MoS2-Nanosheet Based Optoelectronic Synaptic Transistor with Integrated Computing for Environment-Adaptive Artificial Retina

Abstract

Recent advances in information processing have led to the development of innovative devices that mimic multiple biological sensory systems. The visual system plays a crucial role in information acquisition, with approximately 80% of environmental information being processed through human vision. The demand for energy-efficient and multifunctional devices in complex applications has fueled the exploration of optoelectronic synapses. Here, we demonstrate a MoS₂-nanosheet based optoelectronic synaptic transistor with Au nanocrystal memory cells, which integrates sensing, memory, and processing functions into a single device. This transistor exhibits significant performance in both electrical and optical operation modes, including the endurance of 1000 electrical storage/erasure cycles, remarkable electrical retention time for 10 years, over 2000 s optical storage time, and light current-to-dark current ratio exceeding 10⁴. The device features both homosynaptic and heterosynaptic plasticity, enabling the simulation of short-term memory, long-term memory, and synaptic weight regulation through electrical and optical pulse modulation. Additionally, we have achieved ambient light adaptation and target object recognition. This study presents a promising strategy for advancing artificial intelligence visual systems.