The ABINIT Users Mailing List ( CLOSED )

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

Huanggy04, I think you underestimate the power of abinit!

If you consider the example of ZnO, that you mentioned, the paper of 
Michael Rohlfing tells you that you should absolutely treat as valence 
states the semicore 3s 3p electrons.
That means that you should have a pseudopotential with
1s, 2s , 2p in the core
and
3s, 3p, 3d, 4s in the valence.

The main problem here is that you will have to create your own 
pseudopotential! Because in the pseudopotentials you find usually on the 
web site of abinit, the valence is restricted to the very last electrons 
3d and 4s.

If you do so, everything is going to be allright.

There exist a whole series of paper concerning this type of materials:
_ my paper on Cu2O, kindly cited by Xavier.
_ the paper of Patrick Rinke, I mentioned in my previous mail
_ a paper of the group of Steven Louie on ZnS and ZnSe, PRB 66 195215 (2002)
_ a paper of the team of Kotani and van Shilfgaarde PRB 66 125101 (2002)


Good luck.

Fabien



huanggy04@mails.tsinghua.edu.cn wrote:
> Thanks.I have read that paper.
>
> Though it was printed in 1995,the conclusion "a straightforward application of GWA to II-VI semiconductors including cationic semicore d states yields relatively poor results with deviations of more than 1eV between the calculated quasiparticle energies and experiment." may still be effective.Because I have never seen a straightforward application of GWA to II-VI semiconductors(such as ZnO)including cationic semicore d states. 
>
> I suspect that Abinit's gw method is unable to calculate the II-VI 
> semiconductors including cationic semicore d states