Crucial role of the ligand of silyl Lewis acid in the Mukaiyama aldol reaction

Article abstract of DOI:10.1039/b203838b

The Me₃SiX-induced Mukaiyama aldol reaction proceeds through each catalytic cycle under the influence of X-: the silyl group of Me₃SiNTf₂ does not release from -NTf₂ and that of silyl enol ether intermolecularly transfers to the product, while the silyl group of Me₃SiOTf remains in the product and that of the silyl enol ether becomes the catalyst for the next catalytic cycle.

Full text of DOI:10.1039/b203838b

Crucial role of the ligand of silyl Lewis acid in the Mukaiyama aldol
reaction†
Kazuaki Ishihara, Yukihiro Hiraiwa and Hisashi Yamamoto*
Graduate School of Engineering, Nagoya University, SORST, Japan Science and Technology Corporation
(JST), Furo-cho, Chikusa, Nagoya 464-8603, Japan. E-mail: yamamoto@cc.nagoya-u.ac.jp;
Fax: (+81)52-789-3222
Received (in Cambridge, UK) 19th April 2002, Accepted 6th June 2002
First published as an Advance Article on the web 19th June 2002
The Me3SiX-induced Mukaiyama aldol reaction proceeds
through each catalytic cycle under the influence of X–: the
ocarbenium ions 3 and 8 could be much faster than that induced
by Me3SiNTf2. In fact, Me3SiNTf2 is an extraordinarily active
catalyst,1,2 and its high turnover frequency can be reasonably
explained by a +SiR3-induced cascade process which avoids the
regeneration of Me3SiNTf2.
–
silyl group of Me3SiNTf2 does not release from NTf2 and
that of silyl enol ether intermolecularly transfers to the
product, while the silyl group of Me3SiOTf remains in the
product and that of the silyl enol ether becomes the catalyst
for the next catalytic cycle.
Ghosez1a and Mikami1b have independently introduced Me-
3SiNTf2 as a much more powerful carbonyl-activating reagent
than Me3SiOTf. Very recently, we demonstrated the extremely
high activity of Me3SiNTf2 as a catalyst for the Mukaiyama
aldol and Sakurai–Hosomi allylation reactions of not only
aldehydes but also ketones.2,3 We report here that the Me3SiX-
induced Mukaiyama aldol reaction proceeds through each
catalytic cycle under the influence of its ligand (X): specifically,
there is a significant difference between –NTf2 and –OTf, in that
the silyl group of Me3SiNTf2 does not release from its ligand
and that of silyl enol ether intermolecularly transfers to the
product, while the silyl group of Me3SiOTf remains in the
product and that of the silyl enol ether becomes the catalyst for
the next catalytic cycle.4
(1)
(2)
For a crossover experiment,4 two silyl enol ethers 9 and 13
with comparable steric and electronic properties were combined
with benzaldehyde [eqn. (3)]. Benzaldehyde was added to a
solution of a 1+1 mixture of 9 and 13 (0.6 equiv. of each) in the
presence of 1 mol% of Me3SiNTf2 at 278 °C. The product
composition was determined by GC analysis. A mixture of
The three plausible reaction pathways for the Me3SiX-
induced Mukaiyama aldol reaction are outlined in Fig. 1.5 In the
first stage of catalysis, the addition of a trialkylsilyl enol ether 2
to a Me3Si-activated aldehyde 1 generates a siloxocarbenium
ion intermediate 3. In path A, R3SiX is released into the reaction
medium to give the trimethylsilyl aldolate 4 by the intra-
molecular transfer of X–.6 Thus, R3SiX becomes the catalyst for
the next catalytic cycle.7 In path B, Me3SiX is released into the
reaction medium to give the trialkylsilyl aldolate 5 by the
intramolecular transfer of R3Si+.6,8 In path C, co-ordinated
aldehyde as in 6 or 7, which is activated by penta- or tetra-
coordination4,9 of R3Si+, reacts with 2 through the inter-
molecular silyl transfer of R3Si+. In this catalytic process, 5 may
be obtained as a major product.10
First, the mechanistic details of the Me3SiNTf21a,11-induced
reaction were investigated. To clarify whether transformation
from 2 to a silyl aldolate occurs through –NTf2 transfer (path A)
or +SiR3 transfer (path B or C), the reaction of tert-
butyldimethylsilyl enol ether 9 derived from acetophenone with
benzaldehyde was carried out in the presence of 1 equiv. of
Me3SiNTf2 at 2100 °C for 0.5 h. Regardless of the order of the
addition of substrates and Me3SiNTf2, only tert-butyldime-
thylsilyl aldolate 10 was obtained in high yield [eqn. (1)]. In
contrast, the reaction of trimethylsilyl enol ether 11 with
benzaldehyde in the presence of 1 equiv. of t-BuMe2SiNTf2
gave only trimethylsilyl aldolate 12 in moderate yield [eqn. (2)].
Control experiments for eqn. (1) showed that neither 9 and
Me3SiNTf2 nor 12 and t-BuMe2SiNTf2 exchanged under
similar conditions. These experimental results exclude path A.
Although path C is also unlikely since 12 was not detected by
GC analysis [eqn. (1)], we can not exclude path C because the
equilibration between benzaldehyde and 1 may be quite fast,
and the aldol reaction induced by tert-butyldimethylsilox-
† Electronic supplementary information (ESI) available: experimental
section. See http://www.rsc.org/suppdata/cc/b2/b203838b/
Fig. 1 An outline of the three possible mechanisms for the Mukaiyama aldol
reaction induced by Me3SiX.
1564
CHEM. COMMUN., 2002, 1564–1565
This journal is © The Royal Society of Chemistry 2002

scrambled silyl ethers was isolated in a ratio of 27+27+24+22
(10+14+15+16). Control experiments showed that neither the
starting silyl enol ethers nor the silylated aldol products
exchanged under similar reaction conditions. Based on these
observations, path B was unambiguously precluded. Therefore,
it is likely that the Me3SiNTf2-induced reaction occurs through
path C.
of X– from 3 is expected to occur by electrophilic attack of the
‘R3Si–O+ silicon’ of 3. In the Me3SiOTf-induced reaction,
R3SiOTf would be generated by electrophilic attack of the
‘R3Si–O+ silicon’ to the ‘S = O oxygens’ or the ‘S–O oxygen’ of
–OTf (path A). In the Me3SiNTf2-induced reaction, on the
contrary, less nucleophilicity14 and/or more bulkiness of –NTf2
may suppress the electrophilic attack of the ‘R3Si–O+ silicon’ to
–
the nitrogen11 or oxygen atoms13 of NTf2, and may increase
Lewis acidity of siloxocarbenium ions 3 and 8 (path C).
Lastly, Sakurai–Hosomi allylation reaction induced by
+
Me3SiNTf2 also occurs through SiR3 transfer, according to
results of a stoichiometric experiment [eqn. (5)].2
(3)
(5)
The syn/anti selectivities in the aldol reactions of (E)- and
(Z)-silyl enol ethers induced by silyltriflylimides were investi-
gated (Table 1). Similar syn/anti ratios of silyl aldolates, 19 and
20, were obtained independent of the trialkylsilyl group of the
silyltriflylimides. On the other hand, the ratios were correlative
to the trialkylsilyl group of silyl enol ethers. These results also
support path C.
These findings may provide a basis for the future develop-
ment of not only chiral silyl Lewis acid catalysts but also other
chiral metal catalysts for carbon–carbon bond-forming reac-
tions of silyl nucleophiles with carbonyl compounds.15
Next, we performed a mechanistic study of the Me3SiOTf-
induced Mukaiyama aldol reaction.12 Reactions analogous to
eqn. (1) were conducted using Me3SiOTf [eqn. (4)]. A mixture
of 10 and 12 was obtained in 24% yield at a molar ratio of 1+99
under the same conditions with eqn. (1). The starting materials
were not completely consumed upon prolonged stirring at 278
°C with a concentration three-fold higher than that in eqn. (1),
and a mixture of 10 and 12 was obtained in 61% yield at a molar
ratio of 17+83 after 5 h at 278 °C. The reaction of 9 with
benzaldehyde did not proceed at 2100 °C in the presence of t-
BuMe2SiOTf in a control experiment for eqn. (4). Further
control experiments for eqn. (4) showed that neither 9 and
Me3SiOTf nor 10 and Me3SiOTf exchanged under similar
conditions. These experimental results strongly support the
mechanism (path A) proposed by Hollis and Bosnich.4d
Notes and references
1 (a) B. Mathieu and L. Ghosez, Tetrahedron Lett., 1997, 38, 5497; (b) A.
Ishii, O. Kotera, T. Saeki and K. Mikami, Synlett, 1997, 1145.
2 K. Ishihara, Y. Hiraiwa and H. Yamamoto, Synlett, 2001, 1851.
3 (a) For carbon–carbon bond forming reactions catalyzed by HNTf2, see:
N. Kuhnert, J. Peverley and J. Robertson, Tetrahedron Lett., 1998, 39,
3215; (b) J. Cossy, F. Lutz, V. Alauze and C. Meyer, Synlett, 2002,
45.
4 (a) For mechanistic studies on the Mukaiyama aldol reaction induced by
metal triflates, metal perchlorates, or trityl salts, see: E. M. Carreira, in
Comprehensive Asymmetric Catalysis III, eds. E. N. Jacobsen, A. Pfaltz
and H. Yamamoto, Springer-Verlag, Berlin, Heidelberg, 1999, p. 997;
(b) E. M. Carreira and R. A. Singer, Tetrahedron Lett., 1994, 45, 4323;
(c) S. E. Denmark and C.-T. Chen, Tetrahedron Lett., 1994, 35, 4327;
(d) K. T. Hollis and J. Bosnich, J. Am. Chem. Soc., 1995, 117, 4570.
5 There may be a rapid equilibration between 3 and an oxetane
intermediate: W. W. Ellis and B. Bosnich, Chem. Commun., 1998,
193.
(4)
6 The intermolecular transfer of X– or R3Si+ is unlikely due to unfavorable
entropic and steric factors.
7 This means that chiral silyl Lewis acids cannot be used as enantiose-
lective catalysts.
8 The catalytic cycle through path B is essential for the rational design of
chiral silyl Lewis acid catalysts.
The present observations indicate that the ligand of silyl
Lewis acid plays a crucial role in the aldol reaction. The transfer
Table 1 Syn/anti selectivity in the Mukaiyama aldol reaction
9 (a) For the importance of pentacoordinated silicon species as reactive
intermediates, see: R. Damrauer, C. H. DePuy and V. M. Bierbaum,
Organometallics, 1982, 1, 1553; (b) A. R. Bassindale and T. Stout, J.
Organomet. Chem., 1982, 238, C41; (c) J. C. Sheldon, R. N. Hayes and
J. H. Bowie, J. Am. Chem. Soc., 1984, 106, 7711; (d) R. J. P. Corriu, C.
Guérin and J. J. E. Moreau, Top. Stereochem., 1984, 15, 43.
10 It may be relatively difficult to design chiral silyl Lewis acid catalysts
for the catalytic cycle through path C.
11 Me3SiNTf2, which was prepared from Me3SiCl and AgNTf2 in
dichloromethane, was purified by distillation (80–84 °C , 7 torr). Its N-
silyl structure has been determined on the basis of 13C and 29Si NMR
spectra. A. Vij, Y. Y. Zheng, R. L. Kirchmeier and J. M. Shreeve, Inorg.
Chem., 1994, 33, 3281.
12 S. Murata, M. Suzuki and R. Noyori, Tetrahedron, 1988, 44, 4259.
13 Me3SiNTf2 exists in an N-silyl structure while t-BuMe2SiNTf2 and i-
Pr3SiNTf2 exist in an O-silyl structure according to 19F NMR analysis.
G. Simchen and S. Jonas, J. Prakt. Chem., 1998, 340, 506.
14 According to a report about the relative gas-phase acidities of Brønsted
superacids, HNTf2 (Gacid = 291.8 kcal mol21) is a stronger acid than
TfOH (Gacid = 299.5 kcal mol21). Therefore, it is expected that the
nucleophilicity of –NTf2 is lower than that of –OTf by delocalization: I.
A. Koppel, R. W. Taft, F. Anvia, S.-Z. Zhu, L.-Q. Hu, K.-S. Sung, D. D.
DeaMarteau, L. M. Yagupolskii, Y. L. Yagupolskii, N. V. Ignat’ev, N.
V. Kondratenko, A. Y. Volkonskii, V. M. Vlasov, R. Notario and P.-C.
Maria, J. Am. Chem. Soc., 1994, 116, 3047.
Silyl aldolates
Silyl enol ethera Catalyst
Yield (%)
Syn+Anti
17a
17a
17b
17b
18a
18a
18b
18b
a
Me3SiNTf2
19a, 97
t-BuMe2SiNTf2 19a, > 99
60+40
62+38
52+48
51+49
Me3SiNTf2
t-BuMe2SiNTf2 19b, > 99
Me3SiNTf2
19b, > 99
20a, 90 (1)b[9]c 61+39 (48+52)b
t-BuMe2SiNTf2 20a, 89 (2)b[12]c 58+42 (45+55)b
Me3SiNTf2
20b, > 99
t-BuMe2SiNTf2 20b, > 99
87+13
87+13
b Data of alcohol adducts are given in parentheses. c Yields of dimeric ethers
of alcohols are given in brackets.2
15 In this paper, diethyl ether was used uniformly as a solvent in order to
make it easy to control the catalytic activity of silyl Lewis acids.
CHEM. COMMUN., 2002, 1564–1565
1565