Volatile and Nonvolatile Programmable Iontronic Memristor with Lithium Imbued TiO for Neuromorphic Computing...

Area of Science:

• Materials Science
• Neuroscience
• Computer Engineering

Background:

• Neuromorphic computing aims to mimic brain functionalities.
• Existing devices often require complex fabrication or operational conditions.
• Developing efficient hardware for simulating neural plasticity is crucial.

Purpose of the Study:

• To demonstrate a lithium-infused titanium dioxide iontronic device with inherent synapse-like short-term plasticity.
• To explore the device's potential for neuromorphic computing applications.
• To investigate the tunability and stability of the device's memory characteristics.

Main Methods:

• Fabrication of a lithium (Li) imbued TiO2 iontronic device using LiPON solid-state electrolyte.
• Characterization of short-term plasticity phenomena (paired pulse facilitation, post-tetanic potentiation).
• Evaluation of device performance in a voice recognition task.

Main Results:

• The device exhibits volatile, synapse-like short-term plasticity without a forming process or compliance current.
• Tunable memory time scales and self-relaxation characteristics were observed.
• Achieved 94.4% accuracy in voice recognition using the device as a neuromorphic training platform.

Conclusions:

• The Li-infused TiO2 iontronic device provides a versatile platform for neuromorphic computing, enabling both volatile reservoir and nonvolatile memory functions.
• The device's inherent plasticity and tunable characteristics are suitable for processing temporal information.
• This technology offers a promising approach for efficient hardware implementation of artificial neural networks.