Reconfigurable, non-volatile neuromorphic photovoltaics

Abstract

Recon gurable image sensors for the recognition and understanding of real-world objects are now becoming an essential part of machine vision technology. The neural network image sensor -which mimics neurobiological functions of the human retina -has recently been demonstrated to simultaneously sense and process optical images. However, highly tunable responsivity concurrently with non-volatile storage of image data in the neural network would allow a transformative leap in compactness and function of these arti cial neural networks (ANNs) that truly function like a human retina. Here, we demonstrate a recon gurable and non-volatile neuromorphic device based on twodimensional (2D) semiconducting metal sul des (MoS 2 and WS 2 ) that is concurrently a photovoltaic detector. The device is based on a metal/semiconductor/metal (M/S/M) two-terminal structure with pulse-tunable sulfur vacancies at the M/S junctions. By modulating sulfur vacancy concentrations, the polarities of short-circuit photocurrent -can be changed with multiple stable magnitudes. Device characterizations and modeling reveal that the bias-induced motion of sulfur vacancies leads to highly recon gurable responsivities by dynamically modulating the Schottky barriers. A convolutional neuromorphic network (CNN) is nally designed for image process and object detection using the same device. The results demonstrated the two-terminal recon gurable and non-volatile photodetectors can be used for future optoelectronics devices based on coupled Ionic-optical-electronic effects for Neuromorphic computing.

Self-Powered Programmable van der Waals Photodetectors with Nonvolatile Semifloating Gate

“…Additionally, the pulsed gating demonstrates exceptional retention with the photocurrent experiencing only a modest reduction to 67.07% after 10 years of operation. With these remarkable photodetection features, combined with its gate-tunability, the SFG-PD holds great promise for diverse applications in programmable optoelectronics , and neuromorphic devices. − …”

“…Recently, there has been growing interest in tunable photovoltaic photodetectors (PDs) utilizing atomically thin two-dimensional (2D) materials. This surge in attention is particularly evident with the advancement of programmable , and neuromorphic − optoelectronics. The photovoltaic response is characterized by attributes such as low-power consumption, rapid response, and low dark current. − This behavior is commonly attributed to the built-in electric field within various junctions, including p + -p, p-n, n + -n, and Schottky junctions, which are typically fabricated through processes such as chemical doping, , materials stacking, , epitaxial growth, , local split gating, − strain engineering, , or asymmetric contacting. − Distinguished by the volatile or nonvolatile nature of these junctions, two strategies involving continuous and pulsed voltage application are typically employed to tune the photovoltaic photoresponse, enabling transitions from negative to positive responses.…”

“…However, it is worth noting that continuous gating consumes power. To address this, alternative nonvolatile structures such as ferroelectric polymers, ,, sulfur-vacancy adjustable MoS 2 , and semifloating gate (SFG) , have been introduced. Among these, the SFG structure stands out due to its well-defined dimensions, prolonged retention time, rapid response, and cost-efficiency. , Nonetheless, a comprehensive investigation into the photodetection capabilities of semifloating gate photodetectors (SFG-PDs) with 2D materials is yet to be conducted, especially in terms of understanding the underlying mechanisms of their photoresponse.…”

Neurotransmitter‐Mediated Plasticity in 2D Perovskite Memristor for Reinforcement Learning

“…With the proliferative application of sensory nodes in multiple and diverse applications, from the Internet of Things (IoT) to artificial intelligence (AI), the demand for realtime information processing is surging. [1][2][3][4] In the contemporary computing design, the analog sensory signals are converted to digital data and transferred back-and-forth between the physically separated central processing unit (CPU) and memory unit. As a result, the architecture of the separated sensory nodes, memory unit, and local processing units results in insufficient communication bandwidth, high energy consumption and data-accessing latency.…”

Photoinduced Nonvolatile Resistive Switching Behavior in Oxygen-Doped MoS2 for a Neuromorphic Vision System

“…In this work, we propose a light-modulated resistive switching device based on oxygen-doped MoS 2 (O-doped MoS 2 ). Compared to S 2– ions, the O 2– ions have a lower migration energy barrier and are more likely to drift driven by an external field. , The O-doped MoS 2 film, sandwiched between the ITO top electrode and the Pt bottom electrode, exhibits nonvolatile resistive switching behavior and optically controlled synaptic properties. The device exhibits a high switching ratio of up to 10 4 , which is significantly higher than those of other state-of-the-art devices.…”