Reconfigurable Al₂O₃-Based Memristor for All-in-One Artificial Synapse and Nociceptor Neurons

Area of Science:

• Materials Science
• Electronics Engineering
• Neuroscience

Background:

• Memristors are crucial for neuromorphic computing, intelligent sensors, and robotics.
• Existing memristors often lack reconfigurable switching for diverse applications.
• Volatile and nonvolatile operations are critical for different bionic device functionalities.

Purpose of the Study:

• To develop a novel reconfigurable memristor with adjustable switching behavior.
• To integrate artificial synaptic and nociceptor functions into a single bionic device.
• To explore the potential of this device in neuromorphic network applications.

Main Methods:

• Fabrication of a novel Ag/Al2O3/ITO memristor.
• Modulation of compliance current to achieve adjustable volatile and nonvolatile switching.
• Simulation of synaptic functionality and nociceptor characteristics.

Main Results:

• Demonstrated adjustable switching between volatile and nonvolatile states by modulating compliance current.
• Proposed a mechanism controlling conductive filament states for adjustable switching.
• Successfully simulated synaptic and nociceptor functions within the same device.
• Achieved low single-pulse power consumption (0.912 nJ) for nociceptor function.

Conclusions:

• The developed memristor offers reconfigurable switching for multifunctional bionic devices.
• Integration of synaptic and nociceptor capabilities in a single device is feasible.
• The device shows significant promise for advanced neuromorphic computing and sensor applications.