Direct conversion of glucose to 5-(hydroxymethyl)furfural in ionic liquids with lanthanide catalysts

Article abstract of DOI:10.1039/b916354a

The direct conversion of glucose to 5-(hydroxymethyl)furfural (HMF) in ionic liquids with lanthanide catalysts was examined in search of a possibly more environmentally feasible process not involving chromium. The highest HMF yield was obtained with ytter

Full text of DOI:10.1039/b916354a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/greenchem | Green Chemistry
Direct conversion of glucose to 5-(hydroxymethyl)furfural in ionic liquids
with lanthanide catalysts
Tim St a˚ hlberg, Mathilde Grau Sørensen and Anders Riisager*
Received 10th August 2009, Accepted 20th November 2009
First published as an Advance Article on the web 12th January 2010
DOI: 10.1039/b916354a
The direct conversion of glucose to 5-(hydroxymethyl)furfural (HMF) in ionic liquids with
lanthanide catalysts was examined in search of a possibly more environmentally feasible process
not involving chromium. The highest HMF yield was obtained with ytterbium chloride or triflate
together with alkylimidazolium chlorides. Notably, a higher reactivity was observed when the
hydrophobicity of the imidazolium cation was increased, in contrast to analogous chromium
catalyst systems. This indicates a different reaction mechanism for the lanthanides than for the
chromium catalyst systems.
2
1-25
Introduction
to HMF.
More importantly, very good results have been
shown for the direct conversion of glucose to HMF (up to 70%
yield) in 1-ethyl-3-methylimidazolium chloride ([EMIm]Cl) with
a catalytic amount of chromium(II) chloride. In the proposed
mechanism, a complex between ionic liquid and CrCl interacts
with the open chain of glucose, leading both to isomerization to
fructose and direct conversion to HMF (Scheme 2).
An even higher yield of HMF was obtained when using
chromium(II) or chromium(III) chloride complexed with steri-
cally hindered carbenes. Comparable yields were also shown
using a large surplus of metal halides in DMA together with
chromium(II/III) chloride. Here a detailed mechanism of the
actual role of the halide in the dehydration was also proposed.
In this mechanism, fructose initially loses one water molecule,
and the nucleophilic halide is then added to the anomeric
carbon. The intermediate formed is then deprotonated to an
enolic species, which is subsequently dehydrated twice to form
HMF. Alternatively, the halide acts as a base in the second
step and generates the enolic intermediate by deprotonation
(Scheme 3).
In recent years, the search for alternatives to today’s plat-
form chemicals derived from fossil sources has intensi-
fied. One molecule under particular scrutiny has been 5-
16
2
(
hydroxymethyl)furfural (HMF), which is expected to become
an important precursor to fuels, solvents and polymers in the
chemical infrastructure.
1-3
HMF is formed by dehydration of hexoses under elevated tem-
perature (Scheme 1, exemplified for glucose). The mechanism of
the dehydration of hexoses has been subject of some debate, and
proposals include both a mechanism consisting of only cyclic
17
18
4-8
9-13
intermediates as well as an open-chain mechanism.
Scheme 1 Dehydration of glucose to HMF.
Apart from chromium chloride, there are no reports of equally
efficient catalysts for the direct conversion of glucose to HMF. A
very recent publication showed, however, that a dual-function
catalyst composed of sulfated zirconia on alumina support
could convert glucose directly to HMF with a moderate yield,
presumably via glucose isomerization to fructose followed by
The dehydration of fructose is quite facile – even in the
absence of catalyst – in high-boiling solvents such as DMSO,
14,15
DMF and DMA,
for the formation of HMF.
whereas glucose requires a special catalyst
16-19
However, fructose is derived
from glucose in a process that only yields about 50% of
fructose and involves a chromatographic step in order to be
19
dehydration. Furthermore, lanthanide chlorides have been
20
purified. Consequently, in order to find an economical and
environmentally feasible industrial process for the production
of HMF, an efficient direct conversion from glucose would be
most beneficial.
shown to have a catalytic effect on glucose dehydration in
26
supercritical water. Notably, no leuvulinic acid or formic acid
was observed here, as is normally the case when exposing
27
HMF to aqueous acidic conditions. In the study, kinetic and
In the last five years or so, interesting results have been
presented on the dehydration of hexoses by the use of ionic
liquids. As in the case of high-boiling organic solvents, ionic
liquids have proven to be efficient media for converting fructose
spectroscopic experiments suggested weak interaction between
lanthanide and saccharide, resulting in higher reaction rates for
smaller, less hydrated and heavier lanthanide ions.
Our objective in this work was to combine in particular
the heavier lanthanide catalysts with ionic liquids and to
investigate its catalytic effect on the direct conversion of glucose
to HMF. Thus, imidazolium-based ionic liquids were used as
solvents together with various lanthanide chloride catalysts in
28,29
Centre for Catalysis and Sustainable Chemistry, Department of
Chemistry, Building 207, Technical University of Denmark, DK-2800,
Kgs. Lyngby, Denmark. E-mail: ar@kemi.dtu.dk; Tel: +45 45252233
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 321–325 | 321

View Article Online
16
Scheme 2 Possible mechanism for the chromium-catalyzed dehydration of glucose to HMF in ionic liquids.
18
Scheme 3 Proposed mechanism for the dehydration of fructose to HMF with halides.
a
18
Table 1 HMF degradation in ionic liquids
Raines, where the chloride acts as a nucleophile and promotes
the dehydration of the furanose fructose. Consequently, our
study focused on the alkylimidazolium chlorides as solvents.
Ionic liquid
Conversion of HMF (%)
[
BMIm]Cl or [EMIm]Cl converted only fructose to HMF
[
[
[
[
[
[
[
[
[
[
[
[
MIm]Cl
EMIm]Cl
BMIm]Cl
9
3
29
9
11
> 99
> 99
6
69
59
9
16
without additives (as also reported by Zhao et al. ) and not
glucose. The addition of lanthanide chlorides, on the other hand,
revealed a catalytic effect on glucose conversion (but not on
fructose conversion), even though temperatures above 140 C
were required to avoid slow reactions and negligible yields.
Furthermore, methylimidazolium chloride, which previously has
given an excellent conversion of fructose to HMF, gave less
than one percent HMF from glucose with any lanthanide.
In the experiments with lanthanides (Table 2), cerium had
very little catalytic effect in both [BMIm]Cl and [EMIm]Cl,
whereas the other lanthanides showed quite different behaviours
depending on the liquid. In the case of [EMIm]Cl, the reactivity
for the lanthanide chlorides was the highest for promethium and
then decreased almost linearly through the heavy lanthanides to
HMIm]Cl
OMIm]Cl
EMIm]OAc
BMIm]OAc
Choline][dmp]
MIm][HSO
EMIm][N(CN)
EMIm][C OSO
EMIm][AlCl
◦
b
4
]
23
2
]
2
H
5
3
]
4
]
77
a
Reaction conditions: 1.0 g ionic liquid, 100 mg (0.56 mmol) glucose,
◦
b
1
00 C, 8 h. dmp = dimethyl phosphate.
◦
the temperature range 120–200 C. Additionally, the investiga-
tion was expanded to include even stronger Lewis acids such as
ytterbium triflate.
YbCl . Moreover, when using the more Lewis-acidic ytterbium
3
Results and discussion
Table 2 Lanthanide-catalyzed dehydration of glucose in [EMIm]Cl
a
and [BMIm]Cl
We initially screened the stability of HMF in several ionic liquids
at elevated temperature to make sure that the ionic liquid used
was not detrimental to the desired product. As seen from the
data in Table 1, all ionic liquids induced some degradation of
HMF during the stability tests. The acidic and basic ionic liquids
proved to promote highest degree of degradation, whereas the
[EMIm]Cl
[BMIm]Cl
Catalyst
Yield (%) Selectivity (%) Yield (%) Selectivity (%)
CeCl
PrCl
NdCl
3
3
13
12
10
5
3
27
24
19
7
3
7
8
10
12
24
4
22
23
23
15
37
3
3
imidazolium chlorides, [EMIm][C
were the most benign.
2
H
5
OSO ] and [Choline][dmp]
3
DyCl
YbCl
3
3
On the basis of the initial experiments, we further discovered
that there was no HMF formed from glucose in ionic liquids
that did not have chloride or other halides as anion. This is
in agreement with the mechanism proposed by Binder and
Yb(OTf)3
10
16
a
Reaction conditions: 1.0 g ionic liquid, 100 mg (0.56 mmol) glucose,
.056 mmol catalyst, 140 C, 6 h.
◦
0
3
22 | Green Chem., 2010, 12, 321–325
This journal is © The Royal Society of Chemistry 2010

View Article Online
triflate, a comparable yield to that of dysprosium chloride was
obtained.
In [BMIm]Cl a completely different pattern was observed.
Here the yield increased gradually from CeCl
3
to YbCl . This
3
is consistent with the earlier relative catalytic effect found in
26,28,29
supercritical water.
Furthermore, ytterbium chloride and
ytterbium triflate gave significantly higher yields in [BMIm]Cl
than in [EMIm]Cl. This is the opposite of what has been shown
earlier using chromium catalysts, where the yields were markedly
16
better in [EMIm]Cl, thus clearly suggesting different catalyst–
ionic liquid interaction in the systems.
Increasing the catalyst concentration of YbCl in [EMIm]Cl,
3
up to as much as 30 mol%, did not reveal any improvement in
HMF yield. In contrast, the amount of catalyst had a significant
effect on the HMF yield in [BMIm]Cl, thus further signifying
differences between the catalytic systems (Fig. 1).
Fig. 2 Dehydration of glucose in [BMIm]Cl catalyzed by YbCl
3
at
◦
1
60 C as a function of the reaction time. The reaction contained 1.0 g
3
ionic liquid, 100 mg (0.56 mmol) glucose and 16 mg (0.056 mmol) YbCl .
◦
Fig. 1 Dehydration of glucose in [EMIm]Cl and [BMIm]Cl at 140 C
for 6 h with different amount of YbCl
1
3
3
catalyst. The reaction contained
.0 g ionic liquid, 100 mg (0.56 mmol) glucose and 16 mg (0.056 mmol),
2 mg (0.112 mmol) and 48 mg (0.168 mmol) YbCl , respectively.
3
Fig. 3 Dehydration of glucose in [BMIm]Cl catalyzed by YbCl
3
at
◦
◦
During dehydration in [BMIm]Cl, the selectivity towards
1
60 C and 200 C as a function of time. Both reactions contained 1.0 g
HMF reached a maximum already after 10 min and a maximum
yield after 30 min. The HMF was then slowly degraded
due to the high temperature (Fig. 2). At an even higher
ionic liquid, 100 mg (0.56 mmol) glucose and 16 mg (0.056 mmol) YbCl3.
18
Binder and Raines, who combined LiCl with DMA in order to
◦
temperature of 200 C the yield was slightly increased, but
produce weakly ion-paired chloride ions for sugar dehydration.
In order to examine the effect of ion-pairing more closely,
additional experiments with even longer alkyl chains, such as
at the expense of an even faster degradation of the product.
In this case maximum yield was obtained after five minutes,
and HMF was completely degraded after two hours (Fig. 3).
All reactions formed substantial amounts of humins, but no
other by-products were detected, as the essential anhydrous
conditions (small amount of water is formed upon dehydration)
prevented rehydration of HMF into leuvulinic acid and formic
acid.
1
3
-hexyl-3-methylimidazolium chloride ([HMIm]Cl) and 1-octyl-
-methylimidazolium chloride ([OMIm]Cl), were performed. In
these experiments, a faster conversion and a small increase in
yield could actually be seen when using [HMIm]Cl or [OMIm]Cl
as solvents (Fig. 4), even though it was less pronounced
than the difference between [EMIm]Cl and [BMIm]Cl. The
decrease in selectivity for [HMIm]Cl and [OMIm]Cl compared
to [BMIm]Cl could be a result of the faster formation of
humins.
When it comes to finding an explanation why a longer alkyl
chain on the imidazolium ring gave higher reaction rates for
YbCl
3
and Yb(OTf) catalysts, a possible reason could be related
3
to ion pairing in the solvent. Cations with more hydrophobic
character would be expected to have a weaker association with
the chloride ion, which in turn would become more reactive.
Earlier work has already shown that rates of reactions in ionic
liquids involving halides are highly dependent on the nature of
Conclusions
We have shown that lanthanides catalyze the conversion of
glucose to HMF in dialkylimidazolium chlorides. The strongest
Lewis acids, YbCl
30
the cation. This was also one of the rationales in the work of
3
and Yb(OTf)
3
, gave the highest yields, even
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 321–325 | 323

View Article Online
Dehydration reactions
A 40 mL reaction tube was charged with ionic liquid (1 g) and
◦
lanthanide(III) chloride (0.056 mmol) and heated at 100 C for
1
h. Glucose (100 mg, 0.56 mmol) was then added and the
◦
solution stirred for 3 h at 140 C. After reaction the reaction
tube was cooled in an ice bath and water (5 mL) added. The
solids were filtered off and the filtrate analyzed by HPLC
(
7
0
Agilent 1200 series, Bio-Rad Aminex HPX-87H, 300 mm ¥
◦
.8 mm pre-packed column, 0.005 M H
2
SO
4
mobile phase, 60 C,
-
1
.6 mL min ).
Definitions of yield and selectivity
The yields and selectivities were based on conversion of glucose
and confirmed by calibration of standard solutions of the
products and reactants involved. With a known molar amount
of all components, the conversion, yield and selectivity were
calculated from the equations below:
Fig. 4 Dehydration of glucose in [EMIm]Cl, [BMIm]Cl, [HMIm]Cl
and [OMIm]Cl at 160 C for 1 h. All reactions contained 1.0 g ionic
liquid, 100 mg (0.56 mmol) glucose and 16 mg (0.056 mmol) YbCl .
3
◦
⎛
⎞
Amount of glucose
⎜
⎟
Glucose conversion = ⎜1−
⎟×100%
⎟
16
though these were moderate (24%) compared to CrCl
2
(70%).
⎜
⎟
Starting amount of glucose⎠
⎝
Our results also suggest that the mechanism of the reaction
might be different to that of chromium-catalyzed dehydration
of glucose.
(
1)
Amount of HMF
In the chromium system, highest yields are obtained with
Yield of HMF =
×100%
(2)
(3)
Starting amount of glucose
[
EMIm]Cl, whereas the yield decreases with more hydrophobic
16
imidazolium cations. In this reaction, chromium(II) chloride is
believed to form a complex with the ionic liquid and then asso-
Yield of HMF
Selectivity of HMF =
×100%
16
ciate with the open chain form of glucose. Ytterbium showed
a different reaction pattern, whereby the yield increased as the
hydrophobic character of the imidazolium ring increased. We
believe that the increase in reaction rates are due to weaker ion
pairing between the chloride and imidazolium cation, making
this more reactive in the dehydration mechanism. Furthermore,
ytterbium could be less prone to form complexes with the
imidazolium chlorides, which would explain the difference in
reactivity compared to chromium.
Glucose conversion
Acknowledgements
The reported work was supported by the Danish National Ad-
vanced Technology Foundation in cooperation with Novozymes
A/S. Special gratitude is due to BASF for providing the ionic
liquids.
Further studies on the actual complexation between lan-
thanide chlorides and imidazolium chlorides are needed to
clarify the exact reaction pattern with glucose. Modeling of the
various structures and intermediates together with the metals
could also give new insights into the mechanism. Combined,
this information could provide valuable directions on how
to improve the lanthanide–ionic liquid systems to become
competitive with present chromium systems.
References
1
T. Werpy and G. Petersen, US DOE, 2004, No. DOE/GO-102004-
992, http://www.nrel.gov/docs/fy04osti/35523.pdf.
A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107(2411), 2502.
1
2
3 J. N. Chheda, Y. Roman-Leshkov and J. A. Dumesic, Green Chem.,
2007, 9, 342–350.
4
5
J. U. Nef, Justus Liebigs Ann. Chem., 1910, 376, 1–119.
W. N. Haworth, E. L. Hirst and V. S. Nicholson, J. Chem. Soc., 1927,
1
513–1526.
6
7
W. N. Haworth and W. G. M. Jones, J. Chem. Soc., 1944, 667.
M. J. Antal, W. S. L. Mok and G. N. Richards, Carbohydr. Res., 1990,
1
99, 91–109.
Experimental
8
9
A. S. Amarasekara, L. D. Williams and C. C. Ebede, Carbohydr. Res.,
008, 343, 3021–3024.
M. L. Wolfrom, E. G. Wallace and E. A. Metcalf, J. Am. Chem. Soc.,
942, 64, 265–269.
2
Materials and apparatus
1
Glucose (99.5%), 1-hexyl-3-methylimidazolium chloride (98%),
1
1
1
0 M. L. Wolfrom, R. D. Schuetz and L. F. Cavalieri, J. Am. Chem. Soc.,
1948, 70, 514–517.
1 M. L. Wolfrom, R. D. Schuetz and L. F. Cavalieri, J. Am. Chem. Soc.,
1
-octyl-3-methylimidazolium chloride (98%) and ytterbium tri-
flate (99%) were purchased from Aldrich. [EMIm][N(CN)
98%), [Choline][dmp] (98%) and [EMIm][C OSO ] (98%)
2
]
1
949, 71, 3518–3523.
(
2
H
5
3
2 E. F. L. J. Anet, J. Am. Chem. Soc., 1960, 82, 1502–1502.
were purchased from Solvent Innovation. All other ionic liquids
were obtained from BASF (>95%). Lanthanide chlorides (p.a.)
were purchased from Rare Earth Ltd. All experiments were
performed under nitrogen atmosphere using a Radley Carousel
13 E. Anet, Aust. J. Chem., 1961, 14, 295–301.
14 Y. Nakamura and S. Morikawa, Bull. Chem. Soc. Jpn., 1980, 53,
3
705–3706.
1
1
5 R. M. Musau and R. M. Munavu, Biomass, 1987, 13, 67–74.
6 H. Zhao, J. E. Holladay, H. Brown and Z. C. Zhang, Science, 2007,
316, 1597–1600.
1
2 Plus Basic System.
3
24 | Green Chem., 2010, 12, 321–325
This journal is © The Royal Society of Chemistry 2010

View Article Online
1
1
1
2
7 G. Yong, Y. Zhang and J. Y. Ying, Angew. Chem., Int. Ed., 2008, 47,
345–9348.
8 J. B. Binder and R. T. Raines, J. Am. Chem. Soc., 2009, 131, 1979–
985.
9 H. Yan, Y. Yang, D. Tong, X. Xiang and C. Hu, Catal. Commun.,
009, 10, 1558–1563.
0 S. Bhosale, M. Rao and V. Deshpande, Microbiol. Rev., 1996, 60,
80–300.
24 S. Hu, Z. Zhang, Y. Zhou, B. Han, H. Fan, W. Li, J. Song and Y. Xie,
9
Green Chem., 2008, 10, 1280–1283.
25 Q. Bao, K. Qiao, D. Tomida and C. Yokoyama, Catal. Commun.,
2008, 9, 1383–1388.
1
26 H. Ishida and K.-i. Seri, J. Mol. Catal. A: Chem., 1996, 112, L163–
2
L165.
27 H. P. Teunissen, Recl. Trav. Chim. Pays Bas, 1930, 49, 784–826.
28 K.-i. Seri, Y. Inoue and H. Ishida, Bull. Chem. Soc. Jpn., 2001, 74,
1145–1150.
2
2
2
1 C. Fayet and J. Gelas, Carbohydr. Res., 1983, 122, 59–68.
2 C. Lansalot-Matras and C. Moreau, Catal. Commun., 2003, 4, 517–
29 K.-i. Seri, T. Sakaki, M. Shibata, Y. Inoue and H. Ishida, Bioresour.
5
20.
Technol., 2002, 81, 257–260.
2
3 C. Moreau, A. Finiels and L. Vanoye, J. Mol. Catal. A: Chem., 2006,
30 N. L. Lancaster, P. A. Salter, T. Welton and G. B. Young, J. Org.
Chem., 2002, 67, 8855–8861.
2
53, 165–169.
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 321–325 | 325