Machine Learning Molecular Dynamics Insight into High Interface Stability and Fast Kinetics of Low-cost Magnesium Chloride Amine Electrolyte for Rechargeable Magnesium Batteries

The development of rechargeable magnesium batteries (RMBs) is hindered by the lack of long-lifespan and low-cost electrolytes. Moreover, due to lacking of an in-depth understanding of accurate dynamic solvation structures, the relationship between the interface kinetics behavior and a stable anode interface is still unclear. Herein, we develop a novel low-cost electrolyte containing the MgCl2 salt and the 3-methoxypropylamine solvent (denoted as the MgCl2-S2 electrolyte), which exhibits higher interface stability and faster kinetics compared to the reference Mg(TFSI)2-S2 electrolyte. The powerful machine learning molecular dynamics (MLMD) is employed to systematically investigate the differences of the fine solvation structures and explore the fundamental reason for better interfacial dynamics processes. The results show that the degree of dissociation in MgCl2-S2 electrolyte is less than 1%, much lower than 95.1% in MgTFSI2-S2 electrolyte. The solvation structures dominated by Cl− ions can induce Mg2+ ions to preferentially acquire electrons in both kinetics and thermodynamics, which thus improving the deposition kinetics and enhancing the stability of the anode interface. This work inspires a new paradigm from MLMD to investigate the relationship between interface stability and accurate solvation structures for RMBs.