Catalytic conversion of xylose to furfural over the solid acid SO 42-/ZrO2-Al2O3/SBA-15 catalysts

Article abstract of DOI:10.1016/j.carres.2011.01.001

Al-promoted SO₄^2--/ZrO₂/SBA-15 catalysts were prepared and characterized by XRD, BET, ICP and NH₃-TPD techniques. The influence of introducing aluminum on the structure and surface properties of the catalyst and the catalytic activity for dehydration of xylose to furfural has been investigated. The introduction of the Al stabilizes the tetragonal phase of the ZrO₂ and thus increases the number and intensity of acid sites. Based on the characterization of the deactivated catalyst, the accumulation of byproducts is the main reason for the deactivation of the catalyst. Regeneration with H₂O₂ can completely recover the catalytic activity of the deactivated catalyst.

Full text of DOI:10.1016/j.carres.2011.01.001

Carbohydrate Research 346 (2011) 480–487
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Catalytic conversion of xylose to furfural over the solid acid SO42À/ZrO –Al O /
2
2 3
SBA-15 catalysts
a
a,
⇑
a
b
a
a,
⇑
Xuejun Shi , Yulong Wu , Panpan Li , Huaifeng Yi , Mingde Yang , Gehua Wang
a
Institute of Nuclear and New Energy Technology, Tsinghua University, Beijing 100084, China
Petrochina Xinjiang oilfield company, Karamay 834000, China
b
a r t i c l e i n f o
a b s t r a c t
Al-promoted SO 2 /ZrO /SBA-15 catalysts were prepared and characterized by XRD, BET, ICP and NH –
À
Article history:
4
2
3
Received 31 October 2010
Received in revised form 1 January 2011
Accepted 4 January 2011
TPD techniques. The influence of introducing aluminum on the structure and surface properties of the
catalyst and the catalytic activity for dehydration of xylose to furfural has been investigated. The intro-
duction of the Al stabilizes the tetragonal phase of the ZrO
2
and thus increases the number and intensity
Available online 7 January 2011
of acid sites. Based on the characterization of the deactivated catalyst, the accumulation of byproducts is
the main reason for the deactivation of the catalyst. Regeneration with H
catalytic activity of the deactivated catalyst.
2 2
O can completely recover the
Keywords:
Xylose
Furfural
Ó 2011 Elsevier Ltd. All rights reserved.
Dehydration
Acid catalysis
Sulfated zirconia
2
À1
1
. Introduction
specific surface area usually in the range 80–100 m g and a lack
of ordered mesoporosity and textural homogeneity, making it suit-
able for traditional vapor-phase reactions involving small molecules
Acid-catalyzed dehydration of saccharides into furfural has
1
0
been the focus of many studies because of the possibility of devel-
but less amenable to liquid-phase reactions. The mesoporous
SBA-15 material is a good candidate as a support for the dispersion
of sulfated zirconia because of its unique surface, pore structure,
and hydrothermal stabilities. In particular, it has tunable uniform
hexagonal channels ranging from 4 to 30 nm and thicker pore walls
in the range of 3 to 6 nm. Therefore, it provides excellent support
compared with its counterparts such as microporous zeolites be-
cause it is free from kinetic diffusion limitations in the case of larger
molecules.
It is well known that the Al, Mn, Fe, Ni, Pt etc. are effective pro-
moter to increase the stability and/or activity of sulfated zirconia
catalysts. The role of the promoter is to improve stabilization of
the active tetragonal phase of zirconia, as well as higher surface
areas, higher sulfate contents, and modification of the surface
1
,2
oping renewable chemicals from biomass. Homogenous acids,
such as mineral acids, have been employed broadly in these studies
and it has been found that these are effective and give furfural
1
1
2
yields of 40–60%. However, the mineral acids have serious draw-
1
1
backs in terms of separation and recycling, as well as material cor-
rosion. On the other hand, heterogeneous acid catalysts show
superior behavior over the homogenous acids in terms of selectiv-
ity, recyclability, and material corrosion although solid acid cata-
lysts tend to give low conversions (30–60%) and require reaction
times greater than 2 h. Many types of solid acid catalysts have been
studied in this process, such as inorganic oxides, heteropolyacids,
metal phosphates, and sulfated metal oxides,3 employing a range
of reaction media including water, anhydrous organic solvents or,
more recently, biphasic water/organic co-solvent systems.
–6
1
0
acidity.
In this paper, a series of Al-promoted SO42 /ZrO
/SBA-15 cata-
2
À
Sulfated zirconia is a well-known solid acid catalyst and has been
used in various commercially-important reactions, such as acyla-
tions, alkylations, nitrations, etherifications, esterifications, and iso-
lysts were prepared, and the chemical and structural properties
of these catalysts were characterized. The catalytic activities of
the catalysts were tested in the liquid-phase dehydration of xylose
to furfural. The deactivation and regeneration capacities of the cat-
alysts were also investigated during the dehydration of xylose. This
work focuses on the influences of Al on the structure and surface
acidity of SO4 /ZrO /SBA-15 and assesses the properties of these
2
solid acid catalysts and their actual practical application in the
dehydration of xylose to furfural.
7
–9
merizations.
However, conventional sulfated zirconia has a low
2
À
⇑
Corresponding authors. Tel.: +86 10 89796086; fax: +86 10 69771464.
E-mail addresses: wylong@tsinghua.edu.cn (Y. Wu), wanggehua@tsinghua.
edu.cn (G. Wang).
0
008-6215/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2011.01.001

X. Shi et al. / Carbohydrate Research 346 (2011) 480–487
481
2
2
. Experimental sections
.1. Catalyst preparation
Rad) coupled to a Shimadzu RID-10A differential refractive index
detector (for xylose and furfural). The mobile phase was 0.005 M
À1
2
H SO
4
. The analysis conditions were a flow rate of 0.6 mL min
and a column temperature of 65 °C. Authentic samples of
D-xylose
SBA-15 molecular sieves were synthesized according to the
and furfural were used as standards, and calibration curves were
used for quantification. With toluene and water as the mixed sol-
vent, the furfural present in the organic phase was quantified with
a Techcomp 7890II Gas chromatograph and a PEG-20M column
coupled with an FID detector. The carrier gas was nitrogen. Analy-
sis of the products was done thrice before the averaged result was
obtained.
method described in the literature.11 The SO4 /ZrO
1
2À
2
2
–Al O
3
/SBA-
5 samples with different Zr contents were prepared as follows.
A mixture of Al(NO O (0.06 g) and ZrOCl Á8H O (0.97–
Á9H
.42 g) was impregnated onto SBA-15 molecular sieves (3 g) using
3
)
3
2
2
2
2
an incipient wetness method. The impregnated sample was dried
at 120 °C for 1 h. The dried samples were hydrolyzed in 0.5 M
NH
ride ions and dried for 2 h at 120 °C. The ZrO
ples were treated with 0.5 M (NH SO , filtered, dried at 110 °C for
4 h and then calcined at 650 °C for 3 h.
3
ÁH
2
O for 3 h and then washed with water to remove the chlo-
2
–Al /SBA-15 sam-
2
O
3
3. Results and discussion
3.1. XRD
4
)
2
4
2
2
.2. Catalyst characterization
The low-angle XRD patterns of the calcined SBA-15 and the
SO42 /ZrO
À
2
2
–Al O
3
/SBA-15 samples are shown in Figure 1. The cal-
The X-ray powder diffraction (XRD) patterns of the catalysts
were obtained on a Rigaku D/max-RB 2500 diffractometer using
Cu K radiation with the following parameters: 40 kV, 50 mA, 2h
scanning from 0.6° to 5.0° for the low-angle XRD and 40 kV,
00 mA, 2h scanning from 10° to 90° for high-angle XRD.
The N sorption isotherms and specific surface area were deter-
mined by using a Nova 3200e. The samples were pretreated at
20 °C for 5 h, and the specific surface area of the samples was
determined using the Brunauer–Emmett–Teller (BET) method.
The pore volume and pore size distribution were derived from
the desorption profiles of the isotherms using the Barrett–Joy-
ner–Halanda (BJH) method.
cined SBA-15 support (Fig. 1a) shows three well-resolved diffrac-
tion peaks in the 2h range of 0.7–2.0°, which correspond to the
diffraction of the (1 0 0), (1 1 0), and (2 0 0) planes. These peaks
are characteristic of the hexagonally ordered structure of SBA-
a
2
11
2À
1
5. The d100 peaks were still observed for the SO4 /ZrO
2
–
2
Al /SBA-15 samples, indicating that the hexagonal symmetry
2 3
O
of the SBA-15 support was preserved. However, the intensity of
the d100 peak was attenuated compared with that of the calcined
SBA-15. The attenuation of the d100 peaks is probably due to the
reduction in the X-ray scattering contrast between the silica walls
and the pore-filling material rather than a loss of structural or-
1
12
der. Compared with that of the calcined SBA-15, the (100) dif-
The NH
3
temperature programmed desorption (NH
3
–TPD)
2À
fraction of the SO4 /ZrO
–Al
2 2
O
3
/SBA-15 sample is slightly
experiments were carried out using a Thermo Electron Corporation
TPD/R/O 1100 series catalytic surfaces analyzer equipped with a
TCD detector. The sample (100 mg) was charged into a quartz reac-
tor and treated in helium flow (30 mL/min) at 350 °C for 50 min
2À
shifted to higher angles. It is possible that in the SO4 /ZrO
2
–
2 3
Al O /SBA-15 samples, Zr species are embedded into the channel
of SBA-15 support. Although the d100 peak further shifts to higher
angle, the three diffraction peaks also exist. This indicates that the
well-defined hexagonally ordered structure of the SBA-15 support
was not destroyed during the process of the catalyst preparation.
and then exposed to NH
3
(30 mL/min) at room temperature for
1
h and purged by helium (to remove gas-phase and physically ad-
3 3
sorbed NH ). Once a stable baseline of NH was obtained, the sam-
ple was heated from room temperature to 800 °C at a rate of 10 °C/
min and treated in a helium flow (30 mL/min).
The S element percentage present in the catalysts was deter-
mined using a ICP–AES spectrometer (Thermo Electron, IRIS Intre-
pid II XSP DUO).
The high-angle XRD patterns of the SBA-15 support and the
2À
SO4 /ZrO
2
2
–Al O
3
/SBA-15 samples are shown in Figure 2. The
SBA-15 support shows a broad diffraction peak in the range of
5–30 °C, which is the characteristic peak of amorphous silica
Fig. 2b). The peaks at 30.28, 35.20, 50.40 and 60.28° are the dif-
1
(
13
fractions of (1 1 1), (2 0 0), (2 2 0) and (1 3 1) of tetragonal ZrO
2
,
respectively (Fig. 2a), whereas the peak at 28.28° is the diffraction
of (1 1 1) of monoclinic ZrO (Fig. 2g). For the SO4 /ZrO –Al O /
2 2 2 3
2
.3. Catalytic activity tests
ꢀ
2À
SBA-15 samples no independent lines due to crystalline alumina
were observed in the present study indicating that alumina is
Batch catalytic experiments were performed under nitrogen in
a 50 mL airtight magnetically stirred stainless steel high-pressure
reactor and heated with electricity. In a typical procedure, 0.75 g
xylose, 0.5 g powdered catalyst, and 25 mL mixed solvent (7.5 mL
2
H O and 17.5 mL toluene, all the activity tests were carried out
in the mixed solvent) were poured into the reactor. Time zero
was taken to be the instant the reaction temperature reached
1
60 °C.
After the reaction, the deactivated catalysts were regenerated
by two methods. One was regenerated by calcination at 650 °C
for 3 h, the other was regenerated by treatment with H in a Tef-
2 2
O
lon-lined autoclave and magnetically stirred for 4 h at 100 °C. The
product was filtered and air dried at 100 °C overnight. The deacti-
vation and reusability tests for all samples were carried out at
1
60 °C. The products in the aqueous and organic phases were
quantified using HPLC equipment and GC, respectively.
For the experiments carried out with toluene and water as
the mixed solvent, xylose and furfural were quantitatively deter-
mined with a Shimadzu LC-20AD HPLC pump and an Aminex
HPX-87H Column 300 Â 7.8 (id) mm ion-exchange column (Bio-
2
À
Figure 1. Low-angle XRD patterns of the samples: (a) SBA-15; (b) SO /6% ZrO
2
–
4
2À
2À
Al
2
O
3
/SBA-15; (c) SO4 /9% ZrO
2
–Al
2
O
3
/SBA-15; (d) SO4 /12% ZrO
2
–Al
2
O
3
/SBA-15;
2À
2À
(
e) SO4 /15% ZrO
2
–Al
2
O
3
/SBA-15; (f) SO4 /12% ZrO
2
/SBA-15.

4
82
X. Shi et al. / Carbohydrate Research 346 (2011) 480–487
2
À
Figure 3. Nitrogen sorption isotherms and pore size distribution of the sample: (a)
SBA-15; (b) SO42À/6% ZrO –Al O /SBA-15; (c) SO 2À/9% ZrO –Al O /SBA-15; (d)
Figure 2. High-angle XRD patterns of the samples: (a) SO /ZrO
2
–Al
2
O
3
; (b)SBA-
4
5; (c) SO42 /6% ZrO
À
–Al
O
/SBA-15; (d) SO /9% ZrO
2À
–Al
O
2À
/
/
2
2
3
2
2 3
1
1
2
2
3
2
2
3
/SBA-15; (e) SO4
4
4
2
À
2À
SO42À/12% ZrO –Al O /SBA-15; (e) SO 2À/15% ZrO –Al O /SBA-15; (f) SO 2À/12%
2% ZrO –Al
2
2
O
3
/SBA-15; (f) SO4 /15% ZrO
2
–Al
2
O
3
/SBA-15; (g) SO4 /12% ZrO
2
2 2 3 2 2 3
4 4
SBA-15.
2
ZrO /SBA-15.
homogeneously mixed with zirconia and is present in an amor-
branch of isotherm from the parent SBA-15. The large tail exists in
the hysteresis of nitrogen adsorption–desorption isotherms and
the decrease in the pore volume, which might be due to the forma-
phous state. The diffraction peak of the tetragonal ZrO
2
can be ob-
served at 2h = 30.28° for the SO42 /ZrO
À
2
2
–Al O
3
/SBA-15 samples
with ZrO
ZrO gradually increases with higher ZrO
increasing ZrO
ters (present as the monoclinic ZrO
2
loading above 9 wt %, and the intensity of the tetragonal
2
tion of clustered layers of tetragonal ZrO phase in the internal pore
2
1
loading. However, with
clus-
) were formed on the surface of
/SBA-15 catalyst. However, compared
/SBA-15 catalyst, the broadened diffrac-
surface or partial blockage of pores. The pore size distribution is
shown inFigure 3 as an inset. It shows a very narrow pore size distri-
bution in the case of parent SBA-15, whereas the pore size distribu-
tion of the supported catalysts is changed and the uniformity of the
2
2
2
content (P15%), a very small amount of ZrO
2
2
2À
the SO4 /15% ZrO
with the SO4 /12% ZrO
2
–Al
2 3
O
2
À
mesopores decrease with an increase of ZrO
2
content. A small shoul-
2
2
À
tion peak of ZrO
gests that the SO4 /12% ZrO
crystallite size than that of the SO4 /12% ZrO
2
in SO /12% ZrO
2 2 3
–Al O /SBA-15 catalyst sug-
der peak centers on 3.3 nm in the pore size distribution curve of the
4
2
À
2À
2À
/SBA-15 catalyst has larger zirconia
SO
ples. The appearance of the peak is related to the fixed closure of
the hysteresis loop at a relative pressure close to 0.50 on the N
adsorption–desorption isotherm (Fig. 3e and f). According to the lit-
4
/15% ZrO
2
–Al
2
O
3
/SBA-15 and SO
4
2
/12% ZrO /SBA-15 sam-
2
2À
2
2 3
–Al O /SBA-15 cat-
alyst. This result demonstrates that the presence of the alumina
2
inhibited the growth of zirconia crystallite in the SO42 /ZrO
À
2
–
2
2
Al
2
O
3
/SBA-15 samples.
erature, the location of the peak does not represent the true pore
size of the sample, but it is characteristic of the adsorption. Thus,
what we can derive from the data is simply that non-uniform and fi-
ner pores are formed in the samples, and the irregular reduction in
pore size is probably caused by blocking of the mesoporous channels
2À
For the SO4 /12% ZrO
the monoclinic ZrO also can be seen in 2h = 28.28°. However, only
the tetragonal phase exists in the SO4 /12% ZrO
sample, indicating that the transformation of ZrO from metastable
2 2
/SBA-15 sample, besides tetragonal ZrO ,
2
2
À
2 2 3
–Al O /SBA-15
2
tetragonal phase to monoclinic phase is retarded in the presence of
the alumina element. This suggests that the introduction of the alu-
with ZrO
2
crystallites formed at high loading. The range of pore size
loading; this phenomenon
distribution is increased with higher ZrO
2
minum can stabilize the tetragonal phase of the ZrO
2
, which is an
can be explained that the pore wall restructuring might have taken
place at the time of calcination treatments.
The textural characteristics of the catalyst are depicted in Table
1. The BET surface area and pore volume of the samples decrease
ideal crystalline phase for exhibiting superacidity for sulfated zirco-
6
nia. The enhanced catalytic performance of the alumina-promoted
sulfatedzirconia was attributedto the retardationfrom a metastable
tetragonal zirconia phase to the thermodynamically favored mono-
gradually with the higher ZrO
the surface area decreases with increasing ZrO
may be due to pore blockage. It has been observed that the surface
2
loading. It is quite obvious that
1
4
15
clinic phase. Canton et al. have examined the stabilization effect
of alumina on the phase and size of zirconia with detailed Rietvelt
analysis of X-ray diffraction results. Many researchers reported that
the tetragonal phase of zirconia has more catalytic activity than that
2
loading, which
2
À
area of the SO
/12% ZrO
2
–Al
2 3
O /SBA-15 sample is higher than
4
2
À
that of the SO
the formation of Zr–O–Al linkages in the former sample.
/12% ZrO /SBA-15 sample, which is caused by
2
4
1
6,17
23
of the monoclinic phase.
tetragonal ZrO is a necessary condition for the formation of super-
acidity in sulfated zirconia and its analogs, and good catalytic
Consequently, the presence of a stable
2
3
3.3. NH –TPD
1
8,19
activity is considered to be connected to tetragonal zirconia.
.2. BET
The nitrogen adsorption–desorption isotherms of the SBA-15
The NH
3
–TPD measurement was performed to characterize the
–TPD spectra of
3
acid strength distribution of the catalysts. The NH
3
the samples are shown in Figure 4. All the samples show one major
desorption peak that appears around 125 °C, which is due to
desorption of NH bound to the surface of the samples. The neutral
3
2
À
2 2 3
support and the SO4 /ZrO –Al O /SBA-15 samples are shown in
Figure 3. All isotherms show typical type IV features and have
H1-type hysteresis loop, indicating that the mesoporous texture
can be largely maintained.11 The adsorption branch of isotherm
exhibits a sharp inflection at a relative pressure of ca. 0.65 for the
SBA-15 shows no desorption peak above 200 °C (Fig. 4a), indicating
2À
that there is no acidity on the SBA-15. All the SO4 /ZrO
–Al
2 2
O
3
/
SBA-15 samples afford broad desorption peaks in the range of
1
80 to 800 °C, suggesting a broad heterogeneity of acid site. The
parent SBA-15 support. It is a characteristic of capillary condensa-
peak below 600 °C is very broad and represents the acidic sites of
weak and medium strength and that peak above 600 °C
corresponds to NH released from strong acidic and super strong
3
tion within uniform pores.20 For the SO4 /ZrO
2À
–Al
O
3
/SBA-15 sam-
2
2
ples, there is a decrease in the inflection point on its adsorption

X. Shi et al. / Carbohydrate Research 346 (2011) 480–487
483
Table 1
Catalytic performance, physical characteristics, and elemental analysis data of the samples
a
BET (m2
À1
b
BJH (cm3
À1
c
Sd (%)
Catalyst
Conv. (%)
Sel. (%)
Yield (%)
A
g
)
V
g
)
D
BJH (nm)
None
37.2
38.7
98.5
12.9
13.5
41.6
4.8
5.2
41.0
—
584
—
—
1.09
—
—
8.09
—
—
—
—
SBA-15
SO42 /ZrO
À
–Al O
2 2 3
SO42 /6% ZrO
À
–Al
–Al
O
O
/SBA-15
/SBA-15
96.3
97.6
98.7
98.6
98.0
43.7
46.6
53.4
48.7
44.5
42.1
45.5
52.7
48.0
43.6
402
384
276
267
245
0.73
0.65
0.43
0.35
0.32
8.08
6.61
6.60
4.88
4.94
0.77
1.12
1.55
2.12
1.96
2
2
3
3
SO42 /9% ZrO
À
2
2
SO42 /12% ZrO
À
–Al
–Al
O
O
/SBA-15
/SBA-15
2
2
2
2
2
3
SO42 /15% ZrO
À
3
SO42 /12% ZrO
À
/SBA-15
Reaction conditions: 17.5 mL toluene, 7.5 mL water, 0.75 g xylose, 0.5 g catalyst, 4 h.
a
BET surface area.
Pore volume.
Pore size.
Sulfur content.
b
c
d
SBA-15 sample is larger than that of the SO42 /12% ZrO
À
/SBA-15
2
sample, suggesting that the former has a higher concentration of
2
7
acid sites. Tanabe et al. proposed that acidity of a mixed oxide
system is more than that of single oxide component due to the
charge imbalance imposed by the major oxide component upon
the minor oxide component. In the present study, the increase in
the acidity may be due to the formation of mixed oxides with inter-
action between Zr and Al atoms through oxygen linkages
(
Zr–O–Al), which complies with Tanabe’s model for mixed oxides.
3
3
.4. Catalytic activity
.4.1. Catalytic activity
All the catalysts were employed in the batch experiments for
the dehydration of xylose to furfural. The results obtained with tol-
uene and water as a mixed solvent at 160 °C are summarized in
Table 1. In the products, beside furfural, a few byproducts, such
as pyruvaldehyde, glycolaldehyde, and dihydroxyacetone were de-
tected. For the pure SBA-15 support, the xylose conversion was
roughly the same as that obtained in the homogeneous phase
without any catalyst, indicating that the SBA-15 support is essen-
2
À
2À
/
Figure 4. NH
3
–TPD of the samples: (a) SBA-15; (b) SO /ZrO
2
–Al
2
O
3
; (c) SO4
–Al /SBA-
2 2 3
/SBA-15; (g) SO4 /15% ZrO –Al O /SBA-15.
4
2
À
2À
1
1
2% ZrO
2
/SBA-15; (d) SO /6% ZrO
2
–Al
2
O
3
/SBA-15; (e) SO /9% ZrO
2
2 3
O
4
4
5; (f) SO42 /12% ZrO
À
–Al
2À
2
2
O
3
acidic sites.24–26 For the SO4 /ZrO
2À
O
3
/SBA-15 samples, the
number of the weak and medium strong acid sites decline gradu-
ally, while that of the strong and very strong acid sites increase
2
–Al
2
2À
tially inert (Table 1). For the SO4 /ZrO –Al O catalyst, the xylose
2
2 3
progressively with higher ZrO
2
content from 6 to 15 wt % (Table
conversion and furfural selectivity was lower than that of the
SO4 /12% ZrO –Al O /SBA-15 catalysts. This might also be attrib-
2 2 3
2À
2). This is understandable as strong acidity and superacidity origi-
nate from sulfated zirconia. As presented in Figure 4, the peak cor-
responding to the strong acid sites is shifted to higher temperature
uted to the higher concentrations of acid sites presented in the
2À
SO4 /12% ZrO –Al O /SBA-15 catalysts, as revealed by the NH –
2
2
3
3
2
À
in the case of the SO4 /12% ZrO
2
–Al
2
O
3
/SBA-15 sample compared
TPD results (Table 2). From Table 1, we can see that the content
of S has a good correlation with ZrO2 content. With increasing
ZrO2 content, the xylose conversion and furfural selectivity in-
creased and the maximum furfural yield could be obtained on
the catalyst with 12 wt % ZrO2 loading, on which the xylose
2
À
with that of the SO4 /12% ZrO
indicates that the addition of Al element to the SO4 /12% ZrO
2
/SBA-15 sample. This observation
2
À
2
/
SBA-15 sample leads to the stronger and stable acid sites.
À
Moreover, the NH
3
desorbed on the SO42 /12% ZrO
2
2
–Al O
3
/
Table 2
NH
3
–TPD data of the catalysts
Catalyst
NH
3
desorbed (mmol/g)
180–600 °C
600–800 °C
Total
SO42 /ZrO
À
–Al O
0.43
0.63
0.59
0.55
0.52
0.51
0.32
0.53
0.62
0.58
0.50
0.15
0.34
0.35
0.36
0.38
0.37
0.28
0.27
0.29
0.24
0.19
0.58
0.97
0.94
0.91
0.90
0.88
0.61
0.80
0.91
0.82
0.69
2 2 3
SO42 /6% ZrO
À
–Al
–Al
O
O
/SBA-15
/SBA-15
2
2
3
3
SO42 /9% ZrO
À
2
2
SO42 /12% ZrO
À
–Al
–Al
O
O
/SBA-15
/SBA-15
2
2
2
2
2
2
2
2
2
2
3
SO42 /15% ZrO
À
3
SO42 /12% ZrO
À
/SBA-15
SO42 /12% ZrO
À
–Al
–Al
–Al
–Al
–Al
O
O
O
O
O
/SBA-15 (used)
2
2
2
2
2
3
3
3
3
3
SO42 /12% ZrO
À
/SBA-15 (regeneration-calcined)
SO42 /12% ZrO
À
/SBA-15 (regeneration-H
/SBA-15 (regeneration-H
/SBA-15 (regeneration-H
-First)
2
O
2
O
2
O
2
2
2
SO42 /12% ZrO
À
-Second)
-Third)
SO42 /12% ZrO
À

4
84
X. Shi et al. / Carbohydrate Research 346 (2011) 480–487
conversion and furfural selectivity were 98.7% and 53.4%, respec-
precursors of coke formation, which are deposited on the catalyst
surface with an increase in reaction time, and then they decrease
the acid sites of the catalyst. At the same time, the accumulation
of byproducts on the catalyst surface leads to a decrease in pore
size and then to a slower diffusion of the reactant/products
tively. Further increasing the ZrO
2
content caused a slight decrease
in the conversion and selectivity. In the SO42 /15% ZrO
À
2
2
–Al O
3
/
SBA-15 sample, the presence of the basic character of the mono-
clinic ZrO (above XRD results) is in favor of furfural condensation
2
8
12
reaction, as a result, more deposition of condensation products
took place on the surface of the catalyst which led to a fast deacti-
vation. From Table 1, it also can be seen that the furfural yield of
through the porous structure. Therefore, allowing a longer resi-
dence time for secondary reactions may affect the distribution of
products. From Figure 5, it can be seen that the optimized reaction
temperature and time is 160 °C and 4 h, respectively, for the reac-
tion of xylose dehydration to furfural.
2
À
the SO /12% ZrO
2
/SBA-15 sample was less than that of the
/SBA-15 sample. This may be due to that
–Al /SBA-15 sample has higher concentra-
4
SO42 /12% ZrO
À
2 2 3
–Al O
the SO42 /12% ZrO
À
2
2 3
O
2
À
tions of acid sites than that of the SO4 /12% ZrO
In addition, the above XRD and NH –TPD results indicate that the
addition of Al species can stabilize the tetragonal ZrO phase and
2
/SBA-15 sample.
3.4.2. Deactivation and regeneration
3
In the dehydration of xylose to furfural, the degradation of acid
concentration and the accumulation of byproducts are the key fac-
tors that cause the deactivation of the catalyst. The catalytic per-
2
increase the number of acid sites in the catalysts, which can well
account for their different initial catalytic activity for the dehydra-
tion of xylose.
2À
formance over the SO /12% ZrO
2
–Al
/SBA-15 (regeneration-calcined) and the
/SBA-15 (regenerated-H ) (regenerated
) was investigated, and the results are shown in Table 3.
It can be seen that the S content and the NH desorption of the
2 3
O /SBA-15 (used), the
4
2À
SO4 /12% ZrO
2
–Al
2 3
O
2À
The effect of reaction temperature and time on the catalytic per-
SO4 /12% ZrO
by H
2
–Al
2
O
3
2 2
O
2À
formance of the SO4 /12% ZrO
–Al
2 2
O
3
/SBA-15 sample is shown in
2 2
O
Figure 5. The conversion of xylose increases with the higher reac-
tion temperature. At 180 °C, the xylose conversion is 99.6%.
Although the higher reaction temperature is beneficial to the xy-
lose dehydration to furfural, the furfural selectivity decreases when
the reaction temperature is above 160 °C. This might be due to the
coke deposits on the outer surface of the particle or within the
pores, which leads to a decrease in catalyst acidity.28 The reaction
time also has an important effect on the catalytic activity of the
catalyst. After 24 h reaction, the furfural selectivity decreased to
3
used catalyst decreases, compared with the fresh catalyst (Table
2). Thus, the decrease of the activity might be attributed to the
leaching of sulfur. In addition, it should be pointed out that the
deactivation of the catalyst is also observed when the catalyst
was reused without regeneration, and the yield of furfural de-
creases from 52.7% to 19.1%. At the same time, the activity of the
catalyst can be recovered by calcination and a furfural yield of
46.0% is obtained. This result indicates that the observed progres-
sive catalyst deactivation might be related to the accumulation of
organic matter, which might cover the surface of the catalyst, lead-
ing to the passivation of the acid site. It should be noted that the
4
2.7% at 99.8% xylose conversion, and the solution of the reaction
became darker. At the same time, the color of the catalyst became
darker with an increase in reaction time, which may be due to the
degradation of furfural by secondary loss reactions. The loss reac-
tions during the dehydration of xylose to furfural process are
2
À
activity of the SO4 /12% ZrO
calcined) sample is recovered insufficiently. The presence of the
monoclinic ZrO phase in the XRD pattern directly indicates the
reduction of surface acidity,
2 2 3
–Al O /SBA-15 (regeneration-
2
shown in Scheme 1. Furfural loss reactions involve both furfural
2
1
0,29
resinification and furfural condensation, which lead to the
formation of larger molecules. These larger molecules may act as
which accounts well for
the decreasing activity of the SO42 /12% ZrO
À
O
3
/SBA-15
2
–Al
2
Figure 5. Catalytic performance of the SO 2 /12% ZrO
À
O
3
/SBA-15 sample under different reaction temperature and time. Reaction conditions: 17.5 mL toluene, 7.5 mL
2
–Al
2
4
water, 0.75 g xylose, 0.5 g catalyst.

X. Shi et al. / Carbohydrate Research 346 (2011) 480–487
485
Resinification
SO42-/ZrO -Al O /SBA-15
2
2
3
O
O
O
OH
OH
-H O
2
OH
OH
o
CH
Toluene/water(160 C)
O
HO
OH
HO
OH
Condensation
Scheme 1.
Table 3
Catalytic performance, the physical characteristics, and elemental analysis data for the catalysts
a
BET (m2
À1
b
BJH (cm3
À1
c
Sd (%)
Catalyst
Conv. (%)
Sel. (%)
Yield (%)
A
g
)
V
g
)
D
BJH (nm)
SO42 /12% ZrO
À
O
O
O
O
O
O
/SBA-15
98.7
80.6
93.0
98.6
94.1
92.1
53.4
23.7
49.5
53.5
50.2
42.8
52.7
19.1
46.0
52.8
47.2
39.4
276
241
260
271
259
254
0.43
0.38
0.39
0.41
0.38
0.35
6.60
6.21
6.59
6.59
6.27
5.9
1.55
0.56
0.57
0.57
0.56
0.54
2
–Al
2
–Al
2
–Al
2
–Al
2
–Al
2
–Al
2
2
2
2
2
2
3
3
3
3
3
3
SO42 /12% ZrO
À
/SBA-15 (used)
SO42 /12% ZrO
À
/SBA-15 (regeneration-calcined)
SO42 /12% ZrO
À
/SBA-15 (regeneration-H
/SBA-15 (regeneration-H
/SBA-15 (regeneration-H
-First)
2
O
2
O
2
O
2
2
2
SO42 /12% ZrO
À
-Second)
-Third)
SO42 /12% ZrO
À
Reaction conditions: 17.5 mL toluene, 7.5 mL water, 0.75 g xylose, 0.5 g catalyst, 4 h.
a
BET surface area.
Pore volume.
Pore size.
Sulfur content.
b
c
d
(
regeneration-calcined) sample. In comparison, after first regener-
ation with H , the color of the used catalyst returned to white,
2 2
O
and the catalytic activity recovered completely (the xylose conver-
sion and furfural selectivity are 98.6% and 53.5%, respectively). In
addition, the NH
3
–TPD results also indicate that the acid strength
and amount of the acid sites in the SO42 /12% ZrO
À
/SBA-15
2
2
–Al O
3
2
À
(
regenerated-H -First) sample is higher than that of the SO4
2
O
2
/
1
2% ZrO –Al O /SBA-15 (regeneration-calcined) sample (Table 2).
2
2 3
Thus, the catalytic activity of the catalyst is strongly affected by
the acid strength and concentrations of the acid sites, which seem
to play an important role in catalytic activity. Comparing the above
two regeneration methods, the regeneration with H
2
O
2
is an effec-
tive method for the used SO42 /12% ZrO
À
–Al /SBA-15 sample.
2
2 3
O
The recyclability of the SO42 /12% ZrO
À
–Al O /SBA-15 catalyst
2
2 3
was further studied and the results are shown in Table 3. From
the second run, the sulfur is not further leaching from the catalyst
and the content is about 0.55% (Table 3). In sulfated zirconia cata-
lysts, the existence of covalent S@O bonds in sulfur complexes
formed on metal oxides is necessary for the generation of the acid-
ity. Therefore, the sulfur is accountable for the strong acid sites on
2À
Figure 6. Low-angle XRD patterns of the samples: (a) SO /12% ZrO
2
–Al
2
O
3
/SBA-
4
15; (b) SO42À/12% ZrO –Al O /SBA-15 (used); (c) SO 2À/12% ZrO –Al O /SBA-15
2
2
3
4
2
2 3
2
À
(regeneration-calcined); (d) SO
First); (e) SO4 /12% ZrO
2% ZrO –Al
/12% ZrO
/SBA-15 (regeneration-H
/SBA-15 (regeneration-H -Third).
2
–Al
2
O
3
/SBA-15 (regeneration-H
2
O
2
-
/
4
2
À
2À
2
–Al
2
O
3
2
O
2
-Second); (f) SO4
1
2
2
O
3
2 2
O
3
0
the catalyst. In the recycling study, XRD results indicated that the
monoclinic ZrO
monoclinic ZrO
2
was detected from the second run (Fig. 7). The
with basic character can decrease the acid
2
catalysts is partly changed after the reaction. From the high-angle
XRD patterns, it can be seen that the peak intensity of ZrO in the
used and regenerated samples have a slight increase compared
with that of the fresh sample (Fig. 7). This may be caused by the
strength and acid concentration of the catalyst. This agrees with
the NH –TPD results (Table 2). From Table 3, it can be seen that
2
3
the furfural yield of the catalyst decreased with more recycling
tests. The catalyst deactivation is mainly caused by the presence
of monoclinic ZrO
conglomeration of the ZrO
peaks of the tetragonal phase of ZrO
60.28° are observed on the used and the first regenerated catalyst
samples by H . The result suggests that the tetragonal phase of
ZrO does not transform to the cubic or monoclinic ZrO phase in
the first run. However, the monoclinic ZrO phase was detected
in the second and third regenerated catalysts. In addition, mono-
clinic ZrO phase also had been detected on the catalyst regener-
ated by calcination. The above results indicate that a part of the
tetragonal phase of ZrO transforms to the monoclinic phase of
ZrO in the process of the dehydration reaction and catalyst
regeneration.
2
in the reaction. Only the characteristic
2
, which decreases the acid strength and acid
2
at 2h = 30.28°, 50.40°, and
concentration of the catalyst.
The XRD patterns of the fresh, used and regenerated SO42À/12%
2 2
O
ZrO
2
2
–Al O
3
/SBA-15 samples are shown in Figures 6 and 7. For the
2
2
samples, one major characteristic peak of SBA-15 at about 0.9–1.1°
is observed. This result suggests that the pore structure of SBA-15
is retained in the used and regenerated catalysts. However, the
peak intensity of the used and regenerated catalysts decreases,
and the position of the peak shifts to a lower 2h value relative to
that of the fresh catalyst. This indicates that the regularity of the
hexagonal mesoporous structure of the used and regenerated
2
2
2
2

4
86
X. Shi et al. / Carbohydrate Research 346 (2011) 480–487
2
À
Figure 7. High-angle XRD patterns of the samples: (a) SO /12% ZrO
2
–Al
2
O
3
/SBA-
2À
2À
/SBA-15; (b) SO4 /
4
Figure 9. NH
12% ZrO –Al
calcined); (d) SO4 /12% ZrO
3
–TPD of the samples: (a) SO4 /12% ZrO
2
–Al
2
O
3
2À
2À
1
5; (b) SO4 /12% ZrO
2
–Al
2
O
3
/SBA-15 (used); (c) SO /12% ZrO
2
–Al
2
O
3
/SBA-15
2À
4
2
2
O
3
/SBA-15 (used); (c) SO4 /12% ZrO
/SBA-15 (regeneration-H
12% ZrO –Al O /SBA-15 (regeneration-H O -Second); (f) SO4 /12% ZrO –Al O
2 2 3 2 2 2 2 3
2
–Al
2
O
3
/SBA-15 (regeneration-
2
À
(
regeneration-calcined); (d) SO4 /12% ZrO
2
–Al
/SBA-15 (regeneration-H
/SBA-15 (regeneration-H -Third).
2
O
3
/SBA-15 (regeneration-H
2
O
2
-
/
2À
2À
2
–Al
2
O
3
2
O
2
2À
-First); (e) SO4
/
/
2
À
2À
First); (e) SO4 /12% ZrO
2% ZrO –Al
2
–Al
2
O
3
2
O
2
-Second); (f) SO4
1
2
2
O
3
2 2
O
2 2
SBA-15 (regeneration-H O -Third).
SO42 /12% ZrO –Al O /SBA-15 (regeneration-H O ) sample, the
À
The nitrogen adsorption–desorption isotherms of the fresh,
2
2
3
2 2
used, and regenerated SO42 /12% ZrO
À
O
3
/SBA-15 samples are
2
–Al
2
SBET, DBJH, and VBJH changed little compared with the fresh sample
displayed in Figure 8. All isotherms show typical type IV features
and have H1-type hysteresis loop, indicating that the mesoporous
texture can be largely maintained. Compared with that of the fresh
indicates that the regeneration by H O has little effect on the
structure of the catalyst.
2
2
The NH –TPD spectra of the fresh, used, and regenerated sam-
3
catalyst, the N
2
isotherm of the used catalyst has some change in
ples are shown in Figure 9. Before the reaction, the fresh catalyst
shape, wherein the hysteresis inflection is less sharp. This indicates
that the pore size of the material is less ordered and uniform.
Although nitrogen adsorption–desorption isotherms of the
regenerated catalyst recover to some extent, the opening of the
shows a broad desorption peaks in the range of 180–800 °C, indi-
cating a broad heterogeneity of acid sites. After reaction, the
amount of the desorbed NH3 decreases, suggesting that the acid
sites decrease. At the same time, the temperature of the desorption
peak at high temperature shifts to a lower temperature (655 °C)
compared with that of fresh catalyst (750 °C). This indicates that
the acid strength of the used catalyst deceases after reaction. After
regeneration of the used catalyst, the acid strength and the des-
0
hysteresis loop at lower P/P relative pressure is also observed, wit-
31
nessing the pore size diminution on the sample. The textural
properties of the fresh, used, and regenerated catalysts are given
in Table 3. For the used sample, the hysteresis loop exhibits defor-
mations, and the adsorbed volume decreases clearly compared
with that of the fresh catalyst. SBET and DBJH also show an obvious
decrease. This may be due to the fact that during the reaction, the
accumulated organic matter blocks the mesochannels and causes
some collapse of the pore walls. However, the hexagonal mesopor-
ous structure is still present in the used catalyst. The observed re-
sults agree well with those obtained by XRD as presented above.
orbed NH amount recovered insufficiently. The above results indi-
3
cate that the acid strength and the amount of acid sites changed
after reaction and regeneration. Compared with that of the
2
À
SO4 /12% ZrO –Al O /SBA-15 (regeneration-H O -First) sample,
2
2
3
2 2
2
À
the desorption peak at high temperature on the SO4 /12% ZrO2–
Al O /SBA-15 (regeneration-calcined) sample moves to low tem-
2
3
perature region. At the same time, the NH3 desorption content of
the latter is lower than that of the former. This suggests that both
the acid strength and the concentration of the acid sites on the
2
À
2 2 3
For the SO4 /12% ZrO –Al O /SBA-15 (regeneration-calcined)
sample, the SBET, DBJH, and VBJH recovered to some extent. The
above results suggest that the structure of the used sample after
regeneration by calcination is not completely recovered. For the
2
À
SO4 /12% ZrO –Al O /SBA-15 (regeneration-calcined) sample is
2
2 3
2
À
lower than those of the SO4 /12% ZrO –Al O /SBA-15 (regenera-
2
2 3
tion-H
monoclinic ZrO
stable tetragonal one, as a result, the acid strength and the concen-
2
O
2
-First) sample. This may be caused by the presence of
2
phase which is more basic than that of the meta-
2
À
tration of the SO4 /12% ZrO
2 2 3
–Al O /SBA-15 (regeneration-
calcined) sample are decreased.
4
. Conclusion
The SO42 /ZrO
À
2 2 3
–Al O /SBA-15 catalysts were prepared with
well-ordered mesoporous and super strong acidic sites. The addi-
2
À
tion of Al element to the SO4 /ZrO
the tetragonal ZrO phase and enhances the catalytic activity of
the catalysts. The 98.7% xylose conversion and 53.4% furfural selec-
2
/SBA-15 catalysts stabilizes
2
2
À
tivity can be obtained over the SO4 /12% ZrO
ple at 160 °C. The good catalytic performance of the SO4 /12%
ZrO –Al /SBA-15 catalyst is attributed to an integrated effect
2 2 3
–Al O /SBA-15 sam-
2
À
2
2 3
O
of a suitable amount of acid and super strong acid site. The loss
reaction of furfural is the main cause for the deactivation of the
Figure 8. Nitrogen sorption isotherms and pore size distribution of the sample:
2
À
2À
2 3 2 2 3
O /SBA-15; (b) SO4 /12% ZrO –Al O /SBA-15 (used); (c)
(
a) SO /12% ZrO
2
–Al
/SBA-15 (regeneration-calcined); (d) SO4 /12% ZrO
SBA-15 (regeneration-H -First).
4
2À
2
À
2À
2 2 3
SO4 /ZrO –Al O /SBA-15 catalyst. The regeneration results of
SO4 /12% ZrO
2
–Al
2
O
3
2 2 3
–Al O /
2 2
O
the used catalysts indicate that the catalytic activity of the catalyst

X. Shi et al. / Carbohydrate Research 346 (2011) 480–487
487
can be completely recovered by the H
2
O
2
regeneration. The acid
10. Dias, A. S.; Lima, S.; Pillinger, M.; Valente, A. A. Catal. Lett. 2007, 114,
51–160.
1
strength and the acid concentrations play an important role for
the dehydration activity of the catalysts. In addition, the hexagonal
mesoporous structure of SBA-15 is still present in the used and
11. Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.; Chmelka, B. F.; Stucky,
G. D. Science 1998, 279, 548–552.
12. Dias, A. S.; Pillinger, M.; Valente, A. A. J. Catal. 2005, 229, 414–423.
13. Zhao, J.; Yue, Y. H.; Hua, W. M.; Gao, Z. Catal. Lett. 2007, 116, 27–34.
2
À
2 2 3
regenerated SO4 /ZrO –Al O /SBA-15 catalysts, and the frame-
1
1
4. Gao, Z.; Xia, Y.; Hua, W.; Miao, C. Top. Catal. 1998, 6, 101–106.
5. Canton, P.; Olindo, R.; Pinna, F.; Strukul, G.; Riello, P.; Meneghetti, M.; Cerrato,
G.; Morterra, C.; Benedetti, A. Chem. Mater. 2001, 13, 1634–1641.
6. Yamaguchi, T. Appl. Catal. 1990, 61, 1–25.
work of SBA-15 has some constriction for the used catalyst.
1
1
Acknowledgments
7. Morterra, C.; Cerrato, G.; Pinna, F.; Signoretto, M. J. Catal. 1995, 157,
1
09–123.
18. Yang, X.; Jentoft, F. C.; Jentoft, R. E.; Girgsdies, F.; Ressler, T. Catal. Lett. 2002, 81,
5–31.
This project was supported by National Basic Research Program
of China (973 Program) (No. G2006CB705809), National Natural
Science Foundation of China (No. 20506011), and PetroChina Inno-
vation Foundation (No. 2010D-5006-0406).
2
1
2
9. Farcasiu, D.; Li, J. Q.; Cameron, S. Appl. Catal., A: Gen. 1997, 154, 173–184.
0. Yashima, M.; Arashi, H.; Kakihana, M.; Yoshimura, M. J. Am. Ceram. Soc. 1994,
77, 1067–1071.
21. Zhang, X.; Yue, Y. H.; Gao, Z. Catal. Lett. 2002, 83, 19–25.
22. Hua, W. M.; Yue, Y. H.; Gao, Z. J. Mol. Catal. A: Chem. 2001, 170, 195–202.
23. Reddy, B. M.; Reddy, G. K.; Rao, K. N.; Katta, L. J. Mol. Catal. A: Chem. 2009, 306,
References
6
2–68.
1
2
.
.
Chheda, J. N.; Leshkov, Y. R.; Dumesic, J. A. Green Chem. 2007, 9, 342–350.
Sugar Series; Zeitsch, K. J., Ed., 1st ed.The Chemistry and Technology of Furfural
and its Many By-products; Elsevier: The Netherlands, 2000; Vol. 13, pp 3–22.
Corma, A. Chem. Rev. 1995, 95, 559–614.
Okuhara, T. Chem. Rev. 2002, 102, 3641–3666.
Clark, J. H. Acc. Chem. Res. 2002, 35, 791–797.
Yadav, G. D.; Nair, J. J. Microporous Mesoporous Mater. 1999, 33, 1–48.
Tyagi, B.; Mishra, M. K.; Jasra, R. V. Catal. Commun. 2006, 7, 52–57.
Yu, G. X.; Zhou, X. L.; Tang, C.; Li, C. L.; Wang, J. A.; Novaro, O. Catal. Commun.
24. Yadav, G. D.; Murkute, A. D. P. Adv. Synth. Catal. 2004, 346, 389–394.
25. Benitez, V. M.; Yori, J. C.; Vera, C. R.; Pieck, C. L.; Grau, J. M.; Parera, J. M. Ind.
Eng. Chem. Res. 2005, 44, 1716–1721.
26. Chen, W. H.; Ko, H. H.; Sakthivel, A.; Huang, S. J.; Liu, S. H.; Lo, A. Y.; Tsai, T. C.;
Liu, S. B. Catal. Today 2006, 116, 111–120.
27. Tanabe, K.; Sumiyoshi, T.; Shibata, K.; Kiyoura, T.; Kitagawa, J. Bull. Chem. Soc.
Jpn. 1974, 47, 1064–1066.
28. Lourvanij, K.; Rorrer, G. L. Ind. Eng. Chem. Res. 1993, 32, 11–19.
29. Li, X.; Nagaoka, K.; Olindo, R.; Lercher, J. A. J. Catal. 2006, 238, 39–45.
30. Sohn, J. R. J. Ind. Eng. Chem. 2004, 10, 1–15.
3.
4.
5.
6.
7.
8.
2
008, 9, 1770–1774.
Qi, X.; Watanabe, M.; Aida, T. M.; Smith, R. L., Jr. Catal. Commun. 2009, 10, 1771–
775.
9
.
31. Anderson, J. A.; Fergusson, C.; Rodríguez-Ramos, I.; Guerrero-Ruiz, A. J. Catal.
2000, 192, 344–354.
1