OULU 2007

C 283

ACTA
Virpi Krger

U N I V E R S I T AT I S O U L U E N S I S

TECHNICA

POISONING OF AUTOMOTIVE EXHAUST GAS CATALYST COMPONENTS

THE ROLE OF PHOSPHORUS IN THE POISONING PHENOMENA

FACULTY OF TECHNOLOGY, DEPARTMENT OF PROCESS AND ENVIRONMENTAL ENGINEERING, MASS AND HEAT TRANSFER PROCESS LABORATORY, UNIVERSITY OF OULU

ACTA UNIVERSITATIS OULUENSIS

C Te c h n i c a 2 8 3

VIRPI KRGER

POISONING OF AUTOMOTIVE EXHAUST GAS CATALYST COMPONENTS

The role of phosphorus in the poisoning phenomena

Academic dissertation to be presented, with the assent of the Faculty of Technology of the University of Oulu, for public defence in Kuusamonsali (Auditorium YB210), Linnanmaa, on November 10th, 2007, at 12 noon

O U L U N Y L I O P I S TO, O U L U 2 0 0 7

Copyright 2007 Acta Univ. Oul. C 283, 2007

Supervised by Professor Riitta L. Keiski

Reviewed by Professor Edd Anders Blekkan Professor Jose Luis Garca Fierro

ISBN 978-951-42-8607-0 (Paperback) ISBN 978-951-42-8608-7 (PDF) http://herkules.oulu.fi/isbn9789514286087/ ISSN 0355-3213 (Printed) ISSN 1796-2226 (Online) http://herkules.oulu.fi/issn03553213/ Cover design Raimo Ahonen

OULU UNIVERSITY PRESS OULU 2007

Krger, Virpi, Poisoning of automotive exhaust gas catalyst components. The role of phosphorus in the poisoning phenomena
Faculty of Technology, University of Oulu, P.O.Box 4000, FI-90014 University of Oulu, Finland, Department of Process and Environmental Engineering, Mass and Heat Transfer Process Laboratory, University of Oulu, P.O.Box 4300, FI-90014 University of Oulu, Finland Acta Univ. Oul. C 283, 2007 Oulu, Finland

Abstract
The aim of this thesis project was to gain new knowledge on the effect of phosphorus on the catalytic activity and characteristics of automotive exhaust gas catalyst components. The simultaneous roles of phosphorus and calcium were also studied. The first test series of powdery catalyst samples contained Rh and oxide (Test series 1) and the second, Pt and oxide or ZSM-5 (Test series 2). The catalysts were analyzed when fresh and after two ageing and phosphorus poisoning procedures developed in this work. The procedures consisted of adding poison via impregnation in an aqueous solution (for Test series 1) and in the gaseous phase under hydrothermal conditions (for Test series 2). The poison compounds formed and the changes in the washcoat were studied by using physisorption analyses, SEM, TEM, XRD, and FTIR-ATR. The poison content of the samples was determined by ICP-OES and XRF. Laboratory-scale activity measurements were done to investigate the catalytic activity. Thermodynamic calculations were used to obtain information about ageing conditions and phosphorus compounds formed during ageing. Phosphorus decreased the catalytic activity and the characteristic surface areas of the catalysts. Addition of calcium to a phosphorus-poisoned catalyst was found to have even a regenerating effect on the catalysts' activity. The poisoning methods developed in this study resulted in the same phosphorus compounds as can be found in vehicle-aged catalysts. Phosphorus was identified as cerium, zirconium, aluminium, and titanium phosphates. Phosphorus was detected in zeolites, but phosphorus-containing compounds were not observed. Phosphorus poisoning takes place in the gas phase at high operating temperatures and with high oxygen and water contents. It was also shown that the role of phosphorus poisoning was more pronounced than the role of hydrothermal ageing alone. Phosphorus poisoning mainly affects the oxide components used in this study, not the noble metals. The results can be utilized in the development of catalytic materials and catalyst compositions that can better tolerate phosphorus poisoning under hydrothermal conditions. The results can also be applied in evaluating the effects of phosphorus on different catalyst compositions and in estimating the age of commercial catalysts.

Keywords: calcium, catalyst activity, deactivation, diesel catalysts, phosphorus, platinum, poisoning, rhodium, three-way catalysts, zeolite catalyst

To my parents

Preface
This study was carried out in the Department of Process and Environmental Engineering at the University of Oulu during the years 20032006. During this time the author was also a student in the Graduate School on Functional Surfaces. In addition, the work was a part of the Catalytical Materials: Characterization and Control of the Surface Poisoning Phenomena (POISON) project funded by the Academy of Finland. The study was carried out in co-operation with Ecocat Oy, the Department of Chemistry at the University of Oulu, and the Institute of Materials Science at Tampere University of Technology. Part of the work was performed in the School of Chemical Engineering and Industrial Chemistry at the University of New South Wales, Australia, during a research visit in 2005. I cordially thank Prof. Riitta Keiski, the supervisor of this thesis project, for the opportunity to work in her research group in the Mass and Heat Transfer Process Laboratory. I also highly appreciate her expert knowledge on catalysis and the support she gave me during the work. My warm thanks are due to the advisor of the work, Prof. Ulla Lassi, whose help and encouragement were invaluable. M.Sc. Marko Hietikko, M.Sc. Tomi Kanerva, Dr. Minnamari Vippola, Prof. Toivo Lepist, Prof. Risto Laitinen, Dr. Katariina Rahkamaa-Tolonen, M.Sc. Aslak Suopanki, Dr. Kauko Kallinen, Dr. Dennys Angove, Dr. David French, Lic.Tech. Juha Ahola, and M.Sc. Esa Turpeinen are acknowledged for their successful co-operation in the research work. I also express my thanks to Prof. David Trimm, Prof. Noel Cant, and Dr. Yun Lei for increasing my knowledge in the field of catalysis and for the unforgettable time I spent in Australia. I thank Mr. Jorma Penttinen and Ms. Hannele Nurminen for their invaluable help in developing the test procedures and in carrying out the experiments. I also thank M.Sc. Katri Kynknniemi for her assistance in the laboratory work. I present my gratitude to Prof. Edd Anders Blekkan and Prof. Jose Luis Garca Fierro, who reviewed the manuscript of my thesis. Mr. Keith Kosola is acknowledged for linguistic corrections. I warmly thank Dr. Tanja Kolli, Dr. Mika Huuhtanen, and Dr. Satu Ojala for the help and support they have given me during the past years. Thanks are also due to the other members of the Mass and Heat Transfer Process Laboratory staff. Special thanks are expressed to M.Sc. Riitta Raudaskoski for fruitful co-operation in both work and leisure activities.

Finally, I present loving thanks to my family. I thank my spouse Toni Kallio for bringing joy and a positive challenge into my life. I most cordially thank my mother Aira, father Heikki, sister Anne, and brother Vesa for their continuous care and support. This thesis is dedicated to my parents, who always encouraged me in my studies. The work was funded by the Graduate School on Functional Surfaces and the Academy of Finland. I also gratefully acknowledge the financial support given by the Faculty of Technology at the University of Oulu; Fortum Foundation; Oulu University Scholarship Foundation; Foundation of Technology; Jenny and Antti Wihuri Foundation; The Finnish Cultural Foundation, The Regional Fund of Northern Ostrobothnia; Finnish Konkordia Fund; and Finnish Catalysis Society. Ecocat Oy is acknowledged for donating catalysts for the laboratory tests. Oulu, August 2007 Virpi Krger

List of abbreviations
A/F ratio ATR BET BJH DOC EDS EDTA EGR EU FTIR HC ICP-OES OSC PGM PM SCR SEM SOF T50 TEM TPD TWC XRD XRF ZDDP Air-to-fuel ratio Attenuated Total Reflection Brunauer-Emmet-Teller theory Barrett-Joyner-Halenda theory Diesel Oxidation Catalyst Energy Dispersive X-ray Spectrometer Ethylene Diamine Tetraacetic Acid Exhaust Gas Recirculation European Union Fourier Transform Infrared Spectroscopy Hydrocarbon Induced Coupled Plasma Optical Emission Spectroscopy Oxygen Storage Capacity Platinum Group Metal Particulate Matter Selective Catalytic Reduction Scanning Electron Microscopy Soluble Organic Fraction The temperature of 50% conversion of the feed component Transmission Electron Microscopy Temperature Programmed Desorption Three-way Catalyst X-ray Diffraction X-ray Fluorescence Zinc Dialkyldithiophosphate

10

List of original papers

This thesis is based on the following publications, which are referred to in the text by their Roman numerals:
I Krger V, Hietikko M, Lassi U, Ahola J, Kallinen K, Laitinen R & Keiski RL (2004) Characterization of the effects of phosphorus and calcium on the activity of Rhcontaining catalyst powders. Topics in Catalysis 30/31: 469473. II Krger V, Lassi U, Kynknniemi K, Suopanki A & Keiski RL (2006) Methodology development for laboratory-scale exhaust gas catalyst studies on phosphorus poisoning. Chemical Engineering Journal 120: 113118. III Krger V, Hietikko M, Angove D, French D, Lassi U, Suopanki A, Laitinen R & Keiski RL (2007) Effect of phosphorus poisoning on catalytic activity of diesel exhaust gas catalyst components containing oxide and Pt. Topics in Catalysis 4243: 409413. IV Krger V, Kanerva T, Lassi U, Rahkamaa-Tolonen K, Vippola M & Keiski RL (2007) Characterization of phosphorus poisoning on diesel exhaust gas catalyst components containing oxide and Pt. Topics in Catalysis 45: 153157. V Krger V, Kanerva T, Lassi U, Rahkamaa-Tolonen K, Lepist T & Keiski RL (2007) Phosphorus poisoning of ZSM-5 and Pt/ZSM-5 zeolite catalysts in diesel exhaust gas conditions. Topics in Catalysis 4243: 433436. VI Kanerva T, Krger V, Rahkamaa-Tolonen K, Vippola M, Lepist T & Keiski RL (2007) Structural changes in air aged and poisoned diesel catalysts. Topics in Catalysis 45: 137142.

The manuscripts for the publications (Papers IV) were written by the author of this thesis. In Paper VI, the author was responsible for the chemical ageing procedure, activity measurements, and surface area studies. The manuscript was written in close collaboration with the first author.

11

12

Contents
Abstract Preface 7 List of abbreviations 9 List of original papers 11 Contents 13 1 Introduction 15 1.1 Background ............................................................................................. 15 1.2 Purpose of the work ................................................................................ 16 2 Automotive exhaust gas catalysis 19 2.1 General.................................................................................................... 19 2.2 Three-way catalysis................................................................................. 20 2.3 Diesel catalysis........................................................................................ 22 2.3.1 Oxidation catalysis ....................................................................... 22 2.3.2 NOx reduction............................................................................... 24 2.3.3 Particulate filtration ...................................................................... 25 3 Catalyst deactivation 27 3.1 Overview................................................................................................. 27 3.2 Chemical deactivation............................................................................. 29 3.2.1 Poisoning by phosphates .............................................................. 30 3.2.2 Poisoning by sulphur compounds................................................. 33 3.2.3 Poisoning by other compounds .................................................... 34 3.3 Thermal deactivation............................................................................... 34 3.4 Mechanical deactivation ......................................................................... 35 3.5 Catalyst regeneration............................................................................... 35 4 Experimental 37 4.1 Catalysts.................................................................................................. 37 4.2 Phosphorus poisoning procedures........................................................... 38 4.2.1 Poisoning by impregnation........................................................... 39 4.2.2 Gaseous phase poisoning.............................................................. 39 4.3 Catalyst characterisation techniques ....................................................... 40 4.3.1 Physisorption analyses ................................................................. 41 4.3.2 Scanning Electron Microscopy..................................................... 42 4.3.3 Transmission Electron Microscopy .............................................. 42 4.3.4 Chemical analyses ........................................................................ 42 4.3.5 X-ray diffraction........................................................................... 42
13

4.3.6 Fourier Transform Infrared-Attenuated Total Reflection ............. 43 4.3.7 X-ray Photoelectron Fluorescence................................................ 43 4.3.8 Activity measurements ................................................................. 43 4.3.9 Thermodynamic calculations........................................................ 45 5 Poisoning-induced changes in the catalyst components 47 5.1 Loss in catalytic surface area .................................................................. 47 5.2 Accumulation of phosphorus .................................................................. 51 5.2.1 Phosphorus content....................................................................... 51 5.2.2 Phosphorus compounds ................................................................ 55 5.3 Loss of catalytic activity ......................................................................... 63 5.3.1 Effect of phosphorus..................................................................... 63 5.3.2 Effect of phosphorus and calcium ................................................ 68 5.4 Changes in the active metal particles and washcoat grain size ............... 69 6 Evaluation of the poisoning procedures 71 6.1 Integration of the results ......................................................................... 71 6.2 Evaluation and utilization of the results.................................................. 72 6.3 Suggestions for further research.............................................................. 73 7 Summary and conclusions 75 References 77 Original papers 85

14

1
1.1

Introduction
Background

Motor vehicles play a major role in urban air quality problems. Incomplete burning of petrol or diesel leads to formation of pollutants that have several harmful effects on the environment and human health. These pollutants include carbon monoxide (CO), hydrocarbons (HCs), nitrogen oxides (NOx), and particulate matter (PM) (Kalantar Neyestanaki et al. 2004, Wallington et al. 2006). Catalysis is one of the key technologies used in controlling air pollution. Catalytic abatement of exhaust gas components has been used to reduce automobile emissions since the 1970s, and it has been a success story in the development of environmental catalysts. The increasing motor vehicle population and tightening emission standards create a major challenge for the development of more durable and effective automotive catalysts. Demand for a high performance catalyst and requirements for long catalyst life have increased the amount of research activity focused on automotive exhaust gas catalysis. (Greening 2001, Kapar et al. 2003, Twigg 2007) Catalytic materials used in future three-way catalyst (TWC) and diesel catalyst applications will have to be more effective in controlling emissions. Thermal stability and good tolerance of poisons such as sulphur (S), which originates from fuel, and phosphorus (P), zinc (Zn), calcium (Ca), and magnesium (Mg), which originate from lubricating oils, are also needed. (Heck & Farrauto 1997, Farrauto & Heck 2000, Shelef & McCabe 2000, Lassi 2003) Many challenges remain before researchers and the catalyst industry can fulfil such demands. Robust research efforts are required to clarify the functioning of noble metals and washcoat components and to control deactivation phenomena. Further improvements are also needed in the efficiency and optimization of the entire system of engine, fuel, and exhaust gas after-treatment. (Wallington et al. 2006) Phosphorus and sulphur are present in aged catalytic converters in higher concentrations than other poisons, which increases the probability of their interaction with catalyst components such as aluminium oxide (Al2O3) and cerium oxide (CeO2) (Cabello Galisteo et al. 2004). The role of phosphorus in the deactivation of automotive catalysts has been discussed in several studies, but the mechanism of phosphorus poisoning is not yet completely known.
15

Nowadays, over 95% of the worlds transportation fuel originates from fossil fuels (Wallington et al. 2006). However, due to the quest to reduce dependency on fossil fuels and to control climatic change, interest in forms of energy that would replace traditional crude oil-based fuels is growing (Hamelinck & Faaij 2006, Romm 2006, Mittelbach 1995, Sagar 1995). According to EU directive 2003/30/EC, the EU countries are supposed to replace the use of gasoline and diesel with a renewable form of energy to the extent that by the year 2010, 5.75% of energy, calculated according to energy content, would come from renewable sources (EU 2003). Therefore, production and use of biodiesel and other fuels derived from biomass is rapidly increasing (Mittelbach 1995). This emphasizes the role of phosphorus in exhaust gas catalysis. When using traditional fossil fuels, lubricating oils are normally a major source of phosphorus. In the case of biodiesel, phosphorus as well as potassium (K) and other possible poison compounds may be present in various concentrations, depending on the origin of the biomass (Grosser & Bowman 2006). Despite the increasing risk of phosphorus poisoning in automotive catalysis, the number of publications dealing with the effect of phosphorus on separate catalyst components is still very limited. Therefore, this study was carried out with simplified model catalysts in order to obtain detailed knowledge on the poisoning phenomena of catalyst components. Research work in this area is essential in order to meet emission limits in the future. Research on deactivation of exhaust gas catalysts is widely carried out by studying vehicle-aged, engine-aged, and oven-aged catalysts on the laboratory scale (Koltsakis & Stamatelos 1997, Xu et al. 2004, Uy et al. 2003, Cabello Galisteo et al. 2004). Vehicle- and engine-ageing methods provide realistic information about the overall deactivating effect caused by oil- and fuel-derived contaminants and high operating temperatures. Their disadvantages are that they are relatively expensive and time-consuming methods (Koltsakis & Stamatelos 1997). It may also be difficult to determine the relevance of a single mechanism of deactivation to changes in catalytic activity and characteristics. Therefore, procedures for studying chemical ageing on the laboratory scale were developed in this study. 1.2 Purpose of the work

In the present work, chemical deactivation of catalyst powders was studied. The mechanism of phosphorus poisoning in automotive catalysts is not yet completely
16

known, and therefore, phosphorus was chosen to be the poisoning compound in ageing. The purpose of this thesis project was to examine the effect of phosphorus on the catalytic activity and characteristics of components widely used in automotive exhaust gas catalyst applications. As a special case, the combined effect of phosphorus and calcium was also studied. To understand poisoninginduced changes in the washcoat, alterations in the surface of powdered catalyst components, caused by phosphorus poisoning and hydrothermal ageing, were studied using several characterization techniques. In addition, differences between phosphorus poisoning of catalysts with and without a precious metal were investigated. The characteristics of aged samples containing different oxides were also compared. Vehicle-ageing and engine-ageing procedures are widely used in studying poisoning phenomena of exhaust gas catalysts. However, these methods are often relatively expensive and slow. In addition, studying the effects of a single deactivation mechanism on a decrease in catalytic activity may be complicated. The mechanism of phosphorus poisoning is not completely known, and therefore studies carried out with simplified model components are required. One objective of this study was to develop a methodology for accelerated chemical deactivation studies that can be carried out in a time-saving, affordable, safe, and controlled way on the laboratory scale. In this research, deactivation of automotive catalysts caused by phosphorus poisoning was studied by developing two different chemical ageing procedures on the laboratory scale. To obtain detailed knowledge about the mechanism of poisoning, the studies were carried out with simplified model components. Information about the formation and stability of poison compounds was gained by using thermodynamic calculations. These calculations were also used as a tool for planning ageing conditions. Another purpose of the work was to study the mechanism of phosphorus poisoning with different catalyst components used in present-day catalytic converters. The work also focused on obtaining information about the conditions in which poisoning takes place and finding out whether some components are more sensitive to phosphorus than others. A further purpose of the work was to differentiate between the role of thermal ageing and the role of phosphorus poisoning in catalyst deactivation. This thesis is structured as follows. Chapter 2 describes the composition and function of three-way and diesel catalysts used in automotives. In Chapter 3 the causes leading to deactivation of automotive catalysts, i.e. chemical, thermal and mechanical deactivation, are reviewed. Possible methods for regenerating
17

automotive catalysts are also discussed. Chapter 4 presents the experimental part of the work. The catalyst samples, ageing procedures, and characterization techniques are described. The research results of the work are presented in Chapter 5 and further integrated and evaluated in Chapter 6. Chapter 6 also discusses suggestions for further studies. Finally, the results are summarized and conclusions are drawn in Chapter 7.

18

2
2.1

Automotive exhaust gas catalysis

General

Automotive exhaust catalysts were introduced in the 1970s. These early converters were used only to oxidize HC and CO emissions (Heck & Farrauto 2001). The first automotive catalysts to convert HC, CO, and NOx were dual-bed systems in which NOx was reduced in the first bed. Secondary air was injected between the beds, enabling oxidation of HC and CO in the second bed. The stricter NOx standards in the 1980s led to development of three-way catalysts that simultaneously catalyze three types of reactions: oxidation of CO and HC and reduction of NOx. (Gandhi et al. 2003) Modern automotive catalytic converters are very effective in reducing gaseous pollutants in exhaust gas to levels that meet emissions standards. A converter (Figure 1) typically consists of a metallic or ceramic monolithic substrate that provides a high geometric surface area with a low pressure drop. The monolithic substrate is coated with a high-surface-area carrier material such as aluminium oxide (Al2O3). In order to maintain the high surface area with varying temperatures and gas compositions, the carrier is stabilized with lanthanum (La), barium (Ba) or silicon (Si) oxides, for instance. Precious metals, i.e. platinum (Pt), rhodium (Rh), and palladium (Pd), are added to the catalyst in varying compositions to provide catalytic reduction and oxidation of pollutants to more harmless compounds. (Heck & Farrauto 2001, Wallington et al. 2006, Seymor & O'Farrelly 2005) Rare earth metal oxides are used to stabilize the precious metals against sintering. Different oxides like cerium (Ce) and zirconium (Zr) oxides are used to improve catalytic efficiency by storing oxygen in lean conditions, i.e. an excess of oxygen, and releasing it when the engines operating conditions change to a low oxygen phase. (Farrauto & Heck 1999, Wallington et al. 2006)

19

Fig. 1. Typical structure of a metallic monolithic catalyst.

An exhaust gas after-treatment system consists of a catalytic converter and an electronically controlled air-fuel management system. Air intake and fuel injection are controlled to provide an appropriate ratio between oxygen and fuel (air-to-fuel ratio, A/F). The oxygen content of the exhaust gas is measured by an oxygen or lambda sensor placed immediately before the catalyst in the exhaust manifold. (Farrauto & Heck 1999) 2.2 Three-way catalysis

Automotive exhaust gases formed in gasoline engines normally contain carbon monoxide (CO), nitrogen oxides (NOX) and hydrocarbons (HCs). The typical composition of gasoline engine exhaust gases is presented in Table 1. A three-way catalyst is used to simultaneously remove these three major pollutants from a gasoline engines exhaust gases (Heck & Farrauto 2001). Under lean conditions the catalytic oxidation reactions of CO and hydrocarbons are favoured, whereas reduction of NOX takes place under rich conditions (Lox & Engler 1997).

20

Table 1. Composition of gasoline engine exhaust gas (Heck & Farrauto 2001).
CO (vol-%) 0.5 HC (ppm) 350 NOx (ppm) 900 H2 (vol-%) 0.17 H2O (vol-%) 10 CO2 (vol-%) 10 O2 (vol-%) 0.5 Balance N2

In ideal burning of gasoline, the following simplified reaction takes place:

gasoline + O2 (air ) CO2 + H 2 O + heat

(1)

Combustion in a real engine is incomplete, leading to formation of a number of combustion products. An automotive catalyst promotes reactions between these compounds according the following equations (Heck & Farrauto 2001): Oxidation of HC:
n n C y H n + ( y + )O2 yCO2 + H 2 O 4 2

(2)

Oxidation of CO:
1 CO + O2 CO2 2 CO + H 2 O CO2 + H 2

(3) (4)

Reduction of NOx:
1 N 2 + CO2 2 1 NO / NO2 + H 2 N 2 + H 2 O 2 n n n (2 + ) NO / NO2 + C y H n (1 + ) N 2 + yCO2 + H 2 O 2 2 4 NO / NO2 + CO

(5) (6) (7)

In addition to the oxidation and reduction reactions mentioned above, other significant reactions also take place in a TWC. These reactions include water-gas shift and steam reforming reactions (Kapar et al. 2003):
CO + H 2 O CO2 + H 2 HC + H 2 O CO2 + H 2

(4) (8)

21

2.3

Diesel catalysis

Unlike a spark-ignited engine, a diesel engine employs a compression-ignited process in which the fuel is injected into a highly compressed charge of air. Therefore, oxygen is always in excess in a diesel engine. (Zelenka et al. 1990, Zelenka et al. 1996, Twigg 2006, Frost & Smedler 1995) Furthermore, the requirements for diesel exhaust gas after-treatment are different from those of gasoline vehicles due to the unique demands of emissions standards, vehicle application, and high durability (Clerc 1996). The primary target in diesel emission control is reduction of particulates and NOx (Clerc 1996, Twigg 2003). Diesel emissions include all three phases of matter. Solid emissions consist of dry carbon and soot. The liquid phase includes hydrocarbons derived from unburned fuel and engine lubricating oil (collectively Soluble Organic Fraction, SOF), as well as liquid sulphates. (Zelenka et al. 1996) The size of a diesel exhaust gas particle is between 0.01 and 1 m in diameter. Especially the ultra(<100 nm) and micro-scale (<50 nm) particles are potentially hazardous to health because they are able to enter deep into the respiratory track and possibly even into cell membranes. (Votsmeier et al. 2005) Gaseous emissions contain CO and HC derived primarily from incompletely burned fuel, NOx and sulphur (Zelenka et al. 1996). The typical composition of diesel engine exhaust gas is presented in Table 2. There are three main after-treatment approaches to reducing harmful components in diesel exhaust gas: oxidation catalysis, NOx control, and particulate filtration (Twigg 2006).
Table 2. Composition of diesel engine exhaust gas (Neste Oil Oyj 2006, Roduit et al. 1998).
CO (vol-%) 490 HC (ppm) 30 NOx (ppm) 1135 H2O (vol-%) 4-5 CO2 (vol-%) 7.1 O2 (vol-%) 10.5 SO2 (ppm) 30 Particles (wt-%) <0.02 Bal. N2

2.3.1 Oxidation catalysis Diesel oxidation catalysts (DOC) are used to oxidate HC, CO, aldehydes, SOF, and odorous compounds (Zelenka et al. 1990, Zelenka et al. 1996). The temperature of diesel exhaust gas is relatively low, especially during low-speed driving. This results in the challenge of achieving and maintaining sufficient catalytic performance at low temperatures. Typical DOCs contain Pt in a highly
22

dispersed form. Zeolites are used to adsorb HCs, thus preventing them from inhibiting active Pt sites. (Twigg 2006) Some zeolite catalysts, such as platinum-containing ZSM-5, have demonstrated their effectiveness in reducing NOx with hydrocarbons in diesel exhaust gas with a high oxygen content (Guo et al. 1995). The temperature of diesel exhaust gas is normally relatively low, leading easily to a decrease in the operating temperature of a catalyst. Therefore, addition of a HC adsorbent is needed. Zeolites have the ability to adsorb HCs, which also improves NOx conversion with increasing temperature. (Shinjoh 2006, Farrauto & Voss 1996) With zeolite-containing catalysts, HC removal is effective over the entire exhaust temperature range with very low platinum loading. At low catalyst temperatures (below 200C), removal of gas phase HCs is attributed exclusively to zeolites. However, at higher temperatures (300C and above), HC conversion depends on how the platinum is functioning in the catalyst. (Farrauto & Voss 1996) The desired main oxidation reactions in diesel exhaust catalysis are (Heck & Farrauto 2001): Oxidation of SOF:
SOF + O2 CO2 + H 2 O

(9)

Oxidation of HC:
n n C y H n + (1 + )O2 yCO2 + H 2 O 4 2

(2)

Oxidation of CO:
1 CO + O2 CO2 2

(3)

Depending on the sulphur content of the fuel, the exhaust gas contains a varying amount of sulphur oxides (SOx). Sulphuric acid (H2SO4) is formed in the reaction between sulphur trioxide (SO3) and water. Therefore, oxidation of SO2 to SO3 should be minimized in the oxidation catalyst (Heck & Farrauto 2001): 1 SO2 + O2 SO3 (10) 2

23

2.3.2 NOx reduction An improved combustion process and exhaust gas recirculation (EGR) are used to reduce NOx content in diesel exhaust gas (Zelenka et al. 1996). Decomposition of NO to its elements is thermodynamically favoured between ambient and relatively high temperatures (Goralski & Schneider 2002):
NO U 1 1 N 2 + O2 2 2

(11)

However, under lean conditions the catalytically active metal surface becomes covered by strongly adsorbed oxygen, thus preventing NO adsorption (Twigg 2006). Formation of NO2 becomes significant at moderate temperatures and with excess oxygen (Goralski & Schneider 2002):
1 NO + O2 U NO2 2

(12)

In order to comply with emission standards under lean conditions, NOx reduction catalysis, i.e. selective catalytic reduction (SCR), or NOx trapping is needed in exhaust gas after-treatment (Twigg 2006). In SCR, a reducing agent such as alcohol, hydrocarbon, ammonia (NH3) or urea ((NH2)2CO) is required (Lox & Engler 1997, Zelenka et al. 1996). High NOx conversion rates can be achieved by adding the reducing agent to the exhaust gas before the catalyst (Zelenka et al. 1996). Using HC as a reducing agent involves an advantage compared with other reductants because for instance, the fuel itself can be employed as a source of the reducing agent. The following reactions of NO with HC or alcohol are possible (Lox & Engler 1997):

(6m + 2) NO + 2 Cm H 2 m + 2 U (3m + 1) N 2 + 2mCO2 + (2m + 2) H 2 O 6m NO + 2 Cm H 2 m +1OH U 3m N 2 + 2mCO2 + (2m + 2) H 2 O

(13) (14)

Besides HC and alcohol, NH3 and (NH2)2CO can be used as reducing agents in SCR (Koebel et al. 2000, Madia et al. 2002). (NH2)2CO is a compound that forms NH3 in the reaction with water (Koebel et al. 2000):
( NH 2 ) 2 CO + H 2 O 2 NH 3 + CO2

(15)

The main reaction of NO with NH3 and oxygen is (Koebel et al. 2000):
4 NH 3 + 4 NO + O2 4 N 2 + 6 H 2 O

(16)

24

With equimolar amounts of NO and NO2, the reaction rate is faster than the rate of reaction (16) (Koebel et al. 2000):
4 NH 3 + 2 NO + 2 NO2 4 N 2 + 6 H 2 O

(17)

In contrast, the reaction with NH3 and NO2 (18) is slower than reactions (16) and (17) (Koebel et al. 2000):
7 4 NH 3 + 3 NO2 N 2 + 6 H 2 O 2

(18)

Formation of N2O is possible at temperatures above 400C (Koebel et al. 2000):

4 NH 3 + 4 NO + 3 O2 4 N 2 O + 6 H 2 O

(19)

At very high temperatures, an undesired reaction in which NH3 is oxidized to NO may take place (Koebel et al. 2000):
4 NH 3 + 5O2 4 NO + 6 H 2 O

(20)

In the trapping method, NOx is storaged as NO3 during the lean driving period. When the trapping material is saturated, the stored NOx is released by changing the exhaust gas composition to be slightly reducing for a short period. Then, NOx is reduced to N2 over a Rh-containing component. (Twigg 2006) 2.3.3 Particulate filtration Soot is present in exhaust gas as an aerosol with a concentration that is too low to be burned in a catalytic converter. Furthermore, the residence time of soot in the catalyst is not long enough for oxidation. (van Diepen et al. 1999) Recent advances in the combustion processes of modern passenger car diesel engines have resulted in a reduction of PM emissions. Despite that, elimination of soot by filtration is still needed to prevent possible health effects of diesel PM. (Twigg 2006) Therefore, soot is collected in particulate filters from which it is finally removed (Koltsakis & Stamatelos 1997, van Diepen et al. 1999). A very commonly used filter structure, a wall-flow monolith, is presented in Figure 2.

25

Fig. 2. A wall-flow monolith particulate filter (Bissett 1984).

Trapping material is based on ceramic or metallic substrates that allow up to 95% capture of dry soot. To maintain engine efficiency, the filtering surface needs to be sequentially regenerated. Regeneration methods can be divided into two categories: active and passive processes. In active regeneration the trap is cleaned with a fuel burner or an electrical heater. Passive regeneration methods include catalytic fuel additives and catalytic coating before or inside the trap. (Zelenka et al. 1996, Stamatelos 1997)

26

3
3.1

Catalyst deactivation
Overview

Deactivation of automotive catalysts is a problem that has not been totally solved with current catalyst technology. Catalyst deactivation is a complicated phenomenon due to the several possible factors causing it. Catalysts may loose their activity as a result of a number of unwanted chemical and physical changes. (Butt & Petersen 1988, Forzatti & Lietti 1999, Bartholomew 2001, Moulijn et al. 2001, Richardson 1989) The types of deactivation phenomena can be divided into three categories: chemical, thermal and mechanical ageing (Bartholomew 2001, Bartholomew 2004). Table 3 shows one possible way to categorize these deactivation types and the related mechanisms of deactivation.
Table 3. Types and mechanisms of catalyst deactivation and their descriptions (Bartholomew 2004).
Deactivation type Chemical Deactivation mechanism Poisoning Description of the mechanism Strong chemisorption of compounds on the catalytic surface, blocking of the active reaction sites. Vaporization Vapour-solid and solid-solid reactions Thermal Thermal degradation Reaction of compounds with the catalyst phase to produce volatile compounds. Reaction of vapour, support, or promoter with the catalytic phase to produce an inactive phase. Thermally induced decrease in the catalytic surface, support surface area, and active phase-support reactions. Mechanical Fouling Attrition/crushing Physical deposition of species from the gas phase onto the catalytic surface and pores. Loss of catalytic material due to abrasion of the internal surface area caused by mechanical crushing of the catalyst.

High temperatures and temperature gradients, the presence of impurities, and fluctuating gas phase compositions and flow rates in a catalytic converter increase the possibility of deactivation (Forzatti & Lietti 1999, Bartholomew 2001, Moulijn et al. 2001). To give an example of different forms of deactivation,
27

Figure 3 presents deactivation of a diesel catalyst in different temperature ranges. Mechanical deactivation, which takes place for instance in the form of physical breakage or crushing, is also a relevant deactivation phenomenon (Koltsakis & Stamatelos 1997). Deactivation can be induced by one cause or a combination of several deactivating factors. In any case, the net effect is a reduction in catalytically active sites. (Forzatti & Lietti 1999)
T (C) 1000 Catalyst temperature range 800 Washcoat phase changes 600 Reaction of Zn and P with washcoat Exhaust gas temperature range Operation range

400 Formation and chemisorption of SO3 200 Adsorption of SO2 Adsorption of H2O and particulates 0

Fig. 3. Deactivation of diesel catalysts in different temperature ranges (adopted and modified from Lox & Engler 1997).

Deactivation detected in present-day catalytic converters during normal vehicle operation is typically a result of chemical and thermal mechanisms, while fouling and mechanical factors are more rarely the main reason for deactivation. (Koltsakis & Stamatelos 1997, Taylor 1984) This thesis focuses on the mechanisms of chemical deactivation, i.e. poisoning. Once an exhaust gas catalyst has become deactivated, regenerating it is a very difficult task. Some methods for restoring its catalytic activity have been studied. However, the most viable method for prolonging the life of a catalyst is to prevent deactivation. This can be done by decreasing the amount of poisons in fuel and lubricating oils to a more acceptable level. Another approach is to develop catalyst components that are more resistant to poisoning and thermal ageing.
28

3.2

Chemical deactivation

Exposure to catalyst poisons aggravates the loss of activity in automotive exhaust gas catalysts. Deactivation by poisoning is often related to high operating temperatures and sudden temperature changes. Poisons contaminate the washcoat and precious metals and reduce the catalytically active surface by blocking active sites (Butt & Petersen 1988, Angelidis & Sklavounos 1995). The activity of an automotive catalyst declines gradually over the course of time as the harmful components of fuels, lubricants and other impurities accumulate on the catalysts surface (Koltsakis & Stamatelos 1997). Poisoning is defined as a loss of catalytic activity caused by chemisorption of impurities on the active sites of the catalyst. Poisons are substances whose interaction with active sites is strong and often irreversible. (Butt & Petersen 1988, Forzatti & Lietti 1999) Catalyst poisons can also be classified as selective or non-selective, depending on whether the poisoning is universal or takes place only at certain sites or crystal faces or affects only certain adsorbed species. Poisoning by a certain compound can also be a reaction-specific phenomenon. In such a case, the poison is selective in one reaction, but non-selective in another. (Butt & Petersen 1988) Poisoning of an exhaust gas catalyst caused by accumulation of impurities on the active sites is typically a slow and irreversible phenomenon. Poisons accumulated on the catalyst surface block the access of reactants to the active sites. (Butt & Petersen 1988) Catalyst poisoning is possible even with a low level of impurities (Forzatti & Lietti 1999). S, P, Zn, Ca, and Mg are typical oil- and fuel-derived contaminants found in used catalytic converters. These contaminants have been widely studied (Williamson et al. 1984, Williamson et al. 1985, Inoue et al. 1992, Liu & Park 1993, Angelidis & Sklavounos 1995, Culley et al. 1996, Rokosz et al. 2001, Uy et al. 2003, Cabello Galisteo et al. 2004, Fernndes Ruiz et al. 2002). In the first generation of catalytic converters, lead (Pb) was the major cause of poisoning (Williamson et al. 1979a, Williamson et al. 1979b, Monroe 1980). Nowadays, the use of unleaded or low-lead gasoline has diminished the role of lead as a catalyst poison. However, it still has an impact on the deactivation of TWCs (Fernndes Ruiz et al. 2002, Larese et al. 2005, Lpez Granados et al. 2005). Developing and choosing catalyst materials appropriate for current applications is a feasible way to prevent thermal and chemical deactivation. The durability of the catalyst can also be improved by placing the active components
29

in separate washcoat layers. (Laurikko 1994, Laurikko 1995) Furthermore, precious metals used to control exhaust gas emissions have different levels of resistances against poisoning. For instance, it has been shown that Pt and Rh are less sensitive than Pd to poisoning by sulphur and lead. (Taylor 1993, Lox & Engler 1997) Despite the numerous studies carried out with poisoned catalysts, the individual and combined effects of poisons as well as the influence of simultaneous poisoning and thermal deactivation are still not sufficiently known. Especially knowledge about effects related to deactivation of zeolites is still lacking. It is known that deactivation of zeolite-containing catalysts can occur at both low and high temperatures, especially in moist conditions (van Kooten et al. 2000). 3.2.1 Poisoning by phosphates Zinc dialkyldithiophosphate (ZDDP) is an antiwear, antioxidant, and corrosioninhibiting additive that is used in engine lubricating oils, hydraulic oils, and other lubricants. ZDDP is produced by first letting an alkyl or aryl alcohol react with phosphorous pentasulphide (P2S5), and then neutralizing the resultant acid with zinc oxide. The frictional heat formed by two steel surfaces rubbing together in the engine causes ZDDP to react chemically with the surfaces. As a result, very strong wear-resistant lubricating films are formed. The additive is also a secondary oxidation inhibitor that decomposes hydroperoxides in the branching step of hydrocarbon oxidation. In addition, ZDDP is an effective bearing corrosion inhibitor. (Dickey 2005) In automotive applications ZDDP is a common source of typical catalyst poisons, phosphorus and zinc (Kumar et al. 2004). Table 4 shows typical characteristics of ZDDP. Caracciolo & Spearot (1976) have reported that catalytic activity decreases linearly as the amount of phosphorus detected in the catalyst increases. The phosphorus content of modern lubricating oils is typically 0.08 0.15 wt-% (Bunting et al. 2004). According to Kalantar Neyestanaki et al. (2004), a possible way to lengthen catalyst life is to use lubricants with a lower P content.

30

Table 4. Typical characteristics of zinc dialkyldithiophosphate (Research Institute of Petroleum Industry 2002).
Characteristics Density (g/ml) at 15.5C Viscosity (c.st) at 100C Sulphur content (wt-%) Phosphorus content (wt-%) Zinc content (wt-%) pH Typical value 1.09 10.3 19 10 11 56

The chemical composition in which phosphorus can be found in catalytic converters depends on many factors, such as the chemical composition of oil and oil additives, the age of the oil used, and conditions in the motor (Xu et al. 2004). Phosphorus contamination can be observed on the surface of catalytic converters as several different compounds (Carduner et al. 1988, Xu et al. 2004). Firstly, as illustrated in Figure 4, phosphorus can form a phosphate-containing overlayer that is a mixture of Zn, Ca, and Mg phosphates on the washcoat (Liu & Park 1993, Rokosz et al. 2001). These compounds are very stable on the catalytic surface (Andersson et al. 2007). Low exhaust temperatures are favourable for the formation of zinc pyrophosphate (Zn2P2O7) (Williamson et al. 1985). According to Liu and Park (1993), formation of a glassy, amorphous phase of Pb, Zn, and Ca phosphates on the catalyst surface has been observed. Ca and Mg have also been reported to prevent phosphorus poisoning. Ueda et al. (1994) demonstrated that phosphorus poisoning of the catalyst and the oxygen sensor can be restrained by using Ca and Mg sulphonates as oil additives. Furthermore, Williamson et al. (1980) reported that fuel P (derived from cresyl diphenyl phosphate) can neutralize Pb deactivation by the formation of lead phosphates.

31

Fig. 4. Scanning electron microscope (SEM) image of a layer containing Zn, Ca, and Mg phosphates on a catalytic surface, magnified 5000 times (Rokosz et al. 2001).

Secondly, phosphate compounds can also be formed directly with washcoat components. In that case, the resulting compounds are, for instance, aluminium and cerium phosphates (Rokosz et al. 2001, Uy et al. 2003, Cabello Galisteo et al. 2004). In the case of aged diesel catalytic converters, the adsorbed phosphorus species have also been identified as phosphates (Beckmann et al. 1992, Bunting et al. 2005). For instance, the interaction between phosphorus and Al2O3, resulting in the formation of AlPO4, is expected to be partially responsible for deactivation observed in aged diesel catalysts (Cabello Galisteo et al. 2004). Furthermore, formation of CePO4 is also responsible for deterioration of the oxygen storage and release properties of cerium oxide (Battistoni et al. 1999, Lpez Granados et al. 2006, Larese et al. 2003, Larese et al. 2004a, Larese et al. 2004b). Phosphorus is usually concentrated in the forward-most section of monolithic converters (Culley et al. 1996, Beck et al. 1997a, Beck et al. 1997b, Angove & Cant 2000, Webb et al. 2003). A decrease in catalytic activity and changes in characteristics, such as a loss of surface area, in the front section of a converter have been associated with extensive phosphorus deposition. (Beck et al. 1997, Angove & Cant 2000) Joy et al. (1985) have shown that as low an amount as 0.4 wt-% of phosphorus on the catalyst surface is enough to significantly reduce catalytic activity. In their study, no crystalline phosphorus compounds were detected. Instead, the presence of amorphous zinc phosphorus compounds was observed in SEM analyses.
32

3.2.2 Poisoning by sulphur compounds Sulphur poisoning is a complex phenomenon that can involve significant changes in the structural, morphological, and electronic properties of a catalyst (Rodriguez & Hrbek 1999). Gasoline and diesel contain sulphur in small amounts, but the cumulative effect of poisoning during the catalysts life can be substantial. In the case of TWCs, fast poisoning by sulphur can be to some extent reversible and the poisoned catalyst can be regenerated. (Yu & Shaw 1998) The role of sulphur poisoning is especially substantial in diesel catalysts due to the comparatively low operating temperatures involved. Despite the reduction in the sulphur content of fuel that has taken place during the past decades, sulphur is still present in diesel oil and gasoline in small amounts (Yu & Shaw 1998). Sulphur has a negative effect on the activity and oxygen storage capacity (OSC) of the catalyst (Boaro et al. 2001, Yu & Shaw 1998). The presence of sulphur can cause formation of new inactive compounds on the catalysts surface. Furthermore, morphological changes in the catalyst are also possible. (Yu & Shaw 1998) Depending on the temperature and partial pressure of oxygen, sulphur can be present in an exhaust gas catalyst in the form of sulphates, sulphides or oxysulphides (Karjalainen et al. 2005). In the fuel combustion process that takes place in the engine, sulphur oxidizes to SO2 and SO3, which are then adsorbed by the precious metal sites on the catalysts surface at temperatures below 300C. In the reaction with aluminium oxides, aluminium sulphates are formed. These compounds deactivate the catalyst by reducing the active surface. Sulphur poisoning is a severe problem at low temperatures, whereas at temperatures above 1000C adsorption of sulphur species is almost negligible. In addition, the effect of sulphur poisoning on TWCs has been observed to be more significant in oxidizing than in reducing conditions. (Butt & Petersen 1988, Heck & Farrauto 1996) Under reducing conditions sulphur forms H2S, which poisons metal surfaces and negatively affects oxidation of HCs (Rabinowitz et al. 2001). In the case of NOx storage, Fridell et al. (2001) have reported that sulphur deactivation is more severe when SO2 is added during a rich period than during the lean phase. Again, very stable sulphates, which are unattacked by reductants even at 1000C, can be formed especially in the absence of water (Mahzoul et al. 2001).

33

3.2.3 Poisoning by other compounds Automotive catalysts can also be deactivated by compounds commonly used in the structure of an engine. For instance, iron (Fe) is a poison for platinum group metals (Kalantar Neyestanaki et al. 2004). Iron detected on the catalytic surface is often assumed to originate from the corrosion of metal components in the engine (Cabello Galisteo et al. 2004). Other similar impurities may be copper (Cu), nickel (Ni), and chromium (Cr). Chromium and nickel are added to the construction materials in order to improve the thermal stability of engines. Copper originates from engine bearings, for example. (Angelidis & Sklakovounos 1995) 3.3 Thermal deactivation

High temperatures and uncontrollable temperature changes cause sintering, formation of new compounds, phase changes in existing compounds, and volatilization of components, leading to a loss of active surface area of the catalytic materials and the washcoat (Richardson 1989, Mooney 2005). The operating temperature of light-duty diesel catalysts is normally relatively moderate (below 500C), and therefore the contribution of thermal ageing to deactivation of these catalysts is of lesser importance compared with TWCs. Depending on the load, exhaust gas temperatures may vary from 100C up to 700C. Therefore, the influence of thermal ageing cannot be completely excluded. (Cabello Galisteo et al. 2004, Koltsakis & Stamatelos 1997) Sintering is a thermally activated process in which the surface area is decreased by structural modification of the catalyst. As a result of these modifications, active catalytic materials and also the washcoat can be affected. (Butt & Petersen 1988) Two different models have been proposed for the sintering of metal crystallites: the atomic migration model and the crystallite migration model. In atomic migration atoms migrate from one crystallite to another via the surface or the gas phase, causing an increase in crystallite size and thereby a decrease in surface area. In crystallite migration sintering occurs due to crystallite migration over the surface, resulting in collisions and coalescence of the crystallites. (Forzatti & Lietti 1999, Bartholomew 2001) Thermal degradation of washcoat materials also increases the risk of deactivation of the catalyst. Phase transformations and sintering of the washcoat decrease the catalytically active

34

surface area. In addition, changes in the catalyst structure can cause encapsulation of active metal particles. (Butt & Petersen 1988, Lassi 2003) 3.4 Mechanical deactivation

A catalytic converter can be damaged by a strong mechanical impact. In such a case, the monolith is crushed and as a result, the back pressure of the exhaust gas is increased. If the damage is not noticed in time, local overheating can cause thermal deactivation and finally melting of the honeycomb. Metallic catalysts can endure mechanical deactivation better than ceramic ones. (Laurikko 1995) Mechanical failure of the catalyst may also be a result of high temperatures. Two types of factors can cause high catalyst temperatures. The first one is engine operating conditions. Secondly, as the result of misfiring, non-combusted fuel-air mixtures undergo catalytic combustion, which can produce very strong exothermic reactions within the catalyst. At about 1200C, the metallic foil starts to soften and probably shrink. If the catalyst temperature exceeds 1400C, the foil will melt, after which it is destroyed. (Mooney 2005) 3.5 Catalyst regeneration

It is usually easier to prevent catalyst deactivation than to restore the original activity of the catalyst (Bartholomew 2001). However, some studies have been carried out to investigate the possibilities of regenerating activity that has decreased due to thermal ageing and poisoning. According to van Kooten et al. (2002), catalytic activity can be partially restored by heating the catalyst under oxygen-rich conditions, which removes some of the contaminant species (coke, HCs, nitrogen). However, van Kooten et al. (2002) reported that depositions of P, Zn, Ca, and S were very hard to remove with this method, and therefore they are associated with irreversible deactivation. Reactivation of sulphated oxidation catalysts has been performed by Cabello Galisteo et al. (2007), who decomposed aluminium sulphate in a deactivated catalyst with a reductive treatment. Furthermore, amorphous zinc pyrophosphate has been removed from the catalytic surface at high temperatures and in reducing conditions. The method is not viable from a practical point of view because as a result of sintering, the net activity of the catalyst is not restored. (Williamson et al. 1984)
35

Washing a poisoned catalyst with organic acids has also been tested. Rokosz et al. (2001) and Christou et al. (2006) have shown that it is possible to efficiently remove phosphorus accumulated in the catalyst using oxalic acid. With this method, catalytic activity was restored to a level corresponding to only thermal deactivation (Rokosz et al. 2001). Similar results have been reported by Angelidis & Sklavounos (1995), who treated catalysts containing Ca, P, S, Pb, Fe, and Zn compounds with acetic acid. According to Christou et al. (2006), by washing a catalyst with EDTA (ethylene diamine tetraacetic acid), depositions of Pb, Zn, Ca, Mn, Fe, Cu, Ni, and S, but not P, could be removed. Washing a poisoned DOC with a dilute solution of citric acid has been demonstrated to remove a large amount of P and most of the S accumulated in the catalyst (Cabello Galisteo et al. 2005a). Birgersson et al. (2006) showed in their study that combined thermal and liquid chlorine treatments can increase the number of catalytically active noble metal sites on the washcoat surface. Furthermore, P and S could be partially removed, which led to regaining of micropores. As a result, catalytic activity was restored. Catalyst regeneration is also possible in the case of thermal deactivation. Catalytic activity lost because of chemical interaction between catalytic materials can sometimes be restored by using suitable reaction conditions. For instance, oxidation of Rh/Al2O3 at 800 C caused diffusion of Rh into the aluminium oxide, resulting in a decrease in CO oxidation activity, which was reported by Wong & McCabe (1989). The activity could be restored with a reductive treatment (H2 at 800C) followed by reoxidation at 800C (Wong & McCabe 1989). GonzalesVelasco et al. (2000) have presented results that are congruent with those of Wong & McCabe. Birgersson et al. (2004) and Anderson et al. (2006) demonstrated that a TWC that has been deactivated in terms of activity and metal dispersion can be regenerated by adding chlorine to an oxygen stream. This oxychlorination treatment also seems to be a potential procedure for regenerating DOCs if thermal ageing and sintering are the main causes of deactivation (Cabello Galisteo et al. 2005b) In general, it should be pointed out that the interactions between precious metals and washcoat materials differ from each other. Therefore, the catalytic material used significantly affects the possibility of regenerating the catalyst. (Wong & McCabe 1989) Finally, in cases where regeneration of the catalyst is not possible, methods for recovering platinum group metals (PGMs) from spent TWCs have been developed (Angelidis 2001).
36

4
4.1

Experimental
Catalysts

This work consists of two test series and two different sets of catalyst samples: powders containing Rh (2.0 wt-%) or Pt (1.5 wt-%), which represent components used in commercial exhaust gas catalysts. In the latter test series washcoats without a precious metal were also studied. The catalysts were donated by Ecocat Oy. Figure 5 shows examples of the powdery components used in this study. The samples are listed in Table 5. The samples were studied when fresh and after the ageing and phosphorus poisoning procedures described in Chapter 4.2.

CeO2

Pt/ZrO2

Pt/CeO2

Al2O3

Pt/Al2O3

Fig. 5. Examples of powdery components used in automotive catalysts.

In addition to the samples presented in Table 5, three vehicle-aged metallic catalyst monoliths were used as industrial reference catalysts for the poisoning: a Rh/Pd/Al2O3/La2O3/OSC catalyst (aged for 100,000 km in a petrol-driven vehicle) for Rh-loaded samples, and Pt/Al2O3/La2O3/OSC and Pt/Al2O3/OSC catalysts (aged for 80,000 and 360,000 km in a diesel-driven vehicle, respectively) for Pt-loaded samples. The monoliths were divided into different test zones in the axial (front and rear) direction in order to evaluate the positiondependent role of poisons in catalytic activity. The role of chemical deactivation
37

is most significant in the forward-most section of monolithic converters. Therefore, the front zones of the commercial catalysts were used as reference material in this study.
Table 5. List and description of the powdery catalyst samples.
Catalyst sample Test series 1 Rh/Zr-CeO2 Rh/Ce-ZrO2 Rh/-Al2O3/OSC Test series 2 Al2O3 TiO2 CeO2 ZrO2 Zr-CeO2 Ce-ZrO2 ZSM-5 zeolite Pt/Al2O3 Pt/TiO2 Pt/ZSM-5 zeolite Pt/CeO2 Pt/ZrO2 Pt/Ce-ZrO2 Pt/Zr-CeO2 Zr-stabilized Ce oxide Ce-stabilized Zr oxide Si/Al2 29 Pt 1.4 wt-% Pt 1.5 wt-% Pt 0.85 wt-% Pt 1.5 wt-% Pt 1.5 wt-% Pt 1.5 wt-% in Ce-stabilized Zr oxide Pt 1.5 wt-% in Zr-stabilized Ce oxide Rh 2.0 wt-% in Zr-stabilized Ce oxide Rh 2.0 wt-% in Ce-stabilized Zr oxide Rh 2.0 wt-% in the OSC material, 0.5 wt-% in the entire washcoat Description

4.2

Phosphorus poisoning procedures

Two different methods were utilized for phosphorus ageing. In the first stage of the methodology development, the test series consisted of Rh-loaded catalyst powders representing components used in TWC applications. Phosphorus and as a special case, calcium, were added via impregnation. (Papers III) In the other poison treatment, the ageing procedure was further developed to better correspond to the extreme conditions in a catalytic converter during real driving (Papers IIVI). Phosphorus was added to a sample under continuous air flow and at a high temperature with the purpose of simulating poisoning in hydrothermal ageing conditions. Gas phase poisoning was carried out with a test series consisting of Pt-loaded catalyst powders that represented components used in diesel catalyst applications.
38

4.2.1 Poisoning by impregnation The catalysts were pre-treated in a H2 flow at 900C for 24 hours prior to poisoning, which was performed with an incipient wetness impregnation method. Aqueous salt solutions of (NH4)2HPO4 and Ca(NO3)24H2O were used as sources of phosphorus and calcium, respectively. The final concentrations of phosphorus and calcium in the catalyst powders were 1 wt-%. For XRD analysis the poison concentration was increased to 5 wt-% to obtain better detectability. (Paper I) After poisoning with one poison, the catalyst powders were calcined in air for 30 minutes at 100C, followed by another 30 minutes at 200C and, finally, for one hour at 500C. Subsequently, some of the catalysts were poisoned with the other poison (P+Ca or Ca+P) in order to study the combined effects of calcium and phosphorus. (Paper I) 4.2.2 Gaseous phase poisoning The ageing temperature of 700C was chosen to correspond to the peak temperature in diesel catalytic converters (Koltsakis & Stamatelos 1997). The phosphorus content detected in vehicle-aged diesel catalysts varies depending on the literature source and the method used for the detection. The phosphorus content detected in the front part of converters is normally less than 4 wt-% (Cant & Angove 1998). The amount of phosphorus fed into a catalyst sample during ageing was set to correspond to the typical phosphorus content detected in aged catalysts. The aim was to develop an ageing procedure that is safe and easy to use. Therefore, different equipment constructions and materials were tested during the development process. Tubular quartz, steel and ceramic reactor materials were tested. The steel and ceramic reactors were severely corroded by the repeated exposure to hydrothermal conditions and phosphorus addition. The quartz reactor was found to be resistant to corrosion even after several ageing treatments. Vertical and horizontal reactor positions were tested in the construction. Due to plugging caused by moisture in the sample in the vertical position, the horizontal reactor position was chosen. The sample was placed in the reactor between two pieces of inert quartz wool. A thermoelement was also installed inside the reactor for temperature indication. (Paper II) A schematic illustration of the gaseous phase poisoning equipment is presented in Figure 6.

39

Mass flow controller Air 250ml/min

Quartz reactor

Temperature indicator

Peristaltic pump

Sample 1.5g Furnace

Moisture collector

Aqueous salt solution of (NH4)2HPO4 0.01ml/min

700C / 4 hours

Fig. 6. Schematic illustration of the equipment used for gaseous phase poisoning studies.

In order to simulate phosphorus poisoning in lean hydrothermal conditions, an aqueous salt solution of (NH4)2HPO4 was fed into the quartz reactor with a peristaltic pump. It was demonstrated that less than 70% of the phosphorus adhered to the sample. In addition, the final phosphorus content in the samples varied depending on the catalyst. To ensure sufficient poisoning, the ageing time (4 hours) as well as the (NH4)2HPO4 content (0.13 mol/l) and feeding rate of the solution (0.1 ml/min) were set to correspond to a 6 wt-% target content of phosphorus in the sample. One aim of the studies for which this ageing procedure was developed was to compare the effects of phosphorus poisoning on samples containing different oxides. Therefore, ageing was carried out with all the samples according to the same procedure without changing the ageing time or the amount of phosphorus in the feed. Air (250 ml/min) was used as a carrier gas for the (NH4)2HPO4 feed. The air stream was controlled by a mass flow controller (Brooks 5850TR). After four hours of ageing the (NH4)2HPO4 feed was turned off and the sample was dried in the air stream at 500C for one hour. 4.3 Catalyst characterisation techniques

In this study, several techniques were used to obtain information about the characteristics of the catalysts. The poisoning-induced changes in the washcoat
40

and the poison compounds formed were studied using physisorption analyses, SEM, TEM, XRD, and FTIR-ATR. The poison content of the samples was determined by chemical analyses (ICP-OES) and XRF. Laboratory-scale activity measurements were applied to study catalytic activity. Thermodynamic calculations were used to obtain information about ageing conditions and phosphorus compounds formed during ageing. 4.3.1 Physisorption analyses To characterize the catalysts before and after ageing, physisorption measurements were carried out with a Micromeritics ASAP 2020 unit (Figure 7) and a Coulter Omnisorp 360CX unit. Specific surface areas (m2/g) were calculated according to the BET theory. The BJH theory was used to evaluate the total pore volume (ml/g) and average pore size (nm) of the samples. The internal surface areas were determined using the t-test method. The calculations were carried out by using N2 adsorption isotherms at 196C and assuming that the pores had a cylindrical shape. It should be mentioned that, for the calculations of the numerical values, certain approximations were made concerning pore shape, for instance. However, the pores may have different structures. Furthermore, they may also be interconnected with one another. (Webb & Orr 1997)

Fig. 7. Micromeritics ASAP 2020 apparatus used for physisorption measurements.

41

4.3.2 Scanning Electron Microscopy SEM-EDS was used to detect contaminant levels in the poisoned catalysts. Possible spatial differences in poison accumulation were also studied. The samples were cast in epoxy resin and polished and coated with carbon. The equipment used for the SEM studies was Jeol JSM-6400. 4.3.3 Transmission Electron Microscopy Transmission electron microscopy (TEM) was used to characterize the composition and microstructure of the samples and to determine changes in the washcoat and Pt particle sizes after the ageing treatments. The measurements were carried out with an analytical transmission electron microscope, JEM 2010 from Jeol, operated at 200 kV and equipped with a Noran Vantage energydispersive spectrometer. The TEM samples were prepared from the powder samples by crushing the samples between glass slides. The crushed powder was then dispersed onto a perforated carbon-covered Cu grid with ethanol. The TEM studies were carried out at Tampere University of Technology. 4.3.4 Chemical analyses The phosphorus content of the samples was determined with ICP-OES. Samples containing cerium were digested by heating them in a solution of (NH4)2SO4 solid salt, H2SO4, and HCl. The other samples were mixed with a solution of HNO3, HCl, and HF and digested by using microwaves. After digestion, the samples were diluted with water and analyzed with a GBC Integra XMP ICP-OES analyzer. In order to avoid a systematic error in the results, the ICP-OES backgrounds of the above-mentioned solutions were measured. 4.3.5 X-ray diffraction X-ray diffraction (XRD) was used to identify phosphorus compounds in the aged catalyst samples. The X-ray diffraction diagrams were recorded from the catalyst powders on a plastic sample holder with a Siemens D5000 or Philips XPERT diffractometer employing nickel-filtered Cu K radiation ( = 1.5406 , 40 kV, 30 mA) at 0.020 intervals in the range of 12 2 75 with 1 s count accumulation per step. Diffraction patterns were assigned using a PDF database
42

supplied by International Centre for Diffraction Data (PDF-2 Powder Diffraction File Database). 4.3.6 Fourier Transform Infrared-Attenuated Total Reflection FTIR-ATR studies were carried out to identify phosphorus compounds and to detect possible ageing-induced changes in the catalyst structure. The measurements were carried out with a Perkin Elmer Spectrum One FTIR-ATR. 4.3.7 X-ray Photoelectron Fluorescence X-ray photoelectron fluorescence was used to determine the phosphorus content of the poisoned catalyst samples. The XRF measurements were carried out with a Philips MagiX apparatus containing Super Q Analytical software. 4.3.8 Activity measurements Activity measurements were carried out by using laboratory-scale light-off experiments to study poisoning-induced changes between fresh and aged catalyst powder samples. Catalytic activities were determined in lean reaction conditions by using model reactions. The test gas mixtures are shown in Table 6. The gas mixture for Test series 1 (samples containing oxide and Rh) was used to simulate lean gasoline exhaust gas conditions. The gas mixture for Test series 2 (samples containing oxide and Pt and ZSM-5) simulated diesel exhaust gas conditions. Both gas mixtures were balanced with N2.
Table 6. Composition of the test gas mixtures for the activity measurements.
Gas component Test series 1 CO NO Test series 2 C3H6 NO O2 1000 ppm 1000 ppm 10 vol-% 800 ppm 1000 ppm Concentration

43

Before the measurements the samples in Test series 1 were reduced in a 500 ml/min hydrogen flow for 10 minutes at 500C, followed by 15 minutes at 550C. The samples in Test series 2 were oxidized in a 50 ml/min air flow for 20 minutes at 500C. The pre-treatments and measurements were carried out at atmospheric pressure in a tubular furnace using a quartz reactor. The weight of a powder sample was 0.25 g for oxides and 0.10 g for zeolites. The samples were mixed with quartz sand to prevent agglomeration of the catalyst, and therefore, to prevent diffusional effects in the catalyst bed. The samples were placed in the reactor tube with a support of quartz wool. Gas flow was controlled by mass flow controllers (Brooks 5850TR), and the total gas flow during the experiments was 1 l/min. The temperature of the catalyst was increased from room temperature up to 500C, with a linear heating rate of 1015C/min. The concentrations of the feed and product gases were measured as a function of temperature every 5 seconds with a FTIR gas analyzer (GasmetTM CR2000). Figure 8 illustrates the experimental setup. 1. Mass flow controllers 2. Thermocouple 3. Tubular furnace 4. Quartz reactor 5. Temperature controller 6. FT-IR Gasmet Analyzer

1 3 2 4

Fig. 8. Experimental setup for activity measurements, equipped with a GASMET Analyzer.

TM

Gas

44

4.3.9 Thermodynamic calculations Thermodynamic equilibrium calculations were carried out using commercial HSC Chemistry 5.11 software. The software can be used as a tool for calculating chemical equilibriums between pure components in ideal solutions and, in some cases, in non-ideal solutions. Equilibrium compositions are calculated with a solver that uses the Gibbs energy minimization method. The HSC Chemistry 5.11 software includes a thermochemical database that contains enthalpy, entropy and heat capacity values. Thermodynamic modelling does not provide any information on reaction kinetics, but physically possible phenomena can be estimated by using thermochemical calculations. Phase stability diagrams present the stability areas of condensed phases in ternary systems, either as a function of temperature or in isothermal conditions. These diagrams are useful for a quick estimation of the prevailing phases.

45

46

Poisoning-induced changes in the catalyst components

The research results and the changes caused by poisoning and hydrothermal ageing detected in the catalyst components are classified in this chapter into four sections. The sections are: loss in catalytic surface area, accumulation of phosphorus, loss of catalytic activity, and changes in the active metal particles and washcoat grain size. 5.1 Loss in catalytic surface area

As a result of poisoning and ageing, alterations in the specific surface areas of the catalyst samples were detected. The BET surface areas of all the phosphoruspoisoned catalyst powders were significantly diminished compared with the fresh samples. Figures 9 and 10 present the specific surface areas of the catalysts in Test series 2 (samples containing oxide and Pt and ZSM-5).
Surface area (m2/g) 400 Fresh H2O P+H2O 288

346

323 263

300 203 200 156 100 101 67 25 15 0 Pt/Al2O3 Pt/TiO2

188

93 61 6 2 Pt/CeO2 Pt/ZrO2 Pt/Zr-CeO2 Pt/Ce-ZrO2 Pt/ZSM-5 62 38 58 4128

Fig. 9. Specific surface areas of fresh, hydrothermally aged (H2O), and phosphoruspoisoned (P+H2O) catalyst samples with Pt.

47

Surface area (m2/g) 400 Fresh H2O P+H2O 401 330 294 260 217 200 166 145 79 68 2421 0 Al2O3 TiO2 CeO2 46 6 ZrO2 Zr-CeO2 Ce-ZrO2 ZSM-5 60 42 61 40 33 201

300

100

Fig. 10. Specific surface areas of fresh, hydrothermally aged (H2O), and phosphoruspoisoned (P+H2O) catalyst samples without Pt.

The effect of the ageing treatments was most significant for the surface areas of the ceria-based samples, i.e. CeO2, Pt/CeO2, Zr-CeO2, and Pt/Zr-CeO2 (Figures 9 and 10), and Rh/Zr-CeO2 (not shown). For instance, the specific surface area of the poisoned Pt/Zr-CeO2 was one fifth of the surface area of the fresh sample. Furthermore, for the samples in Test Series 1 (samples containing oxide and Rh), the surface area of the phosphorus-treated Rh/Zr-CeO2 was less than one tenth (12 m2/g) of the surface area of the fresh sample (191 m2/g). The respective values for Rh/Zr-CeO2 were 41 m2/g and 55 m2/g (Papers II and IV). The specific surface areas of Pt/ZrO2 and ZrO2 (Figures 9 and 10) were also strongly affected by ageing. As a result, the surface area of the phosphorus-poisoned Pt/ZrO2 sample as well as the hydrothermally aged and phosphorus-poisoned ZrO2 samples could not be determined. (Paper IV) In the case of zeolites, the BET results showed that the fresh ZSM-5 and Pt/ZSM-5 samples had rather high specific surface areas (Figures 9 and 10). The higher specific surface area of the fresh ZSM-5 may explain the larger phosphorus content (Figure 15 and 16) measured in this sample compared with the sample containing platinum. (Paper V) The effects of ageing were detected as changes in the average pore sizes and total pore volumes of the samples. The poisoning-induced changes in the
48

observed specific surface areas of the samples could be explained by decreases in the pore volumes and increases in the pore sizes detected in these samples. Figures 1112 illustrate the total pore volumes and Figures 1314 the average pore sizes of the catalysts in Test series 2 (samples containing oxide and Pt and ZSM-5).
Total pore volume (ml/g) 0.6 0.49 0.49 0.5 0.4 0.33 0.3 0.2 0.1 0.0 Pt/Al2O3 Pt/TiO2 Pt/CeO2 0.12 0.08 0.27 0.19 0.17 0.13 0.01 0.01 Pt/ZrO2 Pt/Zr-CeO2 Pt/Ce-ZrO2 Pt/ZSM-5 0.18 0.14 0.10 0.20 0.19 0.25 0.25 0.21

Fresh H2O P+H2O

0.13

Fig. 11. Total pore volumes of fresh, hydrothermally aged (H2O), and phosphoruspoisoned (P+H2O) catalyst samples with Pt.

Total pore volume (ml/g) 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Al2O3 TiO2 CeO2 ZrO2 Zr-CeO2 Ce-ZrO2 ZSM-5 0.11 0.10 0.49 0.50 0.41 0.29 0.19 0.17 0.13 0.01 0.19 0.17 0.22 0.25 0.25 0.19 0.14 0.19 Fresh H2O P+H2O

0.13

Fig. 12. Total pore volumes of fresh, hydrothermally aged (H2O), and phosphoruspoisoned (P+H2O) catalyst samples without Pt.

49

Average pore size (nm) 30 Fresh H2 O P+H2O 20 16 13 10 10 14

20

21

19 20 14 9 7 8 3 8 4 3 3 3 11

13

0 Pt/Al2O3 Pt/TiO2 Pt/CeO2 Pt/ZrO2 Pt/Zr-CeO2 Pt/Ce-ZrO2 Pt/ZSM-5

Fig. 13. Average pore sizes of fresh, hydrothermally aged (H2O), and phosphoruspoisoned (P+H2O) catalyst samples with Pt.

Average pore size (nm) 30 Fresh H2O P+H2O 19 19 20 15 12 12 10 9 10 11 7 3 1 0 Al2O3 TiO2 CeO2 ZrO2 Zr-CeO2 11 13 15

20 18

3 3 3

Ce-ZrO2

ZSM-5

Fig. 14. Average pore sizes of fresh, hydrothermally aged (H2O), and phosphoruspoisoned (P+H2O) catalyst samples without Pt.

It can be seen from Figures 1114 that the effect of phosphorus was more significant than the effect of hydrothermal treatment alone, and only a few exceptions to this were observed. The average pore size of the hydrothermally aged Ce-ZrO2 was higher than the value of the poisoned sample (Figure 14). The
50

smallest pore volume was, however, calculated for the poisoned sample, which was in agreement with the specific surface area determined for the Ce-ZrO2 samples. In the case of Al2O3 and TiO2, the average pore sizes of the hydrothermally aged and poisoned samples were equal. The effect of phosphorus was clearly observed in the pore volumes of the Al2O3 and Pt/Al2O3 samples, which did not decrease as a result of hydrothermal ageing. This phenomenon was not observed in the specific surfaces or average pore sizes of the Al2O3-based samples. After phosphorus poisoning, a decrease in the pore volumes of these samples was detected. (Paper IV) For ZSM-5 and Pt/ZSM-5 the total pore volumes of the fresh samples were not affected by hydrothermal ageing, as can be seen in Figures 11 and 12. In contrast, phosphorus poisoning caused a decrease in the pore volumes of both samples, which was congruent with the results from the BET analyses and activity measurements. Any further explanation of the changes in the specific surface areas or activity could not be found in the pore size distributions, giving the same average pore width for all six samples (3 nm) (Figures 13 and 14). (Paper V) 5.2 Accumulation of phosphorus

XRF, SEM-EDS and elemental analyses were used to detect phosphorus accumulation in the aged catalyst powders. Phosphorus compounds formed in the samples were identified with XRD and FTIR-ATR. In addition, thermodynamic calculations were used to support these methods. 5.2.1 Phosphorus content The XRF analyses showed that phosphorus was adsorbed almost quantitatively on the powdery samples during the incipient wetness impregnation method. The semi-quantitative amount of phosphorus measured by XRF for Test Series 1 (the Rh-loaded powder samples) was 1 wt-%, which was the target phosphorus content in this procedure. (Paper I) The amount of phosphorus adsorbed in the powdered catalyst samples during gas phase ageing varied depending on the catalyst composition. Figures 15 and 16 present the phosphorus contents of the samples treated with the gaseous phase poisoning method, determined by ICP-OES. The difference between the amounts of phosphorus detected in the samples can be explained by the dissimilar
51

tendencies of different materials to accumulate impurities. The very low specific surface areas of the fresh ZrO2 and Pt/ZrO2 are probably the reason for the small phosphorus content measured in these samples. However, no correlation between the characteristics of the fresh samples and the phosphorus content detected in the samples after ageing was observed in this study. Further studies are needed to explore the possible attributes that give a correlation between certain characteristics and the catalysts tendency to adsorb poisoning compounds. (Paper IV)
Phosphorus content (wt-%) 5.0 4.4 3.9 4.0

3.8

3.0 2.1 2.0 2.3

2.6

1.0 0.4 0.0 Pt/Al2O3 Pt/TiO2 Pt/CeO2 Pt/ZrO2 Pt/Zr-CeO2 Pt/Ce-ZrO2 Pt/ZSM-5

Fig. 15. Phosphorus contents of the samples with Pt, treated with the gaseous phase poisoning method.

52

Phosphorus content (wt-%) 5.0 4.0 3.0 2.0 1.0 0.0 Al2O3 TiO2 CeO2 ZrO2 Zr-CeO2 Ce-ZrO2 ZSM-5

3.5 3.0 2.4 2.3 1.9 2.2

0.4

Fig. 16. Phosphorus contents of the samples without Pt, treated with the gaseous phase poisoning method.

In contrast, in the impregnation ageing procedure, the final phosphorus content of the sample is equal to the amount of phosphorus impregnated into the sample. Therefore, the gas phase ageing procedure can be used as a tool for studying possible distinctions in the behaviour of different catalyst components in chemical ageing conditions. (Paper II) Poisoning can be considered a relevant deactivation mechanism for the TWC reference catalysts (Rh/Pd/Al2O3/La2O3/OSC catalyst aged for 100,000 km in a petrol-driven vehicle). According to the XRF analyses, the reference catalyst was poisoned by phosphorus, calcium and zinc. However, sulphur was not detected. The first 2 cm from the inlet of the catalyst comprised the most severely poisoned zone. The elemental analyses (XRF) showed that the catalyst contained 3.5 wt-% of phosphorus. Figure 17 presents the elemental analysis data from the inlet region of the TWC, including a back-scattered electron image and selected elemental distribution maps (X-ray maps) obtained by EDS. As can be seen in Figure 17, the poisons were accumulated on the top overlayer of the vehicle-aged catalyst. According to the chemical analyses, the contaminated layer contained phosphates, mainly Ca3(PO4)2. Furthermore, penetration of phosphorus and calcium into the upper areas of the washcoat was observed. (Paper II)

53

A)

B)

C)

D)

Fig. 17. Elemental analyses (X-ray maps) taken at the inlet of the vehicle-aged TWC catalyst; A) backscattered electron image showing the metal foil, washcoat, and contaminant overlayer; EDS elemental maps of B) Al, C) P, and D) Ca (Paper II).

XRF analyses showed that the phosphorus content of the diesel reference catalyst (Pt/Al2O3/La2O3/OSC aged for 80,000 km in a diesel-driven vehicle) was 0.4 wt-%. The catalyst was mainly poisoned by sulphur. Small concentrations of calcium and zinc were also detected. No contaminated overlayer was observed in the SEM-EDS studies (Figure 18). Instead, phosphorus had accumulated in the top overlayer of the catalyst. Due to the overlapping peaks of phosphorus and zirconium, quantitative EDS analyses were used to indicate that phosphorus was located only on the upper layers of the catalyst. In contrast, zirconium was detected throughout the catalyst. Unlike phosphorus, sulphur had penetrated evenly throughout the entire depth of the washcoat. (Paper II)

54

A)

B)

C)

D)

Fig. 18. Elemental analyses (X-ray maps) taken at the inlet of the vehicle-aged diesel catalyst; A) backscattered electron image showing the metal foil and contaminated washcoat; EDS elemental maps of B) Al, C) P, and D) S (Paper II).

It should be pointed out that the deactivation of the reference catalysts is the result of a combination of several ageing phenomena and not phosphorus poisoning alone. However, based on the activity measurements carried out with the vehicle-aged TWC reference (Lassi 2003), poisoning contributes to approximately 20% of the total deactivation, with phosphorus being the major contaminant. 5.2.2 Phosphorus compounds

XRD studies Phosphorus compounds formed with OSC components could be detected with XRD after both chemical ageing procedures. CePO4 was present in Rh/Zr-CeO2,
55

Rh/Ce-ZrO2 and the TWC reference samples (Rh/Pd/Al2O3/La2O3/OSC catalyst aged for 100,000 km in a petrol-driven vehicle) (Figure 19). Formation of CeAlO3 was also observed in the TWC reference due to the reactions of Al2O3 with the pure CeO2 in the catalyst's washcoat. (Paper I) According to the XRD studies, phosphorus had adhered to the ceria-based samples (CeO2, Zr-CeO2, Pt/CeO2, Pt/Zr-CeO2) as CePO4 (monazite-Ce). CePO4 was also found in Ce-ZrO2 but not in Pt/Ce-ZrO2. (Paper III)

300

d d d (iii d

Intensity (a.u.)

a
200

a a c

(ii

100

c b&c b&c c a c b&c Rh a b&c (i

0 13 20 30 40 50 60 7

2-Theta - Scale
Fig. 19. XRD diffractograms of (i) P-poisoned Rh/Al2O3/OSC, (ii) P-poisoned Rh/ZrCeO2, and (iii) P+Ca-poisoned Rh/Zr-CeO2. a = Ce(PO4), b = Zr-rich CeXZr1-XO2 mixed oxide, c = ZrO2 (monoclinic), d = Ce-rich CeXZr1-XO2 mixed oxide, and Rh = rhodium (Paper I).

The Zr-rich mixed oxides (Ce-ZrO2 and Pt/Ce-ZrO2) contained zirconium phosphate (Zr(P2O7)). However, zirconium phosphates were not detected in the samples containing pure zirconium oxide (ZrO2 and Pt/ZrO2). This may be due to the fact that the phosphorus content of these two samples was low and therefore could not be detected by XRD. (Paper III) The presence of aluminium phosphate (AlPO4) was expected in Rh/Al2O3/OSC, but due to the small peak intensities and the overlapping peaks of CePO4, its presence could not be confirmed (Paper I). In the case of Pt/Al2O3,
56

phosphorus appeared as aluminium phosphate AlPO4, as illustrated in Figure 20. By contrast, no AlPO4 or any other phosphorus-containing compounds could be observed in the Al2O3 sample. Phosphorus was detected in both TiO2 and Pt/TiO2 as titanium pyrophosphate (TiP2O7). (Paper III)

c
600

c
500

c d

c c iv)

Intensity (a.u.)

400

iii)
300

Pt a a

Pt

200

a ii)

100

i)
0 10 20 30 40 50 60 70

2-Theta - Scale
Fig. 20. XRD diffractograms of (i) fresh Pt/Al2O3, (ii) P-poisoned Pt/Al2O3, (iii) fresh Pt/CeO2, and (iv) P-poisoned Pt/CeO2. a = Al2O3, b = AlPO4, c = CeO2, d = CePO4, and Pt = platinum.

The presence of calcium phosphate Ca3(PO4)2 in the Test series 1 samples (the Rh-loaded powders) treated with both phosphorus and calcium (Ca+P or P+Ca) was also considered. However, it could not be verified due to the small intensities and overlapping peaks of CePO4. (Paper I) The diffraction peaks of the ZSM-5 structure and Pt were identified according to JCPDS File Cards nos. 440003 and 040802, respectively. No indication of phosphorus or phosphate formation was found in the patterns of the poisoned samples. (Paper V) Phosphorus was detected in the diesel reference catalyst (360,000 km) as phosphorous oxide (P4O6) and elemental phosphorus. However, no presence of phosphates was observed. One explanation for this could be that the reference
57

sample contained a substantial amount of sulphur (3.2 wt-%), which may have covered phosphorus on the surface of the catalysts. It is also possible that the catalyst used in this study did not reach temperatures high enough to convert phosphorus oxide to phosphates with the washcoat components. (Paper III) FTIR-ATR studies FTIR-ATR studies were carried out in addition to XRD to produce supplementary information about the poison compounds. Phosphate bands can be detected at 11001000 cm-1 in the IR spectra. The presence of AlPO4 was detected in both the Pt/Al2O3 (Figure 21) and Al2O3 samples, although phosphorus compounds could not be detected with XRD in the latter sample. In the case of TiO2 and Pt/TiO2 (Figure 21), a difference between the spectra of fresh and poisoned samples was clearly observed, since phosphate bands were identified at 1066 cm-1. In the same way with XRD, phosphates were observed in the ceria-based samples (CeO2, Zr-CeO2, Pt/CeO2, Pt/Zr-CeO2). Phosphates that could not be detected with XRD were detected with IR in the ZrO2 and Pt/ZrO2 samples. (Paper III)

58

120
a) b)
1394

110 100 90
1066

c) d) e)

80
1078

70 60

3600

3100

2600 cm
-1

2100

1600

1100

600

Fig. 21. FTIR-ATR spectra of a) fresh Pt/Al2O3, b) P-poisoned Pt/Al2O3, c) fresh Pt/TiO2, d) P-poisoned Pt/TiO2 and e) vehicle-aged reference catalyst (Paper III).

No phosphorus or any ageing-induced structural changes could be detected in the ZSM-5 and Pt/ZSM-5 samples with FTIR-ATR. This is an interesting result that may differentiate the phosphorus poisoning mechanism of these zeolites from other catalyst components, such as Zr-CeO2 mixed oxides, in which phosphorus was detected as phosphates after the same poisoning procedure. Still, this does not completely exclude the possibility of the presence of phosphates in the poisoned ZSM-5 and Pt/ZSM-5 samples. (Paper V) In the case of the diesel reference catalyst (360,000 km), phosphorus could not be detected with certainty with IR. The presence of phosphorus compounds was expected on the basis of the elemental analyses and XRD studies, but could not be confirmed due to the overlapping bands of the metal oxides and sulphates. The latter phases show an IR adsorption band at around 1078 cm-1. (Paper III) Thermodynamic studies Thermodynamic equilibrium calculations were used to explore the reactions between phosphorus compounds and catalyst components and to find new
59

Absorbance (a.u.)

3666

2988

information about the stability of the phosphorus compounds. The influence of reaction conditions on the poisoning phenomena was also studied. The calculations with CeO2 and Al2O3 in the presence of a (NH4)2HPO4-H2O-air mixture, which all were considered bulk materials, were carried out with commercial HSC Chemistry 5.11 software. (Krger et al. 2007) Equilibrium compositions and phase stability diagrams were calculated as a function of temperature and partial pressures of phosphorus and oxygen. The oxidizing conditions were chosen to simulate the conditions of the feed gas mixture in the laboratory-scale chemical ageing procedure developed in this work (Paper II). In the thermodynamic calculations the input corresponded to the total amount of solid oxide, aqueous salt solution of (NH4)2HPO4 and air fed into the system during the four hours of ageing. The following mixture was used to simulate the feed gas: 0.405 ml (NH4)2HPO4 (in the liquid phase), 23.595 ml H2O (in the liquid phase), 13.2 l O2, 46.8 l N2. The solid input included 1.5 g cerium oxide or aluminium oxide. The results from the thermodynamic equilibrium calculations were given as figures that present the mole fractions of the compounds as a function of temperature within a range of 0C to 1000C. Cerium and aluminium oxides were also studied with phase stability diagrams, changing the partial pressures of oxygen and phosphorus at different temperatures. (Krger et al. 2007) Mole fractions of phosphorus-containing compounds as a function of temperature in the presence of Al2O3 or CeO2 and the feed gas mixture containing (NH4)2HPO4, H2O, O2 and N2 are presented in Figs. 22 and 23, respectively. As can be seen in the figures, formation of AlPO4 and CePO4 starts at approximately 70C and the levels of these compounds are balanced at temperatures above 300C. According to the calculations, the relative amount of CePO4 formed at the chosen ageing temperature (700C) is higher than the amount of AlPO4 at the same temperature. This can be seen by comparing the molar proportions of CePO4 with CeO2 and AlPO4 with Al2O3, which are 0.54 and 0.23, respectively. At temperatures above 100C, the gas phase is composed almost exclusively of gaseous water and air (N2 and O2) and only very small mole fractions of NH3 and H2 are present. (Krger et al. 2007)

60

Fig. 22. Mole fractions of phosphorus-containing compounds as a function of temperature in the presence of Al2O3 and the feed gas mixture containing (NH4)2HPO4, H2O, O2 and N2 (Krger et al. 2007).

Fig. 23. Mole fractions of phosphorus-containing compounds as a function of temperature in the presence of CeO2 and the feed gas mixture containing (NH4)2HPO4, H2O, O2 and N2 (Krger et al. 2007).

Phase stability diagrams of the Al-O-P and Ce-O-P systems at 700C as a function of partial pressures of phosphorus and oxygen are presented in Figures 24 and 25. The figures show that both AlPO4 and CePO4 are formed in oxidizing conditions. It can also be seen that with the same partial pressure of phosphorus, formation of
61

CePO4 starts already at a lower partial pressure of oxygen than does formation of AlPO4. (Krger et al. 2007)

Fig. 24. Phase stability diagrams of the Al-O-P system at 700C as a function of phosphorus and oxygen partial pressure (Krger et al. 2007).

Fig. 25. Phase stability diagrams of the Ce-O-P system at 700C as a function of phosphorus and oxygen partial pressure (Krger et al. 2007).

62

According to the calculations, formation of phosphates with the washcoat components is thermodynamically favourable in the oxidizing conditions and with the oxide samples and phosphorus source used in this study. In addition, it was shown that aluminium and cerium phosphates can be formed at the peak temperature of diesel exhaust gas catalytic converters, which was chosen to be the temperature used in the ageing procedure developed and evaluated in this work. (Krger et al. 2007) 5.3 Loss of catalytic activity

5.3.1 Effect of phosphorus In activity measurements, the light-off temperatures (T50) of the fresh and aged catalyst components were compared with each other. T50 is defined as the temperature of 50% conversion of the feed component. Phosphorus poisoning has a significant deactivating effect on the activities of most of the samples. Some of the aged samples were strongly deactivated and their exact light-off temperatures could not be determined due to the vicinity of the autoignition temperature of the feed gas. Figures 26 and 27 illustrate that the C3H6 light-off temperatures of all the phosphorus-treated samples were significantly higher than those of the fresh catalysts in Test series 2 (samples containing oxide and Pt and ZSM-5). In addition, it was observed that already a low phosphorus content had a considerable deactivating impact on the activity of the catalysts even though the poison was evenly distributed in the catalyst powders and not only on the surface. (Papers III and V)

63

T50 (C) 500 Fresh H2O P+H2O 269 243 228 227 240 179 185 292

400

309 261 214

290

300

297 259

304

262 247 199

302 231

325

200

100

0 Pt/Al2O3 Pt/TiO2 Pt/CeO2 Pt/ZrO2 Pt/Zr-CeO2 Pt/Ce-ZrO2 Pt/ZSM-5

Fig. 26. C3H6 light-off temperatures of the fresh, hydrothermally aged, and poisoned samples with Pt.

T50 (C) 500 Fresh H2O P+H2O 383 350 362 364 388 369 346 370 367

400

300 228 200

100

0 Al2O3 TiO2 CeO2 ZrO2 Zr-CeO2 Ce-ZrO2 ZSM-5

Fig. 27. C3H6 light-off temperatures of the fresh, hydrothermally aged, and poisoned samples without Pt.

64

In the case of Test series 1 (samples containing oxide and Rh), light-off temperature of CO for Rh/Zr-CeO2 was 186C when fresh and 232C when poisoned and for Rh/Ce-ZrO2 173C and 229C, respectively. When comparing the results from Test series 1 and 2, it can be seen that the increase in the light-off temperatures for the cerium-rich Rh/Zr-CeO2 and Pt/Zr-CeO2 (see Fig. 26) samples after phosphorus poisoning was smaller compared with the results with the zirconium-rich Rh/Ce-ZrO2 and Pt/Ce-ZrO2 (see Fig. 26) samples. Therefore, it can be concluded that the cerium-rich mixed oxides studied in this work can better sustain the deactivating impact of phosphorus than can the zirconium-rich mixed oxides in terms of catalytic activity. The CO light-off temperatures of the Rh/Ce-ZrO2 and Rh/Zr-CeO2 samples were approximately the same after phosphorus impregnation, but the relative decrease in the activities of the phosphorus-treated Rh/Zr-CeO2 and Pt/Zr-CeO2 was almost equal. (Paper II) The activities in the propene oxidation reaction (Test series 2, Figures 26 and 27) were significantly higher in the case of samples containing platinum than were the activities of the same catalyst components without the precious metal. The propene light-off temperatures of the samples without Pt were expectedly high even with the fresh samples, the only exception being Al2O3. The light-off temperatures of the fresh TiO2, CeO2, ZrO2, Zr-CeO2, and Ce-ZrO2 samples were all between 350 and 370C. Hydrothermal ageing considerably decreased the activities of the samples. Most of the samples without Pt were strongly deactivated, and the exact propene light-off temperatures of these samples could not be determined due to the vicinity of the autoignition temperature of propene (460C). The most substantial effect was detected in the case of Al2O3, where the propene light-off temperature (>400C, could not be determined) had increased by over 170C compared with the fresh sample. After phosphorus treatment the parent samples were strongly deactivated and the light-off temperatures of all the samples without the precious metal were above 400C. (Paper III) The Al2O3-based sample showed the highest activity of all the samples, including those containing Pt. In addition, Pt/TiO2 as well as the zirconia-based samples, Pt/ZrO2 and Pt/Ce-ZrO2, had low propene light-off temperatures. Figure 28 shows an example of C3H6 conversion curves for the Pt/ZrO2 samples. The light-off temperatures of the ceria-based samples, Pt/CeO2 and Pt/Zr-CeO2, were higher. Similar to the samples without a precious metal, the platinum-containing samples were also substantially deactivated as a result of hydrothermal ageing. Notably, their light-off temperatures increased after hydrothermal ageing and poisoning, but in both cases they were less than the autoignition temperature of
65

propene (460C). After phosphorus ageing, the increase in the light-off temperature was highest for Pt/ZrO2 and lowest for Pt/CeO2 compared with the fresh samples. In the hydrothermally aged case, the most substantial effect of phosphorus poisoning also occurred in the Pt/ZrO2 sample, but unlike in the comparison with fresh samples, the smallest increase in the light-off temperature occurred with the Pt/Zr-CeO2 sample. (Paper III)
100 80 60 40 P+H2O, T50 = 290C 20 0 0 100 200 T (C) 300 400 500

C3H6 conversion (%)

Fresh, T50 = 214C

H2O, T50 = 262C

Fig. 28. C3H6 conversion as a function of temperature and light-off temperatures (T50) of the fresh, hydrothermally aged, and phosphorus-poisoned Pt/ZrO2 samples (Paper III).

With zeolites, the activities in propene oxidation were higher in the Pt/ZSM-5 samples than were the activities of the samples without the precious metal. The propene light-off temperature of fresh Pt/ZSM-5 was significantly lower than that of fresh ZSM-5 (Figure 26 and 27). The fresh zeolite samples showed the ability to adsorb propene at low temperatures. The presence of a precious metal did not affect the amount of HC adsorbed. Propene desorbed in the temperature range of 80190C. At temperatures higher than that, platinum started to affect propene conversion. As a result, the light-off temperature of the fresh Pt/ZSM-5 sample was relatively low, and a steep rise in the propene conversion curve could be observed, as shown in Figure 29. In contrast, the propene conversion curve of fresh ZSM-5 was not as steep, resulting in a considerably higher light-off temperature compared with the sample containing Pt. (Paper V)
66

100 80 C 3H 6 conversion (%) 60 40 P+H2O, T50 = 325C 20 0 0 -20 100 200 300 T (C) 400 500

Fresh, T50 = 231C

H2O, T50 = 302C

Fig. 29. C3H6 conversion as a function of temperature and light-off temperatures (T50) of the fresh, hydrothermally aged, and phosphorus-poisoned Pt/ZSM-5 samples (Paper V).

The ageing treatments had a deactivating effect on the catalytic activity of the zeolite samples. Hydrothermal ageing increased the light-off temperature of Pt/ZSM-5 and the effect of phosphorus ageing was even more substantial. After hydrothermal and phosphorus ageing, the ZSM-5 samples were heavily deactivated and the exact propene light-off temperatures of these samples could not be determined due to the vicinity of the autoignition temperature of propene (460C). The ageing treatments also affected the hydrocarbon adsorption capacity of the zeolites. After both hydrothermal and phosphorus ageing the adsorption behaviour of the Pt/ZSM-5 and ZSM-5 samples for propene, observed with the fresh samples, was almost completely lost. (Paper V) NO and NOx (combined NO and NO2) conversions remained low with the test gas mixture and diesel catalyst samples used in this work. After hydrothermal ageing and phosphorus poisoning, the maximum conversions of NO and NOx decreased and the temperatures of the maximum conversions increased compared with the fresh samples. The hydrothermally aged and phosphorus-poisoned samples were deactivated, resulting in very low NO and NOx conversions. The reasons behind the low NO and NOx conversions were the lean gas mixture and the relatively high gas flow in the activity tests. (Paper V)
67

In summary, the phosphorus treatment deactivated the catalyst samples more than the hydrothermal treatment alone. Deactivation appears to be associated with phosphorus compounds formed with the washcoat components. 5.3.2 Effect of phosphorus and calcium The role of calcium as a catalyst poison has been of a lesser interest in the literature compared with, for instance, phosphorus and sulphur. However, as described next, calcium has a deactivating effect on some Rh-containing catalyst powders. Calcium poisoning carried out with the impregnation method deactivated Rh/Ce-ZrO2 and Rh/Zr-CeO2 to a small extent, whereas in the case of powder Rh/Al2O3/OSC, no deactivation was observed. In all the samples the deactivating effect of calcium was clearly smaller than that caused by P poisoning. The CO light-off temperatures (T50) of the fresh, P, P+Ca, Ca, and Ca+P-treated Rh/ZrCeO2 and Rh/Al2O3/OSC samples, measured in lean reaction conditions, are presented in Figure 30. (Paper I)
T50 (C) 300 232 216 200 186 196 196 179 163 158 157 186

100

Fresh P P+Ca Ca Ca+P

0 Rh/Zr-CeO2 Rh/-Al2O3/OSC

Fig. 30. CO light-off temperatures of the fresh, P, P+Ca, Ca, and Ca+P-treated Rh/ZrCeO2 and Rh/-Al2O3/OSC samples.

Calcium treatment after phosphorus poisoning (P+Ca poisoning) had a regenerating effect on the light-off temperatures (Paper I). Ueda et al. (1994) have
68

also reported congruent results according to which the addition of calcium was observed to reduce the poisoning effect of phosphorus by forming calcium phosphates that prevent accumulation of phosphorus on the surfaces of the catalysts. Poisoning treatments carried out in a reversed order (Ca+P poisoning) showed that phosphorus treatment also had a regenerating effect on Ca-poisoned Rh/Ce-ZrO2 and Rh/Zr-CeO2. However, this was not the case with Rh/Al2O3/OSC, and phosphorus treatment of calcium-poisoned Rh/Al2O3/OSC reduced catalytic activity. It is possible that the more complex composition and the presence of Al2O3 in Rh/Al2O3/OSC affect the reaction between calcium and phosphorus. (Paper I) As demonstrated in this study, calcium can decrease the catalytic activity of components used in automotive catalysts. Therefore, the role of calcium in catalyst deactivation should be considered more profoundly, and further studies of poisoning mechanisms are needed. 5.4 Changes in the active metal particles and washcoat grain size

The TEM analysis indicated considerable structural changes in the catalyst washcoat and Pt particles after both ageing treatments. The grain size of the Pt/Al2O3 washcoat had not increased and no transition from the -phase to the phase had occurred. The platinum particles had grown in both treatments. Some large particles (>100 nm) were also detected. The grain size of the Pt/TiO2 washcoat had grown from about 1015 nm to 3040 nm after the treatments. Detection of platinum was very difficult, and no clear indications of Pt particle growth were observed in the treated catalysts. The grain size of both the Pt/CeO2 and Pt/Zr-CeO2 washcoats had increased from under 10 nm to 1020 nm. Detection of platinum was very difficult also in these two samples, and only occasionally were large Pt particles observed in the aged catalysts. According to the analysis, the washcoat of the zirconia-based Pt/ZrO2 and Pt/Ce-ZrO2 samples had gone through only minor changes due to the treatments. Once again, platinum particles were very hard to detect, thus only a few large Pt particles were observed in the treated catalysts. (Paper VI) Figure 31 presents TEM images of the fresh, hydrothermally aged, and phosphorus-poisoned Pt/ZSM-5 catalysts. Platinum particles appear in black in the images. Clustering of large Pt crystallites was observed in the phosphoruspoisoned Pt/ZSM-5. The Pt particles have grown larger in the poisoned Pt/ZSM-5
69

than in the hydrothermally aged Pt/ZSM-5. The hydrothermal ageing and phosphorus poisoning treatments did not cause any clear changes in the microstructures of the ZSM-5 and Pt/ZSM-5 zeolite washcoats. However, the platinum particle size distribution measurements of the Pt/ZSM-5 zeolite showed an increase in Pt particle size after both hydrothermal ageing and poisoning. The number of small Pt particles (diameter of 025 nm) is higher in the fresh Pt/ZSM5 than in the treated samples. The growth of the Pt particles in the hydrothermally aged Pt/ZSM-5 occurred at the expense of small Pt particles. Clusters of large Pt crystallites were observed in the phosphorus-poisoned Pt/ZSM-5. (Paper V)

Fig. 31. TEM images of the Pt/ZSM-5 catalyst when (a) fresh, (b) hydrothermally aged, and (c) phosphorus-poisoned (Paper V).

Phosphorus ageing affected the characteristics of the samples more than did hydrothermal ageing alone. Both atomic migration and crystallite migration are possible mechanisms of Pt cluster formation. The changes in the characteristics observed in the physisorption studies are consistent with the microstructural changes detected by TEM in the Pt-containing samples after ageing. The accumulation of phosphorus, the increase in the washcoat grain size and the growth of Pt crystallites cause a loss of active surface area needed for catalytic reactions to take place. (Paper VI)

70

6
6.1

Evaluation of the poisoning procedures

Integration of the results

This work was carried out in order to study the mechanisms of phosphorus poisoning on different automotive catalyst components. The aim was to obtain information about the poisoning conditions that catalyst components should be able to withstand during motor operation. In this study, two procedures for studying chemical ageing on the laboratory scale were developed. It was shown that the role of phosphorus poisoning is relevant in the deactivation of automotive catalyst components. Phosphorus added to the ageing treatments decreased the characteristic surface areas of all the powder catalyst samples and reduced their catalytic activity. The changes were associated with formation of compounds between phosphorus and the catalyst components or plugging of pores on the catalytic surfaces. Pore plugging was also observed by Joy et al. (1985), who suggested that it accounts for some of the deactivation. Furthermore, the results from the surface characterizations of the powder samples contaminated by phosphorus compounds formed during the ageing procedures are congruent with the results observed with the reference catalysts used in this work as well as with the results from earlier studies (Lassi 2003, Hietikko et al. 2004, Rokosz et al. 2001, Uy et al. 2003, Xu et al. 2004, Cabello Galisteo et al. 2004) carried out with aged catalysts. In the case of Test series 1 (Rh-loaded powder samples), phosphorus had a deactivating impact on the activity of the catalyst powders. The deactivating effect of calcium was not as pronounced as that of phosphorus. In contrast, adding calcium to a phosphorus-poisoned catalyst was found to have even a regenerating effect on its activity. In addition, the order in which calcium and phosphorus were added significantly affected the properties of the catalyst powders. Phosphorus poisoning also had a deactivating effect on the activity of catalyst components in Test series 2 (Pt and oxides and ZSM-5). An increase in Pt particle size and washcoat grain size was also detected after phosphorus poisoning. In the case of zeolites, the samples ability to adsorb propene was lost after the phosphorus treatment. Furthermore, the effect of pure phosphorus poisoning could be differentiated from the effect of hydrothermal ageing conditions by comparing the results from the activity tests carried out with the samples poisoned in hydrothermal conditions and with the hydrothermally aged samples.
71

According to the results, hydrothermal ageing conditions alone deactivated the samples, but to a lesser extent than did phosphorus ageing. 6.2 Evaluation and utilization of the results

Several characterization techniques were used in this thesis to gain information about phosphorus poisoning of automotive catalyst components. Physisorption analyses were carried out to characterize the poisoning and ageing-induced changes in the specific surface area (BET theory) of the samples. We evaluated the changes in total pore volume and average pore size by using the BJH theory. With this method we demonstrated that the specific surface area of most of the samples was significantly affected by phosphorus poisoning. Useful information about the contaminant levels in the poisoned samples was gathered with SEM-EDS. In addition, it was shown that there were spatial differences in the accumulation of poisons in the reference catalysts. Phosphorus was observed to accumulate in the top overlayer of the catalysts, whereas sulphur was observed also in the inner layers of the catalyst. XRF was a good tool for quick determination of the phosphorus content of the poisoned catalyst samples. ICP-OES was shown to be a feasible and accurate method for determining the phosphorus content of the powdery samples. TEM studies were used to characterize the composition and microstructure of the samples. The method also produced new information about changes in washcoat and Pt particle sizes. An increase in washcoat grain size and growth of Pt crystallites were detected. However, the TEM studies did not provide any information about accumulation of phosphorus in the samples. XRD was used to identify phosphorus compounds in the aged catalyst samples. The method was applicable for studying poison compounds in both powdery and metallic catalyst samples. The information about poison compounds obtained with XRD was supplemented by carrying out FTIR-ATR studies. Laboratory scale light-off experiments were conducted to study the poisoning-induced changes between the fresh and aged catalyst powder samples. A significant decrease in catalytic activity was detected when comparing the activities of the fresh and poisoned samples. The procedure was shown to be a viable method for laboratory-scale catalyst studies. New information about the formation and stability of phosphorus compounds was obtained by using thermodynamic calculations. The calculations showed that formation of phosphates is thermodynamically possible at both the temperatures
72

of normal diesel catalyst operation (below 500C) and the peak temperature used in this study (700C). The results demonstrated that poisoning by both methods used in this work, namely incipient wetness impregnation with an aqueous salt solution of (NH4)2HPO4 and addition of an aqueous salt solution of (NH4)2HPO4 into a gaseous phase, produces the same poison compounds found in vehicle-aged catalysts. Phosphorus can be added to samples in a repeatable, time-saving, affordable, and safe way with both of these methods. Therefore, the methods are considered to be potential ageing procedures when studying accelerated chemical poisoning of powdery automotive exhaust catalyst components on the laboratory scale. 6.3 Suggestions for further research

To prevent poisoning and to prolong the life of automotive catalysts, further studies are needed, especially to obtain more information about the poisoning mechanisms of zeolites. It was shown that accumulation of phosphorus combined with hydrothermal ageing conditions caused a decrease in the catalytic activity of zeolites. However, the form in which the poison adhered to the sample could not be identified with the characterization techniques used in this work. According to the results, calcium can decrease the catalytic activity of catalyst components. Furthermore, calcium also has a regenerating effect on catalysts poisoned by phosphorus. Therefore, the role of calcium in catalyst deactivation and in preventing the deactivating effects of other poison compounds should be studied in more detail. In the future, gaseous phase ageing can be carried out by using different poison sources to determine their separate and simultaneous effect on the behaviour of a catalyst. For instance, zinc, calcium or magnesium could be used to simulate the accumulation of other lubricating oil-derived impurities on exhaust gas catalysts. Furthermore, depending on the real driving conditions that are to be simulated, it would also be possible to vary the gas composition and flow rate as well as the ageing time and temperature.

73

74

Summary and conclusions

Both poisoning methods used in this study produced the same phosphorus compounds that can be found in vehicle-aged catalysts. Phosphorus was detected in the poisoned samples as phosphates of Ce, Zr, Al and Ti. By contrast, in zeolites phosphorus was detected with elemental analysis, but no phosphoruscontaining compounds were observed. According to the results, phosphorus poisoning takes place at high operating temperatures and with high oxygen and water content in the gas phase. It was also shown that the role of phosphorus poisoning was more pronounced than the role of hydrothermal ageing alone. During the gaseous phase phosphorus ageing procedure, less than 70% of the phosphorus fed into the reactor accumulated on the samples. Furthermore, the phosphorus content varied between the catalyst samples. However, it was difficult to find correlations between the amount of phosphorus attached to the sample and the characteristics of the sample observed in this study. In the case of zirconium oxides, the very low specific surface area of the fresh ZrO2 and Pt/ZrO2 must have been the reason for the small phosphorus content measured in these samples. The results indicate that phosphorus poisoning mainly affects the oxide components used in this study instead of affecting the noble metals. Regardless of whether the component contained a noble metal or not, the effect of phosphorus poisoning was detected in the catalytic activity and characteristics of the catalyst. As a result of poisoning, most of the samples that did not contain any noble metal suffered from a very significant decrease in catalytic activity. Despite the decrease in the activity of the catalysts containing a noble metal, these samples maintained some of their activity in the ageing procedures. According to the results, calcium can decrease the catalytic activity of components used in automotive catalysts. The role of calcium in catalyst deactivation should be studied further. The regenerating effect that calcium had on some catalyst components is also a very interesting phenomenon that merits an in-depth study. The methods used in the deactivation studies (i.e. activity measurements, physisorption analyses, SEM, TEM, XRD, FTIR-ATR, ICP-OES, XRF, and thermodynamic calculations) of this thesis project are relevant and timely methods for gathering information about deactivation phenomena and for catalyst characterization in catalyst deactivation studies. The role of laboratory-scale activity measurements, physisorption studies, and XRD combined with FTIR75

ATR is most relevant. These methods are recommended to be used as tools in the engineering approach to analysis of deactivation of exhaust gas catalysts. The results of this work can be utilized in the development of catalytic materials and catalyst compositions that can withstand phosphorus poisoning at high temperatures and in moist ageing conditions better than existing ones do. Other areas where the results can be applied include evaluation of the effects of phosphorus poisoning on different catalyst compositions and prediction of the age of commercial catalysts. To obtain more accurate information about the causes of deactivation and their influence on catalyst activity and selectivity, a deeper analysis of catalyst activity versus catalyst characteristics should be done to be able to design catalysts from the point of view of materials and to specify their operating conditions, e.g. the presence of impurities and temperature windows, more precisely.

76

References
Andersson J, Antonsson M, Eurenius L, Olsson E & Skoglundh M (2007) Deactivation of diesel oxidation catalysts: Vehicle and synthetic aging correlations. Applied Catalysis B: Environmental 72: 7181. Anderson JA, Daley RA, Christou SY & Efstathiou AM (2006) Regeneration of thermally aged Pt-Rh/CexZr1-xAl2O3 model three-way catalysts by oxychlorination treatments. Applied Catalysis B: Environmental 64: 189200. Angelidis TN (2001) Development of a laboratory scale hydrometallurgical procedure for the recovery of Pt and Rh from spent automotive catalysts. Topics in Catalysis 16/17: 419423. Angelidis TN & Sklavounos SA (1995) A SEM-EDS study of new and used automotive catalysts. Applied Catalysis A: General 133: 121132. Angove DE & Cant, NW (2000) Position dependent phenomena during deactivation of three-way catalytic converters on vehicles. Catalysis Today 63: 371378. Bartholomew CH (2004) Catalyst deactivation and regeneration. In: Kirk-Othmer Encyclopedia of Chemical Technology. 5th ed. John Wiley & Sons, Inc, vol. 5, 255 322. Bartholomew CH (2001) Mechanisms of catalyst deactivation. Applied Catalysis A: General 212: 1760. Battistoni C, Cantelli V, Debenedetti M, Kaiulis S, Mattogno G & Napoli A (1999) XPS study of ceramic three-way catalysts. Applied Surface Science 144145: 390394. Beck DD, Sommers JW & DiMaggio CL (1997) Axial characterization of catalytic activity in close-coupled lightoff and underfloor catalytic converters. Applied Catalysis B: Environmental 11: 257272. Beckmann R, Engeler W, Mueller E, Engler BH, Leyrer BH, Lox ES & Ostgathe K (1992) A new generation of diesel oxidation catalysts. SAE Technical Paper Series 922330: 101117. Birgersson H, Eriksson L, Boutonnet M & Jrs SG (2004) Thermal gas treatment to regenerate spent automotive three-way exhaust gas catalysts (TWC). Applied Catalysis B: Environmental 54: 193200. Birgersson H, Boutonnet M, Klingstedt F, Murzin DYu, Stefanov P & Naydenov A (2006) An investigation of a new regeneration method of commercial aged three-way catalysts. Applied Catalysis B: Environmental 65: 93100. Bissett EJ (1984) Mathematical model of the thermal regeneration of a wall-flow monolith diesel particulate filter. Chemical Engineering Science 39: 12331244. Boaro M, de Leitenburg C, Dolcetti G, Trovarelli A & Graziani M (2001) Oxygen storage behaviour of ceria-zirconia-based catalysts in the presence of SO2. Topics in Catalysis 16/17: 299306. Brett PS, Neville AL, Preston WH & Williamson J (1989) An investigation into lubricant related poisoning of automotive three-way catalysts and lambda sensors. SAE Technical Paper Series 890490: 112.

77

Bunting BG, More K, Lewis S & Toops T (2005) Phosphorus poisoning and phosphorus exhaust chemistry with diesel oxidations catalysts. SAE Specific Publications SP1942: 329336. Butt JB & Petersen EE (1988) Activation, deactivation and poisoning of catalysts. Academic Press, Inc., New York. Cabello Galisteo F, Larese C, Mariscal R, Lpez Granados M, Fierro JLG, Frnandez-Ruiz R & Furio M (2004) Deactivation on vehicle-aged diesel oxidation catalysts. Topics in Catalysis 30/31: 451456. Cabello Galisteo F, Mariscal R, Lpez Granados M, Fierro JLG, Brettes P & Salas O (2005a) Reactivation of a commercial diesel oxidation catalyst by acid washing. Environmental Science & Technology 39: 38443848. Cabello Galisteo F, Mariscal R, Lpez Granados M, Fierro JLG, Daley RA & Anderson JA (2005b) Reactivation of sintered Pt/Al2O3 oxidation catalysts. Applied Catalysis B: Environmental 59: 227233. Cabello Galisteo F, Mariscal R, Lpez Granados M, Zafra Poves MD, Fierro JLG, Krger V & Keiski RL (2007) Reactivation of sulphated Pt/Al2O3 catalysts by reductive treatment in the simultaneous oxidation of CO and C3H6. Applied Catalysis B: Environmental 72: 272281. Cant N & Angove D (1998) The performance of catalytic converters retrofitted to diesel vehicles. NSW EPA report number MV-A-38-appendix. Caracciolo F & Spearot JA (1976) The effect of phosphorus containing engine oil additives on exhaust oxidation catalyst degradation. SAE Technical Paper Series 760562: 1898 1910. Carduner KR Chattha MS & Gandhi HS (1988) Phosphorus poisoning of the three-way catalyst studied by 31P NMR. Journal of Catalysis 109: 3740. Christou SY, Birgersson H & Efstathiou AM (2006) Reactivation of severely aged commercial three-way catalysts by washing with weak EDTA and oxalic acid solutions. Applied Catalysis B: Environmental 71: 185198. Culley SA, McDonnell TF, Ball DJ, Kirby CW & Hawes SW (1996) The impact of passenger car motor oil phosphorus levels on automotive emissions control systems. SAE Technical Paper Series 961898: 1321. Clerc JC (1996) Catalytic diesel exhaust aftertreatment. Applied Catalysis B: Environmental 10: 99115. Dickey JR (2005) Lubrication and lubricants. In: Kirk-Othmer Encyclopedia of Chemical Technology. 5th ed. John Wiley & Sons, Inc, vol. 15, 182. van Diepen AE, Makkee M & Moulijn JA (1999) Emission control from mobile sources: Otto and diesel engines, in Janssen FJJG & van Santen RA eds., Environmental Catalysis, Vol 1, London, Imperial College Press, Chapter 12, p. 257291 (Catalytic Science Series). EU (2003) Directive 2003/30/EC of the European Parliament and of the Council of 8 May 2003 on the promotion of the use of biofuels and other renewable fuels for transport. Official Journal of the European Union L 123, 17/05/2003 P. 4246.

78

Farrauto RJ & Voss KE (1996) Monolithic diesel oxidation catalysts. Applied Catalysis B: Environmental 10: 2951. Farrauto RJ & Heck RM (1999) Catalytic converters: state of the art and perspectives. Catalysis Today 51: 351360. Farrauto RJ & Heck RM (2000) Environmental catalysis into the 21st century. Catalysis Today 55: 179187. Fernndes Ruiz R, Furi M, Cabello Galisteo F, Larese C, Lpez Granados M, Mariscal R & Fierro JLG (2002) Chemical analysis of used three-way catalysts by Total Reflection X-ray Fluorescense. Analytical Chemistry 74: 54635469. Forzatti P & Lietti L (1999) Catalyst deactivation. Catalysis Today 52: 165181. Fridell E, Persson H, Olsson L, Westerberg B, Amberntsson A & Skoghlund M (2001) Model studies of NOx storage and sulphur deactivation of NOx storage catalysts. Topics in Catalysis 16/17: 133137. Frost JC & Smedler G (1995) Control of NOx emissions in diesel powered light vehicles. Catalysis Today 26: 207214. Gandhi HS, Graham GW & McCabe RW (2003) Automotive exhaust catalysis. Journal of catalysis 216: 433442. Gonzles-Velasco JR, Botas JA, Ferret R, Gonzles-Marcos MP, Marc J-L & GutirrezOrtiz MA (2000) Thermal Aging of Pd/Pt/Rh automotive catalysts under a cycled oxidizing-reduzing environment. Catalysis Today 59: 395402. Goralski CT Jr. & Schneider WF (2002) Analysis of the thermodynamic feasibility of NOx decomposition catalysis to meet next generation vehicle NOx emissions standards. Applied Catalysis B: Environmental 37: 263277. Greening P (2001) European vehicle emission legislation present and future. Topics in Catalysis 16/17: 513. Grosser Z & Bowman M (2006) The Importance of Biodiesel Analysis. Laboratory News Online. http://www.labnews.co.uk/laboratory_article.php/1532/5/the-importance-ofbiodiesel-analysis (28.3.2007) Guo J, Konno M, Chikahisa T, Murayama T & Iwamoto M (1995) NOx reduction characteristics of Pt-ZSM-5 catalyst with diesel engine exhaust. JSAE Review 16: 21 25. Hamelinck CN & Faaij APC (2006) Outlook for advanced biofuels. Energy Policy 34: 32683283. Heck RM & Farrauto RJ (1997) The automobile catalyst: Its present and future family tree. Cattech 2: 117124. Heck RM & Farrauto RJ (2001) Automobile exhaust gas catalysts. Applied Catalysis A: General 221: 443457. Hietikko M, Lassi U, Kallinen K, Savimki A, Hrknen M, Pursiainen J, Laitinen R & Keiski RL (2004) Effect of the ageing atmosphere on catalytic activity and textural properties of Pd/Rh exhaust gas catalysts studied by XRD. Applied Catalysis A: General 277: 107117.

79

Inoue K, Kurahashi T, Negishi T, Akiyama K, Arimura K & Tasaka K (1992) Effects of phosphorus and ash contents of engine oils on deactivation of monolithic three-way catalysts and oxygen sensors. SAE Technical Paper Series 920654: 19. Joy GC, Molinaro FS and Homeier EH (1985) Influence of phosphorus on threecomponent control catalysts: Catalyst durability and characterization studies. SAE Technical Paper Series 852099: 5364. Kalantar Neyestanaki A, Klingstedt F, Salmi T & Murzin DYu (2004) Deactivation of postcombustion catalysts, a review. Fuel 83: 395408. Karjalainen H, Lassi U, Rahkamaa-Tolonen K, Krger V & Keiski RL (2005) Thermodynamic equilibrium calculations of sulfur poisoning in CeOS and LaOS systems. Catalysis Today 100: 291295. Kapar J, Fornasiero P, Hickey N (2003) Automotive catalytic converters: current status and some perspectives. Catalysis Today 77: 419449. Koebel M, Elsener M & Kleemann M (2000) Urea-SCR: a promising technique to reduce NOx emissions from automotive diesel engines. Catalysis Today 59: 335345. Koltsakis GC & Stamatelos AM (1997) Catalytic automotive exhaust aftertreatment. Progress in Energy and Combustion Science 23: 139. van Kooten WEJ, Krijnsen HC, van den Bleek CM and Calis HPA (2000) Deactivation of zeolite catalysts used for NOx removal. Applied Catalysis B: Environmental 25: 125 135. Krger V, Turpeinen E, Hietikko M, Lassi U & Keiski RL (2007) Evaluation of a novel laboratory-scale phosphorus poisoning procedure for exhaust gas catalysis. To be submitted. Kumar SV, Rogalo J, Deeba M, Burk PL & Sardesal P (2004) Simulation of phosphorus poisoning and deactivation of TWC catalysts. SAE Technical Paper Series 280023: 136152. Larese C, Cabello Galisteo F, Lpez Granados M, Mariscal R, Fierro JLG, Furi M & Fernndes Ruiz R (2003) Deactivation of real three way catalysts by CePO4 formation. Applied Catalysis B: Environmental 40: 305317. Larese C, Cabello Galisteo F, Lpez Granados M, Mariscal Lpez R, Fierro JLG, Lambrou PS & Efstathiou AM (2004a) Effects of calcination temperature on the stability of CePO4 detected in vehicle-aged commercial three-way catalysts. Applied Catalysis B: Environmental 48: 113123. Larese C, Cabello Galisteo F, Lpez Granados M, Mariscal R, Fierro JLG, Lambrou PS & Efstathiou AM (2004b) Effects of the CePO4 on the oxygen storage and release properties of CeO2 and Ce0.8Zr0.2O2 solid solution. Journal of Catalysis 226: 443456. Larese C, Lpez Granados M, Cabello Galisteo F, Mariscal R & Fierro JLG (2005) TWC deactivation by lead: A study of the Rh/CeO2 system. Applied Catalysis B: Environmental 62: 132143. Lassi U (2003) Deactivation Correlations of Pd/Rh Three-way Catalysts Designed for Euro IV Emissions Limits: Effect of Ageing Atmosphere, Temperature and Time. Oulu, Oulu University Press, Acta Universitatis Ouluensis Technica C 180, Dissertation. 139 p. Available on line at: http://herkules.oulu.fi/isbn9514269543 80

Laurikko J (1994) Emissions performance of current TWC vehicles at low ambient temperature over FTP and ECE test cycles. SAE Technical Paper Series 940933: 273 287. Laurikko J (1995) Ik painaakolhut vaurioittavat. Katalysaattorin kestoik ja siihen vaikuttavat tekijt. Suomen autolehti 1: 1216. (In Finnish) Liu DR & Park J-S (1993) Electron microprobe characterization of phosphorus containing deposits on used automotive catalyst surfaces. Applied Catalysis B: Environmental 2: 4970. Lpez Granados M, Cabello Galisteo F, Lambrou PS, Mariscal R, Sanz J, Sobrados I, Fierro JLG & Efstathiou AM (2006) Role of P-containing species in phosphated CeO2 in the deterioration of its oxygen storage and release properties. Journal of Catalysis 239: 410421. Lpez Granados M, Larese C, Cabello Galisteo F, Mariscal R, Fierro JLG, Fernndez-Ruz R, Sanguino R & Luna M (2005) Effect of mileage on the deactivation of vehicleaged three-way catalysts. Catalysis Today 107108: 7785. Lox ESJ & Engler BH (1997) Environmental catalysis - Mobile sources. In: Ertl G, Knzinger H & Weitkamp (eds) Handbook of Heterogeneous Catalysis, Vol. 4. Wiley-VCH, Weinheim, p. 15591695. Madia G, Koebel M, Elsener M & Wokaun A (2002) The effect of an oxidation precatalyst on the NOx reduction by ammonia SCR. Industrial & Engineering Chemistry Research 41: 35123517. Mahzoul H, Gilot P, Brilhac J-F & Stanmore BR (2001) Reduction of NOx over NOx-trap catalyst and the regeneration behaviour of adsorbed SO2. Topics in Catalysis 16/17: 293298. Mittelbach M (1996) Diesel fuel derived from vegetable oils, VI: Specifications and quality control of biodiesel. Bioresource Technology 56: 711. Monroe DR (1980) Phosphorus and lead poisoning of pelleted three-way catalysts. SAE Technical Paper Series 800859: 112. Mooney JJ (2005) Emission control, automotive. In: Kirk-Othmer Encyclopedia of Chemical Technology. 5th ed. John Wiley & Sons, Inc, vol. 10, 3067. Moulijn JA, van Diepen AE & Kapteijn F (2001) Catalyst deactivation: Is it predictable? What to do? Applied Catalysis A: General 212: 316. Research Institute of Petroleum Industry (2002) Zinc dialkyl dithio phosphate (ZDDP), NIOC-RIPI. http://engineering.ripi.ir/ (15.12.2006) Neste Oil Oyj (2006) Dieselpolttoaineopas. http://www.nesteoil.fi/default.asp?path=35,52,63,310,355 (20.11.2006) (In Finnish) PDF-2 Powder Diffraction File Database, International Centre for Diffraction Data, 12 CampusBoulevard Newtown Square, PA 190733273 USA. Rabinowitz HN, Tauster SJ & Heck RM (2001) The effects of sulfur and ceria on the activity of automotive Pd/Rh catalysts. Applied Catalysis A: General 212: 215222. Richardson JT (1989) Principles of catalyst development. Plenum Press, New York, USA.

81

Rodriguez JA & Hrbek J (1999) Interaction of sulfur with well-defined metal oxide surfaces: Unraveling the mysteries behind catalyst poisoning and desulfurization. Accounts of Chemical Research 32: 719728. Roduit B, Baiker A, Bettoni F, Baldyga J & Wokaun A (1998) 3-D modeling of SCR of NOx by NH3 on vanadia honeycomb catalysts. AiChE Journal 44: 27312744. Roine, Outokumpu (2002) HSC Chemistry for Windows, Chemical reaction and equilibrium software with extensive thermochemical database. Users guide. Version 5.1. Outokumpu Research Oy, Pori. Rokosz MJ, Chen AE, Lowe-Ma CK, Kucherow AV, Benson D, Paputa Peck MC & McCabe RW (2001) Characterization of phosphorus-poisoned automotive exhaust catalysts. Applied Catalysis B: Environmental 33: 205215. Romm J (2006) The car and fuel of the future. Energy Policy 34: 26092614. Sadtler Spectral Databases, Bio-Rad Laboratories, Maxted Road, Hemel Hempstead, Heartfordshire HP2 7DX, United Kingdom; Database DB 1000 Selected Inorganics. Sagar AD (1995) Automobiles and global warming: Alternative fuels and other options for carbon dioxide emissions reduction. Environmental Impact Assessment Review 15: 241274. Seymor RJ & O'Farrelly JI (2005) Platinum-group metals. In: Kirk-Othmer Encyclopedia of Chemical Technology. 5th ed. John Wiley & Sons, Inc, vol. 10, 3067. Shelef M & McCabe RW (2000) Twenty-five years after introduction of automotive catalysts: What next? Catalysis Today 62: 3550. Shinjoh H (2006) Rare earth metals for automotive exhaust catalysts. Journal of Alloys and Compounds 408412: 10611064. Skoglundh M, Thormhlen P, Fridell E, Hajbolouri F & Jobson E (1999) Improved lightoff performance by using transient gas compositions in the catalytic treatment of car exhaust. Chemical Engineering Science 54: 45594566. Stamatelos AM (1997) A review of the effect of particulate traps on the efficiency of vehicle diesel engines. Energy Conversion and Management 38: 8399. Taylor C (1984) Automobile catalytic converters, in: Anderson JR & Boudart M (Eds.), Catalysis: Science and technology, Springer-Verlag, Berlin, Chapter 2, pp. 119170. Taylor KC (1993) Nitric oxide catalysis in automotive exhaust systems. Catalysis Reviews - Science and Engineering 35: 457481. Twigg MV (2003) Automotive exhaust emissions control. Platinum Metals Review 47(4): 157162. Twigg MV (2007) Progress and future challenges in controlling automotive exhaust gas emissions. Applied Catalysis B: Environmental 70: 215. Twigg MV (2006) Roles of catalytic oxidation control of vehicle exhaust emissions. Catalysis Today 117: 407418. Ueda F, Sugiyama S, Arimura K, Hamaguchi S & Akiyama K (1994) Engine oil additive effects on deactivation of monolithic three-way catalysts and oxygen sensors. SAE Technical Paper Series 940746: 197206.

82

Uy D, ONeill AE, Xu L, Weber WH & McCabe RW (2003) Observation of cerium phosphate in aged automotive catalysts using Raman spectroscopy. Applied Catalysis B: Environmental 41: 269278. Wallington TJ, Kaiser EW & Farrell JT (2006) Automotive fuels and internal combustion engines: a chemical perspective. Chemical Society Reviews 35: 335347. Webb CC, Bartley GJJ, Bykowski BB, Fransworth G & Riley M (2003) Catalyst aging evaluation with exposure to 0.06 and 0.11 percent phosphorus oils using the FOCASTM burner system. JSAE 20030269: 112. Webb PA & Orr C (1997) Analytical methods in fine particle technology. Micromeritics Instrument Corporation, Norcross, USA. 53153. Williamson WB, Gandhi HS, Heyde ME & Zawacki GA (1979a) Deactivation of threeway catalysts by fuel contaminants: lead, phosphorus and sulfur. SAE Technical Paper Series 790942: 19. Williamson WB, Stepien HK, & Gandhi HS (1980) Poisoning of platinum-rhodium automotive three-way catalysts: Behavior of single-component catalysts and effects of sulfur and phosphorus. American Chemical Society 14(3): 319324. Williamson WB, Stepien HK, Watkins WLH & Gandhi HS (1979b) Poisoning of platinum-rhodium automotive three-way catalysts by lead and phosphorus. Environmental Science and Technology 13: 11091113. Williamson WB, Perry J, Goss RL, Gandhi HS & Beason RE (1984) Catalyst deactivation due to glaze formation from oil-derived phosphorus and zinc. SAE Technical Paper Series 841406: 110. Williamson WB, Perry J, Gandhi HS & Bomback JL (1985) Effects of oil phosphorus of monolithic three-way catalysts. Applied Catalysis 15: 277292. Wong C & McCabe RW (1989) Effects of high-temperature oxidation and reduction on the structure and activity of Rh/Al2O3 and Rh/SiO2 catalysts. Journal of Catalysis 119: 4764. Voss K, Yavuz B, Hirt C & Farrauto R (1994) Performance characteristics of a novel diesel oxidation catalyst. SAE Technical Paper Series 940239: 81100. Votsmeier M, Kreuzer T & Lepperhoff G (2005) Automobile exhaust control. In: Ullmann's Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, pp. 115. Xu L, Guo G, Uy D, ONeill AE, Weber WH, Rokosz MJ & McCabe RW (2004) Cerium phosphate in automotive exhaust catalyst poisoning. Applied Catalysis B: Environmental 50: 113125. Yu T-C & Shaw H (1998) The effect of sulfur poisoning on methane oxidation over palladium supported on -alumina catalysts. Applied Catalysis B: Environmental 18: 105114. Zelenka P, Ostgathe K & Lox E (1990) Reduction of diesel exhaust emissions by using oxidation catalysts. SAE Technical Paper Series 902111: 111. Zelenka P, Cartellieri W & Herzog P (1996) Worldwide diesel emissions standards, current experiences and future needs. Applied Catalysis B: Environmental 10: 328.

83

84

Original papers
I Krger V, Hietikko M, Lassi U, Ahola J, Kallinen K, Laitinen R & Keiski RL (2004) Characterization of the effects of phosphorus and calcium on the activity of Rhcontaining catalyst powders. Topics in Catalysis 30/31: 469473. II Krger V, Lassi U, Kynknniemi K, Suopanki A & Keiski RL (2006) Methodology development for laboratory-scale exhaust gas catalyst studies on phosphorus poisoning. Chemical Engineering Journal 120: 113118. III Krger V, Hietikko M, Angove D, French D, Lassi U, Suopanki A, Laitinen R & Keiski RL (2007) Effect of phosphorus poisoning on catalytic activity of diesel exhaust gas catalyst components containing oxide and Pt. Topics in Catalysis 4243: 409413. IV Krger V, Kanerva T, Lassi U, Rahkamaa-Tolonen K, Vippola M & Keiski RL (2007) Characterization of phosphorus poisoning on diesel exhaust gas catalyst components containing oxide and Pt. Topics in Catalysis 45: 153157. V Krger V, Kanerva T, Lassi U, Rahkamaa-Tolonen K, Lepist T & Keiski RL (2007) Phosphorus poisoning of ZSM-5 and Pt/ZSM-5 zeolite catalysts in diesel exhaust gas conditions. Topics in Catalysis 4243: 433436. VI Kanerva T, Krger V, Rahkamaa-Tolonen K, Vippola M, Lepist T & Keiski RL (2007) Structural changes in air aged and poisoned diesel catalysts. Topics in Catalysis 45: 137142.

I, III, IV, V and VI Reprinted with kind permission of Springer Science and Business Media. II Reprinted with permission of Elsevier. Original papers are not included in the electronic version of the dissertation.

85

86

ACTA UNIVERSITATIS OULUENSIS

SERIES C TECHNICA

266. 267. 268. 269. 270. 271. 272. 273. 274. 275. 276. 277. 278. 279. 280. 281. 282.

Harri Haapasalo & Pivi Iskanius (Eds.) (2007) The 1st Nordic Innovation Research ConferenceFinnkampen Pikka, Vesa (2007) A Business Enabling Network. A case study of a high-tech network; its concepts, elements and actors Noora Ervasti, Elina Pietikinen, Eva Wiklund, Maarit Hiltunen ja Esa Vesmanen (2007) Yksillliset toiveet asuntotuotannossa Fabritius, Tapio (2007) Optical method for liquid sorption measurements in paper Miao, Honglei (2007) Channel estimation and positioning for multiple antenna systems Wang, Lingyun (2007) The key activities of partnership development in Chinaa study of Sino-Finnish partnerships Aikio, Janne P. (2007) Frequency domain model fitting and Volterra analysis implemented on top of harmonic balance simulation Oiva, Annukka (2007) Strategiakeskeinen kyvykkyyden johtaminen organisaation strateginen valmius. Kahden johtamismallin testaus ja

Jokinen, Hanna (2007) Screening and cleaning of pulpa study to the parameters affecting separation Sarja, Tiina (2007) Measurement, nature and removal of stickies in deinked pulp Tth, Gza (2007) Computer modeling supported fabrication processes for electronics applications Nsi, Jari (2007) Intensified use of process measurements in hydrometallurgical zinc production processes Turtinen, Markus (2007) Learning and recognizing texture characteristics using local binary patterns Sarpola, Arja (2007) The hydrolysis of aluminium, a mass spectrometric study Keski-Sntti, Jarmo (2007) Neural networks in the production optimization of a kraft pulp bleach plant Hamada, Atef Saad (2007) Manufacturing, mechanical properties and corrosion behaviour of high-Mn TWIP steels Rahtu, Esa (2007) A multiscale framework for affine invariant pattern recognition and registration

Book orders: OULU UNIVERSITY PRESS P.O. Box 8200, FI-90014 University of Oulu, Finland

Distributed by OULU UNIVERSITY LIBRARY P.O. Box 7500, FI-90014 University of Oulu, Finland

UNIVERSITY OF OULU P.O. Box 7500 FI-90014 UNIVERSITY OF OULU FINLAND

A C TA

U N I V E R S I TAT I S

O U L U E N S I S

S E R I E S

E D I T O R S

A B C D E F G

SCIENTIAE RERUM NATURALIUM Professor Mikko Siponen HUMANIORA

Professor Harri Mantila

TECHNICA

Professor Juha Kostamovaara

MEDICA

Professor Olli Vuolteenaho

SCIENTIAE RERUM SOCIALIUM Senior Assistant Timo Latomaa SCRIPTA ACADEMICA Communications Officer Elna Stjerna OECONOMICA Senior Lecturer Seppo Eriksson

EDITOR IN CHIEF

Professor Olli Vuolteenaho

EDITORIAL SECRETARY Publications Editor Kirsti Nurkkala

ISBN 978-951-42-8607-0 (Paperback) ISBN 978-951-42-8608-7 (PDF) ISSN 0355-3213 (Print) ISSN 1796-2226 (Online)