Metal-Organic Framework-Enabled Bimodal Sensory System for Gas Leak Detection and Localization

Conventional single-modality gas monitoring systems only trigger alarms when the concentration exceeds thresholds, failing to locate leak sources. This significantly increases the difficulty and hazards of the leak investigation. To address this limitation, this study develops a bimodal sensory perception system (BSPS), which ingeniously leverages the porous nature of the metal-organic framework (MOF). On the one hand, the MOF acts as a sensing unit for efficient NH₃ capture; on the other hand, it serves as the memristive medium that modulates conductive filament migration (coefficient of variation = 0.022). In this way, the core functional components are realized by using a single material. By integrating in-sensor computing and near-sensor computing, the system independently senses, encodes, and preprocesses NH₃ concentration gradients and airflow signals. Finally, a spiking neural network (SNN) fuses the bimodal information to visually identify both the ammonia concentration and the leak source location (a recognition rate of 95%). This approach enables advanced gas-sensing platforms with enhanced selectivity and real-time leak localization capabilities.