Liquid electrolytes in commercial lithium-ion batteries (LIBs) normally consist of more than two solvents to balance multiple requirements such as high ionic conductivity and low solution viscosity (A. Wang, et al., Cell Reports Physical Science, vol. 3, no. 9, 2022). However, only one solvent is considered in the widely used Doyle-Fuller-Newman (DFN) model for LIBs, based on the assumption that any solvents in the electrolyte move together as single entities. In a recent work by Wang et al. (A. Wang, et al., Cell Reports Physical Science, vol. 3, no. 9, 2022), the segregation of ethylene carbonate (EC) and ethyl-methyl carbonate (EMC) molecules was observed in a Hittorf experiment. They found ~5% change in the EC:EMC ratio post-Hittorf polarization and predicted over 50% change near the interface with a transport model (Fig. 1). Such changes in solvent distribution may affect the formation of solid electrolyte interface (SEI), one main reason for the long-term capacity fade of LIBs. However, there have not been any works which consider two solvents in a physics-based battery model.
To fill this research gap, we revise the traditional DFN model with a four species electrolyte (, , EC, and EMC) (C. W. Monroe, Journal of The Electrochemical Society, vol. 164, no. 11, pp. E3547-E3551, 2017) and couple it to the interstitial diffusion limited SEI model (L. von Kolzenberg, et al., ChemSusChem, vol. 13, no. 15, pp. 3901-10, May 18 2020). The main characteristics of our model can be summarized into four aspects. (1) We add a mass transport equation for EC. (2) A cross diffusion term appears in the transport equations of both EC and to consider their dragging effect. (3) In the MacInnes equation, to avoid errors brought by identifying thermodynamic factors, we directly use the experimentally measured liquid junction potential (LJP) as a function both EC and concentrations. (4) We assume the EC concentration has a linear effect on the SEI current density.
In Fig. 2 (a), the EC concentration gradient is opposite to that of Li+, which is similar to the result of Wang et al. (A. Wang, et al., Cell Reports Physical Science, vol. 3, no. 9, 2022). The concentration at (the negative electrode – current collector interface) of the double model is slightly higher than that of the DFN model due to the dragging effect from EC, which will further lead to the differences in electrolyte conductivity and diffusivity. All these effects alter the electrolyte potential and therefore terminal voltage. In summary, the revised model can capture more complicated transport mechanisms in the electrolyte.