[Pw_forum] Projection, convergence and imaginary phonon frequencies

[John Kendrick]

Hi,

Yes Happy New Year.

I only gave the first few soft modes for each case.  It is these soft 
modes I am interested in.  There are 174 modes (58 atoms) altogether.  I 
listed 13 in one case as there was more of a match with the intensities 
from the other projection methods.

The comment about cutoff is interesting as I had found that the 
optimiser (BFGS) was having problems.  This would also be consistent 
with  getting more consistent results with the USP potentials.

Thanks

John

Axel Kohlmeyer wrote:
> On Mon, 12 Jan 2009, Nicola Marzari wrote:
> 
> NM> Dear John,
> 
> dear nicola,
> 
> happy new year from cold philly.
> 
> NM> if it is a crystal, the most critical parameter should be k-point
> NM> sampling. Increase that, maybe using shifted meshes of the form
> NM> 2n 2n 3n 1 1 1 , with n=1,2,3...
> 
> but it is a _molecular_ crystal, so i guess there are some 
> areas with comparatively low electron density and quite weak 
> interactions, as seen by the long list of rather soft modes.
> i would suspect that will need to be converged very well wrt 
> to the basis set to not see any ripples on the potential surface.
> 
> for TM oxygen pseudos i would go to 120ry or higher to get
> tight convergence. for MD 80-85ry is usually sufficient 
> (the more traditional 70ry has been shown to be very sloppy),
> but if you want a tight geometry convergence you will have
> to be more careful. BFGS can be quite sensitive and will 
> essentially refuse to converge if you have "ripples" due
> to a very flat potential surface unless you crank up the
> cutoff significantly.
> 
> NM> Norm-conserving and ultrasoft should matter little (unless
> NM> there is a problem with one pseudo), while GGAs should be a bit
> NM> tougher to converge than LDA.
> NM> 
> NM> Do "crystal" and "simple" really lead to a different number of
> NM> modes (10 vs 13) ??
> 
> those are obviously only the lowest few modes. 
> it is a 9x9x6 angstrom cell!
> 
> cheers,
>    axel.
> 
> NM> 
> NM> 			nicola
> NM> 
> NM> 
> NM> John Kendrick wrote:
> NM> > Hi,
> NM> > 
> NM> > I have been doing some calculations recently on molecular crystals. 
> NM> > Because of our interest in the lowest phonon frequencies I have been 
> NM> > concerned that translational invariance is maintained as far as 
> NM> > possible.  To this end I have increased the accuracy of the optimisation 
> NM> > and the scf convergence, to a point where I cannot see how to converge 
> NM> > the calculation any tighter.  I noticed that the norm conserving 
> NM> > pseudopotentials were having greater difficulties than the others.  In 
> NM> > particular for one molecule I have seen significant differences between 
> NM> > the different projection schemes that ensure translational invariance. 
> NM> > Particularly the intensities in the infrared a very different.  Below is 
> NM> > some output from a representative calculation.
> NM> 
> NM> ---------------------------------------------------------------------
> NM> Prof Nicola Marzari   Department of Materials Science and Engineering
> NM> 13-5066   MIT   77 Massachusetts Avenue   Cambridge MA 02139-4307 USA
> NM> tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu
> NM> _______________________________________________
> NM> Pw_forum mailing list
> NM> Pw_forum at pwscf.org
> NM> http://www.democritos.it/mailman/listinfo/pw_forum
> NM> 
>