Kinetics and mechanism of heterogeneous palladium-catalyzed coupling reactions of chloroaryls in water

Article abstract of DOI:10.1039/a905625d

Coupling reactions of substituted chlorobenzenes to biphenyls catalyzed by palladium on carbon are performed in water using sodium hydroxide and sodium formate in the presence of a surface active agent. Thus, chlorobenzene, p-chloro-o-xylene, p-chloro-1,1,1-trifluorotoluene, p-chloroanisole, and p-chlorotoluene are coupled under moderate conditions to the respective biaryls. A competing reduction process occurs (e.g. chlorobenzene is reduced to benzene), which can be minimized by altering conditions. The relationship of product selectivity to reaction temperature, formate concentration, base concentration, and surfactant type is examined. The roles of formate, Pd catalyst, and surfactant are discussed. It is proposed that the reduction is dependent on the participation of palladium hydride [Pd²⁺(H^-)₂], while the coupling occurs via single electron-transfer from Pd⁰ to the substrate, with subsequent decomposition of the chloroaryl radical anions to obtain aryl radicals and chloride anions. This mechanism is supported by experiments with stoichiometric and sub-stoichiometric amounts of palladium which indicate that selective coupling can occur also in the absence of hydrogen (providing that reduced palladium Pd⁰, is present in sufficient amount), and by kinetic investigations which indicate that the coupling is actually a first-order reaction, for which the rate-determining step may be the dissociation of the chloroaryl radical anion.

Full text of DOI:10.1039/a905625d

View Article Online / Journal Homepage / Table of Contents for this issue
Kinetics and mechanism of heterogeneous palladium-catalyzed
coupling reactions of chloroaryls in water
Sudip Mukhopadhyay, Gadi Rothenberg,*† Diana Gitis, Harold Wiener and Yoel Sasson*
Casali Institute of Applied Chemistry, Hebrew University of Jerusalem, Jerusalem 91904, Israel
Received (in Cambridge, UK) 12th July 1999, Accepted 13th September 1999
Coupling reactions of substituted chlorobenzenes to biphenyls catalyzed by palladium on carbon are performed in
water using sodium hydroxide and sodium formate in the presence of a surface active agent. Thus, chlorobenzene,
p-chloro-o-xylene, p-chloro-1,1,1-trifluorotoluene, p-chloroanisole, and p-chlorotoluene are coupled under moderate
conditions to the respective biaryls. A competing reduction process occurs (e.g. chlorobenzene is reduced to benzene),
which can be minimized by altering conditions. The relationship of product selectivity to reaction temperature,
formate concentration, base concentration, and surfactant type is examined. The roles of formate, Pd catalyst, and
surfactant are discussed. It is proposed that the reduction is dependent on the participation of palladium hydride
[Pd2ϩ(HϪ)2], while the coupling occurs via single electron-transfer from Pd0 to the substrate, with subsequent
decomposition of the chloroaryl radical anions to obtain aryl radicals and chloride anions. This mechanism is
supported by experiments with stoichiometric and sub-stoichiometric amounts of palladium which indicate that
selective coupling can occur also in the absence of hydrogen (providing that reduced palladium Pd0, is present in
sufficient amount), and by kinetic investigations which indicate that the coupling is actually a first-order reaction,
for which the rate-determining step may be the dissociation of the chloroaryl radical anion.
Introduction
Preparation of biaryl compounds is of interest to many organic
chemists and relates to numerous applications, the biaryl
moiety being a key building block of many agrochemicals,
pharmaceuticals, and natural products.1 For example, substi-
tuted biphenyls are the backbone of new angiotensin-blocking
drugs called sartans for control of high blood pressure.2 To
meet market needs several synthetic approaches towards
biphenyl coupling have been developed, from the classic general
Ullmann3 and Suzuki4 reactions to custom-tailored catalytic
transformations aimed at specific molecules.
Various Pd-catalyzed biaryl syntheses are described in the
literature as alternatives to the stoichiometric Ullmann reac-
tion with copper. In several cases, stoichiometric amounts
of amines were used as hydrogen donors to regenerate the
Pd catalyst.5 Yields are generally moderate to good. The homo-
geneous Buchwald catalyst allows high-yield Suzuki coupling
of arylborons and aryl chlorides to obtain biphenyls at room
chloroaryl were stirred together with a slight excess of sodium
formate, in the presence of sodium hydroxide, a phase-transfer
catalyst (PTC) such as cetyltrimethylammonium bromide
temperature.2 Recently, Hassan et al. demonstrated6 that
homogeneous palladium acetate can catalyze the homo-
coupling of aryl and heteroaryl halides, but the homogeneous
(CTAB), a catalytic amount of Pd/C, and water. Biaryl was
catalyst could not be recycled.
found to be the major product, together with some hydro/
dehalogenation product. No other products were observed.
It is also possible to use a heterogeneous mixture of
palladium on carbon as the catalyst together with a sodium
Moderate to good yields of the coupling products were
formate solution as the terminal reductant. This approach is
obtained using various substrates, as shown in Table 1. The
palladium catalyst was easily recycled by simple filtration
followed by a short reflux in MeOH. In three such recycles
no appreciable reduction in catalyst activity was observed.
especially appealing as an industrial solution, as the potential
of catalyst recycling seems high. The utility of this simple sys-
tem was recognized over twenty years ago.7 In this paper we
present findings that pertain to the various catalytic coupling
Chlorobenzene 1a was chosen as a model substrate for the
and reduction transformations which occur in this palladium/
kinetic and process parameter studies.
formate/aryl system, and discuss possible mechanisms for the
Formate concentration was found to influence both substrate
reactions involved.
conversion and product selectivity (Fig. 1). Optimal yields of
the coupling product were obtained when [HCO2Ϫ] was
between 3 and 5 M.
Substrate conversion and selectivity towards the coupling
product also increased (from 46% conversion and 38% selectiv-
Results
In a typical reaction [eqn. (1)], two molar equivalents of
ity to 92–3% conversion and 87% selectivity) with the addition
of hydroxide. This effect reached a plateau at an NaOH concen-
tration of 25% w/v (2.8 equivalents relative to substrate).
† Present address: Centre for Clean Technology, Department of
Chemistry, The University of York, Heslington, York, UK YO10 5DD.
E-mail: gr8@york.ac.uk
J. Chem. Soc., Perkin Trans. 2, 1999, 2481–2484
This journal is © The Royal Society of Chemistry 1999
2481

View Article Online
Table 1 Palladium-catalyzed coupling and reduction of chloroaryls
Coupling
(%)b,c
Reduction
(%)b
Entry
Substrate
Time/ha
1
2
3
4
5
6
1a
1ad
1b
1c
1d
1e
1 1/2
1 1/4
4
1 1/4
5
87 (83)
88 (83)
25 (20)
82 (75)
21 (20)
67 (60)
13
12
75
18
79
33
2
Reaction conditions: 80–100 mmol substrate, 150 mmol NaHCO2, 250
mmol NaOH, 0.7 mmol of 10% Pd/C (1 mol% of Pd relative to sub-
strate), 3–5 mol% CTAB, 40 ml H2O, 110 ЊC, autoclave. a Time required
for 100% substrate conversion. b Yield based on GC area, corrected by
the presence of an internal standard. c The numbers in parentheses
represent isolated yields. d KHCO2 was used instead of NaHCO2.
Fig. 2 1a conversion profiles at different temperatures. Reaction
conditions: 88 mmol C6H5Cl, 250 mmol NaOH, 150 mmol NaHCO2,
1 mol% Pd/C, 9 mol% CTAB, 40 ml H2O.
Table 2 Effects of different PTCs on 1a conversion and product
selectivity
Conversion of
C6H5Cl (%)
Selectivity to
biphenyl (%)
Entry PTC type
1
2
3
4
5
6
7
8
9
None
60
93
93
92
79
86
64
85
79
42
48
62
29
87
71
41
51
63
53
31
51
29
30
73
(Me)3(C16H33)NϩBrϪ
(n-Bu)4NϩBrϪ
(n-Bu)4NϩFϪ
(n-Bu)4NϩIϪ
(n-Bu)4Nϩ(OAc)Ϫ
(n-Bu)4PϩBrϪ
(Et)3(Bz)NϩClϪ
(n-Hex)4NϩIϪ
10
11
12
(Me)3(C6H5)PϩBrϪ
[(Et)3Si(CH2)3](Me)3NϩBrϪ
PEG-400
Fig. 3 Selectivity towards coupling profiles at different temperatures.
Reaction conditions: 88 mmol C6H5Cl, 250 mmol NaOH, 15 mmol
NaHCO2, 1 mol% Pd/C, 9 mol% CTAB, 40 ml H2O.
Reaction conditions: 88 mmol C6H5Cl, 150 mmol NaHCO2, 250 mmol
NaOH, 0.7 mmol of 10% Pd/C (1 mol% of Pd relative to C6H5Cl), 40
ml H2O, 110 ЊC, autoclave, 1 h.
cheaper, more stable, and more easily separated than the
quaternary ammonium salts.
Performing the reaction at higher temperatures led to a faster
reaction rate with respect to substrate conversion, as expected.
However, the selectivity towards the coupling product increased
with temperature until a plateau was reached at 110 ЊC (cf. con-
version and selectivity profiles in Figs. 2 and 3). Note too that
product selectivity remains roughly constant throughout the
reaction.
Remarkably, the kinetics of this coupling reaction were
found to fit a first-order profile (e.g. kobs = 29 × 10Ϫ2 minϪ1, with
r2 = 0.996 for 9 observations). The calculated Arrhenius energy
of activation was Ea = 50 kJ molϪ1 (12.5 kcal molϪ1, r2 = 0.983
for four measurements at 90, 100, 110, and 120 ЊC), a typical
value for a chemically controlled rate-determining step.
When reactions were performed using variable catalyst load-
ings it was found that increasing the amount of palladium
catalyst also increases the selectivity towards the coupling
product. Thus, selectivities of 38, 60, 83, and 85% for biphenyl
were observed when 0.25, 0.5, 0.9, and 1 mol% of palladium
(relative to substrate) were employed. The conversion and
selectivity profiles were similar to those obtained with increased
temperatures, i.e. a constant product ratio was reached at a low
substrate conversion and maintained.
Evidently, the coupling and the reduction reactions are com-
peting transformations in this system. However, product select-
ivity can be shifted towards exclusive coupling owing to the fact
that the reduction reaction requires a hydrogen atom, while the
coupling requires only electrons that can be provided by the
reduced palladium catalyst. This was evidenced from a reaction
between 1a and stoichiometric palladium [eqn. (2)], in which a
Fig. 1 Effect of formate concentration on 1a conversion and product
selectivity. Reaction conditions: 88 mmol C6H5Cl, 250 mmol NaOH, 1
mol% Pd/C, 9 mol% CTAB, 40 ml H2O.
Performing the reaction in the presence of catalytic amounts
of various PTCs gave inconclusive results. While addition of a
PTC always improved the substrate conversion, the selectivity
towards biphenyl ranged from 87% using CTAB to 31% using
benzyltriethylammonium chloride (Table 2). This is in accord-
ance with the results reported by Bamfield and Quan, who
found that different phase-transfer agents and surfactants were
needed to maximize yield, depending on the substrate
employed.7 The nonionic surfactant polyethylene glycol 400
(PEG-400) also improved substrate conversion and product
selectivity, and might suit applications on a larger scale as it is
2C6H5Cl ϩ Pd0 → C6H5–C6H5 ϩ Pd2ϩ ϩ 2ClϪ (2)
2482
J. Chem. Soc., Perkin Trans. 2, 1999, 2481–2484

View Article Online
quantitative yield for 2a was obtained. Moreover, similar results
were obtained when a semi-stoichiometric setup was employed.
Thus, when the reaction was performed using semi-stoichio-
metric amounts of palladium (20 mol% Pd), a 20% yield of
biphenyl was measured after 1 h. At this time, 20 mol% of
formate were added, and a further 20% yield of biphenyl was
recorded. Adding a further batch of formate resulted in
100% conversion of the chlorobenzene and a quantitative
yield of biphenyl was obtained. Biphenyl was also formed as
the sole product when a similar semi-stoichiometric reaction
was carried out in the absence of base, but the reaction rate was
ten times slower.
(i.e. k4 ӷ k5). This is why a quantitative yield of biphenyl was
obtained when the formate was added in batches equivalent to
the amount of palladium catalyst present.
Conversely, when a sufficient amount of formate is present so
that enough of the Pd2ϩ is reduced to Pd0, eqn. (4) becomes
important and the reduction of chlorobenzene to benzene
[eqn. (6)] also takes place.
Pd2ϩ(HϪ)2 ϩ C6H5Cl → Pd2ϩ(HϪ)(ClϪ) ϩ C6H6 (6)
No chlorobiphenyls were observed, negating the possibility
of an attack of the phenyl radicals on chlorobenzene [eqn. (7)].
Discussion
Competition between reduction and coupling reactions
When a stoichiometric amount of palladium is employed (in
the absence of formate), quantitative conversion of chloro-
benzene to biphenyl is observed. However, when a catalytic
amount of palladium is used together with formate as the ter-
minal reductant, some reduction of chlorobenzene to benzene
is also observed. We assume that this reduction is caused by the
presence of dihydrogen.
Further support for this is obtained from the fact that 2b–e were
the only coupling products observed when 1b–e were employed
as substrates.
Suggested catalytic cycle leading to aryl coupling
We have previously shown that formate reacts with water in
the presence of a palladium catalyst to give bicarbonate and
dihydrogen [eqn. (3)]. In fact, the combination of formate and
Aryl halides are known as effective electron acceptors in
electron-transfer reactions with donors such as alkali metals,
carbanions, and various organometals.8 By analogy to homo-
geneous processes with palladium and nickel complexes, the
formation of a radical anion through a single electron transfer
process, followed by dissociation of this radical to give a halide
and an aryl radical, could represent a viable pathway for the
coupling reaction observed. The catalytic cycle may begin with
the coordination of a chlorobenzene molecule to the palladium
Pd0
HCO2Ϫ ϩ H2O
HCO3Ϫ ϩ H2
(3)
water can be regarded as a source of hydrogen in the presence
of palladium. Dihydrogen can reduce Pd2ϩ to Pd0 [eqn. (4)], and
k4
H2 ϩ Pd2ϩ
Pd0 ϩ 2Hϩ
(4)
cluster surface (Scheme 1, top), followed by a single electron
Ϫ
transfer to form [C6H5Cl] coordinated to a metastable Pdϩ
᝽
can also dissociate on the surface of Pd0 [eqn. (5)]. However, the
cation. Because of the inherent instability of this system, after
two substrate molecules interact with a palladium cluster elec-
trons in the metal cluster would “jump” to allow the dispropor-
k5
H2 ϩ Pd0
Pd2ϩ(HϪ)2
(5)
tionation of two Pd1ϩ atoms to Pd0 and Pd2ϩ
.
Each radical anion would now eject a chloride anion leaving
behind a phenyl radical (Scheme 1, bottom). The two phenyl
radicals which are still adjacent to the palladium cluster can
fact that batch addition of the formate gives quantitative yield
of the coupling product, and practically no benzene (no
reduction product) shows that the latter reaction is much slower
Scheme 1 Suggested pathways for palladium-catalyzed biaryl coupling: (top) catalytic cycle; (bottom) anion dissociation and phenyl radical
coupling.
J. Chem. Soc., Perkin Trans. 2, 1999, 2481–2484
2483

View Article Online
couple to form biphenyl, and Pd2ϩ can be reduced back to Pd0
in a reaction with one mole of dihydrogen [eqn. (3) and (4)].
Indeed, the observation of higher selectivity towards the
coupling product as reaction temperature is increased can be
explained by assuming that the coupling reaction (for which
Ea = 12.5 kcal molϪ1) benefits from the higher temperature,
while this change is not so significant for the reduction reaction
which has a lower activation energy, and indeed can occur
selectively at temperatures as low as 50 ЊC.9
Support for the single-electron-transfer process proposed
above can be found in the work of Munavalli et al. who
employed an analogous system10 for the coupling of 2-chloro-
3-trifluoromethylpyridine. Despite differences in reaction con-
ditions (especially in the surfactant and PTC amounts), it may
well be that dihydrogen was formed in that system, too. This
would account for the formation of large amounts of 3-
trifluoromethylpyridine, as these cannot be explained satis-
factorily by free-radical reaction with the solvent (toluene),
owing to the low quantities of benzyl-containing products
observed.
catalyst could be recycled, by vacuum filtration followed by
reflux in 25 ml of MeOH, and reused.
The substituted biphenyls 2b–e were similarly prepared: 2b
Isolated yield 20% based on 1b, mp 75 ЊC (from hot EtOH)
(lit.,14 76–77 ЊC). Found: C, 91.31; H, 8.68. C16H18 requires C,
91.42; H, 8.57%. 2c Isolated yield 75% based on 1c, mp 75 ЊC
(from EtOH–H2O) (lit.,15 93–94.5 ЊC). Found: C, 57.83; H,
3.05; F, 39.11. C14H8F6 requires C, 57.93; H, 2.75; F, 39.31%.
δH (CDCl3; Me4Si) 7.71 (8H, m, ArH) (lit.,16 7.67). 2d Isolated
yield 20% based on 1d, mp 175 ЊC (from hot EtOH) (lit.,12 176–
177 ЊC). Found: C, 78.36; H, 6.80. C14H14O2 requires C, 78.50;
H, 6.54; O, 14.95%. δH (CDCl3; Me4Si) 3.84 (6H, s, 2 × OCH3),
6.97 (4H, dt, ArH), 7.47 (4H, dt, ArH), good agreement was
found with literature values.17 2e Isolated yield 60% based on
1e, mp 118.5 ЊC (from CH2Cl2) (lit.,14 120.7–121.5 ЊC). Found:
C, 91.42; H, 7.92. C14H14 requires C, 92.30; H, 7.69%.
Experimental procedure for kinetic studies
Example: 88 mmol 1a, 250 mmol NaOH, 150 mmol NaHCO2,
0.8 mmol of 10% Pd/C and 7 mmol of CTAB were mixed in 40
ml of water at 110 ЊC in an autoclave (vide supra). Reaction
progress was monitored by GC. The following parameters were
studied: (i) initial substrate concentration, using xylene as the
organic solvent (three experiments at 10% w/v, kobs = 3.5 × 10Ϫ2
minϪ1, r2 = 0.996 for 7 observations: 15% w/v, kobs = 2.9 × 10Ϫ2
The exact role played by the base in this system is still
unclear.11 Addition of base may facilitate the coupling reaction
by neutralizing the acid produced in the reduction cycle of
Pd2ϩ
.
minϪ1, r2 = 0.996 for 9 observations; and 20% w/v, kobs
=
Conclusions
2.7 × 10Ϫ2 minϪ1, r2 = 0.996 for 9 observations); (ii) catalyst
loading (4 experiments using 0.25 mol% Pd, kobs = 2.4 ×
10Ϫ2 minϪ1, r2 = 0.998 for 8 observations; 0.5 mol% Pd,
kobs = 2.9 × 10Ϫ2 minϪ1, r2 = 0.987 for 7 observations; 0.9 mol%
Pd, kobs = 4.1 × 10Ϫ2 minϪ1, r2 = 0.992 for 7 observations; and 1
mol% Pd, kobs = 4.3 × 10Ϫ2 minϪ1, r2 = 0.999 for 7 observations);
(iii) reaction temperature (4 experiments at 90, 100, 110, and
120 ЊC); (iv) NaOH concentration (4 experiments for 1 h, at 0,
15, 25, and 35% w/v NaOH; conversions were 46, 78, 92, and
93% respectively; selectivities to 2a were 38, 69, 87, and 87%,
respectively).
Homocoupling of chloroaryls under moderate conditions can
be achieved in the presence of a heterogeneous palladium
catalyst. Competition between the reduction and coupling
pathways may well depend on the concentration of available
dihydrogen, which is supplied here as formate. The ease of
product separation and catalyst recycling promote this method
as a possible alternative to existing homogeneous catalytic
coupling techniques.
Experimental
Melting points were measured in glass capillaries using an
1
Electrothermal 9100 instrument. H NMR spectra were meas-
ured on a Bruker AMX 300 instrument at 300.13 MHz. GC
and GCMS analyses were performed using a HP-5890 gas
References
1 For reviews on biaryl preparation methods and applications
see: (a) G. Bringmann, R. Walter and R. Weirich, Angew. Chem., Int.
Ed. Engl., 1990, 29, 977; (b) M. Sainsbury, Tetrahedron, 1980, 36,
3327.
2 S. C. Stinson, Chem. Eng. News, 1999, 69.
3 F. Ullmann, Ber., 1903, 36, 2389.
4 (a) N. Miyamura, T. Yanagi and A. Suzuki, Synth. Commun., 1981,
11, 513; (b) A. Suzuki, Pure Appl. Chem., 1991, 63, 419; (c) for
a recent mechanistic study see M. Moreno-Mañas, M. Pérez and
R. Pleixats, J. Org. Chem., 1996, 61, 2346.
5 M. Brenda, A. Knebelkamp, A. Greiner and W. Heitz, Synlett, 1991,
809.
chromatograph with
a 50% diphenyl–50% dimethylpoly-
siloxane packed column (25 m × 0.53 mm). Unless stated
otherwise, chemicals were purchased from commercial firms
(>98% pure) and used without further purification. Products
1
were either isolated and identified by comparison of their H
NMR spectra with standard samples, or identified by MS data
and comparison of their GC retention times with standard
samples. Reactions below 100 ЊC were carried out in glass
apparatus, and reactions at 100 ЊC and above were performed
in a stainless steel Parr autoclave.
6 J. Hassan, V. Penalva, L. Lavenot, C. Gozzi and M. Lemaire,
Tetrahedron, 1998, 54, 13793.
7 P. Bamfield and P. M. Quan, Synthesis, 1978, 537.
8 (a) T. T. Tsou and J. K. Kochi, J. Am. Chem. Soc., 1979, 101, 6319;
(b) D. G. Morell and J. K. Kochi, J. Am. Chem. Soc., 1975, 97, 7262.
9 H. Wiener, PhD Thesis, The Hebrew University of Jerusalem, 1988.
10 S. Munavalli, D. I. Rossman, L. L. Szafraniec, W. T. Beaudry, D. K.
Rohrbaugh, C. P. Ferguson and M. Grätzel, J. Fluorine Chem., 1995,
73, 1.
11 cf. H. Zhang and K. S. Chan, Tetrahedron Lett., 1996, 37, 1043.
12 Y. Tamura, M.-W. Chun, K. Inoue and J. Minamikawa, Synthesis,
1978, 822.
13 N. Kamewaza, J. Magn. Reson., 1973, 11, 88.
14 A. McKillop, L. F. Elsom and E. C. Taylor, Tetrahedron, 1970, 26,
4041.
15 Dictionary of Organic Compounds, 6th edn., Chapman and Hall,
London, 1996, vol. 1, p. 899.
General procedure for coupling of chloroaryls
Example: biphenyl 2a from chlorobenzene 1a. In a 300 ml stain-
less steel autoclave equipped with a six-bladed impeller, an
external heating mantle and a gas cooling system, were placed
10.0 g (88 mmol) of 1a, 10.0 g (150 mmol, 170 mol%) of
NaHCO2, 10.0 g (250 mmol, 280 mol%) of NaOH, 1.5–1.8 g
(4–5 mmol, ca. 4 mol%) of CTAB, 0.85 g of 10% w/w Pd/C (0.8
mmol Pd, <1 mol%), and 40 ml H2O. Reaction progress was
monitored by GC. The autogeneous reaction pressures ranged
between 1.2–1.5 atm. The mixture was stirred at 100 ЊC for 2 h,
cooled and extracted with 40 ml of CH2Cl2. Solvent evapor-
ation and recrystallizing afforded 5.63 g (83 mol% based on 1a)
of biphenyl, mp 69 ЊC (from cold EtOH) (lit.,12 69–71 ЊC).
Found: C, 93.20; H, 6.80. C12H10 requires C, 93.46; H, 6.54%. δH
(CDCl3; Me4Si) 7.37 (2H, tt, aromatic 4,4Ј-H), 7.46 (4H, qt,
aromatic 3,3Ј,5,5Ј-H), 7.61 (4H, dq, aromatic 2,2Ј,6,6Ј-H), good
agreement was found with literature values.13 The palladium
16 B. M. Trost and H. C. Arndt, J. Am. Chem. Soc., 1973, 95, 5288.
17 C. J. Pouchert, The Aldrich Library of NMR Spectra, 2nd edn.,
Aldrich Chemical Co., Milwaukee, 1983, vol. 1, p. 854B.
Paper 9/05625D
2484
J. Chem. Soc., Perkin Trans. 2, 1999, 2481–2484