Synaptic Transistors Exhibiting Gate-Pulse-Driven, Metal-Semiconductor Transition of Conduction

Area of Science:

• Materials Science
• Electrical Engineering
• Computer Science

Background:

• Neuromorphic devices are being explored as alternatives to conventional semiconductor devices.
• Existing neuromorphic devices have limitations that hinder widespread adoption.
• Deep trap interfaces offer a novel approach for neuromorphic device design.

Purpose of the Study:

• To propose and characterize a new neuromorphic device utilizing deep trap interfaces.
• To evaluate the synaptic properties and stability of the proposed device.
• To assess the potential of this device to replace conventional semiconductor devices.

Main Methods:

• Fabrication of the device using in-situ atomic layer deposition (ALD) for sequential deposition of charge-inducing dielectrics (CID) and oxide semiconductors.
• Application of gate bias pulses to observe changes in conducting states.
• Analysis of device states, persistence, linearity, and symmetry for excitatory and inhibitory behaviors.

Main Results:

• Observed an abrupt change in conducting states from semiconductor to metal upon gate bias pulse application.
• Demonstrated the implementation of numerous intermediate states based on the number of deposition cycles.
• Showcased excellent long-term memory with state persistence for 10,000 s and superior linearity and symmetry for synaptic behaviors.

Conclusions:

• The developed neuromorphic device exhibits unique and stable synaptic properties.
• The device's performance suggests its potential as a future replacement for conventional semiconductor devices.
• The use of deep trap interfaces and ALD fabrication offers a promising pathway for advanced neuromorphic computing.