Near-infrared Real-Time Trace NH3 Sensor Based on WM-OA-ICOS and EEMD Assisted Optical Denoising

In this paper, a near-infrared NH3 sensor is developed using Wave-length Modulated Off-Axis Integrated Cavity Output Spectroscopy (WM-OA-ICOS) technology. Design a cavity with a length of 15 cm and an effective absorption diameter of 330.5 m. The Ensemble Empirical Mode Decomposition (EEMD) algorithm is adopted as the spectra preprocessing method to reduce the noise, and Allan deviation analysis is conducted by comparing the original signal. It is found that the limit of detection (LoD) of the sensor could reach ~1.54 ppb with an integration time of 43 s. The concentration calibration experiment shows that the sensor can achieve double range. To enhance accuracy and sensitivity in NH3 concentration inversion, the Cuckoo Search algorithm with the Elman neural network (CS-Elman) was designed. Through experimental validation, the sensor achieves rapid detection of trace NH3. The outstanding performance of the NH3 detection device in monitoring trace gases has been demonstrated by field deployment trials.