Applied Catalysis B: Environmental 65 (2006) 21–30

www.elsevier.com/locate/apcatb

Synthesis and characterisation of La1
xCaxFeO3 perovskite-type oxide

catalysts for total oxidation of volatile organic compounds
Bibiana P. Barbero a,1, Julio Andrade Gamboa b, Luis E. Cadús a,*
a
Instituto de Investigaciones en Tecnologı́a Quı́mica (INTEQUI), UNSL, CONICET, Casilla de Correo 290, 5700 San Luis, Argentina
b
Comisión de Energı́a Atómica (CNEA), 8400 San Carlos de Bariloche, Rı́o Negro, Argentina
Received 21 May 2005; received in revised form 28 September 2005; accepted 21 November 2005
Available online 2 February 2006

Abstract
La1
xCaxFeO3 perovskite-type oxides with x = 0, 0.2 and 0.4 were prepared by the citrate method and characterised by means of X-ray
diffraction (XRD), X-ray fluorescence (XRF), surface area measurement BET, X-ray photoelectron spectroscopy (XPS), Fourier transformed
infrared spectroscopy (FT-IR), laser Raman spectroscopy (LRS), oxygen temperature-programmed desorption (O2-TPD) and temperature-
programmed reduction (TPR). The citrate method shows to be simple and appropriate to obtain single phases avoiding segregation and/or
contamination. Moreover, controlling the calcination temperature, specific surface areas adequate for catalysts to be used in oxidation reactions are
achieved. The structure refinement by using the Rietveld method indicates that the partial calcium substitution modifies the orthorhombic structure
of the LaFeO3 perovskite towards a less distorted one. From XRF and XPS, a slight surface enrichment in lanthanum and calcium was detected.
XRD, FT-IR and TPR results indicated that the electronic debalance caused by the partial substitution for La3+ by Ca2+ is compensated by an
oxidation state increase of a part of Fe3+ to Fe4+. O2-TPD results revealed that at a substitution level higher than x = 0.2, oxygen vacancies are also
formed to preserve the electroneutrality. Finally, an improvement of the catalytic activity in propane and ethanol combustion was observed on the
substituted perovskites. Correlating this with the characterisation results, the active sites would be associated to the Fe4+ ions.
# 2006 Elsevier B.V. All rights reserved.

Keywords: Lanthanum iron perovskite; Calcium; XRD; XRF; XPS; FT-IR; Raman spectroscopy; O2-TPD; TPR; Rietveld refinement; Propane combustion;
Ethanol combustion

1. Introduction
An increasing interest has been shown in catalytic
combustion processes during the last decades since they are
a convenient way for emission prevention (the control of
nitrogen oxides NOx and unburned hydrocarbons in heat and
power generation plants) as well as clean-up (volatile organic
compounds removal, automobile exhaust converters) [1].
Perovskite-type oxides display prominent catalytic activities
in many fields such as the total oxidation of methane and of
volatile organic compounds [2]. This activity, coupled with a
high thermal stability, postulates to pervoskite-type oxides as
potential catalysts in substitution of very active noble metals such
* Corresponding author. Fax: +54 2652 426711.
E-mail addresses: bbarbero@unsl.edu.ar (B.P. Barbero),
lcadus@unsl.edu.ar (L.E. Cadús).
1
Fax: +54 2652 426711.
as Pt and Pd, which are more expensive and do not resist
operating at high temperatures [1–3]. Perovskite oxides have
general formula ABO3, where the 12-coordinated A sites may be
occupied by rare-earth, alkaline-earth, alkali or other large ions
and the 6-coordinated B sites are usually filled with transition
metal cations. A large number of metallic cations can occupy the
A and the B sites. Furthermore, the great stability of the
perovskite framework allows partial substitution at the A sites
and/or the B sites modifying the catalytic, redox and structural
properties. The substitution at A site with ions having lower
valence can allow the formation of structural defects such as
anionic or cationic vacancies and/or a change in the oxidation
state of the transition metal cation to maintain the electro-
neutrality of the compound. When the oxidation state of B cation
increases, the relative ease of the redox process generates larger
quantities of available oxygen at low temperature and the overall
oxidation activity enhances. Moreover, the oxygen vacancies
favour the catalytic activity in oxidation reaction because they
increase the lattice oxygen mobility.

0926-3373/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcatb.2005.11.018
22 B.P. Barbero et al. / Applied Catalysis B: Environmental 65 (2006) 21–30
Perovskites with lanthanum in A site and transition metals in
B site (LaBO3) with B = Mn, Co, Fe, Ni have received much
attention. Partial substitution of divalent ions, in particular Sr2+,
for lanthanum has been studied producing interesting results.
For example, La1
xSrxFeO3 perovskites with x = 0.1 and 0.2
are more active in propane oxidation than the corresponding
unsubstituted LaFeO3 [4]. The calcium substitution in A site
has been studied from the structural viewpoint [5–7] but few
results about the catalytic applications have been reported.
Ciambelli et al. [8] investigated a La1
xCaxFeO3 series as
catalysts for methane combustion and they observed that the
catalytic performance of the substituted samples has not been
better than that of unsubstituted sample. In contrast, we have
studied the effect of calcium substitution in La1
xCaxCoO3
perovskites for propane oxidation and it has been beneficial [9].
On the basis of these results, in this work we reported a
study carried out on La1
xCaxFeO3 perovskites (x = 0, 0.2, 0.4)
which have been evaluated in total oxidation of propane
and ethanol. The aspects related to the synthesis process have
been carefully controlled. The structural and redox features
are determined from the results of the textural and physico-
chemical characterisation. Furthermore, an explication of the
catalytic behaviour is proposed.
2. Experimental
2.1. Catalyst preparation
La1
xCaxFeO3 perovskites with x = 0, 0.2, and 0.4 were
prepared by the citrate method [10]. La(NO3)36H2O (Fluka),
Ca(NO3)24H2O (Fluka), Fe(NO3)39H2O (Aldrich) and citric
acid (Mallinckrodt) were used as reagents. The aqueous
solutions of the metal nitrates were added to an aqueous
solution of citric acid with a 10% excess over the number of
ionic equivalents of cations. The resulting solution was agitated
for 15 min and concentrated slowly by evaporating water under
vacuum in a rotavapor at 75 8C until the formation of a gel. This
gel was dried in an oven, increasing slowly the temperature up
to 250 8C and keeping it overnight, to yield a solid amorphous
citrate precursor. The obtained precursor was milled and then
decomposed in air at 400 8C for 30 min and calcined in air at
700 8C for 2 h. The perovskite catalysts are expressed by
La1
xCaxFeO3 where x = 0, 0.2, and 0.4, although the actual
composition may be non-stoichiometric in relationship to the
oxygen and may not be of single phase.
2.2. Catalyst characterisation
2.2.1. Specific surface area
The specific surface area (SSA) of the catalysts was
calculated by the BET method from the nitrogen adsorption
isotherms obtained at 77 K on samples outgassed at 250 8C
using a Micromeritics Accusorb 2100E apparatus.
2.2.2. X-ray diffractometry (XRD)
XRD patterns were recorded at room temperature by using a
Rigaku diffractometer operated at 30 kV and 20 mA, employ-
ing Ni-filtered Cu Ka radiation (l = 0.15418 nm). The data
were collected at 0.028 with a counting time of 5 s per step, in
the (2u) range from 208 to 908. The crystalline phases were
identified by reference to the PDF data employing standard
spectra software. The lattice parameters and the structure of the
catalysts have been estimated from Rietveld’s powder structure
refinement analysis [11].
2.2.3. X-ray fluorescence (XRF)
The composition of every sample was determined by X-ray
fluorescence on a PW1400 Philips instrument. The calibration
was made by using calibration samples prepared with known
amounts of La2O3, Fe2O3 and CaCO3.
2.2.4. X-ray photoelectron spectroscopy (XPS)
XPS analyses were performed on a SSX 100/206
photoelectron spectrometer from Surface Science Instruments
with a monochromatized microfocused Al X-ray source.
Spectra were registered after purging the samples at ambient
temperature in vacuum. The residual pressure in the analysis
chamber during the analysis was about 10
6 Pa. The flood gun
energy was adjusted at 10 eV. The data treatment was
performed with appropriate software. Binding energies were
calibrated with regard to the C(C, H) component of the C 1s
peak fixed at 284.8 eV. The atomic ratios were calculated
by using the atomic sensitivity factors provided by the
manufacturer.
2.2.5. Raman spectroscopy
The Raman spectra were collected by using two different
spectrometers: (a) JASCO TRS600SZP multichannel mono-
chromatic spectrometer, and (b) Dilor LabRam Instruments. In
both cases, the samples were pressed into self-supporting
wafers and the spectra were recorded at ambient temperature.
The excitation source was: (a) an Ar ion laser (514.5 nm)
operated at a power of 50 mW, and (b) a He–Ne laser
(632.8 nm) operated at a power of 10 mW.
2.2.6. FT-IR spectroscopy
FT-IR spectra were registered by using a Nicolet Protegé 460
spectrometer in KBr pellets. The spectra were the result of
averaging out 32 scans obtained at ambient temperature in
wavelength ranging from 4000 to 225 cm
1.
2.2.7. Oxygen temperature-programmed desorption
(O2-TPD)
O2-TPD experiments were performed in a quartz reactor
using a TCD as detector. In each analysis, 500 mg samples were
pre-treated with helium gas increasing the temperature from
ambient temperature up to 700 8C at 10 8C min
1. The samples
were oxidised with a 20% O2/He mixture at a total flow rate of
30 ml min
1 at 700 8C for 30 min. Then, they were cooled
down to ambient temperature in the oxidising mixture and
flushed by a stream of purified He for 30 min. The desorption
was carried out in the same conditions as the pre-treatment,
maintaining the temperature at 700 8C until the baseline of the
chromatograph was stabilised.
 B.P. Barbero et al. / Applied Catalysis B: Environmental 65 (2006) 21–30 23

Table 1
Specific surface areas, elemental composition from XRF (nominal values in
parentheses), and Fe4+ amount from the TPR results
Catalyst SBET (m2/g) La (%) Fe (%) Ca (%) Fe4+ (%)
LaFeO3 18.5 61.8 (57.2) 20.2 (23.0) 0 (0) 2.8
La0.8Ca0.2FeO3 13.0 56.1 (49.8) 21.9 (25.0) 3.7 (3.6) 15.9
La0.6Ca0.4FeO3 16.7 47.5 (41.0) 24.0 (27.5) 6.1 (7.9) 21.5

2.2.8. Temperature-programmed reduction (TPR)

TPR experiments were performed in the same apparatus
used for O2-TPD on the samples after the O2-TPD. The
reducing atmosphere was a 5% H2/N2 mixture at a total flow
rate of 30 ml min
1. The temperature was increased at Fig. 1. XRD results of La1
xCaxFeO3 perovskites and PDF data.
10 8C min
1 from ambient temperature to 700 8C. The
calibration of the TCD was performed by injection of know
amounts of N2 in the H2/N2 flow, thus simulating H2 characteristic of orthorhombic symmetry (PDF 37-1493), space
consumption. group Pnma (#62). The lattice parameters estimated by the
Rietveld method are a = 5.564 Å, b = 7.849 Å, and c = 5.551 Å
2.3. Catalytic test (Table 2). The X-ray patterns of the substituted perovskites are
also monophasic and they almost show the same lines that
The catalysts (300 mg, 0.5–0.8 mm particle diameter) LaFeO3, although a XRD line shift to higher 2u angle values
diluted with glass particles of the same size in a ratio 1:5 when the calcium amount increases is observed. The lattice
were tested in a fixed-bed tubular reactor operated at parameters for the substituted samples are presented in Table 2.
atmospheric pressure. The temperature, measured with a The structure was refined assuming a rhombohedral symmetry,
coaxial thermocouple, was increased from 150 to 200 8C to space group R3c (#167).
attain total conversion of VOC (propane or ethanol). The feed The unit cell volume (normalised to account for the
was a mixture of 2:20:78 for C3H8:O2:He or 1:20.8:78.2 for difference in Z) decreases with the increase of the calcium
C2H5OH:O2:He. The total flow rate was 100 ml min
1 amount (Table 2).
measured at ambient temperature. The reactants and reaction
products were alternately analysed on-line by gas chromato- 3.3. Bulk composition: X-ray fluorescence
graphy. The conversion of VOC, X%, is defined as the
percentage of VOC feed that has reacted, i.e.: XRF results are shown in Table 1. The experimentally
determined lanthanum content is higher than the expected one
ðVOCÞin
ðVOCÞout while iron amount is approx. 13% lower than the theoretical
X ð%Þ ¼  100
ðVOCÞin one in all the cases. Calcium amount detected in La0.8Ca0.2-
FeO3 is in agreement with the theoretical value but in
The specific activity (conversion/m2) is calculated as the VOC La0.6Ca0.4FeO3 the experimental result is quite lower than the
conversion per surface area unit. theoretical value.
Data obtained under differential conditions with propane
conversions below 10% were used in the calculus of the
reaction rates. The apparent activation energies were estimated Table 2
from Arrhenius type plots (ln r versus 1/T). Lattice parameters for La1
xCaxFeO3 perovskite oxides
Sample a (Å) b (Å) c (Å) a (8) V (Å3)/Z
3. Results
LaFeO3 5.564 7.849 5.551 – 61
La0.8Ca0.2FeO3 5.524 – – 60.1 60
3.1. Specific surface area La0.6Ca0.4FeO3 5.490 – – 60.2 59

Table 1 shows the specific surface areas measured by the

BET method. The specific surface areas of LaFeO3 and Table 3
La0.6Ca0.4FeO3 are similar within the experimental error XPS results
whereas that of La0.8Ca0.2FeO3 is lower. Binding energies (eV) Atomic ratios
La 3d5/2 Ca 2p Fe 2p La/Fe Ca/Fe Ca/(La + Ca)
3.2. X-ray diffraction
LaFeO3 833.8 709.9 2.22 0 0
La0.8Ca0.2FeO3 833.5 347.0 710.1 1.40 0.54 0.28
XRD results revealed the formation of perovskite-type
La0.6Ca0.4FeO3 833.2 346.4 710.0 1.05 0.83 0.44
single phases (Fig. 1). The diffractogram of LaFeO3 sample was
24 B.P. Barbero et al. / Applied Catalysis B: Environmental 65 (2006) 21–30
3.4. Surface composition: X-ray photoelectron
spectroscopy
Table 3 shows the XPS results. La 3d5/2 peak is detected at
833.5  0.3 eV. According to the reported [12–14], this
corresponds to La3+. Ca 2p peak is at 346.7  0.3 eV and it
is Ca2+ [13]. Fe 2p peak appears at 710.0  0.1 eV. This
binding energy would be characteristic of Fe3+ [13,14]. Binding
energy shifts due to calcium substitution are not observed.
(La + Ca)/Fe atomic ratios almost double the theoretical value.
Furthermore, Ca/(La + Ca) atomic ratio is slightly higher than
the theoretical one. This indicates a surface enrichment in
lanthanum and calcium.
3.5. Raman spectroscopy
Raman spectra of La1
xCaxFeO3 perovskites are shown in
Fig. 2a (spectra acquired with Jasco spectrometer) and Fig. 2b
(by using LabRam Instruments). The spectra obtained with
Jasco spectrometer (Fig. 2a) are not well defined. Very few
Raman spectra for LaFeO3 perovskite have been found in
literature. Some of them, similar to ours, have been reported by
Popa et al. [15,16] and Bontempi et al. [14]. According to their
assignments, the crystal symmetry of LaFeO3 is orthorhombic.
In the Raman spectra of the substituted samples, the band at
417 cm
1 disappears as well as the other weak ones and the
band at 635 cm
1 decreases and widens.
The spectra shown in Fig. 2b are very different from those in
Fig. 2a. LaFeO3 spectrum presents well-defined bands at
Fig. 2. Raman spectra acquired with (a) Jasco spectrometer and (b) LabRam
Instruments.
107 cm
1 (medium), 113 cm
1 (medium), 277 cm
1 (weak),
432 cm
1 (very strong), 455 cm
1 (shoulder), 497 cm
1
(medium), 608 cm
1 (medium), 711 cm
1 (strong),
739 cm
1 (weak), 918 cm
1 (very weak), and 1108 cm
1
(weak). Bontempi et al. [17] obtained a very similar spectrum
for LaNiO3 perovskite, which has rhombohedral structure. On
the other hand, our spectrum could also be compared with that
of La0.5Ca0.5MnO3 recorded at 8 K by Granado et al. [18] who
assigned it orthorhombic symmetry. The assignment to one or
other symmetry by comparison with the previously published
data is not straightforward, because of the different experi-
mental conditions (laser line, laser power, etc.) used in each
work. However, it is worth noting in the Raman spectra of the
substituted samples, the band intensities decrease when calcium
amount increases.
3.6. FT-IR spectroscopy
Fig. 3 presents the IR spectra of the La1
xCaxFeO3
perovskites. All of them show two strong and well-defined
absorption bands, typical of perovskite oxides ABO3 [19–21].
Because the Fe–O bonds of the BO6 octahedral units are
undoubtedly stronger than those of the 12-coordinated La(III)–
O units, one may predict that the BO6 units dominate the
spectroscopic behaviour [21]. The bands at higher wavenum-
bers (500–730 cm
1) are assigned to the stretching modes of
Fig. 3. FT-IR spectra of La1
xCaxFeO3 perovskites.
 B.P. Barbero et al. / Applied Catalysis B: Environmental 65 (2006) 21–30 25

the BO6 octahedra and those at lower wavenumbers (250–

500 cm
1), to the deformation modes of these same polyhedra
[19].
IR spectrum of LaFeO3 presents an intense band around
567 cm
1 with a shoulder at 623 cm
1. The main band can be
assigned to antisymmetric stretching vibration of the BO6
octahedra [20,21] whereas the shoulder could be assigned to the
symmetric stretching vibration of these octahedra. The bands in
the low-wavenumbers region (250–500 cm
1) correspond to
the deformational modes of the octahedra. Bands associated to
a La(III)–O stretching motion can also appear in this region
[21]. Couzi and Huong [20] reported a similar spectrum of
LaFeO3.
In the IR spectra of calcium-substituted samples, the same
main bands that those for LaFeO3 are found although a slight
shift towards higher wavenumbers when the calcium amount
increases is observed. Furthermore, the shoulders at 623 and
453 cm
1 are undetected in the substituted samples and a band
around 877 cm
1 appears whose intensity increases when the
calcium amount increases.

3.7. Temperature-programmed reduction

Fig. 4 shows the curves of the TPR experiments. TPR profile

of LaFeO3 shows two very weak signals with the maxima at
about 255 and 385 8C. Ciambelli et al. [8,22] and Spinicci et al.
[23] reported a similar curve. TPR profiles of substituted
samples show an intense reduction signal at about 355 8C and
Fig. 5. O2-TPD results.

another weak signal at 410 8C. The intensity of the first signal
increases when the calcium amount increases. A further
reduction starts at temperature above 620 8C and although it is
not completely resolved under the experimental conditions
used in this work, the signal intensity increases with the
calcium substitution.

3.8. Temperature-programmed desorption of oxygen

Fig. 5 shows the O2-TPD results. The curves present a first

peak between 150 and 350 8C whose maximum shifts towards
higher temperatures when the calcium amount increases.
Furthermore, a second signal above 400 8C is observed only for
the substituted samples. Nitadori and Misono [4] observed
curves with the same shape and trend in their O2-TPD results of
La1
xSrxFeO3 perovskites. The first signal is generally denoted
as a oxygen species and they correspond to oxygen species
adsorbed on surface oxygen vacancies. The second signal at
higher temperature is known as b oxygen species. These
oxygen species release from the oxide lattice [24–26].

3.9. Catalytic activity

The results are presented in Figs. 6–8 and Table 4. Catalytic

performance was evaluated using two different molecules:
propane and ethanol. The main products during the total
Fig. 4. TPR profiles of La1
xCaxFeO3 perovskites. oxidation of propane were CO2 and water. At reaction
26 B.P. Barbero et al. / Applied Catalysis B: Environmental 65 (2006) 21–30

Fig. 8. Specific activity (VOC conversion per surface area unit) as a function of
Fig. 6. Catalytic activity in total oxidation of propane (solid line, closed the reaction temperature. Identification as in Fig. 6.
symbols) and ethanol (dotted line, open symbols) on (^) LaFeO3, ( )
La0.8Ca0.2FeO3, and ( ) La0.6Ca0.4FeO3.

Table 4
Catalytic activity results in propane oxidation
Catalyst Tonseta (8C) T50b (8C) T95b (8C) Eac (kcal/mol) rwd (mmol g
1 h
1) rsd (mmol m
2 h
1)
LaFeO3 225 385 428 26.2 76.96 4.16
La0.8Ca0.2FeO3 220 392 445 22.9 88.93 6.84
La0.6Ca0.4FeO3 220 378 420 23.4 108.75 6.51
a
Temperature of reaction start.
b
Temperatures corresponding to 50% and 95% conversion, respectively.
c
Apparent activation energies.
d
Reaction rates per catalyst mass unit and per surface area unit, respectively, at reaction temperature = 300 8C.

temperatures higher than 370 8C, propylene was also detected
and above 420 8C, traces of CO and C2H4 were found. Other
oxygenated products such as acrolein, acrylic acid, acetaldehyde
were not found. The total oxidation of ethanol occurs through an
intermediate, acetaldehyde. In Fig. 7, ethanol conversion and
selectivities to acetaldehyde and CO2 as a function of the reaction
temperature on LaFeO3 are shown. The curves for the other
catalysts have the same shape that for LaFeO3. The carbon
balance was closed in 5% for both reactions.
Propane oxidation starts at around 220 8C on all the catalysts
while ethanol oxidation begins at lower reaction temperature
with the increase of the calcium amount (Fig. 6). At low
reaction temperatures, all the catalysts have a similar behaviour
during propane oxidation and when the reaction temperature
increases, different catalytic performance is observed with
different calcium amount.
Considering the reaction temperature corresponding to 50%
and 95% conversion of VOC (T50 and T95, Table 4), the activity
for propane oxidation increases in the order: La0.8Ca0.2-
FeO3 < LaFeO3 < La0.6Ca0.4FeO3 while for ethanol oxidation
the activity increases with increasing x. However, when the
VOC conversion per surface area unit (VOC conversion/m2) is
compared (Fig. 8), the substituted samples show performances
very similar among them and higher than that of LaFeO3, both
for propane and ethanol oxidation.
By analysing the kinetic parameters (apparent activation
energies and reaction rates, Table 4), the best catalytic
behaviour of the substituted perovskites against LaFeO3
perovskite is also observed.

4. Discussion

4.1. On the synthesis and characterisation of

La1
xCaxFeO3 perovskites

La1
xCaxFeO3 perovskites with x = 0, 0.2, and 0.4 were

Fig. 7. Ethanol conversion and selectivities to acetaldehyde and CO2 as a prepared by the citrate method which is recognised as an
function of the reaction temperature on LaFeO3. appropriate method to obtain pure solids since the precursor gel
 B.P. Barbero et al. / Applied Catalysis B: Environmental 65 (2006) 21–30
is highly homogeneous. Furthermore, specific surface areas
adequate for materials to be used as catalysts in oxidation
reactions are obtained since considerably lower calcination
temperatures than those employed in ceramic methods are
sufficient to obtain the suitable phase. The perovskites
synthesised in this work have specific surface areas between
13 and 18 m2/g (Table 1). Similar SSA (19 m2/g) was obtained
by Daturi et al. [27] for LaFeO3 prepared as aerogels via the so-
called supercritical drying technique. In contrast, Ciambelli
et al. [8] prepared a series of La1
xCaxFeO3 perovskites with
specific surface areas much lower than those obtained by us (3–
6 m2/g range). Although the same preparation method was
employed, they calcined at 800 8C (100 8C higher than us). This
could be the reason for the decrease of the specific surface area.
As it has been shown in the study on La1
xCaxCoO3 perovskites
[9], an increase of the calcination temperature from 700 to
800 8C strongly reduces the specific surface area. Another
difference with regard to our samples is the SSA variation with
the calcium content. Ciambelli et al. [8] found that the specific
surface areas increased for low x values (x = 0.1–0.2) and then
decreased but we observed the contrary tendency.
The synthesis success was corroborated by XRD (Fig. 1)
where a single perovskite phase was observed for all the
samples. No segregated phase or contamination was detected.
This confirms how appropriate is the synthesis method for
avoiding the phase segregation and obtaining pure phases.
Nevertheless, the presence of other phase(s) in very low
concentration and/or as amorphous phase cannot be discarded
because of the limitations of the XRD technique.
The elemental composition results from XRF and XPS
indicate lanthanum excess for all of the samples. The excess
detected by XPS is a bit higher than that by XRF. Considering
that the result from XRF is related to the bulk composition
while that from XPS is about the surface composition, some
lanthanum-rich phase could exist on the surface. Another
relevant insight is that the (La + Ca)/Fe surface atomic ratio is
almost constant for all of the catalysts. This indicates that there
is no substitution limit of calcium for lanthanum in
La1
xCaxFeO3 perovskites, at least when x varies between 0
and 0.4. Thus, the obtained catalysts should consist mainly of
the single perovskite phase with a surface enrichment in
lanthanum and calcium.
The X-ray diffractograms of La1
xCaxFeO3 perovskites
studied in this work show that the diffraction lines are broaden
and their intensities decrease when the calcium amount
increases. This makes that the less intense diffraction lines
become undetected. The prevailing lines in the La0.6Ca0.4FeO3
perovskite are in agreement with those of the PDF 75-0541 of
the cubic LaFeO3 perovskite, although they are shifted to
higher 2u angles. Then, the calcium substitution could induce
changes in the crystalline structure modifying the orthorhombic
structure tending to a nearly cubic structure. The structure
refinement by means of Rietveld method confirms that the
substituted perovskite structure are less distorted than that of
LaFeO3.
The structural modification should be easily detected by
Raman spectroscopy, which is a very sensitive technique to
27
structure distortion. However, the perovskite structure can
change due to either stoichiometry modifications as well as by
the effect of thermal treatments or changes of the oxygen partial
pressure in the surrounding atmosphere [20]. Moreover, as it
can be deduced from the comparison of the spectra obtained
with two different spectrometers (Fig. 2a and b), the results
depend strongly on the experimental conditions employed in
the measurements. While the spectrum shown in Fig. 2a can be
assigned to an orthorhombic structure [15–18], the spectrum of
Fig. 2b could correspond to a rhombohedral [17] or
orthorhombic structure [18]. It has been reported in the
literature [28] that the orthorhombic phase is the stable phase at
ambient temperature for lanthanum orthoferrite and phase
transition to rhombohedral structure occurs at about 977 8C.
Probably, a local overheating of the sample by laser effect
caused the phase change of the samples analysed with the
LabRam Instrument (Fig. 2b). Further on these differences, it is
important to note that in both cases (Fig. 2a and b) the band
intensity decreases when the calcium amount increases. This
could be due to the disorder induced by the calcium
substitution. Considering that a cubic symmetry perovskite
should not show any Raman bands [18,20], a nearly cubic
symmetry in the calcium-substituted perovskites would be
obtained confirming the one observed by XRD.
In summary, La1
xCaxFeO3 perovskites with adequate
specific surface area to be applied in oxidation reactions can
be obtained by a simple method such as the citrate method,
controlling the calcination temperature. The catalysts are single
perovskite phases with a slight surface enrichment in
lanthanum and calcium. The calcium substitution in the A
sites of the LaFeO3 perovskite produces a structural change
from the orthorhombic structure towards a less distorted one.
Furthermore, XRD and Raman spectroscopy results presented
here clearly show that the perovskite structure changes by
composition modifications as well as by exposition at high
temperatures. The study of the surrounding atmosphere on the
perovskite structure is beyond the aim of this work. However,
we can alert about how careful we should be when correlating
the crystalline structure of a perovskite with catalytic activity,
since it is expected that the reaction atmosphere (especially in
the case of oxidation reactions) can modify the crystalline
structure.
4.2. Effect of calcium on the electroneutrality
of the perovskites
The calcium substitution causes a notable change on the
lattice parameters determined by Rietveld method (Table 2).
The shift of diffraction lines to higher 2u angle values when the
calcium amount increases, indicates a lower interplanar
distance and, consequently, a decrease of the cell volume.
This is corroborated by the results of unit cell volumes
(Table 2). The ionic radius of Ca2+ 12-coordinated (1.34 Å)
[29] is very similar to that of La3+ 12-coordinated (1.36 Å) [29];
so, the cell volume should be very slightly affected by the
calcium substitution. Nevertheless, since calcium has lower
oxidation state than lanthanum, an increase of the iron
28 B.P. Barbero et al. / Applied Catalysis B: Environmental 65 (2006) 21–30
oxidation state from Fe3+ to Fe4+ can occur or oxygen vacancies
can be generated to maintain the structure electroneutrality. The
Fe4+ formation in substituted or unsubstituted AFeO3
perovskites has been reported [22]. The ionic radius of Fe4+
(0.585 Å) is smaller than that of Fe3+ (0.645 Å); then, this could
be the reason for the cell volume decrease.
The FT-IR results are in line with these observations.
Particularly, the FT-IR band intensity appearing at around
877 cm
1 increases when the calcium amount increases. In
La1
xSrxCoO3 perovskites, the appearance of a new band at
around 670 cm
1 whose intensity increased with the increase of
strontium ions content was observed by Berger et al. [30]. They
assigned this band to the increase of Co–O constant force as a
result of charge increase of cobalt ions (Co4+ ions formation) to
conserve the electroneutrality of the perovskite structure.
According to this, we can assign the band at 877 cm
1 to
structural modifications produced by the substitution with
calcium and probably associated to the Fe4+ ions formation.
The results from TPR also indicate the Fe4+ formation. The
curves shown in Fig. 4 represent the hydrogen consumption due
to sample reduction. To determine the final state of the sample,
when the reduction finished, the reducing gaseous mixture was
shifted to helium and the sample was cooled down to ambient
temperature in helium flow. Immediately after, it was analysed
by XRD and it was observed that the perovskite phase was
preserved. Only in the diffractogram of the reduced La0.6Ca0.4-
FeO3 sample, the most intense peak of metallic iron was weakly
detected. Furthermore, a slight peak shift towards lower 2u
angles was found (Fig. 9). This indicates that the reduction
causes an expansion of the unit cell volume which is predictable
from the difference between the ionic radii of Fe4+ and Fe3+.
These results conduce us to assign the two first signals to the
reduction of Fe4+ to Fe3+. Ciambelli et al. [8] observed two
overlapped peaks in the TPR curves of La1
xCaxFeO3
perovskites and they assigned the shoulder at lower temperature
to more easily reducible surface Fe4+ ions. The good resolution
of the two peaks in our TPR profiles induce us to propose a
reduction of Fe4+ to Fe3+ in two steps: Fe4+ ! Fe4+/
Fe3+ ! Fe3+ rather splitting due to a diffusion phenomena.
The reduction step starting at 620 8C, which is not completely
resolved under the experimental conditions used in this study,
Fig. 9. XRD patterns of La0.6Ca0.4FeO3 perovskite both fresh and after TPR.
(*) metallic Fe.
can be assigned to the total reduction of iron towards the
metallic state. This deep reduction would be accomplished by
the destruction of the perovskite phase. Due to the intensity of
that signal increases when the calcium content increases, it is
possible to suggest that the calcium substitution decreases the
stability of the perovskite structure. This can be also deduced
from the FT-IR results since the shift of the bands toward higher
wavenumbers (Fig. 2) can be attributed to a weakening of the
metal–oxygen bonds.
A decisive measurement to confirm the Fe4+ formation can
be obtain by using Mössbauer spectroscopy (work in progress).
The singlet with isomeric shift of 0.12 mm/s detected in the
spectra at 298 K is in good agreement with a signal of Fe4+.
In summary, the XRD, FT-IR and TPR results indicate that
the electronic debalance caused in the perovskites by the partial
substitution for lanthanum (3+) with calcium (2+) is
compensated by the oxidation of a part of Fe3+ ions to Fe4+.
Taking into account that the ionic radius of Fe4+ is smaller than
that of Fe3+ and that La3+ and Ca2+ have similar ionic radii, it is
expected that theptolerance
ffiffiffi factor defined by Goldschmidt [31]
t ¼ ðRA þ RO Þ= 2ðRB þ RO Þ, decreases when the calcium
substitution increases. The perovskite structure exists in oxides
within the t range 0.75 < t < 1 [2]. The value 1 is for cubic
perovskite and it decreases when the structure is distorted.
Then, the increase of value t when increasing the calcium
substitution level would indicate a tendency of the structure
towards a nearly cubic one, confirming that previously
observed by XRD and Raman.
If the increase of the iron oxidation state is sufficient to
conserve the perovskite electroneutrality, the formed Fe4+
amount would increase linearly with the calcium amount. In
order to verify this, the Fe4+ amount was estimated from the
TPR results (Table 1) and it was plotted as a function of the
calcium amount (Fig. 10). Furthermore, considering that
the shift of the XRD lines is proportional to the Fe4+ amount,
the position of the XRD main peak was also plotted as a
function of the calcium amount in Fig. 10. As it can be seen, the
relationship is not linear. Even if it is considered that
the calcium amount determined by XRF is lower than the
theoretical value, the relationship is not linear. Then, it is
predictable that besides Fe4+, oxygen vacancies are formed to
Fig. 10. Position of the XRD main peak, Fe4+ amount from TPR results and
area under the first O2-TPD peak per surface area unit as a function of the
calcium amount.
 B.P. Barbero et al. / Applied Catalysis B: Environmental 65 (2006) 21–30
maintain the electroneutrality, especially when the substitution
level increases.
A simple and direct way to determinate oxygen vacancies is
by means of O2-TPD experiments. The first peak of the curves
shown in Fig. 5 is generally assigned to oxygen adsorbed on
surface oxygen vacancies. Thus, the area under the curve of this
signal would be a measure of surface oxygen vacancies amount.
The first signal intensity increases in the order La0.8Ca0.2-
FeO3 < LaFeO3 < La0.6Ca0.4FeO3. However, considering that
they are ‘‘surface vacancies’’ it would be more convenient to do
the comparison taking into account the surface area of the
samples. When the first signal intensity was divided by the
surface area, an increase of the surface oxygen vacancy amount
when increasing the calcium amount was observed (Fig. 10).
This indicates that the electronic debalance caused by the
calcium partial substitution for lanthanum is formerly
compensated by the Fe4+ ion formation and, when the calcium
amount increases, oxygen vacancies are also formed.
A TPD signal such as that observed above 350 8C, which is
not completely resolved under the experimental conditions
used in this work, has also been observed by Nitadori and
Misono [4]. They assigned it to partial reduction of Fe4+ to Fe3+.
In our experiment, it is not very probable that the second signal
corresponds to Fe4+ reduction. We carried out the temperature-
programmed reduction following the TPD after cooling down
the sample to ambient temperature under helium flow without
reoxidation. Since the TPR results shown a clear reduction
signal, we discarded the possibility that the second TPD peak be
associated to the Fe4+ reduction. Then, we agree with other
several authors [24–26] who have assigned that signal to
oxygen coming from the catalyst bulk and probably associated
to inner oxygen vacancies. The desorption of these species
could be completed at much lower temperatures during the
temperature-programmed reduction. In fact, a very weak signal
below 200 8C is observed in the TPR curves. This signal could
be due to oxygen release not associated to iron reduction. The
amount of these vacancies is difficult to estimate but an increase
of the intensity when the calcium amount increases is observed.
In conclusion, the partial substitution for lanthanum by
calcium in La1
xCaxFeO3 perovskites causes an oxidation state
increase of a part of the iron ions and at higher substitution
levels (x > 0.2) oxygen vacancies are also formed.
4.3. On the catalytic performance of the La1
xCaxFeO3
perovskites
It is widely known that the catalytic activity of the perovskite
oxides is related to unusual oxidation states of the transition
metal ions, the amount of non-stoichiometric oxygen and the
structural defects of lattice. Depending on the nature of the
VOC molecule that wants to be oxidised, one of these aspects
will be more relevant for the catalytic performance and will
give an idea about how the reaction occurs. The increase of the
iron oxidation state could improve the catalyst performance in a
redox cycle, in which the iron is reduced supplying the
necessary oxygens for the oxidation reaction and then it is
reoxidized taking oxygen from the gas phase. The oxygen
29
vacancies play an important role, favourable for the oxidation
process because they accelerate the dissociation of oxygen
molecule on the surface and also, increase the mobility of
lattice oxygen.
To identify which is the phenomenon governing the reaction,
we intent to find a linearity between the catalytic behaviour and
the perovskite characterisation results. In Fig. 6 and Table 4 it is
observed that the calcium substitution modifies the perfor-
mance of the LaFeO3 perovskite but in a different way if the
oxidised molecule is a hydrocarbon (propane) or an alcohol
(ethanol). In the propane oxidation, the conversion increases
in the order La0.8Ca0.2FeO3 < LaFeO3 < La0.6Ca0.4FeO3
whereas the ethanol conversion increases when the calcium
amount increases. Taking into account that the Fe4+ ion amount
also increases when the calcium amount increases, we can
suggest that the ethanol oxidation proceeds through a redox
mechanism where the par Fe4+/Fe3+ plays an important role. In
contrast, the propane oxidation could be governed by another
mechanism in which the surface oxygen vacancy concentration
is relevant. According to the O2-TPD results, the intensity of the
a-oxygen signal follows the same order that the propane
conversion (La0.8Ca0.2FeO3 < LaFeO3 < La0.6Ca0.4FeO3).
However, when the VOC specific conversion (conversion
per surface area unit) is considered, very similar results are
obtained with both molecules. Furthermore, the calcium-
substituted samples are more active than the unsubstituted
perovskite but no difference is observed with several
substitution levels. This tendency is coherent with the apparent
activation energy (Ea) values. Ea for LaFeO3 perovskite in
propane oxidation was 26.2 kcal/mol while for La0.8Ca0.2FeO3
and La0.6Ca0.4FeO3 perovskites were 22.9 and 23.4 kcal/mol,
respectively. Existing no steric limitations due to the model
molecule sizes, it can be assumed that the calcium substitution
generates complex active sites. Probably, these sites are related
to Fe4+ ions that, as discussed above, are generated to conserve
the electroneutrality but not further than a certain amount.
However, the site concentration would not only depend on the
Fe4+ ions content. The catalyst capacity to complete the redox
cycle by means of the oxygen activation to re-oxidise itself also
plays an important role. The equilibrium between these factors
seems to be determinant for the improvement of the catalytic
activity. Probably, thus can be explained the methane
combustion results obtained by Ciambelli et al. [8].
5. Conclusions
La1
xCaxFeO3 perovskite-type oxides have been obtained
by means of a very simple method, such as the citrate method,
with adequate specific surface area for oxide catalysts to be
used in oxidation reactions, simply controlling the calcination
temperature. The catalysts are single perovskite phases with a
slight surface enrichment in lanthanum and calcium. The
calcium substitution in the A sites of the LaFeO3 perovskite
produces a structural change from the orthorhombic structure
towards a less distortioned one. Furthermore, XRD and Raman
spectroscopy results clearly show that the perovskite structure
30 B.P. Barbero et al. / Applied Catalysis B: Environmental 65 (2006) 21–30
changes by composition modifications as well as by exposition
at high temperatures.
The structure electroneutrality of La1
xCaxFeO3 perovskites
is achieved by an oxidation state increase of a part of the iron
ions from Fe3+ to Fe4+ and at higher substitution levels
(x > 0.2), oxygen vacancies are also formed.
Finally, an improvement of the catalytic activity in propane
and ethanol combustion was observed on the substituted
perovskites. Correlating this with the characterisation results,
the active sites would be associated to the Fe4+ ions.
Acknowledgements
The financial support from CONICET, Universidad Nacio-
nal de San Luis, Fundación Antorchas and SeCyT (Argentina),
FNRS (Belgium) cooperation project is gratefully acknowl-
edged. The Raman spectra (a) were acquired thanks to a grant
of Japan International Cooperation Agency to CENACA
(Argentina). The authors also thank to Pierre Eloy (Université
Catholique de Louvain, Louvain-la-Neuve, Belgium) for his
help with the XPS analysis.
References
[1] R.L. Garten, R.A. Dalla Betta, J.C. Schlatter, in: G. Ertl, H. Knozinger, J.
Wertkamp (Eds.), Handbook of Heterogeneous Catalysis, vol. 4, VCH,
Weinheim, Germany, 1998, p. 1668.
[2] L.G. Tejuca, J.L.G. Fierro, Adv. Catal. 36 (1989) 237.
[3] L. Simonot, F. Garin, G. Maire, Appl. Catal. B 11 (1997) 167.
[4] T. Nitadori, M. Misono, J. Catal. 93 (1985) 459.
[5] M.A. Alario Franco, M.J.R. Henche, M. Vallet, J.M. González Calbet, J.C.
Grenier, A. Wattiaux, P. Hagenmuller, J. Solid State Chem. 46 (1983) 23.
[6] M.A. Alario Franco, J.M. González Calbet, M. Vallet Regi, J.C. Grenier, J.
Solid State Chem. 49 (1983) 219.
[7] M. Vallet Regi, J. González Calbet, M.A. Alario Franco, J.C. Grenier, P.
Hagenmuller, J. Solid State Chem. 55 (1984) 251.
[8] P. Ciambelli, S. Cimino, L. Lisi, M. Faticanti, G. Minelli, I. Pettiti, P. Porta,
Appl. Catal. B 33 (2001) 193.
[9] N.A. Merino, B.P. Barbero, P. Grange, L.E. Cadús, J. Catal. 231 (2005)
232.
[10] P. Courty, H. Ajot, C. Marcilly, B. Delmon, Power Technol. 7 (1973) 21.
[11] R.A. Young, A. Sakthivel, T.S. Moss, C.O. Paiva-Santos, J. Appl. Cryst. 28
(1995) 366.
[12] T.L. Barr, J. Phys. Chem. 82 (16) (1978) 1801.
[13] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, in: G.E. Muilenberg
(Ed.), Handbook of X-ray Photoelectron Spectroscopy, first ed., Perkin-
Elmer Corporation (Physical Electronics Division), 1979.
[14] D. Briggs, M.P. Seah, 2nd ed., Practical Surface Analysis, vol. 1, John
Willey & Sons, 1993.
[15] M. Popa, J. Frantti, M. Kakihana, Solid State Ion. 154/155 (2002)
135.
[16] M. Popa, J. Frantti, M. Kakihana, Solid State Ion. 154/155 (2002) 437.
[17] E. Bontempi, C. Garzella, S. Valetti, L.E. Depero, J. Eur. Ceram. Soc. 23
(2003) 2135.
[18] E. Granado, N.O. Moreno, A. Garcı́a, J.A. Sanjurjo, C. Rettori, I. Torriani,
S.B. Oseroff, J.J. Neumeier, K.J. McClellan, S.W. Cheong, Y. Tokura,
Phys. Rev. B 58 (17) (1998) 11435.
[19] S.D. Ross, Inorganic Infrared and Raman Spectra, McGraw-Hill, London,
1972.
[20] M. Couzi, P.V. Huong, Ann. Chim. 9 (1974) 19.
[21] A.E. Lavat, E.J. Baran, Vibrat. Spectrosc. 32 (2003) 167.
[22] P. Ciambelli, S. Cimino, S. De Rossi, L. Lisi, G. Minelli, P. Porta, G.
Russo, Appl. Catal. B 29 (2001) 239.
[23] R. Spinicci, A. Tofanari, A. Delmastro, D. Mazza, S. Ronchetti, Mater.
Chem. Phys. 76 (2002) 20.
[24] R.J.H. Voorhoeve, J.P. Remeika, P.E. Freeland, B.T. Mattias, Science 177
(1972) 353.
[25] N. Yamazoe, Y. Teraoka, T. Nakamura, Chem. Lett. (1981) 1767.
[26] T. Seiyama, N. Yamazoe, K. Eguchi, Ind. Eng. Chem. Prod. Res. Dev. 24
(1985) 19.
[27] M. Daturi, G. Busca, R.J. Willey, Chem. Mater. 7 (1995) 2115.
[28] L. Bornstein, Numerical Data and Functional Relationships in Science and
Technology, vol. III/4a, Springer-Verlag, Berlin, 1970.
[29] R.D. Shannon, Acta Cryst. A 32 (1976) 751.
[30] D. Berger, V. Fruth, I. Jitaru, J. Schoonman, Mater. Lett. 58 (2004) 2418.
[31] V.M. Goldschmidt, Skr. Nor. Videnk-Akad., Kl. 1: Mat.-Naturvidensk. Kl.
No. 8 (1926).