Efficient structural models to predict absorption rate and absorption amount for different types of amine

Excessive carbon dioxide (CO₂) emissions are contributing to climate change. Amine solutions are commonly used in industries for the purpose of CO₂ absorption. Nevertheless, conducting experiments on different amines is costly and time-consuming. In this study, the quantitative structure–property relationship methodology was used to develop models for revealing the relationship between different amine structures, including ring-shaped, non-ring-shaped, and the relevant process properties such as absorption rate (AR) and absorption amount (AA). Molecular geometries were optimized using density functional theory (DFT) at levels B3LYP/6-31G (d, p) and semi-empirical PM6 methods. Both methods produced similar results, indicating the independence of some descriptors from the optimization methods. Linear and nonlinear models were developed using genetic algorithm–multi-linear regression and least-squares support vector machines. The correlation coefficient (R²) values for linear and nonlinear models in AR were determined to be 0.87 and 0.98, respectively while it was 0.91 for AA in the linear model. Descriptors analysis showed the hydroxyl group, types of amines, and carbon chain length impact on amine absorption. Ultimately, based on the developed models, it was determined that the placement of atom N between alkyl chains and the hydroxyl group can significantly influence both AR and AA.