Cerium effect on the phase structure, phase stability and redox properties of Ce-doped strontium ferrates

Article abstract of DOI:10.1016/j.jssc.2006.06.027

Nanostructured perovskite-type Sr_1-aCe_aFeO_3-x, (0≤a<0.15) powders have been prepared by citrate-nitrate smoldering auto-combustion. Their phase structure and stability, surface and morphological properties, reduction behavior and interaction with oxygen have been investigated by X-ray Powder Diffraction combined with Rietveld Analysis, ⁵⁷Fe M?ssbauer and X-ray Photoelectron Spectroscopies, N₂-adsorption method, Temperature Programmed Reduction and Oxidation experiments. Our results reveal that citrate-nitrate auto-combustion method is effective in obtaining single phase Sr_1-aCe_aFeO_3-x. The Sr_1-aCe_aFeO_3-x structure is cubic only for a≥0.06, while for a<0.06 remains tetragonal. Moreover, for a≥0.06 after semi-reductive treatment under inert gas, an expanded cubic phase is obtained instead of the brownmillerite-type structure, which is known to have ordered vacancies. Stabilization of octahedral Fe³⁺ by cerium doping appears to be the main factor in determining the structural properties of Sr_1-aCe_aFeO_3-x. The highest oxygen consumption for Ce-doped SrFeO₃ occurs for a=0.06. Preliminary impedance measurements show that Sr_0.94Ce_0.06FeO_3-x has the lowest area-specific resistance.

Full text of DOI:10.1016/j.jssc.2006.06.027

ARTICLE IN PRESS
Journal of Solid State Chemistry 179 (2006) 3406–3419
www.elsevier.com/locate/jssc
Cerium effect on the phase structure, phase stability and redox
properties of Ce-doped strontium ferrates
Ã
F. Deganelloa, , L.F. Liottaa, A. Longoa, M.P. Casalettoa, M. Scopellitib
aIstituto per lo Studio dei Materiali Nanostrutturati del C.N.R, Sezione di Palermo, Via Ugo La Malfa 153, 90146-Palermo, Italy
`
Dipartimento di Chimica Inorganica e Analitica’’S. Cannizzaro’’, Universita di Palermo Parco d’Orleans, Viale delle Scienze, 90128-Palermo, Italy
b
Received 2 May 2006; received in revised form 5 June 2006; accepted 9 June 2006
Available online 29 June 2006
Abstract
Nanostructured perovskite-type Sr1ꢀaCeaFeO3ꢀx, (0pao0.15) powders have been prepared by citrate–nitrate smoldering auto-
combustion. Their phase structure and stability, surface and morphological properties, reduction behavior and interaction with oxygen
have been investigated by X-ray Powder Diffraction combined with Rietveld Analysis, 57Fe Mossbauer and X-ray Photoelectron
¨
Spectroscopies, N2-adsorption method, Temperature Programmed Reduction and Oxidation experiments. Our results reveal that
citrate–nitrate auto-combustion method is effective in obtaining single phase Sr1ꢀaCeaFeO3ꢀx. The Sr1ꢀaCeaFeO3ꢀx structure is cubic
only for aX0.06, while for ao0.06 remains tetragonal. Moreover, for aX0.06 after semi-reductive treatment under inert gas, an
expanded cubic phase is obtained instead of the brownmillerite-type structure, which is known to have ordered vacancies. Stabilization of
octahedral Fe3+ by cerium doping appears to be the main factor in determining the structural properties of Sr1ꢀaCeaFeO3ꢀx. The highest
oxygen consumption for Ce-doped SrFeO3 occurs for a ¼ 0.06. Preliminary impedance measurements show that Sr0.94Ce0.06FeO3ꢀx has
the lowest area-specific resistance.
r 2006 Published by Elsevier Inc.
Keywords: Ce-doped strontium ferrates; Perovskite-type materials; Smoldering auto-combustion synthesis; Octahedral Fe3+
1. Introduction
ferrates release large amounts of oxygen on heating in inert
gas flow, after pre-adsorption of oxygen [8]. However, as
the oxygen vacancy concentration increases, the probabil-
ity of forming much less oxygen-conducting ordered
structures also increases. For example, the brownmiller-
ite-type structure ABO2.5 can be obtained from the
perovskite-type structure ABO3 under inert gas treatment
at high temperature [9]. The crystal structure of the
brownmillerite-type phase is related to the perovskite
phase: while the ABO3 perovskite phase consists of BO6
octahedra, the ABO2.5 brownmillerite phase has BO6
octahedra alternating with BO4 tetrahedra along the
b-axis, resulting in an orthorhombic symmetry, containing
Fe3+, with ordered oxygen vacancies. The orthorhombic
brownmillerite-type phase transforms into higher oxygen-
conducting cubic phase at T4800 1C, due to a thermal
disordering of the oxygen vacancies [9]. Knowing the
behavior of strontium ferrate in a large temperature and
oxygen pressure range is of outmost importance for their
application in SOFC and oxygen permeation membranes.
Strontium ferrates are strontium–iron mixed oxides with
ABO3 perovskite-type structure. Iron in strontium ferrates
is found mainly as Fe4+, but oxygen deficiency is usually
present so that some Fe3+, with larger ionic radius, is
formed depending on the synthesis conditions. Pure and
doped strontium ferrates have attracted much interest as
sensors [1], magnetotransport [2,3] and oxygen permeation
membrane materials [4]. Recently they have been studied
for their mixed ionic–electronic conducting properties,
which make them promising candidate materials for
oxygen-conducting devices such as solid oxide fuel cell
(SOFC) cathodes [5–7]. The large oxygen non-stoichiome-
try range that these perovskite materials are able to
accommodate is an important factor to achieve high
oxygen mobility values. It has been reported that strontium
Ã
Corresponding author. Fax: +39 091 6809399.
E-mail address: francesca.deganello@ismn.cnr.it (F. Deganello).
0022-4596/$ - see front matter r 2006 Published by Elsevier Inc.
doi:10.1016/j.jssc.2006.06.027

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3407
Proper substitution of the A or B site with metal cations
2. Experimental procedure
may improve the phase and structure stability and oxygen
permeability by affecting the iron oxidation state. Among
the dopant cations investigated in literature, lanthanum
has been reported to produce rhombohedral or tetragonal
structure, depending on the dopant content, and to form
brownmillerite structure after high-temperature treatment
at low oxygen partial pressure [10]. Other authors reported
a cubic structure for La-doped SrFeO3, although their
powders were prepared under high oxygen pressure
thermal annealing [11]. On the contrary, Pr [12] and Bi-
doped [4] SrFeO3 are usually cubic or tetragonal, depend-
ing on the dopant content, and form expanded cubic
perovskite-type structure after oxygen loss at low oxygen
partial pressure. In contrast with other A-site-doped
SrFeO3, Ce-doped SrFeO3 show some differences in the
full stoichiometric structures: for example, while BiFeO3
and PrFeO3 do exist, CeFeO3 can be obtained only in
harsh reductive environments or by using particular
synthesis condition [13]. To our knowledge, only few
papers have been published on Ce-doped strontium
ferrates [5,14,15]. In particular, Trofimenko and Ullmann
[14] demonstrated the mixed ionic–electronic conductivity
of (Sr,Ce)(Fe,Co)O3ꢀx both by oxygen permeation and
oxide ion and electronic conductivity measurements. They
showed that a Sr1ꢀaCeaFeO3ꢀx solid solution with a cubic
structure forms for ap0.15 by substituting Sr2+ ion
(rSr2þ ¼ 1:44 ) with Ce3+ ion (rCe3þ ¼ 1:34 ). Cerium doping
causes oxygen vacancies suppression and modifies some of
the properties of strontium ferrates like oxygen and
electronic conduction [14].
This paper describes the preparation and characterization
of Sr1ꢀaCeaFeO3ꢀx perovskite-type materials in order to find
new structure–properties relationships and useful indications
for the application of these materials in low-temperature
oxygen-conducting devices. The preparation of the full
range of Sr1ꢀaCeaFeO3ꢀx perovskite-type mixed oxides was
carried out by a citrate–nitrate smoldering auto-combustion
method. The powder generally obtained by solution combus-
tion methods has an higher degree of phase purity with
improved powder characteristics, such as narrow crystal size
distribution, nanometer-sized particles, high surface area and
good sinterability [16]. All these features may dramatically
affect both the electronic and ionic transport properties,
leading to improved performances [17]. The structure of the
Ce-doped strontium ferrate powders has been studied by
X-ray powder diffraction (XRD) combined with Rietveld
2.1. Powders preparation
For the preparation of the powdered samples, a method
based on the citrate–nitrate auto-combustion has been
employed. A detailed description of the powder prepara-
tion will be published in a forthcoming paper [18]. Briefly,
stoichiometric amounts of strontium nitrate (Carlo Erba X
99.0%), iron nitrate (Carlo Erba 98%) and cerium nitrate
(Sigma Aldrich 99.99%) have been dissolved in distilled
water. Then citric acid (Merk 99%), in a twofold molar
excess of the total metal ions, has been dissolved into the
metal nitrates solution and the pH of the resulting mixture
has been adjusted to 9 by NH3 solution (30%). Finally
ammonium nitrate (Merk X99%), has been added to
obtain a citrate/nitrate ratio of 0.46. The combustion
process is mainly regulated by this ratio, where citrate is the
fuel, whereas nitrates, coming from metal precursors and
from ammonium nitrate, are the oxidants.
The combustion of the homogeneous solution has been
carried out in a high-temperature silicon oil bath. The
water has been evaporated at 70 1C under constant
mechanical stirring, until a green gel was obtained. The
temperature of the bath has been then gradually raised to
200 1C until self-ignition. The spontaneous combustion
lasted for about 10–20 s resulting in a fluffy powder. This
powder has been then fired in stationary air at 1000 1C for
4 h. The powders obtained after firing at 1000 1C have been
named ‘‘as prepared powders’’. In this paper,
Sr1ꢀaCeaFeO3ꢀx perovskite-type materials are designated
by the abbreviation SFC followed by a number that refers
to the nominal Ce molar percentage in the material.
2.2. Powders characterization
X-ray diffraction measurements have been carried out
on a PHILIPS PW1830 X-ray powder diffractometer using
Ni-filtered Cu Ka radiation (40 kV/40 mA). A step size of
0.0512y, an integration time of 3 s per step and a scan range
from 20 to 9812y were used. The diffraction patterns have
been analyzed by Rietveld refinement using the GSAS
package [19]. Chebyschev polynomials and Pearson VII
functions have been chosen for the background and for the
peak profile fitting, respectively. In the structure refine-
ment, lattice constants, Debye Waller factors, microstrain
and full width half maximum (FWHM) have been
considered as variables parameters. For all the fitting
analysis [20], parameters have been considered. From
fitting results, the structural parameters of the investigated
compounds and, in particular, the cell edge lengths and the
relative phase composition were obtained. Thermal para-
meters were optimized and their values were found in good
agreement with the literature. The agreement factors,
reported in the tables as quality indicators for the goodness
of the Rietveld fitting, are the weighted-profile value (wRp),
the expected profile value (Rp) and the Bragg-intensity
analysis; the iron electronic environment by Mossbauer
¨
measurements; surface area and pore distribution by N2
absorption method; surface modifications by X-ray photo-
electron spectroscopy (XPS) and the redox properties
by temperature programmed reduction (TPR) and
oxidation (TPO) experiments. Preliminary electrochemical
impedance spectroscopy (EIS) measurements have been also
performed to get a first insight of the cerium effect on the
electrochemical performance of these compounds as cathodes
for oxygen reduction.

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3408
value (R2F), whose detailed definitions are given elsewhere
in the literature [20]. All the agreement factors are
expressed as percentage values and they should be lower
than 10%. An estimation of the crystal size values has been
obtained from Sherrer equation in agreement with the
GSAS package procedure.
has been repeated twice, after a reoxidation at 1000 1C (rate
of 10 1C/min) and He cooling. The standard deviation for
the overall hydrogen uptake is 715% and the error on the
peak temperature is 715 1C.
TPO experiments have been performed after temperature
programmed desorption (TPD) pre-treatment. TPD pre-
treatment has been carried out in He flow by heating up the
as-prepared powders to 1000 1C and cooling them down to
RT in He flow. In a typical TPO experiment, immediately
after the TPD pre-treatment, the powder was heated from
RT to 1000 1C by flowing O2(5%)/He (30 mL/min) until
1000 1C. In order to check the reversibility of process, the
TPD–TPO cycle has been repeated twice.
57Fe Mossbauer spectra have been collected by a
¨
conventional spectrometer with triangular drive signal
(FG2 function generator and MR250 velocity transducer,
Munchen Wissenschaftliche Elektronik Germany; propor-
¨
tional Kr/CH4 counter, ADC 1683; memory unit/multi-
channel analyzer 269, Takes, Italy). For all experiments, a
57Co/Rh source (10 mCi, Ritverc GmbH, Russian Federa-
tion) has been used. The velocity calibration has been done
by measuring the hyperfine splitting of a-57Fe foil (4 mm,
Ritverc GmbH). All isomer shifts are referred to a-Fe.
Spectra have been evaluated using Lorentz and Voigt
lineshapes in a non-linear MEM curve fitting software.
Since the temperature dramatically influences the oxygen
content of the samples, and thus their structure and
geometry [21], only room temperature (RT) spectra were
collected.
XPS analysis has been performed by a VG Microtech
ESCA 3000 Multilab spectrometer, equipped with a
standard Al Ka excitation source (hn ¼ 1486.6 eV), a nine-
channeltrons detection system and an electron hemisphe-
rical analyzer operating at a constant pass energy of 20 eV.
The binding energy (BE) scale has been calibrated by
measuring C 1s peak (BE ¼ 285.1 eV) from the ubiquitous
surface layer of adventitious carbon. The accuracy of the
BEs location is 70.1 eV. A non-linear least-square peak
fitting routine has been used for the analysis of XPS
spectra, separating elemental species in different oxidation
states, after a Shirley [22] background subtraction from all
the spectra. Relative concentrations of chemical elements
have been calculated by a standard quantification routine,
including Wagner’s energy dependence of attenuation
length [23] and a standard set of VG Escalab sensitivity
factors.
Surface area (BET method) [24] and mesopore size
distributions (BJH method) [25] have been determined by
nitrogen adsorption/desorption measurements at 77 K,
using a Sorptomatic 1900 Carlo Erba Instrument. A pre-
treatment under vacuum at 250 1C for 1.5 h has been used.
Reduction properties of the perovskite oxides have been
studied by TPR measurements. Experiments have been
carried out with a Micromeritics Autochem 2910 appara-
tus, equipped with a thermal conductivity detector and a
mass quadrupole (ThermostarTM, Balzers). A weighted
portion of the samples of about 100 mg has been pre-
treated in a quartz reactor in 5% O2 in He (30 mL/min) at
500 1C for 1 h, followed by purging and cooling to RT in
Ar flow. Then, the flow has been switched to the mixture
5% H2 in Ar (30 mL/min) and the temperature has been
increased linearly up to 1000 1C (rate of 10 1C/min). In
order to investigate the effects of sample aging on the
reduction properties of the samples, the TPR experiment
2.3. EIS characterization
Preliminary EIS measurements have been performed on
symmetric cells of Sr1ꢀaCeaFeO3ꢀx (a ¼ 0, 0.06, 0.15) on a
Ce0.8Sm0.2O2ꢀx electrolyte to investigate the effect of Ce
doping on the area-specific resistance (ASR).
Dense electrolyte pellets for impedance measurements
have been prepared by weighting 0.8 g of Ce0.8Sm0.2O2ꢀx
powders, prepared by citrate–nitrate auto-combustion
method, and by forming the powders by uniaxially pressing
(2 tons for 1 min). A final diameter of E11 mm and a
thickness of E1.4 mm have been obtained. Pellets have
been pre-sintered at 1100 1C for 5 h and then further
sintered for 10 h at 1500 1C. The pellet surfaces have been
carefully polished before the electrode deposition.
SFC powders have been mixed with PEG (MW 400) and
ethanol. Then homogenization has been ensured by mixing
for 30 min in an ultrasonic bath. Alcohol evaporation has
been performed in an oven at 80 1C. Coating of the
electrodes on the electrolyte has been performed by screen
printing of the SFC powder pre-sintered at 1000 1C. After
coating the first side of the pellet, the pellet has been left for
30 min at 160 1C to dry the coating layer. Then the other
side has been coated and dried again at 160 1C. The
sintering temperature has been set to 1000 1C for 1 h.
EIS measurements have been carried out in the
600–800 1C temperature range using a frequency response
analyzer (FRA, Solartron 1260 Impedance/Gain-phase)
coupled with a dielectric interface (Solartron 1296), with an
AC amplitude of 10 mV in the frequency range from
0.01 Hz to 13 MHz. Electrical current collectors have been
made of gold wires on one side and a net on the other side
to ensure contact area between gold and cathodes. The
interfacial resistance of cathodes has been measured in the
two-electrode symmetric cell configuration under air.
3. Results and discussion
3.1. As-prepared powders
Phase composition and structure of the as-prepared
samples have been determined by XRD after firing at
1000 1C. In order to investigate cerium effect on strontium

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3409
ferrates structure, an accurate Rietveld analysis on the
XRD data has been performed and the results are
summarized in Table 1.
Fig. 1 shows some representative Rietveld fittings. In this
work, the undoped strontium ferrate (SF) has a tetragonal
I4/mmm structure (Fig. 1a,d), similar to Sr8Fe8O23 already
Table 1
Rietveld refinement results for Sr1ꢀaCeaFeO3ꢀx samples with different Ce content
Sample name
Phase compositiona (% wt)
XRD space group (structure)
b
Cell parameters (A)
Crystal size (nm)
Agreement factors (%)
˚
wRp ¼ 1.74
Rp ¼ 1.22
R2F ¼ 8.68
SF
Sr8Fe8O23 (100)
I4=mmm (tetragonal)
a0 ¼ b0 ¼ 10.932; c0 ¼ 7.701
248(20)
115(5)
wRp ¼ 2.13
Rp ¼ 1.39
R2F ¼ 8.14
SFC2
Sr0.98Ce0.02FeO3ꢀx(100)
P4=mmm (tetragonal)
a0 ¼ b0 ¼ 3.864
c0 ¼ 3.854
wRp ¼ 2.21
Rp ¼ 1.45
R2F ¼ 5.72
SFC6
Sr0.94Ce0.06FeO3ꢀx (100)
Pm ꢀ 3m (cubic)
92(5)
101(5)
89(5)
70(5)
a0 ¼ b0 ¼ c0 ¼ 3.872
wRp ¼ 2.11
Rp ¼ 1.37
R2F ¼ 3.68
SFC10
SFC15
SFC25
Sr0.90Ce0.10FeO3ꢀx (100)
Pm ꢀ 3m (cubic)
a0 ¼ b0 ¼ c0 ¼ 3.877
wRp ¼ 1.86
Rp ¼ 1.30
R2F ¼ 5.39
Sr0.85Ce0.15FeO3ꢀx (99.0(2))
CeO2 (o1)
Pm ꢀ 3m (cubic)
a0 ¼ b0 ¼ c0 ¼ 3.884
wRp ¼ 2.21
Rp ¼ 1.52
R2F ¼ 4.73
Sr0.85Ce0.15FeO3ꢀx (91.8(3))
CeO2 (8.2(1))
Pm ꢀ 3m (cubic)
a0 ¼ b0 ¼ c0 ¼ 3.884
a‘‘a’’ values are deduced from the nominal Ce content and from the phase composition.
b
˚
Standard deviation for cell parameters is 70.003 A.
(b)
(c)
Obs
Calc
diff
Obs
Calc
diff
SFC2
SFC6
20
40
60
80
20
(d)
40
60
80
(a)
Obs
Calc
diff
SF
SFC2
SFC6
SF
20
40
60
80
77.5
78
78.5
79
2Θ (°)
2Θ (°)
Fig. 1. Representative XRD structures of fresh Ce-doped strontium ferrates: (a) Rietveld refinement of SF with I4=mmm structure; (b) Rietveld
refinement of SFC2 XRD pattern with P 4=mmm structure; (c) Rietveld refinement of SFC6 XRD pattern with Pm ꢀ 3m structure; (d) enlargement of the
SF, SFC2 and SFC6 XRD patterns with different structures.

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3410
described in literature [26]. It is already known in the
literature that undoped strontium ferrates (SrFeO3ꢀx) may
have different XRD structures depending on their oxygen
stoichiometry [26], which is in turn determined by the
synthesis conditions. The more the oxygen content, the
more symmetric the structure, i.e. for a minimum oxygen
content (x ¼ 0.5), the structure is orthorhombic, whereas
for a maximum oxygen content (x ¼ 0) the structure is
pure cubic. An intermediate oxygen content is usually
obtained, whereas the extreme oxygen content can be
obtained only in particular synthesis conditions [26].
For a 2 mol % Ce content (SFC2), the XRD pattern can
be well described by a tetragonal P4=mmm structure with
c0/a0o1 (Fig. 1b,d), as already found for Sr0.9Bi0.1FeO3ꢀx
[27]. For a Ce content X6 mol%, a cubic Pm3m structure
forms, as shown for the SFC6 sample in Fig. 1c and Fig.
1d. The same cubic structure for Ce-doped strontium
ferrates was found by Trofimenko and Ullmann [14],
although they reported cubic structures for the whole range
of dopant content. This discrepancy is probably due to the
different preparation and sintering conditions used by
these authors [14].
By increasing the cerium content in Sr1ꢀaCeaFeO3ꢀx, the
cell parameter increases, as indicated by the shift of the
diffraction peak toward lower angles in Fig. 2a. For higher
cerium content, a CeO2 phase segregates, as clearly shown
by the phase composition of the most concentrated sample
(SFC25) in Table 1. Due to no weight loss during the
spontaneous combustion and formation of single-phase
perovskite, the SFC stoichiometry (‘‘a’’ value), which has
been reported in Table 1, has been assumed to be that of
the metal nitrates precursors in the starting solution, except
for SFC15 and SFC25, where some segregated CeO2 is also
present and ‘‘a’’ value corresponds to the maximum cerium
solubility. Fig. 2b shows that a linear correlation between
cell volume (normalized to the number of atoms in the cell
unit) and cerium nominal content exists until 15 mol%, in
good agreement with literature results [14]. From these
results, the limit of Ce solubility in Sr1ꢀaCeaFeO3ꢀx
corresponds to 14–15 mol%. This value is very close to
that already found by Trofimenko and Ullmann [14].
The cell expansion observed for increasing Ce content
can be ascribed to the formation of Fe3+ cations, with
higher ionic radius (rFe3þ ¼ 0:645 [28]) than Fe4+
(rFe4þ ¼ 0:585 [28]). Fe3+ formation has been confirmed
by RT Mossbauer spectroscopy measurements on fresh SF,
¨
SFC6 and SFC15 samples (Table 2 and Fig. 3). These
samples are paramagnetic at RT, and all contain Fe4+
alongside with Fe3+, even if some discussion is needed. In
agreement with the above discussed XRD results, Moss-
¨
bauer results showed that the structure of SF corresponds
to Sr8Fe8O23 (or SrFeO2.87) and that our experimental
spectra are identical to the literature [29]. Our interpreta-
tion (Fig. 2 and Table 2), however, is slightly different: we
found octahedral, high spin Fe4+, the unusual oxidation
state Fe3.5+ (very common in SrFeO2.86) and a distorted
Fe3+ site respect an octahedral geometry, with distortion
caused by either geometric reasons (tetragonal site) or by a
coordinative insaturation (CNo6). If we assume that the
unusual state (3.5+) is in fact a continuous electron
exchange between Fe4+ and Fe3+, too fast for Mossbauer
¨
to be unequivocally determined, we can deduce the amount
of Fe4+ and Fe3+ originally present in the sample. These
data can be correlated with the amount on Fe4+ and Fe3+
found in SF6 and SFC15 (Fig. 2 and Table 2), for which
the Mossbauer analysis showed octahedral high-spin Fe4+
¨
and octahedral Fe3+ sites. This correlation is linear, and
Fe4+ decreases with the increase of cerium content,
analogously to other A-site doped ferrates [4,11]. Appar-
ently, thus, cerium is able to stabilize the formation of
octahedral Fe3+ (Table 2), suggesting a ‘‘preference’’—for
the doping element—for cubic environments vs. tetragonal
ones (Table 1).
From Rietveld refinements, high microstrain parameter
˚
values have been obtained (between 6 and 12 A). These
unusual values could be essentially due to the presence of
two different iron oxidation states. The crystallite size of
the investigated powders decreases from 250 nm, in the
pure strontium ferrate (SF), to 70–120 nm, in the Ce-doped
samples (Table 1).
Differences in surface area and pore volume for different
cerium doping content have been found by N2 adsorption
58.8
SF
58.6
58.4
58.2
58.0
57.8
57.6
57.4
SFC2
SFC6
SFC10
SFC15
SFC25
47
2Θ (˚)
47.5
-2
0
2
4
6
8
10 12 14 16 18 20 22 24 26
46.5
(a)
(b)
Ce nominal content (%)
Fig. 2. (a) Expanded XRD pattern of Ce-doped strontium ferrates between 45.5 and 48.512y; (b) linear increasing of cell volume with cerium nominal
content.

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3411
Table 2
Mossbauer parameters for SF, SFC6 and SFC15 samples
¨
Assignment
d (mm sꢀ1
)
Da or 2eb (mm sꢀ1
)
G (mm sꢀ1
)
H (T)
Spectral area (%)
SF
Fe4+
Fe3.5+(o)
Fe3+
0.05570.002
0.15870.005
0.7870.02
0.4570.01
0.41270.008
0.3070.01
3971
5671
571
0.6670.01
0.5270.05
SFC6
Fe4+
Fe3+(o)
ꢀ0.04970.001
0.32370.003
0.48970.003
0.54670.008
6571
3571
SFC15
Fe4+
Fe3+(o)
ꢀ0.09770.002
0.43770.004
0.55570.007
0.57870.009
5671
4471
SFC6/He
SFC15/He
SF/H2
Fe3+(o)
0.27970.003
0.16270.005
ꢀ0.06170.006
0.10370.009
0.7770.01
1.3570.01
48.8570.03
41.870.1
4871
5271
Fe3+(t or p)
Fe3+(o)
Fe3+(t or p)
0.25970.003
0.14270.002
ꢀ0.06170.003
0.0470.01
0.79470.009
1.2870.02
48.8570.03
41.870.1
5871
4271
Fe0
Fe2+c
Fe3+
ꢀ0.006770.0008
0.35770.002
0.00970.001
0.19570.002
ꢀ0.33670.002
0.29870.002
0.3070.01
0.42670.004
33.03170.006
3771
1171
5271
0.28270.001
38.8270.01
SFC15/H2
Fe0
Fe2+c
Fe3+
ꢀ0.10270.001
ꢀ0.4607
ꢀ0.00270.002
0.30370.003
1.170.2
1.3570.02
33.01170.007
3671
1271
5271
0.5470.01
ꢀ0.5270.01
ꢀ0.0270.01
35.0670.08
aQuadrupole splitting (paramagnetic phase).
bQuadrupole splitting (magnetically ordered phase) o ¼ octahedral; t ¼ tetrahedral; p ¼ penta-coordinated.
cFe2+ as FeO.
Fig. 3. Fits of Mossbauer spectra at RT for fresh SF (a), SFC6 (b) and SFC15 (c) samples: the fitted peak at lower d value is related to Fe4+ ions; the fitted
¨
peak at higher d value to Fe3+ ions.
measurements. For SF and SFC2, both surface area
(ffi10 m2/g) and porosity (ffi0.01 cm3/g) are quite low; for
higher Ce content, the surface area doubles and porosity
increases up to 0.025 cm3/g. SFC25, which contains 8 wt%
of segregated CeO2 shows the highest surface area
(ffi35 m2/g) and pore volume (ffi0.033 cm3/g).
XPS quantitative results of the surface characterization of
the samples are listed in Table 3. The relative chemical

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3412
Table 3
Binding energies (BE) of the XPS peaks and relative surface chemical composition of the investigated samples
Sample
Sr 3d
O 1s
C 1s
Fe 2p
Ce 3d
BE (eV)
SF
SFC6
133.8
11.4
7.3
529.0
39.3
25.4
32.1
32.2
24.9
42.0
44.2
285.1
39.3
61.5
46.4
46.6
61.2
31.4
29.6
710.1
10.0
5.8
881.8
—
Sr/Fe
1.1
1.3
n. d.
2.7
SFC10
SFC15
SFC25
SFC15/He
SFC15/H2
8.6
7.9
10.2
9.8
0.8
0.8
3.5
2.4
5.1
6.4
0.8
0.9
11.0
14.1
12.6
8.7
2.9
3.4
1.6
Elemental concentrations are expressed as atomic percentage (atom. %).
Fig. 4. XPS curve-fitting of the following photoelectron peaks: (a) Sr 3d, (b) C 1s, (c) O 1s and (d) Fe 2p. The scattered line refers to the raw data, while the
solid line to the curve-fitting results.
composition is expressed as atomic percentage (at%). The
surface (Sr/Fe) atomic ratio ranges from 0.8 to 1.3 for all the
investigated as-prepared samples, as reported in Table 3.
This (Sr/Fe) value corresponds to the stoichiometric bulk
ratio (Sr/Fe ¼ 0.85–1.00), at least within the experimental
error. Representative XPS curve-fittings of the main
photoelectron peaks in the samples are shown in Fig. 4.
The Sr 3d region in Fig. 4a can be fitted by two doublets (Sr
3d3/2 and Sr 3d1/2) located at BE ¼ 132.7 and 134.1 eV,
which can be assigned to surface Sr–O and Sr–CO3 species,
respectively [30–32]. SrCO3 formation on the surface is
around 23–37% with respect to the total Sr peak area (taken
as 100%) in the perovskite samples. By analyzing the
samples composition as a function of the ageing time, an
almost linear correlation between the carbonate amount
formed on the surface and the time elapsed from the
synthesis of the sample has been found (not shown). Freshly
prepared samples are carbonate-free and the surface
carbonate amount increases with time up to 30% after 1
month, as a result of interaction with air. These surface
carbonate species can be easily removed from the sample by
a mild temperature treatment. The presence of SrCO3
species on the samples surface is also confirmed by the C 1s
and O 1s peak fittings. C 1s spectra can be fitted by three
components at BE ¼ 285.1, 286.6 and 289.8 eV, which are
attributable to aromatic hydrocarbons from surface con-
tamination, alcohoxy (CH2Ox) and surface COꢀ3 2 groups,
respectively [30–32], as shown in Fig. 4b. In all the samples,
O 1s spectra consist of three components reported in Fig. 4c,
located at BE ¼ 529.0, 531.2 and 532.5 eV, which can be
assigned to bulk oxide species (O2ꢀ), surface Sr–O and
Sr–COꢀ3 2 groups, respectively [30,31].

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3413
The XPS region in the range 700–740 eV shown in Fig.
3.2. He-treated powders
4d reveals the complex spectral features of the Fe 2p
photoelectron peak. The multiplet splitting effect is
responsible of the broadening of the peaks and the presence
of satellites structures can be ascribed to shake-up
processes [33]. The core-level Fe 2p spectrum consists of
When the samples have been treated under He at 1000 1C
and cooled under the same atmosphere, some structural
transformations occurred. For SF and SFC2 samples,
XRD showed that orthorhombic brownmillerite-type
structures have been formed (Fig. 5a and Table 4). Minor
amounts of Sr4Fe4O11 (Fig. 5a and Table 4) have been also
detected, in agreement with the literature [35]. Our results
for Ce content X 6 mol % show that after high-
temperature He treatment, the structure remains cubic
the typical spin-orbit doublet: Fe 2p3/2 and Fe 2p1/2
,
separated by 13.1 eV. The Fe 2p3/2 peak is centered at
BE ¼ 710.1 eV and shows satellite shake-up features. Since
very close BE values are detected by XPS for Fe2+ and
Fe3+ species, it is rather difficult to quantify their relative
distribution [32]. The curve fitting result and the compar-
ison with a pure sample of Fe2O3 powder taken as
reference (see below the paragraph ‘‘H2-treated powders’’)
allow the assignment to the contemporary presence of
Fe2+ and Fe3+ species for the corresponding core-level Fe
˚
(Fig. 5b,d) although a0 expands from 3.861 to 3.884 A, for
˚
the as-prepared samples (Table 1), to around 3.90 A, for
the He-treated samples (Table 4). When the powders are
treated in He at high temperature, the microstrain
˚
parameter tends to decrease to around 4 A for any Ce
2p3/2 BEs [30–32]. Although Mossbauer results on fresh
content, since only Fe3+ is present. Crystallite size values
do not change very much after this treatment (Table 4).
Formation of cubic structures with expanded lattice has
already been evidenced in literature for Pr and Bi-doped
SrFeO3, and it has been attributed to Fe3+ formation
under inert or reducing conditions [4,12]. In this work,
Fe3+ formation in the expanded phases has been confirmed
¨
samples show also the presence of bulk Fe4+ species, a
partial iron reduction occurs during the experimental
conditions (ultrahigh vacuum and X-ray irradiation) of
XPS measurements, leading only to the detection of the
more stable surface Fe3+ and Fe2+ oxidation state. This is
not surprising since it is known in the literature that
undoped [32,34] and doped [10] strontium ferrates loose
oxygen under low oxygen partial pressure at around
600–700 1C, by forming structures which contain just
Fe3+. The same reduction probably occurred under
vacuum at RT under XPS experimental conditions.
by Mossbauer spectroscopy measurements on He-treated
¨
SFC6 and SFC15, which showed also that, after He
treatment, no detectable amount of Fe4+ was present.
Moreover, the presence of two distinct magnetically
ordered Fe3+ sites was evidenced, an highly symmetric
(d)
(a)
Obs
SF (He)
Calc
diff
SFC15 (He)
Sr4Fe4O11
SF after He treatment
Sr4Fe4O11
20
40
60
80
31
31.5
32
32.5
33
33.5
(b)
(c)
Obs
Obs
Calc
diff
Calc
diff
SFC15 after H2 treatment
αFe
SFC15 after He treatment
CeO2
20
30
40
50
60
70
80
20
40
60
80
2Θ (˚)
2Θ (˚)
Fig. 5. Representative XRD structures of Ce-doped strontium ferrates after He or H2 treatment: (a) Rietveld refinement of SF XRD pattern after He
treatment with brownmillerite-type structure and some Sr4Fe4O11; (b) Rietveld refinement of SFC15 XRD pattern after He treatment with cubic structure;
(c) Rietveld refinement of SFV15 sample after H2 treatment with Ruddlesden Popper Sr3Fe2O6 structure and metallic iron; (d) enlargement of the SFC6
and SFC15 XRD patterns with different structures.

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3414
Table 4
Rietveld refinement results for He treated (after cooling from 1000 1C under He flow) Sr1ꢀaCeaFeO3ꢀx samples with different Ce content
a
Phase composition (% wt)
Sample name
X-ray space group (structure) Cell
b
parameters (A)
Crystal size (nm)
Agreements
factors (%)
˚
SF/He
Sr2Fe2O5 (89.4(8))
Sr4Fe4O11 (10.6(2))
Icmm (orthorhombic)
a0 ¼ 5.669; b0 ¼ 15.587; c0 ¼ 5.530
Cmmm (orthorhombic)
177(5)
256(30)
wRp ¼ 1.51
Rp ¼ 1.20
R2F ¼ 8.23
a0 ¼ 10.990; b0 ¼ 7.716;c0 ¼ 5.483
Icmm (orthorhombic)
SFC2/He
Sr2(1ꢀa)Ce2aFe2O5 (84.4(1))
Sr4Fe4O11 (15.6(4))
173(10)
101(10)
wRp ¼ 1.83
Rp ¼ 1.35
R2F ¼ 7.79
a0 ¼ 5.658; b0 ¼ 15.626; c0 ¼ 5.532
Cmmm (orthorhombic)
a0 ¼ 11.013; b0 ¼ 7.732;c0 ¼ 5.494
SFC6/He
FC10/He
FC15/He
C25/He
Sr0.94Ce0.06FeO3ꢀx (100)
Sr0.90Ce0.10FeO3ꢀx (100)
Pm3m (cubic)
a0 ¼ b0 ¼ c0 ¼ 3.909
124(10)
76(5)
wRp ¼ 2.35
Rp ¼ 1.60
R2F ¼ 6.10
wRp ¼ 2.21
Rp ¼ 1.43
R2F ¼ 5.87
wRp ¼ 2.23
Rp ¼ 1.55
R2F ¼ 8.11
wRp ¼ 1.79
Rp ¼ 1.32
R2F ¼ 3.80
Pm3m (cubic)
a0 ¼ b0 ¼ c0 ¼ 3.902
Sr0.85Ce0.15FeO3ꢀx (99.0(2))
CeO2 (o1)
Pm3m (cubic)
a0 ¼ b0 ¼ c0 ¼ 3.901
86(5)
Sr0.85Ce0.15FeO3ꢀx (91.3(2))
CeO2 (8.7(1))
Pm3m (cubic)
a0 ¼ b0 ¼ c0 ¼ 3.899
85(5)
bis Rietveld refinement results for H2 treated (after TPR and cooling from 1000 1C under He flow) Sr1ꢀaCeaFeO3ꢀx samples with 0 and 15 mol% Ce
SF/H2
Sr3Fe2O6 (88(2))
Fe (12(3))
I4=mmm (Tetragonal)
a0 ¼ b0 ¼ 3.892; c0 ¼ 20.038
Im ꢀ 3m (Cubic)
100(5)
110(10)
wRp ¼ 2.60
Rp ¼ 1.98
R2F ¼ 6.87
a0 ¼ b0 ¼ c0 ¼ 2.864
SFC15/H2
Sr3(1ꢀa)Ce3aFe2O6 (83.3(1))
CeO2 containing Sr3Fe2O6
(5.6(5))
I4=mmm (Tetragonal)
a0 ¼ b0 ¼ 3.901; c0 ¼ 20.096
106(5)
79(10)
wRp ¼ 2.67
Rp ¼ 1.76
Fe (11.1(3))
I4=mmm (Tetragonal)
a0 ¼ b0 ¼ 3.852; c0 ¼ 21.232
Im ꢀ 3m (Cubic)
126(10)
R2F ¼ 8.93
a0 ¼ b0 ¼ c0 ¼ 2.867
a‘‘a’’ values are deduced from the nominal Ce content and from the phase composition.
b
˚
Standard deviation for cell parameters is 70.003 A.
octahedral one (characterized by an higher hyperfine
magnetic field) and another one, with lower coordination
number (probably 4) (Table 2). These results are very
similar to those obtained for Pr-doped SrFeO3, although in
the literature the contribution of tetra-coordinated iron is
separated from that of penta-coordinated iron [12]. The
octahedral/tetrahedral ratio increases with Ce content (Fig.
6 and Table 2), similarly to what has been reported in the
literature for La-doped strontium ferrates [10]. From Table
2, by comparing the results for fresh and He-treated
samples, it is easy to deduce that Fe3+ with coordination
the octahedral site of Fe3+ in the fresh samples, leads to
affirm that cerium also determines the fraction of
octahedral Fe3+ present in the system after He treatment.
From these observations and from the previously discussed
XRD results, it may be concluded that when the octahedral
Fe3+ concentration in the fresh samples reaches 35 mol%,
the structure of the fresh samples is cubic and a
brownmillerite-type phase formation is surely hindered,
at the condition that cerium is present as dopant to
stabilize the cubic structure (Table 2). Cerium effect on Ce-
doped strontium ferrates could be mainly ascribed to an
interaction between the redox couples Ce4+–Ce3+ and
Fe4+–Fe3+ in the solid solution, which are in contact
through oxygen bridges. It is worth noticing that in La-
doped SrFeO3, where La possess only one oxidation state
(La3+), brownmillerite-type structure has been reported to
form under low oxygen pressure thermal treatment [10],
lower than 6 probably originates from reduction of Fe4+
which gives rise to around 80% of Fe3+ with coordination
lower than 6 and to around 20% of octahedral Fe3+
,
,
whereas octahedral iron was already present in the fresh
samples and now is incremented of around 20%. This fact,
according to the original hypothesis that cerium stabilizes

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3415
Fig. 6. Fits of Mossbauer spectra of (a) SFC6 and (b) SFC15 samples after He treatment: the sextet at higher magnetic field is due to octahedral Fe3+; the
¨
sextet at lower magnetic field is due to Fe3+ in a lower coordination.
SF
SFC6
SFC15
70
60
50
40
30
From Mössbuer
From TPR
20
10
0
2
4
6
8
10 12
14
100 200 300 400 500 600 700 800 900
T/˚C
(b)
(a)
Ce content (%)
SF
SFC6
SFC15
100 200 300 400 500 600 700 800 900
T/˚C
(c)
Fig. 7. TPR profiles for some representative Sr1–aCeaFeO3–x samples: (a) TPR (1st cycle); (b) linear correlation vs. cerium content of the Fe4+ mol%
deduced from the first TPR peak compared with Fe4+ mol% deduced from Mossbauer; (c) TPR (2nd cycle) after reoxidation at 1000 1C.
¨
whereas other cations, like praseodymium and bismuth,
that can change quite easily between two oxidation states,
behaves quite similarly to cerium [4,12]. We may exclude
that formation of cubic expanded phases depends on the
ionic radius of the dopant, since Ce3+ and La3+ have very
similar ionic radius, and, on the other side, Bi3+, which is
able to form cubic expanded phases, has lower ionic radius
than Ce3+ [26]. From another point of view, we believe
also that the excess of oxygen atoms in the crystal
structure, caused by Ce doping through the oxygen
vacancy suppression, hinders the formation of ordered
brownmillerite-type phase, as other authors observed for
similar systems [36].
From TPR profiles, information has been obtained on
the He-treated or partially treated samples: the first TPR
has been acquired after an oxidation pre-treatment at
500 1C, followed by a cooling under inert gas flow, whereas
the second TPR has been acquired after reoxidation at
1000 1C and He-cooling to RT. TPR profiles show two
main groups of peaks (Fig. 7a): the first group (low-T
peaks) between 300 and 500 1C and the second one (high T
peaks) between 700 and 1000 1C.
According to the literature [32], the low-T peaks
represent the H2 amount required to reduce all the Fe4+
ions to Fe3+ ions in two consecutive steps. In particular,
after the first step, a mixed-valence perovskite-type

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3416
structure is formed, whereas at the end of the second step, a
brownmillerite-type structure is obtained for the undoped
strontium ferrate [32]. According to our results on He-
treated samples (Table 4 and Fig. 5a,b,d), it is reasonable
to infer that during the TPR experiment, at the end of the
low-T peaks, brownmillerite-type structure forms only for
SF and SFC2 samples, while for the other Ce-doped
samples the cubic expanded phases is obtained. Moreover,
from Fig. 7a and Table 5, we observe that the first TPR
peak’s area, decreases with Ce doping. From the first TPR
peak’s area, it is easy to calculate the correspondent Fe4+
percentage, which results in linear correlation with Ce
content (Fig. 7b). The trend depicted in Fig. 7b is in
similarly to the bulk CeO2 [39]. TPR profiles of SFC6 and
SFC15 samples show the appearance of some small peaks
between 500 and 700 1C, probably due to some easily
available oxygen, created by lattice perturbation for Ce
doping, as described in literature for non-stoichiometric
LaFeO3 [40].
The high-temperature TPR peaks (700–1000 1C) repre-
sent the H2 amount consumed in the reduction of Fe3+ to
Fe2+ and eventually to Fe0. For pure SF and SFC2
samples, TPR profiles are very similar to those in the
literature, but for the other Ce-doped strontium ferrates,
some significant differences have been observed. In SFC6
and SFC15 (Fig. 7a), the peak around 700 1C disappeared,
indicating that it is probably related with the Fe3+
reduction in brownmillerite-type structure to Fe2+. More-
over the H2 uptake (mL/g) in the second TPR peak group
should be equal to the theoretical value necessary to reduce
all the Fe3+ to Fe0 (3 electron process ¼ 3 Â 59 mL H2/g),
but the observed value is considerably lower (Table 5),
since the reduction does not complete at 1000 1C (Fig. 7a),
as already found for LaFeO3 [41].
The second TPR cycle, which occurs after a reoxidation
at 1000 1C followed by cooling from 1000 1C to RT under
inert gas flow, shows some different features (Fig. 7c): the
low TPR peaks almost disappear, indicating that only
some residual Fe4+ remains, as found in the literature for
LaFeO3, which should contain only Fe3+ [40,41]. The H2
volume consumed in the second TPR peak group remains
unchanged from the first TPR to the second TPR cycle
(Fig. 7a,c and Table 5), demonstrating that the reduction
process was totally reversible.
The TPO experiments after TPD give us an indication on
the reoxidation of the phase (brownmillerite-type or
expanded) which is formed after TPD as well as on the
oxygen adsorption capacity of the investigated samples. In
Fig. 8a, the TPO curve of a representative sample (SFC6) is
shown. The reoxidation peak starts at 100 1C and is
complete at 350 1C. The TPD–TPO process is totally
reversible. In Fig. 8b, the mL O2 consumed in the TPO
process have been correlated with the Ce content: oxygen
consumption increases from 0 until 6 mol% (17 mL O2/g),
but then decreases again for higher Ce content. This trend
can be explained by the concurrence of two effects: the
decrease of oxygen vacancies with increasing Ce content
for the formation of the cubic expanded disordered phase,
and the increased oxygen mobility due to stabilization of
Fe3+ in the octahedral coordination. When the oxygen
vacancies are decreased too much, like in SFC15, the
advantage of the geometric effect of iron coordination
cannot compensate any more the decrease of vacancies,
and the interaction with oxygen decreases consequently.
Previous studies by Trofimenko and Ullmann [14] demon-
strated similar trend for Ce-doped SrFe0.8Co0.2O3 in
electronic conductivity which decreased for Ce contents
higher than 5 mol%, suggesting that the oxygen consump-
tion in these materials could be controlled by the electronic
conductivity of the system.
agreement with the previously shown Mossbauer observa-
¨
tions (Table 2) that Fe4+ decreases with Ce doping.
Nevertheless the Fe4+ values in SF, SFC6 and SFC15
calculated from the first TPR peak are all systematically
lower than the values calculated from Mossbauer measure-
¨
ments (Fig. 6b). If strontium ferrates would contain only
Fe4+, the theoretical amount of H2 should correspond to
around 59 mL H2/g (equivalent to the reduction of 1 g of
SrFeO3 by a 1 electron-process). It is known, however,
from the literature [37,38] and confirmed by our Moss-
¨
bauer results (Table 2) that even the undoped strontium
ferrate, Sr8Fe8O23, contains around 70% of Fe4+. Hence,
we should expect for SF sample that at least a 70% fraction
of 59 mL H2/g (43 mL/g) has been consumed in the first
TPR peak. It is quite clear that some iron reduction already
occurred during the pretreatment when, after the oxidation
treatment at 500 1C, the sample has then been cooled down
under inert gas flow from 500 1C to RT. Although we have
already shown that our samples are reduced under He flow
at 1000 1C, there is some evidence in the literature that
evolution of oxygen starts already at around 400 1C [32].
Finally, the reduction temperature of the first TPR peak
decreases from about 440 to 350 1C for increasing Ce
content (Table 5 and Fig. 7a). We exclude in the first TPR
peak a contribution of Ce4+ which reduces to Ce3+. This
reduction will probably occur at higher temperatures,
Table 5
H2 uptake (715%) consumed in the TPR experiment (1st and 2nd cycle)
by Sr1ꢀaCeaFeO3ꢀx samples with different Ce content
1st TPR cycle
ml H2/g low-T
peaks
1st TPR cycle
ml H2/g high-T
peaks
2nd TPR cycle
a
ml H2/g high-
T peaks
Sample name
SF
SFC2
23 (437 1C)
20 (415 1C)
19 (403 1C)
15 (375 1C)
12 (365 1C)
12 (346 1C)
52
46
32
32
43
43
52
46
34
33
43
45
SFC6
SFC10
SFC15
SFC25
Temperature values in brackets refer to the maximum of the indicated
reduction peak.
aThe low-T peaks of the 2nd TPR cycle are all around 5 ml H2/g and
have not been reported in the table.

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3417
18
SFC6
17
16
15
14
13
12
11
10
9
8
0
2
4
6
8
10 12 14 16
100 200 300 400 500 600 700 800 900 1000
(a)
°
(b)
Ce content (mol %)
Temperature ( C)
Fig. 8. (a)TPO after TPD curve for SFC6 as a representative Sr1–aCeaFeO3–x sample. (b) O2 consumption values in ml O2/g (from TPO after TPD) vs. Ce
content. The black or white triangles refer to the two different TPD–TPO cycles.
Fig. 9. Fits of Mossbauer spectrum of (a) SF and (b) SFC15 samples after H2 treatment: the sharper sextet is metallic a-Fe; the other two sub-spectra are
¨
from Fe3+ (less sharp sextet) and from Fe2+(central sub-spectrum).
3.3. H2-treated powders
consider it as the result of the interaction, under reducing
conditions, of SrFeO3 with the 1 wt% segregated CeO2
phase. This hypothesis is supported by the experimental
observations that in SFC15, the CeO2 XRD peak
disappeared after reduction and appeared again after
reoxidation (data not shown) and that in SF this additional
The phase composition of an undoped and a Ce-doped
samples after TPR (by cooling the sample under He) has
been investigated in details by XRD, Mossbauer and XPS
¨
measurements. XRD analysis shows that the main phase
composition of SF and SFC15 after TPR consists of a
Ruddlesden Popper (RP) phase of the type Sr3Fe2O6 and
some metallic iron (Fig. 5c and Table 4bis). Although some
authors found that for undoped strontium ferrate only
FeO and SrO and finally metallic iron are formed at the
end of the reduction process [32], in this work, in the
reduced SF sample, the RP phase was detected, in
agreement with other authors [42]. RP phases derived
from perovskite-type structure and are formed by alter-
nating layers of ABO3ꢀx and oxide of the A cation [34]. In
the SF sample, the cell parameter of the RP phase (Table
4bis) is in agreement with the literature [43]. In SFC15, the
presence of cerium as a dopant could be responsible for the
slightly larger cell parameters (Table 4bis) with respect to
RP phase is absent (Table 4bis). Mossbauer results seem to
¨
substantially confirm XRD results: SF and SFC15 contain
mainly Fe3+, which is very probably square pyramidal [43],
and metallic iron. From Mossbauer results, some Fe2+ was
¨
also present (Table 2 and Fig. 9), probably as FeO. Since
FeO was not evidenced by XRD, it is probably amorphous
or very well dispersed. Nevertheless some differences are
evident by comparing the two spectra in Fig. 9a and b. In
SFC15, quite large G values suggest a larger distortion in
the structure symmetry. This perturbation could be
provoked by the introduction of cerium in the RP phase.
Regarding metallic a-iron, a quantitative comparison
between the two techniques was not possible due to the
phase separation occurring after the reduction, which
˚
˚
˚
the literature (3.901 A vs. 3.894 A for a0 and b0; 20.096 A
misleads Mossbauer results overestimating metallic iron.
¨
˚
vs. 20.039 A for c0). Few percent of an additional RP phase
are also present in SFC15 sample, as it is demonstrated by
XRD analysis in Table 4bis, and it is reasonable to
The XPS analysis performed on the SFC15 sample after
TPR evidences the presence of metallic iron on the surface.
Fe 2p spectra in the SFC15 sample, both before and after

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3418
Fe 2p3/2
Fe 2p1/2
(c)
Fe0
(b)
Fig. 11. Log (ASR) vs. 1000/T(K) for SF, SFC6 and SFC15 taken from
preliminary E IS experiments.
(a)
4. Conclusions
In this work, structure–properties relationships of Ce-
doped SrFeO3 prepared by citrate–nitrate auto-combustion
synthesis have been investigated to evaluate the effect of
Ce-doping in the structure, phase stability, reducibility and
interaction of these materials with oxygen.
700
705
710
715
720
725
730
735
740
Binding Energy (eV)
Our results show that citrate–nitrate smoldering auto-
combustion method is effective in obtaining single-phase
Sr1ꢀaCeaFeO3ꢀx (0pao0.15) powders, with a surface area
of ffi20 m2/g and a crystallite size of ffi100 nm. Structural
and reduction properties of cerium-doped strontium
ferrates are greatly improved by Ce substitution and are
highly dependent on cerium concentration in the solid
solution. The Sr1ꢀaCeaFeO3ꢀx structure is cubic only for
aX0.06, while for ao0.06 remains tetragonal. Moreover,
for aX0.06 after semi-reductive treatment under inert gas,
an expanded cubic phase is obtained instead of the less
conducting brownmillerite-type structure. The improved
properties are mainly correlated with the stabilization of
octahedral Fe3+ site and cubic environment operated by
cerium doping. This stabilization occurs probably through
an interaction between the redox couples Ce4+–Ce3+ and
Fe4+–Fe3+ connected in the solid solution by oxygen
bridges. In the presence of cerium, when the octahedral
iron concentration in the fresh samples is around 35%,
expanded cubic phases are favored after He treatment at
high temperatures and brownmillerite-type phase forma-
tion is hindered. The reduction temperature of Fe4+ to
Fe3+ in Ce-doped SrFeO3 decreases with Ce doping.
During the reduction process of Ce-doped SrFeO3 at high
temperatures, formation of RP phase occurs, where cerium
is inserted in the RP structure.
Fig. 10. XPS spectra in the Fe 2p region: curve (a) the XPS spectrum of
Fe2O3 taken as reference; curve (b) SFC15 sample before and curve (c)
after TPR (H2 treatment).
TPR (H2 treatment), are shown respectively in the curve b
and curve c of Fig. 10. The only difference in these Fe 2p
spectra is represented by the small shoulder at the low BE
side of the spectrum in curve c. This feature can be fitted by
a Fe 2p3/2 component at BE ¼ 706.5 eV, which can be
assigned to Fe0 species [30,31], as also confirmed by the Fe
2p1/2 doublet component indicated by an arrow in curve c.
The surface modifications underwent by the SFC15 sample
upon the TPR treatment are reported in Table 3, showing
an higher Sr/Fe ratio due to a higher Sr content in the
surface, which indicates the presence of Sr3Fe2O6 on the
surface.
3.4. Preliminary electrochemical impedance results
In Fig. 11, the Log (ASR) vs. 1000/T(K) has been
reported for 3 different Ce content (0, 6 and 15 mol%). The
highest ASR has been measured for SF, the lowest one for
SFC6, whereas SFC15 lays in between. From this trend
and from the TPO trend of Fig. 8b, we may tentatively
conclude that the oxygen adsorption could be considered
as the rate determining step of the oxygen reduction
process in Ce-doped SrFeO3.
The best results in TPO (after TPD) experiments have
been obtained for a ¼ 0.06, whereas for higher Ce content,
the beneficial geometric effect of octahedral iron cannot

ARTICLE IN PRESS
F. Deganello et al. / Journal of Solid State Chemistry 179 (2006) 3406–3419
3419
compensate the gradual suppression of oxygen vacancy
concentration induced by Ce. Preliminary EIS results
showed that the lowest ASR has been obtained for the
same cerium content (6 mol%). Oxygen adsorption has
been identified as the rate determining factor of the overall
cathode process in Ce-doped strontium ferrates. All these
findings are important for the application of Ce-doped
SrFeO3 materials in low-temperature oxygen-conducting
devices. Further experiments on the electrochemical
properties of these materials are in progress.
[15] N.E. Trofimenko, H. Ullmann, J. Paulsen, R. Mueller, Solid State
Ionics 99 (1997) 201–214.
[16] R.N. Basu, F. Tietz, E. Wessel, H.P. Buchkremer, D. Stover, Mater.
Res. Bull. 39 (2004) 1335–1345.
[17] S.B. Adler, Chem. Rev. 104 (2004) 4791–4843.
[18] F. Deganello, G. Deganello, unpublished work.
[19] A.C. Larson, R.B. Von Dreele, LANSCE, MS-H 805, Los Alamos
National Laboratory, Los Alamos, NM 87545, USA, 1998.
[20] L.B. McCusker, R.B. Von Dreele, D.E. Coxd, D. Louer, P. Scardi,
J. Appl. Crystallogr. 32 (1999) 36–50.
[21] Y.M. Zhao, M. Hervieu, N. Nguyen, B. Raveau, J. Solid State Chem.
153 (2000) 140–144.
[22] M.P. Seah, in: X. Briggs and X. Seah (Eds.), Practical Surface
Analysis, second ed., vol. 1, Wiley, Chichester, 1990, p. 202.
[23] C.D. Wagner, L.E. Davis, W.M. Riggs, Surf. Interface Anal. 2 (1980)
53–55.
Acknowledgments
The authors are thankful to Prof. G. Deganello for
important suggestions. Moreover the Italian Ministry of
Foreign Affairs (MAE), under the frame of the Italy–Japan
Joint Lab on ‘‘Nanostructured Materials for Environment
and Energy’’ (NaMaTEE), is greatly acknowledged for
supporting the preliminary impedance tests.
[24] S. Brunauer, P.H. Emmet, E. Teller, J. Am. Chem. Soc. 60 (1938)
309–319.
[25] E.P. Barret, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951)
373–380.
[26] J.P. Hodges, S. Short, J.D. Jorgensen, X. Xiong, B. Dabrowski, S.M.
Mini, C.W. Kimball, J. Solid State Chem. 151 (2000) 190–209.
[27] J.-F. Huang, X. Ni, J.-C. Bao, J.-H. Wu, Jiegon Huaxue 13 (1994)
350–359.
[28] R.D. Shannon, Acta Crystallogr. A 32 (1976) 751–767.
[29] P.K. Gallagher, J.B. MacChesney, D.N.E. Buchanan, J. Chem. Phys.
41 (1964) 2429–2434;
References
[1] C. Liang, D. Yang, Z. Yang, F. Hou, M. Xu, Surf. Coatings Technol.
200 (2005) 2515–2517.
L. Fournes, Y. Potin, J.C. Grenier, G. Demazeau, M. Pouchard,
Solid State Commun. 62 (1987) 239–244;
[2] V.P.S. Awana, J. Nakamura, J. Linde
´
n, M. Karpinnen, H.
M. Takano, N. Okita, N. Nakayama, Y. Bando, Y. Takeda, O.
Yamamoto, J.B. Goodenough, J. Solid State Chem. 73 (1988)
140–150.
Yamauchi, Solid State Commun, 119 (2001) 159–162.
[3] Y.M. Zhao, X.J. Yang, Y.F. Zheng, D.L. Li, S.Y. Chen, Solid State
Commun. 115 (2000) 365–368.
[30] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, in: J. Chastain
Jr., R.C. King (Eds.), Handbook of X-ray Photoelectron Spectro-
scopy, Phys. Electronics Inc., Eden Prairie, USA, 1995.
[31] P.A.W. Van der Heide, Surf. Interface Anal. 33 (2002) 414–425.
[32] H. Falcon, J.A. Barbero, J.A. Alonso, M.J. Martinez-Lope, J.L.G.
Fierro, Chem. Mater. 14 (2002) 2325–2333;
[4] S. Zongping, Y. Cong, G. Xiong, S. Sheng, W. Yang, Sci. China
(Series B) 43 (2000) 421–427.
[5] M.T. Colomer, B.C.H. Steele, J.A. Kilner, Solid State Ionics 147
(2002) 41–48.
[6] H. Ullmann, N.E. Trofimenko, F. Tiez, D. Stover, A. Ahmad-
¨
Khanlou, Solid State Ionics 138 (2000) 79–90.
[7] A. Mai, V.A.C. Haanappel, S. Uhlenbruck, F. Tietz, D. Stover, Solid
¨
C.-T. Wang, S.-H. Ro, J. Non-Cryst. Solids 352 (2006) 35–43.
[33] A.P. Grosvenor, B.A. Kobe, M.C. Biesinger, N.S. McIntyre, Surf.
Interface Anal. 36 (2004) 1564–1574.
State Ionics 176 (2005) 1341ꢀ1350.
[8] A.A. Leontiou, A.K. Ladavos, T.V. Bakas, T.C. Vaimakis, P.J.
Pomonis, Appl. Catal. A: Gen. 241 (2003) 143–154.
[9] S. Shin, M. Yonemura, H. Ikawa, Mater. Res. Bull. 13 (1978)
1017–1021.
[34] M.V. Patrakeev, I.A. Leonidov, V.L. Kozhevnikov, V.V. Kharton,
Solid State Sci. 6 (2004) 907–913.
[35] C. Haavik, T. Atake, H. Kawaji, S. Stolen, Phys. Chem. Chem. Phys.
3 (2001) 3863–3870.
[10] M.V. Patrakeev, J.A. Bahteeva, E.B. Mitberg, I.A. Leonidov, V.L.
Kozhevnikov, K.R. Poeppelmeier, J. Solid State Chem. 172 (2003)
219–231.
[36] F. Prado, N. Grunbaum, A. Caneiro, A. Manthiram, Solid State
Ionics 167 (2004) 147–154.
[37] K.V. Pokholok, I.A. Presnyakov, V.A. Ketsko, N.N. Oleinikov, N.T.
Kuznetsov, Russ. J. Coord. Chem. 27 (2001) 632–635.
[38] G. Shirane, E.D. Cox, S.L. Ruby, Phys. Rev. 125 (1962) 1158–1165.
[39] A.M. Venezia, G. Pantaleo, A. Longo, G. Di Carlo, M.P. Casaletto,
L.F. Liotta, G. Deganello, J. Phys. Chem. B 109 (2005) 2821–2827.
[40] R. Spinicci, A. Tofanari, A. Del Mastro, D. Mazza, S. Ronchetti,
Mater. Chem. Phys. 76 (2002) 20–25.
[11] S.E. Dann, D.B. Currie, M.T. Weller, M.F. Thomas, A.D. Al-
Rawwas, J. Solid State Chem. 109 (1994) 134–144.
[12] M. Stange, J. linden, A. kjekshus, N. binsted, M.t. weller, B.c.
hauback, H. fjellvag, J. Solid State Chem. 173 (2003) 148–163.
[13] Y.D. Tretyakov, V.V. Sorokin, A.R. Kaul, A.P. Erastova, J. Solid
State Chem. 18 (1976) 253–261;
H.N. Pandya, R.G. Kulkarni, P.H. Parsania, Mater. Res. Bull. 25
(1990) 1073–1077;
[41] F.J. Berry, J.R. Gancedo, J.F. Marco, X. Ren, J. Solid State Chem.
177 (2004) 2101–2114.
M. Robbins, G.K. Wertheim, A. Menth, R.C. Sherwood, J. Phys.
Chem. Solids 30 (1969) 1823–1825.
[42] M. Wyss, A. Reller, H.R. Oswald, Solid State Ionics 101–103 (1997)
547–554.
[14] N.E. Trofimenko, H. Ullmann, J. Eur. Ceram. Soc. 20 (2000)
1241–1250.
[43] S.E. Dann, M.T. Weller, D.B. Currie, J. Solid State Chem. 97 (1992)
179–185.