[Wien] SP+SO calculation

[Shunli Shang]

Dear Prof. Peter Blaha and wien2k users,

For my case, Fe6C1(SG_162), the "symmetso" doesn't change the symmetry.
I tried your suggestions, but once the SO coupling included, the program
say: Energy to separate semicore and valencestates:-999.00000.

To yushan: I tried to set Fe-3s as core state, but I got worse results:
about 0.015 core charge leaks out of MT-sphere, and the QTL-B value is about
10 for the SP+SO calculation.

I want to attach the structure file and inst file of Fe6C1, hope someone
could help to solve the problem. Thanks.
Other parameters I used are as follows, besides the default values.
ENERGY to separate core and valence states: -7Ry.
RK=8.0
K points in full BZ: 1500.
mixer factor: 0.10
It takes about 50 minutes and one hour for one cycle on the Athlon XP3000+
machine for the SP and SP+SO calculation respectively.

Best regards,
Shunli Shang

----- Original Message ----- 
From: "Peter Blaha" <pblaha at zeus.theochem.tuwien.ac.at>
To: <wien at zeus.theochem.tuwien.ac.at>
Sent: Wednesday, July 21, 2004 8:37 AM
Subject: Re: [Wien] SP+SO calculation


> > SP+SO calculation (some strange results shown in case.output2up file for
the
> > first cycle of SP+SO calculation after the convergent SP calculation).
> > number of occupied bands:         103
> >    highest energy:  0.873929976078531
> >   Bandranges (emin - emax):
> > band           1  -9.98505085982580       -5.64126602440761
> > band           2  -8.59789957521298       -5.64059028470795
> >
> This tells me, that already in the first iteration the eigenvalues are
> completely wrong. You should not have a "band" from -9 to -5 Ry. This is
> an almost dispersionless 3s state.
> Most likely you find in the scf file even other problems
> like "eigenvalues below XX Ryd", QTL-B's,....
>
> Possible cuases I can only guess:
>
> Has the symmetry been changed by symmetso ?
> If yes, maybe it did not transfer the CLM's properly.
>
> Run symmetso, but the try to restart with   x dstart -c (and -up/dn)
> Are the bands then ok ?
>
> If not, check with nn for proper distances,....
>
> I'll be gone on vacations soon, so I cannot give any further help.
> Contact me, if your problem still persists in 3 weeks.
>
>                                       P.Blaha
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Fe6C1.inst
Type: application/octet-stream
Size: 209 bytes
Desc: not available
Url : http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20040721/2557dbca/Fe6C1.obj
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Fe6C1.struct
Type: application/octet-stream
Size: 1963 bytes
Desc: not available
Url : http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20040721/2557dbca/Fe6C1-0001.obj