Ti-oxo radicals and product selectivity in olefin oxidations over titanosilicate molecular sieves

Article abstract of DOI:10.1524/zpch.219.7.905.67088

The O-O bond of Ti-oxo species (Ti-peroxo, Ti-hydroperoxo or Ti-superoxo) generated on titanosilicate molecular sieves contacted with H₂O ₂ cleaves either heterolytically or homolytically. While the former type of O-O cleavage (generating non-radical reactive oxygen intermediates) leads to selective epoxide products, the latter (generating reactive oxygen radical intermediates, O₂^-? and HOO?) results in non-selective, allylic oxidation products in cyclohexene oxidation. These radicals were detected using EPR spin trapping techniques; 5,5-dimethyl-1- pyrroline-N-oxide (DMPO) was used as the spin trap. Radical quenchers like hydroquinone suppressed their concentration and enhanced epoxide selectivity. by Oldenbourg Wissenschaftsverlag, Muenchen.

Full text of DOI:10.1524/zpch.219.7.905.67088

Z. Phys. Chem. 219 (2005) 905–920
by Oldenbourg Wissenschaftsverlag, München

Ti-Oxo Radicals and Product Selectivity
in Olefin Oxidations over Titanosilicate
Molecular Sieves
∗
By Vasudev N. Shetti, D. Srinivas, and P. Ratnasamy
National Chemical Laboratory, Pune 411 008, India
Dedicated to Prof. Dr. Helmut Knözinger on the occasion
th
of his 70 birthday
(Received March 30, 2005; accepted April 11, 2005)
Titanosilicates / Molecular Sieves / Titanium-Oxo Radicals / Reactive
Oxygen Species / Oxidations / Product Selectivity / EPR-Spin Trap
Experiments / Cyclic Voltammetry of Titanosilicates
The O–O bond of Ti-oxo species (Ti-peroxo, Ti-hydroperoxo or Ti-superoxo) generated
on titanosilicate molecular sieves contacted with H O cleaves either heterolytically
2 2
or homolytically. While the former type of O–O cleavage (generating non-radical
reactive oxygen intermediates) leads to selective epoxide products, the latter (generating
−•
•
reactive oxygen radical intermediates, O
2
and HOO ) results in non-selective, allylic
oxidation products in cyclohexene oxidation. These radicals were detected using EPR
spin trapping techniques; 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) was used as the spin
trap. Radical quenchers like hydroquinone suppressed their concentration and enhanced
epoxide selectivity.
1. Introduction
The enormous interest in titanosilicate molecular sieves is due to their remark-
able catalytic activity in oxidation reactions using the environmentally benign
aqueous H O as the oxidant [1–3]. In the dehydrated state, the surface of
2
2
a well-prepared titanosilicate molecular sieve like TS-1, is known [4–8] to con-
4+
4+
tain mainly two types of Ti ions (1) those wherein the Ti ions are 4-fold
coordinated with lattice (–OSi) groups, Ti(OSi) , in near tetrahedral symme-
4
try, called the tetrapodal Ti ions, and (2) those wherein they are tripodally
coordinated to the surface, Ti(OSi) (OH), an extra-lattice OH group forming
3
*
Corresponding author. E-mail: p.ratnasamy@ncl.res.in

9
06
V. N. Shetti et al.
the fourth ligand. The tetrapodal species dominates in TS-1, TS-2 and Ti-beta
while the tripodals are more prevalent in Ti-MCM-41 and other mesoporous
materials [4–8]. On interaction with H O , (H + O ) or alkyl hydroperoxides
2
2
2
2
(ROOH), the Ti ions expand their coordination number to 5 or 6 and form
Ti-hydroperoxo/peroxo [9–16] and superoxo [17–24] complexes, which cat-
alyze the oxidation reactions. What is the influence of the nature and mode
of cleavage of the O–O bond (in the oxo-Ti species) on product selectivity?
In many oxidations catalyzed by titanosilicates it has been noted [1–3] that
product selectivity depends on a variety of factors including the type of crys-
talline structure (TS-1, Ti-beta, Ti-MCM-41), solvent, temperature and oxidant
(
H O , alkyl hydroperoxide, (H + O ), urea-H O etc.). For example, in olefin
2
2
2
2
2
2
oxidations, both epoxidation and allylic oxidations occur, their relative im-
portance varying with the factors mentioned above [1]. In the present paper,
an attempt is made to explore the relationship between (1) the structure of
the Ti-oxo complex, (2) the mode of cleavage of the O–O bond in Ti-oxo
species and (3) product selectivity in olefin epoxidation reactions over dif-
ferent titanosilicate molecular sieves. The redox behavior of Ti in TS-1 and
Ti-MCM-41 is investigated by cyclic voltammetric (CV) measurements. The
stability of the oxo-Ti species is investigated by UV-visible and EPR spectro-
•
•
−•
•
scopic studies. The O-centered radical species such as HO , HOO and O
2
−9
generated by homolytic O–O cleavage are very short lived (10 s for HO ,
for example). We have used the EPR-spin trapping method to detect such free
radicals. The O-centered radicals are reacted with a spin trap, 5,5-dimethyl-
1-pyrroline-N-oxide (DMPO), to generate relatively long-lived nitroxide free
radicals (spin adducts), easily detected by EPR spectroscopy. The spectral fea-
tures and the spin Hamiltonian parameters (superhyperfine coupling constant)
are highly sensitive to the type of radical adduct (DMPO–OH, DMPO–O2
and DMPO–OOH). Changes in radical concentrations in different solvents and
in the presence of known radical scavengers, like hydroquinone, were also
studied. Product selectivities in cyclohexene and styrene oxidations are found
to be influenced by the concentration of such radicals.
2. Experimental
2.1 Materials and characterization
TS-1, Ti-MCM-41, Ti-SBA-15 and ETS-10 were prepared and character-
ized as reported earlier [22, 25]. TS-1: Si/Ti = 33 (XRF); BET surface area
2
−1
≈
400 m /g; FT-IR characteristic band at 964 cm ; cuboidal morphology; par-
ticle size = 0.3 µm (SEM); Ti-MCM-41: Si/Ti = 46 (XRF); BET surface area
2
−1
=
960 m /g; FT-IR characteristic band at 968 cm ; average pore diameter =
3
3
nm; pore volume = 0.9 cm ; Ti-SBA-15: Si/Ti = 30 (XRF); BET surface
2
3
area = 790 m /g, pore volume = 1.7 cm and average pore diameter = 8.3 nm,
3
pore volume = 1.7 cm .

Ti-Oxo Radicals and Product Selectivity in Olefin Oxidations . . .
907
EPR spectra were recorded on a Bruker EMX spectrometer operating at
X-band (ν ≈ 9.45 GHz) frequency and 100 kHz field modulation [22]. EPR
spectral simulations were done using the Bruker Simfonia software package.
Spectral intensity calculations (by double interaction method) and other manip-
ulations were done using the WINEPR package.
Cyclic voltammetric measurements were performed using a CHI 660 A
(USA) electrochemical instrument. A conventional three-electrode cell assem-
bly was used. Saturated Ag/AgCl as reference and Pt as auxiliary electrodes
were used for this purpose. Zeolite-modified graphite electrode was used as
working electrode. In the preparation of this electrode, powder samples of ze-
olite and graphite were mixed together by gentle grinding in a weight ratio
of 60 : 40. It was then made into a pellet (diameter = 10 mm and thick-
ness ∼ 1 mm) and fixed to a polished silver wire. Silver paint was diluted
with iso-amyl acetate and a small drop of it was used for making electri-
cal contact between the pellet and wire. It was dried under an IR lamp for
30 min. An adhesive, Araldite gum was applied to hold the pellet and the
wire covering the silver paint. The modified electrode was kept for 16 h for
drying before it was used in the experiment. 0.1 M KNO solutions were
3
used as electrolyte. Prior to the CV experiments, dissolved oxygen in the
electrolyte solution was removed by purging with argon for 30 min. The
oC fV 5 0s cma Vn s/ sw. ere made in the potential range +1 to −1 V with a scan rate
2.2 EPR-spin trapping and sample preparation
−9
In this technique, the lifetime of radicals is prolonged (from ∼ 10 s to
•
∼
30 s, for HO radicals, for example) by forming spin adducts with spin
traps such as nitrones or nitroso compounds. We have used 5,5-dimethyl-1-
pyrroline-N-oxide (DMPO, Aldrich Co.) as the spin trap. In a typical sample
preparation procedure, to DMPO solution in water (0.5 mmol; 0.2 ml), 0.5 ml
of solvent (DMSO, CH CN or C H OH) was added and the pH was ad-
3
2
5
justed to 7.4 using 0.1 M phosphate buffer (K HPO + KH PO ). Separately,
2
4
2
4
TS-1 (8 mg) was treated with dilute H O (30% aqueous H O was diluted
2
2
2
2
by 10 times; 0.5 ml), warmed for 5 min at 333 K and then centrifused. The
spin trap solution prepared as above at 298 K was added to the centrifugate.
It was immediately taken in an EPR quartz flat cell and its EPR measured
at 298 K.
2.3 Reaction procedures and product analysis
Epoxidation of cyclohexene. 0.1 g of titanosilicate (TS-1, Ti-MCM-41, Ti-
SBA-15 or ETS-10), 0.82 g of cyclohexene, 5 ml of solvent were taken in
a glass flask fitted with a water-cooled condenser and placed in a constant
temperature (333 K) oil bath. H O ([cyclohexene]/[H O ] = 3 mol/mol) was
2
2
2
2

9
08
V. N. Shetti et al.
added drop-wise over a period of 10 min and then the reaction was carried out
for 4 h. The products were analyzed by gas chromatography (CHROMPACK
CP9001; 50 m-long × 0.32 mm-i.d. × 0.3 µm thick Hewlett-Packard fused
silica capillary column) and identified by GC-MS (Shimadzu QP-5000;
3
0 m-long × 0.25 mm-i.d. × 0.25 mm-thick capillary DB-1 column).
Epoxidation of styrene. In a 100 ml Teflon-lined steel autoclave, 0.1 g of
TS-1, 10 g of solvent, 8.63 mmol of styrene and 17.2 mmol of H O (aqueous)
2
2
were taken and the reaction was conducted at 333 K in a rotating hydrother-
mal synthesis reactor (rotation speed = 30 rpm) for 8 h. The products were
analyzed by gas chromatography/mass spectrometry.
Certain oxidations were conducted in the presence of radical scavengers
such as mannitol, catechol, thiourea, sodium acetate, sodium salicylate,
p-benzoquinone, hydroquinone and sodium formate. In those experiments
a known quantity of the radical scavenger (0–0.018 mmol) was added to the
reaction mixture and the reaction conducted in a similar manner.
3. Results and discussion
4
+
3
.1 Structure and redox behavior of Ti ions
The titanosilicate samples, TS-1, Ti-MCM-41 and Ti-SBA-15, show an intense
UV band at 200–230 nm corresponding to a charge transfer transition from
2−
4+
O
to Ti . This band was attributed [1–3] to dispersed-Ti ions in a psuedo-
tetrahedral coordination. Spectral deconvolution studies revealed that this band
comprises of two bands with maximum at 206 and 228 nm corresponding to
tetrapodal Ti species, Ti(OSi) and tripodal Ti species, Ti(OSi) (OH), respec-
4
3
tively. Intensity of the 206 nm band was more in the case of TS-1. Ti-MCM-41
and Ti-SBA-15 contained mainly the band at 228 nm corresponding to the
tripodal Ti. Similar conclusions on the presence of two different tetracoordi-
nated Ti species were drawn earlier [26] from the EPR investigations. The two
4+
Ti species differed in their reducibility in H , with the tripodal Ti being more
easily reducible than the tetrapodal Ti ions [26].
2
4+
Cyclic voltammetry (CV) is a highly sensitive technique for detect-
ing changes in the oxidation state and/or redox potential of ions. TS-1
4
+
−
3+
showed two reduction peaks, A and B (corresponding to Ti + e → Ti )
at −0.51 and −0.27 V (Fig. 1a), with the former being more intense than
4+
the latter redox peak. They represent Ti ions in two different struc-
tural environments and, hence, with different reduction potentials. Ti-MCM-
4
1 exhibited (Fig. 1b) four reduction peaks (A, B, C and D) at −0.52,
0.21, −0.22 and −0.05 V, respectively, with intensities A : B : C : D =
: 33 : 33 : 33, respectively. The reduction peak A by analogy to TS-1, arises
−
1
from a tetrapodal Ti while B arises from the tripodal Ti; C and D, originate,
perhaps, from penta- and hexa-coordinated defect open structures, respec-
tively. The less negative reduction potential of the tripodal Ti (reduction

Ti-Oxo Radicals and Product Selectivity in Olefin Oxidations . . .
909
Fig. 1. Cyclic voltammographs of TS-1, Ti-MCM-41 and TS-1 contacted with aqueous
. Peaks A–D correspond to different types of Ti species in TS-1 and Ti-MCM-41.
2 2
H O
peak B) compared to the tetrapodal species (reduction peak A) indicates
that the former could be reduced more easily. This observation is in agree-
4+
ment with our earlier EPR studies [26] wherein the tripodal Ti could be
reduced with dry H at 673 K while the tetrapodal Ti were reduced only
2
above 823 K.
Upon contacting TS-1 with H O , a new reduction peak was observed at
2
2
−
0.69 V (Fig. 1c). The original reduction peaks of TS-1 disappeared com-
−•
2−
pletely. The new peak is attributed to the reduction of Ti(O ) to Ti(O2 ).
2
Molecular oxygen and free superoxide ion are reduced only beyond −0.9 V.
Superoxide ions bound to Ti are, hence, more easily reduced than free super-
oxide ions.

9
10
V. N. Shetti et al.
Fig. 2. EPR spectra (at 80 K) of superoxo-Ti on titanosilicates contacted with aqueous
. Dotted line indicates the g component signal of superoxo species A and B.
H
2
O
2
z
3.2 Oxo-Ti species and reactivity
On contact with aqueous H O , TS-1 generated different types of Ti-oxo
2
2
species (hydroperoxo, peroxo- and superoxo-Ti) [20, 22–24]. These oxo
species showed a characteristic charge transfer band in the DRUV-visible spec-
trum in the region 300–500 nm [22, 23]. While the peroxo- and hydroperoxo-
Ti were diamagnetic, the superoxo-Ti species were paramagnetic and exhibited
characteristic rhombic-type EPR spectrum. Two types of superoxo-Ti species
A and B were identified on TS-1 [20, 22–24]. They differed mainly in their gz
parameter; g and g values being identical (A-type: g = 2.0260, g = 2.0090
x
y
z
y
and g = 2.0023; B-type: g = 2.0235, g = 2.0090 and g = 2.0023). Their
x
z
y
x
concentration decreased in the order: A > B. Ti-MCM-41 and Ti-SBA-15 con-
tained mainly the B-type superoxo-Ti species (Fig. 2). By analogy, the A and
B-type superoxo species are attributed to originate from the tetrapodal Ti struc-
tures (showing the UV band at 206 nm) and the tripodal Ti structures (showing
the UV band at 228 nm), respectively.
In situ EPR studies provided evidence for the participation of these oxo-
Ti species in epoxidation reactions (Fig. 3). In allyl alcohol epoxidation, the
oxo-Ti species (A-type) generated from the tetrapodal Ti were consumed faster
than those (B-type) generated from the tripodal Ti (Fig. 3a). Participation of
A and B-type species was also observed in the epoxidations of cyclohexene and
styrene (Fig. 3b,c). In the latter two cases, organic C-centered radicals (mul-
nd
tiplet pattern indicated by asterisk; see the 2 derivative plots; g = 2.0022)
were also detected. Such organic radicals were not observed in allyl alcohol

Ti-Oxo Radicals and Product Selectivity in Olefin Oxidations . . .
911
Fig. 3. Reactivity of Ti-oxo species on TS-1 + H
2 2
O system in substrate oxidation. Re-
actions were conducted at 333 K for a known period of time (0–25 min) and spectra
recorded at 80 K. Asterisk (*) indicates signals due to substrate-based radicals. These
nd
could be seen more clearly in the 2 derivative plots.
epoxidations (Fig. 3a). The rate of consumption of the A and B-type species
varied with the type of substrate. In the oxidation of cyclohexene, the A-type
2
species were consumed 4.6×10 times faster (estimated from the EPR signal
intensity as a function of time) than the B-type species. In the case of styrene,

9
12
V. N. Shetti et al.
a
Table 1. Influence of Ti structure: oxidation of cyclohexene with aqueous H
2
O
2
.
Catalyst
Cyclohexene
conv. mol%
TOF
Product selectivity (mol%)
(Epoxide + diol)
selectivity, %
(h )b Epoxide Diol
−1
Allylic
(-ol + -one)
TS-1
18.6
32.4
52.5
12.3
10
17
25
1
55.6
12.9
12.0
8.5
28.7
60.3
62.0
64.6
15.7
26.8
26.0
26.9
84.3
73.2
74.0
73.1
Ti-MCM-41
Ti-SBA-15
ETS-10
a
Reaction conditions: catalyst, 0.1 g; cyclohexene, 0.82 g (10 mmol); H
2
O
2
(30% aque-
b
ous), 3.33 mmol; CH
3
CN, 5 ml; temperature, 333 K; reaction time, 4 h. Turnover fre-
−1
quency (TOF, h ) = moles of cyclohexene converted (experimental) per mole of Ti per
hour.
2
the A species were consumed 0.63 ×10 times faster than B. In allyl alco-
hol epoxidation, the A type species were consumed twice faster. The higher
reactivity of the A type oxo-Ti species can be correlated with the reversible
reduction–oxidation behavior of the corresponding Ti ions (Fig. 1). In cyclic
voltammographs, a reversible redox couple was observed only for the A-type
species. The B-type species showed only a reduction peak, the corresponding
oxidation peak could not be detected. Hence, the facile reduction and oxidation
behavior of the tetrapodal-Ti species, A type, is responsible for their supe-
rior activity. The irreversible redox behavior of the tripodal Ti leads to weak
activity.
3
.3 Catalytic activity and selectivity
Oxidation of olefins containing allylic hydrogens can give rise to (1) epoxides,
2) allylic oxidation products (alcohol or ketone) or (3) cleavage of the double
(
bond. In addition, the epoxide can undergo further acid-catalyzed isomeriza-
tion (for example, styrene oxide to phenyl acetaldehyde) or hydration to diol
(ethylene oxide to ethylene glycol).
3.3.1 Oxidation of cyclohexene
The product distribution in cyclohexene oxidation (Scheme 1) over different
titanosilicate catalysts is listed in Table 1. Olefin conversion is higher over
the mesoporous material Ti-MCM-41 and Ti-SBA-15 due to the greater ease
of diffusion of reactant molecules in their pores. However, the epoxide is fa-
vored only over TS-1. The large amount of diols formed over the mesoporous
material and ETS-10 is due to their relatively greater acidity and consequent
hydration of the epoxide to the diols. In the absence of solvents, cyclohex-
ene conversion is significantly lower; -enol and -enone (from allylic oxidation)
are the major products (Table 2). There is no epoxide/diol formation. The

Ti-Oxo Radicals and Product Selectivity in Olefin Oxidations . . .
913
conversions are also higher in the presence of aprotic solvents. Selectivity to
epoxidation products (epoxide + diol) is however, higher in alcohol solvents.
A similar trend is observed in acetonitrile–methanol (Table 2).
Scheme 1. Oxidation of cyclohexene.
3.3.2 Oxidation of styrene
Styrene oxide, phenyl acetaldehyde, benzaldehyde and benzoic acid were the
products in styrene oxidation (Table 3). Secondary reactions of styrene oxide
also yielded the diol and methylated diol products. A significant influence
of solvent on the reactivity and product selectivity is noted in this case also
(Table 3).
3.4 Oxo radicals – EPR-spin trap studies
In view of the observation of Ti-oxo radicals on contact of TS-1 with H O2
2
the role of radicals, if any, in the oxidation mechanism of TS-1, was investi-
gated through EPR-spin trapping experiments with DMPO. Upon reaction with
•
•
−•
various O-centered radicals (like HO , HOO or O , generated through ho-
2
molytic O–O cleavage), DMPO forms paramagnetic adduct complexes, which
show characteristic EPR spectra differing significantly in their superhyperfine
coupling constants (aN and a ) (Fig. 4). The superoxo and hydroperoxo-
H
DMPO adducts (B) show similar EPR spectral patterns. The hydroxo-DMPO
adduct (A), on the other hand, shows a completely different pattern (Fig. 4)
β
−•
(
A: DMPO–OH: g = 2.0063, a = 14.9 G, a = 14.9 G. B: DMPO–OOH/O
:
N
H
2
β
γ
g = 2.0063; a = 14.2 G, a = 11.2 G; a = 1.25 G). The EPR spectrum of
N
H
H
DMPO on contact with (H O + TS-1) can, hence, be used to differentiate the
2
2
oxo radicals formed in the H O –TS-1 system.
2
2
DMPO alone did not exhibit any EPR spectrum; radicals were also not de-
tected in the absence of TS-1; the self-decomposition of H O into O-centered
2
2
radical species was negligible in the absence of TS-1. Contact of H O with
2
•
2
−•
TS-1 gave rise to the EPR spectrum of DMPO typical of O2 /HOO radical

9
14
V. N. Shetti et al.
adducts (compare Figs. 4 and 5). No radicals were observed in ethanol. The
−•
•
concentration of O2 /HOO radicals (estimated using a weak-pitch standard
supplied by Bruker) decreased in the following order: water–DMSO (3.3×
1
3
13
13
1
>
0 spins) > water–CH CN (1.2×10 spins) > water alone (1.0×10 spins)
3
11
water–C H OH (7.5×10 spins). Solvents influence the identity of the
−• • •
2 2 2
2
5
radicals formed from H O (O /HOO vs. HO , for example). Keeping in
mind that higher epoxidation selectivities were observed in alcoholic solvents
Table 2), wherein the concentration of such O-centered radicals was minimal,
an influence of such radicals on the formation of allylic oxidation products
enol and enone) in cyclohexene oxidation reactions may be inferred.
(
(

Ti-Oxo Radicals and Product Selectivity in Olefin Oxidations . . .
915

9
16
V. N. Shetti et al.
Fig. 4. Reaction of DMPO with O-centered radicals and the EPR spectra of the corres-
ponding radical adducts.
Fig. 5. EPR spectra (at 298 K; normalized intensity) of DMPO-oxygen centered radical
adducts in TS-1 + H
2 2
O system – influence of solvents.

Ti-Oxo Radicals and Product Selectivity in Olefin Oxidations . . .
917
Fig. 6. EPR spectra (at 298 K) of (TS-1 + H
and DMPO + hydroquinone.
2 2
O ) contacted with DMPO, hydroquinone
3.5 Reactions in presence of radical scavengers
Radical-mediated reactions are suppressed in the presence of radical scav-
engers. We have carried out the epoxidation tests in the presence of various
radical scavengers. In the presence of a spin trap (DMPO), characteristic sig-
−•
•
nals due to O2 /HOO radical adducts are seen (Fig. 6). When a radical
scavenger, hydroquinone, was added to this reaction mixture, the EPR pattern
of DMPO-oxo radical was not seen. Instead, the quintet line pattern (indi-
cated by asterisks; Fig. 6) due to the formation of semiquinone radicals was
detected; the hydroquinone had preferentially reacted with the O-centered rad-
icals and formed the semiquinone. When both hydroquinone and DMPO were
added simultaneously to the reaction slurry, hydroquinone reacted faster and
quenched the radicals even before DMPO, the spin trap, could form the spin
adducts. In this case, only the signals due to the semiquinone were observed.
Table 4 presents the influence of hydroquinone, on product selectivities in
the oxidation of cyclohexene and styrene over TS-1. In the absence of hy-
droquinone, oxidation occurs at both olefinic and allylic carbons (Scheme 1).
Upon hydroquinone addition, the selectivity for epoxide increased to 100%.
Allylic products were suppressed indicating that radical intermediates are prob-
ably involved in their formation. In the case of diols, poisoning of the acid
sites responsible for the hydration of the epoxide to diols by hydroquinone is

9
18
V. N. Shetti et al.
Table 4. Influence of hydroquinone addition on the epoxidation of cyclohexene and
styrene over TS-1.
a
Epoxidation of cyclohexene
Hydroquinone Cyclohexene
Product selectivity (mol%)
(mmol)
conv. mol% Epoxide
Diol
Allylic oxidation (-ol + -one)
Nil
18.6
15.3
55.6
100
28.7
0.0
15.7
0.0
0.018
b
Epoxidation of styrene
Product selectivity
Hydroquinone
Styrene
(mmol)
conversion
Styrene
oxide
Phenyl
acetaldehyde
Benzaldehyde
Benzolic
acid
(mol%)
0
16.4
11.5
7.9
37.8
45.5
59.0
62.4
12.2
12.9
12.8
13.2
49.3
39.8
27.2
23.3
0.7
1.8
1.0
1.1
0.009
0.018
0.027
8.0
a
Reaction conditions for epoxidation of cyclohexene: catalyst, 100 mg; cyclohexene,
0 mmol; HQ, 0 or 0.018 mmol; Hb (30% aqueous), 3.33 mmol; CH CN, 5 ml; tem-
1
2
O
2
3
perature, 333 K; reaction time, 4 h. Reaction conditions for epoxidation of styrene: TS-1,
00 mg; styrene, 8.63 mmol; H (30%, aqueous), 17.2 mmol; CH CN, 10 g; tempera-
1
2
O
2
3
ture, 333 K; reaction time, 8 h. Reaction was carried out in a rotating steel autoclave (Hiro
Company, Japan model KH 02; 30 rpm).
probably a reason for their absence in the products. A similar trend is also
seen in the oxidation of styrene. Table 5 shows the influence of various radical
scavengers on the product selectivity in cyclohexene oxidation. The selectivity
for the epoxidation products increased with different radical scavengers in the
order: mannitol < catechol < thiourea < sodium acetate < sodium salicylate
<
para-benzoquinone < sodium formate < hydroquinone.
4. Conclusions
Various Ti-oxo radical species are generated on reaction of H O with titanosil-
2
2
icate molecular sieves. The role of such radicals in the oxidation mechan-
ism of TS-1 has been investigated using EPR-spin traps, like 5,5-dimethyl-
1
-pyrroline-N-oxide (DMPO) and radical scavengers like hydroquinone. The
−•
EPR spectra of DMPO on contact with H O -TS-1 indicate that O
and
2
2
2
•
OOH radicals are predominant. Their concentrations vary with the solvent.
The oxo radicals influence product selectivity in olefin epoxidation reactions.
Addition of radical scavengers, like hydroquinone, suppresses the concentra-
tion of such radicals, and significantly enhances the selectivity for the epoxide

Ti-Oxo Radicals and Product Selectivity in Olefin Oxidations . . .
919
Table 5. Oxidation of cyclohexene in the presence of radical scavengers.
Radical scavenger Cyclohexene
Product selectivity, %
Epoxide + Diol
selectivity, %
conversion, % Epoxide Diol -enol -enone
None
Mannitol
Catechol
Thiourea
Sodium acetate
Sodium salicylate
p-Benzoquinone
Hydroquinone
Sodium formate
18.6
14.1
14.4
9.9
26.4
29.1
11.4
15.3
26.1
55.6 28.7
57.9 17.2 11.5
62.5 20.1 11.0
43.9 42.3
73.1 14.8
53.1 38.0
5.7
10.0
13.4
6.4
7.1
9.8
6.0
3.5
0
84.1
75.1
82.6
86.2
87.9
91.1
84.5
100
6.7
2.3
2.9
12.0
0
84.5
100
94.3
0
0
3.8
0.3
1.6
98.1
Reaction conditions: catalyst (TS-1; 100 mg), cyclohexene (0.82 g), aq. H
.268 g), acetonitrile (5 ml), radical scavenger (0.018 mmol in 0.1 ml water), temperature
333 K), reaction time (4 h); cyclohexene/H
= 3 (mol/mol).
2 2
O (30%,
0
(
2 2
O
at the expense of the allylic oxidation products. A novel method to control
product selectivity in oxidations over titanosilicate molecular sieves is demon-
strated.
Acknowledgement
One of the authors (V.N. Shetti) acknowledges the Council of Scientific and
Industrial Research (CSIR), New Delhi for the award of a Senior Research
Fellowship.
References
1
2
3
4
. P. Ratnasamy, D. Srinivas, and H. Knözinger, Adv. Catal. 48 (2004) 1.
. G. N. Vayssilov, Catal. Rev.-Sci. Eng. 39 (1997) 209.
. B. Notari, Adv. Catal. 41 (1996) 253.
. C. Lamberti, S. Bordiga, D. Arduino, A. Zecchina, F. Geobaldo, G. Span o´ , F. Genoni,
G. Petrini, A. Carati, F. Villain, and G. Vlaic, J. Phys. Chem. B 102 (1998) 6382.
. V. Bolis, S. Bordiga, C. Lamberti, A. Zecchina, A. Carati, F. Rivetti, G. Spanò, and
G. Petrini, Langmuir 15 (1999) 5753.
5
6
. D. Gleeson, G. Sankar, C. R. A. Catlow, J. M. Thomas, G. Spanò, S. Bordiga,
A. Zecchina, and C. Lamberti, Phys. Chem. Chem. Phys. 2 (2000) 4812.
. J. M. Thomas and G. Sankar, Acc. Chem. Res. 34 (2001) 571.
. S. Bordiga, A. Damin, F. Bonino, G. Ricchiardi, A. Zecchina, R. Tagliapietra, and
C. Lamberti, Phys. Chem. Chem. Phys. 5 (2003) 4390.
7
8
9
. G. Tozzola, M. A. Mantegazza, G. Ranghino, G. Petrini, S. Bordiga, G. Ricchiardi,
C. Lamberti, R. Zulian, and A. Zecchina, J. Catal. 179 (1998) 64.
1
0. G. Sankar, J. M. Thomas, C. R. A. Catlow, C. M. Barker, D. Gleeson, and N. Kalt-
soyannis, J. Phys. Chem. B 105 (2001) 9028.

9
20
V. N. Shetti et al.
1
1
1. C. M. Barker, D. Gleeson, N. Kaltsoyannis, C. R. A. Catlow, G. Sankar, and J. M.
Thomas, Phys. Chem. Chem. Phys. 4 (2002) 1228.
2. S. Bordiga, A. Damin, F. Bonino, G. Ricchiardi, C. Lamberti, and A. Zecchina,
Angew. Chem. Int. Ed. 41 (2002) 4734.
1
3. W. Lin and H. Frei, J. Am. Chem. Soc. 124 (2002) 9292.
1
4. F. Corà, M. Alfredsson, C. M. Barker, R. G. Bell, M. D. Foster, I. Saadoune, A. Sim-
perlar, and C. R. A. Catlow, J. Solid State Chem. 176 (2003) 496.
5. D. H. Wells, Jr., W. N. Delgass, and K. T. Thomson, J. Am. Chem. Soc. 126 (2004)
1
1
1
2956.
6. F. Bonino, A. Damin, G. Ricchiardi, M. Ricci, G. Spanò, R. D’Aloisio, A. Zecchina,
C. Lamberti, C. Prestipino, and S. Bordiga, J. Phys. Chem. B 108 (2004) 3573.
7. F. Geobaldo, S. Bordiga, A. Zecchina, E. Giamello, G. Leofanti, and G. Petrini,
Catal. Lett. 16 (1992) 109.
1
1
8. A. Tuel and Y. Ben Taârit, Appl. Catal. A: General 110 (1994) 137.
9. Q. Zhao, X.-H. Bao, Y. Wang, L.-W. Lin, G. Li, X.-W. Guo, and X.-S. Wang, J. Mol.
Catal. A: Chem. 157 (2000) 265.
2
2
0. K. Chaudhari, D. Srinivas, and P. Ratnasamy, J. Catal. 203 (2001) 25.
1. L. Bonoldi, C. Busetto, A. Congiu, G. Marra, G. Ranghino, M. Salvalaggio, G. Span o´ ,
and E. Giamello, Spectrochim. Acta A 58 (2002) 1143.
2
2. D. Srinivas, P. Manikandan, S. C. Laha, R. Kumar, and P. Ratnasamy, J. Catal. 217
(
2003) 160.
23. V. N. Shetti, P. Manikandan, D. Srinivas, and P. Ratnasamy, J. Catal. 216 (2003) 461.
24. V. N. Shetti, D. Srinivas, and P. Ratnasamy, J. Mol. Catal. A: Chem. 210 (2004) 171.
25. P. Wu, T. Tatsumi, T. Komatsu, and T. Yashima, Chem. Mater. 14 (2002) 1657.
26. R. Bal, K. Chaudhari, D. Srinivas, S. Sivasanker, and P. Ratnasamy, J. Mol. Catal.
A: Chem. 162 (2000) 199.