Enhanced conversion of carbohydrates to the platform chemical 5-hydroxymethylfurfural using designer ionic liquids

Article abstract of DOI:10.1002/cssc.201301368

5-Hydroxymethylfurfural (HMF) is a key platform chemical that may be obtained from various cellulosic (biomass) derivatives. Previously, it has been shown that ionic liquids (ILs) facilitate the catalytic conversion of glucose into HMF. Herein, we demonstrate that the careful design of the IL cation leads to new ionic solvents that enhance the transformation of glucose and more complex carbohydrates into HMF significantly. In Situ NMR spectroscopy and computational modeling pinpoint the key interactions between the IL, catalyst, and substrate that account for the enhanced reactivities observed. Solvents by design: 5-Hydroxymethylfurfural (HMF) is a key platform chemical that may be obtained from various cellulosic (biomass) derivatives. Ionic liquids (ILs) facilitate the catalytic conversion of glucose into HMF. We demonstrate that the careful design of the IL cation leads to new ionic solvents that enhance the transformation of glucose and more complex carbohydrates into HMF.

Full text of DOI:10.1002/cssc.201301368

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201301368
Enhanced Conversion of Carbohydrates to the Platform
Chemical 5-Hydroxymethylfurfural Using Designer Ionic
Liquids
Sviatlana Siankevich,[a] Zhaofu Fei,[a] Rosario Scopelliti,[a] Gabor Laurenczy,[a]
Sergey Katsyuba,[b] Ning Yan,*[c] and Paul J. Dyson*[a]
5-Hydroxymethylfurfural (HMF) is a key platform chemical that
may be obtained from various cellulosic (biomass) derivatives.
Previously, it has been shown that ionic liquids (ILs) facilitate
the catalytic conversion of glucose into HMF. Herein, we dem-
onstrate that the careful design of the IL cation leads to new
ionic solvents that enhance the transformation of glucose and
more complex carbohydrates into HMF significantly. In Situ
NMR spectroscopy and computational modeling pinpoint the
key interactions between the IL, catalyst, and substrate that ac-
count for the enhanced reactivities observed.
Introduction
The transformation of lignocellu-
losic biomass into commodity
chemicals is of increasing inter-
est to reduce the dependence of
mankind on petrochemical feed-
stocks.[1] A plethora of routes
and compounds are under inves-
tigation and, of these, the con-
Scheme 1. Dehydration of glucose to HMF.
version of various biomass deriv-
atives into 5-hydroxymethylfur-
fural (HMF) has been identified as a principle route.[1c,2] Cur-
rently, fructose dehydration in the presence of an acid catalyst
is the main approach for HMF production.[3] However, HMF can
be obtained from glucose through isomerization to fructose as
well as directly from cellulose,[4] and recently the production of
HMF from biomass using a hydrothermal carbonization (HTC)
process has been commercialized.[5]
temperatures in the absence of a catalyst to achieve a low
yield, ꢀ10%.[7] Under similar conditions, but in the presence of
a catalyst, for example, H3PO4/Nb2O5, yields of ꢀ50% are ob-
tained.[8] Heterogeneous catalysts used to transform glucose
into HMF generally do not afford the product in more than
30% yield,[9] although in organic solvents higher yields of HMF
can be obtained compared to reactions in water, which is not
Numerous studies that describe the transformations of glu-
cose have been reported.[6] Glucose undergoes dehydration to
HMF (Scheme 1) in aqueous solution at high pressures and
unexpected for a
dehydration reaction.[2] For example,
a DMSO/water/SO42ꢁ/ZrO2/Al2O3 system affords HMF from glu-
cose in 56% yield[10] and a homogeneous DMA/LiBr/CrBr3
system affords HMF in 80% yield.[11]
[a] S. Siankevich, Dr. Z. Fei, Dr. R. Scopelliti, Prof. G. Laurenczy, Prof. P. J. Dyson
Institut des Sciences et Ingꢀnierie Chimiques
Ecole Polytechnique Fꢀdꢀrale de Lausanne (EPFL)
1015, Lausanne (Switzerland)
The abovementioned systems have various limitations relat-
ed to the yield and selectivity of HMF obtained, catalyst and
solvent stability/recycling, and/or corrosion problems.[2] For ex-
ample, under acidic conditions, HMF decomposes to levulinic
and formic acids, and the former is particularly difficult to sep-
arate from HMF.[12] Organic solvents pose other problems con-
nected to recyclability and hence cost,[13] and thus far, the bi-
phasic systems reported do not give adequate yields of
HMF.[14]
Fax:(+41)21-693-97-80
E-mail: paul.dyson@epfl.ch
[b] Prof. S. Katsyuba
A.E. Arbuzov Institute of Organic and Physical Chemistry of
Kazan Scientific Center of Russian Academy of Sciences
Arbuzov Street 8, 420088, Kazan (Russia)
[c] Prof. N. Yan
Nevertheless, the use of volatile and toxic organic solvents
for this reaction is not ideal, consequently, ionic liquids (ILs),
which are nonvolatile and exhibit relevant solvation proper-
ties,[15] have been explored as alternatives. Indeed, the synthe-
sis of HMF from fructose and glucose may be enhanced if cer-
tain ILs are employed as the reaction media.[1a,c,d,16] Notably,
Department of Chemical and Biomolecular Engineering
National University of Singapore
4 Engineering Drive 4, 117576 (Singapore)
Fax:(+65)67-79-19-36
E-mail: ning.yan@nus.edu.sg
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201301368.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
1
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 1. ILs evaluated in the dehydration of glucose into HMF.
Results and Discussion
chloride-based ILs were found to have an exceptionally high
capacity to dissolve carbohydrates, attributable to the exten-
sive hydrogen-bonding interactions that disrupt the structure
of carbohydrates.[17] Combined with transition-metal catalysts,
efficient production of HMF from sugars is possible,[3a] for
which chromium chloride provides the best results (see
below).[1c]
1-Ethyl-3-methylimidazolium chloride (1) combined with the
CrCl2 precatalyst has been shown to catalyze the transforma-
tion of glucose into HMF previously,[1c,16] and through the use
of this combination we obtained HMF in 57% yield after 3 h
(see Experimental Section for further details). IL 1 was subse-
quently replaced with various ether- and hydroxyl-functional-
ized imidazolium chlorides (Figure 1), and the reaction was
evaluated under the same conditions (Figure 2) to explore the
effect of the functional group(s) on the transformation.
Imidazolium-based ILs display a high degree of 3D structural
organization relative to other ILs because of the three acidic
hydrogen atoms on the imidazolium cation that aid the direc-
tionality of hydrogen bonds and
other weak forces.[15] Conse-
quently, the introduction of
some specific hydrogen-bond
donor/acceptor groups, and po-
tentially other types of function-
ality, should enhance the ability
of the IL to interact with a sub-
strate. Such groups also influ-
ence the intermolecular interac-
tions between ions, which favor
solvent–substrate
interactions
and hence activation,[18] and nas-
cent studies indicate that func-
tionalized ILs can enhance bio-
mass transformations.
Figure 2. Dehydration of glucose to HMF in different ILs. Conditions: IL (500 mg), glucose (0.28 mmol, 50 mg),
CrCl2 (6 mol%) at 1008C. IL 9* comprises a mixture of 4 (450 mg) and 9 (50 mg).
Herein, we describe a study
concerned with the application of IL/CrCl2 systems in which hy-
drogen-bond donor/acceptor groups on the imidazolium
cation of the IL allow HMF to be obtained from glucose in con-
siderably higher yields compared to simple imidazolium chlor-
ides. Insights into the underlying mechanisms for the enhance-
ments are provided by NMR spectroscopy and DFT calcula-
tions. Moreover, the new system also converts more complicat-
ed substrates such as cellobiose and maltotetraose into HMF
in good yield.
The ether-modified IL 2 is unstable under the reaction condi-
tions,[19] and after 3 h the yield of HMF is only ꢀ2%, despite
a high conversion, whereas 3 combined with CrCl2 affords
HMF in 65% yield, a significant improvement on 1. The same
comparatively high yield, that is, 65%, is also obtained with
the hydroxyl-derivatized IL 4.
To better understand the role of the hydrogen bonding, IL 5,
which contains a methyl group in place of the most acidic imi-
dazolium ring proton, was evaluated in the dehydration of glu-
cose, and HMF was obtained in 62% yield. Although it is not
unreasonable to assume that the acidic ring proton should en-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
2
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
hance the reaction through the formation of noncovalent in-
teractions with the substrate, the relatively minor decrease in
yield indicates that the role of this proton is not particularly
significant, at least in the presence of the hydroxyl group,
which has been shown to form stronger H bonds than the imi-
dazolium ring protons.[20] Indeed, the functional group at-
tached to the imidazolium cation has a greater influence and,
consequently, two other ILs with hydroxyl groups were evalu-
ated in the dehydration reaction. The slightly more hydropho-
bic IL 6 resulted in a reduced conversion of glucose and
a lower yield of HMF, that is, 27% after 3 h. It has been shown
that as the length of the alkyl chain on an IL cation increases
the yield of HMF in this reaction decreases,[1c,16,21] although the
opposite trend has been observed for systems that do not
employ a catalyst.[22] IL 7, which contains a diol group, showed
no advantage over 4, and HMF was isolated in 52% yield. For
IL 8, the yield of HMF is only 37% after 3 h.
in combination with 4 (1:9, w/w) exhibits exceptionally activity
in the CrCl2-catalyzed dehydration of glucose to HMF with
a yield of 81% obtained. Other combinations were evaluated
but in all cases they were less effective than the combination
with 4.
Dehydration of cellobiose and maltotetraose
Based on the excellent yields of HMF obtained from glucose,
the transformation of cellobiose (a dimer of glucose linked by
a b (1!4) glycosidic bond) into HMF was studied (Scheme 2).
ILs 1, 3, 4, 5, 6, and 7 were evaluated in combination with
CrCl2 and, as expected, the highest yield (68%) was obtained
with 4 after 4 h reaction. With this substrate, the combination
of 4 and 9 led to only a modest improvement in the yield, that
is, 70% after 3 h (Figure 3).
As the combination of 4 with the CrCl2 precatalyst leads to
the highest conversion of glucose and also provides HMF in
the highest yield, this IL was studied further. The catalyst load-
ing and reaction temperature influence the final yield (Ta-
bles S1 and S2, respectively), and the maximum yield of 65%
and maximum conversion of 93% may be obtained after only
2 h.
One of the key advantages of ILs in the dehydration of glu-
cose is that they are nonvolatile and can be recycled in a rela-
tively facile manner. However, as glucose can be transformed
into HMF in organic solvents in good yield we explored the
effect of organic co-solvents [toluene, dichloroethane, tetra-
chloroethane, chlorobenzene, and dimethylacetamide (DMA)]
with 4 in the presence of CrCl2 (Table 1). In general, the yield
Table 1. Effect of co-solvent on the 4/CrCl2-catalyzed reaction of glucose
to HMF.[a]
Scheme 2. Dehydration of cellobiose and maltotetraose into HMF.
Co-solvent
Amount [wt%]
Yield [%]
The conversion of maltotetraose, a tetramer of glucose
linked by an a (1!4) glycosidic bond, to HMF was also evalu-
ated in 1, 4, 5, and 7 in the presence of the CrCl2 precatalyst.
The influence of the hydroxyl group on activity is particularly
significant with this more complex substrate, with only 5% of
HMF obtained with 1 compared to 60% with 4 after 5 h. As far
as we are aware, this is the first report that describes the con-
version of this more complex substrate. The use of 9 as a co-
solvent with 4 accelerates the rate of the reaction such that
the maximum yield of 60% is obtained in a shorter time with
a lower catalyst loading (Figure 4). The conversion of polysac-
charides is considerably more challenging. Nevertheless, if we
use 4 and 9 as the solvent, HMF can be obtained in 40% yield
from starch after 4 h at 1408C and in 10% yield from cellulose
after 30 min reaction at 1608C. In comparison, no reaction is
observed with 1.
2 h
3 h
toluene
10
20
10
20
10
10
10
68
73
67
67
65
69
40
75
75
73
72
67
69
45
dichloroethane
tetrachloroethane
chlorobenzene
DMA
[a] Conditions: 4 (500 mg), co-solvent, glucose (0.28 mol, 50 mg), CrCl2
(6 mol% with respect to glucose) at 1008C.
of HMF is higher than that obtained in the neat IL. Notably,
with 10 wt% of toluene, HMF was obtained in 75% yield in
3 h. Higher concentrations of toluene did not increase the
yield of HMF further and HMF was not observed in neat tolu-
ene. DMA, previously shown to be a good solvent for this reac-
tion in the presence of LiBr and CrBr3,[11] leads to a decrease in
the yield of HMF if used as a co-solvent with 4.
As the presence of toluene in 4 increases the yield of HMF,
the structure of 4 was modified by replacing the methyl group
with a CH2CH2Ph group to give the bifunctionalized IL 9. IL 9
Mechanistic study
The 4/CrCl2 system efficiently catalyzes the transformation of
glucose to fructose and subsequently to HMF. Different mecha-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
3
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
hydroxyethyl-3-methylimidazolium) that was charac-
terized by X-ray crystallography (Figure 5). The IL
ꢁ
cation is coordinated to a CrCl3 anion through the
hydroxyl group with a Cr(1)ꢁO(1) bond length of
2.073(7) ꢁ. The CrꢁCl distances are essentially the
same at 2.401(2)–2.409(3) ꢁ, somewhat longer than
those in CrCl2 (mean=2.37 ꢁ),[25] CrCl3 (mean=
2.37 ꢁ),[26] and [Cr(h5-C5Me5)Cl3]ꢁ (mean=2.32 ꢁ).[27]
The crystal of [Cr(C2OHmim)Cl3] also contains inter-
molecular hydrogen bonds between the OH group
ꢁ
and the Cl atoms in the CrCl3 anion with distances
Figure 3. Dehydration of cellobiose to HMF in different ILs. Conditions: ILs (1 g), cello-
biose (0.15 mmol, 50 mg), CrCl2 (0.098 mmol, 12 mg) at 1008C. Except for 9* in which 4
(900 mg), 9 (100 mg), cellobiose (0.15 mmol, 50 mg), and CrCl2 (0.065 mmol, 8 mg) were
reacted at 1008C.
between the atoms from 3.023(7)–3.043(7) ꢁ.
With the X-ray structure as a starting point, possi-
ble reactive species were examined by quantum-
chemical methods. In the most energetically stable
variant of the optimized structures formed by the in-
teraction of [Cr(C2OHmim)Cl3] with 4, an additional
Clꢁ anion coordinates to the CrCl3 unit to form
a slightly distorted square-pyramidal geometry about
the Cr ion (Figure 6, top). In addition, the chloride
ligand is H bonded to the closest C2H aromatic frag-
ment of the imidazolium ring (H···Cl distance=
2.54 ꢁ). The addition of a molecule of glucose to the
model, placed in various positions near Cr center, re-
veals that the substitution of the [C2OHmim]+ ligand
by glucose is strongly energetically favorable (see
Figure 6 center, for the lowest energy variant of the
optimized structure). The energy of the interaction
(DEi) between the glucose molecule and the [Cr-
Figure 4. Dehydration of maltotetraose to HMF in different ILs. Conditions: ILs (1 g), mal-
toteraose (0.075 mmol, 50 mg), CrCl2 (0.163 mmol, 20 mg) at 1008C. Except for 9* in
which 4 (900 mg), 9 (100 mg), maltoteraose (0.075 mmol, 50 mg), and CrCl2 (0.130 mmol,
16 mg) were reacted at 1008C.
(C2OHmim)Cl4][C2OHmim] adduct
amounts to
12.4 kcalmolꢁ1 (1 kcal=4.2 kJ; see the Supporting In-
formation for details), in which the OH···O hydrogen
nistic routes have been proposed for the transformation of
glucose to fructose,[1c,4] although they all essentially involve
two main steps, first ring opening and second proton transfer
between C2 and C1 (Scheme 3). From fructose there are two
pathways described in the literature that relate to HMF forma-
tion, which involve either the dehydration of fructose by an
open-chain pathway through two 1,2-eliminations and one
1,4-elimination of water or a pathway that involves a cyclic
fructofuranosyl intermediate.[23]
A catalytic cycle that involves CrCl2 has been proposed[24]
that comprises the two steps mentioned above for the conver-
sion of glucose to fructose as well as the additional main step
in the fructose ring closure, which is also observed in our
study. If we take the reported catalytic cycle as a starting
point, to understand the role of the hydroxyl-functionalized IL
4, our attention was focused initially on the interactions be-
tween the solvent, IL 4, the catalyst, and the substrate. As
mentioned above, CrCl2 has been used in combination with ILs
for the conversion of glucose to HMF, however, the nature of
the possible IL–CrCl2 adduct has not been investigated in
detail.
Figure 5. The molecular structure of [Cr(C2OHmim)Cl3] in the solid state.
Only one of the two complexes in the asymmetric unit is shown. Selected
bond lengths [ꢁ] and angles [8]: Cr(1)ꢁO(1) 2.073(7), Cr(1)ꢁCl(2) 2.401(2),
Cr(1)ꢁCl(1) 2.407(3), Cr(1)ꢁCl(3) 2.409(3), O(1)ꢁC(6) 1.433(12), N(1)ꢁC(1)
1.333(13), N(1)ꢁC(3) 1.385(12), N(1)ꢁC(5) 1.446(13), N(2)ꢁC(1) 1.346(13), N(2)ꢁ
C(2) 1.400(13), N(2)ꢁC(4) 1.476(15), O(1)ꢁCr(1)ꢁCl(2) 174.6(2), O(1)ꢁCr(1)ꢁ
Cl(1) 89.8(2), Cl(2)ꢁCr(1)ꢁCl(1) 90.58(9), O(1)ꢁCr(1)ꢁCl(3) 85.7(2), Cl(2)ꢁCr(1)ꢁ
Cl(3) 94.59(9), Cl(1)ꢁCr(1)ꢁCl(3) 171.61(11), C(6)ꢁO(1)ꢁCr(1) 128.8(6), C(1)ꢁ
N(1)ꢁC(3) 106.9(8), C(1)ꢁN(1)ꢁC(5) 124.8(8), C(3)ꢁN(1)ꢁC(5) 127.7(9), C(1)ꢁ
N(2)ꢁC(2) 106.7(9), C(1)ꢁN(2)ꢁC(4) 125.0(9), C(2)ꢁN(2)ꢁC(4) 128.2(9), N(1)ꢁ
C(1)ꢁN(2) 110.4(9).
In the absence of the substrate, equimolar quantities of 4
and CrCl2 were reacted in ethanol under a N2 atmosphere to
afford a zwitterionic complex [Cr(C2OHmim)Cl3] (C2OHmim=1-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
4
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Scheme 3. Reaction pathways for the dehydration of glucose into fructose followed by HMF formation through an open-chain pathway[23d] (beige) and a path-
way that involves a cyclic fructofuranosyl intermediate[23e] (blue). The labeling scheme for the carbon atoms in [C2OHmim]Cl is also provided.
bond between the coordinated OH group of glucose and the
OH group of one of the [C2OHmim]+ cations contributes to at
least part of the value.[20] The [Cr(C2OHmim)Cl4][C2OHmim]
adduct is also able to form a rather weak complex with tolu-
ene through an h2 bond between the C2C3 atoms of the aro-
matic ring (DEi =1.3 kcalmolꢁ1, see Figure 6 bottom and Sup-
porting Information for further details) to the CrII ion. Such an
interaction is highly reversible but stabilizes the active catalyst
to prevent deactivation and/or further interactions of the cata-
lyst with HMF that could lead to the formation of humins. No-
tably, all the calculations discussed above correspond to a CrII
ion in the high-spin (HS) state (S=4/2). The choice of the spin
state is supported by previous studies on related systems,[24a]
and a comparison of the energetic stabilities of the HS and
low-spin (LS) forms of several computed models. In all cases
the HS forms were considerably more stable than the LS
forms. Thus, the weak, reversible interaction of the catalyst
with toluene and, presumably, the aromatic ring in 9, becomes
especially important as the substrate is consumed. In keeping
with these calculations, the interaction of the aromatic ring in
9 with the Cr ion is not observed by NMR spectroscopy (Fig-
ure S1), and the formation of a stronger p-bound system may
also be excluded based on these spectra.
troscopic data (Figure S2) indicate that glucose interacts with
the IL ring protons and hydroxyl group through C6, C5, and
C1. Following the introduction of CrCl2 into the system, the in-
teraction between the hydroxyl group on 4 and C1 of glucose
is enhanced in agreement with the calculations (see above).
This strong hydrogen bond presumably activates the glucose
to a greater extent than that observed with 1, which would
form a weaker H bond that involves only ring protons of the
imidazolium cation. Thus, the functional group in 4 improves
the ability of the IL to interact with the substrate. In addition,
computational fitting revealed the presence of an extra peak,
which may be attributed to an interaction between the Cr ion
and C2 of glucose. Combined, these two interactions should
enhance the opening of the glucopyranose ring and hence the
formation of the linear form of glucose (evidenced in the
13C NMR spectrum) and the subsequent proton transfer from
C1 to C2.
The system was then heated to 878C to monitor the forma-
tion of HMF (Figure 7). This low temperature was used so that
the reaction could be monitored over a 12 h period. As expect-
ed, the signals that correspond to HMF increase in relative in-
tensity and those attributable to glucose and fructose de-
crease. No discernible carbonyl signal for fructose was ob-
served, which indicates that the amount of the linear form of
fructose is negligible during the reaction and suggests that
HMF formation proceeds through the cyclic fructofuranosyl
pathway. Based on the integration of the NMR signals, the rate
constant of the reaction was estimated by using the same rate
NMR spectroscopy was used to investigate the interactions
of 4 with 13C-enriched glucose, initially in the absence of CrCl2.
It was assumed that any carbon atom directly or indirectly
linked to another component in the reaction mixture would
result in the splitting of the signal in the spectrum. The spec-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
5
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 7. In situ 13C NMR spectra of the transformation of 13C-enriched glu-
cose to HMF in 4 with CrCl2; t in h.
Conclusions
The role of substrate–ionic liquid (IL) interactions in the activa-
tion of reactions is of considerable interest as it is imperative
that reactions conducted in ILs are superior to their counter-
parts conducted in water or organic solvents.[28] Herein, we de-
scribe a new functionalized ILs that efficiently catalyse the de-
hydration of glucose, cellobiose, and maltotetraose to 5-hy-
droxymethylfurfural in combination with the CrCl2 precatalyst
under relatively mild conditions (simple ILs are essentially inac-
tive with cellobiose and maltotetraose). The enhancements in-
duced by 4 were traced to two main factors. First, in the active
catalyst species, [Cr(C2OHmim)Cl4]ꢁ, the coordinated hydroxyl
group is readily displaced by the glucose substrate and,
second, specific hydrogen bonds between 4 and 9 have been
delineated at a molecular level.
Experimental Section
All starting materials were obtained from commercial sources and
used as received. IL 1 was purchased from Aldrich, and ILs 2–5 and
7 were prepared using literature procedures.[19,29] All ILs were pre-
pared under an inert atmosphere in dry solvents using Schlenk
techniques. ESI-MS were recorded by using a ThermoFinnigan LCQ
Deca XP Plus quadrupole ion trap instrument in positive ion
mode.[30] NMR spectra were recorded by using a Bruker DMX 400
instrument. The spectra for mechanistic investigations were fitted
with WINNMR, GNMR 4.0, and NMRICMA/MATLAB programs on
a PC (nonlinear least-squares fit to determine the spectral parame-
ters).
Figure 6. Computed structures: top: [Cr(C2OHmim)Cl4][C2OHmim] adduct de-
rived from the reaction of the zwitterion [Cr(C2OHmim)Cl3] with 4. Center:
the [Cr(C2OHmim)Cl4][C2OHmim] adduct and a OH group of glucose that
form a distorted trigonal bipyramid geometry in which one of the two
[C2OHmim]+ cations forms a hydrogen bond with the (acidic) H atom of the
coordinated OH group. Bottom: the [Cr(C2OHmim)Cl4][C2OHmim]–toluene
adduct, which involves an h2 interaction.
constants for the formation of HMF and glucose decomposi-
tion, which gave a pseudo-first-order rate constant of (0.18ꢂ
0.02) hꢁ1 (Figure S4).
Synthesis of [C2OHbim]Cl (6)
1-Butylimidazole (0.074 mol, 9.23 g) and 2-chloroethanol
(0.074 mol, 6 g) were stirred at 808C for 48 h. The resulting IL was
washed with diethyl ether (3ꢂ10 mL) and dried under vacuum
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
6
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
1
(yield=90%). H NMR (400 MHz, D2O, 258C): d=8.71 (s, 1H), 7.47
the appropriate time. After the reaction, the mixture was cooled to
RT and diluted with water and methanol (3:1, 10 mL) for further
analysis.
(s, 2H, overlapped), 4.26 (t, J=4 Hz, 2H), 4.15 (t, J=4 Hz, 2H), 3.87
(t, J=4 Hz, 2H), 1.79 (m, J=8 Hz, 2H), 1.26 (m, J=8 Hz, 2H),
0.85 ppm (t, J=8 Hz, 3H); 13C NMR (D2O, 258C): d=122.5, 59.8,
51.6, 49.4, 31.2, 18.8, 12.7 ppm; MS (ESI): m/z: 169.08 [cation]+; ele-
mental analysis calcd (%) for C9H17ClN2O (204.10): C 52.81, H 8.37,
N 13.69; found: C 53.4, H 8.37, N 14.09.
Dehydration of starch
In a typical reaction, starch (0.31 mmol, 50 mg), CrCl2 (0.130 mmol,
16 mg), 4 (900 mg), and 9 (100 mg) were placed in a glass vial. The
vial was sealed, and the reaction mixture was heated to 1408C for
4 h. After the reaction, the mixture was cooled to RT and diluted
with water and methanol (3:1, 10 mL) for further analysis.
Synthesis of [(C2OH)2im]Cl (8)
IL 8 has been reported previously,[31] however, we obtained a solid-
state product that is analytically pure. A mixture of trimethylsilyli-
midazole (0.10 mol, 14.04 g) and ClCH2CH2OSi(CH3)3 (0.20 mol,
30.60 g) was heated at 608C for 24 h. Diethyl ether (50 mL) was
added, and the solid was collected by filtration. The solid was then
added to an aqueous HCl solution (50 mL, 4.0m) in five portions.
The resulting mixture was washed with diethyl ether (3ꢂ30 mL)
and the remaining aqueous phase was evaporated under vacuum
(1008C, 24 h) to give an off-white solid (yield=95%). 1H NMR
(400 MHz, D2O, 258C): d=8.79 (s, 1H), 7.48 (s, 2H), 4.26 (t, J=
5.0 Hz, 4H), 3.85 ppm (t, J=5.0 Hz, 4H); 13C (D2O, 258C): d=136.1,
122.6, 59.7, 51.6 ppm; MS (ESI): m/z: 157.0 [cation]+; elemental
analysis calcd (%) for C7H13ClN2O2 (192.07): C 43.64, H 6.80, N 14.54;
found: C 43.14, H 6.80, N 14.08.
Dehydration of cellulose
In
a typical reaction, cellulose (0.62 mmol, 100 mg), CrCl2
(0.049 mmol, 6 mg), 4 (900 mg), 9 (100 mg), and HCl (2 mL) were
placed in a glass vial. The vial was sealed, and the reaction mixture
was heated to 1608C for 30 min. After the reaction, the mixture
was cooled to RT and diluted with water and methanol (3:1,
10 mL) for further analysis.
Analysis
HMF was analyzed by HPLC by using an Agilent 1260 Infinity in-
strument equipped with a UV/Vis detector and a Poroshell 120 EC-
C18 column (3.0ꢂ100 mm, 2.7 mm) using water/acetonitrile (97:3) as
the mobile phase at a flow rate of 0.5 mLminꢁ1 at 308C. Glucose
was analyzed by using an evaporative light scattering detector
Synthesis of [C2OHC2Phim]Cl (9)
(2-Chloroethyl)benzene (0.089 mol, 12.5 g) and 1-(2-hydroxyethyl)i-
midazole (0.089 mol, 10 g) were stirred together at 908C for 72 h.
The resulting IL was washed with diethyl ether (5ꢂ100 mL), and
a yellow solid was recovered (yield=95%). 1H NMR (400 MHz,
DMSO, 258C): d=9.21 (s, 1H), 7.78 (s, 1H), 7.75 (s, 1H), 7.22–7.33
(m, 5H), 5.35 (t, J=4.4), 4.46 (t, J=8 Hz, 2H), 4.20 (t, J=4,4 Hz,
2H), 3.70 (s, 1H), 3.15 ppm (t, J=8 Hz, 2H); 13C NMR (DMSO, 258C):
d=137.4, 136.8, 129.2, 129.1, 127.3, 123.2, 122.7, 59.7, 52.1, 50.2,
35.9 ppm; MS (ESI): m/z: 217.13 [cation]+; elemental analysis calcd
(%) for C13H17ClN2O (252.10): C 61.78, H 6.78, N 11.08; found: C
60.76, H 6.58, N 11.53.
(ELCD) and
a COSMOSIL Packed column Sugar-D (3.0 I.D.ꢂ
250 mm) using a water/acetonitrile (30:70) mobile phase at a flow
rate of 0.5 mLminꢁ1 at 408C. Cellobiose and maltotetraose were
determined with a refractive index detector and a Hi-Plex H,
300*7.7 column at 608C using acidified water (0.004m H2SO4) as
the eluent (0.5 mLminꢁ1 flow rate).
Computations
DFT calculations were performed by using the ORCA suite of pro-
grams.[32] The hybrid functional PBE0[33] was used together with the
small def2-SV(P)[34] basis set for optimization of all the structures.
As DFT is not able to account for van der Waals dispersion forces,
which contribute substantially to the weak intermolecular interac-
tions, dispersion corrections were applied within the recently de-
veloped approach DFT-D3[35] together with the Becke–Johnson (BJ)
damping function[36] (indicated by the -D3 appended to the func-
tional name). All stationary points were characterized as minima by
analysis of the Hessian matrices. After that, the single point (SP)
calculations within the COSMO continuum solvation model[37] to
simulate a “real” reaction environment were performed for the op-
timized structures within the same functional and large def2-TZVP
basis set.[38] COSMO parameters of ethanol, as implemented in the
ORCA program package, were chosen for the IL. The SP energy
values were used for the estimation of the interaction energies DEi
[Eq. (1)].
Dehydration of glucose
In a typical reaction, glucose (0.28 mmol, 50 mg), CrCl2 (6 mol%
with respect to glucose), and IL (500 mg) were placed in a glass
vial. The vial was sealed, and the reaction mixture was heated to
1008C for the appropriate time. After the reaction, the mixture was
cooled to RT and diluted with water and methanol (4:1, 25 mL) for
further analysis.
Dehydration of cellobiose
In
a typical reaction, cellobiose (0.15 mmol, 50 mg), CrCl2
(0.098 mmol, 12 mg), and IL (1 g) were placed in a glass vial. The
vial was sealed, and the reaction mixture was heated to 1008C for
the appropriate time. After the reaction, the mixture was cooled to
RT and diluted with water and methanol (3:1, 20 mL) for further
analysis.
Two [C2OHmim]Cl ion pairs per Cr(II) center were explicitly included
in the models to account for the influence of the IL. The addition
of a molecule of glucose or toluene to the model resulted in stable
structures (Figure 6). The energy of the interaction (DEi) between
the glucose/toluene molecule (M) and the [Cr(C2OHmim)Cl4]-
[C2OHmim] adduct (A) was calculated as the difference of the
PBE0-D3/def2-TZVP SP energies of the whole complex (C) and the
molecule and the adduct:
Dehydration of maltotetraose
In a typical reaction, maltotetraose (0.075 mmol, 50 mg), CrCl2
(0.163 mmol, 20 mg), and IL (1 g) were placed in a glass vial. The
vial was sealed, and the reaction mixture was heated to 1008C for
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
7
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[15] J. D. Scholten, B. C. Leal, J. Dupont, ACS Catal. 2012, 2, 184–200.
[16] a) H. Tadesse, R. Luque, Energy Environ. Sci. 2011, 4, 3913–3929.
[17] a) R. P. Swatloski, S. K. Spear, J. D. Holbrey, R. D. Rogers, J. Am. Chem.
Soc. 2002, 124, 4974–4975; b) D. A. Fort, R. C. Remsing, R. P. Swatloski,
P. Moyna, G. Moyna, R. D. Rogers, Green Chem. 2007, 9, 63–69; c) H.
Wang, G. Gurau, R. D. Rogers, Chem. Soc. Rev. 2012, 41, 1519–1537;
d) J. S. Moulthrop, R. P. Swatloski, G. Moyna, R. D. Rogers, Chem.
Commun. 2005, 1557–1559; e) T. Stꢄhlberg, W. Fu, J. M. Woodley, A.
Riisager, ChemSusChem 2011, 4, 451–458.
[18] S. Tang, G. A. Baker, H. Zhao, Chem. Soc. Rev. 2012, 41, 4030–4066.
[19] Z. Fei, W. H. Ang, D. Zhao, R. Scopelliti, E. E. Zvereva, S. A. Katsyuba, P. J.
Dyson, J. Phys. Chem. B 2007, 111, 10095–10108.
[20] S. A. Katsyuba, M. V. Vener, E. E. Zvereva, Z. Fei, R. Scopelliti, G. Laurenc-
zy, N. Yan, E. Paunescu, P. J. Dyson, J. Phys. Chem. B 2013, 117, 9094–
9105.
[21] C. Shi, Y. Zhao, J. Xin, J. Wang, X. Lu, X. Zhang, S. Zhang, Chem.
Commun. 2012, 48, 4103–4105.
[22] Y. Qu, C. Huang, Y. Song, J. Zhang, B. Chen, Bioresour. Technol. 2012,
121, 462–466.
[23] a) C. Li, Z. K. Zhao, A. Wang, M. Zheng, T. Zhang, Carbohydr. Res. 2010,
345, 1846–1850; b) M. J. Antal, Jr., W. S. L. Mok, G. N. Richards, Carbo-
hydr. Res. 1990, 199, 91–109; c) R. S. Assary, P. C. Redfern, J. Greeley,
L. A. Curtiss, J. Phys. Chem. B 2011, 115, 4341–4349; d) C. Moreau, R.
Durand, S. Razigade, J. Duhamet, P. Faugeras, P. Rivalier, P. Ros, G. Avig-
non, Appl. Catal. A 1996, 145, 211–224; e) Y.-N. Li, J.-Q. Wang, L.-N. He,
Z.-Z. Yang, A.-H. Liu, B. Yu, C.-R. Luan, Green Chem. 2012, 14, 2752–
2758.
DEi ¼ EðCÞꢁEðMÞꢁEðAÞ
ð1Þ
The total charge of all considered models was zero.
XRD analysis
Diffraction data were collected at low temperature (140(2) K) using
MoKa radiation by using a mar345dtb system in combination with
a Genix Hi-Flux small-focus generator (marmX system). Data reduc-
tion was performed by using automar.[39] The solution and refine-
ment was performed by using SHELX.[40] The crystal structure was
refined by using full-matrix least-squares based on F2 with all non-
hydrogen atoms defined anisotropically. Hydrogen atoms were
placed in calculated positions by means of the riding model. Pseu-
domerohedral twinning was found during the last stages of refine-
ment and treated by the TWINROTMAT algorithm of PLATON[41] to
obtain two BASF values (parameters representing the fractional
contributions of the various twin components): 0.29(5) and
0.044(2). CCDC 988678 contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[24] a) E. A. Pidko, V. Degirmenci, R. A. van Santen, E. J. M. Hensen, Angew.
Chem. Int. Ed. 2010, 49, 2530–2534; Angew. Chem. 2010, 122, 2584–
2588; b) E. A. Pidko, V. Degirmenci, E. J. M. Hensen, ChemCatChem 2012,
4, 1263–1271.
[25] H. R. Oswald, Helv. Chim. Acta 1961, 44, 1049–1054.
[26] N. Wooster, Z. Kristallogr. Kristallgeom. Kristallphys. Kristallchem. 1930,
74, 363–374.
Acknowledgements
This work was supported by the Swiss National Science Founda-
tion National Research Programme (NRP 66). We thank Kim
Meyer for preparing the Table of Contents image.
[27] S. Aldridge, M. Shang, T. P. Fehlner, Acta Crystallogr. Sect. C 1998, 54,
47–49.
Keywords: biomass · carbohydrates · hydrogen bonds · ionic
liquids · reaction mechanisms
[28] a) J. Zhang, H. Zhang, J. Wu, J. Zhang, J. He, J. Xiang, Phys. Chem. Chem.
Phys. 2010, 12, 1941–1947; b) T. G. A. Youngs, J. D. Holbrey, C. L. Mullan,
S. E. Norman, M. C. Lagunas, C. D’Agostino, M. D. Mantle, L. F. Gladden,
D. T. Bowron, C. Hardacre, Chem. Sci. 2011, 2, 1594–1605; c) N. Yan, Y.
Yuan, R. Dykeman, Y. Kou, P. J. Dyson, Angew. Chem. Int. Ed. 2010, 49,
5549–5553; Angew. Chem. 2010, 122, 5681–5685.
[29] L. C. Branco, J. N. Rosa, J. J. Moura Ramos, C. A. M. Afonso, Chem. Eur. J.
2002, 8, 3671–3677.
[30] P. J. Dyson, J. S. McIndoe, D. Zhao, Chem. Commun. 2003, 508–509.
[31] Y.-B. Xiong, H. Wang, Y.-J. Wang, R.-M. Wang, Polym. Adv. Technol. 2012,
23, 835–840.
[32] F. Neese, ORCA – an ab initio, density functional and semiempirical
program package, Ver. 2.8 (Rev 17), University of Bonn, Germany 2010..
[33] a) B. V. Adamo, J. Chem. Phys. 1999, 110, 6158–6170; b) J. P. Perdew, K.
Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865–3868.
[34] a) W. R. Wadt, P. J. Hay, J. Chem. Phys. 1985, 82, 270–283; b) W. R. Wadt,
P. J. Hay, J. Chem. Phys. 1985, 82, 284–298; c) W. R. Wadt, P. J. Hay, J.
Chem. Phys. 1985, 82, 299–310.
[35] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132,
154104.
[36] S. Grimme, S. Ehrlich, L. Goerigk, J. Comput. Chem. 2011, 32, 1456–
1465.
[1] a) P. Gallezot, Chem. Soc. Rev. 2012, 41, 1538–1558; b) J. C. Serrano-Ruiz,
R. Luque, A. Sepulveda-Escribano, Chem. Soc. Rev. 2011, 40, 5266–5281;
c) M. E. Zakrzewska, E. Bogel-Łukasik, R. Bogel-Łukasik, Chem. Rev. 2011,
111, 397–417; d) S. Liu, L. P. Abrahamson, G. M. Scott, Biomass Bioenergy
2012, 39, 1–4; e) D. M. Alonso, J. Q. Bond, J. A. Dumesic, Green Chem.
2010, 12, 1493–1513.
[2] 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.
[3] a) H. Zhao, J. E. Holladay, H. Brown, Z. C. Zhang, Science 2007, 316,
1597–1600; b) Y. Romꢃn-Leshkov, J. N. Chheda, J. A. Dumesic, Science
2006, 312, 1933–1937; c) J. C. Serrano-Ruiz, J. A. Dumesic, Energy Envi-
ron. Sci. 2011, 4, 83–99.
[4] A. A. Rosatella, S. P. Simeonov, R. F. M. Frade, C. A. M. Afonso, Green
Chem. 2011, 13, 754–793.
[5] http://www.ava-biochem.com/pages/en/home.php.
[6] S. Ghazali-Esfahani, H. Song, E. Paunescu, F. D. Bobbink, H. Liu, Z. Fei, G.
Laurenczy, M. Bagherzadeh, N. Yan, P. J. Dyson, Green Chem. 2013, 15,
1584–1589.
[7] B. R. Caes, M. J. Palte, R. T. Raines, Chem. Sci. 2013, 4, 196–199.
[8] Y.-B. Huang, Y. Fu, Green Chem. 2013, 15, 1095–1111.
[9] J. Ma, W. Yu, M. Wang, X. Jia, F. Lu, J. Xu, Chin. J. Catal. 2013, 34, 492–
507.
[10] N. Shi, Q. Liu, Q. Zhang, T. Wang, L. Ma, Green Chem. 2013, 15, 1967–
1974.
[11] J. B. Binder, R. T. Raines, J. Am. Chem. Soc. 2009, 131, 1979–1985.
[12] H. Ma, B. Zhou, Y. Li, D. S. Argyropoulos, BioResources 2012, 7, 533–544.
[13] V. Choudhary, S. H. Mushrif, C. Ho, A. Anderko, V. Nikolakis, N. S. Marin-
kovic, A. I. Frenkel, S. I. Sandler, D. G. Vlachos, J. Am. Chem. Soc. 2013,
135, 3997–4006.
[37] F. Eckert, A. Klamt, AIChE J. 2002, 48, 369–385.
[38] a) F. Weigend, M. Hꢅser, H. Patzelt, R. Ahlrichs, Chem. Phys. Lett. 1998,
294, 143–152; b) F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005,
7, 3297–3305.
[39] automar, release 2.8.0, Marresearch GmbH, Germany 2011.
[40] G. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122.
[41] A. Spek, Acta Crystallogr. Sect. D 2009, 65, 148–155.
Received: December 17, 2013
Revised: February 20, 2014
Published online on && &&, 0000
[14] C. M. Cai, T. Zhang, R. Kumar, C. E. Wyman, Green Chem. 2013, 15,
3140–3145.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
8
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
Solvents by design: 5-Hydroxymethyl-
furfural (HMF) is a key platform chemical
that may be obtained from various cel-
lulosic (biomass) derivatives. Ionic liq-
uids (ILs) facilitate the catalytic conver-
sion of glucose into HMF. We demon-
strate that the careful design of the IL
cation leads to new ionic solvents that
enhance the transformation of glucose
and more complex carbohydrates into
HMF.
S. Siankevich, Z. Fei, R. Scopelliti,
G. Laurenczy, S. Katsyuba, N. Yan,*
P. J. Dyson*
&& – &&
Enhanced Conversion of
Carbohydrates to the Platform
Chemical 5-Hydroxymethylfurfural
Using Designer Ionic Liquids
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
9
&
ÞÞ
These are not the final page numbers!