A Monolithic 3D/2D Perovskite Memristor Enabling Ultralow-Voltage Neuromorphic Computing and Biomimetic Sensing

Area of Science:

• Materials Science
• Neuroscience
• Electronics

Background:

• Perovskite memristors offer potential for neuromorphic computing and bioinspired sensing.
• Current limitations include high operating voltages, environmental instability, and restricted functional integration.

Purpose of the Study:

• To develop a robust memristor overcoming current limitations in perovskite technology.
• To integrate multiple functionalities onto a single platform for advanced applications.

Main Methods:

• Fabrication of a memristor using a 3D/2D halide perovskite heterostructure.
• Characterization of device performance including switching voltages, ON/OFF ratio, retention, and endurance.
• Evaluation of neuromorphic and sensory functions, including synaptic plasticity emulation, temporal pattern processing, and artificial nociception.

Main Results:

• Achieved ultralow switching voltages (∼+0.18 V SET, ∼-0.4 V RESET) and high ON/OFF ratio (>10^4).
• Demonstrated excellent retention (>10^4 s) and endurance (>600 cycles) under ambient conditions.
• Successfully emulated synaptic plasticity (90.77% MNIST accuracy), temporal pattern processing (perfect accuracy), and artificial nociception.

Conclusions:

• The 3D/2D perovskite heterostructure enables low-voltage, stable, and multifunctional memristor devices.
• The synergistic mechanism at the heterointerface enhances performance and suppresses ion migration.
• This versatile platform serves as a prototype for next-generation intelligent systems and adaptive human-machine interfaces with ultralow power consumption (36 pJ/switch).