Synthesis of bulky, electron rich hemilabile phosphines and their application in the Suzuki coupling reaction of aryl chlorides

Article abstract of DOI:10.1039/b106784b

Novel electron rich, amine functionalised phosphines have been prepared and shown to belong to an unusual class of ligands that can activate palladium complexes to catalyse the Suzuki coupling reaction of chloroarenes.

Full text of DOI:10.1039/b106784b

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis of bulky, electron rich hemilabile phosphines and their
application in the Suzuki coupling reaction of aryl chlorides
Matthew L. Clarke, David J. Cole-Hamilton and J. Derek Woollins*
School of Chemistry, University of St. Andrews, St. Andrews, Fife, UK KY16 9ST.
E-mail: jdw3@st-andrews.ac.uk; Fax: 44-1334-463384
Received 27th July 2001, Accepted 5th September 2001
First published as an Advance Article on the web 17th September 2001
Novel electron rich, amine functionalised phosphines have
been prepared and shown to belong to an unusual class of
ligands that can activate palladium complexes to catalyse
the Suzuki coupling reaction of chloroarenes.
Fig. 1 Electronic contributions of amino and alkyl groups to phos-
phine basicity.
In recent years, the use of transition metal complexes of
electron rich phosphines has led to several important advances
in homogeneous catalysis, and they are therefore a topic of
ever growing interest.1 One particularly high profile example
has been the use of bulky electron rich ligands in the palladium
catalysed Suzuki cross-coupling reaction, a reaction of con-
siderable importance. These reactions are traditionally carried
out with aryl iodides, bromides or triflates as substrates. How-
ever, in 1998 Fu and Littke showed that tBu3P and palladium(0)
catalysts could, if used in a ratio of 1 : 1, be used to catalyse
Suzuki coupling of cheaper and more widely available aryl
chlorides (complete conversion at 80 ЊC using 1.5 mol% Pd2-
dba3ؒCHCl3, dba = dibenzylideneacetone).2 It was proposed
that the reaction proceeded through Pd complexes that only
contain one phosphine ligand, and are therefore very reactive
in both oxidative addition and transmetallation steps. A
more active catalytic system was simultaneously developed by
Buchwald and co-workers, who demonstrated that palladium
complexes of ligand 1 were particularly suited to this reaction.3
Further study revealed that the more readily synthesised ligand
2 could be used at least as effectively, even at room tempera-
ture.4 Another publication appearing at this time showed that
the P,O chelate ligand 3 could also catalyse the coupling of aryl
chlorides (1 mol% Pd2dba3ؒCHCl3, 105 ЊC).5 The latter three
ligands have three features in common. They are electron rich,
hemilabile, and bulky (ligand 2 could well form P∩arene–olefin
chelate complexes). The ligands have also shown utility in
Pd catalysed amination and etherifications of aryl chlorides.6
Catalysis of the Suzuki reaction continues to be an area of
intense interest as demonstrated by four recent publications
appearing after the completion of this work.10
two dialkylamino groups in the parent tris(dialkylamino)-
phosphines actually contributing to the basicity of the phos-
phine. The remaining amino group merely acts as an electron
withdrawing group. Alkyl- and dialkyl-phosphino-amines are
therefore especially electron rich phosphines (Fig. 1).7
Since a huge range of amines are very readily available, and
the P–N bond forming reactions are often quantitative under
mild conditions, we felt that electron rich, bulky hemilabile
phosphino-amines would make ideal ligand candidates for this
type of catalysis. We report here our preliminary results in this
area.
Ligand 4 was prepared by the addition of iPr2PCl to an Et2O
solution of triethylamine and N-methylpiperazine at room
temperature. Removal of Et3NHCl gave 4 as a moisture and air
sensitive liquid that is indefinitely stable under an atmosphere
of dry nitrogen. Thus it is possible to convert cheap and com-
mercially available starting materials into the pure ligand (as
determined by 31P, H, 13C NMR and HRMS) in about an
1
hour.12
It was found that Pd2dba3ؒCHCl3 and 4 (L : Pd = 2 : 1) cata-
lyse the Suzuki coupling of the relatively electron poor aryl
chlorides with PhB(OH)2 at 90 ЊC to give complete conversion
to the desired biaryls with no unwanted side products (Table 1,
entries 1–5).13
We believe that bulky electron rich hemilabile ligands will rise
still further in importance as they can simultaneously activate
metal complexes to oxidative addition, transmetallation/
insertion and reductive elimination. There are many existing
reactions that could be improved/expanded by the use of such
ligands, and probably an equally large number of reactions that
will not proceed with normal ligands, yet await discovery if
more highly reactive metal catalysts can be discovered. We have
recently proposed that the electron donating character of
phosphino-amines has been underestimated as a result of only
Although the reactions were generally run overnight for
convenience, it was found that in the case of m-nitrochloro-
benzene the reaction was complete within six hours. Despite
several attempts under various conditions, the reaction of
electron rich p-chlorotoluene (and chlorobenzene) (using
Pd2dba3ؒCHCl3/4 catalyst) always stopped prior to complete
conversion (50–70% yield). As it is known that electron rich
chloroarenes are more reluctant to undergo oxidative addition
to palladium(0) complexes, we presume that this is the problem-
atic step for the 4/Pd catalyst system. A sluggish oxidative
DOI: 10.1039/b106784b
J. Chem. Soc., Dalton Trans., 2001, 2721–2723
This journal is © The Royal Society of Chemistry 2001
2721

View Article Online
Table 1 Palladium catalysed Suzuki coupling of phenylboronic acid
with a selection of aryl chlorides
stage) a less reactive catalytic system than that developed by
Buchwald and co-workers (using ligand 2).
In summary, we have prepared two new electron rich,11
potentially hemilabile ligands, and shown them to belong to
a rare class of ligands that catalyse the Suzuki reaction of
aryl chlorides. The evidence suggests that the presence of a
potentially donating amine substituent is critical to their
performance, and also suggests that a cyclohexyl-substituted
phosphine may be more suited than isopropyl phosphines as a
ligand. This study represents the first time that phosphino-
amines have been utilised in the Suzuki coupling reaction,
and given that derivatives of these ligands should be easily
accessible, should allow us to optimise catalytic performance of
this and other reactions. An investigation into the co-ordination
chemistry that is diagnostic of a successful catalyst is also
underway.
Entry
Liganda
ArЈ
Base
Conversionbc (%)
1
2
3
4
5
6
7
8
9
4
4
4
4
4
5
8
9
9
9
9
p-F3CC6H4
m-NO2C6H4
p-NCC6H4
p-MeC6H4
p-MeC6H4
p-MeC6H4
p-MeC6H4
p-MeC6H4
K3PO4
K3PO4
K3PO4
K3PO4
CsF
CsF
CsF
CsF
CsF
CsF
CsF
100
100
100
d
45–60 (6 runs)
ca. 70
0
0
93
100
100
100
e
p-CH3COC6H4
m-CHOC6H4
p-F3CC6H4
e
10
11
f
a Reactions were carried out using 1 mol% Pd2dba3ؒCHCl3, with L : Pd
ratio of 2 : 1, toluene, 90 ЊC, 16 hours. b Conversions calculated by
GCMS, using napthalene as an internal standard. c Some of the reac-
tions contained ≈2% of biphenyl isomers. d Reaction time = 6 h. e 0.5
mol% catalyst. f 0.2 mol% catalyst.
Acknowledgements
The authors wish to thank the EPRSC for financial support
(M. L. C.), Johnson Matthey for loan of palladium() chloride,
the EPRSC mass spectrometry service (Swansea), and the
JREI for equipment grants. The authors also wish to thank
Dr D. F. Foster for performing some of the GC analysis.
addition reaction would result in the catalyst spending more
time as a zerovalent palladium complex, which would reduce
catalyst stability. This is somewhat surprising as Buchwald and
co-workers have shown that the transmetallation is the rate
determining step using their ligands.
We have also prepared ligand 5, which has greater conform-
ational freedom than 4. Somewhat to our surprise, a com-
bination of 5 and Pd2dba3ؒCHCl3 does not catalyse the Suzuki
reaction of p-chlorotoluene at all. It is likely that a difference in
co-ordination chemistry for the two ligands accounts for this
lack of catalytic activity (Table 1, entry 6).
Notes and references
1 (a) M. C. Simpson and D. J. Cole-Hamilton, Coord. Chem. Rev.,
1996, 155, 163 and refs. therein; (b) T. Sakakura, T. Sodeyama,
K. Sasaki, K. Wada and M. Tanaka, J. Am. Chem. Soc., 1990, 112,
7221 and refs. therein; (c) J. K. MacDougall, M. C. Simpson,
M. J. Green and D. J. Cole-Hamilton, J. Chem. Soc., Dalton
Trans., 1996, 1161 and refs. therein; (d ) Y. Ben-David, M. Portnoy
and D. Milstein, J. Chem. Soc., Chem. Commun., 1989, 1816;
(e) M. Portnoy and D. Milstein, Organometallics, 1993, 12, 1665.
2 A. F. Littke and G. C. Fu, Angew. Chem., Int. Ed., 1998, 37,
3387.
At this stage, we have not compared the co-ordination
chemistry of 4 and 5 to any great extent. Preliminary experi-
ments suggest that the difference between the ligands is rather
subtle; ligands 4 and 5 were further characterised by conversion
to trans-(η1-L)2PtCl2 complexes (6 and 7 respectively) by
reaction with (PhCN)2PtCl2 or Zeise’s salt in a 2 : 1 ratio. In
both compounds the 31P NMR spectra are sharp and show no
sign of the fluxionality which would occur if the ligand were
forming a transient chelate species in solution. When one
equivalent of ligand 4 was added to (COD)PtCl2 (COD = cyclo-
octa-1,5-diene), a complex mixture (containing 6 and lots of
unidentified products) results. Reaction of 5 with (COD)PtCl2,
also does not occur cleanly. However, the major species can be
identified as [η2-5]PtCl2.8 Thus it is possible that an excessively
strong chelate may account for the lack of catalytic activity
when using this ligand. Burrows and co-workers have prepared
related ligand N-diphenylphosphino-NЈ,NЈ,NЈ-trimethyl-
ethylenediamine and shown that it readily forms chelates with
platinum.9
To further establish the requirements for a successful ligand
in this reaction, we have prepared and tested N-diisopropyl-
phosphinopiperidine, 8, which presumably shares steric and
electronic properties with 4, but lacks the auxiliary nitrogen
atom. This ligand was tested in the Pd catalysed Suzuki reaction
of chlorotoluene under identical conditions to those used for 4.
Ligand 8 only gives a conversion of 30% after 20 hours reaction
time. This is further evidence that the potentially donating
nitrogen atom in 4 plays some role in its catalytic properties.
N-Dicyclohexylphosphino-N-methylpiperazine, 9 could also
be prepared by the same general method in excellent purity, and
was tested (in combination with Pd2dba3ؒCHCl3) as a catalyst
for the Suzuki coupling of phenylboronic acid and p-chloro-
toluene. It was pleasing to find that this ligand is superior
to N-diisopropylphosphino-N-methylpiperazine, 4 and that
essentially complete conversion into the desired product could
be achieved (Table 1, entry 8). This catalyst system could also
be used to couple p-chloroacetophenone with phenylboronic
acid using 0.5 mol% catalyst and p-chlorobenzotrifluoride
using 0.2 mol% catalyst (Table 1, entries 9–11). However, if the
reactions were run at lower temperatures (50 ЊC) conversions
were always less than 50% which suggests that this is (at this
3 D. Old, J. P. Wolfe and S. L. Buchwald, J. Am. Chem. Soc., 1998, 120,
9722.
4 J. P. Wolfe, R. A. Singer, B. H. Yang and S. L. Buchald, J. Am. Chem.
Soc., 1999, 121, 9550.
5 X. Bei, H. Turner, W. H. Weinberg and A. S. Gurran, J. Org. Chem.,
1999, 64, 6797.
6 (a) X. Bei, H. Turner, W. H. Weinberg and A. S. Gurran, Tetrahedron
Lett., 1999, 40, 1237; (b) A. Aranyos, D. W. Old, A. Kiyomori,
J. P. Wolfe, J. P. Sadighi and S. L. Buchwald, J. Am. Chem. Soc.,
1999, 121, 4369; (c) D. Old, J. P. Wolfe and S. L. Buchwald, J. Am.
Chem. Soc., 1998, 120, 9722.
7 (a) M. L. Clarke, D. J. Cole-Hamilton, A. M. Z. Slawin and
J. D. Woollins, Chem. Commun., 2000, 2065; (b) M. L. Clarke,
A. M. Z. Slawin and J. D. Woollins, unpublished work.
8 The structure of this chelate complex, [η2-P, N -5]PtCl2, has been
recently verified by an X-ray crystal structure determination;
M. L. Clarke, A. M. Z. Slawin and J. Derek Woollins, unpublished
work. Analytical data for [η2-P, N -5]PtCl2: Found: C, 27.63;
H, 5.63; N, 5.60. C11H27N2P1Pt1Cl2 requires C, 27.28; H, 5.62;
N, 5.78%.
9 A. D. Burrows, M. F. Mahon and M. T. Palmer, J. Chem. Soc.,
Dalton Trans., 2000, 3615.
10 (a) G. Y. Li, Angew. Chem., Int. Ed., 2001, 40, 1513; (b) S. Gibson,
G. R. Eastham, D. F. Foster, R. P. Tooze and D. J. Cole-Hamilton,
Chem. Commun., 2001, 779; (c) T. E. Pickett and C. J. Richards,
Tetrahedron Lett., 2001, 42, 3767; (d ) A. Furstner and A. Leitner,
Synlett, 2001, 290.
11 We have additionally prepared [trans-L2Rh(CO)Cl] complexes from
ligands 4 and 9. We were surprised to find the position of ν(CO)
for these compounds to be at 1960 cmϪ1 [for trans-(4)2Rh(CO)Cl]
and 1964 cmϪ1 [for trans-(9)2Rh(CO)Cl]. This suggests that these
two ligands are not especially strong donor ligands.7 Our more
recent work suggests that bulky alkyl groups can diminish the N
P
donation, and reduce phosphine basicity.7b.
12 All ligands were characterised by multinuclear NMR, IR and (high
resolution) mass spectroscopies. Their structure was further con-
firmed by the preparation of the platinum and rhodium complexes
described, which gave satisfactory chemical analyses. Selected
31P NMR and analytical data (CDCl3; chemical shift in ppm relative
to external phosphoric acid) for compounds discussed here.
4 δP 82.8; 5 δP 86.3; 6 δP 77.3 (1JP–Pt = 2676 Hz). Found: C, 38.05;
2722
J. Chem. Soc., Dalton Trans., 2001, 2721–2723

View Article Online
H, 7.16; N, 7.86; C22H50N4P2Pt1Cl2 requires C, 37.82; H, 7.21;
N, 8.02%. 7 δP 82.2 (1JP–Pt = 2641 Hz). Found: C, 38.20; H, 8.29;
N, 7.80; C22H54N4P2Pt1Cl2 requires C, 37.61; H, 7.75; N, 7.97%. 8 δP
84.5. 9 δP 75.8. trans-(4)2Rh(CO)Cl: δP 102.1 (1JP–Rh = 133 Hz).
Found: C, 46.12; H, 8.41; N, 9.35. C23H50P2N4O1Rh1Cl1 requires C,
46.37; H, 8.76; N, 9.36%.
to a second Schlenk flask containing Pd2dba3ؒCHCl3 (31 mg,
0.030 mmol, 1 mol%), napthalene internal standard and CsF (1.39 g,
9.12 mmol, 3 equiv.). Chlorotoluene was then added by syringe
(0.385 g, 3.04 mmol). A sample for GC analysis was taken at this
time before introduction of phenylboronic acid (0.556 g, 4.55 mmol,
1.5 equiv.). The reactions were sampled periodically by removing
0.05 mL by syringe, and analysing by GC. The GC was set at a ramp
program from 60 to 250 ЊC. Products can be readily separated
from the more volatile aryl chlorides with retention times between
4 and 20 minutes.
13 Representative procedure for Suzuki coupling reactions: To a
Schlenk flask containing ligand 9 (36 mg, 0.121 mmol, 4 mol%)
under a nitrogen atmosphere was added dry toluene (10 mL). This
was then degassed by two freeze–thaw cycles. This was then added
J. Chem. Soc., Dalton Trans., 2001, 2721–2723
2723