Stereoselective aldol additions of achiral ethyl ketone-derived trichlorosilyl enolates

Article abstract of DOI:10.1021/jo034092x

Methods for the preparation of geometrically defined enoxy(trichlorosilanes) derived from ethyl ketone enolates have been developed. The addition of enoxy(trichlorosilanes) (trichlorosilyl enolates) to aldehydes proceeds with good yields in the presence of catalytic amounts of chiral phosphoramides. The reaction of Z-trichlorosilyl enolates to aryl aldehydes affords aldol products with good to excellent diastereo- and enantioselectivities. Phosphoramide-catalyzed aldol additions lacked substrate generality providing modest selectivities with unsaturated and aliphatic aldehydes. In all cases, the phosphoramide-catalyzed aldol addition of E-trichlorosilyl enolates to aldehydes provided good yields with moderate to good stereoselectivities.

Full text of DOI:10.1021/jo034092x

Ster eoselective Ald ol Ad d ition s of Ach ir a l Eth yl Keton e-Der ived
Tr ich lor osilyl En ola tes
Scott E. Denmark* and Son M. Pham
Department of Chemistry, Roger Adams Laboratory, University of Illinois, Urbana, Illinois 61801
denmark@scs.uiuc.edu
Received J anuary 24, 2003
Methods for the preparation of geometrically defined enoxy(trichlorosilanes) derived from ethyl
ketone enolates have been developed. The addition of enoxy(trichlorosilanes) (trichlorosilyl enolates)
to aldehydes proceeds with good yields in the presence of catalytic amounts of chiral phosphoramides.
The reaction of Z-trichlorosilyl enolates to aryl aldehydes affords aldol products with good to excellent
diastereo- and enantioselectivities. Phosphoramide-catalyzed aldol additions lacked substrate
generality providing modest selectivities with unsaturated and aliphatic aldehydes. In all cases,
the phosphoramide-catalyzed aldol addition of E-trichlorosilyl enolates to aldehydes provided good
yields with moderate to good stereoselectivities.
In tr od u ction
developed. Although this burgeoning field is highly
promising, excessively long reaction times and narrow
substrate classes currently prevent its broad application.5
The aldol addition is among one of the most important
methods for the stereocontrolled construction of carbon-
carbon bonds.1 The utility and application of this trans-
formation is illustrated by the vast number of reviews
written on the subject over the past decade.2 1,3-Oxygen-
ated carbon chains as afforded by the aldol reaction are
common structural motifs observed in many natural
products.3 The formation of these carbon chains is typi-
fied by the addition of an enol or enolate to an aldehyde
to form two new stereogenic centers in the case of
substituted nucleophiles. Controlling the relative and
absolute configuration of the new stereogenic centers has
been the primary focus of most new developments in this
field. The nonenzymatic methods that are currently
utilized often provide the desired structural motifs with
excellent stereocontrol. Unfortunately, these processes
are typically restricted to the use of either stoichiometric
amounts of the chiral modifying agent, limited substrate
classes, or poor stereochemical correlation between the
starting enolate donor and aldol product.4 Stereoselective
routes employing unactivated ketones have also been
Ba ck gr ou n d
Recent reports from these laboratories have docu-
mented the utility of trichlorosilyl enol ethers as a new
(4) For examples of chiral auxiliaries see: (a) Evans, D. A.; Bartroli,
J .; Shih, T. L. J . Am. Chem. Soc. 1981, 103, 2127. (b) Crimmins, M.
T.; King, B. W.; Tabet, E. A.; Chaudary, K. J . Org. Chem. 2001, 66,
894. (c) Crimmins, M. T.; King, B. W.; Tabet, E. A. J . Am. Chem. Soc.
1997, 119, 7883. (d) Abiko, A.; Liu, J .-F.; Masamune, S. J . Am. Chem.
Soc. 1997, 119, 2586. For examples of chiral metalloids see: (e) Evans,
D. A.; Rieger, D. L.; Bilodeau, M. T.; Urp´ı, F. J . Am. Chem. Soc. 1991,
113, 1047. (f) Duthaler, R. O.; Hafner, A. Chem. Rev. 1992, 92, 807.
(g) Roos, G. H. P.; Haines, R. J .; Raab, C. E. Synth. Commun. 1993,
23, 1251. (h) Slough, G. A.; Bergman, R. G.; Heathcock, C. H. J . Am.
Chem. Soc. 1989, 111, 938. (i) Corey, E. J .; Kim. S. S. J . Am. Chem.
Soc. 1990, 112, 4976. (j) Fujimura, O. J . Am. Chem. Soc. 1998, 120,
10032. (k) Sodeoka, M.; Tokunoh, R.; Miyazaki, F.; Hagiwara, E.;
Shibasaki, M. Synlett 1997, 463. (l) Sodeoka, M.; Ohrai, K.; Shibasaki,
M. J . Org. Chem. 1995, 60, 2648. (m) Kru¨ger, J .; Carreira, E. M. J .
Am. Chem. Soc. 1998, 120, 837. For examples of chiral Lewis acids,
see: (n) Kiyooka, S.-I. Rev. Heteroat. Chem. 1997, 17, 245. (o) Parmee,
E. R.; Tempkin, O.; Masamune, S.; Abiko, A. J . Am. Chem. Soc. 1991,
113, 9365. (p) Furata, K.; Maruyama, T.; Yamamoto, H. J . Am. Chem.
Soc. 1991, 113, 1041. (q) Ishimaru, K.; Monda, K.; Yamamoto, H.;
Akiba, K.-Y. Tetrahedron 1998, 54, 727. (r) Carreira, E. M.; Singer, R.
A.; Lee, W. J . Am. Chem. Soc. 1994, 116, 8837. (s) Singer, R. A.;
Carreira, E. M. J . Am. Chem. Soc. 1995, 117, 12360. (t) Kim, Y.; Singer,
R. A.; Carreira, E. M. Angew. Chem., Int. Ed. 1998, 37, 1261. (u) Evans,
D. A.; Kozlowski, M. C.; Burgey, C. S.; MacMillan, D. W. C. J . Am.
Chem. Soc. 1997, 119, 7893. (v) Kobayashi, S.; Nagayama, S.; Busuji-
ma, T. Tetrahedron 1999, 55, 8739. (w) Yanagisawa, A.; Matsumoto,
Y.; Nakashima, H.; Asakawa, K.; Yamamoto, H. J . Am. Chem. Soc.
1997, 119, 9319. (x) Yamashita, Y.; Saito, S.; Ishitani, H.; Kobayashi,
S. J . Am. Chem. Soc. 2002, 124, 3292. (y) Ishitani, H.; Yamashita, Y.;
Shimizu, H.; Kobayashi, S. J . Am. Chem. Soc. 2000, 122, 5403.
(5) (a) Yoshikawa, N.; Yamada, Y. M. A.; Das, J .; Sasai, H.;
Shibasaki, M. J . Am. Chem. Soc. 1999, 121, 4168. (b) Yamada, Y. M.
A.; Yoshikawa, N.; Sasai, H.; Shibasaki, M. Angew. Chem., Int. Ed.
Engl. 1997, 36, 1871. (c) Yamada, Y. M. A.; Shibasaki, M. Tetrahedron
Lett. 1998, 39, 5561. (d) List, B.; Lerner, R. A.; Barbas, C. F., III J .
Am. Chem. Soc. 2000, 122, 2395. (e) Sakthivel, K.; Notz, W.; Bui, T.;
Barbas, C. F., III J . Am. Chem. Soc. 2001, 123, 5260. (f) Notz, W.;
List, B. J . Am. Chem. Soc. 2000, 122, 7386. (g) Trost, B. M.; Ito, H. J .
Am. Chem. Soc. 2000, 122, 12993. (h) Trost, B. M.; Silcoff, E. R.; Ito,
H. Org. Lett, 2001, 3, 2497. (i) Trost, B. M.; Ito, H.; Silcoff, E. R. J .
Am. Chem. Soc. 2001, 123, 3367.
* Corresponding author.
(1) Norcross, R. D.; Paterson, I. Chem. Rev. 1995, 95, 2041.
(2) (a) Carreira, E. M. In Modern Carbonyl Chemistry; Otera, J .,
Ed.; Wiley-VCH: Weinhiem, Germany, 2000; Chapter 8. (b) Palomo,
C.; Oiarbide, M.; Garc´ıa, J . M. Chem. Eur. J . 2002, 8, 37. (c)
Machajewski, T. D.; Wong, C.-H. Angew. Chem., Int. Ed. 2000, 39,
1352. (d) Carreira, E. M. In Comprehensive Asymmetric Catalysis;
J acobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer-Verlag:
Heidelberg, Germany, 1999; Vol. III, Chapter 29. (e) Saito, S.; Yama-
moto, H. Chem. Eur. J . 1999, 5, 1959. (f) Nelson, S. G. Tetrahedron:
Asymmetry 1998, 357. (g) Gro¨ger, H.; Vogl, E. M.; Shibasaki, M. Chem.
Eur. J . 1998, 4, 1137. (h) Paterson, I.; Cowden, C. J .; Wallace, D. J . In
Modern Carbonyl Chemistry; Otera, J ., Ed.; Wiley-VCH: Weinheim,
Germany, 2000; Chapter 9. (i) Cowden, C. J .; Paterson, I. Org. React.
1997, 51, 1. (j) Heathcock, C. H.; Kim, B. M.; Williams, S. F.;
Masamune, S.; Paterson, I.; Gennari, C. In Comprehensive Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford,
UK, 1991; Vol. 2.
(3) (a) Nicolaou, K. C.; Vourloumis, D.; Winssinger, N.; Baran, P.
S. Angew. Chem., Int. Ed. 2000, 112, 46. (b) Mukaiyama, T. Tetrahe-
dron 1999, 55, 8609.
10.1021/jo034092x CCC: $25.00 © 2003 American Chemical Society
Published on Web 06/03/2003
J . Org. Chem. 2003, 68, 5045-5055
5045

Denmark and Pham
class of aldol reagents.6 These species react spontane-
ously with aldehydes at or below ambient temperature
to afford aldol adducts in good to high diastereoselectivity
for substituted enolates. For unsubstituted enolates, e.g.
methyl ketone enolate 1, reaction with various aldehydes
occurs cleanly at room temperature to afford aldol
adducts in 86-93% yield in the absence of external
agents.
Most importantly, however, these reactions are ac-
celerated by the addition of catalytic quantities of Lewis
bases, in particular chiral phosphoramides, Scheme 1.7
By the use of these Lewis basic additives, excellent
enantioselectivities can be achieved from a variety of
methyl and cyclic ketone enolates with various aldehydes.
A distinct advantage over the use of chiral Lewis acids
is that the geometric configuration in conformationally-
restricted, substituted enolates is transferred to the
relative diastereoselectivity of the aldol products. Thus,
reactions of trichlorosilyl enolates are believed to proceed
through closed transition structures organized around
silicon. This report details our current investigations
regarding the stereoselective preparation of achiral trichlo-
rosilyl enolates derived from ethyl ketones and their
subsequent additions to achiral aldehyde acceptors.
to the corresponding trichlorosilyl enol ether was initi-
ated. The metal catalysts were evaluated on the basis of
their effectiveness in catalyzing the trans-silylation and
if applicable the level of geometric control in the trichlo-
rosilyl enol ethers. To more effectively evaluate the level
of geometric control, it was desirable to begin with highly
enriched E- or Z-trimethylsilyl enol ethers. Thus, 4 was
prepared in a 30/1 E/Z mixture by using lithium tetra-
methylpiperidide following the method of Collum.9 The
Z-isomer was prepared without purification according to
the method of Evans10 through the formation of the boron
enolate with dibutylboron triflate11 and subsequent trap-
ping with chlorotrimethylsilane (TMSCl). Although the
TMS enol ether was generated with high Z-selectivity
based upon 1H NMR analysis, attempts to isolate and
purify it proved unsuccessful. A survey of metal salts and
metal complexes was therefore undertaken with the
E-enolate. As a general protocol, TMS enol ether 4
(30/1, E/Z) was added dropwise to a mixture of SiCl4 and
metal catalyst in CH2Cl2 at room temperature, Table 1.
1
The progress of the reaction was easily monitored by H
NMR by using the diagnostic signals at δ 0.2 (TMS-O)
and 0.4 (TMS-Cl).
TABLE 1. Su r vey of Meta l Sa lt Ca ta lysts for th e
Tr a n s-Silyla tion of 4 to 5
SCHEME 1
MXn
time, h
result
(E/Z)a
Hg(OAc)2
Pd(OAc)2
PdBr2
Rh2(OAc)4
NiCl2
PtCl2
(PhCN)2PtCl2
ZnCl2
2
2
trans-silylation
trans-silylation
isomerization
isomerization
isomerization
isomerization
isomerization
isomerization
5, 1/2
5, 1/6
4, 1/3
4, 1/6
4, 1/5
4, 1/8
4, 1/4
4, 5/1
24
24
24
24
24
24
b
a
b
Determined by 1H NMR analysis. Used 2.5 mol % of the
Resu lts
metal catalyst.
A. P r ep a r a tion of En ola tes. The preparation of
methyl and cyclic ketone trichlorosilyl enolates by metal-
catalyzed trans-silylation has been described in detail.8
In these foregoing studies the geometry of the resultant
trichlorosilyl enolate either was fixed within a ring or
was not an issue as in the case of methyl ketones. Thus,
the effect of the reaction conditions in controlling the
geometry of trichlorosilyl enol ethers had not been of
concern. To begin this investigation, an exhaustive survey
of metal salts and metal complexes for the trans-silyla-
tion of 3-pentanone-derived trimethylsilyl enol ether 4
Interestingly, only the mercury and palladium acetate
salts were able to affect trans-silylation and afforded
predominantly the Z-trichlorosilyl enolate. It is important
to note that isomerization of the trimethylsilyl enol ether
1
was observed by H NMR spectroscopy at early reaction
times for both Hg(OAc)2 and Pd(OAc)2. Although this was
an unexpected result, it provided an important insight
into the mechanism of metal-catalyzed trans-silylation.
The remainder of the metal salts evaluated in this
initial survey could only affect isomerization of the TMS
enol ether to afford predominantly Z-TMS enol ether
without participating in trans-silylation. Although previ-
ous attempts to selectively prepare pure samples of (Z)-4
were unsuccessful, the results from the preliminary
catalyst survey could be used advantageously to prepare
the opposite geometric isomer. Thus, the 30/1, E/Z
(6) (a) Denmark, S. E.; Stavenger, R. A. Acc. Chem. Res. 2000, 33,
432. (b) Denmark, S. E.; Stavenger, R. A.; Wong, K.-T.; Su, X. J . Am.
Chem. Soc. 1999, 121, 4928. (c) Denmark, S. E.; Wong, K.-T.;
Stavenger, R. A. J . Am. Chem. Soc. 1997, 119, 2333. (d) Denmark, S.
E.; Winter, S. B. D.; Su, X.; Wong, K.-T. J . Am. Chem. Soc. 1996, 118,
7404.
(7) (a) Denmark, S. E.; Stavenger, R. A. J . Am. Chem. Soc. 2000,
122, 8837. (b) Denmark, S. E.; Stavenger, R. A.; Wong, K.-T. J . Org.
Chem. 1998, 63, 918.
(8) (a) Denmark, S. E.; Stavenger, R. A.; Wong, K.-T.; Su, X. J . Am.
Chem. Soc. 1999, 121, 4982. (b) Denmark, S. E.; Stavenger, R. A.;
Winter, S. B. D.; Wong, K.-T.; Barsanti, P. A. J . Org. Chem. 1998, 63,
9517.
(9) Hall, P. L.; Gilchrist, J . H.; Collum, D. B. J . Am. Chem. Soc.
1991, 113, 9571.
(10) Evans, D. A.; Nelson, J . V.; Vogel, E.; Taber, T. R. J . Am. Chem.
Soc. 1981, 103, 3099.
(11) Mukaiyama, T.; Inoue, T. Bull Chem. Soc. J pn. 1980, 53, 174.
5046 J . Org. Chem., Vol. 68, No. 13, 2003

Achiral Ethyl Ketone-Derived Trichlorosilyl Enolates
TABLE 3. Con ver gen ce of Geom etr ic Selectivity
mixture of trimethylsilyl enol ethers was isomerized and
purified to afford a 1/4, E/Z mixture with PtCl2.
For the remainder of the survey, a 3/1, E/Z mixture of
trimethylsilyl enol ethers, prepared with lithium diiso-
propylamide (LDA), was used due to the observed isomer-
ization prior to trans-silylation. Summarized in Table 2
are the results for all successful trans-silylations at-
tempted with various metal salts and metal complexes.
With the exception of the trans-silylation with Pd(OAc)2,
all reactions were heterogeneous.
MXn
4, E/Z
yield, %a
5, E/Zb
Hg(OAc)2
Hg(OAc)2
Hg(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(TFA)2
Pd(TFA)2
30/1
3/1
1/4
30/1
3/1
1/4
65
72
60
62
76
69
70
43
1/2
1/2
1/2
1/6
1/6
1/6
1/6
1/6
TABLE 2. Resu lts fr om Su ccessfu l Tr a n s-Silyla tion s
3/1
1/4
a
b
Yield of distilled material. Determined by 1H NMR analysis.
MXn
time, h
yield, %a
5, E/Zb
Hg(OAc)2
Hg(TFA)2
HgCl2
Pd(OAc)2
Pd(TFA)2
Tl(TFA)3
5
5
72
60
31
76
70
20
1/2
1/2
1/2
1/6
1/7
1/3
the final ratio of geometrical isomers is metal dependent
and under kinetic control, but that there is also a
preequilibrium that causes the starting materials to
converge (vide infra).
48c
5
15
24c
Having completed a general survey of the trans-
silylation of 3-pentanone TMS enol ether 4 to trichlo-
rosilyl enolate 5, the scope and efficiency of mercury- and
palladium-catalyzed trans-silylation was then studied
with use of various ethyl ketone-derived TMS enol ethers.
To evaluate a variety of steric demands for the spectator
group (the nonparticipating and nonenolized substituent
from the corresponding ketone), butanone (methyl),
2-methyl-3-pentanone (isopropyl), 2,2-dimethyl-3-pen-
tanone (tert-butyl), and propiophenone (phenyl) were
subsequently studied. Additionally, propanal was in-
cluded as a test substrate where the spectator group is
simply hydrogen.
The trimethylsilyl enol ethers of propanal (6) and
2-butanone (7) were prepared by treatment of the car-
bonyl compounds with chlorotrimethylsilane and triethyl-
amine in hot DMF as developed by House.12 Unfortu-
nately the products contained significant amounts of
hexamethyldisiloxane that could not be removed by
fractional distillation. The desired compounds were ul-
timately purified sacrificially with use of silica gel
chromatography. Trimethylsilyl enol ether 8 was pre-
pared by enolization with LDA followed by subsequent
trapping of the lithium enolate with TMSCl. The bulkier
substrates 9 and 10 were prepared by enolization and
trapping with in-situ-generated iodotrimethylsilane in
acetonitrile.
a
b
Yield of distilled material. Determined by 1H NMR analysis.
c Trans-silylations were incomplete.
Of the metal catalysts tried, only Hg(OAc)2, Hg(TFA)2,
Pd(OAc)2, and Pd(TFA)2 affected complete trans-silyla-
tion. The use of PdCl2, Pd(dba)2, Ru(COD)Cl2, Rh(COD)-
Cl, CdBr2, Zn(OAc)2, and CsCl, led only to isomerization
of the trimethylsilyl enol ether. On the other hand, the
use of Pd(COD)Br2, Ni(OAc)2, (CH3CN)2PtCl2, AgOAc,
AgTFA, CdI2, AlCl3, Ph2GeCl2, LaCl3‚7H2O, SnCl2, SnCl4,
Me2SnCl2, Ph2SnCl2, and Pb(OAc)2 resulted in hydrolytic
protodesilylation of the trimethylsilyl enol ether to afford
3-pentanone (vide infra). Finally, HgCl2, Tl(TFA)3, Pd-
(PPh3)4, TiCl4, and BCl3 afforded small amounts of
trichlorosilyl enol ether, but the rate of trans-silylation
was extremely slow compared with that of the other
metal salts.
Interestingly, starting with either a 30/1 or 4/1, E/Z
mixture of trimethylsilyl enol ethers, trans-silylation
afforded predominantly (Z)-5 in a 1/2 or 1/6, E/Z mixture
for mercury and palladium catalysts, respectively. This
suggests that although an equilibration between the E/Z
isomers exists, the final geometric selectivities for metal-
catalyzed trans-silylations of this type are kinetically
controlled with the ratio being dictated by the metal
employed. Metatheses were complete within 5 h when 5
mol % of Hg(OAc)2, Hg(TFA)2, and Pd(OAc)2 were used.
However, Pd(TFA)2 was less effective, requiring 15 h for
complete conversion. It is notable to mention that certain
batches of commercially available Pd(TFA)2 were able to
generate (Z)-5 in as high as 1/16, E/Z. Attempts to clarify
these results by either directly preparing Pd(TFA)2 or
recrystallizing commercial material were unsuccessful.
Although these results were not the norm, they were
applied to the selective preparation of (Z)-5 used in
subsequent sections.
The various TMS enol ethers were each subjected to
trans-silylation with 5 mol % of Hg(OAc)2, Pd(OAc)2, and
Pd(TFA)2, Table 4. Curiously, the rates of metathesis in
all cases were slower than that of enol ether 4. In fact,
initial attempts to affect trans-silylation of 9 and 10
resulted in significant hydrolysis of the product during
the long reaction times. However, it was determined that
the levels of hydrolysis could be minimized by increasing
the amounts of metal catalyst to 10 mol %.
The results in Table 4 show that the size of the
spectator group has a profound influence on the geomet-
When starting with a 1/4, E/Z mixture of 4, metal-
catalyzed trans-silylation afforded 5 with the same
selectivity as when either a 30/1 or 3/1, E/Z mixture was
used, Table 3. Taken together, these results indicate that
(12) House, H. O.; Czuba, L. J .; Gall, M.; Olmstead, H. D. J . Org.
Chem. 1969, 34, 2324.
J . Org. Chem, Vol. 68, No. 13, 2003 5047

Denmark and Pham
TABLE 5. Elem en ta l An a lysis of th e Meta l Resid u e
Ta k en F ollow in g Tr ea tm en t w ith SiCl4
ric composition of the trichlorosilyl enolate. Relatively
large substituents such as tert-butyl and phenyl afford
the Z isomer almost exclusively, for all metals used,
whereas smaller substituents such as methyl exert little
to no bias on the stereochemical outcome. Interestingly,
when the spectator group is H, the selectivities are
improved relative to methyl regardless of the metal
employed.
M(L)2
time
results, % composition of M(X)2
Hg(OAc)2 5 min C, 0.42; H, 0.21; Cl, 17.17; Hg, 64.09
theoretical for HgCl2:
C, 0.00; H, 0.00; Cl, 26.12; Hg, 73.88
theoretical for Hg(OAc)2:
C, 15.08; H, 1.90; Hg, 62.94
TABLE 4. Tr a n s-Silyla tion of Va r iou s Keton e-Der ived
TMS En ol Eth er s to Tr ich lor osilyl En ola tes
Pd(OAc)2 5 min C, 1.26; H, 0.52; Cl, 31.92; Pd, 51.07
theoretical for PdCl2:
C, 0.00; H, 0.00; Cl, 39.99; Pd, 60.01
theoretical for Pd(OAc)2:
C, 21.40; H, 2.69; Pd, 47.40
Pd(TFA)2 5 min C, 11.86; H, 0.81; Cl, 0.00; F, 16.03; Pd, 35.75
theoretical for PdCl2:
C, 0.00; H, 0.00; Cl, 39.99; Pd, 60.01
theoretical for Pd(TFA)2:
C, 14.45; H, 0.00; F, 34.29; Pd, 47.40
C, 0.12; H, 0.14; Cl, 37.46; F, 0.21; Pd, 55.17
theoretical for PdCl2:
C, 0.00; H, 0.00; Cl, 39.99; Pd, 60.01
theoretical for Pd(TFA)2:
enol
entry ether
R
MXn
enolate time, h yield, %a Z/Eb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
6
6
6
7
7
7
4
4
4
8
8
8
9
9
9
H
H
H
Hg(OAc)2
Pd(OAc)2
Pd(TFA)2
11
11
11
12
12
12
5
16
16.5
36
16.5
16.5
24
5
5
15
18
18
20
24
29.5
22
18
23
22
50
25
35
58
73
53
72
76
70
65
58
54
55
50
55
66
65
55
8/1
8/1
4/1
2/1
3/1
8/1
2/1
6/1
7/1
Pd(TFA)2 12 h
c
Me Hg(OAc)2
Me Pd(OAc)2
Me Pd(TFA)2
Et
Et
Et
i-Pr Hg(OAc)2
i-Pr Pd(OAc)2
i-Pr Pd(TFA)2
t-Bu Hg(OAc)2
t-Bu Pd(OAc)2
t-Bu Pd(TFA)2
C, 14.45; H, 0.00; F, 34.29; Pd, 47.40
Hg(OAc)2
Pd(OAc)2
Pd(TFA)2
5
5
ence of basic compounds during trichlorosilylation, the
lithium enolates were generated amine free via addition
of MeLi to the corresponding TMS enol ether.13 The rate
of silyl cleavage can be monitored by NMR spectroscopy,
noting the disappearance of the O-TMS peak at roughly
δ 0.2 ppm. Gratifyingly, the substrate was not subject to
isomerization during this procedure and could provide
(E)-5 in essentially the same geometric mixture as the
corresponding E-TMS enol ether with typical yields
between 60 and 75%, Scheme 2.
13
13
13
14
14
14
15
15
15
8/1
9/1
c
17/1
>20/1
>20/1
>20/1
99/1
99/1
99/1
c
c
c
10 Ph
10 Ph
10 Ph
Hg(OAc)2
Pd(OAc)2
Pd(TFA)2
c
c
Yield of distilled material. Ratio determined by 1H NMR
a
b
analysis. c Using 10 mol % of the metal salt.
SCHEME 2
To better understand the role of the counterion in the
mercury- and palladium-catalyzed trans-silylation, the
fate of the metal species during the course of the reaction
needed to be determined. Thus, Hg(OAc)2, Pd(OAc)2, and
Pd(TFA)2 were independently treated with 20 equiv of
SiCl4 under standard reaction conditions to elucidate the
nature of the catalytic species. After evaporation, the
resulting residues were then isolated, washed with CH2-
Cl2, and dried for elemental analysis. In each case,
analysis revealed little to no carbon and high levels of
chloride, suggesting the catalytic species may be the
corresponding metal chlorides. Interestingly, both
Hg(OAc)2 and Pd(OAc)2 completely exchange to the
chloride salts within 5 min of treatment with SiCl4 as
inferred by the low percentages of carbon and hydrogen
observed in entries 1 and 2, Table 5. However, unlike
Pd(OAc)2, Pd(TFA)2 required several hours to fully
exchange to PdCl2, compare entries 3 and 4. Even so,
exchange to the chloride adduct is significantly faster
than trans-silylation.
As with (E)-5, (E)-12 was prepared by trapping of the
configurationally defined lithium enolate generated from
(E)-7. Initial attempts to obtain (E)-7 by chromatographic
separation of mixtures of (E)- and (Z)-7 resulted in
hydrolysis to the ketone. Following the method of Vedejs,
(E)-7 could be generated in as high as 22/1, E/Z.14
Unfortunately, yields for this carbenoid rearrangement
were typically poor, between 35 and 45%. Nonetheless,
(E)-7 could be prepared and was treated with MeLi to
generate the amine-free lithium enolate in situ. Trapping
with 10 equiv of SiCl4 afforded a 65% yield of the
corresponding trichlorosilyl enolate with a 20/1, E/Z ratio,
Scheme 3.
B. Ald ol Ad d it ion s. Optimization studies of (Z)-5
were conducted with benzaldehyde as the aldehyde ac-
ceptor. Additions were done at 10 and 15 mol % catalyst
loadings at -78 °C in CH2Cl2, Table 6, entries 1 and 2.
Although it was hoped that E-trichlorosilyl enolates
could be prepared via a similar metal-catalyzed trans-
silylation, the observation of stereoconvergence from both
(E)- and (Z)-4 deemed this unlikely. Alternatively, at-
tempts at direct trichlorosilylation with either SiCl4 or
SiCl3OTf from lithium enolates generated with amide
bases afforded intractable mixtures. To avoid the pres-
(13) Denmark, S. E.; Ghosh, S. K. Angew. Chem., Int. Ed. 2001, 40,
4759.
(14) Vedejs, E.; Larsen, S. D. J . Am. Chem. Soc. 1984, 106, 3030.
5048 J . Org. Chem., Vol. 68, No. 13, 2003

Achiral Ethyl Ketone-Derived Trichlorosilyl Enolates
SCHEME 3
though a slight improvement from 79 to 83% yield was
observed for reactions run with 20 and 50 mol % of the
catalyst as opposed to 5 mol % of the catalyst. Likewise,
enantioselectivities were significantly improved from 5/1
to 8/1 er when the catalyst loadings were increased from
5 to 20 mol %. Curiously, the diastereoselectivities
remained relatively constant at roughly 6/1, syn/anti
throughout the loading study. In any case, no improve-
ment was observed when increasing the amount of
catalyst beyond 20 mol %.
Although reactions with 10 mol % of the catalyst were
equally stereoselective, the survey was performed with
15 mol % of the catalyst because benzaldehyde is often
the most reactive electrophile. Indeed, the addition of 5
to trans-cinnamaldehyde was incomplete after 2 h. A brief
investigation revealed that reactions with aldehydes
other than benzaldehyde, Chart 1, required 8 h for
complete conversion. This time scale would be used in
all subsequent aldol additions.
TABLE 7. Ald ol Ad d ition of (Z)-12 to Ben za ld eh yd e,
Usin g Va r yin g Am ou n ts of P h osp h or a m id e 2
CHART 1
loading, mol %
dr, syn/antia
er (syn)b
yield, %c
2.5
5
10
15
20
4/1
5/1
5/1
10/1
7/1
16/1
15/1
1.2/1
1.2/1
1.2/1
1.5/1
1.7/1
28/1
53
75
79
65
86
82
85
50
100
21/1
a
b
Determined by 1H NMR analysis. Determined by CSP SFC
analysis. c Yield of chromatographically homogeneous material
containing a mixture of diastereomers.
To study the generality with respect to the aldehyde
donor, trichlorosilyl enolate (Z)-5 was combined with
various aldehydes in the presence of 15 mol % of (S,S)-2
and the reaction was quenched after 8 h. With the
exception of cyclohexanecarboxaldehyde the yields in all
cases were good. Unfortunately, both diastereoselectivi-
ties and enantioselectivities were greatly attenuated with
respect to benzaldehyde, especially when employing
nonaryl-based aldehydes such as trans-cinnamaldehyde
and furfuraldehyde, Table 6.
To evaluate the influence of the spectator group, aldol
additions were carried out with 2-butanone enolate
(Z)-12. A reaction with 10 mol % of the catalyst with
benzaldehyde revealed that both diastereo- and enantio-
selectivities were attenuated compared to (Z)-5. Attempts
to optimize this reaction with benzaldehyde further
revealed a dramatic catalyst loading effect unobserved
with the 3-pentanone-derived enolate. Although the
reaction is responsive to the effects of the chiral catalyst,
at least 50 mol % of the phosphoramide was required to
provide good stereoselectivities. Certainly nonideal, this
is, however, a dramatic improvement over reactions run
at standard catalyst loadings with (Z)-12. Further at-
tempts to improve stereoselectivities by modifying the
chiral catalyst will be discussed later.
TABLE 6. Su r vey of Va r iou s Ald eh yd es in th e
Ster eoselective Ald ol Ad d ition of (Z)-5 Ca ta lyzed by 2
Aldol additions with E-enolates were performed in the
same manner as the Z-enolates as described above.
Beginning with (E)-5, a survey of aldehydes was con-
ducted at 1 M concentration in CH2Cl2, using 15 mol %
of catalyst, and were quenched after 8 h. The results are
summarized in Table 8. Interestingly, the results for the
aldol addition to benzaldehyde were much poorer when-
compared to those with (Z)-5. Not only was the enantio-
selectivity attenuated, but the diastereoselectivity was
among the lowest of the group. Considering the remain-
ing unsaturated aldehydes, the results are similar to
those obtained with (Z)-5, with aryl aldehydes affording
higher selectivities, although now with the anti product
being favored. Again with the exception of cyclohexane-
carboxaldehyde, the yields in all cases were good.
In light of the results obtained for (Z)-12, a loading
study was performed for (E)-12, using 25 and 50 mol %
of catalyst, to determine if stereoselective aldol additions
were viable under the current conditions. Benzaldehyde
loading,
mol %
dr,
er
yield,
%
c
entry aldehyde
product syn/antia (syn)b
1
2
3
4
5
6
a
a
b
c
d
e
10
15
15
15
15
15
16a
16a
16b
16c
16d
16e
16/1
16/1
8/1
4/1
5/1
19/1
21/1
9/1
5/1
3/1
85
84
80
79
85
45
1/2
1/1
Determined by 1H NMR analysis. Determined by CSP SFC
a
b
analysis. c Yield of chromatographically homogeneous material
containing a mixture of diastereomers.
To optimize the phosphoramide-catalyzed aldol addi-
tions with (Z)-5 for substrates other than benzaldehyde,
a loading study was performed with 2-furfuraldehyde,
Table 7. Other than moderating the catalyst loadings,
the reaction concentrations were increased to 1 M. The
yields of isolated adducts were similar for all runs,
J . Org. Chem, Vol. 68, No. 13, 2003 5049

Denmark and Pham
CHART 2
was again selected as the representative substrate with
reactions being conducted at 1 M concentration in
CH2Cl2 and were quenched after 8 h. The yields were
similar in both cases, ranging from 76 to 80% for 25 and
50 mol % loadings, respectively. Unfortunately, even with
50 mol % of the catalyst, the reactions only afforded an
8/1, syn/anti mixture of the undesired diastereomer in
racemic form.
TABLE 8. Su r vey of Ald eh yd e Accep tor s for th e
Ad d ition of (E)-5 Ca ta lyzed by 2
aldehyde
product
dr, syn/antia
er (syn)b
yield, %c
a
b
c
d
e
16a
16b
16c
16d
16e
1/1
1/8
1/2
1/5
1/13
8/1
12/1
3/1
3/1
1/1
86
82
76
82
53
a
b
Determined by 1H NMR analysis. Determined by CSP SFC
analysis. c Yield of chromatographically homogeneous material
containing a mixture of diastereomers.
TABLE 9. Su r vey of Dim er ic P h osp h or a m id es in th e
Ald ol Ad d ition of (Z)-12 to Ben za ld eh yd e
The aldol additions with (Z)-12 clearly showed an in-
crease in enantioselectivity by increasing the loading of
catalyst from 5 to 50 mol %. To further improve on this
result and to reduce the amount of catalyst required, sev-
eral dimeric catalysts based on 2 were prepared and their
ability to catalyzed aldol additions was evaluated. The
use of dimeric phosphoramides is predicated on the know-
ledge that two phosphoramides are required in the tran-
sition state to provide high levels of stereoselectivity.15
It was envisioned that covalently tethered Lewis basic
residues would enhance the relative concentration of the
second coordinating species, allowing more facile forma-
tion of the two-phosphoramide complex. Although similar
dimeric catalysts have been employed successfully in al-
lylations with allyltrichlorosilanes,16 aldehyde-aldehyde
aldol additions,13 and silyl ketene acetal additions to ke-
tones,17 no attempts have been made thus far to evaluate
their efficacy in ketone-aldehyde aldol additions.
catalyst
dr, syn/antia
er (syn)b
yield, %c
18
15/1
16/1
15/1
8/1
1.3/1
1.2/1
1.8/1
2.8/1
2.1/1
1/1
69
48
64
78
63
85
62
19
20
21d
22d
23e
24e
8/1
2/1
1/1
N/A
a
b
Determined by 1H NMR analysis. Determined by CSP SFC
analysis. c Yield of chromatographically homogeneous material
d
containing a mixture of diastereomers. Began with an 8/1, Z/E
mixture. e 15 mol % of the catalyst used.
The effectiveness of the dimeric phosphoramides in
aldol addition reactions was evaluated with (Z)-12 and
benzaldehyde. As before, reactions were run at 1 M
concentration in CH2Cl2, using 5 mol % of the catalyst.
The reactions were allowed to proceed for 8 h before being
quenched with saturated, aqueous sodium bicarbonate
solution and CH2Cl2. The results of the catalyst survey
are summarized in Table 9. Although with only a single
example, the dimers 18-22 were superior to the linked
binaphthyldiamine-derived phosphoramide, 23, which
had been employed with great success in previous allyl-
ation and crossed-aldehyde aldol additions. Even though
the enantioselectivities were extremely modest, the di-
astereoselectivities were essentially perfect for reactions
run with the chiral stilbene-1,2-diamine-derived dimers
The dimeric catalysts were prepared as previously
described, Scheme 4. Preparation of the chiral phosphoryl
chloride followed by treatment with the lithiated di-
methyl diamine linker afforded the tethered phosphora-
mides as stable white powders, Chart 2.
SCHEME 4
(15) (a) Denmark, S. E.; Pham, S. M. Helv. Chim. Acta 2000, 53,
1846. (b) Denmark, S. E.; Su, X.; Nishigaichi, Y. J . Am. Chem. Soc.
1998, 120, 12990.
(16) (a) Denmark, S. E.; Fu, J . Org. Lett. 2002, 4, 1951. (b) Denmark,
S. E.; Fu, J . J . Am. Chem. Soc. 2001, 123, 9488. (c) Denmark, S. E.;
Wynn, T. A. J . Am. Chem. Soc. 2001, 123, 6199.
(17) Denmark, S. E.; Fan, Y. J . Am. Chem. Soc. 2002, 124, 4233.
5050 J . Org. Chem., Vol. 68, No. 13, 2003

Achiral Ethyl Ketone-Derived Trichlorosilyl Enolates
SCHEME 6
18-22. Interestingly, for this class of linked phosphora-
mides a tether length of seven methylene units appears
to be optimal with respect to enantioselectivity. This
differs from the allylations and aldehyde-aldehyde al-
doladditions where a tether length of five methylene units
was found to be optimal.
Although the pentanone-derived enolates did not dis-
play a selectivity enhancement with respect to catalyst
loading, a survey of the dimeric phosphoramides was also
performed with (Z)-5 and 1-naphthaldehyde. Again,
reactions were conducted at 1 M concentration in CH2-
Cl2 with 5 mol % of the catalyst for 8 h. Yields for
reactions catalyzed by dimers 19-22 ranged from 68 to
71%. Unfortunately, reactions of (Z)-5 in the presence of
the dimeric phosphoramides demonstrated poor diaste-
reoselectivities, 2.7/1 to 3.0/1 syn/anti, with little to no
enantioselectivity, even with the optimal tether length.
between the rate of formation of the Z-isomer (kZ) and
the rate of formation of the E-isomer (kE) as governed by
the nature of the metal, Scheme 6.
Further studies with various ethyl ketone derived TMS
enol ethers have shown that substitution on the spectator
side has a profound influence on selectivity during trans-
silylation. In fact, large substituents effectively diminish
the difference in selectivities observed previously for the
pentanone-derived silyl enol ether 4 when mercury and
palladium are used. Likewise, there is little difference
between mercury and palladium when substituents are
smaller than ethyl as seen in the trans-silylation of 6 (R
) H) and 7 (R ) Me). For each case, the metatheses
catalyzed by Hg(OAc)2 and Pd(OAc)2 gave virtually
identical results. Interestingly, increasing steric bulk
from methyl to phenyl increases the level of Z-selectivity
for silicon metathesis, such that trimethylsilyl enol ether
10 (R ) Ph) metathesizes to give only (Z)-15, regardless
of the metal catalyst. These steric effects are likely
manifested by the minimization of 1,2-eclipsing interac-
tions between the spectator group and the methyl sub-
stituent of the metal enolate. As the spectator group
increases in size, the equilibrium between the conformers
shifts to avoid unnecessary steric interactions.
Discu ssion
A. Meta l-Ca ta lyzed Tr a n s-Silyla tion . A variety of
ethyl ketone-derived trichlorosilyl enolates were prepared
via mercury(II)- or palladium(II)-catalyzed trans-silyla-
tion beginning with the corresponding trimethylsilyl enol
ethers. Although a diverse assortment of Lewis acidic
metal salts and complexes was studied, only the acetate
and trifluoroacetate salts of mercury and palladium gave
consistently high levels of silicon metathesis. In a few
cases, such as PtCl2, (PhCN)2PtCl2, and NiCl2, isomer-
ization of the trimethylsilyl enol ether was observed
without trans-silylation. The remaining metal salts
provided significant levels of hydrolyzed product. Al-
though it is possible that formation of the ketone could
occur via protodesilylation of the silyl enol ether, it is
more likely that adventitious water present in the salt
interrupted the metathetical pathway by quenching the
metal enolate, Scheme 5. Interestingly since Pd(TFA)2
was the slowest acting metathesis catalyst, isomerization
was observed well before significant amounts of trans-
silylation had taken place (vide infra).
The increase in steric bulk of the spectator also reduces
the rate of trans-silylation such that the metathesis of
10 required an increase in catalyst concentration even
when Hg(OAc)2 was used, the most effective trans-
silylation catalyst. Curiously, a decrease in steric bulk
from ethyl to hydrogen resulted in slower trans-silyla-
tions as well, possibly from a decrease in nucleophilicity
of the aldehyde enolate. Thus, the metathesis of 6 and 7
required longer reactions times compared to 4 regardless
of the catalyst. As of yet, the origin of the rate attenuation
remains unclear. In all cases, longer reaction times
resulted in lower yields of trichlorosilyl enol ethers. It is
presumed that this is due to hydrolysis of the trichlo-
rosilyl enolate as it remains in the reaction flask during
long reaction times.
SCHEME 5
The results from these studies provided a key insight
into the mechanism of trans-silylation. It may be postu-
lated that Lewis acidic metals rapidly and reversibly
combine with trimethylsilyl enol ethers to generate an
organometallic species capable of allowing isomerization
of the enol ether. The nature of this species has not been
investigated, although it is possible to envision isomer-
ization via a zwitterionic oxycarbenium adduct, Figure
1.
By the use of various mixtures of geometric isomers of
4, it appeared that isomerization occurs prior to, or is
concomitant with trans-silylation as alluded above. The
use of a 30/1 or a 1/4, E/Z-mixture of (E/Z)-4 afforded 5
with similar geometric ratios depending upon the choice
of metal, but regardless of the initial geometric mixture.
This result indicates that metal-catalyzed trans-silylation
affords a mixture of both E- and Z-trichlorosilyl enolates
with the final ratio being a function of the partitioning
It is likely that complexation of the metal catalyst gives
rise to organometallic species of type I. In fact, the
formation of R-mercurioketones by this route is well
J . Org. Chem, Vol. 68, No. 13, 2003 5051

Denmark and Pham
Pd(OAc)2-catalyzed dehydrosilylations of silyl enol ethers
to generate R,â-unsaturated ketones.20 The authors men-
tion that PdCl2 was less effective than Pd(OAc)2 due to
its poor solubility in acetonitrile. Likewise for trans-
silylations, the in-situ formation of metal chlorides may
be advantageous due to its increased dispersity and
higher solubility.
The striking observation that the metal can influence
the stereoselectivity of trans-silylation warrants special
comment. In both cases, the stereodetermining event is
most likely the loss of metal chloride to produce the
trichlorosilyl enolate. The conformation of the metallo-
species prior to the loss of metal salt is crucial for
determining the geometry of the resultant trichlorosilyl
enol ether. From a stereoelectronic standpoint, the
carbon-metal bond should be orthogonal to the atomic
plane so as to maximize overlap with the carbonyl
π-system.
It is well-established that Group 11 and 12 metal
enolates such as mercury exist primarily in the carbon-
bound form.18,21 This gives two limiting conformations for
the loss of metal salt, Figure 2. The more favorable
conformation places the hydrogen substituent on the
metal-bearing carbon synclinal to the spectator group,
IV, resulting in minimal gauche interactions between
these two substituents. Demetalation from this confor-
mation affords the Z-trichlorosilyl enol ether. The less
favorable conformation places the methyl substituent on
the metal-bearing carbon synclinal to the spectator group,
V, causing larger steric interactions. Subsequent demeta-
lation from this conformation would afford the E-trichlo-
rosilyl enol ether. This explanation would clearly explain
the effect that bulkier substituents22 have on the stereo-
selectivity of trans-silylation. Thus, substituents such as
tert-butyl and phenyl provide large steric interactions
such that demetalation provides virtually only the Z-
isomer.
F IGURE 1. Proposed trans-silylation mechanism.
precedented in the work of both House18 and Yamamoto.19
In either case, the intermediacy of I allows for rapid
isomerization of the trimethylsilyl enol ether.
The next step in this process is postulated to involve
the loss of a metal counterion followed by removal of the
TMS group to afford either an R-mercurioketone (vide
supra) or an R-palladioketone. Complexation of the
R-metalloketone by SiCl4 with concomitant displacement
of chloride generates a new trichlorosilyl oxycarbenium
ion adduct, III. Attack of chloride on the metal affords
the trichlorosilyl enolate and regenerates a form of the
metal catalyst.
From the proposed mechanism, some of the metal
chlorides are required to catalyze the trans-silylation
after two turnover events regardless of the results for
the exchange reactions. However, the exchange studies
clearly demonstrate that counterion exchange is rapid
compared to trans-silylation, suggesting that a metal-
(II) chloride is the active catalyst that is formed in situ.
Thus, the effect of the counterion on the selectivity of
trans-silylation would be minimal at best for both Hg-
(OAc)2 and Pd(OAc)2 because these exchange so rapidly.
Conversely, it is conceivable that counterion exchange
with Pd(TFA)2 may be competitive with trans-silylation,
although as stated previously, full exchange to the more
reactive metal(II) chloride would be expected after two
turnover events.
Nonetheless, the effect of the counterion on the rate of
metathesis can be profound if Pd(OAc)2 and Pd(TFA)2 are
compared. For the trans-silylation of enol ether 4, Pd-
(TFA)2 is nearly a factor of 8 slower than Pd(OAc)2. The
exchange studies also determined that Pd(TFA)2 ex-
changes much more slowly than Pd(OAc)2, supporting the
hypothesis that PdCl2 is the active catalyst. Therefore,
when beginning with Pd(TFA)2, trans-silylation does not
occur until sufficient quantities of PdCl2 are produced.
It is likely that chloride is required to displace the TMS
group as TFA may not be of sufficient nucleophilicity to
affect this process. This is further evidenced by the
presence of TMSCl during the early stages of reaction,
although a rapid ligand exchange between TMSOTf and
adventitious chloride cannot be completely overlooked.
Interestingly, trans-silylations are extremely slow when
beginning with either HgCl2 or PdCl2. This apparent
contradiction was also observed by Saegusa et al. in their
F IGUR E 2. Proposed mechanism for mercury-catalyzed
trans-silylation.
Interestingly, these steric arguments do not govern the
stereoselectivity observed for the preparation of 6 (R )
H). In fact, trans-silylation of 6 afforded the trichlorosilyl
(20) Ito, Y.; Hirao, T.; Saegusa, T. J . Org. Chem. 1978, 43, 1011.
(21) (a) Bluthe, N.; Malacria, M.; Gore, J . Tetrahedron 1984, 40,
3277. Drouin, J .; Boaventura, M.-A.; Conia, J .-M. J . Am. Chem. Soc.
1985, 107, 1726. (b) Drouin, J .; Boaventura, M.-A.; Conia, J .-M. J . Am.
Chem. Soc. 1985, 107, 1726.
(18) House, H. O.; Auerbach, R. A.; Gall, M.; Peet, N. P. J . Org.
Chem. 1973, 38, 514.
(19) Yamamoto, Y.; Maruyama, K. J . Am. Chem. Soc. 1982, 104,
2323.
(22) For a discussion of steric effects in organic chemistry as well
as a listing of Taft steric parameters, see: Newman, M. S., Ed. Steric
Effects in Organic Chemistry; Wiley: New York, 1956.
5052 J . Org. Chem., Vol. 68, No. 13, 2003

Achiral Ethyl Ketone-Derived Trichlorosilyl Enolates
enol ether with greater Z-selectivity than for the methyl
derivative 7. Although counterintuitive, the preference
for silyl enol ethers of aldehydes to adopt a Z-configura-
tion is well documented.23 In the preparation of a variety
of aldehyde enoxysilanes, Cazeau reported that the
Z-isomer was the major product in each case.23b Although
the author does not provide a rationalization for the
observed stereoselectivity, he does comment that these
enoxysilanes were prepared under thermodynamic con-
trol.
complex. Complexation to mercury affords carbon-bound
enolates. The level of stereoselection is directed by the
magnitude of the gauche interactions between the methyl
group and the spectator group for the two limiting
conformations available for demetalation. On the other
hand, complexation to palladium provides enolates that
may be considered as an oxo-π-allylpalladium complex.
In this case, the degree of stereoselection is directed by
the magnitude of the A1,2 strain between the methyl
group and the spectator group during demetalation,
allowing for greater levels of stereoselection in the case
of some substrates.
Unlike Group 11 and 12 metal enolates, Group 7
through 10 enolates have been proposed to exist dynami-
cally as η3-oxaallylmetallo enolates and 2-oxoalkylmetallo
enolates.20,24 Although the isolation and characterization
of an oxo-π-allylpalladium complex has only recently been
reported,25 speculation of their existence has been pro-
posed extensively in the literature.26 Thus, revising the
proposed mechanism for palladium catalysis to incorpo-
rate the η3-oxaallylpalladium intermediate would prove
insightful, Figure 3. For these intermediates an unfavor-
able conformation exhibits severe A1,2 strain from the
positioning of the methyl substituent on the metal-
bearing carbon with the spectator group in a synplanar
fashion, VI. Therefore, demetalation would occur pref-
erentially through a structure similar to that of VII to
position the methyl substituent antiperiplanar with
respect to the spectator group. This would lead to the
formation of Z-trichlorosilyl enol ethers. It follows that
the magnitude of A1,2 strain increases with larger specta-
tor groups, such that trans-silylations of 9 (R ) t-Bu) and
10 (R ) Ph) are nearly completely selective for the
Z-isomer. As was the case for mercury catalysis, explana-
tions for the observed results with 6 (R ) H) lie outside
steric arguments alone. Thus it is conceivable that
stereoelectronic factors play a large role in determining
the stereochemical outcome of trans-silylations with
aldehyde TMS enol ethers.
B. Ald ol Ad d ition s. Aldol additions of trichlorosilyl
enolates to aldehydes catalyzed by phosphoramide 2 are
believed to proceed through a two-phosphoramide mech-
anism characterized by chairlike transition structures.
This affords diastereomeric products related to the start-
ing enolate geometry as predicted by the Zimmerman-
Traxler model.27,28 For the addition of (Z)-5 (R ) Et) to
benzaldehyde, essentially perfect diastereocontrol was
achieved in that a 15/1, Z/E enolate mixture afforded
aldol adduct 16a in a 16/1, syn/anti ratio. The level of
stereocontrol is further highlighted by the high enanti-
oselectivity. Unfortunately, attenuated selectivities were
observed when employing 1-naphthaldehyde as the elec-
trophile. In this example, only an 8/1, syn/anti ratio was
obtained suggesting a failure in the activated complex
to sustain a chairlike transition structure, which may be
manifested in either of two ways. First, the ability to
achieve a 2:1 phosphoramide:enolate complex may be
diminished when aldehydes other than benzaldehyde are
used in reactions with 5. This 1:1 complex would now
have a propensity to react through a boatlike transition
structure affording anti products.15a Second, the parti-
tioning between chairlike and boatlike transition struc-
tures for the 2:1 complex of this enolate may allow the
boat form to be competitive, Figure 4. Both scenarios are
possible though it is unclear why 2:1 complexes would
be less favorable for this class of trichlorosilyl enolate or
why the 2:1 complex should be less selective for chairs.
The use of smaller electrophilic partners such as
2-furfuraldehyde and trans-cinnamaldehyde had little
effect in improving stereoselectivities. Indeed, it is be-
lieved that phosphoramide-catalyzed aldol additions with
trichlorosilyl enolates require relatively bulky aldehydes
(24) (a) Goetz, R. W.; Orchin, M. J . Am. Chem. Soc. 1963, 85, 2783.
(b) Prince, R. H.; Raspin, K. A. J . Chem. Soc. (A) 1969, 612. (c) Noyori,
R.; Umeda, I.; Ishigami, T. J . Org. Chem. 1972, 37, 1542. (d) Bennett,
M. A.; Watt, R. Chem. Commun. 1971, 95. (e) Yoshimura, N.;
Murahashi, S.-I.; Moritani, I. J . Organomet. Chem. 1973, 52, C58.
(25) (a) Fujii, A.; Hagiwara, E.; Sodeoka, M. J . Am. Chem. Soc. 1999,
121, 5450. (b) Sodeoka, M.; Tokunoh, R.; Miyazaki, F.; Hagiwara, E.;
Shibasaki, M. Synlett 1997, 463.
(26) (a) Sodeoka, M.; Ohrai, K.; Shibasaki, M. J . Org. Chem. 1995,
60, 2648. (b) Albe´niz, A. C.; Catalina, N. M.; Espinet, P.; Redo´n, R.
Organometallics 1999, 18, 5571. (c) Nokami, J .; Mandai, T.; Watanabe,
H.; Ohyama, H.; Tsuji, J . J . Am. Chem. Soc. 1989, 111, 4126. (d) Ito,
Y.; Aoyama, H.; Hirao, T.; Mochizuki, A.; Saegusa, T. J . Am. Chem.
Soc. 1979, 101, 494.
F IGURE 3. Proposed mechanism for palladium-catalyzed
trans-silylation.
(27) Zimmerman, H. E.; Traxler, M. D. J . Am. Chem. Soc. 1957,
79, 1920.
In summary, the difference in stereoselectivity exhib-
ited by the various metal salts can be attributed mainly
to the different manners in which enol ethers may
(28) The configuration at silicon in the putatitve cationic two-
phosphoramide transition structure depicted in Figure 4 is purely
speculative, but based on the following rationale: (1) the observed cis-
coordination of phosphoramides in SnCl4 complexes (ref 29), (2) the
preferred coordination of a complexed benzaldehyde molecule trans
to a chlorine atom in a phosphoramide‚SnCl4 complex (ref 29), and (3)
the expected activating effect of a stronly donating phosphoramide
ligand trans to the enolate (ref 30).
(23) (a) Miller, R. D.; McKean, D. R. Synthesis 1979, 730. (b) Cazeau,
P.; Buboudin, F.; Moulines, F.; Babot, O.; Dunogues, J . Tetrahedron
1987, 43, 2075. (c) Fataftah, Z. A.; Ibrahm, M. R.; Abdel-Rahman, H.
N. Bull. Chem. Soc. J pn. 1991, 64, 671.
J . Org. Chem, Vol. 68, No. 13, 2003 5053

Denmark and Pham
12. Again, 1-naphthaldehyde remained superior com-
pared with additions to the smaller unsaturated alde-
hydes as well as cyclohexanecarboxyaldehyde.
Although competition between 2:1 and 1:1, phosphora-
mide:enolate pathways may sometimes be problematic,
it is unlikely that such a competition may explain the
less than ideal results observed for aldol additions with
enolate 5. The absence of a loading effect, albeit with
modest diastereoselectivity for the addition of (Z)-5 to
2-furfuraldehdye, suggests that 2:1 complexation is readily
achieved. This further explains why no enhancement was
obtained when attempting additions to 1-naphthaldehyde
with the dimeric phosphoramides. Although the tethered
catalyst certainly increases the relative concentration of
the second phosphoramide, it is likely that the presence
of the linker prevents ideal binding of the Lewis basic
ends. Unfortunately in the monomeric phosphoramide 2,
the nature of the complex involving enolate 5 allows
flexibility for the transition structure to rest in either a
chair-form or boat-form resulting in diminished diaste-
reoselectivities.
Interestingly, switching to a smaller spectator group
in enolate 7 (R ) Me) afforded dramatically different
results. Severely diminished selectivities were observed
for the addition of (Z)-7 to benzaldehyde with typical
catalyst loadings of 10-15 mol %. Unlike (Z)-5, (Z)-7 was
highly responsive to catalyst concentration, achieving
excellent results when utilizing 50 mol % of phosphora-
mide 1. This suggests that binding of the second phos-
phoramide with enolate 5 may be less facile compared
to enolate 7, allowing aldolization through a one-phos-
phoramide pathway to become competitive. Thus, in-
creasing the relative concentration of the second phos-
phoramide through the use of dimeric catalysts afforded
aldol products with essentially perfect correlation of
enolate geometry to diastereoselectivity. Interestingly,
the optimal tether length was seven methylene units,
unlike previous examples where five methylene units
proved to be superior, Figure 6. This suggests that the
relative orientation around silicon is different for allyl-
ation reactions than for aldol additions, although this
may not be entirely the case as the optimal lengths were
determined with different catalyst structures. For the
aldol additions, tether lengths of four and five methylene
units afforded poor enantioselectivities. This may be the
result of the methylene tether restricting coordination
of the Lewis basic units to a nonoptimal orientation.
Increasing the tether length to six and seven methylene
units may allow added flexibility to achieve a more
F IGURE 4. Competing transition structures which afford
diastereomeric products.
F IGURE 5. Competing chairlike and boatlike transition
structures for the addition of (E)-12 to aldehydes.
SCHEME 7
to enhance the energetic differences between the chair
and boat transition structures. Not unexpectedly, the
addition of (Z)-5 to cyclohexanecarboxaldehyde was poor
with regard to both stereoselectivity and yields. Although
not typically problematic with unsaturated aldehydes,
the reversible formation of chlorohydrins often leads to
diminished yields with aliphatic aldehydes, Scheme 7.
The lack of stereocontrol is evident by the formation of
the undesired diastereomer in racemic fashion.
Surprisingly, use of (E)-5 (R ) Et) resulted in severely
attenuated selectivities for additions to benzaldehyde.
Unlike the near-perfect correlation of geometric informa-
tion with (Z)-5, the E-counterpart afforded aldol adducts
with no diastereoselectivity with benzaldehyde. A lack
of chair-boat selectivity may further explain the reduced
enantioselectivities in that addition of the minor geo-
metric isomer through a boatlike transition structure
should afford the desired diastereomer albeit with low
enantioselectivity, Figure 5. The results for the remaining
aldehydes reflect those observed for additions with (Z)-
F IGURE 6. Correlation of enantioselectivity vs catalyst tether
length.
5054 J . Org. Chem., Vol. 68, No. 13, 2003

Achiral Ethyl Ketone-Derived Trichlorosilyl Enolates
F IGURE 7. Overlay of the 2:1 monomer complex and five methylene tether phosphoramide complex.
desirable coordination geometry. A decrease in enantio-
selectivity at eight methylene units may imply an en-
tropic consequence that disfavors chelation. Although
improved enantioselectivities could not be achieved,
enhancement of chair selectivity through the use of as
little as 5 mol % of a dimeric phosphoramide clearly
demonstrates the utility of linked catalysts for increasing
the relative concentration of the second binder. Unfor-
tunately, results with (E)-7 were disappointing. Although
only a small loading study was conducted, formation of
the undesired diastereomer in racemic form was observed
even when 50 mol % of the catalyst was used.
To better understand the nature of the activated
complex, a computational modeling was done with X-ray
crystal structures of the tin complexes for the monomeric
phosphoramide 2 and the dimeric, five methylene tether
phosphoramide 19 as templates.29 The X-ray crystal
structures of 2 and 19 with tin tetrachloride were viewed
and manipulated with Cerius2. Initial observation of the
tin complexes revealed that the orientation of the phos-
phoramides was distinctly different between the 2:1
monomeric complex and the 1:1 dimeric complex. By
aligning the vertices of the octahedral complex, it is
observed that there exists little overlap between the
second phosphoramides of the two complexes, Figure 7.
Analysis of the crystal structures shows a rotation of the
second phosphoramide unit, in the dimeric complex, away
from the preferred orientation established by the crystal
structure of the 2:1 complex. It is likely that this was
necessary to avoid bond distortion in the methylene
tether.
unsuccessful, observing that the complexes each relaxed
back to their original conformation. Thus it appears that
regardless of the length, the methylene tether distorts
the dimer complexation geometry away from the pre-
sumed idealized geometry established with the 2:1
complex, using the monomeric phosphoramide.
Con clu sion s
Z-Trichlorosilyl enolates can be prepared by mercury-
(II) and palladium(II) catalysis from the corresponding
trimethylsilyl enol ethers. Although Pd(TFA)2 could also
catalyze the trans-silylation, its activity was attenuated
by slow exchange to the chloride salt. From this process,
good to excellent stereoselectivity was observed for large
spectator groups (Ph, t-Bu); however, only moderate
stereoselection was achieved for smaller groups (Me, Et).
The scope of aldol additions involving trichlorosilyl
enolates has been expanded to include ethyl ketone
enolates, with reactions proceeding in good to high yields
for all reactions with the exception of aliphatic aldehydes.
Excellent stereoselectivities were observed for the addi-
tion of (Z)-5 and (Z)-7 to benzaldehyde. Unfortunately,
the reactions with these substituted enolates lack sub-
strate generality with stereoselectivities becoming at-
tenuated when electrophilic partners other than benzal-
dehyde are used. Efforts to improve the overall generality
of aldol additions of substituted trichlorosilyl enolates to
aldehydes by the design of second-generation dimeric
phosphoramides and the development of alternate reac-
tion manifolds are currently underway.
Methylene units were then sequentially added and the
resulting structure was minimized to evaluate the effect
of increasing tether lengths on the orientation of the
phosphoramides. Unfortunately this provided little in-
sight as each of the minimized complexes adopted nearly
the same phosphoramide conformation for five methylene
up to eight methylene unit tethers. Attempts to access
other minima by severely distorting the tether were also
Ack n ow led gm en t. We are grateful to the National
Science Foundation for generous financial support (CHE
9803124-0105205). S.M.P. thanks the Eastman-Kodak
Company for a graduate fellowship.
Su p p or tin g In for m a tion Ava ila ble: Detailed procedures
for the preparation and aldol addition reactions of trichlorosilyl
enolates along with spectroscopic characterization of the
enolates and aldol products. This material is available free of
charge via the Internet at http://pubs.acs.org.
(29) Denmark, S. E.; Su, X. Tetrahedron 1999, 55, 8727.
(30) Coe, B. J .; Glenwright, S. J . Coord. Chem Rev. 2000, 203, 5.
J O034092X
J . Org. Chem, Vol. 68, No. 13, 2003 5055