Advances in Molecular Dynamics Studies of Molten Salts Based on Machine Learning
Достижения в исследованиях расплавленных солей методом молекулярной динамики на основе машинного обучения
2023-01-01
SCID: 54.1/q4dxz99x
Discuss with AI
coordination chemistrymachine learning potentialsmolecular dynamicsmolten saltsneural networks
Figures from the paper
Abstract (AI)
Molten salt is a kind of molten material with important application value.However, the relationship between microstructure and macroscopic properties of molten salt has not been fully investigated, so it is of great significance to carry out extensive molecular dynamics studies.For high-temperature molten salts, molecular dynamics studies mainly relied on the development of force fields in classical molecular dynamics and first principles molecular dynamics previously.Due to the accelerated development of machine learning and neural networks, the potential for molten salts based on machine learning has been developed recently, and great progress has been made in exploring the coordination chemistry and accurately predicting physical properties.Herein, the latest research progress of molecular dynamics related to molten salt was firstly reviewed, especially the development status of machine learning potential.Secondly, the application progress of machine learning potentials in the research of molten salt was summarized.Finally, the research prospect of machine learning potentials for molten salt was discussed, and some suggestions were given.
Key Findings
1
Machine-learning potentials enable more accurate prediction of molten-salt physical properties, according to the reviewed research.
2
Machine-learning-based potentials have recently emerged for molecular dynamics simulations of molten salts, complementing classical force fields and first-principles methods.
3
The microstructure–macroscopic-property relationship in molten salts remains insufficiently investigated, motivating further molecular-dynamics studies.
4
The review summarizes recent developments and applications of machine-learning potentials for molten salts and identifies future research directions and recommendations.
5
These machine-learning potentials have substantially advanced investigations of molten-salt coordination chemistry.
Research Object
high-temperature molten salts
Research Subject
the relationship between molten-salt microstructure, coordination chemistry, and macroscopic physical properties as investigated using molecular dynamics and machine-learning potentials
Publication Details
Publication Date
2023-01-01
Journal
Publisher
ISSN
Cited by
11
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest