Machine learning lattice parameters of monoclinic double perovskites

Abstract

AbstractMonoclinic double perovskites, such as A2B′B″O6, exhibit a wide range of unique electronic, magnetic, and optical properties, depending on the alloying elements, ordering, and lattice distortion. Four lattice constants, a, b, c, and β, can vary due to the ionic radii of alloying elements, which represent the compression, expansion, and tilting of octahedra in the crystal structure. In the current study, we develop the Gaussian process regression model to shed light on relationships between effective ionic radii and lattice constants for monoclinic double perovskites. The model is highly accurate and stable, contributing to fast, low‐cost estimations of lattice constants.