Synthetic Methods for Perovskite Materials – Structure and Morphology

Crystallization is used at some stage in nearly all process industries as a method of production, purification or recovery of solid materials. In recent years, a number of new applications have also come to rely on crystallization processes such as the crystallization of nano and amorphous materials. The articles for this book have been contributed by the most respected researchers in this area and cover the frontier areas of research and developments in crystallization processes. Divided into five parts this book provides the latest research developments in many aspects of crystallization including: chiral crystallization, crystallization of nanomaterials and the crystallization of amorphous and glassy materials. This book is of interest to both fundamental research and also to practicing scientists and will prove invaluable to all chemical engineers and industrial chemists in the process industries as well as crystallization workers and students in industry and academia.


Introduction
Solid state chemistry thrives on a rich variety of solids that can be synthesized using a wide range of techniques.It is well known that the preparative route plays a critical role on the physical and chemical properties of the reaction products, controlling the structure, morphology, grain size and surface area of the obtained materials (Cheetham & Day, 1987;Rao & Gopalakrishnan, 1997).This is particularly important in the area of ABO 3 perovskite compounds given that they have for long been at the heart of important applications.From the first uses of perovskites as a white pigments, PbTiO 3 in the 1930's (Robertson, 1936) to the MLC capacitors (mostly based in substituted PbTi 1-x Zr x O 3 or BaTiO 3 materials) in which today's computers rely on to operate, synthetic methods have been a key factor in the optimization of their final properties (Pithan et al., 2005).
Traditionally, most of these ceramic materials have been prepared from the mixture of their constituent oxides in the so called solid state reaction, "shake and bake" or ceramic method, a preparative route for which high temperature is a must in order to accelerate the slow solid-solid diffusion (Fukuoka et al., 1997;Inaguma et al., 1993;Safari et al., 1996).Despite its extended use in practically all fields in which perovskite-structured materials are needed, not all applications are better off with this method since the low kinetics and high temperature also yield samples with low homogeneity, with the presence of secondary phases and with uncontrolled (and typically large) particle size of low surface area which are undesired for some applications such as in gas sensors or in catalysis where small particles and high surface area are needed (Bell et al., 2000).This conventional route, however, is widely employed due to its simplicity and low manufacturing cost.Nevertheless, with appropriate optimisation, as when softmechanochemical processing is used prior to calcinations at high temperature (Senna, 2005), the method results in high quality single phase perovskites that can be used in electroceramic applications.
On the other hand, combustion methods (glycine-nitrate, urea based, and other modifications) have been proposed as one of the most promising methods to prepare perovskite oxide powders to be used as cathode materials in Solid Oxide Fuel Cell technology.(Bansal & Zhong, 2006;Berger et al., 2007;Dutta et al., 2009;Liu & Zhang, 2008).This method consist on a highly exothermal self-combustion reaction between the fuel (usually glycine, urea or alanine) and the oxidant (metal nitrates), that produces enough heat to obtain the ceramic powders.Compared to the ceramic method this synthetic route has much faster reaction times and lower calcination temperatures leading to powders with large compositional homogeneity and nanometric particle sizes, which are desired characteristics for this type of application.
For some applications, such as in multiferroic materials based devices, it is the crystal symmetry of the multiferroic what matters.In these materials the presence or not of a centre of symmetry is crucial for the observation of ferroelectricity.With this regard, there are cases, as in some AMnO 3 perovskites (A=Y or Dy), in which the synthetic route determines whether an orthorhombic compound with a centre of symmetry (i.e.non ferroelectric) or a hexagonal phase without the centre (i.e.ferroelectric) is formed (Carp et al., 2003;Dho et al., 2004).Consequently, preparative conditions have to be carefully selected in order to obtain crystal phases with the adequate structure.The use of more than one synthesis method is thus worth trying in all cases.
In this work three different groups of perovskite compounds have been prepared and their crystal structure and microstructure have been studied using X-ray diffraction (XRD) and scanning electron microscopy (SEM).Each group of samples had its own structural characteristics so, prior to choosing one synthetic approach, trials were carried out using different methods.In all cases, the final method chosen was the one that maximised phase purity and resulted in better properties.Here we demonstrate that phase pure samples susceptible to be compared depending on the desired characteristics can be obtained using different synthetic methods.

Experimental
Up to four different synthetic methods (solid state reaction, glycine-nitrate route, sol-gel and freeze-drying) have been used to synthesize a group of 14 perovskite compounds (Figure 1), which have the potential for their use in different applications.www.intechopen.com

Compound
These compounds have been divided into three groups and their compositional details are summarised in Table 1.Preparation procedures are detailed below.The selection of each method and composition for each series of perovskites was based on the desired applications that are described later.

Ceramic solid state reaction
The compounds prepared via the ceramic route were obtained from mixing stoichiometric amounts of the raw oxides with 2-propanol in an agate mortar.Starting materials were always oxides of high purity such as La 2 O 3 (99.99%),Sm 2 O 3 (99.999%),Pr 2 O 3 (99.9%),Gd 2 O 3 (99.99%),BaO (99.99%),SrO (99.9%),CaO (99.9%) and Fe 2 O 3 (99.98%).Afterwards, these mixtures were shaped into pellets and were fired in air at 950ºC for 2h.The products obtained were ground, pelletized again and fired at higher temperatures.The flowchart shown in Figure 1a details the heat treatments required in each case until pure samples were obtained.

Glycine-nitrate route
The preparations of the perovskite compounds with the general composition Ln 0.5 A 0.5 FeO 3-δ by glycine-nitrate route involved the use of some nitrates instead of the oxides as starting materials: Ba(NO 3 ) 2 (99.99%),Sr(NO 3 ) 2 (99.9%) and Fe(NO 3 ) 3 (99.98%).The oxides used were La 2 O 3 (99.99%),Sm 2 O 3 (99.999%),Gd 2 O 3 (99.99%).The oxides were dissolved in diluted nitric acid to obtain the corresponding nitrates and the metal nitrates dissolved in distilled water.The solutions were mixed in a 1 litre glass beaker, under constant stirring and placed on a hot plate to evaporate the water excess.After a significant reduction of the solution volume the glycine was added.The amount of glycine used was calculated in order to obtain a glycine/nitrate molar ratio of 2:1.This amino acid acts as complexing agent for metal cations and as the fuel for the combustion reaction.The resulting viscous liquid was auto-ignited by putting the glass beaker directly in a preheated plate (at ~400ºC).The obtained powders were pelletized and fired at 800ºC for 2 hours to remove the carbon residues.The heatings at temperatures above 800ºC were repeated until pure phases were obtained.A flowchart with more details and heat treatments involved in this synthesis is shown in Figure 1b.

Sol-gel
The sol-gel method was used for the oxides of general composition Nd 0.8 Sr 0.2 (Mn 1-x Co x )O 3 (x = 0.1, 0.2 and 0.3).Initially, the oxide Nd 2 O 3 (99.9%)was dissolved in aqueous nitric acid followed by the addition of Sr(NO 3 ) 2 (99%), Co(NO 3 ) 2 .6H 2 O (99%) and Mn(C 2 H 3 O 2 ) 2 .4H 2 O (99%).Citric acid was then used as the quelating agent and ethylene glycol as the sol forming product.The solution was slowly evaporated in a sand bath for 24 h and the gel obtained was subjected to successive heat treatments at the temperature of 850ºC (with intermediate grindings).Each firing was of 10 h and was carried out under nitrogen atmosphere.The flowchart for this synthesis is shown in Figure 1c.
The mixture for freeze-drying method was frozen drop-by-drop under liquid nitrogen and subjected to freeze drying at P = 5.10 -2 mbar.Thermal decomposition was achieved by slow heating in air up to 600ºC.The pure phases were obtained after repeated heatings at 850ºC (with intermediate grindings), each of 10 h, under nitrogen atmosphere.A flow chart for this method is shown in Figure 1d.

Characterization
Room temperature X-ray powder diffraction data were collected in the 18≤2θ≤110° range with an integration time of 10 s/0.02° step.A Bruker D8 Advance diffractometer equipped with a Cu tube, a Ge (111) incident beam monochromator ( = 1.5406Å) and a Sol-X energy dispersive detector were used for the samples obtained by glycine-nitrate route and LSSB-ss.
A Philips X'Pert-MPD X-ray diffractometer with secondary beam graphite monochromated Cu-Kα radiation was used for the samples obtained by ceramic solid-state, sol-gel and freeze-drying techniques.
All samples were single phase without detectable impurities.The crystal structure was refined by the Rietveld method (Rietveld, 1959) from X-ray powder diffraction data using GSAS software package (Larson & Von Dreele, 1994).
Microstructural analysis was carried out in a JEOL JSM 6400 scanning electron microscope (SEM) using a secondary electron detector at 30 kV and 1.10 −10 A for the LPS20, LNSC30, LPSC40, LPSC50, LPSC60, LPSC70 samples and a JEOL JSM-7000F at 3 kV and 11.10 -12 A for the rest of samples.

Characterization of Ln 1−x A
x FeO 3−δ (Ln=La, Nd, Pr; A=Sr, Ca) perovskites with 0.2≤x≤0.8 The (La 1−x Sr x )FeO 3-δ (LSF) perovskite system exhibits high electronic and oxide ion conductivities at high temperatures, which make it attractive for solid oxide fuel cell (SOFC) cathodes.Several works (Ecija et al., 2011;Rodríguez-Martínez & Attfield, 1996;Vidal et al., 2007 and references therein) have shown that the physical properties of these perovskite materials are very sensitive to changes in the doping level (x) the average size of the A cations (<r A >) and the effects of A cation size disorder (σ 2 (r A )).
The synthesis of these compounds allows us to study the effect of the variation of the doping level x on the properties of the perovskites with general formula Ln 1−x A x FeO 3−δ (Ln=La, Nd, Pr; A=Sr, Ca) applied as SOFC cathodes.This has been achieved by keeping constant both the average size (<r A >) and the size mismatch (σ 2 (r A )) to 1.22 Å and 0.003 Å 2 , respectively.
For the preparation of this series the solid state reaction route has been chosen due to its simplicity to obtain perovskite phases in the same synthetic conditions.

Structural study
The room temperature X-ray powder diffraction patterns of these compounds are shown in Fig. 2a.A structural transition from orthorhombic symmetry (space group Pnma) for samples with x≤0.4 to rhombohedral symmetry (space group R-3c) for 0.5≤x≤ 0.7 compounds, and finally, to a mixture of rhombohedral R-3c and cubic Pm-3m perovskite phases for the x = 0.8 composition was observed.Representative Rietveld fits to the X-ray diffraction data for the samples LPS20, LNSC50 and LNSC80 are shown in Fig. 2b, 2c and 2d, respectively.
The dependence with the doping level x of the cell parameters and cell volume per formula unit and atomic distances and bond angles for all samples are represented in Figure 3a and b, respectively.There is a systematic decrease in volume, cell parameters and <Fe-O> distances with increasing doping level across the series.Given that in the system studied the A-site mean ionic radius <r A > has been kept constant as the doping level increases, the observed effect can be solely associated to a reduction of the Fe-site mean ionic radius as it www.intechopen.comoxidises from Fe 3+ to Fe 4+ , with smaller radius (<r Fe >, r Fe 3+ =0.645 Å and r Fe 4+ =0.585 Å) (Shannon, 1976).Details of these effects are given elsewhere (Vidal et al., 2007).
The increase of the <A-O> distances and <Fe-O-Fe> bond angles with increasing doping level (x) can be explained due to the structural transition produced with x: when passing from orthorhombic (Pnma, LPS20) to a mixture of rhombohedral and cubic (R-3c + Pm-3m, LSC80) the octahedra that compose the perovskite structure reduce its cooperative tilting and the structure "expands".
These results are in nice agreement with other structural studies of related perovskites in which similar structural transitions with doping level were observed (Blasco et al., 2008;Dann et al., 1994;Tai et al., 1995).

Morphological study
Microstructure of bulk samples was study by scanning electron microscopy (SEM).Images of the sintered bars at 1300ºC are shown in Fig. 4.
These micrographs show different particle size distributions with grain sizes ranging between 0.33 and 2.83 m for sample LPS20, to 5 and 37 m for sample LSC70.
The dispersion in particle sizes is larger for values of x≥0.4.As observed previously (Liou, 2004a(Liou, , 2004b) ) this increase in particle size with the doping level seems to be a result of a change in the melting point of the samples that decreases increasing alkaline-earth cation content.According to Kharton et al. (Kharton et al., 2002) this effect results in a liquid-phase assisted by sintering and an enhanced grain growth.As in the case of the previous family of iron perovskites, this new series of compounds are of interest for their use as mixed ionic electronic conducting materials, mainly from the point of view of cathodes for Solid Oxide Fuel Cells (SOFC), although they could also be used as ceramic membranes for oxygen separation.In the present case, the degree of lanthanide substitution was fixed to x=0.5 given that previous studies have shown that it is precisely at this degree of substitution when electronic and ionic conductivity are maximised (Hansen, 2010;Vidal et al., 2007).
In the present series of compounds the effect of the <r A > variation on the properties of four different phases, Ln 0.5 A 0.5 FeO 3-δ (Ln=La, Nd, Sm; A=Ba, Sr) was evaluated.For this series (Table 1), <r A > has been varied between 1.34 and 1.25 Å keeping x and σ 2 (r A ) constant, with values of 0.5 and 0.0161 Å 2 , respectively.
Prior to choosing a synthesis method for all samples, two different methods were tried for one of the samples, La 0.05 Sm 0.45 Sr 0.18 Ba 0.32 FeO 3-δ , for which the structural parameters and morphology were evaluated.The ceramic and glycine-nitrate routes were used in the present case.

3.2.1
Influence of the synthetic method on the structure and morphology of the La 0.05 Sm 0.45 Sr 0.18 Ba 0.32 FeO 3-δ Laboratory X-ray diffraction data at room temperature for La 0.05 Sm 0.45 Ba 0.5 FeO 3-δ obtained by ceramic and glycine-nitrate routes were extremely similar, both samples being pure to the detection limits of the technique.XRD patterns were fitted by the Rietveld method (Figure 5) considering a rhombohedral symmetry (space group R-3c) in both cases.From the fits to the XRD data it was observed that both phases are nearly identical: they crystallise in the same space group and does not show significant difference among lattice parameters, cell volume or density values.
The morphological study, however, shows a different picture.As shown in Figure 6, where SEM micrographs taken at the same amplification for both samples are presented, their microstructure is quite different.The sample prepared following the ceramic route presents a microstructure with heterogeneous grain sizes and shapes, in which particles range from ~ 2 to approximately 8 µm in diameter.On the other hand, the average grain size of the sample obtained by glycine-nitrate route is about 200 nm, almost an order of magnitude smaller.The higher calcination temperature and longer reaction time required to obtain the samples in the ceramic process can explain the bigger grain size showed for these samples (Melo Jorge et al., 2001).For the application as SOFC cathodes, samples with small and homogeneous particle sizes are usually preferred.As a consequence, the glycine-nitrate process was considered a more appropriate technique for preparing the perovskite samples of this series: Ln 0.5 A 0.5 FeO 3-δ (Ln= La, Nd, Sm; A= Ba, Sr) with 1.25≤ <r A > ≤1.34 Å.

Structural study
Room temperature X-ray diffraction patterns of the LB134, LNSB131, LNSB128, LSSB125-gn samples show that all the samples are single phase compounds (Figure 7a).A shift of the diffraction maxima to lower diffraction angles (2θ) with decreasing <r A > anticipates an increase of the lattice parameters.X-ray powder diffraction patterns were indexed using a cubic symmetry (Pm-3m space group) for the LB134 and LNSB131 samples and a rhombohedral symmetry (R-3c space group) in the case of LNSB128 and LSSB125-gn compounds.Figure 7b shows the Rietveld fits to the XRD patterns for two samples.Given that the doping level has been fixed (x=0.5), changes in the ratio Fe 3+ /Fe 4+ and, therefore, in the Fesite ionic radius (<r Fe >), are not expected.Consequently, the decrease in lattice parameters is ascribed to the variation of <r A >.A more detailed analysis is given elsewhere (Ecija et. al, 2011).This also explains the decrease of the <A-O> and <A-Fe> mean distances with decreasing <r A > (Figure 8b).Although <r Fe > has been kept constant, there is a slight reduction of the <Fe-O> distances as <r A > decreases, which is a consequence of the tilting in the FeO 6 octahedra due to the rhombohedral distortion.The decrease of the <Fe-O-Fe> bond angles is the result of the same effect (Woodward, 1998).

Morphological study
Figure 9 shows the SEM micrographs of the LB134, LNSB131, LNSB128, LNSB125-gn bulk samples after the last heating at 1050ºC.
In this series of samples no significant differences can be found in the morphology and average particle size.All samples present some agglomeration and fine grain size.Image analysis of the micrographs has allowed us to determine that the average grain size of the samples is in the range of 150-250 nm.The hole-doped perovskite manganese oxides with general formula Ln 1-x A x MnO 3 (Ln= La, Pr, Nd; A= Ca, Sr, Ba, Pb; x<0.5) draw considerable attention in the late 1990´s due to their colossal magneto-resistance (CMR) effect at low temperatures (Rao, 1998).In the search for new CMR materials it was observed that doping on Mn site by other transition metal elements, such as Cr, Fe, Co and Ni, was an effective way to obtain new materials, which also helped to understand the new phenomenon (Tai, 2000;Takeuchi, 2002;Ulyanov, 2007).Phase morphology, highly dependent on preparative conditions, was also observed to play an important role in the effect: low temperature CMR was improved as the grain size was reduced (Das, 2002).In order to help in this area, it was decided to study the effects of the synthesis method in a series of perovskite compounds with the general formula Nd 0.8 Sr 0.2 (Mn 1-x Co x )O 3 (0.1≤x≤0.3).The sol-gel and freeze-drying techniques were used in order to compare their structural, morphological and magnetic properties.Details of the later (magnetic properties) can be found elsewhere (Vidal et al., 2005).

Synthesis and characterization of
Fig. 9. SEM micrographs of the surface of the LB134, LNSB131, LNSB128, LSSB125-gn bulk samples.

Structural study
The X-ray powder diffraction patterns (XRPD) of all the compounds studied in this section were indexed in the orthorhombic space group Pmna irrespective of the synthesis method used.Fig. 10 shows the XRD and Rietveld refinement profiles for all phases with the formula Nd 0.8 Sr 0.2 (Mn 1-x Co x )O 3 .
When the lattice parameters and cell volume are compared a slight decrease with increasing cobalt content is observed in both cases.This effect is related to the changes in sizes of the B site atoms upon doping (Meera et al., 2001;Pollert et al., 2003).The oxidation states of Mn and Co in the AMn 1-x Co x O 3 systems has been for long debated (Goodenough et al., 1961;1997;Toulemonde et al., 1998;Troyanchuk et al., 2000), and different mixtures of Mn 4+ -Mn 3+ and Co 3+ -Co 2+ have been proposed.The most likely combination, based in spectroscopic and magnetic results, seems to indicate that the cobalt is introduced as Co 2+ (and not as Co 3+ ) thus causing a mixed state (4+ and 3+) in manganese.This mixture of oxidation states has also been proved useful for Pollert et al. to explain magnetic and electrical properties of the series Nd 0.8 Na 0.2 Mn 1-x Co x O 3 (x≤0.1),Pr 0.8 Na 0.2 Mn 1-x Co x O 3 (x≤0.2) (Pollert et al., 2003a(Pollert et al., , 2003b) ) and La 0.8 Na 0.2 Mn 1-x Co x O 3 (x≤0.2) (Pollert et al., 2004).On the other hand, as the Co content increases, the LSMC materials would shift from the oxygen excessive region to the oxygen deficient region, which would also contribute to the decrease in the unit volume.Wandekar et al. (Wandekar et al., 2009) carried out a detailed study on the crystal structure and conductivity of Co substituted LSM system and showed that the change in the ion radius of B-site element plays a predominant role at low Co content and the increase in oxygen vacancy becomes dominant at high Co content.
Fig. 10.Rietveld fits to the X-ray diffraction data in the orthorhombic Pnma space group for the Nd 0.8 Sr 0.2 Mn 1-x Co x O 3 compounds.The labels "sg" and "fd" indicates that samples were prepared using the sol gel and freeze-drying techniques, respectively.In each case, lattice parameters (a, b, c), unit cell volume (V) and theoretical density (ρ) data are inset.
In consequence, considering only the high spin states of these elements and assuming nearly oxygen stoichiometric phases at room temperature, the observed reduction of the lattice volume in the present samples is consistent with the gradual appearance of Mn 4+ (of smaller size than Mn 3+ and Co 2+ (Shannon, 1976) and so with the reduction of the mean B-site ionic radii.This explanation is also satisfactory for the same effect in the compounds La 0.7 Na 0.3 Mn 1- Table 2 shows details of the averge interatomic <B-O>, <A-O> and <A-B> distances together with <B-O-B> bond angles.As it can be observed <B-O> and <A-B> distances decrease as the cobalt concentration increases.These tendencies can be explained following the same reasoning indicated before, which basically concludes that cobalt enters in the structure in an oxidation state with smaller radius than manganese.No significant changes are observed in the <A-O> distances or in the <B-O-B> angles.This indicates that, despite the observed changes in the B size, the cobalt doping does not cause a noticeable distortion in the perovskite structure.Table 2. Mean atomic distances and bond angles for the series of ABO 3 (A = Nd 0.8 Sr 0.2 ; B = Mn 1-x Co x ; x = 0.1, 0.2 and 0.3) samples prepared by the two methods described in the text.

Variables
In the same way, no significant changes are observed between the compounds synthesised by the sol-gel or the freeze-drying techniques.

Morphological study
The SEM micrographs of the present series of compounds are shown in Fig. 11.All pictures were taken on sintered bars after they were prepared by the sol-gel and freeze-drying methods.
As observed, the Nd 0.8 Sr 0.2 (Mn 1-x Co x )O 3 samples prepared by the sol-gel route show a homogeneous particle size morphology distributed in small agglomerates.The grain size of the nearly spherical particles decreases with Co content from ~ 200 nm for the sample with x = 0.1 to ~ 100 nm for x = 0.3.In the case of the compounds obtained by the freeze-drying technique the morphology of the particles is also spherical and homogeneous but the grain size is quite stable in all phases and slightly higher (~250 nm) than in the previous case.
According to some studies (Kuharuangrong et al., 2004), Co doping of 20% does not usually influence the grain size of LSM, but 40 % mol Co significantly reduces the grain size (about 5-10 µm) of the La 0.84 Sr 0.16 Mn 1-x Co x O 3 (x= 0, 0.2, 0.4) samples prepared by conventional oxide mixing process.However, some reduction in the particle size was observed in the case of La 0.67 Pb 0.33 Mn 1-x Co x O 3 (x= 0 and 0.3) compounds, which were also prepared by the ceramic route (Dhahri et al., 2010).The present observation of a reduction in the particle size of the Nd 0.8 Sr 0.2 (Mn 1-x Co x )O 3 phases in one of the methods but not in the other would point towards the synthesis method as the responsible for this effect, rather than the amount of cobalt.In this case, however, the observed decrease of the grain size is less significant than in previous studies.

Conclusion
Different synthetic methods were used to prepare several perovskite type compounds.Phase-pure perovskites were obtained in all cases, irrespective of the method of synthesis and even when more than one method was used.It is to note that no differences in the overall crystal structure were observed in any of the cases when the method was changed.Moreover, changes in the crystal structure of the compounds within each series could be perfectly explained considering only compositional (and resulting structural) variables such as the mean A-site ionic radius or the A-site disorder, none of them being influenced by the synthesis route.This is true for the Ln 1−x A x FeO 3−δ series where changes from orthorhombic to cubic are consistent with the change in the doping level x and also for the Ln 0.5 A 0.5 FeO 3 where the change was observed from rhombohedral to cubic symmetry when the mean Asite radius, <r A >, was varied.
At least in the present group of compounds this seems to indicate that, properly used, and with no other factors present (such as in the case of the growth of samples on crystallographycally oriented substrates), the structural characteristics can be studied and properly addressed without having into account the synthesis method.
It is no news that, compared with the solid state reaction, the glycine-nitrate route, sol-gel and freeze-drying techniques require lower calcination temperatures to yield pure crystals of perovskite phases, with the resulting energy saving (very important nowadays).The disadvantage is, however, the fact that they are more time consuming and require more controlled synthesis conditions.In the same way, it is clearly observed that the ceramic method always yields phases with higher particle sizes but this is not a disadvantage if the final compounds are not to be used in an application that requires high specific area.It is to note that in the case of the Nd 0.8 Sr 0.2 (Mn 1-x Co x )O 3 , grain size was observed to change depending on the synthesis route.Although no significant differences were observed in the crystal structure, sol-gel method resulted in smaller grain sizes than the freeze drying method.In consequence, this seems to be a consequence of the preparative method which reinforces the idea that, if the microstructure of the sample is the key issue, the use of several synthesis methods is always worth trying.

Fig. 1 .
Fig. 1.Flowcharts for the: (a) ceramic, (b) glycine-nitrate, (c) sol-gel and (d) freeze-drying methods used to obtain the perovskite compounds shown in the present work.

Fig. 2 .
Fig. 2. (a) X-ray diffraction patterns for the series Ln 1−x A x FeO 3−δ with x=0.2 to x=0.8, all obtained by the ceramic route.Rietveld fits to the X-ray diffraction data for samples LPS20 (b), LPSC50 (c) and LSC80 (d).

Fig. 3 .
Fig. 3. (a) Variation of the unit cell parameters and volume per formula unit with doping, x.(b) Dependence of the mean atomic distances and bond angles with doping, x.Shaded area indicates the x range where the structural transition takes place.

Fig. 6 .
Fig. 6.SEM micrographs taken on the surface of the La 0.05 Sm 0.45 Sr 0.18 Ba 0.32 FeO 3−δ phases obtained by the glycine-nitrate (gn) and solid state reaction (ss) methods.

Fig. 7 .
Fig. 7. (a) X-ray powder diffraction patterns at room temperature for all samples.(b) and (c) show the Rietveld fits to the X-ray powder diffraction patterns at room temperature for LB134-ss and LSS125-ss samples.The <r A > dependence of the lattice parameters, unit cell volume, main bond distances and Fe-O-Fe bond angles are shown in Figure8.As it can be observed, lattice parameters and unit cell volume decrease with decreasing the average A-site ionic radius.Given that the doping level has been fixed (x=0.5), changes in the ratio Fe 3+ /Fe 4+ and, therefore, in the Fesite ionic radius (<r Fe >), are not expected.Consequently, the decrease in lattice parameters is ascribed to the variation of <r A >.A more detailed analysis is given elsewhere(Ecija et. al, 2011).This also explains the decrease of the <A-O> and <A-Fe> mean distances with decreasing <r A > (Figure8b).Although <r Fe > has been kept constant, there is a slight reduction of the <Fe-O> distances as <r A > decreases, which is a consequence of the tilting in

Fig. 8 .
Fig. 8. <r A > dependence of (a) the unit cell parameters and volume per formula unit; and (b) the main average bond lengths (<A-Fe>, <A-O>, <Fe-O>) and Fe-O-Fe bond angle.Shaded area indicates the x range where the structural transition takes place.

Table 1 .
Nominal compositions, synthetic methods, and labels of the studied perovskites.