[17p-P04-1~10] 23 Joint Session N "Informatics" (Poster)

[17p-P04-1] Stability predictions of stable anode materials for battery systems by graph neural network potentials

〇Kohei Tada^1,2, Hiroyuki Ozaki², Tetsu Kiyobayashi² (1.Osaka Univ., 2.AIST)

Keywords:

neural network potential,LIB,SIB

First-principles calculations based on density functional theory (DFT calculations) are widely used in the study for battery materials. However, due to the computational cost of DFT, it is difficult to comprehensively investigate systems that require a large number of models, such as investigation on the stability with cation mixing. In this study, we investigate whether the neural network potentials (M3GNET, and CHGNET) can estimate the influence of Li cation mixing on the structural stability of stable anode materials for batteries.