Deep Learning Enhanced in Situ Atomic Imaging of Ion\nMigration at Crystalline–Amorphous Interfaces

Improving\nthe performance of energy storage, neuromorphic computing,\nand more applications requires an in-depth understanding of ion transport\nat interfaces, which are often hindered by facile atomic reconfiguration\nat working conditions and limited characterization capability. Here,\nwe construct an in situ double-tilt electric manipulator inside an\naberration-corrected scanning transmission electron microscope. Coupled\nwith deep learning-based image enhancement, atomic images are enhanced\n3-fold compared to traditional methods to observe the potassium ion\nmigration and microstructure evolution at the crystalline–amorphous\ninterface in antimony selenide. Potassium ions form stable anisotropic\ninsertion sites outside the (Sb<sub>4</sub>Se<sub>6</sub>) chain,\nwith a few potassium ions present within the moieties. Combined experiments\nand density functional theory calculations reveal a reaction pathway\nof forming a novel metastable state during potassium ion insertion,\nfollowed by recovery and unexpected chirality changes at the interface\nupon potassium ion extraction. Our unique methodology paves the way\nfor facilitating the improvement and rational design of nanostructured\nmaterials.