Seasonal variation and gas/particle partitioning of PCBs in air from central urban area of an industrial base and coastal city –Tianjin, China

Abstract

The 26 air samples were collected from Tianjin urban areas from July 2010 to May 2011 for the analysis of 86 PCB congeners to investigate the gas/particle partitioning of PCBs. The total PCBs concentration (in pg/m 3) ranged from 31.28 to 650.33 (mean: 273.77) and showed a significantly seasonal variation as summer > autumn > spring > winter. Gas PCBs accounted for 80.79% of total PCBs. The fraction of each gas PCB homolog increased with the decreasing chlorine atoms. The contribution of gas PCBs to total PCBs showed a strong seasonal variation with summer (94.07%) > autumn (78.91%) > spring (78.19%) > winter (56.02%). Tri-, tetra- and penta-PCBs were the predominantly homologs in Tianjin atmosphere. The natural logarithm of partial pressure of PCBs were well correlated with the inverse absolute environmental temperature (R = 0.79). The steep slope of –4933.64 obtained from Clausius–Clapeyron (C.C) plot indicated the influence of local PCBs sources. In generally, the C.C slopes and the surface-air exchange enthalpy (ΔH SA) for different PCB homologs increased with the increasing chlorine atoms. The fraction of particle bound PCBs were underestimated for PCBs with P L 0 10 –3.85 Pa and overestimated for PCB congeners with P L 0 < 10<sup>–3.85 Pa by Junge-Pankow model. The logarithm of particle/gas partitioning coefficients (log KP) of PCBs were well correlated with the logarithm of respective sub-cooled vapor pressures (log P L 0) (R = 0.72) and the octanol-air partition coefficient (log K OA) (R = 0.77). The slopes for log K P against log P L 0 and log K OA plots were –0.44 and 0.48, deviated significantly from the expected –1 and +1 for the equilibrium state of gas/particle partitioning, possibly explained by the emission of PCBs and particles from the local industries and other sources. The TEQ concentrations (fg/m 3) for 12 dioxin-like PCB congeners ranged from 35.628 to 612.796 with the mean value as 144.878.