Cermet cathodes for strontium and magnesium-doped LaGaO3-based solid oxide fuel cells

Materials Chemistry and Physics 114 (2009) 356–361

Contents lists available at ScienceDirect

Materials Chemistry and Physics journal homepage: www.elsevier.com/locate/matchemphys

Cermet cathodes for strontium and magnesium-doped LaGaO3 -based solid oxide fuel cells Pradyot Datta a,d,∗ , D.I. Bronin b , P. Majewski c , F. Aldinger d a

Technische Universität Clausthal, Institut für Metallurgie, 42 Robert-Koch Strasse, 38678 Clausthal-Zellerfeld, Germany Institute of High-Temperature Electrochemistry of Russian Academy of Sciences, Ekaterinburg 620219, S. Kovalevskoz 22, Russia c University of South Australia, School of Advanced Manufacturing and Mechanical Engineering, Mawson Institute, Mawson Lakes, South Australia 5095, Australia d Max-Planck-Institut für Metallforschung and Institut für Nichtmetallische and Anorganische Materialien, Universität Stuttgart, Pulvermetallurgisches Laboratorium, Heisenbergstrasse 3, Stuttgart 70569, Germany b

a r t i c l e

i n f o

Article history: Received 14 March 2008 Received in revised form 31 August 2008 Accepted 13 September 2008 Keywords: Oxide materials Thermal expansion Photoelectron spectroscopies X-ray diffraction Polarization conductivity

a b s t r a c t To check the suitability of La0.9 Sr0.1 Ga0.85 Mg0.15 O3−ı –Ag cermets as cathode material for solid oxide fuel cell (SOFC) with Sr- and Mg-doped LaGaO3 electrolyte a series of cermets with different Ag contents were prepared by conventional sintering process. The chemical compatibility between La0.9 Sr0.1 Ga0.85 Mg0.15 O3−ı (LSGM) and Ag was investigated by X-ray diffraction, scanning electron microscopy, and X-ray photoelectron spectroscopy. Thermal expansion coefficient of the cermets was measured as a function of Ag content and was found to increase with increasing metallic content. Oxygen adsorption at the surface of the cermets could be detected but no reaction or solid solubility between LSGM and Ag was found. It was noticed that a minimum of 30 wt.% Ag is needed to form a cermet with percolating network. From impedance spectroscopy measurement activation energy for the polarization conductance was found to be around 110 kJ mol−1 . © 2008 Elsevier B.V. All rights reserved.

1. Introduction Solid oxide fuel cells (SOFC) have offered an alternative source of producing electricity with the potential usage in a variety of commercial and industrial applications because of its high efficiency, emission of low pollutants, low noise and potential for cogeneration. Oxygen ion conducting electrolyte materials are the key components of a SOFC. Among the solid electrolytes yttria-stabilized zirconia (YSZ) has been most extensively studied. However, YSZ requires operation temperature as high as about 1000 ◦ C, which is associated with problems like expensive constructional and interconnect materials. This has lead to the development of intermediate temperature SOFC operation at 600–800 ◦ C. A plethora of research activities focus the reduction of the operation temperature of the SOFC below 800 ◦ C without reducing the efficiency of SOFC. Perovskite LaGaO3 doped with Sr2+ and Mg2+ at the A- and B-site, respectively (LSGM) is the most promising material for the intermediate temperature operation of SOFC because of its higher ionic conductivity as compared to YSZ [1,2].

∗ Corresponding author at: Technische Universität Clausthal, Institut für Metallurgie, 42 Robert-Koch Strasse, 38678 Clausthal-Zellerfeld, Germany. Tel.: +49 5323 723688; fax: +49 5323 723184. E-mail address: [email protected] (P. Datta). 0254-0584/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.matchemphys.2008.09.038

However, to tap the full potential of this electrolyte proper electrode materials need to be provided, because lowering the temperature increases the overpotential of the electrodes [3]. A fundamental requirement for the successful operation of SOFC is the thermodynamic stability between the cathode and the electrolyte [4]. Normally used cathode material for high temperature operation La1−x Srx MnO3 (LSM) suffers from very low ionic conductivity [5]. In order to increase the efficiency of the cell other materials with mixed conductivity should be chosen so that the oxygen reaction can be extended over a larger surface of the cathode rather than limiting it only at the triple phase boundary area. La1−x Srx CoO3 (LSC) satisfies the criterion of having higher oxygen ion conductivity and therefore a higher rate of surface oxygen exchange but it has higher thermal expansion coefficient (TEC) as compared to the electrolyte. Moreover, the ionic conductivity of this material drops rapidly with decreasing temperature necessitating the need of composite cathode for the operation of intermediate temperature SOFC [6]. The performance of La(Sr)MnO3 (LSM)–YSZ composite cathode material for SOFC operating temperatures above about 800 ◦ C was reported to be improved due to the suppression of the growth of LSM particles by YSZ particles thereby maintaining the porosity and increasing the triple phase boundary length (TPBL) [7,8]. Tanner et al. [9] using a model reported the effect of porous composite electrodes on the overall charge-transfer process and predicted that composite electrodes can significantly improve the performance of

P. Datta et al. / Materials Chemistry and Physics 114 (2009) 356–361

fuel cells. Indeed, the overall polarization of the cathode is reported to be decreased in composite cathodes [10]. Not only that the oxygen permeability of composite cathode materials is found to be higher than that of pure oxides [11,12]. Ideally, the use of porous mixtures of an electro-catalyst and electrolyte material as cathode material should improve the performance as a result of increase in the TPBL. In principle, platinum, gold, and palladium could be incorporated into electrode materials. While the cost of these metals is prohibitive for commercial applications, silver is much less expensive and sufficiently active for oxygen reduction [13]. BaCe0.8 Gd0.2 O3 (BCG)–silver composites were reported to be promising cathode materials for intermediate temperature SOFC using BaCeO3 -based electrolytes [14,15]. Results indicated that the electrochemical properties of these composites are quite sensitive to the composition and microstructure of the electrodes. Wang and Barnett [16] reported that the resistivity of the composite cathodes like Ag–La(Sr)CoO3 and Ag–LSM decreases with increasing Ag content and the performance of the composites is superior to that of pure silver. Other authors [17–21] also reported encouraging results with silver-electrolyte composite cathodes. In this work, the possibility of using LSGM–Ag cermet as cathode material is explored as it satisfies all the conditions necessary to be a cathode material. Oxygen permeation in a composite is realized by the means of oxygen transport through the oxide phase and of electron migration through the metal phase. Silver is chosen as it is reported to be catalytically active towards surface oxygen exchange, which is one of the most important requirements that must be met by the metal phase. One of the most important criteria for successful operation of the fuel cell is the compatibility between the cathode material and the electrolyte. Any chemical reaction or inter-diffusion between the ceramic and the metallic phase would shrink the triple phase boundary area, which is the preferable cathode reaction site, and this in turn would adversely affect the performance of the SOFC. The chemical compatibility between LSGM and Ag has been investigated by means of scanning electron microscopy (SEM), energy dispersive X-ray analysis (EDX) and X-ray photoelectron spectroscopy (XPS). TEC of LSGM–Ag cermets as a function of Ag content has also been determined. The reported electrical and electrochemical characteristics behaviour of the cermets has been measured by dc four probe and ac impedance spectroscopy. 2. Experimental procedure Samples with overall nominal composition of La0.90 Sr0.10 Ga0.85 Mg0.15 O3−ı were prepared by solid state synthesis starting from powders of La2 O3 (99.99%, Sigma–Aldrich, Steinheim, Germany) SrCO3 (98%+, Sigma–Aldrich, Steinheim, Germany), Ga2 O3 (99.99%, Sigma–Aldrich, Steinheim, Germany) and MgO (98%+, Merck, Darmstadt, Germany). Detailed sample preparation procedure is given in another report [22]. Along with the obtained LSGM powder with a particle size of few microns varying amounts of silver powder (10–50 wt.%) were mixed separately and milled in a WC ball mill (average particle size <160 ␮m) for 10 min to get a homogenize mixture. The powder mixtures were then isostatically cold pressed at 625 MPa for 60 s into rectangular compacts with the dimension of 15 mm × 5 mm × 5 mm. The pressed samples were then annealed at different temperatures starting from 600 to 900 ◦ C for durations of 20–100 h in static air atmosphere. Heating rate for all cases was 5 ◦ C min−1 and all samples were furnace cooled to room temperature. The phase distribution of the calcined powders as well as sintered samples was studied by powder X-ray diffraction analysis (XRD) using Cu K␣1+2 radiation with 40 kV acceleration voltage, 25 mA filament current (D5000, Kristalloflex, Siemens, Germany). Crushing and grinding of the compacts was performed in an agate mortar. Fine grained powder, was sprayed evenly on a substrate. Diffraction data were smoothened and the background and Cu K␣2 component were removed using the Siemens software package, DiffracAT, EVA5.0 rev.1. Microstructural characterization was performed by a scanning electron microscope (SEM Model REM S200, Cambridge Instruments, UK) using a Tungsten anode with an acceleration voltage of 20 kV. Chemical analysis was carried out by a built-in EDX spectrometer (Model AN10000 pentafect detector, Link Systems, High Wycombe, UK). Specimens for SEM were embedded in a carbon containing epoxy

357

polymer and polished with 1 ␮m diamond suspension on a soft polishing cloth (DP-NAP, Struers, Copenhagen, Denmark). The density of the sintered samples was measured by Helium Pycnometry (Micromeritics, AccuPyc 1330, Australia), whereas the porosity was measured by conventional mercury porosimetry (Denver Instrument GmbH, Göttingen, Germany). Linear TEC of the samples was measured after annealing at 600 ◦ C for 100 h. The measurement was carried out using a bar of sapphire of the dimension of 10 mm × 4 mm × 4 mm in air atmosphere using a push rod type differential dilatometer (Model 802, Bähr-Thermoanalyse GmbH, Hüllhorst, Germany) in the temperature range from room temperature to 900 ◦ C. For the dc four probe measurements electrolyte samples were prepared in the form of 10 mm × 4 mm × 4 mm pellets. Two current electrodes (porous Pt) were made by painting porous Pt on the two opposite faces of the samples. The potential probes were made of Pt wires (0.2 mm diameter) spooled around the sample. Potential probes were smeared with Pt paste in order to provide better electrical contact with the sample. Resistance of an electrolyte was calculated by means of Ohm law: R=

U+ + U− , I+ + I−

where U+ and U− are absolute values of the voltages, which were measured when current was passed in forward (I+ ) and reverse (I− ) directions through the sample. Specific conductivity of electrolyte was calculated by equation:

 

=

L S

× R−1 ,

where L is the distance between the potential probes; S is the cross-section area of the sample. Experimental error of this method mainly depends on the precision of measurements of geometrical parameters of the sample and the distance between measuring probes, and it was not more than 5%. For the impedance spectroscopy measurements the sintered LSGM pellets were 1 mm thick and 18 mm in diameter. LSGM–Ag powders along with around 5 wt.% Bi2 O3 were thoroughly mixed in ethyl alcohol with polyvinyl butyral as binder to make a paste. The electrode paste was painted smoothly on both the surfaces of LSGM electrolyte. The pellets were first air-dried. Then they were fired in air at 900 ◦ C for 1 h. Bi2 O3 was added as sintering additive and adhesion of the electrode to electrolyte was good. Ag meshes were attached to both electrodes and co-sintered with them so as to act as current collectors. Lead wires of Pt were used to connect the Ag mesh current collectors to the ac impedance spectrometer (IM6, Zahner Elektrik). Electrochemical measurements were carried out in a frequency range 1 Hz to 100 kHz with an applied variable voltage of 10 mV. All the measurements were taken in air atmosphere in a temperature range of 450–800 ◦ C. The electrochemical cell was connected to impedancemeter by the twoelectrode four-wire mode that allowed to exclude the impedance of current carrying leads from a total impedance of the system. Specific values of polarization conductivity were calculated by the formula:  = 2[S(Rdc − Rhf )]

−1

,

where S is the area of an electrode, Rdc is the total resistance of an electrochemical cell to a direct current and Rhf is the resistance determined by extrapolation of a high-frequency part of impedance hodograph on real axis, which corresponds to resistance of an electrolyte. XPS technique was employed to identify minute reactions between the ceramic phase and Ag. The XPS analysis was performed with a Thermo VG Thetaprobe system operating in the parallel data acquisition mode using monochromatic Al K␣ (h = 1486.68 eV; spot size 400 ␮m). Experimental detail is given elsewhere [23].

3. Results 3.1. XRD and SEM study LSGM–Ag powder mixtures compacted and annealed at various temperatures. Fig. 1 shows the XRD patterns of the LSGM–Ag cermets after annealing at 600 ◦ C for 100 h. The LSGM parent phase, which has a perovskite structure, is marked as p in the pattern. Ag is also indexed. In addition to the peaks of these two phases some small ones are also visible which are indexed as LaSrGa3 O7 (JCPDS 45-637) and LaSrGaO4 (JCPDS 24-1208). A typical SEM picture of LSGM–50Ag after 100 h of annealing is shown in Fig. 2. The percolation of Ag phase in the LSGM is clearly visible. Through EDX analysis the dark phase is identified as LSGM and the bright one is Ag. No trace of Ag in the LSGM matrix or any of the constituent elements of LSGM in Ag is found. It is noted here that purpose of doing SEM was to find out the different phases present in the picture and

358

P. Datta et al. / Materials Chemistry and Physics 114 (2009) 356–361

Fig. 1. XRD patterns of LSGM–Ag cermets after annealing at 600 ◦ C for 100 h. (a) LSGM–20Ag, (b) LSGM–30Ag, (c) LSGM–40Ag and (d) LSGM–50Ag (p = perovskite, * = Ag, ◦ = LaSrGa3 O7 , and 1 = LaSrGaO4 ).

to do EDX in various phase fields so as to check the homogeneity of that phase. The figure clearly serves that purpose. For the sake of brevity SEM picture of only one representative composition is shown which clearly demonstrates that silver and LSGM electrolyte do not agglomerate and they are uniformly distributed in the electrode composite. To study the effect of temperature on the cermets annealing was also performed at 800 and 900 ◦ C. After 800 ◦ C annealing, the phase composition of the cermets does not change as can be seen from Fig. 3. To accelerate processes of possible interaction between composite phases it was decided to anneal the cermets at 900 ◦ C, which is close to the melting point of silver (960.5 ◦ C). At such a high temperature even a short time annealing will give a fair idea about the stability of the material for long time at lower temperatures. XRD patterns of cermets after annealing at 900 ◦ C for 20 h are shown in Fig. 4. It shows no indication of the formation of any other phase other than the ones already visible at lower temperatures. SEM picture also does not show any pronounced grain growth.

Fig. 3. XRD patterns of LSGM–Ag cermets after annealing at 800 ◦ C for 100 h. (a) LSGM–20Ag, (b) LSGM–30Ag, (c) LSGM–40Ag, and (d) LSGM–50Ag (p = perovskite, * = Ag, ◦ = LaSrGa3 O7 , and 1 = LaSrGaO4 ).

Fig. 4. XRD patterns of LSGM–Ag cermets after annealing at 900 ◦ C for 20 h. LSGM–10Ag, (b) LSGM–20Ag, (c) LSGM–30Ag, (d) LSGM–40Ag and (e) LSGM–50Ag (p = perovskite, * = Ag, ◦ = LaSrGa3 O7 , and 1 = LaSrGaO4 ).

3.2. Thermal expansion coefficient (TEC) Matching TEC of all the components of a SOFC is an important requirement for smooth running and long-term stability of such a system. TEC of LSGM–Ag cermets was measured up to a temperature of 900 ◦ C and the result is given in Fig. 5. The figure indicates

Fig. 2. SEM photograph of LSGM–50Ag cermet after annealing at 600 ◦ C for 100 h. (Bright phase is silver).

that as the temperature increases TEC also increases for all the cermets irrespective of their metallic contents. One more noticeable characteristic of the TEC curves is that the rate of increment is higher at lower temperature levels than at temperatures higher

Fig. 5. Variation of TEC of LSGM–Ag cermets with temperature. (a) LSGM–10Ag, (b) LSGM–20Ag, (c) LSGM–30Ag, (d) LSGM–40Ag and (e) LSGM–50Ag.

P. Datta et al. / Materials Chemistry and Physics 114 (2009) 356–361

Fig. 6. Variation of TEC of LSGM with variation of Ag contents at (a) 600 ◦ C and (b) 800 ◦ C.

359

Fig. 7. Variation of porosity of LSGM with silver after annealing at 600 ◦ C at various durations.

than about 250 ◦ C. Also, as the intended operation temperature of SOFC is 600–800 ◦ C, the absolute values of the TEC of the cermets along with the TEC value of LSGM are plotted as a function of Ag contents at these two temperatures (Fig. 6). It is observed that as the Ag contents increase TEC values decrease first and then when Ag content is 30 wt.% or more it starts to increase again at both temperatures. 3.3. Porosity As porosity is an important parameter of the cathode material because it facilitates the supply of oxygen to triple phase boundaries, the porosity and the effect of metallic contents of the cermets on the porosity has been studied. The theoretical density of LSGM was measured by helium pycnometry and it was found to be 6.56 g cm−3 . The value of the density calculated from the lattice parameters of LSGM (orthorhombic structure with a = 5.5392 Å, b = 5.4975 Å and c = 7.7484 Å, Z = 4) [24] is found to be 6.8 g cm−3 . However, as the calculated value is for pure single phase LSGM and as the possibility of other phases present in LSGM cannot be ruled out, the measured pycnometry density of LSGM is taken for further calculation. This is done for avoiding overestimation of porosity contents as density of the normally present secondary phases are quite less than LSGM (LaSrGaO4 = 6.39 g cm−3 , LaSrGa3 O7 = 5.56 g cm−3 [25]). The theoretical density of silver is taken as 10.5 g cm−3 [26]. The density of the mixtures was calculated using the rule of mixture assuming there is no interaction between the constituents of the mixtures. Fig. 7 shows the porosity of the cermets as a function of Ag contents for the samples, which were annealed at 600 ◦ C. At least 5 samples of each composition were measured and the average value is plotted. Error bar was not given, as there was a very little difference in the measured porosity of the samples. It is observed from the figure that the porosity increases up to 30 wt.% Ag and then decreases when the Ag content is further increased. This trend is the same irrespective of the annealing time. It is also noticed that there is no significant change of the porosity with the increase in annealing time though a slight decrease can be seen in Fig. 7(a–c). The effect of annealing temperature on porosity is shown in Fig. 8. The figure clearly indicates that as the annealing temperature increases the porosity slightly decreases.

Fig. 8. Variation of porosity of LSGM with silver after annealing at various temperatures for 100 h.

analysis. Core level spectra for the elements with highest photo ionization cross-section were only recorded. For example, Ag 3d has higher photo ionization cross-section value than Ag 4d. So, only the 3d peaks were recorded for extracting more reliable information. Ag 3d peaks of LSGM–30Ag sample which was annealed for 100 h at 600 ◦ C are given in Fig. 9. Two distinct curves are extracted by the non-linear curve fitting method.

3.4. X-ray photoelectron spectroscopy (XPS) The composition of the electrolyte as well as electrode surface and the chemical state of the elements were determined by XPS

Fig. 9. Core level XPS spectra of Ag 3d5/2 from LSGM–30Ag after 100 h annealing at 600 ◦ C.

360

P. Datta et al. / Materials Chemistry and Physics 114 (2009) 356–361

cal conductivity of other electrolytes [27]. However, the role of the additive on electrochemical property was beyond the scope of this work. 4. Discussion 4.1. Interaction between LSGM and Ag

Fig. 10. Arrhenius plot of LSGM–Ag cermets after annealing at 600 ◦ C.

3.5. Electrical conductivity The Arrhenius plot of some of the LSGM–Ag cermets is shown in Fig. 10. In the figure the behaviour of LSGM is also shown. It is noticed that as Ag content is increased up to 20 wt.% the conductivities of the cermets become lower than that of LSGM. But, when the added Ag content was 30 wt.% then the conductivity curve shows an abruptly higher value than parent LSGM as well as the cermets with lower Ag contents. 3.6. Polarization resistance A typical Arrhenius plot using the symmetrical cell arrangement with LSGM–Ag/LSGM/LSGM–Ag is shown in Fig. 11. Samples investigated in this study showed a single depressed arc. Given that the electrodes were thin, the amplitude of the ac applied voltage was very small (10 mV) and the measurement was taken in continuous flow of air, it is most likely that the contribution of the concentration polarization is negligibly small and the effective charge-transfer resistance dominates the polarization for the electrodes. Thus, polarization resistance is essentially equal to the charge-transfer resistance. The activation energy from the slope of the curves was determined which is around 110 kJ mol−1 . It is noted here that 5 wt.% Bi2 O3 was added as sintering additive. Previously it was reported that the addition of Bi2 O3 improved the electri-

As in cermets chemical compatibility between the metallic and the ceramic material is very important they were annealed over a wide range of temperature and time. It seems LSGM does not interact with Ag irrespective of annealing time and temperature. XRD patterns taken up to 900 ◦ C and after an annealing time of even 100 h do not show any other phase. Thus, it can be conclusively said Ag and LSGM phases are chemically compatible. This wide range of temperature of annealing was undertaken to check the chemical compatibility and the inter-diffusion of any of the constituent elements of LSGM with Ag, especially so, as it is reported that Ga interacts with other metals like Pt. However, the radius of Ag2+ is 0.94 Å [28] which is not favorable for forming solid solution with any of the constituent elements of LSGM as the radii are either too high (La3+ and Sr2+ ) or too low (Ga3+ and Mg2+ ), i.e. both differ by around 30–40%. Chemical reaction of Ag with La, Sr, Ga or Mg may not be possible because of the high cohesive energy of silver due to filled 4d electrons [29]. 4.2. Porosity When Ag is added porosity of the cermets decreases with the increase of Ag content irrespective of the annealing time or temperature (Figs. 7 and 8). This is due to an increased compressibility of the LSGM mixture during powder compaction with increasing amounts of ductile Ag. Similar observations were made with Ce0.9 Gd0.1 O1.95 –Ag cermets [17]. 4.3. TEC TEC of a composite material is given by the following equation: ˛c = ˛A VA + ˛B VB where Vi and ˛i are volume fractions and TEC of the respective phases. As the TEC of silver is 19.1 × 10−6 K−1 [26], it is obvious from the above equation that with the increase of Ag content the TEC of the LSGM–Ag cermets will increase as it is found with other electrolyte–metal system [30]. 4.4. XPS

Fig. 11. Polarization conductivity as a function of reciprocal temperature for different LSGM–Ag cermets measured in air after sintering at 900 ◦ C.

There is a negative shift of Ag 3d5/2 of about 1 eV for the silver state of LSGM–30Ag as well as one additional peak is found at 368.4 eV. The high-energy peak is due to metallic silver. The low energy peak is due to the inducement of ionized silver atoms by adsorbed oxygen as the chemical characteristics of the adsorbed oxygen is atomic in nature and ionic in character like that of Ag–O bonding [31]. Adsorbed oxygen state can be related with that of surface oxide like film as proposed by Li et al. [32]. When the electronic states of silver and oxygen overlap they will hybridize and form bonding and antibonding states. If only the bonding state becomes occupied, the hybridization energy will be attractive and will counteract the orthogonalization energy loss. On the contrary, if both bonding and antibonding state become occupied then there will be no gain of energy due to hybridization and the energy loss due to orthogonalization will prevail. Though earlier results indicate bulk

P. Datta et al. / Materials Chemistry and Physics 114 (2009) 356–361

361

nobleness of silver, but it seems the surface is not that noble. Surface nobleness is determined by two factors, i.e. the degree of filling of the antibonding states on adsorption and the degree of orbital overlap with the adsorbate. Though Ag has a filled antibonding state [33] but as the orthogonalization energy loss is not much the net result of these two does not make the silver surface very noble and this is probably the reason for adsorption of oxygen on the surface of silver. It can be concluded that silver irrespective of the amount of its addition to LSGM does not interact in any way with any constituent elements of LSGM other than the modification of the surface by adsorbing oxygen, which is beneficial from the point of view of catalytic activity silver [32].

5. Conclusions

4.5. dc conductivity

Financial support from the Deutsche Forschungsgemeinschaft (DFG) for carrying out this work is gratefully acknowledged.

LSGM–30 wt.% Ag has much higher conductivity than LSGM and the other two composites (Fig. 10). This is attributed to the electron conduction path percolation, which leads to higher conductivity. This reveals that Ag has formed a percolative network in the LSGM matrix. SEM picture also verifies this. LSGM–20 wt.% Ag and LSGM–10 wt.% Ag have less conductivity than LSGM. This gives a concrete proof that Ag contents of these compositions are not sufficient to form a percolating structure. A minimum of 30% Ag silver is needed to have a percolating microstructure. The purpose of dc 4 probe measurement was to find out the percolating limit of silver. It is clear from the figure that the conductivity of LSGM–30Ag is few orders of magnitude higher than LSGM–20Ag, which means the conductivity of silver, is dominant in that composition. As the percolating limit is ascertained, compositions with more than 30% silver were not measured, as that was redundant. This result is consistent with Chen et al. [12] who got similar results with YSZ–Pd composites. 4.6. Electrode polarization and kinetics In case of a mixed conducting electrode material, in addition to the triple phase boundaries at the electrode–electrolyte surface, oxygen exchange can also occur over the entire pore surface area of the electrode though the nature of this oxygen exchange is debatable. Whether this is either charge-transfer or a chemical exchange process is not clear [34]. Upon increasing the Ag content decreases the overall porosity of the cermets, which should have increased polarization resistance of the LSGM–Ag. Very little difference in the activation energy of the cermets containing 50 and 60 wt.% of silver clearly indicates that the same kinetics is operative for both the cases. This also proves that the rate-controlling step for the electrodes investigated in this study may not be gas pore diffusion. Only these two compositions were chosen for measurement of polarization resistance as they are well above the percolating limit of silver, which is ascertained by dc measurement given in the previous section. One to one comparison of LSGM–Ag with literature is not possible as till date there is no reported data on the performance of LSGM–Ag cathode with any electrolyte.

LSGM does not have any interaction with Ag irrespective of the amount of silver and annealing temperature. Long-term annealing of the cermets does not show any significant degradation of the properties even up to a temperature as high as 900 ◦ C. Polarization conductivity study shows that LSGM–Ag is a suitable cathode material for LSGM electrolytes. From conductivity study and due verification of microstructure it is found that a minimum of 30 wt.% Ag is needed for the formation of a percolating microstructure. Acknowledgement

References [1] T. Ishihara, H. Matsuda, Y. Takita, J. Am. Chem. Soc. 116 (1994) 3801–3803. [2] M. Feng, J.B. Goodenough, Eur. J. Solid State Inorg. Chem. 31 (1994) 663–672. [3] A. Endo, H. Fukunaga, C. Wen, K. Yamada, Solid State Ionics 135 (2000) 353– 358. [4] J.A.M. van Roosmalen, E.H.P. Cordfunke, Solid State Ionics 52 (1992) 303–312. [5] S. Carter, A. Selcuk, R.J. Chater, J. Kajda, J.A. Kilner, B.C.H. Steele, Solid State Ionics 53–56 (1992) 597–605. [6] V. Dusastre, J.A. Kilner, Solid State Ionics 126 (1999) 163–174. [7] M.J.L. Ostergard, C. Clausen, C. Bagger, M. Mogensen, Electrochim. Acta 40 (1995) 1971–1981. [8] M. Juhl, S. Primdahl, C. Manon, M. Mogensen, J. Power Sources 61 (1996) 173–181. [9] C.W. Tanner, K.Z. Fung, A.V. Virkar, J. Electrochem. Soc. 144 (1997) 21–30. [10] T. Kenjo, M. Nishiya, Solid State Ionics 57 (1992) 95–302. [11] C.S. Chen, H. Kruidhof, H.J.M. Bouwmeester, H. Verweij, A.J. Burggraff, Solid State Ionics 86–88 (1996) 569–572. [12] C.S. Chen, B.A. Boukamp, H.J.M. Bouwmeester, G.Z. Gao, H. Kruidhof, A.J.A. Winnbust, A.J. Burggraff, Solid State Ionics 76 (1995) 23–28. [13] K. Sasaki, K. Hosoda, T.N. Lan, K. Yasumoto, S. Wang, M. Dokiya, Solid State Ionics 174 (2004) 97–102. [14] H. Hu, M. Liu, J. Electrochem. Soc. 143 (1996) 859–864. [15] Z.L. Wu, M. Liu, J. Am. Ceram. Soc. 81 (1998) 1215–1220. [16] L.S. Wang, S.A. Barnett, Solid Sate Ionics 76 (1995) 103–113. [17] P. Datta, P. Majewski, F. Aldinger, Mater. Chem. Phys. 107 (2008) 370–376. [18] C. Xia, Y. Zhang, M. Liu, Appl. Phys. Lett. 82 (2003) 901–903. [19] C. Xia, M. Liu, Adv. Mater. 14 (2002) 521–523. [20] M. Camaratta, E. Wachsman, Solid State Ionics 178 (2007) 1411–1418. [21] A. Jaiswal, E.D. Wachsman, Solid State Ionics 177 (2006) 677–685. [22] P. Datta, P. Majewski, F. Aldinger, Mater. Chem. Phys. 102 (2007) 125–131. [23] P. Datta, P. Majewski, F. Aldinger, Mater. Res. Bull. 43 (2008) 1–8. [24] P. Datta, P. Majewski, F. Aldinger, J. Alloys Compd. 438 (2007) 232–237. [25] A.A. Khanlou, F. Tietz, D. Stöver, Solid State Ionics 135 (2000) 543–547. [26] C.J. Smithells, E.A. Brandes (Eds.), Metals Reference Book, 5th ed., Butterworths, London/Boston, 1976. [27] M. Hirano, T. Odaa, K. Ukai, M. Yasunobu, Solid State Ionics 158 (2003) 215– 223. [28] R.D. Shannon, Acta Crystallogr. A32 (1976) 751–767. [29] L.H. Tjeng, M.B.J. Meinders, J. Van Elp, Elp.J. Ghijsen, G.A. Sawatzky, Phys. Rev. B 41 (1990) 3190–3199. [30] P. Datta, P. Majewski, F. Aldinger, J. Alloys Compd. 455 (2008) 454–460. [31] A.I. Boronin, S.V. Koscheev, G.M. Zhidomirov, J. Electron. Spectrosc. Relat. Phenom. 96 (1998) 43–51. [32] W.X. Li, C. Stampfl, M. Scheffler, Phys. Rev. Lett. 90 (2003) 256102-1–256102-4. [33] B. Hammer, J.K. Norskov, Nature 376 (1995) 238–240. [34] M. Liu, Z. Wu, Solid State Ionics 107 (1998) 105–110.