AlScN-based ferroelectric memristor for electrical synapse emulation and light-stimulated reservoir computing.

Area of Science:

• Materials Science
• Neuroscience
• Computer Engineering

Background:

• Neuromorphic computing aims to mimic the human brain's structure and function for efficient information processing.
• Ferroelectric memristors offer promising solutions for synaptic devices due to their non-volatile memory properties.
• Integrating optical functionalities with electronic devices can enhance neuromorphic system capabilities.

Purpose of the Study:

• To present a multifunctional indium tin oxide (ITO)/aluminum scandium nitride (AlScN)/n+ silicon ferroelectric memristor.
• To investigate its potential for integrated electrical-optical neuromorphic computing.
• To demonstrate its performance in synaptic behavior and reservoir computing applications.

Main Methods:

• Fabrication of the ITO/AlScN/n+ Si memristor using radio frequency sputtering.
• Characterization of ferroelectric properties, including remanent polarization and endurance.
• Electrical and optical measurements to assess synaptic behaviors like potentiation, depression, and photoresponse.
• Implementation of a physical reservoir computing system for data processing.

Main Results:

• The memristor exhibited robust ferroelectricity (48.46 μC/cm2) and high endurance (>10^5 cycles).
• Demonstrated core synaptic behaviors (potentiation, depression) with improved linearity and accuracy.
• Showcased volatile photoresponse for optically induced synaptic plasticity.
• Achieved 96.35% classification accuracy in a reservoir computing system using dual electrical-optical modulation.

Conclusions:

• The ITO/AlScN/n+ Si memristor is a viable candidate for next-generation optoelectronic neuromorphic systems.
• Its multifunctional nature enables efficient and compact integrated electrical-optical computing.
• The device holds potential for energy-efficient artificial intelligence hardware.