Bimetallic Gold-Palladium vapour derived catalysts: The role of structural features on their catalytic activity

Article abstract of DOI:10.1016/j.jcat.2011.11.007

A procedure to synthesize new Au-Pd bimetallic catalysts using Au and Pd vapours as reagents (metal vapour synthesis, MVS) is reported. The simultaneous co-condensation of Au and Pd vapours with acetone vapour affords Au-Pd/acetone-solvated metal atoms wh

Full text of DOI:10.1016/j.jcat.2011.11.007

Journal of Catalysis 286 (2012) 224–236
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Bimetallic Gold–Palladium vapour derived catalysts: The role of structural features
on their catalytic activity
b
b
b,c
Claudio Evangelisti a, , Eleonora Schiavi , Laura Antonella Aronica , Anna Maria Caporusso
,
⇑
Giovanni Vitulli c, Luca Bertinetti d, Gianmario Martra d, Antonella Balerna e, Settimio Mobilio f
a CNR, Institute of Molecular Science and Technologies (ISTM), Via G. Fantoli 15/16, 20138 Milano, Italy
b Department of Chemistry and Industrial Chemistry, University of Pisa, Via Risorgimento 35, 56126 Pisa, Italy
c Advanced Catalysts S.r.l., Via Risorgimento 35, 56126 Pisa, Italy
d Department of Chemistry IFM & NIS Interdipartimental Centre of Excellence, Via P. Giuria 7, 10125 Torino, Italy
e INFN – Frascati National Laboratories, Frascati, Via E. Fermi 40, 00044 Frascati, Roma, Italy
f Department of Physics, University of Roma TRE, Via della Vasca Navale 84, 00146 Roma, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
A procedure to synthesize new Au–Pd bimetallic catalysts using Au and Pd vapours as reagents (metal
vapour synthesis, MVS) is reported. The simultaneous co-condensation of Au and Pd vapours with ace-
tone vapour affords Au–Pd/acetone-solvated metal atoms which have been used to deposit Au–Pd bime-
Received 29 August 2011
Revised 7 November 2011
Accepted 8 November 2011
Available online 16 December 2011
tallic nanoparticles on c-alumina and titanium oxide supports. Transmission electron microscopy (TEM)
analysis determined the nanoparticles dimensions (dm = 2.2–2.4 nm) and size distribution while the X-
ray absorption spectroscopy (XAS) analysis showed the presence of small bimetallic Au–Pd nanoparticles
with a large amount of Au–Pd bonds. The bimetallic co-condensed systems, tested in the selective oxida-
tion of benzyl alcohol with molecular oxygen both in toluene solvent and in solvent-free conditions,
showed higher catalytic activity and selectivity than the corresponding monometallic systems as well
as of the analogous systems obtained by separate evaporation of the two metals.
Keywords:
Au–Pd catalysts
Metal vapour synthesis
TEM and EXAFS analyses
Benzyl alcohol oxidation
Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction
solid solutions at any ratio [15] and depending on their preparation
method, Au–Pd bimetallic systems can contain different structural
Bimetallic systems are of great interest from both scientific and
technological point of view. Their catalytic, electronic and optical
properties are different from those of their pure constituent metals
and are influenced by different factors: not only from particles size
and shape, as in monometallic nanoparticles, but also from the
atomic ratio of the metals and the electronic and chemical structure
of bimetallic nanoparticles [1–3]. Among the numerous metals
which were investigated in combination, gold–palladium systems
have been extensively studied because of their peculiar catalytic
properties [4]. Bimetallic gold–palladium catalysts have been
recently reported as promoters of processes such as oxidation of
alcohols to corresponding aldehydes or ketones [5–7], production
of hydrogen peroxide [8,9], production of vinyl acetate [10,11],
hydrogenation of unsaturated compounds [12,13] and hydrode-
chlorination of chloro- and chlorofluorocarbons [14].
features including core–shell structure [8,16–18], segregated Au
and Pd monometallic domains [19], uniform Au–Pd alloys [20,21]
or a mixture of them [22,23].
In this work, we report the preparation of supported Au–Pd
bimetallic systems by using Pd and Au vapours as reagents (metal
vapour synthesis, MVS) [24] following two different synthetic pro-
cedures. In the first one, Au and Pd vapours were co-condensed
simultaneously with acetone vapour in order to obtain a close
interaction between the two metals, leading to a bimetallic Au–
Pd/acetone solvated metal atoms (SMAs) solution. Then, Au–Pd/
acetone SMA was used as starting material to synthesize bimetallic
[Au–Pd] catalysts supported on
c-alumina and titanium oxide,
respectively. In the second one, Au and Pd were evaporated sepa-
rately in two different evaporation reactions; therefore, Pd/acetone
and a Au/acetone SMA were achieved. Afterwards, Pd and Au SMA
were mixed together and used as starting materials to obtain
supported [Au][Pd] bimetallic systems.
In order to have a further insight into the morphology and the
structural properties of the Au–Pd bimetallic nanoparticles and to
rationalize the role played by the two different MVS procedures
Recently, several investigations towards the relationship be-
tween catalytic activity and structural and electronic features of
Au–Pd bimetallic catalysts have been reported. Au and Pd can form
used, Au–Pd bimetallic systems supported on c-alumina and on tita-
nium oxide were investigated by the transmission electron micros-
⇑
Corresponding author. Present address: CNR, Institute of Molecular Science and
Technologies, Via C. Golgi 19, 20133 Milano, Italy. Fax: +39 050 2219260.
copy (TEM) and X-ray absorption spectroscopy (XAS) analyses.
E-mail address: claudio@dcci.unipi.it (C. Evangelisti).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.11.007

C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
225
The catalytic activity of Au–Pd bimetallic systems was studied
in the selective oxidation of benzyl alcohol to benzaldehyde, cho-
sen as test reaction, evidencing a close relationship between the
structural features of these bimetallic systems, induced by the dif-
ferent preparation procedures used, and their catalytic behaviour.
The selective oxidations of alcohols to their respective aldehydes
or ketones are a processes of great relevance from a laboratory
and industrial point of view [25,26]. It is well known that both
highly dispersed palladium [27] and gold [28] catalysts promote
these processes using molecular oxygen as environmental friendly
reagent respect to stoichiometrical reagents such as permanganate
or dichromate, which are traditionally employed [29,30].
Recently, Au–Pd bimetallic catalysts have shown to be very
active in these reactions evidencing a beneficial interaction be-
tween the two metals depending on the structural characteristics
of these bimetallic systems. Significant advances in the field has
been reported by Dimitratos et al. [31] and Enache et al. [6] which
achieved high catalytic efficiency under solvent and solvent-free
conditions, respectively, with supported Au–Pd bimetallic nano-
particles obtained by reduction in Au and Pd salts.
di. TEM analysis was performed on both monometallic and bime-
tallic samples supported on Al2O3 and TiO2. The instrument condi-
tion at the time of the measurements resulted in a minimum cross
section of the electron beam during EDX analysis of ca. 5 nm.
The X-ray absorption spectroscopy (XAS) measurements at the
Au L3 edge (11,919 eV) and at the Pd K-edge (24,350 eV) were per-
formed at the GILDA beamline [33] of the European Synchrotron
Radiation Facility (ESRF, Grenoble). The X-ray beam was mono-
chromatized and horizontally focused by using a double crystal
Si(311) monochromator [34]; two Pd mirrors at the Au L3 edge
and two Pt mirrors at the Pd K-edge were used to reject the higher
harmonics and focus the beam in the vertical direction [33]. XAS
measurements at the Au L3 and at the Pd K edges were performed
on the two reference samples (Au and Pd foils) and on a complete
set of two monometallic (Au (1%) and Pd (0.5%) and two bimetallic
[Au][Pd], [AuPd] samples supported on
c-Al2O3. Two bimetallic
[Au][Pd], [AuPd] samples supported on TiO2 were measured only
at the Au L3 edge. The catalytic samples were prepared for the X-
ray absorption spectroscopy (XAS) experiments mixing their fine
powders with boron-nitride and pressing them into pellets. The
different sample amounts were calculated [35] in order to achieve
appropriate edge jumps.
2. Experimental
Both reference samples and the catalysts at the Au L3 edge were
measured in transmission mode with two Ar-filled ionization
chambers to detect the incident (I0) and transmitted (IT) X-ray
beam. The catalysts at the Pd K-edge were measured in fluores-
cence mode (IF) by using a 13-element high-purity Ge multi-detec-
tor coupled to an X-ray pulse digital analyser for a more accurate
dead-time effect correction [36].
2.1. General
All operations involving the MVS products were performed
under a dry argon atmosphere. The co-condensation of gold,
palladium and acetone vapours was carried out in a static reactor
similar to those previously described [32]. In the case of co-evapo-
ration experiments, the MVS apparatus was equipped with two
alumina-coated tungsten crucibles heated by Joule effect with
two Rial AEJ2 generators, respectively, with a maximum power
of 2 kW that allowed to evaporate independently two metals.
The acetone-solvated metal atoms solutions were handled under
argon atmosphere with the use of the standard Schlenk techniques.
The amount of palladium and gold in the above solutions was
determined by inductively coupled plasma–optical emission spec-
trometers (ICP–OES) with a Spectro-Genesis instrument equipped
with a software Smart Analyser Vision. For ICP–OES, a sample
(1 mL) of SMA solutions 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 palladium and gold was 2 ppb.
Acetone was purified by conventional methods, distilled and
The absorption coefficients were calculated as
l(E) = ln(I0/IT) in
transmission mode and as (E) = (IF/I0) in fluorescence mode. All
l
spectra were recorded at 20 K in order to reduce the thermal
effects.
2.2. Preparation of the [AuPd] catalysts
In a typical experiment (Scheme 1), palladium and gold vapours
generated by resistive heating of two different alumina-coated
tungsten crucibles filled with ca. 150 mg of Pd powder and 300
mg of Au pellets, respectively, were co-condensed simultaneously
at liquid nitrogen temperature with acetone (100 mL) in the glass
reactor chamber of the MVS apparatus for a chosen time. The
reactor chamber was warmed to the melting point of the solid
matrix, and the resulting red–brown solution was siphoned at a
low temperature into a Schlenk tube and kept in a refrigerator at
ꢀ40 °C. The metal amount recovered in SMA was about the
80 wt.% of the evaporated metal. The palladium and gold content,
obtained by ICP–OES analysis, were 0.5 mg/mL of Pd and 1.0 mg/
mL of Au, corresponding to a Pd/Au molar ratio = 1. 70 mL of the
[Au–Pd]/acetone SMA (0.33 mmols, 35 mg of Pd and 0.35 mmols,
stored under argon. Commercial
c-Al2O3 (Chimet product, type
49, 3.1
l
m average particle size, surface area 110 m2 gꢀ1), TiO2
(Degussa product, powder containing anatase (80 wt.%) and rutile
(20 wt.%), P99.5%, 21 nm average particle size, surface area
50 15 m2 gꢀ1) were dried in a static oven before use. Benzyl
alcohol was supplied from Aldrich and used as received. The GLC
analyses were performed on a Perkin–Elmer Auto System gas
chromatograph, equipped with a flame ionization detector (FID),
70 mg of Au) was added to a suspension of
c-Al2O3 or TiO2
(7.00 g) in acetone (30 mL). The mixture was stirred for 12 h at
room temperature. Au–Pd bimetallic nanoparticles were quantita-
tively deposited on the supports; the colourless solution was
removed, and the light-brown solid was washed with n-pentane
and dried under reduced pressure. By this way Au–Pd bimetallic
catalysts containing 1 wt.% of Au and 0.5 wt.% of Pd, represented
using a SiO2 column (BP-1, 12 m ꢁ 0.3 mm, 0.25
lm) and helium
as carrier gas.
Electron micrographs were obtained by a Jeol 3010-UHR high-
resolution transmission electron microscope (HRTEM) equipped
with a LaB6 filament operating at 300 kV) equipped with an Oxford
Inca Energy TEM 300 EDS X-rays analyser. Before the introduction
in the instrument, the samples, in the form of powders, were ultra-
sonically dispersed in isopropyl alcohol and a drop of the suspen-
sion was deposited on a copper grid covered with a lacey carbon
film. Histograms of the particle size distribution were obtained
by counting onto the micrographs at least 300 particles; the mean
particle diameter (dm) was calculated by using the formula
as [AuPd]/c-Al2O3 and [AuPd]/TiO2, respectively, were obtained.
2.3. Preparation of the [Au][Pd] catalysts
In a typical experiment (Scheme 2), palladium vapour gener-
ated by resistive heating of an alumina-coated tungsten crucible
filled with ca. 200 mg of Pd powder was co-condensed at liquid
nitrogen temperature with acetone (100 mL) in the glass reactor
chamber of the MVS apparatus for a chosen time. The reactor
dm = Rdini/Rni where ni was the number of particles of diameter

226
C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
Stable at low
temperature (-40°C
)
Acetone (v)
Au - Pd clustering
Acetone solvated
Au-Pd nanoparticles
Au (atoms) + Pd (atoms)
Au-Pd Acetone
upon melting and
warming to -40°C
Frozen Matrix
(-196°C)
vaporization
Support
25°C
(1)
Au
Pd
bulk metal
bulk metal
[AuPd]/Support
Support: γ-Al2O3, TiO2
Scheme 1.
Stable at low
temperature (-40°C)
Acetone solvated
Au nanoparticles
Au / acetone
Support
25°C
(2)
[Au][Pd]/Support
Support: γ-Al2O3, TiO2
Acetone solvated
Pd nanoparticles
Pd / acetone
Scheme 2.
chamber was warmed to the melting 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 palla-
dium content of the Pd/acetone was 0.81 mg/mL. Similarly, 200 mg
of Au pellets were co-condensed at liquid nitrogen temperature
with acetone (100 mL), obtaining a red–purple solution containing
0.83 mg/mL of Au. Then, the 25 mL of Pd/acetone SMA (0.19
mmols, 20.2 mg of Pd) was added at low temperature (ꢀ30 °C) to
Au/acetone SMA (49 mL, 0.2 mmols, 40.6 mg of Au). The resulting
solution, represented as [Au][Pd]/acetone SMA containing the a
autoclave was then purged three times with molecular oxygen
leaving the mixture at the desired oxygen pressure. The reaction
was stirred (1200 rpm) at 100 °C. The reaction was monitored by
GLC analysis of liquid samples removed by the tap. In the case of
recycling experiments, after 8 h, the solid catalyst was separated
from the mixture by filtration (0.4 lm Teflon filter) and the solution
was analysed by ICP–OES to reveal the Pd and Au amount. Where
indicated, the recovered catalyst was washed with acetone and
re-used in further batch runs.
Au:Pd molar ratio = 1, was added to a suspension of c-Al2O3 or
3. Results and discussion
TiO2 (4.06 g) in acetone (20 mL). The mixture was stirred for 12 h
at room temperature. The colourless solution was removed and
the light-brown solid was washed with n-pentane and dried under
reduced pressure. Similarly to those previously reported, these sys-
tems contained 1 wt.% of Au and 0.5 wt.% of Pd.
3.1. TEM results
The analysis of the size distribution and dispersion of the
metallic particles was performed using transmission electron
microscopy (TEM). The data related to the observations carried
2.4. Catalytic oxidation of benzyl alcohol in toluene
out on the monometallic catalysts supported on
reported in Fig. 1.
c-Al2O3 are
In a typical experience, in two-necked round-bottom flask
(100 mL) equipped with a stirring magnetic bar, a silicon stopper
and a glass cannula, connected to the oxygen reservoir, catalyst
(60 mg), toluene (10 mL) and benzyl alcohol (0.31 mL, 3.0 mmol)
were added. The reaction mixture was kept at 0.1 MPa of O2, at
60 °C and the stirrer was set at 1250 rpm. The progress and the
composition of the reaction mixture were determined by GLC anal-
ysis of liquid samples taken from the stoppered side neck with a
syringe.
Concerning the Pd/c-Al2O3 catalyst, for the most part of the
sample regions inspected, EDX analysis revealed the presence of
Pd, but no metal particles large enough in size (P1.0 nm) to be
detected by TEM were observed (Fig. 1A), indicating a very high
dispersion of the metal phase. Only in few regions, some Pd
particles with sizes in the (1.5–5.5) nm range were found (their
number was too low to obtain a statistically significant size distri-
bution) so they represent a very minor fraction of the supported
particles (Fig. 1A). Conversely, a number of metal particles quite
homogeneously distributed on the alumina grains appeared in
2.5. Catalytic oxidation of benzyl alcohol in solvent-free conditions
the images of Au/c-Al2O3, exhibiting a kind of bimodal size
distribution, with the 97% of gold particles having a size distribu-
tion between 1.5 and 7.0 nm (dm = 4.5 0.2 nm), and the rest
between 7.5 and 10.0 nm (Fig. 1B), in agreement with the results
reported in Ref. [46].
Moving to the bimetallic catalysts supported on alumina
(Fig. 2), for both [AuPd]/c-Al2O3 (section A) and [Au][Pd]/c-Al2O3
The oxidation experiments were performed at 0.5 MPa starting
molecular oxygen pressure using a 25 mL stainless-steel autoclave
equipped with a Teflon removable cylinder Jacker, a magnetic bar, a
stainless-steel tap and a manometer (0.25 MPa end scale). The reac-
tions were performed by heating the autoclave in an oil bath at
100 °C ( 1 °C). In a typical experiment, the autoclave was charged
with catalyst (125 mg) and benzyl alcohol (10 mL, 0.971 mol). The
(section B) samples, the support appeared densely populated by
nanometric metal particles. As indicated by the size distributions,

C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
227
A
B
20 nm
20 nm
Al
0
2
4
6
keV
Fig. 1. TEM/EDX data obtained for Pd/
-Al2O3 catalyst; inset: image representative of the few zones with metal particle observable by TEM; lover panel: EDX spectrum of the region displayed in the main frame
(Cu signals are due to the sample holder). Data in section B: image representative of Au/ -Al2O3 and histogram of the metal particle size distribution obtained for this catalyst.
c-Al2O3 (section A) and Au/c-Al2O3 (section B). Data in section A: main frame, image representative of the overwhelming part of the Pd/
c
c
Fig. 2. TEM micrographs and metal particle size distributions of [Au][Pd]/c-Al2O3 and [AuPd]/c-Al2O3 (sections A and B, respectively).

228
C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
their sizes lie in a quite narrow range (1–6 nm), but while in the
[AuPd] case, the nanoparticles are homogeneously and symmetri-
cally distributed around a mean size of about 2.4 0.1 nm, and in
the [Au][Pd] case, the distribution is quite different. In the [Au][Pd]
sample about the 70% of the nanoparticles exhibits a size distribu-
tion between 2 nm and 3 nm and a dm of 2.6 0.1 nm but there is
also a minor population of larger particles with sizes between 4 nm
and 5.5 nm (about 10% in number, corresponding to about 40% in
volume), indicating the presence of at least two families of nano-
particles. The presence of pure Au in these largest particles, consis-
(Fig. 4B) but the EDX signals of the metals were detected. These re-
sults are similar to ones achieved in the Pd/ -Al2O3 case, whereas
the difference with respect to the Au/ -Al2O3 ones shows a signif-
icant role of the support in the formation of gold nanoparticles.
Previously reported studies on Au particle size growth in solution
[37] showed as the differences in metal particle sizes, evidenced
by TEM analyses, can be related to the different deposition time
of metal particles on the support.
The metal dispersion appeared to be quite high also in the case
of [Au][Pd]/TiO2 (Fig. 5A), where most of the titania particles
seemed to be ‘naked’ in TEM images, but not in the EDX spectra,
and only in some cases, few metal particles with size ranging from
1.0 to 7.0 nm were observed (Fig. 5A), clearly not enough to obtain
a statistically significant size distribution.
c
c
tently with that observed in the monometallic Au/c-Al2O3 sample,
could not be excluded.
Noticeably, the limits and the mean values of the size distribu-
tions found for the bimetallic samples are shifted towards values
smaller than the ones found for Au/
c
-Al2O3, while they are defi-
Differently, supported particles appeared more abundant in the
[AuPd]/TiO2 sample (Fig. 5B), although less densely distributed on
the support than in the [AuPd]/c-Al2O3 case. By considering that
nitely larger than in the Pd/Al2O3 case, where most of the metallic
phase was too small in size to be detected by TEM. It can be surely
inferred that in both bimetallic catalysts preparation procedures,
nucleation and growth of supported particles are affected by the
co-presence of the two metals. As for the localization of the two
metals, EDX analysis indicated that both Pd and Au were present
in the regions exhibiting supported particles. Unfortunately, the
particles appeared too close to each other to prevent separated
analysis. Thus, the possibility to obtain information on the nature
of metal particles by their structural features was attempted, by
analysing images taken in high magnification–high resolution
(HR) conditions (Fig. 3).
For both catalysts, supported particles exhibited interference
fringes between 0.223 and 0.237 nm. The fringes are extended over
the whole particles, indicating the single nanocrystals nature of the
supported metals, and the spacing values are compatibles with
d111 of the two metals that is 0.235 nm vs. 0.224 nm for Au and
Pd, respectively. Unfortunately, the difference in the d111 of the
Au and Pd was too small to allow a distinction by HRTEM between
the two metals or between a metal and a possible bimetallic alloy
(the highest possible resolution we could achieve was 0.012 nm/
pixel, resulting in the possibility to determine lattice spacings with
a precision of about 5%).
the specific surface area of the titania powder used is about one
half of that of alumina, if the [AuPd]/TiO2 supported particles ob-
served in TEM images should correspond to the total amount of
metals present on the support, a much higher particle density
should be present. In this respect, it could be considered that
although the mass density of TiO2 and alumina can be similar, that
latter is often constituted by thin particles, in the form of ‘platelets’
[38], resulting more transparent to the electron beam than TiO2
particles.
Thus, on the basis of the higher dispersion of the monometallic
catalysts, and on the asymmetry of the size distribution of particles
observed by TEM on [AuPd]/TiO2, it can be proposed that a signifi-
cant number of metal particles too small to be detected by TEM may
be present, likely representing the majority of the supported phase.
As for the nature of the observed particles, they were not still sep-
arated enough to obtain unambiguous EDX results, and the analysis
of structural data suffered of the limitation reported above.
3.2. XAS results
The XAS measurements of the Au and Pd foils were used to cal-
ibrate the energy scales, to align the XAS spectra and to evaluate
and fix the respective S20 values.
It is of interest to notice that [AuPd] particles are mostly
isolated and only rarely they touch each other, eventually giving
rise to limited contact interfaces, while in the [Au][Pd] case,
agglomerates of 3–10 particles can be observed much more
frequently, resulting in more extended contact interfaces.
As for the catalysts supported on TiO2, the monometallic ones
exhibited an high metal phase dispersion because no particles
were observed neither for Pd/TiO2 (Fig. 4A) nor for Au/TiO2
The extended X-ray absorption fine structure spectra (EXAFS),
usually indicated as
v(k), are given by the normalized oscillations
superimposed onto the measured absorption coefficient
Quantitatively they are defined as follows:
l(k).
v
ðkÞ ¼ ½ð ðkÞ ꢀ l0ðkÞꢂ=l0ðkÞ
l
Fig. 3. HR-TEM micrographs: interference fringes observed for [Au][Pd] c-Al2O3 and [AuPd]//c-Al2O3 (sections A and B, respectively).

C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
229
A
B
20 nm
20 nm
0
2
4
6
0
2
4
6
keV
keV
Fig. 4. TEM/EDX data obtained for Pd/TiO2 (section A) and Au/TiO2 (section B). Cu signals in EDX spectra are due to the sample holder.
where
l0(k) is the total atomic absorption coefficient and k is the
In Fig. 7, significant differences are clearly visible between the
[Au][Pd] and the [AuPd] samples, showing that their structures
are quite different.
In Fig. 8 (lower panel), the Fourier transforms of the Pd K-edge
EXAFS spectra are shown; also in this case, the first peak of the Pd
foil matches the 12 first neighbours that are at an effective distance
of about 2.75 Å in this case. In the Fourier transforms of the sam-
photoelectron wave vector; it is given by
2
1=2
k ¼ ½4
p
mðE ꢀ E0Þ=h ꢂ
where E is the incoming photon energy and E0 is the threshold
energy; as usually it was determined as the inflection point of the
measured absorption edge.
ples supported on c-Al2O3, the presence of a peak between 1 Å
The EXAFS spectra were extracted from the experimental
X-ray absorption spectra according to standard procedures [39];
they are reported in Fig. 6 and in the upper panel of Figure for
Au L3 and Pd K edges, respectively. In Fig. 7 and in the lower
panel of Fig. 8, their Fourier Transforms (FT) are reported. FT
are composed of a series of peaks: the first peak centred at about
2.75 Å for Au and at about 2.5 Å for Pd corresponds to the first 12
neighbours of the fcc structures and the other peaks correspond
to the outer coordination shells. While the first peak comes only
from single scattering contributions, the others are due to the
superposition of single and multiple scattering contributions. To
perform an accurate data analysis [40], we took into account all
these multiple scattering contributions [40]; data were best fitted
to simulated spectra built using theoretical amplitudes and
phases provided by the FEFF8 code [41,42]. Data fitting was per-
formed in both R-space and k-space with k2 (Pd) and k3 (Au)
weights; also an estimation of the accuracy of the obtained struc-
tural parameters, compatible with the data quality and k range
used [43], was achieved.
and 2 Å is clearly visible: in this case, it is due to a Pd–O contribu-
tion. Relevant differences are clearly visible between the bimetallic
samples also in the FT spectra at the Pd K-edge.
In the fitting procedures of the Au bulk and Au/c-Al2O3 (1 wt.%)
spectra, we included single scattering (SS) e multiple scattering
(MS) contributions, up to the fourth coordination shell. In the case
of Pd bulk and of the monometallic Pd/c-Al2O3 (0.5 wt.%), we con-
sidered contributions up to the second Pd–Pd shell. Finally, for the
bimetallic samples only the first coordination shell was studied. In
all the spectra of the catalysts at the Pd K-edge also the peak due to
the Pd–O contribution was fitted.
To reduce the number of free parameters in the fitting of the Au
bulk and Au/c-Al2O3, the bond lengths of the higher order coordi-
nation shells were all linked to that of the second one according
to a fcc structure, and the MS path lengths were assumed equal
to the corresponding SS path lengths. In the case of Au bulk also
the coordination numbers were fixed to the values of the fcc struc-
ture. The fitting curves obtained for Au bulk and the monometallic
catalyst supported on Al2O3 are compared with the experimental
spectra in the values of coordination numbers (N), nearest-neigh-
In Fig. 9, the first peak in the Fourier Transform of the Au foil
matches the 12 first neighbours at an effective distance of about
2.88 Å in the bulk. The small peak at its left is due the shape of
the Au back-scattering amplitude and on the used FT range and k
weighting factor.
bours distances (R) and Debye–Waller factors (r
2) obtained are re-
ported in Table 1. We underline the excellent agreement achieved
and that the values found for the reference Au bulk sample are in
good agreement with those expected and also with those found

230
C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
A
B
10 nm
5 nm
10 nm
0
2
4
6
keV
Fig. 5. TEM/EDX data obtained for [Au][Pd]/TiO2 (section A) and [AuPd]/TiO2 (section B). Data in section A: main frame, image representative of the overwhelming part of the
[Au][Pd]/TiO2 catalyst; inset: image representative of the few zones with metal particle observable by TEM; lover panel: EDX spectrum of the region displayed in the main
frame (Cu signals are due to the sample holder). Data in section B: image representative of [AuPd]/TiO2 and histogram of the metal particle size distribution obtained for this
catalyst.
Fig. 6. Au L3 edge EXAFS spectra of samples supported on Al2O3 (right panel) and on TiO2 (left panel).
in a very detailed study [44] performed on gold bulk thermal
effects.
nano-particle size (D) [45]: assuming spherical nanoparticles, we
get a diameter D = 4.6 nm, in agreement with the present and pre-
viously reported TEM results [46]. Also the reduction in the inter-
atomic distance and the increase in the Debye Waller factor are
compatible with the values reported for Au nanoparticles of such
size [39,47].
The case of the monometallic Pd is more complicated because in
this case Pd atoms are bound not only to other Pd atoms to form
small nanoparticles as suggested by the reduced Pd–Pd first shell
coordination number but also to oxygen atoms. This indicates a
partial oxidation of the Pd clusters and/or some interaction with
the oxygen atoms of the support. The presence of Pd–O and
The spectra and the values found, for all samples supported on
Al2O3, at the Au L3 edge and at the Pd K-edge, are reported in
Fig. 10, Tables 1 and 3; the ones related to the samples supported
on TiO2 are reported in Fig. 11 and Table 2. The results achieved
show interesting correlations and indicate a different environment
for Au and Pd. In all samples, the first shell coordination number is
lower than 12, the value of a bulk fcc metal, showing the presence
of relatively small nanoparticles.
In the monometallic Au/
c-Al2O3 case, the reduction in the
coordination numbers, N, can be used to estimate the average

C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
231
Fig. 7. Au L3 edge k3 weighted Fourier transforms in the k range 3.5–14 Åꢀ1
.
Fig. 9. Amplitudes of the Fourier transforms of the experimental signals at 20 K and
the best fitting simulated signals (dotted lines).
Table 1
Distances (R), coordination numbers (N), distances and Debye Waller factors (r2
)
achieved for the Au bulk sample, for the Au monometallic catalyst and for the
bimetallic samples on
c-Al2O3 at the Au L3 edge.
Sample
Shell number
N
R (Å)
r2 (Å2)
Au bulk
R-factor = 0.002
1
2
3
4
12(fixed)
6(fixed)
24(fixed)
12(fixed)
2.875(5)
4.067(7)
4.986(7)
5.758(7)
0.0020(5)
0.0026(7)
0.0031(7)
0.0032(7)
Fig. 8. Pd K-edge EXAFS spectra and corresponding k2 weighted Fourier transforms
in the k range 3–12 Åꢀ1
.
Au/
Au (1%)
R-factor = 0.006
c
-Al2O3
1
2
3
4
10.9(4)
5.1(2)
20.3(8)
10.5(4)
2.867(5)
4.061(7)
4.976(7)
5.749(7)
0.0031(8)
0.0043(9)
0.0046(9)
0.0050(9)
Pd–Pd coordinations prevents us from evaluating the Pd cluster
dimensions using the Pd–Pd coordination number as in the case
of Au; as a matter of fact, the value found (NPd–Pd = 5.2) is not the
real coordination number of the Pd atoms nanoparticles but is
the product of the unknown atomic fraction of Pd atoms in this
configuration with the real coordination number.
[Au][Pd]
R-factor = 0.006
Au–Au
Au–Pd
10.2(9)
0.9(2)
2.854(7)
2.793(7)
0.0062(9)
0.0063(9)
[AuPd]
R-factor = 0.005
Au–Au
Au–Pd
5.5(5)
3.6(5)
2.815(8)
2.794(8)
0.0087(9)
0.0071(9)
We can, however, reach a qualitative conclusion by comparing
the results reported in Tables 1 and 3 and by noting in Fig. 8 (lower
panel) that in this monometallic Pd sample, differently from the Au
case, the higher order coordination shell signals are very small
compared with the Pd foil ones: we conclude that we are in the
presence of monometallic Pd nanoparticles of dimensions rather
smaller than in the monometallic Au case, in agreement with
TEM results (see Section 3.1). Such small nanoparticles have an
fcc structure because, as given by the EXAFS data analysis, the ratio
between the second shell distance and the first one R2/R1 is 1.40,
the value expected for an fcc structure.
As clearly visible in Table 3, small and comparable amounts of
oxidized Pd are present in all the catalysts supported on Al2O3,

232
C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
Fig. 10. Amplitudes of the Fourier transforms of the experimental EXAFS signals at 20 K and the best fitting simulated signals (dotted lines) achieved at the Au L3 edge and at
the Pd–K-edge (left panel and right panel, respectively) for the samples supported on Al2O3.
and this can probably exclude any significant effect on the quite
different catalytic activities of the bimetallic [AuPd] and [Au][Pd]
samples.
In the bimetallic samples, supported on c-Al2O3 and TiO2, we
clearly observe (Tables 1–3) the presence of Au–Pd bonds; their
contributions are quite different in [Au][Pd] samples compared to
the [AuPd] ones. In the first case, the number of Pd atoms among
the Au neighbours (and the number of Au atoms among the Pd
neighbours) is rather small (ꢃ10% and ꢃ20% respectively); on the
other hand, in the [AuPd] case, the presence of Au–Pd bonds is
much higher.
In the [Au][Pd]/c-Al2O3 sample, the values of the Au–Au and
Pd–Pd coordination numbers found are compatible with those
found on the monometallic samples but due the co-presence of
Au and Pd atoms as nearest neighbours they cannot be used, as
in the pure monometallic case, to determine the effective size
of the Au and Pd nanoparticles. Only a small amount of hetero-
Fig. 11. Amplitudes of the Fourier transforms of the experimental EXAFS signals at
20 K and the best fitting simulated signals (dotted lines) achieved at the Au L3 edge
for the samples supported on TiO2.
atomic Au–Pd bonds is present in the [Au][Pd]/c-Al2O3 sample
and it can be due to the interactions between the monometallic
nanoparticles. A possible explanation could be that probably the
smaller Pd nanoparticles sit on the surface of the larger Au ones
giving rise to some alloying only in the boundary regions. In
the [Au][Pd] sample supported on TiO2 similar results were
achieved.
Table 2
Results achieved on the bimetallic samples on TiO2 at the Au L3 edge.
Sample
Shell
N
R (Å)
r2 (Å2)
In the case of the [AuPd] samples, we found that the Pd–Pd
coordination numbers are smaller than the Pd–Au ones while the
interatomic distances are comparable one with the other. The ab-
sence of a Pd–Pd distance of 2.74 Å suggests the absence of pure
Pd nanoparticles while the presence of Pd–Pd distances compara-
ble to the Pd–Au ones suggests the presence of a well-defined
Pd–Au phase. Since the Pd:Au molar ratio is equal to 1 in the pres-
ence of a homogenous PdAu alloy, the Pd–Pd and Pd–Au coordina-
tion numbers should be the same but this is not the case because
the number of Pd and Au atoms around the Pd absorber is different
(NPd–Pd – NPd–Au).
[Au][Pd]
R-factor = 0.006
Au–Au
Au–Pd
9.5(9)
0.9(2)
2.850(7)
2.794(7)
0.0068(9)
0.0075(9)
[AuPd]
R-factor = 0.003
Au–Au
Au–Pd
5.0(5)
3.4(5)
2.814(8)
2.793(8)
0.0045(9)
0.0031(9)
Table 3
Results achieved at the Pd K-edge.
Sample
Shell number
N
R (Å)
r2 (Å2)
Let us now consider the situation around Au; in this case, the
Au–Au distances of about 2.82 Å are different from the Au–Pd
and the Pd–Pd ones and are contracted with respect to the Au bulk
value. This, together with the value of the Au–Au coordination
numbers rather higher than the Au–Pd ones, indicates the presence
of small Au-rich cores. From the value of the Au–Au distances [47],
we quote a dimension of about 12 Å corresponding to a nanoparti-
cle of about 55 atoms. From all these considerations, we conclude
that in our [AuPd] samples, bimetallic Au–Pd nanoparticles, com-
posed of an Au rich core surrounded by a AuPd alloyed shell, are
present.
Pd bulk
R-factor = 0.003
Pd–Pd
Pd–Pd
12(fixed)
6(fixed)
2.745(7)
3.885(7)
0.0022(6)
0.0029(6)
Pd/c
-Al2O3
Pd–O
Pd–Pd
Pd–Pd
0.9(2)
5.2(6)
2.5(4)
2.00(1)
2.74(1)
3.84(1)
0.002(1)
0.007(1)
0.009(1)
Pd (0.5%)
R-factor = 0.012
[Au][Pd]
R-factor = 0.013
Pd–O
Pd–Pd
Pd–Au
1.1(2)
4.3(6)
1.2(3)
2.00(1)
2.74(1)
2.79(1)
0.002(1)
0.008(1)
0.012(1)
[AuPd]
R-factor = 0.009
Pd–O
Pd–Pd
Pd–Au
0.7(2)
2.5(5)
3.6(5)
2.01(1)
2.80(1)
2.79(1)
0.003(2)
0.005(2)
0.006(2)

C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
233
A physical reason of such a structure could be attributed to
kinetically more favourable nucleation steps of Au atoms in ace-
tone solution during the melting of the acetone solid matrix, fol-
lowed from a slower aggregation of Au and Pd atoms on pre-
formed Au seeds but also by the high thermodynamic stability at
low temperatures (<500 K) of Au–Pd bimetallic particles having a
Au rich shell [16,48]. It can be related to the different stabilization
properties of acetone towards Pd and Au atoms, as observed also in
alumina-supported samples derived from mononuclear Pd and Au
SMA, where Pd nanoparticles are smaller than those of Au.
Comparing the results reported in Tables 1 and 2, the main
structural difference between the [AuPd] samples on Al2O3 and
TiO2 comes from their very different Debye–Waller factors (r
2).
While the Au–Au and Au–Pd coordination numbers and distances
are quite similar between their Debye–Waller factors, there is a
factor 2. This difference can probably be related to the different
fractions of metallic phase and size distributions found by the
TEM analysis (see Section 3.1).
The different structural results achieved by the EXAFS data
analysis for the [AuPd] and [Au][Pd] samples can be also qualita-
tively observed in the normalized experimental X-ray Absorption
Near Edge Structure (XANES) spectra reported in Figs. 12.
It is well known that the Au L3 edge probes electronic transi-
tions from 2p to 5d states.
Fig. 12b. XANES spectra at the Pd K-edge of the Pd foil and of the catalysts
supported on Al2O3.
The interpretation of the Pd XANES spectra (Fig. 12b) is compli-
cated by the presence of oxygen. Taking into account the two fea-
tures A and B present above the edge jump, it is clearly visible a
reduction in their relative intensities in moving from the Pd foil
to the nanostructured samples. These changes are surely related
to the nanoparticles reduced dimensions and oxidation that pro-
duces electronic and structural effects [50]. The small shifts to
In Fig. 12a, the peak A is known as white line and its intensity is
related to the presence of unoccupied 5d states (d-holes) [49]. In
gold bulk, its presence is due to the s–p–d atomic level hybridiza-
tion, which results in a partial depletion of the atomic filled 5d10
orbitals. In the nanoparticles, the reduced number of Au–Au bonds
reduces the hybridization as well resulting in an increase in the 5d
level occupancy and hence in a decrease in the intensity of the fea-
ture A; this is visible in Fig. 12a comparing the bulk spectrum with
the [Au] monoatomic cluster case. As shown by Liu et al. [49] in
Au–Pd clusters, a further reduction in the intensity of the white
line (feature A) can be due to a charge transfer from Pd to Au.
Therefore, the strong electronic effect observed in Fig. 12a compar-
ing the [Au][Pd] sample to the [AuPd] one can be related to the in-
creased Au–Pd coordination, which causes a further filling of the
Au d band. These two effects observed by looking at the Au L3 edge
XANES spectra when moving from Au bulk to the [AuPd] sample
that is the decrease in the intensity of white line when moving
from the bulk to the Au nanoparticles and a further decrease when
moving to the [AuPd] nanoparticles are the electronic counterpart
of what we have found in the EXAFS spectra.
higher energies of these XANES features observed in the Pd/c-
Al2O3 and [Au][Pd] spectra result from lattice contraction. The
opposite shift observed in the [AuPd] sample is associated with
the large amount of Au present around Pd in this sample, which,
as seen by EXAFS data analysis, results in an increase in the coor-
dination distances. Such trends are in agreement with Ref. [17]
where the experimental spectra of a Pd and an Au–Pd foil are
reported.
3.3. Catalytic behaviour in selective oxidation of benzyl alcohol
The Au–Pd bimetallic systems obtained by above-described two
different procedures, and supported on
c-alumina and titanium
oxide, were tested in the selective oxidation of benzyl alcohol to
Fig. 12a. XANES spectra at the Au L3 edge of the Au foil and of the catalysts supported on Al2O3 (left panel); XANES of the Au foil and of the catalysts supported on TiO2 (right
panel).

234
C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
Table 4
Selective catalytic oxidation of benzyl alcohol over MVS-derived supported Au, Pd and Au–Pd bimetallic catalysts.
Run
Catalyst
Solvent
Time (h)
Conversion (%)
Selectivity (%)
Specific activityb (ꢁ10ꢀ3 sꢀ1
)
2
3
4
1
2
3
Pd/
Au/
[Au][Pd]/
1 wt.% Au–0.5 wt.% Pd
[AuPd]/ -Al2O3
1 wt.% Au–0.5 wt.% Pd
Pd/TiO2 0.5 wt.%
Au/TiO2 1 wt.%
c
c
-Al2O3 0.5 wt.%
-Al2O3 1 wt.%
-Al2O3
Toluene
Toluene
Toluene
3
3
3
4.5
2.5
12
14
46
81
86
45
19
–
9
–
2.1
1.2
5.5
c
4
c
Toluene
3
6
3
3
3
57
98
10
6
84
91
53
45
72
14
8
45
55
28
2
1
2
–
–
26.4
22.7
4.7
2.8
5.8
5
6
7
Toluene
Toluene
Toluene
[Au][Pd]/TiO2
11
1 wt.% Au–0.5 wt.% Pd
[AuPd]/TiO2
1 wt.% Au–0.5 wt.% Pd
8
Toluene
3
12
2
2
2
8
2
2
2
31
87
3
90
93
44
41
74
93
57
46
86
96
8
6
51
48
14
5
38
45
13
4
2
1
5
11
12
2
5
9
14.4
10
62.5
38.5
250.6c
9
10
Pd/
Au/
[AuPd]/
c
c
-Al2O3 0.5 wt.%
-Al2O3 1 wt.%
-Al2O3
–
–
–
2
11a
c
25
75
5
4
22
73
1 wt.% Au–0.5 wt.% Pd
Pd/TiO2 0.5 wt.%
Au/TiO2 1 wt.%
[AuPd]/TiO2
1 wt.% Au–0.5 wt.% Pd
12
13
–
–
–
104.2
77.1
14a
1
–
220.3c
8
Reaction conditions: solvent = 10 ml, benzyl alcohol = 0.3 M, substrate /metal molar ratio = 500, T = 60 °C, P(O2) = 1 atm.
a
Reaction conditions: benzyl alcohol 10 ml (0.088 mol), catalyst 125 mg, T = 100 °C, P(O2) = 5 atm.
b
Mols of products/[(mols of Au + mols of Pd) ꢁ hour].
c
Calculated at t = 2 h.
benzaldehyde as model reaction and their catalytic properties
were compared with those of the corresponding monometallic
systems. Catalytic tests were carried out in mild conditions
(60–100 °C and 0.1–0.5 MPa of molecular oxygen) in both organic
solvent (toluene) and solvent-free conditions as reported in Table
4. The presence of benzyl benzoate and benzyl ether was observed
as by-products, whereas benzoic acid was absent.
components, as well as geometric effect due to changes in lattice
constants induced by a direct interaction between Pd and Au
atoms. As evidenced by XANES data analyses, the increased Au–
Pd coordination in [AuPd] samples leads to a charge transfer from
Pd to Au causing a further filling of the Au–Pd bond, on the other
hand EXAFS data showed an increased coordination distances in
these samples. Conversely, [Au][Pd]/c-Al2O3 and [Au][Pd]/TiO2 sys-
Considering the catalytic tests in toluene solution (runs 1–8),
bimetallic [AuPd] systems deriving from the simultaneous co-con-
densation of Au and Pd vapours and then supported on c-Al2O3 and
TiO2 (see Scheme 1) (run 4 and 8, respectively) showed remarkable
higher catalytic activity (SA = 26.4 ꢁ 10ꢀ3 sꢀ1, run 4, and SA =
14.4 ꢁ 10ꢀ3 sꢀ1, run 8, respectively) than the corresponding mono-
metallic Au and Pd systems (runs 1, 2 and runs 5, 6, respectively). It
is worthy to note that, among the monometallic samples, Au and
Pd systems supported on TiO2 showed slightly higher catalytic
tems obtained from interactions of previously formed Pd and Au
nanoparticles having comparable metal particle size distributions
with respect to the analogous [AuPd] samples (see TEM analyses,
Section 3.1) and a smaller amount of hetero-atomic Au–Pd bonds
(see EXAFS analyses, Section 3.2) did not show any increase in cat-
alytic activity, although higher selectivity, with respect to the
monometallic systems.
The comparison between the catalytic results obtained with
[AuPd] bimetallic catalysts on the different supports shows that,
although HRTEM analyses indicated rather smaller metal particle
activity than analogous samples deposited on c-Al2O3; these differ-
ent behaviours could be related to the different size distributions
sizes in TiO2-supported system, the catalytic activity of [AuPd]/c-
found by the TEM analysis (see Section 3.1).
Al2O3 system (run 4) was higher than that of [AuPd]/TiO2 (run 8).
The above-mentioned [AuPd] bimetallic catalysts were also
significantly more active than bimetallic systems prepared by
MVS technique following the second procedure (see Scheme 2)
(SA = 5.5 ꢁ 10ꢀ3 sꢀ1, run 3 and SA = 5.8 ꢁ 10ꢀ3 sꢀ1, run 7). More-
However, TiO2-supported systems showed higher selectivity than
the analogous sample supported on
c-Al2O3. The difference in
selectivity between the two catalysts appears more evident at
low benzyl alcohol conversion, where [AuPd]/TiO2 sample led to
the formation of lower amount of benzyl benzoate. The mechanism
by which benzyl benzoate are formed was previously described
[6,51] and involves a condensation reaction between the generated
aldehyde and the initial alcohol which is catalysed by the acid–ba-
sic sites present on the catalyst. These data seem to indicate that
the weaker acid sites of titanium oxide-supported catalyst limit
the formation of benzyl benzoate at low benzyl alcohol conversion.
Moreover, it is interesting to note that in both cases the selectivity
over, [AuPd]/c-Al2O3 (run 4) and [AuPd]/TiO2 (run 8) catalysts
showed the highest selectivity towards benzaldehyde (run 4,
Sel. = 91% at 98% of benzyl alcohol conversion and run 8, Sel. =
93% at 87% at benzyl alcohol conversion, respectively). Similarly
to that previously observed [6,17,31], the above-reported catalytic
results together with XAS analyses show that the high catalytic
efficiency, in terms of activity and selectivity, of Au–Pd bimetallic
systems is mainly caused by electronic interaction between the

C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
235
[AuPd]/TiO2
[AuPd]/γ-Al2O3
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
1
2
3
4
1
2
3
4
Cycle
Cycle
Fig. 13. Recycling tests in solvent-less batch runs of [AuPd]/
mmols), catalyst 125 mg, T = 100 °C, P(O2) = 5 atm, t = 8 h).
c-Al2O3 and [AuPd]/TiO2 left side and right side, respectively. (Reaction conditions: benzyl alcohol 10 ml (96
to benzaldehyde increased at high alcohol conversion whereas a
decrease in by-product formation, including benzyl benzoate,
was observed.
the samples prepared by simultaneous vaporization of the two
metals (route 1), nanoparticles composed of an Au rich core sur-
rounded by an Au–Pd alloyed shell have been evidenced. In the
samples prepared by route 2, monometallic Au and Pd nanoparti-
cles and a small presence of hetero-atomic Au–Pd bonds were
found. A possible explanation of the presence of hetero-atomic
bonds could be given by the sitting of the smaller Pd nanoparticles
on the surface of the larger Au ones giving rise to some alloying
only in the boundary regions.
The catalytic activity of the samples obtained by routes 1 and
2, evaluated in the oxidation of benzyl alcohol with molecular
oxygen, as a reference reaction, is quite different; [AuPd]/c-
Al2O3 and [AuPd]/TiO2 samples obtained by route 1 resulted lar-
gely more active and selective than analogous samples prepared
by route 2, pointing out the relevance of the different structural
features of catalysts prepared in different ways in such reaction.
These results are in agreement with previously reported studies
of high catalytic activity in Au–Pd systems with comparable
structure [4,17,31].
Taking into account the high catalytic efficiency of [AuPd]/c-
Al2O3 and [AuPd]/TiO2 systems, they were also tested in the same
reaction under solvent-free conditions comparing their catalytic
efficiency with those of the corresponding monometallic systems.
Similarly to that observed for reactions performed in toluene
solution, the bimetallic [AuPd] systems (run 11 and run 14,
respectively) are significantly more active and selective than the
corresponding monometallic ones (runs 9 and 10, and runs 12 and
13, respectively). The catalysts showed high activity (SA =
250.6 ꢁ 10ꢀ3 sꢀ1 and SA = 220.3 ꢁ 10ꢀ3 sꢀ1, respectively) and high
selectivity towards benzaldehyde at about 75% of alcohol conver-
sion. Again, at low benzyl alcohol conversion, the [AuPd]/TiO2 sam-
ple showed higher selectivity than [AuPd]/c-Al2O3 which led to the
formation of dibenzylether as main by-product that, similarly to
the above-described benzyl benzoate formation, could be catalysed
by the acid–basic sites of the catalyst [51].
Leaching tests, performed by ICP–OES analyses, on [AuPd]/
c
-
In any case, bimetallic catalysts prepared by both routes are
more efficient than the corresponding similarly prepared monome-
tallic Au and Pd samples.
Al2O3 and [AuPd]/TiO2 systems in solvent-free conditions, revealed
gold and palladium amounts in solution lower than 0.2% of the to-
tal available metal. Accounting the very low metal leached during
The bimetallic [AuPd]/c-Al2O3 and [AuPd]/TiO2 supported cata-
the reaction, recycling tests with [AuPd]/
c
-Al2O3 and [AuPd]/TiO2
lysts (route 1) have been reused four batch runs in solvent-free
conditions without loss in catalytic activity and selectivity and
very low amounts of gold and palladium leached in solution
(<0.2% of the total available metal).
The results here reported show an interesting additional exam-
ple of a strong relation between catalytic properties and structural
features in Au–Pd bimetallic catalysts. Moreover, they could sug-
gest the simultaneous vaporization of two metals by metal vapour
synthesis technique, a valuable preparative route to bimetallic me-
tal particles which, differently from other preparation ways, can be
easily deposited in mild conditions on a wide range of inorganic
and organic supports.
catalysts in subsequent solvent-free batch runs were carried out.
As showed in diagrams reported in Fig. 13, both catalysts were
recycled for four batch runs without no significant loss in catalytic
activity and selectivity.
4. Conclusions
Bimetallic Au–Pd nanostructured systems supported on Al2O3
and TiO2 have been conveniently prepared using Au and Pd va-
pours as reagents. Two alternative routes have been adopted: (1)
direct vaporization of the two metals ([AuPd]) and (2) mixing sep-
arately prepared Au and Pd precursors ([Au][Pd]).
TEM analysis has shown the high metal phase dispersion
mainly present in the monometallic samples on TiO2 and on the
monometallic Pd/c-Al2O3 sample. Both bimetallic [AuPd] and
Finally, the co-evaporation of Pd and Au metals allows an inter-
action at the atomic level between Au and Pd vapour as mono- or
oligoatomic species in the absence of additional stabilizers, which
are often expected in most of other synthetic procedures.
[Au][Pd] systems deposited on alumina show metal particles rang-
ing 1.0–6.0 nm in diameter while the analogous systems deposited
on titania present a significant number of metal particles too small
to be detected by TEM (<1.0 nm) as confirmed also by XAS analysis.
XAS investigations on the bimetallic catalysts derived from
routes 1 and 2 have shown quite different structural features. In
Acknowledgment
This work was partially supported by the Italian Ministry of
University and Scientific Research (MIUR) under the PRIN 2008
Program (2008SXASBC_001).

236
C. Evangelisti et al. / Journal of Catalysis 286 (2012) 224–236
[26] U.R. Pillai, E. Sable-Demessie, Appl. Catal. A: Gen. 245 (2003) 103.
References
[27] K. Mori, T. Hara, T. Mizugaki, K. Ebitani, K. Kaneda, J. Am. Chem. Soc. 26 (2004)
10657.
[28] A. Abad, P. Conception, A. Corma, H. Garcia, Angew. Chem. Int. Ed. 44 (2005)
4066.
[29] F.M. Menger, C. Lee, Tetrshedron Lett. 22 (1981) 1655.
[30] D.G. Lee, U.A. Spitzer, J. Org. Chem. 35 (1970) 3589.
[31] N. Dimitratos, A. Villa, D. Wang, F. Porta, D. Su, L. Prati, J. Catal. 244 (2006) 113.
[32] J.S. Bradley, in: G. Schmid (Ed.), Clusters and Colloids. From Theory to
Applications, VCH, Weinheim, 1994, p. 459.
[33] S. Pascarelli, F. D’Acapito, G. Antonioli, A. Balerna, F. Boscherini, R. Cimino, G.
Dalba, P. Fornasini, G. Licheri, C. Meneghini, F. Rocca, S. Mobilio, ESRF Newslett.
23 (1995) 17.
[34] S. Pascarelli, F. Boscherini, F. D’Acapito, J. Hrdy, C. Meneghini, S. Mobilio, J.
Synch. Rad. 3 (1996) 147.
[35] P.A. Lee, P.H. Citrin, P. Eisenberger, B.M. Kincaid, Rev. Mod. Phys. 53 (1981) 769.
[36] G. Ciatto, F. d’Acapito, F. Boscherini, S. Mobilio, J. Synch. Rad. 11 (2004) 278.
[37] C. Evangelisti, P. Raffa, G. Uccello-Barretta, G. Vitulli, L. Bertinetti, G. Martra, J.
Nanosci. Nanotechnol. 11 (2011) 2226–2231.
[1] N. Toshima, T. Yonezawa, New J. Chem. (1998) 1179. and references therein.
[2] J. Ferrando, J. Jellinek, R.L. Johnston, Chem. Rev. 108 (2008) 845.
[3] L. Guzci, Catal. Today 101 (2005) 53. and references therein.
[4] G.J. Hutchings, Chem. Commun. (2008) 1148.
[5] D.I. Enache, J.K. Edwards, P. Landon, B. Solsona-Espriu, A.F. Carley, A.A. Herzing,
M. Watanabe, C.J. Kiely, D.W. Knight, G.J. Hutchings, Science 311 (2007) 362.
[6] D.I. Enache, D. Barker, J.K. Edwards, S.H. Taylor, D.W. Knight, A.F. Carley, G.J.
Hutchings, Catal. Today 122 (2007) 407.
[7] A. Villa, C. Campione, L. Prati, Catal. Lett. 115 (2007) 133.
[8] J.K. Edwards, B.E. Solsona, P. Landon, A.F. Carley, A. Herzing, C.J. Kiely, G.J.
Hutchings, J. Catal. 236 (2005) 69.
[9] F. Menegazzo, P. Burti, M. Signoretto, M. Manzoli, S. Vankova, F. Boccuzzi, F.
Pinna, G. Strukul, J. Catal. 257 (2008) 369.
[10] D. Kumar, M.S. Chen, D.W. Goodman, Catal. Today 123 (2007) 77.
[11] M.S. Chen, D. Kumar, C. Yi, D.W. Goodman, Science 310 (2005) 291.
[12] B. Pawelec, A.M. Venezia, V. La Parola, E. Cano-Serrano, J.M. Campos-Martin,
J.L.G. Fierro, Appl. Surf. Sci. 242 (2005) 380.
[13] V.I. Parvulescu, V. Parvulesco, U. Undruschat, G. Filoti, F.E. Wagner, C. Kubel, R.
Richards, Chem. Eur. J. 12 (2006) 2343.
[14] M. Bonarowska, A. Malinowski, W. Juszczyk, Z. Kapinski, Appl. Catal. B –
Environ. 30 (2001) 187.
[15] H. Okamoto, T.B. Massalki, Bull. Alloy Phase Diagrams 6 (1985) 7.
[16] A.A. Herzing, A.F. Carley, J.K. Edwards, G.J. Hutchings, C.J. Kiely, Chem. Mater.
20 (2008) 1492.
[38] L. Marchese, S. Bordiga, S. Coluccia, G. Martra, A. Zecchina, J. Chem. Soc.
Faraday Trans. 89 (1993) 3483–3489.
[39] T. Comaschi, A. Balerna, S. Mobilio, Phys. Rev. B 77 (2008) 075432.
[40] J.J. Rehr, R.C. Albers, Rev. Mod. Phys. 72 (2000) 621.
[41] A.L. Ankudinov, B. Ravel, J.J. Rehr, S.D. Conradson, Phys. Rev. B 58 (1998) 7565.
[42] M. Newville, B. Ravel, J. Rehr, E. Stern, Y. Yacoby, Physica B 208–209 (1995)
154.
[43] G.G. Li, F. Bridges, C.H. Booth, Phys. Rev. B 52 (1995) 6332.
[44] T. Comaschi, A. Balerna, S. Mobilio, J. Phys.: Condens. Matter 21 (2009) 325404.
[45] A. Balerna, L. Liotta, A. Longo, A. Martorana, C. Meneghini, S. Mobilio, G.
Pipitone, Eur. Phys. J. D 7 (1999) 89.
[46] L.A. Aronica, E. Schiavi, C. Evangelisti, A.M. Caporusso, P. Salvadori, G. Vitulli, L.
Bertinetti, G. Martra, J. Catal. 266 (2) (2009) 250.
[47] A. Balerna, E. Bernieri, P. Picozzi, A. Reale, S. Santucci, E. Burattini, S. Mobilio,
Phys. Rev. B 31 (1985) 5058.
[17] S. Marx, A. Baiker, J. Phys. Chem. C 113 (2009) 6191.
[18] C.H. Chen, L.S. Sarma, J.M. Chen, S.C. Shih, G.R. Wang, G.G. Liu, M.T. Tang, J.F.
Lee, B.J. Hwang, ACS Nano 1 (2007) 114.
[19] M. Harada, K. Asakura, N. Toshima, J. Phys. Chem. 97 (1993) 5103.
[20] D. Wang, A. Villa, F. Porta, L. Prati, D.S. Su, J. Phys. Chem. C 112 (2008) 8617.
[21] Y. Ding, F. Fan, Z. Tian, Z. Lin Wang, J. Am. Chem. Soc. 132 (2010) 12480.
[22] A. Beck, A. Horvath, Z. Schay, G. Stefler, Z. Koppany, I. Sajo, O. Geszti, L. Guczi,
Top. Catal. 44 (2007) 115.
[23] S. Devarajan, P. Bera, S. Sampath, J. Colloid Interf. Sci. 290 (2005) 117.
[24] G. Vitulli, C. Evangelisti, A.M. Caporusso, P. Pertici, N. Panziera, S. Bertozzi, P.
Salvadori, in: B. Corain, G. Schmid, N. Toshima (Eds.), Metal Nanoclusters in
Catalysis and Materials Science. The Issue of Size-Control, Elsevier,
Amsterdam, 2008 (Chapter 32).
[48] H.B. Liu, U. Pal, A. Medina, C. Maldonado, J.A. Ascencio, Phys. Rev. B 71 (2005)
075403.
[49] F. Liu, D. Wechsler, P. Zhang, Chem. Phys. Lett. 461 (2008) 254.
[50] C.-M. Lin, T.-L. Hung, Y.-H. Huang, K.-T. Wu, M.-T. Tang, C.-H. Lee, C.T. Chen,
Y.Y. Chen, Phys. Rev. B 75 (2007) 125426.
[51] G. Li, D.I. Enache, J. Edwards, A.F. Carley, D.W. Knight, G.J. Hutchings, Catal.
Lett. 110 (2006) 7–13.
[25] R.A. Sheldon, J.K. Kochi, Metal-Catalyzed Oxidations of Organic Compounds,
Academic Press, New York, 1981.