Bioresource Technology 102 (2011) 9111–9120

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

A novel mechanism and kinetic model to explain enhanced xylose yields from
dilute sulfuric acid compared to hydrothermal pretreatment of corn stover
Jiacheng Shen a,c,⇑, Charles E. Wyman a,b
a
Center for Environmental Research and Technology, Bourns College of Engineering, University of California, 1084 Columbia Avenue, Riverside, CA 92507, USA
b
Chemical and Environmental Engineering Department, Bourns College of Engineering, University of California, 1084 Columbia Avenue, Riverside, CA 92507, USA
c
Northwest Irrigation and Soils Research Laboratory, Agricultural Research Service, United States Department of Agriculture, 3793 N. 3600 E. Kimberly, ID 83341, USA

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

Article history: Pretreatment of corn stover in 0.5% sulfuric acid at 160 °C for 40 min realized a maximum monomeric
Received 14 January 2011 plus oligomeric xylose yield of 93.1% compared to a maximum of only 71.5% for hydrothermal (no added
Received in revised form 31 March 2011 mineral acid) pretreatment at 180 °C for 30 min. To explain differences in dilute acid and hydrothermal
Accepted 1 April 2011
yields, a fast reacting xylan fraction (0.0889) was assumed to be able to directly form monomeric xylose
Available online 7 April 2011
while a slow reacting portion (0.9111) must first form oligomers during hydrothermal pretreatment. Two
reactions to oligomers were proposed: reversible from fast reacting xylan and irreversible from slow
Keywords:
reacting xylan. A kinetic model and its analytical solution simulated xylan removal data well for dilute
Corn stover
Dilute sulfuric acid
acid and hydrothermal pretreatment of corn stover. These results suggested that autocatalytic reactions
Hydrothermal pretreatment from xylan to furfural in hydrothermal pretreatment were controlled by oligomeric xylose decomposi-
Kinetic model tion, while acid–catalytic reactions in dilute acid pretreatment were controlled by monomeric xylose
Xylose decomposition.
Ó 2011 Published by Elsevier Ltd.

1. Introduction controlled pH liquid hot water (Mosier et al., 2005). A vital consid-
eration is that conditions to maximize the total yield of xylose from
Corn stover is an abundant agricultural residue in the United hemicellulose and glucose from cellulose in the combined opera-
States (about 75 million dry tones in 2005) (http://pdf.wri.org/ tions of pretreatment and enzymatic hydrolysis are different from
corn_stover_for_ethanol_production.pdf), and has significant poten- those that maximize the yield of either sugar alone (Lloyd and
tial as a feedstock for conversion to liquid fuels such as ethanol and Wyman, 2005).
biodiesel (Kadam and McMillan, 2003). However, for biological con- In addition to enhancing sugar yields for fermentation to etha-
version, corn stover, like most lignocellulosic materials, must first be nol and other products, pretreatment can generate reactive com-
pretreated to increase cellulose accessibility by cellulase (Wyman pounds from which to make other products from the cellulose
et al., 2005a,b). Hydrothermal pretreatment, in which biomass is and hemicellulose portions of biomass. For example, Jae et al.
treated with either hot-water or steam without adding a mineral (2010) recently proposed a process to utilize xylan, glucan, and lig-
catalyst, is attractive due to the simplicity and minimization of nin in biomasses to produce furfural, 5-(Hydroxymethyl)furfural
chemical additives, but yields tend to be limited. However, dilute (HMF), and pyrolysis products, respectively. In such an approach,
sulfuric acid pretreatment has proven to be a very effective in recov- the pretreatment objective would be to maximize xylose yields
ering most of the hemicellulose sugars from corn stover as well as and minimize glucose losses to maximize furfural and HMF yields.
other biomasses during pretreatment, while exposing cellulose to In addition, this application also requires minimizing lignin losses
enzymes for high yields from subsequent hydrolysis (Wyman during pretreatment to maximize availability for downstream
et al., 2005a, 2008). Some recent applications of these pretreatments pyrolysis.
to corn stover include dilute-sulfuric acid in a high-solids percola- Kinetic models are valuable for testing mechanisms and process
tion reactor (Zhu et al., 2005), steam (Varga et al., 2004), and design, and several have been applied to pretreatment to describe
hemicellulose release and degradation such as xylan decomposi-
tion during dilute acid pretreatment of cellulosic biomass
⇑ Corresponding author. Address: Northwest Irrigation and Soils Research (Jacobsen and Wyman, 2000). Some models assumed that hemicel-
Laboratory, Agricultural Research Service, United States Department of Agriculture,
lulose follows homogeneous reaction kinetics to form xylose and
3793 N. 3600 E. Kimberly, ID 83341, USA. Tel.: +1 208 423 6565; fax: +1 208 423
6555. other sugars (Aguilar et al., 2002; Bustos et al., 2003; Morinelly
E-mail address: [email protected] (J. Shen). et al., 2009). Another type of hemicellulose decomposition model

0960-8524/$ - see front matter Ó 2011 Published by Elsevier Ltd.

doi:10.1016/j.biortech.2011.04.001
9112 J. Shen, C.E. Wyman / Bioresource Technology 102 (2011) 9111–9120

Nomenclature

A pre-exponential factors for reaction (L/(mol min)) Rh modified reaction ordinate including the effect of proton
cp glucose or xylose concentration measured by HPLC (g/L) (dimensionless)
Ea activation energies for reaction (J/mol) r conversion factor from glucose and xylose to glucan and
FX concentration of fast reacting xylan (mol/L) xylan (0.9 and 0.88)
FSX total amount of xylan in corn stover SX concentration of slow reacting xylan (mol/L)
FU concentration of furfural (mol/L) T reaction temperature (K)
fFX fraction of fast reacting xylan in the total xylan (dimen- Tb is a base temperature (100 °C)
sionless) Tr pretreatment temperature (°C)
fSX fraction of slow reacting xylan in the total xylan ts reaction time (s)
(dimensionless) vp liquid volume including moisture of the raw corn stover
fxsp xylan fractions in the pretreated residue (dimension- (L)
less) XY concentration of monomeric xylose (mol/L)
f0 total carbohydrate content (glucan plus xylan) in the yFU furfural yield (dimensionless)
raw corn stover (dimensionless) yFX fraction of the remaining fast reacting xylan (dimen-
H concentration of proton (mol/L) sionless)
K first-order rate constant for reaction integrating the pro- yOSX yield of oligomeric xylose (dimensionless)
ton concentration (min1) ySX fraction of slow reacting xylan in the total xylan
00
k rate constant not including the proton concentration (L/ (dimensionless)
(mol min)) yXY yield of monomeric xylose (dimensionless)
m0 weight of the dried corn stover added to the reactor (g) yp sugar yields (dimensionless)
mp weight of pretreated dry corn stover solid (g) ysp solid yield (dimensionless)
OFX concentrations of oligomeric xylose formed by the fast
xylan reaction (mol/L) Subscript
OSX concentration of oligomeric xylose formed by the slow I reaction number
xylan reaction (mol/L) 0 initial state
R gas constant (J/(mol K))

partially accounted for the heterogeneous structure of hemicellu- hemicellulose hydrolysis based on these fast and slow reactions,
lose in biomass by considering two types of hemicelluloses, one and (4) determine parameters for the kinetic model to best de-
of which reacts faster than the other (Kobayashi and Sakai, scribe experimental hemicellulose hydrolysis data for both hydro-
1956), and several investigators have applied this type of model thermal and dilute acid pretreatment of corn stover
(Jacobsen and Wyman, 2002; Schell et al., 2003; Lu and Mosier,
2008). The fast and slow fractions in these models were postulated
to form oligomeric xylose and then monomeric xylose (called just 2. Modeling hemicellulosic hydrolysis kinetics
xylose in following text) and finally furfural by irreversible, first-
order reactions, as in Eq. (1): 2.1. Definition and determination of initial fractions of fast and slow
reacting hemicelluloses
Hemicellulose
(fast) Although the concept has been applied for several decades
Oligomeric Xylose Furfural
ð1Þ (Kobayashi and Sakai, 1956), the fast and slow reacting hemicellu-
xylose C5H10O5 C5H4O2
Hemicellulose loses in biomass have not been distinctly defined. Because of com-
(slow) plex hemicellulose structure and limitations in analytical methods
to understand its structure, it is still difficult to identify the fast and
However, this model does not distinguish differences in reaction slow reacting hemicelluloses at the molecular level. In a review,
mechanisms between fast and slow reacting hemicelluloses, i.e., Wyman et al. (2005c) reported that the fast and slow fractions of
both are irreversible first-order. Hemicelluloses in biomass are hemicellulose were around 65% and 35%, respectively, for many
polysaccharides with a backbone of pentoses to which are attached biomasses, based on the observation that the hydrolysis rate slo-
different side-groups of hexoses and pentoses. As the side-groups wed when the hemicellulose fraction dropped below 35%. Maloney
are hydrolytically removed, the longer chained xylans and others et al. (1985) determined the percent of fast reacting xylan in white
in the backbone can be hydrolyzed more effectively. These side- birch treated by sulfuric acid as between 59.5% and 72.4% by fitting
groups may directly form monomeric sugars. the data to their model. A summary of data by Esteghlalian et al.
In this paper, data was developed on xylose release from corn (1997) showed the fast reacting xylan fraction to be between
stover hemicellulose for both hydrothermal and dilute sulfuric acid 67% and 84.3%. Based on their experimental data, Esteghlalian
pretreatments to support development of a novel mechanism and et al. (1997) also calculated the fast reacting xylan portion to make
kinetic model that could explain why dilute acid gives higher xy- up 64.4%, 76.8%, and 83.8% of the total xylan in corn stover, switch-
lose yields. On that basis, our objectives for this study were to grass, and poplar, respectively, by fitting the data to a model that
(1) experimentally measure sugar release from corn stover in assumed direct conversion of xylan to xylose without considering
hydrothermal and dilute acid pretreatments to maximize total xy- oligomer intermediates.
lose yields for subsequent furfural production or fermentation to It may be noted that the values of fast and slow reacting xylan
ethanol or other products, (2) determine the proportions of fast fractions obtained by fitting of constants were not consistent with
and slow reacting hemicelluloses in corn stover according to our experimental break points in the decomposition of two types of xy-
proposed definition, (3) develop kinetic rate expressions for lans. For example, Kim and Lee (1987) and Chen et al. (1996)
 J. Shen, C.E. Wyman / Bioresource Technology 102 (2011) 9111–9120 9113

reported the average percent of fast reacting xylan in hardwood by an autocatalytic mechanism plus the portion that remains in
was 69.7% using best-fit technology, but the break point between the solid hemicellulose in biomass. Therefore, our definition of
the two straight lines from their experiments was between 2% fast reacting and slow reacting hemicelluloses combined may
and 4%, leaving open the possibility that the fraction of fast react- be equivalent to the fast reacting hemicellulose defined by other
ing xylan in hardwood was between 96% and 98%. Kobayashi and investigators, and our unreacted hemicellulose may be equiva-
Sakai (1956), the originators of the two fraction model, reported lent to the slow reacting hemicellulose defined by others. In
the break points for dilute sulfuric acid pretreatment of hardwood any event, our definition provides a clear approach to determine
were lower than 10%. the fast and slow reacting hemicellulose fractions in biomass.
In addition to this inconsistency, two additional considerations Based on these definitions and our hydrothermal experimental
led us to not apply the best-fit approach to estimate the fractions data, the percent of fast and slow reacting xylan in the corn stover
of fast and slow reacting xylans (corresponding to the parameters we used were 8.9% and 91.1%, respectively. Fig. 1 presents the fast
fFX,0 and fSX,0 (= 1  fFX,0) in Eqs. (21)–(25)). First, because fFX,0 and and slow reacting portions of xylans in the raw corn stover and
fSX,0 are properties of biomass, which should be initial conditions their changes during hydrothermal and dilute acid pretreatments.
in the model, and not kinetic parameters (such as rate constants), The sum of the fast plus slow reacting fractions of xylan that even-
they should not be derived from a fitting program because the model tually reacted equals 0.715 (= 0.0889 + 0.633), leaving 0.285 as the
cannot be used to determine initial conditions, but determined by unreacted fraction, close to the values for the data (0.65 and 0.35)
experimental data. Second, increasing the number of model param- summarized by Wyman et al. (2005c).
eters to be fit to the data in addition to the rate constants, for exam-
ple, from 4 parameters in our model to 5 parameters, increases the 2.2. Development of a hemicellulose hydrolysis kinetic model
uncertainty in the fitted parameters values.
In this paper, we hypothesize that the response of hemicellu- Because the fast reacting hemicellulose reacts more rapidly
lose to different environments could be used as an indicator of than the slow reacting portion, some smaller macromolecules
hemicellulose structure. In this context, two catalytic mecha- of fast reacting hemicelluloses, such as side groups, may directly
nisms can breakdown hemicellulose: hydrothermal (water only) form monomers such as xylose by a irreversible first-order reac-
pretreatment in which protons released from acetyl groups in tion, and some larger macromolecules may first form oligomers
biomass and from water at high temperatures promote hemicel- such as oligomeric xylose by a reversible first-order reaction.
lulose hydrolysis, and dilute acid pretreatment in which protons These oligomers could quickly equilibrate with the larger macro-
released from added mineral acids accelerate hydrolysis. Hydro- molecules, and then further breakdown to monomers such as
thermal pretreatment of hemicellulose in corn stover forms xylose by an irreversible first-order reaction that becomes rate-
three products: monomeric sugars such as xylose, a much larger controlling. Therefore, our kinetic model assumes that chemical
portion of oligomeric sugars such as oligomeric xylose, and unre- equilibrium is rapidly reached between the fast reacting hemi-
acted hemicelluloses left in the biomass. Most of the sugar re- celluloses in biomass and the oligomers, such as oligomeric xy-
leased from dilute acid pretreatment is as monomers. We lose, by a pseudo-homogeneous, reversible first-order reaction.
categorized the portion released easily in hydrothermal pretreat- Further conversion from oligomers to monomers, such as xylose,
ment as the fast reacting hemicellulose. On the other hand, the follows an irreversible first-order reaction, which is slower than
extra hemicellulose sugars released by dilute acid were consid- the reversible reaction, and controls the reaction rate. The slow
ered as coming from the slow reacting fraction. The hemicellu- reacting hemicellulose is converted into mono-molecular com-
lose remaining in the solids after dilute acid pretreatment pounds, such as xylose, through oligomers, such as oligomeric
were taken to be recalcitrant (unreacted) hemicellulose. Thus, xylose, by two irreversible first-order reactions. Reactions from
we can defined the fast reacting hemicelluloses to be the portion monomers, such as xylose, to decomposition products, such as
that directly forms monomeric sugars (reaction 20 in Eq. (2) dot- furfural, are taken to be irreversible first-order reactions. The
ted line) by an autocatalytic mechanism, and the slow reacting proton concentration that catalyzes reactions is integrated into
hemicelluloses to be the portion that reacts to oligomeric sugars all rate constants. Another key assumption is that resistance to
mass transfer in the interior of biomass particles and mass trans-
fer from the bulk liquid to the solid surface is small for the agi-
120
tated reactors used at high reaction temperatures.
Itwouldbedesirabletoapplythebelowreactionmechanismtothe
100
entire hemicellulose and not just the xylan portion. However, we fo-
cused onjust determiningthefast and slow reactingxylan,oligomeric
80 xylose, xylose, and furfural in the following text because only xylose
63.3 anditsderivativescanbemeasuredaccuratelyenoughformeaningful
60 81.4 analysis,andtheresultsarelikelyagoodproxyforfastandslowreacting
91.11 hemicellulosesandtheiroligomers,mono-compounds,anddecompo-
40
sitionproducts.Inanyevent,thefollowingreactionmechanismresults
fromtheseassumptions(Eq.(2)):
28.5
20 r1, H+, k1
9.72 Oligomeric
Xylan (fast)
8.89 8.98 8.98 xylose (fast)
(Solid)
0 r-1, H+, k-1
Raw CS HWR AR r2, H+, k2
k2 ’
Fig. 1. Fast and slow reacting fractions of xylan in raw corn stover and their r4, H+, k4
changes in residues following pretreatments. Column 1: black: fast reacting xylan; r3, H+, k3 r5, H+, k5
white: slow reacting xylan in raw corn stover (Raw CS). Column 2: black: reacted Oligome- Xylose
Xylan (slow) ric xylose C5H10O5 Furfural
fast reacting xylan; white: remaining slow reacting xylan; texture: reacted slow
(Solid) (slow) C5H4O2
reacting xylan in residue after hydrothermal pretreatment (HWR). Column 3: black:
reacted fast reacting xylan; white: remaining slow reacting xylan; texture: reacted ð2Þ
slow reacting xylan in residue after 0.5% sulfuric acid pretreatment (AR).
9114 J. Shen, C.E. Wyman / Bioresource Technology 102 (2011) 9111–9120

 
According to the law of mass action, the rate equations for these k4 k3 SX;0 expðk3 tÞ  expðk5 tÞ expðk4 tÞ  expðk5 tÞ
XY ¼ 
reactions can be, respectively, expressed as k4  k3 k5  k3 k5  k4
0
k2 F X;0  0 
dF X þ 0 expðk2 tÞ  expðk5 tÞ
¼ k1 F X þ k1 OFX ð3Þ k5  k2
dt
dOFX ð16Þ
¼ k1 F X  k1 OFX  k2 OFX ð4Þ
dt 8 h i 9
expðk5 tÞ
dSX k5 k4 k3 SX;0 < k5 k3
1
k5
 expðk
k3
3 tÞ =
¼ k3 SX ð5Þ FU ¼ h i
dt k4  k3 :  1 expðk 5 tÞ
 k4 4 ;
expðk tÞ
k5 k4 k5
dOSX " #
¼ k3 SX  k4 OSX ð6Þ 0 0
dt k5 k2 F X;0 expðk5 tÞ expðk2 tÞ
dX Y þ 0  0 ð17Þ
¼ k4 OSX þ k2 OFX  k5 X Y ð7Þ k5  k2 k5 k2
dt
dF U The total amount of xylan FSX0 in biomass initially must equal the
¼ k5 X Y ð8Þ
dt sum of the initial fast and slow reacting fractions:
 
00 Ea FSX;0 ¼ F X;0 þ SX;0 ð18Þ
ki ¼ ki H ¼ Ai H exp  i¼15 ð9Þ
RT
Thus, the fractions fFX,0 and fSX,0 of fast and slow reacting xylans in
where FX, SX, OFX, OSX, XY, FU, and H are the concentrations of fast the total xylan in biomass at an initial time, were, respectively, de-
reacting xylan, slow reacting xylan, oligomeric xylose formed by fined as:
the fast xylan reaction, oligomeric xylose formed by the slow xylan F X;0 or SX;0
reaction, xylose, furfural, and protons, respectively, all in mol/L. The fFX;0 or SX;0 ¼ ð19Þ
F X;0 þ SX;0
first order rate constants for each reaction are k1, k1, k2, k3, k4, and
k5 in units of min1 with the proton concentrations integrated into As the reaction proceeds, the fractions of the remaining fast reacting
00
each. On the other hand, ki represents the corresponding rate con- xylan as yFX, and slow reacting xylan as ySX, in the total xylan of bio-
stants that do not include the proton concentration, with units of L/ mass, and yields of oligomers as yOSX, xylose as yXY, and furfural as
(mol min) and subscripts i = 1–5 corresponding to the appropriate yFU can be, respectively, expressed as:
reaction step in expression (4). Furthermore, Ai and Eai are the
F X or SX or OSX or X Y or F U
Arrhenius pre-exponential factors and the activation energies for yFX; or SX; or OSX or XY or FU ¼ ð20Þ
F X;0 þ SX;0
each reaction with units of L/(mol min) and J/mol, respectively.
Finally, R is the gas constant with units of J/(mol K), and T is the
Therefore, the corresponding fractions of fast and slow reacting
reaction temperature in K. Because reaction 2 is assumed to be
xylans, and yields of oligomeric xyloses, xylose, and furfural can be
rate-controlling step, the pseudo steady state hypothesis can be
determined at any time by dividing the results from Eqs. (13)–(17)
applied to Eq. (4), and substituting the result into Eqs. (3) and (7),
by the total amount of xylan contained initially in the raw biomass.
respectively, results in:
0
yFX ¼ fFX;0 expðk2 tÞ ð21Þ
dF X 0
¼ k1 F X ð10Þ
dt ySX ¼ fSX;0 expðk3 tÞ ð22Þ
dX Y 0
¼ k4 OSX þ k2 F X  k5 X Y ð11Þ
dt k3 fSX;0
yOSX ¼ ½expðk3 tÞ  expðk4 tÞ ð23Þ
k4  k3
where
 
k4 k3 fSX;0 expðk3 tÞ  expðk5 tÞ expðk4 tÞ  expðk5 tÞ
k1 k2 yXY ¼ 
0 0
k1 ¼ k2 ¼ ð12Þ k4  k3 k5  k3 k5  k4
k1 þ k2 0
k2 fFX;0 0
þ 0 ½expðk2 tÞ  expðk5 tÞ
Application of the pseudo-steady state hypothesis to the revers- k5  k2
ible first-order reaction makes the two-step reaction equivalent to ð24Þ
an irreversible first-order reaction. Therefore, both pathways of di-
rect and indirect formation of monomers can be combined as one k5 k4 k3 fSX;0 n 1 hexpðk5 tÞ expðk3 tÞi 1
h
expðk5 tÞ expðk4 tÞ
io
yFU ¼ k5 k3 k5
 k3  k5 k k
 k
pathway (dotted line), as in Eq. (2). Moreover, an analytical solu- k4  k3 4 5 4
" #
tion is more easily developed after application of the pseudo- 0
k5 k2 fFX;0 expðk5 tÞ expðk2 tÞ
0

steady state hypothesis. þ 0  0

k5  k2 k5 k2
Integration of Eqs. (5), (6), (8), (10), and (11), in that order, with
the initial conditions FX = FX,0, SX = SX,0, and OSX = XY = 0 at t = 0 ð25Þ
results in the analytical expressions given by Eqs. (13)–(17) to
describe fast reacting xylan, slow reacting xylan, oligomeric xylose,
3. Methods
xylose, and furfural concentrations over time.

0
3.1. Feedstock and reagents
F X ¼ F X;0 expðk2 tÞ ð13Þ
Raw corn stover supplied by the National Renewable Energy
SX ¼ SX;0 expðk3 tÞ ð14Þ Laboratory (NREL) to the Biomass Refining Consortium for Applied
Fundamentals and Innovation (CAFI) was used as the feedstock,
k3 SX;0 and sulfuric acid was purchased from Sigma–Aldrich (St. Louis,
OSX ¼ ½expðk3 tÞ  expðk4 tÞ ð15Þ
k4  k3 MO). The composition of the raw corn stover as determined in
 J. Shen, C.E. Wyman / Bioresource Technology 102 (2011) 9111–9120 9115

our laboratory by standard NREL methods was 36.1% glucan, 21.4% The rest of the experimental procedures were the same as the
xylan, 1.72% arabinan, 0.93% galactan, 13.6% lignin, 2.8% ash, and hydrothermal pretreatment described above.The sugar yields yp
23.5% others. D-(+)-glucose and xylose (both Sigma–Aldrich, (dimensionless) from pretreatment was defined by the following
99.5%) were used as standard materials for HPLC analysis. Calcium equation:
carbonate (Fisher Scientific, 99.0%) was used to neutralize acid in rcp v p
samples before HPLC analysis. yp ¼ ð26Þ
m0 f0

3.2. Pretreatment of corn stover in hydrothermal and dilute sulfuric where cp is the glucose or xylose concentration measured by HPLC
acid (g/L), vp is the liquid volume including moisture of the raw corn sto-
ver (L), m0 is the weight of the dried corn stover added to the reactor
The raw corn stover was ground to less than 1 mm using a lab- (g), r is the conversion factor from glucose and xylose to glucan and
oratory mill with an internal sieve (model 4, Arthur H. Thomas xylan (0.9 and 0.88), respectively, and f0 is the total carbohydrate
Company, Philadelphia, PA). The milled corn stover was stored in content (glucan plus xylan) in the raw corn stover (dimensionless).
plastic bags in a freezer at 18 °C until used for pretreatments. The solid yield yps (dimensionless) from pretreatment is defined as:
A one-liter stainless steel Parr reactor with two stacked pitched mp
yps ¼ ð27Þ
blade impellers (Model 4525, Parr Instruments Company, Molline, m0
Illinois) was used for monomeric and oligomeric sugar production
where mp is the weight (g) of pretreated dry corn stover solids fol-
from the raw corn stover by hydrothermal and dilute sulfuric acid
lowing pretreatment for a particular time.
pretreatments. The reactor was heated in 4 kW fluidized sand
The modified severity factor including pH was calculated to
baths (Model SBL-2D, Techne Co., Princeton, NJ), and the internal
facilitate comparison of the effect of various dilute sulfuric acid
temperature was monitored with a type K thermocouple (Omega
pretreatment conditions on sugar yields (Chum et al., 1990):
CASS-18U-12, Omega Engineering Co., Stamford, CT). For hydro-
Z ts  
thermal pretreatments, typically, 50 g of dry corn stover and Tr  Tb
logðRh Þ ¼ log exp dt þ logðHþ Þ
650 mL of deionized water were loaded into the Parr reactor to 0 14:75
provide a total suspension volume of 700 mL in the reactor, and   
Tr  Tb
the Parr reactor was sealed. The temperature of the sand bath ¼ log ts exp  pH ð28Þ
14:75
was taken to 450 °C, and the Parr reactor was lowered into it. After
about 5 min, the temperature of the Parr reactor approached to where Rh is the modified reaction ordinate including the effect of
within 5 °C of the target temperature, and the Parr reactor was protons, ts is the reaction time (s), Tr is the pretreatment tempera-
raised to keep the temperature stable at the target temperature, ture (°C), and Tb is a base temperature (100 °C). In Eq. (28) it was as-
at which point the reaction time was set as zero. When the desired sumed that the pretreatment temperature Tr is constant, i.e., the
reaction time was reached, the Parr reactor was quickly moved to a time for temperature rise could be ignored in the initial pretreat-
water bath to cool it and stop the reaction. When ambient pressure ment period for heating-up reactor. Final pH values measured at
was reached, the Parr reactor was opened, and the pretreated corn room temperature after pretreatment were used in Eq. (28).
stover slurry was filtered through filter paper (Fisher Brand Glass
Fiber Filter Circles). The filtrates were analyzed by HPLC as 3.3. Analysis of components in the liquid fraction
described below. The pretreated solids were transferred to a glass
filter and washed three times with hot-deionized water. The Xylose and glucose concentrations in the liquid fraction were
washed solids were measured for moisture using an automatic sys- measured using an Agilent HPLC equipped with an RI detector
tem (HB43-S, Mettler Toledo) and weighed to calculate solid yields. and a Bio-Rad Aminex HPX-87H column (Bio-Rad Laboratories,
For dilute sulfuric acid pretreatment, the milled corn stover was Hercules, CA) at 65 °C. The mobile phase was 0.005 M sulfuric acid
soaked in a predetermined solution volume with the required with a flow rate of 0.6 mL/min. The identities of the compounds
dilute sulfuric acid concentration (for example, 0.5 wt.%) overnight. were authenticated by comparing their retention times with those

7
A 30 B

6
25

5
20
Concentration (g/L)

Concentration (g/L)

4
15
3

10
2

5
1

0 0
0 5 10 15 20 25 30 35 40 45
0 5 10 15 20 25 30 35 40 45
Time (min)
Time (min)

Fig. 2. Total xylose (A) and glucose (B) concentrations over reaction time. Closed points show results before post hydrolysis, and open points show the results after post
hydrolysis, with the difference being oligomers. Square points: hydrothermal at 180 °C; rhombus: 0.5% sulfuric acid at 160 °C; and triangles: 0.5% sulfuric acid at 180 °C.
9116 J. Shen, C.E. Wyman / Bioresource Technology 102 (2011) 9111–9120

of pure compounds (Sigma–Aldrich, St. Louis, MO). Before injecting and slow reacting xylans as a function of time during pretreatment
into the HPLC, the acid in the filtrates was neutralized with calcium (yFX, and ySX in Eqs. (21) and (22) vs. time). For dilute acid pretreat-
carbonate to pH 5–6, and then the suspensions were centrifuged at ment, these could be directly calculated because the xylan frac-
15,000 rpm for 5 min through a 0.2 lm member filter. The super- tions fxsp (Fig. 4 A) in the pretreated residues and solid yields ysp
natants were used for HPLC analysis. (Fig. 4 E) were measured vs. time, i.e.,

0:0889f xsp ysp

3.4. Analysis of components in the solids yFX ¼ ð29Þ
FSX0
The components in the solids were analyzed according to NREL 0:911f xsp ysp
ySX ¼ ð30Þ
standard procedures (Selig et al., 2008; Sluiter et al., 2008, 2005) FSX0
with the carbohydrates expressed in terms of the corresponding
in which FSX0 is equal to 15.2 for hydrothermal pretreatment and
sugars. Briefly, about 0.3 g of the pretreated residue or the raw corn
21.4 for dilute acid pretreatment at 160 °C (0.214 of xylan fraction
stover was added to 3 mL of 72 wt.% sulfuric acid, and the suspen-
in the raw corn stover times the initial corn stover weight in the
sion was hydrolyzed at 50 °C for 1 h. Then, the slurry was diluted to
Parr). For hydrothermal pretreatment, yFX and ySX were calculated
4 wt.% sulfuric acid by adding 84 mL deionized water followed by
by subtracting monomeric and oligomeric xylose yields vs. time
autoclaving at 121 °C for 1 h. After cooling, the slurry was filtered
from the fast and slow reacting xylans (0.0889 and 0.9111) in the
through a crucible, and the filtrate was used in HPLC analysis. A
raw corn stover (Fig. 4 C), i.e.,
Waters HPLC model 2695 system equipped with a 2414 refractive
detector, a Waters 2695 autosampler, a Bio-Rad Aminex HPX-87P yFX ¼ 0:0889  monomeric xylose yield in liquid fraction ð31Þ
column (Bio-Rad Laboratories, Hercules, CA), and Millenium32 ySX ¼ 1  0:0889  oligomeric xylose yield in liquid fraction ð32Þ
chromatography manager 3.2 software (Waters Co., Milford, MA)
were employed for analyzing the compounds. The column temper- The yields of oligomeric xylose required to be fitted in Eq. (23) were
ature was 85 °C, and the mobile phase was deionized water at a determined by the xylose yields after post hydrolysis minus the xy-
flow rate of 0.6 mL/min. lose yields before post hydrolysis.

3.5. Post hydrolysis 4. Results and discussion

After pretreatments of the raw corn stover, 72% sulfuric acid 4.1. Effect of pretreatment time and temperature on concentrations
was added to samples of the liquid fractions derived from both and yields of glucose and xylose
hydrothermal and dilute acid pretreatments to make a 4% acid
solution according to NREL standard procedures (Sluiter et al., The raw corn stover was pretreated at the following three
2006). The solution was autoclaved at 121 °C for 1 h and further conditions: (1) hydrothermal (just water without acid added) at
prepared for HPLC analysis of the sugars as described above. a solids concentration of 71 g/L and temperature of 180 °C, (2)
100 g/L solids with 0.5 wt.% sulfuric acid at 160 °C, and (3) 71 g/L
3.6. Determination of the remaining fractions of fast and slow reacting solids with 0.5 wt.% sulfuric acid at 180 °C. Fig. 2 presents xylose
xylans over the pretreatment period and glucose concentrations before and after post hydrolysis plotted
against time for these three cases. As expected, the relative concen-
An important aspect of modeling xylan decomposition based on trations of glucose were low, while the total xylose (xylose plus
experimental data was to determine the remaining fractions of fast oligomeric xylose) concentrations for both pretreatments were

5
A 100 93.1
B
91.4
89.2
90 87.8 4.5

80 4
71.5

70 3.5
pH and Severity factor

60 3
Yield (%)

50 2.5

40 2

30 1.5

20 1
13 14.1
12.4
10.9
10 6.48 6.23
0.5
0.682
0 0
hot water bp hot water ap acid 160oC bp acid 160oC ap acid 180oC bp acid 180oC ap
0 5 10 15 20 25 30 35 40 45
Pretreatement condition Time (min)

Fig. 3. (A) Yields of total glucose and xylose at the optimal conditions. Black bars refer to glucose, and white bars refer to xylose. Columns 1 and 2: hydrothermal at 180 °C and
30 min before post hydrolysis (bp). Columns 3 and 4: hydrothermal at 180 °C and 30 min after post hydrolysis (ap). Columns 5 and 6: 0.5% sulfuric acid at 160 °C and 40 min
before post hydrolysis. Columns 7 and 8: 0.5% sulfuric acid at 160 °C and 40 min after post hydrolysis. Columns 9 and 10: 0.5% sulfuric acid at 180 °C and 10 min before post
hydrolysis. Columns 11 and 12: 0.5% sulfuric acid at 180 °C and 10 min after post hydrolysis. (B) Final pH values and severity factors vs. time at three pretreatment conditions:
hydrothermal at 180 °C and 30 min (squares); 0.5% sulfuric acid at 160 °C and 40 min (rhombus); 0.5% sulfuric acid at 180 °C and 20 min (triangle). Solid symbols refer to
measured pH, and open symbols are the severity factor.
 J. Shen, C.E. Wyman / Bioresource Technology 102 (2011) 9111–9120 9117

higher because hemicellulose in corn stover, like most plants, was 0.5% sulfuric acid increased the maximum xylose concentration
more easily decomposed than cellulose. For hydrothermal to a much higher value of 25.2 g/L for reaction at 160 °C for
pretreatment, the maximum total xylose concentration was about 40 min and 17.1 g/L after pretreatment for 10 min at 180 °C. Three
13.4 g/L following reaction for 30 min at 180 °C. However, adding reasons could explain the lower xylose concentration at 180 °C

A B

C D

Fig. 4. (A and B) Change in composition of corn stover solids following pretreatment with 0.5% sulfuric acid at 160 °C vs. time: (A) squares: glucan; rhombus: xylose; (B):
triangle: lignin; multi symbol: ash; cross: others. (C and D) The change in composition of corn stover solids following pretreatment with 0.5% sulfuric acid at 180 °C vs. time:
(C) squares: glucan; rhombus: xylose; (D) triangle: lignin; multi symbol: ash; cross: others. (E) Solid yields following three pretreatments: squares: hydrothermal
pretreatment at 180 °C; rhombus: 0.5% sulfuric acid at 160 °C; triangle: 0.5% sulfuric acid at 180 °C.
9118 J. Shen, C.E. Wyman / Bioresource Technology 102 (2011) 9111–9120

than at 160 °C: (1) a higher initial corn stover concentration of pretreatment and 0.5% sulfuric acid pretreatment at the two tem-
100 g/L at 160 °C vs. 71 g/L at 180 °C, which accounts for most peratures. Increasing pretreatment time reduced the final pH for all
of the difference; (2) differences in the modified severity factor three cases, likely due to greater proton release from acetyl com-
of 3.29 for 40 min at 160 °C vs. 2.77 for 10 min at 180 °C; and (3) pounds in hemicellulose. The final proton concentrations in the li-
more degradation products at higher temperatures even for the quid fractions for the three cases increased by 1.86, 1.05, and 1.29
same severity. The times to achieve the maximum total glucose times from 10 to 40 min, respectively. Hydrothermal pretreatment
(monomer plus oligomer) concentrations (10 min for hydrother- was more sensitive to longer reaction times than dilute acid pre-
mal, 40 min for dilute acid at 160 °C, and 20 min for dilute acid treatment because the former depended on proton release from
at 180 °C) for both hydrothermal and dilute acid pretreatments acetyl compounds and not from external proton addition.
were different than those for xylose (30 min for hydrothermal,
20 min for dilute acid at 160 °C, and 15 min for dilute acid at 4.2. Composition of corn stover solids pretreated with 0.5% sulfuric
180 °C). acid at 160 and 180 °C
Fig. 3 A summarizes the highest xylose yields as well as glucose
yields at the same times, before and after post hydrolysis for Fig. 4 presents the composition of acid pretreated corn stover
hydrothermal and dilute acid pretreatments; as noted above, these solids at 160 °C (A and B) and 180 °C (C and D) over the reaction
do not correspond to the highest glucose yields due to the different times. Generally, xylan (xylan fraction fxsp) and lignin contents in
optimal times for xylose and glucose. From this figure, we can see the pretreated corn stover solids at both temperatures decreased
that the highest total xylose yield for hydrothermal pretreatment and increased monotonically, respectively, with increasing
was about 71.5% and the highest total glucose yield was 6.51% pretreatment time for the period studied due to (1) significant
(data not shown in Fig. 3), with about 91% of the total xylose and release of hemicellulose, mainly xylan, from corn stover into the
65% of the total glucose as oligomers. However, oligomers made liquid phase; (2) limited release of glucan; (3) limited lignin loss;
much lower contributions to the total xylose and glucose yields and (4) likely formation of pseudo-lignin by degradation. The max-
for dilute acid pretreatment (5.6% and 16% for 160 °C, and 2.4% imum glucan contents in the solids occurred at intermediate
and 12% for 180 °C, respectively). Furthermore, dilute acid pro- pretreatment times (62.5% at 160 °C for 25 min and 63.8% at
duced higher total sugar yields than hydrothermal pretreatment, 180 °C for 15 min), and the highest lignin contents were 24.5% at
with the highest total xylose and glucose yields being 93.1% and 160 °C for 40 min and 25.6% at 180 °C for 20 min. Higher tempera-
13% at 40 min and 160 °C and the highest total xylose and glucose ture was more effective in xylan removal from corn stover, with
yields being 91.4% at 10 min and 17.6% (data not shown in Fig. 3) at almost all of the xylan removed at 180 °C after 10 min (corre-
20 min and 180 °C. The maximum xylose yield of 93.1% from dilute sponding to a log of the modified severity factor of 3.36), while
acid pretreatment was close to those (90% and 95%) reported by xylan was still detected in the pretreated solids at 160 °C even
Lloyd and Wyman (2005) and Tucker et al. (2003), respectively. for 40 min of reaction time (a log of the modified severity factor
However, our highest xylose yield of 71.5% from hydrothermal of 3.39). This outcome implies that temperature was the most
pretreatment is much greater than the 46.8% value reported in significant factor for xylan removal among the three parameters
the limited study of these conditions by Lloyd and Wyman of temperature, time, and proton concentration that comprise the
(2005). The reason may be that they used steam explosion for modified severity factor (Eq. (28)) and that the same severity does
hydrothermal pretreatment, an open system in which, some xylose not necessarily realize the same pretreatment effect.
might be released the air due to the sudden drop in pressure, while Fig. 4E presents the total solid yields for the three pretreat-
we used a closed Parr reactor. In addition, increasing temperature ments run. The 60.8% solid yield from hydrothermal pretreatment
reduced reaction times needed to reach maximum xylose yields at 30 min was higher than any of the total solid yields for the two
because higher temperature accelerated hemicellulose degrada- dilute acid pretreatments. In addition, the solid yields for lower
tion. These results confirmed that the maximum xylose yield is temperature (160 °C) pretreatment with dilute acid was higher
governed by both temperature and time, as expected. than for dilute acid pretreatment at the higher temperature
Fig. 3 B shows final pH values measured after pretreatment and (180 °C). Both outcomes can be attributed to sulfuric acid and high
severity factors calculated from Eq. (28) vs. time for hydrothermal temperature increasing solubilization of corn stover.

53.4 kg solid
60.8 kg solid 31.1 kg glucan
A 33.0 kg glucan B 1.61 kg xylan
5.36 kg xylan 13.1 kg liginin
13.5 kg liginin 3.12 kg ash
2.54 kg ash Parr reactor, 10% solids,
Parr reactor, 7.1% solids, 4.50 kg others
6.45 kg others 160oC, 40 min
180oC, 30 min
3 1 3
1 Hot Water 0.5% H2SO4
Pretreatment
Pretreatment
100 kg corn stover 100 kg corn stover
36.1 kg glucan 36.1 kg glucan
21.4 kg xylan 2 Liquid 21.4 kg xylan 2 Liquid
13.6 kg lignin 13.6 kg lignin
0.274 kg glucose 3.98 kg glucose
2.8 kg ash 2.8 kg ash
0.23 kg oligomeric glucose 0.754 kg oligomeric glucose
24.3 kg others 24.3 kg others
1.58 kg xylose 21.3 kg xylose
Glucose yield 6.24% 15.8 kg oligomeric xylose Glucose yield 11.8% 1.29 kg oligomeric xylose
Xylose yield 71.5% Xylose yield 93.1%
Glucan closure 97.5% Glucan closure 98.4%
Xylan closure 96.6% Xylan closure 100.6%

Fig. 5. Xylan and glucan mass balances for corn stover pretreatment for hydrothermal at 180 °C (A) and 0.5% sulfuric acid at 160 °C (B).
 J. Shen, C.E. Wyman / Bioresource Technology 102 (2011) 9111–9120 9119

100
A B 100
90 90

80 80

70 70

60 60
Yield (%)

Yield (%)
50 50

40 40

30 30

20 20

10 10

0 0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Time (min) Time (min)

Fig. 6. Experimental data and model results for xylan decomposition by 180 °C hydrothermal (A) and 160 °C, 0.5% sulfuric acid (B) pretreatments. Fast reacting xylan: data
(solid points) and model (solid lines); slow reacting xylan: data (crosses) and model (dot lines); oligomeric xylose: data (squares) and models (dot/dash lines); xylose
monomer: data (rhombus) and models (dashed lines).

4.3. Mass balances of sugars from pretreated corn stover ments and Table 1 reporting the fitted parameters for each
pretreatment. The rate constant k4 in Table 1 is the smallest among
Fig. 5 presents mass balances for hydrothermal pretreatment of all four rate constants for pretreatment without added acid, imply-
corn stover at 180 °C for 30 min (A) and 0.5% sulfuric acid pretreat- ing that reaction of oligomeric xylose to monomers controlled the
ment at 160 °C for 40 min (B), both being conditions that realized rate of xylan conversion to furfural for hydrothermal pretreatment.
the highest xylose release. Based on projecting our results to a However, the rate constant k4 in Table 1 is greater than k5 for dilute
basis of 100 kg of dry corn stover, 17.4 kg of total xylose (1.58 kg acid pretreatment, suggesting the acid addition resulted in the
of monomeric xylose and 15.8 kg of oligomeric xyloses), 2.50 kg rate-controlling step conversion to xylose decomposition to furfu-
of total glucose (0.274 kg of monomeric glucose and 2.23 kg of ral from oligomeric xylose decomposition in hydrothermal
oligomeric glucose), and 60.8 kg of total solids would result from pretreatment, which explains the reason of higher yield of dilute
hydrothermal pretreatment at 180 °C. Application of 0.5% sulfuric acid pretreatment than hydrothermal pretreatment from reaction
acid at 160 °C increased the amounts released into solution to mechanism. The results also demonstrate that autocatalysis
22.6 kg of total xylose (21.3 kg of monomeric xylose and 1.29 kg occurred in hydrothermal pretreatment has a different rate-
of oligomeric xylose) and 4.73 kg of total glucose (3.98 kg of mono- controlling mechanism with acid–catalysis occurred in dilute acid
meric glucose and 0.754 kg of oligomeric glucose), while reducing pretreatment. Furthermore, the rate constant k4 (0.0021 min1) for
the total solids left to 53.4 kg. Thus, dilute acid pretreatment pro- hydrothermal pretreatment was only about 1/100 of that
duced much higher glucose and xylose yields (11.8% and 93.1%) (0.2165 min1) for dilute acid pretreatment, suggesting that reac-
than hydrothermal pretreatment (6.24% and 71.5%). Weight losses tion from oligomeric xylose to xylose was much slower for hydro-
of 20.9% and 20.2% in the overall mass balances of the solid and li- thermal than for dilute acid pretreatment and further providing a
quid fractions combined were also measured for hydrothermal and quantitative explaining of why dilute acid pretreatment gives
dilute acid, respectively, possibly because the filter paper was not higher yields than hydrothermal pretreatment. The rate constant
0
fine enough to capture all the particles in the slurry after pretreat- k2 (0.090 min1 for hydrothermal pretreatment and 0.4413 min1
ment. In addition, some solids were adsorbed in the filter paper for dilute acid pretreatment) was greater than k3 and k4 (0.0375
and funnel wall during filtration and could not be collected. An- and 0.0021 min1 for hydrothermal pretreatment, and 0.1177
other possible cause of losses could be lack of detection by HPLC and 0.2165 min1 for dilute acid pretreatment), confirming that
of unknown compounds formed by decomposition of hemicellu-
lose and cellulose. Despite the losses in overall solids, most of
the xylan and glucan in the raw corn stover fed to pretreatment Table 1
Parameters for models in Eqs. (21)–(24) for hydrothermal pretreatment at 180 °C and
could be accounted for, with deviations of only 3.4% for xylan dilute acid pretreatment at 160 °C.
and 2.5% for glucan for hydrothermal pretreatment and 0.6% of
xylan and 2.1% of glucan for 160 °C dilute acid conditions in Rate constant k2 (min1) k3 (min1) k4 (min1) k5 (min1) SSD

Fig. 5. Hydrothermal 0.090 0.0375 0.0021 0.0156 552

Dilute acid 0.4413 0.1177 0.2165 0.0042 141

4.4. Simulation of fast and slow reacting xylan hydrolysis k2

00 00
k3
00
k4 k5
00

L/(mol min) L/(mol min) L/(mol min) L/(mol min)

Once the fractions of the remaining fast and slow xylan (yFX and Hydrothermal 1033 431 24.1 179
ySX), and the yields of oligomers and xylose (yOSX and yXY) were Dilute acid 25.4 6.77 12.4 0.242
known over time from the experiments reported above, the rate SSD: sum of square deviation.
constants in Eqs. (21)–(24) were fit to these data, with Fig. 6 show- The initial fractions of fast and slow reacting hemicelluloses are 0.0889 and 0.9111,
ing the results for hydrothermal (A) and dilute acid (B) pretreat- respectively.
9120 J. Shen, C.E. Wyman / Bioresource Technology 102 (2011) 9111–9120

0
the pathway represented by k2 is the reaction path for the fast Jacobsen, S.E., Wyman, C.E., 2000. Cellulose and hemicellulose hydrolysis models for
application to current and novel pretreatment processes. Applied Biochemistry
reacting xylan. On the other hand, the value of the rate constant
0 and Biotechnology 84 (6), 81–96.
k2 for dilute acid pretreatment is 4.9 times that for hydrothermal Jacobsen, S.E., Wyman, C.E., 2002. Xylose monomer and oligomer yields for
pretreatment, meaning that the reaction rate for the path from uncatalyzed hydrolysis of sugarcane bagasse hemicellulose at varying solids
xylan to xylose was faster for dilute acid than for hydrothermal concentration. Industrial & Engineering Chemistry Research 41, 1454–1461.
00 Jae, J.H., Tompsett, G.A., Lin, Y.C., Carlson, T.R., Shen, J.C., Zhang, T.Y., Yang, B.,
pretreatment. Interestingly, the k2 value that excludes the proton Wyman, C.E., Conner, W.C., Huber, G.W., 2010. Depolymerization of
concentration with dilute acid was only 2.5% (25.4/1033%) of that lignocellulosic biomass to fuel precursors: maximizing carbon efficiency by
for hydrothermal conditions. From this, we could conclude that combining hydrolysis with pyrolysis. Energy & Environmental Science 3, 358–
365.
external protons supplied by dilute acid are the primary reason Kadam, K.L., McMillan, J.D., 2003. Availability of corn stover as a sustainable
that reaction 20 is so much faster for dilute acid pretreatment. feedstock for bioethanol production. Bioresource Technology 88, 17–25.
Kim, S.B., Lee, Y.Y., 1987. Kinetics in acid–catalyzed hydrolysis of hardwood
hemicellulose. Biotechnology and Bioengineering Symposium 17, 71–84.
Kobayashi, T., Sakai, Y., 1956. Hydrolysis rate of pentosan of hardwood in dilute
5. Conclusions sulfuric acid. Bulletin of the Agricultural Chemical Society of Japan 20, 1–7.
Lloyd, T.A., Wyman, C.E., 2005. Combined sugar yields for dilute sulfuric acid
The maximum total xylose yield was only 71.5% after 30 min for pretreatment of corn stover followed by enzymatic hydrolysis of the remaining
solids. Bioresource Technology 96, 1967–1977.
hydrothermal pretreatment of corn stover in a one-liter Parr reac-
Lu, Y.L., Mosier, N.S., 2008. Kinetic modeling analysis of maleic acid–catalyzed
tor at 180 °C. However, the maximum total xylose yield increased hemicellulose hydrolysis in corn stover. Biotechnology and Bioengineering 101,
significantly to 93.1% for pretreatment with 0.5% sulfuric acid at 1170–1181.
160 °C for 40 min. A model with a fast xylan reaction to monomeric Maloney, M.T., Chapman, T.W., Baker, A.J., 1985. Dilute acid–hydrolysis of paper
birch – kinetics studies of xylan and acetyl-group hydrolysis. Biotechnology and
xylose and a slow xylan reaction to monomeric xylose via oligo- Bioengineering 27, 355–361.
meric xylose could fit experimental xylan release data well for both Morinelly, J.E., Jensen, J.R., Browne, M., Co, T.B., Shonnard, D.R., 2009. Kinetic
hydrothermal and dilute acid pretreatments. The results predicted characterization of xylose monomer and oligomer concentrations during dilute
acid pretreatment of lignocellulosic biomass from forests and switchgrass.
that xylose formation from oligomers was rate-controlling for Industrial & Engineering Chemistry Research 48, 9877–9884.
hydrothermal pretreatment, while xylose decomposition to furfu- Mosier, N., Hendrickson, R., Ho, N., Sedlak, M., Ladisch, M.R., 2005. Optimization of
ral was rate-controlling for dilute acid pretreatment. pH controlled liquid hot water pretreatment of corn stover. Bioresource
Technology 96, 1986–1993.
Schell, D.J., Farmer, J., Newman, M., McMillan, J.D., 2003. Dilute-sulfuric acid
pretreatment of corn stover in pilot-scale reactor – investigation of yields,
Acknowledgements kinetics, and enzymatic digestibilities of solids. Applied Biochemistry and
Biotechnology 105, 69–85.
We thank the Defense Advanced Research Projects Agency Selig, M., Weiss, N., Ji, Y., 2008. In: Laboratory Analytical Procedure No. TP-510-
42629. National Renewable Energy Laboratory, Golden, CO.
(DARPA) and LOGOS Technologies for providing financial support Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., Crocker, D., 2008.
for this project. We are also grateful to the Center for Environmen- In: Laboratory Analytical Procedure No. TP-510-42618. National Renewable
tal Research and Technology of the Bourns College of Engineering Energy Laboratory, Golden, CO.
Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., 2005. In:
(CE-CERT) at the University of California, Riverside (UCR) for
Laboratory Analytical Procedures No. TP-510-42622. National Renewable
providing key equipment and facilities. Dr. Wyman is particularly Energy Laboratory, Golden, CO.
grateful to the Ford Motor Company for funding the Chair in Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., 2006. In:
Laboratory Analytical Procedures No. TP-510-42623. National Renewable
CE-CERT that augments support for many projects such as this.
Energy Laboratory, Golden, CO.
The views, opinions, and/or findings contained in this article/pre- Tucker, M.P., Kim, K.H., Newman, M.M., Nguyen, Q.A., 2003. Effects of temperature
sentation are those of the author/presenter and should not be and moisture on dilute-acid steam explosion pretreatment of corn stover and
interpreted as representing the official views or policies, either cellulase enzyme digestibility. Applied Biochemistry and Biotechnology 105 (1–
3), 165–178.
expressed or implied, of the Defense Advanced Research Projects Varga, E., Reczey, K., Zacchi, G., 2004. Optimization of steam pretreatment of corn
Agency or the Department of Defense. Approved for Public Release, stover to enhance enzymatic digestibility. Applied Biochemistry and
Distribution Unlimited. Biotechnology 113, 509–523.
Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y., 2005a.
Coordinated development of leading biomass pretreatment technologies.
Bioresource Technology 96, 1959–1966.
References Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y., 2005b.
Comparative sugar recovery data from laboratory scale application of leading
Aguilar, R., Ramirez, J.A., Garrote, G., Vazquez, M., 2002. Kinetic study of the acid pretreatment technologies to corn stover. Bioresource Technology 96, 2026–
hydrolysis of sugar cane bagasse. Journal of Food Engineering 55, 309–318. 2032.
Bustos, G., Ramirez, J.A., Garrote, G., Vazquez, M., 2003. Modeling of the hydrolysis Wyman, C.E., Decker, S.R., Himmel, M.E., Brady, J.W., Skopec, C.E., Viikari, L., 2005c.
of sugar cane bagasse with hydrochloric acid. Applied Biochemistry and Hydrolysis of cellulose and hemicellulose. In: Dumitriu, S. (Ed.),
Biotechnology 104, 51–68. Polysaccharides: Structural Diversity and Functional Versatility. CRC Press,
Chen, R.F., Lee, Y.Y., Torget, R., 1996. Kinetic and modeling investigation on two- Boca Raton, FL, pp. 995–1034.
stage reverse-flow reactor as applied to dilute-acid pretreatment of agricultural Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y.,
residues. Applied Biochemistry and Biotechnology 57–8, 133–146. Mitchinson, C., Saddler, J.N., 2008. Comparative sugar recovery and
Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment catalyst fermentation data following pretreatment of poplar wood by leading
effects and the combined severity parameter. Applied Biochemistry and technologies. Biotechnology Progress 25, 333–339.
Biotechnology 24–25, 1–14. Zhu, Y.M., Lee, Y.Y., Elander, R.T., 2005. Optimization of dilute-acid pretreatment of
Esteghlalian, A., Hashimoto, A.G., Fenske, J.J., Penner, M.H., 1997. Modeling and corn stover using a high-solids percolation reactor. Applied Biochemistry and
optimization of the dilute-sulfuric-acid pretreatment of corn stover, poplar and Biotechnology 121, 1045–1054.
switchgrass. Bioresource Technology 59, 129–136.