New QSPR Models to Predict the Critical Micelle Concentration of Sugar-Based Surfactants

Abstract

Sugar-based surfactants represent a fruitful field of research in the context of sustainable chemistry since they can be obtained from renewable resources. In this work, new quantitative structure property relationships (QSPR) models for the critical micelle concentration (CMC) dedicated to sugar-based surfactants are proposed in order to reduce testing in a screening perspective. An important literature compilation allowed the constitution of a data set of 83 sugar-based surfactants for which accurate CMC values were found. Then, a series of QSPR models were developed based on molecular descriptors of the whole molecule and of the hydrophobic and hydrophilic fragments taken separately. Different models were considered by including quantum-chemical descriptors with hope to access physically based models and by using only simple constitutional descriptors to favor fast and easy prediction. The best QSPR model was obtained including quantum-chemical descriptors of the whole molecular structure with a root-mean-square error (RMSE) of 0.32 (log) evaluated on a validation set of 27 molecules. A simpler model with good performances was also found (with a RMSE of 0.36 (log) on the validation set), including only constitutional-based fragment descriptors, that can be easily computed from the 2-dimension structure of the hydrophilic and hydrophobic fragments.