A catalytic and mechanistic investigation of a PCP pincer palladium complex in the Stille reaction

Article abstract of DOI:10.1039/b417908k

In an improved procedure, the complex {2,6-bis[(diphenylphosphino)methyl] benzene}chloropalladium(II) (1) was synthesised as its THF adduct and the structure was determined by X-ray crystallography. The catalytic properties of the derivative {2,6-bis[(diphenylphosphino)methyl]benzene}(trmuoroacetato) palladium(II) (2) was investigated in the Stille reaction. Complex 2 proves to be an excellent catalyst for the C-C cross-coupling between trimethyl phenyl stannane and aryl bromides using a very low catalyst loading (0.1-0.0001%), giving high turnover numbers (TONs) up to 6.9 × 10⁵. A kinetic investigation of the catalytic reaction suggests a heterogeneous colloidal palladium catalyst formed from the PCP Pd(II) pre-catalyst. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b417908k

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
A catalytic and mechanistic investigation of a PCP pincer palladium
complex in the Stille reaction†
Daniel Olsson, Patrik Nilsson, Mostafa El Masnaouy and Ola F. Wendt*
Inorganic Chemistry, Department of Chemistry, Lund University, P. O. Box 124, S-221 00,
Lund, Sweden. E-mail: ola.wendt@inorg.lu.se
Received 26th November 2004, Accepted 19th April 2005
First published as an Advance Article on the web 29th April 2005
In an improved procedure, the complex {2,6-bis[(diphenylphosphino)methyl]benzene}chloropalladium(II) (1) was
synthesised as its THF adduct and the structure was determined by X-ray crystallography. The catalytic properties of
the derivative {2,6-bis[(diphenylphosphino)methyl]benzene}(trifluoroacetato)palladium(II) (2) was investigated in the
Stille reaction. Complex 2 proves to be an excellent catalyst for the C–C cross-coupling between trimethyl phenyl
stannane and aryl bromides using a very low catalyst loading (0.1–0.0001%), giving high turnover numbers (TONs)
up to 6.9 × 105. A kinetic investigation of the catalytic reaction suggests a heterogeneous colloidal palladium catalyst
formed from the PCP Pd(II) pre-catalyst.
cycle was proposed,7,18 based on the fact that the catalyst is often
recovered unchanged afterwards, but this notion has not been
unchallenged. Regarding the related so-called orthopalladated
(bidentate-P,C cf. Chart 1: 5, 6) catalysts there now seems to be
Introduction
In recent years, tremendous progress has been reported in the
development of powerful new methods for carbon–carbon bond
formation and palladium catalysed cross-coupling reactions
some consensus for a classical cycle, where the Pd(II) pre-catalyst
have been important and flexible components in this process.1
is reduced to Pd(0) in an initiation step.19–22 However, for the
The Stille reaction, the cross-coupling of aryl halides or
PCP catalysts there has been arguments both for and against
triflates with aryl stannanes, is extremely versatile. Today, this
the Pd(II)/Pd(IV) cycle but so far there is little experimental
process can be viewed as a routine synthetic tool, but its study
evidence for either.19,23,24 During the preparation of this work a
continues, and recent work has also provided a greater mech-
study of the mechanism of the Heck reaction catalysed by PCP
anistic understanding of the reaction.2–4 A major development
complexes appeared. The author concludes that the reaction
has been the successful activation of aryl chlorides for the Stille
is probably heterogeneous and that the PCP complexes only
reaction.5
act as precatalysts for catalytically active metallic palladium(0)
During the last few years phosphapalladacycles including
species.25
PCP pincer complexes (e.g. Chart 1: 1–3) have emerged as
Here we report on the catalytic activity of {2,6-bis[(diphenyl-
very promising catalysts for C–C bond forming reactions.
phosphino)methyl]benzene}(trifluoroacetato)palladium(II) (2)
Although pincer complexes were synthesised already in 1976
in the Stille reaction between aryl bromide and trimethylphenyl-
by Moulton and Shaw,6 they were not applied in catalysis
tin, together with a kinetic and mechanistic investigation of the
until 1997 when Milstein and co-workers used PCP com-
catalytic reaction. Further, we present an improved synthetic
plexes in the Heck reaction obtaining good yields and high
strategy for {2,6-bis[(diphenylphosphino)methyl]benzene}chlo-
turnover numbers.7 Later, applications of PCP complexes in
catalysis include dehydrogenation,8 Karasch addition,9,10 ketone
reduction,11 asymmetric aldol,12,13 and Heck reaction.7,14–17
ropalladium(II) (1).
Experimental
General considerations
All experiments involving synthesis of metal complexes and
phosphine ligands were carried out under argon or nitrogen in a
Braun glove-box equipped with an inert gas purifier or using
standard Schlenk and/or high vacuum techniques. All non-
deuterated solvents were refluxed over sodium/benzophenone
ketyl or CaH2 and freshly distilled under nitrogen, except when
noted. Commercially available reagents were used as received.
Compound 2 was prepared according to the literature.7,26 NMR
spectra were recorded on a Varian Unity (299.79 MHz for 1H)
1
or Varian Unity INOVA (499.77 MHz for H) spectrometer.
Chemical shifts are given in ppm downfield from TMS using the
residual 1H NMR solvent peak, H3PO4 (31P d 0) or CClF3 (19F d
0) as internal and external reference, respectively.
Chart 1
The mechanism of operation of the PCP complexes in catalysis
has been the subject of controversy. Early on a Pd(II)/Pd(IV)
Preparations
† Electronic supplementary information (ESI) available: Reaction condi-
tions and observed rate constants for reaction (1) at 160 ◦C in benzene-d6
solvent. Reaction conditions and yields for reaction (1) in toluene sol-
vent. Reaction schemes and derived rate constants for the homogeneous
reactions. Crystallographic details and molecular structure of complex
1. See http://www.rsc.org/suppdata/dt/b4/b417908k/
One-pot synthesis of {2,6-bis[(diphenylphosphino)methyl]phe-
nyl}chloropalladium (1·THF). A solution of n-BuLi in hexanes
(1.8 mL, 1.6 M, 2.89 mmol) was added at −40 C◦ to a
solution of HPPh2 (0.54 g, 2.89 mmol) in THF (15 mL). After
completed addition the reaction mixture was stirred at the same
1 9 2 4
D a l t o n T r a n s . , 2 0 0 5 , 1 9 2 4 – 1 9 2 9
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
©

View Article Online
temperature during 2 h, and then allowed to warm to room
temperature during 1 h. After decreasing the temperature to
−70 ◦C, a solution of 1,3-bis(chloromethyl)benzene (0.25 g,
1.45 mmol) in THF (5 mL) was added dropwise. The reaction
mixture was allowed to warm to room temperature and stirred
overnight. Without purification, a THF (5 ml) suspension of
Pd(TFA)2 (0.48 g, 1.45 mmol) (Aldrich) (TFA = trifluoroacetate)
was added directly to the reaction mixture and stirred at room
temperature for five hours. The mixture was filtered to remove
unreacted Pd(TFA)2, and the solvent was removed from the
filtrate in vacuo. The resulting solid was dissolved in diethyl
ether (30 mL), a white precipitate, which was filtered off, was
observed and the ether was evaporated from the filtrate in vacuo
yielding a yellow precipitate. Recrystallisation from THF gave
yellow crystals of pure 1·THF (0.56 g, 60%). 1H NMR (CDCl3):
d 1.86 (m, 4H, THF; CH2), 3.75 (m, 4H, THF; CH2O) 3.96
(vt, 2JP–H = 4.65 Hz, 4H, CH2P), 6.98–7.04 (m, 1H, back bone),
7.11 (d, 3J = 7.5 Hz, 2H, back bone), 7.37–7.42 (m, 12H, meta-
2, 4-bromoanisole, Me3SnPh, ferrocene (as internal standard)
and solvent and placed in an oil bath at 160 C. The reaction
◦
was monitored by following the changes of the chemical shift
at 3.17 and 3.56 ppm, belonging to the methoxy group of 4-
bromoanisole and 4-phenylanisole, respectively. The product
was characterised in situ. All reactions were studied under
pseudo-first-order conditions with an excess of either the tin
compound ((2.16–864) × 10−3 mol dm−3) with respect to the
aryl bromide ((9.62–96.2) × 10−3 mol dm−3), or vice versa. To
assure catalytic reaction conditions the concentration of the
palladium(II) complex was at least one order of magnitude less
((0.15–0.52) × 10−3 mol dm−3) than that of Me3SnPh and 4-
bromoanisole. Stock solutions of all reagents in benzene-d6 were
used, except for the tin compound, which was administrated
as received. Inhibition experiments were performed by adding
approximately 0.1 g (5 × 10−4 mol) of elemental mercury to the
reaction mixture.27,28 In addition, experiments using cyclooc-
tatetraene (COT) as catalyst inhibitor were also performed. The
amount of COT was approximately a two-fold excess (0.35 ×
10−3 mol dm−3) relative to the amount of the catalyst, and was
added at the start of the reaction.29 The 1H NMR signals were
integrated generating kinetic traces, which were fitted to single
exponentials using the KaleidaGraph software.
1
and para-PPh), 7.83–7.90 (m, 8H, ortho-PPh). 31P{ H} NMR
(CDCl3): d 34.4 (s). Found: C, 63.36; H, 5.25. C36H35ClOP2Pd
requires C, 62.89; H, 5.13%.
Preparation of {2,6-bis[(diphenylphosphino)methyl]phenyl}-
bromopalladium (3). In a J. Young NMR-tube 2.2 mg (3.2 ×
10−6 mol) of 2 was dissolved in 0.500 mL benzene-d6. 23.3 mg
(2.3 × 10−4 mol) of NaBr was added and the TFA for bromide
substitution in 2 was followed using NMR spectroscopy. In
another preparation a J. Young NMR-tube was charged with
2.8 mg (4.0 × 10−6 mol) of 2. The complex was dissolved in
0.500 mL benzene-d6 and 3 mg (1.2 × 10−5 mol) of Me3SnBr was
added. Again, an almost instant TFA for bromide substitution
was observed, generating an identical spectrum as in the previous
preparation. In both cases a total consumption of the starting
material was seen. The products were not separated and isolated,
Results and discussion
Synthesis
Our preparation of complex 2 using the method described by
Milstein from the ligand 1,3-bis[(diphenylphosphino)methyl]-
benzene (4) and Pd(TFA)2 in THF (Scheme 1) resulted in very
low yield (20–30%), probably due to the tedious purification
procedure, contrary to the reported high yield of 96%.7,26 In an
alternative procedure, we decided to synthesise complex 2 in one-
pot without the isolation of phosphine ligand 4. Pd(TFA)2 was
directly added to the crude ligand mixture. After purification,
we unexpectedly obtained complex 1·THF in reasonably good
yield (60%) as the only product. This result might easily be
justified by the formation of 2 which then reacts with LiCl
present in the reaction mixture. The structure of 1·THF was
1
2
but characterised in situ. H NMR (C6D6): d 3.53 (vt, JP–H
=
4.80 Hz, 4H, CH2), 6.85–7.05 (m), 7.99 (m). 31P{ H} NMR
1
(C6D6): d 35.3 (s).
Typical procedure for Stille reactions
1
identified by H and 31P NMR spectroscopy and confirmed by
All equipment was rinsed with aqua regia prior to use. A
toluene solution (10 mL) of 1 equivalent of aryl halide, 0.5
equivalents of 2-methylnaphthalene (as internal standard), 1.5
equivalents of Me3SnPh and 2 (added from a 10−5 M stock
solution) were placed in a Schlenk tube with a Teflon-lined
screw cap together with a magnetic stirring bar. The mixture
was stirred vigorously at the temperature and during the time
indicated in Table 2. Samples were withdrawn after 1, 2, 3, 4,
18, 19 or 40 h depending on when no further improvement
in the yield was detected. For GC-analysis 10–15 drops of the
reaction mixture were taken out and dissolved in 2 mL of
Et2O. The organic phase was washed with 2 mL H2O, separated
and sealed in a GC-vial. Gas chromatography analyses for the
determination of the yield were performed with a Varian 3300
gas chromatograph with FID and fitted with a 30 m CP-Sil 5
CB capillary column. In some cases the product was isolated:
when the reaction was finished, the toluene was evaporated and
the crude reaction mixture was extracted with diethyl ether and
washed with water. The organic phases were dried over MgSO4
and filtered. The solvent was evaporated and the resulting
crude mixture was purified by column chromatography on silica.
Biphenyl was eluted with hexane, 4-acetylbiphenyl with hexane–
ether (1 : 2), 2-acetylbiphenyl with hexane–ether (1 : 5) and 4-
phenylbenzaldehyde with hexane–ethyl acetate (1 : 9). All of the
products are known and commercially available.
X-ray crystallography. Complex 1 was prepared for the first time
by Rimml and Venanzi30 from ligand 4 and [PdCl2(CH3CN)2]
in CH2Cl2, and its crystal structure was determined later by
the same author.31 The crystals were obtained from a CH2Cl2
solution and the structure was solved in the monoclinic space
group P21/n. We obtained pale yellow, octahedral blocks
suitable for X-ray crystallography by slow crystallisation in THF
at room temperature.‡ In this case, 1 co-crystallises with one
molecule of THF in the tetragonal space group P41212.
Scheme 1
‡ Crystal data: C32H27ClP2Pd·C4H8O, M = 687.43, tetragonal, a = b =
3
˚
˚
15.016(2), c = 14.129(3) A, V = 3185.7(9) A , space group P41212 (no.
92), Z = 4, 32386 reflections measured, 5069 unique (Rint = 0.0514) which
were used in all calculations. The final wR(F2) was 0.0740 (all data) and
the R(F) was 0.0299 (I > 2r(I)) using 185 parameters. CCDC reference
number 256981. See http://www.rsc.org/suppdata/dt/b4/b417908k/
for crystallographic data in CIF or other electronic format.
Kinetics
The kinetic experiments for the catalytic reaction were per-
1
formed in benzene-d6 solution using H NMR spectroscopy.
In a typical experiment, a J. Young NMR-tube was loaded with
D a l t o n T r a n s . , 2 0 0 5 , 1 9 2 4 – 1 9 2 9
1 9 2 5

View Article Online
Half a molecule forms the asymmetric unit with the C4–
C1–Pd–Cl1 atoms along the twofold axis. This axis generates
the molecular structure shown in ESI Fig. S1†together with
selected bond distances and angles. As determined by symmetry
the coordination geometry is perfectly planar but with distorted
angles. Similar to other PCP complexes the P–Pd–P angle is
much lower than 180◦, in this case at 165.73(3)◦. It is also clear
that packing effects are not negligible since the Pd–Cl and Pd–C1
distances in 1·THF are substantially longer than those reported
by Venanzi for the monoclinic structure.31
entry 6. These results convinced us that we had good catalytic
activity for activated aryl halides and no further test with
bromoacetophenones were performed.
We then undertook a thorough study aimed at comparing
different aryl halides under low catalyst load and high tem-
perature to determine the versatility of complex 2. The results
are summarised in Table 2 and ESI Table 2.† From entry 1 it
is seen that the yield and TON with phenyl chloride is very
poor with our system. The highest yield of 17% was achieved
with a catalyst load of 0.1%, at 160 ◦C for 40 h. For phenyl
bromide the best yield was 91%, entry 2, and no improvement
in the yield was detected after 40 h. Phenyl bromide gives the
highest TON, 6.9 × 105, in all of the catalytic experiments,
using a catalyst load of 0.0001%, entry 5. This is to the best
of our knowledge by far the highest TON reported for a Stille
coupling of an unactivated aryl bromide. The phenyl bromide
reactions performed at 110 ◦C, entry 3, gave poor results. Entries
8–11 concern functionalised aryl bromides. The results of these
experiments rule out the existence of homo-coupling. Contrary
to expectations, 4-bromobenzaldehyde, 4-bromotoluene and
4-bromoanisole turned out to be about as reactive as the
unsubstituted phenyl bromide when the catalytic load was 0.01%
and the temperature 160 ◦C. The anisole was even slightly
more active than bromobenzene. With lower catalyst loads of
0.0001%, however, the substituted bromobenzenes fared worse
than the unsubstituted one. As indicated by the preliminary
experiments in Table 1 steric effects were not dramatic and the
catalyst works fairly well also for ortho-substituted substrates.
Examples of conversion plots at 110 ◦C and 160 ◦C are given in
Fig. 1.
Catalysis
PCP palladium(II) complexes have been widely used in different
metal mediated organic transformations. However, these types
of complexes have not been used in Stille coupling. Therefore,
we wished to see whether complex 2 would show the same
high activity in the Stille reaction as was earlier demonstrated
in the Heck reaction.7 It should be mentioned that other
phosphapalladacycles have been used as Stille pre-catalysts.
For example Louie and Hartwig32 reported that palladacycle 5
catalysed the coupling reaction between 4-bromoacetophenone
and Me3SnPh with a TON of 1650. The highest TON of 8.4 ×
105 has been reported by Bedford and co-workers33 who used
the same aryl bromide but Bu3SnPh as transmetallating agent
and complex 6 as catalyst, cf. Chart 1.
[Pd]=2
Ar–X + Ph–SnMe3 −−−→ Ar–Ph + X–SnMe3
(1)
A preliminary survey of pre-catalyst 2 with various aryl
bromides in reaction (1) gave very promising results, cf. Table 1.
In all cases but one, good to excellent GC and isolated yields
were obtained. With a catalyst load of 0.001%, a temperature of
Entries 8–11 in Table 2 show that the simple rule that electron
withdrawing substituents on the arene give higher yields and
electron donating groups gives lower do not apply in the present
reaction. The fact that no dramatic change was observed when
changing from phenyl bromide to phenyl iodide or activated
and deactivated aryl bromides suggest that the actual rate
determining step is probably not oxidative addition (vide infra).
◦
160 C and a reaction time of 120 h the yield was quantitative
according to GC, Table 1, entry 4. The low catalyst load gave
an exceptionally high TON of 105. Activated and sterically
hindered substrates gave high yields, whereas the use of a
deactivated electron rich substrate only gave a 60% yield, Table 1,
Table 1 Results of the Stille reaction mediated by catalyst precursor 2 under various reaction conditions
Entry
ArX
[Pd]a (%)
tb/h
T/◦C
TONc
TOFd/h−1
Yielde (%)/f (%)
1
2
3
4
5
6
PhBr
PhBr
4-BrC6H4Ac
4-BrC6H4Ac
2-BrC6H4Ac
4-BrC6H4OMe
1
0.1
0.01
0.001
0.01
0.1
6
24
48
120
48
48
110
110
110
160
110
110
96
930
9300
16
40
190
830
180
13
90/95
84/92
87/93
95/100
75/85
50/60
1 × 105
8500
600
a [Pd]/[ArX] × 100. b Refers to the time after which no improvement of the yield was detected. c mol product/mol catalyst, based on GC yield.
d mol product/(mol catalyst × time), based on GC yield. e Isolated yield after column chromatography. f Conversion determined by GC based on the
starting material.
Table 2 Results of the Stille reaction mediated by catalyst precursor 2 under various reaction conditions
Entry
ArXa
[Pd]b (%)
tc/h
T/◦C
TONd
TOFe/h−1
Yieldf (%)
1
2
3
4
5
6
7
8
9
PhCl
PhBr
PhBr
PhBr
PhBr
PhI
PhI
4-BrC6H4CHO
4-BrC6H4Me
4-BrC6H4OMe
2-BrC6H4OMe
0.1
0.1
40
19
18
18
18
19
18
18
18
18
18
160
160
110
160
160
160
160
160
160
160
160
170
920
1400
7600
690000
850
210000
7300
7500
8700
5400
4.4
48
75
17
91
13
76
68
85
21
73
75
86
54
0.01
0.01
0.0001
0.1
0.0001
0.01
0.01
0.01
0.01
420
38000
45
11500
400
420
10
11
480
300
a 1.0 × 10−3 mol. b [Pd]/[ArX] × 100. c Refers to the time after which no improvement of the yield was detected. d mol product/mol catalyst, based
on GC yield. e mol product/(mol catalyst × time), based on GC yield. f Yield determined by GC based on the product.
1 9 2 6
D a l t o n T r a n s . , 2 0 0 5 , 1 9 2 4 – 1 9 2 9

View Article Online
have been varied and relatively good fits are generated when
the kinetic traces are fitted to a single exponential. In Fig. 2
typical kinetic traces for the reaction under catalytic conditions
are displayed. A first order dependence is indicated for both
substrates, but clearly there are large errors. All observed pseudo-
first-order rate constants are given in ESI Table 1.†
Fig. 1 Conversion plots of Entry 2 (ꢀ) and 3 (᭺) in Table 2. Solid lines
denote the best fit to a single exponential using KaleidaGraph.
Kinetics and mechanism
Preliminary investigation. In line with the catalysis exper-
iments, monitoring reaction (1) in benzene-d6 solution by
NMR spectroscopy measurements shows that the reaction
between Me3SnPh and PhBr is catalysed by 2 and exclu-
sively gives biphenyl as the product. Simultaneously, and
within 30 min, 2 is converted to [2,6-bis[(diphenylphosphino)-
methyl]phenyl]bromopalladium(II) (3). No intermediates were
identified and no decomposition could be noted. Complex 3
was identified by an independent synthesis from 2 and NaBr
in benzene-d6 solution, cf. Scheme 2. Compound 3 forms
almost instantly in a TFA for bromide substitution as seen
by for example a shift of the characteristic virtual triplet,
Fig. 2 Observed intensities of 4-bromoanisole (ꢀ) and 4-phenylanisole
[Me3SnPh] = 6.5 × 10−1 mol dm−3; [4-bromoanisole]0 = 9.6 ×
10−3 mol dm−3; [ferrocene] = 3.7 × 10−3 mol dm−3. Solid lines denote
the best fit to a single exponential using KaleidaGraph.
◦
(᭺) as a function of time at 160 C: [2] = 1.6 × 10−4 mol dm−3
;
To confirm the first order dependence a number of reactions
were run where the aryl bromide was used in excess compared to
the tin compound; in this case data could be collected using the
methyl shifts corresponding to either the Me3SnPh decay or the
4-phenylanisole or Me3SnBr formation. Again, there are large
errors and a fairly low consistency among the data, but there is
no systematic variation with tin concentration, again supporting
a first order dependence in [Sn], cf. Table 3. In another series
of experiments the tin concentration was kept in excess and
varied. The points in the corresponding plot (Fig. 3) were
very scattered indicating non-reproducible kinetics. The trend,
however, indicates increasing rate with increasing concentration,
in line with the other results.
corresponding to the methylene protons, from 3.32 (2JP–H
=
4.50 Hz) to 3.53 (2JP–H = 4.80 Hz) ppm. Also the possibility
of formation of 3 from Me3SnBr and 2 has been investigated
by NMR spectroscopy. Thus, 2 and Me3SnBr were mixed in
benzene-d6 solution and within 20 min. all of the starting
material had been converted to 3.
Scheme 2
No reaction was observed between 2 and either Me3SnPh or
PhBr when they were mixed separately, i.e. both of the substrates
are required to activate the catalyst. The only reaction observed
in the absence of aryl bromide is an alkyl/aryl scrambling
reaction of the tin compound. Such a rearrangement has also
been found to be catalysed by various platinum compounds.34
The same rearrangement is observed under catalytic conditions
in the kinetic experiments (vide infra), but only to a minor
extent so that any influence on the concentration of Me3SnPh in
the catalytic experiment could be ignored when Me3SnPh is in
excess.
Fig. 3 Observed pseudo-first-order rate constants for reaction (1) as a
function of [Me3SnPh] in benzene-d6 solution at 160 C◦. The values are
reported as means of the reactant decay and the product formation.
Complete data and further reaction conditions are reported in ESI
Table 1.†
Kinetic investigation. All kinetic experiments were done
with the aryl bromide and trimethylphenyltin in large excess
compared to the palladium(II) catalyst at 160 ◦C. Due to the
relatively high concentration of the tin compound, the phenyl
region of the NMR spectrum became very crowded. Using 4-
bromoanisole as substrate instead of phenyl bromide generates
a strong singlet at d 3.17, which appears at lower field (d 3.56)
as the formation of the product, 4-phenylanisole, takes place.
Under the reaction condition used the reaction is finished within
12 h. The concentrations of both 4-bromoanisole and Me3SnPh
To determine the reaction order with respect to the catalyst,
the concentration of 2 was varied causing a change of the
reaction rate in a fairly linear manner (Fig. 4). Overall, this
gives a somewhat vague picture of the rate law but within
single runs there seems to be a first order dependence in both
D a l t o n T r a n s . , 2 0 0 5 , 1 9 2 4 – 1 9 2 9
1 9 2 7

View Article Online
Table 3 Reaction conditions and observed pseudo-first-order rate constants for reaction (1) at 160 ◦C in benzene-d6 solvent when 4-bromoanisole
was used in excess. Constants for both the reactant decay and product formation are displayed.a
105k1/s−1
103[2]/M
103[4-bromoanisole]/M
103[Me3SnPh]/M
Me3SnPh
4-Bromoanisole
Me3SnBr
31 11
0.15
0.15
0.15
96.2
96.2
96.2
2.16
5.40
10.8
43
44 19
24
8
27
16
8
4
14
6
5
14.8 2.9
9.2 1.8
a [Ferrocene] = 3.66 mM in all measurements.
and catalyst. At least two homogeneous possibilities can explain
these observations; a Pd(II)/Pd(IV) cycle as has been proposed
for the Heck reaction and a classical Pd(0)/Pd(II) cycle. The
catalytic cycles and the derivation of the rate laws from these
are outlined in the ESI† showing that under reversible oxidative
addition the derived rate laws agree with the experimental one.
However, the mercury test speaks in favour of a heterogeneous
form of Pd rather than a homogenous one27,28 although the
latter cannot be ruled out.§ Also the very small inhibition seen
on addition of cyclooctatetraene makes the involvement of any
soluble Pd(0) species less likely. Just as in the homogeneous case
a heterogeneous catalysis, cf. Scheme 3, is compatible with the
experimental rate law. From the reaction scheme the following
rate law is obtained:
Rate = k2K1[Pd][Sn][ArBr]/(1 + K1[ArBr])
and assuming that the pre-equilibrium lies to the left, i.e. most
active sites are not occupied by a substrate, this reduces to the
following:
Fig. 4 Observed pseudo-first-order rate constants for reaction (1) as a
function of [2] in benzene-d6 solution at 160 ◦C. The values are reported
as means for the reactant decay and the product formation. Complete
data and further reaction conditions are reported in ESI Table 1.†
Rate = k2K1[Pd][Sn][ArBr]
substrates and in catalyst. Finally, the addition of Hg(l) to
the reaction mixture diminished the catalytic activity almost
totally. In this case less than 20% of the 4-bromoanisole
was converted to 4-phenylanisole after 12 h. To verify the
Hg(l) poisoning experiment, also experiments with addition of
cyclooctatetraene (COT) were performed. Agreement between
the two tests provides a good indication that the results are
valid.29 Thus, since the catalytic activity was diminished less than
20% when COT was added (cf. Fig. 5) the active catalyst species
can be considered as heterogeneous rather than homogenous. In
addition, it has been shown that already at 1 equivalent of added
dibenzo[a,e]cyclooctatetraene (dct), some inhibition (less then
20%) occurred in the case of a heterogeneous catalyst species.27
Scheme 3
From our initial experiments it is clear that all the precatalyst
is recovered within detection limits. Probably, most of the
palladium, which we retrieve as 3, is not involved in catalysis and
only a minor fraction forms a highly active form of Pd metal that
opens up for a heterogeneous cycle with reversibility in the first
step. Clearly, initiation takes place only at high temperatures and
it is entirely possible that the number of catalytically active sites
is a complicated function of the concentration of catalyst, sub-
strate, water and other impurities. The fact that we get scattered
data could then be due to variations in the amount of initiation.
That the PCP complex is only a pre-catalyst is also in keeping
with the fact that phosphinite PCP complexes are more active
and suffer from more massive decomposition during catalysis
Fig. 5 Observed intensities of 4-bromoanisole as a function of time at
160 ◦C with addition of Hg(l) (᭺), with addition of COT (ꢀ) and (᭛)
without any additive: [2] = 1.6 × 10−4 mol dm−3; [Me3SnPh] = 6.5 ×
10−2 mol dm−3 in the case of added Hg(l) and 2.2 × 10−2 mol dm−3
in the case of added COT or non-additive; [4-bromoanisole] = 9.6 ×
10−3 mol dm−3; [ferrocene] = 3.7 × 10−3 mol dm−3
.
§ One should note that also in the presence of mercury there is some
background reactivity that could be due to a slow, parallel homogeneous
path.
Reaction mechanism. Clearly, there can be several mechanis-
tic explanations for a rate law that is first order in both substrates
1 9 2 8
D a l t o n T r a n s . , 2 0 0 5 , 1 9 2 4 – 1 9 2 9

View Article Online
than the corresponding phosphine ones.25 In the phosphinite
case a substantial induction period was reported, whereas we
observe no or a very short induction, cf. Figs. 1 and 2. This seems
to indicate that once a small fraction of the pre-catalyst has been
transformed to an active catalytic site it remains active during
the whole course of the reaction. It is also clear that the low
catalyst loading and the fractional initiation keeps the palladium
from aggregating and forming palladium black. de Vries et al.
have shown that the turnover is proportional to catalyst loading
(using Pd(OAc)2 that forms nano-particles in solution) up to
a threshold value where aggregation becomes rapid, palladium
black precipitates and the catalyst activity drops to almost zero.35
It seems that a colloidal Pd catalyst formed in small and varying
amounts from the (PCP) pre-catalyst can explain most of our
observations.
2 A. L. Casado and P. Espinet, J. Am. Chem. Soc., 1998, 120,
8978.
3 A. L. Casado, P. Espinet and A. M. Gallego, J. Am. Chem. Soc.,
2000, 122, 11771.
4 P. Espinet and A. M. Echavarren, Angew. Chem., Int. Ed., 2004, 43,
4704.
5 A. F. Littke and G. C. Fu, Angew. Chem., Int. Ed., 2002, 41, 4176.
6 C. J. Moulton and B. L. Shaw, J. Chem. Soc., Dalton Trans., 1976,
1020.
7 M. Ohff, A. Ohff, M. E. van der Boom and D. Milstein, J. Am. Chem.
Soc., 1997, 119, 11687.
8 M. Gupta, C. Hagen, W. C. Kaska, R. E. Cramer and C. M. Jensen,
J. Am. Chem. Soc., 1997, 119, 840.
9 R. A. Gossage, L. A. van de Kuil and G. van Koten, Acc. Chem. Res.,
1998, 31, 423.
10 L. A. van de Kuil, D. M. Grove, R. A. Gossage, J. W. Zwikker, L. W.
Jenneskens, W. Drenth and G. van Koten, Organometallics, 1997, 16,
4985.
The conversion of 2 into 3, which does not happen in a
heterogenous (or homogeneous Pd(0)/Pd(II)) cycle, is easily
explained by the side reaction between 2 and the product
Me3SnBr. One final point that needs to be addressed is the fact
that variation of the aryl halide had little effect on the activity
in the catalysis experiments. This seemingly contradicts the fact
that the reaction is first order in electrophile. However, it is
entirely possible that any differences in equilibrium constants,
K1, are off-set by the reactivity in the following transmetallation
and the overall rate will be determined by the total barrier for
the latter, rate determining step. We attempted to observe any
colloidal palladium particles by TEM, but so far this has been
unsuccessful.
In conclusion, we have observed 2 to be an effective catalyst in
the Stille coupling reaction. Our mechanistic investigation points
strongly to a heterogeneous mechanism involving palladium
particles formed from the PCP Pd(II) pre-catalyst in small
amounts, but the absence of TEM evidence makes our evidence
less conclusive. Also, the mechanism of initiation is unknown to
us at this point. We argue that our observations strongly point
towards a heterogeneous active catalyst but clearly a number of
issues need to be resolved before we have a complete picture. We
are now pursuing such investigations in our laboratory.
11 P. Dani, T. Karlen, R. A. Gossage, S. Gladiali and G. van Koten,
Angew. Chem., Int. Ed., 2000, 39, 743.
12 J. M. Longmire, X. Zhang and M. Shang, Organometallics, 1998, 17,
4374.
13 F. Gorla, A. Togni, L. M. Venanzi, A. Albinati and F. Lianza,
Organometallics, 1994, 13, 1607.
14 D. Morales-Morales, C. Grause, K. Kasaoka, R. Redon, R. E.
Cramer and C. M. Jensen, Inorg. Chim. Acta, 2000, 300–302, 958.
15 D. Morales-Morales, R. Redon, C. Yung and C. M. Jensen, Chem.
Commun., 2000, 1619.
16 I. P. Beletskaya, A. V. Chuchurjukin, H. P. Dijkstra, G. P. M. van
Klink and G. van Koten, Tetrahedron Lett., 2000, 41, 1075.
17 S. Sjo¨vall, O. F. Wendt and C. Andersson, J. Chem. Soc., Dalton
Trans., 2002, 1396.
18 B. L. Shaw, S. D. Perera and E. A. Staley, Chem. Commun., 1998,
1361.
19 R. B. Bedford, Chem. Commun., 2003, 1787.
20 I. P. Beletskaya, A. N. Kashin, N. B. Karlstedt, A. V. Mitin, A. V.
Cheprakov and G. M. Kazankov, J. Organomet. Chem., 2001, 622,
89.
21 C. S. Consorti, M. L. Zanini, S. Leal, G. Ebeling and J. Dupont, Org.
Lett., 2003, 5, 983.
22 C. Rocaboy and J. A. Gladysz, New J. Chem., 2003, 27, 39.
23 I. P. Beletskaya and A. V. Cheprakov, Chem. Rev., 2000, 100, 3009.
24 J. Dupont, M. Pfeffer and J. Spencer, Eur. J. Inorg. Chem., 2001,
1917.
25 M. R. Eberhard, Org. Lett., 2004, 6, 2125.
26 R. Cohen, M. E. van der Boom, L. J. W. Shimon, H. Rozenberg and
D. Milstein, J. Am. Chem. Soc., 2000, 122, 7723.
27 D. R. Anton and R. H. Crabtree, Organometallics, 1983, 2, 855.
28 C. Paal and W. Hartmann, Ber. Dtsch. Chem. Ges., 1918, 51, 711.
29 T. Malmstro¨m and C. Andersson, J. Mol. Catal., 1997, 116, 237.
30 H. Rimml and L. M. Venanzi, J. Organomet. Chem., 1983, 259, C6.
31 F. Gorla, L. M. Venanzi and A. Albinati, Organometallics, 1994, 13,
43.
Acknowledgements
We thank Ms Patricia Becel for helpful assistance in the
laboratory. We also thank Ms Lisa Karlsson for valuable help
with the TEM investigation. Financial support from the Swedish
Research Council, the Knut and Alice Wallenberg Foundation,
the Crafoord Foundation and the Royal Physiographic Society
in Lund is gratefully acknowledged.
32 J. Louie and J. F. Hartwig, Angew. Chem., Int. Ed., 1996, 35, 2359.
33 D. A. Albisson, R. B. Bedford, P. N. Scully and S. E. Lawrence, Chem.
Commun., 1998, 2095.
34 C. Eaborn, K. Kundu and A. Pidcock, J. Chem. Soc., Dalton Trans.,
References
1981, 1223.
1 T. N. Mitchell, in Metal Catalysed Cross-coupling Reactions, ed.
F. Diederich and P. J. Stang, Wiley-VCH, New York, 1998.
35 A. H. M. de Vries, J. M. C. A. Mulders, J. H. M. Mommers, H. J. W.
Henderickx and J. G. de Vries, Org. Lett., 2003, 5, 3285.
D a l t o n T r a n s . , 2 0 0 5 , 1 9 2 4 – 1 9 2 9
1 9 2 9