Database of migration barriers for Li/Na-containing compounds by high-throughput NEB workflow

Solid-state electrolytes (SSEs) are vital for advancing high-performance all-solid-state batteries. Since the energy barrier of SSEs is a main property that governs the ionic transport performance, it is important to construct a database of such barriers. While density functional theory (DFT) based nudged elastic band (NEB) method provides fundamental insights into lithium-ion migration barriers, the high computational costs make large-scale calculations challenging. Some existing migration barrier databases based on bond valence method exhibit limited accuracy and lack local descriptions of specific migration paths. To address these limitations, we have developed a high-throughput NEB workflow that integrates a fine-tuned CHGNet potential, enabling the efficient and accurate calculation of barriers across all inequivalent migration paths. This work presents a database containing 6928 materials with totally 20320 inequivalent migration paths and their corresponding energy barriers. Furthermore, by extracting local environment descriptors from these paths in lithium-containing compounds, we constructed machine learning models capable of effectively predicting energy barriers directly from the structural features of the migration pathways.