Catalytic asymmetric epoxidation of aldehydes. Optimization, mechanism, and discovery of stereoelectronic control involving a combination of anomeric and cieplak effects in sulfur ylide epoxidations with chiral 1,3-oxathianes

Article abstract of DOI:10.1021/ja9812150

A range of 1,3-oxathianes based on camphorsulfonic acid have been prepared and tested in the catalytic asymmetric epoxidation of carbonyl compounds. It was found that the 1,3-oxathiane derived from acetaldehyde 5b gave the highest yield and enantioselectivity in the epoxidation process. The enantioselectivity was independent of the solvent and metal catalyst used (although yields were dependent on both). The optimum conditions were applied to a range of aldehydes, and good enantioselectivities and diastereoselectivities were observed. The origin of the enantioselectivity was probed, and in particular the role of the oxygen of the 1,3-oxathiane was investigated. Thus, the sulfur and carbon analogues of the camphorsulfonic acid based 1,3-oxathiane (derived from formaldehyde) were prepared (i.e., 1,3-dithiane and thiane analogues). With this series of analogues the steric effects are minimized so that the electronic effects can be investigated. The series of compounds was reacted in the catalytic cycle with benzaldehyde and gave stilbene oxides with 44% ee (sulfur analogue), 41% ee (1,3-oxathiane), and 20% ee (carbon analogue). Thus, it was concluded that the oxygen of the 1,3-oxathiane exerted a significant electronic effect in controlling the face selectivity of the ylide reactions. This electronic effect was a result of combined anomeric (higher with the sulfur analogue, not present with the carbon analogue) and Cieplak effects. A strong anomeric effect was observed in the X-ray structures of one of the 1,3-oxathianes, and an even greater one was observed in the corresponding sulfoxide (this was used as an electronic analogue of the ylide). The face selectivity of the ylide was believed to be complete in reactions with 5b. The minor enantiomer resulted from reaction of the minor conformer of the ylide, reacting again with high face selectivity. This was proven by using a more substituted diazo compound, which was expected to give much less of the minor conformer. Indeed, reaction with mesityldiazomethane gave the corresponding epoxide in essentially enantiomerically pure form.

Full text of DOI:10.1021/ja9812150

8328
J. Am. Chem. Soc. 1998, 120, 8328-8339
Catalytic Asymmetric Epoxidation of Aldehydes. Optimization,
Mechanism, and Discovery of Stereoelectronic Control Involving a
Combination of Anomeric and Cieplak Effects in Sulfur Ylide
Epoxidations with Chiral 1,3-Oxathianes
Varinder K. Aggarwal,*,† J. Gair Ford,† S´ılvia Fonquerna,† Harry Adams,†
Ray V. H. Jones,‡ and Robin Fieldhouse‡
Contribution from the Department of Chemistry, UniVersity of Sheffield, Sheffield S3 7HF, England, and
Zeneca Process Technology Department, Earls Road, Grangemouth, Stirlingshire FK3 8XG, U.K.
ReceiVed April 10, 1998
Abstract: A range of 1,3-oxathianes based on camphorsulfonic acid have been prepared and tested in the
catalytic asymmetric epoxidation of carbonyl compounds. It was found that the 1,3-oxathiane derived from
acetaldehyde 5b gave the highest yield and enantioselectivity in the epoxidation process. The enantioselectivity
was independent of the solvent and metal catalyst used (although yields were dependent on both). The optimum
conditions were applied to a range of aldehydes, and good enantioselectivities and diastereoselectivities were
observed. The origin of the enantioselectivity was probed, and in particular the role of the oxygen of the
1,3-oxathiane was investigated. Thus, the sulfur and carbon analogues of the camphorsulfonic acid based
1,3-oxathiane (derived from formaldehyde) were prepared (i.e., 1,3-dithiane and thiane analogues). With this
series of analogues the steric effects are minimized so that the electronic effects can be investigated. The
series of compounds was reacted in the catalytic cycle with benzaldehyde and gave stilbene oxides with 44%
ee (sulfur analogue), 41% ee (1,3-oxathiane), and 20% ee (carbon analogue). Thus, it was concluded that the
oxygen of the 1,3-oxathiane exerted a significant electronic effect in controlling the face selectivity of the
ylide reactions. This electronic effect was a result of combined anomeric (higher with the sulfur analogue,
not present with the carbon analogue) and Cieplak effects. A strong anomeric effect was observed in the
X-ray structures of one of the 1,3-oxathianes, and an even greater one was observed in the corresponding
sulfoxide (this was used as an electronic analogue of the ylide). The face selectivity of the ylide was believed
to be complete in reactions with 5b. The minor enantiomer resulted from reaction of the minor conformer of
the ylide, reacting again with high face selectivity. This was proven by using a more substituted diazo compound,
which was expected to give much less of the minor conformer. Indeed, reaction with mesityldiazomethane
gave the corresponding epoxide in essentially enantiomerically pure form.
Introduction
However, although highly efficient and elegant, these methods
have limitations. The Sharpless epoxidation requires an allylic
alcohol, and the Jacobsen/Katsuki epoxidation generally requires
cis-substituted alkenes bearing a π-stabilizing substituent. Very
recently Shi has made a major breakthrough and found that
simple sugar-based ketones can epoxidize a wide range of
alkenes with high enantioselectivity.9-11
An alternative to oxidative processes for the synthesis of
epoxides is the reaction of sulfur ylides with aldehydes and
ketones.12-14 Sulfur ylide epoxidation is a carbon-carbon bond
forming reaction and as such offers a complementary method
to the oxidative processes described above.
The search for efficient methods for the preparation of
enantiomerically pure epoxides continues unabated, particularly
as chiral epoxides are important intermediates in the synthesis
of pharmaceuticals and agrochemicals. Most attention has
focused on asymmetric oxidations of alkenes, and the methods
developed by Sharpless and Jacobsen/Katsuki have stood the
test of time and emerged as the best. Indeed, these methods
represent landmark achievements in asymmetric synthesis.1-8
† University of Sheffield.
‡ Zeneca Process Technology Department.
The first attempts at carrying out asymmetric epoxidations
using chiral sulfur ylides were performed by Trost in 1973, using
(1) Berrisford, D. J.; Bolm, C.; Sharpless, K. B. Angew. Chem., Int. Ed.
Engl. 1995, 34, 1059-1070.
(2) Besse, P.; Veschambre, H. Tetrahedron 1994, 50, 8885-8927.
(3) Finn, M. G.; Sharpless, K. B. In Asymmetric Synthesis; Morrison, J.
D., Ed.; Academic Press: New York, 1985; Vol. 5, pp 247-308.
(4) Jacobsen, E. N. In Catalytic Asymmetric Synthesis; Ojima, I., Ed.;
VCH: New York, 1993; pp 159-202.
(5) Johnson, R. A.; Sharpless, K. B. In Catalytic Asymmetric Synthesis;
Ojima, A., Ed.; VCH: New York, 1993; pp 103-158.
(6) Katsuki, T. Coord. Chem. ReV. 1995, 140, 189-214.
(7) Katsuki, T.; Martin, V. S. Org. React. 1996, 48, 1-299.
(8) Rossiter, B. E. In Asymmetric Synthesis; Morrison, J. D., Ed.;
Academic Press: New York, 1985; Vol. 5, pp 193-246.
(9) Wang, Z.-X.; Tu, Y.; Frohn, M.; Shi, Y. J. Org. Chem. 1997, 62,
2328-2329.
(10) Wang, Z.-X.; Tu, Y.; Frohn, M.; Zhang, J.-R.; Shi, Y. J. Am. Chem.
Soc. 1997, 119, 11224-11235.
(11) Tu, X.; Wang, Z.-X.; Shi, Y. J. Am. Chem. Soc. 1996, 118, 9806-
9807.
(12) For a review see: Li, A.-H.; Dai, L.-X.; Aggarwal, V. K. Chem.
ReV. 1997, 97, 2341-2372.
(13) Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1962, 84, 867.
(14) Johnson, A. W.; Lacount, R. B. J. Am. Chem. Soc. 1961, 83, 417.
S0002-7863(98)01215-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/08/1998

Catalytic Asymmetric Epoxidation of Aldehydes
J. Am. Chem. Soc., Vol. 120, No. 33, 1998 8329
Scheme 1
Scheme 3
Scheme 2
an adamantyl-based chiral sulfur ylide (Scheme 1).15 Despite
the large difference in size between the groups attached to sulfur,
no asymmetric induction was obtained. This negative result
presumably discouraged further work in asymmetric carbonyl
epoxidation for many years, especially as it must have been
difficult to conceive how the groups attached to sulfur could
be further differentiated in order to improve the levels of
enantioselectivity.
More recently a number of chiral sulfides have been
developed and those of Durst16,17(1) and Solladie´-Cavallo18 (2)
provide very impressive levels of enantioselectivity in benzyl-
idine ylide epoxidation (Scheme 2). In contrast to benzylidine
ylide epoxidation, methylene ylide epoxidation using 1 (methyl
instead of benzyl attached to S) was much less effective (less
than 4% ee).16 No doubt if Trost had attempted to carry out
benzylidine rather than methylene ylide epoxidation, higher
levels of enantioselectivity would have been achieved and
asymmetric sulfur ylide epoxidations would have been born
much earlier.
Scheme 4
The asymmetric ylide epoxidations developed by Durst and
Solladie´-Cavallo, while giving high levels of asymmetric
induction, suffer from requiring stoichiometric amounts of
sulfide, and attempts have been made to make the process
catalytic. The reaction of a sulfur ylide with an aldehyde gives
epoxide and returns sulfide. Thus, to render the epoxidation
process catalytic in sulfide, it is necessary to form the sulfur
ylide in situ. This was achieved, originally by Furukawa19 and
more recently by Dai,20 by sulfide alkylation and deprotonation
in the presence of an aldehyde (Scheme 3). Using chiral
sulfides, epoxides were obtained in variable yields but also with
only moderate enantioselectivity. This process is, however,
limited to simple, non-enolizable aldehydes.
Indeed, we recently reported the successful application of this
strategy to carbonyl epoxidation using catalytic quantities of
sulfide (Scheme 4).23 As these reaction conditions do not
require the use of strong base, this method can be applied to
readily enolizable and even base-sensitive aldehydes.24,25 In
this paper we describe the design and optimization of chiral
sulfides for use in the catalytic cycle (Scheme 4) to produce
nonracemic epoxides26 and our studies into the mechanism of
the enantioselectivity.
Results and Discussion
An alternative method for ylide formation involves the
Optimization of Reaction Conditions. In the design of the
chiral sulfides two factors were deemed important. First, a
single sulfur ylide should be produced from the reaction of the
sulfide with the metal carbenoid. If diastereomeric sulfur ylides
were formed, they would certainly react with different and
possibly opposite selectivities thereby eroding the enantio-
reaction of a sulfide with a carbene or metal carbenoid.21,22
(15) Trost, B. M.; Hammen, R. F. J. Am. Chem. Soc. 1973, 95, 962-
964.
(16) Breau, L.; Durst, T. Tetrahedron: Asymmetry 1991, 2, 367-370.
(17) Breau, L.; Ogilvie, W. W.; Durst, T. Tetrahedron Lett. 1990, 31,
35-38.
(18) Solladie´-Cavallo, A.; Diep-Vohuule, A.; Sunjic, V.; Vinkovic, V.
Tetrahedron: Asymmetry 1996, 7, 1783-1788.
(19) Furukawa, N.; Sugihara, Y.; Fujihara, H. J. Org. Chem. 1989, 54,
4222-4224.
(23) Aggarwal, V. K.; Abdel-Rahman, H.; Jones, R. V. H.; Lee, H. Y.;
Reid, B. D. J. Am. Chem. Soc. 1994, 116, 5973-5974.
(24) Aggarwal, V. K.; Abdel-Rahman, H.; Jones, R. V. H.; Standen, M.
C. H. Tetrahedron Lett. 1995, 36, 1731-1732.
(25) Aggarwal, V. K.; Abdel-Rahman, H.; Fan, L.; Jones, R. V. H.;
Standen, M. C. H. Chem.sEur. J. 1996, 2, 212-218.
(26) For a preliminary account see: Aggarwal, V. K.; Ford, J. G.;
Thompson, A.; Jones, R. V. H.; Standen, M. C. H. J. Am. Chem. Soc. 1996,
118, 7004-7005.
(20) Li, A.; Dai, L.; Hou, X.; Huang, Y.; Li, F. J. Org. Chem. 1996, 61,
489-493.
(21) Doyle, M. P. In ComprehensiVe Organometallic Chemistry; Abel,
E. W., Stone, F. G. A., Wilkinson, G., Eds.; Elsevier: Oxford, 1995; Vol.
12, pp 421-468.
(22) Padwa, A.; Hornbuckle, S. F. Chem. ReV. 1991, 91, 263-309.

8330 J. Am. Chem. Soc., Vol. 120, No. 33, 1998
Aggarwal et al.
Scheme 5
Scheme 6
Scheme 7
surprised to obtain greatly reduced yields of the epoxide, with
concomitant increased amounts of stilbenes (Scheme 6).
Stilbenes arise from the reaction of the metal carbenoid with
phenyldiazomethane,31,32 a reaction that competes with the
formation of the intermediate sulfur ylide (Scheme 7). In-
creased amounts of stilbene and reduced yields of epoxide
indicate that the former reaction is dominating over the latter.
As we had previously shown that 0.2 equiv of sulfide were
sufficient to obtain good yields of epoxides in the catalytic
process,25 the much reduced yields obtained with thioacetal 5b
suggested that the thioacetal was undergoing decomposition,
perhaps catalyzed by Cu(acac)2, thereby reducing the amount
available for epoxidation. This was tested by subjecting
thioacetal 5b to Cu(acac)2 in reagent grade CH2Cl2, and
complete hydrolysis was observed after a few seconds. Under
strictly anhydrous conditions the rate of hydrolysis could be
reduced, but the desired improvement in yield was not obtained.
Table 1. Preparation of Sulfides 5a-5l from Mercaptoisoborneol
entry
precursor
thioacetal yield/%
1
2
3
4
5
6
7
8
9
10
11
12
13
14
acetaldehyde
isobutyraldehyde
pivaldehyde
benzaldehyde
5b
5c
5d
5e
5f
5 g
5h
5i
5j
5k
5l
5m
5r
5s
99
100
95
84
98
85
81
83
97
94
81
93
0
phenylacetaldehyde
phenoxyacetaldehyde dimethyl acetal
methoxyacetaldehyde dimethyl acetal
benzyloxyacetaldehyde dimethyl acetal
2,2-dimethoxypropane
cyclobutanone
glycolaldehyde diethyl acetal
chloral hydrate
The solution to the problem was eventually found by using
Cu(acac)2 that was prepared ourselves from copper oxide and
acetylacetone and purified by sublimation33 instead of the
commercial material. When this batch of Cu(acac)2 was used,
little hydrolysis was observed in our test experiments, indicating
that the commercial material contained small quantities of
impurities that showed high levels of catalytic activity in
thioacetal hydrolysis. When the purified Cu(acac)234 was used
in the catalytic cycle with 0.2 equiv of thioacetal 5b, stilbene
oxide was obtained in good yield and with the same enantio-
selectivity. Upon discovery of suitable conditions for the
catalytic epoxidation process, the other thioacetals were tested
(Scheme 8), and the results are summarized in Table 2.
trifluoroacetaldehyde dimethyl acetal
trimethyl orthoformate
0
selectivity obtained.27 This could best be achieved by incor-
porating the sulfide within a rigid cyclic structure and making
one of the lone pairs more accessible than the other. Indeed,
the best chiral sulfides all have the sulfide moiety in a cyclic
structure, and single diastereomeric sulfonium salts are formed
upon alkylation. Second, we wanted to design chiral sulfides
that could be easily modified, to explore both electronic and
steric effects in the epoxidation process. Cyclic thioacetals of
general structure 5 fulfill these requirements. A single sulfur
ylide should be formed in the reaction of the sulfide with the
metal carbenoid as the equatorial lone pair is much more
accessible than the axial lone pair, which is hindered by the
bridging geminal dimethyl group. The presence of the thioacetal
moiety enables the R and R′ groups to be easily varied in order
to probe steric and electronic effects by the use of different
carbonyl compounds.
Sulfide 5b (R ) Me, entry 2) proved to be the most effective
catalyst both in terms of enantioselectivity and yield. Increasing
the size of the R group did not give an increase in enantio-
selectivity but instead lowered the yield (entries 2-4). For R
) But (5d, entry 4), no epoxide was obtained; stilbenes were
formed instead. Evidently the rate at which the sulfide reacts
with the metal carbenoid (k1) is dependent upon the size of R
and R′ (Scheme 7), and if either is too hindered, as in the case
of sulfide 5d, k1 slows down to such an extent that k2
dominates.25
A range of thioacetals were easily prepared from camphor-
sulfonyl chloride via the known hydroxy thiol (Scheme 5, Table
1).28-30 Thioacetal 5b was initially tested in the epoxidation
process using benzaldehyde and slow addition of phenyldiazo-
methane. Using stoichiometric amounts of 5b, trans-stilbene
oxide was obtained in good yield and in very high enantiomeric
excess. With catalytic amounts of the sulfide, however, we were
For R ) Ph (5e, entry 5), no epoxide was formed but only
limited amounts of stilbenes were obtained. In this case, ring
expanded thioacetals 7 were believed to have been formed via
(31) Shankar, B. K. R.; Shechter, H. Tetrahedron Lett. 1982, 23, 2277-
2280.
(27) Aggarwal, V. K.; Kalomiri, M.; Thomas, A. P. Tetrahedron:
Asymmetry 1994, 5, 723-730.
(32) Doyle, M. P.; Griffin, J. H.; Bagheri, V.; Dorow, R. L. Organo-
metallics 1984, 3, 53-61.
(28) De Lucchi, O.; Lucchini, V.; Marchioro, C.; Valle, G.; Modena, G.
J. Org. Chem. 1986, 51, 1457.
(33) Cu(acac)2 was prepared by adding saturated sodium carbonate to a
solution of copper oxide and acetylacetone. See: Bryant, B. E.; Fernelius,
W. C. Inorg. Synth. 1957, 5, 115.
(29) Eschler, B. M.; Haynes, R. K.; Ironside, M. D.; Kremmydas, S.;
Ridley, D. D.; Hambley, T. W. J. Org. Chem. 1991, 56, 4760.
(30) Eliel, E. L.; Frazee, W. J. J. Org. Chem. 1979, 44, 3598.
(34) Commercial copper acetylacetonate that had been purified by
sublimation was equally effective.

Catalytic Asymmetric Epoxidation of Aldehydes
J. Am. Chem. Soc., Vol. 120, No. 33, 1998 8331
Scheme 8
Scheme 10
Table 3. Reactions of Sulfide 5b with Benzaldehyde in the
Catalytic Cycle Using Different Metal Salts
Table 2. Reactions of Sulfides 5 in the Catalytic Cycle with
Benzaldehyde
entry
catalysta
yield/%b
ee/%c
entry
sulfide
R
R′
yield/%a
ee/%b
1
2
3
4
5
copper acetylacetonate
copper tetramethylheptanedionate
copper hexafluoropentanedionate
copper bronze
73
64
0
35
61
93
92
1
2
3
4
5
6
7
8
9
10
11
12
13
14
5a
5b
5c
5d
5e
5f
5m
5g
5h
5i
H
H
83
71
57
0
0
56
0
43
70
71
68
9
11
18
41
93
93
Me
Pri
But
Ph
CH2Ph
CCl3
CH2OPh
CH2OMe
CH2OBn
CH2OAc
TMS
H
H
H
H
H
H
H
H
H
H
H
Me
91
92
rhodium acetatec
a 5 mol% copper salts or 1 mol% Rh2(OAc)4. b Only trans-stilbene
oxide was obtained (>98:2 trans/cis). c Ee values were measured by
HPLC using a Chiralcel OD column. The (R,R) enantiomer was the
major product in each case.
88
83
92
90
93
62
70
89
Scheme 11
5n
5q
5j
Me
5k
spiro-cyclobutyl
a Only trans-stilbene oxide was obtained (>98:2 trans/cis). b Ee
values were measured by HPLC using a Chiralcel OD column. The
(R,R) enantiomer was the major product in each case.
Scheme 9
Table 4. Reactions of Sulfide 5b with Benzaldehyde in the
Catalytic Cycle in Different Solvents
entry
solvent
yield/%a
ee/%b
1
2
3
4
5
6
CH2Cl2
ButOMe
EtOAc
toluene
MeCN
THF
64
35
39
42
31
43
92
94
94
93
92
89
a Stevens rearrangement.35 The Stevens rearrangement of the
ylide competes with carbonyl epoxidation and in this case is
facilitated by the stability of the intermediate phenyl-substituted
radical (Scheme 9). A single heteroatom in the side chain can
be tolerated (entries 8-11), and the similar enantioselectivities
obtained compared to R ) Me showed that there was no
substantial electronic effect. The trichloro derivative 5m (entry
7) gave no epoxide, only stilbenes. This sulfide is either
sterically too hindered or electronically too deactivated by the
chlorines to react with the metal carbenoid, and so diazo
dimerization is observed instead (k2 > k1). We expected the
TMS derivative (sulfide 5q, entry 12) to be less sterically
hindered than sulfide 5d and to be electronically activated to
react with the metal carbenoid, and indeed it was, although only
marginally so; epoxide was obtained but in low yield and with
reduced enantioselectivity. The presence of two alkyl groups
adjacent to sulfur gave much reduced yields of epoxide, again
because of increased steric hindrance, but the enantioselectivity
was also low (entries 13, 14).
We next decided to investigate whether the choice of metal
salt would have any effect on the yield or diastereo- or
enantioselectivity of the epoxidation process. Reactions were
carried out in dichloromethane using 0.2 equiv of our optimum
sulfide 5b with benzaldehyde and a range of metal catalysts
(Scheme 10, Table 3). All the metal salts used gave essentially
the same enantioselectivity. These results showed that the
choice of metal salt used for the diazo decomposition had little
effect on the enantioselectivity of the process. It also showed
a Only trans-stilbene oxide was obtained (>98:2 trans/cis). b Ee
values were measured by HPLC using a Chiralcel OD column. The
(R,R) enantiomer was the major product in each case.
that the metal did not participate in any way in the reaction of
the sulfur ylide with the aldehyde. The optimum metal salt
was copper acetonylacetonate.
The effect of solvent on the epoxidation process was also
investigated. Reactions were carried out using copper tetra-
methylheptanedionate as the metal catalyst because we had
found that this salt was soluble in all of the solvents we wished
to investigate, unlike copper acetylacetonate (Scheme 11, Table
4), which was not completely soluble in toluene, tBuOMe, THF,
or EtOAc. The table shows that all the solvents gave essentially
the same enantioselectivity and that the highest yields were
obtained using CH2Cl2.
Having optimized both the epoxidation conditions and sulfide
catalyst choice (5b), we then performed epoxidation reactions
with a range of aldehydes (Scheme 12, Table 5). It was found
that good yields and high diastereo- and enantioselectivities were
obtained for other aromatic and unsaturated aldehydes. How-
ever, lower yields and diastereoselectivities were observed with
aliphatic aldehydes, but the level of enantioselectivity was
essentially maintained. Reaction of valeraldehyde gave con-
siderably lower enantioselectivity (see below).
Origin of Diastereoselectivity. To account for the origin
of the enantioselectivity/diastereoselectivity, we needed to know
whether the sulfur ylide reactions were under kinetic or
(35) It was not possible to isolate a pure sample of 7 to categorically
prove its structure. The data are consistent with this product though.

8332 J. Am. Chem. Soc., Vol. 120, No. 33, 1998
Aggarwal et al.
Scheme 12
Scheme 14
Table 5. Yields, Enantioselectivities and Ratios of Epoxides
Formed from Aldehydes Using 0.2 equiv of Sulfide 5b
entry
aldehyde
benzaldehyde
p-chlorobenzaldehyde
p-tolualdehyde
cinnamaldehyde
valeraldehyde
yield/%
ee/%a
trans/cis
1
2
3
4
5
6
73
72
64
55c
35
32
94 (R,R)b
92(R,R)b
92(R,R)b
89d
>98:2
>98:2
>98:2
>98:2
92:8
68d
cyclohexanecarbaldehyde
90d
70:30
a Enantiomeric excesses were determined by chiral HPLC using a
Chiralcel OD column. b Absolute configurations were determined by
comparison of [R]D values with literature values.36 c We originally
reported a 73% yield, but this was incorrect. d By analogy we have
assumed that the same absolute configuration (R,R) was obtained.
sufficient to promote reversibility in syn betaine formation with
aliphatic aldehydes however.
Origin of Enantioselectivity. Having established that trans
epoxides are derived from irreversible formation of anti betaines,
we needed information on the transition state leading to their
formation in order to understand the origin of the enantio-
selectivity. As this was not possible, we focused on gaining
information on the structure of the ylide. We believe a single
sulfonium ylide is formed, as we have previously shown that
alkylation of the related oxathiane 5a only gave the equatorial
sulfonium salt.38 Ylide conformation has been studied by X-ray
diffraction,39,40 NMR,41-46 and computation.47,48 All of these
studies indicate that the preferred conformation of sulfur ylides
is one in which the filled orbital on the ylide carbon is
orthogonal to the lone pair on sulfur. The barrier to rotation
around the C-S bond of the semistabilized ylide, dimethyl
Scheme 13
sulfonium fluorenide, has been found to be 42 ( 1.0 kJ mol-1 46
.
This information implies that the ylide derived from 5b will
adopt conformations 8A and 8B and that these will be in rapid
equilibrium at room temperature. Of these two, conformation
8B will be favored, as 8A suffers from 1,3-diaxial interactions
between the phenyl ring and the axial groups. The aldehyde
can then approach either face of the ylide, but the re face is
more accessible as the si face is hindered by the equatorial
methyl group (Scheme 14).
The aldehyde can react in an end-on or [2 + 2] mode, but
there is no evidence, experimental or theoretical, to indicate
which is preferred. While most examples invoke the end-on
mode, there is circumstantial evidence that the [2 + 2] mode
may be favored. This circumstantial evidence comes from a
study of the asymmetric induction observed in methylene
transfer from the C2 symmetric sulfur ylide 9 to benzaldehyde.
thermodynamic control. From crossover experiments we had
found that the addition of benzylsulfonium ylide to aldehydes
was remarkably finely balanced.37 The trans epoxide was
derived directly from irreVersible formation of the anti betaine
or indirectly from reVersible formation of the syn betaine. The
cis epoxide was derived from partial reversible formation of
the syn betaine. The higher trans selectivity observed in
reactions with aromatic aldehydes compared to that with
aliphatic aldehydes was due to greater reversibility in the
formation of the syn betaine (Scheme 13).
Reactions of simple sulfides (Me2S, tetrahydrothiophene) in
the catalytic cycle with benzaldehyde give stilbene oxide as an
84:16 ratio of trans/cis epoxides.25 However, epoxidation using
our camphor-derived 1,3-oxathiane 5b gave only trans epoxides
with a range of aromatic and unsaturated aldehydes. This higher
selectivity must be due to an increase in k-3 relative to k4. An
increase in k-3 relative to k4 would be expected for sulfides of
increasing steric hindrance or where the ylide had increased
stability. In the case of the 1,3-oxathiane, we believe that the
corresponding ylide shows increased stability relative to simple
benzyl sulfonium ylides as a result of the anomeric effect (vide
infra). The positive charge on sulfur can be delocalized over
the oxygen, and this will lead to increased stability and therefore
greater reversibility and thus greater trans selectivity. The
moderate increase in stability of the ylide is evidently not
(38) Aggarwal, V. K.; Thompson, A.; Jones, R. V. H.; Standen, M.
Tetrahedron: Asymmetry 1995, 6, 2557-2564.
(39) Christensen, A. T.; Witmore, W. G. Acta Crystallogr. B 1969, 25,
73-78.
(40) Christensen, A. T.; Thom, E. Acta Crystallogr. B 1971, 27, 581-
586.
(41) Ratts, K. W. Tetrahedron Lett. 1966, 4707-4712.
(42) Cook, A. F.; Moffatt, J. G. J. Am. Chem. Soc. 1968, 90, 740-747.
(43) Matsuyama, H.; Minato, H.; Kobayashi, M. Bull. Chem. Soc. Jpn.
1977, 50, 3393-3396.
(44) Galloy, J.; Watson, W. H.; Craig, D.; Guidry, C.; Morgan, M.;
McKellar, R.; Ternay, A. L.; Martin, G. J. Heterocycl. Chem. 1983, 20,
399-405.
(45) Johnson, A. W.; Amel, R. T. J. Org. Chem. 1969, 34, 1240-1247.
(46) Aggarwal, V. K.; Schade, S.; Taylor, B. J. Chem. Soc., Perkin Trans.
1 1997, 2811-2813.
(47) Bernardi, F.; Schlegel, H. B.; Whangbo, M.-Y.; Wolfe, S. J. Am.
Chem. Soc. 1977, 99, 5633-5636.
(36) Imuta, M.; Ziffer, H. J. Org. Chem. 1979, 44, 2505.
(37) Aggarwal, V. K.; Calamai, S.; Ford, J. G. J. Chem. Soc., Perkin
Trans. 1 1997, 593-599.
(48) Eades, R. A.; Gassman, P. G.; Dixon, D. A. J. Am. Chem. Soc.
1981, 103, 1066-1068.

Catalytic Asymmetric Epoxidation of Aldehydes
J. Am. Chem. Soc., Vol. 120, No. 33, 1998 8333
Scheme 15
Figure 1.
If this was the case, enantioselectivity should be highly
dependent upon the size of the equatorial substituent. However,
increasing the size of this substituent from Me to Pri did not
result in a concomitant increase in selectivity (Table 2, entries
2 and 3). Indeed, the enantioselectivity was essentially the same
for a range of substituents suggesting that the facial selectivity
was essentially complete.50 Even in the absence of a substituent,
good re-face selectivity was still observed (5a gave 41% ee:
Table 2, entry 1). This latter result in particular suggested that
the oxygen of the oxathiane was affecting the facial selectivity
of the ylide, and we believe it is exerting this effect through a
combination of the anomeric51 and Cieplak effects.52-54
A resonance form of the ylide is shown in Figure 1. If there
is a contribution from this resonance form to the ground-state
structure of the ylide, then the C4-S bond should be more
electron rich than the C2-S bond and this should affect the
face selectivity of the ylide. The resonance form shown in
Figure 1 is a result of the anomeric effect and should manifest
itself in shortening of the C4-O bond relative to the C6-O bond
and lengthening of the C4-S bond relative to the C2-S bond.
Indeed, this was observed in the X-ray structure of oxathiane
5p (Table 6, Figure 2).55 The anomeric effect should be even
greater in the ylide, but it was not possible to prepare and isolate
this reactive intermediate. We therefore prepared the corre-
sponding sulfoxide, an electronic analogue of the ylide, and were
able to grow crystals of 10 (Figure 2).55 The sulfoxide did
indeed show even greater differences in bond lengths, demon-
strating an even greater anomeric effect (Table 6).56 As a
control, an X-ray of the thiane sulfoxide 1755 (see later for
synthesis) was also obtained and showed essentially no differ-
ences in bond lengths (Table 6).56 This proved that the
differences in bond lengths observed with oxathiane 10 were
due to a significant anomeric effect. The X-ray data provide
strong evidence for a significant contribution of the resonance
form depicted in Figure 2 to the structure of the ylide. This
Scheme 16
(50) Lower selectivities were observed in cases containing an aromatic
ring in the side chain of the equatorial substituent. It is possible that aromatic
groups offer π-stacking opportunities to an incoming aromatic aldehyde
and result in some si-face attack.
(51) For a comprehensive review on the anomeric effect, including a
discussion on second row elements, see (a) Juaristi, E.; Cuevas, G.
Tetrahedron 1992, 48, 5019-5087 and also (b) Kirby, A. J. The Anomeric
Effect and Related Stereoelctronic Effects at Oxygen, 1st ed.; Springer-
Verlag: Berlin, 1983.
In this process styrene oxide was obtained with essentially no
enantioselectivity (Scheme 15).49 Analysis of the two possible
modes shows that an end-on approach would be expected to
furnish styrene oxide with high enantioselectivity, as only one
of the two approaches would be expected to be favored. In
contrast, both approaches in the [2 + 2] addition mode are
equally accessible and so this would give styrene oxide with
low enantioselectivity, as observed. Thus both end-on and [2
+ 2] modes should be considered.
End-on and [2 + 2] transition states in the reaction of ylide
8 with benzaldehyde, leading to trans epoxides, are shown in
Scheme 16. From analysis of molecular models of these
transition states it is clear that they can all be accommodated.
Thus, our own results do not provide any further evidence as
to which mode (end-on versus [2 + 2]) is favored. The
transition states shown in Scheme 16 account for the high
enantioselectivities observed.
(52) Cieplak, A. S. J. Am. Chem. Soc. 1981, 103, 4540-4552.
(53) Cieplak, A. S.; Tait, B. D.; Johnson, C. R. J. Am. Chem. Soc. 1989,
111, 8447-8462.
(54) Gung, B. W. Tetrahedron 1996, 52, 5263-5301.
(55) The author has deposited atomic coordinates for 5p, 10, and 17
with the Cambridge Crystallographic Data Centre. The coordinates can be
obtained on request from the director at Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB2 1EZ, UK.
(56) Second row elements show a substantial anomeric effect, often
greater than first row elements.51a Indeed, nS are better donors than nO, and
51a
σ*C-S is a more effective acceptor orbital than σ*C-O
.
In oxathianes 5
two anomeric effects will be in operation (nS f σ*C-O and nO f σ*C-S
)
and differences in bond length will result if one of the two anomeric effects
is more powerful, and evidently one is (nO f σ*C-S). In the sulfoxide,
only one anomeric effect is in operation (nO f σ*C-S). As this overlap of
orbitals is likely to be even more effective than in the 1,3-oxathiane and
because the anomeric effect in the reverse direction is essentially zero, very
significant differences in bond lengths result. Indeed, the observed differ-
ences in bond lengths of oxathiane 10 are some of the largest differences
recorded.
We were curious as to the origin of the minor enantiomer
and considered the possibility that it arose from si-face attack.
(49) This work was carried out by J. K. Whitesell and discussed in the
following review: Durst, T.; Breau, L.; Ben, R. N. Phosphorous Sulfur
Relat. Elem. 1993, 74, 215-232.

8334 J. Am. Chem. Soc., Vol. 120, No. 33, 1998
Aggarwal et al.
Figure 2.
Table 6. Bond Lengths from X-ray Structures of 5p, 10, and 17
com- C2-S/ S-C4/ C4-O/ O-C6/ C4-C5/ C6-C5/
Scheme 17
entry pound
Å
Å
Å
Å
Å
Å
1
2
3
5p
10
17
1.798 1.814 1.406 1.439
1.815 1.854 1.402 1.447
1.809 1.801
1.537
1.524
resonance form will result in the C4-S bond being more electron
rich than the C2-S bond. According to Cieplak, nucleophilic
reactions on π systems occur on the face opposite the better
donor.57 Thus, the ylide derived from 5a should preferentially
react on the face opposite to the C4-S bond, that is, the re
face, and this is observed.
Scheme 18
To gain further evidence of this combined anomeric and
Cieplak effect, sulfur and carbon analogues (13 and 16
respectively) of oxathiane 5a were studied. These substrates
were chosen to minimize steric effects in the epoxidation process
and to allow us to focus on the above electronic effects.
The sulfur analogue 13 was prepared as shown in Scheme
17. Dithiol 1258,59 was treated with paraformaldehyde and tosic
acid under Dean-Stark conditions to give the desired 1,3-
dithiane 13. The carbon analogue 16 was prepared as shown
in Scheme 18. Addition of vinylcerium60,61 to 11 gave the exo
(57) The only other example in which an anomeric or anomeric-type
effect controlled the outcome of a reaction in a manner similar to that
described above is provided by Le Noble (Hahn, J. M.; le Noble, W. J. J.
Am. Chem. Soc. 1992, 114, 1916-1917). He studied the addition of
nucleophiles to 5-azaadamantone and found a slight preference for addition
of MeLi anti to nitrogen. BH4 reduction in MeOH gave addition syn to
nitrogen possibly because of hydrogen bonding or metal chelation with the
nitrogen lone pair. These experiments do not provide clear-cut evidence
for a combined anomeric and cieplak effect.
alcohol, which upon treatment with 9-BBN gave the corre-
sponding 4-hydroxythiane.62 The six-membered ring is ob-
tained, as the addition of thiyl radicals to olefins is reversible
and the thiane results from formation of the more stable
secondary radical.63-67 Attempts to prepare the half oxalate
chloride of 14 for subsequent radical deoxygenation68,69 were
unsuccessful as treatment with oxalyl chloride resulted in
(61) Imamoto, T.; Takiyama, N.; Nakamura, K.; Hatajima, T.; Kamiya,
Y. J. Am. Chem. Soc. 1989, 111, 4392-4398.
(62) Masuda, Y.; Hoshi, M.; Nunokawa, Y.; Arase, A. J. Chem. Soc.,
Chem. Commun. 1991, 1444-1445.
(63) Henrick, C. A.; Willy, E.; Baum, J. W.; Baer, T. A.; Garcia, B. A.;
Mastre, T. A.; Chang, S. M. J. Org. Chem. 1975, 40, 1-7.
(64) Schulte-Elte, K. H.; Ohloff, G. HelV. Chim. Acta 1968, 51, 548-
553.
(58) Montenegro, E.; Echarri, R.; Claver, C.; Castillon, S.; Moyano, A.;
Pericas, M. A.; Riera, A. Tetrahedron: Asymmetry 1996, 7, 3553-3558.
(59) Oae, S.; Togo, H. Bull. Chem. Soc. Jpn. 1983, 56, 3802-3812.
(60) Imamoto, T.; Takiyama, N.; Nakamura, K. Tetrahedron Lett. 1985,
26, 4763-4766.

Catalytic Asymmetric Epoxidation of Aldehydes
J. Am. Chem. Soc., Vol. 120, No. 33, 1998 8335
Table 7. Yields and Enantioselectivities of Epoxides Formed
from Benzaldehyde Using 0.2 equiv of Sulfide
Scheme 19
Scheme 20
entry
sulfidea
yield/%
ee/%d
1
2
3
16
5b
13
C
O
S
96b
83c
19c
20
41
44
a The element in the 5 position of the thiane is indicated (thiane 16,
oxathiane 5b, and dithiane 13). b 96:4 trans/cis ratio was observed.c Only
trans-stilbene oxide was observed (>98:2 trans/cis). d Ee values were
measured by HPLC using a Chiralcel OD column. The (R,R) enantiomer
was the major product in each case.
Scheme 21
Figure 3.
elimination to give 15 instead. The unsaturated thiane was
reduced to the thiane 16 by hydrogenation,70 but a mixture of
inseparable products were obtained. Attempts to purify this
mixture by formation of the HgCl2 salt71 and subsequent
recrystallization were unsuccessful. Oxidation to the sulfoxide
17 led to a substrate that we were able to recrystallize, and
subsequent reduction gave pure 16.
In the epoxidation process, the sulfur analogue 13 should
show enhanced selectivity (greater anomeric effect), and the
carbon analogue 16 should show reduced selectivity (no
anomeric effect). In the event, the sulfur analogue 13 gave only
slightly higher selectivity (and only moderate yield72), while
the carbon analogue 16 gave much reduced selectivity (Table
7).
The stereochemical outcome of the ylide reaction is deter-
mined by two factors: the face selectivity of the ylide and the
conformation of the ylide (Figure 3). The sulfur analogue 13
of the oxathiane was expected to give an increase in selectivity
due to an increase in the anomeric effect,73 which should result
in an increase in face selectivity. However, replacement of
oxygen by sulfur results in a more distorted chair and therefore
reduced 1,3-diaxial interactions, which will result in an increase
in the population of conformer 19B (Figure 3). Thus, the ylide
conformations may react with increased face selectivity, but a
greater proportion of conformer 19B will result in a reduction
in enantioselectivity. The observed enantioselectivity results
from a balance between these two effects. The carbon analogue
16 gave reduced selectivity. As the thiane adopts a very similar
conformation to the 1,3-oxathiane (C-O and C-C have similar
bond lengths), the ratio of conformers 18A/18B and 20A/20B
will be similar, and the difference in enantioselectivity must
result from a difference in face selectivity of the ylide. The
oxygen of the 1,3-oxathiane therefore exerts a significant
stereoelectronic effect in promoting reaction on the face opposite
to this group.
We therefore believe that there is a very high preference for
re-face attack as a result of combined steric and electronic
effects, which both act in concert and reinforce each other. The
lower enantioselectivity observed with valeraldehyde compared
to the other aldehydes could be due to its smaller size, which
in turn may allow some si-face attack on the ylide. To test
this, the size of the equatorial substituent on the sulfide was
increased. Using sulfide 5c higher enantioselectivity was indeed
obtained (Scheme 19). This confirmed that facial selectivity
in the reaction of sulfide 5b is dependent on the size of the
aldehyde; R-branched, aromatic, and unsaturated aldehydes react
with essentially complete facial selectivity, while unbranched
aldehydes react with moderate facial selectivity.
As facial selectivity appears to be complete, the source of
the minor enantiomer could be reaction of the minor conforma-
tion of the ylide 8A (Scheme 14). To reduce the amount of
this minor conformation, we prepared thioacetals bearing axial
substituents 5j (R ) Me) and 5l (R ) R1 ) spiro-cyclobutyl)
to increase the 1,3-diaxial interactions. However, instead of
increased selectivity, a decrease in enantioselectivity was
observed. A study of the conformation of these thioacetals
revealed the origin of this reduction in selectivity. NOE
experiments carried out on 5b and 5j revealed that 5j existed
in both chair and boat forms, while 5b adopted the chair form
only. Reaction of the corresponding ylide from the boat
conformer of 5j should give the opposite enantiomer to that
from the chair form (Scheme 20). The boat conformation may
be favored because of severe 1,3-diaxial interactions in the chair
form. The related thioacetal 21 has also been shown to exist
in both chair (21A) and boat (21B) forms (Scheme 21).74
An alternative way to increase the 1,3-diaxial interactions
and favor ylide conformer 8B over 8A would be to increase
the bulk of the aromatic ring. Thus mesityldiazomethane 22
(65) Thalmann, A.; Oertle, K.; Gerlach, H. Org. Synth. 1984, 63, 192-
197.
(66) Walling, C.; Helmreich, W. J. Chem. Soc. B 1959, 81, 1144-1148.
(67) Claus, P. K.; Vierhapper, F. W.; Willer, R. L. J. Org. Chem. 1977,
42, 4016-4023.
(68) Barton, D. H. R.; Crich, D. J. Chem. Soc., Perkin Trans. 1 1986,
1603-1611.
(69) Barton, D. H. R.; Crich, D. J. Chem. Soc., Chem. Commun. 1984,
774-775.
(70) Attempts to carry out ionic reduction of 15 failed due to competing
rearrangements of the camphor skeleton. Hydrogenation using diimide and
hydroboration of 15 were also unsuccessful.
(71) Whitehead, E. V.; Dean, R. A.; Fidler, F. A. J. Am. Chem. Soc.
1951, 73, 3632-3635.
(72) The low yield obtained with 13 is typical of 1,3-dithianes.
1,3-Dithiane itself gave 30% yield of stilbene oxide in the catalytic cycle.
Also, ylide formation is quite sensitive to the steric hindrance of the sulfide,
and only the more accessible sulfur should react.
(73) Juaristi, E.; Cuevas, G. Tetrahedron 1992, 48, 5019-5087.
(74) De Lucchi, O.; Lucchini, V.; Marchioro, C.; Modena, G. Tetrahedron
Lett. 1985, 26, 4539-4542.

8336 J. Am. Chem. Soc., Vol. 120, No. 33, 1998
Aggarwal et al.
Scheme 22
ity observed with valeraldehyde is a result of reduced facial
selectivity of the ylide due to reduced steric effects.
Experimental Section
Proton chemical shifts were recorded on the δ scale and were
measured relative to the residual signal of chloroform at δ 7.25. All
coupling constants are measured in hertz. The deuterated solvents used
were as specified in the Experimental Section. 13C chemical shifts were
measured from the central peak of chloroform at δ 77.0. [R]20D values
are given in 10-1 deg cm2 g-1
. Enantiomeric excesses (ee) were
determined by chiral HPLC using a Chiralcel OD column.
Typical Procedure for the Preparation of 1,3-Oxathianes. (a)
Preparation of (1R,4R,6R,8R)-4,11,11-Trimethyl-5-oxa-3-thia-tricyclo-
[6.2.1.01,6]undecane (5b). To a cooled solution (ice bath) of (+)-(10)-
mercaptoisoborneol30 (270 mg, 1.45 mmol) and acetaldehyde (0.24 mL,
4.29 mmol) in dichloromethane (2 mL) under argon was added boron
trifluoride etherate (0.20 mL, 1.63 mmol). After a few minutes, the
reaction mixture was loaded directly onto a silica gel column and eluted
with 50% dichloromethane in petroleum ether (40-65 °C) to give the
desired sulfide 5b as a clear oil (305 mg, 99%) (Found: C, 67.71; H,
was prepared and used in the catalytic cycle with thioacetal 5b
and benzaldehyde (Scheme 22). Greatly increased enantio-
selectivity was observed (but in only low yield). It therefore
seems likely that the minor enantiomer formed in the reactions
involving phenyldiazomethane arises from reaction of the minor
conformer 8A. Attempts to reduce the proportion of the minor
conformer 8A by lowering the temperature were unsuccessful.
Reaction of trans-cinnamaldehyde75 with thioacetal 5b, phen-
yldiazomethane, and Rh2(OAc)476 at 0 °C furnished the epoxide
with similar enantioselectivity. Evidently, the small difference
in temperature was not sufficient to make a significant difference
in enantioselectivity. At even lower temperatures the diazo
compound did not decompose.
9.26; S, 15.30. C12H20OS requires C, 67.87; H, 9.49, S, 15.10): [R]20
D
-140.4 (c 1.14 in CHCl3); υmax (thin film)/cm-1 2940, 2872, 1097,
1072; δH (250 MHz, CDCl3) 0.77-1.91 (16H, m), 2.68 (1H, d, J )
14.5), 3.08 (1H, d, J ) 14.5), 3.58 (1H, dd, J ) 3.5, 8.0), 4.73 (1H, q,
J ) 6.5); δC (63 MHz, CDCl3) 20.48, 21.75, 23.28, 27.29, 28.67, 34.51,
37.85, 41.73, 45.57, 46.60, 77.62, 85.25; m/e (EI) 212 (M+, 86%), 135
(95), 93 (100) (Found: M+, 212.1238. C12H20OS requires 212.1235).
(b) Preparation of (1R,4R,6R,8R)-4-(Acetoxymethyl)-11,11-di-
methyl-5-oxa-3-thia-tricyclo[6.2.1.01,6]undecane (5n). To a solution
of sulfide 5l (118 mg, 0.52 mmol) and acetic anhydride (60 µL, 0.64
mmol) in ether (1 mL) under nitrogen was added pyridine (46 µL, 0.57
mmol). The stirred solution was left overnight before loading directly
onto a silica gel column and eluting with 10% ethyl acetate in petroleum
ether (40-65 °C) to give the desired sulfide 5n as a clear oil (129 mg,
Conclusion
Oxathiane 5b has been found to be a highly efficient sulfide
for carbonyl epoxidation. It is easily prepared in two steps from
camphorsulfonyl chloride (which is available in both enantio-
meric forms) and can be used in substoichiometric amounts in
our catalytic cycle. In reactions with aromatic and unsaturated
aldehydes only trans epoxides are obtained, and this high
selectivity is due to irreversible formation of the anti betaine
and greater reversibility in the formation of the syn betaine.
This is believed to result from the enhanced stabilization of the
ylide due to an anomeric effect from the ring oxygen. Indeed,
replacement of the ring oxygen by carbon results in reduced
diastereoselectivity (96:4 compared to >98% de).
Two factors control the enantioselectivity of the epoxidation
process: ylide conformation and face selectivity of the ylide.
The sulfur ylide derived from oxathiane 5b can exist in two
conformations 8A and 8B. Conformer 8A suffers from 1,3-
diaxial interactions, and so conformation 8B is favored. The
ylide can potentially react on either face, but it is believed that
there is very high selectivity for re-face attack with aromatic,
unsaturated, and R-branched aldehydes. The high face selectiv-
ity is a result of steric effects (the equatorial methyl group) and
electronic effects (a combination of anomeric and Cieplak
effects), which both favor re-face attack, and they therefore
reinforce each other. The minor enantiomer is believed to come
from the reaction of the less favored ylide conformation 8A.
When more hindered diazo compounds (mesityldiazomethane)
are used, the proportion of the minor conformation 8A is reduced
further as a result of increased 1,3-diaxial interactions and the
enantiomeric excess is essentially complete. The lower selectiv-
93%): [R]20 -120.6 (c 6.47 in CHCl3); υmax (thin film)/cm-1 2953,
D
2879, 1746, 1220, 1071; δH (250 MHz, CDCl3) 0.79-1.09 (5H, m),
1.28 (3H, s), 1.38-1.54 (3H, m), 1.61-1.75 (3H, m), 1.84-1.95 (1H,
m), 2.07 (3H, s), 2.76 (1H, d, J ) 14.0), 3.07 (1H, d, J ) 14.0), 3.60
(1H, dd, J ) 3.0, 8.0), 4.17-4.23 (2H, m), 4.91 (1H, dd, J ) 4.5, 6.0);
δC (63 MHz, CDCl3) 20.41, 20.88, 23.09, 27.23, 28.25, 34.23, 37.72,
42.47, 45.49, 46.70, 65.82, 79.35, 85.17, 170.66; m/e (EI) 210 (M+
-
AcOH, 100%), 168 (47), 135 (76), 121 (37) (Found: M+ - AcOH,
210.1078. C12H18OS requires 210.1078).
(c) Preparation of (1R,4R,6R,8R)-4-(p-Nitrobenzoylmethyl)-11,11-
dimethyl-5-oxa-3-thia-tricyclo[6.2.1.01,6]undecane (5p). To a slurry
of p-nitrobenzoyl chloride (100 mg, 0.54 mmol) in pyridine (1 mL)
was added a solution of sulfide 5l (94 mg, 0.41 mmol) in pyridine (0.3
mL) at 0 °C under nitrogen. After 1.5 h the reaction mixture was
allowed to warm to room temperature and was left for 5 h. The reaction
mixture was poured into water, the resulting mixture was extracted
with dichloromethane (3 times), and the combined organic extracts were
dried (MgSO4). After filtration and removal of the solvent under
reduced pressure, the residue was purified by flash chromatography,
eluting with 10% dichloromethane in petroleum ether (40-65 °C) to
give the desired sulfide 5p as a yellow crystalline solid (114 mg, 73%),
mp 67-69 °C (Found: C, 60.22; H, 6.15; N, 3.51; S, 8.24. C19H23-
NO3S requires C, 60.46; H, 6.14; N, 3.71; S, 8.49): [R]20 -83.9 (c
D
5.96 in CHCl3); υmax (KBr disk)/cm-1 2962, 1724, 1522, 1274; δH (250
MHz, CDCl3) 0.77-1.19 (5H, m), 1.32 (3H, s), 1.41-1.56 (1H, m),
1.63-1.79 (3H, m), 1.83-1.96 (1H, m), 2.80 (1H, d, J ) 14.0), 3.12
(1H, d, J ) 14.0), 3.65 (1H, dd, J ) 3.0, 8.0), 4.51 (2H, d J ) 5.0)
5.07 (1H, d, J ) 5.0), 8.15-8.33 (4H, m); δC (63 MHz, CDCl3) 20.4,
23.2, 27.2, 28.3, 34.2 37.7, 42.5, 45.5, 46.8, 66.9, 79.2, 85.4, 123.6,
131.0, 135.3, 150.6, 164.3; m/e (CI) 378 (M+, 8%), 395 (M+ + NH3,
100), 211 (40), 169 (12) (Found: M+, 378.1359. C19H23NO3S requires
378.1375).
(75) This aldehyde was chosen as it gave a slightly lower enantioselec-
tivity compared to other aldehydes, thereby allowing us to monitor changes
more easily.
(76) Use of Cu(acac)2 was unsuccessful. No epoxides, only stilbenes,
were formed.
(d) Preparation of (1R,4R,6R,8R)-4-(Trimethylsilyl)-11,11-di-
methyl-5-oxa-3-thia-tricyclo[6.2.1.01,6]undecane (5q). To a solution
of sulfide 5a (285 mg, 1.44 mmol) in tetrahydrofuran (3 mL) at -78
°C under nitrogen was added sec-BuLi (1.1 mL of a 1.3 M solution in

Catalytic Asymmetric Epoxidation of Aldehydes
J. Am. Chem. Soc., Vol. 120, No. 33, 1998 8337
cyclohexane). After a few minutes, chlorotrimethylsilane (0.3 mL, 2.36
mmol) was added to the yellow solution. After 2.5 h the reaction was
quenched at -78 °C with water before allowing the mixture to warm
to room temperetaure and pouring it onto brine. The resulting mixture
was extracted with dichloromethane (3 times), and the combined organic
extracts were dried (MgSO4). After filtration and removal of the solvent
under reduced pressure, the residue was purified by flash chromatog-
raphy, eluting with 50% dichloromethane in petroleum ether (40-65
°C) to give the desired sulfide 5q as a white crystalline solid (314 mg,
81%), mp 36-39 °C (Found: C, 62.02; H, 9.49; S, 11.62. C14H26-
OSSi requires C, 62.16; H, 9.69; S, 11.85): [R]20D -140.1 (c 11.49 in
CHCl3); υmax (KBr disk)/cm-1 2960, 2790, 1244, 1069; δH (250 MHz,
CDCl3) 0.07 (9H, s), 0.78-1.06 (5H, m), 1.22-1.47 (4H, m), 1.54-
1.87 (4H, m), 2.69 (1H, d, J ) 14.0), 3.00 (1H, d, J ) 14.0), 3.38 (1H,
dd, J ) 3.0, 8.0), 4.44 (1H, s); δC (63 MHz, CDCl3) -3.64, 20.51,
23.37, 27.49, 29.02, 35.19, 38.02, 42.64, 45.60, 46.58, 74.86, 85.42;
m/e (EI) 270 (M+, 11%), 197 (7), 165 (69), 121 (100) (Found: M+,
270.1469. C14H26OSSi requires 270.1474).
Preparation of (1S,4R,6R,8R)-(11,11-Dimethyl-5-oxa-3-thia-
tricyclo[6.2.1.01,6]undec-4-yl)-methanol-3-oxide (10). A solution of
3-chloroperoxybenzoic acid (125 mg, 0.73 mmol) in dichloromethane
(0.5 mL) was added to a solution of sulfide 5l (141 mg, 0.62 mmol) in
dichloromethane (1 mL) under nitrogen. After 15 h, the solvent was
removed under reduced pressure and the residue was loaded onto a
silica gel column eluting with ethyl acetate. The second fraction, eluted
with acetone, gave sulfoxide 10 as a white crystalline solid (136 mg,
96%), mp 145-146 °C: [R]20 -151.5 (c 1.65 in CHCl3); υmax (KBr
D
disk)/cm-1 3377 (OH), 3275 (OH), 2955, 1022, 998; δH (250 MHz,
(D3C)2SO) 0.91 (3H, s), 1.01 (3H, s), 0.97-1.12 (1H, m), 1.12-1.25
(1H, m), 1.44-1.83 (5H, m), 2.72 (1H, d, J ) 12.5), 3.50 (1H, d, J )
12.5), 3.66-3.81 (2H, m), 3.81-3.90 (1H, m), 4.14 (1H, dd, J ) 5.0,
2.0), 5.18 (1H, dd, J ) 6.0, 5.5); δC (63 MHz, (D3C)2SO) 20.49, 22.85,
27.19, 33.00, 36.98, 45.47, 46.81, 51.62, 52.49, 59.00, 85.02, 96.61.
Preparation of (1S,4R)-10-Mercaptomethyl-7,7-dimethyl-bicyclo-
[2.2.1]heptan-2-one (11). Camphorsulfonyl chloride (12.0 g, 48 mmol)
and triphenylphosphine (50.1 g, 191 mmol) were refluxed in a mixture
of water (40 mL) and 1,4-dioxane (160 mL) for 1 h under nitrogen.
After the reaction mixture had cooled it was extracted with petroleum
ether (200 mL and 3 × 100 mL). The combined organic extracts were
washed with water (2 × 100 mL) and brine (100 mL) before being
dried over MgSO4. After filtration and removal of the solvents under
reduced pressure were complete, the resulting oil was loaded directly
onto a silica gel column and eluted with 5% ethyl acetate in petroleum
ether to give thiol 11 as a white crystalline solid (7.3 g, 82%), mp
62-65 °C [lit. 59, 65-66 °C]: δH (250 MHz, CDCl3) 0.89 (3 H, s),
1.00 (3 H, s), 1.21-2.01 (6H, m), 2.07 (1H, t, J ) 4.6), 2.26-2.43
(2H, m), 2.85 (1H, dd, J ) 13.7, 6.7) [lit. 59, δH (CCl4) 0.93 (3 H, s),
1.05 (3 H, s), 1.2-2.6 (8H, m), 2.73 (1H, d, J ) 6), 2.95 (1H, d, J )
6)]; δC (63 MHz, CDCl3) 19.83, 20.31, 21.40, 26.63, 27.06, 29.92,
43.29, 43.67, 47.85, 60.65.
Preparation of (1R,2R,4R)-1-Mercaptomethyl-7,7-dimethyl-
bicyclo[2.2.1]heptane-2-thiol (12). A solution of (1R,4R)-dithio-
benzoic acid 7,7-dimethyl-2-thioxo-bicyclo[2.2.1]hept-1-ylmethyl ester58
(0.52 g, 1.6 mmol) in ether (9 mL) was added dropwise to a suspension
of lithium aluminum hydride (62 mg, 1.6 mmol) in ether (9 mL) at
room temperature under nitrogen. After a few seconds, the red color
of the starting material had been quenched. After 30 min, the reaction
was quenched with ethyl acetate (0.5 mL) before hydrochloric acid (1
mL of a 3% solution) was added and the suspension was filtered. The
solids were washed with ether, and the combined organics were dried
over MgSO4. After filtration of the mixture and removal of the solvents
under reduced pressure, the residue was loaded onto a silca gel column
and eluted with 5% ethyl acetate in petroleum ether to give a 10:1
mixture of exo/endo dithiols (0.24 g, 72% of 12) together with
benzylmercaptan (0.14 g, 67%): δH (250 MHz, CDCl3) 0.85 (3H, s),
1.02 (3H, s), 1.12-2.06 (10H, m), 2.55 (1H, dd, J ) 13.0, 7.0), 2.98
(1H, dd, J ) 13.7, 9.0).
General Procedure for the Epoxidation of Benzaldehyde Using
Sulfides 5, 13, and 16. By use of a syringe pump, phenyldiazomethane
(0.75 mmol in 0.25 mL of dichloromethane) was added to a stirred
solution of the desired sulfide (0.1 mmol), Cu(acac)2 (0.025 mmol),
and benzaldehyde (0.5 mmol) in dichloromethane (0.25 mL) under
nitrogen at room temperature over a period of 3 h. After being stirred
for an additional 1 h, the solvent was removed under reduced pressure
and the residue was chromatographed on silica gel to give trans-stilbene
oxide: δH (250 MHz, CDCl3) 3.85 (2H, s), 7.35 (10 H, m) (lit. 25).
Epoxidation of Benzaldehyde Using Sulfide 5b in Dichloro-
methane with Different Metal Catalysts. By use of a syringe pump,
phenyldiazomethane (0.75 mmol in 0.25 mL of dichloromethane) was
added to a stirred solution of sulfide 5b (0.1 mmol), the desired copper
salt (0.025 mmol) or rhodium acetate (0.01 mmol), and benzaldehyde
(50 µL, 0.5 mmol) in dichloromethane (0.25 mL) under nitrogen at
room temperature over a period of 3 h. After being stirred for an
additional 1 h, the solvent was removed under reduced pressure and
the residue was chromatographed on silica gel to give trans-stilbene
oxide: δH (250 MHz, CDCl3) 3.85 (2H, s), 7.35 (10 H, m) (lit. 25).
Epoxidation of Benzaldehyde Using Sulfide 5b in Different
Solvents. By use of a syringe pump, phenyldiazomethane (0.75 mmol
in 0.25 mL of the desired solvent) was added to a stirred solution of
sulfide 5b (0.1 mmol), Cu(Me3COCHCOMe3)2 (0.025 mmol), and
benzaldehyde (0.5 mmol) in the same solvent (0.25 mL) under nitrogen
at room temperature over a period of 3 h. After being stirred for an
additional 1 h, the solvent was removed under reduced pressure and
the residue was chromatographed on silica gel to give trans-stilbene
oxide: δH (250 MHz, CDCl3) 3.85 (2H, s), 7.35 (10 H, m) (lit. 25).
Typical Procedure for the Epoxidation of Aldehydes Using
Sulfide 5b. By use of a syringe pump, phenyldiazomethane (1.5 mmol
in 0.5 mL of dichloromethane) was added to a stirred solution of sulfide
5b (0.2 mmol), Cu(acac)2 (0.05 mmol), and the aldehyde (1 mmol) in
dichloromethane (0.5 mL) under nitrogen at room temperature over a
period of 3 h. After being stirred for an additional 1 h, the solvent
was removed in vacuo and the residue was chromatographed on silica
gel to give the desired epoxide. trans-Stilbene oxide: δH (250 MHz,
CDCl3) 3.85 (2H, s), 7.35 (10 H, m) (lit. 25). trans-2-(4-Chlorobezene-
yl)-3-phenyloxirane: δH (250 MHz, CDCl3) 3.82 (1H, d, J ) 2.0),
3.85t (1H, d, J ) 2.0), 7.04-7.50 (9H, m) (lit. 25). trans-2-(4-Tolyl)-
3-phenyloxirane: δH (250 MHz, CDCl3) 2.37 (3H, s), 3.83 (1H, d, J
) 1.5), 3.86 (1H, d, J ) 1.5), 7.16-7.44 (9H, m) (lit. 19 and 36).
trans-2-(trans-2-Phenylethylene)-3-Phenyloxirane: δH(250 MHz,
CDCl3) 3.52 (1H, dd, J ) 2.0,8.0), 3.88 (1H, d, J ) 2.0), 6.06 (1H,
dd, J ) 8.0,16.0), 6.72 (1H, d, J ) 16.0), 7.10-7.60 (10H, m) (lit.
77). 2-Cyclohexyl-3-phenyloxirane: δH (250 MHz, CDCl3) 0.76-
2.09 (10H, m), 2.76t (1H, dd, J ) 7.0, 2.0), 2.91c (1H, dd, J ) 9.0,
4.0), 3.68t (1H, d, J ) 2.0), 4.08c (1H, d, J ) 4.0), 7.20 (5H, m) (lit.
25). 2-n-Butyl-3-phenyloxirane: δH (250 MHz, CDCl3) 0.90-1.90
(7H, m), 2.95t (1H, dt, J ) 2.0, 5.5), 3.15c (1H, m), 3.61t (1H, d, J )
2.0), 4.08c (1H, d, J ) 4.5), 7.35 (5H, m) (lit. 25).
Preparation of (1R,6R,8R)-11,11-Dimethyl-3,5-dithia-tricyclo-
[6.2.1.01,6]undecane (13). The 52:48 mixture of sulfide 12 and benzyl
mercaptan (145 mg, total mass), paraformaldehyde (145 mg, 2.0 mmol),
and p-toluenesulfonic acid (6.3 mg, 0.03 mmol) were refluxed in toluene
(1 mL) in a Dean-Stark trap under nitrogen. Water was added after
a few hours, and after separation, the organic layer was washed with
sodium hydroxide (2 M) and water before being dried over MgSO4.
After filtration of the mixture and removal of the solvents under reduced
pressure were complete, chromatography with 10% dichloromethane
in petroleum ether gave dithiane 13 (57 mg, 72% based on 12) and a
mixture of 13 and its C6 epimer: [R]20 -39.41 (c 2.03 in CHCl3);
D
υ
max (thin film)/cm-1 2952, 2879; δH (250 MHz, CDCl3) 0.91 (3H, s),
1.12-1.47 (2H, m), 1.31 (3H, s), 1.52-1.88 (5H, m), 2.59 (1H, d, J )
14.5), 3.01 (1H, d, J ) 14.5), 2.98-3.18 (1H, m), 3.61 (1H, d, J )
12.5), 3.87 (1H, d, J ) 12.5); δC (63 MHz, CDCl3) 20.81, 21.77, 27.17,
29.86, 30.69, 35.88, 37.36, 44.92, 45.59, 46.27, 48.26; m/e (EI) 214
(M+, 100%), 168 (27), 93 (26) (Found: M+, 214.0843. C11H18S2
requires 214.0850).
Preparation of (1S,2S,4R)-1-Mercaptomethyl-7,7-dimethyl-2-
vinyl-bicyclo[2.2.1]heptan-2-ol. Cerium chloride heptahydrate (6.08
g, 16.3 mmol) was dried under vacuum at 150 °C for 2.5 h before
being suspended in tetrahydrofuran (20 mL) under nitrogen. This
(77) Padwa, A.; Gasdaska, J. R. Tetrahedron 1988, 44, 4147-4156.

8338 J. Am. Chem. Soc., Vol. 120, No. 33, 1998
Aggarwal et al.
suspension was submitted to sonication for 1 h and then stirred at room
temperature for 1 h. After cooling at -78 °C, vinylmagnesium bromide
(16.5 mL of a 1 M solution in tertrahydrofuran) was added. After 2 h,
ketone 11 (1 g, 5.35 mmol) was added portionwise to the rapidly stirred
suspension (over 7 min). The reaction mixture was allowed to warm
to room temperature overnight before being quenched with water;
hydrochloric acid (3 M aqueous) was added to the mixture to dissolve
the solids. The resulting mixture was extracted with petroleum ether
(100 mL and 3 × 50 mL), and the combined extracts were dried over
MgSO4. After filtration of the mixture and removal of the solvents
under reduced pressure were complete, chromatography with 50%
dichloromethane in petroleum ether gave the desired thioalcohol (864
mg, 75%): υmax (thin film)/cm-1 3478, 2951, 2582, 1732; δH (250 MHz,
CDCl3) 0.91 (3H, s),1.02-1.21 (1H, m), 1.16 (3H, s), 1.35 (1H, t, J )
7.5), 1.51-1.84 (6H, m), 1.99 (1H, ddd, J ) 13.0, 3.5, 3.5), 2.55 (1H,
dd, J ) 13.5, 7.5), 2.99 (1H, dd, J ) 13.5, 7.5), 5.08 (1H, dd, J )
10.5, 1.0), 5.26 (1H, dd, J ) 17.5, 1.0), 6.36 (1H, dd, J ) 17.5, 10.5);
δC (63 MHz, CDCl3) 21.35, 21.66, 22.78, 26.38, 27.89, 45.49, 45.90,
50.74, 55.74, 81.82, 112.00, 145.10; m/e (EI) 212 (M+, 35%), 194 (34),
108 (100) (Found: M+, 212.1242. C12H20OS requires 212.1235).
Preparation of (1S,6R,8R)-11,11-dimethyl-3-thia-tricyclo[6.2.1.01,6]-
undecan-6-ol (14). 9-Borabicyclo[3.3.1]nonane (0.61 mL of a 0.5 M
solution in tetrahydrofuran, 0.31 mmol) was added to an ice-cooled
solution of (1S,2S,4R)-1-mercaptomethyl-7,7-dimethyl-2-vinyl-bicyclo-
[2.2.1]heptan-2-ol (0.643 g, 3.03 mmol) in tetrahydrofuran (5.5 mL)
under nitrogen. After 4 h, the solution was allowed to warm to room
temperature and was left overnight. The reaction mixture was diluted
with petroleum ether and filtered through a short silica plug, which
was then washed with dichloromethane. After removal of the solvents
under reduced pressure, the residue was loaded onto a silica gel column
and eluted with 5% ethyl acetate in petroleum ether to give sulfide 14
(506 mg, 79%) as a white solid, mp 34-35 °C: [R]20D +57.0 (c 10.00
in CHCl3); υmax (solution, CHCl3)/cm-1 3473, 2934, 1066, 911; δH (400
MHz, CDCl3) 0.88 (3H, s), 0.99 (1H, ddd, J ) 11.5, 9.5, 4.5), 1.13
(3H, s), 1.29 (1H, d, J ) 13.5), 1.43 (1H, br s), 1.53 (1H, ddd, J )
14.0, 12.0, 4.5), 1.64 (1H, m), 1.76 (1H, dddd, J ) 14.0, 4.5, 2.5, 1.0),
1.78 (1H, dd, J ) 4.5, 4.0), 2.03 (1H, ddd, J ) 13.5, 5.0, 3.0), 2.06
(1H, dd, J ) 12.0, 1.0), 2.19 (1H, ddd, J ) 14.0, 13.5, 5.0), 2.25 (1H,
ddd, J ) 14.0, 9.5, 3.5), 2.33 (1H, dddd, J ) 12.0, 5.0, 2.5, 1.0), 3.14
(1H, ddd, J ) 13.5, 12.0, 4.5), 3.19 (1H, d, J ) 12.0); δC (101 MHz,
CDCl3) 20.91, 22.45, 23.18, 26.55, 27.09, 30.57, 34.93, 45.14, 46.96,
48.63, 50.78, 76.91; m/e (EI) 212 (M+, 52%), 194 (21), 108 (100)
(Found: M+, 212.1230. C12H20OS requires 212.1235).
Preparation of (1S,8R)-11,11-Dimethyl-3-thia-tricyclo[6.2.1.01,6]-
undec-5-ene(15). (1S,6R,8R)-11,11-Dimethyl-3-thia-tricyclo[6.2.1.01,6]-
undecan-6-ol (14) (578 mg, 2.73 mmol) was added to a solution of
oxalyl chloride (1.16 mL, 13.53 mmol) in benzene (2.5 mL) under
nitrogen. After 5 h, the solvents and excess reagent were removed
under high vacuum, and the residue was eluted through a short silica
gel column with 10% ethyl acetate in petroleum ether to give sulfide
15 (392 mg, 74%) as a colorless oil: [R]20D -44.6 (c 1.57 in CHCl3);
υmax (thin film)/cm-1 2943, 2874, 1686, 1448; δH (400 MHz, CDCl3)
0.76 (3H, s, H13), 0.93 (3H, s), 1.20 (1H, ddd, J ) 12.0, 9.5, 4.5), 1.61
(1H, ddd, J ) 12.0, 12.0, 4.5), 1.73 (1H, dd, J ) 4.5, 4.5), 1.76-1.94
(3H, m), 2.33-2.41 (1H, m), 2.41 (1H, dd, J ) 13.0, 1.0), 2.68 (1H,
d, J ) 13.0), 2.81 (1H, dm, J ) 17.0), 3.28 (1H, dm, J ) 17.0), 5.50
(1H, dm, J ) 5.5); δC (101 MHz, CDCl3) 18.52, 19.49, 25.57, 27.37,
28.46, 32.92, 36.43, 43.82, 47.95, 48.94, 113.77, 146.87; m/e (EI) 194
(M+, 100%), 179 (37), 151 (47) (Found: M+, 194.1128. C12H18S
requires 194.1129).
Preparation of (1R,6S,8R)-11,11-Dimethyl-3-thia-tricyclo[6.2.1.01,6]-
undecane (16). Palladium on 10% activated carbon (3.44 g) was added
to a solution of (1S,8R)-11,11-dimethyl-3-thia-tricyclo[6.2.1.01,6]undec-
5-ene (15) (344 mg, 1.77 mmol) in anhydrous methanol (80 mL). This
mixture was submitted to hydrogenation at 15-20 atm at room
temperature for 22 h. After the mixture was filtered through Celite
and the solvent removed, the residue was loaded onto a silica gel column
and eluted with petroleum ether to give sulfide 16 (291 mg, 89%) as
a mixture of diastereomers (3:1). To a solution of the diastereomeric
mixture of sulfide 16 (229 mg, 1.17 mmol) in dichloromethane (1.5
mL) was added a solution of 3-chloroperoxybenzoic acid (258 mg, 1.40
mmol) in dichloromethane (1.5 mL) at 0 °C under nitrogen. After 1 h
the mixture was diluted with dichloromethane (15 mL) and washed
with aqueous sodium hydrogencarbonate solution (1 × 15 mL) and
brine (1 × 15 mL) before drying over MgSO4. After filtration and
removal of the solvent were complete, the residue (244 mg, 98%) was
recrystallized in petrol/dichloromethane to give sulfoxide 17 (124 mg,
50%): [R]20D -97.5 (c 0.4 in CHCl3); υmax (thin film)/cm-1 2839, 2872,
1464, 1016; δH (400 MHz, CDCl3) 0.96 (3H, s), 1.18-1.24 (3H, m),
1.47-1.75 (8H, m), 2.06-2.09 (1H, m), 2.47-2.53 (1H, m), 2.55-
2.59 (1H, d, J ) 12.9), 3.26-3.32 (1H, m), 3.32-3.36 (1H, dd, J )
12.9, 2.3); δC (101 MHz, CDCl3) 20.9, 23.1, 27.4, 29.6, 35.8, 40.5,
42.4, 46.4, 47.3, 51.0, 54.9; m/e (EI) 212 (M+, 38%), 195 (24), 163
(100). A solution of trifluoroacetic anhydride (0.06 mL, 0.40 mmol)
in acetone (0.5 mL) was added slowly to a mixture of sulfoxide 17 (50
mg, 0.24 mmol) and dried sodium iodide (85 mg, 0.57 mmol) in acetone
(1 mL) at 0 °C under nitrogen. After 20 min the mixture was diluted
with ethyl acetate (20 mL) and washed with sodium thiosulfate solution
(1 × 20 mL) and brine (1 × 20 mL). After being dried over MgSO4
and having the solvents removed, the residue was eluted through a
short silica gel column with petroleum ether to give sulfide 16 (41 mg,
89%): [R]20D -92.9 (c 0.7 in CHCl3); υmax (thin film)/cm-1 2949, 2877,
1463, 1385, 1251, 1150, 953; δH (400 MHz, CDCl3) 0.89 (3H, s), 1.04-
1.21 (2H, m), 1.21 (3H, s), 1.46-1.70 (6H, m), 1.97-2.10 (1H, m),
2.44-2.59 (4H, m), 2.69-2.75 (d, J ) 14.4, 1H); δC (63 MHz, CDCl3)
21.1, 23.4, 27.4, 28.4, 29.5, 34.4, 36.8, 40.4, 43.5, 44.3, 46.2, 47.4;
m/e (EI) 196 (M+, 100%).
Preparation of Mesitaldehyde Tosylhydrazone.78 To a slurry of
p-toluenesulfonhydrazide (5.0 g, 27 mmol) in methanol (12 mL) was
added mesitaldehyde (4 mL, 27 mmol). Crystallization of the resultant
white cake from methanol gave the desired hydrazone as a white
crytalline solid, mp 159-161 °C (lit. 79, 158-160 °C) (Found: C,
64.47; H, 6.38; N, 8.89; S, 10.06. C17H20N2O2S requires C, 64.53; H,
6.37; N, 8.85; S, 10.13): υmax (KBr disk)/cm-1 3203, 1608, 1557, 1326,
1165; δH (250 MHz, CDCl3) 1.95 (3H, s), 2.25 (6H, m), 2.41 (3H, m),
6.84 (2H, m), 7.25-7.35 (2H, m), 7.55 (1H, br d, CH), 7.80 (2H, m);
δC (63 MHz, CDCl3) 21.1, 21.3, 21.6, 127.1, 128.1, 129.6, 135.3, 137.9,
139.3, 144.2, 148.0; m/e (EI) 316 (M+, 37%) 161 (100), 132 (96), 91
(77).
Preparation of Mesityldiazomethane 22.80 A suspension of
mesitaldehyde tosylhydrazone (2.41 g, 7.62 mmol) in benzene (12 mL)
was heated with sodium hydroxide (12 mL of a 14% solution) and
benzyl triethylammonium chloride (0.32 g, 1.39 mmol) at 60-70 °C
for 1 h. After cooling, the organic layer was separated and washed
with sodium hydroxide (14%, 2 × 15 mL) and water (15 mL) before
being dried over sodium sulfate. Filtration and removal of the solvent
in vacuo gave the desired diazo compound as an orange oil. The diazo
compound was dissolved in dichloromethane and used without further
purification. The concentration of the solution was determined by
titration of 50 µL of the solution with p-toluic acid (0.25 mmol) in
dichloromethane and comparison of the peaks at 5.38 (ester CH2) and
2.43 (p-toluic acid CH3) ppm.
Epoxidation of Benzaldehyde Using Sulfide 5b and Mesityldiazo-
methane. To a stirred solution of sulfide 5b (19.0 mg, 0.09 mmol),
Cu(acac)2 (6.1 mg, 0.023 mmol), and benzaldehyde (50 µL, 0.49 mmol)
in dichloromethane (0.25 mL) under nitrogen was added a solution of
mesityldiazomethane 22 (0.75 mmol in 0.25 mL of dichloromethane)
at room temperature over a period of 3 h using a syringe pump. After
being stirred for an additional 1 h, the solvent was removed in vacuo
and the residue was chromatographed on silica gel to give the desired
epoxide as a white solid (13.8 mg, 12%), mp 66-68 °C (lit. 81, 67-
68 °C) (Found: C, 85.45; H, 7.58. C17H18O requires C, 85.67; H,
7.61): υmax (thin film)/cm-1 2971, 2918, 1607, 890, 792; δH (250 MHz,
CDCl3) 2.28 (3H, s), 2.41 (9H, s), 3.82 (1H, d, J ) 2.0), 3.90 (1H, br
d), 6.86 (2H, s), 7.26-7.54 (5H, m); δC (63 MHz, CDCl3) 19.88, 20.95,
(78) Creary, X. Org. Synth. 1985, 64, 207-216.
(79) Nozaki, H.; Noyori, R.; Sisido, K. Tetrahedron 1964, 20, 1125-
1132.
(80) Wulfman, D. S.; Yosefian, S.; White, J. M. Synth. Commun. 1988,
18, 2349-2352.
(81) Fuson, R. C.; Denton, J. J.; Best, C. E. J. Org. Chem. 1943, 8,
64.

Catalytic Asymmetric Epoxidation of Aldehydes
J. Am. Chem. Soc., Vol. 120, No. 33, 1998 8339
60.00, 62.11, 125.50, 128.26, 128.61, 128.72, 130.96, 137.05, 137.51;
m/e (EI) 238 (M+, 26%), 223 (48), 132 (100), 117 (97%) (Found: M+,
238.1357. C17H18O requires 238.1358).
Supporting Information Available: Experimental proce-
dures, including data for the preparation of compounds not given
below, tables of crystallographic data for 5p, 10, and 17, and
methods for the enantiomeric excess determination of epoxides
(35 pages, print/PDF). See any current masthead page for
ordering information and Web access instructions.
Acknowledgment. We thank Sheffield University and
Zeneca for support of a studentship (J.G.F.) and the European
HCM program (EU Fellowship ERBFMBICT972053) for a
fellowship (S.F.).
JA9812150