Catalytic hydrolysis of cellulose into furans

Article abstract of DOI:10.1134/S0036024416130045

Chromium chloride in 4-(3-methylimidazolium-1-yl)butane-1-sulfonic acid hydrogen sulfate (IL-1) was found to effectively catalyze the hydrolysis of microcrystalline cellulose (MCC) at 150°C for 300 min to achieve 87.8% conversion to a slate of products. With a catalytic amount of CrCl₃, the yields of 5-hydroxymethyl furfural (HMF) and furfural were up to 32.4 and 15.2%, respectively, small molecules levulinic acid (LA, 10.8%) and the total reducing sugars (TRS, 10.7%) were also generated. Through LC-MSD analysis and mass spectra, dimer of furan compounds as the main by-products were speculated, and the components of gas products were methane, ethane, CO, CO₂, and H₂. We suggested that IL-1 and CrCl₃ exhibited a coordination interaction; the formation of the intermediate via the hydride shift played a key role in the formation of HMF. The catalyst was recycled and exhibited constant activity for five successive trials.

Full text of DOI:10.1134/S0036024416130045

ISSN 0036-0244, Russian Journal of Physical Chemistry A, 2016, Vol. 90, No. 13, pp. 2508–2516. © Pleiades Publishing, Ltd., 2016.
CHEMICAL KINETICS
AND CATALYSIS
Catalytic Hydrolysis of Cellulose into Furans1
Chengmei Shi, Furong Tao*, and Yuezhi Cui
Shandong Provincial Key Laboratory of Fine Chemicals, Qilu University of Technology, Jinan 250353, P.R. China
*
e-mail: taofurong17329@126.com
Received October 9, 2015
Abstract—Chromium chloride in 4-(3-methylimidazolium-1-yl)butane-1-sulfonic acid hydrogen sulfate
(
IL-1) was found to effectively catalyze the hydrolysis of microcrystalline cellulose (MCC) at 150°C for
3
00 min to achieve 87.8% conversion to a slate of products. With a catalytic amount of CrCl , the yields of
3
5
-hydroxymethyl furfural (HMF) and furfural were up to 32.4 and 15.2%, respectively, small molecules
levulinic acid (LA, 10.8%) and the total reducing sugars (TRS, 10.7%) were also generated. Through
LC-MSD analysis and mass spectra, dimer of furan compounds as the main by-products were speculated,
and the components of gas products were methane, ethane, CO, CO , and H . We suggested that IL-1 and
2
2
CrCl exhibited a coordination interaction; the formation of the intermediate via the hydride shift played a
3
key role in the formation of HMF. The catalyst was recycled and exhibited constant activity for five successive
trials.
Keywords: microcrystalline cellulose, hydrolysis, HMF, ionic liquids, CrCl3
DOI: 10.1134/S0036024416130045
1. INTRODUCTION
lenge today either from an environmental or an effi-
ciency point of view.
Biomass is an important resource that can be con-
verted to energy [1], chemicals [2], foods and feed-
stocks [3]. As the global petroleum economy draws to
a close, the portfolio of derived products from biomass
will need to be diversified in order to satisfy feedstock
requirements across a range of markets. Cellulose,
which is the most abundant natural polymeric carbon
source in the word, occupies 30–40% of biomass.
However, cellulose is a high molecular polymer of D-
anhydroglucopyranose units joined together in long
chains by β-1,4-glycosidic bonds. The tight hydrogen
bonding network and van der Waals interactions
greatly stabilize cellulose [4], making it notoriously
resistant to hydrolysis. Thus, the development of a
reliable non-fermentative process for the conversion
of cellulose to bio-based products remains important.
Due to the increased interest and demand for
industrial application, more efficient catalytic systems
were recently developed; the utilization of metal ions
in ionic liquids (ILs) was demonstrated to be efficient
for the conversion of cellulose into furans. Wu et al.
[
19] studied the hydrolysis of microcrystalline cellu-
lose, using double-functionalized ionic liquids
Cr([PSMIM]HSO ) as catalyst, BmimCl as solvent,
4
3
the yields of TRS and HMF were up to 94 and 53%,
respectively. Binder et al. [20] showed transformation
of lignocellulosic biomass into furanes in DMA-LiCl
system at 140°C with the yield of HMF up to 48%. In
our previous work [21, 22], we also suggested that
ionic liquids containing metal ions are the promising
systems for cellulose hydrolysis. Based on an early
report describing effective conversion of glucose into
HMF over chromium catalysts [23], herein, we
describe an efficient method for the direct conversion
of microcrystalline cellulose (MCC) into 5-
hydroxymethylfurfural (HMF) and furfural in ionic
Recently, extensive research and development pro-
grams have been initiated worldwide to convert ligno-
cellulosic biomass into valuable products, such as
monosaccharides, ethylene glycol, ethanol, 5-HMF
liquid IL-1 in the presence of CrCl .
3
[
5–9]. Up to now, the processes of cellulose hydrolysis
to chemicals including catalysis by mineral acids [10,
1
1], enzyme-driven reactions [12, 13], the use of sub-
2. EXPERIMENTAL
critical and supercritical water [14, 15] and solid acid
catalysts [16–18]. These processes generally require
harsh conditions and cause several troubles, such as
2.1. Materials
Microcrystalline cellulose (extra pure, 90 μm,
high temperature and pressure, so it is still a great chal- J and K chemical company) was dried for 24 h at 90°C
prior to cellulose hydrolysis; 4-methyl-2-pentanone
(AR, >90%), furfural (CP, >90%) and acetonitrile
1
The article is published in the original.
2508

CATALYTIC HYDROLYSIS OF CELLULOSE INTO FURANS
2509
−
O
O
O
SO3
H2SO4
SO3H
S
N
N
N
N
N
N
−
HSO4
Scheme 1. Synthesis of IL-1.
(
HPLC) were purchased from Tianjin chemical pure ethanol; the aqueous phase and organic phase
reagent company (Tianjin, China); 1-methylimidaz- were recovered by extraction and separation using sep-
ole (CP, >99%), 1,4-butanesultone (>99%), 1-chlo-
robutane (CP, >99.5%), 1-bromobutane (CP,
arating funnel; the gaseous products were recovered by
filling N after reaction; the solid residue recovered by
2
>
99.5%) were from Alfa Aesar and were used as
filtration was dried at 100°C for 8 h, then weighed to
calculate the yield of liquefaction. After filtration,
extraction and separation, organic and water phases
were collected to characterize the products. All results
were replicated at least three times. The range of
experimental errors for the HPLC analysis and TRS
analysis were ±1 and ±2%, respectively.
received; 5-HMF (>99%) was from Aldrich; levulinic
acid (>98%) was from DaoCheng chemical company
(
Hong Kong, China). All other reagents and solvents
were reagent grade and were used as received.
2.2. Preparation of the Ionic Liquids
1
-Methylimidazole (16.4 g, 0.2 mol) and 27.28 g
The weight yields of the products were calculated
based on weight of cellulose. We considered that the
recovered solid residues were all cellulose. Cellulose
conversion was determined by the change of cellulose
(
(
0.2 mol) of 1,4-butanesultone were mixed in a flask
250 mL) and stirred at 42–45°C for 17 h. The
obtained white solid was ground, washed repeatedly
with small amount ether, filtered and vacuum dried for
4
h in the room temperature. Equimolar quantity of weight before and after the reaction. The yield of prod-
the obtained product was added to 98% H SO , and ucts was calculated from the equation: yield (wt %) =
2
4
the mixture was stirred at 80°C for 6 h. The obtained (weight of products)/(initial weight of cellulose) ×
viscous liquid was washed with ether for three times 100%.
and dried in vacuum to form ionic liquid—
4
-(3-methylimidazolium-1-yl)butane-1-sulfonic acid
hydrogen sulfate (IL-1). The synthesis of IL-1 was
2.4. Sugars Analysis
1
indicated in Scheme 1. H-NMR (400 MHz, D O): δ
2
1
.53 (m, 2H), 1.80 (m, 2H), 2.72 (t, 2H), 3.67 (s, 3H),
The yield of the total reducing sugars (TRS) was
4
.02 (t, 2H), 7.24 (d, 2H), 8.47 (s, 1H); ESI-MS: m/z analyzed by phenol-sulfuric acid method [25]. A mix-
(
+)218.6, m/z (–)96.3.
ture containing 0.1 mL reaction sample (water-soluble
Other ionic liquids, such as C mimBF , portion), 0.9 mL deionized water, 1 mL 5% phenol
4
4
C mimPF , C mimBr, C mimCl, C mimH PO , (freshly distilled), and 5 mL 98% concentrated sulfuric
4
6
4
4
4
2
4
were synthesized according to the method reported by acid was prepared. The analysis was performed on a
Dupont et al. [24].
HP8453 UV–Vis spectrophotometer at about 490 nm
with a slit width of 0.06 mm. The concentration of
TRS was calculated based on the standard curve
obtained with glucose. The yield of TRS was calculated
from the equation: yield (wt %) = (initial weight of
TRS)/(initial weight of cellulose) × 100%.
2
.3. Typical Procedures for MCC Hydrolysis in IL-1
Experiments were carried out in a glass-lined stain-
less steel autoclave (100 mL). The reactor was loaded
with 0.5 g of microcrystalline cellulose, 2.0 g of the
catalyst, 1 mL of H O (0.2 M CrCl aq.), and 8 mL of
2
3
4
-methyl-2-pentanone (MIBK). MIBK was used as
2.5 HMF, Furfural and Levulinic Acid Analysis
phase modifier to extract continuously the products
from aqueous phase [21]. An electrical heating jacket
was employed to heat the reactor to the desired tem-
perature. After the appointed reaction time reached,
the reactor was removed from the electrical heating
jacket and quickly quenched in a cool water bath at
atmospheric temperature and pressure. Then MCC
was fractionated into four portions: aqueous phase,
organic phase, gaseous products and solid residue.
The liquid products (HMF, furfural, and levulinic
acid) were analyzed by high-performance liquid chro-
matography (HPLC) on a Waters Alliance 2695 chro-
matograph equipped with Waters 2996 PDA detector
and an Intersil ODS-EP C18 reversed-phase column
(
250 × 4.6 mm, 5 μm) thermostatted at 30°C. The
mobile phase (water-acetonitrile 15 : 85, vol/vol) flow
The liquid and solid fractions were collected by rate was 0.5 mL/min. UV detection was performed at
repeatedly washing the reactor with distilled water and 280 nm for HMF and furfural, and 250 nm for
RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A Vol. 90 No. 13 2016

2510
CHENGMEI SHI et al.
levulinic acid. The sample volume was 10 μL. The Table 1. Hydrolysis of MCC with various catalysts: 0.5 g
MCC, 2.0 g catalysts, 1 mL H O (0.2 M CrCl aq.), 8 mL
concentration of HMF, furfural and levulinic acid
were calculated based on the standard curves obtained
with standard substances.
2
3
MIBK, T = 150°C, t = 300 min, P = 2.5 atm
Yield (wt %)
Conv.
(%)
Entry
Catalysts
levu-
linic TRS
acid
2
.6. Dimeric Furan Compounds Analysis
fur-
fural
HMF
Dimeric furan compounds were analyzed using
high performance ion trap mass spectrometeric detec-
tor (LC-MSD) on an Agilent 1100 series LC-MSD
Trap VL equipped with Agilent 1100 Detector and a
ZORBAX Eclipse Plus C18 reversed-phase column
1
2
None
0
0
0
0
0
1.42 3.90
2.63 6.48
Al2O3
1.98
(
150 × 4.6 mm, 5 μm). During this process, the col-
3a HCl
4a HNO3
5a H PO
10.71 4.53 4.08 25.62 57.43
7.45 4.01 3.54 19.83 44.37
6.93 3.46 2.19 21.37 47.56
12.96 5.02 3.96 24.18 52.18
5.89 1.46 2.04 1.58 14.46
15.43 7.45 2.56 10.24 70.00
umn temperature remained constant at 25°C, while
the mobile phase applied was water-acetonitrile
(
50 : 50, vol/vol) at the flow rate of 1.0 mL/min, with
UV detection at 250 nm [26], and the volume for each
injection was 5 μL. The dry gas was N at the flow of
3
4
6a H SO
2
2
4
5
L/min and dry temperature was 350°C. With the
7
C mimH PO
analysis of mass spectrum, the hit gas was He.
4
2
4
In addition, the characterization of ionic liquids,
NMR spectra were recorded on an INOVA-400 spec-
trometer; ESI-MS analyses were performed using a
Waters micromass ZQ Alliance spectrometer. The gas
fractions were analyzed by mass spectrometry (Infi-
con, Transpector 2). The mass spectrum analysis of
liquid phase was performed on micrOTOF-Q II elec-
trospray ionization mass spectrometer. Molecular
weight distributions of liquid phase products was mea-
sured through gel permeation chromatography
GPC), which was performed on a Waters Alliance
695 chromatograph equipped with Waters 2996 PDA
Detector. During this process, the column tempera-
ture remained constant at 25°C. The mobile phase was
trichloromethane at the flow rate of 1.0 mL/min, with
UV detection at 270 nm.
8
IL-1
9b None
0
0
0
0
0
0
0
0
0
3.07 5.72
1.75 4.13
1.18 3.16
1.76 2.37
0b C mimBF
1
0.92
0.73
0.19
4
4
6
11b C mimPF
4
1
2b C mimBr
4
3b C mimCl
1
1.72 1.35 0.92 2.37 8.72
6.09 4.37 2.96 1.93 19.46
32.41 15.18 10.83 10.73 87.80
4
1
4b C mimH PO
(
2
4 2 4
5b IL-1
1
a
b
With 2.0 g 0.1 M inorganic acid: HCl, HNO , H PO , H SO .
3
3
4
2
4
With 1 mL 0.2 M CrCl aq.
3
(
entries 3–6) were also non-ideal, the conversion of
2
.7. FT-IR Analysis
MCC was only 44–58%, and the main products were
the reducing sugars. When ionic liquids (ILs) were
used as catalysts (entries 7–8), the activity of IL-1
After the catalyzed hydrolysis reaction was com-
pleted, the organic phase was separated and distilled in
vacuum. The low-boiling fraction (25–27°C) was
(
70%) was much better than C mimH PO (14.46%).
4 2 4
studied by Fourier transform infrared spectroscopy Collectively, the results above confirmed that strong
(
FT-IR). All the FT-IR spectra were collected on an acidity of solution promoted the hydrolysis of MCC
FT-IR spectrometer (Nicolet Nexus 870) with a reso- effectively. In addition, the hydrolysis results for
–1
lution of 4 cm and 64 scans in the region of 4000– MCC in different ILs with catalytic amount of CrCl
3
–1
4
00 cm , the test temperature was 25°C and humidity (entries 9–15) were also investigated. As shown in
was 50%.
the table, the conversion of MCC in C mimBF ,
4
4
C mimPF , C mimBr was lower than in the absence
4
6
4
of ionic liquids (entry 9), this is in accordance with
the results of Rogers et al. [27]. Meanwhile, we
referred to 4-(3-methylimidazolium-1-yl)butane-1-
sulfonic acid hydrogen sulfate as IL-1, and other
ionic liquids, which were referred using their chemi-
3
. RESULTS AND DISCUSSION
3
.1. Hydrolysis of MCC with Various Catalysts
The effects of different catalysts on the reactions
are summarized in Table 1. The conversion of MCC
was only 3.9% in the absence of catalysts (entry 1); the cal formula, are ionic liquids themselves. Also, we
presence of Al O (entry 2), having both acidic and suggested that acidity of solution was very important
2
3
basic active sites, increased the conversion to 6.48%. for MCC hydrolysis, the main reason for lower activ-
Also, the reactions catalyzed by protic mineral acid ity of C mimBF , C mimPF , C mimBr was that
4 4 4 6 4
RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A
Vol. 90
No. 13
2016

CATALYTIC HYDROLYSIS OF CELLULOSE INTO FURANS
2511
(
a)
b)
(
1715.67
3500
3000
2500
2000
1500
1000
500
1
Wavenumber, cm−
Fig. 1. The FT-IR spectra of sample catalyzed with HCl (a) and IL-1 (b): 0.5 g MCC, 2.0 g catalysts, 1 mL 0.2 M CrCl aq., 8 mL
3
MIBK, T = 150°C, t = 300 min, P = 2.5 atm.
maybe they could not provide enough acidic sites the
hydrolysis reaction needed. It was obviously that
MCC conversion (hydrolysis) ranging from 70.0 to
3.2. The FT-IR Spectra of Hydrolysis Products
Catalyzed with HCl and IL-1
Figure 1 shows the FT-IR spectra of the products
8
7.8% was achieved when the reaction was catalyzed of MCC hydrolysis catalyzed by HCl (a) and IL-1 (b),
by IL-1 in the same conditions. Based on the fact respectively. There are some obviously different vibra-
tional bands in the two spectra. The absorbance at
that CrCl has almost no catalytic effect in MCC
3
–1
around 3346 and 3122 cm is associated with the
hydroxyl group stretching vibrations; around 2968 and
hydrolysis in the absence of other catalysts, we sug-
gested that IL-1 and CrCl had a synergistic interac-
3
–1
1
368 cm are methyl, methylene and methyne
tion, which could promote the hydrolysis of MCC
efficiently. We considered that IL-1 in the reaction
had double function, not only it could dissolve the
MCC like other ionic liquids, its acidity could
stretching vibrational bands; the absorbance at 1370,
–1
1
282, and 1165 cm correspond to the characteristic
C–O–C stretching vibrational bands; collectively,
these data suggest clearly that sample (a) shows the
weaken the glycosidic bonds through binding with a basic framework of sugars. In addition, it is also clear
glycosidic oxygen atom, then promote the hydrolysis that spectrum (b) shows lots of different bands. The
–1
of MCC effectively.
band at 1715 cm is assigned to C=O stretching vibra-
tions; the absorbance at 1641 cm– is C=C stretching
1
–1
vibrational bands; and 973, 1165, 1208 cm bands
correspond to furane ring [28]. Based on results of
FT-IR spectra, with our present hydrolysis method,
the products of MCC hydrolysis catalyzed by HCl are
mainly the reducing sugars, but the spectra also
showed the presence of furanes in hydrolysis products
when the reaction was catalyzed by IL-1.
Table 2. The influence of CrCl concentration on the
3
hydrolysis of MCC: 0.5 g MCC, 2.0 g IL-1, 1 mL CrCl aq.,
3
8
mL MIBK, T = 150°C, t = 300 min, P = 2.5 atm
Yield (wt %)
C
Conv.
(%)
(
mol/L)
levulinic
acid
TRS
HMF furfural
3
.3. Influence of Different CrCl Concentration
3
None
10.24
11.89
11.06
10.73
9.75
15.43
27.38
29.45
32.41
25.51
24.03
7.45
13.05
14.72
15.18
15.97
13.88
2.56
5.09
5.73
70.00
83.76
85.47
87.80
83.45
79.91
on the Hydrolysis of MCC
0
.05
.1
To confirm the promoting action of CrCl , we
3
0
studied the effect of chromium chloride’s concentra-
tion on the reaction. As shown in Table 2, comparing
0.2
0.3
0.5
10.83
7.06
5.98
with the reaction in the absence of CrCl (70.0%), the
3
addition of CrCl could promote the reaction effec-
3
tively, the conversion of MCC increased by 10–18%.
9.03
However, the concentration of CrCl was higher, the
3
RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A Vol. 90 No. 13 2016

2512
CHENGMEI SHI et al.
3
3
2
5
0
5
35
30
25
20
15
10
5
(
a)
(b)
HMF
HMF
Furfural
20
TRS
Furfural
TRS
1
5
0
5
1
Levulinic acid
Levulinic acid
0
0
100
120
140
160
180
200
100
200
300
400
500
600
Temperature, °C
Time, min
Conv.
9
0
0
0
0
0
0
0
0
0
HMF
Furfural
(
c)
90
(d)
HMF
Furfural
Levulinic acid
8
7
6
5
4
3
2
1
Levulinic acid
Conv.
80
70
60
50
40
30
20
10
0
0
8
: 4
8 : 3
8 : 2
8 : 1
0.5 1.0
2.0
3.0
4.0
V(MIBK) : V(CrCl
3
)
m(IL-1), g
Fig. 2. The effect of hydrolysis conditions on the reaction: 0.5 g MCC, 2.0 g IL-1, 1 mL 0.2 M CrCl aq., 8 mL MIBK, T = 150°C,
3
t = 300 min, P = 2.5 atm. (a) Reaction temperature; (b) reaction time; (c) content of water; (d) dosage of IL-1.
results were not corresponding better. When the CrCl
results; the yield of HMF and furfural reached 32.4
3
concentration was 0.2 M aq., we get the optimum and 15.2%, respectively. From Fig. 2b, after a 600 min
results, the yields of HMF and furfural reached 32.4
and 15.2%, respectively, and small amounts of
levulinic acid (10.8%) and TRS (10.7%) were also
generated. We considered that maybe the coordination
reaction, the yield of TRS decreased obviously, but
oppositely, the yield of levulinic acid increased from
1
.2 to 15.38%. Figure 2a shows similar pattern to the
products, which indicated that with the increasing of
reaction temperature and prolonging of reaction time,
of IL-1 and CrCl played an important role on MCC
3
hydrolysis. To support the connection between IL-1
and CrCl , we performed the dehydration of glucose the dehydration of glucose was favored. These are
3
in 0.2 M H SO . The results showed that in this case consistent with the hydrolysis path of MCC in this sys-
2
4
after the addition of catalytic amount of CrCl , the tem, which is shown in Scheme 2. Glucose was dehy-
3
yields of HMF and furfural increased by 8.3 and 4.5%, drated with formation of HMF, which subsequently is
respectively, which confirmed that the CrCl pro-
3
fragmented into levulinic and formic acids. When the
further decomposition reaction goes on, these com-
pounds are eventually converted to gaseous products.
Also, our results proved that the formation of furfural
was from the loss of formaldehyde in HMF. Mean-
while, Fig. 2a also shows that the hydrolysis results
were slightly better at 180°C compared to 150°C, how-
ever, when the reaction system was carried out above
moted the dehydration of glucose effectively.
3.4. The Effects of Hydrolysis Conditions on the Reaction
Figure 2 summarizes the influences of hydrolysis
conditions on the reaction yields. The temperature
and time dependences of the yields of products for the
CrCl -catalyzed hydrolysis of MCC in IL-1 were
3
1
50°C, an odoriferous compound was generated, it
shown in Figs. 2a, 2b, respectively. As shown in the
figures, when the reaction time was 300 min and the
was black, viscous, fluffy, and did not dissolve in
reaction temperature was 150°C, we get the optimal water. We did not consider that it was the degradation
RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A
Vol. 90
No. 13
2016

CATALYTIC HYDROLYSIS OF CELLULOSE INTO FURANS
2513
OH
OH
OH
OH
O
CrCl3
IL-1
O
O
O
HO
HO
OH
OH
O
OH
MCC
HO
OH
O
n
O
Glucose
OH
+
IL-1 CrCl3
H
OH
O
Levulinic acid Formic acid
O
HO
O
HMF
O
CHO
Furfural
Scheme 2. Schematic illustration of HMF and furfural production from MCC.
products of cellulose or IL-1. At present, we were not exactly structure of the main by-products, however,
sure what it was about the substance, but maybe it is we inferred that the products were furans, and
one subject of future study, so we chose 150°C as an according to the molecular weight, the by-products
optimum treatment temperature.
According to the previous reports [29], water is
generally thought to have a negative effect on the
should have two furan rings at least, so we specu-
lated them to be dimer of furans compounds.
Trace of the gas products were also analyzed using
dehydration of glucose to HMF. Our work also con- mass spectrometry and the compounds detected were
firmed this point, as shown in Fig. 2c, when the ratio methane, ethane, CO, CO , and H for hydrolysis of
2
2
of MIBK to CrCl was 8 : 1, the conversion of MCC
3
MCC at 150°C for 300 min reaction. This meant that
reached 87.8%, much higher than the results of the the long chain hydrocarbons gaseous products were
reaction in which the ratio was 8 : 4 (37.7%) and 8 : 3 not generated, and we suggested MCC did not gasify
(
49.1%), and the selectivity of HMF and furfural were directly under this condition in the reaction system. So
higher too when V(MIBK) : V(CrCl ) was 8 : 1, prob- we speculated that the generation of gas products may
3
ably, the reason was that the increasing ratio of MIBK be due to the further decomposition of the primary
to the aqueous layer decreases side reactions of the reaction products.
HMF product. The dosage of IL-1 was also studied
which was shown in Fig. 2d. When the dosage of IL-1
3
.6. Mechanism of MCC Hydrolyze to HMF
was 3.0 g, we get the optimal results, the yields of
HMF and furfural could reach 36.5 and 17.4%,
respectively. However, considering the cost of reaction
and green engineering, we chose the dosage of IL-1
was 2.0 g.
in Cr-IL-1 System
In most cases, the glycosidic bonds in cellulose
were weakened by Brønsted acid attack. A Lewis acid
would also weaken the glycosidic bonds through bind-
ing with a glycosidic oxygen atom in a similar manner
to protonic acid, leading to the hydrolysis of polysac-
charide to produce monosaccharide. We propose a
3.5. The Analysis of By-products
Except for the main products, small amount of
important by-products were detected. Through gel-
permeation chromatography (GPC), it was found
that molecular weights of the hydrolysis products
were distributed below 2000. Using LC-MSD anal-
ysis, the main protonated molecules with m/z 410.4
and m/z 354.4, were detected. Then the second
mechanism (Fig. 3) in which CrCl in IL-1 forms
3
2n–
[
CrCl (SO ) ]
complexes in a similar manner to
LnCl as reported by Rogers et al. [30]. These com-
3
4 n
3
plexes would promote rapid conversion of the α-ano-
mers of glucose to the β-anomers through hydrogen
bonding between the chloride anions or oxygen atom
2
−
order mass spectra analysis was performed to deter- in SO4 and the hydroxyl groups. Then the ring
mine their structure clearly. For m/z 354.4, the aldoses would reverse to straight chain form, combin-
main fragment peaks were m/z 103.9, 110.8, 125.0, ing with the chromium to form an intermediate on the
1
92.9, 229.1, 254.0, 278.5, 322.2, 337.1 and for m/z way to HMF via the hydride shift [31, 32]. The hydride
4
10.4, the main fragment peaks were m/z 134.9, shift would enable conversion of the aldoses into
2
13.9, 258.9, 284.9, 311.9, 351.9, 366.9. Analyzing ketoses, followed by dehydration to produce HMF.
these two groups fragment peaks, we can not get the Then with the help of acidic solution, the HMF was
RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A Vol. 90 No. 13 2016

2514
CHENGMEI SHI et al.
OH
OH
2n−
OH
(2n −2)−
O
O
O
CrCl2(SO4)n
O
HO
CrCl3(SO4)n]2n−
[
HO
HO
O
H
HO
HO
HO
or
S
O
OH
Cl
OH
O
O
OH
OH
O
α or β
H
2
n−
(2n −2)−
OH
OH
H
H
Cl
O
S
O
O
O
HO
HO
HO
HO
CrCl (SO )
2
4 n
O
O
OH
OH
O
O
CrCl3(SO4)n −1
_
CrCl3(SO4)n]2n−
_
[CrCl3(SO4)n]2n−
[
OH
OH OH OH
O
O
HO
HO
OH
OH
OH OH
H
β or α
Cr3+
+
H
Cr3+
OH
OH
O
OH OH
O
H
H
hydride shift
O
OH OH OH
OH OH
H
_
3H2O
O
HO
O
HMF
-
1
IL
HC
H
O
O
O
OH
+
HCOOH
O
O
Formic acid
Furfural
Levulinic acid
Fig. 3. Putative mechanism of CrCl promoted conversion of glucose into the main products.
3
combined with water to generate levulinic acid or loss
of formaldehyde to produce furfural. In addition, we
conducted the ESI-MS analysis using solution after
3.7. Recycling of IL-1 Catalysts
In the principles of green engineering, catalyst
recycling is always important in metal-catalyzed liq-
uid-phase reactions. Therefore, the reusability of
the reaction, let M = IL-1 + CrCl , the m/z (–) 529.13
3
in the negative ion mode indicated the existence of
CrCl and IL-1 catalysts was tested over five runs.
3
–
[
M + Cl + H O] , which supported the mechanism After the first reaction run for MCC, the catalyst IL-1
2
we put forward above on the other hand.
and CrCl were only existed in aqueous phase, so the
3
RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A
Vol. 90
No. 13
2016

CATALYTIC HYDROLYSIS OF CELLULOSE INTO FURANS
2515
Yield, wt %
HMF
90
80
70
60
50
40
30
20
10
0
Furfural
Levulinic acid
Conversion
1
2
3
4
5
Runs no.
Fig. 4. Recycling of IL-1 and CrCl for the hydrolysis of MCC: 0.5 g MCC, 2.0 g IL-1, 1 mL 0.2 M CrCl aq. 8 mL MIBK, T =
3
3
,
150°C, t = 300 min, P = 2.5 atm.
organic phase was separated. Then 1.5 g of distilled taneous pressure with a catalytic amount of 0.2 M
water was added into the reaction mixture to decrease CrCl aq. in the reaction catalyzed by IL-1. HMF, fur-
3
the viscosity of ionic liquid and facilitate the fural and LA yields were up to 32.4, 15.2, and 10.8%,
extraction of main products. The HMF and furfural respectively, small amount of TRS (10.7%) were also
were then separated from aqueous phase by extracting generated. With the increasing of reaction temperature
8
times with 5 mL of ethyl acetate. After extraction, and prolonging of reaction time, the dehydration of
CrCl was still in the aqueous phase, then the solution glucose proceed more thoroughly, the yield of TRS
3
was heated at 60°C in a vacuum drier until the water decreased but levulinic acid yield increased. Our
and residual ethyl acetate were removed absolutely. results also confirmed that water had a negative effect
The dried IL-1 catalyst with CrCl was used directly in on the dehydration of TRS to HMF. Using FT-IR
3
the next run by adding fresh feedstock under the same analysis, the formation of furans compounds were
reaction conditions. It can be seen from Fig. 4, confirmed in our reaction system. Collectively, our
although a decrease of the conversion from 88.2 to method provided a new opportunity for transforma-
7
4.6% was observed in the second run, no further tion of cellulose through a non-fermentative process
decrease in activity was observed in the subsequent into bio-based products. However, a detail mecha-
runs, that is to say, an average MCC conversion of nism involved in the cooperation of IL-1 and CrCl
3
about 72.5% could be obtained. Also, the yields of and the separation, purification of the products are
HMF, furfural and levulinic acid decrease only slightly remaining subjects of future study.
over five repeated runs. The decrease of activity in the
second run may be attributed to non-completely
ACKNOWLEDGMENTS
extraction of little by-products from the liquid-phase,
which may be restrained the conversion of MCC and
the further dehydration of glucose. How to improve
the reusability of ionic liquids remained a problem for
future study.
The author gratefully acknowledge the promotive
research fund for young and middle-aged scientists of
Shandong province (BS2014NJ012), the financial
support of the National Natural Science Foundation
of China (nos. 21276149 and 21406121) and the Pro-
gram for Scientific Research Innovation Team in Col-
leges and Universities of Shandong Province.
4
. CONCLUSION
In summary, we demonstrated that CrCl and IL-1
3
system was an excellent combination for the conver-
sion of MCC into small molecular weight compounds
REFERENCES
(
HMF, furfural, LA). By comparing different acidic
1. Nur A. S. Ramli and Nor A. S. Amin, Fuel Proces.
catalysts, we indicated that acidic ionic liquids had
higher activity than protonic acid, and IL-1 showed
higher catalytic activity among different ILs. High
conversion, usually 87–88%, was realized under spon-
Technol. 128, 490 (2014).
2
. F. Schüth, R. Rinaldi, N. Meine, M. Käldström,
J. Hilgert, and M. D. Kaufman Rechulski, Catal.
Today 234, 24 (2014).
RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A Vol. 90 No. 13 2016

2516
CHENGMEI SHI et al.
3
4
5
. Q. Lu, X. Ye, Z. Zhang, C. Dong, and Y. Zhang, 18. S. Shen, C. Wang, Y. Han, B. Cai, and H. Li, Fuel 134,
Biores. Technol. 171, 10 (2014).
573 (2014).
. X. Chen, J. Chen, T. You, K. Wang, and F. Xu, Carbo- 19. L. Zhou, R. Liang, Z. Ma, T. Wu, and Y. Wu, Biores.
hydr. Polym. 125, 85 (2015).
Technol. 129, 450 (2013).
. K. Barbera, P. Lanzafame, A. Pistone, S. Millesi,
G. Malandrino, A. Gulino, S. Perathoner, and
G. Centi, J. Catal. 323, 19 (2015).
2
0. J. B. Binder and R. T. Raines, J. Am. Chem. Soc. 131,
1979 (2009).
2
2
2
1. F. Tao, H. Song, and L. Chou, Biores. Technol. 102,
6
7
8
9
. F. Deng, L. Fu, and M. Ma, Carbohydr. Polym. 121,
9
000 (2011).
163 (2015).
2. F. Tao, H. Song, and L. Chou, J. Mol. Catal. A: Chem.
57, 11 (2012).
3. H. B. Zhao, J. E. Holladay, H. Brown, and
. G. Shen, X. Zhang, Y. Shen, S. Zhang, and L. Fang,
3
Anal. Biochem. 471, 38 (2015).
. S. Xiao, B. Liu, Y. Wang, Z. Fang, and Z. Zhang,
Z. C. Zhang, Science 316, 1597 (2007).
Biores. Technol. 151, 361 (2014).
. K. Zhuo, Q. Du, G. Bai, C. Wang, Y. Chen, and 24. J. Dupont, C. S. Consorti, P. A. Z. Suarez, and
J. Wang, Carbohydr. Polym. 115, 49 (2015).
0. S. M. Shaik, P. N. Sharratt, and R. B. H. Tan, J. Anal. 25. Q. Dong, L. Y. Zheng, and J. N. Fang, Chin. Pham. J.
Appl. Pyrolys. 104, 234 (2013). 31, 550 (1996).
R. F. Souza, Org. Synth. 10, 184 (2004).
1
1
1. Z. Zhang, B. Liu, and Z. Zhao, Polym. Degrad. Stabil. 26. E. Ferrer, A. Alegríaa, R. Farré, P. Abellán, and
9
7, 573 (2012).
2. M. A. Haque, K. M. Cho, D. N. Barman, M. K. Kim,
and H. D. Yun, Process Biochem. 50, 807 (2015).
F. J. Romero, J. Chromatogr. A 947, 85 (2002).
1
2
2
2
3
7. R. P. Swatloski, S. K. Spear, J. D. Holbrey, and
R. D. Rogers, J. Am. Chem. Soc. 124, 4974 (2002).
13. S. K. Wahono, V. T. Rosyida, C. Darsih, D. Pratiwi,
8. L. H. Yao, C. A. Su, L. Q. Qi, C. J. Liu, Y. M. Hu, and
A. Frediansyah, and Hernawan, Energy Proc. 65, 331
H. Zhao, Spectrosc. Spectr. Anal. 1, 32 (1999).
(
2015).
9. X. H. Qi, M. Watanabe, T. M. Aida, and R. L. Smith,
14. N. Ding, R. Azargohar, A. K. Dalai, and J. A. Kozinski,
Green Chem. 10, 799 (2008).
Fuel 118, 416 (2014).
0. C. C. Hines, D. B. Cordes, S. T. Griffin, S. I. Watts,
15. L. K. Tolonen, M. Juvonen, K. Niemelä, A. Mikkel-
V. A. Cocalia, and R. D. Rogers, New J. Chem. 32, 872
son, M. Tenkanen, and H. Sixta, Carbohydr. Res. 401,
(2008).
16 (2015).
1
6. H. Lei, L. Lin, Z. Wu, S. Zhou, and S. Liu, Appl. Catal. 31. J. B. Binder, A. V. Cefali, J. J. Blank, and R. T. Raines,
B: Environ. 174–175, 225 (2015).
Energy Environ. Sci. 3, 765 (2010).
7. X. Wei, Z. Wang, Y. Wu, Z. Yu, J. Jin, and K. Wu, 32. R. W. Nagorski and J. P. Richard, J. Am. Chem. Soc.
J. Anal. Appl. Pyrolys. 107, 150 (2014). 123, 794 (2001).
1
RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A
Vol. 90
No. 13
2016