Comparison of the octanol–air partition coefficient and liquid-phase vapor pressure as descriptors for particle/gas partitioning using laboratory and field data for PCBs and PCNs

The conventional Junge–Pankow adsorption model uses the sub-cooled liquid vapor pressure (pLo) as a correlation parameter for gas/particle interactions. An alternative is the octanol–air partition coefficient (Koa) absorption model. Log–log plots of the particle–gas partition coefficient versus pLo were previously made for partitioning data from controlled laboratory studies, resulting in separate trend lines for different ortho-substituted PCB classes. The same plots applied to field data for PCBs and PCNs resulted in separate regression lines with slopes that were statistically different at the 99% confidence level. When Koa is used as the correlation parameter, these differences are resolved showing the ability of the model to reduce variability both within a compound class and between compound classes. The Koa model is also preferred because it uses parameters that can be measured directly (Koa and fom), unlike the parameters of the Junge–Pankow model which must be estimated.