Transport properties are essential for the understanding and modeling of electrochemical cells, in particular complex systems like lithium-ion batteries. In this study, we demonstrate how a certain degree of freedom in the choice of variables allows us to efficiently determine a complete set of transport properties. We apply the entropy production invariance condition to different sets of electrolyte variables and obtain a general set of formulas. We demonstrate the application of these formulas to an electrolyte typical for lithium-ion batteries, 1M lithium hexafluoro-phosphate in a 1:1 wt. % mixture of ethylene and diethyl carbonates. While simplifications can be introduced, they provide inadequate predictions of conductivity and transport numbers, and we argue that a full matrix of Onsager coefficients is needed for adequate property predictions. Our findings highlight the importance of a complete set of transport coefficients for accurate modeling of complex electrochemical systems and the need for careful consideration of the choice of variables used to determine these properties.

A systematic description of microscopic mechanisms is necessary to understand mass transport in solid and liquid electrolytes. From Molecular Dynamics (MD) simulations, transport properties can be computed and provide a detailed view of the molecular and ionic motions. In this work, ionic conductivity and transport numbers in electrolyte systems are computed from equilibrium and nonequilibrium MD simulations. Results from the two methods are compared with experimental results, and we discuss the significance of the frame of reference when determining and comparing transport numbers. Two ways of computing ionic conductivity from equilibrium simulations are presented: the Nernst−Einstein approximation or the Onsager coefficients. The Onsager coefficients take ionic correlations into account and are found to be more suitable for concentrated electrolytes. Main features and differences between equilibrium and nonequilibrium simulations are discussed, and some potential anomalies and critical pitfalls of using nonequilibrium molecular dynamics to determine transport properties are highlighted.

No abstract