Lewis base activation of Lewis acids: Catalytic, enantioselective addition of silyl ketene acetals to aldehydes

Article abstract of DOI:10.1021/ja047339w

The concept of Lewis base activation of Lewis acids has been reduced to practice for catalysis of the aldol reaction of silyl ketene acetals and silyl dienol ethers with aldehydes. The weakly acidic species, silicon tetrachloride (SiCl₄), can b

Full text of DOI:10.1021/ja047339w

Published on Web 02/23/2005
Lewis Base Activation of Lewis Acids: Catalytic,
Enantioselective Addition of Silyl Ketene Acetals to
Aldehydes
Scott E. Denmark,* Gregory L. Beutner, Thomas Wynn, and Martin D. Eastgate
Contribution from the Roger Adams Laboratory, Department of Chemistry, UniVersity of Illinois,
Urbana, Illinois 61801
Received May 6, 2004; E-mail: denmark@scs.uiuc.edu
Abstract: The concept of Lewis base activation of Lewis acids has been reduced to practice for catalysis
of the aldol reaction of silyl ketene acetals and silyl dienol ethers with aldehydes. The weakly acidic species,
silicon tetrachloride (SiCl4), can be activated by binding of a strongly Lewis basic chiral phosphoramide,
leading to in situ formation of a chiral Lewis acid. This species has proven to be a competent catalyst for
the aldol addition of acetate-, propanoate-, and isobutyrate-derived silyl ketene acetals to conjugated and
nonconjugated aldehydes. Furthermore, vinylogous aldol reactions of silyl dienol ethers are also
demonstrated. The high levels of regio-, anti diastereo-, and enantioselectivity observed in these reactions
can be rationalized through consideration of an open transition structure where steric interactions between
the silyl cation complex and the approaching nucleophile are dominant.
Introduction
The appeal of Lewis acid catalysis as a platform for the
development of asymmetric transformations can be attributed
to the fact that the electron-deficient metal center retains its
electrophilicity regardless of the identity or structure of its
ligands.1 The exchange of simple halide or alkoxide ligands for
enantiopure alcohols or amines provides easy access to well-
defined metal species with reduced electrophilicity and increased
steric demands when compared to the parent metal complexes.
The fundamental mechanism by which a Lewis acid promotes
reaction at an organic functional group is through electrophilic
activation.2 The binding of an electron-deficient metal complex
to the nonbonding lone pair of some functional group polarizes
the adjacent bonds, activating them toward nucleophilic attack
(Figure 1). In this context, Lewis acid activation has been
applied to the reactions of a number of functional groups,
including carbonyl compounds, imines, and epoxides.3 Save for
cases of Lewis acid-catalyzed pericyclic reactions, the coordina-
tion complex is transformed into a covalent complex upon
formation of the new bond.
Figure 1. General picture of Lewis acid-catalyzed processes.
required for this step in the catalytic cycle is present either as
part of the solvent or as part of the nucleophile, as in the case
of latent silylated nucleophiles. Performing Lewis acid-catalyzed
reactions in either highly polar or protic solvents is a simple
method for promoting catalyst turnover. Protonation of the
product complex, as demonstrated in the hydrolytic kinetic
resolution of epoxides, is effective for release of the cobalt-
(salen) complex from the ring-opened product.4 Similarly, the
This change in the nature of the bond between the catalyst
and substrate along the reaction coordinate introduces the major
challenge facing Lewis acid-catalyzed processes, namely,
catalyst turnover. Because the strength of the metal-substrate
bond increases upon changing from the substrate to the product
complex, release of the catalyst must be assisted by a subsequent
bond formation (Figure 1). In most successful cases, the reagent
(1) Yamamoto, H. Lewis Acids in Organic Synthesis; Wiley-VCH: Weinheim,
Germany, 2001.
(2) Santelli, M.; Pons, J.-M. Lewis Acids and SelectiVity in Organic Synthesis;
CRC Press: Boca Raton, FL, 1996.
(4) Schaus, S. E.; Brandes, B. D.; Larrow, J. F.; Tokunaga, M.; Hansen, K.
B.; Gould, A. E.; Furrow, M. E.; Jacobsen, E. N. J. Am. Chem. Soc. 2002,
124, 1307-1315.
(3) Otera, J. Modern Carbonyl Chemistry; Wiley-VCH: Weinheim, Germany,
2000.
9
3774
J. AM. CHEM. SOC. 2005, 127, 3774-3789
10.1021/ja047339w CCC: $30.25 © 2005 American Chemical Society

Lewis Base Activation of Lewis Acids
A R T I C L E S
Background
use of propionitrile increases the rate of catalyst turnover in
the boryl complex-catalyzed aldol reaction of silyl ketene
acetals.5
1. Lewis Base Activation of Lewis Acids. The need for
preformation of a chiral Lewis acidic catalyst is linked to the
inherent Lewis acidity of electron-deficient metal centers.
Changes in the ligands can have a strong, but often unpredictable
effect on the electrophilicity of the metal center. Therefore, care
must be taken to ensure that the parent metal species is removed
from the prepared catalyst complex to avoid reaction catalyzed
by this species. Development of a Lewis acid-catalyzed al-
dolization that eliminates this concern could take a number of
forms: (1) preformation and purification of the chiral metal-
ligand complex, (2) highly favorable equilibrium formation of
the metal-ligand complex, or (3) binding of the ligand in such
a way as to significantly enhance the reactivity of the metal
center. In the latter case, the contribution of the nascent Lewis
acidic precursor to the overall reaction rate could be minimized.
In the case of chiral ligands, this latter strategy would increase
the observed rate of reaction through the chiral metal complex
relative to the achiral one and presents an attractive possibility
for the design of a novel catalytic method.
In the case of silylated nucleophiles, such as the silyl ketene
acetals mentioned above, it is silylation of the product aldolate
that actually completes the turnover event in the catalytic cycle.
Several studies highlight the importance of this step, especially
when considering the development of asymmetric methods. If
release of the catalyst by silylation is slow compared to
intervention of a silyl cation-catalyzed process, the overall
mechanism of the reaction shifts from being one that is chiral
metal complex-catalyzed to one that is chiral metal complex-
initiated (Scheme 1). The titanium(IV) Schiff base complex
developed by Carreira incorporates an achiral salicylate ligand
since it can act as a silyl shuttle in the catalyst turnover step.6
Alternatively, the use of additives can minimize catalyst
inhibition or the involvement of a competitive, achiral silyl
cation promoted pathway, such as in the use of trimethyl-
silyl triflate in the silyl ketene acetal aldol reactions developed
by Evans and co-workers.7
In the binding of a Lewis basic ligand to a Lewis acid, there
is a net flow of electron density from the donor to the acceptor
moiety. Therefore, the reactivity of the electrophilic acceptor
moiety decreases, and this should preclude any kind of ligand-
accelerated catalysis.9 However, only considering the changes
in the overall electron density upon adduct formation is a
simplification of the true affects of a Lewis acid-base interac-
tion, as pointed out by Gutmann.10 Although the overall electron
density on the acceptor may increase, this electron density is
not distributed evenly among its constituent atoms. Gutmann
has shown by careful examination of the X-ray crystal structures
of antimony pentachloride, SbCl5 (1), and its complex with
tetrachloroethylene carbonate (2) that significant changes in
bond lengths occur upon complexation (Figure 2).10a Gutmann
proposed that the alternating pattern of these bond changes is
indicative of changes in electron density throughout the complex
that compensate for changes at the donor and acceptor atoms.
The build-up of electron density in the acceptor fragment leads
to a lengthening of the peripheral ligand-metal bonds that
renders the metal center of the complex more electropositive.
Scheme 1
The aldol addition of ester-, ketone-, and aldehyde-derived
enolates is a major focus in asymmetric catalysis, due to the
potential to evaluate mechanisms for control of diastereo- and
enantioselectivity in this single carbon-carbon bond-forming
reaction.8 Advances in the aldol reaction have always been tied
to conceptual advances, the most recent and widespread being
the arrival of Lewis acid catalysis. This report details a unique
method of catalysis that capitalizes on the generality of Lewis
acid catalysis, yet avoids the problems of that strategy, namely,
the requirement for preformation of the chiral Lewis acid
complex and a mechanism for catalyst turnover. The application
of this new method to the aldol addition of silyl ketene acetals
to aldehydes will be disclosed in full.
Figure 2. Effects of Lewis acid-base interactions on bond lengths and
electron density.
Support for this conclusion can be derived from calculations
performed with relevant Lewis acid-base adducts of silicon
tetrachloride (Scheme 2). Gordon and co-workers have studied
the binding of chloride ion to SiCl4 (3) to form penta- and
hexacoordinate silicates (4 and 5) at the 6-311++G(d,p) level
of theory and observed changes in bond lengths and electron
densities consistent with the Gutmann analysis.11 Binding of
(5) Parmee, E. R.; Tempkin, O.; Masamune, S.; Abiko, A. J. Am. Chem. Soc.
1991, 113, 9365-9366.
(6) (a) Carreira, E. M.; Singer, R. A. Tetrahedron Lett. 1994, 35, 4323-4326.
(b) Hollis, T. K.; Bosnich, B. J. Am. Chem. Soc. 1995, 117, 4570-4581.
(7) (a) Evans, D. A.; Kozlowski, M. C.; Burgey, C. S.; Macmillan, D. W. C.
J. Am. Chem. Soc. 1997, 119, 7893-7894. (b) Evans, D. A.; Tedrow, J.
S.; Shaw, J. T.; Downey, C. W. J. Am. Chem. Soc. 2002, 124, 392-393.
(c) Evans, D. A.; Downey, C. W.; Shaw, J. T.; Tedrow, J. S. Org. Lett.
2002, 4, 1127-1130.
(8) (a) Carreira, E. M. In ComprehensiVe Asymmetric Catalysis; Jacobsen, E.
N., Pflatz, A., Yamamoto, H., Eds.; Springer-Verlag: Heidelberg, Germany,
1999; Vol. III, Chapter 29. (b) Nelson, S. G. Tetrahedron: Asymmetry
1998, 9, 357-384. (c) Cowden, C. J.; Paterson, I. Org. React. 1997, 51,
1-200. (d) Mukaiyama, T.; Kobayashi, S. Org. React. 1994, 46, 1-104.
(e) Mukaiyama, T. Org. React. 1982, 28, 203-332. (f) Nielsen, A. T.;
Houlihan, W. J. Org. React. 1968, 16, 1-438.
(9) Berrisford, D. J.; Bolm, C.; Sharpless, K. B. Angew. Chem., Int. Ed. Engl.
1995, 34, 1059-1070.
(10) (a) Gutmann, V. The Donor-Acceptor Approach to Molecular Interactions;
Plenum Press: New York, 1978. (b) Jensen, W. B. The Lewis Acid-Base
Concept; Wiley-Interscience: New York, 1980.
9
J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005 3775

A R T I C L E S
Denmark et al.
the first chloride ion is exothermic by 40.8 kcal/mol, but more
interestingly leads to an increase in the partial positive charge
at silicon by +0.051. A corresponding increase in the partial
negative charge at the chlorine atoms accompanies this change.
A greater degree of the negative charge accumulates at the axial
chlorine atoms when compared to the equatorial chlorine atoms
due to their involvement in a hypervalent three-center/four-
electron bond. Binding of the second chloride ion, although now
an endothermic process by 48.3 kcal/mol, further accentuates
this polarization as the partial positive charge at silicon increases
by another +0.310. Similar trends have been observed by
Sakurai and co-workers in their computational studies of
allyltrifluorosilanes and the corresponding fluorosilicates.12 If
enhanced electrophilicity at the metal center can be achieved
upon simple coordination of a Lewis acid with a Lewis base, is
this change kinetically significant? Only a few examples exist
where simple coordination of a neutral donor to a metal species
leads to enhanced reactivity.13
opening of meso-epoxides reported from this laboratory in
1998.14 In the presence of 10 mol % of the chiral phosphoramide
and a stoichiometric amount of SiCl4, high yields and moderate
enantioselectivities were obtained for the formation of vicinal
chlorohydrins.
Although the desymmetrization of meso-epoxides is an
important synthetic process,15 the main objective was the
extension of this in situ-generated, chiral Lewis acid concept
to the addition of main group organometallic nucleophiles to
carbonyl compounds in analogy to earlier studies with trichlo-
rosilyl nucleophiles.16 Initial studies focused on the addition of
tri-n-butylallylstannane to aldehydes.17 High yields and selec-
tivities are observed under the influence of the chiral, dimeric
phosphoramide.18
The extension of this phosphoramide-catalyzed/SiCl4-medi-
ated system to other nucleophiles was the next logical step.
Considering the nucleophilicity scales developed by Mayr and
co-workers,19 it was believed that more reactive trialkylsilyl enol
ethers derived from esters could probe the limits of the
asymmetric environment imposed by the catalyst. Issues of both
diastereo- and enantioselectivity could be assessed in the
reactions of acetate-, propanoate-, and isobutyrate-derived silyl
ketene acetals. Earlier studies from these laboratories had shown
that trichlorosilyl ketene acetals gave the aldol adducts derived
from aldehydes in low enantioselectivities, most likely because
of the hyper-reactivity of these species.20
Scheme 2
2. Catalytic Asymmetric Aldol Reactions. Despite the
intense interest and notable successes in catalytic asymmetric
aldol reactions of silyl ketene acetals, several limitations are
still apparent: (1) a lack of anti diastereoselective methods for
R-substituted silyl ketene acetals, (2) a low degree of stereo-
convergence with respect to silyl ketene acetal geometry, and
(3) poor substrate scope in vinylogous aldol reactions of silyl
dienol ethers.
Most Lewis acid-catalyzed aldol reactions of silyl ketene
acetals are highly syn diastereoselective.21 This syn diastereo-
selectivity has been a commonly observed theme throughout
the development of these methods, beginning with the early
Taken to the extreme, polarization of the adjacent bonds in
the metal fragment of the adduct leads to ionization of one of
the X ligands and generation of a cationic metal center (Scheme
3). A poorly Lewis acidic species, incapable of promoting a
given reaction by itself, could be ionized upon binding of a
chiral Lewis base, thus generating a cationic metal species with
enhanced activity. This process would represent a unique
direction in the development of Lewis acid catalysts because it
not only addresses the problem of catalyst preformation but also
holds potential for solving the problem of catalyst turnover. The
use of a chiral ligand that is not covalently bound to the
electrophilic metal species shifts the requirements of the turnover
event from cleavage of the strong metal-oxygen bond in the
product to dissociation of the ligand from the metal center.
Another important difference compared to traditional Lewis acid
catalysis is that the stoichiometry of the reaction changes from
being catalytic in Lewis acid to stoichiometric in Lewis acid.
Following this scheme, in each catalyst cycle, the Lewis acidic
fragment would become incorporated into the product. There-
fore, such a reaction would be Lewis base-catalyzed and Lewis
acid-mediated.
(11) (a) Gordon, M. S.; Carroll, M. T.; Davis, L. P.; Burggraf, L. W. J. Phys.
Chem. 1990, 94, 8125-8128. (b) Gordon, M. S. Private communication.
(12) Kira, M.; Sato, K.; Sakurai, H.; Hada, M.; Izawa, M.; Ushio, J. Chem.
Lett. 1991, 387-390.
(13) (a) Nelson, S. G.; Wan, Z. Org. Lett. 2000, 2, 1883-1886. (b) Nelson, S.
G.; Kim, B.-K.; Peelen, T. J. J. Am. Chem. Soc. 2000, 122, 9318-9319.
(c) Evans, D. A.; Johnson, J. S. Org. Lett. 1999, 1, 595-598.
(14) (a) Denmark, S. E.; Barsanti, P. A.; Wong, K.-T.; Stavenger, R. A. J. Org.
Chem. 1998, 63, 2428-2429. (b) Nakajima, M.; Saito, M.; Uemura, M.;
Hashimoto, S. Tetrahedron Lett. 2002, 43, 8827-8829. (c) Tao, B.; Lo,
M. M.-C.; Fu, G. C. J. Am. Chem. Soc. 2001, 123, 353-354.
(15) (a) Hoffman, R. W. Angew. Chem., Int. Ed. 2003, 42, 1096-1109. (b)
Jacobsen, E. N.; Wu, M. H. In ComprehensiVe Asymmetric Catalysis;
Jacobsen, E. N., Pflatz, A., Yamamoto, H., Eds.; Springer-Verlag: Heidel-
berg, Germany, 1999; Vol. III, Chapter 35.1.
(16) (a) Denmark, S. E.; Fujimori, S. In Modern Aldol Reactions; Mahrwald,
R., Ed.; Wiley-VCH: Weinheim, Germany, 2004; Chapter 7, p 229. (b)
Denmark, S. E.; Stavenger, R. A. Acc. Chem. Res. 2000, 33, 432-440.
(17) Denmark, S. E.; Wynn, T. J. Am. Chem. Soc. 2001, 123, 6199-6200.
(18) This reaction is mechanistically distinct from those of allyltrichlorosilane:
(a) Denmark, S. E.; Fu, J. Chem. Commun. 2003, 167-170. (b) Denmark,
S. E.; Fu, J. Org. Lett. 2002, 4, 1951-1953. (c) Denmark, S. E.; Fu, J. J.
Am. Chem. Soc. 2001, 123, 9488-9489. (d) Iseki, K.; Mizuno, S.; Kuroki,
Y.; Kobayashi, Y. Tetrahderon 1999, 55, 977-988. (e) Nakajima, M.; Saito,
M.; Shiro, M.; Hashimoto, S. J. Am. Chem. Soc. 1998, 120, 6419-6420.
(f) Denmark, S. E.; Coe, D. M.; Pratt, N. E.; Griedel, B. D. J. Org. Chem.
1994, 59, 6161-6163.
Scheme 3
(19) (a) Mayr, H.; Bug, T.; Gotta, M. F.; Hering, N.; Irrang, B.; Janker, B.;
Kempf, B.; Loos, R.; Ofial, A. R.; Remennikov, G.; Schimmel, H. J. Am.
Chem. Soc. 2001, 123, 9500-9512. (b) Burfeindt, J.; Patz, M.; Mu¨ller,
M.; Mayr, H. J. Am. Chem. Soc. 1998, 120, 3629-3634.
The potential for development of a chiral Lewis base-
catalyzed/Lewis acid-mediated process was first demonstrated
in the phosphoramide-catalyzed/SiCl4-mediated asymmetric ring
(20) Denmark, S. E.; Winter, S. B. D.; Su, X.; Wong, K.-T. J. Am. Chem. Soc.
1996, 118, 7404-7405.
9
3776 J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005

Lewis Base Activation of Lewis Acids
A R T I C L E S
studies by Mukaiyama22 and continuing into the studies of the
highly selective catalysts developed by Masamune.23 These
studies show that the levels of diastereo- and enantioselectivity
observed in the reactions of E- and Z-isomers of the silyl ketene
acetals are often not comparable. Furthermore, only a few
isolated examples of anti diastereoselectivity have been re-
ported.24,25
Scheme 4
Another interesting class of silyl ketene acetals that have
received only limited attention are silyl dienol ethers.26 The
vinylogous aldol reaction is important because it provides rapid
access to larger fragments useful for the synthesis of polypro-
pionate-derived natural products. This type of aldol reaction
introduces the added challenge of site selectivity along with
the issues of diastereo- and enantioselectivity. In contrast to the
vinylogous aldol reaction of metallodienolates, the reactions of
silyl dienol ethers occur preferentially at the γ-carbon because
these reactions typically operate under frontier molecular orbital
control (Figure 3).27 Most asymmetric methods provide high
regio- and enantioselectivities only for a limited group of
lactone- and dioxanone-derived silyl dienol ethers.28 The reac-
tions of simple ester-derived silyl dienol ethers are limited, and
products are only obtained with moderate enantioselectivity.29
an investigation of the reaction with respect to aldehyde structure
will be undertaken to illustrate substrate scope and functional
group compatibility.30
Results
1. Acetate-Derived Silyl Ketene Acetal Additions. The
promising results from the addition of the weakly nucleophilic
tri-n-butylallylstannane (N ) 5.82)19 to aldehydes promoted by
the phosphoramide/SiCl4 catalyst system suggested that the more
nucleophilic silyl ketene acetals would also prove to be viable
substrates (N ) 9.49 for 3).19 However, the high reactivity of
these species raised immediate concerns about the possibilities
of (1) competitive background reaction, (2) silyl metathesis to
trichlorosilyl enol ethers, and (3) very early transition structures,
all of which would lead to reduced levels of selectivity. Initial
experiments with the commercially available silyl ketene acetal,
1-methoxy-1-tert-butyldimethylsilyloxyethene (1a), were con-
ducted using the optimal conditions developed previously for
the allylation reaction.17 The bulky tert-butyldimethylsilyl ketene
Figure 3. Site selectivity of vinylogous aldol reactions.
Thus, despite the enormous number of asymmetric aldol
reactions of silyl ketene acetals on record, we felt that a
comparison of the phosphoramide/SiCl4 catalyst system with
existing methods would provide an interesting test for the
concept and may also address some of the remaining synthetic
challenges in this field. The primary objective of this study was
to assay the catalyst system in the reactions of three classes of
ester-derived silyl ketene acetals (Scheme 4). Initial studies
would focus on the reactions of acetate-derived silyl ketene
acetals to establish proof of principle and to elucidate the
reactivity and enantioselectivity of this catalyst system. The
survey would then be expanded to include propanoate-derived
silyl ketene acetals and silyl dienol ethers to investigate
additional levels of site and stereochemical complexity. Finally,
(28) For boron, see: (a) Kiyooka, S.-I.; Hena, M. A.; Yabukami, T.; Murai, K.;
Goto, F. Tetrahedron Lett. 2000, 41, 7511-7516. (b) Sato, M.; Sunami,
S.; Sugita, Y.; Kaneko, C. Chem. Pharm. Bull. 1994, 42, 839-845. For
titanium, see: (c) Paterson, I.; Davies, R. D. M.; Heimann, A. C.; Marquez,
R.; Meyer, A. Org. Lett. 2003, 5, 4477-4480. (d) DeRosa, M.; Acocella,
M. R.; Villano, R.; Soriente, A.; Scettri, A. Tetrahedron: Asymmetry 2003,
14, 2499-2502. (e) DeRosa, M.; Acocella, M. R.; Soriente, A.; Scettri, A.
Tetrahedron: Asymmetry 2001, 12, 1529-1531. (f) Soriente, A.; DeRosa,
M.; Stanzione, M.; Villano, R.; Scettri, A. Tetrahedron: Asymmetry 2001,
12, 959-963. (g) DeRosa, M.; Soriente, A.; Scettri, A. Tetrahedron:
Asymmetry 2000, 11, 3187-3195. (h) Szlosek, M.; Figadere, B. Angew.
Chem., Int. Ed. 2000, 39, 1799-1801. (i) Soriente, A.; DeRosa, M.;
Apicella, A.; Scettri, A.; Sodano, G. Tetrahedron: Asymmetry 1999, 10,
4481-4484. (j) Anne, S.; Yong, W.; Vandewalle, M. Synlett 1999, 1435-
1437. (k) Kim, Y.; Singer, R. A.; Carriera, E. M. Angew. Chem., Int. Ed.
1998, 37, 1261-1263. (l) Pichon, M.; Jullian, J.-C.; Figadere, B.; Cave,
A. Tetrahedron Lett. 1998, 39, 1755-1758. (m) Sato, M.; Sunami, S.;
Sugita, Y.; Kaneko, C. Heterocycles 1995, 41, 1435-1444. For copper,
see: (n) Fettes, A.; Carreira, E. M. Angew. Chem., Int. Ed. 2002, 41, 4098-
4101. (o) Evans, D. A.; Kozlowski, M. C.; Murry, J. A.; Burgey, C. S.;
Campos, K. R.; Connell, B. T.; Staples, R. J. J. Am. Chem. Soc. 1999,
121, 669-685. (p) Pagenkopf, B. L.; Kru¨ger, J.; Stojanovic, A.; Carreira,
E. M. Angew. Chem., Int. Ed. 1998, 37, 3124-3126. (q) Kru¨ger, J.; Carreira,
E. M. Tetrahedron Lett. 1998, 39, 7013-7016. (r) Kru¨ger, J.; Carreira, E.
M. J. Am. Chem. Soc. 1998, 120, 837-838. (s) Singer, R. A.; Carreira, E.
M. J. Am. Chem. Soc. 1995, 117, 12360-12361. For chromium, see: (t)
Shimada, Y.; Matsuoka, Y.; Irie, R.; Katsuki, T. Synlett 2004, 57-60. (u)
Onitsuka, S.; Matsuoka, Y.; Irie, R.; Katsuki, T. Chem. Lett. 2003, 32,
974-975. (v) Matsuoka, Y.; Irie, R.; Katsuki, T. Chem. Lett. 2003, 32,
584-585.
(21) (a) Mukaiyama, T.; Inubushi, A.; Suda, S.; Hara, R.; Kobayashi, S. Chem.
Lett. 1990, 1015-1018. (b) Carreira, E. M.; Singer, R. A.; Lee, W. J. Am.
Chem. Soc. 1994, 116, 8837-8838. (c) Evans, D. A.; Murry, J. A.;
Kozlowski, M. C. J. Am. Chem. Soc. 1996, 118, 5814-5815.
(22) (a) Kobayashi, S.; Furuya, M.; Ohtsubo, A.; Mukaiyama, T. Tetrahedron:
Asymmetry 1991, 2, 635-638. (b) Kobayashi, S.; Fujishita, Y.; Mukaiyama,
T. Chem Lett. 1990, 1455-1458.
(23) (a) Masamune, S.; Sato, T.; Kim, B.; Wollman, T. A. J. Am. Chem. Soc.
1986, 108, 8279-8281. (b) Parmee, E. R.; Tempkin, O.; Masamune, S.;
Abiko, A. J. Am. Chem. Soc. 1991, 113, 9365-9366. (c) Parmee, E. R.;
Hong, Y. P.; Tempkin, O.; Masamune, S. Tetrahedron Lett. 1992, 33,
1729-1732.
(24) (a) Mikami, K.; Matsukawa, S.; Sawa, E.; Harada, A.; Koga, N. Tetrahedron
Lett. 1997, 38, 1951-1954. (b) Evans, D. A.; MacMillan, D. W. C.;
Campos, K. R. J. Am. Chem. Soc. 1997, 119, 10859-10860. (c) Mikami,
K.; Matsukawa, S. J. Am. Chem. Soc. 1993, 115, 7039-7040.
(25) The zirconium(IV) BINOL system achieves high anti diastereoselectivity
with a synthetically useful degree of stereoconvergence for a wide variety
of substrates: (a) Yamashita, Y.; Ishitani, H.; Shimizu, H.; Kobayashi, S.
J. Am. Chem. Soc. 2002, 124, 3292-3302. (b) Ishitani, H.; Yamashita, Y.;
Shimizu, H.; Kobayashi, S. J. Am. Chem. Soc. 2000, 122, 5403-5404.
(26) Casiraghi, G.; Zanardi, F.; Appendino, G.; Rassu, G. Chem. ReV. 2000,
100, 1929-1972.
(29) (a) Bluet, G.; Campagne, J.-M. J. Org. Chem. 2001, 66, 4293-4298. (b)
Brennan, C. J.; Campange, J.-M. Tetrahedron Lett. 2001, 42, 5195-5197.
(c) Bluet, G.; Bazan-Tajeda, B.; Campange, J.-M. Org. Lett. 2001, 3, 3807-
3810. (d) Bluet, G.; Campagne, J.-M. Tetrahedron Lett. 1999, 40, 5507-
5509.
(30) Portions of this work have been communicated: (a) Denmark, S. E.; Wynn,
T.; Beutner, G. L. J. Am. Chem. Soc. 2002, 124, 13405-13407. (b)
Denmark, S. E.; Beutner, G. L. J. Am. Chem. Soc. 2003, 125, 7800-7801.
(27) (a) Fleming, I. Frontier Orbitals and Organic Reactions; Wiley-Inter-
science: New York, 1996. For a notable exception, see: (b) Saito, S.;
Nagahara, T.; Shiozawa, M.; Nakadai, M.; Yamamoto, H. J. Am. Chem.
Soc. 2003, 125, 6200-6210.
9
J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005 3777

A R T I C L E S
Denmark et al.
acetal provided ready availability and ease of handling compared
to the more commonly used trimethylsilyl analogue. In the
addition of 1a to benzaldehyde (5a) promoted by 5 mol % of
the dimeric phosphoramide (R,R)-6, an extremely rapid reaction
was observed at -78 °C (<30 s), consistent with the high
reactivity expected for this species. Despite our initial concerns,
the desired product was obtained in high yield and enantioselec-
tivity (Table 1, entry 1). Determination of the absolute configura-
tion of the aldol adduct 7aa by comparison of its optical rotation
to literature values revealed that the 3R-isomer had been form-
ed.31 The sense of selectivity observed in this reaction was the
same as that observed in the allylation reaction, clearly indicating
that the catalyst complex enforced the same sense of facial
differentiation at the aldehyde. Throughout these studies, the
use of a slight excess of the silyl ketene acetal (1.2 equiv) and
SiCl4 (1.1 equiv) aided the reproducibility of the reaction, as
was the case in earlier studies of the allylation. The requirement
for an excess of these reagents is likely related to a small amount
of unproductive trans-silylation that occurs upon addition of the
silyl ketene acetal to the reaction mixture (vide infra).
of aromatic aldehydes, such as 4-methoxybenzaldehyde (5e) and
2-furaldehyde (5i), preserved the high selectivity of the reaction
(entries 5 and 9). Substrates at the opposite end of the electronic
spectrum, such as 4-trifluoromethylbenzaldehyde (5f), also
afforded high yield and selectivity (entry 6). Gratifyingly,
olefinic aldehydes gave high yields although subtle effects on
selectivity became apparent when comparing cinnamaldehyde
(5g) and R-methylcinnamaldehyde (5h) (entries 7 and 8). Much
to our delight, the two aliphatic aldehydes investigated in the
initial survey, hydrocinnamaldehyde (5k) and cyclohexanecar-
boxaldehyde (5j), gave the â-hydroxy ester products in high
yields and enantioselectivities (entries 10 and 11). Unfortunately,
their rate of reaction was greatly attenuated when compared to
that of either benzaldehyde or cinnamaldehyde. ReactIR moni-
toring of the reaction with benzaldehyde revealed that it required
less than 30 s to go to completion, whereas the reaction of
aliphatic aldehydes required close to 3 h.
Other acetate-derived silyl ketene acetals were briefly inves-
tigated to assay the effect of the steric demands of the alkoxy
group on enantioselectivity (Table 2). As expected from a sys-
tematic study conducted with propanoate-derived silyl ketene
acetals (vide infra), the tert-butyl acetate-derived tert-butyldi-
methylsilyl ketene acetal (1b) provided the corresponding â-hy-
droxy ester products with higher levels of enantioselectivity
when compared to 1a. For each of the problematic aldehydes
examined in the initial survey (Table 1), 1-naphthaldehyde (1b),
R-methylcinnamaldehyde (5h), and hydrocinnamaldehyde (5k),
there were considerable improvements. Although these reactions
clearly proceeded to completion based on consumption of the
aldehyde, the isolated yields of the products were found to be
consistently lower than those observed in the additions of 1a.
Careful analysis of the crude reaction mixtures revealed that
varying amounts of the tert-butyldimethylsilyl ester 8 had been
formed. The ratio of the esters was quite sensitive to the structure
of the aldehyde. Cleavage of this silyl ester during the typical
KF/KH2PO4 workup led to loss of the corresponding carboxylic
acid in the aqueous phase, accounting for the lower yield. At-
tempts to modify the reaction and workup conditions were un-
able to suppress this undesired side reaction, suggesting it occur-
red during the reaction. The use of a mild, sodium bicarbonate
workup allowed for isolation of a mixture of both the tert-butyl
and tert-butyldimethylsilyl esters in a high overall conversion.
Table 1. Aldol Reaction with
1-(tert-Butyldimethylsilyloxy)-1-methoxyethene (1a)a
entry
R
product
yield, %b
erc
1
2
3
4
5
6
7
8
C6H5 (5a)
1-naphthyl (5b)
2-naphthyl (5c)
4-CH3C6H4 (5d)
4-CH3OC6H4 (5e)
4-CF3C6H4 (5f)
7aa
7ab
7ac
7ad
7ae
7af
7ag
7ah
7ai
97e,f
98
98
97
97
96.3:3.7
90.2:9.8
96.8:3.2
97.2:2.8
98.7:1.3
95.7:4.3
96.9:3.1
72.7:27.3
93.4:6.6
94.2:5.8
90.7:9.3
97
(E)-C6H5CHdCH (5g)
(E)-C6H5CHdC(CH3) (5h)
2-furyl (5i)
95e
98
9
10
11
94
cyclohexyl (5j)d
PhCH2CH2 (5k)d
7aj
7ak
86e
72e
a All reactions employed 1.1 equiv of SiCl4, 1.2 equiv of 1a, and 0.05
equiv of (R,R)-6 at 0.2 M in CH2Cl2 at -78 °C for 15 min. b Yields of
analytically pure material. c Determined by CSP-SFC. d Reaction time of
6 h. e Chromatographically homogeneous material. f (3R)-7aa absolute
configuration.
In the subsequent survey of aldehyde structure, the reactivity
of aliphatic aldehydes was of particular interest. Traditionally,
aliphatic aldehydes have demonstrated poor reactivity and only
moderate levels of selectivity in the Lewis base-catalyzed
reactions of trichlorosilyl nucleophiles.16 Only the highly
reactive methyl acetate-derived trichlorosilyl ketene acetal was
able to participate in a productive reaction with these substrates
under Lewis base catalysis.
Table 2. Aldol Reaction with
1-(tert-Butyldimethylsilyloxy)-1-tert-butyloxyethene (1b)a
entry
R
7:8b
yield, %c
er d
The aldol reaction of 1a with a wide range of aldehydes
provided consistently high yields and enantioselectivities. In a
series of aromatic aldehydes, 2-naphthaldehyde (5c) and 4-tolu-
aldehyde (5d) gave yields and selectivities comparable to those
of benzaldehyde (entries 2 and 4). However, the isomeric
1-naphthaldehyde (5b) showed a considerable drop in selectivity
(entry 3). The electronic nature of the aldehyde seemed to have
little effect on reactivity or selectivity. The electron-rich nature
1
2
3
4
C6H5 (5a)
1-naphthyl (5b)
4.4:1
6:1
2.2:1
ND
68
79
58
46
97.9:2.1
96.9:3.1
85.5:14.5
96.9:3.1
(E)-C6H5CHdC(CH3) (5h)
PhCH2CH2 (5k)e
a Reactions employed 1.1 equiv of SiCl4, 1.2 equiv of 1b, and 0.01 equiv
of (R,R)-6 at 0.2 M in CH2Cl2 at -78 °C for 1 h. b Determined by 1H NMR
analysis. c Yields of analytically pure material. d Determined by CSP-SFC.
2. Propanoate-Derived Silyl Ketene Acetal Additions. 2.1.
Survey of Propanoate Structure. The results of the aldol
addition of the acetate-derived silyl ketene acetal 1a demon-
(31) Lutz, G. P.; Du, H.; Gallagher, D. J.; Beak, P. J. Org. Chem. 1996, 61,
4542-4554.
9
3778 J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005

Lewis Base Activation of Lewis Acids
A R T I C L E S
strated that the catalyst could exert a high degree of control
over the approach of a nucleophile to one face of the activated
aldehyde. In the next stage of this study, a series of propanoate-
derived silyl ketene acetals (2) reacted with benzaldehyde 5a
under the standard conditions mentioned above to afford high
yields of the desired products. Furthermore, high levels of anti
diastereoselectivity were consistently obtained (Table 3). It
should be noted that these high levels of anti diastereoselectivity
were not dependent on the specific catalyst structure. Similar
results were obtained from reactions run to obtain the racemates
required for CSP-SFC analysis with HMPA. The structure of
the alkoxy group seemed to have little effect on diastereose-
lectivity, although its influence on enantioselectivity was
significant. Enantioselectivity increased with the size of the
alkoxy group in the order of Me < Et < i-Pr < Ph , t-Bu. In
the reaction of the (E)-tert-butyldimethylsilyl ketene acetal 2e
with benzaldehyde, nearly complete selectivity for (2S,3R)-9ea
was observed in the presence of only 1 mol % of the catalyst
(R,R)-6. The product configuration was established by cleavage
of the ester group and by comparison of the 1H NMR spectrum
and optical rotation of the carboxylic acid to the data reported
in the literature.32 This result indicates that Re face attack on
the aldehyde had occurred, as was the case in the acetate-derived
silyl ketene acetal additions described above.
2.2. Scope of Aldehyde Structure in the Additions of the
Silyl Ketene Acetal (E)-2e. Encouraged by the high levels of
enantioselectivity and the anti diastereoconvergence observed
in the reactions of (E)-2e, we investigated the scope of the
reaction with a wide range of aldehydes. In all cases, the
reactions were slower than the corresponding acetate additions,
but high yields, anti diastereo-, and enantioselectivities were
generally observed (Table 4). A variety of aromatic aldehydes,
ranging from electron-rich substrates, such as 4-methoxyben-
zaldehyde (5e) and 2-furaldehyde (5i), to electron-poor aromatic
aldehydes, such as 4-trifluoromethylbenzaldehyde (5f), reacted
with consistently high selectivities (entries 4-6). Other aromatic
aldehydes, such as 1- and 2-naphthaldehyde (5c) and thiophene-
2-carboxaldehyde (5l), gave comparable yields and selectivities
(entries 2, 3, 6, and 7).35 Olefinic aldehydes, such as cinna-
maldehyde (5g) and R-methylcinnamaldehyde (5h), also per-
formed well under the standard reaction conditions (entries 8
and 9). However, the reactivity patterns of all conjugated
aldehydes were not similar. Variable yields and selectivities were
obtained with crotonaldehyde (5m) and 3-phenylpropargylal-
dehyde (5n) (entries 10 and 11). Comparison of these selectivi-
ties with those observed for the reactions of acetate-derived silyl
ketene acetal 1a showed that a drop in selectivity is not observed
with aldehydes such as 1-naphthaldehyde and R-methylcinna-
maldehyde.
Table 3. Aldol Reaction with Propanoate-Derived
tert-Butyldimethylsilyl Ketene Acetals (2)a
Table 4. Aldol Reaction with
(E)-1-(tert-Butyloxypropenyloxy)-tert-butyldimethylsilane (2e)a
entry
R
E:Zb
product
yield, %
drb
S/Nc
erd
entry
R
product yield, %b
dr c
S/N d
er e
1
2
3
4
5
6
7
Me (2a)
Et (2b)
Et (2b)
i-Pr (2c)
Ph (2d)
82:18
95:5
7:93
71:29
94:6
9aa
9ba
9ba
9ca
9da
9ea
9ea
98e
99:1
98:2
85:15
98:2
94:6
99:1
99:1
981
85.4:14.6
88.8:11.2
86.7:12.4
91.1:8.9
93.9:6.1
78e,f
76e
1
2
3
4
5
6
7
8
9
C6H5 (5a)
9ea
9eb
9ec
9ee
9ef
9ei
93f
98
95
88
93
96
90
98
90
55
92
99:1 311 >99.9:0.1
1-naphthyl (5b)
2-naphthyl (5c)
4-CH3OC6H4 (5e)
4-CF3C6H4 (5f)
2-furyl (5i)
2-thiophenyl (5l)
(E)-PhCHdCH (5g)
(E)-PhCHdC(CH3)(5h) 9eh
>99:1 669
>99:1 711
>99:1 209
>99:1 889
98:2
97.2:2.8
99.6:0.4
99.2:0.8
96.0:4.0
96.9:3.1
94.3:5.7
80e
264
92
98e
t-Bu (2e) 95:5
t-Bu (2e) 12:88
93f,g
73e,g
311 >99.9:0.1
162 99.5:0.5
9el
95:5
a All reactions employed 1.1 equiv of SiCl4, 1.2 equiv of silyl ketene
9eg
>99:1 98 >99.9:0.1
acetal, and 0.01 equiv of (R,R)-6 at 0.2 M in CH2Cl2 at -78 °C for 3 h.
>99:1
>99:1
96:4
96.1:3.9
98.2:1.8
84.5:15.5
b Determined by H NMR analysis. c Determined by H NMR analysis of
HC(3). d Determined by CSP-SFC. e Chromatographically homogeneous
material. f Analytically pure material. g (2S,3R) absolute configuration.
1
1
10 (E)-CH3CHdCH (5m)
11 phenyl propargyl (5n)
9em
9en
a Reactions employed 1.1 equiv of SiCl4, 1.2 equiv of 2e, and 0.01 equiv
of (R,R)-6 at 0.2 M in CH2Cl2 at -78 °C for 3 h. b Yields of analytically
pure material. c Determined by 1H NMR analysis. d Determined by 1H NMR
analysis of HC(3). e Determined by CSP-SFC. f (2S,3R) absolute config-
uration.
Although the size of the alkoxy substituent was crucial for
obtaining high enantioselectivity, the geometry of the silyl
ketene acetal was unimportant. Nearly identical levels of
diastereo- and enantioselectivity were observed regardless of
the geometry of the silyl ketene acetal. In the case of (Z)-2b,
high, but not complete, anti diastereoconvergence was obtained
when compared to the reaction of (E)-2b (Table 3, entries 2
and 3).33 Examination of both the Z- and E-isomers of the tert-
butyl propanoate-derived silyl ketene acetals demonstrated high
anti diastereoconvergence (entries 6 and 7).34 The major product
(9ea) isolated from the reaction of (Z)-2e was found to have
the same absolute and relative configuration as the product
derived from the reaction of (E)-2e.
When the reaction of aliphatic aldehydes was investigated
with the silyl ketene acetal (E)-2e, virtually no product was
observed, even after extended reaction times (∼24 h). Although
the reason was unclear at the time, aliphatic aldehydes did prove
to be reactive with the less sterically encumbered silyl ketene
(34) Kamimura, A.; Mitsudera, H.; Asano, S.; Kidera, S.; Kakehi, A. J. Org.
Chem. 1999, 64, 6353-6360.
(35) The diastereoselectivity reported in this work for the aldol product derived
from 1-naphthaldehyde 5b and 2e (9eb) is higher than that reported in the
previous communication. The species originally assigned as the syn
diastereomer of 9eb was, in fact, the aldol product derived from 2-naph-
thaldehyde (9ec). Commercially available 1-naphthaldehyde is contaminated
with 2-4% 2-naphthaldehyde (Aldrich). This isomer could not be removed
by simple distillation. The observed enantioselectivity is unaffected as the
peaks for 9eb and 9ec were well separated in the CSP-SFC analysis.
(32) Van Draanen, N. A.; Arseniyadis, S.; Crimmins, M. T.; Heathcock, C. H.
J. Org. Chem. 1991, 56, 2499-2506.
(33) Davies, S. G.; Edwards, A. J.; Evans, G. B.; Mortlock, A. A. Tetrahedron
1994, 50, 6621-6642.
9
J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005 3779

A R T I C L E S
Denmark et al.
Table 5. Aldol Reaction of
(E)-1-(Ethoxypropenyloxy)-tert-butyldimethylsilane (2b) with
Aliphatic Aldehydes
Table 6. Aldol Reaction with R-Substituted tert-Butyldimethylsilyl
Ketene Acetals
entry
R
product
yield,%a
drb
S/Nc
erd
Me (2b)e
9ba
78f
75
82
88
98:2
92:8
85:15
97:3
981
88.8:11.2
83.6:16.4
85.8:14.2
86.3:13.7
entry
R
product
yield, %a
drb
S/Nc
erd
1
2
3
4
Et (10a)e
11aa
11ba
11ca
1
2
3
4
5
6
7
PhCH2CH2 (5k)e
(CH3)2CHCH2 (5o)e
(CH3)2CHCH2 (5o)g
TBSO(CH2)5 (5p)g
BnO(CH2)5 (5q)g
BzO(CH2)5 (5r)g
cyclohexyl (5j) h
9bk
9bo
9bo
9bp
9bq
9br
9bj
71
65f
76
51
70
60
40
91:9
234 94.9:5.1
i-Pr (10b)e
i-Bu (10c)e
339
89:11 105 93.0:7.0
91:9
96:4
95:5
95:5
72 89.7:10.3
96.7:3.3
a Yields of analytically pure material. b Determined by 1H NMR analysis.
c Determined by 1H NMR analysis of HC(3). d Determined by CSP-SFC.
e Reactions employed 1.1 equiv of SiCl4, 1.2 equiv of the silyl ketene acetal,
and 0.01 equiv of (R,R)-6 at 0.2 M in CH2Cl2 at -78 °C for 3 h.
f Chromatographically homogeneous material.
96.7:3.3
95.4:4.5
89:11 430 67.8:32.2
a Yields of analytically pure material. b Determined by 1H NMR analysis.
c Determined by 1H NMR analysis of HC(3). d Determined by CSP-SFC.
e Reactions employed 1.1 equiv of SiCl4, 1.2 equiv of 2b, 0.05 equiv of
(R,R)-6, and 0.1 equiv of TBAI at 0.4 M in CH2Cl2 at -78 °C for 24 h.
f Chromatographically homogeneous material. g Reactions employed 1.1
equiv of SiCl4, 1.2 equiv of 2b, 0.05 equiv of (R,R)-6, and 0.05 equiv of
DIPEA at 0.4 M in CH2Cl2 at -78 °C for 24 h. h Reaction employed 1.1
equiv of SiCl4, 1.2 equiv of 2b, 0.1 equiv of (R,R)-6, and 0.1 equiv of
TBAI at 0.8 M in CH2Cl2 at -40 °C for 24 h.
Scheme 5
acetal (E)-2b (Table 5). Further optimization showed that the
combination of longer reaction times, higher catalyst loadings,
and the addition of either tetrabutylammonium iodide (TBAI)
or diisopropylethylamine (i-Pr2EtN) to the solution could give
synthetically useful yields of the aldol adducts (compare entries
2 and 3). In the case of the unbranched hydrocinnamaldehyde
(5k) and the â-branched isovaleraldehyde (5o), moderate yields
could be obtained under optimized conditions (entries 1-3).
The levels of selectivity observed in these reactions, although
high, fell short of the levels attained in the aldol reactions of con-
jugated aldehydes, such as benzaldehyde (5a) and cinnamalde-
hyde (5g). Disappointingly, the reaction with the R-branched
cyclohexanecarboxaldehyde (5j) remained sluggish. Even with 10
mol % of the dimeric phosphoramide catalyst (R,R)-6 at -40 °C,
only moderate yields and selectivities were obtained (entry 7).
In view of the anticipated oxophilicity of the putative silyl
cation intermediate, a brief investigation of the stability and
influence of commonly used silyl-, ester-, and ether-protecting
groups was undertaken. The reactions of 6-(tert-butyldimethylsilyl-
oxy)-, 6-benzyloxy-, and 6-benzoyloxyhexanal (5p-r) revealed
that comparable yields could be obtained under the modified
conditions developed for simple aliphatic aldehydes. In the
presence of Lewis basic functional groups, such as ethers and
esters, similar levels of diastereo- or enantioselectivity were
obtained when compared to simple, unfunctionalized aliphatic
aldehydes (entries 4-6).
2.3. Scope of Silyl Ketene Acetal Structure in the Addition
to Benzaldehyde. Although the steric demands of the alkoxy
substituent in the silyl ketene acetal have a strong effect on the
enantioselectivity of the aldol reaction, the influence of the
structure of the ester-derived portion had not been investigated.
Thus, a series of ethyl ester-derived silyl ketene acetals were
prepared bearing methyl (2b), ethyl (10a), isopropyl (10b), and
isobutyl (10c) substituents and were then subjected to the
optimized reaction conditions with benzaldehyde. In this series,
yields and selectivities were generally high and comparable to
those obtained with (E)-2b (Table 6, entry 1). The isovalerate-
derived silyl ketene acetal 10b, bearing an R-branched sub-
stituent, did show a marked decrease in diastereoselectivity.
However, in the presence of a larger substituent, such as the
isobutyl group of the isocaproate-derived silyl ketene acetal 10c,
comparable levels of diastereo- and enantioselectivity were
restored when compared to 2a and 10a.
An R,R-disubstituted methyl ester-derived silyl ketene acetal
3 was also prepared to investigate the effect of R-disubstitution.
Silyl ketene acetal 3 exhibited severely attenuated reactivity
(Scheme 5). Synthetically useful yields for the addition of the
isobutyrate-derived silyl ketene acetal 3 to benzaldehyde 5a
could be obtained by increasing the reaction times, increasing
the catalyst loading of (R,R)-6 to 5 mol %, and using 5 mol %
of i-Pr2EtN as an additive. Under these conditions, moderate
yields and enantioselectivities were obtained for the â-hydroxy
ester product 12. The absolute configuration of this adduct was
determined to be 3S36 by optical rotation, which despite a priority
change in assigning this center, still represents a Re face attack,
as observed in the case of the propanoate addition product 9ea
and the acetate addition product 7aa.
3. Vinylogous Aldol Reactions of Silyl Dienol Ethers. The
low reactivity of the isobutyrate-derived silyl ketene acetal 3,
when compared to that of the other, less sterically demanding
silyl ketene acetals examined in this survey, suggests that the
degree of substitution at the nucleophilic carbon center has a
significant effect on the rate of the aldol reactions. This
sensitivity of the chiral phosphoramide/SiCl4 catalyst system
to steric factors might be used to augment the inherent
γ-regioselectivity of vinylogous aldol reactions of silyl dienol
ethers. If high regioselectivity could be achieved without
affecting the high diastereo- and enantioselectivity observed in
the reactions of simple ketene acetals, a general and highly
selective vinylogous aldol process could be obtained.
3.1. Synthesis of the Silyl Dienol Ethers. Although the silyl
dienol ethers for the studies of the Lewis base-catalyzed
vinylogous aldol reaction are not thoroughly described in the
(36) Kobayashi, S.; Ishitani, H.; Yamashita, Y.; Ueno, M.; Shimizu, H.
Tetrahedron 2001, 57, 861-866.
9
3780 J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005

Lewis Base Activation of Lewis Acids
A R T I C L E S
Table 7. Vinylogous Aldol Reactions of Simple, Ester-Derived Silyl Dienol Ethers 4a-d with Aldehydes 5a-k
entry
dienolate
R1
R2
R3
R4
R5
product
yield, %a
γ
/
Rb
drb
erc
1
2
3
4
5
6
7
8
4ad
4ad
4af
4bd
4bd
4bf
4cd
4cd
4cf
Ph (5a)
Et
Et
Et
Me
Me
Me
Et
Et
Et
H
H
H
Me
Me
Me
H
H
H
H
H
H
H
H
H
Me
Me
Me
H
H
H
H
H
H
H
H
H
14aa
14ag
14ak
14ba
14bg
14bk
14ca
14cg
14ck
14da
14dg
14dk
89e
84e
68g
93e
88
>99:1
>99:1
>99:1
>99:1
>99:1
ND
>99:1
>99:1
>99:1
99:1
98.8:1.2
98.3:1.7
95.3:4.7
99.5:0.5
99.6:0.4
ND
96.1:3.9
94.0:6.0
97.5:2.5
94.5:5.5
91.2:8.8
ND
PhCHdCH (5g)
PhCH2CH2 (5k)
Ph (5a)
PhCHdCH (5g)
PhCH2CH2 (5k)
Ph (5a)
PhCHdCH (5g)
PhCH2CH2 (5k)
Ph (5a)
PhCHdCH (5g)
PhCH2CH2 (5k)
91e
97
-
9
H
73
10
11
12
4dd
4dd
4df
t-Bu
t-Bu
t-Bu
H
H
H
Me
Me
Me
92h
71
>99:1
>99:1
ND
H
H
99:1
ND
a Yields of analytically pure material. b Determined by 1H NMR analysis. c Determined by CSP-SFC. d Reactions employed 1.1 equiv of SiCl4, 1.2
equiv of dienolate, and 0.01 equiv of (R,R)-6 at 0.2 M in CH2Cl2 at -78 °C for 3 h. e (R) absolute configuration. f Reactions employed 1.1 equiv of SiCl4,
1.2 equiv of dienolate, 0.05 equiv of (R,R)-6, and 0.05 equiv of i-Pr2EtN at 0.2 M in CH2Cl2 at -78 °C for 24 h. g (S) absolute configuration. h (2S,3R)
absolute configuration.
literature, the use of established synthetic methods allows for
easy access to these compounds. The enolization procedure
developed by Schlessinger and co-workers proved to be a
versatile and reliable method.37 In this procedure, the addition
of 1 equiv of HMPA to the solution of LDA prior to addition
of the unsaturated carbonyl compound is required to suppress
competing conjugate addition of the amide base.
The modified protocol allowed for the synthesis of several
tert-butyldimethylsilyl dienol ethers in good yields. The products
could be purified by distillation and stored for 1-2 weeks at
-15 °C without significant decomposition. In most cases, the
products were not obtained as single geometrical isomers. The
only case that did provide high selectivity was the ethyl sene-
cioate-derived silyl dienol ether 4c (Figure 4). In some cases,
the configuration of the major isomer of the silyl dienol ethers
could be determined by measuring their 1H nOe NMR spectra.
The major isomers obtained from ethyl crotonate and tert-butyl
pentenoate are (1Z)-4a and (1Z)-4d, respectively. The prepara-
tion of the tert-butyldimethylsilyl analogue of Grunwell’s diene38
was also achieved under these conditions, providing 13 in good
yield. In 4d, the major isomer was assigned by inspection of
the 1H NMR coupling constants. The major isomer of the tert-
butyl pentenoate-derived silyl dienol ether 4d, bearing a methyl
substituent at C(4), was determined to be the 1Z,3Z-isomer.39
3.2. Vinylogous Aldol Reactions of Simple Ester-Derived
Silyl Dienol Ethers. The reaction of the ethyl crotonate-derived
silyl dienol ether 4a with benzaldehyde was chosen to address
the fundamental question of site selectivity with these nucleo-
philes. Without a substituent that might provide a steric bias for
reactivity at the terminal γ-position, only the inherent site selec-
tivity provided by orbital control would be operative. In the pres-
ence of only 1 mol % of the chiral dimeric phosphoramide (R,R)-
6, the γ-addition product 14aa was obtained exclusively in high
yield and enantioselectivity (Table 7, entry 1). Similar results
were obtained in the addition of 4a to cinnamaldehyde as well
as the aliphatic aldehyde, hydrocinnamaldehyde (5k) (entries 2
and 3). These results demonstrate that the catalyst complex
provides sufficient differentiation, even between mono- and
unsubstituted positions, giving nearly exclusive γ-site selectivity.
A survey of silyl dienol ethers bearing methyl substituents
at different sites was then undertaken. In the case of the methyl
tiglate-derived silyl dienol ether 4b, with a competition between
di- and unsubstituted reactive centers, high regio- and enanti-
oselectivities were observed in the additions to benzaldehyde
and cinnamaldehyde (entries 4 and 5). However, this more
highly substituted substrate did not react with the aliphatic
aldehyde 5k. The â-methyl-substituted silyl dienol ether 4c
derived from ethyl senecioate showed more general reactivity,
providing the γ-adduct in good yield, regio-, and enantioselec-
tivity for all three aldehydes in the survey (entries 7 and 9). It
is interesting to note that despite poor geometrical selectivity
in the formation of the silyl dienol ethers, the E-isomer of the
product was consistently obtained as a single isomer. Only in
the reactions of the ethyl senecioate-derived silyl ketene acetal
4c could any of the corresponding Z-isomer be detected by 1H
NMR analysis (E/Z 97:3).
Clearly, the high levels of enantioselectivity observed in the
additions of simple silyl ketene acetals to aldehydes translated
(37) Herrman, J. L.; Kieczykowski, G. R.; Schlessinger, R. H. Tetrahedron Lett.
1973, 14, 2433-2436.
(38) Grunwell, J. R.; Karipides, A.; Wigal, C. T.; Heinzman, S. W.; Parlow, J.;
Surso, J. A.; Clayton, L.; Fleitz, F. J.; Daffner, M.; Stevens, J. E. J. Org.
Chem. 1991, 56, 91-95.
(39) The other isomers present could not be unambiguously identified by 1H
NMR analysis.
Figure 4. Silyl dienol ethers used in the vinylogous aldol reactions.
9
J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005 3781

A R T I C L E S
Denmark et al.
Figure 5. Correlation of the configuration of vinylogous aldol products to known compounds.
Table 8. Table 8. Vinylogous Aldol Reactions of
Dioxanone-Derived Silyl Dienol Ether 13 with Aldehydes 5a-k
into high enantioselectivities for the additions of silyl dienol
ethers. The use of a 2-pentenoate-derived silyl dienol ether, bear-
ing a methyl group at the terminal γ-position, would address
whether the high anti diastereoselectivity observed previously
would also be maintained. Disappointingly, an ethyl 2-penteno-
ate-derived silyl dienol ether provided a statistical mixture of
the two possible isomeric aldol adducts. However, increasing the
size of the alkoxy substituent from ethyl to tert-butyl restored the
high levels of regioselectivity observed in other cases (entries
10 and 11). The γ-addition product 14da could be obtained in
good yields as well as excellent anti diastereo- and enantioselec-
tivities for aldol reactions with benzaldehyde and cinnamal-
dehyde. Despite several attempts, no reaction was observed with
the aliphatic aldehyde 5k. This behavior was similar to the
reactivity pattern observed with the methyl tiglate-derived silyl
dienol ether 4b. The fact that these subtle changes in structure
led to such dramatic changes in reactivity attests to the sensitivity
of the catalyst complex to the steric demands of the nucleophile.
3.3. Vinylogous Aldol Reactions of a Dioxanone-Derived
Silyl Dienol Ether. Simple ester-derived silyl dienol ethers, such
as 4a-d, comprise one class of commonly used nucleophiles
in vinylogous aldol reactions. The other class, three-heteroatom-
substituted silyl dienol ethers, such as Grunwell’s diene,38 were
also investigated to provide a point of comparison with other
catalytic asymmetric methods.28 Reactions of the dioxanone-
derived silyl dienol ether 13 gave uniformly high regioselec-
tivities and yields (Table 8). However, the enantioselectivities
of these reactions were extremely sensitive to the structure of
the aldehyde. The selectivity obtained with the aliphatic
aldehyde 5k proved to be the highest in the series (er(5k) >
er(5g) > er(5a)). This is particularly noteworthy because this
trend is reversed to that usually observed in this series of
aldehydes. Determination of the absolute configuration of these
aldol products (vide infra) revealed that the product derived from
Re face attack on the aldehyde had been produced, even though
entry
R
product
yield, %a
γ/
Rb
erc
1
2
3
Ph (5a)d
15a
15g
15k
92e
88e
83g
>99:1
>99:1
>99:1
87.2:12.8
88.9:11.1
94.6:5.4
PhCHdCH (5g)d
PhCH2CH2 (5k)f
a Yields of chromatographically homogeneous material. b Determined by
1H NMR analysis. c Determined by CSP-SFC. d Reactions employed 1.1
equiv of SiCl4, 1.2 equiv of dienolate, and 0.01 equiv of (R,R)-6 at 0.2 M
in CH2Cl2 at -78 °C for 3 h. e (R) absolute configuration. f Reactions
employed 1.1 equiv of SiCl4, 1.2 equiv of dienolate, 0.05 equiv of (R,R)-6,
and 0.05 equiv of DIPEA at 0.2 M in CH2Cl2 at -78 °C for 24 h. g (S)
absolute configuration.
in some cases the product possessed a 5S stereocenter due to
Cahn-Ingold-Prelog priority changes.
4. Determination of the Absolute Configuration of the
Products of the Vinylogous Aldol Reaction. Unlike the simple
â-hydroxy esters produced in the reactions of acetate- and
propanoate-derived silyl ketene acetals, the absolute configura-
tions of many of these vinylogous aldol products could not be
determined by direct correlation to known examples in the
literature. Instead, a synthetic route needed to be devised to relate
these structures to a set of compounds for which the absolute
configurations had been unambiguously assigned. A straight-
forward chemical correlation involved cleavage of the
C(2)-C(3) double bond and formation of a 1,3-diol, the absolute
configuration of which has been established by Masamune and
others,23a,40 in analogy to the stereochemical model provided
by the Sharpless asymmetric epoxidation.41
In the cases of the aldol adducts 14aa, 14ak, 14ba, and 14da,
ozonolysis followed by reduction of the ozonide with a solution
9
3782 J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005

Lewis Base Activation of Lewis Acids
A R T I C L E S
of lithium aluminum hydride in tetrahydrofuran provided the
desired diols in moderate yields (Figure 5, eqs 1, 3, 5, and 7).
Comparison of their optical rotations with those reported in the lit-
erature revealed that in each case, the aldol adduct derived from
Re face attack on the aldehyde had been produced, even though
in some cases, the product possessed a 3S stereocenter due to
priority changes. This sense of asymmetric induction is con-
sistent with that observed in all other cases reported in this study.
Two cases required different routes due to their substitution
patterns. The aldol adduct 14ag would not be amenable to this
method because of concern about competitive ozonolysis of the
C(6)-C(7) double bond. Therefore, an alternate route was em-
ployed that would not require differentiation of the two double
bonds contained in the molecule. Both double bonds of 14ag
were hydrogenated, generating the alkanoate 17. This compound
was then compared to identical alkanoate prepared from the
now known (5R)-14ak to establish absolute configuration, and
the optical rotations were similar (compare eqs 2 and 4).
Finally, in the case of the ethyl senecioate-derived aldol
adduct 14ca, correlation to the known â-acetoxy methyl ketone
was made by acetylation of 14ca with acetic anhydride followed
by cleavage of the C(2)-C(3) double bond with ruthenium
tetraoxide (eq 5). Again, the product corresponding to the (3R)
enantiomer was obtained.42
ketene acetal 2b. Furthermore, the amount of this species
initially formed did not change, even after 6 h. The formation
of an equivalent amount of TBSCl was also observed, consistent
with trans-silylation between SiCl4 and the silyl ketene acetal
2b. Analysis of these spectra and comparison to literature reports
led to the conclusion that this new species was the R-trichlo-
rosilyl ethyl propanoate 21. Subsequent in situ 1H NMR studies
of the aldol reactions between 2b and benzaldehyde or hydro-
cinnamaldehyde revealed that this newly formed species was
completely unreactive and remained even after consumption of
the aldehyde. The loss of a small amount of the silyl ketene
acetal to this unproductive side reaction helps to understand
the need for a slight excess of both the silyl ketene acetal and
SiCl4 in the optimal procedure.
Scheme 6
The attenuated reactivity of aliphatic aldehydes in the Lewis
base-catalyzed reactions of trichlorosilyl enol ethers and allyl-
trichlorosilanes had long been an issue of concern.16 Because
the highly reactive acetate-derived silyl ketene acetal 1a did
react with aliphatic aldehydes, an examination of reactivity of
this substrate was prompted. Monitoring the reaction of silyl
ketene acetal 1a with cyclohexanecarboxaldehyde (5j) in the
presence of SiCl4 and 5 mol % of (R,R)-6 by ReactIR revealed
the immediate disappearance of the aldehyde signal without a
corresponding increase in the ester signal derived from the
product. This reaction required approximately 1 h to reach
completion, as indicated by monitoring the formation of the ester
product 6aj. This observation was surprising, considering that
under similar conditions, the aldol reaction of 1a and benzal-
dehyde required less than 30 s. To probe the unexplained
disappearance of the aldehyde signal, further studies were
performed. Upon mixing 5j and SiCl4 in CDCl3 in the presence
of 10 mol % of HMPA, examination of the 1H NMR spectrum
revealed that the aldehydic proton had completely disappeared
only to be replaced by a new set of signals, assigned as the
R-chloro trichlorosilyl ether 23 (Scheme 7). The structure of
23 was confirmed by comparison to similar species in the
literature43 and to a related compound that had previously been
identified in these laboratories.44 Similar 1H NMR experiments
5. Spectroscopic Investigations of the Reaction Intermedi-
ates. To gain information regarding the identity of the active
enolate equivalent as well as the cause of the reactivity problems
associated with aliphatic aldehydes, several studies were
undertaken using both ReactIR and 1H NMR spectroscopy. To
investigate the existence of the putative silyl cation intermediate,
1
a series of low temperature H and 29Si NMR spectroscopic
studies were undertaken. 1H NMR investigations on the stability
of the silyl ketene acetals in the presence of the phosphoramide
and SiCl4 could identify whether isomerization of the silyl ketene
acetal or trans-silylation to the trichlorosilyl ketene acetal was
occurring prior to aldolization. A clear understanding of the
identity of the active enolate equivalent in these reactions is
essential to the development of a rationale for the peculiar anti
diastereoconvergence of this process. Because the chemical shift
values for several O- and C-trichlorosilyl carbonyl compounds
have been documented in the literature, 1H NMR spectroscopy
appeared to be the technique of choice for these studies.
The stability of the ethyl propanoate-derived O-tert-butyldi-
methylsilyl ketene acetal 2b was first studied. Upon mixing
equimolar amounts of 2b and SiCl4 in CDCl3 at 0.2 M at -65
°C in the presence of 10 mol % of HMPA, conditions similar
to those used in the synthetic process but in the absence of the
aldehyde, little change could be observed in the 1H NMR
spectrum of 2b (Scheme 6). The silyl ketene acetal remained
unchanged, with no noticeable erosion in the initial E/Z ratio,
even after extended times. No evidence of signals corresponding
to the O-trichlorosilyl ketene acetal could be observed. The only
change was the appearance of a quartet signal at 2.16 ppm.
1
performed with benzaldehyde showed little change in the H
NMR spectrum and no evidence for formation of an analogous
R-chloro trichlorosilyl ether.
Scheme 7
1
Careful examination of the H and 13C NMR spectra of this
reaction mixture revealed that only a small amount of this new
species had been formed upon mixing. This new signal
represented approximately 10% of the initial amount of silyl
(40) Nun˜ez, M. T.; Martin, V. S. J. Org. Chem. 1990, 55, 1928-1932.
(41) (a) Katsuki, T.; Martin, V. S. Org. React. 1996, 48, 1-300. (b) Katsuki,
T.; Sharpless, K. B. J. Am. Chem. Soc. 1980, 102, 5974-5976.
Crotonaldehyde (5m) also exhibited uncharacteristically low
reactivity and provided low yields of the aldol product when
(42) Nair, M. S.; Joly, S. Tetrahedron: Asymmetry 2000, 11, 2049-2052.
9
J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005 3783

A R T I C L E S
Denmark et al.
compared to the closely related substrate cinnamaldehyde. Upon
mixing 5m and SiCl4 in the presence of 10 mol % of HMPA at
-70 °C in CDCl3, 1H NMR analysis revealed a similar phenom-
enon to that observed with cyclohexanecarboxaldehyde (Scheme
8). The signal corresponding to the aldehydic proton decreased
in intensity, coincident with the appearance of several new
signals, assigned to the R-chlorosilyl ether 24 and the R-chloro
trichlorosilyl ether 25. After 15 min, these three species were
found to exist in approximately a 2:1:1 ratio of 5m:24:25.
and co-workers for HMPA-bound trimethylsilyl cations.50 The
second signal at -206 ppm is assigned as the hexachlorosilicate
dianion (27) by consideration of its extremely low chemical
shift (29Si δ (SiF62-) ) -197 ppm).51 Thus, the problem of
poor anion solvation in CH2Cl2 (ꢀ ) 8.93)52 is partially
alleviated.
Scheme 9
Scheme 8
29Si NMR spectroscopic experiments conducted with the
dimeric phosphoramide catalyst (R,R)-6 revealed the presence
of similar species. However, it appears that the species are under
rapid exchange under these conditions. The signal corresponding
to the putative silyl cation is rather broad, obscuring the
coupling, whereas the signal at -206 ppm could not be
observed. Although these data are not as convincing as those
obtained in the case of the HMPA-ligated silyl cation, it must
be remembered that the dimeric phosphoramide (R,R)-6 is a
much weaker Lewis base than HMPA due to the electron-
withdrawing effect of the naphthyl rings.53 Regardless of these
differences, observation of the 31P NMR spectra of the samples
prepared from SiCl4 and either phosphoramide reveals a
dramatic upfield shift of the phosphoramide signals from 27 to
19 ppm. This change in the chemical shift is suggestive of
complexation to a metal center and is consistent with the upfield
shifts in the 31P NMR of related chiral phosphoramides induced
upon binding to SnCl4.54
Finally, the existence of the putative silyl cation intermediate
could be addressed by NMR spectroscopy. This species is
proposed to be the active Lewis acid that mediates these aldol
processes. Preliminary kinetic studies show that the reaction is
zero order in SiCl4.45 Thus, the equilibrium binding of the
phosphoramide to SiCl4 is quite strong, and at any time, the
phosphoramide catalyst is saturated with SiCl4 (or other
chlorosilane species). If the catalyst resting state was indeed a
silyl cation complex, it should be spectroscopically observable.
29Si NMR experiments were performed under conditions that
approximated those used in the synthetic protocol. Upon mixing
SiCl4 and HMPA in a 1:1 ratio in CDCl3 at -60 °C, several
signals could be observed (Scheme 9). The major signal
corresponded to free SiCl4 (-20 ppm), with new signals
appearing at -110 and -206 ppm. The signal at -110 ppm,
which lies in the region for a five-coordinate silicon species,46
is split into a triplet with a coupling constant of J ) 9 Hz. On
the basis of these observations, this signal was tentatively
assigned the structure SiCl3(HMPA)2+ (26). The chemical shift
value of the putative trichlorosilyl cation is consistent with
values reported for other Lewis base adducts of heteroatom-
substituted silyl cations.47 The value is at lower field compared
to the adduct between allyltrichlorosilane and DMF reported
by Kobayashi and co-workers (δ ) -170 ppm). However, a
chemical shift of -170 ppm more likely represents the neutral,
hexacoordinate adduct rather than a pentacoordinate silyl
cation.48 A closer analogy is provided by the pentacoordinate
trichlorosilyl amidinate complexes reported by Karsch and co-
Discussion
The chiral Lewis base-catalyzed/SiCl4-mediated aldol addi-
tions of silyl ketene acetal nucleophiles are characterized by
rapid rates and high levels of regio-, diastereo-, and enantiose-
lectivity. To formulate a unified picture of this process, the
trends derived from the structural changes in both the aldehyde
and the silyl ketene acetal will first be highlighted. In particular,
the significant influence of the steric demands of the nucleo-
philes will be considered. These trends will then be used as the
basis for an open transition structure model to explain the high
levels of anti diastereoselectivity observed in this reaction. With
the aid of computational studies performed at the PM3 level of
theory, this model will then be elaborated to include an analysis
of the factors influencing enantioselectivity.
2
workers (-89 < δ < -99 ppm).49 The value of the JSi-P
coupling, although small, is similar to those observed by Cremer
1. Reactivity Trends. The reactivity patterns observed with
respect to silyl ketene acetal structure stand in contrast to the
reactivity scales developed by Mayr and co-workers for main
group organometallic nucleophiles.20 In an extensive study of the
reactivity of silyl enol ethers, Mayr found that an isobutyrate-
(43) (a) Lokensgard, J. P.; Fisher, J. W.; Bartz, W. J.; Meinwald, J. J. Org.
Chem. 1985, 50, 5609-5611. (b) Gundersen, L.-L.; Benneche, T. Acta
Chem. Scand. 1991, 45, 975-977.
(44) Chloro-1-trichlorosiloxyheptane: Ghosh, S. K. Unpublished results from
these laboratories.
(45) Examination of the rate of reaction of 1-naphthaldehyde with the silyl ketene
acetal 2e in the presence of SiCl4 (1-4 equiv) and 1 mol % of (R,R)-6 by
ReactIR revealed a zero-order dependence on the concentration of SiCl4
(m ) 0.051 ( 0.025, r2 ) 0.798).
(50) Arshadi, M.; Johnels, D.; Edlund, U.; Ottoson, C.-H.; Cremer, D. J. Am.
Chem. Soc. 1996, 118, 5120-5131.
(51) Marsmann, H. In NMR Basic Principles and Progress, Volume 17; Diehl,
P., Fluck, E., Kosfeld, R., Eds.; Springer-Verlag, Berlin, 1981; p 65.
(52) Reichardt, C. SolVent and SolVent Effects in Organic Chemistry; Wiley-
VCH: Weinheim, Germany, 1988; p 407.
(46) Kennedy, J. D.; McFarlane, W. In Multinuclear NMR; Mason, J., Ed.;
Plenum Press: New York, 1987; Chapter 11, p 305.
(47) (a) Bassindale, A. R.; Jiang, J. J. Organomet. Chem. 1993, 446, C3-C5.
(b) Swamy, K. C. K.; Chandrasekhar, V.; Harland, J. J.; Holmes, J. M.;
Day, R. O.; Holmes, R. R. J. Am. Chem. Soc. 1990, 112, 2341-2348.
(48) Kobayashi, S.; Nishio, K. Tetrahedron Lett. 1993, 34, 3453-3456.
(49) Karsch, H. H.; Segmueller, T. In Organosilicon Compounds V: From
Molecules to Materials; Auner, N., Weis, J., Eds.; Wiley-VCH: Weinheim,
Germany, 2003; p 270.
(53) 31P and 119Sn NMR studies have shown that the 1,1′-binaphthyl-2,2′-
diamine-derived phosphoramide 6 can be displaced from SnCl4 by a
bispyrrolidine-derived phosphoramide: Schleusner, M. Unpublished results
from these laboratories.
(54) (a) Denmark, S. E.; Fu, J. J. Am. Chem. Soc. 2003, 125, 2208-2216. (b)
Denmark, S. E.; Xu, S. Tetrahedron 1999, 55, 8727-8738.
9
3784 J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005

Lewis Base Activation of Lewis Acids
A R T I C L E S
derived silyl ketene acetal is more reactive than a related acetate-
derived silyl ketene acetal (N ) 9.49 versus 8.58).19 According-
ly, one might have predicted that the R,R-disubstituted ketene
acetal 3 would be more reactive than the unsubstituted 1a in the
aldol process. However, a reversed trend is observed wherein
the unsubstituted acetate-derived silyl ketene acetal 1a is far
more reactive than the isobutyrate-derived silyl ketene acetal 3
(Figure 6). Furthermore, the presence or absence of an R-sub-
stituent plays a far greater role in determining reac-
tivity than its size or structure. In a series of ethyl ester-derived
silyl ketene acetals 10a-c, bearing a variety of alkyl substit-
uents, little reactivity difference could be discerned by ReactIR
monitoring (Table 6). These reactions were typically complete
in less than 30 min. Clearly, in this reaction, the use of a
nucleophilicity scale based on electronic factors alone does not
provide a good model for rationalizing reactivity. The observed
trend suggests that steric factors have a dominant influence on
the reactivity of silyl ketene acetals with aldehydes under
catalysis by an in situ-generated, chiral phosphoramide-bound,
trichlorosilyl cation.
contrast to studies by Guthrie that have shown that aliphatic
aldehydes are considerably more reactive than conjugated,
aromatic aldehydes toward a wide variety of neutral and anionic
nucleophiles.55 Using Jencks’ free energy relationship for
carbonyl addition reactions (γ), acetaldehyde is shown to be at
least 2 orders of magnitude more reactive than benzaldehyde.56
An alternative explanation for this trend can be formulated
1
from the “in situ protection” of aliphatic aldehydes. H NMR
spectroscopic analysis revealed that both cyclohexanecarbox-
aldehyde and crotonaldehyde were rapidly transformed into
R-chloro trichlorosilyl ethers. Presumably, this process occurs
through attack of the ionized chloride on the silyl cation complex
of the aldehyde 28 (Scheme 10). Therefore, to explain the poor
reactivity of aliphatic aldehydes, we posit that an equilibrium
exists between the aldehyde that is bound in the putative silyl
cation 28 and an unreactive R-chloro trichlorosilyl ether 29.
Because of its low dielectric constant, dichloromethane (ꢀr )
8.93)52 is poorly capable of supporting the ionic species 28.
Collapse of 28 to the R-chloro trichlorosilyl ether generates a
more stable, neutral species. Consequently, a low equilibrium
concentration of the activated aldehyde complex 28 is available
for aldolization. To compensate for the apparent loss in reactivity
of these substrates, a powerful nucleophile must be employed
that can still react at an appreciable rate with the extremely low
concentration of activated aldehyde present in solution at any
time. Less nucleophilic species, although capable, cannot
undergo reactions at rates that provide a synthetically useful
process. The importance of steric factors in this reaction helps
explain the higher reactivity of the ethyl propanoate-derived 2b
with aliphatic aldehydes when compared to that of the tert-
butyl propanoate-derived 2e. The fact that conjugated aldehydes
are resistant to this side process may be due to the unfavorable
loss of resonance stabilization in the extended conjugated
system. This hypothesis can be confirmed through calculations
performed on the addition of HCl to the two representative
aldehydes, benzaldehyde and cyclohexanecarboxaldehyde, using
the PC version of GAMESS(US) QC package employing the
6-31*(p,d) basis set.57 In the case of benzaldehyde, formation
of the chlorohydrin is an endothermic process (∆Ho ) +2.36
kJ/mol), a result consistent with the resistance of this substrate
to chlorohydrin formation under the reaction conditions. How-
ever, in the case of cyclohexanecarboxaldehyde, chlorohydrin
formation is predicted to be highly exothermic (∆Ho ) -18.75
kJ/mol), again consistent with experimental observations.58
Figure 6. Comparison of reaction rates for substituted silyl ketene acetals.
The proximity of the R-substituents to the reactive center at
C(2) of the silyl ketene acetal imparts the greatest significance
to steric effects at this position. However, remote substituents
still affect reactivity. In a series of propanoate-derived silyl
ketene acetals 2a-e, a slight, but noticeable, decrease in reaction
rate was observed as the size of the alkoxy substituent was
increased from methyl to a sterically more demanding tert-butyl
ester. Although these reactivity differences were not as drastic
as those observed in the case of varying the degree of R-sub-
stitution, the role of these substituents in determining selectivity
proved more important (Table 3). Similar effects from remote
substituents were observed in the vinylogous aldol reaction. For
example, the presence or absence of a 2-methyl group, although
remote from the reactive γ-position of the silyl dienol ether,
greatly influenced the scope of the reaction (Table 7).
Scheme 10
When considering reactivity trends with respect to the
electrophilic partner in the reaction, steric effects are clearly
not as important as they are for the nucleophilic partner. Instead,
a different trend becomes apparent that is based on the structure
of the electrophile. Conjugated aldehydes bearing aromatic and
olefinic substituents show high reactivity in this aldol reaction.
No differences were observed between the reactions of electron-
rich and electron-poor conjugated aldehydes. However, aliphatic
aldehydes proved to be unreactive. The low reactivity of this
class of aldehydes had been reported in Lewis base-catalyzed
reactions of trichlorosilyl enol ethers and allylic trichlorosi-
lanes.16,18 When viewed in the larger context of aldol reactions
in general, this trend is puzzling because it stands in direct
The poor reactivity observed for crotonaldehyde in the aldol
reaction was surprising but can be explained using a similar
rationale. Again, the reduced concentration of the activated
aldehyde complex 28 translates into reduced reaction rate for
the aldolization. However, the low yields associated with this
9
J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005 3785

A R T I C L E S
Denmark et al.
reaction are likely due to the formation of trichlorosilyl ether
25. It is possible that side reactions involving the trichlorosilyl
enol ether 25 can lead to undesired products, contributing to
the reduced yield of the desired aldol product 9em.
decreases in the following order: 1a > 3 > 2a, which clearly
does not follow their reactivity (Figure 6). However, a clearer
relationship between the size of the alkoxy substituent and
enantioselectivity is apparent in the progressive improvement
from methyl to tert-butyl esters (Table 3).
Although the acetate-derived silyl ketene acetal 1a reacted
at a reasonable rate with aliphatic aldehydes, the tert-butyl pro-
panoate-derived silyl ketene acetal 2e proved completely un-
reactive. By using the less sterically demanding ethyl propan-
oate-derived silyl ketene acetal 2b and adding TBAI or i-Pr2EtN
in substoichiometric quantities, synthetically useful rates and
yields could be obtained (Table 5, entries 1 and 2). The bene-
ficial effect of the addition of tetrabutylammonium salts to the
reactions of other trichlorosilyl nucleophiles has been investigat-
ed by Berrisford and co-workers.59 It is believed that these addi-
tives enhance reactivity by increasing the ionic strength of the
solution. An increase in ionic strength should shift the R-chloro
trichlorosilyl ether equilibrium back toward the reactive, ionized
species (Scheme 10). The specific role of i-Pr2EtN in promoting
the reactions of aliphatic aldehydes remains unclear at this time.60
The marginal reactivity of aliphatic aldehydes is further
attenuated by the presence of R-substituents. Whereas linear
and â-branched aldehydes, such as 5k and 5o, retain some
reactivity under modified conditions, the R-branched substrate
cyclohexanecarboxaldehyde (5j) remains relatively unreactive,
most likely due to steric factors affecting approach of the
nucleophile (Table 1, compare entries 1, 2, and 7).
2. Selectivity Trends. 2.1. Trends in Enantioselectivity.
With regard to reactivity, steric effects play a dominant role in
this chiral phosphoramide/SiCl4 catalyst system. Examination
of the corresponding trends in enantio-, diastereo-, and regi-
oselectivity reveals a similar pattern. The most significant trend
with respect to enantioselectivity was the consistent observation
of aldol products derived from Re face attack on the aldehyde.
The ability of the catalyst (R,R)-6 to select for nucleophilic
attack at this prochiral face of the aldehyde regardless of
nucleophile structure even extends beyond the use of silyl ketene
acetals to reactions of other nucleophiles, such as allylic
stannanes,17 silyl enol ethers,61a and isocyanides.61b This uniform
facial selectivity suggests that it is the interaction between the
catalyst-bound aldehyde and the incoming nucleophile that is
the major contributor to the observed enantiofacial selectivity
in this family of Lewis base-catalyzed reactions.
The dramatic reactivity differences observed between con-
jugated and aliphatic aldehydes are also evident with respect to
diastereoselectivity. In almost all cases, the selectivity trend
observed for a particular nucleophile shows that 5a > 5g .
5k. It is only in the case of the dioxanone-derived silyl enol
ether 13 that a reversed trend is observed (Table 8).
2.2. Trends in Diastereoselectivity. The consistently high
anti diastereoselectivity is at once a great advantage synthetically
and also provides helpful insights into the overwhelming control
features in relative topicity of attack in the aldol addition.
Interestingly, it was observed that even the structure of the
catalyst had little effect as both HMPA and (R,R)-6 both gave
high anti diastereoselectivity. A rationale for this behavior will
be presented in detail in a subsequent section.
2.3. Trends in Regioselectivity. Although the reactivity of
silyl dienol ethers favors reaction at the terminal γ-position,
this selectivity is often variable, especially in the vinylogous
aldol reaction of simple ester-derived silyl dienol ethers.26 The
high regioselectivity observed under the influence of the chiral
phosphoramide/SiCl4 catalyst system appears to be a conse-
quence of the steric influence provided by the catalyst complex
during the aldolization. In a direct comparison of differentially
substituted methyl ester-derived silyl ketene acetals, there are
large rate differences based on the degree of R-substitution
(Figure 6). The reaction of a silyl dienol ether can then be
considered as an intramolecular competition where, in this case,
reaction generally occurs at the less substituted γ-position. This
can be observed in the additions of the ethyl crotonate-derived
silyl dienol ether 4a, wherein the question of regioselectivity
can be framed as an intramolecular competition between acetate-
like (C(4)) and propanoate-like (C(2)) silyl ketene acetals (Table
7, entries 1-3). A more heavily biased system is seen in the
case of the tiglate-derived silyl dienol ether 4b, wherein the
observed regioselectivity can be thought of as an intramolecular
competition between acetate-like (C(4)) and isobutyrate-like
(C(2)) silyl ketene acetals (Table 7, entries 4-6).
The problems encountered in the case of the 2-pentenoate-
derived silyl dienol ethers further illustrate this trend and validate
the hypothesis for the role of steric effects. Here, both the C(2)
and C(4) reactive centers are monosubstituted, propanoate-like
positions. In initial studies with the ethyl pentenoate-derived
silyl dienol ether, a near equal distribution of products was
observed. Only when a bulky tert-butyl ester substituent was
introduced could high regioselectivity be recovered (Table 7,
entries 10-12). The role of the alkoxy substituent at this point
remains unclear, but consideration of the conformation of the
With respect to nucleophile structure, two important contribu-
tors to enantioselectivity are (1) the presence of R-substituents
and (2) the size of the alkoxy substituent. The enantioselectivity
for the methyl ester-derived silyl ketene acetals 1a, 2a, and 3
(55) Guthrie, J. P. Can. J. Chem. 1978, 56, 962-973.
(56) Sander, E. G.; Jencks, W. P. J. Am. Chem. Soc. 1968, 90, 6154-6162.
(57) (a) Granovsky, A. A. (http://classic.chem.msu.su/gran/gamess/index.html).
(b) 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.;
Windus, T. L.; Dupuis, M.; Montgomery, J. A., Jr. J. Comput. Chem. 1993,
14, 1347-1363.
1
silyl dienol ether provides some insight. From H NMR nOe
(58) It is reasonable to assume that the aldol reaction occurs through the R-chloro
trichlorosilyl ether because products derived from reactions with aliphatic
aldehydes give similar levels of selectivity and the same sense of asymmetric
induction as that observed with other substrates.
(59) Short, J. D.; Attenoux, S.; Berrisford, D. J. Tetrahedron Lett. 1997, 38,
2351-2354.
(60) Initially, i-Pr2EtN was added to scavenge adventitious HCl in the reaction
mixture over extended reaction times. Thus, small amounts of the
ammonium chloride would be generated which could also alter the ionic
strength of the solution. Moreover, i-Pr2EtN may also facilitate catalyst
turnover as was seen in the reactions of allyltrichlorosilanes (ref 18).
(61) (a) Denmark, S. E.; Heemstra, J. R., Jr. Org. Lett. 2003, 5, 2303-2306.
(b) Denmark, S. E.; Fan, Y. J. Am. Chem. Soc. 2003, 125, 7825-7827.
studies, the geometry of the major isomer of 4d was determined
to be 1Z,3Z. If the pinwheel effect is considered,62 this would
place the alkoxy substituent in an s-cis position, directed toward
C(2), possibly accounting for the enhanced regioselectivity
observed in this case (Figure 7). This analysis, if true, will have
interesting consequences for regioselectivity in the vinylogous
aldol reactions of ketone- and aldehyde-derived silyl dienol ethers.
(62) Wilcox, C. S.; Babston, R. E. J. Org. Chem. 1984, 49, 1451-1453.
9
3786 J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005

Lewis Base Activation of Lewis Acids
A R T I C L E S
be the catalyst resting state, low temperature 29Si NMR
investigations, as outlined in Scheme 9, provided direct evidence
for the existence of this species. The spectroscopic detection
of this species provides the first direct evidence for the
intermediacy of a silyl cation in any Lewis base-catalyzed
process and lends support to ideas about the dual roles of both
the Lewis basic and Lewis acidic species in this catalytic
asymmetric aldol reaction.
Figure 7. Conformation of tert-butyl pentenoate-derived silyl dienol ether
4d.
2.4. Origin of Double Bond Geometries in the Products
of the Vinylogous Aldol Reaction. Despite the fact that all of
the silyl dienol ethers employed in these phosphoramide-
catalyzed vinylogous aldol reactions were mixtures of geo-
metrical isomers, high selectivity for the thermodynamically
favored (E)-R,â-unsaturated ester product was observed through-
out. Consideration of the silyl dienol ethers suggests that they
most likely exist in and react through the s-trans conformation
to avoid allylic strain in the s-cis conformation and to maximize
conjugation.63 This conformation should lead to the E product,
regardless of the geometry of the silyl dienol ether.
4. Mechanism of Stereoselection and Proposed Catalyst
Structure. 4.1. Rationale for the Observed Diastereoselec-
tivity. The high anti diastereoselectivity observed in these
reactions stands in stark contrast to that of most other Lewis
acid-catalyzed aldol reactions of silyl ketene acetals which
produce the syn diastereomer with high selectivity.8 Variable
levels of stereoconvergence with respect to silyl ketene acetal
geometry are observed with different catalyst systems. The trend
toward syn diastereoselectivity is also apparent in simple,
catalytic aldol reactions of silyl ketene acetals. Only two reports,
one on the aldol reactions of tert-butyl thiopropanoate-derived
silyl ketene acetals64 and another on the use of trityl perchlorate
as a catalyst for the additions of silyl enol ethers,65 consistently
provide the aldol adducts with synthetically useful levels of anti
diastereoselectivity. The fact that this reaction is anti diaste-
reoselective, as well as diastereoconvergent, is particularly
noteworthy and deserves comment.25
3. Proposed Mechanism. The proposed catalytic cycle
involves the intermediacy of a highly electrophilic, phosphora-
mide-bound silyl cation. As discussed in the Introduction, the
active species in these reactions is a chiral Lewis acid. HoweVer,
1
these are not Lewis acid-catalyzed reactions. H NMR spec-
troscopic analysis of reaction mixtures revealed that the direct
products of aldolization are trichlorosilyl ethers rather than the
tert-butyldimethylsilyl ethers that would be formed by a process
that would be catalytic in silicon tetrachloride. Examination of
the catalytic cycle shown in Figure 8 reveals that each molecule
of SiCl4 that enters the catalytic cycle ends up incorporated into
the product, while the Lewis basic phosphoramide catalyst is
released to participate in subsequent turnovers. Therefore, these
reactions are phosphoramide-catalyzed and SiCl4-mediated.
The cycle is initiated by binding of the phosphoramide to
the weakly Lewis acidic SiCl4. This binding leads to polarization
and eventual ionization of a chloride ion, generating a chiral,
trichlorosilyl cation i. This species can then bind the aldehyde
to form the complex ii. Depending on the nature of the bound
aldehyde, this intermediate may be in equilibrium with the
unreactive chlorohydrin species iii. Attack of the silyl ketene
acetal on the activated aldehyde then generates a species, which
after cleavage of the tert-butyldimethylsilyl group and dissocia-
tion of the catalyst, forms the product as the trichlorosilyl ether.
The key species in this cycle is the putative silyl cation.
Because initial kinetics studies suggested that this species might
The analysis of the diastereoselectivity of this reaction
requires, first, knowledge of the structure of the active enolate
equivalent. Because no trans-silylation to a reactive trichlorosilyl
enol ether is observed with 2b and because the direct products
of the reaction are clearly trichlorosilyl ethers rather than tert-
butyldimethylsilyl ethers, the reaction should proceed through
an open transition structure. Furthermore, because the silyl
ketene acetal does not lose its stereochemical integrity under
the reaction conditions, this analysis must also explain the
diastereoconvergence of both (E)- and (Z)-2e.
With this in mind, several open transition structures were
considered to rationalize the diastereoselectivity of this reaction.
In each enantiomorphic series, six possible transition structures
can be envisioned for each silyl ketene acetal isomer. Within
those six structures, there are three ((+)-synclinal, antiperiplanar,
(-)-synclinal) that lead to the anti product and three that lead
to the syn product. These transition structures can be further re-
fined if the conformation of the silyl ketene acetal is also con-
Figure 8. Proposed catalytic cycle.
9
J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005 3787

A R T I C L E S
Denmark et al.
sidered. According to Wilcox et al., the geometry of the double
bond and the position of the R-substituent are the relevant factors
in understanding the conformation of these molecules (Figure 9).62
substituent lies in the sector adjacent to the bound silyl cation
((Z)-apanti and (E)-apanti). The reverse is true for the syn pathway
((Z)-apsyn and (E)-apsyn). In his studies of the additions of silyl
enol ethers to aldehydes catalyzed by trityl perchlorate, Mu-
kaiyama et al. rationalized the high level of anti selectivity
obtained for both the E- and Z-isomers of the tert-butyldim-
ethylsilyl enol ether of 3-pentanone on the basis of interaction
between the R-substituent and the bound trityl cation.65 Mo-
lecular models of the silyl cation/phosphoramide catalyst
complex (vide infra) attest to the size of the ligated silyl cation.
Therefore, we propose that it is the interaction between the
R-substituent and the bound silyl cation complex that is
responsible for the high degree of anti diastereoconvergence
observed in these aldol reactions.
This rationalization gains support from the precipitous loss
in reactivity that occurs when an R,R-disubstituted silyl ketene
acetal is employed (Figure 6). Under the assumption that the
sector containing the silyl cation complex is the most sterically
congested, there are only two transition structures that place
the R-hydrogen substituent on the silyl ketene acetal in this
position for the approach of (E)-2e, (E)-scsyn, and (E)-apanti
(Figure 9). Exchange of this R-hydrogen for a methyl group,
as would be the case upon changing from the propanoate-derived
2e to the isobutyrate-derived 3, forces a substituent into that
region of space and accounts for the significant loss in reactivity.
Examination of the influence of this additional methyl group
on any of the other transition structures reveals that it would
appear in a relatively uncongested sector and the loss in
reactivity for 3 would be difficult to explain.
Figure 9. Possible open transition structures for the reaction of (E)-2e.
The next step in the analysis is to narrow down the number
of possible transition structures that are consistent with the
observations. The (+)-synclinal (+sc) and (-)-synclinal (-sc)
transition structures were eliminated on the basis of unfavorable
dipole-dipole interactions, as explained by Heathcock et al. for
the diastereoselectivity in silyl enol ether aldol reactions.66 This
hypothesis is supported by stereochemical studies from these
laboratories in which a rigid model shows a clear preference for
the antiperiplanar (ap) transition structure in the presence of a
wide variety of Lewis acid catalysts.67 Although no cationic Lewis
acids were considered in these studies, the validity of extending
this analysis will be borne out by other observations based on
steric rather than electronic considerations (vide infra). There-
fore, two ap transition structures, one leading to the anti and
another to the syn product, had to be considered for the reactions
of (E)- and (Z)-silyl ketene acetals, respectively (Figure 10).
The antiperiplanar transition structure is also helpful in
understanding the change in the degree of diastereoconvergence
upon changing from an ethyl to a tert-butyl propanoate-derived
silyl ketene acetal. Although the major difference between the
apanti and apsyn transition structures in either the Z- or E-series
is the position of the R-substituent, a second difference is also
apparent. This difference, based on careful consideration of the
pinwheel conformation, was originally suggested by Gennari
and co-workers in studies of the diastereoconvergent aldol
reactions of thioketene acetals and appears applicable to this
system.64 In the Z-series, there is a gauche interaction between
the alkoxy group and the aldehyde R group in (Z)-apsyn that is
absent in (Z)-apanti (Figure 10). A similar gauche interaction is
also present the E-series, but now it is between the silyloxy
substituent and the aldehyde R group in (E)-apsyn. On the basis
of this analysis, a decrease in the steric demands of the alkoxy
substituent would have a greater affect on the degree of anti
diastereoselectivity in the Z-series than in the E-series. This is
consistent with the observed trends in diastereoselectivity, where
a lesser degree of diastereoconvergence is observed with 2b as
compared to 2e, lending further support to the conclusion that
an apanti transition structure is operative in this aldol addition
(Table 3, entries 2 and 3).
Figure 10. Stereochemical analysis of the aldol reaction of silyl ketene
acetals.
The most obvious difference between the transition structures
leading to the syn and those leading to anti is the position of
the R-substituent. In the transition structures for the anti pathway
from both (Z)- and (E)-silyl ketene acetals, the R-methyl
The assumption that the phosphoramide-bound trichlorosilyl
cation complex acts as an extremely large Lewis acid not only
rationalizes the observed anti diastereoconvergence but also is
consistent with earlier conclusions regarding the sensitivity of
this catalyst system to the steric demands of the incoming
nucleophile.
(63) Eliel, E. L.; Wilen, S. H.; Mander, L. N. In Stereochemistry of Organic
Compounds; John Wiley and Sons: New York, 1994; p 567.
(64) Gennari, C.; Bernardi, A.; Cardani, S.; Scolastico, C. Tetrahedron Lett.
1985, 26, 797-800.
(65) Kobayashi, S.; Murakami, M.; Mukaiyama, T. Chem. Lett. 1985, 1535-
1538.
(66) Heathcock, C. H.; Davidsen, S. K.; Hug, K. T.; Flippin, L. A. J. Org. Chem.
1986, 51, 3027-3037.
(67) Denmark, S. E.; Lee, W. J. Org. Chem. 1994, 59, 707-709.
4.2. Rationale for the Observed Enantioselectivity. Al-
though the diastereoselectivity in this reaction can be rationalized
9
3788 J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005

Lewis Base Activation of Lewis Acids
A R T I C L E S
through use of general models for the aldol transition structure,
more information about the structure of the catalyst complex is
required to rationalize the high enantioselectivity. Despite
numerous attempts at crystallization, a structure determination
for the catalyst complex is lacking. However, from 29Si NMR
studies aided by computer modeling, some predictions can be
made that are consistent with the observed trends in enantiose-
lectivity.
approach through (E)-apanti. Smaller, unbranched R-substituents,
such as the ethyl substituent in 10a or the isobutyl substituent
in 10c, can be rotated into a gauche orientation that would direct
them away from an unfavorable interaction with the catalyst.
However, the isopropyl substituent in 10b cannot be rotated
into a conformation that avoids destabilizing approach through
an (E)-apanti transition structure.
In aldehydes bearing R-substituents, unfavorable interactions
with the catalyst system may be responsible for reduced
selectivity. For cyclohexanecarboxaldehyde, this unfavorable
interaction is particularly clear. The preferred conformation of
R-branched aldehydes has one of the alkyl substituents eclipsed
with the carbon-oxygen double bond.71 Therefore, the remain-
der of the ring system experiences steric interactions with the
catalyst backbone which allow a lower energy, syn-selective
pathway for nucleophilic attack.
Calculations on a complex of benzaldehyde and the dimeric
phosphoramide (R,R)-6-bound trichlorosilyl cation were per-
formed using the PC version of GAMESS(US) QC package
employing the PM3 basis set. The aldehyde is bound trans to
one of the phosphoramides, consistent with a consideration of
the nature of the hypervalent bonds in the ligand field around
silicon.68 This constellation places the aldehyde against one of
the binaphthyl units. In this model, the exposure of the Re face
of the aldehyde is determined by two factors: interaction with
the N-methyl group of the 1,1′-binaphthyl-2,2′-diamine back-
bone and interactions with one of the naphthyl rings of the 1,1′-
binaphthyl-2,2′-diamine. The N-methyl groups, previously
thought to play little role in the reactivity or selectivity of the
catalyst, protrude far into the binding pocket. This substituent
effectively blocks the approach of the nucleophile from the Si
face (Figure 11). The structure also suggests the possibility of
a stabilizing, edge-to-face π-π interaction for this conforma-
tion.69 This interaction may help to rationalize the higher
selectivity observed for conjugated compared to that for aliphatic
aldehydes.70
Conclusion
The use of chiral Lewis acids generated by Lewis base
catalysis has allowed for the development of a novel and general
aldol reaction between a wide range of silyl ketene acetals and
aldehydes. The putative, in situ-generated chiral silyl cation
provides high levels of diastereo- and enantioselectivity. This
process represents one of the few examples of an anti stereo-
convergent aldol reaction of silyl ketene acetals. Furthermore,
the extension of this method to silyl dienol ethers derived from
esters shows exceptionally high levels of regioselectivity in
addition to comparable diastereo- and enantioselectivities.
Analysis of trends in reactivity and selectivity shows that the
catalyst binding pocket is congested, and steric considerations
are predominant influences on diastereo- and enantioselectivity.
Conclusions drawn from computational studies of the aldehyde-
bound catalyst complex support the hypothesis of steric control.
Considering that this chiral phosphoramide/SiCl4 catalyst system
has also proven to be useful in a number of other carbonyl
addition processes, it appears that the Lewis base-catalyzed
reactions of main group organometallic nucleophiles represent
a general method. Further studies are underway to optimize and
structurally define the catalyst structure and to further expand
the scope of Lewis base-catalyzed, carbon-carbon bond-
forming processes.
Figure 11. Calculated structure of the aldehyde-silyl cation complex
optimized in GAMESS(US) QC package at the PM3 level and visualized
using Chem3D, with selected hydrogens omitted for clarity.
Acknowledgment. We are grateful to the National Science
Foundation for generous financial support (NSF CHE0105205
and CHE 0414440). G.L.B. thanks Boehringer Ingelheim
Pharmaceuticals, Inc. and the Johnson and Johnson Pharma-
ceutical Research Institute for graduate fellowships. M.D.E.
thanks Glaxo-Smith Kline for a postdoctoral fellowship.
The shape of the binding pocket also helps to explain the
dramatic influence of branching R to the reactive center on the
silyl ketene acetal or the aldehyde. In the case of larger, branched
R-substituents on the nucleophile, interactions between such a
substituent and the adjacent naphthyl ring may disfavor the anti
Supporting Information Available: Experimental details for
the preparation and full spectroscopic and analytical character-
ization of all compounds reported along with coordinates for
the catalyst-aldehyde complex in Figure 11. This material is
available free of charge via the Internet at http://pubs.acs.org.
(68) Tandura, S. N.; Voronkov, M. G.; Alekseev, N. V. Top. Curr. Chem. 1986,
113, 99-189.
(69) (a) Jones, G. B. Tetrahedron 2001, 57, 7999-8016. (b) Janiak, C. J. Chem.
Soc., Dalton Trans. 2000, 3885-3896. (c) Hunter, C. A.; Sanders, J. K.
M. J. Am. Chem. Soc. 1990, 112, 5525-5534.
JA047339W
(70) Although these interactions are known to be weak, especially in nonpolar
media, they are enthalpically favorable: Smithrund, D. B.; Diederich, F.
J. Am. Chem. Soc. 1990, 112, 339-343.
(71) Karabatsos, G. J.; Verdini, A. S. Top. Stereochem. 1970, 5, 167-203.
9
J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005 3789