Solvent strength effects on the anti-solvent ability of carbon dioxide with ionic liquid/organic mixtures

ILs show potential for different types of applications, yet removing compounds such as reaction products and impurities can be difficult. This work describes the solvent strength effects on the anti-solvent behavior of CO2 with ILs and IL/organic mixtures. The anti-solvent behavior is measured for both liquid and solid solutes in ILs. A systematic study was done to highlight the effect of specific interactions.;For liquid solutes, the anti-solvent ability is measured in terms of the Lower Critical End Point (LCEP). For all the IL/organic mixtures measured, the LCEP increases as the concentration of IL increases above ∼10 mole%. The ease of separation depends on the CO2 solubility in the pure organic; 2-butanone is the easiest to separate and has the highest CO 2 solubility. Another measure of the anti-solvent ability of CO 2 is the nucleation pressure of solid [Cu(acac)(tmen)][BPh4] dissolved in IL/organic mixtures. It is not possible to separate [Cu(acac)(tmen)][BPh 4] from [hmim][Tf2N] using CO2; a cosolvent is needed. For organic/[Cu(acac)(tmen)][BPh4] mixtures, the nucleation pressure depends on specific interactions and solute solubility. The presence of IL is significant, causing an increase in the nucleation pressure.;The results for the systematic study of the solvent strength of IL/organic mixtures indicate that ILs in general preferentially solvate polar, protic compounds. The solvent strength of ILs is unaffected by CO2, even at large mole fractions. However, the organics are affected. The IL/organic mixtures with CO2 behave intermediary between that of IL/CO 2 and organic/CO2. The solvent strength behavior can be used to explain the antisolvent ability of CO2 for liquid organic solutes and the solid [Cu(acac)(tmen)][BPh4] solute.;The solvent strength behavior can also be used to predict the anti-solvent behavior in terms of the LCEP. A correlation based on the Linear Solvation Energy Relationship (LSER) equation modified to take into account two components was used. The fit is excellent, and can also be used to predict data of other mixtures within 10% for ILs other than [bmim][PF6]. This is excellent considering that only pure component parameters for pi*, alpha, and beta were used.