Presentation Information

[25a-12F-1]Development of Autonomous Exploration System for Thin-film Materials

〇Yota Suzuki1,2, Kazunori Nishio1, Akira Aiba1, Kei Takihara1, Shigeru Kobayashi3, Ryo Nakayama3, Ryota Shimizu3, Yasuyuki Nagamine2, Taro Hitosugi1,3 (1.Tokyo Tech, 2.Tokyo Tech OFC, 3.UTokyo)

Keywords:

Autonomous Exploration System,Machine Learning,Robotics

In the considerable material exploration space, it is essential for material development to find new functional materials efficiently. In order to efficiently explore such new materials, we have developed an autonomous experiment system utilizing machine learning and robotics, which automatically performs measurement and analysis of thin film materials fabricated by RF magnetron sputtering, including XRD, Raman, UV-Vis, and SEM/EDS. The system then autonomously searches the material search space until the desired physical properties are obtained, repeating synthesis, evaluation, and prediction of the subsequent synthesis conditions by Bayesian optimization.