A group contribution model for prediction of the viscosity with temperature dependency for fluorine-containing ionic liquids

• A large and comprehensive dataset was used to develop a model.
• A linear GC model was developed to predict the viscosity of fluorine-containing ionic liquids.
• The model can estimate the reduction in viscosity by insertion of fluorine atom into the anion structure.
• New ionic liquids can be designed for specific range of viscosity.

One of the barriers against application of ionic liquids (ILs) for industrial separations is their relatively high viscosity. The viscosity of the IL can however be decreased by inserting the fluorinated moieties into the anion structure. Hence, if one could find a relationship between chemical structure, temperature, and viscosity one could design fluorine-containing ILs (F-ILs) with desired specifications. In this communication, two extensive dataset of F-ILs were gathered; one with reliable sources consisting of 863 experimental viscosities for 85 F-ILs, and the other including unreliable sources consisting of 1160 experimental data points for 332 F-ILs. Using the reliable dataset an efficient group contribution (GC) model was developed, which employs a total of 35 functional groups to estimate the temperature-dependent viscosities of F-ILs. The proposed GC model produced a low average absolute relative deviation (AARD) of 3.23%, root-mean-square error (RMSE) of 0.19, and coefficient of determination (R2) of 0.977. The other model developed using the second database only parameterized in supplementary materials.

The functional groups of cations and anions were calculated from their corresponding chemical structure. Thereafter, the functional groups as well as the experimental data of viscosity were gathered into a dataset. Finally a 35-parameter group contribution model was developed.Figure optionsDownload as PowerPoint slide