High Performance DUV Synaptic Oxide Thin Film Transistor with Rapid Response

Delang Lin,Dongxiang Luo,Fion Sze Yan Yeung,Hoi Sing Kwok,Rongsheng Chen
DOI: https://doi.org/10.1109/led.2024.3441608
IF: 4.8157
2024-01-01
IEEE Electron Device Letters
Abstract:Neuromorphic systems that enable parallel processing and self-learning are promising solutions to overcome the limitations of Von Neumann architectures. Synaptic devices are fundamental in neuromorphic computing and are being developed for optical stimulation to improve performance. In this paper, high-performance deep-ultraviolet (DUV) synaptic transistors based on the ITMO-ITZO-ITMO channel are proposed. The response speed and current of the fabricated devices were greatly improved by the stacked active layer, achieving postsynaptic currents exceeding 40 nA under weak DUV pulse stimulation (20 ms pulse width and 20 μW/cm2 light intensity). Moreover, these devices showed an ultra-high paired-pulse facilitation index of 410% with a 20 ms interval. Additionally, the DUV synaptic transistors exhibited various synaptic plasticity, successfully simulated the learning behavior of the human brain, and enabled 4-bit time-series signal recognition. This marks a significant advancement toward efficient neuromorphic systems with enhanced decision-making capabilities.
What problem does this paper attempt to address?