Emulating Artificial Synaptic Plasticity Characteristics from SiO₂-Based Conductive Bridge Memories with Pt...

Area of Science:

• Materials Science
• Neuroscience
• Computer Engineering

Background:

• The rapid advancement of information technology demands electronic devices for efficient neuromorphic computing.
• Artificial neural networks require synaptic elements that can emulate biological synaptic properties for effective learning.

Purpose of the Study:

• To investigate the impact of platinum nanoparticle layers on bipolar switching in silicon dioxide conductive bridge memories.
• To explore the emulation of diverse artificial synaptic functionalities using these engineered memory devices.

Main Methods:

• Fabrication of SiO2-based conductive bridge memories with a dense layer of Pt nanoparticles as the bottom electrode.
• Utilized a numerical model to analyze the influence of bottom electrode thermal conductivity on filament formation.
• Implemented an intermediate switching transition slope during the SET process.

Main Results:

• Demonstrated that Pt nanoparticles significantly influence the bipolar switching effect in SiO2 memory devices.
• Identified the role of bottom electrode thermal conductivity in the conducting filament growth mechanism.
• Successfully emulated various forms of synaptic plasticity, including short-term (facilitation, depression) and long-term plasticity, as well as spike-dependent plasticity.

Conclusions:

• The developed Pt nanoparticle-based synaptic elements offer a pathway to multifunctional, low-power neuromorphic computing.
• These elements exhibit biological-like behavior, crucial for advanced artificial intelligence and brain-inspired computing.
• The findings provide critical insights for designing next-generation synaptic devices for efficient information processing.