Solid acid-catalyzed cellulose hydrolysis monitored by in situ Atr-Ir spectroscopy

Article abstract of DOI:10.1002/cssc.201100631

The solid acid-catalyzed hydrolysis of cellulose was studied under elevated temperatures and autogenous pressures using in situ ATR-IR spectroscopy. Standards of cellulose and pure reaction products, which include glucose, fructose, hydroxymethylfurfural (HMF), levulinic acid (LA), formic acid, and other compounds, were measured in water under ambient and elevated temperatures. A combination of spectroscopic and HPLC analysis revealed that the cellulose hydrolysis proceeds first through the disruption of the glycosidic linkages of cellulose to form smaller cellulose molecules, which are readily observed by their distinctive C-O vibrational stretches. The continued disruption of the linkages in these oligomers eventually results in the formation and accumulation of monomeric glucose. The solid-acid catalyst accelerated the isomerization of glucose to fructose, which then rapidly reacted under hydrothermal conditions to form degradation products, which included HMF, LA, formic acid, and acetic acid. The formation of these species could be suppressed by decreasing the residence time of glucose in the reactor, reaction temperature, and contact with the metal reactor. The hydrolysis of regenerated cellulose proceeded faster and under milder conditions than microcrystalline cellulose, which resulted in increased glucose yield and selectivity. Vibrating acids: The solid acid-catalyzed hydrolysis of cellulose has been studied by using in situ ATR-IR spectroscopy. It is possible to monitor the formation and consumption of important products and intermediates by their distinctive vibrational characteristics under the reaction conditions. Copyright

Full text of DOI:10.1002/cssc.201100631

DOI: 10.1002/cssc.201100631
Solid Acid-Catalyzed Cellulose Hydrolysis Monitored by
In Situ ATR-IR Spectroscopy
Joseph Zakzeski,[a] Ruud J. H. Grisel,*[b] Arjan T. Smit,[b] and Bert M. Weckhuysen*[a]
The solid acid-catalyzed hydrolysis of cellulose was studied
under elevated temperatures and autogenous pressures using
in situ ATR-IR spectroscopy. Standards of cellulose and pure re-
action products, which include glucose, fructose, hydroxyme-
thylfurfural (HMF), levulinic acid (LA), formic acid, and other
compounds, were measured in water under ambient and ele-
vated temperatures. A combination of spectroscopic and HPLC
analysis revealed that the cellulose hydrolysis proceeds first
through the disruption of the glycosidic linkages of cellulose
to form smaller cellulose molecules, which are readily observed
by their distinctive CÀO vibrational stretches. The continued
disruption of the linkages in these oligomers eventually results
in the formation and accumulation of monomeric glucose. The
solid-acid catalyst accelerated the isomerization of glucose to
fructose, which then rapidly reacted under hydrothermal con-
ditions to form degradation products, which included HMF, LA,
formic acid, and acetic acid. The formation of these species
could be suppressed by decreasing the residence time of glu-
cose in the reactor, reaction temperature, and contact with the
metal reactor. The hydrolysis of regenerated cellulose proceed-
ed faster and under milder conditions than microcrystalline cel-
lulose, which resulted in increased glucose yield and selectivity.
Introduction
Biomass is an important feedstock for the production of fuels,
chemicals, and energy. In the United States and the European
Union, the highest proportion of renewable energy is obtained
from biomass, which surpasses hydropower, geothermal, wind,
and solar power. Moreover, biomass is currently the only re-
newable source of liquid transportation fuels.[1,2] Based on the
continued depletion of fossil fuel reserves and the continued
increase in chemical and energy demand, the use of biomass
can only be expected to increase in the foreseeable future.
One particular opportunity to satisfy the energy and raw mate-
rial requirements of society is through the valorization of ligno-
cellulosic biomass, which is inexpensive, abundant, and renew-
able. Comprising the inedible component of most plant
matter, lignocellulosic biomass has the added advantage that
it does not directly compete with agricultural food production.
The annual production of lignocellulosic biomass is enormous
and, with only modest changes in land use, capable of satisfy-
ing virtually all of the raw material requirements of the chemi-
cal industry provided that suitable processes for its selective
conversion can be developed.[3,4]
Disrupting the innate crystalline structure in a controlled
manner is, therefore, critical for the valorization of cellulose.
One method to disrupt the crystalline structure in cellulose
is through the use of ionic liquids. Pioneering work by Rogers
et al. revealed that certain ionic liquids, such as 1-butyl-3-meth-
ylimidazolium chloride, were capable of dissolving cellulose.[6]
Chloride anions penetrated the cellulose structure and disrupt-
ed the network of hydrogen bonds, and several catalytic pro-
cesses have since been developed for cellulose valorization
using this ionic liquid as the solvent.[7–10] Dissolution of cellu-
lose in an ionic liquid and subsequent precipitation using an
antisolvent, such as water or methanol, results in significant al-
terations to the cellulose microenvironment. The material ob-
tained as a result of this pretreatment process, termed regen-
erated cellulose, is significantly more susceptible to conversion
by enzymatic or catalytic processes, such as hydrolysis, relative
to untreated cellulose.[11,12]
Of the catalytic processes, hydrolysis is particularly attractive
as a method to produce glucose, which is a useful platform
chemical from which many other valuable chemicals can be
This impetus has prompted immense interest in the devel-
opment of processes that readily convert the components of
lignocellulosic biomass, namely, cellulose, hemicellulose, and
lignin, into useful fuels and chemicals. In particular cellulose,
the most abundant component,[5] has attracted considerable
attention. Cellulose is a recalcitrant material and has poor solu-
bility in most common solvents, which can be largely ex-
plained by its crystalline structure that is energetically stabi-
lized by a vast network of inter- and intramolecular hydrogen
bonding. This extensive bonding also protects cellulose from
chemical transformation by hindering the access of catalysts,
enzymes, and other reagents to key structural components.
[a] Dr. J. Zakzeski, Prof. Dr. B. M. Weckhuysen
Inorganic Chemistry and Catalysis Group, Utrecht University
Universiteitsweg 99, 3584 CG Utrecht, (The Netherlands)
Fax: (+31)302-534328
E-mail: b.m.weckhuysen@uu.nl
[b] Dr. R. J. H. Grisel, A. T. Smit
Biomass, Coal & Environmental Research
Energy Research Centre of the Netherlands (ECN)
Westerduinweg 3, PO Box 1, 1755 ZG Petten, (The Netherlands)
Fax: (+31)224-568487
E-mail: grisel@ecn.nl
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201100631.
430
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2012, 5, 430 – 437

Solid Acid-Catalyzed Cellulose Hydrolysis
made. The hydrolysis of cellulose to glucose and subsequent
conversion to products such as hydroxymethylfurfural (HMF)
and levulinic acid (LA) have been reported using surface-sup-
ported Lewis-acid, Brønsted-acid,[6] and solid-acid catalysts[13–20]
in water. Kinetic and mechanistic studies on the conversion of
cellulose derivates and sugars to HMF and LA have provided
considerable insight into the types of transformations that
occur in these systems.[21,22] Often, the performance of these
processes depends strongly on reaction conditions, and strate-
gic selection of the appropriate conditions vastly influences
product yield and selectivity. The ability to monitor the
changes that occur during a process under the operating con-
ditions is especially advantageous to maximize product conver-
sion and selectivity.
Results and Discussion
A possible route through which glucose and the various other
chemicals are formed by the acid-catalyzed hydrolysis of cellu-
lose is depicted in Scheme 1. The steps in this route are sup-
ported by previous mechanistic studies on acid-catalyzed cellu-
lose hydrolysis[21,22] and the spectroscopic results presented
below. The vibrations and the assignments associated with
these species are presented in Table 1.[29] The process begins
with the cellulose starting material, which is hydrolyzed in the
presence of the solid-acid catalyst and water to form cellulose
with a reduced chain length. The increased solubility of this
material relative to the starting cellulose allows the observa-
tion of several vibrations, which include the most prominent
CÀO stretch and C1ÀH bend at n˜ =1105 and 1026 cmÀ1, re-
spectively. Continued hydrolysis results in the formation and
accumulation of glucose, which is observed by vibrations at
n˜ =1074 and 1038 cmÀ1. Isomerization of glucose to fructose
and subsequent dehydration results in the formation of HMF,
The products that are formed during the conversion of cellu-
lose to glucose and sequential degradation, such as HMF or
LA, contain a wide range of acid, aldehyde, and alcohol func-
tional groups, and the characteristic vibrations associated with
these functional groups are detectable by using IR spectrosco-
py. The principle challenges in the online monitoring of these
processes by using IR spectroscopy arise not only from the ele-
vated temperatures and pressures under which these process-
es occur, but also because of the strong absorptions of the
solvent, particularly bulk water, which often interfere with de-
tection of the relevant products and reactants.
which is best observed by the C=C stretch at n˜ =1586 cmÀ1
.
The HMF aldehyde vibration is not as pronounced under the
reaction conditions likely because of the reaction of water with
the aldehyde to form a geminal diol, the equilibrium of which
is thermodynamically favored at elevated temperatures. The
appearance of a vibration at around n˜ =1710 cmÀ1 indicates
the formation of LA and formic acid (or other acid species),
which form as a result of HMF hydrolysis.
Attenuated total reflection (ATR)-IR spectroscopy has been
described as a promising technique for monitoring chemical
reactions on line.[23–25] Because most of the absorption comes
from the first few layers that are in contact with the probe[26]
rather than from the bulk, ATR-IR spectroscopy circumvents
the interference caused by the strong water absorptions,[18]
which allows the observation of the vibrations associated with
the substrate and product functional groups. To the best of
our knowledge, only ATR-IR spectroscopic studies of cellulose
involved the determination of cellulose structural characteris-
tics or cellulose–water interactions.[27,28] In this work, we dem-
onstrate the use of an autoclave reactor equipped with an
ATR-IR spectroscopic probe as a useful means to monitor the
transformations that occur during the solid-acid-catalyzed hy-
drolysis of cellulose in water in situ.
Pure component analysis
The IR spectra of several pure standards, which are the expect-
ed products of cellulose hydrolysis, such as glucose, fructose,
HMF, LA, acetic acid, and formic acid, were measured in water
at room and reaction temperature (typically 423 K) using the
ATR-IR spectroscopic probe. The results for glucose, the target-
ed product for cellulose hydrolysis, are given in Figure 1. The
most intense vibrations both at room and reaction tempera-
ture appeared at approximately n˜ =1034 and 1079 cmÀ1
,
which were assigned to C1ÀH and CÀO bends, respectively.[29]
Other prominent vibrations included CÀO stretching vibrations
Scheme 1. Proposed reaction pathway for cellulose hydrolysis.
ChemSusChem 2012, 5, 430 – 437
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
431

B. M. Weckhuysen et al.
1407 cmÀ1, respectively, and the
symmetric bending vibrations
for both compounds appear at
n˜ =1370 cmÀ1. These results indi-
cate that the observation of the
formation of these acids is possi-
ble by tracking the appearance
of the acid vibration at n˜ =
1710 cmÀ1, but identification of
the acid species requires obser-
vation of the vibrations that
appear between n˜ =1410 and
1170 cmÀ1.
Table 1. Selected vibrations and assignments of species depicted in Scheme 1.
Species
Frequency
[cmÀ1
Assignment
Species
Frequency
[cmÀ1
]
Assignment
]
A
B
cellulose
ꢀ1213
1161
1105
1026
1396
1362
1153
1101
1074
1038
900
1415
1347
1258
1157
1086
1064
multiple vibrations
CÀO str[b]
D
HMF
1664
1586
1556
1526
1400
1194
1060
1022
970
C=O ald[a] str
C=C str
C=C str
C=C str
CÀO str
C1ÀH bend
C2ÀH bend
CÀH bend
CÀO str
glucose
fructose
C=C str
ring=CÀOÀC=def
CÀH def[c]
CÀO str
ring def
CÀO, C-1-H bend
C1ÀH bend
CÀOÀC bend
CÀH bend
CÀH bend
multiple vibrations
CÀO str
CÀH op[d] def
CÀH op def
C=O acid str
C=O ald str
CH3 asym bend
CH3 sym bend
CÀO acid str
C=O acid str
CÀO acid str
799
E
F
LA
1710
C
1631[e]
1407
1369
1172
1710
1195
The IR spectra of HMF at room
temperature and 443 K are given
in Figure S4 in the Supporting
Information. At room tempera-
ture, a vibration assigned to a
C=O aldehyde stretch was ob-
served at n˜ =1664 cmÀ1. A series
of vibrations in the range ex-
CÀO str
formic acid
CÀH bend
[a] Aldehyde. [b] Stretch. [c] Deformation. [d] Out-of-plane. [e] Not observed at elevated temperatures.
pected for C=C stretches were observed at n˜ =1586, 1526, and
1400 cmÀ1. A =CÀOÀC= ring vibration was detected at n˜ =
1194 cmÀ1, a CÀH in-plane deformation at n˜ =1071 cmÀ1, a
ring deformation at n˜ =970 cmÀ1, and CÀH out-of-plane defor-
mations at n˜ =970 and 772 cmÀ1. Upon heating to 443 K, the
C=C vibrations at n˜ =1586 and 1556 cmÀ1 became significantly
more pronounced, and the appearance of vibrations that re-
semble those of LA, which include C=O and CÀO acid stretches
at n˜ =1706 and 1161 cmÀ1, respectively, suggested that HMF
decomposes to LA under hydrothermal conditions.[21,22]
The vibrational frequencies of fructose, which is likely
formed as an intermediate during HMF formation, exhibited
several similar vibrations as observed with glucose, particularly
at n˜ =1415, 1347, and 1157 cmÀ1. The most intense fructose vi-
brations appeared at n˜ =1064 and 1086 cmÀ1, which are shift-
ed to slightly higher wavenumbers than those observed with
glucose. Fructose was considerably less stable than glucose at
elevated temperatures even in the absence of a catalyst. The
time-resolved IR spectra of the reaction of fructose in water at
423 K are given in Figure 2, which shows the changes in vibra-
tional intensity with time. The spectrum recorded of the fruc-
tose solution in water at room temperature was used as a ref-
erence and was subtracted from the other spectra to empha-
size the changes that occur during heating. The most signifi-
cant vibrations appeared at n˜ =1707 and 1586 cmÀ1, which
corresponded to the formation of acids and HMF (see above).
These vibrations increased in intensity with time and those as-
sociated with fructose (around n˜ =1056 cmÀ1) decreased in in-
tensity. HPLC analysis of the solution after 3 h (Table 2, entry 1)
revealed the presence of acetic acid (3.2%), formic acid (2.3%),
LA (1.2%), and relatively large amounts of HMF (11.2%), which
supported the spectroscopic results. Large quantities of un-
Figure 1. IR spectra of glucose measured in water. Spectra at 423 K offset for
clarity.
at n˜ =1153 and 1109 cmÀ1, weaker vibrations assigned to a
CÀOÀC bend at n˜ =918 cmÀ1, and C2ÀH vibrations at n˜ =1396
and 1362 cmÀ1. Heating the solution to 423 K resulted in an in-
crease in the vibrational intensity of several bands, especially
at n˜ =1038 and 1074 cmÀ1; however, the frequency of these vi-
brations remained almost unchanged at elevated tempera-
tures. In the region between n˜ =4000 and 1900 cmÀ1 (not
shown), weak vibrations assigned to CÀH stretches (n˜ =
ꢀ2930 cmÀ1) and strong, broad water OÀH stretching vibra-
tions (n˜ =3700–3000 cmÀ1) were also observed.
The Supporting Information contains the IR spectra acquired
for LA, acetic acid, and formic acid, which all exhibited a C=O
acid stretch at around n˜ =1710 cmÀ1. The vibration from the
ketone component of LA appeared at n˜ =1631 cmÀ1 at room
temperature, but was not observed at elevated temperatures,
possibly as a result of the formation of a geminal diol. The CÀ known products were detected, which may include unconvert-
O acid stretch for formic acid, acetic acid, and LA appeared at
n˜ =1213, 1280, and 1172 cmÀ1, respectively. The acetic acid
and LA CH3 asymmetric bends are observed at n˜ =1396 and
ed fructose and soluble condensation products formed be-
tween HMF and fructose. A small amount of solid material was
432
www.chemsuschem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2012, 5, 430 – 437

Solid Acid-Catalyzed Cellulose Hydrolysis
Table 2. HPLC analytical results for cellulose, glucose, and fructose hydrolysis.[a]
Entry
Substrate
Glucose
[%]
2-Furalde-
hyde [%]
HMF
[%]
Acetic
acid [%]
Formic
acid [%]
LA
[%]
Unknown
[%]
Recovered solid
material[a] [%]
Glucose
selectivity [%]
1[c,d]
2[e,f,g]
3[h]
4[c,h]
5[c]
d-fructose
a-d-glucose
a-d-glucose
a-d-glucose
cellulose
cellulose
cellulose
cellulose
0.0
18.8
78.6
90.6
0.2
46.6
17.1
54.1
0.3
0.2
0.0
0.0
0.0
0.7
0.2
0.0
11.2
1.8
0.4
4.7
0.9
3.4
0.5
1.7
3.2
0.5
0.0
0.0
0.1
0.4
0.0
0.3
2.3
3.2
0.4
0.0
0.3
0.8
0.9
0.6
1.2
12.4
0.0
0.0
0.0
1.9
0.0
0.8
78.9
62.4
20.6
3.5
17.5
19.1
27.4
27.8
2.9
0.0
0.0
0.0
18.6
78.6
91.7
1.0
63.9
37.1
63.5
1.2
80.9
27.2
53.9
14.8
6
7[i]
8[j]
[a] Reaction conditions unless noted otherwise: regenerated cellulose (0.25 g), active carbon catalyst (0.25 g), H2O (25 g), T=423 K at autogenous pressure,
t=24 h. [b] Not including catalyst weight. [c] Without catalyst. [d] Regenerated cellulose substituted with d-fructose (0.31 g). [e] Regenerated cellulose sub-
stituted with a-d-glucose (5.0 g). [f] Activated carbon catalyst substituted with SZR (0.56 g). [g] T=453 K, t=4 h. [h] Regenerated cellulose substituted with
a-d-glucose (0.28 g). [i] T=413 K. [j] Teflon-lined reactor.
Figure 2. Time-resolved differential IR spectra of the reaction of d-fructose in
water. Reaction conditions: d-fructose (0.31 g), H2O (25 g), 423 K.
Figure 3. Time-resolved differential IR spectra of the acid-catalyzed hydroly-
sis of glucose in water. Reaction conditions: d-glucose (5 g), SZR (0.58 g),
H2O (25 g), 453 K.
collected, but no glucose was detected in the reaction mixture
at the conclusion of the reaction.
acetic acid (0.5%), and small quantities of HMF (1.8%), fructose
(0.3%), mannose (0.4%), and 2-furaldehyde (0.2%). Glucose
likely degraded to (in)soluble humins and other products not
detected by HPLC, which accounts for the relatively high
amount of unknown products. The analytical results corre-
sponded well with the vibrations observed by using ATR-IR
spectroscopy, with the most prominent vibrations correspond-
ing to the formation of LA and consumption of glucose.
Reactions of glucose with solid-acid catalysts
The time-resolved IR spectra of glucose in water in the pres-
ence of a sulfated zirconia catalyst (SZR) are given in Figure 3.
Upon reaching reaction temperature [time (t)=0 h], the
most prominent vibrations were those associated with glucose
at n˜ =1074 and 1038 cmÀ1, which decreased in intensity as the
reaction proceeded. A weak vibration appeared in the acid
region at n˜ =1710 cmÀ1, and the intensity of the vibration in-
creased with time. Moreover, vibrations at n˜ =1407 and
1366 cmÀ1 were observed, which could be attributed to either
formic acid or the LA CH3 asymmetric and symmetric bends,
respectively. A vibration at n˜ =1175 cmÀ1 resembled the CÀO
acid stretch in LA. A weak vibration at n˜ =1593 cmÀ1 also in-
creased slightly with increasing reaction time, which was at-
tributed to HMF accumulation (see above). HPLC results at the
conclusion of the reaction indicated relatively high LA concen-
trations (12.4%) based on initial carbon (Table 2, entry 2).
Other products included glucose (18.8%), formic acid (3.2%),
The results of glucose degradation in the presence of the ac-
tivated carbon catalyst at 423 K are given in Table 2, entry 3.
Under these milder conditions, a larger proportion of the origi-
nal glucose was recovered at the conclusion of the reaction.
Small quantities of HMF and formic acid were detected, in ad-
dition to a large quantity of unknown products (20.6%). The
analytical results of glucose hydrolysis under similar conditions,
but in the absence of a catalyst, are given in Table 2, entry 4.
Under these conditions, more than 90% of the glucose starting
material was recovered. HMF (4.7%) and small quantities of
solid material were observed, but in less than half the amount
compared to fructose. In the absence of a catalyst, no acid
products were detected. Moreover, a significant reduction in
ChemSusChem 2012, 5, 430 – 437
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
433

B. M. Weckhuysen et al.
the quantity of unknown products was observed in the ab-
sence of a catalyst. These results suggest that the catalyst ac-
celerates the formation of glucose degradation by-products.
nonregenerated microcrystalline Avicel cellulose, except that
considerably lower conversions were observed (<5%), and vir-
tually no change in absorption intensity was observed after
the solution had reached reaction temperature, which indi-
cates that very little cellulose hydrolysis occurred in the ab-
sence of a catalyst.
Acid-catalyzed hydrolysis of cellulose
The time-resolved ATR-IR spectra of the regenerated cellu-
lose hydrolysis in the presence of a solid-acid catalyst are
given in Figure 5a, and the change in absorbance of selected
The hydrothermal treatment of regenerated cellulose was stud-
ied at 423 K in the absence of a catalyst. The time-resolved
ATR-IR spectra are given in Figure 4. The spectrum of the cellu-
lose solution in water at room temperature was used as a ref-
erence and was subtracted from the other spectra to empha-
size the changes that occurred during heating. The only peaks
observed in the reference spectrum were attributed to water
Figure 4. Time-resolved differential IR spectra of cellulose in water. Reaction
conditions: regenerated cellulose (0.25 g), H2O (25 g), 423 K.
vibrations, which included a broad peak at n˜ =3335 cmÀ1, as-
signed to the water OÀH stretch, and a peak at n˜ =1639 cmÀ1
,
assigned to the water OÀH bend. Peaks associated with cellu-
lose were not observed because of its insolubility in water.
Heating the regenerated cellulose/water mixture to 423 K re-
sulted in the appearance of very weak vibrations at n˜ =1213,
1161, 1105, and 1026 cmÀ1, which grew slightly in intensity
with time. These vibrations differ to those assigned to glucose
(see above), and instead are likely to arise from the slow, par-
tial hydrothermal depolymerization of cellulose to smaller cel-
lulose chains that are slightly more water soluble and, there-
fore, more readily detected by using ATR-IR spectroscopy.
In agreement with the spectroscopic results, HPLC analysis of
the solution after treatment indicated the presence of only
trace quantities of glucose (0.2% based on initial carbon), and
similarly small quantities of other products, that is, formic acid
(0.3%), acetic acid (0.1%), and HMF (0.9%) were detected after
24 h (Table 2, entry 5). Relatively low solid-substrate mass
losses occurred, which corresponded to 19.1% of the starting
material. The discrepancy between the detected materials and
residual recovered solid cellulose likely corresponds to the for-
mation of water-soluble, oligomeric glucose that was not
detected by using HPLC. Similar results were observed using
Figure 5. a) Time-resolved differential IR spectra of the solid acid-catalyzed
hydrolysis of cellulose in water. Reaction conditions: regenerated cellulose
(0.25 g), active carbon catalyst (0.25 g), H2O (25 g), 423 K. b) Change in ab-
sorbance of selected vibrations vs. time.
vibrations are given in Figure 5b. In contrast to the uncata-
lyzed reaction, significant changes in the IR spectra were ob-
served. In the initial 3 h at reaction temperature, the most
prominent vibrations were observed at n˜ =1105 and
1153 cmÀ1, which continued to grow in intensity with increas-
ing reaction time. These vibrations resemble the weak vibra-
tions observed during heating in the absence of a catalyst
except that the intensity increased at a significantly faster rate.
After approximately 4 h, additional vibrations that strongly re-
semble the vibrations observed with pure samples of glucose
in water (see above) began to appear at n˜ =1075, 1038, and
892 cmÀ1 and continued to increase in intensity until about
16 h, at which point the intensities began to decrease slightly.
The appearance of these vibrations also coincided with the
appearance of vibrations near n˜ =1582 cmÀ1, which also in-
434
www.chemsuschem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2012, 5, 430 – 437

Solid Acid-Catalyzed Cellulose Hydrolysis
creased in intensity with time. These vibrations likely corre-
spond to the formation and accumulation of HMF, which ex-
hibits C=C vibrations in this region. After 4 h, vibrations in the
aldehyde and acid region at approximately n˜ =1709 cmÀ1
began to appear and became significant after 9–10 h. These re-
sults are indicative of the formation of LA and possibly formic
acid, both of which have a C=O acid vibration at around
n˜ =1710 cmÀ1. HPLC analysis of the solution after the reaction
revealed high solid-substrate mass losses corresponding to
72.8% of the starting material, confirmed the presence of high
quantities of glucose (46.6%), in agreement with the large
changes observed in the IR spectra (Table 2, entry 6). LA
(1.9%), formic acid (0.8%), and acetic acid (0.4%) were also de-
tected in the solution along with HMF (3.4%) and 2-furalde-
hyde (0.7%). Small quantities of insoluble humins were also
observed. Taken together, the spectroscopic and analytical re-
sults suggest that oligomeric glucose (i.e., cellulose with re-
duced chain length) is formed from cellulose from the time
the system reaches reaction temperature. After approximately
4 h, monomeric glucose and some HMF begin to appear in the
reactor, and several more hours are required until sufficient
acid concentrations are present to allow their detection.
ditions, prevention of the degradation of glucose to degrada-
tion products, such as HMF, cannot be avoided completely, but
the formation of acids and other degradation products can be
suppressed by operating at shorter reaction times. Reducing
the reaction temperature (Table 2, entry 7) results in slightly re-
duced byproduct formation, with the glucose/byproduct yield
ratio increasing from 6.5 at 423 K to 10.9 at 413 K (Figure S7 in
the Supporting Information). Both the overall glucose yield
and selectivity are lower at 413 than at 423 K, and a higher
proportion of unknown products has been detected. These un-
known products likely consist of soluble glucose oligomers
that were not completely depolymerized under the milder re-
action conditions, which is supported by the presence of
strong vibrations at n˜ =1061 and 1027 cmÀ1.
Reactions of microcrystalline cellullose
The solid acid-catalyzed hydrolysis of nonregenerated micro-
crystalline Avicel cellulose typically requires higher operating
temperatures (>443 K) for the reaction to proceed than re-
quired for regenerated cellulose. At elevated temperatures, the
presence of stainless steel, from which the ATR-IR sentinel is
constructed, may contribute to the decomposition of glu-
cose[30] as a decreased glucose stability was observed and rela-
tively large amounts of insoluble humins were formed at
453 K. A teflon liner was used to minimize contact between
the solution and the reactor walls. Although the stainless steel-
promoted degradation was not as prominent with the regener-
ated cellulose at lower temperatures, the use of a teflon liner
nevertheless resulted in higher overall glucose yields and a
glucose/byproduct ratio of 16.4 (Table 2, entry 8). Minimizing
contact of the solution with the metal is, therefore, necessary
for high glucose selectivity, especially at higher temperatures.
The time-resolved ATR-IR spectra of Avicel cellulose at 453 K
are given in Figure 6. Only very weak vibrations were observed
˜
The cellulose starting material, which is insoluble in the
aqueous solution, was initially not observed in the ATR-IR spec-
tra most likely because of poor contact of the cellulose with
the probe. In the absence of a catalyst, the glycosidic linkages
in cellulose remained largely intact; hence, significant changes
were not observed in the IR spectra. In the presence of a suit-
able catalyst, these linkages are susceptible to disruption, and
the initial vibrations observed by using ATR-IR spectroscopy
are attributed to smaller cellulose fragments that formed
through hydrolysis of some of the interior glycosidic linkages
in the original cellulose material. Continued hydrolysis of the
glycosidic bonds in the smaller fragments eventually resulted
in the formation of glucose, the accumulation of which is signi-
fied by the vibrations at n˜ =1075 and 1038 cmÀ1. The C1ÀH vi-
brations of the larger fragments, which still contain a consider-
able quantity of glycosidic linkages, are distinct from the vibra-
tions observed with glucose, which lack these linkages and
allow the rate of glucose formation to be monitored by using
in situ ATR-IR spectroscopy.
at n=1071 and 1026 cmÀ1 until the solution reached 453 K, at
which point a maximum vibrational intensity was observed.
The formation of HMF, which proceeds rapidly with the in-
creased accumulation of glucose in the reaction solution, likely
proceeds through the formation of fructose from glucose as
suggested by previous mechanistic studies on HMF forma-
tion.[16,17] Previous studies have indicated that the dehydration
of fructose to form HMF occurs rapidly,[16] which supports the
very low concentrations of fructose observed at the conclusion
of the reaction and the spectroscopic results depicted above.
The strongest fructose vibrations, which appear at n˜ =1086
and 1064 cmÀ1, overlap significantly with those from the cellu-
lose fragments and glucose, which precludes the direct obser-
vation of these materials by using ATR-IR spectroscopy. The
delay in the formation of the acids, indicated by the vibrations
at n˜ =1710 cmÀ1, supports earlier reports that HMF is an inter-
mediate product in acid-catalyzed glucose degradation.[16,17]
These spectroscopic results suggest that, under the given con-
Figure 6. Time-resolved differential IR spectra of the acid-catalyzed hydroly-
sis of cellulose in water. Reaction conditions: Avicel cellulose (5 g), SZR
(0.58 g), H2O (25 g), 453 K.
ChemSusChem 2012, 5, 430 – 437
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
435

B. M. Weckhuysen et al.
Even under these conditions, oligomeric cellulose and glucose
were less stable than at lower temperatures, evidenced by the
continual decrease in vibrational intensity at n˜ =1071 and
1026 cmÀ1. A gradual increase in intensity of vibrations at n˜ =
1710 and 1586 cmÀ1 indicated the accumulation of small quan-
tities LA and HMF. HPLC analysis of the solution and solids
after the reaction revealed that approximately 36% of the orig-
inal cellulose was converted with 25.7% selectivity to glucose.
LA and HMF were formed with 7 and 5% selectivity, respec-
tively. The remainder of the material consisted of humins and
unconverted cellulose (55.6% selectivity) and small quantities
of acetic acid, formic acid, fructose, mannose, and 2-furalde-
hyde (7% selectivity together).
Experimental Section
The cellulose used was microcrystalline Avicel PH 101 (Fluka, parti-
cle size ꢀ50 mm), and the ionic liquid used was 1-butyl-2-methyli-
midazolium chloride (BMImCl, Basionic ST 70, purity ꢁ95%). Both
were purchased from Sigma–Aldrich. Sulfated zirconia (SZR) was
supplied by Saint-Gobain NorPro, and activated carbon CAP SUPER
was kindly provided by Norit. The carbon was activated using
phosphoric acid with an acid density of approximately 1.1 mmol
H+ equivalents per gram catalyst as determined by using titration.
A suspension of 40 mgmLÀ1 per g of activated carbon in deminer-
alized water resulted in a solution of pH 2.3. All compounds were
used without further purification or pretreatment, unless stated
otherwise.
For cellulose pretreatment, BMImCl was heated at 423 K for 2 h
under a nitrogen flow before the addition of cellulose. After an in-
cubation period of 30 min, heating ceased, and the cellulose was
precipitated by adding 5 parts (v/v) of hot demineralized water
(>353 K). The liquid BMImCl solutions were stirred at all times. The
solutions were stored at 48C overnight to allow the precipitate to
settle. The precipitates were separated by centrifugation (Thermo
Scientific SL 40R, 10 min at 4000 rpm) and washed three times
with hot (> 353 K) and once with cold demineralized water. The re-
sultant solid is referred to as regenerated cellulose. The moisture
content of the regenerated cellulose was adjusted to 95%, which
resulted in a colloidal suspension that was stable for weeks. The
suspension was stored at 277 K until tested.
Conclusions
Solid acid-catalyzed hydrolysis represents an attractive method
for the valorization of cellulose into glucose and other valuable
platforms chemical. In situ ATR-IR spectroscopy is a powerful
technique that can be used to monitor the transformations
that occur during this process, which includes the formation of
the reaction products and by-products. The most relevant spe-
cies, which include glucose and glucose oligomers, fructose,
HMF, and several acids, are readily observed and generally dis-
tinguishable based on differences in vibrational characteristics
even in the presence of water under elevated temperatures
and pressures. In the case of regenerated cellulose with the ac-
tivated carbon catalyst, relatively short reaction times (ꢀ5 h)
at 423 K minimized the formation of glucose degradation
products.
Hydrolysis reactions were conducted in a 40 mL Parr stainless steel
autoclave equipped with an MT ReactIR 45 m attenuated total re-
flection (ATR)-IR sentinel with a diamond probe. To prevent catalyt-
ic breakdown of some of the products by the reactor wall at ele-
vated temperatures (>443 K), a Teflon insert was required. The
temperature was monitored by using a thermocouple, and stirring
was conducted by using a magnetic driver equipped with an im-
pellor at approximately 750 rpm. The head space of the reactor
was flushed with nitrogen prior to heating (2 KminÀ1). All experi-
ments were performed at autogenous pressure. After the reaction,
the solids were separated by centrifugation. The conversion of cel-
lulose was calculated by the weight difference of cellulose before
and after reaction. The supernatants were analyzed for typical
products, such as glucose, fructose, HMF, furfural, formic acid, LA,
and acetic acid (HPLC, Agilent 1100 series, equipped with an IR an-
alyzer, UV detector and a Biorad AMINEX HPX-87H column).
Compounds were identified by using elution times and quantified
by using calibration standards. The solid residue was dried (323 K,
vacuum) until no further mass change was observed. The products
yields are defined as the molar ratio of carbon in the product over
the mol of carbon in the charged cellulose (ꢁ100%).
A combination of analytical and spectroscopic results re-
vealed key aspects of the catalytic system. The hydrolysis of
cellulose in the absence of a catalyst is slow, but still observ-
able by using ATR-IR spectroscopy. The presence of a solid-acid
catalyst significantly increases the rate of cellulose hydrolysis
to form glucose oligomers, the continued hydrolysis of which
eventually results in the accumulation of monomeric glucose.
The formation of other by-products proceeds through the iso-
merization of glucose to fructose, which is less hydrothermally
stable and rapidly reacts to form by-products, such as HMF.
The formation of HMF and other subsequent degradation
products, which include LA and formic acid, are also signifi-
cantly accelerated by the presence of solid-acid catalysts.
Therefore, operation at reduced glucose residence times re-
sults in improved glucose selectivity. Harsher conditions (i.e., Acknowledgements
higher temperatures and the presence of stronger acid cata-
lysts) tend to result in the increased formation of acid products
and decreased glucose selectivity. Microcrystalline Avicel cellu-
lose hydrolysis requires harsh conditions; therefore, higher acid
content and lower glucose selectivity are typically observed
during the hydrolysis of Avicel cellulose relative to regenerated
cellulose, which is more susceptible to hydrolysis at lower reac-
tion temperatures. Minimizing the contact to metal in the reac-
tor, especially at higher temperatures, is important to decrease
the formation of degradation products.
J.Z. and B.M.W. gratefully thank the National Science Foundation
International Research Fellowship Program for support of this re-
search under Award No. 0856754. This research has been per-
formed within the framework of the CatchBio program. The au-
thors gratefully acknowledge the support of the Smart Mix Pro-
gram of the Netherlands Ministry of Economic Affairs and the
Netherlands Ministry of Education, Culture and Science. Thanks
are also extended to Pieter C. A. Bruijnincx for many useful
discussions.
436
www.chemsuschem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2012, 5, 430 – 437

Solid Acid-Catalyzed Cellulose Hydrolysis
[16] J. Pang, A. Wang, M. Zheng, T. Zhang, Chem. Commun. 2010, 46, 6935–
6937.
[17] S.-J. Kim, A. A. Dwiatmoko, J. W. Choi, Y.-W. Suh, D. J. Suh, M. Oh, Biore-
sour. Technol. 2010, 101, 8273–8279.
Keywords: cellulose · glucose · heterogeneous catalysis ·
hydrolysis · ir spectroscopy
[18] J. Hegner, K. C. Pereira, B. DeBoef, B. L. Lucht, Tetrahedron Lett. 2010, 51,
2356–2358.
[19] A. Takagaki, C. Tagusagawa, K. Domen, Chem. Commun. 2008, 5363–
5365.
[1] R. D. Perlack, L. L. Wright, A. F. Turhollow, R. L. Graham, B. J. Stokes, D. C.
Erbach, U. S. Department of Energy, Biomass as Feedstock for a Bioener-
gy and Bioproducts Industry: The Technical Feasibility of a Billion-Ton
Annual Supply, 2005.
[20] S. Suganuma, K. Nakajima, M. Kitano, Y. Daizo, H. Kato, S. Hayashi, M.
Hara, J. Am. Chem. Soc. 2008, 130, 12787–12793.
[21] B. Girisuta, L. P. B. M. Janssen, H. J. Heeres, Chem. Eng. Res. Des. 2006,
84, 339–349.
[22] J. B. Binder, A. V. Cefali, J. J. Blank, R. T. Raines, Energy Environ. Sci. 2010,
3, 765–771.
[2] Renewable Energy Road Map 2007, http://europa.eu/legislation_sum-
maries/energy/renewable_energy/127065_en.htm.
[3] J. J. Bozell, J. E. Holladay, D. Johnson, J. F. White, Top Value Added Candi-
dates from Biomass, Volume I: Results of Screening for Potential Candi-
dates from Sugars and Synthesis Gas, Pacific Northwest National Labora-
tory: Richland, WA, 2007.
[23] K. Kervinen, M. Allmendinger, M. Leskelꢃ, T. Repo, B. Rieger, Phys. Chem.
Chem. Phys. 2003, 5, 4450–4454.
[24] K. Kervinen, H. Korpi, J. G. Mesu, F. Soulimani, T. Repo, B. Rieger, M. Le-
skelꢃ, B. M. Weckhuysen, Eur. J. Inorg. Chem. 2005, 2591–2599.
[25] J. Zakzeski, P. C. A. Bruijnincx, B. M. Weckhuysen, Green Chem. 2011, 13,
671–680.
[26] Z. Zhang, G. E. Ewing, Anal. Chem. 2002, 74, 2578–2583.
[27] H. Shinzawa, S.-I. Morita, K. Awa, M. Okada, I. Noda, Y. Ozaki, H. Sato,
Appl. Spectrosc. 2009, 63, 501–506.
[28] A. Watanabe, S. Morita, S. Kokot, M. Matsubara, K. Fukai, Y. Ozaki, J. Mol.
Struct. 2006, 799, 102–110.
[4] J. J. Bozell, G. R. Petersen, Green Chem. 2010, 12, 539–554.
[5] Kirk–Othmer Encyclopedia of Chemical Technology, 4th Ed. (Ed.: J. Krosch-
witz), Wiley, New York, 1993, 5, 476–496.
[6] R. P. Swatloski, S. K. Spear, J. D. Holbrey, R. D. Rogers, J. Am. Chem. Soc.
2002, 124, 4974–4975.
[7] H. B. Zhao, J. E. Holladay, H. Brown, Z. C. Zhang, Science 2007, 316,
1597–1599.
[8] F. Tao, H. Song, L. Chou, ChemSusChem 2010, 3, 1298–1303.
[9] F. Tao, H. Song, L. Chou, Carbohydr. Res. 2011, 346, 58–63.
[10] Y. Su, H. M. Brown, X. Huang, Z. C. Zhang, Appl. Catal. A: Gen. 2009, 361,
117–142.
[29] G. Socrates, Infrared and Raman Characteristic Group Frequencies: Tables
and Charts, Wiley, New York, 2001.
[11] M. B. Turner, S. K. Spear, J. D. Holbrey, R. D. Rogers, Biomacromolecules
2004, 5, 1379–1384.
[30] Q. Xiang, Y. Y. Lee, R. W. Torget, Appl. Biochem. Biotechnol. 2004, 115,
1127–1138.
[12] H. Zhao, C. L. Jones, G. A. Baker, S. Xia, O. Olubajo, V. N. Person, J. Bio-
technol. 2009, 139, 47–54.
[13] M. Kitano, D. Yamaguchi, S. Suganuma, K. Nakajima, H. Kato, S. Hayashi,
M. Hara, Langmuir 2009, 25, 5068–5075.
Received: October 10, 2011
[14] A. Onda, T. Ochi, K. Yanagisawa, Green Chem. 2008, 10, 1033–1037.
[15] R. Rinaldi, F. Schꢂth, ChemSusChem 2009, 2, 1096–1107.
Published online on February 7, 2012
ChemSusChem 2012, 5, 430 – 437
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
437