Assessing the Effect of Explicit Polarizability on Models of Carbon Dioxide Solvation in Ionic Liquids

ZJ Huo and LE Smith and RJ Goudy and K Ndiaye and S Kaiser and ME Nikolov and E Silva and TA Parrack and S Garrett-Roe and CA Daly, ACS OMEGA, 10, 56349-56363 (2025).

DOI: 10.1021/acsomega.5c08258

Ionic liquids are an important possible carbon capture material because of their anomalously high sorption selectivity for carbon dioxide over other gases common in air. Many research groups have investigated the molecular origins of this property and provided important insights, including using 1D and 2D-IR spectroscopy. Molecular dynamics simulations have been indispensable to the interpretation of these experiments. In prior molecular dynamics simulation work, charge-scaled force fields have typically been used to provide a mean-field treatment of effects vital to ionic liquid systems such as charge transfer and polarization. Here, we compare models of carbon dioxide solvated in ionic liquids with explicit polarization to models of the same with implicit polarizability through charge-scaling. We calculate structural, dynamical, and spectroscopic properties, and make comparisons to the same items measured in experiment. In this study, we focus on two ionic liquids: 1-butyl-3-methylimidazolium (BMIM+) paired with bis(trifluoromethane sulfonyl imide) (Tf2N-) and 1-butyl-3-methylimidazolium (BMIM+) paired with hexafluorophosphate (PF6 -). We find that many structural, dynamical, and spectroscopic properties are changed when polarization is modeled explicitly. We also find that explicit polarizability softens local ion cages around the carbon dioxide and that the long-time diffusion of the carbon dioxide is gated by the reorganization of the ionic liquid molecules. Comparisons to experiment show modest improvement of many observables compared with experiment for the explicitly polarizable model over the charge-scaled model. Overall, our results show that charge-scaled force fields are likely sufficient to compute spectroscopic properties of carbon dioxide in ionic liquids and suggest some interpretive rules for understanding their structural and dynamical properties. Those using charge-scaled force fields should generally assume that the ion cages around solutes such as carbon dioxide are too stiff and cation-rich in their models and adjust their interpretations and predictions accordingly.

Return to Publications page