Behavior of bimetallic Pd-Cr/Al2O3 and Pd-Cr/(Ce,Zr)Ox/Al2O3 catalysts for CO and NO elimination

Article abstract of DOI:10.1016/S0021-9517(02)00145-8

A series of Pd-Cr bimetallic catalysts supported on a (Ce,Zr)O _x/Al₂O₃ mixed support or Al₂O ₃ alone have been characterized using a combination of X-ray diffraction, electron paramagnetic resonance, and Raman spectroscopy and employing in situ diffuse reflectance infra-red Fourier transform and X-ray near-edge structure spectroscopy to analyze the redox and chemical processes taking place during the course of the CO+NO+O₂ reaction. The catalytic behavior of these bimetallic systems was strongly affected by the nature of the support. In the case of the alumina support, evidence for interactions between Pd and Cr was observed in the calcined state. Under reaction conditions, formation of a mixed oxide phase containing Pd(I) and Cr(III) appears to be responsible for the improvement in CO oxidation and for the detrimental effect in the NO reduction process with respect to a monometallic, Pd reference system. In the case of the (Ce,Zr)O _x/Al₂O₃ mixed support, the addition of Cr was less influential in terms of catalytic activity. Results for the latter are rationalized mainly on the basis of the absence of interaction between the metal components in the calcined state and the presence of surface entities containing Cr alone when metallic-like Pd particles are formed during the course of reaction.

Full text of DOI:10.1016/S0021-9517(02)00145-8

Journal of Catalysis 214 (2003) 220–233
www.elsevier.com/locate/jcat
Behavior of bimetallic Pd–Cr/Al O and Pd–Cr/(Ce,Zr)O /Al O
2 3
2
3
x
catalysts for CO and NO elimination
a,∗
a
a
a
b
M. Fernández-García, A. Martínez-Arias, A. Iglesias-Juez, A.B. Hungría, J.A. Anderson,
a
a
J.C. Conesa, and J. Soria
a
Instituto de Catálisis y Petroleoquímica, CSIC, Campus Cantoblanco, 28049 Madrid, Spain
b
Division of Physical and Inorganic Chemistry, University of Dundee, DD14HN, Dundee, Scotland, UK
Received 29 July 2002; revised 16 October 2002; accepted 24 October 2002
Abstract
A series of Pd–Cr bimetallic catalysts supported on a (Ce,Zr)Ox/Al O mixed support or Al O alone have been characterized using
2
3
2 3
a combination of X-ray diffraction, electron paramagnetic resonance, and Raman spectroscopy and employing in situ diffuse reflectance
infra-red Fourier transform and X-ray near-edge structure spectroscopy to analyze the redox and chemical processes taking place during the
course of the CO + NO + O reaction. The catalytic behavior of these bimetallic systems was strongly affected by the nature of the support.
2
In the case of the alumina support, evidence for interactions between Pd and Cr was observed in the calcined state. Under reaction conditions,
formation of a mixed oxide phase containing Pd(I) and Cr(III) appears to be responsible for the improvement in CO oxidation and for the
detrimental effect in the NO reduction process with respect to a monometallic, Pd reference system. In the case of the (Ce,Zr)Ox/Al O
2
3
mixed support, the addition of Cr was less influential in terms of catalytic activity. Results for the latter are rationalized mainly on the basis
of the absence of interaction between the metal components in the calcined state and the presence of surface entities containing Cr alone
when metallic-like Pd particles are formed during the course of reaction.

2003 Elsevier Science (USA). All rights reserved.
Keywords: Pd–Cr bimetallic catalysts; Pd–Cr mixed oxides; Pd–Cr alloys; CeO –ZrO ; Al O ; CO oxidation; NO reduction; XRD; TEM-XEDS; EPR;
2
2
2 3
in situ DRIFTS; in situ XANES
1
. Introduction
pacity (OSC) after thermal sintering, which could potentially
decrease the cold-start emissions, mainly by allowing the
catalyst to be located in positions closer to the engine mani-
fold with minimum system deactivation being produced [4].
Optimum promoter properties are achieved at Ce/Zr atomic
ratios close to one and when pseudocubic (tetragonal sym-
metry for oxygen sublattice) or cubic structures of the mixed
oxide are formed, which are presumably stabilized as nano-
sized particles [4–7]. When supported on alumina, these
chemical/structural characteristics have been optimized by
employing synthesis methods based on reverse microemul-
sions [7], rather than more classical methods based on co-
impregnation [4,8,9]. The advantages induced by the use of
adequate preparation methods are particularly evident when
the supports are subjected to aging treatments, which de-
grade the tetragonal structure of mixed oxides with Ce:Zr
atomic ratio close to 1 in a disproportion reaction which pro-
duces two (Ce-rich and Zr-rich) phases [7–10].
Three-way catalysts (TWC) have been widely used to di-
minish pollutant emissions from gasoline engine powered
vehicles [1]. Classical components of these systems usually
include Rh, Pt, and/or Pd as active metals, ceria as promoter,
and high-surface alumina as support [1,2]. More recently, the
classical promotion by ceria has been extended to include
other oxide components in order to increase or maintain the
durability of the TWC while decreasing the toxic emissions
produced during the cold-start (or light-off) period, which
may represent a considerable portion of the total emissions
produced during any driving cycle [1,3]. Among the latter,
Ce–Zr mixed oxide systems have been considered as substi-
tutes for ceria on the basis of their greater oxygen storage ca-
*
Corresponding author.
E-mail address: m.fernandez@icp.csic.es (M. Fernández-García).
0021-9517/03/$ – see front matter  2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0021-9517(02)00145-8

M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
221
The use of Pd as the single active metal component
2. Experimental
.1. Catalyst preparation
A ceria–zirconia/alumina support (CZA) of 10 wt%
in TWCs has recently received considerable attention as a
result of the high cost and scarcity of Rh, the availability
of cleaner fuels, and its remarkable activity for oxidation
reactions [1,11]. However, some limitations are apparent
for Pd-only systems with respect to their performance in
NO reduction reactions [11–13]. It would be desirable to
promote Pd with lower cost base metals instead of using
Rh as an additional active metal, despite the latter’s well-
known activity for NO elimination [14]. As detailed in the
review by Coq and Figueras [15], promising results have
been obtained by using Mo [16,17], Mn [18], Cr [19], or
Cu [20–22], which have been attributed to the formation
of the corresponding alloys (with consequent perturbation
of the Pd electronic properties) or, in the case of Mo
2
of mixed oxide—expressed as Ce0.5Zr0.5O2—was prepared
by modification of the microemulsion method used previ-
ously for preparing the unsupported oxide [6]. Full details
of the procedure can be found elsewhere [7]. After dry-
ing overnight at 373 K, this support was calcined under
2
−1
air at 773 K for 2 h (SBET = 186 m g ). According to
ICP-AES chemical analysis, it has a Zr/Ce atomic ratio of
1
.0 ± 0.1. This CZA support was coimpregnated (incipient
wetness method) with aqueous solutions of palladium (II)
and chromium (III) nitrates (from Alfa Aesar and Merck,
respectively; purities > 99.99%) in an amount to give 1.0
wt% of the precious metal and varying quantities (0.25, 0.5,
and 1 wt%) of the base metal. These amounts correspond to
Pd:Cr atomic ratios of 0.5, 1, and 2 respectively. Monometal-
lic references containing 1 wt% of Pd or Cr were also pre-
pared. Three additional systems of 1 wt% of Pd, 1 wt% of Cr,
and the corresponding bimetallic with 1 wt% of each metal
were also prepared on the parent alumina (A; from Condea
SBET = 180 m g ). The catalysts were calcined follow-
ing the same drying/calcination procedure described above
for the supports and are referred to as PdA, PdCZA, CrA,
CrCZA, PdxxCrA, and PdxxCrCZA, where xx denotes the
weight percentage of Cr.
or Mn, to a mixture of effects involving alloy formation
δ+
(
in the case of Mn) and/or the existence of Pd–Mo
,
δ+
Pd–Mn interactions where the redox capabilities of the
non-noble metal facilitate both CO activation and NO
dissociation. In the case of copper, the promoting effect
on both CO oxidation and NO reduction activities induced
by alloying may be explained, according to theoretical
studies [23], by the strong modification produced in the Pd
valence electronic density by charge injection into the sp
subband.
2
−1
The case of Cr is particularly interesting as the interaction
between the two active components, Pd and Cr, is present
in both the oxidized and reduced states. As in the case of
Cu [24], the existence of Pd–Cr mixed oxides [25] and al-
loys [26] is well known. However, a drastic difference can
be observed between these two binary systems when the
existence and stability of zerovalent solid solution phases,
which are the common phases detected in TWC systems
during treatment with reductive or stoichiometric mixtures
are compared [19,20,22,24]. While complete mixing is ob-
tained in the Pd–Cu case, Pd–Cr displays limited solubility
in a broad region around equimolar concentration at work-
ing temperatures (below 873 K). It is clear that the extent
of interaction between Pd–Cr in TWCs is largely unknown
and requires attention. In this work, the objective was to ad-
dress this aspect using ceria–zirconia promoted Pd–Cr cat-
alysts, focusing on determining which are the main phys-
ical/chemical parameters affecting catalytic performance in
reactions of interest for the TWC. A series of Pd–Cr bimetal-
lic catalysts supported on Al2O3 and on (Ce,Zr)Ox/Al2O3
mixed supports have been studied using a multitechnique ap-
proach involving XRD, TEM-XEDS, Raman, and EPR for
catalyst characterization and catalytic activity in combina-
tion with in situ DRIFTS and XANES techniques to evalu-
ate the catalytic performance, along with the physicochemi-
cal processes which take place during the course of the stoi-
chiometric CO + O2 + NO reaction.
2.2. Catalytic tests
Catalytic tests using stoichiometric mixtures of 1% CO +
−1
0
.45% O2 + 0.1% NO (N2 balance) at 30,000h were
performed in a Pyrex glass flow reactor system. Details of
the experimental conditions employed for these tests can
be found elsewhere [27]. Gases were regulated with mass
flow controllers and analyzed using an on-line Perkin–Elmer
1
725X FTIR spectrometer coupled with a multiple reflec-
tion transmission cell (Infrared Analysis Inc.; path length 2.4
m). Oxygen concentrations were determined using a para-
magnetic analyzer (Servomex 540A). Experimental error in
the CO and/or NO conversion values obtained in these con-
ditions is estimated as ± 7%. Prior to catalytic testing, in
situ calcination was performed in diluted oxygen (2.5% O2
in N2) at 773 K, followed by cooling under the same at-
mosphere and subsequent N2 purging at room temperature
(
RT). A typical test consisted of increasing the temperature
−1
from 298 to 823 K at 5 K min
.
2
.3. Characterization techniques
Powder X-ray diffraction (XRD) patterns were recorded
on a Siemens D-500 diffractometer using nickel-filtered
Cu-Kα radiation operating at 40 kV and 25 mA and
◦
with a 0.025 step size. Transmission electron microscopic
(TEM) analysis of materials was done with a JEOL 200

222
M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
5 K min 1 temperature ramp up to 673 K, in the presence
of the CO + NO + O2 flowing mixture (similar to that em-
ployed for catalytic activity tests).
−
FX (0.31-nm point resolution) equipped with a LINK
(
(
AN10000) probe for energy dispersive X-ray spectroscopy
XEDS).
Raman spectra were acquired using a Renishaw Dis-
Results obtained in XANES experiments were subjected
to normalization using standard procedures and analyzed
using principal component analysis (PCA) [28–30], which
assumes that the absorbance in a set of spectra can be
mathematically modeled as a linear sum of individual
components, called factors, which correspond to each of
the palladium or chromium species present in a sample
plus noise [31]. To determine the number of individual
components, an F-test based on the variance associated with
factor k and the summed variance associated with the pool
of noise factors is performed. A factor is accepted as a
“pure” chemical species (i.e., a factor associated with signal
and not noise) when the percentage of significance level of
the F-test, %SL, is lower than a test level set in previous
studies at 5% [28,29]. The ratio between the reduced
eigenvalues, R(r), which approaches one for noise factors,
was also used in determining the number of factors. Once the
number of individual components was set, XANES spectra
corresponding to individual Pd or Cr species and their
concentration versus temperature profiles were generated
by an orthogonal rotation (varimax rotation), which should
align factors (as closely as possible) along the unknown
concentration profiles, followed by iterative transformation
factor analysis (ITFA). ITFA starts with delta function
representations of the concentration profiles, associated with
each chemical species, located at temperatures predicted by
the varimax rotation, which are then subjected to refinement
by iteration until error in the resulting concentration profiles
is lower than the statistical error extracted from the set of
raw spectra [28–30].
persive system 1000, equipped with a single monochroma-
tor, a holographic Notch filter, and a cooled TCD. Sam-
ples were excited using the 415 nm Ar line in an in situ
cell (Linkam TS-1500), which allowed treatment under con-
trolled atmospheres. Samples were analyzed as stored and
after drying at 773 K using synthetic air; spectra consisted
of 100 accumulations during 15 min acquisition time, using
a typical running power of 25 mW.
EPR spectra were recorded at 77 K and room temperature
(
RT) with a Bruker ER 200 D spectrometer operating in
the X-band and calibrated with a DPPH standard (g =
.0036). Portions of ca. 40 mg of sample were placed inside
2
a quartz probe with greaseless taps, calcined at 773 K
for 1 h under 200 Torr of oxygen, evacuated at RT (1 ×
−4
1
0
Torr), and measured at liquid nitrogen temperature.
Quantitative estimation of the amount of paramagnetic
species was performed by double integration of the spectra
and comparison with a copper sulfate pentahydrate standard.
DRIFTS analysis of adsorbed species present on the
catalyst surface under reaction conditions was carried out
using a Perkin–Elmer 1750 FTIR fitted with an MCT
detector. Analysis of the NO conversion at the outlet of
the IR chamber was performed by chemiluminescence
(
Thermo Environmental Instruments 42C). Additional post-
catalyst gas analysis was performed using a Baltzers Prisma
Quadropole mass spectrometer. The DRIFTS cell (Harrick)
was fitted with CaF2 windows and a heating cartridge that
allowed samples to be heated to 773 K. Samples of ca. 80 mg
were calcined in situ at 773 K (with synthetic air, 20% O2
in N2) and then cooled to 298 K in synthetic air before the
introduction of the reaction mixture and subsequent heating
−1
−1
at 5 K min up to 673 K, recording one spectrum (4 cm
3. Results
resolution, average of 20 scans) generally every 10–15 K.
The gas mixture (using the same concentrations as those
employed for laboratory reactor tests) was prepared using a
computer-controlled gas blender with 75 cm min passing
through the catalyst bed.
3.1. Activity results
3
−1
The CO and NO conversion profiles for Pd and Pd–Cr
catalysts containing 1 wt% of Cr are presented in Figs. 1A
and 1B, respectively. A CrCZA reference system has also
been included for comparison purposes. Below 350 K, con-
version of NO corresponds to adsorption/desorption phe-
nomena, as calculated by mass balance of the outlet gases.
NO2 was never detected after contact of the reaction mix-
ture with the samples. Concerning N-selectivity of the re-
duction process, a maximum of N2O of roughly 30% was
obtained for all catalysts at temperatures below full conver-
sion. For PdA, PdCZA, and PdxCrCZA full selectivity to
N2 is reached around 480–520 K, while Pd1CrA only gets
to that point above 580 K. With the exception of PdA, CO
oxidation over all samples proceeds at a lower temperature
than NO reduction. A comparison of light-off behavior for
all samples may be obtained by comparing the temperatures
(Table 1) required to attain 50% conversion. As can be de-
XANES experiments at the Cr and Pd K-edges were per-
formed on line 7.1 at SRS, Daresbury (Cr K-edge), and on
line BM29 of the ESRF synchrotron at Grenoble (Pd K-
edge). Si(111) (for Cr) and Si(311) (for Pd) double-crystal
monochromators were used in conjunction with a Pt-coated
focalized mirror (7.1) or rejection mirror (BM29) to mini-
mize the harmonic content of the beam. Additionally, a de-
tuning of 30% was used for experiments conducted on Sta-
tion 7.1. Transmission experiments were carried out using
N2/O2 (Cr) or Kr/Ar-filled (Pd) ionization chambers. The
energy scale was simultaneously calibrated by measuring
the corresponding metal foil inserted before a third ioniza-
tion chamber. Samples as self-supported discs (absorbance
0
.5–2.0) were placed in a controlled-atmosphere cell for
treatment. XANES spectra were taken every 15 K during a

M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
223
Fig. 2. Raman spectra of CrA (a), Pd1CrA (b), CrCZA (c), and Pd1CrCZA
d) samples. Dashed lines, hydrated samples; full lines, calcined samples.
(
supported samples, with a moderate decrease in CO light-
off and negligible influence on NO reduction. In this case,
Cr loading had only a slight influence on CO conversion and
did not influence NO reduction performance. Cr-only cat-
alysts display poor performance in both reactions, the best
performance being achieved using CrCZA, Fig. 1.
3
.2. Sample characterization
Calcined samples were analyzed using XRD, TEM-
XEDS, Raman, and EPR spectroscopies. A summary of
XRD/TEM results is presented in Table 2 and mainly re-
flects information concerning the ceria–zirconia (CZ) com-
ponent; significant differences were not apparent from com-
paring the Cr-containing samples analyzed here and the
monometallic Pd reference samples reported earlier [27].
Briefly, the CZ-supported component has a Ce:Zr atomic
ratio of ∼ 1, with an average crystallite size of 2 nm, and
is irregularly dispersed over the alumina support [7]. CZ
patches can be classified in two simple groups: bidimen-
sional patches with strong interaction with alumina and tridi-
mensional patches with properties approaching those of the
bulk material [7,10]. In our case, bidimensional CZ entities
are dominant [7].
Characterization of Cr and Pd in the oxidized state was
performed using Raman and EPR spectroscopies. Mono
and bimetallic systems containing 1 wt% of Cr have Ra-
man spectra shown in Fig. 2. Spectra are shown for hy-
drated (as stored) and de-hydrated (calcined) samples; the
strong changes observed between these states provide ev-
Fig. 1. CO (A) and NO (B) conversion profiles for the CO + NO + O reac-
2
tion over palladium–chromium (closed symbols) samples and monometal-
lic references (open symbols). Circles, Pd1CrCZA and PdCZA; squares,
Pd1CrA and PdA; triangles, CrCZA.
duced from Fig. 1 and Table 1, the temperature difference
between the two reactions ranges from about 25 K for Pd-
CZA to 75 K for Pd05CrCZA. The addition of Cr exhibits
catalytic effects with a marked dependence on the support;
in the case of alumina, it appears that Cr decreases the CO
light-off by 75 K and increases the NO value by ca. 45 K.
An additional effect is the existence of a plateau in NO con-
version (Fig. 1B) at temperatures from ca. 450 to 550 K.
A lesser effect, on the other hand, is observed for the CZA
Table 1
−1
50% isoconversion temperatures (T , K) observed for the bimetallic and monometallic reference catalysts in the CO + O + NO reaction at 30,000 h
50
2
Pd1CrCZA
Pd1CrA
Pd05CrCZA
Pd025CZA
PdCZA
PdA
T
T
(CO conv.)
(NO conv.)
391
458
403
458
384
458
403
458
432
455
479
412
5
0
0
5

224
M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
Table 2
Summary of results for the ceria–zirconia component from XRD and TEM/ED/XEDS analysis
b
(Ce + Zr/Al)at. ratiob,c
Sample
Cell parameter
Å)
Average particle
size (nm)
(Ce/Zr)
at. ratio
a
(
d
d
PdCZA
Pd1CrCZA
a
5.30
5.30
2
2
0.8–1.2
0.8–1.3
0.25–0.45
0.2–0.5
Obtained from dark-field TEM images using the (111) ring of Cex Zr
XEDS results.
Only zones presenting detectable amounts of Ce,Zr are included.
From electron diffraction results.
O .
2
1
− x
b
c
d
idence for the presence of highly dispersed species rather
than well-developed Cr-containing phases [32–34]. The Ra-
man profiles are dominated in all cases by bands at ca. 1005
or antiferromagnetic phase [38]. The presence of a weak,
narrow signal at g = 4.3 was also noted (see the inset in
Fig. 3), whose intensity decreased with increasing recording
temperature, and which most likely corresponds to residual
−1
and 875 cm , ascribed to terminal Cr=O and bridging
Cr–O–Cr vibrations, respectively, and attributed to Cr(VI)
species in the form of chromates with different degrees
of oligomerization [32–34]. These bands appear with di-
minished intensity in the presence of Pd. Additional mi-
3
+
Fe impurities present in the alumina support material.
3.3. In situ XANES analysis
−1
nor bands are detected at 1050 cm in the case of hy-
Factor analysis results at the Pd K-edge (Table 3) for
drated CrZCA, ascribed to the Cr=O stretch of isolated
Pd1CrCZA, Pd1CrA, PdCZA, and PdA obtained during a
−1
species [32–34], at 360 cm for hydrated CrA, which can
be attributed to a Cr–O deformation mode [32–34], and at
temperature-programmed run under CO + NO + O are
2
consistent with the existence of two different chemical
−1
6
25 cm for calcined Pd1CrCZA. The last is typical of
alumina-supported Cr materials [32,34] and may be asso-
ciated with the presence of amorphous or poorly crystalline
Cr2O3. In connection with this band, it is interesting to note
−1
the absence of a sharp band close to 550 cm due to crys-
talline Cr2O3 [32–34]. An alternative assignment of bands
−1
around 650 cm is the B2g mode of PdO, since this mate-
rial has a resonance-enhanced Raman signal peaking at this
−1
energy [35]. However, as the band is centered at 625 cm
and the excitation line used here is 415 nm it is unlikely that
a resonance enhanced signal should be detected.
EPR spectra of the samples (Fig. 3) show a common,
fairly symmetrical signal at g = 1.970 and with ꢀHpp ≈
4
5 G, whose integrated intensity represents a number of
spins equivalent to ca. 5.0, 1.1, 1.7, and 1.6 wt% of the
entire chromium content of the samples for CrA, Pd1CrA,
CrZCA, and Pd1CrZCA, respectively. Additionally, a broad
and poorly defined signal (inset of Fig. 3) appears at low
magnetic field. It shows the highest intensity for Pd1CrA,
decreasing significantly for CrZCA and Pd1CrZCA and
was almost absent for CrA. The main difference between
both signals concerns their behavior towards recording
temperature: while a roughly threefold decrease is produced
in the intensity of the signal at g = 1.970, indicating
that the corresponding paramagnetic species approximately
follows the Curie law, the intensity of the low-magnetic
field signal essentially does not change or undergoes a
slight increase when the spectrum is recorded at RT. The
behavior of the signal at g = 1.970 is typical of isolated
Cr(V) square pyramidal species [36,37]. The influence of
the recording temperature of the low-field signal, along with
its relatively large linewidth, suggests that it corresponds to
species coupled with Cr(III) or Cr(IV) immersed in a ferro-
Fig. 3. EPR spectra at 77 K of CrA (a), Pd1CrA (b), CrCZA (c), and
Pd1CrCZA (d). The inset shows a large magnetic field range spectrum of
PdCrA.

M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
225
Table 3
Principal component analysis of Pd K-edge XANES results
Factor
Eigenvalue
%SL
R(r)
Variance
A. PdA
0.00
0.00
1
2
3
4
5
6
7
8
491.351
0.64410
0.00217
0.00095
0.00066
0.00041
0.00032
0.00021
701.23
272.15
2.08
1.29
1.46
1.13
1.33
1.03
99.868
0.021
0.001
0.001
8.03
17.90
20.61
27.85
29.84
37.27
B. Pd1CrA
1
2
3
4
5
6
7
8
546.3
0.00
0.00
0.07
21.35
22.53
28.41
34.48
36.98
1104.32
53.83
13.04
1.17
1.36
1.29
99.914
0.014
0.002
0.45848
0.00779
0.00055
0.00042
0.00028
0.00019
0.00015
1.08
0.98
C. PdCZA
1
2
3
4
5
6
7
8
458.49
0.31594
0.00149
0.00092
0.00040
0.00029
0.00021
0.00018
0.00
0.00
1334.23
194.34
1.45
2.06
1.23
1.22
1.01
1.10
99.930
0.070
0.001
0.001
10.24
12.38
25.48
29.50
34.36
35.02
D. Pd1CrCZA
1
2
3
4
5
6
7
8
762.732
0.53430
0.00366
0.00261
0.00102
0.00048
0.00038
0.00036
0.00
0.00
7.24
1334.65
135.34
1.29
2.34
1.88
1.15
0.98
1.08
99.929
0.070
0.001
Fig. 4. Concentration profiles during the temperature-programmed reaction
run for Pd- (A) and Cr-containing species (B). Initial species (filled sym-
bols; species number 1 and 2); intermediate species (half-filled symbols;
species number 3), and final species (open symbols; species number 4).
Circles, Pd1CrCZA; rhombs, PdCZA; triangles, Pd1CrA; squares, PdA.
5.87
16.84
30.13
33.46
33.07
−
3
Variances lower than 10 are not reported.
samples in their oxidized state shows only slight differences
which may be a consequence of the limited particle size
of the oxidic Pd-containing entities. This suggests similar
local symmetry (D2h) and oxygen first-shell coordination
distances among the samples and also in the PdO reference.
Similarly, the XANES spectra of the final state detected in
the reaction (Fig. 5B) do not display significant differences
between samples and can be readily identified as zerovalent
Pd. Slight differences when compared with the reference
Pd foil (Fig. 5B,a) include the smaller intensity of the
continuum resonances (CRs), particularly for Pd1CrCZA,
and a shift to lower energy (ca. 1.4 eV; similar in all
catalysts) for the 4f CR (located around 24,390 eV).
As Pd–Cr alloys have reduced fcc lattice parameters with
respect to Pd [39] and the positions of CRs have an
inverse dependence with the square of the coordination
distance, the formation of such a binary phase would
be consistent with a shift to higher energies, as occurs
with the Pd LIII-edge White line [40]. It appears that
species except in the case of Pd1CrA, where three chemical
species are detected. The number of species is clearly
indicated by sharp variations in the % SL and R(r) values
displayed in Table 3. In Fig. 4A the fraction of these
chemical species as a function of sample temperature are
depicted. This plot shows that Pd is present initially as only
one chemical form when the reaction starts. This compound
is then directly converted to a final state in samples where
only two species in total were detected, while for Pd1CrA,
which showed evidence for three species, conversion from
initial to final sates involved an intermediate species. The
temperature at which the maximum rate of initial-state
(
chemical species) disappearance occurred was lower for
Pd1CrA (ca. 450 K) but very similar for the other samples
ca. 490 K for PdA and 510 K for Pd1CrCZA and PdCZA).
(
In all cases, only one Pd species was detected above 625 K.
When compared with the reference spectrum of PdO
(
Fig. 5A,a), the shape of the XANES spectrum of the

226
M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
Fig. 5. XANES spectra for PCA predicted chemical species. (A) Pd initial, oxidized species; (B) Pd final, reduced species; (C) Pd species detected for
the Pd1CrA sample; and (D) Cr species detected for the Pd1CrA sample. Lower case labels: (a) reference compounds, PdO or Pd; (b) PdA; (c) Pd1CrA;
(d) PdCZA, (e) Pd1CrCZA. Labels 1 to 4 as in Fig. 4.
the presence of Cr in the Pd network must be limited at
most to a small percentage and that differences with the
reference foil mostly reflects the small particle size and/or
an interaction with the support component (particularly with
the CZA support). The XANES spectrum of the intermediate
compound detected for Pd1CrA is presented in Fig. 5C
together with the spectra of the other two Pd-containing
species and the PdO reference. Although its shape is rather
complex to interpret in detail, the edge position (which is
located at lower energy than the shoulder d-like resonance
observed at the adsorption edge region) indicates a Pd(I)-
like partially reduced state. This interpretation is based upon
the universal linear relationship that has been shown to
exist between the formal oxidation state and the edge or
CR position (see, for example, Ref. [30]). The presence
of a Pd partially reduced state is further confirmed by the
decreasing intensity (with respect to the calcined material
species numbered as 1 in Fig. 5C) of the first (5sp) CR
located ca. 24,365 eV.
Further information concerning the chemical nature of
the intermediate was obtained by studying the Cr K-edge
of the Pd1CrA (and CrA) sample(s). In this case, Table 4
gives evidence for the existence of four species. Fig. 4B
shows the presence of two initial species in Pd1CrA, which
Table 4
Principal component analysis of Cr K-edge XANES results for Pd1CrA
Factor
Eigenvalue
%SL
R(r)
Variance
Pd1CrA
0.00
0.00
0.82
0.63
1
2
3
4
5
6
7
8
851.332
2.64623
0.01261
0.00521
0.00117
0.00035
0.00021
0.00020
285.33
193.15
2.22
4.06
3.00
1.46
0.93
1.08
99.674
0.323
0.002
0.001
7.07
25.88
34.90
33.33
−
3
Variances lower than 10 are not reported.

M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
227
are converted to an intermediate state above ca. 350 K, giv-
ing another, final state above ca. 520 K. Consistent with re-
sults at the Pd K-edge, transformation of the Cr species was
complete by 625 K. In Fig. 5D, the corresponding XANES
spectra of the Cr species are presented along with some ref-
erence compounds. Comparison of the 1s → 3d pre-edge
feature (located around 5995 eV) and edge shape indicates
the chemical nature of the two species present when the
Pd1CrA was initially exposed to the reaction mixture; a non-
centrosymmetric tetrahedral or subgroup Cr(VI) compound,
with some similarities to the distorted-tetrahedral K2Cr2O7
reference, and another one similar to Cr(IV) in an oxidic
environment. The presence of both Cr chemical states for
alumina-supported materials calcined above 573 K has been
shown on the basis of spectroscopic characterization stud-
ies [32–34,41,42]. It can be noted, however, that, according
to the XANES data, the sample here seems to contain pre-
dominantly Cr(VI) after calcination (result not shown). The
main point to stress here is that an intermediate compound
with maximum concentration about ca. 520 K is also ob-
served at the Cr K-edge. Comparison with reference mate-
rials suggests that the compound contains Cr(III). Note that
the huge edge shift (ca. 2–3 eV) displayed between Cr for-
mal oxidation states differing in one unit is extremely use-
ful in providing an accurate description of the oxidation
state [30,41]. As this intermediate evolves with temperature,
a Cr(VI) state in an oxidic environment (CrO3-like) is re-
covered. The CrA reference, on the other hand, also displays
the presence of a CrO3-like species in the calcined state, in
this case showing stability throughout the temperature ramp
under the reactive atmosphere.
lower-frequency band, which also shows a red shift with
increasing temperature, is assigned to bridging carbonyl
species adsorbed onto Pd(0) [27,44–46]. For both carbonyl
species, the red shifts observed result from a continuous
decrease in CO coverage of Pd particles resulting either
from desorption or from reaction with oxygen at these
centers. The magnitude of the shift may reflect differences
between the relative adsorption enthalpies of CO on these
sites [46], between adsorbed molecule distances and/or
between different catalytic activities for such centers. The
observed shift was greater for the linearly adsorbed species.
As a general rule, the intensity of carbonyls adsorbed
onto Pd(0) sites was decreased in the presence of Cr.
−1
An additional band at 2155–2158 cm
assigned to CO
adsorbed onto Pd(II) sites [27,44–47] was detected in all
cases except PdA. In the case of PdA and less so for
−1
the other samples, a band at ca. 2120–2130 cm , also
related to the presence of surface Pd(II) [25,44–47], was
observed. The temperature dependence of the adsorbed
Pd(II) carbonyl species was very similar among the samples
studied. DRIFTS spectra also contained bands at 2250 and
−1
2232 cm , which were not detected in the absence of NO
in the reactive mixture [7,21,27], and can be ascribed to
isocyanate species adsorbed at octahedral and tetrahedral
3
+
Al cations, respectively [48]. This species results from
NO decomposition followed by N–CO combination on
metal particles and spillover onto support sites, and the large
intensity of the νasNCO band and thus its ease of detection
may be used to roughly determine the onset temperature for
NO dissociation [42–48].
To examine the state of the metal surfaces on completion
of the reaction runs and in the absence of thermally driven
adsorption effects, samples were flushed at 623 K, cooled
in N2 to room temperature, and subsequently exposed to
1% CO/N2 before the cell was again flushed with N2.
Spectra recorded during this treatment are displayed in
3
.4. In situ DRIFTS
Spectra recorded under the CO + NO + O2 mixture for
PdA, Pd1CrA, PdCZA, and Pd1CrCZA samples are shown
in Figs. 6A–6D, respectively. A rough correlation can be
observed for all samples between evolution of bands due
to gaseous CO (barely discernible doublet in the 2200–
−1
Fig. 7. A band centred at 1990–1980 cm , present in
all samples and attributed to bridging carbonyls adsorbed
onto metallic Pd particles, was observed with attenuated
intensity for the CZA-supported samples. The presence of
Cr also affected the intensity of the latter, but this was
more marked for the alumina-supported samples. As a
shoulder on the band due to bridging carbonyls, a broad
− −1
1
2
050 cm region) and CO2 (2400–2300 cm ) and the
CO conversion profiles (Fig. 1A). In general, the spectra
give evidence for the presence of different carbonyl species
mainly adsorbed onto Pd. Cr carbonyl contributions, which
−1
−1
may be detected around 2160–2170 cm
[43], are not
band centered ca. 1925 cm was observed for PdA and
observed, even for the CrCZA reference catalyst (result
not shown). This may be attributed to the low adsorption
enthalpy of CO on Cr [43]. Analysis of Pd carbonyls is
made on the basis of previous results for the monometallic
references [27,44,45].
with very small intensity for the remaining samples and
attributed to CO adsorbed in three-fold hollow sites on
Pd(0) [27,44–47]. Linear Pd(0) carbonyls (band ca. 2095–
−1
2100 cm ) were detected for the monometallic references,
with greater intensity for PdCZA, but were almost absent
in the presence of Cr. This linear carbonyl was the most
affected by the purging step in accordance with its low
adsorption enthalpy. Additional bands at 2118–2124 and
At low temperatures, a shoulder (above the CO gas
−
1
contribution) located at or below 2100 cm , and a broad
band centered around 1980–1970 cm
−1
were observed
−1
for all Pd-containing samples (Figs. 6A–6D). The former,
which displays a significant red shift with temperature,
2152 cm
can be attributed to adsorption onto Pd(I)
and Pd(II), respectively [21,22,27,44–47]. It should be
mentioned, however, that the presence of Cr carbonyls
cannot be excluded in the presence of CO gas, although
−1
was observed down to ca. 2050 cm and is attributed to
CO adsorbed onto Pd(0) on-top sites [27,43,44,46]. The

228
M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
Fig. 6. DRIFTS spectra of (A) PdA, (B) Pd1CrA, (C) PdCZA, and (D) Pd1CrCZA samples in a flow of 1% CO, 0.45% O , 0.1% NO, N balance at (a) 303,
2
2
(b) 333, (c) 363, (d) 393, (e) 423, (f) 453, (g) 483, (h) 513, (i) 543, and (j) 573 K.

M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
229
(
CrO3-like) oxidation states. The EPR spectra show that
the Cr EPR-visible species are influenced by the presence
of the Pd, decreasing the amount of Cr(V) and indicating
the presence of a phase containing magnetically coupled
cations in the Pd1CrA catalyst but not present for CrA. This
influence is also evident in the evolution of the XANES Cr
K-edge spectra immediately upon contact with the reactive
mixture at RT; as depicted in Fig. 8I B and justified by
the XANES study, the reaction mixture changes the Cr
phase distribution and chemical state, giving a tetrahedral
Cr(VI) plus a Cr(IV), CrO2-like chemical phase. For the
sake for simplicity, Fig. 8I does not differentiate between
the two Cr(VI) phases already mentioned. Although the
extent and nature of the Pd–Cr interaction is unclear, both
techniques do show that Pd and Cr are in sufficiently
close proximity to alter the behavior of the Pd1CrA system
significantly with respect to the monometallic reference. The
close proximity of both active elements in the calcined state
induces the formation of an intermediate in the ca. 350–
625 K temperature range (Fig. 8I C), this phase containing
Pd(I) and Cr(III) as the principal oxidation states and,
in terms of its local order, showing similarity to PdCrO2
oxide [25]. This chemical compound has Pd–Pd distances of
Fig. 7. DRIFTS spectra following experiments shown in Fig. 6 after
exposure to CO (full lines) and subsequent flushing with N (dashed lines).
2
(a) Pd1CrA, (b) PdA, (c) Pd1CrCZA, and (d) PdCZA.
their absence was noted in the CrCZA case. Results in Fig. 7
give firsthand evidence that the Pd morphology/particle size
is very different for A and CZA monometallic samples.
This is supported by the intensity ratio of the on-top to
bridge Pd(0) carbonyl bands. On the other hand, Cr has
two clear effects. First, in A-supported specimens a drastic
influence on morphology is shown by the strong decrease
in Pd(0) bridging carbonyls. This can be mainly related (as
described below) to the genesis step of the Pd zerovalent
phase. Second, the intensity of bands related to oxidized
Pd centers clearly indicates that the presence of Cr favors
the formation of such centers, at the surface, independent
of the nature of the support. This appears to selectively
influence the Pd, centers where CO is adsorbed in the ontop
mode and located in closed packed (111)-like planes, having
much lesser impact in those corresponding to open (100)-
like planes which are mainly (although not exclusively)
associated with the CO bridging adsorption mode [21,22,
2
1
.83 Å, very close to metallic dimensions (2.75 Å), giving a
-D (dimension) electronic conductor [25], with some Pd5s
electron delocalization, resembling the metal.
Around 520 K this binary oxide is transformed, yielding
two monometallic single phases containing Pd(0) and Cr(VI)
(
Fig. 8I.D). It should be stressed that the evolution of this
chemical compound gives the same Pd and Cr phases as the
monometallic references at temperatures above 520 K. No
evidence is detected that would indicate the formation of a
binary Pd–Cr alloy, in spite of the existence of the binary
oxidic precursor. The presence of a miscibility gap for bulk
Pd–Cr alloys has been reported, and typically a maximum
of ca. 15 at% Cr is known to be substitutionally present
in the fcc lattice structure of Pd [26,39]. The absence of
a significant alloying effect for samples here indicates that
the Pd–Cr bulk alloy binary phase diagram obtained for bulk
phases [26] is also appropriate for nanosized materials and
that the surface contribution to the total energy may further
lower the limit of Cr solubility in the Pd. An additional
influence of the gas phase on the resulting zerovalent phase
composition may also need to be considered. In any case, it
is clear that Pd–Cr alloy phases have been described under
2
7,44–47,49]. Of course, the analysis of the second effect
is somewhat masked by the first in the case of the alumina-
supported samples.
4
. Discussion
(
CO + NO) stoichiometric mixtures only in prereduced (H2
4
.1. Structural and morphological nature of samples
at 870 K) materials [19], indicating that the presence of Cr(0)
before and/or at the beginning of the reaction may be of
importance in obtaining/maintaining the binary zerovalent
phase. The formation of metallic Pd particles from the
Pd–Cr oxidic intermediate has, however, a clear impact
on the final zerovalent Pd phases formed. Comparison
of IR results for CO-adsorption on postreaction samples
mainly shows a lowering of the bridge and threefold Pd(0)
carbonyls as well as a larger intensity of Pd(II) in the
presence of Cr. The first fact clearly indicates a smaller
As is evident from Section 3, a clear influence of the
support on the evolution of samples with temperature is
observed under reaction conditions. In the case of alumina-
supported samples, a degree of interaction between the
noble metal and the Cr is apparent, even in the calcined
material. This is schematically represented in Fig. 8I A,
where according to XANES, both metals are mainly inserted
into oxidic matrixes with Pd(II) (PdO-like) and Cr(VI)

230
M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
Fig. 8. Schematic view of the main Pd- and Cr-containing chemical phases detected in the Pd1CrA (I) and Pd1CrCZA (II) systems. (A) sketch visualizes the
calcined materials; (B), (C), and (D) sketches of systems under the reaction mixture at temperatures below 400 K, at 500 K, and above the latter.
particle size of the monometallic Pd particles, while the
second gives evidence for the coexistence of metallic and
oxidized Pd sites at the surface, which may be driven by the
presence at the high (reaction) temperatures of small oxidic
Cr entities (ions or small aggregates) at the surface of the
noble metal particles. The presence of such oxidized entities
has been also mentioned for Mn or Mo (Pd-) bimetallic
catalysts [16–18].
In the case of CZA-supported samples, there was no
evidence from EPR, Raman, or XANES for any significant
interaction between Pd, either in the oxidized, Pd(II), or
partially reduced, Pd(I), state, and Cr. This absence of
interaction between the components was found for Pd–Cr
catalysts with Pd:Cr atomic ratios from 0.5 to 2. It seems that
the presence of the CZ component prevents, to a large extent,
the close proximity of these components in these novel
ceria–zirconia TWC catalysts. As Raman spectra (Fig. 2)
suggest broadly similar Cr distributions on A and CZA
after calcination, the absence of Cr–Pd interactions could be
attributed to the Pd distribution, which tends to locate Pd
at the interface between the CZ patches and the A carrier
in the CZA-supported sample [27]. This is schematically
shown in Fig. 8II A, where Pd particles are located at
the mentioned interface between CZ patches and alumina,
and Cr entities can be present in both the alumina carrier,
as Cr(VI) oligomers, and CZ patches, as amorphous or
poorly crystalline Cr2O3-like entities (which, according to
Raman results, is the Cr species exclusive of CZA materials).
This Pd1CrCZA system does not evolve until the reaction
temperature exceeds 400 K (Figs. 8II B and C). The Pd
zerovalent particles formed are not significantly affected in
terms of either their size or morphological characteristics by
the presence of Cr, although the surfaces of these particles
display a significant quantity of Pd(II) sites after reaction.
Again, this suggests the presence of small Cr ions or
aggregates located at the surface of the noble metal particles
(Fig. 8II D). The Pd(II)/Pd(0) surface equilibrium is shifted
to the oxidized state to a greater extent for CZA- than for
A-supported Cr-containing materials, but this effect may
also be observed for the monometallic Pd references (Fig. 7)
and has been previously reported for such reference Pd
samples [44].
4.2. CO oxidation and NO reduction
In Table 5 IR and XANES evidence for the onset
temperatures of reduced Pd (Pd(0) and Pd(I)) state formation
has been summarized and the results compared with the
initial (T ) and T temperatures corresponding to 50%
0
50
conversion of CO or NO. Note, of course, that TIR values in
Table 5, which point to the presence of surface Pd(0) entities,
are subject to COg availability and, thus, depend on CO
conversion level. For NO reduction, catalytic activity data
are shown only for the 50% conversion level, as the onset

M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
231
Table 5
existence of a Pd–CZ interaction where additional sites for
Onset temperatures for IR detection of NCO (T
, K) and Pd(0) carbonyls
NCO
oxidant (NO/O ) activation are readily generated [27,44,51]
2
(
TIR, K) and for detection of Pd(I) and Pd(0) by XANES (T
, K) and
XANES
and, in some cases, by the ease with which metallic Pd is
initial and 50% conversion temperatures (T or T , K) for PdA, Pd1CrA,
0
50
formed [44]. Here, the presence of Cr plays only a secondary
role, decreasing the light-off temperature by less than half
of what is detected for the alumina system. As only small
changes in particle size are expected on going from PdCZA
to Pd1CrCZA (DRIFTS results in Fig. 7), the effect of
this parameter on light-off performance can be eliminated.
A more likely explanation can be proposed by considering
that the Pd(II)/Pd(0) surface distribution shows more bias
toward oxidized states in the presence of Cr. This would
PdCZA, and Pd1CrCZA samples
Samples
Pd(0) (Pd(I)) onset
T , T CO
T
T , NO
50
0
50
NCO
TIR
T
XANES
420
PdA
363
423
363
423
350, 479
321, 403
352, 432
321, 391
423
453
453
483
412
Pd1CrA
PdCZA
Pd1CrCZA
470 (350)
410
458
458
458
410
is not readily discernible due to the adsorption–desorption
phenomena at low temperatures and, in any case, differences
at the onset of the light-off are roughly equivalent at higher
levels of conversion.
suggest enhanced NO and/or O dissociation assisted by the
2
Pd–Cr interface, which would lead to a reduced CO coverage
of the metallic particles and, consequently, to a reduction
of the CO inhibition effect on light-off with respect to the
alumina-supported samples.
The behavior of PdA for CO oxidation is consistent with a
Langmuir–Hinshelwood mechanism, as observed for model
Pd/SiO2 systems [46]. In the model case mentioned, at tem-
peratures below initial light-off CO coverage of the metal-
lic Pd particles is close to unity and the rate-determining
step is the rate of creation of vacant sites by desorption
of CO, where oxidants (NO/O2) can be activated [46]. Al-
though this model has widespread applicability, Hoffman et
al. [49] have recently shown that the CO-poisoning effect of
the metal surface is strongly diminished for small particles.
For preoxidized materials (such as those used here), the cat-
alytic activity is very low until extensive reduction of Pd by
CO can take place [27,44,45] and a rough correlation can be
established for PdA between the temperature of appearance
of Pd(0) sites at the surface by DRIFTS, when NO is used
as oxidant and the corresponding temperature of appearance
of bulk Pd(0) by XANES when oxygen is used (Table 5). In
the case of the alumina-supported system, the preference of
CO for NO or O2 as oxidant depends to a large extent on
the Pd particle size which, in turn, drives the type (open vs
closed and number of edge and steps) of surface exposed to
the reactants, and is mainly based on the fact that NO disso-
ciation and reduction are structure-sensitive processes which
are favored in systems with large particle size [44–47].
The presence of Cr in the A-supported catalysts dramat-
ically decreases the temperature for CO oxidation. In this
case, the correlation is with the presence of the Pd(I)–Cr(III)
phase (Table 5). This is likely due to the metallic-like charac-
ter of the Pd 5sp band, which is still able to activate the CO
molecule and to initiate the reaction. In this case, the absence
With regards to NO reduction, Table 5 displays a rough
correlation between the appearance of NCO and the T ,
5
0
NO temperature, indicating that this molecule can be used
as a simple probe for the initiation of the NO reduction
process. For this reaction, the introduction of Cr into the
catalyst has detrimental effects on A-supported systems.
A correlation can be established between the evolution of
the Pd(I)–Cr(III) intermediate concentration (Fig. 4) and NO
conversion (Fig. 1A), particularly for the plateau detected in
the conversion plot (450–550 K) and the region of maximum
concentration of the intermediate. It therefore appears that
the lower performance of the palladium alumina-supported
in the presence of Cr can be ascribed to a lower activity
of the Pd(I)–Cr(III) intermediate for NO activation (always
with respect to the metallic, Pd reference). In fact, small
variations in the Pd valence density of metallic-like systems
have either no effect or a moderate influence on CO
activation, while the corresponding NO process can be
strongly favored or hindered [52]. Moreover, the interaction
between Pd and Cr O appears to have little impact on the
2
3
CO desorption temperature and C–O stretching frequency of
such molecules when adsorbed onto the noble metal [53].
After the disappearance of the Pd(I)–Cr(III) and in the
presence of Pd(0) the NO conversion profile reaches 100%.
For CZA-supported materials, no significant influence of
Cr is detected. This despite the clear enhancement in NO
(and/or O ) dissociation already mentioned on the basis of
2
the IR results and attributed to the presence of Cr entities
at the Pd surface. Although these observations are not eas-
ily reconciled, they might be rationalized by suggesting an
influence of Cr species on the N-coupling step(s) of the re-
duction process. Whether such step occurs near or at the Pd–
CZ interface and/or is modified directly by Cr presence at
such sites or indirectly through modification of interfacial Ce
(or Zr) centers remains unknown. Additionally, comparison
with PdA indicates that the Pd–CZ interface does not appear
to promote this reaction in our system, contrarily to previ-
ous literature reports for similar reactions [43,54]. However,
a comparison between PdA and PdCZA must be affected by
(
or further decrease) of the CO-poisoning effect (character-
istic of the noble, zerovalent phase) can be also claimed to
contribute to the reduction in light-off temperature. Note, on
the other hand, that classical Pd(I) and Pd(II) states (in their
monometallic oxide form) are usually believed to be less ef-
fective for CO oxidation that metallic-like states [26,27,44,
4
5,50].
It is well known that the presence of a reducible oxide
such as CZ decreases the CO oxidation light-off temperature
with respect to alumina-supported catalysts due to the

232
M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
particle size effects (Fig. 7), which exert a very important in-
fluence on the NO dissociation process [44–47]. The Pd size
distribution and interaction with the support is known to fa-
vor NO reduction in suitably prepared alumina systems with
respect to ceria or ceria–zirconia promoted catalysts [55].
References
[
[
[
1] E.S.J. Lox, B.H. Engler, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.),
Environmental Catalysis, Wiley–VCH, 1999, p. 1.
2] A. Trovarelli, Catal. Rev. Sci. Eng. 38 (1996) 97, and references
therein.
3] D.S. Lafyatis, G.P. Ansell, S.C. Bennett, J.C. Frost, P.J. Millington,
R.R. Rajaram, A.P. Walker, T.H. Ballinger, Appl. Catal. B 18 (1998)
123.
[
[
4] J. Kašpar, P. Fornasiero, M. Graziani, Catal. Today 50 (1999) 351.
5] Y. Nagai, T. Yamamoto, T. Tanaka, S. Yoshida, T. Nonaka, T. Oka-
moto, A. Suda, M. Sugira, Catal. Today 74 (2002) 225.
5
. Conclusions
[
6] A. Martínez-Arias, M. Fernández-García, V. Ballesteros, L.N. Sala-
manca, C. Otero, J.C. Conesa, J. Soria, Langmuir 15 (1999) 4796.
7] M. Fernández-García, A. Martínez-Arias, A. Iglesias-Juez, C. Belver,
A.B. Hungría, J.C. Conesa, J. Soria, J. Catal. 194 (2000) 385.
The behavior of a series of Pd–Cr bimetallic catalysts in
CO oxidation and NO reduction processes has been tested
and compared with that of monometallic Pd references.
The catalytic properties display strongly dependence on the
degree of interaction existing between both metals in the
calcined state and this is in turn dependent on the nature of
the support.
[
[8] M.H. Yao, N.J. Blaird, F.W. Kunz, J. Catal. 166 (1997) 67.
9] M. Ozawa, K. Matuda, S. Suzuki, J. Alloys Comp. 303–304 (2000) 56.
[
[
10] M. Fernández-García, A. Martínez-Arias, A. Iglesias-Juez, C. Belver,
A.B. Hungría, J.C. Conesa, J. Soria, Phys. Chem. Chem. Phys. 4
(2002) 2473–2481.
In the case of alumina, both metals are closely located
following calcination. Under reaction conditions, this prox-
imity induces the formation at low or intermediate temper-
atures of a Pd(I)–Cr(III) mixed oxide phase which alters
the catalytic response with respect to the monometallic Pd
reference. This phase is apparently capable of activating
CO and its presence at lower temperatures than zerovalent
Pd in the reference system significantly reduces CO light-
off temperature. In contrast, this phase is detrimental for
NO reduction. It appears that such a Pd–Cr mixed oxide
could be used to enhance Pd activity in oxidation reactions.
For (Ce,Zr)Ox/Al2O3 supported systems, no evidence was
found of significant interactions between the metals in the
calcined state in a series of catalysts with Pd:Cr atomic ratios
ranging from 0.5 to 2. However, under the course of the reac-
tion, formation of Pd(0) particles with Cr entities at the sur-
face was detected at high temperature. This may enhance NO
activation/dissociation and moderately favor CO oxidation
but does not have any effect on NO reduction. This could be
rationalized by considering that Cr may affect N-coupling
steps.
[11] R. van Yperen, D. Lindner, L. Mubmann, E.S. Lox, T. Kreuzer, Stud.
Surf. Sci. Catal. 116 (1998) 51.
[
12] Z. Hu, C.Z. Wan, Y.K. Lui, J. Dettling, J.J. Steger, Catal. Today 30
1996) 83.
(
[13] M. Skoglundh, H. Johansson, L. Löwendahl, K. Jansson, L. Dahl, B.
Hirschauer, Appl. Catal. B 7 (1996) 299.
[14] K.C. Taylor, Catal. Rev. Sci. Eng. 35 (1993) 457.
[15] B. Coq, F. Figueras, J. Mol. Catal. A 173 (2001) 117.
[16] T.E. Hoost, G.W. Graham, M. Shelef, O. Alexeev, B.C. Gates, Catal.
Lett. 38 (1996) 57.
[17] M. Schmal, M.A.S. Baldanza, M.A. Vannice, J. Catal. 185 (1999) 138.
[18] J.F. Trillat, J. Massadier, B. Morawek, H. Praliaud, A.J. Renouprez,
Stud. Surf. Sci. Catal. 116 (1998) 103.
[19] A. El Hamdaoui, G. Bergeret, J. Massadier, M. Primet, A.J. Re-
nouprez, J. Catal. 148 (1994) 47.
[20] M. Fernández-García, A. Martínez-Arias, C. Belver, J.A. Anderson,
J.C. Conesa, J. Soria, J. Catal. 190 (2000) 387.
[
[
[
[
21] M. Fernández-García, A. Martínez-Arias, J.A. Anderson, J.C. Conesa,
J. Soria, Stud. Surf. Sci. Catal. 130 (2000) 1325.
22] 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.
23] M. Fernández-García, J.C. Conesa, A. Clotet, J.M. Ricart, N. López,
F. Illas, J. Phys. Chem. B 102 (1998) 141.
24] M. Fernández-García, J.A. Anderson, G.L. Haller, J. Phys. Chem. 100
(1996) 16247.
[25] R.D. Shannon, D.B. Rogers, C.T. Prewitt, Inorg. Chem. 10 (1971) 713.
[26] T.B. Massalki, Binary Alloy Phase Diagram, Vol. 1, 2nd ed., ASM
International, 1993.
[27] M. Fernández-García, A. Martínez-Arias, A. Iglesias-Juez, A.B. Hun-
gría, J.A. Anderson, J.C. Conesa, J. Soria, Appl. Catal. B 31 (2001)
Acknowledgments
39.
[28] M. Fernández-García, C. Márquez-Alvarez, G.L. Haller, J. Phys.
Chem. 99 (1995) 12565.
We thank the Comunidad de Madrid for grants received
[
29] C. Márquez-Alvarez, I. Rodríguez-Ramos, A. Guerrero-Ruiz, G.L.
Haller, M. Fernández-García, J. Am. Chem. Soc. 119 (1997) 2905.
(
to A.B.H. and A.I.-J.), with which this work has been
carried out, and for financial assistance (to A.B.H.) under the
Ayudas para estancias breves en centros de investigación
extranjeros” program. We are also grateful to the scientific
and technical staff at Daresbury SRS (beam time Ref.
[30] M. Fernández-García, Catal. Rev. Sci. Eng. 44 (2002) 59.
[
31] E.R. Malinowsky, Factor Analysis in Chemistry, Wiley, New York,
991.
32] M.I. Zaki, N.E. Foudad, J. Leyrer, H. Knozinger, Appl. Catal. 21
1986) 359.
[33] B.M. Weckhuysen, I.E. Wachs, J. Phys. Chem. B 101 (1997) 2793.
“
1
[
(
3
8285) and the ESRF Synchrotron (BM29) for their help
[34] B. Grzybowska, J. Sloczynski, R. Grabowski, K. Wcislo, A. Koz-
during the XANES experiments, to Mr. F. Sánchez for
EPR measurements, and to Dr. M.A. Bañares for recording
Raman data. Financial help by CICYT (Project MAT2000-
lowska, J. Stoch, J. Zielinski, J. Catal. 178 (1998) 687.
[
[
35] J.R. McBridge, K.C. Kass, W.H. Weber, Phys. Rev. B 44 (1991) 5016.
36] D. Cordishi, M.C. Campa, V. Indovina, M. Occhiozzi, J. Chem. Soc.
Faraday Trans. 90 (1994) 207.
1
467) is also acknowledged.

M. Fernández-García et al. / Journal of Catalysis 214 (2003) 220–233
233
[
37] Z. Zhu, M. Hartmann, E.M. Maes, R.S. Czernuszevich, L. Kevan,
[47] A. Bensalem, J.C. Muller, D. Tessier, F. Bozon-Verduraz, J. Chem.
Soc. Faraday Trans. 92 (1996) 3233.
J. Phys. Chem. B 104 (2000) 4690.
[
[
38] O.F. Gorriz, L.E. Cadús, Appl. Catal. A 180 (1999) 247.
39] D. Wang, J.D. Cleway, T.B. Flanegan, J. Alloys. Comp. 325 (2000)
[48] J.A. Anderson, C. Márquez-Alvarez, M.J. López-Muñoz, I. Rodrí-
guez-Ramos, A. Guerrero-Ruiz, Appl. Catal. B 14 (1997) 189.
[49] J. Hoffman, I. Meusel, J. Hartmann, J. Libuda, H.-J. Freund, J. Ca-
tal. 204 (2001) 378.
[50] A.L. Tarasov, L.K. Przheval’skaya, V.A. Shevts, V.B. Kazanskii, Ki-
net. Katal. 29 (1989) 1020.
151.
[40] Y.S. Lee, K.Y. Lim, C.N. Whang, Y. Jeon, Appl. Phys. A 70 (2000) 59.
[41] A. Ellison, G. Diakun, P. Worthington, J. Mol. Catal. 46 (1988) 131.
[42] P.G. Harrison, W. Daniell, Chem. Mater. 13 (2001) 1708, and refer-
ences therein.
[51] P. Bera, V. Jayaram, G.N. Subbana, M.S. Hegde, J. Catal. 196 (2000)
293.
[43] M. Pykavy, V. Staemmler, O. Seiferth, H.J. Freund, Surf. Sci. 479
(
2001) 11, and references therein.
[52] F. Illas, N. López, A. Clotet, J.M. Ricart, J.C. Conesa, M. Fernández-
García, J. Phys. Chem. B 102 (1998) 8017.
[53] K. Wolter, H. Kuhlenbeck, H.-J. Freund, J. Phys. Chem. B 106 (2002)
6723.
[
44] A. Martínez-Arias, M. Fernández-García, A. Iglesias-Juez, A.B. Hun-
gría, J.A. Anderson, J.C. Conesa, J. Soria, Appl. Catal. B 31 (2001)
5
1.
[
45] M. Fernández-García, A. Martínez-Arias, L.N. Salamanca, J.M. Coro-
nado, J.A. Anderson, J.C. Conesa, J. Soria, J. Catal. 187 (1999) 474.
46] X. Xu, D.W. Goodman, J. Phys. Chem. 97 (1993) 7711.
[54] R. Di Monte, P. Fornasiero, J. Kašpar, P. Rumori, G. Gubitosa, M.
Graziani, Appl. Catal. B 24 (2000) 157.
[55] R. Farrauto, R.M. Heck, Catal. Today 51 (1999) 351.
[