The mechanistic basis for electronic effects on enantioselectivity in the (salen)Mn(III)-catalyzed epoxidation reaction

Article abstract of DOI:10.1021/ja973468j

Enantioselectivity in the (salen)Mn-catalyzed asymmetric epoxidation reaction correlates directly with the electronic properties of the ligand substituents, with complexes bearing electron-donating substituents affording highest ee's. Several lines of evi

Full text of DOI:10.1021/ja973468j

948
J. Am. Chem. Soc. 1998, 120, 948-954
The Mechanistic Basis for Electronic Effects on Enantioselectivity in
the (salen)Mn(III)-Catalyzed Epoxidation Reaction
Michael Palucki, Nathaniel S. Finney, Paul J. Pospisil, Mehmet L. G u1 ler,
Toyohisa Ishida, and Eric N. Jacobsen*
Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
ReceiVed October 3, 1997
Abstract: Enantioselectivity in the (salen)Mn-catalyzed asymmetric epoxidation reaction correlates directly
with the electronic properties of the ligand substituents, with complexes bearing electron-donating substituents
affording highest ee’s. Several lines of evidence point to a single factorscontrol of the position of the transition
state along the reaction coordinatesas being responsible for the electronic effects on enantioselectivity. Analysis
of the epoxidation of cis-â-deuteriostyrene reveals that electron-rich catalysts display a more pronounced
q
secondary inverse isotope effect than electron-deficient catalysts. A strong correlation between ∆∆H and the
electronic character of the catalyst is also observed. The conclusion that enantioselectivity is tied to the position
of a transition state along the reaction coordinate may hold general implications for the design of asymmetric
catalysts, particularly those that effect reactions without substrate precoordination.
Introduction
Scheme 1
Effective stereochemical communication between substrate
and catalyst is essential for attaining high enantioselectivities
in asymmetric catalytic reactions. Enzymatic processes achieve
this, at least in part, by inducing substrate precoordination to
catalyst prior to reaction, thereby minimizing the degrees of
freedom in the critical transition state and maximizing selectiv-
ity-determining interactions between the catalyst’s asymmetric
1
environment and the substrate. As might be anticipated, many
of the most successful nonenzymatic asymmetric catalyst
systems also operate on this principle: archetypal examples of
2
such effective substrate-directed catalytic reactions include
ity undergoing transformation. The (salen)Mn-catalyzed olefin
epoxidation is among the most useful and widely applicable
reactions that do not involve substrate precoordination. With
3
rhodium-catalyzed asymmetric hydrogenation, the Sharpless
epoxidation of allylic alcohols, and the Noyori Ru(BINAP)-
catalyzed hydrogenation of functionalized olefins and ketones.
In contrast to this enzymatic motif, a number of practical
and effective asymmetric catalyst systems have been discovered
that do not require a specific precoordinating group in the
substrate in order to impart high enantioselectivities. Such
systems offer the potential advantage of enhanced generality,
with reaction scope limited only by the nature of the functional-
4
6b
5
the appropriate choice of ligand and reaction conditions, it has
proven effective for virtually every class of unfunctionalized
7
conjugated olefin. Scheme 1 illustrates the simplest mechanism
for this reaction which is consistent with the available data. In
this reaction epoxidation proceeds through two sequential C-O
bond-forming steps: addition of substrate alkene to a (salen)-
Mn(V) oxo species generates a radical intermediate, which in
turn partitions between collapse and rotation/collapse processes
to provide the observed mixture of cis and trans epoxides.8
6
(
1) Walsh, C. Enzymatic Reaction Mechanisms; W. H. Freeman and
Company: New York, 1979.
2) For a definitive discussion of substrate-directed reactions, see: (a)
Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. ReV. (Washington, D.C.)
993, 93, 1307.
(
In 1991, we reported the observation of dramatic catalyst
electronic effects on the enantioselectivity of (salen)Mn-
1
(
3) Halpern, J. Science 1982, 217, 401.
(4) Johnson, R. A.; Sharpless, K. B. In Catalytic Asymmetric Synthesis;
(7) (a) Jacobsen, E. N. In ComprehensiVe Organometallic Chemistry II;
Wilkinson, G., Stone, F. G. A., Abel, E. W., Hegedus, L. S., Eds.;
Pergamon: New York, 1995; Vol. 12, Chapter 11.1. (b) Katsuki, T. Coord.
Chem. ReV. 1995, 140, 189.
(8) (a) Finney, N. S.; Pospisil, P. J.; Chang, S.; Palucki, M.; Konsler, R.
G.; Hansen, K. B.; Jacobsen, E. N. Angew. Chem., Int. Ed. Engl. 1997, 36,
1720. (b) Jacobsen, E. N.; Deng, L.; Furukawa, Y.; Mart ´ı nez, L. E.
Tetrahedron 1994, 50, 4323. (c) Hosoya, N.; Hatayama, A.; Yanai, K.;
Fujii, H.; Irie, R.; Katsuki, T. Synlett. 1993, 641. (d) Srinivasan, K.; Michaud,
P.; Kochi, J. K. J. Am. Chem. Soc. 1986, 108, 2309. For an alternative
mechanistic interpretation, see: (e) Linde, C.; Arnold, M.; Norrby, P.-O.;
Akermark. B. Angew. Chem., Int. Ed. Engl. 1997, 36, 1723 and references
therein.
Ojima, I., Ed.; VCH: New York, 1993; Chapter 4.1.
5) Takaya, H.; Ohta, T.; Noyori, R. In Catalytic Asymmetric Synthesis;
Ojima, I., Ed.; VCH: New York, 1993; Chapter 1.
6) For example, asymmetric dihydroxylation: (a) Kolb, H. C.; Van-
Nieuwenhze, M. S.; Sharpless, K. B. Chem. ReV. (Washington, D.C.) 1994,
4, 2483. Asymmetric epoxidation: (b) Jacobsen, E. N. In Catalytic
(
(
9
Asymmetric Synthesis; Ojima, I., Ed.; VCH: New York, 1993; Chapter 4.2.
Asymmetric aziridination: (c) Li, Z.; Conser, K. R.; Jacobsen, E. N. J.
Am. Chem. Soc. 1993, 115, 5326. (d) Evans, D. A.; Faul, M. M.; Bilodeau,
M. T.; Anderson, B. A.; Barnes, D. M. J. Am. Chem. Soc. 1993, 115, 5328.
Asymmetric cyclopropanation: Doyle, M. P. In Catalytic Asymmetric
Synthesis; Ojima, I., Ed.; VCH: New York, 1993; Chapter 3.
S0002-7863(97)03468-9 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/08/1998

(Salen)Mn(III)-Catalyzed Epoxidation Reaction
J. Am. Chem. Soc., Vol. 120, No. 5, 1998 949
Scheme 2
Figure 1. Schematic energy diagram illustrating the proposed effect
of ligand substituents on the (salen)Mn-catalyzed epoxidation reaction.
More electron-donating substituents stabilize the Mn(V)oxo relative
to the Mn(IV) radical intermediate resulting in a later, more product-
like transition state.
plexes provide strong support for the mechanistic interpretation
represented schematically in Figure 1.
9
catalyzed epoxidation of cis-disubstituted olefins, and since that
discovery, the importance of electronic effects in asymmetric
1
0
catalytic reactions has been increasingly appreciated. In the
case of the (salen)Mn-catalyzed epoxidation, electron-donating
substituents on the ligand were found to lead to higher levels
of asymmetric induction, while electron-withdrawing substitu-
Results and Discussion
1
. Correlation of Enantioselectivity and the Electronic
Character of the Catalyst. One of the appealing features of
salen-based catalysts is that the ligands may be tuned both
sterically and electronically in a synthetically straightforward
manner by variation of the corresponding diamines and salicyl
aldehyde ligand precursors (Scheme 2). While the first
investigations on the (salen)Mn-catalyzed asymmetric epoxi-
dation established that steric bulk at the 3,3′-position of the salen
ligand was essential in order to achieve high enantioselectivi-
9
ents led to decreased enantioselectivity. In the original report,
we suggested that these effects might be interpreted according
to a Hammond Postulate argument, wherein ligand substituents
influence enantioselectivity by modulating the reactivity of the
high-valent (salen)Mn oxo intermediate. Thus, electron-
withdrawing substituents were proposed to lead to a more
reactive (salen)Mn oxo intermediate, which adds to olefin in a
comparatively early transition state and affords lower levels of
enantioselectivity; conversely, electron-donating groups attenu-
ate the reactivity of the oxo species, leading to a comparatively
late transition state and concomitantly higher enantioselectivity
11
ties, subsequent studies revealed that variations in the elec-
tronic properties of the substituents at the 5,5′-position of the
9
catalysts can have an equally profound effect.
To quantify the influence of the 5,5′-substituent on the
enantioselectivity of the epoxidation reaction, we evaluated the
catalysts depicted in Scheme 2 in the epoxidation of each of
three model substrates: 2,2-dimethylchromene (4), a member
of a synthetically important class of olefins that is epoxidized
with particularly high enantioselectivity by (salen)Mn catalysts;12
cis-â-methylstyrene (5), an aryl-substituted alkene which is
among the most commonly employed model substrates for
enzymatic and nonenzymatic epoxidation systems; and cis-2,2-
dimethyl-3-hexene (6), a nonconjugated olefin. Epoxidation of
(Figure 1).
Herein we present a detailed analysis of these catalyst
electronic effects. The diastereo- and enantioselectivities, kinetic
isotope effects, and temperature profiles of the epoxidation
reaction catalyzed by a series of substituted (salen)Mn com-
(
9) (a) Jacobsen, E. N.; Zhang, W.; G u¨ ler, M. L. J. Am. Chem. Soc.
991, 113, 6703.
10) (a) Rajanbabu, T. V.; Casalnuovo, A. L. J. Am. Chem. Soc. 1992,
14, 6262. (b) Inoguchi, K.; Sakuraba, S.; Achiwa, K. Synlett 1992, 169.
c) Nishiyama, H.; Yamaguchi, S.; Kondo, M.; Itoh, K. J. Org. Chem. 1992,
7, 4306. (d) Nishiyama, H.; Yamaguchi, S.; Kondo, M.; Itoh, K. J. Org.
Chem. 1992, 57, 4306. (e) Li, Z.; Conser, K. R.; Jacobsen, E. N. J. Am.
Chem. Soc. 1993, 115, 5326. (f) Chang, S.; Heid, R. M.; Jacobsen, E. N.
Tetrahedron Lett. 1994, 35, 669. (g) RajanBabu, T. V.; Ayers, T. A.;
Casalnuovo, A. L. J. Am. Chem. Soc. 1994, 116, 4101. (h) Mashima, K.;
Kusano, K.; Sato, N.; Matsumura, Y.; Nozaki, K.; Kumobayashi, H.; Sayo,
N.; Hori, Y.; Ishizaki, T.; Akutagawa, S.; Takaya, H. J. Org. Chem. 1994,
1
(
1
3
1
these alkenes under biphasic NaOCl conditions revealed the
same trend for each substrate, with electron-donating groups
on the catalyst leading to higher enantioselectivities. The
observed enantioselectivities correlated directly with the σp
values of the 5,5′-substituents on the catalyst, as reflected in
the Hammett plots in Figure 2.
(
5
Although the nature of the correlation was maintained for
all three substrates, the magnitude of the dependence varied
significantly: 6 undergoes epoxidation with low enantioselec-
tivity (26-37% ee), whereas for 4 the observed ee ranges
5
9, 3064. (i) Casalnuovo, A. L.; Rajanbabu, T. V.; Ayers, T. A.; Warren,
T. H. J. Am. Chem. Soc. 1994, 116, 9869. (j) RajanBabu, T. V.; Ayers, T.
A. Tetrahedron Lett. 1994, 35, 4295. (k) Legros, J. Y.; Fiaud, J. C.
Tetrahedron 1994, 50, 465. (l) RajanBabu, T. V.; Casalnuovo, A. L. Pure
Appl. Chem. 1994, 7, 1535. (m) Park, S. B.; Murata, K.; Matsumoto, H.;
Nishiyama, H. Tetrahedron: Asymmetry 1995, 6, 2487. (n) Vander Velde,
S. L.; Jacobsen, E. N. J. Org. Chem. J. Org. Chem. 1995, 60, 5380. (o)
Schnyder, A.; Hintermann, L.; Togni, A. Angew. Chem., Int. Ed. Engl. 1995,
1
4
from 22% (catalyst 2e) to 96% (catalyst 2a).
This latter
example corresponds to a remarkable difference between
(11) Zhang, W.; Loebach, J. L.; Wilson, S. R.; Jacobsen, E. N. J. Am.
Chem. Soc. 1990, 112, 2801.
3
4, 931. (p) Zhang, H. C.; Xue, F.; Mak, T. C. W.; Chan, K. S. J. Org.
Chem. 1996, 61, 8002. (q) Gravert, D. J.; Griffin, J. H. Inorg. Chem. 1996,
(12) Lee, N. H.; Muci, A. R.; Jacobsen, E. N. Tetrahedron Lett. 1991,
32, 5053.
(13) Zhang, W.; Jacobsen, E. N. J. Org. Chem. 1991, 56, 2296.
(14) Katsuki has reported a case where a MeO-substituted chiral (salen)-
Mn complex affords lower epoxidation enantioselectivity than the corre-
sponding catalyst bearing a H-substituent: Hamada, T.; Fukuda, T.;
Imanishi, H.; Katsuki, T. Tetrahedron 1996, 52, 515.
3
1
5, 4837. (r) RajanBabu, T. V.; Casalnuovo, A. L. J. Am. Chem. Soc. 1996,
18, 6325. (s) Halterman, R. L.; Jan, S. T.; Abdulwali, A. H.; Standlee, D.
J. Tetrahedron 1997, 53, 11277. (t) RajanBabu, T. V.; Ayers, T. A.;
Halliday, G. A.; You, K. K.; Calabrese, J. C. J. Org. Chem. 1997, 62, 6012.
(
3
u) Reetz, M. T.; Waldvogel, S. R.; Goddard, R. Tetrahedron Lett. 1997,
8, 5967.

950 J. Am. Chem. Soc., Vol. 120, No. 5, 1998
Palucki et al.
Figure 2. Correlation of the enantioselectivity of epoxidation of alkenes 4-6sexpressed as the log of the ratio of epoxide enantiomerssagainst
values of the substituent X in catalysts 2a-e.
σ
P
q
diastereomeric transition state energies, ∆∆∆G , of 2.0 kcal/
mol. Catalysts 3a-e display analogous trends, affording slightly
higher enantioselectivities. For example, epoxidation of 4
q
afforded 66% ee with 3e and 98% ee with 3a (∆∆∆G ) 2.0
kcal/mol). The linearity of the correlations indicates that
changes in enantioselectivity are indeed due to the electronic
character of the catalysts and that the steric perturbations
imposed by variation of the 5,5′-substituents do not play a
significant role.
2
. Mechanistic Possibilities. In principle, these sizable
electronic effects may be ascribed to any of several factors. The
,5′-substituent could induce significant conformational distor-
5
tion of the reactive (salen)Mn(V) oxo intermediate. While this
possibility is difficult to assess directly, since to date no (salen)-
15
Mn(V) oxo species have yielded to structural characterization,
the X-ray crystal structures of a number of chiral (salen)Mn-
III) catalysts indicate that the complexes maintain a square
P
Figure 3. Plot of E1/2 for the Mn(II)/Mn(III) redox couple vs σ values
of the substituent X in catalysts 3b-e. The CV measurements were
(
-3
carried out at room temperature under argon: [complex] ) 10 M in
planar salen ligand conformation regardless of the electronic
properties of the ligand substituents.16,17 A second possibility
is that the substituents may effect changes in the Mn-oxo bond
length in the active species, in turn altering the nonbonding
substrate/ligand interactions in the relevant transition states.
However, such effects on metal-oxo bond length are typically
2 2 4 4
CH Cl ; [Bu NBF ] ) 0.1 M; scan speed 200 mV/s; working electrode,
glassy carbon; auxilliary electrode, Pt wire; reference electrode, Ag/
AgCl/KCl.
catalyst. Electron-donating substituents on the catalyst would
be expected to stabilize the high-valent Mn(V) oxo intermediate,
attenuating its reactivity and thus generating a relatively milder
oxidant. Similarly, electron-withdrawing substituents on the
catalyst are expected to destabilize the Mn(V) oxo intermediate,
1
8
small. Further, the observed ee trend runs contrary to what
one might predict from this hypothesis, since electron-withdraw-
ing substituents should shorten the Mn-oxo bond length and
thus lead to an increase, rather than a decrease, in the
enantioselectivity of oxo transfer.
1
9
making it a more reactive oxidant. The results of electro-
chemical studies on catalysts 3a-e lend credence to this
An alternative explanation is that the electronic effect on
enantioselectivity derives from changes in the reactivity of the
Mn(V) oxo intermediate imparted by the substituents on the
supposition: there is a clear correlation between σ of the 5,5′-
p
substituent and the value of E1/2 for the Mn(II)/Mn(III) redox
couple (Figure 3). Accurate measurement of the redox poten-
tials for higher Mn oxidation states was prevented by the
apparent instability of Mn(IV) and Mn(V) species under the
electrochemical reaction conditions.
In accord with the Hammond postulate, a milder oxidant
should lead to an oxygen transfer to alkene via a more product-
(15) For recent efforts directed toward the isolation of reactive Mn(V)
oxo complexes, see: (a) Groves, J. T.; Lee, J.; Marla, S. S. J. Am. Chem.
Soc. 1997, 119, 6269. (b) Feichtinger, D.; Plattner, D. A. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1718.
(
16) Pospisil, P. J.; Carsten, D. H.; Jacobsen, E. N. Chem. Eur. J. 1996,
, 974.
17) Katsuki has demonstrated recently that sparteine can induce moderate
enantioselectivity in the epoxidation of certain olefins catalyzed by achiral
salen)Mn complexes and has proposed a mechanism involving complex-
2
(
(19) The most direct test of this hypothesissdirect comparison of the
rates of reactions catalyzed by electron-rich and elctron-poor catalystssis
precluded by the fact that mass transport of the aqueous oxidant into the
organic layer is the rate-limiting step under biphasic conditions: Su a´ rez,
A.; Pospisil, P. J.; Jacobsen, E. N. Unpublished results. See also: (a)
Senanayake, C. H.; Smith, G. B.; Liu, J.; Fredenburgh, L. E.; Ryan, K. M.;
Hughes, D. L.; Larsen, R. D.; Verhoeven, T. R.; Reider, P. J. Tetrahedron
Lett. 1995, 36, 3993. (b) Senanayake, C. H.; Smith, G. B.; Ryan, K. M.;
Fredenburgh, L. E.; Liu, J.; Roberts, F.; Hughes, D. L.; Larsen, R. D.;
Verhoeven, T. R.; Reider, P. J. Tetrahedron Lett. 1996, 37, 3271. (c) Hughes,
D. L.; Smith, G. B.; Liu, J.; Dezeny, G. C.; Senanayake, C. H.; Larsen, R.
D.; Verhoeven, T. R.; Reider, P. J. J. Org. Chem. 1997, 62, 2222.
(
ation of the amine to the catalyst resulting in a nonplanar reactive ligand
conformation: Hashihayata, T.; Ito, Y.; Katsuki, T. Tetrahedron 1997, 53,
9
541. However, the very low yields of epoxide obtained in these reactions
(0.1-0.5 equiv relative to sparteine) raise the possibility that any enanti-
oselectivity might be due instead to secondary reactions such as kinetic
resolution of epoxide mediated by the chiral amine. This notion is supported
by the fact that epoxide yields decrease and enantioselectivities increase
with added amine or with added water in the presence of amine.
(
18) Nugent, W. A.; Mayer, J. M. Metal-Ligand Multiple Bonds;
Wiley: New York, 1988; pp 148-152.

(Salen)Mn(III)-Catalyzed Epoxidation Reaction
J. Am. Chem. Soc., Vol. 120, No. 5, 1998 951
Scheme 3
Table 1. Epoxidation of cis-â-Deuteriostyrene Catalyzed by 3a-e
enantiofacial
selectivity
a
cis/transa
entry
catalyst
k
H
/k
D
1
2
3
4
5
3a (X ) OMe)
3b (X ) Me)
3c (X ) H)
0.82
0.84
0.86
0.90
0.95
5.4
4.8
4.4
4.2
2.0
5.1
5.7
6.3
6.9
7.1
Figure 4. Correlation of the kinetic isotope effect in the epoxidation
of styrene vs cis-â-deuteriostyrene as a function of the σ
the substituent X in catalysts 3a-e.
P
values of
3d (X ) Cl)
2
3e (X ) NO )
a
1
Determined by H NMR in the presence of chiral shift reagent
(
)
Eu(hfc)
[cis + trans (major enantiomers)]/[cis + trans (minor enantiomers)].
See: Zhang, W.; Lee, N. H.; Jacobsen, E. N. J. Am. Chem. Soc. 1994,
16, 425.
3
according to the following equation: enantiofacial selectivity
1
like transition state. Since this transformation involves the
encounter of two reactants which, at the origin of the reaction
coordinate, do not interact at all, one might expect a higher
degree of stereochemical communication in a later transition
state. Oxo transfer from a more reactive oxidant, on the other
hand, should proceed via a more reactant-like transition state,
with greater spatial separation between substrate and catalyst
and concomitantly poorer differentiation of the diastereomeric
transition structures (Figure 1). As outlined in the following
discussion, support for this mechanistic proposal has been
provided through isotopic labeling studies and the temperature
profile of the epoxidation.
Figure 5. Correlation of the enantiofacial selectivity in the first C-O
bond-forming step in the epoxidation of cis-â-deuteriostyrene with the
kinetic isotope effect using catalysts 3a-e.
3
. A. Correlation of Deuterium Isotope Effects with the
C-O bond-forming step, the higher the enantioselectivity in
Electronic Character of the Catalyst. A direct evaluation of
the relative position of the transition state on the reaction
coordinate was obtained by competitive epoxidation of styrene
and cis-â-deuteriostyrene with catalysts 2a-e. On transforma-
tion to the radical intermediate, the â-carbon of styrene
undergoes a formal rehybridization from sp to sp which, in
principle, should lead to the observation of an inverse secondary
isotope effect (kH/kD < 1) for the epoxidation (Scheme 3).
Further, the magnitude of kH/kD should vary in concert with
the epoxidation reaction.
3
. B. Electronic Effects on Cis/Trans Partitioning. As
an additional mechanistic handle provided by the isotope
labeling study outlined above, the cis/trans partitioning in the
second C-O bond-forming step could be assessed directly by
measuring the relative proportions of cis- and trans-â-deuteri-
ostyrene oxide produced. The cis/trans ratio represents relative
rate of the collapse and rotation/collapse pathways outlined in
Scheme 1 (eq 1), in turn reflecting the energetic barrier to ring
closure. A higher barrier should lead to a longer lived radical
intermediate, and thus a greater extent of rotational isomerization
and a lower cis/trans ratio.
2
3
2
0,21
the position of the transition state: later transition states, in
which the â-carbon has more sp character, should exhibit
3
smaller values of kH/kD. The experimental data reveal a direct
correlation between kH/kD and σp (Table 1, Figure 4), indicating
that the electronic character of the catalyst does indeed alter
the degree of rehybridization at the â-carbon and thus the
position of the transition state leading to formation of the radical
intermediate. Perhaps more relevant, a direct correlation is
observed between kH/kD and the enantioselectivity of epoxidation
k /k
) (% cis epoxide)/(% trans epoxide)
(1)
cis trans
A correlation was revealed between kcis/ktrans and σp (Table
, Figure 6) wherein electron-poor catalysts lead to higher cis/
1
trans ratios relative to catalysts bearing electron-rich substituents.
This observed effect can be rationalized in a straightforward
manner: as ring closure of the radical intermediate to generate
the second C-O bond of the epoxide occurs with formal
reduction of Mn(IV) to Mn(III), this process should be favored
by more electron-withdrawing substituents on the metal ligand,
leading to an increase in the cis/trans ratio relative to more
electron-rich ligands. Thus, the electronic character of the
(salen)Mn epoxidation catalysts has been demonstrated to
(
Figure 5). This result supports in a compelling and quantitative
manner the notion that the later the transition state in the first
(
20) Isaacs, N. Physical Organic Chemistry, 2nd ed.; John Wiley &
Sons: New York, 1995; pp 296-301.
21) In contrast, there is no change in hybridization at the R carbon. As
(
would be expected, the experimentally determined kinetic isotope effect
for the epoxidation of styrene vs R-deuteriostyrene, kH/kD, ) 1. It should
be noted that this observation is inconsistent with a mechanism for
epoxidation involving oxametallacycle formation. See refs 8a and 8e.

952 J. Am. Chem. Soc., Vol. 120, No. 5, 1998
Palucki et al.
Figure 6. Correlation between the cis/trans product ratio in the
P
epoxidation of cis-â-deuteriostyrene and the σ values of the substituent
X in catalysts 3a-e.
Figure 7. Correlation between the enantioselectivity in the epoxidation
influence both formation and ring closure of the radical
intermediate and consequently affect the stereoselectivity of both
C-O bond-forming steps in the oxygen-atom-transfer process.
of indene at various temperatures with the σ
X in catalysts 2a-c and 2e-g.
P
values of the substituent
reactions. The observation of an isokinetic relationship is
generally taken as evidence that the reaction parameter being
varied (e.g., temperature, solvent, or, in this case, σp) is
influencing only one type of interaction in the system. The
isoelectronic relationship seen in the present case may thus be
taken as evidence that variation of the electronic nature of the
3
. C. Temperature Effects in the Epoxidation. We have
recently developed an anhydrous, low-temperature protocol for
the (salen)Mn-catalyzed epoxidation reaction employing the
combination of m-chloroperbenzoic acid (m-CPBA) and N-
methylmorpholine N-oxide (NMO) as the terminal oxidant
system.22 These reactions are monophasic, anhydrous, and
5,5′-substituents of the ligand influences only one central aspect
homogeneous, allowing access to a broader temperature range
for the epoxidation. With the ability to carry out epoxidations
at -78 °C, significant improvements in the enantiomeric
of the reaction, namely the degree of the first C-O bond
formation and thus the position of the enantiodifferentiating
transition state along the reaction coordinate.
excesses have been attained for many substrates, most notably
terminal olefins such as styrene.2
2b
As noted earlier, direct kinetic analysis of the biphasic
epoxidation reaction is complicated by the zero-order rate
dependence on olefin that results from rate-limiting transfer of
In addition, it has become
possible to evaluate the temperature-dependence of enantiose-
lectivity for (salen)Mn-catalyzed epoxidations, allowing further
insight into the nature of the electronic effect.
1
9c
oxidant into the organic layer.
Although it was hoped that
the single-phase mCPBA/NMO conditions would be more
amenable to the measurement of absolute rate constants,
epoxidations of indene carried out at -78 °C were, remarkably,
typically complete within 1 min, again precluding straightfor-
The epoxidation of indene was carried out under low-
temperature conditions with catalysts 2a-c,e,f. Indene is
epoxidized without partitioning to cis and trans diastereomers,
and as such the observed ee is directly representative of the
inherent facial selectivity of the first C-O bond-forming step,
2
5
ward rate measurements. Making the conservative estimate
2
3
that t1/2 ) 30 s and applying the Eyring equation to the rate
as long as this first step is completely irreversible (eq 2).
2
6
expression in eq 3, this places an upper limit of ca. 12-13
q
kcal/mol on ∆G at [indene] ) 1 M, testifying to the very high
enantiofacial selectivity ) kmajor/kminor
)
1
5
reactivity of the (salen)Mn oxo intermediate.
(% major epoxide)/(% minor epoxide) (2)
rate ) [(salen)Mn(V)oxo][indene]
(3)
The relationship between enantioselectivity in the epoxidation
of indene under the m-CPBA conditions and the σP values of
the 5,5′-substituents of different catalysts was investigated over
a range of different temperatures (Figure 7). The Hammett plot
at each temperature is linear, consistent with the results obtained
in the epoxidations of 4-6 at room temperature using NaOCl
as the terminal oxidant. A notable feature of this temperature
profile is the existence of an “isoelectronic point”sa single point
at or near which all of the Hammett plots intersectsin analogy
to the well-established isokinetic effect.24 The isokinetic effect
is characterized by the intersection at a single point (the
isokinetic point) of the Eyring plots for a series of related
Analysis of the temperature dependence of the observed
enantioselectivities (Figure 8) provided a means of evaluating
q
q
∆∆H and ∆∆S for the two diastereomeric transition states
leading to formation of the major and minor enantiomeric indene
oxide products (Table 2). Two opposing trends emerge on
q
inspection of these data. The enthalpic contribution (∆∆H )
dominates in the more enantioselective reactions effected by
catalysts bearing electron-rich substituents. Conversely, the
enthalpic contribution diminishes and the entropic contribution
q
(∆∆S ) becomes more important with more electron-withdraw-
ing catalyst substituents, to the point where, for the nitro-
substituted catalyst 2e, the enthalpic component marginally
favors the minor enantiomer, but is overcome by the entropic
contribution. This latter case leads to an unusual situation in
(22) (a) Palucki, M.; Pospisil, P. J.; Zhang, W.; Jacobsen, E. N. J. Am.
Chem. Soc. 1994, 116, 9333. (b) Palucki, M.; McCormick, G. J.; Jacobsen,
E. N. Tetrahedron Lett. 1996, 37, 5457.
(23) This assumption is based on the general observation that conjugated
cis-alkenes are observed not to undergo cis/trans isomerization under the
conditions of epoxidation.
(25) Efforts to lower the observed rate of epoxidation by effecting the
reactions under high dilution conditions were thwarted by the observation
of low catalyst turnovers under these conditions.
(24) (a) Isaacs, N. Physical Organic Chemistry, 2nd ed.; John Wiley &
Sons: New York, 1995; pp 116-7. (b) Leffler, J. E.; Grunwald, E. Rates
and Equilibria in Organic Reactions; Wiley: New York, 1963.
(26) In this analysis, [(salen)Mn(V)oxo] is taken to be equal to the initial
catalyst concentration.

(Salen)Mn(III)-Catalyzed Epoxidation Reaction
J. Am. Chem. Soc., Vol. 120, No. 5, 1998 953
formation and thus transition state position) is influenced by
the 5,5′-substituent.
(d) Eyring analyses revealing that enantioselectivity in
epoxidations with the most selective catalysts relies almost
entirely on enthalpic factors, while less enantioselective catalysts
are subject to greater entropic influence.
The notion that the position of the transition state along the
reaction coordinate is the critical element in the (salen)Mn-
catalyzed asymmetric epoxidation reaction is perhaps quite
reasonable in retrospect, but it could hardly have been antici-
pated in the initial design of these systems. It seems likely that
(salen)Mn-catalyzed epoxidations are not the only reactions that
can exhibit a correlation between enantioselectivity and the
position of the enantioselectivity-determining step along the
reaction coordinate. The observations made in this study may
therefore hold general implications for the design of asymmetric
catalysts, particularly those that proceed without substrate
precoordination.
Figure 8. Eyring plots for the epoxidation of indene with catalysts
a-c and 2e-g.
Experimental Section
2
General. Enantiomeric ratios measured by HPLC were determined
using the Daicel Chiralcel O-series column with a mobile phase of
10% isopropyl alcohol in hexane unless otherwise noted. Tetrahydro-
furan (THF), diethyl ether, benzene, and toluene were distilled from
the sodium benzophenone ketyl. Dichloromethane, hexane, and
acetonitrile were distilled from calcium hydride. All other reagents
and solvents were reagent grade and were used without further
purification unless otherwise specified. 4-Methylmorpholine N-oxide
Table 2. Relative Actication Parameters for the Formation of
Indene Oxide Enantiomers Catalyzed by Complexes 2a-c, e-g
q
q
∆
∆H
∆∆S
(kcal mol-1)
(cal mol-1 K-1
)
entry
catalyst
substituent
i
1
2
3
4
5
6
2g
2a
2b
2c
2f
OSi( Pr)
3
1.70 ( 0.08
1.40 ( 0.16
0.66 ( 0.05
0.29 ( 0.10
0.11 ( 0.09
-0.28 ( 0.02
-0.69 ( 0.35
-0.07 ( 0.76
2.90 ( 0.21
3.80 ( 0.44
3.60 ( 0.40
2.50 ( 0.10
OCH
3
CH
H
3
(NMO) and indene were obtained from Aldrich Chemical Co. and were
used as received. m-Chloroperbenzoic acid was purified according the
Br
NO
2
literature procedure.27
2e
General Procedure for the Epoxidation of Indene. A solution of
indene (24 µL, 0.20 mmol), NMO (117 mg, 1.00 mmol), and catalyst
2
a (7.7 mg, 0.01 mmol) in 2 mL of dry CH
appropiate temperature for 10 min before the addition of solid m-CPBA
72 mg, 0.40 mmol). After stirring for 1 h at the appropiate temperature,
mL of 3 N NaOH and 2 mL of H O were added to the solution. The
organic phase was separated, washed with 4 mL of brine, and dried
over Na SO . Removal of the drying agent followed by concentration
2 2
Cl was cooled to the
which the observed degree of asymmetric induction increases,
rather than decreases, with increasing temperature.
(
4
The straightforward correlation between the relative activation
parameters for the diastereomeric transition states and the
electronic properties of the catalysts appears to be quite
consistent with the Hammond postulate arguments presented
above. If electron-donating substituents do indeed lead to a
later transition state with a more highly developed C-O bond,
the energetic differences between the diastereomeric transition
states should be more dependent on the degree of bond
2
2
4
yielded the epoxide which was dissolved in hexane and filtered through
a pad of Celite. Complete removal of the solvent from the resulting
filtrate provided indene oxide (96% ee, 95% yield) that was >95%
1
1
pure by GC and H NMR analysis. H NMR (CDCl ) δ 2.97 (dd, 1H,
3
J ) 3.0, 17.7 Hz), 3.21, (d, 1H, J ) 18.0 Hz), 4.13 (t, J ) 2.7 Hz),
4.26, (d, 1H, J ) 2.7 Hz), 7.10-7.30 (m, 3H), 7.50 (d, 1H, J ) 6.9
Hz).
General Procedure for the Epoxidation of cis-â-Deuteriostyrene.
To a solution of cis-â-deuteriostyrene (0.30 mmol), catalyst (8 mol
q
formation (reflected by ∆∆H ). Similarly, in the case of
electron-deficient catalysts, if the transition state is earlier, less
ordered, and characterized by a lower degree of C-O bond
formation, it is reasonable for whatever enantioselectivity that
is observed to derive primarily from entropic factors.
%
), in 1 mL of CH
pH ) 11.3, 1.50 mL, 0.90 mmol) at room temperature. The mixture
was stirred for 1 h before the addition of 2 mL of H O and 2 mL of
CH Cl . The organic layer was separated, washed with brine (3 mL),
and dried over Na SO . Removal of the drying agent and solvent
followed by chromatography (SiO ) afforded pure cis-â-deuteriostyrene
oxide (78% yield).
2 2
Cl was added a solution 0.55 M buffered bleach
(
2
Conclusions
2
2
A wide range of experimental data support the proposal that
the electronic character of chiral (salen)Mn catalysts influences
the enantioselectivity of the catalytic epoxidation by altering
the position of the transition state in the first C-O bond-forming
step. The critical experimental results are as follows:
2
4
2
General Procedure for the Measurement of the Kinetic Inverse
Isotope Effect. To a solution of cis-â-deuteriostyrene (1 equiv), styrene
(
(
1
1 equiv), and catalyst (4 mol %) in CH
0.2 equiv) at room temperature. The mixture was allowed to stir for
h before the addition of 2 mL of H O and 2 mL of CH Cl . The
organic layer was separated, washed with brine (3 mL), and dried over
2 2
Cl was added buffered bleach
(a) Linear Hammett plots correlating enantioselectivity in the
epoxidation of cis disubstituted olefins with the σp values of
the 5,5′-substituents on the catalyst.
2
2
2
1
(b) Secondary deuterium isotope effects correlating the value
2 4
Na SO . GC and H NMR analysis of the reaction mixture provided
of kH/kD at Câ, and thus the degree of rehybridization, with both
σp and the observed enantioselectivity.
a measure of the relative rates of epoxidation for styrene and cis-â-
deuteriostyrene.
Preparation of Chiral Ligands and the Corresponding (salen)-
Mn(III) Complexes. Catalysts 2e-f, 3a-c, and their ligand precursors
(c) Linear Hammett plots correlating enantioselectivity in the
epoxidation of indene with σp over a 100 °C temperature range,
including the observation of an “isoelectronic” point indicating
that a single reaction parameter (such as the degree of bond
(27) Bortilini, O.; Campestrini, S.; DiFuria, F.; Modena, G. J. Org. Chem.
1987, 52, 3.

954 J. Am. Chem. Soc., Vol. 120, No. 5, 1998
Palucki et al.
were have been reported previously.16,28 The synthesis and character-
ization of ligand precursors to 2a-d and 3d-e are provided as
Supporting Information.
m/z calcd (C36
H
36Cl
MnN
O
) 688.1223, found 688.1220. Anal. Calcd
3
2
2
for C36 36Cl MnN
H
3
2
O
2
: C, 62.67; H, 5.26; N, 4.06; Cl, 15.41; Mn, 7.96.
Found: C, 62.98; H, 5.28; N, 4.08; Cl, 15.38; Mn, 7.92.
Chloro-(R,R)-[[2,2′-[(1,2-diphenyl-1,2-ethanediyl)bis(nitrilometh-
ylidyne)]bis[4-methoxy-6-(1,1-dimethylethyl)phenolato]]-N,N′,O,O′]-
manganese(III) (2a). A three-necked round-bottom flask equipped
with a reflux condensor, an addition funnel, and a septum was charged
with the salen ligand (1.00 g, 1.69 mmol) and 100 mL of EtOH. The
Chloro-(R,R)-[[2,2′-[(1,2-diphenyl-1,2-ethanediyl)bis(nitrilometh-
ylidyne)]bis[4-(triisopropylsiloxy)-6-(1,1-dimethylethyl)phenolato]]-
N,N′,O,O′] manganese(III) (2g). To a solution of diphenylethylene-
diamine (244 mg, 1.15 mmol) in 90 mL of EtOH was added 3-tert-
butyl-5-(triisopropylsiloxy)salicylaldehyde16 (800 mg, 2.30 mmol). The
resulting yellow solution was heated to reflux for 25 min before the
addition of a solution of Mn(OAc) ‚4H O (0.564 g, 2.30 mmol) in 5
mixture was heated to reflux, and a solution of Mn(OAc)
2
2
‚4H O
dissolved in 5 mL of H O was added whereupon the solution
2
2
2
immediately turned brown. After 30 min of reflux, air was bubbled
into the solution and refluxed for an additional 30 min. Brine (3 mL)
was added and the mixture allowed to cool to room temperature. The
mL of H O. The resulting brown solution was heated to reflux for 30
2
min, after which air was bubbled through the solution via a needle for
an additional 30 min. Brine (5 mL) was added, and the mixture was
further heated at reflux for 30 min before being cooled to ambient
temperature. The solvent volume was reduced to ca. 15 mL under
vacuum, and CH Cl (100 mL) and water (100 mL) were added. The
solvents were removed under vacuum, and then 100 mL of CH
was added. The solution was washed with 100 mL of brine and 100
mL of H O and dried over anhydrous Na SO . Removal of the drying
agent followed by concentration of filtrate resulted in a brown powder
which was chromatographed (EtOH/CH Cl ) to afford the catalyst
0.882 g, 54% yield) as a brown solid. IR (KBr) 3281, 2942, 1605,
2 2
Cl
2
2
4
2
2
organic phase was separated, washed with brine (100 mL), and dried
over Na SO . After solvent removal, the residue was purified by
2
2
2
4
(
2 2 2
chromatography (SiO , 5% EtOH/CH Cl ) to afford 0.854 g of product
-
1
1
2
6
541, 1454, 1418, 1350, 1296, 1206, 1156, 1057, 818 cm ; mp 282.5-
(77% yield): mp 280-280.8 °C.; IR (CH Cl ) 2947, 2868, 1600, 1535,
2
2
-1
83 °C; HRMS (FAB) m/z calcd (C38
H
42ClMnN
2
O
4
) 680.2214, found
1464, 1409, 1343, 1230, 1041, 1010, 963, 882, 869 cm ; HRMS (FAB)
+
80.2212.
m/z calcd (C H ClMnN O Si ) 929.4881 (M - Cl ), found 929.4914.
5
4
78
2
4
2
Chloro-(R,R)-[[2,2′-[(1,2-diphenyl-1,2-ethanediyl)bis(nitrilometh-
Chloro-(R,R)-[[2,2′-[(1,2-cyclohexanediyl)bis(nitrilomethylidyne)]-
bis[4-nitro-6-(1,1-dimethylethyl)phenolato]]-N,N′,O,O′]manganese-
(III) (3d). A three-necked round-bottom flask equipped with a reflux
condenser, an addition funnel, and a septum was charged with the ligand
precursor to 3d (0.413 g, 0.815 mmol) and 60 mL of EtOH. The
mixture was heated to reflux and a solution of Mn(OAc) ‚4H O (0.400
ylidyne)]bis[4-methyl-6-(1,1-dimethylethyl)phenolato]]-N,N′,O,O′]
manganese(III) (2b). To a solution of the salen ligand (561 mg, 1.00
mmol) in 10 mL of 95% EtOH was added dropwise a soution of Mn-
(
OAc)
2
2
‚4H O (490 mg, 2.00 mmol) dissolved in 5 mL of 95% EtOH.
The resulting brown solution was heated at reflux for 30 min before
the addition of a solution of LiCl (212 mg, 5.00 mmol) in 5 mL of
EtOH was added. The mixture was heated at reflux for an additional
2
2
g, 1.630 mmol) dissolved in 5 mL of H O was added, whereupon the
2
solution immediately turned brown. After the mixture was heated to
reflux for 30 min, air was bubbled into the solution and reflux was
continued for an additional 30 min. Brine (3 mL) was added, and the
mixture was allowed to cool to room temperature. Solvents were
removed under vacuum, and the residue was dissolved 100 mL of CH2-
3
0 min. Water (5 mL) was added, and the mixture was cooled slowly
to room temperature. The precipitate was collected by filtration to
afford the brown catalyst (564 mg, 87% yield): IR (KBr) 2952, 1615,
1
539, 1454, 1429, 1343, 1310, 1235, 1204, 1169, 1005, 855, 822, 700
-
1
cm ; mp 279-280 °C; HRMS (FAB) m/z calcd (C38
H
42ClMnN
42ClMnN : C,
0.31; H, 6.52; N, 4.32; Cl, 5.46; Mn, 8.46. Found: C, 70.30; H, 6.53;
2
O
2
)
Cl . The solution was then washed with 100 mL of brine and 100 mL
2
6
7
48.2315, found 648.2320. Anal. Calcd for C38
H
2
O
2
of H O and dried over anhydrous Na SO . Removal of the drying agent
2
2
4
followed by concentration of the filtrate resulted in a brown powder
which was purified by chromatography (5% EtOH/CH Cl ), affording
N, 4.22; Cl, 5.59; Mn, 8.28.
2
2
Chloro-(R,R)-[[2,2′-[(1,2-diphenyl-1,2-ethanediyl)bis(nitrilometh-
ylidyne)]bis[6-(1,1-dimethylethyl)phenolato]]-N,N′,O,O′] manganese-
the product (0.482 g, 84% yield): mp 338-339 °C; IR (CCl ) 2965,
4
2949, 2284, 2249, 2245, 1618, 1593, 1536, 1410, 1332, 1307, 1265,
(
III) (2c). To a hot solution of the salen ligand precursor to 2c (2.644
g, 5.0 mmol) in 100 mL of 95% EtOH was added Mn(OAc) ‚4H
2.451 g, 10.00 mmol) as a solid in one portion. The mixture was
1175, 771, 752, 745 cm-1; HRMS (FAB) m/z calcd (C H Cl MnN O )
28
34
2
4
2
2
O
555.1378 (M - Cl ), found 555.1368.
+
2
(
Chloro-(R,R)-[[2,2′-[(1,2-cyclohexanediyl)bis(nitrilomethylidyne)]-
heated to reflux for 30 min under an atmosphere of air and then cooled
to room temperature. A solutioin of LiCl (1.060 g, 25.00 mmol) in
EtOH (20 mL) was added, and the mixture was heated to reflux for an
additional 30 min. The mixture was allowed to cool to room
temperature, affording 2.365 g of product as dark brown crystals. A
bis[4-nitro-6-(1,1-dimethylethyl)phenolato]]-N,N′,O,O′]manganese-
(III) (3e). A three-necked round-bottom flask equipped with a reflux
condenser, an addition funnel, and a glass stopper was charged with
the salen ligand (0.209 g, 0.400 mmol) and 20 mL of EtOH. The
mixture was heated to reflux, and a solution of Mn(OAc)
2
2
‚4 H O
second crop of product (0.490 g) was obtained by adding H
2
O to the
dissolved in 2 mL of 50% EtOH/H O was added whereupon the solution
2
filtrate (total yield 91%): mp 271.5-272 °C; IR (KBr) 2950, 1593,
immediately turned brown. After 30 min of refluxing, 1 mL of brine
was added and the resulting solution refluxed for an additional 30 min.
After the mixture was cooled to room temperature, the solvents were
-
1
1
543, 1414, 1389, 1302, 1196, 1145, 870 cm ; HRMS (FAB) m/z
calcd (C36 38ClMnN ) 620.2002, found 620.1999. Anal. Calcd for
38ClMnN : C, 68.62; H, 6.24; N, 4.45; Cl, 5.63; Mn, 8.72.
H
2 2
O
C
36
H
2 2
O
removed and 50 mL of CH
washed with 50 mL of brine followed by 50 mL of H
over anhydrous Na SO . Removal of the drying agent followed by
concentration of the filtrate resulted in a brown powder which was
chromatographed (5% EtOH/CH Cl ), affording the product (0.220 g,
2
Cl
2
was added. The solution was then
Found: C, 68.45; H, 6.43; N, 4.15; Cl, 5.87; Mn, 8.66.
2
O and dried
Chloro-(R,R)-[[2,2′-[(1,2-diphenyl-1,2-ethanediyl)bis(nitrilometh-
ylidyne)]bis[4-chloro-6-(1,1-dimethylethyl)phenolato]]-N,N′,O,O′] man-
ganese(III) (2d). To a hot solution of the salen ligand precursor to
2
4
2
2
2
4
d (0.602 g, 1.0 mmol) in 40 mL of 95% EtOH was added Mn(OAc)
2
‚-
9
0% yield): IR (CCl
4
) 2955, 1628, 1598, 1565, 1510, 1330, 1311, 1296,
2
H O (0.490 g, 2.00 mmol) as a solid in one portion. The reaction
-1
1271, 1116, 706, 698, 690 cm ; mp >400 °C; HRMS (FAB) m/z calcd
mixture turned brown immediately. The mixture was heated to reflux
for 30 min under an atmosphere of air and then cooled to room
temperature. A solution of LiCl (0.212 g, 5.00 mmol) in EtOH (20
mL) was added, and the mixture was heated to reflux for an additional
+
(C
28
H34ClMnN
4
O
6
) 577.1859 (M - Cl ), found 577.1849.
Acknowledgment. This work was supported by the NIH
(GM43214) and by a postdoctoral fellowship to N.S.F. (NIGMS).
3
0 min. Water (10 mL) was added until no more precipitate formed.
The resulting brown solid was collected by filtration, washed with 20
mL of 30% EtOH, and dried in air to give 0.628 g of product (91%
yield): mp 297.5-298 °C; IR (KBr) 2950, 2868, 1613, 1534, 1495,
Supporting Information Available: Synthesis and charac-
terization of the ligand precursors to 2a-d and 3d-e (4 pages).
See any current masthead page for ordering and Internet access
instructions.
-
1
1
454, 1426, 1404, 1389, 1304, 1173, 1005, 816 cm ; HRMS (FAB)
(28) Larrow, J. R.; Jacobsen, E. N.; Gao, Y.; Hong, Y.; Nie, X.; Zepp,
C. M. J. Org. Chem. 1994, 59 1938.
JA973468J