Evaluation of a conceptual model for gas-particle partitioning of polycyclic aromatic hydrocarbons using poly-parameter linear free energy relationships

A model for gas-particle partitioning of polycyclic aromatic hydrocarbons (PAHs) was evaluated using polyparameter linear free energy relationships (ppLFERs) following a multiphase aerosol scenario. The model differentiates between various organic (i.e., liquid water-soluble (WS)/organic soluble (OS) organic matter (OM), and solid/semisolid organic polymers) and inorganic phases of the particulate matter (PM). Dimethyl sulfoxide and polyurethane were assigned as surrogates to simulate absorption into the above-mentioned organic phases, respectively, whereas soot, ammonium sulfate, and ammonium chloride simulated adsorption processes onto PM. The model was tested for gas and PM samples collected from urban and nonurban sites in Europe and the Mediterranean, and the output was compared with those calculated using single-parameter linear free energy relationship (spLFER) models, namely Junge-Pankow, Finizio, and Dachs-Eisenreich. The ppLFER model on average predicted 96 ± 3% of the observed partitioning constants for semivolatile PAHs, fluoranthene, and pyrene, within 1 order of magnitude accuracy with root-mean-square errors (RMSE) of 0.35–0.59 across the sites. This was a substantial improvement compared to Finizio and Dachs-Eisenreich models (37 ± 17 and 46 ± 18% and RMSE of 1.03–1.40 and 0.94–1.36, respectively). The Junge-Pankow model performed better among spLFERs but at the same time showed an overall tendency for overestimating the partitioning constants. The ppLFER model demonstrated the best overall performance without indicating a substantial intersite variability. The ppLFER analysis with the parametrization applied in this study suggests that the absorption into WSOSOM could dominate the overall partitioning process, while adsorption onto salts could be neglected