Gaussian Approximation Potential for Hexagonal Boron Nitride (hBN-GAP)
Repository URI
Repository DOI
Change log
Authors
Thiemann, Fabian L
Rowe, Patrick
Müller, Erich A
Description
This dataset contains the Gaussian Approximation Potential (GAP), a potential to describe the interatomic interactions in crystalline hexagonal boron nitride (hBN). Moreover, training and validation data obtained from DFT calculations are available too. The directory Potential contains the XML files for the hBN-GAP machine learning model. The file r6_innercut_hBN.xml contains the potential parameter for the semi-analytical 2b potential. The selected training and validation configurations are available in xyz format and named hBN_GAP_Training.xyz and hBN_GAP_Validation.xyz, respectively. An example job for running a hBN-GAP simulation in LAMMPS can be found in the directory Example_LAMMPS_job.
Version
Software / Usage instructions
This potential can be used in molecular dynamics simulations. It requires the QUIP software package github.com/libAtoms/QUIP. It can also be used directly in LAMMPS and an example input file is provided.
Keywords
DFT, GAP, Machine Learning, hBN