Unique solvent effect of microporous crystalline titanosilicates in the oxidation of 1-hexene and cyclohexene

Article abstract of DOI:10.1016/j.jcat.2008.03.001

We investigated the influence of solvent on the activity, epoxide selectivity, and H₂O₂ efficiency of three types of zeolites-TS-1, Ti-Beta, and Ti-MWW-in the liquid-phase epoxidation of 1-hexene and cyclohexene. We found that the effect of solvent not only was related to the hydrophilicity/hydrophobicity of titanosilicates, but also was highly dependent on substrates. Acetonitrile should be the solvent of choice for the oxidation of 1-hexene over Ti-Beta and Ti-MWW, whereas methanol is preferred by TS-1. In contrast, for cyclohexene oxidation, methanol is favorable for Ti-Beta, whereas acetonitrile is the best for TS-1 and Ti-MWW. The effect of hydrophilicity/hydrophobicity is demonstrated by a series of catalytic results, but this cannot explain the solvent effect on the oxidation of cyclohexene over Ti-Beta. The H₂O₂ efficiency depends strongly on the aprotic/protic nature and polarity of solvents, as well as on the properties of titanosilicates.

Full text of DOI:10.1016/j.jcat.2008.03.001

Journal of Catalysis 256 (2008) 62–73
www.elsevier.com/locate/jcat
Unique solvent effect of microporous crystalline titanosilicates in the
oxidation of 1-hexene and cyclohexene
a,b
c
a,∗
Weibin Fan , Peng Wu , Takashi Tatsumi
a
Catalytic Chemistry Division, Chemical Resources Laboratory, Tokyo Institute of Technology, Nagatsuta 4259, Midori-ku, Yokohama 226-8503, Japan
b
State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, P.O. Box 165, 27 South Taoyuan Road,
Taiyuan 030001, China
c
Shanghai Key Laboratory of Green Chemistry and Chemical Processes, Department of Chemistry, East China Normal University, North Zhongshan Road 3663,
Shanghai 200062, China
Received 8 November 2007; revised 25 January 2008; accepted 4 March 2008
Available online 14 April 2008
Abstract
We investigated the influence of solvent on the activity, epoxide selectivity, and H O efficiency of three types of zeolites—TS-1, Ti-Beta,
2
2
and Ti-MWW—in the liquid-phase epoxidation of 1-hexene and cyclohexene. We found that the effect of solvent not only was related to the
hydrophilicity/hydrophobicity of titanosilicates, but also was highly dependent on substrates. Acetonitrile should be the solvent of choice for the
oxidation of 1-hexene over Ti-Beta and Ti-MWW, whereas methanol is preferred by TS-1. In contrast, for cyclohexene oxidation, methanol is
favorable for Ti-Beta, whereas acetonitrile is the best for TS-1 and Ti-MWW. The effect of hydrophilicity/hydrophobicity is demonstrated by a
series of catalytic results, but this cannot explain the solvent effect on the oxidation of cyclohexene over Ti-Beta. The H O efficiency depends
2
2
strongly on the aprotic/protic nature and polarity of solvents, as well as on the properties of titanosilicates.
2008 Elsevier Inc. All rights reserved.
©
Keywords: Cyclohexene; 1-Hexene; Oxidation; Solvent effect; Ti-Beta; Ti-MWW; TS-1; Titanosilicate
1
. Introduction
It has been shown that the structure of titanosilicates, synthe-
sis methods, reaction temperature, diffusion of substrate mole-
cules and the solvents used for oxidation strongly influence the
catalytic performance [4–8]. In general, the role of solvents is
to homogenize the liquid phase, thus avoiding mass-transfer
problems to accelerate the interaction of reactants with cata-
lysts. Methanol as a solvent favors the oxidation of alkanes
and alkenes over TS-1, due in part to its hydrophobic charac-
ter [5,9], whereas the catalytic activity of relatively hydrophilic
Ti-Beta is improved in acetonitrile solvent in the oxidation of
Since the discovery of TS-1, titanosilicate molecular sieves
have been attracting much attention, because these materials are
capable of serving as highly efficient catalysts for the oxida-
tion of various organic substrates (i.e., alkanes, alkenes, alco-
hols, aromatics) with H2O2 as an oxidant under mild conditions
[
1–3]. However, the nature of the active intermediate titanium
species formed during the oxidation process is not yet fully un-
derstood. In addition, the factors that determine the activity and
selectivity of various titanosilicates in specific reactions are not
yet clear, although considerable progress toward this has been
made. A full understanding of these factors will be favorable
for the development of new potential catalysts and design of
optimum oxidation conditions.
1
-hexene and cyclohexanol [5]. This is widely accepted as re-
sulting from different active species. Species I, with a stable
five-membered ring structure formed by the coordination of
ROH to Ti centers and hydrogen bonding to Ti-peroxo com-
plex, is believed to be the active intermediates in protic al-
cohol solvents (Scheme 1), whereas species II is assumed to
contribute to the oxidation of substrates in aprotic solvents. Re-
2
*
cently, Lamberti and co-workers found that an end-on η Ti
Corresponding author. Fax: +81 45 9245282.
E-mail address: ttatsumi@cat.res.titech.ac.jp (T. Tatsumi).
hydroperoxo complex was generated (likely by the reversible
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.03.001

W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
63
method using highly deboronated B-MWW as a silica source
◦
[
18]. The as-synthesized samples were calcined at 550–600 C
for 8 h. For Ti-MWW, the as-synthesized material was first
◦
washed with 2 M HNO3 aqueous solution at 100 C for 24 h
before calcination, whereas TS-1 was treated with 1 M HCl
aqueous solution (liquid/solid = 50 mL/g) for 20 h at room
temperature after calcination. Some samples (1 g) were sily-
lated by being suspended in the boiling solution of 1,1,1,3,3,3-
hexamethyldisilazane (0.2 g) and toluene (20 mL) for about
Scheme 1. Structural scheme of proposed intermediate Ti species.
4
h.
The crystalline phase and purity of the samples were identi-
rupture of Ti–O–Si bridge under anhydrous H2O2 conditions),
and that this complex was reversibly transformed to a side-on
η Ti peroxo complex after addition of water, indicating that
fied with a MAC Science M3X 1030 X-ray diffractometer using
CuKα radiation. Crystal morphology and size were evaluated
on a Hitachi S-5200 field-emission scanning electron micro-
scope (FE-SEM). IR spectra in the OH stretching vibration
region were measured on a Perkin-Elmer 1600 FTIR spec-
trometer. Before the spectra were recorded, the samples were
2
water molecules play an active role in determining the relative
concentration of Ti peroxo to hydroperoxo species present on
the working catalyst [10–12]. Despite such findings and inter-
pretations, however, aprotic acetone has been reported to be the
solvent of choice in terms of both activity and selectivity for
the epoxidation of styrene and allyl alcohol on TS-1 [13,14].
In contrast, in the hydroxylation/oxidation of toluene, anisole,
benzyl alcohol and cyclohexanol over TS-1, the catalytic ac-
tivity and selectivity were significantly enhanced in water com-
pared with in acetonitrile solvent [15], even though the reactants
were in two immiscible liquids (triphasic, not biphasic, condi-
tions).
◦
evacuated at 500 C for 2 h under high-vacuum conditions. The
diffuse reflectance (DR) UV–vis spectra were recorded against
a white halon reference standard on a Jasco V-550 UV–vis spec-
trophotometer equipped with an integration sphere. The chemi-
cal compositions of the calcined samples were determined by
a Shimadzu ICPS-8000E inductively coupled plasma atomic
emission spectrometer.
Actually, the role of the solvent is not so simple. The apro-
tic/protic nature and polarity of solvents, the solubility of re-
actants and products in solvents, and diffusion and counterdif-
fusion effects also may play important roles [5,6]. In addition,
the interaction of the solvent with the oxidant and/or the in-
termediate species possibly affects the reaction pathway [16].
Up to now, although several types of interesting titanosili-
cates (e.g., TS-1, TS-2, Ti-MTW, Ti-Beta and Ti-MWW) have
been synthesized and found to be potential catalysts in fine
chemistry, the effect of solvent on the catalytic performance
remains unclear despite the reasonable explanations given to
date [4,5].
2
.2. Catalytic measurements
The oxidation of 1-hexene and cyclohexene was carried out
◦
at 60 C for 2 h under stirring condition in a 20-mL round-
bottomed flask immersed in a thermostatted water bath and
equipped with a condenser. The stirring was sufficiently strong
to rule out the effect of external diffusion on catalytic per-
formance for all of the batches. In a typical batch, 0.05 g of
catalyst, 10 mL of solvent, 10 mmol of alkene, and 10 mmol of
H2O2 (31% aqueous solution) were used. The product was ana-
lyzed using a Shimadzu GC-14B gas chromatograph equipped
with a 50-m OV-1 capillary column and a flame ionization
detector. The amount of the unconverted H2O2 was deter-
mined with 0.1 M Ce(SO4)2 aqueous solution using the titration
method.
In this paper, we report a comprehensive investigation into
the solvent effect on the catalytic performance of three types of
representative titanosilicates—TS-1, Ti-MWW, and Ti-Beta—
for the oxidation of 1-hexene and cyclohexene with H O ,
2
2
demonstrating that the catalytic activity can be enhanced when
appropriate amounts of mixed solvents are present in the oxi-
dation system, compared with single solvent. In particular, con-
verse results are obtained when the same titanosilicate is used
for catalyzing the oxidation of 1-hexene and cyclohexene. Thus,
new insight into the effect of solvent is needed.
3
. Results
3
.1. Physicochemical characteristics of TS-1, Ti-Beta, and
Ti-MWW
XRD and FE-SEM measurements confirmed the crystalline
phase and purity of TS-1, Ti-Beta, and Ti-MWW [7,8,19–22].
The FE-SEM images showed that the crystals of TS-1 and Ti-
Beta had an irregular spherical shape with diameters of about
0.15–0.2 and 0.2–0.3 µm, respectively, whereas those of Ti-
MWW exhibited a habit of thin hexagonal platelets of about
0.2–0.5 µm long and 0.05–0.1 µm thick (not shown). The find-
ing of similar crystal sizes in these three types of titanosilicate
basically eliminates any effect of particle size on catalytic per-
formance. The diffuse reflectance UV–vis spectroscopy find-
ings demonstrated only a single band at about 205 for TS-1
2
. Experimental
2
.1. Synthesis and characterization of different titanosilicates
TS-1 was synthesized through the method reported by
Thangaraj and Sivasanker [17], with modifications. Ti-Beta was
synthesized with the seeding method by using fumed silica
as a silica source following the procedures reported by Cam-
blor et al. [7]. Ti-MWW was prepared by the postsynthesis

64
W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
(A)
(A)
(B)
Fig. 2. Relationship of 1-hexene conversion obtained in acetonitrile and
methanol solvents over TS-1 (A) and Ti-MWW (B) and reaction time. (Re-
◦
action conditions: 60 C, 0.05 g catalyst, 10 mL solvent, 10 mmol substrate
and 10 mmol H O (30% aqueous solution).)
(B)
2 2
Fig. 1. (A) DR UV–vis spectra and (B) IR spectra in the OH region of (a)
Ti-MWW, (b) TS-1, and (c) Ti-Beta catalysts.
3
.2. Oxidation of 1-hexene and cyclohexene over various
titanosilicates in acetonitrile and methanol
and at 210 nm for Ti-MWW, and one band at 215 nm with
a small shoulder around 230 nm for Ti-Beta (Fig. 1A). This
demonstrates that all of the Ti atoms were isolated and had
a tetrahedral coordination in both TS-1 and Ti-MWW [8,21–
Fig. 2A shows the dependence of 1-hexene conversion ob-
tained over TS-1 on the reaction time. In agreement with the
results reported in the literature [1,3,5], the use of methanol as
a solvent resulted in substantially higher catalytic activity com-
pared with the use of acetonitrile; the initial reaction rates were
2
3], whereas the Ti atoms in Ti-Beta were in a fourfold to
sixfold coordination [5]. The absence of bands above 250 nm
indicates that extra-framework octahedral Ti species and TiO2
were negligible in all three materials [5,8,21–23]. The IR spec-
tra in the OH region showed two much more intense bands
0
.093 mmol/min in methanol solvent and 0.036 mmol/min in
acetonitrile. As expected, the difference in activity decreased
with decreasing Ti content, because the number of active sites
decreased (Fig. 3). However, with respect to the oxidation of
cyclohexene, a converse solvent effect was observed (Fig. 4);
when acetonitrile was used as the solvent, the conversion was
nearly fourfold greater than that obtained with methanol as
solvent. Almost no dependence of conversion on Ti content
was observed at a Ti/(Si + Ti) molar ratio >0.0086, likely
because the reaction occurred mainly on the external surface
−
1
at about 3740 and 3550 cm (Fig. 1B), attributed to termi-
nal Si–OH and hydrogen-bonded silanol groups at defect sites,
respectively [22,24,25], in Ti-Beta compared with TS-1 and Ti-
MWW, implying that Ti-Beta contained much more defect sites
and thus was more hydrophilic. The hydrophilicity of these
titanosilicates decreased in the order Ti-Beta > Ti-MWW >
TS-1.

W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
65
Fig. 5. Dependence of conversion on the Ti content of Ti-MWW in the oxida-
tion of 1-hexene in methanol and acetonitrile solvents. (Reaction conditions:
60 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol substrate and 10 mmol
Fig. 3. Dependence of 1-hexene conversion on the Ti content of TS-1 in the oxi-
◦
dation of 1-hexene in methanol and acetonitrile solvents. (Reaction conditions:
◦
6
0 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol substrate and 10 mmol
H O (31% aqueous solution).)
2 2
H O (31% aqueous solution).)
2
2
Fig. 6. Dependence of conversion on the Ti content of Ti-MWW in the oxida-
tion of cyclohexene in methanol and acetonitrile solvents. (Reaction conditions:
60 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol substrate and 10 mmol
Fig. 4. Dependence of conversion on the Ti content of TS-1 in the oxidation of
cyclohexene in methanol and acetonitrile solvents. (Reaction conditions: 60 C,
◦
◦
4
h, 0.05 g catalyst, 10 mL solvent, 10 mmol substrate and 10 mmol H O (31%
H O (31% aqueous solution).)
2 2
2
2
aqueous solution).)
and near the pore mouth, because cyclohexene has nearly
the same molecular size as cyclohexane, which hardly dif-
fuses into the channel of MFI-type materials. The increased
Ti content incorporated into the internal surface did not con-
tribute significantly to the enhanced activity in the oxidation
of cyclohexene. In contrast to TS-1, Ti-MWW exhibited much
higher activity for the oxidation of 1-hexene in acetonitrile sol-
vent than in methanol (initial reaction rate, 0.48 mmol/min vs
(Ti-YNU-1) had a 12 membered-ring (MR) interlayer pore-
opening connected to supercages, whereas high Ti content led
to the formation of a typical Ti-MWW structure with distorted
10-MR pore openings in the interlayer space [21,22]. Thus,
more Ti species would be accessible to bulky substrate mole-
cules in the low-content sample than in the high-content sam-
ple. The latter is highly active only for linear alkenes, because
of the difficulty in accommodating cyclic alkenes. When 1-
hexene was oxidized over Ti-Beta, the conversion in acetonitrile
solvent was about twice that attained in methanol; however,
for cyclohexene oxidation, methanol was clearly superior to
acetonitrile as a solvent (Table 1), in conflict with the results
reported by Corma et al. [5] but consistent with our previous
findings [26].
0
.14 mmol/min) (Figs. 2B and 5), whereas it gave slightly
higher conversion in the oxidation of cyclohexene in acetoni-
trile solvent than in methanol (Fig. 6). The conversion of 1-
hexene increased almost linearly with the amount of Ti in the
framework (Fig. 5), whereas a converse trend was observed in
the oxidation of cyclohexene (Fig. 6). The low-content sample

66
W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
Table 1
Catalytic results for the oxidation of 1-hexene and cyclohexene over various titanosilicates
a
b
Entry
Titanosilicate
Si/Ti
Solvent
Substrate
Conv. (%)
TON
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
TS-1
TS-1
TS-1 silylated
TS-1 silylated
TS-1 poisoned
TS-1
60
60
60
60
60
60
60
60
60
60
60
35
35
35
35
35
35
35
35
44
44
44
44
44
44
44
44
44
44
44
MeCN
MeOH
MeCN
MeOH
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
1-Hexene
1-Hexene
1-Hexene
1-Hexene
11.3
18.4
11.0
19.2
16.9
4.9
0.9
4.7
1.4
1.6
0.7
17.0
7.3
8.7
4.6
16.5
31.0
7.7
83.2
135.4
81.0
141.3
124.4
36.1
6.6
34.6
10.3
11.8
5.2
74.1
31.8
37.9
20.1
71.9
135.2
33.6
15.7
371.5
119.7
323.1
108.3
332.8
81.6
81.1
33.2
57.7
18.5
22.8
c
1-Hexene
Cyclohexene
Cyclohexene
Cyclohexene
Cyclohexene
Cyclohexene
Cyclohexene
1-Hexene
1-Hexene
1-Hexene
1-Hexene
Cyclohexene
Cyclohexene
Cyclohexene
Cyclohexene
1-Hexene
1-Hexene
1-Hexene
1-Hexene
1-Hexene
TS-1
TS-1 silylated
TS-1 silylated
TS-1 poisoned
TS-1 poisoned
Ti-Beta
c
c
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
Ti-Beta
Ti-Beta silylated
Ti-Beta silylated
Ti-Beta
Ti-Beta
Ti-Beta silylated
Ti-Beta silylated
Ti-MWW
3.6
68.3
22.0
59.4
19.9
60.6
15.0
14.9
6.1
Ti-MWW
Ti-MWW silylated
Ti-MWW silylated
Ti-MWW poisoned
Ti-MWW
c
Cyclohexene
Cyclohexene
Cyclohexene
Cyclohexene
Cyclohexene
Cyclohexene
Ti-MWW
Ti-MWW silylated
Ti-MWW silylated
Ti-MWW poisoned
Ti-MWW poisoned
10.6
3.4
4.2
c
c
a
◦
Reaction conditions: 60 C, 2 h (4 h for TS-1 and its modified materials for the oxidation of cyclohexene), 0.05 g catalyst, 10 mL solvent, 10 mmol alkene and
1
0 mmol H O (31% aqueous solution).
2
2
b
Turnover number: the mole of substrate converted per mole of Ti present in the catalysts.
c
The poisoning of samples was conducted by adding 1 mmol of 2,4-dimethylquinoline into the reaction mixture.
3
.3. Catalytic performance of various titanosilicates after
In contrast, the H2O2 efficiency on Ti-MWW (>90%) and sily-
silylation and poisoning
lated Ti-MWW (ca. 80%) in acetonitrile solvent was similar
to that in methanol irrespective of the alkenes oxidized. On
the other hand, the poisoning by 2,4-dimethylquinoline led to
a drastic decrease in H2O2 efficiency in the oxidation of cy-
clohexene. For Ti-Beta, the H2O2 efficiency depended on the
substrates and modifications. Relatively high H2O2 efficiency
was obtained with acetonitrile as the solvent in the oxidation of
Table 1 summarizes the catalytic results of silylated and
poisoned TS-1, Ti-Beta, and Ti-MWW for the oxidation of
1
-hexene and cyclohexene. Clearly, silylation of TS-1 had no
significant effect on the oxidation of 1-hexene and cyclohex-
ene regardless of the solvents used (entries 1–4, 6–9); however,
the silylation of Ti-Beta drastically decreased the activity both
in methanol and in acetonitrile (entries 12–19). This phenom-
enon also was observed in the oxidation of cyclohexene on
silylated Ti-MWW, although no significant effect was produced
on the oxidation of 1-hexene (entries 20–23, 25–28). On the
other hand, when TS-1 and Ti-MWW were poisoned by bulky
1
-hexene, whereas methanol exhibited much greater efficiency
than acetonitrile in the oxidation of cyclohexene (80% vs 50%).
In contrast, silylated Ti-Beta exhibited greater H2O2 efficiency
in methanol solvent than in acetonitrile (44–55% vs <30%), re-
gardless of the substrates.
3
.4. Effects of the aprotic/protic nature and polarity of the
2
,4-dimethylquinoline, they were hardly active for the oxida-
solvents
tion of cyclohexene (entries 10, 11, 29 and 30), although they
still gave high conversion in the epoxidation of 1-hexene (en-
tries 5 and 24).
In addition, the modification of titanosilicates had a marked
solvent effect on H2O2 efficiency. In the case of TS-1, irrespec-
tive of substrates or modifications, the use of methanol as the
solvent always gave higher H2O2 efficiency in 1-hexene oxida-
tion compared with the use of acetonitrile (>85% vs ca. 70%).
Tables 2 and 3 give an overview of the influence of the
aprotic/protic nature and polarity of different solvents on the
catalytic performance of various titanosilicates for the oxida-
tion of 1-hexene and cyclohexene, respectively. In the oxidation
of 1-hexene, the activity of Ti-Beta increased with the increase
in polarity for both the aprotic and protic alcohol solvents, al-
though methanol solvent gave a lower conversion than ethanol

W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
67
Table 2
Influence of the aprotic/protic nature and polarity of solvents on the catalytic performance of different titanosilicates for the oxidation of 1-hexene
a
Solvent
Dielectric
constant
TS-1 (Si/Ti = 53)
-Hexene
Conv. TON
%)
Ti-Beta (Si/Ti = 30)
Ti-MWW (Si/Ti = 44)
1
H O
1-Hexene
b
H O
1-Hexene
b
H O
2 2
2
2
2
2
b
c
d
c
d
c
d
Selec.
(%)
Conv. Effi.
(%)
Conv. TON
(%)
Selec.
(%)
Conv. Effi.
(%)
Conv. TON
(%)
Selec.
(%)
Conv. Effi.
(%)
(
(%)
(%)
(%)
MeCN
MeCOMe 20.7
MeCOEt
MeOH
EtOH
37.5
14.5
17.8
19.1
21.2
17.3
9.2
94.5 100
116.1 97.9
124.5 99.2
138.2 96.6
112.8 81.1
60.0 100
22.8
22.9
25.6
25.3
19.6
12.9
63.8
79.7
77.6
91.0
90.4
72.8
17.7
7.1
5.0
8.7
11.7
4.7
66.5
26.7
18.8
32.7
44.0
17.7
100
40.6
27.2
32.5
19.6
22.6
18.7
43.6
26.1
15.6
38.1
53.1
25.8
68.3
42.3
49.4
21.9
16.0
14.0
371.5 100
230.1 99.2
268.7 99.7
119.1 91.0
87.0 92.1
76.2 99.7
69.1
47.6
52.9
30.2
21.6
21.7
97.3
89.6
95.1
74.2
75.3
65.4
72.3
78.9
36.3
34.8
67.0
18.5
32.7
24.5
10.9
t-BuOH
a
◦
Reaction conditions: 60 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol 1-hexene and 10 mmol H O (31% aqueous solution).
2
2
b
c
d
Turnover number: the mole of substrate converted per mole of Ti present in the catalysts.
Epoxide selectivity.
Efficiency.
Table 3
Influence of the aprotic/protic nature and polarity of solvents on the catalytic performance of Ti-Beta and Ti-MWW for the oxidation of cyclohexene
a
Solvent
Dielectric Ti-Beta (Si/Ti = 30)
Ti-MWW (Si/Ti = 44)
constant
b
b
Conv. (%) TON Product selectivity (%)
H O
2
Conv. (%) TON Product selectivity (%)
H O
2 2
2
c
d
e
c
d
e
Epo. Diol Ether Others Conv. Effi.
%) (%)
Epo. Diol Ether Others Conv. Effi.
(
(%)
(%)
MeCN
MeCOMe 20.7
MeCOEt 18.5
MeOH
EtOH
t-BuOH
37.5
17.7
7.1
11.6
40.4
13.0
4.8
66.5 80.0 18.8
26.7 12.3 78.1
–
–
–
1.2
9.6
1.9
1.0
0
36.2
32.9
33.8
51.8
22.9
16.4
52.2
23.5
36.3
80.1
89.1
30.3
15.0
3.8
5.3
14.9
9.5
3.6
81.6
20.7
28.8
81.1
51.7
19.6
70.0 20.4
39.3 60.0
13.7 84.4
–
–
–
9.6
0.7
1.9
1.7
2.5
0
18.4
20.1
21.0
18.6
15.8
7.4
80.8
18.7
25.9
79.6
88.3
53.4
43.6
151.9
48.9
3.8 94.3
0.6 16.1 82.3
4.3 43.2 52.6
32.7
24.5
10.9
6.4
6.7 85.2
11.7 42.0 43.8
15.6 76.8 7.6
18.0 12.7 81.5 5.0
0.8
a
◦
Reaction conditions: 60 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol cyclohexene and 10 mmol H O (31% aqueous solution).
2
2
b
c
d
e
Turnover number: the mole of substrate converted per mole of Ti present in the catalysts.
Epoxide.
Monomethyl glycol ether.
Efficiency.
solvent. This also held true for Ti-MWW, with the exception
of lower activity in acetone solvent than in methyl ethyl ketone
solvent. In contrast, the activity of TS-1 increased with decreas-
ing polarity of the aprotic solvents, whereas the opposite effect
was seen for the protic alcohol solvents. The aprotic solvents
favored the formation of epoxide regardless of the titanosili-
cate catalysts used, whereas alcohol solvents (except t-BuOH)
led to a significant reduction in epoxide selectivity, especially
over Ti-Beta, where hydrolysis/solvolysis of 1-hexene epoxide
occurred. In addition, compared with TS-1 and Ti-MWW, Ti-
Beta also showed very low selectivity for epoxide in acetone
ing solvent polarity for both the aprotic and protic alcohol se-
ries, with the exception that methyl ethyl ketone solvent gave
greater conversion than acetone solvent. As the solvent polar-
ity decreased, the epoxide selectivity decreased dramatically
for the aprotic solvents but increased for the protic alcohol sol-
vents.
3
.5. Effect of water on oxidation activity
Figs. 7 and 8 depict the catalytic results of the three titanosil-
icates for the oxidation of 1-hexene and cyclohexene when
water was added to the reaction system using acetonitrile as
the solvent. The volume ratio of water to acetonitrile in the
reaction mixture was kept below 13% mL/mL to avoid any
serious mass-transfer problems. The activity of Ti-MWW for
the oxidation of 1-hexene increased with increasing amounts
of water in the system. The addition of a small amount of wa-
ter also slightly promoted substrate conversion over Ti-Beta.
In contrast, no positive effect was observed in the oxidation
of cyclohexene over these two materials. As for TS-1, oxi-
dation of neither 1-hexene nor cyclohexene was significantly
affected by the addition of water under the present study condi-
tions.
and methyl ethyl ketone solvents. The efficiency of H O in the
2
2
oxidation of 1-hexene over TS-1 was higher in protic alcohol
solvents than in aprotic solvents, whereas the opposite effect
was seen for Ti-MWW. The H O efficiency was much lower
2
2
for Ti-Beta than for TS-1 and Ti-MWW. It is noteworthy that
H O efficiency generally improved with increasing 1-hexene
2
2
conversion for all three materials.
For cyclohexene oxidation, because the conversion over
TS-1 was very low, we focused on studying the effect of sol-
vents on the catalytic properties of Ti-Beta and Ti-MWW.
For both of these catalysts, activity decreased with decreas-

68
W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
Table 5
Influence of the MeCN/MeOH ratio in the solvent mixture on the catalytic prop-
a
erty of Ti-Beta (Si/Ti = 30) for the oxidation of 1-hexene
b
MeCN/MeOH
Conversion of
TON
Selectivity (%)
(mL/mL)
1-hexene (%)
Epoxide
Others
0
1
4
6
8
1
/10
8.7
15.7
17.7
20.8
20.5
17.7
32.7
59.0
66.5
78.2
77.1
66.5
36.3
97.0
98.8
99.3
100
63.7
3.0
1.2
0.7
0
.7/8.3
.0/6.0
.0/4.0
.5/1.5
0/0
100
0
a
◦
Reaction conditions: 60 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol
1
-hexene and 10 mmol H O (31% aqueous solution).
2 2
b
Turnover number: the mole of substrate converted per mole of Ti present in
the catalysts.
Fig. 7. Effect of the water amount, including the water contained in H O aque-
2
2
3
.6. Effect of mixed solvents on the catalytic performance of
ous solution, in the reaction mixture on the conversion of 1-hexene over various
◦
various titanosilicates
titanosilicates. (Reaction conditions: 60 C, 2 h, 0.05 g catalyst, 10 mL acetoni-
trile, 10 mmol substrate and 10 mmol H O (31% aqueous solution).)
2
2
Table 4 shows the catalytic results of TS-1 for the oxidation
of 1-hexene in a mixed solvent of methanol and acetonitrile
with the total volume of solvent kept constant (10 mL). With
increasing amounts of MeCN up to a MeCN/MeOH ratio of
1
.5/8.5 (mL/mL), not only was the activity enhanced, but also
the selectivity was increased to nearly 100%. But a further in-
crease in the MeCN/MeOH (mL/mL) ratio to >6/4 produced
in a significant decrease in conversion. Actually, the conversion
of 1-hexene also increased when acetone or methyl ketone in-
stead of acetonitrile was used as a mixed solvent, even though
the epoxide selectivity declined. In a manner similar to that for
TS-1, the activity of Ti-Beta for the oxidation of 1-hexene in
a mixed acetonitrile and methanol solvent (MeCN/MeOH =
6
/4–8.5/1.5 mL/mL) was greater than that obtained in a single
solvent while maintaining high epoxide selectivity (Table 5).
In contrast, adding a small amount of acetonitrile to methanol
(
e.g., MeCN/MeOH = 1.5/8.5 mL/mL) drastically lowered the
conversion of cyclohexene on Ti-Beta, close to that obtained
in pure acetonitrile solvent (Table 6). For the Ti-MWW cata-
lyst, the conversion of 1-hexene was significantly reduced as
increasing amounts of methanol were co-added into the ace-
tonitrile solvent (Table 7).
Fig. 8. Effect of the water amount, including the water contained in H O aque-
2
2
ous solution, in the reaction mixture on the conversion of cyclohexene over
◦
various titanosilicates. (Reaction conditions: 60 C, 2 h, 0.05 g catalyst, 10 mL
acetonitrile, 10 mmol substrate and 10 mmol H O (31% aqueous solution).)
2
2
Table 4
3.7. Decomposition of hydrogen peroxide
Influence of the MeCN/MeOH ratio in the solvent mixture on the catalytic prop-
erty of TS-1 (Si/Ti = 53) for the oxidation of 1-hexene
a
The results of the decomposition of H2O2 over various ti-
b
◦
MeCN/MeOH
Conversion of
1-hexene (%)
TON
Selectivity (%)
Epoxide
tanosilicates at 60 C in the absence of substrates are illustrated
(mL/mL)
Others
in Fig. 9. The figure shows that regardless of the solvents used,
the Ti-Beta catalyst produced much more rapid decomposition
of H O than either TS-1 or Ti-MWW. In general, compared
0
0
0
1
4
6
8
1
/10
21.2
26.9
26.5
27.4
26.8
24.2
17.1
14.5
138.2
175.4
172.8
178.6
174.7
157.8
111.5
94.5
96.4
98.9
98.9
99.6
100
100
100
100
3.6
1.1
1.1
0.4
0
0
0
0
.4/9.6
.8/9.2
.5/8.5
.0/6.0
.0/4.0
.5/1.5
0/0
2
2
with methanol, MeCN solvent strongly promoted the decom-
position of H2O2, although the difference was not apparent for
TS-1.
4. Discussion
a
◦
Reaction conditions: 60 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol
-hexene and 10 mmol H O (31% aqueous solution).
1
The effect of solvent on catalytic performance is a quite
complicated but important area in studies of heterogeneous cat-
alytic systems involving titanosilicate catalysts. The structure
2
2
b
Turnover number: the mole of substrate converted per mole of Ti present in
the catalysts.

W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
69
Table 6
a
Influence of the MeCN/MeOH ratio in the solvent mixture on the catalytic property of Ti-Beta (Si/Ti = 30) for the oxidation of cyclohexene
MeCN/MeOH
b
Conversion of
TON
Selectivity (%)
H O
2 2
(mL/mL)
1-hexene (%)
c
Epoxide
Others
Conv. (%)
Effi. (%)
0
1
4
6
8
1
0
/10
40.4
21.9
21.5
21.1
17.7
17.7
36.9
151.9
82.3
80.8
79.3
66.5
66.5
138.7
0.6
30.0
51.4
56.8
66.5
80.0
1.0
99.4
70.0
48.6
43.2
33.5
20.0
99.0
51.8
35.7
33.2
33.7
33.0
36.2
41.2
80.1
64.6
67.8
64.2
62.1
52.2
93.6
.5/8.5
.0/6.0
.0/4.0
.5/1.5
0/0
d
/10 (HQ)
a
◦
Reaction conditions: 60 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol cyclohexene and 10 mmol H O (31% aqueous solution).
2
2
b
c
d
Turnover number: the mole of substrate converted per mole of Ti present in the catalysts.
Efficiency.
1
mmol of hydroquinone was added into the reaction mixture.
Table 7
Influence of the MeCN/MeOH ratio in the solvent mixture on the catalytic prop-
a
erty of Ti-MWW (Si/Ti = 44) for the oxidation of 1-hexene
b
MeCN/MeOH
Conversion of
TON
Selectivity (%)
(mL/mL)
1-hexene (%)
Epoxide
Others
0
1
4
6
8
1
/10
22.0
35.4
43.9
53.2
60.5
68.3
119.7
192.6
238.8
289.4
329.1
371.5
87.6
98.8
99.4
100
100
100
12.4
1.2
0.6
0
.5/8.5
.0/6.0
.0/4.0
.5/1.5
0/0
0
0
a
◦
Reaction conditions: 60 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol
1
-hexene and 10 mmol H O (31% aqueous solution).
2 2
b
Turnover number: the mole of substrate converted per mole of Ti present in
the catalysts.
and surface properties of the catalysts and the aprotic/protic na-
ture, polarity, and basicity of solvents, along with the solubility
and diffusion effects of reactants, intermediates, and products in
solvents, all need to be considered when choosing a suitable sol-
vent. It has been proposed that peroxo-compounds are formed
as intermediates by the reaction of group V and VI metal oxides
or metal complexes with either organic hydroperoxides or hy-
drogen peroxide [9]. In contrast, in the liquid-phase hydrogen
peroxide oxidation of organic substrates catalyzed by titanosili-
cates, species I and II (Scheme 1) with a five-membered ring
of hydrogen bonded structure are presumed to be the active
species in alcohols and aprotic solvents, respectively [4,5].
Fig. 9. Decomposition of H2O2 over different titanosilicates in methanol and
◦
acetonitrile solvents at 60 C in the absence of substrate molecules. (Reaction
◦
conditions: 60 C, 2 h, 0.05 g catalyst, 10 mL solvent and 10 mmol H O (31%
2
2
aqueous solution).)
firm this hypothesis, a TS-1 sample was poisoned by adding
,4-dimethylquinoline into the reaction mixture. The conver-
sion of cyclohexene drastically declined in acetonitrile solvent
Table 1, entry 10), whereas the conversion of 1-hexene was not
2
(
significantly changed in methanol solvent (Table 1, entry 5).
This finding demonstrates that the oxidation of cyclohexene
was indeed catalyzed primarily by Ti sites on the exterior sur-
face and/or near the pore mouth.
4
.1. Effect of solvent on the catalytic performance of TS-1 in
This finding is further supported by Fig. 4, which shows
that the Ti content in TS-1 had no significant influence on
the catalytic activity for cyclohexene oxidation; there was lit-
tle difference in the numbers of Ti sites on the external surface
between the low-content and high-content samples when the
Ti/(Si + Ti) molar ratio in the TS-1 catalyst was >0.0086. This
is probably because Ti cations did not randomly substitute Si,
but instead preferably occupied T , T , T , and possibly T
the epoxidation of 1-hexene and cyclohexene
Compared with species II, species I is more likely to form
on TS-1. This is believed to be due to this species’ hydrophobic
character [4], which allows methanol to approach Ti sites more
readily than water, leading to the formation of numerous active
sites of species I [4]. Therefore, when 1-hexene is oxidized on
TS-1, methanol should be the solvent of choice. But because
cyclohexene has a large molecule size, it is oxidized mainly on
the exterior surface and/or near the pore mouth, where much
more Si–OH and Ti–OH groups are present than inside the
channels, making these areas relatively hydrophilic. To con-
6
7
11
10
sites [25], the predominating defect sites in silicalite-1 [27].
Another contributing factor may be the much larger internal
surface area compared with external surface area (ca. 20-fold
larger). Species II should form more easily on the external
surface and act as an active intermediate when acetonitrile is

70
W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
used as the solvent [5]. Moreover, species II is more intrin-
sically active than species I, due to its higher electrophilic
character [5]. These characteristics make acetonitrile a better
solvent in the oxidation of cyclohexene (which occurs mainly
on the external surface of TS-1) than in the oxidation of 1-
hexene over TS-1. This is supported by the finding that sily-
lation slightly increased the conversion of cyclohexene over
TS-1 when methanol was used as solvent (Table 1, entry 9).
The selective silylation of the external surface of TS-1 by bulky
1
,1,1,3,3,3-hexamethyldisilazane could increase the hydropho-
bicity, favoring the formation of species I. It has been shown
that the silylation of Ti-MCM-41 led to a significant increase
Scheme 2. Mechanistic pathway for the oxidation of cyclohexene to cyclo-
hexanediol.
4
3
in the Q /Q ratio, which significantly increased its hydropho-
bicity and enhanced its activity for the oxidation of 1-hexene
and n-hexane [28]. Nevertheless, after silylation, considerable
amounts of OH groups remained on the exterior surface of TS-1
as a result of the steric limitation due to the large molecular
size of 1,1,1,3,3,3-hexamethyldisilazane. This made the exter-
nal surface of TS-1 still relatively hydrophilic even after sily-
lation. The preponderant methanol solvent strongly competes
with water for adsorption on Ti sites, resulting in lower activity
compared with that found with acetonitrile solvent.
the entrance of side pockets with such a large-molecule agent
produced a considerable steric constraint on the diffusion of
substrate and product molecules into and out of these side pock-
ets, giving rise to a slight decrease in activity (Table 1, entries
2
7 and 28).
4
.3. Effect of solvent on the catalytic performance of Ti-Beta
in the epoxidation of 1-hexene and cyclohexene
4
.2. Effect of solvent on the catalytic performance of Ti-MWW
As suggested previously [5], in Ti-Beta the concentration
of species II is higher than that of species I, due to the cat-
alyst’s hydrophilicity. This property may reasonably account
for the positive solvent effect of acetonitrile on the oxidation
of 1-hexene. However, with respect to cyclohexene oxidation,
this interpretation contradicts our finding that methanol solvent
gave much higher conversion than acetonitrile. The silylation of
Ti-Beta caused a drastic decrease in activity for the epoxidation
of alkenes, likely due to the strong influence of silylation on the
pore mouth, causing narrowing or blocking of the pore open-
ings. This is supported by the more profoundly decreased cy-
clohexene conversion in methanol solvent; the larger molecular
size of species I than species II led to a more severe geometric
constraint on the diffusion of cyclohexene molecules into the
channels once the intermediate species were formed near the
pore mouth by the interaction of species I with cyclohexene. As
a result, Ti species in the channels were less accessible to the
substrate molecules when methanol was used as the solvent, re-
sulting in very low conversion of cyclohexene.
The contrast in solvent effect between cyclohexene oxida-
tion and 1-hexene oxidation on Ti-Beta possibly arises from
different reaction mechanisms. Indeed, recently it was shown
that oxidation of cyclohexene occurred in two ways over TAPO-
5 (Scheme 2) [30]: trans-cyclohexanediol was produced from
the acid-catalyzed ring opening of the epoxide through nucle-
ophilic substitution, whereas the cis-isomer was formed via
a radical mechanism. Thus, the higher conversion attained
by using methanol as solvent may have resulted from the
radical mechanism, because polar MeOH can stabilize the
superoxo-titanium species, whereas MeCN can stabilize the
hydroperoxo/peroxo-titanium species [31]. To explore this pos-
sibility, we carried out cyclohexene oxidation in methanol sol-
vent in the presence of 1 mmol hydroquinone (a free-radical
in the epoxidation of 1-hexene and cyclohexene
In agreement with our previous results [8,18,21,22], Ti-
MWW was found to be much more active in acetonitrile than
in methanol in the epoxidation of 1-hexene because of its hy-
drophilicity (Table 1, entries 20 and 21) [18,22]. Silylation had
no marked effect on the catalytic results, because the reac-
tion occurred mainly inside the channels with distorted 10-MR
pore openings, hindering the silylation of internal Ti species
by bulky 1,1,1,3,3,3-hexamethyldisilazane. The epoxidation of
cyclohexene over Ti-MWW occurred mainly on its external
surface (similar to that for TS-1), due to the slightly smaller
pore openings of MWW-type materials compared with MFI-
type materials [29], and the poisoning by 2,4-dimethylquinoline
resulted in a drastically reduced activity for oxidation of cyclo-
hexene regardless of the solvent used (Table 1, entries 29 and
3
0) but did not significantly affect the activity in 1-hexene epox-
idation (Table 1, entry 24). The presence of numerous 12-MR
side pockets on the external surface of Ti-MWW led to a cy-
clohexene conversion between those obtained with TS-1 and
Ti-Beta. Note that the Ti species inside the side pockets were
located on the intracrystalline surface and thus had an envi-
ronmental state similar to that of the Ti species in TS-1; these
Ti species were relatively hydrophobic. As a result, methanol
solvent gave almost the same activity as acetonitrile solvent,
likely because both the hydrophilic Ti species on the exter-
nal surface and hydrophobic Ti species in the side pockets
could serve as active sites for the epoxidation of cyclohexene
(Table 1, entries 25 and 26). On silylation with 1,1,1,3,3,3-
hexamethyldisilazane, the external surface silanol groups were
silylated; therefore, the sites active for the cyclohexene epoxi-
dation became totally hydrophobic, making methanol the sol-
vent of choice (Table 1, entry 28). But the silylation close to

W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
71
Scheme 3. Interaction of MeCN with acid sites (T = Si or Ti).
Consequently, there must be a third reason why methanol
is the favored solvent in the oxidation of cyclohexene over Ti-
Beta. This factor is not clear at the moment; we can assume the
possible existence of a synergetic effect between Ti species and
acid sites resulting from defect sites in Ti-Beta when methanol
solvent is used. This effect may be too subtle to enhance the
rate of 1-hexene oxidation but may be strong enough to oxidize
more reactive cyclohexene as an internal alkene. These contra-
dictory results can be reasonably interpreted; the addition of a
certain amount of acetonitrile poisoned some of the acid sites
Fig. 10. Effect of trans-cyclopentanediol added to the reaction mixture on the
conversion of cyclohexene over Ti-Beta in acetonitrile solvent. (Reaction con-
◦
ditions: 60 C, 2 h, 0.05 g catalyst, 10 mL solvent, 10 mmol substrate and
1
0 mmol H O (31% aqueous solution).)
2 2
(
as shown in Scheme 3) [5], leading to a drastic diminution of
scavenger), and found that the conversion indeed declined (Ta-
ble 6), but only slightly. This demonstrates that the higher ac-
tivity obtained in methanol solvent itself does not demonstrate
the presence of the radical mechanism.
the synergetic effect and thus a marked reduction in activity.
The poisoning of acid sites by acetonitrile is strongly supported
by the increased epoxide selectivity with an increasing amount
of acetonitrile in the mixed solvents.
The competition of alcohol substrates with protic alcohol
solvents for adsorption on the Ti sites of titanosilicates can be
extended to the competition of the water molecules with alco-
hol substrates on hydrophilic Ti-Beta with the use of aprotic
solvents, as supported by the decreasing activity for the oxi-
dation of cyclohexanol on Ti-Beta with decreasing polarity of
aprotic solvents [5]. This suggests that another possible reason
may be that the formed diol competed with water present in
4
.4. Effect of the protic/aprotic nature of solvent on the
catalytic performance of titanosilicates
As reported previously, with respect to the protic alcohol
solvents, the activity of TS-1 for the epoxidation of 1-hexene
decreased in the order MeOH > EtOH > t-BuOH due to dimin-
ishing electrophilicity and increasing steric constraints [9]. Ba-
sically similar trends also were found for Ti-Beta and Ti-MWW,
but with small differences, possibly due to the larger pores of
Ti-Beta compared with TS-1 and the presence of side pock-
ets on the exterior surface of Ti-MWW. The increased epoxide
selectivity in the order MeOH < EtOH < t-BuOH can be at-
tributed to the increased steric hindrance to the alcoholysis of
epoxide with increasing alcohol molecular size; this was more
evident for cyclohexene oxidation, even though the epoxide se-
lectivity was not high (Table 3). When 1-hexene was oxidized
in aprotic solvents, the activity of Ti-Beta decreased with de-
creasing polarity, in agreement with results reported by Corma
et al. [5], in contrast with the trend in oxidation over TS-1.
This difference may be attributed to the strong organophilicity
of TS-1, which hinders formation of species II in strongly po-
lar aprotic solvents. Ti-MWW behaves like Ti-Beta and unlike
TS-1.
If species II were indeed the active intermediate, as hy-
pothesized previously for the oxidation of organics in apro-
tic solvents [5], then it would be reasonable to conclude that
the number of species II would increase with increasing wa-
ter content in the reaction mixture until saturation is reached.
Ti-MWW showed much higher activity in the oxidation of lin-
ear alkenes in MeCN solvent than in MeOH solvent, suggesting
that Ti-MWW was rather hydrophilic. Thus, it is not unexpected
that Ti-MWW gave a higher conversion of 1-hexene with in-
the added H O aqueous solution for readsorption on Ti sites,
2
2
leading to the decreased activity with the MeCN solvent due
to the very small amount of water in the reaction system. In
contrast, in methanol solvent, because diol and monomethyl
glycol ether were formed in much smaller amounts than the
methanol, and because the molecular sizes of these two prod-
ucts are much larger than that of methanol, they might not be
easy to coordinate to the Ti center to form the five-membered
cycle. Consequently, their effect on the oxidation of cyclohex-
ene would be negligible. To further confirm this point, we added
different amounts of trans-cyclopentanediol to the cyclohex-
ene oxidation system with acetonitrile solvent. As expected,
the conversion of cyclohexene decreased as the amount of cy-
clopentanediol increased (Fig. 10). This suggests that the ad-
sorption of product cyclohexanediol on Ti sites indeed may be a
factor in the low conversion of cyclohexene in acetonitrile sol-
vent. But the reduction was not significant enough to account
for the remarkable difference in solvent effect; the cyclohexene
conversion in methanol solvent was twice that in acetonitrile
solvent even after subtracting the effect of radical-based oxida-
tion. Moreover, this finding is not in agreement with the previ-
ous finding that addition of small amounts of acetonitrile into
the methanol solvent (e.g., MeCN/MeOH = 1.5/8.5 mL/mL)
resulted in a drastic decrease in activity similar to that obtained
in pure MeCN solvent.

72
W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
creasing water content in the reaction mixture. Corma et al.
reported that the higher the water content, the greater the ox-
idation activity of Ti-Beta for the epoxidation of 1-hexene [5].
Our result is consistent with that observation, although the ac-
tivity increase was greater for Ti-MWW than for Ti-Beta. Be-
cause TS-1 is organophilic, water in less than large amounts
did not affect the types and amounts of active species. Irrespec-
tive of the titanosilicates, due to the stronger nucleophilicity
of alcohol compared with water, protic alcohol solvents gen-
erally gave a lower epoxide selectivity compared with apro-
tic solvents (except for t-butyl alcohol, which demonstrated a
very high epoxide selectivity as a result of serious steric con-
straints).
Compared with a single solvent of methanol or acetonitrile,
mixed solvents composed of appropriate amounts of these two
agents enhanced the oxidation of 1-hexene over TS-1 and Ti-
Beta. This finding can be explained by assuming that although
TS-1 is basically hydrophobic, a certain amount of hydrophilic
sites originate from defect sites and surface silanol groups;
therefore, Ti sites located at these positions would give much
higher activity in acetonitrile than in methanol. As a result,
a suitable mixture of solvents makes all Ti sites form active
species, leading to the increase in activity. In contrast, Ti-Beta
is essentially hydrophilic, due to the presence of numerous de-
fect sites; however, some of the Ti sites are in hydrophobic
lattice sites, which favor methanol as a solvent. With respect
to Ti-MWW, the decreased activity with increasing amount of
methanol in the oxidation of 1-hexene possibly can be attributed
to the presence of lamellar structure and/or many defect sites
arising from deboronation, making almost all of the Ti sites rel-
atively hydrophilic.
5. Conclusion
Our findings demonstrate that the activity and selectivity
of titanosilicates for the epoxidation of alkenes depend on the
substrate as well as the nature and polarity of solvents. In
contrast to TS-1, both Ti-MWW and Ti-Beta exhibited much
higher activity and selectivity in acetonitrile than in methanol
for the oxidation of 1-hexene. The addition of water to the
reaction mixture increased the catalytic activity of Ti-MWW
and Ti-Beta for the oxidation of 1-hexene, whereas the mixed
solvents of methanol and acetonitrile were superior to an op-
timum single solvent for TS-1 and Ti-Beta. The differences in
these titanosilicates can be accounted for by their hydrophilic-
ity/hydrophobicity (except for the result of the oxidation of
cyclohexene over Ti-Beta). Acetonitrile solvent was more suit-
able for the epoxidation of cyclohexene on TS-1 than methanol,
because the reaction occurred mainly on the hydrophilic exter-
nal surface. Except for TS-1 in aprotic solvents, the titanosil-
icates demonstrated increased catalytic activity with increas-
ing solvent polarity. Acetonitrile solvent favored the forma-
tion of epoxide. H2O2 efficiency was affected by the solvent
and the catalyst structure; Ti-Beta exhibited much lower H2O2
efficiency than TS-1 and Ti-MWW regardless of the solvent
used.
Acknowledgments
This work was supported by Core Research for Evolutional
Science and Technology of JST. W.F. thanks the Japan Society
for Promotion of Science (JSPS) and Japan Science and Tech-
nology Agency (JST) for a postdoctoral fellowship. W.F. is also
grateful for the support of “Hundred Talents Project” of the Chi-
nese Academy of Sciences.
The conversion of hydrogen peroxide occurs by two routes:
the effective oxidation of substrate and simple decomposition.
Generally, the decomposition rate was lower in methanol than
in acetonitrile, although this was not evident for TS-1. This is in
agreement with previous findings [5] and may be ascribed to the
different active species; species II is more active than species I
for the decomposition of H2O2 [5]. In addition, the decompo-
sition of H2O2 also was seen to be related to the structure of
the titanosilicates. Ti-Beta promoted significantly greater de-
composition of H2O2 compared with Ti-MWW and TS-1. The
efficiency of H2O2 is a combined result of substrate oxidation
and decomposition; the lower the decomposition, the more the
amount consumed for substrate oxidation, and hence the higher
the efficiency. This property can account for the much higher
efficiency obtained in aprotic solvents (e.g., MeCN) compared
with protic alcohol solvents in the oxidation of alkenes con-
ducted on Ti-MWW. In contrast, TS-1 showed higher H2O2
efficiency in methanol than in acetonitrile, as a result of the high
activity and low decomposition rate. Although MeCN solvent
gave higher H2O2 selectivity on Ti-Beta than MeOH solvent,
the efficiency of H2O2 on Ti-Beta was still much lower than
that on Ti-MWW and TS-1, due to its intrinsic properties. In
summary, regardless of the solvent used, high oxidation activ-
ity led to high H2O2 efficiency.
References
[
[
1] M.G. Clerici, G. Bellussi, U. Romano, J. Catal. 129 (1991) 159.
2] A. Corma, M.A. Camblor, P. Esteve, A. Martinez, J. Perez-Pariente, J.
Catal. 145 (1994) 151.
[3] P. Wu, T. Tatsumi, T.T. Komatsu, T. Yashima, J. Catal. 202 (2001) 245.
[4] G. Bellussi, A. Carati, M.G. Clerici, G. Maddinelli, R. Millini, J.
Catal. 133 (1992) 220.
[
[
5] A. Corma, P. Esteve, A. Martinez, J. Catal. 161 (1996) 11.
6] T. Tatsumi, K. Asano, K. Yanagisawa, Stud. Surf. Sci. Catal. 84 (1994)
1861.
[
7] M.A. Camblor, M. Costantini, A. Corma, L. Gilbert, P. Esteve, A. Mar-
tinez, S. Valencia, Chem. Commun. (1996) 1339.
8] P. Wu, T. Tatsumi, J. Phys. Chem. B 106 (2002) 748.
9] M.G. Clerici, P. Ingallina, J. Catal. 140 (1993) 71.
[
[
[10] F. Bonino, A. Damin, G. Ricchiardi, M. Ricci, G. Spanò, R. D’Aloisio, A.
Zecchina, C. Lamberti, C. Prestipino, S. Bordiga, J. Phys. Chem. B 108
(2004) 3575.
[
11] C. Prestipino, F. Bonino, S. Usseglio, A. Damin, A. Tasso, M.G. Clerici,
S. Bordiga, F. D’Acapito, A. Zecchina, C. Lamberti, Chem. Phys. Chem. 5
(2004) 1799.
[12] S. Bordiga, F. Bonino, A. Damin, C. Lamberti, Phys. Chem. Chem. Phys. 9
(
2007) 4854.
13] A. Bhaumik, R. Kumar, P. Ratnasamy, Stud. Surf. Sci. Catal. 84 (1994)
883.
[14] S.B. Kumar, S.P. Mirajkar, G.C.G. Pais, P. Kumar, R. Kumar, J. Catal. 156
1995) 163.
[
1
(

W. Fan et al. / Journal of Catalysis 256 (2008) 62–73
73
[
[
[
[
[
[
15] A. Bhaumik, R. Kumar, J. Chem. Soc. Chem. Commun. (1995) 349.
16] M.G. Clerici, Top. Catal. 15 (2001) 257.
[24] G.P. Heitmann, G. Dahlhoff, W.F. Hölderich, J. Catal. 186 (1999) 12.
[25] C. Lamberti, S. Bordiga, A. Zecchina, G. Artioli, G. Marra, G. Spanò, J.
Am. Chem. Soc. 123 (2001) 2204.
[26] Y. Goa, P. Wu, T. Tatsumi, Chem. Commun. (2001) 1714.
[27] G. Artioli, C. Lamberti, G. Marra, Acta Crystallogr. B 56 (2000) 2.
[28] T. Tatsumi, K.A. Koyano, N. Igarashi, Chem. Commun. (1998) 325.
[29] Ch. Baerlocher, L.B. McCusker, D.H. Olson, Atlas of Zeolite Framework
Types, sixth revised ed., Elsevier, 2007, pp. 213, 235.
17] A. Thangaraj, S. Sivasanker, J. Chem. Soc. Chem. Commun. (1992) 123.
18] P. Wu, T. Tatsumi, Chem. Commun. (2002) 1026.
19] R. Millini, E.P. Massara, G. Perego, G. Bellussi, J. Catal. 137 (1992) 497.
20] C. Lamberti, S. Bordiga, A. Zecchina, A. Carati, A.N. Fitch, G. Artioli, G.
Petrini, M. Salvalaggio, G.L. Marra, J. Catal. 183 (1999) 222.
[21] W. Fan, P. Wu, S. Namaba, T. Tatsumi, Angew. Chem. Int. Ed. 43 (2004)
2
36.
[30] S. Lee, R. Raja, K.D.M. Harris, J.M. Thomas, B.F.G. Johnson, G. Sankar,
Angew. Chem. Int. Ed. 42 (2003) 1520.
[31] D. Srinivas, P. Manikandan, S.C. Laha, R. Kumar, P. Ratnasamy, J.
Catal. 217 (2003) 160.
[
[
22] W. Fan, P. Wu, S. Namaba, T. Tatsumi, J. Catal. 243 (2006) 183.
23] G. Ricchiardi, A. Damin, S. Bordiga, C. Lamberti, G. Spanò, F. Rivetti, A.
Zecchina, J. Am. Chem. Soc. 123 (2001) 11409.