Combining Molecular Dynamics and Experimental Methods for the Parametrization of Binary Carbonate-Based Electrolytes

Abstract

Modelling the ionic transport in battery cells requires precise parametrization of the involved electrolytes. For carbonate-based electrolytes, however, the evaluation of their parameters suffers from interphase effects between the bulk electrolyte and the Li metal electrode, commonly present in the usual electrochemical polarization experiments. In this work, we combine measurements on conductivity and concentration cells with molecular dynamic simulations, avoiding these difficulties and thus, allowing for a more accurate determination of the parameters. We determine the conductivity, the transference number, the thermodynamic factor and the salt diffusion coefficient for three different electrolytes, i.e mixtures of ethylene carbonate (EC), ethyl methyl carbonate (EMC), methyl propionate (MP), dimethyl carbonate (DMC) and propylene carbonate (PC), containing LiPF$_6$ at various concentrations and temperatures. In order to validate the simulated transference numbers, we employ electrophoretic Nuclear Magnetic Resonance spectroscopy (eNMR).

Figures (7)

• Figure 1: Conductivities $\kappa$ obtained by MD simulations and electrochemical measurements, together with their corresponding fit functions (solid and dashed lines). The experimental conductivities for LiPF$_6$ in EC:EMC and in EC:DMC:PC were taken from Landesfeind et al.Landesfeind2019 and Valø en et al.Valoen2005
• Figure 2: Transference numbers $t^\mathrm{VOL}_+$ in the VOL reference frame obtained by MD simulations, electrochemical experiments and eNMR measurements, together with their corresponding fit functions (solid and dashed lines). The transference numbers of LiPF$_6$ in EC:EMC and in EC:DMC:PC were taken from Landesfeind et al.Landesfeind2019 and Valø en et al.Valoen2005
• Figure 3: Thermodynamic factors $\mathit{TDF}^\mathrm{VOL}$ in the VOL reference frame determined by combining concentration cell measurements with MD simulations or polarization experimentsLandesfeind2019Valoen2005. The solid and dashed lines show their corresponding fit functions. Note that we multiply the literature data by a factor of 2, as mentioned in the main text.
• Figure 4: Diffusion coefficients $D^\mathrm{VOL}_\pm$ in the VOL reference frame calculated with MD simulations, the Advanced Electrolyte Model (AEM)Logan2019Gering2006Gering2017 or electrochemically measuredLandesfeind2019Valoen2005 together with their corresponding fit functions. Note that the AEM calculations determine the diffusion coefficient of a similar electrolyte LiPF$_6$ in EC:DMC (3:7, weight) over molal concentrations, slightly deviating from molar concentrations (see SI Figure S14).
• Figure 5: Calculated viscosity $\eta$ using MD simulations. The experimental data points $\eta_\mathrm{exp}$ correspond to the Arrhenius fits of the experimental values from Logan et al.Logan2018 (see SI Figure S13a).