Assessment of nano-QSPR models of organic contaminant absorption by carbon nanotubes for ecological impact studies

Abstract

Adsorption of organic contaminants on carbon nanotubes is a critical ecological criterion. Consequently, a predictive model for this endpoint is useful from the point of view of ecological risk assessment. Quantitative Structure–Property Relationships (QSPRs) built by the CORAL software package (http://www.insilico.eu/coral) are used to develop predictive models for adsorption (log K_∞) of organic contaminants by multi-walled carbon nanotubes (MWCNTs). The statistical characteristics of a CORAL model for external validation are: n = 30; r² = 0.8878; and RMSE = 0.475 (mg/g). The probabilistic scheme of the definition of the domain of applicability for the CORAL models are suggested and the practical aspects of using the CORAL software are discussed.