Ultrawide-bandwidth boron nitride photonic memristors.

Area of Science:

• Materials Science
• Nanotechnology
• Device Physics

Background:

• Photonic memristors are key for energy-efficient artificial vision, but current devices have limited spectral ranges and single operating modes.
• Existing technologies hinder complex computing applications due to narrow spectral responses and limited functional modes.

Purpose of the Study:

• To develop novel photonic memristor arrays with broad spectral response and reconfigurable operating modes.
• To integrate opto-sensing, data storage, and processing capabilities for advanced artificial vision systems.

Main Methods:

• Fabrication of wafer-scale hexagonal boron nitride (hBN)/silicon (Si) heterostructures using low-temperature, large-area growth.
• Characterization of opto-reconfigurability across ultraviolet to near-infrared spectrum by adjusting laser power.
• Analysis of light-induced reconfigurability mechanism involving hydrogen ions and photogenerated electrons.

Main Results:

• Demonstrated opto-reconfigurable photonic memristors with broad spectral response (UV to NIR).
• Achieved reconfigurable modes (non-resistive-switching, volatile, non-volatile) by controlling laser power.
• Exhibited high performance: switching ratio > 10^9, retention > 40,000 s, endurance > 10^6 cycles, thermal stability up to 300°C.

Conclusions:

• The hBN/Si photonic memristors offer a scalable solution for integrated sensing-storage-computation artificial vision.
• The developed devices are fully compatible with silicon-based semiconductor technologies.
• This work advances the development of ultrascalable and energy-efficient artificial vision systems.