Insights into the heck reaction with PCP pincer palladium(II) complexes

Article abstract of DOI:10.1021/ol049430a

The Heck reaction of phenyl halides with styrene using a series of related PCP pincer palladium(II) complexes was studied in order to evaluate the effect of ligand structure and electronics on the catalytic activity and to investigate the nature of the catalyst species. We suggest these pincer complexes are precatalysts for highly active forms of metallic palladium. This conclusion is based on kinetic studies (induction periods, sigmoidal kinetics), Hg drop tests, quantitative poisoning experiments, and NMR studies.

Full text of DOI:10.1021/ol049430a

ORGANIC
LETTERS
2004
Vol. 6, No. 13
2125-2128
Insights into the Heck Reaction with
PCP Pincer Palladium(II) Complexes
Michael R. Eberhard
Van’t Hoff Institute for Molecular Science, UniVersity of Amsterdam,
Nieuwe Achtergracht 166, 1018 WV Amsterdam, The Netherlands
chmre@science.uVa.nl
Received March 26, 2004
ABSTRACT
The Heck reaction of phenyl halides with styrene using a series of related PCP pincer palladium(II) complexes was studied in order to evaluate
the effect of ligand structure and electronics on the catalytic activity and to investigate the nature of the catalyst species. We suggest these
pincer complexes are precatalysts for highly active forms of metallic palladium. This conclusion is based on kinetic studies (induction periods,
sigmoidal kinetics), Hg drop tests, quantitative poisoning experiments, and NMR studies.
One of the success stories of PCP pincer complexes is their
application in the C-C coupling of aryl halides with alkenes
in the presence of base, namely, the Heck reaction.1,2 Milstein
et al. first reported the reaction of a series of aryl iodides
and bromides with olefin using PCP pincer palladium
complexes such as [PdTFA{(C6H3)(CH2)-1-(CH2PPri2)2-
2,6}].3 Shibasaki et al. obtained extraordinarily high TONs
of up to 8 900 000 for the reaction of iodobenzene with
n-butyl acrylate using bis(phosphito) pincer complex [PdI-
{(C6H3)(OP(OC6H4OMe-4)2)2-2,6}].4 High activity for form-
ing trisubstituted olefins from acrylates and aryl halides,5 as
well as a unique activity for the conversion of aryl chlorides
with styrene to trans-stilbenes,6 was observed with [PdCl-
{(C6H3)(OPPri2)2-2,6}] 1.
attributed this to electronic effects, as well as a larger
P-Pd-P angle compared to 1. With this in mind, and since
complexes 1-4 are closely related, we became interested in
evaluating the effect of ligand structure and electronics on
the activity in the Heck reaction between iodo-, bromo-, and
chlorobenzene with styrene and Cs2CO3 (eq 1).
The chemistry of palladacycles and their application in
homogeneous catalysis has been reviewed.8 Several research
groups,9 who have recently investigated palladacycles for
C-C coupling reactions, reported that in many cases neither
the palladacycles nor the pincer palladium complexes are
(3) Ohff, M.; Ohff, A.; van der Boom, M. E.; Milstein, D. J. Am. Chem.
Soc. 1997, 119, 11687.
(4) Miyazaki, F.; Yamaguchi, K.; Shibasaki, M. Tetrahedron Lett. 1999,
40, 7379.
(5) Morales-Morales, D.; Grause, C.; Kasaoka, K.; Redo´n, R.; Cramer,
R. E.; Jensen, C. M. Inorg. Chim. Acta 2000, 300-302, 958.
(6) Morales-Morales, D.; Redo´n, R.; Yung, C.; Jensen, C. M. Chem.
Commun. 2000, 1619.
(7) (a) Wang, Z.; Eberhard, M. R.; Jensen, C. M.; Matsukawa, S.;
Yamamoto, Y. J. Organomet. Chem. 2003, 681, 189. (b) Eberhard, M. R.;
Matsukawa, S.; Yamamoto, Y.; Jensen, C. M. J. Organomet. Chem. 2003,
687, 185.
We previously reported on the synthesis of complexes
2-4,7 as well as compared 1 and 3 in allylic alkylation
catalysis.7a We found that 3 was more active than 1 and
(1) Beletskaya, I. P.; Cheprakov, A. V. Chem. ReV. 2000, 100, 3009.
(2) van der Boom, M. E.; Milstein, D. Chem. ReV. 2003, 103, 1759.
(8) (a) Dupont, J.; Pfeffer, M.; Spencer, J. Eur. J. Inorg. Chem. 2001,
1917. (b) Bedford, R. B. Chem. Commun. 2003, 1787.
10.1021/ol049430a CCC: $27.50 © 2004 American Chemical Society
Published on Web 04/24/2004

the true catalysts but instead palladium nanoparticles or low-
ligated palladium species. Nowotny et al. presented kinetic-
based evidence that cyclopalladated imine catalysts were
thermally labile generating finely divided metallic palladium,
which was active in the Heck coupling of iodobenzene with
styrene.9a Fluorous imine palladacycles synthesized by Ro-
caboy and Gladysz were found to be precursors for highly
active Heck and Suzuki catalysts.9b On the basis of kinetics
and TEM, these palladacycles release catalytically highly
active soluble colloidal palladium nanoparticles. A series of
NC-palladacycles were reported by Beletskaya et al. as very
effective cheap precursors for the phosphine-free Heck
reaction.9c The authors proposed a noncatalytic Heck reaction
that leads to the release of catalytically active low-ligated
palladium(0). Dupont et al. showed that chloropalladated
propargylamine was a simple and efficient catalyst precursor
for the Heck reaction.9d Although the authors were not able
to characterize the proposed catalytically active palladium
nanoparticles by TEM, they used the Hg drop test as a means
to support their hypothesis. Bedford et al. found that imine-
based palladacycles liberate Pd(0) species that are active in
the Suzuki reaction.9e This was supported by a mechanistic
study using GC-MS analysis. Most recently, de Vries et al.
presented kinetic- and MS-based evidence showing that the
Herrmann-Beller palladacycle behaves similarly to ligand-
free Pd(OAc)2 in Heck catalysis.9f They came to the
conclusion that the palladacycle merely serves as a source
of ligand-free monomeric palladium in Heck reactions of
aryl bromides.
sigmoidal-shaped kinetics with induction periods between
15 and 45 min.
After 6 h at 100 °C, complex 1 gave 72% and 3 gave
61% conversion for bromobenzene/styrene (Figure 2). Sys-
Figure 2. Conversion vs time: PhBr/styrene/Cs2CO3 at 100 °C
with complexes 1-4 in anhydrous dioxane under argon.
tems with complexes 2 and 4 show low activity (8 and 5%
yields, respectively). We observe again an induction period
of 15 min for systems using 1 and 3 and S-shaped kinetic
curves.
For the chlorobenzene/styrene system at 180 °C (Figure
3), complex 1 reaches 65% and 3 achieves 43% conversion
We would now like to report our results regarding the
structure-activity study, as well as present observations that
complexes 1-4 may be sources for catalytically active
metallic palladium based on kinetic studies (induction
periods, sigmoidal kinetics), Hg drop tests, quantitative
poisoning experiments, and NMR studies.
Figure 1 shows the result for the iodobenzene/styrene
systems, with complex 3 giving 55% conversion and complex
Figure 3. Conversion vs time: PhCl/styrene/Cs2CO3 at 180 °C
with complexes 1-4 in anhydrous dioxane under argon.
within the first 30 min. After 6 h, 75 and 63% conversions
are obtained, respectively. These highly active systems do
not show induction periods. Complex 2 shows an induction
period of ca. 15 min and then a gradual increase in
Figure 1. Conversion vs time: PhI/styrene/Cs2CO3 with complexes
1-4 in anhydrous dioxane under argon.
(9) (a) Nowotny, M.; Hanefeld, U.; van Koningsveld, H.; Maschmeyer,
T. Chem. Commun. 2000, 1877. (b) Rocaboy, C.; Gladysz, J. A. New J.
Chem. 2003, 27, 39. (c) Beletskaya, I. P.; Kashin, A. N.; Karlstedt, N. B.;
Mitin, A. V.; Cheprakov, A. V.; Kazankov, G. M. J. Organomet. Chem.
2001, 622, 89. (d) Consorti, C. S.; Zanini, M. L.; Leal, S.; Ebeling, G.;
Dupont, J. Org. Lett. 2003, 5, 983. (e) Bedford, R. B.; Cazin, C. S. J.;
Hursthouse, M. B.; Light, M. E.; Pike, K. J. Wimperis, S. J. Organomet.
Chem. 2001, 633, 173 (f) de Vries, A. H. M.; Mulders, J. M. C. A.;
Mommers, J. H. M.; Henderickx, H. J. W.; de Vries, J. G. Org. Lett. 2003,
5, 3285.
1 yielding 26% product after 4 h at 100 °C, while complexes
2 and 4 achieved 54 and 25% conversion after 4 h at 150
°C, respectively. At 100 °C, systems with complexes 2 and
4 showed little activity. Experiments with 1, 2, and 4 exhibit
2126
Org. Lett., Vol. 6, No. 13, 2004

conversion, rising to 58% conversion after 6 h. Complex 4
is inactive toward chlorobenzene (only 5% yield after 6 h).
Deactivation occurs for systems with 1 and 3 after 30 min,
which is substantiated by the observation of Pd black.
Induction periods, S-shaped kinetic curves, and formation
of Pd black suggest that complexes 1-4 are precatalysts and
that palladium(0), possibly in the form of palladium nano-
particles, may be the active catalytic species. We tested this
hypothesis as follows.
First, we subjected all 12 experiments to mercury drop
tests.10 Crabtree10a and Whitesides10b found that adding an
excess of metallic mercury (with respect to the metal com-
plex) to the reaction mixture will lead to the amalgamation
of the surface of a heterogeneous metal particle, thus
poisoning it, but will not affect a homogeneous catalyst.
When we added up to 300 equiv of Hg(0) (relative to the
complex used) to the reaction mixtures at t ) 0 min, no
catalytic activity at all was observed in any of the 12
experiments.11,12 Figure 4 further illustrates these observations
This method has been described in detail in a review by
Widegren and Finke.13 The authors state that if a catalyst
can be poisoned completely with ,1 equiv of the added
ligand (per metal atom), then this is highly suggestive
(kinetic-based) evidence of a heterogeneous catalyst. A het-
erogeneous catalyst only has a fraction of the metal atoms
on the surface; hence, even if every surface atom is active,
,1 equiv of ligand will be sufficient to poison the catalyst.
Ligands suitable for quantitative poisoning experiments are,
for example, CS2, thiophene, or PPh3.
At t ) 15 min, catalytic activity of system PhCl/styrene/
Cs2CO3/180 °C with complex 1 was effectively stopped when
any of the following four ligands was added: 0.5 equiv of
CS2, 0.1 or 0.3 equiv of thiophene, or 0.03 equiv of PPh3
(relative to complex 1) (Figure 5a-d, respectively).14
Figure 5. Conversion vs time: PhCl/styrene/Cs2CO3 at 180 °C
with complex 1 (a-d) and 3 (e): (a) 0.5 equiv of CS2 at 15 min;
(b) 0.1 equiv of thiophene at 15 min; (c) 0.3 equiv of thiophene at
15 min; (d) 0.03 equiv of PPh3 at 15 min; (e) 0.03 equiv of PPh3
at 15 min.
Figure 4. Conversion vs time: (a) PhCl/styrene/Cs2CO3/180 °C/1
or 3; Hg at 15 min; (b) PhI/styrene/Cs2CO3/150 °C/2 or 4; Hg at
75 min.
Similarly, for the system PhCl/styrene/Cs2CO3/180 °C with
complex 3, 0.03 equiv of PPh3 only (relative to complex 3)
added at t ) 15 min was sufficient to suppress the catalytic
activity (Figure 5e).
When 0.03, 0.5, 1, 2, or 4 equiv of PPh3, respectively,
were added at t ) 0 min to PhCl/styrene/Cs2CO3/180 °C
with complex 1, only 10, 7, 5, 3, and 2% conversions to
trans-stilbene were observed, respectively. 31P{1H} NMR
spectroscopy also showed the formation of Ph3PO.
for 4 of the 12 systems. Metallic mercury was added to a
reaction mixture at a certain time after the reaction had
started, and a catalytically active system had formed. For
systems with 1 or 3 and PhCl/styrene/Cs2CO3/180 °C, we
added Hg(0) at 15 min; for systems with 2 or 4 and
PhI/styrene/Cs2CO3/150 °C, we added Hg(0) at 75 min. All
four catalytic reactions were suppressed instantly.
(13) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A 2003, 198, 317. See
also: Lin, Y.; Finke, R. G. Inorg. Chem. 1994, 33, 4891. Weddle, K. S.;
Aiken, J. D., III; Finke, R. G. J. Am. Chem. Soc. 1998, 120, 5653.
(14) It is important to note that quantitative poisoning experiments
reported in the literature are usually conducted at temperatures <50 °C,
whereas our experiments are conducted at 180 °C. In some reported cases,
higher temperatures lead to dissociation of the ligand (poison) from the
heterogeneous catalyst (thus regenerating a catalytically active system).
However, in our case, the addition of CS2, PPh3, and thiophene and
conducting the experiment at 180 °C is effective in poisoning the
catalytically active system. Catalysis ceases, and the observed conversions
are approximately 25-30% lower for 1 and 45% lower for 3 (Figure 5) as
compared to an unpoisoned system where the conversion reaches 73% for
1 and 60% for 3 after 4 h (Figure 3). One possible explanation is that in
our case, these ligands do not dissociate from the palladium particle even
under these forcing conditions or, alternatively, that addition of these ligands
accelerates growth of the metal particles, resulting in faster production of
inactive Pd black.
Detailed quantitative poisoning experiments were con-
ducted for systems with 1 or 3 and chlorobenzene/styrene.
(10) (a) Anton, D. R.; Crabtree, R. H. Organometallics 1983, 2, 855.
(b) Foley, P.; DiCosimo, R.; Whitesides, G. M. J. Am. Chem. Soc. 1980,
102, 6713
(11) It is interesting to note in this context that Crabtree reported that
the Hg(0) drop test is negative for pincer CNC (C ) carbene) palladium
complexes used in Heck catalysis. The stability of the catalyst was attributed
to the great strength of the carbene-palladium bond. Peris, E.; Loch, J. A.;
Mata, J.; Crabtree, R. H. Chem. Commun. 2001, 201.
(12) In most cases, Hg(0) will not affect a homogeneous catalyst but
forms an amalgam with a heterogeneous catalyst thereby poisoning it.
However, it cannot be completely excluded that Hg(0) interferes with
catalysis in a way other than killing the true heterogeneous catalyst. Hg(0)
would possibly also react with a PCP Pd(0) pincer complex (which to the
best of our knowledge is not known yet).
Org. Lett., Vol. 6, No. 13, 2004
2127

Furthermore, when we investigated the residues of all
catalysis experiments by 31P{1H} NMR spectroscopy, we
found that none of the experiments had the original complex
or a species related to it present.15 In all experiments, we
observed species that resonate between 45 and 55 ppm, the
major species giving a signal at 51.0 ppm. We believe that
these species are phosphine oxides. The shift at 51.0 ppm is
reminiscent of the reported shift for Pri2PhP(O) at 51.4 ppm
in CDCl3.16 GC-MS of the catalysis samples also shows a
species with m/z 210. M+ for Pri2PhP(O) is 210.16
NMR studies) forming Pd black and a phosphine oxide
(similar shift to the species observed in the catalytic residue).
Bis(phosphinito) pincer complexes 1 and 3 decompose
rapidly, whereas 2 and 4 are more stable and show slower
signs of decomposition.
At this point, we cannot completely exclude that the pincer
palladium(II) complexes studied here catalyze in part the
Heck reaction. However, all observations combined, i.e.,
induction periods, S-shaped kinetic curves, formation of Pd
black, poisoning experiments (Hg drop tests, quantitative
poisoning with ,1 equiv of CS2, thiophene, and PPh3), and
NMR studies, suggest that metallic palladium(0) plays a
significant role in this catalysis. The pincer palladium
complexes serve as precursors for the true catalytic species.
Complexes 1 and 3, possessing only phosphinito donors,
decompose more easily and therefore lead to more active
sources of metallic palladium than either complexes 2 or 4.
The introduction of a But2P group leads to more stable
complexes but less active catalyst precursors.
It has been found by other researchers that formation of
phosphine oxides is an integral part in the prereduction/-
activation stage of the Heck catalysis with Pd(II) and
phosphines leading to a Pd(II)/Pd(0) catalytic cycle.1 The
reduction of Pd(II) to Pd(0) occurs with relative ease in the
presence of phosphines and oxygenated ligands such as
2-
NO3-,17a AcO-,17b oxide, or OH-.17c It is possible that CO3
may act in a similar way, but an alternative explanation is
that adventitious traces of H2O and OH- decompose the
pincer complex.18
Acknowledgment. Experiments were carried out by
Zhaohui Wang at the University of Hawaii. We thank Dr.
Gadi Rothenberg for valuable comments.
The pincer complexes 1-4 are otherwise stable for
prolonged periods of time in solution up to 200 °C; thus,
their decomposition must be triggered by one of the reactants.
Indeed, experiments show that these complexes are also
stable (in dioxane/150-180 °C) in the presence of phenyl
halides, styrene, or Hg(0) (or any combination thereof) but
decompose in the presence of Cs2CO3 (indicated by 31P{1H}
Note Added after ASAP. Zhaohui Wang was deleted
from the author list of the version posted ASAP on April
24, 2004; a corrected version including a sentence in the
Acknowledgment indicating his contribution to the paper was
posted on May 21, 2004.
(15) 31P{1H} NMR shifts for complexes 1-4 and derivatives are between
150 and 200 ppm for -OPPri2 and 70-80 ppm for -PBut2.5,7
(16) Keglevich, G.; Tama´s, A.; Parlagh, G. To¨ke, L. Heteroatom Chem.
2001, 12, 38.
(17) (a) Malatesta, L.; Angoletta, M. J. Chem. Soc. 1957, 1186. (b)
Amatore, C.; Jutand, A. Organometallics 1992, 11, 3009. (c) Grushin, V.
V.; Alper, H. Organometallics 1993, 12, 1890.
Supporting Information Available: Experimental pro-
cedures. This material is available free of charge via the
Internet at http://pubs.acs.org.
(18) Although every precaution was taken to exclude H2O from the
catalytic systems, traces may still be present.
OL049430A
2128
Org. Lett., Vol. 6, No. 13, 2004