Highly sensitive detection of atmospheric HONO using mid-infrared spectroscopy combined with deep learning

Gaseous nitrous acid (HONO) is an important source of hydroxyl radicals (OH) in the atmosphere, significantly influencing atmospheric oxidation capacity and the formation of secondary pollution. However, its extremely low environmental concentration, combined with considerable spatial and temporal variations, presents challenges for high-precision monitoring. This study employs a quantum cascade laser (QCL) with a central wavelength of 1280 cm-1, utilizing highly sensitive Tunable Laser Absorption Spectroscopy (TLAS) and Cavity Ring-Down Spectroscopy (CRDS) techniques to measure HONO. Initially, high-precision calibration measurements were conducted on the HONO absorption lines within this wavelength range. Subsequently, the acquired spectral line data were used to carry out highly sensitive measurements and noise reduction on the HONO spectral lines, employing CRDS technology alongside time convolutional neural networks. The findings of this study indicate that mid-infrared spectroscopy, in combination with deep learning analysis, provides an efficient and reliable new technological approach for real-time, high-precision monitoring of atmospheric trace HONO.