[10p-S202-1~14] 15.6 Group IV Compound Semiconductors (SiC)

[10p-S202-1] Optimization of furnace temperature distribution in SiC sublimation process using machine learning

〇(M2)Yoshiki Inoue¹, Tomoaki Furusho², Kentaro Kutsukake^2,3, Shunta Harada^1,2, Miho Tagawa^1,2, Toru Ujihara^1,2,4 (1.Nagoya Univ., 2.IMaSS Nagoya Univ., 3.AIP RIKEN, 4.GAN-OIL AIST)

Keywords:SiC, simulation, machine learning

SiC is expected to be an energy-saving semiconductor material for power devices, but its high cost is an issue, and it is necessary to fabricate low-defect-density crystals at high growth rates. We have reported that machine learning can be used to speed up the thermo-fluid analysis of the SiC sublimation method. In this study, in addition to speeding up the thermo-fluid analysis using machine learning, we optimized the temperature distribution to suppress polycrystalline precipitation at the periphery of the single crystal in order to obtain a low defect density crystal.