[18a-S2_204-1~6] 23.1 Joint Session N "Informatics"

12:30 PM - 12:45 PM JST(3:30 AM - 3:45 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[18a-S2_204-5] Real-Time Temperature Measurement of SiC Wafers During Thermal-Plasma-Jet Annealing Using Machine-Learning-Assisted Optical-Interference Contactless Thermometry

〇Jiawen Yu¹, Hiroaki Hanafusa¹, Seiichiro Higashi¹ (1.Hiroshima Univ.)

Keywords:

machine learning,temperature measurement,plasma processing

In this presentation, we report a machine-learning-assisted optical-interference contactless thermometry (OICT) method for real-time temperature analysis during millisecond thermal processing using a thermal plasma jet. By extracting characteristic features from optical thickness variations, the proposed approach sequentially reconstructs transient temperature–depth distributions inside the wafer. Compared with conventional manual fitting and database search methods, the machine learning approach maintains comparable temperature accuracy and analysis speed while significantly improving generality across a wide range of heating conditions. Experimental validation demonstrates reliable real-time temperature measurement under various plasma heating parameters, highlighting the potential of this method for advanced plasma process monitoring and control in next-generation semiconductor manufacturing.