Mimicking the Synaptic Weights and Human Forgetting Curve Using Hydrothermally Grown Nanostructured CuO Memristor...

Area of Science:

• Materials Science
• Neuroscience
• Electronics

Background:

• Neuromorphic computing aims to mimic biological neural networks.
• Memristors are crucial for developing artificial synapses.
• Copper oxide (CuO) is a promising material for memristor fabrication.

Purpose of the Study:

• To fabricate CuO thin film memristors via a hydrothermal method.
• To investigate the synaptic properties of CuO memristors for neuromorphic applications.
• To compare the performance of as-prepared and annealed CuO memristors.

Main Methods:

• Hydrothermal synthesis of CuO thin films.
• Fabrication of memristor devices.
• X-ray diffraction (XRD) for structural analysis.
• Electrical characterization of synaptic behavior and retention properties.

Main Results:

• XRD confirmed polycrystalline CuO films with monoclinic structure.
• Devices exhibited analog memory and synaptic properties.
• As-prepared CuO memristors showed improved synaptic weights and rectifying behavior.
• Annealed devices had inferior synaptic performance.
• Retention loss matched human memory forgetting curves.

Conclusions:

• Hydrothermally grown CuO thin film memristors show potential for neuromorphic computing.
• As-prepared CuO devices demonstrate superior computational and learning capabilities.
• The study highlights the impact of post-fabrication annealing on memristor performance.