Fireworks promote the secondary formation of atmospheric imidazoles rather than their direct emission: Insights from firework source samples and machine learning
Abstract. Firework emissions release large quantities of reactive nitrogen‑ and carbon‑containing species that can modify aerosol chemical evolution, yet their impacts on imidazole‑containing particles remain poorly constrained. In this work, single‑particle aerosol mass spectrometry was deployed to probe imidazole evolution during the firework (FW) episode in Liaocheng, China, combined with source sample experiments and multi‑model machine learning analysis. Field observations revealed pronounced elevations in particle counts and fractional contributions of imidazoles during the FW episode, along with a shift from bimodal to trimodal particle-size distribution and strengthened internal mixing with sulfate, nitrate, and ammonium. Analysis of firework source samples confirmed no direct imidazole emissions from fireworks, yet abundant precursors of imidazoles (including amines, ammonium, and carbonyl compounds) were detected. Machine learning attribution identified carbonyl compounds and amines as predominant drivers for imidazole formation, with modified particle acidity acting as a secondary modulator, whereas inorganic salts, relative humidity, and NOx exerted largely indirect regulatory effects. Photochemical degradation driven by O3 constituted a major atmospheric sink for imidazoles. This study demonstrates that fireworks predominantly boost the secondary production of imidazoles rather than direct primary emissions, which advances mechanistic insights into episodic brown‑carbon evolution and nitrogen‑heterocycle aerosol pollution driven by intense short‑term combustion events.