cis-Stilbene and (1α,2β,3α)-(2-ethenyl-3-methoxycyclopropyl)benzene as mechanistic probes in the MnIII(salen)-catalyzed epoxidation: Influence of the oxygen source and the counterion on the diastereoselectivity of the competitive concerted and radical-type oxygen transfer

Article abstract of DOI:10.1021/ja0177206

cis-Stilbene (1) has been epoxidized by a set of diverse oxygen donors [O×D], catalyzed by the Mn^III(salen)X complexes 3 (X = Cl, PF₆), to afford a mixture of cis- and trans-epoxides 2. The cis/trans ratios range from 29:71 (extensive isomerization) to 92:8, which depends both on the oxygen source [O×D] and on the counterion X of the catalyst. When (1α,2β,3α)-(2-ethenyl-3-methoxycyclopropyl)-benzene (4) is used as substrate, a mechanistic probe which differentiates between radical and cationic intermediates, no cationic ring-opening products are found in this epoxidation reaction; thus, isomerized epoxide product arises from intermediary radicals. The dependence of the diastereoselectivity on the oxygen source is rationalized in terms of a bifurcation step in the catalytic cycle, in which concerted Lewis-acid-activated oxygen transfer competes with stepwise epoxidation by the established Mn^V(oxo) species. The experimental counterion effect is attributed to the computationally assessed ligand-dependent reaction profiles and stereoselectivities of the singlet, triplet, and quintet spin states available to the manganese species.

Full text of DOI:10.1021/ja0177206

Published on Web 04/11/2002
cis-Stilbene and
(1r,2â,3r)-(2-Ethenyl-3-methoxycyclopropyl)benzene as
Mechanistic Probes in the MnIII(salen)-Catalyzed Epoxidation:
Influence of the Oxygen Source and the Counterion on the
Diastereoselectivity of the Competitive Concerted and
Radical-Type Oxygen Transfer
Waldemar Adam,† Konrad J. Roschmann,*,† Chantu R. Saha-Mo¨ller,† and
Dieter Seebach‡
Contribution from the Institute of Organic Chemistry, UniVersity of Wu¨rzburg, Am Hubland,
D-97074 Wu¨rzburg, Germany, and Laboratorium fu¨r Organische Chemie, Eidgeno¨ssische
Technische Hochschule, ETH Ho¨nggerberg, HCI H 315, CH-8093 Zu¨rich, Switzerland
Received December 10, 2001
Abstract: cis-Stilbene (1) has been epoxidized by a set of diverse oxygen donors [OxD], catalyzed by the
MnIII(salen)X complexes 3 (X ) Cl, PF6), to afford a mixture of cis- and trans-epoxides 2. The cis/trans
ratios range from 29:71 (extensive isomerization) to 92:8, which depends both on the oxygen source [OxD]
and on the counterion X of the catalyst. When (1R,2â,3R)-(2-ethenyl-3-methoxycyclopropyl)-benzene (4)
is used as substrate, a mechanistic probe which differentiates between radical and cationic intermediates,
no cationic ring-opening products are found in this epoxidation reaction; thus, isomerized epoxide product
arises from intermediary radicals. The dependence of the diastereoselectivity on the oxygen source is
rationalized in terms of a bifurcation step in the catalytic cycle, in which concerted Lewis-acid-activated
oxygen transfer competes with stepwise epoxidation by the established MnV(oxo) species. The experimental
counterion effect is attributed to the computationally assessed ligand-dependent reaction profiles and
stereoselectivities of the singlet, triplet, and quintet spin states available to the manganese species.
substituted olefins.2,3 To account for this loss of stereoselectivity,
Introduction
a radical intermediate has been proposed, which leads to cis/
Direct oxygen transfer to olefins is a well-established and
popular route to prepare epoxides, valuable building blocks in
synthetic organic chemistry.1 In recent years, there has been
much effort to conduct this transformation selectively under
catalytic conditions.2 To date, the best known method to
epoxidize unfunctionalized olefins enantioselectively is the
Jacobsen-Katsuki epoxidation, in which optically active MnIII-
(salen) complexes are employed as catalysts and PhIO or NaOCl
as oxygen sources, with the MnV(oxo) species as the active3
oxidant.2-4
Although the synthetic value of this reaction is undisputed,
its mechanism is currently under intensive debate.3 Substrate
isomerization has been of considerable concern, in which cis-
olefins afford a mixture of cis- and trans-epoxides (Scheme 1),
a process which is particulary prone to occur for phenyl-
trans-epoxides through isomerization by simple bond rotation.3,5
Alternatively, the trans-epoxides may be formed from the cis-
olefins through carbocationic intermediates.3a To distinguish
between these two options, we selected (1R,2â,3R)-(2-ethenyl-
3-methoxycyclopropyl)benzene (4) as a mechanistic probe for
the Jacobsen-Katsuki epoxidation. A similar cyclopropane
derivative (a methyl instead of the vinyl substituent) has already
been used by Newcomb and co-workers6 to elucidate the
mechanism of the iron-catalyzed CH oxidation; they found that
cationic as well as radical intermediates are involved. Evidently,
(3) (a) Dalton, C. T.; Ryan, K. M.; Wall, V. M.; Bousquet, C.; Gilheany, D.
G. Top. Catal. 1998, 5, 75-91. (b) 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, 1720-1723. (c) Palucki, M.; Finney, N.
S.; Pospisil, P. J.; Gu¨ler, M. L.; Ishida, T.; Jacobsen, E. N. J. Am. Chem.
Soc. 1998, 120, 948-954. (d) Linker, T. Angew. Chem., Int. Ed. Engl.
1997, 36, 2060-2062.
(4) (a) Feichtinger, D.; Plattner, D. A. Angew. Chem., Int. Ed. Engl. 1997, 36,
1718-1719. (b) Feichtinger, D.; Plattner, D. A. J. Chem. Soc., Perkin Trans.
2 2000, 1023-1028. (c) Feichtinger, D.; Plattner, D. A. Chem.-Eur. J. 2001,
7, 591-599.
(5) (a) Srinivasan, K.; Michaud, P.; Kochi, J. K. J. Am. Chem. Soc. 1986, 108,
2309-2320. (b) Jacobsen, E. N.; Deng, L.; Furukawa, Y.; Mart´ınez, L. E.
Tetrahedron 1994, 50, 4323-4334. (c) Linde, C. Ph.D. Thesis, Royal
Institute of Technology, Stockholm, 1998.
* To whom correspondence should be addressed. E-mail: adam@
chemie.uni-wuerzburg.de. Fax: (internat.) +49-931/888-4756.
† University of Wu¨rzburg.
‡ Eidgeno¨ssische Technische Hochschule.
(1) (a) Johnson, R. A.; Sharpless, K. B. In Catalytic Asymmetric Synthesis;
Ojima, I., Ed.; VCH: New York, 1993; Chapter 4.1. (b) Jacobsen, E. N.
In Catalytic Asymmetric Synthesis; Ojima, I., Ed.; VCH: New York, 1993;
Chapter 4.2.
(6) (a) Newcomb, M.; Le Tadic-Biadatti, M.-H.; Chestney, D. L.; Roberts, E.
S.; Hollenberg, P. F. J. Am. Chem. Soc. 1995, 117, 12085-12091. (b)
Newcomb, M.; Shen, R.; Choi, S.-Y.; Toy, P. H.; Hollenberg, P. F.; Vaz,
A. D. N.; Coon, M. J. J. Am. Chem. Soc. 2000, 122, 2677-2686.
(2) (a) Katsuki, T. Coord. Chem. ReV. 1995, 140, 189-214. (b) Katsuki, T. J.
Mol. Catal. A 1996, 113, 87-107. (c) Ito, Y. N.; Katsuki, T. Bull. Chem.
Soc. Jpn. 1999, 72, 603-619.
9
5068
J. AM. CHEM. SOC. 2002, 124, 5068-5073
10.1021/ja0177206 CCC: $22.00 © 2002 American Chemical Society

cis-Stilbene as a Mechanistic Probe in Epoxidation
A R T I C L E S
Scheme 2. Vinylcyclopropane 4 as a Mechanistic Probe To
Scheme 1. Possible Mechanism of the Isomerization in the
Jacobsen-Katsuki Epoxidation of cis-Stilbene (1)
Distinguish between Radical and Cation Intermediates
the advantage of the vinylcyclopropane 4 over the simpler
probes that have been employed so far in the Jacobsen-Katsuki
epoxidation7 is the fact that it differentiates between cationic
and radical intermediates (Scheme 2).
Among the factors that influence the diastereoselectivity in
the epoxidation of cis-stilbene (1), it was recently shown that
ligation of the counterion in the Mn(salen)X complex 3 plays
an important role.8 Thus, the cis/trans-epoxide ratio was ca.
30:70 (extensive isomerization) for cis-stilbene, when the
complexes 3 were employed with the ligating counterions Cl-,
Br-, and AcO-. In contrast, the cis/trans ratio is ca. 75:25
(moderate isomerization) for the complexes 3 with the non-
ligating counterions BF4-, PF6-, and SbF6-. This counterion
effect was rationalized mechanistically in terms of the two-state-
reactiVity model.8,9
That also the oxygen source may affect the selectivity of
metal-catalyzed oxidations has recently been demonstrated by
Nam and co-workers, who have investigated the mechanism of
the epoxidation by iron complexes.10 When peroxidic oxygen
donors (OxD) such as hydrogen peroxide, tert-butyl hydroper-
oxide, and mCPBA, were employed with the Fe(porph) com-
plexes, initially an FeOOR(porph) adduct is formed, which
oxidizes the substrate by the three different pathways A-C
(Scheme 3). First, the FeOOR(porph) adduct may epoxidize
olefins directly, which constitutes Lewis-acid activation of the
oxygen donor ROOH (path A); alternatively, the O,O bond of
Scheme 3. Three Pathways A-C for the Reaction of the
FeOOR(porph) Adduct in the Iron-Catalyzed Epoxidation
the oxygen donor ROOH may be cleaved either heterolytically
(path B) to yield an Fe(V)oxo species [actually the Fe(IV)oxo/
porphyrin radical cation11] or homolytically (path C) to give an
Fe(IV)oxo complex. Whereas the Fe(V)oxo species affords high
yields of epoxides, the Fe(IV)oxo complex preferably performs
allylic oxidation. The ratio between heterolytic and homolytic
cleavage depends on the type of oxygen donor ROOH and the
axial ligand L. Evidently, this study provides clear-cut evidence
for the participation of several metal-activated intermediates as
active oxidants.
In view of the experimental facts for the Fe(porph) catalyst,
it was of mechanistic relevance to examine also the influence
of the oxygen donor on the diastereoselectivity of the MnIII-
(salen)-catalyzed epoxidation of cis-stilbene (1). Besides the
routinely used iodosyl benzene (PhIO) and bleach (NaxOClQ),
we decided to employ also iodosyl pentafluorobenzene (C6F5-
IO),12a periodate (TBAxIO4Q),12b ozone, hydrogen persulfate
(TBAxHSO5Q),12c and dimethyldioxirane (DMD)12d as oxygen
(7) (a) Fu, H.; Look, G. C.; Zhang, W.; Jacobsen, E. N.; Wong, C.-H. J. Org.
Chem. 1991, 56, 6497-6500. (b) Linde, C.; Arnold, M.; Norrby, P.-O.;
Åkermark, B. Angew. Chem., Int. Ed. Engl. 1997, 36, 1723-1725. (c)
Adam, W.; Mock-Knoblauch, C.; Saha-Mo¨ller, C. R.; Herderich, M. J.
Am. Chem. Soc. 2000, 122, 9685-9691.
(8) Adam, W.; Roschmann, K. J.; Saha-Mo¨ller, C. R. Eur. J. Org. Chem. 2000,
3519-3521.
(9) Schro¨der, D.; Shaik, S.; Schwarz, H. Acc. Chem. Res. 2000, 33, 139-145.
(10) (a) Nam, W.; Lim, M. H.; Lee, H. J.; Kim, C. J. Am. Chem. Soc. 2000,
122, 6641-6647. (b) Nam, W.; Han, H. J.; Oh, S.-Y.; Lee, Y. J.; Choi,
M.-H.; Han, S.-Y.; Kim, C.; Woo, S. K.; Shin, W. J. Am. Chem. Soc. 2000,
122, 8677-8684. (c) Nam, W.; Lim, M. H.; Moon, S. K.; Kim, C. J. Am.
Chem. Soc. 2000, 122, 10805-10809. (d) Nam, W.; Lim, M. H.; Oh, S.-
Y.; Lee, J. H.; Lee, H. J.; Woo, S. K.; Kim, C.; Shin, W. Angew. Chem.,
Int. Ed. 2000, 39, 3646-3649.
(11) Dawson, J. H. Science 1988, 240, 433-439.
(12) (a) Schmeisser, M.; Dahmen, K.; Sartori, P. Chem. Ber. 1967, 100, 1633-
1637. (b) Pietika¨inen, P. Tetrahedron Lett. 1995, 36, 319-322. (c)
Pietika¨inen, P. Tetrahedron 2000, 56, 417-424. (d) Ferrer, M.; Gibert,
M.; Sa´nchez-Baeza, F.; Messeguer, A. Tetrahedron Lett. 1996, 37, 3585-
3586.
9
J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002 5069

A R T I C L E S
Adam et al.
Table 1. Counterion and Oxygen-Donor Effects on the Cis/Trans
Ratios of the Epoxides 2 in the MnIII-Catalyzed Epoxidation of
cis-Stilbene (1)
Scheme 4. Synthesis of the Epoxide 5 and the Cationic
Ring-Opening Products 6 and A-C
epoxidesb
cis/trans
entry
[OxD]
PhIO
catalyst
convna [%]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
3a (X ) Cl)
3b (X ) PF6)
3a (X ) Cl)
3b (X ) PF6)
3a (X ) Cl)
3b (X ) PF6)
3a (X ) Cl)
3b (X ) PF6)
3a (X ) Cl)
3b (X ) PF6)
3a (X ) Cl)
3b (X ) PF6)
3a (X ) Cl)
3b (X ) PF6)
43
64
89
95
62
67
47
86
21
20
14
26
39
31
29:71
76:24
47:53
82:18
38:62
40:60
56:44
90:10
57:43
77:23
75:25
69:31
75:25
92:8
C6F5IO
Q
TBAxIO4
c
O3
Q
TBAxHSO5
NaxOClQ
DMDd
cis-stilbene (1); yet even at this low temperature, the oxidative
cleavage of stilbene to benzaldehyde was the main reaction (ca.
75%). Nevertheless, the cis/trans ratios obtained with ozone
demonstrate again a pronounced counterion effect with the
catalysts 3a and 3b, although the cis-epoxide dominates in both
cases (entries 7 and 8).
a Conversion is relative to cis-stilbene (1); mass balances were >80%.
b The cis/trans ratios were determined by 1H NMR analysis directly on the
crude product mixture; all cis/trans ratios have been run at least in duplicate
and are reproducible within an error of (5% of the stated values. c The
reaction was performed at -78 °C and gave benzaldehyde as the main
product (ca. 75%) due to ozonolysis. d Dimethyldioxirane, used as acetone-
free CH2Cl2 solution (0.185 M), see ref 12d.
In the case of TBAxHSO5Q, more isomerization takes place
with catalyst 3a than with catalyst 3b (entries 9 and 10). When
compared to PhIO, less isomerization is observed with
sources. As catalysts, the Mn(salen)X complexes 3a (X ) Cl)
and 3b (X ) PF6) were used to evaluate - as in the case of
PhIO - whether a counterion effect operates on the diastereo-
selectivity also for these oxygen donors.
Q
TBAxHSO5 for catalyst 3a (entries 1 and 9), whereas for
catalyst 3b the cis/trans ratio stays about the same within
experimental error (entries 2 and 10). Significantly, the oxygen
donor NaxOClQ affords almost the same cis/trans diastereo-
selectivity for both catalysts 3a and 3b (entries 11 and 12), with
the cis-epoxide preferred.
Results
The manganese-catalyzed epoxidation of cis-stilbene (1) was
carried out with 10 mol % of catalyst 3 and 1 equiv of the
oxygen donor (OxD). The results are summarized in Table 1.
For the set of oxygen donors examined herein, the general trend
displayed by catalyst 3a (X ) Cl) is that the cis/trans ratio ranges
from mainly trans-2 (entry 1) to mainly cis-2 (entries 11 and
13), whereas with catalyst 3b (X ) PF6), the cis-epoxide prevails
for all oxygen donors, except TBAxIO4Q (entry 6). Specifically,
when iodosyl benzene (PhIO) is used as oxygen donor, the
MnIII(salen)Cl catalyst 3a leads mostly to isomerization (cis/
trans 29:71, entry 1), whereas the MnIII(salen)PF6 catalyst 3b
affords mostly the cis product (cis/trans 76:24, entry 2).
Similarly, the related iodosyl pentafluorobenzene (C6F5IO) also
displays a counterion effect on the diastereoselectivity, but less
pronounced. Thus, the cis/trans ratio of 47:53 for catalyst 3a
with C6F5IO (entry 3) indicates less isomerization than with
PhIO (entry 1), while for the catalyst 3b (X ) PF6), the cis/
trans ratio is nearly constant for both iodosyl oxygen donors
(entries 2 and 4).
The least isomerization is exhibited by dimethyldioxirane
(DMD) as oxygen source for both catalysts 3a and 3b (entries
13 and 14). Direct epoxidation by DMD rather than metal-
catalyzed oxygen transfer by the manganese catalyst 3 is
unlikely, since DMD in combination with the Jacobsen Mn-
(salen*)Cl catalyst epoxidizes 2,2-dimethyl-2H-chromenes highly
enantioselectively (83-93% ee).13 If the achiral DMD were the
oxidant, necessarily racemic epoxide would be formed.
The (1R,2â,3R)-(2-ethenyl-3-methoxycyclopropyl)benzene
(4) was synthesized according to the literature procedure;6a its
epoxide 5 was prepared by DMD epoxidation as a 54:46 mixture
of diastereomers (Scheme 4). The epoxide 5 is extremely
sensitive to acids, and it decomposes on attempted purification
by chromatography on deactivated silica gel and even on Florisil.
This presented severe difficulties in the workup of the epoxi-
dation mixture that is obtained with the Jacobsen catalyst Mn-
(salen)Cl (3a), which requires removal of the paramagnetic
manganese species by silica gel chromatography to enable
product analysis by NMR spectroscopy. Fortunately, the epoxide
5 persists exposure to the Mn(salen)Cl complex 3a (established
by a control experiment), and the polymer-bound Mn(salen)
The tetrabutylammonium periodate (TBAxIO4Q) as oxygen
donor is exceptional, for which no counterion effect is observed.
Not only is the cis/trans ratio the same (ca. 40:60) within
experimental error for both catalysts 3a and 3b (entries 5 and
6), but TBAxIO4 is the only oxygen donor for which
isomerization to the trans diastereomer dominates also for
complex 3b.
Q
In contrast to the other oxygen sources, the epoxidation with
ozone was performed at -78 °C to minimize the ozonolysis of
(13) Adam, W.; Jeko, J.; Le´vai, A.; Maier, Z.; Nemes, C.; Patonay, T.; Pa´rka´nyi,
L.; Sebok, P. Tetrahedron: Asymmetry 1996, 7, 2437-2446.
9
5070 J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002

cis-Stilbene as a Mechanistic Probe in Epoxidation
A R T I C L E S
Scheme 5. Catalytic Cycle for the MnIII-Catalyzed Epoxidation of
cis-Stilbene (1) by Lewis-Acid Activation (Path 1) versus MnV(oxo)
Species (Path 2)
catalyst p-MnCl was employed to circumvent these problems
by simple filtration (see Supporting Information).14
Furthermore, for the same reasons, the usual HPLC analysis
on a silica gel column, to assess the product distribution in the
epoxidation mixture, had to be conducted on a reversed-phase
column. Although under these HPLC conditions there is also
some decomposition of the epoxide 5, the resulting products
did fortunately not elute to interfere with the product analysis.
The authentic mixture of the cationic ring-opening products
6 was prepared by treatment of the epoxide 5 with catalytic
amounts of Cr(salen)PF6 as Lewis acid15 (see Supporting
Information). Whereas 12% of the cyclic acetal 6a was isolated
after column chromatography (Figure S1), no aldehyde 6b was
detected; instead, as outlined in Scheme 4, a mixture of three
unknown products A, B, and C (HPLC analysis, Figure S2)
was formed in 77% yield (determined gravimetrically).
In a test reaction, cis-stilbene (1) was treated with p-MnCl,
4-phenylpyridine N-oxide (PPNO), and iodosyl benzene (PhIO)
to yield a 63:37 mixture of the diastereomeric epoxides 2 at a
conversion of 67%. When the probe (1R,2â,3R)-(2-ethenyl-3-
methoxycyclopropyl)benzene (4) was epoxidized under these
conditions, 83% of the olefin was consumed to give 2% of
epoxide 5 and 33% of ring-opened products, at a mass balance
of 51% (see Figure S3). No cationic ring-opening products 6
or A-C were detected by RP-HPLC analysis (see Figures S4
and S5). Consequently, the observed ring-opening products
derive from radical intermediates, and, hence, the following
mechanistic discussion shall be limited to such species.
to the cis-epoxide or undergo C,C-bond rotation and cyclize to
the trans-epoxide. If the MnV(oxo) complex were the only
oxidant in this epoxidation, irrespective of which oxygen donor
is used to generate the reagent, the cis/trans selectivity should
be the same. Yet this is not the case (Table 1); besides some
qualitative similarities, each oxygen donor displays a distinct
cis/trans-epoxide ratio. Thus, the Jacobsen-Katsuki catalytic
cycle must be extended to accommodate this divergence
mechanistically (Scheme 5). The fact that the type of oxygen
donor affects the diastereoselectivity of the epoxidation requires
that besides the MnV(oxo) species, at least one other oxidant
must be involved in this catalytic process. Consequently, we
propose the following unprecedented diastereoselectivity-
controlling bifurcation step in the catalytic cycle for this
epoxidation: After ligation between the MnIII catalyst and the
oxygen donor (OxD) to result in the MnIII(OLG) adduct, the
MnV(oxo) oxidant is released by splitting off the leaving group
LG, and subsequent unselective epoxidation of the substrate
(pathway 1) affords a mixture of cis- and trans-epoxides 2.
Alternatively, the ligated MnIII(OLG) adduct functions directly
as Lewis-acid-activated epoxidant by concerted oxygen transfer
to give the cis-epoxide (pathway 2). The latter alternative has
already been suggested for the Mn(salen)-16a and Fe(porph)-
catalyzed10,17 epoxidation with mCPBA as oxygen donor and
for the manganese-catalyzed sulfimidation.16b The transition
structure for the concerted oxygen transfer by the MnIII(OxD)
Discussion
From the data listed in Table 1, it may be concluded that not
only the counterion of the Mn(salen)X complexes 3a,b (X )
Cl, PF6), but also the oxygen donor [OxD], has an influence on
the cis/trans diastereoselectivity of the Jacobsen-Katsuki ep-
oxidation of cis-stilbene (1). We shall first address the influence
of the oxygen donor on the diastereoselectivity and subsequently
the counterion effect. Although we mechanistically interpret the
stereochemical trends separately in terms of the oxygen-donor
and counterion effects, it should be kept in mind that both factors
are intimately intertwined and cause the complexity of the
observed diastereoselectivity.
The Effect of the Oxygen Donor. It is generally accepted
that the first step in the Jacobsen-Katsuki epoxidation is the
formation of the MnV(oxo) species as the actual oxidant.2-4
After addition of the MnV(oxo) species to cis-stilbene (1), the
resulting radical intermediate may either collapse immediately
Q
adduct with HSO5 as oxygen donor may be similar to the
“butterfly” structure for peracid epoxidations.
An appropriate admixture of the stepwise epoxidation by the
established MnV(oxo) species (path 1 affords a mixture of the
cis- and trans-epoxides2,3) and of the concerted epoxidation (path
2 yields only cis-epoxide18) accounts for the stereochemical data
in Table 1.19
As shown in the bifurcation step of Scheme 5, the formation
of the MnV(oxo) species depends on the leaving group LG of
the oxygen donor; thus, the type of oxygen donor [OxD] is
responsible for the observed cis/trans diastereoselectivity. This
is most apparent in the cis/trans ratios for catalyst 3a, which
(16) (a) Bryliakov, K. P.; Babushkin, D. E.; Talsi, E. P. J. Mol. Catal. A 2000,
158, 19-35. (b) Ohta, C.; Katsuki, T. Tetrahedron Lett. 2001, 42, 3885-
3888.
(17) (a) Vaz, A. D. N.; McGinnity, D. F.; Coon, M. J. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 3555-3560. (b) Toy, P. H.; Newcomb, M.; Coon, M. J.;
Vaz, A. D. N. J. Am. Chem. Soc. 1998, 120, 9718-9719.
(18) Yang, Y.; Diederich, F.; Valentine, J. S. J. Am. Chem. Soc. 1991, 113,
7195-7205.
(14) Sellner, H.; Karjalainen, J. K.; Seebach, D. Chem.-Eur. J. 2001, 7, 2873-
2887.
(15) (a) Mart´ınez, L. E.; Leighton, J. L.; Carsten, D. H.; Jacobsen, E. N. J. Am.
Chem. Soc. 1995, 117, 5897-5898. (b) Larrow, J. F.; Schaus, S. E.;
Jacobsen, E. N. J. Am. Chem. Soc. 1996, 118, 7420-7421. (c) Lebel, H.;
Jacobsen, E. N. Tetrahedron Lett. 1999, 40, 7303-7306.
9
J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002 5071

A R T I C L E S
Adam et al.
vary from dominant trans selectivity for the PhIO oxygen donor
(entry 1) to dominant cis selectivity for DMD (entry 13). For
this reason, we shall analyze the effect of the oxygen donor for
catalyst 3a and thereby keep the counterion constant, i.e., X )
Cl.
Therefore, more cis-epoxide is observed with O3 as compared
to with PhIO (entries 1 and 7).
Effect of the Counterion. While for all oxygen donors a
counterion effect on the diastereoselectivity is observed, this is
Q
not the case for TBAxIO4 (entries 5 and 6) and NaxOClQ
Evidently, based on the cis/trans ratios, the oxygen donors
fall into two classes: PhIO, C6F5IO, TBAxIO4Q, and O3 afford
the MnV(oxo) species as the dominant oxidant with extensive
isomerization (pathway 1) through the stepwise radical process
(cis/trans ratio 29:71 to 56:44 for catalyst 3a, entries 1, 3, 5,
and 7); for TBAxHSO5Q, NaxOClQ, and DMD, the concerted
process through Lewis-acid catalysis (pathway 2) is also
operative, as reflected by the higher cis selectivity with catalyst
3a (cis/trans ratio 57:43 to 75:25, entries 9, 11, and 13).
However, if for the first set of oxygen donors the MnV(oxo)
were the principal oxidant, the same diastereoselectivity should
be displayed by them since the leaving group of the oxygen
donor is no longer involved. This is definitively not the case,
because all cis/trans ratios differ (entries 1, 3, 5, and 7).
For both C6F5IO and PhIO, the MnV(oxo) species is the main
oxidant, but the small differences in the diastereoselectivity for
the ClQ as counterion (see entries 1 versus 3) may be explained
by the fact that the former is capable of oxidizing the chloride
counterion of complex 3a to the hypochlorite ion but PhIO is
not. This has been confirmed by a control experiment with
TEBAxClQ as the chloride source (see Supporting Information).
Consequently, the in-situ-generated OClQ competes as oxygen
donor with C6F5IO. Because from the data in Table 1 (entry
11) we suppose that the hypochlorite ion epoxidizes mainly by
the concerted Lewis-acid-catalyzed pathway 2, a higher cis
selectivity is observed for C6F5IO, as compared to PhIO (entries
(entries 11 and 12). For these oxygen donors, the cis/trans ratios
are about the same (within the experimental error) for both
catalysts 3a and 3b. This may be rationalized by the fact that a
ligating anion (IO3Q and ClQ) is released from the oxygen donor
Q
(IO4 and OClQ), which then coordinates to the positively
charged manganese catalyst. Thus, both catalysts 3a and 3b are
ligated by the same counterion, causing similar stereoselectivi-
ties. The ligating properties of IO3 and ClQ have been
Q
confirmed by means of control experiments (see Supporting
Information).
The counterion effect on the cis/trans selectivity in the
manganese-catalyzed epoxidation of cis-stilbene (1) was previ-
ously rationalized in terms of the two-state-reactiVity model.8
While our results (Table 1) are consistent with this model, recent
computations23 on the mechanism of the oxygen transfer in the
Jacobsen-Katsuki epoxidation are informative, which show that
the singlet, triplet, and quintet spin states are all accessible in
the MnV(oxo) complex. From the qualitative energy profiles
(Figure 1), it is evident that for the neutral MnV(oxo) species23c
with chloride as counterion, the singlet state is lowest in energy
(Figure 1a), whereas in the case of the cationic MnV(oxo)
species23a (Figure 1b), the triplet is the ground state; clearly,
the energies of the spin states in the “naked cation” are spaced
more closely. For the final MnIII(epoxide) adduct, the situation
is reversed. For the chloride as counterion, the energies of the
spin states fall within a much narrower range than for the
hexafluorophosphate; in both cases, the quintet is favored as
ground state. According to these calculations, the nature of the
counterion X affects drastically the relative energy ordering of
the spin states in the MnV(oxo) complex as well as of the final
MnIII(epoxide) adduct.
Q
1 and 3). As for the oxygen donor TBAxIO4 , the slightly higher
cis selectivity (entry 5) may be caused in the same way, since
the reaction of periodate with chloride to form iodate and
hypochlorite is feasible (according to their oxidation poten-
tials20), whereas for TBAxHSO5Q (entry 9), this possibility was
ruled out by a control experiment. Also for DMD (entry 13),
the OClQ oxygen donor is generated in situ by oxidation of
ClQ,21 as is evident from the 75:25 cis/trans ratio.
The somewhat higher cis selectivity for ozone (entry 7) versus
PhIO (entry 1) with catalyst 3a may be explained in terms of a
temperature effect since the experiments with O3 were performed
at -78 °C, and those with PhIO at ca. 20 °C (a comparative
run with PhIO was not possible at -78 °C, because no reaction
occurs at this low temperature). At -78 °C, the C,C-bond
rotation responsible for cis/trans isomerization is sufficiently
slowed22 such that ring closure becomes the faster process.
Moreover, the computations reveal a mechanistically impor-
tant feature with respect to the effect of the counterion on the
stereoselective behavior of the three spin states. Whereas the
singlet state of MnV(oxo)Cl is predicted to epoxidize cis-alkenes
concertedly (the energy curve displays no minimum) and hence
diastereoselectively, the triplet and quintet states involve step-
wise oxygen transfer (the energy curves display minima) and
should undergo extensive cis/trans isomerization. As for the
MnV(oxo)x complex, the triplet und singlet states should display
a similar stereoselectivity as the neutral MnV(oxo)Cl species;
that is, the triplet should be unselective (stepwise) and the singlet
selective (concerted), but the quintet state should transfer its
oxygen atom concertedly in a diastereoselective manner.
(19) The participation of a MnIV(oxo) species may also be proposed to explain
the observed diastereoselectivity, since it is known to cause isomerization.
[(a) Groves, J. T.; Stern, M. K. J. Am. Chem. Soc. 1987, 109, 3812-3814.
(b) Groves, J. T.; Stern, M. K. J. Am. Chem. Soc. 1988, 110, 8628-8638.
(c) Lee, R. W.; Nakagaki, P. C.; Bruice, T. C. J. Am. Chem. Soc. 1989,
111, 1368-1372. (d) Arasasingham, R. D.; He, G.-X.; Bruice, T. C. J.
Am. Chem. Soc. 1993, 115, 7985-7991. (e) Groves, J. T.; Lee, J.; Marla,
S. S. J. Am. Chem. Soc. 1997, 119, 6269-6273.] On the basis of our
experimental data, this mechanistic alternative cannot be the dominant
pathway: When cis-stilbene (1) is treated with the authentic MnIV(oxo)
species in EtOAc, that is, under stoichiometric conditions, a 36:64 mixture
of the cis- and trans-epoxides 2 is formed;7c however, a control experiment
under our catalytic conditions in EtOAc gave even more isomerization (cis/
trans 25:75). Moreover, the authentic MnIV species leads to substantial
amounts of chlorinated products in CH2Cl2, which we do not observe.
(20) Lide, D. R. Handbook of Chemistry and Physics, 76th ed; CRC Press: Boca
Raton, FL, 1995; pp 8-21-8-26.
The qualitative energy profiles in Figure 1 suggest that for
hexafluorophosphate as counterion, the epoxidation is expected
to proceed mainly on the quintet surface; the triplet and the
singlet channels constitute only minor reaction pathways.
Because the MnV(oxo)x quintet transfers the oxygen atom
diastereoselectively, it follows that for the Mn(salen)PF6 catalyst
(23) (a) Linde, C.; Åkermark, B.; Norrby, P.-O.; Svensson, M. J. Am. Chem.
Soc. 1999, 121, 5083-5084. (b) Cavallo, L.; Jacobsen, H. Angew. Chem.,
Int. Ed. 2000, 39, 589-592. (c) Abashkin, Y. G.; Collins, J. R.; Burt, S.
K. Inorg. Chem. 2001, 40, 4040-4048. (d) Strassner, T.; Houk, K. N. Org.
Lett. 1999, 1, 419-421. (e) El-Bahraoui, J.; Wiest, O.; Feichtinger, D.;
Plattner, D. A. Angew. Chem., Int. Ed. 2001, 40, 2073-2076.
(21) Montgomery, R. E. J. Am. Chem. Soc. 1974, 96, 7820-7821.
(22) Lau, H. H. Angew. Chem. 1961, 73, 423-432.
9
5072 J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002

cis-Stilbene as a Mechanistic Probe in Epoxidation
A R T I C L E S
Figure 1. Qualitative energy profiles of the singlet (s, [), triplet (t, 9), and quintet (q, 2) trajectories for chloride (a) and hexafluorophosphate (b) as
counterions (ref 23).
3b the cis-epoxide 2 is formed preferentially. When the chloride
ion is the counterion, all three spin states of MnV(oxo)Cl
participate in the oxygen-transfer process. Thus, since only the
singlet state reacts diastereoselectively through a concerted
epoxidation and the triplet and quintet states give rise to
isomerization through the stepwise pathway, for catalyst 3a (X
) Cl) more trans-epoxide 2 is expected as compared to catalyst
3b (X ) PF6), as experimentally observed (Table 1). Indeed,
for the latter catalyst, the highest cis/trans ratio (92:8) is given
by DMD (entry 14), which follows almost exclusively the
stereoselective quintet pathway.
branching, the unselective epoxidation by way of the stepwise
radical process with the MnV(oxo) complex (path 1) competes
with the concerted Lewis-acid-activated selective epoxidation
by the MnIII(OxD) adduct (path 2). The counterion effect on
the stereoselectivity may be explained in terms of ligand-
dependent reaction profiles for the three available spin states
(singlet, triplet, and quintet) of the MnV(oxo)X species.
Acknowledgment. This work was supported by the Deutsche
Forschungsgemeinschaft
(Sonderforschungsbereich
347
“Selektive Reaktionen metallaktivierter Moleku¨le”) and the
Fonds der Chemischen Industrie. We are grateful to Dr. A.
Heckel and Dr. H. Sellner for providing us with the polymer-
bound salen ligand. Profs. M. Newcomb (Department of
Chemistry, University of Illinois at Chicago, IL) and P.-O.
Norrby (Department of Chemistry, Technical University of
Denmark, Lyngby) as well as Dr. M. Bro¨ring (Institute of
Inorganic Chemistry, University of Wu¨rzburg, Germany) are
thanked for helpful discussions. This work is dedicated to
Professor Albert Padwa, a helpful friend and excellent scientist,
on the occasion of his 65th birthday.
Conclusion
We have demonstrated that the oxygen donor (OxD) and the
counterion X of the Mn(salen)X catalysts 3a and 3b have a
definite influence on the diastereoselectivity of the MnIII-
catalyzed epoxidation of cis-stilbene (1). By application of the
mechanistic probe (1R,2â,3R)-(2-ethenyl-3-methoxycyclopro-
pyl)-benzene (4), which enabled us to distinguish between
radical and carbocationic intermediates through selective open-
ing of the cyclopropane ring, the participation of cationic species
in the Jacobsen-Katsuki epoxidation could be ruled out.
Consequently, the formation of stereochemically isomerized
epoxide product is attributed to radical pathways. To account
for the complex behavior of the diverse oxygen donors, the
generally accepted catalytic cycle had to be extended to
incorporate a bifurcation step (Scheme 5). For the latter product
Supporting Information Available: Experimental details and
spectral data (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
JA0177206
9
J. AM. CHEM. SOC. VOL. 124, NO. 18, 2002 5073