Visible-light driven photoelectric synaptic transistors based on ZnO/SnO2 heterostructure for neuromorphic computing

Abstract

Photoelectric synaptic transistors (PSTs) exhibit tremendous potential in artificial intelligence systems, as they exhibit the ability to combine the neuromorphic computing and the visual perception neural functions. For visual information processing and computing, the utilization of the visible-light signals is crucial for the development of artificial neuromorphic systems. In this report, the ZnO/SnO2 heterostructure was fabricated by spin coating and the PSTs based on the heterostructure were integrated. The visible-light driven synaptic functions of the PSTs were demonstrated. The oxygen vacancy content and charge separation efficiency under light excitation are improved through the designed ZnO/SnO₂ heterostructure, consequently enhancing the photoresponsivity. Under the irradiation with 532nm green light, various synaptic behaviors of the PSTs, including short-term memory, long-term memory, and paired-pulse facilitation, have been simulated by the PSTs based on ZnO/SnO2 heterostructure. Moreover, the exceptional performance of the PSTs in handwritten digit image recognition, with an impressive recognition rate of up to 92.88%, is demonstrated by the PSTs based on ZnO/SnO2 heterostructures. This result highlights its enormous applications potentials in next generation computer systems.