2022年第69回応用物理学会春季学術講演会

講演情報

一般セッション(口頭講演)

FS フォーカストセッション「AIエレクトロニクス」 » FS.1 フォーカストセッション「AIエレクトロニクス」

[22a-E102-1~8] FS.1 フォーカストセッション「AIエレクトロニクス」

2022年3月22日(火) 09:00 〜 11:15 E102 (E102)

木村 睦(龍谷大)

09:15 〜 09:30

[22a-E102-2] High Symmetry and Linearity in Weight Update with GeTe/Sb2Te3 Superlattice-like Structure for Artificial Synapse

〇(D)Shinyoung Kang1、Soo-Min Jin2、Tae-Hun Shim2、Jea-Gun Park2、Yun-Heub Song2、Yuji Sutou1,3 (1.Tohoku Univ. (Eng)、2.Hanyang Univ.、3.Tohoku Univ. (AIMR))

キーワード:Artificial synapse, Interfacial phase change memory, Neuromorphic

Artificial intelligence (AI), which allows us to act and think like humans, has been extensively studied not only in academia but also in society. Therefore, ‘neuromorphic computing’ was introduced that a new computational method that mimics the biological brain. Recently, there are device structural and material efforts for PCM for artificial synaptic applications. Due to the mechanism of PCM, problems such as high current and asymmetrical resistance change to achieve high resistance still exist. To solve this problem, an interfacial phase change memory (iPCM), in which a superlattice-like structure is created by alternately depositing a GeTe thin film and a Sb2Te3 thin film, has emerged. We implemented a linear and symmetrical conductance transition based on the unique properties of GeTe/Sb2Te3 iPCM deposited by sputtering. This will be of great help in utilizing artificial synapses that constitute neuromorphic computing requiring analog transition.