Forming-free chemi-memristive gas sensing for artificial olfactory system

This work introduces a novel chemi-memristive gas sensor for artificial olfactory systems that operate at room temperature without requiring external heaters. The sensor utilizes a forming-free Ti/TiO₂/Pt memristive device, where oxygen-vacancy clusters enable stable and repeatable resistive switching between high-resistance and low-resistance states. Gas sensing is achieved by selectively modulating these vacancies: exposure to oxidizing gases (e.g., NO₂) disrupts the conductive filament and increases resistance, while reducing gases (e.g., NH₃) promote filament formation with minimal resistance changes. The device exhibits rapid response and recovery times, reliable endurance over multiple cycles, and the unique ability to store gas-exposure history while adjusting risk levels through voltage tuning. Integration with an artificial neural network further validates the platform's high recognition accuracy for varying gas concentrations, which underscores its potential for low-power, in-sensor-computing applications in environmental monitoring and healthcare.