[25p-31A-1~9] FS.1 Focused Session "AI Electronics"

2:15 PM - 2:30 PM JST(5:15 AM - 5:30 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation AwardEnglish Presentation

[25p-31A-5] Improvement in Computing Capacities of HfZrO₂ FeFET-based Reservoir Computing through Operating Voltage Optimization

〇(D)SHINYI MIN¹, Toprasertpong Kasidit¹, Eishin Nako¹, Ryosho Nakane¹, Mitsuru Takenaka¹, Shinichi Takagi¹ (1.The Univ. of Tokyo)

Keywords:

Reservoir Computing,FeFET

The reservoir in a reservoir computing (RC) system nonlinearly transforms time-series inputs into high-dimensional data which allows efficient training for real-time information processing. Hf_0.5Zr_0.5O₂ (HZO)-based ferroelectric field-effect transistor (FeFET) is promising for RC systems through ferroelectric polarization, domain dynamics, and non-linear interaction with channel carriers. We have reported the fundamental characteristics of the HZO FeFET-based RC system. However, the impact of the bias conditions needs to be examined, since the FeFET is fundamentally a multi-terminal device and operating voltage optimization can enhance the RC performance. In this study, we investigate how the parameters of input V_g(t) and V_d impact the performance of the HZO FeFET-based RC system and have optimized to maximize the computing capacities.