Heterogeneous or homogeneous? A case study involving palladium-containing perovskites in the Suzuki reaction

Article abstract of DOI:10.1002/adsc.200404331

The utility of a series of palladium-containing perovskite catalysts in the Suzuki reaction is described; turnover numbers of up to 400,000 are reported. A detailed investigation into the mode of action of these catalysts encompassing kinetic studies, catalyst poisoning, microscopy and three-phase tests demonstrate that these heterogeneous materials are pre-catalysts that operate by a solution-phase mechanism.

Full text of DOI:10.1002/adsc.200404331

FULL PAPERS
Heterogeneous or Homogeneous? A Case Study Involving
Palladium-Containing Perovskites in the Suzuki Reaction
Stephen P. Andrews,a Antonia F. Stepan,a Hirohisa Tanaka,b Steven V. Ley,a
Martin D. Smitha,*
a
Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, CB2 1EW, UK
E-mail: mds44@cam.ac.uk
b
Materials R&D Division, Technical Centre, Daihatsu Motor Co. Ltd. Ryuo, Gamo-gun, Shiga 520-2593, Japan
Received: October 27, 2004; Accepted: January 28, 2005
Supporting Information for this article is available on the WWW under http://asc.wiley-vch.de or from the author.
Abstract: The utility of a series of palladium-contain- strate that these heterogeneous materials are pre-cat-
ing perovskite catalysts in the Suzuki reaction is de- alysts that operate by a solution-phase mechanism
scribed; turnover numbers of up to 400,000 are report-
ed. A detailed investigation into the mode of action of Keywords: cross-coupling; heterogeneous catalysis;
these catalysts encompassing kinetic studies, catalyst homogeneous catalysis; palladium; perovskite; Suzuki
poisoning, microscopy and three-phase tests demon- reaction
Introduction
The utility of a palladium-containing perovskite
LaFe0.57Co0.38Pd0.05O3 (LaPd*, a catalyst originally de-
ꢀHomogeneousꢁ[1] palladium catalysts are widely used in veloped for automotive emission control)[29] in the Suzu-
many cross-coupling reactions, and systems capable of ki cross-coupling,[30] has been disclosed earlier.[31] In this
demonstrating astonishingly high turnover numbers paper, investigations into the reactivity of a series of Pd-
for deactivated substrates at room temperature have containing perovskites in the Suzuki reaction are out-
been disclosed.[2–7] Whilst these techniques are certainly lined, and a range of experiments designed to ascertain
state-of-the-art in terms of their substrate tolerance, cat- whether this process proceeds by way of a ꢀhomogene-
alyst loading and reactivity, their complexity may pre- ousꢁ or a ꢀheterogeneousꢁ mechanism are delineated.
clude application on a large scale. Additionally, the re-
moval of residual palladium and associated ligands
from these reactions is often difficult, and as such, effec- Results and Discussion
tive and economical catalysts free from these constraints
are essential for the translation of laboratory methods to It was considered important to investigate whether the
larger scale operations in industry.[8] ꢀHeterogeneousꢁ identity of the metals other than palladium in the perov-
catalysts can be removed from reaction media by a sim- skite structure was an important influence on reactivity
ple filtration, but relatively few examples of successful and turnover number in cross-coupling reactions. With
and highly active ꢀheterogeneousꢁ catalysts for use in this in mind, seven palladium-containing perovskites
cross-coupling reactions have been reported. A range were screened for their activity in the Suzuki reaction
of approaches has been adopted in an attempt to solve of 4-bromoanisole with phenylboronic acid under our
this problem including encapsulation[9–11] or incarcera- previously reported conditions (Scheme 1).
tion,[12,13] the use of polymeric supports,[14–16] other in-
The kinetic profiles of all seven catalysts are pictured
soluble palladium sources such as Pd/C[17–20] and transi- in Figure 1.
tion metal-free cross-coupling reactions.[21,22]As such,
the search for an efficient, cheap and practical ꢀhetero-
geneousꢁ catalytic system is a major challenge. However,
desorption of the transition metal from the ꢀheterogene-
ousꢁ support can occur, resulting in its entry into a con-
ventional solution-phase catalytic cycle.[23–25] This has
prompted investigations into the nature of the catalyti-
cally active species in these and other systems.[26–28]
Scheme 1. Suzuki reaction.
Adv. Synth. Catal. 2005, 347, 647–654
DOI: 10.1002/adsc.200404331
ꢂ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
647

Stephen P. Andrews et al.
FULL PAPERS
bromide.[25,31] In an attempt to clarify, and further inves-
tigate, this phenomenon a series of mechanistic investi-
gations was undertaken. Distinguishing whether a cata-
lyst operates by a homogeneous or a heterogeneous
mechanism is non-trivial, and it has been demonstrated
in a series of elegant studies [35–37] that a combination of
kinetics, poisoning studies, microscopy and other meth-
ods offer the best prospect of making a definitive judg-
ment. Taking this into account, it was decided to initiate
these studies with an examination of the kinetics of the
coupling of 4-bromoanisole and phenylboronic acid.
Kinetic Experiments
Figure 1. Kinetic profiles of palladium-containing perovskite
catalysts in the Suzuki reaction of phenylboronic acid with
4-bromoanisole.
Kinetic studies are one of the most effective methods of
investigating whether a catalyst is actually a precatalyst
as such systems often display sigmoidal kinetics with
an identifiable induction period. Monitoring the cou-
All seven chosen catalysts are effective in this Suzuki pling of 4-bromoanisole and phenylboronic acid under
reaction, with relatively little difference between them, the standard Suzuki conditions (sampling every 20 sec-
although the La0.9Ce0.1Fe0.57Co0.38Pd0.05O3-catalysed re- onds for the first four minutes) demonstrated that there
action went to completion in the shortest period of was a short, but measurable induction period of approx-
time. Attempts to perform the reaction with perovskites imately four minutes, Figure 2.
deficient in palladium (such as LaCe0.1Fe0.6Co0.4O3,
During this period the reaction rate was considerably
LaFeO3, LaCoO3, or LaFe0.6Co0.4O3) were unsuccessful. lower than it was in the following time of approximately
It was also important to establish the lifetimes of the cat- one hour. This induction period probably correlates
alysts, and so turnover numbers were measured for all with production of a solution-phase Pd species, which
seven catalysts for the same reaction, Table 1.
may occur during an oxidative insertion step. It has
This demonstrated that turnover numbers varied from been shown (using a three-phase test)[25] that the pres-
151,000 (for La0.9Ce0.1Al0.95Pd0.05O3) to almost 400,000 ence of an aryl halide in solution is necessary to generate
for LaFe0.95Pd0.05O3. At this stage it was decided a solution-phase active catalyst.
to concentrate our studies on a single catalyst –
During its preparation, LaFe0.57Co0.38Pd0.05O3 is heat-
LaFe0.57Co0.38Pd0.05O3 – as this is prepared on an industri- ed in oxygen at 8008C to afford the final perovskite cat-
al scale, and is also the catalyst for which our generic Su- alyst containing PdIII. It has been demonstrated that un-
zuki conditions have been developed. In our previous der the conditions of the combustion engine exhaust sys-
paper a three-phase test[33,34] indicated that the reaction tem (that operates over 8008C) the palladium reversibly
was proceeding via the intermediacy of a soluble cata- moves out of the perovskite lattice to form a face-cen-
lyst, generated in the presence of a solution-phase aryl
Table 1. Turnover numbers for palladium-containing perovs-
kites.
Catalyst
TON (Pdꢀ1
)
LaFe0.57Co0.38Pd0.05O3
La0.9Ce0.1Fe0.57Co0.38Pd0.05O3
LaFe0.95Pd0.05O3
243,000[32]
351,000
399,000
213,000
151,000
177,000
190,000
LaCo0.95Pd0.05O3
La0.9Ce0.1Al0.95Pd0.05O3
LaFe0.57Mn0.38Pd0.05O3
LaMn0.95Pd0.05O3
Substrate/catalyst ratio¼27000: 1 (bromoanisole:perov-
skite)¼0.00019 mol% Pd. The reactions were run until
LCMS conversion did not increase by more than 1% over
24 h (typically approx. 2 weeks). At this point conversions
were estimated based on 1H NMR analysis and TON calcula-
ted accordingly.
Figure 2. Kinetic investigation of LaFe0.57Co0.38Pd0.05O3 in the
first 7 minutes of the reaction between phenylboronic acid
and 4-bromoanisole.
648
ꢂ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 647–654

Palladium-Containing Perovskites in the Suzuki Reaction
FULL PAPERS
Table 2. Maitlis filtration tests.
Conversion before
Filtration (%)*
Maximum Conversion
after Filtration (%)*
4
30
73
78
83
96
99
96
* Conversion based on LC-MS.
catalyst, even when the reaction was stopped after only
4% conversion, Table 2.
Figure 3. Kinetics of oxidized, reduced and re-oxidized
LaFe0.57Co0.38Pd0.05O3 catalysts.
The results of these tests support the formation of
soluble Pd and demonstrate not only that catalytic activ-
ity was partially retained after filtration but also that the
active solution-phase species has a lifetime that is less
than the total reaction time. This test does not preclude
tred cubic arrangement of metallic palladium, or palla- the presence of nanoclusters,[39,40] which would not be re-
dium-cobalt clusters at the surface of the perovskite.[29] moved by filtration through media of this nature.
These redox conditions can be mimicked by the treat-
ment of the catalyst with hydrogen at 8008C to produce
reduced material that can be regenerated by re-oxida- Catalyst Poisoning
tion in oxygen at 8008C. Kinetic plots of the Suzuki reac-
tion catalysed by the standard oxidized material (al- CS2 as well as other ligands like PPh3 or thiophene bind
ready described in Figure 1), reduced material and re- strongly to metal centres, thereby blocking access of the
oxidized catalyst are pictured in Figure 3.
substrate to the active site. This poisoning can be per-
The reduced form of the catalyst is an active, but infe- formed quantitatively to permit determination of the
rior Suzuki catalyst displaying sigmoidal kinetics with an number of active metal atoms.[41] If the catalyst can be
induction period significantly longer than that displayed poisoned with less than 1.0 equiv. of the added ligand
by the oxidized catalyst. However, it is interesting to (per metal atom), that suggests that not all metal atoms
note that this catalyst, in which the palladium is zero-val- are participating in catalysis. This is generally consid-
ent, is a significantly worse catalyst for the Suzuki reac- ered evidence for a heterogeneous catalyst, which usual-
tion than the catalyst in which the palladium is predom- ly only has a fraction of the metal atoms available for cat-
inantly in a high-valent state (PdIII). This is almost coun- alysis. On the contrary, more than 1.0 equiv. is usually re-
ter-intuitive as it can be expected that the palladium quired to completely poison a homogeneous catalyst.[42]
must be reduced from PdIII prior to any oxidative inser- A drawback of CS2 poisoning experiments is that they
tion event; it is clear then that the Pd0 species produced must be performed at temperatures below 508C, be-
during the reaction must be distinct from that produced cause the ligands will begin to reversibly dissociate
under the redox conditions described herein. The re-oxi- from the catalyst above this temperature.[36]
dized catalyst displays kinetics identical to those of the
The poisoning experiment was conducted using phe-
original catalyst, consistent with the reversible migra- nylboronic acid and 1-fluoro-4-iodobenzene as coupling
tion of palladium in and out of the perovskite lattice.[29] partners under our usual Suzuki conditions [1 mol %
LaPd*, 3 equivs. K2CO3, 2-propanol/water (1:1, v/v)].
The activated iodide was chosen in order that the reac-
Maitlisꢀ Filtration Test
tion could progress at the lower temperature (408C)
necessary to prevent ligand dissociation.[41] After the re-
In order to probe the formation of soluble Pd species, action had proceeded 10–30% (as adjudged by LC-
Maitlisꢁ filtration experiments were performed and the MS), CS2 was added to the reaction mixture (Scheme 2).
catalytic activity of the resulting solution measured.[38]
It was found that only 0.01 equiv. of CS2 (vs. palladi-
This involved stopping the reaction after a range of con- um) was necessary to poison LaPd* (average over 39 ex-
versions (as adjudged by LC-MS) and filtering the reac- periments). The stoichiometry of this poisoning study
tion through a 0.45 mm syringe filter to remove the bulk indicates that the true total turnover number is signifi-
catalyst. It was found that the filtrates were catalytically cantly higher than the 243,000 originally indicated.[43]
active, and that the reactions progressed to significantly
When the kinetic experiment was repeated with the
higher conversions in the absence of the solid perovskite original ꢀoxidizedꢁ catalyst in the presence of triphenyl-
Adv. Synth. Catal. 2005, 347, 647–654
asc.wiley-vch.de
ꢂ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
649

Stephen P. Andrews et al.
FULL PAPERS
Scheme 3. PS-thiophenol three-phase poisoning study.
Scheme 2. CS2 poisoning experiment.
phosphine (50 molar equivalents with respect to Pd), the ki reaction but the oxidized form, which is a perovskite
additive was found to inhibit the reaction by prolonging lattice incorporating PdIII shows high activity. A mecha-
the induction period by some 40 minutes (Figure 4). nistic proposal that incorporates the information dis-
However, the maximum reaction rate attained by the cussed thus far is pictured below, Figure 5.
catalyst is still comparable to the reaction in the absence
The first step in the reaction could be expected to in-
of the added ligand. This can be attributed to the phos- volve a reduction step in which PdIII (or other high-val-
phine competing with bromoanisole for ligation to a ent Pd species) in the perovskite is reduced to a Pd0 spe-
Pd0 species that is still bound to the perovskite surface. cies still bound to a solid surface (possibly by the aque-
The binding of the phosphine is reversible and so it is ous alcoholic solvent) which can then undergo oxidative
when oxidative insertion of the aryl halide occurs that addition with an aryl halide. It is thought that this oxida-
the generation of a soluble Pd species is initiated.
tive addition is the step at which a soluble catalyst is pro-
A three-phase test designed to catch and poison a duced on the basis of a previously described three-phase
soluble Pd source was performed with polymer-support- test. Triphenylphosphine increases the period of time it
ed thiophenol as the immobilized poison. It was found takes for the active catalyst to be produced, and hence it
that if the thiol was added to a Suzuki reaction that is believed to be competing with the aryl halide at this
had been started in the usual manner, the reaction prog- stage for a solid-bound Pd0 species. The actual mecha-
ress would stop at the point of addition, Scheme 3.
nism by which the palladium enters the solution after
Also, if the reagents were added to a sample of perov- the oxidative insertion is unknown; the mechanism of
skite, which was pre-mixed with PS-thiophenol in sol- oxidative addition itself is still the subject of much de-
vent at 808C, no reaction occurred. Due to the site-iso- bate despite sophisticated studies.[44] Once in solution,
lation principle, this test is evidence of a catalytically ac- it is envisaged that the catalyst can enter a standard
tive soluble Pd source that is trapped by the strongly cross-coupling catalytic cycle, producing a soluble Pd0
binding sulfur ligand. This is complementary to a previ- species. It is possible that this Pd0 intermediate may
ous three-phase test conducted on this system.[31]
form halide-stabilized nanoparticles, which could ag-
Mechanism
The reduced form of LaPd* has been shown by kinetic
experiments to be a relatively poor catalyst for the Suzu-
Figure 4. Comparison of the kinetics of LaFe0.57Co0.38Pd0.05O3
in the Suzuki reaction with and without added triphenylphos-
phine.
Figure 5. Putative mechanistic scheme for the generation of a
soluble active catalyst (PS¼polymer supported, Ar¼aryl).
650
ꢂ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 647–654

Palladium-Containing Perovskites in the Suzuki Reaction
FULL PAPERS
glomerate to generate palladium black in the absence of The mechanism by which this occurs involves the gener-
another oxidative addition, although these species have ation of a soluble, catalytically active palladium species
not been observed by TEM. This is consistent with other that is desorbed from the bulk inorganic phase and prob-
ligand-free systems in which high substrate to catalyst ably redeposited on a modified solid phase after the re-
ratios produce high TONs.[45–47] The Pd0 species can car- action is complete. The nature of the soluble palladium
ry on the catalytic cycle or be recaptured by the bulk in- species is not clear; attempts to observe palladium nano-
organic phase (as suggested by ICP measurements that clusters in these systems via TEM investigations have
detected a residual Pd level of approx. 2 ppm in reaction proven fruitless. Only a small number of perovskites
mixture filtrates). The presence of such a species in sol- were screened for activity in this study, but in principle
ution is indicated by the ease with which a poly- a huge range of perovskite materials containing alterna-
mer-supported thiol poison can hinder the reaction, tive metal combinations could be prepared for potential
and the prevailing catalytic activity of filtrates after solid applications in organic synthesis programmes. The apti-
catalyst removal.
tude of different metal-containing perovskite systems is
currently being explored.
TEM Studies
In an attempt to ascertain whether palladium was being
redeposited on the surface of the perovskite after reac-
tion a series of TEM studies was initiated. The surface
of the catalyst was examined before, and after reaction,
Figure 6.
Whereas the surface of the unused catalyst is relative-
ly uniform and appears amorphous, the post-reaction
catalyst has undergone a profound morphological
change and appears to consist predominantly of a new
Experimental Section
General
All reagents and solvents were bought from commercial sour-
ces and used as supplied. Reactions were performed in Rad-
leyꢁs CarouselTM tubes in air unless otherwise stated. Kinetic
experiments were conducted with React ArrayTM SK233
equipment and analysed by in-line high-pressure liquid chro-
crystalline phase. The identity of this crystalline material matography HPLC (vide infra). Thin layer chromatography
(TLC) was performed with Merck 60 F254 silica gel plates
and viewed under UV radiation (254 nm). Liquid chromatog-
raphy/mass spectrometry (LC-MS) experiments were conduct-
ed on HP1100/HP MSD with diode array detection. Column:
Mercury MS Luna 3 mm C8(2), 20ꢁ2.0 mm. Eluent: A: water,
is not known at present, though preliminary X-ray dif-
fraction data indicate that it is neither the perovskite cat-
alyst or simple lanthanum oxide or hydroxide.
0.1% TFA, B: acetonitrile, 0.1% TFA. Flow rate: 0.6 mL minꢀ1
.
Conclusion
Method: gradient 5–95% B in A over 2.5 minutes, run time 8
minutes. Ionization technique: atmospheric pressure chemical
ionization (APCI).
It has been demonstrated that palladium-containing
perovskites are a general class of ꢀheterogeneousꢁ re-
High-resolution mass spectrometry was conducted with
agents possessing utility in cross-coupling reactions. Kratos Concept spectrometer at the University of Cambridge.
Infrared spectra were recorded by Perkin-Elmer Spectrum
One spectrometer equipped with an attenuated total reflec-
tance sampler. NMR spectra were recorded with Bruker
DPX-400 spectrometer operating at 400 MHz and 100 MHz
for measurement of 1H and 13C shifts, respectively, with
CDCl3 solvent. COSY spectra were obtained and used to as-
sign 1H signals. HMQC and DEPT experiments were used to
aid the assignment of 13C signals. All quoted 1H and 13C shifts
are measured in ppm, with residual CHCl3 calibrated to 7.27
and 77.0 ppm respectively.
Microwave heating was performed in 5 mL Smith vials by ir-
radiation with a Personal Chemistry Smith SynthesiserTM
.
TEM images were obtained at the Multi-Imaging Centre, Uni-
versity of Cambridge. Samples were dehydrated in three
changes of dry ethanol, followed by three changes of dry aceto-
nitrile and incubated overnight in 50/50 acetonitrile and
Spurrꢁs epoxy resin. The samples were aged for 8 hours in 25/
75 acetonitrile and Spurrꢁs resin followed by 5 days in 100 resin
under vacuum and finally cured at 608C for 24 hours. Sections
(60 nm) were cut on a Leica Ultracut UCT using a diamond
knife, mounted on 400 mesh copper grids and viewed on a
FEI-Philips CM100 operated at 80 kVusing a 100 mm condens-
Figure 6. TEM images of the catalyst before (left) and after
(right) reaction.
Adv. Synth. Catal. 2005, 347, 647–654
asc.wiley-vch.de
ꢂ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
651

Stephen P. Andrews et al.
FULL PAPERS
er aperture and a 30 mm objective aperture. Images were re-
corded on Kodak 4489 cut film.
Data Analysis
Calibration: Standard solutions of 4-bromoanisole (2.0 mL
mLꢀ1), naphthalene-2-carboxamide (2.0 mg mLꢀ1) and 4-me-
thoxybiphenyl (2.0 mg mLꢀ1) were prepared with acetonitrile
in volumetric flasks and serially diluted with acetonitrile. A set
of six solutions of varying concentration was thus obtained for
each of the three compounds and these were injected into the
HP1100 and the HPLC trace recorded at 220 nm. HPLC trace
areas were plotted against sample concentration to produce a
calibration curve.
Reactions were performed using a React ArrayTM SK233 au-
tomated workstation subjected to in-line HPLC analysis on
Hewlett Packard HP1100. Column: SupercosilTM ABZþ
3.3 cmꢁ4.6 mm, 3 mm. Eluent: A: water, 0.1% TFA, B: 95%
acetonitrile, 5% water, 0.05% TFA. Flowrate: 1 mL minꢀ1. De-
tection: UV (220 nm). Method: gradient 10–95% B in A over 7
minutes, run time 10 minutes. All solvents were commercially
available and used as supplied.
All reactions were performed in air at 808C in 25-mL vessels
with fitted condensers and stirrer bars. 10 mL aliquots were
sampled at specified intervals and diluted serially by a factor
of 100 before being subjected to in-line HPLC analysis. The ali-
quots were dispensed into 1.5-mL vials, quenched with acetoni-
trile (1 mL) and mixed [75% by volume, (aspirated and dis-
pensed at 10 mL minꢀ1)]. 100 mL of these solutions were trans-
ferred to clean 1.5-mL vials, diluted with acetonitrile (900 mL)
and mixed [75% by volume, (aspirated and dispensed at 10 mL
minꢀ1)] to give analytical samples ready for analysis by HPLC
(20 mL injected).
Monitoring Suzuki Reaction Progress (Raw and
processed data available in Supporting Information)
During the Suzuki reaction, the HPLC trace areas for naphtha-
lene-2-carboxamide, 4-bromoanisole and 4-methoxybiphenyl
were recorded and normalized with respect to the internal
standard (naphthalene-2-carboxamide). Areas of 4-bromoani-
sole and 4-methoxybiphenyl were then extrapolated back to
their concentrations using the calibration curves. Percentages
of 4-bromoanisole remaining in solution and 4-methoxybi-
phenyl formed were calculated and reaction progress assessed
by plotting conversion [¼(% 4-methoxybiphenyl)/(% 4-
bromoanisoleþ% 4-methoxybiphenyl)] against time.
Typical Procedure for React ArrayTM Suzuki Study
4-Bromoanisole (500 mL, 4 mmol), phenylboronic acid
(744 mg, 6 mmol), K2CO3 (1.66 g, 12 mmol) and naphtha-
lene-2-carboxamide (64.8 mg, 379 mmol) were added to a reac-
tion vessel and diluted with 2-propanol/water (12 mL, 1:1, v/v).
The SK233 was initiated and once the mixture had reached
808C and homogeneous phase mixing, the perovskite catalyst
(10 mg, 40 mmol) was added and the timer started. The reac-
tions were run for 5 hours and 30 minutes. Aliquots (10 mL)
were withdrawn automatically, diluted and injected into
HP1100 for HPLC analysis at t¼1, 6, 11, 16, 21, 26, 31, 36,
41, 46, 51, 56, 61, 75, 90, 105, 120, 135, 150, 165, 180, 210, 240,
270, 300, 330 minutes.
Typical Procedure for Determination of Catalyst
Turnover Number
LaFe0.57Co0.38Pd0.05O3 (0.50 mg, 2.0 mmol) was added to a stir-
ring solution of 4-bromoanisole (10.0 g, 53.5 mmol), phenyl-
boronic acid (9.8 g, 80 mmol) and K2CO3 (22.2 g, 120 mmol)
in a mixture of 2-propanol/water (100 mL, 1:1, v/v) at 808C.
The resulting mixture was stirred until LC-MS (220 nm) re-
vealed no change in the ratio of 4-bromoanisole:4-methoxybi-
phenyl between two aliquots that had been removed at 24-hour
intervals (this takes approximately two weeks).
Typical Procedure for CS2 Poisoning Experiment
A 20-mL tube was charged with phenylboronic acid 1 (180 mg,
1.50 mmol), 1-fluoro-4-iodobenzene 25 (0.11 mL, 1.00 mmol),
LaFe0.57Co0.38Pd0.05O3 (3.00 mg, 1 mol %, 0.05 mol % Pd),
K2CO3 (0.40 mg, 2.90 mmol) and water/2-propanol (1:1, v/v,
5 mL), and the mixture was stirred at 08C until LC indicated
the reaction had proceeded 10–30%. At 08C CS2 [0.05–
0.4 mL of a freshly prepared 0.02 nM solution in water/2-prop-
anol (1:1 v/v)] was then added. This mixture was stirred at
408C and the conversion followed by LC over 2 days.
Small-Scale Work-Up (1 mL of the crude reaction
mixture extracted at 808C with vigorous stirring to
maintain phase-homogeneity)
The sample was partitioned between Et2O (10 mL) and water
(5 mL). The separated aqueous phase was back-extracted with
Et2O (10 mL) and the combined organic phases were washed
with brine (5 mL), dried (MgSO4), filtered and evaporated un-
1
der reduced pressure. The residue was analysed by H NMR
and the MeO peak integrals measured and compared to pro-
vide the definitive ratio of 4-bromoanisole:4-methoxybiphenyl
and from this, the catalyst turnover number.
Typical Procedure for Maitlisꢀ Test
A 20-mL tube was charged with phenylboronic acid 1 (180 mg,
1.50 mmol), 1-fluoro-4-iodobenzene 25 (0.11 mL, 1.00 mmol),
LaFe0.57Co0.38Pd0.05O3 (3.00 mg, 1.00 mol %, 0.05 mol % Pd),
K2CO3 (0.40 g, 2.90 mmol) and water/2-propanol (1:1, v/v,
5 mL), and the mixture was stirred at 08C, room temperature
or 408C until LC indicated the reaction had proceeded 5–
70%. The mixture was then filtered through a 0.45 mm syringe
filter and stirred at 408C. The conversion was followed by LC
over 1 day.
Acknowledgements
The authors would like to thank the EPSRC and GlaxoSmithK-
line for a CASE award (to SPA), Novartis [for a CASE award
(to AFS) and the Novartis Research Fellowship (to SVL)], Dai-
652
ꢂ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 647–654

Palladium-Containing Perovskites in the Suzuki Reaction
FULL PAPERS
hatsu Motor Company for their kind gift of perovskite catalysts,
Hokko Chemical Industry Co. Ltd., for their assistance in sam-
ple preparation and the Royal Society for a URF (to MDS).
[24] D. A. Conlon, B. Pipik, S. Ferdinand, C. R. LeBlond,
J. R. Sowa, B. Izzo, P. Collins, G. J. Ho, J. M. Williams,
Y. J. Shi, Y. K. Sun, Adv. Synth. Catal. 2003, 345, 931.
[25] I. W. Davies, L. Matty, D. L. Hughes, P. J. Reider, J. Am.
Chem. Soc. 2001, 123, 10139.
[26] A. Biffis, M. Zecca, M. Basato, Eur. J. Inorg. Chem. 2001,
1131.
References and Notes
[27] Y. Na, S. Park, S. B. Han, H. Han, S. Ko, S. Chang, J. Am.
Chem. Soc. 2004, 126, 250.
[28] C. A. Jaska, I. Manners, J. Am. Chem. Soc. 2004, 126,
9776.
[29] Y. Nishihata, J. Mizuki, T. Akao, H. Tanaka, M. Uenishi,
M. Kimura, T. Okamoto, N. Hamada, Nature 2002, 418,
164.
[30] N. Miyaura, T. Yanagi, A. Suzuki, Synth. Commun. 1981,
11, 513.
[31] M. D. Smith, A. F. Stepan, C. Ramarao, P. E. Brennan,
S. V. Ley, Chem. Commun. 2003, 2652.
[32] This is an improvement on our previously disclosed value
of 27,000.
[33] F. G. J. Rebek, S. Zimmerman, J. Am. Chem. Soc. 1975,
97, 454.
[1] The classical, solubility-based definition equates “hetero-
geneous” to “insoluble” and “homogeneous” to “solu-
ble”. When referring to this definition, quotation marks
will be used (ꢀhomogeneousꢁ, ꢀheterogeneousꢁ). A newer
definition deals with the active site of the catalytic sys-
tem; see J. Schwartz, Acc. Chem. Res. 1985, 18, 302. If
the active site is uniform, the system is called homogene-
ous; if the active site consists of irregular sites, then the
catalyst is called heterogeneous. When referring to this
definition no quotation marks will be used.
[2] R. B. Bedford, S. L. Hazelwood, M. E. Limmert, Chem.
Commun. 2002, 2610.
[3] R. B. Bedford, S. L. Welch, Chem. Commun. 2001, 129.
[4] J. P. Wolfe, R. A. Singer, B. H. Yang, S. L. Buchwald, J.
Am. Chem. Soc. 1999, 121, 9550.
[34] F. G. J. Rebek, J. Am. Chem. Soc. 1974, 96, 7112.
[35] J. A. Widegren, M. A. Bennett, R. G. Finke, J. Am.
Chem. Soc. 2003, 125, 10301.
[36] J. A. Widegren, R. G. Finke, J. Mol. Catal. A: Chem.
2003, 198, 317.
[37] B. H. Lipshutz, S. Tasler, W. Chrisman, B. Spliethoff, B.
Tesche, J. Org. Chem. 2003, 68, 1177.
[38] P. M. Maitlis, J. Mol. Cat. 1980, 7, 543.
[39] M. T. Reetz, J. G. de Vries, Chem. Commun. 2004, 14,
1559.
[40] M. T. Reetz, E. Westermann, Angew. Chem. Int. Edit.
2000, 39, 165.
[41] B. J. Hornstein, J. D. Aiken, R. G. Finke, Inorg. Chem.
2002, 41, 1625.
[42] In cases where a heterogeneous precatalyst generates a
catalytically active soluble catalyst, the stoichiometry of
such poisoning studies are useful for the calculation of
turnover numbers but are not particularly informative
for the investigation of homogeneous versus heterogene-
ous mechanisms.
[5] S. D. Walker, T. E. Barder, J. R. Martinelli, S. L. Buch-
wald, Angew. Chem. Int. Edit. 2004, 43, 1871.
[6] A. F. Littke, G. C. Fu, Angew. Chem. Int. Edit. 2002, 41,
4176.
[7] A. F. Littke, C. Y. Dai, G. C. Fu, J. Am. Chem. Soc. 2000,
122, 4020.
[8] C. E. Garrett, K. Prasad, Adv. Synth. Catal. 2004, 346,
889.
[9] C. Ramarao, S. V. Ley, S. C. Smith, I. M. Shirley, N.
DeAlmeida, Chem. Commun. 2002, 1132.
[10] S. V. Ley, C. Mitchell, D. Pears, C. Ramarao, J. Q. Yu,
W. Z. Zhou, Org. Lett. 2003, 5, 4665.
[11] S. Kobayashi, R. Akiyama, Chem. Commun. 2003, 449.
[12] R. Akiyama, S. Kobayashi, J. Am. Chem. Soc. 2003, 125,
3412.
[13] K. Okamoto, R. Akiyama, S. Kobayashi, Org. Lett. 2004,
6, 1987.
[14] B. M. Trost, E. Keinan, J. Am. Chem. Soc. 1978, 100,
7779.
[15] A. Datta, K. Ebert, H. Plenio, Organometallics 2003, 22,
4685.
[43] The true turnover number takes into account the number
of catalytically active metal sites in the molecule, and for
the catalyst described here is 100 times greater than first
indicated (giving a true turnover number of 24.3 million).
For a soluble catalyst generated from an insoluble preca-
talyst, it is to be expected that the soluble palladium con-
tent of the reaction mixture may vary as the reaction pro-
gresses. In the studies presented herein, there was rela-
tively little variation in the amount of CS2 required to
poison the reaction over a range of experiments, and
no correlation between product conversion (when the
poison was added) and the volume of CS2 required to
stop any catalysis.
[16] Y. M. A. Yamada, K. Takeda, H. Takahashi, S. Ikegami,
J. Org. Chem. 2003, 68, 7733.
[17] K. Kçhler, R. G. Heidenreich, J. G. E. Krauter, M.
Pietsch, Chem. Eur. J. 2002, 8, 622.
[18] R. G. Heidenreich, E. G. E. Krauter, J. Pietsch, K. Kçh-
ler, J. Mol. Catal. A: Chem. 2002, 182, 499.
[19] R. G. Heidenreich, K. Kçhler, J. G. E. Krauter, J. Pietsch,
Synlett 2002, 1118.
[20] Z. Novak, A. Szabo, J. Repasi, A. Kotschy, J. Org. Chem.
2003, 68, 3327.
[21] N. E. Leadbeater, M. Marco, Angew. Chem. Int. Edit.
2003, 42, 1407.
[22] N. E. Leadbeater, M. Marco, J. Org. Chem. 2003, 68,
5660.
[23] S. S. Prockl, W. Kleist, M. A. Gruber, K. Kçhler, Angew.
Chem. Int. Edit. 2004, 43, 1881.
[44] C. Amatore, A. Jutand, Acc. Chem. Res. 2000, 33, 314.
[45] D. Zim, A. L. Monteiro, J. Dupont, Tetrahedron Lett.
2000, 41, 8199.
Adv. Synth. Catal. 2005, 347, 647–654
asc.wiley-vch.de
ꢂ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
653

Stephen P. Andrews et al.
FULL PAPERS
[46] A. H. M. de Vries, J. Mulders, J. H. M. Mommers,
H. J. W. Henderickx, J. G. de Vries, Org. Lett. 2003, 5,
3285.
[47] Y. J. Deng, L. Z. Gong, A. Q. Mi, H. Liu, Y. Z. Jiang,
Synthesis 2003, 337.
654
ꢂ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 647–654