[gpaw-users] How to get reasonable magnetic moments ?

[Jens Jørgen Mortensen]

On Fri, 2010-11-05 at 14:25 +0100, Wei-Bing Zhang wrote:
> 2010/11/5 Jens Jørgen Mortensen <jensj at fysik.dtu.dk>:
> > On Fri, 2010-11-05 at 11:26 +0100, Wei-Bing Zhang wrote:
> >> Hi all
> >>
> >> Recently, I try to do some test calculations about bulk Nickel using
> >> gpaw. And I found some strange results about magnetic moments. The
> >> magnetic moment change largely with smearing parameters of FermiDirac
> >> method and seems to be smaller than the results in literature.
> >
> > What is the literature value?  What functional are you using and what
> > lattice constant?
> 
> Thanks for this reply. It seems that i made a mistake. I compared the
> result(0.39) obtained using smearing parameters 0.2 to the
> reference(0.63, PBE, a:3.522, vasp code) and then get the previous
> wrong statement(possibly, a smearing parameter 0.2 is normal value in
> other code ). Sorry.
> The result(0.61) obtained using small smearing
> parameter(Kpoints:16x16X16,h=0.18, The default parameter in ASE) seems
> to be comparable.I also check the convergence about k-points which
> suggest 16x16x16 is enough.
>  Could you give me some simple clue about variance of magnetic moment
> with respect to smearing parameters? Thanks .

Let's say the majority spin has a large peak in the DOS right below the
Fermi level.  If the width of the Fermi-Dirac smearing functions is
small, then the weight of the peak will be close to one and if the width
is very large, the the weight will be close to 0.5.

Jens Jørgen

> >>  And I
> >> check the gpaw-setup and find that the PAW potential did not include
> >> Non linear core correction, which should be important to magnetic
> >> system.
> >
> > The PAW method works with the full core density, so it doesn't need a
> > NLCC - those are for pseudo potentials.
> Good, Thanks.
> 
> 
> 
> >
> >> parameter moments
> >> 0.01 [ 0.61319955]
> >> 0.02 [ 0.61328128]
> >> 0.03 [ 0.61175437]
> >> 0.04 [ 0.6088879]
> >> 0.05 [ 0.60540485]
> >> 0.1 [ 0.58561697]
> >> 0.15 [ 0.53254744]
> >> 0.2 [ 0.39047028]
> >>
> >>
> >> wei-bing zhang
> >> ----------------------------------------------------------------------
> >> _______________________________________________
> >> gpaw-users mailing list
> >> gpaw-users at listserv.fysik.dtu.dk
> >> https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
> >
> >
> 
> Wei-Bing Zhang