Catalytic dehydration of carbohydrates suspended in organic solvents promoted by ALCL3/SiO2 coated with choline chloride

Article abstract of DOI:10.1002/cssc.201402761

We show that the coating of choline chloride on silica-supported AlCl₃ allows the dehydration of carbohydrates to successfully proceed in low boiling point organic solvents. The concept is based on the in situ formation of a deep eutectic liquid phase on the catalyst surface, thus facilitating the interaction between the solid catalyst and insoluble carbohydrate.

Full text of DOI:10.1002/cssc.201402761

CHEMSUSCHEM
COMMUNICATIONS
DOI: 10.1002/cssc.201402761
Catalytic Dehydration of Carbohydrates Suspended in
Organic Solvents Promoted by AlCl3/SiO2 Coated with
Choline Chloride
Jie Yang,[a, b] Karine De Oliveira Vigier,[a] Yanlong Gu,*[b] and FranÅois Jꢀrꢁme*[a]
We show that the coating of choline chloride on silica-support-
ed AlCl3 allows the dehydration of carbohydrates to successful-
ly proceed in low boiling point organic solvents. The concept
is based on the in situ formation of a deep eutectic liquid
phase on the catalyst surface, thus facilitating the interaction
between the solid catalyst and insoluble carbohydrate.
the most efficient ionic liquids are also still prohibitively expen-
sive for use on a large scale. In addition, the sensitivity of cur-
rent solid catalysts to the change of ionic strength and the
long term recycling of ionic liquids are two major problems.
The possibility to conduct heterogeneously-catalyzed con-
version of carbohydrates in a low boiling point organic solvent
is of huge interest from an industrial point of view because of
their ease of recovery and recycling at the end of the process.
However, the insolubility of carbohydrates and solid catalysts
in such reaction medium requires overcoming low solid–solid
interaction that restricts the reaction rate to an unacceptable
level. For this reason, use of low boiling point organic solvents
for heterogeneously catalyzed conversion of carbohydrates
was scarcely investigated.[4] In most cases, high boiling point
organic solvents capable of dissolving large amount of carbo-
With the depletion of fossil resources, carbohydrates are now
rapidly emerging as an important class of renewable raw mate-
rials from which valuable chemicals and fuels can be pro-
duced.[1] Heterogeneously catalyzed conversion of carbohy-
drates is however a challenging task and requires many obsta-
cles to be tackled such as the low solubility of carbohydrates
in organic solvents, activation at low temperature, impurities
of different nature, and so on.[2]
In the field of carbohydrates
processing, water is often used
as a solvent because of its ability
to dissolve large amount of car-
bohydrates and also because of
its low price and biocompatibil-
ity. One should, however, men-
tion that utilization of water as
a reaction medium poses strin-
gent problems such as (1) the
deactivation of solid catalysts by
water, (2) the recovery of the
products of the reaction which is
often energy-consuming and,
(3) wastewater treatment, which
Scheme 1. Illustration of the work.
is one of the main problems
faced by biorefineries. Besides
water, many recent studies have also highlighted the possible
use of ionic liquids for carbohydrate processing.[3] However,
hydrates such as dimethysulfoxide or dimethylformamide are
used.[5] Using low boiling point organic solvents, water is often
added to facilitate the dissolution of carbohydrates.[6]
Here, we wish to propose another approach based on the
impregnation of solid catalysts with choline chloride (ChCl). In
particular, in organic solvents, we show that the deposition of
ChCl on a catalytic surface facilitates the contact between the
solid catalysts and the insoluble carbohydrates. The concept is
inspired by the ability of ChCl to readily form deep eutectic or
low melting mixtures[7] with carbohydrates. In particular, we
hypothesized that such phenomenon may also occur on a cata-
lyst surface (Scheme 1). This kind of solubilization/liquefaction
of carbohydrates on the catalyst surface results in a significant
improvement of the catalytic reaction rate, thus opening an at-
[a] J. Yang, K. De Oliveira Vigier, F. Jꢀrꢁme
Institut de Chimie des Milieux et Matꢀriaux de Poitiers
Universitꢀ de Poitiers, ENSIP
1 rue Marcel Dorꢀ, 86073 Poitiers cedex 9 (France)
E-mail: francois.jerome@univ-poitiers.fr
[b] J. Yang, Y. Gu
Key Laboratory for Large-Format Battery, Materials and System
Ministry of Education, School of Chemistry and Chemical Engineering
Huazhong University of Science and Technology
Wuhan 430074 (China)
E-mail: klgyl@hust.edu.cn
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201402761.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 6
&
1
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
COMMUNICATIONS
www.chemsuschem.org
tractive route to convert carbohydrates in conventional low
boiling solvents.
Table 1. Catalytic dehydration of fructose to HMF in MeTHF.
For the demonstration of this concept, we first focused on
the well-known acid-catalyzed dehydration of fructose to 5-hy-
droxymethylfurfural (HMF) as a model reaction.[8] Glucose and
xylose, which are two more interesting feeds, are studied at
the end of the manuscript. As low boiling point organic sol-
vents, we selected methyltetrahydrofuran (MeTHF) and methyl-
isobutylketone (MIBK) because of their ease of elimination at
the end of the reaction (boiling point <1208C), their accepted
biocompatibility and the insolubility of ChCl in such media.
First of all, AlCl3.6H2O was impregnated over silica gel. To
this end, 60 mg of AlCl3.6H2O was dissolved in 3.0 mL of
MeOH containing 1.2 g of silica in suspension. The solution
was stirred for 10 min prior to removing methanol under re-
duced pressure yielding AlCl3@SiO2 (4.8 wt% of AlCl3.6H2O).
TEM analysis showed the coating of SiO2 with AlCl3 (Fig-
ure S1a). ICP (inductively coupled plasma) analysis confirmed
the loading of 0.53 wt% of Al species on SiO2 (i.e., 4.8 wt%
AlCl3.6H2O) (Table S1). Note that energy dispersion X-ray (EDX)
analysis also revealed that part of Al species had diffused into
the bulk. This was further confirmed by X-ray photoelectron
spectroscopy (XPS) analysis that revealed an aluminium con-
tent lower than the detection limit of the apparatus on the
AlCl3@SiO2 surface (on a layer of 10 nm). Then, AlCl3@SiO2 was
tested as a solid catalyst in the dehydration of fructose to HMF
using first MeTHF as a reaction medium. Typically, 1.8 g of fruc-
tose was suspended in 18 mL of MeTHF and the amount of
added AlCl3@SiO2 was calculated to obtain 2.5 mol% of AlCl3
relative to the amount of fructose. The solution was then
heated under reflux (808C) in a Dean–Stark apparatus to con-
tinuously remove water from the catalytic phase. The reaction
progress was monitored by HPLC and the efficiency of solid
catalyst was determined by the initial production rate (IPR) of
HMF defined by the number of mol of HMF produced per
mole of Al and per hour. Results are summarized in Table 1.
As expected, in MeTHF, the production rate of HMF was
rather low in the presence of AlCl3@SiO2 due to low solid–solid
interaction between AlCl3@SiO2 and the insoluble fructose. For
instance, after 1 hour of reaction, HMF was produced with only
8% yield which corresponded to an IPR of 3.2 hÀ1 (Table 1,
entry 1). Other products were mostly black soluble and insolu-
ble materials.
[a]
Entry ChCl loading on
t
Fructose
HMF yield [%] IPR [hÀ1
]
AlCl3@SiO2 [wt%] [h] conv. [%]
1
2
0
4
1
1
1
1
1
1
4
2
2
2
1
90
91
91
92
93
94
98
nd
nd
nd
96
8
14
17
20
25
51
55
58
28
17
26
3.2
5.9
7.0
3
7
4
5
6
7
8[b]
9[c]
10[d]
11[e]
17
28
39
39
39
39
39
-
9.6
11.2
38.4
38.4
nd
nd
nd
nd
[a] Initial production rate of HMF; [b] first recycling; [c] second recycling;
[d] third recycling; [e] in a biphasic AlCl3-ChCl/MeTHF system.
which is higher than the temperature required for catalytic ex-
periments (Figure S2b). In particular, DTA/TGA analysis was
characterized by two endothermic peaks located at 858C (loss
of water) and 2368C (loss of ChCl). The wt% loss of ChCl is
clearly proportional to the amount of ChCl initially impregnat-
ed. For instance, for the AlCl3-ChCl17@SiO2 catalyst, the loss of
ChCl was 16.73 wt% (vs 17 wt% for the expected value), fur-
ther confirming that ChCl was not degraded during the im-
pregnation step (Figure S2b). XPS analyses were also per-
formed to determine the possible diffusion of ChCl and alumi-
nium in the bulk. As observed for AlCl3@SiO2, the aluminium
content of the AlCl3-ChCl17@SiO2 catalyst was lower than the
detection limit of the apparatus, confirming that aluminium
had also diffused into the bulk during impregnation. Regarding
ChCl, the nitrogen amount determined by XPS was found to
be 11 wt%, which is lower than the amount obtained by ele-
mental analysis (17 wt%). This difference suggests that ChCl
has also partly diffused into the bulk. In particular, we may
conclude that only 65 wt% of the initial amount of ChCl is
present on the AlCl3-ChCl17@SiO2 surface (in a layer of 10 nm).
Catalytic results are summarized in Table 1. Interestingly,
a clear improvement of the initial production rate of HMF was
observed when ChCl was impregnated on AlCl3@SiO2. In partic-
ular, when the amount of impregnated ChCl was increased
from 4 to 39 wt%, the initial production rate of HMF was con-
comitantly enhanced from 3.2 hÀ1 to 38.4 hÀ1, further demon-
strating the contribution of ChCl to the catalyst activity (en-
tries 1–6 and Figure 1). By means of ICP measurements, we
found that the leaching of Al in MeTHF was below the detec-
tion limit of the apparatus, which is not surprising when one
considers that AlCl3 is highly soluble in ChCl-derived DES (deep
eutectic solvent) liquid phase. On the other hand, analysis of
Next, ChCl was impregnated over AlCl3@SiO2. To this end,
the desired amount of ChCl was co-added during the impreg-
nation of AlCl3.6H2O on silica, yielding the so-called AlCl3-
ChClX@SiO2 (X=wt% of ChCl on SiO2). TEM analysis of the
AlCl3-ChCl17@SiO2 catalyst is provided in the supporting infor-
mation (SI, Figure S1b) as a selected example and also evi-
denced a deposition on the silica support. The ratio of C/N de-
termined by elemental analysis (4.23) remained similar to that
of neat ChCl (C/N=4.29) indicating that ChCl was not degrad-
ed during the impregnation step. In addition, the wt% of ChCl
determined by elemental analysis was also in accordance with
the expected value (17 wt%). By means of DTA/TGA (differen-
tial thermal analysis/thermogravimetric analysis) analyses, the
AlCl3-ChClX@SiO2 catalyst were found to be stable up to 2368C
the MeTHF by H and 13C NMR did not reveal any contamina-
1
tion with ChCl, which is in perfect agreement with the very
low solubility of ChCl in such organic solvent.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 6
&
2
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
COMMUNICATIONS
www.chemsuschem.org
ed experimental evidences that suggest the formation of
a DES on the catalyst surface during the reaction, one should
comment that the full characterization of this in-situ-formed
DES was difficult to perform due to problems of separation be-
tween the solid catalyst and the suspension of unreacted car-
bohydrates at low conversions.
Clearly, at such high ChCl loading (39 wt%), the utility of
silica is questionable since reaction can also occur in a simple
biphasic solution of fructose-ChCl DES/MeTHF.[9] Therefore,
a similar reaction was conducted without silica but keeping all
other components unchanged. In such a case, a biphasic
system was clearly formed in accordance with previous reports.
The lower (and minor) phase was composed of the liquid fruc-
tose/ChCl DES containing AlCl3 and MeTHF was the upper
phase. Under these biphasic conditions, HMF was however
produced with only 26% yield versus 55% in the presence of
AlCl3-ChCl39@SiO2 (Table 1, entry 11). Note that, for other ChCl
loading (4–28 wt%), no appreciable yield of HMF was detected
in the absence of SiO2. The lower yield of HMF obtained when
removing silica was presumably due to mass-transfer limita-
tion. The DES is indeed highly viscous and dispersed with diffi-
culty in the MeTHF phase. In the presence of silica, we assume
that SiO2 dispersed the in-situ-produced DES in MeTHF, thus fa-
voring a better contact between fructose and the catalytic
phase (Figure 2). It is noteworthy that, at the end of the reac-
tion, the MeTHF phase contained only 60 wt% of produced
HMF in the case of the biphasic system versus more than
96 wt% when the reaction was conducted in the presence of
the AlCl3-ChCl@SiO2 solid catalyst, presumably due to its better
dispersion in MeTHF. This represents a significant advantage
over biphasic systems, which require several extraction runs to
fully recover HMF, a detrimental aspect for the viability of
these biphasic systems.10
Figure 1. Kinetic profiles for AlCl3-ChCl@SiO2 and AlCl3@SiO2 in MeTHF at
808C.
The AlCl3-ChCl39@SiO2 catalyst was found to be the most
productive one (HMF was initially produced 12 times faster
than over AlCl3@SiO2). Kinetic profiles showed that a maximum
yield of HMF of 55% can be obtained when prolonging the re-
action time to 4 h (Figure 1, Table 1, entry 7). Other products
were mostly soluble black materials commonly named humins.
Because the reaction was performed in an organic solvent, no
side formation of levulinic and formic acid (stemming from the
acid-catalyzed rehydration of HMF) was evidenced, as is gener-
ally the case in water.[8] As an evidence, one can note that
HMF is rather stable in our conditions because no appreciable
decrease of the yield of HMF was observed after prolonged re-
action time as commonly observed in water (Figure 1).
Although the AlCl3-ChCl39@SiO2 catalyst afforded the best re-
sults, one should comment that such high loading of ChCl
hampered the convenient recycling of the catalyst. Indeed,
whereas at a loading lower than 28 wt%, the catalyst can be
conveniently recovered as a powdery and free-flowing solid at
the end of the reaction, this is not the case when the ChCl
loading reached 39 wt%. In such a case, the catalyst was
coated by a highly viscous liquid (consistent with the in situ
formation of a DES) making its full recovery and recycling tech-
nically difficult (sticky solid). In addition, in the second cycle, it
was very difficult to adequately re-disperse the recovered
AlCl3-ChCl39@SiO2 catalyst in MeTHF. For these reasons, a signifi-
cant drop of the HMF yield was observed during the second
and third run (Table 1, entries 8–10). Although we have collect-
To ensure the recycling of the AlCl3-ChClX@SiO2 catalyst, it is
clear that the system needs to be optimized with the aim of
being operational at a lower ChCl loading. In this context,
MeTHF was replaced by MIBK. The catalytic procedure was
similar to that described above with MeTHF. To facilitate the re-
covery of a powdery and free-flowing solid at the end of the
reaction, the loading of ChCl was kept at 17 wt%.
In MIBK, it was possible to increase the reaction temperature
up to 1208C (compared to only 808C in MeTHF), resulting in
Figure 2. Pictures of the reaction media (a) in a biphasic ChCl39/Fructose DES and (b) in the presence of AlCl3-ChCl39@SiO2.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 6
&
3
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
COMMUNICATIONS
www.chemsuschem.org
As observed above in MeTHF, more than 95% of produced
HMF was recovered in the MIBK phase, whereas the partition
coefficient of HMF in a water/MIBK or fructose-ChCl DES/MIBK
(1/1) biphasic systems is close to 1.[10]
Table 2. Catalytic dehydration of fructose to HMF in MIBK.
As expected, the reaction temperature has a significant
effect on the initial production rate of HMF. For instance, the
yield of HMF obtained after 4 h of reaction dropped from 65%
to 53% and 33% when the temperature was concomitantly
decreased from 1008C to 908C and 808C, respectively (Table 2,
entries 4–6). Similarly, a decrease of the amount of catalyst
from 2.5 mol% of AlCl3 to 1.3 mol% and 0.6 mol% was obvi-
ously accompanied by a decrease of the reaction rate. For in-
stance, HMF was produced with 52%, 56%, and 16% yield
after 1 h, 8 h and 16 h of reaction in the presence of 2.5 mol%,
1.3 mol%, and 0.6 mol% of AlCl3, respectively (Table 2, en-
tries 4, 7, 8). It should be noted that an increase of the fructose
content by a factor 2 (from 10 wt% to 20 wt%) led to a de-
crease of the maximum yield of HMF to 30%. Please note that
this experiment was performed at a catalyst loading of
1.3 mol% (compared to 2.6 mol%); otherwise, the system was
strongly limited by mass-transfer problems. Hence, 1008C,
10 wt% of fructose and 2.5 mol% of AlCl3 were selected as the
optimal conditions.
Entry ChCl loading on
T
t
Fructose HMF yield [%] IPR [hÀ1
]
AlCl3@SiO2 [wt%] [8C] [h] conv. [%]
1
2
0
4
100
100
100
100
90
4
4
4
4
4
4
8
96
96
95
96
95
95
94
48
54
52
65
53
33
56
16
30
57
50
52
15.2
19.2
22.8
49.6
nd
3
9
4
5
6
7[a]
8[b]
9[c]
10[d]
11[e]
12[f]
17
17
17
17
17
17
17
17
17
80
nd
nd
100
100 16 97
100
100
100
100
nd
3
2
2
2
nd
94
nd
nd
nd
49.6
nd
nd
[a] 1.3 mol% of AlCl3 relative to fructose, [b] 0.6 mol% relative to fructose,
[c] at 20 wt% of fructose and 1.3 mol% of AlCl3 relative to fructose,
[d] first cycle after 2 h of reaction, [e] second cycle, [f] third run.
an improvement of the initial production rate of HMF by
a factor 5, as compared to the above described experiments
performed in MeTHF (Table 2, entry 4). Note that, at 808C, no
significant difference of HMF yield was, however, observed be-
tween MeTHF and MIBK after 4 h of reaction, suggesting that
there is no beneficial effect of MIBK on the reaction selectivity
(Table 2, entry 6 and Figure 1). As observed in the case of
MeTHF, we were pleased to see that an impregnation of the
AlCl3@SiO2 catalyst with ChCl has also a significant effect on
the initial production rate of HMF (Table 2, entries 1–4).
Next, recycling experiments were undertaken. As described
above, after 2 h of reaction (i.e., 57% yield), the AlCl3-
ChCl17@SiO2 catalyst was successfully recovered by filtration
and reused as collected. The MIBK phase was removed under
reduced pressure and also recycled. By means of elemental
analysis performed on the used catalyst, we found that after
the first cycle the C/N ratio was increased from 4.23 to 14.59
(Table S1). The increase of the C wt% observed after reaction
could be due to the presence of (1) fructose (DES phase),
(2) HMF, and/or (3) other carbonaceous materials, presumably
humins (a common side-product observed in the chemistry of
carbohydrates) on the recovered AlCl3-ChCl17@SiO2 catalyst sur-
face. To check this, the recovered AlCl3-ChCl17@SiO2 catalyst
was washed with water. Analysis of the aqueous wash by HPLC
revealed the presence of ChCl, fructose, and HMF. The amount
of fructose and HMF was 2.8 mol% and 1.3 mol%, respectively
(relative to the initial amount of fructose). Although presence
of fructose and HMF on the catalyst surface contributes to in-
crease the C wt%, it cannot, however, totally explain the C/N
ratio observed after reaction, suggesting that other carbona-
ceous materials, presumably humins, were also deposited on
the catalyst surface. In addition, DTA/TGA analyses of used
AlCl3-ChCl17@SiO2 also revealed two peaks at 2468C and 3258C.
If 2468C may be ascribed to the thermal degradation of ChCl
(Figure S2), fructose and humins are still under decomposition
at higher temperature, which could be consistent with the
peak observed at 3258C. Interestingly, this deposition of
humins did not affect the efficiency of the AlCl3-ChCl17@SiO2
catalyst that was successfully recycled in MIBK at least three
times without significant change of its activity (Table 2, en-
tries 10–12). One may hypothesize that absorbed humins are
partly extracted by MIBK in the subsequent catalytic runs.
Different hydrated metal chlorides were also evaluated
under our reaction conditions. As summarized in Table 3, AlCl3
As also illustrated by the kinetic profiles (Figure 3), the initial
production rate of HMF over the AlCl3@SiO2 catalyst was im-
proved by a factor 3.2 when it was impregnated with 17 wt%
of ChCl. Interestingly, the yield of HMF reached a maximum of
65% after 4 h of reaction, which is a quite remarkable yield for
reaction in an organic solvent (Figure 3). Interestingly, whereas
water is released, no appreciable leaching of AlCl3 and ChCl
was observed in MIBK. This was ascribed to the low amount of
theoretical water released (540 mg) during the reaction which
is too low to dissolve ChCl and AlCl3 in MIBK.
Figure 3. Kinetic profiles for AlCl3-ChClx@SiO2 and AlCl3 in MIBK at 1008C.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 6
&
4
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
COMMUNICATIONS
www.chemsuschem.org
the catalytic dehydration of glucose to HMF is much more
complex and requires first an isomerization of glucose to fruc-
tose prior to dehydration to HMF.[11] In this context, AlCl3 and
CrCl3 were selected as catalytic sites since both metal chlorides
are known to be capable of catalyzing the conversion of glu-
cose to HMF.[10c] In addition, from Table 3, both metal chlorides
were also found to be the most efficient catalysts. As expected,
the reaction of glucose suffered from a lack of selectivity and
HMF was obtained with 10% and 20% yield after 5 h of reac-
tion in the presence AlCl3 and CrCl3, respectively (Table 4, en-
tries 2, 3). An increase of the amount of metal chloride from
2.5 mol% to 5.0 mol% did not significantly change the yield of
HMF obtained after 5 h (Table 4, entries 4, 5). Upon prolonged
reaction time (24 h), the yield of HMF was, however, slightly
improved to 25% (Table 4, entry 6). As observed above in the
case of fructose, an increase of the ChCl content from 17 wt%
to 28 wt% and 39 wt% led to an increase of the HMF yield up
to 31% in the best case of CrCl3 (Table 4, entries 7–9). After
24 h of reaction at a ChCl content of 28 wt%, the yield of HMF
increased to 38% in the presence of AlCl3 and CrCl3 (Table 4,
entries 10–11). The reaction also proceeded from xylose, anoth-
er valuable feed derived from hemicellulose. In this case, furfu-
ral was obtained in MIBK with 43% yield in the presence of
AlCl3 (Table 4, entry 12). Here again, more than 95 wt% of pro-
duced furfural was in the MIBK phase, whereas in a biphasic
system, furfural was extracted with difficulty from the catalytic
phase.
Table 3. Influence of Lewis acids in the catalytic dehydration of fructose
to HMF in MIBK.
Entry
Metal chloride
Conv. [%]
HMF yield [%]
1
2
3
4
5
6
7
8
AlCl3.6H2O
CrCl3.6H2O
CuCl2.2H2O
FeCl3.6H2O
LiCl
FeCl2.4H2O
CoCl2.6H2O
AlCl3
94
93
92
94
–
–
–
92
59
50
38
22
–
–
–
40
and CrCl3 were the most efficient ones, affording HMF with
59% and 50% yield of HMF, respectively, after 1 h of reaction
(Table 3, entries 1–2). As a general trend, efficiency of metal
chlorides can be classified as follows: AlCl3 ~CrCl3 >CuCl2 >
FeCl3. However, no reaction was observed in the presence of
LiCl, FeCl2, or CoCl2. It is noteworthy that when anhydrous
AlCl3 was used, yield of HMF was lower. Indeed, although the
AlCl3-ChCl17@SiO2 catalyst afforded HMF with 59% yield after
1 h of reaction when prepared with AlCl3.6H2O, the HMF yield
dropped to 40% when anhydrous AlCl3 was employed
(Table 3, entries 1 and 8).
Finally, inulin, a biopolymer of fructose was also tested. After
only 2 h of reaction, we were pleased to see that HMF can be
produced with 45% yield in MIBK, which is a competitive yield
considering that two successive steps are required to convert
inulin to HMF (hydrolysis and dehydration) (Table 4, entry 13).
Cellulose, a valuable feed, was also tested. As described in the
literature, cellulose is much more recalcitrant and remained un-
altered during the catalytic process. A ball-milling of cellulose
for 24 h prior catalytic experiments, however, slightly improved
its reactivity and, after 3 h of reaction, 8% yield of HMF was
obtained (Table 4, entry 14).
In a last set of experiments, this concept was applied to a dif-
ferent feed of valuable carbohydrates such as glucose and
xylose, derived from cellulose and hemicellulose, respectively.
Results are summarized in Table 4. In comparison to fructose,
Table 4. Heterogeneously-catalyzed dehydration of valuable carbohy-
drate feeds to furfural derivatives.
In conclusion, we report here that the coating of a silica-sup-
ported AlCl3 with ChCl allowed the dehydration of carbohy-
drates to furanic derivatives to successfully proceed in a low
boiling point organic solvent. The scientific line of this work is
based on the in situ formation of a liquid deep eutectic mix-
ture on the catalyst surface, which may greatly facilitate the
contact between the two suspensions of silica and carbohy-
drates, resulting in a significant enhancement of the reaction
rate. This concept can be applied to different valuable feeds
such as fructose, glucose, xylose, and inulin. Although similar
yields of HMF can be obtained in classical biphasic systems,
our work is distinguished in that 1) the organic solvent con-
tains more than 95 wt% of HMF thus by-passing the costly
and energy-consuming extraction stage required in the case of
biphasic systems (low partition coefficient of furfural and
HMF), 2) the possible recovery of the solid catalyst at the end
of the reaction, and 3) the use of a catalytic amount of ChCl.
Entry Metal
Catalyst
ChCl
[wt%]
Carbohydrates t Conv HMF
[h] [%] yield [%]
chloride [mol%]
1
2
3
4
5
6
7
8
AlCl3
AlCl3
CrCl3
AlCl3
CrCl3
AlCl3
CrCl3
AlCl3
CrCl3
CrCl3
AlCl3
AlCl3
AlCl3
AlCl3
2.5
2.5
2.5
5.0
5.0
5.0
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
17
17
17
17
17
17
28
28
39
28
28
17
17
17
fructose
glucose
glucose
glucose
glucose
glucose
glucose
glucose
glucose
glucose
glucose
xylose
4
5
5
5
5
96
91
91
93
94
65
10
20
14
23
25
28
16
31
38
38
43
45
8
24 96
5
5
5
92
93
94
9
10
11
12
13
14
24 96
24 96
24 95
2
3
inulin
cellulose[a]
nd
nd
[a] cellulose was pre-treated in
a planetary ball-milling for 24 h at
200 rpm prior catalytic experiments.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 6
&
5
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
COMMUNICATIONS
www.chemsuschem.org
ers, P. A. Jacobs, B. F. Sels, ChemCatChem 2011, 3, 82–94; d) P. L. Dhepe,
A. Fukuoka, ChemSusChem 2008, 1, 969–975; e) Y.-C. Lin, G. W. Huber,
Energy Environ. Sci. 2009, 2, 68–80.
Experimental Section
Typical procedure for the preparation of AlCl3-ChClx@SiO2: 1.2 g
of silica gel was suspended in 3 mL of methanol containing 60 mg
AlCl3.6H2O and the desired amount of choline chloride (60 mg to
800 mg). The suspension was stirred at room temperature for
10 min. Then, methanol was removed under reduced pressure
yielding a white solid named AlCl3-ChClx@SiO2 with x=4–39. Note
that x represents the mass fraction of ChCl on SiO2 and was deter-
mined as follow x=(mChCl/(mSiO2 +mAlCl3 +mChCl))ꢃ100 where m rep-
resents the mass. SiO2 used in this study was purchased from Carlo
Erba (ref. P2050027, Silica 60 ꢄ). SiO2 has a particle size within the
range of 40–63 mm and a water content lower than 6 wt%.
Typical procedure for the catalytic production of HMF: In a typi-
cal procedure, 1.8 g of fructose was suspended in 18 mL of MeTHF
or MIBK. The amount of catalyst introduced into the flask was cal-
culated to obtain 2.5 mol% of AlCl3.6H2O relative to the amount of
fructose. In the case of MeTHF, a Dean–Stark apparatus was used.
The solution was then stirred at 808C in the case of Me-THF and
1008C in the case of MIBK. At the desired time, the reaction media
was cooled down to room temperature and then centrifuged at
3000 rpm for 5 min. The organic phase containing HMF was ana-
lyzed by HPLC. The solid phase was also analyzed to determine the
amount of unreacted fructose. To this end, the solid phase was
washed with water and analyzed by HPLC.
[3] a) H. Wang, G. Gurau, R. D. Rogers, Chem. Soc. Rev. 2012, 41, 1519–
1537; b) H. Tadesse, R. Luque, Energy Environ. Sci. 2011, 4, 3913–3929;
c) A. Brandt, J. Grasvik, J. P. Hallett, T. Welton, Green Chem. 2013, 15,
550–583; d) H. Olivier-Bourbigou, L. Magna, D. Morvan, Appl. Catal. A
2010, 373, 1–56.
[4] as recent examples, please see a) R. Weingarten, A. Rodriguez-Beuer-
man, F. Cao, J. S. Luterbacher, D. M. Alonso, J. A. Dumesic, G. W. Huber,
ChemSusChem, 2014, DOI: 10.1002/cctc.201402299; b) L. Lai, Y. G.
Zhang, ChemSusChem 2011, 4, 1745–1748; c) S. Saravanamurugan, M.
Paniagua, J. A. Melero, A. Riisager, J. Am. Chem. Soc. 2013, 135, 5246–
5249; d) L. Qi, I. T. Horvath, ACS Catal. 2012, 2, 2247–2249.
[5] as selected examples, please see a) M. Ohara, A. Takagaki, S. Nishimura,
K. Ebitani, Appl. Catal. A 2010, 383, 149–155; b) X. Qi, M. Watanabe,
T. M. Aida, R. L. Smith, Ind. Eng. Chem. Res. 2008, 47, 9234–9239; c) K.
Seri, Y. Inoue, H. Ishida, Catal. Lett. 2000, 1, 22–23; d) F. Wang, A.-W. Shi,
X.-X. Qin, C.-L. Liu, W. S. Dong, Carbohydr. Res. 2011, 346, 982–985;
e) C. B. Rasrendra, J. N. M. Soetedjo, I. G. B. N. Makertihartha, S. Adisas-
mito, H. J. Heeres, Top. Catal. 2012, 55, 543–549.
[6] M. H. Tucker, A. J. Crisci, B. N. Wigington, n. Phadke, R. Alamillo, J.
Zhang, S. L. Scott, J. A. Dumesic, ACS Catal. 2012, 2, 1865–1876 and ref-
erences therein.
[7] a) Q. Zhang, K. De Oliveira Vigier, S. Royer, F. Jꢀrꢁme, Chem. Soc. Rev.
2012, 41, 7108–7146; b) P. Dominguez de Maria, Z. Maugeri, Curr. Opin.
Chem. Biol. 2011, 15, 220–225; c) D. Carriazo, M. C. Serrano, M. C. Gu-
tierrez, M. L. Ferrer, F. del Monte, Chem. Soc. Rev. 2012, 41, 4996–5014;
d) M. Francisco, A. van den Bruinhorst, M. C. Kroon, Angew. Chem. Int.
Ed. 2013, 52, 3074–3085; Angew. Chem. 2013, 125, 3152–3163; e) C.
Russ, B. Koenig, Green Chem. 2012, 14, 2969–2982; f) A. P. Abbott, D.
Boothby, G. Capper, D. L. Davies, R. K. J. Rasheed, J. Am. Chem. Soc.
2004, 126, 9142–9147.
[8] a) A. A. Rosatella, S. P. Simeonov, R. F. M. Frade, C. A. M. Alfonso, Green
Chem. 2011, 13, 754–793; b) M. E. Zakrzewska, E. Bogel-Lukasik, R.
Bogel-Lukasik, Chem. Rev. 2011, 111, 397–417; c) R. J. van Putten, J. C.
van der Waal, E. de Jong, C. B. Rasrendra, H. J. Heeres, J. G. de Vries,
Chem. Rev. 2013, 113, 1499–1597; d) T. Stꢆhlberg, W. J. Fu, J. M. Wood-
ley, A. Riisager, ChemSusChem 2011, 4, 451–458.
Typical procedure for recycling experiments: At the end of the
reaction, the solid phase recovered after centrifugation was
washed with diethylether to remove trace of HMF from the catalyst
surface. The catalyst was then dried at 608C under vacuum. Then,
it was reused in the next catalytic cycles using fresh fructose and
recycled MeTHF or MIBK.
Acknowledgements
Authors are grateful to the CNRS, the University of Poitiers and
the French Ministry of Research for their financial supports. Jie
Yang also thanks the support of the China Scholarship Council
(CSC). Yanlong Gu thanks for National Natural Science Founda-
tion of China for the financial support (21373093) and the Fun-
damental Research Funds for the Central Universities of China
(2014ZZGH019). Catherine Canaff is also acknowledged for XPS
analyses.
[9] a) F. Ilgen, D. Ott, D. Kralisch, C. Reil, A. Palmberger, B. Konig, Green
Chem. 2009, 11, 1948–1954; b) S. Hu, Z. Zhang, Y. Zhou, J. Song, H. Fan,
B. Han, Green Chem. 2009, 11, 873–877; c) F. Liu, J. Barrault, K. De Oli-
veira Vigier, F. Jꢀrꢁme, ChemSusChem 2012, 5, 1223–1226; d) K. De Oli-
veira Vigier, A. Benguerba, J. Barrault, F. Jꢀrꢁme, Green Chem. 2012, 14,
285–289.
[10] a) Y. Romꢇn-Leshkov, J. A. Dumesic, Top Catal. 2009, 52, 297–303; b) B.
Saha, M. M. Abu-Omar, Green Chem. 2014, 16, 24–38; c) F. Liu, A. Ma¨ıtꢀ,
K. De Oliveira Vigier, D. Cartigny, J.-M. Clacens, M. F. Costa Gomes,
A. A. H. Pꢇdua, F. De Campo, F. Jꢀrꢁme, Green Chem. 2013, 15, 3205–
3213.
[11] a) Y. J. Pagꢇn-Torres, T. Wang, J. M. R. Gallo, B. H. Shanks, J. A. Dumesic,
ACS Catal. 2012, 2, 930–934; b) V. Choudhary, S. H. Mushrif, C. Ho, A.
Anderko, V. Nikolakis, N. S. Marinkovic, A. I. Frenkel, S. I. Sandler, D. G.
Vlachos, J. Am. Chem. Soc. 2013, 135, 3997–4006 and references there-
in.
Keywords: 5-hydroxymethylfurfural
carbohydrates · catalysis · choline chloride
·
biomass
·
[1] a) A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411–2502;
b) G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–4098;
c) J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. Int. Ed. 2007,
46, 7164–7183; Angew. Chem. 2007, 119, 7298–7318; d) R. Rinaldi, F.
Schꢅth, Energy Environ. Sci. 2009, 2, 610–626; e) P. Gallezot, Chem. Soc.
Rev. 2012, 41, 1538–1558.
Received: July 28, 2014
Revised: September 5, 2014
Published online on && &&, 0000
[2] a) X. M. Tong, L. Y. Yang, Appl. Catal. A 2010, 385, 1–13; b) R. Rinaldi, F.
Schueth, ChemSusChem 2009, 2, 1096–1107; c) S. Van De Vyver, J. Gebo-
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 6
&
6
&
ÞÞ
These are not the final page numbers!

COMMUNICATIONS
Nice coat! Coating SiO2/AlCl3 with chol-
ine chloride is an efficient way to facili-
tate a better contact between a solid
catalyst and carbohydrates suspended
in a low boiling point organic solvent.
Under optimized conditions, the coating
of SiO2/AlCl3 with choline chloride re-
sulted in an improvement of the initial
production rate of 5-hydroxymethylfur-
fural by a factor of 12.
J. Yang, K. De Oliveira Vigier, Y. Gu,*
F. Jꢀrꢁme*
&& – &&
Catalytic Dehydration of
Carbohydrates Suspended in Organic
Solvents Promoted by AlCl3/SiO2
Coated with Choline Chloride
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 6
&
7
&
ÞÞ
These are not the final page numbers!