Kinetics and mechanism of the epoxidation of alkyl-substituted alkenes by hydrogen peroxide, catalyzed by methylrhenium trioxide

Article abstract of DOI:10.1021/jo951774e

Epoxidations of alkyl-substituted alkenes, with hydrogen peroxide as the oxygen source, are catalyzed by CH₃ReO₃ (MTO). The kinetics of 28 such reactions were studied in 1:1 CH₃CN-H₂O at pH 1 and in methanol. To accommodate the different requirements of these reactions, ¹H-NMR, spectrophotometric, and thermometric techniques were used to acquire kinetic data. High concentrations of hydrogen peroxide were used, so that diperoxorhenium complex CH₃Re(O)(η²-O₂)₂(H ₂O), B, was the only predominant and reactive form of the catalyst. The reactions between B and the alkenes are about 1 order of magnitude more rapid in the semiaqueous solvent than in methanol. The various trends in reactivity are medium-independent. The rate constants for B with the aliphatic alkenes correlate closely with the number of alkyl groups on the olefinic carbons. The reactions become markedly slower when electron-attracting groups, such as halo, hydroxy, cyano, and carbonyl, are present. The rate constants for catalytic epoxidations with B and those reported for the stoichiometric reactions of dimethyldioxirane show very similar trends in reactivity. These findings suggest a concerted mechanism in which the electron-rich double bond of the alkene attacks a peroxidic oxygen of B. These data, combined with those reported for the epoxidation of styrene (a term intended to include related molecules with ring and/or aliphatic substituents) by B and by the monoperoxo derivative of MTO, suggest that all of the rhenium-catalyzed epoxidations occur by a common mechanism. The geometry of the system at the transition state can be inferred from these data, which suggest a spiro arrangement.

Full text of DOI:10.1021/jo951774e

J . Org. Chem. 1996, 61, 3969-3976
3969
Kin etics a n d Mech a n ism of th e Ep oxid a tion of Alk yl-Su bstitu ted
Alk en es by Hyd r ogen P er oxid e, Ca ta lyzed by Meth ylr h en iu m
Tr ioxid e
Ahmad M. Al-Ajlouni and J ames H. Espenson*
Ames Laboratory and Department of Chemistry, Iowa State University, Ames, Iowa 50011
Received September 28, 1995X
Epoxidations of alkyl-substituted alkenes, with hydrogen peroxide as the oxygen source, are
catalyzed by CH
3
ReO
3
(MTO). The kinetics of 28 such reactions were studied in 1:1 CH
3
2
CN-H O
1
at pH 1 and in methanol. To accommodate the different requirements of these reactions, H-NMR,
spectrophotometric, and thermometric techniques were used to acquire kinetic data. High
2
concentrations of hydrogen peroxide were used, so that diperoxorhenium complex CH
(H
3
Re(O)(η -
O
2
)
2
2
O), B, was the only predominant and reactive form of the catalyst. The reactions between
B and the alkenes are about 1 order of magnitude more rapid in the semiaqueous solvent than in
methanol. The various trends in reactivity are medium-independent. The rate constants for B
with the aliphatic alkenes correlate closely with the number of alkyl groups on the olefinic carbons.
The reactions become markedly slower when electron-attracting groups, such as halo, hydroxy,
cyano, and carbonyl, are present. The rate constants for catalytic epoxidations with B and those
reported for the stoichiometric reactions of dimethyldioxirane show very similar trends in reactivity.
These findings suggest a concerted mechanism in which the electron-rich double bond of the alkene
attacks a peroxidic oxygen of B. These data, combined with those reported for the epoxidation of
styrene (a term intended to include related molecules with ring and/or aliphatic substituents) by
B and by the monoperoxo derivative of MTO, suggest that all of the rhenium-catalyzed epoxidations
occur by a common mechanism. The geometry of the system at the transition state can be inferred
from these data, which suggest a spiro arrangement.
In tr od u ction
efficiently catalyzes the epoxidation of olefins, in semi-
aqueous or organic solvents, at low and moderate tem-
peratures with high activity and catalytic turnovers.
The rhenium peroxides A and B are the catalytically
active intermediates. In the epoxidation of styrenes, both
The epoxidation of olefins is a subject with a consider-
9-11
1
-5
able history, amply explored in books and reviews. The
oxygen source may be hydrogen peroxide, an alkyl
hydroperoxide, or oxygen itself. Epoxides are useful
intermediates, and many of their ring-opened products,
5
6
such as glycols and glycol ethers, are of practical value.
A catalyst, often a transition metal complex, is required
in nearly every case. We have limited the scope of this
research to the use of hydrogen peroxide as the oxygen
source. The various transition elements that catalyze
epoxidation reactions may differ as to specificity, rate,
and mechanism. Compounds of some group IV-VI
metals in high oxidation states, such as Ti(IV) and V(V),
are converted to peroxo complexes. The intermediate
from V(V) gives rise to radical pathways resulting in
nonselective oxidations and catalyst decomposition,
whereas Ti(IV) complexes give mainly stereospecific
reactions.7 Generally speaking, the best catalysts in this
group contain Mo(VI) and W(VI).8
A and B are comparably active catalysts,12 as they are
13
3
for oxygen transfer to other substrates such as PR ,
14
- 15
- 16
2
R S, Br , and Cl . Having established several times
that A and B are both effective catalysts, we carried out
the present study at high concentrations of hydrogen
peroxide, to make B the only rhenium species present at
an appreciable concentration.
We have studied the kinetics of epoxidation of 28 alkyl-
substituted alkenes in two media, acidic acetonitrile-
water and methanol, using three techniques for obtaining
kinetic data. Without water (other than that introduced
with the peroxide or formed in the reaction), excellent
yields of epoxides are obtained in methanol and unacidi-
3 3
Methylrhenium trioxide (CH ReO or MTO), a rheni-
um(VII) derivative and thus a powerful Lewis acid,
X
Abstract published in Advance ACS Abstracts, May 15, 1996.
(
1) Sheldon, R. A. Top. Curr. Chem. 1993, 164, 23-28.
(9) Herrmann, W. A.; Fischer, R. W.; Marz, D. W. Angew. Chem.,
Int. Ed. Engl. 1991, 30, 1638.
(2) Simandi, L. I. Catalytic Activation of Dioxygen by Metal Com-
plexes; Kluwer Academic Publishers: Dordrecht, 1992; p 109.
3) Drago, R. S. In Dioxygen Activation and Homogeneous Catalytic
Oxidation; Elsevier: New York, 1991; Vol. 66, p 83.
4) Laszlo, P.; Levart, M.; Bouhlel, E.; Montaufier, M.; Singh, G. P.
In Catalytic Selective Oxidation; American Chemical Society: Wash-
ington, DC, 1993; Chapter 4.
(10) Herrmann, W. A.; Fischer, R. W.; Scherer, W.; Rauch, M. H.
Angew. Chem., Int. Ed. Engl. 1993, 32, 1157.
(11) Herrmann, W. A.; Fischer, R. W.; Rauch, M. U.; Scherer, W. J .
Mol. Catal. 1994, 86, 243.
(12) Al-Ajlouni, A.; Espenson, J . H. J . Am. Chem. Soc. 1995, 117,
9243.
(
(
(
5) Strukul, G. Catalytic Oxidations with Hydrogen Peroxide as
Oxidant; Kluwer Academic Publishers: Dordrecht, 1992.
6) Inoue, M.; Nakayama, E.; Nakamura, Y. Bull. Chem. Soc. J pn.
991, 64, 3442.
(13) Abu-Omar, M. M.; Espenson, J . H. J . Am. Chem. Soc. 1995,
117, 272.
(14) Vassell, K. A.; Espenson, J . H. Inorg. Chem. 1994, 33, 5491.
(15) Pestovsky, O.; vanEldik, R.; Huston, P.; Espenson, J . H. J .
Chem. Soc., Dalton Trans. 1995, 227.
(
1
(
7) H o¨ ft, E. Top. Curr. Chem. 1993, 164, 63-78.
(8) J acobson, S. E.; Tang, R.; Mares, F. Inorg. Chem. 1973, 17, 3055.
(16) Hansen, P. J .; Espenson, J . H. Inorg. Chem. 1995, 34, 5389.
S0022-3263(95)01774-9 CCC: $12.00 © 1996 American Chemical Society

3
970 J . Org. Chem., Vol. 61, No. 12, 1996
Al-Ajlouni and Espenson
Sch em e 1
fied, neat acetonitrile. The more water present, however,
particularly in the acidic medium, the greater the yield
of the ring-opened diol originating from the primary
epoxide product. From the kinetic data we can now
present a detailed picture of the reactions, including a
description of their steric and electronic requirements.
Exp er im en ta l Section
Ma ter ia ls. The solvents used for quantitative kinetic
studies were 1:1 CH
3
CN-H
2 3
O at pH 1 and CD OD. Methanol-
d
4
(Cambridge Isotopes) and HPLC-grade acetonitrile (Fisher)
were used. Water was purified by a Millipore-Q water
purification system. When required, 0.1 M perchloric acid,
diluted from the 70% reagent (Fisher), was used to maintain
3 3
pH. Occasional experiments were done in CH OH or CD CN.
within a Parr Model 1455 precision calorimeter, equipped with
a thermistor in a platinum resistance probe. The solutions
were separately allowed to reach temperature equilibration
for 20-30 min. The two solutions were then mixed and stirred
for 1-2 min. The stirrer was then stopped, and the temper-
ature was recorded at 5-10 s intervals by a computer with
Lotus 1-2-3 software. For each alkene, experiments were done
1
7
Methylrhenium trioxide was purified first by sublimation,
then by recrystallization from methylene chloride-hexane, and
finally by a second sublimation. Stock solutions of MTO in
methanol-d
or 2 weeks. The concentration was determined spectropho-
tometrically after dilution into a large volume of water: 239
4
or acetonitrile were stored at 5 °C and used within
1
-
1
-1
-1
-1 18
nm (ꢀ 1900 L mol cm ), 270 nm (ꢀ 1300 L mol cm ).
with 1.0 M H
2 2 T
O , with [Re] (which is essentially equal to [B]
at this high a peroxide concentration) being varied in the range
We purchased the alkenes (Aldrich); those which proved to
1
be pure by H NMR were used as such. trans-4-Octene,
0
.5-5.0 mM, and with [alkene] 10-50 mM. The reaction
1
-octene, and 1-hexene were purified by distillation. Aqueous
followed first-order kinetics, with temperature-time curves
independent of the reaction stoichiometry; in these systems
solutions of hydrogen peroxide were prepared from the 30%,
or occasionally the 3%, reagent. The concentrated reagent,
which proved quite stable over time, was standardized by
iodometric titration; the diluted peroxide solutions were much
less stable and were made up daily by dilution.
∆
T was within the range 0.2-0.8 °C. The temperature profile
is in addition affected by a slow thermal leakage in the
adiabatic calorimeter. With the thermal leakage expressed
as a first-order function with a constant k
cooling), the expression for the temperature at a given time
T
, (Newton’s law of
Kin etics. Three methods were applied, as required by the
range of alkenes and solvents. In every case the temperature
was controlled at 25 °C. (a) The reactions studied in methanol
20
(
T
t
) is given by eq 1, where the constants m
1
, m
2
, and m
3
are
(1)
related to the temperature factors of the experiment.
1
were followed by H NMR in a total volume of 0.6 mL. These
experiments used 0.50 M H
2 2
O , 1-50 mM MTO, and 0.05-
-kψt
-kTt
T ) m + m e
+ m e
3
0
.1 M alkene, added last. The relative amounts of peroxide
t
1
2
and alkene were chosen with a concern for the requirements
of the kinetic analysis which was carried out by a first-order
kinetic equation. The proton spectrum was recorded at 2-20
min increments over the 2-5 h reaction time. The changes
in intensity (Y) of the alkene signal(s) were fit to a single-
Resu lts
Ca ta lytic In ter m ed ia tes. As cited in the Introduc-
tion, quantitative kinetic data have already been reported
for oxygen-transfer reactions in the MTO/H O system.
2 2
exponential decay: Y
b) A spectrophotometric method was used for some of the
reactions carried out in 1:1 CH O, in which the acid
t
) Y
∞
+ (Y
0
- Y
∞
) × exp(-k
ψ
t).
(
Because of that, we shall rely on the established course
of these systems to provide the initial guidance for the
kinetic scheme that might apply to epoxidation reactions
of the alkyl-substituted alkenes. In this scheme, either
or both of the previously-depicted rhenium peroxides, A
or B, may transfer an oxygen atom to the alkene.
In the event that the monoperoxide A is reactive, then
MTO is immediately recovered along with the epoxide.
Similarly, if an oxygen atom is transferred from the bis-
peroxide B, then A is generated along with the epoxide.
This sequence of events in shown in Scheme 1.
Prior research has established that the conversion of
MTO to A upon reaction with hydrogen peroxide consti-
tutes a reversible equilibrium, likewise the conversion of
A to B. With the notation in Scheme 1, these are the
applicable equilibrium constants at 25.0 °C:
3
CN-H
2
concentration and the ionic strength was maintained at 0.100
M unless specified otherwise. The acid concentration needs
2 2
to be fairly high, because MTO-H O solutions otherwise
1
9
decompose fairly rapidly, unlike MTO itself. The cuvettes
had short optical paths, 0.01-0.2 cm, because the alkenes to
which the method was applicable had high molar absorptivities
and the other reagents contributed a large absorbance back-
ground at the wavelengths used. The reaction mixtures were
prepared in the reaction cuvette which was held in a thermo-
statted water bath, with the last component added being the
alkene. Air was not excluded, since controls showed there was
no need. Hydrogen peroxide was used in large excess over
the alkene, such that the bis-peroxide B was maintained at
constant concentration throughout. The data were obtained
by following the loss of the alkene absorption in the range
2
00-225 nm. The absorbance-time data were fitted by
nonlinear least-squares to a first-order kinetic equation: A
+ (A - A t).
) × exp(-k
c) A number of the reactions of the alkyl-substituted
t
K1/L mol-
1
K2/L mol-1
)
A
(
∞
0
∞
ψ
medium
13
1
2
1
:1 CH3CN-H2O, pH 1
1.3 × 10
1.36 × 10
alkenes were not accompanied by a useful absorbance change
at an accessible wavelength (>200 nm). These reactions, along
with several known systems that we used as controls, were
studied in the aqueous acetonitrile medium by a thermometric
21
2
2
CH3OH
2.61 × 10
8.14 × 10
Kin etic Design . The fact that the peroxo complex
formation reactions are reversible certainly does not
mean that the concentrations of the three species are
necessarily sustained at equilibrium under all circum-
stances in which the catalytic cycles are in operation. In
many of the cases already considered, especially those
2
0
method. The components were placed inside two reservoirs
(
17) Herrmann, W. A.; K u¨ hn, F. E.; Fischer, R. W.; Thiel, W. R.;
Ramao, C. C. Inorg. Chem. 1992, 31, 4431.
18) Yamazaki, S.; Espenson, J . H.; Huston, P. Inorg. Chem. 1993,
2, 4683.
19) Abu-Omar, M. M.; Hansen, P. J .; Espenson, J . H. J . Am. Chem.
Soc. 1996, 118, 4966.
20) Kustin, K.; Ross, E. W. J . Chem. Educ. 1993, 70, 454.
(
3
13
with reactive substrates such as phosphines and sul-
(
(
(21) Zhu, Z.; Espenson, J . H. J . Org. Chem. 1995, 60, 1326.

Epoxidation of Alkyl-Substituted Alkenes by H
2
O
2
J . Org. Chem., Vol. 61, No. 12, 1996 3971
fides,14 MTO, A, and B are interconverted, but they are
not equilibrated rapidly enough to maintain the equilib-
ria among them. These reactions must be considered as
kinetic steps, along with the oxygen-transfer step. This
means that the concentrations of A and B used in
analysis of the kinetic data are those given by the steady-
state approximation, not those described by the equilib-
rium constants listed.
That is to say, the immediate fate of a newly-born A is
not necessarily to react further with H O to form B; it
2 2
might instead react with the substrate to yield product,
thereby regenerating MTO. Indeed, instead of either of
these, A might revert to MTO upon addition of water and
loss of peroxide. The choice a given molecule of A
exercises depends upon the relative rates of three reac-
F igu r e 1. Kinetic data for the epoxidation of 1-octene (0.10
tions and therefore on the relative values of k
substrate], and k-1
One method of incorporating all of these considerations
2 2 2 3
[H O ], k -
M, with 50 mM MTO) and trans-4-octene (0.10 M, with 25 mM
[
.
1
MTO) and 0.50 M H
2
O
2
in methanol-d
4
at 25 °C. The H NMR
signal intensities were recorded for an olefinic proton of
-octene at δ 5.6 and of trans-4-octene at δ 5.4.
is to formulate a general expression for the rate of
reaction in terms of the three independent concentration
variables, which include the total catalyst concentration,
1
solvents. In all of these experiments, [H
2 2
O ] was fixed
between 0.3 and 1.0 M. At the lower value, >97% of the
total MTO is present as B in aqueous acetonitrile and
[Re]
T
, defined by [Re]
T
) [MTO] + [A] + [B]. The
resulting rate equation appears to have such a complex
form (eq 2) that it might be difficult to apply in a form
that would convince one that it represented a unique
>
99% in methanol. One approximation made in reducing
eq 2 to eq 3 is clearly satisfied. In addition, we rely on
the fact that k is not large compared to k ; indeed k
and k are of the same order of magnitude in any number
solution. This equation has, in fact, been applied suc-
3
4
3
cessfully in other studies,1
2,14,22
where the conditions were
4
such that none of the terms could be neglected to simplify
the expression. The use of eq 2 (alk ) alkene or other
substrate) was feasible in such cases because the rate
1
2
of other systems, the case of the styrenes being the most
germane.
A further advantage is gained by working at constant
[H O ], since the amount of water present in the solution
2 2
remains nearly constant. This is important since the
rates of epoxidation of styrene are sensitive to the amount
of water.12
1
constants for the rhenium-peroxide equilibrations (k ,
, and k-2) have been determined independently.12,14,22
-1, k
2
k
d[epoxide]
)
dt
Kin etic Da ta . The results from the three kinetic
2
k k k [Re] [alk][H O ]
1
2
4
T
2
2
techniques will be presented in turn, and then the
k k [Re] [H O ][alk] +
1
3
T
2
2
1
k4[alk] + k-2
comparisons made between them. The H NMR method
(2)
was applied to reaction mixtures with 0.1 M alkene, 0.50
M hydrogen peroxide, and 1-50 mM MTO. Figure 1
displays the results of two typical reactions, with 1-octene
and trans-4-octene. The signal intensity for an olefinic
proton absorption of each alkene is displayed as a
function of time. The smooth curve drawn through the
experimental points is the nonlinear least-squares fit to
the exponential decay function given before. The values
of the pseudo-first-order rate constants so obtained were
divided by the total catalyst concentration. This quotient,
2
k k [H O ]
1
2
2
2
k-1 + k [alk] + k [H O ] +
3
1
2
2
k4[alk] + k-2
The numerical fitting of initial rate data to this
equation was the procedure used in some of the experi-
ments carried out in our earlier research dealing with
the epoxidation of styrene. On the other hand, having
validated this approach and the kinetic model now in
several cases, including styrene, the most germane
example in the present instance, it seemed preferable to
move directly in the case at hand to a set of conditions
in which one or another limiting form of eq 2 would be
realized. This would allow the direct evaluation of the
epoxidation rate constants through the use of a simplified
but entirely valid kinetic treatment. In particular,
consider the limiting form at a concentration of hydrogen
peroxide sufficiently high that only the last term in the
numerator, and the last in the denominator, is significant
compared to the rest. The simplified form is
k
ψ T 4
/[Re] , represents the value of the rate constant k ,
which applies to the reaction between B and the alkene
(Scheme 1). These values, in methanol-d at 25 °C, are
4
presented in Table 1.
The nonconjugated alkenes, with λmax e 200 nm, could
not be studied by the spectrophotometric method. Other
alkenes, such as â-pinene, R-pinene, and limonene, have
absorption maxima with quite large molar absorptivities
in the range 200-225 nm. Since MTO, H
absorb significantly in this range, short-path optical cells
0.01-0.2 cm) were used. A series of experiments with
[H ] g 0.3 M, [alkene] ) 5-20 mM, and [Re] varying
between 0.2 and 5 mM were carried out in acidic
acetonitrile. Note that the value of [H ] in this range
is immaterial, in that its concentration does not affect
the reaction rate (eq 3); note also that [Re] was adjusted
2 2
O , A, and B
(
d[epoxide]
d[alkene]
O
2
2
T
)
-
) k [Re] [alkene] (3)
4 T
dt
dt
2
O
2
The concentration of hydrogen peroxide needed to
attain this limit accurately is different in the two
T
to bring the reaction to a convenient time scale and that
it was varied for each alkene to confirm that the direct
(22) Espenson, J . H.; Pestovsky, O.; Huston, P.; Staudt, S. J . Am.
Chem. Soc. 1994, 116, 2869.
T
proportionality of rate on [Re] (eq 3) was followed. The

3
972 J . Org. Chem., Vol. 61, No. 12, 1996
Al-Ajlouni and Espenson
Ta ble 1. Secon d -Or d er Ra te Con sta n ts a t 25 °C for th e Ep oxid a tion of Olefin s w ith CH3Re(O)(O2)2(H2O), B, by Va r iou s
Meth od s in Tw o Solven ts
k4/L mol- s-1
1
CH3OD
1:1 CH3CN-H2O,
olefin
(method)a
pH 1 (method)a
styrene (1)
0.019 ( 0.001 (N)
0.11 ( 0.01 (S)
0
.021 ( 0.002 (S)
4
-methoxystyrene (2)
0.62 ( 0.05 (T)
0
.57 ( 0.04 (S)
1
-methylcyclohexene (3)
0.32 ( 0.01 (N)
0.35 ( 0.01 (N)
4.20 ( 0.16 (T)
3.63 ( 0.10 (T)
3.88 ( 0.20 (S)
2.21 ( 0.15 (S)
2.79 ( 0.13 (S)
â-pinene (4)
0
.34 ( 0.01 (S)
R-pinene (5)
b
(R)-(+)-limonene (6)
0.15 (N)
c
0
.09 (N)
norbornylene (7)
cyclohexene (8)
methylenecyclohexane (9)
0.250 ( 0.001 (N)
0.108 ( 0.003 (N)
0.096 ( 0.001 (N)
1.33 ( 0.03 (N)
0.260 ( 0.007 (N)
0.061 ( 0.001 (N)
0.106 ( 0.008 (N)
0.039 ( 0.001 (N)
0.048 ( 0.003 (N)
1.88 ( 0.04 (T)
1.06 ( 0.02 (T)
0.97 ( 0.06 (T)
8.93 ( 0.55 (T)
3.92 ( 0.04 (T)
0.55 ( 0.06 (T)
2
2
2
,3-dimethyl-2-butene (10)
-methyl-2-pentene (11)
-methyl-1-pentene (12)
cis-3-hexen-1-ol (13)
trans-3-hexen-1-ol (14)
trans-4-octene (15)
0.253 ( 0.003 (T)
0
.289 ( 0.005 (S)
1
1
3
5
4
-octene (16)
-hexene (17)
,3-dimethyl-1-butene (18)
-hexen-2-one (19)
-hexen-3-one (20)
0.0126 ( 0.001 (N)
0.0124 ( 0.0002 (N)
0.0087 ( 0.0003 (N)
0.098 ( 0.021 (T)
0.066 ( 0.003 (S)
0.0038 ( 0.0003 (S)
allyl alcohol (21)
allyl phenyl ether (22)
0.0033 ( 0.0002 (N)
0.0052 ( 0.0001 (N)
2
-buten-1,2-diol (23)
0.0326 ( 0.0004 (S)
0.012 ( 0.001 (S)
0.0051 ( 0.0003 (S)
cis-1-bromopropene (24)
trans-1-bromopropene (25)
1
1
-chloro-2-methyl-propene (26)
-bromo-2-methyl-propene (27)
0.0160 ( 0.0005 (N)
0.0098 ( 0.0003 (N)
methyl acrylate (28)
,4-dicyano-2-butene (29)
acrylonitrile (30)
<0.001 (S)
<0.001 (S)
<0.001 (S)
1
a
By NMR (N), spectrophotometric (S), or thermometric (T) methods. b For the epoxidation of the internal olefinic group. c For the
epoxidation of the external olefinic group (see text).
F igu r e 2. Spectrophotometrically-determined values of the
F igu r e 3. Typical curves of temperature versus time are
shown for the epoxidation of 2,3-dimethyl-2-butene (∼20 mM)
pseudo-first-order rate constant k
(
(
ψ
for the reaction of â-pinene
5 mM) with hydrogen peroxide (0.4 M) as a function of [Re]
which is ∼[B]). The data are linear through the origin as in
T
with hydrogen peroxide (1.0 M) at [Re] ) 0.45 and 0.75 mM
T
in aqueous acetonitrile at pH 1 and 25 °C.
eq 3. The medium is 1:1 acetonitrile-water at pH 1 and 25
°
C.
Typical temperature-time data are shown in Figure 3
for the epoxidation of 2,3-dimethyl-2-butene in acidic
acetonitrile-water. The data were analyzed according
to first-order kinetics, after allowance for a temperature
solutions of hydrogen peroxide and MTO were allowed
to equilibrate for about 5-10 min before the reaction was
initiated by the addition of the alkene.
T
drift term, designated exp(-k t) in eq 1. The values of
The absorbance-time data were fit to first-order
the pseudo-first-order reaction rate constants, k , are
ψ
kinetic functions. The values of k
of [Re]
affords k
are also listed in Table 1.
Experiments carried out using the thermometric tech-
nique all employed 1.0 M H . The other concentration
ranges were 10-50 mM alkene and 0.2-5.0 mM MTO.
ψ
are a linear function
shown as a function of [B] (= [Re] ) in Figure 4 for
T
T
, as shown for â-pinene in Figure 2. The slope
. The values of k determined by this method
norbornylene and cyclohexene. The values of k , given
4
4
4
by the slope of this graph, are given in Table 1.
To show that the three methods give the same values
of the rate constants, several compounds were examined
by the NMR (N), spectrophotometric (S), and thermo-
metric (T) methods. In acidic, aqueous acetonitrile and
2
O
2

Epoxidation of Alkyl-Substituted Alkenes by H
2
O
2
J . Org. Chem., Vol. 61, No. 12, 1996 3973
F igu r e 5. Time progression of concentrations for trans-4-
octene, showing the alkene, the epoxide, and the diol. Condi-
F igu r e 4. Thermometrically-determined rate constants for
the epoxidation of cyclohexene (solid circles) and norbornylene
tions: 0.10 M alkene, 0.5 M H
methanol-d at 25 °C.
2 2
O , and 25 mM MTO in
T
(open circles) shown to be linear functions of [Re] , in agree-
4
ment with eq 3. The data were taken in aqueous acetonitrile
at pH 1 and 25 °C.
the conversion of epoxide to diol; it is not a catalyst,
however, for direct hydration of the alkene to the diol by
a pathway that would bypass the epoxide.
(L mol- s-1) at 25 °C to
1
in methanol, the values of k
4
which the different methods were applied are as follows:
Neat acetonitrile was also used to examine the prod-
ucts, even though it was not one of the media for kinetics;
here the epoxide is stable. In CH
3
OD, under conditions
used for kinetic studies, formation of the epoxide was
faster than its hydrolysis to the 1,2-diol. For trans-4-
1
:1 CH3CN/H2O, pH 1
3.63 ( 0.10 (T)
methanol
â-pinene
â-pinene 0.34 ( 0.01 (S)
0.35 ( 0.01 (N)
styrene 0.021 ( 0.001 (S)
0.019 ( 0.001 (N)
3
.88 ( 0.20 (S)
4
-methoxystyrene 0.62 ( 0.05 (T)
0
.57 ( 0.04 (S)
octene, when [Re]
pseudo-first-order rate constant of the epoxidation step
T
) 25 mM and [H
2 2
O ] ) 0.5 M, the
trans-4-octene
0.253 ( 0.003 (T)
0
.289 ( 0.005 (S)
-
3
-1
is 1.34 × 10
s . Under the same conditions, the
The thermometric experiments, by definition, are
adiabatic, not isothermal. The temperature rise was
always <0.6 °C, and usually ∼0.2-0.4 °C. We may take
as a guide to the effect of this variation the increase in
the rate constant for 4-methoxystyrene. For this com-
pound, values of k
temperature, leading to the value ∆H ) 42.8 kJ mol
With that, a 0.6° increase will cause k to increase by
.8%, and even less for smaller temperature increases.
pseudo-first-order rate constant for the ring opening is
-4 -1
1
.70 × 10
together than by MTO alone.
The case of limonene (6) is an interesting one, in that
s , more strongly enhanced by MTO and
2 2
H O
it has two sites of unsaturation, either of which can be
epoxidized. The kinetics were studied by recording
4
were determined as a function of
q
-1 12
.
1
intensity of the H NMR signal as a function of time. The
resonance of H was quite distinct, since it was far
removed from the water peak; H and H were also used,
but were less reliable because of that interference. The
4
a
4
b
c
We take this effect as fairly insignificant within the
experimental errors and have therefore treated each
determination by isothermal kinetics.
-1 -1
value of k
ψ
gave k
4
) 0.235 L mol
s ; the mathematics
23
of this situation is just like that for parallel reactions,
in that the rate constant is the sum of those for the
separate pathways, irrespective of the fact that a reso-
nance associated with just one of the double bonds was
observed. To resolve the rate constant into its compo-
nents, the ratio of the two different epoxides (6e, 6i) was
determined, giving [6i]/[6e] ) 1.9. This measurement
was carried out when only ca. 34% of the starting
material had been consumed, to minimize the influence
of formation of the double epoxide and the diols derived
from all three epoxides. With this ratio, and the total
1
P r od u cts were identified by H NMR for every alkene
studied in methanol. In each case the epoxide was
formed first. The epoxide then underwent ring opening
to the diol upon reaction with water, of which there are
two sources, the 30% stock solution of hydrogen peroxide
and that produced in the overall reaction. These steps
are as follows:
-1
rate constant, the separated values of k
4
are 0.09 L mol
-
1
-1 -1
s
(external) and 0.15 L mol
s
(internal).
In the second stage, three processes are at work:
uncatalyzed hydrolysis of the epoxide, which we have
shown is negligibly slow (>2 d) under these conditions,
catalysis by MTO alone, which certainly occurs albeit
2 2
slowly, and catalysis jointly by MTO and H O , which is
the most effective pathway of the three.1
1,12
Discu ssion
Kin etic Mod el. The adherence of the kinetic data to
The data supporting the sequence of events in eq 4 are
displayed for trans-4-octene in Figure 5. The consump-
tion of the alkene and increase of the epoxide occur
concomitantly, as do the consumption of the epoxide and
the buildup of the diol. These studies establish that MTO
is a catalyst for both the epoxidation of the alkene and
the rate equations, here and in related instances, espe-
(23) Espenson, J . H. Chemical Kinetics and Reaction Mechanisms;
2nd ed.; McGraw-Hill Inc.: New York, 1995; p 277.

3
974 J . Org. Chem., Vol. 61, No. 12, 1996
Al-Ajlouni and Espenson
Sch em e 2
4
F igu r e 6. Correlation of the rate constants (k ), for the
reactions between aliphatic alkenes and B, with N, the number
of aliphatic substituents on the double bond.
or more of the H’s at an olefinic position. When we group
the alkenes accordingly, they will have between one (N
)
1) and four (N ) 4) alkyl groups attached to the olefinic
carbons. We limit ourselves to comparing the acyclic
compounds without aryl, hydroxy, or halogen substitu-
ents to avoid the mixing in of other factors. For com-
cially for the family of styrenes,12 serves to support the
pounds 10-12 and 15-18 of Table 1, the plot of log (k
4
)
kinetic model and the general pattern of reactivity given
against N is shown in Figure 6. The positive correlation
in each solvent is evident; indeed there appears to be an
almost linear relation over some 2 orders of magnitude
in reactivity, although we do not attach particular
significance to the correlation being linear. The implica-
tion seems apparent: the more electron rich an alkene,
other factors aside, the more rapidly it is epoxidized in
in Scheme 1. In the earlier studies1
2-16
much effort was
devoted to defining separately the reactivity of A (the
monoperoxorhenium(VII) complex) as compared to B, the
bis-peroxo complex. In nearly every case, including
styrene, the differences in reactivity between A and B
were not remarkable. Having made that distinction, we
therefore decided to simplify the present study by limit-
ing the effort to B alone. It should be noted that this
usually cannot be achieved by using a synthesized and
isolated sample of B. This limitation is so because the
the MTO/H
2 2
O system. The effect of an additional
substituent in this case is, however, less marked than in
2
4-26
the reactions of peroxy acids and R-azo hydroperoxides.
The marked retarding effect of halogen substituents
is evident by this comparison of values of rate constants
step by which H
tive with that by which the alkene reacts; that is, it is a
matter of the relative values of k-2 and k [alkene]. With
2 2
O is released from B is often competi-
in methanol (throughout the Discussion, values of k
4
(L
4
-
1
-1
mol
s ) are shown as parenthetical numbers without
this point in mind, the limited reaction scheme applicable
under the conditions employed here, and which therefore
involves the recycling of B and A (and not A and MTO),
is given in Scheme 2. This will be the basis on which
the molecular mechanism will be discussed.
further identification):
Kin etic Effects of Su bstitu en ts. The major issue
is to infer from these and other data how the key step,
oxygen transfer from B to the alkene, occurs. We are
proposing that a peroxidic oxygen is attacked by the
electron-rich double bond, to which it is ultimately
transferred. This intermediate (or transition state, we
cannot be sure) is shown by the diagram suggested for it
in Scheme 2. Indeed, we have already verified by 18O-
labeling that exactly this happens in the MTO-catalyzed
A somewhat analogous comparison, this time in the
semiaqueous medium, is shown here, although it is less
exact since it takes CH
n-C
3
as equivalent, in effect, to
6
13
H :
2 2 3
reaction between H O and PhSCH : the sulfoxide
derives its oxygen neither from a RedO group nor from
a water molecule in the semiaqueous medium but from
the peroxide.14 We have not carried out the labeling
experiment here, but none of the data give us reason to
suspect a change in mechanism.
The effect of other electron-attracting groups can be
seen by comparing the low rates of those molecules with
carbonyl groups (19, 20) and hydroxy groups (14, 15, 21,
and 23) to comparable alkenes that are strictly aliphatic.
The limit of this trend is seen in the rates of the last three
compounds in Table 1. They have very strongly electron-
attracting groups. In these cases the rate constants are
For the families Ar
3 2 2 2 2
P, Ar PR, Ar S, ArSR, R S, ArNMe ,
and ArNH , etc., we have shown numerous instances in
2
which the addition of an electron-donating (or electron-
attracting) substituent on the aryl ring increases (or
decreases) the values of k and k ; likewise an alkyl group
3 4
in place of an aryl group or a hydrogen atom increases
the rate. In all of the systems studied, no exceptions to
this trend have been noted.
too small for evaluation; we place an upper limit k <
4
-3
-1 -1
1
1
0
L mol
s
for the reactions of methyl acrylate (28),
,4-dicyano-2-butene (29), and acrylonitrile (30).
In this diverse set of alkenes, a similar pattern
emerges. This pattern can be seen first by examining
the kinetic effect of substituting an alkyl group for one
(
24) Baumstark, A. L. J . Bioorg. Chem. 1986, 14, 326.
(
25) Baumstark, A. L.; Pilcher, R. S. J . Org. Chem. 1982, 47, 1141.
(26) Baumstark, A. L.; Vasquez, P. C. Tetrahedron Lett. 1983, 123.

Epoxidation of Alkyl-Substituted Alkenes by H
2
O
2
J . Org. Chem., Vol. 61, No. 12, 1996 3975
All of these trends, without exception, show that the
more electron rich an alkene is, the more rapidly it reacts
with B. We also note that, uniformly, the rate of
epoxidation of an alkene by B is 5-10 times slower in
methanol than in aqueous acetonitrile. The same holds
true for styrene;12 as before, we attribute this trend to
the increase in the activity of water and to an increase
in solvent polarity, stabilizing a polar arrangement at
the transition state.
Steric effects seem relatively mild, which allows us to
make the comparisons of electronic effects just given
without reference to such factors. Note, for example,
these rate constants in methanol:
F igu r e 7. Correlation of the rate constants for the epoxidation
of several olefins by B, in 1:1 CH
for dimethyldioxirane in acetone.
3
2
CN-H O at pH 1, with those
0
.987 and 0.982. The strong correlation allows certain
comparisons to be drawn between the structures of the
systems at the transition states, as presented in the next
section.
Str u ctu r e of th e Tr a n sition Sta te. Guided by
previous results for epoxidation, we envisage two possible
And, in any event, the correlation, presented in Figure
, of reaction rate with the number of alkyl groups is
exactly counter to what the trend would have been were
steric effects playing a major factor. Perhaps the most
cautious statement is this: electronic factors outweigh
steric ones.
6
modes by which the olefin might react with the
metalloperoxide: a direct attack3
0-32
of the double bond
on a peroxidic oxygen of B or prior binding of the olefin
to the rhenium, followed by the slower insertion into the
M-O bond to yield a five-membered peroxometalacycle,
which would then rapidly yield the epoxide.33 We dia-
gram the two possibilities as
Consider a pair of alkenes that are 1,1-disubstituted,
â-pinene (4) and methylenecyclohexane (9). Reaction of
the first is about 4-fold faster than the latter. â-Pinene
has a higher angle strain because it is a bicyclic system.
This factor may account for the rate enhancement, since
in the transition state for epoxidation the olefinic carbons
2
3
change hybridization from sp to sp . Similar compari-
sons could be made between norbornylene (7) and cyclo-
hexene (8), with relative rates of 2.4 and 1.8 in the two
solvents, and between norbornylene and trans-4-octene
(15), with relative rates of 7.4.
A Lin ea r -F r ee-En er gy Cor r ela tion . An interesting
analogy can be drawn between two cyclic peroxides, B
and dimethyldioxirane, DMDO.27 The kinetics of epoxi-
From these data and earlier reports, we argue against
the second formulation. The reasons, which offer differ-
ent degrees of cogency, are as follows: (1) The second
mechanism seems suitable for later transition metals and
has been invoked for them.34 An electron-rich metal
favors olefin coordination by back-bonding. (2) No evi-
dence was obtained for any intermediate forming tran-
siently during the repeated NMR scans of any of the 19
alkenes studied by this technique. Although this is not
a definitive argument, seeing that it depends on the rates
of successive steps that would occur along this pathway,
it is suggestive. (3) A five-membered peroxometallacycle
has been isolated only in the instance of Pt and Rh peroxo
complexes with tetracyanoethylene.35 Such intermedi-
ates have not been found for high-valent early transition
metal peroxides of Groups IV-VI.32 The Re(VII) species,
entirely lacking in d electrons, seems likely to fall in the
same category. (4) One piece of evidence for prior binding
of an olefin to a metal peroxide is that the reaction is
inhibited by the presence of basic (coordinating) solvents
and ligands.36 In the present, the reactions of B are
dation with DMDO, eq 5, have been reported for a
significant number of alkenes.2
8,29
The similarities are striking. The epoxidation reac-
tions of both B and DMDO are stereospecific. Both
proceed at rates that increase with the nucleophilicity
of the olefin, that are but slightly sensitive to steric effects
1
2
(
see the report on styrene ), and that are enhanced by
the solvent polarity and an increase in the activity of
+
water in the solution. Moreover, Hammett constants (σ )
are similar: F ) -0.92 (B) and -0.90 (DMDO).
Similar kinetic factors run throughout the entire series.
4
Figure 7 displays a plot of log k (B) versus log k(DMDO).
(
30) Sharpless, K. B.; Townsend, J . M.; Williams, D. R. J . Am. Chem.
This figure records an excellent correlation for (sepa-
rately) seven alkenes and four styrenes. The respective
slopes are 0.84 and 1.18, with correlation coefficients of
Soc. 1972, 94, 295.
(31) Trost, M. K.; Bergman, R. B. Organometallics 1991, 10, 1172.
(
(
(
32) J ørgenson, K. A. Chem. Rev. 1989, 89, 431-458.
33) Mimoun, H. Angew. Chem., Int. Ed. Engl. 1982, 21, 734.
34) Mimoun, H.; Mignard, M.; Brechot, P.; Saussine, L. J . Am.
(
27) Adam, W.; Hadijiarapoglou, L. Top. Curr. Chem. 1993, 164, 45-
Chem. Soc. 1986, 108, 3711.
(35) Sheldon, R. A.; VanDoorn, J . A. J . Am. Chem. Soc. 1975, 94,
115.
(36) Chaumette, P.; Mimoun, H.; Saussine, L.; Fischer, J .; Mistchler,
A. J . Organomet. Chem. 1983, 250, 291.
6
3
2.
(
28) Baumstark, A. L.; McCloskey, C. J . Tetrahedron Lett. 1987, 28,
311.
29) Baumstark, A. L.; Vasquez, L. J . Org. Chem. 1988, 53, 3437.
(

3
976 J . Org. Chem., Vol. 61, No. 12, 1996
Al-Ajlouni and Espenson
enhanced by the change to a more polar solvent (metha-
nol to aqueous acetonitrile) and in the case of styrene12
by the addition of water. Epoxidation reactions of DMDO
are also accelerated by water.29 (5) The very close
parallel with DMDO, for which coordination is not an
option, suggests that the first scheme is the valid one
for B.
peroxo oxygen atoms and the rhenium atom. The data
offer mild support for the spiro transition state. In the
isomers of â-methylstyrene, the phenyl ring and the
olefinic carbon atoms are coplanar and rotation about the
Ph-C bond is hindered owing to π-interaction. Thus
there should not be a significant difference in steric
hindrance between this pair of isomers; the rate ratio is
Allyl alcohols are epoxidized more rapidly than normal
olefins by early transition metal peroxides. An allyl
alcohol may bind to the metal through its OH group prior
to oxygen transfer.37 This feature contradicts our find-
ings: for B, allyl alcohols and allyl ethers are less reactive
than the corresponding olefins. It appears that this is
the only major difference between B and other early
transition metal peroxides. Perhaps the OH group does
not bind; perhaps the electronic factors are dominant.
The geometrical configuration that might represent the
reaction at its transition state is worthy of consideration.
Consider these two geometries, the first a cyclic planar
k
cis/ktrans ) 1.3. In the case of the isomers of 3-hexen-1-
ol, with kcis/ktrans ) 2.5, given the free rotation around
the single bond between the alkyl group and its olefinic
carbon atom, attack should be more sterically hindered
1
1
for R ) n-Pr than for R ) H. The rate ratio of 2.5 is,
however, less striking than that for DMDO,2 where
ratios of 7-9 were found, perhaps because the Re-O
bonds are longer than the C-O bonds, which reduces the
interactions referred to. This geometry also agrees with
the finding that the 1,1-disubstituted alkenes and the 1,2-
disubstituted alkenes have similar reactivities; the sub-
stituents, whether H or alkyl, do not greatly affect the
attack of the olefin.
8
3
8
(
or “butterfly”) one and the other a “spiro” arrange-
3
9,40
ment
Ack n ow led gm en t. This research was supported by
the U.S. Department of Energy, Office of Basic Energy
Sciences, Division of Chemical Sciences, under contract
W-7405-Eng-82. We gratefully acknowledge the as-
sistance of Prof. Kenneth Kustin toward implementing
the thermometric technique used in these studies.
J O951774E
(37) NIST MS Library from Magnum, 1990, Finnegan MAT: San
J ose, CA.
(
(
38) Bartlett, P. D. Rec. Chem. Prog. 1950, II, 47.
39) Plesnicar, B. in The Chemistry of Peroxides; Wiley: New York,
1
983; p 521.
40) Applications of Molecular Orbital Theory in Organic Chemistry;
(
In the spiro form, the olefinic carbon atoms are
coplanar with the three-membered ring formed by the
Sharpless, K. B.; Verkoeven, T. R., Eds.; Elsevier: Amsterdam, 1977;
p 221.