PdCu alloy nanoparticles on alumina as selective catalysts for trichloroethylene hydrodechlorination to ethylene

Article abstract of DOI:10.1016/j.apcata.2012.12.019

PdCu bimetallic catalysts supported on alumina were prepared by different common impregnation protocols and by a polyol nanoparticles synthesis route. These materials were studied in order to establish the relationship between structure, particle size, Pd-Cu interaction and catalytic activity/selectivity in the hydrodechlorination of trichloroethylene (TCE) in gas phase. The surface properties and the interaction between Cu and Pd in the bimetallic particles, as well as their catalytic behaviour are strongly influenced by the preparation protocol. Depending on the synthesis, PdCu alloy formation or isolated phases were observed. The presence of isolated Pd particles leads to high ethane selectivity, whereas upon alloy formation or strong interaction between Pd and Cu a higher ethylene yield was obtained. By the nanoparticles polyol synthesis, Pd-Cu alloy formation was obtained, leading to total selectivity to ethylene in the TCE hydrodechlorination reaction. On the Pd monometallic catalysts larger particle size resulted in higher levels of ethylene formation.

Full text of DOI:10.1016/j.apcata.2012.12.019

Applied Catalysis A: General 453 (2013) 130–141
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
PdCu alloy nanoparticles on alumina as selective catalysts for
trichloroethylene hydrodechlorination to ethylene
Beteley T. Mesheshaa,b, Noelia Barrabésa,∗, Jordi Llorcac, Anton Dafinovb, Francisco
Medinab, Karin Föttingera,∗
a Institute of Materials Chemistry, Vienna University of Technology, Getreidemarkt 9/BC/01, 1060 Vienna, Austria
b Departament d’Enginyeria Química, Universitat Rovira i Virgili, Av/Països Catalans 26, 43007 Tarragona, Spain
c Institut de Tècniques Energètiques and Centre de Recerca en NanoEnginyeria, Universitat Politècnica de Catalunya, Diagonal 647, 08028 Barcelona, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
PdCu bimetallic catalysts supported on alumina were prepared by different common impregnation pro-
tocols and by a polyol nanoparticles synthesis route. These materials were studied in order to establish
the relationship between structure, particle size, Pd–Cu interaction and catalytic activity/selectivity in
the hydrodechlorination of trichloroethylene (TCE) in gas phase. The surface properties and the inter-
action between Cu and Pd in the bimetallic particles, as well as their catalytic behaviour are strongly
influenced by the preparation protocol. Depending on the synthesis, PdCu alloy formation or isolated
phases were observed. The presence of isolated Pd particles leads to high ethane selectivity, whereas
upon alloy formation or strong interaction between Pd and Cu a higher ethylene yield was obtained. By
the nanoparticles polyol synthesis, Pd-Cu alloy formation was obtained, leading to total selectivity to
ethylene in the TCE hydrodechlorination reaction. On the Pd monometallic catalysts larger particle size
resulted in higher levels of ethylene formation.
Received 13 August 2012
Received in revised form
13 November 2012
Accepted 11 December 2012
Available online xxx
Keywords:
PdCu bimetallic catalyst
Nanoparticles
Alloy
High resolution TEM
Temperature-programmed reduction
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
leading to valuable olefins [9]. Since then, a number of transition
metal additives have been studied such as Pt–Cu [10–13], Pt–Sn
Chlorinated organic compounds, despite their wide spread
industrial applications, have been categorized as high priority envi-
ronmental pollutants. They represent hazardous ozone depleting
compounds with high carcinogenic activity. Catalytic hydrodechlo-
rination (HDC) is a promising technology to transform them
into more useful, but less harmful hydrochlorocompounds [1–4],
obtaining valuable products that could be applied in other pro-
cesses. Trichloroethylene (TCE) is one of the harmful pollutants
for humans that must be removed from the environment. The gas
phase catalytic hydrodechlorination of TCE can lead to a wide vari-
ety of products with added value. Mostly, noble metal catalysts
have been used for this reaction converting the TCE to the fully
hydrogenated product ethane [5–8].
[14] or Pd–Ag [15,16]. Several explanations have been proposed
to explain the change in the selectivity towards the olefin using
bimetallic systems, related with an electronic modification of the
noble metal [12], or structure-sensitivity [1,2,17] versus structure-
insensitivity [4,18]. In general, the high ability of the noble metals
for the hydrogenolysis of C Cl bonds results in the fully hydro-
genated products. In the case of non-noble metal like Cu, the
dechlorination step occurs selectively leading to the olefin, in the
case of TCE to ethylene.
Previous works [19,20] showed the requirement of alloy for-
mation of the metals in order to obtain highly active and selective
catalysts towards ethylene, avoiding deactivation. Barrabés et al.
[19,21,22] showed that the metal–metal interaction strongly
affected the catalytic behaviour in the TCE HDC reaction on cop-
per hydrotalcite-derived catalysts. In this work, the influence of
the particle size and of the metal amount was not studied. By
studying the effect of the particles size and the metals ratio, while
maintaining complete alloy formation and minimizing the noble
metal amount required, optimization of such catalysts for the
TCE hydrodechlorination process should be achieved. The inter-
action of the metals and the dispersion on the surface is directly
related with the preparation method employed. Most of the meth-
ods used to prepare bimetallic catalysts on alumina are based on
In the last decade, it was discovered that the modification of
the noble metal catalysts by the addition of a second metal modi-
fies the catalytic performance in the hydrodechlorination reactions
∗
Corresponding author at: Institute of Materials Chemistry, Vienna University of
Technology, Getreidemarkt 9 BC, 1060 Vienna, Austria. Tel.: +43 1 58801 165110;
fax: +43 1 58801 16599.
E-mail addresses: noe.barrabes@gmail.com (N. Barrabés),
karin.foettinger@tuwien.ac.at (K. Föttinger).
0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2012.12.019

B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
131
co-impregnation, sequential impregnations, co-exchange with the
salts of the two metals or redox reaction [23]. An alternative route
employed in the last years in catalysis, in order to control the sur-
face chemistry of the catalysts, is the use of metal nanoparticles
(NPs) prepared in solution. NPs have unique physical and chemi-
cal properties different from bulk materials due to the reduction
of particle size. Extremely high surface areas and therefore high
fractions of metal atoms occupying positions on or close to the sur-
face as well as the presence of various types of active sites make
such nanoparticles ideal candidates as key components of catalytic
systems [24].
The present work aims at the optimization of bimetallic catalysts
for the TCE hydrodechlorination by comparing several prepara-
tion protocols including the synthesis of bimetal nanoparticles in
solution. The study of the structure and surface chemistry of the
catalysts achieved by various preparation procedures and the influ-
ence on the catalytic behaviour will allow the elucidation of the
optimal parameters.
In the first step, Pd-Cu bimetallic catalysts supported on alumina
have been prepared by different impregnation procedures from dis-
solved metal salts as precursors. Controlled surface reaction (CSR),
sequential impregnation (SI), sequential reduction impregnation
(SRI) and co-impregnation (COI) protocols have been employed.
From the different preparations, we expected different degrees of
metal–metal interaction, such as isolated metals and alloy forma-
tion. Then, in order to achieve control over the particle size and
composition and to optimize the metals amount and ratio in the cat-
alysts, Pd and Pd-Cu alloy nanoparticles were prepared following
a polyol-method using 1-hexadecylamine, as reported previously
by Meshesha et al. [25]. This method leads to uniform, nanosized
and spherical metal particles, which were subsequently supported
on alumina. All catalysts were characterized by several techniques
in order to study the structure and surface properties and tested in
the gas phase TCE hydrodechlorination reaction.
nitrate salt under hydrogen flow (CSR). The COI catalysts were pre-
pared by direct impregnation with a solution containing the two
metals (Pd(NO3)2·2H2O and Cu(NO3)2·3H2O). Then the samples
were dried, calcined at 300 ◦C for 2 h, and finally reduced at 400 ◦C
under hydrogen flow for 1 h.
2.1.2. Synthesis of PdCu/Al2O3np nanoparticles catalysts
A modified procedure of the previously described protocol [25]
was adopted in order to prepare 1-hexadecylamine (HDA) capped
Pd and Pd-Cu nanoparticles. Analytically pure Pd acetyl acetonate
(Pd(AcAc)2) and Cu(NO3)2·2H2O were used as the starting mate-
rial without further purification. In a typical synthesis procedure,
100 ml of ethylene glycol solution of the respective metal salt(s)
was mixed with an appropriate amount of HDA in a round-bottom
flask equipped with a magnetic stirrer, refluxing device and ther-
mometer. The mixture was mixed thoroughly in an ice bath to
prevent rapid reduction of the noble metals in ethylene glycol and
HDA solutions at room temperature. The solution was then heated
rapidly to 140 ◦C and kept there for 12 h. The nanoparticles were
synthesized maintaining HDA/metal molar ratio of three. Nanopar-
ticles with different Pd/Cu molar ratios were synthesized.
A solution of Pd and Pd-Cu nanoparticles in toluene was used
to impregnate the particles onto the ␥-alumina support. After
evaporation of the solvent in a rotavapor the catalyst was dried
overnight at 100 ◦C, calcined in air at 350 ◦C for 2 h and reduced
by flowing hydrogen at 300 ◦C for 2 h prior to reaction. The cal-
cination and reduction of the catalysts were performed at the
lowest temperature required to remove the organic species from
the catalyst surface and fully reduce particles. The materials synthe-
sized from the nanoparticle precursors are labelled as Pd1/Al2O3np,
Pd1Cu1/Al2O3np (Cu/Pd molar ratio 1.64), Pd1Cu5/Al2O3np (Cu/Pd
molar ratio 8.33), and Pd1.6Cu1/Al2O3np (Cu/Pd molar ratio 1). For
example Pd1Cu1/Al2O3np contains 1 wt% Pd and 1 wt% Cu metal
impregnated on alumina.
For direct comparison, an additional bimetallic catalyst was
prepared from the same precursor salts as the nanoparticles and
calcined and reduced under identical conditions. This material
contains 1 wt% of both, Pd and Cu, and was prepared by simul-
taneous impregnation of Cu(NO3)2 and Pd(AcAc)2 salts followed
by calcination at 350 ◦C. This reference catalyst is referred to as
Pd1Cu1/Al2O3ref.
2. Experimental
2.1. Catalyst synthesis
Two different approaches were utilized to prepare ␥-alumina
supported bimetallic Pd-Cu catalysts: (1) materials prepared
by impregnation from a metal salt precursor solution will be
referred to as “impregnation catalysts”; (2) catalysts prepared from
(bi)metallic nanoparticles by a polyol method followed by loading
the nanoparticles onto the alumina support will be referred to as
“nanoparticles catalysts” in this work. Details on the preparation
are described in the following.
2.2. Characterization techniques
The chemical composition of the samples was determined by
ICP-OES with a PerkinElmer Plasma 400 instrument.
The structure and crystallinity of the supported catalysts after
calcination was investigated by X-ray diffraction carried out on a
Siemens D5000 diffractometer (Bragg–Bentano for focusing geom-
etry and vertical ꢀ–ꢀ goniometer) fitted with a grazing incident (ω:
0.52◦) attachment for thin film analysis and scintillation counter as
a detector.
Selected catalysts were characterized with TEM and HRTEM
techniques. TEM was operated at 100 kV (JEOL JEM-2000 EX II).
High-Resolution Transmission Electron Microscopy (HRTEM) was
carried out at 200 kV with a JEOL 2010F instrument equipped with
a LaB6 source. The point-to-point resolution of the microscope was
0.20 nm. Samples were deposited on holey-carbon-coated grids
from alcohol suspensions. More than 100 individual particles were
used in each sample for particle size determination.
2.1.1. PdCu/Al2O3i catalysts synthesis by impregnation
Four different impregnation-based procedures were applied to
prepare bimetallic Pd2Cu1.2/Al2O3i (2 wt% Pd, 1.2 wt% Cu, molar
ratio 1:1) catalysts, which are illustrated in Scheme 1: SI, SRI, CSR,
and COI on ␥-alumina (Sasol Germany GmbH, Puralox SBA 200, spe-
cific surface area 250 m2/g). Furthermore, monometallic catalysts
were prepared as well. In the first step, the monometallic 2 wt%
Pd/Al2O3 was prepared by impregnation using an aqueous solu-
tion of the precursor salt, Pd(NO3)2·2H2O. This step consists of a
cationic exchange between the metallic complex and the alumina
surface in an aqueous solution at pH 11 obtained by the addition of
ammonia. After evaporation of the aqueous phase, the catalyst was
dried at 120 ◦C and calcined in air at 450 ◦C.
The temperature programmed reduction (TPR) studies were
performed in a ThermoFinnigan (TPORD 110) apparatus equipped
with a thermal conductivity detector (TCD). Before the TPR analy-
sis, the samples (around 100 mg) were dried under flowing argon at
120 ◦C for 24 h. The analysis was carried out using 3% H2/Ar flowing
After that, the monometallic Pd catalyst was directly impreg-
nated with the copper salt (SI), or reduced under hydrogen flow
at 300 ◦C for 1 h before impregnation with the copper salt solution
(SRI), or introduced to an aqueous solution containing a copper

132
B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
Scheme 1. Synthesis protocols for Pd2Cu1.2/Al2O3i impregnation catalysts.
at 20 ml/min by heating from room temperature to 900 ◦C with a
ramp of 10 ◦C min−1
Pd particles over the alumina support with particles size of around
2 nm. Lattice fringe images recorded of individual Pd particles
exhibited spacing of metallic Pd at 2.3 A corresponding to Pd(1 1 1).
.
˚
The Fourier transform infrared (FT-IR) spectra were recorded
on a Bruker IFS 28 spectrometer equipped with an MCT detector
at a resolution of 4 cm−1. The measurement cell is connected to
a vacuum system working in the 10−6 mbar range and allowing
in situ pre-treatments of samples. The catalysts were pressed into
self-supporting wafers and were reduced in 500 mbar of pure H2 at
300 ◦C. After keeping the catalyst in hydrogen atmosphere at that
temperature for 30 min, the cell was evacuated for another 30 min
at the reduction temperature. CO adsorption measurements were
carried out at 20 ◦C in 5 mbar of pure CO static atmosphere, followed
by evacuation of the gas phase CO at 20 ◦C.
A general view of the COI sample is depicted in Fig. 1(A). The metal
particle size distribution is broader compared to that of pure Pd,
and the main diameter is around 3.5 nm. Fourier transforms anal-
ysis (FFT) of high resolution images of individual particles show
˚
spots at 2.2 A, which is lower than the Pd(1 1 1) spacing at 2.3,
suggesting the possibility of alloying between Pd and Cu in the
sample. Both Pd and Cu have fcc crystal structure, and the lattice
parameter of Pd is larger than that of Cu. For that reason, alloying
Cu with Pd results in a reduction of the (1 1 1) lattice parame-
ter.
The sample CSR presents a rather broad size distribution his-
togram (Fig. 2). The mean particle size is centred at about 5.8 nm.
The lattice-fringe image/FFT of a representative particle (Fig. 2(B-
2.3. Catalytic activity and selectivity measurements
˚
The catalysts were tested for gas-phase hydrodechlorination
reaction of TCE using a continuous fixed-bed quartz glass reac-
tor setup described previously [19]. The gas feed at a total flow of
60 ml/min was obtained by flowing an inert gas (He) and hydrogen
through a saturator kept at 25 ◦C by a thermostat, containing TCE in
liquid phase. The TCE partial pressure was 92 mbar and the TCE flow
was 5.5 ml/min. The molar ratio of H2/TCE was maintained at 3.5,
near the stoichiometric amount required to obtain ethylene. The
gas flows were adjusted by mass flow controllers (Brooks Instru-
ment) and introduced into the reactor, which is placed in a furnace
coupled with a temperature controlling system. The outlet of the
reactor is connected by a six-way valve to a gas chromatograph
(HP 5890 series II, HP Poraplot column, FID). For all measurements
100 mg of catalyst were charged in the reactor. Catalytic conversion
and selectivity were calculated by analyzing the calibrated peak
areas of TCE and the respective products.
a)) presents spots at 2.3 A characteristic of pure Pd metal (1 1 1),
and their relative position indicates that the Pd crystallite is ori-
ented along the [1 1 0] crystallographic direction. Besides pure Pd
particles, the sample also contains Pd-Cu alloy particles. Fig. 1(B-b)
˚
shows a high resolution image with a spot at 2.2 A being identi-
fied and the relative position of the spots in the FT pattern reveals
again the [1 1 0] crystallographic orientation. On the other hand,
the SRI sample (Fig. 1(C)) exhibits well-dispersed metal particles
on the support with particles size of 4.0 nm. Spots in the FT pat-
˚
tern (Fig. 1(C-a)) at 2.2 and 1.3 A are significantly lower than for
Pd(1 1 1) and Pd(2 2 0) values, respectively, indicating the alloy for-
mation between Pd and Cu. The angle at 90◦ between both spots
indicates that the crystallite is oriented along the [1 1 2] crystallo-
graphic direction. Fig. 1(C-b) also shows indicative images of alloy
formation. The FT image of the particle marked by an arrow shows
˚
spots at 2.2 and 1.8 A, again at lower values than for Pd(1 1 1) and
Pd(2 0 0), indicative of Pd-Cu alloying.
˚
Fig. 1(C-c) shows spots in the FT pattern at 2.6 A that can only
3. Results and discussion
be ascribed to (1 1 0) planes of Cu3Pd (tetragonal structure). Conse-
quently, for this sample the existence of alloying between Pd and Cu
is clear. The same is observed for the SI sample (Fig. 1(D)) that con-
3.1. Characterization results
˚
tains metal particles of about 5.0 in diameter with spots at 2.2 A.
3.1.1. PdCu/Al2O3 impregnation catalysts
It is interesting to note that no core–shell structures have been
observed in any of the samples, that is, no copper shells develop
on the Pd particles. In contrast, alloying of Pd with Cu creates true
alloy particles with sharp lattice edges. However, HRTEM cannot
HRTEM analysis was performed in order to study the particle
size distribution (Fig. 1) and alloy formation of the metallic parti-
cles obtained from different preparation protocols. HRTEM results
of monometallic Pd catalyst (not shown) contain well-dispersed

B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
133
Fig. 1. HRTEM micrographs of Pd2Cu1.2/Al2O3i catalysts prepared by impregnation (A: COI, B: CSR, C: SRI and D: SI).
reveal the presence of surface layers enriched in Cu of a few atom
layers thickness.
For obtaining more insights into the metal–metal interaction,
TPR analysis was performed (Fig. 3). The CuO/Al2O3 reduction is
characterized by a peak 280 ◦C [26], related with the reduction of
the CuO phase as well as the reduction of Cu+ to Cu0. The reduction
of monometallic Pd catalysts occurs at lower temperatures, around
100 ◦C. The addition of copper (bimetallic catalysts) modifies the
reduction profile. Previous studies demonstrated that the forma-
tion of a Pd-Cu alloy arises from the interaction between copper and
palladium oxides. Fig. 3 shows that in all the samples a new reduc-
tion peak appears between 125 and 200 ◦C, that can be attributed
to the reduction of Cu and Pd from the mixed oxide, PdxCuyO [26].
For the COI preparation method the interaction between palladium
and copper also occurs with a reduction peak at 145 ◦C, but the
presence of isolated palladium particles leading to peaks between
50 and 100 ◦C is additionally observed in the TPR profile. When the
CSR protocol was used, low intensity and broad peaks are observed,
which is related with the fact that the metals are already reduced
during the preparation [27], thus no reduction of isolated Pd par-
ticles is observed, leading to the lower hydrogen consumption. SI
and SRI samples show reduction peaks at 130 ◦C and 160 ◦C related
with PdCu alloy, as observed by HRTEM.
In order to study the surface composition and accessibility of
the metals at the surface, CO adsorption on Cu-Pd catalysts was
examined by FT-IR spectroscopy (Fig. 4). CO adsorbed at room
Fig. 2. Particle size histograms.

134
B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
Table 1
Physical characteristics of Pd-Cu nanoparticle catalysts.
Catalyst
Metal loading
Pd (wt%)
Particle sizea (nm)
Cu (wt%)
Pd1/Al2O3np
0.8
0.9
0.9
1.6
0.7
–
10
15
–
15
–
Pd1Cu1/Al2O3np
Pd1Cu5/Al2O3np
Pd1.6Cu1/Al2O3np
Pd1Cu1/Al2O3ref
1.1
4.7
1.0
0.8
a
Determined by TEM.
CO adsorbed on Pd2/Al2O3i. This range is related with CO species
multiply-bonded on neighbouring Pd0 sites, indicating no signifi-
cant dilution of the Pd surface by Cu in the COI sample. In addition,
the signal of CO on Cu2+ is rather high as compared to the rela-
tive intensities observed on the other bimetallic catalysts, which
hints to more particles with oxidized Cu species at the surface.
Therefore, it can be concluded that isolated Pd and Cu particles are
present on the COI sample in addition to alloyed particles, which is
in agreement with the TPR results.
The catalysts prepared by CSR, SRI and SI exhibit main CO bands
at 2142, 2115, 2065 and 1978 cm−1. CO adsorption on metallic
copper leads to a single CO vibration located between 2080 and
2120 cm−1 depending on the crystal plane and the nature of the
support [27]. In this way the band around 2115 cm−1 is attributed
to Cu0-CO species, while the band at 2065 cm−1 is likely related
with carbonyls on zero-valent Pd. Previous studies assigned this
band to CO linearly bonded on Pd-Cu alloy [28]. The band of bridge-
bonded CO at 1978 cm−1 suggests that Pd metal ensembles may
also exist at the surface. The vibrational band visible at 2142 cm−1
can be assigned to CO adsorbed on oxidized Cu+. This would indi-
cate the presence of oxidized Cu species in addition to reduced Cu,
even though the TPR measurements revealed that most of the Cu is
reduced at this temperature (300 ◦C). An alternative assignment is
adsorption to Cu species alloyed with Pd. Borovkov et al. [10] inves-
tigated CO adsorption on Pt-Cu catalysts by FTIR spectroscopy and
attributed a band at 2135 cm−1 not to CO associated with oxidized
Cu species, but with metallic Cu modified by Pt. The blue-shift by
about 40 cm−1 of the frequency of CO on Cu was attributed to the
modification of the electronic properties upon alloying with Pt [10].
The assignment of CO on Cu species is commonly based on the sta-
bility of the CO adsorption complexes against evacuation. The band
at 2142 cm−1 was stable against evacuation, which can be an indi-
cation for CO adsorbed on Cu+. However, a higher stability of the
CO on Cu that is alloyed with Pd was obtained by DFT calculations
for the Pd-Cu bimetallic system by Lopez and Nørskov [29]. Such a
stabilization effect can therefore also be relevant for the present Pd-
Cu alloy samples. In light of the potential alternative assignment of
the band at 2142 cm−1 the Cu species should be fully reduced under
these conditions, which is indeed in good agreement with the TPR
measurements.
Fig. 3. TPR profiles of Pd2Cu1.2/Al2O3i catalysts prepared by impregnation: (a) CSR,
(b) COI, (c) SRI, and (d) SI.
temperature leads to the appearance of several bands between
1920 and 2150 cm−1 related with CO on palladium and copper
particles. The region from 1750 to 1900 cm−1 is assigned to CO
adsorption on three-fold hollow Pd(1 1 1) sites [28]. The region
from 1900 to 2000 cm−1 is attributed to two-fold bridge CO adsorp-
tion. The bands between 2000 and 2100 cm−1 are assigned to
linearly bonded CO on Pd and Cu0. Several bands at higher fre-
quencies >2100 cm−1 which appear additionally in the spectra of
the bimetallic catalysts are attributed to CO on oxidized Cu species
(Cu+, Cu2+).
The spectrum (a) is characteristic for CO on metallic Pd. In
all bimetallic catalysts, clear changes in the available adsorption
sites are visible. The COI catalyst (spectrum (b)) shows a broad
CO band between 1700 and 2000 cm−1, similar to the spectrum of
Overall, the FTIR studies of the bimetallic catalyst showed that
the Pd-Cu bimetallic catalysts prepared by redox reaction (SRI, SI
and CSR) exhibited a preferential deposition of Cu on the surface
of Pd breaking up larger ensembles that can adsorb CO on hollow
sites. In the work of Melendrez et al. [30], in contrast, Cu was likely
deposited at the low coordination sites (edges and corners).
3.1.2. PdCu/Al2O3 nanoparticles catalysts
The PdCu nanoparticles impregnated on ␥-alumina have been
characterized by different techniques. The molar compositions of
the metals were investigated using ICP spectroscopy and the results
are summarized in Table 1.
As expected a little deviation from the originally employed
metal amount was observed. The decrease in weight percentage
Fig. 4. FTIR-CO spectra (5 mbar CO, 20 ◦C) of bimetallic catalysts prepared by
impregnation: (a) Pd2/Al2O3i, (b) COI, (c) CSR, (d) SRI and (e) SI. Samples were
reduced at 300 ◦C prior to CO adsorption.

B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
135
Fig. 5. XRD patterns of (a) ␥-Al2O3, (b) Pd1/Al2O3np, (c)Pd1Cu1/Al2O3np, (d)Pd1Cu5/Al2O3np (e) Pd1Cu1/Al2O3ref, and (f) Pd1.6Cu1/Al2O3np nanoparticle catalysts after
calcination. Reflections of PdO (ꢀ) and Pd (ꢀ) are indicated.
of the active components arises during the preparation and sep-
aration of the nanoparticles from the polyol solution. In general,
the extraction method isolated more than 90% of the nanoparticles
from the polyol solution, showing the efficiency of the preparation
methodology.
probably due to an overlap with the diffraction peak of alumina
at around 40◦. However, there is a clear distortion of the alumina
diffraction line around 37.8◦ probably due to the influence of CuO
species.
Two
representatives
of
the
bimetallic
catalysts,
The presence of crystalline Pd or/and Cu containing phases
within calcined Pd-Cu/Al2O3np samples was investigated by XRD.
The diffractograms are shown in Fig. 5. The pure alumina support
shows the typical pattern of transition alumina, assigned to ␥-
Al2O3 (JCPDS-01-079-1558). This pattern is evident in all samples.
Broad and diffuse diffraction lines of the support sample indicate
poor crystallinity of ␥-alumina. In agreement with the previously
reported results, it is confirmed that low to moderate total metal
loading on high surface area Al2O3 support calcined gives highly
dispersed Cu and Pd species which are difficult to detect by X-ray
diffraction [26]. The Pd and Cu loading influences the intensity of
the alumina peaks between 25◦ and 60◦, due to possible diffusion
of Cu and Pd atoms into the ␥-alumina lattice [11].
In order to clarify the existence and state of Pd and Cu species,
the XRD profiles were magnified between 30◦ and 50◦ as shown
in Fig. 5. In the calcined Pd1/Al2O3np catalyst, PdO reflections
(JCPDS-00-043-1024) at around 34.4◦, 42◦ and 55◦ were observed.
This species is also present in all bimetallic catalysts. Additional
reflections of metallic Pd (JCPDS-01-087-0641) were observed in
Pd1.6Cu1/Al2O3np at 40.5◦ and 47.3◦. The characteristic reflec-
tions of CuO were not clearly visible in all the bimetallic catalysts
Pd1.6Cu1/Al2O3np and Pd1Cu1/Al2O3np reduced at 300 ◦C
were characterized using TEM and HRTEM integrated with Fourier
transform analysis (Fig. 6). The TEM images for the Pd1/Al2O3np
catalyst after reduction (not shown) show palladium particles
uniformly dispersed on the support with an average particle size of
10 nm. The Pd nanoparticles prepared in solution before impreg-
nation onto the alumina were characterized by a smaller particle
size with an average diameter of 4 nm. The increase in particle size
for Pd1/Al2O3np is probably due to the heat treatments applied in
the calcination and reduction.
Pd1.6Cu1/Al2O3np sample is comprised of particles with differ-
ent particle sizes. As a general trend, large particles of about 100 nm
in diameter are present, while most of the particles are smaller,
varying between 10 and 20 nm. Fig. 6a and b shows bright-field
TEM images at low magnification, where both types of particles
are easily recognized. High-resolution TEM (HRTEM) cannot be
performed over the large particles because they are too thick for
lattice-fringe analysis. Fig. 6c and d therefore correspond to HRTEM
images of the small particles. The particle depicted in Fig. 6c shows
˚
a lattice spacing at 2.62 A, as deduced from Fourier Transform
(FT) analysis (inset). This corresponds to (1 1 0) planes of Cu3Pd

136
B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
Fig. 6. HRTEM images: Pd1.6Cu1/Al2O3np (a–d) and Pd1Cu1/Al2O3np (e–h).
alloy. The structure of pure Pd and Cu is cubic (Fm-3m). The same
symmetry is found for the CuPd alloy, but the Cu3Pd alloy is tetrago-
nal (P4mm) and, for that reason, can be easily recognized. Fig. 6d
shows a HRTEM image of another particle. In this case, the particle
shows two domains, indicated in the figure with labels “a” and “b”
with the FT analysis of both domains attached. Domain “a” shows
The coexistence of Pd and Cu3Pd phases is commonly encoun-
tered in the sample. A low-magnification view of Pd1Cu1/Al2O3np
catalyst is shown in Fig. 6e. In this case, the particle size distribu-
tion spans from 10 to 25 nm in diameter, with the mean particle size
centred at about 14–15 nm. A careful lattice fringe analysis of these
particles carried out with HRTEM is shown in Fig. 6f–h. In Fig. 6f, the
FT analysis shows the occurrence of spots at 1.94 and 2.24 A, which
correspond to (2 0 0) and (1 1 1) planes of pure Pd metal, respec-
tively. Another representative particle is shown in Fig. 7g, along
˚
˚
˚
spots at 2.62 and 2.76 A. Again, crystallographic planes at 2.62 A can
˚
only be ascribed to (1 1 0) planes of Cu3Pd alloy. The spots at 2.76 A
in domains “a” and “b” correspond to (1 1 0) planes of pure Pd.

B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
137
Fig. 8. Catalytic activity of impregnation catalysts in the TCE hydrodechlorination
reaction at 300 ◦C after 180 min (ꢁ: TCE conversion (%); ꢀ: ethane selectivity (%) and
᭹: ethylene selectivity (%)). Reaction conditions: total flow 60 ml/min, TCE partial
pressure 92 mbar, H2/TCE molar ratio 3.5, amount of catalyst 100 mg.
Fig. 7. TPR profile of (a) Pd1/Al2O3np, (b) Pd1Cu1/Al2O3ref, (c) Pd1Cu1/Al2O3np, (d)
Pd1.6Cu1/Al2O3np, (e) Pd1Cu5/Al2O3np.
could depict that the surface is enriched with Pd species. The broad
peak located >150 ◦C can be assigned to the reduction of isolated
CuO species. The TPR profile for Pd1Cu5/Al2O3np in general shows
a uniform broad peak with a maximum around 197 ◦C. No peak
related to the reduction of isolated PdO species at lower tempera-
tures was detected for this catalyst. The reduction peak at 197 ◦C is
related to reduction of CuO species. These species are likely affected
by the presence of Pd as they are reduced at lower temperatures.
The TPR profile for Pd1.6Cu1/Al2O3np catalyst did not reveal a uni-
form reduction peak but reveals the existence of different species.
The profile consists of distinct peaks located around 96 ◦C, 123 ◦C
and 208 ◦C, respectively. The first peak at 96 ◦C is associated to the
reduction of isolated PdO species, similar as Pd1/Al2O3np. The main
reduction peak located at around 123 ◦C is assigned tentatively to
Pd-Cu oxide species that are in close vicinity, whereas the high
temperature reduction peak at 208 ◦C is related to reduction of iso-
lated CuO species as observed also for the Pd1Cu5/Al2O3np catalyst.
The existence of different reduction peaks shows the heterogeneity
of the surface species in the bimetallic nanoparticle catalysts. The
presence of different TPR reduction peaks indicate that the bimetal-
lic catalyst prepared using the polyol synthesis protocol did not
result in uniform alloy formation. Instead, isolated Cu and Pd and
Pd-Cu alloys exist, whereas the Pd1Cu1/Al2O3ref catalyst shows a
rather uniform reduction peak associated with PdCu species.
˚
with its FT analysis. In this case, several spots at 2.24 and 2.62 A
are recognized in the FT image, which are ascribed to (1 1 1) crys-
tallographic planes of Pd and (1 1 0) planes of Cu3Pd, respectively.
The coexistence of these two phases in the particle induces struc-
tural stress, which is observed as poor-defined lattice fringes in
the HRTEM image. Finally, Fig. 6h shows an individual Cu3Pd alloy
˚
particle exhibiting lattice fringes at 2.62 A, corresponding to (1 1 0)
crystallographic planes.
The degree of Pd–Cu interactions was further investigated by
H2-temperature programmed reduction analysis as shown in Fig. 7.
The consumption of hydrogen and reducibility of supported metal
catalysts depends on the metal-support interaction, metal–metal
interaction and particle size distribution.
The TPR profile for Pd1/Al2O3np catalyst showed a main reduc-
tion peak at around 107 ◦C. Different TPR profiles were found for
the bimetallic catalysts due to a different degree of interaction, and
the respective features are assigned tentatively. As shown in the
Fig. 7, the Pd1Cu1/Al2O3ref catalyst from salt precursors showed
a main reduction peak at around of 151 ◦C that can be assigned
to the reduction of CuO, promoted by nearby Pd. A small reduc-
tion peak at around 274 ◦C is assigned to the reduction of isolated
copper species. The main reduction peak for this catalyst shifted
to high temperature when compared to the Pd1/Al2O3np catalyst.
The decrease in the reduction temperature of CuO in bimetallic cat-
alysts, induced by the presence of palladium, indicates that a close
proximity between copper and palladium species was achieved.
The consumption of hydrogen of the Pd1Cu5/Al2O3np catalyst is
much higher than that of Pd1/Al2O3np. From the ICP analysis shown
in Table 1, it is found that Pd1 and Pd1Cu5 contain similar Pd loading
but differ in Cu nominal loading. Converting the weight % into Cu/Pd
mole ratio, it was found that copper is around 8-fold higher than
Pd. This implies the catalyst is highly enriched in Cu and/or Pd-Cu
alloys. Consequently, when Cu is promoted by introducing small
amount of a noble metal like Pd, its hydrogen uptake capacity is
expected to improve by a higher magnitude as shown in the figure.
The sample Pd1.6Cu1/Al2O3np and likely also Pd1/Al2O3np contain
already reduced Pd as observed in XRD (presence of reflections of
Pd and PdO).
3.2. Catalytic activity
3.2.1. PdCu/Al2O3 impregnation catalysts
All materials were tested in the gas phase catalytic hydrodechlo-
rination of TCE at 300 ◦C using a molar ratio TCE/H2 of 3.5. The
results are shown in Fig. 8. Whereas the Pd monometallic cata-
lyst (Pd2/Al2O3i) leads to high conversion of TCE to mainly ethane
(not shown), the use of bimetallic catalysts resulted in strongly
increased selectivity towards ethylene, in agreement with [9,10].
A different catalytic behaviour is obtained depending on the syn-
thesis protocol employed for bimetallic Pd1.2Cu1/Al2O3i catalysts.
The catalysts showed conversion levels between 50% and 75% under
these conditions. CSR and SI protocols led to the higher conversion
(70% and 75%, respectively), whereas COI and SRI protocols exhib-
ited somewhat lower conversion (64% and 49%, respectively). In
terms of selectivity, all the catalysts showed selectivity values to
ethylene higher than 75%. However, the selectivity to ethylene was
improved even more for the SRI and SI materials (between 90 and
95%). This can be explained taking into account the close interac-
tion between Pd and Cu, leading to mostly bimetallic active sites
The TPR profile of Pd1Cu1/Al2O3np is characterized by broad
peaks around 107 ◦C and 125 ◦C, respectively. The peak around
107 ◦C is probably due to the reduction of isolated PdO species. The
main reduction peak located at maximum of 125 ◦C is likely related
to the reduction of Pd-Cu oxide species that are in close vicinity.
The lower reduction temperature for the bimetallic PdCu phase

138
B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
the conversion decreased by 66% on Pd1Cu1/Al2O3ref. The
Pd1Cu1/Al2O3ref catalyst prepared by impregnation from metal
salts for comparison showed the highest rate of deactivation
in comparison to the nanoparticles catalysts. In addition, the
degree of deactivation on the bimetallic nanoparticles catalysts
Pd1.6Cu1/Al2O3np and Pd1Cu1/Al2O3np was less as compared to
the monometallic Pd1/Al2O3np catalyst.
Deactivation is an interesting issue in HDCl. It is well known that
during the hydrodechlorination reaction of chlorinated compounds
using supported metal catalysts, the chloride species formed dur-
ing the reaction can interact with the metallic sites of the catalyst
leading to deactivation. This was a main topic of a recent work
of us on Pd-Cu catalysts [19]. In the previous work Cl poisoning
of the Cu was identified as the origin of deactivation of bimetal-
lic Pd-Cu catalysts. The deactivation was less severe the better the
interaction between Pd and Cu was, and did not play a role on
well-alloyed samples exhibiting constant catalytic performance.
Flushing in H2 at elevated temperatures (i.e. 300 ◦C) fully restored
the conversion and selectivity of deactivated samples. Cl poisoning
on samples with incomplete alloying was suggested to be due to
less efficient removal of Cl from poisoned Cu species. In the present
work the conversion decreased only at the very beginning from the
first measurement after 5 min TOS to the conversion determined
after 30 min on stream, after this initial period a stable conver-
sion was observed. However, due to the rather large particle size of
the np samples (with mean particle sizes between 10 and 20 nm)
TEM measurements did not reveal significant changes in particle
size of used catalysts therefore giving no clear indication whether
sintering could occur. Based on our previous work [19] and in agree-
ment with other studies dedicated to deactivation (e.g. [31]) the
most likely cause of the initial activity loss followed by a period of
constant activity is Cl poisoning. Only the sample Pd1Cu1/Al2O3ref
shows a continuous decrease in conversion, which is probably due
to a less efficient interaction between Cu and Pd. The most pro-
nounced loss of activity is observed on the Pd1Cu5 catalyst. This can
be explained by a large fraction of isolated copper surface without
nearby noble metal particles, which are able to efficiently regen-
erate Cu from chloride species on the surface deposited during
the dechlorination reaction. When an optimum composition of the
metals is reached (such as Pd1Cu1) leading to extensive alloy for-
mation and avoiding isolated metal particles a high stability with
time of stream is obtained.
Considering the degree of conversion, an increase of the Pd
amount resulted in a higher activity (Pd1.6Cu1/Al2O3np catalysts).
All the catalysts containing 1 wt% Pd showed almost similar initial
conversion. However, the Pd amount does not affect the stability
of the catalysts. Pd1.6Cu1/Al2O3np and Pd1Cu1/Al2O3np catalysts
showed the best performance in terms of stability.
The main products formed on the nanoparticle catalysts
were ethylene, ethane, chloroethylene and also some traces of
dichloroethylenes (<0.5%). The selectivity evolution with reaction
time is shown in Fig. 10.
Unexpectedly, monometallic Pd1/Al2O3np exhibits a high selec-
tivity towards ethylene as shown in Fig. 10a. While the bimetallic
samples exhibited a very stable selectivity (Fig. 10b–e), the ethylene
selectivity on the monometallic Pd1/Al2O3np catalyst increased
with time on stream (Fig. 10a). 63% ethylene and 28% selectivity
to ethane was obtained at 5 min time of reaction. Ethylene selec-
tivity increases during time on stream to 87%, and ethane formation
declined to 10% of ethane formation after 180 min.
This result indicates that the use of nanoparticles synthesis
changes completely the catalytic behaviour of the monometallic
catalyst Pd1/Al2O3np with respect to the selectivity, as compared
to the Pd2/Al2O3i catalyst (Section 3.2.1). An increasing produc-
tion of ethylene was found in comparison with that previously
reported for Pd impregnation catalysts, where full hydrogenation
Fig. 9. TCE conversion at 300 ◦C over monometallic Pd and PdCu nanoparticles cat-
alysts with time on stream. The reaction was carried out at a total flow of 60 ml/min
with a TCE partial pressure of 92 mbar and a H2/TCE molar ratio of 3.5. 100 mg of
catalyst were employed.
as observed by TPR and HRTEM. The bimetallic sites are respon-
sible for the selective HDC of TCE to ethylene. Characterization
results reported in Section 3.1.1 have shown that SRI and SI synthe-
sis protocols lead to narrow particles size distribution around 5 nm
and Pd-Cu alloy formation. In contrast, the COI and CSR synthe-
ses resulted in a less homogeneous particle size distribution. CSR
and COI catalysts exhibited isolated monometallic phases beside
bimetallic particles. Thus, in the catalytic reaction a higher selec-
tivity to ethane was observed attributed to pure Pd particles. High
selectivity to ethylene is related to the presence of bimetallic par-
ticles.
As described in Section 3.1.1 alloy formation was confirmed
by HRTEM and TPR. The Pd-Cu alloy formation created the active
sites responsible for ethylene formation. The SI catalyst showed
higher activity than SRI. In the TPR, major reduction processes were
observed at 160 ◦C for the SRI sample, and around 135 ◦C for the SI
material. This temperature difference in reduction might be related
with a different alloy composition depending on the synthesis pro-
tocol.
3.2.2. PdCu/Al2O3 nanoparticles catalysts
In order to investigate whether the catalytic performance in the
gas phase HDC of TCE depends on the Pd/Cu molar ratio, degree
of interaction and particles size, PdCu nanoparticles deposited on
Al2O3 were tested for this reaction. Fig. 9 shows the catalytic activ-
ities of the catalysts vs. time on stream.
The monometallic Pd1/Al2O3np catalyst showed an initial con-
version of 55%, and after 180 min on stream the activity decreased
to 20%. The catalytic performance of Pd-Cu bimetallic catalysts,
with different Pd/Cu weight loading, showed similar trends as
the monometallic Pd catalyst (Fig. 9), with also some deactivation
going on with time on stream. The highest initial catalytic activity
was observed for Pd1.6Cu1/Al2O3np catalysts showing an initial
conversion of 70% mainly due to the higher Pd metal loading. In
contrast, Pd1Cu1/Al2O3np, Pd1Cu1/Al2O3ref, and Pd1Cu5/Al2O3np
exhibited moderate catalytic activities of around 52%, 48% and
46%, respectively. After 3 h of reaction the conversion decreased
over all the catalysts, but to different extents. The decrease in
conversion was similar on the nanoparticle catalysts with various
metal ratios, amounting to around 35% for Pd1Cu1/Al2O3np,
37% for Pd1Cu5/Al2O3np, and 38% for Pd1.6Cu1/Al2O3np, while

B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
139
Fig. 10. Selectivity vs. time on stream (min) at 300 ◦C for Pd1/Al2O3np (a), Pd1Cu1/Al2O3ref (b), Pd1Cu1/Al2O3np (c), Pd1Cu5/Al2O3np (d), Pd1.6Cu1/Al2O3np (e) catalysts.
The reaction was carried out at a total flow of 60 ml/min with a TCE partial pressure of 92 mbar and a H2/TCE molar ratio of 3.5. 100 mg of catalyst were employed.
was observed [6,8]. Possible reasons are differences in particle size
and morphology, as well as different reaction conditions such as
temperature and H2:TCE ratio. Nevertheless, the bimetallic cata-
lysts always showed higher ethylene selectivity than monometallic
Pd catalysts. The ratio between the metals (Pd/Cu) and the
total metal loading seem to affect the catalytic behaviour. As is
shown in Fig. 10, the Pd-Cu bimetallic catalysts with similar Pd
contents but different copper:palladium molar ratios exhibit dif-
ferent selectivity. Pd1Cu1/Al2O3np (Fig. 10c) and Pd1Cu5/Al2O3np
catalysts (Fig. 10d) showed total ethylene formation at initial
reaction time (5 min), whereas Pd1Cu1/Al2O3ref (Fig. 10b) and
Pd1.6Cu1/Al2O3np (Fig. 10e) produce ethylene at selectivities
around 80% and 90%, respectively, with additional formation of
ethane.
also increases accompanied by the initial loss of catalytic activity.
The increase in ethylene selectivity is most pronounced on the
monometallic Pd sample. A likely reason for the change in selec-
tivity can be coke formation which reduces the hydrogen supply
on the Pd. This was recently demonstrated by Srebowata et al.
[31] who detected a PdCx phase by XRD on a used Pd catalyst after
HDCl of 1,2-dichloroethane. The improved selectivity could also
be related with a dilution of the noble metal by the presence of Cl
species on the surface decreasing the concentration of hydrogen
available at the surface. A similar effect but less pronounced can
be valid for the Pd-rich Pd1.6Cu1/Al2O3np catalyst.
3.2.3. Comparison of the catalytic behaviour obtained using
different synthesis protocols
As the time of reaction increases the selectivity to ethylene over
Pd1/Al2O3np, Pd1Cu1/Al2O3ref and Pd1.6Cu1/Al2O3np catalysts
The use of bimetallic nanoparticles preparation in solution in
comparison with the conventional impregnation synthesis routes

140
B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
Table 2
Comparison of the catalytic performance resultsa over impregnated and nanoparticles catalysts.
Catalyst
Mean particle size (nm)
(Bi)metal state
TCE conversion (%)
Selectivity (%)
Ethylene
Ethane
Chloroethylene
Pd2/Al2O3i
2.0
5.0
5.0
5.8
3.5
10
15
–
Pd
84
75
49
70
64
55
52
46
69
48
32
90
95
80
75
68
10
5
20
25
28
0
0
10
15
0
0
0
0
0
9
0
0
0
3
Pd2Cu1.2/Al2O3i SI
Pd2Cu1.2/Al2O3i SRI
Pd2Cu1.2/Al2O3i CSR
Pd2Cu1.2/Al2O3i COI
Pd1/Al2O3np
Pd1Cu1/Al2O3np
Pd1Cu5/Al2O3np
Pd1.6Cu1/Al2O3np
Pd1Cu1/Al2O3ref
Cu3Pd alloy
PdCu alloy
Pd, PdCu alloy
Pd, Cu, PdCu alloy
Pd
Cu3Pd alloy
–
Pd, Cu, Cu3Pd alloy
–
63
100
100
90
15
–
81
a
Reaction conditions: 300 ◦C reaction temperature, total flow 60 ml/min, TCE partial pressure 92 mbar, H2/TCE molar ratio 3.5, amount of catalyst 100 mg.
to prepare Pd/Al2O3 and PdCu/Al2O3 catalysts leads to different cat-
alytic behaviour in the TCE hydrodechlorination reaction. Table 2
shows a summary of the catalytic results obtained over both types
of catalysts.
The common impregnation catalysts always show higher TCE
conversion than the nanoparticles catalyst, although the differ-
ences in the metal content and particle sizes must be taken into
account.
explained taking into consideration that small metal particles (par-
ticles with high interaction with the support) exhibit a high amount
of atoms with low coordination that are very active for HDC reac-
tion, but with a poor selectivity to olefins. The low-coordinated sites
present on small nanoparticles bind ethylene more strongly than
terrace sites and simultaneously such low-coordinated sites exhibit
lower hydrogenation barriers than terrace sites. In order to increase
the selectivity to ethylene it is necessary to decrease or deactivate
these low-coordinated atoms. It seems that the introduction of cop-
per forming Pd-Cu alloy plays an important role bringing about this
effect. As described previously, hydrogenolysis of a C Cl bond may
require larger Pt ensembles and the addition of Cu decreases the
number of sufficiently large Pt islands suitable for the bond scis-
sion [12]. A dilution of Pd ensembles by Cu can also be deduced from
the FTIR spectra of adsorbed CO, as described in Section 3.1.1. Even
though the literature concerning the structure-sensitivity of C Cl
bond dissociation on Group VIII metals is contradictory [10], here
we can see clearly the effect of the particle size on the selectivity
of Pd monometallic catalysts towards ethylene formation, despite
the metal loading difference.
The monometallic Pd2/Al2O3i catalyst from salt precursor
showed very high selectivity to ethane (68%). Such catalytic
behaviour of noble metals was reported previously owing to the
noble metals ability to hydrogenate the C C double bond. How-
ever, the monometallic Pd1/Al2O3np catalyst prepared from the
nanoparticle precursor, showed a higher selectivity to ethylene
(62%) and a lower conversion (55%). The main differences between
the impregnation and the nanoparticle catalyst are a higher metal
loading in the former catalyst (2 wt% instead of 0.8 wt%) and a dif-
ferent mean particle size of 2 nm and 10 nm, respectively. Another
possible explanation could be differences in morphology.
The introduction of copper increases the selectivity to ethyl-
ene in all the catalysts. The best catalytic performance within
the nanoparticles catalysts with high and stable activity and
selectivity to ethylene was obtained on Pd1Cu1/Al2O3np and
Pd1.6Cu1/Al2O3np. Within the series of catalysts prepared by
impregnation from salt precursor solutions, the sequential impreg-
nation and sequential reduction impregnation routes led to
materials with 90% and 95% ethylene selectivity at initial conver-
sion levels of 75% and 49%, respectively.
PdCu alloyed nanoparticles, mainly of the composition Cu3Pd,
were obtained both by the nanoparticle precursor synthesis and
by the impregnation route. Depending on the synthesis protocol,
isolated Pd and/or Cu particles were present in some samples in
addition to the PdCu alloy, which affected activity and selectivity.
Upon PdCu alloy formation Cu modifies the electronic properties
of Pd [32,33] and has thereby been proposed to weaken ethylene
adsorption. A lower bonding strength of ethylene on Pt-Cu/SiO2
catalysts was observed in microcalorimetric measurements in [34].
Consequently, ethylene would desorb more easily from Pd in PdCu
than hydrogenate further to ethane.
4. Conclusion
Different preparation procedures for Pd-Cu bimetallic catalysts
supported on ␥-alumina were compared in this work, including
conventional impregnation and preparation of bimetallic nanopar-
ticles from a polyol solution. The catalyst synthesis protocol affects
the catalytic properties in the hydrodechlorination of TCE. On
all materials, better ethylene selectivity was achieved than over
monometallic Pd catalysts. Differences in catalyst performance
were connected with differences in metal–metal interaction and
particle size as main factors governing the selectivity.
The highest selectivity to olefins with values up to 100% was
reached on catalysts where Pd-Cu alloy formation was detected by
HRTEM and TPR. Catalysts exhibiting isolated monometallic phases
in addition to bimetallic particles, such as the Pd1Cu1.2/Al2O3i
catalysts prepared by CSR and COI route, yielded more ethane,
which was attributed to the reactivity of the Pd particles for total
hydrogenation. SRI and SI catalysts composed of uniform bimetal-
lic particles with a narrow particle size distribution around 5 nm
reached ethylene selectivities higher than 90%.
Aiming to have control over the particle size and interac-
tion between the metals, bimetallic Pd-Cu nanoparticles with
different Pd:Cu ratios were prepared by a polyol method using 1-
hexadecylamine as stabilizing agent. The catalytic activity mainly
depended on the amount of the noble metal, with the highest activ-
ity on Pd1.6Cu1/Al2O3np, and the selectivity on the amount of
copper. In general, increasing copper content resulted in an increase
in ethylene selectivity at comparable TCE conversion.
In this sense, isolated and small Pd particles are very active
for the hydrogenation of TCE leading to ethane as the main prod-
uct. Emphasizing on the monometallic Pd catalysts prepared by
the two different protocols, the effect of the interaction of the Pd
nanoparticles with the support Al2O3 plays an important role in
the selectivity to ethane. The monometallic Pd catalyst obtained
from metal nanoparticles is characterized by larger particle size
(10 nm) and hence possesses lower interaction with the support
as compared to the monometallic Pd catalyst obtained from salt
impregnation with small particles of around 2 nm. As a result, the
former catalyst yields higher selectivity to ethylene. This could be

B.T. Meshesha et al. / Applied Catalysis A: General 453 (2013) 130–141
141
Another issue that was clearly observed is the influence of the
Pd particle size on the performance of monometallic Pd catalysts.
Larger Pd particles resulted in higher selectivity to ethylene, as
observed by comparing Pd2/Al2O3i with 2 nm mean particle size
with Pd1/Al2O3np exhibiting ∼10 nm sized particles. This may be
related to the stronger interaction of small particles with the sup-
port and the presence of a higher concentration of low coordinated
sites with a high activity for ethane formation. On small nanopar-
ticles more low-coordinated sites are present which bind ethylene
more strongly and exhibit lower hydrogenation barriers than ter-
race sites. Alloying with Cu can modify the electronic properties
of Pd thus weakening ethylene adsorption. In addition hydrogen
supply is reduced by diluting the Pd surface. Overall, on large Pd
particles and on PdCu alloy particles, ethylene formation is strongly
enhanced.
[10] V.Y. Borovkov, D.R. Luebke, V.I. Kovalchuk, J.L. d’Itri, J. Phys. Chem. B 107 (2003)
5568–5574.
[11] D.R. Luebke, L.S. Vadlamannati, V.I. Kovalchuk, J.L. d’Itri, Appl. Catal. B: Environ.
35 (2002) 211.
[12] L.S. Vadlamannati, D.R. Luebke, V.I. Kovalchuk, J.L. d’Itri, Stud. Surf. Sci. Catal.
130 (2000) 233–238.
[13] L.S. Vadlamannati, V.I. Kovalchuk, J.L. d’Itri, Catal. Lett. 58 (1999) 173–178.
[14] W.D. Rhodes, K. Lazar, V.I. Kovalchuk, J.L. d’Itri, J. Catal. 211 (2002) 173–182.
[15] B. Heinrichs, J.P. Schoebrechts, J.P. Pirard, J. Catal. 200 (2001) 309–320.
[16] P.P. Kulkarni, V.I. Kovalchuk, J.L. d’Itri, Appl. Catal. B: Environ. 36 (2002)
299–309.
[17] M.A. Aramendia, V. Borau, I.M. Garcia, C. Jimenez, F. Lafont, A. Marinas, J.M.
Marinas, F.J. Urbano, J. Mol. Catal. A: Chem. 184 (2002) 237–245.
[18] F.H. Ribeiro, C.A. Gerken, G.A. Somorjai, C.S. Kellner, G.W. Coulston, L.E. Manzer,
L. Abrams, Catal. Lett. 45 (1997) 149–153.
[19] N. Barrabes, D. Cornado, K. Foettinger, A. Dafinov, J. Llorca, F. Medina, G. Rup-
prechter, J. Catal. 263 (2009) 239–246.
[20] V.I. Avdeev, V.I. Kovalchuk, G.M. Zhidomirov, J.L. d’Itri, Surf. Sci. 583 (2005)
46–59.
[21] N. Barrabes, K. Föttinger, A. Dafinov, F. Medina, G. Rupprechter, J. Llorca, J.E.
Sueiras, Appl. Catal. B: Environ. 87 (2009) 84–91.
[22] N. Barrabes, K. Föttinger, J. Llorca, A. Dalinov, F. Medina, J. Sa, C. Hardacre, G.
Rupprechter, J. Phys. Chem. C 114 (2010) 17675–17682.
[23] J. Batista, A. Pintar, J.P. Gomilsek, A. Kodre, F. Bornette, Appl. Catal. A: Gen. 217
(2001) 55–68.
[24] Y. Wang, J. Ren, K. Deng, L. Gui, Y. Tang, Chem. Mater. 12 (2000) 1622–1627.
[25] B.T. Meshesha, N. Barrabes, F. Medina, J.E. Sueiras, Polyol mediated synthesis &
characterization of Cu nanoparticles: effect of 1-hexadecylamine as stabilizing
agent, World Scientific and Engineering Academy and Society, Athens, 2009.
[26] J. Batista, A. Pintar, D. Mandrino, M. Jenko, V. Martin, Appl. Catal. A: Gen. 206
(2001) 113–124.
Acknowledgements
This work was financially supported by the MEC Spain, the
Austrian Exchange Service (Acciones Integradas, ES 05/2009), the
Generalitat de Catalunya (ICREA ACADEMIA award) and the Marie
Curie project PIEF-GA-2009-236741ENVIROCATHYDRO.
References
[27] J. Sa, H. Vinek, Appl. Catal. B: Environ. 57 (2005) 247–256.
[28] T. Wadayama, H. Osano, H. Yoshida, S. Oda, N. Todoroki, Appl. Surf. Sci. 254
(2008) 5380–5384.
[1] B. Coq, G. Ferrat, F. Figueras, J. Catal. 101 (1986) 434–445.
[2] Z. Karpinski, K. Early, J.L. d’Itri, J. Catal. 164 (1996) 378–386.
[3] L.E. Manzer, M.J. Nappa, Appl. Catal. A: Gen. 221 (2001) 267–274.
[4] F.H. Ribeiro, C.A. Gerken, G. Rupprechter, G.A. Somorjai, C.S. Kellner, G.W. Coul-
ston, L.E. Manzer, L. Abrams, J. Catal. 176 (1998) 352–357.
[5] J.W. Bae, E.D. Park, J.S. Lee, K.H. Lee, Y.G. Kim, S.H. Yeon, B.H. Sung, Appl. Catal.
A: Gen. 217 (2001) 79–89.
[6] F.D. Kopinke, K. Mackenzie, R. Kohler, Appl. Catal. B: Environ. 44 (2003) 15–24.
[7] M. Legawiec-Jarzyna, A. Srebowata, W. Juszczyk, Z. Karpinski, React. Kinet.
Catal. Lett. 87 (2006) 291–296.
[29] N. Lopez, J.K. Nørskov, Surf. Sci. 477 (2001) 59–75.
[30] R. Melendrez, G. Del Angel, V. Bertin, M.A. Valenzuela, J. Barbier, J. Mol. Catal.
A: Chem. 157 (2000) 143–149.
´
[31] A. Sre˛bowata, W. Lisowski, J.W. Sobczak, Z. Karpin´ ski, Catal. Today 175 (2011)
576–584.
[32] M. Fernández-García, J.A. Anderson, G.L. Haller, J. Phys. Chem. 100 (1996)
16247–16254.
[33] N. Mårtensson, R. Nyholm, H. Calén, J. Hedman, B. Johansson, Phys. Rev. B 24
(1981) 1725.
[34] L. Li, X. Wang, A. Wang, J. Shen, T. Zhang, Thermochim. Acta 494 (2009)
99–103.
[8] S. Ordonez, H. Sastre, F.V. Diez, Appl. Catal. B: Environ. 40 (2003) 119–130.
[9] V.I. Kovalchuk, J.L. d’Itri, Appl. Catal. A: Gen. 271 (2004) 13–25.