Reactions of silyl-stabilised sulfur ylides with organoboranes: Enantioselectivity, mechanism, and understanding

Article abstract of DOI:10.1039/b718496d

The reaction of trimethylsilyl-substituted sulfonium ylides with organoboranes (Ph₃B, Et₃B) has been studied and although homologated products were obtained in good yield (after oxidation to the corresponding alcohols), the enantiomeric excesses were low with our camphor-based chiral sulfide (up to 40% ee, cf. corresponding phenyl-substituted sulfonium ylides gave >95% ee). Cross-over experiments were conducted to ascertain the nature of this difference in selectivity. Thus, aryl- and silyl-substituted sulfonium ylides (1 equiv.) were (separately) reacted with Et₃B (1.5 equiv.) followed by Ph₃B (1.5 equiv.) The experiments were repeated changing the order of addition of the two boranes. It was found that the aryl-substituted sulfonium ylide only trapped the first borane that was added indicating that ate complex formation was non-reversible and so was the selectivity determining step. In contrast the silyl-substituted sulfonium ylide only trapped Ph₃B (it is more reactive than Et ₃B) indicating that ate complex formation was reversible and so 1,2-migration was now the selectivity determining step. The reactions have been studied computationally and the experimental observations have been reproduced. They have further revealed that the cause of reversibility in the case of the silyl-substituted sulfonium ylides results from ate complex formation being less exothermic and a higher barrier to 1,2-migration. This journal is The Royal Society of Chemistry.

Full text of DOI:10.1039/b718496d

View Article Online / Journal Homepage / Table of Contents for this issue
www.rsc.org/obc
Volume 6 | Number 7 | 7 April 2008 | Pages 1129–1300
ISSN 1477-0520
FULL PAPER
Dean Howells et al.
Reactions of silyl-stabilised sulfur
ylides with organoboranes:
enantioselectivity, mechanism, and
understanding
1477-0520(2008)6:7;1-A

PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Reactions of silyl-stabilised sulfur ylides with organoboranes:
enantioselectivity, mechanism, and understanding†
a
b
a
a
c
Dean Howells, Rapha e¨ l Robiette, Guang Y. Fang, Luke S. Knowles, Michael D. Woodrow,
a
a
Jeremy N. Harvey* and Varinder K. Aggarwal*
Received 29th November 2007, Accepted 28th January 2008
First published as an Advance Article on the web 28th February 2008
DOI: 10.1039/b718496d
The reaction of trimethylsilyl-substituted sulfonium ylides with organoboranes (Ph
3
B, Et B) has been
3
studied and although homologated products were obtained in good yield (after oxidation to the
corresponding alcohols), the enantiomeric excesses were low with our camphor-based chiral sulfide (up
to 40% ee, cf. corresponding phenyl-substituted sulfonium ylides gave >95% ee). Cross-over
experiments were conducted to ascertain the nature of this difference in selectivity. Thus, aryl- and
silyl-substituted sulfonium ylides (1 equiv.) were (separately) reacted with Et
3
B (1.5 equiv.) followed by
Ph B (1.5 equiv.) The experiments were repeated changing the order of addition of the two boranes. It
3
was found that the aryl-substituted sulfonium ylide only trapped the first borane that was added
indicating that ate complex formation was non-reversible and so was the selectivity determining step. In
contrast the silyl-substituted sulfonium ylide only trapped Ph
3
B (it is more reactive than Et B)
3
indicating that ate complex formation was reversible and so 1,2-migration was now the selectivity
determining step. The reactions have been studied computationally and the experimental observations
have been reproduced. They have further revealed that the cause of reversibility in the case of the
silyl-substituted sulfonium ylides results from ate complex formation being less exothermic and a higher
barrier to 1,2-migration.
Introduction
In a desire to extend the scope of the reaction we considered the
possibility of using silyl-substituted ylides. Such ylides are known
and their reactions with carbonyl compounds result in alkene
formation (Peterson reaction) rather than epoxidation (Corey–
Chaykovsky reaction). Further reactions of silyl-substituted
ylides with boranes have not been described although Matteson
The reaction of sulfur ylides with organoboranes to give ho-
mologated boranes is emerging as a useful synthetic method in
asymmetric synthesis. So far aryl-stabilised ylides 1 have been
reacted with alkyl, aryl and vinyl boranes to give the corresponding
homologated boranes with very high enantioselectivity (>95% ee).
1
3
has reported the reactions of silyl-substituted carbenoid 2 with
These reactions occur via intermediate ate complexes followed by
4
chiral borate esters (Scheme 2). This gave the corresponding
2
1
,2-metallate rearrangement (Scheme 1).
homologated product 3 but with low diastereoselectivity (73 : 27)
and thus after oxidation the silylcarbinol 4 was obtained with low
enantiomeric excess (42% ee).
Scheme 1 Reaction of aryl-stabilised ylides with organoboranes.
Scheme 2 Matteson’s homologation of boronic esters with silyl-stabilised
a
carbenoid.
University of Bristol School of Chemistry, University of Bristol, Bristol, BS8
1
TS, UK. E-mail: V.Aggarwal@Bristol.ac.uk; Fax: +44 (0)117 9298611;
We expected the silyl-stabilised sulfur ylide to react with boranes
with similarly high enantioselectivity as the phenyl-stabilised
Tel: +44 (0)117 954 6315
b
Unit e´ de Chimie organique et m e´ dicinale, Universit e´ catholique de Louvain,
Pl. L. Pasteur 1, B-1348 Louvain-la-Neuve, Belgium
ylide because: (i) both ylides can be expected to have similar
c
II CEDD Discovery Chemistry, GlaxoSmithKline R&D, Stevenage, SG1
5
pK
a
’s and so one can expect both reactions to be non-reversible.
2
NY, UK
6
7
8
(
ii) Ph and Me Si have similar A values (2.7 and 2.4–2.6
3
†
Electronic supplementary information (ESI) available: Synthesis and
respectively) and so ylide conformation, a key factor in achieving
high enantioselectivity, should be similarly well controlled. Indeed,
characterisation of all compounds and computational details. See DOI:
0.1039/b718496d
1
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 1185–1189 | 1185

we have found that in the case of the phenyl-stabilised ylide 1 there
reacting much slower than the minor conformer (an example of
11
is a very large preference for one ylide conformation over the other
the Curtin–Hammett principle ), since in the TS the bulky Me Si
3
9
(
Scheme 3).
group is pushed toward the bulky camphor moiety (Scheme 5).
Indeed, DFT calculation of the phenyl stabilised ylide–borane
reaction and the corresponding ylide-aldehyde reaction revealed
that there is substantially greater pyramidalisation in the TS of
1c
the former reaction. The greater degree of pyramidalisation in
the TS of the ylide–borane reaction compared to the epoxidation
reaction results in the ylide substituents being pushed further
towards the sulfonium scaffold in the ylide–borane reaction which
in turn means that the reaction will be more sensitive to the steric
environment around the ylidic carbon.
Scheme 3 Energy difference between ylide conformers.
In this paper we describe our studies on the use of silyl-
substituted ylides and the surprising finding that their addition
to boranes is, in fact, reversible.
Results and discussion
The required sulfonium salts (5, 6) were easily prepared
by reaction of the appropriate sulfides with the commercially
10
available triflate (Scheme 4). Reactions of the salts 5 and 6 with
LHMDS and two different boranes (Ph B and Bu B) furnished
the silylcarbinols in moderate yield after oxidative work up
Table 1). However, we were surprised and disappointed at the
3
3
(
low enantioselectivity obtained with the chiral sulfonium salt 6 in
both cases (entry 2,4), and so further optimisation of yield was
not conducted.
We considered two possible reasons for the low enantioselec-
tivity. Our first consideration was that the two ylide conformers
might react with vastly different rates—the major conformer
Scheme 5 Potential origin of low enantioselectivity in ylide reaction.
Although difficult to prove, we tested this proposition by
using an alternative chiral sulfide 7 for which we had previously
established good control over ylide conformation and face se-
lectivity (Scheme 6: the phenyl stabilised ylide 8 gave 95 : 5
12
er in epoxidation ). In this case however, there is less steric
encumbrance around the ylidic carbon and so the rates of reactions
of the two ylide conformers can be expected to be more equal.
However, in the event even lower enantioselectivity was observed
in reaction of the silyl-stabilised ylide 9 with Ph B (entry 5).
3
These results suggest that the origin of the low enantioselectivity
is unlikely to originate from the major ylide conformer reacting
more slowly than the minor ylide conformer.
Scheme 4 Synthesis of sulfonium salts.
Table 1 Reaction of silyl-stabilised ylides with organoboranes
Entry
Salt
R
Yield (%)
Er (%)
—
Scheme 6 Demonstration of control of ylide conformation and face
selectivity through epoxidation.
1
2
3
4
5
5
6
5
6
9
Bu
Bu
Ph
Ph
Ph
63
72
61
66
41
a
60 : 40
—
A second possibility was that the addition of the silyl-stabilised
ylide onto the borane was reversible and that the two diastereomers
of the ate complex underwent migration at similar rates. Although
3
b
b
71 : 29
57 : 43
a
The enantiomeric ratio was determined by formation of the Mosher’s
a reversibility of the addition did not seem that likely since Me Si
b
ester. The enantiomeric ratio was determined by chiral HPLC using
chiracel OD column under the following conditions: 5% IPA in hexane,
5,13
is slightly less able to stabilise a carbanion than a Ph group
−
1
(so reversal of ate complex formation should be less favoured) it
could be easily tested by cross-over experiments. The experiment
1
ml min . Retention time: minor, 6.96 min; major, 10.53 min.
1
186 | Org. Biomol. Chem., 2008, 6, 1185–1189
This journal is © The Royal Society of Chemistry 2008

would entail addition of borane A to an ylide at low temperature
and after complete consumption of the ylide addition of a second
borane B. If ate complex formation is non-reversible this should
furnish the product derived solely from borane A but if ate complex
formation is reversible a mixture of products derived from boranes
A and B would be produced.
To establish the temperature and time required for the ylide to
react with the borane, the reaction of ylides 5 and 10 with Et B
3
1
1
was monitored by B NMR (Scheme 7). In fact, within 10 minutes
◦
at −78 C ate complex formation from the reaction of ylide 5 with
Et
3
B was essentially complete although rearrangement only took
◦
place upon warming to −10 C. In contrast, the reaction of the
phenyl-stabilised ylide 10 could not be monitored effectively. At
◦
−
78 C substantial migration had already occurred and it was
difficult practically to monitor the reaction at lower temperature.
Circumstantial evidence suggested to us that ate complex forma-
◦
tion was very fast at −105 C but that no significant migration
occurred at that temperature: when we conducted reactions of
ylide 10 with Et
3
B/Ph
3
B at this temperature, no products derived
◦
from a second homologation were observed whereas at −78 C
approximately 10% yield of the second homologation adducts were
obtained.
Scheme 8 Cross-over experiments of sulfur ylides with boranes.
ylide, either the barrier to alkyl/aryl group migration must be
raised or the barrier to reversion is lowered, or both. In order
to distinguish between these possibilities and to understand why
there should be such differences in the reactions we carried out
calculations to try to shed light on this unusual situation.
In a previous study, we have studied the reaction of non-
14
substituted and phenyl-substituted ylides with BMe
3
.
These
Scheme 7 Establishing reactivity of sulfur ylides with boranes.
calculations revealed a high exothermicity of the ate complex
−
1
formation (≈30 kcal mol ), which occurred without enthalpic
−
1
In the event reaction of the phenyl-stabilised ylide with Et
followed by Ph B only furnished the product from Et B whilst
under the same conditions the silyl-stabilised ylide only furnished
the product from Ph B (Scheme 8, equations 3 and 1). Changing
the order of addition of the two boranes resulted in the phenyl-
stabilised ylide trapping the BPh and the silyl-stabilised ylide
again trapping the Ph B (Ph B is evidently considerably more
reactive than Et B). This indicated that the phenyl-stabilised
3
B
barrier, and a low barrier to migration (12–15 kcal mol ). This ac-
counts for the non-reversibility of the addition step in these cases.
This study showed also that the MP2/6-311+G**//B3LYP/6-
31G* method was the more adequate to describe these systems. It
is however important to make the point that at this level of theory
the energy of reactants (and hence exothermicity of ate complex
formation) is overestimated by ca. 10 kcal mol− (see SI of ref. 14).
The computed energy profile for the reaction of silyl-substituted
ylide is reported in Fig. 1. Calculations are in good agreement with
cross-over experiments, supporting a reversibility of ate complex
formation (if one takes into account the error involved in the
method). Examination of the profile shows that the reason for this
reversibility is a combination of a decrease in exothermicity of
ate complex formation and an increase (as compared to phenyl-
stabilised ylide) in barrier to alkyl migration (Table 2).
3
3
3
3
1
3
3
3
ylide added non-reversibly to Et
reacted reversibly.
3
B whilst the silyl-stabilised ylide
This means that in the reaction of silyl-stabilised ylides with
organoboranes, addition may well occur faster for one conformer
of the ylide (see Scheme 5), as for phenyl-stabilised ylide, but since
this step is reversible the two diastereomers of the ate complex are
accessible and the selectivity-determining step is the 1,2-migration.
The low enantioselectivity observed in these cases must therefore
result from a similar rate of migration in the two diastereomers of
the ate complex.
The fact that the Me
methyl migration barrier (barriers with R = H and Me
comparable) is unexpected given that Me Si group is reported
2 reaction at an a-carbon. However, an electronic
3
Si group (unlike Ph) has no effect on
3
Si are
3
15
We were intrigued as to why the silyl-stabilised ylide should add
reversibly to the borane whilst the phenyl-stabilised ylide reacts
to activate S
N
analysis reveals the development of a positive charge on the ylidic
carbon at the TS (d charge on the carbon is 0.29 a.u. in non-
substituted case; see Supporting Information). This indicates a
non-reversibly because the silyl-stabilised ylide is expected to be a
5
slightly poorer leaving group (since it has a lower pK
b
) than the
phenyl-stabilised ylide. Clearly, in the case of the silyl-substituted
significant S
N
1 character of the reaction which can explain the
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 1185–1189 | 1187

Scheme 9 Computational analysis of the isodesmic reaction to probe
stability of ate complex.
geminal effects (strong donor and strong acceptor orbitals cause
15
the greatest antibonding geminal destabilising interactions) and
vicinal steric interactions. Stabilisation of the ylide is thus not
counter-balanced by a stabilisation of the ate complex in this case,
which accounts for the decrease in exothermicity of ate complex
formation.
In summary, in the case of phenyl-substituted ylide, a large
stabilisation of the ate complex, which counter-balances stabilisa-
tion of the ylide, along with the activation of the migration step
make the addition onto the organoborane non-reversible. For silyl-
stabilised ylide, a combination of a decrease in exothermicity of
ate complex formation, due to low stabilisation of the ate complex,
and an absence of activation of the migration step, due to a
Fig. 1 Computational potential energy surface (relative energies in kcal
−
1
mol ) for the reaction of Me
3
B with Me
3
SiCHSMe
2
(the dotted curve
significant S 1 character of the 1,2-migration reaction, renders
N
corresponds to qualitative energy profile taking into account the estimated
error in the method).
the addition of the ylide to the borane reversible.
a
Conclusions
Table 2 Influence of substitution on the energy profile
In conclusion, it has been found that in contrast to aryl-stabilised
ylides which react with boranes to give homologated products
with very high enantioselectivity, silyl-stabilised ylides only give
low enantioselectivity. Through cross-over experiments it has been
shown that the two classes of ylides show quite different reactivity
with boranes. Aryl-stabilised ylides react with boranes non-
reversibly to give intermediate ate complexes which undergo stere-
ospecific 1,2-metallate rearrangement leading to homologated bo-
ranes with high enantioselectivity. However, silyl-stabilised ylides
react with boranes reversibly and so now selectivity is determined
by the equilibrium ratio of diastereomeric ate complexes that are
formed and their relative rates of subsequent 1,2-migration. The
origin of the difference in reactivity and change in the selectivity-
determining step has been analysed computationally. The main
factors responsible for the silyl-stabilised ylide reacting reversibly
with the borane are a reduction in exothermicity of ate complex
formation and an increase in the activation barrier to migration
of the borane substituent.
R
Reactants
Ate complex
TS(barrier)
Product
H
Ph
TMS
0.0
0.0
0.0
−34.6
−34.6
−24.3
−19.2(15.4)
−22.1(12.5)
−8.4(15.9)
−57.9
−54.5
−51.4
a
Energies are obtained at the MP2/6-311+G**//B3LYP/6-31G* level of
theory and are given in kcal mol respective to reactants.
−
1
absence of activation by the Me
to have little effect on S 1 reactions ).
3
Si group (Me Si has been shown
3
15c
N
The substantial decrease in exothermicity of ate complex
formation is even more surprising. Indeed, based on the slightly
lower stabilisation of the silyl-stabilised ylide compared to the
phenyl-stabilised ylide, and the similar steric encumbrance of the
two groups (vide supra), one would expect, if anything, the opposite
trend. A detailed analysis (see Supporting Information) shows that
ate complex formation is highly exothermic in the case of phenyl-
stabilised ylide because the phenyl group is able to stabilise the ate
complex (hyperconjugation) to a similar extent as the ylide. This
stabilisation of the ate complex can be illustrated by the energy
of the isodesmic reaction reported in Scheme 9. For the silyl-
substituted ylide the isodesmic reaction indicates a much lower
stabilisation of the ate complex, probably due to destabilising
Computational details
The MP2/6-311+G**//B3LYP/6-31G* method has been se-
lected to be the more adequate for the studied system after
investigation of the model reaction of BMe
variety of different levels of theory, including QCISD(T)/6-31G*
and G3(MP2)//B3LYP methods (see reference 14).
All species have been fully geometry optimised using the Jaguar
6.0 pseudospectral program package. Geometry optimisation
and NBO analysis were carried out using the well established
B3LYP hybrid density functional and the standard split valence
polarised 6-31G* basis set.
3
with CH
2
SMe at a
2
16
1
188 | Org. Biomol. Chem., 2008, 6, 1185–1189
This journal is © The Royal Society of Chemistry 2008

Single point calculations at the MP2/6-311+G** level have been
carried out on every transition state structure using the Gaussian
2202. However, we have addressed this issue computationally through
a series of isodesmic reactions (SI, pp. 7–8) and have found that in the
ate complexes TMS and Ph have similar steric effects.
M. E. Squillacote and J. M. Neth, J. Am. Chem. Soc., 1987, 109, 198.
8 W. Kitching, H. A. Olszowy, G. M. Drew and W. Adcock, J. Org. Chem.,
1982, 47, 5153.
17
03 program package.
7
Frequency calculations have been carried out on every transition
state structure in order to check the correct nature of the point on
the potential energy surface.
9
V. K. Aggarwal, J. Charmant, L. Dudin, M. Porcelloni and J.
Richardson, Proc. Natl. Acad. Sci. U. S. A., 2004, 5467–5471.
1
1
0 E. Vedejs and G. R. Martinez, J. Am. Chem. Soc., 1979, 101, 6452.
1 J. I. Seeman, Chem. Rev., 1983, 83, 83–134.
Acknowledgements
12 D. M. Badine, C. Hebach and V. K. Aggarwal, Chem. Asian J., 2006,
, 438.
13 For calculations on relative stabilisation of a range of sulfonium ylides
RCHSMe , see: V. K. Aggarwal and C. Hebach, Org. Biomol. Chem.,
005, 3, 1419. For electronic effects of silyl groups see:; J. B. Lambert,
1
We thank GSK and EPSRC for support of this work. V. K. A.
thanks the Royal Society for a Wolfson Research Merit Award
and EPSRC for a Senior Research Fellowship. R.R. is a Charg e´ de
recherches FNRS.
2
2
Tetrahedron, 1990, 46, 2677–2689; J. B. Lambert, Y. Zhao, R. W.
Emblidge, L. A. Salvador, X. Liu, J. H. So and E. C. Chelius, Acc.
Chem. Res., 1999, 32, 183–190.
1
4 R. Robiette, G. Y. Fang, J. N. Harvey and V. K. Aggarwal, Chem.
Commun., 2006, 741–743.
References
15 (a) H. K. Oh, C. H. Shin and I. Lee, J. Chem. Soc., Perkin Trans. 2, 1993,
411–2414; (b) Y. Apeloig, R. Biton and A. Abu-Freith, J. Am. Chem.
2
1
(a) V. K. Aggarwal, G. Y. Fang and A. T. Schmidt, J. Am. Chem. Soc.,
005, 127, 1642; (b) G. Y. Fang and V. K. Aggarwal, Angew. Chem.,
Soc., 1993, 115, 2522–2523; (c) P. J. Stang, M. Ladika, Y. Apeloig, A.
Stanger, M. D. Schiavelli and M. R. Hughey, J. Am. Chem. Soc., 1982,
104, 6852–6854.
2
Int. Ed., 2007, 46, 359; (c) G. Y. Fang, O. A. Wallner, N. D. Blasio, X.
Ginesta, J. N. Harvey and V. K. Aggarwal, J. Am. Chem. Soc., 2007,
16 Jaguar 6.0, LLC, Schr o¨ dinger, Portland, Oregon, 2005.
1
29, 14632.
17 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, J. A. Montgomery, Jr. T. Vreven, K. N. Kudin, J. C.
Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci,
M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E.
Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,
C. Pomelli, J. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P.
Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D.
Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S.
Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz,
I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. G. Johnson,
W. Chen, M. W. Wong, C. Gonzalez and J. A. Pople, GAUSSIAN 03
(Revision B.04), Gaussian, Inc., Pittsburgh, PA, 2003.
2
3
D. S. Matteson and D. Majumdar, J. Am. Chem. Soc., 1980, 102, 7588;
P. Kocienski and C. Barber, Pure Appl. Chem., 1990, 62, 1933.
F. Cooke, P. Magnus and G. L. Bundy, J. Chem. Soc., Chem. Commun.,
1
978, 714; C. Fleischmann and E. Zbiral, Tetrahedron, 1978, 34, 317.
4
5
D. J. S. Tsai and D. Matteson, Organometallics, 1983, 2, 236.
Phenyl is usually able to stabilise a carbanion slightly better than Me Si.
3
The following examples of pK
page at http://www.chem.wisc.edu/areas/organic/index-chem.htm,
a
’s illustrate this (taken from Reich’s web-
th
accessed at 25 /April/2007). PhSO
28.7); (EtO) P(O)CH Ph (27.6), (EtO)
FlCHPh (17.9), FlCHSiMe
(21.5). Fl = fluorenyl.
2
CH
2
Ph (23.4), PhSO
2
CH
2
SiMe
(28.7);
3
(
2
2
2
P(O)CH SiMe
2
3
3
6
Although Ph and TMS have similar A values, this doesn’t always result
in the same steric influence due to the shape of the Ph group but they are
nevertheless a reasonable guide. For a discussion of this issue see: N. L.
Allinger and M. T. Tribble, Tetrahedron Lett., 1971, 3259; W. F. Bailey,
H. Connon, E. L. Eliel and K. B. Wiberg, J. Am. Chem. Soc., 1978, 100,
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 1185–1189 | 1189