Presentation Information

[9p-N322-15]Study on the Performance Improvement of Optical-Interference Contactless Thermometry Based on Machine Learning

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

Keywords:

Temperature Measurement,Machine Learning,AI

High-accuracy real-time temperature measurement is essential for ensuring uniformity and reproducibility in semiconductor processes. We have developed a unique Optical-Interference Contactless Thermometry (OICT) technique, which enables precise temperature measurement even under challenging conditions such as plasma processing and millisecond annealing, by analyzing the changes in optical film thickness induced by wafer temperature variations. In previous work [1], we achieved real-time temperature measurement by combining OICT with image processing and database searching. However, this approach lacks generalizability and requires substantial computational and storage resources to improve accuracy. In this study, we focused on the high adaptability and generalization capability of artificial intelligence and machine learning (ML), and explored a new method to realize real-time temperature measurement by integrating ML into the OICT framework.