The 83rd JSAP Autumn Meeting 2022

Presentation information

Symposium (Oral)

Symposium » In-materio AI computing

[20p-M206-1~11] In-materio AI computing

Tue. Sep 20, 2022 1:30 PM - 6:15 PM M206 (Multimedia Research Hall)

Hirofumi Tanaka(Kyushu Inst. of Tech.), Takuya Matsumoto(Osaka Univ.)

2:30 PM - 2:45 PM

[20p-M206-4] Impacts of Chemical Phase Segregation on the Stochastic Properties of Resistive Switching in Amorphous TaOx

Atsushi Tsurumaki-Fukuchi1, Takayoshi Katase2, Hiromichi Ohta3, Masashi Arita1, Yasuo Takahashi1 (1.Faculty of IST, Hokkaido Univ., 2.MSL, Tokyo Tech., 3.RIES, Hokkaido Univ.)

Keywords:memristor, analog resistive switching, amorphous metal oxide

Toward understanding the physical mechanisms of analog and stochastic resistive switching in amorphous TaOx, which are key phenomena in the development of current neuromorphic hardware, direct analyses of the causal ion migration were conducted in its atomically flat thin films by conductive atomic force microscopy (C-AFM). In the analyses, both analog- and stochastic-type resistive switching was directly measured by C-AFM and the ion migration account for the occurrence of resistive switching was visualized in ångström scale. The measurements indicated that the mechanism of the analog type of resistive switching, where the resistance was gradually changed by voltage applications, is based on the reversible redox reactions of amorphous TaOx for 2.0 < x < 2.5. Meanwhile, in the stochastic-type resistive switching, where the resistance changes were suddenly caused with temporal stochasticity, segregation of a metastable phase of amorphous TaO2 and its important impacts on the switching stochasticity was suggested. Based on these results, direct clues for controlling the switching stochasticity of amorphous TaOx were provided in this study.