Rapid, High-Yield Fructose Dehydration to 5-Hydroxymethylfurfural in Mixtures of Water and the Noncoordinating Ionic Liquid [bmim][OTf]

Article abstract of DOI:10.1002/cssc.201901529

The noncoordinating ionic liquid [bmim][OTf] (bmim=1-butyl-3-methylimidazolium) is an effective and versatile solvent for the high-yield dehydration of fructose to the platform chemical 5-hydroxymethylfurfural (HMF) over short reaction times. In contrast to prior studies in which low yields were obtained for this transformation in ionic liquids (ILs) with noncoordinating anions, this contribution reveals that the water content is an essential parameter for an efficient reaction in ILs. Achieving the optimum amount of water can increase the yield dramatically by regulating the acidity of the catalyst and partially suppressing the side reaction caused by self-condensation of HMF. Using acid catalysis in [bmim][OTf] with 3.5 % water content, yields above 80 % can be achieved at 100 °C in only 10 min, even at high (14 %) fructose loading. These results also suggest that [bmim][OTf] represents a superior medium for solvent extraction of HMF compared to halide-based ILs, allowing the option of isolation or further valorization of the HMF formed.

Full text of DOI:10.1002/cssc.201901529

Accepted Article
Title: Rapid, High-Yield Fructose Dehydration to HMF in Mixtures of
Water and the Non-Coordinating Ionic Liquid [bmim][OTf]
Authors: Amir Al Ghatta, James Wilton-Ely, and Jason Hallett
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.201901529
Link to VoR: http://dx.doi.org/10.1002/cssc.201901529
A Journal of
www.chemsuschem.org

10.1002/cssc.201901529
ChemSusChem
Rapid, High-Yield Fructose Dehydration to HMF in Mixtures of Water and the Non-
Coordinating Ionic Liquid [bmim][OTf]
Amir Al Ghatta,[a,b] James D. E. T. Wilton-Ely,*[b] and Jason P. Hallett*[a]
[a] A. Al Ghatta, Prof. Dr. J. P. Hallett
Department of Chemical Engineering
Imperial College London
South Kensington Campus, London SW7 2AZ (UK)
E-mail: j.hallett@imperial.ac.uk
[b] A. Al Ghatta, Prof. Dr. J. D. E. T. Wilton-Ely
Department of Chemistry
Imperial College London
Molecular Sciences Research Hub
White City Campus, London W12 0BZ (UK)
E-mail: j.wilton-ely@imperial.ac.uk
Keywords: biorenewables, catalysis, ionic liquids, platform chemicals, sustainable chemistry
ABSTRACT: The non-coordinating ionic liquid [bmim][OTf] is an effective and versatile solvent for the high-yield
dehydration of fructose to the platform chemical HMF over short reaction times. In contrast to literature reports,
which report low yields for this transformation in ionic liquids (ILs) with non-coordinating anions, this contribution
reveals that the water content is an essential parameter for an efficient reaction in ILs. Achieving the optimum
amount of water can increase the yield dramatically by regulating the acidity of the catalyst and partially suppressing
the side reaction caused by self-condensation of HMF. Using acid catalysis in [bmim][OTf] with 3.5% water content,
yields above 80% can be achieved at 100 °C in only 10 minutes, even at high (14%) fructose loading. These results
suggest that [bmim][OTf] represents a superior medium for solvent extraction of HMF compared to halide-based
ILs, allowing the option of isolation or further valorisation of the HMF formed.
INTRODUCTION
The depletion of oil reserves and the increase in regulation of CO2 emissions have increased the demand for
renewable chemicals that can replace oil-derived products.[1] While specific targets have been established for
biofuels, little planning has been proposed for bio-derived chemicals, perhaps due to the small fraction of the total
oil production used for chemical products compared to that used for transport fuel. The replacement of common oil
derivatives with renewable platform chemicals is a significant challenge mainly due to the high cost of the
technologies involved. In this field, sugars are seen as the most suitable feedstock for the production of a wide
range of green chemicals. This sector has experienced substantial investments in research but little commercial
activity. In a report submitted to the European Commission Directorate on General Energy, a consortium led by
E4tech, RE-CORD and WUR[2] discussed 94 green products, including chemicals, biofuels and polymers, which
were identified as accessible from sugars obtained from the pretreatment of sustainable biomass feedstocks. These
products could replace the majority of all oil-derived chemicals. The sugars identified as having the potential to be
converted into valuable products are glucose, fructose, galactose, xylose, arabinose, ribose, lactose, sucrose and
maltose. Of the 94 products obtained from these sugars, 25 were selected as involving feasible technologies that
are suitable for large scale prototype testing in a pilot plant.[2] 5-hydroxymethylfurfural (HMF) represents a promising
intermediate product from sugars, capable of replacing a large variety of petrochemical derivatives, so its
generation in high-yield is an important aspect in the scale up of the production of this chemical.[3] The main route
to form HMF is by the dehydration of sugars,[4] and important results have been achieved for HMF production using
ionic liquids.[5-9] The use of ionic liquids to produce biorenewable chemicals has illustrated many of the advantages
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
associated with ionic liquid media, such as their low volatility and non-flammability, even if concerns have been
raised more recently over the toxicity of these solvents.[10] Deep eutectic solvents and water have also been
investigated for the dehydration of fructose into HMF but improved yields are still needed.[11-13] Other research has
focused on optimising heterogeneous catalysts for the dehydration reaction both involving metal catalysts[14-20] and
supported ionic liquids.[21,22]
In particular, ionic liquids with halide counteranions have proved to be very efficient for the production of
HMF from fructose, glucose and cellulose with both Brønsted and Lewis acid catalysts.[23-25] Due to the ability of
halide-based ionic liquids to dissolve sugars and cellulose through their hydrogen bonding ability,[26] these media
have proved very attractive for dehydration reactions. In many cases, superior performance is observed when
compared to organic solvents and water in terms of loading, reaction time and process conditions.[27,28] In particular,
chromium and tin chloride catalysts have proved extremely efficient in chloride and bromide ionic liquids.[29] While
catalytic systems that use glucose as a substrate are limited, more options are reported for fructose dehydration.
In an acidic ionic liquid with a coordinating anion, [bmim][HSO4] / CrCl3·6H2O has been reported as a very efficient
system to obtain a quantitative yield of HMF from fructose at high loadings (1:7 with respect to the ionic liquid).[30]
Alternatively, [bmim]Br and [Hmim]Cl can dehydrate fructose in high yield with no added catalyst. Of the two
systems, [Hmim]Cl proved to be more efficient, achieving 92% yield at 90 °C in 45 min while [bmim]Br required
heating to 100 °C and 1 hour to obtain the same yield.[31,32] The ionic liquid [bmim]Cl can perform catalyst-free high
yield dehydration (97%) under microwave irradiation at high loading,[33] however this heating methodology is not
suitable for scale-up. The addition of an acid can enhance the reaction in non-protic ionic liquids. Li and co-workers
managed to achieve near quantitative yields of HMF in [bmim]Cl or [bmim]Br with HCl in 10 minutes at 80 °C, while
the dehydration was unsuccessful with HCl in non-halogenated and non-coordinating ionic liquids, such as
[bmim][BF4] and [bmim][PF6], achieving less than 10% yield in 30 minutes.[34] A 52% yield was obtained in 3 hours
using Amberlyst-15 in [bmim][PF6] at 80 °C, with the process showing improvement (80% yield) on addition of
DMSO.[35] A triphasic system based on water and the hydrophobic ionic liquid [bmim][NTf2] led to 81% HMF yield
in 24 hours using a vanadium phosphate catalyst.[36] However, in general, satisfactory results have not been
obtained in ionic liquids with non-coordinating anions such as [BF4]-, [PF6]-, [OTf]- and [NTf2]-. Shi and co-workers
analysed the effect of anions on the dehydration of fructose and, through a molecular dynamics study, showed that
chloride anions, in contrast to triflate, tend to form a centralised structure around the fructose molecule through
hydrogen bonds.[37] This finding is in agreement with others who used IR spectroscopy and molecular dynamics
calculations to show that non-coordinating anions in ionic liquids, such as tetrafluoroborate, do not contribute to
hydrogen bonding with the hydroxyl group of HMF.[38] While successful results have been reported for ionic liquids
with chloride and bromide anions, the applicability of these reactions has always proved problematic due to the
difficulties encountered in the separation of HMF from the medium.[39,40] Different approaches have been proposed
for HMF recovery, such as the use of supercritical CO2 or high-vacuum entrained distillation, but the feasibility of
applying these techniques on a large scale is doubtful on account of the high cost associated with handling CO2 or
high vacuum.[41–43] Moreover, a better compromise between reaction time, heating method and catalyst loading still
needs to be achieved. Some researchers have focused on the oxidation of HMF in the reaction media directly to
2,5 furandicarboxylic acid (FDCA), which can be separated by water addition,[44,45] but this reaction proved to be
non-selective. A further issue is that the transition metal catalysts, which have proved to be the most selective for
this oxidation[46] are deactivated in the coordinating halide-rich environment.[47] In contrast, the oxidation of alcohols
is extremely efficient in non-coordinating ionic liquids with superior yields and selectivity to those reported in
conventional organic solvents.[48–52] However, the dehydration of sugars to HMF in these solvents is inefficient with
this poor performance being ascribed to the lack of hydrogen bonding. Therefore, an efficient chemical pathway
that enhances the HMF yield from sugars in non-coordinating solvents would make the process much more
attractive for both isolation of HMF and its further catalytic transformation. The absence of hydrogen bonding
between the hydroxyl group of HMF and the solvent decreases the affinity of HMF for these solvents, making its
extraction with organic solvents far easier, reducing the amount of extraction solvent needed. Therefore, the aim of
this contribution is to focus on improving the yield of HMF in non-halogenated and non-coordinating ionic liquids,
such as [bmim][OTf] (Figure), which are more suitable media for further reaction of HMF. In order to increase the
yield over short reactions times, analysis was performed on a series of different reaction parameters. Water content
2
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
proved to be the most crucial parameter for solvent design as it stabilises HMF towards side reactions. This report
demonstrates that the reason behind the low selectivity achieved so far in non-coordinating ionic liquids lies in the
instability of HMF under acidic conditions. Water addition suppresses side reactions and stabilises the HMF, leading
to high yields of HMF from the dehydration of fructose. Furthermore, this is achieved over short reaction times and
with high substrate loadings, making [bmim][OTf] as efficient for this transformation as the halide-based ionic liquids
reported in literature.
Figure 1. The cations and anions used in the ionic liquids in this contribution.
EXPERIMENTAL SECTION
Materials. Fructose, 36% hydrochloric acid, calcium hydride, phosphorus pentoxide, lithium triflate and potassium
hydroxide and [emim][OTf] were purchased from Sigma Aldrich. 5-Hydroxymethylfurfural (99%, chromatography
standard) was obtained from Fluorochem and purified by crystallisation at 3 °C from ethyl acetate and hexane. N-
methylimidazole, N-methylpyrrolidine and chlorobutane for ionic liquid synthesis were purchased from Alfa Aesar.
Methylimidazole was purified at reduced pressure (2.1 mbar. 40 °C reflux temperature) in the presence of calcium
hydride. Chlorobutane was dried at atmospheric pressure over phosphorus pentoxide. Methylpyrrolidine was
distilled at atmospheric pressure over KOH. Acetonitrile, ethyl acetate and hexane were purchased from VWR and
further purified by distillation over calcium hydride.
Synthesis of [bmim]Cl and [bmpyr]Cl. [bmim]Cl was synthesised on a large scale (350 g). Manipulation of all
chemicals was performed by cannulation under inert conditions. 1-methylimidazole (160 mL, 2 mol) was transferred
to a three neck flask equipped with an air condenser followed by the addition of acetonitrile (100 mL). The mixture
was heated at 60 °C under stirring and chlorobutane (230 mL, 2.2 mol, 1.1 eq) was then added dropwise using a
dropping funnel. The reaction mixture was heated for 3 days at 70 °C. Complete consumption of 1-methylimidazole
was confirmed by 1H NMR spectroscopy. [bmim]Cl was precipitated by gradual addition of ethyl acetate at 3 °C.
The resultant white crystals were washed under inert conditions with portions of ethyl acetate (2 L in total), dried
under vacuum and stored in a glovebox. [bmpyr]Cl was synthesised by the same procedure using N-
methylpyrrolidine. Data for the products were in good agreement with those reported previously.[30,53]
Synthesis of [bmim][OTf] and [bmpyr][OTf]. [bmim]Cl (30 g, 0.172 mol) was dissolved in water (100 mL) and
Li[OTf] (30 g, 0.179 mol) was dissolved in water (50 mL). The lithium triflate solution was added dropwise to the
chloride solution under vigorous stirring and the reaction left overnight. The product was extracted from aqueous
solution using a total of 800 mL dichloromethane. The organic extracts were washed repeatedly with water (5 mL)
until the silver nitrate test revealed no halide ions (5-6 washes). The organic phase was dried with magnesium
sulfate and then evaporated to dryness under reduced pressure. The ionic liquid, [bmim][OTf] was further dried
under reduced pressure overnight at 60 °C (final yield 96%) and handled thereafter under an inert atmosphere.
3
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
[bmpyr][OTf] was synthesised by the same procedure starting with [bmpyr]Cl. Data for the products were in good
agreement with those reported previously. See Supporting Information for spectra.[54,55]
Fructose conversion to HMF. Experiments were performed in triplicate in 1.5 mL vials. A larger quantity of the IL
was prepared with the desired amount of water and catalyst. This was divided into 1 g portions, which were used
for each experiment. Each vial was loaded with an appropriate amount of fructose followed by the addition of the
solvent to be tested in the desired fructose/solvent ratio. The vials were placed in a preheated heating block for the
desired reaction time. At the end of the reaction, the reaction mixture was rapidly quenched with 1 mL of water and
then further diluted 6:1 and filtered before HPLC analysis. Kinetic analysis was performed by halting the reaction
by addition of water at the desired time and recording the data.
Quantification of HMF yield was performed using a Shimadzu Prominence preparative HPLC (Shimadzu,
Kyoto, Japan) with refractive index (RI) and ultraviolet (UV) detectors and equipped with an Aminex HPX-87H
column using 0.005 M H2SO4 as the mobile phase at 55 °C and flow rate of 0.6 mL/min. See Supporting Information
for numerical data.
The isolation of HMF from [bmim][OTf] was performed on a larger scale, using 1 g of ionic liquid and 14%
fructose loading. 1H NMR analysis of the ionic liquid/water mixture after reaction showed only resonances for the
IL and HMF and no precipitation was observed on water addition. Extraction with diethyl ether from the IL phase
followed by solvent removal resulted in a viscous liquid (1H NMR spectrum, Figure S17 in the Supporting
Information). Washing with sodium carbonate solution was avoided due to the high affinity of HMF for the water
phase and its instability at high pH. Instead the ionic liquid phase was neutralised to pH 7 after addition of water
(20 mL). The solution was then washed with dichloromethane until the organic layer became clear (5 x 1 mL). The
purified water layer was then extracted with diethyl ether (3 x 100 mL) and concentrated under vacuum to obtain a
yellow liquid, which crystallised at – 4 °C to give 71% yield of HMF from fructose (Figures S18-20 in the Supporting
Information). Following this procedure, the ionic liquid was extracted from the water phase mentioned above using
ethyl acetate as a light brown viscous liquid. Analysis by 1H NMR spectroscopy and mass spectrometry showed no
degradation of the ionic liquid with only slight traces of HMF remaining in the ionic liquid phase (as evidenced by
1H NMR spectroscopy).
Partition coefficient measurements. The ionic liquids (300 mg) were mixed with HMF (30 mg) and heated at 60
°C for 30 minutes to obtain a homogeneous solution and then cooled to room temperature. Diethyl ether (1 g) was
added to the mixture and vigorously stirred for one hour. Aliquots from the organic phase were removed and
analysed using a ThermoFischer GC ULTRA with external standard calibration. The partition coefficient was
calculated using the following formula:
표푟ꢀ푎푛ꢁ푐 푝ℎ푎푠푒
퐻푀퐹
퐶
퐾 =
ꢁ표푛ꢁ푐 푙ꢁ푞푢ꢁ푑
퐻푀퐹
퐶
RESULTS AND DISCUSSION
Stability of HMF in the ionic liquid under acidic conditions
Literature studies on the mechanism of fructose dehydration appear to agree that a hydrogen bonding environment
is necessary for dehydration to be achieved.[37] However, the potential for hydrogen bonding is reduced in an ionic
liquid with a non-coordinating anion. To understand the behaviour of HMF in a non-coordinating environment, this
study began by analysing the stability of HMF in [bmim][OTf] under acidic conditions at elevated temperatures with
a range of water content. The stability of HMF has proven to be a key issue for the effective valorisation of the
dehydration reaction.[56] A solvent system capable of stabilising HMF is needed to improve the yield of reaction and
to avoid self-condensation of HMF over longer reaction times. Water has the potential to act as a hydrogen bond
acceptor, contributing to the stabilisation to HMF. In this study, p-toluenesulfonic acid (PTSA) was added to the
HMF in the IL at 90 °C for 10 minutes. These conditions were chosen since p-toluenesulfonic acid has already been
4
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
shown to be a good reagent for fructose dehydration in many different solvent systems when PTSA is used in its
immobilised form (Amberlyst-15).[57-63]
Figure 2. Degradation of HMF at 90 °C in [bmim][OTf] with a range of water content and 10% PTSA loading (with
respect to HMF) and 10% HMF loading (with respect to the ionic liquid).
The data in Figure 2 appeared to confirm our hypothesis as HMF proved to be unstable in the pure ionic
liquid, undergoing 43% degradation under the conditions tested. The gradual addition of water hindered the
formation of side products, decreasing the decomposition to just 14%. The origin of this behaviour could be linked
either to the hydrogen bonding of water with HMF or the decreasing activity of the protons on addition of increasing
amounts of water. However, no trace of further hydration to formic and levulinic acids is observed, suggesting that
the ionic liquid is contributing to the stability of HMF. To further investigate the role of water in the ionic liquid, a
stability test was performed at 1.3% water content with HCl and H2SO4, which exhibit a lower pKa than PTSA. The
results showed slight changes in the degree of degradation with less degradation observed for HCl (18.4 ± 0.8%)
and slightly more for H2SO4 (23.2 ± 0.2%) compared to PTSA (20.3%). The significant difference in acidity of these
Brønsted acid catalysts shows no strong correlation to the difference in degradation observed. This implies that the
presence of water is not decreasing the acidity of these catalysts substantially, but has more of an effect on the
stabilisation of HMF through hydrogen bonding.
Presence of intermediates
During the experiments on fructose dehydration, it was observed that HMF yield does not follow a first order
relationship to the substrate, fructose, suggesting the formation of intermediates during the dehydration process.
In all the conditions reported here, fructose conversion was almost complete in less than 1 minute at temperatures
higher than 90 °C, while HMF yield increases over a longer reaction time (3 to 10 min). The formation of
intermediates is indicated by the appearance of two peaks at 7.7 min and 8.1 min (compared to fructose at 9.3 min)
in the HPLC chromatogram. By analysing the kinetics of the reaction, it was observed that HMF yield does not
increase further when these two peaks were no longer present, indicating that the reaction is complete and the
substrate and the intermediates have been consumed. A mechanistic study of fructose dehydration in ethanol was
performed by Zhang et al[64] and the authors demonstrated that two principal intermediates were formed prior to
HMF formation, according to the mechanisms depicted in Scheme 1. Their analysis was performed by in situ NMR
spectroscopy and mass spectrometry, strongly supporting this reaction path. Due to the difficulty in quantifying the
5
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
intermediates, the rate limiting steps in the dehydration will not be discussed here. Instead, the reaction was
considered complete when the intermediate peaks in the chromatogram were no longer present.
Scheme 1. Fructose dehydration scheme proposed in previous work.[45]
Selection of the acid catalyst
According to the stability test described above, it was decided to operate under conditions where the decomposition
of HMF is minimised. However, the dehydration of fructose involves the generation of 3 moles of water per mole of
fructose, so it was decided to operate with a lower initial water content to avoid a possible excess that could reduce
the rate of reaction or lead to further undesired hydrolysis of HMF. An initial water content of 3.5% and a fructose
loading of 14% was selected. Under these conditions, complete dehydration of fructose will lead to a water content
of 5%. The acids, PTSA, H2SO4 and HCl were chosen for catalyst screening. The success of PTSA for the
dehydration of HMF has already been established in different systems and HCl has been shown to be an effective
catalyst in a wide variety of solvents. For example, Tuerke and co-workers obtained 82% yield at 30% fructose
loading in 1 minute at 185 °C and 17 bar pressure in a microbatch reactor using HCl in a biphasic system of
water/MIBK.[65] A study published in 2011 reported an 82% HMF yield in isopropanol in 1 hour at 120 °C,[66] while
a near quantitative yield was obtained in [bmim]Cl at 80 °C.[34] Sulfuric acid proved to be an efficient catalyst in
[bmim]Cl and DMSO, reaching over 93% yield.[34,67]
The results from the screening of different acids are shown in Figure 1. Hydrochloric acid proved to be the
best acid for this purpose, reaching 74% yield in 10 minutes at 14% fructose loading. The greater acid strength can
ensure complete dehydration of fructose intermediates while chloride ions can contribute to the catalysis by
providing hydrogen bonding to the hydroxyl group. It was thus decided to continue the study with HCl as the catalyst.
6
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
Figure 1. Acid selection for fructose dehydration in [bmim][OTf]. Fructose loading = 14%, T = 90 °C, initial water
content = 3.5%, acid catalyst loading 10%, t = 10 min.
Effect of the water content on the yield and kinetics of the reaction
In order to understand the role of water in the reaction medium, a range of water content in the ionic liquid was
investigated. The results in Figure 4 clearly suggest that water has a significant effect on the yield of HMF in the
reaction.
Figure 4. Effect of water content on HMF yield from fructose in [bmim][OTf], fructose loading = 10%, T = 90 °C, t =
10 min, 10 % acid catalyst loading (HCl).
Gradual addition of water led to a maximum in HMF yield (79%) at 3.5% water content, while using pure
water as solvent resulted in no reaction at all, suggesting that the ionic liquid increases the activity of H+ or enhances
the hydration of the chloride anion, hindering its role in dehydration. Low water concentration led to a poor yield,
which may be due to the overly aggressive activity of the H+ ions, leading to side reactions of HMF, such as self-
condensation or reaction with fructose (to produce humins). High yields are obtained at 3.5% water content,
suggesting a good compromise between stabilisation and dehydration of HMF. In this medium, water plays a
7
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
different role compared to the dehydration in halogenated ionic liquids as [bmim]Cl, [bmim]Br and [Hmim]Cl,[31,32,34]
where the only effect is to decrease the reaction rate. In other catalytic systems, water contamination leads to a
drastic decrease in selectivity with overhydration to levulinic acid and formic acid.[68,69] In our case, the water content
has a beneficial effect on the yield while an excessive amount leads to lower efficiency due to the deactivation of
the protons acting as catalyst in the medium. This is further suggested by the relative amounts of water and HMF,
which in the case of 3.5% water content is equivalent to 2.5 moles of water per mole of HMF.
In order to further understand the role of water content on the stability of HMF and the rate of reaction,
kinetic analysis was performed at two different water contents at higher temperature (100 °C).
Figure 5. Effect of water content on the kinetics of HMF production from fructose in [bmim][OTf] at 100 °C, fructose
loading = 10 %, catalyst loading = 10% (HCl).
The data in Figure 5 suggest that the water content contributes to both the stability of HMF and the kinetics
of reaction. At 2.5% water content, the reaction was complete after 3 minutes, reaching a maximum of 76% followed
by a gradual decrease in HMF yield of 6% by 10 minutes. In contrast, at 3.5% water content, the reaction achieves
a steady state at 82% HMF yield. This trend suggests that water contributes to an increase in the yield of reaction
by stabilising HMF towards self-condensation side reactions. Compared to the previous results (Figure), the yield
at 3.5% water content was improved, suggesting that the yield was limited by some intermediates that did not
undergo complete dehydration. In contrast, at lower water content, the instability of HMF in the system decreases
the yield by the 10 minute time point, reaching a maximum of 75%. From these observations, it can be seen that
the contribution of water is distinct from studies in which it exerts a favourable catalytic effect as proton bridge for
the dehydration of glucose in DMSO/water mixtures,[70,71] or as a contaminant that leads to side reactions.[72]
A
consistent reduction in yield can be observed due to the side reaction between fructose and HMF (or the
intermediates formed) and this mainly occurs in the initial stages, when the fructose and the intermediates are
present in high concentration. This is supported by the following investigations on fructose loading.
Effect of fructose loading
The next step in the study was to analyse the effect of fructose loading on the HMF yield (Figure 6). This revealed
that the yield of HMF seemed to be affected significantly by the substrate loading. At 90 °C, the yield decreases
from 86% under very dilute conditions (2% fructose loading) to 72% at 14% loading. The decrease in yield can be
traced to secondary reactions that arise from the instability of HMF and the reaction of HMF with fructose. To
analyse this phenomenon, a kinetic study was conducted at 14% and 10% loading to determine whether an
optimum set of conditions would emerge (Figure 6). At 90 °C, it was observed that even at 14% loading and 3.5%
water content, HMF does not undergo any significant decomposition, reaching a plateau (the yield decreases by
only 2% from 8 to 10 minutes). At 10% fructose loading, the HMF yield increases to 82% in 10 minutes, showing a
8
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
trend of increasing yield. The decrease in rate can be attributed to the increase in the amount of water in the
reaction mixture, which lowers the activity of the protons present. The fact that the yield of HMF does not increase
beyond 8 minutes at 14% fructose loading, in contrast to the slight yield increase with 10% loading, suggests that
the kinetics of the side reactions are more rapid in the former case. Moreover, the plateau observed suggests that
HMF is stable under these operating conditions, indicating that the major loss in yield is actually due to the reaction
of HMF with fructose or the intermediates.
Figure 6. a) The HMF yield from fructose in [bmim][OTf], at different fructose loadings and b) kinetic analysis of the
same reaction at two fructose loadings (right). Reaction conditions: T = 90 °C, 3.5% water content, t = 10 min, 10%
HCl loading.
Effect of temperature
This can lead to the rate of reaction being increased remarkably but it can also favour side reactions. The
optimisation for temperature was performed at 10% fructose loading, as shown in Figure 7.
Figure 7. Effect of temperature on HMF yield from fructose in [bmim][OTf], 10% fructose loading, 10% HCl loading,
3.5% water content.
9
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
At 100 °C, the HMF yield reaches a steady state after 6 min at 81% with slight improvements compared to
the result at 90 °C (79%), indicating that, at the lower temperature, the reaction is still limited by the dehydration of
the intermediates. At 80 °C, it was observed that the reaction is limited not only by the conversion of intermediates
but also by fructose dehydration. Higher temperatures were not analysed so as to avoid loss of water from the
reaction mixture. According to these results, 100 °C is the optimal temperature for the reaction.
In order to improve the economics of the process, a kinetic analysis was carried out at 14% fructose loading
at 100 °C (Figure 8). By increasing the temperature, it was possible to improve the yield from 72% to 81%, making
the dehydration at high concentration a more feasible process from a techno-economic viewpoint.
Figure 8. Kinetic analysis of the dehydration of fructose to HMF in [bmim][OTf] at two different temperatures and
14% fructose loading, 10% HCl loading, 3.5% water content.
Effect of the cation
According to a previous study,[64] the cation plays an important role in the dehydration of fructose at low temperature
through the coordination between the protons of the cations and the hydroxyl groups of fructose. This effect was
investigated by changing the cation of the ionic liquid solvent. The pyrrolidinium ionic liquid, [bmpyr][OTf], has a
lower hydrogen donating ability compared to the imidazolium cation, according to their Kamlet-Taft parameters.[73]
However, the results reported in Table 1 show no significant effect of the cation on HMF yield, suggesting that no
interactions arise with the fructose molecule at higher temperatures.
Table 1. Effect of the cation on HMF yield from fructose in different ILs. Reaction conditions: 14% fructose loading,
T = 100 °C, t = 10 min, 10% HCl loading, 3.5% water content.
Ionic Liquid
[bmim][OTf]
[emim][OTf]
[bmpyr][OTf]
HMF yield (%)
80.8 ± 2.0
77.4 ± 0.7
80.1 ± 0.5
Dehydration from glucose to HMF
Glucose represents a valuable feedstock for HMF synthesis since it is considerably cheaper than fructose. The
dehydration of this sugar in [bmim][OTf] with HCl proved to be completely unselective, mainly due to the very low
10
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
solubility of glucose in [bmim][OTf] and the incompatibility of the Brønsted acid catalyst with the substrate. In order
to improve the yield, CrCl3·6H2O was used as the catalyst and the solvent composition was modified with [bmim]Cl,
thus increasing the solubility of glucose. Figure 9 shows the yield at different [bmim]Cl concentrations after 30 min
at 10% glucose loading and 10 mol% catalyst loading. It can be seen that a minimum proportion of 10% [bmim]Cl
is required to maintain the yield of HMF above 50%. In order to improve the yield, the solvent composition was
varied while maintaining a 3.5% water content, but an improvement in yield of only 5% was achieved. While water
can improve the dehydration of fructose in non-coordinating ionic liquids, the same is not true for glucose.
Figure 9. Glucose dehydration to HMF at 120 °C in [bmim][OTf]/[bmim]Cl mixtures using CrCl3·6H2O (10 mol%
loading with 10% glucose loading.
Extraction of HMF from the ionic liquid
While there are many reports of catalysis in non-coordinating ionic liquids, the ability to isolate HMF from media
such as [bmim][OTf] is also an important factor in the use of these systems. This is facilitated by the lack of hydrogen
bonding between the hydroxyl group and the solvent, favouring the partitioning of HMF towards the extracting
solvent. In order to analyse the effect of a coordinating environment on the partition coefficient, the extraction of
HMF by diethyl ether was performed from mixtures with varying proportions of [bmim]Cl. Figure 10 shows a
remarkably negative impact of chloride anions on the partition coefficient, even at low concentrations. This
illustrates the impracticality of extracting HMF from chloride-based ionic liquid media. In contrast, in the absence of
chloride anions, the partition coefficient increases by an order of magnitude, facilitating the extraction into the
organic phase.
11
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
Figure 10. Partition coefficient of HMF in diethyl ether and ionic liquid mixtures with different chloride concentrations.
By exploiting this feature, the isolation of HMF from this solvent was performed on a larger scale, using 1
g of ionic liquid and 14% fructose loading, giving 71% isolated yield of HMF (Figures S18-20 in the Supporting
Information). The ionic liquid was extracted from the water phase using ethyl acetate as a light brown viscous liquid.
Analysis by 1H NMR spectroscopy and mass spectrometry showed no degradation of the ionic liquid (Figure S21-
23 in the Supporting Information).
CONCLUSIONS
As far as we are aware, the results presented here represent the highest yields yet reported for the rapid
dehydration of fructose in non-coordinating ionic liquids. It was demonstrated that high HMF yields (>80%) could
be obtained at high fructose loading (14%) in a short reaction time (<10 min) making this reaction system very
competitive. This is despite the widespread belief that non-coordinating ionic liquids are unsuitable reaction media
for this transformation compared to halide-based ionic liquids (e.g. [bmim]Cl). We have shown that the water content
is a key parameter for efficient dehydration, providing stability to the HMF product. The analysis presented here
shows that the dehydration of fructose to HMF in non-coordinating ionic liquids is compromised by the instability of
HMF in the (pure) ionic liquid under acidic conditions. However, this behaviour can be suppressed by regulating
the amount of water present, stabilizing HMF towards self-condensation. By harnessing this approach we were
able to achieve efficient dehydration without significant loss in yield due to the reaction of fructose with HMF in the
initial stages. The overall reaction appears to be partially limited by the intermediates generated and their successful
conversion to the desired HMF product. By increasing the temperature, it was possible to improve the yield of
reaction by dehydrating all the intermediates present. Further research is underway in order to optimise the water
content in the system and increase the fructose loading still further.
ACKNOWLEDGEMENTS
Imperial College London is thanked for the award of a President’s PhD Scholarship (to A.A.G.).
CONFLICT OF INTEREST
The authors declare no conflict of interest.
12
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
REFERENCES
[1]
[2]
G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–4098.
E4tech, RE-CORD, WUR, From the Sugar Platform to Biofuels and Biochemicals, 2015,
https://ec.europa.eu/energy/sites/ener/files/documents/EC%20Sugar%20Platform%20final%20report.pdf
[3]
A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Cairney, C. A. Eckert, W. J. Frederick Jr.,
J. P. Hallett, D. J. Leak, C. L. Liotta, J. R. Mielenz, R. Murphy, R. Templer, T. Tschaplinski, Science 2006,
311, 484–489.
[4]
[5]
R.-J. 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.
H. Wang, C. Zhu, D. Li, Q. Liu, J. Tan, C. Wang, C. Cai, L. Ma, Renew. Sustain. Energy Rev. 2019, 103,
227–247.
[6]
[7]
[8]
T. Ståhlberg, W. Fu, J. M. Woodley, A. Riisager, ChemSusChem 2011, 4, 451–458.
T. Wang, M. W. Nolte, B. H. Shanks, Green Chem. 2014, 16, 548–572.
S. Siankevich, Z. Fei, R. Scopelliti, G. Laurenczy, S. Katsyuba, N. Yan, P. J. Dyson, ChemSusChem. 2014,
7, 1647-1654.
[9]
S. Siankevich, Z. Fei, R. Scopelliti, P. G. Jessop, J. Zhang, N. Yan, P. J. Dyson, ChemSusChem 2016, 9,
2089-2096.
[10]
[11]
[12]
[13]
[14]
C. J. Clarke, W. C. Tu, O. Levers, A. Bröhl, J. P. Hallett, Chem. Rev. 2018, 118, 747–800.
T. Istasse, L. Bockstal, A. Richel, ChemPlusChem 2018, 83, 1135–1143.
C. Antonetti, S. Fulignati, D. Licursi, A. Galletti, ACS Sustain. Chem. Eng. 2019, 7, 6830–6838.
X. Haung, S. Kudo, J. Sperry, J. Hayashi, ACS Sustain. Chem. Eng. 2018, 7, 5892–5899.
M. Du, A. Agrawal, S. Chakraborty, S. J. Garibay, R. Limvorapitux, B. Choi, S. T. Madrahimov, S. T.
Nguyen, ACS Sustain. Chem. Eng. 2019, 7, 8126–8135.
[15]
[16]
T. Yamada, K. Kamata, E. Hayashi, M. Hara, S. Uchida, ChemCatChem 2019, 11, 1-6
Q. Hou, M. Zhen, W. Li, L. Liu, J. Liu, S. Zhang, Y. Nie, C. Bai, X. Bai, M. Ju, Appl. Catal. B Environ. 2019,
253, 1–10.
[17]
[18]
[19]
J. J. Wiesfeld, R. Gaquere, E. J. M. Hensen, ACS Sustain. Chem. Eng. 2019, 7, 7552–7562.
T. Zhang, W. Li, H. Xin, L. Jin, Q. Liu, Catal. Commun. 2019, 124, 56–61.
A. Rusanen, R. Lahti, K. Lappalainen, J. Kärkkäinen, T. Hu, H. Romar, U. Lassi, Catal. Today 2019,
doi.org/10.1016/j.cattod.2019.02.040
[20]
[21]
[22]
[23]
[24]
[25]
F. D. Bobbink, Z. Huang, F. Menoud, P. J. Dyson, ChemSusChem, 2019, 12, 1437–1442
K. B. Sidhpuria, A. L. Daniel-Da-Silva, T. Trindade, J. A. P. Coutinho, Green Chem. 2011, 13, 340–349.
V. Degirmenci, E. A. Pidko, P. C. M. M. Magusin, E. J. M. Hensen, ChemCatChem 2011, 3, 969–972.
M. E. Zakrzewska, E. Bogel-łukasik, R. Bogel-łukasik, Chem. Rev. 2011, 111, 397–417.
S. An, Z. Wang, H. Zhang, H. N. Mira, Y.-F. Song, ChemCatChem 2019, 11, 1–12.
R. Pramod V, M. B. Rajendra, C. Wook-Jin, J. R. Amol, K. Hern, Catal. Letters 2019, 149, 672–687.
13
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
[26]
[27]
R. P. Swatloski, S. K. Spear, J. D. Holbrey, R. D. Rogers, J. Am. Chem. Soc. 2002, 124, 4974–4975.
G. Yong, Y. Zhang, J. Y. Ying, Angew. Chem. 2008, 120, 9485–9488; Angew. Chem. Int. Ed. 2008, 47,
9345–9348.
[28]
[29]
Z. C. Zhao, H., Holladay, J.E., Brown, H., Zhang, Science 2007, 316, 1597–600.
Q. Hou, W. Li, M. Zhen, L. Liu, Y. Chen, Q. Yang, F. Huang, S. Zhang, M. Ju, RSC Adv. 2017, 7, 47288–
47296.
[30]
[31]
S. Eminov, J. D. E. T. Wilton-Ely, J. P. Hallett, ACS Sustain. Chem. Eng. 2014, 2, 978–981.
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.
[32]
[33]
[34]
[35]
[36]
C. Moreau, A. Finiels, L. Vanoye, J. Mol. Catal. A Chem. 2006, 253, 165–169.
C. Li, Z. K. Zhao, H. Cai, A. Wang, T. Zhang, Biomass Bioenergy 2011, 35, 2013–2017.
C. Li, Z. K. Zhao, A. Wang, M. Zheng, T. Zhang, Carbohydr. Res. 2010, 345, 1846–1850.
C. Lansalot-Matras, C. Moreau, Catal. Commun. 2003, 4, 517–520.
C. Tian, X. Zhu, S. H. Chai, Z. Wu, A. Binder, S. Brown, L. Li, H. Luo, Y. Guo, S. Dai, ChemSusChem
2014, 7, 1703–1709.
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
C. Shi, Y. Zhao, J. Xin, J. Wang, X. Lu, X. Zhang, S. Zhang, Chem. Commun. 2012, 48, 4103–4105.
H. Wang, S. Liu, Y. Zhao, H. Zhang, J. Wang, ACS Sustain. Chem. Eng. 2016, 4, 6712-6721.
L. Hu, Y. Sun, L. Lin, Ind. Eng. Chem. Res. 2012, 51, 1099–1104.
G. Tian, X. Tong, Y. Cheng, S. Xue, Carbohydr. Res. 2013, 370, 33–37.
Z. Wei, Y. Liu, D. Thushara, Q. Ren, Green Chem. 2012, 14, 1220–1226.
X. Sun, Z. Liu, Z. Xue, Y. Zhang, T. Mu, Green Chem. 2015, 17, 2719–2722.
C. Shi, J. Xin, X. Liu, X. Lu, S. Zhang, ACS Sustain. Chem. Eng. 2016, 4, 557–563.
D. Yan, J. Xin, Q. Zhao, K. Gao, X. Lu, G. Wang, S. Zhang, Catal. Sci. Technol. 2018, 8, 164–175.
D. Yan, J. Xin, C. Shi, X. Lu, L. Ni, G. Wang, S. Zhang, Chem. Eng. J. 2017, 323, 473-482.
M. Sajid, X. Zhao, D. Liu, Green Chem. 2018, 20, 5427–5453.
T. Stahlberg, E. Eyjolfsdottir, Y. Y. Gorbanev, I. Sdaba, A. Riisager, Catal. Letters 2012, 142, 1089–1097.
V. Farmer, T. Welton, Green Chem. 2002, 4, 97–102.
I. A. Ansari, R. Gree, Org. Lett. 2002, 4, 1507–1509.
A. Fall, M. Sene, M. Gaye, G. Gómez, Y. Fall, Tetrahedron Lett. 2010, 51, 4501–4504.
Z. G. Wang, Y. Jin, X. H. Cao, M. Lu, New J. Chem. 2014, 38, 4149–4154.
L. Lin, J. Liuyan, W. Yunyang, Catal. Commun. 2008, 9, 1379–1382.
T. P. Wells, J. P. Hallett, C. K. Williams, T. Welton, J. Org. Chem. 2008, 73, 5585–5588.
I. Correia, T. Welton, J. Chem. Soc. Dalt. Trans. 2009, 35, 4115–4121.
A. Tsurumaki, H. Ohno, Chem. Commun. 2018, 54, 2–9.
[56]
K. I. Galkin, V. P. Ananikov, ChemSusChem 2019, 12,1-8
14
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
[57]
J. Jeong, C. A. Antonyraj, S. Shin, S. Kim, B. Kim, K. Y. Lee, J. K. Cho, J. Ind. Eng. Chem. 2013, 19, 1106–
1111.
[58]
[59]
G. Sampath, S. Kannan, Catal. Commun. 2013, 37, 41–44.
G. Morales, J. A. Melero, M. Paniagua, J. Iglesias, B. Hernández, M. Sanz, Chin. J. Catal. 2014, 35, 644–
655.
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
A. Takagaki, M. Ohara, S. Nishimura, K. Ebitani, Chem. Commun. 2009, 6276–6278.
M. Ohara, A. Takagaki, S. Nishimura, K. Ebitani, Appl. Catal. A Gen. 2010, 383, 149–155.
F. Ilgen, D. Ott, D. Kralisch, C. Reil, A. Palmberger, B. König, Green Chem. 2009, 11, 1948–1954.
A. A. Assanosi, M. M. Farah, J. Wood, B. Al-Duri, RSC Adv. 2014, 4, 39359–39364.
J. Zhang, X. Yu, F. Zou, Y. Zhong, N. Du, X. Huang, ACS Sustain. Chem. Eng. 2015, 3, 3338–3345.
T. Tuercke, S. Panic, S. Loebbecke, Chem. Eng. Technol. 2009, 32, 1815–1822.
L. Lai, Y. Zhang, ChemSusChem 2011, 4, 1745–1748.
C. Sievers, I. Musin, T. Marzialetti, M. B. V. Olarte, P. K. Agrawal, C. W. Jones, ChemSusChem 2009, 2,
665–671.
[68]
[69]
[70]
[71]
[72]
[73]
F. Tao, H. Song, L. Chou, RSC Adv. 2011, 1, 672–676.
X. Qi, M. Watanabe, T. M. Aida, R. L. Smith, Green Chem. 2008, 10, 799–805.
F. Zhou, X. Sun, D. Wu, Y. Zhang, H. Su, ChemCatChem 2017, 9, 2784–2789.
X. Cao, S. P. Teong, D. Wu, G. Yi, H. Su, Y. Zhang, Green Chem. 2015, 17, 2348–2352.
K. ichi Shimizu, R. Uozumi, A. Satsuma, Catal. Commun. 2009, 10, 1849–1853.
B. Kirchner, Ionic Liquids, Springer, 2009.
15
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901529
ChemSusChem
For ToC use:
Rapid, High-Yield Fructose Dehydration to HMF in Mixtures of Water and the Non-
Coordinating Ionic Liquid [bmim][OTf]
Amir Al Ghatta, James D. E. T. Wilton Ely* and Jason P. Hallett*
In contrast to literature reports, the non-coordinating ionic liquid [bmim][OTf] is an effective and versatile solvent for
the high-yield dehydration of fructose to the platform chemical HMF over short reaction times. The study reveals
that water content is an essential parameter and that judicious addition of water can increase the yield dramatically.
16
This article is protected by copyright. All rights reserved.