Acid-catalyzed dehydration of fructose and inulin with glycerol or glycerol carbonate as renewably sourced co-solvent

Article abstract of DOI:10.1002/cssc.201000162

Ionic liquids (ILs) can be partially substituted by glycerol or glycerol carbonate as cheap, safe, and renewably sourced co-solvents in the acid-catalyzed dehydration of fructose and inulin to 5-hydroxymethylfurfural (HMF). In the particular case of glycerol, we found that HMF can be conveniently extracted from the IL/glycerol (65:35) mixture with methylisobutylketone, limiting the reactivity of glycerol with HMF and allowing the recovery of HMF with a high purity (95-%). Influences of the fructose content, temperature, and the nature of the ionic liquid are also discussed. The possible use of industrial-grade glycerin is also investigated. We demonstrate that by using glycerol carbonate, up to 90-wt-% of the IL can be successfully substituted, decreasing the environmental costs of traditional IL-based processes. Super subs: The amount of the ionic liquid [BMIM]Cl required for the acid-catalyzed dehydration of fructose and inulin into HMF, over Amberlyst-70 resin as solid acid catalyst, can be reduced by substituting it (up to 90-wt-%) with large amounts of glycerol or glycerol carbonate; cheap co-solvents from renewable sources.

Full text of DOI:10.1002/cssc.201000162

DOI: 10.1002/cssc.201000162
Acid-Catalyzed Dehydration of Fructose and Inulin with
Glycerol or Glycerol Carbonate as Renewably Sourced Co-
Solvent
Maud Benoit, Yoan Brissonnet, Erwan Guꢀlou, Karine De Oliveira Vigier, Joꢁl Barrault, and
[a]
FranÅois Jꢀrꢂme*
Ionic liquids (ILs) can be partially substituted by glycerol or
glycerol carbonate as cheap, safe, and renewably sourced co-
solvents in the acid-catalyzed dehydration of fructose and
inulin to 5-hydroxymethylfurfural (HMF). In the particular case
of glycerol, we found that HMF can be conveniently extracted
from the IL/glycerol (65:35) mixture with methylisobutylketone,
limiting the reactivity of glycerol with HMF and allowing the
recovery of HMF with a high purity (95%). Influences of the
fructose content, temperature, and the nature of the ionic
liquid are also discussed. The possible use of industrial-grade
glycerin is also investigated. We demonstrate that by using
glycerol carbonate, up to 90wt% of the IL can be successfully
substituted, decreasing the environmental costs of traditional
IL-based processes.
Introduction
Since fossil carbon reserves are predicted to disappear and
with growing concerns about global warming, the use of bio-
mass as raw material for energy and fine chemistry has in
recent years emerged as a fascinating and promising ap-
the IL phases with high purity. Although these IL-based pro-
cesses have allowed the isolation of HMF with high yields, the
toxicity hazards and high prices of ILs are currently major ob-
stacles, that have to be circumvented.
[
1]
proach. In this context, interest in the acid-catalyzed dehydra-
tion of hexoses to 5-hydroxymethylfurfural (HMF) has grown.
Indeed, from HMF a new generation of biofuels (e.g., dimethyl-
furan) and a wide range of fine chemicals can be obtained.
Several comprehensive review articles on the use of HMF can
One of the possible solutions for decreasing the environ-
mental and economical impact of IL-based processes involves
substituting a reasonable amount of the IL by a cheap and re-
newably sourced co-solvent. For the success of this strategy, it
is clear that a few issues need to be addressed. In particular
the choice of the renewably sourced co-solvent is crucial, be-
cause it should be (1) capable of dissolving large amounts of
carbohydrates, (2) very cheap and safe, and (3) miscible with
ILs. Obviously, the addition of a co-solvent should not have a
detrimental effect on the reaction selectivity and the catalyst
activity.
[
2]
be found in the literature.
In the acid-catalyzed dehydration of hexoses, the nature of
the solvent is of prime importance. For economical and envi-
ronmental reasons, the use of water as solvent for the produc-
[
3]
tion of HMF has been widely explored. Although very good
works have been reported, aqueous processes still suffer from
a lack of selectivity due to the possible side rehydration of
HMF to levulinic and formic acid. Dimethylsulfoxide (DMSO)
and dimethylformamide (DMF) have also been reported as sol-
Recently, we and others have shown that glycerol can be
À1
used as a cheap (0.5 Ekg ) and renewable solvent for cataly-
[8]
sis and organic chemistry. In recent reports, glycerol has been
proved to have a beneficial effect on the rate of various organ-
ic reactions, making this natural liquid polyol an attractive can-
didate for our study. Inspired by these preliminary works, we
show here that glycerol and glycerol carbonate can be used as
renewably sourced co-solvents for the acid-catalyzed dehydra-
tion of fructose and inulin into HMF. In particular, we found
that up to 90 wt% of the ILs can be substituted by glycerol
carbonate without significant alterations of the HMF yield, thus
decreasing the cost and the environmental impact of tradition-
al IL-based processes. The possible extraction of HMF from IL/
[4]
vents for the acid-catalyzed dehydration of hexoses to HMF.
These solvents have the property to dissolve hexoses while
being capable of diluting the released water, thus limiting the
HMF rehydration side reaction. However, even though HMF
yields higher than 90% have been achieved in these solvents,
the extraction of HMF, the toxicity of DMF, and the possible
formation of sulfurized products (with DMSO) remain some
major drawbacks.
Ionic liquids (ILs) are widely used in carbohydrate chemistry
owing to their unique ability to dissolve a large number of dif-
[
5]
ferent mono-, oligo-, and polysaccharides. In this context,
their use as solvent for the acid-catalyzed dehydration of hexo-
[a] M. Benoit, Y. Brissonnet, Dr. E. Guꢀlou, Dr. K. De Oliveira Vigier,
Dr. J. Barrault, Dr. F. Jꢀrꢁme
[
6]
ses has been extensively investigated. Compared to DMSO
and DMF, a few ILs are known to be poorly miscible with
Laboratoire de Catalyse en Chimie Organique, Universitꢀ de Poitiers-CNRS
4
0 avenue du recteur Pineau, 86022 Poitiers (France)
[7]
methylisobutylketone (MIBK), a recognized green solvent,
thus offering a convenient way to selectively extract HMF from
Fax: (+33)05 49 45 33 49
E-mail: francois.jerome@univ-poitiers.fr
1
304
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 1304 – 1309

Glycerol and Glycerol Carbonate as Renewably Sourced Co-Solvent
glycerol and IL/glycerol carbonate mixtures with MIBK is also
discussed.
We next checked the catalytic activity of various Brønsted
acids in different 65:35 IL/glycerol mixtures. The reactions were
performed by starting from 1 g of fructose, dissolved in 2.5 g
of solvent mixture. Regardless of which homo- or heterogene-
ous Brønsted acid was used in the [BMIM]Cl/glycerol mixture,
similar HMF yields (60–70%) were obtained after nearly 10 min
of reaction at 1108C (Table 2, entries 1–6). These results are in
Results and Discussion
In a first set of experiments, we investigated the acid-catalyzed
dehydration of fructose in 1-butyl-3-methyl imidazolium chlo-
ride ([BMIM]Cl), which is known to be an efficient medium for
[
6]
the production of HMF. In a typical procedure, 1 g of fructose
was dissolved in 2.5 g of neat [BMIM]Cl and heated at 1108C in
the presence of Amberlyst 70 (A70; 0.1 equiv H ). In agree-
Table 2. Dehydration of fructose in different IL/glycerol mixtures, cata-
lyzed by different Brønsted-acid catalysts.
[
a]
+
ment with existing literature, HMF was produced in a yield of
94% after 15 min of reaction (Table 1, entry 1). Then, [BMIM]Cl
was partially substituted by glycerol. From the results present-
[
b]
[c]
Entry
IL
Catalyst
A70
t
[min]
HMF [%]
1
2
3
4
5
6
7
8
9
[BMIM]Cl
[BMIM]Cl
[BMIM]Cl
[BMIM]Cl
[BMIM]Cl
[BMIM]Cl
[BMIM]Cl
[BMIM]Cl
[MDIM]PF6
[HPYR]Cl
[HPYR]Br
10
5
70
62
69
68
62
66
<10
66
<5
64
41
H
2
SO
4
Table 1. Acid-catalyzed dehydration of fructose in [BMIM]Cl/glycerol mix-
tures.
CF
3
SO
3
H
5
HCl
A15
10
5
[
d]
[
e]
Carb-SO
3
H
10
10
10
10
10
10
[
f]
A70
A70
A70
A70
A70
[
g]
[
b]
[c]
Entry
Amount of
BMIM]Cl [wt%]
Amount of
glycerol [wt%]
t
HMF
[%]
[
[min]
10
11
1
2
3
100
80
65
65
65
65
65
50
20
0
20
35
35
35
35
35
50
80
15
10
8
10
40
210
12
8
94
82
70
71
60
52
72
42
7
[
1
a] Conditions: 1 g fructose dissolved in 2.5 g of solvent mixture (65:35),
10 8C, 10 mol% of H . [b] Time required to reach the maximum HMF
yield. [c] Determined by HPLC. [d] H -exchange capacity 4.70 mmolg
e] Vulcan was used as carbonaceous support; H -exchange capacity
+
[
[
[
[
d]
e]
f]
+
À1
4
5
6
7
8
9
.
+
[
0
À1
.58 mmolg . [f] Reused in a second catalytic cycle. [g] Reused after reac-
g]
tivation with HCl.
2
[a] Conditions: 1 g of fructose dissolved in 2.5 g of a [BMIM]Cl/glycerol
mixture. [b] Time required to reach the maximum HMF yield. [c] Deter-
mined by HPLC. [d] In biphasic [BMIM]Cl/glycerol–MIBK. [e] 908C. [f] 708C.
perfect agreement with recent reports by Schꢄth et al. and
Corma et al., who independently reported that a release of
protons occurs when using [BMIM]Cl and Brønsted acids (i.e.,
cation exchange between the imidazolium moiety and the
proton, leading to the liberation of HCl in the reaction
[g] Glycerin (80 wt% glycerol, 15 wt% water, 5 wt% soap) was used.
[10]
ed in Table 1, it appears that substituting [BMIM]Cl by glycerol
leads to a faster fructose conversion rate, but lowers the HMF
yield. The HMF yield remained acceptable up to a glycerol-for-
medium). This hypothesis was further confirmed by recycling
the A70 catalyst. Indeed, when A70 was reused in a fresh
[BMIM]Cl/glycerol mixture, a HMF yield of less than 10% was
obtained (Table 2, entry 7). After treatment of the used A70
with a solution of HCl the A70 recovered its initial activity, fur-
ther evidencing the leaching of protons during the catalytic
process (entry 8).
[
BMIM]Cl substitution level of 35 wt%, (70%; entry 3); however,
when using a higher glycerol content the HMF yield dropped
to unacceptable levels (entries 8–9). More details about the re-
activity of glycerol in our reaction are provided in the follow-
ing paragraphs. On the basis of these first results, the 65:35
In the literature, [BMIM]Cl is considered as one of the best
solvents for the acid-catalyzed dehydration of hexoses, but
never in the presence of a co-solvent such as glycerol. We se-
lected other ILs in combination with glycerol as a co-solvent in
this study: 1-methyl-3-dodecyl imidazolium hexafluorophos-
phate ([MDIM]PF6), hexylpyridinium chloride ([HPYR]Cl), and
hexylpyridinium bromide ([HPYR]Br). As reported in the litera-
ture, chloride-based ILs gave the highest HMF yields (Table 2,
[
BMIM]Cl/glycerol mixture was selected as medium for the ex-
periments described in this manuscript.
With the aim of minimizing the energy consumption of our
process, we then decreased the reaction temperature. As ex-
pected, a decrease of the reaction temperature from 1108C to
08C required an increase of the reaction time, from 10 to
0 min. In this case, the selectivity of the process slightly de-
9
4
creased: HMF was produced in a yield of 60% yield (vs 70% at
10 8C, Table 1, entries 3 and 5). The reaction could even be
carried out at 708C, but at the expense of the reaction selectiv-
entries 1 and 10). In the [HPYR]Br/glycerol and [MDIM]PF /glyc-
6
1
erol mixtures, the HMF yields dropped to 41% and less than
5%, respectively (entries 9 and 11). These lower yields are con-
sistent with a release of protons in the reaction medium.
ity as a maximum HMF yield of 52% was obtained after
210 min (entry 6).
Indeed, HBr and HPF are more acidic than is HCl; therefore,
6
ChemSusChem 2010, 3, 1304 – 1309
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
1305

F. Jꢀrꢁme et al.
cation exchange between imidazolium and the supported
proton is much less favorable with PF - and bromide-based ILs
6
than with chloride-based ILs. Thus, it is reasonable to conclude
that our reaction is mainly catalyzed by the released HCl.
We also performed viscosity measurements because, like
[BMIM]Cl, glycerol is a highly viscous liquid and mass-transfer
problems may occur. The results are shown in Figure 1. Inter-
Figure 2. Influence of the fructose content on the HMF yield. Conditions:
+
2
.5 g of [BMIM]Cl/glycerol (65:35), 1108C, and A70 (0.1 equiv H ).
initial amount of fructose, analysis of the reaction progress by
HPLC clearly showed that nearly 20 mol% of glycerol was con-
sumed within 10 min (Figure 2). Like fructose, glycerol is a nat-
ural polyol that can be dehydrated under acidic conditions, or
can react either with fructose, HMF, or intermediates. Indeed,
compared to neat [BMIM]Cl the formation of few unidentified
secondary products was detected by HPLC. The presence of
these secondary products made the purification of HMF at the
end of the reaction rather complex.
Figure 1. Viscosity of a) neat glycerol, b) neat [BMIM]Cl, c) [BMIM]Cl/glycerol
(65:35), d) [BMIM]Cl/glycerol (65/35)+1 g of fructose, and e) [BMIM]Cl/glyc-
erol (65/35)+1 g of fructose after 8 min of reaction in the presence of A70.
Viscosity measurements were performed at 1108C.
To get more insight on the reactivity of glycerol a few coun-
ter experiments were undertaken. First, a [BMIM]Cl/glycerol
mixture (65:35) was heated, without fructose and Amber-
lyst 70, to 1108C. Glycerol was not consumed in these condi-
tions, confirming the stability of glycerol in [BMIM]Cl. Next, a
[BMIM]Cl/glycerol (65:35) mixture was heated to 1108C in the
presence of Amberlyst 70, but without fructose, for 60 min.
Again, no reaction took place since 100% of the glycerol was
recovered, showing that glycerol is stable in these conditions.
It should be noted that when Amberlyst 70 is used as solid cat-
alyst oligomerization or dehydration of glycerol may occur, but
only after a prolonged reaction time or at a reaction tempera-
ture higher than 1108C. Therefore, any consumption of glycer-
ol at this stage can be ascribed to a side reaction with fructose
or HMF, or intermediates.
estingly, it appears that the viscosity of glycerol (99.9% purity)
at 1108C is five times lower than that of neat [BMIM]Cl
(41.7 MPas for [BMIM]Cl vs. 7.5 MPas for glycerol). The substi-
tution of 35 wt% of [BMIM]Cl by glycerol decreased the viscos-
ity of the reaction medium at 1108C from 41.7 to 31.8 MPas.
When 1 g of fructose was dissolved in 2.5 g [BMIM]Cl/glycer-
ol (65:35), the viscosity of the reaction medium increased from
31.8 to 53.0 MPas. Even if at 1108C the viscosity of the reaction
medium is initially high, it does not really impact the reaction
progress because during the reaction the viscosity rapidly
drops due to the release of water (3 mol water per fructose),
which rapidly makes the reaction medium more “fluid.” For ex-
ample, after addition of the Amberlyst 70 resin the viscosity of
the reaction medium dropped from 53.0 to 18.0 MPas within
8
min (corresponding to the time required to reach the maxi-
Similar to [BMIM]Cl, glycerol is not miscible with MIBK.
Therefore, we explored the possibility of converting fructose
into HMF in biphasic [BMIM]Cl/glycerol (65:35)–MIBK. Such a
methodology is expected to (1) limit the possible reactivity of
glycerol with HMF, and (2) allow the convenient and selective
recovery of HMF from the [BMIM]Cl/glycerol (65:35) mixture.
Such a strategy has proved to be efficient especially for the
mum HMF yield; Table 1, entry 3).
Because glycerol is capable of dissolving large amounts of
fructose, we investigated the influence of the fructose content
on the HMF yield. As shown in Figure 2, at 1108C, a decrease
of the amount of fructose dissolved in 2.5 g of [BMIM]Cl/glyc-
erol (65:35), from 1 g to 0.5 and 0.25 g, did not affect the HMF
yield. However, an increase of the fructose content from 1 g to
[9]
acid-catalyzed dehydration of fructose to HMF in water.
5
g led to a drop of the HMF yield, from 67% to 50%. A highly
concentrated solution of fructose can be used in the 65:35
BMIM]Cl/glycerol mixture (up to 9 g in 2.5 g). However, with 7
As shown in Figure 3, a continuous extraction of HMF with
MIBK suppressed the consumption of glycerol. Whereas glycer-
ol was continuously consumed without assistance of MIBK, we
found that no glycerol was consumed in biphasic [BMIM]Cl/
glycerol–MIBK (Figure 3). This result shows that (1) the side
consumption of glycerol can be ascribed to its reaction with
HMF, and (2) side reactions between glycerol and fructose are
negligible. It should be also noted that the HMF yields ob-
tained with or without assistance of MIBK are similar, showing
that the side reaction between glycerol and HMF is not a dom-
[
and 9 g of fructose the selectivity of the process dropped
owing to the formation of insoluble black material (presumably
humins), and in these cases the maximum HMF yields was
35% and 27%, respectively.
We next investigated the reactivity of glycerol in our pro-
cess, in order to clarify why a high glycerol content (>35 wt%)
is detrimental to the selectivity of the reaction. Whatever the
1
306
www.chemsuschem.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 1304 – 1309

Glycerol and Glycerol Carbonate as Renewably Sourced Co-Solvent
it is clear that the direct use of glycerin is much more desirable
from the viewpoint of green chemistry. To this end, we investi-
gated the acid-catalyzed dehydration of fructose to HMF in a
[
BMIM]Cl/glycerin 65:35 mixture. Typically, 1 g of fructose was
dissolved in 2.5 g of [BMIM]Cl/glycerin 65:35 and heated to
+
1
10 8C in the presence of A70 (0.1 equiv H ). The glycerin used
here was kindly provided by Valagro and comes from a biodie-
sel unit located in the western part of France (Chalandray).
À1
This glycerin (0.3 Ekg ) is brown-colored and contains
80 wt% glycerol, 15 wt% water, and 5 wt% of soap. Using
glycerin, we found that a 72% yield of HMF was produced
after 12 min of reaction, showing that glycerin is also eligible
for use in our process (Table 1, entry 7).
Figure 3. Conversion of glycerol in [BMIM]Cl/glycerol (65:35) and in biphasic
[
BMIM]Cl/glycerol–MIBK [results collected from 5 g fructose, 2.5 g [BMIM]Cl/
+
glycerol (65:35), 1108C, and A70 (0.1 equiv H )].
Although the use of glycerin as a renewably sourced co-sol-
vent is indeed attractive, the extraction of HMF is more prob-
lematic. Indeed, when the catalytic reaction was performed in
biphasic [BMIM]Cl/glycerin (65:35)–MIBK, the purity of the re-
covered HMF was lower compared to the use of pharmaceuti-
cal-grade glycerol, owing to the concomitant extraction of im-
purities initially present in glycerin.
inating reaction (at the initial stage of our process) as com-
pared to the dehydration of fructose to HMF (Table 1, entries 3
and 4).
It should be noted that, starting from 5, 7, and 9 g of fruc-
tose, the use of MIBK as an extraction solvent not only allowed
to inhibit the reactivity of glycerol, but also the recovery of
HMF with a very high purity (95%), thereby considerably sim-
With the aim of further limiting the dependency of our pro-
cess on [BMIM]Cl, we moved on to the possible utilization of
glycerol carbonate as a renewably sourced co-solvent. Glycerol
carbonate is readily available from glycerol by reaction with
1
plifying the work-up procedure. Figure 4 shows HNMR spectra
[11]
[12]
other renewable raw materials, such as CO2 or urea. Glyc-
erol carbonate can be also prepared by transcarbonatation
[13]
with diethylcarbonate, an ecofriendly carbonatation agent.
Like glycerol, glycerol carbonate is cheap and recognized as a
green solvent, but to date examples of its use as a green sol-
vent for catalysis remain scarce. As compared to glycerol, the
presence of only one hydroxyl group was expected to limit its
reactivity with HMF.
Similar to the experiments described above with glycerol,
[
BMIM]Cl was partially substituted by glycerol carbonate and
the yield of HMF was monitored by HPLC. As expected, up to
0 wt% of [BMIM]Cl could be substituted. In these conditions,
8
HMF was produced with a yield higher than 75% (Table 3, en-
tries 1–4). When the catalytic process was performed in the
presence of MIBK, the glycerol carbonate content could be fur-
ther increased from 80 to 90 wt%. For example, in biphasic
1
Figure 4. HNMR spectra (300 MHz, 258C, CDCl
3
) of HMF recovered with
MIBK from [BMIM]Cl/glycerol (65:35) [obtained from 5, 7, or 9 g of fructose
dissolved in 2.5 g of [BMIM]Cl/glycerol (65:35)].
[
BMIM]Cl/glycerol carbonate (10:90)–MIBK, the HMF yield was
[14]
increased from 60% to 70% (Table 3, entries 5 and 6). Simi-
larly, when the amount of fructose was decreased from 1 g to
0.5 g, a HMF yield of 72% was obtained in [BMIM]Cl/glycerol
carbonate (10:90) (Table 3, entry 7). These last experiments
show that 90 wt% of the [BMIM]Cl could be replaced by glyc-
erol carbonate without a dramatic effect on the HMF yield,
thus considerably limiting the dependence of our process on
[BMIM]Cl.
of the recovered HMF and confirms the purity of the recovered
HMF (only the remaining MIBK was detected as a contami-
nant). Considering that chloride-based ILs provide better re-
sults, [BMIM]Cl is more attractive than [HPYR]Cl because of its
very low solubility in MIBK, allowing the isolation of high-
purity HMF. Indeed, when using [HPYR]Cl a significant contami-
nation of HMF with this IL occurred during the extraction
stage with MIBK.
In addition, whereas after total consumption of fructose
HMF was consumed in the [BMIM]Cl/glycerol (65:35) mixture,
mainly due to side reactions with glycerol, the HMF yield re-
mained stable in the [BMIM]Cl/glycerol carbonate mixture
(65:35), demonstrating the greater stability of HMF in glycerol
carbonate (Figure 5). All attempts to dehydrate fructose to
HMF in neat glycerol carbonate failed, and only a 10% yield of
HMF was obtained in these conditions (Table 3, entry 8).
The glycerol used in this study has a purity of 99.9% (phar-
maceutical grade). However, glycerol produced from the man-
ufacture of biodiesel (named glycerin) is actually a mixture of
glycerol, water, and soap (stemming from the catalyst neutrali-
zation step). Even though pharmaceutical-grade glycerol is
À1
cheap (0.5 Ekg ), the purification of industrially produced
glycerin is a costly and energy-consuming process. Therefore,
ChemSusChem 2010, 3, 1304 – 1309
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
1307

F. Jꢀrꢁme et al.
Table 3. Acid-catalyzed dehydration of fructose in [BMIM]Cl/glycerol car-
[
a]
bonate mixtures.
[
b]
[c]
Entry
Amount of
BMIM]Cl [wt%]
Amount of
glycerol carbonate [wt%]
t
HMF
[%]
[
[min]
1
2
3
4
5
100
65
50
20
10
10
10
0
0
35
50
80
90
90
90
100
15
35
30
25
20
30
40
20
95
98
91
75
60
70
72
10
[
d]
e]
6
7
8
[
[a] 1 g of fructose dissolved in 2.5 g of a [BMIM]Cl/glycerol carbonate mix-
ture. [b] Time required to reach the maximum HMF yield. [c] Determined
by HPLC. [d] In the presence of MIBK. [e] Starting from 0.5 g of fructose
dissolved in 2.5 g of a [BMIM]Cl/glycerol carbonate mixture.
Scheme 1. Acid-catalyzed production of HMF from inulin in [BMIM]Cl/glycer-
ol carbonate (10:90): 0.5 g inuline, Amberlyst 70 (0.1 equiv H ), 2.5 g
[BMIM]Cl/glycerol carbonate (10:90), 1108C.
+
lar case of glycerol, its reactivity with HMF can be circumvent-
ed by the addition of MIBK. In this case side reactions involving
glycerol are not only suppressed but HMF can be also conven-
iently recovered with a purity close to 95%, considerably sim-
plifying the work-up procedure.
Experimental Section
Chemicals
Fructose was purchased from Sigma-Aldrich. Glycerol (purity
9
9.9%) was kindly provided by Stꢀarinerie-Dubois. Glycerol carbon-
ate and [BMIM]Cl were synthesized as described in Refs. [13] and
18], respectively. Amberlyst 70 was provided by Rohm&Haas and
Figure 5. Acid-catalyzed dehydration of fructose to HMF in [BMIM]Cl/glycerol
[
(
65:35) and [BMIM]Cl/glycerol carbonate (65:35).
used as-received. Amberlyst 70 is a macroporous polystyrenic-type
À1
resin containing 2.55 mmolg sulfonic groups, has a water con-
2
À1
tent of 53–59 wt%, a surface area of 36 m g , and a particle size
of 0.5 mm. Viscosity measurements were collected on a Rotational
Viscometer “VISCO ELITE L” from Fungilab S.A.
Finally, we investigated the acid-catalyzed dehydration of
inulin in [BMIM]Cl/glycerol carbonate (10:90) to show the ver-
satility of our approach (Scheme 1). Inulin is a biopolymer of
fructose, extracted from chicory or dahlia tubers. Direct pro-
duction of HMF from inulin is even more complex and first re-
quires a hydrolysis of inulin to fructose, followed by dehydra-
Analysis
[
15,16]
The amounts of HMF were calculated by external calibration at
tion of fructose to HMF.
The catalyst Amberlyst 70 contains
2
58C using a HPLC equipped with a nucleosil 100–5 C18 column
5
4–59 wt% of water and was used here without any drying in
(250ꢅ4.6 mm), a Varian Prostar UV detector (320 nm), Varian Pros-
order to promote the hydrolysis of inulin to fructose. Using the
same procedure as that described above for the acid-catalyzed
dehydration of fructose, a 60% yield of HMF was obtained at
tar pumps (model 210), using acetonitrile/water (10:90) as mobile
phase (0.8 mLmin ). Fructose and glycerol were quantified by ex-
ternal calibration at 258C using a HPLC equipped with a Varian
À1
1
10 8C from inulin in [BMIM]Cl/glycerol carbonate (10:90),
NH -column, a Shimadzu LC-20AT pump, a Shimadzu RID-10A de-
2
[
17]
tector, and acetonitrile/water (80:20) as mobile phase at
showing the convenience of our methodology.
À1
0
.8 mLmin .
Conclusions
General procedure for acid-catalyzed dehydration
We have shown that glycerol and glycerol carbonate can be
used as cheap and renewably sourced co-solvents for the acid-
catalyzed dehydration of fructose and inulin into HMF. In par-
ticular, when using glycerol carbonate we found that up to
Fructose (0.5 or 1 g) was dissolved in 2.5 g of a mixture [BMIM]Cl/
glycerol or [BMIM]Cl/glycerol carbonate. Then the mixture was
heated under air at 1108C in the presence of Amberlyst 70
+
(
[
0.1 equiv H ). All attempts to recycle the [BMIM]Cl/glycerol or
BMIM]Cl/glycerol carbonate medium failed. Indeed, even if an
90 wt% of [BMIM]Cl can be substituted without affecting the
HMF yield, thus considerably limiting the costs and environ-
mental impact of traditional IL-based processes. In the particu-
HMF yield of nearly 70% was produced in such media, the remain-
ing 30% was mainly composed of black soluble and insoluble ma-
1
308
www.chemsuschem.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 1304 – 1309

Glycerol and Glycerol Carbonate as Renewably Sourced Co-Solvent
terials, making the recycling of the medium almost impossible. The
same phenomenon was encountered in neat [BMIM]Cl, for which
cumbersome and energy-consuming purification processes were
required.
Note: In the mixture [BMIM]Cl/glycerol (65:35), the reaction could
be scaled up without significant change of yield. For example,
starting from 8 g of fructose dissolved in 20 g of [BMIM]Cl/glycerol
[4] a) A. Takagaki, M. Ohara, S. Nishimura, K. Ebitani, Chem. Commun. 2009,
276–6278; b) A. S. Amarasekara, L. D. Williams, C. C. Ebede, Carbohydr.
6
Res. 2008, 343, 3021–3024; c) X. Qi, M. Watanabe, T. M. Aida, R. L.
Smith, Ind. Eng. Chem. Res. 2008, 47, 9234–9239; d) X. Tong, Y. Li, Chem-
SusChem 2010, 3, 350–355.
[
5] a) Q. Liu, M. H. A. Janssen, F. van Rantwijk, R. A. Sheldon, Green Chem.
2
005, 7, 39–42; b) O. A. El Seoud, A. Koschella, L. C. Fidale, S. Dorn, T.
Heinze, Biomacromolecules 2007, 8, 2629–2647; c) S. Murugesan, R. J.
Linhardt, Curr. Org. Synth. 2005, 2, 437–451.
(
65:35), a 69% yield of HMF was obtained after 15 min of reaction
+
at 1108C in the presence of Amberlyst 70 (0.1 equiv H ).
[6] a) G. Yong, Y. Zhang, J. Y. Ying, Angew. Chem. 2008, 120, 9485–9488;
Angew. Chem. Int. Ed. 2008, 47, 9345–9348; b) H. Zhao, J. E. Holladay, H.
Brown, Z. C. Zhang, Science 2007, 316, 1597–1599; c) X. Qi, M. Wata-
nabe, T. M. Aida, R. L. Smith, ChemSusChem 2009, 2, 944–946; d) S. Hu,
Z. Zhang, Y. Zhou, B. Han, H. Fan, W. Li, J. Song, Y. Xie, Green Chem.
General procedure for the acid-catalyzed dehydration of
fructose in biphasic [BMIM]Cl/glycerol (65:35)–MIBK
2
008, 10, 1280–1283; e) C. Moreau, A. Finiels, L. Vanoye, J. Mol. Catal.
1
0 mL of MIBK was added to the reaction mixture. Below 10 mL of
A: Chem. 2006, 253, 165–169; f) X. Qi, M. Watanabe, T. M. Aida, R. L.
Smith, Green Chem. 2009, 11, 1327–1331; g) T. S. Stꢆhlberg, M. G. Soren-
sen, A. Riisager, Green Chem. 2010, 12, 321–325; h) S. Hu, Z. Zhang, J.
Song, Y. Zhou, B. Han, Green Chem. 2009, 11, 1746–1749; i) Q. Bao, K.
Qiao, D. Tomida, C. Yokoyama, Catal. Commun. 2008, 9, 1383–1388;
j) C. Lansalot-Matras, C. Moreau, Catal. Commun. 2003, 4, 517–520.
7] The European Medicines Agency has classified MIBK as a solvent with
low toxic potential, in the same range as ethanol. Document reference:
CVMP/VICH/502–99/FINAL. Available from http://www.emea.europa.eu/
pdfs/vet/vich/050299en.pdf (accessed September 2010).
MIBK, the HMF extraction from the glycerol phase was not com-
plete (e.g., isolated yield of HMF with only 5 mL of MIBK=30%).
The HMF/MIBK phase was then recovered by simple phase decant-
ation, and MIBK was separated from HMF by distillation using a
rotary evaporator. Because the extraction with MIBK is highly selec-
tive to HMF, the distilled MIBK could be recycled for other extrac-
tion cycles. It should be noted that because glycerol is insoluble in
most volatile organic solvents, we also tried to use ethyl acetate
for the extraction of HMF from the [BMIM]Cl/glycerol phase. How-
ever, in this case the extraction was much less selective than with
MIBK and HMF was recovered with a lower purity.
[
[8] For a recent review, see Y. Gu, F. Jꢀrꢂme, Green Chem. 2010, 12, 1127–
138, and references cited therein.
1
[
9] a) C. Carlini, M. Giuttari, A. M. Raspolli Galletti, G. Sbrana, T. Armaroli, G.
Busca, Appl. Catal. A: Gen. 1999, 183, 295–302; b) F. Benvenuti, C. Carli-
ni, P. Patrono, A. M. Raspolli Galletti, G. Sbrana, M. A. Massucci, P. Galli,
Appl. Catal. A: Gen 2000, 193, 147–153; c) C. Moreau, R. Durand, S. Razi-
gade, J. Duhamet, P. Faugeras, P. Rivalier, P. Ros, G. Avignon, Appl. Catal.
A: Gen. 1996, 145, 211–224; d) C. Moreau, R. Durand, C. Pourcheron, S.
Razigade, Ind. Crops Prod. 1994, 3, 85–90.
General procedure for the acid-catalyzed conversion of
inulin in [BMIM]Cl/glycerol carbonate (10:90)
0
.5 g of inulin was dissolved in 2.5 g [BMIM]Cl/glycerol carbonate
[
[
[
10] a) R. Rinaldi, N. Meine, J. vom Stein, R. Palkovits, F. Schꢄth, ChemSu-
(
10:90). The resulting solution was heated in air at 1108C in the
sChem 2010, 2, 266–276, b) N. Villandier, A. Corma, Chem. Commun.
+
presence of A70 (0.1 equiv H ).
2
010, 46, 4408–4410.
11] a) J. George, Y. Patel, M. Pillai, P. Munshi, J. Mol. Catal. A: Chem. 2009,
04, 1; b) M. Aresta, A. Dibenedetto, F. Nocito, C. Pastore, J. Mol. Catal.
3
Acknowledgements
A: Chem. 2006, 257, 149.
12] a) J. W. Yoo, Z. Mouloungui, Stud. Surf. Sci. Catal. 2003, 146, 75; b) S.
Claude, Z. Mouloungui, J.-W. Yoo, A. Gaset, US Pat. US6025504, 2000;
c) M. J. Climent, A. Corma, P. De Frutos, S. Iborra, M. Noy, A. Velty, P.
Conception, J. Catal. 2010, 269, 140–149; d) M. Aresta, A. Dibenedetto,
F. Nocito, C. Ferragina, J. Catal. 2009, 268, 106–114.
The authors are grateful to the CNRS and the French Ministry of
Research for financial support. M.B. also thanks the “Agence Na-
tionale de la Recherche” (ANR-09-CP2D-18–01) for the funding of
her PhD. The authors acknowledge Rodolphe De Sousa for pro-
viding the glycerol carbonate used in this study.
[13] a) Y. Patel, J. George, S. Muthukumaru, P. Munshi, Green Chem. 2009, 11,
056–1060; b) N. Pasquier, H. Keul, E. Heine, M. Moeller, Biomacromole-
1
cules 2007, 40, 2874–2882; c) N. Fricke, H. Keul, M. Mçller, Macromol.
Chem. Phys. 2009, 210, 242–255; d) H. R. Pfaendler, F. X. Muller, Synthe-
sis 1992, 350.
Keywords: carbohydrates · catalysis · ionic liquids · renewable
resources · solvent effects
[
14] Contrary to what was observed in the [BMIM]Cl/glycerol, the HMF re-
covered by extraction with MIBK was, in this case, contaminated with
glycerol carbonate.
[
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. H. Clark, Green Chem. 2006, 8, 17–21; d) R. Rinaldi, F. Schꢄth, Chem-
SusChem 2009, 2, 1096–1107; e) P. Gallezot, ChemSusChem 2008, 1,
[
[
[
[
15] F. Ilgen, D. Ott, D. Kralisch, C. Reil, A. Palmberger, B. Kçnig, Green Chem.
2
009, 11, 1948–1954.
16] For a selected recent example, seea) S. Hu, Z. Zhang, Y. Zhou, J. Song,
H. Fan, B. Han, Green Chem. 2009, 11, 873–877.
17] In agreement with previous works, a very low HMF yield (<5%) was
observed starting from glucose.
18] J. G. Huddleston, H. D. Willauer, R. P. Swatloski, A. E. Visser, R. D. Rogers,
Chem. Commun. 1998, 1765–1766.
734–737; f) J. Zakzeski, P. C. A. Bruijnincx, A. L. Jongerius, B. M. Weck-
huysen, Chem. Rev. 2010, 110, 3552–3599; g) J. J. Bozell, G. R. Petersen,
Green Chem. 2010, 12, 539–554.
2] Selected reviews: a) J. Lewkowski, Arkivoc 2001, 17–54; b) K. De Oliveira
Vigier, F. Jꢀrꢂme, Top. Curr. Chem. 2010, 1–30.
[
[
3] a) X. Qi, M. Watanabe, T. M. Aida, R. L. Smith, Green Chem. 2008, 10,
799–805; b) F. S. Asghari, H. Yoshida, Carbohydr. Res. 2006, 341, 2379–
2387; c) F. S. Asghari, H. Yoshida, Ind. Eng. Chem. Res. 2006, 45, 2163–
2173; d) X. Qi, M. Watanabe, T. M. Aida, R. L. Smith, Catal. Commun.
Received: June 10, 2010
2008, 9, 2244–2249; e) J. N. Chheda, Y. Roman-Leshkov, J. A. Dumesic,
Revised: July 15, 2010
Green Chem. 2007, 9, 342–350.
Published online on September 30, 2010
ChemSusChem 2010, 3, 1304 – 1309
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
1309