Experimental evidence for multiple oxidation pathways in the (salen)Mn-catalyzed epoxidation of alkenes

Article abstract of DOI:10.1002/1521-3765(20020603)8:11<2568::AID-CHEM2568>3.0.CO;2-Z

The substrate electronic effects on the selectivity in the catalytic epoxidation of para-substituted cis stilbenes 2a-i were investigated by using (R,R)-[N,N′-bis(3,5-di-tBu-salicylidene)-1,2-cyclohexanediamine] manganese(III) chloride 1 in benzene as the catalyst with iodosobenzene as the terminal oxidant. A Hammett study of the selectivity results reveals a stronger electrophilic character than previously assumed in the (salen)Mn-catalyzed reaction. In general, the best correlations with the experimental values were obtained by using the Hammett σ⁺ values, which gave ρ = -1.37 for the rate of cisepoxide formation and ρ = -0.43 for the rate of the stepwise process leading to the corresponding trans product. The reaction involves two separate pathways as indicated also by the competitive breakdown of the intermediate on the path to trans epoxide for methoxy-substituted substrates. The asynchronicity in the concerted pathway leading to cis epoxide is apparent for 4-methoxy-4′-nitrostilbene, which yields cis epoxide with 75% ee entirely as a result of electronic effects.

Full text of DOI:10.1002/1521-3765(20020603)8:11<2568::AID-CHEM2568>3.0.CO;2-Z

FULL PAPER
Experimental Evidence for Multiple Oxidation Pathways
in the (salen)Mn-Catalyzed Epoxidation of Alkenes
Christian Linde,*[a] Nordine KoliaÔ,[a] Per-Ola Norrby,[b] and Bjˆrn äkermark[c]
Abstract: The substrate electronic ef-
fects on the selectivity in the catalytic
epoxidation of para-substituted cis stil-
benes 2a i were investigated by using
(R,R)-[N,N'-bis(3,5-di-tBu-salicylidene)-
1,2-cyclohexanediamine]manganese(iii)
chloride 1 in benzene as the catalyst with
iodosobenzene as the terminal oxidant.
A Hammett study of the selectivity
results reveals a stronger electrophilic
character than previously assumed in
the (salen)Mn-catalyzed reaction. In
general, the best correlations with the
experimental values were obtained by
using the Hammett s values, which
gave 1 ˆ À1.37 for the rate of cis-
epoxide formation and 1 ˆ À0.43 for
the rate of the stepwise process leading
to the corresponding trans product. The
reaction involves two separate pathways
as indicated also by the competitive
breakdown of the intermediate on the
path to trans epoxide for methoxy-sub-
stituted substrates. The asynchronicity in
the concerted pathway leading to cis ep-
oxide is apparent for 4-methoxy-4'-nitro-
stilbene, which yields cis epoxide with
75% ee entirely as a result of electronic
effects.
‡
Keywords: asymmetric catalysis
¥
epoxidation
relationships
¥
linear free energy
manganese
¥
¥
reaction mechanisms
Isolated alkenes are believed to react in a concerted
manner (Scheme 1, path A).[6] On the other hand, the
formation of a mixture of cis and trans epoxides in the
(salen)Mn-catalyzed epoxidation of conjugated cis alkenes,[7]
as well as nonlinear behavior in Eyring-type plots of the
Introduction
The selective epoxidation of unfunctionalized alkenes has
been a major challenge in organic synthesis for many years.[1, 2]
The successful development of optically active (salen)Mn
catalysts has provided a useful tool for this transformation.[2, 3]
In particular, conjugated cis-1,2-disubstituted alkenes are
epoxidized with high to excellent enantioselectivity. The high
degree of asymmetric induction displayed by (salen)Mn-
based catalysts has initiated extensive mechanistic studies of
O
+
C
Mn
O
B
A
Mn
the reaction,[2 which, to a large extent, parallels earlier
investigations of porphyrin-based systems.[5]
4]
‡
O
Mn
O
O
B'
Mn
Mn
[a] Dr. C. Linde, N. KoliaÔ
A
Department of Chemistry, Organic Chemistry
Royal Institute of Technology, 10044 Stockholm (Sweden)
Fax : (‡46)8791-2333
E-mail: linde@orgchem.kth.se
O
O
O
Mn
[b] Prof. Dr. P.-O. Norrby
Mn
Mn
Department of Chemistry, Organic Chemistry
Danish Technical University, Building201, Kemitorvet
2800 Lyngby (Denmark)
Scheme 1. Previously proposed reaction paths for (salen)Mn-catalyzed
epoxidation.
Fax : (‡45)4525-2123
E-mail: pon@kemi.dtu.dk
enantioselectivity temperature relationship,[8, 9] indicate a
stepwise process. One proposal for a stepwise mechanism
involves an irreversibly formed radical intermediate that
collapses to syn and anti products (Scheme 1, path B).[3a]
Another proposal involves a manganaoxetane that undergoes
reductive elimination to the syn product or is opened to a
radical intermediate that is transformed into the anti product
[c] Prof. Dr. B. äkermark
Department of Organic Chemistry, Arrhenius Laboratory
Stockholm University, 10691 Stockholm (Sweden)
Fax : (‡46)8154908
E-mail: bjorn.akermark@organ.su.se
Supporting information for this article is available on the WWW under
http://wiley-vch.de/home/chemistry/ or from the author. This contains
the derivation of Hammett expressions and experimental details.
2568
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2568 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 1 1

2568 2573
(Scheme 1, path C).[8, 10] However, experimental evidence has
been presented that cannot be rationalized unambiguously by
any of these mechanistic models. The results from a study of
kinetic isotope effects are not consistent with the manganaox-
etane pathway,[3a] but strong indications that syn-product
formation does not proceed via a radical intermediate were
found in another study by using cyclopropyl-substituted
conjugated alkenes as radical traps.[11] An alternate model is
needed to solve the conflict between the experimental results
and these mechanistic proposals. Based upon a computational
study, we presented instead a model, in which the timing of
spin change along the reaction coordinate determines the
choice of reaction path and, thus, the syn anti-product
distribution in the reaction.[12] This model provides a rationale
for the contradictory experimental observations. It also
accommodates the suggestion of Jacobsen and co-workers
that the cis and trans epoxides are formed following a
common selectivity-determining first step (Scheme 1, path
B'),[3a] since the results from our study suggest that the spin
change might occur either before or after the rate-limiting
step.
N
O
N
O
Mn
Cl
tBu
tBu
tBu
(R,R)-1
PhIO, benzene
tBu
X
X
Y
O
Y
X
Y
2a-i
3a-i
NO2 NO2
NO2
H
Br
Cl
NO2 MeO
Br
H
a
b
c
d
e
f
g
h
i
H
H
H
H
MeO
MeO
MeO MeO
Scheme 2. Reagents and products in the Hammett study.
Table 1. Epoxidation of para-substituted cis-stilbenes by using (R,R)-1 and
PhIO in benzene.[14]
Results and Discussion
[a, b]
Substrate krel
yield[b, c] cis:trans[b] ee cis[d] [%] ee trans[d] [%]
[f]
1
2
3
4
5
6
7
8
9
2a
2b
2c
2d
2e
2 f
2g
2h
2i
0.4
nd[e]
96
99
98
98
84
85
72
64
4:96
14:86
25:75
27:73
29:71
59:41
58:42
69:31
86:14
85
81
75
77
76
75
68
65
26
We have undertaken an experimental study of the electronic
effects in the title reaction as part of our ongoing effort to
elucidate the mechanism of the reaction. We have chosen
para-substituted cis-stilbenes as model substrates in our study.
These are ideal substrates since the electronic properties at
both carbons of the double bond can be modified by the para-
substituents, while similar steric influence is maintained at
both ends. Furthermore, the gross electronic properties of the
two carbons are similar and allow the formation of a benzylic
radical intermediate irrespective of the regioselectivity in the
attack of the active oxidant. Separate Hammett expressions
for the relative rates of cis- and trans-epoxide formation from
these substrates can be obtained through competitive experi-
ments. Conclusions can be drawn about the asynchronicity of
the attack as well as the timing of the bifurcation to syn and
anti-addition products along the reaction path, for example, if
the common transition state postulate is feasible (Scheme 1,
path B'). The substrate electronic effects on the title reaction
have been investigated in previous Hammett studies. How-
ever, none of the studies include both competitive experi-
ments with a reference substrate and determination of relative
formation rates of syn- and anti-addition products individu-
ally.[7, 13] Our study proves this information to be critical for a
detailed investigation of the mechanism of the reaction.
The reagents utilized in the study are shown in Scheme 2.[14]
The substituents were chosen to cover a wide range of
electronic properties and include symmetrically disubstituted
compounds (2a and 2i) as well as push pull systems (2 f). The
observed rates of formation of epoxide relative to that of the
unsubstituted product, 3c, as well as yields and selectivities,
are shown in Table 1. The relative overall rate constants for
epoxide formation are within the same order of magnitude for
all substrates, in agreement with previous reports by Kochi
et al.[7] Qualitatively, the electronic influence on the reaction
0.7
1.0
1.0
1.1
1.5
1.7
1.7
2.9
nd[e]
[f]
25
23
75
61
48
[f]
[a] The relative rate constants for epoxidation were determined in
competing experiments by using 0.2 equiv of PhIO with 0.5 equiv each of
the substituted and unsubstituted cis stilbenes. [b] Determined by 1H NMR
spectroscopy and chiral HPLC on
a Daicel ChiralcelOD column.
[c] Determined after addition of 1equiv of oxidant. [d] Determined
by chiral HPLC on a Daicel ChiralcelOD column. [e] Not determined.
[f] Not chiral.
rates fits into a general reaction mechanism, in which the
alkene reacts with an electrophilic oxidation reagent.[15]
We have performed a systematic analysis of the results in
Table 1by using the Hammett equation for linear free energy
relationships [Eq. (1)]. Several s scales have been tested, but
log(kX/k0) ˆ 1sX
(1)
‡
all conclusions in this work are based on the s values since
they consistently gave the best correlation with the exper-
imental data.[16] The better fit could be obtained intuitively by
considering the similarity in magnitude of the changes in the
overall rates of epoxide formation krel for the symmetrically
substituted substrates 2i and 2a versus 2c (Table 1, entries 1,
3, and 9). These rate changes require the s values to be similar
in magnitude, but opposite in sign, for the methoxy and nitro
‡
substituents. This is the case for the s values (À0.78 and
‡0.79, respectively), but not for the other s scales.[16]
Chem. Eur. J. 2002, 8, No. 11
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2569 $ 20.00+.50/0
2569

C. Linde et al.
FULL PAPER
We first made the simple assumption that the reaction rates
of formation of the individual cis- and trans-epoxide isomers
both vary with the average of the Hammett constants of the
two substituents, that is, that the transition state is perfectly
synchronous, and the two possible alkene face attacks are
influenced equally by the two substituents [Eq. (2)].[16] The
relative rates of formation of the isomeric epoxides [(kXY/k0)cis
To analyze the rate variations in more detail, we postulated
a kinetic model involving a concerted and a stepwise pathway
as shown in Scheme 3. We assumed that the first step of the
radical process is irreversible and rate determining since no
isomerized alkene was observed in the reaction after partial
‡
and (kXY/k0)trans] are plotted against the average s values in
á
A
B
B
A
Figure 1.[17] The rates are presented also in Table 2 together
+
‡
O
M
O
M
with the s values employed.
A
B
kc
log(kXY/k0) ˆ 1(sX ‡ sY)/2
(2)
kr
A
B
O
A
B
B
A
ignored
O
M
O
M
+
A
B
A
O
M
B
O
M
A
B
O
+
Scheme 3. Postulated paths and constants employed in the kinetic model.
conversion. We also ignored the formation of cis product in
À
the radical process since the equilibrium for C C bond
rotation is shifted far toward the intermediate leading to
trans product (Table 1, entry 1). Our previous radical trapping
study provided additional support for this assumption.[11] The
complete suppression of anti-addition products from such
traps strongly indicates that the ring closure of the inter-
mediate radical to epoxide is substantially slower than trap
opening. The high yields of nonisomerized epoxides in the
same study show that the syn addition does not go through a
radical intermediate. Thus, cis product is formed only by
concerted addition (rate constant kc), and the rate of trans-
product formation is determined by the initial formation of
the radical intermediate (rate constant kr). Thereby, the rate
ratio between the concerted and stepwise processes (kc/kr) is
proportional to the cis trans-epoxide product ratio since the
two products are formed from the same alkene reactant. Note
that no assumption was made at this point whether syn and
anti products are formed via a common transition state
(Scheme 1, path B or B').
‡
Figure 1. Hammett plot based on average
[Eq. (2)].
s
substituent constants
Table 2. Hammett constants and relative rates of formation of single
products.[17]
‡
‡
‡
‡
Product
X
Y
sX
sY
(sX ‡ sY†/2 (k/k0)cis (k/k0)trans
1
2
3
4
5
6
7
8
9
3a
3b
3c
3d
3e
3f
3g
3h
3i
NO2 NO2
0.79
0.79
0
0.15
0.11
0.79
0.79
0.395
0
0.075
0.055
0.005
À 0.315
À 0.39
À 0.78
0.064
0.392
1
0.512
0.803
NO2
H
Br
H
H
H
H
0
0
0
0
1
1.080
1.276
3.540
3.944
4.692
9.976
0.973
1.041
0.820
0.952
0.703
0.541
Cl
NO2 MeO
0.79 À 0.78
Br
H
MeO
MeO
0.15 À 0.78
0
À 0.78
MeO MeO À 0.78 À 0.78
The simplified description of the electronic effects by using
an effective Hammett constant expressed as the average of
the two para-substituents [Eq. (2)] is valid only for a perfectly
synchronous transition state. Perfect synchronicity is possible
in a concerted process but unequal influence from the
substituents would be expected in an asynchronous process,
which could be either stepwise or concerted. Therefore, the
regular Hammett expression was modified to describe a
reaction with nonequal influence from the para-substituents
of an unsymmetrical alkene. The possibility of two sterically
identical but electronically dissimilar regiochemical attacks
leads to the bis-exponential in Equation (3),[16] in which the
It is shown in Figure 1that formation of the cis epoxides is
strongly dependent on the substituents, whereas no clear
effect can be seen on the formation of trans epoxide. We
interpreted these observations as a first indication that cis-
and trans-epoxide formation might occur following separate
pathways. It is clear that the rate of cis-product formation is
not influenced equally by the two substituents in cis-3 f as it is
formed 3.5 times faster than cis-3c, although the average
‡
s values for the two substrates are similar (Table 2, entries 3
and 6). Apparently, the formation of cis-3 f follows a path that
maximizes the effect of the activating methoxy substituent,
and this gives a higher relative rate of formation than would
be expected from the simplified expression in Equation (2).
‡(1Àf)sY
]
‡(1Àf)sX
log(kXY/k0) ˆ log[(101[fs
‡ 101[fs
])/2]
(3)
X
Y
2570
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2570 $ 20.00+.50/0 Chem. Eur. J. 2002, 8, No. 1 1

(Salen)Mn-Catalyzed Epoxidation of Alkenes
2568 2573
fraction f describes the relative influence of each substituent
in each regiochemical attack. This value is to be determined in
nonlinear fitting together with the reactivity coefficient 1.
Equation (3) is solved easily for the two unknowns (1 and f)
by standard numerical methods for over-determined non-
linear equation systems.[18] The experimental log(kXY/k0) val-
ues for cis-epoxide formation were reproduced within exper-
radical character were developed in the transition state as all
substituents stabilize a radical better than hydrogen. Non-
linear behavior could also be indicative of a change in the
rate-determining step. In our case, such a change would imply
reversible formation of the intermediate radical (Scheme 3)
and would result in cis trans isomerization of the alkene.
However, no stilbene isomerization was observed during the
reactions. We recognized instead that the mass balance in
Table 1is incomplete for the methoxy-substituted stilbenes,
and this implies that the rate of stilbene conversion is
substantially faster than the rate of epoxide formation. The
decrease in epoxide yields for electron-rich stilbenes was
interpreted as an indication of a competing pathway leading
to decomposition of the radical intermediate to non-epoxide
products for these substrates.[20] To test this idea, we postu-
lated that the rate of formation of the radical intermediate is
equal to the rate of disappearance of stilbene, corrected for
the rate of formation of cis epoxide, which was assumed to be
formed following the concerted pathway (Scheme 3).[17] The
resulting data are presented in Table 3. Now a fair fit was
achieved between experimental and calculated rate constants,
‡
imental error (r2 ˆ 0.98, Table 3) by using the s values,
whereas an attempted fit to other s scales resulted in much
‡
Table 3. The logarithmic values (log(kXY/k0)), calculated based on s for
the experimental and calculated relative rates of epoxide formation and
radical intermediate formation.[a]
cis
trans
rad
Product
exp
Eq. (3)
sEff
exp
exp
Eq. (3)
sEff
3a
3b
3c
3d
3e
3f
3g
3h
3i
1
À 1.194 À 1.082
À 0.407 À 0.376
0.790 À 0.291 À 0.297 À 0.338
0.274 À 0.095 À 0.081 À 0.147
0.790
0.344
0
0
0
0
0
0
0
0.033 À 0.096
0.070 À 0.012 À 0.006 À 0.0310.073
0.106 À 0.072
0.052 0.018 0.054
0.024 À 0.023
0.549
0.596
0.6710.696
0.999
0.504 À 0.368 À 0.086
0.651 À 0.476 À 0.021
0.074
0.125
0.194
0.428
0.079 À 0.186
0.165 À 0.385
0.188 À 0.439
0.334 À 0.780
‡
again using the s values. The 1 value is low, only À0.43, and
À 0.508 À 0.153
the degree of asynchronicity is strong (f ˆ 0.91).
1.068 À 0.780 À 0.267
[b]
À 1.37
À 0.43
The nonlinear enantioselectivity temperature dependence
observed by Katsuki et al.,[8] as well as the results from the
study of radical traps,[11] are accommodated by the reaction
mechanism in Scheme 3. In the study of Katsuki et al., the
amounts of epoxides formed via a radical intermediate cannot
f
r2
0.87
0.98
0.91
0.94
[a] In this analysis, it is postulated that a radical intermediate is formed
from the fraction of the alkene that is not converted to cis epoxide. This
radical intermediate is only partly converted to trans epoxide; see text.
[b] No positive correlation was obtained.
À
be estimated since either C C bond rotation is not permitted
due to the cyclic structures (1,2-dihydronaphthalene and 1,3-
cyclooctadiene), or the alkenes are monosubstituted (styrene
and p-nitrostyrene). On the other hand, it is predicted from
the reaction mechanism in Scheme 3 that nonlinear diaste-
reoselectivity temperature dependence and nonlinear enan-
tioselectivity temperature dependence for the anti-addition
product would be expected in a system, in which the rates of
the competing pathways are similar but with different
temperature dependence.
The results obtained in this study are not easily depicted in
a classical Hammett sense, since they depend on two
interacting substituents for each stilbene. For illustration
purposes, we have plotted in Figure 2 the experimental
relative rates of cis- and trans-epoxide formation and the
postulated rate of radical intermediate formation against an
™effective Hammett substitution constant∫, sEff [Eq. (4)]. The
constant is calculated from Equation (3) as log(kXY/k0)calcd/1
for the two successful fits in Table 3 by using f ˆ 0.87 for
cis epoxide and f ˆ 0.91for trans epoxide and radical forma-
tion.
worse correlation. The adjusted parameters indicate a sub-
stantial influence of the substituents on the rate of formation
of cis-stilbene oxide (1 ˆ À1.37) and a stronger electrophilic
character than has been assumed before in the (salen)Mn-
catalyzed reaction.[19] It is clear that the formation of
cis epoxide is an asynchronous process (f ˆ 0.87), with the
dominant influence coming from the more electron-donating
substituent. Analysis of the push pull system 2 f provides an
additional confirmation about the asynchronicity of the
reaction. The electronic effects of the two substituents in this
substrate would be cancelled in a symmetric transition state
‡
(equal magnitude but opposite sign of the s values). How-
ever, the asynchronicity of the transition state allows the more
activating substituent to dominate the reactivity in cis-epoxide
formation, and thus it explains the high relative rate of cis-3 f
formation (Table 2, entry 6). Interestingly, high enantioselec-
tivity is observed in the formation of this product even though
the substituents are sterically similar (Table 1, entry 6), and
the selectivity indicates that electronic effects play an
important role in determining the alkene face selectivity in
syn addition.
log(kXY/k0) ˆ 1sEff {sEff ˆ log(kXY/k0)calcd/1 [from Eq. (3)]}
(4)
The analysis of the experimental rates of trans-epoxide
formation is less tractable, and the rates could not be
reproduced by using Equation (3).[16] It is evident from the
data in Table 2, and also from Figure 1, that the rates of trans-
epoxide formation show a nonlinear Hammett behavior, for
which all substituents have a negative influence on the rates.
This trend is opposite to what had been expected if substantial
The rate of concerted syn addition and the rate of the
postulated radical intermediate formation both correlate well
‡
with the Hammett s values. The transition states for both
processes are highly asynchronous where the buildup of
charge on one carbon of the alkene in the rate-determining
step is stabilized by through resonance from the closest para-
substituent. The concerted pathway shows a large electronic
Chem. Eur. J. 2002, 8, No. 11
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2571 $ 20.00+.50/0
2571

C. Linde et al.
FULL PAPER
Acknowledgements
We gratefully acknowledge the Danish Technical Research Council, the
Swedish Natural Science Research Council, the Swedish Research Council
for Engineering Science, and the Carl Trygger Foundation for financial
support. C.L. thanks the Royal Institute of Technology for a research
¬
scholarship. N.K. thanks the ERASMUS program and Universite d×Aix-
Marseilles III for financial support. We also thank Professor Waldemar
Adam, Dr. Declan Gilheany, and Dr. Peter Brandt for valuable discussions.
[1] K. A. J˘rgensen, Chem. Rev. 1989, 89, 431 458.
[2] a) T. Katsuki, ™Asymmetric Epoxidation of Unfunctionalized Olefins
and Related Reactions∫ in Catalytic Asymmetric Synthesis, 2nd ed.,
(Ed.: I. Ojima), Wiley, New York, 2000, pp. 287 326; b) E. N.
Jacobsen, ™Asymmetric Catalytic Epoxidation of Unfunctionalized
Olefins∫ in Catalytic Asymmetric Synthesis (Ed.: I. Ojima), VCH, New
York, 1993, pp. 159 202.
[3] a) M. Palucki, N. S. Finney, P. J. Pospisil, M. L. G¸ler, T. Ishida, E. N.
Jacobsen, J. Am. Chem. Soc. 1998, 120, 948 954; b) T. Katsuki, J. Mol.
Catal. A: Chem. 1996, 113, 87 107; c) T. Katsuki, Coord. Chem. Rev.
1995, 140, 189 214.
Figure 2.
constants [Eq. (4)].
Experimental
rates
versus
effective
substituent
[4] a) C. T. Dalton, K. M. Ryan, V. M. Wall, C. Bousquet, D. G. Gilheany,
Top. Catal. 1998, 5, 75 91; b) T. Linker, Angew. Chem. 1997, 109,
2150 2152; Angew. Chem. Int. Ed. Engl. 1997, 36, 2060 2062.
[5] See, for example: a) M. Sono, M. P. Roach, E. D. Coulter, J. H.
Dawson, Chem. Rev. 1996, 96, 2841 2887; b) D. Ostovic, T. C. Bruice,
Acc. Chem. Res. 1992, 25, 314 320.
[6] H. Fu, G. L. Look, W. Zhang, E. N. Jacobsen, C.-H. Wong, J. Org.
Chem. 1991, 56, 6497 6500.
[7] K. Srinivasan, P. Michaud, J. K. Kochi, J. Am. Chem. Soc. 1986, 108,
2309 2320.
[8] T. Hamada, T. Fukuda, H. Imanishi, T. Katsuki, Tetrahedron 1996, 52,
515 530.
[9] H. Buschmann, H.-D. Scharf, N. Hoffmann, P. Esser, Angew. Chem.
1991, 103, 480 518; Angew. Chem. Int. Ed. Engl. 1991, 30, 477 515.
[10] P.-O. Norrby, C. Linde, B. äkermark, J. Am. Chem. Soc. 1995, 117,
11035 11036.
[11] C. Linde, M. Arnold, P.-O. Norrby, B. äkermark, Angew. Chem. 1997,
109, 1802 1803; Angew. Chem. Int. Ed. Engl. 1997, 36, 1723 1725.
[12] C. Linde, B. äkermark, P.-O. Norrby, M. Svensson, J. Am. Chem. Soc.
1999, 121, 5083 5084.
[13] E. N. Jacobsen, L. Deng, Y. Furukawa, L. E. MartÌnez, Tetrahedron
1994, 50, 4323 4334.
dependence (1 ˆ À1.37) indicative of a late transition state.
The transition state of the radical pathway is much earlier
(1 ˆ À0.43) and has no significant radical character as
indicated by the influence from the para-substituents. The
stepwise anti-addition process displays a nonlinear behavior,
interpreted as a competitive breakdown of the radical
intermediate into non-epoxide products when electron-do-
nating substituents are present on the stilbene.
Apparently cis- and trans-epoxide formation in the title
reaction follow separate pathways. We cannot exclude the
possibility of multiple oxidizing species in the reaction
mixture, in which each species follows preferentially one of
the reaction pathways.[21] These oxidants would show reaction
rates, which have very different electronic dependence. In
such a case the oxidants must be in rapid equilibrium under
catalytic conditions, since the product distribution is affected
significantly by the electronic properties of the substrate. The
current results could be interpreted also as an attack on the
alkene of a common electrophilic oxidizing species, but with
subsequent branching to a radical intermediate before the
transition state for the concerted syn addition, for example, by
spin-surface crossing. In any case the common transition state
hypothesis (Scheme 1, path B') is excluded, and the fate of the
alkene is sealed already in the first irreversible step.
[14] The para-substituted cis-stilbenes and (R,R)-1 were prepared accord-
ing to published procedures: a) H. Yamataka, K. Nagareda, K. Ando,
T. Hanafusa, J. Org. Chem. 1992, 57, 2865 2869; b) W. Zhang, E. N.
Jacobsen, J. Org. Chem. 1991, 56, 2296 2298; c) J. F. Larrow, E. N.
Jacobsen, J. Org. Chem. 1994, 59, 1939 1942.
[15] K. G. Rasmussen, D. S. Thomsen, K. A. J˘rgensen, J. Chem. Soc.
Perkin Trans. 1 1995, 2009 2017.
[16] Only the simplified kinetic expressions for the successful fits to the
‡
experimental data by using the s values are presented in this paper.
The derivation of these expressions from the complete kinetic
expressions are available in the Supporting information together with
‡
the various s values used in the evaluation. Standard s and s values
for all substituents were obtained from a) M. B. Smith, J. March in
March×s Advanced Organic Chemistry, 5th ed., Wiley Interscience,
New York, 2001. Two different types of s ¥ values were tested: the
Creary scale: b) X. Creary, M. E. Mehrsheikh-Mohammadi, S. McDo-
nald, J. Org. Chem. 1987, 52, 3254 3263; and the Jackson scale: c) H.
Agirbas, R. A. Jackson, J. Chem. Soc. Perkin Trans. 2 1983, 2, 739
742; d) R. A. Jackson, J. Organomet. Chem. 1992, 437, 77 83.
Experimental Section
Experimental details and analytical data are available in the Supporting
information. General procedure for epoxidation: (R,R)-1 (12 mg,
0.02 mmol, 4 mol%), para-substituted stilbene 2 (0.5 mmol, 1equiv), and
iodosobenzene (110 mg, 0.5 mmol, 1 equiv) were stirred in benzene (4 mL)
for 12 hours at room temperature. The catalyst residues were separated on
a short silica column. The cis trans ratio and enantiomeric purity were
determined by 1H NMR spectroscopy and chiral HPLC on a Daicel
Chiralcel-OD column by using IPA/IB (isopropanol/hexanes). Competi-
tion experiments were performed starting with a 1:1 mixture of unsub-
stituted and substituted cis-stilbenes with oxidant (0.2 equiv).
[17] The relative rate constants were calculated from the experimental
data in Table 1by using the following expressions; kcis ˆ kc ˆ krel
Â
(%cis)/100, ktrans ˆ krel  (%trans)/100, kr ˆ krel  100/(%yield) À kc.
[18] The ™solver∫ in Microsoft Excel98 for Macintosh was used to obtain
the values of 1 and f giving the smallest sum of squares of deviations
between experimental and calculated log(krel).
[19] In a previous report by Jacobsen et al. (see: ref. [13]), correlation was
found between the relative ratio of cis- and trans-epoxide formation
2572
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2572 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 1 1

(Salen)Mn-Catalyzed Epoxidation of Alkenes
2568 2573
[log(kcis/ktrans) ˆ 1s] and the standard s values. However, the individ-
ual rates of formation of the individual epoxide isomers could not be
determined since no competitive experiments were performed against
the unsubstituted substrate. We predict that such measurements would
lead to conclusions similar to those presented here.
Angew. Chem. 2001, 113, 2131 2134; Angew. Chem. Int. Ed. 2001, 40,
2073 2076; Y. G. Abashkin, J. R. Collins, S. K. Burt, Inorg. Chem.
2001, 40, 4040 4048); or Lewis acid activation of the stoichiometric
oxidant by coordination to a manganese(iii) complex in competition
with an oxomanganese(v) complex (W. Adam, personal communica-
tion. See also: J. P. Collman, A. S. Chien, T. A. Eberspacher, J. I.
Brauman, J. Am. Chem. Soc. 2000, 122, 11098 11100; W. Adam, C.
Mock-Knoblauch, C. R. Saha-Mˆller, M. Herderich, J. Am. Chem.
Soc. 2000, 122, 9685 9691; K. P. Bryliakov, D. E. Babushkin, E. P.
Talsi, J. Mol. Catal. A: Chem. 2000, 158, 19 35; W. Adam, K. J.
Roschmann, C. R. Saha-Mˆller, Eur. J. Org. Chem. 2000, 3519 3521;
D. Feichtinger, D. A. Plattner, Chem. Eur. J. 2001, 7, 591 599).
[20] The cis trans ratio does not increase during the reaction. Therefore,
the decrease in trans-epoxide formation cannot be explained by
further oxidation of the trans epoxide to secondary products. Instead,
use of excess oxidant leads to preferential loss of cis epoxide.
[21] Current suggestions of the nature of these oxidants involve oxoman-
ganese(v) complexes in different spin states (see: ref. [12]; T.
Strassner, K. N. Houk, Org. Lett. 1999, 1, 419 421; L. Cavallo, H.
Jacobsen, Angew. Chem. 2000, 112, 602 604; Angew. Chem. Int. Ed.
2000, 39, 589 592; H. Jacobsen, L. Cavallo, Chem. Eur. J. 2001, 7,
800 807; J. El-Bahraoui, O. Wiest, D. Feichtinger, D. A. Plattner,
Received: January 2, 2002 [F3773]
Chem. Eur. J. 2002, 8, No. 11
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0811-2573 $ 20.00+.50/0
2573