Experimental determination of partial solubility parameters in deep eutectic solvents.

Abstract

This study compares the solubility predictions of the Flory-Huggins-HSP (FH-HSP) model with those of openCOSMO-RS and PC-SAFT, for evaluating the screening capabilities of the Hansen solubility parameters (HSP) using the relative energy difference (RED) parameter. Two deep eutectic solvents, thymol + L-menthol (TM) and thymol + cyclohexanone (TC), and their precursors were investigated. The three models performed reasonably well in predicting the solubility of thymol and L-menthol in organic solvents but showed significant deviations in predicting solid solubility in cyclohexanone. For solubility predictions in DES compared to available experimental data, the root- mean-square logarithmic deviation (RMSLD) values obtained with FH-HSP, openCOSMO-RS, and PC-SAFTwere 0.86, 0.27, and 0.93, respectively. Using FH-HSP with HSP obtained from an ideal mixing rule of the parameters of pure compounds resulted in a deviation of 1.13. The screening assessment of the studied DES revealed a 70.4% match rate between the RED parameter and ln𝛾∞ 𝑖, with 6.9% false negatives. Using HSP obtained fromthe ideal mixing rule resulted in similar scores, with a 68.5% match rate and 10.4% false negatives. The results suggest that the HSP theory can aid DES screening processes by identifying promising DES-solute pairs, thus allowing the targeting of more robust methods of computational resources in smaller sets of compounds.