Chenyu Wang¹, Wei Tian¹, Ke Zhou^1*

1 College of Energy, SIEMIS, Soochow University, Suzhou 215006, China

* Corresponding authors emails: zhouke@suda.edu.cn

DOI10.24435/materialscloud:89-2k [version v1]

Publication date: Aug 29, 2024

How to cite this record

Chenyu Wang, Wei Tian, Ke Zhou, Ab-initio simulation of liquid water without artificial high temperature, Materials Cloud Archive 2024.131 (2024), https://doi.org/10.24435/materialscloud:89-2k

Description

Comprehending the structure and dynamics of water is crucial in various fields such as water desalination, ion separation, electrocatalysis, and biochemical processes. While reported works show that the ab-initio molecular dynamics (AIMD) can accu- rately portray water’s structure, the artificial high temperature (AHT) from 120K to 30K is needed to mimic the quantum nature of hydrogen-bond network from GGA, metaGGA to hybrid functionals. The AHT proves to be an inadequate approach for systems involving aqueous multiphase mixtures, such as water-solid interfaces and aque- ous solutions. This is due to the activation of additional phonons in other phases, which can lead to an overestimation of the dynamics for nearby water molecules. In this work, we find the regularized SCAN (rSCAN) functional can well capture both the structure and dynamics of liquid water at ambient conditions without AHT. Moreover, rSCAN can well match the experimental results of hydration structures for alkali, alkali earth and halide ions. We anticipate that the versatile and accurate rSCAN functional will emerge as a key tool based on ab-initio simulation for investigating chemical processes in aqueous environments.

Materials Cloud sections using this data

Files

File name	Size	Description
NEP_dataset.zip MD5md5:1a7d02c730c0a12d3835973a1550fd2d	118.3 MiB	This data repository includes training sets and neuroevolution potentials (NEPs) for liquid water, utilizing various functionals such as PBE+D3, optB88-vdW, SCAN, and rSCAN.

License

Files and data are licensed under the terms of the following license: Creative Commons Attribution 4.0 International.
Metadata, except for email addresses, are licensed under the Creative Commons Attribution Share-Alike 4.0 International license.

External references

Keywords

water AIMD machine learning potential metaGGA density functional theory