Heterogeneously catalyzed asymmetric C=C hydrogenation: Origin of enantioselectivity in the proline-directed Pd/isophorone system

Article abstract of DOI:10.1021/ja061104y

We have studied the proline-directed, Pd-catalyzed enantioselective hydrogenation of isophorone in the liquid state using a variety of methods. Our results unambiguously reveal the true reaction pathway and demonstrate that all earlier mechanistic hypothe

Full text of DOI:10.1021/ja061104y

Published on Web 05/12/2006
Heterogeneously Catalyzed Asymmetric CdC Hydrogenation:
Origin of Enantioselectivity in the Proline-Directed Pd/
Isophorone System
Alexander I. McIntosh, David J. Watson, Jonathan W. Burton, and
Richard M. Lambert*
Contribution from the Department of Chemistry, UniVersity of Cambridge,
Cambridge CB2 1EW, United Kingdom
Received February 15, 2006; E-mail: RML1@cam.ac.uk
Abstract: We have studied the proline-directed, Pd-catalyzed enantioselective hydrogenation of isophorone
in the liquid state using a variety of methods. Our results unambiguously reveal the true reaction pathway
and demonstrate that all earlier mechanistic hypotheses are wrong: although a proline/isophorone
condensation product is formed, it is merely a spectator and not a key reaction intermediate in subsequent
heterogeneous hydrogenation. Enantioselectivity is the result of kinetic resolutionsa process that occurs
homogeneously in solution and not at the metal surface. Racemic 3,3,5-trimethylcyclohexanone (TMCH)
is produced by initial heterogeneous hydrogenation of isophorone; proline then reacts homogeneously,
preferentially with one enantiomer of TMCH, leaving an excess of the other. Thus in complete contrast to
the case of ketoester asymmetric hydrogenation, the metal surface is not involved in the crucial enantio-
differentiation step. The mechanism we propose also explains why the maximum attainable yield of
enantiopure TMCH cannot exceed 50%.
Introduction
is undoubtedly more to learn, our understanding of the
fundamentals of enantioselective CdO hydrogenation may be
The development of homogeneous chiral transition metal
catalysts opened up a major new field of chemistrysthe
synthesis of pure enantiomers from achiral precursors.1-3 The
academic and technical consequences of these advances have
transformed synthetic chemistry. By comparison, effective
heterogeneously catalyzed enantioselective reactions are rarities,
despite their huge potential importance which derives from the
major operational advantages offered by heterogeneous over
homogeneous catalysis. In addition to the fundamental chal-
lenges that it poses, this area is of foremost significance to the
pharmaceutical, fine chemicals, and advanced materials indus-
tries. Two systems have received intensive study: the enanti-
oselective hydrogenation of R- and â-ketoesters catalyzed by
modified platinum metals4-11 and by Ni.12-15 Although there
regarded as relatively well developedsa key point being that,
irrespective of details, the critical enantio-differentiation step
occurs at the surface of the metal catalyst.
In marked contrast, despite its potential importance in organic
synthesis, heterogeneously catalyzed asymmetric hydrogenation
of CdC bonds has received very little attention. Thus far,
essentially all the work in this area has been carried out by
Tungler and co-workers16,17 who focused on the metal-catalyzed,
proline-directed, enantioselective hydrogenation of isophorone
(1) to dihydroisophorone (2) (3,3,5-trimethylcyclohexanone,
hereafter TMCH), where relatively modest enantiomeric ex-
cesses (ee’s) have been achieved. They concluded that enanti-
oselectivity arises from the initial (homogeneous) formation of
a proline/isophorone condensation product which then adsorbs
on the metal surface where it undergoes heterogeneous asym-
metric (diastereoselective) hydrogenation. Hydrolysis of the
TMCH-proline hydrogenation product then delivers enantioen-
riched TMCH. However, there are significant apparent incon-
sistencies and gaps in their analysis of the proline/isophorone
system. Accordingly, we investigated this system in order to
(i) test earlier proposals and (ii) clarify key aspects of the
mechanism. We find that the proline/isophorone condensation
(1) Knowles, W. S. Angew. Chem., Int. Ed. 2002, 41, 1999-2007.
(2) Noyori, R. Angew. Chem., Int. Ed. 2002, 41, 2008-2022.
(3) Sharpless, K. B. Angew. Chem., Int. Ed. 2002, 41, 2024-2032.
(4) Blaser, H. U.; Baiker, A. In Handbook of Heterogeneous Catalysis; Ertl,
G., Kno¨tzinger, H., Weitkamp, J., Eds.; Wiley-VCH: New York, 1997;
Vol. 5, p 2422.
(5) Blaser, H. U.; Jalett, H. P.; Muller, M.; Studer, M. Catal. Today 1997, 37,
441-463.
(6) Baiker, A. J. Mol. Catal. A-Chem. 1997, 115, 473-493.
(7) Wells, P. B.; Wilkinson, A. G. Top. Catal. 1998, 5, 39-50.
(8) Jenkins, D. J.; Alabdulrahman, A. M. S.; Attard, G. A.; Griffin, K. G.;
Johnston, P.; Wells, P. B. J. Catal. 2005, 234, 230-239.
(9) Attard, G. A.; Ahmadi, A.; Jenkins, D. J.; Hazzazi, O. A.; Wells, P. B.;
Griffin, K. G.; Johnston, P.; Gillies, J. E. ChemPhysChem 2003, 4, 123-
130.
(14) Jones, T. E.; Baddeley, C. J. J. Mol. Catal. A-Chem. 2004, 216, 223-231.
(15) Lorenzo, M. O.; Baddeley, C. J.; Muryn, C.; Raval, R. Nature 2000, 404,
376-379.
(10) Bonello, J. M.; Williams, F. J.; Lambert, R. M. J. Am. Chem. Soc. 2003,
125, 2723-2729.
(11) Bonello, J. M.; Lambert, R. M.; Kunzle, N.; Baiker, A. J. Am. Chem. Soc.
2000, 122, 9864-9865.
(16) Tungler, A.; Ma´the´, T.; Petro´, J.; Tarnai, T. J. Mol. Catal. 1990, 61, 259-
267.
(12) Jones, T. E.; Baddeley, C. J. Surf. Sci. 2002, 519, 237-249.
(13) Jones, T. E.; Baddeley, C. J. Surf. Sci. 2002, 513, 453-467.
(17) Tungler, A.; Kajta´r, M.; Ma´the´, T.; To´th, G.; Fogassy, G.; Petro´, J. Catal.
Today 1989, 5, 159-171.
9
10.1021/ja061104y CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 7329-7334
7329

A R T I C L E S
McIntosh et al.
product is merely a spectator and not a key reaction intermediate.
Moreover, enantioselectivity is the result of kinetic resolutions
a process that occurs homogeneously in solution and not at the
metal surface. Racemic TMCH (()-2 is produced by initial
heterogeneous hydrogenation of isophorone (1), proline (3), then
reacts homogeneously, preferentially with one enantiomer of
TMCH, leaving an excess of the other. Thus in complete contrast
to the case of ketoester asymmetric hydrogenation, the metal
surface is not involved in the crucial enantio-differentiation step.
The mechanism we propose also explains why the maximum
attainable yield of enantiopure TMCH cannot exceed 50%.
Experimental Methods
Materials. L-Proline (99+%), D-proline (99+%), isophorone (97%),
and 3,3,5-trimethylcyclohexanone (TMCH) (98%) were supplied by
Aldrich. HPLC grade methanol was supplied by Fisher Scientific, and
the 10% Pd/C catalyst was purchased from Alfa Aesar.
Hydrogenation Procedure. Hydrogenation was carried out at
atmospheric pressure using standard Quickfit glassware with continuous
stirring of the reaction mixture. Proline + isophorone reactions were
carried out with 0.004 mol of proline and 0.004 mol of isophorone in
8 mL of methanol in the presence of 20 mg of Pd/C catalyst. Similarly,
proline + TMCH reactions were carried out with 0.004 mol of proline
and 0.004 mol of TMCH in 8 mL of methanol with 20 mg of Pd/C
catalyst. After hydrogenation, the catalyst was removed by filtration
using 0.45 µm grade filters, and the filtrate was combined with an equal
volume of dichloromethane. Next, the unreacted proline was extracted
by scavenging through a macroporous polystyrene sulfonic acid column
(Argonaut Technologies Inc.); the latter was then washed with a 1:1
mixture of methanol and dichloromethane to recover any remaining
isophorone and/or TMCH. The combined eluent, which contained all
the TMCH product along with unreacted isophorone, was then analyzed
by gas chromatography.
Figure 1. UV-vis spectra of 1:1 isophorone:L-proline reaction mixtures
as a function of elapsed time τ. Samples were being withdrawn at 24 h
intervals for examination by UV-vis spectroscopy.
from the hydrogenation of one of the condensation products
resulting from the homogeneous reaction of proline 3 with
isophorone 1.16,17 We therefore studied the time dependence of
the homogeneous interaction of a 1:1 mixture of L-proline (S)-3
and isophorone (1) in methanol solution at room temperature,
samples being withdrawn at 24 h intervals for examination by
UV-vis spectroscopy. The results are shown in Figure 1, where
τ is the time elapsed between starting the homogeneous reaction
and making the measurement.
It is apparent that the intensity of the absorbance at 235 nm,
which corresponds to the πfπ* transition of the R,â-unsaturated
carbonyl system of isophorone,18 decreased with increasing
reaction time. This was accompanied by a corresponding
progressive increase in the intensity of an absorbance at 277
nm that we ascribe to the condensation product formed by
reaction of L-proline (S)-3 and isophorone (1). Note that the
occurrence of an isosbestic point provides a clear indication
that only one major species is formed by this condensation
reaction. Analysis of the solution by means of liquid chroma-
tography-mass spectrometry (LC-MS) and electrospray ion-
ization mass spectrometry (ESI-MS) showed formation of a
molecule resulting from the condensation of one molecule of
isophorone (1) and one molecule of L-proline (S)-3 with
elimination of one molecule of water (ESI-MS m/z found: (M
+ H)+, 236.1642. C14H22NO2 requires M, 236.1651). Scheme
1 illustrates the various possible condensation products.
We assign the condensation product to the linear and/or cross-
conjugated dienamines 4 by comparison with the UV absorp-
tions for the dienamines formed between morpholine and
isophorone (λmax 288 nm) and between piperidine and isophor-
one (λmax 274 nm).19-21
Analysis. Reaction mixtures were analyzed using a Hewlett-Packard
5890 Series II gas chromatograph equipped with a â-cyclodextrin
capillary column (Chirasil-Dex CB, Varian, Inc.). Chromatograms were
acquired using the following temperature program: 90 °C for 10 min
then 5 °C/min to 110 °C. Enantiomeric excesses expressed in % terms
were calculated according to
[A] - [B]
[A] + [B]
ee )
× 100
where [A] is the concentration of one enantiomer and [B] that of the
other enantiomer. Absolute conversions were calculated using a decane
internal standard that was added to worked-up solutions prior to GC
analysis. Ultraviolet-visible (UV-vis) spectra were recorded with a
Cary 100 Bio UV-visible spectrometer using 50 µM initial concentra-
tion for both reactants.
Results and Discussion
In this relatively complex system, identifying the separate
roles of homogeneous and heterogeneous chemistry is of
paramount importance. Accordingly, we first studied the time
dependence of the homogeneous reaction between L-proline
(S)-3 and isophorone (1). Then we examined the heterogeneous
hydrogenation of the resulting reaction mixtures as a function
of elapsed homogeneous reaction time. Finally, we investigated
the homogeneous reaction of L-proline (S)-3 with the primary
hydrogenation product racemic TMCH (()-2 and then the
subsequent heterogeneous hydrogenation of the resulting con-
densation product.
(18) Flego, C.; Perego, C. Appl. Catal. A-Gen. 2000, 192, 317-329.
(19) Nozaki, H.; Yamaguti, T.; Ueda, S.; Kondoˆ, K. Tetrahedron 1968, 24,
1445-1453.
(20) We cannot rule out that the intermediate is the zwitterionic species
corresponding to 4 for the following reason: the UV spectrum of aniline
shows large shifts in the K (230 nm) and B bands (280 nm) compared
with benzene (K band, 203.5 nm; B band, 254 nm), whereas N-protonated
aniline shows no such shifts (K band, 203 nm, B band 254 nm).21 An
ammonium substituent therefore has little effect on the absorption
wavelength of the conjugated system to which it is attached; hence the
condensation product formed between proline and isophorone may be the
zwitterionic species corresponding to 4.
Homogeneous Reaction of Proline with Isophorone. Tun-
gler et al. proposed that enantioselectivity in this system arises
9
7330 J. AM. CHEM. SOC. VOL. 128, NO. 22, 2006

Heterogeneously Catalyzed Asymmetric CdC Hydrogenation
A R T I C L E S
Scheme 1. Possible Products from the Condensation of
Isophorone with Proline
Figure 2. The time dependence of hydrogen uptake during hydrogenation
of the various prereacted solutions.
Table 1. The τ Dependence of TMCH Enantiomeric Excess
Produced by Hydrogenation up to One Molar Equivalent H2 of
L-Proline (S)-3 and Isophorone 1 Reaction Mixtures
prereaction
time ( ), h
ee (%)
3%
τ
±
0
24
48
72
96
49
54
39
33
19
Heterogeneous Hydrogenation of the Reaction Mixture
as a Function of τ. A corresponding approach was used to
examine the τ dependence of the heterogeneously catalyzed
hydrogenation of the reaction mixtures resulting from the
homogeneous condensation of L-proline (S)-3 and isophorone
1; that is, mixtures of L-proline and isophorone in methanol
were stirred for a series of times τ (the same times as used to
acquire the UV-vis data in Figure 1), placed under 1 atm
hydrogen gas, and the Pd/C catalyst added. In each case, after
consumption of 1 molar equiv of hydrogen, the reaction was
stopped and the ee of the resulting TMCH 2 was measured by
chiral GC. The results are presented in Table 1, from which it
is apparent that the ee of the TMCH decreased with increasing
homogeneous condensation time (τ).
This is a key observation as it demonstrates that enantiose-
lective formation of TMCH does not involve heterogeneous
asymmetric hydrogenation of the condensation product 4 formed
between L-proline and isophoronesas was previously sug-
gested.16,17 In other words, if enantioselective formation of
TMCH 2 did involve heterogeneous hydrogenation of the
initially formed condensation product 4, the ee of the resulting
TMCH should either increase with τ, if the hydrogenation of
isophorone is competitive with the hydrogenation of the proline/
isophorone condensation product 4, or be independent of τ, if
there is no competitive hydrogenation of isophorone. Both of
these scenarios are in direct contradiction to observation. (As
we shall show later, the condensation product 4 does undergo
hydrogenation, but this is a side reaction and it does not
contribute to TMCH formation, racemic or otherwise.)
Figure 3. UV absorption spectra of hydrogenation reaction mixtures
corresponding to τ ) 72 h. Samples taken (a) before hydrogenation; (b)
halfway through regime A (see Figure 2); (c) at the turning point between
regime A and regime B (see Figure 2); and (d) after consumption of 1
molar equiv of hydrogen.
The time dependence of hydrogen uptake during hydrogena-
tion of the L-proline/isophorone reaction mixtures described
above is shown in Figure 2. The uptakes for τ g 24 h show
distinctive features: it is apparent that there are two dissimilar
regions, designated A and B, characterized by very different
reaction rates. The amount of hydrogen consumed in regime A
increases with τ and closely follows the increase in intensity of
the 277 nm absorbance (Figure 1). This strongly suggests that
regime A is associated with the relatively rapid hydrogenation
of the 277 nm species 4.
To test this hypothesis, a reaction mixture corresponding to
τ ) 72 h was hydrogenated, samples being taken for UV
spectroscopy (a) prior to hydrogenation, (b) halfway through
regime A, (c) at the turning point between regime A and regime
B, and (d) after consumption of 1 molar equiv of hydrogen.
The results are shown in Figure 3.
It is apparent that case (a) is equivalent to the τ ) 72 h
spectrum shown in Figure 1, as would be expected. The
spectrum taken in the middle of regime A, and especially that
taken at the A/B turning point, clearly demonstrates that during
this initial phase it is principally the 277 nm species 4 that is
hydrogenated. Control experiments carried out with (a) hydrogen
in the absence of catalyst and (b) with catalyst in the absence
(21) Jaffe´, H. H.; Orchin, M. Theory and Applications of UltraViolet Spectros-
copy; Wiley: New York, 1962.
9
J. AM. CHEM. SOC. VOL. 128, NO. 22, 2006 7331

A R T I C L E S
McIntosh et al.
of hydrogen gave a null result, confirming that disappearance
of the 277 nm species 4 was indeed due to its hydrogenation
and not to some other process. The ee of the TMCH produced
by hydrogenation was also measured at the A/B turning point:
this TMCH was racemic. The UV spectrum of the solution
obtained after consumption of 1 molar equiv hydrogen (spectrum
d, Figure 3) shows that, in contrast to regime A, substantial
hydrogenation of isophorone 1 occurs in regime B. Moreover,
even after the consumption of 1 molar equiv of hydrogen, a
significant amount of unreacted isophorone remained. Control
experiments showed that under our conditions neither L-proline
alone nor TMCH alone underwent catalytic hydrogenation. This
indicates that the 277 nm condensation product 4 can consume
more than 2 equiv of dihydrogen. Support for this proposal was
gained by the isolation and full characterization of (2S)-1-
[(1S,5R)-3,3,5-trimethylcyclohexyl]pyrrolidine-2-carboxylic acid
7 from the above reaction mixture, the result of hydrogenation
of 4 with 2 mol of hydrogen. The amino acid 7 was isolated as
Figure 4. Dependence of the ee of TMCH 2 on the ee of the proline
chiral auxiliary at the point corresponding to the consumption of 1 molar
equiv of hydrogen. The solid line corresponds to exact linearity: eeproduct
) eemax × eeaux, where eemax is equal to 49%.
1
to justify this radically new interpretation of the proline-directed
enantioselective heterogeneous hydrogenation of isophorone
according to which step II is the critical homogeneous process
that leads to enantio-differentiation by kinetic resolution of
racemic TMCH (()-2 formed by racemic heterogeneous
hydrogenation of isophorone 1 (step I). First we describe the
results of experiments deigned to address certain specific
mechanistic issues.
the major component (13:1 by H NMR) of an inseparable
mixture along with the diastereomer 7′ (for structural assignment
and full characterization, see Supporting Information).
To simplify the further discussion of this relatively complex
system, we now introduce a reaction scheme that accounts for
all our observationsssome yet to be described and discusseds
as shown below. The key features of this scheme, and indeed
The possible occurrence of nonlinear effects during the
catalytic hydrogenation of racemic TMCH (()-2 and proline 3
was investigated as such behavior has been observed in a
number of other catalytic asymmetric processes.22 Clearly, there
are no nonlinear effects, consistent with the mechanism proposed
in Scheme 2.
Scheme 2. Proposed Reaction Scheme for the Kinetic Resolution
of TMCH 2 by L-Proline (S)-3. Intermediate 8 May Be the
Regioisomeric Enamine, the Corresponding Zwitterionic Species,
or the Corresponding Iminium Ion
Figure 4 shows the dependence of the ee of the hydrogenation
product (TMCH 2) on the enantiomeric excess of the proline
chiral auxiliary (varied by mixing L- and D-proline) at the point
corresponding to the consumption of 1 molar equiv of hydrogen.
(The solid line in Figure 4 corresponds to exact linearity:
eeproduct ) eemax × eeaux/100, where eemax is equal to 49%.)
Reaction of Proline with the Primary Hydrogenation
Product (TMCH). Next we examined the proposed steps II
and III in Scheme 2, namely, the hydrogenation of mixtures of
L-proline (S)-3, (()-TMCH (()-2. A mixture of L-proline (S)-
3, (()-TMCH (()-2, and Pd/C was found to consume significant
amounts of hydrogen even at τ ∼ 0. Given the control
experiments which show that neither L-proline nor TMCH alone
undergo Pd-catalyzed hydrogenation, we may conclude that
L-proline and TMCH undergo a relatively fast reaction (step
II) resulting in a product (8) that undergoes subsequent
hydrogenation to yield 7 (step III). Indeed, analysis of hydro-
genated reaction mixtures revealed the presence of a species
corresponding to the fully hydrogenated condensation product
of TMCH and L-prolinesspecies 7 [m/z (ES) found: (M + H)+,
240.1971. C14H26NO2 requires M, 240.1964] identical to material
isolated from the τ ) 72 h reaction mixture above (for full
structural assignment of 7, see the Supporting Information).
Most importantly, chiral GC measurements showed that the
residual TMCH exhibited significant enantiomeric excess
[predominantly the (S)-enantiomer]; together, these observations
establish step II as the stereo-determining step.
the principal message of this paper, are readily summarized.
Steps I, III, and V are catalytic hydrogenations and necessarily
involve the metal surfaceshowever, they do not result in the
formation of enantioenriched TMCH. Step II is the stereo-
determining reactionshowever, it does not involve the metal
surface; it consists of competitive homogeneous reactions
involving L-proline (S)-3 and both enantiomers of TMCH 2.
Clearly, these two competing reactions must have intrinsically
different rate constants. Hydrogenation of enantiopure (S)-
TMCH, (S)-2, and (R)-proline (R)-3 did indeed reveal a much
faster hydrogen uptake than the comparable reaction of enan-
tiopure (S)-TMCH, (S)-2, and (S)-proline (S)-3. We shall proceed
(22) Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed. 1998, 37, 2923-2959.
9
7332 J. AM. CHEM. SOC. VOL. 128, NO. 22, 2006

Heterogeneously Catalyzed Asymmetric CdC Hydrogenation
A R T I C L E S
In summary, we have shown that in the proline-directed
asymmetric hydrogenation of isophorone the catalyst surface
does not participate in the key enantio-differentiating event,
which actually occurs homogeneously in the solution phase (step
II). The catalyst does, of course, enable the initial (racemic)
hydrogenation (step I, Scheme 2); it also plays an important
role in driving the kinetic resolution equilibrium step to
completion by irreversibly removing (R)-TMCH (R)-2 via 8 to
the amino acid 7 by hydrogenation (steps II and III, Scheme
2). Therefore, in the absence of side reactions, the best possible
overall performance of this system and systems analogous to it
corresponds to ∼100% ee in the product TMCH 2, but at a
maximum yield of ∼40%, relative to the starting material. Step
V (Scheme 2) formerly thought to be the critical step involved
in surface-mediated enantio-differentiation is actually an un-
desired side reactionsits effects are minimized by operating at
τ ) 0, that is, zero pre-equilibration of reactant and chiral
additive before commencing hydrogenation.24
Figure 5. Dependence of TMCH ee on TMCH conversion during the
catalytic hydrogenation of a mixture initially containing 0.004 mol TMCH
and 0.004 mol L-proline in 8 mL of methanol.
Accordingly, further experiments were conducted to confirm
beyond doubt that enantioselectivity in this system is the result
of the kinetic resolution of racemic TMCH (()-2 formed in
the initial racemic heterogeneous hydrogenation of isophorone.
Figure 5 shows the dependence of the enantiomeric excess of
TMCH 2 on TMCH conversion during the catalytic hydrogena-
tion of a mixture initially containing 0.004 mol TMCH and
0.004 mol L-proline (S)-3 in 8 mL of methanol. The data are
very strikingsthey show that as TMCH is consumed the ee of
the remaining (unreacted) TMCH rises rapidly, approaching
100% ee [of the (S)-enantiomer] at ∼60% conversion, thereafter
remaining constant until all the TMCH has reacted. This is a
clear indication that the (R)-enantiomer of TMCH reacts much
more rapidly with L-proline than does the (S)-enantiomer,
quickly leading to a situation in which effectively all the (R)-
enantiomer is consumed leaving behind essentially just the (S)-
enantiomer (∼100% ee), which reacts slowly with L-proline.
Equation 1 relates the efficiency of the kinetic resolution (the
S-factor) to the ee and the conversion.23
Interestingly, the same major diastereomer of the amino acid
7 is formed from the hydrogenation of isophorone 1 or racemic
TMCH (()-2 in the presence of L-proline (S)-3.25 The implica-
tion from this result is that the hydrogenation of 4sthe
condensation product of L-proline (S)-3 and isophorone 1s
occurs at the alkene double bond first and then, in a subsequent
step, at the enamine double bond.26
The highly diastereoselective reduction of 4 to yield 7,
coupled with the proposed reduction of the alkene double bond
in 4 in the presence of the enamine/iminium ion double bonds,
bodes well for the future design and discovery of a truly
heterogeneous metal-catalyzed hydrogenation of alkenes which
does not operate via a homogeneous kinetic resolution mech-
anism. However, to develop truly heterogeneous systems for
metal-catalyzed asymmetric hydrogenation, it will be necessary
to anchor the chiral agent to the metal surface more strongly
than any competing adsorbate. Interestingly, our results27 for
the separate adsorption from solution of isophorone and proline
onto single-crystal surfaces of Pt show that isophorone adsorbs
∼105 times faster than proline. This is entirely consistent with
what we find here: isophorone adsorption and hydrogenation
is so fast that proline itself never gets a chance to become
directly involved in a surface reaction step. Work is in progress
to develop a truly heterogeneous catalytic asymmetric hydro-
genation of olefins.
kFAST
ln[(1 - c)(1 - ee)]
ln[(1 - c)(1 + ee)]
S )
where S ) kREL
)
(1)
kSLOW
The solid line in Figure 5 shows the fit of eq 1 to the data,
from which we calculate S ) 17.8 for the ratio of the overall
reaction rate constants for removal of the (R)- and (S)-isomers
of TMCH by reaction with L-proline.
A control experiment was performed to establish that kinetic
resolution depends on equilibrium II being followed by step
III. First, (()-TMCH and L-proline were mixed and allowed
to react under hydrogen but without catalyst. After 30 min at
room temperature, the ee of the remaining TMCH reached 7.7%
[excess of (S)-2] and remained constant thereafter. This observa-
tion is fully consistent with equilibrium II being followed by
step III. The (R)-enantiomer in racemic TMCH (()-2 reacts
(homogeneously) somewhat preferentially with L-proline to form
condensation product 8, hence yielding a slight excess of the
(S)-enantiomer. However, in the presence of both hydrogen and
catalyst, the irreversible hydrogenation step III drives the
equilibrium between L-proline and (R)-TMCH (R)-2 (step II)
toward complex 8, thus enabling high conversion and much
larger ee.
(24) Steps IV and V consume one molecule of L-proline and one molecule of
isophorone, so the only limit to a high ee of (S)-TMCH is that the reaction
is routinely run until only 1 equiv of H2 is consumedsstep V just reduces
the amount of (S)-TMCH which can be produced, not the ee of the TMCH
which is controlled by hydrogen availability, that is, conversion.
(25) Purification of the hydrogenation mixture from a τ ) 72 h experiment
which was halted at the turning point between regimes A and B gave a
13:1 mixture of 7:7′.
(26) If the hydrogenation occurred in one surface-bound event, the two
substituents on the cyclohexane in 7 would be cis and not trans to one
another. The trans stereochemistry of the cyclohexane substituents in 7 is
readily accounted for; hydrogenation of 8 will likely occur from the least
hindered face of the enamine (or imine) double bond from the lowest energy
conformer in which two of the three methyl groups are pseudoequatorial.
(23) Kagan, H. B.; Fiaud, J. C. Top. Stereochem. 1988, 18, 249.
(27) McIntosh, A. I.; Watson, D. J.; Lambert, R. M. Manuscript in preparation.
9
J. AM. CHEM. SOC. VOL. 128, NO. 22, 2006 7333

A R T I C L E S
McIntosh et al.
Acknowledgment. A.I.M. acknowledges financial support by
the UK Engineering and Physical Sciences Research Council.
J.W.B. gratefully acknowledges the Royal Society UK for the
award of a University Research Fellowship.
Supporting Information Available: Full structural assign-
ments and characterization. This material is available free of
charge via the Internet at http://pubs.acs.org.
JA061104Y
9
7334 J. AM. CHEM. SOC. VOL. 128, NO. 22, 2006