120th General Meeting of the KCS

Type Poster Presentation
Area Polymer Chemistry
Room No. Exhibition Hall 2+3
Time 10월 20일 (금요일) 13:00~14:30
Code POLY.P-30
Subject QSPR modeling for solubility parameter of polymers
Authors KYUSUNG LEE, Byeong Hun Lee, Yong Seok Kim1, Sung Kwang Lee*
Department of Chemistry, Hannam University, Korea
1Center for Chemical Materias, Korea Research Institute of Chemical Technology, Korea
Abstract The solubility parameter is a characteristic of a liquid which is a measure of component mixing. The solubility parameter applies the principle of "like dissolve like". In other words, if the interactions between substances are similar, they are similar in properties to each other and can dissolve well, so this is the best way to compare the solubility parameters. The solubility parameter concept based on regular solution theory has been extensively used in practical applications of polymers in various solvents and solvent mixtures. It has proven to be a valuable empirical tool in the paint and lacquer industry and in the prediction of adhesive interactions. A polymer is composed of long chains of regular repeating units due to the chemical reaction of monomers. Polymer materials are used in variety of area such as aerospace, energy and IT industry. Several physicochemical properties are used to characterize these polymers. In this study, we have developed QSPR(quantitative structure-property relationship) models for predicting solubility parameter of polymer from 2D chemical structure. Multiple linear regression and support vector machines were used to learn the model using descriptors related to solubility parameters. And y-scrambling and cross validation were used to verify the stability and predictability of the model. Finally, applicability domain of models could be used to specify the scope of the models. These models show that they can be used to predict the solubility parameter of a wide range of polymers.
E-mail lyk3766@naver.com