Understanding correlation effects for ion conduction in polymer electrolytes

TL;DR

This paper investigates ion correlation effects in polymer electrolytes using all-atom simulations, develops an extended phenomenological model to describe ionic dynamics, and highlights the need for additional information beyond ion-pair fractions.

Contribution

The study introduces an extended SO model incorporating simulation-derived parameters to better describe ion correlations and dynamics in polymer electrolytes.

Findings

Correlation effects are linked to transient ionic aggregates.

Parameters for the extended model deviate slightly from unity, indicating additional factors.

The model can estimate free and non-free ion dynamics from diffusivity data.

Abstract

Polymer electrolytes typically exhibit diminished ionic conductivity due to the presence of correlation effects between the cations and anions. Microscopically, transient ionic aggregates, e.g. {\it ion-pairs}, {\it ion-triplets} or higher order ionic clusters, engender ionic correlations. Employing {\it all-atom} simulation of a model polymer electrolyte comprising of poly(ethylene oxide) and lithium iodide, the ionic correlations are explored through construction of elementary functions between pairs of the ionic species that qualitatively explains the spatio-temporal nature of these correlations. Furthermore, commencing from the exact Einstein-like equation describing the collective diffusivity of the ions in terms of the average diffusivity of the ions (i.e. the self terms) and the correlations from distinct pairs of ions, several phenomenological parameters are introduced to keep track of the simplification procedure that finally boils down to the recently proposed phenomenological model by Stolwijk-Obeidi (SO) [N. A. Stolwijk and S. Obeidi, Phys. Rev. Lett. 93, 125901, 2004]. The approximation parameters, which can be retrieved from simulations, point to the necessity of additional information in order to fully describe the correlation effects apart from merely the fraction of ion-pairs which apparently accounts for the correlations originating from only the nearest neighbor structural correlations. These parameters are close to but not exactly unity as assumed in the SO model. Finally, as an application of the extended SO model one is able to estimate the dynamics of the free and non-free ions as well as their fractions from the knowledge of the single particle diffusivities and the collective diffusivity of the ions.