Memristive Devices from CuO Nanoparticles.

Area of Science:

• Materials Science
• Electronics Engineering
• Computer Science

Background:

• The von Neumann bottleneck hinders computational efficiency in artificial neural networks (ANNs).
• Memristive systems offer in-situ data processing to overcome this limitation.
• High-reliability memristor devices are essential for scalable ANN implementations.

Purpose of the Study:

• To develop and characterize a novel Ag/CuO/SiO2/p-Si heterostructure memristor.
• To investigate the device's resistive switching (RS) and negative differential resistance (NDR) properties.
• To understand the role of the native oxide layer in device performance.

Main Methods:

• Fabrication of Ag/CuO/SiO2/p-Si and Ag/CuO/ITO memristors using ultra-spray pyrolysis (USP).
• Material characterization using Raman spectroscopy and scanning electron microscopy (SEM).
• Electrical characterization through current-voltage (I-V) measurements.

Main Results:

• The Ag/CuO/SiO2/p-Si device exhibited stable RS and NDR.
• The native silicon oxide layer improved reset variability and facilitated NDR.
• Resistive switching was attributed to traps in amorphous CuO, with set process dominated by trap-controlled space-charge-limited (SCLC) conduction and Poole-Frenkel.
• Comparison with Ag/CuO/ITO device highlighted the role of the native oxide.

Conclusions:

• The developed memristor shows promise for efficient ANNs by addressing the von Neumann bottleneck.
• The native oxide layer plays a critical role in enhancing device reliability and enabling NDR.
• This research paves the way for improved memristive devices with combined RS and NDR functionalities.