Abstract
Atomic arrangements and local sub-structures fundamentally influence emergent material functionalities. The local structures are conventionally probed using spatially resolved studies and the property correlations are usually deciphered by an operator based on sequential explorations and auxiliary information, thus limiting the throughput efficiency. Here we demonstrate a Bayesian deep learning based framework that automatically correlates material structure with electronic properties using scanning tunneling microscopy (STM) measurements in real-time. Its predictions are used to autonomously navigate the sample and direct exploration toward regions that maximize a given material property. This autonomous method is deployed on the low-temperature ultra-high vacuum STM to understand the structure-property relationship in a Europium-based semimetal, EuZn2As2, which has an anti-ferro-magnetic ordering and exhibits a characteristic bandgap. The DKL framework employs a sparse sampling approach to efficiently construct the scalar property space using a minimal number of measurements, about 1 - 10 % of the data required in standard hyperspectral imaging methods. We further demonstrate a target property-guided active learning of structures within a multiscale DKL implemented across length scales in a cascaded fashion for the autonomous discovery of structural origins for an observed material property. This framework offers the choice to select and derive a suitable scalar property from the spectroscopic data to steer exploration across the sample space. Our findings reveal correlations of the electronic properties unique to surface terminations, local defect density, and point defects.
