Selective and recyclable depolymerization of cellulose to levulinicacid catalyzed by acidic ionic liquid

Article abstract of DOI:10.1016/j.carbpol.2014.09.091

Cellulose depolymerization to levulinic acid (LA) was catalyzed by acidic ionic liquids (ILs) selectively and recyclably under hydrothermal conditions. The effects of reaction temperature, time, water amount and cellulose intake were investigated. Dilution effect becomes more pronounced at lower cellulose intake, dramatically improving the yield of LA to 86.1%. A kinetic model has been developed based on experimental data, whereby a good fit was obtained and kinetic parameters were derived. The relationships between IL structure, polymeric structure and depolymerization efficiency were established, shedding light on the in-depth catalytic mechanism of IL, inclusive of acidity and hydrogen bonding ability. The LA product can be readily separated through extraction by methyl isobutyl ketone (MIBK) and IL can be reused over five cycles without loss of activity. This environmentally friendly methodology can be applied to selective production of LA from versatile biomass feedstocks, including cellulose and derivatives, glucose, fructose and HMF.

Full text of DOI:10.1016/j.carbpol.2014.09.091

Carbohydrate Polymers 117 (2015) 569–576
Contents lists available at ScienceDirect
Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol
Selective and recyclable depolymerization of cellulose to levulinic
acid catalyzed by acidic ionic liquid
Huifang Rena, Buana Girisutac, Yonggui Zhoua, Li Liub,∗
a Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China
b Dalian University of Technology, Dalian 116024, China
c Institute of Chemical and Engineering Sciences, Jurong Island 627833, Singapore
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 17 March 2014
Received in revised form 1 July 2014
Accepted 26 September 2014
Available online 15 October 2014
Cellulose depolymerization to levulinic acid (LA) was catalyzed by acidic ionic liquids (ILs) selectively and
recyclably under hydrothermal conditions. The effects of reaction temperature, time, water amount and
cellulose intake were investigated. Dilution effect becomes more pronounced at lower cellulose intake,
dramatically improving the yield of LA to 86.1%. A kinetic model has been developed based on experimen-
tal data, whereby a good fit was obtained and kinetic parameters were derived. The relationships between
IL structure, polymeric structure and depolymerization efficiency were established, shedding light on the
in-depth catalytic mechanism of IL, inclusive of acidity and hydrogen bonding ability. The LA product can
be readily separated through extraction by methyl isobutyl ketone (MIBK) and IL can be reused over five
cycles without loss of activity. This environmentally friendly methodology can be applied to selective
production of LA from versatile biomass feedstocks, including cellulose and derivatives, glucose, fructose
and HMF.
Keywords:
Biomass
Cellulose
Depolymerization
Ionic liquids
Levulinic acid
© 2014 Published by Elsevier Ltd.
1. Introduction
As the most abundant biomass and sustainable raw materi-
als, cellulose consists of d-glucopyranoside units linked together
Exploring how to produce chemicals and fuels from biomass
alternatively has come to prominence due to diminishing fossil
resources and increasing concern about sustainable development,
for biomass is the only sustainable source of organic carbon. As
analogous to petrochemical refinery, biorefinery strategy has been
proposed by US National Renewable Energy Laboratory (NREL) as
a facility that integrates conversion processes and equipments to
produce fuels, power, and chemicals from biomass (Kamm, Gruber,
& Kamm, 2006). Moreover, 12 platform chemicals have been iden-
tified by the US Department of Energy (Werpy & Petersen, 2004),
which could be major outputs from an integrated biorefinery. As
one of the platform chemicals, levulinic acid (LA) is a versatile
building block for fuel additives, solvents, resin and polymer pre-
cursors, flavor substances, pharmaceutical agents and chemical
intermediates (Rackemann & Doherty, 2011). Recent work further
demonstrated that LA can serve as initial feedstock for existing
petrochemical processing operations, bridging the biomass and
petroleum processing together (Bozell, 2010). Thereby production
of LA from biomass has become one of the key steps for biorefinery.
by ␤-1,4 glycosidic bonds. Cellulose is generally regarded as a
difficult material to work with due to its densely packed struc-
ture and insolubility in water (Jarvis, 2003). Alternative media
capable of dissolving cellulose are required to facilitate biomass
processing, whereas the attribute has been demonstrated by ionic
liquids (ILs), which has been widely recognized as green media
due to their non-flammability and non-volatility. Swatloski, Spear,
Holbrey, and Rogers (2002) firstly found out [C4mim]Cl was able
to dissolve up to 25% of cellulose under microwave irradiation,
which opened up vast possibilities to biomass utilization, taking
advantage of the remarkable designability and functionality of ionic
liquids (Welton, 1999). A number of studies have been reported
on depolymerization of cellulose using ionic liquids ever since
(Amarasekara & Owereh, 2009; Binder & Raines, 2009; Dee & Bell,
2011; Kim et al., 2011; Kuo, Suzuki, Yamauchi, & Wu, 2013; Li &
Zhao, 2007; Liu, Zhang, & Zhao, 2013; Rinaldi, Palkovits, & Schüth,
2008; Song, Fan, Ma, & Han, 2013; Su et al., 2009; Tao, Song, & Chou,
2010; Vanoye, Fanselow, Holbrey, Atkins, & Seddon, 2009; Xiong,
Zhang, Wang, Liu, & Lin, 2014). Li and Zhao (2007) firstly applied
ionic liquids in cellulose conversion towards glucose, where sul-
phuric acid in [C4mim]Cl resulted in a yield of 43% for glucose.
Later Su et al. (2009) applied ionic liquids in cellulose conversion
towards 5-hydroxymethylfurfural (HMF), where copper chloride
∗
Corresponding author. Tel.: +86 411 84795945; fax: +86 411 84795945.
E-mail address: lliu@dlut.edu.cn (L. Liu).
http://dx.doi.org/10.1016/j.carbpol.2014.09.091
0144-8617/© 2014 Published by Elsevier Ltd.

570
H. Ren et al. / Carbohydrate Polymers 117 (2015) 569–576
and chromium chloride in [C2mim]Cl led to a yield of 55% for HMF.
In contrast, application of ionic liquids as catalysts in recycle fashion
has been far less explored in cellulose conversion.
chloride ([C3SO3Hmim]Cl) were prepared according to literature
and characterized by NMR prior to use (Supplementary data)
(Cole et al., 2002). The NMR spectra were recorded on Bruker
Avance III 500 MHz and Bruker DRX 400 MHz spectrometers.
UV–vis spectra were recorded on Jasco UV-550 spectrophotome-
ter. The infrared spectra were recorded on Bruker Equinox 55
infrared spectrometer. Scanning electron microscopy images were
obtained on Nova NanoSEM 450 field emission scanning electron
microscopy, where samples were coated with a thin Aurum
film.
Depolymerization of cellulose to LA has been attempted by var-
ious catalysts, including mineral acids (Girisuta, Janssen, & Heeres,
2007; Szabolcs, Molnar, Dibo, & Mika, 2013; Tabasso, Montoneri,
Carnaroglio, Caporaso, & Cravotto, 2014), metal chlorides (Seri,
Sakaki, Shibata, Inoue, & Ishida, 2002), solid acid catalysts (Alonso,
Gallo, Mellmer, Wettstein, & Dumesic, 2013; Zakzeski, Grisel, Smit,
& Weckhuysen, 2012; Zuo, Zhang, & Fu, 2014) and acidic polymers
(Vyver et al., 2011), whereas the challenging issues were focused on
improving the selectivity of LA and developing efficient strategies
for product separation and catalyst recovery. Acidic ionic liquids
have large potentials in replacing conventional acidic catalysts for
they are flexible, recyclable, and could be used as dual solvents
and catalysts (Cole et al., 2002). Applications of acidic ionic liquids
as catalysts for cellulose conversion have been reported with 62%
(Amarasekara & Owereh, 2009) and 99% (Liu, Xiao, Xia, & Ma, 2013)
yields towards total reducing sugar (TRS) product; 37% (Tao, Song,
Yang, & Chou, 2011), 53% (Zhou, Liang, Ma, Wu, & Wu, 2013), 66.5%
(Shi et al., 2013), and 69.7% (Ding et al., 2012) yields towards HMF
product, in the presence of metal chloride as co-catalyst or IL as
solvent. Recently, we employed SO3H-functionalized acidic ionic
liquids in the direct conversion of cellulose to LA under microwave
irradiation, whereas 55% of LA has been obtained (Ren, Zhou, &
Liu, 2013). Alternatively, hydrothermal conditions have attracted
increasing attentions due to the unique inherent properties of
high-temperature water, which provides environmentally friendly
reaction medium (Jin & Enomoto, 2011). In this work, we developed
selective and recyclable methodology under hydrothermal condi-
tions, dramatically improving the yield of LA to 86.1%. Continued
efforts were devoted to understand the basic nature of cellulose
depolymerization towards LA.
2.2. General procedure for cellulose conversion
In a typical procedure, cellulose, ionic liquid and de-ionized
water were mixed in a stainless steel autoclave with Teflon lining
and heated in an oil bath maintained at the desired tempera-
ture for a specified length of time with magnetic stirring. The
reaction was quenched in an ice bath. After filtration of insol-
ubles and removal of water by rotary evaporation at 50 ◦C for
5 min, the crude product was analyzed by 1H NMR using ionic
liquid as internal standard. The yields of products were calcu-
lated from the equation: yield (%) = (mol of the product)/(mol
of glucose unit in cellulose) × 100%. All the results were repli-
cated at least three times. The analytical error was evaluated
to be within 5% based on standard samples of known concen-
trations. The insoluble humins were collected from the reaction
mixture by filtration, oven-dried, and weighed, whereby the yields
of humins in weight percentage were calculated relative to cellu-
lose.
2.3. Reuse of IL
In the first run, the reaction of cellulose (550 mg),
[C3SO3Hmim]HSO4 (1.000 g) and de-ionized water (6.000 g)
were heated in a 20 mL autoclave at 170 ◦C for 5 h. After filtration
of insolubles and removal of water by rotary evaporation at 50 ◦C
for 5 min, the crude product was analyzed by 1H NMR using
ionic liquid as internal standard. The aqueous solution of ionic
liquid was extracted by MIBK 60 mL × 3, and dried under high
vacuum. Afterwards, the IL was adjusted to 1.000 g by adding
fresh IL of no more than 5% and reused in the next cycle as
above.
2. Experimental
2.1. Materials and equipments
Microcrystalline cellulose (average particle size 50 ␮m,
DP = 237, which was used in cellulose experiments unless oth-
erwise specified) and HMF (98%) were purchased from Acros
Organics (USA). ␣-Cellulose (Cat. No. C8002, DP = 100) and
sigmacell cellulose (Cat. No. S6790, DP = 400–500) were pur-
chased from Sigma (St. Louis, USA). Carboxymethylcellulose
(MW 90,000, DS = 0.7, viscosity 50–100 mPa s; MW 250,000,
DS = 0.7, viscosity 1500–3100 mPa s; MW 250,000, DS = 1.2, vis-
cosity 1500–3100 mPa s) were supplied by Aladdin Reagent Co.,
Ltd. (Shanghai, China). Cellulose and derivatives were dried
under vacuum at 80 ◦C for 24 h to diminish trace amount of
moisture prior to use. Fructose (99%) and glucose (p.a.) were
purchased from local suppliers. 1,3-Propane sultone and 1-
methylimidazole were purchased from Wuhan Fengfan Chemical
Co., Ltd. (Wuhan, China) and freshly distilled before use. Nine
ionic liquids including 1-methyl-3-(3-sulfopropyl)imidazolium
3. Results and discussion
3.1. The effects of reaction temperature and time on cellulose
depolymerization
The experiments for determining the effects of reaction temper-
ature and time on cellulose conversion were carried out in 10 mL
autoclaves at temperatures ranging from 160 to 180 ◦C, during
the reaction course of 0.5–5 h, with 1.000 g of IL 1-methyl-3-(3-
sulfopropyl)imidazolium hydrogen sulfate ([C3SO3Hmim]HSO4),
2.000 g of H2O and 250 mg of cellulose. As shown in Fig. 1a, the
yield of LA increased with the rise of reaction temperature from
160 to 180 ◦C within 2 h. Then the yield of LA reached ca. 50% at
170 and 180 ◦C, slightly higher than 160 ◦C, suggesting higher reac-
tion temperature is favorable to LA formation during short reaction
time. When the reaction time was prolonged over 4 h, the yield of LA
increased to 53.7% at 160 ◦C and levelled off afterwards, surpassing
170 and 180 ◦C. It can be learnt that through extending the reaction
time, the yield of LA can be further improved at lower temperature.
Therefore the cellulose conversion is sensitive to the effects of reac-
tion temperature and time in the presence of 2 g of water, which
hydrogen
sulfate
([C3SO3Hmim]HSO4),
1-methyl-3-(4-
sulfobutyl)imidazolium hydrogen sulfate ([C4SO3Hmim]HSO4),
N-(3-sulfopropyl)pyridinium hydrogen sulfate ([C3SO3HPy]HSO4),
N,N,N-triethyl-N-(3-sulfopropyl)ammonium hydrogen sulfate
([C3SO3HN222]HSO4),
hydrogen sulfate
(3-sulfopropyl)imidazolium
([C3SO3Hmim]H2PO4), 1-methyl-3-(3-sulfopropyl)imidazolium
methanesulfonate
3-(3-sulfopropyl)imidazolium
([C3SO3Hmim]1-NS), and 1-methyl-3-(3-sulfopropyl)imidazolium
triphenyl(3-sulfopropyl)phosphonium
([C3SO3HPPh3]HSO4),
dihydrogen
1-methyl-3-
phosphate
([C3SO3Hmim]CH3SO3),
1-methyl-
1-naphthalenesulfonate

H. Ren et al. / Carbohydrate Polymers 117 (2015) 569–576
571
Fig. 2. The effect of cellulose intake at different temperatures: (ꢀ) 160 ◦C; (ꢁ) 170 ◦C;
(ꢂ) 180 ◦C. Conditions: 1.000 g [C3SO3Hmim]HSO4, 5 h, (a) 3.000 g H2O, (b) 6.000 g
H2O.
Fig. 1. The effects of time and water amount at different temperatures: (ꢀ) 160 ◦C;
(ꢁ) 170 ◦C; (ꢂ) 180 ◦C. Conditions: 250 mg cellulose, 1.000 g [C3SO3Hmim]HSO4, (a)
2.000 g H2O, (b) 5 h.
substantial catalytic activity is necessary to reflect the positive dilu-
tion effect.
seems to proceed faster initially at higher temperature and result
in higher yield eventually at lower temperature.
3.2. The effect of water amount on cellulose depolymerization
3.3. The effect of cellulose intake on cellulose depolymerization
As green solvent, water plays a crucial role in the hydrothermal
conversion of cellulose to LA. Fig. 1b shows the effect of water
amount on the yield of LA at different temperatures. The yield of
LA varied in analogous tendency at 170 and 180 ◦C, different from
160 ◦C. When the water amount ranged from 2 g to 6 g, the yield
of LA increased from 52.2% to 62.1% at 170 ◦C, and from 51.5%
to 60.8% at 180 ◦C, which decreased slightly to 60.3% and 58.7%
in the presence of 7 g of water, respectively. With water amount
increased, dilution effect suppresses the formation of dark colored
solid byproducts, known as humins, leading to higher yield of LA.
Howeverat 160 ◦C, the yield of LA increasedfirst and then decreased
from 58.3% to 37.5% with increasing water amount from 4 g to 7 g,
ascribing to severely lowered catalytic activity which could not be
compromised by the dilution effect. Meanwhile, 15.4% of glucose
was observed in the presence of 7 g of water at 160 ◦C. Apparently,
the temperature has dramatic effect on cellulose conversion. With
water amount enlarged, higher reaction temperature is needed
to convert the glucose intermediate completely to LA, suggesting
From an economical point of view, the cellulose digestibility is
an important perspective of the conversion process. Fig. 2 shows
the effect of cellulose intake on the yield of LA with the addition of
3 g (Fig. 2a) and 6 g of water (Fig. 2b), respectively. In the presence
of 3 g of water, the yield of LA decreased slightly from 60% to 50%
in similar trends with increasing cellulose intake from 150 mg to
550 mg at explored temperatures, attributing to that higher cellu-
lose dosage induces humin formation and consequently restrains
LA production. In the presence of 6 g of water, the yield of LA
decreased from 51.7% to 48.1% at 160 ◦C, from 64.3% to 56.4% at
170 ◦C, and from 61.2% to 57.6% at 180 ◦C, with increasing cellu-
lose intake from 150 mg to 550 mg. In contrast, 160 ◦C is not high
enough to convert cellulose efficiently to LA when 6 g of water was
added. While higher temperatures of 170 and 180 ◦C resulted in
higher yields of LA with the addition of 6 g of water rather than
3 g of water, confirming that water can restrain the side prod-
ucts of humins to considerable extent. It confirms that dilution
effect shows up in combination with higher temperature especially

572
H. Ren et al. / Carbohydrate Polymers 117 (2015) 569–576
Fig. 3. (a) IR spectrum and (b) SEM images of humins after complete conversion of cellulose. Conditions: 150 mg cellulose, 1.000 g [C3SO3Hmim]HSO4, 6.000 g H2O, 170 ◦C,
5 h.
170 ◦C for substantial activation, which becomes more pronounced
at lower cellulose intake.
thus provides simple and efficient paradigm for IL application in
biorefinery.
To further look into the distinctive dilution effect on the yield of
LA observed at low cellulose intake, the cellulose conversion was
subjected to a broad range of cellulose intake under the same con-
dition (1.000 g IL, 6.000 g H2O, 170 ◦C, 5 h). With cellulose intake
reduced to 20 mg, the yield of LA improved up to 86.1% (Table A.1,
Supplementary data). It verified that low cellulose intake can spur
the transformation of cellulose to LA effectively. On the other hand,
with increase of cellulose loadings, the yield of LA decreased. Nev-
ertheless, when the ratio of cellulose to IL enhanced to 220%, 49.7%
yield of LA could still be reached, demonstrating improved cellu-
lose digestibility of IL. Nowadays, for applications of ionic liquids
in cellulose conversion, increasing cellulose digestibility is highly
demanding. When ionic liquids act as solvent (Binder & Raines,
2009; Dee & Bell, 2011; Kim et al., 2011; Kuo et al., 2013; Li & Zhao,
2007; Rinaldi et al., 2008; Su et al., 2009; Vanoye et al., 2009; Xiong
et al., 2014), the cellulose/IL ratio is usually in the range of 5–10 wt%,
whereas additional catalysts of acid or metal chloride are needed.
When ionic liquids act as catalyst (Amarasekara & Owereh, 2009;
Liu, Xiao, et al., 2013; Tao et al., 2010), the cellulose/IL ratio is mostly
in the range of 20–50 wt%, however normally co-catalyst or IL as
solvent is required to improve the selectivity of the desired prod-
uct. Herein, cellulose has been converted to LA in high selectivity
with [C3SO3Hmim]HSO4 acting as catalyst, exclusive of co-catalyst
and [C4mim]Cl as solvent, whereas ca. 50% yield of LA could be
obtained at cellulose/IL ratio up to 220%. The current methodology
3.4. Mechanism and kinetics of cellulose depolymerization to LA
In fact, it is a complicated multistep process of cellulose con-
version into LA, which involves several important intermediates.
As generally recognized, cellulose depolymerizes to glucose first,
and glucose transforms to 5-hydroxymethylfurfural (HMF), which
further rehydrates to form LA. In our catalytic process towards the
target product of LA, other than glucose, only tract amount of HMF
was observed under incomplete conversion of cellulose. Since all
our experiments were carried out at temperatures <180 ◦C, this
observation is in line with the previous study (Girisuta et al., 2007),
which has shown that significant amount of HMF would form only
at reaction temperatures >200 ◦C.
As mentioned before, we observed the parallel formation of
byproduct humins, which are suggested to proceed through the
cross-polymerization or condensation of glucose and HMF (Patil,
Heltzel, & Lund, 2012; Sievers et al., 2009). The infrared spec-
trum of humins from cellulose (Fig. 3a) shows the absorbance
bands at 3400 cm−1 attributed to O H stretching vibrations,
2929 cm−1 attributed to aliphatic C H stretching vibrations, 1701
and 1618 cm−1 corresponding to C O and C C vibrations, respec-
tively. The bands in the 1000–1450 cm−1 region are assigned to C
O
stretching and O H bending vibrations. The bands at 750–875 cm−1
are ascribed to aromatic C H out-of-plane bending vibrations. No

H. Ren et al. / Carbohydrate Polymers 117 (2015) 569–576
573
Table 1
The kinetic parameters of cellulose depolymerization to LA.
a
Reaction
kR (min−1
)
E (kJ/mol)
Cellulose → Glucose
Glucose → LA
0.134 0.013
0.258 0.016
0.206 0.014
121.78 7.19
202.55 5.34
166.85 6.51
Glucose → Humins
a
The reference temperature was set at 170 ◦C.
As only trace amount of HMF was observed throughout our
kinetics experiments, the formation reaction of HMF and its subse-
quent decomposition reactions were not included in the kinetic
model. Accordingly, we proposed that glucose was converted
directly to LA and simultaneously decomposed to the undesired
product humins, whereby their reaction rates were defined as Eqs.
(2) and (3), respectively:
RG1 = kG1CGLC
RG2 = kG2CGLC
(2)
(3)
The following equation was used to define the reaction rate
constants given in Eqs. (1)–(3):
ꢀ ꢁ
ꢂꢃ
E
T−T
i
R
R
T T
R
ki = kR,iexp
i ≡ C, G1, G2
(4)
The kinetic parameters defined in Eq. (4) were estimated by min-
imizing the sum of square error between the experimental data and
the predicted values obtained from the solution of these differential
Eqs. (5)–(7):
dCCEL
= −RC
(5)
(6)
(7)
dt
dCGLC
dt
= RC − RG1 − RG2
Fig. 4. Reaction pathway of cellulose depolymerization to LA catalyzed by IL.
dCLA
dt
= RG1
absorbance peaks ascribed to IL were observed, indicating no IL
remained in humin byproduct, which was confirmed by the ele-
mental analysis of humins (C, 64.560; H, 4.421; N, 0.000). The
scanning electron microscopy (SEM) images of humins show two
types of particles, including big isolated spherical particles (ca.
2 ␮m) and small agglomerated particles less than 1 ␮m (Fig. 3b).
In our catalytic system, the yields of humins, in the range of 14–22
wt%, decreased with increase of water amount after complete con-
version of cellulose at 170 and 180 ◦C (Fig. A.1, Supplementary
data), whereas the quantification of humins seemed unlikely at
160 ◦C due to the mixing with unconverted cellulose starting mate-
rial. Furthermore, the dilution effect at 170 ◦C is more notable than
that at 180 ◦C, confirming 170 ◦C is the optimum reaction temper-
ature.
A kinetic study was carried out to gain an in-depth insight on
the cellulose depolymerization to LA catalyzed by ionic liquid. The
first systematic kinetic study on biomass hydrolysis to glucose
was developed by Saeman (1945). Afterwards, kinetic studies on
H2SO4-catalyzed hydrolysis of cellulose to LA have been reported
by Girisuta et al. (2007). In this work, we proposed the following
kinetic model for the IL-catalyzed cellulose depolymerization to LA,
including the depolymerization of cellulose to glucose and the sub-
sequent conversion of glucose to LA and humins (Fig. 4), whereas
the reaction rates were approximated using the pseudo-first-order
approach.
A MATLAB optimisation routine fminsearch was used to obtain
the best estimates of the kinetic parameters, which are given in
Table 1. A good fit between the experimental data and kinetic model
was obtained, as shown in Fig. 5.
The slowest reaction rate constant at the reference tempera-
ture (kR) was observed for the depolymerization of cellulose to
glucose (0.134 min−1), which suggests that this reaction could be
the rate-limiting step in the whole steps of cellulose conversion to
LA. The highest reaction rate constant was observed in the reaction
of glucose to LA (0.258 min−1), which indicates that the conversion
of glucose to LA was the fastest reaction compared to the other
reactions. Based on the value of kR for glucose to humins reac-
tion (0.206 min−1), it can be calculated that about 55.6% of the
glucose converted would form LA as the desired product selec-
tively, which is very close to our experimental data using glucose
as the starting material (60.0%). The depolymerization reaction
of cellulose to glucose catalyzed by IL has the lowest activation
energy (121.78 kJ/mol), which is relatively lower than 151.5 kJ/mol
reported for H2SO4 catalyst (Girisuta et al., 2007). This result indi-
cates that depolymerization of cellulose catalyzed by IL is more
favored energetically.
3.5. The effect of IL structures on cellulose depolymerization
One of the advantages of ionic liquid is designability, whereas
the desired properties can be obtained through adjusting the
cationic and anionic structures of ionic liquids. Some insights were
put into the effect of IL structures on cellulose conversion to LA,
whereby the Brønsted acidities of ILs were determined by Hammett
method (Gu, Zhang, Duan, & Deng, 2005) using UV–vis spectropho-
tometer with 4-nitroaniline as indicator. The maximal absorbance
The first step in the depolymerization of cellulose was the cleav-
age of ␤-(1 → 4)-glycosidic bond to glucose. The reaction rate for
this depolymerization step (RC) was expressed by Eq. (1), whereas
the concentration of cellulose (CCEL) was defined as the number of
available anhydroglucose units:
RC = kCCCEL
(1)

574
H. Ren et al. / Carbohydrate Polymers 117 (2015) 569–576
Table 2
Cellulose conversion to LA catalyzed by acidic ionic liquids.a
IL
Amax
[I] (%)
[IH+] (%)
H0
YLA (mol%)
–
0.401
0.249
0.266
0.270
0.259
0.264
0.370
0.267
0.323
0.354
100
0
–
0
[C3SO3Hmim]HSO4
[C4SO3Hmim]HSO4
[C3SO3HN222]HSO4
[C3SO3HPy]HSO4
[C3SO3HPPh3]HSO4
[C3SO3Hmim]H2PO4
[C3SO3Hmim]CH3SO3
[C3SO3Hmim]1-NS
[C3SO3Hmim]Cl
62.1
66.4
67.4
64.6
65.8
92.2
66.6
80.6
88.2
37.9
33.6
32.6
35.4
34.2
7.8
33.4
19.4
11.8
1.204
1.285
1.305
1.251
1.273
2.063
1.289
1.607
1.863
63.0
57.6
54.5
56.9
55.2
16.7
53.9
53.8
66.2
a
All the experiments were conducted at 170 ◦C for 5 h with 150 mg of cellulose,
3.31 mmol of IL and 6.000 g of H2O.
> pyridinium > triphenylphosphonium > triethylammonium,
whereas elongating the alkyl chain linked with the sulfonic acid
group from propyl to butyl also causes small decrease of acidity. In
contrast, for ILs with the same cationic structure of [C3SO3Hmim],
the acidities of ILs are more dependent on the anions and decrease
in the order: HSO4− > CH3SO3− > 1-NS > H2PO4−. Consequently,
the stronger the acidity of ILs, the higher yield of LA was obtained,
indicating the essential role of the acidity in the catalytic cleavage
of cellulose, possibly through protonation of glycosidic oxygen.
Surprisingly, when [C3SO3Hmim]Cl was tested, the highest yield
of LA (66.2%) was achieved, although its acidity is much lower
than that of [C3SO3Hmim]HSO4. Rogers and co-workers (Remsing,
Swatloski, Rogers, & Moyna, 2006; Swatloski et al., 2002) reported
that hydrogen bonding exists between the carbohydrate hydroxyl
protons and the IL Cl−, which helps to break the extensive hydro-
gen bonding network of cellulose and facilitate its dissolution.
Herein, it shed light on that stronger hydrogen bond acceptors as
IL anions could presumably increase accessibility of the polymer
to the acid sites and then promote the catalytic performance of IL
in cellulose conversion, apart from acidity factor.
3.6. Application to versatile biomass feedstocks
The catalytic methodology of IL can be applied to produce
LA product from versatile biomass feedstocks. When organic
molecules such as glucose, fructose and HMF were used as starting
materials instead of cellulose under the same reaction conditions,
as shown in Table 3, the yields of LA increase in the order glu-
cose (60.0%, entry 1) < fructose (76.7%, entry 2) < HMF (93.4%, entry
3). Furthermore, by using a variety of cellulose and derivatives as
starting materials, the relationship between polymeric structure
and yield of LA was established for the first time. The results are
summarized in Table 3.
Firstly, the effect of degree of polymerization (DP) of cellulose
on the yield of LA was investigated, ranging from 100 to 500. The
yield of LA reached the highest of 65.6% for sigmacell cellulose
Table 3
Conversion of different biomass feedstocks into LA.a
Fig. 5. Comparison of experimental data (ꢀ: Glucose concentration; ᭹: LA concen-
tration) and kinetic model (solid and dashed lines). Conditions: 250 mg cellulose,
1.000 g [C3SO3Hmim]HSO4, 2.000 g H2O, (a) 160 ◦C, (b) 170 ◦C, (c) 180 ◦C.
Entry
Feedstock
DP
DS
YLA (mol%)
1
2
3
4
5
6
7
8
9
Glucose
Fructose
HMF
Cellulose
Cellulose
Cellulose
CMC (MW 90,000)
CMC (MW 250,000)
CMC (MW 250,000)
–
–
–
100
237
400–500
–
–
–
–
–
–
0
0
0
0.7
0.7
1.2
60.0
76.7
93.4
56.4
63.0
65.6
44.0
43.9
40.1
of the unprotonated form of the indicator was observed at 381 nm
in H2O, which decreased after adding acidic ionic liquid. Under
the same concentration of 4-nitroaniline (4 mg/L) and ionic liquids
(50 mmol/L) in H2O, Hammett acidity functions (H0) of ILs with dif-
ferent cations and anions were calculated and compared thereafter.
As summarized in Table 2, the type of cations has slight effect on
the acidities of ILs which decrease in the order: imidazolium
a
All the experiments were conducted at 170 ◦C for 5 h with 0.925 mmol of feed-
stock, 1.000 g of [C3SO3Hmim]HSO4 and 6.000 g of H2O.

H. Ren et al. / Carbohydrate Polymers 117 (2015) 569–576
575
(DP = 400–500, entry 6), followed by 63.0% for microcrystalline cel-
Acknowledgments
lulose (DP = 237, entry 5) and 56.4% for ␣-cellulose (DP = 100, entry
4) in order. It seems that the yield of LA increases slightly with
higher DP. The depolymerization efficiency towards LA is hindered
by the parallel formation of humins. The higher DP, the harder
and slower hydrolysis of cellulose (Fig. A.2, Supplementary data).
Accordingly there were less glucose and HMF accumulated in the
conversion process of cellulose with higher DP than lower DP,
which may reduce the polymerization and condensation reactions
to form humins, hence leading to higher yield of LA.
Secondly, the effect of degree of substitution (DS) of cellu-
lose derivatives on the yield of LA was evaluated. Three kinds
of carboxymethylcellulose (CMC) with different DS and molecu-
lar weight (MW) were compared under the same conditions. CMC
(MW 90,000) and CMC (MW 250,000) with the same DS of 0.7 gave
similar yields of LA (entry 7 and 8), whereas CMC (MW 250,000)
with DS of 0.7 gave a higher yield of LA than CMC (MW 250,000)
with DS of 1.2 (entry 9), indicating DS has a more perceptible effect
than DP. The more hydroxyl groups were substituted, the more
difficultly the cellulose derivatives were depolymerized to LA, high-
lighting the key role of hydroxyl groups during the breakdown of
polymeric backbones. It is likely due to hydrogen bonding between
hydroxyl protons and IL anions increasing accessibility of the poly-
mer to acid sites and promoting the catalytic process, also in line
with the result that IL with stronger hydrogen bond acceptor as
anion led to higher yield of LA.
This project was supported by the Fundamental Research Funds
for the Central Universities. L. L. acknowledges DICP-100-Talents
Program for financial support.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.carbpol.
2014.09.091.
References
Alonso, D. M., Gallo, J. M. R., Mellmer, M. A., Wettstein, S. G., & Dumesic, J. A. (2013).
Direct conversion of cellulose to levulinic acid and gamma-valerolactone using
solid acid catalysts. Catalysis Science and Technology, 3, 927–931.
Amarasekara, A. S., & Owereh, O. S. (2009). Hydrolysis and decomposition of cellulose
in brönsted acidic ionic liquids under mild conditions. Industrial and Engineering
Chemistry Research, 48, 10152–10155.
Binder, J. B., & Raines, R. T. (2009). Simple chemical transformation of lignocellulosic
biomass into furans for fuels and chemicals. Journal of the American Chemical
Society, 131, 1979–1985.
Bozell, J. J. (2010). Connecting biomass and petroleum processing with a chemical
bridge. Science, 329, 522–523.
Cole, A. C., Jensen, J. L., Ntai, I., Tran, K. L. T., Weaver, K. J., Forbes, D. C., et al. (2002).
Novel brønsted acidic ionic liquids and their use as dual solvent-catalysts. Journal
of the American Chemical Society, 124, 5962–5963.
Dee, S. J., & Bell, A. T. (2011). A study of the acid-catalyzed hydrolysis of cellulose
dissolved in ionic liquids and the factors influencing the dehydration of glucose
and the formation of humins. ChemSusChem, 4, 1166–1173.
Ding, Z.-D., Shi, J.-C., Xiao, J.-J., Gu, W.-X., Zheng, C.-G., & Wang, H.-J. (2012). Catalytic
conversion of cellulose to 5-hydroxymethyl furfural using acidic ionic liquids
and co-catalyst. Carbohydrate Polymers, 90, 792–798.
3.7. Separation of LA and reuse of IL
Girisuta, B., Janssen, L. P. B. M., & Heeres, H. J. (2007). Kinetic study on the acid-
catalyzed hydrolysis of cellulose to levulinic acid. Industrial and Engineering
Chemistry Research, 46, 1696–1708.
The recyclability of IL was investigated with 550 mg of cellu-
lose, 1.000 g of [C3SO3Hmim]HSO4 and 6.000 g of H2O at 170 ◦C for
5 h. After the reaction, a variety of organic solvents were screened
to extract LA product from the aqueous solution of IL and methyl
isobutyl ketone (MIBK) proved to be optimum extraction solvent,
where the extraction efficiency was as high as 98% and over 95%
of the IL can be recovered. The yield of LA did not decrease dur-
ing the recycle, but increased gradually from 58.5% to 65.7% after
five cycles (Fig. A.3, Supplementary data), demonstrating that the
IL maintained high catalytic activity. Through comparison of 1H
NMR spectra among IL recycled after five times, fresh IL and its
zwitterionic precursor, it can be seen that the IL catalyst remains
stable under the reaction conditions and does not decompose to
the corresponding zwitterion (Fig. A.4, Supplementary data). It is
assumed that the increase of LA yield was attributed to the residual
oligomers of cellulose dissolved in the reaction mixture which can
further transform to LA in the subsequent cycle. Moreover, LA can
be routinely distilled as pure product from MIBK which is reusable
too.
Gu, Y., Zhang, J., Duan, Z.,
& Deng, Y. (2005). Pechmann reaction in non-
chloroaluminate acidic ionic liquids under solvent-free conditions. Advanced
Synthesis and Catalysis, 347, 512–516.
Jarvis, M. (2003). Cellulose stacks up. Nature, 426, 611–612.
Jin, F., & Enomoto, H. (2011). Rapid and highly selective conversion of biomass
into value-added products in hydrothermal conditions: Chemistry of acid/base-
catalysed and oxidation reactions. Energy and Environmental Science, 4, 382–397.
Kamm, B., Gruber, P. R., & Kamm, M. (2006). Biorefineries—Industrial processes and
products. Weinheim: Wiley-VCH.
Kim, B., Jeong, J., Lee, D., Kim, S., Yoon, H.-J., Lee, Y.-S., et al. (2011). Direct trans-
formation of cellulose into 5-hydroxymethyl-2-furfural using a combination of
metal chlorides in imidazolium ionic liquid. Green Chemistry, 13, 1503–1506.
Kuo, I-J., Suzuki, N., Yamauchi, Y., & Wu, K. C.-W. (2013). Cellulose-to-HMF con-
version using crystalline mesoporous titania and zirconia nanocatalysts in ionic
liquid systems. RSC Advances, 3, 2028–2034.
Li, C., & Zhao, Z. K. (2007). Efficient acid-catalyzed hydrolysis of cellulose in ionic
liquid. Advanced Synthesis and Catalysis, 349, 1847–1850.
Liu, B., Zhang, Z., & Zhao, Z. K. (2013). Microwave-assisted catalytic conversion of
cellulose into 5-hydroxymethylfurfural in ionic liquids. Chemical Engineering
Journal, 215, 517–521.
Liu, Y., Xiao, W., Xia, S., & Ma, P. (2013). SO3H-functionalized acidic ionic liquids as
catalysts for the hydrolysis of cellulose. Carbohydrate Polymers, 92, 218–222.
Patil, S. K. R., Heltzel, J.,
& Lund, C. R. F. (2012). Comparison of structural
features of humins formed catalytically from glucose, fructose, and 5-
hydroxymethylfurfuraldehyde. Energy and Fuels, 26, 5281–5293.
Rackemann, D. W., & Doherty, W. O. S. (2011). The conversion of lignocellulosics to
levulinic acid. Biofuels Bioproducts and Biorefining, 5, 198–214.
Remsing, R. C., Swatloski, R. P., Rogers, R. D., & Moyna, G. (2006). Mechanism of cel-
lulose dissolution in the ionic liquid 1-n-butyl-3-methylimidazolium chloride:
A 13C and 35/37Cl NMR relaxation study on model systems. Chemical Communi-
cations, 1271–1273.
Ren, H., Zhou, Y., & Liu, L. (2013). Selective conversion of cellulose to levulinic acid
via microwave-assisted synthesis in ionic liquids. Bioresource Technology, 129,
616–619.
Rinaldi, R., Palkovits, R., & Schüth, F. (2008). Depolymerization of cellulose using
solid catalysts in ionic liquids. Angewandte Chemie International Edition, 47,
8047–8050.
Saeman, J. F. (1945). Kinetics of wood saccharification—Hydrolysis of cellulose and
decomposition of sugars in dilute acid at high temperature. Industrial and Engi-
neering Chemistry, 37, 43–52.
Seri, K., Sakaki, T., Shibata, M., Inoue, Y., & Ishida, H. (2002). Lanthanum (III)-catalyzed
degradation of cellulose at 250 ◦C. Bioresource Technology, 81, 257–260.
Shi, J., Gao, H., Xia, Y., Li, W., Wang, H., & Zheng, C. (2013). Efficient process for
the direct transformation of cellulose and carbohydrates to 5-(hydroxymethyl)
4. Conclusions
A highly selective and recyclable approach to depolymerize cel-
lulose to levulinic acid was developed by catalysis of acidic ionic
liquids under hydrothermal conditions. The optimum reaction con-
ditions were 1 g of IL and 6 g of water at 170 ◦C for 5 h, whereas
dilution effect becomes more pronounced at lower cellulose intake
of 20 mg, dramatically improving the yield of LA up to 86.1%. Nev-
ertheless, at a higher cellulose/IL ratio of 220%, ca. 50% yield of LA
could be reached, demonstrating improved cellulose digestibility
of IL. The catalytic mechanism of IL was in-depth explored, inclu-
sive of acidity and hydrogen bonding ability, which mostly depend
on anions. The acidic IL-catalyzed methodology provides green
and simple paradigm to selective production of LA from versatile
biomass feedstocks.

576
H. Ren et al. / Carbohydrate Polymers 117 (2015) 569–576
furfural with dual-core sulfonic acid ionic liquids and co-catalysts. RSC Advances,
3, 7782–7790.
Sievers, C., Musin, I., Marzialetti, T., Olarte, M. B. V., Agrawal, P. K., & Jones, C. W.
(2009). Acid-catalyzed conversion of sugars and furfurals in an ionic-liquid
phase. ChemSusChem, 2, 665–671.
Song, J., Fan, H., Ma, J., & Han, B. (2013). Conversion of glucose and cellulose
into value-added products in water and ionic liquids. Green Chemistry, 15,
2619–2635.
Su, Y., Brown, H. M., Huang, X., Zhou, X., Amonette, J. E., & Zhang, Z. C. (2009).
Single-step conversion of cellulose to 5-hydroxymethylfurfural (HMF), a ver-
satile platform chemical. Applied Catalysis A: General, 361, 117–122.
Swatloski, R. P., Spear, S. K., Holbrey, J. D., & Rogers, R. D. (2002). Dissolution of cellu-
lose with ionic liquids. Journal of the American Chemical Society, 124, 4974–4975.
Szabolcs, A., Molnar, M., Dibo, G., & Mika, L. T. (2013). Microwave-assisted conversion
of carbohydrates to levulinic acid: An essential step in biomass conversion. Green
Chemistry, 15, 439–445.
Tabasso, S., Montoneri, E., Carnaroglio, D., Caporaso, M., & Cravotto, G. (2014).
Microwave-assisted flash conversion of non-edible polysaccharides and post-
harvest tomato plant waste to levulinic acid. Green Chemistry, 16, 73–76.
Tao, F., Song, H., & Chou, L. (2010). Hydrolysis of cellulose by using catalytic amounts
of FeCl2 in ionic liquids. ChemSusChem, 3, 1298–1303.
Vanoye, L., Fanselow, M., Holbrey, J. D., Atkins, M. P., & Seddon, K. R. (2009). Kinetic
model for the hydrolysis of lignocellulosic biomass in the ionic liquid, 1-ethyl-
3-methyl-imidazolium chloride. Green Chemistry, 11, 390–396.
Vyver, S. V., Thomas, J., Geboers, J., Keyzer, S., Smet, M., Dehaen, W., et al. (2011).
Catalytic production of levulinic acid from cellulose and other biomass-derived
carbohydrates with sulfonated hyperbranched poly(arylene oxindole)s. Energy
and Environmental Science, 4, 3601–3610.
Welton, T. (1999). Room-temperature ionic liquids. Solvents for synthesis and catal-
ysis. Chemical Reviews, 99, 2071–2083.
Werpy, T., & Petersen, G. (2004). Top value-added chemicals from biomass, volume
I: Results of screening for potential candidates from sugars and synthesis gas.
Xiong, Y., Zhang, Z., Wang, X., Liu, B., & Lin, J. (2014). Hydrolysis of cellulose in ionic
liquids catalyzed by a magnetically-recoverable solid acid catalyst. Chemical
Engineering Journal, 235, 349–355.
Zakzeski, J., Grisel, R. J. H., Smit, A. T., & Weckhuysen, B. M. (2012). Solid acid-
catalyzed cellulose hydrolysis monitored by in situ ATR-IR spectroscopy.
ChemSusChem, 5, 430–437.
Zhou, L., Liang, R., Ma, Z., Wu, T., & Wu, Y. (2013). Conversion of cellulose to HMF in
ionic liquid catalyzed by bifunctional ionic liquids. Bioresource Technology, 129,
450–455.
Zuo, Y., Zhang, Y., & Fu, Y. (2014). Catalytic conversion of cellulose into levulinic
acid by sulfonated chloromethyl polystyrene solid acid catalyst. ChemCatChem,
6, 753–757.
Tao, F., Song, H., Yang, J., & Chou, L. (2011). Catalytic hydrolysis of cellulose into
furans in MnCl2-ionic liquid system. Carbohydrate Polymers, 85, 363–368.