[15a-S4_202-1~8] 9.3 Nanoelectronics

9:45 AM - 10:00 AM JST(12:45 AM - 1:00 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[15a-S4_202-3] Machine Learning Technique for Non-destructive Characterization of Embeded Nano-structures in Si MOSFET

〇(D)Renxiang Lyu¹, Hyoto Yamaguchi¹, Zenji Yatabe¹, Seiya Kasai¹ (1.RCIQE, Hokkaido Univ.)

Keywords:

Semiconductor,Machine learning,Nano-structures

For practical application of the nano-artifact metrics using random nanostructures, we proposed an electrical structure discrimination technique using Si MOSFET and we experimentally examined the concept. However, there was a large difference in the electrical characteristics between the device simulation and the fabricated devices. To understand this result, we have investigated the nondestructive characterization of the embedded nanostructure in the device using the machine learning. In this paper, we consider how the differences among the neural network structures and the electrical characteristics affect the structure inference accuracy.