An Antagonistic Photovoltaic Memristor for Bioinspired Active Contrast Adaptation

Area of Science:

• Materials Science
• Optoelectronics
• Neuromorphic Engineering

Background:

• Machine vision systems are crucial for intelligent applications but struggle with limited dynamic range and fixed photoresponsivity.
• High-contrast scenes challenge current systems, leading to loss of image fidelity due to saturation and poor shadow detail.

Purpose of the Study:

• To develop a novel photosensor with an adaptive dynamic range for improved image capture in challenging lighting.
• To demonstrate a photovoltaic memristor capable of autonomous light response adjustment.

Main Methods:

• Fabrication of a photovoltaic memristor with two antagonistic photovoltaic junctions.
• Investigation of the dynamic photodoping effect at an asymmetrical p-n junction.
• Characterization of light-intensity-dependent switchable photovoltaic behaviors and contrast adaptation.

Main Results:

• The photovoltaic memristor demonstrated autonomous adjustment of its response to varying light stimuli.
• Photocurrent polarities switched with increasing light intensity, consistent with Weber's law.
• An 11x11 memristor array achieved human-like active contrast adaptation (94 dB dynamic range in under 1.2 seconds).

Conclusions:

• The developed photovoltaic memristor offers a solution to the dynamic range limitations in current machine vision systems.
• This technology enables state-of-the-art active visual adaptation in photosensors.
• The findings pave the way for advanced neuromorphic device designs.