Highly selective, recyclable epoxidation of allylic alcohols with hydrogen peroxide in water catalyzed by dinuclear peroxotungstate

Article abstract of DOI:10.1002/chem.200400352

The highly chemo-, regio-, and diastereoselective and stereospecific epoxidation of various allylic alcohols with only one equivalent of hydrogen peroxide in water can be efficiently catalyzed by the dinuclear peroxotungstate, K₂[{W(=O)(O₂)₂(H₂O)} ₂(μ-O)]· 2H₂O (1). The catalyst is easily recycled while maintaining its catalytic performance. The catalytic reaction mechanism including the exchange of the water ligand to form the tungsten-alcoholate species followed by the insertion of oxygen to the carbon-carbon double bond, and the regeneration of the dinuclear peroxotungstate with hydrogen peroxide is proposed. The reaction rate shows first-order dependence on the concentrations of allylic alcohol and dinuclear peroxotungstate and zero-order dependence on the concentration of hydrogen peroxide. These results, the kinetic data, the comparison of the catalytic rates with those for the stoichiometric reactions, and kinetic isotope effects indicate that the oxygen transfer from a dinuclear peroxotungstate to the double bond is the rate-limiting step for terminal allylic alcohols such as 2-propen-1-o1 (1a).

Full text of DOI:10.1002/chem.200400352

FULL PAPER
Highly Selective, Recyclable Epoxidation of Allylic Alcohols with Hydrogen
Peroxide in Water Catalyzed by Dinuclear Peroxotungstate
[
a]
[a, b]
[a, b]
Keigo Kamata, Kazuya Yamaguchi,
and Noritaka Mizuno*
Abstract: The highly chemo-, regio-,
and diastereoselective and stereospecif-
ic epoxidation of various allylic alco-
hols with only one equivalent of hydro-
gen peroxide in water can be efficiently
catalyzed by the dinuclear peroxotung-
state, K [{W(=O)(O ) (H O)} (m-O)]·
holate species followed by the insertion
of oxygen to the carbon–carbon double
bond, and the regeneration of the dinu-
clear peroxotungstate with hydrogen
peroxide is proposed. The reaction rate
shows first-order dependence on the
concentrations of allylic alcohol and di-
nuclear peroxotungstate and zero-
order dependence on the concentration
of hydrogen peroxide. These results,
the kinetic data, the comparison of the
catalytic rates with those for the stoi-
chiometric reactions, and kinetic iso-
tope effects indicate that the oxygen
transfer from a dinuclear peroxotung-
state to the double bond is the rate-
limiting step for terminal allylic alco-
hols such as 2-propen-1-ol (1a).
2
2
2
2
2
2
H O (I). The catalyst is easily recy-
2
cled while maintaining its catalytic per-
formance. The catalytic reaction mech-
anism including the exchange of the
water ligand to form the tungsten–alco-
Keywords: alcohols · allylic com-
pounds
· epoxidation · reaction
mechanisms · tungsten
Introduction
are an important example for the epoxidation of allylic alco-
hols while the stereoselective epoxidation needs optically
[12]
Oxidations of organic compounds are important in industry
active tartrate. Although many methods for the epoxida-
tion of allylic alcohols have been developed, catalytic pro-
cesses with expensive and environmentally-unacceptable ox-
idants such as peracids and hydroperoxides in explosive,
hazardous, and carcinogenic organic solvents are still widely
[1–3]
and synthetic chemistry.
Stoichiometric oxidants such as
dichromate, permanganate, and manganese dioxide are
[1–3]
often used for the transformations.
The stoichiometric
use and disposal of such oxidants are undesirable from eco-
nomical and environmental viewpoints. Therefore, much at-
tention has been paid to the use of transition-metal catalysts
to achieve the effective oxidation with molecular oxygen
[13]
used. In this context, the development of efficient catalyt-
ic processes by using hydrogen peroxide or molecular
oxygen as a green oxidant in non-explosive solvent, especial-
ly water, achieve the economical and environmental bene-
[4–7]
and hydrogen peroxide as oxidants.
Epoxidation of allylic alcohols is of great importance
[8,9]
[4–7,14–17]
fits, and remain challenging.
because the epoxides have been used as raw materials for
epoxy resins and building blocks for the synthesis of biologi-
cally important compounds including natural products,
Many tungsten-based catalysts have been reported to be
active for the epoxidation of allylic alcohols with hydrogen
[10,11]
[18–28]
peroxide.
Most of them need strict pH control with
and chiral carbons are formed by the epoxidation. Katsu-
ki–Sharpless systems with tert-butyl hydroperoxide (TBHP)
amines or buffers and/or use biphasic procedures because of
decomposition of both the allylic alcohols and the epox-
[18–28]
ides.
We reported very recently an efficient and simple
route for epoxidation of both internal and terminal olefins
with hydrogen peroxide catalyzed by the divacant lacunary
[
a] K. Kamata, Dr. K. Yamaguchi, Prof. N. Mizuno
Core Research for Evolutional Science and Technology (CREST)
Japan Science and Technology Agency
[29]
silicotungstate, [(n-C H ) N] [g-SiW O (H O) ].
During
4
9
4
4
10 34
2
2
4
-1-8 Honcho, Kawaguchi, Saitama, 332-0012 (Japan)
the course of our investigation of tungsten-catalyzed oxida-
tion, we found that the simple dinuclear peroxotungstate,
K [{W(=O)(O ) (H O)} (m-O)]·2H O (I, Figure 1), could act
E-mail: tmizuno@mail.ecc.u-tokyo.ac.jp
[
b] Dr. K. Yamaguchi, Prof. N. Mizuno
Department of Applied Chemistry, School of Engineering
The University of Tokyo
2
2
2
2
2
2
as an effective catalyst for the epoxidation of allylic alcohols
using hydrogen peroxide in water under mild reaction condi-
7
-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656 (Japan)
[30]
tions [Eq. (1)]. To our knowledge, the isolated I itself has
never been used for the catalytic epoxidation of allylic alco-
hols using hydrogen peroxide in water without additives al-
Fax : (+81)3-5841-7220
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
4728
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200400352
Chem. Eur. J. 2004, 10, 4728 – 4734

4
728 – 4734
oxidation of 1a (entry 1). The catalyst precursor of K WO4
2
showed low catalytic activity due to the non-productive de-
composition of hydrogen peroxide, and the efficiency of the
hydrogen peroxide utilization was low (entry 2). In the case
of H WO , the successive hydrolysis of the product of epoxy
2
4
alcohol to the corresponding triol 1d was dominant while
the conversion of 1a was >99% (entry 3). The selectivity to
1
b was increased by the addition of triethylamine to buffer
Figure 1. Molecular structure of the anion part of
O)} (m-O)]·2H O (I).
2 2 2
K [{W(=O)(O ) -
[18,19]
(
H
2
2
2
the acidity of the reaction medium (entries 4 and 5).
However, even in these cases
the decomposition of hydrogen
peroxide proceeded to some
extent due to the basicity of
triethylamine.
Table 2 shows the results of
though I is a previously known compound. In this paper, we
report the dinuclear peroxotungstate-catalyzed epoxidation
of allylic alcohols with hydrogen peroxide in water, as well
as the kinetic and mechanistic aspects of the present epoxi-
dation.
epoxidation of various allylic alcohols with 30% aqueous
hydrogen peroxide in the presence of I without use of the
buffer or biphasic system. The pH value of aqueous phase
was in the range of 4–5 and no hydrolysis, cleavage, rear-
rangement of oxirane ring, and allylic oxidation were ob-
served. The catalyst I was intrinsically stable in the pH
1
83
range of 2.5–7 as was confirmed by the W NMR and IR
spectra. The efficiency of the hydrogen peroxide utilization
was more than 90% in each case. Oxidation of simple pri-
mary allylic alcohols (1a–5a) proceeded almost quantita-
tively and chemoselectively to afford the epoxy alcohols
without formation of the corresponding aldehydes and car-
boxylic acids (entries 1, 2, and 5–7). Larger scale (100 mmol
scale) epoxidations of allylic alcohols (I:H O :substrate =
Results and Discussion
Catalytic epoxidation of allylic alcohols: Epoxidation of 1a
as a model substrate using 30% aqueous hydrogen peroxide
in various solvents was carried out in the presence of I.
Water was the most effective solvent for the present epoxi-
dation: Epoxy alcohol 1b was formed in a 95% yield with
7% efficiency of hydrogen peroxide utilization under the
conditions in Table 1. The use of non-polar toluene, ben-
zene, and dichloromethane solvents (organic/aqueous bipha-
sic system) gave 1b yields of 90, 80, and 63%, respectively.
Water-miscible polar acetonitrile (5% yield) and methanol
[
31]
2
2
ꢀ
1
9
1:5000:5000) showed turnover frequency (TOF) of 594 h
ꢀ
1
and turnover number (TON) of 4200 for 2a, 442 h and
ꢀ
1
3580 for 3a, and 167 h and 3378 for 5a [Eq. (2)]. These
values are higher than those reported to be active for the
tungsten-catalyzed epoxidation of internal allylic alcohols
with hydrogen peroxide: Na WO -NH CH PO H -[CH (n-
(
<1% yield) were poor solvents probably because of the
2
4
2
2
3
2
3
[32]
ꢀ1
[24]
strong coordination to the tungsten center.
C H ) N]HSO in toluene, 86 h (TOF) and 43 (TON);
8 17 3 4
ꢀ
1
The catalytic activity of I for the epoxidation of 1a was
compared with other tungsten compounds such as H WO
and K WO as shown in Table 1. The oxidation did not pro-
ceed in the absence of I (entry 6). Among tungsten catalysts
tested, I showed the highest yield of the corresponding ep-
oxide (1b): 96% conversion, 99% selectivity to 1b, and
[C H N(n-C H )] PW O
12 40
in 1,2-dichloroethane, 5 h ,
[(n-C H ) N] [PhPO {WO(O ) }] in 1,2-dichloro-
{WZn[M(H O)] (ZnW O ) }
Zn , Mn , Ru , Fe , etc.) in 1,2-dichloroethane, 167 h ,
5
[
5
22]
16 33
3
20;
2
4
4
9
4
2
[23]
3
2 2
ꢀ
1
qꢀ
ethane, 80 h , 40;
(M:
2
4
2
2
9
34
2
2
+
2+
3+
3+
ꢀ1
[27,28]
1000;
Na WO -phosphate buffer-b-d-fructopyranoside
2
ꢀ
4
1
[26]
in water, 0.4 h , 10. For the epoxidation of cis- and trans-
allylic alcohols (2a, 5a, and 6a) the configurations around
the C=C moieties were retained
9
7% efficiency of hydrogen peroxide utilization for the ep-
in the corresponding epoxy al-
[
a]
Table 1. Oxidation of 1a with hydrogen peroxide catalyzed by tungsten compounds.
cohols (entries 2, 7, and 8).
Such a stereospecific epoxida-
tion suggests that the free-radi-
cal intermediates are not in-
volved in the epoxidation.
The epoxidation of secondary
b,b-disubstituted allylic alcohol
(1,3-allylic strained alcohol) of
Entry
Catalyst
Yield [%]
2 2
H O efficiency
[%]
1
b
1c
1d
<1
<1
>99
7
1
2
3
4
5
6
K
2
[{W(=O)(O
2
)
2
(H
WO
WO
2
O)}
2
(m-O)] (I)
95
30
<1
65
54
1
97
K
H
2
4
<1
<1
<1
<1
32
>99
76
57
–
2
4
H
H
2
2
WO
WO
4
/Et
/Et
3
N (0.29m)
N (0.49m)
7
a proceeded diastereoselec-
4
3
1
tively to form mainly the threo-
epoxy alcohol (entry 9). Allylic
no catalyst
no reaction
[
3
a] Reaction conditions: 1a (5 mmol), catalyst (W: 200 mmol, 4 mol%), 30% aq. H
2
O
2
(5 mmol), water (6 mL),
1
alcohols
strain (8a and 9a) gav eerythro
2 2
O (mol) ” 100. rich epoxy alcohols (entries 10
without
1,3-allylic
05 K. Yields were determined by gas chromatograpy and H NMR with an internal standard technique, and
3
+
4+
based on 1a. Residual H
2
O
2
after the reaction was estimated by potential difference titration of Ce /Ce
(SO ·2H O)·H efficiency (%) = products (mol)/consumed H
(
0.1m of aqueous Ce(NH
4
)
4
4
)
4
2
2
O
2
Chem. Eur. J. 2004, 10, 4728 – 4734
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4729

N. Mizuno et al.
FULL PAPER
remote unfunctionalized 6,7-
double bond in the case of 6a.
The analogous formation of
metal–alcoholate species is sug-
gested at the oxygen transfer
step in the allylic alcohol epoxi-
and 12). In addition, epoxidation of allylic alcohol 10a, in
which 1,3- and 1,2-allylic strains compete with each other,
was more erythro selective (entry 13). 2-Cyclohexen-1-ol
gave the corresponding epoxy alcohol (>95% syn configu-
ration) in a 40% yield together with a 49% yield of 2-cyclo-
hexen-1-one after 12 h under the same conditions as those
in Table 2. It is noted that the epoxidation of 2-cyclohexen-
dation systems of VO(acac) /TBHP, Ti(OiPr) /TBHP,
2
4
qꢀ
{WZn[M(H O)] (ZnW O ) } /H O ,
and
H WO /
2
2
9
34
2
2
2
2
4
[
27,28,38]
H O .
By contrast, non-metal catalyzed mCPBA and
2
2
dimethyldioxirane systems prefer the epoxidation at the 6,7-
double bond of 6a. The selectivity has been explained by
the formation of the characteristic hydrogen bonding in the
[27,28]
oxygen transfer step.
In the present I-catalyzed system
1
-ol was highly stereoselective to give the corresponding
(see above), the epoxidation of 1,3-allylic strained alcohol
proceeded diastereoselectively to give the threo diaste-
reoisomer preferentially. The erythro selectivity was a little
higher in the case of the allylic alcohol without 1,3-allylic
epoxy alcohol with oxirane ring cis to hydroxyl group (syn
configuration), while large amounts of a,b-unsaturated ke-
tones were produced for the cyclic allylic alcohols. All these
results for the epoxidations of secondary allylic alcohols are
strain. These facts are similar to those of VO(acac) /TBHP
2
[
9,33–36]
similar to those for VO(acac) /TBHP,
{WZn-
and H WO /H O (metal–alcoholate binding mechanism),
2
2
4
2
2
qꢀ
[27,28]
[36]
[M(H O)] (ZnW O ) } /H O ,
and H WO /H O
2
sys-
and different from those of mCPBA and dimethyldioxirane
(hydrogen-bonding mechanism). Further, it is noted that the
threo/erythro ratio in the epoxidation of 10a catalyzed by I
2
2
9
34
2
2
2
2
4
2
tems in which formation of the metal–alcoholate species in
the oxygen transfer step is suggested, and different from
those with mCPBA,
urea-hydrogen peroxide
tive epoxidation of 6a took place at the electron-deficient
allylic 2,3-double bond to afford only 2,3-epoxy alcohol in
the high yield (entry 8).
No significant changes in the in situ IR spectra were ob-
served during the catalytic epoxidation of 1a by I with hy-
[
5,36]
[5,36]
dimethyldioxirane,
and TS-1/
(threo/erythro 34:66) lies between those for VO(acac) /
2
[5,36]
[28]
qꢀ
systems. Further, the regioselec-
TBHP (33:67)
and {WZn[M(H O)] (ZnW O ) } /H O
2 2 9 34 2 2
2
[28]
(45:55). The stereochemical data suggest that the dihedral
angle between the p plane of the double bond and the hy-
droxy group of the allylic alcohol in the transition-state ge-
ometry of the oxygen transfer step for I-catalyzed epoxida-
tion is 50–708C, in accord with the angle reported for the
1
83
[28]
drogen peroxide. In addition, the W NMR spectrum of
tungsten-catalyzed epoxidation.
the catalyst I after the epoxidation of 1a showed a signal at
On the basis of these results, we propose a possible cata-
lytic cycle for the present epoxidation (Scheme 1). First, the
water ligand of peroxotungstate I is exchanged by an allylic
alcohol to form the tungsten–alcoholate species II (step 1).
The deprotonation of the OꢀH bond of an allylic alcohol
ꢀ
704.5 ppm, which was observed for the as-synthesized I.
IR and UV/Vis spectra of recovered catalyst also suggest
the retention of the structure of I. These facts show that I is
stable under the reaction conditions. The first-order depen-
dence of the reaction rate on the concentration of I as
shown in Figure 3, supports the idea.
followed by the proton transfer to the peroxo ligand is in-
cluded in the step 1. The reaction rate for the epoxidation of
1a decreased with increase in the concentration of protons
(Figure S5), in accord with the presence of step 1 in
The products could easily be isolated by the simple ex-
traction by using dichloromethane or n-pentane after the ox-
idation since catalyst I was completely insoluble in these sol-
vents. Actually, no leaching of tungsten into the organic
phase was confirmed by inductively coupled plasma atomic
emission analysis (ICP-AES). Therefore, the aqueous phase
including the catalyst could be recovered without loss of the
tungsten species. It is notable that I can be reused without
loss of the catalytic activity, stereospecificity, and chemo-,
regio-, and diastereoselectivity for the epoxidation (en-
tries 3, 4, and 10).
[39]
Scheme 1.
Next, the epoxy alcohol and III are formed
(step 2). Finally, I is regenerated by the reaction of III with
1
hydrogen peroxide (step 3). The H NMR spectrum of the
reaction solution of ethanol (870 mm) catalyzed by I
(50 mm) in D O at 305 K after 1 h showed a intense reso-
2
nance of ethanol together with a set of signals at d 1.44 (t,
3
3
J=5.1 Hz) and d 2.21 (d, J=2.5 Hz) due to the methyl
3
protons, at d 5.35 (q, J=5.1 Hz) due to the methylene pro-
tons, and at d 9.79 (q, J=2.5 Hz) due to the formyl
proton. The signals at d 2.21 and 9.79 can be assigned to
acetaldehyde. The H NMR signals of the a-protons of tung-
3
[40]
1
Kinetics and mechanism: The present regioselective epoxi-
dation of allylic alcohol 6a shows that the allylic hydroxyl
function plays an important role in the epoxidation. The
need of the hydroxyl functionality is further demonstrated
by the complete lack of epoxidation reactivity of the ester
sten–alcoholate species have been reported in the range of d
[41,42]
4.8–5.8.
Therefore, the signal at d 5.35 can be assigned
to the methylene protons of the tungsten–ethoxide species
formed through the reaction of ethanol with I. The signal at
d 1.44 can also be assigned to the methyl protons of the
[37]
and ether derivatives of allyl acetate
and 3-methoxy-1-
1
3
propene. In this epoxidation, it is probable that the allylic
hydroxyl groups ligate to the tungsten center to form tung-
sten–alcoholate bond, which prefers the oxygen transfer to
the proximal 2,3-allylic double bond rather than to the
tungsten–ethoxide species. The C NMR spectrum showed
a resonance at d 87.9 due to the methylene carbon of the
[41,42]
tungsten–ethoxide species.
These results also suggests
the formation of II during the present epoxidation.
4730
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4728 – 4734

Epoxidation of Allylic Alcohols
4728 – 4734
Figure 3. Dependence of reaction rate on concentration of I: 2a (0.40m),
I (0.92–4.12 mm), D
values were determined from the reaction profiles at low conversions
<10%) of both 2a and hydrogen peroxide. Slope =1.03.
2 2 2 2 0
O/H O (0.25/0.75 mL), H O (0.56m), 303 K. R
(
Scheme 1.
Figure 4. Dependence of reaction rate on concentration of hydrogen per-
oxide: 2a (0.40m), I (2.22 mm), D O/H O (0.25/0.75 mL), H (0.34–
.78m), 303 K. R values were determined from the reaction profiles at
low conversions (< 10%) of both 2a and hydrogen peroxide. Slope =
0.01.
2
2
2 2
O
0
0
Figure 2. Dependence of reaction rate on concentration of 2a: 2a (0.11–
ꢀ
0
.81m), I (2.22 mm), D
2 2 2 2 0
O/H O (0.25/0.75 mL), H O (0.56m), 303 K. R
values were determined from the reaction profiles at low conversions
(
<10%) of both 2a and hydrogen peroxide. Slope = 1.02.
d ½allylic alcoholŠ
R ¼ ꢀ
0
dt
ð3Þ
1
1
0
¼
k ½catalyst IŠ ½allylic alcoholŠ ½H O Š
2
2
The kinetic studies for the epoxidation of an internal allyl-
ic alcohol 2a showed the reasonable first-order plots for the
loss of concentration of 2a (0.11–0.81m) (Figure 2). In addi-
tion, the first-order dependences of the reaction rate on the
concentration of catalyst I (0.92–4.12 mm) (Figure 3) and the
zero-order dependence on the concentration of hydrogen
peroxide (0.34–0.78m) (Figure 4) were observed. Kinetic
studies on the epoxidation of a terminal allylic alcohol of 1a
also show the first-order dependence of the reaction rate on
both concentrations of 1a (0.31–0.69m) and catalyst I (3.8–
The dependence of the reaction rate of 2a on the tempera-
ture (Arrhenius plots, 296–328 K) is shown in Figure 5. The
good linearity of the Arrhenius plots was observed to afford
ꢀ
1
the following activation parameters:
E
=55.9 kJmol ,
a
°
ꢀ1
°
ꢀ1 ꢀ1
lnA=18.1, DH
=53.4 kJmol , DS
=ꢀ137.4 Jmol K ,
305 K
305 K
°
ꢀ1
and DG3 =95.3 kJmol . The dependence of the reaction
05 K
rate of 1a on the temperature was also examined to give the
ꢀ
1
following activation parameters: E
=68.1 kJmol , lnA=
a
°
ꢀ1
°
ꢀ1 ꢀ1
11.4 mm), and the zero-order dependence on the concentra-
20.5, DH
=65.6 kJmol , DS
=ꢀ117.1 Jmol K , and
3
05 K
305 K
°
ꢀ1
tion of hydrogen peroxide (0.14–0.56m). Therefore, the reac-
DG3 =101.3 kJmol . The parameters for 2a was different
05 K
tion rate equation was expressed by following equation
from those for 1a while the activation energies (E
a
) are in
ꢀ
1
[
Eq. (3)]. R and k are reaction rate and rate constant, re-
the range of 49–86 kJmol reported for tungsten-catalyzed
0
[43,44]
spectively. This means that the transition state is composed
of one molecule of I and one of an allylic alcohol, but not of
hydrogen peroxide.
epoxidation.
When the regeneration rate (step 3) was estimated by
tracing the reaction of III with hydrogen peroxide by using
Chem. Eur. J. 2004, 10, 4728 – 4734
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4731

N. Mizuno et al.
FULL PAPER
Figure 5. Arrhenius plots for the epoxidation of 2a: 2a (0.40m),
I
(
2.22 mm), D
2
O/H
2
O (0.25/0.75 mL), H
2
O
2
(0.56m), 295–328 K. The ob-
Figure 6. Dependence of epoxidation rate of 1a on the content of D
(x = [D O]/([D O] + [H O])): 1a (0.46m), I (11.1 mm), D O + H
(1 mL), H (0.56m), 303 K. R
2
2
O
O
served rate constants (kobs) were calculated with the initial rates by using
Equation (3).
2
2
2
2
2
O
2
0
values were determined from the reac-
tion profiles at low conversions (<10%) of both 1a and hydrogen perox-
ide.
an in situ IR spectrometer, the regeneration was completed
within 1 min and the rate was larger than that for the cata-
lytic epoxidation reaction under the same conditions. The
zero-order dependence on the concentration of hydrogen
peroxide was observed. These results show that the regener-
ation of tungsten species III with hydrogen peroxide (step 3)
smoothly proceeds and is not a rate-limiting step.
The stoichiometric epoxidation of 2a (1.9 mmol, 0.40m)
with I (40 mmol) produced 166 mmol of 2b, showing that I
has 4 equivacti ve oxygen [Eq. (4)]. The rate of stoichiomet-
Conclusion
In summary, the dinuclear peroxotungstate I is found to be
an effective homogeneous catalyst for the epoxidation of al-
lylic alcohols in water with high efficiency of hydrogen per-
oxide utilization. Further, I can be reused with retention of
the high catalytic activity, stereospecificity and chemo-,
regio-, and diastereoselectivity for the epoxidation. The ki-
netic, spectroscopic, and mechanistic investigation show that
the dinuclear peroxotungstate-catalyzed allylic alcohol epox-
idation proceeds via the OꢀH bond deprotonation of allylic
alcohols followed by the proton transfer to peroxotungstate/
alcoholate formation mechanism.
Experimental Section
ꢀ
3
ꢀ1
ric epoxidation of 2a was 8.1”10 mmin and fairly agreed
with that of the catalytic epoxidation under the same condi-
tions (8.8”10 mmin ); this shows that step 3 is not a rate-
limiting step.
Instrumentation: GC analyses were performed on Shimadzu GC-14B
with a ionization detector equipped with a TC-WAX capillary column.
Mass spectra were determined on Shimadzu GCMS-QP2010 at an ioniza-
tion voltage of 70 eV. NMR spectra were recorded on JEOL JNM-EX-
ꢀ3
ꢀ1
Figure 6 shows the dependence of the epoxidation rate of
1
13
2
70. H and C NMR spectra were measured at 270 and 67.5 MHz, re-
1
a on the content of D O (x=[D O]/([D O] + [H O])). The
2 2 2 2
spectively, in CDCl or D O with TMS as an internal or external stan-
3
2
1
83
reaction rate was not affected by the presence of D O and
dard. W NMR spectra of I were measured at 11.2 MHz in D
2
O with
2
Na WO (2m D O solution) as an external standard. UV/Vis spectra
2
4
2
the proton of the OH group in 1a was isotopically in equi-
librium with the proton in water. These facts show that the
deprotonation of OꢀH bond of an allylic alcohol followed
were recorded on a Perkin–Elmer Lambda 12 spectrometer. IR spectra
ꢀ1
were measured on Jasco FT/IR-460 Plus using KBr (400–4000 cm ) or
ꢀ1
polyethylene disks (below 400 cm ). In situ IR spectra were measured
by the protonation of the peroxo ligand is fast and in equi-
librium and that the oxygen transfer step from peroxotung-
state to an allylic double bond (step 2) is included in a rate-
limiting step in the case of a terminal allylic alcohol. On the
other hand, the rate for the epoxidation of an internal allylic
on a Mettler Toledo React IR 4000 spectrometer.
Materials: Compound I was synthesized according to the procedure in
1
83
ref. [45] and characterized by elemental analysis, IR, UV/Vis, and
W
NMR spectroscopy, and X-ray crystallographic structural analysis. The
characterization results are as follows: elemental analysis calcd (%) for
K
2
[{W(=O)(O
2 2 2 2 2
) (H O)} (m-O)]·2H O: H 1.16, O 34.6, K 11.27, W 52.99;
ꢀ
1
alcohol 2a decreased with increasing the contents of D O
found: H 1.12, K 11.98, W 53.19; IR (400–4000 cm : KBr disk; below
400 cm : polyethylene disk): n˜ = 966 n˜ (W=O), 854 n˜ (O-O), 764 n˜ asym(W-
2
ꢀ
1
(
Figure S6). The decrease in the rate by the presence of
ꢀ
1
O-W), 615 n˜ sym(W(O
O): l (e) 243 nm (608m cm );
WO
2
)), 566 n˜ asym(W(O
2
ꢀ1
), 332 cm n˜ (W(OH
2
)); UV/Vis
D O is probably explained as follows: The rate for the 2a
2
ꢀ
1
183
(
H
2
=
W NMR (11.2 MHz, D O,
2
epoxidation was 30 times larger than that for 1a and there-
fore the forward process in step 1 would become contribu-
tive to a rate-limiting step. The idea is supported by the dif-
ferent activation parameters between 1a and 2a.
Na
2
4
, 0.3m, pH 2.5): d = ꢀ704.5.
Molecular structure of the anion part of I is depicted in Figure 1. The
method for the preparation of the tetrahexylammonium derivative of
[{W(=O)(O ) (H O)} (m-O)] was modified (i.e., [(C H ) N] replaced
2 2 2 2 6 13 4
[45]
2
ꢀ
+
4732
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4728 – 4734

Epoxidation of Allylic Alcohols
4728 – 4734
+
by [(n-C12
H
25)(CH
3
)
3
N] ) for the synthesis of dodecyltrimethylammoni-
Table 2. Epoxidation of various allylic alcohols with hydrogen peroxide
in water catalyzed by I.
[
a]
um derivative of I. The desired dodecyltrimethylammonium derivative
was obtained with a 50% yield. Elemental analysis calcd (%) for
ꢀ1
Entry Substrate
t [h
]
Product
Yield
%]
[
(n-C12
H
25)(CH
3 3 2 2 2 2 2
) N] [{W(=O)(O ) (H O)} (m-O)]: C 34.35, H 7.00, N
[
2
9
.70, W 35.48; found: C 35.53, H 6.92, N 2.66, W 35.43; IR (KBr): n˜ =
ꢀ
1
63, 936, 911, 838, 770, 720, 603, 569, 531cm
.
[
b]
1
2
10
2
95
Allylic alcohols 7a, 9a, and 10a were synthesized and confirmed by GC
1
13
analysis in combination with mass and H and C NMR spectroscopy as
[
46–49]
[c]
reported previously.
All epoxy alcohols are known and have been
96
1
13
identified by comparison of their H and C NMR signals with the litera-
[
[
c,d]
[
50–57]
3
4
reuse 1
reuse 2
2
2
97
97
ture data.
c,d]
Oxidation of allylic alcohols: The epoxidation was carried out in a glass
vial containing a magnetic stir bar. A typical procedure for the epoxida-
tion of allylic alcohol is as follows: Into a glass vial were successively
placed I (20 mmol), 2a (5 mmol), hydrogen peroxide (30% aqueous solu-
tion, 5 mmol), and water (6 mL). The reaction mixture was stirred at
5
6
7
2
4
5
97
90
98
3
05 K for 2 h. After the reaction was finished, the products were extract-
ed by using dichloromethane or n-pentane, and the yield and product se-
1
lectivity were determined by GC or H NMR analysis. Recovered aque-
ous phase was allowed to recycling. Reaction systems are homogeneous
except for high-molecular weight alcohol 6a (entry 8 in Table 2). When
the epoxidation of 6a was carried out with I under the conditions in
Table 2, only a 20% yield of 6b was obtained for 10 h. The epoxidation
[e]
[e]
[f]
8
9
12
6
85
85
of 6a efficiently proceeded by using
a lipophilic salt, [(n-C12H25)-
(
CH N] [{W(=O)(O (H O)} (m-O)], as a catalyst instead of I. Large-
3
)
3
2
2
)
2
2
2
scale reactions (100 mmol scale) were performed via the same procedures
as those above described. The turnover frequencies (TOFs) were deter-
mined for the initial stage of the epoxidation (~1 h). Reaction conditions
for the larger scale reactions were as follows: Allylic alcohol (100 mmol),
2 2
I (20 mmol), 30% aq. H O (100 mmol), water (120 mL), 305 K, 24 h.
[
[
b]
[g]
[h]
1
0
1
10
10
4
83
89
83
Kinetic studies: A NMR tube (5 mm diameter) was used as a reactor for
the kinetic studies (spin rate, 15 Hz). It was confirmed that the reaction
rates were not affected by the spin rates from 10 to 20 Hz. The reaction
b,d]
1
reuse 1
1
was periodically monitored by the H NMR spectra. Reaction conditions
[i]
are given in the Figure captions (Figure 2–5). Reaction rates (R
kinetic studies were determined from the slopes of reaction profiles
[substrate] vs time) at low conversions (<10%) of the sub-
ꢀ[substrate]
strate and hydrogen peroxide (initial rate method). Reaction profiles and
/[substrate] vs time) were shown in
0
) for the
12
(
0
t
1
3
5
77
the first-order plots (ꢀln([substrate]
t
0
Figure S1–S4, Supporting Information. The rate constants determined by
the slopes were in good agreement with those determined by the initial
rate method.
[
3
a] Reaction conditions: Allylic alcohol (5 mmol), I (20 mmol, 0.4 mol%),
0% aq. H (5 mmol), water (6 mL), 305 K. Yields were based on al-
lylic alcohol which were determined by GC and H NMR using an inter-
2 2
O
Stability of catalyst I: The in situ IR spectra of the reaction solution (1a
1
(
2 2
2.0m), I (0.1m), H O (2.0m), water (10 mL), 298 K) were measured with
nal standard technique. [b] I (100 mmol, 2 mol%). [c] Substrate: cis/trans
an in situ IR spectrometer to confirm the structure stability of I. The
band positions and intensities characteristic of I, that is, n˜ (W=O) (966)
1
3:87, epoxy alcohol: cis/trans = 13:87. [d] Reuse experiment. [e] Isolat-
ed yield. [f] [(n-C12 25)(CH N] [{W(=O)(O (H O)} (m-O)] (20 mmol)
was used as a catalyst. [g] 3-Buten-2-one was produced as a by-product
7% yield). [h] 3-Buten-2-one was formed as a by-product (9% yield).
i] 3-Penten-2-one was formed as a by-product (9% yield).
H
3
)
3
2
2
)
2
2
2
ꢀ1
and n˜ (O-O) (854 cm ), were periodically monitored. The n˜ (W-O-W)
band could not be observed because of overlap with the intense back-
ground of water absorption. No substantial changes in the IR spectra
were observed during the catalytic epoxidation. More concentrated solu-
(
[
tion (1a (2.0m), I (0.3m), H
2
O
2
(2.0m), D
W NMR measurement because of the lower sensitivity. Under the con-
ditions, the epoxidation of 1a was completely finished for 20 min. The
2
O (3 mL)) was used for the
1
83
from the Ministry of Education, Culture, Sports, Science, and Technology
of Japan.
1
83
W NMR spectrum showed a signal at ꢀ704.5 ppm for the acquisition
during catalysis (500 scans, 20 min).
After the epoxidation was completed (reaction conditions in Table 1), the
products were separated by the extraction with dichloromethane or n-
pentane and the volume of the aqueous phase was reduced to the half by
the evaporation. To the solution, ethanol (20 mL) was added and the pre-
cipitate was recovered by the filtration. The stability of the recovered I
was confirmed by the IR (KBr disks) and UV/Vis spectra (0.1 mm in
[1] R. A. Sheldon, J. K. Kochi, Metal Catalyzed Oxidations of Organic
Compounds, Academic Press, New York, 1981.
[2] C. L. Hill in Advances in Oxygenated Processes, Vol. 1 (Eds.: A. L.
Baumstark), JAI Press, London, 1988, pp. 1–30.
[3] M. Hudlucky, Oxidations in Organic Chemistry, ACS Monograph
Series, American Chemical Society, Washington, DC, 1990.
2
H O).
[
4] P. T. Anastas, J. C. Warner, Green Chemistry: Theory and Practice,
Oxford University Press, 1998.
[
[
5] J. H. Clark, Green Chem. 1999, 1, 1.
6] R. A. Sheldon, Green Chem. 2000, 2, G1.
Acknowledgement
[7] P. T. Anastas, L. B. Bartlett, M. M. Kirchhoff, T. C. Williamson,
Catal. Today 2000, 55, 11.
This work was supported in part by the Core Research for Evolutional
Science and Technology (CREST) program of Japan Science and Tech-
nology Corporation (JST) and a Grant-in-Aid for Scientific Research
[8] K. A. Jørgensen, Chem. Rev. 1989, 89, 431.
[9] W. Adam, T. Wirth, Acc. Chem. Res. 1999, 32, 703.
[10] P. A. Bartlett, Tetrahedron 1980, 36, 2.
Chem. Eur. J. 2004, 10, 4728 – 4734
www.chemeurj.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4733

N. Mizuno et al.
FULL PAPER
[
[
[
11] P. Besse, H. Veschambre, Tetrahedron 1994, 50, 8885.
dation of allyl acetate did not proceed at all, while that of 1a pro-
ceeded to give a 26% yield of 1b under the above conditions.
Therefore, we conclude that the complete lack of epoxidation reac-
tivity of allyl acetate is not caused by the low solubility into an
aqueous phase.
12] T. Katsuki, K. B. Sharpless, J. Am. Chem. Soc. 1980, 102, 5974.
13] A. S. Rao, Comprehensive Organic Synthesis, Vol. 7 (Eds.: B. M.
Trost), Pergamon Press, Oxford, 1991, pp. 357–387.
[
[
[
14] G.-J. ten Brink, I. W. C. E. Arends, R. A. Sheldon, Science 2000, 287,
1
636.
15] Thematic issue on “Green Chemistry”: Acc. Chem. Res. 2002, 35,
85.
16] Thematic issue on “Organic Reaction in Water”: Adv. Synth. Catal.
002, 3–4, 219.
[38] W. Adam, C. M. Mitchell, C. R. Saha-Mꢁller, J. Org. Chem. 1999,
64, 3699.
6
[39] The effect of methanol on the epoxidation of 1a was examined. As
shown in Figure S7, the epoxidation was inhibited by the co-exis-
tence of methanol and was completely inhibited in the methanol sol-
vent, suggesting that binding of an alcohol functional group to I is a
key step in Scheme 1.
2
[
[
[
[
17] R. Noyori, M. Aoki, K. Sato, Chem. Commun. 2003, 1977.
18] Z. Racizsewski, J. Am. Chem. Soc. 1960, 82, 1267.
19] D. Prat, B. Delpech, R. Lett, Tetrahedron Lett. 1986, 27, 711.
20] J. Prandi, H. B. Kagan, H. Mimoun, Tetrahedron Lett. 1986, 27,
[40] The signals at d 1.44 and 5.35 sharply increased followed by the in-
crease in the signals at d 2.21 and 9.79. The signal at d 2.04 (s) ap-
peared with an induction period and can be assigned to acetic acid.
2
617.
21] M. Quenard, V. Bonmarin, G. Gelbard, Tetrahedron Lett. 1987, 28,
237.
22] Y. Ishii, K. Yamawaki, T. Ura, H. Yamada, T. Yoshida, M. Ogawa,
J. Org. Chem. 1988, 53, 3587.
23] G. Gelbard, F. Raison, E. R. Lachter, R. Thouvenot, L. Ouahab, D.
Grandjean, J. Mol. Catal. A 1996, 114, 77.
24] K. Sato, M. Aoki, M. Ogawa, T. Hashimoto, R. Noyori, Bull. Chem.
Soc. Jpn. 1997, 70, 905.
25] D. Hoegaerts, B. F. Sels, D. E. de Vos, F. Verpoort, P. A. Jacobs,
Catal. Today 2000, 60, 209.
26] C. Denis, K. Misbahi, A. Kerbal, V. Ferrieres, D. Plusquellec, Chem.
Commun. 2001, 2460.
27] W. Adam, P. L. Alsters, R. Neumann, C. R. Saha-Mꢁller, D. Slobo-
da-Rozner, R. Zhang, Synlett 2002, 2011.
28] W. Adam, P. L. Alsters, R. Neumann, C. R. Saha-Mꢁller, D. Slobo-
da-Rozner, R. Zhang, J. Org. Chem. 2003, 68, 1721.
29] K. Kamata, K. Yonehara, Y. Sumida, K. Yamaguchi, S. Hikichi, N.
Mizuno, Science 2003, 300, 964.
1
13
[
[
[
[
[
[
[
[
[
[
The H and C NMR spectra of the reaction solution of I (50 mm)
with 1a (870 mm) in D O in the range of 278–305 K showed only
2
2
resonances of 1a and 1b, and no resonances due to tungsten-alco-
holate species were observed.
[41] W. Clegg, R. J. Errington, P. Kraxner, C. Redshaw, J. Chem. Soc.
Dalton Trans. 1992, 1431.
[42] D. V. Baxter, M. H. Chisolm, S. Doherty, N. E. Gruhn, Chem.
Commun. 1996, 1129.
[43] G. G. Allan, A. N. Neogi, J. Catal. 1970, 16, 197.
[44] D. V. Deubel, J. Phys. Chem. A 2001, 105, 4765.
[45] N. J. Cambell, A. C. Dengel, C. J. Edwards, W. P. Griffith, J. Chem.
Soc. Dalton Trans. 1989, 1203.
[46] S. Krishnamurthy, H. C. Brown, J. Org. Chem. 1975, 40, 1864.
[47] S. Krishnamurthy, H. C. Brown, J. Org. Chem. 1977, 42, 1197.
[48] B. Morgan, A. C. Oehlschlager, T. M. Stokes, J. Org. Chem. 1992,
57, 3231.
[49] H. O. House, S. S. Ro, J. Am. Chem. Soc. 1958, 80, 2428.
[50] W. Adam, M. Braun, A. Griesbeck, V. Luccini, E. Staab, B. Will, J.
Am. Chem. Soc. 1989, 111, 203.
30] K. Kamata, K. Yamaguchi, S. Hikichi, N. Mizuno, Adv. Synth. Catal.
2
003, 345, 1193.
[51] M. J. Kurth, M. A. Abreo, Tetrahedron 1990, 46, 5085.
[52] H. Pettersson, A. Gogoll, J.-E. Bꢂckvall, J. Org. Chem. 1992, 57,
6025.
[53] H. Fuji, K. Oshima, K. Utimoto, Chem. Lett. 1992, 967.
[54] D. F. Taber, J. B. Houze, J. Org. Chem. 1994, 59, 4004.
[55] D. C. Dittmer, Y. Zhang, R. P. Discordia, J. Org. Chem. 1994, 59,
1004.
[
[
[
[
[
31] See the Supporting Information (Table S1).
32] H. Mimoun, I. S. De Roch, L. Sajus, Tetrahedron 1970, 26, 37.
33] K. B. Sharpless, R. C. Michaelson, J. Am. Chem. Soc. 1973, 95, 6136.
34] K. B. Sharpless, T. R. Verhoeven, Aldrichimica Acta 1979, 12, 63.
35] T. Ito, K. Jitsukawa, K. Kaneda, S. Teranishi, J. Am. Chem. Soc.
1
979, 101, 159.
36] W. Adam, R. Kumar, T. I. Reddy, M. Renz, Angew. Chem. 1996,
08, 578; Angew. Chem. Int. Ed. Engl. 1996, 35, 533.
[
[
[56] Y. F. Zheng, D. S. Dodd, A. C. Oehlschlager, P. G. Hartman, Tetrahe-
dron 1995, 51, 5255.
1
37] Water/acetonitrile (1:2 v/v) was also used for the epoxidation to
solubilize both the catalyst I and allyl acetate. The reaction condi-
tions were as follows: Substrate (5 mmol), I (20 mmol), 30% aq.
[57] W. Adam, A. Corma, T. I. Reddy, M. Renz, J. Org. Chem. 1997, 62,
3631.
Received: April 9, 2004
H
2
O
2
(5 mmol), water/acetonitrile (2+4 mL), 305 K, 24 h. The epoxi-
Published online: August 6, 2004
4734
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4728 – 4734