The impact of dilute sulfuric acid on the selectivity of xylooligomer depolymerization to monomers

Article abstract of DOI:10.1016/j.carres.2007.10.022

The disappearance of xylose and xylooligosaccharides with degrees of polymerization (DP) ranging from 2 to 5 was followed at 160 °C with sulfuric acid added to adjust the pH from near neutral to 1.45, and the impact on the yields of lower DP xylooligomers and xylose monomer was determined. In addition, the experimental data for the disappearance of these xylooligomers was kinetically modeled assuming first-order reaction kinetics for xylose degradation and xylooligomer hydrolysis to evaluate how the pH affected the selectivity of monomer formation from xylooligomers and direct oligomer degradation to unknown products. The yield of xylose from xylooligomers increased appreciably with increasing acid concentration but decreased with increasing xylooligomer DP at a given acid concentration, resulting in more acid being required to realize the same xylose yields for higher DP species. For example, the maximum xylose yields were 49.6%, 28.0%, 13.2% and 3.2% for DP values of 2, 3, 4, and 5, respectively, at pH 4.75. Kinetic modeling revealed that all the xylooligomers disappeared at a higher rate compared to xylose monomer and the disappearance rate constant increased with DP at all pH. The kinetics for lower DP oligomers of 2 and 3 showed that these species directly degrade to unknown compounds in the absence of acid. On the other hand, higher oligomers of DP 4 and 5 exhibited negligible losses to degradation products at all pH. Therefore, only xylooligomers of DP 2 and 3 were found to directly degrade to undesired products in the absence of acid, but more work is needed to determine how higher DP species behave. This study also revealed that the source of water and the material used for the construction of the reactor impacted xylose degradation kinetics.

Full text of DOI:10.1016/j.carres.2007.10.022

Available online at www.sciencedirect.com
Carbohydrate Research 343 (2008) 290–300
The impact of dilute sulfuric acid on the selectivity of
xylooligomer depolymerization to monomers
Rajeev Kumar and Charles E. Wyman*
Thayer School of Engineering, Dartmouth College, NH 03755, USA
Received 12 July 2007; received in revised form 12 October 2007; accepted 25 October 2007
Available online 1 November 2007
Abstract—The disappearance of xylose and xylooligosaccharides with degrees of polymerization (DP) ranging from 2 to 5 was
followed at 160 ꢀC with sulfuric acid added to adjust the pH from near neutral to 1.45, and the impact on the yields of lower
DP xylooligomers and xylose monomer was determined. In addition, the experimental data for the disappearance of these xyloo-
ligomers was kinetically modeled assuming first-order reaction kinetics for xylose degradation and xylooligomer hydrolysis to eval-
uate how the pH affected the selectivity of monomer formation from xylooligomers and direct oligomer degradation to unknown
products. The yield of xylose from xylooligomers increased appreciably with increasing acid concentration but decreased with
increasing xylooligomer DP at a given acid concentration, resulting in more acid being required to realize the same xylose yields
for higher DP species. For example, the maximum xylose yields were 49.6%, 28.0%, 13.2% and 3.2% for DP values of 2, 3, 4,
and 5, respectively, at pH 4.75. Kinetic modeling revealed that all the xylooligomers disappeared at a higher rate compared to xylose
monomer and the disappearance rate constant increased with DP at all pH. The kinetics for lower DP oligomers of 2 and 3 showed
that these species directly degrade to unknown compounds in the absence of acid. On the other hand, higher oligomers of DP 4 and
5
exhibited negligible losses to degradation products at all pH. Therefore, only xylooligomers of DP 2 and 3 were found to directly
degrade to undesired products in the absence of acid, but more work is needed to determine how higher DP species behave. This
study also revealed that the source of water and the material used for the construction of the reactor impacted xylose degradation
kinetics.
ꢁ
2007 Elsevier Ltd. All rights reserved.
Keywords: Depolymerization; Degradation; Selectivity; Xylooligomers; Xylose; Yield
6
,7
1
. Introduction
highly digestible cellulose in the residual solids. How-
ever, although sulfuric acid is low in cost itself, the over-
all process is still quite expensive due to the combined
costs for the acid used, the base needed to neutralize
it, the requirements for hydrolyzate conditioning, the
expected exotic materials of construction, and the losses
in sugars during pretreatment and hydrolyzate condi-
tioning. Thus, it is valuable to better understand the role
that the acid plays in hemicellulose hydrolysis if we are
to reduce its use while still achieving high yields essential
to economic success.
Pretreatment is essential to realize the high yields vital to
economic success for biological processing of lignocellu-
losic biomass to fuels (e.g., ethanol) and chemicals, and
dilute acid treatments have been favored by many for
1
–5
near term applications.
At relatively low concentra-
tions of 0.5–4.0% and temperatures of 140–220 ꢀC,
dilute acid hydrolyzes the hemicellulose in biomass to
release primarily xylose monomers and some soluble
xylooligosaccharides with high yields and produce
Numerous studies of hemicellulose hydrolysis using
just water or with sulfuric and other acids give some
indications of the effect of acid on hemicellulose sugar
*
Corresponding author at present address: Department of Chemical
and Environmental Engineering and Center for Environmental
Research and Technology, University of California, 1084 Columbia
Avenue, Riverside, CA 92507, USA. Tel.: +1 951 781 5703; fax: +1
3
,8,9
yields.
In addition, many models have been devel-
oped for reaction kinetics considering various possible
routes to the formation and the degradation of
951 781 5790; e-mail: cewyman@engr.ucr.edu
0
008-6215/$ - see front matter ꢁ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2007.10.022

R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
291
oligomers and monomer in the complex hemicellulose
were determined. In addition, xylooligomer degradation
was modeled with and without incorporating a parallel
degradation pathway, and the influence of reactor mate-
rials and water source on xylose degradation was deter-
mined based on the concerns that the materials of
1
0,11
hydrolysis pathway.
Extensive studies of xylose deg-
radation in an acidic medium showed its loss to be con-
1
2–14
sistent with a unimolecular reaction mechanism.
However, hemicellulose hydrolysis involves a complex
insoluble solid breaking down to release numerous solu-
ble oligomers, monomers, lignin decomposition prod-
ucts, and other compounds. Thus, following the
concentrations of all the species involved is difficult, if
not impossible, limiting insight into controlling
mechanisms.
1
8
construction could impact the results.
2. Materials and methods
2.1. Materials
In an alternate approach, following the reaction of
soluble oligomers of known DP provides a model system
that allows us to accurately track the loss and the forma-
tion of oligomers and monomers that can then be ap-
plied to understand hemicellulose hydrolysis pathways.
One such study focused on determining how the reactiv-
ity of xylooligomer bonds changed with the degree of
polymerization (DP) for the hydrolysis of low DP xyloo-
ligomers from xylobiose to xylopentose as model com-
pounds with dilute sulfuric acid (0.025, 0.05, and
Xylobiose (lot 10702), xylotriose (lot 51003), xylotetrose
(lot 10804), and xylopentose (lot 10205) were obtained
from Megazyme International Ireland, Limited (Bray
Business Park, Bray, Co. Wicklow, Republic of Ireland).
Monomeric xylose was obtained from Fisher Biotech
(lot 006640, Fair Lawn, NJ, 07410).
2.2. Analytical procedures
1
5
0
.25 M) at low temperatures (80 and 100 ꢀC). Under
All chemical analyses of xylooligomers and xylose were
performed using a Waters Alliance HPLC system (Mod-
el 2695, Waters Corporation, Milford, MA) employing
Aminex-HP 42 A and HPX-87P columns (Bio-Rad,
Hercules, CA) for oligomer and xylose quantification,
respectively, equipped with appropriate accessories.
Samples were processed at an eluent flow rate of
0.60 mL/min using a refractive index (RI) detector
(model 2414, Waters Corporation, Milford, MA).
these conditions, xylose degradation was negligible,
and it was determined that the xylosidic linkage adjacent
to the non-reducing terminal group reacted almost 1.8
1
5
times faster than interior links. Research by our group
with low DP soluble oligomers at a higher temperature
of 200 ꢀC where xylose degradation is important showed
that the selectivity of xylose monomer formation from
oligomers was low in the absence of added sulfuric acid
and that the low yields of xylose monomer could not be
adequately explained without the incorporation of reac-
tions for direct degradation of oligomers prior to mono-
2.3. Parameter estimation
1
6
mer formation. More recent research on xylobiose
degradation revealed that sulfuric acid dramatically
increased the ratio of the depolymerization rate constant
to the overall disappearance rate constant to close to
one, significantly enhancing the selectivity of xylose
formation. In addition, the optimum acid concentration
to maximize xylose monomer recovery during acid
hydrolysis of xylan rich hemicellulose was determined
Parameter estimation was performed by fitting the data
by least-square method using the Excel Solver Routine
and Berkeley Madonna software from the University
of California, Berkeley.
2.4. Experimental system
Xylooligomer degradation experiments were performed
using clear glass crimp vials with flat bottoms and
aluminum crimp tops with TFE/silicone liners
(7 mm · 40 mm, 800 lL, Alltech Associates Inc; Deer-
field, IL) into which 500–600 lL of xylooligomer solu-
tion (0.25–1 mg/mL) was pipetted and sealed using a
quarter-inch crimping tool. To prevent rupturing, the
TFE/silicone septa glass reactor caps at high reaction
pressures, a disc punched from a 25 lm thick brass sheet
1
7
to be between pH 2 and 3.
The results for xylooligomer reactions reported to
date suggest that acid enhances the selectivity of xyloo-
ligomer breakdown to monomers through either dra-
matically reducing the role of side reactions that
directly degrade oligomers or by significantly speeding
the reactions to monomers relative to these direct degra-
dation reactions. However, because published results are
limited to a few temperatures, acid concentrations, and
oligomer chain lengths, it is difficult to draw broad con-
clusions. Thus, in this study, xylose, xylobiose, xylotri-
ose, xylotetrose, and xylopentose decomposition were
followed at 160 ꢀC with sulfuric acid added to cover a
pH range from near neutral to 1.45, and the selectivity
and the yields of xylose formation from xylooligomers
1
7
was inserted between the cap and the septa. In addi-
tion, 316 stainless steel tubes of 0.5 in. OD, 5 in. length,
and 0.035 in. wall thickness from Swagelok (Bangor,
ME) were employed as reactors to study the effect of
materials of construction and xylose concentration on
xylose degradation. In this case, 7 mL of xylose solution
(1–10 mg/mL) was pipetted into these tubes which were

292
R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
made leak proof by wrapping their threads with Teflon
tape and applying vacuum grease before sealing with a
bolt fitted with a Teflon stopper. Both the glass vials
and stainless steel reactor tubes were heated to reaction
temperature by immersion in a 22.8 cm ID · 35 cm deep
1
0
.95
5
mg/ml
0
.9
4
kW model SBL-2D fluidized sand bath (Techne Cor-
poration, Princeton, NJ). To hold the reactors in place,
promote uniform heating, and simplify removal, a
screen tied to a 3.5 cm diameter iron rod was immersed
about half way into the depth of the sand bath. Heating
times of 30 and 60 s for the glass and steel reactors,
respectively, were adequate to reach reaction tempera-
tures as determined by a thermocouple. The reaction
vessels were removed from the sand bath at different
reaction times with tongs and quenched in a cold water
tub held at room temperature. To reduce breakage and
loss of contents, the glass vials were air cooled for
2
mg/ml
0.85
1
mg/ml
10
0
.8
0
.75
0
20
30
40
50
Time, minute
Figure 1a. Fraction of initial xylose left in solution over reaction time
at 160 ꢀC for reactors made from glass vials.
1
0–20 s before immersing them in the tub. The contents
of the reactors were then removed with a 1 mL syringe
Dickinson and Company, Franklin Lakes, NJ) and
(
filtered through 0.2 lm nylon filter vials (Alltech Associ-
ates Inc., Deerfield, IL). About 100–150 mg of calcium
carbonate was added to the filter vials to neutralize
acidic samples before they were centrifuged using
Eppendorf Centrifuge MiniSpin at 13,400 rpm for
1
0
0
.95
5
mg/ml
mg/ml
2
min. Centrifugates were then pipetted into 500 lL
0
.9
2
polyethylene HPLC vials.
.85
1
mg/ml
0
.8
3. Results
3
.1. Xylose degradation
0.75
0
10
20
30
40
50
60
70
All experiments were run at 160 ꢀC, the temperature that
Time, minute
is found to often give the highest xylose yields for dilute
Figure 1b. Fraction of initial xylose left in solution over reaction time
at 160 ꢀC for reactors made from stainless steel.
4
,19,20
acid hydrolysis of hemicellulose.
First, we evalu-
ated how the choice of reactor material (stainless steel
or glass) influenced degradation kinetics. As shown in
Figure 1a, the rate of xylose degradation in glass vials
appeared to decrease somewhat with increasing xylose
concentration. On the other hand, as shown in Figure
Table 1. Xylose degradation rate constants (k
concentrations and pH
d
) at different xylose
Xylose concn
mg/ml)
pH neutral
pH 2.75
1
b, little effect of xylose concentration was evident over
(
Steel tube Glass vial Steel tube Glass vial
the same concentration range in stainless steel reactors.
Table 1 shows the xylose degradation constants esti-
mated by fitting the experimental data using Excel
Solver assuming first-order reaction kinetics, shown
below, for three different sugar concentrations and two
different pH values (neutral and 2.75).
1
2
5
0.0044
0.0050
0.0051
0.0048
0.013
0.0085
0.0058
0.0042
0.0062
0.0057
0.0052
0.0036
0.0048
0.0059
0.0042
0.0077
Average
kd
At neutral pH, the degradation rate constants for xy-
lose, Table 1, were higher for glass vials than steel reac-
tors, and the variation in constants with xylose
concentration was more apparent as well. However, at
lower pH this variation in glass vials was less prominent,
and the rate constants for xylose degradation were
Xylose ƒ! Degradation products
dX
dt
1
¼
ꢀk
d
ꢁ X
1
X
1
¼
1
expðꢀk ꢁ tÞ
d
X 0

R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
293
higher in steel reactors than glass vials. Based on these
results, we decided to use glass vials for all subsequent
experiments because we thought that steel might cata-
lyze degradation. Furthermore, to avoid the impact of
water quality on the results, deionized water and water
adjusted to different pH’s with sulfuric acid was kept
refrigerated for all subsequent experiments.
xylose concentration and depends only on acid concen-
1
3,21,22
tration.
Yet, appreciable differences have been
observed for degradation studies between batch and
1
2–14
flow systems.
In the current study, the degradation
rate constants did not change appreciably at xylose con-
centrations of 2 mg/mL (13.3 mM ) until a pH of 2.75
was reached and then became higher, consistent with
trends observed in a previous study at our laboratory
1
7
as shown in Table 2. However, compared to the stud-
ies performed at our laboratory, the ratio of degradation
constants, calculated using models reported in the liter-
ature, was quite high over the pH range, as Table 2
shows. In addition, the ratio of degradation constants
at pH 1.45 to those at neutral condition was very high,
as predicted by models in the literature. Figure 4 shows
the variation of xylose degradation constants with pH
and comparisons with the literature values.
3
.2. Xylooligomers disappearance
To see the impact of dilute sulfuric acid on xylooligomer
degradation, experiments were performed in glass vials
over a pH range from neutral down to 1.45 at 160 ꢀC,
the pH was adjusted with sulfuric acid. At higher pH
and for a given time, oligomers with DP of 4 and 5 gave
lower xylose yields than xylobiose or xylotriose as illus-
trated in Figure 2 for a pH of 4.75. However, the yield
and the selectivity of xylose formation increased appre-
ciably as acid was added for all the oligomers studied as
shown in Figure 3. However, the degradation of xylose
becomes appreciable if too much acid was added to
3.3.2. Xylooligomer disappearance. Overall, disappear-
ance rate constants for xylooligomers were determined
using the Solver Routine in Excel to minimize the differ-
ences between experimental and predicted residual con-
centrations of xylose and xylooligomers for the
following kinetic expressions for all DP oligomers:
1
7
the reaction media. Furthermore, yields seemed to
drop with increasing DP at any acid concentration for
these five oligomers, as also shown in Figure 3.
dX
dt
n
¼
ꢀk ꢁ X
n
n
3
3
.3. Kinetic study
X
n
¼
expðꢀk ꢁ tÞ
n
Xn0
.3.1. Xylose monomer degradation. As discussed in
Section 3.2, the rate of xylose degradation seemed to
depend inversely on its concentration at higher pH val-
ues, especially in glass vials. On the other hand, most
previous kinetic studies were performed at highly acidic
conditions where these effects were not seen, leading to
the conclusion that xylose degradation is first order in
in which n = 1–5; and k = rate constants to describe the
n
overall disappearance of each xylooligomer. The rate
constants for overall disappearance of different xylooli-
gomers increased with DP as shown in Figure 5. The
ratios of xylooligomer disappearance rate constants to
xylose degradation rate constant were greater than 2
for all the oligomers, and this ratio increased with
decreasing pH, as shown in Table 3, contrary to the pre-
1
5
diction by the Kamiyama and Sakai model (developed
for lower temperatures and higher acid molar concentra-
tion) shown in Table 3 in parenthesis.
0
.3
0
0
0
.25
3
.3.3. Xylooligomer hydrolysis versus degradation. The
disappearance rate constants account for the loss of
the oligomers but not the selectivity of xylose formation
versus loss to degradation products. As discussed
earlier, previous research by our group showed that
the selectivity of xylose formation from oligomers was
low near neutral pH values. In addition, calculated rate
constants based on following just xylooligomer degrada-
tion could not adequately describe the formation of
xylose monomer or degradation products from soluble
xylooligomers based on first-order degradation of
0
.2
Xylobiose
Xylotriose
.15
0
.1
Xylotetrose
.05
0
Xylopentose
1
6
0
20
40
60
80
100
xylooligomers in series. From this observation, it was
hypothesized that direct degradation of oligomers also
occurs in parallel with the formation of xylose and lower
DP oligomers as follows:
Time, minutes
Figure 2. Fractional xylose yields from xylobiose up to xylopentose for
reaction at 160 ꢀC and a pH of 4.75.

294
R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
a
1.25
b
1
1
0
.25
.00
.75
pH- Neutral
pH- 4.75
pH-2.75
pH- Neutral
pH- 4.75
pH- 3.75
pH-2.75
1
0
0
0
0
.00
.75
.50
.25
.00
pH-1.45
pH-1.45
0.50
0.25
0.00
0
10
20
30
40
0
10
20
30
40
50
60
70
80
90
Time, minutes
Time, minutes
1
c
1.25
d
pH- Neutral
pH- Neutral
pH- 4.75
pH- 3.75
pH-2.75
pH-1.45
pH- 4.75
pH- 3.75
pH-2.75
pH-1.45
1
0
0
0
0
.00
.75
.50
.25
.00
0
0
.75
0
.5
.25
0
0
10
20
30
40
50
60
70
80
90
0
10
20
30
40
50
60
Time, minutes
Time, minutes
Figure 3. Fractional xylose yields at 160 ꢀC and various pH values for (a) xylopentose, (b) xylotetrose, (c) xylotriose, and (d) xylobiose.
Without acid:
Depolymerization
Oligomers
Lower oligomers
Direct degradation
Degradation products
Xylose monomer
Degradation products
However, it was assumed that direct degradation of
oligomers is reduced or eliminated by adding acid either
by accelerating hydrolysis reactions or reducing degra-
dation as follows:
ically modeled xylooligomer disappearance based on
this mechanism assuming the degradation of oligomers
to shorter chain oligomers and degradation products
followed first-order kinetics.
With acid:
Depolymerization
Lower oligomers Xylose monomer
Direct degradation
Degradation products
Oligomers
Degradation products
Consistent with this hypothesis, another study by our
group showed that sulfuric acid dramatically increased
the xylobiose depolymerization rate constant compared
to the overall disappearance rate constant, shifting the
3.3.4. Xylobiose hydrolysis. As shown in Figure 6,
experimental and modeled xylose concentrations, using
the k determined for xylobiose, were not the same if
2
only the reaction of xylobiose to xylose was considered,
while incorporation of direct degradation of xylobiose
1
7
ratio closer to one as the pH dropped. Thus, we kinet-

R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
295
a
Table 2. Xylose degradation rate constants at different pH and their comparison to the literature values
n
pH
k
d
Ref. 21
k
d
Ref. 12
k
d
Ref. 16
d
k Current study
(
ratio kdn/kdnꢀ1
)
(ratio kdn/kdnꢀ1
)
(ratio kdn/kdnꢀ1
)
(ratio kdn/kdnꢀ1)
1
2
Neutral
4.75
8.47Eꢀ06
1.26Eꢀ04
3.90Eꢀ08
1.91Eꢀ06
(50)
1.88Eꢀ03
1.88Eꢀ03
(1)
5.9Eꢀ03
6.6Eꢀ03
(1.1)
(
14.9)
3
4
5
k
3.75
2.75
1.45
6.18Eꢀ04
4.9)
3.0Eꢀ03
4.9)
2.39Eꢀ02
7.9)
2.8E2
1.91Eꢀ05
1.89Eꢀ03
5.2Eꢀ03
(
(10)
(1)
(0.8)
1.91Eꢀ04
(10)
2.0Eꢀ03
(1.06)
5.5Eꢀ03
(1.06)
(
3.80Eꢀ03
4.34Eꢀ03
9.8Eꢀ03
(
(20)
(2.16)
(1.80)
d 1.45/kd neutral
10 E5
2.30
1.60
a
d
Values are degradation rate constants reported by references noted, while there in () are ratio of the value of k shown to that for previous at higher
pH value.
in which k l is the xylobiose loss constant; k d, the
Current study
Ref. 17
2
2
0
.1
xylobiose depolymerization constant; k = (k l + k d),
2
2
2
Ref. 22
the xylobiose overall disappearance constant; and
k1, the xylose degradation constant of 13.3 mM
xylose.
Ref. 13
0
.01
Parameters were obtained by minimizing the differ-
ence between experimentally measured and predicted
concentrations of xylose monomer and oligomers using
Excel Solver. As Table 4 shows, the rates of both
xylobiose depolymerization to xylose monomer and
degradation to unknown products increased with acid
concentration. However, the ratio of depolymerization
rate constants to degradation rate constants
increased and the ratio of degradation constants to
overall disappearance constant decreased with acid
concentration.
0
.001
0
.0001
0
.00001
0
1
2
3
4
5
6
7
pH
Figure 4. Xylose degradation rate constants and comparison to the
literature values at different pH levels and 160 ꢀC.
3
.3.5. Xylotriose hydrolysis. The experimental and
into the model resulted in predicted concentrations
based on disappearance rate constants that closely
agreed with observed xylose concentrations. As shown
in Table 4 for xylobiose, the ratio of depolymerization
constants to the overall disappearance constant was
found to increase with acid concentration and approach
one at lower pH, for the following kinetic expressions:
modeled xylose and xylobiose concentrations were not
the same if only the reaction of xylotriose to xylobiose
and the reaction of xylobiose to xylose were considered,
as shown in Figure 7a, while incorporation of direct deg-
radation of xylotriose and xylobiose into the model
resulted in predicted concentrations based on the
disappearance constants that closely agreed with those
for the observed xylobiose and xylose concentrations,
as Figure 7b shows. For parameter estimation, the
least-square method of the Berkeley Madonna software
was used with the reaction modeled as xylotriose break-
ing down to xylobiose and xylose and degradation prod-
ucts in parallel reactions. In addition, xylobiose
decomposed to xylose and degradation products, and
xylose further decomposed to degradation products.
The kinetic parameters determined previously were used
k2d
X2
2X1
k
2
l
Degradation
Degradation
k1
X1
to describe xylotriose hydrolysis with k = 0.0059,
1
ꢀ1
dX
dt
2
k d = 0.0059, k l = 0.0062, and k = 0.024 min
at
2
2
3
¼
¼
ꢀðk d þ k lÞ ꢁ X
2
2
2
neutral pH. The stoichiometric equations and corre-
sponding rate laws for xylotriose hydrolysis are as
follows:
dX1
k
2
d ꢁ X
2
ꢀ k
1
ꢁ X
1
dt

296
R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
a
b
0
0
0
0
0
0
0
.07
.06
.05
.04
.03
.02
.01
0
0.04
0
0
0
.03
.02
.01
0
c
d
0
.2
7
6
5
4
3
2
1
0
0
0
.15
0
.1
.05
0
Figure 5. Overall disappearance rate constants for xylose and its oligomers up to xylopentose at 160 ꢀC and pH (a) neutral, (b) 4.75, (c) 2.75, and (d)
1.45.
n
Table 3. Ratio of overall degradation rate constant (k ) of oligomer to
k
3
d
k
l
2
d
a
X3
X2 + X1
X2
2X1
1
xylose degradation rate constant (k ) at various pH
pH
X
2
X
3
X
4
X
5
k
3
l
k
2
a
Neutral
2.1 (2.3)
2.4 (2.2)
4.1 (3.4)
2.9 (3.5)
26.9 (3.6)
3.1 (4.7)
4.8 (4.7)
26.2 (4.8)
10.7 (6.0)
5.3 (5.9)
35.3 (5.9)
391.8 (5.9)
Degradation
Degradation
4
2
1
.75
.75
.45
23.6 (2.3)
165.3 (2.2)
k
1
138.8 (3.5)
625.5 (4.7)
X1
Degradation
a
Values in parenthesis are ratios for degradation rate constants pre-
dicted by the Kamiyama and Sakai model.
1
5
dX
3
2
¼
¼
¼
ꢀðk
3
d þ k
3
lÞ ꢁ X
3
dt
dX
2=3 ꢁ k d ꢁ X ꢀ ðk d þ k lÞ ꢁ X
k and k d to k l were found to increase with increasing
3 3 3
acid concentration. In addition, losses were observed to
decline with acid concentration.
3
3
2
2
2
dt
dX1
1=3 ꢁ k
3
d ꢁ X
3
þ k
2
d ꢁ X
2
ꢀ k
1
ꢁ X
1
dt
in which k l is the xylotriose loss constant; k d, the xylo-
3.3.6. Xylotetrose hydrolysis. Assuming first-order
reactions, xylotetrose could depolymerize to shorter
oligomers by two possible mechanisms and also decom-
pose directly to degradation products. Xylotetrose dis-
appearance could lead to the formation of xylotriose
and xylose and two xylobiose molecules simultaneously.
Further, xylotriose, xylobiose, and xylose could all
3
3
triose depolymerization constant; and k = (k d + k l),
3
3
3
the xylotriose disappearance rate constant.
Modeling of xylotriose hydrolysis and parameter esti-
mation were performed for experimental data at other
pH values as well, with the model parameters and statis-
2
tical coefficients R shown in Table 5. The ratio of k d to
3

R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
297
degrade as per the mechanisms described earlier. The
1
.8
.6
parameters determined at different pH values for xylo-
tetrose, xylotriose, and xylobiose disappearance and
xylose degradation were used in the xylotetrose degrada-
tion model, and the new parameters estimated by the
least-square method. Thus k = 0.1438, k = 0.148,
Observed X2/ X20
Observed X1 yield
Modeled X2/ X20
X1 yield- with parallel deg.
X1yield- without parallel deg.
0
0
4
3
k d = 0.1394, k l = 9.0Eꢀ3, k = 0.128, k d = 0.116,
3
3
2
2
ꢀ
1
k l = 0.012, and k = 0.0066 min for a pH value of
2
0.4
1
2
.75. The kinetic expressions and corresponding rate
laws are shown below
0
.2
0
41
k d
3
k d
X
4
X
3
+ X1
+ X2
X3
X2 + X1
0
20
40
60
80
100
42
k d
3
k l
Time, minutes
4
k l
X
2
Degradation
Figure 6. Xylobiose disappearance and xylose formation at 160 ꢀC and
neutral pH.
Degradation
X2
2
k d
k1
1.0
2X1
X1
Degradation
Exp.X3 total yield
Predicted X3 total yield
Exp.X3 yield
PredictedX3 yield
Exp.X2 yield
Predicted X2 yield
Exp.X1 yield
Predicted X1 yield
k
2
l
0.8
Degradation
0
0
0
.6
.4
.2
dX4
dt
¼
¼
¼
¼
ꢀðk41d þ k42d þ k
3=4k41d ꢁ X
4
lÞ ꢁ X
4
dX
dt
dX
3
2
4
ꢀ ðk
3
d þ k
3
lÞ ꢁ X
3
k d ꢁ X þ 2=3 ꢁ k d ꢁ X ꢀ ðk d þ k lÞ ꢁ X
42
4
3
3
2
2
2
dt
dX1
dt
0.0
1=4k41d ꢁ X
þ k d ꢁ X
4
þ 1=3 ꢁ k
3
d ꢁ X
3
0
20
40
60
80
Time, minute
2
2
ꢀ k
1
ꢁ X
1
Figure 7a. Xylotriose degradation at 160 ꢀC and neutral pH without
incorporating direct losses of xylobiose and xylotriose.
in which k d and k d are the xylotetrose depolymeriz-
4
1
42
ation constants; k l is the xylotetrose loss const.; and
4
similar, except at pH 4.75, with and without considering
direct loss of xylotetrose. Thus, direct loss of xylotetrose
at higher pH was unimportant.
k = (k d + k d + k l) is the xylotetrose overall disap-
4
41
42
4
pearance constant.
The experimental and predicted concentrations are
shown in Figure 8 at 3.75 pH. The values of the param-
eters and the residual sum of square (RSS) determined
at different pH values are shown in Table 6. The sum
of the parameters k d, k d, and k l was constrained
3.3.7. Xylopentose hydrolysis. Xylopentose depolymer-
ization could take place via three routes assuming all the
reactions are of first order: xylopentose may break down
to xylotetrose oligomer and xylose monomer; xylotriose
and xylobiose oligomers; and directly to degradation
products, as shown below. The parameters estimated
individually for lower oligomers and xylose monomer
4
1
42
4
to be equal to k , and the difference between oligomers
4
and xylose concentration was minimized using the
least-square method. The experimental and predicted
concentration of oligomers and xylose monomer were
Table 4. Xylobiose overall disappearance (k
2
), depolymerization (k
2
d) rate, and loss (k
2
l) rate constants and their ratios at different pH values
a
pH
k
2
k
2
d
k
2
l
k
2
d/k d/k l/k
2
l
k
2
2
k
2
2
Neutral
0.0121
0.0156
0.0092
0.128
1.62
0.0059
0.0068
0.0056
0.116
1.5
0.0062
0.0088
0.0036
0.012
0.12
0.95
0.77
1.56
9.67
0.49
0.44
0.61
0.91
0.93
0.51
0.56
0.39
0.094
0.074
4
3
2
1
.75
.75
.75
.45
12.5
a
k
2
l = k
2
2
ꢀ k d.

298
R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
Exp. X3 total yield
Predicted X3 total yield
Exp. X3 yield
pearance rate constant (k5) and minimizing the
difference between experimental and modeled concentra-
tions of oligomers and xylose by using the least-square
method of Berkeley Madonna software for the data at
all pH values. The stoichiometric relationships and rate
laws are as follows:
Predicted X3 yield
Exp. X2 yield
1
.0
Predicted X2 yield
Exp. X1 yield
Predicted X1 yield
0
0
0
0
0
.8
.6
.4
.2
.0
k
51
d
41
k d
X
5
X
4
+ X1
+ X2
X
4
X
3
+ X1
+ X2
k
52
d
42
k d
k
5
l
k
4
l
X
3
X
2
0
20
40
60
80
Time, minute
Figure 7b. Xylotriose degradation at 160 ꢀC and neutral pH including
the direct losses of xylobiose and xylotriose.
Degradation
Degradation
3
k d
2
k d
X3
X2 + X1
X2
2X1
Table 5. Depolymerization (k
3
d), loss (k
3
l), and overall disappearance
3
k l
2
k l
(k
3
) rate constants and their ratios for xylotriose at different pH values
Degradation
Degradation
Parameter pH
3.75
k
1
Neutral
4.75
2.75
0.148
0.1394
0.009
0.941
0.059
1.45
X1
Degradation
k
3
3
3
3
3
3
0.024
0.018
0.005
0.77
0.22
3.46
0.020
0.017
0.003
0.87
0.13
6.84
0.029
0.020
0.009
0.68
0.32
2.13
1.362
1.362
0.00001
0.99
k
k
k
k
k
d
dX
dt
5
l
¼ ꢀðk d þ k d þ k lÞ ꢁ X
5
1
52
5
5
d/k
l/k
d/k
3
dX4
3
1Eꢀ5
¼
¼
¼
4=5k51d ꢁ X
5
ꢀ ðk41d þ k42d þ k
4
lÞ ꢁ X
4
3
l
15.9
0.162
dt
dX
dt
dX2
dt
RSS
0.0632
0.066
0.065
0.3244
3
3=5k d ꢁ X þ 3=4k d ꢁ X ꢀ ðk d þ k lÞ ꢁ X
5
2
5
41
4
3
3
3
were used in modeling xylopentose hydrolysis, for exam-
ple, at neutral pH k5 = 0.0629, k4 = 0.0180,
k d = 0.0148, k d = 0.0032, k = 0.024, k d = 0.0183,
2=5k52d ꢁ X
ðk d þ k
5
þ k42d ꢁ X
4
þ 2=3 ꢁ k
3
d ꢁ X
3
4
1
42
3
3
ꢀ
2
2
lÞ ꢁ X
2
ꢀ
1
k2 = 0.0121, k2d = 0.0059, and k1 = 0.0059 min
.
dX1
dt
Three kinetic parameters (k d, k d, and k l) known
¼ 1=5k d ꢁ X þ 1=4k d ꢁ X þ 1=3 ꢁ k d ꢁ X
5
1
52
5
51
5
41
4
3
3
to account for xylopentose degradation were estimated
by constraining their sums to equal the overall disap-
þ k
2
d ꢁ X
2
ꢀ k
1
ꢁ X
1
1
.1
.0
.9
.8
.7
.6
.5
.4
.3
.2
.1
.0
1
0
0
Exp. X4 total yield
Predicted X4 total yield
Exp. X4 yield
Predicted X4 yield
Exp. X3 yield
Predicted X3 yield
Exp. X2 yield
Predicted X2 yield
Exp. X1 yield
0
0
0
0
0
Predicted X1 yield
0
0
0
0
10
20
30
40
50
60
70
Time, minute
Figure 8. Xylotetrose disappearance at 160 ꢀC and pH 3.75.

R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
299
Table 6. Depolymerization (k41d and k42d), loss (k
4
l), and overall
Table 7. Depolymerization (k51d and k52d), loss (k
5
l), and overall
disappearance (k ) rate constants and their ratios for xylotetrose at
4
disappearance (k ) rate constants and their ratios for xylopentose at
5
different pH values
Parameter
different pH values
Parameter
pH
3.75
0.0323 0.022
0.0166 0.016562 0.080454 1.7800
0.0051 0.0054
pH
Neutral 4.75
2.75
1.45
6.13
Neutral
4.75
2.75
1.45
k
k
k
k
k
k
4
0.0184
0.0148
0.0032
0.00032 0.003
1.0
0.1438
k
k
k
k
k
5
0.0629
0.04186
0.0210
0.0000
1.0
0.0350
0.0280
0.0071
0.0000
1.0
0.19465
0.110791
0.0839
0.0000
1.0
3.8386
1.9140
1.9246
0.0000
1.0
41
42
d
d
51
52
d
d
0.0633
0.009
1.0
0.00
0.426
4.3510
0.00001
1.0
0.00
0.55
4
4
4
l
0.009
1.0
0.00
0.18
5
l
d/k
l/k
4
0.67
0.33
0.172
d/k
5 5
4
0.00
0.136
RSS
0.334
0.649
0.85
0.55
RSS
in which k d and k d are the xylopentose depolymer-
tion in this study is in agreement with Laboratory Ana-
lytical Procedure (LAP-02, paragraph-10.1.7, NREL,
USA) and research done at NREL (Dr. Bonnie Hames,
personal communication) that monomeric sugar degra-
dation depends on their concentration and may be af-
fected by aqueous media as well.
5
1
52
ization constants; k l is the xylopentose loss rate con-
5
stant; k = (k d + k d + k l) is the xylopentose
5
51
52
5
overall disappearance constant.
The modeled and experimental concentration values
were found to agree closely, as shown by Figure 9 at
pH 4.75. However, direct degradation of xylopentose
was not evident with the degradation rate constant
found to be being equal to zero at all pH values, as
shown in Table 7.
The observation that xylose degradation rate con-
stants at lower pH are higher in steel reactors compared
to glass vials could result from the combined catalytic
effects of metal and acid. We have also observed that xy-
lose degradation rate constants can vary significantly
and sometimes by several fold when repeated on differ-
ent days possibly due to changes in the amount of min-
4. Discussion
2
6
erals and salts in the water.
This study showed that xylose degradation depends on
xylose concentration, especially in glass vial reactors at
low concentrations of sulfuric acid. This finding will
be important for pretreatments such as steam explosion,
controlled pH, flowthrough, and dilute acid, where little
The selectivity of monomeric xylose formation from
oligomers increased with acid concentration. However,
this effect is observed until xylose concentrations build
to the point that its degradation becomes significant,
not determined in this study. In addition, the overall
selectivity of oligomer breakdown to monomeric xylose
1
7
2
3–25
if any acid is added.
However, based primarily on
studies conducted at very acidic conditions in stainless
decreased with increasing DP at
concentration.
a given acid
1
1,12,18
steel reactors,
xylose degradation was shown to
follow first-order kinetics, with the rate depending just
on the amount of acid and temperature. The observa-
The rate constants for xylooligomer disappearance
increased with DP over the range of oligomers studied
1
1
0
0
0
0
0
.2
.0
.8
.6
.4
.2
.0
Exp.X5 total yield
Predicted X5 total yield
Exp.X5 yield
Predicted X5 yield
Exp.X4 yield
Predicted X4 yield
Exp.X3 yield
Predicted X3 yield
Exp. X2 yield
Predicted X2 yield
Exp. X1 yield
Predicted X1 yield
-4
0
4
8
12 16 20 24 28 32 36 40
Time, minute
Figure 9. Xylopentose disappearance at 160 ꢀC and pH 4.75.

300
R. Kumar, C. E. Wyman / Carbohydrate Research 343 (2008) 290–300
from xylobiose to xylopentose. Xylooligomers degraded
faster than monomeric xylose at all the pH values and
the ratio of xylooligomer disappearance rate constants
to xylose degradation rate constant increased with
decreasing pH. On the other hand, in a study of xyloo-
ligomer degradation at 43 ꢀC using alkaline medium,
Sartori et al. showed the rate constants for xylooligo-
mers degradation increased compared to the monosac-
charide as the DP increased from 2 to 3 and then
2. Nguyen, Q. A.; Keller, F. A.; Tucker, M. P.; Lombard, C.
K.; Jenkins, B. M.; Yomogida, D. E.; Tiangco, V. M.
Appl. Biochem. Biotechnol. 1999, 77–79, 455–471.
3
. Nguyen, Q. A.; Tucker, M. P.; Keller, F. A.; Eddy, F. P.
Appl. Biochem. Biotechnol. 2000, 84–86, 561–576.
4. Saha, B. C.; Iten, L. B.; Cotta, M. A.; Wu, Y. V.
Biotechnol. Prog. 2005, 21, 816–822.
5
6
. Sun, Y.; Cheng, J. J. Bioresour. Technol. 2005, 96, 1599–
606.
1
. Wyman, C. E.; Bain, R. L.; Hinman, N. D.; Stevens, D. J.
Ethanol and Methanol from Cellulosic Biomass. In
Renewable Energy: Sources for Fuels and Electricity;
Johansson, H. K. T. B., Reddy, A. K. N., Williams, R.
H., Eds.; Island Press: Washington, DC, 1993.
2
1
decreased slightly for oligomers studied of DP 4.
16,17
The hypothesis
that oligomers directly degrade at
low acid concentration was also tested for a range of
oligomers from xylobiose to xylopentose. Observations
with xylobiose and xylotriose strongly suggest that
direct degradation of oligomers occurs along with depo-
lymerization to lower oligomers and monomer and is
minimized by the presence of acid. However, xylotetrose
and xylopentose modeling contradicts the hypothesis,
and direct degradation of these oligomers at low and
high acid concentrations was found to be negligible.
Thus, it may be interpreted that the direct degradation
may only occur with lower oligomers such as xylobiose
and xylotriose. Consistent with the finding in this
research, Garrote et al. assumed for kinetic modeling
of corn cob autohydrolysis that xylan breaks down to
higher DP oligomers and that it then forms lower DP
oligomers, which can degrade directly to furfural and
form xylose. However, this was an empirical fit in that
the length of xyloligomers that directly degrade was not
defined. Therefore, it can be assumed that oligomers of
DP 2–3 are decomposed directly to degradation prod-
ucts at high pH while DP 4 and 5 do not. However,
further validation is needed with higher oligomers to
determine if this trend is followed by higher DP
oligomers.
7
. Lloyd, T. A.; Wyman, C. E. Bioresour. Technol. 2005, 96,
1
967–1977.
8
. Lamptey, J.; Robinson, C. W.; Moo-Young, M. Biotech-
nol. Lett. 1985, 7, 531–536.
9. Rubio, M.; Tortosa, J. F.; Quesada, J.; G o´ mez, D.
Biomass Bioenergy 1998, 15, 483–491.
0. Kobayashi, T.; Sakai, Y. Bull. Agric. Chem. Soc. Jpn.
1
1
1
956, 20, 1–7.
1. Garrote, G.; Dom ´ı nguez, H.; Paraj o´ , J. C. Holz Roh
Werkst. 2001, 59, 53–59.
12. Saeman, J. F. Ind. Eng. Chem. 1945, 37, 43–52.
3. Root, D. F. Ph.D. Thesis, University of Wisconsin, WI,
956.
4. Kwarteng, K. I. Ph.D. Thesis, 1983, Dartmouth College,
Hanover, NH, 1983.
1
1
1
15. Kamiyama, Y.; Sakai, Y. Carbohydr. Res. 1979, 73, 151–
158.
1
1
6. Li, X. M.S. Thesis, Dartmouth College, Hanover, NH,
2
7
2002.
7. Lloyd, T. A. Ph.D. Thesis, Dartmouth College, Hanover,
NH, 2005.
18. Qian, X.; Nimlos, M. R.; Davis, M.; Johnson, D. K.;
Himmel, M. E. Carbohydr. Res. 2005, 340, 2319–
2
327.
1
2
2
2
9. Knappert, D.; Grethlein, H.; Converse, A. Biotechnol.
Bioeng. 1980, 22, 1449–1463.
0. Maloney, M. T.; Chapman, T. W.; Baker, A. J. Biotechnol.
Bioeng. 1985, 27, 355–361.
1. Sartori, J.; Potthast, A.; Ecker, A.; Sixta, H.; Rosenau, T.;
Kosma, P. Carbohydr. Res. 2003, 338, 1209–1216.
2. Converse, A. O.; Kwarteng, I. K.; Grethlein, H. E.;
Ooshima, H. Appl. Biochem. Biotechnol. 1989, 20/21, 63–
Acknowledgments
7
8.
The authors would like to thank the National Institute of
Standards and Technology for supporting this research
through award 60NANB1D0064 and the Thayer
School of Engineering at Dartmouth College for provid-
ing the equipment and facilities vital to this research.
2
3. Mosier, N.; Wywan, C. E.; Dale, B.; Elander, R.; Lee, Y.
Y.; Holtzapple, M.; Ladisch, M. Bioresour. Technol. 2005,
9
4. Liu, C.; Wyman, C. E. Bioresour. Technol. 2005, 96, 1967–
1977.
5. Liu, C.; Wyman, C. E. Ind. Eng. Chem. 2004, 43, 2781–
6, 673–686.
2
2
2
2
788.
6. Liu, C.; Wyman, C. E. Carbohydr. Res. 2006, 341, 2550–
556.
References
2
1
. Wooley, R.; Ruth, M.; Sheehan, J.; Ibsen, K.; Majdeski,
H.; Galvez, A. NREL Report No. TP-580-26157, July
27. Garrote, G.; Dom ´ı nguez, H.; Paraj o´ , J. C. Process
Biochem. 2001, 36, 571–578.
1999; p 130.