Investigation of the Rh-catalyzed asymmetric reductive aldol reaction. Expanded scope based on reaction analysis

Article abstract of DOI:10.1021/ol049591e

(Equation Presented) A series of experiments are described that suggest that the Rh-catalyzed reductive aldol reaction proceeds by addition of a Rh(I) hydride, generated in situ, to the reacting acrylate followed by a stereochemistry-controlling aldol add

Full text of DOI:10.1021/ol049591e

ORGANIC
LETTERS
2004
Vol. 6, No. 14
2309-2312
Investigation of the Rh-Catalyzed
Asymmetric Reductive Aldol Reaction.
Expanded Scope Based on Reaction
Analysis
Albert E. Russell, Nathan O. Fuller, Steven J. Taylor, Pauline Aurriset, and
James P. Morken*
Department of Chemistry, Venable and Kenan Laboratories, The UniVersity of North
Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290
morken@unc.edu
Received March 4, 2004
ABSTRACT
A series of experiments are described that suggest that the Rh-catalyzed reductive aldol reaction proceeds by addition of a Rh(I) hydride,
generated in situ, to the reacting acrylate followed by a stereochemistry-controlling aldol addition reaction. On the basis of this hypothesis,
reaction conditions are engineered that allow for increased substrate scope.
As part of a program focused on the catalytic asymmetric
transformation of simple organic molecules into value-added
products, we have been investigating the late transition-metal-
catalyzed reductive aldol reaction between aldehydes and
unsaturated esters. This transformation, first demonstrated
by Revis and later by Mukaiyama, Kiyooka, and Murai,
provides access to aldol adducts in a mild and catalytic
fashion.1 Development of an asymmetric reductive aldol
reaction is complicated by the fact that group 9 and group
10 catalysts, which are effective for the reaction itself, also
exhibit reactivity in carbonyl and acrylate hydrosilylation.
Thus, our initial work on the introduction of an asymmetric
version of the Rh-catalyzed reaction was facilitated by the
aid of high-throughput screening. We expected that subse-
quent design of improved reaction conditions and engineering
of new reaction pathways would be facilitated by a more
detailed understanding of this reaction. Of particular concern
is the nature of the transition metal complex that is involved
in the C-C bond-forming step and whether it is a Rh(I) or
Rh(III) complex. In this communication, we describe a series
of experiments that suggest a plausible mechanism for the
Rh-catalyzed asymmetric reductive aldol reaction and provide
a working hypothesis that allows for expansion of the
substrate scope.
The asymmetric reductive aldol reaction may be ac-
complished in an enantioselective manner with [(cod)RhCl]2/
binap as the precatalyst and with Et2MeSiH as the reductant.
Early experiments indicated that the reaction produces silyl-
protected aldol adducts from phenyl acrylate and aliphatic
aldehydes and that these may be hydrolyzed during workup
to afford the derived â-hydroxy ester.2 An informative
observation on the nature of the asymmetric reductive aldol
(1) (a) Revis, A.; Hilty, T. K. Tetrahedron Lett. 1987, 28, 4809. (b)
Isayama, S.; Mukaiyama, T. Chem. Lett. 1989, 2005. (c) Matsuda, I.;
Takahashi, K.; Sato, S. Tetrahedron Lett. 1990, 31, 5331. (d) Kiyooka, S.;
Shimizu, A.; Torii, S. Tetrahedron Lett. 1998, 39, 5237. For an excellent
recent review of this area, see: Huddleston, R. R.; Krische, M. J. Synlett
2003, 12.
(2) (a) Taylor, S. J.; Duffey, M. O.; Morken, J. P. J. Am. Chem. Soc.
2000, 122, 4528. For an iridium-catalyzed asymmetric reductive aldol
reaction, see: (b) Zhao, C. X.; Duffey, M. O.; Taylor, S. J.; Morken, J. P.
Org. Lett. 2001, 3, 1829.
10.1021/ol049591e CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/05/2004

reaction arises from analysis of minor reaction products and
is shown in Scheme 1. Analysis of the reaction mixture after
that the organosilane may not be involved in the stereo-
chemistry-controlling step of the cycle.
In an effort to learn more about the nature of transition
metal species generated under the reaction conditions, (R-
binap)Rh(cod)BF4 (3, Scheme 3) was subjected to each of
Scheme 1
Scheme 3
the reaction components in concentrations reflective of
reaction conditions ([binapRhCl]2 shows little detectable
reaction with any of the reaction components). While neither
acrylate nor aldehyde was reactive with the precatalyst,
reaction with PhMe2SiH provides free cyclooctadiene and a
single new C2-symmetric species 4 (Scheme 3) in about 20%
conversion after 30 min. {1H}31P NMR (doublet, 38.6 ppm,
J ) 147 Hz), 1H NMR (broad overlapping triplet of quintets,
-8.2 ppm, JRh-H ) 17 Hz, JP-H ) 17 Hz), and {31P}1H NMR
spectra (triplet, -8.2 ppm, J ) 17 Hz) are consistent with
structure 4, a dinuclear bridged Rh(I) hydride.3 In addition
to 4, 19F NMR indicates that PhMe2SiF is also formed in
this transformation.4 While efforts to prepare 4 by ligand
exchange with [(cod)RhH]4 and by NaEt3BH addition to
[(binap)RhCl]2 were unsuccessful, 4 was detected in the
reaction of 3 with (EtO)Me2SiH. Under similar conditions,
treatment of 3 with Et2MeSiH did not lead to a detectable
hydride-containing product. That 4 is a competent catalyst
or precatalyst was determined by allowing conversion of 3
to 4 to reach completion prior to introduction of acrylate
and aldehyde. This was accomplished by allowing 3 to react
with 28 equiv of PhMe2SiH for 2 h at which point 31P NMR
showed >95% 4 relative to all other phosphorus-containing
compounds. Upon introduction of acrylate and aldehyde, the
reductive aldol adduct was generated at a slightly elevated
rate and identical enantioselectivity relative to reactions
initiated with 3.
hydrolytic workup reveals that, in addition to the major
product 1, a second aldol-type reaction product is formed in
which the â-hydroxyl group is acylated (2, Scheme 1).
Control experiments indicate that 2 does not likely arise from
the silyl ether of 1 under the reaction conditions; inclusion
of the silylated benzaldehyde reductive aldol adduct, in a
reductive aldol reaction with propionaldehyde, does not lead
to acylation of the benzaldehyde adduct. Significant to our
mechanistic hypothesis, both aldol adducts 1 and 2 are
formed with the same level of stereoselectivity and the
product ratio is dependent on reagent concentration: doubling
the aldehyde concentration and halving the silane concentra-
tion results in a 1:1 ratio of 1:2.
Reductive aldol reactions with (R-binap)Rh(cod)BF4 (3)
provide the acylated aldol adduct 2 in a similar manner as
reactions with the neutral catalyst; however, hydrosilylation
of the acrylate predominates when an equimolar ratio of
silane, acrylate, and aldehyde are employed. When a 4:2:1
ratio of aldehyde:acrylate:silane was employed with the
cationic catalyst, then 2 is the dominant reaction product with
either PhMe2SiH or Et2MeSiH. As shown in Scheme 2, the
Scheme 2
A mechanism that is consistent with the experimental
observations is depicted in Scheme 4. We suspect that
dissociation of the bridged dimer 4 provides a Rh(I) hydride
that reacts with the acrylate to provide a Rh(I) enolate.5 After
(3) For other dinuclear µ-H Rh(I) phosphine complexes, see: (a) Fryzuk,
M. D.; Piers, W. E.; Einstein, F. W. B.; Jones, T. Can. J. Chem. 1989, 67,
883. (b) Fryzuk, M. D. Organometallics 1982, 1, 408. (c) Day, V. W.;
Fredrich, M. F.; Reddy, G. S.; Sivak, A. J.; Pretzer, W. R.; Muetterties, E.
L. J. Am. Chem. Soc. 1977, 99, 8091.
(4) Transformation of 3 to 4 likely proceeds by oxidative addition of
silane followed by fluoride abstraction from BF4, thereby generating a Rh-
ratio of 1 to 2 appears to be dependent on the structure of
the silane, and with (EtO)Me2SiH, only the acylation adduct
is observed. Significantly, in each experiment the syn
diastereomer of compound 2 is produced in the same level
of enantioselection as is the syn diastereomer of 1. Addition-
ally, the enantioselection appears to be relatively insensitive
to significant changes in silane structure (enantiomer ratio
varies from ∼89:11 to 91:9 for PhMe2SiH, (EtO)Me2SiH,
and Et2MeSiH). Collectively, these observations suggest that
1 and 2 may arise from the same catalytic intermediate and
(I) hydride, PhMe2SiF, and BF3. We have been unable to detect BF3 by 19
F
NMR, although, in the absence of Lewis bases, it can be anticipated that it
would readily outgas from solution. For a discussion of fluoride abstraction
with electrophilic complexes, see: Brookhart, M. S.; Grant, B.; Volpe, A.
F., Jr. Organometallics 1992, 11, 3920.
(5) Such compounds are precedented in the literature and have been
observed to exist as either O-bound enolates or oxo-π-allylmetal complexes
depending on the ligand environment. (a) Hayashi, T.; Takahashi, M.;
Takaya, Y.; Ogasawara, M. J. Am. Chem. Soc. 2002, 124, 5052. (b) Slough,
G. A.; Hayashi, R.; Ashbaugh, J. R.; Shamblin, S. L.; Aukamp, A. M.
Organometallics 1994, 13, 890. (c) Slough, G. A.; Bergman, R. G.;
Heathcock, C. H. J. Am. Chem. Soc. 1989, 111, 938.
2310
Org. Lett., Vol. 6, No. 14, 2004

Scheme 4
Table 1. Asymmetric Reductive Aldol Reaction with
Unsaturated Aldehyde Substratesa
the enolate participates in an aldol reaction with the alde-
hyde,6 a bifurcation in the reaction pathway might then
provide the silylated adduct (by oxidative addition/reductive
elimination with silane) or the acylated adduct (by addition/
elimination with the aldehyde7) in identical levels of stereo-
selectivity. The lack of dependence of enantioselection on
silane structure and the dependence of the product ratio (1:
2) on silane structure and concentration are accommodated
by the proposed mechanism where Rh(I) enolates are
involved and the organosilane is not involved in the C-C
bond-forming step.8
a Conditions: 1.0 M aldehyde, 0.2 M phenyl acrylate, 0.35 M Et2MeSiH,
room temp, 12 h. b Isolated yield of purified material. Percent yield based
on acrylate.
provides optimal yields of the silylated adduct. We therefore
decided to employ this combination of catalyst and reducing
agent with R,â-unsaturated aldehydes since these substrates
As expected on the basis of the mechanism proposed in
Scheme 4, other aldehyde substrates lead to differing ratios
of acylation:silylation products. For instance, benzaldehyde
provides a 4:1 ratio of silylation:acylation products (Scheme
5). Consistent with the observations described above, both
Table 2. Asymmetric Reductive Aldol Reaction with Crotonate
Derivativesa
Scheme 5
reaction products obtained from benzaldehyde were obtained
in similar levels of enantioselection.
The experiments described above indicate that [(cod)Rh-
(binap)]BF4 is converted to the putative reactive hydride with
higher efficiency than the neutral catalyst and that Et2MeSiH
(6) (a) Yoshida, K.; Ogasawara, M.; Hayashi, T. J. Am. Chem. Soc. 2002,
124, 10984. (b) See ref 5c.
(7) For Tishchenko reactions with Rh(I) alkoxides, see: (a) Krug, C.;
Hartwig, J. F. J. Am. Chem. Soc. 2002, 124, 1674.(b) Slough, G. A.;
Ashbaugh, J. R.; Zannoni, L. A. Organometallics 1994, 13, 3587.
(8) NMR analysis of the catalyst during the course of the reaction is not
informative, as 3 and 4 are the predominant detectable species. Kinetic
analysis would clearly assist our understanding of the reaction mechanism;
however, interpretation of these experiments is hampered by competitive
silylation and hydrosilylation of the acrylate at increased silane concentra-
tion.
a Conditions: 0.5 M aldehyde, 0.6 M phenyl acrylate, 2.5 M Et2MeSiH,
room temp, 48 h. b Isolated yield of purified material. c Number in
parentheses is yield based on recovered starting material.
Org. Lett., Vol. 6, No. 14, 2004
2311

do not participate in reductive aldol reactions when [(cod)-
RhCl]2/binap is used as the catalyst. With 1 equiv each of
aldehyde, acrylate, and silane, low conversion was observed;
however, the reaction was highly selective for the silylated
product. Since R,â-unsaturated aldehydes provide only the
silylated reductive aldol adduct and not the acylated adduct,
higher concentrations of these aldehydes were employed,
thereby maximizing catalyst turnover, while avoiding forma-
tion of the acylated adduct. Under these reaction conditions,
previously unreactive unsaturated aldehydes participate in
reductive aldol reactions providing acceptable reaction yields
and modest levels of stereoinduction (Table 1).
Reductive aldol reactions with phenyl crotonate in place
of phenyl acrylate proceed with lower reaction yields (15%
combined yield of acylated and silylated products, 1:1 ratio,
data not shown) when the neutral catalyst is used. As
mentioned above, it appears that with the neutral catalyst
derived from [(cod)RhCl]2 and binap, only a small amount
of catalyst enters the catalytic cycle such that [binapRhCl]2
is the sole species observed by 31P NMR. To increase the
amount of putative reactive hydride and to favor the silylation
pathway in reactions with crotonate derivatives, these reduc-
tive aldol reactions were conducted with an increased
concentration of silane.9 As observed in Table 2, with 5 equiv
of Et2MeSiH and reaction for 48 h, reductive aldol adducts
can be isolated in moderate yields and with selectivity that
mirrors that of the acrylate derivatives. The geometry of the
acrylate does not have a significant impact on reaction
stereoselectivity, and both (E)- and (Z)-configured crotonates
provide the same level of stereoinduction. Reactions with
trisubstituted acrylates do not proceed; however, arenes, silyl
ethers and alkenes could be accommodated in the reacting
substrate.
In conclusion, stereochemical analysis of reaction products
from diverging reaction pathways suggests a plausible
common catalytic intermediate in the Rh(I)/binap-catalyzed
reductive aldol reaction. This observation, combined with
other data, suggests that the catalytic cycle may be composed
primarily of Rh(I) intermediates and that an aldol addition
of a Rh(I) enolate to aldehyde substrates is the stereochem-
istry-controlling step of the reaction.10
Acknowledgment. We thank NIGMS (GM064451), As-
tra-Zeneca, Bristol-Myers Squib, DuPont, GlaxoSmithKline,
and the David and Lucile Packard Foundation for financial
support.
Supporting Information Available: Characterization
data for all new compounds and experimental procedures.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL049591E
(10) An elegant approach to catalytic intra- and intermolecular reductive
aldol reactions with molecular hydrogen as a reductant was reported recently.
These reactions may proceed through a Rh(I) or Rh(III) hydride. See: (a)
Huddleston, R. R.; Krische, M. J. Org. Lett. 2003, 5, 1143. (b) Jang, H. Y.;
Huddleston, R. R.; Krische, M. J. J. Am. Chem. Soc. 2002, 124, 15156.
(9) Unlike reactions of acrylates, reactions of crotonates exhibit lower
enantioselection when 3 is used as a catalyst (40% ee, 97% yield for entry
1, Table 2).
2312
Org. Lett., Vol. 6, No. 14, 2004