Prediction of the adsorption capability onto activated carbon of liquid aliphatic alcohols using molecular fragments method

Quantitative structure-property relationship (QSPR) for estimating the adsorption of aliphatic alcohols onto activated carbon were developed using substructural molecular fragments (SMF) method. The adsorption capacity of activated carbon (gr/100grC) for 150 aliphatic alcohols onto activated carbon (AC) is studied under equilibrium conditions. Forward and backwards stepwise regression variable selection and multilinear regression (MLR) are combined to describe the effect of molecular structure on the adsorption capacity of activated carbon according to the QSPR method. To quantitatively relate adsorption capacity (Qe) with the molecular structure MLR analysis is performed on the set of 15 substructural molecular fragments (SMF) provided by the software ISIDA. The five fragments selected by variable subset selection, all belonging to the subfragments, adequately represent the structural factors influencing the affinity of alcohols to AC in the adsorption process. Finally, a QSPR model is selected based on leave-one-out cross-validation and its prediction ability is further tested on 30 representative compounds excluded from model calibration. The prediction results from the MLR models are in good agreement with the experimental values. By applying MLR method we can predict the test set (30 compounds) with squared cross validated correlation coefficient (Q2ext) of 0.9538 and root mean square error (RMSE) of 2.0832.