New palladium catalysts on polyketone prepared through different smart methodologies and their use in the hydrogenation of cinnamaldehyde

Article abstract of DOI:10.1016/j.apcata.2012.09.003

For the first time, polyketone (poly-3-oxotrimethylene, PK) has been employed as support for Pd nanoparticles. Three different approaches have been adopted: (i) reduction of Pd(OAc)₂ in EtOH under microwave irradiation, (ii) hydrothermal reduct

Full text of DOI:10.1016/j.apcata.2012.09.003

Applied Catalysis A: General 447–448 (2012) 49–59
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
New palladium catalysts on polyketone prepared through different smart
methodologies and their use in the hydrogenation of cinnamaldehyde
Anna Maria Raspolli Gallettia,∗, Luigi Toniolob, Claudia Antonettia, Claudio Evangelistic, Claudia Forted
a Department of Chemistry and Industrial Chemistry, University of Pisa, Via Risorgimento 36, 56126 Pisa, Italy
b Department of Molecular Sciences and Nanosystems, University Ca’ Foscari of Venice, Dorsoduro 2137, 30123 Venice, Italy
c ICCOM CNR, Via G. Moruzzi 1, 56124 Pisa, Italy
d CNR ISTM, Via Fantoli 16, 20138 Milano, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 June 2012
Received in revised form 31 August 2012
Accepted 1 September 2012
Available online 13 September 2012
For the first time, polyketone (poly-3-oxotrimethylene, PK) has been employed as support for Pd nanopar-
ticles. Three different approaches have been adopted: (i) reduction of Pd(OAc)2 in EtOH under microwave
irradiation, (ii) hydrothermal reduction of the same precursor in EtOH or n-BuOH employing traditional
heating and (iii) vaporization of Pd metal through metal vapour synthesis. These supported systems have
been characterized by solid state NMR, TEM, ICP, XPS, BET and tested in the selective hydrogenation of
cinnamaldehyde (CAL) to hydrocinnamaldehyde (HCAL) as probe reaction. Their catalytic performances
have been compared in relation to their preparation procedure and morphology. It is remarkable that
these catalysts do not show significant leaching even after several catalytic recycles, thus indicating their
high stability and recyclability, ascribed to a stabilizing capacity of this particular polymeric support.
© 2012 Elsevier B.V. All rights reserved.
Keywords:
Pd nanoparticles
Polyketone
Cinnamaldehyde
Hydrocinnamaldehyde
1. Introduction
[PdCl2(PhCN)2] complex on copolymers of styrene/divinylbenzene
with N-vinylpyrrolidone or 1-vinylimidazole monomers via ligand
The synthesis of metal nanoparticles for catalytic applications,
characterized by high catalytic activity and selectivity, is an attrac-
tive research area of great current interest, not only restricted to
the catalysis field, but also for applications in electronics, biotech-
nology, medical science and magnetism [1,2]. A broad range of
approaches is suitable for their preparation, from top-down to
bottom-up ones and combinations thereof [3–9]. It is important to
highlight that during their preparation the use of stabilizing agents
represents a useful and widely applied approach in order to prevent
their agglomeration toward the bulk metal [10–14]. In this context,
the use of polymeric stabilizers, such as poly-(1-vinylpyrrolidone),
polypyrrole, polyvinylpyridine, poly(ethyleneglycol), polystyrene,
nafion and also natural polymers has been deeply investigated
[15–21]. These “quasi” homogeneous metal nanoparticles offer
the advantage of high catalytic activity as the homogeneous ones
but, on the other side, they are not easily separated and recycled
as the heterogeneous ones. In order to improve this last draw-
back, they are generally supported on a solid inorganic carrier. On
the contrary, the applications of polymers as support for metal
nanoparticles are less investigated [22–25]. In particular, Pd cat-
alysts were supported and thus “heterogenized” by anchoring the
exchange reaction. The obtained systems were reduced during
hydrogenation tests or by reaction with sodium borohydride and
employed in the hydrogenation of cinnamaldehyde to cinnamyl
alcohol [22]. In the work of Zhan, hydrophilic interpenetrating poly-
mer networks and Pd nanoparticles were obtained by simultaneous
crosslinking of polyvinyl alcohol and polyacrylamide, leading to
very promising samples for Heck coupling reactions [23]. The syn-
thesis of supported Pd complexes by the easy and quantitative
coordination of Pd(OAc)2 to a cross-linked styrene/divinylbenzene
resin functionalized with bis(diphenylphosphino)methane ligands
has been reported and these systems were tested in the hydro-
genation of nitrocompounds and ␣,␤-unsaturated aldehydes under
mild reaction conditions resulting active, versatile and recyclable
heterogenized catalysts [24].
The use of polyketones as supports for metal nanoparticles
(MNPs) has been almost neglected. A recent investigation on
polyketones is reported by Gomez and co-workers. The use of
isotactic and chiral CO/styrene copolymers as stabilizers for Pd,
Pt and Ru nanoparticles was studied with the aim of apply-
ing those systems in asymmetric catalytic processes. Platinum
and ruthenium nanoparticles stabilized in this way were tested
in the hydrogenation of ethyl pyruvate giving high activities
without enantioselectivity induction. In this study the authors
highlight the main role of stabilizing agent played by this
polymer [25].
∗
Corresponding author. Tel.: +39 050 2219290; fax: +39 050 2219260.
E-mail address: roxy@dcci.unipi.it (A.M. Raspolli Galletti).
0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2012.09.003

50
A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
On the basis of these results, we decided to test as a new sup-
particle sizes of the metals, selective poisonings, presences of other
metals, preparation methods, catalyst precursors, temperatures of
reduction, pressures, etc., have been proposed to influence the
intramolecular hydrogenation selectivity [35].
Now, for the first time, to the best of our knowledge, we report
PK as a support for Pd nanoparticles in order to test its potential
double role of support and stabilizing agent for metal nanoparti-
cles. The new systems were tested in the selective hydrogenation
of cinnamaldehyde to hydrocinnamaldehyde.
port for Pd nanoparticles the polyketone (poly-3-oxotrimethylene,
PK) obtained by the perfectly alternating CO–ethene copolymeriza-
tion catalyzed by Pd(II)-diphosphine complexes [26]. These PKs are
characterized by a useful combination of mechanical, high temper-
ature and chemical resistance [27], which prompted us to take them
into consideration as suitable candidates for support and stabilizer
of Pd nanoparticles.
Three methodologies were adopted: microwave synthesis, ther-
mal reduction in the autoclave and metal vapour synthesis.
Microwave (MW)-assisted synthesis of metal nanoparticles rep-
resents an innovative sustainable and green approach because
of environmental friendly and less energy-demanding condi-
tions with respect to conventional methods. The application of
microwave heating is a fast growing research area, where high reac-
tion rates and high selectivities can be achieved together with a
significant reduction of the reaction time, often by orders of mag-
nitude [28–31]. The hydrothermal reduction of metal precursors in
the autoclave is another valid procedure for the synthesis of metal
nanoparticles, previously deeply investigated by us for ruthenium
catalysts [32]. The third approach namely metal vapor synthesis
(MVS) is another versatile procedure that ensures the deposition
of very reactive mono- and bi-metallic metal nanoparticles that
require no activation steps before use [33,34]. This method is based
on the co-condensation of metal vapors with vapors of weakly
stabilizing organic solvents (acetone, tetrahydrofuran, toluene and
n-pentane) on the cold walls (−196 ◦C) of glass or steel commer-
cially available reactors, under high vacuum. The reaction affords
metal nanoparticles stabilized by the excess of the organic solvent
and stable at low temperature (T < −30 ◦C), i.e. the solvated metal
atoms (SMAs), which are excellent starting aggregates for the gen-
tle deposition of ligand-free metal nanoparticles on a wide range of
organic and inorganic supports. This last method ensures the depo-
sition of very reactive mono- and bi-metallic metal nanoparticles
that require no activation steps before use.
2. Experimental
2.1. Chemicals
Palladium powder and palladium acetate (Pd(OAc)2) were
purchased from Aldrich and used as received; 1,3-bis[bis(2-
methoxyphenyl)phospino]propane (dapp) was provided by
Hyosung Corporation. PK was prepared by CO–ethene copolymer-
ization as reported in [36], but using (dapp) and H2SO4 in place
of 1,3-bis(diphenylphospino)propane and TsOH, respectively,
and in the presence of benzoquinone (Pd = 2.5 × 10−2 mmol l−1
,
Pd/dapp/H2SO4/BQ = 1/1.2/60/60) in 80 ml MeOH/H2O (10/1, v/v).
At 80 ◦C, under 11 MPa total pressure (CO/ethene = 1.4/1), after 2 h
reaction, 10.1 g of PK were obtained, showing an average molecular
weight higher than 50 kDa according to solid state NMR spectra,
a surface area (BET) of 16 m2 g−1 and a melting point of 257 ◦C
determined by DSC.
All other reagents and standard products (MeOH, EtOH, n-
BuOH, acetone, mesitylene, 1-hexene, n-pentane, cinnamaldehyde,
hydrocinnamaldehyde, cinnamyl alcohol and diethyl ether) were
obtained from Aldrich, purified according to standard procedures
and stored under argon.
2.2. Preparation of the Pd nanoparticles supported on PK
The new Pd/PK systems were tested in the selective hydrogena-
tion of cinnamaldehyde (CAL) to hydrocinnamaldehyde (HCAL) (see
below Scheme 1), which is a conventional test reaction, already
investigated by us and many other authors [28,35–43] because of its
industrial, biological and pharmacological interest. Recently, HCAL
has been also found an important intermediate for the preparation
of pharmaceuticals used in the treatment of HIV [44].
Although many efforts have already been devoted to improve
the selectivity in the heterogeneous hydrogenation of CAL to HCAL,
high selectivity and reusability and negligible metal leaching still
remain targets to be accomplished. These are highly demand-
ing tasks due to the complexity of the system involving a rather
complicated reaction mechanism, including competitive/non-
competitive and dissociative/non-dissociative adsorption, side
reaction, adsorption of solvent, etc. Support materials and
Table 1 reports all reaction conditions adopted for the three dif-
ferent types of preparations, carried out using the same PK support.
The MW-assisted synthesis of Samples 1, 1 bis and 2 was per-
formed using a CEM Discover S-class oven with a PC control. The
instrument consists of a single-mode, self tuning cavity where the
correct position of the reaction vial is automatically tuned to ensure
reproducible conditions at each run. The oven operates with a
continuous power generator capable of supplying an irradiation
power from 0 to 300 W (with 1 W-step increment) with a fine and
automated control of the sample temperature based on the CEM
PowerMax technology. The oven allows two different operating
ways: at constant temperature (by varying the irradiation power
in order to maintain the desired temperature value) or at con-
stant power. The 80 ml glass reactor (Sealed Vessel accessory) was
equipped with a magnetic drive stirrer and a vertically focused IR
temperature sensor, which provided an accurate temperature con-
trol during the process (maximum operating temperature 250 ◦C).
This Sealed Vessel accessory was equipped with a pressure line
and a pressure probe, the latter allowing the control of the inter-
nal pressure throughout each run (maximum operating pressure
1.4 MPa). The selected amounts of Pd(OAc)2 and PK were dissolved
in EtOH, under stirring at room temperature. The resulting mixture
was MW-irradiated for 5 min at 100 ◦C under nitrogen atmosphere.
When the reaction was complete, the reactor was rapidly cooled to
room temperature, the solvent was removed by filtration under
nitrogen and the supported catalysts were washed several times
with acetone and ether, dried under vacuum and stored under dry
nitrogen atmosphere.
According to the thermal reduction procedure, for the prepa-
ration of Samples 3, 4, 5 and 5 bis, the precursor was reduced in
Scheme 1. The main reaction pathway of the hydrogenation of cinnamadehyde.

A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
51
a 300 ml mechanically stirred Parr 4560 autoclave equipped with
a P.I.D. controller 4843. The chosen amount of PK was put in the
autoclave before introducing the alcoholic solution of the metal
precursor, prepared by dissolving the desired quantity of Pd(OAc)2
in the alcoholic medium (EtOH or n-BuOH) at room tempera-
ture under magnetic stirring in nitrogen atmosphere in a 200 ml
Schlenk. The resulting mixture was introduced by suction in the
autoclave previously evacuated to 0.5 mm Hg. The autoclave was
pressurized with nitrogen to 3 MPa and heated up to 100 ◦C for 3 h,
then it was rapidly cooled and the gas discharged. Finally the mix-
ture was worked-up and, after solvent removal, the supported Pd
catalyst was washed with acetone and ether, dried under vacuum
and stored under nitrogen atmosphere.
All operations involving the MVS products (Samples 6 and 7)
were performed under a dry argon atmosphere. Pd vapour, gen-
erated by resistive heating of alumina-coated tungsten crucibles
filled with ca. 120 mg of Pd powder, was co-condensed at liq-
uid nitrogen temperature with a mixture of 1-hexene (30 ml)
and mesitylene (30 ml) in a previously described glass reactor
[33] for ca. 1 h. The reactor chamber was warmed to the melt-
ing point of the solid matrix, and the resulting brown solution
was siphoned at a low temperature into a Schlenk tube and kept
in a refrigerator at −40 ◦C. The Pd content, obtained by ICP-OES
analysis, resulted 1.6 mg (Pd)/ml. 40 ml of Pd SMA (0.60 mmol,
64 mg of Pd) were added to a suspension of PK (2.00 g) in mesity-
lene (30 ml). The resulting mixture was stirred for 6 h at room
temperature. Pd nanoparticles were quantitatively deposited on
the support, the colorless solution was removed and the brown
solid was washed with n-pentane and dried under reduced
pressure obtaining a Pd/PK system with 3.2 wt% Pd (Sample 6)
as confirmed by ICP-OES analysis. An analogous approach was
adopted for the synthesis of Sample 7, characterized by lower
loading.
2.3. Characterization of the catalysts
TEM images of all samples were taken using a JEOL 3010-UHR
transmission electron microscope with an acceleration potential of
300 kV. Histograms of the particles size distribution were obtained
by counting at least 500 particles onto the micrographs. The mean
diameter (dm) was calculated as dm = ꢀdini/ꢀni where ni was the
number of particles of diameter di. The counting was carried out on
electron micrographs obtained starting from 50,000 up to 100,000
magnifications, where Pd particles well contrasted with respect to
the support were clearly detected. The graduation of the particle
size scale was 0.5 nm.
Solid state NMR spectra were taken on a Bruker AMX300WB
spectrometer working at 75.47 MHz for carbon-13 and equipped
with a 4 mm CP/MAS probe head. All experiments were performed
at 23 ◦C and spinning at a rate of 8 kHz. 13C direct excitation magic
angle spinning (DEMAS) NMR spectra were acquired with a pulse
length of 2 ␮s, corresponding to a 45◦ flip angle, using a short (1 s)
and a long (60 s) recycle delay acquiring 20,000 and 2000 scans,
respectively.
The 13C CPMAS NMR spectrum was acquired with a 1H 90◦ pulse
length of 3.2 ␮s using a contact time (tCP) of 2.2 ms, a spin-lock field
of 78 kHz and a recycle delay of 10 s; 12,000 scans were accumu-
lated.
Variable contact time experiments were performed with tCP val-
ues ranging from 100 ␮s to 50 ms, acquiring 400 scans for each
experiment. The peak intensity vs. tCP curves were fitted using the
equation [45,46]:
/T
I(0) e−t
TCH
− e−t
/T
CP 1␳H
CP CH
I(tCP) =
(1/TCH) − (1/T1␳H
)

52
A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
HP 3396 integrator, a flame ionization detector and a PONA cap-
illary column (50 m × 0.2 mm × 0.5 ␮m) with a stationary phase
100% dimethylpolysiloxane (carrier gas nitrogen, flow 1 ml/min).
Activity was calculated as converted moles of CAL (gPd h)−1 mea-
sured after 30 min of reaction and at 100 mol% CAL conversion. The
reproducibility of repeated catalytic runs was within 5%. Recycling
experiments were carried out in a similar manner: after removing
the liquid reaction mixture through the sample valve and evacu-
ating the autoclave and then by charging again with the fresh CAL
solution for the subsequent catalytic cycle.
Leaching tests were carried out as follows: at the end of the first
hydrogenation runs the solid catalyst was filtered off under inert
atmosphere and the recovered solution was reused for a second
hydrogenation. In addition, in order to verify the recyclability of the
catalysts and to confirm the leaching tests, the recovered catalyst
was also employed in other successive catalytic cycles, up to three
times.
where T1␳H is the proton spin lattice relaxation time in the rotating
frame, TCH is the 1H–13C cross-polarization time, and I(0) repre-
sents the theoretical maximum magnetization which would be
obtained with an infinitely short TCH and an infinitely long T1␳H. For
an individual peak, I(0) is directly proportional to the corresponding
number of nuclei.
13C spin-lattice relaxation times in the laboratory frame (13C T1)
were measured using the Torchia pulse sequence [47] with
tCP = 2.2 ms and variable delay values ranging from 10 ms to 60 s;
the intensity trends obtained were fitted with a biexponential
decay.
The BET surface area was determined by nitrogen adsorption,
using a single point ThermoQuest Surface Area Analizer Qsurf S1.
The content of Pd was determined by inductively coupled
plasma-optical emission spectrometers (ICP-OES) with a Spectro-
Genesis instrument equipped with a software Smart Analyzer
Vision. For ICP-OES, each sample was heated over a heating plate
in a porcelain crucible in the presence of aqua regia (2 ml) for four
times, dissolving the solid residue in 0.5 M aqueous HCl. The limit
of detection (lod) calculated for Pd was 2 ppb.
3. Results and discussion
The X-ray photoelectron spectroscopy analyses were performed
with a VG Microtech ESCA 3000 Multilab spectrometer, using the
unmonochromatised Al K␣ source (1486.6 eV) run at 14 kV and
15 mA. Survey spectra were collected at constant pass energy of
50 eV. Individual peak energy regions were collected at constant
pass energy of 20 eV. Samples were mounted on a stub holder
using double-sided adhesive tape. In order to avoid the contact with
air, the sample mounting was performed inside a glove bag filled
with N2 and then the holder was loaded inside the XPS machine
under a N2 atmosphere. The C1s peak, set at 285.1 eV, arising from
adventitious carbon, was used as reference for the binding energy
values. Differential surface charging was ruled out by checking the
reproducibility of XPS results in repeated scans under different
X-ray exposures. Analyses of the peaks were performed with the
software provided by VG, based on non-linear least-square fitting
routine using a weighted sum of Lorentzian and Gaussian compo-
nent curves after background subtraction according to Sherwood
[48].
DSC melting curves of the PK were recovered with a Perkin-
Elmer Pyris Diamond DSC instrument. The sample was heated up
to 275 ◦C at a scanning speed of 10 ◦C/min and held at that temper-
ature for 3 min to remove the effects of the thermal history. Then,
the sample was cooled at the same rate. After crystallization, the
sample was reheated and melted again.
Infrared spectra were recorded at 298 K on a Fourier trans-
form infrared (FT-IR) Perkin-Elmer Spectrum One spectrometer
equipped with a ATR accessory.
3.1. Catalyst characterization
The 13C-CPMAS NMR spectrum of the PK is reported in
Fig. 1A.
The two narrow resonances at 35.3 and 209.8 ppm are assigned
to the backbone methylene and carbonyl carbons, respectively.
Upon closer inspection, a broader shoulder is observed for both
peaks, at higher and lower chemical shift, respectively, for the
methylene and carbonyl resonances.
The 13C-DEMAS NMR spectrum of the PK acquired with a long
recycle delay (60 s) is shown in Fig. 1B.
The spectrum shows the same two resonances observed in the
CPMAS spectrum overlapping two broader resonances at 36.8 and
208.7 ppm, as highlighted by deconvolution of the DEMAS spec-
trum performed using the SPORT-NMR software [49], shown in the
same figure. The relative intensities of the carbonyl and methy-
lene signals are in the ratio 1:2, as expected. Since in CP NMR
spectra 13C signals of rigid carbons are enhanced, whereas DE
spectra recorded with sufficiently long relaxation delay show all
carbons present in the sample, the narrow peaks, which are rela-
tively much more intense in the CP spectrum, can be ascribed to
crystalline PK chains; on the other hand, the broader peaks, only
partially detected in the CP spectrum, are ascribable to amorphous
chains. The spectral deconvolution, as well as the comparison of
the intensities of the DEMAS spectra acquired with short and long
recycle delay, indicates that 30 5% of the PK is in a crystalline
form.
The analysis of the variable contact time CP spectra, performed
using the equation above reported, allowed T1␳H and TCH values
to be determined; a T1␳H value of 26 3 ms was found for both
the carbonyl and methylene peaks, as expected since the only
protons present are those from the methylene groups. TCH val-
ues of 1.2 0.1 ms and 80 20 ␮s were found for carbonyl and
methylene carbons, respectively, the latter displaying a sensibly
shorter value being directly bonded to two protons. The low TCH
values confirm the high rigidity of the carbons detected in the CP
spectrum.
The trend of the intensities of the two peaks in the Torchia exper-
iments could be fitted only assuming a biexponential decay. Both
carbonyl and methylene peaks resulted to have a short relaxing
component, accounting for 30 2% of the total intensity, character-
ized by a 13C T1 of 900 100 ms, ascribable to less rigid domains,
and a long relaxing one, characterized by a 13C T1 value of 75 5 s
in the case of the methylene carbons, and over 100 s for the car-
bonyl carbons, for which a precise value could not be determined
2.4. Hydrogenation of cinnamaldehyde to hydrocinnamaldehyde
Hydrogenation reactions were carried out in a stainless steel
300 ml mechanically stirred Parr 4560 autoclave equipped with a
P.I.D. controller 4843. In a typical procedure the proper amount of
the selected catalyst always corresponding, for every run, to 2.6 mg
of Pd(0), was introduced in the reactor under inert atmosphere. The
autoclave was then closed, evacuated up to 65 Pa and a solution
of 3.8 ml of cinnamaldehyde in 100 ml of decalin was introduced
by suction. The reactor was then pressurized with hydrogen to
2 MPa and heated up to 100 ◦C, maintaining a stirring speed of
500 rpm, a value which was ascertained to assure the absence of
mass transfer limitations. The pressure value was manually held
constant at 2 MPa by repeated hydrogen feeds. The course of the
reaction was monitored by periodically sampling the liquid through
a valve and analyzing it by gas chromatography. GC analyses were
performed with a HP 5890 gas chromatograph equipped with a

A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
53
Fig. 1. Part A: 13C-CPMAS NMR spectrum of PK. Asterisks and filled circles indicate spinning sidebands of the carbonyl and methylene peak, respectively. Part B: 13C-DEMAS
NMR spectrum of PK acquired with a 45◦ pulse and a 60 s recycle delay. Expansions show the two different contributions obtained for each signal from spectral deconvolution.
considering that a maximum delay value of 60 s was used in the
experiments.
The higher boiling point n-BuOH was then employed for the
preparation of Samples 4 and 5, at higher (4.0 wt%) and lower
(0.6 wt%) Pd loading respectively (Samples 4 and 5, Table 1). TEM
images with their corresponding particle size distributions are
shown in Fig. 4.
Signals ascribable to the expected terminal groups (CH3CH2CO–
and CH3OCO–) [26] were not detected; on the basis of the sig-
nal to noise ratio, it can be safely asserted that the average
molecular weight is larger than 50 kDa. Samples 1, 1 bis and
2 were prepared adopting the MW technique (Table 1). TEM
measurements show nanoparticles with an average diameter of
11.3 nm for Sample 1 (Pd loading 0.9%) and 4.9 nm for Sample
2 (Pd loading 0.6%); standard deviation 4.4 and 1.8, respectively
(Fig. 2).
Both samples show a homogeneous dispersion of Pd particles
on PK, Sample 2 being characterized by a particle size distribu-
tion relatively quite narrower, probably related to the lower Pd
loading. The surface areas of both these catalysts resulted about
10 m2 g−1 (Table 1). These values are very small if compared with
the more traditional inorganic supports, such as active carbon or
alumina and this is surely a crucial point to be considered when
their catalytic application is evaluated. These samples were also
characterized by XPS analysis; in Table 1 the binding energies rel-
ative to Pd 3d5/2 photoelectron transition characterized by two
components, respectively at 335.6 and 337.6 eV, are reported. The
low binding energy component at 335.6 eV is attributed to metal-
lic Pd(0), whereas the high energy component at 337.6 eV is due to
oxidized Pd2+ [50]. Both samples (Samples 1 and 2) show an almost
complete Pd reduction, as evidenced by the 98% and 99%, derived
from the peak at 335.6 eV.
Sample 1 was synthesized in a duplicate preparation (Sample
1 bis, Table 1) giving highly reproducible results, as confirmed
by TEM, XPS and BET analysis of the new sample, as reported
in Table 1. Thus, it is possible to underline that MW synthe-
sis offers a simple, convenient, rapid, reproducible preparation
route to Pd catalysts starting from a commercially available Pd(II)
precursor.
The second type of Pd catalysts preparation followed the
hydrothermal reduction under nitrogen overpressure [32]. This
approach was adopted for the synthesis of Samples 3, 4, 5 and 5 bis
working at 100 ◦C for 3 h in the presence of low-boiling point alco-
hols (EtOH or n-BuOH) acting as solvents/reducing agents. EtOH
was adopted for the preparation of Sample 3. Pd nanoparticles char-
acterized by an average diameter of 5.5 nm and a standard deviation
of 1.8 was obtained (Fig. 3).
In comparison with the samples prepared via the fast MW irra-
diation technique, these last catalysts reveal a more homogeneous
particle size distribution. From the comparison of Sample 3 with
Sample 4, an effect of the nature of the solvent can be deduced:
EtOH seems to favor the formation of Pd particles of smaller mean
diameter than those obtained by n-butanol. On the other hand, the
comparison among the three Samples 3–5 pointed out that similar
morphology of Pd particles is obtained, although the particle size
distribution of Sample 5, with lower metal loading, is narrower
than those of the other two samples, ranging from about 1 nm
upto 10 nm. This result evidences that for this preparation tech-
nique the metal loading clearly influences the Pd dispersion. Also
for these samples, the almost total reduction of Pd was ascertained
by XPS spectroscopy (Table 1). When the preparation of Sample
5 was repeated, the resulting Sample 5 bis evidenced very similar
features as reported as Table 1.
Finally, Samples 6 and 7, containing a Pd loading of 3.2 wt% and
1.0 wt% respectively, were prepared by depositing Pd nanoparti-
cles, obtained by MVS technique [33,34], on PK (Scheme 2). Using
the MVS apparatus, Pd atoms, produced by evaporation of the metal
under high vacuum, are co-condensed at liquid nitrogen tempera-
ture (−196 ◦C) with a 1:1 mixture of 1-hexene and mesitylene on
the cooled wall of the reactor. On warming until the solid matrix
melts (−30 ◦C/−40 ◦C), a brown solution, named called solvated Pd
atoms solution, was obtained. On warming the solution to room
temperature in the presence of PK, a quantitative deposition of Pd
nanoparticles on the polymeric support occurred.
TEM analysis showed in both cases a metal particle size distribu-
tion narrower than those of the samples prepared by the other two
methods, having an average diameters of 4.9 nm and 4.6 nm and a
standard deviation of 1.5 and 1.4 for Samples 6 and 7, respectively
(Fig. 5).
It is important to note that, differently from the above catalysts,
in these samples prepared by MVS procedure, the Pd loading did
not significantly affect the particle size distribution. It is remarkable
that these samples resulted completely reduced at XPS analysis, as
evidenced in Table 1.

54
A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
Fig. 2. TEM and particle size distribution of Samples 1 and 2.
Fig. 3. TEM and particle size distribution of Sample 3.

A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
55
Fig. 4. TEM and particle size distribution of Samples 4 and 5.
3.2. Catalytic tests
in decalin in the absence of CAL. Again, no consumption of hydro-
gen was observed. Moreover, the recovered PK resulted completely
identical to the fresh one (FT-IR and solid state NMR analysis). The
inertness of the PK made us confident of using it as a support.
3.2.1. Preliminary experiments to test the inertness of PK
As reported in Section 2, the PK was prepared by CO–ethene
copolymerization catalyzed by a Pd(II) complex, so that it might
be contaminated by trace amounts of Pd. Therefore, in prelim-
inary hydrogenation tests, we used the bare PK suspended in
decalin with the usual concentration of CAL. Under the typical reac-
tion conditions (Table 2) no consumption of hydrogen and of CAL
was ascertained, thus proving the complete inactivity of the bare
support. The inertness of the support was proved in additional pre-
liminary experiments, in which the Pd/PK catalyst was suspended
3.2.2. Hydrogenation of CAL
All supported Pd catalysts were screened in the selective
hydrogenation of cinnamaldehyde (CAL) to hydrocinnamaldehyde
(HCAL), which consists of a parallel and consecutive reduction of
different functional groups, C C and C O double bonds. All cat-
alytic CAL hydrogenation tests were carried out under comparable
conditions, using identical amounts of substrate and of Pd (2.6 mg of
Scheme 2. Schematic description of the synthesis steps for the preparation of Pd/PK catalysts according to MVS technique.

56
A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
Fig. 5. TEM and particle size distribution of Samples 6 and 7.
Pd) suspended in decalin, working at 100 ◦C under constant hydro-
gen pressure (2 MPa). It is remarkable that the unique ascertained
by-product was always 3-phenyl-1-propanol, whereas only traces
of cinnamyl alcohol were detected in the reaction mixture.
The catalytic performances are reported in Table 2, where the
activity after 30 min, the activity and the selectivity to HCAL at
total substrate conversion and the turnover frequency after 1 h of
reaction are reported.
Table 2
Catalytic performances of supported Pd in the hydrogenation of cinnamaldehyde to hydrocinnamaldehyde for Samples 1, 2, 3, 4, 5, 6 and 7 (reaction conditions: 100 ◦C, 2 MPa
of H2, 2.6 mg of Pd, 3.8 ml of cinnamaldehyde in 100 ml of decalin as solvent).
TOF (h−1
)
a
b
c
Run
Sample sample description
A30
A100%
Sel. to HCAL at 100 mol%
conversion (mol%)
min
(mol (gPd h)−1
)
(mol (gPd h)−1
)
1
2
3
4
5
6
7
Sample 1
0.9 wt% Pd/PK, MW
Sample 2
0.6 wt% Pd/PK, MW
Sample 3
3.4 wt% Pd/PK autoclave, EtOH
Sample 4
4.0 wt% Pd/PK autoclave, n-BuOH
Sample 5
0.6 wt% Pd/PK autoclave, n-BuOH
Sample 6
3.2 wt% Pd/PK MVS
Sample 7
8.33
1.16
76.8
79.6
83.2
78.2
80.3
79.8
87.7
444
954
11.31
6.62
1.79
1.37
1.29
3.87
11.61
5.81
648
4.13
690
15.04
22.87
21.25
1040
1236
1214
1.0 wt% Pd/PK MVS
a
Activity measured at 30 min of reaction.
b
c
Activity measured at 100 mol% CAL conversion.
Turnover frequency after 1 h of reaction, expressed as moles of converted cinnamaldehyde (moles of Pd h)−1
.

A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
57
The most active systems resulted those with lower Pd loading,
characterized by small particle sizes and quite narrow particle size
distributions, as found with other catalysts [28,35,37,51]. The small
particles with a narrow size distribution result more active and
selective toward the formation of HCAL than big ones, whose sur-
face leads to an enhanced repulsion between the phenyl ring of
the rigid molecule of CAL and the metal nanoparticles. This effect
increases the distance between C C bond and the catalyst surface
causing a lowering of the selectivity toward the C C bond hydro-
genation. The influence of the morphology and the presence of
multi-facet nanocrystals on the surface of the catalyst have been
also found by other authors both for Pd and other metals, such
as gold [52,53]. The geometric morphology, the close interaction
between metal and the support and the presence of multi-faceted
crystals affect the adsorption mode of cinnamaldehyde on the sur-
face of metal nanocrystals. In particular, the elongated shape of
nanoparticles together with the presence of multi-faceted crys-
tals weaken the steric effect of benzene ring in cinnamaldehyde
molecule. Consequently, the adsorption of C C bond is promoted.
It is also rational to suppose that the portion of multi-facet crys-
tal increases by the reduction of Pd particles. A confirmation of
this statement can be deduced from the comparison between the
performances of Samples 1 and 2 with those of Samples 4 and 5,
although these two last catalysts are characterized by a marked
different Pd loading (compare run 1 with run 2 and run 4 with
run 5). This result perfectly matches with the statement above
reported: the smaller the nanoparticles are, the higher the turnover
frequency results. Samples 2 and 5 show an average diameter of
4.9 and 5.8 nm respectively against their corresponding systems at
higher Pd loading of 11.3 and 8.1 nm respectively (Samples 1 and
4). This conclusion is supported by TEM results. In fact, Samples 2
and 5 with a lower Pd loading are characterized by more homo-
geneous nanoparticle size dispersions in comparison with their
corresponding ones.
Fig. 6. CAL conversion (mol%) and selectivity to HCAL (mol%) for Samples 3 and
4 (reaction conditions: 100 ◦C, 2 MPa of H2, 2.6 mg Pd, 3.8 ml cinnamaldehyde in
100 ml of decalin as solvent): (ꢀ), Sample 3; (ꢁ), Sample 4.
Samples 6 and 7 were prepared according to the MVS tech-
nique, showed higher catalytic activity (A = 22.87 h−1, run 6, and
A = 21.25 h−1, run 7, Table 1) than the above reported samples
prepared by different procedures. Once again, these results are
consistent with TEM analyses of these samples, which showed
a smaller nanoparticle size and a narrower metal particle size
distribution, thus again supporting the above underlined strict cor-
relation between catalytic activity and structural features of the
catalytic systems. It is worthy to note that Sample 7, containing the
smallest dimensions show the highest selectivity (87.7%) toward
HCAL.
An interesting comparison between the different adopted
preparation techniques is reported in Fig. 7 where the relations
between the activity after 30 min of reaction and the selectivity to
HCAL, in function of Pd particle sizes are reported.
When MW and autoclave preparation procedures are compared,
the highest activity and selectivity are ascertained for samples char-
acterized by surface Pd atoms of the smallest particles, according
to the published researches [28,35,37,51], confirming that the par-
ticle dimensions play the major role in this catalytic test, under the
adopted reaction conditions. This trend is also found for the relation
between selectivity to HCAL and Pd particle size for MVS samples
(Samples 6 and 7). Regarding the catalytic activity for these sam-
ples it is possible to observe that they show comparable activity
Samples 3 and 4, with approximately the same Pd loading (3.4
and 4.0 wt% respectively) were synthesized according the same
solvothermal route in the autoclave adopting a different alcohol,
EtOH for Sample 3 and n-BuOH for Sample 4. Sample 3 resulted
slightly more active than Sample 4 at every CAN conversion
(Fig. 6).
Again, the catalytic results can be related to a different Pd parti-
cle size, Sample 3 showing an average diameter of 5.5 nm, whereas
Sample 4 a larger one of 8.1 nm. These data suggest that also the
reducing agent/solvent employed in the catalyst preparation is an
important factor for the nanoparticles morphology and thus for the
catalytic performances.
Fig. 7. Dependency of activity30
Autoclave Samples, 4 and 5; (ꢁ), MVS Samples, 6 and 7.
(mol (gPd h)−1) and selectivity to HCAL (mol%) at 100 mol% CAL conversion on Pd metal particle size: (ꢀ), MW Samples, 1 and 2; (
),
min

58
A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
illustrated in Fig. 9, after 1 h of reaction. Only a small decrease of
the yield to HCAL was ascertained for the second recycle of sample
7, probably due to a certain loss of the catalyst during the filtration
and recycle steps. Thus, it is rational to claim a promising stability
and consequent reusability of these systems.
In addition, in order to exclude metal leaching from the sup-
ported Pd systems leading to highly active soluble Pd species, we
tested the solution recovered by filtration after the catalytic runs
of Samples 2, 5 and 7. No further catalytic activity of the liquid
phase was ascertained. This result allowed us to ascribe the activ-
ity of these Pd/PK catalysts to the heterogeneous phase and not
to Pd species leached into solution, thus confirming the stabilizing
effect of this unusual polymeric support, which helps not only the
Pd dispersion but also its anchorage.
Finally, it is also worth pointing out that these Pd/PK cata-
lysts lead to better or at least comparable results with respect to
those previously obtained by employing Pd catalysts supported
on functionalized styrene/divinylbenzene under similar reaction
conditions [24].
Fig. 8. CAL conversion (mol%) and selectivity to HCAL (mol%) for Samples 2, 5 and
7 in order to compare the three different preparation techniques (reaction condi-
tions: 100 ◦C, 2 MPa of H2, 2.6 mg Pd, 3.8 ml cinnamaldehyde in 100 ml of decalin as
solvent): (ꢀ), Samples 2; (
), Sample 5; (ꢁ), Sample 7.
4. Conclusions
and this is in agreement with their Pd particle size, that are very
close (4.9 and 4.6 nm for Samples 6 and 7, respectively).
It is also noteworthy the relation between the catalytic behavior
and the preparation technique of three samples (Samples 2, 5 and 7)
characterized by the lower Pd loading but synthesized respectively
according MW irradiation, autoclave and MVS preparations (Fig. 8).
On the basis of these results, we deduce that the MVS tech-
nique is very effective and significantly better than the other two
procedures, which however, present an advantage, i.e. a shorter
preparation time. In addition, these methodologies allow the use
of milder reducing agents in comparison with the most traditional
H2, MBH4.
PK can be considered a useful candidate as support and stabi-
lizing agent in the preparation of Pd nanoparticles. This statement
has been verified by adopting three different synthetic approaches.
All the prepared samples show interesting morphological features
and promising catalytic performances in the hydrogenation of cin-
namaldehyde to hydrocinnamaldehyde, even when they are used
in successive catalytic cycles. This is the first time that, to the best of
our knowledge, the synthesis of Pd catalysts supported on PK has
been reported and they have been used in the hydrogenation of
CAL to HCAL. These preliminary results highlighted that, not with-
standing the low surface area, PK could be a very interesting and
encouraging support in order to obtain efficient, stable, leaching-
free metal catalysts. The presented results suggest that the Pd/PK
system and, more in general, metal/PK systems may be useful cat-
alysts for many other reactions of industrial interest where metal
leaching is a critical point, not yet solved. Work is in progress in
order to extend the applications of this new family of systems to
other catalytic processes.
3.2.3. Resuse of the catalysts
Notably, the reuse of Samples 2, 5 and 7 recovered from the
catalytic runs 2, 5 and 7 respectively leads to almost reproducible
results in terms of both activity and selectivity, indicating that the
catalysts appear can be almost completely recovered unchanged.
The successive recycle runs of Sample 2 and Sample 5 gave
quite analogous results, without loss of activity and selectivity, as
Acknowledgments
Prof. G. Martra and Dr. M. Manzoli of the Department of Chem-
istry IFM at University of Turin are gratefully acknowledged for
some TEM analyses. We are also very grateful to Mr. Marco Mar-
tinelli for his precious technical assistance. In addition, the authors
thank the project PRIN 2008 Prot. 2008 SXASBC 002 for financial
support.
References
[1] R. Ferrando, J. Jellinek, R.L. Johnston, Chem. Rev. 108 (2008) 845–910.
[2] X. Wang, J. Zhuang, Q. Peng, Y.D. Li, Nature 437 (2005) 121–124.
[3] R.G. Chaudhuri, S. Paria, Chem. Rev. 112 (2012) 2373–2433.
[4] H. Bonnemann, K.S. Nagabhushana, Encyclopedia of Nanoscience and Nano-
technology, vol. 1, American Scientific, Los Angeles, 2004, pp. 779–815.
[5] A.G. Boudjahem, A. Redjel, T. Mokrane, J. Ind. Eng. Chem. 18 (2012) 303–308.
[6] Y.I. Qingfeng, N.I.U. Fengjuan, Rare Metals 30 (2011) 332–336.
[7] R.J. White, R. Luque, V.L. Budarin, J.H. Clark, D.J. Macquarrie, Chem. Soc. Rev. 38
(2009) 481–494.
[8] L.D. Pachon, G. Rothenberg, Appl. Organomet. Chem. 22 (2008) 288–299.
[9] J.P. Wilcoxon, B.L. Abrams, Chem. Soc. Rev. 35 (2006) 1162–1194.
[10] M. Guerrero, J. García-Antón, M. Tristany, J. Pons, J. Ros, K. Philippot, P. Lecante,
B. Chaudret, Langmuir 26 (2010) 15532–15540.
[11] N. Wangoo, K.K. Bhasin, S.K. Mehta, C. Raman Suri, J. Colloid Interface Sci. 323
(2008) 247–254.
[12] M. Ganesan, R.G. Freemantle, S.O. Obare, Chem. Mater. 19 (2007) 3464–3471.
[13] D. Astruc, Inorg. Chem. 46 (2007) 1884–1894.
Fig. 9. CAL conversion (mol%) and selectivity to HCAL (mol%) after 1 h of reaction
in the presence of Samples 2, 5 and 7 (first hydrogenation runs, runs 2, 5 and 7
respectively and successive recycles of the solid catalysts).

A.M. Raspolli Galletti et al. / Applied Catalysis A: General 447–448 (2012) 49–59
59
[14] H. Lee, S.E. Habas, S. Kweskin, D. Butcher, G.A. Somorjai, P.D. Yang, Angew.
Chem. Int. Ed. 45 (2006) 7824–7828.
[33] C. Evangelisti, N. Panziera, P. Pertici, G. Vitulli, P. Salvadori, C. Battocchio, G.
Polzonetti, J. Catal. 262 (2009) 287–293.
[15] M.R. Jones, K.D. Osberg, R.J. Macfarlane, M.R. Langille, C.A. Mirkin, Chem. Rev.
111 (2011) 3736–3827.
[16] S.B. Yin, S.C. Mu, H.F. Lv, N.A.C. Cheng, M. Pan, Z.Y. Fu, Appl. Catal. B: Environ.
93 (2010) 233–240.
[17] Y. Zhang, H. Zhang, Y. Zhang, Y. Ma, H. Zhong, H. Ma, Chem. Commun. (2009)
6589–6591.
[18] B. Corain, G. Schmid, N. Toshima, B. Corain, G. Schmid, N. Toshima (Eds.), Metal
Nanoclusters in Catalysis and Materials Science: The Issue of Size Control, Else-
vier, Amsterdam, 2008, pp. 201–232.
[34] C. Evangelisti, E. Schiavi, L.A. Aronica, A.M. Caporusso, G. Vitulli,
L. Bertinetti, G. Martra, A. Balerna, S. Mobilio, J. Catal. 286 (2012)
224–236.
[35] P. Gallezot, D. Richard, Catal. Rev. Sci. Eng. 40 (1,2) (1998) 81–126.
[36] A. Vavasori, L. Toniolo, J. Mol. Catal. A: Chem. 110 (1996) 13–23.
[37] Z.C. Zhang, X. Zhang, Q.Y. Yu, Z.Z.C. Liu, C.M. Xu, J.S. Gao, J. Zhuang, X. Wang,
Chem. Eur. J. 18 (2012) 2639–2645.
[38] K.J.A. Raj, M.G. Prakash, T. Elangovan, B. Viswanathan, Catal. Lett. 142 (2012)
87–94.
[19] K. Rahme, F. Gauffre, J.D. Marty, B. Payré, C. Minotaud, J. Phys. Chem. C 111
(2007) 7273–7279.
[39] C. Stangel, G. Charalambidis, V. Varda, A.G. Coutsolelos, I.D. Kostas, Eur. J. Inorg.
Chem (2011) 4709–4716.
[20] Y. Borodoko, M. Humphrey, T. Don Tilley, H. Frei, G.A. Somorjai, J. Phys. Chem.
C 111 (2007) 6288–6295.
[40] A. Ungureanu, B. Dragoi, A. Chirieac, S. Royer, D. Duprez, E. Dumitriu, J. Mater.
Chem. 21 (2011) 12529–12541.
[21] N. Pérignon, J.D. Marty, A.F. Mingotaud, M. Dumont, I. Rico-Lattes, C. Mintotaud,
[41] I. Biondi, G. Laurenczy, P.J. Dyson, Inorg. Chem. 50 (2011)
8038–8045.
Macromolecules 40 (2007) 3034–3041.
[22] K. Strzelec, K. Wa˛sikowska, M. Cypryk, P. Pospiech, e-Polymers 024 (2011) 1–12.
[23] K. Zhan, H. You, W. Liu, J. Lu, P. Lu, J. Dong, React. Funct. Polym. 71 (2011)
756–765.
[42] E.V. Ramos-Fernández, J. Ruiz-Martínez, J.C. Serrano-Ruiz, J. Silvestre-Albero,
A. Sepúlveda-Escribano, F. Rodríguez-Reinoso, Appl. Catal. A: Gen. 402 (2011)
50–58.
[24] F. Benvenuti, C. Carlini, M. Marchionna, A.M. Raspolli Galletti, G. Sbrana, J. Mol.
Catal. A: Chem. 145 (1999) 221–228.
[25] I. Favier, D. Picurelli, C. Pradel, J. Durand, B. Milani, M. Gomez, Inorg. Chem.
Commun. 13 (2010) 766–768.
[26] E. Drent, P.H.M. Budzelaar, Chem. Rev. 96 (1996) 663–681.
[27] J.M. Lagaron, M.E. Vickers, A.K. Powell, N.S. Davidson, Polymer 41 (2000)
3011–3017.
[43] M. Boudart, Nature 372 (1994) 320.
[44] S. Mahmoud, A. Hammoudeh, S. Gharaibeh, J. Melsheimer, J. Mol. Catal. A:
Chem. 178 (2002) 161–167.
[45] L.B. Alemany, D.M. Grant, R.J. Pugmire, T.D. Alger, K.W. Zilm, J. Am. Chem. Soc.
105 (1983) 2133–2141.
[46] E.O. Stejskal, J. Schaefer, M.D. Selcik, R.A. McKay, Macromolecules 14 (1981)
275–279.
[28] A.M. Raspolli Galletti, C. Antonetti, A.M. Venezia, G. Giambastiani, Appl. Catal.
A: Gen. 386 (2010) 124–131.
[29] A.M. Raspolli Galletti, C. Antonetti, I. Longo, G. Capannelli, A.M. Venezia, Appl.
Catal. A: Gen. 350 (2008) 46–52.
[30] C. Antonetti, A.M. Raspolli Galletti, I. Longo, Int. J. Chem. Reactor Eng. 8 (2010)
1–17.
[47] D.A. Torchia, J. Magn. Reson. 30 (1978) 613–616.
[48] P.M.A. Sherwood, in: D. Briggs, M.P. Seah (Eds.), Practical Surface Analysis,
Wiley, New York, 1990, p. 181.
[49] M. Geppi, C. Forte, J. Magn. Reson. 137 (1999) 177–185.
[50] V.Z. Radkevich, T.L. Senko, K. Wilson, L.M. Grishenko, A.N. Zaderko, V.Y. Diyuk,
Appl. Catal. A: Gen. 335 (2008) 241–251.
[31] C. Antonetti, M. Oubenali, A.M. Raspolli Galletti, P. Serp, G. Vannucci, Appl. Catal.
A: Gen. 421 (2012) 99–107.
[32] A.M. Raspolli Galletti, C. Antonetti, S. Giaiacopi, O. Piccolo, A.M. Venezia, Top.
Catal. 52 (2009) 1065–1069.
[51] P. Gallezot, A. Giroir-Fendler, D. Richard, Catal. Lett. 5 (1990) 169–174.
[52] A. Giroir-Fendler, D. Richard, P. Gallezot, Catal. Lett.
175–182.
5
(1990)
[53] F. Delbecq, P. Sautet, J. Catal. 152 (1995) 217–236.