Biotechnology Advances 27 (2009) 542–550

Contents lists available at ScienceDirect

Biotechnology Advances
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / b i o t e c h a d v

Research review paper

Water-based woody biorefinery

Thomas E. Amidon, Shijie Liu ⁎
Department of Paper and Bioprocess Engineering, SUNY College of Environmental Science and Forestry, 1 Forestry Drive, Syracuse, NY 13210, United States

a r t i c l e i n f o a b s t r a c t

Available online 22 April 2009 The conversion of biomass into chemicals and energy is essential in order to sustain our present way of life.
Fossil fuels are currently the predominant energy source, but fossil deposits are limited and not renewable.
Keywords: Biomass is a reliable potential source of materials, chemicals and energy that can be replenished to keep pace
Acetic acid with our needs. A biorefinery is a concept for the collection of processes used to convert biomass into
Biorefinery materials, chemicals and energy. The biorefinery is a “catch and release” method for using carbon that is
Cellulose
beneficial to both the environment and the economy. In this study, we discuss three elements of a wood-
Hemicellulose
Hydrolysis
based biorefinery, as proposed by the SUNY College of Environmental Science and Forestry (ESF): hot-water
Hot-water extraction extraction, hydrolysis, and membrane separation/concentration. Hemicelluloses are the most easily
Nano-filtration separable main component of woody biomass and thus form the bulk of the extracts obtained by hot-
Woody biomass water extraction of woody biomass. Hot-water extraction is an important step in the processes of woody
Xylan biomass and product generation, replacing alternative costly pre-treatment methods. The hydrolysis of
Xylose hemicelluloses produces 5-carbon sugars (mainly xylose), 6-carbon sugars (mainly glucose and mannose),
and acetic acid. The use of nano-filtration membranes is an efficient technology that can be employed to
fractionate hot-water extracts and wood hydrolysate. The residual solid mass after hot-water extraction has a
higher energy content and contains fewer easily degradable components. This allows for more efficient
subsequent processing to convert cellulose and lignin into conventional products.
© 2009 Elsevier Inc. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542
2. SUNY College of Environmental Science and Forestry (ESF) biorefinery: water-based technology. . . . . . . . . . . . . . . . . . . . . . . 543
3. Hot-water extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545
4. Hydrolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547
5. Separation/purification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 548
6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 550
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 550
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 550

1. Introduction
Food, energy and materials are the basic needs of humans and Fig. 1
illustrates the flow from various sources to meet these requirements.
Compared with the available solar energy, our other primary energy
sources—geothermal, nuclear, petroleum, coal, natural gas, and minerals—
are limited. Biomass was historically the primary source for all human
needs, but we have learned to utilize fossil energy sources, such as
petroleum, coal and natural gas, to significantly increase our standard of
living.
⁎ Corresponding author.
E-mail address: sliu@esf.edu (S. Liu).
Fossil carbon has become the dominant energy and chemical
source for mankind since the industrial revolution (Amidon et al.,
2008). However, there is only a finite amount of available fossil energy.
World Oil (EIA, 2006) puts the known recoverable reserves, as of
January 1, 2006, at 1,119.615 billion barrels for crude oil; 6,226.555 tril-
lion cubic feet for natural gas; and 997,748 million tons for coal. Klass
(1998) estimated that world crude could last until 2027, based solely
on the 1990s proven conventional reserve. His most optimistic
estimate was that the oil could last until 2100, based on the current
growth in energy consumption and assuming that other potential oil
sources, such as new discoveries and oil sands, are developed. Using
the longest time estimate, it would take less than 280 years to
completely deplete all the oil that ever existed on Earth. We must

0734-9750/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.biotechadv.2009.04.012
 T.E. Amidon, S. Liu / Biotechnology Advances 27 (2009) 542–550
Fig. 1. Energy and material flows to meet human needs.
therefore look for alternatives to petroleum, since the time scale
required for petroleum to recycle or naturally replenish is, at best, in
the order of 280 million years (Table 1). Fossil fuel replenishes when a
carbonaceous age occurs on the Earth, resulting in the preservation of
a large amount of organic matter, but the chances of a high carbon-
storing event occurring while higher life forms survive is unlikely.
Nevertheless, the amount of reserves is too small relative to the
280 million-year rotation, even assuming mankind could survive the
fossil replenishment era. Thus, petroleum and fossil energy resources
are considered to be effectively nonrenewable. Societal awareness of
environmental impacts, as well as problems in stability and sustain-
ability of energy supply, has made the development and implementa-
tion of bio-based chemicals/energy urgent. Domestic energy security
and rural economies could both benefit from a plant-derived
chemical/energy economic base.
Biomass has been an important energy source since the beginning
of civilization. Ligno-cellulosic biomass is the most abundant organic
source on earth, with an annual production of about 170 billion metric
tons (Klass, 1998). Tapping into the chemical energy of biomass and
restoring its historically important position in energy generation and
transportation is imperative to the sustainability of the world
economy. Forests cover about 9.5% of the Earth's surface, or about
32% of the land area, but account for 89.3% of the total standing
biomass and 73 billion metric tons/year, or 42.9%, of the total annual
biomass production. Savanna and grasslands come second, accounting
for 11% of total biomass production. In energy terms, forests alone
could produce 1030 quadrillion British Thermal Unit (Btu)/year
(Klass, 1998), which is equivalent to more than double the world's
total primary energy consumption of about 460 quadrillion Btu in
2005 (EIA, 2003). Research and development is required to boost the
energy conversion efficiency from plant and woody biomass to meet
industrial and residential energy and commodity chemical require-
ments. Plants synthesize chemicals from solar energy, as shown in
Fig. 2. The chemical energy stored by the biomass can be converted
into energy and chemicals that can then be utilized, also shown in
Fig. 2. The use of dedicated or managed energy crops (agriforest and/
or agricultural biomass) could further increase the available biomass
(Keoleian and Volk, 2005; Ragauskas et al., 2006).
Renewable forest material is carbon neutral, i.e. utilizing forest
material will not create a carbon imbalance over the life cycle of the
forest, which is an extended 5–80 years for managed forests, as shown
in Table 1. “Catch and release” is the key to biomass utilization. Carbon
dioxide is drawn from the atmosphere to allow the plants to grow,
while planting, management, the conversion of biomass into bio-
products, and the utilization and decomposition of bio-products will
all produce carbon dioxide. In an optimally-balanced operation,
carbon dioxide is simply being recycled through plant growth and
bio-products, as illustrated in Fig. 2 (Liu et al., 2006). The net effect is
that solar energy and atmospheric carbon dioxide can be converted
into energy and materials that can be utilized by humans.
543
Because the nonrenewable energy sources predominantly used
today will cease to be available in the future, it is necessary to consider
expanding the production and use of plant biomass, especially forest
material, as a sustainable energy and chemical source (Amidon,
2002). The biorefinery is a manufacturing concept for converting plant
biomass into energy and chemicals, in which plant biomass can be
separated into a number of compounds and energy is recovered.
Utilization of the heterogeneous, bonded composite structure is a key
for the future supply of chemicals and energy. The use of renewable
carbon eliminates the fossil-derived carbon dioxide burden to the
environment and reduces the greenhouse gas (GHG)-driven global
climate change. Managed forests thus have significant potential for
reducing GHG emissions through conversion of the forest material
into liquid fuels, electricity and other products that are currently
derived from nonrenewable carbon.
2. SUNY College of Environmental Science and Forestry (ESF)
biorefinery: water-based technology
In addition to its direct utilization as a building material, the major
uses of wood today are for making paper and for generating energy by
burning/combustion. Large volumes of wood, in the order of
1000 metric tons/day or more, are consumed at many industrial
sites. In a pulp mill, residual wood chips and shavings from nearby
lumber mills and/or wood chips from round wood are either
chemically or mechanically disintegrated into fibers. In a chemical
pulp (Kraft) mill, aqueous NaOH and Na2S solutions are used to cook
the wood chips. Lignin and a large fraction of the hemicelluloses are
dissolved in the aqueous phase, referred to as black liquor. The black
liquor is then burned to generate energy and recover the NaOH and
Na2S. A pulp mill thus already acts as a biorefinery, producing energy
and cellulosic fibers from wood and transporting them to consumers.
In a combined heat and power (CHP) facility, wood or woody
biomass is burned or combusted to generate thermal energy (steam)
and power (electricity). Both steam and electricity are then delivered
to other industrial installations. Like a Kraft pulp mill, a CHP plant is
also a simplistic biorefinery. With an already established large woody
biomass handling capacity, it is possible to envision the use of a CHP
model as a starting point for the development of the next generation
of biorefineries.
Xylan, or pentosan (i.e. 5-C sugar polymer), is the dominant
component in the hemicellulose fraction of angiosperm woody
biomass, typically comprising 20–35% of the dry wood mass. This is
currently an under-utilized renewable resource that has potential
utility for the production of bio-based fuels, chemicals and polymeric
materials (Liu et al., 2006; Amidon, 2006; Stipanovic et al., 2006), as
shown in Fig. 3. Xylan is a polymer made of β-xylopyranose units
linked through (1→4)-glycosidic bonds, where arabinose, acetyl
groups, and uronic acids are also present as lateral constituents, as
shown in Table 2. The three major monomers of xylan are xylose,
arabinose and acetic acid. Xylan is the most easily separable of the
three major components of hardwood biomass: cellulose (glucan),
hemicellulose (xylan), and lignin. Lignin has the highest heating
value, and cellulose is of the highest current commodity value (as
fibers). Therefore, the development of an alternative use for xylan is
n potentially attractive starting point for a more sophisticated
Table 1
Recycling (or renewing) times for chemical feedstocks.
Chemical Recycling time
Algae 1 month
Agricultural crops 3 month–1 year
Grasses 1 year
Shrubs 1–5 years
Trees 5–80 years
Oil, gas and coal 280 million years
544 T.E. Amidon, S. Liu / Biotechnology Advances 27 (2009) 542–550

Fig. 2. “Catch and release”: the photosynthesis-consumption cycle by which CO2 is drawn from atmosphere and water, mostly from soil, generating O2 while producing biomass using
energy from sunlight. Then the biomass can then be reacted with O2 to produce CO2 and H2O. This cyclic process produces no net CO2 or H2O.

biorefinery process. Through SUNY ESF, the expansion of a Kraft pulp
mill and/or CHP plant to convert xylan into commercial ethanol and
biodegradable plastics has become the main focus of the biorefinery
effort in New York. Hot-water extraction prior to chemical pulping or
burning of woodchips allows the potentially profitable utilization and
commercialization of the wood extraction procedure and is expected
to improve the niche market advantage of the pulps produced (Kraft
pulp mill), or the economics of the green electrical power produced
(CHP plant).
Kraft pulping is the dominant chemical pulping process used
today. Under alkaline conditions, the net heating values of the
carbohydrates are reduced due to the consumption of alkali to
neutralize the acetyl groups. The recoverable heating value as energy
is also reduced by the amount of energy required to convert
hydroxides to oxides during combustion. Degraded hemicelluloses
and cellulose in the black liquor have a heating value of 13.5 MJ/kg,
which is about half of that of lignin (25 MJ/kg) (Vakkilainen, 2000).
Due to the alkali-consuming degradation of carbohydrates, it requires
more NaOH to maintain the required level of alkalinity during Kraft
pulping, even though the degraded carbohydrates contribute less net
chemical energy to the recovery process than their mass implies.
Therefore, removing hemicelluloses before pulping could improve the
energy efficiency of a Kraft pulp mill.
There are 5–9 main components in the ESF biorefinery (Amidon
et al., 2008; Liu et al., 2006; Amidon, 2006): 1) hot-water extraction; 2)
hydrolysis of hot-water extracts; 3) separation of xylan, sugars and
acetic acid; 4) fermentation of sugars to ethanol, butanol, bioplastics,
other products; 5a) pulping of hot water-extracted woodchips followed

Fig. 3. Schematic of biorefinery.

 T.E. Amidon, S. Liu / Biotechnology Advances 27 (2009) 542–550 545

Table 2
Major components of hemicellulose and extractives (compiled from Hakkila, 2000 and Fengel and Wegener, 1989).

G: Glucose unit, Ga: Galactose unit, M: Mannose unit, X: Xylose unit, A: Arabinose unit, Ac: Acetyl group (CH3CO), Gu: 4-O-methyl-glucuronic acid. All the identified connections (\)
are through \O\.

by bleaching and papermaking; 5b) burning of the hot water-extracted
residual woody biomass and organics for CHP use; 5c) use of extracted
woodchips/woody biomass for reconstituted wood products such as
particle board; 5d) use of the extracted woody biomass to produce wood
fuel pellets; or 5e) further separation/hydrolysis of the extracted woody
biomass to platform chemicals, such as aromatics and wood sugars.
These components mean that the process can be readily adapted to
existing industrial operations and can eventually lead to purpose-built
biorefineries. This study focused on the first three components.
3. Hot-water extraction
Woody biomass is composed of four main components: extractives,
hemicellulose, lignin and cellulose. Cellulose provides the structure and

Fig. 4. Distribution of main wood components before and after hot-water extraction of sugar maple woodchips. “Unidentified” represents the sum of the sugars that have been
converted into compounds that were not identified during the chemical analysis and the minor components derived from woody biomass: such as methanol, formic acid and ash
(Amidon et al., 2008).
546 T.E. Amidon, S. Liu / Biotechnology Advances 27 (2009) 542–550
Fig. 5. pH and temperature profile during hot-water extraction.
strength, while hemicellulose and lignin provide bonding to the
structure. Extractives are extractable compounds of the woody biomass
that can be readily dissolved with organic solvents or water at room
temperature and under atmospheric conditions. Inorganic compounds
are also present in the woody biomass. There are over 70 metal, earth
elements and inorganic compounds in woody biomass, of which K and
Ca are the main elements, followed by Mg and P (Hakkila, 2000). In
hardwood, for example, Ca comprises 0.08–0.2% of dry stem wood mass
and 0.85–3.05% stem bark mass. Mg comprises 0.02–0.04% of dry hard
wood stem mass and 0.07–0.11% dry hardwood stem bark mass. Table 2
shows the major components in the extractive and hemicellulose wood
fractions. The extraction of the readily hydrolysable carbohydrates with
hot water in the absence of added mineral acids or bases facilitates the
utilization and recovery of the hydrolyzate components. When the
heating values of the residual chips are to be recovered after extraction,
there is no reduction in value due to NaOH or metal hydroxides, and no
increase in corrosion due to mineral acids. Furthermore, adverse
Fig. 6. Variation of dissolved solids with autocatalytic extraction time for sugar maple
wood chips in a 1.84 m3-digester. The temperature profile is shown in Fig. 5.
Fig. 7. Effect of extraction temperature on mass removal for sugar maple wood chips.
The experimental data were collected using a 6-l M/K® digester and the water:wood
chip ratio was 5.92:1.
environmental and recovery effects are avoided because no NaOH or
Na is added to the process streams and no by-products from mineral acid
neutralization are produced from the extraction process.
The hot-water extraction process can be catalyzed by bases, acids
or enzymes. Without the addition of NaOH, hydrolyzed hemicelluloses
produce acetic acid, as well as polysaccharides and many other minor
components. Fig. 4 shows the typical mass distributions resulting from
the hot-water extraction of sugar maple wood chips. Cellulose
(glucan) and lignin (klason lignin) were mostly retained by the
residual wood chips, whereas the other components were mostly
found in the extraction liquor. The carbohydrate compositions were
measured by sulfuric acid hydrolysis followed by 1H nuclear magnetic
resonance (NMR).
Fig. 5 shows the variations in pH during hot-water extraction of
sugar maple wood chips in a 1.84 m3 digester with a wood:liquor ratio
of about 1:4. The digester was heated by direct steam injection. The pH
values were measured in liquor samples taken from the digester and
cooled to 30 °C. The pH profile in the extraction liquor dropped from
the initial, near-neutral conditions (around pH 6.3) to acidic
conditions. The desired extraction temperature of 160 °C was reached
at 150 min, which coincided with the leveling off of the pH values.
These acidic conditions during the hot-water extraction process
further catalyzed the extraction and hydrolysis reactions. The hot-
water extraction reactions were therefore autocatalytic. The extrac-
tion initially proceeded slowly and became faster as the pH fell during
Fig. 8. Variations in acetic acid and polymer acetyl group concentrations in the wood
extract with autocatalytic extraction time for sugar maple wood chips in a 1.84 m3-
digester. The temperature profile is shown in Fig. 5.
 T.E. Amidon, S. Liu / Biotechnology Advances 27 (2009) 542–550
Fig. 9. Reducing sugar concentration as a function of hydrolysis time for a concentrated
wood extract subjected to various xylanase products (B, C and D) when pH values were
not adjusted.
the extraction. Efficient reactors could be designed to perform the hot-
water extraction operations using existing wood chip processing units
as model systems.
Fig. 6 shows the concentrations of dissolved solids in the extraction
liquor (or extract) as a function of the extraction time. The
concentration of dissolved solids increased with increasing duration
of extraction. The initial increase in concentration of dissolved solids
was slow during the first 100 min of extraction, but increased more
quickly thereafter. As shown in Fig. 7, the total mass removal from
wood chips increased with increasing extraction time and tempera-
ture. The hot-water extraction proceeded faster and more components
were dissolved at higher extraction temperatures. Extraction effi-
ciency increased as the extraction time was extended.
Fig. 8 shows the concentrations of acetic acid and acetyl groups
associated with polysaccharides in the wood extract as functions of
hot-water extraction time in a 1.84 m3 digester. As seen in Fig. 8, the
acetic acid concentration increased with extraction time from the start
of the hot-water extraction. However, no measurable acetylated
polysaccharides were detected until after almost 100 min, at which
time the liquor pH was close to the final pH, as shown in Fig. 5. The
concentration of acetyl groups quickly overtook that of acetic acid in
the wood extract, though the acetic acid concentration was much
higher, as shown in Fig. 8, at 0.053 mol/L. There were nearly three
acetic acid molecules dissolved in the liquor for every one xylose
molecule dissolved in the wood extract, but the concentration of
Fig. 10. Reducing sugar concentration as a function of hydrolysis time for a concentrated
wood extract subjected to various xylanase products (B, C and D) when the pH values
were adjusted to 5.5.
547
Fig. 11. Reducing sugar concentration as a function of hydrolysis time for a concentrated
wood extract subjected to various xylanase products (B, C and D) when the pH values
were adjusted to 7.0.
acetyl dissolved, but associated with polysaccharides, was higher than
that of acetic acid.
The wood extracts obtained from the hot-water extraction process
consisted mainly of xylo-oligmers, sugars, and acetic acid. The longer-
chain xylo-oligmers (or xylan) could be separated for use as
biopolymers. Hydrolysis or saccharification to produce monosacchar-
ides or sugars is required to further utilize the xylo-oligmers as a
fermentable feedstock.
4. Hydrolysis
The wood extracts after hot-water extraction consisted of mono-
saccharides, polysaccharides, acetic acid, degraded lignin, and other
low-molecular weight extractable substances. Hydrolysis of oligomers
to monosaccharides or sugars could be achieved by increasing the
residence time for the wood extracts at relatively high temperatures
under acidic conditions, or by enzymatic digestion.
Figs. 9–11 show the effects of pH on the hydrolysis or reducing-
sugar end group formation as a function of time for three selected
xylanase products. More sugar end groups were formed with
increasing time. Product C appeared to perform the best overall,
while product D was advantageous when the wood extract pH was not
adjusted. Each xylanase product was added at the same rate.
Fig. 9 shows that the hydrolysis was faster with xylanase product
D, while product C catalyzed hydrolysis for the longest duration. The
hydrolysis performed with no pH adjustments produced an almost
44% increase in reducing end groups compared with the original
concentrated wood extract.
Fig. 10 shows that the hydrolysis proceeded at about the same rate
for the three xylanase products at a pH value of 5.5. However, for the
same amount (volume) of different xylanase products added, the
activities were higher for product C than for B. Product D had the
lowest activity. The extent of hydrolysis was also highest for product C,
with nearly 68% reducing-sugar end groups added to the original
concentrated wood extract after 4 days of hydrolysis. However,
Table 3
Molecular diameters of nonelectrolytes (Reid et al., 1977).
Molecular weight, kg/kmol Molecular diameter σs, 10− 10 m
10 2.9
100 6.2
1000 13.2
10,000 28.5
100,000 62.0
1,000,000 132
548 T.E. Amidon, S. Liu / Biotechnology Advances 27 (2009) 542–550
Fig. 12. A schematic of the membrane separation unit.
product D resulted in only 28% addition of the sugar end groups to the
original concentrated wood extracts after the same period.
Fig. 11 shows that the hydrolysis of all three xylanase products
proceeded very slowly at neutral pH. The best performance was
shown by product C, with a 35% reducing-sugar end group addition to
the original concentrated wood extract after 4 days of treatment.
The results shown in Figs. 9–11 suggest that the three xylanase
products were not very effective at catalyzing the hydrolysis, as the
hydrolysis process took several days to complete. Product D, however,
showed promise as it demonstrated a relatively higher activity than
the other two products when the pH values were left unadjusted.
Because of the relatively slow hydrolysis using the xylanase
products, we examined acid-catalyzed hydrolysis. Because S, N and
P are required nutrients for the growth of micro-organisms, sulfuric,
nitric, and phosphoric acid were considered to be the most suitable
acids as catalysts for the hydrolysis. It is preferable to use a low acid
concentration, and a strong acid is thus required. In this context, both
sulfuric and nitric acids are potential choices.
5. Separation/purification
Wood extracts after hot-water extraction consisted of monomeric
and oligomeric sugars (hexoses and pentoses), acetic acid, degraded
lignin (or aromatics), furfurals and other low-molecular weight
Fig. 13. Variations in xylose concentration in the concentrate and permeate streams as
functions of time during the semi-batch membrane separation.
Fig. 14. Variations in methanol concentrations, Δ in concentrate stream and ∇ in
permeate stream, as functions of time during the semi-batch membrane separation
unit.
extractable substances. Nano-filtration membranes have been used
to analyze these. The separation of sugars and acetic acid is a key step
in the biorefinery process, as acetic acid is a potential inhibitor of
fermentation processes.
The choice of membrane depends on the physical and chemical
properties of the compounds to be separated. For non-interacting
species, molecular size is the most important factor. Table 3 shows the
variation in molecular diameter with molecular weight; increasing
molecular weight increases the molecular size. When the molecular
size is greater than the membrane pore size, the molecules are not
able to permeate through and the molecular weight that corresponds
to the pore size is termed the molecular weight cut-off. Because of the
non-uniform pore size distributions in an industrial membrane, the
molecular weight cut-off refers to the corresponding average pore
size.
Fig. 12 shows a schematic of the membrane separation unit used in
our pilot plant. The raffinate (or concentrate) and permeate streams
were discharged at one end, whereas the wood extracts or hydro-
lyzates were fed into the membrane unit by a pump at the other end.
The permeate stream was collected in one tank and the raffinate
stream was returned to the feed tank.
The membrane used in this study had a molecular weight cut-off of
100 g/mol. However, filtration was not entirely governed by molecular
weight, as molecular shape and hydrophobicity can also affect
penetration. Sugars have a ring structure and thus have a relatively
large diameter. Thus, so long as the pore size is small enough to
prevent the sugar ring structure from permeating through, all the
sugars are likely to be retained in the concentrate stream. This is the
Fig. 15. Variations in furfural concentrations, Δ in concentrate stream and ∇ in permeate
stream, as functions of time during the semi-batch membrane separation unit.
 T.E. Amidon, S. Liu / Biotechnology Advances 27 (2009) 542–550
Fig. 16. Variations in hydroxylmethylfurfural concentrations, Δ in concentrate stream
and ∇ in permeate stream, as functions of time during the semi-batch membrane
separation unit.
primary consideration in selecting sugar concentrating and washing
membranes.
Fig. 13 shows the concentrations of xylose in the concentrate and the
permeate streams as a function of time during the concentration
experiments. At the start of the experiment, both the permeate and the
concentrate streams were returned to the feed tank, and no overall
concentration changes were expected. The initial time period was
employed to ensure that the system reached a steady state. The
permeate stream was directed to the permeate holding tank from
83.6 min. The lines illustrate the average concentration change. The
xylose concentrations in both the concentrate stream and the permeate
streams increased with time once the separation was started. Less than
0.7% of the xylose permeated through the membrane in this experiment.
Fig. 14 shows the concentrations of methanol in the concentrate
and permeate streams as functions of time. As seen in Fig. 14, the
concentration of methanol in the permeate stream remained
relatively unchanged, whereas the concentration in the concentrate
stream increased slightly with time, suggesting that the methanol was
nearly equally distributed between the permeate and concentrate
streams. The slightly higher methanol concentration in the concen-
trate stream could have been due to the methoxyl group in the lignin
fragments.
Fig. 15 shows the concentrations of furfural in the concentrate and
permeate streams as functions of time. There were no differences
between the concentrate and the permeate streams, within the accuracy
of the NMR, and there were no changes in concentrations with time.
Thus furfural was equally distributed between the concentrate and
permeate streams.
Fig. 17. Variations in formic acid concentrations, Δ in concentrate stream and ∇ in
permeate stream, as functions of time during the semi-batch membrane separation
unit.
549
Fig. 18. Rejection efficiencies (Rs) of xylose and aromatic compounds from the permeate
stream.
Like furfural, hydroxymethylfurfural (HMF) showed a peak around
9.5 ppm and two other peaks in the aromatic region. The peaks of HMF
were close to but separate from those of furfural. The peak near
9.5 ppm was chosen to quantify HMF. Fig. 16 shows the concentrations
of HMF in the concentrate and permeate streams as functions of time.
There were slight increases in the concentration with time, and a
slightly higher HMF concentration in the concentrate stream than in
the permeate stream. Consecutive washing is therefore expected to
wash off the HMF from the sugar or wood extract solution, through
not as effectively as furfural.
Fig. 17 shows the formic acid concentrations in the concentrate and
permeate streams. More formic acid was found in the permeate than
in the concentrate stream, suggesting preferential removal of formic
acid from the concentrate stream.
Fig. 18 shows the rejection efficiencies of xylose and aromatic
compounds from the permeate stream. They were relatively
unchanged with changes in the concentration and/or separation
time. Sugars (e.g., xylose) were mostly retained in the raffinate or
concentrate stream, as the rejection efficiency was about 0.993.
Aromatic compounds were also mostly retained, with a rejection
efficiency of 0.855.
Fig. 19 shows the rejection efficiencies of acetate, methanol,
furfural and HMF from the permeate stream. The rejection efficiencies
of these compounds were b0.5, suggesting that they could be washed
out by repeated dilution and concentration. These results indicate that
the tested membrane was able to effectively separate the sugar stream
from acetic acid, furfural, methanol and other light molecules.
Fig. 19. Rejection efficiencies (Rs) of Acetate, methanol, furfural and HMF from the
permeate stream.
550 T.E. Amidon, S. Liu / Biotechnology Advances 27 (2009) 542–550
6. Conclusions
(1) Hot-water extraction of sugar maple wood chips at 160 °C for
2 h removed approximately 23% of the woody biomass.
Cellulose and lignin mostly remained within the residual
wood chips, but more than half of the wood acetyl groups
were present in the extraction liquor. The extraction liquor pH
decreased with increasing extraction time and leveled off 1 h
after the desired extraction temperature was reached. The final
pH was about 3.5. The total measured dissolved solids in the
extract increased with increasing extraction time, especially
after the desired temperature was reached. The total mass
removal from the woody biomass increased with increasing
temperature and extraction time. Xylose was the major
component of the wood extract sugars. Acetate and acetyl-
attached to xylo-oligomers also increased with increased
extraction time. At longer extraction times, more acetyl-
attached oligomers were found in the wood extracts, compared
with free acetic acid. Volatile chemicals, such as formic acid and
methanol, increased with increasing extraction time, and their
concentrations were similar to those of the secondary mono-
meric sugars, glucose and mannose, in the wood extracts.
Furfural and volatile chemicals contributed to the difference
between the total mass removal and the total dissolved solid
measurements when oven drying the extract.
(2) Hydrolysis of wood extracts with xylanase products was less
efficient than with mineral acids. Xylanase or enzyme hydro-
lysis did not produce undesired by-products, but required
longer reaction times. More efficient xylanase products are still
required. Acid hydrolysis, however, was fast and complete,
though some by-products were produced.
(3) Nano-filtration membrane separation can be adapted to
fractionate wood extracts and/or hydrolyzates into desired
chemical components. The separation efficiency depends on
membrane performance. Using a membrane with a molecular
weight cut-off of 100, almost all the sugars were retained on
the retentate or concentrate side. Free acetate and volatile
chemicals, or small molecules such as methanol, formic acid,
and furfural, passed through the membrane almost as readily as
water. Membrane technology can therefore be utilized to purify
sugars, as well as to separate other chemicals.
Acknowledgments
The authors are thankful to NYSERDA and US DOE for financial
support for this study. The authors are indebted to the Biorefinery
Research Institute (BRI) for financial and research support.
References
Amidon TE. New forest-based materials, Part 1. Solut Mag 2002;85(4):50–2.
Amidon TE. The biorefinery in New York: woody biomass into commercial ethanol. Pulp
Pap Can 2006;107(6):47–50.
Amidon TE, Wood CD, Shupe AM, Wang Y, Graves M, Liu S. Biorefinery: conversion of
woody biomass to chemicals, energy and materials. JBMBE 2008;2(2):100–20.
Energy Information Administration (EIA). 2003, 2004 and 2006, Website: www.eia.
doe.gov/iea.
Fengel D, Wegener D. Wood: chemistry, ultrastructure, reactions. New York: Walter de
Gruyter; 1989.
Hakkila P. Chapter 4 structure and properties of wood and woody biomass. In:
Gullichsen J, Paulapuro H, editors. Forest resources and sustainable management,
papermaking science and technology. Helsinki: Fapet Oy; 2000. p. 117–82.
Keoleian GA, Volk TA. Renewable energy from willow biomass crops: life cycle energy,
environmental and economic performance. Crit Rev Plant Sci 2005;24:385–406.
Klass DL. Biomass for renewable energy, fuels, and chemicals. New York: Academic
Press; 1998.
Liu S, Amidon TE, Francis RC, Ramarao BV, Lai Y-Z, Scott GM. Ind Biotechnol 2006;2(2):113–20.
Ragauskas AJ, Williams CK, Davidson BH, Britovsek G, Cairney J, Eckert CA, et al. The
path forward for biofuels and biomaterials. Science 2006;311:484–9.
Reid RC, Prausnitz JM, Sherwood TK. The properties of gases and liquids. New York:
McGraw-Hill; 1977.
Stipanovic AJ, Haghpanah JS, Amidon TE, Scott GM, Mishra K. Opportunities for
hardwood hemicellulose in biodegradable polymer blends. In: Argyropoulos DS,
editor. Materials, chemicals and energy from forest biomass; 2006. Washington,
DC: ACS Symposium Series.
Vakkilainen E. Chapter 11 chemical recovery. In: Gullichsen J, Fogelbolm CJ, editors.
Chemical pulping, Book 6B, papermaking science and technology. Helsinki: Fapet
Oy; 2000. p. B6-B33.