A crucial ingredient in lithium (Li) and sodium (Na)-ion batteries (LIBs and NIBs) is the electrolytes. The use of Li-metal (Na-metal) as anode in liquid electrolyte LIBs (NIBs) is constrained...

Ion transport in materials is routinely probed through several experimental techniques, which introduce variability in reported ionic diffusivities and conductivities. The computational prediction of ionic diffusivities and conductivities helps in identifying good ionic conductors, and suitable solid electrolytes, thus establishing firm structure-property relationships. Machine-learned potentials are an attractive strategy extending the capabilities of accurate \emph{ab initio} molecular dynamics to longer and larger simulations, enabling simulations of ion transport at low-temperature. However, being machine-learned potentials in their infancy, critical assessments of their predicting capabilities are seldom. Here, we identified the main factors controlling the quality of machine-learning potentials based on the moment tensor potential when applied to the properties of ion transport in solid electrolytes. Our results underline the importance of high-quality training sets in fitting moment tensor potentials, diverse training sets, and the importance of intrinsic defects which may occur in solid electrolytes. We demonstrate the limitations posed by short-time scale and high-temperature AIMD simulations to predict the room-temperature properties of materials.

A crucial ingredient in lithium (Li) and sodium (Na)-ion batteries (LIBs and NIBs) is the electrolytes. The use of Li-metal (Na-metal) as anode in liquid electrolyte LIBs (NIBs) is constrained by several issues including thermal runway and flammability, electrolyte leakage, and limited chemical stability. Considerable effort has been devoted toward the development of solid electrolytes (SEs) and all-solid-state batteries, which are presumed to mitigate some of the issues of Li-metal(Nametal) in contact with flammable liquid electrolytes. However, most SEs, such as Li 3 PS 4 , Li 6 PS 5 Cl and Na 3 PS 4 readily decompose against the highly reducing Li-metal and Na-metal anodes. Using first-principles calculations we elucidate the stability of more than 20 solid||solid interfaces formed between the decomposition products of Li 3 PS 4 , Li 6 PS 5 Cl (and Na 3 PS 4 ) against the Li-metal (Nametal) electrode. We suggest that the work of adhesion needed to form a hetereogenous interfaces is an important descriptor to quantify the stability of interfaces. Subsequently, we clarify the atomistic origins of the resistance to Li-ion transport at interfaces of the Li-metal anode and selected decomposition products (Li 3 P, Li 2 S and LiCl) of SEs, via a high-fidelity machine learned potential (MLP). Utilising an MLP enables nano-second-long molecular dynamics simulations on 'large' interface models (here with 8320 atoms), but with similar accuracy to first-principles approaches. Our simulations demonstrate that the interfaces formed between Li-metal and argyrodite (e.g., Li 6 PS 5 Cl) decomposition products are resistive to Li-ion transport. The implications of this study are important since binary compounds are commonly found in the vicinity of Li(Na)-metal upon chemical and/or electrochemical decomposition of ternary and quaternary SEs.