Two-Dimensional MoS₂-Based Anisotropic Synaptic Transistor for Neuromorphic Computing by Localized Electron Beam...

Area of Science:

• Materials Science
• Computer Engineering
• Neuroscience

Background:

• Neuromorphic computing addresses the von Neumann bottleneck for advanced computing.
• Artificial synapses mimic biological systems for complex tasks.
• Current anisotropic materials for synapses have limited variety and adjustable conductivity.

Purpose of the Study:

• To create anisotropic artificial synapses on isotropic materials.
• To overcome limitations of existing anisotropic synapse materials.
• To demonstrate the potential for next-generation computing and sensing.

Main Methods:

• Localized doping of isotropic materials using electron beam irradiation (EBI).
• Purposeful induction of trap sites within the material.
• Construction of artificial neural networks (ANNs) using the fabricated synapses.

Main Results:

• Achieved anisotropic artificial synapses on isotropic materials.
• Demonstrated variable neuromorphic task accomplishment with optimized performance.
• Expanded the family of axon-multisynapse devices.

Conclusions:

• Localized doping offers a novel approach to creating advanced artificial synapses.
• This method holds significant potential for future neuromorphic computing and sensing systems.
• The technique overcomes limitations of traditional anisotropic materials.