Reconfigurable logic in nanosecond Cu/GeTe/TiN filamentary memristors for energy-efficient in-memory computing

Area of Science:

• Materials Science
• Nanotechnology
• Computer Engineering

Background:

• In-memory computing with memristors is crucial for overcoming von Neumann architecture limitations.
• Developing high-performance memristive devices and logic methodologies for energy-efficient computing remains a challenge.

Purpose of the Study:

• To report a filamentary Cu/GeTe/TiN memristor with high performance characteristics.
• To elucidate the resistive switching mechanisms.
• To demonstrate reconfigurable logic functions and potential applications in energy-efficient in-memory computing.

Main Methods:

• Fabrication and characterization of filamentary Cu/GeTe/TiN memristor devices.
• Analysis of current-voltage and resistance-temperature behaviors to determine conduction mechanisms.
• Experimental implementation of logic functions (IMP, NOT, OR, COPY) and simulation of a one-bit full adder.

Main Results:

• The Cu/GeTe/TiN memristor exhibits nanosecond switching speed (<60 ns), low voltage operation (<2 V), high endurance (>10^4 cycles), and good retention (>10^4 s @85 °C).
• Schottky emission and hopping transport mechanisms were identified for high and low resistance states, respectively.
• Reconfigurable logic functions and a one-bit full adder were successfully implemented and validated.

Conclusions:

• The developed memristor demonstrates significant potential for energy-efficient in-memory computing applications.
• Understanding the resistive switching mechanisms is key to optimizing memristive device performance.
• The proposed reconfigurable logic method offers a promising approach for future computing architectures.