Construction of QSPR Models on Properties of Amine Compounds Based on Semi-Empirical Quantum Chemical Calculation and Suggestion of New Amine Compounds

Abstract

Carbon dioxide capture and storage (CCS) is widely performed as one of measures against global warming. In the present CCS, however, cost of carbon dioxide capture is high, and therefore it is strongly required to develop more effective absorbing method. Chemical absorption method with amine compounds which have hydroxyl grope is one of the potent methods for carbon dioxide capture. For decreasing the cost of carbon dioxide capture, novel amine compounds with high capability of absorption and diffusion is required. Hence, the objective of this study is to explore such amine compounds. To construct regression models for predicting the absorption rate and the amount of diffusion of amine compounds, we evaluated the amine compounds whose properties are measured, applying semi-empirical molecular orbital method. By using this method, steric effect and electric nature can be considered. After constructing regression models by the GAPLS (genetic algorithm based-partial least squares) method in which choice of predictive variables is carried out, we evaluated amine compounds which are virtually generated with a structure generator. By using their properties estimated with the GAPLS models, we explored novel amine compounds with desirable properties.