Selective glucose transformation by titania as a heterogeneous Lewis acid catalyst

Article abstract of DOI:10.1016/j.molcata.2013.09.012

The Lewis acidity of phosphate-immobilized anatase TiO₂ (phosphate/TiO₂) has been studied to develop novel environmentally benign reaction systems. Fourier transform infrared (FT-IR) measurements suggested that most Lewis acid sites on bare and phosphate/TiO₂ surface function even in water. phosphate/TiO₂ exhibits high catalytic performance for selective 5-(hydroxymethyl)furfural (HMF) production from glucose in THF/water (90/10 vol.%) solution. This is attributed to water-tolerant Lewis acid sites on phosphate/TiO₂ that promote step-wise conversion of glucose into HMF. The catalyst was easily recovered from reaction solution by simple decantation or filtration, and can be used repeatedly without significant loss of original activity for subsequent reactions.

Full text of DOI:10.1016/j.molcata.2013.09.012

GModel
MOLCAA-8866; No. of Pages6
ARTICLE IN PRESS
Journal of Molecular Catalysis A: Chemical xxx (2013) xxx–xxx
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Selective glucose transformation by titania as a heterogeneous
Lewis acid catalyst
Kiyotaka Nakajimaa,b, Ryouhei Nomaa, Masaaki Kitanoc, Michikazu Haraa,d,∗
a Materials and Structures Laboratory, Tokyo Institute of Technology, Nagatsuta-cho 4259-R3-33, Midori-ku, Yokohama 226-8503, Japan
b Japan Science and Technology (JST) Agency, PRESTO, 4-1-8 Honcho, Kawaguchi 332-0012, Japan
c Materials Research Center for Element Strategy, Tokyo Institute of Technology, Nagatsuta-cho 4259-S2-16, Midori-ku, Yokohama 226-8503, Japan
d Japan Science and Technology (JST) Agency, ALCA, 4-1-8 Honcho, Kawaguchi 332-0012, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 May 2013
Received in revised form 8 August 2013
Accepted 11 September 2013
Available online xxx
The Lewis acidity of phosphate-immobilized anatase TiO2 (phosphate/TiO2) has been studied to develop
novel environmentally benign reaction systems. Fourier transform infrared (FT-IR) measurements
suggested that most Lewis acid sites on bare and phosphate/TiO2 surface function even in water.
phosphate/TiO2 exhibits high catalytic performance for selective 5-(hydroxymethyl)furfural (HMF) pro-
duction from glucose in THF/water (90/10 vol.%) solution. This is attributed to water-tolerant Lewis acid
sites on phosphate/TiO2 that promote step-wise conversion of glucose into HMF. The catalyst was easily
recovered from reaction solution by simple decantation or filtration, and can be used repeatedly without
significant loss of original activity for subsequent reactions.
Keywords:
Lewis acid
Biomass conversion
Glucose
© 2013 Elsevier B.V. All rights reserved.
HMF
1. Introduction
remain for this catalytic system with respect to the effective use of
glucose and energy costs, the results of this reaction system sug-
The effective conversion of glucose, a key component of cel-
lulosic biomass, into HMF is an attractive route to sustainable
chemical production [1]. HMF is a versatile and key platform chem-
ical because it can be further converted into various polymers such
as polyesters and polyamides, and pharmaceuticals [1,2]. However,
the lack of a simple and highly efficient process for HMF produc-
tion from glucose has been an obstacle to the utilization of HMF, a
so-called “sleeping giant” [2]. A proposed reaction mechanism for
HMF formation from glucose is shown in Scheme 1, where glucose
derived from cellulosic biomass or starch is converted into HMF
through isomerization, followed by dehydration in the presence of
appropriate catalysts [2]. Zhao and co-workers demonstrated that
CrCl2, a soluble homogeneous Lewis acid catalyst, dissolved in an
ionic liquid functions as the most efficient catalytic system for the
reaction, although the maximum HMF yield in the reaction system
is still only in the order of 60% [3]. While serious problems, such as
HMF selectivity, separation of the catalyst and HMF from the ionic
liquid, reuse of the catalyst, and handling in practical processes
gest that Lewis acids function as effective catalysts for the efficient
production of HMF from glucose.
Lewis acid catalysts such as AlCl3, BF3, and transition metal
halides are essential for the production of industrially important
chemicals, including polymers and pharmaceuticals [4]. However,
most Lewis acids decompose or are ineffective in the presence of
water, and thereby require dehydrated environment, which results
in significant energy consumption. In addition, they have serious
drawbacks, such as the production of waste and corrosion of equip-
ment, in addition to requiring separation from the product [5].
Although a few exceptions (rare earth metal triflate complexes [6,7]
and Sn4+-incorporated zeolites [8,9]) are known to exhibit Lewis
acid catalysis in water, the scarcity of rare earth metals in the former
and the narrow reaction space available in the latter limit their
utility. Thus, heterogeneous Lewis acid catalysts that are ubiqui-
tous, insoluble, easily separable from products, and highly active for
various reactions in water would be applicable to environmentally
benign chemical production.
Recently, we reported that a part of NbO4 tetrahedra present
in insoluble niobic acid (Nb2O5·nH2O) act as Lewis acids even in
water [10]. Many metal oxides of groups 4 and 5, including niobic
acid, are composed of octahedral MO6 (M: metal) units with satu-
rated coordination spheres, and polyhedral MOx with unsaturated
coordination spheres, such as tetrahedral MO4, are also present on
∗
Corresponding author at: Materials and Structures Laboratory, Tokyo Institute
of Technology, Nagatsuta-cho 4259-R3-33, Midori-ku, Yokohama 226-8503, Japan.
Tel.: +81 45 924 5311; fax: +81 45 924 5381.
E-mail address: mhara@msl.titech.ac.jp (M. Hara).
1381-1169/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2013.09.012
Please cite this article in press as: K. Nakajima, et al., J. Mol. Catal. A: Chem. (2013), http://dx.doi.org/10.1016/j.molcata.2013.09.012

GModel
MOLCAA-8866; No. of Pages6
ARTICLE IN PRESS
2
K. Nakajima et al. / Journal of Molecular Catalysis A: Chemical xxx (2013) xxx–xxx
Scheme 1. Reaction pathway for acid-catalysed conversion of glucose into HMF in water.
the surface. Unsaturated coordination MO4 tetrahedra act as Lewis
acids; however, MO4 species are considered to not function as well
in water as other Lewis acids. Niobic acid with NbO4 species as
water-tolerant Lewis acid sites suggests that anatase TiO2, a ubiqui-
tous material, with TiO4 species on the surface would also function
as an insoluble, easily separable, and water-tolerant Lewis acid cat-
alyst. Therefore, the potential of anatase TiO2 as a heterogeneous
water-tolerant Lewis acid was investigated in this study.
2.3. HMF production from glucose
THF/aqueous solution (2.0 mL (THF, 1.8 mL; distilled water,
0.2 mL)) containing d-glucose (0.02 g) and catalyst (0.05 g) was
heated in a sealed Pyrex tube for 2 h at 393 K. After filtration, the
solutions were analyzed using high performance liquid chromatog-
raphy (HPLC; LC-2000 plus, Jasco) equipped with refractive index
(RI) and photodiode array (PDA) detectors. Aminex® HPH-87H col-
umn (300 mm × 7.8 mm, Bio-Rad Laboratories) with diluted H2SO4
solution (5 mM) of eluent, 0.5 mL min−1 of flow rate, and 308 K of
column temperature was adopted in HPLC analysis.
2. Experimental
2.1. Preparation of anatase TiO2 and phosphate/TiO2
3. Results and discussion
Anatase TiO2 was synthesized by the addition of 20 mL Ti(i-pro)4
to 100 mL distilled water, followed by stirring at room temperature.
After 6 h, the filtrated white precipitate was stirred in 200 mL of
1 M HCl solution for 2 h to complete the hydrolysis of residual Ti-
OCH(CH3)2 species. The obtained powder was repeatedly washed
with distilled water (ca. 1000 mL) until the pH of the filtrate became
neutral. The resulting material was dried overnight at 353 K and
then used as the anatase TiO2 catalyst.
Phosphate/TiO2 was prepared by immobilizing H3PO4 on
anatase TiO2. 5 g of TiO2 was stirred in 200 mL of 1 M H3PO4
solution. After stirring for 48 h, the collected sample was washed
repeatedly with distilled water until phosphate ions were no longer
detected. The resulting material was dried overnight at 353 K and
then used as the phosphate/TiO2 catalyst.
3.1. Structure of anatase TiO2 and phosphate/TiO2
Structural information for the anatase TiO2 and phosphate/TiO2
catalysts was obtained by XRD and N2 adsorption analyses. Fig. 1
shows XRD patterns and (B) N2 adsorption–desorption isotherms
for (a) anatase TiO2 and (b) phosphate/TiO2. Diffraction peaks due
to anatase TiO2 are evident in the XRD patterns for anatase TiO2
and phosphate/TiO2, which indicates that both samples are mainly
composed of anatase TiO2. There was no significant difference
in the XRD patterns of anatase TiO2 and phosphate/TiO2; there-
fore, phosphoric acid modification of anatase TiO2 does not change
the original anatase TiO2 structure. N2 adsorption–desorption
isotherms of the samples are similar to the type-IV pattern, which is
typical of mesoporous solids. The Brunauer–Emmett–Teller (BET)
surface areas and pore volumes of anatase TiO2 and phosphate/TiO2
were estimated to be 252 m2 g−1 and 0.31 mL g−1, and 266 m2 g−1
and 0.25 mL g−1, respectively. SEM images revealed that the pre-
pared TiO2 sample is composed of 10–20 nm TiO2 particles. There
was no significant difference in the morphology of TiO2 and
phosphate/TiO2. The amount of immobilized phosphate on anatase
TiO2 was estimated by inductively coupled plasma-atomic emis-
sion spectroscopy (ICP-AES), which revealed that 0.77 mmol of
phosphate ions were tightly fixed on 1 g of TiO2 by ester formation
between phosphoric acid and neutral OH groups.
2.2. FT-IR measurement and estimation of the amounts of Lewis
acid sites for anatase TiO2 and phosphate/TiO2
Lewis acid densities on anatase TiO2 and phosphate/TiO2 were
estimated for pyridine-adsorbed samples at 298 K by FT-IR mea-
surements. The samples were pressed into self-supporting disks
(20 mm diameter, ca. 20 mg) and placed in an IR cell attached
to a closed glass-circulation system (0.38 dm−3). The disk was
dehydrated by heating at 423 K for 1 h under vacuum to remove
physisorbed water and was exposed to pyridine vapor at 423 K.
The intensities of the bands at 1445 cm−1 (pyridine coordina-
tively bonded to Lewis acid sites, molecular absorption coefficient:
4.86 ␮mol cm−1) were plotted against the amounts of pyridine
adsorbed on the Lewis acid sites of the samples.
In the case of the sample in the presence of saturated water
vapor, the disk placed in the IR cell was exposed to saturated H2O
vapor (20–25 Torr) at room temperature for 60 min. 4.2 layers of
H2O molecules were adsorbed on the TiO2 and phosphate/TiO2
surfaces, as estimated from water vapor–adsorption–desorption
isotherms. Pyridine vapor was then added to the reaction system,
and the intensity of the 1445 cm−1 band (pyridine coordinatively
bonded to Lewis acid sites) increased with increasing amount of
introduced pyridine, reaching a plateau.
3.2. Lewis acid sites on anatase TiO2
Difference Fourier transform infrared (FT-IR) spectra for pyri-
dine adsorption on dehydrated anatase TiO2 and phosphate/TiO2
are shown in Fig. 2, where pyridine is employed as a basic probe
molecule for characterization of the acid sites [11]. Dehydrated
TiO2 exhibits several bands (Fig. 2(A)), but there is no signal
for pyridinium ions formed on Brønsted acid sites (1540 cm−1),
because TiO2 has no Brønsted acid sites. The intensities of
the two bands at 1445 and 1440 cm−1, which are assigned to
adsorbed pyridine on Lewis acid sites (TiO4) and physisorbed pyri-
dine [12], respectively, increase with the amount of introduced
Please cite this article in press as: K. Nakajima, et al., J. Mol. Catal. A: Chem. (2013), http://dx.doi.org/10.1016/j.molcata.2013.09.012

GModel
MOLCAA-8866; No. of Pages6
ARTICLE IN PRESS
K. Nakajima et al. / Journal of Molecular Catalysis A: Chemical xxx (2013) xxx–xxx
3
(A)
(B)
200
(a)
(b)
100
(b)
(a)
80
0
20
60
40
2 / degree
0.0 0.2 0.4 0.6 0.8 1.0
Relative pressure (P/P0)
Fig. 1. (A) XRD patterns and (B) N2 adsorption–desorption isotherms for (a) anatase TiO2 and (b) phosphate/TiO2.
pyridine, and the band intensities due to acid sites reach plateaus
at Ppyridine > 2.8 × 10−2 kPa. Fig. 2B shows difference FT-IR spectra
for pyridine-adsorbed anatase TiO2 and phosphate/TiO2. The sam-
ples were treated with saturated pyridine vapor and subsequently
evacuated at room temperature for 60 min to remove physisorbed
pyridine. Evacuation of the samples results in no change of the
band for pyridine adsorbed on Lewis acid sites, whereas the sig-
nal for physisorbed pyridine is eliminated. Thus, the maximum
intensity of the band at 1445 cm−1 corresponds to the total amount
of Lewis acid sites with adsorbed pyridine. The amounts of Lewis
acid sites on the dehydrated TiO2 and phosphate/TiO2 samples
were estimated to be 0.25and 0.23 mmol g−1, respectively, from the
maximum band intensities and molecular absorption coefficients
at 1445 cm−1. Fig. 3(A) and (B) shows FT-IR spectra for pyridine-
adsorbed TiO2 and phosphate/TiO2, respectively, in the presence
of saturated water vapor. Large amounts of physisorbed H2O
molecules are presented on the TiO2 and phosphate/TiO2 surface
prior to pyridine introduction under the experimental conditions;
therefore, two large and diffuse bands due to water are observed at
2500–3700 and 1500–1700 cm−1 in Fig. 3(A)(a) and (B)(d). Even in
the case of such water-saturated samples, the band due to pyridine
adsorbed on Lewis acid sites (1445 cm−1) appears in Fig. 3(A)(b) and
3(B)(e), in addition to pysisorbed pyridine (1440 cm−1). Although
the band for pysisorbed pyridine disappears after evacuation of
the samples at room temperature, the signal for Lewis acid sites
remains unchanged. The effective Lewis acid densities of TiO2
and phosphate/TiO2 in the presence of saturated water vapor
(water-tolerant Lewis acid density) were 0.24 and 0.22 mmol g−1
,
respectively. This implies that most unsaturated coordination TiO4
tetrahedra still act as Lewis acid sites, even in water, because water
does not deactivate these Lewis acid sites, and all unsaturated coor-
dination TiO4 tetrahedra still act as Lewis acid sites even in water
and the Lewis acid sites are not covered with phosphate ions on
phosphate/TiO2.
Fig. 2. FT-IR spectra for pyridine-adsorbed dehydrated TiO2 and phosphate/TiO2 at 298 K. (A) Difference FT-IR spectra for dehydrated TiO2. Prior to pyridine adsorption, the
TiO2 sample was heated at 473 K for 1 h under vacuum to remove physisorbed H2O. Gas phase pyridine pressure: (a) 1.4 × 10−2; (b) 6.7 × 10−2; (c) 1.3 × 10−1; (d) 2.8 × 10−2
;
(e) 6.7 × 10−1; (f) 1.4 kPa. (B) Difference FT-IR spectra for pyridine-adsorbed (g) TiO2 and (h) phosphate/TiO2. Prior to pyridine adsorption, the samples were heated at 473 K
for 1 h under vacuum to remove physisorbed H2O. After pyridine adsorption, the samples were evacuated at 298 K for 60 min.
Please cite this article in press as: K. Nakajima, et al., J. Mol. Catal. A: Chem. (2013), http://dx.doi.org/10.1016/j.molcata.2013.09.012

GModel
MOLCAA-8866; No. of Pages6
ARTICLE IN PRESS
4
K. Nakajima et al. / Journal of Molecular Catalysis A: Chemical xxx (2013) xxx–xxx
Fig. 3. Difference FT-IR spectra for (A) TiO2 (B) phosphate/TiO2 in the presence of saturated H2O vapor. (a) Hydrated TiO2 in saturated H2O vapor, (b) TiO2 obtained by the
introduction of pyridine (2.7 kPa) after H2O adsorption under H2O vapor, (c) TiO2 sample obtained by evacuation (room temperature) for 60 min after (b), (d) a hydrated
phosphate/TiO2 in saturated H2O vapor, (e) phosphate/TiO2 obtained by the introduction of pyridine (2.5 kPa) after H2O adsorption under H2O vapor, (f) phosphate/TiO2
obtained by evacuation (room temperature) for 60 min after (e).
3.3. Catalytic studies
gas chromatography–mass spectrometry (GC–MS) [10]. Anatase
TiO2 also converts most of the glucose into complex by-products,
including polymerized species, as with rare earth metal triflate
complex catalysts. In contrast to bare TiO2, TiO2 modified with
H3PO4 (phosphate/TiO2), where OH groups on TiO2 are esterified
into O-PO(OH)2 by phosphoric acid, exhibits significant catalytic
performance for HMF production; the HMF yield exceeded 80%
within 2 h. There is large difference between levulinic acid and
formic acid yield on phosphate/TiO2. One possible explanation is
adsorption of the evolved levulinic acid on phosphate/TiO2 [17].
After reaction for 2 h at 393 K, the phosphate/TiO2 particles could be
readily separated from the solution by decantation and the recov-
ery of the catalyst exceeded 99%. The results for reuse experiments
of this catalyst are presented in Fig. 4. The recovered catalyst was
repeatedly evaluated for HMF production from glucose for 2 h at
393 K. No decrease in activity was observed, even after 4 reuses
of the catalyst sample and the total turnover number reached ca.
40. It was also confirmed that the phosphate ions are not des-
orbed from the sample during reaction. Therefore, phosphate/TiO2
functions as a stable and highly selective heterogeneous catalyst
for simple HMF production from glucose. Fig. 5 shows the time
courses for the conversion of glucose into HMF on phosphate/TiO2.
The catalysis of anatase TiO2 was evaluated through the con-
version of glucose into HMF in water. Table 1 summarizes results
for the conversion of glucose into HMF in an aqueous solution
of tetrahydrofuran (THF) at 393 K. The efficiency for HMF pro-
duction from glucose has been reported to be improved by using
THF aqueous solution as a solvent [9]. Homogeneous (HCl and
H3PO4) and heterogeneous Brønsted acids (SO3H-bearing resins:
Nafion®NR50 and Amberlyst-15) that were also tested exhibit
moderate glucose conversion and HMF yield. Brønsted acids are
not effective for the conversion of glucose into HMF at such low
reaction temperatures [13], and the HMF produced is decomposed
by hydration in water without extraction by appropriate organic
solvents [13–15]. Although rare earth metal triflate complexes
(Sc(OTf)3 and Yb(OTf)3) exhibit high glucose conversion, most of
the reacted glucose is converted into unknown species. This can
be attributed to the formation of polymerized species [10,15,16].
Complex intermolecular side reactions, such as aldol condensa-
tion, among reducing saccharides with formyl groups ( CHO) in
the presence of acid catalysts result in the formation of complex
polymers as unknown species that cannot be detected by HPLC and
Please cite this article in press as: K. Nakajima, et al., J. Mol. Catal. A: Chem. (2013), http://dx.doi.org/10.1016/j.molcata.2013.09.012

GModel
MOLCAA-8866; No. of Pages6
ARTICLE IN PRESS
K. Nakajima et al. / Journal of Molecular Catalysis A: Chemical xxx (2013) xxx–xxx
5
Table 1
Catalytic activity for HMF production from glucose over acid catalysts.a
b
Catalyst
Aciddensity(mmol g−1
)
Conv. (%)d
Product yield (%)e
BAS
LAS
Fru
HMF
FA
0.0
0.0
0.0
0.0
0.0
0.0
0.0
10.5
LA
0.0
0.0
24.9
0.0
0.0
0.0
0.0
0.0
Unknown
HCl
H3PO4
9.9
10.2
4.8
0.9
–
–
–
–
0.0
–
–
56.2
9.2
0.0
0.0
2.0
0.0
1.5
0.0
0.0
0.0
29.4
1.2
12.0
3.1
13.9
10.6
8.5
26.9
8.0
Amberlyst-15
Nafion NR50
Sc(OTf)3
Yb(OTf)3
TiO2
69.9
28.7
>99.9
89.4
>99.9
98
31.0
25.6
84.6
78.8
91.5
8.3
–
2.0
1.9
0.24c
0.22c
phosphate/TiO2
81.2
a
Reagents and conditions: distilled water, 0.2 mL; THF, 1.8 mL; d-glucose, 0.02 g (0.11 mmol); catalyst weight, 0.05 g; temperature, 393 K; time, 2 h.
BAS, Brønsted acid sites; LAS, Lewis acid sites.
Water tolerant acid sites.
Glucose conversion for 2 h.
Fru, fructose; FA, formic acid; LA, levulinic acid; unknown, undetectable species.
b
c
d
e
Although a small amount of fructose is observed at the initial
also performed to clarify the recovery of the catalyst. THF/aqueous
solution (2.0 mL (THF, 1.8 mL; distilled water, 0.2 mL)) containing
d-glucose (0.2 g) and catalyst (0.5 g) was heated in a sealed Pyrex
tube at 393 K for 30 min. There was no difference in glucose con-
version and HMF yield between small and large-scale experiment,
and more than 99% of catalyst was successfully recovered by simple
decantation and filtration.
It should be noted that only water-tolerant Lewis acids on
phosphate/TiO2 efficiently catalyze the conversion of glucose into
HMF. Sc(OTf)3 catalyzes various reactions as an excellent water-
tolerant Lewis acid. However, only a small part of the reacted
glucose is converted into HMF on the catalyst. Therefore, Lewis
acid catalysis on this complex promotes intermolecular side reac-
tions, including aldol condensation among reducing saccharides,
in preference to the isomerization of glucose to fructose and the
intramolecular dehydration of fructose. Bare anatase TiO2 also
results in significant glucose conversion and low HMF selectivity
similar to Sc(OTf)3. As a result, highly efficient HMF production
on phosphate/TiO2 cannot be solely explained by only Lewis acid
sites workable in aqueous solution. Phosphoric acid modification
on anatase TiO2 causes a significant improvement in HMF produc-
tion by decreasing side reactions, and the Lewis acid sites are not
covered with phosphate ions, as the IR experimental results indi-
cated. This suggests a synergistic effect between Lewis acid sites
stage of the reaction (15–30 min), beyond 30 min, fructose cannot
be detected. This suggests that produced fructose is immediately
consumed for the subsequent reaction and glucose is converted
into HMF through isomerization, followed by dehydration even
on phosphate/TiO2 as the proposed reaction mechanism shown
in Scheme 1. Phosphoric acid itself cannot work as an effective
catalyst for the conversion of glucose as shown in Scheme 1: phos-
phoric acid has much smaller glucose conversion and HMF yield
than that bare TiO2 and phosphate/TiO2. Therefore, highly efficient
HMF production on phosphate/TiO2 may be due to catalysis of TiO2
Lewis acid sites that interact with reactants even in water. In order
to compare the rate of HMF formation from glucose and fructose
over phosphate/TiO2, HMF production from fructose was carried
out under same reaction condition. THF/aqueous solution (2.0 mL
(THF, 1.8 mL; distilled water, 0.2 mL)) containing d-fructose (0.02 g)
and catalyst (0.05 g) was heated in a sealed Pyrex tube at 393 K for
30 min. Rate of HMF formation from fructose was estimated to be
0.026 mmol g−1 min−1 which is slightly larger than that from glu-
cose (0.017 mmol g−1 min−1). This result also supports that glucose
is converted into HMF via fructose. Large-scale experiment was
100
Glucose
HMF
80
60
40
Formic acid
Fructose
60
20
0
0
30
90
120
Reaction time / min
Fig. 4. Catalytic activity of reused phosphate/TiO2 for HMF production from d-
glucose. Reagents and conditions: catalyst, 0.05 g; water: 0.2 mL; THF, 1.8 mL;
d-glucose: 0.02 g; temperature, 393 K.
Fig. 5. Time courses for the conversion of glucose into HMF on phosphate/TiO2.
Reagents and conditions: catalyst, 0.05 g; water: 0.2 mL; THF, 1.8 mL; d-glucose:
0.02 g; temperature, 393 K.
Please cite this article in press as: K. Nakajima, et al., J. Mol. Catal. A: Chem. (2013), http://dx.doi.org/10.1016/j.molcata.2013.09.012

GModel
MOLCAA-8866; No. of Pages6
ARTICLE IN PRESS
6
K. Nakajima et al. / Journal of Molecular Catalysis A: Chemical xxx (2013) xxx–xxx
and phosphate ions essentially changes the acid catalysis on TiO2.
The details are currently under investigation.
References
[1] J.N. Chhede, G.W. Huber, J.A. Dumesic, Angew. Chem. Int. Ed. 46 (2007)
7164–7183.
[2] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411–2502.
[3] H. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Science 316 (2007) 1597–1600.
[4] D. Schinzer, Selectivities in Lewis Acid Promoted Reactions, Kluwer Academic
Publishers, Dordrecht, 1989.
[5] P.T. Anastas, M.M. Kirchhoff, Acc. Chem. Res. 35 (2002) 686–694.
[6] S. Kobayashi, I. Hachiya, J. Org. Chem. 59 (1994) 3590–3596.
[7] S. Kobayashi, C. Ogawa, Chem. Eur. J. 12 (2006) 5954–5960.
[8] A. Corma, L.T. Nemeth, M. Renz, S. Valencla, Nature 412 (2001) 423–425.
[9] E. Nikolla, Y. Roman-Leshkov, M. Moliner, M.E. Davis, ACS Catal. 1 (2011)
408–410.
[10] K. Nakajima, Y. Baba, R. Noma, M. Kitano, S. Hayashi, M. Hara, J. Am. Chem. Soc.
133 (2011) 4224–4227.
[11] R. Buzzoni, G. Ricchiardi, C. Lamberti, A. Zecchina, Langmuir 12 (1996)
930–940.
[12] E. Spano, G. Tabacchi, A. Gamba, E. Fois, J. Phys. Chem.
21651–21661.
[13] Y. Román-Leshkov, J.N. Chheda, J.A. Dumesic, Science 312 (2006)
1933–1937.
[14] B. Girisuta, L.P.B.M. Janssen, H.J. Heeres, Ind. Eng. Chem. Res. 46 (2007)
1696–1708.
[15] M.J. Antal Jr., W.S.L. Mok, Carbohydr. Res. 199 (1990) 91–109.
[16] H.E. van Dan, A.P.G. Kieboom, H. van Bekkum, Starch/Stärke 38 (1986)
95–101.
[17] We performed adsorption experiment of levulinic acid on phosphate/TiO2.
THF/aqueous solution (2.0 mL (THF, 1.8 mL; distilled water, 0.2 mL)) containing
levulinic acid (0.11 mmol) and catalyst (0.05 g) was stirred at 393 K for 30 min.
The amount of adsorbed levulinic acid was estimated to be ca. 0.10 mmol g−1
which corresponds to ca. 5% yield.
4. Conclusion
In summary, anatase TiO2, an abundant and inexpensive mate-
rial, has water-tolerant Lewis acid sites. Bare anatase TiO2 cannot
function as an efficient heterogeneous catalyst for selective trans-
formation of glucose into HMF, requiring selective isomerization
of glucose into fructose and intramolecular dehydration of fruc-
tose, because of intermolecular side reactions. On the other hand,
TiO2 modified with H3PO4 (phosphate/TiO2), where OH groups
on TiO2 are esterified into O-PO(OH)2 by phosphoric acid, exhibits
high HMF yield (ca. 80%) in THF–water mixture (THF/H2O = 90/10).
Such high HMF yield can be achieved under diluted glucose solu-
tion (ca. 1 wt%) and high catalyst/glucose ratio (50/20 wt%). No
decrease in original activity for subsequent reactions demonstrated
that phosphate/TiO2 can function as a stable and reusable hetero-
geneous catalyst for HMF production.
B 110 (2006)
Acknowledgements
This work was supported by the Core Research for Evolutional
Science and Technology (CREST) and Advanced Low Carbon Tech-
nology Research and Development (ALCA) programs of the Japan
Science and Technology Agency (JST).
Please cite this article in press as: K. Nakajima, et al., J. Mol. Catal. A: Chem. (2013), http://dx.doi.org/10.1016/j.molcata.2013.09.012