Dendritic Network Implementable Organic Neurofiber Transistors with Enhanced Memory Cyclic Endurance for...

Area of Science:

• Materials Science
• Neuroscience
• Electronics Engineering

Background:

• Current neuromorphic electrochemical transistors (NETs) suffer from short retention and unstable endurance.
• Biological neurons utilize dendritic networks for complex information processing and memory.
• Developing artificial synaptic devices with enhanced stability and memory is crucial for advanced AI.

Purpose of the Study:

• To propose and demonstrate novel organic neurofiber transistors (ONTs) with a dendritic network architecture.
• To enhance memory cyclic endurance and spatiotemporal iterative learning capabilities.
• To achieve stable, long-term multilevel memory characteristics in artificial synaptic devices.

Main Methods:

• Fabrication of fibrous organic electrochemical transistors using a double-stranded assembly of electrode microfibers and an iongel gate insulator.
• Utilizing carboxylic-acid-functionalized polythiophene as the semiconductor channel material.
• Characterization of synaptic junction implementation, memory characteristics, and cyclic endurance.

Main Results:

• The proposed ONTs exhibit stable gate-field-dependent multilevel memory characteristics with long-term stability and cyclic endurance.
• Carboxylic acid dissociation enables reversible doping/dedoping, stabilizing ions and ensuring reliable device performance.
• Demonstrated successful speech recognition (88.9% accuracy) using the ONTs with iterative spiking neural network learning.

Conclusions:

• The dendritic network architecture and functionalized polythiophene enable highly sensitive, stable artificial synaptic junctions.
• The ONTs overcome limitations of conventional NETs, offering superior memory retention and cyclic endurance.
• These findings pave the way for advanced neuromorphic computing applications, including real-time speech recognition.