Laser-based selective BTEX sensing using deep neural networks

A mid-infrared absorption-based laser sensor is developed for selective and simultaneous BTEX (benzene, toluene, ethylbenzene and xylenes) measurements under ambient conditions. The sensor is based on a distributed feedback inter-band cascade laser emitting near 3.3 µm. Wavelength tuning and deep neural networks were employed to differentiate the broadband absorbance of BTEX species. The sensor was validated with gas mixtures and real-time measurements were demonstrated at a temporal resolution of 1 second. Minimum detection limit of the sensor is 13 ppm of BTEX in air. This sensor can be utilized to monitor BTEX emissions in petrochemical, rubber and paint industries to avoid hazardous health effects.