Silica and polymer-tethered Pd-SCS-pincer complexes: Evidence for precatalyst decomposition to form soluble catalytic species in Mizoroki-Heck chemistry

Article abstract of DOI:10.1016/j.jcat.2004.05.015

Palladium(II)-SCS-pincer complexes have been covalently immobilized on porous silica and polynorbornene supports and evaluated in the Mizoroki-Heck reaction of iodobenzene and n-butyl acrylate. Kinetic experiments and poisoning studies indicate that the pincer complexes decompose under reaction conditions via rupture of the palladium-ligand bonds to liberate active Pd(0) homogeneous species. There is no evidence for catalysis by the intact Pd(II) pincer complex under any conditions evaluated, ruling out a Pd(II)-Pd(IV) catalytic mechanism that has been postulated for palladacycle catalysts under the conditions used here.

Full text of DOI:10.1016/j.jcat.2004.05.015

Journal of Catalysis 226 (2004) 101–110
www.elsevier.com/locate/jcat
Silica and polymer-tethered Pd–SCS-pincer complexes:
evidence for precatalyst decomposition to form soluble catalytic species
in Mizoroki–Heck chemistry
Kunquan Yu a, William Sommer b, Marcus Weck b,∗, Christopher W. Jones a,b,∗
a
School of Chemical and Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA
b
School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, GA 30332, USA
Received 12 March 2004; revised 7 May 2004; accepted 11 May 2004
Available online 15 June 2004
Abstract
Palladium(II)–SCS-pincer complexes have been covalently immobilized on porous silica and polynorbornene supports and evaluated in the
Mizoroki–Heck reaction of iodobenzene and n-butyl acrylate. Kinetic experiments and poisoning studies indicate that the pincer complexes
decompose under reaction conditions via rupture of the palladium–ligand bonds to liberate active Pd(0) homogeneous species. There is no
evidence for catalysis by the intact Pd(II) pincer complex under any conditions evaluated, ruling out a Pd(II)–Pd(IV) catalytic mechanism
that has been postulated for palladacycle catalysts under the conditions used here.
2004 Elsevier Inc. All rights reserved.
Keywords: Heck; Leaching; Organometallic; Silica; Catalyst; Pincer; Palladium
1. Introduction
12,14–16]. Polymer-immobilized Pd complexes [17–19] or
metallic Pd nanoparticles [20–23] have also been employed
as homogeneous or quasi-homogeneoussystems. With many
heterogeneous catalysts, there is substantial evidence that
leached/dissolved Pd species (that could be colloidal [24] or
molecular [25]) are the true active species [9–14] while in
other studies it is maintained that all the data support insol-
uble active species [6,7,15]. One issue that complicates this
analysis is that there is evidence that in some cases soluble
palladium species are produced during the reaction and that
after all reactants are consumed, these species redeposit back
onto the support with varying efficiencies during Mizoroki–
Heck and Suzuki–Miyaura reactions [13,26]. Thus, if stan-
dard tests for homogeneous catalysts such as (i) filtration
tests and (ii) elemental analysis of the solution for Pd after
reaction are the only methods used to discriminate between
soluble and insoluble catalysts, it is possible that they could
provide misleading evidence that rules out catalysis by sol-
uble species [10–13,27]. Indeed, in a related system, when
Ni/charcoal is used as a catalyst for Kumada couplings, there
is evidence that homogeneous catalyst redeposition onto the
support is rapid and easily missed [28]. If a truly heteroge-
neous catalyst that can be conclusively shown to impart no
The Heck reaction is among the most important carbon–
carbon bond forming reactions in use today, especially in
the fine chemical and pharmaceutical industries. A number
of different precatalysts have been utilized, with palladium
being the metal of choice in most systems. Highly active,
robust catalysts can be created from a number of differ-
ent precatalysts and a catalytic mechanism involving Pd(0)
and Pd(II) intermediates is well-established. Despite this, in
many cases, the nature of the active species is not under-
stood.
Both homogeneous and heterogeneous precatalysts have
been employed and there is conflicting information in the
literature concerning whether truly heterogeneous metallic
Pd is capable of promoting the Heck reaction. Traditional
heterogeneous catalysts have been employed, such as Pd(II)-
exchanged oxides or resins [1–9] or Pd(0) on standard sup-
ports such as activated carbon [8,10–14] or oxides [4,7–
*
Corresponding authors.
E-mail addresses: marcus.weck@chemistry.gatech.edu (M. Weck),
cjones@chbe.gatech.edu (C.W. Jones).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.05.015
2004 Elsevier Inc. All rights reserved.

102
K. Yu et al. / Journal of Catalysis 226 (2004) 101–110
Scheme 1.
leached metal species into solution is developed, this would
be a major advance, as residual metal contamination is the
primary drawback of current Heck catalysis techniques.
A large variety of homogeneous precatalysts have been
used in the Heck reaction. Simple Pd(II) phosphine com-
plexes are common precatalysts that are in situ reduced to ac-
tive Pd(0) species. Recently, the development of stable pal-
ladacycle complexes (Scheme 1) such as 1 (referred to here
as a half pincer) or 2a (referred to here as a pincer complex)
or 2b (referred to here as a CNC-pincer carbene complex)
that are active in Heck catalysis has spurred a debate about a
possible new mechanism in Heck catalysis [29–33]. Because
these complexes are active for Heck catalysis and have very
stable Pd–C bonds, it was surmised that perhaps they could
potentially operate by a new Pd(II)–Pd(IV) catalytic process
[31,34–37] instead of the generally accepted Pd(0)–Pd(II)
cycle. New investigations of the half-pincer catalysts have
shown that it is likely that these complexes are reduced to
form soluble or insoluble Pd(0) catalytic species [33,38–43]
and thus the Pd(II)–Pd(IV) cycle is unlikely [36,44]. How-
ever, for the pincer complexes, there is substantially less data
available on the nature of the catalytic site and no clear pic-
ture of the catalytic cycle has emerged.1
A number of different Pd pincer complexes have been
used as Heck precatalysts including SCS–Pd, NCN–Pd,
PCP–Pd, and CNC–Pd complexes as shown in Scheme 1
[17–19,34,43,44,46–52]. The stability of these complexes
in solvents tracks with the strength of the coordinating lig-
and, decreasing in the order CNC > PCP > NCN > SCS,
with the weakest of the pincer complexes still being sub-
stantially more stable than most other Pd(II) complexes with
nonchelating ligands [53,54]. In an effort to establish re-
coverable, recyclable catalysts, Pd pincer complexes have
been covalently immobilized on polymeric [17–19,46–49,
55] and silica [56] supports and ionically immobilized on
clays [57]. In these cases, catalysis with immobilized cata-
lysts was claimed and the catalysts were effectively recycled
and reused [17,18,46–49,56,57]. Here we report a detailed
study of covalently immobilized Pd–SCS-pincer complexes
on polymeric and porous silica supports and show that for
the systems studied here, there is no evidence for any het-
erogeneous catalysis2 under the conditions reported.
2. Experimental
2.1. Materials and instrumentation
DMF was distilled from calcium hydride, and NEt3 was
distilled over molecular sieves. All other reagents were ob-
tained from commercial sources and generally used with-
out further purification. Gas chromatographic analyses were
performed on a Shimadzu GC 14-A gas chromatograph
equipped with a flame-ionization detector and an HP-5 col-
umn (length = 30 m, inner diameter = 0.25 mm, and film
thickness = 0.25 µm). The temperature program for GC
analysis was the following: heating from 50 to 140 ◦C at
30 K/min and heating from 140 to 300 ◦C at 40 K/min un-
der constant pressure with inlet and detector temperatures
kept constant at 330 ◦C. H NMR (300 MHz) spectra were
1
recorded on a Mercury VX instrument. A Netzsch Ther-
moanalyzer STA 409 was used for simultaneous thermal
analysis combining thermogravimetric analysis (TGA) and
differential scanning calorimetry (DSC) with a heating rate
of 10 ◦C/min in air. FT-Raman spectra were obtained on
a Bruker FRA-106. At least 128 scans were collected for
each spectrum, with a resolution of 2–4 cm−1. X-ray pho-
toelectron spectra (XPS) were collected using a Physical
Electronics (PHI) Model 1600 XPS system equipped with
a monochromator with an Al-Kα source (hν = 1486.8 eV).
Ejected photoelectrons were detected by a hemispherical
analyzer and the pressure in the spectrometer was below
5 × 10−9 Torr. For energy calibration, the carbon 1s bind-
ing energy was set to 285.0 eV. Transmission electron mi-
croscopy (TEM) images were obtained using a JEOL 100C
operated at 100 kV.
SBA-15 (100 Å pore size) was synthesized following a
literature method [61]. The as-prepared material was cal-
cined using the following temperature program: (1) increas-
ing the temperature at 1.2 ◦C/min to 200 ◦C, (2) heating
at 200 ◦C for 1 h, (3) increasing at 1.2 ◦C/min to 500 ◦C,
and (4) holding at 550 ◦C for 6 h. Prior to functionalization,
the SBA-15 was dried under vacuum at room temperature
overnight and then at 120 ◦C/min for 3 h and stored in a dry
2
Here we use the traditional definitions of homogeneous (soluble) and
heterogeneous (insoluble) catalysts. The alternative definitions introduced
by Schwartz [58] and refined by Finke and co-workers [59,60] are not used
(homogeneous, single site; and heterogeneous, multisite).
1
While this manuscript was under review, a manuscript was submitted
that also addresses this issue with the related PCP-pincer system [45].

K. Yu et al. / Journal of Catalysis 226 (2004) 101–110
103
box. The pore size and BET surface areas of SBA-15 were
2.2.4. Synthesis of 4-iodobenzylidene-3-trimethoxyl-
silanylpropylamine (6)
calculated from data gathered from N2 physisorption exper-
iments at 77 K using a Micromeritics ASAP 2010 system.
TGA/DSC and elemental analyses were used to determine
the organic loadings of precatalyst 5.
4-Iodobenzaldehyde (1.0 g, 4.310 mmol) and 3-amino-
propyltrimethoxylsilane (740 mg, 4.134 mmol) were dis-
solved in anhydrous methanol (20 ml). The amount of 0.5 g
of 4 Å molecular sieves was used as water scavenger and
the mixture was refluxed for 20 h. After cooling to room
temperature, solvent was removed under reduced pressure
and the product was extracted with CH2Cl2. Subsequently,
the mixture was concentrated and the product was recrys-
tallized from CH2Cl2 and hexanes. Yield: 1.6 g (95%).
1H NMR (CDCl3): δ 9.96 (1H, s, CH =N), 7.75 (2H, d,
J = 8.4 Hz, ArH ), 7.45 (2H, d, J = 8.4 Hz, ArH ), 3.57
(9H, s, Si(OMe)3), 1.81 (4H, m, –CH2–), 0.71 (2H, m,
–CH2Si) ppm. Anal. Calcd for C13H20O3NISi: C, 39.70; H,
5.13; N, 3.56. Found: C, 39.83; H, 5.01; N, 3.45.
2.2. Catalyst synthesis
2.2.1. Palladation of 3,5-bis(phenylsulfanylmethyl)
phenol (3)
The SCS–OH pincer ligand was synthesized in five steps
according to previously reported procedures [62,63]. SCS–
OH (407 mg, 1.204 mmol) was dissolved in a 1:2 mix-
ture of CH2Cl2 (5 ml)/CH3CN (10 ml) and Pd(PhCN)2Cl2
(462 mg, 1.204 mmol) was then added in one portion. The
dark orange mixture was stirred for 30 min under argon,
at which point AgBF4 (470 mg, 3.333 mmol) was added
and the solution was stirred for an additional 30 min in
the dark. The solvent was then evaporated to dryness under
reduced pressure, n-Bu4NCl (2.0 g, 7.20 mmol) in 20 ml
CH2Cl2 was then added, and the resulting mixture was
stirred under argon for 48 h. The silver salt was removed
by passing the orange-red mixture through an alumina col-
umn. Crude complex 3 was obtained after solvent removal,
which was used for the next step without further purifica-
tion.
2.2.5. Preparation of silica-immobilized 4-iodobenzyli-
dene-3-trimethoxylsilanylpropylamine (7)
Complex 6 (1.6 g, 3.89 mmol) was added into a toluene
(30 ml) suspension of dry SBA-15 (1.5 g) and the mix-
ture was stirred for 24 h at room temperature. After fil-
tering and washing with hexanes and dichloromethane, the
collected solid was mixed with 1.0 g of HN(SiMe3)2 in
toluene (40 ml) and stirred at room temperature overnight.
The silica-immobilized 7 was collected by filtration and
washed extensively with hexanes and dichloromethane and
then dried under high vacuum. FT-Raman: δ 1644 (aro-
matic C=C), 1585 (C=N), 3056 (aromatic C–H), 2850–
2.2.2. Synthesis of Pd–Cl 3-(3,5-bis(phenylsulfanylmethy-
phenoxy)propyl-trimethoxysilane (4)
2960 (aliphatic C–H), 1053 (Ar–I) cm−1
.
Complex 3 (250 mg, 0.528 mmol) was mixed with 3-
iodopropyltrimethoxylsilane (150 mg, 0.517 mmol) and
Cs2CO3 (350 mg, 1.074 mmol) in dry DMF (15 ml). The
reaction mixture was stirred at 80 ◦C for 16 h, and then the
DMF was distilled off and the product was extracted into
CH2Cl2. Recrystallization from CH2Cl2/hexane yielded 4
as an orange yellow powder. Yield: 306 mg (93%). 1H NMR
(d6-DMSO): δ 7.43–7.27 (10H, m, Ph), 6.51 (2H, s, ArH ),
4.74 (4H, s, br, SCH2), 4.09 (2H, m, OCH2), 3.59 (9H, s,
Si(OMe)3), 1.75 (4H, m, CH2), 0.72 (2H, m, SiCH2) ppm.
Anal. Calcd for C26H31S2O4ClPdSi: C, 48.67; H, 4.87; S,
10.00. Found: C, 49.14; H, 4.43; S, 9.70.
2.2.6. Preparation of polynorbornene-immobilized
precatalyst (8)
The polymer-immobilized complex was synthesized in
3 steps following a previously reported procedure [64].
2.3. Catalysis
In a typical Heck reaction, precatalyst 5 (30 mg, 3.6 ×
10−3 mmol), iodobenzene (0.56 g, 2.75 mmol), n-butyl
acrylate (0.70 g, 5.47 mmol), NEt3 (0.40 ml, 2.87 mmol),
and dodecane (0.4 g, GC internal standard) were mixed in
4 ml DMF in a three-neck round-bottom flask equipped with
a condenser and rubber septum. The reaction assembly was
then connected to the manifold of a Schlenk line and was
gently purged with argon for 10 min before it was sealed.
The reaction mixture was then allowed to heat in an oil bath
at 120 ◦C with constant stirring. Aliquots were taken from
the reaction mixture periodically, diluted with THF, and con-
version was determined by GC chromatography.
2.2.3. Preparation of silica-immobilized precatalyst (5)
Complex 4 (200 mg, 0.329 mmol) was mixed with SBA-
15 (1.0 g) in dry toluene (50 ml) and the mixture was
stirred at 80 ◦C for 24 h, at which point the mixture was fil-
tered while hot and washed extensively with THF, acetone,
DMSO, and DCM to remove any physiabsorbed complex 4.
The recovered solid was then suspended in 40 ml toluene and
excess HN(SiMe3)2 (0.5 g) was added to cap any remaining
surface silanols. The resulting mixture was stirred at room
temperature for 24 h and then filtered. Silica-supported 5
was obtained after washing with DCM and THF and dried
under high vacuum.
3. Results and discussion
3.1. Catalyst synthesis and characterization
Reaction between SCS-pincer palladium(II) complex 3
with 3-iodopropyl-trimethoxylsilane in DMF yields com-

104
K. Yu et al. / Journal of Catalysis 226 (2004) 101–110
Scheme 2.
Fig. 2. Conversion of iodobenzene as a function of time for the first (A) and
third run (B).
Precatalyst 5 exhibited high activity for the Heck reac-
tion, and high turnover numbers (TON) were obtained
(> 700 mol product/mol Pd). The catalyst was recycled
three times, and while high yields were achieved even af-
ter the third run (Table 1), a longer induction period was
observed in the third run. The presence and elongation of
the induction period (Fig. 2) between subsequent runs gave
the first indication the silica immobilized 5 acts as a precata-
lyst, with the real active species generated in situ during the
reaction [38].
Fig. 1. X-ray photoelectron spectrum of the Pd(II) 3d binding energies in
precatalyst 5.
plex 4 with a Si(OMe)3 end group. The homogeneous SCS
pincer palladium(II) complex 4 is covalently immobilized
onto mesoporous silica support, e.g., SBA-15, utilizing a re-
action between the –Si(OMe)3 group with surface silanols
to form Si–O–Si bonds with the concomitant release of
methanol (Scheme 2). The remaining accessible surface
silanols were capped by reacting at room temperature with
excess hexamethyldisilazane. A typical loading for 5 sup-
ported on SBA-15 is approximately 0.12 mmol Pd per gram
support. The FT-Raman spectrum of 5 (not shown) includes
the aromatic C–H stretchings at 3056 and the aliphatic ones
at 2850–2987 cm−1, indicating the presence of the pincer
ligand. The XPS spectrum (Fig. 1) contains two binding
energy peaks at 337 (3d5/2) and 343 (3d3/2) eV, which is
characteristic of a palladium(II) species [65].
3.2.2. Filtration tests
To demonstrate that all the SCS-pincer Pd(II) complex 5
was covalently bonded to the silica surface and there were no
physisorbed species left behind, a filtration/leaching test was
performed before each batch of catalyst was used in a reac-
tion. Precatalyst 5 was heated in DMF at 120 ◦C for 30 min
and then filtered while still hot. The filtrate was then kept
at 120 ◦C and iodobenzene, n-butyl acrylate, triethylamine,
and GC internal standard n-dodecane were added and the
resulting mixture was stirred at 120 ◦C. GC analysis indi-
cated that only a trace amount of Heck reaction product is
produced, with less than 2% conversion of iodobenzene and
n-butyl acrylate after 90 min. Compared to the high activity
from a regular Heck reaction with precatalyst 5 (Table 1 and
Fig. 2), it appears that essentially all the SCS-pincer pal-
ladium(II) in complex 5 is covalently bonded to the silica
surface.
3.2. Heck catalysis
3.2.1. Catalytic reactions
The silica-immobilized SCS-type palladium(II) complex
5 was used as a precatalyst for the Mizoroki–Heck coupling
of iodobenzene and n-butyl acrylate. The reaction was car-
ried out at 120 ◦C, using triethylamine as base in DMF as
solvent:
With many heterogeneous catalysts reported in the litera-
ture, there is substantial evidence that leached/dissolved Pd
species (colloidal [24] or molecular [25]) are the true ac-
tive species [9–14]. One way to gather information on the

K. Yu et al. / Journal of Catalysis 226 (2004) 101–110
105
Table 1
a
Heck reaction of iodobenzene and n-butyl acrylate with 5
c
b
Run
No.
Time
(h)
TON
STY
Yield
(%)
1
2
3
1.5
2
3
> 99
97
94
764
733
718
509
367
239
a
−3
mmol) of pre-
Reactions were conducted using 30 mg (3.6 × 10
catalyst 5, 5.47 mmol butyl acrylate, 2.87 mmol NEt , and 2.75 mmol of
3
iodobenzene in 4 ml of DMF.
b
GC yield.
c
Site-time yield (STY) is the TON divided by overall reaction time,
Fig. 3. Conversion of iodobenzene as a function of time under standard con-
ditions (A) and in the presence of 100 eq of PVPy added after 20 min (B).
which is only a rough approximation of average activity, especially in view
of the presence of induction period.
(leached) palladium metal centers could contribute to the
Heck reaction. However, if another insoluble ligand that has
the capability to bind strongly to soluble palladium is added
during the reaction, one can potentially shut down the activ-
ity from the leached, soluble palladium species. By compar-
ing two Heck reactions, one in the presence of the insoluble
homogeneous Pd trap and one in the absence of the Pd trap,
it is possible to quantify the fraction of the total observed
activity that stems from both the supported and the leached
palladium species. Pyridines are known to bind strongly to
Pd(II) [67]. For this reason, commercially available poly(4-
vinylpyridine) (PVPy) (2% cross-linked) was chosen to test
this hypothesis. This polymer is insoluble under the reac-
tion conditions. In one experiment, precatalyst 5 is mixed
with iodobenzene, n-butyl acrylate, triethylamine along with
PVPy (molar ratio PVPy:Pd is 100:1) in DMF and heated at
120 ◦C. GC analysis indicates that less than 1.0% conver-
sion (based on iodobenzene) was obtained even after 3 h.
In contrast, as shown in Fig. 3, in the absence of PVPy,
nearly quantitative conversion of iodobenzene is observed
after only 30–40 min. In another experiment, a Heck reaction
was performed under standard conditions and after achiev-
ing 34% conversion based on iodobenzene, excess PVPy
(∼ 120 eq) was added into the reaction mixture. In this case,
GC analysis indicates that the addition of PVPy almost in-
stantaneously quenched the reaction (Fig. 3). The fact that
PVPy suppresses the activity of precatalyst 5 strongly sug-
gests that the Heck activity from complex 5 comes from
leached/unsupported Pd that is soluble in DMF, since the
probability for a high molecular weight insoluble polymer to
interact with the SCS-pincer palladium metal centers that are
immobilized inside the mesoporous channels of the SBA-15
is negligible. Thus, there is no evidence for catalysis by SCS-
pincer Pd under the conditions reported here.
nature of the catalytic site when a solid catalyst is involved
is through the use of a simple filtration test during the re-
action. For example, a regular Heck reaction performed us-
ing precatalyst 5 was interrupted, and the solid was filtered
off while hot after 61% conversion of iodobenzene. Subse-
quently, fresh reagents were added into the filtrate and it was
observed that the Heck coupling reaction continued at a rate
only slightly slower than the rate prior to the solid removal.
When the recovered solid was subjected to identical Heck re-
action conditions with new reagents, a substantially slower
conversion of the reagents was observed. Furthermore, el-
emental analysis of the solid before and after the first use
indicated that ∼ 17% of the immobilized palladium was lost.
This filtration test clearly suggests the presence of active sol-
uble Pd species in solution. However, it does not prove that
all the catalysis is solely attributable to soluble species. Since
the palladium is covalently bonded to the support before the
reaction, one reasonable conclusion is that some of the sup-
ported Pd metal is leached out into the solution during the
reaction.
3.2.3. Polyvinylpyridine poisoning
One of the goals of covalently supporting the SCS-pincer-
type palladium(II) complex onto silica supports was that it
would allow for the determination of whether the observed
reactivity was due to true palladium pincer catalytic species
or from palladium species that have been leached/lost into
the solution due to pincer ligand decomposition. This is of-
ten not possible for palladium catalysts that are just physi-
sorbed on solid supports such as charcoal, since the palla-
dium species will come off the solid during the reaction at
high temperature and redeposit back onto the support [28,
66]. In addition, for soluble polymer-supported systems [17,
18,46–48], it is extremely hard to distinguish any leached
soluble Pd species from the polymer-supported moieties as
both species can be soluble and complete recovery of low
molecular weight soluble polymer from solution may some-
times be impossible.
Furthermore, the above results indicate that the true cat-
alytic species may in fact be homogeneous molecular Pd,
as it is difficult to envision how insoluble PVPy could com-
pletely poison a reaction that is occurring on a nanoparticle
surface.3 Indeed, as the pincer complex decomposes, molec-
Since the SCS-pincer palladium(II) complex is covalently
bonded onto the insoluble solid support, if during the course
of the Heck reaction, some of the palladium species are
leached into solution, then both supported and unsupported
3
This assumes that for studies involving Pd nanoparticles, it is the
nanoparticle surface that is the catalyst. This conclusion is reached in sev-

106
K. Yu et al. / Journal of Catalysis 226 (2004) 101–110
Scheme 3. Schematic drawing of the three-phase test. Solid and hollow balls represent silica-immobilized pincer-Pd(II) precatalyst and silica-immobilized
iodobenzene analogue, respectively.
Scheme 4.
ular Pd species are liberated that could act as the true catalyst
or they could aggregate into soluble colloids that could act
as the catalyst. A useful control experiment is to prepare pre-
formed stabilized Pd(0) colloids and utilize these in the Heck
reaction4 both in the presence and absence of PVPy. Four
otherwise identical Heck reactions (Pd, 2.2 × 10−3 mmol;
butyl acrylate, 1.8 mmol; iodobenzene, 0.9 mmol; NEt3,
1.0 mmol; and internal GC standard) were carried out, one
in the absence of PVPy and others with 1, 30, and 100 eq of
PVPy relative to Pd metal. Each reaction mixture was kept at
120 ◦C under constant stirring and after 70 min, GC analyses
indicated conversions of iodobenzene were 95, 95, 12, and
8%, respectively. Closer observation of the kinetics of the
latter experiment indicates that the reaction proceeds quickly
to 8% conversion in the first few minutes of the reaction and
then the reaction proceeds no further. This observation in-
dicates that PVPy is not capable of completely poisoning a
system with a nanoparticle precatalyst under the conditions
used here, although a large excess of PVPy can eventually
shut down the catalytic productivity. These results neither
conclusively prove nor rule out either catalysis by the sur-
face of Pd nanoparticles or molecular Pd(0) species, as the
stoichiometry of poisons on nanoparticle surfaces can be a
complex issue [70].
Nonetheless, the above results clearly demonstrate that
soluble Pd species are the true catalysts in the Mizoroki–
Heck reaction using a silica-supported SCS-pincer Pd pre-
catalyst. However, the nature of the soluble Pd species is still
unknown.
3.2.4. Three-phase test
Although the above results argue strongly for homoge-
neous catalysis, a more convincingly set of experiments to
support this hypothesis is the use of a three-phase test. The
“three-phase” test, proposed by Rebek and co-workers [71,
72] and used by Davies et al. [66] and others [28,73,74], en-
tails anchoring one of the reagents onto a solid in addition to
using a separate solid-supported catalyst (Scheme 3). In this
way, if the Heck reactivity completely comes from the sup-
ported metal center, no transformation should be observed
for the anchored reagent, as the two solids cannot access
each other on the molecular level. In contrast, if during the
course of the reaction some active metal species are leached
into the solution, the anchored reagent will be converted to
some degree. With this in mind, we prepared a supported
analogue of iodobenzene that is covalently immobilized onto
eral studies, although to the best of our knowledge it has not been con-
clusively shown that catalysis by molecular Pd species that leach from the
nanoparticles (as in cases where supported Pd metal is used [13,26]) can be
ruled out. Pd nanoparticles immobilized on PVPy [68] or polyvinylpyrolli-
done (PVP) [24,69] spheres have been used as Heck catalysts and in these
cases, the nanoparticles are implied to be the catalysts.
4
PVP–palladium nanoparticles were synthesized according to a liter-
ature method [69]. The nanoparticles were isolated by evaporation of the
solvent followed by washing with water and ethanol and then dried under
high vacuum.

K. Yu et al. / Journal of Catalysis 226 (2004) 101–110
107
Fig. 4. FT-Raman spectra of 7 (A) and a mixture of 7, the used precatalyst 5, and the Heck coupling product (B).
Scheme 5.
the silica surface (Scheme 4). Leaching tests indicated that
3.2.5. Polymer-supported catalyst
the immobilized iodobenzene in complex 7 is all covalently
bonded to the silica surface and FT-Raman investigations al-
low for identification of all the key bands of the supported
organic (vide infra).
The results described so far indicate that a soluble
Pd species is responsible for the Heck catalysis with the
silica-supported system. However, it is unclear whether the
leached species result from pincer ligand–Pd bond rupture
or via Si–O–Si bond rupture to liberate a free Pd pin-
cer complex. To probe this, catalysis by polynorbornene-
immobilized SCS–Pd pincer species can be considered.
Polynorbornene-immobilized catalyst (8) does not contain
any Si–O–Si linkages between the pincer complex and the
support (Scheme 5). Precatalysts of this type have been
previously shown to be capable of promoting the Heck reac-
tion under similar conditions [75]. However, in the presence
of 100 eq of PVPy, Heck catalysis was completely extin-
guished. The reaction was monitored every 5 min for 1 h,
then every 8 h for 48 h. No activity was observed. The
polynorbornene-based polymer is soluble under the con-
ditions of the reaction; thus, it may be feasible for the
nonsoluble PVPy to interact with some of the supported
complexes and poison activity associated with them. How-
ever, a complete loss of activity over prolonged periods of
time is likely not consistent with activity from immobilized
sites, as extinguishing all sites with the insoluble polymer
is statistically unlikely unless the active species is homo-
geneous. Thus, these observations are consistent with Pd
leaching through decomposition of the ligand–Pd bonds.
When the Heck reaction of n-butyl acrylate (1.56 mmol)
and 7 (400 mg) using complex 5 (37 mg) as catalyst is per-
formed with NEt3 as base in DMF, GC analysis indicated
that after 3.5 h, butyl acrylate conversion was ∼ 12%, cor-
responding to ∼ 59% conversion of the anchored iodoben-
zene. The FT-Raman spectrum of the recovered solids shows
an extra Raman band at 1601 cm−1, characteristic of C=O
stretch of the coupling product, and decreased intensity of
the Ar–I stretch at 1053 cm−1, indicative of the partial con-
version of 7 to the coupling product. Another new peak
at 1637 cm−1 and its shoulders are most likely due to the
overlap from the C=C stretch of the acrylate in the cou-
pling product and the aromatic C=C stretches (the coupling
product and unreacted 7). We were not able to conclusively
attribute any bands associated with the immobilized precat-
alyst 5 in Fig. 4B, since 5 accounted for less than 10% in
the recovered solid mixture. The successful coupling reac-
tion between n-butyl acrylate and anchored iodobenzene 7
clearly indicates the presence of active soluble palladium
species, in agreement with the more traditional catalytic ex-
periments described above.

108
K. Yu et al. / Journal of Catalysis 226 (2004) 101–110
Scheme 6.
Hence, the leaching observed with the silica-supported cat-
alysts is likely also attributable to a similar leaching mecha-
nism.
a Heck reaction using the homogeneous SCS-pincer com-
plex 9. To ascertain whether Pd nanoparticles exist in so-
lution after reaction, TEMs of both the reaction solution
and the catalyst precursor were obtained. The images ver-
ified the presence of nanoparticles in the solution (particle
sizes ∼ 2–4 nm) but not in the catalyst precursor. This sup-
ports the notion that palladium comes off the pincer ligand
during the induction period and forms soluble Pd species.
How Pd(0) is generated and the nature of the soluble ac-
tive palladium species are still unknown. In the case of some
NC-palladacycle precatalysts, Beletskaya et al. observed that
prereduction of Pd(II) with a simultaneous disassembly of
a palladacycle could be effected by the olefin in a non-
catalytic Heck reaction [38,82]. In another study, Biffis et
al. demonstrated that under Heck conditions, heterogeneous
palladium metal catalysts underwent significant leaching in
the presence of aryl halides [9]. In order to ascertain which
agent(s) caused the metal leaching in the case of our silica
immobilized precatalyst 5, several control experiments were
performed. Precatalyst 5 was allowed to contact one or two
reagents in DMF at 120 ◦C for 5–30 min, and the reaction
mixture was filtered while hot under argon and any miss-
ing reagents were then added into the filtrate. The resulting
solid-free mixture was allowed to stir at 120 ◦C and kinetics
was monitored through GC analysis. We found that neither
iodobenzene nor NEt3 alone caused any palladium leaching,
while a combination of both caused significant metal leach-
ing. For example, 5 (3.6 × 10−3 mmol) was mixed with
NEt3 (2.87 mmol) and iodobenzene (2.75 mmol) in DMF
(5 ml) and allowed to stir at 120 ◦C for 30 min, at which
point, the solid was filtered off under argon. When n-butyl
acrylate (5.47 mmol) was added into the filtrate and the re-
sulting mixture was allowed to stir at 120 ◦C, GC analysis
indicated a conversion (based on iodobenzene) of 99% after
93 min. Moderate metal leaching was also observed with a
combination of NEt3 and n-butyl acrylate. While palladium
leaching in the presence of iodobenzene/NEt3 might be the
direct consequence of the oxidative addition of the organic
halide to the metal center [9,66,83], the leaching mechanism
from n-butyl acrylate/NEt3 is less straightforward. The real
nature of the active leached species, whether in the form of
colloidal nanoparticles/clusters and/or molecular palladium
species, is still unclear. Further examination is warranted.
3.2.6. Mercury poisoning test
The ability of Hg(0) to poison metal particle hetero-
geneous catalysts, by amalgamating the metal or adsorb-
ing on the metal surface, has been known for many years
[74,76,77]. Hg(0) is especially effective in poisoning Pt, Pd,
and Ni metals in forming an amalgam [76]. However, Hg(0)
is not expected to have a poisoning effect on molecular ho-
mogeneous organometallic complexes containing metals in
high oxidation states that are tightly bound by protective lig-
ands, especially pincer complexes that utilize a hypothetical
Pd(II)–Pd(IV) cycle. When a Heck reaction is performed in
the presence of excess Hg(0) using precatalyst 5, no activity
is observed even after 2 h. This provides evidence that the
SCS-pincer-type Pd(II) precatalyst decomposes under reac-
tion conditions to form soluble Pd(0) species and that Hg(0)
interactions with these Pd(0) moieties completely kill the re-
action. In a similar experiment, in the presence of excess
Hg(0), the homogeneous SCS-pincer Pd(II) complex 9 also
showed no Heck reactivity. This is consistent with a break-
down of pincer ligand–Pd bonds to liberate Pd(0) species.
3.2.7. Insights into the Heck reaction mechanism
Although a catalytic mechanism involving Pd(0) and
Pd(II) intermediates is well-established with a number of
different precatalysts, the real nature of the active species
in SCS-pincer Pd catalysis is still not clear. Indeed, the
controversial issue of a Pd(II) and Pd(IV) intermediacy
in phosphine-based palladium-catalyzed Heck reactions at-
tracted much research interest and is still under debate [31,
37,78–81]. However, there is still no definitive evidence in
the literature to support this mechanism. The results from
our investigation indicate that unsupported Pd species have
been leached into the solution (perhaps the cause of the in-
duction period), and the supported Pd metal center (those
still attached to the pincer ligand) is not active for the Heck
coupling reaction. Furthermore, the Hg test and polymer-
supported catalytic results show that the leached species are
very likely decomposition products of the pincer complex
and not simply intact pincer species leached due to Si–O–Si
bond rupture. These results support a catalytic mechanism
involving Pd(0) and Pd(II) intermediates.
3.2.8. Nature of the true catalytic species
Soluble colloidal palladium nanoparticles often impart a
reddish-orange tint in DMF [42]. The same color change
from an initial light yellow to reddish is also observed during
All of the experiments here, when taken together, support
the existence of soluble, Pd(0) species as homogeneous cat-
alysts. There is no evidence for catalysis by Pd–SCS species

K. Yu et al. / Journal of Catalysis 226 (2004) 101–110
109
under the conditions evaluated here. In considering cataly-
sis by soluble Pd(0) colloids or Pd(0) molecular species, the
evidence here supports catalysis by molecular Pd(0) species
that are present in solution in equilibrium with Pd(0) col-
loids. The color of the solution (from orange to deep red) is
consistent with the presence of Pd colloids. Furthermore, Pd
nanoparticles have been observed in solution after reaction
by TEM. In addition, after all of the substrate has been con-
sumed, Pd black is sometimes observed to form when the
reaction is carried out under argon. Prior to this, however,
the solution appears homogeneous. Although catalysis by
Pd colloids cannot be conclusively ruled out, the observation
that the insoluble PVPy polymer effectively extinguishes re-
actions (essentially completely) may be more consistent with
catalysis by molecular species than with catalysis mediated
by the surface of Pd colloids [24,68,69]. Poisoning by Hg(0)
is believed to be consistent with either catalysis by “naked”
molecular Pd(0) or Pd(0) colloids. If molecular Pd(0) species
are the true catalysts, these results are consistent with those
of de Vries et al. [25], Kohler and co-workers [12,13], and
Arai and co-workers [10,11] who have found that catalysis
by molecular Pd(0) is likely when starting with Pd(OAc)2 or
supported Pd(0) as precatalyst sources.
Thus, in many cases where Heck catalysis is reported
by “recyclable”, “recoverable” immobilized molecular cat-
alysts, it is highly likely that homogeneous species account
for some or all of the observed catalytic productivity. Espe-
cially in cases where the only test for catalyst stability that
was carried out is the capability to recover and reuse the
catalyst (and where reaction yields are reported instead of
kinetics), one needs to exercise caution in labeling the cata-
lysts as truly recoverable and recyclable systems [56,84–86].
In many cases, it is possible that these are simply recover-
able precatalyst sources that leach active species into solu-
tion in each use. For example, with the catalysts described
here, the immobilized Pd complex is simply a recoverable
Pd reservoir. After one reaction, 17% of the Pd is lost into
the solution as active Pd(0) species. Furthermore, XPS spec-
tra of the used catalyst still show the characteristic peaks for
Pd(II) species, with no additional peaks for Pd(0) or Pd(IV)
species. Thus, 83% of the Pd precatalyst sites remain after
one use and this solid can be reused until all of the Pd is
depleted from the solid. Hence, without the crucial control
and poisoning studies utilized here (and outlined by Finke
and co-workers [74] and others [28,66,73]), one cannot read-
ily conclude that a recyclable and recoverable catalyst is in
hand.
are a substantial distance away from the Pd–ligand bonds.
For example, within the SCS–Pd pincer family, two types
of catalysts have been prepared as shown in Scheme 6. The
Pd–SCS–O pincer complexes (ether linkage to the support
tether, Scheme 6a) that are described here have been con-
clusively shown to decompose to catalytically active Pd(0).
This observation is in line with earlier work by Bergbreiter
et al. on the polymer supported systems, where traces of Pd
black were observed with similar complexes [48]. Surpris-
ingly, simply changing the linkage to the support tether to an
amide may drastically alter the behavior of the complexes.
These Pd–SCS–N complexes (Scheme 6b) are claimed to be
stable under reaction conditions, yielding recoverable, recy-
clable catalysts [17,18,46–48]. Additional studies of silica-
supported versions of this Pd–SCS–N system as well as Pd–
PCP systems are currently underway in our laboratories.
4. Conclusions
In summary, SCS-pincer-type Pd(II) complexes have
been successfully immobilized onto polymeric and meso-
porous silica supports through covalent bonds. The sup-
ported pincer Pd(II) complexes were utilized as precatalysts
in the Mizoroki–Heck reaction. It is demonstrated that there
is no evidence for any heterogeneous catalysis and all the
Heck activity comes from leached soluble palladium species
that likely result from decomposition of the pincer lig-
and. Further studies concerning the nature of these leached
species and efforts toward the development of stable solid
catalyst systems are underway.
Acknowledgments
The authors thank the U.S. Department of Energy, Ba-
sic Energy Sciences, for financial support through Catalysis
Science Grant/Contract DE-FG02-03ER15459. The authors
thank Mr. Jie Diao and Professor Dennis Hess at Georgia
Tech for assistance with the XPS measurements and Profes-
sor Robert Davis of the University of Virginia for helpful
discussions.
References
[1] T.H. Bennur, A. Ramani, R. Bal, B.M. Chandra, S. Sivasankar, Catal.
Commun. 3 (2002) 493.
[2] A. Molnar, A. Papp, K. Miklos, P. Forgo, Chem. Commun. (2003)
2626.
[3] C. Mehnert, D.W. Weaver, J.Y. Ying, J. Am. Chem. Soc. 120 (1998)
12289.
[4] L. Djakovitch, K. Kohler, J. Mol. Catal. A 142 (1999) 275.
[5] R.S. Varma, K.P. Naiker, P.J. Liesen, Tetrahedron Lett. 40 (1999) 2075.
[6] A. Corma, H. Garcia, A. Leyva, A. Primo, Appl. Catal. A 257 (2004)
77.
[7] M. Dams, L. Drijkoningen, B. Pauwels, G. Van Tendeloo, D.E. De Vos,
P.A. Jacobs, J. Catal. 209 (2002) 225.
3.2.9. Implications for Heck catalysis with other pincer
complexes
Other studies of immobilized PCP [56], SCS [17,18,46–
48], and NCN [49] Pd pincer complexes have implied that
these complexes are stable under reaction conditions and
thus the supported catalysts are recyclable. It is interesting
that rather subtle changes of ligand structure may drastically
change the behavior of these complexes, even changes that

110
K. Yu et al. / Journal of Catalysis 226 (2004) 101–110
[8] A. Nejjar, C. Pinel, L. Djakovitch, Adv. Synth. Catal. 345 (2003) 612.
[9] A. Biffis, M. Zecca, M. Basato, Eur. J. Inorg. Chem. 5 (2001) 1131.
[47] D.E. Bergbreiter, P.L. Osburn, J.D. Frels, J. Am. Chem. Soc. 123
(2001) 11105.
[48] D.E. Bergbreiter, P.L. Osburn, Y.-S. Liu, J. Am. Chem. Soc. 121
(1999) 9531.
[10] F. Zhao, B.M. Bhanage, M. Shirai, M. Arai, Chem. Eur. J. 6 (2000)
843.
[49] H.P. Dijkstra, N. Ronde, G.P.M. Van Klink, D. Vogt, G. Van Koten,
Adv. Synth. Catal. 345 (2003) 364.
[50] E. Peris, J.A. Loch, J. Mata, R.H. Crabtree, Chem. Commun. (2001)
201.
[51] S. Sjovall, O.F. Wendt, C. Andersson, J. Chem. Soc., Dalton Trans.
(2002) 1396.
[52] E. Diez-Barra, J. Guerra, V. Hornillos, S. Merino, J. Tejeda, Organo-
metallics 22 (2003) 4610.
[11] F. Zhao, B.M. Bhanage, M. Shirai, M. Arai, J. Catal. 194 (2000) 479.
[12] R.G. Heidenreich, J.G.E. Krauter, J. Pietsch, K. Kohler, J. Mol. Catal.
A 182–183 (2002) 499.
[13] K. Kohler, R.G. Heidenreich, J.G.E. Krauter, J. Pietsch, Chem.
Eur. J. 8 (2002) 622.
[14] K. Kohler, M. Wagner, L. Djakovitch, Catal. Today 66 (2001) 105.
[15] B.M. Choudary, S. Madhi, N.S. Chowdari, M.L. Kantam, B. Sreedhar,
J. Am. Chem. Soc. 124 (2002) 14127.
[53] R.A. Gossage, L.A. van de Kuil, G. Van Koten, Acc. Chem. Res. 31
(1998) 423.
[54] A. Weisman, M. Gozin, H.-B. Kraatz, D. Milstein, Inorg. Chem. 35
(1996) 1792, and references cited therein.
[16] R.L. Augustine, S.T. O’Leary, J. Mol. Catal. A 95 (1995) 277.
[17] D.E. Bergbreiter, Chem. Rev. 102 (2002) 3345.
[18] D.E. Bergbreiter, P.L. Osburn, A. Wilson, E.M. Sink, J. Am. Chem.
Soc. 122 (2000) 9058.
[55] B.M.J.M. Suijkerbuijk, L. Shu, R.J.M. Klein Gebbink, A.D. Schlüter,
G. Van Koten, Organometallics 22 (2003) 4175.
[56] R. Chanthateyanonth, H.A. Alper, J. Mol. Catal. A 201 (2003) 23.
[57] M. Poyatos, F. Marquez, E. Peris, C. Claver, E. Fernandez, New J.
Chem. 27 (2003) 425.
[19] H.P. Dijkstra, M.Q. Slagt, A. McDonald, C.A. Kruithof, R. Kreiter,
A.M. Mills, M. Lutz, A.L. Speck, W. Klopper, G.P.M. Van Klink,
G. Van Koten, Eur. J. Inorg. Chem. (2003) 830.
[20] R. Sablong, U. Schlotterback, D. Vogt, S. Mecking, Adv. Synth.
Catal. 345 (2003) 333.
[58] J. Schwartz, Acc. Chem. Res. 18 (1985) 302.
[59] Y. Lin, R.G. Finke, Inorg. Chem. 33 (1994) 4891.
[60] J.D. Aiken, Y. Lin, R.G. Finke, J. Mol. Catal. A 114 (1996) 29.
[61] D. Zhao, Q. Hou, J. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem.
Soc. 120 (1998) 6024.
[62] J.M. Pollino, M. Weck, Synthesis 9 (2002) 1277.
[63] W.T.S. Huck, F.C.J.M. van Veggel, D.N. Reinhoudt, J. Mater. Chem. 7
(1997) 1213.
[64] J.M. Pollino, L.P. Stubbs, M. Weck, Macromolecules 36 (2003) 2230.
[65] S.M. Islam, B.K. Palit, D.K. Mukherjee, C.R. Saha, J. Mol. Catal.
A 124 (1997) 5.
[66] I.W. Davies, L. Matty, D.L. Hughes, P.J. Reider, J. Am. Chem.
Soc. 123 (2001) 10139.
[67] F.R. Hartley, Coord. Chem. Rev. 35 (1981) 143.
[68] S.P.M.T. Greci, R.C. Kwong, K. Mercado, G.K.S. Prakash, G.A. Olah,
M.E. Thompson, Chem. Mater. 12 (2000) 1985.
[69] R. Narayanan, M.A. El-Sayed, J. Am. Chem. Soc. 125 (2003) 8340.
[70] B.J. Hornstein, J.D. Aiken, R.G. Finke, Inorg. Chem. 41 (2002) 1625.
[71] J. Rebek, F. Gavina, J. Am. Chem. Soc. 96 (1974) 7112.
[72] J. Rebek, D. Brown, S. Zimmerman, J. Am. Chem. Soc. 97 (1975) 454.
[73] C. Baleizão, A. Corma, H. Garcia, A. Leyva, J. Org. Chem. 69 (2004)
439.
[74] J.A. Widegren, R.G. Finke, J. Mol. Catal. A: Chem. 198 (2003) 317.
[75] J.M. Pollino, M. Weck, Org. Lett. 4 (2002) 753.
[76] G.M. Whitesides, M. Hackett, R.L. Brainard, J.P.P.M. Lavalleye,
A.F. Sowinski, A.N. Izumi, S.S. Moore, D.W. Brown, E.M. Staudt,
Organometallics 4 (1985) 1819.
[21] E.H. Rahim, F.S. Kamounah, J. Frederiksen, J.B. Christensen, Nano
Lett. 1 (2001) 499.
[22] L.K. Yeung, R.M. Crooks, Nano Lett. 1 (2001) 14.
[23] K.P. Gopidas, J.K. Whitesell, M.A. Fox, Nano Lett. 3 (2003) 1757.
[24] J. Le Bars, U. Specht, J.S. Bradley, D.G. Blackmond, Langmuir 15
(1999) 7621.
[25] A.H.M. de Vries, J.M.C.A. Mulders, J.H.M. Mommers, H.J.W. Hen-
drickx, J.G. de Vries, Org. Lett. 5 (2003) 3285.
[26] D.A. Conlon, B. Pipik, S. Ferdinand, C.R. Leblond, J.R. Sowa Jr.,
B. Izzo, P. Collins, G.-J. Ho, J.M. Williams, Y.-J. Shi, Y. Sun, Adv.
Synth. Catal. 345 (2003) 931.
[27] J.A. Widegren, R.G. Finke, J. Mol. Catal. A 198 (2003) 317.
[28] B.H. Lipshutz, S. Tasler, W. Chrisman, B. Spliethoff, B. Tesche, J. Org.
Chem. 68 (2003) 1177.
[29] J. DuPont, M. Pfeffer, J. Spencer, Eur. J. Inorg. Chem. (2001) 1917.
[30] W.A. Hermann, V.P.W. Bohm, C.-P. Reisinger, J. Organomet.
Chem. 576 (1999) 23.
[31] B.L. Shaw, S.D. Perera, E.A. Staley, Chem. Commun. (1998) 1361.
[32] A.S. Gruber, D. Zim, G. Ebeling, A.L. Monteiro, J. DuPont, Org.
Lett. 2 (2000) 1287.
[33] M. Beller, T.H. Riermeier, Eur. J. Inorg. Chem. (1998) 29.
[34] D. Morales-Morales, C. Grause, K. Kasaoka, R. Redon, R.E. Cramer,
C.M. Jensen, Inorg. Chim. Acta 300–302 (2000) 958.
[35] A. Sundermann, O. Uzan, J.M.L. Martin, Chem. Eur. J. 7 (2001) 1703.
[36] V.P.W. Bohm, W.A. Hermann, Chem. Eur. J. 7 (2001) 4191.
[37] B.L. Shaw, New J. Chem. 22 (1998) 77.
[38] I.P. Beletskaya, A.N. Kashin, N.B. Karlstedt, A.V. Mitin, A.V. Chep-
rakov, G.M. Kazankoc, J. Organomet. Chem. 622 (2001) 89.
[39] W.A. Hermann, K. Ofele, D.V. Preysing, S.K. Schneider, J. Organo-
met. Chem. 687 (2003) 229.
[40] C.S. Consorti, M.L. Zainini, S. Leal, G. Ebeling, J. DuPont, Org.
Lett. 5 (2003) 983.
[41] C. Rocaboy, J.A. Gladysz, New J. Chem. 27 (2003) 39.
[42] C. Rocaboy, J.A. Gladysz, Org. Lett. 4 (2002) 1993.
[43] R.B. Bedford, C.S.J. Cazin, M.B. Hursthouse, M.E. Light, K.J. Pike,
S. Wimperis, J. Organomet. Chem. 633 (2001) 173.
[44] J.A. Loch, M. Albrecht, E. Peris, J. Mata, J.W. Faller, R.H. Crabtree,
Organometallics 21 (2002) 700.
[77] C. Paal, W. Hartmann, Chem. Ber. 51 (1918) 711.
[78] D. Morale-Morales, R. Redon, C. Young, C.M. Jensen, Chem. Com-
mun. (2000) 1619.
[79] F. Miyazaki, K. Yamaguchi, M. Shibasaki, Tetrahedron Lett. 40 (1999)
7379.
[80] M.E. van der Boom, D. Milstein, Chem. Rev. 103 (2003) 1759.
[81] M. Ohff, A. Ohff, M.E. van der Boom, D. Milstein, J. Am. Chem.
Soc. 119 (1997) 11687.
[82] J. Tsuji, Acc. Chem. Res. 2 (1969) 144.
[83] Q. Wang, H. Liu, M. Han, X. Li, D. Jiang, J. Mol. Catal. A 118 (1997)
145.
[84] C. Venkatesan, A.P. Singh, Catal. Lett. 88 (2003) 193.
[85] S.F. Zhao, R.X. Zhou, X.M. Zheng, J. Mol. Catal. A 211 (2003) 139.
[86] M. Lagasi, P. Moggi, J. Mol. Catal. A 182–183 (2002) 61.
[45] M.R. Eberhard, Z. Wang, Org. Lett. (2004), in press.
[46] D.E. Bergbreiter, J. Li, Chem. Commun. (2004) 42.