[15a-A205-1~10] 23.1 Joint Session N "Informatics"

▲ [15a-A205-6] Design of thermodynamically stable perovskites using machine learning

Zhilong Song¹, Xiwen Chen¹, 〇Sae Dieb², Masashi Ishii² (1.Soochow university, 2.MaDIS, NIMS)

Keywords:Materials informatics

Perovskite materials have attracted much attention in the past few years and have been widely used in solar cells, light emitting diodes, lasers, photocatalysis. In this work, we use Monte Carlo tree search (MCTs) to design a thermodynamically stable perovskites materials. Two versions of MCTS were adopted; basic and with policy gradient. MCTS efficiently found the most thermodynamically stable double perovskite structure in the pre-calculated data using only elements type information.