Near ideal synaptic functionalities in Li ion synaptic transistor using Li₃PO_xSe_x electrolyte with high ionic...

Area of Science:

• Materials Science
• Solid-state ionics
• Neuromorphic Engineering

Background:

• Solid-state lithium-ion transistors are key for artificial neural networks in neuromorphic computing.
• Existing electrolytes exhibit slow ionic conduction, limiting synaptic transistor performance.

Purpose of the Study:

• To investigate the impact of electrolyte ionic conductivity on synaptic transistor performance.
• To enhance ionic conduction for improved artificial synapse functionality.

Main Methods:

• Systematic exploration of electrolyte ionic conductivity effects.
• Isovalent chalcogenide (Se) substitution in Li3PO4 to reduce Li ion migration activation energy.
• Fabrication and characterization of LiCoO2 channel synaptic transistors.

Main Results:

• Selenium substitution reduced activation energy for Li ion migration from 0.35 to 0.253 eV, enabling fast ionic conduction.
• Optimized devices exhibited linear conductance switching, discrete nonvolatile states, and good retention.
• Achieved a low nonlinearity ratio of 0.12 and a high on/off ratio of 19.
• Lower conductivity Li3PO4 showed asymmetric and nonlinear weight-update characteristics.

Conclusions:

• Facilitating Li ionic conduction in solid-state electrolytes is crucial for high-performance artificial synapse devices.
• Enhanced ionic conductivity leads to linear and nonvolatile synaptic behavior.
• This research paves the way for developing advanced neuromorphic computing hardware.