Deep-learning interatomic potentials of the ɛ-ZrX_{2} series (X=H, D, and T)

Area of Science:

• Materials Science
• Computational Chemistry
• Nuclear Engineering

Background:

• Epsilon-zirconium dihydride (ɛ-ZrH₂) is crucial for nuclear energy, hydrogen storage, and catalysis.
• Accurate simulation of material properties is essential for advancing these applications.

Purpose of the Study:

• Develop accurate deep-learning interatomic potentials (DP) for ɛ-ZrH₂.
• Enhance molecular dynamic (MD) simulations for zirconium-based hydrides.

Main Methods:

• Trained deep-learning potentials using density functional theory (DFT) data for ɛ-ZrH₂.
• Incorporated the H-isotope effect for irradiation simulations.
• Compared DP performance against modified embedded atom method (MEAM) potentials.

Main Results:

• Developed DP accurately predict structural, mechanical, and thermodynamic properties of ɛ-ZrH₂ at DFT accuracy.
• DP demonstrate superior performance compared to conventional MEAM potentials.
• DP facilitate MD simulations of ɛ-ZrX₂ (X=H, D, T) under irradiation.

Conclusions:

• The developed DP provide a significant advancement for simulating zirconium-based hydrides.
• These potentials enable more reliable MD simulations for hydrogen storage and nuclear applications.
• The inclusion of isotopic effects broadens the applicability of MD simulations for materials under irradiation.