A highly enantioselective one-pot sulfur ylide epoxidation reaction

Article abstract of DOI:10.1016/S0040-4039(02)01072-9

Enantiomerically pure tricyclic C₂-symmetric sulfide 1 can be prepared in high yield and only three steps from D-mannitol. When used in a one-pot sulfur ylide mediated epoxidation reaction, between 94 and 98% enantiomeric excess was obtained; the highest reported to date for the catalytic process. Molecular modelling studies of the ylide conformation provides a basis for understanding the stereochemical outcome of the reaction.

Full text of DOI:10.1016/S0040-4039(02)01072-9

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 5427–5430
A highly enantioselective one-pot sulfur ylide epoxidation
reaction
Caroline L. Winn, Benjamin R. Bellenie and Jonathan M. Goodman*
Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, UK
Received 22 March 2002; revised 24 May 2002; accepted 7 June 2002
Abstract—Enantiomerically pure tricyclic C2-symmetric sulfide 1 can be prepared in high yield and only three steps from
D
-mannitol. When used in a one-pot sulfur ylide mediated epoxidation reaction, between 94 and 98% enantiomeric excess was
obtained; the highest reported to date for the catalytic process. Molecular modelling studies of the ylide conformation provides
a basis for understanding the stereochemical outcome of the reaction. © 2002 Elsevier Science Ltd. All rights reserved.
Our aim was to design and synthesise a chiral sulfide
that could be easily prepared in a few, high-yielding
steps from a readily available and reasonably priced
starting material, and which could find application in
enantioselective epoxidation reactions and other areas
of asymmetric synthesis.1 Chiral sulfide 1 (a crystalline,
odourless compound) can be prepared in three steps
93% enantiomeric excess, respectively, and a high
degree of diastereoselectivity.
Another elegant solution to the problem of sulfide
alkylation was reported by Aggarwal et al. who used an
elegant rhodium-catalysed route towards enantiomeri-
cally enriched epoxides, aziridines and cyclopropanes.7
This involved direct addition of a carbene to the sulfide,
alleviating the need for an extra deprotonation step and
facilitating alkyl addition to more hindered and less
nucleophilic substrates.
and 76% overall yield from D-mannitol, a readily avail-
able and cheap starting material.2 High degrees of
enantioselectivity were obtained when this sulfide was
used in a one-pot sulfur ylide mediated epoxidation
reaction.
O
BnBr, PhCHO, NaOH
Some reports in the literature described problems that
had been encountered in the formation or isolation of
the sulfonium salt which is an intermediate in the
reaction pathway. In our case, there was concern that
the steric hindrance from the groups a to the sulfur
atom would hinder addition of an alkyl group, and the
presence of four heteroatoms in the molecule would
reduce the nucleophilicity of the sulfur to such an
extent that alkylation could not occur.3
room temp.
Ph
Ph
S
(S,S), 93% ee
3
Ph
Ph
O
O
O
BnBr, PhCHO, NaOH
room temp.
O
O
Ph
Ph
S
(R,R), 94-98% ee
1
A catalytic process eliminates the need to isolate the
intermediate sulfonium salt. The first example of this
was reported by Furukawa,4 although this method suf-
fered from relatively low enantioselectivities. Metzner et
al.5,6 used a similar one-pot epoxidation method
(Scheme 1). Using sulfides 2 and 3 (Fig. 1), stilbene
oxide was produced in good yield with up to 88% and
Scheme 1. One-pot epoxidation reactions used by Metzner
(top) and in this study (bottom).
Ph
Ph
O
O
O
O
S
S
S
1
2
3
Figure 1. Chiral sulfides used in one-pot epoxidation reac-
* Corresponding author. Fax: +44-1223-336362; e-mail: j.m.goodman
@ch.cam.ac.uk
tions.
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)01072-9

5428
C. L. Winn et al. / Tetrahedron Letters 43 (2002) 5427–5430
Following the procedure of Metzner (Scheme 1), it was
possible to form the desired product in 42% yield after
2 days stirring at room temperature in an acetoni-
trile:water mixture (9:1). The enantiomeric excess was
found to be 94% (R,R) while the trans:cis ratio
observed was 1:0.1 (Table 1, entry 1). With a longer
reaction time of 7 days, the yield could be increased to
59% with no loss of enantioselectivity; 97% ee was
observed (entry 3). In all cases the trans isomer was
formed preferentially and chiral HPLC identified the
major isomer as (R,R)-diphenyloxirane. In each case
the sulfide was recovered in good yield, indicating that
the reaction could be carried out catalytically. With 0.1
or 0.2 equiv. of sulfide (entries 4 to 8), similar high
enantioselectivities were observed, but the reaction rate
and yields were reduced.
the desired product, albeit with a slightly lower enan-
tioselectivity of 84% (entry 2). para-Chloro- and para-
fluorobenzaldehyde could also be employed in the same
reaction, resulting in epoxides with a very high degree
of enantiomeric excess (>95%, entries 3, 4), although in
these cases the reaction yield is low. These results were
as expected with respect to the relative reactivities of
the aldehydes; the more reactive para-nitrobenzalde-
hyde gave a higher product yield but a lower selectivity
than the less reactive halo-derivatives.
The enantioselectivity achieved in the epoxidation reac-
tions may be influenced by the ylide geometry as shown
in Fig. 2. Molecular mechanics calculations using the
MM2* forcefield within MacroModel8,9 were carried
out on sulfide 1 to determine the lowest energy confor-
mation. The global minimum conformation found
using a Monte Carlo conformational search was used
to build models of the ylide, which were minimised at
RHF/6-31G**//RHF/6-31G** using GAMESS-US.10
The enantioselectivities observed are approximately
constant, and average 96%. In a further experiment,
aliquots were taken at regular intervals from a larger
scale reaction set up using 0.1 equiv. of sulfide, recrys-
tallised twice from toluene. No drop in ee was observed
over 7 days (enantiomeric excesses observed range from
97 to 99%), and the trans:cis ratio was greater than
1:0.12 in all cases. The small variation in diastereoselec-
tivity noted in Table 1 is probably not significant, as the
NMR measurement of the ratio is less precise than the
HPLC method used for the ee.
The lowest energy and hence most populated ylide
conformation has the phenyl group pointing away from
the ring (1b), minimising unfavourable steric interac-
tions between the ring substituents and the aromatic
group. It may be expected that the attack of the ylide
onto the aldehyde would occur from the re face, leading
to the formation of the (R,R) epoxide, as is observed
experimentally (Fig. 3).
The investigation was further extended to a range of
aldehydes (Table 2). Unfortunately, no reaction
We regard this as a mnemonic for the sense of selectiv-
ity, as it does not represent a full account of the
processes going on in the reaction. It is qualitatively
consistent with Metzner’s results for sulfide 2: attack
occurred
when
para-methoxybenzaldehyde
was
employed (presumably due to its low reactivity, entry
1), but para-nitrobenzaldehyde succeeded in affording
Table 1. Initial results of epoxidation studiesa
Entry
Sulfide equiv.
Time (days)
% Yieldb
% Sulfide recovered
trans:cisc
%eed
1
2
3
4
5
6
7
8
1
1
1
0.2
0.1
0.1
0.1
0.1
2
5
7
3
2
2
2
4
42
45
59
27
20
16
14
41
81
70
93
85
96
86
80
90
1:0.10
1:0.11
1:0.08
1:0.07
1:0.20
1:0.11
1:0.17
1:0.11
94 (R,R)
96
97
95
94
98
96
97
a Reactions were conducted on a 0.1–2 mmol scale using 1 equiv. benzaldehyde, 2 equiv. benzyl bromide, 2 equiv. sodium hydroxide, in a 9:1
mixture of acetonitrile and distilled water.
b Isolated yield of trans epoxide
c Diastereomeric excess determined by NMR of product mixture.
d Enantiomeric excess determined by chiral HPLC using a Chiracel™ OD column.
Table 2. Results of epoxidation studies using a range of aldehydesa
Entry
R(R
6
CHO)
Time (days)
% Yield
% Sulfide recovered
%de
%ee
1
2
3
4
p-MeOC6H4
p-NO2C6H4
p-ClC6H4
2
2
2
3
–
71
20
19
75
77
86
83
–
100
100
100
–
84
96
98
p-FC6H4
a 1 equiv. of sulfide 1 was used in all reactions except entry 4, where 0.1 equiv. of sulfide were used. Reaction conditions and methods were the
same as for Table 1.

C. L. Winn et al. / Tetrahedron Letters 43 (2002) 5427–5430
5429
remarkably close to that observed experimentally.
However, the calculated energy difference between the
ylide conformations for sulfide 2 corresponds to an ee
of just 17%; a much smaller enantioselectivity than that
observed in the reaction. This indicates that the ylide
conformation is not the only important factor in con-
trolling the stereoselectivity of these reactions.12 The
close agreement of the initial calculation may be fortu-
itous; a higher level of calculation including solvent
effects may enable further insights.13
In conclusion, we have shown that readily accessible,
crystalline C2-symmetric sulfide 1 can be used in cata-
lytic, transition-metal free, sulfur ylide mediated epoxi-
dation reactions. The enantioselectivities obtained
(averaging 96%) are, as far as we are aware, the highest
reported to date for this reaction, except for the work
of Solladie´-Cavallo et al., which requires a two-step
epoxidation procedure.14 Future prospects include the
extension of the scope of this reaction and the applica-
tion of sulfide 1 to other asymmetric reactions.
Acknowledgements
This work was supported by grants from the Royal
Society and the Engineering and Physical Sciences
Research Council.
Figure 2. Two possible ylide geometries, with the phenyl ring
pointing towards (1a and 2a) or away (1b and 2b) from the
ring. Hydrogens have been omitted.
References
1. Li, A.-H.; Dai, L.-X.; Aggarwal, V. K. Chem. Rev. 1997,
97, 2341–2372.
2. (a) Rao, A. V. R.; Reddy, K. A.; Gurjar, M. K.; Kunwar,
A. C. J. Chem. Soc., Chem. Commun. 1988, 1273–1274;
(b) Winn, C.; Goodman, J. M. Tetrahedron Lett. 2001,
42, 7091–7093.
3. Durst, T.; Breau, L.; Ben, R. N. Phosphorus, Sulfur,
Silicon 1993, 74, 215–232.
4. Furukawa, N.; Sugihara, Y.; Fujihara, H. J. Org. Chem.
1989, 54, 4222–4224.
5. (a) Julienne, K.; Metzner, P.; Henyron, V.; Greiner, A. J.
Org. Chem. 1998, 63, 4532–4534; (b) Julienne, K.; Met-
zner, P.; Henyron, V. J. Chem. Soc., Perkin Trans. 1
1999, 731–735; (c) Julienne, K.; Metzner, P.; Greiner, A.;
Henyron, V. Phosphorus, Sulfur, Silicon 1999, 341–342;
(d) Metzner, P.; Alayrac, C.; Julienne, K.; Nowaczyk, A.
Actualite´ Chim. 2000, 54–59.
Figure 3. For sulfide 1, attack from the re face is less hin-
dered than attack from the si face. The expected, lower energy
ylide conformation (with the phenyl group pointing away
from the ring) is shown. For Metzner’s sulfide 2, si face attack
is less hindered. Hydrogens have been omitted.
6. Zanardi, J.; Leriverend, C.; Aubert, D.; Julienne, K.;
Metzner, P. J. Org. Chem. 2001, 66, 5620–5623.
7. (a) Aggarwal, V. K.; Alonso, E.; Hynd, G.; Lydon, K.
M.; Palmer, M. J.; Porcelloni, M.; Studley, J. R. Angew.
Chem., Int. Ed. Engl. 2001, 8, 1430–1433; (b) Aggarwal,
V. K.; Alonso, E.; Fang, G.; Ferrara, M.; Hynd, G.;
Porcelloni, M. Angew. Chem., Int. Ed. Engl. 2001, 8,
1433–1436.
occurs from the si face of the favoured ylide conforma-
tion 2b (Fig. 3) leading to the formation of the (S,S)
epoxide. The larger substituents on sulfide 1 (with
respect to the methyl groups adjacent to the sulfur
atom in Metzner’s compound 2) are responsible for the
increased enantioselectivity observed.
8. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.;
Liskamp, R.; Lipton, M.; Caufield, C.; Chang, C.; Hen-
drickson, T.; Still, W. C. J. Comp. Chem. 1990, 11,
440–467.
The energy difference between the ylide conformations
1a and 1b was calculated to be 8.5 kJ/mol. Using
Boltzmann factors11 to compute relative populations of
the conformations at 298 K, this energy difference
corresponds to a limiting enantiomeric excess of 94%;
9. Allinger, N. L. J. Am. Chem. Soc. 1977, 99, 440–467.

5430
C. L. Winn et al. / Tetrahedron Letters 43 (2002) 5427–5430
10. Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S.
T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga,
N.; Nguyen, K. A.; Su, S. J.; Windus, T. L.; Dupuis, M.;
Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347–
1363.
12. Other computational approaches to prediction of enan-
tioselectivity include: (a) Myllima¨ki, V. T.; Lindvall, M.
K.; Koskinen, A. M. P. Tetrahedron 2001, 57, 4629–4635;
(b) Lindvall, M. K.; Koskinen, A. M. P. J. Org. Chem.
1999, 64, 4596–4606.
13. Aggarwal, V. K.; Harvey, J. N.; Richardson, J. J. Am.
11. Goodman, J. M.; Kirby, P. D.; Haustedt, L. O. Tetra-
hedron Lett. 2000, 41, 9879–9882 on line calculation of
population from relative energies can be found at http://
www.ch.cam.ac.uk/magnus/boltz.html.
Chem. Soc. 2002, 124, 5747–5756.
14. Solladie´-Cavallo, A.; Diep-Vohuule, A.; Sunjic, V.;
Vinkovic, V. Tetrahedron: Asymmetry 1996, 7, 1783–1788.