Prediction of the sorption coefficient for the adsorption of PAHs on MWCNT based on hybrid QSPR-molecular docking approach

In this study, quantitative structure-property relationship (QSPR) methodology and molecular docking employed for modeling of the sorption coefficient (log KCNT) of 13 polycyclic aromatic hydrocarbons (PAHs) on multiwall carbon nanotube (MWCNT) adsorbent. A molecular docking simulation used to present a reliable and accurate QSPR model with defining the distance and best orientation of PAHs structures on nano-adsorbent. A genetic algorithm-multiple linear regression (GA-MLR) method was employed for implementation of QSPR model. In this model, the square of correlation coefficients (R2) was 0.945 and 0.890, and the root mean square errors (RMSE) were 0.07 and 0.14 for the training and test sets, respectively. Also, inspection to selected descriptors indicates the electrostatic and steric parameters of PAHs are the predominant factors responsible on the log KCNT values. These results can be used for prediction of sorption coefficient of other PAHs by MWCNT and modify the surface of the adsorbent for improving the log KCNT.