Zirconium-mediated synthesis of a new class of 1,4-bis(diphenylphosphino)-1,3-butadiene-bridged diphosphine, NUPHOS: Highly efficient catalysts for palladium-mediated cross couplings

Full text of DOI:10.1021/ja010090n

5110
J. Am. Chem. Soc. 2001, 123, 5110-5111
Scheme 1
Zirconium-Mediated Synthesis of a New Class of
1,4-Bis(diphenylphosphino)-1,3-butadiene-Bridged
Diphosphine, NUPHOS: Highly Efficient Catalysts
for Palladium-Mediated Cross Couplings
Simon Doherty,*,‡ Julian G. Knight,* Edward G. Robins,
Tom H. Scanlan, Paul A. Champkin, and William Clegg
Department of Chemistry
Bedson Building, Newcastle UniVersity
Newcastle upon Tyne, NE1 7RU, UK
ReceiVed January 10, 2001
Over the past two decades 2,2′-bis(diphenylphosphino)-1,1′-
binaphthyl (BINAP) has proven to be an extremely versatile
bidentate phosphine which, in combination with an appropriate
transition metal, forms catalysts that are highly active and selective
for a range of reactions.1 For reactions that do not involve an
asymmetric transformation it would be more practical to find an
alternative less expensive but structurally similar diphosphine.
One such diphosphine is 2,2′-bis(diphenylphosphino)-1,1′-biphen-
yl (dpbp).2 This ligand has several features in common with
BINAP in that it is a bidentate triaryl phosphine with two
diphenylphosphino groups bridged by a four-carbon tether,
composed of sp2-hybridized atoms, and forms a skewed seven-
membered ring when coordinated to a metal center.
conium to copper are commonplace,5 the application of this
procedure to the synthesis of bidentate phosphines has not yet
been reported. Herein, we report details of the synthesis of a new
class6 of 1,4-bis(diphenylphosphino)-1,3-butadiene-bridged diphos-
phine (NUPHOS) from the corresponding zirconacyclopentadiene.
Results of preliminary catalyst testing have demonstrated that
palladium complexes of 1,4-bis(diphenylphosphino)-1,2,3,4-tetra-
phenyl-1,3-butadiene (1,2,3,4-Ph4-NUPHOS) are an order of
magnitude more active for cross couplings than any catalyst
system previously reported. One of the most attractive features
of the chemistry described herein is the scope for structural
modification, since the basic building block is an alkyne or diyne.
On the basis of Fagan’s synthesis of heterocycles via the
reductive coupling of alkynes and diynes, we reasoned that
treatment of zirconacycle 1 with diphenylchlorophosphine would
liberate the corresponding diphosphine, 3, directly. Unfortunately,
we have been unable to isolate any phosphorus-containing species
from this reaction. Remarkably though, transmetalation of zir-
conacyclopentadiene 1 with copper(I) chloride prior to the addition
of diphenylchlorophosphine gave 1,4-bis(diphenylphosphino)-
1,2,3,4-tetraphenyl-1,3-butadiene copper complex, which can be
liberated by repeated extraction of a dichloromethane solution
with aqueous ammonia (Scheme 1). Crystallization from toluene-
hexane gave 1,4-bis(diphenylphosphino)-1,2,3,4-tetraphenyl-
1,3-butadiene (1,2,3,4-Ph4-NUPHOS, 3a) as colorless prisms.7
Similarly, when but-2-yne is used in the reductive coupling step,
the product 1,4-bis(diphenylphosphino)-1,2,3,4-tetramethyl-1,3-
butadiene (1,2,3,4-Me4-NUPHOS, 3b) is isolated in 53% yield.
Alternatively, 3a can be prepared by transmetalation of 1,4-
dilithiotetraphenylbutadiene8 with copper(I) chloride followed by
quenching with diphenylchlorophosphine. The use of zircona-
cyclopentadienes to prepare C4-bridged diphosphines offers
several attractive features including: (i) a one-pot straightforward
synthesis, (ii) variability since a wide range of alkynes and diynes
are known to undergo reductive coupling by Negishi’s reagent,
and (iii) access to a range of chiral diphosphines.
Hayashi and co-workers have recently reported the first results
of catalyst testing using dpbp.2 They showed that the selectivity
of palladium-based catalysts in the cross coupling of sec-BuMgCl
with vinyl halides was higher than that obtained with more
common diphosphines such as dppe and comparable to 1,1′-bis-
(diphenylphosphino)ferrocene (dppf)-based catalysts.3 This ligand
has also been shown to form highly efficient palladium catalysts
for the amination of aryl bromides and the rhodium-catalyzed
conjugate addition of boronic acid to enones. Given the effective-
ness of catalysts based on dpbp and their obvious potential in a
range of platinum group-catalyzed reactions there is likely to be
considerable interest in the preparation and applications of related
diphosphines, especially if their synthesis is straightforward and
versatile.
Several years ago Fagan and co-workers demonstrated that
1-phenylphospholes could be liberated directly from the corre-
sponding zirconacycle using dichlorophenylphosphine.4 We have
been investigating the synthesis of bidentate diphosphines by this
route and have found that (i) reaction of zirconacycle 1 with
chlorodiphenylphosphine does not give the corresponding diphos-
phine and (ii) diphosphines can be prepared by transmetalation
of the zirconacycle to copper before quenching with chlorodi-
phenylphosphine. While examples of transmetalation from zir-
The reaction of 3a-b with [(COD)PdCl2] in dichloromethane
affords [(NUPHOS)PdCl2] (4a-b). An X-ray analysis9 of 4a
(Figure 1) clearly shows that the palladium atom is distorted away
from square planar, as indicated by the dihedral angle of 12.1°
between the planes containing Pd(1), P(1), P(2) and Pd(1), Cl(1),
‡ Present address: The School of Chemistry, David Keir Building,
Stranmillis Road, The Queen’s University of Belfast, Belfast, Northern Ireland,
BT9 5AG.
(1) (a) Noyori, R.; Takaya, H. Acc. Chem Res. 1990, 23, 345. (b) Miyashita,
A.; Yasuda, A.; Takaya, K.; Torumi, T.; Ita, T. Sauchi, T.; Noyori, R. J. Am.
Chem. Soc. 1980, 102, 7933. (c) Cho, S. Y.; Shibasaki, M. Tetrahedron Lett.
1998, 39, 1773.
(5) Takahashi, T.; Sun, W.-H.; Nakajima, K. Chem. Commun. 1999, 1595.
(6) There has been a single report of a related four-carbon-bridged
diphosphine, 2,2′-bis(diphenylphosphino)-3,3,4,4,3′,3′,4′,4′-octafluoro-bicy-
clobutyl-1,1′-diene, isolated from the reaction between diphenylphosphine and
2,2′-dichlorooctafluoro(bi-1-cyclobuten-1-yl). For details see: Cullen, W. R.;
Williams, M. J. Fluorine Chem. 1979, 14, 85.
(2) Ogasawara, M.; Yoshida, K.; Hayashi, T. Organometallics 2000, 19,
1567.
(3) Hayashi, T.; Konishi, M.; Koboni, Y.; Kumada, M.; Higuchi, T.;
Hirotsu, K. J. Am. Chem. Soc. 1984, 106, 158.
(4) Fagan, P. J.; Nugent, W. A.; Calabrese, J. C. J. Am. Chem. Soc. 1994,
116, 1880.
(7) Full details of the single-crystal X-ray analysis of 3a are given in the
Supporting Information.
(8) Chisholm, M. H.; Jansen, R. M.; Huffman, J. C. Organometallics 1990,
9, 2305.
10.1021/ja010090n CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/04/2001

Communications to the Editor
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 5111
Interestingly, the activity of catalysts formed from NUPHOS-
type diphosphines depends markedly on the substitution pattern
of the C4-tether. Specifically, substitution of the phenyl groups
with methyl groups results in a dramatic decrease in catalyst
activity to 270 (mol product) (mol palladium)-1 h-1, although it
should be noted that the activity of catalysts formed from 3b is
comparable to values reported for most common bidentate
phosphines. A major problem associated with cross-coupling
reactions involving secondary Grignard reagents is isomerization
of the alkyl group: in the case of sec-butylmagnesium chloride
to give n-butyl derivatives via a â-hydride elimination pathway.
One of the most effective catalyst systems capable of achieving
high iso/n selectivity is [(dppf)PdCl2] reported by Hayashi. The
high selectivity was attributed to the large natural bite angle of
dppf and its concomitantly small Cl-Pd-Cl angle that accelerates
reductive elimination of the diorganopalladium intermediate. In
this regard, palladium catalysts based on 3a-b are also particu-
larly effective, the cross coupling of sec-butylmagnesium chloride
with bromobenzene giving exclusively sec-butylated products,
with none of the n-butyl product detected by GC. Notably though,
the natural bite angle formed by NUPHOS-based diphosphines
(∼ 92°) is significantly smaller than that of dppf.
Figure 1. Molecular structure of [(1,2,3,4-Ph4-NUPHOS)PdCl2] (4a).
Table 1. Cross-Coupling Reactions with sec-Butyl Magnesium
Chloridea
L2PdCl2
R-Br + sec-BuMgCl
8 sec-BuR
(1)
entry substrate (R-Br) diphosphine catalyst (mol %) TOFb GC yield %c
1
2
3
4
5
6
7
8
bromobenzene
bromobenzene
bromobenzene
bromobenzene
2-bromopropene
2-bromopropene
2-bromopropene
2-bromopropene
3a
3b
dppf
BINAP
3a
3b
dppf
BINAP
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
6900 100 (3h)
270 78 (17 h)
145 100 (22 h)
In a comparative study, catalysts formed from 3a and 3b also
show disparate activities toward 2-bromopropene, with initial TOF
of 2200 and 620 (mol product) (mol palladium)-1 h-1, respectively
(entries 5 and 6). The activity of 2200 (mol product) (mol
palladium)-1 h-1 for catalysts based on [(dppf)PdCl2] is similar
to that of 3a and somewhat unexpected, considering the poor
activity of the same system in the reaction between bromobenzene
and sec-butylmagnesium chloride (entry 3). Such high activities
have not previously been reported for dppf-based catalysts,
presumably because bromobenzene is the most common substrate
for screening new catalysts for cross-couplings.
260
2200
620
64 (22 h)
99 (2 h)d
65 (2 h)e
2200 100 (1 h)
220
61 (4 h)f
a Reaction conditions, 1.0 equiv of RBr, 2 equiv of sec-butylmag-
nesium chloride (2.0 mol dm-3), 0.1 mol % palladium, 0.5 equiv
n-decane internal standard, diethyl ether 1 mL/5 mmol bromide. b Initial
turn over frequency, determined by GC analysis. c Determined by
GC analysis of the reaction mixture, based on product. The reaction
time is given in parentheses. d 1% n-BuCMedCH2 detected. e 7%
n-BuCMedCH2 detected. f 6% n-BuCMedCH2 detected.
Kamer has clearly demonstrated a relationship between the
natural bite angle of Xanthphos-type diphosphines and their
activity and selectivity in palladium-catalyzed cross-couplings.
Both the rate and selectivity of the cross-coupling were found to
increase with increasing bite angle, reaching a maximum value
with DPEphos (102.7°).12 In contrast, NUPHOS-type diphos-
phines, which have much smaller natural bite angles, typically
close to 92°, form catalysts that are considerably more active. It
is clearly difficult to reconcile the differences between the activity
of catalysts formed from 3a, biaryl, and Xanthphos diphosphines
with bite angle effects alone, and a more credible explanation is
likely to involve a combination of factors.
Cl(2), presumably enforced by the torsional twist about C(2)-
C(3). Similar distortions have been reported for [(BINAP)PdCl2]10
and [(dpbp)PdCl2],2 which are based on structurally related four-
carbon atom tethers. The natural bite angle of 92.066(15)° is
comparable to that reported for [(BINAP)PdCl2] (92.69(8)°) and
[(dpbp)PdCl2] (92.24(4)°) and is characteristic of diphosphines
bridged by a four-carbon atom sp2-hybridized tether. In contrast,
the natural bite angle of diphosphines with four-carbon atom sp3
tethers is generally much larger and close to 100°.
In summary, we have shown that a new range of potentially
versatile C4-bridged diphosphines can be prepared directly from
zirconacycles. Results of preliminary catalyst testing are extremely
encouraging, and it is clear that this new class of diphosphine
may be effective in a host of platinum group-catalyzed reactions.
Further studies are underway to understand the origin of the high
catalyst activity achieved with NUPHOS-type diphosphines, to
establish the scope of this synthetic protocol, and to exploit the
application of these ligands in catalysis, in particular, the
development of chiral versions and their application in asymmetric
transformations.
Preliminary results have revealed that catalysts formed from
[(NUPHOS)PdCl2] are highly active and selective for the cross-
coupling of sec-butylmagnesium chloride with aryl and alkenyl
bromides, with TOF superior to those obtained with catalysts
based on BINAP and dppf (Table 1). For example, the reaction
of sec-butylmagnesium chloride with bromobenzene, carried out
in the presence of 0.1 mol % of [(1,2,3,4-Ph4-NUPHOS)PdCl2]
(4a) in Et2O, begins to reflux within 5 min, and reaction is
complete within 30 min (entry 1). Under the same conditions,
catalysts based on BINAP and dppf are significantly less active,
the reaction requiring 5 and 7 h, respectively, to reach completion.
The initial TOF of 6900 (mol product) (mol palladium)-1 h-1 for
catalysts based on 3a is an order of magnitude greater than the
initial TOF of 145 and 260 (mol product) (mol palladium)-1 h-1
for catalysts based on dppf and BINAP, respectively. For
comparison, Hayashi reported that coupling of bromobenzene and
sec-butylmagnesium chloride, catalyzed by 1 mol % of [(dpbp)-
PdCl2], requires 6 h to reach 93% completion.
Acknowledgment. We gratefully acknowledge the EPSRC for funding
(E.G.R. and P.A.C.) and Johnson Matthey for loans of palladium salts.
Supporting Information Available: Synthetic procedures and char-
acterization data for compounds 3a,b and 4a,b and tables of crystal data,
structure solution and refinement, atomic coordinates, bond lengths and
angles, and anisotropic thermal parameters for 3a and 4a and catalyst
testing procedures (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org. Observed and calculated structure factor
tables are available from the authors upon request.
(9) Crystal data for 4a: C53H42Cl4P2Pd, M ) 989.01, monoclinic space
group P21/n, a ) 12.7804(4) Å, b ) 22.2135(6) Å, c ) 16.3605(5) Å, â )
101.657(2)°, V ) 4548.9(2) Å3, Z ) 4, Dc ) 1.444 g cm-3, µ(Mo KR) )
0.750 mm-1, F(000) ) 2016, T ) 160(2) K, 10900 independent reflections;
R1 ) 0.0252, wR2 ) 0.0651.
(10) Ozawa, F.; Kubo, A.; Matsumoto, Y.; Hayashi, T.; Nishioka, E.;
Yanagi, K.; Moriguchi, K. Organometallics 1993, 12, 4188.
JA010090N
(11) Steffen, W. L.; Palenik, G. J. Inorg. Chem. 1976, 15, 2432.
(12) Leeuwen, P. W. N. M.; Kamer, P. C. J.; Reek, J. N. H.; Dierkes, P.
Chem. ReV. 2000, 100, 2741.