Direct Production of 5-Hydroxymethylfurfural via Catalytic Conversion of Simple and Complex Sugars over Phosphated TiO2

Article abstract of DOI:10.1002/cssc.201500395

A water-THF biphasic system containing N-methyl-2-pyrrolidone (NMP) was found to enable the efficient synthesis of 5-hydroxymethylfurfural (HMF) from a variety of sugars (simple to complex) using phosphated TiO₂ as a catalyst. Fructose and glucose were selectively converted to HMF resulting in 98 % and 90 % yield, respectively, at 175 °C. Cellobiose and sucrose also gave rise to high HMF yields of 94 % and 98 %, respectively, at 180 °C. Other sugar variants such as starch (potato and rice) and cellulose were also investigated. The yields of HMF from starch (80-85 %) were high, whereas cellulose resulted in a modest yield of 33%. Direct transformation of cellulose to HMF in significant yield (86 %) was assisted by mechanocatalytic depolymerization - ball milling of acid-impregnated cellulose. This effectively reduced cellulose crystallinity and particle size, forming soluble cello-oligomers; this is responsible for the enhanced substrate-catalytic sites contact and subsequent rate of HMF formation. During catalyst recyclability, P-TiO₂ was observed to be reusable for four cycles without any loss in activity. We also investigated the conversion of the cello-oligomers to HMF in a continuous flow reactor. Good HMF yield (53 %) was achieved using a water-methyl isobutyl ketone+NMP biphasic system.

Full text of DOI:10.1002/cssc.201500395

DOI: 10.1002/cssc.201500395
Full Papers
Direct Production of 5-Hydroxymethylfurfural via Catalytic
Conversion of Simple and Complex Sugars over
Phosphated TiO2
Luqman Atanda,[a] Abhijit Shrotri,[a, b] Swathi Mukundan,[a] Qing Ma,[a] Muxina Konarova,[a]
and Jorge Beltramini*[a]
A water–THF biphasic system containing N-methyl-2-pyrroli-
done (NMP) was found to enable the efficient synthesis of 5-
hydroxymethylfurfural (HMF) from a variety of sugars (simple
to complex) using phosphated TiO2 as a catalyst. Fructose and
glucose were selectively converted to HMF resulting in 98%
and 90% yield, respectively, at 1758C. Cellobiose and sucrose
also gave rise to high HMF yields of 94% and 98%, respective-
ly, at 1808C. Other sugar variants such as starch (potato and
rice) and cellulose were also investigated. The yields of HMF
from starch (80–85%) were high, whereas cellulose resulted in
a modest yield of 33%. Direct transformation of cellulose to
HMF in significant yield (86%) was assisted by mechanocatalyt-
ic depolymerization—ball milling of acid-impregnated cellu-
lose. This effectively reduced cellulose crystallinity and particle
size, forming soluble cello-oligomers; this is responsible for the
enhanced substrate–catalytic sites contact and subsequent
rate of HMF formation. During catalyst recyclability, P–TiO2 was
observed to be reusable for four cycles without any loss in ac-
tivity. We also investigated the conversion of the cello-oligo-
mers to HMF in a continuous flow reactor. Good HMF yield
(53%) was achieved using a water–methyl isobutyl ketone+
NMP biphasic system.
Introduction
5-Hydroxymethylfurfural (HMF) is a vital platform chemical that
could revolutionize the biorefinery industry. It is a key compo-
nent in the production of high-value chemicals, polymers, and
fuels.[1] For example, HMF can be readily oxidized to 2,5-furan-
dicarboxylic acid (FDCA),[2] a potential replacement for petrole-
um-based terephthalic acid, relevant to the plastic industry for
the production of polyethylene terephthalate (PET).[3] Rehydra-
tion of HMF gives levulinic acid from which g-valerolactone
can be produced, a useful chemical that serves as a solvent
and fuel additive.[4] Another important derivative of HMF is 2,5-
dimethylfuran (DMF), a potential biofuel.[5] DMF has an energy
density comparable to gasoline (31.5 vs. 35 MJLÀ1) and approx-
imately 40% higher than ethanol (23 MJLÀ1),[6] making it suit-
able as a replacement for ethanol in gasoline–ethanol blends.
2-Methylfuran, which is an intermediate of HMF hydrogenoly-
sis, also possesses excellent fuel qualities.[5] Therefore, critical-
to-efficient production of bio-fuels and chemicals as well as
a replacement of petroleum-derived counterparts requires
a cost-effective process for selective production of HMF from
biorenewable feedstock.[7]
HMF can be obtained by catalytic dehydration of simple
sugars such as fructose and glucose or directly from cellulose
(a complex sugar) as shown in Scheme 1. In spite of excellent
production of HMF achievable from fructose and glucose,[8] in
practice it is more economical and sustainable to use a bio-
mass-derived sugar that is readily available and accessible. Cel-
lulose is readily obtained from industry, forestry, or agricultural
residues.[9] Unlike starch, which is consumed by humans, cellu-
lose is inedible; thus, it has the added advantage of not im-
pacting negatively on agricultural food production. Hence,
a one-pot chemical transformation of cellulose to HMF is a po-
tential strategy to effectively utilize this enormous biomass re-
serve to supply the next generation of chemicals and fuels.
Nonetheless, chemical conversion of cellulose faces a major
challenge, that is, its robust structure.[10] Cellulose has a high
crystallinity and chemical stability and is insoluble in water,
which leads to poor reactivity. A possible solution to overcome
this issue is to solubilize cellulose prior to reaction.
[a] L. Atanda, Dr. A. Shrotri, S. Mukundan, Dr. Q. Ma, Dr. M. Konarova,
Dr. J. Beltramini
Australian Institute for Bioengineering&Nanotechnology
Nanomaterials Center and School of Chemical Engineering
The University of Queensland
St Lucia, QLD 4072 (Australia)
E-mail: j.beltramini@uq.edu.au
[b] Dr. A. Shrotri
Although cellulose is known to be insoluble in most conven-
tional solvents, recent advances in the dissolution of cellulose
in the presence of ionic liquids[11] or subcritical/supercritical
fluids[12] have been proposed. These techniques depolymerize
cellulose into soluble oligosaccharides, thus allowing subse-
quent production of useful chemical compounds in high
yields. Despite the success and viability of these strategies,
Catalysis Research Center
Hokkaido University
Kita 21 Nishi 10, Kita-Ku, Sapporo 001-0021 (Japan)
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201500395.
This publication is part of a Special Issue on the “Future Energy” confer-
ence in Sydney, Australia. To view the complete issue, visit:
http://onlinelibrary.wiley.com/doi/10.1002/cssc.v8.17/issuetoc.
ChemSusChem 2015, 8, 2907 – 2916
2907
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
the only existing report of HMF
production from cellulose utiliz-
ing a flow reactor is by McNeff
et al.[18] In their approach,
a fixed-bed porous-metal-oxide-
based catalytic process was uti-
lized. Dissolution of cellulose
was achieved by introducing
a preheated water–organic mix-
ture into a solubilization cham-
ber containing cellulose. The
cascade process of a preheater–
solubilization chamber–catalytic
reactor allowed the continuous
depolymerization of cellulose,
which resulted in 87% conver-
sion of cellulose to produce HMF
in 35% yield.
Herein, we report the catalytic
production of HMF from sugars
applying phosphated TiO2 cata-
Scheme 1. Catalytic reaction pathway for conversion of sugar to HMF.
lyst under batch as well as con-
tinuous flow conditions. The
high costs associated with recycling of ionic liquids[13] or high
energy input required to generate sub- or super-critical
fluids[14] pose a number of challenges for process economics
and sustainability. Alternatively, cellulose can be solubilized
through mechanocatalytic depolymerization.[15] This technique
involves a contact between cellulose and an acid catalyst, fol-
lowed by a mechanically assisted solid-state depolymerization.
This approach of cellulose dissolution is scalable, amenable to
multi-feedstock, involves less waste, and can be easily integrat-
ed into existing biorefineries.[15a] Since the pioneering work of
Hick et al.,[15a] mechanocatalytic depolymerization of cellulose
using acid catalysts has become of great interest for the com-
plete dissolution of cellulose.[15b,c]
main aim is to selectively transform cellulose to HMF. However,
because of the complexity of working with cellulose, a para-
metrical activity study for the catalytic conversion of glucose
to HMF using pure and modified TiO2 with phosphorus, tung-
sten, molybdenum, and vanadium oxides was performed. In
addition, the effect of phase modifiers on the aqueous phase
of the biphasic water–organic medium was examined. The ver-
satility of the batch process was checked by testing other
sugar variants such as fructose, cellobiose, sucrose, starch, and
cellulose. Dissolution rate and reactivity of cellulose was en-
hanced through a mechanocatalytic treatment using the pro-
cedure described by Shrotri et al.[15c] Lastly, the viability of con-
verting the water-soluble cello-oligomers to HMF in a biphasic
continuous reactor was explored.
Due to the realization of the depolymerization of cellulose
into soluble oligomers, it would be highly desirable to develop
an efficient heterogeneous catalytic route for HMF production
under continuous flow conditions. A plethora of homogeneous
and heterogeneous catalytic systems have been described in
the literature for the conversion of solubilized cellulose to HMF
in a batch process.[16] For example, cellulose dissolution with 1-
butyl-3-methyl-imidazolium chloride ([BMIM]Cl) was demon-
strated to facilitate the conversion of cellulose to sugars and
further to HMF in high yields (62%) catalyzed by a paired
CrCl3/LiCl catalyst system.[16c] Shi et al.[16d] described the forma-
tion of 53% HMF by direct degradation of cellulose in a bipha-
sic system using concentrated NaHSO4 and ZnSO4 as a co-cata-
lyst. Nandiwale et al.[16e] reported the catalytic conversion of
cellulose using bimodal micro-/mesoporous H-ZSM-5 zeolite, in
which HMF was formed in 46% yield and 67% cellulose con-
version. Mascal and Nikitin[16f,17] reported another route of cel-
lulose deconstruction in an aqueous HCl–LiCl solution, yielding
mainly 5-chloromethylfurfural, which could subsequently be
converted to high HMF in high yield (72%) by performing
a simple hydrolytic reaction.[16f] To the best of our knowledge,
Results and Discussion
A neutral amine sol–gel method was used to prepare TiO2
nanoparticles. A previous report of the material synthesis
showed that the TiO2 nanoparticles were of anatase phase
with a mesoscopic structure.[19] The material possessed Lewis
acidity and was found active for the dehydration of glucose to
HMF. In addition, the catalytic performance of TiO2 was pro-
moted by phosphate treatment due to enhanced surface acidi-
ty. Indeed, increasing the surface area and the accessibility of
the active sites on TiO2 was also reported by others to improve
the catalytic performance of TiO2 to efficiently catalyze the de-
hydration of sugars into HMF.[20] In this study, we investigate
the effect of tungstate, molybdate, and vanadate on the prop-
erties of TiO2, especially on the surface acidity in comparison
to that of phosphate. For this purpose, hydrated titanium hy-
droxide initially prepared through the neutral amine sol–gel
method was impregnated with ammonium salts of phospho-
rus, tungsten, molybdenum, and vanadium.
ChemSusChem 2015, 8, 2907 – 2916
www.chemsuschem.org
2908
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Catalyst characterization
Table 1. Physical and acid properties of pure and modified TiO2 nanopar-
ticles.
Powder X-ray diffractometry (XRD) was used to examine the
crystalline structure of the as-synthesized materials. Figure 1
presents the XRD patterns of TiO2 nanoparticles before and
after modification.
[a]
Entry Sample SBET
Pore volume[b] Crystal size[c] Acidity[d] [mmolgÀ1
]
[m2 gÀ1] [cm3 gÀ1
]
[nm]
PyL
PyB
1
2
3
4
5
TiO2
54.7
0.22
0.43
0.32
0.33
0.02
27.3
6.12
9.54
9.36
53.0
3.68
0.72
1.73
–
–
P-TiO2 151.0
W-TiO2 112.5
Mo-TiO2 115.2
V-TiO2 9.34
3.81
5.35
18.2
0.97
0.20
[a] BET surface area. [b] Determined by Barrett–Joyner–Halenda (BJH)
measurement at P/P0 =0.99. [c] Measured by XRD using Debye–Scherrer
equation for the (101) plane. [d] Determined using the Lambert–Beer
Law.
(Figure S1e). The isotherm of V-TiO2 resembles a typical materi-
al with low surface area. The reason for this can be ascribed to
the phase transition from anatase to rutile,[22] which causes
a drastic reduction in the surface area and corresponding crys-
tal growth.
TEM micrographs of the samples are shown in Figure 2. The
morphological features of TiO2 are influenced by the presence
of other oxides. We clearly observed that the particle size of
TiO2 (Figure 2a) is far larger than that of P-TiO2, W-TiO2, and
Mo-TiO2 (Figure 2b–d), which can be used to explain the low
specific surface area of TiO2 in comparison with these three
samples. The phosphate, tungstate, and molybdate ions in
TiO2 are believed to stabilize the TiO2 structure and prevent
sintering during thermal treatment, which could have led to
grain growth.[19,23] Contrarily, a remarkable increase in crystal
size is observed with V-TiO2, suggesting crystal growth (Fig-
Figure 1. XRD measurement of (a) TiO2, (b) V-TiO2, (c) P-TiO2, (d) W-TiO2, and
*
(e) Mo-TiO2. (*) Vanadium oxide and ( ) rutile.
We observed a distinct peak reflection at 2q value of 29.58,
which could be matched with the (101) plane of the JCPDS file
no. 21-1272. This indicates the formation of the anatase TiO2
polymorph in all samples. It is also of interest that no evidence
of the crystalline phases of the phosphorus, tungsten, and mo-
lybdenum oxides were observed in Figure 1c–e. This suggests
that these oxides could be present in an amorphous phase or
as crystals with very small sizes that are not detectable by
XRD. Contrarily, weak reflections of crystalline vanadium oxide
and the onset of rutile TiO2 formation were identified in Fig-
ure 1b, which can be correlated with JCPDS file numbers 01-
089-0612 and 21-1276, respectively. The anatase-to-rutile trans-
formation was described to proceed via the surface reaction of
vanadium oxide with titanium sites governed by a grain
growth mechanism.[21] This fact is supported by the estimated
crystal size determined by applying Scherrer equation to the
(101) plane. It is noted that the average crystal size of TiO2 in-
creased approximately twofold from 27.3 to 53 nm on addition
of vanadium ions, indicating crystal growth (Table 1). On the
other hand, according to the XRD line broadening analysis of
other modified TiO2 samples, the crystal size of TiO2 was re-
duced and estimated to be between 6 and 10 nm.
The result of the N2 adsorption–desorption isotherm of the
TiO2 samples is also presented in Table 1. The isotherm of pure
TiO2 shown in Figure S1a in the Supporting Information exhib-
its a sharp increase in the adsorbed N2 volume at a relative
pressure of 0.65, which is characteristic for capillary condensa-
tion of a material with uniform mesopore structures. The iso-
therms of the modified TiO2 samples (Figure S1b–d) are similar
to the parent material, and the onset of the condensation step
shifts to lower values (P/P0 ꢀ0.6), with the exception of V-TiO2
Figure 2. TEM micrographs of (a) TiO2, (b) P-TiO2, (c) W-TiO2, (d) Mo-TiO2, and
(e) V-TiO2.
ChemSusChem 2015, 8, 2907 – 2916
www.chemsuschem.org
2909
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ure 2e). Together with the XRD results, a mechanism of crystal
growth can be deduced to occur by lattice strain relaxation,
which involves reconfiguration of the TiO2 crystal geometry
and reduction of the induced strain arising from lattice distor-
tion caused by the presence of vanadium ions.[24] This behavior
is facilitated by thermal treatment that increases mobility and
segregation of vanadium oxide as well as transformation of
the TiO2 structure to the thermodynamically more stable rutile
phase.[24]
interaction between the surface of titanium hydroxide and mo-
lybdenum oxide. In the case of the other modified TiO2 sam-
ples, the presence of weak band at 1450 cmÀ1 implies that
a relatively small amount of Lewis acid sites are still retained.
Catalyst evaluation
Simple sugars to HMF
Pyridine adsorption experiments on the pure and modified
TiO2 nanoparticles were conducted to probe the presence of
Lewis or Brønsted surface acid sites. As shown in Figure 3a,
the pyridine infrared (py-IR) spectrum of TiO2 is dominated by
absorption bands in the regions of 1400–1450 cmÀ1 and 1570–
1620 cmÀ1, typical of pyridine molecules coordinated with
Lewis acid sites. Therefore, TiO2 is defined here as a pure Lewis
acid catalyst.
Previously, we reported that P-TiO2 exhibits a better catalytic
performance than TiO2 in the transformation of glucose to
HMF in a water–butanol reaction medium.[19] Herein, glucose
dehydration is carried out in a water–THF biphase system, and
the catalytic performances of TiO2 and P-TiO2 are compared.
Preliminary experiments using 2 wt% glucose concentration as
the reaction feed were conducted and the results are present-
ed in Table 2. The outcome of the experiment confirmed the
Table 2. Catalytic conversion of glucose to HMF using TiO2 and modified
TiO2 nanoparticles.[a]
Entry
Sample
Glucose conc.
[wt%]
Glucose conv.
[%]
HMF yield
[%]
1
2
3
4
5
6
TiO2
2
2
5
5
5
5
90.4
93.6
96.5
99.9
98.5
99.9
72.8
83.4
62.8
35.5
27.5
17.0
P-TiO2
P-TiO2
V-TiO2
W-TiO2
Mo-TiO2
[a] Reaction conditions: glucose/cat. 4:1 w/w ratio, 100 mL solvent
(water/THF=1:4 v/v), 4 g NaCl, 105 min reaction time, 1758C reaction
temperature, 20 bar Ar gas.
Figure 3. FTIR spectra of pyridine adsorbed on (a) TiO2, (b) P-TiO2, (c) W-TiO2,
(d) Mo-TiO2, and e) V-TiO2.
improved activity of P-TiO2 over TiO2. A conversion of 90.4%
and 72.8% HMF yield was achieved with TiO2 (Table 2, entry 1).
However, the HMF yield increased to 83.4% at 93.6% conver-
sion when P-TiO2 was used as catalyst (entry 2). Other identifia-
ble products were fructose, levoglucosan, levulinic acid, formic
acid, and acetic acid; their corresponding yields are listed in
Table S1.
The surface interaction of TiO2 with the oxides of phospho-
rus, tungsten, molybdenum, and vanadium leads to considera-
ble modification of the nature of its surface acid sites. The de-
tection of a band in the region of 1500–1560 cmÀ1 on TiO2
after modification with these oxides (Figure 3b–e) allows us to
identify the presence of Brønsted acid sites. The variation of
the intensity of this band indicates a change in concentration
of the available Brønsted acid sites. The estimated concentra-
tions of Brønsted (PyB) and Lewis (PyL) acid sites are presented
in Table 1. The result shows that the Brønsted acidity decreases
in the order: Mo-TiO2 >W-TiO2 >P-TiO2 >V-TiO2. More so, Mo-
TiO2 appears to be a pure Brønsted acid catalyst because of
the disappearance of all the bands associated with Lewis acid
sites. Similarly, Damyanova et al.[25] observed that pyridine ad-
sorbed on TiO2-supported 12-molybdophosphate existed pre-
dominantly in the protonated form (Brønsted acidity) when
the sample was pretreated at 2508C; however, Lewis acidity
appeared on the sample pretreated at 3508C due to the pres-
ence of isolated Mo6+ cations of molybdenum oxide phase(s).
Hence, a plausible reason for the acidity change of TiO2 from
Lewis to Brønsted in Mo-TiO2 may be as a result of the strong
However, from an industrial point of view, it is of interest to
work at a relatively high glucose concentration to achieve
a more economical and sustainable process. When conducting
the reaction with 5 wt% glucose concentration on P-TiO2
under similar reaction conditions, the HMF yield dropped to
62.8% (entry 3). This phenomenon can be explained as fol-
lows: at high concentration, glucose forms oligosaccharides
that contain reactive hydroxy groups, which give rise to higher
rates of cross-polymerizations with reactive intermediates and
HMF.[26] This hypothesis is supported by a color change of the
catalyst from yellowish-white to dark brown at the end of the
reaction (Figure S2). The color change can be ascribed to the
deposition and coverage of the catalyst surface with humin
compounds. Therefore, it is essential to develop a catalyst
system that is resistant to humins or a reaction medium that
hinders the formation of humins to achieve effective conver-
sion of high glucose concentration feedstocks into HMF with
ChemSusChem 2015, 8, 2907 – 2916
www.chemsuschem.org
2910
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
good selectivity and high conversion. Hence, TiO2 was modi-
fied by the introduction of molybdenum, vanadium, and tung-
sten oxides in a similar procedure as that for phosphorus
oxide.[19] This attempt was unsuccessful as lesser yields of HMF
were achieved (entries 4–6). The catalytic performance of these
materials can be understood with respect to their surface acidi-
ty. On the basis of the py-IR results, Mo-TiO2 is characterized
mainly by Brønsted acidity and very low HMF yield could be
achieved. This is because the isomerization of glucose to fruc-
tose, crucial to HMF formation, is catalyzed by Lewis acidity.
On the other hand, W-TiO2 and V-TiO2 have both Brønsted and
Lewis acid functionalities just as P-TiO2. The band intensities of
W-TiO2 suggest the presence of a higher concentration of acid
sites than P-TiO2. The reduced HMF yield on W-TiO2 may then
be ascribed to the presence of excess acid sites, which could
be responsible for favoring unwanted reactions. There is a gen-
eral consensus that interaction of bulk vanadium oxide with
the surface of TiO2 forms a thin monolayer of vanadium
oxide.[27] We reason that the segregated vanadium oxide led to
partial monolayer coverage of the TiO2 surface. As a result, the
Brønsted acidity of the vanadium oxide comes into contact
with the glucose substrate and only a few Lewis acid sites are
available to play a relevant role during the isomerization of
glucose to fructose. However, V-TiO2 results in a higher HMF
yield in comparison to W-TiO2 (35.5% vs. 27.5%), which can be
ascribed to the acid site concentration. W-TiO2 has a higher
concentration of acid sites (especially Brønsted sites), which
can easily promote rapid degradation of HMF.
tioning of HMF into the organic layer. Horvat et al.[29] explained
the mechanism of HMF transformation in aqueous medium.
The authors stated that addition of water to the 2,3-carbon po-
sitions on HMF is responsible for the undesired polymerization
reactions to humin, whereas water added to 4,5-carbon posi-
tions gave way to levulinic acid formation via decarboxylation
to produce formic acid. Therefore, we can say that the benefi-
cial role of NMP on HMF formation is related to its ability to
minimize ring opening of these carbon atoms. The less intense
color change of the catalyst after reaction also supports the
fact that NMP helps to reduce humin formation and deposition
on the catalyst (Figure S3) in comparison to the reaction with-
out NMP.
Furthermore, we attempted to favor the formation of HMF
by increasing the volume of NMP added to the water–THF re-
action medium. We observed that HMF yield increased to
a maximum value of 90.5% at 60:20 THF/NMP volume ratio
(Figure 4b). Further increase in NMP volume caused a drastic
reduction in HMF yield. This is because a single phase reaction
system formed at the end of the reaction. Since there is a lack
of partitioning, the already formed HMF undergoes further re-
action. Thus, the optimal reaction medium for subsequent ex-
perimental design is 20:60:20 water/THF/NMP volume ratio.
Using these reaction conditions, glucose was replaced with
fructose as the feed in the production of HMF. The results are
presented in Table 3 and show that conversion of fructose into
HMF proceeds significantly faster than with glucose, achieving
98.6% HMF yield within 30 min of reaction time (Table 3,
entry 2). This suggests that the mechanism for the transforma-
tion of glucose to HMF occurs via the rate-determining step
(formation of the intermediate fructose). Other simple sugars
also investigated for the production of HMF are cellobiose and
sucrose. Cellobiose and sucrose are dimers consisting of glu-
cose–glucose and glucose–fructose monomers, respectively. It
is expected that the presence of additional functional groups
in each of these dimers will increase the multistep reactions,
beginning with hydrolytic reaction to produce their constitu-
ent monomer units, followed by isomerization and dehydra-
tion reaction steps. Therefore, we carried out the reaction with
the dimers at a higher temperature to facilitate the
The other approach was to modify the reaction medium to
minimize HMF degradation. Undesired side reactions can be
suppressed through the addition of phase modifiers.[28] In this
study, organic solvents such as acetone, acetonitrile (AN), N-
methyl-2-pyrrolidone (NMP), and toluene were examined. Fig-
ure 4a shows that selectivity to HMF was significantly im-
proved by the addition of NMP. For example, addition of
10 mL of NMP to the water–THF medium enabled about 80%
HMF yield. According to a report from Romµn-Leshkov et al.,[28]
NMP acts as an aqueous phase modifier that can suppress
humin formation in water and simultaneously enhance parti-
hydrolysis step. Identical results of HMF formation
were obtained for both cellobiose and sucrose (en-
tries 3 and 4). However, the higher HMF yield ob-
tained from sucrose in comparison with cellobiose
(98.2% vs. 94.2%) is attributed to the presence of
fructose in the dimer structure, which can undergo
dehydration faster than glucose.
Complex sugars to HMF
Considering the successful production of HMF in
high selectivity from simple sugars, we investigated
a much more practical and sustainable production of
Figure 4. (a) Influence of organic co-solvents on the selective conversion of glucose to
HMF; 10 mL organic co-solvent, 70 mL THF. (b) Volume ratio effect of THF/NMP on glu-
cose-to-HMF reaction. Reaction conditions: 4:1 glucose/cat. w/w ratio, 1.25 g P-TiO2 cata-
lyst, 100 mL solvent (water/organic=1:4 v/v), 4 g NaCl, 105 min reaction time, 1758C re-
HMF from complex sugars. Thus, we explored the re-
activity of starch and cellulose towards HMF forma-
tion. The conversion of starch into HMF proceeds sig-
nificantly faster and with higher selectivity than cellu-
*
*
action temperature, 20 bar Ar gas. ( ) Glucose conversion and ( ) HMF yield.
ChemSusChem 2015, 8, 2907 – 2916
www.chemsuschem.org
2911
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
cose to HMF as a representative reaction. The reaction was
performed in a water/(THF+NMP) medium at 1758C for
105 min. After each reaction, the catalyst was recovered by fil-
tration and the spent catalyst was washed with acetone and
kept under vacuum at 808C overnight for drying. The recov-
ered catalyst was reused without any post treatment. This pro-
cedure was followed for four subsequent cycles, and the re-
sults are shown in Figure 5. The activity of the catalyst resulted
in >90% yield of HMF and remained fairly constant even after
its repeated use for four times. As reported previously, catalyst
regeneration was performed because of deactivation resulting
from catalyst surface coverage by deposited humins, albeit
structural stability of the catalyst was maintained.[19] With NMP
serving as an aqueous phase modifier, unwanted reaction to
Table 3. Catalytic conversion of a variant of sugars to HMF over P-TiO2
catalyst.[a]
Entry
Substrate
T
[8C]
Conversion
[%]
HMF yield
[%]
1
2
3
4
5
6
7
8
fructose[b]
glucose
cellobiose
sucrose
starch-rice
starch-potato
cellulose
175
175
180
180
180
180
180
180
99.9
98.2
99.7
99.8
99.7
99.8
56.7
99.9
98.6
90.5
94.2
98.2
80.7
84.6
33.0
86.2
pretreated cellulose
[a] Reaction conditions: substrate/cat. 4:1 w/w. ratio, 100 mL solvent
(water/THF+NMP=1:4 v/v), 4 g NaCl, 105 min reaction time, 20 bar Ar
gas. [b] 30 min reaction time.
lose (Table 3, entries 5–7). Starch from rice and potato gave
rise to good yields of HMF (80–85%), whereas a modest HMF
yield (33%) was obtained from cellulose. The reason for the
disparity in the reactivity of starch and cellulose can be attrib-
uted to the structural difference of the two polymers. Starch is
a polymer of glucose units linked together through a-1,4 or a-
1,6 linkages whereas cellulose is a polymer of glucose units
linked together in an unbranched b-1,4 fashion, which are
densely packed because of strong inter-chain hydrogen
bonds.[9b] Due to the presence of strong glycosidic bonding
between the sugar moieties of cellulose, it is more crystalline
and less soluble than starch. On this basis, reactivity is low due
to poor interaction between cellulose and the catalyst, which
are both present in solid state. However, cellulose is more ideal
as a starting material for HMF production, as it is the most
abundant naturally occurring feedstock. Conversion of cellu-
lose to HMF in significant yields can be achieved through a me-
chanocatalytic depolymerization of cellulose, which should
reduce cellulose crystallinity and simultaneously facilitate the
cleavage of b-1,4-glycosidic bond linkages.[15a–e] By acid impreg-
nation of cellulose, formation of soluble cello-oligomers during
ball milling is facilitated. Entry 8 of Table 3 shows that conver-
sion rate of cellulose was significantly promoted after the pre-
treatment process, producing 86.2% yield of HMF. Our re-
search group has previously shown through X-ray diffraction
and scanning electron microscopy analyses that the structure
of the pretreated cellulose is amorphous and has a reduced
particle size.[15c] Furthermore, liquid state nuclear magnetic res-
onance results suggest the formation of oligomers comprising
mainly C1–C6 carbons. All these factors are responsible for the
enhanced solubility and reactivity of the pretreated cellulose.
In this way, a good HMF yield can be effectively produced
from cellulose.
Figure 5. Catalyst recyclability test. Reaction conditions: glucose/cat. 4:1 w/
w, 100 mL solvent (water/THF+NMP=1:4 v/v), 4 g NaCl, 105 min reaction
*
time, 1758C reaction temperature, 20 bar Ar gas. ( ) Glucose conversion and
&
( ) HMF yield.
humins was greatly inhibited. This was further supported
through a quantitative analysis of deposited humins on the
catalyst. The spent catalysts after the reaction, with and with-
out NMP in the reaction system, were recovered, washed with
deionized water, filtered, and kept under vacuum at 808C over-
night for drying. The carbon content of a known amount of
samples was analyzed using a CHNS-O elemental analyzer. The
carbon content of the fresh catalyst sample was also analyzed
for reference. The result shows that <0.4% carbon was pres-
ent on the fresh sample. For the reaction system with NMP,
4.1% carbon was found on the spent catalyst, which increased
to 21.2% on the spent catalyst from the reaction system with-
out NMP. The measured carbon content can be related to the
amount of humins deposited on the catalyst surface, which is
relatively high in the reaction system without NMP. Thus, we
can deduce that the inhibitory effect of NMP on humin deposi-
tion on the catalyst surface seems to explain the observed sta-
bility of the catalyst activity.
Catalyst recyclability
HMF production under continuous flow conditions
To minimize cost and environmental impact of catalytic indus-
trial processes, it is desirable that the catalyst is stable and can
be easily recycled once the reaction ends. Recycling efficiency
of P-TiO2 catalyst was investigated using the conversion of glu-
To further improve the attractiveness of HMF production and
simulate an industrial scenario, we investigated the catalytic
performance of phosphated TiO2 for the conversion of cellu-
ChemSusChem 2015, 8, 2907 – 2916
www.chemsuschem.org
2912
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
lose in a biphasic continuous flow reactor. The reactor system
consisted of a U-shaped stainless steel tubular reactor, two
HPLC pumps (Alltech 426), heating oil bath, back pressure reg-
ulator, and a reservoir to collect the reaction product. A sche-
matic diagram of the reactor setup is shown in Figure 6.
The biphasic system consists of a sugar solution (aqueous
solution of the soluble cello-oligomers obtained from treated
cellulose) and an organic solvent mixture [methyl isobutyl
ketone (MIBK)/NMP 3:1 v/v]. The choice of MIBK in replacing
THF as the organic extracting solvent is based on its very low
Figure 7. Effect of the reaction parameters (a) LHSV, (b) reaction tempera-
ture, (c) catalyst loading amount on the production of HMF in a flow reactor
using P-TiO2 as catalyst. Reaction conditions: 5 g cello-oligomer in 100 mL
water, 3:1 MIBK/NMP v/v ratio , 1:1 water/MIBK+NMP v/v ratio, and 60 bar
back pressure.
Figure 6. Schematic diagram of the flow reactor setup for the conversion of
cello-oligomers to HMF.
miscibility with water even at high reaction temperature.[18]
The two feed streams were delivered by the HPLC pumps, and
then connected by a ‘T’ connection to form a single premixed
stream prior to entering the flow reactor. A back pressure reg-
ulator was connected to the reactor outlet to control the pres-
sure of the reaction system. Under a steady flow rate at 60 bar
(back pressure), the preloaded reactor with P-TiO2 catalyst
(400 mg) was immersed in the heating oil bath, which was
maintained at the reaction temperature. A short induction
period was noticed, which may be attributable to back mixing
of liquid in the pressure regulator. A factor used to measure
the efficiency of the reactor is liquid hourly space velocity
(LHSV). Herein, it is defined as the ratio of the volumetric flow
rate of the feed solution (in mLhÀ1) to the heated reactor
volume (in mL).
time on stream. This suggests that a higher temperature accel-
erates glycosidic bond cleavage of the cello-oligomers, and
thus reactivity is enhanced. However, a further increase of the
reaction temperature to 2308C was detrimental for HMF yield
as shown in Figure 7b. This may be due to an unwanted reac-
tion producing humins. Another factor that may enhance reac-
tion performance is catalyst loading. Under a fixed LHSV of
18.9 hÀ1 and reaction temperature of 2208C, we studied the
effect of catalyst loading amount between 350 and 450 mg.
Based on the results illustrated in Figure 7c, we observed HMF
yield to increase with an increase in catalyst loading from 350
to 400 mg. The enhanced productivity can be ascribed to
more active sites being accessible by the reactant. With further
increase in catalyst loading, the yield of HMF declined. We can
deduce that catalyst loading of 400 mg provides the required
acid sites for the multi-step reactions in the conversion of the
cello-oligomers to HMF. Above 400 mg, excessive acid sites can
promote degradation of HMF, resulting in the decline of HMF
yield. In spite of a considerable activity decay as the reaction
progressed beyond 40 min time on stream, which may be due
to glucose fractions of the cello-oligomers turning into char or
humins, we were able to demonstrate the feasibility of the
continuous production of HMF from cellulose in a relatively
good yield. Further optimization of reactor configuration and
process conditions may effectively enhance the application of
this biphasic continuous reactor for the production of HMF.
The role of LHSV on the catalytic transformation of the cello-
oligomers was investigated. The different LHSVs were obtained
by changing the total flow rates of the feed streams between
0.2 and 0.4 mLminÀ1, which corresponds to a LHSV of 12.6–
25.2 hÀ1. The flow rate of the two feed streams was kept at
a ratio 1:1. As shown in Figure 7a, an initial experiment at
2208C showed that a maximum HMF yield of 53% could be
reached at a total flow rate of 0.3 mLminÀ1, corresponding to
a LHSV of 18.6 hÀ1. At higher LHSVs, the yield of HMF started
to decline, which is attributable to shorter residence time
within the reactor and, hence, reduced contact time of the
sugar substrate with the catalyst.
The effect of reaction temperature was also investigated,
and the experiments were carried out between 210–2308C at
a fixed LHSV of 18.9 hÀ1. In Figure 7b, we observed that the
temperature has a significant effect on the HMF yield. When
the temperature rose from 210 to 2208C, the HMF yield in-
creased significantly from about 40% to 53% within 60 min
ChemSusChem 2015, 8, 2907 – 2916
www.chemsuschem.org
2913
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
from Elga ultra pure apparatus was used for catalyst synthesis and
reaction.
Conclusions
TiO2 and modified TiO2 nanoparticles were prepared according to
our previously developed method.[19] Typically, titanium hydroxide
hydrate was prepared by the neutral amine sol–gel technique
using titanium(IV) butoxide as the TiO2 precursor. Excess solvent
was evaporated and the sample was dried at 808C overnight.
Phosphating was accomplished by treating the titanium hydroxide
hydrate with ammonium phosphate monobasic. Titanium hydrox-
ide was slurred in a minimum amount of aqueous ammonium
phosphate solution. The mixture was stirred for 4 h before the re-
moval of excess water by evaporation and dried at 808C overnight.
TiO2 containing 15 wt% phosphate was employed in this study.
Addition of molybdate, vanadate, and tungstate to TiO2 was car-
ried out in a similar procedure outlined above using ammonium
molybdate, ammonium metavanadate, and ammonium metatung-
state, respectively. Samples were then calcined at 6008C for 4 h.
Catalyst samples were designated as x-TiO2, where x represents P,
Mo, V, and W heteroatoms.
We described a batch and continuous process of catalytic con-
version of sugars to HMF in a biphasic system. The effective-
ness of the catalytic process to selective formation of HMF was
enhanced by adding N-methyl-2-pyrrolidone (NMP) to the re-
action medium. This effect was the consequence of the sup-
pression of undesired polymerization of HMF to humins.
Phosphated TiO2 was found to be an efficient and versatile
solid acid catalyst in the selective conversion of a variety of
sugars towards HMF formation. The reaction system with
water–THF+NMP medium and P-TiO2 as catalyst operated as
a batch reaction process afforded fructose and glucose conver-
sion efficiencies up to 98% and 90% HMF yields. Furthermore,
cellobiose and sucrose conversions achieved 94% and 98%
HMF yields, respectively. The minor difference in their reactivity
was ascribed to differences in the dehydration rate of the
monomeric units of the dimers. Sucrose contains fructose,
which is more reactive than glucose. Similarly, a higher HMF
yield (80–85%) could be obtained from starch (rice, potato)
than from cellulose (33%), which was mainly attributed to the
hydrolysis rate as glucose units of cellulose are strongly linked
by b-1,4-glycosidic bonds. Mechanocatalytic depolymerization
of cellulose was used as an efficient pretreatment process for
the production of soluble cello-oligomers that can be easily hy-
drolyzed to glucose units, subsequently achieving 86% HMF
yield. A catalyst recyclability study showed that the P-TiO2 cata-
lyst could be easily recovered and that reproducible and stable
activity was achieved.
Catalyst characterization
Structural analysis of the samples were characterized by X-ray dif-
fraction (XRD) using a Rigaku Miniflex diffractometer with a filtered
monochromatic CoKa radiation. The diffraction patterns were col-
lected in the range of 108ꢂ2qꢂ908 with a step size of 0.02. Spe-
cific surface area was determined by carrying out N2 adsorption at
À1968C using a Micromeritics TriStar II 3020 surface area and po-
rosity analyzer. Prior to analysis, the samples were outgassed at
2008C for at least 8 h under vacuum to remove the surface ad-
sorbed species. The surface area was calculated by using the Brun-
auer–Emmett–Teller (BET) method. Total pore volume was estimat-
ed using the volume of N2 gas adsorbed at a relative pressure (P/
P0) of 0.99. The morphology of the synthesized particles was inves-
tigated by transmission electron microscopy (TEM, JEOL JSM 2100),
operated at an acceleration voltage of 200 kV. Pyridine infrared
spectroscopy analysis was used to identify the nature of surface
acid sites. Prior to pyridine adsorption, the catalysts were activated
at 2008C under vacuum for 1 h and then cooled to 1508C. Pyridine
was then added to the system to saturate the exposed catalyst sur-
face (50 mg, 25 mm thickness). Chemisorption of pyridine was
maintained at 1508C for 30 min. Gaseous and physisorbed pyridine
were then evacuated under a N2 flow at 1508C for another 30 min.
FTIR spectra of the samples were recorded at room temperature
using a Nicolet 6700 (Smart Orbit Accessory). The concentration of
Brønsted and Lewis acid sites were estimated using the Lambert–
Beer Law in the form C=A/(e1), where C is the concentration of
the vibrating species (mmolgÀ1), A is the intensity of the band
(cmÀ1), e is the integration extinction coefficient (cmmmolÀ1), and
1 is the sample thickness (gcmÀ2).[30] Values of 1.67 cmmmolÀ1 and
2.22 cmmmolÀ1 were used as the integrated molar extinction coeffi-
cients for pyridine bands at 1545 (PyB) and 1455 cmÀ1 (PyL), re-
spectively.[31] A CHNS-O elemental analyzer (FLASH EA 1112 series,
Thermo Electron Corporation) was used to analyze the carbon con-
tent of fresh and spent catalyst samples. 2–3 mg of each sample
was placed in a tin container, which was combusted in a furnace
at 9008C. The gaseous products were separated chromatographi-
cally and analyzed using a thermal conductivity detector (TCD).
A flow reactor system was also used to demonstrate the ca-
pability of a continuous production of HMF. Utilizing the solu-
ble oligomers obtained from pretreated cellulose in a water–
MIBK+NMP biphasic system and P-TiO2 as catalyst, a reasona-
bly good yield of HMF (53%) was obtained. Thus, the pretreat-
ment of cellulose to give soluble oligomers appears to be ad-
vantageous and applicable for the continuous production of
HMF using a flow reactor. Hence, this approach is amenable to
direct transformation of real biomass for production of HMF in
scalable quantities.
Experimental Section
Materials and catalyst preparation
The following chemicals were used: glucose (ꢁ99.5%, Sigma–Al-
drich), fructose (99%, Sigma–Aldrich), cellobiose (ꢁ98%, Sigma–
Aldrich), cellulose (Sigmacell Type 20, 20 mm), starch-rice (Sigma–
Aldrich), starch-potato (Sigma–Aldrich), 5-hydroxymethyfurfural
(ꢁ99%, Sigma–Aldrich), titanium(IV) butoxide (97%, Sigma–Al-
drich), n-butanol (ꢁ99.4%, Sigma–Aldrich), ammonium phosphate
monobasic (ꢁ98%, Sigma–Aldrich), ammonium molybdate tetra-
hydrate (81–83% MoO3 basis, Sigma–Aldrich), ammonium metava-
nadate (ꢁ99%, Sigma–Aldrich), ammonium metatungstate hydrate
(99.99%, Sigma–Aldrich), aqueous ammonia solution (28 wt%,
Sigma–Aldrich), tetrahydrofuran (99.9%, Merck), methyl isobutyl
ketone (99.5%, Sigma–Aldrich), N-methyl-2-pyrrolidone (99.5%,
Sigma–Aldrich), acetonitrile (99.9%, Merck), toluene (99.9%,
Mechanocatalytic depolymerization of cellulose
Water-soluble, cellulose-based oligomers were produced by using
methods described elsewhere.[15c] In a typical method, H2SO4
Merck), and acetone (99.9%, Merck). Ultra pure water (18 MWcmÀ1
)
ChemSusChem 2015, 8, 2907 – 2916
www.chemsuschem.org
2914
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ꢁ Full Papers
ꢀ
(2.5 mmol) was diluted to a volume of 40 mL. Sigmacell microcrys-
talline cellulose (10 g) was then added to this solution, and the so-
lution was stirred for a few minutes. The resulting slurry was dried
using a rotary evaporator, followed by drying overnight in air at
508C. The acidified cellulose powder thus obtained was then
milled in a planetary ball mill using 5 mm stainless steel balls, with
a cellulose-to-ball weight ratio of 1:10. The mill was operated at
300 rpm, with a 20 min pause after every 15 min of continuous
milling. The pause allowed dissipation of heat generated during
milling, which prevented overheating of reactants. The milling time
reported refers only to the active milling time.
ni
nC60
Product yield ½mol%Š ¼
 100%
ð2Þ
0
where nC6 and nC6 denote number of moles of C6 sugars in the
product and feed, respectively, and ni is the number of moles of
identified products (HMF, levulinic acid, levoglucosan, etc.).
Acknowledgements
The authors gratefully acknowledge the financial support of the
Nanomaterials center at The University of Queensland, Australia.
The authors also acknowledge the facilities, scientific, and techni-
cal assistance of the Australian Microscopy and Microanalysis Re-
search Facility at The University of Queensland. The first author
sincerely appreciates the sponsorship from the International Post-
graduate Research Scholarship (IPRS) and UQ Centennial Scholar-
ship (UQ Cent).
Catalytic reactions
Batch transformation reaction of sugars to HMF was carried out in
a two-phase reaction system consisting of water/THF (1:4 v/v). In
a typical experimental run, substrate (5 g), catalyst (1.25 g), and sol-
vent (100 mL) were charged into a 300 mL reactor vessel provided
by Parr Instrument Company. NaCl (4 g) was added to the reaction
medium to maintain a biphasic reaction medium as well as im-
prove the efficiency of HMF extraction by the organic layer. The re-
actor was purged with Ar (99.9%) and then pressurized to 20 bar.
The temperature and stirring were controlled by a 4843 Controller
provided by Parr. Temperature in the reactor was monitored by
a thermocouple in the solution, and a constant stirring rate of
500 rpm was used for the reaction. After the reaction was com-
plete, the product mix was collected, centrifuged, and the superna-
tant was collected for analysis. Catalytic runs were repeated with
an experimental error of Æ2%.
Keywords: biphasic systems
· heterogeneous catalysis ·
mesoporous materials · solvents · sugars
[1] a) J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. Int. Ed. 2007,
46, 7164–7183; Angew. Chem. 2007, 119, 7298–7318; b) J. J. Bozell,
G. R. Petersen, Green Chem. 2010, 12, 539–554.
[2] O. Casanova, S. Iborra, A. Corma, ChemSusChem 2009, 2, 1138–1144.
[3] a) C. Moreau, M. Belgacem, A. Gandini, Top. Catal. 2004, 27, 11–30;
b) A. Gandini, A. J. D. Silvestre, C. P. Neto, A. F. Sousa, M. Gomes, J.
Polym. Sci. Part A 2009, 47, 295–298.
Continuous flow reactions were performed on a custom-built
fixed-bed reactor system. Two Alltech HPLC pumps were used for
feeding the sugar solution and the organic solvent into the fixed-
bed reactor. A U-shaped fixed-bed reactor was designed by using
1/4 in (1 in=2.54 cm) OD SS316 Swagelok tubing and tube fittings.
The powdered P-TiO2 was used as is without any pelletisation. The
catalyst was loaded into the reactor, and a small amount of quartz
wool was inserted from both sides of the reactor to hold the cata-
lyst in place. A 7 mm Swagelok inline filter was connected to the re-
actor exit, which was followed by a Swagelok back pressure regula-
tor. A steady flow was established through the system until the de-
sired pressure was achieved. The reactor was then dipped slowly
into a stirred oil bath set at the reaction temperature, and the reac-
tion time was noted as 0 min. Reaction conditions were typically
5 g cello-oligomer dissolved in 100 mL water, 3:1 MIBK/NMP v/v
ratio, 1:1 water/MIBK+NMP v/v ratio, and 60 bar back pressure
while the flow rate, reaction temperature, and catalyst dosage
[4] a) D. M. Alonso, S. G. Wettstein, J. A. Dumesic, Green Chem. 2013, 15,
584–595; b) I. T. Horvµth, H. Mehdi, V. Fabos, L. Boda, L. T. Mika, Green
Chem. 2008, 10, 238–242; c) . Bereczky, K. Lukµcs, M. Farkas, S. DóbØ,
Nat. Resour. 2014, 5, 177–191.
[5] Y. Romµn-Leshkov, C. J. Barrett, Z. Y. Liu, J. A. Dumesic, Nature 2007, 447,
982–985.
[6] J. B. Binder, R. T. Raines, J. Am. Chem. Soc. 2009, 131, 1979–1985.
[7] X. Tong, Y. Ma, Y. Li, Appl. Catal. A 2010, 385, 1–13.
[8] a) G. Yong, Y. Zhang, J. Y. Ying, Angew. Chem. Int. Ed. 2008, 47, 9345–
9348; Angew. Chem. 2008, 120, 9485–9488; b) R. Huang, W. Qi, R. Su, Z.
He, Chem. Commun. 2010, 46, 1115–1117; c) J. N. Chheda, Y. Roman-
Leshkov, J. A. Dumesic, Green Chem. 2007, 9, 342–350; d) H. Zhao, J. E.
Holladay, H. Brown, Z. C. Zhang, Science 2007, 316, 1597–1600; e) Y.
Romµn-Leshkov, J. Dumesic, Top. Catal. 2009, 52, 297–303; f) J. Wang,
W. Xu, J. Ren, X. Liu, G. Lu, Y. Wang, Green Chem. 2011, 13, 2678–2681;
g) Z. Zhang, Q. Wang, H. Xie, W. Liu, Z. Zhao, ChemSusChem 2011, 4,
131–138; h) X. Qi, M. Watanabe, T. M. Aida, J. R. L. Smith, Green Chem.
2008, 10, 799–805.
[9] a) M. Stçcker, Angew. Chem. Int. Ed. 2008, 47, 9200–9211; Angew. Chem.
2008, 120, 9340–9351; b) P. L. Dhepe, A. Fukuoka, ChemSusChem 2008,
1, 969–975.
[10] a) Y. Nishiyama, P. Langan, H. Chanzy, J. Am. Chem. Soc. 2002, 124,
9074–9082; b) Y. Nishiyama, J. Sugiyama, H. Chanzy, P. Langan, J. Am.
Chem. Soc. 2003, 125, 14300–14306; c) D. Klemm, B. Heublein, H.-P.
Fink, A. Bohn, Angew. Chem. Int. Ed. 2005, 44, 3358–3393; Angew.
Chem. 2005, 117, 3422–3458.
[11] a) Y. Su, H. M. Brown, G. Li, X.-d. Zhou, J. E. Amonette, J. L. Fulton, D. M.
Camaioni, Z. C. Zhang, Appl. Catal. A 2011, 391, 436–442; b) Y. Zhang,
H. Du, X. Qian, E. Y. X. Chen, Energy Fuels 2010, 24, 2410–2417; c) E.
Ohno, H. Miyafuji, J. Wood Sci. 2014, 60, 428–437; d) M. P. Foley, L. M.
Haverhals, D. K. Klein, W. B. McIlvain, W. M. Reichert, D. W. O’Sullivan,
H. C. De Long, P. C. Trulove, Am. J. Biomass Bioenergy 2014, 3, 68–82;
e) R. P. Swatloski, S. K. Spear, J. D. Holbrey, R. D. Rogers, J. Am. Chem.
Soc. 2002, 124, 4974–4975; f) Y. Cao, J. Wu, J. Zhang, H. Li, Y. Zhang, J.
He, Chem. Eng. J. 2009, 147, 13–21; g) O. A. El Seoud, A. Koschella, L. C.
Fidale, S. Dorn, T. Heinze, Biomacromolecules 2007, 8, 2629–2647.
were varied between 0.2–0.4 mLminÀ1
450 mg, respectively.
, 210–2308C, and 350–
Liquid products were analyzed using a Shimadzu Prominence
HPLC equipped with both refractive index (RID-10) and UV/Vis
(SPD-M20A) detectors using a Bio-Rad Aminex HPX-87H as the ana-
lytical column. The column was operated at 508C using 5 mm
H2SO4 as the mobile phase at a flow rate of 0.6 mLminÀ1 for the
analysis of both the aqueous and organic phases. The concentra-
tions of glucose, fructose, HMF, and other identifiable products
were quantified by means of HPLC analysis through the external
standard method and calibration curves of commercially available
standard substrates. Sugar conversion (Conv.) and product yield
were calculated according to Equations (1) and (2):
ꢀ
ꢁ
nC6
nC60
Conv: ½mol%Š ¼ 1 À
 100%
ð1Þ
ChemSusChem 2015, 8, 2907 – 2916
www.chemsuschem.org
2915
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
´
[12] a) Z. Fang, J. A. Kozinski, Proc. Combust. Inst. 2000, 28, 2717–2725;
[20] a) A. Chareonlimkun, V. Champreda, A. Shotipruk, N. Laosiripojana, Fuel
2010, 89, 2873–2880; b) S. De, S. Dutta, A. K. Patra, A. Bhaumik, B. Saha,
J. Mater. Chem. 2011, 21, 17505–17510; c) S. De, S. Dutta, A. K. Patra,
B. S. Rana, A. K. Sinha, B. Saha, A. Bhaumik, Appl. Catal. A 2012, 435–
436, 197–203; d) S. Dutta, S. De, A. K. Patra, M. Sasidharan, A. Bhaumik,
B. Saha, Appl. Catal. A 2011, 409–410, 133–139; e) A. Dutta, A. K. Patra,
S. Dutta, B. Saha, A. Bhaumik, J. Mater. Chem. 2012, 22, 14094–14100;
f) C.-H. Kuo, A. S. Poyraz, L. Jin, Y. Meng, L. Pahalagedara, S.-Y. Chen,
D. A. Kriz, C. Guild, A. Gudz, S. L. Suib, Green Chem. 2014, 16, 785–791;
g) K. Nakajima, R. Noma, M. Kitano, M. Hara, J. Mol. Catal. A 2014, 388–
389, 100–105.
[21] M. A. BaÇares, L. J. Alemany, M. C. JimØnez, M. A. Larrubia, F. Delgado,
M. LopØz Granados, A. Martínez-Arias, J. M. Blasco, J. L. G. Fierro, J. Solid
State Chem. 1996, 124, 69–76.
[22] E. C. Bucharsky, G. Schell, R. Oberacker, M. J. Hoffmann, J. Eur. Ceram.
Soc. 2009, 29, 1955–1961.
b) L. K. Tolonen, M. Juvonen, K. Niemelä, A. Mikkelson, M. Tenkanen, H.
Sixta, Carbohydr. Res. 2015, 401, 16–23; c) S. Kumar, R. B. Gupta, Ind.
Eng. Chem. Res. 2008, 47, 9321–9329; d) M. Sasaki, Z. Fang, Y. Fukushi-
ma, T. Adschiri, K. Arai, Ind. Eng. Chem. Res. 2000, 39, 2883–2890; e) Y.
Ishikawa, S. Saka, Cellulose 2001, 8, 189–195; f) A. Koriakin, H. Van N-
guyen, D. W. Kim, C.-H. Lee, J. Supercrit. Fluids 2014, 95, 175–186;
g) F. S. Asghari, H. Yoshida, Carbohydr. Res. 2010, 345, 124–131.
[13] D. Klein-Marcuschamer, B. A. Simmons, H. W. Blanch, Biofuels Bioprod. Bi-
orefin. 2011, 5, 562–569.
[14] K. T. Tan, K. T. Lee, Renewable Sustainable Energy Rev. 2011, 15, 2452–
2456.
[15] a) S. M. Hick, C. Griebel, D. T. Restrepo, J. H. Truitt, E. J. Buker, C. Bylda,
R. G. Blair, Green Chem. 2010, 12, 468–474; b) N. Meine, R. Rinaldi, F.
Schüth, ChemSusChem 2012, 5, 1449–1454; c) A. Shrotri, L. K. Lambert,
A. Tanksale, J. Beltramini, Green Chem. 2013, 15, 2761–2768; d) R. Carra-
squillo-Flores, M. Käldstrçm, F. Schüth, J. A. Dumesic, R. Rinaldi, ACS
Catal. 2013, 3, 993–997; e) J. Hilgert, N. Meine, R. Rinaldi, F. Schüth,
Energy Environ. Sci. 2013, 6, 92–96; f) M. Käldstrçm, N. Meine, C. Fares,
F. Schüth, R. Rinaldi, Green Chem. 2014, 16, 3528–3538; g) P. Dornath,
H. J. Cho, A. Paulsen, P. Dauenhauer, W. Fan, Green Chem. 2015, 17,
769–775.
[23] A. K. Dalai, R. Sethuraman, S. P. R. Katikaneni, R. O. Idem, Ind. Eng. Chem.
Res. 1998, 37, 3869–3878.
[24] L. E. Depero, P. Bonzi, M. Musci, C. Casale, J. Solid State Chem. 1994, 111,
247–252.
[25] S. Damyanova, M. L. Cubeiro, J. L. G. Fierro, J. Mol. Catal. A 1999, 142,
85–100.
[16] a) W.-H. Peng, Y.-Y. Lee, C. Wu, K. C. W. Wu, J. Mater. Chem. 2012, 22,
23181–23185; b) Y. Su, H. M. Brown, X. Huang, X.-d. Zhou, J. E. Amo-
nette, Z. C. Zhang, Appl. Catal. A 2009, 361, 117–122; c) P. Wang, H. Yu,
S. Zhan, S. Wang, Bioresour. Technol. 2011, 102, 4179–4183; d) N. Shi, Q.
Liu, Q. Zhang, T. Wang, L. Ma, Green Chem. 2013, 15, 1967–1974; e) K. Y.
Nandiwale, N. D. Galande, P. Thakur, S. D. Sawant, V. P. Zambre, V. V.
Bokade, ACS Sustainable Chem. Eng. 2014, 2, 1928–1932; f) M. Mascal,
E. B. Nikitin, Green Chem. 2010, 12, 370–373.
[26] B. F. M. Kuster, Starch/Staerke 1990, 42, 314–321.
[27] G. Centi, Appl. Catal. A 1996, 147, 267–298.
[28] Y. Romµn-Leshkov, J. N. Chheda, J. A. Dumesic, Science 2006, 312, 1933–
1937.
ˇ
´
´
[29] a) J. Horvat, B. Klaic, B. Metelko, V. Sunjic, Tetrahedron Lett. 1985, 26,
2111 –2114; b) J. Horvat, B. Klaic, B. Metelko, V. Sunjic, Croat. Chem. Acta
1986, 59, 429–438.
[30] A. A. Budneva, E. A. Paukshtis, A. A. Davydov, React. Kinet. Catal. Lett.
1987, 34, 63–67.
[31] C. A. Emeis, J. Catal. 1993, 141, 347–354.
[17] a) M. Mascal, E. B. Nikitin, Angew. Chem. Int. Ed. 2008, 47, 7924–7926;
Angew. Chem. 2008, 120, 8042–8044; b) M. Mascal, E. B. Nikitin, Chem-
SusChem 2009, 2, 859–861.
[18] C. V. McNeff, D. T. Nowlan, L. C. McNeff, B. Yan, R. L. Fedie, Appl. Catal. A
2010, 384, 65–69.
[19] L. Atanda, S. Mukundan, A. Shrotri, Q. Ma, J. Beltramini, ChemCatChem
2015, 7, 781–790.
Received: March 19, 2015
Revised: May 14, 2015
Published online on August 3, 2015
ChemSusChem 2015, 8, 2907 – 2916
www.chemsuschem.org
2916
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim