We have developed a hierarchy of DFT-based approaches to understand and design electrode/buffer/SE interfaces in all-solid-state batteries. Even with relatively efficient thermodynamics approximations, we reveal that S-O exchange reactions between oxides (O^2-) and thiophosphates (PS₄^3-) resulting in the formation of phosphates (PO₄^3-) are responsible for large reaction driving force between the common NaMO₂ (M = Co, Ni, Fe, Mn) cathodes and Na₃PS₄ SE. Such reactions, with their associated large volume changes, can be mitigated by careful selection of the cathode/SE combination, or the use of buffer layers with good stability between both materials. We have also identified several promising binary oxides with similar or better chemical compatibility with most electrodes and SEs than the commonly-used Al₂O₃.

Despite the success of thermodynamics-based approximations, we further demonstrate that explicit modelling of the electrode/SE interface via ab initio molecular dynamics (AIMD) simulations sometimes yield different and more realistic predictions of interfacial reaction products. For example, AIMD predicts that formation of SO₄^2– are kinetically favored over the formation of PO₄^3– at the NaCoO₂/Na₃PS₄ interface model, in contrast to the predictions of the thermodynamic models. These observations have been validated experimentally. We have also extended such studies to the LiNi_0.85Co_0.1Al_0.05O₂/Li₆PS₅Cl interface, with and without buffer layers, and the predicted interfacial reactions are in good agreement with experimental characterizations.