Presentation Information
[18a-PA1-13]Deep Learning for Predicting and Analyzing the Electrical Properties of Polycrystalline Ge Thin Films
〇(M1)Takenori Nakajima1, Takamitsu Ishiyama1, Kaoru Toko1 (1.Univ. of Tsukuba)
Keywords:
materials informatics,polycrystalline thin film,deep learning
Recently, materials informatics integrating materials science with machine learning—has been developing rapidly. In this study, we focus on polycrystalline Ge thin films, whose electrical properties are strongly influenced by grain characteristics. By training a deep learning model on EBSD images of locally patterned microfabricated regions together with Hall measurement results, we aim to elucidate how grain-boundary structures and crystallographic orientations affect the electrical properties.