Entrainer-intensified vacuum reactive distillation process for the separation of 5-hydroxylmethylfurfural from the dehydration of carbohydrates catalyzed by a metal salt-ionic liquid

Article abstract of DOI:10.1039/c2gc16671b

As more and more novel catalyst systems are being developed for the dehydration of carbohydrates, especially glucose, an effective way to separate the dehydration product 5-hydroxylmethylfurfural (5-HMF) is also required for industrial manufacturing. In this paper, for the first time, we have developed a process called EIVRD (entrainer-intensified vacuum reactive distillation) to separate 5-HMF from the dehydration solutions of carbohydrates catalyzed by a metal chloride/1-methyl-3-octyl imidazolium chloride ([OMIM]Cl) ionic liquid, in which high vacuity and entrainers were applied to intensify the distillation of 5-HMF as well as the dehydration of fructose or glucose. In such an EIVRD process, the average recoveries of 5-HMF dehydrated from fructose and glucose are around 93% and 88%, respectively. The recycling of the catalyst system in the EIVRD process is so convenient that the recovery and actual yield of 5-HMF is successfully repeated during the whole five recycled reactions.

Full text of DOI:10.1039/c2gc16671b

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Green Chemistry
Cite this: Green Chem., 2012, 14, 1220
www.rsc.org/greenchem
PAPER
Entrainer-intensified vacuum reactive distillation process for the separation of
-hydroxylmethylfurfural from the dehydration of carbohydrates catalyzed
by a metal salt–ionic liquid†
5
a
b
a
a
Zuojun Wei,* Yingxin Liu, Dilantha Thushara and Qilong Ren
Received 26th December 2011, Accepted 30th January 2012
DOI: 10.1039/c2gc16671b
As more and more novel catalyst systems are being developed for the dehydration of carbohydrates,
especially glucose, an effective way to separate the dehydration product 5-hydroxylmethylfurfural
(5-HMF) is also required for industrial manufacturing. In this paper, for the first time, we have developed
a process called EIVRD (entrainer-intensified vacuum reactive distillation) to separate 5-HMF from the
dehydration solutions of carbohydrates catalyzed by a metal chloride/1-methyl-3-octyl imidazolium
chloride ([OMIM]Cl) ionic liquid, in which high vacuity and entrainers were applied to intensify the
distillation of 5-HMF as well as the dehydration of fructose or glucose. In such an EIVRD process, the
average recoveries of 5-HMF dehydrated from fructose and glucose are around 93% and 88%,
respectively. The recycling of the catalyst system in the EIVRD process is so convenient that the recovery
and actual yield of 5-HMF is successfully repeated during the whole five recycled reactions.
Introduction
as acetonitrile, dimethylsulphoxide (DMSO), N,N-dimethylfor-
5,24,26,31,32
mamide (DMF), and N,N-dimethylacetamide (DMA)),
organic–water mixtures (such as water–acetone, water–butanol
In recent years, the efficient utilization of biomass has received
increasing attention as a promising alternative to petroleum for
21,25,32
28–30
and water–DMSO mixtures)
as well as ionic liquids
1
–4
the production of fuels and chemicals. Among the many poss-
ible biomass-derived chemicals, 5-hydroxylmethylfurfural (5-
HMF) prepared from the dehydration of carbohydrates is con-
sidered to be a key bio-based platform compound, as it can be
converted to 2,5-furandicarboxyl acid, 2,5-dihydroxymethyl-
furan, 2,5-bis(hydroxymethyl) tetrahydrofuran, dimethylfuran,
and other liquid alkanes by oxidation, hydrogenation, hydroge-
have been mainly studied for carbohydrate dehydration. From
the ecological and technological point of view, aqueous pro-
cesses are the most favored but, unfortunately, are inefficient,
because 5-HMF can further react with water to form levulinic
17,33,34
acid and formic acid.
Whereas in an ionic liquid medium,
the formed amounts of levulinic acid and formic acid are very
low, thus make it possible to obtain a high selectivity to 5-HMF.
Moreover, the ionic liquid is likely to serve as a co-catalyst,
especially in the dehydration of glucose which is catalyzed by
1
,2,5–10
nolysis or aldol condensation.
Being the basic building units of the carbohydrates such as
sucrose, inulin, cellulose and other polysaccharoses, fructose and
glucose are always applied as model substrates for the dehy-
dration of carbohydrates into 5-HMF. A lot of catalysts, includ-
30,35
CrCl and SnCl .
By far, many ionic liquids, especially imi-
2
4
dazolium-based ionic liquids have been demonstrated to be the
most effective catalyst–solvent system for the conversion of fruc-
11–14
24,30
ing mineral acids (such as HCl, H SO and H PO ),
organic
2
4
3
4
tose and glucose into 5-HMF,
owing to their outstanding
liquid acids (such as oxalic acid, levulinic acid and maleic
characteristic properties such as low vapor pressure, good
thermal stability and a range of tunable hydrophobicity/hydro-
philicity. For example, a record yield of 5-HMF as high as 92%
from fructose is achieved within 15–45 minutes at 90 °C using
1
5–17
acid),
metal salts and metal oxides (such as γ-TiP, C-ZrP O,
2
1
8–23
TiO , HTiNbO –MgO, NbOPO and ZrP),
ion exchange
2
5
4
1
2,24,25
12,16,26,27
28–30
resins,
zeolites,
as well as ionic liquids
have
been applied for the dehydration of carbohydrates.
Additionally, the selection of a suitable reaction solvent is of
1
-H-3-methylimidazolium chloride ([HMIM]Cl) as the catalyst–
36
solvent system. Compared with the dehydration of fructose, the
dehydration of glucose is rather difficult. However, such a
problem has been overcome by Zhang’s group, who have
obtained a high 5-HMF yield of ca. 80% from glucose for the
1
2,18,19,22,23
the same importance. Water,
organic solvents (such
a
Department of Chemical and Biological Engineering, Institute of
Pharmaceutical Engineering, Zhejiang University, Hangzhou 310027,
P.R. China. E-mail: weizuojun@zju.edu.cn; Tel: +86-13588810769
first time using CrCl –1-ethyl-3-methyl-imidazolium chloride
2
30,37
(
[EMIM]Cl) as the catalyst–solvent system.
even higher yield of 5-HMF is obtained from glucose using a N-
heterocyclic carbene complexed with CrCl /3-octyl-1-methyl
Afterwards, an
b
College of Pharmaceutical Science, Zhejiang University of Technology,
Hangzhou 310032, P.R. China
†
1
Electronic supplementary information (ESI) available. See DOI:
0.1039/c2gc16671b
2
38
imidazolium chloride ([OMIM]Cl) as a catalyst/solvent. This
1220 | Green Chem., 2012, 14, 1220–1226
This journal is © The Royal Society of Chemistry 2012

View Article Online
new methodology enables high 5-HMF yields of 96% and 81%
form fructose and glucose, respectively. Hu et al. have found
that SnCl4 in 1-ethyl-3-methylimidazolium tetrafluoroborate
reactivity for the dehydration of carbohydrates. However, in
order to avoid being distilled out together with 5-HMF under
vacuum, ionic liquids having longer carbon side chains are more
suitable to the distillation process. Based on the thermodynamic
3
5
(
[EMIM]BF ) is also effective for the conversion of glucose and
4
11,30,39
45
glucose unit-containing carbohydrates.
Although the cata-
formula proposed by Valderrama and Robles, the boiling point
lytic mechanism remains elusive, some insights on the reaction
indicate that the special ionic liquid forms complexes with the
metal chloride to promote rapid mutarotation of α-glucose to
β-glucose, and leads to isomerization of glucose to fructose,
of [OMIM]Cl and [BMIM]Cl are estimated as 638.9 K and
548.9 K, respectively. Considering that the boiling point of 5-
HMF is ca. 564 K, it is rational to choose [OMIM]Cl as the
reaction medium for the dehydration of carbohydrates in our fol-
lowing distillation process.
6
30,35,40
while the latter can be much easier transformed to 5-HMF.
In spite of many advantages in the dehydration of carbo-
hydrates, the use of ionic liquids for the production of 5-HMF in
industry has been impracticable up to now. The main problems
are the separation of the products as well as the recycling of the
ionic liquid-containing reaction medium. Similar to the conven-
tionally used solvent such as DMSO and DMF, ionic liquids
always have very high boiling points, however, 5-HMF is ther-
mally sensitive. Therefore, it is neither economical nor techni-
cally feasible to separate the products out by simple solvent
distillation. Alternatively, the extraction of 5-HMF from ionic
Experimental
Materials
Anhydrous [OMIM]Cl was synthesized from n-octyl chloride
and 1-methylimidazole according to the literature. Its structure
46
1
was confirmed by FT-IR and H-NMR analysis, the same as our
47
previous study. Fructose (purity: 99%) and glucose (purity:
9%) were bought from Sino-pharm chemical reagent Co. Ltd.
Standard 5-HMF was bought from Tengzhou Runlong Flavours
Fragrances Co. Ltd (>99.9%). Iridium trichloride was bought
from Shanxi Kaida Chemical Engineering Co. Ltd (molecular
formula: IrCl ·(1–2)H O, iridium content ≥58.3%). CrCl ·6H O
2
9
35–38
liquids using organic solvents has been frequently reported.
&
Authors claimed that such a solvent extraction method can
recover more than 90% of 5-HMF and the ionic liquid can be
3
6
3
2
3
reused more than 5 times. For example, Moreau et al. reported
that 5-HMF can be completely extracted in a continuous manner
with diethyl ether (50 ml for 8 h for 0.5 g of 5-HMF in 2 g of
ionic liquid) or in a stepwise manner (6 × 150 ml for 1.2 g of 5-
HMF in 2 g of ionic liquid), and the yield of 5-HMF decreases
and the organic solvents were all bought from Sino-pharm
chemical reagent Co. Ltd.
A sliding-vane rotary vacuum pump (2XZ-2) was bought
from Zhejiang Huangyan Qiujing Vacuum pump Co. Ltd. The
vacuum pressure sensor (Model 760) for vacuity measurement
was bought from Setra Systems, Inc.
37
little, while the ionic liquid is reused for five times. Yu et al.
used methyl isobutyl ketone (MIBK) to recover 5-HMF from
CuCl /CrCl /[EMIM]Cl at 100 °C. Three successive 8 h runs
2
2
were conducted, and more than 90% of 5-HMF is recovered. Hu
et al. conducted the dehydration of fructose in ethyl acetate/
Typical work-up procedure
5
choline chloride (ChoCl)-based ionic liquid to promote the yield
of 5-HMF by continuously extracting it from the ionic liquid
phase. These solvent extraction processes are both time-consum-
ing and require large amounts of solvent, which prevents their
industrial use. Besides, the applied ionic liquids are always
viscous and miscible with 5-HMF, therefore it is not as easy to
extract 5-HMF as expected; the researchers themselves claim that
such a method “is difficult, at the present time, to be transposed
The reaction set-up is shown in Fig. 1. Dehydration of the carbo-
hydrates was carried out in a 100 ml magnetically stirred flask at
36
on a pilot scale”. So far, exploring an effective separation
method is much more urgent in order to industrially manufacture
5
-HMF from the dehydration of carbohydrates in ionic liquids.
Reactive distillation is a combination of chemical reaction and
product separation by a distillation process within a single unit,
4
1–43
and it has a wide range of industrial applications.
Inspired
by our previous study on the Fischer esterification reactions cata-
lyzed by ionic liquid in which we have easily separated the ester
44
products and recovered the ionic liquid by distillation process,
the objective of the present study is focused on the reactive dis-
tillation process for the separation of 5-HMF from the metal
chloride–ionic liquid catalyst system. Considering that 5-HMF
has a much higher boiling point (114–116 °C, 133.3 Pa), two
conventional chemical engineering strategies have been applied:
Fig. 1 Experimental set-up for the vacuum reactive distillation of the
fructose and glucose dehydrations. Codes: 1) oil bath heater; 2) magnetic
rotator; 3) two-neck flask for vapor generation; 4) nitrogen cylinder; 5)
ring-type gas distributor; 6) three-neck flask; 7) connector with heat
keeping jacket; 8) to vacuum oil pump; 9) two-neck flask for product
collection; 10) ice bath.
(1) a vacuum pump is added to accelerate the evaporation of 5-
HMF; (2) nitrogen and other organic vapours are used as entrai-
ners to increase the evaporation efficiency. As reported, imidazo-
lium ionic liquids with shorter carbon side chains show higher
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1220–1226 | 1221

View Article Online
different temperatures (Fig. 1, code: 6). In a typical experiment,
% (mol ratio based on fructose or glucose) of CrCl or IrCl3
necessary to further study the thermal stability of 5-HMF at
those higher temperatures first.
7
3
and 20 g of [OMIM]Cl were successively added into the flask
Fig. 1, code: 6) and heated to the predetermined temperature by
As shown in Fig. 2, 5-HMF is quite thermally unstable at
120 °C in [OMIM]Cl when no metal salt is added, with only
68% recovery after heating for 3 hours. In contrast, it tends to be
(
an oil bath. Then, fructose or glucose was added to initiate the
reaction. A vacuum pump and an accurate pressure sensor were
used to depressurize the system and record the value. The entrai-
ner was introduced to the bottom of the flask (Fig. 1, code: 6)
from the left inlet with a ring-type gas distributor (Fig. 1, code:
much more stable in the presence of either 7 mol% of IrCl or
3
CrCl , in which only 1–3% degradation is found. Unfortunately,
3
when the temperature increases, 5-HMF starts to degrade again
despite the metal chloride being present. For IrCl –[OMIM]Cl,
3
5
). A nitrogen cylinder with a stable outlet pressure of 0.1 MPa
about 85% and 75% of 5-HMF are recovered after 3 hours of
heating at 150 °C and 180 °C, respectively. 5-HMF appears to
was used as a nitrogen entrainer supplier. While hexane or
MIBK were used as entrainers, ca. 150 ml of the corresponding
liquid solvent was loaded in a 500 ml flask (Fig. 1, code: 3) and
heated to its ambient boiling point (69 °C for hexane and 118 °C
for MIBK) to generate saturated vapor with 0.1 MPa pressure,
and was then introduced to the gas distributor (Fig. 1, code: 5).
The flow rate of the gas is controlled by the working power of
the vacuum pump. The generated 5-HMF was vaporized and
entrained out, and settled down on the inner surface of the con-
nector (Fig. 1, code: 7) and flask (Fig. 1, code: 9). Two samples
were analyzed for carbohydrates and 5-HMF after the reaction
was terminated: one is the 5-HMF in the ionic liquid reaction
solution in flask (Fig. 1, code: 6) and the other one is the 5-HMF
powder in connector (Fig. 1, code: 7) and flask (Fig. 1, code: 9)
after rinsing with pure water.
Analysis
Carbohydrates were analyzed by HPLC (Detector: Waters 410
®
Differential Refractometer; Aminex HPX-87H column 9 μm,
−
1
3
00 × 7.8 mm), using 5 mmol l H SO in ultrapure water as
2 4
−1
mobile phase at a flow rate of 0.6 ml min and a column temp-
erature of 60 °C. 5-HMF was analyzed by HPLC (UV Detector:
TM
Waters 2487; wavelength: 284 nm; Sunfire C18 column 5 μm;
2
50 × 4.6 mm), using 70% methanol in ultrapure water as
−1
mobile phase at a flow rate of 1 ml min and a column tempera-
ture of 35 °C.
Recovery of 5-HMF was defined as the molar ratio of the 5-
HMF collected in the connector (Fig. 1, code: 5) and the flask
(Fig. 1, code: 6) to the total 5-HMF that added or could theoreti-
cally produced from. The yield of 5-HMF was defined as the
molar ratio of the total amount of formed 5-HMF to the carbo-
hydrates added.
Results and discussion
Thermal stability of 5-HMF in IrCl –[OMIM]Cl and
3
3
CrCl –[OMIM]Cl
IrCl and CrCl were adopted as catalysts respectively for the
3
3
dehydration of fructose and glucose in [OMIM]Cl, as they have
been demonstrated to be the most effective metal salts for the
4
4,47
same catalytic reactions in our previous study,
as well as in
Fig. 2 Thermal stability of 5-HMF in IrCl
OMIM]Cl (b) catalyst systems at 100 °C (◯), 120 °C (□), 150 °C (□)
and 180 °C (4). × is 5-HMF in only [OMIM]Cl at 120 °C for compari-
son. Reaction conditions: 2.0 g of [OMIM]Cl and 7 mol% of IrCl or
3 3
–[OMIM]Cl (a) and CrCl –
30,48
other reports. IrCl3 and CrCl3 are regarded as singularly
thermal stabilizers for 5-HMF at 100–120 °C,
[
3
0,47
which may
inhibit the unwanted degradation of 5-HMF during the following
vacuum distillation process. Since the vacuum reactive distilla-
tion experiment will be carried out at 150 °C or 180 °C, it is
3
CrCl (or added only [OMIM]Cl for comparison) were first heated to the
3
predetermined temperature, and then 140 mg of 5-HMF was added.
1222 | Green Chem., 2012, 14, 1220–1226
This journal is © The Royal Society of Chemistry 2012

View Article Online
Table 1 Vacuum distillation of 5-HMF from IrCl
3
–[OMIM]Cl without entrainer
Mass of 5-HMF (mg)
Entry
Temp. (°C)
120
Time (min)
Residue in IL phase
Distilled out
Recovery (%)
Vacuity (Pa)
20–50
1
2
3
4
5
6
15
30
15
30
15
30
<5
<5
334
217
255
135
195
40.4
ND
ND
152
271
224
343
277
431
195
487
30.4
54.2
44.7
68.5
55.4
86.1
97.5
97.4
150
90–120
180
150–180
a
7
120
120
130–150
130–150
b
8
3 2
Operation conditions: 0.5 g 5-HMF and 0.065 g IrCl ·(1–2)H O (ca. 7 mol% based on 5-HMF) dissolved in 20 g [OMIM]Cl. The vacuity is the as-
a b
shown value while the vacuum pump was running at its full capacity. 0.2 g 5-HMF. 0.5 g 5-HMF heated with the vacuum pump running at its full
capacity; ND is the abbreviation of not detected.
be more stable in CrCl –[OMIM]Cl than in IrCl –[OMIM]Cl, in
3
Table 2 Vacuum distillation of 5-HMF from IrCl /[OMIM]Cl with
nitrogen as an entrainer
3
3
which nearly 90% of 5-HMF can be recovered in the former
when heated at 180 °C for 3 hours. It can be seen from Fig. 2,
that the degradation of 5-HMF can be controlled within 5%, in
the range of experimental temperatures, if the heating time is less
than 30 minutes. These results give a guide to estimate the degra-
dation amount of 5-HMF during the following vacuum distilla-
tion process.
Mass of 5-HMF (mg)
Temp.
Entry (°C)
Time
(min)
Residue in
IL phase
Distillated
out
Recovery
(%)
1
2
3
4
5
6
7
8
9
120
150
180
5
287
167
66
204
55.7
8.5
90.5
5.5
2.5
203
322
428
288
438
478
403
474
469
40.5
64.2
85.5
57.5
87.5
95.5
80.5
94.7
93.8
10
30
5
10
30
5
Vacuum distillation of pure 5-HMF from IrCl –[OMIM]Cl
3
In the first stage, pure 5-HMF, instead of the carbohydrate, was
10
30
introduced into the IrCl –[OMIM]Cl system, in order to test the
3
feasibility of the distillation process. Since the saturated vapor
pressure of 5-HMF is rather low (e.g. 133.3 Pa at 114–116 °C),
and the fraction of 5-HMF in [OMIM]Cl is always less than
Operation conditions: 0.5 g of 5-HMF and 0.065 g of IrCl
3 2
·(1-2)H O
(
ca 7 mol% based on 5-HMF) were dissolved into 20 g of [OMIM]Cl.
Nitrogen was introduced from the left neck of the three-neck flask (code
4 in Fig. 1). The vacuity is controlled at ca. 300 Pa.
1
0% in the dehydration of carbohydrate, leading to even lower
partial vapor pressure in the gas phase, it is therefore necessary
to add a vacuum pump to intensify the distillation. Compared
with the distillation of pure 5-HMF, extra forces are needed to be
overcome besides the intermolecular force of 5-HMF for suc-
water steam or nitrogen to remove the substances with lower
boiling points from triglycerides such as long-chain fatty acids,
cessful distillation of 5-HMF from the IrCl –[OMIM]Cl system,
3
49–52
i.e. the interaction forces between 5-HMF and [OMIM]Cl, as
steroids and odor components.
Principally, the addition of
well as those between 5-HMF and IrCl . As a result, the distilla-
tion process should be much more difficult than that of pure 5-
HMF.
an entrainer can increase the gas–liquid contact area, which will
enhance the mass transfer of the target component from the
liquid phase to the gas phase. Besides, the affinity from the
entrainer may also accelerate the distillation of the target com-
ponent. As a result, a higher recovery of the target component
could be obtained in the gas phase at milder operation con-
ditions. As shown in Table 2, with the addition of nitrogen as an
3
As expected, only 54.2% of 5-HMF is distilled from the
IrCl –[OMIM]Cl system at 120 °C in 30 minutes when 0.2 g of
3
5
1
-HMF was boiled and evaporated rapidly in a few minutes at
20 °C at a vacuity of 130–150 Pa (Table 1, entries 2 and 7).
Higher recoveries of 5-HMF are obtained at higher temperatures:
i.e. 68.5% at 150 °C and 86.1% at 180 °C in 30 minutes, as
increasing the temperature can raise the partial vapor pressure of
entrainer, the recoveries of 5-HMF from IrCl –[OMIM]Cl
3
are remarkably higher than those when no entrainer was
introduced (Table 1). Furthermore, a higher recovery of 5-HMF
can be obtained at a lower temperature with a longer degassing
time and at a higher temperature with a lower degassing time.
For example, 95.5% of 5-HMF is recovered at 150 °C in
30 minutes with nitrogen as an entrainer (Table 2, entry 6) while
a similar higher value, 94.7% of 5-HMF recovery is obtained at
180 °C in 10 minutes with the same entrainer (Table 2, entry 8).
An excessive distillation time of 30 minutes is unnecessary for
improving the recovery (Table 2, entry 9), indicating that a gas–
liquid equilibrium of 5-HMF has already been reached in
5
-HMF. These recoveries are fairly satisfying. However, in con-
sideration of the generally less than 50% yield of 5-HMF from
the dehydration of glucose, the final recovery of 5-HMF would
be even less by the reactive distillation process. Therefore, in
order to obtain a satisfactory recovery of 5-HMF or to operate
the process at milder conditions, we attempted to add an entrai-
ner to improve the efficiency of the vacuum distillation process.
The idea of adding entrainers originated from the classic
refinery industry of edible oil, which uses high-temperature
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1220–1226 | 1223

View Article Online
Table 3 Vacuum Reactive distillation of 5-HMF from the dehydration of carbohydrates with different entrainers
Entry
Carbohydrate
Frustose
Entrainer
Temp. (°C)
Time (min)
Recovery (%)
Yield (%)
1
2
3
4
5
6
7
8
9
N
2
180
150
180
150
180
150
180
150
120
180
150
180
150
180
150
180
150
100
10
30
10
30
10
30
10
30
20
10
30
10
30
10
30
10
30
180
90.6
89.4
94.6
93.6
95.0
93.6
—
94.4
93.2
93.8
93.7
95.1
94.6
92.5
90.8
89.0
68.5
65.2
70.4
64.8
71.8
66.2
51.3
42.1
35.1
Hexane
MIBK
a
—
—
—
a
—
—
a
1
0
Glucose
N
2
88.3
85.3
87.5
88.6
88.2
88.4
—
1
1
12
13
14
15
16
17
18
Hexane
MIBK
a
a
a
—
—
—
—
—
3 2 3 2
Operation conditions: 2 g of fructose/glucose and 0.065 g of IrCl ·(1-2)H O/0.049 g of CrCl ·6H O (ca 7 mol% based on fructose) were dissolved
into 20 g of [OMIM]Cl. Nitrogen or organic vapor was introduced from the left neck of the three-neck flask (code 4 in Fig. 1). The vacuity is
a
controlled at ca. 300 Pa. Reaction carried out without vacuum distillation.
1
0 minutes, and further degassing time leads to the decompo-
based on our results (Table 3). However, compared with the
dehydration reactions without vacuum distillation, the yields of
5-HMF from glucose in our EIVRD process are greatly increased
by more than 20% (e.g. Table 3, entries 11 vs. 17). This could be
explained that the EIVRD process removes 5-HMF out from the
reaction bulk soon after it generates, which (1) avoids further
decomposition of the produced 5-HMF; (2) avoids the cross-
reactions between the reactive aldehyde groups in 5-HMF, fruc-
sition of 5-HMF.
These results demonstrate that the distillation of 5-HMF from
the metal salt–[OMIM]Cl system under vacuum is feasible, and
the favorable operation conditions are at 150 °C for 30 minutes
and 180 °C for 10 minutes.
43
tose and glucose which was mentioned in our previous work,
Entrainer-intensified vacuum reactive distillation (EIVRD) for
and (3) depletes fructose much more rapidly and therefore
pushes the right shift of the isomerization equilibrium reaction of
5
-HMF
39
As vacuum distillation with an entrainer is effective in the recov-
ery of 5-HMF, as mentioned above, the next step was to apply
the EIVRD process to the dehydration reactions of the carbo-
hydrates in the metal chloride–[OMIM]Cl system. For either
fructose or glucose, three entrainers, i.e. nitrogen, hexane, and
MIBK, were tried. The results are shown in Table 3.
glucose to fructose. Only a few types of catalyst (base or
53
enzyme ), which promote the isomerization of glucose to fruc-
tose and are effective in the dehydration of glucose to 5-HMF
have been reported, so it is exciting that we have reached the
goal by the simple entrainer-intensified vacuum reactive distilla-
tion process.
As can be seen from Table 3, the EIVRD process is quite
effective for the recovery of 5-HMF from the dehydration reac-
tion bulk. Within the experimental range, the averaged recoveries
of 5-HMF dehydrated from fructose and glucose are around
When fructose was used as a substrate, the yield of 5-HMF
can be as high as 95% with our EIVRD process.
As shown in Table 3, the type of the entrainers used in this
work has little effect on the recovery and the yield of 5-HMF.
We therefore deduce that the distillation of 5-HMF in our case, is
controlled by the surface area of the liquid–gas interface while
the entrainer is bubbled in, rather than the affinity force between
the entrainer and 5-HMF. In all of the entries in Table 3, the con-
version of fructose or glucose is close to 100%.
9
3% and 88%, respectively. These results are very exciting as
they report, for the first time, the separation of the 5-HMF
product from the reaction bulk with high recoveries. Use of a
high temperature (180 °C) can increase the dehydration rate and
the vapor pressure of 5-HMF, as a result, the reaction is almost
complete within 10 minutes and the maximum recovery is
obtained. However, it seems that the contribution of a higher
temperature can be compensated by prolonging the reaction time
at a lower temperature, as it can be seen that similar recoveries
are reached in 30 minutes while the reaction was carried out at
3
Recycle of IrCl –[OMIM]Cl system
Compared with the reported laboratory solvent-extraction
35–38
1
50 °C.
method for 5-HMF removal,
the recycling of the catalytic
It is generally known that glucose can be converted to 5-HMF
systems in our EIVRD process is quite convenient. After one
batch of the reaction is completed, 2 g of fructose or glucose is
added directly into the flask reactor to initiate the next reaction.
As shown in Table 4, both the recovery and the yield of 5-HMF
by two steps, i.e., isomerization into fructose, and subsequent
2
7,39,53
dehydration to 5-HMF.
As a result, the dehydration of
glucose is more difficult than that of fructose. It is the same case
1224 | Green Chem., 2012, 14, 1220–1226
This journal is © The Royal Society of Chemistry 2012

View Article Online
Table 4 Recycling of the catalyst system in the dehydration of fructose and glucose
Fructose used as reactant
Glucose used as substrate
Entry
Recovery
Yield
Purity of obtained HMF
Recovery
Yield
Purity of obtained HMF
1
2
3
4
5
88.6
86.7
87.5
88.6
89.3
91.4
92.2
91.1
93.5
93.1
94.5
95.1
94.8
93.7
94.1
86.2
87.1
84.5
82.6
84.8
65.1
63.8
65.4
66.7
64.8
93.2
92.9
94.6
94.1
92.9
Note: the reaction conditions were similar to those of entry 2 for fructose and entry 11 for glucose in Table 3.
are well repeated during the five cycles of IrCl –[OMIM]Cl for
As reported, the imidazolium ionic liquid, that has shorter
carbon side chains, shows higher reactivity for the dehydration
of carbohydrates. However, in order to avoid being distilled out
together with 5-HMF under vacuum, ionic liquids having longer
carbon side chains are more suitable to this distillation process.
Based on the thermodynamic formula proposed by Valderrama
3
the dehydration of either fructose or glucose. Small amounts of
an insoluble solid is found at the bottom of the reactor, which
may be ascribed to the formation of by products such as humins,
carbon and other polymers at high temperature. Such a solid,
however, can be easily discarded by filtration after several runs
when the amount is no longer negligible.
45
and Robles, the boiling point of 3-octyl-1-methyl imidazolium
chloride ([OMIM]Cl) and 3-butyl-1-methyl imidazolium chlor-
ide ([BMIM]Cl) are estimated as 638.9 K and 548.9 K, respect-
The purity of the obtained 5-HMF by HPLC detection is
around 92.9–95.1%, indicating that 4.9–7.1% of [OMIM]Cl
6
(
most probably) is stripped out with 5-HMF. This can also be
ively. Considering the boiling point of 5-HMF is ca. 564 K, it
1
seen from its H NMR spectrum (Fig. 5s†) in which several
small impurity peaks are visible. However, a crystallization and
re-crystallization process can be easily carried out to further
purify the 5-HMF.
is rational to choose [OMIM]Cl as the reaction medium for the
dehydration of carbohydrate in our following distillation process.
The amount of [OMIM]Cl used has a large impact on the
mass transfer efficiency between the gas and liquid phases
during the EIVRD process. We found that it was very hard to
distribute the entrainer gas uniformly in the liquid phase
(
[OMIM]Cl) when only 2 g of [OMIM]Cl was used as the reac-
Discussion
tion medium during the first stage of the experiment. When the
amount of [OMIM]Cl was increased to 20 g, a simple ring-type
gas distributor could be introduced to the bottom of the reaction
flask to intensify the blending of the entrainer gas and the
As we know, dehydration of fructose is rather easier than that of
glucose. Therefore the first step to achieve the final goal of the
dehydration of biomass is to dehydrate glucose efficiently. The
EIVRD process can improve the yield of glucose remarkably.
More excitingly, it can effectively separate 5-HMF out from the
reaction bulk with high recovery. From an industrial point of
view, the EIVRD process is simple and easy to operate.
[OMIM]Cl,which worked efficiently. However, we did not opti-
mize the economic amount in the present work.
Comparing Tables 1–3, one can see that the direct vacuum dis-
tillation could separate 5-HMF to certain extent, while the intro-
duction of an entrainer, yet independent of its polarity, enabled
the distillation to be performed much more efficiently, indicating
that increasing the surface area of the gas–liquid phase is the
main reason for using an entrainer to improve the separation of
Conclusions
In this paper, for the first time, we have developed a process
called the entrainer-intensified vacuum reactive distillation
(EIVRD) process, which can be used to separate 5-HMF from
the dehydration solutions of carbohydrates catalyzed by a metal
chloride–[OMIM]Cl system. A high vacuity and an entrainer
were applied to intensify the dehydration of fructose or glucose.
The following main conclusions can be drawn:
(1) 5-HMF is thermally instable at 150 °C and 180 °C, while
the presence of IrCl or CrCl can decrease its degradation to
within 5% if the heating time is less than 30 minutes.
(2) Vacuum distillation is effective to separate 5-HMF from
[OMIM]Cl. Additionally, the introduction of nitrogen as an
entrainer can accelerate the distillation of 5-HMF, by which
95.5% and 94.7% recovery of 5-HMF can be achieved at 150 °C
in 30 minutes and at 180 °C in 10 minutes, respectively.
(3) The EIVRD process used in this study can improve the
dehydration of glucose remarkably. More excitingly, it can effec-
tively separate 5-HMF out from the reaction bulk with high
recovery. In the EIVRD process, the average recoveries of
5
-HMF. Consequently, a reactor that can produce a high surface
area of the reaction liquid may have the same effect for the sep-
aration of 5-HMF. Conceptual reaction/separation equipment that
54
is similar to the existing molecular distillation equipment may
be a good choice. In such an equipment, the dehydration of
carbohydrates in the ionic liquid is conducted within the narrow
and high-speed rotating cylinders. The reaction liquid therefore
forms a very thin layer with large specific surface areas adhering
to the inner wall of the outer cylinder, in which 5-HMF having
higher vapor pressure will be evaporated under heat and vacuum
and form light components or the final product, while the metal
salt and ionic liquid, having ignorable vapor pressures, will not
be distilled and consequently form the heavier component. The
heavier component can be re-entered into the cylinder reactor
with more carbohydrates to make the reaction consecutive.
3
3
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1220–1226 | 1225

View Article Online
5
-HMF dehydrated from fructose and glucose are around 93%
21 A. S. Dias, S. Lima, D. Carriazo, V. Rives, M. Pillinger and A.
A. Valente, J. Catal., 2006, 244, 230.
and 88%, respectively. The affinity of the entrainer has little
effect on the recovery of 5-HMF.
2
2
2 F. S. Asghari and H. Yoshida, Carbohydr. Res., 2006, 341, 2379.
3 F. Benvenuti, C. Carlini, P. Patrono, A. M. R. Galletti, G. Sbrana, M.
A. Massucci and P. Galli, Appl. Catal., A, 2000, 193, 147.
(
4) Recycling of the IrCl –[OMIM]Cl catalyst system in the
3
2
2
4 C. Lansalot-Matras and C. Moreau, Catal. Commun., 2003, 4, 517.
5 X. H. Qi, M. Watanabe, T. M. Aida and R. L. Smith, Green Chem., 2008,
EIVRD process is very convenient. Recovery of the catalyst
system and the actual yield of 5-HMF are repeated rather well
during all five recycled reactions of the dehydration of either
fructose or glucose.
10, 799.
26 A. S. Dias, M. Pillinger and A. A. Valente, J. Catal., 2005, 229, 414.
27 C. Moreau, R. Durand, S. Razigade, J. Duhamet, P. Faugeras, P. Rivalier,
P. Ros and G. Avignon, Appl. Catal., A, 1996, 145, 211.
28 Q. X. Bao, K. Qiao, D. Tomida and C. Yokoyama, Catal. Commun.,
2
008, 9, 1383.
Acknowledgements
2
3
9 C. Fayet and J. Gelas, Carbohydr. Res., 1983, 122, 59.
0 H. B. Zhao, J. E. Holladay, H. Brown and Z. C. Zhang, Science, 2007,
316, 1597.
1 G. A. Halliday, R. J. Young and V. V. Grushin, Org. Lett., 2003, 5, 2003.
2 R. M. Musau and R. M. Munavu, Biomass, 1987, 13, 67.
3 T. M. Aida, Y. Sato, M. Watanabe, K. Tajima, T. Nonaka, H. Hattori and
K. Arai, J. Supercrit. Fluids, 2007, 40, 381.
This research was supported by the Zhejiang Provincial Natural
Science Foundation of China (Y4090304, Y4100671) and
National Natural Science Foundation of China (21106134,
3
3
3
2
0936005).
3
3
3
3
3
3
4
4 X. H. Qi, M. Watanabe, T. M. Aida and R. L. Smith, Catal. Commun.,
2008, 9, 2244.
5 S. Q. Hu, Z. F. Zhang, J. L. Song, Y. X. Zhou and B. X. Han, Green
Chem., 2009, 11, 1746.
6 C. Moreau, A. Finiels and L. Vanoye, J. Mol. Catal. A: Chem., 2006,
References
1
2
3
4
5
J. N. Chheda and J. A. Dumesic, Catal. Today, 2007, 123, 59.
A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411.
M. J. Climent, A. Corma and S. Iborra, Green Chem., 2011, 13, 520.
P. Gallezot, Chem. Soc. Rev., 2012, 41, 1538.
S. Q. Hu, Z. F. Zhang, Y. X. Zhou, B. X. Han, H. L. Fan, W. J. Li,
J. L. Song and Y. Xie, Green Chem., 2008, 10, 1280.
253, 165.
7 S. Yu, H. M. Brown, X. W. Huang, X. D. Zhou, J. E. Amonette and Z.
C. Zhang, Appl. Catal., A, 2009, 361, 117.
8 G. Yong, Y. G. Zhang and J. Y. Ying, Angew. Chem., Int. Ed., 2008, 47,
9
345.
9 A. Takagaki, M. Ohara, S. Nishimura and K. Ebitani, Chem. Commun.,
009, 6276.
0 C. Z. Li, Z. H. Zhang and Z. B. K. Zhao, Tetrahedron Lett., 2009, 50,
403.
6
7
8
9
Y. Roman-Leshkov, C. J. Barrett, Z. Y. Liu and J. A. Dumesic, Nature,
2
2
007, 447, 982.
J. A. Widegren and R. G. Finke, J. Mol. Catal. A: Chem., 2003, 191,
87.
5
1
4
4
1 A. Hasabnis and S. Mahajani, Ind. Eng. Chem. Res., 2010, 49, 9058.
2 C. Noeres, E. Y. Kenig and A. Gorak, Chem. Eng. Process., 2003, 42,
C. J. Barrett, J. N. Chheda, G. W. Huber and J. A. Dumesic, Appl. Catal.,
B, 2006, 66, 111.
T. Ståhlberg, W. Fu, J. M. Woodley and A. Riisager, ChemSusChem,
157.
4
4
3 R. Taylor and R. Krishna, Chem. Eng. Sci., 2000, 55, 5183.
4 Z. J. Wei, F. J. Li, H. B. Xing, S. G. Deng and Q. L. Ren, Korean
J. Chem. Eng., 2009, 26, 666.
2
011, 4, 451.
1
0 A. A. Rosatella, S. P. Simeonov, R. F. M. Frade and C. A. M. Afonso,
Green Chem., 2011, 13, 754.
4
4
5 J. O. Valderrama and P. A. Robles, Ind. Eng. Chem. Res., 2007, 46, 1338.
6 P. A. Z. Suarez, S. Einloft, J. E. L. Dullius, R. F. de Souza and J. Dupont,
J. Chim. Phys. Phys.-Chim. Biol., 1998, 95, 1626.
1
1
1
1 Y. Takeuchi, F. M. Jin, K. Tohji and H. Enomoto, J. Mater. Sci., 2007, 43,
472.
2 Y. Roman-Leshkov, J. N. Chheda and J. A. Dumesic, Science, 2006, 312,
933.
3 B. Girisuta, L. P. B. M. Janssen and H. J. Heeres, Chem. Eng. Res. Des.,
006, 84, 339.
2
4
4
4
5
7 Z. J. W. Z. J. Wei, Y. Li, D. Thushara, Y. X. Liu and Q. L. Ren, J. Taiwan
1
Inst. Chem. Eng., 2011, 42, 363.
8 J. A. Chun, J. W. Lee, Y. B. Yi, S. S. Hong and C. H. Chung, Korean
J. Chem. Eng., 2010, 27, 930.
9 P. Decap, S. Braipson-Danthine, A. Vanbrabant, W. De Greyt and
C. Deroanne, J. Am. Oil Chem. Soc., 2004, 81, 611.
2
1
1
4 B. F. M. Kuster and H. S. Vanderbaan, Carbohydr. Res., 1977, 54, 165.
5 J. D. Chen, B. F. M. Kuster and K. Vanderwiele, Biomass Bioenergy,
1
991, 1, 217.
6 J. Jow, G. L. Rorrer, M. C. Hawley and D. T. A. Lamport, Biomass,
987, 14, 185.
0 A. Maza, R. A. Ormsbee and L. R. Strecker, J. Am. Oil Chem. Soc.,
1
1992, 69, 1003.
1
5
5
1 J. Cmolik and J. Pokorny, Eur. J. Lipid Sci. Technol., 2000, 102, 472.
2 G. R. List, J. W. King, J. H. Johnson, K. Warner and T. L. Mounts, J. Am.
Oil Chem. Soc., 1993, 70, 473.
1
1
7 F. S. Asghari and H. Yoshida, Ind. Eng. Chem. Res., 2006, 45, 2163.
8 T. Armaroli, G. Busca, C. Carlini, M. Giuttari, A. M. R. Galletti and
G. Sbrana, J. Mol. Catal. A: Chem., 2000, 151, 233.
53 R. L. Huang, W. Qi, R. X. Su and Z. M. He, Chem. Commun., 2010, 46,
1
9 C. Carlini, M. Giuttari, A. M. R. Galletti, G. Sbrana, T. Armaroli and
G. Busca, Appl. Catal., A, 1999, 183, 295.
1115.
54 P. F. Martins, V. M. Ito, C. B. Batistella and M. R. W. Maciel, Sep. Purif.
Technol., 2006, 48, 78.
2
0 P. Carniti, A. Gervasini, S. Biella and A. Auroux, Catal. Today, 2006,
118, 373.
1226 | Green Chem., 2012, 14, 1220–1226
This journal is © The Royal Society of Chemistry 2012