Technical Report 123


Based on activity undertaken as part of this task force it is suggested that multimedia models based on the activity approach provide a promising framework for ERA of ionisable organic compounds with a down-the-drain emission scenario. In the interim there is a need for improved regressions implemented into existing tools for describing the partitioning behaviour of ionisable organic compounds to solids, such as described above, an observation that is consistent with the experiences of the agrochemicals industry. Whereas it is suggested that preliminary activity along these lines can lead to low-tiered improvements in model parameterisation with respect to handling ionisable organic compounds, there remains the need for additional data aimed at improving mechanistic understanding of processes that may be influencing environmental fate and behaviour. For instance, there is a need for additional work on the sorption of cationic materials to sludge, and improved understanding needed for extrapolating data between sludge, soil, and sediment, which will enable the derivation of more environmentally relevant PECs.

As a general observation it is possible to define a generic model environment for use in lower-tiered predictive risk assessment, based on work described by Franco and Trapp (2010). As an initial screen, Figures 13 and 14 can be used to assess the relative influence of humic acids on the bioavailability of a chemical substance, information that could then be used in the design of ecotoxicological testing. The environmental exposure of organisms and humans to chemicals is implemented in the EU tool for the chemical safety assessment (EUSES). The fate of a chemical released to the environment is predicted via a regional Level III multimedia model (SimpleBox), which is linked to output from SimpleTreat described above. Activity related to this task force has helped initiate an update to the SimpleTreat model that will allow for better handling of ionisable organic compounds (Franco et al. 2013). The updated version of SimpleTreat should be used to derive the concentration in effluent (Clocaleff), which can then be used to provide an improved estimate of local concentration in surface water (Clocalwater). In the EU TGD Clocalwater is derived as:


Where SUSPwater is the concentration of suspended matter in the river, DILUTION is the dilution factor, and Kpsusp is the solids-water partitioning coefficient of suspended matter, and is estimated from the KOC of the substance. As discussed above, caution should be taken in the derivation of Kpsusp. Consequently, improving estimates of partitioning with solids for ionisable organic compounds is recommended as a critical component towards improved projections of bioavailable concentrations in surface waters.

Where refinement of the risk assessment is required, it is suggested that current understanding of mechanistic interactions involving the ionised species of an organic chemical with various environmental matrices are currently not sufficient to provide a definitive ERA.

Thus, in addressing the question defined in the terms of reference for this task force, i.e.: Can recommendations be made regarding what pH and soil properties are most useful for accurately predicting environmental concentrations, i.e. is it possible to define a generic model environment for use in a predictive risk assessment? Is it possible to advise the most appropriate testing conditions? The response would be that it is possible at low-tier assessments, using refinements to existing tools, such as demonstrated by equation 1, or utilising the activity approach to multimedia modelling described by Franco and Trapp (2010). Additionally, empirical studies aimed at assessing the partitioning behaviour of ionisable organic compounds, such as APIs, should be conducted with various different types of sludge, at varying pH, analogous to the approach adopted in the agrochemicals industry. Furthermore, there is a need for improved understanding regarding the extrapolation of data obtained from soil sorption studies towards estimating sorption to sludge or to sediment, as these matrices may have properties that are very different from soils. Again, this is analogous to the approach taken by the agrochemicals industry, whereby sorption behaviour between different types of soils is conducted, and attempts are made towards an improved understanding regarding the mechanisms that are most important for influencing sorption.

Clearly, improved mechanistic understanding is essential to support high-tiered assessments. It is suggested, for instance, that the key factor influencing bioavailability, and therefore the PEC, is sorption to different environmental matrices. Based on the current paradigm this is typically based on measurements or estimates of KOC, which assume that hydrophobic interactions are the dominant sorption process. As illustrated above in Figure 16 (Section 4.2), relationships based on KOC regressed against DOW for ionisable organic compounds, are sufficient to address interactions dominated by hydrophobic interactions, but the correlation is not entirely satisfactory. Thus, there is a need to assess how other parameters might be used to improve the understanding of sorption. For instance, the charge surface area of the molecule might help to differentiate the relative importance of the hydrophobic component of the molecule versus the charge; a complex molecule with significant branching may provide steric hindrance to the charge, resulting in greater hydrophobic behaviour being possible. Whereas the sorption behaviour of a relatively simple molecule with one or more charged functional groups not shielded by hydrophobic chains may likely be more strongly influenced by electronic interactions. Thus introduction of additional parameters such as charged surface area in combination with DOW and the cationic exchange capacity of the sorbent may lead to improved PEC estimates. Additional research directed at improving the ability to predict sorption of ionisable organic compounds to various environmental matrices is thus needed.