Ab initio thermodynamics of liquid and solid water: supplemental materials

Authors: Bingqing Cheng1*, Edgar Engel1, Jörg Behler2, Christoph Dellago3, Michele Ceriotti1

  1. Laboratory of Computational Science and Modeling, Institute of Materials, École Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland
  2. Universität Göttingen, Institut für Physikalische Chemie, Theoretische Chemie, Tammannstr. 6, 37077 Göttingen, Germany
  3. Faculty of Physics, University of Vienna, Boltzmanngasse 5, 1090 Vienna, Austria
  • Corresponding author email: bingqing.cheng@epfl.ch

DOI10.24435/materialscloud:2018.0020/v1 (version v1, submitted on 04 December 2018)

How to cite this entry

Bingqing Cheng, Edgar Engel, Jörg Behler, Christoph Dellago, Michele Ceriotti, Ab initio thermodynamics of liquid and solid water: supplemental materials, Materials Cloud Archive (2018), doi: 10.24435/materialscloud:2018.0020/v1.


Thermodynamic properties of liquid water as well as hexagonal (Ih) and cubic (Ic) ice are predicted based on density functional theory at the hybrid-functional level, rigorously taking into account quantum nuclear motion, anharmonic fluctuations and proton disorder. This is made possible by combining advanced free energy methods and state-of-the-art machine learning techniques. The ab initio description leads to structural properties in excellent agreement with experiments, and reliable estimates of the melting points of light and heavy water. We observe that nuclear quantum effects contribute a crucial 0.2 meV/H2O to the stability of ice Ih, making it more stable than ice Ic. Our computational approach is general and transferable, providing a comprehensive framework for quantitative predictions of ab initio thermodynamic properties using machine learning potentials as an intermediate step.

In this set of supplemental materials, we have included the neural network potential for bulk water, including its training set in two different formats. We have also included the input files for running free energy calculations.

Materials Cloud sections using this data

No Explore or Discover sections associated with this archive entry.


File name Size Description
MD5MD5: 66083ac5ed9489b4a62f0506c24f4a2c
478 Bytes An overview of the data set.
MD5MD5: bcca65b0296e8916615c5c517c9998ee
22.2 KiB The parameters of the water neural network potential based on revPBE0-D3 DFT, and an example on how to use it.
MD5MD5: 57f26ce2a2e717e1ddc66caddde20700
7.4 MiB A whole set of input files for running
* path-integral molecular dynamics simulations ./pimd/
* Free energy estimation of an ice system using thermodynamic integration method using the NN potential ./NN-TI/
* revPBE0-D3 DFT calculations using the CP2K code ./cp2k-input/
* compute the chemical potential difference between ice and liquid water using the interface pinning method ./interface-pinning/
* Thermodynamic integration between the MBPOL water potential and the neural network potential ./mbpol-TI/
* a sample python data analysis notebook ./data-analysis/
MD5MD5: 8cf0da8a72ddcb778529d2869990a53c
17.4 MiB The training set for ML potentials, based on revPBE0-D3 DFT.
1593 bulk liquid water configurations + energy + forces
* input.data: the format for training neural network potentials.
* dataset_1593.xyz: in libatom format.


Files and data are licensed under the terms of the following license: Creative Commons Attribution 4.0 International.

External references

Preprint (Preprint where the data is discussed. The manuscript is also accepted in PNAS.)


Machine learning potential Machine learning potential training set Free energy calculation input files

Version history

04 December 2018 [This version]