Asymmetric hydrogenation of enamides with Rh-BisP* and Rh-MiniPHOs catalysts. Scope, limitations, and mechanism

Article abstract of DOI:10.1021/ja010161i

The asymmetric hydrogenation of aryl- and alkyl-substituted enamides catalyzed by Rh-BisP* complex affords optically active amides with very high ee values. The Rh-MiniPHOS catalyst gives somewhat less satisfactory results. Hydrogenation of the aryl-subst

Full text of DOI:10.1021/ja010161i

5268
J. Am. Chem. Soc. 2001, 123, 5268-5276
Asymmetric Hydrogenation of Enamides with Rh-BisP* and
Rh-MiniPHOS Catalysts. Scope, Limitations, and Mechanism
Ilya D. Gridnev,* Masaya Yasutake, Natsuka Higashi, and Tsuneo Imamoto*
Contribution from the Department of Chemistry, Faculty of Science, Chiba UniVersity,
Inage-ku, Chiba 263-8522, Japan
ReceiVed January 18, 2001
Abstract: The asymmetric hydrogenation of aryl- and alkyl-substituted enamides catalyzed by Rh-BisP* complex
affords optically active amides with very high ee values. The Rh-MiniPHOS catalyst gives somewhat less
satisfactory results. Hydrogenation of the aryl-substituted enamides with (S,S)-BisP*-Rh catalyst gives R-amides,
whereas the t-Bu- and 1-adamantyl-substituted enamides give S-products with 99% ee. Reaction of [Rh(BisP*)-
(CD3OD)2]BF4 (11) with CH2dC(C6H5)NHCOCH3 (5) gives two diastereomers of the catalyst-substrate
complex (12a,b), which interconvert reversibly by both intra- and intermolecular pathways as shown by EXSY
data. Only one isomer in equilibrium with solvate complex 11 was detected for each of the catalyst-substrate
complexes 17 and 18 obtained from CH2dC(t-Bu)NHCOCH3 (6) or CH2dC(1-adamantyl)NHCOCH3 (7).
Hydrogenation of these equilibrium mixtures at -100 °C gave monohydride intermediates 19 and 20,
respectively. In these monohydrides the Rh atom is bound to the â-carbon. A new effect of the significant
decrease of ee was found for the asymmetric hydrogenation of CH2dC(C6H4OCH3-o)NHCOCH3 (21), when
H2 was substituted for HD or D2.
Introduction
we were prompted to examine the synthetic utility of these
complexes in the asymmetric hydrogenation of enamides.
Furthermore, a mechanistic study of asymmetric hydrogenation
of methyl (Z)-R-acetamidocinnamate catalyzed by BisP*-Rh
complex revealed some new aspects of the reaction pathway.17
Most of the previous mechanistic studies in the field of Rh-
catalyzed asymmetric hydrogenation have been done for dehy-
droamino acids.18,19 In the usual explanations of the stereose-
lection mechanism in the catalytic asymmetric hydrogenations
the carboxy group of a dehydroamino acid is regarded as an
important stereoregulating factor.18,20-23 In the structure of an
enamide the carboxy group is absent, but the nature of the
substituent at the R-position can be varied without losing the
high degree of enantioselectivity. Therefore, a mechanistic study
of catalytic asymmetric hydrogenation of enamides can provide
new details of the structure of the intermediates and the
mechanism of stereoselection in the Rh-catalyzed hydrogena-
tions.
Optically active 1-arylalkylamines are useful compounds from
many points of view.1 The possibility of producing chiral amides
directly by asymmetric hydrogenation of enamides was recog-
nized in 1975 by Kagan,2 who achieved 90% optical yields in
some hydrogenations catalyzed by Rh(I)-DIOP.2,3 For a long
time comparable results could not be obtained with other
diphosphine ligands.4 Note, however, that a successful applica-
tion of Ru-catalyzed asymmetric hydrogenation for the synthesis
of isoquinoline alkaloids was reported by Noyori et al.5 The
breakthrough was achieved with the introduction of BPE and
DuPHOS ligands, which gave up to 98.5% enantioselectivity
in Rh-catalyzed asymmetric hydrogenations of enamides.6 Later,
rhodium complexes of such ligands as BICP,7 H8-BINAM,8
TRAP,9 Binaphane,10 and modified DIOP11,12 were also found
to be effective for similar hydrogenation. It was shown recently
in our group that new structurally simple P-chirogenic diphos-
phines abbreviated as BisP*13,14 and MiniPHOS14-16 give Rh
complexes, which are effective catalysts for the asymmetric
hydrogenation of dehydroamino acids providing a wide series
of unnatural amino acids with excellent ee values. Therefore,
(1) Nogradi, M. StereoselectiVe Synthesis, 2nd ed.; VCH: Weinheim,
1995.
(2) Kagan, H. B.; Langlois, N.; Dang, T. P. J. Organomet. Chem. 1975,
90, 353.
(3) Sinou, D.; Kagan, H. B. J. Organomet. Chem. 1976, 114, 325-337.
(4) Morimoto, T.; Chiba, M.; Achiwa, K. Chem. Pharm. Bull. 1992, 40,
2894-2896.
In this paper we report detailed studies of the scope and
limitations of the asymmetric hydrogenation of enamides
(5) Noyori, R.; Ohta, M.; Hisao, Y.; Kitamura, M.; Ohta, T.; Takaya,
H. J. Am. Chem. Soc. 1986, 108, 7117-7119.
(6) Burk, M. J.; Wang, Y. M.; Lee, J. R. J. Am. Chem. Soc. 1996, 118,
5142-5143.
(13) Imamoto, T.; Watanabe, J.; Wada, Y.; Masuda, H.; Yamada, H.;
Tsuruta, H.; Matsukawa, S.; Yamaguchi, K. J. Am. Chem. Soc. 1998, 120,
1635-1636.
(14) Gridnev, I. D.; Yamanoi Y.; Higashi, N.; Tsuruta, H.; Yasutake,
M.; Imamoto, T. AdV. Synth., Catal. 2001, 343, 118-136.
(15) Yamanoi, Y.; Imamoto, T. J. Org. Chem. 1999, 64, 2988-2989.
(16) Gridnev I. D.; Imamoto, T. Organometallics 2001, 20, 545-549.
(17) Gridnev, I. D.; Higashi, N.; Asakura, K.; Imamoto, T. J. Am. Chem.
Soc. 2000, 122, 7183-7194.
(7) Zhu, G.; Zhang, X. J. Org. Chem. 1998, 63, 9590-9593.
(8) Zhang, F.-Y.; Pai, C.-C.; Chan, A. S. C. J. Am. Chem. Soc. 1998,
120, 5808-5809.
(9) Kuwano, R.; Sato, K.; Kurokawa, T.; Karube, D.; Ito, Y. J. Am. Chem.
Soc. 2000, 122, 7614-7615.
(10) Xiao, D.; Zhang, Z.; Zhang, X. Org. Lett. 1999, 1, 1679-1681.
(11) Li, W.; Zhang, X. J. Org. Chem. 2000, 65, 5871-5874.
(12) Yan, Y.; RajanBabu, T. V. Org. Lett. 2000, 4137-4140.
(18) Brown, J. M. Hydrogenation of Functionalized Carbon-Carbon
Double Bonds; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer:
Berlin, 1999; Vol. 1, pp 119-182.
10.1021/ja010161i CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/04/2001

Asymmetric Hydrogenation of Enamides
J. Am. Chem. Soc., Vol. 123, No. 22, 2001 5269
Table 1. Enantioselective Hydrogenation of Aryl- and
Scheme 1
Alkylenamides Catalyzed by t-Bu-BisP*and t-Bu-MiniPHOSa
entry
Ar
R1, R2
precatalyst
ee, %
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
C6H5
C6H5
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
Me, Me
Me, Me
H, H
H, H
H, H
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
99 (R)
66 (R)
50 (R)
47 (R)
97 (R)
87 (R)
96 (R)
86 (R)
46 (R)
16 (R)
93 (R)
71 (R)
99 (R)
88 (R)
99 (R)
97 (S)
99 (R)
98 (R)
99 (S)
99 (S)
99 (S)
2-CH3OC6H4
2-CH3OC6H4
3-CH3OC6H4
3-CH3OC6H4
4-CH3OC6H4
4-CH3OC6H4
2-ClC6H4
2-ClC6H4
3-ClC6H4
3-ClC6H4
4-ClC6H4
Scheme 2
4-ClC6H4
3-CH3OCOC6H4
3-CH3OCOC6H4
C6H5
C6H5
t-Bu
1-Adamantyl
1-Adamantyl
1* b
1
2
1
1
2
compounds. Thus, the successful asymmetric hydrogenation of
3-methoxy-substituted substrate 3 (entries 15 and 16) enables
the effective introduction of an asymmetric center in the
synthesis of the acetylcholinesterase inhibitor SDZ-ENA-713
(4) (Scheme 1).25-29
a Reactions were carried out in MeOH at room temperature under
an initial H2 pressure of 3 atm and S/C ) 100 for 24-36 h. The yields
were quantitative. b 1* is [Rh((R,R)-t-Bu-BisP*)]BF4.
Particularly interesting is the dramatic difference in the sense
of stereoselection observed for all the studied aryl-substiuted
enamides (giving R-products) and the enamides containing bulky
tert-butyl and 1-adamantyl groups (giving S-products with
excellent ee values, entries 19-21). Burk et al. have reported a
similar effect observed in asymmetric hydrogenation of 1-acety-
lamino-1-phenylethene (5), 1-acetylamino-1-tert-butylethene (6),
and 1-acetylamino-1-(1-adamantyl)ethene (7) catalyzed by a
DuPHOS-Rh complex.6,30 We could not employ the previously
reported sophisticated chromatographic conditions for the
separation of the enantiomers of the adamantyl-substituted
hydrogenation product (8).30 The separation was achieved on
HPLC using the UV detector adjusted for 210 nm, but the
absolute configuration of our product could not be assigned at
this stage, since the [R]D values have not been previously
reported. The exact structure of 8, obtained by the hydrogenation
of 7 in the presence of 1, was determined by converting it to
thioamide 9, the S absolute configuration of which was
determined by the Flack parameter obtained in the single-crystal
X-ray study (Scheme 2).
catalyzed by the rhodium complexes of t-Bu-BisP* (1) and t-Bu-
MiniPHOS (2) as well as the mechanistic studies carried out
for a wide range of substrates.24
Results and Discussion
Catalytic Hydrogenation of Enamides with Rh-BisP* and
Rh-MiniPHOS Catalysts. The results of the catalytic asym-
metric hydrogenation of enamides catalyzed by rhodium com-
plexes of t-Bu-BisP* and t-Bu-MiniPHOS are summarized in
Table 1. It should be noted that enantioselectivities approaching
100% have been achieved for a wide range of substrates (entries
1, 5, 7, 11, 13, 15, and 17-21). Almost no solvent dependence
of the stereochemical purity was found (Table S1), and hence
methanol was used throughout the further studies.
The BisP*-based catalyst 1 is optimal; it gave better results
compared to 2 in all the studied cases. The m- and p-methoxy-
and chloro-substituted phenylenamides gave high ee values
(entries 5, 7, 11, and 13), but for their ortho-substituted
analogues the optical yields were drastically decreased (entries
3, 4, 9, and 10). This result suggests that the absence of steric
hindrance around the double bond of aromatic enamide is an
important factor for the successful asymmetric hydrogenation.
Entry 16 demonstrates the applicability of the described
method to the asymmetric synthesis of practically useful
Thus, our data unequivocally indicate that the sense of
enantioselection in the asymmetric hydrogenation of enamides
stands in close correlation to the stereochemical environment
of the double bond of the enamide, leading in extreme cases to
the complete inversion of the sense of enantioselection. To
(19) Landis, C. R.; Hilfenhaus, P.; Feldgus, S. J. Am. Chem. Soc. 1999,
121, 8741-8754.
(25) Brufani, M.; Filocamo, L.; Lappa, S.; Maggi, A. Drugs Future 1997,
22, 397-410.
(26) Cutler, N. R.; Polinsky, R. J.; Sramek, J. J.; Rnz, A.; Jhee, S. S.;
Mancione, L. Acta Neurol. Scand. 1998, 97, 244-250.
(27) Enz, A.; Boddeke, A. H.; Gmelin, G.; Malanowski, J. Prog. Brain
Res. 1993, 98, 431-438.
(28) Ohara, T.; Tanaka, K.; Fukaya, H.; Demura, N.; Iimura, A.; Seno,
N. BehaV. Brain Res. 1997, 83, 229-233.
(29) Tse, F. L. S.; Laplanche, R. Pharm. Res. 1998, 15, 1614-1620.
(30) Burk, M. J.; Casy, G.; Johnson, N. B. J. Org. Chem. 1998, 63,
6084-6085.
(20) Vineyard, B. D.; Knowles, W. S.; Sabacky, G. L.; Bachman, G. L.;
Weinkauff, D. J. J. Am. Chem. Soc. 1977, 99, 5946-5952.
(21) Koenig, K. E.; Sabacky, M. J.; Bachman, G. L.; Christopfel, W.
C.; Barnstorff, H. D.; Friedman, R. B.; Knowles, W. S.; Stults, B. R.;
Vineyard, B. D.; Weinkauff, D. J. Ann. N.Y. Acad. Sci. 1980, 333, 16-22.
(22) Knowles, W. S. Acc. Chem. Res. 1983, 16, 106-112.
(23) Brown, J. M. Chem. Soc. ReV. 1993, 22, 25-41.
(24) Preliminary communication: Gridnev, I. D.; Higashi, N.; Imamoto,
T. J. Am. Chem. Soc. 2000, 122, 10486-10487.

5270 J. Am. Chem. Soc., Vol. 123, No. 22, 2001
GridneV et al.
Scheme 3
R-position (Scheme 3), in accordance with the previous results
of similar experiments carried out with dehydroamino ac-
ids.14,17,31,32 On the other hand, the reaction of t-Bu-substituted
enamide 6 with HD gave a significant predominance of the
â-deuterated product, whereas in the case of adamantyl-
substituted enamide 7 almost equal quantities of R- and
â-deuterated compounds were obtained (Scheme 4). Thus, the
isotope partitioning in asymmetric hydrogenation with HD
evidently correlated with the difference in the stereochemical
outcome of the hydrogenation. We considered that the different
reaction pathways may be responsible for these effects and
studied the structure of the intermediates in the catalytic cycles
of both aromatic and aliphatic enamides.
An additional interesting effect was noticed in these experi-
ments. In most of the studied cases the enantiomeric excesses
of the hydrogenation products from the experiments using HD
and D2 reproduced well the ee values obtained with dihydrogen.
However, a significant decrease of the optical yield was
observed in the case of o-methoxyphenyl-substituted enamide.
Thus, the ee in the hydrogenation of this substrate with H2 was
50% (Table 1, entry 3). In the hydrogenation with HD under
the same conditions the ee was drastically decreased (24% ee
in two independent runs). Use of D2 led again to a drastic
decrease of ee: 12% and 5% ee were obtained in two
independent experiments. To our knowledge, this is the first
observation of the significant dependence of the optical yield
on the isotopic composition of the hydrogenation reagent. Most
likely either kinetic or equilibrium isotope effects, or a combina-
tion of both, are responsible for such drastic change in the ee.
Scheme 4
obtain further insight into the origin of this phenomenon, we
undertook a mechanistic investigation of the asymmetric
hydrogenation of enamides catalyzed by the Rh-BisP* catalyst.
Asymmetric Hydrogenation of Enamides Using D2 and
HD. The preceding research showed that the nature of the
isotope partitioning in the products and intermediates of the
catalytic asymmetric hydrogenation using deuterium hydride
may provide valuable information on the reaction pathways and
the structure of the intermediates.14,17,31,32 On the other hand,
by studying the catalytic deuteration the reversibility of the
catalytic reaction can be checked as well as the possible
intervention of the solvent in the catalytic cycle of the
asymmetric hydrogenation.14,31,32
We studied the catalytic hydrogenations of a series of
enamides using D2 and HD. The complete absence of isotopic
scrambling was found in all deuteration experiments; only the
dideuteration products were observed (Schemes 3 and 4; errors
in the isotopomeric excess determination did not exceed (0.05).
Usually, the same results were obtained when the deuteration
was carried either in methanol or THF. Only in the case of
o-methoxyphenyl-substituted enamide was a 10% admixture of
the monodeuterated product (deuterium in the â-position)
obtained in methanol. This result may be explained by the
relatively slow reductive elimination resulting in the partial
isotope exchange in the long-lived monohydride intermediate
(see below). Thus, the deuteration experiments indicate that,
once the hydrogen (or deuterium) atom is transferred in the
migratory insertion step, the reaction does not reverse to free
H2.31,32
Intermediates in the Asymmetric Hydrogenation of Phe-
nylenamide (5) and p-Clorophenylenamide (10). Addition of
a 2-fold excess of enamides 5 or 10 to a deuteriomethanol
solution of solvate complex 1117 at -20 °C resulted in
immediate formation of two diastereomers of catalyst-substrate
complexes 12 or 13 in a ratio changing from 4:1 at -90 °C to
2:1 at 0 °C (12, ∆H ) -1.3 kcal mol-1, ∆S ) -3.0 cal mol-1
K-1) or from 13:1 at -90 °C to 3:1 at 0 °C (13, ∆H ) -1.5
kcal mol-1, ∆S ) -3.2 cal mol-1 K-1) (Scheme 5). In both
cases, no detectable amounts of the solvate complex 11 could
be found in the spectra in the temperature range from -90 to
+30 °C, and, therefore, the tightness of binding for these
substrates (5 and 10) is comparable to that in the catalyst-
substrate complex of 11 and methyl (Z)-R-acetaminocin-
namate.17 The spectral data for the catalyst-substrate complexes
12 and 13 are given in Tables S2-S4. All spectral properties
of 12 and 13, including the temperature dependence of the NMR
spectra, are essentially similar; in the following section we
describe in detail the spectral characteristics of 12, making
reference to the spectra of 13 only when necessary.
The configurations of major and minor isomers 12a and 12b
were elucidated from the NMR data. A portion of the 2D 1H-
31P correlation spectrum is shown in Figure 1. It reveals a
remarkable difference in the chemical shifts of the alkyl groups
of 12a and 12b. In the major isomer the difference in chemical
shifts of two unequal t-Bu groups is 0.60 ppm, whereas the
methyl groups are separated by only 0.18 ppm. The reverse
situation is evident in the spectrum of 12b: only a small
difference (0.10 ppm) is observed for the signals of the t-Bu
groups, but ∆δ(Me) is as great as 1.05 ppm. Inspection of the
molecular models of 12a and 12b shows that the observed
differentiation in the chemical shifts can be explained by the
shielding effect of the phenyl ring: In 12a one of the t-Bu
groups is shielded, and its chemical shift is, therefore, high-
field shifted. In 12b the same effect was observed for one of
In the hydrogenation using HD an unequal distribution of
deuterium between the R- and â-positions of the hydrogenation
product was observed (Schemes 3 and 4). All aromatic enamides
give predominantly the isomer containing deuterium at the
(31) Brown, J. M.; Parker, D. Organometallics 1982, 1, 950-956.
(32) Brauch, T. W.; Landis, C. R. Inorg. Chim. Acta 1998, 270, 285-
297.

Asymmetric Hydrogenation of Enamides
J. Am. Chem. Soc., Vol. 123, No. 22, 2001 5271
Scheme 5
Scheme 6
unexpected and demonstrates that the phenyl ring can minimize
its steric demands by acquiring favorable conformation.
The structures of the major and minor isomers of 13 (see
Tables S2-S4) were assigned on the basis of the same reasons;
the major isomer 13a has the structure similar to 12a: the p-Cl-
phenyl substituent in 13a is close to one of the t-Bu groups, as
follows from the analysis of the relative chemical shifts.
The 31P NMR spectra of 12 are temperature-dependent. At
temperatures below -40 °C the line shape of the 31P NMR
spectrum of 12 is unaffected by the dynamic effects, whereas
at higher temperatures the reversible broadening of the spectral
lines attests to the reversible interconversion of the diastereomers
12a and 12b (Figure S1). The mechanism of this interconversion
was elucidated from the EXSY spectra. Figure 2 displays the
phase-sensitive 2D 31P-31P EXSY spectra of the equilibrium
mixture of 12a and 12b at two different temperatures. In the
spectrum taken at -40 °C (Figure 2, left) only two cross-peaks
corresponding to the pairwise exchange of P1 and P2 in two
isomers are observed. This attests to the intramolecular exchange
of 12a and 12b without complete dissociation of the substrate.
On the other hand, in the EXSY spectrum taken at -25 °C, all
possible cross-peaks between four signals are observed, although
their intensity is different (Figure 2, right). The two most
intensive cross-peaks are the same as those observed at -40
°C, and, therefore, the intramolecular exchange of 12a and 12b
is still the fastest process at -25 °C. Four much less intensive
cross-peaks indicate the occurrence of intermolecular exchange
via complete dissociation of either 12a or 12b producing a free
substrate and solvate complex 11. In the 1H-1H EXSY spectrum
taken at -30 °C a set of exchange cross-peaks between free
enamide and coordinated enamide was observed, confirming
the dissociation mechanism. Thus, two diastereomers 12a and
12b interconvert reversibly via intra- and intermolecular path-
ways. The interconversion via the intramolecular mechanism
is notably faster at -25 °C. Similar equilibria were previously
observed for various catalyst-substrate complexes of dehy-
droamino acids.14,17,33-35
It should be noted that the rate of intermolecular intercon-
version of 12a and 12b is significantly faster compared to that
in the catalyst-substrate complexes of dehydroamino acids. The
quantification of the EXSY data gives for the intermolecular
interconversion of the diastereomers of 12 free activation energy
∆G248 ) 14 kcal mol-1, which is approximately 5 kcal mol-1
lower compared to that for the complex of 11 and methyl (Z)-
R-acetaminocinnamate.17
We attempted to detect a monohydride intermediate in the
asymmetric hydrogenation of 5 catalyzed by 11. However, either
the addition of 5 to a deuteriomethanol solution of a dihydride
complex 1417 at -100 °C or hydrogenation of 12 at -100 °C
produced directly a catalyst-product complex 15 in equilibrium
with solvate 11 (Scheme 5). The same result (producing complex
16) was obtained when the catalyst-substrate complex 13 was
hydrogenated at -100 °C. The optical yields of the hydrogena-
tion products obtained in the NMR experiments were always
the same as those in the catalytic hydrogenation.
Figure 1. 1H-31P HMBC spectrum of 12 (400 MHz, -20 °C, CD3-
OD).
the methyl groups. We conclude, therefore, that the major isomer
of 12 has the structure 12a (Scheme 6). This is somewhat

5272 J. Am. Chem. Soc., Vol. 123, No. 22, 2001
GridneV et al.
Figure 2. 31P-31P EXSY spectra of 12 (162 MHz, CD3OD): (left) at 223 K, mixing time 1 s; (right) at -25 °C, mixing time 0.1 s.
Scheme 7
Figure 3. DFT-optimized structures (B3LYP/LANL2DZ) of 17′a and
17′b.
in 17 and the noncoordinated one in 6 is only 27 (for the CH2
group) and 13 (for Ctert) ppm (27 and 9 ppm, respectively, in
the case of 18); the corresponding differences are 44 and 48
ppm in 12. These values suggest also that, in contrast to 12
and 13, the double bond is weakly coordinated in 17 and 18.
This conclusion seems to be reasonable in view of the bulkiness
of the t-Bu and 1-adamantyl substituents.
In the phase-sensitive 31P-31P EXSY spectra of 17, the
exchange cross-peaks were observed between the doublet of
Intermediates in the Asymmetric Hydrogenation of tert-
Butylenamide (6) and 1-Adamantylenamide (7). Addition of
a 2-fold excess of t-Bu-substituted enamide 6 or adamantyl-
substituted enamide 7 to the deuteriomethanol solution of solvate
complex 11 resulted in formation of the catalyst-substrate
complexes 17 and 18, respectively, which differ significantly
from 12 and 13 in structure and stability (Scheme 7).
The temperature dependencies of the 31P NMR spectra of 17
and 18 (e.g. Figure S2) show that significant amounts of 11
equilibrate with the excess of either 17 or 18 even at -90 °C.
The thermodynamic parameters of these are ∆H ) -2.1 kcal
mol-1, ∆S ) -3.9 cal mol-1 K-1 (17) and ∆H ) -2.4 kcal
mol-1, ∆S ) -3.3 cal mol-1 K-1 (18). In both cases only one
isomer of the corresponding catalyst-substrate complex was
observed in the temperature range from -90 to +30 °C; neither
the line shape of the 1D NMR spectra nor 2D EXSY experi-
ments carried out at various temperatures showed the presence
of the second possible diastereomer. The relatively weak
coordination of enamides 6 and 7 in the catalyst-substrate
complexes 17 and 18 manifests itself also in the chemical shifts
of the coordinated double bonds in their 13C NMR spectra. Thus,
the difference in chemical shifts of the coordinated double bond
11 and each of the multiplets of 17, as well as between the
multiplets of 17. Very similar results were observed in the EXSY
spectra of the catalyst-substrate complex 18.
We failed to determine experimentally the configurations of
1
17 and 18: all alkyl groups resonate very closely in the H
NMR spectra of these catalyst-substrate complexes, retarding
the chemical shift assignment. We therefore decided to resolve
this problem computationally. DFT calculations of simplified
molecules 17′a and 17′b gave virtually equal energies for these
isomers: the energy difference is less than 0.5 kcal mol-1
(Figure 3). Nevertheless, the computed structures of 17′a and
17′b (Figure 3) give some indication of the possible reasons
for the different stability of the real compounds. Analyzing the
structures of 17′a and 17′b, one can see two competing effects
influencing the stability of a certain isomer. The terminal CH2d
group in 17′a is disposed almost in the plane of the chelate
cycle of the catalyst, thus avoiding close contact with the bulky
t-Bu substituent. On the other hand, in 17′b this close contact
is avoided at the expense of placing the methyl substituent of
the substrate in close proximity to the t-Bu group from the
catalyst (Figure 3). It is clear that the exact nature of the
substituent in the R-position of enamide would determine the

Asymmetric Hydrogenation of Enamides
J. Am. Chem. Soc., Vol. 123, No. 22, 2001 5273
1
Table 2. Parameters of the H and 31P NMR Spectra of the
Monohydride Intermediates 19a,b, 20a,b, and 23a-c
compd
δH (1JRhH, 2JPH
)
δ31P1 (1JRhP
)
δ31P2 (1JRhP
)
19a
19b
20a
20b
23a
23b
23c
-21.2 (26, 17, 31)
-23.1 (32, 15, 32)
-21.3 (27, 16, 27)
-23.5a
-24.7 (39, 17, 28)
-21.5 (28, 14, 28)
-23.2 (32, 16, 32)
53.2 (86)
45.6 (82)
53.4 (86)
45.7 (86)
54.9 (81)
53.4 (89)
45.8 (83)
86.7 (162)
77.6 (152)
86.7 (160)
77.6 (150)
71.2 (133)
87.8 (160)
78.7a
a Couplings were not determined due to the very low concentration
or overlapping with other signals.
relative stabilities of the diastereomers. Our experimental data
show that in the case of flat aryl substituents, the steric repulsion
from the t-Bu group can be minimized by acquiring an
appropriate conformation, and the major diastereomers 12a and
13a have structures similar to 17′b. Apparently, in the case of
t-Bu- and 1-adamantyl-substituted substrates the steric repulsion
from the bulky alkyl groups becomes more important than that
from the terminal CH2 group, and only 17a and 18a are observed
experimentally in solution. Recently, very similar computational
results for a complex of the Rh-DuPHOS catalyst and a model
substrate were reported.34
Hydrogenation of the equilibrium mixture of 11, 6, and 17
(or 11, 7, and 18) at -100 °C for 7 min followed by immediate
placement of the sample in the probe of an NMR spectrometer
precooled to -100 °C led to the observation of the monohydride
intermediates 19 and 20 (Scheme 7). In both cases two isomers
were observed in a 10:1 ratio. Monohydrides 19 and 20 are
relatively stable below -85 °C; at higher temperatures they
decompose rapidly, affording 11 and the corresponding hydro-
genation products. The optical yields and the sense of stereo-
selection of the hydrogenation products obtained in the NMR
experiments were always the same as in the catalytic hydroge-
nations (99% ee S). The spectral characteristics of 19 and 20
are very similar (Table 2).
In the 1H NMR spectrum of 19 the hydride proton resonates
at δ ) -21.2 as an 8-line multiplet (coupling with Rh is 28
Hz; that with phosphorus is 16 and 28 Hz). These values are
close to the NMR spectra of the previously characterized
monohydride intermediates in the asymmetric hydrogenation of
dehydroamino acids.14,17,35,36 However, the 31P NMR spectrum
displays significant difference. Thus, in all monohydride
intermediates characterized previously, both phosphorus atoms
resonate quite near, within the 15 ppm interval. On the other
hand, in 19a the difference in chemical shifts of two phosphorus
atoms is more than 30 ppm (δ(P1) ) 51.9, δ(P2) ) 85.3) (Figure
4a). The chemical shift of the low-field signal is very close to
that in the solvate complex 11 (δ ) 89.8). Therefore, the larger
value of δ(P2) in 19a suggests that the coordination site disposed
trans to this phosphorus atom is occupied by a weakly bound
molecule of the solvent, and the carbonyl group is, accordingly,
not coordinated. The same conclusion follows from the analysis
of the carbonyl region of the 13C NMR spectrum of 19a. Thus,
the signal of the carbonyl carbon of 19a (δ ) 176.3) is only
slightly downfield-shifted respective to the carbonyl signals of
the starting enamide 6 (δ ) 173.1) and the hydrogenation
Figure 4. 31P NMR spectra (162 MHz, -90 °C, CD3OD) of the
reaction mixtures containing monohydride intermediates: (a) 19 and
(b) 19*.
Scheme 8
product (δ ) 172.5). Note also that the signal of the coordinated
carbonyl of the catalyst-substrate complex 17 resonates at δ
) 185.0 and is split in a doublet with 7 Hz coupling. Additional
information on the structure of the monohydride intermediate
19a gives a broadened signal at 61.7 ppm in the 13C NMR
spectrum. This value of chemical shift corresponds to the carbon
atom attached to the nitrogen. In the gated-decoupled experiment
this signal becomes a doublet, and, therefore, it belongs to a
methyne carbon. These data lead to the conclusion that the
monohydride intermediate of the asymmetric hydrogenation of
enamide 6 has the structure in which the amide is bound to Rh
by the CH2 group, and the R-position is hydrogenated.
It was very difficult to locate the signal of the CH2Rh carbon
in the 13C NMR spectrum due to its relatively low intensity
and partial overlap with other much more intensive signals. To
obtain unequivocal confirmation of the structure of monohy-
drides 19 and 20, we prepared â-labeled compounds 6* and 7*
(Scheme 8) and carried out the hydrogenation experiments under
the same conditions as for the unlabeled compounds.
Figure 4b shows the 31P NMR spectrum containing the signals
of the labeled monohydride intermediate 19*. The evident
difference from the spectrum of the unlabeled compound (Figure
4a) is the multiplicity of the high-field signal: in the spectrum
of the labeled compound it resonates as a double doublet with
(33) Bircher, H.; Bender, B. R.; Philipsborn, W. v. Magn. Reson. Chem.
1993, 31, 293-298.
two equal couplings (86 Hz). One of these couplings is 1JRh-P
,
(34) Landis, C. R.; Feldgus, S. Angew. Chem., Int. Ed. 2000, 39, 2863-
and the other one is undoubtedly 2JC,-P, since it is absent in the
spectrum of the unlabeled compound. The presence of a large
coupling 2JC-P with the trans-phosphorus atom in the 31P NMR
spectrum of 19*a unequivocally proves the proposed structure
2866.
(35) Brown, J. M.; Chaloner, P. A. J. Chem. Soc., Perkin Trans. 2 1987,
1583-1588.
(36) Landis, C. R.; Halpern, J. J. Am. Chem. Soc. 1987, 109, 1746-
1754.

5274 J. Am. Chem. Soc., Vol. 123, No. 22, 2001
GridneV et al.
Figure 6. Hydride region of the 1H NMR spectrum of 23 (400 MHz,
CD3OD): (a) at -90 °C and (b) at -50 °C.
Figure 5. Multiplet of the CH2Rh group in the 13C NMR spectrum of
19* (125 MHz, -90 °C, CD3OD).
of the monohydride intermediate. The multiplet of the CH2Rh
group of 19*a (Figure 5) resonates at δ ) 15.8 and displays
2
1
the characteristic couplings JC-P ) 86.5 Hz and JRh-C ) 20
Hz. It is noteworthy that the signal of the high-field phosphorus
of minor isomer 19*b in the 31P NMR spectrum is also split in
a triplet (Figure 4b). This allows assignment of its structure as
the Rh-diastereomer of 19*a (as shown in Scheme 7), since
the observed enantioselectivity is higher than that calculated
based on the relative amounts of 19a and 19b, and the minor
isomer cannot, therefore, have an alternative configuration at
the R-carbon.
Intermediates in the Asymmetric Hydrogenation of o-
Methoxyphenylenamide (21). In all the previous examples we
studied, the intermediates of the catalytic reactions gave very
high optical yields. However, in certain cases much lower optical
yields were obtained in catalytic hydrogenation (entries 3 and
9 in Table 1). We chose to study the reaction pathway for the
hydrogenation of o-methoxyphenyl-substituted enamide 21 to
determine the correlation of the structure and stability of the
observed intermediates with the relatively low enantioselectivity
observed in this case.
The catalyst-substrate complex 22 obtained by addition of
a 2-fold excess of enamide 21 to the deuteriomethanol solution
of the solvate complex 11 exists in solution as a mixture of
two interconverting diastereomers 22a and 22b. The ratio 22a:
22b changes from 15:1 at -90 °C to 6:1 at 0 °C (thermodynamic
parameters of the equilibrium could not be determined accurately
in this case due to the presence of small amounts of 11 and an
additional unidentified complex in equilibrium with 22a,b and
11, detectable in the temperature interval from -90 to -60 °C).
A small amount of 11 equilibrates with 22 even at -90 °C.
The configuration of the main isomer 22a is the same as that
of 12a and 13a; in this case ∆δBut,major ) 0.89, ∆δMe,major
) 0.15and ∆δBut,minor ) 0.20, ∆δMe,minor ) 1.05.
Hydrogenation of the equilibrium mixture of 11, 21, and 22
carried out at -100 °C resulted in the detection of three isomers
of monohydride intermediate 23a-c in 100:34:10 ratio (Figure
6). The structures of 23a-c were determined by using the NMR
data (Table 2). In the hydride region of the 1H NMR spectrum
taken at -90 °C (Figure 6a) three hydride signals were found.
Heteronucleus correlation experiments and selective 31P decou-
plings made possible the assignments in the 31P NMR spectrum.
The NMR spectra of the monohydrides 23b,c are similar to those
of 19 and 20 (Table 2). They also have similar thermal stability
Figure 7. 31P NMR spectra (162 MHz, -95 °C, CD3OD) of the
reaction mixtures containing monohydride intermediates: (a) 23a-c
and (b) 23a-c*.
and decompose rapidly if the temperature of the sample is raised
over -85 °C. On the other hand, the hydride signal of 23a in
1
the H NMR spectrum is significantly high-field shifted and
has an unusually large value of 1JRh-H (39 Hz). In addition, the
1
coupling JRh-P of the low-field phosphorus is about 20%
smaller in value compared to the same couplings in 23b,c and
is almost equal to the couplings in the previously characterized
monohydride intermediates with the R-C atom bound to
rhodium.14,17,37,38 Conclusive evidence of the structures of 23a-c
was obtained in the hydrogenation experiment using a 13C-
labeled enamide 21* prepared similarly to the labeled com-
pounds 6 and 7 (Scheme 8). The 31P NMR spectra of the
samples containing labeled and unlabeled compounds are
compared in Figure 7. The signals of the major monohydride
species 23a remained unchanged: no significant coupling with
(37) Chan, A. S. C.; Halpern, J. J. Am. Chem. Soc. 1980, 102, 838-
840.
(38) Brown, J. M.; Chaloner, P. A. J. Chem. Soc., Chem. Commun. 1980,
344-346.

Asymmetric Hydrogenation of Enamides
J. Am. Chem. Soc., Vol. 123, No. 22, 2001 5275
Scheme 9
Scheme 10
electronic properties with 21, was unsuccessful: the hydrogena-
tion of the corresponding catalyst-substrate complex at -100
°C gave directly the hydrogenation product. Similar stabilization
through terdentate coordination has been reported for structurally
similar Rh17and Ru.39 However, the monohydride intermediate
23a cannot form directly from its likely precursor, dihydride
intermediate 24, since the vacant coordination site after the
migratory insertion is naturally placed in the trans-position
relative to the o-methoxyphenyl substituent (Scheme 9). Ap-
parently, the initially formed monohydride intermediate 23d
undergoes fast rearrangement to 23a via a simple rotation around
the Rh-C bond.
Mechanism of Stereoselection in the Catalytic Asymmetric
Hydrogenation of Enamides. The outcome of the experiments
using HD corresponds well to the structures of the observed
monohydride intermediates. Upon hydrogenation of solvate 11
with HD, the isotopomer of 14d with apical disposition of
deuterium forms predominantly with the factor 1.3((0.1):1.17
On the other hand, the equatorial hydride is transferred in the
migratory insertion step. If the consequence of substrate
coordination and migratory insertion is faster than the inter-
conversion of isotopomers, the partitioning of deuterium in the
products indicates the fashion of the substrate coordination
during the migratory insertion step. Therefore, the predominance
of the R-deuterated product corresponds to coordination of type
A (Scheme 10) and subsequent formation of a monohydride
intermediate with the tertiary carbon atom bound to rhodium,
whereas the predominance of the â-deuterated product implies
the coordination of type B and a monohydride intermediate with
the CH2 group bound to the rhodium atom.
The observation of the monohydride intermediates 19 and
20 with the â-carbon atom attached to Rh is in accord with the
deuterium partitioning in the products of hydrogenation of
enamides 6 and 7 with HD. The opposite sense of deuterium
partitioning in the reactions using aromatic enamides implies
another pathway through A-type coordination (Scheme 10).
There are four different ways for coordination in each of the
pathways A and B. However, we rule out the possibility of
coordination of the double bond of an enamide to the apical
coordination site of rhodium due to the evident steric hindrance
to such coordination arising from the alkyl substituents of the
catalyst. The remaining four isomers of a hypothetic dihydride
the 13C nucleus is observed. This means that in 23a the rhodium
atom is bound to the quaternary carbon atom, and the labeled
carbon should appear as a signal of the methyl group uncoupled
to phosphorus and rhodium, which indeed was found at δ )
31.8 in the 13C NMR spectrum. On the other hand, the high-
field signals in the 31P spectrum belonging to 23b* and 23c*
are split in triplets, attesting to the presence of the *CH2-Rh
bonds in these intermediates, and supporting the structures drawn
in the Scheme 9 (see the above discussion for 6* and 7*).
The monohydrides 23b and 23c can have different configu-
ration either on carbon or on rhodium. These two possibilities
cannot be distinguished from their NMR spectra. Nevertheless,
comparing the relative integral intensities in the 1H NMR
spectrum of the mixture of dihydrides and the experimental ee
obtained after quenching the NMR sample, a definite conclusion
can be made on the structure of 23b and 23c. Indeed, the ratio
([23a] + [23c] - [23b]):([23a] + [23b] + [23c]) (0.53)
corresponds exactly to the ee of the hydrogenation product
obtained after quenching of this sample (53%) and is close to
the ee value obtained under catalytic conditions (50%). Since
the deuteration experiments attest to the irreversibility of the
migratory insertion step, we conclude that 23a and 23c produce
R-amide, whereas 23b is a precursor of S-amide. Therefore, the
configurations of the R-carbon atoms are different in the
diastereomeric monohydrides 23b and 23c.
The relative stability of the monohydride intermediate 23a,
compared to monohydrides from the catalytic cycles of the
hydrogenation of 5 and 10, apparently results from the pos-
sibility of its stabilization with the adjacent o-methoxy group
(Scheme 9). Thus, our attempt to detect a monohydride from
p-methoxy-substituted enamide, which apparently has similar
(39) Wiles, J. A.; Bergens, S. H.; Young, V. G. J. Am. Chem. Soc. 1997,
119, 2940-2941.

5276 J. Am. Chem. Soc., Vol. 123, No. 22, 2001
GridneV et al.
Scheme 11
substituent of the substrate is much farther distanced from the
alkyl groups of the catalyst (intermediates B1 and B2 in Scheme
11). The relative availability of these intermediates is now
determined by the remaining interaction between the bulky
group R (t-Bu or adamantyl) and one of the alkyl groups of the
catalyst: a methyl group in the case of B1 and a t-Bu group in
the case of B2.
In the case of the o-methoxyphenyl-substituted enamide 21,
these two reaction pathways are apparently competitive. Since
the ratio of the observed monohydride intermediates 23a-c is
equal to the ee observed in the catalytic hydrogenation, we may
conclude that other competitive pathways do not contribute
significantly to the flux of catalysis. This, in turn, leads to a
conclusion that the intermediate formation of dihydride A2 is
less favored in this case than the intermediacy of B1 or B2.
Conclusions
The t-Bu-BisP*-Rh catalyst is one of the best catalysts for
the asymmetric hydrogenation of enamides reported so far.
Various aryl-substituted enamides give the corresponding opti-
cally active amides with 96-99% ee, with the exception of
ortho-substituted compounds which afford moderate enantiose-
lectivity. Enamides containing t-Bu and 1-adamantyl groups at
the R-position give excellent ee values, but the configuration
of the hydrogenation products obtained with the use of the same
catalyst is opposite to that of aryl-substituted enamides. This
dramatic difference is explained by the opposite mechanisms
of stereoselection operating in these polar cases. The relatively
low optical yields observed for the o-methoxyphenyl- and
o-chlorophenyl-substituted enamides may be attributed to the
simultaneous running of both reaction pathways.
intermediate are shown in Scheme 11. Similarly to the asym-
metric hydrogenation of dehydroamino acids,17 we conclude that
the main stereoregulating factor in the case of pathway A is
the steric interaction between the chelate ring made by the
substrate and the adjacent alkyl group of the catalyst. In the
dihydride intermediate A1 this interaction is effectively reduced
(the chelate ring is close to the Me group of the catalyst), which
is in accord with R-stereoselection observed in the hydrogenation
of aromatic enamides.
In the case of the enamides containing bulky t-Bu and
adamantyl substituents, intermediate A1 becomes less acces-
sible: reducing the steric repulsion between the chelate cycle
of the substrate and the alkyl groups of the catalyst automatically
leads to drastic hindrance between the t-Bu group of the catalyst
and the t-Bu or adamantyl group of the substrate, and vice versa,
if the latter interaction is avoided in intermediate A2, the close
contact of the chelate cycle and the t-Bu group of the catalyst
increases the energy of this configuration.
Acknowledgment. This work was supported by “Research
for the Future” Program, the Japan Society for the Promotion
of Science, the Ministry of Education, Japan.
Supporting Information Available: Experimental details,
Tables S1-S4, Figures S1 and S2, charts of NMR spectra for
all important products and intermediates, and Cartesian coor-
dinates for 17′a and 17′b (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
Therefore, it is reasonable to conclude that bulky enamides
choose an alternative mode of coordination, in which the
JA010161I