Grain boundary effect unveiled in monolayer MoS2 for photonic neuromorphic applications

Abstract

Layered 2D materials with grain boundaries (GBs) hold promise for various applications. Despite considerable progress, deliberately introducing GBs remains a challenge during 2D material growth. In this study, we explored the significance of GBs in monolayer molybdenum disulfide (ML MoS₂) in optoelectronic functionalities such as photodetection, photonic synaptic behaviours, and adaptive learning analogous to the brain. Our comparative analysis revealed that GB-free MoS₂ devices demonstrated a superior photodetector performance, featuring a remarkably high dark-to-light current ratio of 10⁵ at −1.0 V, with responsivity reaching 10 A W⁻¹, detectivity soaring to 10¹⁵ W⁻¹, and a noise equivalent power of about 10⁻¹⁶ W √Hz⁻¹, outperforming devices containing GBs. Moreover, GBs significantly delayed the response speed, indicating the prevalence of persistent photoconductance. Conversely, GBs facilitated the replication of various optoelectronic synaptic behaviors and adaptive learning of the brain. Our experimental results uncovered the sharp and robust atomic stitching of MoS₂ domains at GBs, which act as charge carrier recombination centers due to the presence of defects. Notably, GBs induced a barrier within the energy bands by shrinking the bandgap, hindering the flow of photogenerated carriers. We systematically elucidated this phenomenon through energy band diagrams and electrostatic force microscopy results from biased samples under illumination.