Integrating Electronic Stopping Effects into MACE Potentials

[nicdieugenio]

Dear all,

I am working on developing a MLP for radiation damage simulations (high-energy collision cascades) in high-temperature superconductors like YBa2Cu3O7. In general, with the current potential we have, we account for the electronic system via an explicit electronic stopping power term in classical MD. As we are now transitioning to using MACE-based machine learning potentials for these simulations, I was curious to know whether you consider doable to:

• Incorporate MACE with an electronic stopping power term (parameterized separately?)
• Account for transient charge redistributions during high-energy events like collision cascades?

Has anyone experimented with coupling MACE potentials with a dynamic electronic stopping model in general?
I would appreciate any suggestions on whether a hybrid approach that integrates electronic stopping into the MACE framework could be a practical way forward for complex materials like the ones I am simulating.

Best regards,
Niccolò Di Eugenio

[Ruoyan0103]

Hi,

LAMMPS has a fix electron/stopping command, I guess perhaps you can use MACE in LAMMPS with this command together.