The effect of Ni in Pd-Ni/(Ce,Zr)Ox/Al2O3 catalysts used for stoichiometric CO and NO elimination. Part 2: Catalytic activity and in situ spectroscopic studies

Article abstract of DOI:10.1016/j.jcat.2005.08.012

Pd-Ni catalysts supported on Al₂O₃, (Ce,Zr)O _x/Al₂O₃, and (Ce,Zr)O_x were examined with the principle objective of determining the effects of Ni on catalytic activity for CO oxidation and NO reduction reactions under stoichiometric conditions. Catalytic activity findings for the CO + O₂ and CO + O₂ + NO reactions were analyzed in conjunction with in situ DRIFTS and XANES results to obtain information on the processes occurring in the catalysts during the course of the reactions. The results reveal a significant dependence on the nature of the support in terms of the catalytic changes produced by nickel. In the absence of significant nickel-induced electronic perturbations of palladium, these are related to indirect effects on palladium distribution over the catalysts or to a certain impediment of the interactions between active palladium and Ce-Zr mixed-oxide components. Significant promotion of CO oxidation was observed for the (Ce,Zr)O_x/Al₂O ₃-supported catalyst, which reveals a relevant role for the particle size of the nanostructured Ce-Zr mixed oxide in this reaction.

Full text of DOI:10.1016/j.jcat.2005.08.012

Journal of Catalysis 235 (2005) 262–271
www.elsevier.com/locate/jcat
The effect of Ni in Pd–Ni/(Ce,Zr)O /Al O catalysts used for
x
2 3
stoichiometric CO and NO elimination.
Part 2: Catalytic activity and in situ spectroscopic studies
a
a
b
a,∗
A.B. Hungría , M. Fernández-García , J.A. Anderson , A. Martínez-Arias
a
Instituto de Catálisis y Petroleoquímica, CSIC, C/ Marie Curie, Campus Cantoblanco, 28049 Madrid, Spain
b
Surface Chemistry and Catalysis Group, Department of Chemistry, University of Aberdeen, AB24 3UE Scotland, UK
Received 18 March 2005; revised 30 June 2005; accepted 16 August 2005
Abstract
Pd–Ni catalysts supported on Al O , (Ce,Zr)Ox/Al O , and (Ce,Zr)Ox were examined with the principle objective of determining the effects of
2
3
2 3
Ni on catalytic activity for CO oxidation and NO reduction reactions under stoichiometric conditions. Catalytic activity findings for the CO + O
2
and CO + O + NO reactions were analyzed in conjunction with in situ DRIFTS and XANES results to obtain information on the processes
2
occurring in the catalysts during the course of the reactions. The results reveal a significant dependence on the nature of the support in terms of
the catalytic changes produced by nickel. In the absence of significant nickel-induced electronic perturbations of palladium, these are related to
indirect effects on palladium distribution over the catalysts or to a certain impediment of the interactions between active palladium and Ce–Zr
mixed-oxide components. Significant promotion of CO oxidation was observed for the (Ce,Zr)Ox/Al O -supported catalyst, which reveals a
2
3
relevant role for the particle size of the nanostructured Ce–Zr mixed oxide in this reaction.
2005 Elsevier Inc. All rights reserved.

Keywords: Pd–Ni catalysts; TWC; CeO –ZrO ; Al O ; CO oxidation; NO reduction; DRIFTS; Pd K-edge XANES
2
2
2 3
1
. Introduction
Part 1 of this work was dedicated to analyzing in detail as-
lysts. The presence of nickel apparently favors a preferential
interaction of palladium with the Ce–Zr mixed-oxide compo-
nent of the support in the (Ce,Zr)Ox/Al2O3-supported system,
whereas nickel interacts preferentially with the alumina compo-
nent. Concerning the chemical nature of palladium and nickel
components in the initial catalysts, these are shown to appear
in the form of small PdO-type or NiO-type particles (in addi-
tion, for alumina-supported samples, a part of the nickel can
form NiAl O -type particles). Overall, the results suggest no
strong interactions between Pd and Ni in the initial calcined
systems, although a certain degree of interaction between both
components in the (Ce,Zr)Ox-supported catalyst cannot be dis-
counted. The dispersion of palladium and the corresponding
characteristic size of the PdO particles were strongly influ-
enced by the support nature. These findings appear to be fairly
similar for systems having alumina as the carrier (Al2O3- and
pects of characterization that could be relevant to understanding
the catalytic properties of bimetallic Pd–Ni catalysts supported
on Al2O3, (Ce,Zr)Ox/Al2O3, and (Ce,Zr)Ox [1]. These can be
related to the amount and nature of the contacts established be-
tween Pd and the different support components, the structural
and compositional characteristics of the Ce–Zr mixed-oxide
component of the support, the dispersion and chemical state of
palladium, or the existence of interactions between Pd and Ni
2
4
[
2–12]. In this respect, as we concluded in Part 1 [1], the pres-
ence of rather homogeneous Ce–Zr mixed-oxide nanostructures
with average particle sizes of 3.4 and 5.4 nm, respectively) was
revealed in (Ce,Zr)Ox/Al2O3- and (Ce,Zr)Ox-supported cata-
(
*
x 2 3
(Ce,Zr)O /Al O -supported systems), in which particle size
Corresponding author.
E-mail address: amartinez@icp.csic.es (A. Martínez-Arias).
averaged ca. 2.5 nm. In contrast, a considerably higher degree
0021-9517/$ – see front matter  2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2005.08.012

A.B. Hungría et al. / Journal of Catalysis 235 (2005) 262–271
263
ture under the reactive flow at a ramp rate of 5 K min 1, with a
glass-sheathed thermocouple placed in the center of the catalyst
bed for temperature control.
−
of dispersion was seen in the (Ce,Zr)Ox-supported catalyst. In
turn, nickel achieved highly dispersed states in all of the sys-
tems.
This second part of the study aims to analyze the catalytic
properties of such Pd–Ni systems, focusing in particular on the
promoting effects induced by the presence of nickel in the cat-
alyst formulation. For this purpose, the catalytic activity for
CO oxidation and NO reduction of the bimetallic Pd–Ni cat-
alysts supported on Al2O3, (Ce,Zr)Ox/Al2O3, and (Ce,Zr)Ox
during light-off tests was examined in comparison with that of
monometallic Pd systems, which have been analyzed in detail
in previous studies [7,9,10]. The present study is complemented
by the analysis of the catalysts under reaction conditions by in
situ DRIFTS and XANES techniques to obtain information on
possible changes occurring during the course of the reaction on
interaction with the reactant mixture.
Diffuse reflectance infrared Fourier transform spectroscopy
(DRIFTS) analysis of adsorbed species present on the cata-
lyst surface under reaction conditions was carried out using a
Perkin–Elmer 1750 FTIR fitted with an MCT detector. On-line
analysis of the NO concentration at the outlet of the infrared
(IR) chamber was performed by chemiluminescence (Thermo
Environmental Instruments 42C). The DRIFTS cell (Harrick)
was fitted with CaF2 windows and a heating cartridge that al-
lowed samples to be heated to 773 K. Samples of ca. 65 mg
were pretreated in situ under diluted O2 (in a manner similar to
that for the catalytic tests), before introduction of the reaction
−1
mixture and heating at 2 K min from 298 to 673 K, recording
−1
one spectrum (4 cm resolution, average of 50 scans) gener-
ally every 10–15 K. The gas mixture (similar to that used for
catalytic tests) was prepared using a computer-controlled gas
2
. Experimental
3
−1
blender with 80 cm min passing through the catalyst bed.
X-Ray absorption near-edge structure (XANES) experi-
ments at the Pd K-edge were performed on beamline BM29
of the ESRF synchrotron and at station 9.3 of the SRS syn-
chrotron. A Si(311) double-crystal monochromator was used in
conjunction with a rejection mirror to minimise the harmonic
content of the beam. Transmission experiments were carried
out using Kr/Ar-filled ionization chambers. The energy scale
was simultaneously calibrated by measuring a Pd foil using
a third ionization chamber. Samples were self-supported (ab-
sorbance, 0.5–1.0) and placed in a controlled-atmosphere cell
for treatment. XANES spectra were taken every 15 K in the
presence of the CO + NO + O2 flowing mixture (similar to the
The catalysts were prepared as indicated in Part 1 [1]. The
bimetallic Pd–Ni catalysts are referred to as PdNiA, PdNiCZA,
and PdNiCZ (with 1 wt% loading for each of the metals) for
the systems supported on Al2O3, CeZrO4/Al2O3 (with 10 wt%
of the Ce–Zr mixed oxide), and CeZrO4, respectively. Another
two bimetallic catalysts with 0.5 and 2 wt% Ni and the same
Pd loading (1 wt%) over the CZA support were prepared by the
same methods described in Part 1 [1] and are denoted here by
Pd0.5NiCZA and Pd2NiCZA, respectively. Monometallic ref-
erence catalysts (with 1 wt% loading in all cases) are identified
with either Pd or Ni as a prefix.
Catalytic activity tests using stoichiometric gas mixtures of
−1
1
% CO + 0.5% O2 or 1% CO + 0.45% O2 + 0.1% NO (N2
one used for catalytic activity tests) during a 5 K min tem-
perature ramp up to 673 K. The series of spectra were analyzed
using principal factor analysis (PCA), details of which can be
found elsewhere [14].
4
−1
balance) at 3 × 10 h were performed in a Pyrex glass fixed-
4
−1
bed flow reactor. Space velocity (3 × 10 h ) and sieving size
(
particles in the 0.125–0.250 mm range) were selected to mini-
mize diffusion effects [13] as estimated from independent tests
performed on the PdNiCZA catalyst for the CO–O2 reaction,
which indicated the absence of significant mass-transfer limi-
tations for particle sizes < ca. 0.5 mm (at least for conversion
values < ca. 50% in this case); in this respect and for compar-
ative purpose, it is worth noting that no significant differences
in textural properties (surface area, pore distribution) were de-
tected when comparing monometallic Pd and bimetallic Pd–Ni
catalysts on each respective support. The gases used (all of
commercial purity: 99.9% CO, 99.99% O2, and 98.5% NO, the
latter with 1% NO2 and 0.5% N2O as the main impurities) were
regulated with mass flow controllers and analyzed on-line us-
ing a Perkin–Elmer 1725X Fourier transform infrared (FTIR)
spectrometer coupled with a multiple reflection transmission
cell (Infrared Analysis Inc.). Oxygen concentrations were deter-
mined using a paramagnetic analyzer (Servomex 540A). Exper-
imental error in the CO and/or NO conversion values obtained
under these conditions is estimated as ± 7%. Before catalytic
testing, a calcination pretreatment was performed in situ under
Further analysis of the catalysts after use in the catalytic tests
mentioned earlier was performed by X-ray diffraction, using the
equipment described in Part 1 [1]. The diffractograms obtained
were analogous to those observed for the initial calcined cat-
alysts, thus revealing the stability of the support components
under reaction conditions [1].
3. Results
3.1. Catalytic activity tests
Fig. 1 shows the conversion profiles obtained for the bimetal-
lic catalysts under CO + O2. The results demonstrate the pro-
moting role of the Ce–Zr mixed-oxide component, in accor-
dance with results reported previously for the monometallic
Pd catalysts [9]. Relatively low apparent activation energies, in
−1
the 25–35 kJ mol range (which should be considered only in
qualitative terms, because the high activity of the catalysts at
room temperature precludes a guarantee of the system’s differ-
ential behavior), are estimated from analysis of low conversion
points of PdNiCZ and PdNiCZA catalysts, suggesting active
sites involving palladium–cerium contacts [15]. Greater promo-
2
.5% O2 (in N2) at 773 K, followed by cooling under the same
atmosphere and a thorough N2 purge at room temperature. Tests
were performed in the light-off mode, increasing the tempera-

264
A.B. Hungría et al. / Journal of Catalysis 235 (2005) 262–271
Fig. 1. CO conversion profiles obtained during the CO–O reaction over the
2
indicated catalysts.
Table 1
Isoconversion temperatures (in K) at 50% conversion obtained during stoi-
4
−1
chiometric CO–O and CO–O –NO reactions (3 × 10
h
GHSV) over the
2
2
bimetallic Pd–Ni catalysts and the corresponding monometallic references
Sample
T
CO
T
CO
T NO
50
2
50
CO+O2
50
CO+NO+O
CO+NO+O
2
PdNi Pd
447 447
Ni
PdNi Pd
Ni
PdNi Pd
Ni
A-supported
479
479
408 410
CZA-supported
0
1
2
.5 wt% Ni 334
406
460
wt% Ni
wt% Ni
334
339
404 637 403
410
433 668 455 456
464
770
Fig. 2. Conversion profiles obtained during the CO–O –NO reaction over the
2
CZ-supported
335 <303 498 356
358 520 448 485 >773
indicated catalysts.
system (Table 1). In contrast, no specific influence of the pres-
ence of Ni could be identified in overall terms in regard to
N2 selectivity in the NO reduction process, in accordance with
previous reports [7,10]). Other details apparent in the NO con-
version profiles shown in Fig. 2 concern the presence of oscil-
lations at relatively low reaction temperatures (particularly for
the CZ-supported system), which, in accordance with previous
analyses [10], are essentially related to adsorption/desorption
of NOx species. Additionally, analysis of the data in Table 1 re-
veals that changing the Ni loading (in the 0.5–2.0 wt% range)
in the bimetallic CZA-supported system produced only mini-
mal effects on the catalytic activity for these reactions, with the
optimum properties achieved in global terms for the 1 wt% Ni
system (PdNiCZA). In contrast, comparison with the catalytic
performance observed over monometallic Ni systems (Table 1)
indicates that the catalytic behavior observed in the bimetallic
catalysts cannot be attributed to effects induced by Ni alone.
tion of CO oxidation activity was achieved in the presence of Ni
for the CZA-supported catalyst, as evidenced by data presented
in Table 1. These data reveal an important promoting effect of
nickel in this system that appears to be almost independent of
the amount of nickel present in the catalyst (Table 1). In con-
trast, the presence of Ni produced an apparent decrease in CO
oxidation activity in the CZ-supported system, but had no sig-
nificant influence on the A-supported system (Table 1).
Different effects were induced by the presence of NO in
the reactant mixture as a function of the support and/or the
presence of Ni in the catalyst (Fig. 2; Table 1). The only ma-
terial for which the presence of Ni did not appear to exert
any significant influence was the A-supported catalyst, which
in turn displayed the greatest NO reduction activity. As ob-
served previously for the monometallic Pd catalysts [7,10], the
presence of NO had a detrimental effect on CO oxidation ac-
tivity in the bimetallic systems. However, in contrast to the
behavior observed for the monometallic systems for which the
magnitude of the NO-induced reduction of the CO oxidation
activity was observed to increase with the CZ content of the
catalyst [10], a relatively greater level of deactivation was ob-
served here for PdNiCZA than for PdNiCZ. In terms of NO
reduction, the results reveal that the presence of Ni somewhat
limited the detrimental effect induced by the establishment of
Pd–CZ contacts [10]. This Ni-induced enhancement of NO re-
duction activity was particularly pronounced for the PdNiCZ
3.2. In situ DRIFTS
The results of DRIFTS experiments recorded under CO+O2
and CO + O2 + NO for the bimetallic systems are shown in
Figs. 3 and 4, respectively. A reasonable correlation is observed
between evolution of bands due to gaseous CO2 (appearing in
−1
the 2400–2300 cm range) and CO conversion in the cat-
alytic tests (Figs. 1 and 2). Unfortunately, the relatively low

A.B. Hungría et al. / Journal of Catalysis 235 (2005) 262–271
265
Fig. 3. In situ DRIFTS spectra obtained during the course of the CO–O reaction for the indicated catalysts. Spectra taken every 20 K from 303 up to 483–543 K
2
(from bottom to top).
Fig. 4. In situ DRIFTS spectra obtained during the course of the CO–O –NO reaction for the indicated catalysts. Spectra taken every 30 K from 303 up to 573 K
2
(from bottom to top).
concentration of NO used in these tests prevents direct moni-
toring from analysis of the NO gas phase bands in the DRIFTS
spectra. Nevertheless, on-line chemiluminescence monitoring
of NOx species during the CO + O2 + NO reaction revealed
good agreement between NO conversion results obtained in the
catalytic tests of Fig. 2 and those obtained using the DRIFTS
cell. The DRIFTS spectra revealed the formation of different
chemisorbed carbonyl (or nitrosyl) species on interaction with
the reactant mixtures. Their characteristics and evolution were
observed to depend on various factors, including the presence
or absence of NO in the mixture and the amount of CZ compo-
nent present in the catalyst.
Results for the CO + O2 reaction show the formation of
an atop carbonyl species chemisorbed on metallic Pd parti-
cles immediately on contact with the reactant mixture at room
−1
temperature (band at 2100–2060 cm ; see Fig. 3) [7,16–19].
Differences in the intensity and frequency of this type of car-
bonyl species as a function of the presence or absence of CZ
in the catalyst are apparent. Whereas for PdNiCZ, maximum
intensity of these species was observed immediately on con-
tact with the reactant mixture at 303 K, decreasing in inten-
sity and red-shifting at higher reaction temperature, for PdNiA
the maximum appeared at an intermediate reaction tempera-
ture (ca. 383 K), showing decreasing intensity along with red-

266
A.B. Hungría et al. / Journal of Catalysis 235 (2005) 262–271
shifting at higher temperatures. Of significance was the corre-
lation between the onset temperatures of the red shift of this
species and the CO oxidation activity, in agreement with pre-
vious results for alumina-supported palladium systems [7]. The
PdNiCZA system represents an intermediate case, because it
shows maximum intensity of the atop species at 303 K, whereas
the band appears to be split into two components at intermediate
reaction temperatures (at least up to 423 K) as a consequence of
the varying extent of perturbation. Analysis of the shifted com-
ponent in this sample shows fairly similar positions as those
observed for PdNiCZ. The band splitting is a consequence of
the presence of two types of palladium species interacting with
the two different support components, CZ and A, in agreement
with results presented in Part 1 [1].
The spectra show also the growth (up to a certain, relatively
high, intermediate temperature) of bands due to two differ-
ent bridging (or bridging and three-fold coordinated, respec-
tively) carbonyls chemisorbed on metallic palladium (bands at
−
1
−1
ca. 1970 cm and a broader one centered at ca. 1920 cm
−1
and extending down to ca. 1800 cm , respectively) [7,16–20].
A decrease in intensity along with a red shift of bands due to
this type of carbonyl, likely related mainly to thermal desorp-
tion effects, was produced at high temperature. The decreasing
atop/bridge carbonyl intensity ratio observed with increasing
reaction temperature can be related in part to the increasing size
of the Pd metallic particles produced during the course of the
reaction [7,21,22]. (Different thermal stabilities should be also
considered [17,19,20].) Indeed, this agrees with the fact that
evolution of the overall intensity of metallic palladium carbonyl
species in PdNiA (Fig. 3) reflects that the degree of palladium
reduction increases with reaction temperature. In situ XANES
results presented later in this report are also in agreement with
a gradual palladium reduction process during the course of the
reaction.
Fig. 5. DRIFTS spectra recorded at the end of the run under CO–O after cool-
2
ing down to 303 K under inert gas followed by admission of 3% CO/N flow
2
(
—) and subsequent purging under N (- - -) for the indicated catalysts.
2
lyst, whereas PdNiCZA and PdNiA showed greater similarities
in this respect. This is in line with the smaller palladium parti-
cle size in the final state of the PdNiCZ system, in agreement
with the greater dispersion of palladium observed in the char-
acterization results of Part 1 [1,21,22].
The presence of NO in the reactant mixture significantly hin-
dered the formation of metallic Pd carbonyls (bands at 2100–
−1
2
060 and 1975–1900 cm ), as can be inferred by comparing
Figs. 3 and 4. In turn, such a comparison reveals that the for-
mation of Pd carbonyls (bands at ca. 2160 and 2130 cm
Another carbonyl species detected during the runs under
CO + O2 corresponds to carbonyl species chemisorbed at ox-
n+
−1
)
was promoted by the presence of NO (Fig. 4). The extent of this
effect apparently increased with an increasing amount of CZ in
the catalyst. Indeed, for PdNiCZ, it became particularly difficult
to detect metallic Pd carbonyls during the temperature ramping.
These observations provide evidence of a CZ-promoted oxi-
dizing effect induced by NO on the palladium component, in
line with results reported for the monometallic palladium cata-
lysts [10]. Other bands observed in these spectra were related
n+
−1
idized Pd sites (band at ca. 2148–2142 cm ) [10,20], ob-
served exclusively (at relatively low reaction temperatures)
for the CZ-containing catalysts (Fig. 3). Careful inspection
of the spectra also reveals the presence of a small band at
−1
ca. 2180 cm at low reaction temperatures that was not de-
tected during similar runs performed using the monometallic
Pd systems [7,9]. This band is most likely related to carbonyls
weakly chemisorbed at exposed Ni2+ cations present in the
−
1
to adsorbed isocyanate species (at ca. 2250 and 2231 cm for
NiO-type oxide that is highly dispersed on the catalyst [1,23].
To analyze the state of the catalyst at the end of the runs un-
der CO + O2, the samples were cooled under N2 before CO
−1
A-containing catalysts and a very weak one at ca. 2241 cm
for the PdNiCZ system [7,10,18]), providing evidence for NO
dissociation. The initial appearance of these bands correlates
reasonably well with the onset of NO reduction. In addition,
(3% CO in N2 flow) was admitted into the DRIFTS cell at
3
03 K, then purged under N2 at the same temperature. The re-
−1
a band at ca. 1850 cm appearing exclusively in the presence
of nickel (also seen when the same experiment was performed
using a reference Ni/Al2O3 catalyst [results not shown]) was
observed at relatively low reaction temperatures and can be as-
sults of these experiments are given in Fig. 5. The spectra show
the presence of the linear and bridging carbonyl species on
the metallic palladium particles, carbonyl species chemisorbed
n+
on Pd for the CZ-containing catalysts, and the band due to
2+
−1
2+
weakly adsorbed carbonyls on Ni cations at ca. 2180 cm
,
signed to nitrosyl species chemisorbed on Ni cations [24,25].
which readily disappears on N2 purging. Note that the ratio of
the intensities of atop and bridging carbonyls chemisorbed on
metallic palladium was particularly high for the PdNiCZ cata-
Formation of this species hindered formation of the carbonyl
species chemisorbed on Ni2 cations, in agreement with the
+
higher adsorption energy of NO over such cations [26]. Re-

A.B. Hungría et al. / Journal of Catalysis 235 (2005) 262–271
267
the PdNiA catalyst in this case. In turn, analysis of the metal-
lic palladium chemisorbed atop/bridge carbonyl intensity ratio
is in agreement with the achievement of a smaller palladium
particle size in PdNiCZ after the runs under CO + O2 + NO.
Analysis of the spectral range corresponding to chemisorbed
COx- or NOx-related species (assignments made on the basis of
differences between experiments performed in the presence and
absence of NO) revealed differences as a function of the support
material used (Fig. 7). PdNiCZ showed the presence of bands
attributable to carbonate or carboxylate species (bidentate car-
−1
bonates at ca. 1580 and 1297 cm and monodentate carbonates
or carboxylates at ca. 1506, 1400, and 1332 [27]). A band due
−1
to chelating-nitrite species (appearing at 1185 cm [28]) was
observed at low temperature and had a maximum at 333 K.
The evolution of this feature correlates well with the previ-
ously mentioned adsorption/desorption phenomena which was
responsible for the shape of the NO conversion profile at low
temperature during the catalytic tests (Fig. 2). Nitrate species,
−1
displaying main bands at 1527 and 1273 cm [29–31], are ob-
served at higher reaction temperatures. PdNiA and PdNiCZA
catalysts showed the formation of hydrogencarbonate bands (at
−
1
−1
Fig. 6. DRIFTS spectra recorded at the end of the run under CO–O –NO after
2
1655, 1438, and 1227 cm [32,33]) and a band at 1546 cm
cooling down to 303 K under inert gas followed by admission of 3% CO/N
2
that can be ascribed to a bidentate carbonate species (a band
similar to this one is observed to grow with increasing reaction
temperature under CO + O2), although nitrate species could
also contribute to this band [33]. Bands attributable to nitrate
flow (—) and subsequent purging under N (- - -) for the indicated catalysts.
2
sults (similar to those presented in Fig. 5) of the postreaction
chemisorption of CO after the runs under CO + O2 + NO are
displayed in Fig. 6. The nature of the carbonyls detected is simi-
lar to that of those observed after the run under CO + O2. How-
−1
species (at 1560 and 1308 cm [29,33]) are observed at low re-
action temperature whereas other types of nitrate species (bands
−
1
at 1610, 1578 and 1460 cm [29,33]) appear at relatively high
temperatures. The presence of nickel does not apparently pro-
duce significant differences in the nature or evolution of these
species, on the basis of comparison with results reported for the
monometallic Pd catalysts [7,34]. The carbonate- and nitrate-
type species detected in these catalysts are similar to those
n+
ever, a relatively higher intensity of Pd carbonyls was de-
tected after the runs under CO + O2 + NO, indicating that they
are apparently favored for CZ-containing catalysts. However,
in contrast to results observed after the CO + O2 run (Fig. 5),
a small band due to this type of species was also detected for
Fig. 7. In situ DRIFTS spectra obtained during the course of the CO–O –NO reaction for the indicated catalysts (zone corresponding to carbonate- or nitrate-type
2
species, see text). Spectra taken every 30 K from 303 up to 573 K (from bottom to top).

268
A.B. Hungría et al. / Journal of Catalysis 235 (2005) 262–271
Fig. 8. XANES spectra for the indicated catalysts. (Left) Pd initial, oxidized species; (right) Pd final, reduced species (see text for details).
observed in the metal-free supports and in samples with other
metals deposited over these supports [31,33,35], thus indicat-
ing that they become stabilized at support surface sites free
from significant interactions with either the palladium or nickel
components. In turn, their evolution as a function of the reac-
tion temperature, along with comparison with the respective
catalytic activities observed for the samples (Figs. 1 and 2),
would eliminate any major relevant catalytic role for this type
of species, in accordance with previous reports [7,31].
3.3. In situ XANES
With the aim of analyzing changes in the chemical state of
palladium in the course of the runs under CO + O2 + NO,
including the possibility that formation of Pd–Ni alloys may
occur under such conditions (as observed previously for analo-
gous Pd–Cu catalysts [3,4]), the two catalysts in which nickel
induces apparent changes in catalytic activity were explored by
XANES spectroscopy. Principal component analysis (PCA) of
the Pd K-edge XANES spectra of PdNiCZA and PdNiCZ un-
der CO+O2 +NO suggested the presence of only two different
chemical species during the course of the runs (Fig. 8). Com-
paring the XANES spectra of the two species with reference
compounds indicates that the palladium appears to be present
Fig. 9. Concentration profiles of species detected by XANES during the run
under CO–O –NO for the indicated catalysts.
2
This can be related to the CZ-promoted reduction of palladium
on interaction with CO, in agreement with previous postula-
tions [9,36]. Such a process can in turn be favored by the higher
palladium dispersion in this sample [1].
4. Discussion
initially in the form of an oxidized Pd2 species with a lo-
cal symmetry similar to that exhibited by PdO (D2h symmetry
group). The second Pd-containing species can be readily as-
+
Promoting effects of nickel in the bimetallic Pd–Ni systems
have been shown to depend strongly on the nature of the support
used (Table 1). According to earlier reports, [2,7,9,10], such dif-
ferences must be linked primarily to a series of factors which in
general terms can involve morphologic/structural or electronic
modifications to palladium and/or the CZ active components of
the systems, as well as changes in the nature or the amount of
contacts between both active components (as a consequence of
changes in the palladium distribution over the supports). Re-
sults given in Table 1 discount significant direct involvement of
nickel or alumina components on the catalytic activity, although
0
cribed to a zero-valent Pd fcc phase by comparison with a Pd
foil. Fig. 9 shows the evolution of the fraction of each of these
species as a function of the reaction temperature. The concen-
tration of the Pd2 species decreased gradually from ca. 360
+
and 410 K for PdNiCZ and PdNiCZA, respectively. The re-
0
duced Pd species appear concurrently, and their concentrations
then grows until the reduction process becomes complete at
about 580 and 630 K, respectively. These results reveal a greater
facility for bulk reduction of palladium in the PdNiCZ catalyst.

A.B. Hungría et al. / Journal of Catalysis 235 (2005) 262–271
269
these components may be indirectly involved by inducing mod-
ifications in any of the properties mentioned. In turn, differ-
ences as a function of the reactant mixture used in each case
must be related to the way in which such different conforma-
tions of active components operate with respect to interference
between reactants or self-poisoning effects due to deactivation
induced on interactions among chemisorbed species [10].
On this basis, the absence of promotional effects of nickel
in the PdNiA catalyst indicates that NiO-type species present
in the system, according to XPS and XANES results presented
in Part 1 [1], do not induce modifications in the characteris-
tics of the active palladium particles and are not involved in
phenomena of reactant chemisorption that could affect the cat-
alytic activity of palladium. This latter finding agrees with the
limited influence of the nickel entities on the formation of NO-
or CO-derived chemisorbed species at temperatures relevant for
the catalytic activity, according to the results described earlier.
In fact, only weakly chemisorbed carbonyl or nitrosyl species
appearing at low temperature could be identified as directly re-
lated to the presence of nickel in the catalysts. Such species
certainly would not be expected to play a significant role in
aspects that could induce important catalytic changes, like pro-
ducing a modification of the characteristics of the reactant mix-
ture at a local level or activating the reactants in any sense. It
must be noted, however, that direct comparison with results of
Yamamoto et al. [5,6], who claimed an involvement of nickel
species in chemisorption phenomena that could facilitate the
palladium activity for this type of reaction [6], yields limited in-
formation because of at least two physicochemical differences
related to the Ni component. First, their catalyst configuration
appears to present nickel fully in the form of a nickel alumi-
nate, according to XRD evidence. In contrast, a mixture of
highly dispersed NiO- and NiAl2O4-type entities is apparently
present in the catalysts examined in this work [1]. Second, we
use a considerably lower amount of nickel, because in fact opti-
mum properties for the reactions examined here are apparently
achieved for such low nickel loading (Table 1). But possible
influences of nickel on palladium at structural/morphologic or
electronic levels can be considered as residual in our case, on
the basis of XANES (Figs. 8 and 9) and STEM analyses [1].
Therefore, the catalytic properties of the PdNiA catalyst can
be assimilated into those of the analogous PdA system, which
were examined in detail in a previous study [7]. Briefly, the
results observed for the CO + O2 reaction are compatible with
the typical Langmuir–Hinshelwood kinetic scheme in which the
reaction is initially limited by thermal desorption of CO from
the metallic Pd particles whose surface atoms constitute the ac-
tive centers for the reaction [7,20,37,38]. This is based mainly
on the correlation observed between the evolution of the fre-
quency of the atop carbonyl species chemisorbed on metallic
Pd particles and the catalytic activity (Fig. 3). Thus the rela-
tively high frequency before reaction onset is due to dipolar
coupling effects, which mainly govern the frequency of metal-
lic palladium carbonyls (those corresponding to atop species
being most sensitive to coupling) [39] in palladium particles
highly covered with CO. The simultaneous triggering of CO
oxidation and the red shift of the atop carbonyl reflects that
desorption of CO allows the reaction to proceed over the par-
ticles by leaving sites free for dissociative oxygen adsorption
and further combination of the oxygen atoms with chemisorbed
CO molecules [7,37,38]. Within this scheme, the irregulari-
ties observed in the CO conversion profile for this catalyst (at
temperatures < ca. 440 K, giving rise to a kind of staircase-
type shape; see Fig. 1) can be attributed to the increase in the
number of active sites resulting from the progressive growth of
metallic particles during the course of the reaction (in accor-
dance with DRIFTS results presented in Fig. 3). The inhibiting
effect of NO on CO oxidation activity (Table 1) can be corre-
lated mainly with the hindering of metallic palladium carbonyl
formation observed in the presence of NO (compare Figs. 3
and 4), indicating the greater difficulty in generating active re-
duced palladium states at the surface of the particles [7]. In
turn, the absence of significant effects of nickel on NO reduc-
tion (Table 1) further confirms the similar nature of the metallic
palladium particles formed during the course of the reaction for
PdA and PdNiA, given the strong structural sensitivity of such
a reaction [7,40,41].
In contrast to the A-supported catalysts, the presence of
nickel significantly affects the catalytic performance of the CZ-
supported system (Table 1). The decreased CO oxidation ac-
tivity and enhanced NO reduction activity observed for the
bimetallic PdNiCZ system with respect to the monometallic
PdCZ system, along with the absence of significant nickel-
induced electronic modifications of palladium (according to
XANES analysis; Figs. 8 and 9), suggest that contacts be-
tween Pd and the CZ are to a certain (probably small) extent
blocked by nickel entities (most likely in the form of NiO-type
species [1]); in fact XPS results presented in Part 1 suggested
that some interactions between Pd and Ni components can oc-
cur in this catalyst [1]. It must be taken into account that the
high CO oxidation activity achieved for PdCZ must be mainly
a consequence of the existence of contacts between both cat-
alyst components [2,9]. In the presence of such contacts, the
direct participation of interfacial oxide anions at the surface of
the CZ support in CO oxidation processes allows the open-
ing up of a new reaction path operating at low temperature
(<423 K), which circumvents the CO-inhibiting effects ob-
served in A-supported catalysts [2,9,11]. In turn, NO reduction
appears to be strongly inhibited in the presence of such con-
tacts as a consequence of the passivation of the active sites at
the surface of metallic Pd particles due to the CZ-promoted
formation of oxidized states of palladium (in agreement with
DRIFTS results described earlier; see Fig. 4). Such an effect
is considered mainly responsible for the poorer response of
CZ-supported palladium catalysts toward NO reduction [10],
in agreement with results presented in Table 1. On this ba-
sis, a certain nickel-induced suppression of the establishment
of contacts between Pd and CZ can explain the observed cat-
alytic activity differences between PdNiCZ and PdCZ (Table 1).
However, another hypothesis that might explain these results
could be related to differences (even if small) in the respec-
tive palladium dispersions (i.e., a relatively larger palladium
crystallite size in PdNiCZ; in any case, within the context of
very highly dispersed states for both systems). Unfortunately,

270
A.B. Hungría et al. / Journal of Catalysis 235 (2005) 262–271
the similar DRIFTS results observed for both samples (results
for the monometallic system available elsewhere [10]) does
not allow a distinction between these hypotheses, although it
highlights the subtle character of the differences between both
systems.
Concerning the activity of PdNiCZA when NO is present in
the reactant mixture, the data show an important NO-induced
inhibition of CO oxidation, stronger than that observed for the
PdNiCZ catalyst (Figs. 1 and 2; Table 1). The behavior of
PdNiCZA in this respect (and in contrast to PdNiCZ, as dis-
cussed earlier) is more representative of the presence of pure
Pd–CZ contacts than those present in PdCZ, which also shows
a large inhibiting effect of NO [10], Table 1. According to
previous studies [10], such an inhibiting effect can be related
to the previously mentioned CZ-promoted formation of oxi-
dized states of palladium (according to DRIFTS results), which
would impede CO adsorption and activation, and/or to NO in-
terference with oxygen adsorption, leading to reoxidation of
interfacial vacancies of CZ [10,28]. However, it is interesting
that the PdNiCZA catalyst still maintains a level of NO reduc-
tion activity similar to that achieved by the analogous PdCZA
reference (Table 1). Because, as mentioned earlier, poorer NO
reduction activity would be expected for palladium sites in con-
tact with the CZ component [10], this result suggests a different
nature for the sites active for such reaction. In fact, palladium
particles in contact with the alumina component are shown to
be most active for this reaction (Table 1), whereas interactions
of these with highly dispersed two-dimensional CZ entities can
also favor such processes [10]. In this sense, no significant
particle size differences were apparent for such noble metal
particles, in comparison with the other CZA- or A-supported
systems of the analogous series, according to the analysis of
Z-contrast images [1]. Nevertheless, considering the prevail-
ing interaction between palladium and CZ in this catalyst, the
nickel-induced morphologic changes in palladium particles in
contact with the CZ component (in particular, achievement of
a larger particle size [1]) can also have an important role in the
NO reduction activity of this catalyst [2,7].
The most striking promoting effects of nickel appear for
the CZA-supported catalyst and are particularly reflected in the
important improvement in CO oxidation activity observed for
PdNiCZA (Table 1). In the absence of appreciable electronic
modifications of palladium (according to XANES results;
Figs. 8 and 9), the most likely hypothesis to account for this
effect must be related to the nickel-induced changes in the pal-
ladium distribution over the support. Thus, on the basis mainly
of the XEDS results presented in Part 1 [1], it has been ob-
served that the presence of nickel favors the formation of con-
tacts between Pd and CZ components of this system. Interest-
ingly, in contrast to observations in the CZ-supported catalysts,
such contacts might not necessarily involve highly dispersed
palladium particles, because the particle size of palladium
particles appears fairly close to that observed for the A-sup-
ported catalysts (on the basis of DRIFTS results discussed ear-
lier and STEM-EELS observations presented in Part 1 [1]). In
addition, it must be considered that a part of the palladium must
be in contact with the alumina component of the catalyst (ac-
cording to DRIFTS and STEM-EELS results [1]) but does not
participate in the low-temperature CO oxidation activity of this
system, in accordance with the absence of frequency shifts for
the corresponding atop carbonyl species (Fig. 3). Taking these
points into account, the fact that the PdNiCZA catalyst displays
a CO oxidation catalytic activity similar to that of PdNiCZ, in
which palladium appears highly dispersed and the amount of
contacts between Pd and CZ components (even if somewhat
damped by nickel entities according to the discussion above)
is maximized, indicates an important influence of CZ parti-
cle size on such activity. In brief, contacts established between
Pd and CZ particles in PdNiCZA (average CZ particle size of
ca. 3.5 nm [1]) appear to be appreciably more active for CO
oxidation than those present in PdNiCZ (average CZ particle
size ≈ 5.5 nm [1]). This suggests an important role of Pd–CZ
interfacial properties on such activity, as has been proposed
for other CZ- or ceria-supported metal catalysts [31,42–44].
Size-dependent differences in the chemical activity of oxide
nanoparticles can be related to various factors, as has been dis-
cussed previously [45,46]. These factors can include electronic
changes (e.g., changes in band gap size or introduction of elec-
tronic states within the oxide band gap, redistribution of atomic
charges) and structural or geometrical changes (e.g., stabiliza-
tion of determinate crystalline, or even amorphous, phases,
particularly at the oxide surface; changes in the amount or dis-
tribution of defects). In the case of ceria-related oxides, it has
been observed that decreasing the size of the nanoparticles can
decrease the reduction temperature of the oxide [47], in agree-
ment with theoretical predictions [48], and also enhance the
electronic conductivity [49,50]. Both aspects can play an im-
portant role in the CO oxidation activity of Pd–CZ contacts,
taking into account the direct participation of oxide anions of
the support in the catalytic process.
5. Conclusions
Pd–Ni catalysts supported on Al2O3, (Ce,Zr)Ox/Al2O3, and
(Ce,Zr)Ox have been examined with regard to their catalytic
activity for CO oxidation and NO reduction under stoichiomet-
ric CO–O2 and CO–O2–NO mixtures, focusing mainly on the
effects induced by the presence of nickel on the catalytic activ-
ity. These are shown to depend strongly on the support used.
Whereas practically no differences are detected for the Al2O3-
supported bimetallic catalyst in comparison with the analogous
monometallic palladium system [7,9,10], apparent differences
appear for (Ce,Zr)Ox-containing catalysts. Because no apparent
modification of the palladium electronic properties (as would
be expected on, for instance, formation of alloys with nickel)
were detected by in situ XANES experiments, and no direct
involvement on the catalytic processes could be attributed to
the highly dispersed oxidized nickel entities present in the ca-
talysts [1], the changes observed are related to indirect effects
involving modifications in the catalyst configurations with re-
spect to the corresponding monometallic palladium references.
The lower CO oxidation and higher NO reduction activities
observed for the (Ce,Zr)Ox-supported system suggest that ac-
tive contacts between highly dispersed Pd and (Ce,Zr)Ox are

A.B. Hungría et al. / Journal of Catalysis 235 (2005) 262–271
271
modified either through direct blocking by NiO-type entities or
through a certain decrease in the palladium dispersion. An im-
portant promotion of CO oxidation activity is observed for the
[16] R. Raval, G. Blyholder, S. Haq, D.A. King, J. Phys.: Condens. Matter 1
1989) SB165.
(
[
[
[
17] X. Xu, P. Chen, D.W. Goodman, J. Phys. Chem. 98 (1994) 9242.
18] K. Almusaiteer, S.C. Chuang, J. Catal. 184 (1999) 189.
19] I.V. Yudanov, R. Sahnoun, K.M. Neyman, N. Rösch, J. Hoffmann, S.
Schauermann, V. Johánek, H. Unterhalt, G. Rupprechter, J. Libuda, H.-J.
Freund, J. Phys. Chem. B 107 (2003) 255, and references therein.
20] X. Xu, D.W. Goodman, J. Phys. Chem. 97 (1993) 7711.
21] K. Wolter, O. Seiferth, H. Kuhlenbeck, M. Bäumer, H.-J. Freund, Surf.
Sci. 399 (1998) 190.
(Ce,Zr)Ox/Al2O3-supported catalyst as a result of Ni-induced
changes in the palladium distribution over the support, which
favors the generation of active Pd–(Ce,Zr)Ox contacts [1]. The
significant extent of such a promoting effect reveals an impor-
tant role of the particle size of the (Ce,Zr)Ox component on
such catalytic processes.
[
[
[
[
22] M. Frank, M. Bäumer, Phys. Chem. Chem. Phys. 2 (2000) 3723.
23] C. Xu, Q. Guo, D.W. Goodman, Catal. Lett. 41 (1996) 21.
Acknowledgments
[24] G. Niu, Y. Huang, Z. Cao, Y. Huang, Q. Li, Appl. Surf. Sci. 141 (1999)
5.
3
[
25] J.B. Peri, J. Catal. 86 (1984) 84.
The authors thank the technical and scientific staff at the
ESRF BM-29 (Dr. G. Subías) and SRS 9.3 (Drs. A.R. Lennie
and I. Harvey) synchrotron beamlines, and Dr. A. Iglesias-Juez
for their help in the recording of XAFS data. They also thank
Professor J.C. Conesa for helpful discussions. A.B.H. thanks
the Comunidad de Madrid for a doctoral studies grant and for
financial support through the “Ayudas para estancias en centros
extranjeros” program. Support for this work was also provided
by the CICyT (project MAT2000-1467).
[26] A. Zecchina, D. Scarano, S. Bordiga, G. Ricchiardi, G. Spoto, F. Teobaldo,
Catal. Today 27 (1996) 403.
[27] C. Li, Y. Sakata, T. Arai, K. Domen, K.-I. Maruya, T. Onishi, J. Chem.
Soc., Faraday Trans. I 85 (1989) 929.
[
28] A. Martínez-Arias, J. Soria, J.C. Conesa, X.L. Seoane, A. Arcoya, R. Cata-
luña, J. Chem. Soc., Faraday Trans. 91 (1995) 1679.
[
29] A.A. Davydov, in: C.H. Rochester (Ed.), Infrared Spectroscopy of Ad-
sorbed Species on the Surface of Transition Metal Oxides, Wiley, New
York, 1990, p. 75.
[30] K. Hadjiivanov, V. Avreyska, D. Kissurski, T. Marinova, Langmuir 18
2002) 1619.
(
[31] A. Martínez-Arias, M. Fernández-García, A.B. Hungría, A. Iglesias-Juez,
O. Gálvez, J.A. Anderson, J.C. Conesa, J. Soria, G. Munuera, J. Catal. 214
References
(
2003) 261.
[
1] A.B. Hungría, N.D. Browning, R.P. Erni, M. Fernández-García, J.C. Co-
nesa, J.A. Pérez-Omil, A. Martínez-Arias, J. Catal., in press, 10.1016/
j.jcat.2005.08.011.
[
[
32] A.M. Turek, I.E. Wachs, E. DeCanio, J. Phys. Chem. 96 (1992) 5000.
33] A. Martínez-Arias, M. Fernández-García, A. Iglesias-Juez, J.A. Anderson,
J.C. Conesa, J. Soria, Appl. Catal. B 28 (2000) 29.
34] A. Iglesias Juez, Ph.D. Thesis, Universidad Autónoma de Madrid (2005).
35] A. Martínez-Arias, A.B. Hungría, M. Fernández-García, J.C. Conesa,
G. Munuera, J. Power Sources (in press).
[
2] A. Martínez-Arias, J.C. Conesa, M. Fernández-García, J.A. Anderson, in:
J.A. Anderson, M. Fernández-García (Eds.), Supported Metals in Cataly-
sis, Imperial College Press, London, 2005, p. 283.
[
[
[3] M. Fernández-García, A. Martínez-Arias, C. Belver, J.A. Anderson, J.C.
Conesa, J. Soria, J. Catal. 190 (2000) 387.
[
36] M. Fernández-García, A. Martínez-Arias, L.N. Salamanca, J.M. Coron-
ado, J.A. Anderson, J.C. Conesa, J. Soria, J. Catal. 187 (1999) 474.
[
4] A.B. Hungría, A. Iglesias-Juez, A. Martínez-Arias, M. Fernández-García,
J.A. Anderson, J.C. Conesa, J. Soria, J. Catal. 206 (2002) 281.
5] S. Yamamoto, K. Matsushita, Nippon Kagaku Kaishi (2000) 553.
6] S. Yamamoto, K. Matsushita, Y. Hanaki, Nippon Kagaku Kaishi (2001)
[37] T. Engel, G. Ertl, Adv. Catal. 28 (1979) 2.
[38] J. Libuda, I. Meusel, J. Hoffmann, J. Hartmann, L. Piccolo, C.R. Henry,
H.-J. Freund, J. Chem. Phys. 114 (2001) 4669.
[
[
[39] P. Hollins, Surf. Sci. Rep. 16 (1992) 51.
1
9.
[40] D.R. Rainer, S.M. Vesecky, M. Koranne, W.S. Oh, D.W. Goodman, J. Ca-
tal. 167 (1997) 234.
[
[
[
7] A. Martínez-Arias, A.B. Hungría, M. Fernández-García, A. Iglesias-Juez,
J.A. Anderson, J.C. Conesa, J. Catal. 221 (2004) 85.
8] R. van Yperen, D. Lindner, L. Mubmann, E.S. Lox, T. Kreuzer, Stud. Surf.
Sci. Catal. 116 (1998) 51.
9] M. Fernández-García, A. Martínez-Arias, A. Iglesias-Juez, A.B. Hungría,
J.A. Anderson, J.C. Conesa, J. Soria, Appl. Catal. B 31 (2001) 39.
[41] L. Piccolo, C.R. Henry, J. Mol. Catal. A 167 (2001) 181.
[42] J.G. Nunan, H.J. Robota, M.J. Cohn, S.A. Bradley, J. Catal. 133 (1992)
309.
[43] A. Martínez-Arias, A.B. Hungría, M. Fernández-García, J.C. Conesa,
G. Munuera, J. Phys. Chem. B 108 (2004) 17983.
[
10] A. Martínez-Arias, M. Fernández-García, A. Iglesias-Juez, A.B. Hungría,
J.A. Anderson, J.C. Conesa, J. Soria, Appl. Catal. B 31 (2001) 51.
11] M. Shelef, G.W. Graham, R.W. McCabe, in: A. Trovarelli (Ed.), Catalysis
by Ceria and Related Compounds, Imperial College Press, London, 2002,
p. 343.
[44] S. Carrettin, P. Concepcion, A. Corma, J.M.L. Nieto, V.F. Puntes, Angew.
Chem.-Int. Ed. 43 (2004) 2538.
[
[45] M. Fernández-García, A. Martínez-Arias, J.C. Hanson, J.A. Rodríguez,
Chem. Rev. 104 (2004) 4063, and references therein.
[46] R. Di Monte, J. Ka sˆ par, J. Mater. Chem. 15 (2005) 633.
[47] J.A. Rodriguez, J.C. Hanson, J.-Y. Kim, G. Liu, A. Iglesias-Juez, M. Fer-
nández-García, J. Phys. Chem. B 107 (2003) 3535.
[
12] R. Di Monte, J. Ka sˆ par, Catal. Today 100 (2005) 27.
[13] R. Cataluña Veses, Ph.D. Thesis, Universidad Politécnica de Madrid
(
2005).
[48] H. Cordatos, D. Ford, R.J. Gorte, J. Phys. Chem. 100 (1996) 18128.
[49] Y.-M. Chiang, E.B. Lavik, I. Kosacki, H.L. Tuller, J.Y. Ying, J. Electroce-
ram. (1997) 7.
[
[
14] M. Fernández-García, Catal. Rev. Sci. Eng. 44 (2002) 59.
15] C.N. Costa, S.Y. Christou, G. Georgiou, A.M. Efstathiou, J. Catal. 219
(2003) 259.
[50] A. Tschöpe, E. Sommer, R. Birringer, Solid State Ionics 139 (2001) 255.