Ogre

Ogre is an open-source package for structure prediction of organic and inorganic epitaxial interfaces by lattice matching followed by surface matching. The lattice matching step produces domain-matched interfaces, where commensurability is achieved with different integer multiples of the substrate and film unit cells. In the surface matching step, Bayesian optimization (BO) is used to find the interfacial distance and registry between the substrate and film. Different strategies are employed for inorganic and organic materials to perform this step efficiently and avoid computationally expensive DFT calculations. For inorganic materials, the BO objective function is a geometric score function, based on the overlap and empty space between atomic spheres at the interface. Preliminary ranking of surface-matched structures is performed with a ranking score function, based on the degree of similarity of the interface environment to the bulk environment of both materials. For organic materials the BO objective function is based on dispersion-corrected deep neural network interatomic potentials, which are also used for preliminary ranking. Ogre includes additional utilities for constructing surface models and converging the surface and interface energies for both organic and inorganic materials. Ogre automatically identifies all symmetrically unique surfaces for the user-specified Miller indices and detects all possible surface terminations. Specifically for molecular crystals, the surface is cleaved from the bulk crystal structure with the molecules on the surface kept intact. A molecular graph representation is used to identify and repair cleaved molecules. Calculated surface energies can be used to predict the Wulff shape of a molecular crystal.