Molecular investigation of amine performance in the carbon capture process: Least squares support vector machine approach

• Published in: The Korean journal of chemical engineering Vol. 37; no. 1; pp. 72 - 79
• Main Authors: Rezaei, Bijan, Riahi, Siavash, Gorji, Ali Ebrahimpoor
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.01.2020; Springer Nature B.V; 한국화학공학회
• Subjects: Amines; Biotechnology; Carbon sequestration; Catalysis; Chemistry; Chemistry and Materials Science; Correlation coefficients; Criteria; Density functional theory; Energy costs; Environmental Engineering; Industrial Chemistry/Chemical Engineering; Materials Science; Mathematical models; Molecular structure; Optimization; Parameters; Reaction mechanisms; Regeneration; Statistical tests; Support vector machines; 화학공학; QSPR; Amine; LS-SVM; Carbon Capture; Cyclic Capacity
• ISSN: 0256-1115; 1975-7220
• DOI: 10.1007/s11814-019-0408-6

The growing threat of global warming has raised more attention towards carbon capture. Current amine plants used for carbon removal suffer from great costs inflicted by high energy demand of the solvent regeneration step. Recently, looking for amines with proper performance in reduced temperatures has been the subject of many researches. Clearly, conducting these researches without any criterion and based only on trial and error wastes large amounts of money and time; thus, it is highly needed that the effect of different amine structural parameters be studied on the amine’s cyclic capacity. Quantitative structure property relationship (QSPR) provides an effective method for predicting amines capacity for CO 2 absorption. In this work, density functional theory (DFT) was employed for optimization of the molecular geometries, and linear and nonlinear models based on parameters related to the molecular structure are presented. The value of the square of the correlation coefficient (R 2 ) for the MLR and SVM models are 0.894 and 0.973, respectively. Developed models can be used as a criterion for amine selection. Reliability and high predictability of the models are confirmed based on statistical tests. Moreover, mechanistic interpretation of models for better understanding of the reaction mechanism of carbon capture was discussed.