Deep Potential Molecular Dynamics Training Dataset for Pentavalent Vanadium Electrolyte Precipitation in Vanadium Flow Batteries

This dataset contains the training data for a deep potential (DP) model developed to investigate the precipitation mechanism of pentavalent vanadium (V(V)) species in vanadium flow battery (VFB) electrolytes using deep potential molecular dynamics (DPMD).The dataset comprises 76,172 DFT-labeled configurations spanning the complete structural evolution from isolated solvated V(V) ions to vanadium oxide precipitates, including hydrated vanadium oxo ions, hydroxylated intermediates with varying protonation degrees, and aggregated structures with various polymerization degrees. All configurations were generated through an active learning workflow (DP-GEN2) with 79 iterations, and labeled by single-point DFT calculations at the PBE/DZVP-MOLOPT-SR-GTH level with D3 dispersion correction using CP2K.