Versatile optoelectronic memristor based on wide-bandgap Ga₂O₃ for artificial synapses and neuromorphic computing

Area of Science:

• Materials Science
• Nanotechnology
• Optoelectronics

Background:

• Optoelectronic memristors are crucial for developing neuromorphic visual systems (NVs).
• Existing devices often face limitations in data storage capacity and mimicking complex neural functions.
• Amorphous wide-bandgap Gallium Oxide (Ga2O3) offers potential for advanced memristive applications.

Purpose of the Study:

• To introduce a novel amorphous wide-bandgap Ga2O3 photoelectric synaptic memristor.
• To demonstrate its capabilities in multi-bit data storage and logic gate implementation.
• To evaluate its synaptic characteristics and performance in artificial neural networks (ANNs).

Main Methods:

• Fabrication of an Ag/Ga2O3/Pt memristive device.
• Utilizing adjustable current compliance (Icc) and variable ultraviolet (UV) light intensities for data storage.
• Implementing "AND" and "OR" logic gates using voltage polarity and UV light.
• Characterizing synaptic behaviors like paired-pulse facilitation (PPF) and spike-timing-dependent plasticity (STDP).
• Integrating the device into an ANN for pattern recognition tasks.

Main Results:

• Achieved 3-bit data storage by tuning Icc and UV light intensity.
• Successfully implemented "AND" and "OR" logic gates.
• Demonstrated stable and diverse synaptic plasticity, including PPF, SIDP, SNDP, STDP, and SFDP.
• The integrated device mimicked optical potentiation and electrical depression with 90.7% pattern accuracy in an ANN.

Conclusions:

• The amorphous Ga2O3 photoelectric synaptic memristor offers multifunctional capabilities for optoelectronic memory.
• The device shows promise for applications in neuromorphic computing and artificial visual perception.
• Its ability to mimic synaptic behavior and achieve high accuracy in ANNs highlights its potential for future intelligent systems.