Precatalyst or dosing-device? The [Pd2{μ-(C6H4) PPh2}2{μ-O2C(C6H5)}2] complex anchored on a carboxypolystyrene polymer as an effective supplier of palladium catalytically active nanoparticles for the Suzuki-Miyaura reaction

Article abstract of DOI:10.1016/j.jcat.2019.10.014

A new catalyst has been synthesized from the precursor [Pd₂{μ-(C₆H₄) PPh₂}₂ {μ-O₂C(C₆H₅)}₂] immobilized on a carboxypolystyrene polymer that exhibits an excellent dispersion of the Pd (II) centers, reusability, and catalytic activity in front of phenyl bromides. The activity of this new material was studied in detail for the Suzuki-Miyaura reaction and compared to that of Pd nanoparticles (NPs) supported on UVM-7 (a mesoporous silica), and Pd NPs stabilized with polyvinylpirrolydone. The homogeneous/heterogeneous character of the catalytic process was determined from the results of the hot-filtration, centrifugation, poisoning, three phases tests, and from differential sensitivity kinetic assays. Changes in the palladium oxidation state, and the size and morphology of the Pd NPs formed during several reusing catalytic cycles were determined by XPS, and by SEM and STEM-HAADF respectively. Kinetic and structural analyses concluded that the new material behaved as a supplier of active Pd(0) NPs to the Suzuki-Miyaura catalytic cycle, which was determined to be driven mainly by the palladium fraction present in solution.

Full text of DOI:10.1016/j.jcat.2019.10.014

Journal of Catalysis 381 (2020) 26–37
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Precatalyst or dosing-device? The [Pd2{
l-(C H ) PPh } {l-O C(C H )} ]
6
4
2 2
2
6
5
2
complex anchored on a carboxypolystyrene polymer as an effective
supplier of palladium catalytically active nanoparticles for the Suzuki-
Miyaura reaction
M. Ángeles Úbeda a,1, Pedro Amorós , Juan F. Sánchez-Royo , Jamal El Haskouri , M. Dolores Marcos ,
Francisco Pérez-Pla
a
b,1
b
b
c
⇑
b, ,1
Departament de Química Inorgànica, Universitat de València, Dr. Moliner 50, 46100 Burjassot, València, Spain
Institut de Cíencia dels Materials (ICMUV), c/Catedrático José Beltrán 2, 46980 Paterna, València, Spain
Centro de Reconocimiento Molecular y Desarrollo Tecnológico (IDM), Unidad Mixta Universitat Politècnica de València - Universitat de València. Departamento de
b
c
Química, Universitat Politècnica de València, Camino de Vera s/n, 46022 València, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
6 4 2 2 2 6 5 2
A new catalyst has been synthesized from the precursor [Pd2{l-(C H ) PPh } {l-O C(C H )} ] immobi-
Received 30 July 2019
Revised 23 September 2019
Accepted 9 October 2019
lized on a carboxypolystyrene polymer that exhibits an excellent dispersion of the Pd (II) centers,
reusability, and catalytic activity in front of phenyl bromides. The activity of this new material was stud-
ied in detail for the Suzuki-Miyaura reaction and compared to that of Pd nanoparticles (NPs) supported
on UVM-7 (a mesoporous silica), and Pd NPs stabilized with polyvinylpirrolydone. The homogeneous/
heterogeneous character of the catalytic process was determined from the results of the hot-filtration,
centrifugation, poisoning, three phases tests, and from differential sensitivity kinetic assays. Changes
in the palladium oxidation state, and the size and morphology of the Pd NPs formed during several reus-
ing catalytic cycles were determined by XPS, and by SEM and STEM-HAADF respectively. Kinetic and
structural analyses concluded that the new material behaved as a supplier of active Pd(0) NPs to the
Suzuki-Miyaura catalytic cycle, which was determined to be driven mainly by the palladium fraction pre-
sent in solution.
Keywords:
Suzuki-Miyaura
Palladium nanoparticles
Carboxypolystyrene polymer
Heterogeneous homogeneous catalysis
Kinetic study
STEM-HAADF
HRTEM
Ó 2019 Elsevier Inc. All rights reserved.
XRD
SEM
1
. Introduction
Palladium catalyzed reactions are involved in many of the pro-
compounds have been immobilized in organic, inorganic, or hybrid
supports [8–10].
Polystyrene (PS) and styrene/divinylbenzene (PS-DVB) copoly-
mers are the most popular organic polymeric insoluble supports
because of their stability, mechanical robustness, low cost, high-
liligthing their facile functionalization [11–16]. Among the differ-
ent strategies used in the complex immobilization: covalent
bonding, ion-pair formation, encapsulation, or entrapment, the
most used corresponds to the covalent binding through functional
groups of the modified polymer since this procedure can limit the
leaching of the supported complex.
The palladium immobilized precatalyst undoubtedly become
nanoparticles (NPs) in the first cycle [16]. The sigmoidal reaction
kinetics very often observed in cross-coupling reactions are
assigned to this transformation that contributes the catalytically
active Pd(0) species [17]. Although some authors consider that
the immobilization of homogeneous complexes on an expensive
support that ends up giving NPs, adds complexity to the catalytic
cesses used in organic synthesis [1–3]. In this context, the palla-
dium carbon-carbon coupling reactions have allowed to
synthesized valuable natural products, complex organic molecules,
pharmaceuticals, agricultural chemicals, or advanced materials [4–
7
].
Palladium complexes have used as precatalysts in homoge-
neous catalysis with high reactivity and selectivity although they
show the difficulty of recycling for reuse or contamination of the
products obtained. With the idea in mind of designing sustainable
and economic processes, the use of heterogeneous systems was
proposed and developed quickly. Since 1970s many palladium
⇑
Corresponding author.
E-mail address: francsisco.perez@uv.es (F. Pérez-Pla).
All these authors have contributed equally.
1
https://doi.org/10.1016/j.jcat.2019.10.014
021-9517/Ó 2019 Elsevier Inc. All rights reserved.
0

M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
27
system and none have used for the industry [18], nowadays it is
relevant to deepen into the knowledge of which species are active
in the heterogeneous catalysis as well as if the catalytic process is
truly homogeneous or heterogeneous [19–22], since it is the start-
ing point for the design of more selective and efficient catalysts.
The ‘‘cocktail-of-species” model is currently being considered
Surface area values were calculated from nitrogen adsorption-
desorption iso-therms (ꢀ196 °C) recorded on a Micromeritics
ASAP-2020 automated analyzer.
X-ray photoelectron spectra (XPS) were recorded with an Omi-
cron spectrometer equipped with an EA-125 hemispherical elec-
tron multichannel analyzer and an unmonochromatized Mg K
a
[
16,23]. Recently Polynski and Anakinov [17] have demonstrated
X-ray source having radiation energy of 1253.6 eV. XPS spectra of
the precatalyst and the catalysts after several cycles were recorded
in order to follow the evolution.
FTIR spectra of the polymeric support and precatalyst were
characterized before and after the cross-coupling reaction using
an Agilent Cary 630 FTIR spectrophotometer.
through a detailed computational study, the validity of this model
for palladium catalysts in ArX oxidative-addition-dependent reac-
tions. Regardless of the palladium precatalyst, three types of spe-
cies were consider: Pd NPs, Pd molecular species (depending on
2
ꢀ
reaction type, for example X[PhPdX]1ꢀ2
X ) and inactive Pd.
Being the catalysis a kinetic phenomenon it should be consider
that kinetic data can provide information about all these questions.
In a previous work, we have studied the activity of a Pd NPs-UVM-
Kinetic studies. HPLC measurements were carried out using a
chromatograph (Jasco) equipped with an analog-digital converter
(Jasco LC-Net II), quaternary gradient pump (PU-2089), diode array
detector (MD-2018), column oven (Jasco co-2065 Plus), manual
7
catalyst synthesized by immobilizing the precursor (3, see
Scheme 1) on UVM-7 silica [24], and subsequent reduction of Pd
II) with hydrazine [25]. The prepared material was relevant
injector (Reodhyne 720i, loop 5
l
L), and a 150 ꢁ 4.6 mm C-18 Kro-
(
masil column (stationary phase 5
l
m C-18 functionalized silica
1
because of the high dispersion, small size, and high activity of
the silica supported Pd NPs. The kinetic study suggested, however,
that most of the catalytic action fell on the solubilized Pd fraction,
which resulted in a high activity in the Suzuki-Miyaura (SM) reac-
tion but causing a rapid deactivation of the material by the aggre-
gation of the solubilized Pd in the liquid phase. The question that
arises is if the catalyst activity can be maintained, but increasing
its reusability simultaneously. With the aim to answer the ques-
tion, the new Pd(II) carboxy-polystyrene supported material
described below was prepared, and its activity compared to that
of the previously silica supported catalyst. Kinetic studies were
also applied and the transformation of the Pd(II) complex immobi-
lized were also studied.
particles from Scharlau). H NMR spectra were recorded at 298 K
with a Bruker 400 and a 500 AMX spectrometers. Samples were
3 2 2
dissolved in CDCl or CD Cl . Chemical shifts (d in ppm) were
referred to Si(CH ) .
3 4
2.3. Synthesis of the Pd(II) precatalyst (4)
Partial deprotonation of the carboxypolystyrene polymer. Car-
boxypolystyrene resin (1 g) was suspended in a solution of KOH
56.1 mg, 1 mmol) in CH OH (5 mL). The suspension was vigor-
ously stirred for 1 h, filtered, and the solid washed with H O:CH
. The white solid was allowed to
(
3
2
3
-
3 2 2
OH (2:1 v/v), CH OH, and CH Cl
air dry.
Immobilization of complex (1) on the carboxypolystyrene polymer.
Complex (1) (50 mg, 0.028 mmol) was suspended in a mixture of
2
. Experimental
CH
gen atmosphere. AgBF
covered with an aluminum foil, and the suspension stirred for
5 min at room temperature. The resulting yellowish solution
2
Cl
2
(3 ml) and CH
3
CN (0.15 mL) in a Schlenk tube under nitro-
was added (23 mg, 0.118 mmol), the tube
2.1. Reagents and materials
4
4
Commercially available 4-substituted phenylhalides, biphenyl
together with the AgBr precipitate formed were exposed to the
sunlight to reduce Ag(I) to Ag(0). In a second Schlenk tube, the
deprotonated carboxypolystyrene resin (1 g) was suspended in
compounds, and the carboxypolystyrene resin (1% cross-linked
functionalized in 0.5–1.5 mmol/g) were supplied by Aldrich. HPLC
grade solvents CH
lau, 96% v/v extrapure), and CH
ther purification. Complexes [Pd
and [Pd -(C ) PPh -O C(C
were prepared according to the method described in the literature.
2
Cl
2
(Baker), CH
CN (Baker) were used without fur-
-Br) -(C ) PPh ] [26] (1),
)} ] [27] (2 (see Scheme 3)
3
OH (Scharlau), ethanol (Schar-
2 2
CH Cl (5 mL) and, handling carefully the vacuum, the content of
3
the first tube was transferred to the second one under nitrogen
atmosphere through a flexible pipe equipped with a filter. The
new suspension was stirred for 1 h, and the yellow solid, the pre-
4
(
l
4
{
l
6
H
4
2 4
}
2
{
l
6
H
4
2
}
2
{
l
2
H
6 5
2
catalyst [Pd
2
{
l-(C
H
6 4
2 2 2 6 4 2
) PPh } {l-O C(C H - PS)} ] (4), was filtered,
washed with CH
2
Cl
2
and allowed to air dry.
2.2. Instruments
2.4. Synthesis of the stabilized Pd NPs
Characterization of Materials. The palladium and phosphorus
contents were determined by energy dispersive X-ray spec-
troscopy (EDX) analysis using a scanning electron microscope (Phi-
lips - SEM-XL 30). For electron microscopy analyses, the samples
were dispersed in ethanol and placed onto a carbon coated copper
microgrid and left to dry before observation.
H
2
PdCl
4
was freshly prepared by reaction of PdCl
2
(22 mg,
0
.124 mmol) and HCl (1.3 mL, 0.2 M).
polyvinylpyrrolidone (PVP, 211 mg) in deionized water (50 mL)
was added to the yellow reaction mixture and stirred for 30 min.
A
solution of
4 4
Then, NaBH (15 mL, 0.053 M) (Pd:NaBH 1:6.4) was added slowly
The TEM and HRTEM microstructural characterizations were
carried out using a JEOL JEM-1010 instrument operating at
under vigorous magnetic stirring, resulting in the immediate for-
mation of brown Pd(0) NPs. The suspension was evaporated to dry-
ness and the solid was subsequently dispersed by sonication in
0 mL of a mixture of H O/ethanol (1:4 v/v). These NPs were pre-
2
pared for comparative purposes dealing with the catalytic
experiments.
1
00 kV and equipped with a CCD camera and a Tecnai G2 F20(FI)
instrument, respectively. STEM-HAADF images were acquired on
a JEOL-2100F microscope operated at 200 kV. Powder X-ray
diffraction (XRD) was carried out using a Bruker D8 Ad-vance
5
diffractometer with monochromatic Cu K
asource operated at
4
0 kV and 40 mA. Patterns were collected in steps of 0.02° (2h) over
the angular range 1–10.0° (2h), with an acquisition time of 25 s per
step. Additionally, XRD patterns were recorded over a wider angu-
lar range, 10–80°, (2h) to determine the presence of segregated
crystalline phases.
2.5. HPLC calibration and analysis procedures
HLPC standards were prepared by diluting with ethanol stock
solutions containing the 4-substituted phenyl halide (10
ꢀ2
M)

2
8
M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
Scheme 1. Synthesis procedure of the precatalyst (4). (i) Partial deprotonation of the carboxylic groups of the polymer. (ii) Synthesis of the cationic complex (3). (iii)
Immobilization of the Pd(II) on the carboxypolystyrene polymer.
ꢀ
2
6
cycles. The experiment was repeated doubling the analytical con-
and its reaction product (0:5 ꢀ 1:0 ꢁ 10 M) in ethanol. The chro-
centration of Pd.
matographic separation of the reaction samples was carried out
3 2
passing as the eluent a solution of CH CN and H O (0.1% v/v in
HCl) in a ratio 70:30 v/v through a C-18 reverse phase column ther-
mostated at 40 °C at a rate flow of 1.0 mL/min.
2
.8. Investigation of the homogeneous vs. heterogeneous nature of
catalysis
2
.6. General procedure for the kinetic study of the Suzuki-Miyaura
Hot filtration test. 4-iodoanisole (118.0 mg, 0.504 mmol),
phenylboronic acid (92.8 mg, 0.761 mmol), CO (104.0 mg,
.753 mmol), and the precatalyst (47.1 mg, ꢃ1% in Pd relative to
reaction
K
2
3
0
In a typical experiment, the 4-substituted phenyl halide (XC
R X = I, Br; R = NO , COCH , H, OCH , CH ; 0.5 mmol), phenyl-
boronic acid (0.75 mmol), CO (0.75 mmol), and ethanol
50 mL, 96% v/v) were introduced into a 2-neck flask equipped
6
-
the substrate) were added to the reactor and allowed to react in
ethanol (50 mL) at 60 °C for 15 min. By this time, a sample of
the reaction mixture (12 mL) was taken, filtered, and the super-
natant poured on a second flask maintained at the work tempera-
ture, allowing the reaction to continue for 24 h. The samples taken
before and after the filtration were analyzed by the HPLC technique
described above.
H
4
2
3
3
3
K
2
3
(
with a cooler and a thermostat jacket attached to a water bath.
After heating at 60 °C, a sample of 0.3 mL was taken (the blank)
and the precatalyst added (typically in 1 mol % with respect to
ꢀ4
the 4-substituted phenyl halide, 45 mg, ½Pdꢂ ¼ 1:0 ꢁ 10 M). The
progress of the reaction was monitored taking aliquots (0.3 mL)
at regular time intervals. After sampling, aliquots were poured into
vials containing ethanol (96%, 2 mL) and HCl (0.15 M, 0.1 mL) as
the quencher. The resulting solution was filtered with a 0.02
pore size microfilter (from Whatman), and portions of 5 L of the
Centrifugation test. 4-iodoanisole (117.2 mg, 0.501 mmol),
phenylboronic acid (93.0 mg, 0.763 mmol), K2CO3 (104.0 mg,
0.752 mmol), and the precatalyst (47.0 mg, ꢃ1% in Pd relative to
the substrate) were allowed to react in ethanol (50 mL) at 60 °C.
A portion of the reactive mixture (20 mL) was removed after
14 min. of reaction, centrifuged, and the supernatant liquid
(12 mL) poured into another flask maintained at the work temper-
ature. The progress of the reaction in both flasks was monitored by
HPLC.
lm
l
filtrate were analyzed according to the HPLC procedure described
above.
2.7. Reusing of the catalyst by centrifugation/washing procedure
Catalyst poisoning. 4-iodoanisole (112.3 mg, 0.480 mg), phenyl-
boronic acid (99.1 mg, 0.812 mmol),
K
2
CO
3
(109.6 mg,
The study of the precatalyst reusing was carried out following
0.790 mmol), and the precatalyst (47.0 mg, ꢃ1% in Pd relative to
the substrate) were allowed to react in ethanol (50 mL) at 60 °C.
QuadraPure TU (500 mg) was added after 36 min. of reaction. The
progress of the reaction was monitored by HPLC.
Three phases test. The three phases test was carried out follow-
ing the procedure published by Crudden et col. [28], Kantman et
col. [29], and Gruber-Woelfler [30]. Amino funtionalized silica gel
the general procedure, using 4-iodoanisole as the substrate. The
first cycle was completed after 120 min, then the catalyst was sep-
arated by centrifugation and washed with ethanol to remove the
reactant and product residues. The recovered solid was dried at
7
0 °C for 24 h and used in the next cycle. The addition of reagents
in each cycle was that specified in the general procedure except for
CO , the mass of which was reduced by half (0.375 mmol) from
K
2
3
(BrPhCONH
described in the literature. 4-iodoanisole (93.3 mg, 0.400 mmol),
phenylboronic acid (104.1 mg, 0.854 mmol), K CO (117.1 mg,
2 2
–SiO ) was synthesized according to the method
the second cycle, as it was not completely removed during wash-
ing. The process was carried out several times reaching a total of
2
3

M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
29
0
.847 mmol), and ethanol (50 mL) were introduced in a reaction
The value of p and q were set up to 1, according to the results of a
flask, and the mixture heated up to 60 °C. After blank extraction,
the precatalyst (46.2 mg, ꢃ1% in Pd) and functionalized
previous kinetic study carried out on the Pd NPs-UVM-7 system
[25]. In Eqs. (6) and (7), k
i
ðtÞ are empirical functions of time taking
BrPhCONH
2
–SiO
2
silica gel (500 mg) were introduced into the reac-
into account the variation of the catalyst activity along the reaction
time interval originated by the change in the concentration of the
active Pd species. Specifically, the mathematical expression of such
functions is given by Eq. (8),
tor. After 60 min. of reaction the solid was separated by filtration,
and washed with ethanol and CH Cl . The solid was hydrolyzed
2
2
following the procedure described in the literature. The 4-
bromobenzoic acid and [1,1’-biphenyl]-4-carboxylic acid obtained
a
i;1 þ ai;2
t
1
were analyzed by H NMR spectroscopy.
kiðtÞ ¼
ð8Þ
1
þ ai;3t2
2.9. Schmidt’s analysis of differential selectivity
The concentration profiles predicted by the model were calculated
from the numerical integration of the ordinary differential equation
system (ODE) depicted by Eq. (5). After substitution of concentra-
tion curves into Eqs. (1) and (2), the theoretical yields predicted
Schmidt’s differential sensitivity studies (SADS) [21,31,32] were
performed using the 4-iodoanisole/4-iodotoluene and bro-
moanisol/bromotoluene substrate pairs with the aim to compare
the catalytic active species of precatalyst (4) with those present
when stabilized Pd NPs, or complex (2) were used as catalysts. In
a typical experiment, the phenyl halide pair (0.25 mmol in each),
i
by the model (g ) were calculated. Finally, the empirical aij param-
^
eters in Eq. (8) were estimated by minimizing the least-squares pre-
dictor / in Eq. (9) with regard these parameters. In Eq. (9), T is the
transpose matrix operator.
phenylboronic acid (0.75 mmol), K
(
(
2
CO
ꢃ1% in Pd with respect to the total substrate), and ethanol
50 mL) were introduced into the reactor at 60°. The reaction kinet-
3
(0.75 mmol), the catalyst
T
/ðaÞ ¼ gg
^
ð9Þ
Calculations were performed with an ad hoc software developed
in our laboratory written in the Julia programming language [33].
Minimization of Eq. (9) was carried out using the Nelder-Mead
method [34] implemented in the NLOPT library [35], and integra-
tion of the non-stiff ODE in Eq. (5) was performed using the
Runge-Kutta like method of Tsitouras [36] implemented in the Dif-
ferentialEquations Julia’s package [37].
ics was followed by HPLC for the iodoanisole/iodotoluene system
according to the general procedure described above. For the 4-
bromoanisole/4-bromotoluene pair, the reaction mixture could
not be efficiently separated by HPLC. In this case, analyses were
carried out by GC (see supplementary information).
3
. Methodology of kinetic data analysis
4
. Results and discussion
Concentrations of biphenyl, 4-substituted phenyl halides, and
their corresponding 4-substituted biphenylic derivatives were esti-
mated from the chromatographic peak areas using the response
factors calculated from the calibration plots. The yield of 4-
4
.1. Precatalyst synthesis and characterization
Scheme 1 summarizes the synthesis procedure. The solvated
substituted biphenyl products ( 1
calculated by means of Eqs. (1) and (2),
g
ðtÞ), and biphenyl (g2ðtÞ) were
complex (3) was immobilized on the carboxypolystyrene material
by reaction with the carboxylate groups that easily replace the ace-
tonitrile molecules. Thus, the palladium complex was covalently
tethered to the polymer through bridging carboxylate groups giv-
ing the precatalyst [Pd {l-(C H ) PPh } {l-O C(C H -PS)} ] labeled
2 6 4 2 2 2 6 4 2
as (4).
The homogeneous pale yellow color of the precatalyst is a good
indication of the complex incorporation on the polymer resin. We
have used EDX to assess the presence and stoichiometry of the
organometallic Pd complex in the precatalyst as well as the chem-
ical homogeneity of the complex along the polymer matrix, given
that a principal objective of our work at this stage was to favor a
good dispersion of the precatalytic active sites.
g1 ¼ ½P
g2 ¼ ½P
1
ꢂ=½Aꢂ
ꢂ=½Bꢂ
ð1Þ
ð2Þ
0
2
0
where [A]
0
and [B]
0
stand for the zero time concentration of
4
1
-substituted phenyl halides and phenylboronic acid, whereas [P ]
2
and [P ] represent those of the corresponding 4-substituted biphe-
nyl products and biphenyl respectively. The experimental yields
were compared with those calculated from the chemical model
shown in Eqs. (3) and (4), which consider the SM cross-coupling
and the boronic acid oxidative homocoupling reactions, both of
them catalyzed by Pd,
The EDX value were averaged from ca. 50 individual measure-
ments, with a final value corresponding to a Pd/P = 0.98 M ratio.
This value fits very well with the expected ratio taking into account
that the real anchored species are dimeric complexes with the Pd
metal centers connected through two phosphine-related ligands.
k ðtÞ
1
A þ B ! P
1
ð3Þ
ð4Þ
Pd
k ðtÞ
2
B þ B ! P
2
Pd
Our data, also confirm that the precatalyst is chemically homoge-
The above formal mechanism has associated the mathematical
model consisting of Eq. (5),
2
neous (at least) at the micro-scale level (spot area of 1
lm ).
Hence, the precatalyst can be considered as monophasic product,
and segregation of the organometallic complex can be discarded.
The covalent anchoring through the carboxylate groups of the
polymer can be inferred when the IR spectra of the resin and the
precatalyst are compared (see Fig. S1 in supplementary informa-
tion). In both spectra it can be appreciated several signals in the
0
1
0
1
½
Aꢂ
ꢀ1
0
ꢀ r
ꢁ
B
C
B
C
d B ½Bꢂ C B ꢀ1 ꢀ2 C
1
2
B
C ¼ B
C ꢁ
ð5Þ
dt @ ½P ꢂ A
@
1
0
0 A
r
1
½
P
2
ꢂ
1
ꢀ1
7
00–800, 1000–1100 and around 1500 cm that can be assigned
in which r
1
2
and r are the reaction rates for processes (3) and (4),
to CAH bending out of plane, CAH in plane and CAC stretching
calculated according to Eqs. (6) and (7),
vibrations, respectively. The starting resin shows an intense peak
ꢀ
1
ꢀ1
p
q
(at 1687 cm ) and an additional signal around 1220 cm that
can be attributed to C@O and CAOH stretching vibrations of the
carboxylate groups. After the complex inclusion, the intensity of
r
r
1
2
¼ k
¼ k
1
2
ðtÞ½Aꢂ ½Bꢂ
ðtÞ½Bꢂ
ð6Þ
ð7Þ
2

3
0
M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
ꢀ
1
1
the signals at ca. 1220 cm
intense peak at 1687 cm
1
strongly decreases. Moreover, the
disappears and two new signals at
544 and 1395 cmꢀ1, associated to
as(OACAO) and (OACAO)
kðtÞ were found to be monotonically decreasing functions with
maximum rate at zero time [25]. In general, the silica supported
catalyst presented higher activities for 4-substituted phenyl
iodides. However, this type of catalysts undergone rapid deactiva-
tion, leading to particularly low yields when using phenyl bromide
ꢀ
m
m
s
vibrations can be observed. This evolution strongly suggests that
the Pd-complexes are effectively anchored on the polymer. The
XPS spectrum confirms that the oxidation state of the Pd is pre-
served after the anchoring step. In fact, the Pd 3d spectrum shows
two signals at binding energy values of 337.5 and 342.8 eV (see
further below Fig. 7 in Section 4.6) that, according to the bibliogra-
phy, are due to Pd(II) metal centers.
The STEM-HAADF image in Fig. 1 supports the Pd(II) oxidation
state. In fact, only an extremely low number of white spots (due
to the presence of Pd(0) nanoparticles) is observed (see below for
details). In any case, the mapping on Pd(0)-free domains reveals
the presence and homogeneous distribution of Pd and P, now at
the nanoscale.
derivatives substituted with electrodonor radicals as substrates
ꢄ
(
g1ð24h; 80 CÞ ¼ 34, 16, 10% for Br-Ph-R with R = H, OCH3,
CH [25]). This issue makes the material of little interest to catalyze
3
reactions involving low activated phenyl halides. Although the
polymeric material reacted more slowly for 4-substituted phenyl
iodides, it was quite more active for 4-substituted phenyl bromides
than the silica-supported material. Even more interesting was the
fact that the yields for the polymer supported system were higher
ꢄ
for all substrates (e.g. g1ð5h; 60 CÞ ¼ 87; 76; 68% for Br-Ph-R with
R = H, OCH , CH ), which makes the use of the polymeric support
3
3
especially attractive when using weakly activated phenyl halides.
The induction period suggested that Pd(II) is not the active cat-
alyst, and hence Pd(II) complexes in material (4) must undergo
reduction. The activity was rationalized assuming that small
amounts of Pd atoms coming from the Pd(II) reduction enter into
the solution, starting the SM catalytic cycle. The activity increases
while the reduction in progress supplies Pd atoms. On the other
hand, the aggregation process leads to a gradual activity loss as
the solubilized Pd species form larger aggregates. Obviously, the
activity curves must go through a maximum as a consequence of
the opposite effect caused by the reduction/aggregation processes.
This behavior differs from that observed for the Pd NPs-UVM-7 cat-
alyst. For this material, Pd(0) enters massively the dissolution as
small Pd NPs when the reactor reaches a critical temperature
(80 °C), resulting in a high initial reaction rate. Nevertheless, the
activity drops rapidly due to the aggregation phenomenon, which
occurs in a greater extent because of the higher Pd(0) concentra-
tion in the solution.
The difference in activity for phenylbromides presented by the
polymeric material and Pd NPs-UVM-7 can be understood by ana-
lyzing the way in which the material introduces the Pd active spe-
cies in the medium. The polymer-immobilized catalyst provides
small amounts of Pd atoms continuously, hence it catalyzes the
reaction as long as the polymer is capable to do this task. In con-
trast, the silica material delivers most of Pd(0) as Pd NPs at the
beginning of the reaction. As mentioned, the Pd NPs aggregate fas-
ter as they are larger and more concentrated, resulting in an earlier
depletion of the most active Pd species. These differences are also
related to the different nature of the materials: porous vs. non-
porous when the silica or the polymer is used as support. In order
to provide Pd NPs, the polymer needs first to be swollen through
the solvent action, then forming the Pd(0) species, and finally dif-
fuse (more hindered when compared to the open bimodal pore
system of the UVM-7). The other determinant factor is the relative
rate of the aggregation respect to the reaction rate. If the aggrega-
tion process is faster than the chemical reaction, the early deple-
tion of the active Pd will cause a decrease in the reaction yield.
In this line, the yield of the slow reacting bromo derivatives is
higher when using the catalyst immobilized on the polymer, as
the active species are not exhausted throughout the course of the
reaction, whereas the decrease in yield was not observed for the
faster reacting 4-iodo phenyl derivatives, all of them reaching near
4
.2. Catalyst activity
The activity of precatalyst (4) was checked for the SM reaction
of boronic acid with the 4-substituted phenylhalides shown in
Scheme 2, following the methodology previously described. The
Pd(II)-carboxypolystyrene material catalyzed efficiently the reac-
tion at 60 °C. This made a difference with the catalyst Pd NPs-
UVM-7 that did not show significant activity below 80 °C [25].
Fig. 2(a) and Fig. S2 (supplementary information) show the kinetic
behavior of 4-substituted phenyl bromides and phenyl iodides
respectively. The reactivity order was that expected for a SM reac-
tion controlled by the transmetallation step (i.e. the reaction
became faster when increasing the substituent electronegativity
for substrates sharing the same halide) [38–40]. As expected, the
rate for 4-substituted phenyl iodides was higher than that of their
corresponding bromo derivatives due to the left-displacement that
occurs in the oxidative addition equilibrium involving the phenyl
bromides. The equilibrium shift was caused by the higher energy
of the CABr bond in relation to that of the CAI bond.
Fig. 2(b) shows the variation in activity of the catalyst quanti-
fied by means of the kðtÞ function, which reflects the change in
the nature and concentration of the Pd species inside the reactor
[
41,42]. The basic idea behind this methodology is the assumption
that all Pd species (in the liquid phase and anchored on the catalyst
surface) have different activities (k ), so that the observed catalytic
i
activity is calculated as the sum of that of all species,
X
d½Aꢂ
ꢀ
¼ ð
k
i
x
i
ðtÞÞ½Pdꢂ ½Aꢂ½Bꢂ ¼ kðtÞ½Pdꢂ ½Aꢂ½Bꢂ
ð10Þ
T
T
dt
i
The first-order dependence of Eq. (10) on the analytical concentra-
tion of Pd (½Pdꢂ ) was checked (see Fig. S3 in supplementary infor-
T
mation). Obviously, if the mole fraction of Pd species (x
i
ðtÞ) varies,
e.g. due to dissolution, aggregation, or deposition processes, the
activity of the catalyst will change over time [43]. The function
inside the summatory requires a detailed knowledge of the distri-
bution of Pd species and their specific activity. Usually, the last
information is not experimentally available and, for this reason,
the empirical kðtÞ function was used in this work to perform a
semi-quantitative analysis of the activity. Table 1 collects the calcu-
lated parameters defining the catalyst activity namely, the value of
1
00% conversion.
kðtÞ (kmax), the formal catalytic constant (kcat ¼ kmax=½Pdꢂ ), the turn-
T
over frequency at the maximum activity (TOF ¼ kmax½Aꢂ½Bꢂ=½Pdꢂ ),
T
and the experimental yields of the SM products and biphenyl at 5 h.
Remarkable differences in activity were found for the silica and
polymer supported catalysts. All reactions catalyzed by the poly-
meric material exhibited an induction period, their activity kðtÞ
curves going through a maximum as shown in Fig. 2(b). This
extreme was absent for the silica-supported material, for which
4.3. Reusing catalyst experiments
Fig. 3(a) and (b) shows the change of yield and activity along 6
reusing cycles. All the graphs indicated that the activity remained
high and constant during the first 4 cycles, observing an abrupt
decrease from the fifth one, and the activity declining until a prac-

M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
31
Fig. 1. STEM-HAADF image of the precatalyst containing the Pd binuclear organometallic complex anchored on the polymer surface. Arrows mark white spots associated with
isolated and sparse Pd NPs. The mapping carried out on the white square shows the presence of Pd and P dispersed along the polymer.
Scheme 2. The Suzuki-Miyaura reaction.
tically null value by the 6th cycle. The observation was consistent
with the palladium-reservoir role played by the material. Correlat-
ing with the activity loss, the appearance of the recovered material
varied from the bright yellow of the catalyst before use (originated
by the phosphine Pd(II) complexes) to the dark gray of the material
recovered after the 6th cycle. This change in color clearly indicated
a progressive reduction of the Pd(II) complexes, and the subse-
quent deposition of Pd(0) on the polymer during the cooling/cen-
trifugation processes carried out between cycles. The effect of
using a polycarboxystyrene polymer with double load of catalyst
is shown in Fig. S4 (see supplementary information). Higher activ-
ity was observed together with a slower deactivation of the mate-
rial along the cycles.
Table 2 collects the relative abundance of P, Pd, and K deter-
mined from scan electron microscopy (SEM) for the first, third,
and sixth reusing cycles (see Figs. S5 and S6 in supplementary
information). The P signal was originated by the phosphine com-
plexes, whereas the Pd signal was originated by both the Pd(II)
bound to phosphorus (ratio P:Pd is 1:1) and the Pd(0) deposited
on the polymer. Consequently, the ratio P/Pd was proportional to
the mole fraction of Pd(II). SEM analysis indicated that Pd was
found mainly as Pd(II) (96%) after the first cycle, this amount
decreasing to 73% after the third one. Composition drastically
changed after the sixth cycle, where the P signal was practically
absent, therefore the catalyst consisting of Pd(0) mostly. These
results also suggested that Pd(II) reduction implied the removal
of the phosphine ligands from polystyrene. The results are indica-
tive that the SM reaction is occurring as long as the catalyst con-
tains a certain pool of Pd(II). Additionally, Pd(0) aggregates
formed along cycles have little or null activity as evidenced by
the lack of activity of 6th cycle, for which Pd(II) is exhausted but
the material is still containing Pd(0). Therefore, kinetic results
and SEM analysis allow concluding that the Pd(0) formed/solubi-
lized along the reaction probably supports most of the catalytic
action.
4.4. Classical tests to check the homogeneous vs. heterogeneous nature
of the catalysis
The question raised now is how Pd(0) develops its catalytic
action. One possibility is that the Pd(0) fixed on the polymer sur-
face is the active catalyst, or, at the other end, the catalysis is basi-
cally driven by the solubilized fraction of Pd(0). With the aim to
investigate this issue, the classical hot-filtrate and centrifugation
tests were carried out according to the procedures described
previously.
Hot filtration and centrifugation experiments. Fig. S7 (see supple-
mentary information) shows the evolution of the reaction mixture
before and after the filtration process when using
4
-methoxyiodobenzene as the substrate. The reaction stops
completely in the filtrate, while it continues in the flask whose
reactants are in contact with the catalyst. Analogous comments
can be made after the observation of Fig. S8 (see supplementary
information), where the results related to the centrifugation test
are shown.

3
2
M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
Fig. 2. (a) Change with time of the reaction yield for the tested 4-substituted phenyl bromides indicated at the plot legend catalyzed by precatalyst (4). Solid lines were
calculated by fitting the experimental yields (symbols) to the mathematical model defined by Eqs. (3)–(5); (b) estimated variation of k
over frequency estimated at the maximum catalyst activity.
1
ðtÞ calculated from Eq. (8); (c) turn-
Table 1
6 4
Kinetic and statistical fitting parameters for the SM reaction for the studied 4-substituted phenylhalides (X–C H
–R) catalyzed by precatalyst (4).a
ꢀ
1
ꢀ1
cat ꢁ 10ꢀ4/minꢀ1
ꢀ2
TOFmax ꢁ10ꢀ3/hꢀ1
g1(5 h)b
g2(5 h)b
LOF(%)c
X
I
R
k
max/min
M
R
k
M
COCH
OCH
CH
H
3
d
23
8.5
6.4
11
23
8.0
6.4
11
2.4
0.7
0.5
1.0
100
99
99
0
4.7
0
0.9991
0.9994
0.997
3.3
2.9
6.5
1.8
3
3
99
0
0.9998
NO
COCH
OCH
2
47
17
2.5
1.6
0.6
44
16
2.4
1.4
0.6
3.9
1.7
0.2
0.1
0.03
100
100
87
76
68
11
0
6.3
24
0
0.9999
0.9993
0.990
0.990
0.995
1.3
3.3
11
12
7.6
Br
3
3
CH
H
3
a
b
c
ꢀ4
In all kinetic runs: [Pd] = 1:0 ꢁ 10 mol/L.
: SM reaction yield, :biphenyl yield.
LOF: lack of fit in %.
g
g
2
1
d
A turn-over number of 125 mol substrate/mol catalyst was calculated for the 4-iodoanislole substrate (see supplementary information).
We may be tempted to conclude that the reaction is mainly dri-
of filtration and centrifugation tests were not conclusive due to the
possibility of Pd elimination from solution during the sample han-
dling. In order to demonstrate the existence of active Pd within the
solution, two additional tests were carried out, namely the Pd-
poisoning and the three phase tests.
Poisoning catalyst test. Fig. S9 (see supplementary information)
shows that the reaction stops just after the Quadrapure TU (poison)
addition (50 min) [31,44]. By this time, the product yield stabilizes
at 64% that differs from the 85% value observed at 50 min for the
reaction carried out in absence of poison. Concurring with the
hypothesis raised above, this result strongly suggests that the sol-
ubilized Pd fraction is the main responsible for the catalytic action.
ven by the Pd(0) clusters existing on the surface of the catalyst.
However, if that is the case, it is difficult to understand why the
material isolated in cycle 6 does not exhibit any catalytic activity.
An alternative explanation is to admit that the solubilized species
of Pd(0) suffer a coalescence process in solution during filtering
upon lowering the temperature that causes their deposition on
the polymeric material (‘‘boomerang system”) [16]. According to
previous assumptions the catalysis would be mainly conducted
by the solubilized Pd(0) and, consequently, the lack of activity of
the filtrate would be a consequence of the Pd removal during the
cooling/filtering/centrifugation processes. It is obvious that results

M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
33
Fig. 3. (a) Time variation of the SM reaction yield for the 4-methoxyphenyl iodide along the several reusing cycles displayed at the plot legend catalyzed by precatlyst (4).
Solid lines were calculated by fitting the experimental yields (symbols) to the mathematical model defined by Eqs. (3)–(5); (b) estimated variation of k
ðtÞ calculated from Eq.
8); (c) turn-over frequency estimated at the maximum catalyst activity.
1
(
Table 2
4.5. Schmidt’s analysis of differential selectivity
Relative abundance of P, Pd, and K atoms (in %) in precatalyst (4) calculated from SEM
analysis for selected reusing cycles.
The SADS test allows comparing the active species of two cata-
Cycle
P
Pd
Kb
P/Pd
lysts. If one of them is immobilized and the other is in solution, the
test will allow differentiating when the catalytic action is carried
out by the solubilized fraction of Pd or by that immobilized on
the polymeric support. The test consists of two kinetic runs carried
out in identical experimental conditions but differing in the cata-
lyst used. For this purpose, the behavior of precatalyst (4) was
compared with the molecular compound (2) (see Scheme 3), and
with the stabilized Pd NPs. These Pd NPs showed sizes lower than
1a
3
6
4.9
22.4
0.04
5.1
30.5
14.9
90
47.1
85.1
0.96
0.73
0.002
a
Catalyst before use.
b
The K signal resulted unhelpful for the catalyst characterization as there were
ꢀ
+
two potassium sources, namely the carboxylate groups of the resin (COO K ), and a
variable amount of K
process.
2 3
CO residues that remained on the surface after the washing
5
nm and were stabilized by the PVP polymer (see Fig. S13 in sup-
plementary information). In each experiment, two phenyl halides
were simultaneously reacted until their complete depletion, and
Three phases test. Fig. S10(b) (supplementary information)
shows H NMR spectrum of the reaction mixture after the hydrol-
ysis treatment, which was fully consistent with the presence of
-bromobenzoic acid (signals at 7.98 and 7.64 ppm, integration
ratio 2:2) and [1,1’-biphenyl]-4-carboxylic acid (signals at 8.19,
.72, and 7.51 ppm, integration ratio 2:2:3) in an approximate
ratio 70:30. The detection of [1,1’-biphenyl]-4-carboxylic acid con-
firmed that the solubilized Pd species are involved in the catalytic
process.
the concentration of the corresponding biaryl products (P
1
and
) was measured along the reaction time. The concentration of
was plotted against that of P so that the coincidence of the tra-
], [P ]) ensured that both catalysts
1
P
P
2
1
2
4
jectories on the phase plane ([P
1
1
shared the same active species. On the contrary, if the catalysis
were driven by dissimilar Pd species, then a notable divergence
between trajectories should be observed.
Fig. 4(a) shows that the trajectories associated with the immo-
bilized material and stabilized Pd NPs consist of straight lines of
7

3
4
M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
The conclusion to be drawn is that the catalytic active species of
the Pd(II) supported polymer and stabilized Pd NPs must be very
similar, i.e. the catalysis must be driven by the solubilized Pd(0),
and hence the Pd(II) supported on the polymer is acting as a mere
Pd(0) supplier. Furthermore, the SADS diagram allows discarding
the supply of active Pd(II) phosphine complexes to solution from
the resin, as in such a case the trajectory associated with the
immobilized catalyst should be superimposed on that of (2). A
SADS experiment using 4-bromotoluene and 4-bromoanisole as
substrates yielded similar results, see Fig. 4(b). For this system,
however, the three catalysts exhibited linear SADS plots suggesting
that both Pd(II) complexes (2 and 4) reduce to Pd(0). This behavior
can be rationalized assuming that the lower reaction rate of bro-
moderivatives allows for the previous reduction of (2) in the
homogeneous phase.
Scheme 3. Structure dinuclear Pd(II) complex.
Fig. S11 (supplementary information) shows the yield in the
biarylic product of the 4-iodoanisole measured from the SADS
experiment when using as catalysts the immobilized material
(4), and stabilized Pd NPs. It is apparent that the difference
between both materials lies in the induction period (present only
in the immobilized catalyst), the conversion, which reaches practi-
cally 100 % in the case of the immobilized complex, but reaching
only 60 % for the stabilized Pd NPs, and the initial reaction rate,
which is clearly greater for the Pd(0) NPs material. Again, the dif-
ferences in activity can be rationalized if it is assumed that the sol-
ubilized fraction of Pd(0) is the one with the highest catalytic
activity. Indeed, Pd NPs exhibit initially a high reaction rate as they
consist basically of stabilized Pd(0) nanoclusters. On the contrary,
an induction period is observed for the immobilized material
caused by the time it takes to reduce some of Pd (II), which slows
down the conversion rate due to the lack of active Pd during the
initial phase of the reaction. Finally, the observed difference in
the biaryl product yields should arise from the different rates of
other processes not related directly to the main catalytic cycle
[
45]. Probably, PVP induces Pd(0) aggregation better than the car-
boxylic resin, whereby the stabilized suspension of Pd NPs deacti-
vates faster than material (4).
4.6. STEM-HAADF characterization of the catalyst along the reusing
cycles
The evolution of the catalyst after consecutive cycles has been
also analyzed by electron microscopy and XPS. Hence, STEM-
HAADF images of the catalyst between the first and the fifth cycle
have been gathered in Fig. 5. No significant differences appear dur-
ing the first four cycles. The polymer surface seems to be decorated
with small white spots associated to the formation of PdNPs.
Together with small pseudospherical particles with sizes in the
5
–7 nm range, larger additional elongated, irregular and curved
spots also appears from the first catalytic cycle. The irregular and
variable shape of these spots is suggestive of nanoparticle aggrega-
tion more than particle growing. In any case, we observe a certain
tendency related to the size of the larger aggregates detected as the
catalytic cycles increases: these clusters evolve from 25–35, to 50–
6
3
5 and up to more than 100 nm for the catalyst after cycles 1 and 2,
and 4, and finally after 5 cycles, respectively. In a parallel way,
the density of small Pd NPs slightly decreases after the fourth cycle
and more abruptly after the fifth cycle. In order to illustrate this
point, the Pd NPs size distribution for each cycle is shown in
Fig. S14 (supplementary information). In addition to the Pd(0)
NPs detected through STEM-HAADF, a certain proportion of Pd
Fig. 4. SADS plots for the SM reaction of (a) 4-iodoanisole/4-iodotoluene and (b) 4-
bromoanisole/4-bromotoluene with phenylboronic acid at 60° in ethanol catalyzed
by materials (4), (2), and Pd NPs stabilized with PVP.
(
probably as anchored complexes) remains trapped inside the
polymer. Its presence is evident in the mapping done in domains
where white spots do not appear. The polymer is not porous and
shows a low surface area of ca. 2 m /g. Then, cannot be discarded
similar slope for a SADS experiment performed with 4-iodoanisole
and 4-iodoacetophenone as substrates. They differ significantly
from the curved trajectory shown by complex (2), see Scheme 3.
2
that a low ratio of Pd organometallic complexes could be anchored

M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
35
Fig. 5. STEM-HAADF images of the catalyst after (a) one, (b) two, (c) three, (d) four
and (e) five catalytic cycles. The white spots correspond to Pd nanoparticles or
aggregates thereof. The persistence of Pd and P in free-Pd NPs areas (without white
spots) is shown in the mapping performed to the catalyst after 5 reaction cycles.
but also trapped inside the polystyrene resin without any possibil-
ity to be chemically reduced and consequently unable to form or
participate in the nucleation and growth of metal nanoparticles.
In this sense, the polymer must be viewed as a reservoir of Pd com-
plexes that during the first four or five catalytic cycles evolves
through the formation of active Pd NPs up to exhaust of the acces-
sible complexes or the aggregation of the nanoparticles. The forma-
tion of particle aggregates can be clearly appreciated in the HRTEM
images shown in Fig. 6. These micrographs are consistent with the
STEM-HAADF data. The observation of aggregates occurs even dur-
ing the firsts cycles leading to relatively small clusters with differ-
ent and irregular distributions in which the real ‘‘building blocks”
are pseudospherical nanoparticles with very small averaged sizes
comprises in the 3–4 nm range. As the number of catalytic cycles
increases, a significant growth of this aggregates occurs, but pre-
serving the size of the primary Pd NPs. This small crystallite size
is consistent with the absence of XRD signals regardless the num-
ber of cycles (and the aggregation degree).
Fig. 6. HRTEM images of the catalyst after (a) one, (b) three and (c) four catalytic
cycles.
certain proportion of Pd was reduced forming Pd NPs. A part of
these Pd NPs could be eliminated during the filtration step, and
the rest remains on the resin support (according to the STEM-
HAADF images). In any case, what is obvious is that a large propor-
tion of Pd organometallic complexes remain anchored to the resin
as seminal catalytic building blocks. From the second cycle, the Pd
3
d XPS signals attributed to Pd(0) can be clearly observed and its
relative intensity increases in a regular way with the catalytic
cycles. This evolution can almost be better understood as a pro-
gressive reduction of the Pd(II) complexes trapped within the resin.
The catalysts act as a reservoir of palladium that due to the non-
porous nature of the resin (in contrast with the Pd NPs-UVM-7 sil-
ica case) shows a certain inertia towards the attack to the complex
what it means to lose the anchorage, the reduction of palladium
and its subsequent nucleation and growth forming Pd NPs
The evolution of the Pd 3d XPS spectra (Fig. 7) is consistent with
the electron microscopy data and also with the kinetic study. After
the first catalytic cycle, the dominant signals in the XPS spectrum
fits very well with the peaks detected for the pre-catalyst. Then, in
the recovered solid after cycle 1, Pd(II) is the dominant oxidation
state, which must corresponds to the existence of Pd organometal-
lic complexes anchored to the resin. This does not exclude that a
(
or aggregates). The Pd NPs in solution and/or decorating the resin
surface must be the effective catalytic sites. Until the fourth cycle,
the STEM-HAADF images are quite similar: a good dispersion of Pd

3
6
M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
Pd. The catalysis was driven mostly by Pd(0) species in solution,
as assessed by the three phases test and the differential sensitivity
kinetic assay results. The activity of material (4) compared to that
of Pd NPs-UVM-7 resulted in slightly lower conversion rates for 4-
phenyliodides, but higher conversions for 4-phenylbromides even
operating at lower temperatures. Comparison with PVP stabilized
Pd NPs revealed a similar behavior as material (4) presented lower
initial rates but enhanced conversion yields. Hence, the ability of
the material (4) to supply continuously Pd(0) to the medium made
it a more suitable catalyst, specially when using weakly activated
phenyl halides as substrates.
The catalyst reusing experiments showed that activity of mate-
rial (4) remained at the same level for the first 4 cycles making a
difference with Pd NPs-UVM-7 [25] and PVP stabilized Pd NPs
(
Pd NPs were remarkably deactivated during the first cycle, see Fig-
ure S11), which lost notably their catalytic activity along the first
cycle. Palladium deposited on the support during the cooling/filter-
ing of the reaction mixture, resulting in filtrates without catalytic
activity. The structural study of the recovered catalyst was consis-
tent with the kinetic study. Thereby, SEM and XPS studies showed
a progressive reduction in the Pd(II)/Pd(0) ratio, and STEM-HAADF
showed that Pd NPs size increased through the cycles, both phe-
nomena being correlated with an activity fall. The structural anal-
ysis suggested that the larger less active Pd NPs were not formed
by crystal growth, but rather by aggregation of smaller Pd NPs,
probably in solution.
Acknowledgments
We thank the Spanish Ministerio de Economía y Competitividad
and European Feder Funds (MAT2015-64139-C4-2-R and RTI2018-
1
00910-B-C44) for funding this research.
Fig. 7. (a) Evolution of the Pd 3d XPS spectra with the number of progressive
catalytic cycles (PRE = precatalyst; RnC, n = 1, 2, 3, 4 and 5, n = number of catalytic
cycles). (b) Evolution of the Pd0/Pd(II) ratio with the number of catalytic cycles.
Appendix A. Supplementary material
Figure S1: FTIR spectra of (4); Figure S2: Catalytic activity of
phenyliodides; Figure S3: Influence of catalyst concentration in
reaction rate; Figure S4: Results of a reusing experiment using a
double load Pd catalyst; Figures S5/S6: SEM results; Figure S7:
hot filtration classical test results; Figure S8: Centrifugation classi-
cal test results; Figure S9: Poisoning test results; Figure S10: Three
phases test results; Figure S11: SADS kinetics for phenyliodide
derivatives. Figure S12: Deactivation of the catalyst by continous
addition of substrate; Figure S13: TEM micrographs of PVP stabi-
lized Pd NPs; Figure S14: NPd NPs distribution size for the diverse
reusing catalyst cycles. Supplementary data associated with this
article can be found, in the online version, at https://doi.org/10.
NPs or small aggregates is observed (although it is true that the
size of the latter and their abundance increase progressively with
the number of cycles). After the fifth cycle, Pd(0) is the dominant
species. However, the catalytic activity drops abruptly. It is also
after the fifth cycle when more important differences are seen in
STEM-HAADF images. As previously commented, the density of
Pd NPs on the surface resin decreases and large aggregates are
observed (these latter are probably not able to act as effective cat-
alytic centers). This evolution can be understood as an exhaustion
of the accessible Pd reserve. A certain amount of Pd complexes will
resist trapped within the resin according to XPS and the mapping
images.
1
016/j.jcat.2019.10.014.
References
5
. Conclusions
[
[
[
[
1] N. Miyaura, A. Suzuki, Palladium-catalyzed cross-coupling reactions of
organoboron compounds, Chem. Rev. 95 (1995) 2457–2483.
2] D.S. Surry, S.L. Buchwald, Dialkylbiaryl phosphines in Pd-catalyzed amidation:
a user’s guide, Chem. Sci. 2 (2011) 27–50.
3] R. Chinchilla, C. Nájera, Recent advances in Sonogashira reactions, Chem. Soc.
Rev. 40 (2011) 5084–5121.
2 6 4 2 2 2 6 4 2
The new material [Pd {l-(C H ) PPh } {l-O C(C H -PS)} ] (4)
containing the organometallic complex (2) covalently linked to a
carboxypolystyrene support was prepared through a synthetic
route that allows the Pd(II) cations to be dispersed homogeneously
on the polymer surface.
4] J. Corbet, G. Mignani, Selected patented cross-coupling reaction technologies,
Chem. Rev. 106 (2006) 2651–2710.
The activity was checked for the Suzuki-Miyaura reaction. The
conversion-time plots exhibited an induction period originated
by the reduction of Pd (II) to Pd (0), and their analysis indicated
that the activity went through a maximum along the reaction time.
This phenomenon was explained by the continuous delivery of Pd
atoms to the solution from the polymeric support and the subse-
quent deactivation by aggregation of the solubilized fraction of
[5] J. Magano, J. Dunetz, Large-scale applications on transitions metal-catlyzed
coupling for synthesis of pharmaceuticals, Chem. Rev. 111 (2011) 2177–2250.
[
6] C. Torbog, M. Beller, Recent applications of palladium-catalyzed coupling
reactions in the pharmaceutical, agrochemical, and fine chemistry industries,
Adv. Synth. Catal. 352 (2009) 3027–3043.
[
[
7] R. Chinchilla, C. Nájera, Chemicals from alkynes with palladium catalysts,
Chem. Rev. 114 (2014) 1783–1826.
8] L. Yin, J. Liebscher, Carbon-carbon coupling reactions catalyzed by
heterogeneous palladium catalysts, Chem. Rev. 107 (2007) 133–173.

M. Ángeles Úbeda et al. / Journal of Catalysis 381 (2020) 26–37
37
[
9] V. Polshettiwar, C. Len, A. Fihri, Silica-supported palladium: Sustainable
catalysts for cross-coupling reactions, Coord. Chem. Rev. 253 (2009) 2599–
[27] F. Estevan, S. Ibañez, P. Hirva, A. Ofori, M. Sanau, M. Úbeda, Benzoato and
thiobenzoate ligands in the synthesis of palladium(III) and (II) compounds:
stability and catalytic applications, Eur. J. Inorg. Chem. (2015) 2822–2832.
[28] C.M. Crudden, M. Sateesh, R. Lewis, Mercaptopropyl-modified mesoporous
silica: a remarkable support for the preparation of a reusable, heterogeneous
palladium catalyst for coupling reactions, J. Am. Chem. Soc. 127 (2005) 10045–
10050.
2
626.
[
10] M. Lamblin, L. Nassar-Hardy, J. Hierso, E. Fouquet, F. Felpin, Recyclable
heterogeneous palladium catalysts in pure water: Sustainable developments
in Suzuki, Heck, Sonogashira and Tsuji-Trost reactions, Adv. Synth. Catal. 352
(
2010) 33–79.
[
11] N.E. Leadbeater, M. Marco, Preparation of polymer-supported ligands and
metal complexes for use in catalysis, Chem. Rev. 102 (2002) 3217–3274.
12] B. Altava, M.I. Burguete, E. Garcí-Verdugo, S.V. Luis, Chiral catalysis
immobilized on achiral polymers: effect of the polymer support on the
performance of the catalyst, Chem. Soc. Rev. 47 (2018) 2722–2771.
13] E. Mohammadi, B. Movassagh, Synthesis of polystyrene-supported Pd(II)-NHC
complex derived from theophylline as an efficient and reusable heterogeneous
catalyst for the Heck-Masuda cross-coupling reaction, J. Mol. Catal. A: Chem.
[29] M.L. Kantam, M. Roy, S. Roy, B. Sreedhar, S.S. Madhavendra, B.M. Choudary, R.L.
De, Polyaniline supported palladium catalyzed Suzuki-Miyaura cross-coupling
of bromo- and chloroarenes in water, Tetrahedron 63 (2007) 8002–8009.
[30] H. Gruber-Woelfler, P.F. Radaschitz, P.W. Feenstra, W. Haas, J.G. Khinast,
Synthesis, catalytic activity, and leaching studies of a heterogeneous Pd-
catalyst including an immobilized bis(oxazoline) ligand, J. Catal. 286 (2012)
30–40.
[31] A.F. Schmidt, A.A. Kurokhtina, Distinguishing between the homogeneous and
heterogeneous mechanisms of catalysis in the Mizoroki-Heck and Suzuki-
Miyaura reactions: Problems and prospects, Kinet. Catal. 53 (2012) 714–730.
[32] A.F. Schmidt, A.A. Kurokhtina, E.V. Larina, Simple kinetic method for
distinguishing between homogeneous and heterogeneous mechanisms of
catalysis, illustrated by the example of ‘‘ligand-free” Suzuki and Heck reactions
of aryl iodides and aryl bromides, Kinet. Catal. 53 (2012) 86–93.
[33] J. Bezanson, S. Karpinski, V.B. Shah, A. Edelman, Julia: A fast dynamic language
for technical computing, CoRR abs/1209.5 (2012).
[34] J.A. Nelder, R. Mead, A Simplex method for function minimization, Comput. J. 7
(1965) 308–313.
[35] S.G. Johnson, The NLOPT nonlinear-optimization package, 2012, http://ab-
initio.mit.edu/nlopt. http://ab-initio.mit.edu/nlopt.
[36] C. Tsitouras, T. Simos, Explicit high order methods for the numerical
integration of periodic initial-value problems, Appl. Math. Comput. 95
(1998) 15–26.
[37] C. Rackauckas, Q. Nie, Differentialequations.jl - a performant and feature-rich
ecosystem for solving differential equations in Julia, J. Open Res. Softw. 5
(2017) 10.
[
[
[
[
[
[
4
18 (2016) 158–167.
14] M. Bakherad, B. Bahramian, S. Jajarmi, A novel 1,2,4-triazine-functionalized
polystyrene resin-supported Pd(II) complex: a copper- and solvent-free highly
efficient catalyst for Sonogashira coupling reactions, J. Organomet. Chem. 749
(
2014) 405–409.
15] V.N. Mikhaylov, V.N. Sorokoumov, K.A. Korvinson, A.S. Novikov, I.A. Balova,
Synthesis and simple immobilization of palladium(II) acyclic diaminocarbene
complexes on polystyrene support as efficient catalysts for Sonogashira and
Suzuki-Miyaura cross-coupling, Organometallics 35 (2016) 1684–1697.
16] V.N. Mikhaylov, V.N. Sorokoumov, D.M. Liakov, A.G. Tskhovrebov, I.A. Balova,
Polystyrene-supported acyclic diaminocarbene palladium complexes in
Sonogashira cross-coupling: Stability vs. catalytic activity, Catalalysts
2018) 141–159.
8
(
17] M.V. Polynski, V.A. Ananikov, Modeling key pathways proposed for the
formation and evolution of ‘‘coctail”-type systems in Pd-catalyzed reactions
involving ARX reagents, ACS Catal. 9 (2019) 3991–4005.
18] S. Hüber, J.G. de Vries, V. Farina, Why does industry not use immobilized
transition metal complexes as catalyts?, Adv Synth. Catal. 358 (2016) 3–25.
19] L. Falivene, S.M. Kozlov, L. Cavallo, Constructing bridges between
computational tools in heterogeneous and homogeneous catalysis, ACS Catal.
[
[
[38] Y. Nishihara (Ed.), Applied Cross-Coupling Reactions, Lecture Notes in
Chemistry, vol. 80, Springer-Verlag, Berlin, Heidelberg, 2013.
8
(2018) 5637–5656.
[39] T.E. Barder, S.D. Walker, J.R. Martinelli, S.L. Buchwald, Catalysts for Suzuki-
Miyaura coupling processes: scope and studies of the effect of ligand structure,
J. Am. Chem. Soc. 127 (2005) 4685–4696.
[
20] L. Liu, A. Corma, Metal catalysts for heterogeneous catalysis: from single atoms
to nanoclusters and nanoparticles, Chem. Rev. 118 (2018) 4981–5079.
21] A. Biffis, P. Centomo, A. Del Zotto, M. Zecca, Pd metal catalysts for cross-
couplings and related reactions in the 21st century: a critical review, Chem.
Rev. 118 (2018) 2249–2295.
22] A.A. Kurokhtina, E.V. Larina, A.F. Shmidt, Study of the differential selectivity of
cross-coupling reactions for elucidating the nature of the true catalyst, Kinet.
Catal. 56 (2015) 191–197.
[
[40] D.A. Conlon, B. Pipik, S. Ferdinand, C.R. LeBlond, J.R. Sowa, B. Izzo, P. Collins, G.
Ho, J.M. Williams, Y. Shi, Y. Sun, Suzuki-Miyaura cross-coupling with quasi-
heterogeneous palladium, Adv. Synth. Catal. 345 (2003) 931–935.
[41] R. Narayanan, C. Tabor, M. El-Sayed, Can the observed changes in the size or
shape of a colloidal nanocatalyst reveal the nanocatalysis mechanism type:
Homogeneous or heterogeneous?, Top Catal. 48 (2008) 60–74.
[42] J.D. Webb, S. MacQuarrie, K. McEleney, C.M. Crudden, Mesoporous silica-
supported Pd catalysts: an investigation into structure, activity, leaching and
heterogeneity, J. Catal. 252 (2007) 97–109.
[43] S. Reimann, J. Stötzel, R. Frahm, W. Kleist, J.D. Grunwaldt, A. Baiker,
Identification of the active species generated from supported Pd catalysts in
heck reactions: an in situ quick scanning EXAFS investigation, J. Am. Chem.
Soc. 133 (2011) 3030–3921.
[
[
23] D.B. Eremin, V.P. Ananikov, Understanding active species in catalytic
transformations: from molecular catalysis to nanoparticles, leaching,
‘
2
‘cocktails” of catalysts and dynamic systems, Coord. Chem. Rev. 346 (2017)
–19.
[
24] J. El Haskouri, J.M. Morales, D. de Zárate, L. Fernández, J. Latorre, C. Guillem, A.
Beltrán, D. Beltrán, P. Amorós, Nanoparticulated silicas with bimodal porosity:
chemical control of the pore sizes, Inorg. Chem. 47 (2008) 8267–8277.
25] P. Alfonso Albiñana, J. El Haskouri, M.D. Marcos, F. Estevan, P. Amorós, M.
Úbeda, F. Pérez-Pla, A new efficient, highly dispersed, Pd nanoparticulate silica
supported catalyst synthesized from an organometallic precursor. Study of the
homogeneous vs. heterogeneous activity in the Suzuki-Miyaura reaction, J.
Catal. 367 (2018) 283–295.
[
[44] J.M. Richardson, C.W. Jones, Poly(4-vinylpyridine) and Quadrapure TU as
selective poisons for soluble catalytic species in palladium-catalyzed coupling
reactions - application to leaching from polymer-entrapped palladium, Adv.
Synth. Catal. 348 (2006) 1207–1216.
[45] A.F. Schmidt, A.A. Halaiqaa, V. Smirnov, Interplays between reactions within
[
26] F. Estevan, A. García-Bernabé, P. Lahuerta, M. Sanaú, M.A. Ubeda, M.C. de
Arellano, Synthesis, reactivity, and X-ray crystallographic characterization of
mono-, di-, and tetranuclear palladium(II)-metalated species, Inorg. Chem. 39
and without the catalytic cycle of the heck reaction as
optimization of the synthetic protocol, Synlett 18 (2006) 2861–2878.
a clue to the
(
2000) 5964–5969.