- Email: [email protected]
Effect of ‘A’-site non stoichiometry in strontium doped lanthanum ferrite based solid oxide fuel cell cathodes
Materials Research Bulletin 72 (2015) 306–315
Contents lists available at ScienceDirect
Materials Research Bulletin journal homepage: www.elsevier.com/locate/matresbu
Effect of ‘A’-site non stoichiometry in strontium doped lanthanum ferrite based solid oxide fuel cell cathodes Koyel Banerjee, Jayanta Mukhopadhyay* , Madhurima Barman, Rajendra N. Basu* Fuel Cell and Battery Division, CSIR-Central Glass and Ceramic Research Institute, Kolkata 700032, India
A R T I C L E I N F O
A B S T R A C T
Article history: Received 13 January 2015 Received in revised form 9 July 2015 Accepted 2 August 2015 Available online 5 August 2015
Effect of A-site non-stoichiometry in strontium doped lanthanum cobalt ferrite (La1xSrxCoyFe1yO3d, x = 0.4; y = 0.2) is studied in a systematic manner with variation of ‘A’ site stoichiometry from 1 to 0.94. The perovskite based cathode compositions are synthesized by combustion synthesis. Powder characterizations reveal rhombohedral crystal structure with crystallite size ranging from 29 to 34 nm with minimum lattice spacing of 0.271 nm. Detailed sintering studies along with total DC electrical conductivities are evaluated in the bulk form with variation of sintering temperatures. The electrode polarizations are measured in the symmetric cell configuration by impedance spectroscopy which is found to be the lowest (0.02 V cm2 at 800 C) for cathode having highest degree of ‘A’-site deficiency. The same cathode composition exhibits a current density of 2.84 A cm2 (at 0.7 V, 800 C) in anode-supported single cell. An attempt has been made to correlate the trend of electrical behaviour with increasing ‘A’site deficiency for such cathode compositions. ã 2015 Elsevier Ltd. All rights reserved.
Keywords: A. Oxides C. Impedance spectroscopy D. Thermal expansion D. Electrical properties D. Electrochemical properties
1. Introduction To increase the efficiency of the intermediate temperature solid oxide fuel cell (SOFC), the mixed ionic and electronic conductor (MIEC) materials, such as, Ba1xSrxCo1yFeO3d (BSCF), Nd-doped BSCF (BSNCF) and La1xSrxCoyFe1yO3d ((LSCF) have proven promising and demonstrated faster oxygen diffusion and improved surface exchange kinetics [1–9]. Among the MIEC materials as mentioned above, LSCF based cathode materials have good electrical conductivity as well as high oxygen surface exchange coefficient and also have a good oxygen self-diffusion coefficient between the temperature range of 600 C and 800 C. The oxygen self-diffusion coefficient of LSCF is 2.6 109 cm2 s1 at 500 C, which is far better in performance to that of LSM having oxygen self-diffusion coefficient value of 1012 cm2 s1 at 1000 C [10]. There are many factors concerning the duration of stability for LSCF based cathodes. One of the probable mechanisms for voltage deterioration is explained to be slow decomposition of the LSCF perovskite due to partial de-mixing of strontium [11–13]. Mai et al. [14] reported that a small A-site deficiency and high ‘Sr’ content in LSCF cathodes increase the cell performance using gadolinium doped ceria (CGO) interlayer [14]. The long term stability of the cell
* Corresponding authors. Fax: +91 33 24730957. E-mail addresses: [email protected], [email protected] (J. Mukhopadhyay), [email protected], [email protected] (R.N. Basu). http://dx.doi.org/10.1016/j.materresbull.2015.08.002 0025-5408/ ã 2015 Elsevier Ltd. All rights reserved.
using LSCF cathode materials with off-stoichiometry in the ‘A’-site have also been tested [15]. The off-stoichiometry in the ‘A’ site lattice of LSCF-based cathode materials leads to the formation of the stable aliovalent states of the metal ions in the ‘B’ site of the perovskite and hence results in higher conduction. Moreover, change of the oxygen off-stoichiometry by inducing such deficiency in the ‘A’ site also enhances such conduction [10]. It has also been found that ‘A’-site vacancy plays an important role for the performance of the LSCF cathode but long-term cell tests lasting up to 3000 h revealed a degradation of the cell. The sintering of the cathode having off-stoichiometry in the ‘A’ site is also found to be prevalent through the formation of a congruent eutectic melt which results in easy process of grain to grain interlocking at low temperature [16] and hence higher grain conductivity. It has been observed that in iron-based cathodes, reactivity with YSZ electrolyte is significantly reduced. In addition, thermal expansion coefficient of the ferrite based–perovskite is relatively close to that of the electrolyte YSZ. On the other hand, LSCF-type perovskites are generally incompatible with YSZ electrolytes due to undesirable interface reactions. Therefore, a CGO diffusion barrier layer is used to prevent the formation of low conductive compounds without negatively affecting the electrochemical performance [17,18]. Besides, the composition and microstructure of cathode materials has a significant impact on the performance of SOFCs. Hence, rational design of materials composition through controlled oxygen non-stoichiometry and consequent increase in defect aspects can enhance the ionic and
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
electronic conductivities as well as the catalytic properties for oxygen reduction in the cathode material. In the present study, systematic variation of lanthanum concentration in ‘A’-site oxygen non-stoichiometry in ferrite based LSCF materials (LSCF; La1xSrxCoyFe1yO3d, x = 0.4; y = 0.2) has been carried out in the range of 0.54–0.6 while the concentration of dopant ‘Sr’ is kept constant at 0.4. Ferrite based LSCF materials have been chosen as cathode material for the application in intermediate temperature solid oxide fuel cell for the thermal compatibility with the electrolyte YSZ. Powder and bulk characterizations including the DC electrical conductivities of the cathode materials have been carried out in details and clinically correlated with the microstructures. Impedance spectroscopy is employed to investigate the electrode polarizations of the cathodes as thick film in symmetric cell configuration. Electrochemical performances of all cathode compositions are evaluated in conjunction with ceriabased interlayer in the form of Ni-YSZ anode-supported single cells and the performances are correlated with the off-stoichiometry of the ‘A’ site of the synthesized perovskites. 2. Experimental The compositions of the cathodes as investigated under this study are given in Table 1. All the four compositions (LS-1, LS-2, LS3 and LS-4) were prepared by the combustion synthesis technique as reported earlier using starting materials lanthanum (III) nitrate hexahydrate (99%, Sisco Res. Lab. Ltd., India), strontium (II) nitrate (99%, s.d.fine, India), iron (III) nitrate nonahydrate (98%, Merck) and cobalt (II) nitrate hexahydrate (99.5%, E. Merck, Germany) and Lalanine as a fuel (99%, SRL, India) [19]. In all the four combustion synthesis, the ratio of metal nitrate and fuel were maintained 1:1 to prevent the undesired hydrolysis of the respective precursor salts. The as-synthesized powders were calcined at 800 C for 4 h in air. The Co-CGO was prepared using combustion synthesized technique and already reported by our group [20]. For the evaluation of thermal analysis of the cathode materials the viscous precursor gels of all the samples were collected. The differential thermal analysis (DTA) and thermogravimetric analysis (TGA) were carried out from room temperature to 800 C in air at a heating rate of 5 C/min using Simultaneous Thermal Analyzer (STA 449C, Netzsch, Germany). Phase purity of the samples was checked by recording X-ray data in the 2u range 10 –80 using X’Pert Pro software (PANalytical, Philips, Holland). For structural characterization, Rietveld refinement of the powder diffraction profiles and quantitative phase analysis of LS samples calcined at 800 C were carried out using X’Celerator operating at 40 kV and 30 mA using CuK a radiation with step size 0.05 ð2uÞ and step time 75 sec from 15 to 90 for all the samples. The BET surface areas of the powder calcined at three different temperatures were carried out using Quantachrome Instruments (version 9.0). The morphology of both the powder was characterized using transmission electron microscopy (TEM, Tecnai G2 30ST) with evaluation of lattice spacing by high resolution TEM (HRTEM). For the densification studies, powders calcined at 800 C were pressed uniaxially and
307
then sintered in the temperature range of 1050 C–1150 C in air in an interval of 50 C. Field emission scanning electron microscopy (FESEM, Gemini Supra 35 Zeiss) was used to examine the microstructure of the fractured surface of the sintered samples. The thermal expansion coefficients (CTE) of all the sintered samples of different compositions under the present investigation were measured by a dilatometer (NETZSCH, DIL402) up to 900 C with a constant heating rate of 10 C/min. The electrical conductivity of the sintered samples (1050 C–1150 C) was evaluated in the temperature range 500 C–800 C in air by 4probe technique using a Power Source (Agilent E3631A) and a Multimeter (Keithley 2002). Electrochemical impedance of symmetric cells having screen printed LSCF cathode on both the sides Co0.01Ce0.79Gd0.20O2d (Co-CGO) electrolyte in the configuration, LSCF / Co-CGO /LSCF, were carried out in the form of circular disc of diameter of 10 mm with the full range swipe of the AC frequency of 101 Hz f 106 Hz. While the sintering temperature of bulk Co-CGO was kept at 1100 C, the thick films of the composite cathode were fired at 900 C [21]. For electrochemical performance evaluation of the cathodes of varied composition, thick film paste prepared using such powders were screen printed onto anodesupported planar half-cell of configuration, NiO-YSZ/YSZ. The halfcell was fabricated by tape casting method which is also reported our group previously [22]. Single cells in the form of coupon cells (16 mm diameter, 1.5 mm thick with active cathode area of 0.3 cm2) of configuration Ni-YSZ/YSZ/CoCGO/LSCF were fabricated after heat treatment of screen printed LSCF-based cathodes at 950 C. Co-CGO was used as an interlayer in between the electrolyte and cathode to reduce the unwanted reaction and thermal compatibility. The electrochemical measurements of such single cells were carried out in the temperature range 700 C– 800 C using an in-house electrochemical measurement setup. Moist H2 (3% H2O) was used as a fuel on the anode side and oxygen was fed on the cathode side. During the measurement, the flow rates for both the fuel gas and O2 were maintained at 100 SCCM. The best performing cathode was also tested with air for comparison with the results of oxygen and the flow rates for the same is maintained at 500 SCCM. 3. Results and discussion The ‘A’-site deficient LSCF-based cathode plays a major role in improving the performance of SOFC by the increase of the electronic conductivity and catalytic behavior. However, the deficiency in ‘A’ site has a limitation with respect to the substitution of Sr2+ in the trivalent site (La3+) of perovskite having formula La1xSrxCo1yFeyO3d. It has been reported that at x > 0.4 makes the perovskite structure metastable and the ‘A’ site deficiencies greater than 5% results a negative effect on measured performance [23]. Keeping the view of the above, the ‘Sr’ content in the LSCF compositions has been kept constant at 0.4 whereas, the ‘A’ site deficit has been made upto 6% with respect to ‘La’ site as shown in Table 1.
Table 1 Compositions of the LSCF powders and relative densification and CTE values of the sintered bulk samples. Sample ID
Composition
LS-1 LS-2 LS-3 LS-4
La0.54Sr0.40Co0.20Fe0.80O3d La0.56Sr0.40Co0.20Fe0.80O3d La0.58Sr0.40Co0.20Fe0.80O3d La0.60Sr0.40Co0.20Fe0.80O3d
Relative densification (%) at the following sintering temperatures
1050 ( C)
1100 ( C)
1150 ( C)
80.1 80.2 79.7 76.2
90.6 87.2 86.4 84.8
94.5 94.1 89.5 88.3
Coefficient of thermal expansion (CTE) of sintered (1150 C) bulk cathodes in the temperature range 30–900 C 106 (K1) 16.88 15.46 15.72 15.76
308
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
3.1. Thermal analysis The thermal analysis of the precursor gels collected during the synthesis for all the four LSCF compositions have been shown in Fig. 1. Mass loss in the TG curve occurs at 158 C may be attributed due to the initial evaporation which is followed by an initiation of thermal decomposition of metal nitrate-alanine complex in the temperature range of 180 C–190 C as also reflected by the presence of the sharp exothermic peak at 182 C in DTA. The decomposition is found to be progressive enough in the temperature range of 248 C–255 C which is also observed by the presence of a shallow exothermic hump in DTA. The final decomposition of the metal-alanine complex takes place in the broad temperature range of 300 C–480 C with a specific exothermic peak appearing in the temperature range of 410 C–416 C as also observed in DTA. 3.2. Powder characterizations The X-ray diffractograms of the all four compositions of LSCF powders calcined at 800 C are shown in Fig. 2. Reitveld analyses for all the powder samples calcined at 800 C are carried out and all the peaks have been indexed accordingly. The calculated lattice parameters as obtained from Rietveld analyses of the X-ray diffractograms are found to be a = b = 5.49 Å and c = 13.40 Å for three cathode compositions viz. LS-1 (La0.54Sr0.4Co0.2Fe0.8O3d), LS-2 (La0.56Sr0.4Co0.2Fe0.8O3d) and LS-3 (La0.58Sr0.4Co0.2Fe0.8O3d). The same is found to be a = b = 5.41 Å and c = 12.96 Å for LS-4 (La0.6Sr0.4Co0.2Fe0.8O3d). The average crystallite sizes of these are within the range between 29.6 nm and 37.9 nm. Since the ionic radius of Co4+ (0.067 nm) is smaller than that of Fe4+ (0.0725 nm), formation of higher concentration of Co4+ as a result of higher degree of substitution in ‘A’ site lattice of the LSCF-based perovskite results in lattice shrinkage. The theoretical densities of the crystallites are found to be 6.10, 6.17, 6.26 and 6.38 g cc1 for cathode composition LS-1, LS-2, LS-3 and LS-4 respectively. ‘A’-site deficiency of the LSCF-perovskite makes a pronounced effect on the surface area of the calcined materials. The powder samples calcined at 800 C was used to surface area measurements which are evaluated using gas adsorption BET isotherm. LS-1 with largest ‘A’-site deficiency in ‘La’ is found to have lowest surface area (6.45 m2/gm), which might be attributed to the agglomeration owing to ease of materials diffusion across the particle boundary because of higher degree of defects in the perovskite. While surface area values for LS-3 and LS-4 are found to be 10.68 m2 gm1 and 10.35 m2 gm1 respectively, it is 7.78 m2 gm1 for the composition LS-2. The powder morphology of the compositions is also
Fig. 2. XRD patterns of the powder samples (a) LS-1, (b) LS-2, (c) LS-3 and (d) LS4 calcined at 800 C.
evaluated using TEM (Fig. 3). It is found that the average particle size ranges in between 50 nm and 100 nm. The particle locking across the boundary is observed to be more prominent in LS1 having highest ‘A’ site deficiency which is also reflected in the value of the lowest surface area for the same composition. The interplanar distances have been calculated from HRTEM images which show the orientation of crystal planes accordingly (Inset of Fig. 3). Selected area electron diffraction patterns shown in insets of Fig. 3 exhibit rhombohedral crystal symmetry. The interplanar spacing is reduced from 0.381 nm to 0.275 nm as the ‘A’-site deficiency is increased from the composition LS-4 (La0.60Sr0.40Co0.20Fe0.80O3d) to LS-1 (La0.54Sr0.40Co0.20Fe0.80O3d). This trend of interplanar spacing is found to have corroboration with the lattice shrinkage as was observed in the Reitveld analyses. 3.3. Densification and thermal expansion studies
Fig. 1. TG and DTA analysis of various LSCF gels in the temperature range of 30– 800 C.
To find out the effect of the off-stoichiometry in the ‘A’-site lattice particularly in the place of ‘La’, the densification study of the samples sintered in the temperature range of 1050 C–1150 C is shown in Table 1. As expected the density of all the samples (LS-1– LS-4) increases with increase in sintering temperature. While, LS-1 and LS-2 reaches near 95% of the theoretical density at 1150 C, the percentage densification remains below 90% for LS-3 and LS-4. The densification data also supports the microstructures of the fractured surfaces of the sintered samples obtained using FESEM (Fig. 4). As reported in the literature the LSCF-based materials tend to get sintered by the presence of liquid phase [16,24]. The presence of higher degree of deficiency in ‘La’ site of LSCF-based perovskite is likely to facilitate the formation of liquid phase during the sintering. It is observed from Fig. 4 that development of well crystalline grains are prevalent in case of LS-1 having highest deficit in ‘A’-site of the perovskite which might be because of the result of recrystallization of the grains from the melt phase
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
Fig. 3. TEM image of the particles (a) LS-1, (c) LS-2, (e) LS-3, (g) LS-4 with HRTEM and Selected area electron diffraction pattern (SAED) image (insets).
Fig. 4. FESEM image of the fractured surface of the bulk samples (a) LS-1, (b) LS-2, (c) LS-3, (d) LS-4 sintered at 1050 C.
309
310
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
(Fig. 4a). As the recrystallization passes through a process of recovery, grain sizes are found to be smaller i.e. in the range of 0.2– 0.5 mm [25]. In contrast, LS-4 having lowest level of ‘A’-site deficit exhibits higher percentage of melt phase in the microstructure and hence well-developed grains are not much prominent in such case evident from Fig. 4d. LS-3 and LS-2 falls in between the extremes and hence the sintered microstructures are also found with the grains having a balance between the presence of melt phase and recrystallized grains. However, the microstructure of LS-3 having higher degree of ‘A’ site deficit is associated with higher melt phase compared to that of LS-2. The thermal expansion co-efficient of the bulk samples sintered at different temperatures have been evaluated which are also presented in Table 1. It is found that irrespective of composition of the cathodes under present investigation, the coefficient of thermal expansion increases with increase in sintering temperature. The highest lattice shrinkages are found with LS-1 having highest degree of ‘A’ site deficit. This attributes to the higher coefficient of thermal expansion (CTE) of the perovskite and is found to be 16.88 106 (K1) in the temperature range 30 C–900 C. The thermal expansion coefficient of the LSCF-based is found least [15.46 106 (K1)] with LS2 sintered at 1150 C. However, CTE is found to be in the range of 15.46–15.76 106 (K1) for the perovskites having lower in ‘A’site deficiency e.g., in LS-2 to LS-4. Such a higher level of thermal mismatch for LSCF-based cathodes with that of the YSZ electrolyte may, however, be minimized by laying down a thin and dense particulate coating of doped ceria based interlayer [26,27].
3.4. Electrical conductivity The DC electrical conductivity of the sintered samples for all the four compositions are examined in the temperature range 500 C– 800 C with an interval of 25 C. The temperature dependent Arrhenius plots for DC electrical conductivity for cathode compositions are shown in Fig. 5. It is found that irrespective of all the compositions, the total electrical conductivity increases with increase in sintering temperature which attributes to the higher degree of sinterability of the samples at high temperature. However, all the compositions exhibit a conductivity transformation from semiconductor to metallic type upon increase of measurement temperature from 500 C–800 C. It is reported in our earlier communications that such shift of electrical conductivity is natural in case of LSCF-based materials [27]. It has also been reported that the charge disproportionation of Co3+ occurs in La1xSrxCo1yFeyO3 based compositions increases the concentration of electronic carriers [28] which is presented in Eq. (1). 2Co3+ ! Co4+ + Co2+
(1)
Defect model associated with the defect structure of the LSCF based materials proposed by different groups has been given by the following Kröger–Vink notations [10,29,30] SrO þ Fe Fe
LaFeO3 Sr0La þFe Fe þO o
!
Fig. 5. Arrhenius plots of the electrical conductivity of the bulk as a function of temperature.
ð2Þ
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
1 2Co Co þ O o ! 2COCo þ V o þ O2 ðgÞ 2
ð3Þ
1 2Fe Fe þ O o ! 2FeFe þ V o þ O2 ðgÞ 2
ð4Þ
311
highest conductivity exhibits the activation energy of 0.04 eV in the semiconducting region whereas, the same is found to be 0.17 eV in the range of metallic type conductivity. 3.5. Electrochemical studies
In Eq. (2) it is observed that when Sr2+ is doped as SrO in the lattice site of La3+, decrease of positive charge gives rise to a negative 0 lattice viz. Sr La . The deficiency in the positive charge thus generated is compensated by the conversion of Co3+ to Co4+ and Fe3+ to Fe4+ which is responsible for the electronic conduction. The charge disproportionation reaction as shown in Eq. (1) is primarily responsible for generating Co Co site or Fe Fe. Eqs. (3) and (4) basically show how at elevated temperature oxygen vacancy is created in the LSCF lattice due to conversion of B-site transition metal elements Co and Fe from their higher oxidation state (+4) to stable lower oxidation state (+3). Presence of higher degree of ‘A’site deficiency leads to formation of higher aliovalent cation of ‘Co’ and hence enhances the chances of disproportionation reaction as mentioned above. LS-1 (La0.54Sr0.40Co0.20Fe0.80O3d) sintered at a temperature of 1150 C shows a maximum electrical conductivity of 320 S cm1 measured at 800 C whereas, LS-4 (La0.60Sr0.40Co0.20Fe0.80O3d) with stoichiometric composition at ‘A’ site and prepared under similar condition exhibits the lowest electrical conductivity of 235 S cm1 measured at 800 C. It is observed that all the compositions of the cathodes exhibits lower activation energy in the regime of semiconducting electrical conduction (lower temperature range of 500 C–650 C) compared to that of the metallic type of conductivity (higher temperature range of 675 C–800 C). LS-1 sintered at 1150 C and having
The electrode polarization of the cathode compositions has been examined using impedance spectroscopy in the form of screen printed thick film applied onto the both sides of Co-CGO disks of thickness 1 mm. It is found that the polarization values decrease as the ‘A’ -site deficiency increases due to higher rate of surface exchange kinetics [31–33]. The characteristics impedance spectra for all the four compositions have been given in Fig. 6. Impedance decreases with increase in the ‘A’-site deficiency which is clear from the impedance spectra obtained from of SYM-LS-4 to SYM-LS-1. It is observed from the figure that impedance shifts towards the higher frequency side much faster for SYM-LS-1–SYMLS-3 compared to that of the stoichiometric composition of SYMLS-4. In the low temperature regime, the impedance spectra though exhibits two incomplete semicircles (particularly observed for impedance spectra at 600 and 700 C for SYM-LS-4), the same is found to be transformed to single semicircle upon increase in the measurement temperature. The inset of Fig. 6 shows the spectra for all the compositions but with a much higher resolution of the scale values. From Fig. 7, it can be observed that at lower temperature regime, the different processes viz. bulk impedance happening at the higher frequency region and electrode phenomena happening in the middle and low frequency field can be differentiated by the presence of two different values of |Z|. The Z1 signifies primarily the charge transfer corresponding to the transfer of active oxygen species involving vacancies/ions or adsorbed atom from the
Fig. 6. Nyquist plots measured at different temperatures for the four cathode compositions.
312
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
Fig. 7. Bode plots measured at different temperatures for the four compositions.
electrode to the electrolyte surface and the next regime wherein combination of Z1 + Z2 attributes to the total contributions involving the solid state diffusion coupled with chemical reaction occurring on the electrode surface [34,35]. LSCF being the mixed ionic and electronic conducting electrode, the electronic contribution primarily comes out of the grain of the LSCF electrode and ionic conduction arises out of the defect conduction and hence happens preferably at the grain boundary junctions. It is also observed from the figure that with increase in measurement temperature 700 C, distinction of these two regions are demolished and a single step ǀZǀ is found to prevail throughout the frequency region. This is also observed from the Nyquist plot (Fig. 6) that the high frequency region of the semicircle is severely depressed upon increasing the temperature and thereby the electrode response takes over the bulk impedance. In case of LS1 having composition La0.54Sr0.40Co0.20Fe0.80O3-d, it is found that the shift of ǀZǀ from Z1 to Z1 + Z2 is found to be in the frequency field of 1.5–1.4 105 Hz compared to 3.6 - 2.8 105 Hz as observed for LS4, the stoichiometric composition (La0.60Sr0.40Co0.20Fe0.80O3d). This indicates that the cathode response involving oxygen diffusion, surface adsorption of molecular oxygen and subsequent reduction to oxygen ion using electron transfer process,
incorporation and transport of oxygen ion from the cathode surface to the electrolyte interface [34] is more effective for the composition having the highest degree of ‘A’ site deficiency. While the high frequency intercepts represents the ohmic polarization (Ro) of the symmetric cell which consists of the ohmic polarization of CoCGO electrolyte and that of the electrode polarization including the interfaces, the low frequency intercept depicts the total resistance (Rt). Cathode polarization (Rp) is calculated from the difference of the high frequency and low frequency intercept. The calculated Rp for all the cathodes are presented in the Table 2. Cathode polarizations are found to decrease with increase in the ‘A’-site deficiency and found to be the lowest (0.02 V cm2) for SYMLS-1 and the highest value for the same is noticed with SYM-LS-4 (0.12 V cm2). SYM-LS-2 and SYM-LS-3 are found with intermediate values of the Rp. The ohmic polarization (Ro) is also found minimum with SYM-LS-1 (2.49 V cm2). Fig. 8 represents the Arrhenius plot for the ohmic and cathodic polarizations for the cathode compositions. SYM-LS-1 and SYM-LS-2 exhibit lower activation energy of ohmic polarizations (0.21 and 0.17 eV for LS-1 and LS-2 respectively) compared to SYM-LS-3 and SYM-LS-4 (0.47 and 0.44 eV for SYM-LS-3 and SYM-LS-4 respectively). The same trend is also followed for the activation energies of cathode
Table 2 Impedance data and electrochemical performance of single cells with different cathode compositions. Sample
LS-1 LS-2 LS-3 LS-4
Impedance data
Electrochemical performance
Ohmic polarization (V cm2)
Electrode polarization (V cm2)
Current density at 0.7 V at 800 C (A cm2) Oxygen
Air
2.49 3.58 3.24 3.12
0.02 0.03 0.04 0.12
2.84 2.49 2.19 2.10
2.30 2.04 1.78 1.70
Cell ASR values in oxygen at 800 C (V cm2)
0.13 0.14 0.17 0.15
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
313
Fig. 8. Arrhenius plots of (a) Electrode polarization (Rp) and (b) Ohmic polarization (Ro) for the various cathode compositions.
polarization viz. SYM-LS-1 and SYM-LS-2 are found with activations energies of 0.63 and 0.56 eV for Rp and the same is 1.42 eV and 1.54 eV for SYM-LS-3 and SYM-LS-4 respectively. It is found from the impedance spectroscopy that a low frequency arc appears in the lower temperature especially in the range 650 C which is primarily associated with the diffusion and charge transfer reaction for the oxygen reduction reaction. In order to investigate the effect of charge transfer and diffusion polarization on the cathode performance having the highest ‘A’ site deficiency, impedance study has been carried out for the Sym-LS-1 in different environment using different oxygen partial pressure at 107 (mixture of Ar and air), 0.21 (air) and 1 atm (in oxygen). It has been observed that at lower temperature range and in lower oxygen partial pressure, the cathode polarization is initially governed by diffusion process followed by the charge transfer being captured at the lower frequency of the impedance spectra. Fig. 9 shows the Nyquist plots at an intermediate temperature of 650 C at different oxygen partial pressure. At a specified temperature of 650 C, with decreasing partial pressure of oxygen for the LSCF-based cathode materials, the sizes of low-frequency arc is found to be enlarged gradually. Simultaneously, it is also observed that the arc is having lower values than that of the highfrequency arc. While the low frequency arc for Nyquist plot is found with total impedance (Rt) of 1561 V cm2 at 5.012 Hz in Ar + air atmosphere, the same is observed to be only 5.19 V cm2 at 1.259 Hz in oxygen rich environment. Rt is found to be intermediate in air atmosphere viz. 5.29 V cm2 at 1.585 Hz (insets of Fig. 9). This is a typical indication that the primary rate-limiting mechanism being related with the charge-transfer processes for oxygen at the air electrode [36]. As the temperature increases (>700 C), the diffusion being faster through the porous electrodes, the lower frequency arc disappears emphasizing the fact that under cathodic polarization the oxygen reduction kinetic process is not only governed by charge-transfer mechanism but also strongly by adsorption–desorption and diffusion of oxygen. At high temperature though the concentration polarization is decreased in oxygen
rich environment and charge transfer polarization is also found to decrease because the activation of the cathode is found with faster kinetics at higher temperature. The exchange current density (io) for SYM-LS-1 is calculated at 800 C from the impedance spectroscopy using the low field approximation technique [21] wherein the rate limiting steps for oxygen reduction is considered as single and the same is found to be 1.16 A cm1, which is significantly large for such MIEC conducting cathodes. Single cell electrochemical performances using each of the cathode compositions have been evaluated using hydrogen as the fuel and oxygen as the oxidant and shown in Fig. 10. A systematic increase in the current density is observed for the compositions with increase in the ‘A’–site deficiency. While LS-1 exhibits a current density as high as 2.84 A cm2 and power density 1.99 W cm2, LS-4 shows the cell performance as 2.10 A cm2 with power density 1.47 W cm2 when measured at 800 C corresponding to cell voltage of 0.7 V. However, the peak current density as well as power density for LS-1 is found to be still higher and exhibits 3.78 A cm2 and 2.27 W cm2 when taken at individual cell voltage of 0.6 V. Cell area specific resistances (ASR) are calculated considering the slope of the linear portion of the I–V curve and is found to be minimum (0.13 V cm2) with cell having LS-1 as the cathode. The highest performing cathode is also tested with the air and a drop of 19% is observed in the peak performance. The peak current density for LS1 with air is found to be 3.06 A cm2 when measured at 800 C with the individual cell voltage of 0.6 V. The same is found to be 2.80 A cm2 using air as the oxidant at cell operating voltage of 0.7 V. All the cells have been tested for 100 h. The cell performances for all the compositions along with cell ASR are given in Table 2. These cell performances overshadow the performance of our earlier reported LSCF-based cathodes prepared by combustion technique [15] and hence establish the effect of the ‘A’-site deficiency in the LSCF-based perovskite lattice. The FESEM images of the cross section of the single cells of all cathode compositions are shown in Fig. 11. From the cross sectional micrographs, well defined porous Ni-YSZ based anode, dense YSZ based electrolyte
Fig. 9. Nyquist plots of LS-1 at different oxygen partial pressure at 650 C.
314
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
Fig. 10. Electrochemical testing of SOFC single cell (coupon cell) at different temperatures with cathode compositions LS-1, LS-2, LS-3 and LS-4.
Fig. 11. FESEM images of the single cells with LS-1, LS-2, LS-3 and LS-4 cathode compositions.
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
and porous cathodes of LSCF compositions in conjunction with CoCGO based interlayer are found vivid. However, cell microstructures with LS-1 and LS-2 having higher degree of ‘A’ site off stoichiometry is found with higher densification for screen printed cathode layer compared to their stoichiometric compositions because of ease of sintering for ‘A’ site deficit cathodes as discussed in the earlier section. 4. Conclusions Nanocrystalline LSCF-based cathode compositions (La1xSrxCoyFe1yO3d, x = 0.4; y = 0.2) varying the A-site stoichiometry from 1 to 0.94 (with a decrement of 0.02) in the ‘La’ site of the perovskite have been synthesized using combustion synthesis technique. Thermal analyses revealed a multistage decomposition. XRD analysis of all the four compositions shows rhombohedral crystal structure upon calcination at 800 C. Surface area evaluation of the calcined powders revealed that with increase in the ‘A’– site deficiency increases the materials diffusion across the particulate boundary and hence increase the particle sizes by forming agglomerates. CTE of the LSCF-based perovskite decreases with decrease in ‘A’ site deficiency and found highest in LS-1 (16.88 106). The CTE is found to be in the range of 15.46– 15.76 106 (K1) for the other perovskites of LS-2–LS-4. However, such mismatch in CTE be minimized when applied with a doped ceria based interlayer. The DC electrical conductivity also follows the same trend and is found to be highest (320 S cm1) with LS1 sintered at a temperature of 1150 C. Though at low temperature and low partial pressure of oxygen, the diffusion polarization is found with low frequency arc, the same is found negligible at high temperature and oxygen rich environment. Dependence of charge transfer reaction of the oxygen reduction is also found by the presence low frequency arc at moderate temperature of 650 C and even observed at oxygen rich environment. The impedance spectra of the screen printed cathode compositions revealed a single step deconvolution when measured 700 C and the calculated exchange current density for the highest ‘A’ site deficit cathode is found to be as high as 1.16 A cm2 at 800 C. The lowest cathode polarization is obtained for LS-1 which is found to be 0.02 V cm2 at the same measurement temperature. The current density for single cell prepared with LS-1 cathode material is observed to be the highest i.e. 2.84 A cm2 when operated with cell voltage 0.7 V at 800 C. Thus, LSCF-based cathode compositions having synthesized with maximum 6% of deficit at ‘La’ site is found to have good structural integrity with effective process of sintering in the bulk conditions, good DC electrical conductivity and the least cathode polarizations resulting in the highest electrochemical performance in the single cell configuration. Acknowledgements One of the authors (KB) is thankful to CSIR for providing Senior Research Fellowship (SRF). The authors are thankful to the Director,
315
CSIR-CGCRI for his kind permission to publish the work. The authors also acknowledge the financial support of Department of Science and Technology. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]
[12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36]
W. Zhou, R. Ran, Z. Shao, J. Power Sources 192 (2009) 231–246. S. Li, Z. Lu, X. Huang, W. Su, Solid State Ionics 178 (2008) 1853–1858. Y. Zeng, Y.S. Lin, S.L. Swartz, J. Membr. Sci. 150 (1998) 87–98. A. Raj, A.S. Nesaraj, M. Kumar, F. Tietz, H.P. Buchkremer, D. Stoever, J. New Mater. Electrochem. Syst. 7 (2004) 145–151. N.J. Simrick, A. Bieberle-Hütter, T.M. Ryll, J.A. Kilner, A. Atkinson, J.L.M. Rupp, Solid State Ionics 206 (2012) 7–16. X. Ding, X. Kong, J. Jiang, C. Cui, X. Guo, Mater. Res. Bull. 45 (2010) 1271–1277. L. Conceicao, A.M. Silva, N.F.P. Ribeiro, M.M.V.M. Souza, Mater. Res. Bull. 46 (2011) 308–314. L. Lu, Q. Shi, Y. Yang, H. Zhang, Mater. Res. Bull. 47 (2012) 1016–1020. L. Lu, Y. Guo, H. Zhang, J. Jin, Mater. Res. Bull. 45 (2010) 1135–1140. C. Sun, R. Hui, J. Roller, J. Solid State Electrochem. 14 (2010) 1125–1144. Z. Feng, Y. Yacoby, M.J. Gadre, Y.L. Lee, W.T. Hong, H. Zhou, M.D. Biegalski, H.M. Christen, S.B. Adler, D. Morgan, Y.S. Horn, J. Phys. Chem. Lett. 5 (2014) 1027– 1034. L. Adijanto, R. Küngas, F. Bidrawn, R.J. Gorte, J.M. Vohs, J. Power Sources 196 (2011) 5797–5802. H. Kishimoto, N. Sakai, T. Horita, K. Yamaji, M.E. Brito, H. Yokokawa, Solid State Ionics 179 (2008) 1338–1342. A. Mai, V.A.C. Haanappel, S. Uhlenbruck, F. Tietz, D. Stover, Solid State Ionics 176 (2005) 1341–1350. F. Tietz, A. Mai, D. Stover, Solid State Ionics 179 (2008) 1509–1515. X. Tan, Z. Wang, K. Li, Ind. Eng. Chem. Res 49 (6) (2010) 2895–2901. A. Mai, V.A.C. Haanappel, F. Tietz, D. Stover, Solid State Ionics 177 (2006) 2103– 2107. D. Waller, J.A. Lane, J.A. Kilner, B.C.H. Steele, Solid State Ionics 86–88 (1996) 767–772. A. Dutta, J. Mukhopadhyay, R.N. Basu, J. Eur. Ceram. Soc. 29 (2009) 2003–2011. A. Dutta, A. Kumar, R.N. Basu, Electrochem. Commun. 11 (2009) 699–701. K. Banerjee, J. Mukhopadhyay, R.N. Basu, Int. J. Hydrogen Chem. 39 (2014) 15754–15759. R.N. Basu, A. Das Sharma, A. Dutta, J. Mukhopadhyay, Int. J. Hydrogen Energy 33 (2008) 5748–5754. T. Striker, J.A. Ruud, Y. Gao, W.J. Heward, C. Steinbruchel, Solid State Ionics 178 (2007) 1326–1336. F. Schulze-Küppers, S. Baumann, F. Tietz, H.J.M. Bouwmeester, W.A. Meulenberg, J. Eur. Ceram. Soc. 34 (2014) 3741–3748. R. Zuo, S. Su, Y. Wu, J. Fu, M. Wang, L. Li, Mat. Chem. Phys. 110 (2008) 311–315. J.P. Martinez, D.M. Lopez, C.S. Baustista, A.J.D.S. Garcia, J.C.R. Morales, J.C. Vazque, P. Nunez, Bol. Soc. Esp. Ceram. 49 (2010) 15–22. J. Mukopadhyay, Rajendra N. Basu, J. Power Sources 252 (2014) 252–263. H. Duanping, X.U. Qin, Z. Feng, C. Wen, L. Hanxing, Z. Jian, J. Wuhan Univ. Technol. Mater. Sci. Ed. 23 (2008) 80–84. E.V. Bongio, H. Black, F.C. Raszewski, D. Edwards, C.J. McConville, V.R.W. Amarakoon, J. Electroceram. 14 (2005) 193–198. A.N. Petrov, O.F. Kononchuk, A.N. Petrov, O.F. Kononchuk, A.V. Andreev, Solid State Ionics 80 (1995) 189–199. E.N. Armstrong, K.L. Duncana, E.D. Wachsman, Phys.Chem. Chem. Phys. 15 (2013) 2298–2308. S.E. Dann, D.B. Currie, M.T. Weller, M.F. Thomas, A.D. Al-Rawwas, J. Solid State Chem. 109 (1994) 134–144. J. Mizusaki, M. Yoshihiro, S. Yamauchi, K. Fueki, J. Solid State Chem. 58 (1985) 257–266. J.W. Erning, T. Hauber, U. Stimming, K. Wippermann, J. Power Sources 31 (1996) 205–211. Stuart B. Adler, Chem. Rev. 104 (2004) 4791–4843. E.P. Murray, M.J. Sever, S.A. Barnett, Solid State Ionics 148 (2002) 27–34.
Contents lists available at ScienceDirect
Materials Research Bulletin journal homepage: www.elsevier.com/locate/matresbu
Effect of ‘A’-site non stoichiometry in strontium doped lanthanum ferrite based solid oxide fuel cell cathodes Koyel Banerjee, Jayanta Mukhopadhyay* , Madhurima Barman, Rajendra N. Basu* Fuel Cell and Battery Division, CSIR-Central Glass and Ceramic Research Institute, Kolkata 700032, India
A R T I C L E I N F O
A B S T R A C T
Article history: Received 13 January 2015 Received in revised form 9 July 2015 Accepted 2 August 2015 Available online 5 August 2015
Effect of A-site non-stoichiometry in strontium doped lanthanum cobalt ferrite (La1xSrxCoyFe1yO3d, x = 0.4; y = 0.2) is studied in a systematic manner with variation of ‘A’ site stoichiometry from 1 to 0.94. The perovskite based cathode compositions are synthesized by combustion synthesis. Powder characterizations reveal rhombohedral crystal structure with crystallite size ranging from 29 to 34 nm with minimum lattice spacing of 0.271 nm. Detailed sintering studies along with total DC electrical conductivities are evaluated in the bulk form with variation of sintering temperatures. The electrode polarizations are measured in the symmetric cell configuration by impedance spectroscopy which is found to be the lowest (0.02 V cm2 at 800 C) for cathode having highest degree of ‘A’-site deficiency. The same cathode composition exhibits a current density of 2.84 A cm2 (at 0.7 V, 800 C) in anode-supported single cell. An attempt has been made to correlate the trend of electrical behaviour with increasing ‘A’site deficiency for such cathode compositions. ã 2015 Elsevier Ltd. All rights reserved.
Keywords: A. Oxides C. Impedance spectroscopy D. Thermal expansion D. Electrical properties D. Electrochemical properties
1. Introduction To increase the efficiency of the intermediate temperature solid oxide fuel cell (SOFC), the mixed ionic and electronic conductor (MIEC) materials, such as, Ba1xSrxCo1yFeO3d (BSCF), Nd-doped BSCF (BSNCF) and La1xSrxCoyFe1yO3d ((LSCF) have proven promising and demonstrated faster oxygen diffusion and improved surface exchange kinetics [1–9]. Among the MIEC materials as mentioned above, LSCF based cathode materials have good electrical conductivity as well as high oxygen surface exchange coefficient and also have a good oxygen self-diffusion coefficient between the temperature range of 600 C and 800 C. The oxygen self-diffusion coefficient of LSCF is 2.6 109 cm2 s1 at 500 C, which is far better in performance to that of LSM having oxygen self-diffusion coefficient value of 1012 cm2 s1 at 1000 C [10]. There are many factors concerning the duration of stability for LSCF based cathodes. One of the probable mechanisms for voltage deterioration is explained to be slow decomposition of the LSCF perovskite due to partial de-mixing of strontium [11–13]. Mai et al. [14] reported that a small A-site deficiency and high ‘Sr’ content in LSCF cathodes increase the cell performance using gadolinium doped ceria (CGO) interlayer [14]. The long term stability of the cell
* Corresponding authors. Fax: +91 33 24730957. E-mail addresses: [email protected], [email protected] (J. Mukhopadhyay), [email protected], [email protected] (R.N. Basu). http://dx.doi.org/10.1016/j.materresbull.2015.08.002 0025-5408/ ã 2015 Elsevier Ltd. All rights reserved.
using LSCF cathode materials with off-stoichiometry in the ‘A’-site have also been tested [15]. The off-stoichiometry in the ‘A’ site lattice of LSCF-based cathode materials leads to the formation of the stable aliovalent states of the metal ions in the ‘B’ site of the perovskite and hence results in higher conduction. Moreover, change of the oxygen off-stoichiometry by inducing such deficiency in the ‘A’ site also enhances such conduction [10]. It has also been found that ‘A’-site vacancy plays an important role for the performance of the LSCF cathode but long-term cell tests lasting up to 3000 h revealed a degradation of the cell. The sintering of the cathode having off-stoichiometry in the ‘A’ site is also found to be prevalent through the formation of a congruent eutectic melt which results in easy process of grain to grain interlocking at low temperature [16] and hence higher grain conductivity. It has been observed that in iron-based cathodes, reactivity with YSZ electrolyte is significantly reduced. In addition, thermal expansion coefficient of the ferrite based–perovskite is relatively close to that of the electrolyte YSZ. On the other hand, LSCF-type perovskites are generally incompatible with YSZ electrolytes due to undesirable interface reactions. Therefore, a CGO diffusion barrier layer is used to prevent the formation of low conductive compounds without negatively affecting the electrochemical performance [17,18]. Besides, the composition and microstructure of cathode materials has a significant impact on the performance of SOFCs. Hence, rational design of materials composition through controlled oxygen non-stoichiometry and consequent increase in defect aspects can enhance the ionic and
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
electronic conductivities as well as the catalytic properties for oxygen reduction in the cathode material. In the present study, systematic variation of lanthanum concentration in ‘A’-site oxygen non-stoichiometry in ferrite based LSCF materials (LSCF; La1xSrxCoyFe1yO3d, x = 0.4; y = 0.2) has been carried out in the range of 0.54–0.6 while the concentration of dopant ‘Sr’ is kept constant at 0.4. Ferrite based LSCF materials have been chosen as cathode material for the application in intermediate temperature solid oxide fuel cell for the thermal compatibility with the electrolyte YSZ. Powder and bulk characterizations including the DC electrical conductivities of the cathode materials have been carried out in details and clinically correlated with the microstructures. Impedance spectroscopy is employed to investigate the electrode polarizations of the cathodes as thick film in symmetric cell configuration. Electrochemical performances of all cathode compositions are evaluated in conjunction with ceriabased interlayer in the form of Ni-YSZ anode-supported single cells and the performances are correlated with the off-stoichiometry of the ‘A’ site of the synthesized perovskites. 2. Experimental The compositions of the cathodes as investigated under this study are given in Table 1. All the four compositions (LS-1, LS-2, LS3 and LS-4) were prepared by the combustion synthesis technique as reported earlier using starting materials lanthanum (III) nitrate hexahydrate (99%, Sisco Res. Lab. Ltd., India), strontium (II) nitrate (99%, s.d.fine, India), iron (III) nitrate nonahydrate (98%, Merck) and cobalt (II) nitrate hexahydrate (99.5%, E. Merck, Germany) and Lalanine as a fuel (99%, SRL, India) [19]. In all the four combustion synthesis, the ratio of metal nitrate and fuel were maintained 1:1 to prevent the undesired hydrolysis of the respective precursor salts. The as-synthesized powders were calcined at 800 C for 4 h in air. The Co-CGO was prepared using combustion synthesized technique and already reported by our group [20]. For the evaluation of thermal analysis of the cathode materials the viscous precursor gels of all the samples were collected. The differential thermal analysis (DTA) and thermogravimetric analysis (TGA) were carried out from room temperature to 800 C in air at a heating rate of 5 C/min using Simultaneous Thermal Analyzer (STA 449C, Netzsch, Germany). Phase purity of the samples was checked by recording X-ray data in the 2u range 10 –80 using X’Pert Pro software (PANalytical, Philips, Holland). For structural characterization, Rietveld refinement of the powder diffraction profiles and quantitative phase analysis of LS samples calcined at 800 C were carried out using X’Celerator operating at 40 kV and 30 mA using CuK a radiation with step size 0.05 ð2uÞ and step time 75 sec from 15 to 90 for all the samples. The BET surface areas of the powder calcined at three different temperatures were carried out using Quantachrome Instruments (version 9.0). The morphology of both the powder was characterized using transmission electron microscopy (TEM, Tecnai G2 30ST) with evaluation of lattice spacing by high resolution TEM (HRTEM). For the densification studies, powders calcined at 800 C were pressed uniaxially and
307
then sintered in the temperature range of 1050 C–1150 C in air in an interval of 50 C. Field emission scanning electron microscopy (FESEM, Gemini Supra 35 Zeiss) was used to examine the microstructure of the fractured surface of the sintered samples. The thermal expansion coefficients (CTE) of all the sintered samples of different compositions under the present investigation were measured by a dilatometer (NETZSCH, DIL402) up to 900 C with a constant heating rate of 10 C/min. The electrical conductivity of the sintered samples (1050 C–1150 C) was evaluated in the temperature range 500 C–800 C in air by 4probe technique using a Power Source (Agilent E3631A) and a Multimeter (Keithley 2002). Electrochemical impedance of symmetric cells having screen printed LSCF cathode on both the sides Co0.01Ce0.79Gd0.20O2d (Co-CGO) electrolyte in the configuration, LSCF / Co-CGO /LSCF, were carried out in the form of circular disc of diameter of 10 mm with the full range swipe of the AC frequency of 101 Hz f 106 Hz. While the sintering temperature of bulk Co-CGO was kept at 1100 C, the thick films of the composite cathode were fired at 900 C [21]. For electrochemical performance evaluation of the cathodes of varied composition, thick film paste prepared using such powders were screen printed onto anodesupported planar half-cell of configuration, NiO-YSZ/YSZ. The halfcell was fabricated by tape casting method which is also reported our group previously [22]. Single cells in the form of coupon cells (16 mm diameter, 1.5 mm thick with active cathode area of 0.3 cm2) of configuration Ni-YSZ/YSZ/CoCGO/LSCF were fabricated after heat treatment of screen printed LSCF-based cathodes at 950 C. Co-CGO was used as an interlayer in between the electrolyte and cathode to reduce the unwanted reaction and thermal compatibility. The electrochemical measurements of such single cells were carried out in the temperature range 700 C– 800 C using an in-house electrochemical measurement setup. Moist H2 (3% H2O) was used as a fuel on the anode side and oxygen was fed on the cathode side. During the measurement, the flow rates for both the fuel gas and O2 were maintained at 100 SCCM. The best performing cathode was also tested with air for comparison with the results of oxygen and the flow rates for the same is maintained at 500 SCCM. 3. Results and discussion The ‘A’-site deficient LSCF-based cathode plays a major role in improving the performance of SOFC by the increase of the electronic conductivity and catalytic behavior. However, the deficiency in ‘A’ site has a limitation with respect to the substitution of Sr2+ in the trivalent site (La3+) of perovskite having formula La1xSrxCo1yFeyO3d. It has been reported that at x > 0.4 makes the perovskite structure metastable and the ‘A’ site deficiencies greater than 5% results a negative effect on measured performance [23]. Keeping the view of the above, the ‘Sr’ content in the LSCF compositions has been kept constant at 0.4 whereas, the ‘A’ site deficit has been made upto 6% with respect to ‘La’ site as shown in Table 1.
Table 1 Compositions of the LSCF powders and relative densification and CTE values of the sintered bulk samples. Sample ID
Composition
LS-1 LS-2 LS-3 LS-4
La0.54Sr0.40Co0.20Fe0.80O3d La0.56Sr0.40Co0.20Fe0.80O3d La0.58Sr0.40Co0.20Fe0.80O3d La0.60Sr0.40Co0.20Fe0.80O3d
Relative densification (%) at the following sintering temperatures
1050 ( C)
1100 ( C)
1150 ( C)
80.1 80.2 79.7 76.2
90.6 87.2 86.4 84.8
94.5 94.1 89.5 88.3
Coefficient of thermal expansion (CTE) of sintered (1150 C) bulk cathodes in the temperature range 30–900 C 106 (K1) 16.88 15.46 15.72 15.76
308
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
3.1. Thermal analysis The thermal analysis of the precursor gels collected during the synthesis for all the four LSCF compositions have been shown in Fig. 1. Mass loss in the TG curve occurs at 158 C may be attributed due to the initial evaporation which is followed by an initiation of thermal decomposition of metal nitrate-alanine complex in the temperature range of 180 C–190 C as also reflected by the presence of the sharp exothermic peak at 182 C in DTA. The decomposition is found to be progressive enough in the temperature range of 248 C–255 C which is also observed by the presence of a shallow exothermic hump in DTA. The final decomposition of the metal-alanine complex takes place in the broad temperature range of 300 C–480 C with a specific exothermic peak appearing in the temperature range of 410 C–416 C as also observed in DTA. 3.2. Powder characterizations The X-ray diffractograms of the all four compositions of LSCF powders calcined at 800 C are shown in Fig. 2. Reitveld analyses for all the powder samples calcined at 800 C are carried out and all the peaks have been indexed accordingly. The calculated lattice parameters as obtained from Rietveld analyses of the X-ray diffractograms are found to be a = b = 5.49 Å and c = 13.40 Å for three cathode compositions viz. LS-1 (La0.54Sr0.4Co0.2Fe0.8O3d), LS-2 (La0.56Sr0.4Co0.2Fe0.8O3d) and LS-3 (La0.58Sr0.4Co0.2Fe0.8O3d). The same is found to be a = b = 5.41 Å and c = 12.96 Å for LS-4 (La0.6Sr0.4Co0.2Fe0.8O3d). The average crystallite sizes of these are within the range between 29.6 nm and 37.9 nm. Since the ionic radius of Co4+ (0.067 nm) is smaller than that of Fe4+ (0.0725 nm), formation of higher concentration of Co4+ as a result of higher degree of substitution in ‘A’ site lattice of the LSCF-based perovskite results in lattice shrinkage. The theoretical densities of the crystallites are found to be 6.10, 6.17, 6.26 and 6.38 g cc1 for cathode composition LS-1, LS-2, LS-3 and LS-4 respectively. ‘A’-site deficiency of the LSCF-perovskite makes a pronounced effect on the surface area of the calcined materials. The powder samples calcined at 800 C was used to surface area measurements which are evaluated using gas adsorption BET isotherm. LS-1 with largest ‘A’-site deficiency in ‘La’ is found to have lowest surface area (6.45 m2/gm), which might be attributed to the agglomeration owing to ease of materials diffusion across the particle boundary because of higher degree of defects in the perovskite. While surface area values for LS-3 and LS-4 are found to be 10.68 m2 gm1 and 10.35 m2 gm1 respectively, it is 7.78 m2 gm1 for the composition LS-2. The powder morphology of the compositions is also
Fig. 2. XRD patterns of the powder samples (a) LS-1, (b) LS-2, (c) LS-3 and (d) LS4 calcined at 800 C.
evaluated using TEM (Fig. 3). It is found that the average particle size ranges in between 50 nm and 100 nm. The particle locking across the boundary is observed to be more prominent in LS1 having highest ‘A’ site deficiency which is also reflected in the value of the lowest surface area for the same composition. The interplanar distances have been calculated from HRTEM images which show the orientation of crystal planes accordingly (Inset of Fig. 3). Selected area electron diffraction patterns shown in insets of Fig. 3 exhibit rhombohedral crystal symmetry. The interplanar spacing is reduced from 0.381 nm to 0.275 nm as the ‘A’-site deficiency is increased from the composition LS-4 (La0.60Sr0.40Co0.20Fe0.80O3d) to LS-1 (La0.54Sr0.40Co0.20Fe0.80O3d). This trend of interplanar spacing is found to have corroboration with the lattice shrinkage as was observed in the Reitveld analyses. 3.3. Densification and thermal expansion studies
Fig. 1. TG and DTA analysis of various LSCF gels in the temperature range of 30– 800 C.
To find out the effect of the off-stoichiometry in the ‘A’-site lattice particularly in the place of ‘La’, the densification study of the samples sintered in the temperature range of 1050 C–1150 C is shown in Table 1. As expected the density of all the samples (LS-1– LS-4) increases with increase in sintering temperature. While, LS-1 and LS-2 reaches near 95% of the theoretical density at 1150 C, the percentage densification remains below 90% for LS-3 and LS-4. The densification data also supports the microstructures of the fractured surfaces of the sintered samples obtained using FESEM (Fig. 4). As reported in the literature the LSCF-based materials tend to get sintered by the presence of liquid phase [16,24]. The presence of higher degree of deficiency in ‘La’ site of LSCF-based perovskite is likely to facilitate the formation of liquid phase during the sintering. It is observed from Fig. 4 that development of well crystalline grains are prevalent in case of LS-1 having highest deficit in ‘A’-site of the perovskite which might be because of the result of recrystallization of the grains from the melt phase
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
Fig. 3. TEM image of the particles (a) LS-1, (c) LS-2, (e) LS-3, (g) LS-4 with HRTEM and Selected area electron diffraction pattern (SAED) image (insets).
Fig. 4. FESEM image of the fractured surface of the bulk samples (a) LS-1, (b) LS-2, (c) LS-3, (d) LS-4 sintered at 1050 C.
309
310
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
(Fig. 4a). As the recrystallization passes through a process of recovery, grain sizes are found to be smaller i.e. in the range of 0.2– 0.5 mm [25]. In contrast, LS-4 having lowest level of ‘A’-site deficit exhibits higher percentage of melt phase in the microstructure and hence well-developed grains are not much prominent in such case evident from Fig. 4d. LS-3 and LS-2 falls in between the extremes and hence the sintered microstructures are also found with the grains having a balance between the presence of melt phase and recrystallized grains. However, the microstructure of LS-3 having higher degree of ‘A’ site deficit is associated with higher melt phase compared to that of LS-2. The thermal expansion co-efficient of the bulk samples sintered at different temperatures have been evaluated which are also presented in Table 1. It is found that irrespective of composition of the cathodes under present investigation, the coefficient of thermal expansion increases with increase in sintering temperature. The highest lattice shrinkages are found with LS-1 having highest degree of ‘A’ site deficit. This attributes to the higher coefficient of thermal expansion (CTE) of the perovskite and is found to be 16.88 106 (K1) in the temperature range 30 C–900 C. The thermal expansion coefficient of the LSCF-based is found least [15.46 106 (K1)] with LS2 sintered at 1150 C. However, CTE is found to be in the range of 15.46–15.76 106 (K1) for the perovskites having lower in ‘A’site deficiency e.g., in LS-2 to LS-4. Such a higher level of thermal mismatch for LSCF-based cathodes with that of the YSZ electrolyte may, however, be minimized by laying down a thin and dense particulate coating of doped ceria based interlayer [26,27].
3.4. Electrical conductivity The DC electrical conductivity of the sintered samples for all the four compositions are examined in the temperature range 500 C– 800 C with an interval of 25 C. The temperature dependent Arrhenius plots for DC electrical conductivity for cathode compositions are shown in Fig. 5. It is found that irrespective of all the compositions, the total electrical conductivity increases with increase in sintering temperature which attributes to the higher degree of sinterability of the samples at high temperature. However, all the compositions exhibit a conductivity transformation from semiconductor to metallic type upon increase of measurement temperature from 500 C–800 C. It is reported in our earlier communications that such shift of electrical conductivity is natural in case of LSCF-based materials [27]. It has also been reported that the charge disproportionation of Co3+ occurs in La1xSrxCo1yFeyO3 based compositions increases the concentration of electronic carriers [28] which is presented in Eq. (1). 2Co3+ ! Co4+ + Co2+
(1)
Defect model associated with the defect structure of the LSCF based materials proposed by different groups has been given by the following Kröger–Vink notations [10,29,30] SrO þ Fe Fe
LaFeO3 Sr0La þFe Fe þO o
!
Fig. 5. Arrhenius plots of the electrical conductivity of the bulk as a function of temperature.
ð2Þ
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
1 2Co Co þ O o ! 2COCo þ V o þ O2 ðgÞ 2
ð3Þ
1 2Fe Fe þ O o ! 2FeFe þ V o þ O2 ðgÞ 2
ð4Þ
311
highest conductivity exhibits the activation energy of 0.04 eV in the semiconducting region whereas, the same is found to be 0.17 eV in the range of metallic type conductivity. 3.5. Electrochemical studies
In Eq. (2) it is observed that when Sr2+ is doped as SrO in the lattice site of La3+, decrease of positive charge gives rise to a negative 0 lattice viz. Sr La . The deficiency in the positive charge thus generated is compensated by the conversion of Co3+ to Co4+ and Fe3+ to Fe4+ which is responsible for the electronic conduction. The charge disproportionation reaction as shown in Eq. (1) is primarily responsible for generating Co Co site or Fe Fe. Eqs. (3) and (4) basically show how at elevated temperature oxygen vacancy is created in the LSCF lattice due to conversion of B-site transition metal elements Co and Fe from their higher oxidation state (+4) to stable lower oxidation state (+3). Presence of higher degree of ‘A’site deficiency leads to formation of higher aliovalent cation of ‘Co’ and hence enhances the chances of disproportionation reaction as mentioned above. LS-1 (La0.54Sr0.40Co0.20Fe0.80O3d) sintered at a temperature of 1150 C shows a maximum electrical conductivity of 320 S cm1 measured at 800 C whereas, LS-4 (La0.60Sr0.40Co0.20Fe0.80O3d) with stoichiometric composition at ‘A’ site and prepared under similar condition exhibits the lowest electrical conductivity of 235 S cm1 measured at 800 C. It is observed that all the compositions of the cathodes exhibits lower activation energy in the regime of semiconducting electrical conduction (lower temperature range of 500 C–650 C) compared to that of the metallic type of conductivity (higher temperature range of 675 C–800 C). LS-1 sintered at 1150 C and having
The electrode polarization of the cathode compositions has been examined using impedance spectroscopy in the form of screen printed thick film applied onto the both sides of Co-CGO disks of thickness 1 mm. It is found that the polarization values decrease as the ‘A’ -site deficiency increases due to higher rate of surface exchange kinetics [31–33]. The characteristics impedance spectra for all the four compositions have been given in Fig. 6. Impedance decreases with increase in the ‘A’-site deficiency which is clear from the impedance spectra obtained from of SYM-LS-4 to SYM-LS-1. It is observed from the figure that impedance shifts towards the higher frequency side much faster for SYM-LS-1–SYMLS-3 compared to that of the stoichiometric composition of SYMLS-4. In the low temperature regime, the impedance spectra though exhibits two incomplete semicircles (particularly observed for impedance spectra at 600 and 700 C for SYM-LS-4), the same is found to be transformed to single semicircle upon increase in the measurement temperature. The inset of Fig. 6 shows the spectra for all the compositions but with a much higher resolution of the scale values. From Fig. 7, it can be observed that at lower temperature regime, the different processes viz. bulk impedance happening at the higher frequency region and electrode phenomena happening in the middle and low frequency field can be differentiated by the presence of two different values of |Z|. The Z1 signifies primarily the charge transfer corresponding to the transfer of active oxygen species involving vacancies/ions or adsorbed atom from the
Fig. 6. Nyquist plots measured at different temperatures for the four cathode compositions.
312
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
Fig. 7. Bode plots measured at different temperatures for the four compositions.
electrode to the electrolyte surface and the next regime wherein combination of Z1 + Z2 attributes to the total contributions involving the solid state diffusion coupled with chemical reaction occurring on the electrode surface [34,35]. LSCF being the mixed ionic and electronic conducting electrode, the electronic contribution primarily comes out of the grain of the LSCF electrode and ionic conduction arises out of the defect conduction and hence happens preferably at the grain boundary junctions. It is also observed from the figure that with increase in measurement temperature 700 C, distinction of these two regions are demolished and a single step ǀZǀ is found to prevail throughout the frequency region. This is also observed from the Nyquist plot (Fig. 6) that the high frequency region of the semicircle is severely depressed upon increasing the temperature and thereby the electrode response takes over the bulk impedance. In case of LS1 having composition La0.54Sr0.40Co0.20Fe0.80O3-d, it is found that the shift of ǀZǀ from Z1 to Z1 + Z2 is found to be in the frequency field of 1.5–1.4 105 Hz compared to 3.6 - 2.8 105 Hz as observed for LS4, the stoichiometric composition (La0.60Sr0.40Co0.20Fe0.80O3d). This indicates that the cathode response involving oxygen diffusion, surface adsorption of molecular oxygen and subsequent reduction to oxygen ion using electron transfer process,
incorporation and transport of oxygen ion from the cathode surface to the electrolyte interface [34] is more effective for the composition having the highest degree of ‘A’ site deficiency. While the high frequency intercepts represents the ohmic polarization (Ro) of the symmetric cell which consists of the ohmic polarization of CoCGO electrolyte and that of the electrode polarization including the interfaces, the low frequency intercept depicts the total resistance (Rt). Cathode polarization (Rp) is calculated from the difference of the high frequency and low frequency intercept. The calculated Rp for all the cathodes are presented in the Table 2. Cathode polarizations are found to decrease with increase in the ‘A’-site deficiency and found to be the lowest (0.02 V cm2) for SYMLS-1 and the highest value for the same is noticed with SYM-LS-4 (0.12 V cm2). SYM-LS-2 and SYM-LS-3 are found with intermediate values of the Rp. The ohmic polarization (Ro) is also found minimum with SYM-LS-1 (2.49 V cm2). Fig. 8 represents the Arrhenius plot for the ohmic and cathodic polarizations for the cathode compositions. SYM-LS-1 and SYM-LS-2 exhibit lower activation energy of ohmic polarizations (0.21 and 0.17 eV for LS-1 and LS-2 respectively) compared to SYM-LS-3 and SYM-LS-4 (0.47 and 0.44 eV for SYM-LS-3 and SYM-LS-4 respectively). The same trend is also followed for the activation energies of cathode
Table 2 Impedance data and electrochemical performance of single cells with different cathode compositions. Sample
LS-1 LS-2 LS-3 LS-4
Impedance data
Electrochemical performance
Ohmic polarization (V cm2)
Electrode polarization (V cm2)
Current density at 0.7 V at 800 C (A cm2) Oxygen
Air
2.49 3.58 3.24 3.12
0.02 0.03 0.04 0.12
2.84 2.49 2.19 2.10
2.30 2.04 1.78 1.70
Cell ASR values in oxygen at 800 C (V cm2)
0.13 0.14 0.17 0.15
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
313
Fig. 8. Arrhenius plots of (a) Electrode polarization (Rp) and (b) Ohmic polarization (Ro) for the various cathode compositions.
polarization viz. SYM-LS-1 and SYM-LS-2 are found with activations energies of 0.63 and 0.56 eV for Rp and the same is 1.42 eV and 1.54 eV for SYM-LS-3 and SYM-LS-4 respectively. It is found from the impedance spectroscopy that a low frequency arc appears in the lower temperature especially in the range 650 C which is primarily associated with the diffusion and charge transfer reaction for the oxygen reduction reaction. In order to investigate the effect of charge transfer and diffusion polarization on the cathode performance having the highest ‘A’ site deficiency, impedance study has been carried out for the Sym-LS-1 in different environment using different oxygen partial pressure at 107 (mixture of Ar and air), 0.21 (air) and 1 atm (in oxygen). It has been observed that at lower temperature range and in lower oxygen partial pressure, the cathode polarization is initially governed by diffusion process followed by the charge transfer being captured at the lower frequency of the impedance spectra. Fig. 9 shows the Nyquist plots at an intermediate temperature of 650 C at different oxygen partial pressure. At a specified temperature of 650 C, with decreasing partial pressure of oxygen for the LSCF-based cathode materials, the sizes of low-frequency arc is found to be enlarged gradually. Simultaneously, it is also observed that the arc is having lower values than that of the highfrequency arc. While the low frequency arc for Nyquist plot is found with total impedance (Rt) of 1561 V cm2 at 5.012 Hz in Ar + air atmosphere, the same is observed to be only 5.19 V cm2 at 1.259 Hz in oxygen rich environment. Rt is found to be intermediate in air atmosphere viz. 5.29 V cm2 at 1.585 Hz (insets of Fig. 9). This is a typical indication that the primary rate-limiting mechanism being related with the charge-transfer processes for oxygen at the air electrode [36]. As the temperature increases (>700 C), the diffusion being faster through the porous electrodes, the lower frequency arc disappears emphasizing the fact that under cathodic polarization the oxygen reduction kinetic process is not only governed by charge-transfer mechanism but also strongly by adsorption–desorption and diffusion of oxygen. At high temperature though the concentration polarization is decreased in oxygen
rich environment and charge transfer polarization is also found to decrease because the activation of the cathode is found with faster kinetics at higher temperature. The exchange current density (io) for SYM-LS-1 is calculated at 800 C from the impedance spectroscopy using the low field approximation technique [21] wherein the rate limiting steps for oxygen reduction is considered as single and the same is found to be 1.16 A cm1, which is significantly large for such MIEC conducting cathodes. Single cell electrochemical performances using each of the cathode compositions have been evaluated using hydrogen as the fuel and oxygen as the oxidant and shown in Fig. 10. A systematic increase in the current density is observed for the compositions with increase in the ‘A’–site deficiency. While LS-1 exhibits a current density as high as 2.84 A cm2 and power density 1.99 W cm2, LS-4 shows the cell performance as 2.10 A cm2 with power density 1.47 W cm2 when measured at 800 C corresponding to cell voltage of 0.7 V. However, the peak current density as well as power density for LS-1 is found to be still higher and exhibits 3.78 A cm2 and 2.27 W cm2 when taken at individual cell voltage of 0.6 V. Cell area specific resistances (ASR) are calculated considering the slope of the linear portion of the I–V curve and is found to be minimum (0.13 V cm2) with cell having LS-1 as the cathode. The highest performing cathode is also tested with the air and a drop of 19% is observed in the peak performance. The peak current density for LS1 with air is found to be 3.06 A cm2 when measured at 800 C with the individual cell voltage of 0.6 V. The same is found to be 2.80 A cm2 using air as the oxidant at cell operating voltage of 0.7 V. All the cells have been tested for 100 h. The cell performances for all the compositions along with cell ASR are given in Table 2. These cell performances overshadow the performance of our earlier reported LSCF-based cathodes prepared by combustion technique [15] and hence establish the effect of the ‘A’-site deficiency in the LSCF-based perovskite lattice. The FESEM images of the cross section of the single cells of all cathode compositions are shown in Fig. 11. From the cross sectional micrographs, well defined porous Ni-YSZ based anode, dense YSZ based electrolyte
Fig. 9. Nyquist plots of LS-1 at different oxygen partial pressure at 650 C.
314
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
Fig. 10. Electrochemical testing of SOFC single cell (coupon cell) at different temperatures with cathode compositions LS-1, LS-2, LS-3 and LS-4.
Fig. 11. FESEM images of the single cells with LS-1, LS-2, LS-3 and LS-4 cathode compositions.
K. Banerjee et al. / Materials Research Bulletin 72 (2015) 306–315
and porous cathodes of LSCF compositions in conjunction with CoCGO based interlayer are found vivid. However, cell microstructures with LS-1 and LS-2 having higher degree of ‘A’ site off stoichiometry is found with higher densification for screen printed cathode layer compared to their stoichiometric compositions because of ease of sintering for ‘A’ site deficit cathodes as discussed in the earlier section. 4. Conclusions Nanocrystalline LSCF-based cathode compositions (La1xSrxCoyFe1yO3d, x = 0.4; y = 0.2) varying the A-site stoichiometry from 1 to 0.94 (with a decrement of 0.02) in the ‘La’ site of the perovskite have been synthesized using combustion synthesis technique. Thermal analyses revealed a multistage decomposition. XRD analysis of all the four compositions shows rhombohedral crystal structure upon calcination at 800 C. Surface area evaluation of the calcined powders revealed that with increase in the ‘A’– site deficiency increases the materials diffusion across the particulate boundary and hence increase the particle sizes by forming agglomerates. CTE of the LSCF-based perovskite decreases with decrease in ‘A’ site deficiency and found highest in LS-1 (16.88 106). The CTE is found to be in the range of 15.46– 15.76 106 (K1) for the other perovskites of LS-2–LS-4. However, such mismatch in CTE be minimized when applied with a doped ceria based interlayer. The DC electrical conductivity also follows the same trend and is found to be highest (320 S cm1) with LS1 sintered at a temperature of 1150 C. Though at low temperature and low partial pressure of oxygen, the diffusion polarization is found with low frequency arc, the same is found negligible at high temperature and oxygen rich environment. Dependence of charge transfer reaction of the oxygen reduction is also found by the presence low frequency arc at moderate temperature of 650 C and even observed at oxygen rich environment. The impedance spectra of the screen printed cathode compositions revealed a single step deconvolution when measured 700 C and the calculated exchange current density for the highest ‘A’ site deficit cathode is found to be as high as 1.16 A cm2 at 800 C. The lowest cathode polarization is obtained for LS-1 which is found to be 0.02 V cm2 at the same measurement temperature. The current density for single cell prepared with LS-1 cathode material is observed to be the highest i.e. 2.84 A cm2 when operated with cell voltage 0.7 V at 800 C. Thus, LSCF-based cathode compositions having synthesized with maximum 6% of deficit at ‘La’ site is found to have good structural integrity with effective process of sintering in the bulk conditions, good DC electrical conductivity and the least cathode polarizations resulting in the highest electrochemical performance in the single cell configuration. Acknowledgements One of the authors (KB) is thankful to CSIR for providing Senior Research Fellowship (SRF). The authors are thankful to the Director,
315
CSIR-CGCRI for his kind permission to publish the work. The authors also acknowledge the financial support of Department of Science and Technology. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]
[12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36]
W. Zhou, R. Ran, Z. Shao, J. Power Sources 192 (2009) 231–246. S. Li, Z. Lu, X. Huang, W. Su, Solid State Ionics 178 (2008) 1853–1858. Y. Zeng, Y.S. Lin, S.L. Swartz, J. Membr. Sci. 150 (1998) 87–98. A. Raj, A.S. Nesaraj, M. Kumar, F. Tietz, H.P. Buchkremer, D. Stoever, J. New Mater. Electrochem. Syst. 7 (2004) 145–151. N.J. Simrick, A. Bieberle-Hütter, T.M. Ryll, J.A. Kilner, A. Atkinson, J.L.M. Rupp, Solid State Ionics 206 (2012) 7–16. X. Ding, X. Kong, J. Jiang, C. Cui, X. Guo, Mater. Res. Bull. 45 (2010) 1271–1277. L. Conceicao, A.M. Silva, N.F.P. Ribeiro, M.M.V.M. Souza, Mater. Res. Bull. 46 (2011) 308–314. L. Lu, Q. Shi, Y. Yang, H. Zhang, Mater. Res. Bull. 47 (2012) 1016–1020. L. Lu, Y. Guo, H. Zhang, J. Jin, Mater. Res. Bull. 45 (2010) 1135–1140. C. Sun, R. Hui, J. Roller, J. Solid State Electrochem. 14 (2010) 1125–1144. Z. Feng, Y. Yacoby, M.J. Gadre, Y.L. Lee, W.T. Hong, H. Zhou, M.D. Biegalski, H.M. Christen, S.B. Adler, D. Morgan, Y.S. Horn, J. Phys. Chem. Lett. 5 (2014) 1027– 1034. L. Adijanto, R. Küngas, F. Bidrawn, R.J. Gorte, J.M. Vohs, J. Power Sources 196 (2011) 5797–5802. H. Kishimoto, N. Sakai, T. Horita, K. Yamaji, M.E. Brito, H. Yokokawa, Solid State Ionics 179 (2008) 1338–1342. A. Mai, V.A.C. Haanappel, S. Uhlenbruck, F. Tietz, D. Stover, Solid State Ionics 176 (2005) 1341–1350. F. Tietz, A. Mai, D. Stover, Solid State Ionics 179 (2008) 1509–1515. X. Tan, Z. Wang, K. Li, Ind. Eng. Chem. Res 49 (6) (2010) 2895–2901. A. Mai, V.A.C. Haanappel, F. Tietz, D. Stover, Solid State Ionics 177 (2006) 2103– 2107. D. Waller, J.A. Lane, J.A. Kilner, B.C.H. Steele, Solid State Ionics 86–88 (1996) 767–772. A. Dutta, J. Mukhopadhyay, R.N. Basu, J. Eur. Ceram. Soc. 29 (2009) 2003–2011. A. Dutta, A. Kumar, R.N. Basu, Electrochem. Commun. 11 (2009) 699–701. K. Banerjee, J. Mukhopadhyay, R.N. Basu, Int. J. Hydrogen Chem. 39 (2014) 15754–15759. R.N. Basu, A. Das Sharma, A. Dutta, J. Mukhopadhyay, Int. J. Hydrogen Energy 33 (2008) 5748–5754. T. Striker, J.A. Ruud, Y. Gao, W.J. Heward, C. Steinbruchel, Solid State Ionics 178 (2007) 1326–1336. F. Schulze-Küppers, S. Baumann, F. Tietz, H.J.M. Bouwmeester, W.A. Meulenberg, J. Eur. Ceram. Soc. 34 (2014) 3741–3748. R. Zuo, S. Su, Y. Wu, J. Fu, M. Wang, L. Li, Mat. Chem. Phys. 110 (2008) 311–315. J.P. Martinez, D.M. Lopez, C.S. Baustista, A.J.D.S. Garcia, J.C.R. Morales, J.C. Vazque, P. Nunez, Bol. Soc. Esp. Ceram. 49 (2010) 15–22. J. Mukopadhyay, Rajendra N. Basu, J. Power Sources 252 (2014) 252–263. H. Duanping, X.U. Qin, Z. Feng, C. Wen, L. Hanxing, Z. Jian, J. Wuhan Univ. Technol. Mater. Sci. Ed. 23 (2008) 80–84. E.V. Bongio, H. Black, F.C. Raszewski, D. Edwards, C.J. McConville, V.R.W. Amarakoon, J. Electroceram. 14 (2005) 193–198. A.N. Petrov, O.F. Kononchuk, A.N. Petrov, O.F. Kononchuk, A.V. Andreev, Solid State Ionics 80 (1995) 189–199. E.N. Armstrong, K.L. Duncana, E.D. Wachsman, Phys.Chem. Chem. Phys. 15 (2013) 2298–2308. S.E. Dann, D.B. Currie, M.T. Weller, M.F. Thomas, A.D. Al-Rawwas, J. Solid State Chem. 109 (1994) 134–144. J. Mizusaki, M. Yoshihiro, S. Yamauchi, K. Fueki, J. Solid State Chem. 58 (1985) 257–266. J.W. Erning, T. Hauber, U. Stimming, K. Wippermann, J. Power Sources 31 (1996) 205–211. Stuart B. Adler, Chem. Rev. 104 (2004) 4791–4843. E.P. Murray, M.J. Sever, S.A. Barnett, Solid State Ionics 148 (2002) 27–34.