Production of solubilized carbohydrate from cellulose using non-catalytic, supercritical depolymerization in polar aprotic solvents

Article abstract of DOI:10.1039/c5gc02071a

We report yields of solubilized and depolymerized carbohydrate from solvent processing of cellulose as high as 94% without use of catalysts. Cellulose was converted using a variety of polar aprotic solvents at supercritical conditions, including 1,4-dioxane, ethyl acetate, tetrahydrofuran, methyl iso-butyl ketone, acetone, acetonitrile, and gamma-valerolactone. Maximum yield of solubilized products from cellulose, defined as both depolymerized carbohydrate and products of carbohydrate dehydration, was 72 to 98% at 350 °C for reaction times of 8-16 min. In all cases solvents were recovered with high efficiency. Levoglucosan was the most prevalent solubilized carbohydrate product with yields reaching 41% and 34% in acetonitrile and gamma-valerolactone, respectively. Levoglucosan yields increased with increasing polar solubility parameter, corresponding to decreasing activation energy for cellulose depolymerization.

Full text of DOI:10.1039/c5gc02071a

View Article Online
View Journal
Green
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: A. Ghosh, R. C.
Brown and X. Bai, Green Chem., 2015, DOI: 10.1039/C5GC02071A.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/greenchem

Page 1 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
Production of solubilized carbohydrate from cellulose using non-catalytic, supercritical
depolymerization in polar aprotic solvents
Arpa Ghosh1, Robert C. Brown2, 3, and Xianglan Bai2*––
1) Department of Chemical and Biological Engineering, Iowa State University, Ames, Iowa,
USA, 50011
2) Department of Mechanical Engineering, Iowa State University, Ames, Iowa, USA, 50011
3) Bioeconomy Institute, Iowa State University, Ames, Iowa, USA, 50011
*Corresponding author. Postal address: Iowa State University, 2070 Black Engineering Building,
Ames, IA 50011; Tel: +1 515 294 6886; Fax: +1 515 294 3091; E-mail address:
bxl9801@iastate.edu (Xianglan Bai)

Green Chemistry
Page 2 of 25
View Article Online
DOI: 10.1039/C5GC02071A
Abstract
We report yields of solubilized and depolymerized carbohydrate from solvent processing of
cellulose as high as 94% without use of catalysts. Cellulose was converted using a variety of
polar aprotic solvents at supercritical conditions, including 1,4-dioxane, ethyl acetate,
tetrahydrofuran, methyl iso-butyl ketone, acetone, acetonitrile, and gamma-valerolactone.
Maximum yield of solubilized products from cellulose, defined as both depolymerized
carbohydrate and products of carbohydrate dehydration, was 72 to 98% at 350 oC for reaction
times of 8-16 min. In all cases solvents were recovered with high efficiency. Levoglucosan was
the most prevalent solubilized carbohydrate product with yields reaching 41% and 34% in
acetonitrile and gamma-valerolactone, respectively. Levoglucosan yields increased with
increasing polar solubility parameter, corresponding to decreasing activation energy for cellulose
depolymerization.
2

Page 3 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
Introduction
Development of alternative technologies based on renewable energy resources has proliferated in
recent times to displace fossil fuels and petroleum-based chemicals. Lignocellulosic biomass has
drawn attention as a sustainable feedstock for producing biobased renewable fuels and
biochemicals.1 Cellulose, the most abundant component of lignocellulosic biomass, is a good
source of fermentable sugars.2 Currently, the most prominent pathway to depolymerize cellulose
to sugars is enzymatic hydrolysis. Although this biochemical route has high selectivity for final
products, the process has several drawbacks such as slow conversion rates, high cost of enzyme,
and end-product inhibitions.3, 4 Biomass depolymerization is slowed by the innate structural
recalcitrance of cellulose. Cellulose is a three-dimensional cross-linked biopolymer of D-glucose
units joined by -1,4 glycosidic bonds, which is intrinsically rigid.5 Additionally, the intensive
inter- and intra-molecular hydrogen bonding in cellulose gives rise to a rigid crystalline structure,
which is highly resistant to enzymatic hydrolysis.6
Alternatively, it is possible to produce fermentable sugars through thermochemical conversion
pathways, such as pyrolysis and solvent liquefaction (also known as solvolysis).7 If optimized,
these technologies could be rapid, robust and economical approaches to convert biomass into
solubilized carbohydrate at high yields. While glucose can be directly produced from cellulose in
the presence of water, thermal depolymerization of cellulose in the absence of water usually
produces levoglucosan (LG) as the main sugar monomer along with other anhydro
monosaccharides and oligosaccharides. These anhydro sugars can be hydrolyzed to glucose and
then fermented to ethanol or the anhydro monosaccharides can be directly fermented to ethanol.8,
9 Levoglucosan is an important precursor chemical in its own right, which can be used to
synthesize pharmaceuticals and biodegradable plastics.10
3

Green Chemistry
Page 4 of 25
View Article Online
DOI: 10.1039/C5GC02071A
Solvent liquefaction is the depolymerization of biomass in the presence of a solvent. Reaction
temperatures are normally in the range of 150 – 400 oC with system pressure elevated to keep the
solvent from boiling.11 Solvent liquefaction of cellulose can be very selective depending upon the
choice of the solvent and/or catalyst. Solvents can dissolve some of the indigenous mineral matter
in biomass that suppresses sugar yields during thermal depolymerization of lignocellulose.12
Moreover, unlike pyrolysis, solvent liquefaction is able to recover non-volatile sugars (i.e.,
solubilized poly- or oligosaccharides) as products. The dilution of sugar products in the solvent
can also potentially reduce secondary reactions that decompose sugars. Solvent phase conversion
makes processing of wet biomass possible, thus eliminating energy intensive drying of raw
feedstock.
Hydrolysis with highly concentrated acid could facilitate depolymerization of cellulose, but
scaling up is challenging due to associated corrosive effects, handling hazards and complexity of
acid recovery.13, 14 Ionic liquids combined with homogenous or heterogeneous catalysts have also
been explored because their exquisite solvation properties promote faster hydrolysis of
16
cellulose.15,
Nevertheless, the progress of this technology has been hindered because
quantitative recovery and reuse (at least 98%) of these expensive solvents has not been solved.17
Conversion of cellulose to solubilized carbohydrates also has been studied using hot and
pressurized protic solvents.18-20 Combinations of processing in supercritical and subcritical water
has been employed to produce a hydrolyzed product.21 Nevertheless, it has been reported that the
presence of water may lower the activity of acid catalyst and also favor formation of undesired
degradation products compared to other polar solvents.22-24 The use of methanol leads to the
formation of methylated oligosaccharides.25
Solvolysis of biomass in polar aprotic solvents results in more desirable product distributions. For
example, up to 38% of LG was produced when cellulose was treated with acetone,26 sulfolane27
or 1, 4-dioxane.28 However, no information about other products was reported in these studies.
4

Page 5 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
Luterbacher et al.29 recently demonstrated high yields of solubilized carbohydrate from
lignocellulosic biomass using gamma-valerolactone in the presence of acid catalyst. In their
o
study, biomass was converted in a continuous flow reactor at 157 – 217 C. Yield of solubilized
carbohydrate from cellulose was as high as 80%, but required times as long as 2 h because of the
low reaction temperatures. Only 90% of GVL was recovered after the reaction since some of the
solvent was consumed during the process.
We have recently produced soluble sugar monomers by depolymerizing acid pretreated
switchgrass in 1, 4-dioxane solution.12 Despite relatively high reaction temperatures (300 – 350
ºC), the sugars were stable. Over 98% of the 1, 4-dioxane solvent was recovered after reaction.
Yields of levoglucosan from cellulose were as high as 50% and oligosaccharides were not
quantified. Other studies also showed that acid catalyzed aprotic solvents could produce
dehydration products of cellulose such as 5-hydroxymethylfurfural, furfural and levoglucosenone
at high yields.30, 31
While previous studies using various polar aprotic solvents suggest solvent liquefaction as an
alternative to enzymatic hydrolysis for production of carbohydrates from lignocellulosic biomass,
they either employed acid catalysts, heterogeneous catalysts or very long reaction times to
achieve high yields of solubilized carbohydrates. Neither does the literature contain much
information on the role of solvent properties on cellulose conversion. The choice of solvent not
only determines product distributions and yields, but also influences the thermal stability of
products and the ease of recycling the solvent. In general, solvents with low boiling points and
high thermal and chemical stability are preferred.
The present work focuses on non-catalytic conversion of cellulose in polar aprotic solvents to
produce solubilized carbohydrates, which are defined as the sum of monosaccharide,
oligosaccharide and polysaccharide products that dissolve in the processing solvent. Although the
5

Green Chemistry
Page 6 of 25
View Article Online
DOI: 10.1039/C5GC02071A
absence of acid simplifies the process, reaction rates and product yields are expected to decrease,
thus requiring operation at higher temperatures and pressures to avoid long reaction times. In the
present study, cellulose was depolymerized in seven different polar aprotic solvents to understand
the effect of solvent properties on the yields and distribution of the cellulose depolymerization
products.
2. Experimental section
2.1 Materials
Microcrystalline cellulose with average particle size of 50 m was purchased from Sigma-
Aldrich. Levoglucosan (purity > 99.6%), cellobiosan (purity > 98.6%) were obtained from
Carbosynth, UK and 5-hydroxymethylfurfural (5-HMF, purity > 99%) and furfural (purity >
99%) from Sigma Aldrich. Glucose (purity > 99%) from Fisher Scientific, cellobiose (purity >
98%) from Acros Organics, and maltotriose, maltotetraose and maltohexaose were obtained from
MP Biomedicals. HPLC grade (submicron filtered) 1,4-dioxane, ethyl acetate, tetrahydrofuran
(THF), methyl iso-butyl ketone (MIBK), acetone, acetonitrile, and gamma-valerolactone (GVL)
were purchased from Fisher Scientific.
2.2 Solvent processing methodology
Solvolysis experiments were performed in mini-reactors from Swagelok (316 SS) assemblies
each consisting of one male connector (NPT type) and two sealed 3/8 inch caps. These reactors
had total capacity of 2.5 mL. Microcrystalline cellulose in the amount of 10-100 mg was placed
in a mini-reactor with 1.2 mL of aprotic solvent. The mini-reactors were tightly sealed and shaken
for 1 h prior to reaction. A fluidized sand bath (Techne Industrial Bed 51) was used as the heating
o
source. The heating bath was operated at 325 to 375 C and reaction times were up to 20 min.
Experiments were performed with 1,4-dioxane, ethyl acetate, tetrahydrofuran (THF), methyl iso-
6

Page 7 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
butyl ketone (MIBK), acetone, acetonitrile, and gamma-valerolactone (GVL) as solvent with each
experiment performed two or three times with average results reported. The standard deviation
was within 5% error from the mean for most of the solvents tested except for 1,4-dioxane, ethyl
acetate and acetone (error slightly over 10%) probably because of evaporation of solvent during
extraction process due to their high volatility or remaining solvent in solid residue. The heating
rate of the reactor was 7 ºC s-1 and the reaction time measured from the time the reactor entered
the preheated sand bath to the time it was removed and immersed in cold water. After one hour,
the cooled reactor was opened for extraction of the liquid and solids contents of the reactor at
room temperature. The liquid contents were transferred from the reactor using a pipette (Fisher
Scientific) and the solids were collected at the bottom. Liquid fraction contained both the original
solvent and solubilized products from cellulose depolymerization. This liquid fraction was
o
filtered using syringe-filters of pore size 0.45 μm. Solid residues were dried in an oven at 50 C
overnight and weighed. Gas production was determined by weighing a reactor before and after
venting non-condensable gases from the cooled reactor.
The following definition was used to calculate the yield of solubilized products:
Solubilized products yield (%) =(
)
(1)
2.3 Analytical methods
A Gas Chromatograph with Mass Spectrometer and Flame Ionization Detector (Agilent 7890B
GC-MS/FID) was used to analyze the liquid fraction. The products in liquid fraction were first
identified by MS and then quantified by FID. The gas chromatograph was equipped with two
identical Phenomenex ZB 1701 (60 m x 0.250 mm and 0.250 μm film thickness) capillary
columns for separation of the products. One of these columns was connected to the MS while the
other was connected to the FID. The injection port and FID back detector in the GC were held at
7

Green Chemistry
Page 8 of 25
View Article Online
DOI: 10.1039/C5GC02071A
o
250 and 300 C, respectively. Helium carrier gas flow was 1 mL min-1. Injection volume for
o
analysis was 1 μL. The oven temperature of GC was ramped from 40 (3 min hold time) to 240 C
o
(4 min hold time) at a heating rate of 3 C min-1. The instrument was quantitatively calibrated
with LG, 5-HMF and furfural. 1,6-anhydro-β-D-glucofuranose (AGF) was quantified using LG as
the standard.
The yield and selectivity of a particular solubilized product were calculated on carbon molar basis
as:
Yield (%) =
(2)
(3)
Selectivity (%) =
GC/MS non-detectable solubilized products, such as solubilized carbohydrates with degree of
polymerization (DP) higher than 1, were characterized by molecular weight using Gel Filtration
Chromatography (GFC). GFC analysis was conducted with a Dionex Ultimate 3000 series HPLC
system using water as the eluent. The organic liquid fraction from solvolysis was diluted to 90%
water content and then tested for GFC. Two columns of the type PL-aquagel-OH-20 5 μm were
o
connected in series at 25 C and DI water passed through as the mobile phase at 0.8 mL min-1.
Refractive index was the basis of detection. Polyethylene glycol (PEG) was used as a standard to
generate a calibration curve of the molecular weight distribution of the solubilized products
relative to PEG. Identification of some of the solubilized products was done by comparing
retention times of individual standards of LG, glucose, cellobiosan and cellobiose. To obtain
more accurate molecular weight distributions and DP of the products, a calibration system was
developed between the known molecular weights of the carbohydrates LG, cellobiosan,
maltotriose, maltotetraose and maltohexaose and their relative molecular weights determined by
PEG standard in GFC. Although malto-oligosaccharides from cellulose are not expected as
8

Page 9 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
products, these compounds show similar DP to their analogous cello-oligosaccharides, thus
justifying this molecular weight estimation for carbohydrates produced from cellulose. The DP
for the oligo- or polysaccharides was then determined by dividing this molecular weight by 162.
The yield of solubilized carbohydrates was calculated according to:
YSC = YSP - YF
(4)
where YSC, YSP, and YSF are the yields of solubilized carbohydrates, solubilized products, and
furans (5-HMF and furfural), respectively.
3. Results and Discussion
3.1. Effect of different polar aprotic solvents on cellulose depolymerization
3.1.1 Solubilized carbohydrate production and conversion efficiency
The evolution of solubilized monomeric products from cellulose is presented in Figure 1 for
o
different solvents at 350 C. For all solvents LG was the major solubilized monomeric product,
achieving maximum yield within 8-16 min of reaction. The highest LG yield of 38% was
achieved in acetonitrile followed by 34% in GVL. The lowest yield was 15% in 1,4-dioxane. For
all solvents, the maximum yield of 5-HMF, the second most dominant solubilized monomeric
product, reached maximum yield at nearly the same time as LG. The maximum yield of 5-HMF
occurred in THF (9%) while the minimum yield occurred in acetonitrile (1%). Other solubilized
co-products, furfural and 1,6-anhydro-β-D-glucofuranose (AGF), showed a combined yield of
less than 3% in all the solvents.
Table 1 shows that all the polar aprotic solvents produced relatively high yields of solubilized
products from cellulose in relatively short times. Overall, the yields of solubilized products at 350
ºC at the reaction time that achieved maximum LG yield were in the range of 72-98% depending
9

Green Chemistry
Page 10 of 25
View Article Online
DOI: 10.1039/C5GC02071A
upon the solvent. Although treating cellulose in MIBK produced the highest solubilized products
yield (98%), the same process in THF, acetone, GVL and acetonitrile also achieved impressive
yields of solubilized products, exceeding 90%. Even 1, 4-dioxane and ethyl acetate, which had
the lowest optimum LG yields, resulted in yields of solubilized products above 72%. The order
of solubilized products yield in various solvents from low to high was 1,4-dioxane
THF <
»
acetone < Ethyl acetate < GVL
acetonitrile < MIBK. Conversion rate of cellulose was
»
estimated as the initial rate at which the unreacted cellulose decomposed. The order of conversion
rate of cellulose from low to high was 1,4-dioxane < THF < acetone < ethyl acetate < acetonitrile
< GVL < MIBK and ranged from 2.07 to 3.53 mol min-1 (Figure S4).
The total mass balance of cellulose conversion at the condition of maximum LG yield is
summarized in Figure 2 where solubilized products include solubilized carbohydrates (LG, AGF
and solubilized carbohydrates with DP > 1) and other dehydration products (5-HMF and
furfural). The solubilized products contained significant amounts of solubilized carbohydrates
with DP > 1, confirmed by GFC analysis, in addition to monomeric carbohydrates (the analysis
will be discussed in detail in a later section). As shown in Table 2, solubilized carbohydrate yield
from processing in acetonitrile and GVL reached 94 and 86%, respectively, compared to 63% for
1,4-dioxane. Furthermore, the selectivity of LG among the solubilized products was found to be
in the range of 16-40%, depending upon the solvent.
Stability of LG in different solvents is also compared in Table 2 using a degradation rate, defined
as the slope of the LG yield curve with time after the point of its maximum yield. The order of
decreasing LG stability is: 1,4-dioxane MIBK > THF > acetone acetonitrile > ethyl acetate >
>
>
GVL. Although GVL and acetonitrile had comparable high yield of LG, GVL had degradation
rate that was three times higher than acetonitrile.
3.1.2 Solvent recovery
10

Page 11 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
The recovery of solvent was quantified for acetonitrile, GVL and THF. For processing at 350 ºC
solvent recovery was very high: 97% for acetonitrile, 98% for THF and 99% for GVL. Except for
ethyl acetate and MIBK, other solvents also had high solvent recovery, as estimated by GC/MS
peak areas for solvent decomposition products.
3.1.3 Effect of reaction conditions
The influence of reaction conditions was further evaluated for THF and acetonitrile as solvents.
Acetonitrile and THF were chosen because they produced the highest LG and 5-HMF yields,
respectively. Both of the solvents have low boiling points, easing their recovery. As shown in
Figure 3 (a) and (b), higher temperatures facilitated the rate of reaction as evidenced by the
shorter time to reach maximum LG yield. Increasing temperature from 325 to 350 oC
significantly enhanced the maximum LG yield for both acetonitrile and THF, but no further
increase in LG yield was seen for temperature above 350 oC. The yield of 5-HMF in THF at 350
o
oC was marginally higher than at 325 C. Further increases in temperature significantly reduced
the yield of 5-HMF (Figure 3 (c)).
The effect of mass loading on the yields of major monomeric products was also studied for
acetonitrile and THF. The optimal mass loadings were strongly dependent on the choice of
solvent as well as the reaction products studied. As shown in Figure 4 (a), maximum LG yield of
o
41% was obtained at 10 mg mass loading of cellulose in acetonitrile at 350 C. An increase in
mass loading of cellulose from 20 mg to 50 mg caused a reduction in LG yield from 38 to 12%.
On the other hand, the optimum mass loading in THF to achieve maximum LG yield (24%) at
o
350 C was 50 mg cellulose (Figure 4 (b)). The increase in LG yield was accompanied by a
decrease in 5-HMF yield from 9 to 7% for THF (Fig. 4 (c)).
11

Green Chemistry
Page 12 of 25
View Article Online
DOI: 10.1039/C5GC02071A
It was further noted that optimal mass loading of cellulose varied by product. For example, the
o
maximum LG yield for solvolysis in THF at 350 C occurred at mass loading of 20 mg while
maximum 5-HMF yield occurred at mass loading of 50 mg. Increasing the mass loading of
cellulose up to 100 mg was detrimental to the production of both LG and 5-HMF possibly due to
mass transfer limitation in the conversion process of cellulose.
3.2 Understanding cellulose depolymerization in polar aprotic solvents
3.2.1 Role of solubility parameter of the solvents on carbohydrate yields and selectivity
In order to maximize the production of solubilized carbohydrates, it is important to identify
solvent properties that play a role in enhancing the depolymerization of cellulose. High solubility
or degree of interaction of cellulose with a solvent may contribute to faster reaction and higher
degree of depolymerization.32 The solubility parameter of a solvent compared to that of cellulose
is potentially a good indicator of the degree of interaction between the two.33-35 However, for the
solvents evaluated in this study, the solubility parameters at ambient conditions were in the range
of 17-26.3 MPa1/2, significantly lower than the solubility parameter for cellulose (39.3 MPa1/2)
(see in Table S3).33 Despite these differences, the solvents showed excellent conversion of
o
cellulose to solubilized products, ranging from 72 to 98% at 350 C. At all reaction conditions
tested, the aprotic polar solvents were operated above their critical points with the possible
exception of GVL, for which critical point data is not available in the literature (see Table S1 and
S2 in the supplemental data for details). The solubility parameter, which is a function of
temperature and pressure, was estimated for the reaction conditions based on thermophysical
properties of the solvents described in the supplementary material (Table S4). At reaction
conditions, the solubility parameters for all solvents investigated were estimated to be in the
range of 25.7-33.8 MPa1/2, which approaches the solubility parameter of cellulose (Table S5).
12

Page 13 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
This similarity is likely responsible for the high level of cellulose conversion to solubilized
products even in the absence of catalyst.
The three major components of solubility parameter are the dispersive solubility parameter, the
polar solubility parameter and the hydrogen bonding solubility parameter.36 Of these, only the
polar solubility parameter, , exhibited a wide variability among the solvents tested (Table S5).
As shown in Figure 5 (a) and (b), the maximum yield of LG and its selectivity among solubilized
products increased linearly with increasing solvent . Although the maximum yields of 5-HMF
did not correlate with the polar interaction parameter of the solvents, the ratio of 5-HMF to LG
maximum yields decreased linearly with solvent polar solubility parameter (Figure 5 (c)),
indicating that polarity of the solvent indirectly influences the selectivity of 5-HMF. These
observations suggest that could be an important parameter in solvent selection.
3.2.2 Activation energy for cellulose solvolysis in different solvents
The activation energies of cellulose decomposition were calculated and compared for several of
the solvents (see Table S6, Figure S1, S2, S3). The activation energies for solvolysis in GVL,
acetonitrile, and THF were 19.7, 20.23 and 26.53 kcal mol-1, respectively. Activation energy for
o
cellulose decomposition during pyrolysis above 300 C was previously reported to be 45-60 kcal
mol-1.37-39 Clearly, solvolysis in polar aprotic solvents substantially reduces the activation energy
of cellulose depolymerization. Thus, it seems likely that the higher polar solubility parameters of
acetonitrile and GVL compared to THF (Table S5) contribute to their lower activation energies
for cellulose depolymerization and, hence, enhanced LG yields.
3.2.3 Formation of solubilized carbohydrate as a function of solvent polarity
Evolution of solubilized carbohydrate with DP > 1 was investigated using GFC analysis. Figure
6 (a) and (b) show GFC chromatograms of solubilized products from solvent processing of
13

Green Chemistry
Page 14 of 25
View Article Online
DOI: 10.1039/C5GC02071A
cellulose in acetonitrile and THF, respectively, for selected reaction times. It can be seen that
solubilized products gradually shifted to lower molecular weight with increasing reaction time.
These high molecular weight species are undoubtedly anhydro oligo- and polysaccharides. In
fact, the complete absence of monosaccharides and disaccharides (glucose and cellobiose) and the
prevalence of anhydro monosaccharides and disaccharides (LG and cellobiosan) support this
supposition. As reaction progressed, the molecular weight of the anhydro polysaccharides
(defined as DP > 10) decreased accompanied by an increase in LG and anhydro oligosaccharides
(defined as 2 ≤ DP ≤ 10). For both acetonitrile and THF solvent, LG and anhydro
oligosaccharides (defined as 2 ≤ DP ≤ 10) were the major solubilized carbohydrates when LG
yield reached maximum (Figure 6 (c)).
Despite the aforementioned similarities, the product distributions of solubilized carbohydrates for
acetonitrile and THF were distinct. For acetonitrile, the concentration of solubilized anhydro
polysaccharides was low during the entire course of reaction (Figure 6 (a)). As time progressed,
the LG peak increased rapidly whereas anhydro oligosaccharides gradually decreased. On the
other hand, the solubilized products in THF contained much higher concentrations of anhydro
polysaccharides with higher average DPs. Furthermore, the concentration of both LG and
anhydro oligosaccharides increased as anhydro polysaccharides decreased in THF as the reaction
progressed (Figure 6 (b)). This suggests that depolymerization of anhydro poly- and
oligosaccharides to LG proceeded more slowly in THF than in acetonitrile. This difference we
attribute to the higher activation energy barrier for cellulose depolymerization when using a lower
polarity solvent like THF compared to acetonitrile.
Solubilized products also include 5-HMF and furfural as minor products. As previously shown in
Figure 1, 5-HMF and furfural increased along with LG and reached their maximum yields at the
same time as LG. This suggests that LG, 5-HMF and furfural were simultaneously produced
rather than 5-HMF and furfural being products of secondary decomposition of LG. In fact, neither
14

Page 15 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
5-HMF nor furfural was produced when LG was reacted in THF and acetonitrile at 350 oC. These
dehydration products could be produced from the reducing ends of anhydro poly- or
oligosaccharides through well-known ring-opening40-42 fragmentation and rearrangement
mechanisms. Based on these observations, we propose Scheme 1 for the depolymerization of
cellulose in aprotic solvents.
Conclusions
We demonstrated that polar aprotic solvents at supercritical condition are capable of rapidly
converting cellulose into solubilized and depolymerized carbohydrate without the use of catalysts.
A wide range of polar aprotic solvents effectively deconstructed cellulose with maximum yields
of solubilized products reaching 72-98% and maximum yields of solubilized carbohydrate
reaching 63-94%. To our knowledge, this is the highest yield of solubilized carbohydrate reported
for non-catalytic solvent processing of cellulose. These high yields are attributed to a close
correspondence in solubility parameters for cellulose and the solvents at the elevated
temperatures and pressures of the experiments. Levoglucosan was the major carbohydrate product
with the highest maximum yield of 41% obtained using acetonitrile as solvent. Solvents with
higher polar solubility parameters lowered the activation energies for cellulose
depolymerization and promoted LG formation whereas anhydro polysaccharides and
oligosaccharides were preferentially produced in solvents with lower polar solubility parameter.
Acknowledgement
Financial support from the Iowa Energy Center is greatly acknowledged. The authors would like
to thank Tanner C. Lewis for his assistance in performing experiments. The expertise of Patrick
Johnston, Marjorie Rover and Patrick Hall in analyzing products is also greatly appreciated.
15

Green Chemistry
Page 16 of 25
View Article Online
DOI: 10.1039/C5GC02071A
References
1.
2.
P. McKendry, Bioresource Technol, 2002, 83, 37-46.
D. Klemm, B. Heublein, H. P. Fink and A. Bohn, Angewandte Chemie International
Edition, 2005, 44, 3358-3393.
3.
B. C. Saha, L. B. Iten, M. A. Cotta and Y. V. Wu, Process Biochemistry, 2005, 40, 3693-
3700.
4.
5.
6.
7.
C. Tengborg, M. Galbe and G. Zacchi, Biotechnol Progr, 2001, 17, 110-117.
M. Jarvis, Nature, 2003, 426, 611-612.
Y. Nishiyama, P. Langan and H. Chanzy, J Am Chem Soc, 2002, 124, 9074-9082.
R. C. Brown and C. Stevens, Thermochemical Processing of Biomass: Conversion into
Fuels, Chemicals and Power, Wiley, 2011.
8.
9.
L. Jarboe, Z. Wen, D. Choi and R. Brown, Appl Microbiol Biot, 2011, 91, 1519-1523.
M. R. Rover, P. A. Johnston, T. Jin, R. G. Smith, R. C. Brown and L. Jarboe,
ChemSusChem, 2014, 7, 1662-1668.
10.
C. Longley and D. C. Fung, in Advances in Thermochemical Biomass Conversion, ed. A.
V. Bridgwater, Springer Netherlands, 1993, DOI: 10.1007/978-94-011-1336-6_120, ch.
120, pp. 1484-1494.
11.
12.
13.
F. Behrendt, Y. Neubauer, M. Oevermann, B. Wilmes and N. Zobel, Chemical
Engineering & Technology, 2008, 31, 667-677.
X. L. Bai, R. C. Brown, J. Fu, B. H. Shanks and M. Kieffer, Energy & Fuels, 2014, 28,
1111-1120.
J. D. Wright and A. J. Power, Comparative technical evaluation of acid hydrolysis
processes for conversion of cellulose to alcohol, Solar Energy Research Inst., Golden,
CO (USA), 1986.
14.
15.
M. J. Taherzadeh and K. Karimi, Bioresources, 2007, 2, 472-499.
R. P. Swatloski, S. K. Spear, J. D. Holbrey and R. D. Rogers, J Am Chem Soc, 2002, 124,
4974-4975.
16.
17.
J. B. Binder and R. T. Raines, P Natl Acad Sci USA, 2010, 107, 4516-4521.
S. M. Sen, J. B. Binder, R. T. Raines and C. T. Maravelias, Biofuel Bioprod Bior, 2012,
6, 444-452.
18.
M. Sasaki, Z. Fang, Y. Fukushima, T. Adschiri and K. Arai, Industrial & Engineering
Chemistry Research, 2000, 39, 2883-2890.
19.
20.
21.
22.
T. Adschiri, S. Hirose, R. Malaluan and K. Arai, J Chem Eng Jpn, 1993, 26, 676-680.
S. Saka and T. Ueno, Cellulose, 1999, 6, 177-191.
K. Ehara and S. Saka, Journal of Wood Science, 2005, 51, 148-153.
E. I. Gurbuz, J. M. R. Gallo, D. M. Alonso, S. G. Wettstein, W. Y. Lim and J. A.
Dumesic, Angew Chem Int Edit, 2013, 52, 1270-1274.
23.
24.
M. A. Mellmer, C. Sener, J. M. R. Gallo, J. S. Luterbacher, D. M. Alonso and J. A.
Dumesic, Angewandte Chemie International Edition, 2014, 53, 11872-11875.
D. M. Alonso, J. M. R. Gallo, M. A. Mellmer, S. G. Wettstein and J. A. Dumesic, Catal
Sci Technol, 2013, 3, 927-931.
25.
26.
Y. Ishikawa and S. Saka, Cellulose, 2001, 8, 189-195.
P. Koll and J. Metzger, Angewandte Chemie-International Edition in English, 1978, 17,
754-755.
27.
28.
29.
H. Kawamoto, W. Hatanaka and S. Saka, J Anal Appl Pyrol, 2003, 70, 303-313.
G. Bao, S. Shiro and H. Wang, Science in China Series B: Chemistry, 2008, 51, 479-486.
J. S. Luterbacher, J. M. Rand, D. M. Alonso, J. Han, J. T. Youngquist, C. T. Maravelias,
B. F. Pfleger and J. A. Dumesic, Science, 2014, 343, 277-280.
F. Cao, T. J. Schwartz, D. J. McClelland, S. H. Krishna, J. A. Dumesic and G. W. Huber,
Energy & Environmental Science, 2015, 8, 1808-1815.
30.
16

Page 17 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
31.
32.
R. Weingarten, A. Rodriguez-Beuerman, F. Cao, J. S. Luterbacher, D. M. Alonso, J. A.
Dumesic and G. W. Huber, Chemcatchem, 2014, 6, 2229-2234.
Y. Su, H. M. Brown, X. Huang, X.-d. Zhou, J. E. Amonette and Z. C. Zhang, Applied
Catalysis A: General, 2009, 361, 117-122.
33.
34.
W. L. Archer, Industrial & Engineering Chemistry Research, 1991, 30, 2292-2298.
A. F. Barton, CRC handbook of solubility parameters and other cohesion parameters,
CRC press, 1991.
35.
C. M. Hansen, The Three Dimensional Solubility Parameter and Solvent Diffusion
Coefficient: Their Importance in Surface Coating Formalation, Danish Technical Press,
1967.
36.
37.
38.
C. M. Hansen, Hansen solubility parameters: a user's handbook, CRC press, 2012.
R. K. Agrawal, Can J Chem Eng, 1988, 66, 413-418.
A. G. W. Bradbury, Y. Sakai and F. Shafizadeh, Journal of Applied Polymer Science,
1979, 23, 3271-3280.
39.
40.
41.
42.
G. Varhegyi, E. Jakab and M. J. Antal, Energy & Fuels, 1994, 8, 1345-1352.
J. B. Paine, Y. B. Pithawalla and J. D. Naworal, J Anal Appl Pyrol, 2008, 83, 37-63.
J. B. Paine, Y. B. Pithawalla and J. D. Naworal, J Anal Appl Pyrol, 2008, 82, 10-41.
M. S. Mettler, A. D. Paulsen, D. G. Vlachos and P. J. Dauenhauer, Green Chem, 2012,
14, 1284-1288.
17

Green Chemistry
Page 18 of 25
View Article Online
DOI: 10.1039/C5GC02071A
Table 1. Solubilized products in different solvents (350 oC and 20 mg mass loading)
Maximum
LG yield
(%)
Time to reach
maximum LG yield
(min)
Solubilized products
yield at maximum LG
yield (%)
Solvent
1,4-dioxane
Ethyl acetate
THF
15
21
22
25
25
34
38
8
8
~12
12
10
8
72
81
91
98
95
93
95
MIBK
Acetone
GVL
Acetonitrile
16
Table 2. Yields of solubilized carbohydrate, and selectivity of LG in solubilized products, both at
the optimum condition for LG production, and also thermal stability of LG in different solvents
(350 oC and 20 mg mass loading)
Solubilized
carbohydrate in
Selectivity of LG Degradation rate of LGa (mol
Solvent
in liquid (%)
min-1)
liquid (%)
1,4-dioxane
Ethyl acetate
THF
63
72
81
91
89
86
94
16
26
28
25
26
36
40
0.063
0.250
0.159
0.098
0.207
0.681
0.216
MIBK
Acetone
GVL
Acetonitrile
a: Calculated after maximum in LG yield
18

Page 19 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
1,4-dioxane
Ethyl acetate
20
15
10
5
25
20
15
10
5
0
0
0
10
20
30
0
10
20
Time of reaction (min)
Time of reaction (min)
MIBK
THF
25
20
15
10
5
30
25
20
15
10
5
0
0
0
10
20
30
0
10
20
30
Time of reaction (min)
Time of reaction (min)
Acetonitrile
Acetone
40
30
25
20
15
10
5
30
20
10
0
0
0
10
20
30
0
10
20
30
Time of reaction (min)
Time of reaction (min)
GVL
40
30
20
10
0
0
10
20
30
Time of reaction (min)
Figure 1. Carbon molar yields of GC/MS detectable solubilized products as a function of time for
several polar aprotic solvents at 350 oC with 20 mg cellulose as feedstock ( LG; X 5-HMF;
Furfural; AGF).
19

Green Chemistry
Page 20 of 25
View Article Online
DOI: 10.1039/C5GC02071A
100
90
80
70
60
50
40
30
20
10
0
Solid residue
Furfural
5-HMF
Soluble sugars
with DP>1
AGF
LG
Figure 2. Product distribution of cellulose in different solvents reacted at 350 oC with 20 mg mass
loading of cellulose. The reaction times are varied for solvents and depend on the time to reach
the maximum LG yield in each solvent (see Table 1). The yield of solubilized carbohydrate with
DP > 1 is determined by subtracting the yields of GC/MS detectable monomers (LG, 5-HMF,
furfural and 1,6-anhydro-β-D-glucofuranose (AGF)) from the yield of solubilized products.
20

Page 21 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
(a) Acetonitrile
40
35
30
25
20
15
10
5
0
0
10
20
30
40
50
Time of reaction (min)
(b) THF
25
20
15
10
5
0
0
10
20
30
40
Time of reaction (min)
(c) THF
10
9
8
7
6
5
4
0
10
20
30
40
Time of reaction (min)
Figure 3. Effect of temperature on product yields as a function of time for cellulose
solvolysis with 20 mg mass loading of cellulose; (a) LG in acetonitrile, (b) LG in THF,
(c) 5-HMF in THF ( 325 oC; 350 oC; 375 oC).
21

Green Chemistry
Page 22 of 25
View Article Online
DOI: 10.1039/C5GC02071A
(a) Acetonitrile
45
40
35
30
25
20
15
10
5
0
0
5
10
15
20
25
30
Time of reaction (min)
(b) THF
30
25
20
15
10
5
0
0
5
10
15
20
25
30
Time of reaction (min)
(c) THF
10
9
8
7
6
5
4
3
2
1
0
0
5
10
15
20
25
30
Time of reaction (min)
Figure 4. Effect of mass loading of cellulose on product yields as function of time for
solvolysis at 350 oC (a) LG yield in acetonitrile; (b) LG yield in THF; (c) 5-HMF yield in
THF ( 10 mg loading; 20 mg loading; 50 mg loading; 100 mg loading).
22

Page 23 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
(a)
40
35
30
25
20
15
10
5
GVL
Acetonitrile
MIBK
THF
Acetone
Ethyl acetate
y = 1.0264x + 12.802
R² = 0.8654
1,4-dioxane
0
0
5
10
15
20
)
25
Polar solubility parameter, δP (MPa1/2
50
45
40
35
30
25
20
15
10
5
(b)
Acetonitrile
GVL
MIBK
THF
Acetone
Ethyl acetate
1,4-dioxane
y = 1.1042x + 15.466
R² = 0.8899
0
0
5
10
15
20
25
Polar solubility parameter, δP (MPa1/2
)
(c)
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1,4-dioxane
y = -0.0218x + 0.5431
R² = 0.8713
Ethyl acetate
MIBK
THF
Acetone
GVL
Acetonitrile
20
0
5
10
15
25
Polar solubility parameter, δP (MPa1/2
)
Figure 5. Correlation between the polar solubility parameter of the solvent and (a) maximum LG
yield; (b) LG selectivity at maximum LG yield; and (c) yield ratio of 5-HMF to LG at maximum
LG yield for cellulose depolymerization at 350 oC using 20 mg cellulose loading.
23

Green Chemistry
Page 24 of 25
View Article Online
DOI: 10.1039/C5GC02071A
LG
(a) Acetonitrile
80
70
60
50
40
30
20
10
0
3 min
5 min
8 min
10
8
DP > 100
6
DP > 10
4
2
0
Cellobiosan
MW = 324
14
19
24
DP > 100
15
DP > 10
20
10
25
Retention time (min)
(b) THF
5 min
LG
80
3 min
DP > 100
8 min
70
60
50
40
30
20
10
0
10
8
DP > 10
6
4
Cellobiosan
MW = 324
2
0
10
15
20
DP > 100
DP > 10
10
15
20
25
Retention time (min)
(c)
80
70
60
50
40
30
20
10
0
LG
THF at 12 min
Acetonitrile at 16 min
Anhydro-oligosaccharides
22
23
24
25
26
27
28
Retention time (min)
Figure 6. Molecular weight distribution of solubilized products from the solvolysis of
o
cellulose at 350 C for different reaction times in (a) Acetonitrile; (b) THF; (c)
Comparison of acetonitrile and THF at 16 min and 12 min, respectively (Peak areas were
normalized by the weight of solubilized products).
24

Page 25 of 25
Green Chemistry
View Article Online
DOI: 10.1039/C5GC02071A
Scheme 1. Proposed depolymerization pathway of cellulose in aprotic polar solvents
Cellulose
Anhydroglucose unit
Reducing end-group unit
n > 8: anhydro polysaccharides
n ≤ 8: anhydro oligosaccharides
…..
n = 1: cellotriosan
n = 0: cellobiosan
5-HMF
LG
Furfural
AGF
25