Unable to reproduce published elastic constants despite matching cohesive energy and lattice constant

LAMMPS
1

I’m trying to reproduce the elastic constants of the published HfNbTaTiZr EAM potential (https://doi.org/10.1016/j.ijplas.2026.104626). My lattice constant and cohesive energy match the paper, and C12 also matches (~109 vs. 108 GPa). However, C11 is consistently ~145 GPa (paper: 170 GPa) and C44 is ~64 GPa (paper: 54 GPa). Paper values are from figure not a table.
I’ve repeated the calculations with different system sizes, relaxation procedures, and corrected atom-type mapping, but the results remain the same. I’ve attached my relaxation and elastic constant scripts along with the log file for elastic constant calculation script.

in.HfNbTaTiZr_pristine_relax.lmp (1.8 KB)

in_HfNbTaTiZr_elastic_constants.lmp (6.3 KB)

log.elastic (123.3 KB)

2

Please note that this is a question about your research and how you validate your potentials and not about LAMMPS. Generally, those kinds of questions have a very low probability of receiving a meaningful answer in a public forum. You need to talk to somebody with experience in your specific area.

It just so happens, though, that I have recently been reviewing a tutorial for simulations of HEAs
and that leads me to some questions / suggestions:

  • you show the results of just one run. what happens if you have a different (random) distribution of atom types? results from simulations always need some error estimates, anyway.
  • please note that generally lattice constant and energy converge much easier than pressure.
  • have you considered using MC, e.g. with fix atom/swap to further equilibrate your system? HEA geometries are not always fully random. In some cases certain atom pairing are more likely to find than others. This can be easily estimated from a series of tests of binary mixtures.
  • what kind of averaging / equilibration have the authors of the paper performed?
3

Thank you for your comment.

  1. Using random distribution also have the same results
  2. I haven’t consider using MC yet.
  3. The issue is author didn’t specify the exact path to how he calculated the elastic constant values,
    Somehow I am stuck. and Can’t figure out the possible way to validate the elastic constant values
4

I would call that a deficiency in the publication and suggest to contact the authors for clarification.
For any further assistance, you will have to find a collaborator with more experience in these matters. There is not much else that can be done from where I stand.

5

Thank you soo much. Will try that.