As epidemiological evidence continues to mount, it has become undeniable that exposure to high levels of airborne pollutants such as SO_x and NO_x are detrimental to human health. In recent years, millions of premature deaths by stroke, coronary heart disease, and lung cancer worldwide have been linked to poor outdoor air quality.

Unfortunately, not all polluting industries have faced the same stringent regulations. For instance, restrictions on harmful pollutant emissions from road vehicles have remained higher than those from marine transport. This discrepancy between sectors is expected to shrink as an increasing number of industries come into the spotlight of regulators. In this rapidly changing landscape, the demand for effective and innovative pollution abatement solutions is rising.

In the present work, our focus is on investigating the viability of a novel airborne pollutant removal concept. In this no-waste design, SO_x and NO_x are trapped into ammonium salt particles that can be then sold as an agricultural fertilizer. The gaseous pollutants are first oxidized by ozone, which is then mixed with ammonia in humid air to allow the ammonium particles to form and grow.

The study of this system requires analyzing the interplay between chemical reactions and the turbulent fluid dynamics that enables them through efficient mixing. To this end, numerical simulations are an invaluable tool that facilitates uncovering detailed knowledge where experimental studies may be intractable. Here, we leverage the use of high-fidelity large-eddy simulations to study reactive and non-reactive flow conditions relevant to this multi-pollutant removal solution. These investigations are supplemented by reactor modeling approaches to analyze specific key chemical processes. Finally, we implement and employ state-of-the-art data-driven methods that provide enhanced insight into our numerical datasets. For this purpose, we apply proper orthogonal decomposition, a machine-learning workflow for automated region identification, and global sensitivity analysis.