The discussion of descriptors for the QSAR model and molecular dynamics simulation of benzimidazole derivatives as corrosion inhibitors

The corrosion inhibition performances of 20 protonated benzimidazole derivatives were studied using theoretical methods. Nuclear Independent Chemical Shift (NICS), the measurement of aromaticity, demonstrated good correlation with inhibition efficiencies and was used as a descriptor. Binding energies were calculated on the basis of molecular dynamics simulations using a realistic corrosive environment. Some improved descriptors correlate well with experimental inhibition efficiencies. A reliable nonlinear quantitative structure-activity relationship model was constructed by a support vector machine approach. The correlation coefficient and root-mean-square error were 0.96 and 6.79%, respectively. Additionally, six new benzimidazole molecules were designed, and their inhibition efficiencies were predicted.