Different displacements can result in different phonon vibration modes

[Mr sun]

Dear sir,
After I learned the PHYSICAL REVIEW B 93,035414 (2016) article.Use ATK software to simulate the results of the article, found that set dynamical_matrix parameter "processes_per_displacement =" 1 or 2, resulting in different phonon vibration mode (such as the picture), what is the basis of this parameter settings? What is the reason for this result? thank you very much

[Jess Wellendorff]

Please supply the scripts you used to obtain these results. Makes it a whole lot easier to help...

[Mr sun]

This is the script for the corresponding image of graphene above (see attached).

[Jess Wellendorff]

I cannot say exactly what may have happened in your calculations, but I have run simplified versions of your scripts on our local supercomputer (ATK 2017), and do not find any difference between 1 and 2 processes per displacement: The vibrational frequencies do not depend on that. See input/output: https://www.dropbox.com/sh/wcctkkqg2sxzhgk/AACmaXLpKAgCCVrJuelQp7fUa?dl=0

[Anders Blom]

You are looking at k=0 where small numerical differences can affect the results since the eigenvalues are zero, but often not exactly, just very close to. For the same reason you actually cannot even identify LA and TA phonon modes at k=0. Try at some finite but small k and probably the results will match no matter how you run.