Assessment of atmospheric distribution of polycyclic aromatic hydrocarbons using a molecular structure model

• PAHs were analysed, as the majority of the urban health risks are associated with PAHs.
• Kp was calculated for 16 PAHs in 39 specific air samples from 19 West Balkan hotspots.
• Kp predicted by SPARC was compared with the experimental values.
• Evaluation of the model predictions by comparison with other partitioning models was conducted.

In this paper, a comparison of experimentally obtained and SPARC software v4.6 modelled values of gas/particle partitioning coefficients was conducted to determine whether the evaluation of atmospheric distribution of PAH molecules can be performed using a molecular structure model. Partitioning coefficients were calculated for sixteen EPA PAHs, in thirty-nine samples of ambient air collected at nineteen urban, industrial, highly contaminated and background sites in the Republic of Serbia and Bosnia and Herzegovina. For obtaining samples of ambient air, the conventional high volume (Hi-Vol) methodology was applied, whereby gaseous and particulate phase data collection was conducted simultaneously by glass fibre filters (GFFs) and polyurethane foam filters (PUFs). The best prediction was for PAHs with 5 or more rings (benzo(b)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, indeno(1,2,3-cd)perylene and benzo(ghi)perylene). For evaluating the applicability of SPARC software predictions of gas/particle partitioning coefficients for the existing conditions, the results were compared with those obtained by applying other frequently used and highly ranked theoretical models of phase distributions, namely Junge–Pankow adsorption model, KOA absorption model, Dachs–Eisenreich dual model and PP-LFER model.