ABSTRACT
40Semivolatile compounds become associated with aerosols by being occluded in the particle during formation, adsorbing to active sites on the surface and absorbing into a liquid-like film on the particle. Experimental estimates of the particle/gas distribution are often made with a high volume air sampler (hi-vol), which employs a filter to retain particulate matter and a sorbent trap for gaseous compounds. Such estimates may be biased by sampling artifacts but are nonetheless useful for gauging the relative particle/gas distribution for different classes of compounds. This chapter reviews the theoretical models for “exchangeable” organic compounds which are adsorbed or absorbed to aerosols and compares theoretical and hi-vol estimates of phase distribution for organochlorine compounds (OCs) and polycyclic aromatic hydrocarbons (PAHs). Hi-vol measurements of the particulate fraction for PAHs in urban and rural air are in fairly good agreement with predictions of the Junge —Pankow adsorption model. In the case of OCs, the hi-vol sampler underestimates (or the model overestimates) the particulate fraction. Results are presented for the speciation of polychlorinated biphenyls (PCBs) in urban air. The most toxic PCBs are those which have no orthosubstituted chlorines (”coplanar” PCBs) and some compounds having only one ortho-chlorine. These non- and mono-ortho PCBs are preferentially found in the particulate fraction during hi-vol sampling. Part of the reason for this is that mono- and non-ortho PCBs have lower vapor pressures than multi-ortho congeners in the same homolog group (”primary” ortho-effect). However correlations of the particle/gas partition coefficient (K p) with liquid-phase vapor pressure (p°L) show that sorption of the toxic PCBs to urban aerosols is enhanced over expectations from lower vapor pressure alone. This “secondary” ortho-effect can be accounted for by the shapes of the PCB molecules, which are more nearly planar for mono- and (especially) non-ortho congeners. The octanol/air partition coefficient (Koa) is suggested as an alternative to p°L for describing absorption of semivolatile compounds to aerosols. Values of Koa for PCBs also show a “secondary” ortho-effect: mono- and non-ortho congeners partition into octanol to a greater extent than multi-ortho compounds of the same volatility. Sorption of PCBs to urban air particulate matter was measured in the laboratory and values of Kp were correlated with p°L or K oa. In the former plot, K p of the PCB congeners increased in the order: multi-ortho < mono-ortho < non-ortho. These class differences were largely resolved when K oa was used as a correlation parameter.
