The ABINIT Users Mailing List ( CLOSED )

[Fabien Bruneval <fabien.bruneval@polytechnique.fr>]

Dear Denis,

A crucial point in GW calculations is the semicore/valence interaction. 
The pseudopotential needed for GW may be different from the ones 
suitable at DFT level.
In particular, if non-linear core corrections are used for the DFT step, 
you may think of changing of pseudopotential in order to include the 
semicore shells in the valence for two reasons:

1) the exchange interaction between semicore and valence can be much 
different if treated in the LDA or treated exactly in the GWA.

2) the non-linear core corrections are NOT implemented in the GW part of 
abinit. There is a WARNING somewhere in the log file.

If your Germanium pseudopotential has non-linear core corrections, the 
code cannot provide correct answers.

Good luck,

Fabien Bruneval




Denis RIDEAU wrote:
> Dear abinit users:
>
>
> I calculated the germanium GWA electronic band structure using ABINT. I 
> performed LDA with theoritical lattice parameter (5.4767 A) and the GW 
> correction was applied as a post-DFT scheme.
> As expected, the GWA germanium band structure I obtained exhibit a poor 
> conduction band, in particular the direct gap is overestimated (1.8 eV). 
> TM pseudopotential and HGH pseudopotential provides similar results.
> What can I do with ABINIT to improve these results?
> Regards
>
> Denis Rideau