Effect of ZrO2 content on textural and structural properties of CeO2-ZrO2 solid solutions made by citrate complexation route

Article abstract of DOI:10.1016/S0020-1693(03)00034-3

Single phase homogeneous CeO₂-ZrO₂ solid solutions with various compositions were synthesized using a citrate complexation route. Investigation of the sintering behaviour disclosed important modifications of the textural and, in particular, structural properties, which apparently create a strongly defective structure that could explain the unusual redox properties of these catalytic systems.

Full text of DOI:10.1016/S0020-1693(03)00034-3

Inorganica Chimica Acta 349 (2003) 217ꢀ226
/
www.elsevier.com/locate/ica
Effect of ZrO content on textural and structural properties of CeO ꢀ
/
2
2
ZrO solid solutions made by citrate complexation route
2
a,
J. Kaspar *, P. Fornasiero , G. Balducci , R. Di Monte , N. Hickey , V. Sergo
a
a
a
a
b
a
Dipartimento di Scienze Chimiche, Universit a` di Trieste, via L. Giorgieri 1, I-34127 Trieste, Italy
b
Dipartimento di Ingegneria dei Materiali e Chimica Applicata, Universit a` di Trieste, Via Valerio 2, I-34127 Trieste, Italy
Received 15 July 2002; accepted 19 September 2002
Dedicated to Professor Mauro Graziani on his 65th birthday.
Abstract
Single phase homogeneous CeO
2 2
ꢀZrO solid solutions with various compositions were synthesized using a citrate complexation
/
route. Investigation of the sintering behaviour disclosed important modifications of the textural and, in particular, structural
properties, which apparently create a strongly defective structure that could explain the unusual redox properties of these catalytic
systems.
#
2003 Elsevier Science B.V. All rights reserved.
Keywords: Exhaust catalysts; CeO ZrO solid solution; CeO ꢀ/ZrO ; Synthesis of mixed oxides
2
ꢀ/
2
2 2
1
. Introduction
Synthesis of new catalytic nanophase materials is
Ce Zr
x
O mixed oxides have a number of impor-
1ꢁx 2
tant catalytic applications: they are extensively em-
ployed in current automotive three-way catalysts
(TWC) as the oxygen storage promoters, i.e. materials
that regulate the oxygen partial pressure through the
attracting strong interest since it has been shown that by
tuning the structural and textural properties of an active
component on a nanoscale level, quantum size effects
may become important, leading to new and unpredicted
chemical and catalytic properties [1]. This is efficiently
illustrated by the case of oxide-supported gold systems.
This metal was traditionally considered catalytically
inactive, given its inability to chemisorb small molecules
3
ꢂ
4ꢂ
Ce /Ce redox couple, thereby keeping the oxidants-
to-reductants ratio in the exhaust at the stoichiometric
value where the highest pollutant conversions are
attained [5]. They exhibit high thermal stability com-
pared to the traditionally employed CeO , as well as an
2
exceptional ability to promote the catalytic activity of
supported noble metals (NM) with respect to conven-
tional supports, such as Al O or SiO , and even with
such as CO, H , etc. [2], until Haruta et al. [3] identified
2
an unexpectedly high activity of a Au/Fe O catalyst in
3
2
3
2
2
respect to CeO itself [6]. Similar advantages of the use
2
CO oxidation at very low temperature (203 K). The
origin of this high activity has been attributed to various
factors (preparation methods, nature of the support and
pre-treatment, morphology of the gold particles), how-
ever, the crucial point is that all of these factors lead to a
nanometer sized dispersion of the supported gold
particles, the maximum activity being observed for
of Ce Zr
x
O as NM support have been identified in a
2
1ꢁx
wide range of applications in the fields of chemicals,
energy production (reformer/WGSR in on-board H2
production for fuel cell, catalytic combustion), catalysts
for abatement of pollutants, etc. [7,8]. Despite their
apparent simplicity, at least in terms of chemical
composition, these materials show rather intriguing
properties. In particular, their redox behaviour, on
which their main function in TWCs is based, has been
shown to depend on several factors, including structure
[9], texture [10,11], phase homogeneity [12] and even the
particles with a diameter of 2ꢀ3 nm [3,4].
/
*
E-mail address: kaspar@units.it (J. Kaspar).
Corresponding author. Fax: ꢂ
/
39-040-558 3903.
0020-1693/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0020-1693(03)00034-3

218
J. Kaspar et al. / Inorganica Chimica Acta 349 (2003) 217ꢀ226
/
pre-treatment [13ꢀ
/
16]. A variety of methodologies have
been applied to the synthesis of Ce Zr1ꢁxO2 [17],
rotary evaporator. First, a transparent gel was obtained
and then, after the evolution of nitrogen oxides, a solid
was recovered. This was heated in air from r.t. to 773 K
at a rate of 3 K min and calcined at this temperature
for 5 h. Henceforth, these samples are indicated as fresh.
Effects of calcination (5 h) were investigated in the
x
however, despite the large amount of work, a number
of fundamental questions related to these systems, such
ꢁ
1
as which is the optimal Ce Zr
x
O composition in
2
1ꢁx
terms of redox and textural stability, still remain open.
A possible interpretation of such variability in their
chemical behaviour is that, due to limited solubility of
ZrO in the CeO lattice, metastable phases and domain-
temperature range from 873 to 1273 K. N adsorption
2
isotherms at 77 K were measured on a Micromeritics
ASAP 2000 analyser.
2
2
type structures may be formed at a nanoscale level in the
synthesis, which cannot be easily detected by conven-
tional characterisation methods [18]. Failure to detect
such inhomogeneities would effectively result in com-
parison of different materials with nominally equal
composition. Co-precipitation and gel-like synthesis
methods indeed often lead to compositionally inhomo-
geneous mixed oxides [19,20]. Accordingly, contradic-
tory results have been reported in the literature on what
Raman spectra were obtained on either a Perkin-
Elmer 2000 FT-Raman spectrometer with a diode
pumped YAG laser (1064 nm) and a r.t. super InGaAs
detector, using a laser power of 200 or 600 mW; or a
ꢂ
Renishaw S1000 system with an Ar ion laser (514.5
nm) operating in the back scattering geometry. In this
latter case the power delivered to the sample was about 5
mW.
SEM micrographs were taken with a conventional
scanning electron microscope (Cambridge Inc.). Before
taking the images the powders were dispersed in water, a
drop of the suspension was deposited onto the specimen
holder and, after water evaporation, the powder was
coated with gold to prevent charging under the electron
beam.
should be the best Ce Zr
x
O composition in terms of
2
1ꢁx
thermal stability and redox properties: either CeO - or
2
ZrO -rich, or even Ce0.5Zr0.5O , have all been indicated
2
2
as the most effective compositions [10,21,22]. Further-
more, remarkably different textural and chemical beha-
viour can be achieved by tuning the synthesis method at
a single sample composition (Ce0.2Zr0.8O ) [23]. Evi-
2
Powder X-ray diffraction patterns were collected on a
Siemens Kristalloflex Mod.F diffractometer using Ni-
filtered Cu Ka. The particle sizes of the powdered
samples were calculated from the diffraction patterns
using the Schrerrer formula [26].
dently, a single synthesis method, reproducibly leading
to homogeneous solid solutions, is needed to truly assess
the effects of the Ce Zr1ꢁxO2 composition. In this
x
work, a citrate complexation synthesis method is applied
to investigate the effects of Ce Zr
O composition on
2
x
1ꢁx
structural and textural properties. The aim is that of
using a single synthesis methodology which allows
production of compositionally homogeneous systems
3. Results and discussion
across the whole Ce Zr
x
O range of compositions, as
2
The use of citric acid as complexing agent, employed
as described in Section 2, leads to formation of
apparently homogeneous gel-like materials at all the
compositions investigated. These materials were then
1ꢁx
it has been shown previously the applicability of such
method for selected compositions, even though single
phase systems were not always obtained [24]. It is shown
here that the citrate route using water as solvent can be
conveniently employed to obtain homogeneous single
phase products over a wide compositional range.
dried and calcined to obtain white (ZrO ) to yellow
2
(CeO ) powders, which were characterized by different
2
techniques to assess the following aspects: (i) homo-
geneity of the mixed oxide, (ii) effects of increasing
calcination temperature on structural and (iii) textural
stability.
2
. Experimental
Ce Zr O (xꢃ
/
0.2, 0.4, 0,5, 0.6, 0.8) solid solutions
3.1. XRD and Raman characterization
x
1ꢁx
2
and pure cerium and zirconium oxides were synthesised
from nitrate precursors by complexing the metal cations
with the citrate [25]. The preparation was carried out as
follows. After dissolving Ce(NO ) ×6H O (32.1% Ce,
Before discussing the results, let us recall some
features of the Ce Zr1ꢁxO2 phase diagram. CeO2
x
¯
3 /
crystallizes in the fluorite structure (space group Fm
/
3
/
3
2
Aldrich) and ZrO(NO ) ×
/
nH O (27.32% Zr, Aldrich) in
m), while ZrO is monoclinic at r.t. (space group P2 /c).
2 1
3
2
2
H O, a solution of citric acid (99.7%, Prolabo) in H O
2 2
Given the relatively large difference (13%) between the
4ꢂ
4
cation radii of Ce (0.097 nm) and Zr
ꢂ
was added, leading to a final concentration of mixed
ꢁ
(0.084 nm), a
limited mutual solubility might be expected [27]. Con-
sistently, only the tetragonal (t-) and cubic (c-)phases,
1
oxide of 40 g l . The resulting solution was stirred at
53 K for 2.5 h to obtain a complete dissolution of the
precursors and then at room temperature (r.t.) for at
3
found respectively at high (ꢀ
/
80 mol.%) and low (B20
/
least 12 h. The solvent was subsequently eliminated on a
mol.%) ZrO contents, are thermodynamically stable,
2

J. Kaspar et al. / Inorganica Chimica Acta 349 (2003) 217ꢀ
/226
219
while two metastable tetragonal phases (t?, tƒ) may be
formed at intermediate compositions [28]. The presence
of metastable phases poses serious difficulties to re-
The evolution of the powder XRD patterns with
increasing temperature of calcination was therefore
investigated. Typical results are exemplified in Fig.
searchers*
/
sometimes underestimated*/regarding as-
1(a) for Ce0.6Zr0.4O . The fresh sample features broad,
2
sessment of the homogeneity of the prepared solid
solution. Vegard’s rule is typically applied for this
purpose. Due to the lower cation radius of Zr(IV)
compared to Ce(IV), a linear decrease of the cell
parameter (or cell volume, when t-phases are included)
is expected and, in fact, observed [29,30]. However, the
as-synthesized high surface area samples typically fea-
ture quite broad XRD peaks, due to the nanometer
dimension of the crystallites. This does not allow an
unambiguous detection of a single phase product. After
symmetric XRD peaks which strongly sharpen as the
calcination temperature increases. In the range investi-
gated, no evidence indicated that phase segregation
could be induced by the increase of the calcination
temperature up to 1273 K. This is an important point
since in our previous investigation of such composition,
evidence for phase inhomogeneity, revealed by asym-
metric peak profiles, was detected [24]. We observed
that, as indicated in Section 2, a complete dissolution of
the precursors must be ensured in the synthesis to obtain
phase homogeneity. In fact, we have used water as
solvent in this work instead of methanol [24]. Similar
symmetrical XRD patterns are observed after calcina-
tion at 1273 K in the cases of the other compositions
calcination 1073ꢀ
/
1273 K for 2ꢀ5 h, significant sharpen-
/
ing of the XRD patterns is normally observed, which is
due to the growth and sintering of the crystallites.
However, such sintering often leads to the appearance of
strongly asymmetric peaks in the XRD spectra, which
has been attributed to metastable nature of the
(
(
Fig. 1(b)), except for the Ce0.4Zr0.6O and ZrO samples
2 2
the latter sample is not reported). In the former case,
phase segregation into a CeO -rich and ZrO -rich
2
Ce Zr
x
O of intermediate compositions, leading to
2
2
1ꢁx
mixture was detected for calcination temperatures at
and above 973 K, whilst in the latter case, transforma-
tion of t-ZrO into m-ZrO with increasing calcination
phase segregation upon calcination at high temperature
31,32]. In principle, the metastable nature of these
[
2
2
mixed oxides makes assessment of structural homoge-
neity difficult.
temperatures was detected.
Except in the above quoted cases for high calcination
temperatures, and to some extent for the Ce0.2Zr0.8O2
sample (see below), Vegard’s rule is nicely followed by
all the compositions here investigated (Fig. 2). This is an
important finding that indicates the versatility of the
present synthesis method for obtaining single phase
solid solutions (to the detection limit of the XRD
technique) for a wide interval of Ce Zr
x
O composi-
2
1ꢁx
tions. For comparison, by the usual co-precipitation
techniques, single phase solid solutions could not be
obtained over a large compositional range, even for
commercial products [33,34]. The data shown in Fig. 2
Fig. 1. XRD patterns of (a) Ce0.6Zr0.4
2) 873 K, (3) 973 K, and (4) 1273 K and (b) (1) Ce0.2Zr0.8
Ce0.4Zr0.6 , (3) Ce0.5Zr0.5 , (4) Ce0.8Zr0.2 , all samples were
calcined at 1273 K for 5 h.
O
2
after calcination at (1) 773 K,
Fig. 2. Vegard rule for the Ce
x
Zr1ꢁx
O
2
(xꢃ
/
0.2, 0.4, 0.6, 0.8) solid
content
(
O
2
, (2)
solutions and CeO detected as variation of d(1 1 1) with CeO
2
2
O
2
O
2
O
2
for samples calcined at the indicated temperatures. Pseudocubic cell is
considered for the tetragonal samples.

220
J. Kaspar et al. / Inorganica Chimica Acta 349 (2003) 217ꢀ226
/
suggest that the phase segregation observed at high
temperatures for Ce0.4Zr0.6O2 should preferably be
attributed to a high tendency of this composition to
segregate, due to existence of a bi-phasic (t and t?)
region in the phase diagram around this composition
even smaller then those here observed [36], suggesting
that some other subtle changes of the structural proper-
ties may also occur in the mixed oxides with increasing
calcination temperature (see below).
Phase attribution and sample modification induced by
the calcination were therefore investigated by Raman
[
35], rather than to the formation of compositional in-
ꢁ
1
homogeneities in the synthesis [12]. The deviation from
the linearity observed for d(1 1 1) in the case of
Ce0.2Zr0.8O2 is attributed to the fact that there is a
significant tetragonal distortion of the pseudo-cubic unit
spectroscopy in the range 100ꢀ
absorption due to the MÃO stretching is observed (Fig.
3). Six Raman-active modes of A ꢂ3E ꢂ2B1g sym-
/
1000 cm , where the
/
/
/
1g
g
metry are observed for tetragonal ZrO (space group
P4 /nmc), while for the cubic fluorite structure (space
2
cell with a ratio of c/aꢃ/1.017. Rietveld fitting per-
2
¯
formed on the sample calcined at 1273 K confirmed the
presence of a single phase product.
group Fm/3/m) only one F2g mode centred at around 490
ꢁ
1
¯
is Raman-active [40]. In pure CeO (Fm3m space
cm
group), the F2g mode is observed at 465 cm
/
/
2
ꢁ
1
A perusal of Fig. 1(a) reveals, however, that there is a
small but continuous shift of the XRD peaks to higher
values of 2u with increasing calcination temperature.
Similar behaviour was also observed for the other
compositions, except ZrO and Ce0.4Zr0.6O (vide su-
.
The Raman spectra reported in Fig. 3 show the
following features: (i) for CeO contents of 100 and 80
2
mol.%, the spectrum is dominated by a single strong
band which shifts to higher frequencies and becomes less
sharp at lower CeO content. This band is attributed to
2
2
pra). Rietveld analysis of the XRD patterns reported in
Fig. 1(a) revealed a contraction of the cell volume from
2
ꢁ
1
the F2g mode. A minor peak at about 145 cm
present in the spectrum of Ce0.8Zr0.2O ; (ii) at high ZrO
is also
3
.15119 to 0.14919 nm as the calcination temperatures
0
2
2
was increased from 773 to 1273 K, corresponding to a
relative decrease of 1.3%. Phase segregation in the
Ce Zr1ꢁxO2 systems leads to CeO - and ZrO -rich
content (E80 mol.%) five of the six bands of the
/
tetragonal ZrO are detected, while the presence of the
monoclinic phase in clearly detected for the pure ZrO2.
2
x
2
2
phases; accordingly, if we assume that segregation of
pure CeO occurred, then segregation of about 5 wt.%
Finally, (iii), at intermediate CeO2 contents (40ꢀ
/
60
mol.%), the spectrum is dominated by a broad and
2
ꢁ
1
of CeO2 can be calculated for the above quoted
variation of unit cell volume, which should have been
detected under our experimental conditions. A further
consideration is that an interpretation of this phenom-
enon in terms of phase segregation cannot be applied to
intense band at 530ꢀ
/540 cm
with shoulder at about
ꢁ1
ꢁ
1
470 cm . A weak band at about 125 cm
present.
The shift of the F2g Raman frequency in Ce0.8Zr0.2O2
is also
to higher values with respect to CeO is due to the
2
CeO , as in fact a perusal of Fig. 2 indeed reveals some
2
decrease of the lattice parameter, resulting in shorter
O bond lengths. The presence of the very weak
feature at 145 cm may suggest the presence of the tƒ-
change of the lattice parameter also in this case. A
possible interpretation is to relate these shifts to the
nanometer particle size of the present catalysts: by
decreasing the size of the particle, the ratio of surface
to bulk atoms, which favours conversion of Ce(IV) to
Ce(III) near the surface region [36], increases. This in
turn leads to lattice expansion due to the higher ionic
MÃ
/
ꢁ
1
phase, i.e. a tetragonal (space group P4 /nmc) phase
with some distortion of the oxygen sublattice with
respect to the cubic fluorite structure and with a ratio
2
of unit cell parameters c/aꢃ1 [28]. For high ZrO2
/
content, the spectral features are consistent with an
essentially regular tetragonal P4 /nmc symmetry, with
3
ꢂ
4ꢂ
radius of Ce
compared to Ce
[37]. In our case
obviously the opposite will occur. The reasons for such
2
respectively short and long MÃ
/
O and MÃ
/
OII bond
I
changes may be understood by considering that the
distances. Among the six characteristic Raman modes of
ꢁ1
4
ꢂ
3ꢂ
standard potential for reduction of Ce to Ce is 1.74
V in solution [38], which indicates that Ce(IV) in
solution is a strong oxidant. In the solid state, the
t-ZrO , the band at approximately 260 cm
2
corre-
O and those at 610 and
640 cm O [41]. Accordingly, when the
amount of CeO2 inserted into the ZrO2 lattice is
sponds to the stretching of MÃ
/
II
ꢁ
1
to that of MÃ
/
I
situation is different. CeO crystallises in the fluorite
2
structure in which each cerium ion is co-ordinated by
eight oxygen neighbours. This co-ordination stabilises
increased from 12 to 24 mol.%, the bands at 610 and
ꢁ
1
640 cm , related to the stretching modes, shift and
4
ꢂ
the Ce
state and makes the reduction of CeO2
tend to overlap [42]. This accounts for the observation
of five bands in the Ce0.2Zr0.8O instead of the predicted
unfavourable. In fact, the fluorite structure of ceria is
a direct result of the ionic nature of ceria and of the
charge and size of the ions. Model calculation has
shown that it is formed when a sufficiently high number
2
six modes.
The observation of a few Raman modes with the
ꢁ
1
strong features at 520ꢀ
/
540 cm
for the intermediate
CeO contents cannot be easily rationalised on the basis
of CeO units (Â50) are clustered together [39].
/
2
2
¯
On the other hand, expansion of the CeO lattice
parameter was observed for particle dimensions that are
of the proposed phase symmetries (P4 /nmc and Fm
2
/
3
/
m)
2
and the possible shifts of the stretching modes. More-

J. Kaspar et al. / Inorganica Chimica Acta 349 (2003) 217ꢀ
/226
221
Fig. 3. Raman spectra of samples calcined at 773 K (a), 973 K (b) and 1173 K (c): ZrO
(5); CeO (6) (laser power 200 mW).
2 2 2 2
(1); Ce0.2Zr0.8O (2); Ce0.4Zr0.6O (3); Ce0.6Zr0.4O (4);
Ce0.8Zr0.2
O
2
2
over, the appearance of the spectra clearly changes with
the temperature of calcination as detailed for
Ce0.6Zr0.4O2 in Fig. 4, suggesting the occurrence of
some important changes in the structural properties of
the sample. Such behaviour would not be expected from
simple sintering. This aspect was further examined in
some detail using different wavelength of the Raman
excitation radiation (Fig. 4). A remarkable dependency
of the Raman spectra on the excitation energy is clearly
denoted by the presence of strong bands at 3360, 2490
ꢁ
1
and 1920 cm that are observed only when IR
excitation is used. The absence of the band at 525
ꢁ
1
cm
used is another noteworthy feature. The attribution of
in the spectrum when the green light source is
ꢁ
1
the band at 3360 cm to surface OH groups is highly
reasonable, substantiated also by the broad pattern of

222
J. Kaspar et al. / Inorganica Chimica Acta 349 (2003) 217ꢀ226
/
detected in mixed fluorite oxides, they are relatively
small and linearly related to the square root of the cell
volume [41]. Thus, for the E mode that was found at
g
ꢁ
1
31 cm in Ce0.12Zr0.88O [43], we can estimate a value
6
of 595 cm
2
ꢁ
1
for Ce0.4Zr0.6O by applying the linear
2
relationship found by Kim et al. for Y O (2 mol.%)ꢀ
/
2
3
CeO2ꢀ
/
ZrO mixed oxides [41]. This is somewhat distant
1
2
ꢁ
from the value of the peak maximum at 550 cm
observed in Fig. 3. On the other hand, it should be noted
that the above relationship has been found to hold for
the t-phase, whilst t? or tƒ phase are detected at CeO
2
content of 40ꢀ60 mol.%. It might be argued that the
/
ꢁ
1
increase of the intensity of the band around 550 cm
could be related to a progressive tetragonalization of the
lattice induced by the increase of the calcination
temperature. This interpretation, however, is not sup-
ported by the Raman spectra obtained using the visible
excitation line. In this case, the strong band around 550
ꢁ
1
cm is not observed, the only apparent modification of
the spectra with increasing calcination temperature
being a sharpening of the peak with maximum at 466
ꢁ
1
ꢁ1
cm
that shifts to 471 cm
(as the calcination
temperature is increased from 973 to 1273 K). Note
that, due to the nanometer dimension of the starting
materials, this sharpening of the peak is fully consistent
with the occurrence of particle sintering [44]. The shift to
ꢁ
1
higher frequencies of the band at 466 cm , attributed
to the F2g mode of the fluorite lattice, is consistent with
the decrease of the lattice parameter observed by XRD.
Generally speaking, decreasing the lattice parameter and
hence cell volume, the lattice moves towards the
structural features of the ZrO2 systems and hence
tetragonalization is expected, but it cannot explain the
ꢁ
1
features due to the band at 530 cm
.
ꢁ
1
Another possibility is the band at 530 cm could be
attributed to luminescence, which accounts for its
absence using the visible excitation [45]. However, even
in this case, the variation of the relative intensity of the
ꢁ
1
band at 530 and the shoulder at 465ꢀ478 cm appears
/
to be a clear indication of a subtle modification of the
structure of the mixed oxide with increasing calcination
temperature. Similar peak attribution could be also
ꢁ
1
made for the bands at 1915 and 2500 cm . None of
these bands seems attributable to the presence of an
impurity as their intensities relative to the peak/shoulder
Fig. 4. Effect of calcination temperature (a) and Raman excitation line
b) on the spectra of Ce0.6Zr0.4 : (a) sample calcined at 773 K (1), 973
ꢁ
1
at about 470 cm
ture of calcination in a similar fashion as the band
are neatly affected by the tempera-
(
O
2
K (2), 1173 K (3) and 1273 K (4); (b) sample calcined at 1273 K, IR
excitation at 1068 nm (laser power 600 mW) (1) and green light
excitation at 514.5 nm (2).
ꢁ
1
around 530 cm . No evidence for significant amounts
of impurities was indeed found by X-ray fluorescence
analysis of the sample.
the band. Conversely, the attribution of the bands at
1
ꢁ1
It should be recalled that a band around 2100 cm
2
ꢁ
2
has been attributed to the F7/2
2
490, 1920 and 525 cm
is more uncertain. Generally
O lattice
l
/
F5/2 electronic
3
ꢂ
speaking, the phonon spectra related to the MÃ
/
transition of Ce
Ce Zr O , and its intensity was related to the
species in a reduced Y O doped
2
3
ꢁ
1
are observed below 1000 cm . Even though shifts in
the phonon spectra due to introduction of foreign cation
and, consequently, change of bond lengths can be
x
1ꢁx
2
3
ꢂ
concentration of these species [46]. Formation of Ce
on calcination at high temperature and in the presence

J. Kaspar et al. / Inorganica Chimica Acta 349 (2003) 217ꢀ
/226
223
of oxygen might seem somewhat surprising but the
presence of defects in Ce Zr1ꢁxO2 has long been
between the pure CeO and the ZrO -doped systems in
2 2
that sintering leads to a highly ordered structure in the
former case while disordered or defective lattice is
x
recognized [47], even though their exact nature is still
matter of debate [9,48]. Recently, Egami and coworkers
suggested that the insertion of ZrO in the CeO lattice
created by the sintering of the Ce Zr
x
O , which can
2
1ꢁx
be attributed to the misfit between the Ce(IV) and
Zr(IV) radii responsible for this lattice strain. This is
observed as long as tƒ- and t?-, i.e. closely related to
2
2
favours generation of Frenkel-type defects, the displace-
ment of lattice oxygen to the octahedral sites possibly
3
ꢂ
generating Ce
have been indicated as the origin of the mobility of the
oxygen in the bulk of the Ce Zr O compared to
sites in the lattice [49]. These defects
CeO , phases are considered, while for the t-
2
Ce0.2Zr0.8O , whose structure is closely related to t-
2
ZrO we do not have evidence for such behaviour.
2
x
1ꢁx
2
CeO , leading to enhanced redox properties. Notice that
2
a displacement of the oxygen atoms from the tetrahedral
sites of the fluorite lattice into the octahedral ones could
possibly lead to decrease of lattice parameter due to the
larger size of the octahedral cavity.
3
.2. Textural and morphological characterization
Textural and morphological properties were investi-
gated by N physisorption, XRD and SEM techniques.
The BET surface areas as a function of temperatures are
reported in Fig. 5. Interestingly, at comparable synthesis
conditions, there is a strong effect of the sample
composition on the initial surface area, which is max-
An alternative view of such processes involving the
3
2
ꢂ
generation of the Ce species is to consider the removal
of one electron from a doubly ionized oxygen vacancy to
4
ꢂ
3ꢂ
reduce Ce
to Ce
(defects in CeO2ꢁx are mainly
doubly ionized vacancies, but the presence of singly
ionized oxygen vacancies was recently detected in
dynamically compacted m-ZrO2 [50]); the standard
imum at a CeO content of 80 mol.%. The effect is quite
2
remarkable as this sample features a BET surface area
which is roughly three times that of either ceria or
enthalpy of this process in CeO is 0.8ꢀ
/
1 eV [51,52].
This energy corresponds to shift of about 2900ꢀ1300
relative to the IR and
2
2
ꢁ1
zirconia (64 vs. 18 and 20 m g ), indicating the
favourable effect of the formation of the solid solution
upon the textural properties. However, upon increasing
the calcination temperature, due to the relatively poor
textural stability that can be achieved by this synthesis
methodology [23], the effects of the ZrO2 addition
become quite small in terms of the surface area
measurements. Notwithstanding the relatively low and
/
ꢁ
1
ꢁ1
cm
and 13 000ꢀ11 400 cm
/
green excitation frequency, respectively. It is then quite
reasonable to assume that the IR excitation frequency is
much closer to a Raman resonant condition, with
ensuing enhanced Raman efficiency for the bands which
ꢁ
1
we observe in the range 1900ꢀ/2500 cm
(Fig. 4).
Usually excitation wavelengths far from Raman reso-
nant conditions do not lead to a complete suppression of
the bands: some intensity can still be recorded. In our
case we do not observe the above quoted band at about
comparable BET areas, the retarding effects of the ZrO2
addition on the rate of sintering are neatly detected by
the XRD technique. The crystallite sizes were calculated
from peak broadening using the Scherrer formula [26].
The data are reported in Fig. 6 and show that insertion
ꢁ
1
2
100 cm , but this can be due to very low concentra-
tion of the Raman resonant system or to low excitation
of ZrO into the CeO lattice remarkably modifies the
2
power; indeed Otake et al. [46] did observe these bands
2
in the ZrO2ꢀ
/
CeO2ꢀY O system, but they examine
/
2 3
3
ꢂ
samples with much larger Ce
concentrations due to
10ꢁ19 Pa) using a laser
power twice that of our case. Despite this, the intensity
ꢁ
13
/
extremely low pO2 (10
ꢀ
ꢁ
1
of the band at about 2100 cm recorded by those
authors is about 30 times weaker than in our case (as
can be inferred from the relative intensity of this band
ꢁ
1
with respect to the band at 470 cm in our Fig. 4, and
in their [46, Figs. 2 and 3]). In summary, we believe that
ꢁ
1
the intense band at 1910 cm
is a resonant effect; as
such it is coupled with an electronic transition close in
energy to the peculiar IR excitation wavelength which
gives rise to it. Even though the observed bands cannot
be unequivocally assigned, a clear indication has been
obtained, that is that the lattice structure and, more
probably, the defect structure of the solid solution is
strongly modified with increasing calcination tempera-
Fig. 5. Surface area measured by the BET method as a function of the
calcination temperature (calcination temperatures and CeO2 content
are indicated).
ture, particularly at the intermediate Ce Zr1ꢁxO
x
2
compositions. There is therefore a striking difference

224
J. Kaspar et al. / Inorganica Chimica Acta 349 (2003) 217ꢀ226
/
different sintering behaviour of the samples should be
mainly related to the effects of composition. As far as
the porosity is concerned, the micrographs at higher
magnification reveals the presence of porosity in the
submicrometer range in CeO and Ce0.4Zr0.6O that is
2
2
absent in pure ZrO2.
4. Conclusions
The use of citrate complexation synthesis method
allowed production of single phase Ce Zr O over a
x
1ꢁx
2
wide interval of compositions that feature consistent
textural properties. This allowed the assessment of
effects of composition on the textural and structural
properties of these important catalytic materials. These
findings can be summarised as follows: (i) sintering of
Fig. 6. Effect of ZrO
at the indicated temperatures, as detected from the Schrerrer formula
average particle size is reported for ZrO and Ce0.4Zr0.6 when phase
separation was detected).
2
addition on the crystallite size after calcination
the Ce Zr
x
O particles strongly depends on the oxide
2
1ꢁx
(
2
O
2
composition, the lowest sintering rate being observed for
intermediate composition; (ii) the sintering process
significantly modifies the structure of solid solutions
sintering process of CeO . High ZrO content signifi-
2
2
by creating a defective solid as more ZrO is inserted
2
cantly decreases the rate of sintering as long as the cubic
or pseudo-cubic (tƒ) phase symmetry is maintained in
the mixed oxide, leading to a pronounced minimum at
intermediate composition, as shown by the data ob-
tained after calcination at 1273 K. At this composition,
reversible transition from the tƒ to the t? phase was
observed [28], leading to a highly defective system. The
presence of this particular structure seems to be
responsible for the unusually low sintering rate of this
sample. It should be noted that, due to the high
homogeneity of the initial precursor, no phase separa-
tion occurred under our calcination as checked by
Rietveld analysis of the sample calcined at 1273 K.
XRD line broadening, leading to apparent low particles
size, is indeed observed when such segregation occurs
into the CeO lattice, provided that enough CeO is
2
2
present to avoid formation of the t-phase. Tetragona-
lisation of the lattice at high ZrO contents can in fact
2
release the lattice strain generated by the insertion of the
small Zr(IV) cation in the lattice, leading to a less
defective solid [9].
Defective structure has been shown to play a funda-
mental role in the catalytic properties of the CeO -based
2
materials [49], accordingly a nano-sized defective CeO2
oxide was significantly more active in redox catalytic
reactions compared to conventional CeO2 powders
featuring larger particles [55]. On the contrary, as here
observed, when the solid solution is sintered, due to the
insertion of ZrO into the CeO lattice the structural and
2
2
textural properties are remarkably modified with respect
to CeO , since generation of defective structures is
[
53]. Even though other factors, such as strain effects
2
induced by the defective structure of local compositional
inhomogeneities could in principle contribute to XRD
line broadening, we believe that such factors should not
significantly affect the present measurements. No sig-
nificant increase of dispersion of bond lengths due to
local compositional inhomogeneities caused by insertion
of ZrO into the CeO lattice could be observed by
induced. This may explain some of the intriguing and
unique properties of these catalysts such as the promo-
tion of the reducibility at low temperatures, which can
be induced by sintering at high temperatures [13,54,56].
Finally, the effects of the Ce Zr
x
O composition
2
1ꢁx
on the sintering behavior suggest some possible guide-
lines for the choice of the appropriate composition to be
employed in the catalytic application of these materials:
due to the fast phase separation observed in the
2
2
EXAFS, when homogeneous systems (as above defined)
were investigated [54].
The morphology of the powders is illustrated in Fig. 7
with some representative SEM micrographs. All micro-
graphs were recorded on samples calcined at 500 8C. At
low magnification it appears that all three powders have
the same grain size dimension and morphology. This is
an indication that the use of citrate synthesis methodol-
ogy leads to comparable textural properties, indepen-
dently of the sample composition. Accordingly, the
Ce0.4Zr0.6O sample and the increase of BET area with
2
CeO content, cerium-rich compositions appear to be
2
good candidates for low to medium temperature cata-
lytic applications (B1273 K). On the other hand when
/
high temperatures are considered, ZrO -rich tetragonal
2
phase could represent a good compromise due to its
intrinsic thermodynamic stability coupled to a relatively
small sintering rate, which allows a dispersed state of the

J. Kaspar et al. / Inorganica Chimica Acta 349 (2003) 217ꢀ
/226
225
2 2 2
Fig. 7. SEM micrographs of (a) CeO , (b) Ce0.4Zr0.6O and (c) ZrO .
[
[
[
7] S. Bernal, J. Kaspar, A. Trovarelli (Eds.), Catal. Today, 20(2)
1999) 173.
8] A. Trovarelli (Ed.), Catalysis by Ceria and Related Materials, vol.
, Imperial College Press, London, 2002.
catalyst to be maintained, even though the redox
properties could be limited by the low cerium content.
(
2
9] P. Fornasiero, R. Di Monte, G. Ranga Rao, J. Kaspar, S.
Acknowledgements
Meriani, A. Trovarelli, M. Graziani, J. Catal. 151 (1995) 168.
10] H. Vidal, J. Kaspar, M. Pijolat, G. Colon, S. Bernal, A.M.
[
[
Cordon, V. Perrichon, F. Fally, Appl. Catal. B Environ. 27 (2000)
The University of Trieste, Fondo Trieste 1999,
4
11] H. Vidal, J. Kaspar, M. Pijolat, G. Colon, S. Bernal, A. Cordon,
9.
MURST*
the environmental impact of mobile source emissions’,
CNR Agenzia 2000, and project ‘Urban Atmosphere’*
/
PRIN 2000 ‘Catalysis for the reduction of
V. Perrichon, F. Fally, Appl. Catal. B Environ. 30 (2001) 75.
/
[12] P. Fornasiero, R. Di Monte, J. Kaspar, T. Montini, M. Graziani,
Stud. Surf. Sci. Catal. 130 (2000) 1355.
Consortium INCA are gratefully acknowledged for
financial support.
[
13] G. Balducci, P. Fornasiero, R. Di Monte, J. Kaspar, S. Meriani,
M. Graziani, Catal. Lett. 33 (1995) 193.
[14] N. Izu, T. Omata, S. Otsuka-Yao-Matsuo, J. Alloys Comp. 270
(
1998) 107.
References
[15] R.T. Baker, S. Bernal, G. Blanco, A.M. Cordon, J.M. Pintado,
J.M. Rodriguez-Izquierdo, F. Fally, V. Perrichon, Chem. Com-
mun. (1999) 149.
[1] J.M. Thomas, Angew. Chem., Int. Ed. 38 (1999) 3589.
[2] G.C. Bond, D.T. Thompson, Catal. Rev. Sci. Eng. 41 (1999) 319.
[3] M. Haruta, N. Yamada, T. Kobayashi, S. Iijima, J. Catal. 115
[16] T. Masui, T. Ozaki, K. Machida, G. Adachi, J. Alloys Comp. 292
(1999) L8.
(
1989) 301.
[17] G.Y. Adachi, T. Masui, Synthesis and modification of ceria-based
materials, in: A. Trovarelli (Ed.), Catalysis by Ceria and Related
Materials, Chapter 3, Imperial College Press, London, 2002, pp.
[
[
4] M. Haruta, M. Date, Appl. Catal. A Gen. 222 (2001) 427.
5] J. Kaspar, P. Fornasiero, M. Graziani, Catal. Today 50 (1999)
2
85.
51ꢀ
/
83.
[18] T. Egami, W. Dmowski, and R. Brezny, SAE, 970461 (1997).
[
6] A. Trovarelli, Catal. Rev. -Sci. Eng. 38 (1996) 439.

226
J. Kaspar et al. / Inorganica Chimica Acta 349 (2003) 217ꢀ226
/
[
[
[
[
19] S. Rossignol, F. Gerard, D. Duprez, J. Mater. Chem. 9 (1999)
615.
[38] T. Moeller, in: J.C. Bailar, H.J. Emeleus, R. Nyholm, A.F.
Trotmandickenson (Eds.), Comprehensive Inorganic Chemistry,
Chapter 44, Pergamon Press, Oxford, 1982, p. 1.
[39] H. Cordatos, D. Ford, R.J. Gorte, J. Phys. Chem. 100 (1996)
18128.
1
20] P. Fornasiero, G. Balducci, R. Di Monte, J. Kaspar, V. Sergo, G.
Gubitosa, A. Ferrero, M. Graziani, J. Catal. 164 (1996) 173.
21] A. Trovarelli, C. de Leitenburg, G. Dolcetti, Chemtech 27 (1997)
32.
[40] V.G. Keramidas, W.B. White, J. Am. Ceram. Soc. 57 (1974) 22.
[41] D.J. Kim, J.W. Jang, H.L. Lee, J. Am. Ceram. Soc. 80 (1997)
1453.
22] M. Fernandez Garcia, A. Martinez Arias, A. Iglesias Juez, C.
Belver, A.B. Hungria, J.C. Conesa, J. Soria, J. Catal. 194 (2000)
3
85.
[42] T. Hirata, E. Asari, M. Kitajima, J. Solid State Chem. 110 (1994)
201.
[
[
[
[
[
[
[
[
[
[
[
[
23] J. Kaspar, R. Di Monte, P. Fornasiero, M. Graziani, H.
Bradshaw, C. Norman, Top. Catal. 16 (2001) 83.
[43] T. Hirata, J. Phys. Chem. Solids 56 (1995) 951.
[44] G.W. Graham, W.H. Weber, C.R. Peters, R.K. Usmen, J. Catal.
130 (1991) 310.
24] P. Vidmar, P. Fornasiero, J. Kaspar, G. Gubitosa, M. Graziani, J.
Catal. 171 (1997) 160.
25] C. Marcilly, P. Courty, B. Delmon, J. Amer. Ceram. Soc. 53
[45] M. Yashima, K. Ohtake, M. Kakihana, H. Arashi, M. Yoshi-
mura, J. Phys. Chem. Solids 57 (1996) 17.
(1970) 56.
26] J.R. Anderson, K.C. Pratt, Introduction to Characterization and
Testing of Catalysts, Academic Press, London, 1985.
27] A.R. West, Basic Solid State Chemistry, 2, Wiley, Chichester,
[46] T. Otake, H. Yugami, H. Naito, K. Kawamura, T. Kawada, J.
Mizusaki, Solid State Ionics 135 (2000) 663.
[47] A.I. Leonov, A.B. Andreeva, E.K. Keler, Izv. Akad. Nauk SSSR,
Neorg. Mater. 2 (1966) 137.
1999.
28] M. Yashima, H. Arashi, M. Kakihana, M. Yoshimura, J. Am.
Ceram. Soc. 77 (1994) 1067.
[48] G. Vlaic, R. Di Monte, P. Fornasiero, E. Fonda, J. Kaspar, M.
Graziani, J. Catal. 182 (1999) 378.
29] C.K. Narula, L.P. Haack, W. Chun, H.W. Jen, G.W. Graham, J.
Phys. Chem. B 103 (1999) 3634.
30] K. Otsuka, Y. Wang, M. Nakamura, Appl. Catal. A Gen. 183
[49] E. Mamontov, T. Egami, R. Brezny, M. Koranne, S. Tyagi, J.
Phys. Chem. B 104 (2000) 11110.
[50] M. Maczka, E.T.G. Lutz, H.J. Verbeek, K. Oskam, A. Meijerink,
J. Hanuza, M. Stuivinga, J. Phys. Chem. Solids 60 (1999) 1909.
[51] R.N. Blumenthal, J.E. Laubach, in: I.W. Vahlidick, S.A. Mersol
(Eds.), Anisotropy in Single Crystal Refractory Compounds,
(1999) 317.
31] C.E. Hori, H. Permana, K.Y.S. Ng, A. Brenner, K. More, K.M.
Rahmoeller, D.N. Belton, Appl. Catal. B Environ. 16 (1998) 105.
32] C. Bozo, F. Gaillard, N. Guilhaume, Appl. Catal. A Gen. 220
Plenum, New York, 1968, pp. 137ꢀ150.
/
(2001) 69.
[52] G. Balducci, J. Kaspar, P. Fornasiero, M. Graziani, M.S. Islam,
J. Phys. Chem. B 102 (1998) 557.
33] J.P. Cuif, G. Blanchard, O. Touret, A. Seigneurin, M. Marczi, E.
Qu e´ mer e´ , SAE, 970463 (1997).
[53] G. Colon, F. Valdivieso, M. Pijolat, R.T. Baker, J.J. Calvino, S.
Bernal, Catal. Today 50 (1999) 271.
34] J.P. Cuif, G. Blanchard, O. Touret, A. Seigneurin, M. Marczi, E.
Qu e´ mer e´ , SAE, 961906 (1996).
[54] P. Fornasiero, E. Fonda, R. Di Monte, G. Vlaic, J. Kaspar, M.
Graziani, J. Catal. 187 (1999) 177.
[
[
35] S. Meriani, Mater. Sci. Eng. A-Struct. Mater. 109 (1989) 121.
36] S. Tsunekawa, R. Sivamohan, S. Ito, A. Kasuya, T. Fukuda,
Nanostruct. Mater. 11 (1999) 141.
[55] A. Tschope, J.Y. Ying, Nanostruct. Mater. 6 (1995) 1005.
[56] S. Otsuka-Yao-Matsuo, N. Izu, T. Omata, K. Ikeda, J. Electro-
chem. Soc. 145 (1998) 1406.
[
37] R.D. Shannon, Acta Crystallogr. A32 (1976) 751.