Controllable high-performance memristors based on 2D Fe₂GeTe₃oxide for biological synapse imitation

Area of Science:

• Materials Science
• Nanotechnology
• Electronics

Background:

• Memristors are crucial for artificial neural networks and high-performance computing.
• Existing two-dimensional material memristors often trade off high performance for low power consumption.
• Modulating memristor performance after fabrication is a significant challenge.

Purpose of the Study:

• To fabricate novel, forming-free memristors with tunable performance.
• To investigate the impact of oxidation time on memristor characteristics.
• To demonstrate the potential for neuromorphic computing applications.

Main Methods:

• Fabrication of Cu/Fe3GeTe2oxide/Fe3GeTe2/Al memristors.
• Systematic variation of oxidation time during fabrication.
• Characterization of electrical properties, including On/Off ratio and set voltage.
• Analysis of conductive filament formation and rupture mechanisms.

Main Results:

• Achieved a large On/Off ratio (1.58 × 10^3) and low set voltage (0.74 V) with optimal oxidation.
• Demonstrated continuous modulation of performance by adjusting oxidation time.
• Identified the formation and rupture of Al conductive filaments as the switching mechanism.
• Showcased imitation of biological synaptic functions using voltage pulses.

Conclusions:

• The developed memristors offer tunable performance through controlled oxidation.
• The devices exhibit potential for efficient neuromorphic computing.
• This work addresses the bottleneck of fixed performance in memristor devices.