Chemical Reactivity and Cathode Properties of LaCoO3 on Lanthanum Silicate Oxyapatite Electrolyte

Chemical reactivity and cathode properties of LaCoO3 were investigated using new oxide ion conductor, lanthanum silicate oxyapatite. The LaCoO3 is found to be good candidate for cathode of the lanthanum silicate oxyapatite solid-electrolyte since no chemical reaction occurred between the LaCoO3 and lanthanum silicate oxyapatite heating at 1273 K for 60 h in air. Based on electrochemical measurements, lower overpotential between the LaCoO3 and lanthanum silicate oxyapatite was confirmed compared to the overpotential at YSZ/LaCoO3 interface. From analysis on the extended interfacial conductivity as function of oxygen activity at the triple phase boundary at fixed temperature, the overpotential evaluated by impedance spectra is the rate limiting process by oxygen diffusion on the LaCoO3 surface. Comparing to the bulk conductivity of LaCoO3, the electrode resistance evaluated by impedance spectra was confirmed to be different from the electrical transport properties of the LaCoO3 bulk.


Introduction
Lanthanum cobaltite with provskite structure (LaCoO 3 ) is famous mixed conductor by electron holes and oxide ions and is the candidate for applications to cathode for solid oxide fuel cell (SOFC) and oxygen permeable membrane [1][2][3][4][5][6][7]. Although the LaCoO 3 shows high performance as SOFC cathode when yttria stabilized zirconia (YSZ) is used as solid electrolyte [1,2], decrease in the cathode performance with continuous SOFC operation condition. Cause of the decrease in the cathode performance is the formation of insulator phase, La 2 Zr 2 O 7 , by the chemical reaction between the YSZ and LaCoO 3 [8][9][10]. Hence, it is necessary to replace the solid electrolyte from YSZ if one wants to utilize the excellent cathode property of LaCoO 3 for SOFC.
Lanthanoid silicate oxyapatites have known as new oxide ion conductors since Nakayama et al. have reported their high oxide ion conductivity from 1995 to 1998 [11][12][13]. In particular, the lanthanum silicate oxyapatite shows high oxide ion conductivity even in the ceramic electrolyte. In addition, the oxide ion conductivity of the lanthanum silicate oxyapatite is higher than the conductivity of YSZ below about 873 K because of its low activation energy [13,14]. Hence, the lanthanum silicate oxyapatite is the candidate of the solid electrolyte material for middle temperature operation SOFC [15]. On the other hand, no research has been done to develop specified electrode materials for the lanthanum silicate oxyapatite. We focused on the LaCoO 3 which could not apply as cathode material on the YSZ electrolyte even though the LaCoO 3 itself has excellent potential.
In this study, the chemical reactivity at the LaCoO 3 /lanthanum silicate oxyapatite interface and fundamental research on the cathode performance was investigated. For investigation of the cathode properties, we applied new analysis theory by extended interfacial conductivity [16,17]. Based on this

Preparations of the cell
Cathode properties of the LCO on LSO electrolyte were investigated by a half cell configuration. In order to fabricate the LCO electrode, both surfaces of the sintered LSO ceramics were polished using #100 to #2000 waterproof abrasive papers. In addition, thin groove was made using a diamond wheel for binding platinum wire.
Two polyester-mask-seal with hole diameter of 6 mm were attached at the center on both surfaces of the LSO disk. Platinum paste (TR-7907) Tanaka Kikinzoku Kougyo, Ltd., Japan) was attached inside of the hole and then, dried the paste at 373 K for 30 min. Then, 5 ml of the LCO ink was pipetted and pulled into the seal hole. After drying the LCO ink, mask seals on both side of the LSO pellet were removed. Then, both electrodes were attached by heating at 1273 K for 5 h in air.
Before the electrochemical measurements, platinum wire was wound at the side of the pellet and then, small amount of the platinum paste was painted between the LSO pellet and the platinum wire in order to make electrical contact among them.

Electrochemical Measurements
Electrochemical measurement was performed by three terminal method in one room condition. One platinum wire attached to the side of the LSO pellet was wired to a platinum lead wire. The cell was sandwiched by two platinum meshes fixed on porous alumina plates. Two platinum wires were welded to each mesh to connect a load circuit and sensing circuit of the potentio/galvanostat (R-600, Gamry, USA). Three electrochemical measurement techniques, cyclic voltammetry, electrochemical impedance with-and without d.c. voltage convolution were used for the analysis. The electrochemical measurements were made at the temperature range between 723 K and 1173 under the oxygen partial pressure from 10 -2 atm to 1 atm. The oxygen partial pressure around the cell was changed using Ar + O 2 gas mixture. The oxygen partial pressure was monitored by zirconia oxygen sensor fixed at the downstream side from the electrochemical cell.
Cyclic voltammetry was measured in air at voltage scan rate of 5 mV/s. Potential difference rage among the working electrode (LCO) and reference electrode (Pt) was between -1.0 V and 1.0 V. Electrochemical impedance was measured in frequency range between 1 MHz and 0.01 Hz depending on the temperature. The electrochemical properties were analyzed based on the extended interfacial conductivity theory [18,19].

Chemical reactivity
By comparison of the XRD patterns of the initial mixed powder and the heated powder at 1273 K for 60 h, the LCO and LSO seemed to be equilibrated as two phase coexisting state because the patterns were found to be almost the same and new phase formation was not observed except for a sharpening of the LCO peaks around 2θ = 33 to 35 as shown in Fig. 1. Peak width of the XRD is known to be related to the crystallinity and grain size. Therefore, the main reason for variation of the XRD peaks by heating at 1273 K for 60 h is improvement of the crystallinity and grain growth of the LCO powder.
On the other hand, XRD pattern of initial mixed powder by LCO and YSZ changed into different one by heating at 1273 K for 60 h by showing in Fig. 2(a) and (b). New phase showing strong intensity peaks were indexed by the peaks of La 2 Zr 2 O 7 [24] as shown in Fig. 2(a). In addition to the peaks by YSZ and La 2 Zr 2 O 7 , weak peaks by Co 3 O 4 and LCO were confirmed ( Fig. 2(b)). It can be concluded that the LCO is very reactive to YSZ and easily decomposed at 1273 K. Hence, it is difficult to use the LCO as cathode of YSZ electrolyte.
In contract to the case of the LCO/YSZ interface, the LCO is stable even though the LCO is attached to the LSO electrolyte at 1273 K and therefore the LCO is found to be unique candidate as cathode material of LSO electrolyte.

Electrochemical impedance under open cell circuit condition
Typical impedance spectra corrected at 1023 K were shown in Fig. 3 as function of oxygen partial pressure. All of the spectra were found to be constructed by two decompressed arcs at high frequency and low frequency region. With respect to the high frequency region arc, it seemed to be distorted signal by the instrument because similar distorted spectrum was measured when the circuit was connected to a test cell constructed by resistors and condenser. Therefore, only the spectra shown in low frequency region were analyzed by the equivalent circuit. The equivalent circuit used in this study is shown in Fig. 4 to fit the obtained spectra. The R b is the bulk resistance of the electrolyte, R 2 and R 3 are the resistance of the electrode response, respectively. The CPE 2 and CPE 3 are the constant phase element of the electrode response, which are the empirical element to give the Cole-Cole type relaxation [27]. Reason for the separation with two elemental parts by R 2 and CPE 2 , and R 3 and CPE 3 is to fit the distorted spectra. It is necessary to emphasize that no clear physicochemical meaning of R 2 , R 3 , CPE 2 , and CPE 3 , and is reasonable to consider that sum of these elemental parts should be treated  [23] and LSO [18] are shown for comparison.  as electrode impedance (Z electrode ) as shown in Fig. 4. Therefore, sum of the R 2 and R 3 are treated as electrode resistance (R E ), hereafter. Based on the interfacial conductivity proposed by Mizusaki et al. [28,29], the interfacial conductivity (σ E ) is defined by where A is the electrode area. The relationships between the σ E and oxygen partial pressure show that the σ E is almost independent of oxygen partial pressure (Fig. 5). It is necessary to emphasize that the σ E is not an electrical conductivity but a chemical one [16,17,28,29]. Although it is difficult to conclude the rate determining step by the p(O 2 ) dependences of σ E because the measured p(O 2 ) range is narrow, the p(O 2 ) dependences on the σ E becomes weak if about a half of the adsorption site on the metallic electrode surface is fulfilled by the adsorbed oxygen [28,29]. On the other hand, if the σ E corresponds to the oxygen permeation by mixed conductive properties in LaCoO 3 , the σ E should be proportional to 1/4 power of p(O 2 ) by considering into the p(O 2 ) dependence of the partial conductivity of oxide ions in LaCoO 3 [30].

Cyclic voltamometry
Relationship between current and potential difference between the working electrode and reference electrode measured by cyclic voltammetry are shown in Fig. 6. In case of solid-state electrochemical system constructed by oxygen ion conductor and electrode, terminal voltage difference between the working electrode and reference electrode (E w-r ) is theoretically given when the following two situations are established; (I) steady-state is achieved, and (II) local equilibrium is established [16,28], where E op , i(O 2-), and η are the terminal voltage at open cell circuit condition, oxygen ionic current, and overpotential, respectively. Furthermore, the η is presented by, R, T, and F are the gas constant, absolute temperature, and Faraday constant, respectively. The a o (w) is the oxygen activity under the oxygen ionic current flow at the triple phase boundary among the working electrode, solid electrolyte, and gas phase. The p(O 2 ) is the oxygen partial pressure at the triple phase boundary under ionic current flow condition. The p(O 2 , w) is the oxygen partial pressure at the working electrode side with no current flow condition. In case of this study, p(O 2 , w) is equal to 0.21. From the impedance analysis described above, the R b value was already evaluated, and therefore, we could convert the relationship between the E w-r and i(O 2-) as shown in Fig. 6 into the relationship between the η and i(O 2-) (Fig. 7). Results on the case at LCO/YSZ cell were also plotted for comparison. For the case of LCO/LSO cell, almost no hysteresis was observed by the scan rate at 5 mV/s. On the other hand, clear hysteresis was found in the case of LCO/YSZ cell. Cathode properties at the LCO/LSO interface is found to be better than the properties at the LCO/YSZ interface by comparison of the η-i(O 2-) relationships at the same η and temperature condition. One reason is that no insulator phase is formed at the LCO/LSO interface as discussed in section 3.1.

Electrochemical impedance with applying d.c. bias
In order to discuss the triple phase boundary situation under steady-state oxygen ionic current flow, impedance spectra were corrected under fixed d.c. bias applied condition. Typical impedance spectra are shown in Fig. 8. The shapes of the spectra are similar to the spectra measured under open cell circuit condition in Fig. 3. Hence, the spectra were analyzed by the same equivalent circuit shown in Fig. 4. The R E value measured under d.c. bias applied condition is indexed as R Ex . Similarly, the interfacial conductivity under d.c. bias applied condition is named as extended interfacial conductivity (σ Ex ) which is presented by [16]; Taking into consideration of the electrochemical impedance model based on the interfacial conductivity theory, the σ Ex is a function of p(O 2 ) at steady-state current flow condition [16]. If the electrode resistance is determined by the neutral oxygen diffusion on the LCO surface obeying the law of classical irreversible thermodynamics, σ Ex and σ E values measured by different experiments become the same with each other at the same a o (w) and temperature. From the plots of σ Ex and σ E versus p(O 2 ) (Fig. 5), σ Ex and σ E show almost the same values. These results indicate that the electrode resistance is not the electric resistance but a kind of chemical resistance by neutral oxygen diffusion by the chemical potential gradient yielding to the classical irreversible thermodynamic process [16]. In order to check an adequacy of this result, we confirmed the total conductivity (σ t ) of the LCO bulk as functions of oxygen partial pressure and temperature (Fig. 9). The p(O 2 ) dependence on the σ t shows similar tendency to the previous report [30,31]. Although the oxygen partial pressure dependence of the σ t of LCO is similar to the p(O 2 ) dependence of the σ E and σ Ex , the temperature dependence of the σ t is opposite to the dependences of the σ E and σ Ex . These facts are also supported that the σ E and σ Ex are not governed by the electric properties in the LCO bulk.

Summary
Lanthanum cobaltite (LaCoO 3 ) is found to show low chemical reactivity to the lanthanum silicate oxyapatite solid-electrolyte at 1273 K. From solid-state electrochemical measurements, high cathode performance was confirmed at the LaCoO 3 /lanthanum silicate oxyapatite interface by comparison to the interface at LaCoO 3 /YSZ. Based on the analysis on the interfacial conductivity theory, the interfacial conductivity evaluated by the impedance spectra measured under open cell circuit condition are found to be the same with the extended interfacial conductivity measured by the d.c. bias applying condition. This result indicates that electrochemical kinetics at the LCO/LSO interface is governed by the LCO surface properties, namely surface diffusion of neutral oxygen and as a result, the relationship between the current and overpotential can be explained based on the interfacial conductivity theory.