Surface-Periodicity-Guided Structure Mapping for Automated Scanning Tunneling Microscopy

Abstract

Automating scanning tunneling microscopy (STM) is essential for scaling atomically precise characterization and, ultimately, fabrication of next-generation materials and devices. A key challenge is to maintain atomic precision across extended surface regions, where local structures must be aligned, identified, and revisited in a common lattice frame. Here we use surface periodicity as an intrinsic reference frame to transform lattice-resolved STM images into lattice-registered unit-cell maps. Using Ag/Si(111)-(7×7) as a model system, the framework preserves lattice indices and neighborhood relationships while standardizing local structural units. This representation enables data-efficient supervised recognition of Ag adsorption configurations from one or a few reference images and supports environment-aware target-site selection. We further implement the framework in a continuous automated STM workflow, demonstrating autonomous scanning, target identification, and structure-specific characterization. Our results show that crystalline surface periodicity can provide a transferable basis for automated STM operations on lattice-resolved surfaces.