AX_2 Fluorite database for MLFF training underlying the publication: Predictive accuracy of on-the-fly Machine Leaning Force Fields for superionic diffusion kinetics in AX_2 Fluorites

DOI:10.4121/0c17c247-8f79-4372-be4a-92d54223b143.v1
The DOI displayed above is for this specific version of this dataset, which is currently the latest. Newer versions may be published in the future. For a link that will always point to the latest version, please use
DOI: 10.4121/0c17c247-8f79-4372-be4a-92d54223b143

Datacite citation style

Menno Bokdam (2025): AX_2 Fluorite database for MLFF training underlying the publication: Predictive accuracy of on-the-fly Machine Leaning Force Fields for superionic diffusion kinetics in AX_2 Fluorites. Version 1. 4TU.ResearchData. dataset. https://doi.org/10.4121/0c17c247-8f79-4372-be4a-92d54223b143.v1
Other citation styles (APA, Harvard, MLA, Vancouver, Chicago, IEEE) available at Datacite

Dataset

AX_2 fluorite database for force field training containing six materials: CaF2, Li2O, PbF2, SrCl2, SrF2, BaF2. The dataset has been used (and could be used by you) to train Machine Learning Force Fields (MLFFs) to simulate the (onset of) superionic phase.


These structures have been selected with the on-the-fly Machine-Learning Force Fields method (as implemented in VASP 6.3 and higher) and described in this reference:

Jinnouchi R., Lahnsteiner J., Karsai F., Kresse G., Bokdam M., "Phase transitions of hybrid perovskites simulated by machine-learning force fields trained on the fly with Bayesian inference", Phys. Rev. Lett. 122, 225701, (2019)


The structures have been automatically selected during a heating run from 800 to 2600 K and the corresponding coordinates, energies, forces and stresses are stored in the ML_AB file. The ML_AB file can be used to generate new force fields with the method of the users preference. The open-source FPdataViewer software can be used to read-in the ML_AB file. It also contains a connection to the Atomic Simulation Environment with which descriptors can be generated.


Caution: There are 2 versions of the ML_AB databases included per fluorite:

ML_AB_full: all data picked up by on-the-fly training, contains however ~20% molten structures with uncoverged DFT labels (ie. energies, forces, stress)

ML_AB_filtered: a curated version of the data above, whereby all structures with lattice vectors deviating by more then 5% from the mean have been filtred out, cleaning up most of the unconverged labels. Files contain the *.out.* in the filename.


FPdataviewer factsheets

A high level overview of the ML_AB databases has been generated using the open-source FPdataViewer software (https://github.com/dynamicsolids/FPdataViewer). Each pdf file contains statistics related to the structures, energies and forces stored in the databases. The factsheet can be used to get a quick overview of the data stored in the database.

History

  • 2025-11-18 first online, published, posted

Publisher

4TU.ResearchData

Format

Zip Compressed Archive of text files, PDF files

Organizations

University of Twente, Faculty of Science and Technology and the MESA+ institute for Nanotechnology

DATA

Files (10)