Metal silicates by a molecular route as catalysts for epoxidation of alkenes with
tert-butyl hydroperoxide
David Juwiler,a Jochanan Blumb and Ronny Neumann*a†
a Casali Institute of Applied Chemistry, Graduate School of Applied Science, The Hebrew University of Jerusalem, Jerusalem,
Israel 91904
b Department of Organic Chemistry, The Hebrew University of Jerusalem, Jerusalem, Israel 91904
Table 1 Epoxidation of cyclooctene catalysed by metal silicatesa
Macroporous, site isolated metal silicates are synthesized by
a molecular route; the molybdenum silicate is especially
Conversion (mol%)
(Oxidant 30% H2O2)
Conversion (mol%)
(Oxidant 6 m TBHP)
active for the selective epoxidation of alkenes with tert-butyl
hydroperoxide.
Metal silicate
TiO2–4SiO2
VO2.5–3SiO2
MoO3–4SiO2
WO3–4SiO2
0
0
5
5
40
90
7
The epoxidation of alkenes with ‘redox molecular sieves’ has
been studied intensively during the last decade.1 The titanium-
substituted silicalite of the MFI structure, TS-1, has been the
flagship of this effort. The small pored TS-1 is remarkably
effective for epoxidation of small linear olefins due to the
isomorphous distribution of titanium in the silicalite framework
leading to isolated Ti–(OSi)4 active centers held responsible for
its unique activity. In the quest to develop similar catalysts
viable for larger substrates, most attention has been directed
towards mesoporous catalysts such as Ti–MCM-41.2 A much
less studied approach has been to use amorphous metallosilicate
aerogels3 and xerogels4 prepared by sol–gel synthesis. The use
of the sol–gel technique has limited the possibility of obtaining
only M–O–Si connectivities (M = Ti, V, Mo and W) because
the rate of reaction of the metal alkoxides in the sol–gel method
is much faster than that of the silicon alkoxide.5 Reduced metal
site isolation considerably reduces the catalytic effectiveness of
amorphous metallosilicates. Recently, a non-aqueous molecular
route has been described for the preparation of metal silicates6
[eqn. (1)] (M = Ti, Zr, Hf, Cu) which by design leads to metal
site isolation within the silicate.
15
a
Reaction conditions: 1.5 mmol cyclooctene, 1.5 mmol metal silicate
(according to formula in table above), 0.1 ml 30% H2O2 + 1 ml acetone or
0.25 ml 6 m TBHP in n-decane, 60 °C, 12 h. Analysis was carried out by
GLC (HP-5890) using a dimethyl polysiloxane column (RTX-1, 30 m, 0.32
mm id, 0.25 mm coating).
substrate (Table 1). Using 30% hydrogen peroxide as oxidant,
reactivity was very low and the metal silicate was dissolved into
the homogeneous phase, presumably by aqueous hydrolysis. On
the other hand, using 6 m tert-butyl hydroperoxide in n-decane
as oxidant, a high yield of cyclooctene oxide was observed for
the molybdenum silicate, MoO3–4SiO2. The vanadium silicate
showed intermediate activity, whereas TiO2–4SiO2 and WO3–
4SiO2 were only slightly active. With both oxidants there was
no cyclooctene conversion without metal silicate. The stability
of MoO3–4SiO2 was tested by filtering the reaction mixture at
reaction temperature. There was no discernible loss in activity
when the heterogeneous silicate was recycled in three consec-
utive runs. No metal leaching into the organic phase was
measurable.¶
D
t
t
?
?
MCl4 + HOSi(OBu )3 — M[OSi(OBu )3]4 —
M[O(SiO)3]4 + CH2CMe2 + H2O
(1)
A more complete examination of the activity of MoO3–4SiO2
was carried out using different alkene substrates. First, the
effect of the TBHP:alkene ratio was studied in the epoxidation
of linear alkenes (Fig. 1). In general, doubling the amount of
The molecular route procedure is based on the synthesis of an
alkoxy intermediate which forms a metal silicate polymer by
thermal elimination of isobutene and water rather than by
hydrolysis and condensation as in the sol–gel technique. We
have now utilized this technique to prepare macroporous
titanium, vanadium, molybdenum and tungsten silicate. The
molybdenum analog was especially active towards the selective
epoxidation of bulky alkenes with tert-butyl hydroperoxide
(TBHP).
Substrate
1 / 1
1 / 2
The metal silicates described in this paper were prepared by
adapting the method reported for the titanium silicate.7 Thus, 40
mmol tri-tert-butoxysilanol7,8 in 100 ml toluene was reacted
with 10 mmol of metal chloride, TiCl4, V(O)Cl3, Mo(O)Cl4 and
W(O)Cl4, respectively, in the presence of 40 mmol Et3N at
room temperature for 2 h. The precipitates were filtered and
then calcinated for 12 h at ~ 250 °C to yield TiO2–4SiO2,
VO2.5–3SiO2, MoO3–4SiO2 and WO3–4SiO2, respectively. The
calcination temperature was chosen after thermogravimetric
measurements showed clean peaks for isobutene and water
elimination at this temperature.‡ N2—physisorption measure-
ments using the BET method showed that the metal silicates
were macroporous with large pore sizes of 130 ± 20 Å and
relatively low surface areas of 25 ± 5 m2 g21. The IR spectra
showed the expected peak9 at 950–960 cm21 attributable to the
Si–O stretching vibration polarized by the metal atom.§
The catalytic activity of the four metallosilicates was first
compared using the reactive but bulky cyclooctene as the model
0
20 40 60 80 100 0 20 40 60 80 100
Conversion (mol%)
Epoxide
Hydroxy-Ether
Allylic Oxidaton
Fig. 1 Epoxidation of linear alkenes catalysed by MoO3–4SiO2 as a function
of the oxidant : substrate ratio. Reaction conditions: 1.5 mmol substrate, 7.5
mmol (0.5 mol%) metal silicate, 1.5 mmol 6 m TBHP in n-decane (left
panel), 3.0 mmol 6 m TBHP in n-decane (right panel), 50 °C, for dec-1-ene
80 °C, 12 h. Analysis was carried out by GLC (HP-5890) using a dimethyl
polysiloxane column (RTX-1, 30 m, 0.32 mm id, 0.25 mm coating).
Unknown products were identified by GC–MSD (HP-GCD) under similar
conditions.
Chem. Commun., 1998
1123
Juwiler, David
