Reconfigurable halide perovskite nanocrystal memristors for neuromorphic computing

Area of Science:

• Materials Science
• Nanotechnology
• Solid-State Electronics

Background:

• In-memory computing requires electronic devices with specific, often reconfigurable, switching characteristics.
• Current memristive devices lack the versatility to meet diverse volatile and non-volatile switching demands, necessitating application-specific material designs.
• The development of "reconfigurable memristors" combining ionic diffusive and drift mechanisms is crucial for universal applications but remains a significant challenge.

Purpose of the Study:

• To present a novel reconfigurable halide perovskite nanocrystal memristor.
• To demonstrate on-demand switching between diffusive/volatile and drift/non-volatile modes.
• To showcase the device's potential in advanced computing architectures.

Main Methods:

• Fabrication of a memristor using halide perovskite nanocrystals and organic capping ligands.
• Utilizing controllable electrochemical reactions to tune device switching behavior.
• Characterization of device performance, including endurance in both volatile and non-volatile modes.

Main Results:

• Achieved on-demand switching between volatile (diffusive) and non-volatile (drift) operating modes.
• Demonstrated state-of-the-art endurance: 2 × 10^6 cycles for volatile mode and 5.6 × 10^3 cycles for non-volatile mode.
• Successfully implemented the perovskite memristor in a benchmark reservoir network, integrating volatile recurrent and non-volatile readout layers.

Conclusions:

• The halide perovskite nanocrystal memristor offers a versatile platform for in-memory computing by enabling reconfigurable switching.
• The device exhibits excellent endurance, making it suitable for complex computational tasks.
• This work paves the way for universal memristive devices by overcoming limitations of current technologies.