High-performance solar-blind optoelectronic synaptic device based on N-doped Ga2O3 nanowire network for neuromorphic computing

Abstract

Optoelectronic synapses use optical signals as modulation inputs and offer advantages such as low-power consumption, non-contact operation, and high parallelism. However, devices operating in the visible and near-infrared regions are susceptible to ambient light interference, leading to signal distortion and high background noise. In contrast, optoelectronic synaptic devices working in a solar-blind region are impervious to external ambient interference, enabling them to achieve stable and high-fidelity signal transmission effectively. Ga2O3 as a wide bandgap with inherent oxygen-vacancy defects has strong deep-ultraviolet light sensitivity, making it an ideal material for solar-blind optoelectronic synapses. Here, high-performance solar-blind synaptic devices were prepared using nanowire network interfacial effects and nitrogen doping to modulate oxygen vacancies. Notably, the paired-pulse facilitation index increased by approximately 1.2 fold after doping. In handwritten digit recognition experiments, the recognition accuracy of the doped device exceeded 90%, representing an improvement of nearly 15% compared to the undoped counterpart. Furthermore, leveraging its enhanced light perception and strong memory characteristics, a precise optical encryption communication scheme with strong anti-interference capability was also designed. The optoelectronic synaptic devices based on element regulated Ga2O3 nanowire network provide a new way for the development of the next generation of high-performance solar-blind optoelectronic neuromorphic systems.