This README file provides a description of the results obtrained in the computational study of 2D Mo-S system on Al2O3 substrate, as published in the "Substrate-aware computational design of two-dimensional materials" paper https://arxiv.org/abs/2310.07604. The data record contains a list of structures from the evolutionary search, a training set of the MLIP used for local relaxation of the structures, the Jupyter Notebook with the scripts used for making the figures for the paper together the accompanying data, and thermochemistry calculations details.
The file notebooks.zip contains a Jupyter Notebook with all the scripts used to build the raw figures in the paper
and a figures_data folder with raw data on convex hulls, band structure caluclations, etc.
The file structures.zip contains results of the evolutionary search stored as a list of files in the the extended XYZ format.
all_EA_structures.xyz)all_EA_structures_withEnv.xyz)convex_hull_structures.xyz)convex_hull_structures_withEnv.xyz)metastable_structures.xyz)metastable_structures_withEnv.xyz)reference_structures.xyz)dft_relaxed_convex_hull_structures.xyz)dft_relaxed_metastable_structures.xyz)dft_relaxed_metastable_structures.xyz)Each structure without substrate contains the following metadata from the EA search:
ID key)energy key)eformCH key)x parameter in $Mo_{1-x}S_{x}$ system (compCH key)energyCCH key)Each structure with substrate contains the following metadata from the EA search:
ID key)adjustedEnergy key)environmentEnergy key)adjustedSupercellFactor key)x parameter in $Mo_{1-x}S_{x}$ system ('ref_compCH' key)refinedAdjustedEnergyCCH key)The file thermochemistry.zip contains a Jupyter Notebook with all the scripts used to build the raw figures for thermochemistry calculations in the paper and a list of folders with raw data on MoO3, S2 and SO2 vibrational frequencies calculations.
The file uspex.zip contatins settings of the modified version of the USPEX code, that was used to perform the evolutionary
search (input.uspex), and the substate file with the atomic structure of the Al2O3 substrate (POSCAR_SUBSTRATE).
The file mlip.zip contains the trained MTP potential and the training dataset with 3 different classes of systems used for training the MLIP (3128 structures in total) within the automatic self-consistent training (ASCT) approach. Details of the ASCT workflow are given in the paper. All the training structures are stored in a single file in two formats: extended XYZ format (.xyz) and MTP configurations format (.cfg). The following structures enter the dataset:
Each structure is accompanied by its corresponding energies, forces and stress tensor computed using VASP. In the case of surface slab configurations, a vacuum layer of 20 Å was added to avoid the periodic surface interactons. Each configuration contains the data on its energy (in eV), interatomic forces (in eV/Å) and stresses (in kbar).
This file lammps.zip contains the input files for the LAMMPS code, that were used for local relaxation of 2D Mo-S
structures during the EA search (lammps.in) and relaxation after joining with substrate (lammps_env.in).
The file vasp.zip contains the INCAR files used for the VASP calculations while training the MLIP and performing the re-evaluation of the stable structures from the EA run.