I’ve been working on calculating dispersion curves for point-mass Argon using the LJ potential, but I’m encountering some issues. The plots I’m getting are quite strange, and some even show imaginary frequencies. Additionally, the frequency scale appears to be off by a factor of around 30 compared to the literature data.
I’ve double-checked the potential parameters and their units, and they seem to be correct. I’m not sure what’s going wrong. Could you please take a look at the attached script and let me know your thoughts?
Hi Erick,
The imaginary modes (and odd frequencies perhaps) are because you haven’t optimised the structure first. Frequencies are only really meaningful for a minimum. If you add “opti conp” before you compute the phonons then things will look much better.
Regards,
Julian
Thanks for the suggestion. I replaced the “conv” tag in the first line with “opti conp” and re-ran the analysis. The dispersion curves have improved, but there’s still an issue. For example: the kinks that appear around 1.2 THz
I verified the potential energy from GULP with LAMMPS using the same structure and potential settings, and both are in a similar ballpark range, indicating that the interactions should be correct.
However, I’m still confused about what might be wrong with the GULP inputs. I’ve attached the modified file here. Could you please take a look and let me know your thoughts?
Hi Erick,
I’ve looked at your curves and don’t see anything unexpected, so I think you need to post what you’re comparing against that makes you think something is wrong. In the original version of your post, the flat regions were just the zone boundaries at the middle of the curve (since you’re sampling the whole BZ along the direction), which is expected. I’m not sure what the kinks are that you’re referring to? Obviously at gamma there are discontinuities at the origin, but I don’t see kinks away from gamma (including images of gamma elsewhere beyond the primitive BZ).
Regards,
Julian