[Pw_forum] strange results for phonon frequencies.

Axel Kohlmeyer akohlmey at gmail.com
Sun Sep 2 08:53:09 CEST 2012

On Sun, Sep 2, 2012 at 6:09 AM, Elie M <elie.moujaes at hotmail.co.uk> wrote:
> Dear all,
> Sorry to ask so many questions about computing phonon frequencies but
> recently i have got some strange results for one of the systems I am working
> on. My work involves computing phonon frequencies for a 42 carbon at the
> Gamma point so in total 126 irreducible representations.


> What is the reason for those LARGE negative values at the beginning and the
> LARGE ones at the end? Is it because mixing_beta was changed?

no. think about it for a bit and what mixing does.
why should mixing_beta change the result?
it only should only affect how fast you converge
and whether you converge at all.

you would get those "strange large" frequencies, when
your system is not in the minimum, which is to be
expected, if you didn't reoptimize the geometry with the
larger density cutoff (which is still on the low side, btw).

the procedure is usually the following:
1) you have to choose your cutoffs, so you have
    well converged forces and derivatives of them
2) you need a well optimized geometry with those
    settings (you may have to break symmetry by
    slightly randomizing positions occasionally)
3) you compute phonons.

you cannot do 1) after 3) and without 2).

> Thanks for the help
> Elie Koukaes
> University of Nottingham
> NG7 2RD
> UK
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum

Dr. Axel Kohlmeyer  akohlmey at gmail.com  http://goo.gl/1wk0
International Centre for Theoretical Physics, Trieste. Italy.

More information about the users mailing list