Prediction of the lower critical solution temperature of poly(N-isopropylacrylamide-co-methoxy triethyleneglycol acrylate) in aqueous salt solutions using support vector regression

Abstract

This study investigates the lower critical solution temperature (LCST) behavior of a novel thermosensitive NIPA-co-MTGA polymer comprising N-isopropylacrylamide (NIPA) and methoxy triethyleneglycol acrylate (MTGA) by inducing hydrophilic/hydrophobic transitions in an aqueous salt solution. The LCST of the NIPA-co-MTGA polymer was in between that of the NIPA and MTGA homopolymers, and depended on the polymer composition. The LCST of all the polymers decreased with increasing salt concentration. The extent of decrease depended on the type of salt, with the order qualitatively corresponding to the Hofmeister series of anions. The LCSTs of the polymers in aqueous salt solutions were modeled and successfully predicted using support vector regression (SVR), a machine learning methodology. This is the first study to predict the LCST of thermosensitive polymers in aqueous salt solutions using SVR.