Catalytic dehydration of fructose to 5-hydroxymethylfurfural over Nb 2O5 catalyst in organic solvent

Article abstract of DOI:10.1016/j.carres.2012.12.021

The catalytic dehydration of fructose to 5-hydroxymethylfurfural (HMF) in DMSO was performed over Nb₂O₅ derived from calcination of niobic acid at various temperatures (300-700 °C). The catalysts were characterized by powder X-ray diffraction, N₂ physical adsorption, temperature-programed desorption of NH₃, n-butylamine titration using Hammett indicators, infrared spectroscopy of adsorbed pyridine, and X-ray photoelectron spectroscopy. It was found that both catalytic activity and surface acid sites decrease with increasing calcination temperatures. The Nb₂O₅ derived from calcination of niobic acid at 400 °C reveals the maximum yield of HMF among all the catalysts, although the amount of acid sites on the catalyst is lower than that on the sample calcined at 300 °C. The results suggest that the presence of larger amounts of strong acid sites on the surface of the Nb₂O₅ calcined at 300 °C may promote side reactions. The Nb₂O₅ prepared at 400 °C shows 100% fructose conversion with 86.2% HMF yield in DMSO at 120 °C after 2 h. The activity of the catalyst decreases gradually during recycle because of coke deposition; however, it can be fully recovered by calcination at 400 °C for 2 h, suggesting that this catalyst is of significance for practical applications.

Full text of DOI:10.1016/j.carres.2012.12.021

Carbohydrate Research 368 (2013) 78–83
Contents lists available at SciVerse ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Catalytic dehydration of fructose to 5-hydroxymethylfurfural over
Nb2O5 catalyst in organic solvent
Fenfen Wang a, Hai-Zhen Wu b, Chun-Ling Liu a, Rong-Zhen Yang a, Wen-Sheng Dong a,
⇑
a Key Laboratory of Applied Surface and Colloid Chemistry (SNNU), MOE, School of Chemistry and Chemical Engineering, Shaanxi Normal University, Xi’an 710062, China
b College of Chemistry Life Science, Weinnan Normal University, Weinan 714000, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 6 October 2012
Received in revised form 19 December 2012
Accepted 21 December 2012
Available online 3 January 2013
The catalytic dehydration of fructose to 5-hydroxymethylfurfural (HMF) in DMSO was performed over
Nb2O5 derived from calcination of niobic acid at various temperatures (300–700 °C). The catalysts were
characterized by powder X-ray diffraction, N2 physical adsorption, temperature-programed desorption of
NH3, n-butylamine titration using Hammett indicators, infrared spectroscopy of adsorbed pyridine, and
X-ray photoelectron spectroscopy. It was found that both catalytic activity and surface acid sites decrease
with increasing calcination temperatures. The Nb2O5 derived from calcination of niobic acid at 400 °C
reveals the maximum yield of HMF among all the catalysts, although the amount of acid sites on the cat-
alyst is lower than that on the sample calcined at 300 °C. The results suggest that the presence of larger
amounts of strong acid sites on the surface of the Nb2O5 calcined at 300 °C may promote side reactions.
The Nb2O5 prepared at 400 °C shows 100% fructose conversion with 86.2% HMF yield in DMSO at 120 °C
after 2 h. The activity of the catalyst decreases gradually during recycle because of coke deposition; how-
ever, it can be fully recovered by calcination at 400 °C for 2 h, suggesting that this catalyst is of signifi-
cance for practical applications.
Keywords:
Fructose
Dehydration
5-Hydroxymethylfurfural
Niobium oxide
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
organic acids, Lewis acids, ionic liquids, and so on.11–19 However,
the catalysts mentioned above have some limitations with respect
Fossil fuels have brought about substantial social and economic
benefits for the whole world in the past few decades. Unfortu-
nately, they are faced with decreasing dilemma and gradual
exhaustion because of excessive consumption. In addition, in con-
sideration of more increasing environmental consciousness and
sustainable development as well as in the pursuit of low carbon
routes, exploiting alternative and renewable resource is considered
as an efficient method to solve fossil crisis and environmental pol-
lution. Biomass which is sustainable and degradable is widely dis-
tributed in nature, and can be expected to be a potential clean
substitute for fossil fuel and industrial feedstock.1–5 5-Hydroxym-
ethylfurfural (HMF), as a valuable biomass-derived platform com-
pound which is usually derived from degradation of cellulose, and
subsequent dehydration of the resultant monosaccharide, can be
converted to numerous various chemical products and liquid fuel,
potentially useful for fine chemicals, pharmaceuticals, the petro-
leum industry, and furanose-based polymers.6–10
to corrosion of equipment, separation from products, cost and
recovery of catalysts, and lower selectivity of products.
Solid acid catalysts as important functional materials are becom-
ing more and more attractive and intriguing due to stronger acidity
and environmentally benign properties as well as simple separation,
less corrosive, and safety, so they are desirable and favorable in the
chemical industry compared with homogeneous acid catalysts.20–22
In particular, niobium oxide (Nb2O5) as a promising solid acid
catalyst, which shows high acid strength and stable activity, has
been widely used for various reactions, such as esterification,
isomerization, hydrolysis, dehydration, etc.23–25 Previously, niobic
acid (Nb2O5ꢀxH2O) and niobium phosphate (NbOPO4) have been
used in the reaction of fructose dehydration in the aqueous
phase or water–2-butanol biphasic system.26,27 The higher
catalytic performance of niobium phosphate than niobic acid was
believed to be related with the higher effective acidity of its
surface.27,28
Recently, diverse homogeneous catalysts have been employed
in the dehydration of fructose to HMF, including mineral acids,
In the present work, the dehydration of fructose to HMF in
DMSO solvent catalyzed by niobium oxides derived from calcina-
tion of commercial niobic acid at various temperatures was inves-
tigated. Our data provide a basis for correlating the catalyst acidity
with the catalytic performance in fructose dehydration.
⇑
Corresponding author. Tel./fax: +86 (0)29 81530806.
E-mail address: wsdong@snnu.edu.cn (W.-S. Dong).
0008-6215/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.carres.2012.12.021

F. Wang et al. / Carbohydrate Research 368 (2013) 78–83
79
Table 1
2. Results and discussion
2.1. XRD
Textural properties of Nb2O5 samples at various calcination temperatures
Calcination
temperature
(°C)
BET surface
area (m2 gꢁ1
Pore volume
Average pore diameter
(nm)
)
(cm3 gꢁ1
)
Figure 1 shows the XRD patterns of the Nb2O5 catalysts derived
from calcination of niobic acid at various temperatures in air. The
sample remains amorphous without distinct diffraction peaks after
calcination at 300 °C. The XRD pattern of Nb2O5 calcined at 400 °C
reveals some weak diffraction peaks at 2h = 22.6°, 28.5°, and 36.7°
which can be assigned to the pseudohexagonal Nb2O5 crystal (TT
phase). With increasing calcination temperatures to 500 and
600 °C, the intensity of the diffraction peaks assigned to the
pseudohexagonal Nb2O5 increases, indicating an increase of crys-
tallinity of sample. For the sample calcined at 700 °C, the diffrac-
tion peaks at 2h = 28.5° and 36.7° were found to be split into two
peaks. This splitting indicates a transformation of the pseudohex-
agonal TT phase to the orthorhombic phase (T phase) of Nb2O5 dur-
ing the calcination at 700 °C. This phenomenon is quite similar to
that already reported by Xu and co-workers.29 The pseudohexago-
nal phase has a lower crystallinity and can be considered a modi-
fication of the orthorhombic phase.
BJH
BJH
adsorption desorption
300
400
500
600
700
194.1
141.7
49.9
17.6
5.7
0.19
0.17
0.12
0.10
0.05
3.6
4.2
9.6
27.2
47.9
3.2
3.6
7.1
15.7
34.8
and strong acid sites. With increasing calcination temperatures
the amounts of desorbed NH3 from the Nb2O5 samples decreases
evidently, indicating the reduction of acid sites and acid strength
on the surface of the samples.
2.4. Measurement of acid strength
The effect of calcination temperature on the acidity and acid site
distribution of Nb2O5 catalyst is shown in Figure 3. Based
on the acid strength, acid sites were divided into three groups:
medium-strong and weak acid (ꢁ3.0 6 H0 6 6.8), strong acid
(ꢁ8.2 6 H0 6 ꢁ3.0), and very strong acid (H0 6ꢁ8.2). None of the
Nb2O5 catalysts shows very strong acid sites (H0 6ꢁ8.2). The total
amount of acid sites in terms of mmol gꢁ1 (H0 66.8) of Nb2O5
samples decreases with increasing calcination temperatures. The
increase of calcination temperature also reduces both acid
strength and acid amount in the different acid strength ranges
(strong acid (ꢁ8.2 6 H0 6 ꢁ3.0), medium-strong and weak acid
(ꢁ3.0 6 H0 6 6.8)). This is consistent with the results of NH3–
TPD. Figure 3 also shows the fraction of acid sites with different
acid strength on the Nb2O5 catalysts. It can be observed that the
fraction of strong acid sites (ꢁ8.2 6 H0 6 ꢁ3.0) decreases with
increasing calcination temperatures while, the fraction of medium
strong and weak acid sites (ꢁ3.0 6 H0 6 6.8) increases with
increasing calcination temperatures.
2.2. Nitrogen physisorption
Textural properties of the Nb2O5 samples derived from the
nitrogen physisorption are shown in Table 1. The results indicate
that the increase of calcination temperature results in a decrease
in the sample surface area from 194.1 to 5.7 m2 gꢁ1 and pore vol-
ume from 0.19 to 0.05 cm3 gꢁ1, but there is an increasing tendency
in average pore diameter. The BET surface area and pore volume
are markedly decreased when the calcination temperature was
P400 °C. At the calcination temperature of 700 °C, the sample pre-
sents quite low surface area (5.7 m2 gꢁ1
) and pore volume
(0.05 cm3 gꢁ1), these results are in accordance with the literature
reported by Tanabe and co-workers.30
2.3. NH3–TPD measurement
The NH3–TPD curves of samples calcined at different tempera-
tures are shown in Figure 2. The sample calcined at 300 °C shows
two desorption peaks centered at around 200 and 400 °C, respec-
tively, indicating that the surface of this sample contains two dif-
ferent acid sites. The acid sites based on desorption temperature
were divided into three categories, weak acid (150–250 °C), med-
ium acid (250–350 °C), and strong acid (350–450 °C).31 The present
results indicate that the Nb2O5 samples contain weak, medium,
2.5. FT-IR of adsorbed pyridine
In order to gain more information about acid sites on the sur-
face of these Nb2O5 catalysts, the infrared spectra of pyridine ad-
sorbed on the Nb2O5 samples calcined at various temperatures
were recorded and the results are shown in Figure 4. It can be seen
e
d
a
b
c
c
b
a
d
e
10
20
30
40
50
60
70
80
100
200
300
400 500
oC
Temp. ( )
600
700
2θ ( deg.)
Figure 1. XRD patterns of Nb2O5 calcined at different temperatures: (a) 300, (b)
400, (c) 500, (d) 600, and (e) 700 °C.
Figure 2. NH3–TPD curves of Nb2O5 calcined at different temperatures: (a) 300, (b)
400, (c) 500, (d) 600, and (e) 700 °C.

80
F. Wang et al. / Carbohydrate Research 368 (2013) 78–83
1.2
100
90
80
70
60
50
40
100
120 oC
120 oC
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
-8.2
-3.0
≤
≤
H0
≤
≤
-3.0
6.8
1.0
0.8
H0
90
80
the total acidity at H0
≤
6.8
90 oC
0.6
-3.0
-8.2
≤
≤
H0
≤
≤
6.8
70
60
50
40
90 oC
H0
-3.0
0.4
0.2
0.0
300
400
500
600
700
300
400
600
700
500
Calcined temp. ( )
oC
Calc. temp.
(
oC
)
Figure 5. Effect of calcination temperature on the conversion of fructose and yield
of HMF: (a) 90 and (b) 120 °C. (Reaction conditions: fructose 0.1 g, DMSO 5.0 g,
catalyst 0.01 g, 2 h.)
Figure 3. Effect of calcination temperature on the acidity and surface acid unit of
Nb2O5 samples.
that both Lewis and Brønsted acid sites are present on the surface
of samples calcined at lower temperatures. The absorption associ-
ated with Brønsted acid sites (ca. 1543, 1650 cmꢁ1) appears to de-
crease remarkably with increasing calcination temperatures and
became invisible when the calcination was P500 °C. This surface
change is considered to be due to a phase transformation of amor-
phous Nb2O5ꢀnH2O to T–T phase Nb2O5 at around 500 °C as ob-
served by XRD.30 The peak intensity assigned to Lewis acid sites
(1440, 1485, 1570, 1610 cmꢁ1) shows a maximum on the sample
calcined at 300 °C, and decreases with the increase of pretreatment
temperature ranging from 300 to 700 °C.
conversion of fructose, which is closely related with its structure
and surface acidity. The sample calcined at 300 °C is amorphous,
and it has the highest BET surface area and the most acid sites.
With increasing calcination temperatures, the amorphous Nb2O5
transforms gradually to the pseudohexagonal phase and further
to the orthorhombic phase, resulting in the decrease of surface
area, acid strength, and acid sites; consequently the conversion
of fructose decreases gradually. On the other hand, the yield of
HMF reaches the maximum for the sample calcined at 400 °C,
and then decreases with further increasing calcination tempera-
ture. Even for the sample calcined at 700 °C, it still reveals 76.4%
fructose conversion and 65.4 HMF yield, suggesting that Lewis acid
sites play an important role in catalytic dehydration of fructose to
HMF in the present cases. For the samples calcined below 400 °C,
the contribution from Brønsted acid sites could not be ruled out,
because Brønsted acid sites eliminated by calcination below
400 °C can be quantitatively regenerated with suitable rehydra-
tion.29,30 However, the regeneration of Brønsted acid sites becomes
impossible once the calcination temperature is increased to
500 °C.30 Nevertheless, our present results show that Lewis acid
sites can catalyze effectively the dehydration of fructose to HMF.
The sample calcined at 300 °C also shows the highest fructose con-
version at 90 °C. However, this sample does not have the highest
HMF yield, suggesting that the presence of larger amounts of
strong acid sites may promote side reactions such as polymeriza-
tion of HMF or cross-polymerization between fructose and HMF,
bringing about numerous polymers and humin matter. The results
revealed that the sample calcined at 400 °C gave a yield of 86.2%
HMF at 120 °C for 2 h, it is the most effective catalyst for the dehy-
dration of fructose to HMF in the present study.
2.6. Catalytic activity
The catalytic dehydration of fructose to HMF in DMSO using
Nb2O5 as catalyst under different conditions was investigated. Fig-
ure 5 shows the reaction results of the Nb2O5 samples calcined at
different temperatures 120 and 90 °C, respectively. Obviously, cat-
alytic performance of the catalysts was noticeably affected by the
calcination temperature. Initially, the reaction was performed at
120 °C, fructose was completely converted over all the catalysts,
the yield of HMF reaches the maximum on the Nb2O5 catalyst cal-
cined at 400 °C, and then decreases with further increasing calcina-
tion temperature. When the reaction was carried out at 90 °C, the
results show that the catalyst calcined at 300 °C gave the highest
e
d
We also compared the catalytic performance of other different
solid acid catalysts for example, niobic acid, NO3ꢁ/PO43ꢁ/Nb2O5,
c
-Al2O3, AmberliteÒ-IRC-50 ion-exchange resin for the dehydra-
tion of fructose to HMF under the same experimental conditions
at 120 °C for 2 h, although these conditions might not be the opti-
mal for some catalysts in terms of their performance. The experi-
mental results are summarized in Figure 6. Nb2O5 exhibited
excellent catalytic activity for the dehydration of fructose among
these tested solid acid catalysts, affording a complete conversion
of fructose and 86.2% HMF yield. For uncalcined niobic acid, the
conversion of fructose was also 100%, the yield of HMF was
80.2%, which is lower than that of Nb2O5 catalyst. In the presence
of NO3ꢁ/PO43ꢁ/Nb2O5 solid acid catalyst, it showed 100% fructose
conversion and the HMF yield reached 78.2%. When applying
AmberliteÒ-IRC-50 ion-exchange resin as catalyst, it gave a HMF
c
b
a
1543
1485
1650
1605
1570
1440
1800 1750 1700 1650 1600 1550 1500 1450 1400
Wavenumber ( )
cm-1
Figure 4. FTIR spectra of pyridine adsorbed on Nb2O5 calcined at different
temperatures: (a) 300, (b) 400, (c) 500, (d) 600, and (e) 700 °C.

F. Wang et al. / Carbohydrate Research 368 (2013) 78–83
81
100
80
60
40
20
0
100
80
60
40
20
0
100
conversion
yield
90
80
70
60
50
40
NO3-/PO43-/Nb2O5 Amberlite resin Al2O3
Niobic acid
Nb2O5
90
100
110
120
130
Reaction temp. ( )
oC
Catalyst
Figure 6. Effects of different solid acid catalysts on the yield of HMF. (Reaction
conditions: fructose 0.1 g, DMSO 5.0 g, catalyst 0.01 g, 120 °C, 2 h.)
Figure 8. Effect of reaction temperature on the conversion of fructose and the yield
of HMF. (Reaction conditions: fructose 0.1 g, DMSO 5.0 g, catalyst 0.01 g, 2 h.)
yield of 73.3%. Although the conversion of fructose was completed
research indicated that fructose had a good solubility in DMSO and
the formation of furanoid is preferred in DMSO.33 Clearly, in our
catalytic system DMSO is also an efficient solvent in promoting
dehydration of fructose to HMF, giving a HMF yield of 86.2%, and
no rehydration byproducts such as levulinic acid and formic acid
were detected. When adding DMA into DMSO (w/w = 3:7), the
yield of HMF dramatically dropped to 60%, indicating that DMA
solvent is undesirable for the dehydration of fructose to HMF
although the detail reason is still not clear. When the reaction
was conducted in ionic liquid [BMIM]Br, the yield of HMF was
78.4%. In the presence of the mixture of equivalent water with
PEG-200, the conversion of fructose was very low and only 8%
HMF yield was obtained. When water was used as solvent, only
49% fructose conversion and 7.3% HMF yield were obtained, indi-
cating that water was not beneficial to the dehydration of fructose
to HMF in the present cases because HMF was prone to decompose
to levulinic acid, formic acid, and other byproducts in the presence
of water.26 Although the present research shows that DMSO is an
efficient solvent in promoting the dehydration of fructose to
HMF, it is difficult to separate from products. Hence much research
still needs to be done to find more appropriate solvent.
The results of fructose conversion and HMF yield as a function
of reaction temperature are depicted in Figure 8. The conversion
of fructose increases from 86.2% to 100% with increasing the reac-
tion temperature from 90 to 120 °C, and the HMF yield increases
from 65.6% to 86.2%. Nevertheless, further increasing reaction tem-
perature led to the reaction solution becoming brown black grad-
ually, and the yield of HMF decreased to 82.4% at 130 °C. This
may be due to the fact that at higher temperatures self-polymeri-
zation of HMF intermolecular or cross-polymerization between
fructose and HMF, brings about numerous polymers and humin
matter, which were not characterized.5 The results suggest the
reaction temperature could be a crucial parameter for formation
of HMF, and 120 °C is the optimum reaction temperature for the
dehydration of fructose.
using
c-Al2O3 as catalyst, it exhibited much lower activity than
Nb2O5, only 55% HMF yield was obtained under the same reaction
condition. So the Nb2O5 calcined at 400 °C was chosen as a catalyst
for further study in detail.
We also investigated the influence of calcination duration time
of Nb2O5 catalyst at 400 °C on the dehydration of fructose. The re-
sults showed that when niobic acid was calcined at 400 °C for 3 h, a
fructose conversion of 100% and a HMF yield of 80.9% were
obtained. Prolonging the duration time to 5 h, the yield of HMF de-
creased to 78.2%. The Nb2O5 calcined at 400 °C for 4 h exhibited the
highest HMF yield.
The synthesis of HMF from biomass has been conducted in var-
ious reaction media for example, water, dimethyl sulfoxide
(DMSO), dimethylformamide (DMA), ionic liquids, isopropyl alco-
hol, or biphasic systems. HMF can be isolated by extraction with
various organic solvents such as methyl isobutyl ketone, dichloro-
methane, ethyl acetate, tetrahydrofuran, diethyl ether, and ace-
tone.32 It has been found that solvent plays a very important role
in the reaction. Hence, in the following section the effects of differ-
ent solvents such as aprotic solvent DMSO, the mixture of DMA
and DMSO (w/w = 3:7), [BMIM]Br, and the mixture of PEG-200
and water on the dehydration of fructose to HMF were examined
at 120 °C for 2 h, and the results are depicted in Figure 7. Previous
100
conversion
yield
80
60
40
20
0
We also investigated the influence of reaction time on the dehy-
dration of fructose at 120 °C. The results show that when reaction
was carried out for 120 min the highest HMF yield could be
obtained.
DMSO/DMA
[BMIM]Br
H2O
DMSO
PEG-200/H2O
2.7. Catalyst recycles
Solvent
The catalyst reusability was tested using Nb2O5 calcined at
400 °C. After reaction, the catalyst was separated from the reaction
solution by centrifugation, and washed with deionized water, dried
Figure 7. Effect of solvent on the yield of HMF. (Reaction conditions: fructose 0.1 g,
DMSO 5.0 g, catalyst 0.01 g, 120 °C, 2 h.)

82
F. Wang et al. / Carbohydrate Research 368 (2013) 78–83
bromide), PEG-200, and the other reagents were all analytical
grade and used without further purification.
100
80
60
40
20
0
Nb2O5 solid acid catalysts with different acidity were prepared
by calcination of commercial niobic acid (Nb2O5ꢀnH2O) at various
temperatures (300–700 °C) in air for 4 h. NO3ꢁ/PO43ꢁ/Nb2O5 cata-
lyst was prepared by impregnating niobic acid with nitric acid
and phosphoric acid, the sample was filtrated and dried at
100 °C, and finally calcined at 400 °C for 4 h.
4.2. Catalyst characterizations
The crystal structures of Nb2O5 samples were characterized by
powder X-ray diffraction (XRD) using a Rigaku D/MAX-III X-ray dif-
fractometer scanning from 10° to 80° (2h) at a rate of 0.02° sꢁ1 with
1
2
3
4
5
a Cu Ka (k = 0.15418 nm) radiation source at 35 kV, 40 mA. BET
Reclycling times
surface areas, pore volumes, and average pore diameters of
Nb2O5 samples were obtained from N2 (77 K) adsorption measure-
ment using a Micromeritics ASAP2020M system. The samples were
pretreated under vacuum at 250 °C for 4 h before the measure-
ment. The average pore diameter data were calculated according
to the Barrett–Joyner–Halenda (BJH) model.
The acidities of samples were determined by ammonia temper-
ature-programed desorption (NH3–TPD) using a Micromeritics
AutoChemII-2920 system. A 100 mg sample (50–80 mesh) was
loaded into the sample tube and pretreated in He at 300 °C for
60 min. The sample was cooled, and NH3 was adsorbed at 120 °C
for 30 min, then NH3 adsorbed was further purged with dry helium
at the same temperature to remove the weakly adsorbed NH3. Fi-
nally, the sample was heated with a linear rate of 10 °C/min under
dry helium up to 700 °C. The effluent was monitored by a quadru-
pole mass spectrometer.
The acid strength and amounts of samples were examined by
the n-butylamine titration method using various Hammett indica-
tors, including anthraquinote (pKa = ꢁ8.2), chalcone (pKa = ꢁ5.6),
dicinnamalacetone (pKa = ꢁ3.0), methyl yellow (pKa = 3.3), and
neutral red (pKa = 6.8). Acid strength was expressed by the Ham-
mett acidity function (H0) that was scaled by the pKa values of
the indicators. Before the measurement, the 50 mg samples were
pretreated at 250 °C for 4 h in flowing nitrogen.
FT-IR spectra of adsorbed pyridine were recorded on an FT-IR
spectrometer (Bruker EQUINX 55). The catalyst sample was com-
pressed into a self-supporting wafer (10 mg/cm2, diameter of
15 mm), and then it was placed into the FTIR cell equipped with
CaF2 windows. The wafers were evacuated under vacuum at
300 °C for 1 h and then cooled to room temperature, followed by
exposure to pyridine vapor. The wafers were allowed to adsorb
pyridine for 30 min. Infrared spectra were recorded at room tem-
perature after subsequent evacuation at 100 °C.
Figure 9. Recycling of Nb2O5 for dehydration of fructose to HMF. (Reaction
conditions: fructose 0.1 g, DMSO 5.0 g, catalyst 0.01 g, 120 °C, 2 h.)
under vacuum at 60 °C for 10 h before the next run. From Figure 9,
it is found that the catalyst can be used repeatedly at least five
times with slight loss of HMF yield. The loss of activity was mainly
caused by the formation of carbon deposition on the catalyst sur-
face. The XPS analysis confirms that carbon content in the catalyst
reused for five recycles clearly increases (7.53%). BET analysis indi-
cates that the specific surface area decreases from 141.7 m2 gꢁ1 for
the fresh catalyst to 75.8 m2 gꢁ1 for the used one. This is probably
due to adsorption of humins, etc. on the surface of the catalyst.
When the used catalyst was reactivated by calcination in air at
400 °C for 2 h, its activity was almost recovered fully and achieved
85.9% yield of HMF.
3. Conclusion
The present study demonstrates that Nb2O5 catalysts prepared
by calcination of niobic acid are effective for the dehydration of
fructose to produce 5-hydroxymethylfurfural. The catalyst perfor-
mance for the dehydration reaction is significantly affected by
the catalyst calcination temperature that induces the changes in
surface acidity and crystallization of Nb2O5. The Lewis acid sites
on the catalysts can catalyze effectively the dehydration of fructose
to HMF. The calcination at 300 °C produces amorphous Nb2O5 cat-
alyst that exhibits the highest acidity compared with those Nb2O5
samples obtained by calcination at 400–700 °C. This sample also
shows the maximum activity for the conversion of fructose, how-
ever, its yield to HMF is lower than that of the Nb2O5 prepared at
400 °C, suggesting that the presence of larger amounts of stronger
acid sites on the surface of the sample prepared at 300 °C could
promote side reactions. The Nb2O5 prepared at 400 °C shows
100% fructose conversion with 86.2% HMF yield at 120 °C for 2 h,
and it can be recycled at least five times with slight loss of catalytic
activity.
X-ray photoelectron spectroscopy (XPS) was performed using a
VG Scientific (ESCALAB 250) photoelectron spectroscopy. The X-ray
source was Al Ka with 1486.6 eV operating at 12 kV and 73 W with
the vacuum about 10ꢁ8 Torr. The binding energy was calibrated
using the C1s peak at 284.8 eV.
4.3. Catalytic reaction
4. Experimental
The batch catalytic experiments were conducted in a 10 mL
thick-walled glass vessel equipped with a reflex condenser. One
hundred milligrams (0.56 mmol) of fructose as substrate and 5 g
(64.1 mmol) of DMSO as solvent were firstly added in the reactor
and then 10 mg catalyst was added into the mixture, the reaction
mixture was heated to desirable temperatures (90 and 120 °C,
respectively) with an oil bath under strong stirring (500 rpm).
Under these conditions the mass transfer effect was eliminated.
After the reaction, the reactor was cooled to room temperature,
4.1. Chemicals and catalysts
Fructose (99%) and niobic acid (99%) were purchased from
Tianjin Chemical Reagents Company. c-Al2O3 was purchased from
Sinopharm Chemical Reagent Co. Ltd (Shanghai, China). 5-Hydrox-
ymethylfurfural (99%) and AmberliteÒIRC-50 ion-exchange resin
were obtained from Sigma–Aldrich. Levulinic acid (98%) was pur-
chased from Alfa Aesar. DMSO (dimethyl sulfoxide), DMA (N,N-
dimethyl-acetamide), [BMIM]Br (1-butyl-3-methylimidazolium

F. Wang et al. / Carbohydrate Research 368 (2013) 78–83
83
2. Roman-Leshkov, Y.; Chheda, J. N.; Dumesic, J. A. Science 2006, 312, 1933–1937.
3. Quan, C.; Guo, X. C.; Guan, J.; Mu, X. D.; Zhang, D. K. Appl. Catal. A 2011, 403, 98–
103.
the catalyst was separated by centrifugation and the post-reaction
sample was diluted with mobile phase solution before analysis.
4. Zhao, H. B.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Science 2007, 316, 1597–
1600.
5. Hansen, T. S.; Woodley, J. M.; Riisager, A. Carbohydr. Res. 2009, 344, 2568–2572.
6. Bao, Q. X.; Qiao, K.; Tomida, D.; Yokoyama, C. Catal. Commun. 2008, 9, 1383–
1388.
7. Chheda, J. N.; Huber, G. W.; Dumesic, J. A. Angew. Chem., Int. Ed. 2007, 46, 7164–
7183.
8. Hu, S. Q.; Zhang, Z. F.; Zhou, Y. X.; Song, J. L.; Fan, H. L.; Han, B. X. Green Chem.
2009, 11, 873–877.
9. Huber, G. W.; Chheda, J. N.; Barrett, C. J.; Dumesic, J. A. Science 2005, 308, 1446–
1450.
10. Yang, Y.; Xiang, X.; Tong, D.; Hu, C.; Abu-Omar, M. M. Bioresour. Technol. 2012,
116, 302–306.
4.4. Analysis
Sample analysis was performed on a Shimadzu HPLC system
equipped with a RID-10A detector and a Bio-Rad Aminex HPX-
87H ion exclusion column (300 ꢂ 7.8 mm). 0.005 M H2SO4 was
used as the mobile phase that had a flow rate of 0.55 mL/min.
The column temperature was 50 °C, and the detector temperature
was set at 45 °C. The amount of fructose and HMF was determined
by using calibration curves.
11. Lourvanij, K.; Rorrer, G. L. J. Chem. Tech. Biotechnol. 1997, 69, 35–44.
12. Asghari, F. S.; Yoshida, H. Ind. Eng. Chem. Res. 2006, 45, 2163–2173.
13. Li, Y. H.; Lu, X. Y.; Yuan, L.; Liu, X. Biomass Bioenerg. 2009, 33, 1182–1187.
14. Mittal, N.; Nisola, G. M.; Chung, W.-J. Tetrahedron Lett. 2012, 53, 3149–3155.
15. Yong, G.; Zhang, Y.; Ying, J. Y. Angew. Chem., Int. Ed. 2008, 47, 9345–9348.
16. Ohara, M.; Takagaki, A.; Nishimura, S.; Ebitani, K. Appl. Catal. A 2010, 383, 149–
155.
17. Ryu, J.; Choi, J.-W.; Suh, D. J.; Ahn, D. J.; Suh, Y.-W. Catal. Commun. 2012, 24, 11–
15.
18. Jadhav, A. H.; Kim, H.; Hwang, I. T. Catal. Commun 2012, 21, 96–103.
19. Qu, Y.; Huang, C.; Zhang, J.; Chen, B. Bioresour. Technol. 2012, 106, 170–172.
20. Okuhara, T. Chem. Rev. 2002, 102, 3641–3666.
4.5. The fructose conversion and HMF yield definitions
The fructose conversion and HMF yield were defined as fol-
lows:Fructose conversion (mol %):
ꢀ
ꢁ
Fructose concentration in product
Fructose concentration in the loaded sample
X ¼ 1 ꢁ
ꢂ 100%
ð1Þ
21. Guo, X.; Cao, Q.; Jiang, Y.; Guan, J.; Wang, X.; Mu, X. Carbohydr. Res. 2012, 351,
35–41.
22. Ordomsky, V. V.; Van der, S. J.; Schouten, J. C.; Nijhuis, T. A. J. Catal. 2012, 287,
68–75.
23. Nowak, I.; Ziolek, M. Chem. Rev. 1999, 99, 3603–3624.
24. Tanabe, K.; Okazaki, S. Appl. Catal. A 1995, 133, 191–218.
25. Tanabe, K. Catal. Today 2003, 78, 65–77.
26. Yang, F.; Liu, Q.; Bai, X.; Du, Y. Bioresour. Technol. 2011, 102, 3424–3429.
27. Carniti, P.; Gervasini, A.; Biella, S.; Auroux, A. Catal. Today 2006, 118, 373–378.
28. Carniti, P.; Gervasini, A.; Biella, S.; Auroux, A. Chem. Mater. 2005, 17, 6128–
6136.
29. Chai, S. H.; Wang, H. P.; Liang, Y.; Xu, B. Q. J. Catal. 2007, 250, 342–349.
30. Iizuka, T.; Ogasawara, K.; Tanabe, K. Bull. Chem. Soc. Jpn. 1983, 56, 2927–2931.
31. Yadav, G. D.; Murkute, A. D. Adv. Synth. Catal. 2004, 346, 389–394.
32. Zhang, L.; Zhang, Y. ChemSusChem 2011, 4, 1745–1748.
33. Chheda, J. N.; Dumesic, J. A. Catal. Today 2007, 123, 59–70.
HMF yield (mol %):
Moles of carbon in product
Moles of carbon loaded as fructose
Y ¼
ꢂ 100%
ð2Þ
Acknowledgment
This work was financially supported by the Program for Chang-
jiang Scholars and Innovative Research Team in University of China
(IRT1070) and the Program for Key Science and Technology Inno-
vation Team of Shaanxi Province (2012KCT-21).
References
1. Qi, X. H.; Watanabe, M.; Aida, T. M.; Smith, R. L. Green Chem. 2008, 10, 799–805.