[Pw_forum] Pt bulk phonon calculated by PWSCF 1.2.0

[Nicola Marzari]

Dear Sampyo,


our experience with Ag(111) slabs was that:

1) you should play it really safe with the number of bands - make
sure that at every k-point the higher electronic eigenvalue is above
the Fermi energy by at least 3 times the smearing (do translate
degauss from Ry to eV)

2) you need tight tolerances both in the SCF calculation and
in the phonon calculation. We used 10^-17 in the SCF (might be overkill, 
but I woudln't go much lower), and 10^-14 / 10^-15 in the phonons. 
Remarkably, 10^-12 in the phonons was not tight enough for the slab 
calculations (I wish we had realized this earlier).

3) be conservative with the vacuum. We now use a little over
9 Angstrom - the scf and phonon runs (v 1.1.2) converge with much
less effort than using larger separations.

4) use inversion symmetry whenever possible - it kills
several "sloshing" modes and makes selfconsistency faster
(v 1.1.2).

5) use ngauss=-1 :-)

All the best,

				nicola

Hong, SamPyo wrote:
> Dear PWSCF users, 
>  
> a couple of days ago I reported a problem with the anomaly in the bulk 
> Pt phonon dispersion,
> but it was beautifully solved  according to the instruction suggested 
> below by Stefano de Gironcoli.
> Really I could reproduce accurately the anomaly in the bulk Pt phonon 
> dispersion.
>  
> Just a small question: during a phonon calculation of a slab consisting 
> several layers,
> I'm seeing the following messages.  Could any of you explain why I'm 
> getting these messages?
> Can this be a effect on the final result of phonons? 
>  
> (I say that most of the parameters I'm using are defaults.)
> ---------------------------------------------------------------------------------------------------------  
>    nbndx  =    56  nbnd   =    55  natomwfc =    56  npwx   =     696
>      nelec  =   92.00 nkb   =   121  ngl    =    7557
>  
> Possibly too few bands at point    1    .00000    .00000    .00000
>  
> Possibly too few bands at point    2    .04167    .07217    .00000
>  
> Possibly too few bands at point    8    .08333    .00000    .00000
>  
> --------------------------------------------------------------------------------------------------------
>  
> Thanks a lot.
>  
> Sampyo

---------------------------------------------------------------------
Prof Nicola Marzari   Department of Materials Science and Engineering
13-5066   MIT   77 Massachusetts Avenue   Cambridge MA 02139-4307 USA
tel 617.4522758  fax 617.2586534  marzari at mit.edu  http://nnn.mit.edu