Herein, we explore the transport properties of a prototypical nanoconfined electrolyte, aqueous NaCl in graphene slit pores, using machine learning-based molecular dynamics simulations. The machine learning potential driving these simulations allows us to reach density functional theory-level accuracy at orders of magnitude lower cost than conventional ab initio molecular dynamics.
We find that ionic conductivity decreases as an electrolyte becomes increasingly confined. By decomposing the total solution conductivity into the constituent Onsager transport coefficients, we show how this trend arises from changes in both ion self-diffusion (Nernst-Einstein transport) as well as non-ideal, correlated ion motion. We observe that self-diffusion varies non-monotonically as a function of the degree of confinement due to changes in solution density as well as a shift towards a more vehicular diffusion mechanism. We additionally demonstrate that non-ideal contributions to transport – which are conventionally neglected in both theoretical and experimental studies of these systems – become significantly more pronounced under confinement due to the formation of long-lived ion pairs. Finally, we show that ion transport varies minimally across the height of the slit, challenging the standard assumption that adsorbed ions exhibit low mobility. This work not only resolves key inconsistencies in prior literature but also establishes a detailed mechanistic understanding of confined electrolyte transport which has until now been inaccessible with available simulation methods. Such insights could guide the future design of optimized nanoconfined environments for efficient and selective ion transport.
KD Fong, CP Grey, A Michaelides. "On the physical origins of reduced ionic conductivity in nanoconfined electrolytes." ACS Nano (2025).
KD Fong, B Sumic, N O’Neill, C Schran, CP Grey, A Michaelides. “The interplay of solvation and polarization effects on ion pairing in nanoconfined electrolytes.” Nano Letters 24.16 (2024): 5024-5030.