rd ed.;
Wiley-VCH: Weinheim, 2007.
FischerTropsch Synthesis: Catalysts and Chemistry 555
48. Kock, A. J. H. M.; Geus, J. W. Prog. Surf. Sci. 1985, 20, 165272.
49. de Smit, E.; Cinquini, F.; Beale, A. M.; Safonova, O. V.; van Beek, W.; Sautet, P.;
Weckhuysen, B. M. J. Am. Chem. Soc. 2010, 132, 1492814941.
50. Raupp, G. B.; Delgass, W. N. J. Catal. 1979, 58, 348360.
51. Niemantsverdriet, J. W.; Van der Kraan, A. M.; Van Dijk, W. L.; Van der Baan, H. S.
J. Phys. Chem. 1980, 84, 33633370.
52. Dry, M. E. In Studies in Surface Science and Catalysis; Steynberg, A. P.,
Dry, M. E., Eds.; 152, Elsevier: Amsterdam, 2004; pp 533600.
53. Luo, M. S.; Hamdeh, H.; Davis, B. H. Catal. Today 2009, 140, 127134.
54. Herranz, T.; Rojas, S.; Perez-Alonso, F. J.; Ojeda, M.; Terreros, P.; Fierro, J. L. G.
J. Catal. 2006, 243, 199211.
55. Dry, M. E. In Catalysis, Science and Technology; Anderson, J. R., Boudart, M.,
Eds.; Springer-Verlag: New York, 1981; Vol. 1, pp 159255.
56. Luo, M. S.; Davis, B. H. Fuel Process. Technol. 2003, 83, 4965.
57. Fiato, R. A.; Soled, S. L. FischerTropsch Hydrocarbon Synthesis with High
Surface Area Copper and Potassium Promoted Reduced-Carbided Iron/
Manganese Spinels. US Patent 4621102A, 1986.
58. Du Toit, E. Ph.D. Thesis, University of the North West, 2002.
59. Niemantsverdriet, J. W.; van der Kraan, A. M. J. Catal. 1981, 72, 385388.
60. Bromfield, T. C.; Visagie, R. Chromium Oxide Incorporation into Precipitated Iron-
Based FischerTropsch Catalysts for Increased Production of Oxygenates and
Branched Hydrocarbons. Patent WO 2005049765A1, 2005.
61. van Steen, E.; Claeys, M. Chem. Eng. Technol. 2008, 31, 655666.
62. Tsakoumis, N. E.; Ronning, M.; Borg, O.; Rytter, E.; Holmen, A. Catal. Today
2010, 154, 162182.
63. Morales, F.; Weckhuysen, B. M. Catalysis 2006, 19, 140.
64. van de Loosdrecht, J.; Bazhinimaev, B.; Dalmon, J. A.; Niemantsverdriet, J. W.;
Tsybulya, S. V.; Saib, A. M.; van Berge, P. J.; Visagie, J. L. Catal. Today 2007,
123, 293302.
65. Saib, A. M.; Moodley, D. J.; Ciobica, I. M.; Hauman, M. M.; Sigwebela, B. H.;
Weststrate, C. J.; Niemantsverdriet, J. W.; van de Loosdrecht, J. Catal. Today
2010, 154, 271282.
66. Claeys, M.; van Steen, E. In Studies in Surface Science and Catalysis;
Steynberg, A. P., Dry, M. E., Eds.; 152, Elsevier: Amsterdam, 2004;
pp 601680.
67. Iglesia, E.; Reyes, S. C.; Madon, R. J.; Soled, S. L. Adv. Catal. 1993, 39,
221302.
68. Diehl, F.; Khodakov, A. Y. Oil Gas Sci. Technol. 2009, 64, 1124.
69. Bezemer, G. L.; Bitter, J. H.; Kuipers, H.; Oosterbeek, H.; Holewijn, J. E.; Xu, X. D.;
Kapteijn, F.; van Dillen, A. J.; de Jong, K. P. J. Am. Chem. Soc. 2006,
128, 39563964.
70. Jacobs, G.; Patterson, P. M.; Zhang, Y. Q.; Das, T.; Li, J. L.; Davis, B. H. Appl.
Catal. A: Gen. 2002, 233, 215226.
71. Karaca, H.; Hong, J. P.; Fongarland, P.; Roussel, P.; Griboval-Constant, A.;
Lacroix, M.; Hortmann, K.; Safonova, O. V.; Khodakov, A. Y. Chem. Commun.
2010, 46, 788790.
72. Kitakami, O.; Sato, H.; Shimada, Y.; Sato, F.; Tanaka, M. Phys. Rev. B 1997, 56,
1384913854.
73. Enache, D. I.; Rebours, B.; Roy-Auberger, M.; Revel, R. J. Catal. 2002, 205,
346353.
74. Iglesia, E.; Soled, S. L.; Fiato, R. A.; Via, G. H. J. Catal. 1993, 143, 345368.
75. Feller, A.; Claeys, M.; van Steen, E. J. Catal. 1999, 185, 120130.
76. Moradi, G. R.; Basir, M. M.; Taeb, A.; Kiennemann, A. Catal. Commun. 2003,
4, 2732.
77. Tan, K. F.; Chang, J.; Borgna, A.; Saeys, M. J. Catal. 2011, 280, 5059.
78. Rytter, E.; Skagseth, T. H.; Eri, S.; Sjastad, A. O. Ind. Eng. Chem. Res. 2010,
49, 41404148.
79. Beuther, H.; Kobylinski, T. P.; Kibby, C. L.; Pannell, R. B. Synthesis Gas
Conversion Using Ruthenium-Promoted Cobalt Catalyst Prepared by
Nonaqueous Impregnation. US Patent 4585798A, 1986.
80. Beuther, H.; Kibby, C. L.; Kobylinski, T. P.; Pannell, R. B. Fluid Bed Catalyst for
Synthesis Gas Conversion and Its Utilization for Preparation of Diesel Fuel. US
Patent 4413064A, 1983.
81. Beuther, H.; Kibby, C. L.; Kobylinski, T. P.; Pannell, R. B. Conversion of Synthesis
Gas to Diesel Fuel and Gasoline. US Patent 4605680A, 1986.
82. Eri, S.; Kinnari, K. J.; Schanke, D.; Hilmen, A.-M. Preparation and Use of
Promoted Cobalt Catalyst with Low Surface Area Alumina for FischerTropsch
Reaction with High Olefin Selectivity. Patent WO 2002047816A1, 2002.
83. Saib, A. M.; Claeys, M.; van Steen, E. Catal. Today 2002,
71, 395402.
84. Van Berge, P. J.; Van De Loosdrecht, J.; Barradas, S. Method of Treating an
Untreated Catalyst Support, and Forming a Catalyst Precursor and Catalyst from
the Treated Support. Patent EP 1 303 350 B1, 2000.
556 FischerTropsch Synthesis: Catalysts and Chemistry
85. Van Berge, P. J.; Van De Loosdrecht, J.; Barradas, S. Production of Fischer
Tropsch Synthesis Produced Wax. Patent EP 1 432 778 B1, 2001.
86. Wei, D. G.; Goodwin, J. G.; Oukaci, R.; Singleton, A. H. Appl. Catal. A: Gen. 2001,
210, 137150.
87. Xiong, H. F.; Motchelaho, M. A. M.; Moyo, M.; Jewell, L. L.; Coville, N. J. J. Catal.
2011, 278, 2640.
88. Stranges, A. N. Germanys Synthetic Fuel Industry 19271945. Presented at the
AIChE 2003, New Orleans, 2003. http://www.fischertropsch.org.
89. Keyser, M. J.; Everson, R. C.; Espinoza, R. L. Appl. Catal. A: Gen. 1998, 171,
99107.
90. Co-Precipitated CobaltZinc Catalysts for FischerTropsch Reaction or
Functional Group Hydrogenation. Patent EP1358934A1, 2003.
91. Puskas, I.; Fleisch, T. H.; Full, P. R.; Kaduk, J. A.; Marshall, C. L.; Meyers, B. L.
Appl. Catal. A: Gen. 2006, 311, 146154.
92. Joustra, A. H.; Scheffer, B. Process for the Preparation of Alumina-Based
Extrudates. Patent EP 455307A1, 1991.
93. Neimark, A. V.; Kheifets, L. I.; Fenelonov, V. B. Ind. Eng. Chem. Prod. Res. Dev.
1981, 20, 439450.
94. Kheifets, L. I.; Neimark, A. V.; Fenelonov, V. B. Kinet. Catal. 1979, 20,
626632.
95. Neimark, A. V.; Fenelonov, V. B.; Heifets, L. I. React. Kinet. Catal. Lett. 1976, 5,
6772.
96. Van Berge, P. J. In Scaling Up of an Alumina Supported Cobalt Slurry Phase
FischerTropsch Catalyst Preparation, CatCon World Wide Catalyst Industry
Conference, Houston, TX, USA, June 1213, 2000.
97. Van Berge, P. J.; Van De Loosdrecht, J.; Caricato, E. A.; Barradas, S.; Sigwebela,
B. H. Impregnation Process for Catalysts. Patent WO 2000020116A1, 2000.
98. Soled, S. L.; Baumgartner, J. E.; Reyes, S. C.; Iglesia, E. In Preparation of
Catalysts VI: Scientific Bases for the Preparation of Heterogeneous Catalysts,
Studies in Surface Science and Catalysis, 1995; Vol. 91, pp 989997.
99. de Jong, K. P. Deposition Precipitation onto Pre-shaped Carrier Bodies,
Possibilities and Limitations; Elsevier: Amsterdam, 1991; Vol. 63.
100. Boutonnet, M.; Jaras, S.; Logdberg, S., Method for Depositing Metal Particles on
a Support. Patent EP 1 985 361, 2008.
101. van de Loosdrecht, J.; Barradas, S.; Caricato, E. A.; Ngwenya, N. G.;
Nkwanyana, P. S.; Rawat, M. A. S.; Sigwebela, B. H.; van Berge, P. J.;
Visagie, J. L. Top. Catal. 2003, 26, 121127.
102. Wolters, M.; Munnik, P.; Bitter, J. H.; De Jongh, P. E.; De Jong, K. P. Method for
Producing a Supported Metal Nitrate. Patent WO 2010109216A1, 2010.
103. Soled, S.L.; Baumgartner, J.E.; Reyes, S.C.; Iglesia, E.; Poncelet, G., J. M. B. D. P.
A. J. a. P. G. In Studies in Surface Science and Catalysis; Elsevier, 1995; Vol. 91;
pp 989997.
104. Hoek, A.; Moors, J. H. Catalyst Activation and Rejuvenation Process. Patent WO
9717137A1, 1997.
105. Behrmann, W. C.; Davis, S. M.; Mauldin, C. H. Method for Preparing Cobalt-
Containing Hydrocarbon Synthesis Catalyst. Patent WO 9206784A1, 1992.
106. Oosterbeek, H. Phys. Chem. Chem. Phys. 2007, 9, 35703576.
107. van Berge, P. J.; Barradas, S.; van de Loosdrecht, J.; Visagie, J. L. Erdol Erdgas
Kohle 2001, 117, 138142.
108. Iglesia, E. Appl. Catal. A 1997, 161, 5978.
109. Barbier, A.; Tuel, A.; Arcon, I.; Kodre, A.; Martin, G. A. J. Catal. 2001, 200,
106116.
110. Bian, G. Z.; Fujishita, N.; Mochizuki, T.; Ning, W. S.; Yamada, M. Appl. Catal. A:
Gen. 2003, 252, 251260.
111. Martinez, A.; Prieto, G. J. Catal. 2007, 245, 470476.
112. Prieto, G.; Martinez, A.; Concepcion, P.; Moreno-Tost, R. J. Catal. 2009, 266,
129144.
113. Borg, O.; Dietzel, P. D. C.; Spjelkavik, A. I.; Tveten, E. Z.; Walmsley, J. C.;
Diplas, S.; Eri, S.; Holmen, A.; Ryttera, E. J. Catal. 2008, 259, 161164.
114. Iglesia, E. Appl. Catal. A: Gen. 1997, 161, 5978.
115. Martinez, A.; Rollan, J.; Arribas, M. A.; Cerqueira, H. S.; Costa, A. F.; S-
Aguiar, E. F. J. Catal. 2007, 249, 162173.
116. Fischer, N.; van Steen, E.; Claeys, M. Catal. Today 2011, 171, 174179.
117. Yang, J.; Tveten, E. Z.; Chen, D.; Holmen, A. Langmuir 2010, 26, 1655816567.
118. Saib, A. M.; Borgna, A.; de Loosdrecht, J. V.; van Berge, P. J.;
Niemantsverdriet, J. W. Appl. Catal. A: Gen. 2006, 312, 1219.
119. Reynhout, M. J. Cobalt-Based Hydrocarbon Synthesis Catalysts Prepared from
Solid Solution Mixture of Metal Compound Precursors. Patent WO
2008061970A2, 2008.
120. White Paper FischerTropsch Catalyst Test on Coal Derived Synthesis Gas,
Syntroleum Corporation. http://www.syntroleum.com.
121. Moodley, D. J.; van de Loosdrecht, J.; Saib, A. M.; Overett, M. J.; Datye, A. K.;
Niemantsverdriet, J. W. Appl. Catal. A: Gen. 2009, 354, 102110.
122. Moodley, D. J.; Saib, A. M.; van de Loosdrecht, J.; Welker-Nieuwoudt, C. A.;
Sigwebela, B. H.; Niemantsverdriet, J. W. Catal. Today 2011, 171, 192200.
123. Moodley, D. J.; van de Loosdrecht, J.; Saib, A. M.; Niemantsverdriet, J. W. Chem.
Ind. 2010, 128, 4981.
124. Ciobica, I. M.; van Santen, R. A.; van Berge, P. J.; van de Loosdrecht, J. Surf. Sci.
2008, 602, 1727.
125. Soled, S. L.; Iglesia, E.; Fiato, R. A.; Baumgartner, J. E.; Vroman, H.; Miseo, S.
Top. Catal. 2003, 26, 101109.
126. Hilmen, A. M.; Schanke, D.; Holmen, A. In Natural Gas Conversion IV;
dePontes, M., Espinoza, R. L., Nicolaides, C. P., Scholtz, J. H., Scurrell, M. S.,
Eds.; Elsevier: Amsterdam, 1997; Vol. 107, pp 237242.
127. Rothaemel, M.; Hanssen, K. F.; Blekkan, E. A.; Schanke, D.; Holmen, A. Catal.
Today 1997, 38, 7984.
128. Craje, M. W. J.; van der Kraan, A. M.; van de Loosdrecht, J.; van Berge, P. J. Catal.
Today 2002, 71, 369379.
129. Kiss, G.; Kliewer, C. E.; DeMartin, G. J.; Culross, C. C.; Baumgartner, J. E.
J. Catal. 2003, 217, 127140.
130. Das, T. K.; Jacobs, G.; Patterson, P. M.; Conner, W. A.; Li, J. L.; Davis, B. H. Fuel
2003, 82, 805815.
131. Li, J. L.; Jacobs, G.; Zhang, Y. Q.; Das, T.; Davis, B. H. Appl. Catal. A: Gen. 2002,
223, 195203.
132. Storsaeter, S.; Borg, O.; Blekkan, E. A.; Holmen, A. J. Catal. 2005, 231, 405419.
133. Huffman, G. P.; Shah, N.; Zhao, J. M.; Huggins, F. E.; Hoost, T. E.; Halvorsen, S.;
Goodwin, J. G. J. Catal. 1995, 151, 1725.
134. Bezemer, G. L.; Remans, T. J.; van Bavel, A. P.; Dugulan, A. I. J. Am. Chem. Soc.
2010, 132, 85408541.
135. Yan, Z.; Wang, Z. J.; Bukur, D. B.; Goodman, D. W. J. Catal. 2009, 268, 196200.
136. den Breejen, J. P.; Sietsma, J. R. A.; Friedrich, H.; Bitter, J. H.; de Jong, K. P.
J. Catal. 2010, 270, 146152.
137. Ronning, M.; Tsakoumis, N. E.; Voronov, A.; Johnsen, R. E.; Norby, P.; van
Beek, W.; Borg, O.; Rytter, E.; Holmen, A. Catal. Today 2010, 155, 289295.
138. Li, J. L.; Zhan, X. D.; Zhang, Y. Q.; Jacobs, G.; Das, T.; Davis, B. H. Appl. Catal. A:
Gen. 2002, 228, 203212.
139. British Intelligence Objectives Sub-Committee, Interrogation of Dr Otto Roelen of
Ruhrchemie A.G., B.I.O.S. Final Report No. 447; Item no 30 (1945). http://www.
fischer-tropsch.org.
140. Zonnevylle, M. C.; Geerlings, J. J. C.; van Santen, R. A. Surf. Sci. 1990, 240,
253262.
141. Tan, K. F.; Xu, J.; Chang, J.; Borgna, A.; Saeys, M. J. Catal. 2010, 274, 121129.
142. Saeys, M.; Tan, K. F.; Chang, J.; Borgna, A. Ind. Eng. Chem. Res. 2010, 49,
1109811100.
143. Jacobs, G.; Sarkar, A.; Ji, Y.; Luo, M.; Dozier, A.; Davis, B. H. Ind. Eng. Chem.
Res. 2008, 47, 672680.
144. Tavasoli, A.; Abbaslou, R. M. M.; Dalai, A. K. Appl. Catal. A: Gen. 2008, 346,
5864.
145. Zhou, W.; Chen, J. G.; Fang, K. G.; Sun, Y. H. Fuel Process. Technol. 2006, 87,
609616.
146. Bartholomew, C. H. Appl. Catal. A: Gen. 2001, 212, 1760.
147. Madon, R. J.; Shaw, H. Catal. Rev. 1977, 15, 69106.
148. Liu, Z. T.; Zhou, J. L.; Zhang, B. J. J. Mol. Catal. 1994, 94, 255261.
149. Bartholomew, C. H.; Bowman, R. M. Appl. Catal. 1985, 15, 5967.
150. Leviness, S. C.; Mart, C. J.; Behrmann, W. C.; Hsia, S. J.; Neskora, D. R. Slurry
Hydrocarbon Synthesis Process with Increased Catalyst Life. Patent WO
9850487A1, 1998.
151. Wilson, J.; De Groot, C. J. Phys. Chem. 1995, 99, 78607866.
152. Luo, M. S.; Davis, B. H. In Catalyst Deactivation 2001, Proceedings, Studies in
Surface Science and Catalysis, 2001; Vol. 139, pp 133140.
153. Thune, P. C.; Weststrate, C. J.; Moodley, P.; Saib, A. M.; van de Loosdrecht, J.;
Miller, J. T.; Niemantsverdriet, J. W. Catal. Sci. Technol. 2011, 1, 689697.
154. Reynhout, M. J. Process for Regenerating a Cobalt Catalyst. Patent EP
1920836A1, 2008.
155. Pichler, H.; Schulz, H. Chemie Ingenieur Technik 1970, 42, 11621174.
156. Lahtinen, J.; Vaari, J.; Kauraala, K. Surf. Sci. 1998, 418, 502510.
157. Beitel, G. A.; Laskov, A.; Oosterbeek, H.; Kuipers, E. W. J. Phys. Chem. 1996,
100, 1249412502.
158. Habermehl-Cwirzen, K. M. E.; Kauraala, K.; Lahtinen, J. Phys. Scripta 2004,
T108, 2832.
159. Ernst, K. H.; Schwarz, E.; Christmann, K. J. Chem. Phys. 1994, 101, 53885401.
160. Bridge, M. E.; Comrie, C. M.; Lambert, R. M. J. Catal. 1979, 58, 2833.
161. Prior, K. A.; Schwaha, K.; Lambert, R. M. Surf. Sci. 1978, 77, 193208.
162. Geerlings, J. J. C.; Zonnevylle, M. C.; Degroot, C. P. M. Surf. Sci. 1991, 241,
315324.
163. Papp, H. Surf. Sci. 1985, 149, 460470.

164. Nowitzki, T.; Borchert, H.; Jurgens, B.; Risse, T.; Zielasek, V.; Baumer, M.
ChemPhysChem 2008, 9, 729739.
165. Habermehl-Cwirzen, K.; Lahtinen, J.; Hautojarvi, P. Surf. Sci. 2005, 598,
128135.
166. Weststrate, C. J.; Gericke, H. J.; Verhoeven, M.; Ciobica, I. M.; Saib, A. M.;
Niemantsverdriet, J. W. J. Phys. Chem. Lett. 2010, 1, 17671770.
167. Steinbach, F.; Kiss, J.; Krall, R. Surf. Sci. 1985, 157, 401412.
168. Denecke, R. Appl. Phys. Mater. Sci. Process. 2005, 80, 977986.
169. Krebs, H. J.; Bonzel, H. P.; Gafner, G. Surf. Sci. 1979, 88, 269283.
170. Dwyer, D. J.; Gland, J.; Albert, M.; Bernasek, S., Intermediates to the Dissociative
Chemisorption of Co and Ch3oh on Fe(100). Abstracts of Papers of the American
Chemical Society 1987, 193, 30.
171. Moon, D. W.; Cameron, S.; Zaera, F.; Eberhardt, W.; Carr, R.; Bernasek, S. L.;
Gland, J. L.; Dwyer, D. J. Surf. Sci. 1987, 180, L123L128.
172. Dwyer, D. J.; Somorjai, G. A. J. Catal. 1978, 52, 291301.
173. Dwyer, D. J.; Hardenbergh, J. H. J. Catal. 1984, 87, 6676.
174. Wedler, G.; Colb, K. G.; McElhiney, G.; Heinrich, W. Appl. Surf. Sci. 1978, 2, 3042.
175. Wedler, G.; Colb, K. G.; Heinrich, W.; McElhiney, G. Appl. Surf. Sci. 1978, 2,
85101.
176. Vink, T. J.; Gijzeman, O. L. J.; Geus, J. W. Surf. Sci. 1985, 150, 1423.
177. van Santen, R. A.; Neurock, M.; Shetty, S. G. Chem. Rev. 2010, 110,
20052048.
178. Hammer, B.; Norskov, J. K. In Advances in Catalysis, Impact of Surface Science
on Catalysis; Academic Press: San Diego, 2000; Vol. 45, pp 71129.
179. Bromfield, T. C.; Ferre, D. C.; Niemantsverdriet, J. W. ChemPhysChem 2005, 6,
254260.
180. Curulla-Ferre, D.; Govender, A.; Bromfield, T. C.; Niemantsverdriet, J. W. J. Phys.
Chem. B 2006, 110, 1389713904.
181. Sorescu, D. C. J. Phys. Chem. C 2008, 112, 1047210489.
182. Sorescu, D. C.; Thompson, D. L.; Hurley, M. M.; Chabalowski, C. F. Phys. Rev. B
2002, 66, 035416.
183. Ojeda, M.; Nabar, R.; Nilekar, A. U.; Ishikawa, A.; Mavrikakis, M.; Iglesia, E.
J. Catal. 2010, 272, 287297.
184. Scheijen, F. J. E.; Ferre, D. C.; Niemantsverdriet, J. W. J. Phys. Chem. C 2009,
113, 1104111049.
185. Inderwildi, O. R.; Jenkins, S. J.; King, D. A. J. Phys. Chem. C 2008, 112,
13051307.
186. Shetty, S.; van Santen, R. A. Phys. Chem. Chem. Phys. 2010, 12, 63306332.
187. Jones, G.; Bligaard, T.; Abild-Pedersen, F.; Norskov, J. K. J. Phys. Condens.
Matter 2008, 20, 064239.
FischerTropsch Synthesis: Catalysts and Chemistry 557
188. Chorkendorff, I.; Niemantsverdriet, J. W. Concepts of Modern Catalysis and
Kinetics. Wiley-VCH: Weinheim, 2003.
189. Ciobica, I. M.; van Santen, R. A. J. Phys. Chem. B 2002, 106, 62006205.
190. Sorescu, D. C. Phys. Rev. B 2006, 73.
191. Lo, J. M. H.; Ziegler, T. J. Phys. Chem. C 2007, 111, 1101211025.
192. Govender, A. Towards a Mechanism for the FischerTropsch Synthesis on
Fe(100) Using Density Functional Theory. Ph.D. Thesis, Eindhoven University of
Technology, Eindhoven, The Netherlands, 2010.
193. Cheng, J.; Hu, P.; Ellis, P.; French, S.; Kelly, G.; Lok, C. M. J. Phys. Chem. C
2010, 114, 10851093.
194. Cheng, J.; Gong, X. Q.; Hu, P.; Lok, C. M.; Ellis, P.; French, S. J. Catal. 2008,
254, 285295.
195. Biloen, P.; Helle, J. N.; Sachtler, W. M. H. J. Catal. 1979, 58, 95107.
196. Biloen, P.; Sachtler, W. M. H. Adv. Catal. 1981, 30, 165216.
197. Ciobica, I. M.; Kramer, G. J.; Ge, Q.; Neurock, M.; van Santen, R. A. J. Catal.
2002, 212, 136144.
198. Zhuo, M. K.; Tan, K. F.; Borgna, A.; Saeys, M. J. Phys. Chem. C 2009, 113,
83578365.
199. Gracia, J. M.; Prinsloo, F. F.; Niemantsverdriet, J. W. Catal. Lett. 2009, 133, 257261.
200. Deng, L. J.; Huo, C. F.; Liu, X. W.; Zhao, X. H.; Li, Y. W.; Wang, J. G.; Jiao, H. J.
J. Phys. Chem. C 2010, 114, 2158521592.
201. Anderson, R. B. In Catalysis; Emmett, P. H., Ed.; Reinhold Publishing Company:
New York, 1956; Vol. IV.
202. Botes, F. G.; van Dyk, B.; McGregor, C. Ind. Eng. Chem. Res. 2009, 48,
1043910447.
203. Botes, F. G. Catal. Rev.: Sci. Eng. 2008, 50, 471491.
204. Botes, F. G. Ind. Eng. Chem. Res. 2009, 48, 18591865.
205. Yang, J.; Liu, Y.; Chang, J.; Wang, Y. N.; Bai, L.; Xu, Y. Y.; Xiang, H. W.; Li, Y. W.;
Zhong, B. Ind. Eng. Chem. Res. 2003, 42, 50665090.
206. van Dijk, H. A. J. Ph.D. Thesis, Eindhoven University of Technology, 2001
207. Botes, F. G.; Breman, B. B. Ind. Eng. Chem. Res. 2006, 45, 74157426.
208. Yates, I. C.; Satterfield, C. N. Energy Fuel 1991, 5, 168173.
209. Donnelly, T. J.; Yates, I. C.; Satterfield, C. N. Energy Fuel 1988, 2, 734739.
210. Lozano-Blanco, G.; Thybaut, J. W.; Surla, K.; Galtier, P.; Marin, G. B. Ind. Eng.
Chem. Res. 2008, 47, 58795891.
211. Yang, Q. Y.; Maynard, K. J.; Johnson, A. D.; Ceyer, S. T. J. Chem. Phys. 1995,
102, 77347749.
212. Govender, N. S.; Botes, F. G.; de Croon, M.; Schouten, J. C. J. Catal. 2008, 260,
254261.
213. Weckhuysen, B. M. Angew. Chem. Int. Ed. 2009, 48, 49104943.
 7.12 Hydrotreating: Removal of Sulfur from Crude Oil Fractions
with Sulfide Catalysts
W Bensch, Christian-Albrechts-Universitat zu Kiel, Kiel, Germany
2013 Elsevier Ltd. All rights reserved.

7.12.1 General Introduction 287

7.12.2 Sulfur Compounds in Diesel Feeds and Reactivities of These Compounds 289
7.12.3 HDS Reaction Pathways and Mechanism 289
7.12.4 Catalyst Components, Formation and Nature of Active Phases 292
7.12.5 Analytical Tools for the Characterization of the Catalysts 294
7.12.6 Formation of the Sulfide Phases 295
7.12.7 Conversion of MoO3 into MoS2 295
7.12.7.1 Conversion of MoO3 into MoS2 Studied with XPS and IRES 295
7.12.7.2 Temperature-Programmed Sulfidation of MoO3 297
7.12.8 Quick EXAFS Spectroscopy Investigation of the Local Environments in Ni-Promoted Molybdenum Sulfide
Catalysts 298
7.12.9 Sulfidation of Co-Promoted Molybdenum Oxide in the Presence of a Chelating Molecule Investigated
with XPS 299
7.12.10 Determination of the State of Co in CoMoS Phase Using In Situ Mossbauer Emission Spectroscopy 300
7.12.11 The Local Structures of Co/Ni and Mo in CoMoS and NiMoS Phases Determined with EXAFS 301
7.12.12 Sulfidation of NiW Catalysts Supported on Al2O3 303
7.12.12.1 x-Ray Absorption Spectroscopy Performed at the W LIII Absorption Edge 303
7.12.12.2 Ni K Edge x-Ray Absorption Spectroscopy 304
7.12.13 Probing the Active Sites: CO and NO Adsorption 305
7.12.14 Structure of the Phases 308
7.12.14.1 The Role of Carbon in HDS Catalysts 310
7.12.14.2 Morphology and Microstructure of Catalysts 312
7.12.15 Choice of Catalysts and Other Catalysts for HDS 315
7.12.16 Poisoning and Deactivation Effects 316
7.12.17 Synthesis Strategies of Catalysts 316
7.12.18 Supported Catalysts and the Support Effect 317
7.12.19 Conclusion 317
References 318

7.12.1 General Introduction
In 2010, about 3914 million tons of crude oil was produced
worldwide. Crude oil is a very complex chemical mixture con-
taining hydrocarbons (alkanes, alkenes, cycloalkanes, and aro-
matics), sulfur compounds (mercaptanes, sulfides, disulfides,
thiophenes, and dibenzothiophenes), nitrogen compounds
(anilines, aliphatic amines, quinoline, acridine, carbazole,
and indole), oxygenates (phenols, ketones, and carboxylic
acids), and metal atoms such as V and Ni that are present as
porphyrin complexes of V4O and Ni2 to mention the
most important constituents. The composition of crude oil
depends on the location where it is produced with the sulfur
level varying between 1 and 6 wt%. Average values in weight%
are for carbon 8387, hydrogen 1014, nitrogen 0.12, oxygen
0.11.5, sulfur 0.56, and metals <0.1.
Different physical and chemical processes such as distilla-
tion, extraction, reforming, hydrogenation, cracking, and
blending are used in a refinery converting crude oil to higher
value products. The main products are liquid petroleum gas,
gasoline, jet fuel, diesel fuel, wax, lubricants, bitumen, and
petrochemicals. The sulfur content increases with increasing
boiling point of the different fractions from crude oil.
Hydrotreating or hydroprocessing refers to a variety of cat-
alytically promoted chemical reactions such as hydrogenation,
which saturate unsaturated hydrocarbons and remove S by
hydrodesulfurization (HDS), N by hydrodenitrogenation
(HDN), O by hydrodeoxygenation (HDO), and metals by
hydrodemetallization (HDM) from different petroleum streams
in a refinery. Hydrotreating is mainly applied and is necessary to
avoid air pollution emissions, to avoid poisoning of noble
metals and acid catalysts used in catalytic reforming and crack-
ing, and to improve fuel quality. The different streams are trans-
formed into products through catalytically promoted chemical
reactions such as hydrogenation, isomerization, aromatization,
alkylation, cracking, and hydrotreating (Scheme 1).
The important products of an oil refinery are diesel fuel and
gasoline. About 1 ton of crude oil delivers 0.25 tons of diesel
fuel. Due to stringent environmental regulations,13 the sulfur
content of diesel fuel must be reduced to ultralow levels

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00723-3 287

288 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

Gas
Amine guard H2 plant Polybed PSA
concentration

From FCC LPG

From FCC Propylene

Light
straight
run Light naphtha
Hydrotreating
isomerization

Crude Hydrotreating Reforming Gasoline

distillation
Naphtha

Kerosene
Kerosene
jet fuel
Middle distillate
Hydrotreating
Diesel

Desalted
crude oil Atmospheric Catalytic
gas oil condensation
C4
isomerization
Alkylation
SHP

FCC Hydrotreating

Cycle oils
Hydrotreating

Hydrocracking

Lube oils
Vacuum
distillation

Asphalt

Fuel oil

Black oil Solid fuel

Refinery adsorbent locations

for contaminant removal
Scheme 1 Process flow from crude oil to final products. http://www.uop.com/refining-flowscheme-2/. Abbreviations: FCC, fluid catalytic cracking;
PSA, pressure swing adsorption; LPG, liquified petroleum gas.

(ultralow sulfur diesel, ULSD, 1015 ppm) to decrease harm-
ful exhaust emissions of diesel engines and to improve air
quality (see also Chapter 7.19). In addition, there is also a
need to strictly reduce the S content because even trace levels of
S can poison other catalysts (e.g., Ni, Cu, Pd, or Pt) used in
further catalytic processes (hydrocracking, reforming, automo-
tive catalysis, fuel cells, etc.). Hydrotreating is one important
catalytic step to convert crude oil streams into diesel fuel
(Scheme 1). In the presence of a catalyst and molecular
hydrogen, double bonds of aromatics are hydrogenated and
S, N, O, and metals are removed. The catalysts widely used are
Mo and/or W-based sulfides promoted with Co and/or Ni (see
also Chapter 5.09). The catalysts are supported on g-Al2O3 or
are supplied as bulk material. Depending on the distinct pro-
cess, typical reaction conditions for hydrotreating are
300400  C and total pressure up to 15 MPa. The actual condi-
tions depend on the boiling point of the feed. The required
severe reduction of the sulfur level is a complicated chemical
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts 289

and technical problem. Many different factors such as the cata-
lysts,46 process parameters,7 feedstock source and quality,8,9
reactivities of sulfur compounds,10 inhibition effects of H2S,11
nitrogen compounds,12 and aromatics13,14 present in the feed
influence the degree of desulfurization of diesel feeds. Thou-
sands of papers and patents were published in the past and
state of the art up to 2002 has been reviewed in Refs. [1517].
The actuality of research in the field of hydrotreating was impres-
sively demonstrated by F. Parlevliet and S. Eijsbouts in 2008:
counting only sulfide-based catalysts, since 1986 more than
3700 patents and about 5300 nonpatents were published.18 In
2010, a review was published where the different scientific and
technological aspects of ULSD production are thoroughly dis-
cussed and 702 references are cited.19 Further reviews and
review-like papers present and discuss the state of the art in
HDS catalysis at different times of development.1517,2036
A short discussion of the development of the knowledge about
HDS catalysts is given in the last section of the chapter.
petroleum. The reactivity of BTs and DBTs decreases with
increasing molecular geometry (Figure 2).
Setting the relative rate for thiophene as 100, benzothio-
phene has a rate of 50, DBT of 30, 4-methyl-DBT of 5, and di-
or trisubstituted DBTs of 1 (see Scheme 2).37
7.12.3 HDS Reaction Pathways and Mechanism
During hydrotreating a complicated set of exothermic reactions
occurs, and the main reactions are hydrogenation of aromatics,
HDS, and HDN. Even for the simple substrates such as thio-
phene, the HDS networks are complex: a considerable number
Compound Relative rate
Thiophene 100
S

7.12.2 Sulfur Compounds in Diesel Feeds and

Reactivities of These Compounds Benzothiophene 50
S
Diesel feeds from different sources were investigated with
different analytic techniques to identify and quantify the S-
containing species. Many different sulfur compounds were Dibenzothiophene 30
identified with the most abundant classes being benzothio-
S
phenes (BTs) and dibenzothiophenes (DBTs). The BTs may
contain alkyl substituents (17 C atoms) and the DBTs such
substituents with 15 C atoms. Whereas sulfur can be relatively Methyldibenzothiophene 5
easily removed from BTs, b-DBTs (called refractory S- S
compounds) with alky groups in position 4 or 4 and 6
CH3
(Figure 1) require drastic catalytic conditions. The distribution
of sulfur compounds mainly depends on the origin of the Dimethyldibenzothiophene 1
S
CH3 CH3

Trimethyldibenzothiophene 1
S CH3
S CH3 CH3
H3C CH3
Scheme 2 Relative rates of different S-containing molecules relative to
Figure 1 The molecular structure of 4,6-methyl-DBT. thiophene.

R Deep HDS
Reactivity

R R
S
S R
R
S

R R
S
R R
S R R
S
Molecular geometry
Figure 2 Decrease of the reactivity of different S-containing compounds in diesel feed. Present state of the art and future challenges in the
hydrodesulfurization (HDS) of polyaromatic sulfur compounds.32
290 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

of elementary steps such as adsorption/desorption equilibria
and hydrogenation and hydrogenolysis reactions, whose kinet-
ics strongly depend on the reaction conditions and on the
structure of the particular substrate, occur. The understanding
of the fundamental steps is further complicated because results
reported by different authors obtained on the same substrate
such as thiophene applying similar catalysts and approximately
identical reaction conditions are often very different, if not
contradictory. Nevertheless, for the HDS reaction some primary
phenomena involved can be summarized as follows:
adsorption (coordination) of the S-containing compound
onto the active site,
dissociative adsorption of hydrogen onto the surface of the
catalyst,
hydrogenation of unsaturated CC bonds,
cleavage of two CS bonds sequentially or simultaneously,
addition of hydrogen to the broken bond of both S and C,
release of the hydrocarbon product from the catalytic site, and
release of H2S from the site.
S active sites of the catalyst surface. Which reaction pathway
S The DDS pathway produces biphenyl (BP), while in the
Scheme 3 The two reaction pathways of HDS of thiophene with the
different desulfurized and hydrogenated products.
One should note that the phenomena might not necessarily
occur in the order shown here (see also Chapter 7.02).
The reaction pathways of HDS of thiophene were inten-
sively studied and it was proposed that the reaction proceeds
via two parallel pathways (Scheme 3). In one reaction route
(bottom path in Scheme 3), the sulfur from thiophene is
removed directly (direct desulfurization, DDS) to form buta-
diene, which is then hydrogenated to give 1-butene and cis-
and trans-2-butene. In a further reaction, butenes are then
hydrogenated to n-butane. In the second pathway (top path
in Scheme 3), the first step is hydrogenation of thiophene to
tetrahydrothiophene, which is further desulfurized to butadi-
ene (hydrogenation pathway, HYD). Tetrahydrothiophene
and butadiene were observed in trace amounts in the reaction
products, and it was suggested that these intermediates are very
reactive and they are transformed rapidly to butenes and
butane.29 Due to this observation, the use of thiophene as
the S-containing model compound is of less interest because
the different reaction pathways cannot be distinguished.
In a simplified picture, the general HDS of BTs and DBTs
consists also of two routes: the DDS and the HYD pathway
(Figure 3), both in the presence of molecular H2. One should
note that both pathways occur in parallel employing different
predominates depends on the nature of the sulfur compounds,
the reaction conditions, and the catalyst used.
HYD route first hydrogenation of a benzene ring occurs fol-
lowed by desulfurization, and in the case of DBT cyclohexyl-
benzene is obtained.38 In addition to hydrogenation of an
aromatic ring before sulfur removal, hydrogenation of an aro-
matic ring may occur after sulfur was removed with a distinct

Benzothiophene

+H2
S S
-H2S
+ H2
S -H2S
+H2

Dibenzothiophene

+H2 S +H2 -H2S

S +H2 -H2S

S S
-H2S
+H2 +H2

+H2

Figure 3 Simplified picture of the two principal pathways during HDS of S-containing molecules.15
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts 291

rate constant for this reaction. This often leads to confusion in
interpreting the results of experimental data as both routes can
produce cyclohexylbenzene as the final product. The hydroge-
nation pathways are subject to thermodynamic equilibrium
constraints and therefore partially hydrogenated intermediates
such as tetrahydrodibenzothiophenes have lower equilibrium
concentrations at higher temperatures. Consequently, a maxi-
mum of HDS via the hydrogenative route is observed as a
function of temperature. Hence, HDS via the HYD pathway
becomes limited at low pressures and high temperatures. In a
different route, isomerization and transmethylation of methyl
groups at the 4- or the 6- position may occur reducing the steric
hindrance of the DBT molecule.
It is assumed that the direct pathway DDS involves the
insertion of a metal atom on the surface of the catalyst into a
sulfurcarbon bond in the S-containing compound. After inser-
tion of the metal atom, several reactions occur (Figure 4) lead-
ing to the removal of sulfur as H2S and the catalyst is restored to
its original state.
The formation of BP in the DDS pathway reflects the cata-
lysts hydrogenolysis function for direct CS bond cleavage,
and the latter is due to the catalysts hydrogenation function
to effect ring hydrogenation followed by CS bond cleavage
(see also Chapters 7.16 and 8.09 ).39 The two different desul-
furization routes imply the presence of two different active sites
in the catalysts.
In Figure 5, the schemes of both reactions including details
of the active sites are illustrated. The active site is a coordina-
tively unsaturated metal site where the S ligand is facile. The
S atom in aromatic rings can coordinate to the active center
of HYD and DDS functions. There is evidence that in the case
of DDS the initial adsorption of the S compound occurs
through s-bonding, while for hydrogenation coordination is
through p-bonding for both S removal and hydrogenation.
The main reason for the low reactivity of b-DBTs seems to be
a steric hindrance due to retarded adsorption onto the
catalyst surface40,41 implying that adsorption is the rate-limiting
step. Two types of chemisorption patterns of 4,6-
dimethyldibenzothiophene (4,6-DMDBT) on MoS2 are
discussed: a flat adsorption and S-m3 type adsorption. The two
molecules DBT and 4,6-DMDBT can interact well with the
(10 1 0) edge of MoS2 by flat chemisorption. However, the
chemisorption of 4,6-DMDBT is more difficult using the S-m3
type coordination due to the steric hindrance of the alkyl
groups. This steric hindrance is expected to increase with increas-
ing size of the alkyl groups (from methyl to ethyl to propyl).
Consequently, an end-on adsorption mode is proposed involving
interactions between the S atom and the active sites. A different
explanation is that steric hindrance slows down the surface CS
bond scission.42 In this case, the rate-limiting step is not the S-
centered adsorption and a side-on adsorption of the DBT mole-
cule involving the p-electrons of the aromatic ring is proposed.
Alkyl substituents in DBTs alter the overall HDS due to reduction
of HDS reactivity and due to a change of the ratio between rates of
the HYD and the DDS pathway. Alkyl substituents in the 4- and/
or 6-positions result in pronounced preference of the HYD
route.7,23,38,40 In a recent study, the reaction sequence of the
DDS route was determined as DBT > 4-MDBT > 4,6-DMDBT

H- L L
L +n-2 S
+n
S M M S
M Hydride attack
(endo or exo) S S
S S
S S Oxidative
addition
I II
[H] V
H IV
H III -
H H+
L H L H
S S H L L
M M +n
M S M S M S
S S S S
S S S S S S

S-ring not opened S-ring opened

Reductive elimination Hydride attack Proton attack

H+ Proton attack(exo)
(endo) (endo or exo) (exo)

H H H
H H
H
+n-2 L - L +
L H M S M S M S
S
M S
S S S S S
S S
H+ proton attack

(a) 2,3-Dihydrothiophene (b) H2S, butadiene, thiobutene

Figure 4 Two different pathways to hydrogenation and ring-opening reactions leading to the products (a) and (b). Note that the mechanisms were
determined with organometallic complexes. L ligands; M Co or Ni.
292 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

Sequential deS Simultaneous deS Hydrogenation

S H S H H H
H H

S S S S S S

Mo S Mo S Mo Mo S Mo S Mo Mo S Mo S Mo

H H

H H H H H H
H H
S S S S S S S S

Mo S Mo S Mo Mo S Mo S Mo Mo S Mo S Mo

H
H H H H H

S S S S S S S S

Mo S Mo S Mo Mo S Mo S Mo Mo S Mo S Mo

H2S
+H2

H H

S S S S S

Mo S Mo S Mo Mo S Mo S Mo

+H2
H2S
= Vacancy
H H H H
S S S

Mo S Mo S Mo

Figure 5 Vacancy model of the HDS mechanism as published in Ref. [32].

being explained by steric hindrance due to the methyl groups
located adjacent to the S atom of DBT preventing s-bonding of S
to the active site of the catalyst. However, no dependence of
reaction rates of HYD pathways on the number of methyl groups
was observed indicating that the alkyl groups do not hamper the
p-adsorption of DBTs on the active sites.43 Deep desulfurization
means to desulfurize the refractory S-containing molecules.
7.12.4 Catalyst Components, Formation and
Nature of Active Phases
Most industrially used catalysts consist of a support with a high
specific surface (e.g., g-Al2O3) and active materials such as
MoS2/WS2 promoted with Co and/or Ni. In nature, MoS2 occurs
as the mineral molybdenite in the form of black, lead-silvery gray
crystallites or aggregates. In the structure of 2H-Mo(W)S2
(H hexagonal, 2 two MS2 slabs in the unit cell, space group
P63/mmc), layers consisting of SMoS slabs are held together by
weak van der Waals forces (Figure 6). The Mo/W atoms are in a
trigonal-prismatic environment of six S atoms and the Mo(W)S6
polyhedra share edges within the slabs. Consecutive layers are
stacked in an . . .ABAB. . . fashion along the crystallographic c-axis.
The MoS bonds (MoS: 2.43 A) within the layers are
covalent and in the trigonal-prismatic ligand field, a two-
electron band of mainly dz2 character is separated from the
main part of the Mo 4d band and hybridizes strongly with the S
3p band. The dz2 band is completely filled and is separated by a
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
a
b
Figure 6 The crystal structure of 2H-MoS2.
gap of Eg  1.3 eV from the empty eg part of the Mo 4d band.
Cleaving MoS2 crystallites between the layers generates a chem-
ically inert (0 0 0 1) surface, while the (1 0 1 0) surface
formed at the step edges of the basal planes exposes in alter-
nating SMoS sandwiches rows of coordinatively unsaturated
S and Mo atoms, and is considered chemically reactive.
In the following discussion, one should distinguish
between the terms active site and active phase: the active of the oxides into the sulfidic phases, and sulfide formation
site is the transition metal ion where the adsorption of the
reactant molecule such as a heterocyclic compound or hydro-
gen occurs through S vacancies or S ions. The active site and the
ensemble of transition metal and sulfur ions are the active
phase making this site active.
The promoted catalysts are often abbreviated as CoMo(W)
S/NiMo(W)S or CoMo(W)/NiMo(W), which are not chemical
formula. The chemical compositions vary and the optimal
promoter content depends on the HDS reaction conditions
and the diesel feed. The Mo(W):S ratio is near 1:2, but non-
stoichiometry has also been reported. In most applications and
test reactions, the Co(Ni)/[Co(Ni) Mo(W)] ratio varies
between 0.2 and 0.4.
It is well documented that the structure and the dispersion
of the catalytically active phase are dramatically altered during
the life cycle of a typical supported CoMo/NiMo catalyst,
which is schematically shown in Figure 7.
The preparation of the catalysts involves several individual
steps: First, a Mo precursor such as (NH4)6Mo7O244H2O (AHM)
(1520 wt%) is impregnated on for example, g-Al2O3 (impreg-
nation route) (see also Chapter 2.10). Then, different methods
are used to introduce the promotors: the AHM-containing
293
support is first calcined at elevated temperatures yielding
MoO3/g-Al2O3 followed by impregnation with the Co or Ni
precursor (mainly Ni2/Co2 salts such as nitrates) (15 wt%).
The impregnated composite is calcined again to form (CoO/NiO)
MoO3/g-Al2O3. There are reports that MoO3/g-Al2O3 is presul-
fided at elevated temperatures and the product is then impreg-
nated with the promotor salt. A critical step is the transformation
for one of the metals prior to the other should be avoided. It is
well documented that Co is sulfided at lower temperatures
than Mo. Hence, often additives such as chelating agents (citric
acid, urea, glycol, nitriloacetic acid (NTA), and ethylene diamine
tetraacetic acid (EDTA)) are added to increase sulfidization
temperature of the promotors thus leading to a better dispersion
of promotors. In addition, such additives lower the interaction
strength between the catalysts and the support. Reactions
between the Mo and Co/Ni precursors and the support should
also be avoided and therefore the temperature of the sulfidiza-
tion step should not be too high. Two strategies are followed to
sulfidize the oxides: in the ex situ approach, the material is treated
at elevated temperatures in, for example, 10% H2S/H2, 10%
H2S/He, or 10% H2S/N2 flow; in the in situ method, the oxides
are transferred to the HDS reaction chamber and sulfidization
occurs in the presence of S-containing molecules. No general
conclusion can be drawn whether the in situ route generates
better catalysts than the ex situ method, and vice versa.
In the as-prepared state, oxidic CoMo/NiMo catalysts
display a high Mo dispersion with well-dispersed MoO3 or
polymolybdate phases and well-dispersed CoO/NiO is located
near Mo particles.4447 In some investigations, the presence of
294 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

a MoO3 b
MoOxSy

c
MoS2

NiSx
MoOxSy NiO

NiOxSy j
g
i

MoS2 decorated
with NiSx d MoS2 decorated
h with NiSx
NiSx

MoS2 decorated
with NiSx
e
f

NiMoNTA
complex
k

Figure 7 Life cycle of a HDS catalyst: (a) Oxidic Mo and Ni/Co are bound to Al2O3; (b) partially sulfided catalysts, Ni/Co are sulfided, Mo oxysulfide
is still bound to the support; (c) partially sulfided catalyst and part of Mo Mo is no longer bound to Al2O3; (d) partially sulfided catalyst consisting of
small MoS2 slabs decorated with promoter Ni/Co; slabs are still anchored to the support; (e) fully sulfided catalyst with Mo being completely
sulfided, MoS2 slabs decorated with sulfidic Ni/Co become mobile; (f) deactivation due to sintering of decorated small MoS2 slabs to form larger slabs,
stacks, and crystals, edge surface area of the crystals is too low to accommodate all Ni/Co and separate Ni/Co sulfide crystals are formed; (g)
regeneration MoS2 crystals and some decorating Ni/Co sulfide get partly redispersed; (h) regeneration large MoS2 crystals are only oxidized on
the surface, Ni/Co sulfide crystals cannot redisperse on oxidation; (i) regeneration small structures are fully oxidized, nearly the original state is
reached; (j) high-temperature sulfidation with Mo being completely sulfided so fast that intermediate steps can be neglected; (k) NTA preparation route
NiMoNTA complex gets completely sulfided, and no linkages with Al2O3 are formed at any stage of the process.31

finely dispersed compounds such as CoMoO4/NiMoO4 was
also observed.48,49 It is assumed that Mo is bound to the
support via MoOAl bridges (see also below).5058 Several
crystalline compounds such as MoO3, CoO, NiO, Al2(MoO4)3,
CoAl2O4, NiAl2O4, CoMoO4, and NiMoO4 could be detected
in x-ray powder patterns but mainly only in very small
amounts if at all.
7.12.5 Analytical Tools for the Characterization
of the Catalysts
A large variety of analytic methods are used to characterize the
catalysts in different states starting from the as-prepared pre-
cursors to catalysts recovered after intense use. In the follow-
ing, some techniques are mentioned and some examples are
discussed in detail demonstrating the kind of information that
can be extracted by applying a distinct analytic method and
combinations of methods.
Chemisorption techniques are applied to receive information
on the active phase dispersion and on the type of sites present on
the catalyst surface. These methods include dynamic oxygen
chemisorption (DOC),59,60 CO chemisorption (COC),6169
and NO chemisorption (NOC).7088
Temperature-programmed techniques such as temperature-
programmed sulfidation (TPS),8995 temperature-programmed
reduction (TPR),96103 and temperature-programmed oxida-
tion (TPO)104107 are suitable methods to understand the
changes occurring during sulfidation of the catalyst and
regeneration.
Temperature-programmed desorption (TPD) or
temperature-programmed surface reaction (TPSR) are some-
times applied.90,93 In a TPD experiment, a spent catalyst is
heated and reaction intermediates and products desorbing
from the surface are detected as a function of temperature.
x-Ray photoelectron spectroscopy (XPS)110115 was also
often used for characterization of catalysts in different states.
It yields information on the elemental composition and the
oxidation state of elements. XPS is a surface-sensitive technique
and, with this method, it is possible to determine how well
particles are dispersed over a support.43,51,116130
Extended x-ray absorption fine structure (EXAFS) spectros-
copy and/or x-ray absorption near-edge spectroscopy (XANES)
are powerful methods to probe the local environment of atoms
and to estimate the oxidation state of the atoms. Hence,
detailed information on the distance, number, and type of
neighbors of the absorbing atom on a subnanometer length
scale can be acquired.131136
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts 295

SIMS, Rutherford backscattering spectroscopy (RBS), infrared 300400  C

MoO3 ! MoS2 [1]
(IR) and transmission electron microscopy (TEM),64,137140 H2 S=H2

Mossbauer emission spectroscopy,141143 and in situ laser

Raman spectroscopy (LRS)138,144,145 were also applied to charac-
terize catalysts in different states.
7.12.6 Formation of the Sulfide Phases
The finely dispersed oxides MoO3/polyoxomolybdates and CoO/
NiO are transformed during the sulfidation,99,104,105,137,146150
and a direct interaction between sulfidic MoS2 and Co/Ni is
generated.151156 If Mo oxide is fully sulfided, the catalyst consists
of well-dispersed MoS2 nanoparticles, and sulfidic Co/Ni deco-
rate the edges of the MoS2 slabs. The formation of large MoS2
and/or Co9S8 (Ni3S2, NiS) crystals is in most cases not
observed.141,150,157162
In cases where well crystalline compounds such as
Al2(MoO4)3, CoAl2O4 (NiAl2O4), CoMoO4 (NiMoO4), CoO
(NiO), and MoO3 are formed during calcinations, a full sulfida-
tion of these compounds is often not achieved57,98,99,163 applying
typical laboratory conditions such as gas phase, 673 K, 0.1 MPa,
10% H2S/H2, or commercial (e.g., liquid phase, /603 K,
/3.0 MPa, light oil spiked with S compounds such as,
dimethyl disulfide)164167 sulfidation conditions.
As outlined above, the different individual steps of the trans-
formation of the oxidic precursors into the final sulfided cata-
lysts were intensively studied with a variety of analytical tools. In
the following, the results of fundamental studies on the sulfida-
tion of oxidic precursors are discussed for selected examples.
However, the structure of the nanocrystalline respectively
amorphous oxidic catalyst precursor deposited on a support
such as g-Al2O3 is not known and one can imagine a large
variety of local MoO coordination environments generated in
an unpredictable way during the preparation of the catalyst
precursor. Hence, the investigation of real catalysts is a chal-
lenge requiring a combination of the results of different
advanced analytical methods.
7.12.7.1 Conversion of MoO3 into MoS2 Studied with XPS
and IRES
To reduce the complexity of the system, the sulfidation of
crystalline MoO3 was investigated with XPS and infrared emis-
sion spectroscopy (IRES).168 Such an investigation allows com-
parison of elementary reaction steps with known structural
features of MoO/MoS compounds. XPS is a surface-sensitive
spectroscopic technique, being sensitive to changes of the oxi-
dation state of the elements involved in the chemical reaction
(see also Chapter 9.16) and IRES supplies information on the
changes in MoO coordination during successive exchange of
O atoms by S (see also Chapter 9.36). During conversion of
MoO3 into MoS2, intermediate Mo oxysulfides are formed and
therefore the thermal decomposition of (NH4)2MoO2S2 as a
model compound was studied which decomposes according
to eqn [2] and forms an {MoOS2} oxysulfide169 on the way
to MoS2:
NH4 2 MoO2 S2 ! fMoOS2 g 2NH3 H2 O
[2]
fMoOS2 g ! 0MoS2 0

7.12.7 Conversion of MoO3 into MoS2 A sequence of XPS spectra of MoO3 after sulfidation in H2S/
H2 at different temperatures is shown in Figure 8, left. The first
Formally, the conversion of Mo oxide into the sulfide is a spectrum of crystalline MoO3 exhibits a single Mo 3d doublet
simple reaction according to eqn [1]: and the binding energy (Ebind) of 232.9 eV for the Mo 3d5/2

Mo 3d S 2s S 2p

400 C

300 C

250 C

200 C

150 C

100 C

50 C

25 C

MoO3
241 237 233 229 225 221 168 166 164 162 160 158
Binding energy (eV) Binding energy (eV)
Figure 8 Evolution of Mo 3d (left) and S 2p (right) x-ray photoelectron spectra of crystalline MoO3 with increasing sulfidation temperature.168
296 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

S S O
O H H H H H H
O O S
VI + H2S
Mo VI VI VI
Mo Mo Mo

Scheme 4 Attack on the terminal O atom by H2S to form intermediate species and release of H2O.

line is in accordance with Mo in formal oxidation state 6. O O S S

Sulfidation at low temperatures leads to pronounced changes VI VI VI VI
Mo Mo + H2S Mo Mo
in the spectra and successively further signals appear in the O O
- H2O
spectra indicating the formation of different intermediate spe- O O
cies. Sulfidation seems to be finished between 250 and 400  C
and the spectrum of the sample treated at the highest temper- S
S S
ature shows essentially one doublet and the binding energy of S
VI VI V V
the Mo 3d5/2 peak of 229.0 eV corresponds to Mo4 in Mo Mo Mo Mo
O O
MoS2.170 In the corresponding S 2p spectrum (Figure 8,
O O
right), there is only one doublet with Ebind 161.8 eV for the
S 2p3/2 core level in agreement with the sulfide S2 anion Scheme 5 Replacement of two terminal O atoms by two S atoms
present in MoS2. followed by oxidation of two S2 to S2 2 and reduction of Mo6 to Mo5.
For different temperatures between 25 and 300  C, all the
Mo 3d spectra can be explained on the basis of mixtures of Mo6 O O S S S S
+ 2H2S
and Mo4 doublets and an additional doublet with a Mo 3d5/2 VI VI IV
Mo 2H2O Mo Mo
binding energy of 231.1 eV, which is already observed after
sulfidation at room temperature (RT). A tentative assignment
was done to Mo5 present in oxysulfide intermediate phases. All Scheme 6 Exchange and subsequent redox reaction of edge metal
centers involving two terminal O2 atoms.
spectra contain Mo 3d signals in the partially sulfided state from
these three states in varying percentages giving evidence for the
stepwise transition of Mo6 in MoO3 to Mo4 in MoS2. MoVI Ot + H2 MoIV + H2O
The S 2p peak in Figure 8, right, can be explained as the
Scheme 7 Direct reaction between terminal O atom and H2 and
superposition of two doublets with Ebind 161.7 and 163.1 eV reduction of Mo6 to Mo4.
for the S 2p3/2 peaks. The signal located at 163.1 eV can be
assigned to a bridging S22 disulfide anion,138,139 while the sulfidation of solid MoO3, a deprotonation of H2S is not possi-
peak at 161.8 eV points to terminal disulfide and/or sulfide ble and H2S is the reactive component. Hence, the terminal O
anions. According to the XPS results, one of the first elementary atom is attacked by H2S to form intermediate species (Scheme 4)
steps during sulfidation involves a metalligand redox reaction and H2O is released, which is the driving force of the reaction.
leading to the formation of Mo5 and oxidized sulfur ligands Note that in this step the oxidation number of Mo is not
(S22). Secondary ion mass spectrometry (SIMS) measure- changed.
ments gave evidence that the sulfidation of MoO3 leads to In the next reaction step (Scheme 5), Mo6 is reduced
the formation of oxysulfides at the surface and Mo(IV) oxides to Mo5 by a metalligand redox reaction in which two
in the interior of the crystallites at sulfiding temperatures S2 anions bonded to adjacent Mo6 centers are oxidized
between RT and 100  C, while MoS2 is formed above 125  C. to S22: 2S22 ! S22 2e and the Mo centers are reduced,
For sulfur, two different chemical states are found and the most 2Mo6 2e ! 2Mo5. The proposed reaction scheme is clearly
likely species are S2 and S22 or SH, but the presence of supported by the XPS data and is in agreement with the IRES
elemental S can be excluded. The S22 or SH species disappear showing characteristic emission from disulfide species.
when the temperature is raised to 150200  C. Rutherford In the low-temperature regime, the occurrence of Mo4 in the
backscattering data yield S:Mo atomic ratios of 11.5 for XPS spectra may be explained by the reaction in Scheme 6,
T < 100  C and of 22.5 above 100  C.137 indicating a subsequent redox reaction with two terminal O22
Combining the results obtained on sulfidation of MoO3 and anions.
the model compound (NH4)2MoO2S2 yields the following pic- In principle, a direct reduction may occur according to
tures of individual reaction steps for the two different tempera- Scheme 7 but as was demonstrated by TPS measurements,97
ture regimes RT to 200  C (generation of Mo oxysulfide) and this type of reaction is more relevant at higher temperatures.
200400  C (transformation of Mo oxysulfide into MoS2). The Despite the S/Mo ratio in the XPS spectra measured after
terminal MoOt groups are relatively unstable and play an sulfidation at 50 and 100  C exhibiting very similar values,
important role in the uptake of sulfur. This instability is well the relative amount of the different sulfur ligands changed.
known from solution chemistry where dissolved molybdate The fraction of S2 and/or terminal S22 anions increased
species react with H2S in NH4OH to form tetrathiomolybdate: because of bridging S22 ligands. In parallel, the proportion
MoO42 4H2S ! MoS42 4H2O. In solution, the reaction of Mo5 increased, as opposed to the contribution of bridging
may proceed via H2S or SH formed in situ by deprotonation S2 anions. A possible explanation is the release of S from S22
of H2S, which depends on NH4OH concentration. For the bridges according to Scheme 8.
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts 297

H
S H
S
S
V V S
+ H2 V
Mo Mo V
O Mo Mo H 2S
O
O
O

S
V V
Mo Mo
O

O
2
Scheme 8 Reaction of a S2 entity with H2 forming two transient SH fragments, stabilization by H transfer (H2S complex), and subsequent emission
of H2S.

First, S22 reacts with H2 yielding two transient SH units.

Such SH units must be kinetically stabilized by, for example, S
S S
bulky ligands and sufficient electron density on Mo, which is S S
S Mo Mo S
highly unlikely. However, stabilization occurs easily by proton Mo Mo
Mo Mo
transfer to form H2S, which then immediately decomposes. Mo S S Mo
S S
After the emission of S from S22 units, the sterical hindrance S S
on the surface of the crystal is reduced so that new S22 groups
can be formed, leading to a higher S/Mo ratio in the XPS Scheme 9 A layer of MoS2 containing a m-S atom as defect.
spectrum of the sample sulfided at 150  C. In the IRES, the
low-intensity MoOt vibration indicates that most of the avail-
able MoOt moieties have reacted at 150  C. At 200  C sulfi-
dation temperature, the analysis of the surface composition TPS of MoO3/Al2O3
using the XPS spectrum demonstrates an increased degree of
sulfidation.
In the temperature range 200400  C, a major change of H2
composition is observed compared to that of the oxysulfides
Partial pressure (a.u.)

detected at lower temperatures. In this region, Mo6 and Mo5

species are mainly reduced to Mo4 and the signal of bridging
S22 in the XPS spectrum is absent. The formation of Mo4 is H2S
mainly due to the reaction of MoOt with H2 (Scheme 8)
followed by S uptake from the sulfiding atmosphere. The
finally formed MoS2 contains imperfections in the sulfur
planes and the IRE spectrum gives evidence that m2-S atoms
are present on MoSMo bridges. Such defects are accompa- H 2O
nied with a S deficiency and leads to a bending of the MoS2
layers (Scheme 9).
Indeed, bent MoS2 nanocrystallites are often seen in TEM
images (see below). It is well known that the formation of perfect 400 600 800 1000 1200
MoS2 requires temperatures of about 800  C. An important con- Temperature (K)
clusion of the study performed on MoO3 is that sulfidization of
Figure 9 TPS pattern of MoO3 supported on Al2O3 pretreated in Ar
MoO3 at about 400  C, typically for catalyst pretreatments, gen-
at 298 K.98
erates MoS2 exhibiting an appreciable amount of defects.

7.12.7.2 Temperature-Programmed Sulfidation of MoO3

The sulfidation of the MoO3 model compound in a H2S/H2/Ar
atmosphere as a function of temperature was studied with
TPS,97,98 supporting the findings presented above. In the TPS
pattern of MoO3 supported on Al2O3, negative peaks indicate
that the corresponding gas is consumed, and positive peaks
mean that the gas is released. As can be seen in Figure 9, at
400 K H2S is consumed and H2O is produced and no uptake of
hydrogen can be detected. The observations suggest that at low
temperatures an exchange of O by S atoms is the main reaction:
MoO3 H2 S ! MoO2 S H2 O
Increasing the temperature to T  500 K, a sharp hydrogen
consumption peak is seen while at the same temperature H2S
and small amounts of H2O are produced. These observations
may be interpreted as hydrogenation of excess sulfur formed by
the decomposition of MoO2S according to the equations:
MoO2 S ! MoO2 S
S H2 ! H2 S
298 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

Up to the final temperature, further oxygen is replaced by
sulfur until MoS2 is formed:
MoO2 2H2 S ! MoS2 2H2 O
One should keep in mind that TP experiments detect only
reactions accompanied by a net production or consumption of
gases. If the following two reactions would occur instantaneously,
no hydrogen signal would be detected:
MoO2 2H2 ! Mo 2H2 O
Mo 2H2 S ! MoS2 2H2
In any case, TPS is a powerful method to study the sulfida-
tion of the catalysts, giving useful information about the step-
wise conversion of the oxidic precursors into the final sulfide.
7.12.8 Quick EXAFS Spectroscopy Investigation of
the Local Environments in Ni-Promoted Molybdenum
Sulfide Catalysts
The changes of the local environment around Mo of a NiMo
catalyst deposited on high surface area silica during sulfidation
were investigated with a temperature-resolved Quick EXAFS
(QEXAFS) study (Figure 10).171 The sample was heated with
a heating rate of 6  C min1 in 10%H2S in H2 atmosphere. The
resulting Fourier transform of the Mo K-edge EXAFS spectra is
displayed in Figure 10.171 The sulfidation process can be
divided into four regions.
Up to about 110  C, Mo is present in the oxidic form; that
is, Mo is mainly surrounded by O atoms as seen by MoO
shells between 0.5 and 1.9 A and a MoMo shell at 3 A (values
not phase-corrected). Between about 110 and 225  C, the lon-
gest MoO shell and the MoMo shell disappeared, whereas
the shorter MoO can still be identified. A new signal appears
at about 2 A (not phase-corrected) caused by the presence of S
indicating the co-existence of MoO and MoS shells, with the
MoO shell corresponding to MoOt groups. Near the 225  C
region, a new signal at 2.5 A (not phase-corrected) develops
belonging to an intermediate product. The signal at about
2.5 A dominates the temperature range between 225 and
280  C and in this region only contributions of the MoS
and the MoMo shells are seen attributed to sulfur-rich phases
such as MoS3, which was also reported in other EXAFS
studies.172 Above about 310  C, the spectrum of the final
product MoS2 starts to form (region IV) and no significant
changes are seen up to the final temperature of 400  C. The
shoulder on the left side of the MoMo shell at about 3 A is
caused by the interference of the MoS and the MoMo shells.
The results are in accordance with those obtained from XPS,
RBS, and SIMS137,173 and slight differences of temperatures at
which the changes occur are due to the different experimental
setups of the methods. The intermediates formed during sulfi-
dation of the Mo precursor are those discussed above.
Due to the low loading of Ni of only about 1.3 wt%, EXAFS
data were collected with a lower heating rate of 3  C min1.
Fourier transform spectra (Figure 11) of the sulfidation of the
fresh catalyst show NiO and NiSi shells at 1.65 and 2.9 A
(not phase-corrected), respectively, while the NiSi shell dis-
appeared at 105  C. The transformation of the Ni oxide to the
5 a.u.
400
390
370
350
330

Surfidation temperature (C)

315
295
|FT[c(k)k3]| (a.u.)

QEXAFS Mo K-edge of NiMo/SiO2

280
260
Tsulf (C) 245
225
Fourier transform of k3 c(k)

400
400 210
400
400 Region IV
380 195
365
345
MoS2 175
330
315 160
295
280 Region III 140
260
245
225
MoS3-like 125
210
195 105
170 Region II
155 oxysulfides 90
140
125 70
110
90 50
70
55 Region I
40
35
Fresh
oxidic
Fresh
0 1 2 3 4
r () 0 1 2 3 4
r ()
Figure 10 Fourier transforms of the Mo K-edge Quick EXAFS functions
obtained during sulfidation of NiMo/SiO2. The four sulfidation regions are Figure 11 Fourier transforms of the Ni K-edge Quick EXAFS functions
shown.171 measured during the sulfidation of NiMo/SiO2.171
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
Ni sulfide is seen in the spectra by a shift of the signal by about
0.1 A in the distance of the first shell. Definitive sulfidation
occurs between 125 and 140  C and no further alteration can
be detected in the spectra up to 400  C. That means full sulfi-
dation is achieved for Ni at a significantly lower temperature
than for the Mo oxide.
7.12.9 Sulfidation of Co-Promoted Molybdenum
Oxide in the Presence of a Chelating Molecule
Investigated with XPS
The situation is similar for CoMoS catalysts: Co is sulfided at
low temperatures and stable bulk sulfides such as Co9S8 are
formed as large crystals while sulfidation of Mo only starts at
moderate temperatures. Hence, it is surprising that a phase is
formed in which Co atoms decorate the edges of MoS2 slabs
and not a mixture composed of separated Co9S8 and MoS2.
To retard sulfidation of Co or Ni, chelating agents such as
NTA, EDTA, ethylenediamine (EN), or 1,2-cyclohexanediamine-
N,N,N,N-tetraacetic acid (CyDTA) are added during catalyst
preparation,171,174178 and the as-prepared precursor mixture
is not calcined in order to leave the complexes formed with
the ligands intact. The effect of NTA addition on the formation
of the sulfides was studied with XPS on a CoMoS model catalyst
deposited on SiO2. In the absence of NTA, Co is transformed
into Co9S8 at very low temperatures (Figure 12, left).
Even at RT, sulfidation of Co in 10% H2S/H2 atmosphere
starts as can be seen by the appearance of a second doublet at
779.3 eV corresponding to the binding energy of bulk Co9S8.
Increasing the temperature leads to disappearance of the dou-
blet at 782.3 eV and only the signal at 779.3 eV remains, which
means that cobalt has completely transformed into the sulfidic
state. The addition of NTA retards sulfidation of Co and the
Sulfidation
temperature Co 2p
400 C
300 C
200 C
150 C
100 C
50 C
25 C
Unsulfided
Shake up Co 2p1/2 Shake up Co 2p
814 809 804 799 794 789 784 779 774 810 805 800 795 790 785 780 775 770
Binding energy (eV)
Figure 12 Co 2p XPS spectra of CoOx/SiO2 model catalysts (left) and Co 2p XPS spectra of Co(NTA)/SiO2 model catalysts as a function of sulfidation
temperature.173
299
transformation to the sulfidic state starts above about 150  C
and is completed at 225  C. Below 150  C, the typical signal of
the oxide is seen and at higher temperatures the doublet of
sulfidic cobalt appears (Figure 12, right). Recording the evolu-
tion of the N 1s signal as a function of sulfidation temperature
demonstrates that the intensity starts to decrease and the signal
is absent when sulfidation of cobalt is complete. These obser-
vations suggest that NTA is decomposed and Co ions are
available for sulfidation.
Sulfidation of Mo oxide calcined at 500  C in the absence of
NTA (Figure 13) starts at a higher temperature than the forma-
tion of Co sulfide. In the as-prepared state, the binding energy
of the Mo 3d5/2 core level matches well with Mo6. On increas-
ing the temperature, MoO3 is successively transformed into
MoS2 as discussed above and the typical spectrum of MoS2 is
seen with a binding energy of 228.8 eV for the Mo 3d5/2 signal.
The Co 2p3/2 and Mo 3d XPS spectra of a CoMo catalyst
with an atomic ratio of 0.3 calcined at 500  C are shown in
Figure 14, right. In the unsulfided state, the binding energy of
the Mo 3d5/2 signal at 232.7 eV is typical for oxidic Mo in
oxidation state 6. After sulfidation at T > 175  C, a Mo 3d
doublet appears in the spectra, typical for MoS2. The results
clearly show that sulfidation of Mo oxide occurs between
T  50 and 175  C and several intermediates are formed con-
taining Mo6, Mo5, and Mo4 species. At about 150  C, the
doublet of Mo4 appears in the spectrum and at T > 175  C the
signal of Mo6 disappears.
The sulfidation behavior of Co is shown in Figure 14. In the
as-prepared state, a Co 2p doublet is seen at 781.8 eV with
shake-up satellites at higher binding energy typical for Co2
oxide. On increasing the temperature above 175  C, the spec-
trum shows a Co 2p doublet at a binding energy of 778.7 eV
typical for Co2 in a sulfidic environment. Shake-up satellites
occurring at higher binding energy are less pronounced
Sulfidation Co 2p
temperature
400 C
300 C
225 C
200 C
150 C
100 C
Unsulfided
3/2
Binding energy (eV)
300 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

compared to oxidic Co. According to the XPS spectra recorded

Sulfidation
temperature Mo 3d over temperature, the sulfidation of Co starts at T > 50  C and
is completed around T 150  C. Summarizing the results, in a
400 C mixed CoMo/SiO2 catalyst Co sulfide is formed between 50
and 150  C and Mo is sulfided between 100 and 200  C. Both
300 C regions overlap to an appreciable extent, but sulfidation of Co
proceeds faster than that of molybdenum.
250 C

200 C
7.12.10 Determination of the State of Co in CoMoS
Phase Using In Situ Mossbauer Emission Spectroscopy
175 C
It is well documented that Co-promoted MoS2 catalysts are in
general much more active for HDS than unpromoted MoS2.
Hence, the key question is to determine the chemical state of
150 C
Co and the structural environment in sulfided CoMoS catalysts
100 C supported on Al2O3. Mossbauer spectroscopy probes the local
environment of the absorber as well as the chemical state
50 C simultaneously. Because Co is not a suitable Mossbauer
nucleus the experiments were performed in the emission
25 C mode, that is, the measured spectrum is not that of Co but of
the decay product Fe. Radioactive 57Co was used as a promoter
Mo 3d3/2 during catalyst preparation and the 57CoMoS catalyst was
Unsulfided
Mo 3d5/2 investigated under in situ conditions so that the Co spectrum
and the changes could be directly correlated to the activity of
240 238 236 234 232 230 228 226 224 222
the catalyst. The Mossbauer spectra of the CoMoS catalyst are
Binding energy (eV) compared with that of Co model compounds for which the
Figure 13 Evolution of the Mo 3d XPS spectra of MoO3/SiO2 model state and the environment of Co are known.141,179 The analysis
catalysts with increasing sulfidation temperature.173 of the Mossbauer spectra of the sulfided CoMoS/Al2O3

Sulfidation Mo 3d Sulfidation
temperature Co 2p
temperature

400 C

400 C
300 C
300 C
250 C
250 C
200 C

200 C 175 C
175 C
150 C 150 C

100 C

100 C
50 C

50 C 25 C
25 C
Unsulfided

Unsulfided

240 238 236 234 232 230 228 226 224 222 814 809 804 799 794 789 784 779 774
(a) Binding energy (eV) (b) Binding energy (eV)
Figure 14 The Mo 3d (a) and Co 2p (b) XPS spectra of a CoOx/MoO3/SiO2 catalyst calcined at 500  C during sulfidation at different temperatures.173
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts 301

Co:Al2O3 Co9S8
Co 40
Mo Co-Mo-S
Co9S8
0.09 (a) 30

mg Co/g Al2O3
20
CoMos
0.27 (b)

10 Co:Al2O3
Absorption

0.53 (c) (a)

140

120
0.75 (d)

kT (cm3 min-1 g-1)

100

80

1.19 (e) 60

40

20

-6 -4 -2 0 2 4 6
Velocity (mm s-1)
Figure 15 In situ Mossbauer emission spectra of 57Co in sulfided
CoMo/Al2O3 catalysts with different Co:Mo ratios.141
catalysts (Figure 15) essentially revealed contributions from
Co ions located in the Al2O3 support, bulk Co9S8 dominating
in catalysts with a high Co content, and a new formerly
unknown state, which is labeled in Figure 15. This new state
is best seen in sulfided CoMoS with a low Co content. The
presence of this new phase correlates with the catalytic activity
for the HDS reaction (Figure 16).
Hence, one may conclude that this CoMoS phase is itself
catalytically active or is closely related to the active site of the
catalyst. This is because the catalytic activity of MoS2 is associ-
ated with edges and it is highly likely that the CoMoS phase is
located at or near the edges. This assumption was verified with
Auger spectroscopy and transmission electron microscopy
clearly demonstrating that the basal planes of MoS2 do not
contain measurable/detectable amounts of Co while the edges
of the crystallites contain Co besides S and small amounts
of O.180,181
In further MES experiments, it was demonstrated that Mo is
not required to obtain the CoMoS MES spectrum. For Co/
Al2O3, CoMo/AI2O3, Co/C, and CoMo/C, Co sulfide species
exhibiting the CoMoS type features such as quadrupole
splitting (QS) were observed (Figure 17).
Furthermore, the temperature dependence of the MES spec-
tra of the CoMoS phase is identical with that of the Co sulfide
phase present in CoMo/AI2O3 and Co/C catalysts.183 The
important conclusion of these observations is that the
(b)
0.2 0.4 0.6 0.8 1.0 1.2 1.4
Co
Mo
Figure 16 Correlation between the activity of CoMo/Al2O3 catalysts for
the HDS reaction, expressed in the reaction rate constant kT, and the
cobalt phases observed in Mossbauer spectra.141
Co sulfide species are essentially the same, and only the parti-
cle sizes and/or ordering may differ.157,183185 Other impor-
tant findings of the MES experiments are that the value of QS of
the Co phase depends on the Co content in the catalysts and on
the sulfidation temperature with large QS values indicating the
presence of very small Co sulfide units and small values are
related to larger well-ordered Co sulfide particles.
7.12.11 The Local Structures of Co/Ni and Mo in
CoMoS and NiMoS Phases Determined with EXAFS
The local environment of Mo and the promoter atoms were
investigated with EXAFS/XANES in different states of the cata-
lysts under in situ conditions allowing in situ sample
presulfiding151,153,154,156,186189 in a mixture of e.g., H2S/H2
at atmospheric pressure and at 673 K. Because industrial
CoMoS catalysts are supported on Al2O3, such catalysts con-
tain at least two Co species such as Co in the Al2O3 lattice and
in the CoMoS phase. Hence, the CoMoS catalysts are deposited
on carbon for the EXAFS investigations.
The radial distance distribution (Fourier transform) of
MoS2, sulfided Mo, and a CoMoS catalyst supported on carbon
is shown in Figure 18.153 Evaluation of the data of MoS2 yields
302 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

Co(2.25)Mo(6.84)/C Co(2.45)/C
4.94 5.18
(S, 373 K)

Intensity (106 counts)

Intensity (106 counts)

4.86 5.11
3.06 2.83
(S, 473 K)

2.99 2.77
3.27 3.68
(S, 573 K)

3.20 3.58
3.27 3.64
(S, 673 K)

3.20 3.54
T = 300 K T = 300 K

-6 -4 -2 0 2 4 -6 -4 -2 0 2 4
1)
Doppler velocity (mm. s Doppler velocity (mm. s1)

Figure 17 MES spectra of carbon supported Co and CoMo catalysts measured at 300 K after various sulfidation treatments.182

30 Mo-Mo
10
MoS2 Co9S8
(3.18 )
Mo/C Co-S + Co-Co (1) Co/C
CoMo/C
Co-Mo/C

Mo-s
(2.41 )
Abs (FT)

Abs (FT)

15

Mo-Mo
(6.32, 6.41 ) Co-Co (2)
Mo-S Mo-Mo
(3.98 ) (5.47 )

0 0
0 2 4 6 8 0 2 4 6 8
r () r ()

Figure 18 Magnitude of the Fourier transforms of the MoS2 reference
compound (solid line), the Mo/C catalyst (dotted line), and the CoMo/C
catalyst (dashed line).153
a coordination number (CN) of 6 for Mo with the S atoms at a
distance of 2.41 A. The next-nearest neighbors are six Mo
atoms at a distance of 3.16 A. The reduced intensity of the
first and the second peak in the spectrum of the two catalysts
is an indication for a reduced number of neighbors; especially
for the MoMo coordination the number is lower than 6. To
explain these observations, a structural model was developed
taking into account the reduced CNs in the catalysts, leading to
the average crystallite size for MoS2 of 1.53 nm. Such small
particle crystallites will expose a relatively large number of
catalytically active edge sites.
The absolute part of the Fourier transform (FT) spectrum of
Co9S8, of Co/C and CoMo/C samples, is displayed in Figure 19.
In the crystal structure of Co9S8, three different CoS coor-
dinations are present, which are denoted in the spectrum as
one overall CoS coordination. The two CoCo distances at
about 3.5 A are given as one overall CoCo(2) coordina-
tion (indexed (2) to distinguish these from the first CoCo
coordination at 2.505 A). The first peak in the spectrum of
Figure 19 Fourier transforms of Co9S8 and the sulfided Co/C, CoMo/
C, and CoMoS/C catalysts.131
Co9S8 is caused by a combined CoS and CoCo(1) coor-
dination and the second peak is attributed only to the Co
Co(2) distance. In the spectra of Co/C and CoMo/C
samples, the first peak is slightly shifted to lower r values.
Overall, the two sulfided catalysts show identical features as
Co9S8 but with reduced amplitudes.
Despite the overlap of the CoCo/CoS nearest neighbors
in the magnitude of the Fourier transform, a clear differentia-
tion between these two contributions is seen in the imaginary
part of the spectrum (Figure 20). In addition, a scattering
contribution from Mo neighbors can also be identified.
The detailed analysis of the spectra yields a CoS distance of
2.2 A with a high sulfur CN of 6.2(1.3). The Co atom is further
surrounded by 1.7(0.35) Mo neighbors at a distance of 2.8 A.
Using these geometric parameters, a model was developed
fitting with a structure where Co is located on the edge of a
MoS2 crystallite in the same plane as the Mo atoms. It was
proposed that Co is coordinated to four S atoms on the edge,
which are slightly displaced from their bulk positions. Accord-
ing to this model, Co is not exactly on the position of a
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
10
Co-S Co9S8
Co-Co (1) Co/C
CoMo/C
CoMoS/C
Co-Co (2)
Im (FT)
0 473 K
b r ()
-10
0 1 2 3 4 Figure 21 W EXAFS Fourier transforms of NiW/Al2O3 after sulfidation at
r ()
Figure 20 Imaginary part of the Fourier transforms of Co9S8, sulfided
Co/C, CoMo/C, and CoMoS/C catalysts.131
Mo atom of a bulk crystal, which would give a CoMo distance
of 3.16 A, but is shifted toward the MoS2 particle yielding the
experimentally determined CoMo separation of 2.8 A. The
S CN of 6 is achieved by placing two S atoms on Co to form
a distorted trigonal prism together with the four S atoms of the
MoS2 edge. A slightly different model was proposed on the
basis of density functional theory (DFT) calculations indicating
that some reconstruction may play a role but the model is also
in agreement with the EXAFS data.
7.12.12 Sulfidation of NiW Catalysts Supported
on Al2O3
The extent of sulfidation of W is crucial for catalytic perfor-
mance.190192 It is well documented that the sulfidation of W
in NiW/Al2O3 occurs at a lower rate than that of Mo of Al2O3-
supported CoMo or NiMo catalysts. It was pointed out that the
relatively strong WOAl bond94 or the stronger WO bond
compared to the MoO bond are responsible for this behavior
and hence higher temperatures are required to fully transform W
oxide to W sulfide.193195 It is well documented that the final
degree of W sulfidation strongly depends on the calcination and
the sulfidation temperatures.191,192,196 At the working tempera-
ture of 673 K, Al2O3-supported NiW catalysts are only partially
converted to WS2 and other species such as WS3 or oxysulfides
are still present.191,192 In contrast to Mo-based HDS catalysts,
the addition of Co as promoter seems to have no positive
effect,193 while Ni addition enhances the catalytic performance
of the W-based catalysts. For NiW-based materials, it was
observed that Ni can easily be sulfided and sulfidation is com-
pleted before W sulfidation has started.193,197 At low sulfidation
temperatures, Ni sulfide phases are present as a separated com-
pound or as small Ni sulfide particles interacting with WOxSy
phases. At higher sulfidation temperatures, WS2 slabs are
formed accompanied by a redispersion of Ni sulfide species
and NiWS-type structures are formed,198201 which was
shown by a 57Co Mossbauer emission study on 57Co-doped
NiW/Al2O3.202 As for Co(Ni)MoS-based catalysts, two types of
303
WS
15
II 673 K
|FT [c(k)k3]| (a.u.)
573 K
10
5 W-O
I
373 K
300 K
0
0 2 4 6 8
indicated temperatures. Intervals I and II: WW contributions of WO3 and
WS3, respectively (R not corrected for phase-shift).203
NiW catalysts are proposed: a type I NiWS phase obtained at
low-temperature sulfidation exhibiting residual oxygen linkages
showing a high hydrogenation activity and a type II phase
formed at higher sulfidation temperatures exhibiting a high
HDS activity.191
The sulfidation behavior of NiW materials supported on dif-
ferent materials (g-Al2O3, carbon, and amorphous silica-alumina
(ASA)) was studied applying different analytic tools and some
selected results are discussed in more detail. The supports were
impregnated using aqueous solutions of ammonium metatung-
state ((NH4)6W12O39xH2O) and nickel nitrate. The resulting
material was dried at 383 K and subsequently calcined at elevated
temperature in air. The sulfidation was done stepwise in a mixture
of 10 vol.% H2S in H2 at 300, 373, 473, 573, 673, and 773 K for
1 h, atmospheric pressure. For catalytic measurements, the cata-
lysts were sulfided at 673 K for 2 h.203
7.12.12.1 x-Ray Absorption Spectroscopy Performed
at the W LIII Absorption Edge
The Fourier transforms of a catalyst that has been calcined at
673 K prior to sulfidation are shown in Figure 21. At low
sulfidation temperature, the spectra are dominated by two
WO contributions at 1.76 and 1.91 A corresponding to the
WO bonds in the first coordination shell of WO3. The peaks
in the region labeled I (Figure 21) are caused by second shell W
and third shell O backscatters. These findings indicate that at
low sulfidation temperatures an oxidic tungsten phase is pre-
sent with a WO3-like structure.
The sulfidation of W starts at 573 K and a WS contribution
at a distance of 2.42 A with a CN of 0.9 can be detected in the
spectrum. In this spectrum, a second WW contribution
(2.71 A, CN 1) is seen indicating the presence of a WS3-type
structure (WS 2.40 A, CN WS 5; WW 2.75 A, CN
WW 2). While at 673 K the WO contributions are clearly
reduced, contributions of the oxidic W phase are still present.
The characteristic WW distance in WO3 at 3.73 A cannot be
detected in the spectrum (T 673 K) suggesting that the origi-
nal W oxide phase has largely disappeared. In addition, an
increase of the WS coordination from 0.9 to 1.6 is observed
304 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
while the WW CN associated with the WS3-type structure
decreased from 1.0 to 0.2. The lower than expected WS CN
for a-WS3 indicates that sulfidation of the WS3-type phase is
not complete at 673 K. A small WW contribution at 3.16 A
points to the onset of WS2 formation, and it can be assumed
that a partial transition from WOxSy and WS3-like phases to
WS2 already occurred before complete sulfidation to WS3 was
achieved. In accordance with results reported earlier,95,193 the
W species formed during the sulfidation procedure may be
regarded as a WOxSy phase.
If a NiW catalyst supported on Al2O3 is sulfided in a single
step at T 673 K for 2 h, much more WS2 is present compared to
the stepwise sulfidation. In addition, the amount of oxygen is
also substantially lower. A higher calcination temperature of
823 K and a single-step sulfidation leads to a lower degree of
sulfidation and the catalyst contains more oxygen and less sulfur.
Several NiW catalysts with a varying Ni:W ratio were cal-
cined at 823 K and then sulfided at T 673 K at a total pressure
of 15 bar (H2S/H2 ratio: 0.1). The analysis of the EXAFS data
indicate that in all catalysts WS2 slabs are present pointing to a
higher degree of sulfidation at higher pressure. With increasing
Ni:W ratio, sulfidation seems to be more complete and at low
Ni:W ratio the low WS CNs suggest the presence of some
oxygen. A larger number of WW CNs at higher Ni:W ratios
may be interpreted as a higher order of the crystallites and a
more complete sulfidation.
7.12.12.2 Ni K Edge x-Ray Absorption Spectroscopy
In Figure 22, the Fourier transform functions at the Ni K edge
of stepwise sulfided Al2O3-supported NiW catalysts (left) and
the near-edge spectra (right) are displayed. After treatment at
T 300 K, the spectrum contains peaks of four different con-
tributions. At a distance of 1.98 A a NiO contribution (CN
4.2), at 2.17 A a NiS contribution (CN 2.3), and two NiNi
contributions at larger distances of 2.60 A (CN 0.8) and 3.02 A
(CN 1.5) are seen. Both the NiO distance of 1.98 A and the
NiNi separation of 3.02 A match well with the geometric
parameters of NiO (NiO 2.08 A, NiNi 2.96 A). The
other observed peaks in the spectrum point to a NiOxSy species
formed during sulfidation. Compared to bulk NiS (Ni
S 2.39 A), the NiS distance of 2.17 A is much smaller. The
NiS
15
|FT [c(k)k3]| (a.u.)
NiNi
673 K
10
573 K
473 K
5 NiO
373 K
NiNi
300 K
0 0.0
0 2 4 6
r ()
Figure 22 (Left) Ni EXAFS Fourier transforms of NiW/Al after sulfidation at indicated temperatures and (right) the XANES region at the Ni K edge during
sulfidation (distance R not corrected for phase-shift).203
occurrence of NiNi contributions indicates that NiO and Ni
(oxy)sulfide are not well dispersed over the support surface. In
the XANES spectra, a continuous decrease of the area of the
white line is accompanied by a decrease of the NiO CN
suggesting a progressive Ni sulfidation. In the spectrum of the
sample sulfided at 573 K, the NiO contribution is absent and
most of Ni is sulfided. In the XANES spectra (Figure 22, right),
a pre-edge peak near 8333 eV characteristic for the 1s to 3d
transition shows an increase of the intensity related to a change
of the symmetry around Ni.
For the Ni sulfide obtained at 573 and 673 K, the bond length
for NiS amounts to 2.21 A and that for NiNi is 2.93 A. Espe-
cially the NiNi separation does not match with any well-known
bulk Ni sulfide such as Ni3S2 (NiNi 2.50 A, 3.80 A), NiS
(NiNi 2.68 A), NiS with millerite structure (NiNi 3.16 A),
NiS2 (NiNi 3.924.12 A), or Ni3S4 (NiNi 2.50 A), and
together with the NiS bond length of 2.21 A the formation of
the NiWS structure may be assumed.
The results of the studies can be schematically summarized
as shown in Figure 23.
Tungsten oxide is more difficult to be sulfided due to the
stronger metaloxygen bond than that of Mo oxide in Co(Ni)
Mo catalysts. Hence, over a broad range of sulfiding conditions
incomplete sulfided tungsten phases are observed and these
may be viewed as oxysulfides WOxSy. For these materials,
indications are found for the presence of WO3, WS3, and WS2
with differing amounts of oxygen and sulfur atoms. Interaction
between WO3 and the support g-Al2O3 makes it more difficult
to sulfide W and calcination even increases the interaction with
the support. Investigations with carbon as support demon-
strate a real support effect and because metalsupport interac-
tions are negligible or even absent, WS2 formation is more or
less complete at 673 K for NiW/C, while alumina or amor-
phous silicaalumina supports require sulfidation at higher
temperatures being accompanied by a loss of WS2 dispersion.
Because sulfidation of W is very slow, Ni sulfide particles tend
to aggregate as larger crystallites, being not segregated from the
WOxSy phase. Under typical sulfiding conditions (673 K,
1 bar), not all NiS species are redispersed to the WS2 slabs.
This is caused by the sulfidation sequence (NiO is sulfided
earlier than WO3) and because not all tungsten has converted
to WS2. Compared to sulfided Co(Ni)Mo catalysts with
0.6
300 K
373 K
473 K
Whiteline
0.4 573 K
mx (a.u.)
673 K
0.2
Pre edge peak
8 8300 8320 8340 8360
E (eV)
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts 305

NiO
WO3
Ni Ni Fresh
Al2O3

NiOxSy
WO3

15 bar sulfidation
Ni Ni
373 K

NiS
1 bar sulfidation

WO3
Ni Ni
473 K

NiSWOxSy NiWS
S S
WS2 W Ni
NiS WOxSy S S

573 K

NiSWOxSy NiWS NiWS

S S

NiS WOxSy Ni WS2 W Ni Ni

WS2 Ni Ni WS2 Ni
S S

673 K

NiS NiWS
Ni
WS2 Ni

923 K

Figure 23 The sulfidation process of g-Al2O3-supported NiW catalyst. Left: sulfidation at atmospheric pressure; right: sulfidation at 15 bar. Fresh
material contains Ni and W oxide; between 373 and 573 K small NiS crystallites are formed being attached to WOxSy (NiSWOxSy), which may be a
precursor to the NiWS phase. At 673 K and 1 bar, NiSWOxSy is partially converted to NiWS. NiS particles are redispersed to smaller crystallites
over the WS2 edges during this transformation process. At 973 K and 1 bar, full sulfidation is achieved and a loss of WS2 dispersion as well as an
increased crystallinity of the WS2 slabs occurs. Sulfidation at 15 bar and T 673 K is more complete, WS2 is reasonably good dispersed, and more
NiWS phase is formed.203

promoter ions being present in the Co(Ni)MoS phase, NiW
catalysts may be regarded as atomically dispersed Ni species in
NiWS phases together with less dispersed NiSWOxSy
phases.
7.12.13 Probing the Active Sites: CO and NO
Adsorption
Different spectroscopic techniques are used to obtain informa-
tion regarding the surface binding sites of HDS catalysts.30 A
powerful approach is to monitor the changes of adsorption of
small probe molecules such as NO7087 or CO6169 with IR
spectroscopy. Applying IR studies of NO adsorption allowed
identifying different adsorption complexes for unpromoted
and promoted catalysts.69,7173 It was also shown by IR NO
adsorption experiments and EXAFS investigations that NO
molecules are adsorbed onto MoS2 edges and not onto the
basal planes in sulfided Mo/Al2O3 catalysts.142
In the following, the results of a CO IR study are discussed in
detail.69 Different g-Al2O3-supported Mo catalysts calcined at
773 K exhibiting different Mo loadings (4.2, 9.8, and 11.9 wt%,
abbreviated as Mo4/Al, Mo9/Al, and Mo12/Al), Co-promoted
catalysts (impregnation with Co nitrate, calcinations at 773 K,
1.6, 3.1, and 4.7 wt% on the Mo9/Al catalyst), and a Mo catalyst
supported on silica (Mo7Si, 7.2 wt% Mo, calcination tempera-
ture 623 K) were pressed into discs and placed in an IR cell.
Then, sulfidation was done (1) under H2S/H2 flow or (2) under
H2S/H2 flow followed by a post-treatment under pure H2 flow.
IR CO adsorption experiments were done at 100 K.
Three absorptions are seen in the IR spectrum of CO on
the sulfided Mo7/Si catalyst (2158, 2120, and 2075 cm1)
(Figure 24) with the band at 2158 cm1 being caused by the
interaction of CO with silanol groups, while the remaining
absorptions are due to CO adsorption onto unpromoted Mo
sites of the sulfide phase. The spectra of CO adsorbed onto
catalysts with different Mo loading are shown in Figure 24.69
The peaks located at 2189 and 2156 cm1 are caused by CO
adsorption onto Al3 vacancies and hydroxyl groups of alu-
mina, while the band at 2110 cm1 is characteristic of CO
interacting with Mo sites. This absorption increases with Mo
loading and in the same direction the two former peaks
decrease because of a higher coverage of the support by MoS
catalyst. A band should appear at 2075 cm1 in the IR
306 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

2110

A 2120

0.1 A
2156
0.01 A
2158 2075
2189

Mo12/Al
2075

Mo9/Al

Mo7/Si
Mo5/Al

2250 2200 2150 2100 2050 2000 1950 2250 2200 2150 2100 2050 2000 1950
(a) Wavenumbers (cm1) (b) Wavenumbers (cm1)

Figure 24 (a) IR spectra of CO adsorbed onto sulfided Mo catalysts supported on silica and (b) on g-Al2O3.69

spectrum for the catalyst with the highest Mo loading (Mo12/

A
Al). Compared to Mo7/Si, the band related to the Mo sites on 0.05 A
Mo/Al catalysts is shifted toward lower wavenumber, whereas
the absorption at 2075 cm1 is not significantly affected.
2156
MoS2-based catalysts often display a hexagonal shape and
two types of crystallographic edge planes (1 0  1 0) and 2190
(1 0 1 0) can be distinguished called the molybdenum and Co5Mo9/Al
the sulfur edges. The nature of these two MoS2 edges may
explain the presence of the two CO absorptions: the absorp-
Co3Mo9/Al
tion at 2075 cm1 is caused by CO in interaction with Mo
atoms on the sulfur edge and the peak at 21102120 cm1 is
related with CO adsorbed onto Mo atoms on the molybdenum 2070
Co1Mo9/Al
edge. These assignments are in agreement with DFT calcula-
tions where a difference of the two absorption bands of
40 cm1 was proposed67,68 in good agreement with 35 cm1 2110
for the Mo/Al catalyst and 40 cm1 for the Mo/Si catalyst.
Mo9/Al
IR CO spectra of the Co-promoted CoMo9/Al catalysts
(Figure 25) show a new peak at 2070 cm1, which increases
in intensity with increasing Co content up to 3.1 wt% Co, while 2250 2200 2150 2100 2050 2000 1950
the band characteristic of unpromoted Mo sites (2110 cm1) Wavenumbers (cm1)
decreases. Increasing the Co content to 5 wt% does not lead to a
Figure 25 IR spectra of CO adsorbed onto CoMo9/Al catalysts.69
further intensity gain of the absorption at 2070 cm1. For a
sulfided Co/Al2O3 catalyst (3 wt% Co), an absorption at
2094 cm1 and a shoulder with low intensity at 2056 cm1 is
observed. Therefore, the band in the IR spectra of the Co- The results demonstrate that for CoMo/Al catalysts the
promoted catalysts can be assigned to Co-promoted Mo sites, concentration of unpromoted Mo sites is reduced with increas-
in agreement with Ref. [63]. For the highest Co content, that is, ing Co content while in the same direction new Co-promoted
for an atomic ratio for Co/(Co Mo) of 0.48, the peak at sites are generated. However, the sum of the amount of unpro-
2110 cm1 is still observable suggesting that promotion of the moted and promoted sites does not vary significantly as func-
MoS2 slabs by Co is not complete in the CoMo/Al2O3 catalysts. tion of Co content. This observation suggests that the creation
The presence of Co also results in a decrease of the absorptions of one promoted site leads to the loss of about one unpro-
caused by CO-support interactions and this decrease is more moted Mo site. Interestingly, by increasing the atomic ratio of
pronounced at higher Co loading. A possible explanation is that Co/(Co Mo) above 0.36 the number of Co-promoted sites is
an increasing amount of cobalt spinel or cobalt sulfide phase is not increased, and the amount of unpromoted Mo sites almost
formed at the higher Co contents. does not decrease. These observations indicate that above a
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
2070
A 2110
2055
0.1 2156
2190 Co5Mo9/Al
Co3Mo9/Al
Co1Mo9/Al
Mo9/Al
2300 2200 2100 2000 1900
(a) Wavenumbers (cm1)
Figure 26 The influence of the H2 post-treatment at 573 K on the CO adsorption behavior on sulfided Mo/Al and CoMo/Al catalysts. (a) IR spectra of
CO adsorbed onto the Mo9/Al catalyst after sulfidation (dotted line) and after sulfidation and H2 treatment (full line). (b) Difference spectra derived
from IR spectra of CO adsorbed after sulfidation and H2 post-treatment minus spectra measured after sulfidation.69
certain Co/(Co Mo) ratio, additional Co atoms are not sig-
nificantly deposited on MoS2 edges.
In Figure 26, the IR CO spectra of the catalysts after a
post-treatment in H2 are shown and the difference spectra
are displayed in Figure 26, right. For Mo9/Al, the treatment
leads to an increased intensity of the band at 2110 cm1 (Mo
sulfide sites) and to the appearance of a new absorption at
2098 cm1. It seems also that the H2 post-treatment generates
additional Al3 vacancies and hydroxyl groups (bands at 2156
and 2190 cm1), which may be explained by breaking of
AlOAl bridges during H2 treatment creating new Al3 vacan-
cies and OH groups. A detailed analysis of the spectra without
and with H2 treatment yields an increase of Mo sites by a factor
2.3 after H2 treatment (Mo12/Al show a slightly smaller
increase). Reversibility of the effects of the H2 treatment was
checked by heating the catalysts in a H2S/H2 atmosphere at
573 K for 2 h. The IR spectra are very similar to those of
catalysts not treated in H2, and especially the peak at
2098 cm1 totally disappeared. The H2 post-treatment at
573 K removes S atoms from the MoS2 edge slabs significantly,
increasing the amount of accessible Mo sites. The observation
that Mo12/Al exhibits only an increase of such sites by a factor
of 1.8 may be due to larger size of the MoS2 slabs that are less
easily reduced by H2.
The H2 post-treatment also generates additional sites on the
CoMo/Al catalysts (Figure 26). Independent of the Co content,
the amount of unpromoted Mo sites increases by a factor of
about 1.8 being slightly lower than for Mo9/Al (factor of 2.3).
This increase can be explained by the generation of Mo sites
characterized by the band at 2098 cm1. For the concentration
of promoted sites, a factor of 1.3 is estimated as being smaller
than for unpromoted sites, and for the catalyst with the highest
Co content the factor is only 1.1. Because the increase of
307
2098
A 2070
2110 2055
0.05
 0.75
2300 2200 2100 2000 1900
(b) Wavenumbers (cm1)
Co-promoted sites by H2 treatment is smaller than that of
unpromoted Mo sites, it is most likely that most of the Co-
promoted sites are already generated during the sulfidation
procedure, whereas unpromoted Mo sites are more difficult
to be created.
In the IR spectra, no additional absorptions could be
detected after the H2 treatment excluding the creation of new
adsorption sites, that is, the environment of Co sites is not
modified by the reductive treatment. Like for unpromoted
catalysts, the number of Al3 vacancies and OH groups
increased after the H2 treatment.
The adsorption properties of sulfided CoMo/Al2O3 catalysts
were also investigated applying in situ Fourier transform infra-
red spectroscopy (FTIR) spectroscopy combined with NO
adsorption (Figure 27).73,204 The IR absorptions associated
with NO adsorbed onto Co sites at 1850 cm1 start to evolve
by the addition of small amounts of Co to a Mo/Al2O3 catalyst
and the intensity of the band associated with NO adsorbed
onto Mo sites at 1690 cm1 starts to decrease. The catalytic
activity in the HDS of thiophene correlates with the absorption
of NO onto Co sites and not onto Mo sites suggesting the
formation of the CoMoS phase upon Co addition to the Mo/
Al2O3 catalyst (Figure 27).
In another in situ IR study, sulfided CoMo/Al2O3 catalysts
were investigated after sulfidation at different temperatures.
On a comparison of the changes of the surface properties by
increasing the sulfiding temperature from 673 to 873 K, differ-
ent types of CoMoS structures were detected. The type II
CoMoS structure being obtained at the higher sulfiding tem-
perature is characterized by none or very few bonds to Al2O3.
At the lower temperature, type I CoMoS is formed having Mo
OAl linkages, which is much less active in the HDS reaction
than the type II catalyst.205
308 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

Absorbance (a.u.)
6
Co
4

2 Moedge

2 4 6 8 10
Absorbance

CoMo (b) wt% Co added

Mo

kHDS  102 (mol h1 g1)

1.5

Co 1.0
1900 1800 1700 1600 1500
(a) Wavenumber (cm1) 0.5

1 2 3 4
(c)
Absorbance of NO on Co
Figure 27 (a) IR spectra of NO adsorbed onto sulfided 2% Co/Al2O3, 8% Mo/Al2O3, and 2%Co 8%Mo/Al2O3. (b) Absorbance of IR bands of NO
adsorbed onto Co (1850 cm1) and Mo (1690 cm1) plotted as function of Co/Mo ratio at constant Mo loading. (c) Thiophene HDS activities versus
absorbance of IR band of NO adsorbed onto Co.73

Atomic-scale insights into the nature of NO adsorption Diameter

onto MoS2 and CoMoS are obtained combining IR spectros-
copy investigations with scanning tunneling microscopy
(STM) experiments and DFT calculations.88 The NO molecule
does not adsorb at fully sulfided MoS2 edges when these edges
do not contain hydrogen. However, it can be expected that
sulfided catalysts contain hydrogen at the edge in the form of
SH groups. The results suggest a pushpull type mechanism
where simultaneously vacancies are created, and NO adsorp-
tion and H2S release occur. Dosing the catalyst with atomic
hydrogen, STM images reveal formation of vacancies and at
these vacancies NO dimers are adsorbed. In the IR spectra
measured during TPD, several NO adsorption complexes are
observed. On the basis of DFT calculations, it was shown that
mononitrosyl species dominate at the Mo edges and stable
dinitrosyl moieties are found at both the unpromoted and
the Co-promoted S-edges.
7.12.14 Structure of the Phases
On the S-terminated (0 0 0 1) plane of MoS2, only very weak
physisorption occurs, and hence the basal plane of the crystal-
lites is more or less inactive in catalytic reactions. Coordina-
tively unsaturated sites (CUS) and/or sulfur ion vacancies at
the corners and edges of the Mo(W)S2 slabs are the predomi-
nantly catalytically active sites (Lewis acid character).23 The
vacancies or coordinatively unsaturated sites are created in a
reaction with hydrogen. Due to the Lewis acid character of
Sites
n Layers
Rim
or
Edge stacking height
Basal
Figure 28 Scheme of the rim-edge model.
CUS, molecules with the unpaired electrons can be adsorbed.
Using probe molecules such as CO, NO, or O2, the number of
CUS could be determined and correlated with catalytic activity.88
Under real reaction conditions (presence of H2, H2S, and
excess substrates and reaction products), it is highly questionable
whether vacancies/CUS remain empty for any length of time.
For unpromoted MoS2 catalysts, a two-site model was
developed for MoS2 to explain the HDY and DDS pathways
(Figure 28). These sites are located on the edge of MoS2 nano-
crystals at different positions. The so-called rim-site terminat-
ing layers of the crystal stack on which HDY and DDS occur are
located at the edge of exterior layers with adjacent basal planes.
The second sites called edge-sites are located on interior layers
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
that have no exposed basal plane surfaces. On these sites only
the desulfurization reaction occurs.206
Promotion by adding Co decreases the strength of the MoS
bond and increases the electron density on the sulfur
atoms.207,208 More electron density at the S centers increases the
basicity of surface S species, which may facilitate CS bond
cleavage. Indeed, supported CoMoS catalysts exhibit a pro-
nounced selectivity toward DDS.209 Because of the electronic
density reorganization, the S atom bonded to Co and Mo exhibits
an intermediate bond strength optimal for HDS activity. The
bond weakening effect due to Co(Ni) promotion is supported
by theoretical calculations as well as experimental results.210,211
For promoted catalysts, the CoMoS (or NiMoS) phase
model proposed by Topse et al.141,179,212,213 is now widely
accepted: the catalysts consist of MoS2 nanocrystals with Co (or
Ni) promoter atoms substituting Mo atoms at the edges of the
MoS2 layers in the same plane as the Mo atoms, but their local
coordination is different. The CN of the promotor atoms is still
under debate and evidence was presented either for CN 4 or
for CN 5. Practical CoMo HDS catalysts are generally com-
posed of several phases: Co9S8 nanoparticles, Co2 species
2 1
3 corrected high-resolution electron microscopy and high-angle
1 3
2 can accept or donate electrons and thus act as a catalyst. Hence,
3 3
Figure 29 Scheme of a CoMoS catalyst composed of Co-
decorated single slab MoS2 (1), Co9S8 nanoparticles (2) and Co in the
support (3).212
Mo edge (100%)
(a) (b)
Figure 30 STM image of a MoS2 nanocrystal on Au(1 1 1) (left) showing the bright brim sites, the ball model of the structure with top view (middle,
top) and side view (middle, bottom), and a model cluster with the S2 groups.226
309
strongly interacting with the support, Mo (oxy)sulfides, as
well as CoMoS, schematically shown in Figure 29.
Two structural types of Co(Ni)MoS catalysts were
identified155: type-I with low activity and type-II with high
activity. Type-I structures are incompletely sulfided and con-
tain remaining MoOAl linkages to the support. The forma-
tion of such MoOAl bonds occurs during catalyst
preparation (see below). Type-II Co(Ni)MoS phases display a
weaker interaction to the support and they are fully sulfided.
Often, type-II catalysts consist of multiple MoS2 slabs that are
not linked to the support. The degree of stacking can be con-
trolled by appropriate selection of support properties and syn-
thesis conditions leading to stable single slabs of the catalytic
materials on g-Al2O3 supports.214,215 A type-III catalyst was
identified on SiO2 support and should consist of Co sulfide
clusters with S2 dimers between the Co atoms located at the
edges of the MoS2 slabs.135,216 A new model concerning hydro-
gen activation by MoS2 was recently presented. Unsupported
overstoichiometric MoS2x catalysts containing S22 species on
the edges react with H2 to form SH groups. These SH groups
participate in the catalytic reaction of thiophene conversion
directly to butane (HYD pathway). Removal of the extra sulfur
leads to a suppression or loss of the HYD function.217 Evi-
dences for the presence of H groups on MoS2 were recently
presented.218,219
Progress in the identification of active sites in MoS2 and
CoMoS/NiMoS model catalysts was achieved by the combina-
tion of theoretical calculations (DFT; see also Chapters 7.15 and
9.02) and advanced analytical tools such as STM, aberration-
annular dark-field scanning transmission electron microscopy
(HAADF-STEM).220225 The local structure of pure MoS2 (grown
on Au(11 1)) has a special electronic state identified as metallic
brim sites (Figure 30). In the brim edge states, Mo is fully
covered by S atoms forming S2 dimers on the edge. The edges
an S-containing molecule may be bonded at the brim site and
when hydrogen is available at the neighboring edge sites as a SH
group, hydrogen transfer and hydrogenation reaction can occur.
The results strongly indicate that the brim sites are catalytically
active for the HYD pathway.
The proposed formation of sulfur vacancies in MoS2 under
H2 atmosphere has also been observed directly by STM.
Mo78S204
cluster
-S 1
-S
2
-S 1
-S
2
-S 2
-S
1
-S 2
- S 1
-S 2
S1
310 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

However, one should keep in mind that STM was performed
under high vacuum conditions in the absence of substrate
molecules, H2/H2S, and at ambient temperature.
STM provided also evidences for significant changes of the
morphology when the Co/Ni promoter is added. Pure MoS2
particles adopt a triangular shape and when Co is added, the
particles become truncated hexagons (Figure 31, left). Two dif-
ferent edges can then be distinguished: one edge is covered by
100% and the second one by 50% S atoms. The situation is more
complex for NiMoS because two different shapes of the nano-
particles are observed. The first one is identical with that found
for CoMoS but the second one exhibits higher indexed edges
exhibiting a lower degree of S coordination (Figure 31, right).
The results suggest that Co/Ni plays two different roles:
promotion of MoS2 with Co/Ni alters the electronic structure
as seen as modified brim sites in STM; lower S coordination on
one type of NiMoS edges is an attractive structural situation for
adsorption of S-containing molecules. The occurrence of the
second edge type in NiMoS clusters may be a first step to
explain the different selectivities of the two catalytic systems
in the hydrotreating processes.
Concerning the activation of H2 on MoS2, it is proposed
that a dissociative chemisorption occurs forming SH groups
at the edges. The most stable position of hydrogen seems to be
on top of the S atom adjacent to the brim states.
The promoter atoms Co in an unsupported CoMoS catalyst
could be located using probe aberration-corrected STEM (Figure
32).225 Due to the lower atomic number, the Co atoms have a
lower intensity in the HAADF-STEM image compared to Mo.
The distances between the MoCoMoMo atoms along
the edge of the nanosized particle determined by the line
profile (Figure 32, right) are 2.34 A (0.1 A) for MoCo,
2.7 A (0.1 A) for CoMo), and 2.7 A (0.1 A) for MoMo
for the atoms labeled ABCD. This result may be explained
by the partial substitution of Mo by Co in the lattice of MoS2.
Similar results were obtained for Co-promoted WS2 and the
experimental finding is summarized in Figure 33.
According to DFT calculations, the 50% Co partial occupa-
tion of the Mo-edge covered by 25% S is stable and the simula-
tions also indicated the existence of two different configurations
with such S-saturations that are close in energy (Figure 33(a) and
33(b)). In these configurations, either MoCoMoCo alter-
nates or a pairing of the atoms in the order MoMoCoCo is
present, and in both cases the Mo-edge structure contains mixed
CoMo sites. The S atoms are located at the top of the Mo atoms.
However, DFT calculations also hinted that S atoms might be
located in bridging positions suggesting that the mobility of S
atoms on the Mo edge is high.228,229 In Figure 33(c), the model
obtained experimentally is schematically shown. While DFT
calculations yield an optimized local CoMo distance between
2.74 and 2.84 A, a pairing of Mo and Co edge atoms (2.34 A) is
observed. The model for Co-promoted WS2 (Figure 33(d))
exhibits a sequence WCoWCoW supporting the idea
that Co atoms are incorporated into the WS2 structure by
substituting W atoms at the edge in an alternating fashion.
7.12.14.1 The Role of Carbon in HDS Catalysts
It was also suggested54,55 that carbon plays an important role
in stabilizing the active phase. It is assumed that excess of sulfur
located on the surface of MoS2 is replaced by carbon, yielding
MoSxCy phases with a carbidic species (see also Chapter 7.14).

NiM
CoM oS
oS(1
010)
(10
10) 0) NiMoS(1120) NiM
10
) 01 oS(1
0) (10
(1 01
01 oS 2 oS 0)
(1 M iM
oS
2 N
M NiMoS(1010)
MoS2(1010) edge MoS2(1010)
(100% S) (100% S)

NiMoS(1010)
(50% S)

NiMoS(1010)
CoMoS2(1010) edge Ni
(50% S) Mo
S( )
(50% S) 10 10
10 (10
) NiMoS(1120) iMoS NiMoS(1120)
N
high index edge

Figure 31 Structure of promoted nanocrystals derived from STM investigations and DFT calculations.227

Mo
Co 140
A 130 Mo
Mo Mo
B 120
Mo 110
C
100 Mo
D 90
80
70 Co
A
60 C D
B
50
0.5 nm 40 0.234 0.27 0.27
30
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
nm

Figure 32 HAADF-STEM image of a Co-doped unsupported MoS2 nanowire catalyst (left) and the corresponding line profile (right).225
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts 311

Mo-edge Mo-edge
(1010) (1010)

S S
Mo Co Mo Co Mo Mo Co Co

S-edge
S-edge
(1010)
(1010)
MoCoMoCo MoMoCoCo

(a) (b)
Mo-edge Mo-edge
Mo-edge W-edge
(1010) (1010)

2.34 2.70 2.70 2.24 2.72 2.72 2.24

S
Mo1 Co Mo2 Mo3 W1 W2 Co2 W3
Co1

S-edge (1010) S-edge (1010)

(c) (d)

Figure 33 Two different views (perspective and top) of (a) Mo-edge, 50% Co in an alternate position and 25%S; (b) Mo-edge with 50% Co, pairing
configuration, 25% S. (c) and (d) perspective view of CoMoS (CoWS) structures for the experimentally observed MoCoMoMo (WCoWCoW)
sequence at the Mo/W-edge. Mo (W) atoms: blue (orange), S atoms yellow.225

800 structure (NEXAFS)232 and the most important findings are

discussed in detail. Catalysts were prepared by thermal decom-
MS833 before N2/DMS treatment position of (NH4)2MoS4 (ATTM) in inert atmosphere according
600 MS833 after N2/DMS treatment to the equation:
Intensity (counts s1)

NH4 2 MoS4 ! MoS3 2NH3 H2 S

400 Above 573 K, MoS2x phases are obtained by autoreduction of

MoS3 with the value x decreasing with increasing temperature.
The MoS2x samples were treated between 573 and 833 K with
a gas flow of CH3SCH3 (DMS) diluted in N2 or H2. A second
200
set of samples was synthesized by heating ATTM with a DBT
solution in decalin. The sample was treated at 623 K in a bomb
for 3 days.
0 The x-ray powder pattern of the catalyst treated with DMS at
0 10 20 30 40 50 60 70 80 90 100 793 K (sample composition: MoS2.05C1.12) shows no new
2q peaks (Figure 34) but a reduction of the absolute intensity
Figure 34 Comparison of the synchrotron XRD patterns of a sample compared to the sample before the treatment (note: identical
before and after DMS treatment (T 793 K).232 sample volumes were used) indicating a lower density of the
material after DMS treatment.
The (1 1 0) plane at 2y 47 exhibits a stronger intensity
The incorporation of carbon and the nature of the carbon relative to the (1 0 0) plane after the DMS treatment. The
species in MoS2-based catalysts were investigated with x-ray change of the intensity can be explained with a more pro-
powder diffractometry, IR spectroscopy, electron energy loss nounced distortion/more folding along the c axis after heating
spectroscopy (EELS), and near-edge x-ray absorption fine the precursor in DMS. The TEM image of a material obtained
312 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
20 nm
Figure 35 TEM micrographs of a MoS2 sample prepared by thermal decomposition of ATTM (left) and of a sample treated with DMS at 793 K
(right).232
s*
C-
K edge p* B
Intensity (a.u.)
A amorphous carbon.
292 294 296 298 300 302 304
a that both the Co(Ni)Mo(W)SC and Co(Ni)Mo(W)S phases
b of great importance and was more or less neglected.233
280 290 300 310 320 330 340 350
Energy loss (eV)
Figure 36 EELS C K-edge spectra of Mo2C (a) and a sample treated
at 793 K with N2/DMS. The inset shows the fine structure at the s*
transition.232
by thermal decomposition of ATTM at 833 K in N2 (Figure 35)
shows that MoS2 particles are present in closely packed groups
of large slabs exhibiting highly folded shapes. The average slab
length was estimated to about 94 A and the average stacking
number as 9.3, which is in reasonable agreement with the
value of 7.8 obtained from the x-ray powder pattern.
In the TEM image of a sample treated with DMS, heteroge-
neous regions with some highly stacked layers and regions
with less stacked and less curved layers are observed. The
average particle size was reduced by a factor of 2 to about
47 A and the stacking number decreased from 9.3 to 4.0.
In the EELS (Figure 36), an intense peak to the p* antibond-
ing state is observed and by a comparison with a spectrum of
graphite a graphitic structure can be excluded. In addition, the
peak shape structure near the ionization edge is also different
from that for amorphous, graphite, or carbide structures. Two
features A and B are observed for graphite and carbide for the s*
transition. Feature A is more intense for graphite than feature B,
while the opposite is observed for carbides and diamond. The
intensity difference is caused by the different density of states for
20 nm
graphitic sp2 carbon bonds and tetrahedral sp3 bonds in car-
bides. The feature B is more intense than A in the MoS2 sample
treated with DMS suggesting that carbon is not present as a
graphitic-like material and that MoC bonds of a carbide-like
structure are present. Because the p* peak is less intense than for
Mo2C, one may conclude that the sample may contain some
The replacement of S by C on edges of MoS2 can be crystal-
lographically accommodated. Hence, it cannot be ruled out
constitute the active catalyst. As mentioned recently, the role
of carbon in understanding the function of the catalysts may be
7.12.14.2 Morphology and Microstructure of Catalysts
In real catalysts, phase segregation was found and Co9S8 and
mainly Ni3S2, besides other Ni sulfides, are formed under reac-
tion conditions. In the remote control concept, Co9S8 and Ni3S2
may play a distinct catalytic role, providing spillover hydrogen
for the MoS2. Real catalysts display different MoS2 morphologies
such as straight and curved MoS2 single slabs, and stacks of
slabs.215 The MoS2 nanoparticles are either parallelly57,234,235
or perpendicularly234,235 oriented to the support.
Spent NiMoS/Al2O3 and CoMoS/Al2O3 catalysts prepared
by gas phase or liquid phase sulfidation (Table 1) of the oxidic
precursors exhibit a large variety of microstructures.215
The x-ray powder patterns of the two catalysts NiMo1-T2-L-
F and NiMo2-T2-L-F show no clear reflection at 2y 14 indi-
cating the absence of significant MoS2 stacking (Figure 37).
The shape and the broadness of the reflections are typical for
MoS2 with a low crystallinity. In addition, no indications are
found for the presence of crystalline Ni sulfides.
However, in about 30% of analyzed TEM regions of NiMo1-
T2-L-F Ni3S2 crystals are visible, which was verified with energy-
dispersive x-ray spectroscopy (EDX), electron diffraction, and
lattice imaging (Figure 38(a)). The TEM image in Figure 38(b)
shows that Ni3S2 crystals are decorated with MoS2 layers.
Ni sulfide particles with diameters between 5 and 50 nm can be
imaged using STEM-EDX. An example is shown in Figure 39
for NiMo3-T2-L-U. From an analysis of the STEM-EDX data,
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts 313

Table 1 Type I/II (T1/2), type II (T2), liquid-phase sulfidation (L), gas-phase sulfidation (G), FCC-PT (F), ULSD (U), FCC-PT canister (F/C), HC-PT
(H), diesel HDS (D)

Catalyst properties

Catalyst annotationa Description Metal loadings Sulfidation Originating fromb

NiMo1-T2-L-F Type 2 NiMo/Al2O3 3.2 at Ni and 5.8 at Mo per nm2 Liquid phase FCC-PT
NiMo2-T2-L-F Type 2 NiMo/Al2O3 1.8 at Ni and 5.6 at Mo per nm2 Liquid phase FCC-PT
NiMo3-T2-L-U Type 2 NiMo/Al2O3 2.5 at Ni and 8.1 at Mo per nm2 Liquid phase ULSD
NiMo3-T2-L-F/C Type 2 NiMo/Al2O3 2.5 at Ni and 8.1 at Mo per nm2 Liquid phase FCC-PT canister
NiMo3-T2-G-H Type 2 NiMo/Al2O3 2.5 at Ni and 8.1 at Mo per nm2 Gas phase HC-PT
NiMo3-T2-G Type 2 NiMo/Al2O3 2.5 at Ni and 8.1 at Mo per nm2 Gas phase
NiMo4-T1/2-L-F Type 1/2 NiMo/Al2O3 1.5 at Ni and 3.1 at Mo per nm2 Liquid phase FCC-PT
CoMo1-T2-L-F Type 2 CoMo/Al2O3 1.7 at Co and 4.8 at Mo per nm2 Liquid phase FCC-PT
CoMo2-T2-L-F Type 2 CoMo/Al2O3 2.4 at Co and 6.7 at Mo per nm2 Liquid phase ULSD
CoMo3-T1/2-L-D Type 1/2 CoMo/Al2O3 1.5 at Co and 3.1 at Mo per nm2 Liquid phase Diesel HDS
a
Type 1/2 (T1/2), Type 2 (T2), liquid phase sulfidation (L), gas phase sulfidation (G), FCC-PT (F), ULSD (U), FCC-PT canister (F/C), HC-PT (H), diesel HDS (D).
b
Fluid catalytic cracker pretreat (FCC-PT), ultralow sulfur diesel (ULSD), hydrocracker pretreat (HC-PT), hydrodesulfurization (HDS).215

* = MoS2
+ = gamma-Al2O3

* +
+
* +
* *
Intensity

+
* NiMo1-T2-L-F

NiMo2-T2-L-F

Reference

5 10 20 30 40 50 60 70 80
2q

Figure 37 x-Ray powder patterns of two liquid phase sulfided type II NiMo/Al2O3 catalysts NiMo1-T2-L-F (top) and NiMo2-T2-L-F (middle) together
with the pattern of a reference sample containing 1.5% Ni3S2 (bottom). Vertical bars: reflection positions of crystalline Ni3S2.215

it can be estimated that 3247% of Ni has segregated in
these catalysts corresponding to 1.33.0 wt% Ni3S2. Com-
bining these data with the MoS2 dispersion estimated from
the TEM micrographs and about 80% of Mo atoms located
on edge and corner (e c) sites, the amount of not segre-
gated Ni is sufficient to reach Ni/Mo (e c) atom ratios of
0.240.29. The Ni sulfide segregation phenomenon is also
detected in type I/II NiMo catalysts.
A detailed analysis also demonstrates that the sulfidation
procedure significantly influences the morphology of type II
catalysts. The MoS2 stacking is more pronounced when sulfi-
dation is done in H2S/H2 atmosphere, while liquid-phase sul-
fidation seems to prevent it. In a H2S/H2 atmosphere, only
few but much larger Ni3S2 crystals are formed than after
liquid-phase sulfidation. This may be related to the exother-
micity of the H2S/H2 sulfidation, an effect that is suppressed if
liquid is present during the sulfidation.
In contrast to Ni-promoted catalysts, no Co-rich particles
are visible in CoMo type II catalysts displaying the same
composition and preparation, when they are used under the
same conditions. It is highly likely that Co9S8 aggregates less
rapidly than Ni3S2. For liquid-phase sulfided type II commer-
cial NiMo catalysts, MoS2 stacking is not pronounced and is
mainly absent in some catalysts (Figure 40). The MoS2 disper-
sion is high after liquid-phase sulfidation and using short
reaction times, while longer reaction times and/or more severe
reaction conditions produce longer MoS2 slabs (i.e., the dis-
persion is lower) (NiMo3-T2-L-F/C, used for 1 year in a unit
314 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

5 nm 10 nm

(a) (b)

Figure 38 TEM micrographs of liquid-phase sulfided type II NiMo/Al2O3 catalysts: (a) shows a lattice image of a Ni sulfide crystallite in NiMo1-T2-L-F;
(b) decoration of a Ni sulfide particle with MoS2 slabs in NiMo2-T2-L-F.215

0.2 mm 0.1 mm
(a) (b)

Figure 39 STEM-EDX images of Ni (red) and Mo (blue) of liquid phase sulfided type II NiMo/Al2O3 catalyst NiMo3-T2-L-U.215

10 nm 5 nm

(a) (b)

Figure 40 TEM micrographs of liquid-phase sulfided type II NiMo/Al2O3 catalyst NiMo3-T2-L-F/C showing long curved MoS2 slabs.215
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
20 nm
(a)
Figure 41 HAADF-STEM images of MoS2/C (a) and WS2/C (b) catalysts. All clusters in the two images are single-layer structures.220
20 nm
Figure 42 Promotion of MoS2 with Co changes the morphology toward
heavily truncated triangles.220
operating at high temperature). Nevertheless, the degree of
stacking is rather limited and most MoS2 slabs consist of only
one or two layers. It seems that MoS2 stacking is not required
for desulfurization of molecules such as DBT. An interesting
feature occurs for this catalyst, namely the curvature of the
MoS2 layers (Figure 40). Type I and II catalysts often show
MoS2 morphologies such as Al2O3 needles or plates coated
with MoS2.
Carbon-supported MoS2, WS2, CoMoS, NiMoS, and NiWS
catalysts were prepared by impregnation with an aqueous
solution containing (NH4)2[MS4] (M Mo or W) and, for
promoted catalysts, also with Me(acetate)2 (Me Co, Ni).
After impregnation, the samples were dried in air at 110  C
and then treated with 10% H2S in H2 for 6 h at 1073 K. The
morphology of the catalysts was investigated with HAADF-
STEM and an example is shown in Figure 41.220 Most of the
single-layered clusters exhibit a regular shape in the form of a
truncated triangle. The angles at the vertices of the nanostruc-
tures amount to about 120 and are consistent with both the
315
20 nm
(b)
(1 0  1 0) (so-called Mo(W) edges) and (1 0 1 0) (so-called S
edges) being exposed.
The addition of promoter atoms to MoS2 modifies the
morphology to more heavily truncated triangles (Figure 42),
that is, a change toward a more hexagonal shape is induced.
This observation is in line with STM images of Co-promoted
MoS2 deposited on Au(1 1 1).236 The Co atoms in the CoMoS
structures are located at the (1 0 1 0) S edges and Co addition
causes exposure of more promoted (1 0 1 0) S-type edges
while the amount of Mo-type edges (1 0 1 0) is reduced. The
situation is similar for NiMoS and NiWS catalysts.
Promoter atoms may have other effects than changing the
degree of truncation. There are particles that are more round
and in some cases additional new higher index edges are
present. For some of the edges, angles of about 150 to the
dominant edges are observed indicating that the promoted
structures also expose (1 1 2 0) edges. This finding is in agree-
ment with atomically resolved STM images of NiMoS struc-
tures for small clusters.221 The present results give evidence that
extended (1 1 2 0) edges may indeed be stable features of the
promoted structures.
7.12.15 Choice of Catalysts and Other Catalysts
for HDS
Whether CoMoS or NiMoS is used depends on several factors
because these catalysts behave differently. Generally, CoMoS/
g-Al2O3 catalysts are very efficient in the HDS process but they are
less efficient for HDN or hydrogenation of aromatic molecules.
NiMoS/g-Al2O3 phases exhibit better HDN activity and are better
hydrogenation catalysts. Hence, they are preferably applied for
feedstock with a high concentration of unsaturated compounds
because more aromatic molecules are saturated leading to a hig-
her cetane number and density. However, this is accompanied by
higher hydrogen consumption, requiring appropriate hydrogen
availability, hydrogen recycle rate, and compression costs. For
aromatic hydrogenation reactions, NiWS-based catalysts are
highly effective but the higher costs compared to NiMoS catalysts
have limited their uses. At low pressures and high temperatures, a
CoMoS catalyst can be a better choice than a NiMoS catalyst.
Other approaches to obtain active catalysts were also
reported and some are mentioned in brief. Using bimetallic
complexes containing, for example, Mo3S4M cores (M e.g.,
316 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts

Ru, Ir, Rh) deposited on g-Al2O3 yields active catalysts. The identification of the Mo edge brim site (see above) as the HYD site
results suggest that the activity is at least partially caused by a explains the low inhibiting effect of H2S on hydrogenation.
better distribution of MoS2 slabs on the support. No clear H2S partial pressure also plays an important role in
evidences were found whether the noble metals are located in poisoning.244 However, at low partial pressure H2S may exert
direct contact with the MoS2 phase.237 a promotional effect.249 Model compound studies have shown
Are there compounds exhibiting better HDS performance? that some basic nitrogen-containing molecules (e.g., acridine)
Yes, as can be seen in Figure 43; the HDS activity scales with are orders of magnitude more inhibiting than other N-
the MS bond strength (M transition metals) and the best containing molecules (e.g., aniline or indole). The differences
catalysts are found for noble metal sulfides. Keeping in mind in the degree of inhibition between different catalysts, different
that more than 120 000 tons of catalysts are used in refineries types of S-compounds, and between DDS and HYD pathways
and that roughly the same amount must be replaced every year have been explained on the basis of different catalytic sites
due to catalyst deactivation, the high prices and restricted involved in DDS and HYD reactions, as well as based on the
availability of the noble metals prevent the use of such sulfides. mode of adsorption of the reactants.
During the last few years, unsupported metal phosphides The HDS reaction is also significantly inhibited by organic
have been studied due to their high activity for HDS and nitrogen compounds. Experimental and theoretical data
HDN.239. Both MoP and WP exhibit only moderate activity suggest that competitive adsorption between N-containing
but Ni2P shows an excellent activity in hydroprocessing. and S-containing compounds occurs on active sites. The HDS
According to results of several studies performed for simulta- reaction is then inhibited due to the strong adsorptive strength
neous HDS and HDN reactions, the overall activity was found of the N-containing molecules. The extent of inhibition,
to be in the order Fe2P < CoP < MoP < WP < Ni2P. however, depends on the type and the concentration of organic
nitrogen compounds.
Catalysts applied in the hydrotreating of light feeds (so-
7.12.16 Poisoning and Deactivation Effects called middle distillates) mainly deactivate due to coke depo-
sition and sintering. Metal deposition as a deactivation source
The poisoning effects of H2S, nitrogen compounds, and is found in hydroprocessing of heavy feeds.250 Coverage of the
aromatics are different for the HYD and DDS routes.240242 active sites by N-compounds and coke formation are the most
According to experimental and theoretical DFT-based investiga- important factors in catalyst deactivation. Basic N-containing
tions, the general trend for poisoning is: nitrogen compounds may adsorb irreversibly onto catalyst acidic sites,
compounds> organic sulfur compounds > polyaromatics oxy- and may be converted to coke. HDS reaction conditions such
gen compounds  H2S > monoaromatics. The H2S produced as T, partial H2 pressure, and H2/oil ratio strongly influence
during HDS may react with the CUS to form saturated sites coke formation. Coke is formed at the beginning of the reac-
containing SH groups.243 Hence, the number of CUS is reduced tion, but the catalyst may retain a substantial portion of its
inhibiting removal of S via DDS, while the HYD pathway is only original activity.251
slightly affected.244248 Different catalysts exhibit different sensi-
tivities against H2S poisoning. NiMoS/g-Al2O3 is more sensitive
than Co-promoted catalysts.39 The inhibition effect is more pro- 7.12.17 Synthesis Strategies of Catalysts
nounced for HDS of DBT than for alkyl-substituted DBTs.244 The
In the literature, numerous different preparation approaches
are reported. Hence, only some of these methods are presented
in brief.
1.00E + 03 A variety of unsupported NiMoWS catalysts are commer-
RuS2
cially used with the commercial trade name NEBULA. The
IrS2
synthesis procedure is patented and therefore the exact prepa-
OsS2
Log (HDS activity per mole of metal)

Rh2S3 ration method is not known.

1.00E + 02 There are studies where the chemical vapor deposition
ReS2 (CVD) route was applied to deposit Co on supported MoS2.
The MoS2 was treated with Co(CO)3NO vapor at RT for a
1.00E + 01
PdS PtS distinct time followed by sulfidization. HDS activity was sub-
MoS2
stantially higher than that of a conventionally prepared catalyst
Cr2S3
WS2 (impregnation route). More interestingly, adding Co by the
Ni CVD technique to a conventionally prepared CoMoS/Al2O3
FeS Co9S8 NbS2
1.00E + 00 VS
TaS2 catalyst, the HDS activity increased significantly.252
MnS Besides the use of oxidic starting compounds, the suitability
of S-containing precursors was investigated. Such precursors
are either ammonium thiomolybdate (ATM)/thiotungstate
1.00E 01 (ATT), tetraethylammonium salts of the type [M0 (MS4)2]n
30 40 50 60 70 80
(n 2 or 3) (M0 Co or Ni, M Mo or W)253, or tetraalkylam-
MS strength bond (kcal mol1) monium thiometalates with formula (R4N)2MS4 (R methyl,
Figure 43 HDS activity of transition metal sulfides plotted against MS ethyl, propyl, pentyl, hexyl, etc.; M Mo, W). These precursors
bond strength.238 are thermally decomposed at relatively low temperatures often
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts 317

yielding finely dispersed unsupported catalysts.230,254256 atoms. In addition, a weak catalyst-support interaction favors
Using alkyl-containing precursors, the decomposed MoS2x formation of type II Co(Ni)Mo(W)S structures with a low degree
contains carbon and small amounts of nitrogen. The carbon of stacking. Several disadvantages of promising supports should
content increases with the size of R. The catalyst is then impreg- be shortly mentioned. An upscale of mesoporous materials such
nated with promotor salts and the Co(Ni)Mo(W)S catalyst is as SBA-15 and MCM-41 yielding kilograms or even hundred of
sulfided either ex situ or in situ.257259 Unpromoted MoS2 kilograms was not possible until now. Zeolites have very small
catalysts with large carbon contents and a good performance pore diameters (less than 2 nm) and dispersion of the catalysts
are obtained by thermal decomposition and different activa- is relatively poor. Moreover, the pores are so small that large
tion modes.260 To avoid the use of any oxidic materials in the DBT molecules cannot enter the pores.
catalyst preparation, a combination of ATTM/(Prop4N)2MoS4
and nickel diethyldithiocarbamate or cobalt dimethylthiocar-
bamate was used.261,262 These catalysts are characterized by a
7.12.19 Conclusion
good dispersion, variable carbon content, and high activity.
Another approach is the use of single source precursors such
Co(Ni)MoS/Al2O3 and related catalysts have been commercially
as (ML)(MoO4) (L amine such as EN, M transition metal),
used for more than 80 years. During this time period, different
for example, Co(en)3MoO4. The precursor is first thermally
models and theories were developed and the evolution of the
decomposed followed by sulfidization at elevated temperature
models/theories is summarized in several reviews.23,2527
in a H2S/H2 atmosphere.10 There are also reports where the
Most models developed are directly related to the layered
oxidic catalyst precursors are sulfided in solution at elevated
crystal structure of MoS2. The first structural model of CoMo/
temperature and high H2 pressure using, for example, dimethyl-
Al2O3 was the monolayer model developed by Schuit and
disulfide (DMDS) dissolved in an inert organic solvent.
Lipsch (Figure 44).275 After calcination, Mo species are on the
surface of the support forming a monolayer. The interaction
between Mo and Al2O3 occurs via MoOAl bridges. Incorpo-
7.12.18 Supported Catalysts and the Support Effect
ration of Mo6 ions was proposed to be charge compensated by
a capping O2 layer located on top of the monolayer, and Co2
For most supported Co(Ni)MoS catalysts g-Al2O3 is used: this
was assumed to substitute Al3 ions in the surface and being in
material has a high specific surface area (200350 m2 g1),
tetrahedral environment. The promotional effect of Co2 was
pore size can be controlled, it has a good affinity to sulfide
suggested to increase the stability of the monolayer.
yielding a high dispersion, it contains acidic and basic sites, can
During sulfidation, the capping oxygen atoms are replaced
be easily formed into desired shapes, mechanical strength is
by S2 anions. The presence of H2 under HDS conditions
suitable, and cost is low. To improve the performance of the
causes a removal of S2 leading to a reduction of Mo6 to
catalysts several support materials such as carbon, SiO2, TiO2,
Mo3, which were assumed to be the catalytically active sites.
ZrO2, MgO, zeolites, and mesoporous silicas are used.263266
Another model was the intercalation model (Figure 45)276,277
Further strategies are modifying g-Al2O3 by mixing with different
assuming that the sulfided catalysts contain MoS2/WS2 on the
materials such as zeolites or other metal oxides and adding
g-Al2O3 and the Co/Ni atoms occupy octahedral sites in the
additives such as P, F, B, Si, Mg, Zn, La, V, and Ga as modifiers,
van der Waals gaps between the MoS2/WS2 layers. Later it was
to alter the acidic and basic properties of g-Al2O3. The con-
shown that intercalation of Co/Ni is energetically not possible
sequences of the application of such mixtures are altered
and it was proposed that the promoter cations are located at
interaction between the impregnation solution and the support,
the edges of the slabs (pseudo-intercalation).278 The model
a change of the dispersion of the active phase, and reducibility of
implicitly assumes a three-dimensional multilayer structure of
the oxidic precursor catalyst. In recent years, phosphorus is
MoS2/WS2 so that intercalation can take place.
mainly used as a modifier and in several reports an improvement
of the performance was reported influencing especially the S S S
S
HDN activity of NiMoS/Al2O3. The addition of boron to g-
Mo Mo
Al2O3 seems to have a positive effect for CoMoS because the
formation of catalytically inactive bulk Co9S8 is minimized.267 O O O O
Some oxides have acidbase properties promoting desul- Co Co
furization of alkyl DBTs. Acidic supports such as zeolites facilitate Co
isomerization and trans-alkylation of alkylated DBT resulting Figure 44 Schematic illustration of the monolayer model.
in an enhanced HDS activity of these refractory sulfur species.
However, the disadvantage of such supports is an extensive
coking shortening the lifetime of the catalyst. To increase the Bulk intercalation
dispersion of the catalytically active phase, mesoporous silica Surface
materials such as SBA-15 or MCM-41 with very large surface MoS
S intercalation
Co Co
areas of about 1000 m2 g1 were tested (see also Chapters 5.06 Mo S
S
and 7.11).268274 Under laboratory conditions, all these mate-
rials showed comparable or even better HDS performance than
industrial catalysts.
Most of the studies focus on a high dispersion of the catalysts Figure 45 Schematic representation of the intercalation pseudo-
enabling a good decoration of Mo(W)S2 edges with promoter intercalation model.
318 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
MoS2 Co9S8
Figure 46 The contact synergy model.
The contact synergy model also assumes the presence of
layered MoS2 (Figure 46).279,280 In this model, Co9S8 and
MoS2 are in tight contact and results in spill-over of hydrogen
from Co9S8 to MoS2 enhancing the intrinsic activity.281
During the last decades, a detailed picture of the CoMoS
phase on the atomic scale was developed based on the findings
and observations made with a variety of analytical tools and
theoretical calculations. The catalyst contains nanosized MoS2
particles decorated with Co on the edges to form the catalyti-
cally highly active CoMoS phase. After impregnation of the
alumina support with molybdenum and cobalt compounds,
followed by drying and/or calcination, well-dispersed oxides of
these elements are obtained. The oxides are then converted
into the sulfides achieving the catalytically active phase. A
more detailed picture of the Co/NiMoS catalysts will be
obtained applying more in situ studies to probe the structure
activity relationship at the individual sites on nanosized
catalysts in their working state. Such studies are challenging
especially if industrial-type conditions are envisaged. The
results of the experiments will help to rationally design cata-
lysts for the different applications in the HDS reaction. HDS
research has moved to the most refractory sulfur compounds,
that is, DBTs with alkyl substituents in the 4- and 6-positions.
In addition, improvements of conventional metal sulfide
catalysts and the development of new materials are required
to meet the increasing demand for energy.
References
1. European regulations: http://www.dieselnet.com/standards/eu/fuel.php.
2. Japan regulations: http://www.dieselnet.com/standards/jp/fuel.php.
3. Ultra Low-sulfur Diesel Regulations Take Effect in US. Green Car Congress: Ultra
Low-sulfur Diesel Regulations, 2006, http://www.greencarcongress.com/2006/
06/ultra_lowsulfur.html.
4. Topse, H.; Hinnemann, B.; Nrskov, J. K.; Lauritsen, J. V.; Besenbacher, F.;
Hansen, P. L.; Hytoft, G.; Egeberg, R. G.; Knudsen, K. G. Catal. Today 2005,
107108, 1222.
5. Fujikawa, T. Top. Catal. 2009, 52, 872879.
6. Breysse, M.; Geantet, C.; Afanasiev, P.; Blanchard, J.; Vrinat, M. Catal. Today
2008, 130, 313.
7. Knudsen, K. G.; Cooper, B. H.; Topsoe, H. Appl. Catal. A: Gen. 1999, 189,
205215.
8. Marafi, A.; Al-Hindi, A.; Stanislaus, A. Fuel Process. Technol. 2007, 88,
905911.
9. Ho, T. C. Appl. Catal. A: Gen. 2003, 244, 115128.
10. Ho, T. C. Catal. Today 2004, 98, 318.
11. Farag, H. Appl. Catal. A: Gen. 2007, 331, 5159.
12. Choi, K. H.; Korai, Y.; Mochida, I.; Ryu, I. W.; Min, W. Appl. Catal. B: Environ.
2004, 50, 916.
13. Liu, Z.; Zhang, Q.; Zheng, Y.; Chen, J. Energy Fuel 2008, 22, 860866.
14. Song, T.; Zhang, Z.; Chen, J.; Ring, Z.; Yang, H.; Zheng, Y. Energy Fuel 2006, 20,
23442349.
15. Babich, I. V.; Moulijn, J. A. Fuel 2003, 82, 607631.
16. Song, C. Catal. Today 2003, 86, 211263.
17. Song, C.; Ma, X. Appl. Catal. B: Environ. 2003, 41, 207238.
18. Parlevliet, F.; Eijsbouts, S. Catal. Today 2008, 130, 254264.
19. Stanislaus, A.; Marafi, A.; Rana, M. S. Catal. Today 2010, 153, 168.
20. Pawelec, B.; Navarro, R. M.; Campos-Martin, J. M.; Fierro, J. L. G. Catal. Sci.
Technol. 2011, 1, 2342.
21. Grange, P.; Vanhaeren, X. Catal. Today 1997, 36, 375391.
22. Brunet, S.; Meya, D.; Perot, G.; Bouchy, C.; Diehl, F. Appl. Catal. A: Gen. 2005,
278, 143172.
23. Topsoe, H.; Clausen, B. S.; Massoth, F. E. In Anderson, J. R., Boudart, M., Eds.;
Catalysis Science and Technology; Springer Verlag: New York, 1996; Vol. 11.
24. Startsev, A. N. Catal. Rev.: Sci. Eng. 1995, 37, 353423.
25. Chianelli, R. R.; Daage, M.; Ledoux, M. J. Adv. Catal. 1994, 40, 177232.
26. Massoth, F. E. Adv. Catal. 1978, 27, 265310.
27. Weisser, O.; Landa, S. Sulfide Catalysts, Their Properties and Applications.
Pergamon: New York, 1973.
28. Prins, R.; DeBeer, V. H. J.; Somorjai, G. A. Catal. Rev. Sci. Eng. 1989, 31, 141.
29. Girgis, M. J.; Gates, B. C. Ind. Eng. Chem. Res. 1991, 30, 20212058.
30. Topse, H.; Clausen, B. S.; Massoth, F. E. Hydrotreating Catalysis. Springer-
Verlag: New York, 1996.
31. Eijsbouts, S. Appl. Catal. A: Gen. 1997, 158, 5392.
32. Whitehurst, D. D.; Isoda, T.; Mochida, I. Adv. Catal. 1998, 42, 345371.
33. Kabe, T.; Ishihara, A.; Qian, W. Hydrodesulfurization and Hydrodenitrogenation.
Wiley-VCH: New York, 1999.
34. Sun, M.; Nicosia, D.; Prins, R. Catal. Today 2003, 86, 173189.
35. Choi, K.; Kunisada, N.; Korai, Y.; Mochida, I.; Nakano, K. Catal. Today 2003, 86,
277286.
36. Breysse, M.; Djega-Mariadassou, G.; Pessayre, S.; Geantet, C.; Vrinat, M.;
Perot, G.; Lemaaire, M. Catal. Today 2003, 84, 129138.
37. Sapre, A. Advanced Distillate Hydroprocessing Technology. In 4th Conference on
Oil Refining and Petrochemicals in the Middle East, Abu Dhabi, January 28, 2003.
38. Gates, B. C.; Topse, H. Polyhedron 1997, 16, 32133217.
39. Egorova, M.; Prins, R. J. Catal. 2004, 225, 417427.
40. Landau, M. V. Catal. Today 1997, 36, 393429.
41. Ma, X. L.; Sakanishi, K.; Mochida, I. Ind. Eng. Chem. Res. 1996, 35, 24872494.
42. Mijoin, J.; Perot, G.; Bataille, F.; Lemberton, L.; Breysse, M.; Kasztelan, S. Catal.
Lett. 2001, 71, 139145.
43. Gao, Q.; Ofosu, T. N. K.; Ma, S.-G.; Komvokis, V. G.; Williams, C. T.; Segawa, K.
Catal. Today 2011, 164, 538543.
44. Apecetche, M. A.; Delmon, B. React. Kinet. Catal. Lett. 1979, 12, 385392.
45. Delmon, B. Appl. Catal. 1985, 15, 132.
46. Carrier, X.; Lambert, J. F.; Che, M. J. Am. Chem. Soc. 1997, 119, 1013710146.
47. Le Bihan, L.; Blanchard, P.; Fournier, M.; Grimblot, J.; Payen, E. J. Chem. Soc.,
Faraday Trans. 1998, 94, 937940.
48. Brito, J. L.; Barbosa, A. L.; Albornoz, A.; Severino, F.; Laine, J. Catal. Lett. 1994,
26, 329337.
49. Damyanova, S.; Spojakina, A. React. Kinet. Catal. Lett. 1993, 51, 465472.
50. Delannay, E. Appl. Catal. 1985, 16, 135152.
51. Hayden, T. F.; Dumesic, J. A. J. Catal. 1987, 103, 366384.
52. Hayden, T. F.; Dumesic, J. A.; Sherwood, R. D.; Baker, T. K. J. Catal. 1987, 105,
299318.
53. Derouane, E. G.; Pedersen, E.; Clausen, B. S.; Gabelica, Z.; Candia, R.; Topse, H.
J. Catal. 1986, 99, 253261.
54. Kisfaludi, G.; Leyrer, J.; Knozinger, H.; Prins, R. J. Catal. 1991, 130, 192201.
55. Louwers, S. P. A.; Craje, M. W.; van der Kraan, A. M.; Geantet, C.; Prins, R.
J. Catal. 1993, 144, 579596.
56. Miciukiewicz, J.; Qader, Q.; Massoth, F. E. Appl. Catal. 1989, 49, 247257.
57. Pratt, K. C.; Sanders, J. V.; Christov, V. J. Catal. 1990, 124, 416432.
58. Saini, A. R.; Johnson, B. G.; Massoth, F. E. Appl. Catal. 1988, 40, 157172.
59. Woodruff, D. P., Delchar, T. A., Eds.; In Modern Techniques of Surface Science;
Cambridge University Press: Cambridge, 1986.
60. Zangwill, A., Ed.; In Physics at Surfaces; Cambridge University Press: Cambridge,
1988.
61. Bachelier, J.; Duchet, J.-C.; Cornet, D. Bull. Soc. Chim. Belg. 1981, 90,
13011309.
62. Peri, J. B. J. Phys. Chem. 1982, 86, 16151622.
63. Bachelier, J.; Tilliette, M. J.; Cornac, M.; Duchet, J.-C.; Lavalley, J.-C.; Cornet, D.
Bull. Soc. Chim. Belg. 1984, 93, 743750.
64. Muller, B.; van Langeveld, A. D.; Moulijn, J. A.; Knozinger, H. J. Phys. Chem.
1993, 97, 90289033.
65. Mauge, F.; Lavalley, J.-C. J. Catal. 1992, 137, 6976.
66. Elst, L. P. A. F.; Eijsbouts, S.; van Langeveld, A. D.; Moulijn, J. A. J. Catal. 2000,
196, 95103.
67. Travert, A.; Dujardin, C.; Mauge, F.; Cristol, S.; Paul, J.-F.; Payen, E.;
Bougeard, D. Catal. Today 2001, 70, 255269.
68. Travert, A.; Dujardin, C.; Mauge, F.; Veilly, E.; Cristol, S.; Paul, J.-F.; Payen, E.
J. Phys. Chem. B 2006, 110, 12611270.
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
69. Dujardin, C.; Lelias, M. A.; van Gestel, J.; Travert, A.; Duchet, J. C.; Mauge, F.
Appl. Catal. A: Gen. 2007, 322, 4657.
70. Okamoto, Y.; Katoh, Y.; Mori, Y.; Imanaka, T.; Teranishi, S. J. Catal. 1981, 70,
445448.
71. Topse, N.-Y.; Topse, H. J. Catal. 1982, 75, 354374.
72. Topse, N.-Y.; Topse, H. J. Catal. 1982, 77, 293296.
73. Topse, N.-Y.; Topse, H. J. Catal. 1983, 84, 386401.
74. Valyon, J.; Hall, W. K. J. Catal. 1983, 84, 216228.
75. Valyon, J.; Schneider, R. L.; Hall, W. K. J. Catal. 1984, 85, 277283.
76. Shuxian, Z.; Hall, W. K.; Ertl, G.; Knozinger, H. J. Catal. 1986, 100, 167175.
77. Koizumi, N.; Iijima, M.; Kasahara, S.; Yamada, M. Chem. Lett. 1996, 9, 815816.
78. Koizumi, N.; Iijima, M.; Mochizuki, T.; Yamada, M. Stud. Surf. Sci. Catal. 1997,
106, 293302.
79. Koizumi, N.; Takahasi, K.; Yahazaki, M.; Yamada, M. Catal. Today 1998, 45,
313318.
80. Yamada, M.; Koizumi, N.; Yamazaki, M. Catal. Today 1999, 50, 38.
81. Yamazaki, M.; Magara, H.; Koizumi, N.; Yamada, M. Stud. Surf. Sci. Catal. 1999,
126, 155162.
82. Nielsen, L. P.; Ibsen, L.; Christensen, S. V.; Clausen, B. S. J. Mol. Catal. A: Chem.
2000, 162, 375379.
83. Mauge, F.; Lamotte, J.; Nesterenko, N. S.; Manoilova, O.; Tsyganenko, A. A. Catal.
Today 2001, 70, 271284.
84. Okamoto, Y.; Kawano, M.; Kubota, T. Chem. Commun. 2003, 10861087.
85. Okamoto, Y.; Kubota, T. Catal. Today 2003, 86, 3143.
86. Okamoto, Y.; Kawano, M.; Kawabata, T.; Kubota, T.; Hiromitsu, I. J. Phys. Chem. B
2005, 109, 288296.
87. Koizumi, N.; Jung, S.; Hamabe, Y.; Suzuki, H.; Yamada, M. Catal. Lett. 2010, 135,
175181.
88. Topse, N.-Y.; Tuxen, A.; Hinnemann, B.; Lauritsen, J. V.; Knudsen, K. G.;
Besenbacher, F.; Topse, H. J. Catal. 2011, 279, 337351.
89. Jayamurthy, M.; Vasudevan, S. J. Phys. Chem. 1994, 98, 67776784.
90. Moon, S. J.; Ihm, S. K. Appl. Catal. 1988, 42, 307322.
91. Mangnus, P. J.; de Beer, V. H. J.; Moulijn, J. A.; van Langeveld, A. D. Appl. Catal.
1991, 68, 161177.
92. Mangnus, P. J.; Moulijn, J. A.; de Beer, V. H. J. Appl. Catal. 1990, 67, 119139.
93. McGarvey, G. B.; Kasztelan, S. Stud. Surf. Sci. Catal. 1992, 73, 310.
94. Scheffer, B.; Mangnus, P. J.; Moulijn, J. A. J. Catal. 1990, 121, 1830.
95. Sun, M.; Burgi, T.; Cattaneo, R.; Prins, R. J. Catal. 2001, 197, 172181.
96. Hurst, N. W.; Gentry, S. J.; Jones, A.; McNicol, B. D. Catal. Rev: Sci. Eng. 1982,
24, 233309.
97. Arnoldy, P.; van den Heijkant, J. A. M.; de Bok, G. D.; Moulijn, J. A. J. Catal.
1985, 92, 3555.
98. Scheffer, B.; Arnoldy, P.; Moulijn, J. A. J. Catal. 1988, 112, 516527.
99. Scheffer, B.; Dekker, N. J. J.; Mangus, P. J.; Moulijn, J. A. J. Catal. 1990, 121,
3146.
100. Arnoldy, P.; Moulijn, J. A. J. Catal. 1985, 93, 3854.
101. Mangnus, P. J.; Riezebos, A.; van Langeveld, A. D.; Moulijn, J. A. J. Catal. 1995,
151, 178191.
102. Scheffer, B.; Molhoek, P.; Moulijn, J. A. Appl. Catal. 1989, 46, 1130.
103. Al-Dalama, K.; Stanislaus, A. Thermochim. Acta 2011, 520, 6774.
104. de Boer, M.; van Dillen, A. J.; Koningsberger, D. C.; Geus, J. W. J. Phys. Chem.
1994, 98, 78627870.
105. Zeuthen, P.; Muegge, B.; Blom, P.; Massoth, F. E. Appl. Catal. 1991, 68, 117130.
106. Yoshimura, Y.; Yokokawa, H.; Sato, T.; Shimada, H.; Matsubayashi, N.;
Nishijima, A. Appl. Catal. 1991, 73, 3953.
107. Massoth, F. E.; Zeuthen, P. J. Catal. 1994, 145, 216222.
108. Burch, R.; Collins, A. Appl. Catal. 1985, 18, 373387.
109. Li, X. S.; Grange, P.; Xin, Q.; Guo, X. X.; Delmon, B. J. Catal. 1992, 137,
385393.
110. Niemantsverdriet, J. W. Spectroscopy in Catalysis, An Introduction. Wiley-VCH:
Weinheim, 2000.
111. Ertl, G., Kuppers, J., Eds.; In Low Energy Electrons and Surface Chemistry; VCH:
Weinheim, 1985.
112. Feldman, L. C., Mayer, J. W., Eds.; In Fundamentals of Surface and Thin Film
Analysis; North-Holland: Amsterdam, 1986.
113. Watts, J. F., Wolstenholme, J., Eds.; In An Introduction to Surface Analysis by
XPS and AES; Wiley VCH: Weinheim, 2003;; Koningsberger, D. C., Prins, R.,
Eds.; In X-ray Absorption: Principles, Applications, Techniques of EXAFS,
SEXAFS, and XANES; Wiley: New York, 1998; p 321.
114. Briggs, D.; Seah, M.P. Practical Surface Analysis, sec. Ed., Vol. 1, Auger and
X-ray Photoelectron, Spectroscopy; 1990.
115. Reinert, F.; Hufner, S. New J. Phys. 2005, 7, 97.
116. Hughes, A. E.; Pratt, K. C.; Tsai, E. Appl. Catal. A 1992, 90, 117129.
319
117. Yoshimura, Y.; Yoshitomi, S.; Kameoka, T.; Sato, T.; Shimada, H. Energ. Fuel.
1994, 8, 435445.
118. Arteaga, A.; Fierro, J. L. G.; Delannay, F.; Delmon, B. Appl. Catal. 1986, 26,
227249.
119. Ramirez, J.; Fuentes, S.; Diaz, G.; Vrinat, M.; Breysse, M.; Lacroix, M. Appl. Catal.
1989, 52, 211224.
120. McGarvey, G. B.; Kasztelan, S. J. Catal. 1994, 147, 476483.
121. Okamoto, Y.; Maezawa, A.; Imanaka, T. J. Catal. 1989, 120, 2945.
122. Gandubert, A. D.; Legens, C.; Guillaume, D.; Rebours, S.; Payen, E. Oil Gas Sci.
Technol.: Rev. IFP 2007, 62, 7989.
123. Jimenez-Gonzalez, A.; Schmeisser, D. J. Catal. 1991, 130, 332346.
124. Spevack, P. A.; Mclntyre, S. Appl. Catal. 1990, 64, 191207.
125. Sajkowski, D. J.; Miller, J. T.; Zajac, G. W.; Morrison, T. I.; Chen, H.; Fazzini, D. R.
Appl. Catal. 1990, 62, 205220.
126. Lamic, A. F.; Daudin, A.; Brunet, S.; Legens, C.; Bouchy, C.; Devers, E. Appl.
Catal. A 2008, 344, 198204.
127. Gandubert, A. D.; Krebs, E.; Legens, C.; Costa, D.; Guillaume, D.; Raybaud, P.
Catal. Today 2008, 130, 149159.
128. Qiu, L.; Xu, G. Appl. Surf. Sci. 2010, 256, 34133417.
129. Albiter, M. A.; Huirache-Acuna, R.; Paraguay-Delgado, F.; Zaera, F.; Alonso-
Nunez, G. J. Nanosci. Nanotechnol. 2008, 8, 64376444.
130. Ramos, M.; Berhault, G.; Ferrer, D. A.; Torres, B.; Chianelli, R. R. Catal. Sci.
Technol. 2012, 2, 164178.
131. Bouwens, S. M. A. M.; van Veen, J. A. R.; Koningsberger, D. C.; de Beer, V. H.;
Prins, R. J. Phys. Chem. 1991, 95, 123134.
132. Kasztelan e, M. W. J.; Louwers, S. P. A.; de Beer, V. H. J.; Kasztelan, S.; Prins, R.;
van der Kraan, A. M. J. Phys. Chem. 1992, 96, 54455452.
133. Sun, M.; Burgi, T.; Cattaneo, R.; van Langeveld, D.; Prins, R. J. Catal. 2001, 201,
258269.
134. Ferdous, D.; Dalai, A. K.; Adjaye, J. J. Mol. Catal. A: Chem. 2005, 234,
169179.
135. Okamoto, Y.; Kato, A.; Usman, N.; Rinaldi, T.; Fujikawa, H.; Koshika, I.;
Hiromitsu, T.; Kubota, T. J. Catal. 2009, 265, 216228.
136. Shido, T.; Prins, R. J. Phys. Chem. B 1998, 102, 84268435.
137. de Jong, A. M.; Borg, H. J.; van IJzendoorn, L. J.; Soudant, V. G. F. M.; de
Beer, V. H. J.; van Veen, J. A. R.; Niemantsverdriet, J. W. J. Phys. Chem. 1993,
97, 64776483.
138. Weber, Th.; Muijsers, J. C.; Niemantsverdriet, J. W. J. Phys. Chem. 1995, 99,
91949200.
139. Muijsers, J. C.; Weber, Th.; van Hardeveld, R. M.; Zandbergen, H. W.;
Niemantsverdriet, J. W. J. Catal. 1995, 157, 698705.
140. Roukoss, C.; Laurenti, D.; Devers, E.; Marchand, K.; Massin, L.; Vrinat, M. C. R.
Chimie 2009, 12, 683691.
141. Wivel, C.; Candia, R.; Clausen, B. S.; Marup, S.; Topsoe, H. J. Catal. 1981,
68(453463), 433452.
142. Topsoe, H.; Candia, R.; Topsoe, N.-Y.; Clausen, B. S. Bull. Soc. Chim. Belg.
1984, 93, 783806.
143. Topsoe, H.; Clausen, B. S. Appl. Catal. 1986, 25, 273293.
144. Payen, E.; Kasztelan, S.; Houssenbay, S.; Szymanski, R.; Grimblot, J. J. Phys.
Chem. 1989, 93, 65016506.
145. Polz, J.; Zeilinger, H.; Mueller, B.; Knozinger, H. J. Catal. 1989, 120, 2228.
146. Mangnus, P. J.; Poels, E. K.; van Langeveld, A. V.; Moulijn, J. A. J. Catal. 1992,
137, 92101.
147. Mangnus, P. J.; Poels, E. K.; Moulijn, J. A. Ind. Eng. Res. 1993, 32, 18181821.
148. Scheffer, B.; van Oers, E. M.; Arnoldy, P.; de Beer, V. H. J.; Moulijn, J. A. Appl.
Catal. 1986, 25, 303311.
149. Scheffer, B.; Amoldy, P.; Moulijn, J. A. J. Catal. 1988, 112, 516527.
150. Vissers, J. P. R.; Scheffer, B.; de Beer, V. H. J.; Moulijn, J. A.; Prins, R. J. Catal.
1987, 105, 277284.
151. Bouwens, S. M. A. M.; Koningsberger, D. C.; de Beer, V. H. J.; Louwers, S. P. A.;
Prins, R. Catal. Lett. 1990, 5, 273284.
152. Bouwens, S. M. A. M.; Koningsberger, D. C.; de Beer, V. H. J.; Prins, R. Catal.
Lett. 1988, 1, 5559.
153. Bouwens, S. M. A. M.; Prins, R.; de Beer, V. H. J.; Koningsberger, D. C. J. Phys.
Chem. 1990, 94, 37113718.
154. Bouwens, S. M. A. M.; Koningsberger, D. C.; de Beer, V. H. J.; Prins, R. ACS Div.
Pet. Chem. Prep. 1988, 33, 596.
155. Bouwens, S. M. A. M.; van Zon, F. B. M.; van Dijk, M. P.; van der Kraan, A. M.;
deBeer, V. H. J.; van Veen, J. A. R.; Koningsberger, D. C. J. Catal. 1994, 146,
375393.
156. Louwers, S. P. A.; Prins, R. J. Catal. 1992, 133, 94111.
157. Craje, M. W. J.; de Beer, V. H. J.; van Veen, J. A. R.; van der Kraan, A. M. J. Catal.
1993, 143, 601615.
320 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
158. Eijsbouts, S.; Heinerrnan, J. J. L.; Elzerman, H. J. W. Appl. Catal. A 1993, 105,
5368.
159. Eijsbouts, S.; Heinerrnan, J. J. L.; Elzerman, H. J. W. Appl. Catal. A 1993, 105,
6982.
160. Eijsbouts, S.; van Leerdam, G. C. Bull. Soc. Chim. Belg. 1995, 104, 347352.
161. van der Kraan, A. M.; Craje, M. W. J.; Gerkema, E.; Ramselaar, W. T. M.; de
Beer, V. H. J. Hyper. Inter. 1989, 46, 567574.
162. van Veen, J. A. R.; Gerkema, E.; van der Kraan, A. M.; Hendriks, P. J. A. M.;
Beens, H. J. Catal. 1992, 133, 112123.
163. Spevack, P. A.; McIntyre, N. S. J. Phys. Chem. 1993, 97, 1103111036.
164. Berrebi, G.; Roumieu, S. Bull. Soc. Chim. Belg. 1987, 96, 967976.
165. van Gestel, J.; Leglise, J.; Duchet, J. C. J. Catal. 1994, 145, 429436.
166. Hallie, H. Oil Gas J. 1982, 80, 6974.
167. Suzuki, T.; Dufresne, E. Stud. Surf. Sci. Catal. 1995, 92, 215.
168. Weber, Th.; Muijsers, J. C.; van Wolput, J. H. M. C.; Verhagen, C. P. J.;
Niemantsverdriet, J. W. J. Phys. Chem. 1996, 100, 1414414150.
169. Prasad, T. P.; Diemann, E.; Muller, A. J. Inorg. Nucl. Chem. 1973, 35,
18951904.
170. Patterson, T. A.; Carver, J. C.; Leyden, D. E.; Hercules, D. M. J. Phys. Chem.
1976, 80, 17021708.
171. Cattaneo, R.; Weber, Th.; Shido, T.; Prins, R. J. Catal. 2000, 191, 225236.
172. Leliveld, R. G.; van Dillen, A. J.; Geus, J. W.; Koningsberger, D. C. J. Catal. 1997,
171, 115129.
173. Coulier, L.; de Beer, V. H. J.; van Veen, J. A. R.; Niemantsverdriet, T. W. Top.
Catal. 2000, 13, 99108.
174. de Jong, A. M.; de Beer, V. H. J.; van Veen, J. A. R.; Niemantsverdriet, J. W.
J. Phys. Chem. 1996, 100, 1772217724.
175. Shimizu, T.; Hiroshima, K.; Honma, T.; Mochizuki, T.; Yamada, M. Catal. Today
1998, 45, 271276.
176. Ohta, Y.; Shimizu, T.; Honma, T.; Yamada, M. Stud. Surf. Sci. Catal. 1999, 127,
161168.
177. Medici, L.; Prins, R. J. Catal. 1996, 163, 3849.
178. Cattaneo, R.; Shido, T.; Prins, R. J. Catal. 1999, 185, 199212.
179. Topse, H.; Clausen, B. S.; Candia, R.; Wivel, C.; Mrup, S. J. Catal. 1981, 68,
433452.
180. Chianelli, R. R.; Ruppert, A. F.; Behal, S. K.; Kear, B. H.; Wold, A.; Kershaw, R.
J. Catal. 1985, 92, 5663.
181. Topse, N. Y.; Topse, H.; Srensen, O.; Clausen, B. S.; Candia, R. Bull. Soc.
Chim. Belg. 1984, 93, 727733.
182. Craje, M. W. J.; de Beer, V. H. J.; van der Kraan, A. M. Hyper. Inter. 1991, 69,
795798.
183. Craje, M. W. J.; de Beer, V. H. J.; van der Kraan, A. M. Appl. Catal. 1991, 70,
L7L13.
184. Craje, M. W. J.; de Beer, V. H. J.; van der Kraan, A. M. Hyper. Inter. 1991, 69,
799802.
185. Craje, M. W. J.; Gerkema, E.; de Beer, V. H. J.; van der Kraan, A. M. Hyper. Inter.
1990, 57, 17951800.
186. Startsev, A. N. ACS Div. Pet. Chem. Prep. 1994, 39, 434.
187. Kampers, E. W. H.; Maas, T. M. J.; van Grondelle, J.; Brinkgreve, P.;
Koningsberger, D. C. Rev. Sci. lnstrum. 1989, 60, 26352638.
188. Kulkarni, G. U.; Rao, C. N. R. Catal. Lett. 1991, 11, 6370.
189. Louwers, S. P. A.; Prins, R. ACS Div. Pet. Chem. Prep. 1990, 35, 211.
190. Reinhoudt, H. R.; Troost, R.; van Langeveld, A. D.; Sie, S. T.; van Veen, J. A. R.;
Moulijn, J. A. Fuel Proc. Technol. 1999, 61, 133147.
191. Breysse, M.; Bachelier, J.; Bonnelle, J. P.; Cattenot, M.; Cornet, D.; Decamp, T.;
Duchet, J. C.; Engelhard, P.; Frety, R.; Gachet, C.; Geneste, P.; Grimblot, J.;
Gueguen, C.; Kasztelan, S.; Lacroix, M.; Lavalley, J. C.; Leclercq, C.; Moreau, C.;
de Mourgues, L.; Olive, J. L.; Payen, E.; Portefaix, J. L.; Toulhoat, H.; Vrinat, M.
Bull. Soc. Chim. Belg. 1987, 96, 829844.
192. Breysse, M.; Cattenot, M.; Decamp, T.; Frety, R.; Gachet, C.; Lacroix, M.;
Leclercq, C.; de Mourgues, L.; Portefaix, J. L.; Vrinat, M.; Houari, M.; Grimblot, J.;
Kasztelan, S.; Bonnelle, J. P.; Housni, S.; Bachelier, J.; Duchet, J. C. Catal. Today
1988, 4, 3955.
193. Vissenberg, M. J.; van der Meer, Y.; Hensen, E. J. M.; de Beer, V. H. J.; van der
Kraan, A. M.; van Santen, R. A.; van Veen, J. A. R. J. Catal. 2001, 198, 151163.
194. van der Vlies, A. J.; Kishan, G.; Niemantsverdriet, J. W.; Prins, R.; Weber, Th.
J. Phys. Chem. B 2002, 106, 34493457.
195. van der Vlies, A. J.; Prins, R.; Weber, Th. J. Phys. Chem. B 2002, 106,
92779285.
196. Reinhoudt, H. R.; van Langeveld, A. D.; Stockmann, R. M.; Prins, R.;
Zandbergen, H. W.; Moulijn, J. A. J. Catal. 1998, 179, 443450.
197. Reinhoudt, H. R.; Troost, R.; van Langeveld, A. D.; van Veen, J. A. R.; Sie, S. T.;
Moulijn, J. A. J. Catal. 2001, 203, 509515.
198. Reinhoudt, H. R.; van Langeveld, A. D.; Mariscal, R.; de Beer, V. H. J.;
Veen, J. A. R.; Sie, S. T.; Moulijn, J. A. Stud. Surf. Sci. Catal. 1997, 106,
263271.
199. Payen, E.; Kasztelan, S.; Grimblot, J.; Bonelle, J. P. Catal. Today 1988, 4,
5770.
200. Reinhoudt, H. R.; van der Meer, Y.; van der Kraan, A. M.; van Langeveld, A. D.;
Moulijn, J. A. Fuel Process. Technol. 1999, 61, 4354.
201. Zuo, D.; Li, D.; Nie, H.; Shi, Y.; Lacroix, M.; Vrinat, M. J. Mol. Catal. A 2004, 211,
179189.
202. Vissenberg, M. J.; Joosten, L. J. M.; Heffels, M. E. H.; van Welsenes, A. J.; de
Beer, V. H. J.; van Santen, R. A.; van Veen, J. A. R. J. Phys. Chem. B 2000, 104,
84568461.
203. Hensen, E. J. M.; van der Meer, Y.; van Veen, J. A. R.; Niemantsverdriet, J. W.
Appl. Catal. A: Gen. 2007, 322, 1632.
204. Topse, N.-Y. Catal. Today 2006, 113, 5864.
205. Candia, R.; Villadsen, J.; Topse, N.-Y.; Clausen, B. S.; Topse, H. Bull. Soc.
Chim. Belg. 1984, 93, 763773.
206. Daage, M.; Chianelli, R. R. J. Catal. 1994, 149, 414427.
207. Chianelli, R. R. Catal. Rev. Sci. Eng. 1984, 26, 361393.
208. Topse, H.; Clausen, B. S.; Topse, N.-Y.; Pedersen, E.; Niemann, W.; Muller, A.;
Bogge, H.; Lengeler, B. J. Chem. Soc., Faraday Trans. 1987, 83,
21572167.
209. Bataille, F.; Lemberton, J.; Michaud, P.; Perot, G.; Vrinat, M.; Lemaire, M.;
Schulz, E.; Breysse, M.; Kasztelan, S. J. Catal. 2000, 191, 409422.
210. Raybaud, P.; Hafner, J.; Kresse, G.; Kasztelan, S.; Toulhoat, H. J. Catal. 2000,
190, 128143.
211. Chianelli, R. R. Oil Gas Sci. Technol. 2006, 61, 503513.
212. Topse, H. Appl. Catal. A: Gen. 2007, 322, 38.
213. Topse, H.; Clausen, B. S. Catal. Rev. Sci. Eng. 1984, 26, 395420.
214. Hinnemann, B.; Norskov, J. K.; Topsoe, H. J. Phys. Chem. B 2005, 109,
22452253.
215. Eijsbouts, S.; van den Oetelaar, L. C. A.; van Puijenbroek, R. R. J. Catal. 2005,
229, 352364.
216. Okamoto, Y.; Hioka, K.; Arakawa, K.; Fujikawa, T.; Ebihara, T.; Kubota, T. J. Catal.
2009, 268, 4959.
217. Afanasiev, P. J. Catal. 2010, 269, 269280.
218. Polyakov, M.; van den Berg, M. W. E.; Hanft, T.; Poisot, M.; Bensch, W.;
Muhler, M.; Grunert, W. J. Catal. 2008, 256, 126136.
219. Polyakov, M.; Poisot, M.; Bensch, W.; Muhler, M.; Grunert, W. J. Catal. 2008,
256, 137144.
220. Brorson, M.; Carlsson, A.; Topse, H. Catal. Today 2007, 123, 3136.
221. Lauritsen, J. V.; Bollinger, M. V.; Laegsgaard, E.; Jacobsen, K. W.; Nrskov, J. K.;
Clausen, B. S.; Topse, H.; Besenbacher, F. J. Catal. 2004, 221, 510522.
222. Lauritsen, J. V.; Kibsgaard, J.; Helveg, S.; Topse, H.; Clausen, B. S.;
Besenbacher, F. Nat. Nanotechnol. 2007, 2, 5358.
223. Lauritsen, J. V.; Besenbacher, F. Adv. Catal. 2006, 50, 97.
224. Lauritsen, J. V.; Kibsgaard, J.; Olesen, G. H.; Moses, P. G.; Hinnemann, B.;
Helveg, S.; Nrskov, J. K.; Clausen, B. S.; Topse, H.; Laegsgaard, E.;
Besenbacher, F. J. Catal. 2007, 240, 218231.
225. Deepak, F. L.; Esparza, R.; Borges, B.; Lopez-Lozano, X.; Jose-Yacaman, M. ACS
Catal. 2011, 1, 537543.
226. Besenbacher, F.; Lauritsen, J. V.; Wendt, S. Nano Today 2007, 2, 3037.
227. Besenbacher, F.; Brorson, M.; Clausen, B. S.; Helveg, S.; Hinnemann, B.;
Kibsgaard, J.; Lauritsen, J. V.; Moses, P. G.; Nrskov, J. K.; Topse, H. Catal.
Today 2008, 130, 8696.
228. Krebs, E.; Silvi, B.; Raybaud, P. Catal. Today 2008, 130, 160169.
229. Krebs, E.; Daudin, A.; Raybaud, P. Oil Gas Sci. Technol. 2009, 64, 707718.
230. Chianelli, R. R.; Berhault, G. Catal. Today 1999, 53, 357366.
231. Kelty, S. P.; Berhault, G.; Chianelli, R. R. Appl. Catal. Gen. 2007, 322, 915.
232. Berhault, G.; Mehta, A.; Pavel, A. C.; Yang, J.; Rendon, L.; Yacaman, M. J.;
Araiza, L. C.; Duarte Moller, A.; Chianelli, R. R. J. Catal. 2001, 198, 919.
233. Chianelli, R. R.; Berhault, G.; Torres, B. Catal. Today 2009, 147, 275286.
234. Sakashita, Y.; Yoneda, T. J. Catal. 1999, 185, 487495.
235. Araki, Y.; Honna, K.; Shimada, H. J. Catal. 2002, 207, 361370.
236. Lauritsen, J. V.; Helveg, S.; Lgsgaard, E.; Stensgaard, I.; Clausen, B. S.;
Topse, H.; Besenbacher, F. J. Catal. 2001, 197, 15.
237. Herbst, K.; Brorson, M.; Carlsson, A. J. Mol. Catal. A: Chem. 2010, 325, 17.
238. Chianelli, R. R.; Berhault, G.; Raybaud, P.; Kasztelan, S.; Hafner, J.; Toulhoat, H.
Appl. Catal. A: Gen. 2002, 227, 8396.
239. Oyama, S. T.; Gott, T.; Zhao, H.; Lee, Y.-K. Catal. Today 2009, 143, 94107.
240. Hensen, E. J. M.; de Beer, V. H. J.; van Veen, J. A. R.; van Santen, R. A. J. Catal.
2003, 215, 353357.
241. Egorova, M.; Prins, R. J. Catal. 2004, 221, 1119.
 Hydrotreating: Removal of Sulfur from Crude Oil Fractions with Sulfide Catalysts
242. Breysse, M.; Djega-Mariadassou, G.; Pessayre, S.; Geantet, C.; Vrinat, M.;
Perot, G.; Lemaire, M. Catal. Today 2003, 84, 129138.
243. Topse, N.-Y.; Topse, H. J. Catal. 1993, 139, 641651.
244. Kabe, T.; Aoyama, Y.; Wang, D.; Ishihara, A.; Qian, W.; Hosoya, M.; Zhang, Q.
Appl. Catal. A: Gen. 2001, 209, 237247.
245. Farag, H.; Sakanishi, K.; Kouzu, M.; Matsumura, A.; Sugimoto, Y.; Saito, J. Catal.
Commun. 2003, 4, 321326.
246. Vogelaar, B. M.; Kagami, N.; van Langeveld, A. D.; Eijsbouts, S.; Moulijn, J. A.
Prep. Pap. Am. Chem. Soc. Div. Fuel Chem. 2003, 48, 548549.
247. Texier, S.; Berhault, G.; Perot, G.; Harle, V.; Diehl, F. J. Catal. 2004, 223,
404418.
248. Kasahara, S.; Shimizu, T.; Yamada, M. Catal. Today 1997, 35, 5964.
249. Guernalec, N.; Geantet, C.; Raybaud, P.; Cseri, T.; Aouine, M.; Vrinat, M. Oil Gas
Sci. Technol. 2006, 61, 515525.
250. Vogelaar, B. M.; Eijsbouts, S.; Bergwerff, J. A.; Heiszwolf, J. J. Catal. Today 2010,
154, 256263.
251. Froment, G. F. Catal. Rev. 2008, 50, 118.
252. Okamoto, Y. Catal. Today 2008, 132, 917.
253. Eltzner, W.; Breysse, M.; Lacroix, M.; Vrinat, M. Polyhedron 1986, 5, 203210.
254. Alonso, G.; Aguirre, G.; Rivero, I. A.; Fuentes, S. Inorg. Chim. Acta 1998, 274,
108110.
255. Poisot, M.; Bensch, W.; Fuentes, S.; Alonso, G. Thermochim. Acta 2006, 444,
3545.
256. Poisot, M.; Bensch, W.; Fuentes, S.; Ornelas, C.; Alonso, G. Catal. Lett. 2007,
117, 4352.
257. Bocarando, J.; Alonso-Nunez, G.; Bensch, W.; Huirache-Acuna, R.; Del Valle, M.;
Cruz-Reyes, J. Catal. Lett. 2009, 130, 301307.
258. Bocarando, J.; Alonso-Nunez, G.; Huirache-Acuna, R.; Bensch, W.; Fuentes, S.;
Petranovskii, V.; Huang, Z.-D. Appl. Catal. A: Gen. 2009, 363, 4551.
259. Nava, H.; Pedraza, F.; Alonso, G. Catal. Lett. 2005, 99, 6571.
260. Romero-Rivera, R.; Camacho, A. G.; Del Valle, M.; Alonso, G.; Fuentes, S.;
Cruz-Reyes, J. Top. Catal. 2011, 54, 561567.
321
261. Huang, Z.-D.; Bensch, W.; Lotnyk, A.; Kienle, L.; Fuentes, S.; Bocarando, J.;
Alonso, G.; Ornelas, C. J. Mol Catal. A: Chem. 2010, 323, 4551.
262. Huang, Z.-D.; Bensch, W.; Kienle, L.; Fuentes, S.; Alonso, G.; Ornelas, C. Catal.
Lett. 2008, 124, 2433.
263. Rayo, P.; Ramirez, J.; Rana, M. S.; Ancheyta, J.; Aguilar-Elguezabal, A. Ind. Eng.
Chem. Res. 2009, 48, 12421248.
264. Tian, R.; Shen, B.; Wang, F.; Lu, C.; Xu, C. Energy Fuel 2009, 23, 5559.
265. Breysse, M.; Afanasiev, P.; Geantet, C.; Vrinat, M. Catal. Today 2003, 86,
516.
266. Breysse, M.; Portefaix, J. L.; Vrinat, M. Catal. Today. 1991, 10, 489505.
267. Saih, Y.; Segawa, K. Appl. Catal. A: Gen. 2009, 353, 258265.
268. Zepeda, T. A.; Pawelec, B.; Fierro, J. L. G.; Montesinos, A.; Olivas, A.; Fuentes, S.;
Halachev, T. Micropor. Mesopor. Mater. 2008, 111, 493506.
269. Huang, Z.-D.; Bensch, W.; Kienle, L.; Fuentes, S.; Alonso, G.; Ornelas, C. Catal.
Lett. 2008, 122, 5767.
270. Huang, Z.-D.; Bensch, W.; Kienle, L.; Fuentes, S.; Alonso, G.; Ornelas, C. Catal.
Lett. 2009, 127, 132142.
271. Huang, Z.-D.; Bensch, W.; Kienle, L.; Alonso, G.; Bocarando, J.; Fuentes, S.;
Ornelas, C. J. Mol. Catal. A. Chem. 2010, 323, 4551.
272. Vradman, L.; Landau, M. V.; Herskowitz, M.; Ezersky, V.; Talianker, M.; Nikitenko, S.;
Koltypin, Y.; Gedanken, A. Stud. Surf. Sci. Catal. 2003, 146, 721724.
273. Landau, M. V.; Vradman, L.; Herskowitz, M.; Koltypin, Y.; Gedanken, A. J. Catal.
2001, 201, 2236.
274. Gutierrez, O. Y.; Klimova, T. J. Catal. 2011, 281, 5062.
275. Lipsch, J. M. J. G.; Schuit, G. C. A. J. Catal. 1969, 15, 179189.
276. Voorhoeve, R. J. H. J. Catal. 1971, 23, 236242.
277. Voorhoeve, R. J. H.; Stuiver, J. C. M. J. Catal. 1971, 23, 228235.
278. Farragher, A. L.; Cossee, P. In Proc. 5th Int. Cong. Catal., North Holland,
Amsterdam; Hightower, J. W., Ed.; , 1977 p 1301.
279. Hagenbach, G.; Delmon, B. C. R. Acad. Sci. Paris, Serie C 1971, t273, 1489.
280. Grange, P. Catal. Rev. Sci. Eng. 1980, 21, 135181.
281. Delmon, B. Bull. Soc. Chim. Belg. 1979, 88, 979987.
 9.16 X-Ray Spectroscopy
S DeBeer, Max Planck Institute for Chemical Energy Conversion, Mulheim an der Ruhr, Germany; Cornell University, Ithaca, NY, USA
F Neese, Max Planck Institute for Chemical Energy Conversion, Mulheim an der Ruhr, Germany
2013 Elsevier Ltd. All rights reserved.

9.16.1 Introduction 427

9.16.2 X-Ray Spectra and Data Interpretation 428
9.16.2.1 Ligand K-Edge Spectra 428
9.16.2.2 Metal K-Edge Spectra 428
9.16.2.3 Metal L-Edge Spectra 429
9.16.2.4 X-Ray Emission 429
9.16.2.5 EXAFS 430
9.16.3 Experimental Methods 431
9.16.4 Theoretical Calculation of X-Ray Spectra 432
9.16.4.1 General Considerations 432
9.16.4.1.1 Transition energies 433
9.16.4.1.2 Transition intensities 433
9.16.4.1.3 Spinorbit coupling 433
9.16.4.1.4 Continuum states 433
9.16.4.1.5 The rising edge 433
9.16.4.2 Time-Dependent DFT 434
9.16.4.3 Ligand K-Edge Spectra 434
9.16.4.4 Metal K-Edge Spectra 435
9.16.4.5 Metal L-Edge Spectra 435
9.16.4.6 X-Ray Emission 437
9.16.5 Conclusion 439
References 439

Abbreviations TDA TammDancoff approximation

DFT Density functional theory TD-DFT Time-dependent density functional theory
ECP Effective core potentials XAS X-Ray absorption spectroscopy
EXAFS Extended X-ray absorption fine structure XES X-Ray emission spectroscopy
FT Fourier transform ZORA Zeroth-order regular approximation
SOC Spinorbit coupling
SOMO Singly occupied molecular orbital

9.16.1 Introduction
X-Ray core-level spectroscopy provides a fundamental experi-
mental means to obtain insight into the geometric and elec-
tronic structure of inorganic systems. These methods can
provide information on metal oxidation state, spin state, and
covalency, as well as identify the nature and distance of
coordinated ligands. In addition, X-ray spectroscopy is
inherently element specific, providing far greater selectivity
than many other spectroscopic methods. This chapter focuses
primarily on X-ray absorption (XAS) and X-ray emission
spectroscopies (XESs), with a focus on the fundamental phys-
ical phenomena, the experimental methods, and the more
detailed information that can be obtained through the appli-
cation of quantum chemical calculations.
XAS involves the excitation of core electrons to bound states
localized on the photoabsorber and the eventual excitation of
the photoelectron to the continuum.1 The resulting spectra are
typically divided into two regions the edge region, which
provides electronic structure information and the extended
X-ray absorption fine structure (EXAFS) region , which provides
information about the distance, number, and type of ligands. By
selectively tuning the incident X-ray energy to the binding en-
ergy of an absorber, one can, in principle, obtain XAS data on
any element of interest. Thus for transition-metal complexes,
measurements can be made both from the perspective of the
ligand and from the metal. Here we discuss the information that
can be obtained utilizing ligand K- and/or metal K-/L-edge XAS.
XES involves monitoring the processes which occur after
ionization of a 1s core electron.2 The most likely event is the

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00918-9 427

428 X-Ray Spectroscopy
observation of a Ka fluorescence line, which occurs when a 2p
electron refills a 1s core hole. Kb (3p1s) and valence-to-core
transitions are also possible, with each region of the XES spec-
trum providing different but complementary information.
For much of X-ray core spectroscopy, the analysis of experi-
mental data has relied largely on empirical observations and
semi-quantitative correlations. Herein, we briefly review the
more qualitative developments. We then discuss the more recent
developments in quantitative analysis. It is the close correlation
of the experimental data to quantum chemistry calculations,
which greatly enhances the information that can be obtained,
providing a detailed picture of the electronic structure. The
complementary information that can be obtained from the
EXAFS region is also briefly discussed. Recent theoretical and
experimental developments are highlighted.
9.16.2 X-Ray Spectra and Data Interpretation
An XAS edge results when a core electron absorbs energy equal to
or greater than its binding energy. XAS edges are labeled accord-
ing to the shell the electron originates from. A K-edge corre-
sponds to a 1s core level; L-edges correspond to 2s and 2p
levels; and M-edges to 3s, 3p, and 3d levels, as shown in Figure 1.
For reference, the Fe K-edge occurs at 7.1 keV, the L-edges at
700 eV, and the M-edges between 50 and 100 eV. The large
range of energies spanned by K-, L-, and M-edge core excitations
require different experimental setups and also provide access to a
much wider range of electronic structure information. Here, the
content of ligand K-, metal K-, and metal L-edges are described,
with a focus on the complementarity of the information that can
be obtained.
9.16.2.1 Ligand K-Edge Spectra
Ligand K-edge XAS involves monitoring the processes occur-
ring after the excitation of a ligand 1s electron. By monitoring
the absorption coefficient, m, as a function of energy, one can
obtain an insight into the effective charge or oxidation state on
an absorbing atom. In addition, lower-energy pre-edge features
provide information about the nature of bound open-shell
metal ions. This was clearly illustrated by the early work of
Hodgson, Hedman, and Solomon, which examined the Cl
K-edge spectra of D2d-[ZnCl4]2 and [CuCl4]2, as shown in
Figure 2.3 The primary edge jump at  2827 eV corresponds to
a dipole-allowed Cl 1s to 4p transition. In the case of a d10
metal system as D2d-[ZnCl4]2, no lower-energy transition is
observed. However, in the case of d9 [CuCl4]2 a pre-edge
feature occurs at 2820 eV. This feature arises from the
L-edges
K-edge
1s 2s 2p 3s 3p 3d Continuum
Energy
Figure 1 Atomic energy level diagram for an absorbing atom and the
corresponding X-ray absorption edges.
presence of an unoccupied molecular orbital, which contains
both Cl 3p and Cu 3d characters. The intensity of the pre-edge
feature can be correlated to the oscillator strength, which can
be factored into two components: (1) the covalency and (2)
the radial transition moment dipole integral.4 If one assumes a
constant transition moment dipole integral, then the intensity
reflects the covalency of the metalligand bond. However, as
the total charge on the ligand changes, the assumption of a
constant transition moment dipole integral fails and a critical
point in the quantitative analysis of pre-edge intensities as a
reflection of covalency is how this factorization is handled.
This has been shown to have a particularly pronounced effect
in the analysis of S K-edge spectra of highly covalent transition-
metal dithiolene complexes.5,6 The details of this analysis are
discussed further in this chapter.
9.16.2.2 Metal K-Edge Spectra
In the case of a metal K-edge for a first-row transition metal, the
intense edge jump results from an electric-dipole-allowed 1s to
4p transition. Superimposed on the rising edge are weak pre-
edge transitions, which correspond to 1s to 3d transitions.
These transitions are formally electric-dipole forbidden, but
quadrupole allowed and can gain intensity through 3d4p
mixing in suitable symmetry.7,8 Hence pre-edges of centrosym-
metric complexes (e.g., Oh) tend to be weak, while the pre-edge
of noncentrosymmetric complexes (e.g., Td) typically have
significantly higher intensities (Figure 3). Metal K-edges are
most frequently used to identify the oxidation state of an
absorbing atom. In a simple picture, as a metal becomes
more oxidized, it takes more energy to ionize a core electron
and the edge shifts up in energy. Figure 4(a) compares the Fe
K-edge of ferrous and ferric FeS4 model complexes. For a
change of one unit of oxidation state, the pre-edge shifts by
1 eV.8,9 However, even when the oxidation state is constant,
changes in the local geometry can induce a pronounced change
2
D2d[ZnCl4]2-
D2d[CuCl4]2-
Normalized absorption
1.5
1
0.5
0
2815 2820 2825 2830
Energy (eV)
Figure 2 Comparison of the Cl K-edge spectra for D2d-[ZnCl4]2 and
[CuCl4]2. Reprinted with permission from Hedman, B.; Hodgson, K. O.;
Solomon, E. I. J. Am. Chem. Soc. 1990, 112, 16431645. Copyright 1990
American Chemical Society.
 X-Ray Spectroscopy 429

in the edge. This is clearly illustrated when comparing the data of thumb are not applicable when the coordination environ-
for ferro- and ferricyanide (Figure 4(b)) to the data for ferrous ment changes. For example, Fe(V) species have been identified
and ferric FeS4 model complexes (Figure 4(a)). in which the pre-edge energy falls in the same range as typical
A detailed analysis of the intensity and energy distributions Fe(IV) species.11,12 Thus care must be taken when using such
of 40 ferrous and ferric model complexes, by Westre et al., has generalized fingerprints for electronic structure. These discrep-
shown that the Fe K-pre-edge features may be correlated di- ancies have motivated our interest in more rigorous interpre-
rectly with the coordination number.8 In general, centrosym- tation of the pre-edge region, in order to obtain more detailed
metric six-coordinate complexes will have much weaker insights, and also to enable these spectra to be used in a
pre-edge intensities than noncentrosymmetric five-coordinate predictive fashion.1517 Recent density functional theory
complexes. These model studies have thus been extended to (DFT)-based approaches to understanding the edge region are
numerous studies of nonheme iron enzyme active sites to described below.
determine the coordination number of active sites at various
stages of an enzymatic reaction.
The shifts in XAS pre-edge energies have also provided 9.16.2.3 Metal L-Edge Spectra
essential information for determining the oxidation states of Metal L-edge XAS corresponds to monitoring dipole-allowed
high-valent metal species. In particular, Fe(IV), Fe(V), and Fe 2p3d transitions of a transition-metal absorber. In contrast to
(VI) species have been characterized using XAS.1013 For a the weak dipole-forbidden 1s3d pre-edge transitions at a
series of complexes with similar coordination environments, metal K-edge, the metal L-edge provides an intense dipole-
an increase in one unit of oxidation state results in an 1 eV allowed probe of the 3d-manifold. Due to spinorbit coupling
increase in the pre-edge energy.8,14 However, such simple rules (SOC) of the 2p hole, 2p3/2 and 2p1/2 final states result, giving
rise to the L3 and L2 edges.18 In a manner similar to the ligand
K-edge, it has been shown that the intensity of these features
1.2 should reflect the total metal 3d character in the unoccupied
molecular orbitals. Therefore, the more covalent a transition-
1.0 metal complex is, the lower the L-edge intensity will be.19,20
Normalized absorption

This is, of course, a very qualitative picture, and again one must
0.8 take care in accounting for the changes in the radial transition-
moment integral which will also contribute to the observed
0.6 intensities. In the case of d1d8 systems, additional splitting of
Edge the L3 and L2 edges due to the multiplet effect is observed.
0.4 1s to 4p
Pre-edge
1s to 3d 9.16.2.4 X-Ray Emission
0.2
XES involves following the emission of photons by the decay of
0.0 electrons after ionization of the metal 1s electron. The most
7105 7110 7115 7120 7125 7130 7135 likely event is a Ka emission line, which corresponds to a
Energy (eV) dipole-allowed 2p1s transition. Experimentally, the Ka line
Figure 3 The edge and pre-edge region for Td (solid line) and Oh will be split into two features due to the SOC of the 2p hole,
(dashed line) ferric complexes. resulting in Ka1,2 and Ka2,3 lines. Approximately one order of

1.2

2.0
1.0
Normalized absorption

Normalized absorption

0.8 1.5

0.6
1.0

0.4

0.5
0.2

0.0 0.0
7105 7110 7115 7120 7125 7130 7135 7105 7110 7115 7120 7125 7130 7135
(a) Energy (eV) (b) Energy (eV)
Figure 4 Comparison of the normalized Fe K-edges for oxidized (solid line) and reduced (dashed line) FeS4 model complexes (a) and
ferri- (solid line) and ferrocyanide (dashed line) (b).
430 X-Ray Spectroscopy
Kb /Kb
2,5
Kb
1,3
Normalized intensity
100
7020 7040 7060 7080 7100
Energy (eV)
Figure 5 Fe Kb XES spectrum for a high-spin ferric model complex. The
box indicates the valence-to-core region of the spectrum, which has been
expanded by 100-fold above. A simplified energy diagram showing the
origin of the Kb1,3 and valence-to-core transitions is given on the right.
Adapted from Lee, N.; Petrenko, T.; Bergmann, U.; Neese, F.; DeBeer,
S. J. Am. Chem. Soc. 2010, 132, 97159727.
magnitude lower in intensity than the Ka lines are the K b lines,
corresponding to dipole-allowed 3p1s transitions (Figure 5).
The Kb1,3 emission line corresponds to an electric-dipole-
allowed 3p1s transition, which will have significant contribu-
tions from 3p3d electron exchange interactions as well as 3p
SOC contributions. To higher energy are transitions that corre-
spond to valence electron transitions into the metal 1s core hole,
the so-called Kb2,5/Kb" (or valence-to-core) region. These tran-
sitions have been previously assigned as ligand np to metal 1s
(Kb2,5) and ligand ns to metal 1s (Kb or satellite) transitions.2
Recently, systematic studies on a series of low-molecular-
weight ferrous and ferric complexes were carried out.21 These
results demonstrate that the Kb main-line spectra are dominated
by spin-state contributions (Figures 6(a) and 6(c)), while the
valence-to-core spectra (Figures 6(b) and 6(d)) show a greater
sensitivity to the chemical environment. Significant efforts have
been placed on understanding the valence-to-core region. The
experimental data demonstrate that the energy and intensity
distributions are affected by spin state, oxidation state, and
coordination environment and thus provide a detailed finger-
print of the local geometric and electronic structure.
9.16.2.5 EXAFS
The EXAFS region primarily provides metrical information for
the number and type of ligands surrounding a photoabsorber,
as also discussed in Chapter 9.05.1 When the photoelectron
has been excited to the continuum, it can be modeled as a
wave, which propagates out from the photoabsorber and is
backscattered by the electron density surrounding the neigh-
boring atoms. The EXAFS region is defined in terms of w(k),
where1
mk  m0 k
wk ,
m0 k
X Nf k
wk e2k s sin 2kR dk
2 2
s
kR 2
and
r
Fe 3d 2me E  E0
k [2]
2
Lp
Ls
In eqn [1], m(k) is the observed absorption coefficient and
m0(k) is the atomic absorption (or the absorption coefficient
that corresponds only to the absorbing atom, without any
Fe 3p
surrounding ligands). For convenience, EXAFSs are expressed
in terms of photoelectron wave vector space (k) rather than
energy space. The relationship between energy (in eV) and k (in
Fe 1s
A1) is given in eqn [2]. Here E0 defines the point at which the
EXAFSs start (i.e., k 0) and represents the ionization thresh-
old energy for the electron.
In its simplest form, EXAFSs can be approximated as pho-
toelectron scattering, where
7120 7140
X ei2kR h i i2kR
idk e
wk s kR
kf ke [3]
kR
This equation can be expressed as a sine function, where
X Nf k
wk s kR2
sin 2kR dk [4]
Here N refers to the number of scatterers of a given type, f(k)
is the amplitude function for the backscattering atom, d(k) is
the phase shift for the absorberbackscatterer pair, and R is the
absorberscatterer (i.e., metalligand) distance.
Averaging over all atoms and introducing the disorder
parameter s2 then gives the EXAFS equation for a single
absorberscatterer pair, as
X Nf k
e2k s sin 2kR dk
2 2
wk s kR2
[5]
The contributions of each parameter to the EXAFS data are
best illustrated through examples.
Figure 7 shows the dependence of the EXAFS data on the
metalligand distance (R) together with the Fourier transform
(FT) of the data into R-space (i.e., A). The FT allows one to
visualize the radial distribution of electron density with respect
to the central absorbing atom. As the FeO bond is elongated
from 2 to 3 to 4 A, the amplitude of the EXAFS decreases and
the frequency increases (Figure 7, inset). This is more readily
visualized by examining the FT of the data, which shows the
intensity of a scattering contribution at a given distance from
the photoabsorber. It should be noted that the FTs are not
phase-shift corrected, and thus the distances cannot be read
directly from the FT. In any case, this plot clearly indicates that
the decrease in intensity as the FeO distance is increased. As
illustrated by eqn [5], the data have a 1/R2 dependence dem-
onstrating that this is a local probe of metrical structure.
As also indicated by eqn [5], the EXAFS will have a linear
dependence on the coordination number. As the coordination
number (N) of a given backscatterer increases, the amplitude of
the EXAFS signal and the intensity of the FT increase linearly. A
similar effect is observed for a change in the disorder parameter
(s2), where an increase in disorder decreases the amplitude of
both the EXAFS and the FT, with an exponential dependence.
The identity of the backscatterer will affect both the phase
[1]
and amplitude of the EXAFS signal. Figure 8 shows the calcu-
lated FTs and EXAFS for FeO, FeS, and FeFe vectors. For
 X-Ray Spectroscopy 431

0.12 1.5

0.1 1-
[Fe(lll)Cl4]

Normalized intensity (X 1000)

0
[Fe(lll)(acac)3]
Normalized intensity

0.08 0 1
[Fe(lll)(tpfp)Cl]

0.06

0.04 0.5

0.02

0 0
7020 7030 7040 7050 7060 7070 7080 7090 7085 7090 7095 7100 7105 7110 7115 7120

(a) Energy (eV) (b) Energy (eV)

0.12 4

3.5
0.1 3-
[Fe(lll)(CN)6]
3
Normalized intensity (X 1000)

3+
[Fe(lll)(tacn)2]
0.08
Normalized intensity

1+
[Fe(lll)(tpp)(ImH)2] 2.5

0.06 2

1.5
0.04

0.02
0.5

0 0
7020 7030 7040 7050 7060 7070 7080 7090 7085 7090 7095 7100 7105 7110 7115 7120

(c) Energy (eV) (d) Energy (eV)

Figure 6 Fe Kb main line XES spectrum for high spin ferric complexes (a). Valence-to-core region for the complexes shown in (a) (b). Fe Kb main
line XES spectrum for low-spin ferric complexes (c). Valence-to-core region for the complexes shown in (c) (d). Adapted from Lee, N.; Petrenko, T.;
Bergmann, U.; Neese, F.; DeBeer, S. J. Am. Chem. Soc. 2010, 132, 97159727.

higher atomic number backscatters, the FT intensity increases
and the amplitude envelope of the EXAFS peaks at a higher
k-value. This allows EXAFS to be used to determine the identity
of the atoms ligated to an absorber. However, it should be
noted that similar backscatterers will have similar phase and
amplitude profiles; thus while EXAFS can distinguish oxygen
from sulfur, it cannot distinguish oxygen from nitrogen. In
general for first-row atoms, atomic numbers of Z  1 cannot
be distinguished. Moving down the periodic table, the required
change in Z becomes larger, as the change in one electron
becomes a smaller and smaller perturbation on the change in
overall electron density surrounding the scatterer.
In fitting EXAFS data, the goal is to determine the sum of all
the absorberscatterer interactions that make up the total
EXAFS signal. Using the EXAFS equation, this may be described
as a sum of damped sine waves in k-space. The radial distribu-
tion of backscatterers with respect to the absorber is most
readily visualized by examination of the FT.
9.16.3 Experimental Methods
A brief overview of the experimental setup and sample require-
ments for XAS and XES measurements is described. Typically
both XAS and XES measurements utilize intense tunable X-ray
sources provided by synchrotron facilities. The experimental
setup varies depending on the energy region of interest, with
<2 keV measurements requiring ultra-high vacuum condition,
25 keV measurements requiring a helium environment,
and > 5 keV measurements being performed in air.
432 X-Ray Spectroscopy

0.4 higher resolution, it is worth commenting on the process of

fluorescence-detected XAS in more detail. After the excitation
2
of a core electron, the core hole will be refilled by an electron
3
from a higher-energy shell and a fluorescent photon will
4
0.3 be produced. This assumes that the number of emitted

3
Ka-photons is proportional to the number of photons absorbed,

EXAFS * k
which is a reasonable assumption in the dilute limit. As a
FT magnitude

standard solid-state detector will have an E/DE of 50, no fine

0.2 structure is observed on the emission lines.
XES measurements also use fluorescence detection, but rely
on high-resolution Bragg optics spectrometers (E/DE of 5000
2 4 6 8 10 12 14
-1
k ( )
or greater) to spectrally analyze the fluorescence, which is then
0.1 detected by a standard solid-state detector. Unlike XAS, XES
typically utilizes a fixed-energy incident beam with a photon
energy well above the absorption of the absorber. The spherical
crystal spectrometer is then scanned in energy in a y2y rela-
0 tionship defined by a Rowland circle. Using a Von Hamos
0 1 2 3 4 5 6 geometry with a cylindrical analyzer together with a position-
R () sensitive detector, dispersive XES measurements are possible.
Figure 7 Calculated FTs and EXAFS (inset) for an FeO interaction as it
Variable energy XES which scans through the incident energy is
is elongated from 2 to 3 to 4 A. also possible, and is more commonly referred to as resonant
XES or resonant-inelastic x-ray scattering.
Sample requirements for XAS and XES measurements in the
hard X-ray range are similar. Samples can be run in almost any
1.6 O form, but are most typically run as dilute solids or solutions.
S Typically concentrations of  1 mM or greater in the absorber
Fe of interest are needed. Required volumes will depend on the
3
EXAFS * k

1.2 individual beam line and the size of beam spot, though sample
FT Magnitude

volumes of 50150 microliters are common.

0.8 9.16.4 Theoretical Calculation of X-Ray Spectra

5 10 15
k (-1)
In recent years it has become increasingly commonplace to
supplement XAS and XES measurements with quantum chem-
0.4
ical calculations. The mostly used method is DFT.22 Hence the
focus of the description will be on DFT-based methods. In this
section, the scope and limitations of the available techniques
0 are briefly described. Since the chapter cannot be comprehen-
0 1 2 3 4 5 6 sive, we focus on molecular systems. Molecular techniques can
R () also beneficially be used to study defects in solids. However,
for calculations on extended solid-state systems and infinite-
Figure 8 Calculated FTs and EXAFS (inset) for FeO, FeS, and FeFe
lattice compounds, special techniques are in use but are out-
interactions all at a distance of 2.5 A.
side the scope of this chapter.

Currently, most XAS measurements are performed by scan-

ning the energy of the incident X-ray beam from just below the
ionization threshold of the absorbing atom to well above the
edge. This is achieved using either a crystal monochromator or a
grating (at lower energies) for energy selection through Bragg
reflection. A standard XAS scan spans an energy range
of 100 eV (to over 1000 eV for EXAFS). At high energies
(> 2 keV), the incident-beam intensity is measured using a
gas-filled ionization chamber, whereas at lower energies it is
often more convenient to measure total electron yield from a
standard (such as a metal grid). Absorption by the sample is
then measured either directly by transmission or indirectly by
fluorescence or electron yield. For fluorescence measurements,
most often a solid-state Ge or Si detector is used. As XES mea-
surements are enabled by the ability to detect fluorescence at
9.16.4.1 General Considerations
X-Ray spectroscopic techniques involve the excitation of core-
electrons from deep-lying core-orbitals into semi-occupied or
unoccupied valence orbitals. Both bound final states and con-
tinuum final states are accessed. In addition, strong core-level
SOC effects can prominently influence the spectra shapes. The
requirements on the theoretical methods that are able to deal
with these challenges are very different. Hence, it is customary
to divide the discussion into three separate areas dealing with
the calculation of transition energies and intensities, SOC ef-
fects, and continuum states.
We start by considering the excitation of a core s-electron
into a semi-occupied or empty valence orbital. In this case,
there are no important effects of SOC and the final state is

bound. The overall goal of the theoretical calculations is the
prediction of: (a) transition energies, (b) transition intensities,
and (c) the correct number of spectral features. While (c)
appears to be a trivial requirement, it is among the most
difficult objectives to meet.
9.16.4.1.1 Transition energies
In calculating transition energies, it is important to determine
beforehand whether one is aiming at accurate, absolute, or
relative transition energies. Relative is, in this context, under-
stood as the transitions occurring in one compound or analo-
gous transitions occurring in a series of compounds. The
accurate calculation of absolute transition energies, on the
other hand, is a formidable challenge. To succeed in this,
undertaking requires to accurately capture all differential en-
ergy contributions between the ground state and the core-hole
excited state. Within DFT, this first of all involves the proper
description of core orbitals in an all-electron treatment. Many
contemporary calculations are based on effective core poten-
tials (ECPs), which require additional approximations and
assumptions in the treatment of core-level excitation events.
Unfortunately, contemporary DFT potentials fall short of being
accurate in the core region (as well as in the outer region of an
atom or molecule) and hence core-orbital energies deviate
from their exact KohnSham values by dozens if not hundreds
of eV. Secondly, the proper incorporation of kinematic (scalar)
relativistic effects into the treatment is required. These spin-free
relativistic effects are known to preferentially stabilize s- and
p-orbitals and destabilize d- and f-electrons.23 The lower the
orbital energies, the larger will be the associated effects. Hence,
scalar relativistic effects will make important contributions to
the energies of, say, s!d excitations. Unfortunately, contem-
porary methods to treat relativistic effects, such as the zeroth-
order regular approximation (ZORA)24 for relativistic effects,
do not treat the core orbitals properly and overstabilize
them.24 To some extent, the introduced errors tend to cancel
with the inherent errors in the DFT potentials. However, such a
cancelation is not systematic or reliable. Third, the treatment
should properly capture the differential correlation energies of
the states, which is not a simple task as the correlation energies
of, say, (s2dn) configurations and (s1dn1) configurations in
transition-metal complexes may be very different. Fourth, the
relaxation of the electronic structure due to the creation of a
core hole must be treated accurately and this is, again, anything
but a simple undertaking. It is probably fair to state that
the systematically accurate calculation of absolute transition
energies is not feasible, except when using the most advanced
theoretical methods (based on correlated wavefunction the-
ory) on the smallest systems (e.g., atoms, ions, and very
small molecules). It will be argued below that the calculation
of transition energies that carry large, but highly systematic
errors is much more feasible and leads to very useful pre-
dictions, provided that consistent computational protocols
are employed.
9.16.4.1.2 Transition intensities
The calculation of transition intensities in core-level spectra is
somewhat more involved than for valence-to-valence transi-
tions in molecular systems. The underlying reason is that the
wavelength of the employed radiation is no longer large
X-Ray Spectroscopy 433
compared to molecular dimensions (a few Angstroms). Thus,
one has to recall that the interaction between light and matter
is based on a multipole expansion that is usually terminated
after the first term, which is represented by the electric-dipole
operator. However, a precise treatment shows that the form of
the electric-dipole operator that arises naturally from the treat-
ment is the dipole-velocity form, while in practice the dipole-
length operator is commonly employed.25 In the limit of exact
molecular Eigen states, the two forms are equivalent, but in an
approximate treatment they are not. Fortunately, this seems to
be of limited relevance for the calculation of X-ray spectra
because the dipole length and velocity forms yield almost
indistinguishable results. The next terms in the multipole ex-
pansion are represented by the electric quadrupole and
magnetic-dipole intensity mechanisms. While both are
expected to be about four orders of magnitude smaller than
the electric-dipole mechanism, it is important that they follow
different selection rules. Thus, the requirement for an allowed
electric-quadrupole transition is that the direct product of
the initial and final states with the six components of the
quadrupole operator (transforming as xx, yy, zz, xy, xz, or yz)
contain the totally symmetric representation.26 Magnetic
dipole transitions are allowed when the direct product with
the molecular-rotation operators (Rx, Ry, and Rz) contain the
totally symmetric representation. In practice, electric-
quadrupole transitions have clearly been observed in XAS
spectra and can be experimentally distinguished from electric-
dipole transitions by the angular dependence of the transition
moment in measurement on oriented crystals.27 Evidence for
dominant magnetic-dipole mechanisms has not been obtained
to the best of our knowledge.
9.16.4.1.3 Spinorbit coupling
Large SOC effects arise if the excitation does not arise from a
core-level s-orbital but from a core-level p-orbital. The in-state
SOC of the p5 configuration leads to very large splittings and
a pronounced multiplet structure in the spectra. This is difficult
to model with good accuracy theoretically. In particular,
multiplet treatments in DFT are scarcely developed and the
most common approach is based on heavily parameterized
quasi-atomic multiplet calculations28,29 as will be further elab-
orated below.
9.16.4.1.4 Continuum states
Excitations into the continuum of unbound states cannot be
described with traditional techniques of quantum chemistry.
Here the language of scattering theory must employed, that is
largely a domain of physics. The necessary techniques are so
different from the apparatus used to calculate the pre-edge
region of XAS spectra that even their cursory description is
outside of the range of this chapter. Most research groups
involved in EXAFS employ the FEFF code.30 The underlying
theory is described in the literature.3035
9.16.4.1.5 The rising edge
The rising edge region of XAS spectra represents weakly-bound
states close to the continuum. These states are difficult to
calculate with either molecular, solid state, or scattering tech-
niques. Since the focus of this chapter is X-ray spectroscopy of
molecular systems, we focus below on the calculation of the
434 X-Ray Spectroscopy

pre-edge region of XAS spectra and the valence-to-core region with

of XES spectra.  
I 8  I  4 I I
Akl oI D 2 dkl  o I Dk Dl [16]
15 15

1
X I I 2
X I
9.16.4.2 Time-Dependent DFT Ck oI Qkl Dl  eklm Ml DmI [17]
5 3
l lm
The protocol outlined above is applied below to ligand and
metal K-edge spectra and is therefore described briefly. In DFT Here eklm is the LeviCivitta tensor and Ml is the magnetic
linear-response theory, the transition frequencies o are calcu- transition-dipole moment. This procedure was followed through-
lated from the nonstandard eigenvalue-type equation:22 out the present work. Neglecting the B-matrix amounts to the
      so-called TammDancoff (TDA) approximation. One then sim-
A B X 1 0 X ply solves
o [6]
B A Y 0 1 Y
AX oX [18]
With the supermatrices
    which resembles the configuration interaction with a single
 
Aia, jb ea  ei dai, jb ia~
f XC jb  cHF ijab [7] excitation approach. The key point in solving eqn [6] is to
    project the solution onto the manifold of core-excited single
 
Bia, jb ia~
f XC jb  cHF jckb [8] excitations. This obviously neglects the coupling of the core-
excited K-edge states to L- and M-edge, as well as valence-
And the effective two-electron exchange-correlation operator excited states. The plausibility of this approximation has been
discussed.17
d2 EXC
f~XC r1 ; r2 r12
1
[9] It is most important to emphasize that due to the factors
drr1 drr2 discussed above, the calculations do not yield accurate abso-
For some two-electron operator g(r1,r2), the integrals are lute transition energies. The errors are large but very systematic
defined as such that they can be compensated for by calibration. Such a
calibration involves the comparison of theory and experiment
  
pqgrs cp r1 cq r1 gr1 ; r2 cr r2 cs r2 dr1 dr2 [10] for a set of test molecules. Obviously, the resulting calibration
curve must be done for each element separately and depends
Labels i,j refer to occupied spinorbitals of the KohnSham on the basis set and the specific-density functional used in the
solution, a,b to unoccupied orbitals (p, q, r, s to general spin calculation and on whether relativistic effects have been in-
orbitals), ep are the orbital energies, and cHF is the fraction of the cluded in the treatment or not. It is therefore necessary to
HartreeFock exchange in the density functional. EXC is the employ a consistent protocol. For example, a recent proposal
exchange correlation functional, r(r1) the ground-state electron involves the BP functional, the def2-TZVP basis set, and the
density at position r1, and r12 is the interelectronic distance. ZORA treatment for relativistic effects.15
The solution of eqn [6] yields the transition frequencies o
and transition amplitudes Xai, Yai from which one can calculate
dipole- and quadrupole-transition moments as 9.16.4.3 Ligand K-Edge Spectra
X  
D Xia Yia iD
^ a [11] Ligand K-edge absorption spectra are obtained by exciting a
ia ligand core-electron into semi-occupied or empty valence or-
X   bitals. In practice, this is best done at the chlorine, sulfur, and
Q Xia Yia iQ
^ a [12]
ia
phosphorous edges.
The most salient points are readily appreciated by taking a
^ k rk
D [13] system with a single hole as an example.4 The starting point is
  the frozen orbital approximation to the initial and final states.
^ kl rk rl  1r 2 dkl
Q [14]
3 The initial state is a single Slater determinant:
where the sum r is the position operator of the electron with
jIi jc1 c1 . . . ck ck . . . cn cn cSOMO j [19]
components (k x, y, z, r (rx2 ry2 rz2)1/2).
The calculation of transition-dipole moments is straightfor- and the final state is the one where an electron is promoted
ward, but the calculation of quadrupole-transition moments from a ligand orbital with 1s character, ck, to the singly occu-
presents some difficulty, since the results depend on the choice pied molecular orbital (SOMO)
of origin. Since the quadrupole transition-moment mechanism
arises from a series expansion, it seems natural to choose the jFk i jc1 c1 . . . ck cSOMO . . . cn cn cSOMO j [20]
origin such that the expansion yields optimum convergence. As In the dipole-length approximation to the transition-dipole
discussed in detail by DeBeer George et al.16, it is possible to operator and using the Fermi golden rule, an electric-dipole
readjust the origin for each transition (I) from the solution of a transition has intensity proportional to:
simple linear equation system as:

  2
 ! 
Dk /  cSOMO  r ck  [21]
AI R CI [15]
 X-Ray Spectroscopy 435

where A sums over !

nuclei and i over electrons and ZA is the instrumental for establishing whether the ligand behaves in a
!
charge of atom A. R A is the position of the A0 th nucleus and r noninnocent way or not.5
denotes an electron position. The ground-state singly occupied
orbital, cSOMO, is written
p 9.16.4.4 Metal K-Edge Spectra
cSOMO 1  a2 jMi  ajLi [22]
The main difference between metal- and ligand K-edge spectra in
where the overlap between the metal-centered part of the terms of the computational protocol is the intensity mechanism
SOMO, |Mi, and the ligand-centered part, |Li, is neglected. that is operative. At the metal K-edge, the dominating transitions
Therefore 100  a2 is the percentage ligand character in the in the pre-edge region are of metal-1s!metal-nd character. These
SOMO which is distributed over the individual ligands that transitions are electric quadrupole allowed and electric-dipole
contribute to the SOMO and 100  (1  a2) is the percentage forbidden. Dipole character is induced in these transitions
metal character in the SOMO. The ligand part will in general be due to mixing of d- and p-orbitals at the metal, which, in turn, is
a linear combination of valence ns and np orbitals induced by deviations from centrosymmetry. These are fairly
subtle effects since the p-mixings are small and effectively compete
p
jLi kjnsi 1  k2 jnpi [23] with the quadrupole-intensity mechanism. These factors were all
systematically addressed in a study dealing with iron K-edge
with hybridization coefficient k. The donor MO, ck, is assumed
spectra.17 It was found that the agreement between theory and
to be a pure ligand 1s orbital (see below). The square of the
experiment is also fairly good (Figure 10). Transition energies are
transition-dipole moment becomes
typically computed to an accuracy of better than 0.5 eV and
D E2 1 
 ! transition intensities correlate with experiment with a correlation
 c jr jc  a2 1  k2 hr i2 [24]
 SOMO k  3 np, 1s coefficient R 0.980.99.17

with

9.16.4.5 Metal L-Edge Spectra
hr inp, 1s Radnp jrjRad1s [25]
where Rad denotes the radial part of the orbitals. The expres-
sion has a straightforward interpretation the pre-edge intensity
increases as the ligand character the SOMO increases (a2), as the
ligand in question makes a larger contribution to the total
ligand character in the SOMO (cSOMO 2
) and as the p-character
of this contribution increases (1  kSOMO2
). In absolute terms, for
a particular ligand atom, the factor 1 =3 hr i2np, 1s is the intrinsic
intensity of an atomic 1s!npx,y,z transition. Its value does, how-
ever, not reflect a real atomic transition because the radial
wavefunctions distort in the molecule according to the atoms
environment. Since 1 =3 hr i2np, 1s depends on the molecular envi-
ronment, Solomon and coworkers have found it necessary to
determine its value for different types of bonding situations
(e.g., for sulfur ligands: RS, RSSR, S2 etc.) separately RussellSaunders (LS) coupling scheme. The problem of transfer-
before values of a2 can be determined from the areas under
the experimental XAS absorption curves. However, neither
2
=3 hr inp, 1s itself nor the value of a is a physical observable.
1 2
Only their product is directly related to the oscillator strength
which is an observable. We have therefore chosen a different
approach and directly calculated the physically observable quan-
tity, for example, the oscillator strengths of the observed transi-
tions using a simple protocol outlined above.
The TDDFT protocol outlined above was successfully ap-
plied to a substantial number of sulfur and chlorine containing
transition-metal complexes and was found to be surprisingly
successful. Transition energies are typically predicted within
0.20.3 eV and the correlation between calculated oscillator
strength and properly normalized experimental intensities is
on the order of 10%. An example for the comparison be-
tween theory and experiment is given in Figure 9.
Following the detailed study of Ray et al.,5 DeBeer,
Wieghardt, and coworkers have studied a number of
transition-metal dithiolene complexes where the comparison
between calculated and experimental sulfur K-edge spectra was
The same TDDFT protocol can obviously not be applied to the
calculation of metal L-edge spectra as it (a) does not contain any
SOC effects and (b) does not properly treat multiplet splittings.
The first shortcoming could be fixed by resorting to two-
component TDDFT methods. However, this approach can still
not be successful because the multiplet effects are not properly
accounted for. The term multiplet effects describe the fact that a
number of final states of different energy arise from a given
electronic configuration. If the number of unpaired electrons in
a given electronic configuration exceeds the value of 2S, where S is
the total spin of the target state, then several spin couplings arise,
all of which lead to a given total spin S. In a similar way, the orbital
angular momenta couple to different total values of L. In atoms
and ions this behavior is well understood and summarized in the
ring this scheme to molecules is that the different Russell
Saunders-like multiplet states are not single determinants, even
if the ground state is well described by a single Slater determinant.
The necessary determinants involved in properly building multi-
plet states may well be double- or triply excited relative to the
electronic ground state. Hence, they are not accessible by TDDFT
that only incorporates single excitations. In fact, a DFT based
theory that would properly include these effects has apparently
not been developed.
In principle, one can of course resort to wavefunction-based
configuration interaction methods in which multiplet effects,
spincouplings, and higher excitations can be incorporated ex-
plicitly. These methods suffer from two defects: (a) the high-
computational cost of the calculations and (b) the poor quality
of the HartreeFocktype reference wavefunction from which
these calculations must be started. Thus, enormous levels of con-
figuration interaction would be necessary to capture the physics
of the problem correctly. These extensive calculations are not
feasible and would also be of questionable quality because of
the size-consistency problems that are inherent in configuration
436 X-Ray Spectroscopy

2.0
D2d-[CuCl4]2-
2-
D4h-[CuCl4]

Normalized absorption
1.5 [NiCl4]2-
[CoCl4]2-
[Fe(II)Cl4]2-
1.0
[Fe(III)Cl4]1-
0
[TiCl4]

0.5

0.0
2818 2819 2820 2821 2822 2823 2824
Energy (eV)

6
Calculated pre-edge intensity

0
2733 2734 2735 2736 2737 2738 2739
Energy (eV)
Figure 9 Comparison of experimental (top) Cl K-pre-edge XAS data to the calculated spectra (bottom). Adapted from DeBeer George, S.; Petrenko, T.;
Neese, F. Inorg. Chim. Acta 2008, 361, 965972.

7115
0.20
1+,0,1-,2-
3-,4-
Calculated pre-edge enery (eV)

Calculated quadrupole + dipole contributions

[FeCl4]1-
7114
0.15

7113

[FeCl4]2-
0.10 [Fe(salen)Cl]
7112

0.05
7111 [Fe(acac)3]
[FeCl6]3-
[Fe(CN)6]3-
[FeCl6]4-
[Fe(prpep)2]+
[Fe(CN)6]4- [Fe(prpep)2]
7110 0.00
7110 7111 7112 7113 7114 7115 0 5 10 15 20 25
Experimental pre-edge energy (eV) Experimental intensity
Figure 10 Correlation of experimental and theoretical transition energies (left) and intensities (right). Adapted from DeBeer George, S.; Petrenko, T.;
Neese, F. J. Phys. Chem. A 2008, 112, 1293612943.
 X-Ray Spectroscopy 437

interaction treatments. While these problems are all severe, they where DN is the total number of absorbing species in the
may be surmountable in the foreseeable future. irradiated volume, and WGI is the transition rate between the
An alternative approach that has seen much use is the ground state |Gi and state |Ii (capital letters are used for many
parametric multiplet approach as prominently exercised by electron states, lowercase letters for orbitals). WGI is related to
Sawatzky, de Groot, and coworkers.28,29,3639 In these appro- the photoionization cross section sGI(o0) at the incident pho-
aches, one essentially performs an atomic configuration interac- ton frequency o0 as
tion calculation that leads to the correct LS multiplets but
WGI sGI o0 I0 [27]
perturbs these multiplets with ligand-field splittings and
charge-transfer interactions. The necessary parameters involve The emission rate corresponding to the transition of state
metalligand covalencies, which lead to anisotropy in the ligand |Ii to state |Fi is given by the Einstein coefficient AIF which is
field and the electronelectron interaction parameters. Together proportional to the emission oscillator strength fIF
with the ligand-field and charge-transfer parameters that enter
the model, there are a substantial number of empirical para- AIF 3fIF gel [28]
meters that enter the calculations. These parameters are not
where gel is the classical radiative decay rate of the single-electron
straightforward to predict from DFT or ab initio calculations,
oscillator at frequency oIF which is equal to the transition
which means that the main domain of application of the multi-
frequency between the initial and final states (CGS units)
plet model is the fitting of known spectra. Good agreement with
experiment can be achieved in this way, as shown in Figure11, 2e2 o2IF
which shows the calculated and experimental L-edge spectra of a gel [29]
3me c3
low-spin Fe(III) model complex.40
Here, me is the mass of the electron, e is the electron charge,
D is the speed of light, DI is the total decay rate of the core-
9.16.4.6 X-Ray Emission excited state |Ii which is the sum of the radiative (DIR) and
nonradiative (DIN) decay rates
A physically rigorous modeling of the XES experiment would
R N
involve a calculation of the emission process following the DI DXI DI
core-level ionization. This is a fairly involved task as it will DRI AIF0 [30]
require the treatment of the electronic relaxation following the F0
ionization event together with possible spin recouplings to
According to eqns [26][30], the relative intensities of dif-
give spin- and space-symmetry adapted final states. In practice,
ferent transitions in the XES spectrum can be written as
it has become evident that a much simpler approach is suc-
cessful that is described below.21 o2 fIF
The total number of photons emitted per unit time IIF that IRIF X IF 2 [31]
oIF fIF
correspond to the transition of an initial core-ionized or core- F 0

excited state |Ii to the final state |Fi in the limit of low concen-
tration of the absorbing species and small path length of the Then the normalized XES spectrum IIR(oem) corresponding
incident beam is given by to emission from state |Ii is
X
AIF IRI oem IRIF Locm ; oIF G; s [32]
IIF WGI DN [26]
DI F0

10 where L(ocm,oIF,G,s) is a lineshape function which, in general,

depends on homogeneous (G) and inhomogeneous (s) broad-
ening parameters. From eqn [7], it follows that the calculated
8
XES spectrum has an integrated area of unity ( IR(oem) 1),
and thus can be directly compared with the normalized exper-
6 imental XES spectrum. The emission oscillator strength fIF is
related to the absorption oscillator fFI strength as fIF  fFI. The
oscillator strength can be expanded as a power series in terms of
4
the dimensionless fine-structure constant a (1/137.03599).
The first term in this expansion represents the electric-dipole
2 oscillator strength (fIFed). For orientationally averaged transition
rates, the term linear in a is zero, and the next leading term  a2
can be written as the sum of the electric quadrupole (fIFeq) and
0
magnetic dipole (fIFmd) oscillator strengths:21,41
700 705 710 715 720 725 730
Energy (eV) eq
fIF fIFed fIFmd fIF [33]
Figure 11 Comparison of calculated (dashed line) and experimental
The mathematical form of the electric-dipole, magnetic-
(full line) L-edge X-ray absorption spectra of [Fe(tacn)2]3. Reprinted
with permission from Wasinger, E. C.; de Groot, F. M. F.; Hedman, B.; dipole, and electric-quadrupole contributions to this oscillator
Hodgson, K. O.; Solomon, E. I. J. Am. Chem. Soc. 2003, 125, strength are the same as described above for the absorption
1289412906. Copyright 2003 American Chemical Society. process.
438 X-Ray Spectroscopy

The one-electron approximation consists of employing a set
of frozen ground-state orbitals for the calculations. These are
conveniently taken to be KohnSham DFT orbitals. Associated
with the KohnSham procedure is the ground-state Kohn
Sham determinant
jGi ji . . . j . . . nj
In the one-electron approach, the initial core-excited and
final states are constructed as
jIi ja . . . j . . . nj
jFi ja . . . i . . . nj
For the specific application at hand, orbital i is a core-
orbital, for example, the metal 1s-derived orbital of a given
absorber atom, a is an arbitrary unoccupied orbital (or a con-
tinuum unbound one-electron level), and j is another semi-
core or valence orbital of the same atom.
Rather than trying to obtain accurate total energies for all of
these states, the simple approximation is adopted for the emis-
sion energy
EsIF osIF esj  esi [37]
This is of course fairly simplistic. In terms of time-dependent
DFT, it can be shown that the orbital energy difference is a well-
defined approximation to the state energy difference if no
HartreeFock exchange is present in the DFT potential.22
[34] Calculations indicate that DFT in conjunction with the one-
electron approach does not provide an accurate modeling of
the Kb main line. This is presumably due to: (a) a general
underestimation of the core-level spin-polarization and (b)
[35] the lack of a proper treatment of the multiplet structure giving
rise to the actual Kb main line peaks.
[36] On the other hand, the one-electron approach has been
found to be very successful in the modeling of the Kb2,5 region,
representing valence-to-core transitions. A comparison be-
tween calculated and measured spectra for a series of Fe(II)
and Fe(III) complexes is shown in Figure 12 and demonstrates
that the main features of the spectrum are all faithfully repro-
duced by the calculations. Thus it appears that this region
mainly reflects the ground-state electronic structure. The same
comments concerning the actual calibration procedure apply

1.5 80
1-
[Fe(III)Cl4]
0 70
[Fe(III)(acac)3]
0
[Fe(III)(tpfp)Cl]
Normalized intensity (X 1000)

60
Calculated intensity

1
50

40

30
0.5

20

10

0 0
7085 7090 7095 7100 7105 7110 7115 7120 7085 7090 7095 7100 7105 7110 7115 7120
(a) Energy (eV) (b) Energy (eV)

4 150
3+
[Fe(III)(tacn)2]
3.5 3-
[Fe(III)(CN)6]
1+
[Fe(III)(tpp)(ImH)2]
Normalized intensity (X 1000)

3
Calculated intensity

100
2.5

1.5
50

0.5

0 0
7085 7090 7095 7100 7105 7110 7115 7120 7085 7090 7095 7100 7105 7110 7115 7120
(c) Energy (eV) (d) Energy (eV)

Figure 12 Experimental valence-to-core spectra for high-spin ferric complexes (a) and the corresponding calculated spectra (b). Experimental
valence-to-core spectra for low-spin ferric complexes (c) and the corresponding calculated spectra (d). A 2.5 eV broadening and a constant shift of
182.5 eV have been to all calculated spectra. Reprinted with permission from Lee, N.; Petrenko, T.; Bergmann, U.; Neese, F.; DeBeer, S. J. Am. Chem.
Soc. 2010, 132, 97159727. Copyright 2010 American Chemical Society.

as in the case of the TDDFT-based XAS calculations. In
the calibration study by Lee et al., 21 it was found, however,
that the errors in the calculated transition energies are somewhat
larger than in the case of XAS and amount to about 1 eV. The
reasons for this increase in the error are not fully understood
and may involve the limitations of the one-electron approach
but also uncertainties in the experimental calibration procedure.
9.16.5 Conclusion
We hope that this chapter has demonstrated the complemen-
tarity of XAS, XES, and EXAFS. When the results of these
techniques are viewed together, one can obtain a fairly detailed
picture of the geometric and electronic structure of the investi-
gated systems. Most importantly, the techniques are highly
element specific and such highly selective measurements can
even be performed on complex systems. In order to employ
synchrotron-based techniques, it is not necessary to use isoto-
pically enriched samples and it is also not necessary for the
species to be paramagnetic. All of these advantages have added
to the growing popularity of synchrotron-based X-ray spectro-
scopic techniques over the past decades in several research
fields ranging from material sciences over biological chemistry
to heterogeneous and homogenous catalyses. In this chapter,
we have provided a cursory overview of the experimental and
theoretical techniques that are presently in large-scale use. We
hope that it became evident or at least plausible that one can
take the analysis of the information content of the experimen-
tal data to the next higher level through the combination with
quantum chemical calculations. These calculations can now
efficiently be performed with user-friendly programs and are
no longer confined to expert territory. Together with the
increasing availability of bright beam lines and ever improving
X-ray detectors at the various synchrotrons worldwide, there
appears to be every reason to be optimistic about the future of
synchrotron-based spectroscopic techniques.
References
1. Koningsberger, D. C.; Prins, R. X-ray Absorption, Principles, Applications
Techniques of EXAFS, SEXAFS, and XANES. Wiley: New York, 1988.
2. Glatzel, P.; Bergmann, U. Coord. Chem. Rev. 2005, 249, 6595.
3. Hedman, B.; Hodgson, K. O.; Solomon, E. I. J. Am. Chem. Soc. 1990, 112,
16431645.
4. Neese, F.; Hedman, B.; Hodgson, K. O.; Solomon, E. I. Inorg. Chem. 1999, 38,
48544860.
5. Ray, K.; DeBeer George, S.; Solomon, E. I.; Wieghardt, K.; Neese, F. Chem. Eur. J.
2007, 13, 27832797.
6. Ray, K.; Petrenko, T.; Wieghardt, K.; Neese, F. Dalton Trans. 2007, 15521566.
7. Roe, A. L.; Schneider, D. J.; Mayer, R. J.; Pyrz, J. W.; Widom, J.; Que, L. J. Am.
Chem. Soc. 1984, 106, 16761681.
8. Westre, T. E.; Kennepohl, P.; DeWitt, J. G.; Hedman, B.; Hodgson, K. O.;
Solomon, E. I. J. Am. Chem. Soc. 1997, 119, 62976314.
9. DuBois, J. L.; Mukherjee, P.; Collier, A. M.; Mayer, J. M.; Solomon, E. I.; Hedman, B.;
Stack, T. D. P.; Hodgson, K. O. J. Am. Chem. Soc. 1997, 119, 85788579.
X-Ray Spectroscopy 439
10. Aliaga-Alcalde, N.; Mienert, B.; Bill, E.; Wieghardt, K.; George, S. D.; Neese, F.
Angew. Chem. Int. Ed. 2005, 44, 29082912.
11. Berry, J. F.; Bill, E.; Bothe, E.; DeBeer George, S.; Mienert, B.; Neese, F.;
Wieghardt, K. Science 2006, 312, 19371941.
12. de Oliveira, F. T.; Chanda, A.; Banerjee, D.; Shan, X.; Mondal, S.; Que, L., Jr.;
Bominaar, E. L.; Munck, E.; Collins, T. J. Science 2007, 315, 835838.
13. Rohde, J.-U.; In, J.-H.; Lim, M. H.; Brennessel, W. W.; Bukowski, M. R.; Stubna, A.;
Munck, E.; Nam, W.; Que, L., Jr. Science 2003, 299, 10371039.
14. DuBois, J. L.; Mukherjee, P.; Stack, T. D. P.; Hedman, B.; Solomon, E. I.;
Hodgson, K. O. J. Am. Chem. Soc. 2000, 122, 57755787.
15. DeBeer George, S.; Neese, F. Inorg. Chem. 2009, 49, 18491853.
16. DeBeer George, S.; Petrenko, T.; Neese, F. Inorg. Chim. Acta 2008, 361, 965972.
17. DeBeer George, S.; Petrenko, T.; Neese, F. J. Phys. Chem. A 2008, 112,
1293612943.
18. de Groot, F. M. F. Coord. Chem. Rev. 2005, 249, 3163.
19. DeBeer George, S.; Metz, M.; Szilagyi, R. K.; Wang, H. X.; Cramer, S. P.; Lu, Y.;
Tolman, W. B.; Hedman, B.; Hodgson, K. O.; Solomon, E. I. J. Am. Chem. Soc.
2001, 123, 57575767.
20. George, S. J.; Lowery, M. D.; Solomon, E. I.; Cramer, S. P. J. Am. Chem. Soc.
1993, 115, 29682969.
21. Lee, N.; Petrenko, T.; Bergmann, U.; Neese, F.; DeBeer, S. J. Am. Chem. Soc.
2010, 132, 97159727.
22. Neese, F. Coord. Chem. Rev. 2009, 253, 526563.
23. Hess, B. A.; Marian, C. M. In Computational Molecular Spectroscopy; Jensen, P.,
Bunker, P. R., Eds.; Wiley: New York, 2000; p 169ff.
24. van Lenthe, E.; Snijders, J. G.; Baerends, E. J. J. Chem. Phys. 1996, 105,
65056516.
25. Steinfeld, J. I. Molecules and Radiation. Harper and Row: New York, 1979.
26. Cotton, F. A. Chemical Applications of Group Theory. Wiley Interscience:
New York, 1990.
27. Solomon, E. I. Comments Inorg. Chem. 1984, 3, 227.
28. Glatzel, P.; Bergmann, U.; de Groot, F.; Cramer, S. P. Abst. Pap. Am. Chem. Soc.
2002, 224, 025-INOR.
29. Degroot, F. M. F. J Electron Spectros Relat Phenom 1993, 62, 111130.
30. Ankudinov, A. L.; Ravel, B.; Rehr, J. J.; Conradson, S. D. Phys. Rev. B 1998, 58,
75657576.
31. Ankudinov, A. L.; Conradson, S. D.; de Leon, J. M.; Rehr, J. J. Physical Review B
1998, 57, 75187525.
32. Rehr, J. J.; Ankudinov, A.; Zabinsky, S. I. Catal. Today 1998, 39, 263269.
33. Ankudinov, A. L.; Rehr, J. J.; Low, J. J.; Bare, S. R. Top. Catal 2002, 18, 37.
34. Ankudinov, A. L.; Bouldin, C. E.; Rehr, J. J.; Sims, J.; Hung, H. Physical Review B
2002, 65, 104107.
35. Ankudinov, A. L.; Rehr, J. J.; Low, J. J.; Bare, S. R. J. Chem. Phys. 2002, 116,
19111919.
36. Esteva, J. M.; Karnatak, R. C.; Fuggle, J. C.; Sawatzky, G. A. Phys. Rev. Lett. 1983,
50, 910913.
37. Kotani, A.; Ogasawara, H.; Okada, K.; Thole, B. T.; Sawatzky, G. A. Physical
Review B 1989, 40, 6573.
38. Vanderlaan, G.; Thole, B. T.; Sawatzky, G. A.; Verdaguer, M. Physical Review B
1988, 37, 65876589.
39. Hocking, R. K.; Wasinger, E. C.; Yan, Y. L.; de Groot, F. M. F.; Walker, F. A.;
Hodgson, K. O.; Hedman, B.; Solomon, E. I. J. Am. Chem. Soc. 2007, 129,
113125.
40. Wasinger, E. C.; de Groot, F. M. F.; Hedman, B.; Hodgson, K. O.; Solomon, E. I.
J. Am. Chem. Soc. 2003, 125, 1289412906.
41. Berry, J. F.; DeBeer George, S.; Neese, F. Phys. Chem. Chem. Phys. 2008, 10,
43614374. http://www.cec.mpg.de Max-Plank-Institut fur Chemische
Energiekonversion.
Relevant Website
http://www.cec.mpg.de Max-Plank-Institut fur Chemische Energiekonversion.
 9.13 Electron Paramagnetic Resonance
ML Munzarova, Masaryk University, Brno, Czech Republic
2013 Elsevier Ltd. All rights reserved.

9.13.1 EPR Parameters from Experiments 360

9.13.1.1 Spins and Magnetic Moments 360
9.13.1.2 Interaction of Spins with Magnetic Fields: The g-Tensor 360
9.13.1.3 The Effect of Magnetic Nuclei: Hyperfine Coupling 361
9.13.1.4 Systems with More Than One Unpaired Electron: The Zero-Field Splitting 362
9.13.2 Hyperfine Coupling from Quantum Chemistry: Nonrelativistic Approach 362
9.13.3 g-Tensors from Quantum Chemistry 366
9.13.4 Hyperfine Coupling from Quantum Chemistry: Relativistic Contributions 369
9.13.5 ZFS from Quantum Chemistry 370
9.13.6 DFT Calculations of the Hyperfine Coupling Tensors 370
9.13.7 Ab initio Post-HartreeFock Calculations of HFCCs 374
9.13.8 DFT Calculations of g-Tensors 375
9.13.9 Ab initio Post-HartreeFock Calculations of g-tensors 376
9.13.10 DFT Calculations of ZFS 376
9.13.11 Ab initio Calculations of ZFS 377
9.13.12 Conclusion 378
Acknowledgments 379
References 379

Abbreviations MO Molecular orbital

A-tensor Hyperfine coupling tensor NEVPT2 Second-order n-electron valence state
ADF Amsterdam Density Functional perturbation theory
AMFI Atomic mean field NMR Nuclear magnetic resonance
CP Coupled-perturbed PK PedersonKhanna approach
DDCI Difference-dedicated configuration PSO Paramagnetic spinorbit
interaction SA-CASSCF State-averaged CASSCF
DFT Density functional theory SO Spinorbit
DZP Double-zeta polarized basis SOC Spinorbit coupling
DZVP Double-zeta valence-polarized basis SOMO Singly occupied molecular orbital
EPR Electron paramagnetic resonance SORCI Spectroscopy-oriented configuration
GGA Generalized-gradient approximation interaction
HF HartreeFock SSC Spinspin coupling
HFCC Hyperfine coupling constant TZP Triple-zeta polarized basis
IGLO Individual gauge for localized orbitals ZFS Zero-field splitting

Nomenclature j n,l Single-electron spinorbit coupling

Ak Hyperfine coupling tensor component constant
parallel to the symmetry axis Adip(N) Dipolar hyperfine coupling constant of
A? Hyperfine coupling tensor component nucleus N
!
perpendicular to the symmetry axis Aiso(N) Isotropic hyperfine coupling constant of
H Magnetic field nucleus N
^S
H The spin Hamiltonian AN Hyperfine coupling tensor of
^ ZFS
H Zero-field splitting term in the nucleus N
Hamiltonian D Zero-field splitting tensor
Pab
m,n Spin density matrix D,E Zero-field splitting parameters
S^ Spin angular momentum operator g g-Tensor
S^Z Operator for the projection of the spin g g-Value
angular momentum into the z-axis ge Free-electron g-value
rab
N Spin density at the nucleus N IN Spin of nucleus N

Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00915-3 359

360 Electron Paramagnetic Resonance
S ^
Electron spin (eigenvalue of S)
Tkl Components of the anisotropic hyperfine
coupling tensor
Dg Deviation of the g-tensor from the free-
electron g-tensor
9.13.1 EPR Parameters from Experiments
9.13.1.1 Spins and Magnetic Moments
The technique of electron paramagnetic resonance (EPR) spec-
troscopy may be regarded as a fascinating extension of the
famous SternGerlach experiment. In 1922, Stern and Gerlach1
observed that a beam of silver atoms is split into two compo-
nents by an inhomogeneous magnetic field. As a theoretical
rationalization of this striking observation that could not be
explained along the lines of quantum mechanics of that time,
Uhlenbeck and Goudsmith2 postulated in 1925 that electrons
possess an intrinsic angular momentumelectron spin. Later,
Diracs formulation of relativistic quantum mechanics showed
the necessity of invoking an internal degree of freedom for an
electron, with this degree of freedom having the properties of an
angular momentum.3 The operator for this so-called spin angu-
lar mometum is a vector operator labeled S.
By a scalar product of S with itself, we can create a scalar
operator S^ . Its eigenvalues are equal to s(s 1), where s 0,
2
1/2, 1, 3/2, . . . is the spin quantum number, and the eig-
envalues represent the observable quantity square of the
spin angular momentum. The magnitude of the angular
momentum itself is given by the square root {s(s 1)}1/2.
The operator S^ commutes with the operator S^Z that corre-
2
sponds to the projection of the spin angular momentum to
the z-axis. We can thus determine wavefunctions that are eigen-
functions of both S^ and S^Z . The eigenvalues of S^Z are denoted
2
SZ, the allowed values running in integral steps of s from s to
s. For an electron, s 1/2 and the values of SZ are 1/2.4,5
The most transparent physical manifestation of electron
spin is that the electron has a magnetic moment. The magnetic
moment is directly proportional to the spin angular momen-
tum and we have, for the magnetic moment operator,
m^e ge be S^ [1]
Here, be is the Bohr magneton, equal to 9.274 0154 
1024 J1 T1, and ge is called the free-electron g-value, which
is equal to 2.0023 193 043 86 and is one of the most accurately
known physical constants.
9.13.1.2 Interaction of Spins with Magnetic Fields:
The g-Tensor
In the presence of an external magnetic field, the degeneracy of
the states with Sz 1/2 and SZ  1/2 is lifted, and their
separation increases linearly with the magnetic field H, as
shown in Figure 1. This is known as the electronic Zeeman
effect. In the case of free electrons, the levels are separated by
DU hu ge be H [2]
Dgxx, Dgyy, Dgzz Principal components of the Dg-tensor
Dg Deviation of the g-value from the
free-electron g-value
be Bohr magneton
bN Nuclear magneton
!
By varying the magnitude of the static magnetic field H , one
may change the energy level separation as indicated. During an
EPR experiment, a microwave radiation is applied in addition
to the magnetic field. Resonance is said to occur when the
applied magnetic field induces a splitting whose energy is just
equal to that of the applied microwave radiation. The spectro-
scopic act is, then, the absorption of a photon and flipping of
the spin between the two quantized levels, SZ  1/2.
In a general case of the S 1/2 system (one unpaired
electron in a molecule), the energy separation becomes
DU hn g be H [3]
Here g, the g-value is characteristic of the sample and is to
be determined from experiments. Comparing eqns [2] and [3],
we see that real systems are behaving as though the spin
magnetic moment differed from the free spin value by an
amount Dg g  ge. For organic radicals, the deviation Dg is
small, usually in the order of parts per thousand (ppt), but for
systems containing heavier atoms, such as transition-metal com-
plexes, the variations can be much larger. For example, the
isotropic g-value is equal to 1.935 for the complex [MoO
(SCN)5]2, so that Dg 1.935  2.0023  0.0673, that is,
67.3 ppt, and is equal to 2.13 for the Cu(acac)2 complex, so
that Dg 2.130  2.0023 0.1277, that is, 127.7 ppt.
So far, we have been discussing the g-value under isotropic
conditions, that is, in liquids with low viscosity, where the
molecules are not ordered and are tumbling freely. In such a
case, the result of a measurement does not depend on the
relative orientation of the magnetic field and the sample.
Much more information on the spin Hamiltonian parameters
can be obtained from solid-state EPR spectroscopy, where the
spectral observable is a 3  3 tensor called the g-tensor. The
g-tensor can be made diagonal through appropriate choice of
the coordinate axes, known as the principal axes system.
Its elements are denoted gxx, gyy, gzz for a rhombic system and
g?, g?, gk for a uniaxial system.
Magnetic resonance spectroscopists seek to characterize and
interpret spectra quantitatively using a concept known as the
spin Hamiltonian. In general, the Hamiltonian is an operator
representing the energy of the system. While conventional
microscopic Hamiltonians contain kinetic and potential
energy operators for all particles involved, the spin Hamilto-
nian is a phenomenological Hamiltonian that contains only
spin operators and applied fields, together with numerical
parameters that serve as coupling constants between the for-
mer. In the presence of one unpaired electron and a magnetic
field, the spin Hamiltonian adopts the form
^
^S ! !
H S gbe H [4]
 Electron Paramagnetic Resonance 361

^
!
where S is the total electron spin vector operator and g is proportional to the anisotropy of the spin density distribution
the g-tensor parameterizing
!
the interaction
!
between S^ and around the magnetic nucleus. Thus, for single crystals, the spec-
!^
magnetic field H . The term S gbe H is often referred to as tra are dependent on orientation, unless the spin density distri-
the electron Zeeman term. bution is completely isotropic. In the liquid phase, the
anisotropy averages to zero. The Fermi-contact interaction is
proportional to the probability of the unpaired electron to be
9.13.1.3 The Effect of Magnetic Nuclei: Hyperfine Coupling found at the point of the nucleus and, thus, to the isotropic part
If the system contains a nucleus with nonzero spin in addition of the spin density distribution.
to the unpaired electron, the electronic and nuclear spins inter- The spin Hamiltonian in the presence of the electron
act. The two Zeeman energy levels are split, as is shown in Zeeman and the hyperfine interactions adopts the form
Figure 2, which causes the so-called hyperfine splitting of the ! X!
^ !^
H ^
^ S SgbeH S AN IN [5]
EPR signal. While the transition that would occur in the
N
absence of the hyperfine splitting is shown by the dashed !^
line, the two solid lines represent the two transitions observed where I N is the nuclear spin vector operator of the nucleus N
in the presence of the hyperfine splitting. Note that each of the and the summation is over all magnetic nuclei. AN is the hy-
two transitions occurs between levels of identical MI value. This perfine coupling tensor, parameterizing the interaction
^
! ^
!
corresponds to the selection rules DMS  1, DMI 0 for an between S and I N , and it adopts a general form
EPR absorption. 0 1
Axx Axy Axz
There are two mechanisms of electron spinnuclear spin AN @ Ayx Ayy Ayz A [6]
interaction (the magnetic dipolar interaction) and the so-called Azx Azy Azz
Fermi-contact interaction. The magnetic dipolar interaction is
For the description of the experimental outcome the term
hyperfine splitting is used; when describing the size of the elec-
Ms
a Ua = +12 ge b eH tron spinnuclear spin interaction, the term hyperfine coupling
is preferred, which stresses the coupling interaction between
+1 the two spins. Hence, we write below of the hyperfine cou-
2
pling tensor and the hyperfine coupling constant (HFCC).
U = hn The hyperfine coupling tensor can be transferred to the so-
U called principal axes system, that is, a system of axes in which
-1 the matrix of the tensor is diagonal. It then adopts a form
2
which can be further divided into the isotropic and the aniso-
b Ub = - 1 ge b eH
H=0 H 2 tropic parts,
0 0 1 0 1
Resonance Axx 0 0 Aiso 0 0
@ 0 Ayy 0 0 A@ 0 Aiso 0 A
field Hr
0 0 Azz 0 0 0 0 A iso 1
Figure 1 Energy level scheme for the simplest system (e.g., free T11 0 0
electron) as a function of applied magnetic field H, showing EPR @ 0 T22 0 A [7]
absorption. Ua and Ub represent the energies of the Ms 1/2 and 0 0 T33
Ms  1/2 states. Along the lines of Figure 1.2 in the work of Weil et al.4
where the former has elements Aiso (Axx0 Ayy0 Azz0 )/3, with
Aiso called the isotropic hyperfine coupling constant, and the
MI = + 1
2 latter is traceless (T11 T22 T33 0). For magnetic nuclei
MI = 1 located on at least threefold symmetry axis, the anisotropic
2
part adopts the form (Adip, Adip, 2Adip), where Adip is the
Ms = + 12 so-called dipolar coupling constant. The complete hyperfine
coupling tensor is obtained as a sum of the anisotropic tensor
and the diagonal tensor of the form (Aiso, Aiso, Aiso). For an
axially symmetric system, the total hyperfine tensor is some-
Ms = 1 times written in the form (A?, A?, Ak), where Ak denotes the
2
H=0 H MI = 1 component parallel with the symmetry axis and A? denotes
2
the component perpendicular to the symmetry axis.
MI = + 1 The Adip is nonzero only for molecules in which the singly
2
occupied molecular orbital (SOMO) contains some p character
in main-group compounds and some d character in transition-
metal complexes. The more p or d character it contains, the
Figure 2 Energy levels of a system with one unpaired electron and one larger the dipolar coupling is observed. In a first approxima-
magnetic nucleus with I 1/2 (the free hydrogen atom) as a function of tion, the Fermi-contact interaction, which is proportional to
the magnetic field, and the splitting of the EPR signal. The dashed the probability of the unpaired electron to be found at the
transition would be observed if the hyperfine splitting was zero. Along the point of the nucleus, requires some s character of the SOMO
lines of Figure 2.1 in the work of Weil et al.4 to be nonzero. However, it turns out that, in reality, it can be
362 Electron Paramagnetic Resonance

significant, even for a pure p- or d-type SOMO. The nonzero Since the trace of D is zero, only two independent energy
spherical spin density can be, in these cases, created by the parameters are required. It is common to designate these as
process of the so-called spin polarization, which is described in
3
detail in Section 9.13.2. D  DZ [12a]
2
1
E  DX  DY [12b]
9.13.1.4 Systems with More Than One Unpaired Electron: 2
The Zero-Field Splitting where D and E represent the s axial and rhombic terms,
respectively.
In transition-metal chemistry, many dn configurations contain
Since the ZFS usually strongly dominates the spectral shape, it
more than one unpaired electron. In such cases, the ground
is useful to distinguish three limiting cases: (1) the weak-field case,
state has a total spin S > 1/2. Consequently, the ground state
in which bBH  D, (2) the high-field case, in which bBH D, and
possesses a 2S 1-fold degeneracy, corresponding to various
(3) the intermediate case, where bBH
D. The high-field regime is
values of the total spin in the z-direction (SZ). In the presence
frequently met by organic radicals. In the intermediate case, the
of the magnetic field along the z-axis, the energy depends on
spectra are rather complicated. The transition-metal complexes on
SZ, and the 2S 1-fold degeneracy is lifted. However, the latter
which we focus here frequently realize the weak-field regime. We
degeneracy can be partially lifted even in the absence of the
will, thus, discuss in the following exclusively the weak-field
magnetic field, due to interactions between the individual
regime, where D is commonly of the order of several wavenum-
electrons with unpaired spins. There are two possible mecha-
bers, which is large compared to the commonly employed reso-
nisms of such interaction: (1) the dipolar spinspin interaction
nance energy at X-band frequency (hn gbBH
0.3 cm1).
within pairs of electrons and (2) the spinorbit (SO) coupling
In the weak-field case, it is important to distinguish
that can be considered to arise from the interaction of the spin
between the systems with even numbers and odd numbers of
magnetic moment of one electron with the magnetic field
unpaired electrons. The reason is that, for an odd number of
produced by the orbital motion of the other electron.6
unpaired electrons, no electrostatic interaction can completely
To describe these interactions, the spin Hamiltonian must
lift the degeneracy, so that at least a twofold degeneracy
contain, in addition to the electronic Zeeman term, the so-
^ ZFS . If there are no remains, the so-called Kramers theorem,9 and one EPR transi-
called zero-field splitting (ZFS) term, H
tion can be observed. The remaining doubly degenerate levels
other interactions present (e.g., the magnetic hyperfine inter-
are referred to as Kramers doublets. In the weak-field regime,
action), the spin Hamiltonian is written as
only transitions within each Kramers doublet can be observed
^ ZFS gbe H T  S^ S^T DS^
! !
H ^ eZ H
^S H [8] by EPR techniques. The transition energies can be modeled
by using the so-called rhombogram method, as described by
where D is the so-called ZFS tensor (a 3  3 matrix), S^ is the F. Neese10:
^
!
column 3  1 matrix of the spin operator with the elements S X ,
^ !
! ^ !T
S Y , S Z , and S is its transpose, row matrix. The ZFS tensor In this method, each Kramers doublet is regarded as an effective
is related to the energy as a function of different spin S 1/2 system with an effective g matrix geff depending on each
doublet. The effective g-values then only depend on the value of
orientations.7,8 The tensor D can be written in the principal- E/D and the true values g. If these are taken to be close to 2.00, as is
axes system, where its matrix adopts a diagonal form. The spin most often the case, a single parameter, namely E/D, determines the
Hamiltonian for isotropic g can thus be written as resonance positions in the EPR experiment. In addition, the differ-
 ent Kramers doublets will be populated according to Boltzmann
^ S gbe HX S^X HY S^Y HZ S^Z DX S^X2 DY S^Y2 DZ S^Z2
! ! ! ! ! ! statistics, which means that an estimate of the inter-doublet spacing
H and therefore D and E/D can be obtained by variable temperature
measurements.
[9]
The energy eigenvalues U of this Hamiltonian for a mag- As an example, Figure 3(b) displays the level splitting intro-
netic field parallel with the z-axis can be obtained by a solution duced by ZFS for an S 5/2 system. For high symmetry, D and E
of the secular determinant equation (see page 159 in the work are small or zero. Any asymmetry, however, results in a ZFS that
of Weil et al.4) can be very large. The six levels are split by an axial ZFS into three
  Kramers doublets. The introduction of a magnetic field lifts the
 1 1 
 gbe HZ DZ  U 0 DX  DY  remaining degeneracies, and intradoublet transitions can be
 2 2 
  observed. For E/D 0, gk remains close to 2, but the effective g?
 
 0 DZ  U 0  0 [10]
 1 1  is close to 6. Another limiting situation is a case in which a fairly
 DX  DY 0 gbe HZ DZ  U 
 isotropic single line appears at ca. g 4.3. This occurs when the
 2 2 
ratio of E and D is close to 1/3, which is a limiting case for
In zero magnetic field, the energies are, considering that the rhombic symmetry, cf. page 142 in the work of Symons.11
tensor D is traceless, that is, DZ DX DY 0,
1
UX DZ  DX  DY  DX [11a] 9.13.2 Hyperfine Coupling from Quantum Chemistry:
2 Nonrelativistic Approach
1
UY DZ DX  DY  DY [11b]
2 Besides the experimental determination of EPR parameters, it is,
UZ DZ [11c] in general, advisable to udertake computational investigations
 Electron Paramagnetic Resonance 363

g-Value
987 6 5 4 3 2

10
8 MS = 5/2
6
(i) g1 = gz E/D = 0 4
g3 2
g1
0
10
g MS = 3/2
8
(ii) v
E/D = 0.05 a 6
g2
l 4
g3 u
e 2
0
10
MS = 1/2
(iii) g1 = g2 = g3 = 4.28 8
E/D = 1/3
6
4
2
0
1000 2000 3000 0.00 0.05 0.10 0.15 0.20 0.25 0.30
(a) Magnetic field (G) (c) E/D

MS = 5/2

4D

MS = 3/2
2D
MS = 1/2

6
(b)
A1 + ZFS + Magnetic field

Figure 3 Zero-field splitting effects in an S 5/2 system with a D parameter that is large compared to the microwave frequency. (a) Appearance of
powder spectra for three different values of E/D. (b) Level splittings caused by the zero-field splitting and indication of intradoublet transitions
observed in an EPR experiment. (c) Rhombograms for S 5/2 showing the effective g-value as a function of the rhombicity E/D. Neese, F. In Calculations
of NMR and EPR Parameters; Kaupp, M.; Buehl, M.; Malkin, V. G., Eds.; 2004, p 544. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with
permission.

in parallel. The reason is that quantum-chemical calculations electrons, at the point of the magnetic nucleus N, descibed by
can be performed on radicals with well-defined structure, which a position vector RN.
may be unknown for the experimental system in question. By
Aiso N AFermi-contact N
the comparison of experimental data with quantum-chemically
4p
b b ge gN hS^Z i rab
1
determined EPR parameters on various structures, the problem N R N [13]
of the experimental system structure can be solved.10,12 For the 3 e N
overview of computational methods, see Chapter 9.01 by Shaik. Here, bN is the nuclear magneton, gN the nuclear g-value, and
From all EPR parameters, the most transparent relation to hS^Z i is the expectation value of the z-component of total elec-
the electronic structure exists for the isotropic HFCC (Aiso). In tronic spin which normalizes the total spin population to 1.
the nonrelativistic approximation, isotropic HFCC of the The principal contribution to the isotropic coupling arises
nucleus N, Aiso(N), is related to the spin density rab
N , that is, from the spin population of the valence-shell s orbital on the
the density of the a electrons minus the density of the b atom of interest.
364 Electron Paramagnetic Resonance
The components Tkl of the anisotropic hyperfine coupling
tensor are, in the nonrelativistic approximation, given by13
1
Tkl N be bN ge gN hS^Z i
1
2
X   partially compensated by the positive contribution of the metal
Pmab 5 2
, n hm jrN rN dkl  3rN, k rN, l jn i [14]
m, n
ab
where rN r  RN, Pm,n is the spin density matrix, and m, n is
any pair of occupied molecular orbitals (MOs). For an inter-
pretation of the latter expression, see the work of Atherton.14
The matrix T is always traceless and may be brought to the
diagonal form as discussed above, which, for axially symmetric
systems, adopts the form (Adip, Adip, 2Adip), where Adip is
the dipolar coupling constant. In order to compute the com-
ponents of the anisotropic tensor or, in a special case, the
dipolar coupling constant, we need to compute the anisotropic
part (with respect to the particular magnetic nucleus) of the
spin density distribution. This, roughly expressed, means that
we need to determine the magnetic atom p or d orbital coeffi-
cient in the singly occupied orbital plus the anisotropic ligand
contributions, since the principal contribution to the dipolar
coupling comes from the spin population of valence-shell
p, d, . . . orbitals. The calculation of the anisotropic part is, in
general, simpler than the calculation of the isotropic part, for
the reasons decribed below.
As shown in eqn [13], the dominating nonrelativistic con-
tribution to the isotropic hyperfine coupling is proportional to
the spin density at the position of the magnetic nucleus. In
transition-metal complexes, the unpaired electrons are occu-
pying MOs consisting of ligand orbitals, metal d orbitals, and,
if symmetry allows, also metal p and s orbitals. Whereas the
unpaired spin in metal s orbitals does provide nonzero spin
density at the position of the transition-metal nucleus, as is
shown in Figure 4(a), this is not the case for the p and d
orbitals that possess nodal surfaces passing the nucleus. Still,
the unpaired spin found in metal d (and, to a lesser extent,
in metal p) orbitals does come into contact with the nucleus
in an indirect way, by the process of the spin polarization.
This process can be theoretically described by employing the
so-called unrestricted approach. Using this approach, we,
instead of each doubly occupied MO, employ a pair of a and
b spin-orbitals, whose spatial properties can be different,
because they differently interact with the SOMO(s) containing
a single electron of spin alpha (according to convention) only.
The process of spin polarization transferring the spin den-
sity from the transition-metal ion to the other atoms in a
molecule has been described in detail by Cano et al.15 by
considering the spin distribution in the valence orbitals of
all participating atoms. In contrast, the mechanism transferring
the spin density to the transition-metal nucleus illustrated in
Figure 4(c) involves, additionally, core orbitals of the transi-
tion ion. The complex [Cr(CO)4] possesses the Td symmetry,
which forbids the mixing of the metal 4s orbital with the metal
3d orbitals. Thus, the five singly occupied orbitals, which are
dominated by the five metal d orbitals, have zero contribution
to the spin density at the magnetic nucleus. However, the spin
distribution in core 1s, 2s, and 3s orbitals as well as in doubly
occupied valence orbitals containing metal 4s contribution is
polarized in the following way. If the spin in SOMOs is a, the a
spin in metal 2s orbitals is withdrawn from the nucleus in
order to increase its exchange interaction with the SOMOs. As
a result, the net b spin is left in the 2s orbital at the point of the
nucleus, which brings the dominating negative contribution to
the spin density at the nucleus. This negative contribution is
3s orbital, which is spin-polarized in a way opposite to 2s spin
polarization, due to the required orthogonality of the 2s and 3s
shells. Metal 1s orbital brings a spin-polarization contribution
that is 12 orders of magnitude smaller than the 2s and 3s
contributions. Finally, there is a positive contribution to the
spin density resulting from valence MOs that are bonding
between the metal and the ligands, and which have metal 4s
orbital contribution.16
The complex [Cr(CO)4] represents an extreme case in that
the SOMO contribution to the spin density at the nucleus is
exactly zero. Another extreme example is the ScO molecule
(cf. Figure 4(a)), for which the SOMO is highly isotropic, in
that it is dominated by the metal 4s orbital. The spin-
polarization contributions do exist but they are two orders of
magnitude smaller than the SOMO contribution; hence, the
metal 4s orbitals are much less capable of polarizing the lower-
lying shells than the 3d orbitals.
An intermediate case between the [Cr(CO)4] and the ScO
extremes is represented by the MnO3 molecule, cf. Figure 4(b).
The SOMO is composed of ca. 70% by metal 3d orbital and ca.
10% by metal 4s orbital.17 The 4s component gives a direct
SOMO contribution of 2.83, whereas the 3d component spin-
polarizes the valence and the core shell to provide significant
contributions to the spin density. Especially pronounced is
the large valence-shell spin-polarization contribution, which
arises due to the large overlap between the antibonding SOMO
and the lower-lying doubly occupied bonding MO. The
pronounced valence-shell spin polarization for MnO3 causes
specific difficulties in the theoretical description connected to
a so-called spin-contamination problem described later in
this chapter.
Figure 4 demonstrates that, depending on the SOMO, the
spin-polarization contributions to spin density can, in general,
be quite variable. This is true for the valence-shell spin polariza-
tion, where specific details of bonding strongly influence the
resulting spin density. On the contrary, for the core shell spin
polarization, several aspects can be generalized, as has been
shown in a detailed quantum-chemical analysis performed for
22 first-row transition-metal systems.16 The metal 2s contribu-
tion, if it is significant (in the order of tenths of a.u.), is always
negative. The spin density contributions from metal 2s and
metal 3s orbitals are always opposite in sign. This fact can be
uderstood considering that the process of spin polarization
must preserve the orthogonality of the orbitals. We note that,
similarly, there are contributions to the dipolar coupling con-
stant, Adip, from the metal 2p and 3p shell, which are opposite
in sign. While the 2s and 2p orbitals maximize their exchange
interaction with the SOMO, the 3s and 3p orbitals are forced to
lose some of their exchange to stay orthogonal to their respective
penultimate shell. The ratio between the 2s and 3s (and between
the 2p and 3p) contributions proved to be constant for 12
complexes of manganese, as is shown in Figure 5, and is, thus,
not expected to depend on the bonding situation. However, the
ratio does depend on the particular transition metal involved, as
can be seen from the comparison of parts b and c of Figure 4.
 Electron Paramagnetic Resonance 365

O
O
C
O Cr +
Sc O C
Mn O
O C C
O O

Mn 3d + 4s
SOMO
contributions:

Sc Cr 3d
4s 1.77 2.83

Spin polarization Antibonding

contributions:

- valence Mo(s) Bonding 0

Mn 3d + 4s
Cr 4s


Sc 3d+4s
0.03 1.10 0.31
-metal core shells

3s
0.02 0.30 0.44

2s
0
0.47 1.36

1s

0.03 0.02 0.03

Total r N: 1.79 1.57 0.61

Figure 4 Orbital contributions to the spin density at the metal nucleus for: (a) a complex with large SOMO contribution and negligible spin
polarization, ScO, (b) a complex with comparable SOMO contribution and spin-polarization contribution, MnO3, and (c) a complex with negligible
SOMO contribution and large spin polarization contribution, [Cr(CO)4].

The process of the spin polarization is important also for
understanding the isotropic hyperfine coupling of the ligands.
As has been explained by Cano et al.,15 the spin population of
ligand orbitals is determined by an interplay of spin-
delocalization and spin-polarization mechanisms. Figure 6
illustrates the spin polarization within the complex [Cr
(CN)5NO]3 leading to the hyperfine coupling of the axial
nitrogen atom of 14.8 MHz (isotropic) and 4.8 MHz
(dipolar).18,19 The SOMO is a metal dxy orbital and the axial
ligand lies in its nodal plane. Thus, there is no direct SOMO
contribution to the ligand coupling. The spin density that exists
at axial ligand nuclei can be understood to result from a two-step
mechanism. First, the SOMO polarizes metal dxz and dyz orbitals,
which contribute to the p-bonding to the axial ligand. Second,
ligand p components of the polarized metalligand p-bonding
MOs spin-polarize ligand s orbitals, which provide a direct
contribution to the spin density at the ligand nuclei.19,20
From what has been said above, it is clear that a precise
description of the spin-polarization mechanism is mandatory
for any meaningful calculations of the hyperfine coupling. This
has to be reflected in an appropriate quantum chemistry method
and both sufficiently large and balanced basis set. The strengths
and weaknesses of current quantum chemistry methods for such
description are discussed in Sections 9.13.6 and 9.13.7.
366 Electron Paramagnetic Resonance
9.13.3 g-Tensors from Quantum Chemistry
Whereas the dominant, nonrelativistic contribution to the
hyperfine interaction discussed so far can be computed from
the ground-state spin density, the g-tensor is a more complex
property. It contains information not only about the ground
state, but also about certain magnetically coupled excited
states, and the consideration of relativistic effects is necessary
for any meaningful g-tensor calculation. The deviations from
the free-spin g-value arise mainly because of the presence of
orbital magnetism that interacts with the spin magnetism via
the process of so-called spinorbit coupling (SOC). The SO
interaction is a relativistic effect and is properly derived using
Diracs relativistic treatment of the electron. Qualitatively, the
origin of SO interaction can be rationalized as follows:
If we imagine ourselves riding on an electron in an atom, from our
viewpoint, the nucleus is moving around the electron (as the sun
appears to move around the earth). This apparent motion of the
nucleus produces a magnetic field that interacts with the intrinsic
(spin) magnetic moment of the electron, giving the SO interaction
term in the Hamiltonian. The interaction energy of a magnetic
moment m with a magnetic field B is given as m B. The electrons
spin magnetic moment mS is proportional to its spin S, and the
magnetic field arising from the apparent nuclear motion is propor-
tional to the electrons orbital angular momentum L. Therefore, the
SO interaction is proportional to L S and depends on the relative
orientation of the two vectors.5
1.0
3s
0.5
0.0
r N (a.u.)*
0.5
2s
1.0
1.5
2.0
0.0 1.0 2.0 3.0 4.0 5.0
metal 3d gross orbital spin population
Figure 5 Coreshell spin-polarization contributions in various
manganese complexes are proportional to the total 3d spin population. The
data used for the figure have been published by Munzarova et al.16
NO
NC CN NC
NC CN NC
CN CN
SOMO Spin polarization
at the metal, spin
delocalization to
ligand orbitals
Figure 6 Mechanism of spin density transfer from the SOMO to the axial ligand of [Cr(CN)5NO]3. The SOMO polarizes orbitals dominated by metal dxz
and dyz orbitals that realize the p-bonding to the axial ligand. The px and py orbitals of nitrogen polarize the 1s and 2s orbitals of nitrogen, the latter
transferring the spin density to nitrogen nucleus.
The leading contribution to Dg arises due to the interplay
between the SO Hamiltonian H ^ SO and the so-called orbital-
Zeeman Hamiltonian H ^ OZ . If the effects are small, perturbation
theory21 can be used to account for the effect of the magnetic
field and of the SO coupling Hamiltonian. The g-shifts can
then be roughly estimated through their link to suitable exci-
tation energies, DE. In a simplified way, the components of the
g-tensor can be written as22
gij 2:0023 dij  2 xn, l Lij [15]
where
X
Lij m
hf0 j^li jfm ihfm j^lj jf0 i
[16]
Em  E0
Here, dij is equal to 1 for i j and is zero otherwise, while xn,l is
the single-electron SO coupling constant, which is, in the
simplified approaches, extrapolated from experiments.22 The
Em  E0 values correspond to the energy differences between
the orbital f0 of the unpaired electron and other nearby filled
or empty orbitals fm that are linked to f0 by the magnetic field.
The elements hf0 j^li jfm i and hfm j^lj jf0 i couple the orbital f0 of
the unpaired electron and other nearby filled or empty orbitals
through the angular momentum operator components.
When eqns [15] and [16] are applied to d transition-metal
complexes (for details, see Mabbs and Collison22), we obtain
expressions for g-tensor components in terms of the d-orbital
coefficients in the singly and nearby doubly occupied or empty
MOs, the energy differences between the singly and nearby
doubly occupied or empty MOs, and the single-electron SO
coupling constant. These expressions are given in Figure 7 for
the particular cases of d1 and d9 configurations and the dx2y2 ,
dxy, or dz2 single occupation.
To consider some specific cases, we start with the d1 com-
plexes, where the experimentally and theoretically most studied
compounds contain the vanadyl unit, VO2.11 Examples are the
oxovanadium acetylacetonato complex11 or the oxovanadium
salicylaldiminato (salen) complexes.2426 The unpaired electron
is found in a nearly nonbonding dxy orbital, with lobes directing
between the basal ligands. Because of the very strong p-bonding
between the dxz and dyz pair and the oxygen pp orbitals, the
antibonding p orbitals involving dxz and dyz will be strongly
destabilized relative to the SOMO. Thus, the corresponding
excitation energies are high and we expect gxx and gyy close to
the free-electron g-value, but definitely less than 2.0023 (due to
CN NC CN NC CN
Cr Cr
CN NC CN NC CN
CN CN CN
Spin polarization
at the axial ligand
 Electron Paramagnetic Resonance 367

d dz2 g 2dyz
(z2) (yz)
g 2dyz g 1dxz
(yz) (xz)
g 1dxz b dxy
(xz) (xy)
adx2-y2 a dx2-y2
(x2 y2) (x2 y2)
bdxy ddz2

a 2b 2x n,d
gzz = 2.0023 8 gzz = 2.0023
(x2 y2)
a 2g 12x n,d d 2g 22x n,d
gxx = 2.0023 2 gxx = 2.0023 6
(xz) (yz)
a 2g 22x n,d d 2g 12x n,d
gyy = 2.0023 2 gyy = 2.0023 6
(yz) (xz)

(a)
d1 configuration, (b)
d1 configuration,
dxy ground state dz2 ground state

adx2-y2 ddz2

d dz2 bdxy (xy)

( z2)

g 2dyz (yz) a dx2-y2 (x2 y2)

g 1dxz g 2dyz (yz)

(xz)
g 1dxz (xz)
b dxy (xy)

a 2b 2x n,d
gzz = 2.0023 + 8 gzz = 2.0023
(xy)
a g 22x n,d
2 d 2g 22x n,d
gxx = 2.0023 + 2 gxx = 2.0023 + 6
(yz) (yz)
a 2g 22x n,d d 2g 12x n,d
gyy = 2.0023 + 2 gyy = 2.0023 + 6
(xz) (xz)
9 d9
(c)
d configuration, (d)
configuration,
dx2 - y2 ground state dz2 ground state

Figure 7 d orbital energy level diagrams and perturbation theory expressions for g-tensor components in selected configurations.22,23 Apart from
the singly occupied orbitals, the relative ordering of the orbitals in these diagrams is arbitrary. The systems of five levels qualitatively represent the
energies of the five molecular orbitals which are dominated by metal dxy, dxz, dyz, dz 2 , and dx 2 y 2 orbitals. The coefficients of the metal d atomic
orbitals in the corresponding molecular orbitals are denoted as a (for dx 2 y 2 ), b (for dxy), g1 (for dxz), g2 (for dyz), and d (for dz 2 ).

the minus sign in the expressions of Figure 7(a)). In reality,
Dgxx
Dgyy
20 ppt for VO(acac)211 and Dgxx
Dgyy
 4 to hence, a stronger bonding. In other words, a stronger bonding
19 ppt for the VO(salen) complexes.24 The parallel, gzz
value deviates from the free-electron g-value likewise toward
smaller values, again due to the minus sign in the expressions
of Figure 7(a), since the excitations are from a singly occupied
into an empty MO. Its deviation is larger than for the two
perpendicular components, Dgzz
 57 ppt for VO(acac)211 the most theoretically studied transition-metal radicals.2832
and Dgzz
 49 to 55 ppt for the VO(salen) complexes.24
This suggests that the excitation energies are smaller than for the
perpendicular components, which suggests that the p-bonding
to the axial oxo ligand is stronger than the sigma bonding to the
acetylacetonato ligand. Indeed, a larger excitation energy means
a higher energy of the antibonding MO, from which follows a
lower energy of the corresponding bonding counterpart and,
causes higher-energy splitting between the bonding and anti-
bonding levels.27 This very strong p-bonding is presumably one
of the reasons for the prevalence of the VO2 unit.11
An example of a d1 system with a single occupation of metal
dz2 orbital (Figure 7(b)) is the TiF3 complex, which is one of
As expected, the experimental gzz value is close to 2.0023
(Dgzz  3.7 ppt),33 since, within the simplified perturbation
theory approach, there is no excitation from dz2 connected with
a Dgzz (Figure 7(b)). The perpendicular components deviate,
however, significantly from the free-electron g-value
(Dgxx Dgyy  123.7 ppt),33 due to the energetically accessible
368 Electron Paramagnetic Resonance
p-antibonding orbitals involving dxz and dyz AOs, resulting as an
antibonding counterpart of a modest TiF p-bonding.
In many ways similar to the d1 configuration is the d9
configuration, which can also be treated as a positive hole
rather than the presence of nine d electrons. The expressions
for g-tensor components can be obtained directly from those
for the d1 configuration upon the replacement of xn,d by  xn,d;
thus, all g-tensor deviations are expected to be of opposite sign
compared to that in the d1 case. The d1x2 y2 configuration
treated in Figure 7(c) is realized in the Cu2 square-planar
complexes, in which the dz2 orbital is stabilized by the removal
of the two axial ligands from the octahedral coordination,
leaving the sigma-antibonding dx2 y2 orbital being the
SOMO.34 The gzz value, resulting from the dxy ! dx2 y2 excita-
tion, deviates strongly from 2.0023 toward more positive
values, resulting in a Dgzz value of 247 ppt for [Cu(NO3)2]
and as much as 264/285 ppt for Cu(acac)2.35
Finally, we introduce an example of a d9 configuration with
dz2 being the SOMO. This is realized in the [Co(CO)4] and [Ni
(CO)3H] pair of trigonally pyramidal complexes, where the
dx2 y2 antibonding orbital is stabilized with respect to an octa-
hedron by one missing equatorial ligand, whereas the dz2
orbital is kept destabilized by the antibonding interaction
with the axial ligand. Thus dz2 is the SOMO. In agreement
with Figure 7(d), the gzz value is close to the free-electron
one (Dgzz 1.9 ppt for the Co complex, Dgzz 3.6 ppt for the
Ni complex).36 The perpendicular values gxx gyy deviate from
the free-electron value by tenths of ppt: Dgxx Dgyy 65.1 ppt
for the Co complex and Dgxx Dgyy 127.6 ppt for the Ni
complex, in agreement with the significantly larger SO cou-
pling constant for Ni(I) compared to Co(0), cf. Table 9.1 in the
work of Mabbs and Collison.22
In rigorous quantum-chemical calculations of g-tensors, the
most conceptually simple approach is based on the so-called
one-component methods, where both the magnetic field and
the SO coupling are treated as perturbations. The original
Hamiltonian used for solving for the energies and MOs
excludes the SO coupling, but may include other relativistic
effects, called the scalar relativistic effects. A general introduc-
tion to relativistic effects can be found in Chapter 9.01 by
Shaik et al. The main perturbative contribution to the Dg
tensor arises from the SO coupling Hamiltonian and the
orbital-Zeeman Hamiltonian. The actual expression employed
for this contribution within the so-called uncoupled one-
electron approximation is37
"occa virta
a2 X X hca jH ^ SO, v jca ihca j^lO, u jca i
k a a k
DgSO=OZ, uv ge a  Ea
2 k a
E k a
X
occb virt
X b ^ SO, v jcb ihcb j^lO, u jcb i
hcbk jH
a a

k a Ebk  Eba
Note that eqns [15] and [16] are simplifications of eqn
[17]. The g-tensor components are expressed in terms of
matrix elements between occupied and virtual orbitals, which
are divided by corresponding energy differences. Instead of
experimentally determined SO coupling constants employed
in eqns [15] and [16], eqn [17] includes the SO effect as
an expectation value of the SO coupling Hamiltonian H ^ SO .
An apparent difference is the use of the unrestricted
approach - using the spin-orbitals instead of doubly occupied
MOs. Thus, in eqn [17], ca,b a,b
k and ca are unperturbed occu-
pied and virtual a/b spin-orbitals, respectively, ek and ea are
their energies, a is the fine-structure constant, and ge is the free-
electron g-value.
Besides DgSO/OZ, further contributions to Dg are given by
the so-called kinetic energy correction to the spin-Zeeman
interaction, DgRMC, and by the so-called diamagnetic gauge-
correction term, DgGC, whose expressions can be found
elsewhere.37 The total result for Dguv is, thus, the sum of the
contributions38
Dguv DgRMC, uv DgGC1e, uv DgGC2e, uv
DgSO=OZ1e, uv DgSO=OZ2e, uv [18]
where the DgGC and the DgSO/OZ terms have been divided into
so-called one-electron and two-electron contributions, in order
to be consistent with most of the quantum chemistry literature.
Within the one-component methods just described, one
starts by solving the one-particle equation (e.g., the Kohn
Sham equations, within the density functional theory, DFT)
with Hamiltonian not including the SO coupling. One solves
for the energies and MOs, and only then one incorporates the
SO coupling as a perturbation. In this approach, the MOs are
always a or b; thus, each MO has only one spin component,
from which comes the term one-component methods. How-
ever, one can adopt a different approach. The SO term can be
included in the original Hamiltonian. Then, the MOs obtained
by solving the Schroedinger equation have twice as many coef-
ficients, half of which correspond to the alpha spin and half to
the beta spin. Thus, the MOs are now of two-component form.
Finally, rather than deciding whether to include scalar rel-
ativistic terms only or in addition to the spin orbit effects and,
thus, to have a one-component or a two-component method,
one can start from the relativistic Dirac equation, which is the
most exact approach. This equation has been derived from the
Schroedinger equation by employing Einsteins relationship
between the total energy, the momentum, and the rest mass
of a particle. It turns out that the operators in the Dirac theory
are 4  4 matrices and the corresponding wavefunctions have
four components. Of the four solutions, only two correspond
to normal electronic behavior and represent the spin-up and
the spin-down states of the electron. The other two solutions
describe a positron. Within the four-component methods, all
relevant relativistic terms are included in the Hamiltonian
operator since the beginning, and no further perturbative treat-
ment is necessary. The four-component formalism has made
important advances over the past 20 years and, nowadays,
# [17] many standard quantum-chemical schemes, for example, CI
or CC methods, are available in a fully relativistic form.39
k
However, the formalism is computationally much more
demanding than standard nonrelativistic treatments,40 thus,
the more approximate two-component and one-component
methods have been developed simultaneously.
Any quantum-chemical calculation of a g-tensor produces
not only the principal components of the tensors, but also the
coordinates of axes (with respect to the working coordinate
system), in which the tensor adopts a diagonal form. Such a
system of axes is called the principal-axes system of the
g-tensor. Likewise, a prinicipal-axes system of the A-tensor

A tensor components,
g tensor components
Figure 8 The structure of the vanadium salicylaldmininato complex
and the orientations of the g-tensor, as well as the hyperfine coupling
tensor components. Atom labeling: orange - V, red - O, blue - N, grey - C,
white - H. The data used for the figure have been published by
Munzarova and Kaupp.24
can be determined. Interestingly, the A-tensor and g-tensor are
not necessarily coaxial (if not enforced by symmetry), and the
actual deviation of the principal-axes system depends on the
molecule of interest. An example of the relative g-tensor and
A-tensor orientation is shown in Figure 8.
9.13.4 Hyperfine Coupling from Quantum Chemistry:
Relativistic Contributions
In the previous section, we have introduced the concept of
relativistic contributions, especially that of the SO contri-
bution, to the g-tensor. Although our first treatment of the
hyperfine coupling in Section 9.13.2 was a nonrelativistic
one, relativistic effects (and especially the SO coupling) do
influence the isotropic and dipolar hyperfine coupling. In
particular, SO effects may manifest themselves in a second-
order pseudocontact contribution to Aiso (APC) and in a
second-order contribution to Adip (Adip,2).31 The isotropic
HFCC can then be written as
Aiso AFermi Apseudocontact [19]
and
Tkl Tkl, nonrelat: Tkl, 2 [20]
In a simplified perturbation-theory approach, the compo-
nents of the A-tensor can be calculated as22
 
Aij P kdij  3clij  2xn, l Lij 3cxn, l Lij [21]
where c 2/[(2l  1)(2l 3)] and is 2/21 for a d electron;
lij 1=2hf0 jli lj lj li jf0 i  1=3 ll 1dij ; dij 1 for i j but is
zero otherwise; Lij has been defined in eqn [16], and
i X X hf0 j^lr jfm ihfm j^lj^lt ^lt ^lj jf0 i
Lij  Eitr [22]
2 t, r m
Em  E0
P on the hyperfine coupling is included along the lines of pertur-
Within the summation t,rEitr, only those terms for which
i, t, r are combinations of x, y, z contribute. |E| 1, but E
changes sign each time x, y, or z have to be changed to produce
the sequence xyz, for example, Exyz 1 but Exzy  1. Combina-
tions such as itr xxy give Exzy 0. The parameter P is given by
the equation:
P 2:0023gN bN be hr3 i [23]
Electron Paramagnetic Resonance 369
where bN is the nuclear magneton and r3 is the expectation
value of r3 calculated for the wavefunction of interest. The
factor P depends, thus, on the radial wavefunction for the
metal ion. It characterizes the anisotropy of the spin density
around the magnetic nucleus. In simple calculations where the
radial wavefunctions are not used directly, tabulated values of
P for free ions are used.22 The dimensionless constant k is the
Fermi-contact parameter AFermi defined in eqn [13]. In simple
estimates, the value of k can be determined from experimental
isotropic HFCC by substracting the relativistic contributions, as
shown on p. 382 in the work of Mabbs and Collison.22
When the g-tensor of a system is known from experiments
(or from theory), a rough estimate of SO contributions to
HFCCs may be obtained. For example, for a d1 configuration
in a trigonally distorted octahedral field (dz2 being the SOMO),
we may use eqns 9.2049.209 in the work of Mabbs and
Collison22 to get

4 1
Ak AFC P d2  Dg? [24]
7 7

2 15
A? AFC P  d2 Dg? [25]
7 14
where d is the dz2 orbital coefficient in the SOMO, and
Dg? ge  g?. Now, setting (2/7)d2P Adip,22 we can express
the components of the hyperfine tensor with the Fermi-contact
term substracted:

Dg?
Ak Ak  AFC Adip 2  2
2d

15Dg?
A? A?  AFC Adip 1 [26]
4d2
 
Since A?  AdipAdip,2 APC and APC 1=3 Ak 2 A? ,
we get
7 Dg? Adip 17 Dg? Adip
APC Adip, 2  [27]
3 d2 12 d2
So, using the nonrelativistic results for Adip together with
the theoretical (or experimental) value for Dg?, rough (semi-
empirical) estimates of the pseudocontact (APC) and the
second-order dipolar (Adip,2) contributions can be made.
Important in eqn [26] is the relationship between the Dg?
value and the relativistic contributions to hyperfine parame-
ters. In the cases where it is known from experiments that the
deviations of the g-tensor from the free-electron g-value are
negligible, it can also be expected that nonrelativistic treatment
of the hyperfine parameters is sufficient.
In rigorous quantum-chemical calculations of relativistic
contributions to the hyperfine coupling, one can similarly to
g-tensor calculations employ the one-component methods, the
two-component methods, and the four-component methods.
With the one-component methods, the influence of relativity
bation theory. The equations actually employed are formally
very close to eqns [21], [16], and [22], and can be found
elsewhere.31 Instead of employing the tabulated free-ion values
for P, true expectation values of the so-called dipolar operator are
computed for the real system of interest.31 Similarly, the Fermi-
contact contribution k is evaluated from the wavefunction along
the lines of expression [13].
370 Electron Paramagnetic Resonance
With the two-component methods, the SO operator is
included in the molecular Hamiltonian and, thus, the SO
contributions to the hyperfine structure result naturally from
the wavefunction, without the need to introduce them pertur-
batively. Finally, four-component methods can be employed
for the relativistic calculations of HFCCs as well.
9.13.5 ZFS from Quantum Chemistry
To the first order in perturbation theory, the ZFS arises from
the direct electronelectron magnetic dipole spinspin cou-
pling (SSC) between unpaired electrons. Additional second-
order contributions arise from SOC of electronically excited
states with the ground state.10 The SSC contribution is most
frequently dominating the ZFS of organic molecules (due to
their generally small SOC constants), while the SOC contribu-
tion usually dominates in the case of transition-metal com-
plexes. The SOC contribution can be physically interpreted
as has been done by Neese in his review of the ZFS theory10:
The spinorbit coupling introduces some nonzero angular momen-
tum into the orbitally nondegenerate electronic ground state wave-
function. The associated orbital motion can be regarded as a current,
which according to the laws of electromagnetism gives rise to a
magnetic dipole moment perpendicular to the current. The total
net magnetic dipole moment of the current distributions then inter-
acts with the total spin angular momentum. The interaction energy
then depends on the relative orientation of the two magnetic dipole
moments. A zero observable effect arises if the orbital motion is
isotropic, since in this case the orientation of the spin angular
momentum is energetically equivalent in all directions. Conse-
quently, it is the anisotropic orbital motion that gives rise to the
observable SOC contribution to the ZFS.
Any quantum-chemical calculation of ZFS, thus, has to deal
with the dipole spinspin interaction, as well as with the SOC
contribution. This is challenging, since several spin states must
be included in the calculation, and the corresponding integrals
are rather complicated. The corresponding mathematical for-
malism is rather complicated and can be found elsewhere.10
Most of the quantum chemistry calculations of ZFS reported
so far have been implemented on the one-component level,
although two two-component41,42 and one four-component42
implementations have been reported as well.
9.13.6 DFT Calculations of the Hyperfine Coupling
Tensors
We are now at the point of reviewing the methodology of EPR
parameter calculations. A general introduction to computational
methods for molecules is given in Chapter 9.01 by Shaik et al.
Any meaningful calculation of the hyperfine coupling tensors
requires a proper description of the electron-correlation effects.
To achieve this in traditional post-HartreeFock ab initio calcu-
lations is far from being trivial, and the existing approaches are
hardly applicable to larger compounds, especially those involv-
ing transition metals. An alternative is provided by the DFT,
which includes electron correlation approximately, at moderate
computational cost.43 However, since within the DFT there is no
established hierarchy of methods with increasing accuracy of
electronic structure description, the performance of the individ-
ual density functionals must be validated with respect to the
property of interest and the type of system under study.
The first parameter entering any EPR parameter calculation
is the molecular geometry. It should be stressed that EPR
parameters turn out to be very sensitive to bond length and
angles, and geometry employed thus ultimately controls the
behavior of density functionals for HFCC calculations.44 Thus,
prior to any production EPR calculations, the molecular geom-
etry should be carefully validated against the available data
and, if necessary, against the EPR parameters themselves, as
has been recently demonstrated in the context of nuclear mag-
netic resonance (NMR).45 The theory of NMR parameter cal-
culations is reviewed in Chapter 9.14 by Truflandier.
The simplest groups of radicals from the view of the hyper-
fine coupling calculation are the hydrocarbon radicals. With a
few exceptions, all hyperfine couplings are reproduced to
within 10% of experiments using the exchangecorrelation
functionals B3LYP and PBE0, along with the basis set of EPR-
III quality.46 The EPR-III basis is a triple-x basis including
diffuse functions, double d-polarizations, and a single set of
f-polarization functions. The s-part is improved to better
describe the nuclear region: (6,2)/[4,2] for H and (11,7,2,1)/
[7,4,2,1] for B to F.47 Reproducing experimental hyperfine
couplings for main-group compounds involving N, O, and
especially F is more difficult, possibly due to the problem
of current exchangecorrelation functionals to describe the
spin-density distribution on atoms bearing lone pairs.48 For
example, for the radical H2O, the spin density at the oxygen
nucleus is underestimated by 20% when using the B3LYP/
EPR-III method. At the same time, based on an extensive
examination of 75 main-group organic and inorganic radical
hyperfine couplings, the B3LYP/EPR-III approach seems to
offer the most accurate theoretical predictions of HFCCs out
of all methods and basis sets tested.49 A particular challenge is
represented by HFCC calculations for atoms in the case of
phosphorus, none of the standard density functionals repro-
duces even the correct sign of the hyperfine coupling.50
The most prominent group of EPR active species from the
inorganic chemists point of view is the transition-metal com-
plexes. As has been shown in a critical validation study of DFT
approaches,17 the ability of current exchangecorrelation func-
tionals to compute accurately transition-metal HFCCs strongly
depends on the type of the SOMO(s). The easiest systems to
describe theoretically are those with the SOMO being of mainly
s-type and, thus, bringing large direct contribution to the spin
density at the nucleus. Such an example is provided by the ScO
molecule discussed in Section 9.13.2, and an isoelectronic sys-
tem TiN. Eight density functionals have been explored: three
generalized-gradient approximation (GGA) functionals BLYP,
BP86, and BPW91; two Beckes three-parameter hybrid
functionals including ca. 20% HartreeFock (HF) exchange,
namely, B3LYP and B3PW91; and three hybrid functionals
including as much as 50% HF exchange: BHLYP, BHP86, and
BHPW91.43 Figure 9 displays, for these two systems, the depen-
dence of the spin density at the metal nucleus on the particular
exchangecorrelation functional. Although it is evident that the
functionals BP86, BPW91, B3PW91, and BHLYP are more suit-
able than the other functionals within the set tested, it is

TiN calc.
2.25
2.00
N/2S (in a.u.)
1.75
1.50
1.25
1.00
TiO calc.
0.75
BLYP
BP86
Figure 9 Spin density rab
N
at the metal nuclei, normalized to the number of unpaired electrons. The data used for the figure have been published
by Munzarova and Kaupp.17
important to realize that all computed values lie within 5% of
the very precise, gas-phase experimental value.
The molecules TiO and VN contain also, in addition to a 3ds-
type SOMO, one 3dd-type SOMO. While for TiO the functional
dependence of the HFCCs is trend-wise similar and magnitude-
wise smaller than for ScO and TiN, the situation is different for
VN. Clearly, the VN results are deteriorated when functionals
with 50% of HF exchange are employed. The reason for this is
due to a phenomenon called the spin contamination. In order to
describe the spin polarization within DFT, the unrestricted
approach must be employed. If a restricted approach was used,
then the wavefunction would be an eigenfunction of the spin
operator S^ : For the unrestricted approach, the wavefunction
2
is no longer an eigenfunction of the spin operator S^ . Still, the
2
expectation value of the spin operator S^ can be computed. This
2
should be close to S(S 1)2, that is, to 1/2  (1/2 1)2 for a
doublet, to 1  (1 1)2 for a triplet, etc. If the deviation of S^
from S(S 1) is substantial, then the wavefunction is said to
2
be significantly spin-contaminated, and such description of an
electronic state should be viewed with suspicion.
The density functionals employing larger percentage of the
HF exchange can (but do not have to) cause a considerable spin
contamination because the unrestricted HartreeFock method
itself tends to cause a spin contamination.5 Whether or not the
wavefunction will be significantly spin-contaminated turns out
to depend on the particular bonding situation, which is
described in more detail below.
After considering the case with dominant direct SOMO
contribution to HFCCs, let us concentrate on the cases where
the direct SOMO contribution is zero and spin polarization
determines the hyperfine coupling. Figure 10 illustrates the
density functional behavior for the isoelectronic complexes
[Cr(CO)4] and [Mn(CN)4]2. Obviously, the best results are
obtained for the functionals containing 50% of HF exchange,
which best describe the spin polarization, while the pure GGA
Electron Paramagnetic Resonance 371
TiN exp.
ScO calc.
ScO exp.
VN calc.
VN exp.
TiO exp.
BPW91 B3PW9 BHP86
B3LYP BHLYP BHPW91
Functional
functionals (BLYP, BP86, BPW91) reproduce only ca. 50% of
the experimental value. At the same time, the spin contamina-
tion is negligible, due to a modest extent of the valence-shell
spin polarization.
Finally, let us concentrate on cases where both the direct and
the spin-polarization contributions are sizable. Figure 11 shows
the functional behavior for the trigonally planar isoelectronic
complexes TiF3 and MnO3. The SOMO is a mixture of metal 3dz2
orbitals perpendicular to the molecular plane, of metal 4s
orbitals, and of ligand orbitals interacting in an antibonding
fashion (cf. Figure 4). The relative contributions to the SOMO
are similar for both complexes (Mulliken population analysis,
using the BLYP functional, gives a 4s/3dz2 population ratio of
0.18/0.76 for TiF3 and of 0.13/0.67 for MnO3).17 For both
complexes, the best agreement with experiments is obtained
with the B3LYP and B3PW91 functionals. The spin densities
2
with the pure GGA functionals (BLYP, BP86, BPW91) are too
large, due to insufficient spin polarization. The spin densities with
the BHLYP and the BHPW91 functionals are too small for both
cases, but much more so for MnO3 than for TiF3. While for TiF3
the spin state is described correctly with any of the functionals
(S^ 0:752  0:753 compared to nominal S^ 0:750), for
2 2
MnO3, the state is heavily spin-contaminated for the BHLYP,
BHP86, and BHPW91 functionals S^ 1:994  2:054. The
2
strong spin contamination in this case (as well as in other cases)
correlates with the strong spin polarization of the valence shell
and, in turn, with a large overlap between the SOMO and a certain
lower-lying, doubly occupied MO (cf. Figure 4(b)). Thus, in
the case of MnO3, the functionals with 50% HF exchange so
successful for [Cr(CO)4] and [Mn(CN)4]2complexes are
clearly unsuitable for the HFCCs description.
Apart from the critical validation study of 21 3d metal
complexes discussed above,17 two further systematic DFT
studies on transition-metal systems exist. Micera and Garribba
studied 22 representative VIVO complexes with several density
372 Electron Paramagnetic Resonance

0.08 [Mn(CN)4]2 calc.

0.10
[Cr(CO)4]+ calc.
N-/2S (in a.u.)
0.12

0.14

0.16 [Cr(CO)4]+ exp.

0.18 [Mn(CN)4]2 exp.

BLYP BPW91 B3PW91 BHP86

BP86 B3LYP BHLYP BHPW91
Functional
Figure 10 Spin density rabN
at the metal nuclei, normalized to the number of unpaired electrons, and the five SOMOs. Ligand contributions are omitted
for simplicity. The data used for the figure have been published by Munzarova and Kaupp.17

2.00
MnO3 calc.
1.80

1.60
MnO3 exp.
rN-/2S (in a.u.)

1.40

1.20

TiF3 calc.
1.00

0.80 TiF3 exp.

0.60

0.40
BLYP BP86 BPW91 B3LYP B3PW91 BHLYP BHP86 BHPW91
Functional
Figure 11 Spin density rab
N
at the metal nuclei, normalized to the number of unpaired electrons. The data used for the figure have been published by
Munzarova and Kaupp.17

functionals and found excellent agreement between experi-
ments and the half-and-half hybrid functionals containing
50% HF exchange.51 Kossmann et al. studied a series of small
radicals and transition-metal complexes employing also,
besides the established functionals, the TPSS meta-GGA func-
tional (together with its hybrid version, TPSSh) and the
B2PLYP double-hybrid functional.52 The results reveal that
TPSS is superior to BP86; the hybrid variant TPSSh is at least
as accurate as or better than the B3LYP functional, and the
B2PLYP functional also leads to accurate predictions and is a
clear improvement for the difficult metal nucleus HFCCs.
In terms of the basis set, for the main-group systems, it
should be first stated that the conventional double-zeta
polarization (DZP) bases (e.g., the 6-31G* basis frequently
employed for geometry optimizations) are not sufficiently
accurate to generate satisfactory HFCCs.46 However, they can
give reliable results if they are uncontracted in the outer core
inner valence region. An example is the 5s2p1d basis set (3s1p
for hydrogen) of Barone called EPR-II.53 The conventional
triple-zeta polarization (TZP) basis (e.g., the 6-311G* basis)
gives reasonable but slightly too large isotropic HFCCs. Of the
partially uncontracted TZP basis sets, the 7s6p2d (4s2p for
hydrogen) set known as individual gauge for localized
orbitals-III (IGLO-III) is considered to be highly suitable for
molecular HFCC calculations.54 The results of IGLO-III quality
are also obtained using the van Duijneveldt series of basis

sets.55 Finally, the 7s4p2d basis (6s2p for hydrogen), known as
EPR-III, has been purposely tailored by Barone for reliable
magnetic properties calculations.53 A set of 75 main-group
radicals has been systematically studied by Hermosilla et al.
using several DFT methods and basis sets.49 It was concluded
that DFT predictions of HFCCs are reliable for B3LYP/TZVP
and B3LYP/EPR-III combinations.
Basis set studies for transition-metal complexes are much
less abundant. For several 3d-metal systems, a large uncon-
tracted 21s15p10d3f basis has been compared to a medium-
sized basis in a (15s11p6d)/[9s7p4d] contraction.17 Due to
systematic error compensation, the 9s7p4d basis simulates
the 21s15p10d3f basis quite well and has been employed in
a number of application studies. Another systematic basis set
study has been performed for 4d-metal molybdenum com-
plexes, where it turns out that a 12s6p5d basis is a good
compromise between cost and accuracy.56 It is self-evident
that, for meaningful HFCC calculations, all electron basis sets
have to be employed and the use of the pseudopotentials
must be avoided, otherwise, the important coreshell spin-
polarization process cannot be described.
So far, we have discussed the nonrelativistic DFT values of
HFCCs and compared these to experiments. It is now natural to
ask what is the expected size of the relativistic contributions to
the hyperfine coupling, that is, in which cases is it important
to take care of the relativistic effects and where can these be
omitted. An illustrative comparison is given in Table 1 in the
work of Autschbach et al.,57 where it is shown that, within
simple organic and inorganic main-group systems, the paramag-
netic spinorbit/spinorbit (PSO/SO) cross-term38 is always
smaller than 1 MHz in the case of the C, H, and O atoms. For
Si, it can reach 47 MHz and approximately 1020% of the total
hyperfine coupling (for SiOH and SiSH, respectively).
For transition-metal complexes, likewise, the PSO/SO term
can be of the order of a few megahertz. For example, for Ti in
TiF3, the PSO/SO term is 2.1 MHz out of a total 233.7 MHz
of isotropic hyperfine coupling.31 However, for other transition-
metal systems (those with a significantly deviating g-tensor
from the free-electron g-value), SO relativistic contributions
450
500
[Cu(KTS)]
550
[Cu(NH3)4]2+
AII (calc.) in MHz
600
650 [Cu(en)2]2+
700
750
800
850
620 600 580
Figure 12 The importance of second-order SO corrections to hyperfine coupling tensors. Parallel component, functional B3PW91. Atomic mean field
(AMFI) approximation, 9s7p4d Cu basis, IGLO-II basis for ligands. The data used for the figure have been published by Remenyi et al.58
Electron Paramagnetic Resonance 373
of the size of tenths of megahertz can be obtained. For
example, a pseudocontact contribution to isotropic HFCC of
MoV complexes is 1418 MHz, depending on the density
functional.56 Figures 12 and 13 illustrate, for the parallel and
perpendicular components of the hyperfine tensor, the impor-
tance of including the second-order relativistic terms for Cu
complexes, which are in the order of a hundred megahertz. Still
larger contributions are obtained for mercury-containing com-
pounds: 190 MHz out of a total 7070 MHz (i.e., 3%) for Hg
in HgH, and as much as 145 MHz compared to the nonrel-
ativistic value of 249 MHz for F in HgF.57 Finally, in NpF6, the
PSO/SO term is equal to as much as 1631 MHz out of
the total 2111 MHz and dominates, thus, by far, the hyper-
fine coupling.59
Apart from the PSO/SO terms, scalar relativistic contribu-
tions can be important too: for main-group systems and for
light nuclei in the neighborhood of heavy nuclei, these typi-
cally represent, at most, a few megahertz.60 For transition-
metal systems, the scalar relativistic contributions typically
correspond to 1520% of isotropic hyperfine coupling, as has
been shown in a careful study of Mo complexes by Fritscher
et al.56 Finally, extreme values of scalar relativistic contribu-
tions have been reported for Cd (hundreds of megahertz) and
Hg (thousands of megahertz) compounds.61
Currently, the most widely employed A-tensor implemen-
tation is the one within the Gaussian package, going from the
current Gaussian09 version62 back to Gaussian94. The Gauss-
ian implementation offers a variety of density functionals, but
it does not include relativistic effects on hyperfine couplings.
From the relativistic implementations, of very frequent use is
the one within the Amsterdam Density Functional (ADF)
program,63 where the hyperfine tensors have been recently
implemented within the two-component method.57 Another
relativistic implementation is within the MAG-ReSpect code,
including one-component, two-component,64 and four-
component65 methods. The relativistic effects on HFCCs are
also included in the ORCA software,66 developed and imple-
mented by Neese.67 Very recently, a new module for DFT
calculations of hyperfine couplings has been implemented in
Ideal agreement
[Cu(dtc)2]
[Cu(iz)4]2+ [Cu(mnt)2]2
[Cu(sac)2]
[Cu(gly)2]
[Cu(acac)2]
Second-order results (+ spin-orbit)
Nonrelativistic first-order results
560 540 520 500 480 460
AII(exp.) in MHz
374 Electron Paramagnetic Resonance
20
40
60
A (calc.) in MHz
80
[Cu(mnt)2]2 [Cu(KTS)]
100
120
140
[Cu(dtc)2]
160
180
200
120
Figure 13 The importance of second-order SO corrections to hyperfine coupling tensors. Perpendicular component, functional B3PW91. AMFI
approximation, 9s7p4d Cu basis, IGLO-II basis for ligands. Data used for the figure have been published by Remenyi et al.58
the open-source NWChem quantum chemistry package.59 This
implementation includes both scalar relativistic and spinorbit
effects, and allows a chemically intuitive analysis in terms of
localized MO contributions. Finally, let us note that DFT cal-
culations of HFCCs have been implemented also for solid-state
systems under periodic boundary conditions within the BAND
module of the ADF program.68
9.13.7 Ab initio Post-HartreeFock Calculations
of HFCCs
Nowadays, an overwhelming majority of calculations of HFCCs
is being done using the DFT. The reason for this is that current
functionals provide astonishingly good isotropic HFCCs for a
large variety of systems, while approaches based on the wave-
function, such as multireference configuration interaction
(MRCI, see pages 576577 in the work of Levine5) or coupled
cluster (CC) approaches, show a slow convergence with respect
to the level of sophistication.69 Nevertheless, there are, in gen-
eral, three reasons for why ab initio post-HartreeFock
approaches are important for hyperfine couplings. First, in
many cases (encompassing especially the transition-metal sys-
tems), the computed HFCCs depend strongly on the functional
employed, and reference data are needed in order to judge the
suitability of the given density functional for the system of
interest. Second, complex bonding situations exist for which
all the current density functionals largely fail. Third, even in
the cases where DFT is successful in computing the HFCCs,
this may partially result from error cancellations, which are
good to analyze and understand for the sake of developing
better density functionals for the difficult cases.
Successful ab initio post-HartreeFock calculations of
HFCCs are based either on the multireference configuration
interaction (MR-CI) or on the CC methods. A validation study
of the CC approaches has been done for HFCCs of eight
transition-metal complexes by Munzarova and Kaupp.17 The
[Cu(sac)2]
[Cu(NH3)4]2+ [Cu(gly)2]
[Cu(acac)2]
[Cu(en)2]2+
Second-order results (+ spin-orbit)
Nonrelativistic first-order results
100 80 60 40
A (exp.) in MHz
results for the largest CCSD(T)/15s10p6d2f calculations for the
isotropic metal-coupling constants have been found to be only
ca. 5% below the experimental value, indicating a slight over-
estimation of the spin polarization. The use of a smaller 9s7p4d
basis set resulted in a further reduction in isotropic HFCCs.
However, the results should still be considered as being very
successful considering, for example, the only 6% underestima-
tion of metal isotropic coupling for MnO3, a case which is very
difficult for density functional approaches (see above).
Still higher accuracy has been achieved with CCSD(T) cal-
culations of HFCCs for small main-group radicals H2CN,
H2CP, NH2, and PH2.70 With a few exceptions, the anisotropic
hyperfine couplings deviate from experiments in the order of
35%. Somewhat surprisingly, this is a poorer accuracy than
that obtained for the isotropic hyperfine couplings, namely, as
little as 12% of experiments.
Whereas the CCSD method is generally successful for com-
puting HFCCs, the configuration interaction singlesdoubles
(CISD) method fails in nearly all cases.69 The reason for this is
that higher than double excitations, even though not directly
contributing to isotropic HFCCs, are important in that they
influence the coefficients of the single excitations in the expan-
sion. However, including higher than double excitations is com-
putationally demanding. This problem has been elegantly solved
by Engels within the MRD-CI/BK approach, which accounts
for the indirect influence of the higher excitations neglected in
the wavefunction.71 Highly accurate HFCCs for main-group and
transition-metal systems have also been obtained by the qua-
dratic configuration interaction theory with single and double
excitations.72 Excellent results have also been obtained for the
metal dipolar HFCC of a case copper complex by the recently
developed spectroscopy-oriented configuration interaction
(SORCI) method.73 Accurate prediction of the metal isotropic
HFCC is challenging, but fairly accurate values can be obtained
when appropriate care is used in the construction of the one-
electron and many-electron bases. In this case, one obtains the
necessary core-level spin polarization, but without any of the

spin contamination which plagues spin-unrestricted HF or DFT
treatments. For this reason, the SORCI method is superior to the
DFT method for metal isotropic HFCCs.73
To the best of our knowledge, the available ab initio
A-tensor implementations are all established within the non-
relativistic framework. The CCSD spin density distributions
necessary for HFCC calculations can be obtained from Gauss-
ian 03 or Gaussian 09.62 A special program directed at CC
approaches is the CFOUR package by Stanton, Gauss et al.,74
which is a follow-up of the older ACES2 package. The SORCI
method has been implemented within the ORCA package.66
Much work has been done recently in the field of solvent
effects on the hyperfine coupling. Their treatment is beyond
the scope of the current work; however, an excellent account of
these can be found in the work of Ciofini.75
9.13.8 DFT Calculations of g-Tensors
A theoretician-oriented review on the DFT approaches for
g-tensor calculations has been published in 2004 by Patchovskii
and Schreckenbach.30 Here, we try to give a practically oriented
guide to density functional g-tensor calculations, focusing
especially on the literature published after 2004. As explained
above, the electronic g-tensor is dominated by the SOC and,
thus, should be viewed as a fundamentally relativistic property.
Computational methods can be divided between those that
treat SOC by perturbation theory (one-component methods)
and those that include SOC already in the zeroth-order wave-
function (two- and four-component methods).
A comparison for d1 transition metal complexes
1-comp BP86
300
2-comp BP86
4-comp BP86
200
100
g|| calc. [ppt]
0
100
200
300
400
400 300 200 100 0 100
g|| exp. [ppt]
Figure 14 Comparison of computed g-shift tensor components (BP86 results in ppt) by one-, two-, and four-component methods with experiments
for the [MEX4]q series. The data are taken from the work of Hrobarik et al.76
Electron Paramagnetic Resonance 375
An illuminating comparison of the one-component, two-
component, and four-component methods for a set of 3d1, 4d1,
and 5d1 transition-metal complexes using identical functionals
and basis sets has been done by Hrobarik et al.76 The advantage
of the multicomponent relativistic approaches is the full inclu-
sion of higher-order SOC to all orders. For all complexes stud-
ied by Hrobarik et al.,76 sizable higher-order SOC effects
beyond the leading order have been found, even for 3d com-
plexes. In particular, for the 4d and 5d complexes, second-order
SOC contributions to the g-tensor tend to be of the same order
of magnitude as the first-order terms. Thus, the one-component
calculations with leading-order perturbation theory give dra-
matic errors for such systems, as is shown in Figure 14. It turns
out that, at least for most of the complexes studied, the higher-
order SO effects are significantly larger than the dependence of
the results on the exchangecorrelation functional.
From what has been just said, it is clear that the validation
and selection of the best functionals for g-tensor calculations
depend on the treatment (or omission) of the higher-order SO
contributions. While for calculations including higher-order
SO contributions g-tensors are reproduced best by hybrid den-
sity functionals with about 3040% exact-exchange admixture
for MoV systems,77 the ideal amount of HF exchange for the
same type of systems is only 1020% if higher-order SO con-
tributions are omitted, while a larger amount of HF exchange
deteriorates the results.44 However, for different bonding situ-
ations, appreciable exact-exchange admixture may turn out
optimal, even for calculations omitting higher-order SO effects,
as has been shown for the case of an eight-coordinated vana-
dium (IV) complex.78 The need to use system-dependent
M
X X
X X
100 [MEX4]q
1-comp BP86
50 2-comp BP86
4-comp BP86
0
50
g calc. [ppt]
100
150
200
250
300
350
200 350 300 250 200 150 100 50 0 50
g exp. [ppt]
376 Electron Paramagnetic Resonance
amounts of HF exchange is unsatisfactory from the theory
point of view. An alternative has been suggested by means of
the so-called local hybrid functionals, which incorporate
position-dependent but system-independent amounts of HF
exchange included in the exchangecorrelation functional. So
far, only little to moderate improvement has been found for
thermochemistry-optimized local hybrids compared to global
hybrids such as B3LYP. Closer analyses point to possible areas
in which the fundamentally more flexible local hybrids may be
improved for the property at hand.79
Regarding basis-set dependence of the g-tensors within DFT,
it has been shown that medium-sized double-zeta valence-
polarized (DZVP) basis sets can be sufficient if they are specif-
ically tailored to the system of interest.44 However, the majority
of studies employs fully uncontracted triple-z quality Slater80 or
Gaussian-type81 basis sets. For g-tensor calculation, the use of
effective core potentials can be of useful accuracy at very limited
computational cost.56
Unlike for the HFCC calculations, with g-tensor calculations,
a fundamentally new practical problem arises due to the depen-
dence of magnetic operators on the choice of the origin of the
coordinate system, which is known as the gauge-origin
problem.82 Common approaches to mitigation of the gauge
dependence are the gauge-including atomic orbitals (GIAO)
and the IGLO methods. Both of these approaches give very
similar results for the g-tensors.30
Coming to implementations of electronic g-tensors within
DFT, starting with the one-component approaches, let us first
mention the ORCA code developed by Neese.66 Due to the
user-friendly style, ORCA is considered to be a helpful tool not
only for computational chemists, but also for chemists, phys-
icists, and biologists that are interested in developing the full
information content of their experimental data with the help of
calculations. Another one-component implementation is
within the Gaussian09 package,62 where, for open-shell spe-
cies, the NMR keyword provides both the NMR chemical shift
tensor and the EPR g-tensor.
Among the two-component methods, a widely used pro-
gram is the ADF package, including the program BAND for
periodic structures.63 This implementation uses Slater-type
orbital basis sets, employs both nonhybrid and hybrid func-
tionals, and is capable of including the spin polarization via
the spin-unrestricted approach. Both two-component and one-
component calculations can be done. Another two-component
implementation with the inclusion of spin polarization, enabling
one-component calculations as well, is the MAG-ReSpect code
developed by Malkin and coworkers.83 Within the latter, also, the
local hybrid functionals have been implemented.79 Very recently,
the first four-component DFT implementation of EPR g-tensors
appeared within the ReSpect code,84 which enables also both
two-component and one-component calculations, with the
inclusion of spin polarization.85
9.13.9 Ab initio Post-HartreeFock Calculations
of g-tensors
Ab initio calculations of g-tensors are less abundant in the
literature than the DFT calculations. In recent years, schemes
have been developed for the calculation of g-tensors using the
second-order multireference perturbation theory (MRPT2),
the MRCI,86 the multiconfigurational SCF (MCSCF), the
difference-dedicated configuration interaction (DDCI),87 and
the CC theory.5 While results in agreement with the experi-
mental data are reported for the MCSCF method,88 its relative
performance to the other ab initio methods cannot be judged
due to the absence of comparative studies. Out of the available
comparisons, the best results are provided by the MRCISD Q
method,89 the DDCI3 Q method,89 and the CC method.90
Slightly inferior are the SORCI method and the multireference
second-order perturbation theory (MRMP2).89
A special case of the MRMP2 method is the complete active
space perturbation theory to second-order (CASPT2), see page
585 in the work of Levine.5 The CASPT2 method is attractive
for applications with large transition-metal complexes, since it
is computationally less demanding than the configuration
interaction and CC methods. It is, thus, the most important
ab initio competitor of DFT methods. The results obtained for
a series of d1 MOX4ncomplexes (M V, Cr, Mo, Tc, W, Re;
X F, Cl, Br) indicate that the CASPT2 approach is superior to
the results from previous DFT calculations,28 as is the MRCI
approach.86 However, for some d9 copper complexes, the com-
plete active space self-consistent field (CASSCF) wavefunction
provides a too ionic description of the CudL bonds, which
leads to the overestimation of the angular momentum and
SO matrix elements. This, in connection with the accurate
CASPT2 excitation energies, leads to an overestimation of
the Dg values. Possible solutions for the inadequacy of the
CASSCF wavefunction are currently under investigation.28
Coming to implementations of ab initio g-tensor calcula-
tions, the CC methods have been implemented within the
CFOUR package.74 The CASSCF/CASPT2 and the CCSDT
methods have been implemented in the MOLCAS package.28,91
Within the ORCA package,66 the following ab initio methods
have been implemented and evaluated for g-tensors: RO-CIS,
MRMP2, SORCI, DDCI3, and MRCISD.89 The Gaussian62 imple-
mentation of the g-tensors supports, besides DFT, only HF and
MP2 methods, which cannot be recommended for accurate
calculations.
Considerable work has been recently done in the field of
solvent effects on the g-tensors. Their treatment is beyond the
scope of the current work; however, an excellent account of
these can be found in the work of Ciofini.75
9.13.10 DFT Calculations of ZFS
As in the case of the g-tensor, it is equally important for
ZFS calculations to adopt not only the most appropriate non-
relativistic ground state electron density, but also to have a
proper account of the relativistic effects. Indeed, in the case of
transition-metal systems, the SOC contribution to the ZFS
usually dominates over the SSC one10 and must, thus, be
correctly reproduced. Very recently, four-component and two-
component MRCI and CC relativistic methods have been
compared to one-component CASPT2 and one-component
DDCI methods.42 For four low-lying states of chalcogen
diatomics, the four-component and two-component methods
provide excellent agreement with experiments, while the
one-component methods have an accuracy of only about

80%. Another two-component study has been reported for a
series of density functional approaches, where, however, the
ZFS of chalcogen diatomics has been found to be systemati-
cally underestimated by a factor of ca. 2.41
Most of the current ZFS calculations are performed on the
one-component level though. In the DFT framework, perturba-
tion theoretical treatment of the SOC term has been suggested by
Pederson and Khanna (PK),92 Reviakine et al.,41 Takeda et al.,93
Aquino and Rodriguez,94 as well as in Neese95 and Schoneboom
et al.96 While Pederson and Khanna have reported results of
apparent high accuracy on large spin-coupled single-molecule
magnets, Reviakine et al. criticized that a calibration study on
mononuclear complexes has neither been reported by these
authors nor by Aquino and Rodriguez. The works by Neese9698
as well by Reviakine et al.41 showed that the PK method has
rather large errors and underestimates the D value by a factor
of 24 upon application to most transition-metal monomers.
However, it has been found that this does not hold for MnII
complexes, where the (small) ZFS is overestimated by the PK
approach, together with the spinspin term.
The problem of how to best calculate the ZFS tensor has
been reconsidered by Neese, who proposed an analytic deriv-
ative theory of the ZFS, implemented it, and evaluated it
through a series of benchmark calculations on diatomic mol-
ecules (coupled-perturbed, CP approach).99,100 Both the PK
method and the CP approach have been analyzed by Schmitt,
Jost, and van Wullen.101 The authors have shown that the PK
method can be improved by properly taking into account the
quantum nature of spin when extracting the ZFS tensor from
the magnetic anisotropy. They further showed that a closed-
shell molecule at a large distance from an open-shell complex
will have a spurious contribution to the ZFS tensor calculated
with the CP method, which can be improved if one correctly
transforms the equations used in wavefunction-based theory to
a sum-over-states type expression before one interprets it as an
energy derivative. van Wullen et al. have shown that, if
improved along these lines, the PK and CP approaches become
identical down to the working equations.
Let us now concentrate on the relation of spin-restricted
and spin-unrestricted one-component DFT calculations of ZFS.
A case study on the Mn(acac)3 complex revealed that spin-
restricted calculations give larger absolute D values, which are
closer to experiments than the spin-unrestricted values.102 So,
while the spin-unrestricted approach is mandatory for exact
calculations of the hyperfine couplings, it may be less appro-
priate for the calculations of the ZFS, probably due to the
problem of spin contamination. The difference between the
spin-restricted and spin-unrestricted results is especially pro-
nounced for the PBE0 functional containing exact exchange,
where, also, the spin-restricted result for D is the best one (92%
of the experimental value).102
Proceeding with the evaluation of the individual density
functionals, a detailed study of the one-component DFT
approaches has been performed for a series of MnIII complexes
by Duboc et al.100 Among the tested functionals, the best pre-
dictions have been obtained with the B3LYP functional, fol-
lowed by the nonhybrid BP86 functional, which, in turn, was
found to be more successful than the meta-hybrid GGA func-
tional TPSSh. A similar hierarchy has been observed for the
series of small main-group radicals, where the theoretical vs.
Electron Paramagnetic Resonance 377
experimental slope of 0.95 was found for B3LYP and the slope
of 0.85 was found for BP86.99 A possible reason for the supe-
rior behavior of the hybrid functionals in these cases is the
better account of the important spin polarization. However,
the hybrid and nonhybrid functionals can, in some cases,
perform very similarly, as has been observed for Mn4 and
Fe3 ions in the layered LiAlO2 oxide,103 or the nonhybrid
functionals can even prove to be superior to the hybrid func-
tionals, as for the hexaqua-chromium(II) complex, where the
BP86 functional (with the CP approach) reproduced 87% of
the experimental value and the B3LYP approach reproduced
only 73% of the experimental value.104 A basis set study for
various GGA functionals has been performed by Zein et al.,
which showed that standard double- (SVP) or triple-z (TZVP)
basis sets are saturated with respect to ZFS results.102
The reason why the hybrid functionals perform better than
the pure ones in some cases, while in others they perform worse,
is similar to in the context of the hyperfine coupling. It is, again,
the spin-polarization vs. spin-contamination dilemma. For sys-
tems which are more susceptible to spin contamination, the use
of hybrid functionals leads to largely spin-contaminated wave-
function and to the overestimation of the D parameter. For
systems which are less susceptible to spin contamination and
spin polarization, the results for D are similar for hybrid
and nonhybrid functionals. This has been nicely analyzed for
high-spin nitrenes and carbenes by Misochko et al.105 Finally,
for systems less susceptible to spin contamination with pro-
nounced spin-polarization contribution to the ZFS, the hybrid
functionals prove superior to the nonhybrid ones.99
An interesting DFT study of ZFS has been presented recently
by Cirera et al.,106 who focused on the relationships between
the molecular geometry and the size of the parameter D. In
particular, the distortions from the octahedral to the trigonal
prismatic geometry (the Bailar twist), from the square pyrami-
dal to the trigonal bipyramidal geometry (the Berry pseudor-
otation), and the planarization of a tetrahedron have been
explored. It has been concluded that the Bailar twist of d5
hexacoordinate complexes with trigonal prismatic geometry
results in relatively high negative D values; the analysis of the
variation of D along the Berry pseudorotation path for d5
pentacoordinate complexes indicates the presence of positive
D values for the tetragonal pyramid, whereas the trigonal bipy-
ramidal structure gives D values close to zero. Finally, the
planarization of tetrahedral structures toward square-planar
coordination in d5 tetranuclear complexes yields a maximum
negative anisotropy in the middle of the interconversion path-
way. For the relation between the spin state and stereochemis-
try, see Chapter 9.17 by Cirera.
9.13.11 Ab initio Calculations of ZFS
Although, in some cases, DFT calculations of ZFS appear suc-
cessful, in general, the DFT calculations of ZFS turn out to be
inferior to the wavefunction-based ab initio methods. This has
been shown for the series of MnIII complexes by Duboc et al.
when comparing several DFT functionals and the ab initio
state-averaged CASSCF (SA-CASSCF) method with a minimum
active space.100 In particular, the TPSSh meta-GGA functional
provided disappointing results, indicating that this functional
378 Electron Paramagnetic Resonance
has a problem with magnetic response properties. Similar
results have been obtained for a series of NiII complexes by
Kowalewski, who compared several density functionals with
the CASSCF and the second-order n-electron valence state per-
turbation theory (NEVPT2) methods.107 Liakos et al. report the
study of the hexaqua-chromium(II) complex, where the SORCI
and the CASSCF methods provided 107% of the small exper-
imental value for D.104At the same time, the DFT result (with
the CP approach) was only 7387% of experimental result,
depending on the functional.104 The underlying reason for the
inferiority of DFT methods to the wavefunction-based
methods is that spin-unrestricted single-determinant methods
do not accurately reproduce the two-electron-reduced electron
density required for the evaluation of two-electron spin-
dependent properties.108
Most of the one-component ab initio calculations of ZFS
are currently being done using a two-step procedure to include,
firstly, dynamical electron correlation effects and, secondly,
spinorbit effects on the relative energies of the low-lying
electronic states. It is, hence, based on the idea that electron
correlation and SO effects are largely decoupled. Two quite
close implementations of this approach are available: within
the MOLCAS package109 based on the development of
Malmquist et al.110 and within the ORCA code.66 Both imple-
mentations start with a CASSCF calculation to treat nondynami-
cal correlation. This is followed by the introduction of dynamic
correlation effects through the evaluation of the single- and
double-excitation contributions in a second-order perturbative
manner (CASPT2) in MOLCAS, whereas the NEVPT2 approach
was used in ORCA. The SOC is then taken into account by
interaction of the CASSCF states using the CASPT2 or NEVPT2
relative energies. For the cases where comparison can be made,
the CASSCF/CASPT2 results from MOLCAS and the CASSCF/
NEVPT32 results from ORCA appear to be comparable.111
The applications of the MOLCAS approach include the 2006
study of ZFS in complexes of Mn, Fe, and Co by De Graaf and
Sousa,40 whose results compared favorably with experimental
data and fully relativistic benchmarks. Maurice et al. studied the
set of three NiII and one CoII complexes and obtained good
results especially when an enlarged active space including ligand
orbitals was employed: three out of four theoretical D values
were found within 100105% of experiments.112 Maurice et al.
further employed the same method to interpret the nonintuitive
ZFS in MnII complexes.113
The ORCA code, which is capable of using DFT methods for
ZFS as well, is currently the most often employed package for
ZFS calculations. Excellent ab initio results (95% of experi-
ments) have been obtained with ORCA for the very small
anisotropy parameter D of copper acetate monohydrate,114
and reliable D and E/D values have been obtained for a set of
Co(II) complexes.115 A surprisingly high quality for the predic-
tion of D with the SA-CASSCF method has been demonstrated
for the series of MnIII complexes by Duboc et al.100 Excellent D
results have been obtained also for the hexaqua-chromium(II/
III) complexes by Liakos et al.104
Apart from the MOLCAS and ORCA codes, other one-
component ZFS implementations exist, cf. the Introduction
of Neese.99 Two-component calculations of ZFS have been
implemented in the DFT-based ReSpect code.83 The implemen-
tation in ReSpect and comparison of a two-component and
one-component methods for ZFS parameters of group 16 triplet
diatomics and for the octet GdH3 molecule have been done by
Reviakine et al.41 Two-component and four-component calcu-
lations of ZFS are possible within the Dirac code116 that enables
a number of both density functional and wavefunction
approaches. An application study using the one-step Dirac and
two-step ORCA calculations has been done for a series of chal-
cogen diatomics by Rota et al., who found a high accuracy for
the two- and four-component methods, while discovering only
80% accuracy for the two-step one-component methods.42
9.13.12 Conclusion
The key parameters of the EPR spin Hamiltonian are: for any
paramagnetic system, the g-tensor (g); for any system involving
a magnetic nucleus, the hyperfine coupling tensor (A); and for
any system possessing more than one unpaired electron, the
zero-field splitting tensor (D). While the A-tensor can be, in
the first approximation, treated at the nonrelativistic level, the
g-tensor and the D-tensor are intrinsically relativistic properties
dominated for transition-metal systems by the SOC.
Any meaningful calculation of the hyperfine coupling ten-
sors requires a proper description of the electroncorrelation
effects. Within the state-of-the-art DFT, so far, no functional
has been identified that would be universally applicable to all
transition-metal systems of interest. The hybrid functionals
often perform better than the GGA functionals due to the better
description of the spin polarization; the results may, however,
be deteriorated when significant spin-contamination sets in. As
a compromise, it is, thus, advisable to employ functionals with
ca. 20% HF exchange, such as the B3LYP or B3PW91. After
careful calibration on a class of systems of interest, the amount
of the HF exchange can be increased or decreased, depending on
the spin-polarization vs. spin-contamination importance. From
the class of more recent functionals, the TPSSh and B2PLYP
functionals can be recommended. The main advantage of the
density functional methods is their low computational cost,
enabling to treat relatively large systems, and a conceptually
simple analysis in terms of contributions of the individual
KohnSham orbitals. However, for systems whose size enables
to employ the MR-CI, the CCSD, and the CCSD(T) methods,
these are the best candidates for a reliable description of the
hyperfine coupling.
DZP bases (e.g., the 6-31G* basis frequently employed for
geometry optimizations) are not sufficiently accurate to gener-
ate satisfactory HFCCs. However, they can give reliable results
if uncontracted in the outer coreinner valence region. The
7s4p2d basis (6s2p for hydrogen), known as EPR-III, has
been purposely tailored by Barone for reliable magnetic prop-
erties calculations. For transition-metal systems, the 9s7p4d
basis set can be recommended.
The available ab initio A-tensor implementations are all
established within the nonrelativistic framework. When rela-
tivistic effects are to be included, it is advisable to employ one
of the DFT packages: the ADF program, the MAG-ReSpect code,
the ORCA software, the NWChem package, or the BAND mod-
ule of the ADF program for solid-state systems.
The situation for the g-tensors is different, since higher-order
SO effects are often significantly larger than the dependence of

the results on the exchangecorrelation functional. It is, thus,
very important at what level the relativistic effects are treated. In
general, the two-component and the four-component methods
are preferable to the one-component approaches. From this, it
follows that the validation and selection of the best functionals
for g-tensor calculations depends on the treatment (or omis-
sion) of the higher-order SO contributions. While for calcula-
tions including higher-order SO contributions g-tensors are
reproduced best by hybrid density functionals with about
3040% exact-exchange admixture, the ideal amount of HF
exchange can be only 1020% if higher-order SO contributions
are omitted, while a larger amount of HF exchange deteriorates
the results. Medium-sized DZVP basis sets can be sufficient if
they specifically tailored to the system of interest. From the
ab initio methods for g-tensor calculations, the best results are
provided by the MRCISD Q method, the DDCI3 Q method,
and the CC method.
In the case of the D-tensor, again, the two-component and
the four-component methods are preferable to the one-
component approaches. However, these cannot always be
employed for the reasons of computational requirements and
because the majority of implementations is done at the one-
component level. Out of the DFT methods, the best predictions
have been obtained with the B3LYP functional, followed by the
nonhybrid BP86 functional, which, in turn, was found to be
more successful than the meta-hybrid GGA functional TPSSh.
However, in general, the DFT calculations of ZFS turn out to be
inferior to the wavefunction-based ab initio methods. Cur-
rently, the most often employed approach is a two-step one-
component scheme combining the CASSCF and CASPT2 or
NEVPT2 methods.
Acknowledgments
This work was supported by the project CEITEC Central
European Institute of Technology from the European
Regional Development Fund (grant number CZ.1.05/1.1.00/
02.0068). The author thanks Martin Kaupp for providing
several figures included in this work, Olga Malkin, Frank
Neese, and Alexei Arbuznikov for helpful comments, and
Dominik Munzar for the critical reading of the manuscript.
References
1. Gerlach, W.; Stern, O. Z. Phys. 1922, 9, 349.
2. Uhlenbeck, G.; Goudsmith, S. Naturwissenschaften 1925, 13, 593 Nature 1926,
117, 263.
3. Reiher, M.; Wolf, A. Relativistic Quantum Chemistry: The Fundamental Theory of
Molecular Science. Wiley-VCH: Weinheim, 2009.
4. Weil, J. A.; Bolton, J. R.; Wertz, J. E. Electron Paramagnetic Resonance:
Elementary Theory and Practical Applications. Wiley and Sons: New York, 1994;
pp 4, 34, 3840.
5. Levine, I. N. Quantum Chemistry, 6th ed.; Pearson/Prentice Hall: Upper Saddle
River, NJ, 2009; pp 283284, 338339, 477.
6. Harriman, J. E. Theoretical Foundations of Electron Spin Resonance. Academic
Press: New York, 1978; pp 8788.
7. Jansen, H. J. F. Phys. Rev. B 1999, 59, 4699.
8. van Wullen, C. J. Chem. Phys. 2009, 130, 194109.
9. Kramers, H. A. Quantum Mechanics. North Holland: Amsterdam, 1957.
10. Neese, F. In Calculation of NMR and EPR Parameters; Kaupp, M., Buehl, M.,
Malkin, V. G., Eds.; Wiley-VCH: Weinheim, 2004; pp 541564, 581591.
Electron Paramagnetic Resonance 379
11. Symons, M. Chemical and Biochemical Aspects of Electron-Spin
Resonance Spectroscopy. Van Nostrand Reinhold: Wokingham, 1978;
pp 129130, 142.
12. Ban, F., Gauld, J. W., Boyd, R. J., Eds.; In Calculation of NMR and EPR
Parameters; Wiley-VCH: Weinheim, 2004; pp 567580.
13. Malkin, V. G.; Malkina, O. L.; Eriksson, L. A.; Salahub, D. R. In Modern Density
Functional Theory: A Tool for Chemistry Theoretical and Computational
Chemistry; Politzer, P., Seminario, J. M., Eds.; Elsevier: Amsterdam, 1995; Vol. 2,
pp 273347.
14. Atherton, N. M. Principles of Electron Spin Resonance. Ellis Horwood: Chichester,
1993; pp 194195.
15. Cano, J.; Ruiz, E.; Alvarez, S.; Verdaguer, M. Comment Inorg. Chem. 1998, 20,
2756.
16. Munzarova, M. L.; Kubacek, P.; Kaupp, M. J. Am. Chem. Soc. 2000, 122,
1190011913.
17. Munzarova, M.; Kaupp, M. J. Phys. Chem. A 1999, 48, 99669983.
18. Kuska, H. A.; Rogers, M. T. J. Chem. Phys. 1964, 40, 910911.
19. Goodman, B. A.; Raynor, J. B.; Symons, M. C. R. J. Chem. Soc. A 1966,
994997.
20. Munzarova. M. L. unpublished results.
21. Lowe, J. P. Quantum Chemistry, 2nd ed.; Academic Press: San Diego, 1993;
pp 392428.
22. Mabbs, F. E.; Collison, D. Electron Paramagnetic Resonance of d Transition Metal
Compounds. Elsevier: Amsterdam, 1992; pp 338441.
23. Drago, R. S. Physical Methods for Chemists. Saunders College Publishing:
Orlando, 1992.
24. Munzarova, M. L.; Kaupp, M. J. Phys. Chem. B 2001, 105, 1264412652.
25. Micera, G.; Garribba, E. J. Comput. Chem. 2011, 32, 28222835.
26. Micera, G.; Garribba, E. Eur. J. Inorg. Chem. 2011, 37683780.
27. Jean, Y.; Volatron, F. An Introduction to Molecular Orbitals. Oxford University
Press: New York, 1993; p 80.
28. Vancoillie, S.; Malmqvist, P.; Pierloot, K. Chem. Phys. Chem. 2007, 8,
18031815.
29. Bolvin, H. Chem. Phys. Chem. 2006, 7, 15751589.
30. Patchovskii, S.; Schreckenbach, G. In Calculation of NMR and EPR
Parameters; Kaupp, M., Buehl, M., Malkin, V. G., Eds.; Wiley: Weinheim, 2004;
pp 505532.
31. Belanzoni, P.; Baerends, E. J.; van Asselt, S.; Langewen, P. B. J. Phys. Chem.
1995, 99, 1309413102.
32. van Lenthe, E.; Wormer, P. E. S.; van der Avoird, A. J. Chem. Phys. 1997, 107,
24882498.
33. DeVore, T. C.; Weltner, W., Jr. J. Am. Chem. Soc. 1977, 99, 4700.
34. Albright, T. A.; Burdett, J. K.; Whangbo, M.-H. Orbital Interactions in Chemistry.
John Wiley and Sons: New York, 1985; pp 295298.
35. (a) Kasai, P. H.; Whipple, E. B.; Weltner, W., Jr. J. Chem. Phys. 1966, 44, 2581;
(b) Wilso, R.; Kivelson, D. J. Chem. Phys. 1966, 44, 4445; (c) Maki, A. H.;
McGarvey, B. R. J. Chem. Phys. 1958, 29, 31 35.
36. (a) Fairhust, S. A.; Morton, J. R.; Preston, K. F. J. Magn. Reson. 1983, 55, 453;
(b) Hanlan, L. A.; Huber, H.; Kundig, E. P.; McGarvey, B. R.; Ozin, G. A. J. Am.
Chem. Soc. 1975, 97, 7054; (c) Morton, J. R.; Preston, K. F. J. Chem. Phys.
1984, 81, 5775.
37. Malkina, O. L.; Vaara, J.; Schimmelpfennig, B.; Munzarova, M.; Malkin, V. G.;
Kaupp, M. J. Am. Chem. Soc. 2000, 122, 92069218.
38. Munzarova, M. Ph.D. Thesis, Universitat Stuttgart, 2001, p 56.
39. Visscher, L. In Relativistic Electronic Structure Theory; Part I: Fundamentals;
Schwerdtfeger, P., Ed.; Elsevier: Amsterdam, 2002; p 291.
40. De Graaf, C.; Sousa, C. Int. J. Quant. Chem. 2006, 106, 24702478.
41. Reviakine, R.; Arbuznikov, A. V.; Tremblay, J. C.; Remenyi, C.; Malkina, O. L.;
Malkin, V. G.; Kaupp, M. J. Chem. Phys. 2006, 125, 054110.
42. Rota, J.-B.; Knecht, S.; Fleig, T.; Ganyushin, D.; Saue, T.; Neese, F.; Bolvin, H.
J. Chem. Phys. 2011, 135, 114106.
43. Koch, W.; Holthausen, M. A Chemists Guide to Density Functional Theory.
Wiley-VCH: Weinheim, 2000; pp 29, 7585.
44. Hadt, R. G.; Nemykin, V. N.; Olsen, J. G.; Basu, P. Phys. Chem. Chem. Phys.
2009, 11, 1037710384.
45. Pawlak, T.; Munzarova, M. L.; Pazderski, L.; Marek, R. J. Chem. Theor. Comput.
2011, 7, 39093923.
46. Munzarova, M. L. In Calculation of NMR and EPR Parameters: Theory and
Applications; Kaupp, M., Buhl, M., Malkin, V. G., Eds.; Wiley-VCH: Weinheim,
2004; pp 466471.
47. (a) Rega, N.; Cossi, M.; Barone, V. J. Chem. Phys. 1996, 105, 11060;
(b) Barone, V. J. Phys. Chem. 1995, 99, 11659.
48. Malkin, V. G.; Malkina, O. L.; Salahub, D. Chem. Phys. Lett. 1994, 221, 91.
380 Electron Paramagnetic Resonance
49. Hermosilla, L.; Calle, P.; Garca de la Vega, J. M.; Sieiro, C. J. Phys. Chem.
A 2005, 109, 11141124.
50. Kaupp, M.; Arbuznikov, A.; Hesselmann, A.; Goerling, A. J. Chem. Phys. 2010,
132, 184107.
51. Micera, G.; Garribba, E. Dalton Trans. 2009, 19141918.
52. Kossmann, S.; Kirchner, B.; Neese, F. Mol. Phys. 2007, 105, 20492071.
53. Barone, V. In Recent Advances in Density Functional Methods, Part 1;
Chong, D. P., Ed.; World Scientific: Singapore, 1995; p 287.
54. Engels, B.; Eriksson, L. A.; Lunell, S. Adv. Quantum Chem. 1996, 27, 297.
55. Suter, H. U.; Pless, V.; Ernzenhof, M.; et al. Chem. Phys. Lett. 1994,
230, 398.
56. Fritscher, J.; Hrobarik, P.; Kaupp, M. J. Phys. Chem. B 2007, 111,
46164629.
57. Autschbach, J.; Patchovskii, S.; Pritchard, B. J. Chem. Theory Comput. 2011, 7,
21752188.
58. Remenyi, C.; Reviakine, R.; Arbuznikov, A.; Vaara, J.; Kaupp, M. J. Phys. Chem.
A 2004, 108, 50265033.
59. Aquino, F.; Prichard, B.; Autschbach, J. J. Chem. Theor. Comput. 2012, 8,
598609.
60. Arbuznikov, A. V.; Vaara, J.; Kaupp, M. J. Chem. Phys. 2004, 120,
21272139.
61. Belanzoni, P.; van Lenthe, E.; Baerends, E. J. J. Chem. Phys. 2001, 114,
44214433.
62. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.;
Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.;
Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.;
Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.;
Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.;
Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.;
Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.;
Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.;
Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.;
Daniels, A. D.; Farkas, O.; Foresman, J. B.; Cioslowski, J.; Fox, D. J. Gaussian 09,
Revision A.1; Gaussian, Inc.: Wallingford, CT, 2009.
63. ADF2010, SCM, Theoretical Chemistry, Vrije Universiteit, Amsterdam, The
Netherlands, http://www.scm.com.
64. Komorovsky, S.; Repisky, M.; Malkina, O. L.; Malkin, V. G.; Malkin, I.; Kaupp, M.
J. Chem. Phys. 2006, 124, 084108.
65. Malkin, E.; Repisky, M.; Komorovsky, S.; Mach, P.; Malkina, O. L.; Malkin, V. G.
J. Chem. Phys. 2011, 134, 044111.
66. Neese, F. ORCA An Ab Initio DFT and Semiempirical Program Package, 2.820
ed.; University of Bonn: Bonn, 2010.
67. Sundararajan, M.; Ganyushin, D.; Ye, S.; Neese, F. Dalton Trans. 2009, 6021.
68. Kadantsev, E. S.; Ziegler, T. Magn. Reson. Chem. 2010, 48, S2S10.
69. Engels, B. In Calculation of NMR and EPR Parameters; Kaupp, M., Buhl, M.,
Malkin, V. G., Eds.; Wiley-VCH: Weinheim, 2004; pp 483492.
70. Puzzarini, C.; Barone, V. J. Chem. Phys. 2010, 133, 184301.
71. (a) Engels, B. J. Chem. Phys. 1994, 100, 1380; (b) Engels, B. Chem. Phys. Lett.
1991, 179, 398.
72. Filatov, M.; Cremer, D. J. Chem. Phys. 2004, 121, 56185622.
73. Neese, F. Magn. Reson. Chem. 2004, 42, S187S198.
74. Stanton, J. F.; Gauss, J.; Harding, M. E.; Szalay, P. G. CFOUR Coupled Cluster
Techniques for Computational Chemistry, a quantum-chemical program package;
with contributions from Auer, A. A.; Bartlett, R. J.; Benedikt, U.; Berger, C.;
Bernholdt, D. E.; Bomble, Y. J.; Christiansen, O.; Heckert, M.; Heun, O.; Huber, C.;
Jagau, T.-C.; Jonsson, D.; Juselius, J.; Klein, K.; Lauderdale, W. J.; Matthews, D.;
Metzroth, T.; ONeill, D. P.; Price, D. R.; Prochnow, E.; Ruud, K.; Schiffmann, F.;
Stopkowicz, S.; Varner, M. E.; Vazquez, J.; Watts, J. D.; Wang, F. and the integral
packages MOLECULE - J. Almlof and P. R. Taylor, PROPS - P. R. Taylor, ABACUS -
T. Helgaker, H. J. Aa. Jensen, P. Jrgensen, and J. Olsen, and ECP routines by A. V.
Mitin and C. van Wullen.; For the current version, see http://www.cfour.de.
75. Ciofini, I. In Calculation of NMR and EPR Parameters; Kaupp, M., Buhl, M.,
Malkin, V. G., Eds.; Wiley-VCH: Weinheim, 2004; pp 191208.
76. Hrobarik, P.; Repisky, M.; Komorovsky, S.; Hrobarikova, V.; Kaupp, M. Theor.
Chem. Acc. 2011, 129, 715725.
77. Fritscher, J.; Hrobarik, P.; Kaupp, M. Inorg. Chem. 2007, 46, 81468161.
78. Remenyi, C.; Munzarova, M. L.; Kaupp, M. J. Phys. Chem. B 2005, 109,
42274233.
79. Arbuznikov, A. V.; Kaupp, M. J. Chem. Theor. Comput. 2009, 5, 29852995.
80. Patchkovskii, S.; Ziegler, T. J. Chem. Phys. 1999, 111, 5730.
81. Fritcher, J.; Hrobarik, P.; Kaupp, M. J. Phys. Chem. B 2007, 111, 4616.
82. van Wullen, C. In Calculation of NMR and EPR Parameters; Kaupp, M., Buehl, M.,
Malkin, V. G., Eds.; Wiley: Weinheim, 2004; pp 85100.
83. Malkin, V. G. et al. ReSpect Program, Version 3.1, 2007.
84. Repisky, M.; Komorovsky, S.; Malkin, E.; Malkina, O. L.; Malkin, V. G. Chem.
Phys. Lett. 2010, 488, 8794.
85. Horbarik, P.; Malkina, O. L.; Malkin, V. G.; Kaupp, M. Chem. Phys. 2009, 356,
229235.
86. Brownridge, S.; Grein, F.; Tatchen, J.; Kleinshmidt, M.; Marian, C. M. J. Chem.
Phys. 2003, 118, 95529562.
87. (a) Miralles, J.; Daudey, J.-P.; Caballol, R. Chem. Phys. Lett. 1992, 198, 555;
(b) Miralles, J.; Castell, O.; Caballol, R.; Malrieu, J.-P. Chem. Phys. 1993, 172,
3343.
88. Engstrom, M.; Vahtras, O.; Agren, H. Chem. Phys. Lett. 2000, 328, 483491.
89. Neese, F. Chem. Phys. Lett. 2003, 380, 721728.
90. Gauss, J.; Kallay, M.; Neese, F. J. Phys. Chem. A 2009, 113, 1154111549.
91. Karlstrom, G.; Lindh, R.; Malmqvist, P.-A.; Roos, B. O.; Ryde, U.; Veryazov, V.;
Widmark, P.-O.; Cossi, M.; Schimmelpfennig, B.; Neogrady, P.; Seijo, L. Comp.
Mater. Sci. 2003, 28, 222.
92. Pederson, M. R.; Khanna, S. N. Phys. Rev. B 1999, 60, 9566.
93. Takeda, R.; Mitsuo, S.; Yamanaka, S.; Yamaguchi, K. Polyhedron 2005, 24, 2238.
94. Aquino, F.; Rodriguez, J. H. J. Chem. Phys. 2005, 123, 204902.
95. Neese, F. J. Am. Chem. Soc. 2006, 128, 10213.
96. (a) Schoneboom, J.; Neese, F.; Thiel, W. J. Am. Chem. Soc. 2005, 128, 5840;
(b) Ray, K.; Weyhermuller, T.; Neese, F.; Wieghardt, K. Inorg. Chem. 2006, 5, 2027.
97. Ganyushin, D.; Neese, F. J. Chem. Phys. 2006, 125, 024103.
98. Berry, J. F.; Bill, E.; Neese, F.; Garcia-Serres, R.; Weyhermuller, T.; Wieghardt, K.
Inorg. Chem. 2006, 45, 2027.
99. Neese, F. J. Chem. Phys. 2007, 127, 164112.
100. Duboc, C.; Ganyushin, D.; Sivalingam, K.; Collomb, M.-N.; Neese, F. J. Phys.
Chem. A 2010, 114, 1075010758.
101. Schmitt, S.; Jost, P.; van Wullen, C. J. Chem. Phys. 2011, 134, 194113.
102. Zein, S.; Neese, F. J. Phys. Chem. A 2008, 112, 79767983.
103. Stoyanova, R.; Barra, A.-L.; Ynocheva, M.; Kuzmanova, E.; Zhecheva, E. Dalton
Trans. 2011, 40, 91069115.
104. Liakos, D. G.; Ganyushin, D.; Neese, F. Inorg. Chem. 2009, 48, 1057210580.
105. Misochko, E. Y.; Korchagin, D. V.; Bozhenko, K. V.; Chapyshev, S. V.;
Aldoshin, S. M. J. Chem. Phys. 2010, 133, 064101.
106. Cirera, J.; Ruiz, E.; Alvarez, S.; Neese, F.; Kortuz, J. Chem. Eur. J. 2009, 15,
40784087.
107. Kowalewski, J.; personal communication.
108. van Gastel, M. J. Chem. Phys. 2009, 131, 124111.
109. Karlstrom, G.; Lindh, R.; Malmquist, P.-A.; Roos, B. O.; Ryde, U.; Veryazov, V.;
Widmark, P.-O.; Cossi, M.; Schimmelpfennig, B.; Neogrady, P.; Seijo, L. Comp.
Mater. Sci. 2003, 28, 222.
110. (a) Malmquist, P.-A.; Roos, B. O. Chem. Phys. Lett. 1989, 155, 189;
Malmquist, P.-A.; Roos, B. O.; Schimmelpfennig, B. Chem. Phys. Lett. 2002,
357, 230.
111. Neese, F. personal communication.
112. Maurice, R.; Bastardis, R.; de Graaf, C.; Suaud, N.; Mallah, T.; Guihery, N.
J. Chem. Theor. Comp. 2009, 5, 29772984.
113. Maurice, R.; de Graaf, C.; Guihery, N. J. Chem. Phys. 2010, 133, 084307.
114. Maurice, R.; Sivalingam, K.; Ganyushin, D.; Guihery, N.; de Graaf, C.; Neese, F.
Inorg. Chem. 2011, 50, 62296236.
115. Maganas, D.; Sottini, S.; Kyritsis, P.; Groenen, E. J. J.; Neese, F. Inorg. Chem.
2011, 50, 87418754.
116. Dirac, a relativistic ab initio electronic structure program, written by T. Saue, L.
Visscher, and H. J. Aa. Jensen, release DIRAC10, R. Bast, K. G. Dyall, U.
Ekstrom, et al.