Ab initio machine-learning unveils strong anharmonicity in non-Arrhenius self-diffusion of tungsten
2025

Machine Learning Reveals Non-Arrhenius Diffusion in Tungsten

publication 10 minutes Evidence: high

Author Information

Author(s): Zhang Xi, Divinski Sergiy V., Grabowski Blazej

Primary Institution: Institute for Materials Science, University of Stuttgart

Hypothesis

Can an efficient ab initio framework accurately compute the Gibbs energy of transition states in vacancy-mediated diffusion?

Conclusion

The study demonstrates that the non-Arrhenius behavior of tungsten self-diffusion is strongly influenced by anharmonicity.

Supporting Evidence

  • The proposed computational framework accurately predicts temperature-dependent self-diffusivity.
  • Strong anharmonicity was observed in the Gibbs energies of vacancy formation and migration.
  • The results align well with experimental data, indicating the robustness of the findings.

Takeaway

Scientists used a special computer method to understand how tungsten atoms move at high temperatures, finding that their movement doesn't follow the usual rules.

Methodology

The study employed a machine-learning-assisted thermodynamic integration approach to calculate Gibbs energies of vacancy formation and migration.

Limitations

The study primarily focuses on tungsten and may not directly apply to other materials without further validation.

Digital Object Identifier (DOI)

10.1038/s41467-024-55759-w

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication