an0c03180_si_001.pdf (1.03 MB)
Reversible Transition of Volatile and Nonvolatile Switching in Ag–In–Zn–S Quantum Dot-Based Memristors with Low Power Consumption for Synaptic Applications
journal contribution
posted on 2021-02-09, 19:42 authored by Nan He, Langyi Tao, Qiangqiang Zhang, Xiaoyan Liu, Xiaojuan Lian, Xiang Wan, Er-Tao Hu, Lin He, Yang Sheng, Feng Xu, Yi TongNeuromorphic
computing systems composed of electronic synapses
and neurons provide an effective choice to construct the next generation
of intelligent computing systems. However, poor resistive switching
uniformity and high power consumption of the reported electronic synapses
are the major obstacles for developing large-scale brain-inspired
computing systems. Herein, quaternary Ag–In–Zn–S
(AIZS) quantum dots (QDs) have been utilized to fabricate a kind of
electronic synaptic devices with crossbar-array structure. The devices
exhibit excellent electrical performance, including uniformly distributed
operation voltages, reliable data retention, robust cycle measurement,
and large memory window. More interestingly, the devices can achieve
reversible transitions between volatile switching and nonvolatile
memory switching in a single QD-based cell by modulating the compliance
current. The power consumption per switching can be achieved as low
as ∼10 pW. Additionally, the Schottky emission and ohmic conduction
can be used to account for the switching mechanisms well. Furthermore,
biological synaptic-like behaviors have been successfully mimicked
by QD-based devices, including short-term potentiation, paired-pulse
facilitation, long-term potentiation/depression, and the preliminary
transition from short-term memory to long-term memory. These results
reveal that the QD-based device with excellent performance might be
a potential candidate for energy-efficient brain-inspired computing
applications.