Monolithic Integration of Sensing, Computing, and Storage in an Anomalous Hall Effect-Based Neuromorphic Device.

Area of Science:

• Materials Science
• Neuroscience
• Electrical Engineering

Background:

• Neuromorphic computing faces challenges due to the physical separation of sensing, memory, and processing units.
• Intrinsic integration of these functions is crucial for advancing neuromorphic architectures.

Purpose of the Study:

• To develop a monolithic device integrating magnetic sensing, memristor-based storage, and analog computation.
• To address the limitations of conventional neuromorphic computing by creating a unified functional unit.

Main Methods:

• Fabrication of a novel Ag/NiO/NiCo2O4 heterostructure.
• Exploitation of the anomalous Hall effect in NiCo2O4 for sensing and computation.
• Utilizing a NiO layer as a nonvolatile memristor for reconfigurable storage via interfacial modulation.

Main Results:

• The device integrates magnetic sensing, memristor storage, and analog computation.
• NiO memristor programming (±2 V) nonvolatily modulates the NiCo2O4's anomalous Hall effect response.
• Programmatic switching between high-sensitivity and wide-dynamic-range modes was achieved.

Conclusions:

• The monolithic Ag/NiO/NiCo2O4 heterostructure offers a viable foundation for energy-efficient neuromorphic sensors.
• This integrated approach supports the development of multifunctional autonomous systems.
• Interfacial modulation via oxygen vacancy migration is key to device functionality.