A promoting effect of dilution of Pd sites due to gold surface segregation under reaction conditions on supported Pd-Au catalysts for the selective hydrogenation of 1,5-cyclooctadiene

Article abstract of DOI:10.1016/j.cattod.2015.07.022

Restructuration of AuPd/CeO₂ catalysts is observed during 1,5-cyclooctadiene hydrogenation, with gold surface segregation in which gold acts as a diluent of Pd surface sites. The composition of the catalyst influences the catalytic behavior, an

Full text of DOI:10.1016/j.cattod.2015.07.022

Catalysis Today 259 (2015) 213–221
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
A promoting effect of dilution of Pd sites due to gold surface
segregation under reaction conditions on supported Pd–Au catalysts
for the selective hydrogenation of 1,5-cyclooctadiene
Patricia Concepcióna,∗, Saray Garcíaa, Juan Carlos Hernández-Garridob,
Jose Juan Calvinob, Avelino Cormaa,c,∗∗
a Instituto de Tecnología Química CSIC-UPV, Universidad Politécnica de Valencia, 46022 Valencia, Spain
b Departamento de Ciencia de los Materiales, Ingeniería Metalúrgica y Química Inorgánica, Universidad de Cádiz, Facultad de Ciencias,
Campus Río San Pedro, Puerto Real, 11510 Cádiz, Spain
c King Fahd University of Petroleum and Minerals, P. O. Box 989, Dhahran 31261, Saudi Arabia
a r t i c l e i n f o
a b s t r a c t
Article history:
Restructuration of AuPd/CeO2 catalysts is observed during 1,5-cyclooctadiene hydrogenation, with gold
surface segregation in which gold acts as a diluent of Pd surface sites. The composition of the catalyst
influences the catalytic behavior, and an optimum is observed at a Pd/Au molar ratio of 0.13, as deter-
mined by X-EDS STEM. Pd–Au/CeO2 shows higher hydrogenation rates (11.2 × 10−5 mol/g s) versus the
unpromoted Pd/CeO2 catalysts (1.7 × 10−5 mol/g s), and the Pd/TiO2 (7.5 × 10−5 mol/g s) catalysts used
as conventional hydrogenation catalysts. Since the active sites are generated under reaction conditions,
characterization of the catalysts after quenching the reaction at specific reaction times was required.
© 2015 Elsevier B.V. All rights reserved.
Received 30 March 2015
Received in revised form 14 July 2015
Accepted 16 July 2015
Available online 11 August 2015
Keywords:
AuPd
CeO2
Selective hydrogenation
Cyclooctadiene
XPS
Site isolation
1. Introduction
characterization techniques, X-ray photoelectron spectroscopy
(XPS) is a very powerful spectroscopic tool which gives chemical
A rational way to design solid catalysts will go through the
fundamental understanding of the generation of the catalytic
active sites and their interaction with reactant and products
under reaction conditions. For achieving that, catalysts are many
times characterized before being subjected to the reaction and
the obtained results are correlated with catalytic activity and
selectivity. Following that methodology it is sometimes difficult to
establish good structure–reactivity correlations since, as has been
reported, structural and chemical modifications of the catalysts
can occur under reaction conditions [1–4]. Moreover these changes
could in certain cases affect only the catalyst surface [5]. Therefore,
knowledge of the catalyst under reaction conditions (operando
spectroscopy) or at least as close as possible to reaction conditions,
becomes very important. From the point of view of catalyst
information from the 6–9 nm outermost surface layers, while
infrared spectroscopy (FTIR) of adsorbed molecule, gives specific
information about the interaction of molecules with active sites
located at the external catalyst surface. Thus, the combination of
both spectroscopies will be helpful for better understanding the
catalytic behavior at the surface of solids.
In the case of catalysts based on metal nanoparticles, several
parameters like particle size and morphology, metal–support inter-
actions, chemical composition i.e. oxidation states, presence of
promoters, etc. are of paramount importance for catalysis [6–9].
Moreover, the addition of a second metal to produce bimetal-
lic nanoparticles, has opened more possibilities for engineering
catalysts with enhanced activity and selectivity. Important techno-
logical areas, including catalytic reforming [10], pollution control,
alcohol oxidation [11–13], selective hydrogenation [8,14–16],
electro-catalysis in fuel cells [17], among other, are based on
bimetallic systems. It has been reported that the synthesis method
can direct the formation of bimetallic materials with different final
structures like core–shell or alloy formation with markedly differ-
ent catalytic behavior [1,18]. In the case of bimetallic materials,
and even more so when the size of the materials are in the range
∗
Corresponding author.
Corresponding author at: Instituto de Tecnología Química CSIC-UPV, Universi-
∗∗
dad Politécnica de Valencia, 46022 Valencia, Spain.
E-mail addresses: pconcepc@upvnet.upv.es (P. Concepción), acorma@itq.upv.es
(A. Corma).
http://dx.doi.org/10.1016/j.cattod.2015.07.022
0920-5861/© 2015 Elsevier B.V. All rights reserved.

214
P. Concepción et al. / Catalysis Today 259 (2015) 213–221
of the nanometer scale, structural and compositional changes can
produce large differences in the catalytic behavior of the nanopar-
ticles. If this is so, one should take into account that metal surface
composition and size and shape of those bimetallic catalysts can
change during the reaction, being the final catalyst different from
the starting one.
In the present study we will show that modifications in the
structure of monometallic Pd and Au and bimetallic AuPd cata-
lysts occur during the hydrogenation of 1,5-cyclooctadiene, with
an impact on the catalytic activity. This has been found by
stopping the reaction during the catalytic process and follow-
ing the possible change by XPS, HRTEM and FTIR spectroscopic
characterization.
on CeO2 and 0.13 and 0.23 wt% Pd supported on TiO2 were pre-
pared by a deposition–precipitation method by mixing CeO2 or
TiO2 powders with appropriate amounts of aqueous solutions of
Pd(NO3)2·xH2O (Aldrich, 99.9%) at a fixed pH (7.6 0.3) adjusted
with 0.35 wt% NH4OH solution. The mixtures were stirred for 12 h
and statically aged for 6 h. Then they were washed with deionized
water, filtered and dried at 120 ◦C in air for 12 h.
A series of Pd–Au/CeO2 catalysts with of 0.2, 0.17, 0.10 and
0.05 wt% of Pd loading and 0.36 wt% Au were prepared by the
deposition–precipitation method of Au/CeO2 with Pd(NO3)2. Typ-
ically, the air dried Au/CeO2 sample was added to an aqueous
solutions of Pd(NO3)2 at a fixed pH (7.6 0.3), adjusted with
0.35 wt% NH4OH. The mixture was stirred for 12 h and statically
aged for 6 h. The samples were then washed, filtered, and dried at
120 ◦C for 12 h. The above samples dried in air at 120 ◦C for 12 h,
are referred to as prepared samples and labeled as X% Pd–Y% AuCe,
where X, Y represent the weight percent of each element as deter-
mined by ICP. The H2 pre-reduced sample was obtained by further
treatment of the air dried sample in flowing H2 at 200 ◦C for 2 h.
Once reduced, the samples were kept under N2 atmosphere until
further use in catalytic tests or spectroscopic characterization tech-
niques. On the other hand, the as prepared Pd–Au/CeO2 samples
were studied immediately after preparation.
The selective hydrogenation of dialkenes and/or alkynes to
monoalkenes, continues to attract attention due to practical impor-
tance and fundamental interest [16,19,20]. While Pd exhibits the
highest activity toward alkenes in dialkenes or alkynes hydroge-
nations, the selectivity toward monoalkenes with Pd significantly
decreases when working at high conversions of dialkenes or
alkynes [20]. Thus, considerable attention has been paid to fac-
tors which can improve the selectivity of Pd catalysts for the above
hydrogenation reactions. The reaction selectivity during dialkenes
hydrogenation to alkenes is determined by the rates of alkene
desorption and hydrogenation. Taking into account that the inter-
action of adsorbed molecules depends on the morphology of the
surface i.e. the coordination number of the surface atoms, which
depends among others, on the metal particle size, addition of
other metals, etc., metal particle size effects [21,22] as well as the
addition of other metals to Pd [16,23,24] have been studied as
one way to improve its catalytic performance. Among bimetallic
Pd catalysts, the Pd–Au system has been extensively investigated
for selective hydrogenations [16,19,20,23–26] as well as for oxi-
dation reactions [11,18,27,28]. The enhanced catalytic behavior
observed in the bimetallic AuPd system has usually been explained
due to formation of a core/shell structure [29,30], alloy forma-
tion [24,31,32], selective poisoning of the Pd surface [16,33,34] and
changes in metal dispersion and morphology of the particle [26].
On the other hand, the impact of metal–support interactions on
the catalytic behavior of supported metal particles is well known
for several processes. In this sense, ceria has been reported as a very
active support for oxidation as well as for hydrogenation reactions
[35–37]. Interestingly, very few studies [38] have been reported
for the hydrogenation of cyclodienes or dialkenes using ceria sup-
ported catalysts, therefore our interest has been driven on this
reaction, as a test reaction for studying monometallic and bimetallic
Au–Pd/CeO2 catalysts.
2.2. Catalyst characterization
Gold and palladium loadings were determined by inductively
coupled plasma atomic emission spectroscopy (ICP-AES).
X-ray photoelectron spectra of the catalysts were recorded with
a SPECS spectrometer equipped with a Phoibos 150MCD multichan-
nel analyzer, using MgK␣ (1253.6 eV) irradiation. The spectra were
recorded with an X-ray power of 50 mW in order to avoid fotore-
duction of the gold species. The residual pressure in the analytical
chamber was maintained below 10−9 mbar during data acquisi-
tion. The binding energies of Au 4f and Pd 3d were corrected for
surface charging by referencing them to the energy of Ce3d5/2 v1
peak of the ceria support set at 882.178 eV in case of CeO2 sup-
ported samples and to the Ti2p3/2 XPS peak set at 458.5 eV in case
of TiO2 supported samples. Peak intensities have been calculated
after nonlinear Shirley-type background subtraction and corrected
by the transmission function of the spectrometer. CasaXPS software
has been used for spectra deconvolution.
FTIR spectra have been collected with a FTS-40A BioRad spec-
trometer equipped with a DTGS detector (4 cm−1 resolution, 32
scans). An IR cell allowing in situ treatments under controlled
atmospheres and temperatures from −176 ◦C to 500 ◦C has been
connected to a vacuum system with gas dosing facility. Self-
supporting pellets (ca. 10 mg cm−2) were prepared from the sample
powders and treated at 50 ◦C under dynamic vacuum (10−4 mbar).
After activation, the samples were cooled down to −176 ◦C fol-
lowed by CO dosing at increasing pressures (0.4–8.5 mbar) and the
IR spectrum recorded after each dosage.
2. Experimental
2.1. Catalyst preparation
Electron microscopy studies (HRTEM, HAADF-STEM, XEDS-
STEM) were performed in a JEOL2010F microscope. The structural
resolution in HRTEM mode of this microscope is 0.19 nm. The
crystallographic information in HRTEM images was analyzed from
Digital Diffraction Patterns calculated from selected regions. Local,
subnanometer scale, chemical analysis of the catalysts was done
by recording X-EDS spectra using an electron probe with a 0.5 nm
diameter.
XPS, electron microscopy and FTIR studies have also been done
on samples after catalytic performance. For that purpose, the reac-
tion has been stopped at specified times, the reactant mixture
removed and analyzed by GC and the catalysts recovered by filtra-
tion followed by drying in vacuum and keeping in a N2 atmosphere
until use in characterization studies.
TiO2 was a commercial sample (Degussa P25, mainly anatase).
Nanocrystalline CeO2 was prepared by thermolysis of an acidi-
fied Ce(NO3)4 solution followed by redispersion, according to Ref.
[36]. The dispersion was purified and concentrated using an ultra-
filtration cell equipped with a 3KD membrane. The purification was
monitored by the residual acidity of the dispersion, determined
by an acid titration of the supernatant after ultra-centrifugation at
50 000 rpm for 6 h. 0.36 wt% Au was deposited on the nanopartic-
ulated cerium oxide by the following procedure: Nanocrystalline
CeO2 powders (2 g) were added into an HAuCl4 aqueous solution
(Aldrich, 99.9%) whose pH was fixed at 7.7 by addition of a 0.2 M
NaOH solution. The mixture was stirred for 12 h, statically aged for
6 h and then washed with deionized water, filtered and dried at
120 ◦C in air for 12 h. Samples of 0.1 and 0.24 wt% Pd supported

P. Concepción et al. / Catalysis Today 259 (2015) 213–221
215
2.3. Selective hydrogenation of 1,5-cyclooctadiene
Since the lack of activity observed with PdCeO2 samples could
be related, in a first approximation, to the absence of reduced Pd
species, one could then increase the amount of reduced Pd species,
and therefore their catalytic activity, by reducing the samples in H2.
Ex situ reduction at 200 ◦C in H2 leads to the formation of Pd0 in
PdCeO2, as confirmed by XPS analysis (Table 1) and, consequently to
an increase in the hydrogenation activity (Fig. 1B). However when
compared with Pd/TiO2, at the same Pd loading (0.1 wt%), the reac-
tion rate of the reduced 0.1PdCe-H2 sample (1.7 × 10−5 mol/g s)
is still lower than the value observed with the 0.13PdTi sam-
ple (4.4 × 10−5 mol/g s) or 0.13PdTi-H2 sample (7.5 × 10−5 mol/g s).
Thus, the different catalytic activity of Pd/CeO2 and Pd/TiO2 sam-
ples cannot be simply explained by a difference in the amount of
Pd0 species under reaction conditions (see Table 1), but other fac-
tors like metal morphology, size and shape, may also play a role,
being those under investigation. The selectivity to products differs
slightly in the 0.13PdTi versus 0.1PdCe sample (Fig. S1, supple-
mentary information). In particular, the selectivity to isomerization
product (1,3-cyclooctadiene) is lower in the PdTi sample while their
selectivity to cyclooctane is higher. In accordance with the above
reported differences in activity, the differences in selectivity agree
with the postulated morphological factors.
The liquid phase hydrogenation of 1,5-cyclooctadiene (COD)
was carried out in a reinforced glass batch reactor (2.5 mL,
SUPELCO) closed with a Teflon septum and equipped with a pres-
sure gauge and a side arm with an on–off lock. A micro-syringe
can be allowed to inject and withdraw liquid or gas through the
lock. Typically, the reactor with 7 mg (80–120 mesh) of the catalyst
and 56.1 mg (0.5 mmol) 1,5-cyclooctadiene in n-octane (1.0 mL)
was purged with hydrogen for 3 times and then hydrogen was
introduced up to 10 bar. The reaction was assumed to start when
the reactor was introduced into the oil bath at 50 ◦C and vigorously
stirred (ca. 1000 rpm). At a desired reaction times, 20 ␮L samples
were taken through the micro-syringe and analyzed with a HP gas
chromatograph equipped with a FID and a Carbowax 20M capillary
column. The mole calibration factors of n-cyclooctane (fCOA = 1.00),
cyclooctene (fCOE = 0.98), 1,3-cyclooctadiene (f1,3COD = 0.77) and
1,5-cyclooctadiene (fCOD = 0.77) were determined by standard solu-
tion of these components. Isomerization of 1,5-cyclooctadiene to
1,3-cyclooctadiene is observed as a primary product on all catalysts.
As the activity and selectivity of Pd can be modified by intro-
duction of promoters, we have explored this possibility here by
synthetizing bimetallic Pd–Au materials.
3. Results and discussion
3.1. Pd/CeO2 catalysts
The as prepared Pd/CeO2 samples show no catalytic activity for
the hydrogenation of COD, in opposite to Pd/TiO2 samples which
give conversions close to 100% at the same Pd loading (Fig. 1A).
XPS characterization of the samples prior to catalytic test (Table 1)
shows in both cases the presence of Pd2+ (BE 337–336 eV) [39,40]
making then difficult to explain their different catalytic behavior
from the analysis of the fresh catalysts. However, when the sam-
ples were analyzed by XPS after stopping the reaction at specific
reaction times, clear differences can be observed between the two
samples. Pd0 (BE 334.5 eV) [39,40] is detected in the PdTiO2 sam-
ple stopping the reaction at 50 min, while only Pd2+ was found in
the PdCeO2 sample even after 160 min reaction time. These results
would be consistent with the high potential of CeO2 to stabilize
cationic species and could help to explain the catalytic behavior
displayed in Fig. 1A by assuming Pd0 sites, generated under reac-
tion conditions, as the catalytic active sites for the hydrogenation
of COD.
3.2. Pd–Au/CeO2 catalysts
An air dried Au/CeO2 (0.36 wt% Au) sample was impregnated
with different Pd loadings (0.05–2 wt%) within a molar ratio Pd:Au
of 0.25–1. XPS data of the as prepared samples (Table 2) show
the presence of Au0 (84–83.6 eV) and Au+ (85.6 eV) [36,41] in both
Au/CeO2 and Pd–Au/CeO2 samples. Meanwhile Pd2+ and Pd0 are
both present in the as prepared Pd–Au/CeO2 samples, while only
Pd2+ was detected in Pd/CeO2 (see Table 1). Thus, the presence of
Au seems to favor the reduction of Pd2+ to Pd0 even during sam-
ple impregnation/drying. On the other hand, the Pd/Au molar ratio
determined by XPS increases (except in the 0.2Pd–0.36AuCe sam-
ple) when increasing the Pd loading, as expected according to the
preparation method in which Pd is added on the previously pre-
pared Au/CeO2 sample.
Ex situ reduction of the bimetallic samples in H2 at 200 ◦C
increased the amount of Pd0 species, and reduced all Au species
A
100
80
60
40
20
0
B
100
80
60
40
20
0
0
40
80
120
0
10
20
30
40
Reaction Time/ min
Reaction Time /min
Fig. 1. (A) COD conversion versus reaction time on 0.1PdCe (ᮀ), 0.13PdTi (ꢀ) and 0.24PdTi (ꢁ) samples. (B) COD conversion versus reaction time on ex situ reduced samples:
0.1PdCe-H2 (᭿) and 0.24 PdCe-H2 (ꢀ) samples.

216
P. Concepción et al. / Catalysis Today 259 (2015) 213–221
Table 1
XPS data of as prepared and ex situ reduced Pd/CeO and Pd/TiO2samples, before and after stopping the reaction at specific reaction times.e
Sample
Pd (wt%)a
Pd/(Me + O)b
Before catalytic test
After catalytic test
t (min)
%Ce
Pd/(Me + O)c
Pd2+ (%)d
Pd0 (%)d
Pd/(Me + O)c
Pd2+ (%)d
Pd0 (%)d
0.1PdCe
0.24PdCe
0.13PdT
0.23PdT
0.096
0.24
0.13
0.23
0.0005
0.0013
0.0003
0.0005
0.001
337.2 (100)
337.2 (100)
336.0 (100)
336.0 (100)
–
–
–
–
n.d.
0.0076
n.d.
n.d.
337.8 (100)
n.d.
n.d.
160
n.d.
50
0.0026
0.0005
0.0013
–
1.03
0.0013
336.5 (61.7)
334.5 (38.3)
97.0
0.1PdCe-H2
0.24PdCe-H2
0.13PdT-H2
0.096
0.24
0.13
0.0005
0.0013
0.0003
0.001
0.0026
0.0005
337.5 (61.4)
n.d.
337.0 (61.4)
334.9 (38.5)
333.6 (38.6)
0.0015
0.0048
0.0008
337.0 (59.4)
337.0 (38.1)
336.6 (61.2)
334.7 (40.6)
334.3 (61.9)
334.2 (38.8)
130
11
10
65
44.83
35.24
a
Pd weight percent determined by ICP.
Surface composition (atomic ratio) determined by ICP; Me = Ti or Ce.
Surface composition (atomic ratio) determined by XPS; Me = Ti or Ce.
BE (eV) of each oxidation state. Values in brackets represent the atomic percent of each oxidation state determined by XPS.
Conversion of COD (%C) at reaction time (t) at which the sample has been characterized.
b
c
d
e
Table 2
XPS data of as prepared and ex situ reduced Pd–Au/CeO2 samples before catalytic test.
Sample
Pd/Aua
Pd/Aub
Pd 3d5/2c
Pd2+ (%)
Au4f7/2c
Au0 (%)
Pd0 (%)
Au+ (%)
Au3+ (%)
0.00Pd–0.36AuCe
0.05Pd–0.36AuCe
0.10Pd–0.36AuCe
0.17Pd–0.36AuCe
0.20Pd–0.36AuCe
0
0
–
–
–
84.1 (76.1)
83.8 (83.0)
83.6 (67.6)
83.6 (67.6)
n.d
85.6 (23.8)
85.2 (17.0)
85.8 (32.4)
85.8 (32.4)
0.25
0.5
0.87
1
1.28
3.5
7.77
5.8
337.4 (100)
337.3 (85.7)
337.3 (88.2)
n.d
–
–
–
334.7 (14.3)
335.3 (11.8)
0.05Pd–0.36AuCe-H2
0.10Pd–0.36AuCe-H2
0.17Pd–0.36AuCe-H2
0.20Pd–0.36AuCe-H2
0.25
0.5
0.87
1
4
336.5 (64.5)
337.0 (53.3)
337.2 (47.7)
337.0 (74.2)
334.0 (35.5)
335.2 (46.7)
335.2 (52.3)
334.8 (25.8)
83.1 (100)
83.3 (100)
83.2 (100)
83.5 (100)
–
–
–
–
–
–
–
–
7.4
3.16
8.1
a
Surface composition (atomic ratio) determined by ICP.
Surface composition (atomic ratio) determined by XPS.
BE (eV) of each oxidation state. Values in brackets represent the atomic percent of each oxidation state determined by XPS.
b
c
to Au0 (Table 2). A change in the Pd/Au metal surface composition
is also observed by XPS, pointing to a re-dispersion of both metals
during H2 reduction. However no clear trend is observed in the
surface distribution of both metals (Pd/Au ratio) during reduction
conditions, with the metal loading. In the XPS core level bind-
ing energy (BE) the 84.1 eV of metallic gold, observed in the Pd
free Au/CeO2 sample, shifts to lower BE in the presence of Pd,
83.6 0.2 eV in the as prepared samples and to 83.2 0.2 eV in the
reduced ones. This shift could be related to particle size effects or
to an electronic interaction between Pd and Au. In the case of pal-
ladium, Pd0 BE between 334.5 and 335.5 eV are observed in the
as prepared as well as in reduced samples, however the different
Pd BE are not clearly associated to the presence of gold or to the
reduction treatment. Thus, according to the XPS results no clear
evidence about the presence of any Au–Pd particle interaction or
alloy formation can be deduced.
The catalytic activity of the Au/CeO2 catalyst for the hydroge-
nation of COD is very low, even after H2 reduction at 200 ◦C (8%
conversion at 160 min). As prepared Pd–Au/CeO2 samples shows
also low activity. However activity increases drastically after ex
situ reduction in H2 (Fig. 2), being this activity higher than on
reduced Pd/CeO2 samples and even higher than on the reduced
Pd/TiO2 samples (at the same Pd loading (0.1 wt%)), with initial
reaction rates of 11.2 × 10−5 mol/g s for the 0.1Pd–0.36AuCe-H2
sample. Moreover the initial reaction rate of reduced Pd–Au/CeO2
samples follows a volcano type curve when plotted versus the
nominal Pd loading (i.e. Pd/Au molar ratio), with a maximum
at a Pd/Au molar ratio of 0.5 (Fig. 3). The selectivity to isomer-
ization and hydrogenation products changes also with catalyst
composition (Fig. S2 in supplementary information). Specifically
the selectivity toward isomerization is lower in all Pd–Au/CeO2
samples versus the un-promoted PdCe sample. Interestingly the
lowest 1,3-cyclooctadiene selectivity is observed on the most active
catalysts (i.e. with a Pd/Au molar ratio of 0.5), while the selectiv-
ity to cyclooctane is higher on this sample. In order to understand
the promoting effect of gold on the catalytic activity observed with
AuPd/CeO2 bimetallic catalysts, and taking into account the results
presented in the first part of this work which showed that the
catalysts underwent dramatic changes under reaction conditions,
100
80
60
40
20
0
0
10
20
30
Reaction Time /min
Fig. 2. COD conversion versus reaction time on as prepared (open symbol) and ex
situ reduced (full symbols) 0.05Pd–0.36AuCe (ꢀ, ᭹), 0.1Pd–0.36AuCe (ᮀ, ᭿) and
0.17Pd–0.36AuCe (ଝ, ꢀ) samples.

P. Concepción et al. / Catalysis Today 259 (2015) 213–221
217
1,2
1,0
0,8
0,6
0,4
0,2
0,0
0,0
0,2
0,4
0,6
0,8
1,0
1,2
Pd/Au molar ratio
Fig. 3. Variation of the reaction rate in reduced Pd–Au/CeO2 samples versus the Pd/Au molar ratio determined by ICP.
characterization of the samples has been performed after stopping
the reaction at specific reaction times.
very suitable for analyzing particle size and morphology and deter-
mining surface properties.
From those data, we can conclude that under reaction condi-
tions, and according to XPS results, a modification of the oxidation
state of the metal surface sites and a surface restructuration of the
AuPd/CeO2 catalysts is observed (see Table 3 and Figs. S3a, b and
S4a, b in supplementary information). The amount of Pd0 species
increases in both the as prepared and ex situ reduced samples, to
around 53% and 70% respectively, due to the in situ reduction of Pd2+
ions under reaction conditions. On the other hand, a gold surface
segregation has been observed after reaction, being more evident
on the reduced samples, giving a Pd/Au molar ratio of ∼2.5 in all
cases. This agrees with the much lower surface energy of Au versus
Pd [42], favoring surface enrichment by Au. However the different
catalytic behavior shown in Fig. 2 for the as prepared and ex situ
reduced samples, cannot be explained according to the XPS data,
as both samples show similar XPS results once exposed to reaction
conditions. Therefore further characterization work was required
for explaining the reactivity behavior. Since different morphology
of the exposed metal particle, such as crystallographic phases, coor-
dination number of the surface atoms, particle size effects, etc. has
been reported to influence the catalytic behavior of Pd samples in
selective hydrogenation reactions, electron microcopy studies as
well as IR studies of CO adsorption as probe molecule have been
performed on the as prepared and ex situ reduced 0.1Pd–0.36AuCe
sample after stopping the reaction at 15 min. Both techniques are
Electron microscopy results show the presence of both large
and very small Au nanocrystals in the as prepared 0.1Pd–0.36AuCe
sample (non-active sample) after stopping the reaction at 15 min
reaction time. Data included in Fig. 4 illustrate the case of the
large Au nanoparticles. The X-EDS spectrum recorded from this
large, very high intensity area confirms that it corresponds to Au.
Likewise, the combination of HRTEM and XEDS results recorded
on the same areas of the catalyst also suggests the occurrence of
very small gold nanocrystals. Thus, Fig. 5 shows a HRTEM image in
which a CeO2 nanoparticle is unambiguously identified from the
spots observed in the corresponding DDP (shown as inset). The X-
EDS spectrum recorded from the same crystallite, using a 0.5 nm
probe in STEM mode, shows Au in addition to Ce and O peaks. Since
no gold particle is clearly visible in the analyzed area, we must
conclude that gold must be present in that area in a very highly dis-
persed state. Image simulation data [43] indicate that the detection
limit in HRTEM images of metal nanoparticles supported on ceria is
roughly below 1 nm. Small patches of material with diameter below
1 nm have been also detected by HRTEM on the border of the CeO2
nanocrystals which suggest the occurrence of subnanometer-sized
Au nanocrystals (Fig. S5, supplementary information). Pd particles
could not be observed in the images of the as-prepared catalyst.
Likewise, X-EDS from large regions in this catalyst did not show
the presence of Pd peaks, this suggesting that probably Pd is in the
Table 3
XPS data of as prepared and ex situ reduced Pd–Au/CeO2 samples after stopping the reaction at specific reaction times.a
Sample
t (min)
C (%)a
Pd/Aub
Pd3d5/2c
Pd2+ (%)
Au4f7/2c
Au0 (%)
Pd0 (%)
Au+ (%)
Au3+(%)
0.05Pd–0.36AuCe
0.10Pd–0.36AuCe
0.17Pd–0.36AuCe
0.20Pd–0.36AuCe
15
15
n.d
n.d.
7.13
3.91
2.4
3
n.d
n.d
337.4 (45.1)
337.2 (47.9)
n.d
334.3 (54.9)
334.7 (52.1)
83.9 (77.6)
83.8 (57.1)
n.d
85.6 (22.4)
85.6 (42.9)
–
–
n.d
n.d
0.05Pd–0.36AuCe-H2
0.10Pd–0.36AuCe-H2
0.17Pd–0.36AuCe-H2
0.20Pd–0.36AuCe-H2
15
15
7
54.04
94.55
39.2
2.2
336.8 (42.1)
337.6 (27.8)
337.2 (28.5)
337.5 (41.7)
334.6 (57.9)
335.0 (72.2)
334.7 (71.5)
334.8 (58.3)
84.0 (68.6)
84.2 (67.8)
83.3 (55.5)
83.8 (77.9)
–
–
–
–
86.0 (31.4)
86.5 (32.2)
86.1 (44.5)
86.1 (22.1)
2.57
2.23
2.4
7
55.35
a
Conversion of COD (C) at reaction time (t) at which the sample has been characterized.
Surface composition (atomic ratio) determined by XPS.
BE (eV) of each oxidation state. Values in brackets represent the atomic percent of each oxidation state determined by XPS.
b
c

218
P. Concepción et al. / Catalysis Today 259 (2015) 213–221
Fig. 4. HAADF-STEM of the as prepared 0.1Pd–0.36AuCe sample after 15 min reaction time. The XEDS-spectrum was recorded on the large crystallite shown as enlargement.
Copper and molybdenum signals arise from the specimen holder and the coated Mo-grids used, respectively.
the 2098 cm−1 IR band could also be related to unsaturated sites
associated to Pd(1 0 0) surface [45–48], HRTEM disregard this pos-
sibility since larger Pd crystallites (with a higher saturation degree)
could be assumed from the HRTEM data.
form of larger crystallites, statistically hard to find due to the very
low Pd content. Particle size distributions could not be obtained
in any case due to intrinsic difficulties associated to the combined
effects of very small metal loadings, the nanocrystalline nature of
the support which results in a huge number of crystal overlaps and
the heavy atomic number of Ce.
IR spectra of CO adsorption performed on the as prepared
0.1Pd–0.36AuCe sample after stopping the reaction at 15 min reac-
tion time (Fig. 6A) shows the presence of an IR band at 2127 cm−1
which should be related to Pd␦+ or Au␦+ (oxo) species [44], and an IR
band at 2098 cm−1 ascribed to unsaturated surface sites, probably
associated to the small Au particles detected by HRTEM [44]. While
A
different picture is clearly observed in the reduced
0.1Pd–0.36AuCe sample (active sample) after stopping the reaction
at 15 min reaction time. In this case, X-EDS spectra analysis of high
intensity areas observed in HAADF-STEM images of this catalyst,
using a tiny (0.5 nm) electron probe, show the presence of both Au
and Pd signals (atomic composition around 88.68%Au, 11.32%Pd,
Pd/Au ratio of 0.13), suggesting the coexistence of both metals in
the same particle as it can be seen in Fig. 7. However confirmation
Fig. 5. HTREM image of the as prepared 0.1Pd–0.36AuCe sample after 15 min reaction time. The DDP inset confirms the cubic fluorite-like phase for CeO2.

P. Concepción et al. / Catalysis Today 259 (2015) 213–221
219
A
0.025
2127
2098
2200
2150
2100
2050
Wavenumber (cm-1)
B
0.025
2103
2095
c
b
a
2200
2150
2100
2050
Wavenumber (cm-1)
Fig. 6. (A) FTIR spectra of CO adsorption at −176 ◦C at increasing CO coverages (0.2–1 mbar) on as prepared 0.1Pd–0.36AuCe sample after 15 min reaction time. Large saturated
IR band at 2155 cm−1 assigned either to Au+, Ce4+, OH groups [51] or to Pd2+ ions [47]. (B) FTIR spectra of CO adsorption at −176 ◦C and 0.2 mbar (a), 0.5 mbar (b) and 1 mbar
(c) on ex situ reduced 0.1Pd–0.36AuCe sample, after 15 min reaction time. Large saturated IR band at 2155 cm−1 assigned either to Au+, Ce4+, OH groups [51] or to Pd2+ ions
[47]
crystals (at ∼2110 cm−1) are detected. This agrees with HRTEM
of a PdAu alloy phase by HRTEM was not possible due to the over-
lap of the image of these small particles with the contrasts coming
from a large number of underlying ceria crystallites. DDPs from
the HREM images could only be indexed as corresponding to cubic
fluorite CeO2. Smaller particles containing only Pd have also been
detected, as it is illustrated by the HAADF-STEM and XEDS-STEM
studies summarized in Fig. 8.
The surface structure of the reduced 0.1Pd–0.36AuCe sample
after stopping the reaction at 15 min reaction time has also been
studied by IR spectroscopy of adsorbed CO (Fig. 6B). IR bands in the
2106–2098 cm−1 region, ascribed to Pd0 and Au0 low coordinated
metal surface sites are observed, while no IR bands due to large
data where no large Pd crystallites were detected on the reduced
samples after catalytic test.
At this point we can say that HRTEM and FTIR-CO results suggest
the coexistence of both Au and Pd metals in the same particle, with
an Pd/Au atomic composition of 0.13, together with the absence
of big crystallites and the stabilization of small Pd particles, after
catalytic test in the most active ex situ reduced 0.1Pd–0.36AuCe
sample, while they were not observed in the non-active sample.
Moreover a strong restructuration of the surface metal species
has been observed after reaction conditions, with gold segrega-
tion toward the surface, disrupting the Pd crystal surface. In fact,

220
P. Concepción et al. / Catalysis Today 259 (2015) 213–221
Fig. 7. HAADF-STEM image and XEDS spectra of the reduced 0.1Pd–0.36AuCe sample after 15 min reaction time.
Fig. 8. HAADF-STEM image and XEDS spectra recorded on different spots of the reduced 0.1Pd–0.36AuCe sample after 15 min reaction time. XEDS spectrum from the region
marked as 1 on the image peaks both Ce and Pd signals whereas only Ce peaks are observed in the spectrum obtained from 2.
XPS data of the ex situ reduced 0.1Pd–0.36AuCe sample shows a
decrease in the Pd/Au molar ratio from 7.4 prior to reaction to
2.6 after reaction, pointing to a decrease of Pd on the surface.
This agrees with the FTIR studies of CO adsorption where well
faceted Pd crystals observed on the ex situ reduced 0.1Pd–0.36AuCe
sample before reaction (Fig. S6, supplementary information) are
not more detected once the sample has been exposed to reac-
tion conditions. On the other hand, since the catalytic activity
shows a volcano type correlation with the Pd/Au molar ratio,
the surface composition of the Au–Pd surface phase should play
an important role for catalytic activity. In our case, X-EDS STEM
analysis shows a Pd/Au molar ratio of 0.13 for the most active sam-
ple. This result implicates isolated Pd surface sites. Moreover the
observed drop in the selectivity to isomerization product agrees
with the presence of isolated Pd surface sites on the most active
sample. However, in detrimental, the selectivity to further hydro-
genated product (cyclooctane) increases. This may be related to the
different interaction strength of intermediate products on isolated
Pd surface sites [49]. According to Goodman and co-workers [50],
the promotional role of Au is to generate isolated Pd monomer
sites. Thus these authors attributed the enhanced catalytic activity
of Au–Pd catalysts for the acetoxylation of ethylene to vinyl acetate
to pairs of Pd monomers.
4. Conclusions
Strong restructuration of AuPd/CeO2 catalysts is observed after
reaction conditions, with gold surface segregation favoring dilution
of Pd atoms by Au at the surface of the crystallites. In this sense,
the promotional role of Au in our catalysts appears to be related
to generate isolated Pd sites, which act as catalytic active sites for
the selective hydrogenation of 1,5-COD. This conclusion opens new
perspectives for the design of more active and selective catalyst,
this being the subject of further investigations.

P. Concepción et al. / Catalysis Today 259 (2015) 213–221
[18] J.K. Edwards, B.E. Solsona, P. Landon, A.F. Carley, A. Herzing, C.J. Kiely, G.J.
221
Acknowledgements
Hutchings, J. Catal. 236 (2005) 69–79.
[19] G.C. Bond, A.F. Rawle, J. Mol. Catal. A 109 (1996) 261–271.
[20] H. Arnold, F. Döbert, J. Gaube, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.),
Handbook of Heterogeneous Catalysis, vol. 5, VCH, Weinheim, 1997, pp.
2165–2186.
[21] J.P. Boitiaux, J. Cosyns, E. Robert, Appl. Catal. 35 (1987) 193–209;
J.P. Boitiaux, J. Cosyns, S. Vasudevan, Appl. Catal. 6 (1983) 41–51.
[22] J. Cosyns, Catalyse par les Metaux, CNRS Edition, Paris, 1984, pp. 371–379.
[23] H. Miura, Catal. Today 28 (1996) 215–221.
The authors thank the project CSD2009-00050 for financial sup-
port. SGG thank the Spanish Ministry of Science and Innovation
for financial support through project MAT2011-28009. JCHG and
JJCG acknowledge financial support from MINECO/FEDER (Projects
MAT2013-40823-R and CSD2009-00013). JCHG acknowledges sup-
port from the Ramón y Cajal Fellowships Program of MINECO
(RYC-2012-10004).
[24] A. Sárkány, O. Geszti, G. Sáfrán, Appl. Catal. A 350 (2008) 157–163.
[25] L. Piccolo, A. Piednoir, J.C. Bertolini, Surf. Sci. 592 (2005) 169–181.
[26] A.K. Krauth, G.H. Bernstein, E.E. Wolf, Catal. Lett. 45 (1997) 177–186.
[27] A.A. Herzing, A.F. Carley, J.K. Edwards, G.J. Hutchings, C.J. Kiely, Chem. Mater.
20 (2008) 1492–1501.
Appendix A. Supplementary data
[28] Z. Suo, C. Ma, M. Jin, T. He, L. An, Catal. Commun. 9 (2008) 2187–2190.
[29] H. Liu, G. Mao, S. Meng, J. Mol. Catal. 74 (1992) 275–284.
[30] K. Okitsu, M. Murakami, S. Tanabe, H. Matsumoto, Chem. Lett. 11 (2000)
1336–1337.
[31] Q. Chen, L. Luo, X. Yang, Indian J. Chem. 47A (2008) 1317–1322.
[32] M.P. Kapoor, Y. Ichihashi, T. Nakamori, Y. Matsumura, J. Mol. Catal. A: Chem.
213 (2004) 251–255.
[33] D. Mei, M. Neurock, C.M. Smith, J. Catal. 268 (2009) 181–195.
[34] H. Miura, M. Terasaka, K. Oki, T. Matsuda, Stud. Surf. Sci. Catal. 75 (1993)
2379–2382.
[35] M. Alhumaimess, Z.J. Lin, W.H. Weng, N. Dimitratos, N.F. Dummer, S.H. Taylor,
J.K. Bartley, C.J. Kiely, G.J. Hutchings, Chem. Sus. Chem. 5 (2012) 125–131.
[36] J.Y. Chane-Ching, EP 208580, 1987.
[37] J.K. Edwards, J. Pritchard, L. Lu, M. Piccinini, G. Shaw, A.F. Carley, D.J. Morgan,
C.J. Kiely, G.J. Hutchings, Angew. Chem. Int. Ed. 53 (2014) 2381–2384.
[38] R.S. Monteiro, F.B. Noronha, L.C. Dieguez, M. Schmal, Appl. Catal. A: Gen. 131
(1995) 89–106.
[39] K.R. Priolkar, P. Bera, P.R. Sarode, M.S. Hegde, S. Emura, R. Kumashiro, N.P.
Lalla, Chem. Mater. 14 (2002) 2120–2128.
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2015.07.
022
References
[1] S. Alayoglu, F. Tao, V. Altoe, C. Specht, Z. Zhu, F. Aksoy, D.R. Butcher, R.J.
Renzas, Z. Liu, G.A. Somorjai, Catal. Lett. 141 (2011) 633–640.
[2] F. Tao, M.E. Grass, Y. Zhang, D.R. Butcher, J.R. Renzas, Z. Liu, J.Y. Chung, B.S.
Mun, M. Salmeron, G.A. Somorjai, Science 322 (2008) 932–934.
[3] N.J. Divins, I. Angurell, C. Escudero, V. Pérez-Dieste, J. Llorca, Science 346
(2014) 620–623.
[4] K. Qadir, S.H. Joo, B.S. Mun, D.R. Butcher, J.R. Renzas, F. Aksoy, Z. Liu, G.A.
Somorjai, J.Y. Park, Nano Lett. 12 (2012) 5761–5768.
[5] W.C. Chueh, A.H. McDaniel, M.E. Grass, Y. Hao, N. Jabeen, Z. Liu, S.M. Haile, K.F.
McCarty, H. Bluhm, F. El Gabaly, Chem. Mater. 24 (2012) 1876–1882.
[6] A. Pulido, P. Concepción, M. Boronat, A. Corma, J. Catal. 292 (2012) 138–147.
[7] Q. Hua, T. Cao, X.K. Gu, J. Lu, Z. Jiang, X. Pan, L. Luo, W.X. Li, W. Huang, Angew.
Chem. Int. Ed. 53 (8) (2014) 4856–4861.
[8] T. Ishida, N. Kinoshita, H. Okatsu, T. Akita, T. Takei, M. Haruta, Angew. Chem.
Int. Ed. 47 (2008) 9265–9268.
[9] S. Arrii, F. Morfin, A.J. Renouprez, J.L. Rousset, J. Am. Chem. Soc. 126 (2004)
1199–1205.
[10] G.W. Huber, J.W. Shabaker, J.A. Dumesic, Science 300 (2003) 2075–2077.
[11] 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
(2006) 362–365.
[12] S. Nishimura, Y. Yakita, M. Katayama, K. Higashimine, K. Ebitani, Catal. Sci.
Technol. 3 (2013) 351–359.
[13] L. Kesavan, R. Tiruvalam, M.I. bin Saiman, D.I. Enache, R.L. Jenkins, N.
Dimitratos, J.A. Lopez-Sanchez, S.H. Taylor, D.W. Knight, C.J. Kiely, G.J.
Hutchings, Science 331 (2011) 195–199.
[14] A. Huidobro, A. Sepúlveda-Escribano, F. Rodríguez-Reinoso, J. Catal. 212
(2002) 94–103.
[40] J. Batista, A. Pintar, D. Mandrino, M. Jenko, V. Martin, Appl. Catal. A: Gen. 206
(2001) 113–124.
[41] J. Guzman, S. Carretin, A. Corma, J. Am. Chem. Soc. 127 (2005) 3286–3287.
[42] A. Jablonski, S.H. Overbury, G.A. Somorjai, Surf. Sci. 65 (1977) 578–592.
[43] S. Bernal, J.J. Calvino, M.A. Cauqui, J.A. Pérez Omil, J.M. Pintado, J.M.
Rodríguez-Izquierdo, Appl. Catal. B: Environ. 16 (1998) 127–138.
[44] M. Boronat, P. Concepción, A. Corma, J. Phys. Chem. C 113 (2009)
16772–16784.
[45] Y. Perez, M.L. Ruiz-González, J.M. González-Calbet, P. Concepción, M. Boronat,
A. Corma, Catal. Today 180 (2012) 59–67.
[46] H. Borchert, B. Jürgens, V. Zielasek, G. Rupprechter, S. Giorgio, C.R. Henry, M.
Bäumer, J. Catal. 247 (2007) 145–154.
[47] D. Tessier, A. Rakai, F. Bozon-Verduraz, J. Chem. Soc. Faraday Trans. 88 (1992)
741–749.
[48] S. Bertarione, D. Scarano, A. Zecchina, V. Johánek, J. Hoffmann, S.
Schauermann, M.M. Frank, J. Libuda, G. Rupprechter, H. Freud, J. Phys. Chem. B
108 (2004) 3603–3613.
[49] D. Childers, A. Saha, N. Schweitzer, R.M. Rioux, J.T. Miller, R.J. Meyer, ACS
Catal. 3 (2013) 2487–2496.
[50] M. Chen, D. Kumar, C.-W. Yi, D.W. Goodman, Science 310 (2005) 291–293.
[51] P. Concepción, S. Carretín, A. Corma, Appl. Catal. A: Gen. 307 (2006) 42–45.
[15] Q. Zhang, J. Li, X. Liu, Q. Zhu, Appl. Catal. A: Gen. 197 (2000) 221–228.
[16] A. Sárkány, A. Horváth, A. Beck, Appl. Catal. A: Gen. 229 (2002) 117–125.
[17] F. Maroun, F. Ozanam, O.M. Magnussen, R.J. Behm, Science 293 (2001)
1811–1814.