Assessing the importance of ab- and adsorption to the gas-particle partitioning of PCDD/Fs

Abstract

The gas-particle partitioning of polychlorinated dibenzo- p-dioxins and dibenzofurans (PCDD/Fs) was examined (i) by re-interpreting results from controlled field experiments and (ii) in diurnal samples taken in the New Jersey (NJ), New York (NY) City region. In the controlled field experiments, aerosol-laden filters were exposed to elevated concentrations of PCDD/Fs. Gas-particle partitioning coefficients, K p's, were significantly correlated with octanol–air partitioning coefficients, K oa's. The regression of all individual datapoints resulted in the following best fit ( r 2=0.74, n=87): log K p,meas=1.00(±0.13) log(10 −12 f om K oa/ ρ oct)−0.15(±0.48). We interpret this as showing that the ability of organic matter (OM) to absorb PCDD/Fs is generally well described by the octanol–air partitioning model ( f om K oa). At the NJ land-based sites, samples were taken and analyzed for organic carbon (OC) and elemental carbon (EC), gaseous and particulate PCDD/Fs. K p's were significantly correlated with the f om K oa approach. Adsorption to the filter and possibly to black carbon (BC), which was estimated based on EC measurements, could have contributed to the observed K p values. Gas-particle predictions based on BC adsorption and OM absorption, with K p= f om K oa/(10 12 ρ oct)+ f BC K BC–air/(10 12 ρ BC) resulted in K p predictions that were close to measured values. Adsorption to the filter might have been the major reason for elevated K p's: The one NJ site with highest K p's was most prone to the filter adsorption sampling artefact because of relatively low sampling volumes and concentrations of total suspended particulates. In addition, while adsorption to BC would result in better prediction of K p values, no influence of f BC or f BC/ f om ratios was seen, suggesting that it was of lesser importance in our sample set.