High-activity, single-site mesoporous WO3-MCF materials for the catalytic epoxidation of cycloocta-1,5-diene with aqueous hydrogen peroxide

Article abstract of DOI:10.1016/j.jcat.2008.03.017

For the first time, the high-activity, single-site mesoporous WO₃-MCF materials were synthesized and characterized by N₂ sorption, TEM, UV-vis DRS, UV-Raman, and XPS. It was found that the dispersion and nature of the tungsten species depend strongly on the tungsten oxide content and the support characteristic. The novel catalyst remains a highly ordered mesostructure of the silica support. The catalytic performance of the materials in the epoxidation of cycloocta-1,5-diene with aqueous H₂O₂ was investigated. The excellent catalytic performance of WO₃-MCF in the selective oxidation of cycloocta-1,5-diene was attributed to the presence of isolated tetrahedral {WO₄} species and the unique pore structure. The novel catalyst can be easily recycled after reaction and reused many times with no significant loss of activity. The good stability can be attributed to the presence of isolated tungsten species anchored on the support through W{single bond}O{single bond}Si covalent bonds.

Full text of DOI:10.1016/j.jcat.2008.03.017

Journal of Catalysis 256 (2008) 259–267
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
High-activity, single-site mesoporous WO3-MCF materials for the catalytic
epoxidation of cycloocta-1,5-diene with aqueous hydrogen peroxide
Ruihua Gao a, Xinli Yang a,b, Wei-Lin Dai a,∗, Yingyi Le a, Hexing Li c, Kangnian Fan a
a
Department of Chemistry and Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Fudan University, Shanghai 200433, PR China
School of Chemistry and Chemical Engineering, Henan University of Technology, Henan 450052, PR China
Department of Chemistry, Shanghai Normal University, Shanghai 200234, PR China
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
For the first time, the high-activity, single-site mesoporous WO3-MCF materials were synthesized and
characterized by N2 sorption, TEM, UV–vis DRS, UV-Raman, and XPS. It was found that the dispersion
and nature of the tungsten species depend strongly on the tungsten oxide content and the support char-
acteristic. The novel catalyst remains a highly ordered mesostructure of the silica support. The catalytic
performance of the materials in the epoxidation of cycloocta-1,5-diene with aqueous H2O2 was investi-
gated. The excellent catalytic performance of WO3-MCF in the selective oxidation of cycloocta-1,5-diene
was attributed to the presence of isolated tetrahedral {WO4} species and the unique pore structure. The
novel catalyst can be easily recycled after reaction and reused many times with no significant loss of
activity. The good stability can be attributed to the presence of isolated tungsten species anchored on the
support through W–O–Si covalent bonds.
Received 4 December 2007
Revised 20 March 2008
Accepted 20 March 2008
Available online 24 April 2008
Keywords:
Isolated tetrahedral {WO4} species
Mesocellular silica foam (MCF) catalyst
Epoxidation
Cycloocta-1,5-diene
© 2008 Elsevier Inc. All rights reserved.
1. Introduction
materials in which the active tungsten species are well isolated
and anchored on the support through W–O–Si covalent bonds.
It is commonly accepted that the catalytic behavior of the sup-
Tungsten-containing materials have been widely used in the
metathesis and isomerization of alkenes [1], selective oxidation
of unsaturated compounds [2], dehydrogenation of alcohols [3],
and hydrodesulfurization and hydrocracking of heavy fractions in
petroleum chemistry [4,5]. In addition, many research groups have
attempted to synthesize mesoporous tungsten oxide materials [6]
or to incorporate tungsten into nonsiliceous mesostructured ma-
terials, including TiO2 [7,8] and siliceous mesoporous molecu-
lar sieves (e.g., silica, M41S, SBA-n) by impregnation [9], coge-
lation [10], insertion [11–14], or grafting [15]. But the structure
of the nonsupported tungsten oxide mesoporous materials is not
thermally stable and will collapse on high-temperature calcination.
Meanwhile, tungsten-containing mesoporous molecular sieves also
have some serious drawbacks, including (1) difficult-to-control dis-
persion of the tungsten species, allowing the preferential formation
of tiny tungsten oxide particles even at relatively low concentra-
tions; (2) structural collapse of the mesoporous materials after
high-temperature calcinations and/or after a certain amount of
metal loading; and (3) an inherent leaching problem, which causes
confusion as to the heterogeneity of the catalysts. To minimize this
leaching problem, much effort has been focused on synthesizing
ported tungsten oxides is sensitive to the intrinsic nature of the
tungsten species, including oxidation state, coordination circum-
stance, dispersion, and stability. On the other hand, the chemical
and structural features of the supports affect catalytic performance
either indirectly, by influencing tungsten dispersion and stabiliza-
tion of the tungsten oxide species, or directly, by forming a shape-
selective catalyst. Therefore, it is very important to study the effect
of the supports and the relationship between the catalytic perfor-
mance of the catalysts and the tungsten oxidation state, coordina-
tion circumstance, dispersion and stability.
There are some other oxidation catalysts with isolated active
sites. The dispersion and the nature of the vanadium oxide species
also affect the catalytic behavior of VOx-based catalysts. Berndt
et al. [16] reported that isolated vanadium oxide species are ac-
tive sites for the partial oxidation of methane. They also reported
that vanadium species catalyzes the oxidative dehydrogenation of
propane in both valence states, V5+ and V4+; however, V4+ seems
to be more selective, although less active, than V5+, due to its
lower oxidation potential [17]. Fe-ZSM-5 zeolites are recognized as
highly efficient catalysts for numerous reactions; however, debate
on the nature of the active Fe sites in these catalysts is ongoing
[18,19]. This is due mainly to the fact that in most cases, depending
on the preparation method and the Fe content, various coexisting
Fe species are created, ranging from isolated Fe ions via dimers and
small oligonuclear FexOy clusters inside the pores to large Fe2O3
Corresponding author. Fax: +86 21 65642978.
*
E-mail address: wldai@fudan.edu.cn (W.-L. Dai).
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.03.017

260
R. Gao et al. / Journal of Catalysis 256 (2008) 259–267
particles on the external surface. According to Santhosh Kumar et
al. [20], for both NH3 and i-butane-SCR (selective reduction of NO),
the concentration of accessible Fe3+ that can participate in a re-
versible redox cycle should be maximized, whereas the formation
of FexOy clusters must be completely (for i-butane-SCR) or largely
(for NH3-SCR) avoided, indicating that Fe3+ may play a major role
in the catalytic cycle.
only with different volumes of sodium tungstate (Na2WO4·2H2O,
0.2 M) aqueous solution. Here 10, 20, and 30% represent the the-
oretical weight ratio of WO3 to silica over the materials before
treatment with AMA. ICP analysis indicated an Si/W molar ratio
of 942 for 10%WO3-MCF-1 after AMA treatment (compared with
55 before AMA treatment), suggesting that more than 94% of the
loosely bonded tungsten species were removed by AMA solution.
In a typical procedure for the synthesis of 10%WO3-SBA-15-1
In recent years, there has been considerable interest in the
catalytic epoxidation of alkenes, which is an efficient and acces-
sible approach to the intermediates and precursors of many useful
chemical products. Epoxidation of cycloocta-1,5-diene (COD) has
been investigated using several metal catalysts under both ho-
mogeneous [21–24] and heterogeneous [25] conditions. Pillai et
al. [26] prepared effective catalyst systems for COD epoxidation
using hydrogen peroxide. Although these systems offer good cat-
alytic activity in the epoxidation of COD, their use of acetonitrile
is a significant limitation. In previous work [27], we found that
WO3-containing mesocellular silica foam (MCF) catalysts showed
high activity in the O-heterocyclization of COD, but that detectable
leaching of tungsten species from 10%WO3-MCF occurred, accord-
ing to ICP-AES analysis. It is known that leaching of tungsten
species is inevitable in most reactions with H2O2; therefore, there
is a strong impetus to find a more stable catalyst with no leach-
ing of tungsten species, to avoid the generation of heavy metal-
containing wastewater. In previous studies, we identified several
tungsten species on the catalyst surface, including isolated tetra-
hedral {WO4} species, low-condensed polymeric tungsten oxide
species, high-condensed polymeric tungsten oxide species, and
small particles of crystalline WO3; however, which of these is
the most important for the oxidation with H2O2 remains unclear.
Based on the widely accepted concept that the isolated tetrahedral
{WO4} species must be the most simple and stable in most reac-
tions, we have designed a novel isolated tetrahedral {WO4} species
containing WO3-MCF catalyst.
catalyst, first 5 g of Pluronic P123 triblock polymer (EO20PO70EO20
,
Mav = 5800, Aldrich) and 28 g of distilled water were added to
◦
150 mL of 2 M HCl and stirred for 4 h at 40 C. After 10 g of TEOS
was added and the mixture was stirred for another 30 min, 6.2 mL
of an aqueous solution of sodium tungstate (Na2WO4·2H2O, 0.2 M)
◦
was added. The resulting mixture was aged at 40 C under moder-
◦
ate stirring for 24 h, and then crystallized at 95 C for 3 days. The
solid product was filtered, washed with distilled water, and then
◦
dried at room temperature. After calcination at 500 C in air for
5 h to remove the template, the product was treated with 1 M
◦
AMA at 80 C for 6 h (designated 10%WO3-SBA-15-1) [28].
The impregnation method-derived 10%WO3/MCF-1, 10%WO3/
SBA-15-1, and 10%WO3/MCM-41-1 catalysts were synthesized as
follows. The prespecified amount of ammonium tungstate was
dissolved in an aqueous solution of ammonia. The MCF materi-
als (prepared following a procedure similar to that reported by
Schmidt-Winkel et al. [29]), SBA-15 [30,31], and MCM-41 [32],
◦
were added into the stirred solution at 60 C. Then the excessive
water was completely evaporated at the same temperature, and
◦
the catalyst was finally obtained after calcination at 550 C in air
◦
for 3 h and then treated with 1 M AMA at 80 C for 6 h (designated
10%WO3/MCF-1, 10%WO3/SBA-15-1, and 10%WO3/MCM-41-1) [28].
2.2. Characterization
The specific surface area, pore volume, and the pore size distri-
bution (PSD) of the samples were measured and calculated accord-
ing to the BET method on a Micromeritics TriStar 3000 analyzer
In the present work, to gain insight into the key role of isolated
tetrahedral {WO4} species on MCF materials in the catalytic oxi-
dation of COD with H2O2, we synthesized the isolated tetrahedral
{WO4} species-doped MCF materials via different methods and sys-
tematically characterized these materials by various analytical and
spectroscopic techniques, including N2 sorption, TEM, UV–vis DRS,
UV-Raman, and XPS.
◦
with liquid nitrogen at −196 C. The BJH method was used to de-
termine PSD. Transmission electron microscopy (TEM) images were
obtained with a Joel JEM 2010 microscope.
The UV-Raman measurements were carried out on a confocal
microprobe Jobin Yvon LabRam Infinity Raman system using the
UV line at 325 nm from a Kimmon IK3201R-F He-Cd laser as the
exciting source. The laser output was 30 mW, and the maximum
incident power at the sample was approximately 6 mW. UV–vis
DRS spectra were collected on a Shimadzu UV-2540 spectrometer
with BaSO4 as a reference.
2. Experimental
2.1. Catalyst preparation
In a typical procedure for the synthesis of WO3-MCF cata-
The XPS spectra were recorded on a Perkin–Elmer PHI 5000C
ESCA system equipped with a dual X-ray source, using the MgKα
lyst [27], first 4.5 g of Pluronic P123 triblock polymer (EO20PO70
-
EO20, Mav = 5800, Aldrich) was added to 170 mL of 1.6 M HCl and
(1253.6 eV) anode and a hemispherical energy analyzer. The back-
◦
−6
stirred for 4 h at 40 C, and then 4.5 g of 1,3,5-trimethybenzene
ground pressure during data acquisition was kept below 10
Pa.
(TMB) was added to the solution and stirred for
2
h. After
Measurements were performed at pass energy of 93.90 eV to en-
sure sufficient sensitivity for the acquisition scan; a pass energy of
23.50 eV was used for the scanning of the narrow spectra of Si2p,
W4f, O1s, and C1s to ensure sufficient resolution. All binding ener-
gies were calibrated using contaminant carbon (C1s = 284.6 eV) as
a reference.
that, 9.9 g of Si(OC2H5)4 (TEOS) was added to the mixture; af-
ter stirring for another 15–20 min, 6.2 mL of sodium tungstate
(Na2WO4·2H2O, 0.2 M) aqueous solution was added. The resulting
◦
mixture was then aged at 40 C under moderate stirring for 24 h
◦
and crystallized at 100 C for another 24 h. The resulting solid
product was filtered, washed with distilled water, dried at room
temperature, and then calcined at 500 C in air for 8 h to remove
The tungsten content was determined by inductively coupled
argon plasma analysis (ICP; IRIS Intrepid, Thermo Elemental Com-
pany) after the samples were solubilized in an HF:HCl solution.
◦
the template. The resulting material, designated WO3-MCF-as-syn,
was finally treated with 1 M of ammonium acetate (AMA) at
◦
80 C for 6 h (designated 10%WO3-MCF-1) [28]. Under such condi-
2.3. Activity test
tions, the tungsten percentage in the solution and in the catalyst
rapidly achieved a balance when the catalyst was treated with
AMA. A sample treated twice is designated 10%WO3-MCF-2; one
treated three times is designated 10%WO3-MCF-3. The 20%WO3-
MCF and 30%WO3-MCF were synthesized by the same method,
◦
The activity test was performed at 60 C with magnetic stir-
ring in a closed 25-mL regular glass reactor using 50% aqueous
H2O2 as the oxygen donor and t-BuOH as the solvent. In a typi-
cal experiment, 0.53 mL of COD (3.6 mmol), 10 mL of t-BuOH, and

R. Gao et al. / Journal of Catalysis 256 (2008) 259–267
261
Scheme 1. Main-product and byproducts from the epoxidation of cyclooc-
ta-1,5-diene by aqueous H2O2.
0.2 g of the WO3-MCF-1 material (10%) were introduced into the
◦
regular glass reactor at 60 C and stirred vigorously. The reaction
was initiated by the addition of 0.7 mL of 50 wt% aqueous H2O2
(11.3 mmol) into the mixture, which was held for 24 h. The main
product (designated MD) and byproducts from the epoxidation of
COD by aqueous H2O2 are listed in Scheme 1. Quantitative analy-
sis of the reaction products was performed by gas chromatography
(GC) and the identification of different products in the reaction
mixture was determined by GC-mass spectrometry (GC-MS) on an
HP 6890GC/5973 mass spectrometer.
(a)
3. Results and discussion
3.1. Catalyst characterization
3.1.1. Pore size distribution from nitrogen adsorption
Fig. 1 shows that the nitrogen adsorption/desorption isotherms
of 10%WO3-MCF-1 and 10%WO3/MCF-1 were of type IV and ex-
hibited steep hysteresis of type H1 at high relative pressures.
The pore size distribution of these two samples unambiguously
demonstrates preservation of the 3-D mesocellular structure of
the support. Meanwhile, the nitrogen isotherms of 10%WO3-SBA-
15-1 and 10%WO3/SBA-15-1 also were type IV, and the pore size
distribution of these two samples clearly indicated retention of
the typical 2-D pore structure of SBA-15. The 10%WO3/MCM-41-
1 also exhibited reversible hysteresis-free type IV isotherms, and
the pore size distribution clearly showed the 2-D pore structure
of MCM-41. The pore size of these samples had the following hi-
erarchy: 10%WO3-MCF-1 ∼ 10%WO3-MCF-1 > 10%WO3-SBA-15-1 ∼
10%WO3/SBA-15-1 > 10%WO3/MCM-41-1.
(b)
3.1.2. Surface morphology from TEM measurements
TEM images (Fig. 2) of 10%WO3-MCF-1 (Fig. 2a) and 10%WO3/
MCF-1 (Fig. 2b) materials reveal a disordered array of silica struts
composed of uniform-sized spherical cells (18–24 nm) intercon-
nected by windows with a narrow size distribution. This is a char-
acteristic structural feature of the MCF materials [29,33].
3.1.3. Isolated sites by UV–vis DRS
The DRS spectra obtained in the UV–vis region of the as-
synthesized WO3-containing MCF samples shown in Fig. 3a pro-
vide information on the chemical nature and coordination states
of tungsten oxide species. This technique is known to be highly
sensitive in distinguishing incorporated metal oxide and extra-
framework metal oxides in different mesostructures [34]. For com-
parison, the figure also shows UV–vis DRS spectra of sodium
tungstate (Na2WO4·2H2O) and bulk WO3. The spectrum of sodium
tungstate, with a spinel structure and isolated [WO4]2− tetrahedral
species, is characterized by a maximum at 230 nm (curve a) [35].
For pure mesocellular silica MCF, no bands are evident in the spec-
trum. After introduction of tungsten oxide into MCF, three Raman
bands appeared at 230, 290, and 430 nm. The broad band at about
430 nm of the 30 wt% WO3-MCF-as-syn. and 20 wt% WO3/MCF-as-
syn. samples can be attributed to tungsten trioxide by comparing
it with the spectrum of bulk WO3 [36]. The second broad band at
about 290 nm represents another kind of tungsten oxide species.
(c)
Fig. 1. Nitrogen adsorption–desorption isotherms for the novel materials (a); pore
size distribution for various samples based on a BJH method, desorption branches
for pore distribution (b); adsorption branches for window distribution (c). The ratio
of WOx to silica over all materials was 10%.
Weber et al. reported that the low-energy absorption edge shifted
toward a lower wavelength with decreasing nuclearity of molybde-
num and tungsten entities [37]; therefore, this broad band should
be assigned to isolated tungsten species or low-condensed poly-
meric tungsten oxide species. The sharp band at 230 nm of the
in situ synthesized sample can be attributed to isolated [WO4]2−
tetrahedral species by comparing it with the structure of sodium

262
R. Gao et al. / Journal of Catalysis 256 (2008) 259–267
Fig. 2. TEM images of different WO3-MCF materials: 10%WO3-MCF-1 (a) and
10%WO3/MCF-1 (b).
tungstate. The absence of a broad band at 430 nm may reflect the
high dispersion of tungsten oxide species and lack of crystalline
WO3 formation in WO3-containing MCF samples with tungsten ox-
ide content <20 wt%. A weak band at 430 nm for the 20 wt%
WO3-MCF sample indicates the formation of only a small amount
of low crystalline WO3 species. Further increases in the tungsten
oxide content up to 30 wt% will lead to the formation of a large
quantity of crystalline WO3 species, resulting in the appearance of
a strong band at 430 nm. The UV–vis DRS spectra further demon-
strate that the chemical nature and coordination states of tungsten
oxide species are determined by the WO3 loadings.
(a)
The DRS spectra recorded in the UV–vis region of the AMA-
treated WO3-containing MCF samples given in Fig. 3b show signif-
icant differences between the spectra of the AMA-treated samples
and the as-synthesized samples. The 10%WO3-MCF-1 sample ex-
hibited only one band at 230 nm, attributed to isolated [WO4]2−
tetrahedral species. For the 20%WO3-MCF-1, two Raman bands ap-
peared at 230 and 290 nm and the weak band at 430 nm dis-
appeared, indicating removal of a small amount of low-crystalline
WO3 species. For the 30%WO3-MCF-1, the band at 430 nm was
much less intense than that for the as-synthesized sample, indi-
cating removal of most crystalline WO3 species by the AMA treat-
ment. Similar investigations by UV–vis spectroscopy reported by
Berndt et al. [16] showed that the vanadium species on 2.8V/MCM-
41 were monomeric and oligomeric tetrahedrally coordinated VV
species, as well as square pyramidal and distorted octahedral
VV species. After the sample was heated in flowing nitrogen to
823 K, only monomeric and oligomeric tetrahedrally coordinated
VV species remained.
(b)
DRS spectra recorded in the UV–vis region of the as-synthesized
and AMA-treated 10%WO3/MCF samples are shown in Fig. 3c. For
the as-synthesized 10%WO3/MCF, three Raman bands appeared at
230, 290, and 430 nm; the very weak peak at 430 nm indicates
the presence of high-condensed polymeric tungsten oxide species.
Only two bands, at 230 and 290 nm, remained on the sample af-
ter one AMA treatment, indicating removal of the high-condensed
polymeric tungsten oxide species. After two AMA treatments, the
peak of 290 nm was much weaker, demonstrating removal of the
low-condensed polymeric tungsten oxide species. After three AMA
treatments, the two bands at 230 and 290 nm were similar to
those of the 10%WO3/MCF-2, indicating that the remaining tung-
sten oxide species were the isolated tetrahedral [WO4]2− species
anchored on the support through W–O–Si covalent bonds, which
cannot be removed by AMA treatment.
(c)
Fig. 3. Diffuse reflectance spectra in the UV–vis region of the as-synthesized
WO3-containing MCF samples (a); the AMA-treated WO3-containing MCF sam-
ples (b); the as-synthesized and AMA-treated 10%WO3/MCF samples (c).
3.1.4. Tungsten percent by ICP
Fig. 4 illustrates the dependence of the residual tungsten per-
cent of the AMA-treated catalysts on the tungsten content of the
as-synthesized WO3-MCF catalysts. As shown, the residual tung-
sten percent increased rapidly initially and maintained a fixed
value of 0.30 from the value at a tungsten loading of 3.94% for
the as-synthesized catalyst. This finding indicates that the MCF
has a fixed maximal value (0.30%) and that the tungsten can be
incorporated very stably into the MCF framework. Based on this
interesting result, we can conclude that for these novel catalysts,
the residual tungsten species are the isolated tetrahedral [WO4]2−
species anchored on the support through W–O–Si covalent bonds.
The loosely bonded tungsten species are completely removed by

R. Gao et al. / Journal of Catalysis 256 (2008) 259–267
263
Table 3
ICP results of the 10%WO3/MCF over the different AMA-treated number
Catalyst
W (%)
Si (%)
Si/W (mol)
10%WO3/MCF-1
10%WO3/MCF-2
10%WO3/MCF-3
1.1
0.70
0.72
42.9
42.8
41.6
253
403
390
Fig. 4. Dependence of the residual tungsten percent of the AMA-treated catalysts on
the tungsten content of the as-synthesized WO3-MCF catalysts.
Table 1
ICP results of the different catalysts over the different supports
Catalyst
W (%)b
W (%)a
Si/W (mol)b
Si/W (mol)a
10%WO3-MCF-1
10%WO3/MCF-1
10%WO3-SBA-1
10%WO3/SBA-1
10%WO3/MCM-1
0.30
1.11
0.29
0.80
1.20
4.61
6.75
4.02
6.74
6.77
941.7
253
975.9
340.5
244.2
54.5
41.5
65.4
41.5
41.4
Fig. 5. UV-Raman spectra of the WMCF materials (*) represent the fundamental vi-
bration of {WO4}, the other modes (") represent the fundamental vibration of silica.
a higher percentage of tungsten remaining on the catalyst, and
also that no matter how much tungsten was added, the percent of
tungsten incorporated into the MCF framework remained a fixed
value, and two AMA treatments are sufficient to obtain the most
stable single-site tungsten-containing catalyst.
a
The as-synthesized value.
The AMA-treated value.
b
Table 2
ICP results of the different catalysts over the different WO3 loadings
The ICP results for the 10%WO3/MCF catalyst after different
AMA treatments, given in Table 3, show that the tungsten percent-
age remained constant after two AMA treatments. This indicates
that the percent tungsten incorporated into the MCF framework re-
mained fixed at a higher level in the impregnation-derived catalyst
than the in situ-derived one. It is interesting to note that although
over the impregnated catalyst more tungsten was incorporated
into the silica framework, the selectivity in the COD selective ox-
idation was much lower, suggesting that the remaining tungsten
species were not the same as single-site {WO4} tetrahedral even
though they cannot be removed by two or more AMA treatments.
This finding agrees well with the conclusions drawn from Tables 1
and 2 and our UV–vis DRS findings. Thus, we can conclude that
the percentage of tungsten incorporated into the silica support is
determined by the preparation method, and that only the in situ
method can lead to the perfect single-site {WO4} tetrahedral.
Catalyst
W (%)b
W (%)a
Si/W (mol)b
Si/W (mol)a
10%WO3-MCF-1
10%WO3-MCF-2
20%WO3-MCF-1
20%WO3-MCF-2
30%WO3-MCF-1
30%WO3-MCF-2
0.3
0.3
0.5
0.3
1.7
0.3
4.61
0.3
11.1
0.5
14.9
1.7
941.7
943
589.0
1062
179.0
998
54.5
941.7
22.7
589.0
14.2
179.0
a
The as-synthesized value.
The AMA-treated value.
b
the AMA treatment. In addition, a tungsten loading of about 4%
was sufficient to obtain an efficient single-site tungsten-containing
catalyst, suggesting that further tungsten loading would lead only
to loosely bounded tungsten species. Even if the 4% loading one,
only 7.6% tungsten species were effective and contributed to the
isolated and stable covalent bonded {WO4} species. This find-
ing also demonstrates that most of the tungsten species in the
tungsten-containing silica catalysts, even those prepared by the in
situ method, were not very stable and could be readily eluted by
the AMA treatment.
Table 1 shows that the tungsten percent was much lower in
the AMA-treated catalyst than in the as-synthesized catalyst in
all cases studied, and that the tungsten percent was lower in the
AMA-treated catalysts synthesized by the in situ method than in
those synthesized by impregnation. However, the similar tungsten
percent in the 10%WO3-MCF-1 and 10%WO3-SBA-1 samples in-
dicate that different silica supports had the same percentage of
tungsten stably incorporated into the silica framework.
3.1.5. Single sites by UV-Raman
The UV-Raman spectra shown in Fig. 5 provide further insight
into the exact molecular structure of our AMA-treated WO3-MCF
−1
materials. The bands at 320, 382, 800, and 970 cm
are assigned
to the vibration of the isolated {WO4} structure [38], whereas
−1
those at 433, 480, 891, and 1078 cm
are associated with vibra-
tions of the siloxane rings of MCF [39]. Clearly, the tungsten clus-
ters on the surface are single-site {WO4} species bound strongly to
silica through W–O–Si covalent bonds, as illustrated in Scheme 2.
Similarly, Berndt et al. [16] detected VIV species incorporated into
the siliceous walls by ESR. The large amounts of loosely bonded
tungsten species on the surface were readily removed on AMA
treatment, and the oxotungsten tetrahedra centers into the frame-
work maintained the structure of pure MCF. The UV-Raman results
provide direct evidence of the presence of single-site tetrahedra
{WO4} species.
The ICP results for the catalysts with different WO3 loadings,
given in Table 2, demonstrate that the tungsten content of these
catalysts differed after one AMA treatment but were similar af-
ter two AMA treatments. This finding suggests that for the in situ
method, one AMA treatment could not remove all of the loosely
bound tungsten species, and that higher tungsten loading led to

264
R. Gao et al. / Journal of Catalysis 256 (2008) 259–267
Table 4
Peak-fitting results of W4f region for the as-synthesized different WO3 loading cat-
alysts and the AMA-treated catalysts
Sample
Binding energy for W4f (eV)
W6+ W6+ W5+ W5+
4f5/2 4f7/2 4f5/2 4f7/2 4f5/2 4f7/2
W4+
W4+
W6+/W5+a
b
WO3
38.0
–
36.0
–
–
–
–
–
–
–
–
36.3
35.8
–
–
–
–
–
34.5
34.0
–
–
b
WO2.5
36.8
36.8
36.8
36.8
37.4
36.9
35.0
35.0
35.0
35.0
35.6
35.1
10%WO3-MCF-as syn. 38.4
20%WO3-MCF-as syn. 38.4
30%WO3-MCF-as syn. 38.5
20%WO3-MCF-1
30%WO3-MCF-1
36.6
36.6
36.7
36.6
36.3
1.7
1.8
2.0
1.0
0.8
38.4
38.1
a
Calculated according to the curve-fitting results of the W4f XPS spectra of cata-
lysts.
b
See reference: Appl. Catal. A 269 (2004) 169–177.
Table 5
Epoxidation of COD over the different catalysts
−1
Catalyst
SMD (%)
TOFa (h
)
H2O2 utilityb (%)
Scheme 2. Incorporating catalytic oxotungsten tetrahedra centers into the frame-
work maintains the structure of MCF.
10%WO3-MCF-1
10%WO3/MCF-1
10%WO3-SBA-15-1
10%WO3/SBA-15-1
10%WO3/MCM-41-1
92.7
68.5
88.2
79.8
81.0
17.8
4.3
10.4
2.2
43.6
20.2
22.9
22.1
8.2
0.56
◦
Reaction conditions: reaction temperature 60 C, COD: 0.53 mL, H2O2: 0.7 mL, reac-
tion time 24 h, t-BuOH: 10 mL, cat.: 0.2 g.
a
Calculated according to the equation: TOF = moles of COD converted per mole
of WO3 per hour.
b
H2O2 utility (%) = (CCOD × SMD)/3CH
(CCOD: conversion of H2O2; SMD: selec-
2O2
tivity of COD epoxide; CH : conversion of H2O2).
2O2
sults from the peak-fitting results of W4f are given in Table 4. The
measured spectra appear to be similar for all of the as-synthesized
samples and show identical positions of the W4f peaks, except for
the minor charging effect observed and corrected for according to
the contaminant carbon (C1s = 284.6 eV). Both W(VI) and W(V)
species at W4f7/2 (of 36.6 and 34.6 eV [40,41]) for the W4f spin–
orbit components were detected for the as-synthesized samples.
It can be seen that the intensity of the W4f XPS peaks also in-
creased with increasing tungsten oxide loading, as illustrated by
the changes in relative peak areas for W4f shown in Fig. 5a. More-
over, the intensity of the W4f XPS peaks was much weaker in
the AMA-treated samples than in the as-synthesized samples. It is
also very interesting that the W(IV) species appeared after AMA
treatment, according to the peak-fitting results of W4f, indicat-
ing that all of the loosely bound tungsten species were removed,
and that the remaining species were the most stable with isolated
structures. According to findings of our previous work, single-site
tungsten species can be present with an oxidation state of IV or
V [42].
(a)
As also shown in Table 4, the W6+/W5+ molar ratio increased
from 1.7 to 2.0 as the WO3 content rose from 10 to 30%, suggesting
a higher content of surface W6+ species and much greater forma-
tion of crystalline WO3 species in the 30 wt% WO3-containing MCF
sample. In the AMA-treated samples, however, the W6+/W5+ mo-
lar ratios dropped to about half of the initial values, demonstrating
that the W6+ species were located mainly on the surface of the
samples and could be removed by AMA treatment. (The XPS spec-
tra of the AMA-treated 10%WO3-MCF-1 also were detected, but the
signal was too weak for ready analysis.) This finding suggests that
high polymeric tungsten species were completely removed after
AMA treatment.
(b)
Fig. 6. XPS spectra of the W4f region for the as-synthesized different WO3 loading
catalysts (a) and the AMA-treated different WO3 loading catalysts (b).
3.1.6. Chemical states by XPS
3.2. Catalytic activity tests in the epoxidation of COD by aqueous H2O2
From the W4f XPS spectra of the WO3-containing MCF and the
AMA-treated samples given in Fig. 6, we can distinguish tungsten
oxide species in different chemical states from the position of the
W4f level by the curve-fitting procedure. Detailed quantitative re-
3.2.1. Effect of the different supports
Table
5 presents the results of COD epoxidation over W-
containing MCF, SBA-15, and MCM-41. Blank experiments per-

R. Gao et al. / Journal of Catalysis 256 (2008) 259–267
265
Table 6
Table 7
Influence of the WO3 loadings and AMA-treated cycles
Influence of the calcination temperature on the catalytic performance of 10%WO3-
MCF
−1
Catalyst
SMD (%)
TOFa (h
)
H2O2 utilityb (%)
◦
−1
)
Temperature ( C)
SMD (%)
TOFa (h
H2O2 utilityb (%)
10%WO3-MCF-1
10%WO3-MCF-2
20%WO3-MCF-1
20%WO3-MCF-2
30%WO3-MCF-1
30%WO3-MCF-2
10%WO3/MCF-1
10%WO3/MCF-2
10%WO3/MCF-3
92.7
91.7
86.4
87.0
59.4
85.0
68.5
94.1
93.2
17.8
16.6
10.4
15.4
4.0
14.0
3.0
3.6
43.6
58.8
53.0
53.4
37.8
28.9
20.2
51.0
50.0
10%WO3-MCF-syn.-1
10%WO3-MCF-300-1
10%WO3-MCF-500-1
10%WO3-MCF-700-1
10%WO3-MCF-900-1
89.0
90.0
92.7
95.0
96.0
0.2
11.3
17.8
3.9
34.3
31.0
43.6
49.5
50.1
3.9
◦
Reaction conditions: reaction temperature 60 C, COD: 0.53 mL, H2O2: 0.7 mL, reac-
tion time 24 h, t-BuOH: 10 mL, cat.: 0.2 g.
3.4
a
Calculated according to the equation: TOF = moles of COD converted per mole
◦
Reaction conditions: reaction temperature 60 C, COD: 0.53 mL, H2O2: 0.7 mL, reac-
of WO3 per hour.
b
tion time 24 h, t-BuOH: 10 mL, cat.: 0.2 g.
H2O2 utility (%) = (CCOD × SMD)/3CH
(CCOD: conversion of H2O2; SMD: selec-
2O2
a
Calculated according to the equation: TOF = moles of COD converted per mole
tivity of COD epoxide; CH : conversion of H2O2).
2O2
of WO3 per hour.
b
H2O2 utility (%) = (CCOD × SMD)/3CH
(CCOD: conversion of H2O2; SMD: selec-
2O2
10%WO3-MCF-2 can be seen to have the highest H2O2 utility of
all of the catalysts according to the same deductive process.
For the AMA-treated 10%WO3/MCF catalysts, the catalytic per-
formance of the 10%WO3/MCF-3 was similar to that of the
10%WO3/MCF-2 and differed from that of the 10%WO3/MCF-1, indi-
cating that the catalytic behavior of these catalysts was affected by
the intrinsic nature of tungsten oxide species. No changes in tung-
sten loading or tungsten species were seen; the most dominant
tungsten species were the isolated tetrahedral [WO4]2− species
and some low-condensed polymeric tungsten oxide species. Con-
sequently, the 10%WO3/MCF-3 and 10%WO3/MCF-2 demonstrated
similar catalytic performance. Meanwhile, for 10%WO3/MCF-1,
the dominant tungsten species were the isolated tetrahedral
[WO4]2− species and some low-condensed polymeric tungsten ox-
ide species; thus, the catalytic performance of the 10%WO3/MCF-1
differed from that of 10%WO3/MCF-3 and 10%WO3/MCF-2. The in
situ-prepared 10%WO3-MCF-2 also had a higher TOF value than
the impregnation-prepared 10%WO3/MCF-3, indicating that the
isolated tetrahedral [WO4]2− species were the most highly ac-
tive sites, although the low-condensed polymeric tungsten oxide
species also showed some activity.
tivity of COD epoxide; CH : conversion of H2O2).
2O2
formed with no catalysts or with purely siliceous supports demon-
strated no conversion of COD (results not given). Comparing
the findings for 10%WO3-MCF-1 and 10%WO3-SBA-15-1 reveals
a greater selectivity to COD epoxide and TOF values in the for-
mer than in the latter, indicating that MCF’s ultra-large pores
and unique 3-D cell-window structure lead to better catalytic
activity and selectivity. Meanwhile, comparing the findings for
10%WO3/MCF-1, 10%WO3/SBA-15-1 and 10%WO3/MCM-41-1 shows
that 10%WO3/MCF-1 was the most effective catalyst. These find-
ings also demonstrate that MCF’s ultra-large pores and the unique
3-D cell-window structure make this more favorable for the title
reaction compared with the SBA-15 and MCM-41, which only have
2-D pore structures.
We also investigated the effect of the two preparation methods.
The in situ-derived 10%WO3-MCF-1 demonstrated better catalytic
performance than the impregnation-derived 10%WO3/MCF-1. This
difference may be attributed to the different states of the tungsten
species over these two catalysts. Similar results also were obtained
for the 10%WO3-SBA-15-1 and 10%WO3/SBA-15-1 catalysts. Table 5
also presents H2O2 utility findings. As shown, the H2O2 utility was
greatest in the 10%WO3-MCF-1, in line with this catalyst’s highest
activity.
3.2.3. Effect of the calcination temperature
Table 7 characterizes the effect of calcination temperature on
the catalytic behavior of the 10%WO3-MCF catalyst. The catalyst
exhibited its best performance to COD epoxide at a calcination
3.2.2. Effect of WO3 loading and AMA-treated cycles
◦
temperature of 500 C. Calcination temperatures above or below
Table
6 presents the reaction results as a function of the
◦
500 C all led to decreased activity. The TOF values unambigu-
WO3 loading and AMA-treated cycles. The TOF value clearly de-
creased with increasing tungsten loading, as did the selectivity to
COD epoxide. The catalytic performance of the three WO3 load-
ing catalysts differed after one AMA treatment but became sim-
ilar after two AMA treatments, indicating that the catalytic be-
havior of the tungsten-containing silica catalysts was affected by
the intrinsic nature of tungsten oxide species. The 10%WO3-MCF-1
and 10%WO3-MCF-2 demonstrated no change in tungsten loading
and the tungsten species. In both catalysts, the dominant tung-
sten species were the isolated tetrahedral [WO4]2− species; thus,
the two catalysts exhibited similar catalytic performance. The TOF
value was higher in the 20%WO3-MCF-2 than in the 20%WO3-
MCF-1. In the 20%WO3-MCF-1, the dominant tungsten species were
the isolated tetrahedral [WO4]2− species and low-condensed poly-
meric tungsten oxide species; however, in the 20%WO3-MCF-2,
the dominant tungsten species were isolated tetrahedral [WO4]2−
species. This indicates that the highly active and selective tung-
sten species for the title reaction were the isolated tetrahedral
[WO4]2− species. The catalytic performance of the 30%WO3-MCF-1
and 30%WO3-MCF-2 also demonstrates that the isolated tetrahe-
dral [WO4]2− species are the most active and selective tungsten
species for the title reaction. Analogously, previous studies have
shown that the isolated tetrahedral vanadium oxide species are
active for the partial oxidation of methane [16]. Moreover, the
ously verify this conclusion; however, it should be noted that the
TOF value was sharply decreased at a calcination temperature of
700 C. These results can be explained by the gradual decomposi-
◦
tion of the intrinsic structure of the active species on the catalyst
◦
◦
at calcination temperatures of 700 C and above. Thus, 500 C was
chosen as the optimum calcination temperature for the 10%WO3-
MCF catalyst, because the tungsten oxide species could not be fully
activated at lower calcination temperature. However, the 10%WO3-
◦
MCF calcined at 700 and 900 C exhibited greater H2O2 utility and
selectivity than the other catalyst even though they had much
lower TOF values, indicating that most of the active tungsten
species were destroyed by treatment at elevated temperatures and
the ones that remained were the most stable isolated tetrahedral
{WO4} species.
3.2.4. Time course of COD oxidation over 10%WO3-MCF-1 catalyst
Scheme 1 shows the main product, COD epoxide, as well as the
main byproducts detected by GC-MS analyses. The finding of no
products from the O-heterocyclization of COD [27] is surprising,
indicating that the isolated tetrahedral {WO4} species played an
important role in the epoxidation of cycloocta-1,5-diene. Fig. 7 il-
lustrates the dependence of COD conversion and the selectivity to
COD epoxide on the reaction time over the 10%WO3-MCF-1 cat-

266
R. Gao et al. / Journal of Catalysis 256 (2008) 259–267
4. Conclusion
The AMA-treated WO3-MCF catalysts exhibited good perfor-
mance, which has been attributed to the ultra-large mesopores
of the catalysts for this reaction. The AMA-treated catalysts re-
tained the special structure of the supports under our treatment
conditions. The most highly isolated tungsten atoms were well em-
bedded into the supports when the catalysts were treated with
AMA, as demonstrated by UV–vis DRS, XPS, and UV-Raman exper-
iments. The UV–vis DRS results also showed that the crystalline
tungsten trioxide species were removed first, followed by the poly-
meric WO3 species. The tungsten percent maintained a fixed value
for the catalysts with different WO3 loadings, as confirmed by
ICP analysis. Different preparation methods led to different final
residual tungsten percents, but only the in situ method led to
the perfect single-site {WO4} tetrahedral species. The recycling ex-
periment demonstrated the excellent stability of the AMA-treated
WO3-MCF catalyst.
Fig. 7. The dependence of the coversion of COD and the selectivity of COD epoxide
on the reaction time over the 10%WO3-MCF-1 catalyst. Reaction condition: reaction
◦
temperature: 60 C, COD: 0.53 mL, H2O2: 0.7 mL, t-BuOH: 10 mL, cat.: 0.2 g.
Acknowledgments
Table 8
This work was supported by the Major State Basic Resource
Development Program (Grant 2003CB615807), the NSFC (Projects
20407006 and 20573024), and the Natural Science Foundation of
Shanghai Science and Technology Committee (Grant 06JC14004).
Reusability of 10%WO3-MCF-1
−1
Entry
SMD (%)
TOFa (h
)
H2O2 utilityb (%)
1
2
3
4
5
92.7
92.0
87.2
88.1
87.0
17.8
16.4
16.9
16.4
17.4
43.6
54.9
46.2
52.8
43.6
References
◦
[1] R.D. Wilson, D.G. Barton, C.D. Baertsch, E. Iglesia, J. Catal. 194 (2000) 175.
[2] B.F. Sels, D.E. De Vos, P.A. Jacobs, Angew. Chem. Int. Ed. 44 (2005) 310.
[3] C.D. Baertsch, K.T. Komala, Y.H. Chua, E. Iglesia, J. Catal. 205 (2002) 44.
[4] T. Kabe, W.H. Qian, A. Funato, Y. Okoshi, A. Ishihara, Phys. Chem. Chem. Phys. 5
(1999) 921.
[5] Y. Rezgui, M. Guemini, Appl. Catal. A 282 (2005) 45.
[6] U. Ciesla, D. Demuth, R. Leon, P. Petroff, G. Stucky, K. Unger, F. Schüth, J. Chem.
Soc. Chem. Commun. (1994) 1387.
Reaction conditions: reaction temperature 60 C, COD: 0.53 mL, H2O2: 0.7 mL, reac-
tion time 24 h, t-BuOH: 10 mL, cat.: 0.2 g.
a
Calculated according to the equation: TOF = moles of COD converted per mole
of WO3 per hour.
b
H2O2 utility (%) = (CCOD × SMD)/3CH
(CCOD: conversion of H2O2; SMD: selec-
tivity of COD epoxide; CH : conversion of H2O2).
2O2
2O2
[7] M.J. Yuan, Z. Shan, B.Z. Tian, B. Tu, P.Y. Yang, D.Y. Zhao, Microporous Meso-
porous Mater. 78 (2005) 37.
[8] X.L. Yang, W.L. Dai, C.W. Guo, H. Chen, Y. Cao, H.X. Li, H.Y. He, K.N. Fan,
J. Catal. 234 (2005) 438.
[9] C. Martin, P. Malet, G. Solana, V. Rives, J. Phys. Chem. B 102 (1998) 2759.
[10] F. Somma, G. Strukul, J. Catal. 227 (2004) 344.
[11] Z.R. Zhang, J.S. Sue, X.M. Zhang, S.B. Li, Chem. Commun. (1998) 241.
[12] E. Briot, J.Y. Piquemal, M. Vennat, J.M. Brégeault, G. Chottard, J.M. Manoli,
J. Mater. Chem. 10 (2000) 953.
alyst. The conversion of COD increased with reaction time over
the first ∼24 h, then remained almost the same with further in-
creases in reaction time. The selectivity of COD epoxide was 100%
initially and decreased slowly thereafter. Based on these findings,
we propose that COD epoxide is an intermediate from which the
byproducts are produced through its further oxidation and further
reaction with water.
[13] W.L. Dai, H. Chen, Y. Cao, H.X. Li, S.H. Xie, K.N. Fan, Chem. Commun. (2003)
892.
[14] X.L. Yang, W.L. Dai, H. Chen, J.H. Xu, Y. Cao, H.X. Li, K.N. Fan, Appl. Catal. A 283
(2005) 1.
[15] J.E. Herrera, J.H. Kwak, J.Z. Hu, Y. Wang, C.H.F. Peden, J. Macht, E. Iglesia,
J. Catal. 239 (2006) 200.
[16] H. Berndt, A. Martin, A. Brückner, E. Schreier, D. Müller, H. Kosslick, G.-U. Wolf,
B. Lücke, J. Catal. 191 (2000) 384.
[17] P. Rybarczyk, H. Berndt, J. Radnik, M.-M. Pohl, O. Buyevskaya, M. Baerns, A.
Brückner, J. Catal. 202 (2001) 45.
[18] M. Santhosh Kumar, M. Schwidder, W. Grünert, A. Brückner, J. Catal. 227 (2004)
384.
[19] J.C. Groen, A. Brückner, E. Berrier, L. Maldonado, J.A. Moulijn, J. Pérez-Ramírez,
J. Catal. 243 (2006) 212.
[20] M. Santhosh Kumar, M. Schwidder, W. Grünert, U. Bentrup, A. Brückner,
J. Catal. 239 (2006) 173.
[21] M.L. Wang, T.H. Huang, React. Kinet. Catal. Lett. 78 (2003) 275.
[22] E. Larsen, K.A. Jøgensen, Acta Chem. Scand. 43 (1989) 250.
[23] G. Legemaat, W. Drenth, J. Mol. Catal. 62 (1990) 119.
[24] R.U. Pillai, S.E. Demessie, Green Chem. 4 (2002) 495.
[25] E. Brule, R.d.Y. Miguel, K.K. Hii, Tetrahedron 60 (2004) 5913.
[26] R.U. Pillai, S.E. Demessie, S.R. Varma, Tetrahedron Lett. 43 (2002) 2909.
[27] R.H. Gao, W.L. Dai, X.L. Yang, H.X. Li, K.N. Fan, Appl. Catal. A 332 (2007) 138.
[28] C.Y. Cheng, K.J. Lin, M.R. Prasad, S.J. Fu, S.Y. Chang, S.G. Shyu, H.S. Sheu, C.H.
Chen, C.H. Chuang, M.T. Lin, Catal. Commun. 8 (2007) 1060.
[29] P. Schmidt-Winkel, W.W. Lukens Jr., P. Yang, D.I. Margolese, J.S. Lettow, J.Y. Ying,
G.D. Stucky, Chem. Mater. 12 (2000) 686.
3.2.5. Reusability of the 10%WO3-MCF-1 catalyst
To investigate the stability of active tungsten species in the
10%WO3-MCF-1 catalyst, the leaching of WOx species into the reac-
tion mixture and the tungsten remaining in the catalyst were also
determined by ICP analysis after five reaction cycles. No detectable
leaching of W-species or obvious loss of tungsten in the 10%WO3-
MCF-1 catalyst was observed, demonstrating the presence of strong
interactions between active tungsten species and the silica-based
matrix with the MCF structure. When the reaction over WO3-MCF
catalyst was carried out for 10 h, the catalyst was removed through
simple filtration, and the reaction solution was stirred for another
14 h. No detectable increases in the conversion and yield of COD
epoxide were found, indicating that the reaction was not influ-
enced by the homogeneous catalysis. Table 8 also gives results of
the selective oxidation of COD to epoxide over the 10%WO3-MCF-1
catalyst with different reaction cycles and the postregeneration
material. After the fifth cycle, the selectivity of the object prod-
uct remained the same, and the TOF values were as high as those
for the fresh catalyst. In conclusion, the 10%WO3-MCF-1 catalyst
demonstrated high stability as well as excellent activity for the
selective oxidation of COD to COD epoxide. The H2O2 utility was
close to 50% for both the fresh and recycled 10%WO3-MCF-1 cata-
lyst.
[30] D.Y. Zhao, J.L. Feng, Q.S. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D.
Stucky, Science 279 (1998) 548.

R. Gao et al. / Journal of Catalysis 256 (2008) 259–267
267
[31] D.Y. Zhao, Q.S. Huo, J.L. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem. Soc. 120
(1998) 6024.
[36] O. Klepel, W. Böhlmann, E.B. Ivanov, V. Riede, H. Papp, Microporous Meso-
porous Mater. 76 (2004) 105.
[37] R.S. Weber, J. Catal. 151 (1995) 470.
[38] J.G. Wu, S.B. Li, J. Phys. Chem. 99 (1995) 4566.
[39] K.J. Chao, C.N. Wu, H. Chang, L.J. Lee, S.F. Hu, J. Phys. Chem. B 101 (1997) 6341.
[40] F. Le Normand, J. El Fallah, L. Hilaire, P. Lègarè, A. Kotani, J.C. Parlebas, Solid
State Commun. 71 (1989) 885.
[32] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Schmitt, C.T.-
W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B. Higgins, J.L. Schlenker,
J. Am. Chem. Soc. 114 (1992) 10834.
[33] P. Schmidt-Winkel, W.W. Lukens Jr., D. Zhao, P. Yang, B.F. Chmelka, G.D. Stucky,
J. Am. Chem. Soc. 121 (1999) 254.
[41] M. Valigi, D. Gazzoli, I. Pettiti, G. Mattei, S. Colonna, S. De Rossi, G. Ferraris,
Appl. Catal. A 231 (2002) 159.
[34] Y. Wang, Q.H. Zhang, Y. Ohishi, T. Shishido, K. Takehira, Catal. Lett. 72 (2001)
215.
[42] X.L. Yang, W.L. Dai, R.H. Gao, K.N. Fan, J. Catal. 249 (2007) 278.
[35] C.W.F.T. Pistorius, J. Chem. Phys. 44 (1966) 4532.