The preparation of trans-2,5- dialkylpyrrolidinylbenzyldiphenylphosphines: New phosphinamine ligands for asymmetric catalysis

Article abstract of DOI:10.1016/S0957-4166(98)00393-0

The preparation of new phosphinamine ligands 3a, 3b, possessing an enantiopure trans-2,5-dialkylpyrrolidinyl unit is described. Of the three synthetic approaches investigated, two proved successful with similar overall yields of 50-60%. One approach forms

Full text of DOI:10.1016/S0957-4166(98)00393-0

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 9 (1998) 3831–3839
The preparation of trans-2,5-dialkylpyrrolidinylbenzyl-
diphenylphosphines: new phosphinamine ligands for asymmetric
catalysis
John P. Cahill,a Frank M. Bohnen,b Richard Goddard,b Carl Krüger b and Patrick J. Guiry a,∗
aDepartment of Chemistry, University College Dublin, Belfield, Dublin 4, Ireland
bMax-Planck-Institut für Kohlenforschung, Kaiser-Wilhelm-Platz 1, D-45470, Mülheim an der Ruhr, Germany
Received 1 September 1998; accepted 25 September 1998
Abstract
The preparation of new phosphinamine ligands 3a, 3b, possessing an enantiopure trans-2,5-dialkylpyrrolidinyl
unit is described. Of the three synthetic approaches investigated, two proved successful with similar overall yields
of 50–60%. One approach forms the trans-2,5-dialkylpyrrolidine unit first and then the diphenylphosphine is intro-
duced whilst the other reverses the order of introduction. In each approach the key step is the cyclocondensation of
an amine with an enantiopure, dialkyl-substituted 1,4-diol cyclic sulfate. An X-ray crystal structure of the palladium
dichloride complex 13 of ligand 3a is presented. © 1998 Elsevier Science Ltd. All rights reserved.
1. Introduction
The broad range of catalytic asymmetric transformations to which transition metal complexes of chiral
phosphinamine ligands may be applied reflects their complexity as an effective class of ligand system.1–3
In contrast to diphosphines and diamines, the asymmetry induced by such heterobidentate systems is
determined by a combination of steric and electronic interactions. In the case of phosphinamines, Aker-
mark obtained spectroscopic evidence for the different electronic properties of the donor atoms in a study
of phosphinamine palladium allyl complexes.4 This concept of electronic disparity was subsequently
used in the design of chiral phosphinamine ligands possessing different chiral elements and backbone
scaffolding. Of the wide range of phosphinamine ligands prepared to date the amino alcohol derived di-
phenylphosphinooxazolines 1, reported independently by the groups of Pfaltz, Williams and Helmchen,5
have been one of the most successful, affording excellent enantioselectivities for a number of reactions.
The effect of having a more basic nitrogen donor atom in a phosphinamine ligand was investigated
∗
Corresponding author. E-mail: patrick.guiry@ucd.ie
0957-4166/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(98)00393-0

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3832
J. P. Cahill et al. / Tetrahedron: Asymmetry 9 (1998) 3831–3839
by Koga, who prepared diphenylphosphinopyrrolidines of type 2.6 Poor enantioselectivities of 11–20%
were obtained using palladium complexes of 2 in the test reaction between sodium dimethylmalonate
and 1,3-diphenylpropenyl acetate, whereas 99% ee was obtained using ligand 1 (R=Bn).5a We were
also interested in phosphinamine ligands possessing the asymmetry inducing, chiral 2,5-disubstituted
pyrrolidine unit.7,8 Therefore, we now report the preparation of ligands of type 3, which possess a more
rigid backbone and in which, upon complexation, would increase the chelate ring size from five to six
when compared to complexes of 2.9
2. Results and discussion
The preparation of 2,5-disubstituted pyrrolidines has recently been reviewed by Figadère, and of the
approaches described,10 the transamination of 1,4-dihydroxy derivatives appeared to be the most attrac-
tive to our needs. Enantiopure 1,4-dimesylates were used by Masamune and Chong to prepare trans-2,5-
dimethyl- and diphenyl-pyrrolidines, respectively,11,12 and by McGrath to prepare trans-2,5-dimethyl-
and diphenyl-pyridinylpyrrolidines.13 1,4-Diacetates were used by Knochel to synthesise trans-2,5-
diferrocenyl pyrrolidines. 1,4-Diol cyclic sulfates were exploited in heterocyclisations most notably
by Burk in the preparation of his DuPhos ligands.14 These dihydroxy derivatives were also used as
pyrrolidine precursors as evidenced by the Machinaga synthesis of trans-2-butyl-5-pentylpyrrolidine,15
an ant venom alkaloid, and by Gallagher in a recent intramolecular example.16 Such literature precedent
prompted our proposed synthesis of 3a, 3b, where the key step is the cyclocondensation of an amine with
enantiopure 2,5-dimethyl- or 2,5-diethyl-1,4-diol cyclic sulfate.17 This differs from Koga’s approach to
prepare 2, as he used a preformed trans-2,5-disubstituted pyrrolidine which was subsequently joined to
the backbone.
The first synthetic route to this class of ligand investigated was the condensation of benzylamine
4 with 2,5-hexanediol cyclic sulfate 5 which afforded the corresponding benzylpyrrolidine 6 in 60%
yield (Scheme 1). The next step in this strategy relied on this benzylic tertiary amine acting as
an ortho-lithiating group for the introduction of the diphenylphosphino group.18 However, despite
numerous attempts, this reaction did not proceed with chlorodiphenylphosphine or indeed a range of
other electrophiles. Hence a different approach to the desired ligand system was required involving the
starting amine, possessing a bromine atom in the 2-position for the proposed lithium–halogen exchange
prior to reaction with chlorodiphenylphosphine. Therefore, 2-bromobenzylamine 7 was reacted with the
(2S,5S)-hexanediol cyclic sulfate 5 in THF for 24 h, after which time a salt, presumed to be a zwitterion,
had precipitated. Subsequent reaction with NaH in THF for a further 24 h afforded the (2R,5R)-1-[(2-
bromophenyl)methyl]-2,5-dimethylpyrrolidine 9 in 67% yield. Reaction of 9 in diethyl ether at −10°C
with n-butyllithium for 30 min and then at ambient temperature for 2 h gave a white precipitate which
was then treated with chlorodiphenylphosphine to afford ligand 3a in 92% yield. The ethyl analogue 3b
was similarly prepared in 89% yield starting from pyrrolidine 10, which was synthesised from (3S,7S)-
octanediol cyclic sulfate 8 and 2-bromobenzylamine 7 in 58% yield.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
J. P. Cahill et al. / Tetrahedron: Asymmetry 9 (1998) 3831–3839
3833
Scheme 1.
Although this second route was successful, we also investigated a third pathway in which the
diphenylphosphino group was already in place prior to the cyclocondensation about the benzylic primary
amine. Therefore, we prepared 2-diphenylphosphinobenzonitrile by reduction of the commercially
available 2-bromobenzonitrile, an approach identical to that used by the group of Seebach.19 The nitrile
11 was reduced with lithium aluminium hydride to 2-diphenylphosphinobenzylamine 12 in 96% yield
(Scheme 2). Subsequent pyrrolidine ring formation after refluxing in THF and treatment with sodium
hydride afforded ligand 3a in a yield of 52%. Interestingly, no salt precipitation occurred during the
heterocyclisation process. Since there is no significant difference in the overall yield obtained from both
successful approaches, either may be used, although the latter, convergent route is to be favoured.
Scheme 2.
Once prepared, we wished to test the chelating ability of these ligands prior to testing them for
their enantiodifferentiating ability in catalysis. As phosphinamines have proved successful in palladium-
mediated carbon–carbon bond forming processes we first prepared a palladium dichloride complex.
Using standard chemistry, one equivalent of ligand 3a was reacted with bis-benzonitrile palladium
dichloride in methylene chloride to afford the palladium dichloride complex 13 in 59% yield (Scheme 3).
Scheme 3.
Once bound, the rotation about the benzylic carbon to nitrogen bond is no longer possible and this is
clearly seen by comparing the 1H NMR spectra of ligand 3a and the palladium dichloride complex 13. In
3a, the methyl groups at the 2,5-positions resonate at 0.90 ppm whilst in 13, one methyl appears at 1.52
and the other at 1.75 ppm. In addition, the methine protons (H-2, H-5) in 3a appear as a multiplet at 2.86
ppm and in 13, resonate at 3.04 ppm for H-5 and 5.47 ppm for H-2, indicating that the latter proton is in
a highly deshielded region of space, presumably in close proximity to the metal atom.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3834
J. P. Cahill et al. / Tetrahedron: Asymmetry 9 (1998) 3831–3839
Figure 1. Crystal structure of the Pd dichloride complex 13. Selected distances (Å) and angles (°): Pd–N 2.138(3), Pd–P
2.237(1), Pd–Cl1 2.286(1), Pd–Cl2 2.370(1), P–Pd–N 92.5(1), Cl1–Pd–Cl2 86.51(4), P–Pd–Cl1 86.21(4), N–Pd–Cl2 94.8(1)
Figure 2. X-Ray crystal structure (side view) highlighting the pyrrolidine ring conformation and the positions of the C24 and
C25 methyl groups
Crystals suitable for X-ray structure determination were grown by isothermal diffusion from methylene
chloride/diethyl ether. The results of this analysis are summarised in Figs. 1 and 2. The coordination of
the phosphinopyrrolidine ligand to the square-planar coordinated Pd atom appears to be relatively strain-
free since the bite angle at the metal atom is close to 90° [92.5(1)°]. There is, however, a clear trans-effect
and the Pd–Cl1 distance of 2.286(1) Å is significantly shorter than the Pd–Cl2 distance of 2.370(1) Å.
As a result of coordination to the metal, the phosphinopyrrolidine ligand 3a adopts a conformation
such that C2 lies approximately in the mean coordination plane of the metal atom (−0.13 Å), with C19
above (+0.74 Å) and C1 below (−0.58 Å) this plane (Fig. 2). A consequence of this geometry is that one
of the methyl groups of the substituted pyrrolidine ring takes up a position in close proximity to the metal
[C24. . .Pd 3.187(5) Å] (sum of van der Waals radii 3.1 Å).
Work is currently in progress investigating the catalytic behaviour of complexes of the ligands 3a
and 3b. Some of the initial results (e.g. in the palladium catalysed asymmetric allylic substitution and
Heck reactions) are promising. These and other applications, in addition to the preparation of analogous
phosphinamine ligands, will form the subject of future publications from this group.20