Chiral pentacyclic phosphines as a new ligand class

Article abstract of DOI:10.1039/b207937b

A new sterically encumbered monophosphine has been prepared stereoselectively from (1R)-camphor and some aspects of its coordination chemistry with palladium examined.

Full text of DOI:10.1039/b207937b

Chiral pentacyclic phosphines as a new ligand class†‡
Robert Haigh, K. M. Abdul Malik and Paul D. Newman*
Department of Chemistry, Cardiff University, Park Place, Cardiff, UK CF10 3TB.
E-mail: newmanp1@cardiff.ac.uk; Fax: +44 29 20874030; Tel: +44 29 20876824
Received (in Cambridge, UK) 14th August 2002, Accepted 23rd September 2002
First published as an Advance Article on the web 4th October 2002
A new sterically encumbered monophosphine has been
prepared stereoselectively from (1R)-camphor and some
aspects of its coordination chemistry with palladium ex-
amined.
performed in THF at rt to give the phosphinodiol 2 as the all
endo isomer. Attempts to cyclise this diol directly to phenop
have thus far been unsuccessful. However, conversion of the
phosphinodiol to the oxide 3 is readily achieved without
compromising the chiral integrity of the compound and reacting
3 with two mol equivalents of toluene-p-sulfonyl chloride in
toluene at reflux gave phenop as the oxide (4) in 73% yield.
Compound 4 was reduced with excess trichlorosilane in
refluxing toluene over 24 h to give phenop in 40% overall yield
(based on camphor).
The isolation of the endo,endo isomer of 1 accords with the
observations of Shaw et al. on the related bicyclic diph-
enylphosphine system; these workers showed that this was the
thermodynamically favoured isomer.7 The subsequent ster-
eoselective LiAlH4 reduction of 1 to the endo,endo-alcohol 2
also accords with previous observations on the bicyclic system.8
The stereochemical outcome of the ring forming reaction (3 ?
4) was less predictable, as no precedent for such a conversion
was evident in the literature, and it came as some surprise that
only a single isomer was observed and isolated in good yield.
Thus, starting from camphor with a predefined stereochemistry
(in our case the 1R enantiomer), a further four chiral centres are
created in a controlled manner during the synthesis of 4.
Assignment of the absolute configuration of these newly
generated centres was made possible through determination of
the crystal structure of 4 (Fig. 1).§
The compound is an unusual pentacyclic phosphine which
has eight stereogenic centres and two sets of fused 5- and
6-membered rings (from the camphor) flanking a central
6-membered oxaphosphinane (oxide) ring. The peripheral
6-membered rings are necessarily boat as dictated by the [2.2.1]
bicyclic structure of camphor, but surprisingly the central
heterocyclic ring also adopts a boat conformation. The phenyl
substituent on the phosphorus projects equatorially and the
oxide is axial. The phosphorus centre may be defined as being
pseudochiral as, although it possesses two structurally equiva-
lent groups, the stereogenic carbon atoms directly bound to the
Metal complexes of electron-rich, bulky monodentate phos-
phines and, in particular, related cyclometallated (P,C) deriva-
tives have proved valuable catalysts for a number of synthet-
ically appealing C–C and C-heteroatom bond forming
reactions.1 Palladium remains the metal of choice for Heck and
Suzuki type coupling, although phosphaplatinacycles have
recently been shown to be highly efficient Suzuki catalysts.2
Whilst considerable interest has focused on these systems, the
development of chiral derivatives for the asymmetric synthesis
of, for example biaryls, has been slow to take off. Buchwald3
and Cammidge4 have employed optically pure binaphthyl- and
ferrocenyl-derived phosphines for the production of various
chiral biaryls with modest to good enantioselectivities. Others
have used similar type ligands5 but have observed little to no
asymmetric induction in Heck couplings, although sulfur
donating metallocycles6 have produced modest enantioselectiv-
ities. Thus, the search for new ligands that may prove effective
in this type of reaction and others remains a focal point for
chemists within this field.
We are interested in developing inherently chiral monop-
hosphines with unusual, rigid, multicyclic structures that
contain alkyl substituents predisposed for cyclometallation with
a view to employing them as ligands in catalysis. To this end we
have synthesised the new ligand (1S,4R,4aS,5aR,6R,9S,9a-
S,10aR)-4,6,11,11,12,12-hexamethyl-10-phenyldodecahydro-
1,4:6,9-dimethanophenoxaphosphinine (phenop) from readily
available (1R)-camphor (Scheme 1).
The reaction of the enolate of camphor with dichlor-
ophenylphosphine gave the diketophosphine 1 in high yield.
Recrystallisation of 1 from ethanol gave a single isomer
determined from NMR data to be endo,endo-1. Lithium
aluminium hydride reduction of the carbonyl functions was
Scheme 1 Reagents: (i) n-BuLi, 0.5 PhPCl2; (ii) LiAlH4; (iii) H2O2; (iv)
TsCl, toluene, reflux; (v) HSiCl3.
† Electronic supplementary information (ESI) available: characterisation
data for the new compounds. See http://www.rsc.org/suppdata/cc/b2/
b207937b/
‡ Dedicated to the memory of Sam M. Liddiard, a true friend and
colleague.
Fig. 1 ORTEP representation of 4.
2558
CHEM. COMMUN., 2002, 2558–2559
This journal is © The Royal Society of Chemistry 2002

phosphorus have the opposite configuration (one is S, the other
R). The phosphorus and oxygen atoms in the ring are disposed
endo- with respect to one of the [2.2.1] bicyclic units and exo-
with regard to the other, i.e. two carbon centres, one adjacent to
oxygen, one to phosphorus, have inverted on conversion of 3 to
4. This suggests that an elimination (dehydration) occurs prior
to cyclisation to give a vinyl phosphine oxide that adds the
remaining alcohol in a Michael-type fashion to yield the
pentacycle 4. Reduction of the phosphine oxide to the
phosphine occurs with retention of stereochemistry at all
centres, i.e. the oxide function has simply been removed. The
resultant phenop can be crystallised from ethanol and is air-
stable in the solid state indefinitely; the half-life for oxidation in
solution is about one week (31P NMR in dichloromethane).
Addition of 1 mole equivalent of phenop per palladium to
Pd3(OAc)6 in toluene gives, after 48 h stirring at rt and removal
of solvent, a colourless air-stable solid in high yield (84%). The
compound was crystallised from toluene in air and an X-ray
structure revealed it to be a bis(cyclometallated ligand)bis(m-
Pd atoms being close to coplanar and the remaining oxygen of
1
the k acetate at the vertex of the flap of the envelope. The two
distinct acetates are readily distinguished in the infrared
2
spectrum of 6 where the k type is identified by ns at 1403 cm21
1
and nas at 1557 cm21 and the k form by stretches at 1285 (ns)
and 1656 (nas) cm21
.
The 31P{1H} NMR spectrum of 6 in d8-toluene is a broad
singlet at d 17 ppm which sharpens on heating to 70 °C. On
cooling to low temperature, the singlet broadens further before
splitting into two separate resonances in a 1+1 ratio at 270 °C.
This behaviour is mirrored in the 1H NMR where the five
distinct methyl resonances split into ten singlets at low
temperature. Broad NMR spectra are typical of these types of
phosphapalladacycle dimers, and have been interpreted as
resulting from an equilibrium between the dimer and a
monomeric species. We do not believe that this is the case here,
rather the central 6-membered dipalladium diacetate ring in 6 is
fluxional, with a particular conformation, such as that in Fig. 2,
being frozen out at low temperature. In this conformation, the
phosphorus donors become inequivalent as do the remaining
ligand fragments in the formally C1 complex and the two
phosphorus nuclei and all proton groups appear as separate
resonances in the respective NMR spectra. Further features of
the coordination chemistry of these new ligands and the
application of the resulting complexes in catalysis will appear in
the near future.
1
2
acetato) dimer, cis-[Pd2(m-k -OAc)(m-k -OAc)(kP,kC-phe-
nop)2], 6 (Fig. 2).§
Several features of the structure are worthy of comment.
Firstly, the phenop ligands have cyclometallated at the
9-methyl position of one of the [2.2.1]-bicycloheptane units of
each ligand and are thus bound as P,C bidentates: such
phosphapalladacycles are well established for other bulky
phosphines notably tri-tert-butylphosphine9 and tri-ortho-tolyl-
phosphine.1f The resulting chelates are fused 6- (pseudo-boat)
and 7-membered (pseudo-chair) rings for each ligand. The Pd–P
bond lengths are relatively short at 2.197 Å (av.) compared to
non-cyclometallated phosphine complexes (2.24–2.35 Å), and
the phosphorus donors are orientated cis with respect to the Pd–
Pd axis. Cyclometallation at the C(9) methyl of each ligand
generates a further chiral centre at the C(7) carbon of the
original camphor units. These stereogenic centres have the
absolute configuration S as dictated by the structure of the
oxaphosphine ligands; the pro-R methyls being disposed away
from the metal centres. Furthermore, the phosphorus atoms
themselves become stereogenic centres on coordination with
the absolute configuration S. Thus, coordination in this mode
generates two further chiral centres (one carbon, one phospho-
rus) in each ligand with absolute stereospecificity. The bridging
acetates have an unusual arrangement with one adopting the
Support for this work from the Department of Chemistry at
Cardiff University is gratefully acknowledged. We thank the
EPSRC for support of the X-ray crystallography facility in
Cardiff.
Notes and references
§
Crystal data for 4: C26H37O2P, M = 412.53, orthorhombic, P212121, a
= 11.656(2), b = 12.060(2), c = 16.579(3) Å, V = 2330.5(7) Å3, Z = 4,
Dc = 1.176 g cm23, m(Mo-Ka) = 0.71073 Å, T = 293 K, 16833 reflections
collected, 5310 independent reflections [R(int) = 0.0465], F2 refinement,
R1 = 0.0365, wR2 = 0.0831, 268 parameters. The absolute structure was
correctly indicated by the Flack parameter being zero within experimental
error [20.07(7)] based on the comparison of 2298 Freidel pairs (hkl and
h¯kl). For 6·0.25C7H8: C57.75H80O6P2Pd2, M = 1144.96, orthorhombic,
P212121, a
= 12.7814(2), b = 18.9991(3), c = 23.0978(5) Å, V =
5608.96(17) Å3, Z = 4, Dc = 1.356 g cm23, m(Mo-Ka) = 0.71073 Å, T =
150 K, 36501 reflections collected, 10215 independent reflections [R(int) =
0.0758], F2 refinement, R1 = 0.0456, wR2 = 0.0944, 621 parameters. The
absolute structure was correctly indicated by the Flack parameter being zero
within experimental error [0.01(3)] based on the comparison of 4534 Freidel
pairs (hkl and h¯kl). CCDC 191755 and 191756. See http://www.rsc.org/
suppdata/cc/b2/b207937b/ for crystallographic data in CIF or other
electronic format.
2
familiar m-k mode through coordination of one oxygen to Pd(1)
and the second oxygen to Pd(2) whereas the other acetate binds
1
in a m-k fashion using a single oxygen to coordinate both
palladium atoms. This produces a central Pd2O3C 6-membered
ring that is hitherto unknown for complexes of this type. The
2
ring has an envelope conformation with the k -acetate and two
1 The literature is extensive. For representative papers and associated
references see: (a) J. P. Wolfe, R. A. Singer, B. H. Yang and S. L.
Buchwald, J. Am. Chem. Soc., 1999, 121, 9550; (b) M. R. Netherton, C.
Dai, K. Neuschütz and G. C. Fu, J. Am. Chem. Soc., 2001, 123, 10099; (c)
G. C. Lloyd-Jones, Angew. Chem., Int. Ed., 2002, 41, 953; (d) S. Lee, N.
A. Beare and J. F. Hartwig, J. Am. Chem. Soc., 2001, 123, 8410; (e) R. B.
Bedford and C. S. J. Cazin, Chem. Commun., 2001, 1540; (f) W. A.
Herrmann, V. P. W. Böhm and C.-P. Reisinger, J. Organomet. Chem.,
1999, 576, 23.
2 R. B. Bedford, S. L. Hazelwood and D. A. Albisson, Organometallics,
2002, 21, 2599.
3 J. Yin and S. L. Buchwald, J. Am. Chem. Soc., 2000, 122, 12051.
4 A. N. Cammidge and K. V. L. Crépy, Chem. Commun., 2000, 1723.
5 M. Shibasaki and E. M. Vogl, in BComprehensive Asymmetric CatalysisB,
eds. E. N. Jacobsen, A. Pfaltz and H. Yamamoto, Springer-Verlag,
Berlin, 1999, vol. 1, ch. 14, pp. 458–487.
6 J. Dupont, A. S. Gruber, G. S. Fonseca, A. L. Monteiro, G. Ebeling and
R. A. Burrow, Organometallics, 2001, 20, 171.
7 S. D. Perera, B. L. Shaw, M. Thornton-Pett and J. D. Vessey, Inorg.
Chim. Acta, 1992, 200, 149.
1
2
Fig. 2 ORTEP representation of cis-[Pd2(m-k -OAc)(m-k -OAc)(kP,kC-
phenop)2], 6. Selected bond lengths (Å) and angles (°): Pd(1)–P(1) 2.195(2),
Pd(2)–P(2) 2.201(2), Pd(1)–C(19) 2.036(5), Pd(2)–C(45) 2.029(6), Pd(1)–
O(3) 2.166(4), Pd(1)–O(5) 2.130(4), Pd(2)–O(4) 2.163(4), Pd(2)–O(5)
2.136(4); P(1)–Pd(1)–C(19) 86.89(16), P(2)–Pd(2)–C(45) 86.59(18), O(3)–
Pd(1)–O(5) 86.35(15), O(4)–Pd(2)–O(5) 88.14(16), P(1)–Pd(1)–O(3)
97.16(12), P(2)–Pd(2)–O(4) 95.83(12), C(19)–Pd(1)–O(5) 89.84(19),
C(45)–Pd(2)–O(5) 89.5(2).
8 R. J. Cross, L. J. Farrugia, P. D. Newman, R. D. Peacock and D. Stirling,
J. Chem. Soc. Dalton Trans., 1996, 4149.
9 R. G. Goel and W. O. Ogini, Organometallics, 1982, 1, 654.
CHEM. COMMUN., 2002, 2558–2559
2559