Experimental methods and prediction with COSMO-RS to determine partition coefficients in complex surfactant systems

Surfactant-based separation processes are a promising alternative to conventional organic solvent processes. A crucial parameter to describe the efficiency of such processes is the partition coefficient between the surfactant aggregates (micelles) and the aqueous bulk phase. In this work, several experimental methods to determine these partition coefficients (micellar liquid chromatography, micellar enhanced ultrafiltration, and cloud point extraction) are evaluated and compared. In addition, these results are compared to predictions with the thermodynamic model COSMO-RS. In particular, systems with the nonionic surfactant TritonX-100 are studied. The partition equilibria of various solutes (pyrene, naphthalene, phenanthrene, phenol, 3-methoxyphenol, and vanillin) and the influence of different additives (alcohols) are investigated. All experimental methods show very good reproducibility. Moreover, the results from different methods are in good agreement, supplementing one another concerning the temperature ranges. Notably, the COSMO-RS model is capable of predicting partition coefficients between micelles and water in the investigated temperature range and at different alcohol concentrations. The results demonstrate the potential of the model COSMO-RS to facilitate the selection of optimized process parameters for a given separation problem. By predicting partition equilibria in multicomponent systems, the selection of surfactant, temperature, and appropriate additives can be facilitated.