All-Optically Modulated In-Sensor Computing Device Based on Ionic-Conducting CuInP₂Se₆

Area of Science:

• Materials Science
• Optoelectronics
• Artificial Intelligence

Background:

• In-sensor computing, inspired by the human visual system, aims to enhance real-time image processing efficiency.
• Existing optoelectronic devices face limitations in heterostructure complexity and optical modulation, hindering practical applications.

Purpose of the Study:

• To develop a simple, efficient in-sensor computing device for neuromorphic vision.
• To explore the potential of 2D CuInP2Se6 for all-optical modulation and advanced image processing.

Main Methods:

• Fabrication of a two-terminal optoelectronic device using 2D CuInP2Se6.
• Characterization of the device's tunable photoresponse across the visible spectrum (400-700 nm).
• Evaluation of bidirectional conductance modulation driven by Cu+ ions and photogenerated electrons.

Main Results:

• The device demonstrated 300 discrete, linear conductance states under red, green, and blue light.
• Achieved color-specific image feature extraction, processing, and recognition across three channels.
• Enhanced color image recognition accuracy by 4.6% when integrated with a convolutional neural network.
• Improved signal-to-noise ratio by 490% for color image preprocessing via bidirectional photoresponse.

Conclusions:

• CuInP2Se6-based devices offer a simple yet effective architecture for in-sensor neuromorphic vision.
• The developed device shows robust performance for artificial intelligence and biomimetic computing applications.
• This work paves the way for advanced, low-power visual processing systems.