[Pw_forum] Fermi energy in 'scf' and 'nscf'(XQ Wang)

[Prasenjit Ghosh]

Dear Wang,

In principle both the scf & nscf should give exactly the same value of Fermi
energy if your calculations are converged w.r.t. the k-point sampling
(since evaluation of Fermi energy involves numerical integration in
k-space). The fact that that the two procedures give different energies
suggest that the k-point mesh is not converged. Since these are numerical
integrations, the more the no of k-points the more accurate is the value of
the integral. Therefore for Guoying's case, the value of Fermi energy
obtained from scf is more accurate than that from nscf as suggested by
Hande.

If the difference in Fermi energy is quite large, what I would suggest is to
increase the no. of k-points in the nscf cal. & redo it.

With regards,

Prasenjit.

2008/8/13 wangxinquan <wangxinquan at tju.edu.cn>

>  >>Dear Guoying
> >>
> >>             scf:           k-mesh grids   10*10*10  (282kpoints in
> >> *.scf.out file)
>
> >>           nscf:           we give 121 kpoints along high symmetry direction
> >>
> >>
> >>In order to find the Fermi energy, the code performs an integration in
> >>k-space and this integration is of course numerically performed over the
> >>kpoints you provide. In the nscf run, you provide a highly heterogeneous
> >>set of kpoints so the Fermi energy you get is not the same as the scf
>
> >>solution where the BZ is sampled more evenly (and also more densely in your
> >>case). As a result, you should trust the scf result.
>
> Dear Guoying,
>
>     In my opinion, you should use the nscf results if you want to calculate
> the band
> structure. According to the example 08 of espresso-4.0.1, the calculation
> of DOS for Ni follows three steps:1. pw.x (scf);2.pw.x(nscf);3.dos.x(dos).
> The calculation
> of nscf modifies the wavefunctions obtained by scf calculations. You will
> see
> that the k-points were identified manually in scf while they were
> automatically selected in
> nscf. So the Fermi energy used in band structure calculation should be the
> result from nscf.
>     I'm not sure that the above suggestions are absolutely right. It's just
> my limited experience.
>  You'd better investigate the README file in the example 08 by yourself.
>     Hope this help.
>
> Cheers,
> XQ Wang
>
>
> =====================================
>
> X.Q. Wang
>
> wangxinquan at tju.edu.cn
>
> School of Chemical Engineering and Technology
>
> Tianjin University
>
> 92 Weijin Road, Tianjin, P. R. China
>
> tel:86-22-27890268, fax: 86-22-27892301
>
> =====================================
>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>


-- 
PRASENJIT GHOSH,
POST-DOC,
ROOM NO: 265, MAIN BUILDING,
CM SECTION, ICTP,
STRADA COSTERIA 11,
TRIESTE, 34104,
ITALY
PHONE: +39 040 2240 369 (O)
+39 3807528672 (M)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20080813/18219ab0/attachment.html>