Abstract
•QSPR/QSAR model was developed for the prediction of activity coefficient of water in ionic liquids for natural gas dehydration.•COSMO-RS sigma-profiles were used as molecular descriptors•Multiple regression was used to develop QSPR/QSAR model.•The QSPR/QSAR model can be used to predict activity coefficient of water in ILs.•The QSPR/QSAR can provide insight into how to design ILs for natural gas dehydration.
Recently, ionic liquids (ILs) have been considered as alternative solvents to glycol in dehydration of natural gas. However, due to the unlimited structural variations and possible combinations of cations and anions of the ILs, selection of potential ILs for this separation process has been a difficult task. Activity coefficient at infinite dilution is one of the most important thermodynamic properties for preliminary selection of suitable liquid desiccants for water absorption and designing of the natural gas dehydration process. In this paper, COSMO-RS based quantitative structure–property/activity relationship (QSPR/QSAR) models were developed for prediction of activity coefficient of water at infinite dilution in ILs over various temperatures. COSMO-RS based descriptors were generated for 53 ILs (318 data points) at various temperatures. Multiple linear regressions were applied to develop the models. The accuracies of the models were verified by different statistical tests. The model provides a better understanding of the effect of the structural variations of ILs on their affinity for water.
[Display omitted]