The ABINIT Users Mailing List ( CLOSED )

[Miguel Angel Salvado Sanchez <mass@sauron.quimica.uniovi.es>]

        Dear Xavier,
        Thanks to solve my question.

        Steve, seems the prayers have been granted, :-)

        Miguel


Miguel Angel Salvadó Sánchez
Dpto. Química Física y Analítica
Universidad de Oviedo
E-mail: mass@sauron.quimica.uniovi.es

-----Mensaje original-----
De: Xavier Gonze [mailto:gonze@pcpm.ucl.ac.be]
Enviado el: miércoles, 22 de octubre de 2003 9:00
Para: forum@abinit.org
Asunto: Re: [abinit-forum] all electron


Dear Miguel,

Miguel Angel Salvado Sanchez wrote:
>       Hello!
>
>       Thank you for the answers.
>       My question was specifically for methods (codes) that solve the 
> Kohn-Sham
> equations
>       without using pseudopotentials or other frozen-core methods.
>       In some moment I have thinked that this could be equivalent in ABINIT
>       to use some odd kind of pseudopotential. For first row elements may be
>       not so unreasonable to use an all-electron method with plane waves as
basis
> functions.

You can use indeed all-electron potentials, with an extremely small smearing
at the origin. The easiest is to enter a Goedecker-type potential file,
with only the local part, see e.g.
~ABINIT/Psps_for_tests , the files 01h.bare and 02he.bare .
I give an example for C below (this might be called 06c.bare).
It is smoothed within 0.001 Bohr of the origin.
Note that for such a potential, you will need extremely high cut-off energy
to be converged.

Xavier

---------------------------------------------------------------------------

Bare (erfc) Goedecker-Teter-Hutter  Wed May  8 14:27:44 EDT 1996
6   6   960508                     zatom,zion,pspdat
2   1   0    0    2001    0.       pspcod,pspxc,lmax,lloc,mmax,r2well
0.0010000  0   0  0  0     rloc, c1, c2, c3, c4
0 0 0                              rs, h1s, h2s
0 0                                rp, h1p