A 2D MoS2-Based Active-Matrix Photosensor Array for Neuromorphic Vision

Abstract

Neuromorphic visual processing hardware mimics key functions of the human visual system, including static image sensing and processing, as well as the perception and prediction of moving targets. However, conventional complementary metal-oxide semiconductor (CMOS) platforms face significant challenges such as power consumption and processing delays. Herein, a 64 × 64 active-matrix photosensor array (AM-PA) is developed with a one-transistor-one-photodetector architecture, by integrating metal-oxide thin-film transistors with monolayer MoS₂ photodetectors. By modulating the intrinsic defects serving as charge-trapping centers within the MoS₂ film, the AM-PA exhibits configurable conductance plasticity in response to optical stimuli. This system enables in-array visual sensing and processing functionalities: static image sensing and pre-processing to reduce noise, prediction of dynamic spatiotemporal patterns, and accurate trajectory forecasting for complex motion scenarios. These results demonstrate the feasibility of the AM-PA as a promising platform for neuromorphic visual systems.