Novel Mg-Doped SrMoO3 Perovskites Designed as Anode Materials for Solid Oxide Fuel Cells

SrMo1−xMxO3−δ (M = Fe and Cr, x = 0.1 and 0.2) oxides have been recently described as excellent anode materials for solid oxide fuel cells at intermediate temperatures (IT-SOFC) with LSGM as the electrolyte. In this work, we have improved their properties by doping with aliovalent Mg ions at the B-site of the parent SrMoO3 perovskite. SrMo1−xMgxO3−δ (x = 0.1, 0.2) oxides have been prepared, characterized and tested as anode materials in single solid-oxide fuel cells, yielding output powers near 900 mW/cm−2 at 850 °C using pure H2 as fuel. We have studied its crystal structure with an “in situ” neutron power diffraction (NPD) experiment at temperatures as high as 800 °C, emulating the working conditions of an SOFC. Adequately high oxygen deficiencies, observed by NPD, together with elevated disk-shaped anisotropic displacement factors suggest a high ionic conductivity at the working temperatures. Furthermore, thermal expansion measurements, chemical compatibility with the LSGM electrolyte, electronic conductivity and reversibility upon cycling in oxidizing-reducing atmospheres have been carried out to find out the correlation between the excellent performance as an anode and the structural features.


Introduction
Solid oxide fuel cells at intermediate temperatures (IT-SOFC) are electrochemical devices able to convert the energy involved in the combustion of a fuel directly into electrical energy. IT-SOFCs work at intermediate temperatures, typically between 700˝C and 850˝C; therefore, the reaction kinetics is extremely favored, and the efficiency of the energy conversion process is very high, compared to other low-temperature fuel cells. The fuel oxidation reaction in SOFC happens in the anode. SOFCs often use anodes based on Ni-YSZ (yttria-stabilized zirconia) and Ni-LDC (lanthanum-dope ceria) cermets. These composite anodes have an excellent catalytic activity for the fuel-oxidation reaction and high electronic and ionic conductivity, but unfortunately, these materials promote carbon formation during the direct oxidation of hydrocarbon fuels and suffer from sintering problems during the cell operation [1][2][3]. Furthermore, Ni-based anodes have the disadvantage of being contaminated with H 2 S traces contained in H 2 [4].
In order to avoid the problems associated with the cermet-based anodes, single-phase active materials have been investigated with the ABO 3 perovskite structure. By suitably choosing stable oxide compounds in reducing atmospheres, these materials can provide enough electronic and ionic conductivity to perform as anodes in IT-SOFC. The SrMoO 3 cubic perovskite with Mo 4+ at the octahedral B positions has an extremely high electrical conductivity at room temperature (10 4 S¨cm´1 [5]); moreover, molybdenum is a very suitable element to catalyze the fuel-oxidation reaction. Unfortunately, this oxygen-stoichiometric oxide cannot exhibit the required the metal atoms and the anisotropic ones for oxygen atoms were also refined for the NPD data. The coherent scattering lengths for Sr, Mo, Mg and O were 7.02, 6.715, 5.375 and 5.805 fm, respectively.
Thermal analysis was carried out in a Mettler TA3000 system equipped with a TC15 processor unit. Thermogravimetric (TG) curves were obtained in a TG50 unit, working at a heating rate of 10˝C¨min´1, in an O 2 flow of 100 mL¨min´1 from 35-900˝C using about 50 mg of sample in each experiment.
Measurements of the thermal expansion coefficient and electrical conductivity required the use of sintered samples. For this purpose, pellets of SrMo 1´x Mg x O 3´δ (x = 0.1, 0.2) were prepared by pressing the powder in dies and sintering in air at 950˝C for 12 h; finally, the pellet was placed in a tube furnace with 5% H 2 /95% N 2 flow for 15 h at 900˝C. The densities of the pellets were around 70%-75% of the crystallographic value, calculated from the mass and geometrical volume. Thermal expansion of the sintered samples was carried out in a dilatometer Linseis L75/H, between 100 and 900˝C in H 2 (5%)/N 2 (95%). The conductivity was measured between 25 and 850˝C in H 2 (5%)/N 2 (95%), by the four-point method in bar-shaped pellets under DC currents of 100 mA. The currents were applied and collected with a Potenciostat-Galvanostat AUTOLAB PGSTAT 302, ECO CHEMIE.
Single-cell tests were made on electrolyte-supported cells with La 0.8 Sr 0.2 Ga 0.83 Mg 0.17 O 3´δ (LSGM) as the electrolyte, SrCo 0.8 Fe 0.2 O 3´δ (SCFO) as the cathode material and SrMo 1´x Mg x O 3´δ (SMMO) as anode material. The LSGM pellets of 20 mm in diameter were sintered at 1450˝C for 20 h and then polished with a diamond wheel to a thickness of 300 µm. La 0.4 Ce 0.6 O 2´δ (LDC) was used as a buffer layer between the anode and the electrolyte in order to prevent the interdiffusion of ionic species between perovskite and electrolyte. Inks of LDC, SMMO and SCFO were prepared with a binder (V-006 from Heraeus, Hanau, Germany). LDC ink was screen-printed onto one side of the LSGM disk followed by a thermal treatment at 1300˝C in air for 1 h. SMMO was subsequently screen printed onto the LDC layer and fired at 1100˝C in air for 1 h. SCFO was finally screen printed onto the other side of the disk and fired at 1050˝C in air for 1 h. The thickness of the anode and cathode was 10 µm. The working electrode area of the cell for both the anode and cathode was 0.25 cm 2 (0.5 cmˆ0.5 cm). Pt gauze with a small amount of Pt paste in separate dots was used as the current collector at both the anodic and the cathodic sides for ensuring electrical contact. The cells were tested in a vertical tubular furnace at 800 and 850˝C; the anode side was fed with pure H 2 , with a flow of 20 mL¨min´1, whereas the cathode worked in air. The fuel-cell tests were performed with an AUTOLAB 302N Potentiostat/Galvanostat by changing the voltage of the cell from 1.2-0.1 V, with steps of 0.010 V, holding 10 s at each step. Current density was calculated by the recorded current flux through the effective area of the cell (0.25 cm 2 ). Each VI (voltage-intensity) scan corresponds to one cycle; the activation of the cell was followed in subsequent cycles until the full power of the single cell was reached.

Crystallographic Characterization
The initial characterization of the products was carried out by XRD. SrMo 1´x Mg x O 3´δ (x = 0.1, 0.2) compounds were obtained as well-crystallized powders. The SrMoO 3 phase was also prepared as a reference. Figure 1 shows the XRD patterns of the SrMo 1´x Mg x O 3´δ (x = 0, 0.1 and 0.2) oxides. The XRD diagrams are characteristic of a cubic perovskite structure with the Pm-3m group. The unit-cell parameters obtained for x = 0, 0.1 and 0.2 are 3.9760(3), 3.9739(4) and 3.9654(2) Å, respectively. No impurity phases were detected in any samples.
In order to perform a more comprehensive structural study for the SrMo 1´x Mg x O 3´δ (x = 0.1 and 0.2) series, an investigation by NPD at room temperature (RT) for the SrMo 1´x Mg x O 3´δ family and high temperature (up to 800˝C) for SrMo 0.9 Mg 0.1 O 3´δ was carried out. The structures were refined in the Pm-3m group (No. 221), with Z = 1. Sr atoms are located at the 1b ( 1 /2, 1 /2 , 1 /2 ) position; Mo and Mg atoms are randomly distributed at 1a (0, 0, 0) sites; and the O oxygen atoms are placed at the 3d ( 1 /2 , 0, 0) position. A small oxygen deficiency was observed at room temperature after refining the occupancy factors of the oxygen atoms.  In order to perform a more comprehensive structural study for the SrMo1−xMgxO3−δ (x = 0.1 and 0.2) series, an investigation by NPD at room temperature (RT) for the SrMo1−xMgxO3−δ family and high temperature (up to 800 °C) for SrMo0.9Mg0.1O3−δ was carried out. The structures were refined in the Pm-3m group (No. 221), with Z = 1. Sr atoms are located at the 1b (½, ½, ½) position; Mo and Mg atoms are randomly distributed at 1a (0, 0, 0) sites; and the O oxygen atoms are placed at the 3d (½, 0, 0) position. A small oxygen deficiency was observed at room temperature after refining the occupancy factors of the oxygen atoms.
Sr2FeMoO6) reported to have Mo 5+ ions, exhibiting Mo 5+ -Mo 6+ mixed valence [12]. Similar unit-cell contraction was observed in previous studies of SrMoO3 doped with 10%, 20% and 30% Fe, where the ionic size of high-spin Fe 3+ is practically the same as Mo 4+ , and the cell is considerably shrunken [6] at room temperature. On the other hand, the oxygen occupancy also evolves with Mg 2+ doping, being slightly deficient for x = 0.1 (2.985(3) O per formula unit) and significantly more deficient for x = 0.2 (2.856(3) per formula unit) at room temperature.  The unit-cell parameters decrease as the amount of Mg in the sample increases. The (Mo,Mg)-O1 bond lengths at room temperature decrease accordingly with Mg-doping from 1.98814(1) Å for the undoped sample to 1.98247(3) Å for the sample with x = 0.2. This happens even though the ionic size of Mg 2+ (0.72 Å) is higher than Mo 4+ (0.65 Å) [8]. This fact may suggest that a unit-cell contraction is happening because oxygen vacancies are being created when Mo is partially replaced by Mg, but it is more probable that this cell contraction is related to a partial oxidation of Mo ions (hole doping effect) as Mg 2+ is introduced into the perovskite, resulting in a mixed-valence state Mo 4+ -Mo 5+ proportional to the doping rate. There are well-known Mo-containing double perovskites (e.g., Sr 2 FeMoO 6 ) reported to have Mo 5+ ions, exhibiting Mo 5+ -Mo 6+ mixed valence [12]. Similar unit-cell contraction was observed in previous studies of SrMoO 3 doped with 10%, 20% and 30% Fe, where the ionic size of high-spin Fe 3+ is practically the same as Mo 4+ , and the cell is considerably shrunken [6] at room temperature. On the other hand, the oxygen occupancy also evolves with Mg 2+ doping, being slightly deficient for x = 0.1 (2.985(3) O per formula unit) and significantly more deficient for x = 0.2 (2.856(3) per formula unit) at room temperature.
The thermal evolution of the crystal structure under the anode conditions of an SOFC was studied by NPD for the x = 0.1 oxide. The NPD patterns are illustrated in Figure 3. No structural transitions in the temperature range under study (25-800˝C) were found. The thermal evolution of the crystal structure under the anode conditions of an SOFC was studied by NPD for the x = 0.1 oxide. The NPD patterns are illustrated in Figure 3. No structural transitions in the temperature range under study (25-800 °C) were found.   Table 2 includes the structural parameters after the refinement of the SrMo0.9Mg0.1O3−δ structure at the different temperatures under study.    The thermal evolution of the crystal structure under the anode conditions of an SOFC was studied by NPD for the x = 0.1 oxide. The NPD patterns are illustrated in Figure 3. No structural transitions in the temperature range under study (25-800 °C) were found.   Table 2 includes the structural parameters after the refinement of the SrMo0.9Mg0.1O3−δ structure at the different temperatures under study.    (3) * Anisotropic betas (ˆ10 4 ); β 12 = β 13 = β 23 = 0. Figure 5a shows the temperature variation of the unit-cell parameters for SrMo 0.9 Mg 0.1 O 3´δ . The unit-cell parameters monotonically increase when heating the sample due to the expansion of the chemical bonds. The thermal evolution of the oxygen content in air was also studied by neutron diffraction. Figure 5b Figure 6 shows the crystal structure of SrMo 1´x Mg x O 3´δ highlighting the evolution of the anisotropic displacements between 200 and 800˝C, with 95% probability for the O nuclear density. At 800˝C, the root mean square (r.m.s.) displacements of O are 0.194 Å perpendicular to the Mo-Mo distance and 0.117 Å parallel to it. The disk-shaped ellipsoids are the result of the strong covalent bonding between Mo 4+ -Mo 5+ and O; SrMoO 3 is well known to exhibit band conduction properties by virtue of the robust covalent mixing between 4d Mo orbitals and O 2p oxygen orbitals, strongly overlapping across 180˝Mo-O-Mo angles. Such strong chemical bonds impede the thermal motion along the bonds, in such a way that O atoms exhibit degrees of freedom in the plane perpendicular to the bonding direction. This is in contrast with the prolate ellipsoids observed in other MIEC oxides, like Ba 0.9 Co 0.7 Fe 0.2 Nb 0.1 O 3´δ [13], which suggests a breathing of the (Co,Fe,Nb)O 6 octahedra upon the migration of the oxygen vacancies across the solid. In that case, the average (Co,Fe) oxidation state varies between 2.84+ and 2.02+ in the 25-800˝C temperature range, thus involving much less covalent chemical bonds within the perovskite octahedra, which make possible the less-frequent prolate kind of thermal ellipsoids. with the prolate ellipsoids observed in other MIEC oxides, like Ba0.9Co0.7Fe0.2Nb0.1O3−δ [13], which suggests a breathing of the (Co,Fe,Nb)O6 octahedra upon the migration of the oxygen vacancies across the solid. In that case, the average (Co,Fe) oxidation state varies between 2.84+ and 2.02+ in the 25-800 °C temperature range, thus involving much less covalent chemical bonds within the perovskite octahedra, which make possible the less-frequent prolate kind of thermal ellipsoids.

Thermal Analysis
The oxidation of the samples by incorporation of oxygen was followed by thermogravimetric analysis carried out in O2 flow from 35-900 °C. Figure 7 shows the TGA curves for the SrMo1−xMgxO3−δ (x = 0.1 and 0.2) samples. The curves indicate an incorporation of 0.67 oxygen atoms per formula unit for the sample with x = 0.1 and 0.49 oxygens for x = 0.2. As the samples are heated, the oxidation of the perovskite compounds is produced, resulting in crystalline phases with a scheelite-type structure. with the prolate ellipsoids observed in other MIEC oxides, like Ba0.9Co0.7Fe0.2Nb0.1O3−δ [13], which suggests a breathing of the (Co,Fe,Nb)O6 octahedra upon the migration of the oxygen vacancies across the solid. In that case, the average (Co,Fe) oxidation state varies between 2.84+ and 2.02+ in the 25-800 °C temperature range, thus involving much less covalent chemical bonds within the perovskite octahedra, which make possible the less-frequent prolate kind of thermal ellipsoids.

Thermal Analysis
The oxidation of the samples by incorporation of oxygen was followed by thermogravimetric analysis carried out in O2 flow from 35-900 °C. Figure 7 shows the TGA curves for the SrMo1−xMgxO3−δ (x = 0.1 and 0.2) samples. The curves indicate an incorporation of 0.67 oxygen atoms per formula unit for the sample with x = 0.1 and 0.49 oxygens for x = 0.2. As the samples are heated, the oxidation of the perovskite compounds is produced, resulting in crystalline phases with a scheelite-type structure.

Thermal Analysis
The oxidation of the samples by incorporation of oxygen was followed by thermogravimetric analysis carried out in O 2 flow from 35-900˝C. Figure 7

Thermal Expansion Measurements
In order to probe the mechanical compatibility of our materials with the other cell components, thermal expansion measurements in dense samples were performed in a 5% H2/95% N2 atmosphere. The dilatometric analysis was carried out between 25 and 900 °C for several cycles; the data were

Thermal Expansion Measurements
In order to probe the mechanical compatibility of our materials with the other cell components, thermal expansion measurements in dense samples were performed in a 5% H2/95% N2 atmosphere. The dilatometric analysis was carried out between 25 and 900 °C for several cycles; the data were

Thermal Expansion Measurements
In order to probe the mechanical compatibility of our materials with the other cell components, thermal expansion measurements in dense samples were performed in a 5% H 2 /95% N 2 atmosphere.
The dilatometric analysis was carried out between 25 and 900˝C for several cycles; the data were only recorded during the heating process. Figure 9 shows the thermal expansion for SrMo 1´x Mg x O 3-δ (x = 0.1 and 0.2) and SrMo 1´x Mg x O 4-δ (x = 0.1 and 0.2). No abrupt changes in the entire temperature measuring range were found. TECs measured in 5% H 2 /95% N 2 atmosphere for perovskite phases and an air atmosphere for scheelite phases between 400 and 850˝C are included in Figure 9. The TEC value for SrMo 0.9 Mg 0.1 O 3´δ is in concordance with that obtained from NPD data in the heating run, of 10.93ˆ10´6 K´1.
Materials 2016, 9,588 10 of 14 only recorded during the heating process. Figure 9 shows the thermal expansion for SrMo1−xMgxO3-δ (x = 0.1 and 0.2) and SrMo1−xMgxO4-δ (x = 0.1 and 0.2). No abrupt changes in the entire temperature measuring range were found. TECs measured in 5% H2/95% N2 atmosphere for perovskite phases and an air atmosphere for scheelite phases between 400 and 850 °C are included in Figure 9. The TEC value for SrMo0.9Mg0.1O3−δ is in concordance with that obtained from NPD data in the heating run, of 10.93 × 10 −6 K −1 . The TECs obtained for the perovskite and scheelite phases are reasonably similar and fit with the general SOFC electrolytes values, so no mechanical compatibility problems should be expected during the oxidation-reduction cycles. For the x = 0.1 compound, the TEC coefficients for SrMo1−xMgxO3−δ and SrMo1−xMgxO4−δ are indeed very similar, exhibiting values of 11.74 × 10 −6 and 11.23 × 10 −6 •K −1 , respectively. For x = 0.2, there is a bigger difference (10.64 × 10 −6 and 13.94 × 10 −6 •K −1 , respectively), which could induce a certain redox instability. Figure 10 shows the thermal variation of the electrical conductivity of SrMo1−xMgxO3−δ (x = 0.1 and 0.2). The resistance was measured by the dc four-probe method; a current of 100 mA was applied, and the potential drop was recorded in an Autolab 302N Potentiostat-Galvanostat. Figure 10 illustrates the reduced phases with the perovskite structure featuring a metallic-like conductivity under reducing conditions in both cases.  The TECs obtained for the perovskite and scheelite phases are reasonably similar and fit with the general SOFC electrolytes values, so no mechanical compatibility problems should be expected during the oxidation-reduction cycles. For the x = 0.1 compound, the TEC coefficients for SrMo 1´x Mg x O 3´δ and SrMo 1´x Mg x O 4´δ are indeed very similar, exhibiting values of 11.74ˆ10´6 and 11.23ˆ10´6¨K´1, respectively. For x = 0.2, there is a bigger difference (10.64ˆ10´6 and 13.94ˆ10´6¨K´1, respectively), which could induce a certain redox instability. Figure 10 shows the thermal variation of the electrical conductivity of SrMo 1´x Mg x O 3´δ (x = 0.1 and 0.2). The resistance was measured by the dc four-probe method; a current of 100 mA was applied, and the potential drop was recorded in an Autolab 302N Potentiostat-Galvanostat. Figure 10 illustrates the reduced phases with the perovskite structure featuring a metallic-like conductivity under reducing conditions in both cases.

Chemical Compatibility
The chemical compatibility of SrMo1−xMgxO3−δ series with La0.8Sr0.2Ga0.83Mg0.17O3−δ (LSGM) electrolyte has been studied by mixing of both powdered samples and heating the mixture at 900 °C under H2/N2 (5%/95%) atmosphere for 24 hours. Figure 11 shows the Rietveld analysis of SrMo0.9Mg0.1O3−δ, consisting of a mixture of both unchanged phases, so no unwanted secondary phases will be formed during the operation in single cells. The same result was obtained for the compound with x = 0.2. Figure 11. Rietveld-refined XRD profiles of a mixture of LSGM and SrMo0.9Mg0.1O3−δ after a thermal treatment at 900 °C in H2(5%)/N2, showing no reaction products between both phases other than the initial reactants. The first and second series of Bragg positions correspond to LSGM and SrMo0.9Mg0.1O3−δ, respectively.

Fuel-Cell Tests
In order to study the behavior of SrMo1−xMgxO3−δ (x = 0.1 and 0.2) as anodes in solid oxide fuel cells, a single cell for each sample was prepared in an electrolyte-supported configuration using a 300 µm-thick LSGM electrolyte, and the output power was measured at 800 and 850 °C. Figure 12

Chemical Compatibility
The chemical compatibility of SrMo 1´x Mg x O 3´δ series with La 0.8 Sr 0.2 Ga 0.83 Mg 0.17 O 3´δ (LSGM) electrolyte has been studied by mixing of both powdered samples and heating the mixture at 900˝C under H 2 /N 2 (5%/95%) atmosphere for 24 h. Figure 11 shows the Rietveld analysis of SrMo 0.9 Mg 0.1 O 3´δ , consisting of a mixture of both unchanged phases, so no unwanted secondary phases will be formed during the operation in single cells. The same result was obtained for the compound with x = 0.2.

Chemical Compatibility
The chemical compatibility of SrMo1−xMgxO3−δ series with La0.8Sr0.2Ga0.83Mg0.17O3−δ (LSGM) electrolyte has been studied by mixing of both powdered samples and heating the mixture at 900 °C under H2/N2 (5%/95%) atmosphere for 24 hours. Figure 11 shows the Rietveld analysis of SrMo0.9Mg0.1O3−δ, consisting of a mixture of both unchanged phases, so no unwanted secondary phases will be formed during the operation in single cells. The same result was obtained for the compound with x = 0.2. Figure 11. Rietveld-refined XRD profiles of a mixture of LSGM and SrMo0.9Mg0.1O3−δ after a thermal treatment at 900 °C in H2(5%)/N2, showing no reaction products between both phases other than the initial reactants. The first and second series of Bragg positions correspond to LSGM and SrMo0.9Mg0.1O3−δ, respectively.

Fuel-Cell Tests
In order to study the behavior of SrMo1−xMgxO3−δ (x = 0.1 and 0.2) as anodes in solid oxide fuel cells, a single cell for each sample was prepared in an electrolyte-supported configuration using a 300 µm-thick LSGM electrolyte, and the output power was measured at 800 and 850 °C. Figure 12

Fuel-Cell Tests
In order to study the behavior of SrMo 1´x Mg x O 3´δ (x = 0.1 and 0.2) as anodes in solid oxide fuel cells, a single cell for each sample was prepared in an electrolyte-supported configuration using a 300 µm-thick LSGM electrolyte, and the output power was measured at 800 and 850˝C. Figure 12 illustrates the cell voltage and power density as a function of current density at these temperatures for the single cell fed with pure H 2 for the x = 0.1 anode. The maximum power densities generated by the cell were 684 and 887 mW/cm 2 , respectively. Materials 2016, 9, 588 12 of 14 illustrates the cell voltage and power density as a function of current density at these temperatures for the single cell fed with pure H2 for the x = 0.1 anode. The maximum power densities generated by the cell were 684 and 887 mW/cm 2 , respectively.  Figure 13 shows the cell voltage and power density as a function of current density at the same temperatures for the anode x = 0.2. The maximum power densities generated by the cell were 555 and 832 mW/cm 2 , respectively. The inset of Figure 13 illustrates a view of the cathode side of the cell. Although both anodes have an exceptional behavior, a slight decrease of the output power of the single cells is observed for x = 0.2 with respect to the x = 0.1 anode. This reduction of the power density could be related to the decrease in the Mo contents of the anode in the x = 0.2 sample, since apparently, molybdenum is responsible for the catalytic oxidation of the fuel, as has been observed in other Mo-containing anodes [6,14]. Additionally, the observed reduction of the electrical conductivity ( Figure 10) in the whole range of measured temperatures also contributes to the deterioration of the output power for this anode material.  Figure 13 shows the cell voltage and power density as a function of current density at the same temperatures for the anode x = 0.2. The maximum power densities generated by the cell were 555 and 832 mW/cm 2 , respectively. The inset of Figure 13 illustrates a view of the cathode side of the cell. illustrates the cell voltage and power density as a function of current density at these temperatures for the single cell fed with pure H2 for the x = 0.1 anode. The maximum power densities generated by the cell were 684 and 887 mW/cm 2 , respectively.  Figure 13 shows the cell voltage and power density as a function of current density at the same temperatures for the anode x = 0.2. The maximum power densities generated by the cell were 555 and 832 mW/cm 2 , respectively. The inset of Figure 13 illustrates a view of the cathode side of the cell. Although both anodes have an exceptional behavior, a slight decrease of the output power of the single cells is observed for x = 0.2 with respect to the x = 0.1 anode. This reduction of the power density could be related to the decrease in the Mo contents of the anode in the x = 0.2 sample, since apparently, molybdenum is responsible for the catalytic oxidation of the fuel, as has been observed in other Mo-containing anodes [6,14]. Additionally, the observed reduction of the electrical conductivity ( Figure 10) in the whole range of measured temperatures also contributes to the deterioration of the output power for this anode material. Although both anodes have an exceptional behavior, a slight decrease of the output power of the single cells is observed for x = 0.2 with respect to the x = 0.1 anode. This reduction of the power density could be related to the decrease in the Mo contents of the anode in the x = 0.2 sample, since apparently, molybdenum is responsible for the catalytic oxidation of the fuel, as has been observed in other Mo-containing anodes [6,14]. Additionally, the observed reduction of the electrical conductivity ( Figure 10) in the whole range of measured temperatures also contributes to the deterioration of the output power for this anode material.
In a previous work [7], an additional test using Au gauze with a small amount of Au paste as the current collector instead of Pt gauze was carried out to check if Pt could promote the catalytic process of O 2 reduction or fuel oxidation as suggested by some authors [15][16][17], increasing the power density and covering up the true activity of the oxides selected as electrodes. In this work, the maximum power densities generated by the cell were even higher than with Pt gauze. Since Au has no catalytic properties, this test implies that the observed activity comes entirely from the anode material.
In order to compare the performance of our SrMo 1´x Mg x O 3´δ (x = 0.1 and 0.2) anodes with other SrMo 1´x M x O 3´δ (M = Fe and Cr) anodes, in previous works [6,7], an identical single cell with these anodes was also made and measured. Similar power outputs were observed in these cases (874 mW/cm 2 for SrMo 0.9 Fe 0.1 O 3´δ and 695 mW/cm 2 for SrMo 0.9 Cr 0.1 O 3´δ at 850˝C), demonstrating that our anodes are even slightly better than these materials. Moreover, in the long-term performance, the Mg 2+ -doped anodes are believed to be superior due to the absence of interdiffusion cationic effects, since Mg is also contained in the LSGM electrolyte.

Conclusions
In this study, we have shown that SrMo 1´x Mg x O 3´δ (x = 0.1 and 0.2) oxides crystallize in a cubic perovskite structure (Pm-3m) where a mixed Mo 4+ -Mo 5+ oxidation state is present at RT; NPD data unveil the creation of an appreciable amount of oxygen vacancies at high temperatures, under the low pO 2 working conditions of an SOFC. The anisotropic displacements for O atoms, conforming flattened ellipsoids, correspond to the highly covalent Mo-O bonds. SrMo 1´x Mg x O 3´δ (x = 0.1 and 0.2) oxides can be successfully used as anode materials in SOFC test cells in an electrolyte-supported configuration using a 300 µm-thick LSGM electrolyte. Excellent maximum output powers of 887 and 832 mW/cm 2 are obtained for x = 0.1, 0.2, respectively, at 850˝C, using pure H 2 as a fuel. The sufficiently large number of oxygen vacancies combined with high thermal displacement factors suggest a high ionic conductivity at the operating temperatures, constituting MIEC-type materials together with the high electronic conductivity associated with the pristine SrMoO 3 sample. In addition, the reversibility of the reduction-oxidation between the Sr(Mo,Mg)O 4´δ scheelite and Sr(Mo,Mg)O 3´δ perovskite phases makes possible the required cyclability of the cells. The obtained TECs, ranging between 13.94ˆ10´6 and 10.64ˆ10´6 K´1, are perfectly compatible with the usual SOFC electrolytes. Finally, excellent chemical compatibility was observed with the electrolyte LSGM for 24 h at 900˝C.