[17p-A33-1~12] FS.1 Focused Session "AI Electronics"

2:30 PM - 2:45 PM JST(5:30 AM - 5:45 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation AwardEnglish Presentation

[17p-A33-4] Experimental Demonstration of Reservoir Computing Using Anti-ferroelectric HZO Capacitors

〇(D)SHINYI MIN¹, Kasidit Toprasertpong¹, Eishin Nako¹, Ryosho Nakane¹, Mitsuru Takenaka¹, Shinichi Takagi¹ (1.The Univ. of Tokyo)

Keywords:

Reservoir Computing,Anti-ferroelectric,HZO

The reservoir computing (RC) system is promising for sequential information processing with extremely-low computation cost by only training the output layer weights. A reservoir in the RC system requires temporal memory properties and nonlinearity to transform time-series inputs into high-dimensional data for efficient learning. We have reported a ferroelectric (FE) Hf_1-xZr_xO₂ (HZO)-based RC system utilizing the temporal-spatial polarization dynamics of FE-HZO. However, the single polarization switching of FE-HZO is insufficient to diversify the output response. Unlike a monotonous switching for FE-HZO, anti-ferroelectric (AFE)-HZO exhibits double polarization switching, which allows complicated response signals. In this study, we propose an effective RC system using an AFE-HZO-based metal/ferroelectric/metal (MFM) capacitor.