Surface reconstruction of supported Pd on LaCoO3: Consequences on the catalytic properties in the decomposition of N2O

Article abstract of DOI:10.1016/j.jcat.2007.10.023

This study deals with the catalytic decomposition of N₂O over noble metal based catalysts. The deposition of palladium on reducible supports such as LaCoO₃ leads to higher activity from temperature-programmed experiments in comparison with conventional supports such as alumina. Such a different catalytic behaviour cannot be completely explained by changes in the metal dispersion but also by the extent of the metal/support interaction. Interestingly, successive reductive (250 °C in H₂) and oxidative thermal treatments in the reactant mixture at high temperature enhance the conversion of N₂O particularly on perovskite support. Additional surface characterisations show a re-dispersion and the stabilisation of palladium species in unusual oxidation states which would originate a rate enhancement in the decomposition of N₂O.

Full text of DOI:10.1016/j.jcat.2007.10.023

Journal of Catalysis 253 (2008) 37–49
www.elsevier.com/locate/jcat
Surface reconstruction of supported Pd on LaCoO3:
Consequences on the catalytic properties in the decomposition of N2O
J.P. Dacquin, C. Dujardin, P. Granger ∗
Université des Sciences et Technologies de Lille, Unité de Catalyse et de Chimie du Solide, Equipe Catalyse Hétérogène, UMR CNRS 8181,
Bâtiment C3, 59655 – Villeneuve d’Ascq Cedex, France
Received 20 July 2007; revised 11 October 2007; accepted 29 October 2007
Abstract
This study deals with the catalytic decomposition of N O over noble metal based catalysts. The deposition of palladium on reducible supports
2
such as LaCoO leads to higher activity from temperature-programmed experiments in comparison with conventional supports such as alumina.
3
Such a different catalytic behaviour cannot be completely explained by changes in the metal dispersion but also by the extent of the metal/support
◦
interaction. Interestingly, successive reductive (250 C in H ) and oxidative thermal treatments in the reactant mixture at high temperature enhance
2
the conversion of N O particularly on perovskite support. Additional surface characterisations show a re-dispersion and the stabilisation of
2
palladium species in unusual oxidation states which would originate a rate enhancement in the decomposition of N O.
2
© 2007 Elsevier Inc. All rights reserved.
Keywords: Greenhouse gas; Nitrous oxide; N O catalytic decomposition; Perovskite; Noble metals; X-ray photoelectron spectroscopy; Surface reconstruction
2
1. Introduction
may depend on the extent of metal/support interaction, i.e. on
the metal dispersion, with the creation of highly active sites
at the metal/support interface. Recently, Uenishi et al. [6] in-
vestigated the redox behaviour of palladium at start up in the
perovskite-type structure LaFe0.95Pd0.05O3 with Pd homoge-
neously distributed in the solid according to a sol–gel route for
the catalyst synthesis. They also reported an interesting cat-
alytic behaviour due to their self-regenerative function under
cycling conditions with typical successive reductive and ox-
idative atmospheres. Under reductive conditions, isolated Pd
species stabilised in the perovskite structure with unusual ox-
idation state +III and/or +IV segregated out onto the surface
from the B-site at relatively low temperature then the nano-
sized Pd0 particles re-oxidise under net oxidising conditions
into Pd2+ at 200–300 ◦C. Successive formation of a solid so-
lution occurs with an increase in temperature. Unfortunately,
such reversible properties have been characterised essentially
on highly-loaded Pd catalysts. Actually, a particular attention
is paid towards the reduction of noble metal contents in the
catalyst formulation or the development of alternative non-
noble metal catalysts. According to those objectives, the use
of sol–gel methods could not be a powerful procedure for op-
timising the surface concentration of noble metals particularly
Potential catalytic applications for perovskite materials
ABO3 have been previously reviewed [1]. Those materials can
be attractive due to lower costs and flexibility of their compo-
sition since they can tolerate significant substitution and non-
stoichiometry. The main drawback is usually related to their
low specific surface area despite significant developments in
the preparation procedures for obtaining solids exhibiting ho-
mogeneous compositions and high specific surface areas [2,3].
Also, those materials can be profitably used as support. The
deposition of active phases, such as noble metals, may consid-
erably enhance the overall activity due to cooperative effects
between noble metals and the support. In this latter case, re-
markable catalytic performances were obtained in the course of
the reduction of NO by hydrogen in O2 excess on 0.1 wt% Pt/
La0.7Sr0.2Ce0.1FeO3 with significant rate and selectivity en-
hancement towards the transformation of NO into nitrogen
below 200 ◦C [4,5]. It was found that this rate enhancement
*
Corresponding author. Fax: +33 3 20 43 65 61.
E-mail address: pascal.granger@univ-lille1.fr (P. Granger).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.10.023

38
J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
under net oxidising conditions with the concomitant formation
of bulk solid solutions which hinders a significant fraction of
noble metals. Nevertheless, the stabilisation of highly dispersed
oxidic palladium species at high temperature with a partial sup-
pression of thermal sintering process offers new perspectives to
those materials particularly for high temperature catalytic ap-
plications.
involving a citrate route. Supported palladium catalysts on
LaCoO3 were prepared according to a classical wet impreg-
nation route using palladium nitrate solutions with adjusted
concentrations in order to obtain 1 wt% Pd. The impregnated
samples were calcined in air at 400 ◦C and then reduced at
250 or 500 ◦C in pure H2 overnight. H2 titration measurements
were performed at 100 ◦C in order to minimise the forma-
tion of palladium hydrides in accordance with previous inves-
tigations [15]. The hydrogen uptake on pre-reduced samples
in H2 at 500 ◦C, was 40.4 and 7.5 µmol per gram of cata-
lyst respectively on Pd/Al2O3 and Pd/LaCoO3 which corre-
sponded to Pd dispersion of approximately 0.86 and 0.16. Pd
dispersion on LaCoO3 pre-reduced at 250 ◦C had not been
considered due to overestimation since hydrogen spill-over ef-
fects during H2 chemisorption measurements took place on
reducible supports [16,17] depending on the reduction tem-
perature particularly under mild conditions. Such an effect
had to be taken into account at moderate reduction tempera-
ture.
Temperature-programmed reduction (TPR) was carried out
in a Micromeritics Autochem II 2920 (5 vol% H2/Ar). In situ
X-ray diffraction (XRD) were performed using a Bruker D8
diffractometer equipped with a CuKα (λ = 0.154 nm) radia-
tion. The sample was in situ reduced under a flow of 3 vol%
H2 in He. The temperature was raised until the desired tem-
perature (3 ◦C min−1) with a plateau of 1 h for data acquisition.
X-ray photoelectron spectroscopy experiments (XPS) were per-
formed using a Vacuum Generators Escalab 220XL spectrom-
eter. A monochromatized aluminium source (1486.6 eV) was
used for excitation. All binding energies were adjusted with
the binding energy (BE) of C 1s (285.1 eV) as internal ref-
erence. Ex situ thermal activation treatments at atmospheric
pressure were performed in a fixed-bed flow reactor and the
aged samples were transferred in the analysis chamber under
ultra high vacuum (∼10−10 Torr) for XPS analysis. In situ ex-
periments were carried out in a catalytic chamber coupled to
the analyser.
Previous investigations dealing with the catalytic decompo-
sition of N2O (in the absence of reducing agent) pointed out
the highest activity of Rh2O3 among the different transition
metal oxides earlier investigated [7]. Rh2O3 is also more ac-
tive than Rh0 but the active sites involved in the decomposition
have not been clearly elucidated. In addition, surface recon-
structions may take place up to a certain limit of oxygen cov-
erage leading to significant changes in the catalytic behaviour
[8,9]. The usual strong oxygen inhibiting effect on the rate of
N2O decomposition can be lowered after deposition and sta-
bilisation of oxidic Rh species in interaction with reducible
supports such as La- and Pr-doped CeO2 [10,11]. The segre-
gation of well-dispersed RhOx species on CeO2 could be a key
parameter [12]. In fact, oxygen from the dissociation of N2O
on Rh could diffuse on reducible supports then desorb with
a lower activation barrier. The accumulation of chemisorbed
O atoms and the competition for adsorption of N2O and NO
have to be taken into account from a practical viewpoint par-
ticularly for end-of-pipe technologies implemented for nitric
acid plants. NO may originate the formation of stable nitrites
and/or nitrates which could also have a strong detrimental effect
on the catalytic performances. Taking into account those con-
siderations, the stabilisation of well-dispersed PdOx entities in
strong interaction with the perovskite structure could be attrac-
tive for promoting the decomposition of N2O and may represent
an interesting practical issue for the replacement of Rh based
catalysts. According to this objective, we have implemented
an alternative strategy than that previously developed by Uen-
ishi et al. [6] for the promotion of well-dispersed Pd/LaCoO3
with lower Pd content and higher surface Pd concentration at
the surface. In fact, it consists of a two-step procedure with
a conventional wet impregnation of Pd precursor and appro-
priate successive reductive and oxidative thermal treatments.
Those thermal treatments are accompanied with surface struc-
tural modifications which favour the re-dispersion of oxidic pal-
ladium species. According to this procedure, a significant rate
enhancement is observed which cannot be related to weak inter-
actions between PdO and LaCoO3 but mainly to the occurrence
of re-dispersion processes which could be governed by surface
reconstructions of the perovskite structure during those succes-
sive thermal ageing. In that case, oxidic Pd species strongly
interacting with LaCoO3 would be stabilised in unusual oxi-
dation states.
2.2. Catalytic measurements
Temperature-programmed and steady state experiments were
performed in a fixed-bed flow reactor using 0.7 g of catalyst in
powder form. The total flow rate was 15 L h−1, corresponding
to a space velocity of approximately 10,000 h−1. The reac-
tant mixture was typically composed of 1000 ppm N2O and
1000 ppm NO. The effluents were analysed with a µGC Var-
ian 4900 chromatograph fitted with two thermal conductivity
detectors and a Balzer mass spectrometer for the detection and
the quantification of O2 and NO2. Prior to quantification, reac-
tants and products were separated on 5 Å molecular sieve and
Poraplot Q columns. The temperature was gradually increased
during temperature-programmed experiments with a constant
2. Experimental
2.1. Catalyst preparation and characterisation
The preparation procedure of LaCoO3 (20 m2 g−1) was de-
scribed elsewhere [13,14] using a so-called sol–gel method
heating rate of 3 ◦C min−1
.

J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
39
◦
−1
Fig. 1. Temperature-programmed reduction experiments (10 C min ) un-
der 5 vol% H diluted in Ar on calcined LaCoO (a) and reference PdO/
2
3
LaCoO (b).
3
Fig. 2. In situ XRD analysis of LaCoO during reductive thermal treatment in
3
◦
◦
◦
3 vol% H diluted in He: calcined LaCoO (a), 100 C (b), 200 C (c), 300 C
2
3
◦
◦
(d), 400 C (e), 500 C (f).
3. Results
3.1. Bulk and surface characterisation of supported palladium
catalysts
3.1.1. Freshly prepared catalysts
The reducibility of LaCoO3 and PdO/LaCoO3 has been in-
vestigated by H2 temperature-programmed reduction. Results
are reported in Fig. 1. The H2 consumption profile vs temper-
ature recorded on LaCoO3 exhibits two consumption ranges
highlighting a two-step reduction process. The low temperature
range, between 200 and 450 ◦C, is associated to the reduction of
Co3+ into Co2+ according to the calculation of the atomic H/Co
ratio equal to 1.04 0.05. The high temperature consumption
range above 450 ◦C, with H/Co = 1.96 0.09, corresponds to
an extensive reduction of Co2+ to metallic Co particles. As ob-
served in Fig. 1b, a significant shift towards lower temperatures
is observed in the presence of palladium which suggests that the
partial reduction of Co3+ into Co2+ could be catalysed by the
presence of metallic palladium particles.
Such observations are in qualitative agreement with in situ
XRD measurements on LaCoO3 under reductive atmosphere
(3 vol% H2 in He). XRD patterns in Fig. 2 reveal a significant
shift of the most intense X-ray line at 2θ = 32.9◦ ascribed to the
rhomboedral structure of LaCoO3 up to 300 ◦C due to the par-
tial reduction of Co(III) located in the B site of the perovskite
structure. A subsequent increase in temperature at 500 ◦C leads
to a strong attenuation and a complete disappearance of X-ray
lines characteristic of LaCoO3. Correlatively, an additional con-
tribution develops at 2θ = 29.9◦ which characterises La2O3.
The reductive pre-treatment induced effect on the surface prop-
erties of LaCoO3 has been investigated by XPS. As shown in
Fig. 3, the Co 2p3/2 spectra recorded on calcined LaCoO3 and
after reduction at 500 ◦C are respectively located at 780.0 and
778.1 eV. An intermediate BE value is obtained after reduction
Fig. 3. XPS Co 2p spectra recorded after different in situ reductive thermal treat-
◦
ments (10 vol% H in N ): calcined LaCoO (a), pre-reduction at 300 C (b),
2
2
3
◦
pre-reduction at 500 C (c).
in mild conditions at 300 ◦C accompanied with the develop-
ment of the characteristic shake up structure at 785.6 eV which
evidences the presence of Co2+. Semi-quantitative analysis in
Table 1 indicates a complete surface reduction of Co3+ to Co2+
at 300 ◦C while Co0 and Co2+ coexist after reduction at 500 ◦C.
Prior to reaction, PdO/LaCoO3 was pre-activated under isother-
mal conditions in pure H2 overnight at 250 or 500 ◦C. As in-
dicated in Table 2, pre-reduction in mild conditions at 250 ◦C
leads to a complete reduction of PdO into Pd0 with BE value for
Pd 3d5/2 core level characteristic of metallic Pd particles [18]
whereas the bulk structural properties of LaCoO3 are preserved
with complete reduction of Co3+ into Co2+ at the surface in
accordance with previous XPS measurements reported in Ta-
ble 1. On the other hand, an extensive bulk reduction of the
perovskite into CoOx and La2O3 was observed at 500 ◦C yield-
ing Pd0/CoOx/La2O3.

40
J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
Table 1
Changes in oxidation state and surface concentration of oxidic cobalt species in LaCoO after exposure to reductive thermal treatments
3
a
Preactivation thermal
treatment
BE (eV)
Relative surface composition (%)
3+
2+
0
Co 2p
core level
Co
Co
Co
3/2
Calcined LaCoO
Pre-reduced in H at 300 C
Pre-reduced in H at 500 C
780
779.7
778.1
100
–
–
–
100
40
–
–
60
3
◦
2
◦
2
a
BE—binding energy.
Table 2
Changes in surface composition of Pd supported on LaCoO under controlled oxidative and reductive atmospheres
3
a
Thermal
pretreatment
XPS analysis
H
H
chemisorption
2
2
c
d
b
Peak width
(eV)
Relative surface atomic composition
uptake
D
BE (eV)
−1
Pd 3d
5/2
(µmol g
)
Pd/La
Co/La
O/La
◦
Calcination in air at 400 C
336.8
335.4
335.1
337.0
337.2
337.3
1.6
1.6
1.6
3.8
1.6
2.0
0.04
0.04
0.04
0.05
0.11
0.03
0.45
0.60
0.62
0.46
0.63
0.81
2.9
3.6
2.4
2.7
2.7
3.6
◦
Reduction in H at 250 C
2
◦
Reduction in H at 500 C
7.5
8.0
0.16
0.17
2
◦
◦
Reduction at 250 C and ageing in N O/NO at 250 C
Reduction at 250 C and ageing in N O/NO at 500 C
Reduction at 500 C and ageing in N O/NO at 400 C
2
◦
◦
4.95
5.0
0.105
0.11
2
◦
◦
2
a
b
c
d
◦
Performed at 100 C.
BE—binding energy.
Relative accuracy equal to 20%.
Metal dispersion.
3.1.2. Influence of the thermal ageing under reactive
conditions on surface properties of pre-reduced supported Pd
on LaCoO3
According to the nature of the activation thermal treat-
ment and successive thermal ageing under oxidative conditions,
metallic or oxidic Pd species can be segregated as Pd(II) into
PdO or in different chemical environments depending on the
extent of interaction with the support and related changes in
structural properties. Under reactive conditions, in the presence
of NO and N2O, different chemical processes can occur on
pre-reduced catalysts at 500 ◦C associated to bulk and surface
re-oxidation as well as catalytic reactions involving both mole-
cules on re-oxidised surfaces. Pre-reduced catalysts have been
aged in the reactive mixture at high temperature overnight in or-
der to accelerate bulk re-oxidation processes and to characterise
the surface properties involved in the catalytic decomposition of
NO and N2O.
3.1.2.1. Bulk characterisation XRD patterns recorded after
ageing in 1000 ppm NO and 1000 ppm N2O diluted in He
overnight at 500 ◦C are reported in Fig. 4. Clearly, the rhom-
boedral structure prevails on the aged samples previously re-
duced at 250 ◦C with characteristic X-ray lines located at
the same position in 2θ as those observed on the reference
PdO/LaCoO3 sample. On the contrary, the perovskite structure
is not evidenced on the aged sample initially reduced at 500 ◦C
with X-ray lines characteristic of La2O3. Bulk Co species are
also detectable at 2θ = 37◦ associated with the segregation of
Co3O4. H2-TPR experiments were performed on the aged cat-
alysts in order to examine changes in the reducibility of oxidic
cobalt species after thermal ageing. The effluents during TPR
experiments were simultaneously analysed by mass spectrome-
Fig. 4. Ex situ XRD analysis of Pd/LaCoO catalysts after successive pre-
3
activation treatment and thermal ageing under reactive conditions in the pres-
◦
ence of 1000 ppm NO and 1000 ppm N O: calcination at 400 C (ref-
2
◦
erence PdO/LaCoO ) (a), pre-reduced at 250 C and aged under N O/NO
at 500 C (b), pre-reduced at 250 C and aged under N O/NO at 250 C (c),
3
2
◦
◦
◦
2
◦
◦
pre-reduced at 500 C and aged under N O/NO at 400 C (d).
2
try. H2 consumption profiles versus temperature are collected
in Fig. 5. As observed the reducibility of oxidic Co species
after ageing strongly depends on the temperature of the pre-
reductive thermal treatment in H2. Under mild conditions, with
a reduction temperature of 250 ◦C, the shape of the profiles
approximately mimics that previously reported on a reference
PdO/LaCoO3 with two distinct H2 consumption ranges corre-

J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
41
consumption range and the development of a broad signal be-
tween 150 and 500 ◦C. Such an observation is well in agreement
with XRD measurements which did not show the restoration of
the rhomboedral structure of LaCoO3 after successive reductive
and ageing thermal treatments respectively at 500 and 400 ◦C.
The calculation of atomic H/Co ratio, equal to 2.87, is slightly
higher than the theoretical one for the bulk reduction of Co3O4.
Coming back to Fig. 5, b and c, relative to pre-reduced samples
in mild conditions (T = 250◦C) then aged under reactive mix-
ture at 250 and 500 ◦C, respectively, subsequent comparisons
with the reference PdO/LaCoO3 provides interesting informa-
tion relative to changes of the reducibility of oxidic palladium
and cobalt species. A particular attention is focused on the low
temperature H2 consumption range in Fig. 5b. As exemplified,
the H2 consumption starts at lower temperature on the aged
sample at 250 ◦C than on the reference PdO/LaCoO3. On the
other hand, such a low temperature process strongly attenuates
on the aged sample at 500 ◦C (see Fig. 5c). Such differences
will be discussed more extensively in the light of additional
surface characterisation. The calculation of the atomic H/Co
ratio leads to an unexpectedly low value of 2.73 for the aged
sample in mild conditions at 250 ◦C in comparison with the the-
oretical value of 3 corresponding to the complete reduction of
LaCoO3 into metallic Co species and La2O3. Such a low value
may underline an incomplete re-oxidation of the catalyst dur-
ing thermal ageing. On the other hand, the value recorded after
ageing at 500 ◦C exceeds the theoretical one. Mass spectrom-
etry analysis provides complementary information. As exem-
plified in Fig. 5, CO2 and N2O, associated to the relative mass
m/z = 44, and CO and N2, corresponding to m/z = 28, are
detected above 500 ◦C. Their presence could be explained by
the accumulation of carbonates during contact in air at room
temperature and/or by the accumulation of nitrate and nitrite
compounds during thermal ageing. Their reduction in the pres-
ence of H2 at high temperature above 500 ◦C may take place
and explain the extra consumption of hydrogen. It is worthwhile
to note that the production of carbon and/or nitrogen-containing
products is more pronounced on the pre-reduced Pd/LaCoO3 at
500 ◦C and aged at 400 ◦C and coincides with the high temper-
ature H2 consumption signal.
3.1.2.2. Surface characterisation XPS analysis provides ad-
ditional data relative to changes in the oxidation state and chem-
ical environment of palladium and cobalt in comparison with
PdO/LaCoO3. Co 2p spectra recorded on the aged samples
are listed in Fig. 6. As illustrated no significant change is dis-
cernable according to the temperature of the pre-reduction and
ageing processes. The binding energy value of the Co 2p3/2
core level associated with the shake up structure respectively
at 780.0 and 789.5 eV essentially characterise the presence of
Co3+. The presence of Co2+ cannot be ruled out particularly
on the pre-reduced and aged sample respectively at 500 and
400 ◦C associated to the segregation of Co3O4. As exemplified
in Fig. 6b, the shake up structure relative to Co2+ at 785.6 eV
may contribute to the overall signal. Fig. 7 shows Pd 3d spectra
recorded after different ex situ thermal treatments under con-
trolled atmospheres. The photopeak Pd 3d5/2 recorded on the
Fig. 5. Temperature-programmed reduction experiments under 5 vol% H di-
2
luted in Ar on aged-samples under reactive conditions in the presence of
1000 ppm NO and 1000 ppm N O: H -consumption profiles recorded on
2
2
◦
◦
reference PdO/LaCoO (a); pre-reduced at 250 C and aged at 250 C (b),
pre-reduced at 250 C and aged at 500 C (c), pre-reduced at 500 C and aged
at 400 C (d). Simultaneous analysis of the effluents by mass spectrometry with
3
◦
◦
◦
◦
m/z = 28 associated to the presence of N and CO and m/z = 44 associated to
2
the formation of CO and N O in the gas phase.
2
2
sponding to a two-step reduction process with the intermediate
formation of Co2+ species. On the other hand, such a compari-
son cannot be achieved after pre-reduction in severe conditions
at 500 ◦C with a strong attenuation of the high temperature H2

42
J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
ageing in 1000 ppm NO and 1000 ppm N2O at 250 ◦C leads
to a significant broadening of the Pd 3d5/2 photopeak which
highlights the segregation at the surface of different palladium
species stabilised in different oxidation states, essentially Pd0
and Pd2+ (Fig. 7d). A more extensive surface oxidation takes
place with an increase in the thermal ageing temperature at
500 ◦C associated with a shift of the Pd 3d5/2 at 337.2 eV
which only characterises the presence of Pd2+. The slightly
higher value than that obtained on PdO/LaCoO3 may reflect
different chemical environment for Pd2+ species. Similar ob-
servations were reported by Uenishi et al. [6] with the obser-
vation of oxidic palladium species exhibiting binding energy
values between those currently reported for Pd2+ in PdO and
Pd4+ in PdO2, respectively, equal to 336.2 and 338 eV [19].
Those authors suggested the stabilisation of unusual Pd(III)
species assisted by the perovskite structure. Interestingly, a sig-
nificant increase in intensity signal is discernible in compar-
ison with that obtained on the reference PdO/LaCoO3 which
indicates a higher surface concentration of palladium species
after successive oxidative/reductive thermal treatments. Spec-
tral features are summarised in Table 2 with semi-quantitative
analysis. Clearly, the surface Pd composition is higher after suc-
cessive reduction and thermal ageing treatment respectively at
250 and 500 ◦C. Regarding the pre-reduced sample at 500 ◦C,
a drop in surface Pd composition is observed. Parallel to this
observation a significant increase in surface Co concentration
is observed which cannot be related to the surface reconstruc-
tion of LaCoO3 but more probably to the segregation of Co3O4
according to XRD observations. The examination of the pho-
topeak O 1s provides additional arguments. As observed in
Figs. 8a and 8b, two contributions arise at 529.1 0.2 eV and
531.3 0.2 eV on the O 1s photopeak which have been previ-
ously assigned to oxygen species from the perovskite structure
and OH groups on the surface [20,21]. Earlier investigations
[22,23] revealed changes in the relative intensity of both con-
tributions according to the extent of reduction. Our observa-
tions are in qualitative agreement with those previous findings
with an increase of OH groups after reduction and the devel-
opment of the 529.1 eV contribution after re-oxidation except
on the pre-reduced sample at 500 ◦C since an additional con-
tribution appears at 533 eV (see Fig. 8b). Complementary ob-
servations can be provided from the examination of the N 1s
photopeak. As exemplified in Fig. 9, XPS spectra are domi-
nated by two contributions on the overall signals at 403.7 and
407.3 eV previously assigned to nitrites and nitrates species
respectively [18,24]. Both contributions observable prior reduc-
tion in H2 on PdO/LaCoO3 attenuate on aged catalysts previ-
ously reduced under mild conditions at 250 ◦C. On the other
hand they intensify on the aged sampled previously reduced
in severe conditions at 500 ◦C. Such an observation is in rel-
ative good agreement with H2-TPR experiments (see Fig. 5)
with the simultaneous formation N-containing products in the
outlet gas mixture above 500 ◦C. This formation more accentu-
ated on the pre-reduced and aged sample, respectively, at 500
and 400 ◦C could be associated to reduction of nitrates and ni-
trites in the presence of hydrogen. Consequently the shoulder
on the O 1s spectra at 533 eV could be partly explained by
Fig. 6. Co 2p spectra recorded on aged Pd/LaCoO catalysts under reactive
3
conditions (1000 ppm NO and 1000 ppm N O): reference PdO/LaCoO (a),
2
3
◦
◦
◦
pre-reduced at 500 C and aged at 400 C (b), pre-reduced at 250 C and aged
at 250 C (c), pre-reduced at 250 C and aged at 500 C (d).
◦
◦
◦
Fig. 7. Pd 3d spectra recorded on Pd/LaCoO catalysts after various ex situ
3
◦
thermal ageing: calcination at 400 C (reference PdO/LaCoO ) (a), pre-reduced
3
◦
◦
◦
at 500 C (b), pre-reduced at 500 C and aged at 400 C under N O/NO (c),
pre-reduced at 250 C and aged at 250 C under N O/NO (d), pre-reduced at
250 C and aged at 500 C under N O/NO (e).
2
◦
◦
2
◦
◦
2
calcined sample in air at 400 ◦C located at 336.8 eV mainly
characterises Pd2+ probably into PdO. The current value ear-
lier reported for PdO deposited on conventional supports such
as alumina are slightly lower around 336.2 eV [18]. Such a shift
could be significant and may reflect a more extensive interac-
tion between Pd2+ species on LaCoO3 than on alumina. A shift
towards lower BE values near 335.1 eV is observed after re-
ductive thermal treatment in H2 at 500 ◦C which characterises
the presence of metallic Pd particles. Similar information can
be obtained after reduction at 250 ◦C (see Table 2). Thermal

J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
43
Fig. 9. N 1s spectra recorded on Pd/LaCoO catalysts: (a) fresh catalyst (refer-
3
(a)
◦
ence PdO/LaCoO ); (b) pre-reduced under hydrogen at 250 C and aged under
3
◦
◦
inlet mixture at 250 C; (c) pre-reduced under hydrogen at 250 C and aged un-
der inlet mixture at 500 C; (d) pre-reduced under hydrogen at 500 C and aged
under inlet mixture at 400 C.
◦
◦
◦
Experiments were achieved at 100 ◦C to minimise the forma-
tion of bulk hydride species [15]. In all cases, the aged samples
were reduced at 500 ◦C for further comparison with the ref-
erence PdO/LaCoO3. The values of the metal dispersion are
reported in Table 2. As observed a loss of metal dispersion is
distinguishable after ageing except for the reduced and aged
sample under mild conditions (D = 0.17). The comparison of
XPS and H2 titration results obtained on the pre-reduced and
aged samples respectively at 250 and 500 ◦C reveals some di-
vergences. As indicated in Table 2 this sample exhibits the
highest atomic Pd/La ratio and the lowest metal dispersion.
As a matter of fact, such a comparison could not be contro-
versial since XPS analysis does not concern the topmost layer
but provides an average composition on a depth representing
approximately ten mono-layers. Also an additional aspect has
to be taken into account suggested by TPR experiments which
show different reducibility according to the temperature of the
thermal ageing process. In this sense, the highest metal dis-
persion obtained after ageing at 250 ◦C could be correlated to
the development of the low H2 consumption range in Fig. 5b
and associated to the segregation of well-dispersed PdO parti-
cles [25]. Now, different explanation could be proposed con-
cerning to disappearance of those species on the aged sample at
500 ◦C which may explain the loss of metal dispersion. Thermal
sintering processes may occur but a more extensive redispersion
phenomenon could also be accounted for with a partial incorpo-
ration oxidic Pd species inside the perovskite structure. Such a
process could generate less reducible oxidic palladium species
and then explain the underestimation of the metallic palladium
dispersion. Such an explanation seems in good agreement with
the shift of the Pd 3d5/2 core level towards higher values which
suggests changes in the chemical environment of oxidic palla-
dium species (see Fig. 7).
(b)
Fig. 8. O 1s spectra recorded on Pd/LaCoO catalysts: (a) fresh catalyst (ref-
3
◦
erence PdO/LaCoO ) (·- · -), pre-reduced under hydrogen at 250 C and aged
3
◦
◦
under inlet mixture at 500 C (· · ·), pre-reduced under hydrogen at 250 C and
◦
aged under inlet mixture at 250 C (—); (b) fresh catalyst (—), pre-reduced un-
◦
◦
der hydrogen at 500 C (·- · -), pre-reduced under hydrogen at 500 C and aged
◦
under inlet mixture at 400 C (· · ·).
the accumulation of strongly chemisorbed nitrites and nitrates
from NO adsorption and successive surface reactions between
those chemisorbed NO and O atoms leading to the formation
of nitrates. One cannot rule out the contribution of carbonates
species on the O 1s spectra as previously observed by Natile et
al. [21]. Nevertheless no noticeable change in intensity of the
photopeak C 1s was observed according to thermal ageing con-
ditions (results not shown) which could explain changes in the
intensity of relative masses m/z = 44 and 28 due to the decom-
position and/or reduction of carbonates (see Fig. 5).
H2 chemisorption measurements were performed on aged
samples in order to examine changes in the metal dispersion.

44
J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
Fig. 10. Temperature-programmed experiment of the catalytic decomposition of
(a)
◦
N O to N and O on Pd/Al O pre-reduced at 250 C and 1000 ppm N O di-
2
2
2
2
3
2
luted in He: N O conversion (Q), N concentration (2), O concentration (!).
2
2
2
3.2. Catalytic performances in the decomposition of N2O on
pre-reduced Pd catalysts supported on Al2O3 and LaCoO3
3.2.1. Freshly prepared Pd/Al2O3 and Pd/LaCoO3
Preliminary experiments were performed in the absence of
NO on reference Pd/Al2O3 catalysts. The conversion profile
and the amount of N2 and O2 produced from the decomposi-
tion vs temperature are reported in Fig. 10. The decomposition
of N2O is accompanied with the simultaneous production of
N2 and O2. N2O conversion starts at 400 ◦C and then reach-
ing a maximum at 500 ◦C of approximately 40%. A signifi-
cant shift of the N2O conversion curve towards higher tem-
perature is observable in the presence of NO (see Figs. 11a
and 11b) which underlines an inhibiting effect of NO on the
rate of N2O conversion. A weak conversion of NO is notice-
able in the whole temperature range of the study which does
not exceed 2.5%. Different observations have been obtained
on pre-reduced Pd/LaCoO3 at 250 ◦C with the occurrence of
two conversion ranges in Fig. 12. The low temperature con-
version range, with no significant observation of oxygen in the
gas phase, can be assigned to the re-oxidation of the solid by
N2O. The catalytic decomposition takes place above 400 ◦C.
It is worthwhile to note that a more extensive conversion than
that recorded on Pd/Al2O3 is obtained of approximately 90%
at 500 ◦C. Complex features are observable in the presence of
NO (results not shown) since N2O and NO compete for the
re-oxidation. In that case a poor production of oxygen was
observed above 470 ◦C which signifies a strong detrimental
effect of NO on the catalytic decomposition of N2O. Similar
temperature-programmed experiments in the presence of NO
and N2O has been performed on LaCoO3 and PdO/LaCoO3
calcined in air at 400 ◦C. In that case, no surface re-oxidation
takes place. As observed in Figs. 13a and 13b, a slightly lower
N2O conversion is observed than that obtained on pre-reduced
Pd/LaCoO3 then re-oxidised under reactive mixture, starting
(b)
Fig. 11. Temperature-programmed experiment of the catalytic decomposition of
N O to N and O in the presence of NO on Pd/Al O pre-reduced at 250 C
and 1000 ppm N O and 1000 ppm NO diluted in He: (a) N O conversion (Q),
NO conversion (1); (b) N concentration (2), O concentration (!).
◦
2
2
2
2 3
2
2
2
2
at the same temperature of 400 ◦C but reaching 60% against
90% in the absence of NO. Parallel to this observation a weak
NO conversion occurs between 200 and 380 ◦C which is not
correlated to the formation of products in the gas phase (see
Fig. 13b). Such observations could be related by the stabilisa-
tion of nitrates and/or nitrites species on PdO/LaCoO3. Further
comparison with calcined LaCoO3 leads to comparable activ-
ity in N2O conversion. The main difference is associated to a
significant attenuation of NO conversion.
3.2.2. On aged supported palladium catalysts
The conversion profiles of N2O vs temperature in the pres-
ence of NO on pre-reduced catalysts then aged overnight

J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
45
(a)
Fig. 12. Temperature-programmed experiment of the catalytic decomposition
◦
of N O to N and O on Pd/LaCoO pre-reduced at 250 C and 1000 ppm
2
2
2
3
N O diluted in He: N O conversion (Q), N concentration (2), O concentra-
2
2
2
2
tion (!).
under reactive conditions at 250 and 500 ◦C are reported in
Fig. 14a. The correlative formation of oxygen in the gas
phase in Fig. 14b emphasises the fact that the catalytic de-
composition of N2O predominantly occurs with no signif-
icant involvement of side reactions associated to bulk re-
oxidation process of the solid. As exemplified, significant
changes in the activity are observable. A qualitative estima-
tion of the overall activity in the decomposition of N2O can
be obtained from the T50 (temperature corresponding to 50%
N2O conversion) which increases according to the following
sequence: PdO/LaCoO3 (T50 = 481◦C) = Pd/LaCoO3 Red.
500 ◦C–Ox. 400 ◦C (T50 = 480◦C) < LaCoO3 Red. 250 ◦C–
Ox. 500 ◦C (T50 = 465◦C) < Pd/LaCoO3 Red. 250 ◦C–Ox.
250 ◦C (T50 = 460◦C) < Pd/LaCoO3 Red. 250 ◦C–Ox. 500 ◦C
(T50 = 433◦C). This comparison leads to different comments.
First, a better catalytic activity in decomposition of N2O is
observed on LaCoO3 after pre-reduction and ageing than on
the calcined LaCoO3 (see Fig. 13a). However the most inter-
esting observation is related to the significant rate enhance-
ment observed on Pd/LaCoO3 Red. 250 ◦C–Ox. 500 ◦C with
N2O conversion starting above 350 ◦C. On the contrary the
loss of the perovskite structure after pre-reduction at high
temperature has a detrimental effect on the activity. Also the
reference PdO/LaCoO3 exhibits a lower activity which under-
lines that surface and bulk reconstructions which take place
during thermal ageing under reactive conditions lead to dif-
ferent surface properties than those obtained on the reference
PdO/LaCoO3 catalyst with the promotion of cooperative ef-
fects between oxidic palladium species and the perovskite
which may originate such an extra conversion. The calcula-
tion of the apparent activation energy from temperature ex-
periments are reported in Table 3. They have been compared
to the intrinsic rates roughly estimated from conversion val-
ues recorded at 460 ◦C and according to the number surface
(b)
Fig. 13. Temperature-programmed experiments of the catalytic decomposition
of N O to N and O in the presence of NO on calcined LaCoO (open sym-
bols) and PdO/LaCoO (full symbols)–1000 ppm N O and 1000 ppm NO
diluted in He: (a) N O conversion (P, Q), NO conversion (!, "); (b) N con-
centration (1, 2), O concentration (E, F).
2
2
2
3
3
2
2
2
2
atoms calculated from H2 titration. As observed significant
changes occur according to the nature of the support. Clearly,
higher values for the apparent activation energy (Eapp) are
obtained on alumina based catalysts which agree with lower
intrinsic rate than that estimated on perovskite based catalysts.
The comparable values obtained for Eapp and the intrinsic rate
on the reference PdO/Al2O3 and after pre-reduction in pure
H2 at 250 ◦C of respectively 93 5 and 89
5 kJ mol−1
,
with variation within the margin of error, suggest that when
the decomposition of N2O takes place at relatively high tem-
perature (up to 400 ◦C), the catalyst sample exhibits quasi-
similar surface properties irrespective of the nature of the
pre-activation thermal treatment. Now regarding Pd supported
on perovskite, significant lower values have been calculated
for Eapp between 42 3 and 55
3 kJ mol−1 for the ref-

46
J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
Table 3
Influence of thermal ageing on the catalytic performances of Pd/Al O and Pd/LaCoO catalysts in the decomposition of N O
2
3
3
2
a
b
Catalyst
Thermal treatment
T
N O conversion
2
at 460 C
Intrinsic rate
(mol h Pd
E
app
50
−1
◦
◦
−1
−1 c
( C)
)
(kJ mol
)
surf
◦
PdO/Al O
Calcination in air at 400 C
510
520
481
460
433
480
0.15
0.04
0.32
0.48
0.74
0.30
1.6
0.4
19
26
66
25
93
89
55
51
42
52
5
5
3
3
3
3
2
3
◦
Pd/Al O
Reduction in H at 250 C
2
3
2
◦
PdO/LaCoO
Calcination in air at 400 C
3
◦
◦
Pd/LaCoO
Pd/LaCoO
Pd/LaCoO
Reduction in H at 250 C ageing in N O/NO at 250 C
3
3
3
2
2
◦
◦
Reduction in H at 250 C ageing in N O/NO at 500 C
2
2
◦
◦
Reduction in H at 500 C ageing in N O/NO at 400 C
2
2
a
◦
Calculated at 460 C from TPR experiments.
b
c
◦
Calculated at 460 C from the number of metallic Pd sites calculated from H titration measurements.
Apparent activation energy.
2
Fig. 15. Steady state catalytic measurements of pre-reduced and aged Pd/
LaCoO in the decomposition of N O in the presence of NO: N O conversion
3
2
2
◦
◦
after pre-reduction at 250 C and thermal ageing under N O/NO at 250 C (!),
pre-reduction at 250 C and thermal ageing under N O/NO at 500 C (") O
2
◦
◦
2
2
◦
production after pre-reduction at 250 C and thermal ageing under N O/NO
2
◦
◦
at 250 C (P), pre-reduction at 250 C and thermal ageing under N O/NO at
2
◦
500 C (Q).
Fig. 14. Ageing induced effect on the catalytic performances in the decompo-
periments emphasise the fact that the decomposition of N2O
predominantly takes place in both cases. Interestingly, a good
agreement is obtained between temperature-programmed and
steady state conversions recorded at 460 ◦C on the aged sample
at 500 ◦C of, respectively, 0.74 and 0.77. Such a convergence
agrees with the fact that both experiments characterise sim-
ilar surface properties. On the other hand, some divergences
arise by examining temperature-programmed and steady state
measurements with a higher conversion on the aged sample
at 250 ◦C obtained in this later case (0.67 against 0.48 from
TP experiments). Such an increase in activity probably reflects
additional surface modifications activated at 460 ◦C. In those
temperature conditions a more extensive re-dispersion and the
stabilisation of oxidic palladium species may originate this ex-
tra conversion.
◦
sition of N O in the presence of NO on LaCoO pre-reduced at 250 C and
2
3
◦
aged under N O/NO at 500 C (2) and on freshly-prepared PdO/LaCoO cal-
cined at 400 C (Q); then pre-reduced at 250 C and aged under N O/NO at
250 C (!); pre-reduced at 250 C and aged under N O/NO at 500 C (F);
pre-reduced at 500 C and aged under N O/NO at 400 C (1); (a) N O con-
2
3
◦
◦
2
◦
◦
◦
2
◦
◦
2
2
version, (b) O concentration.
2
erence PdO/LaCoO3 which highlight a significant beneficial
effect towards the conversion of N2O. More interesting is the
lower value obtained after ageing in the reactant mixture at
500 ◦C on pre-reduced Pd/LaCoO3 in pure H2 at 250 ◦C which
is consistent with the highest conversion observed on this cata-
lyst.
Steady state experiments were performed on the most ac-
tive catalysts obtained after successive reductive (H2 reduction
at 250 ◦C) and ageing thermal treatment, respectively, at 250
and 500 ◦C. The conversion of N2O and the correlative O2 pro-
duction are reported in Fig. 15. As observed, stable conversions
versus time on stream are observable. Similar catalytic features
as those previously reported from temperature-programmed ex-
3.2.3. Comparative study of freshly-prepared and aged
catalysts in the decomposition of N2O
The conversion profile vs temperature recorded on the
freshly-prepared Pd/LaCoO3 earlier reported in Fig. 12 during

J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
47
of supported palladium catalysts in the decomposition of N2O.
Such an extent of interaction may improve the Pd dispersion but
also the intrinsic adsorption properties of palladium sites. These
interpretations are supported by kinetic data reported in Table 3
which show significant changes on the apparent activation en-
ergy and on the estimates of the intrinsic rate according to the
nature of the support. As indicated there is a good agreement
between higher intrinsic rates and lower apparent activation en-
ergies on palladium supported on perovskite.
Coming back to the influence of the nature of the pre-
activation thermal treatment on the overall activity in the de-
composition of N2O, only weak variations in the conversion, in-
trinsic rate and apparent activation energy values are discernible
on PdO/Al2O3 and pre-reduced Pd/Al2O3 which suggest that
both catalysts exhibit similar surface properties at high temper-
ature. Such a tendency could be related to an extensive accu-
mulation of oxygen atoms from the dissociation of NO and/or
N2O at the surface and/or a subsequent surface oxidation of
metallic Pd particles during exposure under reactive conditions
at high temperature. Consequently, oxidic palladium species as
PdOx would be probably involved in the reaction irrespective
of the nature of the pre-activation thermal treatment. As pre-
viously mentioned, the lower apparent activation energies ob-
served when PdO interacts with LaCoO3 than those calculated
on alumina indicate significant modifications of the reactivity
of chemisorbed N2O molecules towards the dissociation and/or
the adsorptive properties of oxidic palladium species in interac-
tion with LaCoO3. Such arguments are in qualitative agreement
with XPS measurements which indicate a significant shift of
the Pd 3d5/2 core level towards higher BE values on the refer-
ence PdO/LaCoO3 catalyst (336.8 eV against 336.1–336.4 eV
on bulk and supported PdO on alumina [18]). Such observa-
tions could be associated to different chemical environments of
Pdn+ species on alumina and perovskite after calcination in air
at 400 ◦C with correlative changes in the adsorptive properties
of those oxidic Pd species. As a matter of fact, such changes
in electronic properties of palladium entities may also reflect
the creation of new active sites at the metal/support interface
involving Pd2+ species and nearest neighbour anionic vacan-
cies which could be potentially active in the decomposition of
N2O. There is no direct evidence of those active sites but H2
temperature-programmed experiment provide additional obser-
vations which underline a significant effect of palladium on the
reduction of Co3+ into Co2+. Based on this observation, palla-
dium species could increase the reactivity of oxygen species
from the perovskite. In the absence of reducing agent, their
desorption could be assisted by the presence of palladium at
high temperature with a subsequent formation of anionic va-
cancies. Such an explanation seems to be in rather good agree-
ment with previous investigation of the CO + O2 reaction on
Pt/CeO2 [26,27]. Those authors proposed two interpretations
for explaining the effect of platinum on the reducibility of ce-
ria by CO: (i) A weakening of the Pt–CO bond when platinum
interact with reduced ceria. (ii) A charge transfer from metal to
ceria accompanied with a small increase in the oxidation state
of the metal and a decrease of the Ce–O bond strength correla-
tively [28]. Both considerations seem to be in good agreement
Fig. 16. Comparison of the catalytic activity of Pd/LaCoO submitted to differ-
3
ent activation treatments before temperature-programmed reaction: conversion
◦
of N O in the absence of NO on pre-reduced catalyst at 250 C (!), conversion
2
◦
of N O in the absence of NO on the pre-reduced catalyst at 250 C and aged
under N O/NO at 500 C (Q), conversion of N O with the presence of NO on
the pre-reduced catalyst at 250 C and aged under N O/NO at 500 C (2).
2
◦
2
2
◦
◦
2
the decomposition of N2O in the absence of NO is compared
with those obtained on aged catalysts in the presence and in
the absence of NO, initially pre-reduced at 250 ◦C and exposed
overnight to the reactive mixture at 500 ◦C (see Fig. 16). Clearly
higher activities in the conversion of N2O are observable on the
aged catalyst irrespective of the feed gas composition, the light-
off temperature shifting from 452 to 405 ◦C in the absence of
NO and being intermediate (T50 = 432◦C) in the presence of
NO. It is also noticeable that no NO conversion related to the
formation of nitrites and/or nitrites significantly occurs on aged
catalyst in comparison with reference PdO/LaCoO3 in Fig. 13.
Such a result agrees with previous observations related to the
inhibiting effect of NO on the rate of N2O conversion earlier
evidenced on PdO/LaCoO3. Probably, this detrimental effect is
attenuated after successive reduction and ageing respectively at
250 and 500 ◦C due to significant redistribution of PdO clus-
ters in strong interaction with the reconstructed perovskite. On
the other hand it could be strengthened after pre-reduction at
500 ◦C because no surface reconstruction of the perovskite is
observed after ageing. Such statements are in relative good
agreement with the accumulation of nitrites and nitrates on this
latter catalyst from XPS observations (see Fig. 9).
4. Discussion
The catalytic performances of Pd/Al2O3 and Pd/LaCoO3 in
the decomposition of N2O in the presence or in the absence
of NO strongly differ. Clearly, freshly-prepared Pd-modified
perovskite catalysts exhibit higher overall activity in the con-
version of N2O in spite of their lower metal dispersion (0.16
against 0.86 on Pd/Al2O3). As a matter of fact, such a difference
is not controversial but simply underlines the significant effect
of the metal/support interaction on the catalytic performance

48
J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
with the tendencies observed on the apparent activation ener-
gies for the decomposition of N2O and on the binding energy
level of the Pd 3d core level on freshly-prepared PdO/Al2O3
and PdO/LaCoO3. The lower apparent activation energy and the
higher intrinsic rate on PdO/LaCoO3 than on PdO/Al2O3 can
be connected to an unexpectedly high value for the binding en-
ergy of the Pd 3d core level in comparison with those currently
reported for PdO on alumina. Such observations may reflect a
charge transfer from Pd to LaCoO3 leading to a small increase
in the oxidation state of Pd and a decrease in the Co–O bond
strength. Such correlations between kinetic and spectroscopic
features suggest a greater ability to generate anionic vacan-
cies via the destabilisation of Co–O bond at the vicinity of Pd
species at high temperature.
nitrites and nitrates species which could also contribute in the
loss of N2O conversion.
5. Conclusion
This study reports the influence of successive reductive and
oxidative pre-activation thermal treatments of supported Pd cat-
alysts in the decomposition of N2O in the absence and in the
presence of NO. A higher activity of palladium sites is ob-
served on LaCoO3 than on Al2O3 which is not only related to
the number of accessible Pd sites but also to the extent of in-
teraction between oxidic palladium species and the support. It
was found that the overall activity strongly depends on the na-
ture of interactions between palladium species and the support
generated during those successive thermal treatments. A high
activity is obtained on PdO/LaCoO3 after reduction in H2 in
smooth conditions at 250 ◦C and re-oxidation after exposure to
NO and N2O at 500 ◦C due to the re-dispersion of palladium
species exhibiting an unusual oxidation state. On the contrary,
a higher reduction temperature has a detrimental effect with
the disappearance of the structure of the perovskite and prob-
ably the growth of palladium particles which affect the extent
of the metal/support interface probably involved in the catalytic
performances. It can be concluded that Pd and the perovskite
have a cooperative effect on the overall activity in the decom-
position of N2O due to the creation of new active sites at the
metal/support interface.
Let us now examine the aged catalysts initially pre-reduced
in pure H2 then re-oxidised overnight under reaction mixture
at high temperature. As previously shown, different catalytic
behaviour has been reported according to the temperature of
the pre-activation treatment in H2. It was found that succes-
sive pre-activation in H2 at 250 ◦C and re-oxidation in reactive
conditions at 500 ◦C promote the catalytic performances which
exceed those recorded on a reference PdO/LaCoO3. On the
other hand, a pre-reduction at 500 ◦C has no significant effect,
the catalytic activity being comparable to that obtained on the
reference PdO/LaCoO3 catalyst. As observed the re-oxidation
under reactive conditions of a pre-reduced catalyst at 500 ◦C
does not lead to the restoration of the perovskite structure but
essentially to the segregation of La2O3 and Co3O4 contrarily
to pre-activation thermal treatment at 250 ◦C. The comparison
of the atomic Pd/La ratio and metal dispersion also indicates
a poorer palladium distribution probably due to thermal sin-
tering in the course of the reduction at 500 ◦C (see Table 2).
More interesting is the examination of the apparent activation
energy calculated on Pd/LaCoO3 Red. 250 ◦C–Ox. 500 ◦C and
the BE of Pd 3d. In fact the previous correlation established
between both kinetic and spectroscopic features is accentuated
with a lower apparent activation energy (42 kJ mol−1 against
55 kJ mol−1 on PdO/LaCoO3) and a higher binding energy for
Pd shifting from 336.8 to 337.2 eV. Consequently, the highest
overall activity of Pd/LaCoO3 Red. 250 ◦C–Ox. 500 ◦C is con-
sistent with a redispersion of less reducible palladium species
stabilised under unusual oxidation state in comparison between
current BE values reported on reference PdO and PdO2 sam-
ples. The reconstruction of the perovskite structure during ther-
mal ageing could be the driving force in determining the extent
of Pd dispersion with a partial incorporation of palladium inside
the structure of the perovskite. Such a configuration would en-
hance the weakening of the Co–O bond with related formation
of anionic vacancies for the dissociation of N2O and explain
the rate enhancement in the conversion of N2O. As observed
pre-reduction at 500 ◦C has a detrimental effect on structural
and catalytic properties due probably to thermal sintering which
would inhibit the reconstruction of the perovskite and related
re-dispersion of Pd species stabilised by the perovskite struc-
ture. In such conditions, oxygen-inhibiting effect could affect
more extensively the rate of N2O decomposition. Oxygen accu-
mulation at the surface would enhance the formation of stable
Acknowledgments
We thank the Institute of Research in Industrial Environment
supported by the Region Nord-Pas-de-Calais and the CNRS
and the ADEME for supporting this research through a grant
(J.P. Dacquin). We gratefully acknowledge Mrs L. Burylo and
Dr. L. Gengembre who conducted XRD and XPS measure-
ments.
References
[1] M.A. Penˇa, J.L. Fierro, Chem. Rev. 101 (2001) 1981.
[2] M. Sadakane, T. Asanuma, J. Kubo, W. Ueda, Chem. Mater. 17 (2005)
3546.
[3] S. Kaliaguine, A. Van Neste, US Patent 6,017,504 (2000).
[4] C.N. Costa, V.N. Stathopoulos, V.C. Belessi, A.M. Efstathiou, J. Catal.
197 (2001) 350.
[5] C.N. Costa, P.G. Savva, C. Andronikou, P.S. Lambrou, K. Polychro-
nopoulou, V.C. Belessi, V.N. Stathopoulos, P.J. Pomonis, A.M. Efstathiou,
J. Catal. 209 (2002) 456.
[6] M. Uenishi, M. Tanigushi, H. Tanaka, M. Kimura, Y. Nishihata, J. Mizuki,
T. Kobayashi, Appl. Catal. B 57 (2005) 267–273.
[7] F. Kapteijn, J. Rodrigues-Mirasol, J.A. Moulijn, Appl. Catal. B 9 (1996)
25.
[8] G. Centi, L. Dall’Olio, S. Perathoner, J. Catal. 192 (2000) 224.
[9] G. Centi, L. Dall’Olio, S. Perathoner, J. Catal. 194 (2000) 130.
[10] V. Boissel, S. Tahir, C.A. Koh, Appl. Catal. B 64 (2006) 234.
[11] A. Bueno-López, I. Such-Basánˇez, C. Salinas-Martínez de Lecea, J. Catal.
144 (2006) 102.
[12] J. Soria, A. Martinez-Arias, J.M. Coranado, J.C. Conesa, Appl. Surf. Sci.
70–71 (1993) 245.

J.P. Dacquin et al. / Journal of Catalysis 253 (2008) 37–49
49
[13] H. Taguchi, S.I. Matsu-ura, M. Nagao, T. Choso, K. Kabata, J. Solid State
Chem. 129 (1997) 60.
[20] S. Kaliaguine, A. Van Neste, V. Szabo, J.E. Gallot, M. Bassir, R. Muzy-
chuck, Appl. Catal. A 209 (2001) 345.
[14] M. Engelmann-Pirez, P. Granger, L. Leclercq, G. Leclercq, Top. Catal.
30/31 (2004) 59.
[21] M.M. Natile, E. Ugel, C. Maccato, A. Glisenti, Appl. Catal. B 72 (2007)
351.
[15] R. Di Monte, P. Fornasiero, J. Kãspar, A. Ferrero, G. Gubitosa, M. Gra-
ziani, Stud. Surf. Sci. Catal. 116 (1998) 559.
[16] J.M. Gatica, R.T. Baker, P. Fornasiero, S. Bernal, J. Kaspar, J. Phys. Chem.
B 105 (2001) 1191.
[17] C.N. Costa, A.M. Efstathiou, J. Phys. Chem. B 108 (2004) 2620.
[18] D. Briggs, M.P. Seah, in: Practical Surface Analysis, vol. 1, second ed.,
Wiley, New York, 1983, p. 613.
[19] K.S. Kim, A.F. Gossmann, N. Winograd, Anal. Chem. 46 (2) (1974)
197.
[22] L.G. Tejuca, J.L.G. Fierro, J.M.D. Tascon, Adv. Catal. 36 (1989) 237.
[23] J.A. Marcos, R.H. Buitrago, E.A. Lombardo, J. Catal. 105 (1987) 95.
[24] I. Twagirashema, M. Engelmann-Pirez, M. Frere, L. Burylo, L. Gengem-
bre, C. Dujardin, P. Granger, Catal. Today 119 (2007) 100.
[25] J.M. Giraudon, A. Elhachimi, F. Wyrwalski, S. Sieffert, A. Aboukais, J.F.
Lamonier, G. Leclercq, Appl. Catal. B 75 (2007) 157.
[26] C. Serre, F. Garin, G. Belot, G. Maire, J. Catal. 141 (1993) 1.
[27] C. Serre, F. Garin, G. Belot, G. Maire, J. Catal. 141 (1993) 9.
[28] H.C. Yao, Appl. Surf. Sci. 19 (1984) 398.