Observation of a stable cis-diphosphine solvate rhodium dihydride derived from PHANEPHOS

Article abstract of DOI:10.1039/b102316k

The methanol solvate rhodium(PHANEPHOS) forms a stable dihydride which has been characterised in solution by NMR as a pair of equilibrating diastereomers.

Full text of DOI:10.1039/b102316k

Observation of a stable cis-diphosphine solvate rhodium dihydride
derived from PHANEPHOS
Hanjo Heinrich,a Ralf Giernoth,b Joachim Bargon*a and John M. Brown*b
a Institute of Physical and Theoretical Chemistry, University of Bonn, Wegelerstr. 12, D-53115 Bonn,
Germany. E-mail: bargon@uni-bonn.de
b Dyson Perrins Laboratory, South Parks Road, Oxford, UK OX1 3QY. E-mail: bjm@ermine.ox.ac.uk
Received (in Cambridge, UK) 12th March 2001, Accepted 24th May 2001
First published as an Advance Article on the web 27th June 2001
The methanol solvate rhodium(PHANEPHOS) forms a
stable dihydride which has been characterised in solution by
NMR as a pair of equilibrating diastereomers.
Early experiments designed to elicit the mechanism of asym-
metric homogeneous hydrogenation provided a contrast be-
tween chelate diphosphine and bisphosphine rhodium com-
plexes. The solvate 1a showed no tendency to react with
Scheme 1 The possible paths for addition of dihydrogen to a dehydroamino
ester; path A: H2 addition prior to substrate (dihydride route). Path B:
substrate addition prior to H2 addition (unsaturate route).
ambient hydrogen, whilst the solvate 2 formed a characterised
dihydride;1 the difference was attributed to a requirement for
trans-diphosphine geometry in the stable solvate with H
correspondingly trans to solvent oxygen. This observation has
generally been sustained until recently. Aside from reversible
ortho–para dihydrogen equilibration by complexes 1a and 1b,2
and the likely mediation of a cis-dihydride in the formation of
the dimeric species 3,3 no further progress had been made prior
to the work of Gridnev, Imamoto and coworkers.4 They
demonstrated that the corresponding solvate 4 from a simple P-
chiral alkylphosphine ligand formed significant quantities of the
cis-dihydride 5 (20% at 295 °C and ambient pressure), with
two diastereomers formed in a ratio of 10+1.5 Further, this
intermediate reacted with the catalytic substrate 6a to form a Rh
alkylhydride,6 which then underwent reductive elimination at
250 °C to give the hydrogenated product. Taken together with
labelling studies, the results are compatible with path A in
Scheme 1, in contrast to the more generally accepted sequence
B.7
We recently demonstrated the presence of an agostic
dihydride intermediate 7 in the hydrogenation cycle of com-
pound 6a by [PHANEPHOS]Rh+, employing para-enriched
hydrogen and the precursor complex 8 (or the NBD analogue)
to identify the transient at 210 to 230 °C by 1H NMR.8 When
hydrogenation is complete and the substrate exhausted a second
species can be observed, however. By carrying out the
hydrogenation of the catalyst precursor in the absence of
substrate the same intermediate is seen, optimally at 240 °C.
The d and J values are entirely consistent with a cis-dihydride
structure 9, with one hydride trans to phosphorus (d ca. 211)
and one trans to one of the two solvent oxygens (d ca. 220).
Both the intensity and persistence of the signals indicate that it
is a relatively robust species. There are two diastereomers 9a
and 9b in 2+1 ratio, and their NMR spectra have been fully
assigned using PHIP++ [Fig. 1(a)].9 The chemical shifts are
very different from the previous case4 where the observed major
diastereomer resonates at d 27.7 and 223.0.
When hydrogenation is carried out under conventional NMR
conditions at 280 °C, the same dihydride 9 may be observed,
along with small amounts of other Rh hydride resonances not
seen in the PHIP spectrum. It is stable up to 240 °C [Fig. 1(b)],
1
and a rough estimate based on integration of the high-field H
NMR signals against the CH2-region of the ligand indicates that
45% of species 9 is formed at equilibrium, making it more
accessible than the previously observed case,4,5 and to higher
1296
Chem. Commun., 2001, 1296–1297
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b102316k

View Article Online
Fig. 2 The PHIP 1H NMR spectrum of the hydrogenation of reactant 6b in
CD3OD in the presence of Rh complex 8 at 227 °C, after 40 5 s pulses of
parahydrogen (98%). Spectra taken earlier in the sequence after 16 pulses
show only traces of complex 9.
observed that when the PHIP experiment was carried out with
6b as substrate at the lower temperature of 227 °C, the solvate
dihydride 9a, b could be observed in significant amount, but
only late in the reaction sequence when the substrate concentra-
tion was depleted (Fig. 2). This opens up the possibility that path
A may contribute to catalytic turnover in the PHANEPHOS
case. Earlier INEPT experiments demonstrated that the agostic
intermediate 7a is in reversible equilibrium with the solvate
complex and substrate.7 This makes the discrimination between
the two pathways quite subtle. Given that both species 7 and 9
are observed in the same experiment under turnover conditions,
the result is accessible in principle and a challenge for further
work.
We thank Philip Pye and Kai Rossen (Merck, Rahway) for a
generous gift of PHANEPHOS, and Johnson-Matthey for the
loan of RhCl3. R. G. thanks BASF AG and Studienstiftung des
Deutschen Volkes for a Fellowship. JMB is very pleased to
acknowledge an unrestricted grant from Merck, Inc. H. H. and
J. B. thank the Deutsche Forschungsgemeinschaft for financial
support.
Notes and references
1 J. Halpern, D. P. Riley, A. S. C. Chan and J. J. Pluth, J. Am. Chem. Soc.,
1977, 99, 8055; D. A. Slack, I. Greveling and M. C. Baird, Inorg. Chem.,
1979, 18, 3125
2 J. M. Brown, L. R. Canning, A. J. Downs and A. M. Forster,
J. Organomet. Chem., 1983, 255, 103.
3 K. Tani, T. Yamagata, Y. Tatsuno, T. Saito, Y. Yamagata and N.
Yasuoka, J. Chem. Soc., Chem. Commun., 1986, 494.
Fig. 1 (a) The PHIP 1H NMR spectrum (CD3OD, 200 MHz) of dihydrides
9a and 9b taken after parahydrogen (98% enriched) passage through a
solution of complex 8 in CD3OD at 240 °C. Minor diastereomer: d 210.87
(JHP 170, 24, JHRh 14, JHH 29.0 Hz), 221.77, (JHP 32, 12, JHRh 21 Hz);
major diastereomer : d 211.46 (JHP 171, 28, JHRh 15.5, 15, JHH 27.5 Hz),
220.91, (JHP 29, 16, JHRh 22.5 Hz). (b) The 1H NMR spectrum (CD3OD,
500 MHz) of dihydrides 9a and 9b formed in the hydrogenation of complex
8 at 280 °C, taken at 240 °C, with comparable J and d values.
4 I. D. Gridnev, N. Higashi, K. Asakura and T. Imamoto, J. Am. Chem.
Soc., 2000, 122, 7183.
temperatures. This may be attributed to the high level of
electron donation ensuing from the [2.2]paracyclophane back-
bone,10 together with the large bite angle of PHANEPHOS,11
which will favour the dihydride at equilibrium. The two
diastereomers are in equilibrium by an unselective mechanism,
as indicated by a selective homodecoupling experiment.12
When the solution containing complex 9 is held at 280 °C
and a solution of compound 6a in MeOH added, rapid formation
of the agostic dihydride 7a occurs. The signals at d 22 and 219
are broad at that temperature, and at 270 °C they decay over
time without formation of any further observable intermediates.
The absence of a ‘classical’ alkylhydride 10 indicates that 7a is
the only accessible intermediate on the hydrogenation pathway.
Further, it must be formed directly from an assumed dihydride
precursor rather than by reinsertion of rhodium into the b-CH of
10 after formation of the latter, since the latter pathway would
vitiate the earlier PHIP experiment by uncoupling the H–H
spins.
5 A stable Rh solvate dihydride has been observed recently: I. D. Gridnev,
N. Higashi and T. Imamoto, Organometallics, 2001, submitted; we
thank Dr Gridnev for a useful exchange of information.
6 I. D. Gridnev, N. Higashi and T. Imamoto, J. Am. Chem. Soc., 2000,
122, 10486; J. A. Ramsden, T. Claridge and J. M. Brown, J. Chem. Soc.,
Chem. Commun., 1995, 2469; J. M. Brown and P. A. Chaloner, J. Chem.
Soc., Chem. Commun., 1980, 344; A. S. C. Chan and J. Halpern, J. Am.
Chem. Soc., 1980, 102, 838.
7 S. Feldgus and C. R. Landis, J. Am. Chem. Soc., 2000, 122, 12714.
8 R. Giernoth, H. Heinrich, N. J. Adams, R. J. Deeth, J. Bargon and J. M.
Brown, J. Am. Chem. Soc., 2000, 122, 12381.
9 Simulations were conducted with the help of the program PHIP++
written by T. Greve (PhD thesis 1996, University of Bonn, Institute of
Physical and Theoretical Chemistry).
10 D. J. Cram and J. M. Cram, Acc. Chem. Res., 1971, 4, 204; Z. Yang, B.
Kovac, E. Heilbronner, S. Eltamany and H. Hopf, Helv. Chim. Acta,
1981, 64, 1991.
11 P. W. Dyer, P. J. Dyson,S. L. James, C. M. Martin and P. Suman,
Organometallics, 1998, 17, 4344; PPdP
= 103.7° in the PdCl2
complex.
In the earlier publication of Gridnev, Imamoto and cowork-
ers,4 it was suggested that path A could be a viable alternative
to the accepted reaction mechanism of path B (Scheme 1). We
12 Selective irradiation at the site of Hb2 demonstrates concurrent loss of
50% intensity at Ha1, Ha2 and Hb1 without selectivity. We thank Dr Tim
Claridge for help with this experiment.
Chem. Commun., 2001, 1296–1297
1297