Suzuki cross-coupling reaction catalyzed by sulfur-containing palladacycles: Formation of palladium active species

Article abstract of DOI:10.1016/j.molcata.2008.02.015

Air-stable cyclopalladated compounds derived from the ortho-metalation of benzylic tert-butyl thioethers efficiently promote C-C coupling reactions and were studied mechanistically in the Suzuki reaction. The catalytic reaction was monitored by ¹⁹F NMR and GC-MS and the decomposition of the S-palladacycle was studied under stoichiometric conditions, showing that these palladacycles serve as a reservoir of zerovalent palladium species. The formation of these species is initiated by attack of the arylboronic acid on the palladacycle resulting in the formation of an arylated palladacycle. Thereafter, this species undergoes a reductive elimination to form the active Pd(0) species. As a minor pathway, the Pd(0) species could also be generated in a pathway involving a reduction process that leads to the decomposition of the palladacycle to Pd(0) affording the ortho hydrogenated thioether. Poisoning studies and TEM analysis indicated the presence of palladium nanoparticles. Competitive experiments showed that electron-withdrawing substituents on the aryl halide and electron-donating substituents on the arylboronic acid facilitated the reaction. They also provide evidence that the soluble species formed from the oxidative addition, that are active for Heck and Suzuki reactions, are the same species.

Full text of DOI:10.1016/j.molcata.2008.02.015

Journal of Molecular Catalysis A: Chemical 287 (2008) 16–23
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Suzuki cross-coupling reaction catalyzed by sulfur-containing palladacycles:
Formation of palladium active species
Danilo Zim1, Sabrina M. Nobre, Adriano L. Monteiro∗
Laboratory of Molecular Catalysis, Instituto de Qu´ımica, UFRGS, Av. Bento Gonc¸ alves, 9500 Porto Alegre 91501-970, CP 15003, RS, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Air-stable cyclopalladated compounds derived from the ortho-metalation of benzylic tert-butyl thioethers
efficiently promote C–C coupling reactions and were studied mechanistically in the Suzuki reaction. The
catalytic reaction was monitored by 19 F NMR and GC–MS and the decomposition of the S-palladacycle was
studied under stoichiometric conditions, showing that these palladacycles serve as a reservoir of zerova-
lent palladium species. The formation of these species is initiated by attack of the arylboronic acid on the
palladacycle resulting in the formation of an arylated palladacycle. Thereafter, this species undergoes a
reductive elimination to form the active Pd(0) species. As a minor pathway, the Pd(0) species could also
be generated in a pathway involving a reduction process that leads to the decomposition of the pallada-
cycle to Pd(0) affording the ortho hydrogenated thioether. Poisoning studies and TEM analysis indicated
the presence of palladium nanoparticles. Competitive experiments showed that electron-withdrawing
substituents on the aryl halide and electron-donating substituents on the arylboronic acid facilitated the
reaction. They also provide evidence that the soluble species formed from the oxidative addition, that are
active for Heck and Suzuki reactions, are the same species.
Received 10 September 2007
Received in revised form 28 January 2008
Accepted 13 February 2008
Available online 4 March 2008
Keywords:
Palladacycles
Nanoparticles
Suzuki reaction
Mechanism
© 2008 Elsevier B.V. All rights reserved.
1. Introduction
atom (1, 3, 4, and 5). Lower activities were observed with a MeS (2)
moiety or with two t-BuS groups (6) ligated to the palladium atom.
The palladium-catalyzed Suzuki cross-coupling reaction is one
of the most efficient methods for the construction of Caryl Caryl
bonds and has found widespread use in organic synthesis [1–4].
Palladacycles, where the ligand is in a position to coordinate to
the metal center through both a metalated carbon and a donor
atom (P, N, S), are one of the most investigated classes of cata-
lyst precursors in particular due to their facile synthesis, thermal
stability and possibility of modulating their steric and electronic
properties [5–8]. Our group have obtained sulfur-containing pal-
ladacycles from the ortho-metalation of benzylic alkyl thioethers
[9] (Scheme 1) and have shown that they are excellent catalyst pre-
cursors for C–C coupling reactions such as Heck [10], Suzuki [11] and
homocoupling [12] processes. The palladacycles were evaluated
for the Suzuki coupling reaction of the relatively deactivated 4-
bromotoluene with phenylboronic acid. All palladacycles depicted
in Scheme 1 were active but the catalytic activity depended on
the palladacycle structure. The most efficient catalyst precursors
present one tert-butyl thioether moiety bonded to the palladium
Moreover, no discernible influence in activity was observed with a
change from chloro (1) to acetate (4) bridges, from methyl to hydro-
gen (3) in the R1 position, and from a phenyl (1) to a naphthyl group
(5) in the palladacycle. Electron-rich and -poor aryl bromides as
well as electron-poor aryl chlorides were efficiently coupled in the
presence of 1 to provide the corresponding biaryl products in excel-
lent isolated yields (>90%). Moreover, palladacycle 1 promotes the
Suzuki coupling even at room temperature although longer reac-
tions times are necessary with a wide variety of functional groups
tolerated (nitro, acetyl, cyano, etc.). In this work we wish to report
our results regarding the nature of the active species involved in
the Suzuki reaction promoted by these palladacycles.
2. Experimental
2.1. General methods
All catalytic reactions were carried out under an argon atmo-
sphere in oven-dried resealable Schlenk tubes. Chemical were
purchased from commercial sources, and were used without
further purification. Palladacycle 1 was obtained using proce-
dure described earlier [9]. Elemental analyses were performed
by the Analytical Central Service of IQ-UFRGS. NMR spectra
∗
Corresponding author. Tel.: +55 51 33086303; fax: +55 51 33087304.
E-mail addresses: danilo.zim@br.rhodia.com (D. Zim),
adriano.monteiro@iufrgs.br (A.L. Monteiro).
1
Present address: Rhodia Poliamida e Especialidades—CPP.
1381-1169/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2008.02.015

D. Zim et al. / Journal of Molecular Catalysis A: Chemical 287 (2008) 16–23
17
Scheme 1. Palladacycles obtained from the ortho-metalation of benzylic alkyl thioethers and the Suzuki cross-coupling reaction.
4.05 (q, J = 7.2 Hz, 1H), 1.55 (d, J = 7.2 Hz, 3H), 1.24 (s, 9H). 13C NMR
(75.4 MHz, CDCl3) ı 161.8 (d, J = 242.9 Hz), 142.6, 128.8 (d, J = 7.5 Hz),
115.5 (d, J = 22.6 Hz), 44.1, 42.0, 31.7, 25.7. GC–MS (EI, 70 eV) m/z (%)
212 (12, M+), 123 (100, M+-StBu), 103 (16), 101 (4), 103 (16), 77 (6),
59 (11), 57 (47).
were recorded on a Varian XL300 spectrometer. Infrared spec-
tra were performed on a Bomem B-102 spectrometer. Mass
spectra were obtained on a gas chromatograph/mass spectrome-
ter (GC/MS) Shimadzu QP-5050 (EI, 70 eV). Gas chromatography
(GC) analyses were performed on a Hewlett-Packard-5890 Gas
Chromatograph with
a flame ionization detector (FID) and
30 m capillary column with a dimethylpolysiloxane stationary
phase.
2.2.3. Synthesis of palladacycle 9
To
a solution of palladium acetate (225 mg, 1 mmol) in
acetic acid (30 mL) was added 1-(1-tert-butylsulfanylethyl)-4-
fluorobenzene (234 mg, 1.1 mmol) at room temperature. The
solution was stirred at 90 ◦C for 15 min. The brown solution thus
obtained was evaporated to dryness and washed with hexanes
(2 × 10 mL). The solid was dissolved in dichloromethane (50 mL),
and the solution was filtered through a plug (5 cm) of alumina
(grade I), and concentrated to ca. 1 mL. Addition of hexanes
(50 mL) afforded a yellow solid, which was collected by filtration,
washed with hexanes (2 × 10 mL), and dried under reduced pres-
sure (151 mg, 40% yield based on Pd): Anal. Calcd for C28H38F2O4S2
Pd2: C, 44.63%; H, 5.08%. Found: C, 44.44%; H, 4.98%. 1H NMR
(300 MHz, DMSO) ı 7.35 (br, d, J = 8.3 Hz, 1H), 6.94 (br, s, 1H),
6.71–6.69 (br, m, 1H), 4.31 (br, m, 1H), 3.34 (s, 3H), 1.58 (d, J = 7.2 Hz,
3H), 1.32 (s, 9H). IR (nujol) ꢀ (cm−1) 1574, 1458, 1443, 1397, 1252,
1229, 1186, 1159, 1112, 865, 811, 753.
2.2. Synthesis of palladacycle 9
2.2.1. 1-(4ꢀ-Fluorophenyl)ethanol
To an ethanol (30 mL) solution of 4ꢀ-fluoroacetophenone (6.1 mL,
50 mmol) was added NaBH4 (3.783 g, 100 mmol) and the mixture
was stirred at room temperature for 1 h then warmed to 80 ◦C for
15 min. After the mixture was cooled to room temperature, the sol-
vent was removed under reduced pressure and the mixture was
diluted with ether (30 mL) and washed with water (3 × 10 mL), and
then dried over MgSO4. After filtration, the solvent was evaporated
to give 1-(4ꢀ-fluorophenyl)ethanol as a colorless liquid (6.567 g,
95%). 1H NMR (300 MHz, CDCl3) ı 7.32–7.27 (m, 2H), 7.04–6.98 (m,
2H), 4.81 (q, J = 6.3 Hz, 1H), 2.90 (br, 1H), 1.43 (d, J = 6.3 Hz, 3H). 13
C
NMR (75.4 MHz, CDCl3) ı 162.3 (d, J = 243.4 Hz), 141.8 (d, J = 3 Hz),
127.4 (d, J = 7 Hz), 115.4 (d, J = 14 Hz), 69.8, 25.5. IR ꢀ (cm−1) 3362,
2975, 1605, 1509, 1222, 1157, 1085, 1014, 900, 836. GC–MS (EI,
70 eV) m/z (%) 140 (19, M), 125 (100), 97 (94), 96 (30), 95 (26), 77
(38), 51 (26), 43 (93).
2.3. Decomposition of palladacycle 1 under stoichiometric
conditions
A mixture of palladacycle 1 (50 mg, 0.15 mmol), phenylboronic
acid (27.5 mg, 0.225 mmol), PPh3 (39.3 mg, 0.15 mmol), unde-
cane (8 mg, internal standard) and K3PO4 (63.6 mg, 0.3 mmol)
in DMF (10 mL) was heated at 50 ◦C for 20 h, then heated
at 80 ◦C for 3 h.GC–MS and GC analysis showed 1-(1-tert-
butylsulfanylethyl)-2-phenylbenzene (7) as the main product (92%)
and (1-tert-butylsulfanylethyl)benzene (8) as a by-product. The
mixture was then allowed to cool to room temperature, taken
up in ether (20 mL) and washed with aqueous NaOH (1 M,
3 × 10 mL) and brine (3 × 10 mL) and then dried over MgSO4.
1-(1-tert-Butylsulfanylethyl)-2-phenylbenzene (7) was isolated
2.2.2. 1-(1-tert-Butylsulfanylethyl)-4-fluorobenzene
A Schlenk flask was charged with 1-(4ꢀ-fluorophenyl)ethanol
(2.800 g, 20 mmol), 2-methyl-2-propanethiol (2.25 mL, 20 mmol),
ZnI2 (2.394 g, 7.5 mmol), dichloromethane (80 mL), and the mix-
ture was refluxed for 1 h. After cooling to room temperature, the
mixture was washed with water (2 × 40 mL), and then dried over
MgSO4. After filtration, the solvent was evaporated and a Kugerloh
distillation under reduced pressure (90 ◦C at 2 mmHg), afforded 1-
(1-tert-butylsulfanylethyl)-4-fluorobenzene as an oil (1.49 g, 35%).
1H NMR (300 MHz, CDCl3) ı 7.38–7.28 (m, 2H), 7.04–6.97 (m, 2H),

18
D. Zim et al. / Journal of Molecular Catalysis A: Chemical 287 (2008) 16–23
at the desired temperature (25, 50 and 80 ◦C) and the conversion
was followed by GC over 3 days.
by chromatography (32 mg, 80%). 1H NMR (300 MHz, CDCl3) ı
7.66–7.16 (m, 9H), 4.17 (q, J = 6.65 Hz, 1H), 1.59 (d, J = 6.87 Hz, 3H),
1.01 (s, 9H). 13C NMR (75.4 MHz, CDCl3) ı 143.8, 141.3, 139.7, 129.7,
129.4, 128.1, 127.9, 127.8, 127.1, 126.2, 37.8, 31.3, 29.7, 25.9. IR (nujol)
ꢀ (cm−1) 3022, 3059, 2963, 2923, 2862, 1597, 1449, 1478, 1364,
1163, 1044, 1009, 762, 746, 701. GC–MS (IE, 70 eV) m/z (%): 270 (8,
M+), 182(12), 181(100), 179(11), 178(8), 166(18), 164(33), 152(6),
57(16).
The same procedure was used for the decomposition of pal-
ladacycle 1 under stoichiometric conditions in the absence of PPh3.
GC–MS and GC analysis after 20 h at 80 ◦C showed the ortho-phenyl
thioether 7 as the main product (70%) and the thioether (1-tert-
butylsulfanylethyl)benzene 8 as a by-product (17%) [9]. GC–MS (EI,
70 eV) m/z (%): 194 (7, M+), 106 (7), 105 (100), 79 (6), 77 (12), 59
(6), 57 (27), 51 (7).
2.7. Transmission electron microscopy analysis
The morphology and the electron diffraction (ED) of the
obtained particles were carried out on a JEOL JEM-2010 equipped
with an energy-dispersive X-ray spectroscopy (EDS) system oper-
ating at accelerating voltage of 200 kV. The sample for TEM was
performed by withdrawing samples of the Suzuki reaction of 4-
bromoacetophenone with phenylboronic acid (0.5 mol% of 1, 2 h,
80 ◦C, 66% conversion) by simply depositing the solution on carbon-
coated copper grids. The histograms of the nanoparticles size
distribution, assuming spherical shape, were obtained from the
measurement of about 500 particles and were reproduced in dif-
ferent regions of the Cu grid, found in an arbitrarily chosen area of
enlarged micrographs.
2.4. Suzuki coupling reaction of 4-trifluoromethylbromobenzene
with phenylboronic acid promoted by palladacycle 9
2.8. Typical experiment using Hammett parameters
Under argon, a resealable NMR tube was charged with pallada-
cyle 9 (1.4 mg, 0.0037), K3PO4 (118 mg, 0.55 mmol), phenylboronic
acid (54.1 mg, 0.44 mmol), 4-trifluoromethylbromobenzene
2.8.1. Effect of the substituent on the aryl halide
An oven-dried resealable Schlenk flask was evacuated and back-
filled with argon and charged with K3PO4 (276 mg, 1.3 mmol),
arylboronic acid (2 mmol), aryl halides (0.02 mmol of each one),
palladacycle 1 (1 ␮mol), and DMF (1.5 mL). The reaction mixture
was stirred at 80 ◦C and the reaction profile was monitored by CG.
(83.5 mg, 0.37 mmol) and DMF (5 mL).
A
capillary tube
(D2O/CF3COOH) was placed into the tube that was introduced into
the NMR spectrometer and the reaction was followed by 19 F NMR
at 50 ◦C for 15 h. After 15 h the mixture was analyzed by GC and
GC–MS showing a 25% yield of the 4-trifluormethylbiphenyl cou-
pling product by comparison with the authentic sample. Otherwise,
GC–MS and GC analysis showed 1-(1-tert-butylsulfanylethyl)-4-
fluoro-2-phenylbenzene [10: 47% yield based on 9; GC–MS (EI,
70 eV) m/z (%) 288(7, M+), 200(16), 199(100), 184(18), 183(42),
179(18), 57(29), 41(44), 39(17)] and 1-(1-tert-butylsulfanylethyl)-
4-fluorobenzene (13% based on 9). The reaction mixture was then
transferred to a resealable Schlenk flask and stirred for 3 h at
130 ◦C, the conversion being improved to 70%.
2.8.2. Effect of the substituent on the arylboronic acid
An oven-dried resealable Schlenk flask was evacuated and back-
filled with argon and charged with K3PO4 (276 mg, 1.3 mmol), aryl
halide (20 mmol), aryl boronic acids (0.25 mmol of each one), pal-
ladacycle 1 (1 ␮mol), and DMF (1.5 mL). The reaction mixture was
stirred at 80 ◦C and the reaction profile was monitored by CG.
3. Results and discussion
Despite the importance of the Pd-catalyzed C–C coupling pro-
cess, the nature of the true active species in most cases is not
well-established. A comprehensive review of Pd-catalyzed Heck
and Suzuki reactions with an eye toward identifying the nature
of the actual active form of palladium was recently published
and, it has been demonstrated that, regardless of the precata-
lyst used, a Pd(0)–Pd(II) cycle is involved [13]. In several reports
on the use of known and new palladacycles as catalyst pre-
cursors for Heck coupling reactions, it has been stated that
palladacycles serve as a reservoir of catalytically active Pd(0)
which might be palladium nanoparticles or low-coordinated pal-
ladium zero species[6,7,13–15]. Bedford et al. have shown that
under stoichiometric conditions phosphinite based palladacycles
[16] and phosphine adducts of N- and S-based palladacycles
[17,18] in the presence of arylboronic acid undergo a transmet-
alation/reductive elimination process to generate Pd(0) species
and the orthoarylated ligand. This process was also observed for
the reaction of a benzodiazepine-containing palladacycle with
excess 3-tolylboronic acid [19]. We have since investigated whether
a similar process occurs with sulphur-containing-palladacycles
not only under stoichiometric conditions but also under catalytic
conditions. Therefore, we first examined the reaction of 1 with
phenylboronic acid in the presence of K3PO4 and DMF as solvent
(Scheme 2 ). Even at room temperature, the formation of palla-
dium black was observed within a few minutes. GC–MS and GC
analysis after 20 h at 80 ◦C showed the ortho-phenyl thioether 1-(1-
tert-butylsulfanylethyl)-2-phenylbenzene (7) as the main product
(70%) and the thioether (1-tert-butylsulfanylethyl)benzene (8) as
a by-product (17%). In order to see whether a phosphine ligand
2.5. Hg(0) poisoning studies
An oven-dried resealable Schlenk flask was charged with
4-bromonitrobenzene (1 mmol), phenylboronic acid (183 mg,
1.5 mmol), K3PO4 (424.5 mg, 2 mmol), 0.5 mol% palladacycle 1
(1.7 mg), DMF (5 mL) and 100 equiv of Hg(0) (relative to 1). The reac-
tion mixture was stirred at 30 ◦C but no conversion was observed
after 20 h. A similar reaction without Hg furnished the coupling
product in quantitative yield after 24 h.
An oven-dried resealable Schlenk flask was charged with K3PO4
(424.5 mg, 2 mmol), phenylboronic acid (183 mg, 1.5 mmol), 4-
bromoacetophenone (199 mg, 1 mmol), palladacycle 1 (1.7 mg,
0.005 mmol) and DMF (6 mL), and the mixture was stirred at
30 ◦C for 3.5 h (20% conversion as judged by GC). Then 100 equiv
of Hg(0) (relative to 1) were added and conversion was fol-
lowed by GC over 24 h showing that the reaction stopped at 20%
conversion. Under the same free conditions the coupling of 4-
bromoacetophenone gave complete conversion to the coupling
product after 24 h.
2.6. CS2 poisoning studies
An oven-dried resealable Schlenk flask was charged with
palladacycle 1 (1.7 mg, 0.005 mmol), K3PO4 (424.5 mg, 2 mmol),
phenylboronic acid (183 mg, 1.5 mmol), 4-bromoacetophenone
(199 mg, 1 mmol) and DMF (6 mL), and the mixture was stirred at
25 ◦C for 2 h (18–22% conversion). A solution of CS2 in DMF was then
added (0, 0.25, 0.375, 0.5, 1, 2 equiv. of CS2). The mixture was stirred

D. Zim et al. / Journal of Molecular Catalysis A: Chemical 287 (2008) 16–23
19
Scheme 2. Decomposition of palladacycle 1 under stoichiometric conditions.
Scheme 3. Suzuki coupling reaction monitored by 19 F NMR and GC–MS.
suppresses or decreases the formation of this by-product the same
reaction was carried out in the presence of triphenylphosphine. The
yield of 7 was improved to 92% and 8 was still formed but in a lower
yield (2%).
As proposed for phosphine adduct palladacycles [16–18], the
generation of Pd(0) species from the palladacycle 1 may be
explained by a process involving a nucleophilic attack of the
boronate species at the metal center followed by a reductive
elimination. However, we cannot exclude that the Pd(0) species
can be also generated in a pathway involving a reduction pro-
cess to generate the organic thioether 8. Similar reduced organic
products have been observed in the stoichiometric studies of
the Pd-catalyzed Stille [20], amination [20] and cycloamida-
tion [21] reactions. For instance, the stoichiometric reaction of
Scheme 4. Postulated mechanism for the formation of active Pd(0) catalysts from the palladacyle 1.

20
D. Zim et al. / Journal of Molecular Catalysis A: Chemical 287 (2008) 16–23
Herrmann’s palladacycle with Me3SnPh gave the ortho-arylated
P(o-Tol)2(C6H4-o-CH2Ph) (73%) and the reduced P(o-Tol)3 (14%)
[20]. The thioethers 7 and 8 formed under stoichiometric conditions
were also observed under catalytic conditions. Thus the coupling
reaction of 4-bromotoluene with phenylboronic acid catalyzed by
2 mol% of 1 at 80 ◦C was followed by GC and GC–MS analysis. The
thioethers 7 and 8 were formed in 50 and 15% yields based on 1,
respectively.
In order to gain more evidence for the palladacycle acti-
vation under catalytic conditions, we have synthesized a new
palladacycle containing a fluorine substituent on the aromatic
ring and followed its reactivity by 19 F NMR. The preparation of
the sulfur-containing palladacycle 9 was accomplished using a
similar procedure described earlier [9]. Thus, the reaction of 1-(1-
tert-butylsulfanylethyl)-4-fluorobenzene with palladium acetate in
acetic acid at 90 ◦C affords the dimeric acetato-bridged palladacycle
9 in moderate yield (45%). This air- and water-stable orange com-
pound has the same activity for the Suzuki coupling reaction of aryl
halides as the other sulfur-containing palladacycles.
Although a positive response to a mercury poisoning test alone can-
not be taken as a confirmation of the presence of heterogeneous
catalytically active palladium species in this reaction, it is reliable
evidence for a catalytic cycle involving a Pd(0) intermediate.
Quantitative poisoning experiments with CS2 have been
employed to determine the number of active metal atoms and
provide evidence of the nature of the catalyst (homogeneous
or heterogeneous) [29]. This method was used to evaluate the
turnover number for a palladium-containing perovskite catalyst
system in the Suzuki reaction [30]. We performed the CS2 poi-
soning experiment for the coupling of 4-bromoacetophenone with
phenylboronic acid (1 mol% 1, 2 equivs K3PO4, DMF). In order to
The coupling of 4-trifluormethylbromobenzene with phenyl-
boronic acid promoted by palladacycle 9 was monitored by 19 F
NMR (Scheme 3) [22]. The 4-trifluormethylbiphenyl coupling prod-
uct was detected within the first half hour (ı −64.9 ppm). During
this time the signal for the palladacycle became broader and after
1 h a new broad signal appeared at −122.0 ppm. As the resonance
for 9 decreased, the resonance at −122.0 ppm increased and the
signal for the palladacycle 9 disappeared after 4.5 h. The only
change in the 19 F NMR spectra after this time was in the relative
intensity of the coupling product trifluormethylbiphenyl. After 15 h
the mixture was analyzed by GC showing a 25% yield of the 4-
trifluormethylbiphenyl coupling product. Otherwise, GC–MS and
GC analysis showed the ortho-phenyl thioether 10 formed from the
coupling reaction between palladacycle 9 and phenylboronic acid
(47% yield based on 9). Again a reduced product, the tert-butyl 1-(4ꢀ-
fluorophenyl)ethyl sulphide, corresponding to the starting ligand,
and analogous to the thioether 8, was observed by GC–MS, but due
to the small concentration (13% based on 9) it was not detected
by 19 F NMR (ı −119.5 ppm). When the reaction mixture was then
transferred to a resealable Schlenk flask and stirred for 3 h at 130 ◦C,
the conversion was improved to 70%, showing that the catalytically
active species retained activity after the palladacycle was com-
pletely consumed. These results clearly indicate that in our system
palladacycles also serve as a reservoir of catalytic species. On the
basis of this spectroscopic information above, we propose an activa-
tion mechanism that is initiated by the attack of the arylboronic acid
on palladium resulting in the formation of the aryl palladacycle that
undergoes a reductive elimination to form the active Pd(0) species
(Scheme 4, path A). However, we cannot exclude that the Pd(0)
species could also be generated in a pathway involving a reduction
process that regenerates the organic ligand (Scheme 4, path B).
The active Pd(0) species generated from the sulphur-containing
palladacycle could be a soluble palladium complex or colloidal
species. Several techniques have been used to assess the homo-
geneity or heterogeneity of catalyst system [13,23]. For instance, Hg
poisoning tests have been used to analyze the possibilities for Heck
coupling reactions involving palladacycle species [24–28]. We have
performed poisoning experiments for the Suzuki reaction. In the
first reaction, we used 4-bromonitrobenzene, phenylboronic acid,
K3PO4, 0.5 mol% palladacycle 1, and 100 equiv of Hg(0) (relative to
1). The reaction mixture was stirred at 30 ◦C but no conversion
was observed after 20 h. A similar reaction without Hg furnished
the coupling product in quantitative yield. Under the same Hg-free
conditions the coupling of 4-bromoacethophenone gave also com-
plete conversion to the coupling product. However, when Hg was
added before complete conversion (∼20%), the reaction terminated.
Fig. 1. (a) Transmission electron microscopy images of a typical sample of palla-
dium nanoparticles formed during Suzuki reaction using 1 as catalyst precursor. (b)
Histogram illustrating the particle size distribution (512 particles count) that can be
reasonably well fitted by a Gaussian curve; the mean diameter of the Pd (0) particles
was estimated to be around 3.3 nm 0.8.

D. Zim et al. / Journal of Molecular Catalysis A: Chemical 287 (2008) 16–23
21
prevent ligand dissociation the experiments were carried out a low
temperature (25 ◦C). Under these conditions a 20% conversion was
achieved after 2 h. For each experiment CS2 in DMF was added
(0, 0.25, 0.375, 0.5, 1, and 2 equiv. of CS2) and the reactions were
analyzed by GC over 3 days. It was found that 0.375 equiv of CS2
were necessary for completely terminating the reaction, indicating
that only part of the palladium added is active as a catalyst. Similar
behavior wasfound for theHeckreactionwithPCPpincerpalladium
complexes [28]. One limitation of the use of small molecule poi-
son for establishing the stoichiometry is the fact that many catalyst
are thought to slowly release active palladium which can deacti-
vate with time, making the total active species count dynamic. We
performed the same CS2 poisoning experiment in a higher conver-
sion (40%) and temperature (50 ◦C) and the result was the same
(0.375 equiv of CS2). More CS2 (0.5 equiv) was necessary for stop-
ping the reaction at higher temperatures (80 ◦C) and conversions
(70%). However, at this temperature the active site counting using
CS2 may be overestimated since CS2 does not bind irreversibly at
higher temperatures (>50 ◦C). Other limitations of the use of small
molecule poison for establishing the stoichiometry in this case is
that we may have a mixture of nanoparticle and soluble active pal-
ladium species, and also for molecular species with labile solvent
ligands more poison molecules may be needed. Nevertheless, the
poison necessary for stopping the reaction is neither characteristic
of a complete molecular system (≥1 equiv) nor comparable with
heterogeneous catalyts («1.0 equiv.) [23], suggesting problably the
presence of both palladium nanoparticles and molecular palladium
species.
3 nm in average size (Fig. 1). Several methods of synthesis of
palladium nanoparticles and their application for the C C bond
formation have been reported [32–34]. Otherwise generation of
palladium nanoparticles during the catalytic coupling reaction was
also detected or proposed [35]. For instance, Pd nanoparticles were
observed during Heck coupling reaction using oxime palladacycles
as catalyst [8]. Whether the nanoparticles are preformed or gen-
erated in situ the evidence suggests that the Pd species leached
from the nanoparticles surface are the true catalysts [35–37]. The
mechanism proposed for the Heck reaction [13,14,38] involving
Pd atom escape from the palladium nanoparticles by a oxidative
addition generating discrete Pd(II) species in solution was also
proposed to the Suzuki reaction [39]. Alternatively, the Pd(II) com-
plexes can be formed by oxidative addition to Pd(0) atoms that have
already leached into solution [37]. Based on our results and the
considerations above we propose that sulfur containing palladacy-
cles decompose to Pd(0), and an equilibrium between palladium
nanoparticles and the actual active species (Pd(0) atoms or Pd(II)
ions) exists during the catalytic reaction.
In order to gain some insight into the oxidative addition and
transmetallation steps we studied the reaction of different aryl
halides and arylboronic acids catalyzed by palladacycle 1. A plot of
the relative reactivity of substituted aryl bromides (Fig. 2a) and aryl
iodides (Fig. 2b) against the ꢁ Hammett parameter gave a value of
ꢂ = 2.34 and 0.65 for the coupling with phenylboronic acid, respec-
tively. However a different effect was observed for the substituents
on the arylboronic acids. Approximately the same linear correla-
tion was observed for the competitive reaction of bromobenzene
(ꢂ = −0.68, Fig. 2c) and iodobenzene (ꢂ = −0.68, Fig. 2d) with sub-
stituted arylboronic acids.
In addition, if the total active species is dynamic, the ratio release
of active molecular palladium/deactivation remains constant under
the conditions studied as proposed by de Vries for the Heck reaction
[31]. Indeed, transmission electron microscopy analysis of aliquots
taken from the Suzuki coupling between 4-bromoacethophenone
and phenylboronic acid under the same conditions as those used
for the CS2 poisoning experiment (0.5 mol% of 1, DMF, 80 ◦C, 66%
conversion) confirmed the presence of palladium nanoparticles
Dupont et al. have proposed a kinetic model for a competitive
Heck reaction involving two differently substituted aryl halides
where the reaction starts with the two aryl halides competing for
the same catalytically active palladium zero species for the forma-
tion of the respective oxidative addition products and the catalytic
cycle proceeds with their corresponding rate constants [24]. This
Fig. 2. Hammett correlation for the Suzuk reaction at 80 ◦C in DMF promoted by palladacycle 1: (a) effects of substituents on aryl bromides (ꢂ = 2.3; R = 0.93); (b) effects of
substituents on aryl iodides (ꢂ = 0.65; R = 0.96); (c) effects of substituents on aryl boronic acids (ꢂ = −0.68; R = 0.98); (d) effects of substituents on aryl boronic acids (ꢂ = −0.65;
R = 0.96).

22
D. Zim et al. / Journal of Molecular Catalysis A: Chemical 287 (2008) 16–23
kinetic model has shown that, whatever the rate limiting step, in
a competitive Heck reaction the changes in the relative reactivity
involving different aryl halides is strictly bound to the changes in
the rate constant for the oxidative addition step. Experimentally,
they performed a Hammett correlation for the Heck competitive
reaction between bromoarenes and n-butylacrylate and, found that
the four different palladacycles evaluated gave the same ꢂ value
(2.4–2.5), indicating that the same active species are formed in the
oxidative addition step. Interestingly, we have found a very close ꢂ
value (2.3 for Ar–Br, Fig. 2a) for the Suzuki reaction promoted by
palladacycle 1. PdCl2(SEt2)2 is the simplest sulfur-containing palla-
dium complex that mediate the Suzuki cross-coupling reaction of
aryl bromides at room temperature [40]. We have studied the reac-
tivity of different aryl bromides for the Suzuki coupling promoted
by this simple sulfur-containing Pd divalent complex. A plot of the
relative reactivity of substituted aryl bromides against the ꢁ Ham-
mett parameter also gave a linear correlation with the same value
of ꢂ (2.3). These results provide evidence that palladacycle 1 and
PdCl2(SEt2)2 form the same active species in the oxidative addition
step in Suzuki reactions and that these species are equivalent to
those for the Heck reaction promoted by palladacycles. The solu-
ble Pd(II) intermediate can be formed by oxidative addition, either
on the nanoparticle surface or through reaction with Pd(0) atoms
that have already leached into solution. The positive slope shows
that the rate of the oxidative addition increases with the presence
of electron-withdrawing substituents on the aryl halide and that
consequently the transition state for the oxidative addition is stabi-
lized for groups that remove electronic density from the Cipso of the
aryl bromide. Since the halogen–carbon bond is weaker for the aryl
iodides, the influence of the substituents is less pronounced and a
ꢂ value of 0.7 was obtained (Fig. 2b); this value is lower than those
obtained for the same palladacycle for the Heck reaction (ꢂ = 1.8)
[24].
arylated palladacycle. This undergoes a reductive elimination to
form the active Pd(0) species. As a minor process, the Pd(0) species
can be generated in a pathway involving a reduction process that
leads to the decomposition of the palladacycle to Pd(0) affording
the ortho hydrogenated thioether. Poisoning studies and transmis-
sion electron microscopy analysis results indicated the presence
of small palladium nanoparticles that release the soluble actual
catalytic active Pd species. Competitive experiments showed that
electron-withdrawing substituents on the aryl halide and electron-
donating substituents on the arylboronic acid facilitate the reaction.
They also provide evidence that sulfur-containing palladacycle and
PdCl2(SEt2)2 form the same active species in the oxidative addition
step in Suzuki reactions and that these species are equivalent to
those for the Heck reaction using the same palladacycle. The solu-
ble Pd(II) intermediate can be formed by oxidative addition, either
on the nanoparticle surface or through reaction with Pd(0) atoms
that have already leached into solution.
Acknowledgments
We thank CNPq and FAPERGS for partial financial support and
CNPq for scholarship (D.Z. and S.M.N.). We also thank Marcos A,
Gelesky and CME-UFRGS for the TEM analysis and Prof. Jairton
Dupont for fruitful discussions.
References
[1] F. Bellina, A. Carpita, R. Rossi, Synthesis (2004) 2419–2440.
[2] N. Miyaura, Top. Curr. Chem. 219 (2002) 11–59.
[3] S. Kotha, K. Lahiri, D. Kashinath, Tetrahedron 58 (2002) 9633–9695.
[4] A. Suzuki, J. Organomet. Chem. 576 (1999) 147–168.
[5] R.B. Bedford, Chem. Commun. (2003) 1787–1796.
[6] J. Dupont, C.S. Consorti, J. Spencer, Chem. Rev. 105 (2005) 2527–2571.
[7] I.P. Beletskaya, A.V. Cheprakov, J. Organomet. Chem. 689 (2004) 4055–
4082.
Compared to the oxidative addition and reductive elimination
steps, little is known about the transmetalation process [41,42].
Miyaura et al. have isolated and characterized the cationic inter-
mediate [Pd(Ph)(dppe)(PPh3)](BF4) by transmetalation reaction of
[Pd(dppe)(PhCN)2](BF4)2 and PhB(OH)2 in the presence of PPh3
[41]. The reaction of a series of para-substituted arylboronic
acids with this intermediate showed a slightly negative ꢂ value
(−0.54), demonstrating an electronic effect that was accelerated
by the electron-donating substituents. It is worthwhile mention-
ing that we have observed a very similar effect of substituents for
arylboronic acids under catalytic conditions for both bromoben-
zene (ꢂ = −0.68) and iodobenzene (ꢂ = −0.65). Similar correlations
were obtained for the Suzuki cross-coupling of arylboronic acids
with vinyl bromide (ꢂ = −1.26) [43] and with (E)-bromostilbene
(ꢂ = −0.71) [44] using Pd(OAc)2/PPh3 as catalyst precursor. As pro-
posed for oxidative addition, in this case also the changes in the
relative reactivity involving different arylboronic acids are strictly
related to the changes in the rate constant for the elementary
step involving the boron species (transmetalation). The negative
slope indicates that the rate of the reaction increases with the
presence of electron-donating substituents on the arylboronic acid
and this behaviour could be explained by the fact that electron-
donating groups enhance the nucleophilicity of the Cipso of the
arylboronic acid or arylboronate anion for arylation of an aryl-Pd-Br
complex.
[8] E. Alacid, D.A. Alonso, L. Botella, C. Najera, M.C. Pacheco, Chem. Rec. 6 (2006)
117–132.
[9] J. Dupont, A.S. Gruber, G.S. Fonseca, A.L. Monteiro, G. Ebeling, R.A. Burrow,
Organometallics 20 (2001) 171–176.
[10] A.S. Gruber, D. Zim, G. Ebeling, A.L. Monteiro, J. Dupont, Org. Lett. 2 (2000)
1287–1290.
[11] D. Zim, A.S. Gruber, G. Ebeling, J. Dupont, A.L. Monteiro, Org. Lett. 2 (2000)
2881–2884.
[12] P.B. Silveira, V.R. Lando, J. Dupont, A.L. Monteiro, Tetrahedron Lett. 43 (2002)
2327–2329.
[13] N.T.S. Phan, M. Van Der Sluys, C.W. Jones, Adv. Synth. Catal. 348 (2006) 609–679.
[14] J.G. de Vries, Dalton Trans. (2006) 421–429.
[15] R.B. Bedford, C.S.J. Cazin, D. Holder, Coord. Chem. Rev. 248 (2004) 2283–2321.
[16] R.B. Bedford, S.L. Welch, Chem. Commun. (2001) 129–130.
[17] R.B. Bedford, C.S.J. Cazin, M.B. Hursthouse, M.E. Light, K.J. Pike, S. Wimperis, J.
Organomet. Chem. 633 (2001) 173–181.
[18] R.B. Bedford, C.S.J. Cazin, M.B. Hursthouse, M.E. Light, E.J.M. Scordia, Dalton
Trans. (2004) 3864–3868.
[19] J. Spencer, D.P. Sharratt, J. Dupont, A.L. Monteiro, V.I. Reis, M.P. Stracke, F.
Rominger, I.M. McDonald, Organometallics 24 (2005) 5665–5672.
[20] J. Louie, J.F. Hartwig, Angew. Chem., Int. Ed. Engl. 35 (1996) 2359–2361.
[21] A.L. Monteiro, W.M. Davis, J. Braz. Chem. Soc. 15 (2004) 83–95.
[22] For the use of 31 P NMR to evaluate the decomposition of a palladated PCP pincer
complex under Heck conditions see: W.J. Sommer, K.Q. Yu, J.S. Sears, Y.Y. Ji, X.L.
Zheng, R.J. Davis, C.D. Sherrill, C.W. Jones, M. Weck, Organometallics 24 (2005)
4351–4361.
[23] J.A. Widegren, R.G. Finke, J. Mol. Catal. A 198 (2003) 317–341.
[24] C.S. Consorti, F.R. Flores, J. Dupont, J. Am. Chem. Soc. 127 (2005) 12054–12065.
[25] K.Q. Yu, W. Sommer, J.M. Richardson, M. Weck, C.W. Jones, Adv. Synth. Catal.
347 (2005) 161–171.
[26] D.E. Bergbreiter, P.L. Osburn, J.D. Frels, Adv. Synth. Catal. 347 (2005) 172–184.
[27] C.S. Consorti, M.L. Zanini, S. Leal, G. Ebeling, J. Dupont, Org. Lett. 5 (2003)
983–986.
[28] M.R. Eberhard, Org. Lett. 6 (2004) 2125–2128.
4. Conclusion
[29] B.J. Hornstein, J.D. Aiken, R.G. Finke, Inorg. Chem. 41 (2002) 1625–1638.
[30] S.P. Andrews, A.F. Stepan, H. Tanaka, S.V. Ley, M.D. Smith, Adv. Synth. Catal. 347
(2005) 647–654.
[31] A.H.M. de Vries, J. Mulders, J.H.M. Mommers, H.J.M. Henderickx, J.G. de Vries,
Org. Lett. 5 (2003) 3285–3288.
In summary, we have demonstrated that sulfur-containing pal-
ladacycles served as a reservoir of zerovalent palladium species.
The formation of these species is initiated by attack of the aryl-
boronic acid on the palladacycle resulting in the formation of an
[32] D. Astruc, F. Lu, J.R. Aranzaes, Angew. Chem. Int. Ed. 44 (2005) 7852–7872.
[33] M. Moreno-Manas, R. Pleixats, Acc. Chem. Res. 36 (2003) 638–643.

D. Zim et al. / Journal of Molecular Catalysis A: Chemical 287 (2008) 16–23
23
[34] M.T. Reetz, R. Breinbauer, K. Wanninger, Tetrahedron Lett. 37 (1996) 4499–4502.
[39] J. Hu, Y.B. Liu, Langmuir 21 (2005) 2121–2123.
[35] D. Astruc, Inorg. Chem. 46 (2007) 1884–1894.
[36] M.B. Thathagar, J.E. ten Elshof, G. Rothenberg, Angew. Chem. Int. Ed. 45 (2006)
2886–2890.
[40] D. Zim, A.L. Monteiro, J. Dupont, Tetrahedron Lett. 41 (2000) 8199–8202.
[41] T. Nishikata, Y. Yamamoto, N. Miyaura, Organometallics 23 (2004) 4317–
4324.
[37] A.V. Gaikwad, A. Holuigue, M.B. Thathagar, J.E. ten Elshof, G. Rothenberg, Chem.-
A Eur. J. 13 (2007) 6908–6913.
[38] C.C. Cassol, A.P. Umpierre, G. Machado, S.I. Wolke, J. Dupont, J. Am. Chem. Soc.
127 (2005) 3298–3299.
[42] N. Miyaura, J. Organomet. Chem. 653 (2002) 54–57.
[43] V.R. Lando, A.L. Monteiro, Org. Lett. 5 (2003) 2891–2894.
[44] C.M. Nunes, A.L. Monteiro, J. Braz. Chem. Soc. 18 (2007) 1443–1447.