Interconversion of monohydride intermediates in Rh(I)-catalyzed asymmetric hydrogenation of dimethyl 1-benzoyloxyethenephosphonate [16]

Full text of DOI:10.1021/ja015611l

J. Am. Chem. Soc. 2001, 123, 4631-4632
Scheme 1. Formation of Catalyst-Substrate Complex 4
Rh(I)-Catalyzed Asymmetric Hydrogenation of
Dimethyl 1-Benzoyloxyethenephosphonate
4631
Interconversion of Monohydride Intermediates in
Ilya D. Gridnev,* Natsuka Higashi, and Tsuneo Imamoto*
Department of Chemistry, Faculty of Science
Chiba UniVersity, Chiba 263-8522, Japan
ReceiVed January 31, 2001
ReVised Manuscript ReceiVed April 7, 2001
intermediates, which are the precursors of the product with the
opposite chirality.
Recently, the interest in the exact mechanism of stereoselection
in the Rh(I)-catalyzed asymmetric hydrogenation of activated
double bonds is renewed due to both the development of the
computational techniques1-3 and the new experimental findings.4-7
Thus, the extensive ab initio DFT calculations were used by
Landis et al. to support the possibility of enantioselection during
the oxidative addition of dihydrogen to a catalyst-substrate
complex of a model Rh-diphosphine complex1 and the real Rh-
DuPHOS catalytic system.2,3 On the other hand, we have observed
almost perfect enantioselection in the low-temperature reaction
of a solvate dihydride [RhH2(t-Bu-BisP*)(CD3OD)2]BF4 (1) with
methyl (Z)-R-acetamidocinnamate,4 demonstrating that a dihydride
pathway can also operate in the Rh(I)-catalyzed asymmetric
hydrogenation. This result was reproduced for other BisP*-Rh
complexes and a series of substrates.5 Dihydride pathway is also
in accord with the dramatic difference of the sense of enantiose-
lection in catalytic asymmetric hydrogenation of aryl- and alkyl-
substituted enamides.6 Brown et al. reported recently the detection
of an agostic intermediate in the asymmetric hydrogenation of
dehydroamino acids catalyzed by the Rh-PHANEPHOS complex.7
Interestingly, this intermediate was computationally predicted to
be inaccessible via the unsaturated pathway;2,3,7 however, its
appearance via the dihydride mechanism is quite possible. In
addition, the reversibility of the formation of the agostic
intermediate with respective to substrate reported by Brown et
al.7 is also best explained if the relative stability of the solvate
dihydride derived from the Rh-PHANEPHOS catalyst is proposed.
We report here the first observation of the structural rearrange-
ment of monohydride intermediates in Rh(I)-catalyzed asymmetric
hydrogenation, and demonstrate the interplay of two possible
reaction pathways in this theoretically and practically important
reaction.
Addition of a 2-fold excess of the phosphonate 2 to a
deuteriomethanol solution of 3 yielded a catalyst-substrate
complex 4 (Scheme 1), which displayed a single set of signals in
the 31P NMR spectrum within the temperature range from -100
to +60 °C. Complexation was fast even at -100 °C; no detectable
equilibrium amounts of free solvate 3 were found in the spectra
even at elevated temperatures. The chemical shift of the carbonyl
carbon atom in 4 (δ ) 178.9, compare to δ ) 165.1 in
uncoordinated substrate) and two vicinal C-P couplings observed
for this signal (5 and 8 Hz) confirm the mode of chelating
coordination by the double bond and the benzoyloxy group of
the substrate. The solution structure of the observed diastereomer
of 4 was determined from the NOE and EXSY data.
We carried out the reaction of dihydride 14 with 2-fold excess
of phosphonate 2 at -100 °C. The signals of two monohydride
intermediates 5a,b in the ratio 100:5 were detected in the hydride
region of the 1H NMR spectra (Figure 1a).10 Raising the
temperature of the sample to -30 °C resulted in disappearance
of the signals of 5a,b and simultaneous growth of new hydride
signals at δ ) -19.7 (6a) and -19.3 (6b); the ratio of 6a:6b
was the same as that of 5a:5b within the experimental accuracy
(Figure 1b-d). Recooling of the sample to -60 °C gave an
additional low-intensity hydride signal at δ ) -22.0 (7a) (Figure
1e).
Attempts to hydrogenate the catalyst-substrate complex 4 in
the temperature interval from -100 to -60 °C were unsuccessful;
no reaction occurred under these conditions. The hydrogenation
of 4 with 2 atm of H2 carried out for 10 min at -30 °C gave a
mixture of the monohydride intermediates 5-7 (Figure 1f)
together with the hydrogenation product. Comparing the spectrum
to that obtained in the previous experiment (Figure 1e), one can
see two main differences: the ratio of 6a:6b changed to 5:1 and
an additional signal at δ ) -21.8 (7b) appeared.
Both experiments were reproduced three times to give es-
sentially the same results. The overnight EXSY experiment carried
out at -60 °C (Figure 2) showed that 6a is interconverting with
7a, and 6b is interconverting with 7b, but no exchange between
6a and 6b or 7a and 7b takes place. These data together with the
correlation of ee values obtained from the NMR samples11 and
relative integral intensities of the monohydride signals in the NMR
spectra testify that the monohydrides 5a, 6a, and 7a are the
precursors of the S-hydrogenation product, whereas 5b, 6b, and
7b give minor R-product after the reductive elimination. The
hydrogenation of the catalyst-substrate complex 4 always gave
markedly lower ee values compared to those observed in the low-
Recently, the asymmetric hydrogenation of protected R,â-
unsaturated R-acyloxyphosphonates was shown to be synthetically
useful.8 We found that the catalytic hydrogenation of dimethyl
1-benzoyloxyethenephosphonate (2) using Rh-(S,S)-t-Bu-BisP*
catalyst gave the corresponding product with 88% ee (S),9 which
is comparable to the best results obtained by Burk et al.8 On the
other hand, the not perfect stereoselectivity observed in this
hydrogenation allows one to analyze minor diastereomers of the
(1) Landis, C. R.; Hilfenhaus, P.; Feldgus, S. J. Am. Chem. Soc. 1999,
121, 8741-8754.
(2) Landis, C. R.; Feldgus, S. Angew. Chem., Int. Ed. 2000, 39, 2863-
2866.
(3) Feldgus, S.; Landis, C. R. J. Am. Chem. Soc. 2000, 122, 12714-12727.
(4) Gridnev, I. D.; Higashi, N.; Asakura, K.; Imamoto, T. J. Am. Chem.
Soc. 2000, 122, 7183-7194.
(5) Gridnev, I. D.; Yamanoi, Y.; Higashi, N.; Tsuruta, H.; Yasutake, M.;
Imamoto, T. AdV. Synth. Catal. 2001, 343, 118-136.
(6) Gridnev, I. D.; Higashi, N.; Imamoto, T. J. Am. Chem. Soc. 2000, 122,
10486-10487.
(10) In the 13C and 31P NMR spectra taken at -95 °C an additional
intermediate was detected. Its spectra resemble those of 4, but it is not the
second diastereomer of the catalyst-substrate complex, since it does not form
in the absence of hydrogen under the same conditions. A molecular hydrogen
complex (see refs 1-3) with a broad undetectable signal in 1H NMR is
possible.
(11) The ee of the product in both experiments was determined by HPLC
analysis after warming the sample to room temperature to complete the
reductive elimination (NMR control) and removing Rh complexes by filtration
through silica gel.
(7) Giernoth, R.; Heinrich, H.; Adams, N. J.; Deeth, R. J.; Bargon, J.;
Brown, J. M. J. Am. Chem. Soc. 2000, 122, 12381-12382.
(8) Burk, M. J.; Stammers, T. A.; Straub, J. A. Org. Lett. 1999, 1, 387-
390.
(9) The detailed account on the asymmetric hydrogenation of R,â-
unsaturated phosphonates using Rh-BisP* and Rh-MiniPHOS catalysts is in
preparation.
10.1021/ja015611l CCC: $20.00 © 2001 American Chemical Society
Published on Web 04/19/2001

4632 J. Am. Chem. Soc., Vol. 123, No. 19, 2001
Communications to the Editor
Scheme 2. Formation and Transformations of Monohydride
Intermediates 5-7
Figure 1. Hydride region of the 1H NMR spectra (400 MHz, CD3OD):
(a) the sample obtained by mixing the CD3OD solutions of 1 and 2 at
-100 °C, spectrum taken at -95 °C; (b) the same sample after raising
the temperature to -70 °C, (c) to -50 °C, (d) and to -30 °C; (e) the
sample after recooling to -60 °C; (f) the sample obtained by hydrogena-
tion of the CD3OD solution of 4 for 10 min at -30 °C, spectrum taken
at -60 °C.
8a and 8b are the most probable unobservable intermediates
formed by complexation of the substrate to the two interconverting
isomers of solvate dihydride 1.4 The failure to observe these
intermediates is in accord with the recent computational data,
which predict extremely low barriers of migratory insertion in
the dihydrides of similar structure.2,3 The occurrence of migratory
insertion through the dihydride intermediate 8a is more favorable
compared to the reaction of 8b due to steric reasons.4
Monohydride intermediates 5a,b are the direct products of the
migratory insertion in 8a,b; in these intermediates the dimethoxy-
phosphinoyl group is located in the trans-position to the vacant
coordination site, and they rearrange to 6a,b in which the
additional coordination of the P(O)(OMe)2 group is possible.
Rearrangement proceeds through the equilibrium amounts of
nonchelating complexes 7a,b, which were also detected in our
experiments. We have suggested such transformation in our
previous work. However, this is the first experimental observation
of this type of rearrangement, as well as of the interconversion
between monohydride intermediates.
Calculations suggest that the oxidative addition of dihydrogen
to 4 should directly lead to 6.2,3 Therefore, when 4 was
hydrogenated at -30 °C, 6a,b might be partially produced via
this pathway. The lower selectivities observed in these experi-
ments are therefore attributable to the lesser effectivity of
enantioselection in the unsaturated pathway compared to dihydride
in this catalytic hydrogenation.
Figure 2. Phase-sensitive 1H-1H EXSY spectrum (400 MHz, CD3OD,
-60 °C) of the equilibrium mixture of monohydride intermediates 5-7.
temperature reaction of 2 with 1, which corresponded to the
different ratios of 5-7a:5-7b in these experiments.
The structures of the monohydride intermediates 5-7 were
elucidated on the basis of their NMR spectra. The chemical shifts
of their hydride protons confirm that in all six compounds the
hydrides are disposed cis respective to both phosphorus atoms.
Experiments applying the R-13C labeled phosphonate 2* proved
that in all observed monohydride intermediates the R-carbon atom
is bound to the rhodium atom. The detailed signal assignments
and measurement of coupling constants were made with the use
of 2D NMR correlation techniques and selective decoupling
experiments. Structures of 5-7 elucidated from these data are
shown in Scheme 2.12
Acknowledgment. This work was supported by the Research for the
Future program, the Japan Society for the Promotion of Science.
We suggest the following sequence of processes explaining
our experimental data (see Scheme 2). The dihydride intermediates
Supporting Information Available: NMR charts of all important
intermediates (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
(12) The low concentration of 7a retarded the observation of the signal of
the carbonyl carbon atom in the 13C NMR spectrum. Nevertheless, in both 5a
and 6a the benzoyloxy group is coordinated, and the fast exchange between
6a and 7a gives a strong support for the suggested structure of 7a.
JA015611L