Quantitative Structure-Property Relationship Studies on the Adsorption of Aromatic Contaminants by Carbon Nanotubes

Liu Fen,Zou Jian-Wei,Hu Gui-Xiang,Jiang Yong-Jun

DOI: https://doi.org/10.3866/pku.whxb201406182

2014-01-01

Abstract:Ab initio calculations have been performed for a group of 59 aromatic compounds at the HF/6-31G* level of theory. Electrostatic potentials (ESPs) and the statistically based structural descriptors derived from ESPs on the molecular surface have been obtained. The linear relationships between the adsorption equilibrium constants of organic contaminants by carbon nanotubes and the theoretical descriptors have been established by multiple linear regression. It is shown that the quantities derived from electrostatic potentials, V-min, sigma(2)(+) and Sigma V-Ind(+) together with the molecular surface area (S) and the energy level of lowest occupied molecular orbital (epsilon(LUMO)) can be used to express the quantitative structure-property relationship (QSPR) of this sample set. All of the descriptors introduced in the QSPR models have definite physical meanings and their reasonability can be explained in terms of intermolecular interactions between the aromatic pollutants and carbon nanotubes or water. The stabilities and predictive powers of the models have been validated by "leave-one-out" and Monte Carlo cross-validation methods. Three nonlinear modeling techniques, namely supported vector machine (SVM), least-square supported vector machine (LSSVM), as well as Gaussian process (GP), have also been used to construct the predictive models. Though the SVM and LSSVM models exhibit strong fitting abilities, their predictive powers are inferior to the other models tested. The GP model yields the best fit and predictive ability among all of the models. Its advantage over the linear model, however, is not as remarkable as expected, which means that the relationship between the molecular structure and the adsorption property for the present system is primarily linear.