New molecular structure based models for estimation of the CO2 solubility in different choline chloride-based deep eutectic solvents (DESs)

In this study, CO2 solubility in different choline chloride-based deep eutectic solvents (DESs) has been investigated using the Quantitative Structure-Property Relationship (QSPR). In this regard, the effect of different structures of the hydrogen bond donor (HBD) in choline chloride (ChCl) based deep eutectic solvents (DESs) has been studied in different temperatures and different molar ratios of ChCl as hydrogen bond acceptor (HBA) to HBD. 12 different datasets with 390 data on the CO2 solubility were chosen from the literature for the model development. Eight predictive models, which contain the pressure and one structural descriptor, have been developed at the fixed temperature (i.e. 293, 303, 313, or 323 K), and the constant molar ratio of ChCl to HBD equal to 1:3 or 1:4. Moreover, two models were also introduced, which considered the effects of pressure, temperature, and HBD structures, simultaneously in the molar ratios equal to 1:3 or 1:4. Two additional datasets were used only for the further external validation of these two models at new temperatures, pressures, and HBD structures. It was identified that CO2 solubility depends on the EEig02d descriptor of HBD. EEig02d is a molecular descriptor derived from the edge adjacency matrix of a molecule that is weighted by dipole moments. This descriptor is also related to the molar volume of the structure. The statistical evaluation of the proposed models for the unfixed and fixed temperature datasets confirmed the validity of the developed models.

Automated summary

The CO2 solubility in different choline chloride-based deep eutectic solvents (DESs) was investigated using the Quantitative Structure-Property Relationship (QSPR). The effect of different structures of the hydrogen bond donor (HBD) in choline chloride (ChCl) based DESs was studied at different temperatures and molar ratios of ChCl as hydrogen bond acceptor (HBA) to HBD. Twelve different datasets with 390 data points on CO2 solubility were chosen for model development.