Particle/gas partitioning behavior of polychlorinated biphenyls (PCBs) in global atmosphere: Equilibrium or steady state?. (1st February 2022)

• Record Type: Journal Article
• Authors: Qiao, Li-Na; Ma, Wan-Li; Liu, Li-Yan; Zhang, Zi-Feng; Song, Wei-Wei; Jia, Hong-Liang; Nikolaev, Anatoly N.; Macdonald, Robie W.; Li, Yi-Fan
• Is Part Of: Atmospheric environment. Volume 270(2022)
• Publication Date: 2022-02-01
• Subjects: PCBs -- Particle/gas partitioning -- Equilibrium -- Steady state -- Octanol-air partition coefficient; Air -- Pollution -- Periodicals; Air -- Pollution -- Meteorological aspects -- Periodicals; 551.51
• DOI: 10.1016/j.atmosenv.2021.118926
• Languages: English
• ISSNs: 1352-2310

Abstract

Abstract: Particle/gas (P/G) partitioning governs the atmospheric transport and fate of semi-volatile compounds (SVOCs). Here, we provide a comprehensive investigation of P/G partitioning of polychlorinated biphenyls (PCBs) by comparing the performance of seven models, (including five equilibrium models: Junge-Pankow model, Harner-Bidleman model, Dachs-Eisenreich model, pp-LFER model, mp-pp-LFER model, one empirical Li-Jia model, and one steady-state Li-Ma-Yang model). For this comparison, 5019 monitoring data points from urban, rural and remote sites for PCBs taken from published papers have been accessed, against which the predictions of these models have been evaluated. Among the 5019 monitoring data points, there are 2791 monitoring data for six selected PCB congeners (CB-28 + 31, −52, −101, −153, −180, −207) and 2228 for dozens of other PCB congeners. We find that in predicting the P/G partition quotient ( K P ) of the six selected PCB compounds, the Li-Ma-Yang model performs best across the entire log K OA (octanol-air partition coefficient) range (from 7.8 to 15.3), with 86 ± 10% of 2791 monitoring data points lying within a deviation of ±1 log unit of the root mean square error (RMSE) of 0.67 ± 0.18 in urban, rural and remote areas. This model is far superior to the other models in predicting K P, especially in the MP Domain (log K OA ≥ 12.50), with 96% of the data points falling within a deviation of ±1 log unit and a RMSE of 0.48, in comparison to the less than 7% for the five equilibrium models and 20% for the empirical Li-Jia model. These findings suggest that when the particle deposition is not negligible, the P/G partitioning of PCBs in ambient atmosphere is not in equilibrium, but in a steady or quasi-steady state. Similar to PBDEs, the properties of individual PCB congeners, such as K OA, are the primary factors affecting the P/G partitioning of PCBs. Graphical abstract: Image 1 Highlights: Seven models have been compared for their performance in predicting K P values of PCBs. Particle deposition is a non-ignorable determining process in P/G partitioning of PCBs. The new equation for size-resolved P/G partition quotients has a good performance for PCBs. Steady state model could be a better choice for predicting K P values of PCBs.