[Wien] about oscilations of scf procedure

[Hua Wu]

what type of solution you have got, metallic or insulating?
if metallic, how wide the conduction band is ?
and what is the point-group symmetry seen by the Re ion?
if you just want to remove the oscilation, you may not need
to resort to inclusion of spin-orbit coupling--which is 
another issue. 

Since the oscilation occurred in your LDA+U calculations,
you firstly need to lower symmetry to allow Hubbard
bands to split, e.g, for a half-filled doublet. 
As for SOC, you can pls refer to the manual
(the initso is helpful !!). You may choose the easy-
magnetization direction to be z-axis, ie. (hkl=001),
as you may not calculate the magnetic anisotropy. -- H. Wu
  
On Thursday 18 August 2005 15:32, zhoukj wrote:
> Dear Wu,
>    I have reduce the mixing parameter and the interation seems
> to oscilate less. So I think spin-orbit should be added to the
> calculation. But i don't know how to set the case.inso due to
> the input file contains h,k,l which describing the direction
> of magnetization.Idon't know what's the direction of
> magnetization? Can you show me how to set .inso for Re? Thank
> you.
>
>
> Sincerely,
>
> ======= 2005-08-18 11:02:00 Original Message：=======
>
> >I would suppose the oscilation occurs for a metallic
> > solution. It seems due to the partial occupation of the
> > narrow Re 4f bands. Firstly, you can try to use a small
> > mixing parameter like 0.01-0.05 or to increase the k-points.
> > Another option is to lower the crystal-field symmetry (e.g,
> > set cubic to orthogonal, but keep the atomic
> >positions), or to include the spin-orbit coupling which
> >is normally important for Re,
> >finally to lift the orbital degeneracy. As a result,
> >the Hubbard U may help you to find a stable solution
> >with some orbitals (others) fully occupied (unoccuoied).
> >
> >On Thursday 18 August 2005 05:02, zhoukj wrote:
> >> Dear all
> >>    I'm now using LDA+U do the calculation of rare earth
> >> oxides. When I put 4f orbital as the correlated orbital, 
> >> and run scf, I found that the scf procedure oscilates, but
> >> it is convergent. What's the problem?  Should I add
> >> spin-orbit in the calculation? How to set it? Thanks a lot.
> >>
> >>
> >>
> >>
> >> Sincerely yours,
> >> Zhou Ke jin
> >>
> >>
> >> -----------------------------------------------------------
> >>--- ----
> >>
> >>                       Beijing Synchrotron Radiation
> >> Facility
> >>
> >>                       IHEP,CAS,100049   P.O.Box  918  2-7
> >>
> >>                       Tel:+86+10+88236375
> >> mail:Kejin.Zhou at ihep.ac.cn
> >>
> >>                       10:31:57    2005-08-18
> >> -----------------------------------------------------------
> >>--- ----
> >>
> >>
> >>
> >>
> >> _______________________________________________
> >> Wien mailing list
> >> Wien at zeus.theochem.tuwien.ac.at
> >> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >
> >--
> >Dr. Hua WU
> >
> >II. Physikalisches Institut
> >der Uni. zu Koeln,
> >Zuelpicher Str. 77,
> >D-50937 Koeln
> >Germany
> >
> >Email: wu at ph2.uni-koeln.de
> >Tel: +49-221-4703599
> >Fax: +49-221-4705178
> >_______________________________________________
> >Wien mailing list
> >Wien at zeus.theochem.tuwien.ac.at
> >http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
> = = = = = = = = = = = = = = = = = = = =
>
>
>
>                            zhou ke jin
>                        zhoukj at ihep.ac.cn
>
>     Beijing Synchrotron Radiation Facility
>
>     IHEP,CAS,100049   P.O.Box  918  2-7
>
>     Tel:+86+10+88236375
>
>     21:28:40    2005-08-18
>
>
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

-- 
Dr. Hua WU

II. Physikalisches Institut
der Uni. zu Koeln,
Zuelpicher Str. 77,
D-50937 Koeln
Germany

Email: wu at ph2.uni-koeln.de
Tel: +49-221-4703599
Fax: +49-221-4705178