[gpaw-users] Ti atom GS convergence problem and wrong Magnetic moment

[Marcin Dulak]

Hi,

Jakob Blomquist wrote:
> I'm having a surprisingly hard time with a simple Ti-atom.
> The magnetic moment doesn't follow Hund's rule (should be a 3F2 state) 
> but end up with +4 if it converges at all.
> It doesn't matter much if I run with PBC = True or not or if the box is 
> symmetric or not.
>
> Is this something known?
>
> scrip run e.g:
>
> *****************
> from ase import Atoms, Atom
> from gpaw import GPAW, FermiDirac, MixerSum
> from ase.parallel import parprint
>
> Ti = Atoms('Ti')
>
> Ti.center(vacuum=6.)
> #Ti.cell[1,1] += 1.0
> #Ti.cell[2,2] += (-1.0)
>
> #mixer = MixerSum(beta=0.1, nmaxold=3, weight=1.0)
> name = 'Ti'
> calc = GPAW(txt=name+'.txt',
>              basis='dz(dzp)',
>              nbands=-4,
>              h = 0.16,
>              xc='PBE',
>              hund=True,
>              )
>
> Ti.set_calculator(calc)
> e = Ti.get_potential_energy()
>
> *****************
>
>   
the system is tricky. Use fixmom=True (as you do) and get rid of pulay 
DIIS mixing as much as possible.
The attached example converges also with eigensolver='rmm-diis' to the 
same energy (-1.33 eV),
however with eigensolver='cg' and mixer = MixerSum(beta=0.25, nmaxold=1) 
to -1.22 eV.
GPAW 0.8.0.8092.
I'm not sure we are getting the right state. One has to be careful about 
converging such systems.
Any news about our old ticket: 
https://trac.fysik.dtu.dk/projects/gpaw/ticket/217 ?

We will hopefully at some point test all isolated atoms in GPAW thanks 
to the new ase tool
https://wiki.fysik.dtu.dk/ase/ase/cmdline.html

Best regards,

Marcin

> ASE rev. 3.5.0.2113
> GPAW rev. 0.8.0.8008
>
>   


-- 
***********************************
 
Marcin Dulak
Technical University of Denmark
Department of Physics
Building 307, Room 229
DK-2800 Kongens Lyngby
Denmark
Tel.: (+45) 4525 3157
Fax.: (+45) 4593 2399
email: Marcin.Dulak at fysik.dtu.dk

***********************************

-------------- next part --------------
A non-text attachment was scrubbed...
Name: test09.py
Type: text/x-python
Size: 574 bytes
Desc: not available
Url : http://listserv.fysik.dtu.dk/pipermail/gpaw-users/attachments/20111018/f323d5c1/attachment-0001.py