Ti-bridged silsesquioxanes as precursors of silica-supported titanium oxide catalysts for the epoxidation of cyclooctene

Article abstract of DOI:10.1016/j.jcat.2010.08.010

Silica-supported titanium oxide catalysts were prepared by the controlled calcination of Ti-bridged polyhedral oligomeric silsesquioxanes (POSS) bearing four Si-O-Ti bonds, Ti[(c-C₅H₉)₇Si ₇O₁₁(OSiMe₂R)]₂ (1a; R = vinyl, 1b; R = methyl) and Ti[(c-C₅H₉)₈Si ₈O₁₃)]₂ (2), supported on various silicas, and these catalysts showed excellent activities for the epoxidation of cyclooctene by ^tBuOOH as an oxidant to selectively give cyclooctene oxide in high yields. On the other hand, catalysts prepared using a corner-capped-type Ti-containing POSS, CpTi[(c-C₅H₉)₇Si ₇O₁₂] (3, Cp = cyclopentadienyl), or tetraisopropoxytitanium (4) showed lower activities. The catalysts prepared from Ti-bridged POSS can be easily separated from the reaction mixture and reused without a significant loss of activity. These catalysts were also applicable for the epoxidation of limonene. A spectroscopic study revealed that calcination of Ti-bridged POSS supported on silica gave isolated tetrahedral Ti(IV) species, which would be responsible for the high activity, while calcination of silica-supported 3 or 4 gave aggregated surface titanium oxide species.

Full text of DOI:10.1016/j.jcat.2010.08.010

Journal of Catalysis 275 (2010) 280–287
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Ti-bridged silsesquioxanes as precursors of silica-supported titanium oxide
catalysts for the epoxidation of cyclooctene
⇑
Shuko Sakugawa, Kenji Wada , Masashi Inoue
Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto University, Katsura, Nishikyo-ku, Kyoto 615-8510, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 19 May 2010
Revised 25 July 2010
Accepted 18 August 2010
Available online 22 September 2010
Silica-supported titanium oxide catalysts were prepared by the controlled calcination of Ti-bridged poly-
hedral oligomeric silsesquioxanes (POSS) bearing four Si–O–Ti bonds, Ti[(c-C
5 9 7 7 2 2
H ) Si O11(OSiMe R)] (1a;
R = vinyl, 1b; R = methyl) and Ti[(c-C Si 13)] (2), supported on various silicas, and these catalysts
H
5 9
)
8
8
O
2
t
showed excellent activities for the epoxidation of cyclooctene by BuOOH as an oxidant to selectively give
cyclooctene oxide in high yields. On the other hand, catalysts prepared using a corner-capped-type Ti-
5 9 7 7
containing POSS, CpTi[(c-C H ) Si O12] (3, Cp = cyclopentadienyl), or tetraisopropoxytitanium (4) showed
Keywords:
Titanium
lower activities. The catalysts prepared from Ti-bridged POSS can be easily separated from the reaction
mixture and reused without a significant loss of activity. These catalysts were also applicable for the
epoxidation of limonene. A spectroscopic study revealed that calcination of Ti-bridged POSS supported
on silica gave isolated tetrahedral Ti(IV) species, which would be responsible for the high activity, while
calcination of silica-supported 3 or 4 gave aggregated surface titanium oxide species.
Ó 2010 Elsevier Inc. All rights reserved.
Silsesquioxane
Silica-supported catalyst
Epoxidation
Cyclooctene
Limonene
1
. Introduction
Siliceous group 4 transition metal catalysts, such as Ti-contain-
both the aerobic oxidation of benzylic alcohols in water [43] and
the hydroformylation of alkenes [44]. In this context, we could ex-
pect that it would be easier to prepare silica-supported monomeric
ing zeolites, are widely used in current chemical industries [1–3].
In particular, titanosilicates such as TS-1 generally show activity
Ti(IV) catalysts using Ti-bridged POSS in place of usual Ti-contain-
i
ing precursors such as Ti(O Pr)
4
[45].
that is superior to that of amorphous TiO
2
–SiO
2
in the epoxidation
We report here the preparation of silica-supported titanium
oxide catalysts using Ti-containing POSS as precursors. These oxide
catalysts showed excellent activities for the epoxidation of cyclo-
of small alkenes [3–5]. However, titanium-containing zeolites gen-
erally have steric limitations regarding applicable substrates be-
cause of their microporous nature. Therefore, the development of
novel titanium catalysts with suitable meso- and/or macropores,
which contain well-dispersed titanium species, is of great
significance.
On the other hand, metal-containing POSS (polyhedral oligo-
meric silsesquioxanes) have attracted attention as well-defined,
homogeneous models of solid siliceous catalysts [6–15]. In partic-
ular, the catalytic activities of Ti-containing POSS for the epoxida-
tion of alkenes have been examined [16–23]. In addition to their
activities as homogeneous catalysts, their application to heteroge-
neous catalysis using metal-containing POSS as precursors has
been examined [10,13,15,24–31]. The preparation of microporous
silicas containing well-dispersed metal oxide species by the con-
trolled calcination of metal-containing POSS has been reported
t
octene by BuOOH. The effects of various titanium sources on the
activities and physical properties of the thus-prepared oxide cata-
lysts by calcination were examined to identify suitable structures
for the required titanium precursors. Changes in the state of
Ti-containing POSS during calcination were also investigated.
2
. Experimental
2.1. Materials and methods
Ti-containing POSS were prepared and subsequent manipula-
tions were performed under an argon atmosphere using standard
Schlenk techniques. Dehydrated toluene (Wako) was used as
t
received. Cyclooctene (Nacalai Tesque), limonene (Wako), BuOOH
[
32–42]. Furthermore, the authors have prepared catalysts pos-
in a decane solution (5.5 M, Fluka), tetrakis(dimethylamido)tita-
nium (Gelest), tetraisopropoxytitanium (4; Nacalai Tesque),
sessing both mesopores and micropores by the impregnation of
metallic salts and POSS ligands onto suitable supports followed
by calcination, and these catalysts showed excellent activities for
5 9 7 7 9 3 5 9 8 8 2
(c-C H ) Si O (OH) (Aldrich), and (c-C H ) Si O11(OH) (Aldrich)
were obtained commercially and used without further purification.
A platinum catalyst (Karsted’s catalyst, (divinyltetramethyldisilox-
ane)platinum(0) (2.1–2.4 wt.%) in xylene) was obtained from Gel-
⇑
Corresponding author. Fax: +81 75 383 2479.
E-mail address: wadaken@scl.kyoto-u.ac.jp (K. Wada).
5 9 7 7 9 2 2
est. (c-C H ) Si O (OH) (OSiMe R) (R = vinyl [26], methyl [46]),
0
021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.08.010

S. Sakugawa et al. / Journal of Catalysis 275 (2010) 280–287
281
Ti[(c-C
18,26]), and CpTi[(c-C
on methods described in previous reports. Dimethylsiloxy-func-
tionalized silica (SiMe (H)-silica, Aldrich) and Cabosil (HS-5, Cabot)
5
H
9
)
7
Si
7
O
11(OSiMe
2
R)]
2
(1a; R = vinyl [26], 1b; R = methyl
O12] (3) [33] were synthesized based
(50 mm inner diameter). A UV cell (10 mm ꢁ 40 mm, inner thick-
ness 1.0 mm) equipped with a branched chamber and a stop valve
was used to avoid any contact with moisture. The catalyst submit-
ted for measurement was heated in the branched chamber at 473 K
under atmospheric pressure for 30 min and evacuated at 673 K fol-
lowed by treatment under 100 Torr of oxygen at 673 K for 1 h. The
catalyst was then transferred to the UV cell, and spectra were taken
in vacuo. Barium sulfate (Nacalai Tesque) was used as a background.
All of the spectra were modified in terms of the Kubelka–Munk
function. Leaching of titanium species from the catalysts during
the reaction was investigated by an ICP atomic emission spectro-
scopic analysis using a Shimadzu ICPS-1000III analyzer.
[
H
5 9
)
7
Si
7
2
were used as received. Mesoporous MCM-41 was prepared based
on the method in the literature [47]. JRC silicas were obtained
from the Catalysis Society of Japan. The BET surface areas of
JRC-SIO-1, -5, -6, -7, and -8 were 104, 174, 103, 88, and
2
ꢀ1
3
16 m g , respectively.
2.2. Synthesis of Ti[(c-C
5
H
9 8 8
) Si O
13)]
2
(2)
To a solution of (c-C
anhydrous toluene (25 cm ), tetrakis(dimethylamido)titanium
168 mg, 0.85 mmol) in anhydrous toluene (20 cm ) was dropwise
5
H
9
)
8
Si
8
O
11(OH)
2
(1481 mg, 1.5 mmol) in
3
2.4. Preparation of an oxide catalyst from silica-tethered Ti-POSS
3
(
added, and the mixture was stirred at room temperature overnight.
SiMe
2 2
(H)-functionalized silica gel (SiMe (H)-silica, 1.0 g) was
After the solvent was evaporated, the resulting off-white solid was
3
treated with a toluene solution (30 cm ) of a Ti-bridged POSS
1a) (103 mg, 0.052 mmol) in the presence of Karsted’s catalyst
0.0022 mmol, 0.040 cm ) under an Ar atmosphere. The mixture
was stirred at 45 °C for 24 h. Elemental analysis of the support (C
.83%, H 1.37%) indicated the presence of ca. 3.2 SiMe (H) groups
in a unit surface area of 1 nm . After filtration, washing of the
resulting solid three times with toluene in air and subsequent vac-
uum-drying gave a tethered catalyst that was designated as 1a-Si-
3
extracted with hexane (35 cm ). Filtration using a 0.45-
lm PTFE fil-
(
(
ter gave a clear solution. After the evaporation of hexane, 2 was ob-
3
tained by recrystallization by slow diffusion of acetonitrile into a
1
toluene solution. Yield 89%. M.p. 260 °C (decomp.). H NMR
3
2
(
400 MHz, C
6
D
6
, 25 °C) d 1.94–1.55 (128 H, br m), 1.22–1.14 (16 H,
br m); C{ H} NMR (75 MHz, C , 25 °C) d 27.81, 27.76, 27.73,
7.59, 27.36, 27.35, 27.31 (CH of Cy), 23.39, 23.02, 22.49 (1: 2: 1,
CH of Cy); Si{ H} NMR (76 MHz, C , 0.02 M Cr(acac) , 25 °C) d
64.83, ꢀ67.65, ꢀ68.22 (1: 2: 1). Anal. Calcd for C80
2019.22): C, 47.59; H, 7.19. Found C, 47.32; H, 6.93.
2
1
3
1
6 6
D
2
2
2
9
1
6
D
6
3
Me
solutions after hydrosilylative condensation, indicating that all of
a was successfully tethered onto silica. The calcination of 1a-Si-
Me (H)-silica in air at 550 °C for 2 h afforded an oxide catalyst,
a-SiMe (H)-silica-cal.
2
(H)-silica. There was no sign of 1a in the filtrate and washing
ꢀ
144
H O26Si16Ti
(
1
2
2.3. Physical and analytical measurements
1
2
Solution-phase NMR spectra were recorded on JEOL JNM-EX-
2
.5. Preparation of oxide catalysts from silica-supported Ti-POSS
1
13
4
00 instruments. H and C spectra were referenced to internal
. Chemical shifts
for the Si nuclei were referenced to an external SiMe resonance.
solvent resonances and reported relative to SiMe
4
A typical preparation procedure of the silica-supported tita-
2
9
4
2
nium oxide catalyst using Ti-POSS is as follows: SiMe (H)-silica
Elemental analyses were performed at the Microanalytical Center
of Kyoto University. The TG-TDA study was performed using a RIG-
AKU TG8120 system. The sample (ca. 5 mg) was heated at a rate of
3
(
1.0 g) was impregnated with a toluene solution (3.0 cm ) of 1a
(
98 mg, 0.050 mmol) at 60 °C in air (however, the impregnation
of 4 was performed in an Ar atmosphere to avoid the hydrolysis
of 4 caused by adventitious contact with moisture). Calcination
of the resulting powder (powder derived from 4, as well) in air at
ꢀ1
3
ꢀ1
1
0 °C min under a stream of air (50 cm min ). Diffuse reflec-
tance infrared Fourier transform (DRIFT) spectra were recorded
using a Nicolet Magna-IR 560 FT-IR spectrometer. An uncalcined
sample (ca. 10 wt.%, diluted with KBr powder) was heated in air
5
50 °C for 2 h afforded an oxide catalyst (Si/Ti molar ratio; 350)
that was designated as 1a@SiMe (H)-silica-cal.
2
ꢀ1
at an approximate rate of 10 °C min , and the spectrum was re-
corded after the sample was held at the prescribed temperature
for 10 min. X-ray fluorescence (XRF) analysis was performed using
a Spectro XEPOS-MS II (25 W).
2.6. Catalytic epoxidation of cyclooctene
All of the reactions were performed with the use of hot stirrers
The oxide catalysts were analyzed by nitrogen gas adsorption,
XRD, XPS, and UV–Vis spectroscopy. Nitrogen adsorption/desorp-
tion isotherms were obtained with a computer-controlled auto-
matic gas sorption system (Quantachrome NOVA 4200e). Samples
were degassed at 573 K for 30 min just before the measurements.
The mean pore diameter was calculated from the pore volume
and the specific surface area assuming that the pore was cylindrical.
equipped with cooling blocks for refluxing the solvent. Prescribed
amounts of the catalyst (0.50 mol% as Ti atom) and cyclooctene
3
(
1.0 mmol) in a toluene solution (1.5 cm ) were taken into a
3
2
0 cm glass Schlenk tube under an Ar atmosphere. This mixture
t
was heated and allowed to equilibrate at 60 °C for 10 min. BuOOH
3
(
1.0 mmol) in decane (1.2 cm ) was then added by the use of a
plastic pipette. The mixture was stirred for 4 h, and the Schlenk
tube was cooled rapidly in an ice bath. The products were identi-
fied by the use of GC–MS (Shimadzu GC–MS Parvum 2, Zebron
ZB-1 capillary column, i.d. 0.25 mm, length 30 m, 50–250 °C) and
quantified by gas–liquid chromatography (GL Sciences GC353,
Inertcap 17 capillary column, i.d. 0.25 mm, length 30 m, 50–
X-ray powder diffraction patterns were recorded using Cu Ka radi-
ation (30 kV, 30 mA) and a carbon monochromator (XRD: XD-D1,
Shimadzu). X-ray photoelectron spectra (XPS) of the catalysts were
acquired using an ULVAC-PHI 5500MT system equipped with a
hemispherical energy analyzer. Samples were mounted on indium
foil and then transferred to an XPS analyzer chamber. The residual
gas pressure in the chamber during data acquisition was less than
2
50 °C) using biphenyl as an internal standard.
ꢀ
8
ꢀ2
1
ꢁ 10 Torr (1 Torr; 133.3 N m ). The spectra were measured
at room temperature using Mg K radiation (1254 eV) generated
a
3. Results and discussion
by an X-ray tube operating at 15 kV, 400 W. The electron take-off
angle was set at 45°. Binding energies were referenced to the C 1s
level of residual graphitic carbon [48]. UV–Vis diffuse reflectance
spectra were measured with a JASCO V-650 spectrophotometer,
in which reflected beams were gathered by an integrating sphere
3.1. Catalytic activity for the epoxidation of cyclooctene
Silica-supported titanium oxide catalysts were prepared by
the calcination (550 °C, 2 h) of Ti-containing POSS tethered or

2
82
S. Sakugawa et al. / Journal of Catalysis 275 (2010) 280–287
supported on various silicas. A silica-tethered catalyst, designated
as 1a-SiMe (H)-silica (see Scheme 1), was prepared by the Pt-cat-
alyzed hydrosilylative condensation of commercially available
Table 1
Catalytic activity of Ti-containing POSS-based catalyst for the epoxidation of
cyclooctene by BuOOH.
2
t
a
ꢀ
dimethylsilyl-functionalized silica with vinylsilyl groups of 1a.
Entry
Catalyst
Reaction
time (h)
Epoxide
TOF (h 1
)
yield (%)b
An oxide catalyst, 1a-SiMe
cination of 1a-SiMe (H)-silica in air at 550 °C for 2 h. On the other
hand, an oxide catalyst designated as 1a@SiMe (H)-silica-cal was
2
(H)-silica-cal, was prepared by the cal-
1
2
3
4
5
6
7
8
9
1a (homogeneous)
1a (homogeneous)
1b (homogeneous)
2 (homogeneous)
1a-SiMe
1a-SiMe
1a@SiMe
1a@SiMe
1b@SiMe
4
41
31
33
29
41
70
75
53
90
86
60
39
21
41
44
39
21
35
38
71
45
43
30
20
2
1.5
1.5
1.5
4
4
4
1.5
4
4
4
4
2
prepared by the usual impregnation method from a toluene solu-
tion of 1a, followed by calcination.
2
(H)-silica
(H)-silica-cal
The activities of these catalysts for the epoxidation of cyclooc-
2
t
2
(H)-silica-cal
(H)-silica-cal
tene were examined by the use of BuOOH as an oxidant (Eq.
2
(
1)). The results are summarized in Table 1. In all of the cases
2
(H)-silica-cal
(H)-silica-cal
examined, cyclooctene oxide was produced selectively, and the
formation of byproducts was negligible. For comparison, the activ-
ity of 1a as a homogeneous catalyst was examined, and this pro-
duced cyclooctene oxide in yields of 31% and 41% when the
reaction was conducted for 1.5 h and 4 h, respectively (entries 1
and 2). Other Ti-containing POSS, 1b and 2 (see Scheme 2), showed
similar activities (entries 3 and 4). The activities of Ti-containing
10
2@SiMe
3@SiMe (H)-silica-cal
4@SiMe
2
11
12
2
2
(H)-silica-cal
a
The reaction conditions are shown in Eq. (1).
Yield based on cyclooctene determined by GLC.
b
Scheme 2. Ti-containing POSS used for the preparation of oxide catalysts.
POSS as homogeneous catalysts for the epoxidation by organic
peroxides have been previously reported [16–23]. Silica-tethered
1
1
a showed a catalytic activity similar to that of 1a, indicating that
a was tethered without a loss of activity (entry 5). An oxide
Scheme 1. Preparation of a silica-tethered catalyst and an oxide catalyst.
2
catalyst, 1a-SiMe (H)-silica-cal, showed a significant activity

S. Sakugawa et al. / Journal of Catalysis 275 (2010) 280–287
283
(
70%; entry 6) than either the homogeneous or tethered Ti-con-
Table 2
Effect of silica support and Ti precursors for the epoxidation of cyclooctene by
BuOOH.
taining POSS catalysts. Furthermore, an oxide catalyst prepared
by the usual impregnation method (1a@SiMe
slightly higher activity (yield of epoxide; 75%) than 1a-SiMe
silica-cal (entry 7), and a high TOF of 71 h was recorded in the
reaction for 1.5 h (entry 8). Note that re-examination of the cata-
lytic run shown in entry 7 using the same batch of the catalyst re-
sulted in 78% yield, and the reaction using another batch of
t
a
2
(H)-silica-cal) showed
(H)-
Entry
Catalyst
Reaction time
(h)
Epoxide yield
(%)
TOF
(h
2
b
ꢀ1
ꢀ
1
)
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
1a@Cabosil-cal
1a@Cabosil-cal
1a@MCM-41-cal
1a@JRC-SIO-1-cal
1a@JRC-SIO-5-cal
1a@JRC-SIO-6-cal
1a@JRC-SIO-7-cal
1a@JRC-SIO-8-cal
1b@Cabosil-cal
2@Cabosil-cal
4
1.5
4
4
4
4
4
4
4
4
4
1.5
4
1.5
4
88
68
81
68
77
71
66
75
87
83
87
58
72
46
53
44
91
41
34
39
36
33
38
44
42
44
77
36
61
27
1
2
a@SiMe (H)-silica-cal afforded the epoxide in 79% yield. These re-
sults indicate good reproducibility of the results in the present
system.
1
1
1
1
1
1
3@Cabosil-cal
3@Cabosil-cal
4@Cabosil-cal
4@Cabosil-cal
ð1Þ
4@MCM-41-cal
a
The reaction conditions are shown in Eq. (1).
Yield based on cyclooctene determined by GLC.
b
The above results indicate that the tethering of 1a through
covalent bonding did not improve the activity of the resulting cat-
alyst. Therefore, we examined the effects of titanium sources on
the activity of the catalysts prepared by the impregnation method.
As shown in entries 9 and 10, oxide catalysts prepared using 1b
and 2 showed higher activities (90% and 86% yields, respectively)
than that prepared using 1a, while the catalyst prepared using a
corner-capped Ti-containing POSS (3) (see Scheme 2) showed a
moderate activity (entry 11), which indicates that Ti-bridged POSS
showed excellent activity to afford the desired epoxide in 81% yield
ꢀ1
(
TOF of 41 h for 4 h, entry 3). This activity is comparable to those
of MCM-41-based titanium catalysts reported in previous studies
[
10,25,30,45]. For example, a Ti-MCM-41 catalyst (Si/Ti molar ra-
ꢀ1
tio; 100) showed a TON of ca. 39 h for the epoxidation of cyclo-
octene by BuOOH for 5 h at 60 °C (TON; ca. 196 for 5 h) [45]. The
t
2
are excellent titanium precursors. The activity of 4@SiMe (H)-sil-
activities of the catalysts prepared using JRC silicas were moderate.
ica-cal was poor, even though impregnation was performed under
an Ar atmosphere to avoid the hydrolysis of 4 caused by contact
with moisture (entry 12). The catalyst prepared by the impregna-
tion of 4 in an air atmosphere gave the epoxide in the lower yield
A Cabosil-supported catalyst derived from 3 showed a slightly low-
er TOF in the reaction for 1.5 h, 77 h (entry 12), compared to that
of the catalyst prepared using 1a (entry 2). Both the Cabosil- and
MCM-41-supported catalysts prepared using 4 as a titanium
source showed lower activities than the catalysts prepared using
ꢀ1
(
26%). On the other hand, Ti-containing POSS can be safely handled
in air, which reflects high stability.
Except for 1a-SiMe (H)-silica, these catalysts did not exhibit
1
a, 1b, or 2 (entries 13–15).
2
It is important to determine whether the reaction actually oc-
activity for epoxidation using an aqueous solution of hydrogen per-
oxide. Severe non-productive decomposition of hydrogen peroxide
was observed in the presence of oxide catalysts such as 1a-Si-
Me (H)-silica-cal and 1a@SiMe (H)-silica-cal. As reported previ-
2 2
ously, Ti-POSS do not show activity as homogeneous catalysts
when hydrogen peroxide is employed as an oxidizing reagent
curs on the solid surface. Hot filtration of the solid catalyst after
the reaction in the presence of 1a@Cabosil-cal was allowed to pro-
ceed at 60 °C for 0.5 h completely suppressed further progress of
the reaction (Fig. 1). This result strongly suggests that the catalyst
worked heterogeneously.
The epoxidation of a larger alkene, limonene, was performed
[
23,27]. In aqueous solutions, Ti-POSS molecules severely aggre-
(
Eq. (2)). The epoxidation selectively occurred at the tri-substituted
gate, which would be one reason for the lack of the catalytic activ-
ity. On the other hand, Ti-POSS catalysts tethered or immobilized
onto polydimethylsiloxane [27], mesoporous silica–polymer com-
posites [28], silyl-functionalized silicas [29], and gels composed
of POSS molecules [31] were reported to show excellent activities
for the epoxidation of alkenes by an aqueous hydrogen peroxide
solution. Note that all of these catalysts are heterogeneous and
strongly hydrophobic because of the presence of organic groups
on their surface. In the present study using oxide catalysts, the ac-
tive Ti species are present on the hydrophilic surface of the sup-
ported oxide catalysts, which would prevent the access of
alkenes to the Ti species, and facilitate non-productive decomposi-
tion of hydrogen peroxide.
C–C double bond [30]. As shown in Table 3, the Cabosil-supported
catalysts prepared using 1a or 1b showed slightly higher activity
than that of 1a as a homogeneous catalyst, whereas the catalyst de-
rived from 4 showed the lowest activity. These results suggest the
Table 2 shows the effect of the silica support on the catalytic
activity for the epoxidation of cyclooctene. 1a was impregnated
onto seven silica supports [MCM-41, Cabosil, and JRC-SIO-1, -5, -
6
, -7, and -8], and this was followed by calcination at 550 °C.
Among the catalysts examined, Cabosil-supported catalysts pre-
pared using Ti-bridged POSS showed the highest activities (83–
ꢀ1
8
8% yields for 4 h, entries 1, 9, and 10). A high TOF of 91 h was
achieved in the reaction for 1.5 h (entry 2). The ICP-AES analysis re-
vealed that no titanium species leached into the solution after the
reaction using 1a@Cabosil-cal (entry 1). 1a@MCM-41-cal also
Fig. 1. Time course of the epoxidation of cyclooctene in the presence of 1a@Cabosil-
cal (a) without and (b) with hot filtration of the catalyst 0.5 h after the reaction
started.

2
84
S. Sakugawa et al. / Journal of Catalysis 275 (2010) 280–287
Table 3
Effect of Ti catalysts for the epoxidation of limonene by BuOOH.a
t
Epoxide yield (%)b
TOF (h )
ꢀ1
Entry
Catalyst
1
2
3
4
5
1a (homogeneous)
1a@Cabosil-cal
1b@Cabosil-cal
2@Cabosil-cal
23
27
24
22
15
12
14
12
11
8
4@Cabosil-cal
a
The reaction conditions are shown in Eq. (2).
Yield based on cyclooctene determined by GLC.
b
applicability of the supported catalysts for the epoxidation of
sterically demanding alkenes.
ð2Þ
One of the major advantages of solid catalysts is the reusability.
After the reaction shown in entry 1 of Table 2, the solid catalyst
was separated from the reaction mixture by centrifugation and
3
washed three times with toluene (2.0 cm ). The vacuum-dried cat-
alyst showed slightly lower activity than the fresh catalyst to give
the epoxide in 78% yield. Subsequent calcination at 550 °C for 2 h
increased the activity of the recovered catalyst and the epoxide
was produced in a yield of 81%, indicating that the catalyst was
reusable.
3.2. Characterization of the oxide catalysts prepared using Ti-POSS
To examine the effects of Ti-POSS on the properties of the oxide
catalysts, characterization was performed by XRD, nitrogen gas
adsorption, UV–Vis spectroscopy, and XPS. The powder XRD pat-
terns of all of the catalysts did not show any peaks due to crystal-
line titanium oxides, indicating that titanium oxide species were
highly dispersed on the surface of the catalysts. The UV–Vis spectra
reported in Fig. 3 (see below) will be more informative on the local
environment of Ti species. Table 4 lists the results of nitrogen
adsorption measurements of selected catalysts and supports, and
Fig. 2 shows the nitrogen adsorption/desorption isotherms of 1a-
Fig. 2. Nitrogen adsorption/desorption isotherms of (a) 1a-SiMe
2
(H)-silica, (b)
1
a-SiMe (H)-silica-cal, (c) 1a@SiMe (H)-silica-cal, (d) Cabosil, and (e) 1a@Cabosil-
2
2
cal. Filled and open symbols show adsorption and desorption data, respectively.
to isolated tetrahedral titanium species [53]. Fig. 3 shows the UV–
Vis diffuse reflectance spectra of selected catalysts under dehy-
SiMe
Cabosil, and 1a@Cabosil-cal. All of the catalysts supported on Si-
Me (H)-silica showed typical type-IV isotherms [49], indicating
2 2 2
(H)-silica, 1a-SiMe (H)-silica-cal, 1a@SiMe (H)-silica-cal,
drated conditions. The oxide catalyst, 1a@SiMe
showed a sharp band at 220 nm. 1a-SiMe (H)-silica-cal showed a
peak that was very similar to that of 1a@SiMe (H)-silica-cal. These
spectra closely resemble those reported for titanosilicates [50,52],
and this result clearly indicates that both 1a-SiMe (H)-silica-cal
and 1a@SiMe (H)-silica-cal contain isolated tetrahedral titanium
2
(H)-silica-cal,
2
2
the presence of mesopores. The isotherms of Cabosil and 1a@Cabo-
sil-cal were of type-II [49], which indicates the presence of both
meso- and macropores. In particular, Cabosil possesses a large
macropore volume, which would be one reason for the superior
activity of Cabosil-supported catalysts, since macropores are
advantageous in terms of diffusion of the substrates and the
products.
2
2
2
species on their surface. The Cabosil- and MCM-41-supported cat-
alysts prepared using 1a also showed absorption at around
2
220 nm. On the other hand, 3@SiMe (H)-silica-cal and 4@Si-
While the formation of micropores by the calcination of POSS
molecules has been previously reported [32–44], an increase in
the BET surface area by the calcination of silica-supported POSS
was not recognized from the results shown in Table 4, probably be-
cause of a low loading level of POSS.
Me (H)-silica-cal both showed a major band at around 260 nm,
2
which has been assigned to octahedral titanium species [51], to-
gether with shoulders at 225 nm and 310 nm. The former shoulder
is considered to be due to isolated tetrahedral titanium species,
and the latter is due to [TiO
the formation of octahedral titanium ions in oxide clusters on the
surface of 3@SiMe (H)-silica-cal and 4@SiMe (H)-silica-cal. The
spectrum of 4@Cabosil-cal showed very similar to those of 3@Si-
Me (H)-silica-cal and 4@SiMe (H)-silica-cal.
The XPS study revealed the presence of Ti(IV) species on the
surface of all of the catalysts examined. Fig. 4 shows selected re-
sults. Remarkably, 1a@SiMe (H)-silica-cal showed a Ti 2p3/2 band
2 n
] clusters [54]. These results suggest
UV–Vis spectroscopy is a very powerful tool for investigating
the local structure around titanium species in TiO
2
–SiO
2
mixed oxi-
2
2
des [50–52]. The LMCT (ligand–metal charge transfer) transition
from O to Ti in isolated tetrahedral titanium species is usually
observed at a shorter wavelength compared to that of octahedral
titanium species. For example, the LMCT transitions of TS-1 and
other titanosilicates are observed at 200–220 nm and are assigned
2
2
2

S. Sakugawa et al. / Journal of Catalysis 275 (2010) 280–287
285
2
1
0
.5
2
(
c)
(
a)
.5
1
(
b)
(
b)
c)
(
d)
(
a)
.5
0
(
2
00
400
600
800
Wavelength (nm)
472
467
462
457
452
447
Binding energy (eV)
1
.8
.6
.4
.2
0
2 2
Fig. 4. Ti 2p spectra of (a) 1a@SiMe (H)-silica, (b) 3@SiMe (H)-silica-cal, and (c)
4
@SiMe (H)-silica-cal.
2
0
0
0
0
(
e)
(
g)
(
f)
2
00
400
600
800
Wavelength (nm)
Fig. 3. UV–Vis diffuse reflectance spectra of (a) 1a@SiMe
2
(H)-silica-cal, (b)
3@SiMe
2 2 2
(H)-silica-cal, (c) 4@SiMe (H)-silica-cal, (d) 1a-SiMe (H)-silica-cal, (e)
1
a@Cabosil-cal, (f) 4@Cabosil-cal, and (g) 1a@MCM-41-cal.
Table 4
Summary of the results of nitrogen adsorption measurements of the selected catalysts
and supports.
Catalyst
BET S.A.
Mean pore
diameter (nm)
Total pore
2
ꢀ1
a
volume (cm3
g )
ꢀ1
(
m
g
)
SiMe
SiMe
2
(H)-silica
(H)-silica-cal
394
420
410
444
394
399
396
413
409
320
276
303
6.0
5.9
5.4
5.9
5.6
5.7
5.6
5.8
5.8
10.8
9.4
9.3
2.6
2.4
0.55
0.62
0.55
0.66
0.55
0.57
0.55
0.60
0.59
0.86
0.65
0.71
0.90
0.83
2
1
1
1
1
2
3
4
a-SiMe
a-SiMe
2
(H)-silica
2
(H)-silica-cal
a@SiMe
b@SiMe
@SiMe
@SiMe
@SiMe
2
(H)-silica-cal
(H)-silica-cal
2
2
(H)-silica-cal
2
(H)-silica-cal
2
(H)-silica-cal
Cabosil
1
4
a@Cabosil-cal
@Cabosil-cal
MCM-41
a@MCM-41-cal
1396
1405
1
Fig. 5. TG-DTA profiles of (a) 1a and (b) 3 in a stream of air.
a
The mean pore diameter was calculated from the pore volume and the specific
surface area assuming that the pore was cylindrical.
Changes in the structures of Ti-POSS during calcination were
at 460.0 eV, which is significantly higher than those of 3@Si-
investigated by TG-DTA, DRIFT, and NMR measurements. Since
several POSS have been reported to be volatile in spite of their rel-
atively large molecular weights [6,7,9], the loss of Ti and Si species
during the calcination of Ti-POSS was examined by treating 1a, 1b,
2, and 3 in air at 550 °C for 2 h in the box furnace. The ceramic
2 2
Me (H)-silica-cal and 4@SiMe (H)-silica-cal (458.5 eV). Since Ti
2
p
3/2 peaks centered at ca. 460 eV and 458–459 eV have been
assigned to a framework tetrahedral titanium and an extraframe-
work octahedral titanium phase for titanosilicates, respectively
[
53,55,56], the present result is consistent with the formation of
2 2
yields of SiO –TiO from 1a (49.6%) and 1b (49.4%) were slightly
tetrahedral titanium species on the surface of 1a@SiMe
2
(H)-sil-
lower than the theoretical values (53.0% and 53.8% for 1a and 1b,
ica-cal.
respectively). The XRF analyses showed that Si/Ti atomic ratios

2
86
S. Sakugawa et al. / Journal of Catalysis 275 (2010) 280–287
ꢀ
1
ꢀ1
of the resulting oxides from 1a and 1b were 14.8(±0.2) and
4.9(±0.1), respectively. These results indicate that by the calcina-
tion of both Ti-POSS, 7(±1)% of Si was lost, while 100(±2)% and
at 2950 cm
and 2870 cm , these bands almost disappeared
1
above 350 °C. This result is consistent with that of the TG-DTA
study, which indicates the combustion of organic substituents in
this temperature range. In the spectrum of 3, a C–H out-of-plane
bending vibration band characteristic of a Cp group is observed
at about 806 cm . This peak disappeared when the sample was
heated to 300 °C. This result supports the argument on the TG
study, which indicated the decomposition of the Cp group of 3 at
this temperature range. Again, C-H stretching vibrations of ali-
phatic C–H bonds disappeared above 350 °C because of the com-
plete combustion of organic groups. Note that a broad, strong
band at about 940 cm is observed for the spectra of both 1a
and 3 above 350 °C, which is often reported to be a good finger-
print of the existence of framework titanium species in titanosili-
cates [51,57,58].
Loss of the Cp group from 3 at around 300 °C was also con-
firmed by H and C NMR studies. The treatment of 3 at 300 °C
9
1
9(±1)% of Ti was remaining in the resulting oxides from 1a and
b, respectively. During the calcination of 1a and 1b, trimethylsilyl
ꢀ1
or dimethylvinylsilyl groups in 1a or 1b would be cleft at relatively
low temperature and then sublime, as revealed by the TG study
(
2
(
see below). On the other hand, the ceramic yields of oxides from
(52.5%) and 3 (51.0%) were very close to the theoretical values
51.6% and 50.8%, respectively). These results indicate the absence
of loss of Ti species during the calcination of POSS at 550 °C for 2 h.
As shown in Fig. 5, the TG-DTA study demonstrated the step-
wise decomposition of 1a and 3 in an air stream. Note that
TG-DTA measurements of supported Ti-POSS did not give any dis-
tinct TG and DTA responses, probably because of the low loading
levels. For 1a, decreases in weight by 7% and 42% were observed
at 300–350 °C and 400–600 °C, respectively. A slightly exothermic
decrease at 300–350 °C would be due to the oxidative cleavage of
dimethylvinylsilyl groups, and a significant loss of weight above
ꢀ
1
1
13
ꢀ1
for 10 min (heating rate 10 °C min ) in an air stream gave a
slightly brownish powder, which was soluble in CDCl except for
a trace amount of brown solid. In the H NMR spectrum, the peak
of a Cp ring (6.43 ppm) completely disappeared, while resonances
attributed to cyclopentyl groups remained. The C NMR spectrum
did not show the peak assignable to the Cp ring of 3.
3
1
4
00 °C is due to the combustion of cyclopentyl groups. The decom-
position of 3 also proceeded in a stepwise manner: an initial endo-
thermic loss of weight by 6% at ca. 315 °C and a significantly
exothermic loss above 320 °C.
1
3
The DRIFT spectra of 1a and 3 treated in air at various temper-
atures are shown in Fig. 6. While the spectrum of uncalcined 1a
showed strong C–H stretching vibrations due to cyclopentyl groups
As shown in the first part of the present study, the oxide cata-
lysts prepared using 1a, 1b, and 2 as titanium sources showed
activities superior to those prepared from 3 and 4. The combined
(a)
1
KM unit
r.t.
o
2
2
3
00 C
o
50 C
o
00 C
o
3
50 C
3100
3000
2900
2800
2700
1200 1100 1000
900
800
700
-
1
-1
Wavenumber (cm )
Wavenumber (cm )
(b)
1
KM unit
r.t.
o
2
2
00 C
o
50 C
o
3
3
00 C
o
50 C
3100
3000
2900
2800
2700
1200 1100 1000
900
800
700
-
1
-1
Wavenumber (cm )
Wavenumber (cm )
Fig. 6. DRIFT spectra of (a) 1a and (b) 3 recorded in air at elevated temperatures. Samples were diluted by KBr to be ca. 10 wt.%, and spectra in the region exceeding three
Kubelka–Munk units are omitted.

S. Sakugawa et al. / Journal of Catalysis 275 (2010) 280–287
287
results of characterization clearly indicate the selective formation
of isolated tetrahedral titanium(IV) species on the surface of the
oxide catalysts from 1a, 1b, and 2, and such titanium species would
be responsible for the high catalytic activity. Therefore, Ti-bridged
POSS are expected to act as precursors of ‘‘titanosilicate-like” spe-
cies. One common feature of 1a, 1b, and 2 is that the titanium atom
is surrounded by bulky POSS cages through four Ti–O–Si bonds.
During calcination, these POSS moieties would be transformed into
small silica clusters, which might act as an inorganic separator of
titanium species [43,44]. This would prevent the aggregation of
titanium oxide species and thus promote the formation of isolated
titanium(IV) species on the silica surface. Note that the TG-DTA
study of 3 revealed that the Cp (cyclopentadienyl) group decom-
posed at around 315 °C, which is a lower temperature than that
for the combustion of cyclopentyl groups (ca. 400 °C). The loss of
the Cp group from 3 in the present study might enable the aggre-
gation of Ti species.
[11] V. Lorenz, A. Fischer, S. Giessmann, J.W. Gilje, Y. Gun’ko, K. Jacob, F.T.
Edelmann, Coord. Chem. Rev. 206–207 (2000) 321.
[
[
12] R. Duchateau, Chem. Rev. 102 (2002) 3525.
13] C. Copéret, M. Chabanas, R.P. Saint-Arroman, J.M. Basset, Angew. Chem. Int. Ed.
42 (2003) 156.
[
[
14] R.W.J.M. Hanssen, R.A. van Santen, H.C.L. Abbenhuis, Eur. J. Inorg. Chem. (2004)
675.
15] K. Wada, T. Mitsudo, Catal. Survey Asia 9 (2005) 229.
[16] H.C.L. Abbenhuis, S. Krijnen, R.A. van Santen, Chem. Commun. (1997) 331.
[
17] T. Maschmeyer, M.C. Klunduk, C.M. Martin, D.S. Shephard, J.M. Thomas, B.F.G.
Johnson, Chem. Commun. (1997) 1847.
[
18] M. Crocker, R.H.M. Herold, A.G. Orpen, Chem. Commun. (1997) 2411.
[19] J.M. Thomas, G. Sankar, M.C. Klunduk, M.P. Attfield, T. Maschmeyer, B.F.G.
Johnson, R.G. Bell, J. Phys. Chem. B 103 (1999) 8809.
[
20] M. Crocker, R.H.M. Herold, A.G. Orpen, M.T.A. Overgaag, J. Chem. Soc. Dalton
Trans. (1999) 3791.
[21] B.F.G. Johnson, M.C. Klunduk, C.M. Martin, G. Sankar, S.J. Teate, J.M. Thomas, J.
Organomet. Chem. 596 (2000) 221.
[
22] P.P. Pescarmona, J.C. van der Waal, I.E. Maxwell, T. Maschmeyer, Angew. Chem.
Int. Ed. 40 (2001) 740.
[23] K. Wada, N. Itayama, N. Watanabe, M. Bundo, T. Kondo, T. Mitsudo,
Organometallics 23 (2004) 5824.
[
24] H.C.L. Abbenhuis, H.W.G. van Herwijnen, R.A. van Santen, Chem. Commun.
1996) 1941.
(
4
. Conclusion
[25] S. Krijnen, H.C.L. Abbenhuis, R.W.J.M. Hanssen, J.H.C. van Hooff, R.A. van
Santen, Angew. Chem. Int. Ed. 37 (1998) 356.
[
[
26] K. Wada, K. Yamada, D. Izuhara, T. Kondo, T. Mitsudo, Chem. Lett. (2000) 1332.
27] M.D. Skowronska-Ptasinska, M.L.W. Vorstenbosch, R.A. van Santen, H.C.L.
Abbenhuis, Angew. Chem. Int. Ed. 41 (2002) 637.
28] L. Zhang, H.C.L. Abbenhuis, G. Gerritsen, N.N. Bhriain, P.C.M.M. Magusin, B.
Mezari, W. Han, R.A. van Santen, Q. Yang, C. Li, Chem. Eur. J. 13 (2007) 1210.
29] K. Wada, N. Watanabe, T. Kondo, T. Mitsudo, Chem. Eng. Sci. 63 (2008) 4917.
Silica-supported titanium oxide catalysts with excellent activi-
t
ties for the epoxidation of cyclooctene by BuOOH were prepared
by the calcination of Ti-bridged POSS supported or tethered on sil-
ica at 550 °C. Based on the results of catalytic runs and spectro-
scopic analyses of the catalysts, highly isolated tetrahedral
titanium(IV) species were revealed to be produced on the catalyst
surface from 1a, 1b, or 2, while aggregated titanium species were
produced from a corner-capped-type Ti-POSS (3) and tetra-
isopropoxytitanium (4). These catalysts were also applicable for
the epoxidation of limonene. The present study clearly indicates
the utility of Ti-bridged POSS as promising precursors for titanosi-
licate-like active titanium species, which can be implanted on a
variety of oxide supports. The present results suggest that com-
pounds that include a titanium atom surrounded by bulky inor-
ganic moieties through four Ti–O–Si bonds may be suitable
precursors for epoxidation catalysts. Further attempts to improve
the catalytic activities of these catalysts and the application to
the epoxidation of much bulky alkenes, as well as the development
of much more efficient titanium precursors, are underway.
[
[
[30] F. Carniato, C. Bisio, E. Boccaleri, M. Guidotti, E. Gavrilova, L. Marchese, Chem.
Eur. J. 14 (2008) 8098.
[
31] K. Wada, K. Hirabayashi, N. Watanabe, S. Yamamoto, T. Kondo, T. Mitsudo, M.
Inoue, Top. Catal. 52 (2009) 693.
32] K. Wada, M. Nakashita, A. Yamamoto, T. Mitsudo, Chem. Commun. (1998) 133.
[
[33] K. Wada, M. Nakashita, M. Bundo, K. Ito, T. Kondo, T. Mitsudo, Chem. Lett.
1998) 659.
[
[
(
34] K. Wada, K. Yamada, T. Kondo, T. Mitsudo, Chem. Lett. (2001) 12.
35] K. Wada, K. Yamada, T. Kondo, T. Mitsudo, J. Jpn. Petrol. Inst. 45 (2002) 15.
[36] K. Wada, K. Tada, N. Itayama, T. Kondo, T. Mitsudo, J. Catal. 228 (2004) 374.
[
[
[
[
[
[
37] N. Maxim, H.C.L. Abbenhuis, P.J. Stobbelaar, B.L. Mojet, R.A. van Santen, Phys.
Chem. Chem. Phys. 1 (1999) 4473.
38] N. Maxim, H.C.L. Abbenhuis, P.C.M.M. Magusin, R.A. van Santen, Chinese J.
Chem. 19 (2001) 30.
39] N. Maxim, P.C.M.M. Magusin, P.J. Kooyman, J.H.M.C. van Wolput, R.A. van
Santen, H.C.L. Abbenhuis, Chem. Mater. 13 (2001) 2958.
40] N. Maxim, A. Overweg, P.J. Kooyman, A. Nagy, R.A. van Santen, H.C.L.
Abbenhuis, J. Mater. Chem. 12 (2002) 3792.
41] N. Maxim, A. Overweg, P.J. Kooyman, J.H.M.C. van Wolput, R.W.J.M. Hanssen,
R.A. van Santen, H.C.L. Abbenhuis, J. Phys. Chem. B 106 (2002) 2203.
42] R. Murugavel, P. Davis, V.S. Shete, Inorg. Chem. 42 (2003) 4696.
Acknowledgment
[43] K. Wada, K. Yano, T. Kondo, T. Mitsudo, Catal. Today 117 (2006) 242.
[
[
44] K. Wada, R. Tomoyose, T. Kondo, T. Mitsudo, Appl. Catal. A: Gen. 356 (2009) 72.
45] C. Galacho, M.M.L. Ribeiro Carrott, P.J.M. Carrott, Micropor. Mesopor. Mater.
This work was supported in part by a Grant-in Aid for Scientific
Research (No. 18360388) from the Ministry of Education, Sports,
Culture, Science, and Technology, Japan. K.W. acknowledges finan-
cial support from the Okura Foundation.
100 (2007) 312. and references therein.
[
[
[
[
46] H.C.L. Abbenhuis, A.D. Burrows, H. Kooijman, M. Lutz, M.T. Palmer, R.A. van
Santen, A.L. Spek, Chem. Commun. (1998) 2627.
47] Q. Cai, W.-Y. Lin, F.-S. Xiao, W.-Q. Pang, X.-H. Chen, B.-S. Zou, Micropor.
Mesopor. Mater. 32 (1999) 1.
48] F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray
Photoelectron Spectroscopy, Perkin-Elmer Co., Eden Prairie, USA, 1992.
49] S.J. Greg, K.S.W. Sing, Adsorption, Surface Area and Porosity, 2nd ed., Academic
Press, London (UK), 1982.
50] X. Gao, I.E. Wachs, Catal. Today 51 (1999) 233. and references therein.
51] C. Murata, H. Yoshida, T. Hattori, Stud. Surf. Sci. Catal. 143 (2002) 845.
52] S. Bordiga, S. Coluccia, C. Lamberti, L. Marchese, A. Zecchina, F. Boscherini, F.
Buffa, F. Genoni, G. Leofanti, G. Petrini, G. Vlaic, J. Phys. Chem. 98 (1994) 4125.
53] T. Blasco, M.A. Camblor, J.L.G. Fierro, J. Pérez-Pariente, Micropor. Mater. 3
References
[
[
1] K. Weissermel, H.J. Arpe, Industrial Organic Chemistry, third ed., VCH,
Weinheim, 1997.
2] I.W.C.E. Arends, R.A. Sheldon, M. Wallau, U. Schuchardt, Angew. Chem. Int. Ed.
[
[
[
3
6 (1997) 1144. and references therein.
[
[
3] B. Notari, Adv. Catal. 41 (1996) 253.
4] S. Bordiga, A. Damin, F. Bonino, G. Ricchiardi, C. Lamberti, A. Zecchina, Angew.
Chem. Int. Ed. 41 (2002) 4734.
[
(
1994) 259.
[
[
54] C. Kormann, D.W. Bahnemann, M.R. Hoffmann, J. Phys. Chem. 92 (1988) 5196.
55] I. Grohmann, W. Pilz, G. Walther, H. Kosslick, V.A. Tuan, Surf. Interface Anal. 22
[
[
5] S. Bordiga, F. Bonino, A. Damin, C. Lamberti, Phys. Chem. Chem. Phys. 9 (2007)
4
854.
6] For a review, see: R.H. Baney, M. Itoh, A. Sakakibara, T. Suzuki, Chem. Rev. 95
1995) 1409.
(
1994) 403.
[
[
[
56] L. LeNoc, D.T. On, S. Solomykina, B. Echchahed, F. Beland, C.C.D. Moulin, L.
Bonneviot, Stud. Surf. Sci. Catal. 101 (1996) 611.
57] G. Perego, G. Bellussi, C. Corno, M. Taramasso, F. Buonomo, A. Esposito, Stud.
Surf. Sci. Catal. 28 (1986) 129.
58] G. Ricchiardi, A. Damin, S. Bordiga, C. Lamberti, G. Spanò, F. Rivetti, A. Zecchina,
J. Am. Chem. Soc. 123 (2001) 11409.
(
[
[
7] F.J. Feher, T.A. Budzichowski, Polyhedron 14 (1995) 3239.
8] R. Murugavel, A. Voigt, M.G. Walawalker, H.W. Roesky, Chem. Rev. 96 (1996)
2
205.
[
9] P.G. Harrison, J. Organomet. Chem. 542 (1997) 141.
[
10] H.C.L. Abbenhuis, Chem. Eur. J. 6 (2000) 25.