[22a-E102-1~8] FS.1 Focused Session "AI Electronics"

[22a-E102-7] Cu diffusion in amorphous-Ta₂O₅ containing H₂O studied with high-dimensional neural network potential

〇(M2)Junichi Okamoto¹, Shimizu Kouji¹, Watanabe Satoshi¹ (1.Univ. Tokyo)

Keywords:synaptic devices, atomic switch, neural network potential

To understand the microscopic processes of Cu diffusion in amorphous Ta₂O₅ (a-Ta₂O₅) toward the development of next-generation memory devices and synaptic devices, in this study we have investigated the effects of density of a-Ta₂O₅ and the presence of H₂O in it on Cu diffusion using a high-dimensional neural network potential (NNP). We have succeeded in constructing a reliable NNP that has the prediction accuracy sufficient for the analysis, and the results of molecular dynamics simulations using the constructed NNP show that the Cu diffusion speed changes depending on the a-Ta₂O₅ density and the presence of H₂O.