Exploring Phononic Properties of Two-Dimensional Materials using Machine Learning Interatomic Potentials

In this manual we provide a guide on the practical implementation and reproduction of the results presented in our publication entitled: "Exploring Phononic Properties of Two-Dimensional Materials using Machine Learning Interatomic Potentials". We specifically discuss the repository gitlab.com/ivannovikov/mlip_phonopy with the MLIP_PHONOPY code—a C++ interface between the MLIP code and the PHONOPY software—which allows one to calculate phonon spectra, group velocities, thermal properties, etc., of a two-dimensional material.

Along with the repository description, this manual contains an instruction on quick installation of the stable branch of the MLIP code, a description of VASP input files for ab initio molecular dynamics (AIMD) calculations (namely, the folders Structures and VASP-inputs) and training set preparation, an instruction on passive training of MomentTensor Potentials (MTPs) using the MLIP code. Finaly, we describe the folders Untrained-MTPs and Exampleswith the additional files available here: dx.doi.org/10.17632/7ppcf7cs27.1.