Conventional lithium-ion batteries, and many next-generation technologies, rely on organic electrolytes with multiple solvents to achieve the desired physicochemical and interfacial properties. The complex interplay between these physicochemical and interfacial properties can often be correlated with the coordination environment of the active cation. We develop a theory for the coordination shell of cations in non-aqueous solvent mixtures¹. Our theory can naturally explain atomistic classical molecular dynamics (MD) simulations of cation solvation in a conventional non-aqueous electrolyte. Moreover, our theory provides a systematic way to report solvation structures, which we hope will bring more transparency to the field. While classical atomistic MD simulations can provide a cheap way to investigate such systems with a large number of atoms, there is often the question of the accuracy and transferability of the force field. Ab initio MD is often a more accurate method, but there are significant limitations to the time and length scales that can be investigated. With recent advances in equivariant neural network interatomic potentials[2,3], there is the prospect to combine the accuracy of DFT with system sizes and time scales of classical MD simulations. I will summarise our progress on learning interatomic potentials for concentrated electrolytes of interest in batteries and supercapacitors.

1. ZAH Goodwin et al. arXiv:2301.07839
2. S Batzner et al. Nat Commun 13 2453 (2022)
3. A Musaelian et al. arXiv:2204.05249