CdS-GeSe Heterostructure-Based Bimodal Memristor with Tunable Synaptic Plasticity and Mixed-Signal Switching for...

Area of Science:

• Materials Science
• Neuroscience
• Computer Science

Background:

• Neuromorphic computing aims to mimic the human brain's structure and function.
• Memristors are key components for developing brain-inspired computing systems.
• Optically triggered devices offer novel pathways for information processing.

Purpose of the Study:

• To develop and characterize an Ag/CdS-GeSe/FTO memristor for optically triggered neuromorphic applications.
• To investigate the device's multifunctional capabilities for synaptic learning and memory storage.
• To evaluate the memristor's performance in image recognition tasks using machine learning datasets.

Main Methods:

• Fabrication of an Ag/CdS-GeSe/FTO memristor device.
• Optoelectrical characterization under electrical and optical stimuli (405 nm pulses).
• Evaluation using Fashion-MNIST and digit MNIST datasets for image recognition accuracy.

Main Results:

• The memristor exhibited analog-to-digital transition and digital switching within ±0.5 V post-electroforming.
• Demonstrated synaptic functions including potentiation, depression, PPF, and PPD.
• Achieved 75% (Fashion-MNIST) and 95% (digit MNIST) accuracy in optoelectrical mode for image recognition.

Conclusions:

• The Ag/CdS-GeSe/FTO memristor shows potential for neuromorphic computing, machine learning, and optical modulation.
• Optoelectrical operation enhances performance compared to electrical mode.
• The device offers stable data retention and excellent reproducibility for reliable applications.