Simple mixed tricyclohexylphosphane-triarylphosphite complexes as extremely high-activity catalysts for the Suzuki coupling of aryl chlorides

Article abstract of DOI:10.1002/1521-3773(20021104)41:21<4120::AID-ANIE4120>3.0.CO;2-7

The most active catalyst yet reported for the Suzuki coupling of activated aryl chlorides, even for electronically deactivated substrates, is generated by addition of tricyclohexylphosphane to the otherwise inactive orthopalladated complex 1 [Eq. (1); for

Full text of DOI:10.1002/1521-3773(20021104)41:21<4120::AID-ANIE4120>3.0.CO;2-7

COMMUNICATIONS
The initial coupling studied was that of the deactivated
substrate 4-chloroanisole with phenylboronic acid using the
catalyst formed in situfrom PCy 3 and the dimeric complex 2a.
As can be seen from the results summarized in Table 1, good
to excellent conversions are obtained at 0.01 mol% Pd
loadings with K3PO4, K2CO3, KF, or a mixture of K3PO4/KF
acting as base (Table 1, entries 2 5, respectively). When cesium
carbonate is used then even higher activity results. Under these
conditions astonishingly high turnover numbers (TONs) of up
to 34000 are seen (Table 1, entries 7 and 8). By comparison, the
previous highest reported TONs observed in this reaction were
8000, which was obtained under the same conditions using
complex 1 as catalyst, and 12800 when the highly sterically
encumbered ligand PBu(Ad)2 (Ad¼ adamantyl) was employ-
ed.[2c,e] Increasing the PCy3:Pd ratio from 1:1 to 2:1 led to a
decrease in catalyst performance. When the orthopalladated
phosphite precursor 2a was replaced by the analogous
phosphinite-containing complexes 2b and then c[5] a slight
decrease in activity was observed with decreasing p-acidity of
the orthopalladated ligand (Table 1, entries 11 and 12).
These results clearly demonstrate that complexes of the
simple, inexpensive, easily handled PCy3 ligand can show
excellent activity, providing that the correct choice of
palladium precursor is made.[2c,f] ™Classical∫ palladium pre-
cursors such as palladium acetate or palladium dibenzylide-
neacetone are generally very poor in this regard.[2f] Indeed we
find that when the coupling is performed with an equimolar
mixture of palladium acetate, ligand 3, and PCy3, substantially
lower activity results (Table 1, entry 13). The use of the
preformed complex 4, which is readily synthesized by reaction
of 2a with an excess of PCy3 (Scheme 2),[6] gave no advantage
over the catalyst formed in situ.
Simple Mixed Tricyclohexylphosphane
Triarylphosphite Complexes as Extremely
High-Activity Catalysts for the Suzuki Coupling
of Aryl Chlorides**
Robin B. Bedford,* Catherine S. J. Cazin, and
Samantha L. Hazelwood (nÿeWelch)
The coupling of aryl halides with aryl boronic acids, the
Suzuki reaction (Scheme 1), is one of the most powerful and
versatile methods for the synthesis of biaryls.[1] There has
[cat]
X
B(OH)2
+
base
R'n
Rn
Rn
R'n
Scheme 1. The Suzuki coupling of aryl halides.
recently been considerable interest in the development of new
catalysts that can couple aryl chlorides because of the lower
cost and greater availability of these substrates compared with
their bromide or iodide counterparts.[2] Unfortunately the
ꢀ
comparatively high C Cl bond strength makes aryl chlorides
difficult to activate. Consequently most catalysts that are able
to catalyze aryl chloride coupling reactions still need to be
used in relatively high loadings. Therefore the advantages
associated with the use of aryl chlorides may be negated by
the high cost of the catalyst systems.
We recently reported that complex 1, based on the
comparatively inexpensive ligands N,N-dimethylbenzylamine
and tricyclohexylphosphane, shows very high activities in
Suzuki coupling reactions of aryl chlorides at low catalyst
loadings.[2c] Previously we had shown that the orthopalladated
The catalyst formed from 2a and PCy3 shows an order of
magnitude higher activity than catalysts formed in situ from
palladium acetate and either PCy2(o-biphenyl) or PtBu3
(Table 1, entries 15 and 16), both of which have previously
been shown to be excellent catalysts for the Suzuki coupling
of aryl chlorides.[2h,d] In addition catalysts formed in situfrom
2a and these phosphanes were not as active as that formed
from PCy3 although they were substantially better than those
formed in situfrom palladium acetate.
It may be anticipated that a catalyst formed in situfrom
complex 1 and one equivalent of ligand 3 would show
essentially identical activity to that formed from 2a and
PCy3. Indeed an increase in conversion at 17 h is observed
compared to that of 1, but it is not as high as that observed
when the phosphite complex 2a is used as the palladium
precursor. Interestingly, it seems as if the added phosphite
ligand decreases the rate of conversion, as exemplified by the
results after two hours compared with those with 1, bu t
increases the catalyst longevity (Table 1, entries 19 22)–
catalyst 1 shows essentially no further activity after 2 h.
Next we investigated the couplings of electronically non-
activated and activated aryl chlorides (Table 1, entries 23 29)
and found that extremely high TONs of up to 1000000 were
observed. By contrast, to the best of our knowledge, the
highest TONs reported for any Suzuki, or indeed any coupling
reaction with an aryl chloride substrate is 100000, obtained
with complex 1.[2c]
NMe2
tBu
tBu
O
PR2
2a: R = OC6H3-2,4-tBu2
b: R = Ph
c: R = iPr
Pd TFA
PCy3
Pd Cl
2
1
complex 2a, which is easily prepared from the very inex-
pensive ligand tris(2,4-di-tert-butylphenyl)phosphite (3)[3]
shows good activity in the Suzuki coupling of aryl bromide
substrates.[4] This latter complex shows no propensity to
couple even electronically activated aryl chlorides such as 4-
chloroacetophenone. Therefore we were highly surprised to
find that mixed PCy3 P(OAr)3 complexes either formed in
situor preformed show, to the best of our knowledge, the
highest activity reported to date in aryl chloride coupling
reactions. Preliminary findings of this study are reported
below.
[*] Dr. R. B. Bedford, C. S. J. Cazin, S. L. Hazelwood (nÿeWelch)
School of Chemistry
University of Exeter
Exeter, EX4 4QD (UK)
Fax : (þ 44)1392-263-434
E-mail: r.bedford@ex.ac.uk
[**] We gratefully acknowledge the EPSRC (studentship for S.L.H.) and
the University of Exeter (studentship for C.S.J.C.) for funding and
Johnson Matthey Chemicals Ltd for funding and the loan of palladium
salts.
4120
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4121-4120 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 21

COMMUNICATIONS
Table 1. Suzuki coupling of aryl chloride substrates.[a]
Entry
ArCl
ArB(OH)2
Catalyst (mol%)
Base
Conv. [%][b]
TON
mol product
per mol Pd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4 4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-4-OMe
C6H4-2-Me
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
2a (0.5) þ PCy3 (1.0)
K3PO4
K3PO4
K2CO3
KF
100
91
82
51
72
100
34
68
99
37
29
26
12
28
4
3
8
11.5
6[c]
6
3[c]
20
10
82
48
88
92
100
100
100
9100
8200
5100
7200
10000
34000
34000
33000
18500
29000
26000
120
28000
4000
3000
8000
11500
6000
2a (0.005) þ PCy3 (0.01)
2a (0.005) þ PCy3 (0.01)
2a (0.005) þ PCy3 (0.01)
2a (0.005) þ PCy3 (0.01)
2a (0.005) þ PCy3 (0.01)
2a (0.0005) þ PCy3 (0.001)
2a (0.001) þ PCy3 (0.002)
2a (0.0015) þ PCy3 (0.003)
2a (0.001) þ PCy3 (0.004)
2b (0.0005) þ PCy3 (0.001)
2c (0.0005) þ PCy3 (0.001)
Pd(OAc)2 (0.1) þ 3 (0.1) þ PCy3 (0.1)
4 (0.001)
K3PO4/KF (1:1)
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Pd(OAc)2 (0.001) þ PCy2(o-biphenyl) (0.002)
Pd(OAc)2 (0.001) þ PtBu3 (0.002)
2a (0.0005) þ PCy2(o-biphenyl) (0.001)
2a (0.0005) þ PtBu3 (0.001)
1 (0.001)
1 (0.001)
6000
3000
1 (0.001) þ 3 (0.001)
1 (0.001) þ 3 (0.001)
20000
100000
82000
480000
880000
920000
1000000
1000000
2a (0.00005) þ PCy3 (0.0001)
2a (0.0005) þ PCy3 (0.001)
2a (0.00005) þ PCy3 (0.0001)
2a (0.00005) þ PCy3 (0.0001)
2a (0.00005) þ PCy3 (0.0001)
2a (0.00005) þ PCy3 (0.0001)
2a (0.00005) þ PCy3 (0.0001)
C6H4-2-Me
C6H4-4-Me
C6H4-4-COMe
C6H4-4-NO2
C6H4-4-COMe
C6H4-4-NO2
Ph
Ph
Ph
C6H4-4-Me
C6H4-4-Me
[a] Reaction conditions: ArCl (10 mmol), ArB(OH)2 (15 mmol), base (20 mmol), 1,4-dioxane (30 mL), 1008C, N2, 17 h. [b] Conversion to Suzuki product,
based on aryl chloride determined by GC (hexadecane standard). [c] Reaction time ¼ 2 h.
tBu
tBu
O
P(OAr)2
a)
2a
Pd PCy3
Cl
4
Scheme 2. a) PCy3, 2 equiv per Pd, CH2Cl2, reflux 18 h.
Regardless of conditions, the mixed PCy3 P(OAr)3 systems
show higher activity than complex 1. Given that the rate-
determining step in the Suzuki coupling of aryl chloride
substrates is generally accepted to be oxidative addition, it is
not surprising that catalysts that usually show good activity
tend to have strongly electron-donating ligands that maximize
electron density on the palladium center.[7] Thus it seems
counter-intuitive that the inclusion of a p-acidic triarylphos-
phite ligand, which shows little or no tendency to facilitate the
coupling of even activated aryl chlorides,[8] should enhance
activity. Plots of conversion against time in the coupling of 4-
chloroanisole with phenylboronic acid catalyzed by 1 and the
catalyst formed in situfrom 2a and PCy3 are shown in
Figure 1. It is immediately apparent that the inclusion of the
triarylphosphite ligand does not increase the rate, it may even
lead to a slight decrease, instead it leads to greatly increased
catalyst longevity compared with complex 1. What is the basis
of this increased longevity? If it is assumed that the rate-
determining step is oxidative addition, then the catalyst
spends most of its time ™resting∫ in a zerovalent state.[9] This
™resting∫ state will be stabilized both thermodynamically and
Figure 1. Plots of conversion [%] versus time (t) in the coupling of 4-
chloroanisole (10 mmol) with phenylboronic acid (15 mmol) catalyzed by 1
~
&
(
) and 0.5 2a þ PCy3 ( ) (both at 0.001 mol% Pd). Reaction conditions:
Cs2CO3 (20 mmol), 1,4-dioxane (27 mL), hexadecane internal standard
(3.0 mL, 0.034m in 1,4-dioxane), 1008C, N2. Conversions determined by
GC.
kinetically by the reversible coordination of the p-acidic
triarylphosphite ligand. Decoordination of the phosphite
ligand would give a low-coordinate, electron-rich complex
readily able to oxidatively add the aryl chloride substrate and
thus enter the catalytic manifold.
In summary, catalysts formed from mixtures of the readily
available, easily handled and inexpensive ligands 3 and PCy3
show by far the highest activity yet reported in the Suzuki
coupling of aryl chlorides.
Received: June 19, 2002 [Z19561]
Angew. Chem. Int. Ed. 2002, 41, No. 21
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4121-4121 $ 20.00+.50/0
4121

COMMUNICATIONS
[1] For recent reviews see: a) N. Miyaura, A. Suzuki, Chem. Rev. 1995, 95,
2457 2483; b) S. P. Stanforth, Tetrahedron 1998, 54, 263 303; c) A.
Suzuki, J. Organomet. Chem. 1999, 576, 147 168.
prises molecules with planarizing distortion of the central
carbon atom. The magnitude of the distortion is dependent on
the size and configuration of the fused rings. Because the
parent, unsubstituted fenestranes are low molecular weight
hydrocarbons, they are not amenable to X-ray crystallo-
graphic analysis. The few X-ray structures on record are of
substituted and functionalized derivatives.[2]
We were intrigued by the possibility of replacing one of the
ring-fusion carbon atoms with a nitrogen atom to facilitate salt
formation and provide an opportunity for X-ray analysis of an
unsubstituted fenestrane. In addition to establishing the full
molecular structure and the extent of the central carbon
planarization, the pyramidal distortion of the nitrogen atom
would also be of interest; to our knowledge, no monoazafe-
nestranes have been prepared. An unusual tetraamino
[5.5.5.5]fenestrane is known and it exists as an equilibrium
mixture of (degenerate) open and closed forms when proto-
nated.[3] We describe herein the first synthesis of an unsub-
stituted 1-azafenestrane 1 along with the synthesis and X-ray
crystallographic analysis of the borane adduct 1¥BH3.
[2] Recent examples of Suzuki coupling reactions with aryl chlorides: a) L.
Botella, C. Nµjera, Angew. Chem. 2002, 114, 187 189; Angew. Chem.
Int. Ed. 2002, 41, 179 181; b) S.-Y. Liu, M. J. Choi, G. C. Fu, Chem.
Commun. 2001, 2408 2409; c) R. B. Bedford, C. S. J. Cazin, Chem.
Commun. 2001, 1540 1541; d) M. R. Netherton, G. C. Fu, Org. Lett.
2001, 3, 4295 4298; e) A. Zapf, A. Ehrentraut, M. Beller, Angew.
Chem. 2000, 112, 4315 4317; Angew. Chem. Int. Ed. 2000, 39, 4153
4155; f) M. G. Andreu, A. Zapf, M. Beller, Chem. Commun. 2000,
2475 2476; g) D. W. Old, J. P. Wolfe, S. L. Buchwald, J. Am. Chem.
Soc. 1998, 120, 9722 9723; h) J. P. Wolfe, R. A. Singer, B. H. Yang, S. L.
Buchwald, J. Am. Chem. Soc. 1999, 121, 9550 9561; i) X. Bei, H. W.
Turner, W. H. Weinberg, A. S. Guram, J. Org. Chem. 1999, 64, 6797
6803; j) C. Zhang, J. Huang, M. L. Trudell, S. P. Nolan, J. Org. Chem.
1999, 64, 3804 3805; k) A. F. Littke, G. C. Fu, Angew. Chem. 1998, 110,
3586 3587; Angew. Chem. Int. Ed. 1998, 37, 3387 3388.
[3] Ligand 3 is cheaper than triphenylphosphane (source: Aldrich Chem-
icals catalogue).
[4] D. A. Albisson, R. B. Bedford, S. E. Lawrence, P. N. Scully, Chem.
Commun. 1998, 2095 2096.
[5] R. B. Bedford, S. L. Welch, Chem. Commun. 2001, 129 130.
[6] Complex 4 could not be isolated pure, but rather contained small
amounts of dimer 2a, even when a large excess of PCy3 was used in the
synthesis. 31P NMR spectroscopy the confirmed cis disposition of P
donors (2JP, P ¼ 40 Hz).
H
B
H
H
[7] Fuand co-workers have recently reported that highly sterically
hindered triarylphosphanes can catalyze couplings of electronically
deactivated aryl chlorides (see ref. [2b]), indicating that more subtle
effects than brute donor-strength can play an important role.
[8] Beller and Zapf have shown that non-orthometalated complexes of
triarylphosphites can couple activated aryl chlorides: A. Zapf, M.
Beller, Chem. Eur. J. 2000, 6, 1830 1833.
N
N
a
d
b
c
a
b
c
H
H
H
H
d
H
1
H
1•BH3
[9] We are assuming that the catalytic cycle proceeds by a Pd0/PdII pathway.
This assumption is made on the basis of studies into the activation of
palladacyclic precatalysts in Suzuki coupling reactions. See reference
[2c] for leading references.
Analysis of the tetracyclic ring structure of 1 reveals that it
contains an embedded pyrrolizidine unit fused to a bicy-
clo[3.3.0]octane system. In recent years, a large number of
pyrrolizidine-based alkaloid natural products in the necine,
alexine and australine families have been synthesized in these
laboratories.[4] The key strategic operation in all of these
syntheses is the tandem [4þ2]/[3þ2] cycloaddition of nitro-
alkenes.[5] This process allows for the facile and stereocon-
trolled construction of highly functionalized nitroso acetals
that serve as precursors for pyrrolizidines upon catalytic
hydrogenolysis.
Synthesis of cis,cis,cis,cis-[5.5.5.5]-1-
Azafenestrane**
Scott E. Denmark,* Laurenz A. Kramps, and
Justin I. Montgomery
The application of the tandem cycloaddition strategy to the
synthesis of 1 is outlined in Scheme 1. Constructing the core of
1 requires the creation of one of the four rings in the tandem
[4þ2]/[3þ2] process by cycloaddition of a C2 dienophile (butyl
vinyl ether (3)) with a cyclopentenyl nitro diene 2 (ring C)
bearing the suitable dipolarophilic tether. The tetracyclic
nitroso acetal 4 is then poised for hydrogenolytic unmasking
to a tricyclic pyrrolizidine (third ring, c) which should undergo
spontaneous lactam formation (!5; fourth ring, d) from the
appended carboxylic ester. Thus, three of the four rings of
[5.5.5.5]-1-azafenestrane can be assembled in two chemical
manipulations. Two-stage reduction of the a-hydroxy lactam 5
leads to the target azafenestrane 1. This approach allows for a
modular synthesis of fenestranes containing rings of different
size at various positions. With regard to the configuration at
the ring fusions, only one relationship was expected to be
variable. In the [4þ2] process, the approach of the dienophile
can take place on the two diastereotopic faces of the
nitroalkene to create cis and trans isomers. The [3þ2] process
is formally in the spiro mode family[6] and thus is expected to
The structural theory of organic chemistry is one of the
most highly evolved constructs in natural science. The ability
to explain and correctly predict the detailed molecular
structure of millions of compounds naturally inspires research
to test the limits of the theory. One important subset of this
field of investigation probes the extent to which a tetracoor-
dinate carbon atom (bearing all carbon substituents) can
deviate from the van©t Hoff/Le Bel tetrahedral geometry. The
interesting family of compounds called fenestranes[1] com-
[*] Prof. Dr. S. E. Denmark, Dr. L. A. Kramps, J. I. Montgomery
Department of Chemistry
University of Illinois, Urbana, IL 61801 (USA)
Fax : (þ 1)217-333-3984
E-mail: denmark@scs.uiuc.edu.
[**] We are grateful to the National Institutes of Health for generous
financial support (R01 GM30938). J.I.M. thanks the Procter and
Gamble Company for a graduate fellowship.
Supporting information for this article is available on the WWW under
http://www.angewandte.org or from the author.
4122
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4121-4122 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 21