Mechanism of Fast Proton Conduction in Layered Solid Electrolytes by Machine-Learning Molecular Dynamics

Abstract

We present the proton conduction mechanism in a layered solid electrolyte (Mg,H)-doped FeOCl, which exhibits a high proton conductivity of approximately 10 mS cm -1 at 393 K under anhydrous conditions, using density functional theory (DFT) calculations and molecular dynamics simulations based on machine-learning force fields. DFT calculations show that O–H configurations are energetically more stable than Cl–H configurations, while finite-temperature MD simulations indicate that most hydrogen atoms are strongly bound to oxygen sites. In contrast, only a small fraction of hydrogen atoms associated with Cl sites—where the binding is relatively weak—participate in long-range proton diffusion. These weak-binding sites form an interconnected network within the layered framework, giving rise to anisotropic proton conduction. Our findings demonstrate that fast proton transport in layered solids is governed by such continuous weak-binding pathways.