Dual catalytic function of 1,3-dialkylimidzolium halide ionic liquid on the dehydration of fructose to 5-hydroxymethylfurfural

Article abstract of DOI:10.1016/j.catcom.2012.03.005

In the fructose dehydration to 5-hydroxymethylfurfural (HMF) in dimethyl sulfoxide, the catalytic function of 1,3-dialkylimidazolium-based ionic liquids with different counter-anions was investigated. It was found that the effects of alkyl chain length and additional alkyl group in Cl-containing ionic liquids were negligible whereas the activity of 1-butyl-3-methylimidazolium-based ionic liquids was considerably changed by varying the anion (particularly, halide ions). The latter finding was confirmed by the control experiment result that the addition of KBr or Ca(CH₃COO)₂ dramatically tuned the catalytic activity of 1-butyl-3-methylimidazolium chloride. From these results and the proposed reaction pathway, it was consequently believed that 1,3-dialkylimidazolium halide contained the dual catalytic function to act as both Br?nsted acid and nucleophile for the fructose dehydration to HMF.

Full text of DOI:10.1016/j.catcom.2012.03.005

Catalysis Communications 24 (2012) 11–15
Contents lists available at SciVerse ScienceDirect
Catalysis Communications
journal homepage: www.elsevier.com/locate/catcom
Short Communication
Dual catalytic function of 1,3-dialkylimidzolium halide ionic liquid on the
dehydration of fructose to 5-hydroxymethylfurfural
Jihye Ryu a,b, Jae-Wook Choi a, Dong Jin Suh a, Dong June Ahn b, Young-Woong Suh c,
⁎
a
Clean Energy Research Center, Korea Institute of Science and Technology, Seoul 136-701, Republic of Korea
Department of Chemical and Biological Engineering, Korea University, Seoul 131-701, Republic of Korea
b
c
Department of Chemical Engineering, Hanyang University, Seoul 133-791, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
In the fructose dehydration to 5-hydroxymethylfurfural (HMF) in dimethyl sulfoxide, the catalytic function of
1,3-dialkylimidazolium-based ionic liquids with different counter-anions was investigated. It was found that
the effects of alkyl chain length and additional alkyl group in Cl-containing ionic liquids were negligible whereas
the activity of 1-butyl-3-methylimidazolium-based ionic liquids was considerably changed by varying the anion
(particularly, halide ions). The latter finding was confirmed by the control experiment result that the addition of
KBr or Ca(CH3COO)2 dramatically tuned the catalytic activity of 1-butyl-3-methylimidazolium chloride. From
these results and the proposed reaction pathway, it was consequently believed that 1,3-dialkylimidazolium
halide contained the dual catalytic function to act as both Brönsted acid and nucleophile for the fructose dehy-
dration to HMF.
Received 9 November 2011
Received in revised form 8 February 2012
Accepted 1 March 2012
Available online 9 March 2012
Keywords:
Fructose
Dehydration
HMF
Alkylimidazolium-based ionic liquid
Halide
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
N,N-dimethylacetamide (DMA) containing LiCl (DMA–LiCl) was as ef-
fective as 1-ethyl-3-methylimidazolium chloride for HMF synthesis.
In order to reduce the dependence of fossil fuels as well as to limit
global warming caused by emission of greenhouse gasses, it is worth-
while developing highly energy-efficient conversion of cellulosic
biomass into renewable liquid fuels and chemicals [1–3]. Among a va-
riety of biomass-derived chemicals, 5-hydroxymethylfurfural (HMF)
is a key intermediate to produce high energy-content fuels (e.g.,
2,5-dimethylfuran) [4] and chemicals (e.g., 2,5-furandicarboxylic
acid as a replacemnt of terephthalic, isophthalic and adipic acids)
[5]. Thus, a great deal of attention has been paid to the dehydration
of fructose using efficient catalytic systems coupled with a down-
stream separation process [6–9].
Since the dehydration reaction proceeds under an acidic condi-
tion, common mineral acids (H2SO4, H3PO4 and HCl) and solid acids
(ion-exchage resins, sulfated zirconia, etc.) are used in many previous
studies [6]. In recent years, various ionic liquids (ILs) have been ap-
plied for the dehydration of sugars, where ILs are utilized as a small
amount of catalyst and/or a reaction medium [10–16]. Furthermore,
the report of Zhang and coworkers [17] stimulated numerous studies
to address ILs' potential for the production of HMF, and also opened
the attempt to search for new solvents that can replace ILs due to
their high cost. As an example, Binder and Raines [16] reported that
Taking a glance at recent results regarding the conversion of fruc-
tose to HMF, we wondered the detailed reaction pathway on the
fructose-to-HMF conversion using imidazolim-based ILs as a catalyst.
As an acidic contribution to the reaction, the proton in imidazolium
cations has a well-known Brönsted acidity. Nevertheless, several ILs
with more and/or higher acidic characters have been developed in
order to increase the reaction rate; for example, Moreau et al. [18]
reported that 1-H-3-methylimidazolium chloride produced HMF in
a yield of 92% within 15–45 min. On the other hand, previous studies
were also conducted to vary ILs' anionic counterpart, where its
nuclephilicity was adjusted or additional acidity was introduced. In
the report by Binder and Raines [16], the effect of halide ions on the
synthesis of HMF from fructose was intensively investigated with
H2SO4 and imidazolum-based ILs being the catalyst and the additive,
respectively. Despite these abundant results, alkylimidzolium-based
ionic liquids may have intrinsic dehydration activities. In other
words, ILs' constituents, cation and anion, may have each specific
contribution to the fructose dehydration reaction.
Thus, the present work has been focused on revealing the catalytic
capability of imidazolium-based ionic liquids on the dehydration of
fructose to HMF. For comparison, Nafion NR50, acetic acid and CaCl2
were used as a catalyst. Then, the effect of chloride-containing cations
with different hydrocarbon substituents attached to the imidazolium
ring was investigated for the reaction. In addition, the effect of the
anion of imidazolium-based ILs was determined with Cl−, Br−, BF4−
PF6−, CH3SO4− and acetate (OAc−). To confirm the anion's effect, the
,
⁎
Corresponding author. Tel.: +82 2 2220 2329; fax: +82 2 2298 4101.
E-mail address: ywsuh@hanyang.ac.kr (Y.-W. Suh).
1566-7367/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.catcom.2012.03.005

12
J. Ryu et al. / Catalysis Communications 24 (2012) 11–15
control experiment was conducted by adding several promoters or
inhibitors to the reaction mixture containing an imidazolium-based
chloride ionic liquid. From these results, the reaction network was de-
scribed on the fructose dehydration and a discussion was formulated
for the roles of the cation and anion of 1,3-dialkylimidzolium halide.
3. Results and discussion
3.1. Catalytic activity of imidazolium-based ionic liquids
First of all, the catalytic activity of C4mimCl was compared with
those of Nafion NR50, CH3COOH and CaCl2 on the dehydration of
fructose into HMF (Table 1). Among the catalysts, Nafion NR50
showed the highest catalytic performance, where the HMF yield
approached ca. 80% at the reaction time of 5 h. The catalyst,
C4mimCl, also converted fructose into HMF at the considerable yield
of 56%. However, acetic acid was believed to have a negligible activity,
because the HMF yield was very similar to that obtained in the blank
test. On the other hand, it should be noted that the activity of CaCl2
was slightly lower than that of C4mimCl, but higher about two-fold
than that of acetic acid. This suggests that the chloride ion acts as
the catalyst for the conversion of fructose into HMF, which is in
agreement with the results of Binder and Raines [16]. From the
above results, it was assumed that both C4mim+ and Cl− may be
responsible for the fructose-to-HMF conversion.
2. Experimental
2.1. Materials
Fructose (Sigma, ≥99%), dimethyl sulfoxide (DMSO, Junsei,
≥99.89%, water max. 0.2%), Nafion NR50 (Wako), acetic acid (Yakuri,
99.7%), CaCl2 (Junsei, ≥95.0%), and Ca(CH3COO)2·xH2O (Aldrich,
99%) were used without purification.
1-butyl-3-methylimidazolium chloride (C4mimCl, >98%),
1-butyl-3-methylimidazolium bromide (C4mimBr, >98%), 1-butyl-
3-methylimidazolium methylsulfate (C4mimCH3SO4, >99%),
1-butyl-3-methylimidazolium acetate (C4mimOAc, >99%), 1-hexyl-
3-methylimidazolium chloride (C6mimCl, >99%), 1-hexyl-2,3-
dimethylimidazolium
chloride
(C6mmimCl,
>99%),
and
As a first assessment to address the intrinsic activity of
1-butyl-2,3-dimethylimidazolium chloride (C4mmimCl, >98%), 1-
octyl-3-methylimidazolium chloride (C8mimCl, >98%), and 1-
octyl-2,3-dimethylimidazolium chloride (C8mmimCl, >98%) were
supplied by C-TRI (South Korea). KBr (≥99%), 1-butyl-3-methylimi-
dazolium tetrafluoroborate (C4mimBF4, ≥98.5%), and 1-butyl-3-
methylimidazolium hexafluorophosphate (C4mimPF6, ≥98.5%)
were purchased from Sigma-Aldrich.
imidazolium-based ILs, the effects of alkyl chain length and additional
alkyl group attached to the imidazolium ring were investigated. Since
imidazolium cations with more hydrophobic characters exhibit a
weaker association with the chloride ion [20], they would be
expected to have a positive effective charge of H atom on the C2
position (i.e., C(2)–H) of imidazolim ring and, in turn, become more
active. Thus, 1-R-3-methylimidazolium chloride and 1-R-2,3-
dimethylimidazolium chloride, where R represents butyl (C4), hexyl
(C6) and octyl (C8), were applied for the dehydration of fructose. As
clearly shown in Fig. 1, the activities of all the Cl-containing ionic
liquids were very similar in terms of the fructose conversion as well
as HMF yield whose trends were sigmoid due to an erroneous zero
time caused by a temperature difference between the reaction
mixture and oil bath. Therefore, the length of the alkyl chain and
the additional alkyl group in imidazolium-based ionic liquids did
not afford any noticeable improvement in the fructose dehydration.
This is related with the pKa values of alkylimidazolium chloride
ionic liquids in DMSO, which will be explained later.
Next, the anion could be supposed to alter the ability of imidazo-
lium cations acting as an acid catalyst and take parts in the dehydra-
tion reaction as a neucleophile. The activities of ionic liquids with
various counter-anions associated with the C4mim cation were thus
investigated in order to rationalize the above supposition. Fig. 2 de-
picts the fructose conversion, HMF yield in a unit of % and mol/mol
whose latter is equal to the amount of HMF produced per that of
ionic liquid (C4mimCl, C4mimBr, C4mimBF4, C4mimPF6, C4mimCH3-
SO4 or C4mimOAc) added to the reaction mixture as a function of
the reaction time. As a result, the anion significantly affected the
catalytic activity of C4mim-based ionic liquids. In comparison with
C4mimCl (HMF productivity=3.1 mol/mol·h at 2 h), C4mimBr in
2.2. Activity test
Prior to the reaction, all ionic liquids were pretreated under vacu-
um at 105 °C for 24 h. In a typical experiment, a pre-dried ionic liquid
was added to the fructose solution in DMSO (2 mL), where the fruc-
tose concentration was 2.5 wt.%. The reaction was initiated by im-
mersing the glass reactor of 10 mL in oil bath pre-heated at 80 °C
(t=0). It took less than 10 min to reach the reaction temperature.
In order for the temperature of whole system to remain constant,
the thermo oil and the reactant solution were stirred using magnetic
bars in the oil bath and glass reactors, respectively. Furthermore,
since the experiments were repeated more than three times for
reproducibility, the average of these data was presented in this
work. The reaction temperature of 80 °C was selected, because
DMSO acts as the catalyst for the dehydration of fructose to
5-hydroxymethylfurfural at 150 °C through formation of a key inter-
mediate, (4R, 5R)-4-hydroxy-5-hydroxymethyl-4,5-dihydrofuran-2-
carbaldehyde [19]. In the preliminary blank test with fructose and
DMSO, only 13% of HMF yield was observed at 80 °C for 3 h.
In several control experiments for confirming the effect of ILs' an-
ions, Ca(CH3COO)2, KBr, CH3COOH or Nafion NR50, whose amount
was calculated and weighed beforehand, was added into the reaction
mixture at the reaction time of 1 h.
After termination of the reaction, the product solution was
quenched in ice water, filtered and finally analyzed by a Perkin
Elmer HPLC equipped with RI detector and Bio-Rad HPX-87 C column
with 0.6 mL/min of distilled water at 80 °C. Due to the formation of a
few byproducts, the quantitative amounts of fructose and HMF (mol)
were measured for quantification. Because 1 mole fructose is con-
verted into 1 mole HMF, the maximum amount of HMF generated is
equal to the amount of initial fructose. Thus, the conversion of
fructose and HMF yield were calculated as follows:
Table 1
Catalytic activities of C4mimCl, Nafion NR50, CH3COOH and CaCl2 for the dehydration
of fructose.a
Reaction
time (h)
Fructose
conversion (%)
HMF yield
(%)
HMF/catalyst
(mol/mol)
C4mimCl
Nafion NR50
CH3COOH
CaCl2
2
5
2
5
2
5
2
5
70
97
76
95
53
74
57
95
31
56
57
79
6.6
29
13
55
6.3
11
10
14
1.3
5.7
2.7
11
FructoseconsumedðmolÞ
Fructose conversionð%Þ ¼
HMF yieldð%Þ ¼
ꢀ 100
Initial f ructoseðmolÞ
HMF producedðmolÞ
a
Reaction condition: fructose (50 mg), catalyst/fructose (0.05 mol/mol;), DMSO
(2 mL), 353 K. In the case of Nafion NR50 having the ion exchange capacity of
0.8 meq/g, the weight of catalyst added was 20 mg.
ꢀ 100
Initial f ructoseðmolÞ

J. Ryu et al. / Catalysis Communications 24 (2012) 11–15
13
Fig. 2. Effect of anions coordinated to the C4mim-based ionic liquids for the fructose de-
hydration under the same condition as presented in Table 1: A) fructose conversion
and B) HMF yield.
Fig. 1. Effect of alkyl chains attached to an imidazolium ring of Cl-containing ionic
liquids for the fructose dehydration under the same condition as presented in
Table 1: A) fructose conversion and B) HMF yield.
DMSO produced more HMF from fructose (productivity=5.0 mol/
mol·h at 2 h), because bromide is a better nucleophile. C4mimCH3SO4
and C4mimPF6 showed the HMF productivities of 2.2 and 0.65 mol/
mol·h at 2 h, respectively, whereas C4mimOAc and C4mimBF4 yielded
only a negligible amount of HMF. For interpretation of these results,
we tried to correlate the above results obtained at 2 h with the disso-
ciation constant of the ionic liquids, since more associated ion pairs
are less effective for acid-catalyzed reactions. As a result, it did not ap-
pear to be meaningful because the pKa values of C4mimCl, C4mimBr,
C4mimBF4, and C4mimPF6 in DMSO were almost similar (22.0–22.1)
[21]. This implies that only the dissociated C4mim cation, which is
well known to be relatively acidic, may not be responsible for the
conversion of fructose into HMF.
In order to verify the effect of anions of C4mim-based ionic liquids
on the fructose dehydration, the control experiment was conducted
to add a certain substance (Ca(CH3COO)2, KBr, CH3COOH or Nafion
NR50) into the reaction mixture containing C4mimCl, fructose and
DMSO at the reaction time of 1 h (Fig. 3). The addition of protons
(i.e., Nafion NR 50) indeed led to a substantial improvement in the
HMF production. In contrast, Ca(CH3COO)2 perfectly inhibited the
activity of C4mimCl, showing no HMF formation. This is considered
that the acetate anion can be strongly associated to C(2)–H of the
C4mim cation, due to its Brönsted basic character. On the other
hand, from the results obtained by the addition of KBr or CH3COOH,
an induction period appeared to be necessary for the fructose-to-
HMF conversion; at 2 h, the HMF productivities (3.5 and 0.9 by the
addition of KBr and CH3COOH, respectively) were lower than that of
C4mimCl. However, as the reaction progressed, the two additives
showed the similar fructose conversion and HMF yield/productivity
to those of C4mimCl. This could be explained in terms of the reaction
mechanism, as described below.
3.2. Reaction network on the fructose dehydration by 1,3-dialkylimidazolium
chloride
In general, there are two different options explaining the reaction
network on the fructose dehydration into HMF: the open-chain [22]
and cyclic pathways [23]. On the basis of the fact that a fructofuranose
is preferred than an open-chain form in the solvent of DMSO [19] and
the previous report in which a sequence of reaction proceeding via a

14
J. Ryu et al. / Catalysis Communications 24 (2012) 11–15
fructofuranosyl-ion intermediate is dominant [24], the cyclic pathway
is taken into account for establishing the reaction mechanism in this
work.
While referring to the reports of Antal et al. [24], and Binder and
Raines [16], we could draw the reaction network shown in Fig. 4. In
the reaction of fructofuranose with C4mimCl, the OH group at C-1
would be attacked by the dissociated C4mim cation of which C(2)–H
is acidic, thus forming the complex 1. Then, the removal of OH leads
to the formation of the intermediate 2 and C4mim+OH− (3). The for-
mer compound is spontaneously converted to the fructofuranosyl
oxocarbenium ions (4 and 5). At this point, two pathways, nucleo-
phile and base, is possible to form the enol (7), as described in Binder
and Raines [16]. According to the nucleophile pathway, the Cl anion
dissociated from C4mimCl can attack the oxocarbenium to form a 2-
deoxy-2-chloro intermediate (6) that is less prone to both side reac-
tions and reversion to fructose. Alternatively, the chloride ion could
form the enol by acting as a base that deprotonates C-1 [16]. The fol-
lowing steps lead to the conversion of the aldehyde intermediate (8)
finally to HMF through the dehydration reaction, which is faster
compared with the formation of the aldehyde 8.
From the above explanation, it could be suggested that, due to a
weak association between C4mim+ and Cl−, the former cation disso-
ciated serves as a Brönsted acid and the latter anion takes part as a
nucelophile in the fructose-to-HMF conversion. This plausible sugges-
tion is supported by the following findings. 1) The variation of alkyl
groups attached to the imidazolium cation assoiated with chloride
did not give any improvement in the HMF production, due to their
similar pKa values of around 22 in DMSO [21]. 2) The dissociation
strength between C4mim+ and anion is very critical to affect the IL's
activity on the fructose conversion into HMF, because the anion in-
deed participates in the reaction as shown in Fig. 4. 3) The nucleo-
phile pathway is considered to be valid from the result that
C4mimBr showed the better catalytic performance than C4mimCl
since bromide is a better nucelophile and leaving group than chloride,
which is in accordance with the result of Binder and Raines [16]. 4)
The finding that the addition of Ca(CH3COO)2 inhibited the HMF pro-
duction explains that the acetate anion becomes associated with
C4mim+, thus being inactive and, further, not triggering the conver-
sion of fructose. This was confirmed by 1H NMR spectroscopy
(600 MHz) that the chemical shift of C(2)–H appeared downfield
(from 9.142 to 9.168 ppm) as a result of Ca(CH3COO)2 addition
(Fig. 5). 5) From the result obtained by the addition of KBr, a nucleo-
phile is necessary to facilitate the HMF production and, specifically,
the formation of 2-deoxy-2-chloro intermediate (6). 6) Finally, the
Fig. 3. Catalytic activities of C4mimCl without and with the addition of Ca(CH3COO)2,
KBr, CH3COOH or Nafion NR50 at the reaction time of 1 h for the fructose dehydration
at 80 °C in DMSO: A) fructose conversion and B) HMF yield.
Fig. 4. Reaction network on the fructose dehydration catalyzed by C4mimCl acting as both Brönsted acid and nucleophile.

J. Ryu et al. / Catalysis Communications 24 (2012) 11–15
15
formed by the Brönsted acid that is C(2)–H of 1,3-dialkylimidazolium-
based ionic liquids. Finally, a better nucleophile (halide ions in this
work) is required to follow the nucleophile pathway via the formation
of 2-deoxy-2-chloro intermediate (6). This step would be very crucial in
acquiring the higher yield of HMF from fructose. Therefore, it was con-
sidered that all the above requirements are fulfilled with the reaction
system in which 1,3-dialkylimidazolium-based ionic liquids with halide
anions in DMSO work for the fructose-to-HMF conversion. This is due to
their dual catalytic function. Consequently, if the above reaction system
is achieved at a similar level by using an industrially feasible approach,
new possibilities would be offered for utilizing HMF as a platform
chemical to produce a variety of biomass-based fuels and chemicals.
Acknowledgments
The financial support by the Korea Ministry of Knowledge Econo-
my (Project No. KC000622) is greatly appreciated. This work is also
the outcome of a Manpower Development Program for Energy & Re-
sources supported by the Ministry of Knowledge and Economy
(MKE).
References
[1] C. Luo, S. Wang, H. Liu, Angewandte Chemie International Edition 46 (2007)
7636–7639.
[2] E.L. Kunkes, D.A. Simonetti, R.M. West, J.C. Serrano-Ruiz, C.A. Gärtner, J.A.
Dumesic, Science 322 (2008) 417–421.
[3] G.W. Huber, S. Iborra, A. Corma, Chemistry Review 106 (2006) 4044–4098.
[4] Y. Román-Leshkov, C.J. Barrett, Z.Y. Liu, J.A. Dumesic, Nature 447 (2007) 982–986.
[5] N.K. Gupta, S. Nishimura, A. Takagaki, K. Ebitani, Green Chemistry 13 (2011)
824–827.
[6] Y. Román-Leshkov, J.N. Chheda, J.A. Dumesic, Science 312 (2006) 1933–1937.
[7] J.N. Chheda, Y. Román-Leshkov, J.A. Dumesic, Green Chemistry 9 (2007) 342–350.
[8] Y. Román-Leshkov, J.A. Dumesic, Topics in Catalysis 52 (2009) 297–303.
[9] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith Jr., Industrial and Engineering Chemistry
Research 47 (2008) 9234–9239.
[10] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith Jr., ChemSusChem 2 (2009) 944–946.
[11] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith Jr., Green Chemistry 12 (2010)
1327–1331.
[12] C. Li, Z.K. Zhao, A. Wang, M. Zheng, T. Zhang, Carbohydrate Research 345 (2010)
1846–1850.
Fig. 5. NMR spectra of C4mimCl (A), and a mixture of C4mimCl and Ca(CH3COO)2 (B) in
deuterated DMSO.
[13] C. Lansalot-Matras, C. Moreau, Catalysis Communications 4 (2003) 517–520.
[14] Q. Bao, K. Qiao, D. Tomida, C. Yokoyoma, Catalysis Communications 9 (2008)
1383–1388.
[15] T. Ståhlberg, M.G. Sørensen, A. Riisager, Green Chemistry 12 (2010) 321–325.
[16] J.B. Binder, R.T. Raines, Journal of the American Chemical Society 131 (2009)
1979–1985.
[17] H. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Science 316 (2007) 1597–1600.
[18] C. Moreau, A. Finiels, L. Vanoye, Journal of Molecular Catalysis A: Chemical 253
(2006) 165–169.
fact that the induction period exists upon the addition of KBr or CH3-
COOH tells that the ion-exchange reaction, resulting in the ion associ-
ation and/or dissociation, gradually occurs compared to the
dehydration reaction.
[19] A.S. Amarasekara, L.D. Williams, C.C. Ebede, Carbohydrate Research 343 (2008)
3021–3024.
4. Conclusions
[20] T. Cremer, C. Kolbeck, K.R.J. Lovelock, N. Paape, R. Wölfel, P.S. Schulz, P.
Wasserscheid, H. Weber, J. Thar, B. Kirchner, F. Maier, H.-P. Steinrück, Chemistry
A European Journal 16 (2010) 9018–9033.
[21] Y. Chu, H. Deng, J.-P. Cheng, Journal of Organic Chemistry 72 (2007) 7790–7793.
[22] M.S. Feather, J.F. Harris, Advances in Carbohydrate Biochemistry 28 (1973)
161–224.
[23] F.H. Newth, Advances in Carbohydrate Chemistry 6 (1951) 83–106.
[24] M.J. Antal Jr., W.S.L. Mok, G.N. Richards, Carbohydrate Research 199 (1990)
91–109.
While investigating the reaction pathway for the fructose dehy-
dration to HMF, several insights can be drawn for further studies.
On a first hand, the furanose form is of advantage rather than the py-
ranose in terms of the conformation of fructose for the dehydration
reaction. This was accomplished by using the DMSO solvent in this
work. Next, for the HMF production, the fructofuranosyl oxocarbe-
nium ions (4 and 5) are a key intermediate, which was successfully