Tripodal titanium silsesquioxane complexes immobilized in polydimethylsiloxane (PDMS) membrane: Selective catalysts for epoxidation of cyclohexene and 1-octene with aqueous hydrogen peroxide

Article abstract of DOI:10.1016/j.jcat.2010.05.003

We investigated the activity, epoxide selectivity, H₂O ₂ efficiency, and recyclability of heterogeneous alkene epoxidation catalysts prepared by encapsulation of tripodal Ti silsesquioxane complexes in polydimethylsiloxane (PDMS) mem

Full text of DOI:10.1016/j.jcat.2010.05.003

Journal of Catalysis 273 (2010) 66–72
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Tripodal titanium silsesquioxane complexes immobilized in polydimethylsiloxane
(PDMS) membrane: Selective catalysts for epoxidation of cyclohexene
and 1-octene with aqueous hydrogen peroxide
Emad H. Aish a, Mark Crocker b,1, Folami T. Ladipo a, ,1
*
a Department of Chemistry, University of Kentucky, Lexington, KY 40506-0055, USA
b University of Kentucky, Center for Applied Energy Research, 2540 Research Park Drive, Lexington, KY 40511-8479, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
We investigated the activity, epoxide selectivity, H2O2 efficiency, and recyclability of heterogeneous
alkene epoxidation catalysts prepared by encapsulation of tripodal Ti silsesquioxane complexes in poly-
dimethylsiloxane (PDMS) membrane. We found that [Ti(NMe2){(c-C6H11)7Si7O12}]/PDMS (3) and [Ti(N-
Me2){(i-C4H9)7Si7O12}]/PDMS (4) displayed high activity, epoxide selectivity (P97%), and H2O2
efficiency (P97%) in cyclohexene- and 1-octene epoxidation with aqueous H2O2. Moreover, these cata-
lysts were highly recyclable. The high H2O2 efficiency can be attributed to the uniformly non-polar envi-
ronment provided about the Ti silsesquioxane complex in 3 and 4 by the PDMS membrane, which
presumably results in low local water concentrations and higher [alkene]:[H2O2] ratios at the Ti center
than in the bulk reaction medium; both of these effects favor the epoxide selectivity and minimal leach-
ing of titanium. Our study of the effect of solvent on catalyst activity revealed that acetonitrile is to be
preferred as solvent, since methanol inhibits the epoxidation at long reaction times.
Received 10 January 2010
Revised 10 May 2010
Accepted 10 May 2010
Available online 8 June 2010
Keywords:
Epoxidation
Ti-silsesquioxane
Encapsulated catalyst
Cyclohexene
1-Octene
Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction
of tripodal Ti vinyl-silsesquioxane complexes by in situ copolymer-
ization on a mesoporous SBA-15-supported polystyrene polymer
Tripodal Ti silsesquioxane complexes have been found to dis-
play excellent activity and selectivity in catalytic epoxidation of
unactivated alkenes with alkyl hydroperoxides [1–3]. Furthermore,
while heterogeneous Ti-based catalysts, such as titania-on-silica
[3,4], titania–silica mixed oxides [5], and titanium silicalite-1
(TS-1) [6,7], invariably contain a range of Ti site types [3,8–10],
the available evidence indicates that the tripodal Ti site is most ac-
tive and selective for alkene epoxidation [1–3]. Consequently, tri-
podal Ti silsesquioxane complexes are attractive precursors for
the preparation of well-defined heterogeneous alkene epoxidation
catalysts. In this context, there have been several reports concern-
ing the immobilization of Ti silsesquioxane complexes. The
impregnation of a tripodal Ti silsesquioxane complex into the
pores of MCM-41, followed by silylation of the outer surface of
the MCM-41 with a bulky silane to prevent leaching of the sils-
esquioxane complex, has been used to prepare a material that
was active in the epoxidation of alkenes with tert-butyl hydroper-
oxide (TBHP) [11]. Heterogeneous Ti silsesquioxane catalysts that
showed activity toward epoxidation of alkenes with aqueous
hydrogen peroxide (H2O2) have been prepared via immobilization
[12], or by grafting onto a siloxane copolymer (via hydrosilylation)
followed by crosslinking of the resulting polymer with a vinyl-ter-
minated siloxane polymer [13]. However, control of the three-
dimensional structure of the polymer, necessary to ensure a hydro-
phobic environment around the Ti complex, appears difficult by
this method. On the other hand, heterogeneous catalysts contain-
ing bipodal Ti sites (such as („SiO)2TiCp2 and („SiO)2Ti(OBu2)
[14,15] and gel catalysts containing tetrapodal Ti sites (i.e.
(„SiO)4Ti) have been reported to display activity toward the epox-
idation of alkenes with H2O2 [16]. However, bipodal and tetrapodal
Ti sites are inherently less efficient in alkene epoxidation than tri-
podal Ti site, as previously mentioned [2]. Hence, improved meth-
ods for the immobilization of tripodal Ti silsesquioxane complexes
remain a critical need.
Herein we describe a simple and effective method for the
preparation of heterogeneous alkene epoxidation catalysts by
encapsulation of tripodal Ti silsesquioxane complexes in poly-
dimethylsiloxane (PDMS) membrane. This type of immobilization
relies on the physical encapsulation of the catalyst. PDMS is an
excellent choice in this regard because it is hydrophobic, chemi-
cally rather inert (and is particularly resistant to oxidation), and
thermally stable to 250 °C. Moreover, it has been successfully ap-
plied to the immobilization of several types of oxidation catalysts,
including Jacobsen’s catalyst, metallo-porphyrins, and zeolites
* Corresponding author. Fax: +1 859 323 1069.
E-mail address: fladip0@uky.edu (F.T. Ladipo).
1
These authors contributed equally to this work.
0021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.05.003

E.H. Aish et al. / Journal of Catalysis 273 (2010) 66–72
67
[17–19]. In this study, we describe the preparation of immobi-
lized Ti silsesquioxane catalysts [Ti(NMe2){(c-C6H11)7Si7O12}]/
PDMS (3) and [Ti(NMe2){(i-C4H9)7Si7O12}]/PDMS (4), which dis-
play high activity toward the epoxidation of cyclohexene and 1-
octene, as well as excellent H2O2 efficiency and epoxide
selectivity.
CH2CH(CH3)2), 0.94 (br d, 42H, CH2CH(CH3)2), 0.56 (br d, 14H,
CH2CH(CH3)2). 13C NMR (CDCl3): d 45.14 (s, NMe2), 26.00, 25.92
(overlapping s, CH2CH(CH3)2), 24.27, 24.16 (overlapping s,
CH2CH(CH3)2), 22.92, 22.81, 22.76 (s, 3:1:3 for CH2CH(CH3)2). 29Si
NMR (CDCl3): d ꢁ65.23, ꢁ67.89, ꢁ68.96, (3:1:3). Anal. Calcd. for
C
30H69NO12Si7Ti: C, 40.93; H, 7.90. Found: C, 40.81; H, 8.01.
2.2. Incorporation of tripodal Ti silsesquioxane complexes into PDMS
membranes
2. Experimental
2.1. Methods and materials
Sylgard184 is sold as two components, prepolymer (RTV 615A)
and crosslinker (RTV 615B), that are mixed in a 10:1 ratio and
heated to fully cure. The PDMS membrane was prepared by mod-
ification of a literature method [22]. The PDMS slabs were dried
overnight at 65 °C. Each slab was cut into irregular pieces with
dimensions ranging from 1 mm to 1 cm; these were later cut into
smaller pieces for ease of handling.
All experiments were performed under dry nitrogen atmosphere
using standard Schlenk techniques or in a Vacuum Atmospheres,
Inc. glovebox. Solvents were dried and distilled by standard meth-
ods before use [20]. Following vacuum distillation from the appro-
priate drying agent, CDCl3 was degassed by repeated freeze–pump–
thaw cycles. All solvents were stored in a glovebox over 4 Å molec-
ular sieves that had been dried in a vacuum oven at 150 °C for at
least 48 h prior to use. Unless otherwise stated, all reagents were
purchased from Aldrich Chemical Company. 1-Octene (>98%) and
cyclohexene were dried over activated 4 Å molecular sieves prior
to use. Uncross-linked PDMS (Sylgard184) was purchased from
Dow Corning Corporation and H2O2 solution (30%) was purchased
from Mallinckrodt Baker Inc. Incompletely condensed silsesquiox-
2.2.1. [Ti(NMe2){(c-C6H11)7Si7O12}]/PDMS (3)
A CH2Cl2 solution (0.8 mL) of [Ti(NMe2){(c-C6H11)7Si7O12}] (1,
0.0530 g, 0.0496 mmol) was added to a Schlenk flask containing
PDMS slabs (1.00 g) and allowed to stand under N2 atmosphere
with periodic shaking. After complete adsorption of CH2Cl2 into
the PDMS slabs, they were placed under vacuum to remove the
CH2Cl2; less solvent than the maximum that would swell into
PDMS was used so that all of the CH2Cl2 containing 1 would adsorb
into the slabs. The slabs were briefly rinsed with CH2Cl2 (3 mL) un-
der N2 atmosphere to remove any catalyst on the surface and dried
again under vacuum. The slabs of 3 (5.03% by weight 1) were
stored in a glove box under N2 at ambient temperature. UV–vis:
kmax = 230 nm. Anal. Calcd. for 3: Ti, 0.23. Found: Ti, 0.27.
anes (trisilanolalkyl-POSS) R7Si7O9(OH)3 (R = i-C4H9 and c-C6H11
)
were purchased from Hybrid Plastics Inc. and dried overnight under
vacuum at 50 °C prior to use. The compounds Ti(NMe2)4 [21] and
[Ti(NMe2){(c-C6H11)7Si7O12}] (1) [2] were prepared according to
the literature methods.
1H, 13C, and 29Si NMR spectra were recorded on a Varian Gem-
ini-200 spectrometer or a Varian VXR-400 spectrometer at room
temperature unless otherwise stated. All chemical shifts are re-
ported in units of d (downfield from tetramethylsilane) and 1H
and 13C chemical shifts were referenced to residual solvent peaks.
29Si NMR spectra were recorded with inverse-gated proton decou-
pling in order to increase resolution and minimize nuclear Overha-
user enhancement effects. To ensure accurate integrated
intensities, [Cr(acac)3] (0.05 M) was added to 13C and 29Si NMR
samples as a shiftless relaxation agent and a delay of at least 5 s
was used between observation pulses for 13C measurements and
10 s for 29Si measurements. GC analyses were performed on a Shi-
madzu GC-17A instrument with flame ionization detection (FID), a
60 m ꢀ 0.32 mm (0.25 mm film thickness) Agilent JW Scientific
DB-5 GC column, and helium as carrier gas. An injection tempera-
ture of 140 °C was employed, which was found to be sufficiently
low to avoid the occurrence of secondary reactions in the injection
port. UV–vis spectra were collected in the absorption mode on an
Agilent 8453 spectrophotometer using a direct insertion method.
Infrared spectra were obtained on a Mattson Galaxy Series FTIR
5000 spectrophotometer. Proton-induced X-ray emission (PIXE)
analyses for Ti were performed by Elemental Analysis Inc., Lexing-
ton, KY. The analytical error is estimated at 1% (relative). Elemen-
tal analysis for C, H, and N was performed by Complete Analysis
Laboratories Inc., Parsippany, NJ.
2.2.2. [Ti(NMe2){(i-C4H9)7Si7O12}]/PDMS (4)
Compound 4 was prepared from [Ti(NMe2){(i-C4H9)7Si7O12}] (2,
0.0530 g, 0.0598 mmol) using the method described for 3. The
slabs of 4 were stored in a glove box under N2 at ambient temper-
ature. UV–vis: kmax = 235 nm.
A sample of 4 containing 7.62% by weight of 2 was analyzed by
PIXE (proton induced X-ray emission). Anal. Calcd. for 4: Ti, 0.41.
Found: Ti, 0.38.
2.3. Procedure for catalytic alkene epoxidation
Epoxidation tests were performed in a magnetically stirred 50-
mL three-necked flask, equipped with a condenser, thermometer
probe and septum for withdrawing samples. In a typical reaction,
solvent (CH3CN or CH3OH, 5 mL), toluene (30 mg, 0.3 mmol as
internal standard), alkene (1-octene or cyclohexene, 0.5 mmol), a
quantity of Ti/PDMS catalyst equivalent to 0.01 mmol of Ti
(2 mol%), and a stirrer bar were placed in the flask. The mixture
was heated to 60 °C and an equimolar or greater amount (relative
to alkene) of aqueous H2O2 was added via syringe. A sample was
immediately taken for analysis (GC and titration), and additional
samples for analysis were taken at regular intervals. The H2O2 con-
centration was determined with aqueous Ce(SO4)2 (0.1 M) and ferr-
oin indicator using the titration method. The selectivity to epoxide
and H2O2 efficiency were determined as follows: epoxide selectiv-
ity = (mol. of epoxide formed/mol. of alkene consumed) ꢀ 100;
and H2O2 efficiency = {[epoxide]/([H2O2]0 ꢁ [H2O2]t)} ꢀ 100.
2.1.1. Synthesis of [Ti(NMe2){(i-C4H9)7Si7O12}] (2)
Ti(NMe2)4 (0.200 g, 0.892 mmol) was added via syringe to a stir-
red solution of (i-C4H9)7Si7O9(OH)3 (0.68 g, 0.859 mmol) in diethyl
ether (10 mL). Stirring the reaction mixture at room temperature
for 18 h afforded a deep yellow mixture, which was filtered. The fil-
trate was reduced to dryness under vacuum and the residue was
dissolved in toluene. Acetonitrile was added dropwise to precipitate
2 as a yellow microcrystalline solid, which was isolated by filtration,
washed with acetonitrile (3 ꢀ 5 mL) and dried under vacuum. Yield:
0.732 g, 97%. 1H NMR (CDCl3): d 3.19 (s, 6H, NMe2), 1.84 (m, 7H,
2.4. Ti silsesquioxane complex leaching studies
The experimental procedure is the same as described in Section
2.3 for catalytic alkene epoxidation with H2O2 at 60 °C. After the
reaction had proceeded for a specified amount of time, a sample
of the reaction mixture was taken for GC analysis. An aliquot

68
E.H. Aish et al. / Journal of Catalysis 273 (2010) 66–72
(about half) of the reaction mixture was then transferred by canula
into a three-necked Schlenk flask equipped with a thermometer
and condenser. Both the original reaction mixture (containing
immobilized catalyst) and the aliquot (absent immobilized cata-
lyst) were heated as before for identical amounts of time. Samples
were taken at regular intervals from each reaction mixture for GC
analysis. Only the reaction mixture containing immobilized cata-
lyst continued to form epoxide product. At completion, the catalyst
was isolated by filtration, washed with acetonitrile (5 mL), and
dried under vacuum for 2 h. The recovered catalyst was then ana-
lyzed for its Ti content by PIXE, as well as characterized by UV–vis
spectroscopy (see Section 3).
(4) were readily prepared by adding a CH2Cl2 solution of the appro-
priate tripodal Ti silsesquioxane complex (1 and 2, respectively) to
the PDMS membrane under N2 atmosphere. The PDMS membranes
were swelled by CH2Cl2, which facilitated diffusion of the Ti sils-
esquioxane complex into the membranes. Removal of CH2Cl2 un-
der vacuum, rinsing of the membranes with CH2Cl2, and drying
again under vacuum afforded the catalysts. The compositions of
3 and 4 were confirmed by elemental analysis using proton-in-
duced X-ray emission (PIXE); in each case, the weight percent of
titanium found was in good agreement with the calculated value.
UV–vis and IR spectroscopic methods were used to characterize
the structure of immobilized catalysts 3 and 4. The UV–vis data
for 3 and 4 are consistent with the retention of a tetrahedral Ti site
in the materials. Specifically, the UV–vis spectra of 3 and 4 con-
tained an intense absorption at 230 and 235 nm, respectively
(see Supplementary material); these absorptions are close to the
range (212–228 nm) previously reported for Ti silsesquioxane
complexes and assigned to a ligand to metal charge transfer tran-
sition involving four-coordinated titanium bearing oxygen ligands
[2]. The absence of bands above 250 nm indicates that octahedral
Ti sites and TiO2 were negligible in the materials [25]. The IR spec-
tra of 3 and 4 contained a partially obscured band of medium
intensity falling in the range 940–960 cmꢁ1; a similar band in the
range 930–970 cmꢁ1 in the IR spectra of tripodal Ti silsesquioxane
complexes [2], titanium silicates such as TS-1 and Ti-MCM-41
[26,27], and amorphous titania–silica mixed oxides [28] was con-
vincingly assigned to a Si–O–Ti stretching vibration [2].
2.5. Catalyst recyclability studies
The experimental procedure is the same as described in Section
2.3 for catalytic alkene epoxidation with H2O2 at 60 °C. After the al-
kene epoxidation had proceeded for a specified amount of time, the
catalyst was removed by filtration, washed with acetonitrile
(5 mL), and dried under vacuum for 2 h. The catalyst was then re-
used for alkene epoxidation as described in Section 2.3.
3. Results and discussion
3.1. Preparation and characterization of tripodal titanium complexes
One of us (MC) has previously reported the synthesis of a series
of tripodal Ti silsesquioxane complexes, including [Ti(NMe2){(c-
C6H11)7Si7O12}] (1), as well as data for the epoxidation of 1-octene
with tert-butyl hydroperoxide (TBHP) using tripodal Ti silsesquiox-
ane complexes as catalysts [2]. In this study, we have similarly pre-
pared [Ti(NMe2){(i-C4H9)7Si7O12}] (2) in excellent yield via
protonolysis of Ti(NMe2)4 with one equivalent of (i-
C4H9)7Si7O9(OH)3 (trisilanolisobutyl-POSS) in diethyl ether (Eq.
(1)). The formulation and structure of 2 were established by micro-
analysis and solution (1H, 13C, and 29Si) NMR data.
3.3. Alkene epoxidation activity
The catalytic activity of [Ti(NMe2){(c-C6H11)7Si7O12}]/PDMS (3)
and [Ti(NMe2){(i-C4H9)7Si7O12}]/PDMS (4) toward the epoxidation
of cyclohexene and 1-octene with H2O2 was investigated as a test
of their epoxidation efficiency. At 60 °C in MeCN, reactions of
cyclohexene or 1-octene with 1 equivalent of aqueous H2O2 in
the presence of 2 mol% of catalyst 3 or 4 proceeded to high conver-
sion with excellent H2O2 efficiency to selectively produce cyclo-
hexene oxide or 1,2-epoxyoctane, respectively (Table 1, entries
1–4; Eqs. (2) and (3)). The highly efficient use of H2O2 by 3 and 4
is remarkable given that Ti silsesquioxane complexes are not active
for epoxidation with H2O2 under homogeneous conditions, due
most likely to excessive coordination of water to titanium. How-
ever, the PDMS membrane provides
ð1Þ
Consistent with the C3v symmetry expected for tripodal
Ti(NMe2)silsesquioxane species, the 29Si NMR spectrum of
2
showed three resonances at d = ꢁ65.23, ꢁ67.89, and ꢁ68.96 ppm
for the silsesquioxane Si atoms in a 3:1:3 ratio. Moreover, consis-
tent with previous observations for Ti silsesquioxane complexes
[23,24], the resonance for the Si atoms bearing OH groups (at
d = ꢁ58 ppm for trisilanolisobutyl-POSS) shifts upfield (by ca.
7 ppm) upon co-ordination of the oxygen with titanium. Analogous
to [Ti(NMe2){(c-C6H11)7Si7O12}] (1), compound 2 is an air- and
moisture-sensitive yellow solid. It displays good solubility in polar
hydrocarbon solvents, such as THF, diethyl ether, chloroform, and
dichloromethane, as well as in aromatic hydrocarbon solvents, such
as benzene and toluene, but shows poor solubility in acetonitrile.
ð2Þ
ð3Þ
a uniformly non-polar environment about the Ti silsesquioxane
complex in 3 and 4, which presumably results in low local water
concentrations, as well as higher [alkene]:[H2O2] ratios at the Ti
center than in the bulk reaction medium (such an effect has been
verified for TS-1, the non-polar character of the zeolite resulting
in a higher concentration of the alkene reactant in the zeolite rela-
tive to the bulk medium) [29,30]. A high alkene:H2O2 ratio is favor-
able for the epoxide selectivity, since the Ti-hydroperoxo
intermediate (Ti-OOH) formed from reaction of the Ti complex with
3.2. Incorporation of tripodal Ti silsesquioxane complexes into PDMS
membranes
Immobilized Ti silsesquioxane catalysts [Ti(NMe2){(c-
C6H11)7Si7O12}]/PDMS (3) and [Ti(NMe2){(i-C4H9)7Si7O12}]/PDMS

E.H. Aish et al. / Journal of Catalysis 273 (2010) 66–72
69
Table 1
Epoxidation of 1-octene (1-C8H14) and cyclohexene (c-C6H10) with H2O2 catalyzed by titanium silsesquioxane complexes immobilized in PDMS.
Entry
Catalyst
Substrate
Solvent
Conv. (%)
Epo. Sel.j (%)
Effi.k (%)
TONl
ref.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
3
4
3
4
3
4
4
4
1-C8H14
1-C8H14
c-C6H10
c-C6H10
1-C8H14
1-C8H14
1-C8H14
1-C8H14
c-C6H10
c-C6H10
1-C6H12
1-C6H12
1-C6H12
1-C6H12
c-C6H10
c-C6H10
c-C6H10
c-C6H10
1-C8H14
1-C8H14
MeCN
MeCN
MeCN
MeCN
MeOH
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeCN
95.1a
89.0a
88.0a
89.0a
68.0a
81.4a
47.0b
35.9b
42.0b
32.0b
14.5c
21.2c
17.7c
8.7c
99.2
99.0
99.1
99.0
98.4
97.3
99.0
99.0
99.0
99.0
100
96.6
100
36.3
–
98.1
99.0
98.3
98.2
97.1
97.0
99.0
98.0
99.0
99.0
63.8
91
470
359
420
320
4
4
TS-1d
TS-1d
Ti-betae
Ti-betae
TS-1f
TS-1f
Ti-betae
Ti-betae
4
94.5
138.2
66.5
32.7
36.1
6.6
25
25
25
25
25
25
25
25
g
g
g
43.6
38.1
–
g
4.9c
0.9c
–
80
0.6
99.0
99.0
–
17.7c
40.4c
44.9h
26.0i
52.2
80.1
99.0
99.0
66.5
151.9
4
a
b
c
Reaction conditions: 0.01 mmol of Ti, 0.5 mmol of alkene, 0.5 mmol of H2O2 (30% aqueous solution), 5 mL solvent, 60 °C, and 40 h.
Reaction conditions: 0.01 mmol of Ti, 10 mmol of alkene, 10 mmol of H2O2 (30% aqueous solution), 10 mL solvent, 60 °C, and 2 h.
Reaction conditions: 60 °C, 2 h (4 h for cyclohexene oxidation with TS-1), 10 mmol of alkene, 10 mmol of H2O2 (31% aqueous solution), 0.05 g of catalyst, and 10 mL
solvent.
d
Si/Ti = 53.
Si/Ti = 30.
Si/Ti = 60.
1-C6H12 = 1-hexene.
e
f
g
h
i
Reaction conditions are same as described for superscript a (above) except 5 mmol of H2O2 (30% aqueous solution) was used.
Reaction conditions are same as described for superscript a (above) except 10 mmol of H2O2 (30% aqueous solution) was used.
Epoxide selectivity = (mol of epoxide formed/mol of alkene consumed) ꢀ 100.
H2O2 efficiency = {[epoxide]/([H2O2]0 ꢁ [H2O2]t)} ꢀ 100.
j
k
l
Turnover number: the mole of substrate converted per mole of Ti present in the catalysts.
the H2O2 can undergo two main processes: oxygen transfer to the
alkene, resulting in formation of the epoxide or decomposition via
homolytic fission of the Ti–OOH bond (in which case no oxygen
transfer occurs) [2,31]. High alkene concentrations increase the
probability that the Ti–OOH intermediate is intercepted by the al-
kene before decomposition can occur. Moreover, limiting the con-
centration of (acidic) H2O2 and H2O at the Ti sites helps to reduce
the occurrence of Ti leaching (vide infra). Thus, the hydrophobic
environment provided by the PDMS membrane facilitates the effi-
cient utilization of H2O2 for alkene epoxidation by Ti silsesquioxane
catalysts.
The use of acetonitrile as solvent for alkene epoxidations with
immobilized catalysts 3 and 4 is particularly suitable, since it is
miscible with olefins and aqueous H2O2 in all proportions. More-
over, it does not swell PDMS to any appreciable extent (the swelling
ratio, S = D/D0, measured experimentally for MeCN at 25 °C is equal
to 1.01, where D is the length of PDMS in the solvent and D0 is the
length of the dry PDMS) [32], and Ti silsesquioxane complexes are
completely insoluble in it (thereby promoting retention of the Ti
silsesquioxane complexes in the membrane). However, given that
it has been shown that the nature of the solvent strongly influences
the performance of titanosilicate catalysts in alkene epoxidation
[7,25,33], we evaluated the effect of a protic solvent on the epoxida-
tion efficiency of 3 and 4. The protic solvent chosen was methanol
because it has comparable polarity (dielectric constant = 32.6) to
acetonitrile (dielectric constant = 37.5) and it has been found not
to appreciably swell PDMS (S = 1.02 at 25 °C) [32]. The epoxidation
of 1-octene with 1 equivalent of aqueous H2O2 was conducted in
methanol at 60 °C in the presence of 2 mol% of immobilized catalyst
3 or 4 for 40 h. As shown in Table 1 (entries 5 and 6), both the H2O2
efficiency and the epoxide selectivity of 3 and 4 were excellent in
methanol (and similar to those achieved in acetonitrile under sim-
ilar conditions, Table 1, entries 1 and 2).
For both catalysts, [Ti(NMe2){(c-C6H11)7Si7O12}]/PDMS (3) and
[Ti(NMe2){(i-C4H9)7Si7O12}]/PDMS (4), the epoxidation of 1-octene
with H2O2 to form 1,2-epoxyoctane proceeded to greater comple-
tion when acetonitrile was the solvent as opposed to methanol
(Fig. 1, see also Table 1, entries 7–10). In this regard, we note that
3 displayed comparable catalytic activity in acetonitrile and meth-
anol during the early stages of the epoxidation reaction (up to
ꢂ25% completion, Fig. 1). Similarly, the epoxidation activity of 4
was comparable in the two solvents, up to ꢂ44% completion of
the reaction (Fig. 1). In contrast, during the later stages of the reac-
tion, the epoxidation activity of both catalysts decreased at a great-
er rate in methanol than in acetonitrile (Fig. 1). Catalysts 3 and 4
likely form different active species in the two solvents: the active
intermediate in MeOH presumably possesses a stable five-mem-
bered ring structure formed by coordination of MeOH to Ti and
Fig. 1. Time profile of the epoxidation of 1-octene with H2O2 catalyzed by catalysts
3 and 4 (reaction conditions: 60 °C, 0.5 mmol of 1-octene, 0.5 mmol of H2O2 (30%
aqueous solution), 0.01 mmol of Ti, and 5 mL solvent).

70
E.H. Aish et al. / Journal of Catalysis 273 (2010) 66–72
Scheme 1. Structures proposed intermediate Ti species.
hydrogen bonding to the Ti–OOH complex (I, Scheme 1). On the
other hand, the active intermediate in acetonitrile is derived via
coordination of water to Ti (II, Scheme 1). In this regard, the puta-
tive active species in TS-1-catalyzed alkene epoxidation in metha-
nol is analogous to I [7,33], while a species equivalent to II is
proposed to contribute in Ti-beta-catalyzed alkene epoxidation in
acetonitrile [7,33]; and the fact that alkene epoxidation over TS-1
is more favorable in MeOH than in MeCN, while alkene epoxidation
over Ti-beta is more favorable in MeCN than in MeOH (e.g. Table 1,
entries 11–12 and 13–14) is widely accepted as resulting from the
presence of these different active species [7,25,33]. Due to the low-
er donor properties of water, an active species similar to II (in
which a water molecule is coordinated to Ti, Scheme 1) would pos-
sess greater electrophilic character and hence exhibit higher intrin-
sic epoxidation activity than an active species similar to I (in which
an alcohol molecule is coordinated to Ti, Scheme 1), and could be
stabilized in the hydrophilic pores of Ti-beta while the hydropho-
bic pores of TS-1 would be much less efficient for its formation.
However, since our catalysts showed comparable 1-octene epoxi-
dation activity in acetonitrile and methanol during early stages
of the reaction (Fig. 1), we believe that in addition to the possible
differences in 1-octene epoxidation activities of species I and II,
inhibition of the reaction due to excessive coordination of MeOH
to Ti contributes to reduced activity of our catalysts in methanol
solvent. Consistent with this suggestion, it has been reported for
a variety of homogeneous and heterogeneous titanium and vana-
dium catalysts that alkene epoxidation with alkyl hydroperoxide
(ROOH) is inhibited by the alcohol (ROH) co-product [31]. The rea-
son why the inhibition appears later in the reaction may be due to
slight swelling of the PDMS at long reaction times (albeit not
enough to cause significant leaching, see Section 3.3.1), thereby
increasing the concentration of MeOH in the hydrophobic PDMS
membrane. In both acetonitrile and methanol, [Ti(NMe2)-
{(i-C4H9)7Si7O12}]/PDMS (4) displayed greater catalytic activity
than [Ti(NMe2){(c-C6H11)7Si7O12}]/PDMS (3) early in the epoxida-
tion of 1-octene with H2O2 (Fig. 1). Presumably, 4 is more active
than 3 because active species derived from 4 possess a more open
coordination environment about the Ti center (and hence are more
reactive) than analogous species derived from 3, given the reduced
steric bulk of isobutyl group versus cyclohexyl group. Consistent
with this explanation, it has previously been observed that cyclo-
pentyl-substituted tripodal Ti silsesquioxane complexes are more
active homogeneous epoxidation catalysts than their bulkier cyclo-
hexyl-substituted analogs [2].
10 mmol of H2O2, 10 mL solvent, 60 °C, and 2 h) [25]. As shown
in Table 1, 4 achieved higher turnover numbers for both cyclohex-
ene- and 1-octene epoxidation with H2O2 (entries 7–10) than
either TS-1 or Ti-beta (entries 11–18) under similar reaction con-
ditions; conversion of cyclohexene over TS-1 was especially low
due to the inability of cyclohexene to enter the TS-1 pore system
(entries 15–16). The high epoxide selectivity (P97%) and out-
standing H2O2 utilization efficiency (P97%) achieved for the epox-
idation of 1-octene and cyclohexene with H2O2 in MeCN and
MeOH using 3 or 4 as catalyst at 60 °C (Table 1, entries 1–10)
are in contrast with typical results of alkene epoxidation with
H2O2 using heterogeneous Ti-based catalysts. For example, 1-hex-
ene epoxidation with H2O2 using TS-1 or Ti-beta as catalyst was
found to occur with high epoxide selectivity but low H2O2 effi-
ciency at 60 °C in MeCN (and in MeOH for Ti-beta) (Table 1, entries
11–14) [25]. And Ti-beta-catalyzed reaction of cyclohexene with
H2O2 was found to occur with moderate epoxide selectivity and
low H2O2 efficiency in MeCN, and with very poor epoxide selectiv-
ity in MeOH (Table 1, entries 17–18) [25].
As discussed earlier, we believe that the uniformly non-polar
environment provided by the PDMS membrane limits the concen-
tration of H2O2 and H2O at the Ti sites in 3 and 4, resulting in
higher [alkene]:[H2O2] ratios at the Ti center than in the bulk
reaction medium. In turn, this effect maintains the prevalence
of tripodal Ti sites (see Section 3.2) in 3 and 4 throughout the
course of the epoxidation reactions, which makes possible the
excellent epoxidation efficiency of the catalysts. In this regard,
based on steric and electronic arguments, tripodal complexes rep-
resent the best comprise between high Lewis acidity on the one
hand, and steric accessibility of the Ti center on the other, and
the available data strongly support the tripodal site being the
most active and selective for epoxidation [1–3]. Consequently, tri-
podal Ti silsequioxane complexes are inherently more efficient
than heterogeneous Ti-based epoxidation catalysts, since the lat-
ter inevitably contain a variety of Ti site types. For example, bipo-
dal ((–SiO)2Ti(OH)2), tripodal ((–SiO)3Ti(OH)), and tetrapodal
((–SiO)4Ti) sites are believed to exist in heterogeneous catalysts,
such as titania-on-silica and TS-1 [2], while the presence of olig-
omeric Ti species has also been inferred [8,9], in addition to crys-
talline TiO2 (anatase and rutile) [10]; all of these various Ti
species show varying degree of activity in catalyzing alkene epox-
idation and/or peroxide decomposition. Presumably, heteroge-
neous Ti silsesquioxane catalysts containing bipodal Ti sites
[14,15] and gel catalysts containing tetrapodal Ti sites [16] dis-
play lower epoxide selectivity and alkene conversion profiles than
In order to assess the epoxidation efficiency of tripodal Ti sils-
esquioxane complexes encapsulated in PDMS membrane in rela-
tion to epoxidation efficiencies of much studied titanosilicate
catalysts, such as TS-1 and Ti-beta, we carried out cyclohexene-
and 1-octene epoxidation with H2O2 using 4 (0.01 mmol Ti) as
catalyst in MeCN or MeOH under conditions analogous to those
previously reported for 1-hexene- and cyclohexene epoxidation
with H2O2 using TS-1 or Ti-beta as catalyst (10 mmol of alkene,
[Ti(NMe2){(c-C6H11)7Si7O12}]/PDMS
(3)
and
[Ti(NMe2){(i-
C4H9)7Si7O12}]/PDMS (4) due to the lower activity and selectivity
of these sites. On the other hand, heterogeneous Ti silsesquioxane
catalysts prepared by grafting of tripodal Ti vinyl-silsesquioxane
complexes onto a polymeric matrix likely possess a less uniform
hydrophobic environment around the tripodal Ti site, resulting in
faster modification and/or deactivation of the catalyst, and hence

E.H. Aish et al. / Journal of Catalysis 273 (2010) 66–72
71
Table 2
Catalyst recyclability studies.a
Solvent
Cat.
Substrate
Conversion (%)
Epoxidation cycled
1
2
3
4
5
MeCN
MeCN
MeOH
3
4
3
1-Octeneb
15.1
43.0
63.4
14.4
42.9
62.9
14.1
42.5
14.1
42.0
14.0
42.0
1-Octeneb
Cyclohexenec
a
Reaction conditions: 60 °C, 0.5 mmol of alkene, 0.5 mmol of H2O2 (30% aqueous
solution), 0.01 mmol of Ti, and 5 mL solvent.
b
Reaction time = 2 h.
Reaction time = 20 h.
Epoxide selectivity P99% and H2O2 efficiency P98%.
c
d
in PDMS. Further support for the latter conclusion was obtained by
isolating the catalyst (3 or 4) via filtration after completion of the
epoxidation reaction, washing the recovered catalyst with acetoni-
trile, drying it under vacuum for 2 h, and then analyzing the recov-
ered catalyst for its Ti content by PIXE. The Ti content found for
used 3 was 0.32% in comparison with 0.27% found for pristine 3,
while the Ti content found for used 4 was 0.32% in comparison
with 0.38% found for pristine 4. These results are consistent with
little loss of Ti from the catalysts under our epoxidation conditions;
we presume the slight increase in Ti content found for used 3 is
reflective of small variation in the Ti content from slab to slab of
3 and 4. We also characterized the recovered catalysts by UV–vis
spectroscopy. Consistent with prevalence of tetrahedral Ti sites
in the materials (and analogous to the pristine catalysts), the
UV–vis spectra of used 3 and 4 contained an intense absorption
at 230 and 235 nm, respectively (see Supplementary material). In
addition, neither spectrum contained a significant band above
250 nm, indicating that octahedral Ti sites and TiO2 were negligible
in the materials [25].
Consistent with retention of the tripodal Ti silsesquioxane com-
plex in the PDMS membrane, the catalysts are recyclable (Table 2).
When the epoxidation of 1-octene with H2O2 (1 equiv.) at 60 °C in
MeCN using 3 as catalyst had proceeded for 2 h (to approximately
15% completion), the catalyst was removed by filtration, washed
with acetonitrile, and dried under vacuum for 2 h. The catalyst
was then reused for 1-octene epoxidation as before. As shown in
Table 2 (entry 1), the activity (% alkene conversion), epoxide selec-
tivity, and H2O2 efficiency of the catalyst were highly reproducible
for the 5 cycles studied. Analogous recycling of 4 in catalytic epox-
idation of 1-octene with H2O2 in MeCN furnished similar results:
the activity, H2O2 efficiency, and epoxide selectivity of catalyst
were highly reproducible for 5 cycles (Table 2, entry 2). Using 3
as catalyst for the epoxidation of cyclohexene with H2O2 in MeOH
revealed that the catalyst can also be recycled in methanol; the
conversion of cyclohexene to cyclohexene oxide and the H2O2 effi-
ciency were comparable after two cycles (Table 2, entry 3). The
aforementioned results clearly support minimal leaching of tita-
nium from the catalyst, due presumably to limited concentration
of (acidic) H2O2 and water (as well as methanol) at tripodal Ti sites
inside the PDMS membrane.
Fig. 2. Time profile of the epoxidation of 1-octene with H2O2 catalyzed by catalysts
3 at different temperatures (Reaction conditions: 0.5 mmol of 1-octene, 0.5 mmol of
H2O2 (30% aqueous solution), 0.01 mmol of Ti, and 5 mL solvent).
reduced epoxide selectivity and alkene conversion profiles [12] in
comparison with 3 and 4.
Detailed studies of the kinetics and mechanism of alkene epox-
idation catalyzed by 3 or 4 are currently underway in our labora-
tory. However, our preliminary data indicate that the epoxidation
reactions are not diffusion-controlled, as indicated by significant
differences in the initial rate of alkene conversion in the tempera-
ture range 313–333 K (Fig. 2); a plot of ln k against 1/T afforded a
straight line and a resulting apparent activation energy of 69 kJ/
mol (not far from the 41 to 60 kJ/mol range reported for epoxida-
tions with TBHP catalyzed by homogeneous tripodal Ti silsesquiox-
ane catalysts). At 60 °C in MeCN, we found that reaction of
1-octene with 10 or 20 molar equivalents H2O2 (30% aqueous solu-
tion) in the presence of 2 mol% of catalyst 4 proceeded with consid-
erably less conversion of 1-octene, albeit the selectivity for
1,2-epoxyoctane remains high (Table 1, entries 19–20). Evidently,
low [alkene]:[water] ratio leads to significant catalyst deactivation,
due presumably to excessive coordination of water to titanium,
making it difficult to determine the dependence of alkene conver-
sion rate on [H2O2].
3.3.1. Studies of Ti silsesquioxane complex leaching and catalyst
recyclability
To establish that the Ti silsesquioxane catalyst remained encap-
sulated in the PDMS membrane and that the epoxidation reactions
occurred in PDMS rather than in the solvents, we examined the
catalytic properties of the reaction mixture and the catalyst recy-
clability. An aliquot (about half) of the reaction mixture was re-
moved when the epoxidation of cyclohexene with H2O2 in MeCN
at 60 °C and using 3 as catalyst was at approximately 63% comple-
tion (after 20 h). The aliquot was placed in a Schlenk flask under N2
and heated as before for 20 h. The reaction did not proceed any fur-
ther in this control experiment, while in the Schlenk flask contain-
ing 3, conversion of cyclohexene into cyclohexene oxide grew to
88%, proving that leaching of the Ti silsesquioxane complex from
PDMS membrane is minimal, if any. Using 4 as catalyst under sim-
ilar reaction conditions as earlier, the epoxidation of cyclohexene
with H2O2 in MeOH was at approximately 31% completion after
2 h. Again, removing an aliquot of the reaction mixture and heating
it at 60 °C for 15 h did not result in further reaction. These control
experiments were repeated multiple times with the same result
each time: the solvent did not contain any appreciable amount of
Ti silsesquioxane complex and the epoxidation reactions occurred
4. Conclusions
The incorporation of Ti silsesquioxane complexes in a PDMS
membrane furnished active heterogeneous catalysts, [Ti(N-
Me2){(c-C6H11)7Si7O12}]/PDMS (3) and [Ti(NMe2){(i-C4H9)7Si7O12}]/
PDMS (4), for the epoxidation of cyclohexene and 1-octene with
aqueous H2O2. These catalysts displayed excellent H2O2 efficiency
(P97%) and epoxide selectivity (P97%), and are highly recyclable.
The high H2O2 efficiency can be attributed to the uniformly non-po-
lar environment provided about the Ti silsesquioxane complex in 3

72
E.H. Aish et al. / Journal of Catalysis 273 (2010) 66–72
[6] M. Taramasso, G. Perego, B. Notari, US Patent 4,410,501, 1983, to Enichem.
[7] M.G. Clerici, P. Ingallina, J. Catal. 140 (1993) 71–83.
[8] A. Tuel, L.G. Hubert-Pfalzgraf, J. Catal. 217 (2003) 343–353.
[9] Q. Yuan, A. Hagen, F. Roessner, Appl. Catal. A 303 (2006) 81–87.
[10] M.C. Capel-Sanchez, J.M. Campos-Martin, J.L.G. Fierro, M.P. de Frutos, A.P. Polo,
Chem. Commun. (Cambridge) (2000) 855–856.
[11] S. Krijnen, H.C.L. Abbenhuis, J.H.C. Van Hoof, R.A. Van Santen, Angew. Chem.,
Int. Ed. 37 (1998) 356–358.
[12] 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–
1221.
[13] M.D. Skowronska-Ptasinska, M.L.W. Vorstenbosch, R.A. van Santen, H.C.L.
Abbenhuis, Angew. Chem., Int. Ed. 41 (2002) 637–639.
[14] K. Wada, T.-a. Mitsudo, Catal. Surv. Asia 9 (2005) 229–241.
[15] P. Smet, J. Riondato, T. Pauwels, L. Moens, L. Verdonck, Inorg. Chem. Commun.
3 (2000) 557–562.
and 4 by the PDMS membrane, which presumably results in low lo-
cal water concentrations, as well as higher [alkene]:[H2O2] ratios at
the Ti center than in the bulk reaction medium; both of these effects
favor the epoxide selectivity and minimal leaching of titanium.
However, given that the principle deactivation pathway of Ti-based
epoxidation catalysts is leaching of Ti into solution [34,35], some
loss of Ti from immobilized tripodal Ti silsesquioxane epoxidation
catalysts is likely upon extended use. If Ti leaching from were to oc-
cur, the tripodal sites could be easily regenerated by the addition of
an appropriate Ti compound, either in situ [36] or ex situ. Since the
incompletely condensed silsesquioxane contains three chemically
identical silanol groups, the tripodal Ti site should be formed
exclusively.
[16] K. Wada, K. Hirabayashi, N. Watanabe, S. Yamamoto, T. Kondo, T.-A. Mitsudo,
M. Inoue, Top. Catal. 52 (2009) 693–698.
[17] R.F. Parton, I.F.J. Vankelecom, D. Tas, K.B.M. Janssen, P.-P. Knops-Gerrits, P.A.
Jacobs, J. Mol. Catal. A: Chem. 113 (1996) 283–292.
Acknowledgments
[18] D.F.C. Guedes, T.C.O. MacLeod, M.C.A.F. Gotardo, M.A. Schiavon, I.V.P. Yoshida,
K.J. Ciuffi, M.D. Assis, Appl. Catal. A 296 (2005) 120–127.
[19] T.C.O. MacLeod, V.P. Barros, A.L. Faria, M.A. Schiavon, I.V.P. Yoshida, M.E.C.
Queiroz, M.D. Assis, J. Mol. Catal. A: Chem. 273 (2007) 259–264.
[20] W.L.F. Armarego, D.D. Perrin, Purification of Laboratory Chemicals, fourth ed.,
Butterworth-Heinemann, Oxford, UK, 1997. 512 pp.
[21] D.C. Bradley, I.M. Thomas, J. Chem. Soc. (1960) 3857–3861.
[22] T. Mwangi Martin, M.B. Runge, B. Bowden Ned, J. Am. Chem. Soc. 128 (2006)
14434–14435.
[23] F.J. Feher, T.A. Budzichowski, K. Rahimian, J.W. Ziller, J. Am. Chem. Soc. 114
(1992) 3859–3866.
[24] I.E. Buys, T.W. Hambley, D.J. Houlton, T. Maschmeyer, A.F. Masters, A.K. Smith,
J. Mol. Catal. 86 (1994) 309–318.
Thanks are expressed to the Chemistry Department and the
University of Kentucky (Office of the Vice President for Research)
for support of this work. Thanks are also expressed to the Egyptian
Government (Ministry of Education) for the award of a fellowship
to E.A. NMR instruments utilized in this research were funded in
part by the CRIF program of the US National Science Foundation
(Grant No. CHE-9974810). Finally, we thank Mr. John Layton for
his assistance with NMR spectroscopy experiments.
[25] W. Fan, P. Wu, T. Tatsumi, J. Catal. 256 (2008) 62–73.
[26] G. Bellussi, M.S. Rigutto, Stud. Surf. Sci. Catal. 85 (1994) 177–213.
[27] B. Notari, Catal. Today 18 (1993) 163–172.
[28] R. Neumann, M. Levin-Elad, J. Catal. 166 (1997) 206–217.
[29] G. Langhendries, D.E. De Vos, G.V. Baron, P.A. Jacobs, J. Catal. 187 (1999) 453–
463.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2010.05.003.
[30] D.E. De Vos, G.V. Baron, F. van Laar, P.A. Jacobs, Nanostruct. Catal. (2003) 311–
328.
References
[31] R.A. Sheldon, J.A. Van Doorn, J. Catal. 31 (1973) 427–437.
[32] N. Lee Jessamine, C. Park, M. Whitesides George, Anal. Chem. 75 (2003) 6544–
6554.
[1] T. Maschmeyer, F. Rey, G. Sankar, J.M. Thomas, Nature (London) 378 (1995)
159–162.
[2] M. Crocker, R.H.M. Herold, A.G. Orpen, M.T.A. Overgaag, J. Chem. Soc., Dalton
Trans. (1999) 3791–3804.
[3] J.K.F. Buijink, J.J.M. van Vlaanderen, M. Crocker, F.G.M. Niele, Catal. Today 93–
95 (2004) 199–204.
[4] H.P. Wulff, British Patent 1,249,079, 1971, to Shell.
[5] R. Hutter, T. Mallat, D. Dutoit, A. Baiker, Top. Catal. 3 (1996) 421–436.
[33] A. Corma, P. Esteve, A. Martinez, J. Catal. 161 (1996) 11–19.
[34] L.J. Davies, P. McMorn, D. Bethell, P.C.B. Page, F. King, F.E. Hancock, G.J.
Hutchings, J. Catal. 198 (2001) 319–327.
[35] W.H. Richardson, V.F. Hodge, J. Org. Chem. 35 (1970) 4012–4016.
[36] M. Crocker, W. Derks, H. Dirkzwager, R.H.M. Herold, F.G.M. Niele, R.J.F.
Wermeling, Patent WO09949972, 1999, to Shell Research.