View Article Online / Journal Homepage / Table of Contents for this issue

Green Chemistry Dynamic Article Links

Cite this: Green Chem., 2011, 13, 2859

www.rsc.org/greenchem PAPER
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

Microwave assisted conversion of carbohydrates and biopolymers to

5-hydroxymethylfurfural with aluminium chloride catalyst in water†
Sudipta De, Saikat Dutta and Basudeb Saha*
Received 13th May 2011, Accepted 20th July 2011
DOI: 10.1039/c1gc15550d

The common Lewis acid AlCl3 has efficiently produced 5-hydroxymethylfurfural (HMF) from
carbohydrate and biopolymer substrates in water, DMSO, and water–methylisobutylketone
biphasic solvents under microwave irradiation. The yield of HMF in different solvents follows an
increasing order from water to water–MIBK biphasic solvent to DMSO. The yield of HMF
increased with an increase in catalyst loading whereas it remains unchanged upon increase of the
carbohydrate concentration. In most reactions, the maximum yield of HMF is recorded within
2 min of reaction time. The mechanism of the AlCl3 catalyzed glucose dehydration reaction is
proposed to proceed through the isomerization of glucopyranose to fructofuranose, followed by a
proton assisted transformation of fructofuranose to HMF.

1. Introduction ([EMIM]BF4 ) at 100 ◦ C.8 Tungsten salts (WCl4 , WCl6 ) acted

as Lewis acid catalysts for glucose conversion in 1-butyl-3-
Irreversible consumption of carbon sources by humankind imidazolium chloride [BMIM]Cl solvent and produced 72%
resulted in diminishing reserves of fossil fuels and global HMF at mild reaction conditions (50 ◦ C).9 In more recent
warming by CO2 emission.1 This issue has prompted a shift studies, GeCl4 in [EMIM]BF4 catalytic system produced 92.1%
of economy to replace fossil-based resources with renewable HMF from fructose at 100 ◦ C.10 The catalytic conversion of
and sustainable ones.2 In this context 5-hydroxymethylfurfural fructose to HMF with boric acid (B(OH)3 ) has been reported
(HMF) has received significant attention as a platform chemical to produce 46% HMF in biphasic water–methylisobutylketone
for synthesizing a broad range of chemicals and liquid trans- (MIBK) solvent at 150 ◦ C.11 The lanthanide salt-based catalytic
portation fuels.3 Carbohydrates constitute 75% of the world’s systems have been reported to produce 24% HMF yield at
renewable biomass and cellulose.4 The dehydration of fructose 140 ◦ C in 6 h from glucose.12 In limited cases, the synthesis of
and glucose serve as model reactions for the synthesis of HMF using mixed Brønsted and Lewis acid catalysts13 and the
HMF from biomass-derived carbohydrates since the conversion microwave-assisted14 reactions from cellulose and sugar deriva-
to HMF always involves these sugars in the final reaction tives were reported to be successful. Among the reports on the
sequence.5 Recently, fructose and glucose have been converted microwave (MW) assisted green synthesis of HMF, concentrated
to HMF in high yields with Cr(II) or Cr(III) halide catalysts HCl catalyzed dehydration of fructose produced 63% HMF in
in imidazolium ionic liquids.6 In recent years, research efforts water. However, the selectivity of HMF decreased when the
for developing sustainable methods of HMF production from reaction was carried out for more than 1 min.14a Microwave
various carbohydrates with Lewis acid catalysts have been assisted glucose dehydration with butyl-3-imidazolium chloride
limited to the use of ionic liquid solvents. Among these efforts, [BMIMCl]/CrCl3 (10 mol%) catalyst produced 71% HMF in
HMF synthesis from carbohydrates, including glucose, was 0.5 min. The yield of HMF decreased to 67% upon continuing
accomplished by various catalytic systems. A catalytic system the reaction for 1 h.14d However, potential drawbacks of these
consisting of Cr(II) or Cr(III)-chlorides and bulky N-heterocyclic methods using ionic liquids is that ionic liquids are expensive
carbene ligands produced high yields of HMF from fructose and hence probably not suitable candidates for developing an
and glucose.7 A SnCl4 catalyzed dehydration reaction of glucose economically favourable and scalable HMF production method.
produced 60% HMF in 1-ethyl-3-imidazolium tetrafluroborate Therefore, it is beneficial to replace expensive ionic liquid sol-
vents with aqueous and/or mixed aqueous biphasic and highly
Laboratory of Catalysis, Department of Chemistry, North Campus, polar aprotic organic solvents for the sustainable synthesis of
University of Delhi, Delhi, India. E-mail: bsaha@chemistry.du.ac.in; HMF from carbohydrates. Regardless of recent developments
Fax: 91 2766 7794; Tel: 011-2766 6646 of metal salt catalyzed synthetic methods for HMF production,
† Electronic supplementary information (ESI) available: Results of it remains highly challenging to develop a sustainable route
HMF yield as a function of fructose concentration, catalyst dosage, reac-
tion time, various aluminium catalysts and UV-Visible and NMR spectra for HMF production from carbohydrates and biopolymers
of isolated HMF products are deposited. See DOI: 10.1039/c1gc15550d in environmentally benign solvents. In a continuation of our

This journal is © The Royal Society of Chemistry 2011 Green Chem., 2011, 13, 2859–2868 | 2859
 View Article Online
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

Scheme 1 The microwave assisted transformation of fructose or starch to HMF with AlCl3 catalyst.

ongoing research program in developing Lewis acid and metal Upon completion of the reaction for the set reaction time, the
salt catalyzed dehydration of biomass derived sugars and reactor was opened. The temperature of the reaction mass was
cellulosic biomass, we have studied the microwave (MW) assisted cooled down to room temperature before analyzing the product
technique of HMF synthesis from various sugar derivatives with using NMR spectroscopic and UV-Visible spectrophotometric
a strong Lewis acid catalyst. AlCl3 is known as an effective Lewis techniques.
acid catalyst for HMF synthesis in ionic liquid ([BMIMCl]).9 In the case of the DMSO solvent mediated reaction, HMF
However, the catalytic effectiveness of AlCl3 for HMF synthesis was isolated from the reaction mixture by distilling out DMSO
has not been reported in aqueous or mixed aqueous solvents at reduced pressure and then extracting the residue as the organic
under microwave irradiation. The present paper describes the layer with diethyl ether after adding water into the DMSO
transformation of carbohydrates (fructose, glucose, sucrose) separated residue. The organic layer was collected in a pre-
and biopolymers (inulin, starch) into HMF in environmentally weighed empty vial and dried under vacuum. After drying off
benign solvents with anhydrous and hydrated AlCl3 , and AlBr3 the solvent, the weight of the vial was recorded. The yield of
Lewis acid catalysts (Scheme 1). isolated HMF was determined by subtracting the weight of the
HMF-containing vial from that of empty vial.
2. Experimental Dehydration of carbohydrates using oil-bath heating. The
2.1 Materials and experimental methods synthesis of HMF from carbohydrates was also carried out in
a round bottom flask by oil-bath heating. For oil bath heating,
Substrates (fructose, glucose, sucrose, starch and inulin) and a glass round bottom flask was charged with substrate, solvent
solvents (DMSO, MIBK) were purchased from Sigma Aldrich and catalyst, and then refluxed in an oil bath. After completion
and used without further purification. The starch sample, was of reaction, the reaction mixture was cooled down to room
dried under vacuum at 100 ◦ C for 24 h before use. Anhydrous temperature and analyzed by UV-Visible spectrophotometric
and hydrated AlCl3 , AlBr3 , and Yb(OTf)3 were purchased from and 1 H NMR spectroscopic techniques.
Sigma Aldrich and used without further purification. SnCl4
was purchased from Spectrochem, India and used without Catalyst life-time study
further purification. Unless otherwise mentioned, distilled water
was used as the aqueous phase and anhydrous AlCl3 was The life-time of the catalyst was studied by recycling the reaction
used as catalyst for all reactions. The catalytic conversion of mixture containing the spent catalyst. Prior to recycling for the
carbohydrates to HMF was performed in a CEM Matthews next run, the HMF component was extracted from the reaction
WC Discover Microwave reactor, (model: Discover System, mixture with diethyl ether and fresh substrate was added into
no. 908010 DV9068) at the standard operating frequency of the reaction mixture. Fresh AlCl3 catalyst was not added to
a microwave synthesis reactor (2.45 GHz, power 300 Watt). 1 H compensate any loss of the catalyst in the prior runs.
NMR spectra were recorded on a JEOL JNM ECX-400 P 400
MHz instrument and NMR data were processed with JEOL Determination of HMF yield
DELTA program version 4.3.6. Progress of the reaction was
The yield of HMF was determined by both 1 H NMR and UV-
monitored by 1 H NMR spectroscopy in limited cases. HMF
Visible spectrophotometric techniques. For 1 H NMR spectro-
yields were measured by both UV-visible spectrophotometric
scopic analysis, HMF was extracted from the reaction mixture
(UV-SPECORD 250 analytikjena) and NMR spectroscopic
with diethyl ether. Pale yellow oily HMF was obtained after
techniques.
removing the solvent in vacuum at room temperature.
HMF production 1
H NMR method. For quantifying the yield of HMF using
1
H NMR spectroscopic technique, a known concentration of
Dehydration of carbohydrate using microwave irradiation
mesitylene (internal standard) was added into the HMF product
The dehydration reactions of carbohydrates, ca. D-fructose, solution in DMSO-d 6 . The percentage of HMF yield was
glucose, sucrose and biopolymer substrates, were carried out calculated by using the integrated values of the aldehyde proton
by charging substrates, solvent and catalyst in a microwave (d = 9.58 ppm) of HMF and three aromatic ring protons
tube under the reaction conditions mentioned in Tables 1 and of mesitylene (d = 6.79 ppm) (Fig S1†). First, a standard
2. The microwave tube was then inserted into the microwave HMF solution of 99% purity was analyzed for correlating
reactor pre-set at the desired temperature and reaction time. the percentage of actual and calculated amount of HMF.

2860 | Green Chem., 2011, 13, 2859–2868 This journal is © The Royal Society of Chemistry 2011
 View Article Online

Table 1 The AlCl3 catalyzed dehydration of fructose to HMF under microwave irradiation

HMF Yield (%)

◦ 1
Entry Fructose (wt%) AlCl3 (mol%) Solvent T ( C) t (min) H NMR UV-Vis Isolated
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

1 5 50 Water/MIBK 130 5 60.1 61.2 —

2a 5 50 Water/MIBK 130 5 61.6 63.3 61.0
3 5 50 Water 120 5 51.4 53.9 50.2
4 5 40 Water 120 5 50.6 51.5 —
5 5 25 Water 120 5 — 40.9 —
6 5 50 Water 120 15 — 55.1 —
7 5 50 Water 120 10 — 54.6 —
8 5 50 Water 120 2 — 52.3 —
9 5 50 Water 120 20 53.3 55.7 52.7
10 5 50 DMSO 100 5 — 42.9 —
11 5 50 DMSO 140 5 70.1 71.3 69.4
12 5 — DMSO 100 5 — 3.5 —
13 5 — DMSO 140 5 21.4 22.0 —
14 5 50 Water 120 5 54.8 55.3 —
15 15 50 Water 120 5 51.0 52.6 50.1
16 30 50 Water 120 5 51.4 52.1 50.8
17a 5 50 Water/THF 100 5 57.9 59.0 —
18 5 50 Water 120 0.5 — 47.8 —
19 5 40 Water 120 2 — 55.1 —
20 5 40 Water 120 15 45.7 47.0 45.5
21 5 25 Water 120 2 — 38.0 —
22 5 25 Water 120 10 — 42.3 —
23 5 25 Water 120 15 42.8 43.4 —
24 10 50 Water 120 5 51.2 53.5 51.9
25 5 — Water 120 5 — 0.8 —
a
NaCl (50 mg). For biphasic solvents aqueous : organic = 1 : 2 volume ratio (mL) was used.

Table 2 Conversion of fructose to HMF catalyzed by AlCl3 under oil-bath heating

HMF Yield (%)

1
Entry Fructose (wt%) AlCl3 (mol%) Solvent H NMR UV-Vis Isolated

1 5 50 Water/MIBK (2/8) 32.2 33.4 —

2a 5 50 Water/MIBK (2/8) 33.5 36.1 33.4
3 5 50 Water 13.0 14.1 —
4 5 50 DMSO — 30.7 —
a
NaCl (100 mg), T = 120 ◦ C, t = 60 min.

Once a good correction was established, the extracted HMF conditions, such as the nature of solvents, catalysts, catalyst
product samples were run and the percentage of HMF yield was concentrations, nature of substrates, substrate concentrations,
calculated. reaction time etc. The purpose for designing these experiments
are: (i) identification of an effective substrate for the synthesis of
UV-Visible spectrophotometric method. The UV-Visible
HMF in an aqueous medium and (ii) optimization of reaction
spectrum of pure HMF solution (Fig. S2†) has a distinct
conditions for maximizing HMF yield. The effect of several
peak at 284 nm with a corresponding extinction coefficient
reaction parameters on HMF yield and catalysts effectiveness
(e) value of 1.66 ¥ 104 M-1 cm-1 . The percentage of HMF in
are described below.
each of the reaction products, as tabulated in Table 1, was
calculated from the measured absorbance values at 284 nm and
the extinction coefficient value. Repeated measurement of the 3.1 Effect of solvents on HMF yield
same solution showed the percentage of error associated with
this measurement was ±3%. The yield of HMF obtained from The dehydration reactions of carbohydrates (fructose, glucose,
two different methods (1 H NMR and UV-Visible) for the same sucrose) with the AlCl3 catalyst were carried out in water, DMSO
reaction product agreed very well and the result was within ±5% and water–MIBK biphasic solvents under microwave irradiation
error (Table 1). to investigate the effect of different solvents on HMF yield.
Details of reaction conditions and the corresponding HMF
yields are summarized in Tables 1 and 2. Under identical reaction
3. Results and discussion
conditions, water, water–MIBK and DMSO solvent mediated
Several experiments were designed for studying the catalytic de- fructose dehydration reactions produced 53.9%, 61.2% and
hydration of carbohydrates to HMF under various experimental 71.3% HMF (entries 1, 3, 11 in Table 1), respectively. The higher

This journal is © The Royal Society of Chemistry 2011 Green Chem., 2011, 13, 2859–2868 | 2861
 View Article Online

Table 3 Results of glucose dehydration catalyzed by AlCl3 under MW irradiation

Entry AlCl3 (mol%) Solvent T (◦ C) t (min) HMF yield (%)

1 50 Water 120 5 37.3

2a 50 Water/MIBK 130 5 43.7
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

3 50 Water/MIBK 130 5 43.0

4 50 DMSO 140 5 6.4
5 50 DMSO 140 5 52.4
6 50 Water 120 2 35.1
7 50 Water 120 10 39.0
8 50 Water 120 15 39.8
9 50 Water 120 20 40.3
10 33 Water 120 5 30.2
11 33 Water 120 15 33.6
12 20 Water 120 5 25.4
a
NaCl (50 mg), Glucose = 5 wt%.

HMF yield in DMSO solvent in comparison to that in aqueous Table 4 Results of 30 wt% fructose dehydration with AlCl3 catalyst at
medium for all carbohydrate substrates can be attributed to the variable catalyst loading
(i) higher microwave absorbing ability of DMSO, as measured a Fructose (wt%) AlCl3 (mol%) HMF yield (%)
by the higher tangent value of DMSO (tan d 0.825) than water
(tan d 0.123),15 (ii) avoidance of the side reactions in DMSO that 1 30 50 52.1
occur in aqueous medium16 and (iii) catalytic ability of DMSO 2 30 40 48.6
3 30 33 46.0
in converting fructofuranose to HMF via a 5-membered cyclic 4 30 25 38.8
mechanism at a higher temperature (150 ◦ C) without any added 5 30 20 35.8
catalyst.17 These positive characteristics of DMSO accounted
Solvent = water (2 mL), T = 120 ◦ C, t = 5 min, MW.
for its use as a solvent so that a comparison of HMF yields in
different solvents can be drawn. The lower HMF yield in water
is due to the rehydration of HMF and the formation of levulinic
and formic acids as side-products. 3.2 Effect of the starting fructose concentration on HMF yield
On the other hand, the higher HMF yield in water–MIBK The literature report suggests that HMF yield and selectivity
biphasic solvent than in water is due to the driving force of depend on the starting fructose concentrations. Humin is one of
the biphasic mixture in which HMF accumulates in the organic the side reaction products of the fructose dehydration reaction.
phase after its formation in the aqueous phase. This method of The percentage of humin formation is reported to increase with
HMF synthesis in a biphasic solvent also produced pure HMF, an increase in starting fructose concentrations, and thereby
as evidenced from the clean 1 H NMR spectrum of the HMF lower the desired HMF yield. The formation of humin can be
product collected from the MIBK layer (Fig. S3†). as high as 35% for 18 wt% fructose solution and 20% for 4.5
It is reported that the salting-out effect18 has a positive wt% fructose solution in water.19 Therefore, a significant loss of
influence in increasing HMF yield in a biphasic solvent mediated HMF yield has been reported in aqueous medium. To study this
dehydration reaction. This effect was observed in boric acid effect, the starting fructose concentrations were varied from 5
catalyzed fructose dehydration in aqueous–organic biphasic wt% to 30 wt%. Under identical reaction conditions of 50 mol%
medium upon addition of NaCl and similar alkali metal of AlCl3 and at 120 ◦ C in water, the yield of HMF changed
salts.11 To check the salting-out effect in the present study, the from 53.9% to 52.1% (Fig. 1(a)). Thus, HMF yield remain
dehydration reactions of 5 wt% fructose and 5 wt% glucose unchanged upon increase in fructose concentrations from 5 wt%
were carried out with an AlCl3 catalyst in biphasic aqueous– to 30 wt%. Additional experiments for the dehydration of highly
MIBK solvent using NaCl salt. The results as tabulated in concentrated fructose (30 wt%) at 120 ◦ C under microwave
Table 1 (entries 1 and 2), Table 2 (entries 1 and 2) and Table assisted heating (Table 4) showed an appearance of insoluble
3 (entries 2 and 3) suggested that the effect of NaCl on HMF brown material in the solution. Attempts were made to analyze
yields was insignificant. A comparison of the results between the brown material using 1 HMR spectroscopy; however, due
entry 2 of Table 1 and entry 2 of Table 2 also revealed to poor solubility of this brown material in most deuterated
that HMF formation from the oil bath heating experiment solvents, signals were not intense (Fig. S4†). Inspection of the 1 H
for the optimized reaction time was significantly lower (36%) NMR spectrum in DMSO-d6 confirms the presence of signals
than that of microwave assisted heating (63%) experiment. To corresponding to two furan ring protons and a –CHO group
further check the salting-out effect, additional experiments were proton (d 9.50 ppm). This brown insoluble material is believed to
performed for a 30% fructose dehydration reaction with an AlCl3 be humin as mentioned in the literature.20 Thus, the probability
catalyst in the presence of Na2 SO4 , NaBr, and NaNO3 salts. The of loss of HMF to the form of humin may be high when the
results confirmed no significant effect of these salts on HMF reaction starts with concentrated fructose solution (ca. 30 wt%)
yield in biphasic aqueous–MIBK solvent. in aqueous medium.20

2862 | Green Chem., 2011, 13, 2859–2868 This journal is © The Royal Society of Chemistry 2011
 View Article Online
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

Fig. 1 (a) The effect of HMF yields on (a) fructose concentrations at fixed AlCl3 loading (50 mol%), and (b) AlCl3 catalyst dosage for 5 wt% fructose
dehydration in water at 120 ◦ C MW heating for 5 min.

3.3 Effect of the catalyst dosage on HMF yield

The dehydration reaction of fructose to HMF was studied

at variable catalyst concentrations for optimizing the reaction
conditions and maximizing the HMF yield. The concentrations
of AlCl3 were varied from 20 mol% to 50 mol% while other
reaction parameters were constant at [fructose] = 5 wt%, T =
120 ◦ C, reaction time = 5 min and solvent = water. As shown in
Fig. 1(b), the yield of HMF increased from 37.2% to 53.9% with
an increase in AlCl3 concentration from 20 mol% to 50 mol%.
The high catalyst loaded experiments improved HMF yield,
meaning that the kinetic of fructose dehydration is fast in the
case of higher catalyst loading. It has been shown in section 3.2
that HMF yields remain unchanged upon increase in starting
fructose concentration from 5 wt% to 30 wt%. This prompted
us to further study the catalyst variation experiments using Fig. 2 The effect of reaction time on HMF yield for the conversion
concentrated fructose solution (ca. 30 wt%). The results, as of carbohydrates (fructose, glucose, sucrose) to HMF in water under
tabulated in Table 4, show the yield of HMF increased from microwave heating at 120◦ . The concentrations of carbohydrate and
35.8% to 52.1% upon increase in catalyst loading from 20 wt% AlCl3 are 5 wt% and 50 mol%, respectively.
to 50 wt% (Table 4). Thus, the trend of HMF formation against
AlCl3 loading for 30 wt% fructose dehydration was similar to and sucrose dehydration reactions showed 5% and 7% increase
that obtained for 5 wt% fructose dehydration. in HMF yield during 2–20 min of reaction time, respectively.
It is important to note that the color of the reaction mixture
gradually turned to deep brown from pale yellow when the
dehydration of fructose was continued for a longer time (20
3.4 Effect of reaction time on HMF yield
min). Perhaps, the color change of the reaction mixture is
The reaction time of the dehydration reaction was varied to associated with the decomposition of the product HMF into
study the rate of HMF formation as a function of time. As side products. To further investigate the HMF decomposition
plotted in Fig. 2, the yield of HMF improved from 47.8% and end product, the reaction mixture was analyzed by 1 H NMR
to 52.3% upon increasing the duration of the reaction from spectroscopy. A 1 H NMR spectral analysis of the vacuum dried
30 s to 2 min for AlCl3 catalyzed dehydration of fructose at reaction mixture in chloroform-d1 confirmed the formation 5-
120 ◦ C. Further increase in reaction time to 20 min under similar chloromethylfurfural (CMF)22 and levulinic acid as side prod-
conditions resulted in 55.7% HMF yield. Thus, the kinetics of ucts (Fig. 3). The formation of levulinic acid can be explained by
HMF formation are rapid in first 0.5 min followed by a slow the rehydration of HMF.20,23 The self-polymerization of HMF
reaction. This behavior of increasing HMF yield with an increase molecules or cross-polymerization of HMF with fructose is also
in reaction time agreed well with the trend observed for fructose possible, particularly due to the fact that the HMF yield is
dehydration in water and water–DMSO solvent using mineral not significantly increased upon prolonged reaction.20 Although
acid and an ion-exchange resin catalyst under MW heating.14a,21 HMF rehydration can be suppressed in non-aqueous solvents,
A similar trend of slow HMF yield after 0.5 min of reaction time considerable rehydration of HMF is possible in aqueous or
was observed for AlCl3 catalyzed dehydration of glucose and mixed aqueous solvents. This hypothesis supports the literature
sucrose as well. Under comparable reaction conditions, glucose data where a lower HMF yield has been reported in aqueous

This journal is © The Royal Society of Chemistry 2011 Green Chem., 2011, 13, 2859–2868 | 2863
 View Article Online
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

Fig. 3 A representative 1 H NMR (CDCl3 ) spectrum of the reaction product obtained from the dehydration reaction of 5 wt% fructose with 50 mol%
AlCl3 in water under microwave heating at T = 120 ◦ C for 15 min reaction time.

medium than in biphasic solvent.24 The integration of proton

signals corresponding to the –CHO group in 1 H NMR spectra
(Fig. 3) of the reaction mixture suggests the formation of
CMF occurs in a 1 : 2 molar ratio of CMF : HMF in water.
The formation of CMF can be attributed to the ability of
AlCl3 to chlorinate the intermediate species formed during the
dehydration of fructose.
As discussed above, detection of levulinic acid as a by-product
in the 1 H NMR spectrum of the reaction mixture in the water
mediated reaction suggested the formation of formic acid as well.
However, the signal for the –COOH group proton of formic acid
(d 8.05 ppm) was too small in comparison to the levulinic acid
signal. This could be due to the fact that the –COOH proton
signal of formic acid did not resolve well in a reaction mixture or
disappeared due to its consumption in undefined side reactions.
In order to investigate the role of formic acid in fructose Fig. 4 The effect of aluminum catalysts and reaction time on HMF
dehydration, a control experiment was performed for the yields for the dehydration of 5 wt% fructose with 50 mol% catalysts in
dehydration of 5 wt% fructose with 0.56 mmol formic acid at water under microwave heating at 120 ◦ C.
120 ◦ C for 5 min under microwave-assisted heating. This control
experiment with formic acid did not produce HMF, which produced about 2–7% less HMF than its anhydrous analogue
precludes the possibility of the participation of a formic acid (Fig. 4), which could be due to the contribution of extra water
by-product as a catalyst in the fructose dehydration reaction. from the hydrated salt.

3.5 Effect of other aluminium catalysts 3.6 Conversion of glucose and sucrose

The dehydration reaction of fructose to HMF was also inves- Although fructose has been the preferred feedstock for HMF
tigated with two other commercially available aluminium salts, production, its occurrence in nature is limited. This drives the
AlBr3 and hydrated aluminium chloride AlCl3 ·6H2 O. The results attention to utilize more abundant carbohydrates, glucose and
(Fig. 4) show that AlBr3 is a less active catalyst for HMF sucrose, as raw materials for HMF synthesis. A recent study
production, which produced 35.1% HMF in 5 min as compared demonstrates that the conversion of glucose to HMF is likely
to 53.9% with the anhydrous AlCl3 catalyst. Under comparable to proceed via consecutive steps, namely, mutarotation and
reaction conditions, the hydrated aluminium chloride catalyst isomerization of glucose into fructose followed by dehydration

2864 | Green Chem., 2011, 13, 2859–2868 This journal is © The Royal Society of Chemistry 2011
 View Article Online
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

Scheme 2 The proposed mechanism for glucose dehydration to HMF with the AlCl3 catalyst in water.

Table 5 Results of sucrose dehydration catalyzed by AlCl3 under MW To understand the mechanism of glucose dehydration reac-
irradiation tion, a 1 H NMR study of a reaction mixture of glucose and AlCl3
Entry AlCl3 (mg) Solvent T (◦ C) t (min) HMF yield (%)
in DMSO-d6 was carried out at room temperature. It shows a
line broadening of –OH proton signals of glucose in the range
1 50 Water 120 5 30.0 of d 4–5 ppm. This line broadening of –OH protons indicates
2 50 Water 120 10 30.4 a H-bonding interaction between a Cl atom of AlCl3 and the
3 50 DMSO 140 5 4.6
4 50 DMSO 140 5 42.5
–OH protons of glucose forming an AlCl3 –glucose adduct. The
5 50 Water 120 2 26.7 line broadening of the –OH protons corresponding –H ◊ ◊ ◊ Cl–
6 50 Water 120 15 31.7 interaction completely disappeared at 100 ◦ C (Fig. S5(d)†) and
7 50 Water 120 20 33.2 a new signal appeared at d 9.49 ppm in the 1 H NMR spectra
8 33 Water 120 5 26.0
9 33 Water 120 15 27.8 after heating the reaction mixture at 100 ◦ C for 40 min. The
10 20 Water 120 5 20.6 new signal at d 9.49 ppm corresponds to the –CHO proton of
HMF. This experiment suggests that the dehydration of glucose
Sucrose = 5 wt%
in DMSO occurs via the formation of an intermediate species
between the –OH proton and the Cl atom of AlCl3 .
The dehydration reaction is believed to be initiated by the
of fructose to HMF.6 In the present study, the conversion of hydrolysis of AlCl3 . The hydrolysis of AlCl3 in water under mi-
glucose and sucrose to HMF was carried out with anhydrous crowave irradiation forms a cationic species [Al(OH)(H2 O)5 ]2+ .
AlCl3 catalyst in water at 120 ◦ C and in DMSO at 140 ◦ C under This active form of the catalyst, [Al(OH)(H2 O)5 ]+ acts as a
MW heating. The yield of HMF was monitored as a function potential electrophile that reacts with a-glucopyranose to form
of reaction time in the range of 2–20 min. The results, as shown the intermediate A (Scheme 2). Intermediate A is possibly
in Tables 3 and 5, reveal the maximum HMF yield from glucose converted to ketohexose B via a hydride transfer, which forms
and sucrose are 52.4% (entry 5 in Table 3) and 42.5% (entry 4 fructofuranose through cyclization. This proposed mechanistic
in Table 5), respectively, in DMSO at 140 ◦ C for 5 min reaction route for the formation of ketohexose B is similar to the
time. Under similar conditions, the HMF yields in water are mechanism described by Davis et al. using a SnCl4 catalyst.25
37.3% and 30.0% from glucose and sucrose, respectively. The The mechanism of HMF formation from fructose has been a
glucose dehydration reaction in water–MIBK biphasic solvent topic of debate over the years. In our case, the transformation of
produced 43% HMF as compared to 37.3% in water. fructofuranose to intermediate compound E occurs via a cyclic

This journal is © The Royal Society of Chemistry 2011 Green Chem., 2011, 13, 2859–2868 | 2865
 View Article Online

mechanism assisted by a proton that is generated during AlCl3 fructose was added into the reaction mixture. Fig. 5 shows that
hydrolysis. It is of note that the possibility of Cl- acting as a the loss of activity of the catalyst in terms of HMF yield is only
base was taken into account for abstracting the proton from 10% after five cycles (53.9% in first cycle versus 44% in fifth
the oxonium ion forming structure D (path A), as proposed by cycle).
Raines et al.26 In path B, Cl- functions as both the nucleophile
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

and base to form intermediate D. The cyclic intermediate E

has been identified recently by NMR study,17 which supports
our proposed mechanism for the formation of E (Scheme 2). A
similar mechanism involving intermediate E via a cyclic pathway
was also proposed by Riisager et al. in a fructose dehydration
reaction with boric acid catalyst.27 Intermediate compound E
then releases a water molecule to form the final product HMF.
This proposed mechanism was similar to a mechanism
previously described for SnCl4 catalyzed dehydration of glucose
in mixed [EMIM]BF4 (EMIM = 1-ethyl-3-methylimidazolium)
and DMSO-d6 solvents (9 : 1 w/w).8 Nevertheless, to understand
the role of other cations in the present reaction, a mechanistic
investigation for glucose (5 wt%) dehydration with SnCl4 catalyst
(20 mol%) was carried out in DMSO-d6 (0.5 mL). 1 H NMR
spectra revealed the formation of HMF upon heating at 80 ◦ C Fig. 5 Catalyst recyclability study for dehydration of fructose with
for 45 min, when the signals of –OH proton corresponding AlCl3 catalyst in water. Reaction conditions: fructose = 5 wt%, AlCl3 =
–H . . . Cl– interactions (interaction between –OH proton and 50 mol%, T = 120 ◦ C, t = 5 min.
Cl atom of SnCl4 ) disappeared. (see supporting information,
Fig. S6†). Further heating at 120 ◦ C for 30 min ensured signifi-
cant conversion of glucose to HMF, as revealed from 1 H NMR 3.8 Conversion of starch and inulin
(Fig. S6(d)†). Attempts were made to investigate the mechanism
The synthesis of HMF from biopolymers, such as starch and
of glucose dehydration with a lanthanide cation, Yb(OTf)3 (Tf =
inulin, was carried out using an anhydrous AlCl3 catalyst
trifluromethanesulfonate) as a catalyst in DMSO-d6 . 1 H NMR
under microwave irradiation. Table 6 shows details of reaction
investigation confirmed the low conversion of glucose into HMF
conditions for HMF synthesis from biopolymers. However,
in the presence of Yb(OTf)3 at 120 ◦ C for 1 h (Fig. S7†).
prerequisite for successful dehydration of starch to HMF is
This result is consistent with a previously reported Yb(OTf)3
the facile depolymerization of biopolymeric units. The depoly-
catalyzed glucose dehydration reaction in ionic liquid solvent
merization of the biopolymeric unit is reported to follow first-
[BMIM]Cl (BMIM = 1-butyl-3-methylimidazolium) where a
order kinetics with respect to catalyst concentration.24 In the
maximum 24% HMF yield was realized at 140 ◦ C for 6 h.12
present work, we used AlCl3 to hydrolyze the glycosidic linkages
between monomeric unit. AlCl3 is believed to be effective for
3.7 Catalyst life-time in water medium
hydrolysis of glycosidic linkages because of its strong Lewis
The life-time of the AlCl3 catalyst was studied for fructose acidic character. As shown in Table 6, a maximum 29.6%
dehydration reaction in aqueous medium by recycling the HMF was obtained from dehydration of starch with AlCl3
reaction mixture containing the spent catalyst. The life time catalyst in DMSO at 140 ◦ C under MW heating. The same
study of the catalyst was carried out with 5 wt% fructose and reaction in aqueous solvent produced about 21% HMF, which
50 mol% AlCl3 in water at 120 ◦ C for 5 min under microwave is clearly less than that obtained in DMSO. The effect of solvent
heating. Prior to recycle for the next run, HMF component was on the HMF yield has been discussed in section 3.1. The
extracted from the reaction mixture with diethyl ether and fresh dehydration of starch to HMF has been reported in the literature

Table 6 Results of starch and inulin dehydration catalyzed by AlCl3 under MW irradiation

Entry Substrate (5 wt%) AlCl3 (mol%) Solvent T (◦ C) t (min) HMF yield (%)

1 Starch 50 Water 120 2 19.6

2 Starch 50 Water 120 5 21.1
3 Starch 50 Water/MIBK 130 5 24.1
4 Starch 50 DMSO 140 5 30.6
5 Starch 50 Water 120 10 21.3
6 Starch 50 Water 120 5 14.0
7 Starch 50 Water 120 10 15.2
8 Inulin 50 Water 120 5 29.4
9 Inulin 50 Water 120 10 31.0
10 Inulin 50 Water/MIBK 130 5 32.5
11 Inulin 50 DMSO 140 5 39.2

Aqueous: organic = 1 : 2 volume (mL)

2866 | Green Chem., 2011, 13, 2859–2868 This journal is © The Royal Society of Chemistry 2011
 View Article Online

with SnCl4 and CrCl2 catalysts in 1-ethyl-3-methylimidazolium 4. Conclusions

chloride ([EMIM]Cl) and 1-octyl-3-methylimidazolium chloride
([OMIM]Cl]) solvent, respectively. These reported reactions In summary, we have demonstrated that AlCl3 is an excel-
produced 47% and 73% HMF with SnCl4 and CrCl2 catalysts, lent catalyst for the rapid conversion of carbohydrates and
respectively28 Although the yield of HMF in water is less than biopolymers into HMF in water, DMSO and water–MIBK
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

that reported in ionic liquid with different Lewis acid catalysts, biphasic solvents under microwave irradiation. The yield of
but our reaction in water is more sustainable in terms of avoiding HMF remains unchanged upon varying the starting fructose
the use of expensive ionic liquid solvents. concentrations from 5 wt% to 30 wt%. The optimized reaction
The direct synthesis of HMF from inulin occurs via hydrolysis conditions were developed to maximize HMF yield by varying
of inulin to fructose followed by dehydration of fructose. The catalyst loading, reaction time, aluminium salts etc. Satisfactory
synthesis of HMF from inulin was carried out with an anhydrous results were obtained when glucose, sucrose, starch, and inulin
AlCl3 catalyst in three different solvents (water, water–MIBK were used as the substrates. The mechanism of the AlCl3
and DMSO) using 1 : 1 molar ratio of inulin: AlCl3 for 5 min catalyzed glucose dehydration reaction is proposed to proceed
microwave heating. The inulin dehydration reaction produced through the isomerization of glucopyranose to fructofuranose
maximum 39% and 31% HMF in DMSO and water, respectively followed by a proton assisted transformation of fructofuranose
(Table 6). In an earlier report, Amberlyst 70 catalyst promoted to HMF. Finally, the Lewis acidic catalyst AlCl3 deserves further
hydrolysis and dehydration of inulin produced 60% HMF at investigation in less expensive ionic liquids to facilitate sustain-
110 ◦ C in [BMIM]Cl/glycerol carbonate (10 : 90).29 Although able conversions of cellulosic materials to value added products.
yield of HMF in aqueous solvent is less, but it is more sustainable
and environmentally friendly process.. Acknowledgements
The authors gratefully acknowledge financial support by the
3.8 Substrate scope of AlCl3 catalyst University Grant Commission (UGC), India and the University
of Delhi. SD thanks UGC, India for a DS Kothari Postdoctoral
From the experimental results demonstrated above, we sought to
Research Fellowship.
explore the substrate scope of AlCl3 catalyst and an estimation
of comparative yields of HMF under microwave heating. This
analysis for all substrates showed a trend of decrease in HMF References
yield from 53.9% to 21.1% in aqueous medium at 120 ◦ C for 1 (a) A. J. Ragauskas, C. K. Williams, B. H. Davidson, G. Britovsek, J.
5 min reaction duration. A similar decreasing trend of HMF Cairney, W. J. Frederick, J. P. Hallett, D. J. Leak, C. L. Liotta, J. R.
yield is also noted in DMSO for the dehydration reaction of Mielenz, R. Murphy, R. Templet and T. Tschaplinski, Science, 2006,
fructose, glucose, sucrose, inulin and starch (Fig. 6). The yields 311, 484; (b) G. Stephanopoulos, Science, 2007, 315, 801.
2 Y. Román-Leshkov, C. J. Barrett, Z. Y. Liu and J. A. Dumesic, Nature,
of HMF from starch and inulin were comparatively lower than 2007, 447, 982.
those obtained from fructose and glucose. However, the direct 3 (a) G. W. Huber, J. N. Chheda, C. J. Barrett and J. A. Dumesic,
synthesis of HMF from starch and inulin certainly is a promising Science, 2005, 308, 1446; (b) J. N. Chheda, G. W. Huber and J. A.
Dumesic, Angew. Chem., Int. Ed., 2007, 46, 7164; (c) A. Boisen, T.
alternative route to HMF production. Nevertheless, the AlCl3 Christensen, W. Fu, Y. Gorbanev, T. Hansen, J. Jensen, S. Klitgaard,S.
catalyst is shown to be effective to convert less-expensive and Pedersen, A. Riisager, T. Ståhlberg and J. Woodley, Chem. Eng. Res.
easily available biopolymers (starch and inulin) to HMF in Des., 2009, 87, 1318.
aqueous medium under microwave irradiation. This indeed 4 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411.
5 T. Ståhlberg, W. Fu, J. M. Woodley and A. Riisager, ChemSusChem,
expands the substrate scope of the catalyst and sustainability. 2011, 4, 451.
6 H. Zhao, J. E. Holladay, H. Brown and Z. C. Zhang, Science, 2007,
316, 1597.
7 G. Yong, Y. Zhang and J. Y. Ying, Angew. Chem., Int. Ed., 2008, 47,
9345.
8 S. Hu, Z. Zhang, J. Song, Y. Zhou and B. Han, Green Chem., 2009,
11, 1746.
9 J. Y. Gerentt and Y. Zhang, ChemSusChem, 2009, 2, 731.
10 Z. Zhang, Q. Wang, H. Xie, W. Liu and Z. K. Zhao, ChemSusChem,
2011, 4, 131.
11 T. S. Hansen, J. Mielby and A. Riisager, Green Chem., 2011, 13, 109.
12 T. Stahlberg, M. G. Sorensen and A. Riisager, Green Chem., 2010,
12, 321.
13 S. Zhao, M. Cheng, J. Li, J. Tian and X. Wang, Chem. Commun.,
2011, 47, 2176.
14 (a) T. S. Hansen, J. M. Woodley and A. Riisager, Carbohydr. Res.,
2009, 344, 2568; (b) Z. Zhang and Z. K. Zhao, Biores. Tech., 2010,
101, 1111; (c) C. Z. Li, Z. H. Zhang and Z. B. K. Zhao, Tetrahedron
Lett., 2009, 50, 5403; (d) X. Qi, M. Watanabe, T. M. Aida and R. L.
Smith Jr., ChemSusChem, 2010, 3, 1071.
15 D. Dallinger and C. O. Kappe, Chem. Rev., 2007, 107, 2563.
16 X. Tong and Y. Li, ChemSusChem, 2010, 3, 350.
17 A. S. Amarasekara, L. D. Williams and C. C. Ebede, Carbohydr. Res.,
Fig. 6 A comparison of HMF yields for dehydration of various 2008, 343, 3021.
substrates in water and DMSO. Reaction conditions: substrate = 5 wt%, 18 Y. Román-Leshkov and J. A. Dumesic, Top. Catal., 2009, 52, 297.
AlCl3 = 50 mol%, T = 120 ◦ C, t = 5 min. 19 B. F. M. Kuster, Starch/Staerke, 1990, 42, 314.

This journal is © The Royal Society of Chemistry 2011 Green Chem., 2011, 13, 2859–2868 | 2867
 View Article Online

20 X. Qi, M. Watanabe, T. M. Aida and R. L. Smith Jr., Green Chem., 25 Y. Román-Leshkov, M. Moliner, J. A. Labinger and M. E. Davis,
2009, 11, 1327. Angew. Chem., Int. Ed., 2010, 49, 8954.
21 X. Zi, M. Watanabe, T. M. Aida and R. L. Smith Jr., Green Chem., 26 J. B. Binder and R. T. Raines, J. Am. Chem. Soc., 2009, 131,
2008, 10, 799. 1979.
22 M. Brasholz, K. von Känel, C. H. Hornung, S. Saubern and J. 27 T. Ståhlberg, S. Rodriguez-Rodriguez, P. Fristrup and A. Riisager,
Tsanaktsidis, Green Chem., 2011, 13, 1114. Chem.–Eur. J., 2011, 17, 1456.
Published on 26 August 2011. Downloaded by Heinrich Heine University of Duesseldorf on 20/05/2013 14:03:37.

23 F. S. Asghari and H. Yoshida, Ind. Eng. Chem. Res., 2007, 46, 7703. 28 J.-A. Chun, J.-W. Lee, Y.-B. Yi, S.-S. Hong and C.-H. Chung,
24 (a) R. Rinaldi, R. Palkovits and F. Scüth, Angew. Chem., Int. Ed., Starch/Staerke, 2010, 62, 326.
2008, 47, 8047; (b) R. Rinaldi, N. Meine, J. vom Stein, R. Palkovits 29 M. Benoit, Y. Brissonnet, E. Guelou, K. D. Oliveira Vigier, J. Barrault
and F. Scüth, ChemSusChem, 2010, 3, 266. and F. Jerome, ChemSusChem, 2010, 3, 1304.

2868 | Green Chem., 2011, 13, 2859–2868 This journal is © The Royal Society of Chemistry 2011