Noise Tailoring in Memristive Filaments.

Area of Science:

• Materials Science
• Solid-State Electronics
• Device Physics

Background:

• Filamentary resistive switching memory devices are crucial for modern electronics.
• Understanding and controlling noise in these devices is essential for performance and novel applications.
• Low-frequency 1/f-type noise is a key characteristic impacting device behavior.

Purpose of the Study:

• To investigate the potential for noise tailoring in filamentary resistive switching memory.
• To study the resistance and frequency scaling of noise properties in mainstream material systems.
• To explore the use of these devices as tunable noise sources.

Main Methods:

• Investigated noise properties in Ta2O5, Nb2O5, and Ag-based resistive switching devices.
• Analyzed the impact of material choice on the noise floor.
• Studied the influence of varying resistance states on noise levels.
• Applied a point-contact noise model to understand noise mechanisms.

Main Results:

• Demonstrated order-of-magnitude lower noise floors in Ta2O5 and Nb2O5 compared to Ag-based devices.
• Showed that resistance state variations allow orders-of-magnitude tuning of noise levels across different materials.
• Identified disorder-induced suppression of noise from remote fluctuators.

Conclusions:

• Noise in resistive switching memory is tailorable through material selection and resistance state control.
• These findings enable the development of multipurpose resistive switching units.
• Devices can function as both tunable memory elements and tunable noise sources for probabilistic computing.