Liquid water polarization dataset

This dataset was used for Fig1.b and Fig.2 in the work "Machine Learning Inference of Molecular Dipole Moment in Liquid Water" by Knijff, Lisanne; Zhang, Chao.

The liquid water dataset was created using a trajectory of ab initio MD simulations. Simulations were done using the CP2K package and BLYP functional. A cubic box with a length of 23.493 Bohr was used, which contained 64 water molecules.