Presentation Information

[7p-S202-4]Experimental demonstration of a data-driven automated floating zone crystal growth furnace using reinforcement learning

〇Sogo Yanagisawa1, Ryohei Matsumoto2, Syogo Sumitani2, Takuya Inagaki3, Eisuke Bannai4, Kentaro Kutsukake1,5, Toru Ujihara1,5, Shunta Harada1,5 (1.Nagoya Univ., 2.Anamorphosis Networks, 3.Sanko, 4.NIMS, 5.IMaSS Nagoya Univ.)

Keywords:

automation,crystal growth,floating zone method

Recent years informatics has applied to materials processing. We already demonstrated automated operation of floating zone (FZ) crystal growth using reinforcement learning within a simulated environment. In this study, we constructed a real, operational FZ furnace prototype to experimentally validate this method, and demonstrated the consistent automated operation. This result implies this method is applicable to other human-dependent material processing techniques.