Regioselective epoxidation of different types of double bonds over large-pore titanium silicate Ti-β

Article abstract of DOI:10.1016/j.molcata.2010.05.024

Regioselective epoxidation of different types of double bonds located within the cyclic and acyclic parts of bulky olefins has been investigated using large-pore titanium silicate Ti-β in the presence of dilute aqueous H ₂O₂ as oxidant under mild liquid-phase conditions. Our experimental results revealed that side-chain vinylic double bonds are selectively epoxidized than those in the cyclohexene-ring. The epoxidation tendency of various bulky olefins with different positional and/or geometric isomers over Ti-β follows the order: terminal -CC- > ring -CC- ≈ bicyclic ring -CC- > allylic C - H bond. Unlike 4-vinyl-1-cyclohexene, epoxidation of an equimolar mixture of cyclohexene and 1-hexene under identical conditions using Ti-β exhibits completely different selectivity and product distributions. Steric factor and accessibility of reactants to active Ti-sites are responsible for the observed regioselectivity of bulky alkenes.

Full text of DOI:10.1016/j.molcata.2010.05.024

Journal of Molecular Catalysis A: Chemical 328 (2010) 60–67
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Heterogeneous catalysis
Regioselective epoxidation of different types of double bonds
over large-pore titanium silicate Ti-␤
Manickam Sasidharana, Asim Bhaumikb,∗
a Department of Chemistry, Faculty of Science and Engineering, Saga University, 1 Honjo-machi, Saga 840-8502, Japan
b Department of Materials Science, Indian Association for the Cultivation of Science, Jadavpur, Kolkata 700 032, India
a r t i c l e i n f o
a b s t r a c t
Article history:
Regioselective epoxidation of different types of double bonds located within the cyclic and acyclic parts
of bulky olefins has been investigated using large-pore titanium silicate Ti-␤ in the presence of dilute
aqueous H2O2 as oxidant under mild liquid-phase conditions. Our experimental results revealed that
side-chain vinylic double bonds are selectively epoxidized than those in the cyclohexene-ring. The epox-
idation tendency of various bulky olefins with different positional and/or geometric isomers over Ti-␤
follows the order: terminal –C C– > ring –C C– ≈ bicyclic ring –C C– > allylic C − H bond. Unlike 4-
vinyl-1-cyclohexene, epoxidation of an equimolar mixture of cyclohexene and 1-hexene under identical
conditions using Ti-␤ exhibits completely different selectivity and product distributions. Steric factor
and accessibility of reactants to active Ti-sites are responsible for the observed regioselectivity of bulky
alkenes.
Received 10 March 2010
Accepted 31 May 2010
Available online 8 June 2010
Keywords:
Catalysis
Epoxidation
Liquid-phase selective oxidation
Titanium silicate
Ti-␤
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
oxidant as in the case of epoxidation of styrene to styrene oxide
using urea/H2O2 adduct as an oxidant over medium-pore TS-1 as
Selective catalytic oxidation reactions have been studied exten-
sively for over three decades for the synthesis of value added fine
chemicals [1–8]. Epoxidation is an important class of reaction as
epoxides are widely utilized as stabilizer, lubricant, additives, sur-
face modifier and so on. The advent of transition metal containing
molecular sieves [3–8] has opened new arenas in the selective oxi-
dation chemistry using dilute H2O2 as oxidant under liquid-phase
conditions. Shape selectivity of molecular sieves or metallosilicates
often depends on their capability to discriminate molecules on the
basis of size with high precision. Those molecules which (both reac-
tant and product) diffuse freely across the channels are found in
higher percentage in the product mixture. Furthermore, this kind
of shape selectivity has been mostly studied in reversible acid-
catalyzed reactions at high-temperature driven by thermodynamic
factors as in the case of medium-pore ZSM-5 zeolite. However,
the oxidation reactions are normally irreversible and influenced
by electronic factor, steric factor and/or restricted transition-state
geometry [7]. Thus the reaction rate can be reduced or inhibited
altogether if the available space is insufficient to accommodate
the steric or geometric demands of the required intermediates or
transition states. Furthermore, occasionally the selectivity of a par-
ticular product has been increased by some external factors as in
the case of hydroxylation of benzene [9] or controlled release of
reported by Laha and Kumar [10].
The selectivity of epoxidation reaction over heterogeneous cat-
alysts is often controlled by hydrophobic microenvironment at
the catalyst surface [11]. On the other hand in the homogeneous
medium the electronic factor is the main criteria. Clerici and
Ingallina [12] have studied the influence of steric factor on the
epoxidation of various butene isomers over TS-1 and found that
the reactivity was decreased when the number of substituents on
the reaction site increases. The steric factor also plays a major role
in the chemoselective transformation of functional groups over TS-
1 [13]. In the case of chemoselective epoxidation of allyl alcohol,
the double bond is selectively epoxidized in the absence of any
substituents situated on it; whereas –OH groups of allyl alcohol
undergo oxidation when a methyl substituent is located on the
double bond. Thus the various factors governing the reactivity of
–C C– is well documented over medium-pore TS-1, whereas large-
pore titanosilicates such as Ti-␤ can show interesting selectivity
features depending upon the nature of the double bond. Compar-
ative study on the epoxidation of allyl alcohol, linear olefins, and
norbornene has been made over Ti-␤ and found that epoxidation
of allyl alcohol is governed by electronic factor, while norbornene
is controlled by steric factor [14]. On the other hand, long chain
unsaturated alcohols undergo selective oxidative cyclization to
tetrahydrofurans/tetrahydropyrans over TS-1/Ti-␤ in high yields
[15]. We report herein the regioselective epoxidation of different
bulky olefins containing two types of double bonds (reaction sites);
one is the terminal double bond in the side-chain alkane and in
∗
Corresponding author.
E-mail addresses: msab@iacs.res.in, msab@mahendra.iacs.res.in (A. Bhaumik).
1381-1169/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2010.05.024

M. Sasidharan, A. Bhaumik / Journal of Molecular Catalysis A: Chemical 328 (2010) 60–67
61
the other –C C– is located in cyclohexene-ring using large-pore
Ti-␤/H2O2 catalytic system. The observed results are corroborated
with electronic and steric factors as well as the interaction of differ-
ent double bonds to active Ti-sites in the channels of titanosilicates.
To obtain in-depth knowledge over the selectivity, epoxidation of
several linear hexenes, hexadienes, and an equimolar mixture of
various terminal alkenes and cyclohexene have also been carried
out under similar conditions to understand various factors involved
in the regioselective epoxidation of bulky olefins over large-pore
Ti-␤.
2. Experimental
The synthesis of various titanosilicates has been carried out
according to the reported procedures. In a typical synthesis of
hydrophobic aluminum-free Ti-␤ [16] under strongly alkaline
medium (pH = 12.6) the following procedure was used: Ti(OBu)4
(Aldrich) was treated with H2O2 at room temperature to form
titanoperoxo complex, which was then mixed with 40% aqueous
tetraethylammonium hydroxide (TEAOH, Aldrich) while stirring
vigorously. Then the required amount of Nipsil VN-3 silica and 3%
of dealuminated seed crystals were added followed by additional
stirring to get homogeneous paste like gel. The resulting gel (chemi-
cal composition: SiO2:0.025 TiO2:0.336 H2O2:0.55 TEAOH:6.6 H2O)
was transferred into a Teflon-lined autoclave, and the crystalliza-
tion was carried out at 413 K for 5 days. After crystallization the
product was centrifuged, and washed thoroughly with distilled
water to remove the organic base. The as-synthesized Ti-␤ was
dried at 373 K and calcined in a controlled flow of O2 for 12 h at
793 K. The medium-pore TS-1 was synthesized according to the
modified procedure by Thangaraj et al. [17]. In a typical synthesis
tetraethyl orthosilicate (21 g, TCI) was hydrolyzed by using 33.8 g
of tetrapropyl ammonium hydroxide (20% aqueous, Aldrich) for
3 h under vigorous stirring. Then 1.15 g tetrabutyl orthotitanate
in 5 g of anhydrous isopropyl alcohol was added dropwise while
stirring. After 1 h of continuous stirring, 27 g water was added into
the resulting clear homogeneous liquid, and the crystallization was
carried out at 443 K for 24 h under agitation. After crystallization,
the solid product was washed, dried, and calcined at 823 K for 12 h.
In the case of TS-2 [18], the synthesis procedure is similar to that
of TS-1 except that tetrabutylammonium hydroxide was used as
organic base, and the crystallization was carried out under static
conditions.
Fig. 1. UV–vis spectra of different titanosiliciates; (bottom to top) TS-1, TS-2, Ti-B-
MCM-22, Ti-␤ and Ti-MCM-41, respectively.
and the contents were vigorously stirred using a magnetic stirrer.
Bromobenzene was used as an internal standard for the estima-
tion of various products. Progress of the reaction was monitored
by analyzing the reaction mixtures at various time intervals with
a capillary gas chromatograph (Shimadzu 14A, OV-1 columns with
flame ionization detectors). The various authentic epoxy products
were purchased commercially and also converted into diol prod-
ucts (to confirm the diol products formed in the reaction mixture)
using H-mordenite catalyst. Furthermore, GCMS splitting pattern
was also employed to identify the unknown products.
3. Results and discussion
3.1. Characterization of the catalysts
The synthesis of microporous Ti-B-MCM-22 [19] was carried
out as follows: an aqueous solution prepared by dissolving piperi-
dine in deionized water was divided into two equal parts. Required
amount of tetrabutyl orthotitanate or boric acid was added to each
piperidine solution under vigorous stirring. Silica was also divided
into two equal parts and added gradually to the solutions con-
taining Ti and B. The gels were then mixed together and stirred
for 2 h to obtain a gel with the following molar composition of
SiO2:0.02 TiO2:0.67 B2O3:1.4 piperidine:19 H2O. The crystalliza-
tion was carried out by keeping the gels in an air-tight autoclave
through progressive heating at 403, 423 and 443 K for 1, 1 and 5
days, respectively. After the crystallization, the solid was dried and
the extra framework B- and Ti-species were removed by treatment
with 2 M HNO3 at 373 K for 20 h. The samples were calcined at
803 K to remove the remaining organic species. The mesoporous
Ti-MCM-41 was synthesized using dodecyltrimethyl ammonium
chloride (TCI) and tetramethyl ammonium hydroxide (20% aque-
ous) according to the reported procedure [20].
All the above-synthesized titanosilicates were characterized
using XRD, FTIR and UV–vis spectroscopy, and scanning electron
microscopy. The powder X-ray diffractograms of various titanosili-
cates and the corresponding scanning electron microscope images
reveal that the samples are phase pure and completely free from
amorphous materials. Furthermore, FTIR spectra of these samples
show characteristic absorption band at 960 cm−1 suggesting the
presence of titanium in the framework position. Fig. 1 shows the
UV–vis spectra of various titanium silicates. The presence of strong
band at 220 nm and the absence of any band either at 260 or 300 nm
indicate the absence of octahedral TiO2 cluster as impurities and
thus titanium sites exist mainly in the tetrahedral positions [21] in
the framework of Ti-␤.
3.2. Effect of reaction temperature and time
The catalytic reactions were carried out in a 25 mL round bottom
glass reactor fitted with a water condenser through which ice-cold
water was circulated. In a typical reaction 5 mmol of substrate,
5 mmol of H2O2 (31 wt% aqueous), 20 wt% of Ti-␤ (Si/Ti = 43) with
respect to substrate and 5 mL of methanol were mixed together
The regioselective epoxidation of 4-vinyl-1-cyclohexene with
double bonds both at side-chain and cyclohexene-ring has been
investigated systematically using Ti-␤ by scrutinizing the reactiv-
ity at different temperature and time. Fig. 2 illustrates the progress
in olefin conversion at different time intervals. Separate experi-

62
M. Sasidharan, A. Bhaumik / Journal of Molecular Catalysis A: Chemical 328 (2010) 60–67
Fig. 3. Effect of temperature on the activity and product selectivity in the partial
oxidation of 4-vinyl-1-cyclohexene over Ti-␤.
Fig. 2. Effect of reaction time on the activity and product distributions in the partial
oxidation of 4-vinyl-1-cyclohexene over Ti-␤.
leading to high selectivity for the side-chain –C C– epoxidation
products. Nevertheless, as the temperature increases from 303
to 338 K, not only the activity of the Ti-␤ improves but it also
increasingly favors cyclohexene epoxidation probably due to easy
access of cyclic double bond. The high temperature also acceler-
ates the rate of consecutive reaction such as the transformation
of monoepoxide into diepoxide, which is the major product in
“ring epoxidation products”. Thus the selectivity for the bulky
cyclohexene-ring epoxidation increased at the expense of side-
chain vinylic epoxy products. The H2O2 selectivity decreased as the
temperature increased as expected.
mental runs were carried out under identical conditions for each
point of the graph to get better mass balance. The various products
formed on epoxidation of 4-vinyl-1-cyclohexene are illustrated in
Scheme 1. The epoxidation of side-chain vinylic –C C– mainly pro-
duced the solvolyzed glycol-monomethyl ether in addition to small
amount of the corresponding epoxide, aldehyde, and ketone when
methanol was used as solvent. For convenience the above mix-
ture of products is called “side-chain epoxidized products” in the
subsequent discussion. In addition, the –C C– located inside the
cyclohexene-ring produces epoxide as the major products; in par-
ticular diepoxide, a product with oxirane moieties situated both at
the vinylic-carbons and cyclohexene-ring is found to be the pre-
dominant product with traces of cyclohexene-monoepoxide; No
corresponding polyol/polymethoxylated products were detected
and combination of these products are called “ring epoxidation
products”. It is seen from Fig. 2 that 4-vinyl-1-cyclohexene con-
version increases with time as expected. Further, the product
distribution pattern suggests that, selectivity for the side-chain
epoxidized products is considerably higher than the cyclohexene-
ring derivatives. At the initial stage of reaction the selectivity for
the “side-chain epoxidation” is considerably high, which decreased
gradually due to successive reaction of primary products to produce
diepoxide or ring epoxidation products. The percentage of these
secondary products increased as the reaction is progressed. The
other products in Fig. 2 refer to the allylic oxidation of C–H bond
at the allylic position of the cyclohexene nucleus. After completion
of reaction in about 2 h, the conversion of H2O2 reached 92% with
selectivity of 78.0% for the epoxidation.
3.3. Effect of various titanium silicates
The epoxidation reaction has also been screened over vari-
ous titanosilicates such as TS-1, TS-2, Ti-B-MCM-22, Ti-␤, and
Ti-MCM-41 with different pore openings under identical reaction
conditions and obtained results are summarized in Table 1. Among
the various catalysts investigated, Ti-␤ (entry 1) with large-pores
coupled with three-dimensional pore connectivity shows the best
activity for the epoxidation of 4-vinyl-1-cyclohexene. The side-
chain vinylic –C C– undergoes facile epoxidation (66.7%) than the
cyclohexene-ring (32.2%). Thus the epoxidation of vinylic double
bond proceeds smoothly to corresponding epoxy products than the
cyclohexene-ring. Unlike Ti-␤, medium-pore titanosilicates such
as TS-1 (ZSM-5 structure, pore width 5.6 Å × 5.3 Å) and TS-2 (ZSM-
11 structure, pore width 5.4 Å × 5.3 Å) show less activity for the
oxidation of 4-vinyl-1-cyclohexene probably due to internal diffu-
sion limitation of bulky reactant and/or products in the channels of
these zeolites [22,23]. Nevertheless the epoxidation of side-chain
–C C– takes place smoothly than that located at the six-membered
cyclic ring which is consistent with the literature reports that
the TS-1 is highly suitable for the epoxidation of linear alkenes
compared to cyclohexene. But in the present study of epoxida-
Fig. 3 exhibits the product distribution in the temperature range
of 303–338 K using Ti-␤/H2O2 catalytic system in methanol solvent.
At low temperature, Ti-␤ shows less activity particularly at 303 and
313 K but the activity increased significantly after 323 K. Further-
more at temperatures below 323 K, the reaction of vinylic –C C–
bond with active sites occurs smoothly than the cyclohexene-ring
Scheme 1. Product distribution for the liquid-phase oxidation of 4-vinyl-1-cyclohexene over Ti-␤.

M. Sasidharan, A. Bhaumik / Journal of Molecular Catalysis A: Chemical 328 (2010) 60–67
63
Table 1
Influence of various titanium silicates for the epoxidation of 4-vinyl-1-cyclohexene.a.
Sr. no.
Catalyst
Particle sizec, ␮m
Conv. mole, %
H2O2 sel., %
Selectivity of products, %
Ratiod
Side-chain oxidized products
Ring oxidized products
Others
1
2
3
4
5
6
7
Ti-␤ (43)b
1.0
0.1–0.2
2–3
0.2–0.5
0.2–0.5
1.0
64
18
16
37
12
21e
19f
75
88
82
92
61
90
46
67
96
97
89
87
79
85
32
4
3
11
13
21
14
1.0
–
–
–
–
0.5
TS-1 (33)b
0.03
0.03
0.12
0.13
0.26
0.16
TS-2 (46)b
Ti-B-MCM-22 (55)b
Ti-MCM-41 (51)b
Ti-␤ (43)b
–
–
Ti-MCM-41 (51)b
0.2–0.5
a
Reaction conditions: 5 mmol of 4-vinyl-1-cyclohexene; 5 mmol H2O2 (31 wt% aqueous solution); 5 mL methanol; catalyst, 20 wt% with respect to substrate; temperature,
338 K; and reaction time 2 h.
b
The figures in the parentheses refers to Si/Ti ratio.
Particle size observed from scanning electron microscope.
Ratio of cyclohexene to vinylic side-chain –C C– oxidation products.
Reaction was performed for 10 min to reduce the conversion.
c
d
e
f
Reaction was performed for 5 h to increase the conversion.
tion of 4-vinyl-1-cyclohexene, it is highly likely that the double
bond in the side-chain or tail can interact with the Ti-sites through
the medium-pores to provide “side-chain epoxidized products”.
Entry 4 exhibits the activity of Ti-B-MCM-22, which is considerably
higher than TS-1 or TS-2; the increased performance may be due
to the presence of Ti-species in the 10-ring channels, supercages,
and the exterior pockets, which permit the easy access of the bulky
olefins to the active sites. However, selectivity for the “side-chain
epoxidized products” is invariably high (89%) irrespective of the
conversion level over these titanosilicates.
formance of Ti-MCM-41 in the epoxidation of linear olefins like
1-hexene was found to be less compared to cyclic or bulky olefins.
Furthermore, intrinsic activity of Ti-MCM-41 could be less because
the mesopores contain a large number of Si–OH groups, which
strongly adsorbs methanol and H2O2 under the reaction condi-
tions and the apolar substrate (4-vinyl-1-cyclohexene) is adsorbed
weakly, thus hindering the intrinsic activity of Ti-MCM-41. More-
over the selectivity for “side-chain epoxidized products” was found
to be 87%. Thus the catalytic performance of Ti-MCM-41 was also
found to be similar to that of microporous titanium counterparts
although former has mesopore architecture. Thus the activity of Ti-
␤ is different in the sense that it resembles to that of medium-pore
TS-1, TS-2 and Ti-B-MCM-22 by forming more side-chain vinylic
epoxidation products contrary to the expectation. In addition,
Ti-MCM-41 (entry 5), the titanium analogue of mesoporous
MCM-41 exhibits still less activity (12.1%) than the other titanosili-
cates in spite of having less diffusion limitation or steric hindrance.
It has been reported in the literature that [24] the catalytic per-
Table 2
Epoxidation of various cyclic olefins over Ti-␤/H2O2 system.a.
Sr. No
Substrate
Conv. mole, %
H2O2 sel., %
Product distribution
Side-chain oxidation products
Ratiod
Ring oxidation products
Others
32.7b
64.1c
58.5
75.3
64.4
33.7
32.5
1.8
1.4
0.52
0.51
1
66.1
28.5b
73.7c
61.0
80.4
60.2
58.5
36.3
37.3
3.4
4.1
0.6
2
3
0.37
23.9b
55.6c
81.5
86.6
95.0
71.5
5.0
–
0.05
0.37
26.6
1.9
26.9b
71.2c
69.0
90.5
43.0
54.6e
52.6
40.3
4.3
4.0
1.22
0.73
4
25.8b
34.4c
38.0
49.0
–
–
67f, 15g, 18h
65f, 21g, 14h
–
–
–
–
5
a
Reaction conditions: 5 mmol of substrate; 5 mmol of H2O2 (31 wt% aqueous solution); catalyst, 20 wt% with respect to substrate; 5 mL solvent; temperature 338 K; and
reaction time, 2 h.
b
Acetonitrile was used as solvent.
Methanol was used as solvent.
Ratio of ring double bond to side-chain epoxidation products.
Epoxy products of exo double bond.
Epoxidation product.
Allylic products.
Unknown products, possibly aromatic oxidation products.
c
d
e
f
g
h

64
M. Sasidharan, A. Bhaumik / Journal of Molecular Catalysis A: Chemical 328 (2010) 60–67
Table 3
Epoxidation of open-chain hexene isomers over Ti-␤/H2O2.a.
Entry
Substrate
Conv. mole,%
H2O2 sel., %
Product distribution, %
Epoxy products
Allylic oxidation products
3.8
Others
4.0
1
2
45.3
52.0
68.6
78.4
92.2
92.7 (80.5% epoxide)
7.3
9.7
4.6
–
3
52.2
83.4
85.3
89.0
5.0
4
61.0
87.6
6.4
a
Reaction conditions: 5 mmol of substrate; 5 mmol of H2O2 (31 wt% aqueous solution); catalyst, 20 wt% with respect to substrate; 5 mL methanol; temperature 338 K; and
reaction time, 2 h. “others” refer to a low boiling compound.
the ratio of “ring epoxidation products” to “side-chain epoxidized
products” is always much higher for large-pore Ti-␤ than either
Ti-MCM-41 or other medium-pore titanosilicates. We have also
attempted to compare the above ratios at similar conversion levels
by increasing the reaction time for the reaction over Ti-MCM-41
(entry 7, conversion = 19.1%), and decreasing that of Ti-␤ (entry 6,
conversion = 20.7%), but the observed product distribution is again
quite high for “side-chain epoxidized products”. Therefore the reac-
tivity and product distribution of these two types of double bonds in
4-vinyl-1-cyclohexene reveal that: (i) the medium-pore titanosil-
icates TS-1, TS-2, and Ti-B-MCM-22 exhibit high selectivity for
vinylic/terminal –C C– epoxidation (ii) large-pore Ti-␤ and meso-
porous Ti-MCM-41 also show preferential epoxidation of vinylic
double bonds and these results contrast with previous reports
that Ti-␤ smoothly epoxidize the cyclohexene-ring. Driven by the
curiosity from the above results, the regioselective epoxidation was
further studied thoroughly with various bulky alkenes containing
two types of double bonds using Ti-␤ as the catalyst as it displays
the best activity.
1-cyclohexene; whereas glycol-monomethyl ether was obtained
predominantly due to further rearrangement of the products in the
presence of methanol solvent (entry 1). In addition the reactivity of
4-vinyl-1-cyclohexene in acetonitrile is decreased almost by one
half compared to that in methanol but the selectivity for “side-
chain epoxidized products” is dominated as observed in the case
of methanol. These two experiments by using different solvents
obviously indicate that the solvents play very little role in deter-
mining the selectivity and/or reactivity of different –C C– bonds.
However, one would expect that the epoxidation of the double bond
in the cyclic ring should be faster than the terminal double bond
and this is indeed the case for the homogeneously catalyzed epox-
idation reaction [25]. Thus the above results indicate that besides
electronic factor, possibly the steric factor also plays dominant role
in the epoxidation.
In the case of limonene (entry 2) both terminal and cyclic –C C–
has a methyl group in its vicinity and therefore terminal –C C–
is again preferentially epoxidized over the cyclic system immate-
rial of the solvents used. The substrate 4-vinyl-1-norbornene with
bicyclic ring skeleton (entry 3) shows very good regioselectivity
towards side-chain epoxidation with subsequent isomerization to
the corresponding aldehyde in 95% yield in acetonitrile and negli-
gible amount of ring epoxidation; whereas for the same substrate
methanol medium produced considerable amount of “ring epoxi-
dation products” (26.6%) at the expense of “side-chain epoxidized
products” (71.5%). The reactivity of 4-ethylidene-1-norbornene
(entry 4), a compound with exo double bond with a methyl sub-
stituent shows increased reactivity of –C C– inside the bicyclic
ring system. The different reactivity pattern of 4-ethylidene-1-
norbornene may be due to conformational and steric hindrance of
3.4. Epoxidation of different bulky olefins with two types of
double bonds over Ti-ˇ
Table 2 exhibits the reactivity and product selectivities toward
different bulky olefins containing two types of double bonds inves-
tigated using Ti-␤/H2O2 catalytic system. In addition to methanol,
we have also employed acetonitrile solvent (entry 1) to observe any
pattern change in the product selectivity. It is relevant to mention
that the experiment involving acetonitrile solely produced epoxide,
aldehyde (after rearrangement), and diol products for 4-vinyl-
Table 4
Epoxidation of hexadiene isomers over Ti-␤/H2O2.a.
Substrate
Conv. mole, % H2O2 sel., % Product distribution, %
Terminal epoxy products Internal epoxy products Diepoxy products Allylic oxidation 2,4-Hexadieneb
48.6
55.0
52.9
49.0
32.8
5.7
–
29.6
24.9
3.8
4.8
33.8
36.5
28.9
46.1
51.0
15.4
37.4
12.5
2.2
32.5
a
Reaction conditions: 5 mmol of substrate; 5 mmol of H2O2 (31 wt% aqueous solution); catalyst, 20 wt% with respect to substrate; 5 mL methanol; temperature 338 K; and
reaction time, 2 h.
b
2,4-Hexadiene is readily formed by isomerization since the conjugated internal diene is highly stable.

M. Sasidharan, A. Bhaumik / Journal of Molecular Catalysis A: Chemical 328 (2010) 60–67
65
Table 5
Epoxidation of equimolar mixture of cyclohexene and various terminal olefins over Ti-␤/H2O2.a.
Entry
Substrates
Terminal alkenes
Conv., mole, %
H2O2 sel., %
Conv., mole, %
23.8
Epoxide, %
42.0
Diol, %
56.6
Othersd, %
1.4
Epoxide, %
83.3
Diol, %
16.7
1b
14.1
83.0
2b
3b
4b
19.5
19.9
19.6
42.6
40.0
29.3
55.6
57.5
69.3
1.8
2.5
1.4
13.3
11.5
10.9
91.3
90.5
91.0
8.6
9.5
9.0
79.0
80.8
82.5
5b
20.5
22.8
75.8
1.4
11.0
91.8
8.2
75.2
6b
7b
19.4
21.0
27.6
27.5
71.5
72.5
0.9
-
10.3
10.6
91.0
92.1
9.0
7.9
76.5
76.1
8c
9c
12.4
11.9
12.5
5.2
7.1
7.2
92.4
89.3
90.8
2.2
3.6
2.0
2.1
1.9
1.1
68.8
72.0
71.0
31.2
28.0
29.0
45.9
47.5
44.0
10c
a
Reaction conditions: 5 mmol of cyclohexene and 5 mmol of terminal alkene; 10 mmol of H2O2 (31 wt% aqueous solution); catalyst, 20 wt% with respect to substrate; 10 mL
acetonitrile; temperature 338 K; and reaction time, 2 h.
b
Ti-␤.
Ti-MCM-41.
Others indicates allylic oxidation products.
c
d
exo double bonds toward the accessibility to active Ti-sites. In addi-
tion, the exo –C C– at the bicyclic ring is flanked by three methyl
substituents, whereas that in the bicyclic ring is strained by two
methyl substituents. As a result the –C C– in the ring is relatively
less hindered and reacts faster by interacting with active sites in
the porous matrix. The substrate 1-phenylcyclohexene (Table 2,
entry 5) illustrates the oxidizing ability of Ti-␤ towards –C C– bond
versus aromatic C–H bond. Our experimental results suggest that
the oxidation predominantly takes place at the –C C– bond fol-
lowed by allylic position, whereas the aromatic C–H bond is least
affected. The observed H2O2 selectivity is slightly higher in the case
of methanol due to increased reactivity than acetonitrile.
suggest that, the vinylic –C C– bonds are easily accessible than the
one at cyclic ring unless the terminal –C C– is highly crowded by
many substituents as in the case of 4-ethylidene norbornene (entry
4) with exo double bond. Thus the oxidation of different double
bonds, allylic –C–H, and aromatic –C–H within the same molecule
∼
follows the order: terminal –C C– > cyclic ring –C C– bicyclic
–C C– > allylic C–H > aromatic C–H.
=
3.5. Comparison of reactivity of various hexene isomers
To gain further insight in the reactivity behavior of side-
chain terminal double bonds,
a series of experiments were
The above investigation over various alkenes with different
types of –C C– bonds clearly reveals that Ti-␤ exerts less diffu-
sion limitation as it smoothly epoxidize both side-chain vinylic as
well as cyclic ring system. However, the observed product selectiv-
ity appears to be influenced mostly by both steric crowding around
the –C C– bond, and conformational flexibility of these substrates
near the active Ti-sites. Thus, when a molecule has more than one
double bonds (reaction sites) in different position, the more acces-
sible one reacts readily than the one, which is not accessible. The
accessibility of the reaction sites is ultimately governed by the
steric hindrance, geometry and/or conformation of the interacting
or reacting molecules with the active Ti-species under the exper-
imental conditions. The observed results over these bulky olefins
carried out with various hexene and hexadiene isomers (lin-
ear olefins) over Ti-␤ under similar experimental conditions and
the results are summarized in Tables 3 and 4. The conver-
sion as well as H2O2 selectivity of 1-hexene (terminal alkene,
Table 3, entry 1) is lower than that of trans-2-hexene (an inter-
nal alkene, entry 2) consistent with the earlier report [14], and
the increased activity of 2-hexene is attributed to the electron
releasing inductive effect of methyl substituents. Entries 3 and
4 (Table 3) exhibit the reaction of trans- and cis-3-hexene (geo-
metrical isomers); the trans-3-hexene is less reactive than the
cis-3-hexene suggesting that the reaction site is less hindered
in the case of later than the former leading to higher conver-
sion. Thus though the Ti-␤ is comprised of large-pore structure

66
M. Sasidharan, A. Bhaumik / Journal of Molecular Catalysis A: Chemical 328 (2010) 60–67
Scheme 2. Transition-state intermediates at the Ti-␤ surface during the liquid-phase oxidation of 4-vinyl-1-cyclohexene over Ti-␤ with dilute aqueous H2O2.
with a free flow of molecules in all the direction along the chan-
nels, the geometry and conformational flexibility of a molecule
vis-à-vis accessibility to the active sites also determines the reac-
tivity.
mer showed more conversion than the later in accordance with the
literature reports [22]. As the length of terminal olefins increases
gradually from C6 to C12 , the conversion of cyclohexene is not
decreased appreciably though it fluctuated within a range. This
result suggests that large-pore Ti-␤ exhibits very little diffusion
limitation for the epoxidation of a mixture of cyclohexene and n-
hexene unlike medium-pore TS-1, where diffusion play a major
role to provide more n-hexene epoxidation [26,27]. Furthermore,
the epoxidation of cyclohexene mainly produced the correspond-
ing cyclohexane-1,2-diol as the major product and the percentage
of diol product also increases with chain length of the linear termi-
nal alkenes, probably due to the increased molecular traffic along
the channels which makes the epoxides prone to hydrolysis before
it is dragged away to the bulk reaction medium.
The conversion of various linear terminal alkenes are invari-
ably lower than cyclohexene and it decreased slightly as the chain
length increases from C6 to C12 ; probably steric effect plays a
role as the hydrophobic alkyl chain length increases. These termi-
nal olefins produced epoxides as the major product with a small
amount of the corresponding diol. Entries 8–10 show the reactivity
over Ti-MCM-41 and also exhibits increased epoxidation towards
cyclohexene-ring rather than terminal olefins but the observed
activity was found to be lower comparable to Ti-␤. The differ-
ence in activity between Ti-␤ and Ti-MCM-41 is presumably due to
amorphous pore wall with high hydrophilic interior and/or inac-
cessibility of titanium active sites located inside the silica walls of
Ti-MCM-41 as reported by Oldroyd et al. [28]. Thus the above exper-
iments clearly reveal that the nature of the active sites of these
two titanosilicates differ intrinsically; presumably the bond length
and/or bond angles of the coordinated Ti atom may play a role in the
formation of titaniumhydroperoxo complex with H2O2, and subse-
quent release of oxygen atom to the organic substrate during the
reaction. Thus the above experiments using an equimolar mixture
of cyclohexene and terminal alkenes confirm the earlier reports
that Ti-␤ is more active towards bulky cyclohexene than the ter-
minal alkenes. However, the selectivity and product distribution
is altered for bulky alkenes with both terminal and cyclic double
bonds as observed in the present study. The catalyst Ti-␤ retains
86% of original activity after five successive reuses after activation
Furthermore, linear 1,5-hexadiene (entry 1, Table 4) with two
–C C– bonds located at the terminal positions shows less reactivity
(48.6%) than other dienes with both terminal as well as inter-
nal –C C– bonds. The higher reactivity of 1,4-hexadiene (55.1%)
over 1,5-hexadiene could be due to the favorable electronic factor
(alkyl substitution) in the former. But the reactivity of conjugated
1,3-hexadiene (entry 3, Table 4) was lower than 1,4-hexadiene,
possibly increased stability due to conjugation may resist the oxi-
dation. Thus the epoxidation of internal –C C– bond dominates
over the terminal counterparts for 1,4- and 1,3-hexadiene isomers.
It is important to note that all three hexadiene isomers undergo iso-
merization to give the most stable conjugated 2,4-hexadiene as the
major side product. This could be facilitated either by acidic tita-
niumhydroperoxo complex or weak Brönsted acidity of Ti-␤ itself.
Thus the reactivity of side-chain and/or internal –C C– is governed
mainly by the electronic factor in the case of hexene and hexa-
diene isomers except 3-hexene geometrical isomers where steric
factor plays a dominant role. The above results over linear olefins
clearly demonstrate that the reactivity of terminal double bond is
always lower than the internal double bonds confirming the ear-
lier reports. To prove the reactivity of 4-vinyl-1-cyclohexene and
similar substrates without any ambiguity, a series of experiments
were performed by externally mixing the individual compounds of
cyclohexene (ring –C C–) and different terminal alkenes (terminal
–C C–) in an equimolar amounts.
3.6. Epoxidation of equimolar mixture of cyclohexene and
different terminal alkenes
The experiments were carried out using Ti-␤/H2O2/acetonitrile
catalytic system under similar experimental conditions and the
reactivity along with product distribution are given in Table 5.
Entries 1–7 (Table 5) show the epoxidation of an equimolar mixture
of various linear terminal alkenes of C6 to C12 with cyclohexene.
Among the bulky cyclohexene and linear terminal alkenes, the for-

M. Sasidharan, A. Bhaumik / Journal of Molecular Catalysis A: Chemical 328 (2010) 60–67
67
at 573 K for 5 h for the epoxidation of 4-vinyl-1-cyclohexene. The
4. Conclusions
loss of activity may be due to the partial removal of framework tita-
nium either by the action of H2O2 or the bulky reaction products
such as diol and glycol-monomethyl ethers as reported by Davis et
al. [29].
Our experimental results suggested that large-pore titanium sil-
icate Ti-␤ exhibited high catalytic activity in the epoxidation of
cyclic and acyclic double bonds. The reactivity of different bulky
olefins has been governed by steric hindrance and/or conforma-
tional flexibility near the active Ti-sites. Among the various bulky
3.7. Mechanistic aspects
olefins investigated using Ti-␤ catalyst, the epoxidation efficiency
∼
The epoxidation of 4-vinyl-1-cyclohexene over Ti-␤/H2O2
catalyst system proceeds smoothly as shown in Fig. 2. The exper-
imental results reveal that primary epoxy products both cyclic-
and acyclic-derivatives undergo consecutive secondary reactions
assisted by either weak Brönsted acid sites of Ti-␤ or acidic titanium
hydroperoxo complex [30,31]. Thus, the acyclic or terminal epox-
ides produce aldehyde and ketone by isomerization; whereas diol
and glycol-monomethyl ethers are formed by the nucleophilic ring
opening reaction with H2O and CH3OH molecules, respectively. In
addition, terminal epoxide also undergoes further epoxidation of
double bond in the cyclohexene-ring to produce diepoxide, which
is the major product among the “Ring epoxidation products”. From
the observed product distribution it is obvious that the stability
of diepoxide is quite higher compared to monoepoxide as seen
from negligible amount of polyol or polymethoxylated product
formation.
follows the order: terminal –C C– > cyclic ring –C C– bicyclic
=
–C C– > allylic C–H. Epoxidation of various hexene and hexadiene
isomers indicated that internal alkenes smoothly epoxidize than
the terminal alkenes. The mesoporous Ti-MCM-41 showed similar
selectivity as that of Ti-␤ but with lower activity in spite of its neg-
ligible diffusion limitation. The difference in the activity between
Ti-␤ and Ti-MCM-41 is attributed to the intrinsic nature of local
Ti-site and crystalline framework of Ti-␤. Thus the present investi-
gation, for olefins containing more than one double bonds (reaction
sites) at different position suggested that the reactivity is ultimately
governed by the steric hindrance, geometry and/or conformational
flexibility of the interacting molecules at the active Ti-active sites.
References
[1] C. Perego, A. Carati, P. Ingallina, M.A. Mantegazza, G. Bellussi, Appl. Catal. A:
Gen. 221 (2001) 63.
The observed reactivity of compounds with cyclic- and acyclic
double bonds over Ti-␤ is however in contrast with the literature,
i.e.–C C– in the cyclic ring should be more reactive than the ter-
minal –C C– bond. Furthermore the epoxidation of other similar
substrates also showed high reactivity towards terminal double
bonds. The observed reactivity can be explained by considering
the formation of possible Ti-species or intermediate that leads to
epoxidation and the kind of interaction between such active inter-
mediate with 4-vinyl-1-cyclohexene is illustrated in Scheme 2. The
interaction of H2O2 with Ti-site produces five coordinated species
(titanium hydroperoxo complex) [30–32], which could be stabi-
lized due to the formation of 5-membered ring. This not only makes
the hydroperoxo complex Brönsted acidic but also the peroxo-
oxygen becomes more electron deficient. Now the interaction of
these activated intermediates in the channels and pores of Ti-␤
with 4-vinyl-1-cyclohexene or other substrates play a crucial role
in determining the reactivity of a particular double bond. Thus, the
interaction of terminal –C C– with active Ti-intermediate (Struc-
ture I, Scheme 2) undergoes less strain as it contains no substituents
on the double bond. However, in the case of cyclohexene-ring of 4-
vinyl-1-cyclohexene, which exist in two different conformational
structures i.e. chair and boat (boat form of cyclohexene is less
stable than the chair form) forms makes the interaction more dif-
ficult (Structure II and III, Scheme 2). The observed low reactivity
vis-à-vis selectivity for cyclic ring oxidation at lower temperature
(Fig. 3) may be the consequence of conformational as well as steric
resistance towards the accessibility at the active Ti-sites. Thus, the
observed product distribution can be explained by considering the
interaction of a molecule as a whole entity instead as ‘terminal’
or ‘cyclic’ double bonds. Therefore, the epoxidation of terminal
–C C– occurs prominently unless it is highly strained by many
substituents as in the case of 4-ethylidene-1-norbornene with exo
double bond. The same explanation can also hold good for geomet-
rical isomers, where the interaction of cis-isomer is more effective
than the trans-isomer.
[2] S. Mukherjee, M. Nandi, K. Sarkar, A. Bhaumik, J. Mol. Catal. A: Chem. 301 (2009)
114.
[3] R.A. Sheldon, J. Dakka, Catal. Today 19 (1994) 215.
[4] A. Bhaumik, T. Tatsumi, Chem. Commun. (1998) 463.
[5] P. Selvam, S.K. Mohapatra, Micropor. Mesopor. Mater. 73 (2004) 137.
[6] W.B. Fan, R.-G. Duan, Y. Yokoi, P. Wu, Y. Kubota, T. Tatsumi, J. Am. Chem. Soc.
130 (2008) 10150.
[7] P. Serna, L.A. Baumes, M. Moliner, A. Corma, J. Catal. 258 (2008) 25.
[8] G.A. Eimer, I. Diaz, E. Sastre, S.G. Casuscelli, M.E. Crivello, E.R. Herrero, J. Perez-
Pariente, Appl. Catal. A: Gen. 343 (2008) 77.
[9] P. Chammingkwan, W.F. Hoelderich, T. Mongkhonsi, P. Kanchanawanichakul,
Appl. Catal. A: Gen. 352 (2009) 1.
[10] S.C. Laha, R. Kumar, J. Catal. 204 (2001) 64.
[11] J. Iglesias, J.A. Melero, J. Sainz-Pardo, J. Mol. Catal. A: Chem. 291 (2008) 75.
[12] M.G. Clerici, P. Ingallina, J. Catal. 140 (1993) 71.
[13] A. Bhaumik, R. Kumar, P. Ratnasamy, Stud. Surf. Sci. Catal. 84 A (1994) 1883.
[14] J.C. Van der Waal, M.S. Rigutto, H. van, Bekkum, Appl. Catal. A: Gen. 167 (1998)
331.
[15] A. Bhaumik, T. Tatsumi, J. Catal. 182 (1999) 349.
[16] T. Blasco, M.A. Camblor, A. Corma, P. Esteve, J.M. Guil, A. Martinez, J.A. Perdigon-
Melon, S. Valencia, J. Phys. Chem. B 102 (1998) 75.
[17] A. Thangaraj, R. Kumar, S.P. Mirajkar, P. Ratnasamy, J. Catal. 130 (1990) 1.
[18] J.S. Reddy, R. Kumar, P. Ratnasamy, Appl. Catal. 58 (1990) L1.
[19] M. Sasidharan, P. Wu, T. Tatsumi, J. Catal. 205 (2002) 332.
[20] P.T. Tanev, M. Chibwe, T.J. Pinnavaia, Nature 368 (1994) 321.
[21] D. Chandra, S.C. Laha, A. Bhaumik, Appl. Catal. A Gen. 342 (2008) 29.
[22] F.Y. Lee, L. Lv, F. Su, T. Liu, Y. Liu, C.H. Sow, X.S. Zhao, Micropor. Mesopor. Mater.
124 (2009) 36.
[23] A. Corma, M.A. Camblor, P. Esteve, A. Martinez, J. Perez-Pariente, J. Catal. 145
(1994) 151.
[24] A. Tuel, Stud. Surf. Sci. Catal. 117 (1998) 159.
[25] S. Liu, J. Xiao, J. Mol. Catal. A: Chem. 270 (2007) 1.
[26] C.E. Ramachandran, H. Du, Y.J. Kim, M.C. Kung, R.Q. Snurr, L.J. Broadbelt, J. Catal.
253 (2008) 148.
[27] W. Fan, P. Wu, T. Tatsumi, J. Catal. 256 (2008) 62.
[28] R.D. Oldroyd, J.M. Thomas, T. Maschmeyer, P.A. MacFaul, D.W. Snelgrove, K.U.
Ingold, D.D.M. Wayner, Angew. Chem. Int. Ed. Engl. 35 (1996) 2787.
[29] L.J. Davis, P. McMorn, D. Bethell, P.C. Bulman Page, G. King, G.E. Hancock, G.J.
Hutchings, J. Mol. Catal. A: Chem. 165 (2001) 243.
[30] A. Bhaumik, P. Kumar, R. Kumar, Catal. Lett. 40 (1996) 47.
[31] Y. Goa, P. Wu, T. Tatsumi, J. Phys. Chem. B 108 (2004) 8401.
[32] U. Wilkenhoner, G. Langhendries, F.V. Laar, G.V. Baron, D.W. Gammon, P.A.
Jacobs, E.V. Steen, J. Catal. 203 (2001) 201.