A three-terminal MoTe2 artificial synapse with electrical and optical dual-mode plasticity.

Abstract

Artificial synaptic devices capable of integrating sensing, memory, and computing functions are highly desirable for next-generation neuromorphic systems. However, most reported artificial synapses rely on single-mode electrical or optical stimulation, limiting their ability to achieve multimodal perception and adaptive information processing. Here, we demonstrate a three-terminal artificial synaptic device based on MoTe2 flakes that enables both electrical and optical synaptic modulation within a unified device architecture. The device exhibits bidirectional electrical synaptic plasticity through gate-controlled charge trapping processes, successfully emulating excitatory/inhibitory responses, short-term memory to long-term memory transition, paired-pulse facilitation, and learning-forgetting-relearning behaviors. Besides, optical stimulation also induces persistent synaptic responses through photocarrier generation and trapping, allowing the device to reproduce adaptive visual memory under different illumination conditions. By regulating optical intensity, pulse duration, and stimulation frequency, controllable visual perception and memory consolidation are achieved. Based on experimentally extracted synaptic characteristics, a convolutional neural network achieves approximately 94.67% accuracy in CIFAR-10 image recognition, demonstrating the ability of the device to perform simulation-based neuromorphic computing. This work provides a multifunctional two-dimensional material-based artificial synapse integrating multimodal sensing, memory storage, and in-memory computing, offering a promising platform for low-power neuromorphic vision systems and intelligent edge computing applications.