High-Performance Artificial Synapse Based on CVD-Grown WSe₂ Flakes with Intrinsic Defects

Area of Science:

• Materials Science
• Neuroscience
• Computer Engineering

Background:

• Advanced neuromorphic systems require high-performance artificial synaptic devices.
• Tungsten diselenide (WSe2) is a promising material for electronic applications.

Purpose of the Study:

• To develop and characterize artificial synaptic devices based on CVD-grown WSe2 flakes.
• To explore the potential of these devices in neuromorphic computing and learning simulations.

Main Methods:

• Fabrication of synaptic devices using CVD-grown WSe2 flakes.
• Characterization of synaptic behaviors including plasticity and light-dependent responses.
• Simulation of an artificial neural network for pattern recognition using MNIST dataset.

Main Results:

• The WSe2 synaptic devices demonstrated robust synaptic behaviors (e.g., paired-pulse facilitation, short- and long-time plasticity).
• Optoelectronic properties enabled light-dosage and wavelength-dependent plasticity, mimicking learning and memory.
• Achieved 92.9% accuracy in MNIST handwritten digit recognition through weight updating.

Conclusions:

• Intrinsic defects in CVD-grown WSe2 are crucial for controllable synaptic plasticity.
• WSe2-based optoelectronic synapses show significant potential for high-performance neuromorphic computation.