Data for "Unfolding the Structural Stability of Nanoalloys via a Combination of Symmetry-Constrained Genetic Algorithm and Neural Network Potential"

PtNi_alloy_eam.db is the dataset (ase.db object) consisting of 55982 intially sampled Pt-Ni alloy structures with EAM energies and forces.

PtNi_alloy_dft.db is the dataset (ase.db object) consisting of the final 6828 resampled Pt-Ni alloy structures with DFT energies and forces calculated by VASP. This is the training set for the NNP, and could be very useful for fitting other machine learning models.

PtNi_nanoalloy_vertices_nnp.db is the dataset (ase.db object) consisting of all the vertices (stable structures) on the convex hulls obtained from NNP-based SCGA runs on 36 Pt-Ni nanoalloy systems. The energies are given by the NNP. Additional information such as mixing energy, motif and symmetry axis are also saved in the dataset and can be queried by the 'data' keyword. An xyz format trajectory of these stable structures is also uploaded.

All the input files and scripts for hybrid MC-MD simulations, QBC resampling, DFT calculations, NNP training, NNP-based SCGA runs and convex hull analysis are provided in inputs_and_scripts.zip.