Integrating FT-ICR MS with rule-based prediction and QSAR to resolve nitrogen-containing organic compounds and their toxicity in wildfire-derived PM2.5.

The intensification of wildfire seasons, driven by global warming, has established wildfires as a primary contributor to aerosol particulate matter (APM) levels. However, the formation mechanisms and molecular composition of nitrogen-containing organic compounds (NOCs) in wildfire-derived APM remain poorly understood. This study characterizes, at the molecular level, wildfire-derived PM2.5 from a wildfire in a temperate coniferous forest in South Korea. Gas chromatography-mass spectrometry analysis revealed that nitrated polycyclic aromatic compound concentrations were approximately 3.4 times higher during the wildfire period (0.122 ng/m3) than the post-wildfire period (0.036 ng/m3), with 2-nitrofluorene and 3-nitrofluoranthene predominating in each period, respectively. Fourier transform ion cyclotron resonance mass spectrometry analysis revealed a relative shift toward CHON compounds across the sampling periods, and rule-based secondary organic aerosol predictions using 12 volatile organic compound precursors matched against observed data, with limonene-derived products showing the highest number of molecular-formula matched compounds. Quantitative structure activity relationship (QSAR)-based lethal concentration (LC50) predictions showed that β-pinene-, α-pinene-, and isoprene-derived compounds exhibited predicted values of 1.0, 1.6, and 8.3 mg/L, respectively. Because these predictions were based on structures assigned at the molecular-formula level, they provide a preliminary hazard ranking of the associated precursor classes rather than confirmed toxicity. Taken together, the observed compositional shifts point to secondary atmospheric processing of NOCs as a factor meriting further consideration in wildfire assessments.