Introduction Much attention has been paid to oxygen electrochemical reactions in alkaline electrolytes such as alkaline fuel cell, alkaline rechargeable metal-air battery and water-splitting devices. Many researchers have been exploring highly-active and cost-effective oxygen electrode catalysts, and one example of good oxygen electrocatalysts is perovskite-type oxide with the general formula of ABO3. Our group focuses on observing intrinsic electrochemical properties of perovskite when used as an oxygen electrode catalyst, using a carbon-free and binder-free perovskite electrode.1, 2 One of our research findings was that whereas oxygen evolution reaction activity of model electrodes were as high as that of composite electrodes consisting of perovskite powder, carbon and binder, oxygen reduction reaction activity of model electrodes was much lower than that of composite electrodes. This contrasting behavior motivated us to conduct further experiments such as hydrogen peroxide reduction reaction activity and cyclic characteristics of oxygen electrochemical reactions by using the model electrode and the composite electrode in terms of surface oxygen non-stoichiometry, which is our present research topic. Experimental section Synthesis and fabrication of La0.8Sr0.2CoO3 (LSCO) crystal electrodes were described in our previous literature.2 Perovskite powders were also synthesized by hydroxide-acid aided process3 and solid state process. The resultant oxide crystals and powders were characterized by using X-ray diffraction (XRD), energy-dispersive X-ray (EDX) spectroscopy, Brunauer-Emmett-Teller for oxide powders, and iodometry for oxide powders. Electrochemical measurements were conducted by using a rotating disk electrode (RDE) in a three-electrode cell. For a working electrode, RDE of LSCO crystal electrodes were assembled by mounting a LSCO crystal in a Teflon disk holder. As for composite electrodes, a catalyst layer consisting of 250 μg cm-2 perovskite powder, 50 μg cm-2 Vulcan XC-72 (Cabot), and 50 μg cm-2 anion conductive ionomer (AS-4, Tokuyama) was formed on glassy carbon (GC) RDE. 50 μg cm-2 Pt(40wt%)/Vulcan (E-TEK) and 50 μg cm-2 anion conductive ionomer on GC RDE was also used as benchmarking of oxygen electrochemical reactions. Pt wire and reversible hydrogen electrode (RHE) were used as counter and reference electrodes, respectively. Oxygen electrochemical activity measurements were conducted using a solution of oxygen-saturated 1.0 mol dm-3 KOH for oxygen reduction reaction (ORR) and oxygen evolution reaction (OER), and argon-purged 1 mol dm-3 KOH containing ca. 5 mmol dm-3 H2O2 for hydrogen peroxide reduction reaction (PRR). For qualitative electronic conductivity measurement, ferricyanide reduction reaction (FRR) was observed in a solution of argon-purged 1.0 mol dm-3 KOH containing 20 mmol dm-3 K3[Fe(CN)6]. Results and discussions XRD and EDX measurements confirmed that resultant LSCO crystals and perovskite powders had perovskite structure and the atomic compositions were quite similar as the desired composition. Figure 1 shows PRR activity on LSCO crystal. In the initial cycles, it shows clear rotation dependence indicating that LSCO crystal was active toward PRR. However, as the cycle number of cyclic voltammetry increased, reduction currents were decreased and finally showed no dependence on rotation. Similar results were observed for solid-state-synthesized LSCO composite electrode suggesting that obtained results by model electrodes were universal for practical electrodes. The decrease in activities was also observed for ORR, while the decrease in OER activity on LSCO crystal was not observed. FRR after PRR and ORR showed that electronic conductivity of LSCO crystal was decreased. The reason of activity decrease in ORR and PRR, and the influence of B-site cation in ORR and PRR in terms of surface oxygen non-stoichiometry will be discussed in the conference.