Occurrence and Distribution Processes of Semivolatile Organic Chemicals in the Atmosphere and Leaves

Abstract

In this study the occurrence and behaviour of semivolatile organic chemicals (SOC5) in the atmospheric environment of Brisbane, Australia are examined. The studied compounds include polycyclic aromatic hydrocarbons (PANs), tetra- to octachiorinated dibenzodioxins and dibenzofiirans (PCDD/Fs), a selection of tn- to octachiorinated biphenyls (PCBs) and hexachlorobenzene (HCB). The characteristics and occurrence of these compounds are outlined and theories and models for vapour/particle and lea±7atmosphere distribution processes are reviewed. An air sampling system developed for sampling SOCs was modified and tested for a series of sampling artefacts. Air samples were collected from seven sites in urban Brisbane. PANs, PCDDIFs and PCBs were detectable in all air and leaf samples. Concentrations and compound profiles were evaluated with respect to the dominant sources. The results show that the main sources for PANs in the air at the seven sites are motor vehicle emissions. For PCDDIFs the homologue profile indicates that a distant source may have significantly contributed to the atmospheric concentrations of these compounds. In a dry atmosphere, SOCs are distributed between the vapour and the particle associated phase. Vapour/particle distribution for SOCs from a given compound group can be correlated with different molecular and physical-chemical descriptors. The vapour/particle distribution results are examined using both an adsorption and absorption based model. The temperature sensitivity of the vapour/particle distribution can be explained using either model. The adsorption model, based on compound subcooled liquid vapour pressure, has a higher predictive capacity. However, a m~or drawback of the adsorption model is that it discriminates between compounds from different compound groups (i.e. PAHs and PCDD/Fs). The absorption model, based on compound KOA values, shows less difference between compounds from different compound groups. Its lower predictive capacity may be the result of more unreliable input parameters. In order to assess leaves as biomonitors, the leaf/atmosphere distribution process of SOCs was studied using predominantly leaves from Melaleuca, an Australian native tree commonly planted along road sides in Brisbane. A comparison of experimentally determined leaf/vapour concentration ratios (ng m3 leaf / ng ni3 air) with calculated leaf/vapour equilibrium bioconcentration factors indicates that PARs with more than 3 rings and PCDD/Fs do not reach equilibrium with the vapour phase in the atmosphere. A kinetic model is applied to identify the main kinetic limitation. The results suggest that for compounds which do not reach equilibrium, the air side resistance is limiting the process. The air side resistance is influenced by a myriad of factors. For predictive purposes it was assumed that the leaf/atmosphere system had reached a steady state situation. This allows the calculation of an empirical loss rate constant which is added to the loss rate constant based on volatiisation. Employing this empirical rate constants, relatively good predictions of leaf concentrations, based on measured atmospheric concentrations, were obtained in a dynamic application of a kinetic leaf/atmosphere distribution model. Additionally, a passive sampling device was developed consisting of fibre glass material coated with tristearin. It seems that SOCs with K0A values larger than lO~ did not reach equilibrium and the kinetics for these compounds is again controlled by the air side resistance. Initial experiments with this sampler show that it has potential for gaining a better insight into the processes which govern the distribution of SOCs between leaves and the atmosphere. Predicting atmospheric concentrations using this device requires further tests and modifications of the sampler.