Email (Optional)
No response
Problem
In the PhononCalc, there is currently a mechanism to resolve imaginary modes that involves random rattling of the atoms followed by re-relaxation. Empirically, this works okay, but it is by no means perfect. The phonon rattle scheme as implemented in hiphive should yield even better results and would be worth considering.
Proposed Solution
Use the phonon-rattle scheme in hiphive in place of a random rattle.
Alternatives
An MC rattle scheme could be used instead, but the choice of MC parameters it not clear in advance.
Code of Conduct
Email (Optional)
No response
Problem
In the
PhononCalc, there is currently a mechanism to resolve imaginary modes that involves random rattling of the atoms followed by re-relaxation. Empirically, this works okay, but it is by no means perfect. The phonon rattle scheme as implemented in hiphive should yield even better results and would be worth considering.Proposed Solution
Use the phonon-rattle scheme in hiphive in place of a random rattle.
Alternatives
An MC rattle scheme could be used instead, but the choice of MC parameters it not clear in advance.
Code of Conduct