Which controls the depolymerization of cellulose in ionic liquids: The solid acid catalyst or cellulose?

Article abstract of DOI:10.1002/cssc.200900281

Cellulose is a renewable and widely available feedstock. It is a biopolymer that is typically found in wood, straw, grass, municipal solid waste, and crop residues. Its use as raw material for biofuel production opens up the possibility of sustainable biorefinery schemes that do not compete with food supply. Tapping into this feedstock for the production of biofuels and chemicals requires-as the first-step-its depolymerization or its hydrolysis into intermediates that are more susceptible to chemical and/or biological transformations. We have shown earlier that solid acids selectively catalyze the depolymerization of cellulose solubilized in 1-butyl-3-methylimidazolium chloride (BMIMCl) at 100°C. Here, we address the factors responsible for the control of this reaction. Both cellulose and solid acid catalysts have distinct and important roles in the process. Describing the depolymerization of cellulose by the equivalent number of scissions occurring in the cellulosic chains allows a direct correlation between the product yields and the extent of the polymer breakdown. The effect of the acid strength on the depolymerization of cellulose is discussed in detail. Practical aspects of the reaction, concerning the homogeneous nature of the catalysis in spite of the use of a solid acid catalyst, are thoroughly addressed. The effect of impurities present in the imidazolium-based ionic liquids on the reaction performance, the suitability of different ionic liquids as solvents, and the recyclability of Amberlyst 15DRY and BMIMCl are also presented.

Full text of DOI:10.1002/cssc.200900281

DOI: 10.1002/cssc.200900281
Which Controls the Depolymerization of Cellulose in Ionic
Liquids: The Solid Acid Catalyst or Cellulose?
Roberto Rinaldi,* Niklas Meine, Julia vom Stein, Regina Palkovits, and Ferdi Schꢀth*[a]
Cellulose is a renewable and widely available feedstock. It is a
biopolymer that is typically found in wood, straw, grass, munic-
ipal solid waste, and crop residues. Its use as raw material for
biofuel production opens up the possibility of sustainable bio-
refinery schemes that do not compete with food supply. Tap-
ping into this feedstock for the production of biofuels and
chemicals requires—as the first-step—its depolymerization or
its hydrolysis into intermediates that are more susceptible to
chemical and/or biological transformations. We have shown
earlier that solid acids selectively catalyze the depolymerization
of cellulose solubilized in 1-butyl-3-methylimidazolium chloride
(BMIMCl) at 1008C. Here, we address the factors responsible
for the control of this reaction. Both cellulose and solid acid
catalysts have distinct and important roles in the process. De-
scribing the depolymerization of cellulose by the equivalent
number of scissions occurring in the cellulosic chains allows a
direct correlation between the product yields and the extent
of the polymer breakdown. The effect of the acid strength on
the depolymerization of cellulose is discussed in detail. Practi-
cal aspects of the reaction, concerning the homogeneous
nature of the catalysis in spite of the use of a solid acid cata-
lyst, are thoroughly addressed. The effect of impurities present
in the imidazolium-based ionic liquids on the reaction perfor-
mance, the suitability of different ionic liquids as solvents, and
the recyclability of Amberlyst 15DRY and BMIMCl are also pre-
sented.
Introduction
Currently, the production of biofuels from carbohydrates is a
field of intense research activity.[1] In particular processes for bi-
ofuels that start from glucose have been attracting immense
attention.[1] However, there is concern that sugars obtained di-
rectly from saccharose (typically produced from sugarcane or
sugar beet) or semi-directly from the hydrolysis of starch (pres-
ent in corn and other crops) can compete with food supply.[2a]
Accordingly, cellulose emerges as a feedstock that is more suit-
able than starch or saccharose. Firstly, this biopolymer is the
most abundant organic compound on the planet.[2] In addition,
the utilization of cellulose, present in lignocellulosic materials
such as wood, straw, grass, municipal solid waste, and crop
residues, does not compete with food supply.[3a] Nevertheless,
the recalcitrance of this biopolymer poses serious challenges
to the chemical and biological processing of cellulose.[3–5]
Many attempts to hydrolyze lignocellulosic materials into
glucose and other fermentable sugars have been reported
during the last century.[5] These processes are commonly re-
ferred to as saccharification of wood. Several mineral acids,
such as H2SO4,[6c,h,i] HCl,[6d,f] or even HF,[6j] have been employed
as liquid acid catalysts for the hydrolysis of cellulose. Some of
these processes were scaled up to pilot-plant level.[6c–e,h,i] Al-
though valuable results were obtained during the pilot tests,
the full-scale production of glucose via saccharification of
lignocellulosic materials was never commercially implemented
because of problems concerning corrosion, energy demand,
catalyst recovery, and decomposition of sugars.[5,6a,b]
line structure of cellulose imposes important restrictions on
the kinetics of the heterogeneous hydrolysis of cellulose.[7] In
practice, a specific pretreatment of the lignocellulosic material
(e.g., by partial chemical degradation, mechanical comminu-
tion, activation by swelling, or by several other processes) is
usually applied to partially disrupt the supramolecular struc-
ture of the cellulosic fibers, enhancing the reactivity of cellu-
lose towards hydrolysis.[8] The solubilization of cellulose in
ionic liquids, however, completely disassembles the supra-
molecular structure of the cellulosic fibers.[9] This strongly im-
proves the reactivity of cellulose.[10]
We have shown earlier that solid acids catalyze the depoly-
merization of cellulose dissolved in 1-butyl-3-methylimidazoli-
um chloride (BMIMCl) under mild conditions.[11] This reaction is
performed efficiently using macroreticulated acid resins
(known by the brand name Amberlyst 15DRY[12]). The total de-
polymerization of cellulose can easily be achieved. In practice,
however, it is much more interesting to stop the process at
the stage in which cellooligomers are the main products.[11]
At first glance, the use of this acidic resin shows several sig-
nificant advantages over molecular acids. Firstly, the break-
down of cellulose by Amberlyst 15DRY in BMIMCl is more con-
[a] Dr. R. Rinaldi, N. Meine, J. vom Stein, Dr. R. Palkovits, Prof. F. Schꢀth
Department of Heterogeneous Catalysis
Max-Planck-Institut fꢀr Kohlenforschung
Kaiser-Wilhelm-Platz 1, 45470 Mꢀlheim an der Ruhr (Germany)
Fax: (+49)208 306 2995
E-mail: rinaldi@mpi-muelheim.mpg.de
schueth@mpi-muelheim.mpg.de
Homepage: http://www.mpi-muelheim.mpg.de/kofo/
In the majority of saccharification processes, the lignocellulo-
sic raw material is typically not solubilized in the reaction
medium.[5] Under such conditions the supramolecular/crystal-
266
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 266 – 276

Depolymerization of Cellulose in Ionic Liquids
trollable.[4] Stopping the reaction at the cellooligomer stage
(Figure 1) allows the separation of the cellooligomers from the
ionic liquid by addition of water.[4] This enables recycling of the
ionic liquid. In contrast, the hydrolysis of cellulose is very fast
when using soluble acids [e.g., mineral acids or p-toluenesul-
in the activation of cellulose towards hydrolysis. Practical as-
pects of the reaction, concerning the homogeneous nature of
the catalysis in spite of using a solid catalyst, are discussed in
detail. We correct our previous conclusion on the heterogene-
ous nature of the catalyst.[11] The effect of impurities in the imi-
dazolium-based ionic liquids on
the reaction performance, the
suitability of different ionic liq-
uids as reaction medium, and
the recyclability of catalyst and
ionic liquid are also presented
here.
Results and Discussion
Effect of the amount of acid
catalyst
The addition of H3O+ species
into a solution of cellulose in
BMIMCl initiates a complex reac-
tion chain, as described in
Figure 1. In the first step, cellu-
lose undergoes depolymeriza-
tion via hydrolysis of 1,4-b-glyco-
sidic bonds. Either smaller 1,4-b-
glucans (cellooligomers) or glu-
cose can be formed at this
stage. In the presence of acidic
species, glucose is likely to be
dehydrated,
producing
a
number of compounds[15] such
as 5-hydroxymethylfurfural (5-
HMF), levulinic acid, formic acid,
and several others (Figure 1).[16,17]
These products are prone to re-
combination with sugars or oli-
gosaccharides via aldol conden-
sation, resulting in polymers
with undefined structures and
stoichiometry called humins.[15,18]
Figure 1. Selected products formed by acid-catalyzed reactions starting from cellulose. Hemicelluloses are typical
impurities of cellulose. Therefore, xylose and other monosaccharides as well as their degradation products (furfu-
ral, furoic acid, and others) are also present in the reaction mixture.
fonic acid (p-TSA)] leading to the formation of a multitude of
dehydration and polymeric byproducts, making work-up of the
product mixture extremely difficult.[13] Up to now, the efficient
extraction of glucose and its dehydration products from the re-
action medium has not been achieved.[4,14]
The formation of the mentioned products in Figure 1 is cata-
lyzed by acid. It is thus expected that by changing the amount
and/or the nature of the acid catalyst, the reaction rates for
the different steps should change differently, altering the distri-
bution of products obtained from cellulose.[13d] The effect of
the amount of catalyst on the depolymerization of cellulose
was investigated by performing the reactions with Amberlyst
15DRY amounts ranging from 0.46 to 6.90 mmol of H3O+ for
the reaction batch. These amounts correspond to concentra-
tions of H3O+ ranging from 4.25 to 63.7 mmolgꢀ1 of reaction
mixture, respectively.
Herein, we comprehensively address the factors responsible
for controlling the depolymerization of cellulose when using
Amberlyst 15DRY in BMIMCl. The effects of catalyst amount
and of temperature on the distribution of the products are in-
vestigated. The description of the reaction by the equivalent
number of scissions occurring in the cellulosic chains allows a
direct correlation between the yield of products and the prog-
ress of the depolymerization reaction. As a result, the roles of
the properties of cellulose and of the solid acid catalyst in con-
trolling of the reaction could be elucidated. Regarding the
properties of the catalyst, acid strength is an important factor
Figure 2 shows the effect of the amount of Amberlyst
15DRY on the reaction performance. As expected, the higher
the loading of Amberlyst 15DRY in the reaction, the faster the
hydrolysis rate. Starting from an amount of Amberlyst 15DRY
equivalent to 0.46 mmol of H3O+, a glucose yield of 0.58%
ChemSusChem 2010, 3, 266 – 276
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
267

R. Rinaldi, F. Schꢀth et al.
strongly enhanced when using 6.9 mmol of H3O+, reaching
7.35% at 3 h.
Figure 2b shows that the glucose/5-HMF ratio reached a
peak at around 2.3 mmol of H3O+. At this value, the yield of
glucose was 8.6 times higher than that of 5-HMF (1008C, 3 h).
In sequence, for the reactions performed with larger amounts
of acid catalyst, there was a marked decrease of the glucose/5-
HMF ratio. This suggests that the formation of 5-HMF is strong-
ly favored when using larger amounts of acid catalyst (Figur-
e 2b). However, as will be discussed in the next sections, an as-
sessment of the product distribution with respect to the
extent of depolymerization is necessary to correctly evaluate
the influence of the amount of Amberlyst 15DRY on the reac-
tion selectivity.
Extrapolating the trends for the yields of glucose and 5-HMF
shown in Figure 2b, a maximum yield of glucose would be ex-
pected when using an amount of Amberlyst 15DRY equivalent
to 11.9 mmol of H3O+. Under these conditions, the predicted
yields of glucose and 5-HMF are 74.8 and 25.2%, respectively.
In practice, however, the reaction produced only 25.0% of glu-
cose and 11.8% of 5-HMF at 1008C for 3 h. These low yields
were due to the substantial formation of humins as byprod-
ucts.
The reaction performed with Amberlyst 15DRY displays an
induction period for the production of glucose. This is not ob-
served using equivalent amounts of p-toluenesulfonic acid (p-
TSA).[11] The induction period depends heavily on the amount
of acid resin used for the reaction (Figure 2c). Here, the induc-
tion period is defined as the reaction time required for a yield
of glucose of 0.02%. The increase in the catalyst amount from
0.46 to 6.9 mmol of H3O+ makes the induction period much
shorter. The induction period decreased from 1.9 h to less than
5 min under the latter reaction conditions (Figure 2c). In other
words, performing the reaction with a large amount of Amber-
lyst 15DRY resembled carrying out the reaction with a soluble
strong acid, for which no induction period for the production
of glucose was noticed.[11]
Assessing the depolymerization of cellulose
The yield of glucose cannot describe fully the extent of the
breakdown of cellulose in the course of the process (Figure 2).
For this purpose, an assessment of the degree of polymeri-
zation (DP) of the cellulose solubilized in the reaction mixture
is needed to describe the number of scissions that have oc-
curred in the cellulosic chains at a given reaction time [Equa-
tion (1)].[19] Figure 3 shows the number of scissions over time
for reactions performed with different amounts of Amberlyst
15DRY.
Figure 2. Depolymerization of cellulose over Amberlyst 15 DRY. a) Effect of
the amount of catalyst on the yield of glucose. b) Yield of glucose and 5-
HMF for reactions performed at 1008C for 3 h using different amounts of
catalyst. c) Dependence of the induction period on the amount of catalyst.
Reaction conditions: cellulose [5 g, corresponding to ca. 31 mmol of anhy-
droglycose units (AGU); C6H10O5: DPo 650 AGU] dissolved in 100 g BMIMCl,
Amberlyst 15DRY (4.6 mmolH3O+ ꢀ1), water (111 mmol), 1008C.
g
was reached after 3 h. Under these conditions the formation of
5-HMF was not detected. Both, yields of glucose and of 5-HMF,
did not increase linearly with the amount of acid used in the
reaction (Figure 2b). Increasing the amount of Amberlyst
15DRY by 15 times (equivalent to 6.9 mmol of H3O+) resulted
in a yield of glucose 50 times higher (28.8%, 3 h) than that ob-
served when using an amount of acid resin equivalent to
0.46 mmol of H3O+. However, the formation of 5-HMF was also
In contrast to the release of glucose, which showed an in-
duction period (Figure 2a), the depolymerization of cellulose
started immediately when Amberlyst 15DRY was added to the
cellulose/BMIMCl solution (Figure 3). Furthermore, the linear
correlation between the scission rates and the catalyst
amounts (Figure 4) showed that depolymerization of cellulose
solubilized in BMIMCl is a first-order reaction with respect to
the amount of acid catalyst. The frequency of scissions at
268
www.chemsuschem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 266 – 276

Depolymerization of Cellulose in Ionic Liquids
Figure 5. Depolymerization of cellulose over Amberlyst 15 DRY. Yield of se-
lected products versus the number of scissions occurring in the cellulosic
chains. These correlations were performed using the whole set of data gath-
ered from the reactions displayed in Figure 2.
Figure 3. Equivalent number of scissions occurring in the cellulose chains
during the course of the depolymerization using Amberlyst 15DRY. Reaction
conditions: 5 g cellulose (corresponding to ca. 31 mmol of AGU; C6H10O5:
DPo 650 AGU) dissolved in 100 g BMIMCl, Amberlyst 15DRY (4.6 mmol
H3O+ ꢀ1), water (111 mmol), 1008C.
g
with the extent of the depolymerization of cellulose. A similar
consistent pattern emerges in the correlation between yield of
5-HMF and the number of scissions occurred in the cellulosic
chains. It is interesting to note that the selectivity for the prod-
ucts depends heavily on the extent of depolymerization.
Therefore, for a fair comparison between yields or selectivities
of different reactions, these values must be taken at the same
extent of reaction.
To determine whether the concentration of acid catalyst al-
tered the product distribution, the glucose/5-HMF ratio, ob-
tained from reactions performed with 4.6 and 6.9 mmol of
H3O+ (Amberlyst 15DRY), were analyzed with respect to the
extent of depolymerization, as shown in Figure 6. Contrary to
expectations, this study did not find a large difference be-
tween the glucose/5-HMF ratios, at a given depolymerization
extent, for the reactions performed with 4.6 and 6.9 mmol of
H3O+ (Amberlyst 15DRY). Increasing the amount of catalyst
from 4.6 to 6.9 mmol of H3O+, that is, by 50%, favors the for-
mation of 5-HMF by only about 10%. In fact, the consistent
trends between yields and reaction extent, as shown in
Figure 5, are only possible because of the small influence of
Figure 4. Correlation between the amount of catalyst and the scission fre-
quency for reactions performed at 1008C using Amberlyst 15DRY as H3O+
source.
1008C was about 17 scissionshꢀ1 (mmolH3O+)ꢀ1 (Amberlyst
15DRY).
The extent of depolymerization and the distribution of
products
Revisiting the set of data shown in Figures 2 and 3, it is appar-
ent that the high concentration of acid not only enhanced the
rate of depolymerization of cellulose but also appeared to
favor the decomposition of glucose into 5-HMF. To understand
how the extent of depolymerization regulates both yields of
glucose and of 5-HMF, the correlation between these yields
and the number of scissions occurring in the cellulosic chains
was plotted, as shown in Figure 5.
The single most striking observation that emerges from the
data comparison shown in Figure 5 is that the formation of
glucose displays a rather consistent trend directly associated
Figure 6. Effect of acid catalyst amount on the glucose/5-HMF ratio.
ChemSusChem 2010, 3, 266 – 276
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
269

R. Rinaldi, F. Schꢀth et al.
the amount of acid catalyst on the product yields at a given re-
action extent.
Effect of temperature
Temperature seriously affects the depolymerization rate of cel-
lulose. The reaction carried out at 1208C was ca. 100 times
faster than a reaction performed at 808C (Figure 7). This en-
Figure 8. Effect of temperature on the glucose/5-HMF ratio.
The nature of Amberlyst 15DRY-catalyzed hydrolysis
Performing the depolymerization of cellulose in a controlled
manner is one of the main motivations for the use of Amber-
lyst 15DRY instead of p-TSA.[11] At first glance, p-TSA seems to
be a more active catalyst than Amberlyst 15DRY. Indeed, the
formation of glucose (Figure 9a) and the cellulose breakdown
(Figure 9b) take place at a faster rate using p-TSA than with
sulfonated acidic resin. Furthermore, the reactions performed
with Amberlyst 15DRY always display an induction period for
the production of glucose, which depends heavily on the
amount of solid acid present in the reaction medium (Figur-
e 2c). In the case of the p-TSA-catalyzed reaction, the induction
period for production of glucose is not evident (Figure 9a). Al-
though Amberlyst 15DRY and p-TSA appear to be different, as
discussed so far, the correlation between yield of glucose and
extent of the depolymerization process (Figure 9c) shows that
the real performance of both catalysts is otherwise the same
considering the amount of glucose released in the progress of
the reactions, as can be seen in Figure 9c.
Figure 7. Effect of temperature on the depolymerization rate of cellulose.
Reaction conditions: cellulose (5 g, corresponding to ca. 31 mmol of AGU;
C6H10O5: DPo 650 AGU) dissolved in 100 g BMIMCl, Amberlyst 15DRY
(1.15 mmol of H3O+), water (111 mmol).
hancement in the reaction rate is comparable to an increase of
the catalyst amount from 1.15 to 6.9 H3O+ mmol in the batch
(see Figure 3). The apparent activation energy for the depoly-
merization of cellulose was determined from the data shown
in Figure 7. The value obtained in BMIMCl, 108ꢁ5 kJmolꢀ1, is
slightly lower than that reported for the dilute-acid hydrolysis
of cellulose in water-based systems (118–131 kJmolꢀ1).[5,20]
However, the present value of apparent activation energy for
the depolymerization of cellulose in BMIMCl is identical to that
reported elsewhere (111ꢁ12 kJmolꢀ1) for the hydrolysis of cel-
lobiose in 1-ethyl-3-methyl-imidazolium chloride (EMIMCl).[13d]
Thus, cellulose dissolved in ionic liquids appears to behave
similarly to cellobiose. This indicates that the solubilization of
cellulose does disassemble the supramolecular structure of this
polymer, fully removing its physical protections against hydro-
lytic processes.[10]
Because Amberlyst 15DRY and p-TSA display similar perform-
ances concerning the production of glucose over the reaction
extent, the question arises: Where does the reaction take
place—on the catalyst surface or in the bulk solution?
Adding Amberlyst 15DRY into BMIMCl starts an ion-ex-
change process, as shown in Figure 10. Our former conclusion
on the heterogeneity of the reaction with Amberlyst 15DRY as
catalyst[11] was disguised due to presence of impurities found
in the technical-grade BMIMCl (vide infra) and thus has to be
corrected. In those experiments,[11] the amount of acid was de-
termined directly by pH measurements of aqueous solutions
of the reaction aliquots. The basic impurities (mainly N-methyli-
midazole) present in the technical-grade BMIMCl establish
buffer equilibria, resulting in much higher pH values than
would be expected without the buffer effect. Thus, reliable re-
sults on the acid amount present in the ionic liquid medium
can be obtained by acid–base titration only.
Although cellulose dissolution in BMIMCl leads to a lowering
of the apparent activation energy for hydrolysis, this value is
still almost identical to that found by us for the acid-catalyzed
degradation of glucose in BMIMCl (114ꢁ6 kJmolꢀ1). This indi-
cates that the further decomposition of glucose should easily
take place parallel to the hydrolysis of cellulose in a batch reac-
tion. Because of the almost identical activation energies, the
effect of temperature on the glucose to 5-HMF ratio, at a given
reaction extent, is not very pronounced, as shown in Figure 8.
Titrating the ionic liquid phase, separated from a suspension
of Amberlyst 15DRY in BMIMCl, revealed that H3O+ was pro-
gressively released in this liquid within the first hour of contact
between both phases (Figure 11). In sequence, the concentra-
270
www.chemsuschem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 266 – 276

Depolymerization of Cellulose in Ionic Liquids
Figure 11. H+ release profile of Amberlyst 15DRY in contact with BMIMCl at
different temperatures.
agreement with the low initial rate of depolymerization found
for the reaction performed with Amberlyst 15DRY (Figure 9b).
Furthermore, because only ca. 80% of the acidic content of
Amberlyst 15 DRY is present after the equilibration time (1 h),
this acidic resin appears to be accordingly less active than p-
TSA. Taking into account the real amount of acid present in
both reaction media, the catalysts display otherwise similar ac-
tivity, as expressed by the frequency of scissions found for the
reactions performed with p-TSA and Amberlyst 15DRY (16 and
17 scissionshꢀ1 (mmolH3O+)ꢀ1, respectively).
Regarding the fact that the depolymerization of cellulose
with Amberlyst 15DRY occurs in the solution phase, an explan-
ation for the slower initial rate for the formation of glucose is
still required. Depolymerization of cellulose solubilized in ionic
liquid takes place preferentially at the larger cellulose mole-
cules, as we have reported recently.[11] Accordingly, the induc-
tion period for the release of glucose is observed. The results
obtained so far showed that the production of cellooligomers
is preferred over the formation of glucose in the earlier stages
of the reaction. This preference can be understood considering
that cellooligomers can be produced by cleaving the polymeric
chain at any position, while formation of glucose requires spe-
cific cleavage at the ends of cellulose. In this manner, the for-
mation of glucose is not favored statistically in the beginning
of the depolymerization process, because the reaction mixture
contains large polymeric chains at this reaction stage. There-
fore, the induction period for production of reducing sugars
and the high selectivity for cellooligomers are features linked
to cellulose rather than to the acid catalyst used.
Figure 9. Depolymerization profiles of cellulose with Amberlyst 15DRY or p-
TSA as catalyst. a) Yield of glucose. b) Equivalent number of scissions. c) Cor-
relation between yield of glucose and equivalent number of scissions. Reac-
tion conditions: cellulose (5 g, corresponding to ca. 31 mmol of AGU;
C6H10O5: DPo 650 AGU) dissolved in 100 g BMIMCl, Amberlyst 15DRY or p-
TSA (2.3 mmol of H3O+), water (111 mmol), 1008C.
Amberlyst 15DRY should be regenerated before reuse, due
to the release of its H3O+ content into the cellulose/BMIMCl
solution. The spent Amberlyst 15DRY possessed no catalytic ac-
tivity for depolymerization of cellulose. However, washing this
material with a H2SO4 solution regenerated its original catalytic
performance, as shown in Figure 12. The regenerated resin
(dried) was able to produce similar number of scissions in the
cellulosic chains per time as the original Amberlyst 15DRY.
Likewise, the yields of glucose obtained with the regenerated
resin (dried) and with Amberlyst 15DRY were very similar, 0.29
and 0.26% at 2 h, respectively.
Figure 10. Ion-exchange process involving BMIM+ and H+ species.
tion of acid remains constant indefinitely. At this stage, ca. 75–
80% of the initial acidic content of Amberlyst 15DRY is present
in the ionic liquid phase. The slow release of H3O+ species is in
ChemSusChem 2010, 3, 266 – 276
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
271

R. Rinaldi, F. Schꢀth et al.
Figure 12. Equivalent number of scissions produced by Amberlyst 15DRY
and by the regenerated resin after its first use. Reaction conditions: cellulose
(5 g, corresponding to ca. 31 mmol of AGU; C6H10O5: DPo 650 AGU) dissolved
in 100 g BMIMCl, Amberlyst 15DRY (1.15 mmol of H3O+), water (111 mmol),
1008C.
Mimicking the solid acidic resin Amberlyst 15DRY
The evidences for the homogeneous nature of the catalysis, in
spite of using a solid acid, indicate that the slow release of
H3O+ species governs the initial rate of the depolymerization
of cellulose using Amberlyst 15DRY. In order to simulate the
H3O+ release process, a reaction was performed by dosing an
HCl solution with a syringe pump over 1 h (2.3 mmol).
Figure 13 shows the profiles for the reactions performed using
Amberlyst 15DRY, dosing HCl over 1 h, and with p-TSA. The de-
polymerization of cellulose and the formation of glucose oc-
curred very rapidly when adding a soluble acid at once (p-TSA,
2.3 mmol) into the reaction medium, as can be seen in
Figure 13. In contrast with this result, the initial profile for the
reaction fed with an HCl solution over 1 h was identical to that
obtained using Amberlyst 15DRY, as H3O+-source (Figure 13a
and b). This confirms the hypothesis that the slow release of
H3O+ is the special feature of the Amberlyst resin for the con-
trol of the depolymerization of cellulose in BMIMCl. At 0.75 h,
however, the reaction fed with HCl started to become faster
than the reaction using Amberlyst 15DRY. This is due to the
fact that Amberlyst 15DRY can release only ca. 75–80% of its
nominal acidic content (Figure 11). Thus, the amount of acid
present in the reaction mixture that was fed continuously with
HCl over 1 h, is higher than that provided by Amberlyst 15DRY.
Consequently, from 0.75 h on, a higher yield of glucose and a
larger number of scissions were observed for the HCl-catalyzed
reaction compared with those values found for Amberlyst 15
DRY (Figure 13a and b).
Figure 13. Profiles for the reactions performed using Amberlyst 15DRY,
dosing HCl over 1 h and with p-TSA. a) Yield of glucose. b) Equivalent
number of scissions. c) Correlation between yield of glucose and equivalent
number of scissions. Reaction conditions: cellulose (5 g, corresponding to
ca. 31 mmol of AGU; C6H10O5: DPo 650 AGU) dissolved in 100 g BMIMCl, Am-
berlyst 15DRY or p-TSA (2.3 mmol of H3O+), water (111 mmol), 1008C. In the
reaction performed with HCl, a 4.6 molLꢀ1 HCl solution was dosed by a sy-
ringe pump over 1 h (0.50 mL, 2.3 mmol of H3O+). The initial amount of
water added in this reaction was 83 mmol.
The profiles for the production of glucose against the
number of scissions are identical for the reactions performed
with p-TSA, Amberlyst 15DRY and HCl (Figure 13c). In fact, in
the mentioned reactions, the catalytic species is H3O+ present
in the ionic liquid phase. Accordingly, the formation of glucose
depends solely on the number of scissions occurring in the cel-
lulosic chains in the depolymerization process.
Revisiting the mechanism of hydrolysis of cellulose
In the last sections, we have shown that the overall rate of de-
polymerization of cellulose depends both on the amount of
H3O+ present in the reaction medium and on the temperature.
The most striking result to emerge from that is the low turn-
over frequency (TOF) of acid-catalyzed depolymerization in
272
www.chemsuschem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 266 – 276

Depolymerization of Cellulose in Ionic Liquids
BMIMCl. We found
a
TOF value of 17 scis-
Both the low scission frequency
observed in the reaction per-
formed using Amberlyst 15DRY
and the reaction rate dependence
on the acid strength are thought-
provoking facts. Analyzing the
mechanism of hydrolysis of cellu-
lose, which was derived mostly
from the experience gained with
the hydrolysis of soluble glyco-
sides,[23] the activation of cellulose
is proposed to proceed through
protonation of the glycosidic
oxygen (Figure 15).[23b,c,e] Sequen-
sionshꢀ1 (mmolH3O+)ꢀ1 at 1008C using Amberlyst 15DRY as H+
source. Unfortunately, for several reasons it is not possible to
directly compare this value with other systems already studied.
Firstly, in the heterogeneous reactions, the hydrolytic process
occurs on the surface of cellulose. Therefore, the reaction rate
is governed by a series of other phenomena (such as adsorp-
tion, swelling of the polymeric structure, diffusion of the reac-
tants, and so on) rather than by the chemical reactivity of cel-
lulose.[3c,5] Secondly, in the homogeneous systems, such as in
those cases in which cellulose is solubilized in mineral
acids,[6d,21] the amount of H3O+ is much too high to establish a
proper comparison with the homogeneous systems based on
cellulose solubilized in imidazolium-ionic liquids, in which cata-
lytic amounts of H3O+ are used. For example, cellulose and
hemicellulose easily hydrolyze to a mixture of water-soluble
oligosaccharides and monosaccharides at room temperature, if
fuming HCl is used as a reactive solvent.[5,6d] In contrast, cellu-
lose dissolved in BMIMCl depolymerizes sluggishly when using
catalytic amounts of H3O+ already at 808C (Figure 7).
tially,
a
cyclic carbocation is
formed by the slow unimolecular
scission of the glycosidic linka-
ge.[23c] The nucleophilic attack of
water on the cyclic carbocation re-
establishes the hydroxyl group at
the C(1) position of the anhydro-
glucose unit.[23]
Figure 15. Proposed mecha-
nism for hydrolysis of cellulo-
se.[23e] Hydrogen, hydroxyl,
and hydroxymethyl groups
are omitted for clarity.
Another interesting feature of the depolymerization of cellu-
lose is the effect of the acid strength on the reaction rate. The
extent of depolymerization of cellulose over time, using molec-
ular acids, with pKa values ranging from ꢀ3 to 14, is directly as-
sociated with the acid strength, as shown in Figure 14. Al-
Although the formation of the
cyclic carbocation is regarded to
be the decisive step of the mecha-
nism (Figure 15),[23] the protonation of the glycosidic oxygen
was already reported to be cumbersome in some electron-defi-
cient acetals.[24] In the case of cellulose, the O-sites can be
roughly classified, according to their basicity, into two groups:
the acetal O-sites and the hydroxyl O-sites. The acetal O-sites
are considerably less basic than the hydroxyl ones. For the
sake of comparison, monoprotonated formaldehyde acetals,
RꢀOꢀCH2ꢀOꢀR, where R is methyl-, ethyl-, and isopropyl, show
pKa values ꢀ4.57, ꢀ4.13, ꢀ3.70, respectively.[25] It is then ex-
pected that the pKa value of the protonated glycosidic O-site
should lie within this range. On the other hand, the protonated
hydroxyl O-sites of cellulose should have a pKa value lying be-
tween those values found for protonated methanol (pKa ꢀ2.4)
and for protonated isopropanol (pKa ꢀ3.2). Therefore, the hy-
droxyl O-sites are expected to be 10 to 100 times more basic
than the glycosidic O-site. In other words, the protonation of
the glycosidic oxygen—as proposed in the mechanism shown
in Figure 15—is favored at high H+ concentration, because the
protonation is more preferable at the hydroxyl groups than at
the glycosidic O-site. In addition, strong acids are required to
activate cellulose towards hydrolysis due to the weak basicity
of the glycosidic O-site. However, in presence of water, the
acidity of strong acids is leveled out to the acidity of H3O+
species (pKa ꢀ1.7). Regarding the acidity of H3O+ species (pKa
ꢀ1.7), the prevalence of the protonated glycosidic O-sites
(pKa ~ꢀ4) is around 1:200, while this prevalence is ca. 1:20 for
the protonated hydroxyl O-sites (pKa ~ꢀ3). Indeed, the low
prevalence of the protonated glycosidic O-sites explains the
observed low scission frequency of the cellulosic chains.
Figure 14. Influence of the acid strength on the number of equivalent scis-
sion occurring in cellulose. Acids tested (pKa values in parentheses): sulfuric
acid (ꢀ3.0), p-TSA (ꢀ3.0), CF3COOH (0.23), oxalic acid (1.23), BMIMHSO4
(1.99), H3PO4 (2.16), malic acid (3.40), benzoic acid (4.19), levulinic acid (4.59),
and water (14). Reaction conditions: cellulose (0.25 g, corresponding to ca.
1.6 mmol of AGU; C6H10O5: DPo 100 AGU) dissolved in 5 g BMIMCl, molecular
acid (0.23 mmol), water (5.5 mmol), 1008C for 2 h. The pKa values were taken
from the CRC Handbook of Chemistry and Physics (79th Ed.).
though acidity in ionic liquids is a concept still in its infancy, it
is already known that the ranking of acid strength of several
organic and inorganic acids—in ionic liquids—resembles gen-
erally that observed in water.[22] Our results show that cellulose
is cleaved extensively in the presence of molecular acids with
pKa <1. However, weaker acids display much lower catalytic ac-
tivity. A comparable trend for the hydrolysis of cellobiose in
EMIMCl was recently reported.[13d]
A serious consequence of the requirement of a strong acid
for the catalytic hydrolysis of cellulose is the restriction on the
number of ionic liquid suitable as solvent for this reaction. Al-
ChemSusChem 2010, 3, 266 – 276
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
273

R. Rinaldi, F. Schꢀth et al.
kylimidazolium ionic liquids containing weakly basic anions -
such as acetate or organophosphates - are great solvents for
cellulose.[26] However, we observed that these solvents are not
appropriate for acid-catalyzed hydrolytic processes. Actually,
acetate or phosphonic anions are protonated easily by strong
acids, resulting in acetic acid (pKa 4.78) or phosphonic acids
(methylphosphonic acid, pKa1 2.38).[27] Consequently, cellulose
cannot be sufficiently protonated to result in its hydrolysis at
reasonable rates in these ionic liquids.
Recycling of ionic liquid
The main concern for ionic-liquid based processes is the reutili-
zation of the solvent because of its high cost. In the depolyme-
rization of cellulose, the cellooligomers can be removed from
the reaction medium by addition of water. Nevertheless, small
sugars, dehydration products, and H3O+/H2O remain solubi-
lized. Therefore, these compounds should be removed prior to
reutilization of the ionic liquid. The acid content was success-
fully neutralized by passing the aqueous solution of BMIMCl
through a neutral alumina column. The neutralization is be-
lieved to occur by the exchange of the labile surface hydroxyl
groups by Clꢀ, forming AlꢀCl surface groups. The liberation of
OHꢀ neutralizes the acidic content of the ionic liquid. Starting
from an initial acid content of ca. 45 mmolH3O+ gꢀ1, the acid
content of the recycled ionic liquid was 3.5 mmolH3O+ gꢀ1.
The content of water in BMIMCl was reduced via vacuum distil-
lation. It is important to point out that the recycled ionic liquid
does not need to be anhydrous to be reused in the process. In
fact, some water is actually required as a reactant for the hy-
drolysis of the glycosidic bond. Cellulose can be solubilized in
BMIMCl even in the presence of 2 wt% water. The recycling of
the ionic liquid was possible without damage to the catalytic
performance. Figure 17 shows the performance of the recycled
Impurities of the ionic liquid
Residues of N-methylimidazole are commonly found as the
main impurity of 1-alkyl-3-methylimidazolium ionic liquids.[28]
To study systematically the effect of this basic impurity on the
depolymerization of cellulose, N-methylimidazole (3.3 mmol)
was added in a reaction performed using Amberlyst 15DRY
(6.7 mmol of H3O+). In a reference reaction carried out with
Amberlyst 15DRY (6.9 mmol of H3O+), no N-methylimidazole
was added. Comparing the results of both reactions
(Figure 16), a decrease in the depolymerization rate from 162
Figure 16. Effect of methylimidazole (added as impurity) and methylimidazo-
lium chloride (used as catalyst) in the depolymerization of cellulose. Reaction
conditions: (reference) cellulose (5 g, corresponding to ca. 31 mmol of AGU,
C6H10O5) dissolved in 100 g BMIMCl, Amberlyst 15DRY (6.9 mmol of H3O+),
water (111 mmol), 1008C; (poisoning with methylimidazole) cellulose (5 g,
corresponding to ca. 31 mmol of AGU, C6H10O5) dissolved in 100 g BMIMCl,
Amberlyst 15DRY (6.7 mmol of H3O+), methylimidazole (3.3 mmol), water
(111 mmol), 1008C; (HMIM as catalyst) cellulose (5 g, corresponding to ca.
31 mmol of AGU, C6H10O5) dissolved in 100 g BMIMCl, HMIMCl (4.6 mmol of
H3O+), water (111 mmol), 1008C.
Figure 17. Recycling of BMIMCl. Reaction conditions: cellulose (5 g, corre-
sponding to ca. 31 mmol of AGU, C6H10O5) dissolved in 100 g BMIMCl, Am-
berlyst 15DRY (4.6 mmol H3O+), water (111 mmol), 1008C.
ionic liquid compared with the fresh ionic liquid. Due to the
small scale of the process, however, a recovery ratio could not
be precisely determined. In our experiments, the recovery ratio
could be estimated to approximately 91%, but this value
could be vastly improved for a larger scale implementation.
The treatment steps described above, however, are not effi-
cient for removing the sugars and other small molecules. The
separation of sugars from the ionic liquid is still the most chal-
lenging issue in the recycling of BMIMCl. Removal of glucose,
cellobiose, and other small sugars from the ionic liquid solu-
tion was attempted by adsorption on alumina or charcoal, but
these were not very effective. Therefore, small sugars and de-
hydration products would be accumulated over successive
cycles of re-utilization of the ionic liquids. This can only be
to 75 scissionshꢀ1 (calculated from the slope at steady state)
was observed in the reaction poisoned with N-methylimidazole
(Figure 16). This result is consistent with the fact that only
3.4 mmol of H3O+ effectively participates in the activation of
cellulose in the reaction poisoned with N-methylimidazole.
Therefore, in the acid-catalyzed depolymerization of cellulose,
the presence of residual N-methylimidazole is an issue because
H+ species are readily neutralized by this organic base. The re-
sulting N-methylimidazolium (HMIMCl, pKa 7.2)[29] cannot acti-
vate cellulose towards hydrolysis as can also be seen in
Figure 16.
274
www.chemsuschem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 266 – 276

Depolymerization of Cellulose in Ionic Liquids
avoided by stopping the reaction at the cellooligomer stage.
However, whether the build-up of impurities would then be a
problem after hundreds of regeneration steps remains to be
seen.
change process, resulting in H+-release, will proceed to some
extent, although differences may be observed, depending on
the pore size of the catalyst and its acid strength and acid site
concentration.
Conclusions
Experimental Section
The factors responsible for the control of depolymerization of
cellulose in ionic liquid using Amberlyst 15DRY as H3O+ source
were determined in detail. Based on these findings the follow-
ing conclusions can be drawn for the Amberlyst 15DRY-cata-
lyzed depolymerization of cellulose:
Materials
a-Cellulose (fibers, Aldrich) and 1-butyl-3-methylimidazolium chlo-
ride (ꢂ99.9%, Iolitec) were used without prior treatment. Amber-
lyst 15DRY (Aldrich) and p-toluenesulfonic acid monohydrate (ꢂ
98.5%, Fluka) were used as received.
(1) Although the amount of catalyst positively affects the de-
polymerization rate, this reaction parameter, surprisingly, has
only a small effect on the distribution of products at a given
reaction extent.
Depolymerization of cellulose
Typically, the solution of cellulose in 1-butyl-3-methylimidazolium
chloride (BMIMCl) was prepared by dissolving 5 g of a-cellulose
[31 mmol, calculated as anhydroglucose unit (AGU) C6H10O5] in
100 g of BMIMCl at 1008C. The mixture was stirred by a mechanical
overhead stirrer. To the cellulose/BMIMCl solution, 2.00 mL of dis-
tilled water (111 mmol) were added under mechanical stirring. The
mixture was stirred for 15 min. Subsequently, the desired amount
(2) Depolymerization of cellulose is a first-order reaction with
respect of the catalyst concentration.
(3) The reaction temperature seriously affects the rates of
the reactions. The activation energy for the depolymerization
of cellulose is 108 kJmolꢀ1. The selectivity at identical conver-
sion of glycosidic bonds is not strongly affected by changes in
reaction temperature, because of the similar activation ener-
gies for hydrolysis of the glycosidic bonds and for further de-
hydration of glucose formed.
of Amberlyst 15DRY (4.6 mmolH3O+ ꢀ1) was added to the mixture
g
containing cellulose. The depolymerization of cellulose was carried
out at 1008C for 3 h. The aliquots were taken at the reaction times
indicated. Cellulose from the aliquots was precipitated by addition
of water. The cellulose was separated from the aqueous solution
by centrifugation. The aqueous solutions were collected and
stored for later HPLC analysis.
(4) The depolymerization of cellulose catalyzed by Amberlyst
15DRY happens in solution. The acidic resin releases H+ into
the solution, controlling the initial rate of depolymerization.
The preferential cleavage of large polymeric molecules occurs
because of statistical reasons. A soluble catalyst is not able to
change this preference. In other words, the initial size of the
cellulose chains is crucial in the control of initial product distri-
bution. Large chains are preferably cleaved into smaller ones
instead of producing glucose, which explains the induction
period observed for the release of glucose or total reducing
sugars.
HPLC analysis of sugars and dehydration products
The chromatographic separations were carried out using a Nucleo-
gel OA HY (300ꢂ7.8 mm, 10 mm) column and an aqueous solution
of sulfuric acid (5 mmolLꢀ1) as mobile phase. The samples were an-
alyzed in isocratic mode at a flow rate of 0.5 mLminꢀ1 at 808C,
using a Perkin–Elmer HPLC 200. Glucose and cellobiose were ana-
lyzed using a refractive index detector. 5-Hydroxymethylfurfural (5-
HMF) was analyzed using a diode-array UV/Vis detector at 254 nm.
The mentioned components were identified by comparison of re-
tention times with those of the corresponding reference com-
pounds. The amounts of products were quantified using external
calibration curves.
(5) A high degree of control of the depolymerization reac-
tion can be achieved also in the case of molecular acids, if
they are added slowly over time to the reaction batch, thus
mimicking the slow ion exchange process occurring over the
solid acids.
(6) Activation of cellulose towards hydrolysis requires a
strong acid. This prohibits the utilization of acetate- or phos-
phonate-based ionic liquids - or any other kind of ionic liquid
composed of a weakly basic anion. These anions capture es-
sentially the available H3O+ species, which prevents the activa-
tion of the glycosidic bonds. Additionally, the presence of N-
methylimidazole, often found at different concentration levels
as impurity in BMIMCl, decreases proportionally the catalytic
performance of this system.
Determination of the apparent degree of polymerization
An aliquot of the reaction mixture (0.5 g) was suspended in DMSO
(5 mL, Fluka). BMIMCl does not interfere in the carbanilation of cel-
lulose.[30] Into the suspension of cellulose in DMSO was added
1.0 mL of phenylisocyanate (>99%, Aldrich). The reaction was car-
ried out for 4–5 h at 808C, resulting in cellulose tricarbanilates
(CTC), which are soluble in DMSO. The cellulose derivative was
then worked up as described in Ref. [31]. The isolated CTC was dis-
solved in THF (2 mgmLꢀ1) and analyzed by gel permeation chro-
matography. The analyses were performed at 508C using a Perkin–
Elmer HPLC 200 using mix-bed GPC columns (2 columns, TSKgel
SuperHZM-M, 4.6 mm IDꢂ15.0 cm) and stabilized THF as eluent
(0.2 mLminꢀ1, LiChrosolv, Merck). For detection, a UV/Vis detector
at 236 nm was used. The system was calibrated with polystyrene
standards (5ꢂ102 to 7ꢂ106 Da, Aldrich). The apparent number-
average and weight-average degree of polymerization, DPn and
(7) Both the ionic liquid and Amberlyst 15DRY can be recy-
cled, the ionic liquid by neutralization with alumina and
vacuum distillation; the acidic resin by washing with sulfuric
acid.
In principle, the homogeneous nature of the reaction with
Amberlyst 15DRY, as solid acid, can be extended to any other
solid Brønsted acid catalyst. For any such catalyst, the ion-ex-
ChemSusChem 2010, 3, 266 – 276
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
275

R. Rinaldi, F. Schꢀth et al.
er, T. H. Wegner, J. I. Zerbe, Two-Stage, Dilute Sulfuric Acid Hydrolysis of
Wood: An Investigation of Fundamentals, Gen. Tech. Rep. FPl-45, US De-
partment of Agriculture, Forest Service, Forest Products Laboratory,
Madison, 1985, p. 73i) R. Wooley, M. Ruth, J. Sheehan, K. Ibsen, Lignocel-
lulosic Biomass to Ethanol Process Design and Economics Utilizing Co-Cur-
rent Dilute Acid Prehydrolysis and Enzymatic Hydrolysis Current and Futur-
istic Scenarios, NREL/TP-580–26157, US Department of Energy, National
Renewable Energy Laboratory, Golden, 1999, p.130; j) R. Erckel, R.
Franz, R. Woernle, T. Riehm, US Patent 4,556,431.
DPw, respectively, were calculated by dividing the respective aver-
age molecular weight values by molar mass of the anhydroglucose
tricarbanilate (519 gmolꢀ1). The equivalent number of scissions
was calculated by using Equation (1):
DP0
DP
ð1Þ
s ¼
ꢀ 1
where DPo and DP are the weight-averaged degree of polymeri-
zation at t=0 and at a given time, respectively.
[7] H. Zhao, C. L. Jones, G. A. Baker, S. Xia, O. Olubajo, V. N. Person, J. Bio-
technol. 2009, 139, 47–54.
[8] a) Y. Sun, J. Cheng, Bioresour. Technol. 2002, 83, 1–11; b) P. Kumar, D. M.
Barrett, M. J. Delwiche, P. Stroeve, Ind. Eng. Chem. Res. 2009, 48, 3713–
3729; c) X. B. Zhao, K. K. Cheng, D. H. Liu, Appl. Microbiol. Biotechnol.
2009, 82, 815–827.
[9] a) R. P. Swatloski, S. K. Spear, J. D. Holbrey, R. D. Rogers, J. Am. Chem.
Soc. 2002, 124, 4974–4975; b) R. C. Remsing, R. P. Swatloski, R. D.
Rogers, G. Moyna, Chem. Commun. 2006, 1271–1273; c) R. C. Remsing,
G. Hernandez, R. P. Swatloski, W. W. Massefski, R. D. Rogers, G. Moyna, J.
Phys. Chem. B 2008, 112, 11071–11078.
Titrations
The acid concentrations given in the text were determined via ti-
tration with a 0.01 molLꢀ1 NaOH solution using a Metrohm Titrino
Plus 848 automated titrator.
Acknowledgements
[10] O. A. El Seoud, A. Koschella, L. C. Fidale, S. Dorn, T. Heinze, Biomacromol-
ecules 2007, 8, 2629–2647.
[11] R. Rinaldi, R. Palkovits, F. Schꢀth, Angew. Chem. 2008, 120, 8167–8170;
Angew. Chem. Int. Ed. 2008, 47, 8047–8050.
We acknowledge the financial support from the Max Planck Soci-
ety. This work was performed as part of the Cluster of Excellence
“Tailor-Made Fuels from Biomass”, which is funded by the Excel-
lence Initiative by the German federal and state governments to
promote science and research at German universities. R.R. and
F.S. thank Prof. Christopher W. Jones (Georgia Institute of Tech-
nology) for his valuable comments on the nonheterogeneity of
sulfonated resins in ionic liquids.
[12] G. Gelbard, Ind. Eng. Chem. Res. 2005, 44, 8468–8498.
[13] a) C. Li, Z. K. Zhao, Adv. Synth. Catal. 2007, 349, 1847–1850; b) C. Li, Q.
Wanga, Z. K. Zhao, Green Chem. 2008, 10, 177–182; c) C. Sievers, M. B.
Valenzuela-Olarte, T. Marzialetti, D. Musin, P. K. Agrawal, C. W. Jones, Ind.
Eng. Chem. Res. 2009, 48, 1277–1286; d) L. Vanoye, M. Fanselow, J. D.
Holbrey, M. P. Atkins, K. R. Seddon, Green Chem. 2009, 11, 390–396.
[14] Q. Liu, M. H. A. Janssen, F. Rantwijk, R. A. Sheldon, Green Chem. 2005, 7,
39–42.
[15] C. Sievers, I. Musin, T. Marzialetti, M. B. V. Olarte, P. K. Agrawal, C. W.
Jones, ChemSusChem 2009, 2, 665–671.
[16] K. Lourvanij, G. L. Rorrer, J. Chem. Technol. Biotechnol. 1997, 69, 35–44.
[17] L. P. Ramos, Quim. Nova 2003, 26, 863–871.
Keywords: catalysis · cellulose · hydrolysis · ionic liquids ·
polymerization
[18] J. A. Rice, Soil Sci. 2001, 166, 848–857.
[19] E. E. Johansson, J. Lind, Polym. Degrad. Stab. 2005, 88, 159–162.
[20] K. Freudenberg, G. Blomqvist, Ber. Dtsch. Chem. Ges. A 1935, 68, 2070–
2082.
[21] B. Kamm, M. Kamm, M. Schmidt, I. Starke, E. Kleinpeter, Chemosphere
2006, 62, 97–105.
[1] a) A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411–2502; b) Y.-
H. P. Zhang, J. Ind. Microbiol. Biotechnol. 2008, 35, 367–375; c) Y. Y.
Romꢃn-Leshkov, C. J. Barrett, Z. Y. Liu, J. A. Dumesic, Nature 2007, 447,
982–985; d) G. W. Huber, J. N. Chheda, C. J. Barrett, J. A. Dumesic, Sci-
ence 2005, 308, 1446–1450; e) Y. Romꢃn-Leshkov, J. N. Chheda, J. A. Du-
mesic, Science 2006, 312, 1933–1937.
[2] a) A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Cairney,
C. A. Eckert, W. J. Frederick, J. P. Hallett, D. J. Leak, C. L. Liotta, J. R. Mie-
lenz, R. Murphy, R. Templer, T. Tschaplinski, Science 2006, 311, 484–489;
b) D. Klemm, B. Heublein, H.-P. Fink, A. Bohn, Angew. Chem. 2005, 117,
3422–3458; Angew. Chem. Int. Ed. 2005, 44, 3358–3393.
[3] a) I. S. Goldstein, Organic Chemicals from Biomass, CRC Press, Boca
Raton, 1981; b) D. Fengel, G. Wegener, Wood: Chemistry, Ultrastructure,
Reactions, Walter De Gruyter, Berlin, 1984; c) H. A. Krꢄssig, Cellulose:
Structure Accessibility and Reactivity Polymer Monographs Vol. 11,
Gordon and Breach Science Publishers, Yverdon, 1992.
[22] D. R. MacFarlane, J. M. Pringle, K. M. Johansson, S. A. Forsyth, M. Forsyth,
Chem. Commun. 2006, 1905–1917.
[23] a) J. T. Edward, Chem. Ind. 1955, 1102–1104; b) T. E. Timell, Can. J. Chem.
1964, 42, 1456–1472; c) J. N. BeMiller, Adv. Carbohydr. Chem. 1967, 22,
25–108; d) B. Capon, Chem. Rev. 1969, 69, 407–498; e) T. P. Nevell, W. R.
Upton, Carbohydr. Res. 1976, 49, 163–174; f) H. H. Jensen, M. Bols, Org.
Lett. 2003, 5, 3419–3421.
[24] A. Kankaanperꢄ, M. Lahti, Acta Chem. Scand. 1969, 23, 2465–2474.
[25] A. Kankaanperꢄ, Acta Chem. Scand. 1969, 23, 1723–1727.
[26] Y. Fukaya, K. Hayashi, M. Wada, H. Ohno, Green Chem. 2008, 10, 44–46.
[27] L. D. Quin, A Guide to Organophosphorous Chemistry, John Wiley & Sons
Inc, New York, 2000.
[4] R. Rinaldi, F. Schꢀth, Energy Environ. Sci. 2009, 2, 610–626.
[28] A. Stark, P. Behrend, O. Braun, A. Mꢀller, J. Ranke, B. Ondruschka, B. Jas-
torff, Green Chem. 2008, 10, 1152–1161.
[29] A. Albert, Transition Met. Chem. 1983, 8, 56–58.
[30] S. Barthel, T. Heinze, Green Chem. 2006, 8, 301–306.
[31] R. Evans, R. H. Wearne, A. F. A. Wallis, J. Appl. Polym. Sci. 1989, 37, 3291–
3303.
[5] R. Rinaldi, F. Schꢀth, ChemSusChem 2009, 2, 1096–1107.
[6] a) P. L. Ragg, P. R. Fields, Philos. Trans. R. Soc. London, Ser. A 1987, 321,
537–547; b) M. R. Ladisch, Proc. Biochem. Soc. 1979, 21–25; c) W. L.
Faith, Ind. Eng. Chem. 1945, 37, 9–11; d) F. Bergius, Ind. Eng. Chem.
1937, 29, 247–253; e) E. E. Harris, E. Beglinger, The Madison Wood-Sugar
Process, Report No. R1617, US Department of Agriculture, Forest Serv-
ice, Forest Products Laboratory, Madison, 1946; f) R. A. Antonoplis, H. W.
Blanch, R. P. Freitas, A. F. Sciamanna, C. R. Wilke, Biotechnol. Bioeng.
1983, 25, 2757–2773; g) D. R. Thompson, H. E. Grethlein, Ind. Eng.
Chem. Prod. Res. Dev. 1979, 18, 166–169; h) J. F. Harris, A. J. Baker, A. H.
Conner, T. W. Jeffries, J. L. Minor, R. C. Pettersen, R. W. Scott, E. L. Spring-
Received: December 3, 2009
Revised: January 22, 2010
276
www.chemsuschem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 266 – 276