Multimodal Tuning of Synaptic Plasticity Using Persistent Luminescent Memitters.

Area of Science:

• Materials Science
• Neuroscience
• Computer Science

Background:

• Mimicking memory processes (encoding, storing, retrieval) is crucial for advancing neuromorphic computing and artificial intelligence.
• Simulating synaptic behavior using various materials (metal oxides, 2D materials, phase change materials) is key for power-efficient computation.

Purpose of the Study:

• To report a novel class of memristive materials, persistent luminescent memitters, for emulating biological synapse functions.
• To demonstrate the capability of these memitters in replicating memory processes and synaptic plasticity.

Main Methods:

• Utilizing persistent luminescent memitters with optical properties similar to biological synapses.
• Precisely controlling excitation frequency, wavelength, pulse number, and power density to emulate memory and plasticity.
• Employing experimental and theoretical analyses to understand the underlying mechanisms of plasticity.

Main Results:

• Successfully emulated memory processes and synaptic plasticity using persistent luminescent memitters.
• Identified electron-coupled trap nucleation and propagation via clustering as the mechanism for experience-dependent plasticity.
• Demonstrated multichannel image memorization and realization of short-term and long-term memory.

Conclusions:

• Persistent luminescent memitters offer a promising new material class for artificial synapses in neuromorphic computing.
• The study enhances the understanding of memory mechanisms through the emulation of synaptic plasticity.
• Findings pave the way for developing novel functional materials for AI and understanding biological memory.