A mechanistic investigation of the asymmetric Meerwein-Schmidt-Ponndorf- Verley reduction catalyzed by BINOL/AlMe3 - Structure, kinetics, and enantioselectivity

Article abstract of DOI:10.1021/jo070563u

(Chemical Equation Presented) The kinetics of the Al-catalyzed asymmetric Meerwein-Schmidt-Ponndorf-Verley (MSPV) reduction are presented. Structural identification of the catalytic precursor formed in situ between (S)-2,2′-dihydroxy-1,1′-binapthyl ((S)-BINOL), AlMe₃, and 2-propanol was established through ¹H and ²⁷Al NMR spectroscopies, and APCIMS. All experimental evidence points toward the formation of a BINOL-chelated, pentacoordinate aluminum species in solution. Ligand-accelerated catalysis was confirmed for the phenolate/AlMe ₃/2-propanol system. The rate law for the catalytic reaction was determined to be nearly unimolecular dependent on aluminum, zero-order dependent on substrate, and inversely dependent on 2-propanol. At the low catalyst loading employed in the BINOL/AlMe₃ system, the inherent reversibility of the MSPV reaction does not affect product yield or enantiomeric excess over time. Systematic ligand studies imply that while a tetrahedral geometry around the aluminum center may result in the most active MSPV reduction catalysts, the enantioselectivity of the reaction is enhanced when the aluminum center allows for a 2-point coordination of the substrate to achieve a pentacoordinate geometry with the fifth ligand weakly coordinated to the axial site of a pseudo square pyramid.

Full text of DOI:10.1021/jo070563u

A Mechanistic Investigation of the Asymmetric
Meerwein-Schmidt-Ponndorf-Verley Reduction Catalyzed by
BINOL/AlMe sStructure, Kinetics, and Enantioselectivity
3
Christopher R. Graves, Hongying Zhou, Charlotte L. Stern, and SonBinh T. Nguyen*
Department of Chemistry and Institute for Catalysis in Energy Processes, Northwestern UniVersity,
2
145 Sheridan Road, EVanston, Illinois 60208-3113
stn@northwestern.edu
ReceiVed March 19, 2007
The kinetics of the Al-catalyzed asymmetric Meerwein-Schmidt-Ponndorf-Verley (MSPV) reduction
are presented. Structural identification of the catalytic precursor formed in situ between (S)-2,2′-dihydroxy-
1
27
1
3
,1′-binapthyl ((S)-BINOL), AlMe , and 2-propanol was established through H and Al NMR
spectroscopies, and APCIMS. All experimental evidence points toward the formation of a BINOL-chelated,
pentacoordinate aluminum species in solution. Ligand-accelerated catalysis was confirmed for the
phenolate/AlMe
unimolecular dependent on aluminum, zero-order dependent on substrate, and inversely dependent on
-propanol. At the low catalyst loading employed in the BINOL/AlMe system, the inherent reversibility
3
/2-propanol system. The rate law for the catalytic reaction was determined to be nearly
2
3
of the MSPV reaction does not affect product yield or enantiomeric excess over time. Systematic ligand
studies imply that while a tetrahedral geometry around the aluminum center may result in the most active
MSPV reduction catalysts, the enantioselectivity of the reaction is enhanced when the aluminum center
allows for a 2-point coordination of the substrate to achieve a pentacoordinate geometry with the fifth
ligand weakly coordinated to the axial site of a pseudo square pyramid.
Introduction
The Meerwein-Schmidt-Ponndorf-Verley (MSPV) reduc-
despite the practical advantages of high chemoselectivity, easily
removed side product, and applicability to both laboratory
and large-scale synthesis, it was largely supplanted in the late
1
-3
tion of carbonyl substrates to their corresponding alcohols,
950s by methods utilizing boro- and aluminum hydrides.7
1
first reported in the mid 1920s, is a highly chemoselective
reaction that can be performed under mild conditions. Utilizing
secondary alcohols, most often 2-propanol, as benign and
inexpensive hydrogen sources, the reaction is mediated by easily
accessible and regenerable aluminum alkoxides.4-6 However,
This is partially due to the fact that traditional MSPV protocols
often required super-stoichiometric amounts of aluminum
alkoxides to obtain satisfactory yields of alcohol in reasonable
5
reaction times. It was not until recent years, almost a century
after its discovery, that catalytic variants of this reduction
have been realized. These examples could be divided into two
8
,9
*
To whom correspondence should be addressed.
classes: (1) protic acid-activated aluminum alkoxides and
(
(
(
1) Verley, M. Bull. Soc. Chim. Fr. 1925, 37, 871-874.
10-16
(
2) “well-defined” aluminum complexes
where the alumi-
2) Ponndorf, W. Angew. Chem. 1926, 39, 138-143.
4
num centers are chelated by multidentate ligands. Both of
3) Meerwein, H.; Schmidt, R. Justus Liebigs Ann. Chem. 1925, 444,
2
21-238.
(
4) Graves, C. R.; Campbell, E. J.; Nguyen, S. T. Tetrahedron: Asymmetry
(7) Aldrige, S.; Downs, A. J. Chem. ReV. 2001, 101, 3305-3366.
2
005, 16, 3460-3468.
(8) Akamanchi, K. G.; Noorani, V. R. Tetrahedron Lett. 1995, 36, 5085-
(
5) Wilds, A. L. Org. React. 1944, 2, 178-223.
5088.
(6) de Graauw, C. F.; Peters, J. A.; van Bekkum, H.; Huskens, J. Synthesis
(9) Kow, R.; Nygren, R.; Rathke, M. W. J. Org. Chem. 1977, 42, 826-
1
994, 1007-1017.
827.
1
0.1021/jo070563u CCC: $37.00 © 2007 American Chemical Society
Published on Web 10/23/2007
J. Org. Chem. 2007, 72, 9121-9133
9121

Graves et al.
SCHEME 1. The MSPV Reduction Catalyzed by Simple
Aluminum Complexes (Left) and the Asymmetric MSPV
Herein, we report a detailed investigation of the Al-catalyzed
asymmetric MSPV reduction of ketones (eq 1) and discuss the
3
Reduction Catalyzed by BINOL/AlMe Precatalyst Mixture
(Right)
key ligand requirements for generating a successful Al-based
enantioselective MSPV reduction. Specifically, we determined
the reaction kinetics, the optimal aluminum coordination sphere,
and the role that reversibility plays in the BINOL/AlMe3-
catalyzed asymmetric MSPV reduction. Notably, we suggest
that the ability of R-halogenated ketone substrates to form
2
-point coordination motifs with the catalyst is critical to
improving the asymmetric outcome of their reduction.
these reported methods suffered from either decreased selectivity
or the use of elaborate, difficult-to-synthesize ligand frame-
works. We recently demonstrated that the catalytic behavior of
the MSPV reduction is highly dependent upon the aggregation
Results and Discussion
Formation of the Proposed Catalytic Species. In our
i
state of the aluminum: while commercial Al(O Pr)3 possesses
1
9
original communication, a series of equivalency tests per-
formed for the Al-catalyzed asymmetric MSPV reduction
established the optimal ligand/metal ratio between BINOL (1)
and AlMe3 to be 1:1. Although somewhat speculative, this set
of experiments led to the proposal that the active catalytic
species formed in situ between BINOL and AlMe3 proceeds
through the initial deprotonation of the alcohols by two of the
basic Al-Me moieties, thereby eliminating 2 equiv of methane
to form (BINOLate)Al(Me) (1a). Upon the addition of 2-pro-
bridging alkoxides and is an inefficient catalyst, freshly pre-
pared, largely non-aggregated aluminum alkoxides are much
better, allowing for high yields of alcohols under mild re-
duction conditions (Scheme 1, left).17 Furthermore, a practical
catalytic asymmetric MSPV reduction of ketones was also
demonstrated: using 2-propanol as the achiral hydrogen
source and AlMe3/enantiopure 2,2′-dihydroxy-1,1′-binapthyl
8
(
BINOL)1 as the precatalyst mixture, good product yields, with
moderate to good enantioselectivites, were obtained (Scheme
i
panol, a third deprotonation occurs to yield (BINOLate)Al(O -
Pr) (1b) as the active monomeric catalytic species (Scheme 2).
1
, right).19
While the kinetic and mechanistic details of the classical
Al(OR)3-promoted MSPV reduction have been examined
SCHEME 2. Proposed Formation of the Catalytic
Precursor 1a and the Proposed Active Catalytic Species 1b
2
0-24
thoroughly,
little work has been performed regarding the
Al-catalyzed variants. We recently reported a combined theo-
retical and experimental study aimed at elucidating the operative
pathway of the hydride transfer, a key mechanistic step of the
BINOL/AlMe3-catalyzed MSPV reaction.25 However, that study
did not focus on ligand effects in catalysis, and experiments
aimed at elucidating key ligand requirements for both rate
acceleration and asymmetric induction were not carried out.
3
in the Reaction of BINOL with AlMe and 2-Propanol
(
10) Ooi, T.; Miura, T.; Maruoka, K. Angew. Chem., Int. Ed. 1998, 37,
347-2349.
11) Ooi, T.; Itagaki, Y.; Miura, T.; Maruoka, K. Tetrahedron Lett. 1999,
0, 2137-2138.
12) Ooi, T.; Miura, T.; Itagaki, Y.; Ichikawa, H.; Maruoka, K. Synthesis
002, 279-291.
13) Ooi, T.; Takahashi, M.; Yamada, M.; Tayama, E.; Omoto, K.;
Maruoka, K. J. Am. Chem. Soc. 2004, 126, 1150-1160.
14) Konishi, K.; Makita, K.; Aida, T.; Inoue, S. J. Chem. Soc., Chem.
Commun. 1988, 643-645.
15) Liu, Y.-C.; Ko, B.-T.; Huang, B.-H.; Lin, C.-C. Organometallics
002, 21, 2066-2069.
16) Ko, B.-T.; Wu, C.-C.; Lin, C.-C. Organometallics 2000, 19, 1864-
869.
2
4
2
(
(
(
(
Newly acquired support for the formation of a species similar
to 1a comes from titration. The H NMR spectrum of a 1:1
mixture of 1 and AlMe3 in THF-d8 exhibits a 4:1 integration
1
(
2
1
2
(
ratio for the aromatics:AlMe protons and no OH signal,
2
6
(
17) Campbell, E. J.; Zhou, H.; Nguyen, S. T. Org. Lett. 2001, 3, 2391-
suggesting a complete deprotonation. In an attempt to obtain
a crystal structure of 1a, equivalent portions of trimethylalu-
minum and (S)-BINOL were reacted together in dry THF under
N2. However, isolation of the crystalline product and subsequent
393.
(
18) Brunel, J. M. Chem. ReV. 2005, 105, 857-897.
(19) Campbell, E. J.; Zhou, H.; Nguyen, S. T. Angew. Chem., Int. Ed.
2
3
2
3
002, 41, 1020-1022.
20) Doering, W. v. E.; Aschner, T. C. J. Am. Chem. Soc. 1953, 75,
(
X-ray analysis yielded a dimeric aluminum structure, (S)-2
93-397.
27
(
Figure 1 and Table 1).
(
21) Shiner, V. J., Jr.; Whittaker, D. J. Am. Chem. Soc. 1963, 85, 2337-
338.
1
(
98.
22) Shiner, V. J., Jr.; Whittaker, D. J. Am. Chem. Soc. 1969, 91, 394-
(26) The H NMR spectrum of a 1:1 mixture of (S)-BINOL and AlMe3
in THF-d8 displayed the following resonances: 7.92 (broad, 3H, naphthyl),
7.32-6.85 (broad, 8H, naphthyl), 5.81 (broad, 1H, naphthyl), -0.76 (s,
3H, AlCH3).
(27) This dimeric aluminum complex is similar in bonding structure to
those reported by the Lin group (see refs 15 and 16).
(
(
(
23) Yager, B. J.; Hancock, C. K. J. Org. Chem. 1965, 30, 1174-1179.
24) Hach, V. J. Org. Chem. 1973, 38, 293-299.
25) Cohen, R.; Graves, C. R.; Nguyen, S. T.; Martin, J. M. L.; Ratner,
M. A. J. Am. Chem. Soc. 2004, 126, 14796-14803.
9122 J. Org. Chem., Vol. 72, No. 24, 2007

Mechanism of the Asymmetric MSPV Reduction
FIGURE 1. An ORTEP depiction of the crystal structure of the isolated aluminum dimer (S)-2 showing the atom labeling scheme and the asymmetric
subunit. Thermal ellipsoids are drawn at 50% probability. Hydrogens are omitted for clarity.
TABLE 1. Selected Bond Lengths (Å) and Angles (deg) for (S)-2
Additional support for the in situ formation of (S)-1b from
both the (S)-BINOL/AlMe3 mixture and (S)-2 comes from mass
spectral studies. In CH2Cl2, the (S)-BINOL/AlMe3 mixture was
treated with excess 2-propanol followed by direct analysis of
the solution via atmospheric pressure chemical ionization mass
spectroscopy (APCIMS). A clean parent ion at m/z 370.8 was
detected with an isotopic pattern similar to the simulated pattern
for C23H19AlO3, the calculated molecular weight of (BINOLate)-
bonding atoms
bond length (Å)
bond angle (deg)
Al1-O1
Al1-O2
Al1-O3
Al1-O2(2)
Al1-C25
1.7549(14)
1.9794(14)
1.9755(13)
1.8249(13)
1.961(2)
O1-Al1-O2
90.63(6)
89.73(6)
121.04(7)
74.50(6)
90.43(6)
122.00(9)
98.14(8)
96.43(8)
116.53(8)
104.37(6)
O1-Al1-O3
i
Al(O Pr) (see the Supporting Information). Direct mass spec-
O1-Al1-O2(2)
O2-Al1-O2(2)
O3-Al1-O2(2)
C25-Al1-O1
C25-Al1-O2
C25-Al1-O3
C25-Al1-O2(2)
Al1-O2-Al1(2)
trometric analysis of the reaction mixture between (S)-2 and
2-propanol in CH2Cl2 also showed a clean parent ion corre-
sponding to the molecular weight of C23H19AlO3, with a similar
isotopic pattern (see the Supporting Information). Presumably
the 2 equiv of 2-propanol that are loosely bound to the Al center
i
in the putative 1b‚2(HO Pr) solvated intermediate were lost upon
i
ionization and only the (BINOLate)Al(O Pr) core remains. These
results support our original model for catalyst formation
(Scheme 2), and suggest that the Al center has two additional
coordination sites available for bonding.
We hypothesized that the dimeric form of (S)-2 is the
energetically favored solid-state structure and that under the
conditions employed for the asymmetric MSPV reduction, the
dimer would break apart into the active monomeric species (S)-
Ligand-Accelerated Catalysis. While the chelation environ-
ment posed by the BINOL framework around Al centers has
been shown to significantly influence our asymmetric MSPV
reduction, further effects of phenoxide-based ligands in the Al-
catalyzed MSPV have not yet been elucidated. The initial rates32
of MSPV reduction of acetophenone with 2-propanol were
determined by using the in situ-formed BINOL/AlMe3 complex,
in situ-formed 2-naphthol (3, 2 equiv)/AlMe3 complex, in situ-
formed 2,2′-biphenol (BIPHEN, 4)/AlMe3 complex, and AlMe3
1
6
1
b, as has been shown to occur for similar aluminum species.
1
The H NMR spectrum of the (S)-BINOL/AlMe3 mixture in
CD2Cl2 showed a complete disappearance of the Al-Me signal
upon the addition of 2-propanol and a growing in of the
isopropoxide group, as was expected for (S)-1b. An equivalent
experiment performed with dimer (S)-2 in CD2Cl2 showed
analogous results, suggesting that the last demethylation step
for the dimer also occurs smoothly to give a species whose NMR
spectrum is identical with that for the (S)-BINOL/AlMe3/2-
propanol mixture. The 27Al NMR spectra for both aforemen-
tioned reactions also support this hypothesis. For the reaction
between the (S)-BINOL/AlMe3 mixture and 2-propanol in
CD2Cl2, a single, broad signal at 28 ppm was detected in its
2
7
27
Al NMR spectrum. Similarly, the Al NMR spectrum of a
solution of (S)-2 and 2-propanol in CD2Cl2 exhibited a single,
broad signal at 30 ppm. As 27Al NMR resonances in the mid-
FIGURE 2. Phenolic compounds employed as ligands in the Al-
catalyzed MSPV reduction of ketones.
2
0-50 ppm range are indicative of pentacoordinate aluminum
28-30
alkoxides,
the aforementioned data imply that similar
alone as catalytic precursors (Figure 3). Acetophenone was
reduced to sec-phenethyl alcohol in the presence of BINOL/
AlMe3 (Figure 3, B) at a rate that is approximately twice as
fast as that observed for AlMe3 (Figure 3, C). The analogous
MSPV reduction with the (2-naphthol)2/AlMe3 and BIPHEN/
AlMe3 mixtures as catalyst precursors proceeded at rates that
pentacoordinate aluminum centers are present under both
i
reaction conditions, presumably 1b‚2(HO Pr), where 2 equiv
of 2-propanol are loosely bound to the Al center in the putative
i
31
(
BINOLate)Al(O Pr) intermediate 1b.
(
28) Kriz, O.; Casensky, B.; Lycka, A.; Fusek, J.; Hermanek, S. J. Magn.
Reson. 1984, 60, 375-381.
(31) While the (S)-BINOL/AlMe3/2-propanol mixture and (S)-2/2-
propanol mixtures have slightly different 27Al resonances (28 and 30 ppm,
respectively), this difference is not significant given the inherent broadness
of the 27Al NMR signals.
(
(
29) Akitt, J. W. Prog. Nucl. Magn. Reson. Spectrosc. 1989, 21, 1-149.
30) Velez, K.; Quinson, J. F.; Fenet, B. J. Sol-Gel Sci. Technol. 1999,
28,29
1
6, 201-208.
J. Org. Chem, Vol. 72, No. 24, 2007 9123

Graves et al.
FIGURE 3. Initial rates of reaction for the MSPV reduction of acetophenone with 2-propanol, using BIPHEN/AlMe
2:1, B), BINOL/AlMe (1:1, C), and AlMe (D) as the catalytic precursors.
3
(1:1, A), 2-naphthol/AlMe
3
(
3
3
TABLE 2. Characterization Data for the in Situ-Generated Complexes between Ligands, AlMe3, and 2-Propanol
proposed solution structure
and 27Al NMR signal
proposed gas-phase structure
and calcd mol wt
APCIMS
results (m/z)
entry
1
ratio
i
i
i
i
1/AlMe3/HO Pr (1:1:40)
(BINOLate)Al(O Pr)(HO Pr)2,
8 ppm
(naphtholate)2Al(O Pr)(HO Pr)2,
4 ppm
(BIPHENate)Al(O Pr)(HO Pr)2,
1 ppm
(BINOLate)Al(O Pr),
370.8
372.9
270.1
2
C23H19AlO3, 370.38 g/mol
i
i
i
i
2
3
3/AlMe3/HO Pr (2:1:40)
(naphtholate)2Al(O Pr),
2
C23H21AlO3, 372.38 g/mol
i
i
i
i
4/AlMe3/HO Pr (1:1:40)
(BIPHENate)Al(O Pr),
3
C15H15AlO3, 270.26 g/mol
are 4 times and 30 times faster, respectively, than that observed
for AlMe3 (Figure 3, A and B). This observed ligand-
accelerated33 catalytic behavior suggests that phenolate ligands
can greatly influence the catalytic activity of aluminum alkoxides
in the MSPV reduction via both electronic and chelation effects.
Not only that decreasing the electron density around the
aluminum center with phenoxide ligands increases its reactivity
toward MSPV reduction, constraining the ligand environment
also increases its activity.
Additional support for similar Al coordination environments
in all three of the aforementioned cases came from mass spectral
studies. In CH2Cl2, the 2-naphthol/AlMe3/2-propanol mixture
exhibited a clean parent ion at m/z 372.9, as was expected for
i
the proposed solvent-free (naphtholate)2Al(O Pr) species (Table
2, entry 3). Analysis of the 4/AlMe3/2-propanol mixture in CH2-
Cl2 gave a similar result: a parent ion at m/z 270.1, the
calculated molecular weight for the solvent-free (BIPHENate)-
i
Al(O Pr) species, was observed with an isotopic pattern similar
To verify that the observed rate differences shown in Figure
were indeed results of the different ligand environments and
to the generated pattern (Table 2, entry 3). Together with the
aforementioned kinetic data, these experiments indicate that not
only is the phenol/AlMe3/2-propanol system ligand-accelerated,
but variation in denticity (mono- versus bidentate) and steric
constraints of the phenolic ligand, also have important rate
consequences in the MSPV reduction.
3
not simply due to differences in aluminum coordination mode,
2
7
we turned to Al NMR spectroscopy. Solutions of [2-naphthol
2 equiv) + AlMe3 + 2-propanol (40 equiv)] and [BIPHEN +
(
AlMe3 + 2-propanol (40 equiv)] in CD2Cl2, similar to the
conditions employed in the BINOL/AlMe3/2-propanol MSPV
reduction chemistry, were prepared and analyzed directly via
Kinetics of the Asymmetric MSPV Reduction Catalyzed
by BINOL/AlMe3. Given the structure of intermediate 1b, a
reasonable empirical rate law for reaction 1 can be expressed
in terms of three concentrations: aluminum catalyst, substrate,
and 2-propanol (eq 3). Through a series of pseudo first-order
kinetic studies, the order with respect to each reactant (x, y, z)
2
7
Al NMR spectroscopy (Table 2, entries 2 and 3, respectively).
27
That the solutions exhibited signals in their Al NMR spectra
at ∼30 ppm, and that these results are equivalent to those
observed for the BINOL/AlMe3/2-propanol combination, sup-
3
4
ports the generation of analogous pentacoordinate aluminum
may be isolated and subsequently determined.
species2
8,30
in all three cases, each of which is presumably
x
y
z
coordinated by 2 equiv of 2-propanol in addition to three
alkoxide/aryloxide ligands.
rate ) ν ) k[catalyst] [substrate] [2-propanol]
(3)
Determination of the Order of Aluminum Catalyst:
Discrimination between Monomeric and Dimeric Catalyst
Activation. While it is generally accepted that the MSPV
(
32) For this kinetic study, the initial rate of reaction was determined by
GC analysis of aliquots taken from the reaction mixture at the appropriate
time. Rate constants were determined from a plot of product yield versus
time up to 20% yield.
(
33) Berrisford, D. J.; Bolm, C.; Sharpless, K. B. Angew. Chem., Int.
(34) Atkins, P. W. Physical Chemistry, 6th ed.; W. H. Freeman and
Company: New York, 1998; p 766.
Ed. 1995, 34, 1059-1070.
9124 J. Org. Chem., Vol. 72, No. 24, 2007

Mechanism of the Asymmetric MSPV Reduction
2
FIGURE 4. Plots of ln[kobs] versus ln[catalyst] for the (BINOL)Al-catalyzed MSPV reduction of R-bromoacetophenone (R ) CH Br, A) and
acetophenone (R ) CH , B) with 2-propanol.
3
FIGURE 5. Investigation for a nonlinear effect in the BINOL/Al-catalyzed MSPV reduction of R-bromoacetophenone (A) and acetophenone (B).
reduction proceeds through a monometallic pathway in which
both the carbonyl substrate and alkoxide are coordinated to a
single aluminum center, recent reports suggest that substrate
is reasonable to conclude that the (BINOLate)Al-catalyzed
MSPV reduction proceeds through a monometallic catalytic
intermediate.
1
0,12
activation by two aluminum centers is also a viable option.
For a number of asymmetric catalytic reactions where two
chiral metal centers act cooperatively, a linear variation of the
catalyst enantiopurity can result in a nonlinear variation in the
To discriminate between these two possibilities for the BINOL/
AlMe3 system, we first determined x, the catalyst order, by
holding the concentration of both substrate and 2-propanol in
high excess relative to catalyst. The initial rates of the (BINOL)-
Al-catalyzed MSPV reduction of acetophenone and R-bromoac-
etophenone with 2-propanol were measured independently at
various catalyst concentrations (12.5-250 mM, see Table S1
in the Supporting Information). Subsequent plots of ln[kobs]
versus ln[catalyst concentration] for both sets of experiments
yielded catalyst orders of 0.89 (Figure 4). As orders close to 1
are most logically explained by a unimolecular dependence, it
3
5
enantiopurity of the product. To this end, a series of MSPV
reductions of R-bromoacetophenone and acetophenone were
carried out independently by using the in situ-generated catalyst
1b with varying enantiomeric purity of BINOL ligand (0-
100%). The linear relationship between ligand and product ee
values (Figure 5) indicates the absence of a nonlinear effect in
the (BINOL)Al-catalyzed MSPV reduction, further supporting
a mechanism in which two metal centers are not acting in
concert.
J. Org. Chem, Vol. 72, No. 24, 2007 9125

Graves et al.
FIGURE 6. Plot of ln[kobs] versus ln[substrate] for the (BINOL)Al-catalyzed MSPV reduction of R-bromoacetophenone (R ) CH
acetophenone (R ) CH , B) with 2-propanol.
2
Br, A) and
3
While we were reasonably confident at this point that the
BINOL)Al-catalyzed MSPV reduction proceeds at a single
and using a constant catalyst loading, the initial reduction rates
were determined at various 2-propanol concentrations (62.5-
437.5 mM, see Table S3 in the Supporting Information). A
subsequent plot of the ln[kobs] versus ln[2-propanol] was
generated (Figure 7), both of which indicate an inverse
dependence on 2-propanol concentration.
(
metal center, isolation of the dimeric species (S)-2 provided an
opportunity to investigate whether such a species can act as a
possible MSPV catalyst. In the presence of DMSO, dimer (S)-2
can be broken into a monomeric ((S)-BINOL)AlMe(DMSO)
complex, similar to that observed by the Lin group for their
This inverse dependence on 2-propanol can be explained via
the equilibrium depicted in Scheme 3. As suggested by the
[
(µ-MMPEP)AlMe]2 dimer (MMPEP ) 2,2-methylenebis(4,6-
15
di(1-methyl-1-phenylethyl)phenolate). Both (S)-2 and ((S)-
BINOL)AlMe(DMSO) can catalyze the reduction of R-bro-
moacetophenone by 2-propanol and the product ee values in
both cases were determined to be 79% in the S configuration,
the same as that found with the in situ-generated catalyst (S)-
27
aforementioned Al NMR experiments, the maximum coordi-
nation number that the aluminum center can obtain under the
(BINOLate)Al-catalyzed MSPV reaction conditions is five.
Once the active catalytic species 1b is formed, excess 2-propanol
can bind to vacant coordination sites on the aluminum center,
1
9
1
b. That the reduction with dimer (S)-2 and ((S)-BINOL)-
i
giving the tetracoordinate intermediate (1b‚(HO Pr)); further
AlMe(DMSO) gave results analogous to the reduction with the
in situ-formed catalyst (S)-1b further supports formation of
identical aluminum species for all cases under typical (BINO-
Late)Al-catalyzed MSPV reaction conditions. Together with the
aforementioned kinetics, NMR, and MS studies, these results
lend credence toward the breakdown of dimer (S)-2 into the
active monomeric catalytic species (S)-1b under MSPV reaction
conditions.
Determination of Reaction Order with Respect to Sub-
strate and 2-Propanol. Convinced that the BINOL/AlMe3-
catalyzed asymmetric MSPV reduction was proceeding through
a monometallic mechanism, we set out to determine the reaction
order with respect to both the ketone substrate and 2-propanol
reductant. At a constant catalyst loading and an excess of
coordination by 2-propanol would give the saturated, pentaco-
i
ordinate species (1b‚2(HO Pr)). As the aluminum center in 1b‚
i
2
(HO Pr) is fully saturated, there are no sites available for
substrate to coordinate and the MSPV pathway shuts down. At
increasing concentrations of 2-propanol, the equilibrium depicted
in Scheme 3 shifts to the right and the inactive aluminum species
becomes more prevalent, decreasing the observed reduction
rate.
The above outlined equilibrium also helps explain the ca.
zeroth-order rate dependence on substrate. At the large 2-pro-
panol concentration employed in the pseudo first-order kinetic
conditions used to determine substrate order, one would expect
the major aluminum species in solution to be the MSPV-inactive
i
1
b‚2(HO Pr). This equilibrium condition suggests that once
2
-propanol, the initial rates of reduction for acetophenone and
acetophenone displaces the excess 2-propanol, reduction is rapid.
At the high 2-propanol concentrations relative to substrate for
all of the reactions used in this study, one would not expect to
observe a major rate difference over this concentration range.
R-bromoacetophenone were determined at various substrate
concentrations (62.5-375.5 mM, see Table S2 in the Supporting
Information). Surprisingly, over this concentration range there
was very little change36 in reaction rate with increasing substrate
concentration, as is further depicted in a plot of the ln[kobs]
versus ln[substrate] (Figure 6), indicating that the reaction rate
is independent of substrate concentration.
Intrigued by these results, we investigated the rate dependence
of 2-propanol. Holding the substrate concentration in high excess
(35) Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of Organic
Compounds; John Wiley and Sons, Inc.: New York, 1994; p 1267.
(36) The slight deviation from linearity observed at higher substrate
concentrations is presumably due to reaction conditions that are no longer
strictly pseudo first-order (ca. only 6-fold excess of 2-propanol).
9126 J. Org. Chem., Vol. 72, No. 24, 2007

Mechanism of the Asymmetric MSPV Reduction
2
FIGURE 7. Plot of ln[kobs] versus ln[2-propanol] for the (BINOL)Al-catalyzed MSPV reduction of R-bromoacetophenone (R ) CH Br, A) and
acetophenone (R ) CH , B) with 2-propanol.
3
FIGURE 8. Reaction yield (2 and 9, grey) and product ee (b and [) as a function of time for the MSPV reduction of R-bromoacetophenone (R
CH Br; yield line A (2); ee line B (b)) and acetophenone (R ) CH ; yield line C (9); ee line D ([)) catalyzed by (S)-BINOL/AlMe
)
2
3
3
.
SCHEME 3. Proposed Equilibrium between the Active Aluminum Alkoxide Catalyst 1b and Excess 2-Propanol
Reversibility in the Asymmetric BINOL/AlMe3-Catalyzed
reaction must effectively overcome the potential for product
racemization through the reversible transfer hydrogenation
processes. While use of excess 2-propanol should shift equi-
MSPV Reduction. Given the inherent reversibility4
,6,37
of the
MSPV reduction, a successful asymmetric variant of this
J. Org. Chem, Vol. 72, No. 24, 2007 9127

Graves et al.
SCHEME 4. Proposed Formation of the Catalytic Precursors 1a, 5a, and 6a and the Proposed Precatalysts 1b, 5b, and 6b
3
Formed between the Reaction of AlMe with (a) BINOL, (b) Monoether-BINOL, and (c) Diether-BINOL, Respectively, in the
Presence of 2-propanol
librium toward product, the potential for reaction reversibility,
especially with the build-up of reaction products, may lead to
a decrease in enantiomeric excess of product over time. To
verify that this was not occurring in the BINOL/AlMe3 system,
the reduction of R-bromoacetophenone catalyzed by in situ-
formed (S)-1b with a 4-fold excess of 2-propanol was carried
out and monitored over time (Figure 8, A and B). The en-
antiomeric purity of product remained unchanged (81 ( 2%)
after 3 days (72 h). Similarly, the reduction of acetophenone,
that exhibit different coordination environments under typical
(BINOLate)Al-catalyzed MSPV reaction conditions. Initial
reaction between AlMe and 5 would release 1 equiv of methane
3
to afford the mono-alkoxide 5a, which yields 5b as a putative
tetracoordinate precatalyst upon addition of excess 2-propanol
(Scheme 4b). The formation of 6b from 6 follows a similar
pathway, except that the addition of AlMe3 to a solution of 6
should not expel methane; rather, it would lead to dative coor-
dination of the ligand to give the pentacoordinate metal species
which is at a lower oxidation potential than R-bromoacetophen-
6
a, which exhibits three Al-Me functionalities. Upon addition
one,3
8-40
is not debilitated by reversibility (Figure 8, C and D).
of 2-propanol, these basic moieties would be protonated to form
The enantiomeric excess of reduction product remained at ∼35%
throughout the reduction and over time, indicating that revers-
ibility, if present, does not influence the asymmetric outcome
of the ((S)-BINOL)Al-catalyzed MSPV reduction under experi-
mental conditions employed for acetophenone-type substrates.
Effect of Ligand Coordinating Ability in the Asymmetric
MSPV Reduction. Having established the importance of bis-
chelating phenol-based ligands in enhancing the activity of the
Al-catalyzed MSPV reduction, we compare the binding mode
the pentacoordinate aluminum alkoxide 6b (Scheme 4c).
We previously reported that the combination of AlMe3 with
enantiomerically pure ligands (S)-5 and (S)-6 produced inferior
catalysts (relative to the parent BINOL/AlMe3 system) for the
MSPV reduction of R-bromoacetophenone with 2-propanol, in
1
9
both yield and selectivity. Given the varying coordination
environments proposed in Scheme 4 for the in situ-generated
catalytic species 1b, 5b, and 6b, we originally proposed three
different intermediates for hydrogen transfer (tetra-, penta-, and
hexacoordinate, Scheme 5, parts a, b, and c, respectively). On
the basis of these models, we inferred that any Al-based MSPV
catalyst that does not employ a tetrahedral geometry at the point
for hydrogen transfer would lead to a decreased enantioselec-
4
1
2
of 1 to the Al center with two other BINOL derivatives: 5
has one phenol moiety paired with a phenol ether, while 6
4
has both of its phenol functionalities masked as ethers (Scheme
). Given our proposal for catalyst formation between BINOL,
4
AlMe3, and 2-propanol (vide supra), reaction of 5 and 6 with
AlMe3 and 2-propanol should result in aluminum intermediates
4
3
tivity. That suggestion is now invalidated by newly acquired
characterization data indicating a pentacoordinate catalytic
species that exists in solution for the BINOL/AlMe3/2-propanol
mixture (vide supra).
(
37) For recent examples of Oppenauer oxidation protocols see: (a) Ooi,
T.; Otsuka, H.; Miura, T.; Ichikawa, H.; Maruoka, K. Org. Lett. 2002, 4,
669-2672. (b) Graves, C. R.; Zeng, B.-S.; Nguyen, S. T. J. Am. Chem.
2
Soc. 2006, 128, 12596-12597.
(
38) Adkins, H.; Cox, F. W. J. Am. Chem. Soc. 1938, 60, 1151-1159.
(43) We originally proposed that a tetracoordinate intermediate was
operative for the BINOL/AlMe3/2-propanol system. On the basis of data
showing reduced ee,19 we suggested a pentacoordinate intermediate for the
MSPV reduction employing the 5/AlMe3/2-propanol combination and a
hexacoordinate intermediate for the 6/AlMe3/2-propanol combination, and
that these higher coordination intermediates resulted in inferior asymmetric
catalysts.
(39) Adkins, H.; Elofson, R. M.; Rossow, A. G.; Robinson, C. C. J.
Am. Chem. Soc. 1949, 71, 3622-3629.
(
(
(
40) Cox, F. W.; Adkins, H. J. Am. Chem. Soc. 1939, 61, 3364-3370.
41) Wipf, P.; Jung, J.-K. J. Org. Chem. 2000, 65, 6319-6337.
42) Takahashi, M.; Ogasawara, K. Tetrahedron: Asymmetry 1997, 8,
3
125-3130.
9128 J. Org. Chem., Vol. 72, No. 24, 2007

Mechanism of the Asymmetric MSPV Reduction
SCHEME 5. Previously Proposed Intermediates19 for the MSPV Reduction of Ketones with 2-Propanol Catalyzed by 1b, 5b,
and 6b
In support of the hypothesis that the solvent-free species 5b
can be formed under the experimental conditions, the APCIMS
spectrum of the 5/AlMe3/2-propanol mixture exhibits a clean
molecular ion at m/z equal to 444.7, with a matching isotopic
indicate a pentacoordinate aluminum center in solution, similar
to that observed for the 1/AlMe3/2-propanol mixture. That
similar pentacoordinate Al-environments exist for both the (S)-1
and (S)-5 cases implies that a tetracoordinate intermediate is
not an absolute prerequisite for a stereoselective MSPV reduc-
tion, and coordination number alone does not sufficiently explain
the observed asymmetric results.
27
pattern to that generated for C27H29AlO4. Further, the Al NMR
spectrum of this mixture exhibits a resonance signifying a
pentacoordinate aluminum species (31 ppm), consistent with a
structure of 5b surrounded by one solvent molecule. While 6 is
not soluble in toluene, addition of AlMe3 affords solubilization
of the ligand, suggesting that a ligand-metal interaction has
A modified explanation for the decreased product ee, as well
as the increased reaction rate, obtained for MSPV reductions
employing the (S)-5/AlMe3 mixture as a precatalyst is as follows.
In solution, the etheric oxygen of (S)-5 is loosely bound to the
aluminum center, similar to that proposed for the (S)-6 ligand
2
7
occurred and supporting the formation of 6a, as does the Al
spectrum of the 6/AlMe3 mixture in CD2Cl2, which exhibits a
signal at 35 ppm indicative of a pentacoordinate aluminum
center. However, upon the addition of a small amount of
(
(
vide supra), and can easily be displaced by the large excess
40 equiv relative to catalyst) of 2-propanol employed in our
2
-propanol, the solution becomes immediately heterogeneous,
MSPV reduction. This shifts the bidentate chiral framework
away from the aluminum center and results in a decrease in
selectivity for the system. However, because the aluminum
center becomes less sterically congested while still maintaining
a phenolate linkage, the rate of acetophenone reduction would
increase relative to that observed for BINOL (Scheme 6). In
this sense, the monoether BINOL ligand (S)-5 essentially acts
as a sterically congested 2-naphthol ligand (vide supra). This
result also implies that enantiomerically pure, monodentate
alcohols may not impart sufficient chirality upon the aluminum
center, and would be inferior ligands for the asymmetric MSPV
reduction.
suggesting that protonation of the Al-Me functionalities to form
the corresponding alkoxides results in expulsion of the datively
bound 6, and the ligand is not a participant in the final aluminum
complex.44
The MSPV reduction of acetophenone with 2-propanol was
carried out employing the in situ-derived species formed
between AlMe3 and either (S)-1, (S)-5, or (S)-6 (Figure 9).
Together with the aforementioned characterization data, the
minimal enantiomeric excess in product obtained in the (S)-6/
AlMe3-catalyzed reduction further supports the dissociation of
the ligand from the aluminum center under the MSPV reaction
conditions. Indeed, the observed rate of reduction with this
ligand (Figure 9, C) was quite close to the rate observed for
the unligated, AlMe3-catalyzed reduction of acetophenone
Substitution of the BINOL Framework. Our experiments
thus far suggest that weak dative ligand coordination, such as
those based on ether linkages between the BINOL ligand and
aluminum, are insufficient for chiral induction in the MSPV
reduction, presumably due to competitive coordination by the
excess 2-propanol reductant. As diprotic BINOL-derived ligands
were most optimal among the phenol-based ligands that we
investigated for the asymmetric MSPV reduction, we hypoth-
esized that incorporation of alkyl groups in the 3 and 3′ positions
of the BINOL framework could potentially increase the steric
bulk surrounding the Al center and lead to both a tighter
transition state and a more selective catalytic species.
(Figure 3, D).
While the (S)-5/AlMe3/2-propanol mixture afforded a much
more active catalyst for the MSPV reduction of acetophenone
than the (S)-BINOL/AlMe3/2-propanol equivalent, product ee
was substantially reduced (Figure 9, A and B, respectively).
As mentioned above, we originally suggested that this decreased
selectivity was a result of deviation from an optimal tetrahedral
aluminum center during the hydrogen transfer process (Scheme
4
4
27
5
a,b). However, this may be an incorrect assessment: Al
NMR data for the 5/AlMe3/2-propanol combination clearly
The MOM-protected analogue of the enantiomerically pure
C1-symmetric, 3-methylated BINOL derivative (S)-7 was syn-
thesized from methoxymethoxy (MOM)-protected (S)-BINOL
(
44) Combinations of 6/AlMe3/2-propanol and AlMe3/2-propanol in
27
CD2Cl2 exhibit similar signals in their Al NMR spectra, further supporting
this hypothesis.
n
via lithiation with 1 equiv of BuLi and subsequent trapping of
J. Org. Chem, Vol. 72, No. 24, 2007 9129

Graves et al.
FIGURE 9. Initial rates of reaction for the MSPV reduction of acetophenone with 2-propanol, using (S)-5/AlMe
S)-6/AlMe (C) as the catalytic precursors.
3 3
(A), (S)-BINOL/AlMe (B), and
(
3
SCHEME 6. Proposed Scheme Showing the Dissociation of the Al-Ether Dative Bond in the Complex Formed between (S)-5
and AlMe in the Presence of Excess 2-Propanol
3
the resulting anion with methyl iodide. Deprotection of the
-methylated MOM-protected alcohol groups afforded (S)-7.
The MSPV reduction of R-bromoacetophenone with 2-pro-
panol was carried out with the in situ-formed complexes between
AlMe3 and ligands (S)-7 through (S)-10 (Table 3). While all
the ligands afforded good yields of the reduced product, the
selectivities of reduction were servely diminished with respect
to the parent BINOL/AlMe3 catalyst. While the C2-symmetric
disubstituted ligands (S)-8 and (S)-10 afforded 50% and 26%
ee of the product, respectively (Table 3, entries 3 and 5), their
monosubstituted counterparts (S)-7 and (S)-9 afforded no
enantioselectivity (Table 3, entries 2 and 4), suggesting that C1-
symmetric ligands are inappropriate choices for the asymmetric
MSPV reduction.
3
The C2-symmetric 3,3′-dimethylated BINOL variant (S)-8 was
n
synthesized via a similar procedure, except that excess BuLi
was used to afford deprotonation in both the 3 and 3′ positions
of the MOM-protected (S)-BINOL, followed by anion trapping
with excess methyl iodide and subsequent deprotection of the
MOM groups. The monoethyl- ((S)-9) and diethyl-BINOL ((S)-
10) derivatives were synthesized in an analogous manner, using
ethyl iodide as the anion trap.
Interestingly, aluminum complexes incorporating (S)-7 through
(S)-10 catalyzed the MSPV reduction of acetophenone with rates
that are more than 1 order of magnitude faster than that for
BINOL itself, albeit with a lower ee (Table 4 and Figure 11).
Given this dramatic rate increase, we initially hypothesized that
the increased steric bulk surrounding the phenol moieties of
FIGURE 10. Alkyl-substituted BINOL derivatives.
9130 J. Org. Chem., Vol. 72, No. 24, 2007

Mechanism of the Asymmetric MSPV Reduction
3 3 3
FIGURE 11. Initial reaction rates of the MSPV reduction of acetophenone with 2-propanol, using (S)-7/AlMe (D), (S)-8/AlMe (A), (S)-9/AlMe
3
(B), (S)-10/AlMe (C), and (S)-BINOL (E) as the catalytic precursors. Line E was generated from the rate data obtained for (S)-BINOL in Figure
3
.
TABLE 3. The MSPV Reduction of r-Bromoacetophenone
Catalyzed by the in Situ- formed Complex between AlMe3 and
Substituted BINOL Derivatives (S)-7 through (S)-10
SCHEME 7. Possible Modes of Reactivity between AlMe3
and the Substituted BINOL Derivative (S)-7 through (S)-10
entry
ligand
GC yield [%]
ee [%]
1
2
3
4
5
(S)-BINOL
(S)-7
(S)-8
(S)-9
(S)-10
99
98
98
97
97
83 (S)
0
50 (S)
0
26 (S)
TABLE 4. Rate and Selectivity Data for the MSPV Reduction of
Acetophenone with Substituted BINOL-derived Ligands
line fit parameter, R2 10 kobs (s-1) ee (%)
4
entry
ligand
1
2
3
4
5
(S)-BINOL
(S)-7
(S)-8
(S)-9
(S)-10
E
9.6 ( 0.1
158 ( 3
333 ( 2
20 ( 2
35
13
15
6
D
A
B
C
0.981
0.993
0.990
0.996
192 ( 1
15
indicates a pentacoordinate geometry under MSPV reaction
2
7
conditions, the Al NMR spectrum of a 1:1:40 mixture of (S)-
(
S)-7 through (S)-10 (relative to BINOL) could potentially inhibit
8
/AlMe3/2-propanol exhibited a signal at 70 ppm, indicative of
the full deprotonation of the ligand by AlMe3, thereby preventing
the formation of a species similar in structure to 1a. Such a
situation could result in the formation of tetracoordinate,
monodeprotonated intermediates (Scheme 7) analogous to that
formed between 5 and AlMe3 (vide supra), and hence may not
result in the most selective catalysts.
2
8,30
a tetracoordinate aluminum species.
found for 27Al NMR experiments conducted with (S)-7 (62
ppm), (S)-9 (65 ppm), and (S)-10 (74 ppm).
Similar results were
The dramatic increase in rate for catalysts derived from (S)-
8
/AlMe3 and (S)-10/AlMe3, relative to the (S)-BINOL/AlMe3
combination, can be explained by their adoption of a tight
tetrahedral geometry in the hydrogen transfer step. With the
BINOL-derived catalyst, the less sterically encumbered alumi-
num site allows for deviation from this “preferred” geometry
to a more sterically bulky pentacoordinate species, a geometry
that previously has been suggested to be less active in the MSPV
To evaluate the possibility that incomplete deprotonation may
have occurred for (S)-7 through (S)-10, we turned to NMR
spectroscopy. The H NMR spectra of a 1:1 mixture of (S)-8
and AlMe3 in THF-d8 exhibited an AlMe peak at 0.95 ppm (10:3
integration ratio for the aromatics:AlMe protons) and does not
contain a resonance that can be assigned to OH protons,
suggesting a complete deprotonation of the ligand (Figure 12).
Unlike the equivalent experiment conducted with 1, which
1
1
5,16
reduction.
Presumably, the tighter tetrahedral arrangement
would place the transferring hydrogen in closer proximity to
J. Org. Chem, Vol. 72, No. 24, 2007 9131

Graves et al.
case. Indeed, such a 2-point binding model between a substrate
i
and (BINOLate)Al(O Pr) has been invoked recently to explain
the high enantioselectivity observed in the MSPV reduction of
4
7
N-phosphinoyl ketimines.
That the enantioselectivities obtained in the aforementioned
stoichiometric imine MSPV reduction were considerably higher
(
93-98%) than that for the analogous, but catalytic, reduction
of the R-halogenated carbonyl substrates can be rationalized via
the relative binding strengths of the two substrate classes to
the Al-center in (BINOLate)Al(O Pr). The N-phosphinoyl imines
i
bind to the metal in a nonreversible, Lewis acid-Lewis base
fashion, as is evidenced by the need for a stoichiometric amount
of the aluminum reagent before the reduced amide product can
be obtained in good yields. This nonlabile two-point binding,
comprising both a strong AlrN and a strong AlrOdP bond
(Figure 13b), results in a highly ordered transition state and
thus a high level of enantioselectivity. Conversely, the R-ha-
logenated carbonyl substrates can be reduced in high yield by
using only 10 mol% of the BINOL/AlMe3 mixture, indicating
a more labile binding of this substrate class to the Al-center
via a weak AlrOdC bond and a weak AlrX bond. While
this reversible binding allows for catalyst turnover, it also holds
the substrate less strongly to the Al-center, leading to an overall
lower selectivity in the reduction. Further experiments to
quantify this situation, as well as the evaluation of other ketone
substrates capable of 2-point coordination, are currently under-
way.
1
FIGURE 12. The H NMR spectrum of the reaction between (S)-8
and AlMe indicating complete deprotonation of the ligand.
3
Conclusion
In conclusion, we have obtained evidence to support the
conjecture that the in situ reaction between 2,2′-dihydroxy-1,1′-
binapthyl, AlMe3, and 2-propanol gives rise to a monomeric,
chiral aluminum alkoxide that acts as the catalyst for the MSPV
reduction of ketones with 2-propanol. Through various kinetic
studies, we came to the conclusion that this catalyzed MSPV
reduction is likely to proceed through a monometallic activation
of the substrates with a direct, concerted hydrogen transfer from
the bound alkoxide to the carbonyl carbon. Importantly,
reversibility does not seem to plague the asymmetric efficacy
of the BINOL/AlMe3 catalyst system under the employed MSPV
experimental conditions, with the enantiomeric excess of the
reduced product remaining constant for many hours after the
reduction. The use of a high excess of 2-propanol inhibits the
aluminum-catalyzed MSPV reduction, presumably locking the
aluminum center in an inactive, substrate-less, pentacoordinate
state.
FIGURE 13. (a) Proposed 2-point coordination of R-halogenated
carbonyl substrates in the asymmetric MSPV reduction catalyzed by
i
(
BINOLate)Al(O Pr), which is similar to (b) the proposed 2-point
i
coordination of N-phosphinoyl ketimines to (BINOLate)Al(O Pr) in its
4
7
stoichiometric reduction by the latter reagent. The models shown
above correctly predict the experimentally observed sense of enan-
tioenrichment for the reduction of both R-halogenated acetophenone
4
7
i
and N-phosphinoyl ketimines at the (BINOLate)Al(O Pr) center.
the carbonyl carbon than in the corresponding pentacoordinate
arrangement, and hence accelerates the reaction rate accordingly.
The aforementioned difference in coordination sphere also
helps to explain the decreased enantioselectivity obtained in the
MSPV reduction of R-bromoacetophenone for the (S)-8/AlMe3-
and (S)-10/AlMe3-catalyzed system relative to the (S)-BINOL/
From a ligand design perspective, the use of bidentate, chiral
phenol-based ligands in aluminum-catalyzed MSPV reduction
not only imparts the necessary asymmetric environment around
the aluminum center, but also dramatically increases the reaction
rate. While BINOL ether ligands can also lead to an active
catalytic aluminum species, the inherent lower coordinating
ability of the ether group, in comparison to the anionic alkoxy
moiety in the parent system, is a liability in the asymmetric
realm, leading to decreased product selectivities. In a similar
vein, crowding the ligand environment around the Al center
with substitutions of the BINOL framework can enforce a
tetrahedral intermediate that may lead to a faster reduction rate
at the expense of selectivity. Most importantly, the ability of
19
AlMe3-catalyzed system. Given the ability of the (BINOLate)-
i
Al(O Pr) species to coordinate two additional solvent molecules
(vide supra), an operative explanation for the high selectivity
observed for (S)-BINOL/AlMe3 with R-halogen-containing
substrates has been the ability for formation of a weak but
important, 2-point coordination between the ketone substrate
and the aluminum center, similar to that proposed for transition-
metal-catalyzed directed hydrogenations,4 thereby facilitating
a more selective reduction (Figure 13a). With the (S)-8/AlMe3/
5,46
2
-propanol and (S)-10/AlMe3/2-propanol systems, where tetra-
coordinate intermediates have been suggested (vide supra), this
-point coordination is not available due to steric crowding, and
the enantiomeric excess of product would be depleted, as is the
i
the (BINOLate)Al(O Pr) center to accommodate 2-point coor-
2
dination with a substrate, as in the case of R-halogenated ketones
or N-phosphinoyl ketimines, is critical to the generation of high
9132 J. Org. Chem., Vol. 72, No. 24, 2007

Mechanism of the Asymmetric MSPV Reduction
enantiomeric excess in product. This multidentate substrate
coordination mode, akin to those seen in substrate-specific
biological catalysts, subtly balances catalytic activity and
enantioselectivity in the BINOL/AlMe3 system and suggests an
additional strategy that can be exploited for developing highly
selective functional group transformations.
) 100 °C, final time ) 2 min; retention times: sec-phenethyl
alcohol ) 4.24 min, acetophenone ) 4.40 min. GC method for
R-bromoacetophenone reductions: start temperature ) 110 °C,
initial time ) 4 min, ramp ) 5 deg/min, final temperature )
1
50 °C, final time ) 2 min; retention times: R-bromomethylbenzyl
alcohol ) 6.07 min, R-bromoacetophenone ) 6.29 min. Chiral GC
method for acetophenone reductions: start temperature ) 80 °C,
initial time ) 20 min, ramp ) 5 deg/min, final temperature )
Experimental Section
170 °C, final time ) 10 min; retention times of sec-phenethyl
General Procedure for the Reduction of Ketone Substrates
by 2-Propanol Catalyzed with the in situ-Formed Complexes
between Ligands and AlMe . In an inert-atmosphere glovebox,
3
an 8-mL vial equipped with a magnetic stirring bar was charged
with ligand (0.1-0.2 mmol) and internal standard-containing
anhydrous toluene (4 mL). AlMe (9.6 µL, 0.1 mmol) was added
3
alcohol enantiomers: 29.72 min (S) and 30.32 min (R). Chiral GC
method for R-bromoacetophenone reductions: start temperature )
100 °C, initial time ) 0 min, ramp ) 2 deg/min, final temperature
) 170 °C, final time ) 10 min; retention times of R-bromometh-
ylbenzyl alcohol enantiomers: 34.95 min (S) and 35.57 min (R).
via syringe and the reaction was stirred. After 0.5 h, 2-propanol
305 µL, 4.0 mmol) was added, the vial was capped with a Teflon-
lined silicone septa, and the reaction was stirred for an additional
.5 h. Either acetophenone (115 µL, 1.0 mmol) or R-bromoac-
Acknowledgment. We gratefully acknowledge financial
support from the NSF and the DOE through a grant administered
by the Northwestern University Institute for Catalysis in Energy
Processes. S.T.N. additionally acknowledges support from the
Packard Foundation and the A. P. Sloan Foundation. C.R.G.
thanks the Natural Sciences and Engineering Research Council
of Canada (NSERC) for an International (PGS-D2) Postgraduate
Fellowship. We thank Ms. Tendai Gadzikwa for assistance with
the APCIMS experiments.
(
0
etophenone (198 mg, 1.0 mmol) was added neat; the reaction was
taken out of the glovebox and stirred at room temperature. At the
indicated time, a 100-µL aliquot of the reaction was collected with
a gas-tight syringe, loaded onto a plug of alumina, rinsed with
methanol (15 mL), and analyzed directly by GC and chiral GC.
GC method for acetophenone reductions: start temperature )
80 °C, initial time ) 4 min, ramp ) 5 deg/min, final temperature
Supporting Information Available: General synthetic proce-
dures, detailed experimental conditions, kinetic data, and charac-
terization data, as well as XYZ coordinates of the crystal structure
(S)-2 (CIF). This material is available free of charge via the Internet
at http://pubs.acs.org.
(
45) King, S. A.; Thompson, A. S.; King, A. O.; Verhoeven, T. R. J.
Org. Chem. 1992, 57, 6689-6691.
46) Mashima, K.; Kusano, K.-H.; Sate; Matsumura, Y.-I.; Nozaki, K.;
(
Kumobayashi, H.; Sayo, N.; Hori, Y.; Ishizaki, T.; Akutagawa, S.; Takaya,
H. J. Org. Chem. 1994, 59, 3063-3076.
(47) Graves, C. R.; Scheidt, K. A.; Nguyen, S. T. Org. Lett. 2006, 8,
1
229-1232.
JO070563U
J. Org. Chem, Vol. 72, No. 24, 2007 9133