Control-Etched Ti₃C₂T MXene Nanosheets for a Low-Voltage-Operating Flexible Memristor for Efficient Neuromorphic...

Area of Science:

• Materials Science
• Electronics Engineering
• Nanotechnology

Background:

• Flexible hardware neural networks are crucial for advanced wearable electronics.
• Developing synaptic plastic networks with low operating voltages remains a challenge.

Purpose of the Study:

• To demonstrate a flexible memristor device based on partially etched MXene (p-MXene) for low-voltage operation.
• To investigate the synaptic plasticity, data storage capabilities, and neuromorphic computing potential of the developed device.

Main Methods:

• Fabrication of a flexible memristor using PET/ITO/p-MXene/Ag structure by controlling aluminum ion etching in Ti3C2Tx MXene.
• Characterization of device performance, including operating voltage, ON/OFF ratio, endurance, data retention, and mechanical/environmental stability.
• Demonstration of synaptic plasticity (potentiation/depression) and Spike-Time-Dependent Plasticity (STDP).
• Simulation of pattern recognition using the MNIST dataset.

Main Results:

• The p-MXene memristor achieved a significantly reduced operating voltage of 1 V compared to 7 V for fully etched MXene devices.
• The device exhibited excellent non-volatile data storage with a ~10^3 ON/OFF ratio, ~10^4 cycle endurance, multilevel states, and ~10^6 s data retention.
• High mechanical stability (~73° bending) and environmental robustness were confirmed.
• Successful mimicry of biological synapse functions, including learning-forgetting patterns and STDP.
• Achieved ~95% accuracy in MNIST pattern recognition with 29 training epochs in simulations.

Conclusions:

• Partially etched MXene is a promising material for low-voltage flexible memristors.
• The developed device shows significant potential for next-generation smart wearable electronics, data storage, and neuromorphic computing.