[Pw_forum] About ecutwfc and ecutrho again

[Axel Kohlmeyer]

On 12/21/07, Huiqun Zhou <hqzhou at nju.edu.cn> wrote:

> Thanks, Nicola and Stefano,
>
> Actually, I was thinking the choice should be the lagest or lagest plus
> alpha.
> The pseudopotentials I'm testing are USPP for iron, chromium, oxygen and
> others, and one of the materials I'm interested is chromite (FeCr2O4). I
> found
> ecutwfc for iron and chromium is indeed around 35 Ry for convergence, but
> over 55 Ry for oxygen. I haven't checked whether I did anything wrong when
> creating potential for oxygen, but it's consistent with the results of
> convergence

if you are using the "bm" vanderbilt oxygen potentials, those were created
with a wavefunction cutoff of 25 to 35 rydberg in mind.

> test for chromite, in which ecutwfc should be set to 55 Ry (ecutrho=600) for
> convergence.
>
> Please comment.

as was already noted. the convergence of the total energy is much
slower as other properties (e.g. forces). so you can get an accurate result
for geometry optimization with much smaller cutoffs for both wavefunction
and density.

to test this, you can build a (distorted) geometry and monitor forces
and/or stress tensor etc. pp.

cheers,
    axel.





>
>
> Dr. Huiqun Zhou
> @Earth Sciences, Nanjing University, China
>
>
> ----- Original Message -----
> From: "Nicola Marzari" <marzari at MIT.EDU>
> To: "PWSCF Forum" <pw_forum at pwscf.org>
> Sent: Friday, December 21, 2007 12:52 PM
> Subject: Re: [Pw_forum] About ecutwfc and ecutrho again
>
>
> >
> >
> > Roughly speaking, the higher ecutwfc among the elements
> > considered.
> >
> > Are these ultrasoft or norm conserving ? If the latter,
> > ecutrho could have been safely kept to 240 Ry.
> >
> > If the former, my rule of thumb is that rarely a cutoff
> > greater than 30 Ry is needed, unless subtle magnetic/spin
> > propertes are sought. Thi sis to say that absolute convergence of
> > the energy is not the most useful criterion - often one looks
> > at lattice parameter, bulk modulus, or phonon frequencies as
> > good indicators of convergence.
> >
> > nicola
> >
> > Huiqun Zhou wrote:
> >> Dear PWscfers,
> >>
> >> I'm creating and testing pseudopotentials for severval
> >> elements myself. I plotted the E-ecutwfc graphs for
> >> each element, in which the ecutwfc was ranging from 20
> >> to 60 with increament of 5 Ry and ecutrho was set to 600.
> >> If take 1 mRy as convergence criterion, I find that the
> >> total energy for some elements will be convergent at
> >> ecutwfc = 35 Ry while some are around ecutwfc = 55 Ry.
> >>
> >> If the material I'm interested in is comprised from elements
> >> with both lower and higher ecutwfc, what ecutwfc should I
> >> set for the material, the higher one, the higher one plus
> >> alpha, or a compromise between the lower and higher?
> >>
> >> Thnaks!
> >>
> >>
> >> Dr. Huiqun Zhou
> >> @Earth Sciences, Nanjing University, China
> >
> >
> > --
> > ---------------------------------------------------------------------
> > Prof Nicola Marzari   Department of Materials Science and Engineering
> > 13-5066   MIT   77 Massachusetts Avenue   Cambridge MA 02139-4307 USA
> > tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu
> > _______________________________________________
> > Pw_forum mailing list
> > Pw_forum at pwscf.org
> > http://www.democritos.it/mailman/listinfo/pw_forum
> >
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>


-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
  Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.