Simulation of Optoelectronic Synaptic Behavior in SnSe2/WSe2 Heterojunction Transistors

Abstract

The extensive research on neuromorphic devices exemplified by transistors is attributed to their capacity to emulate the synaptic plasticity found in the human brain. The unique properties of two-dimensional (2D) materials, such as layered structure, excellent optoelectronic properties, and ability to form heterojunctions, make them promising candidates for synaptic devices. Herein, a transistor based on a SnSe₂/WSe₂ heterojunction structure is investigated, and it can be used to simulate the function of optoelectronic synapses. The heterojunction exhibits bidirectional responses to optical stimuli with different wavelengths (λ = 400 nm/500 nm), enabling the device to emulate both excitatory and inhibitory synaptic behaviors in an all-optical pathway. Besides optical stimulus, gate voltage is used to modulate the synaptic performance of the device under 500 nm illumination, enabling it to mimic light adaption of biological eyes successfully. Based on the wavelength-selective synaptic plasticity, an optically driven artificial neural network (ANN) is proposed to classify handwritten digits with an accuracy of around 80%. This work will be an important step toward the future development of multifunctional optoelectronic synapses.