Ferroelectric modulation of ReS₂-based multifunctional optoelectronic neuromorphic devices for wavelength-selective artificial visual system

Abstract

Neuromorphic optoelectronic vision system inspired by the biological platform displays potential for in-sensor computing. However, it is still challenge to process multiwavelength image in noisy environment with simple device configuration and light-tunable biological plasticity. Here, a prototype visual sensor is demonstrated based on ferroelectric copolymer poly(vinylidene fluoride-trifluoroethylene) (P(VDF-TrFE)) and 2D rhenium disulfide (ReS2) with integration of recognition, memorization, and pre-processing functions in the same device. Such synaptic devices achieve impressive electronic characteristics, including a current on/off ratio of 109 and mobility of 45 cm2V−1s−1. Various synaptic plasticity behaviors have been achieved owing to the switchable ferroelectricity, enabling them to establish an artificial neural network (ANN) with high digit recognition accuracy of 89%. Through constructing optoelectronic device array, object extraction is achieved with wavelength-selective capability in noisy environment, closely resembling human retina for color recognition. Above outcomes bring a notable improvement in the image recognition rate from 72% to 96%. Besides, low energy consumption comparable to single biological event can be realized. With these multifunctional features, this work inspires highly integrated neuromorphic systems and the development of wavelength-selective artificial visual platform.