Epoxidation of α,β-unsaturated carbonyl compounds over various titanosilicates

Article abstract of DOI:10.1006/jcat.2001.3440

The epoxidation of a variety of electron-deficient α,β-unsaturated carbonyl compounds has been studied using dilute H₂O₂ and titanium-containing zeolites under liquid-phase conditions. The influence of the reaction medium and the structure of titanium-containing silicates (Ti-β(OH), Ti-Al-β(OH), Ti-β(F), TS-1, TS-2, Ti-MCM-22, and Ti-MCM-41) have been investigated. The weak basic acetonitrile solvent shows better activity and selectivity for epoxide than for other solvents used in this reaction over large-pore zeolite Ti-β. Among the various titanosilicates studied, aluminum-free Ti-β exhibits the best activity and H₂O₂ selectivity for cyclic α,β-unsaturated ketone, whereas Ti-β and TS-1 exhibit similar activities for open-chain α,β-unsaturated carbonyl compounds. The medium-pore TS-1, TS-2, and Ti-MCM-22 exhibit lower activities for the oxidation of cyclic ketones due to their diffusion limitation. For unsaturated ketones, epoxides are selectively formed, whereas unsaturated aldehydes mainly produce carboxylic acid. Branching at α- and/or β-carbon influences the reactivity of the carbonyl compounds considerably. However, the Ti-β/H₂O₂ catalytic system fails to oxidize substrates like α,β-unsaturated acids, α,β-unsaturated esters, and isophorone.

Full text of DOI:10.1006/jcat.2001.3440

Journal of Catalysis 205, 332–338 (2002)
doi:10.1006/jcat.2001.3440, available online at http://www.idealibrary.com on
Epoxidation of �,�-Unsaturated Carbonyl Compounds
over Various Titanosilicates
M. Sasidharan, Peng Wu, and Takashi Tatsumi1
Division of Material Science and Chemical Engineering, Graduate School of Engineering, Yokohama National
University, 79-5 Tokiwadai, Hodogaya-ku, Yokohama 240-8501, Japan
Received July 9, 2001; revised September 20, 2001; accepted September 25, 2001
oxidation rather than in epoxidation. Homogeneous cata-
a
variety of electron-deficient �,�- lysts such as Pd(OAc)2, with silyloxyalkyl peroxybenzoate
The epoxidation of
unsaturated carbonyl compounds has been studied using dilute as an oxidant (9), produce dimethyl-dioxirane (10) and
H O and titanium-containing zeolites under liquid-phase con-
2
2
cyclohexanone-dioxirane (11). However, the epoxidation
of �,�-unsaturated olefins has not been extensively ex-
plored over heterogeneous catalysts mainly due to the low
reactivity of electron-deficient C–C double bonds. Basic
catalysts such as KF–Al2O3 and KOH–Al2O3 have been
used for the epoxidation of �,�-unsaturated ketones with
TBHP (12). Recently Baiker and co-workers reported the
epoxidation of �-isophorone using TBHP and titania–silica
ditions. The influence of the reaction medium and the structure
of titanium-containing silicates (Ti-�(OH), Ti-Al-�(OH), Ti-�(F),
TS-1, TS-2, Ti-MCM-22, and Ti-MCM-41) have been investigated.
The weak basic acetonitrile solvent shows better activity and se-
lectivity for epoxide than for other solvents used in this reaction
over large-pore zeolite Ti-�. Among the various titanosilicates stud-
ied, aluminum-free Ti-� exhibits the best activity and H O selec-
2
2
tivity for cyclic �,�-unsaturated ketone, whereas Ti-� and TS-1
exhibit similar activities for open-chain �,�-unsaturated carbonyl mixed oxides (13). Here we report the study of epoxidation
compounds. The medium-pore TS-1, TS-2, and Ti-MCM-22 exhibit of �-�-unsaturated ketones over titanium silicates using di-
lower activities for the oxidation of cyclic ketones due to their dif- lute H O in the aqueous phase.
2
2
fusion limitation. For unsaturated ketones, epoxides are selectively
formed, whereas unsaturated aldehydes mainly produce carboxylic
acid. Branchingat�-and/or�-carboninfluencesthereactivityofthe
EXPERIMENTAL
carbonyl compounds considerably. However, the Ti-�/H O cata-
2
2
lytic system fails to oxidize substrates like �,�-unsaturated acids,
The hydrophobic Ti-�(OH) used in the present study was
synthesized in basic medium by modifying the literature
procedure (14). In a typical synthesis, a titanoperoxo com-
plex formed by the addition of Ti(OBu)4 and H2O2 was
added to TEAOH under vigorous stirring. Subsequently
the required amount of Aerosil-200 silica was added and
stirring followed with the addition of 3% dealuminated
seed crystals. The resulting gel (chemical composition:
SiO2 : 0.025 TiO2 : 0.337 H2O2 : 0.55 TEAOH : 6.62 H2O)
was transferred into a Teflon-lined autoclave, and crystal-
lization was carried out at 413 K for 5 days. The aluminum-
containing Ti–Al-�(OH) material was synthesized using
Al(NO3)3 � 9H2O from a gel composition of SiO2 : 0.0044
Al2O3 : 0.025 TiO2 : 0.337 H2O2 : 0.55 TEAOH : 16 H2O.
After crystallization, the product was centrifuged and
washed thoroughly to remove organic base. The Ti-�(F)
was synthesized using HF (47% aqueous) from a gel com-
position of SiO2 : 0.0125 TiO2 : 0.337 H2O2 : 0.56 TEAOH :
�
,�-unsaturated esters, and isophorone. �c 2002 Elsevier Science
INTRODUCTION
The high activity of the oxirane ring makes epoxides an
important group of industrial organic intermediates. Often
epoxidesformedinaninitialstepreactfurthertoprovidein-
dustrially important products such as surfactants, corrosion
protection agents, additives, etc. (1). The �,�-epoxy ketones
are important starting materials for making �-hydroxy ke-
tones and �-hydroxy ketones by reductive cleavage. The �-
hydroxy ketones and the closely related 1,3-diol groups are
found in prostaglandins, taxol (2), and �-hydroxy ketones
in biologically active antitumor agents and drugs (3, 4).
These electron-deficient olefinic compounds are nor-
mally epoxidized with alkaline H2O2, tert-butyl hydroper-
oxide (TBHP), and peracids (5–8). However, the use of
peracids results in enol–ether formation or Bayer–Villiger
0
.56HF : 7.4H2Oandcrystallizedat413Kfor5days. Theas-
synthesized catalysts were dried and calcined in O2 slowly
for 12 h at 793 K. The calcined Ti-� samples thus obtained
were characterized by XRD, FTIR, and UV–vis spectro-
scopiesandsurfaceareameasurements. TS-1andTS-2were
1
To whom correspondence should be addressed. E-mail: ttatsumi@
ynu.ac.jp.
332
0
�
021-9517/02 $35.00
c 2002 Elsevier Science
All rights reserved.

EPOXIDATION OF �,�-UNSATURATED CARBONYL COMPOUNDS
333
prepared according to the literature (15, 16). In a typical
O
O
O
synthesis of TS-1, 42 g of tetraethyl orthosilicate was hy-
drolyzed with 67.7 g of tetrapropylammonium hydroxide
under vigorous stirring for 1 h. Then 2.26 g of tetrabutyl or-
thotitanate dissolved in anhydrous isopropanol was added
dropwise with vigorous stirring. Finally 54 g of water was
added, and the clear solution was crystallized at 443 K for
Ti-beta / H O2
2
OH
OH
O
+
Acetonitrile
SCHEME 1
used as an internal standard. Progress in the reaction was
made by analyzing the products at various times with a cap-
illary gas chromatograph (Shimadzu 14A, OV-1 columns
with flame ionization detectors). The products were identi-
fied by a GC–MS splitting pattern and authentic samples.
3
0 h under static conditions. In the case of TS-2, tetrabuty-
lammonium hydroxide was used, and the crystallization was
carried out at 443 K for 72 h under static conditions.
The medium-pore Ti-MCM-22 was synthesized accord-
ing to the reported procedure (17) from fumed silica (Cab-
o-sil M7D), tetrabutyl orthotitanate (TBOT), boric acid,
and piperidine (PI). Piperidine was dissolved in deion-
ized water and divided into two portions, and the required
amount of TBOT or boric acid was added to each piperi-
dinesolutionundervigorousstirring. Further, silicawasalso
RESULTS AND DISCUSSION
Effect of Various Solvents
The choice of solvent has a crucial effect on the activ-
divided into two equal parts and added to the solutions con- ity and product selectivity for titanium-silicate catalyzed
taining Ti and B, respectively. The gels were stirred for 1 h liquid-phase reactions with dilute H2O2. The epoxidation
and then mixed together. The combined gel was stirred for of 2-cyclohexen-1-one with Ti-� (Si/Ti = 43) synthesized
another 1.5 h to obtain a gel with a molar gel composition in a hydroxide medium using seed crystals was studied
of SiO2 : 0.04 TiO2 : 0.67 B2O3 : 1.4 PI : 19 H2O. The crys- (Scheme 1) using solvents of different polarities and basic-
tallization of the resultant gel was carried out in a stirring ities. As shown in Table 1, the use of protic polar solvents
autoclave at 403 and 423 K each for 1 day and further at such as methanol, ethanol, and 2-propanol (entries 1–3)
4
43 K for 5 days. The solid product was filtered, washed with resulted in high activities of Ti-�, but the oxirane ring
deionized water, and dried at 323 K. The framework boron formed is readily solvolyzed to give the respective glycol
and extra-framework titanium were removed by acid treat- monoalkyl ether. It is known from the literature (19) that
ment with 2 M HNO3 at 373 K for 20 h at a solid-to-liquid the titanium site in Ti-� acts as a Brønsted acid site upon
ratio of 1 g to 20 ml. The samples were calcined at 803 K adsorption of the highly polar protic molecule. During the
to burn off the remaining organic species. The mesoporous reaction, the Brønsted acid sites thus formed can activate
Ti-MCM-41 was also synthesized according to the reported the epoxide product by protonation to give either glycol
procedure (18).
monomethyl ether or monoethyl ether. Furthermore the
The catalytic reaction was carried out in a two-necked fact that hydrophobic Ti-� preferentially adsorbs reactant
glass reactor fitted with a water condenser under N2 atmo- molecules rather than more polar solvent molecules ac-
sphere at 343 K under vigorous stirring. A typical reaction counts for the highest conversion in methanol. The ob-
involves 10 mmol of substrate, 10 mmol of H2O2, (Ti-�, served activity in 2-propanol, as opposed to that in either
Si/Ti = 43) 20 wt% of catalyst with respect to the substrate, methanol or ethanol, is in accordance with its lower polar-
and 10 ml of acetonitrile and cycloheptanone, which was ity. The activity in acetone (entry 4) is moderate with better
TABLE 1
a
Epoxidation of 2-Cyclohexen-1-one over a Ti-�/H O System
2
2
Product selectivities (%)
Conversion
(mol%)
H2O2
selectivity (%)
Entry
Solvent
Epoxide
Diol
Glycol ether
Others
1
2
3
4
5
6
7
8
Methanol
Ethanol
2-Propanol
Acetone
Acetonitrile
Benzene
Chlorobenzene
DMF
70.1
70.5
63.9
46.5
61.2
14.0
12.8
3.0
80.0
83.0
75.2
77.5
87.6
47.0
29.0
15.0
51.0
49.5
68.0
83.0
91.0
100
10.0
13.0
14.0
17.0
8.5
—
—
—
23.5
26.5
13.0
—
—
—
14.5
11.0
5.0
—
0.5
—
100
100
—
—
—
—
a
Reaction conditions: 10 mmol 2-cyclohexen-1-one, 10 mmol H2O2 (31 wt% aqueous), 20 wt% of the catalyst (Ti-�, Si/Ti = 43)
with respect to the substrate, temperature 343 K, and reaction time 5 h.

3
34
SASIDHARAN, WU, AND TATSUMI
TABLE 2
TS-1. Mesoporous Ti-MCM-41 shows a moderately high
Influence of Various Titanium Silicatesa
conversion of 39.5% with high epoxide selectivity. The pre-
vious results reveal that aluminum-free Ti-� is the best cata-
lyst considering its activity and epoxide selectivity. The rela-
tively low activities of TS-1, TS-2, and Ti-MCM-22 clearly
suggest that the cyclic reactant is less accessible to the Ti
site inside the medium-pore zeolites. This is supported by
findings that the activities of Ti-� and TS-1 are similar to
those of open-chain �ꢀ �-unsaturated carbonyl compounds
Product
selectivities (%)
H2O2
Conversion selectivity
Entry
Catalyst
(mol%)
(%)
Epoxide Diol Others
b
1
2
3
Ti-�(OH) (43)
Ti-Al-�(OH) (40)
Ti-�(F) (83)
TS-1 (38)
61.2
57.5
42.9
24.0
20.4
21.2
39.5
9.1
87.6
82.0
80.0
80.0
70.5
71.0
60.0
38.0
45.0
62.0
88.0
91.0
83.0 15.5
96.6
94.0
87.8 10.9
81.0 19.0
96.0
100
100
18.5
32.0
8.5
0.5
1.5
—
0.5
1.1
—
—
—
—
b
c
3.4
5.5
(
entries 8–11).
4
5
6
7
TS-2 (46)
Ti-MCM-22 (53)
Ti-MCM-41 (50)
Ti-� (43)
TS-1 (38)
Ti-� (43)
Epoxidation of Various �,�-Unsaturated Ketones
4.0
—
—
3.8 77.8
2.7 67.0
and Cyclohexene
d
8
9
0
1
d
8.8
36.8
35.4
The results of various cyclic and open-chain �ꢀ�-
unsaturatedketonesoveraTi-�/H O systemarepresented
e
1
1
2
2
e
TS-1 (38)
in Table 3. Entries 1–6 exhibit the epoxidation of cyclic
unsaturated ketones. 2-Cyclohexen-1-one is more active
a
Reaction conditions: 10 mmol 2-cyclohexen-1-one, 10 mmol H2O2
31 wt% aqueous), 20 wt% of the catalyst, 10 mL acetonitrile as solvent,
(
temperature 343 K, and reaction time 5 h.
TABLE 3
b
The catalyst was synthesized in hydroxide medium, and SiO2/
Epoxidation of Various Unsaturated Ketones over Ti-�a
Al2O3 = 200.
c
The catalyst was synthesized in fluoride medium.
The substrate is mesityl oxide.
The substrate is trans-pentenal for entries 10 and 11; others refer to
d
Product
selectivities (%)
e
Conversion
(mol%)
H2O2
carboxylic acid.
Entry Substrate
1
selectivity (%) Epoxide Diol Others
selectivity for epoxide. Interestingly, the use of weak ba-
sic solvent acetonitrile results in high selectivity toward the
epoxide (entry 5). We believe that the weak basic nature
of acetonitrile can effectively neutralize the weak Brønsted
acid sites generated in Ti-� as proposed in the literature
61.2
11.0
87.6
55.0
—
91.0
100
—
8.5
—
0.5
—
—
2
3
N.Ab
—
(20). Thus the highest activities were observed for methanol
and ethanol, while the highest selectivity for epoxide was
observed for acetonitrile. The use of benzene, chloroben-
zene, or DMF as solvents (entries 6–8) decreased the ac-
tivity of Ti-�. The H2O2 selectivity ranges between 75 and
4
5
N.Ab
31.5
—
—
—
—
—
—
84.0
100
8
5% for alcoholic and acetonitrile solvents. These results
clearly show that acetonitrile is the solvent of choice be-
cause it gives good selectivity not only for epoxide but also
for H2O2.
N.Ab
14.0
6
—
—
—
—
7
8
9
44.0
38.0
—
100
100
—
—
—
—
—
—
—
Influence of Various Titanium Silicates
9.1
Table 2 shows the effect of various zeolites over
2
�
-cyclohexen-1-one epoxidation. Ti-�(OH) and Ti–Al-
(OH) exhibit similar activities, but the epoxide selectiv-
N.Ab
ity is higher for the former than for the latter, which con-
tains more acidity due to the presence of aluminum. The
Ti-�(F) synthesized in fluoride medium shows a lower con-
version than other Ti-� zeolites, but the selectivity for epox-
ide is very high. The lower reactivity may be partly due to
a lower Si/Ti ratio. The efficiency of H2O2 ranges between
1
0
1
10.6
44.0
45.0
88.0
78.5
26.0
21.5
61.5
—
12.5c
1
a
Reactionconditions:10mmolsubstrate, 10mmolH2O2 (31wt%aque-
ous), 20 wt% of the catalyst (Ti-�, Si/Ti = 43) with respect to the substrate,
10 mL acetonitrile as solvent, temperature 343 K, and reaction time 5 h.
8
0 and 85% for Ti-� zeolites. The medium-pore TS-1 shows
b
No activity was observed.
Mixture of 2-cyclohexen-1-ol, 2-cyclohexen-1-one, and 2,3-epoxy
moderate activity and good H2O2 as well as epoxide selec-
c
tivity. TS-2 and Ti-MCM-22 are slightly less reactive than cyclohexen-1-one.

EPOXIDATION OF �,�-UNSATURATED CARBONYL COMPOUNDS
335
than the other substrates investigated. The presence of of four coplanar atoms are made possible, giving rise to
an electron-donating methyl group on the olefinic dou- electron-deficient C == C double bonds.
ble bond (entry 2) drastically reduces the activity. Based
Entry 11 shows the epoxidation of cyclohexene over Ti-�
on the electronic effects of substituents increased activity under similar conditions. The activity of cyclohexene is low
would be expected, which suggests that steric effect and consistent with earlier reports (22). Unlike 2-cyclohexen-1-
electron deficiency play important roles. Yamaguchi et al. one, where epoxidation dominates and epoxide is the sole
(21) reported that 3-methyl-2-cyclohexen-1-one oxidation, product, cyclohexene undergoes both epoxidation and al-
using H2O2 and basic hydrotalcites, requires more than 24 h lylic oxidation to give products like 2-cyclohexen-1-ol, 2-
to reach 90% conversion, whereas 2-cyclohexen-1-one cylohexen-1-one, and epoxy ketones. In the case of cyclo-
gives similar conversion in 3 h. Further, the ethoxy group hexene, the active site in Ti-� not only causes oxidation
at the � position (entry 3) of the C–C double bond gives no in different functional groups (such as C == C bond epoxi-
activity using the acetonitrile cosolvent. This is probably dation followed by nucleophilic ring opening of epoxide,
due to the combined effect of the ethoxy group, which not allylic oxidation to give 2-cyclohexen-1-ol) but also causes
only makes the molecule more electron deficient by its many parallel reactions, e.g., oxidation of 2-cyclohexen-1-ol
electron-withdrawing nature but also provides steric hin- to ketone and oxidation of 2-cyclohexen-1-one to the cor-
drance. In addition, isophorone (entry 4) also fails to give responding epoxy ketone. Some of these reaction products,
anyoxidizedproductovertheTi-�/H2O2 systemasopposed polyols in particular, may poison the active site as well. Thus
to that given over the mixed titania–silica/TBHP catalyst the slightly lower activity of cyclohexene may be due to the
system (13). The lack of reactivity of isophorone may be involvement of many parallel reactions rather than to the
due to the hindrance of the �-substituted methyl group and electron deficiency of cyclohexene.
gem-dimethyl groups at the C-5 carbon atom.
To improve the reactivity of these electron-deficient
Entries 5 and 6 show the epoxidation of 2-cyclopenten- olefins, epoxidation was carried out under triphase condi-
-one and 3-methyl-2-cyclopenten-1-one, respectively. The tions using water as reaction medium, and the results are
1
low reactivity of the five-membered ring ketones compared given in Table 4. It is important to mention that the two
to that of the six-membered ring ketones is in agreement different organic and aqueous phases were homogenized
with earlier findings (7–9, 21) and may be due to reso- using acetone before GC analysis. The epoxidation of
nance interactions between the C == C double bond and the �ꢀ�-unsaturated ketones using water as a reaction medium
electron-withdrawing carbonyl (C == O) groups. The copla- considerably increased the activity and the selectivity of
narity of the atoms of the C == C–C == O system is realiz- H2O2. In the presence of water, the primary epoxide prod-
able, leading to electron deficiency of the C == C double uct undergoes a nucleophilic ring opening of the oxirane
bond of 2-cyclopenten-1-one derivatives. In contrast, in 2- ring by water molecules to give a mixture of epoxide
cyclohexen-1-one, the conformations that make the atoms and diol. 2-Cycohexen-1-one and 3-methyl-2-cyclohexen-
of C == C–C == O system present on the same plane would be 1-one (entries 1 and 2) show increased activity and
unfavorable, and therefore the effective overlap of the four selectivity for diol compared to the biphase system which
p-orbitals cannot be attained. Thus the C == C double bond uses acetonitrile as cosolvent. Entry 3 shows the activity of
in 2-cyclohexen-1-one is similar to an isolated C == C double 3-ethoxy-2-cyclohexen-1-one, which undergoes cleavage of
the ethoxy group (–O–C2H5) by the Ti4 /H2O2 system to
+
bond.
The open-chain �ꢀ�-unsaturated ketones (entries 7–10) give 1,3-cyclohexanedione. Very low activity was observed
were less active than the cyclic ketones in the acetonitrile forisophorone(entry4), andtheproductsweremainlyC–H
solvent. The substituent in the � position of olefinic dou- oxidation (mixture of three products, possibly –CH3 oxi-
ble bond suppresses the activity more severely than that dation and allylic oxidation products such as 1,4-dione and
in the � position of the olefinic double bond; the substitu- allylic alcohol) rather than epoxidation products. Unlike
tion of the methyl group on 3-penten-2-one in the � position the six-membered ketone, the substrate 2-cyclopenten-
(entry 8) slightly decreases the conversion, but for that in 1-one exhibits a similar activity under both biphase and
the � position (entry 9) no conversion was observed. Thus, triphase conditions (entry 5). This may be due to the fact
the oxidation system Ti-�/H2O2 is not effective, whereas that cyclopentenone is a rather planar molecule unlike cy-
alkaline H2O2 readily oxidizes these kinds of electron- clohexenone, which has conformational flexibility, and the
deficient ketones (6). The terminal olefin, methyl vinyl expected epoxide and diol products are highly strained and
ketone (entry 10), also exhibits low conversion. The ob- unstable. Furthermore, as under the biphase condition, the
served H2O2 selectivity for open-chain compounds was also presence of a methyl substitutent in the � position (entry 6)
lower (38.0–45.0%) compared to that for cyclic ketones. makes the molecule unreactive due to steric hindrance.
The low reactivity of open-chain ketones is due to the ef-
The 3-Penten-2-one open-chain ketone (entry 7) shows
fective resonance interactions. Because of the flexibility of increased activity in water medium. A similar observation
the molecules, strong electronic interactions over an array was also made by Yamaguchi et al. (21) for open-chain

3
36
SASIDHARAN, WU, AND TATSUMI
TABLE 4
biphase conditions. The water used in the triphase system
not only acts as a dispersion medium but also assists in
driving away the products from the active reaction sites
to the channels, accounting for a higher conversion under
triphase than under biphase conditions using acetonitrile
as cosolvent. It is important to note that under triphase
conditions the reactivity of 3-methyl-2-cyclohenxen-1-one
Epoxidation of Various Unsaturated Ketones under
Triphase Conditionsa
Product
selectivities (%)
Conversion
(mol%)
H2O2
Entry Substrate
1
selectivity (%) Epoxide Diol Others
(
entry 2) is increased slightly more than that of the corre-
sponding open-chain ketone (entry 8), as it provides more
steric hindrance than the open-chain analogue, though
the Ti-� is a large-pore zeolite. Similarly, under triphase
conditions 2-cyclohexen-1-one (entry 1) shows a higher
conversion than the corresponding open-chain compound
3-penten-2-one (entry 7), mainly due to cyclohexenone’s
electronic effect and conformational flexibility, which are
not observed in 3-penten-2-one. Thus, the increased con-
version for open-chain �ꢀ�-unsaturated ketones under
triphase conditions is mainly due to the increase in the con-
tact area of the interface between water and organic phases
80.5
27.0
69.0
90.0
60.0
78.0
6.9
15.0
—
93.1
83.5
—
—
2
3
1.5
100b
4
5
6
4.0
6.0
43.0
—
—
33.0
—
11.0
67.0
—
89.0c
—
29.0
(
i.e., the hydrophobic Ti-� repulses water and adsorbs or-
N.Ad
ganic reactant).
—
Epoxidation of Other �,�-Unsaturated
7
8
59.6
48.9
65.5
71.0
71.2
72.5
28.8
25.0
—
Carbonyl Compounds
2.5
The reactivity and product distribution of different �ꢀ�-
unsaturated aldehydes (entry 1–4) over Ti-� are given in
Table 5. These types of compounds give carboxylic acid
as the product more through the oxidation of the –CHO
9
N.Ad
75.0
—
—
—
—
1
0
81.5
31.2
49.0
19.8e
a
Reaction condition: 10 mmol substrate, 10 mmol H2O2 (31 wt% aque-
ous), 20 wt% of the catalyst (Ti-�, Si/Ti = 43) with respect to the substrate,
TABLE 5
1
0 mL water as solvent, temperature 343 K, and reaction time 5 h.
b
The unstable cleavage product 3-hydroxy-2-cyclohexenone rear-
Epoxidation of Other �ꢀ�-Unsaturated Carbonyl
ranges to 1,3-cyclohexanedione.
a
c
Compounds over Ti-�
Mixture of allylic oxidation products and –CH3 oxidation.
No activity was observed.
Lower boiling product than methyl vinyl ketone, probably a cleavage
d
e
Product
selectivities (%)
product.
H2O2
Conversion selectivity
Carboxylic
Epoxide Diol acid
Entry Substrate
(mol%)
(%)
75.0
62.0
73.5
46.4
—
�ꢀ�-unsaturated ketones under triphase conditions using
quaternary ammonium salts. Entries 8 and 9 show the ef-
fects of methyl group substitution at the � and � posi-
tions, respectively. In a similar manner to the reaction un-
der biphase conditions, the presence of a methyl group at
the � position slightly decreased the activity, but the re-
action was completely inhibited by the methyl substituent
at the � position (entry 9). Methyl vinyl ketone (entry 10)
shows good activity and selectivity in the triphase medium.
Further, the open-chain ketones studied yield more epox-
ide than diol. This may be explained by assuming that
the expected epoxy compounds, which are solvated and
carried away from the active reaction sites by the water
molecules as previously reported (23), could be more hy-
drophilic. Thus, under triphase conditions the interaction
of organic reactants with the active sites of hydropho-
bic titanium silicates is more effective than that under
1
2
3
4
5
45.0
17.0
18.5
3.0
—
4.5
3.8
—
78.0
76.7
97.0
100
—
38.9
41.5
28.0
—
N.Ab
—
—
N.Ab
6
—
—
—
—
a
Reaction conditions: 10 mmol substrate, 10 mmol H2O2 (31 wt%
aqueous), 20 wt% of catalyst (Ti-�, Si/Ti = 43) with respect to the sub-
strate, 10 mL acetonitrile as solvent, temperature 343 K, and reaction
time 5 h.
b
No activity was observed.

EPOXIDATION OF �,�-UNSATURATED CARBONYL COMPOUNDS
337
functional group than through the epoxidation of the
olefinic bond. The increase in the length of the side chain
alkyl group has little effect over the activity, but the pres-
ence of a methyl group at the � position gives low con-
version and carboxylic acid as the sole product (entry 4).
The activity and product distribution also reveal that steric
hindrance plays a major role in controlling the selectivity
of the product. Olefinic compounds containing carboxylic
acid and carboxylate groups (entries 5 and 6) are unreactive
underthereactionconditions. TheTi-� catalystusedwasre-
activated at 673 K for 5 h under O2 flow and reused for epox-
idation of 2-cyclohexen-1-one under biphase conditions.
The activity of Ti-� was reduced by 34% after three suc-
cessive uses with reactivation. The decrease in activity may
1
00
90
80
70
60
Conversion
epoxide
Diol
50
40
30
H O sel.
2
2
2
0
0
0
1
1
2
3
4
5
6
Reaction time, hours
FIG. 2. Product distribution with time for epoxidation of 2-
be due to the partial removal of titanium from the frame- cyclohexen-1-one over Ti-�/H O /H O under triphase conditions using
2
2
2
work either by the action of H2O2 or by the bulky reaction water as reaction medium.
products such as polyols as reported by Carati et al. (25) and
Hutchings and co-workers (26, 27).
than that under biphase conditions when considering the
selectivity of diol at similar conversion levels under biphase
Reaction Pathways
(
the data shown Table 1 is utilized) and triphase condi-
The conversion of 2-cyclohexen-1-one over Ti-�/H2O2 tions. (Fig. 2). The selectivities of diol under biphase condi-
(
under biphase conditions) with reaction time and product tions are 10, 8.5, and 17% at the conversion level of 70.1%
selectivitiesisshowninFig. 1. Theconversionincreaseswith (methanol), 61.2% (acetonitrile), and 46.5% (acetone), re-
time, and the corresponding epoxide decreases from 97.8 to spectively. However, the corresponding diol selectivities
9
1.5% after 5 h. Simultaneously, the amount of diol formed under triphase conditions at similar conversion levels (esti-
by the nucleophilic substitution of the water molecule with mated from the curve in Fig. 2) are 30% (46.5% conv.), 58%
epoxide increases gradually with time at the expense of (61.2% conv.), and 75% (70.1% conv.). Thus, triphase con-
epoxide. Figure 2 depicts product selectivities and the epox- ditions provide a higher diol selectivity than biphase condi-
idation of 2-cyclohexen-1-one over Ti-�/H2O2/H2O under tions using acetonitrile as cosolvent at a similar conversion
triphase conditions. In the initial phase of the reaction, level at the expense of epoxide. The lower diol selectiv-
epoxide is mainly formed, but as time progresses the ini- ity for open-chain ketone under triphase conditions is the
tially formed epoxide is gradually converted into diol.
hydrophilic nature of the corresponding epoxide, which is
Unlike the biphase medium, the triphase medium con- kept intact by easy solvation (hydration) followed by diffu-
sists of an organic ketone phase, an aqueous phase con- sion into the bulk solvent before it interacts with the active
taining H2O2, and solid catalysts. The water used in the site. The diol formation under both biphase and triphase
triphase medium not only assists the hydrophobic Ti-� in conditions is assisted not only by the weak Brønsted acid
selective adsorption of the organic substrate but also helps sites of Ti-� but also by the titanium hydroperoxo complex
to drive away the oxidized products from the channels. The (24) generated from Ti-�/H2O2 as shown in Scheme 2. The
selectivity of epoxide under triphase conditions is lower observed conversion of unsaturated ketones and H2O2 se-
lectivities were also higher under triphase conditions than
1
00
under biphase conditions.
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
O
O
O
Ti-beta / H O2
Conversion
epoxide
Diol
2
Ti-hydroperoxo complex
OH
OH
O
H O
2
2 2
H O sel
H
OSi
O
O
H
SiO
SiO
Ti-beta
SiO
Ti
OSi
Ti
1
2
3
4
5
6
H O / H O
2
O
2
2
SiO
Reaction time, hours
H
OSi
Ti-hydroperoxo complex
FIG. 1. Product distribution with time for epoxidation of 2-
cyclohexen-1-one over Ti-�/acetonitrile.
SCHEME 2

3
38
SASIDHARAN, WU, AND TATSUMI
2. Samuelsson, B., Angew. Chem. Int. Ed. Engl. 22, 805 (1983).
CONCLUSIONS
3
. Tamura, Y., Takura, T., Haruta, J.-H., and Kita, Y., Tetrahedron Lett.
6, 3837 (1985).
2
The large-pore Ti-� can be used as an effective epoxida-
4
. (a) Tamura, Y., Annoura, H., Yamamoto, H., Kondo, H., Kita, Y.,
and Fujioka, H., Tetrahedron Lett. 28, 5709 (1987). (b) Murahashi,
S.-I., Naota, T., and Aanaoka, H., Chem. Lett. 1767 (1993) (and refs.
therein).
tion catalyst for �ꢀ�-unsaturated ketones using dilute H2O2
as oxidant. The use of acetonitrile as solvent resulted in the
best selectivity for epoxide, whereas the highest activity was
obtained for methanol. Ti-� samples are more active for
epoxidation of 2-cyclohexen-1-one than for medium-pore
zeolites such as TS-1, TS-2, and Ti-MCM-22. The presence
of an electron-donating methyl substituent on the C–C dou-
5. Fusion, R. C., Byers, D. J., Sperati, C. A., Foster, R. E., and Warfield,
P. F. , J. Org. Chem. 10, 69 (1945).
6
7
8
. House, H. O., and Ro, R. S., J. Am. Chem. Soc. 80, 2428 (1958).
. Wynberg, H., and Marsman, B., J. Org. Chem. 45, 158 (1980).
. Julian, P. L., Meyer, E. W., and Ryden, I., J. Am. Chem. Soc. 72, 367
ble bond decreases the activity by steric effect, whereas the
electron-withdrawing ethoxy substituent retards the reac-
tion completely by electronic effect under biphase condi-
tions. The observed activity of cyclohexenones is higher
than that of cyclopentenones probably due to the lack
of conformational flexibility coupled with a pronounced
(
1950).
9. Nagata, R., Matsuura, T., and Saito, I., Tetrahedron Lett. 25, 2691
1984).
0. Adam, W., Hajiarapoglou, L., and Nestler, B., Tetrahedron Lett. 31,
31 (1990).
1. Murray, R. M., Singh, M., and Jeyaraman, R., J. Am. Chem. Soc. 114,
346 (1992).
(
1
1
3
1
effect of electron deficiency in the planar ring. Further, 12. Yadav, V. K., and Kapoor, K. K., Tetrahedron Lett. 35, 9481 (1994).
1
1
3. Hutter, R., Mallat, T., and Baiker, A., J. Catal. 157, 665 (1995).
4. Blasco, T., Camblor, M. A., Corma, A., Esteve, P., Guil, J. M.,
Martinez, A., Perdigon-Melon, J. A., and Valencia, S., J. Phys. Chem.
B 102, 75 (1998).
the activity of these unsaturated ketones can be increased
considerably by carrying out the reaction under triphase
conditions (using water as reaction medium). Epoxide, in
addition to a small amount of diol, is the major product un-
der biphase conditions. However, under triphase conditions
cyclic ketones give diol as the major product, but acylic ke-
tones produce a mixture of comparable amounts of diol and
epoxide, probably due to the more hydrophilic nature of
epoxide. Unlike �ꢀ�-unsaturated ketones, �ꢀ�-unsaturated
aldehydes mainly give carboxylic acid, in addition to small
amounts of epoxide and diol. However, �ꢀ�-unsaturated
acid and �ꢀ�-unsaturated esters are inactive toward the Ti-
15. Thangaraj, A., Kumar, R., Mirajkar, S. P., and Ratnasamy, P., J. Catal.
130, 1 (1990).
1
1
6. Reddy, J. S., Kumar, R., and Ratnasamy, P., Appl. Catal. L1, 58 (1990).
7. Wu, P., Tatsumi, T., Komatsu, T., and Yashima, T., Chem. Lett. 774
(
2000).
1
1
8. Koyano, K. A., and Tatsumi, T., Microporous Mater. 10, 259 (1997).
9. (a) Bellussi, G., and Fattore, V., Stud. Surf. Sci. Catal. 69, 79 (1991);
(b) van der waal, J. C., Lin, P., Rigutto, M. S., and van Bekkum, H.,
Stud. Surf. Sci. Catal. 105, 1093 (1997).
2
2
0. Corma, A., Esteve, P., and Martinez, A., J. Catal. 161, 11 (1996).
1. Yamaguchi, K., Mori, K., Mizugaki, T., Ebitani, K., and Kaneda, K.,
J. Org. Chem. 65, 6897 (2000).
�
/H2O2 system.
2
2. Corma, A., Camblor, M. A., Esteve, P., Martinez, A., and Perez-
Pariente, J., J. Catal. 145, 151 (1994).
ACKNOWLEDGMENTS
M.S. is grateful to the Japan Society for the Promotion of Science for
financial assistance. This work was supported in part by a Grant-in-Aid
for Scientific Research from the Ministry of Education, Culture, Sports,
Science, and Technology.
2
2
3. Bhaumik, A., and Tatsumi, T., J. Catal. 176, 305 (1998).
4. Eellussi, G., Carati, A., Clerici, M. G., Maddinelli, G., and Millini, R.,
J. Catal. 133, 220 (1992).
25. Carati, A., Flego, C., Massara, E. P., Millini, R., Carluccio, L., Parker,
W. O., Jr., and Bellussi, G., Microporous Mesoporous Mater. 30, 137
(
1999).
REFERENCES
26. Davis, L. J., McMorn, P., Bethell, D., Bulman Page, P. C., King, F.,
Hancock, F. E., and Hutchings, G. J., J. Catal. 198, 319 (2001).
1
. “Peroxide-Chemie: Langkettige Alpha-Epoxide,” Chapters 1–4, 27. Davis, L. J., McMorn, P., Bethell, D., Bulman Page, P. C., King, F.,
pp. 1–29. H o¨ llriegelskreuth A0.1.4, 1981. Hancock, F. E., and Hutchings, G. J., J. Mol. Catal. A 165, 243 (2001).