ported strategies. It is well established that an increase of the
fructose content generally lowers the selectivity of the reac-
tion. Hence, most works are only applicable for a fructose con-
tent ranging between 5 and 20 wt%.[3] To this end, the fruc-
tose content in ChCl was varied from 5 to 250 wt% (Figure 3).
Remarkably, 100 wt% of fructose could be dehydrated to HMF
in ChCl/CO2 without appreciable loss of yield (66%), increasing
again the interest of such a system for the eco-efficient synthe-
sis of HMF.
Having demonstrated that the ChCl/CO2 system was very
convenient for the dehydration of fructose to HMF, we next in-
vestigated the ability of this system to promote the tandem
hydrolysis/dehydration of inulin, a biopolymer of fructose, to
HMF (Table 2). In the presence of 20 wt% of inulin dissolved in
Table 2. Acid-catalyzed production of HMF from inulin (20 wt%) in chol-
ine chloride (25 g) under 4 MPa CO2 at 1208C.
This surprising tolerance of the ChCl/CO2 system to a high
content of fructose can be attributed to a change of the physi-
co-chemical properties of the ChCl/fructose mixture. As previ-
ously reported, ChCl is capable, even in the presence of water,
of forming various DESs with a wide range of hydrogen-bond
donors such as urea, polyols, or carboxylic acid. One may sus-
pect that when fructose is converted, ChCl and HMF tend to
form a new DES explaining the remarkable stability of HMF
under our conditions. Indeed, previous studies have clearly
shown that when a component is engaged in the formation of
a DES, its reactivity is drastically reduced.[12] To check this hy-
pothesis, various mixtures of ChCl and HMF were heated at
110 8C for 10 min and then cooled to room temperature. Inter-
estingly, from a HMF content of 60 wt% in ChCl, a liquid phase
was obtained at room temperature further pushing forward
the possible formation of a ChCl–HMF eutectic mixture when
working at high concentration of fructose (Figure 4).
Entry
H2O added
[mL]
t[a]
[h]
Yield of
HMF [%]
1
2
3
4
–
5
5
5
1.5
1.5
6
12
22
38
41
15
[a] Time of the experiment.
ChCl (1208C, 4 MPa CO2, 90 min), the yield of HMF was rather
disappointing (12%). This low yield can be ascribed to the low
content of water at the initial stage of the reaction that does
not favor the hydrolysis of inulin to fructose. Hence, we decid-
ed to initially add 16 wt% of water. Under these conditions,
the yield of HMF was improved up to 22%, confirming that
water was not present in a sufficient amount in the ChCl/inulin
system. Upon prolonged reaction time, the yield of HMF was
significantly increased and reached 38% after 6 h of reaction.
HMF was also recovered with 41% yield after 14 h of reaction
further demonstrating the stability of HMF in ChCl as com-
pared to other solvents. Notably, and in accordance with previ-
ous reports, this strategy was not applicable to the dehydra-
tion of glucose, mainly because carbonic acid was not capable
of isomerizing glucose to fructose, a step required prior to de-
hydration to HMF.
In conclusion, we report that fructose can be conveniently
dehydrated to HMF in a ChCl/CO2 system with a yield of up to
72%. In addition to the environmental benefits of this strategy,
we found that HMF is highly stable in the presence of ChCl,
presumably through the formation of a eutectic mixture. This
aspect allows to perform the catalytic process with a high con-
tent of fructose (up to 100 wt%) as compared to traditional
procedures, in which HMF is obtained in yields higher than
60% only from a fructose content lower than 20 wt%. Addi-
tionally, whereas after extraction HMF is unavoidably partly
contaminated with traditional Lewis and Brønsted acids, the
possible decrease of pH upon addition of CO2 allows circum-
venting this problem because carbonic acid is readily convert-
ed to CO2 and water when the CO2 pressure is released. Finally,
we have shown that this system is capable of promoting the
tandem hydrolysis/dehydration of inulin to HMF, further in-
creasing the interest of such methodology. The further applica-
tion of this reaction medium to the conversion of renewable
raw materials is the topic of current investigations in our
group.
Figure 4. Pictures of various ChCl/HMF mixtures at room temperature.
This change of physico-chemical properties can be also ob-
served during the extraction stage of HMF with MIBK. In partic-
ular, whereas HMF can be easily extracted from a ChCl/fructose
mixture (fructose content <40 wt%) or from neat water with
MIBK, the use of ultrasound was required when the fructose
content was higher than 40 wt%, supporting that a strong in-
teraction between ChCl and HMF occurred when working with
a high content of fructose. It should be noted that the more
difficult extraction of HMF when it is produced at high concen-
tration unfortunately led to a concomitant decrease of the
purity of recovered HMF. Indeed, the purity of HMF recovered
after extraction with MIBK decreased from 98% to 93% and
80% when the fructose content was increased from 20 wt% to
80 wt% and 100 wt%, respectively. This change of purity is
also consistent with the formation of a ChCl/HMF DES. Because
of the stabilization of HMF in ChCl, the extraction of impurities
competes with the extraction of HMF.
ChemSusChem 0000, 00, 1 – 4
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
3
&
ÞÞ
These are not the final page numbers!