Search
Menu

Tunable Photomemristor Streamlines Neuromorphic Vision Processing

Facebook X LinkedIn Email
HANGZHOU, China, Feb. 22, 2023 — Researchers at Hangzhou Institute for Advanced Study, State Key Laboratory of Infrared Physics, and Shanghai Institute of Technical Physics have created a two-terminal, nonvolatile photomemristor with a simple architecture and tunable photoresponsivity. The advancement represents a more efficient approach to processing data and reducing power usage than that commonly found in conventional artificial intelligence (AI) machine vision technology, where the frequent movement of large amounts of data between sensors, processors, and memory can result in high power consumption and latency.

The photomemristor is based on 2D graphene/MoS2-xOx/graphene (G/M/G) structures. Photoexcited carriers and oxygen-related ions are coupled in the two-terminal G/M/G architecture, leading to a displaced and pinched hysteresis in the device’s current-voltage characteristic (also known as I-V curve).

The nonvolatile photomemristor can store and read multiple photoresponse states in a nonvolatile mode at zero external voltage. Its switching properties can be jointly controlled by the electric-field-driven migration of ions and photo-induced redox reactions at the asymmetric G/M/G contacts.

By mimicking the biological functions of the human retina and designing specific device structures, the researchers made it possible for the photomemristor to act as a neural network for neuromorphic visual processing and for the implementation of computationally complete photoresponse-stateful logic operations triggered by electrical and light stimuli together.

According to the researchers, the two-terminal photomemristor can support versatile sensing-memory-computing approaches for implementing an in-sensor computing network.

“We designed a two-terminal device with MoS2-xOx and specific graphene for three purposes in one,” the researchers said. “One — to provide low-barrier energy for the migration of oxygen ions; two — to perform as geometry-asymmetric, metal/semiconductor/metal van der Waals heterostructures with multiphotoresponse states; and three — as an extension of a memristor, this device not only provides tunable conductance, but also demonstrates reconfigurable photoresponse for reading at zero bias voltage.”


The researchers said that the tunable short-circuit photocurrent and photoresponse can be increased to 889.8 nA and 98.8 mA/W, respectively. This is a much higher value than other reconfigurable phototransistors based on 2D materials can achieve, the researchers added.

“To reverse the channel polarity and obtain a gate-tunable short-circuit photocurrent, the channel semiconductor must be thin enough. Thus, it is difficult to use the thick film needed to absorb enough light to get a large signal,” the researchers said. “In our case, the mechanism of the two-terminal device rearrangement is based on ion migration, which is not limited by the thickness. We can increase the thickness of the film to absorb more photons and get a large short-circuit photocurrent.

“This new concept of a two-terminal photomemristor not only enables all-in-one sensing-memory-computing approaches for neuromorphic vision hardware, but also brings great convenience for high-density integration.”
A conceptual graph of the all-in-one sensing-memory-computing photo-memristor. In a research advance, reportedly for the first time, a team has developed non-volatile photomemristors that implement computationally complete logic with photoresponse-stateful operations, for which the same photo-memristor serves as both a logic gate and memory, using photoresponse as a physical state variable instead of light, voltage, and memresistance. Courtesy of Xiao Fu, et al.
A conceptual graph of the all-in-one sensing-memory-computing photomemristor. In a research advancement, reportedly for the first time, a team has developed nonvolatile photomemristors that implement computationally complete logic with photoresponse-stateful operations, for which the same photomemristor serves as both a logic gate and memory, using photoresponse as a physical state variable instead of light, voltage, and memresistance. Courtesy of Xiao Fu et al.  
The research was published in Light: Science & Applications (www.doi.org/10.1038/s41377-023-01079-5).

Published: February 2023
Glossary
neuromorphic
Neuromorphic refers to the design and development of computing systems or devices that are inspired by the structure and functioning of the human brain and nervous system. The term is derived from "neuro," meaning relating to the nervous system, and "morphic," referring to form or structure. Neuromorphic computing aims to mimic the parallel processing, connectivity, and energy efficiency observed in biological neural networks. Key characteristics of neuromorphic systems include: Parallel...
vision
The processes in which luminous energy incident on the eye is perceived and evaluated.
photocurrent
The current that flows through a photosensitive device, such as a photodiode, as the result of exposure to radiant power.
neuromorphicvisionvision processingdata processingmemristorsmemorycomputedatainformationphotomemresistorsphotomemresistorphotoresponsephotocurrentsensingAIResearch & TechnologyeducationAsia PacificState Key LaboratoryHangzhou Institute for Advanced Study

We use cookies to improve user experience and analyze our website traffic as stated in our Privacy Policy. By using this website, you agree to the use of cookies unless you have disabled them.