Reliable QSAR for estimating K-oc for persistent organic pollutants: correlation with molecular connectivity indices

Several recent studies have shown that n-octanol/water partition coefficien ts may not be a good predictor for estimating soil sorption coefficients of persistent organic pollutants (POPs), defined here as chemicals with log K -ow greater than 5. Thus, an alternative QSAR model was developed that seem s to provide reliable estimates for the soil sorption coefficients of persi stent organic pollutants. This model is based on a set of calculated molecu lar connectivity indices and evaluated soil sorption data for 18 POPs. The chemical's size and shape, quantified by (1)chi, (3)chi (C) and (4)chi (v)( C) indices, have a dominant effect on the soil sorption process of POPs, Th e developed QSAR model was rationalized in terms of potential hydrophobic i nteractions between persistent organic pollutants and soil organic matrix. Its high predictive power has been verified by an extensive internal and ex ternal validation procedure. (C) 2001 Elsevier Science Ltd. All rights rese rved.