Intermolecular chirality transfer from silicon to carbon: Interrogation of the two-silicon cycle for Pd-catalyzed hydrosilylation by stereoisotopochemical crossover

Article abstract of DOI:10.1021/ja067780h

A "two-silicon cycle" for the highly efficient intermolecular chirality transfer from silicon to carbon (98-99% ct) in the palladium-catalyzed hydrosilylation of norbornenes emerges from an investigation involving catalytic crossover experiments with isotopically labeled silicon-stereogenic silanes. A key outcome of these experiments, which are supported by product-distribution modeling, is the conclusion that the chirality transfer arises from thermodynamically controlled reversible silapalladation of the alkene rather than from kinetic control during irreversible σ-bond metathesis of the resulting β-silyl σ-alkyl palladium complex with chiral silane. Copyright

Full text of DOI:10.1021/ja067780h

Published on Web 12/22/2006
Intermolecular Chirality Transfer from Silicon to Carbon: Interrogation of the
Two-Silicon Cycle for Pd-Catalyzed Hydrosilylation by Stereoisotopochemical
Crossover
Sebastian Rendler,† Martin Oestreich,*,† Craig P. Butts,‡ and Guy C. Lloyd-Jones*,‡
Organisch-Chemisches Institut, Westfa¨lische Wilhelms-UniVersita¨t Mu¨nster, Corrensstrasse 40, D-48149 Mu¨nster,
Germany, and School of Chemistry, UniVersity of Bristol, Cantock’s Close, Bristol BS8 1TS, United Kingdom
Received October 31, 2006; E-mail: martin.oestreich@uni-muenster.de; guy.lloyd-jones@bristol.ac.uk
Despite long-standing interest in silicon-stereogenic reagents,1
it was only recently that highly efficient intra-2 and intermolecular3
Scheme 1. Chirality Transfer (ct) in a Pd-Catalyzed
Hydrosilylation
silicon-to-carbon chirality transfers were achieved. The inter-
molecular process involves the hydrosilylation of norbornene and
norbornene-derivative 1 by chiral silanes 2, employing cationic Pd
precatalyst 3 (Scheme 1).3 Cyclic silanes such as 2a and 2b emerged
as excellent reagents for this process.3 Although these silanes are
readily prepared in very high ee,4 it is the diastereoisomeric ratio
in the quaternary silane products, for example, 4a or 4b, that reveals
the exceptionally high net chirality transfer (98-99% ct).3 We have
also shown that scalemic samples of 2a (1.2 equiv, 16-65% ee)
give rise to a marked asymmetric amplification: the hydrosilylation
product of norbornene (but not 1) being generated in greater ee
than that of the reagent 2.3
Scheme 2. A Two-Silicon Cycle for Hydrosilylation by 2
The Pd-catalyzed hydrosilylation of alkenes5 and dienes6 by
achiral silanes such as Et3SiH has been studied in great detail by
Brookhart et al.5 and Widenhoefer et al.6 However, many facets of
the stereochemistry in such processes remain to be explored. Herein,
we report on the mechanism of the reaction of 1 with 2a or 2b, on
the elucidation of which step(s) in the catalytic cycle induce the
chirality transfer, and on the origin and attenuation (masking) of
asymmetric amplification.
Applying the Brookhart model5 to the reaction of chiral silanes
2 with 1 yields the catalytic cycle outlined in Scheme 2. Key
features are a silane-mediated precatalyst activation (3 f 5),
followed by a three-step propagation (i, ii, and iii). In such a “two-
silicon cycle”, the chirality transfer from silicon to carbon can occur
at one or both of two stages: the C-Si bond forming migratory
insertion step ii (6 f 7) or the C-H bond forming σ-bond
metathesis step iii (7 f 4). Starting from these considerations, we
have designed experiments that both validate the proposed two-
silicon cycle and reveal whether the highly efficient chirality transfer
arises from step ii or step iii, or both.
experiments, in which conversion (11-38%) and [1]/[Pd] (330-
3300) were varied, were analyzed by 29Si{1H} NMR, exploiting a
combination of 1JSiC and γ-2H isotope shifts.7 Under conditions of
high [1] (biasing k1[1][5a] towards 6a) the scrambling of reagents
[13C]-2a/[2H]-2a could be sufficiently suppressed to allow conclu-
sive analysis. A single model, based on the two-silicon cycle
(Scheme 2), in which a PKIE of kH/kD ) 2.5 attends cleavage of
Si-2H in step iii and in which scrambling is inversely proportional
As shown in step iii of Scheme 2, the R3Si and H moieties of
product 4 are delivered from two separate molecules of 2 via two
sequential turnovers of the catalytic cycle. To test for this, we
reacted 1 with a mixture of [13C]-2a and [2H]-2a. A two-silicon
cycle will give rise to a mixture of all four isotopomers (4a, [13C]-
4a, [2H]-4a, and [2H,13C]-4a) with proportions dependent on the
primary kinetic isotope effect (PKIE) attending step iii. For a one-
silicon cycle only two isotopomers ([13C]-4a and [2H]-4a) should
be obtained, thereby allowing distinction on the basis of isotopic
distributions. However, Brookhart has demonstrated that [Pd]-SiR3
species (cf. 5) undergo rapid exchange of the silyl moiety with
R′3Si-H.5 Such a process will have the net effect of scrambling H
2
to [1], gave an excellent fit for the H/13C distributions in 2a and
4a in all five runs.7 The outcome from one run after 38% conversion
(127 turnovers) is given in Scheme 3. The magnitude of the PKIE
(independently determined as kH/kD ) 3.0 ( 0.5)7 suggests small
transfer angles8 in the transition state for step iii, fully consistent
with a σ-bond metathesis process.5,6
With good evidence for propagation via the two-silicon cycle,
we then tested for chirality transfer in step iii by co-reacting
enantioenriched (SiR)-2b9 (97% ee) with an achiral 2H-labeled
silane, [2H]-8 (Scheme 4). These partners were chosen for their
comparable reactivity, and we again employed high [1] to suppress
1H/2H scrambling (vide supra).5 Chiral HPLC analysis followed
2
and H between [13C]-2a and [2H]-2a, thus precluding distinction
of a one-silicon from a two-silicon cycle. A series of five
† Westfa¨lische Wilhelms-Universita¨t Mu¨nster.
‡ University of Bristol.
9
502
J. AM. CHEM. SOC. 2007, 129, 502-503
10.1021/ja067780h CCC: $37.00 © 2007 American Chemical Society

C O M M U N I C A T I O N S
Scheme 3. Analysis of the Two-Silicon Cycle by 2H/13
C
Scheme 5. Match-Mismatch Effects Detected by Crossovera
Crossovera
obsd % (calcd); by chiral HPLC and MS, k3m/ k3 ) 2.5, PKIE )
a
mm
a
obsd % (calcd) at 38% conversion by 29Si{1H} NMR; 127 turnovers.
2.5.
Scheme 4. “Traceless Chirality Transfer” from 2b to Generate 9
explain the lack of amplification with 1. Conclusive evidence for
substantial match-mismatch with 1 was obtained by reaction of
quasiscalemic silane 2b with excess 1. Analysis by chiral HPLC
followed by MS confirmed that the “same-silane” (matched)
products are prevalent over the crossover (mismatched) products
(Scheme 5). Modeling of the product distributions according to the
mm
two-silicon cycle suggests k3m/k3 ) 2.5.7
In summary, the mechanism of intermolecular chirality transfer
via a two-silicon cycle has been investigated by means of
stereoisotopochemical probes in catalytic crossover experiments.
Three important observations emerge: (1) A thermodynamically
controlled diastereo-discrimination in the migratory insertion (K2
vs K2diast) is the major source of chirality transfer by hydrosilylation
(2 f 4), albeit manifested by the irreversible σ-bond metathesis
reaction (step iii). (2) Under appropriate conditions (scalemic
reagent 2, high K1; low [5]), chirality match-mismatch effects in
step iii (k3m/k3mm) can induce asymmetric amplification. (3) Reaction
of a silicon-stereogenic silane with a racemic but racemizing σ-alkyl
palladium(II) species is able to induce remarkable enantioselectivity
(cf. 9, Scheme 4). These observations bode well for the design of
other processes using silicon-based stereoinducers12 in asymmetric
catalysis.
by MS facilitated the quantification of isotopomeric ratios for all
four stereoisomers of 4b and both enantiomers of 9.7
Two distinct conclusions can be drawn from analysis of the
chirality transfer in the crossover products, (2R)-9 and (SiR)-[2H]-
4b. First, there is significant asymmetric induction (60% ee) in the
generation of 9, which is liberated by reaction of a racemic Pd-
alkyl intermediate (cf. 7) with (SiR)-2b, while the solely [2H]-8
derived product [2H]-9 is essentially racemic.7,10 The “traceless
chirality transfer” exerted in generation of 60% ee for (2R)-9
demonstrates the profound stereochemical impact of a single chiral
silicon moiety in the transition state of step iii of the reaction and
confirms reversible silapalladation of alkene 1 by (()-[Pd]-SiPh2-
Me: in effect there is a dynamic kinetic resolution of the Pd-alkyl
intermediate. Second, (SiR)-[2H]-4b, which is liberated by σ-bond
metathesis of 7b with achiral [2H]-8, is generated with the same
chirality transfer as (SiR)-4b (dr ) 99:1), and thus the chirality
transfer arises from step ii alone.
Acknowledgment. We thank the Dr.-Otto-Ro¨hm-Geda¨chtnis-
stiftung, the Deutsche Forschungsgemeinschaft (Oe 249/2-4), the
Fonds der Chemischen Industrie (S.R.), the Aventis Foundation
(M.O.), and AstraZeneca (G.C.L.-J.) for research support. Gerd
Fehrenbach (HPLC), Dr. Manfred Keller (NMR), and Dr. Ju¨rgen
Wo¨rth (MS) provided expert technical assistance.
Supporting Information Available: Full experimental data. This
material is available free of charge via the Internet at http://pubs.acs.org.
When an excess (1.2 equiv) of scalemic silane 2a is employed,
norbornene gives rise to asymmetric amplification3 (vide supra).
Scheme 2 provides a ready explanation for such a phenomenon:
step iii involves two silicon-stereogenic units (7 and 2), and thus
subtle match (m)-mismatch (mm) effects (k3m/k3mm) can kinetically
select for the major enantiomer of the reagent 2a. Modeling of the
two-silicon cycle suggests that k3m/k3mm ) 3 is sufficiently large to
give rise to the observed asymmetric amplification. Given the
equally high chirality transfer obtained with 1, the lack of
asymmetric amplification with this alkene is beguiling. However,
the model also reveals that the exchange of (SiS)-SiR3 with (SiR)-
SiR3 at the stage of the [Pd]-SiR3 species 5 (vide supra)5 can
completely “mask” the amplification despite substantial match-
mismatch effects (see Supporting Information for a full discussion).
Efficient capture of 5 by alkene, so as to minimize silane exchange,
is thus a prerequisite for efficient asymmetric amplification. As
such, the lower reactivity of 1 as compared to norbornene, would
References
(1) Oestreich, M. Chem.sEur. J. 2006, 12, 30.
(2) Schmidt, D. R.; O’Malley, S. J.; Leighton, J. J. Am. Chem. Soc. 2003,
125, 1190.
(3) Oestreich, M.; Rendler, S. Angew. Chem., Int. Ed. 2005, 44, 1661.
(4) Rendler, S.; Auer, G.; Keller, M.; Oestreich, M. AdV. Synth. Catal. 2006,
348, 1171.
(5) LaPointe, A. M.; Rix, F. C.; Brookhart, M. J. Am. Chem. Soc. 1997, 119,
906.
(6) Perch, N. S.; Widenhoefer, R. A. J. Am. Chem. Soc. 2004, 126, 6332.
(7) See Supporting Information for full details.
(8) More O’Ferrall, R. A. J. Chem. Soc. B 1970, 785.
(9) 2b was used instead of 2a because of the slightly diminished diastereo-
selectivity allowing for the detection and separation of all four possible
stereoisomers.
(10) The nonzero ee in [2H]-9 can be ascribed to the generation of traces of
[2H]-2b through silane exchange.
(11) Zhang, Q.; Curran, D. P. Chem.sEur. J. 2005, 11, 4866.
(12) See for example: Rendler, S.; Auer, G.; Oestreich, M. Angew. Chem.,
Int. Ed. 2005, 44, 7620.
JA067780H
9
J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007 503