Flexible wearable/stickable ammonia colorimetric sensors with assistance of deep learning for ammonia leakage monitoring and early warning

Detection and monitoring of ammonia is of great importance in modern industry and environmental protection. In this study, a colorimetric sensing system that is highly sensitive, accurate, adaptable and capable of real-time monitoring of ammonia leakage was developed. An ammonia colorimetric sensor was formed by electrostatically spinning a bromocresol green/polylactic acid nanofibrous membrane deposited on a stretchable adherent nonwoven fabric. The sensor is capable of observing a clear color change with the naked eye at 1 ppm-110 ppm, changing from yellow to blue as the ammonia concentration increases, being stable for more than 6 weeks, and remaining intact for a long time at 60 °C, with the response remaining linear and stable under different humidity tests. The sensor has excellent tensile properties and maintains a stable response after 400 tensile cycles. The system is based on deep learning algorithm and real-time analysis of images to realize the monitoring of ammonia concentration. The results show that the response/recovery time of the sensor is 9 s/7 min, and based on the self-recovery characteristics and long-term stability of the sensor, long-term monitoring and warning can be realized. In addition, we validated the colorimetric sensing system for detecting real-time changes in ammonia concentration in a simulated gas leak. In summary, this work provides effective, rapid, real-time monitoring and early warning of ammonia leakage in hazardous chemical environments.