An artificial synapse based on Sr(Ti, Co)O3 films. (December 2022)

• Record Type: Journal Article
• Authors: Shen, Zhi-Hao; Li, Wen-Hua; Tang, Xin-Gui; Hu, Jia; Wang, Kai-Yuan; Jiang, Yan-Ping; Guo, Xiao-Bin
• Is Part Of: Materials today communications. Volume 33(2022)
• Publication Date: 2022-12
• Subjects: Calcium titanite -- Memristor -- Synapse -- Neuromorphic computing system; Materials science -- Periodicals; 620.11
• DOI: 10.1016/j.mtcomm.2022.104754
• Languages: English
• ISSNs: 2352-4928

Abstract

Abstract: With the explosive growth of the demand for computing power and storage on chips, researchers have proposed an artificial synapse with biological synapse-like functions to achieve low-energy, high-efficiency parallel neuromorphic computation and perform various complex functions efficiently, in response to the shortcomings of "storage and computation separation" in traditional computing systems. In this paper, we fabricated an artificial synapse based on SrTi0.99 Co0.01 O3 calcium titanite. It implements a variety of important synaptic learning and memory functions, including long-term and short-term plasticity, paired-pulse depression, spike-time-dependent plasticity. It has promising applications in the construction of artificial neural networks for processing massive data. Further studies show that the resistance switching ratio of SrTi0.99 Co0.01 O3 is improved by an order of magnitude compared to SrTiO3, and the stability is greatly enhanced, which may be attributed to the increase in the content of oxygen vacancies in the material (from 20.3 % to 23.9 %), which is also more favorable to the synaptic performance. Therefore, Co has an important impact on the enhancement of SrTiO3 properties. This study is a good reference for future full implementation of brain-inspired computing systems. Graphical Abstract: ga1 Highlights: The device implements important synaptic learning and memory functions, such as spike-time-dependent plasticity. we compared Sr(Ti, Co)O3 with SrTiO3 and found that Co doping gives the device a larger switching ratio and better stability. The introduction of Co increases oxygen vacancies from 20.3 % to 23.9 %, promoting device performance.