[Wien] simple question about scf convergence of hematite supercell

[Pavel Novak]

In addition to what Laurence wrote:

look whether your electronic structure is metallic or insulating. If it is
metallic (I guess it is) use LDA+U with nsic=1, convergency will be
faster. For values of U, J look in literature, there are some papers on
Fe2O3 with LDA+U.

Pavel


_________________________________________________
Dr. Pavel Novak
Department of Magnetism and Superconductivity
Institute of Physics AS CR
Cukrovarnicka 10, 162 53 Praha 6, Czech Republic
tel: +420 2 20 318 532
e-mail: novakp at fzu.cz

On Mon, 16 Jun 2008, Laurence Marks wrote:

> The answer to this is that you have to think about the physics of your
> problem, and setup a calculation which is representative of this. If
> you pose the physics very badly in general you will get nowhere; the
> iterations may not coverge, the forces will be inconsistent and the
> results rubbish. Key things:
> 1) Read the UG carefully, as well as the other notes on the web pages
> so you understand the underlying physics.
> 2) Don't just use an RKMAX, think about the plane wave sampling
> (RKMAX/min(RMT)) as well. For instance, this should be at least 3.5
> for a "good" calculation.
> 3) Think about what the k-point sampling means. 1 k-pt for a small
> supercell is not reasonable.
> 4) Think about whether it is reasonable to run Fe2O3 without spin. Fe
> is magnetic, so you are trying to run a spin-polarized system without
> spin which is incorrect physics.
> 5) Think about whether it is reasonable to run Fe2O3 using simple GGA,
> or if this will give incorrect physics.
>
> On Mon, Jun 16, 2008 at 6:19 AM, swati chaudhury <swati at rcais.res.in> wrote:
> > Dear all Wien users,
> >                 I am having problems with convergence of supercell of
> > hematite (Fe2O3, non-magnetic).
> > I have tried by changing so many parameters like mixing factor
> > (PRATT/BROYD/MSEC with 0.05,0.01,0.10,0.40), global E-parameter (0.35,0.30),
> > K-points (1,8) etc. So I am now little bit confused that is there any other
> > parameters, which can help to convergence scf cycle.
> > Basic inputs:
> > Supercell : 2x2x1 with 10 bohr slab
> > Space group=147  (P –3)
> > 13 GGA, -9.0 cut-off value, Rkmax =5.40, Emin = -9.0, Emax = 2.50, 14 - Gmax
> > Please help.
> > Thanks in advance.
> > Regards.
> > Swati.
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >
> >
>
>
>
> --
> Laurence Marks
> Department of Materials Science and Engineering
> MSE Rm 2036 Cook Hall
> 2220 N Campus Drive
> Northwestern University
> Evanston, IL 60208, USA
> Tel: (847) 491-3996 Fax: (847) 491-7820
> email: L-marks at northwestern dot edu
> Web: www.numis.northwestern.edu
> Commission on Electron Diffraction of IUCR
> www.numis.northwestern.edu/IUCR_CED
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>