Computational insights into quinoxaline-based corrosion inhibitors of steel in HCl: Quantum chemical analysis and QSPR-ANN studies

Abstract

The inhibition of mild steel deterioration via organic substances has become popular nowadays. Among the myriads of organic substances applied as potential inhibitors, quinoxalines stand out as toxic-free, cheap and effective compounds in different electrolytes. This report investigates the computational aspects of selected quinoxaline compounds tested as suppressors of mild steel deterioration in HCl medium using quantum chemical method (Density Functional Theory, DFT) and quantitative structure property relationship (QSPR). Feature selection tool was utilized to choose five top molecular descriptors (constitutional indices) that were used to characterize the quinoxaline molecules. Linear (ordinary least squares regression) and nonlinear (artificial neural network) modelling were adopted to correlate the selected constitutional indices of the studied quinoxalines with their experimental inhibition performances. The nonlinear model showed better performance as shown by the obtained results; RMSE of 5.4160, MSE of 29.3336, MAD of 2.3816 and MAPE of 5.0389. The developed models were utilized to determine the inhibition performances of ten new quinoxaline-based corrosion inhibitors which showed excellent inhibition performances of 87.88 to 95.73%.

Keywords

• Corrosion inhibitors
• Quinoxalines
• Molecular descriptors
• Quantitative structure property relationship
• Ordinary least squares regression
• Artificial neural network