Dataset for "Machine Learning Stability and Bandgaps of Lead-Free Perovskites for Photovoltaics"

Datasets used in the publication "Machine Learning Stability and Bandgaps of Lead-Free Perovskites for Photovoltaics" [doi:10.1002/adts.201900178]. All structures were relaxed with the following parameters using Quantumwise QATK 2017: - SG15-GGA norm-conserving (Vanderbilt) pseudopotentials employed in a LCAO-approach (200 Hartree cutoff)<br> - 2x1x2-cubic-perovskite-supercells, relaxed from cubic 11.4Åx5.7Åx11.4Å-structures (forces &lt; 0.01eV/Å)<br> - 300K Fermi-Dirac-smearing<br> - a 6x12x6 k-point grid (Monkhorst-Pack) <br> Specifically, the included files are: <strong>db_2.data: </strong>the actual database used for model building (json-format)<br> <strong>lead_set.data:</strong> the "external" test set used to test predictive power with out of sample compounds (json-format)<br> <strong>load_stanley_c.py:</strong> a python script to parse the .json-files to a python-dictionary including the structures (relaxed and unrelaxed) as ASE-atoms The format of the datafiles is as follows (-1 generally denote values not parsed from the raw data):<br> {<br> "&lt;idstring&gt;" : {<br> "trajectory" : n/a,<br> "energy" : total DFT energy in eV,<br> "rstruc" : relaxed structure, 3-tuple: (cell-vectors, scaled_positions, elements),<br> "gaps" : { "opt_gap", "ind_gap } - both direct and indirect gap,<br> "effective_mass" : n/a,<br> "iterations" : number of relaxation steps,<br> "calc" : some calculation metadata,<br> "ustruc" : unrelaxed input structure,<br> <br> }<br> }<br> Missing ids relate to structures filtered out, because the calculation didn't converge. Some code which works with a different representation of this data can be found at https://github.com/jstanai/Machine-Learning-Perovskite-Properties-for-Photovoltaics