1. Introduction Currently, research on layered and rocksalt Li-rich materials is being actively conducted as a high-capacity positive electrode material for lithium ion batteries, and now it is widely accepted that large capacity originates from oxygen redox reaction.[1] Oxygen redox is proposed to be activated for oxygen ions with a Li-O-Li configuration.[2] It was also reported that oxygen redox is activated for Li-rich manganese spinel-type oxides, i.e., Li[Li x Mn2–x ]O4.[3] Nevertheless, its origin of activation of anionic redox for spinel-type oxides is not known. In this study, LiMg y Ni0.5-y Mn1.5O4, which is solid solution samples between LiNi0.5Mn1.5O4 and LiMg0.5Mn1.5O4, is targeted as positive electrode materials with anionic redox for spinel-type oxides. Mg ions have a high ionic bonding nature with oxygen, and thus better reversibility for anionic redox is anticipated.[1] We report the impact of Mg substitution on reversibility of Ni cationic and O anionic redox in the spinel framework structure, and the possibility of 5 V-class high-voltage electrode materials with excellent capacity retention is discussed. 2. Experimental LiMg y Ni0.5-y Mn1.5O4 (0 ≤ y ≤ 0.5) were synthesized from mixtures of Li2CO3, Mg(OH)2, Ni(OH)2 and Mn2O3 by two-step solid-state reaction at 950 oC for 6 h to 700 °C for 48 h in air atmosphere. Characterization of LiMg y Ni0.5-y Mn1.5O4 were conducted by X-ray diffraction (XRD) and scanning electron microscopy (SEM). For the evaluation of electrochemical properties, composite electrodes consisting of 80 wt% active material, 10 wt% AB, and 10 wt% PVdF, and were casted on aluminum foil. 1.0 M LiPF6 dissolved in ethylene carbonate (EC)/dimethyl carbonate (DMC) (3/7 by volume) solution was used as electrolyte. 3. Results and Discussion XRD patterns of LiMg y Ni0.5-y Mn1.5O4 (0 ≤ y ≤ 0.5) are assigned into the spinel structure with space group of Fd-3m. From SEM images, particle sizes of LiMg y Ni0.5-y Mn1.5O4 are estimated to be ~4 μm. No difference in particle sizes regardless of chemical compositions is observed. Figure 1 compares charge/discharge curves of LiMg y Ni0.5-y Mn1.5O4. After 5 cycles, the non Mg-substituted sample, LiNi0.5Mn1.5O4, delivers 135 mAh g−1 of the discharge capacity in a lithium cell when the cut-off voltage is set to 5.3 V. After 50 cycles test, discharge capacity of LiNi0.5Mn1.5O4 is decreased to 125 mAh g-1. On the other hand, significant improvement of cyclability is evidenced for the partially Mg-substituted samples. No capacity fading and no increase in polarization is observed for LiMg0.06Ni0.44Mn1.5O4 even charge to 5.3 V vs. Li for 50 cycles at a rate of 10 mA g-1. On the basis of these findings, Mg substitution is proposed to be effective to reduce electrolyte decomposition and improve structural stability upon high-voltage exposure. Nevertheless, the further enrichment of Mg ions in the structure results in loss of reversible capacity, and anionic redox seems to be inactive for LiMg0.5Mn1.5O4. From these results, factors affecting electrode performance of high voltage spinel with cationic/anionic redox will be discussed in detail.References(1) N. Yabuuchi, Chem. Rec., 19, 703 (2019).(2) D.-H. Seo, Nature Chem., 8, 692 (2016).(3) E. Iwata et al., and T. Ohzuku, Electrochemistry, 71,1187 (2003). Figure 1