Determination of octanol–air partition coefficients and supercooled liquid vapor pressures of PAHs as a function of temperature: Application to gas–particle partitioning in an urban atmosphere

• Published in: Atmospheric environment (1994) 2006-11, Vol.40 (34), p.6615-6625
• Main Authors: Odabasi, Mustafa, Cetin, Eylem, Sofuoglu, Aysun
• Format: Article
• Language: English
• Subjects: Applied sciences; Atmospheric pollution; Exact sciences and technology; Gas/particle partitioning; Octanol–air partition coefficient; PAHs; Pollutants physicochemistry study: properties, effects, reactions, transport and distribution; Pollution; Vapor pressure
• ISSN: 1352-2310; 1873-2844
• DOI: 10.1016/j.atmosenv.2006.05.051

Octanol–air partition coefficients ( K OA) for 14 polycyclic aromatic hydrocarbons (PAHs) were determined as a function of temperature using the gas chromatographic retention time method. log K OA values at 25° ranged over six orders of magnitude, between 6.34 (acenaphthylene) and 12.59 (dibenz[ a,h]anthracene). The determined K OA values were within factor of 0.7 (dibenz[ a,h]anthracene) to 15.1 (benz[ a]anthracene) of values calculated as the ratio of octanol–water partition coefficient to dimensionless Henry's law constant. Supercooled liquid vapor pressures ( P L) of 13 PAHs were also determined using the gas chromatographic retention time technique. Activity coefficients in octanol calculated using K OA and P L ranged between 3.2 and 6.2 indicating near-ideal solution behavior. Atmospheric concentrations measured in this study in Izmir, Turkey were used to investigate the partitioning of PAHs between particle and gas-phases. Experimental gas–particle partition coefficients ( K p) were compared to the predictions of K OA absorption and K SA (soot–air partition coefficient) models. Octanol-based absorptive partitioning model predicted lower partition coefficients especially for relatively volatile PAHs. Ratios of measured/modeled partition coefficients ranged between 1.1 and 15.5 (4.5±6.0, average±SD) for K OA model. K SA model predictions were relatively better and measured to modeled ratios ranged between 0.6 and 5.6 (2.3±2.7, average±SD).