Quantifying and Interpreting Solvation Power of Cyclic Carbonate by Chemical Calculation and Machine Learning

Ion dynamics in carbonate electrolytes are fundamentally governed by the solvation power of cyclic solvents, a property whose quantification remains elusive because of the intricate competition between electronic and steric effects. We decouple these influences by defining two core descriptors: (i) the nature of the functional groups furnishing solvation sites, encompassing coordinating atom charges and electron localization function (ELF) values, and (ii) structural adaptability, derived from the substituent volume and its distance from the coordination site. Building upon this framework, the quantitative correlations between these descriptors and solvation power, along with their underlying mechanisms, are investigated through an integrated approach of mathematical fitting and machine learning (ML). Notably, functional group properties and structural compatibility comparably contribute to solvation power, challenging the conventional understanding that functional group attributes predominantly dictate the solvation behavior. This work provides a chemical foundation for the rational selection of cyclic carbonate-based electrolytes for battery chemistry.