The ABINIT Users Mailing List ( CLOSED )

[Thomas Maxisch <Thomas.Maxisch@epfl.ch>]

Dear Vincenzo,

> Dear all, on the 3d valence electrons issue
> for GaN (4d for In) I have personally concluded,
> after a great deal of experiments [started with
> PRB 47, 13353 (93) ;)] by myself and valuable
> coworkers including e.g. F. Bernardini, that to stay
> within reasonable cutoff limits and still get high quality
> results, ultrasofts are needed. We have been using various
> uspp versions since 1993 in the nitrides trade, with some
> success. PAW, obviously, is also an answer (coming up in
> abinit, I understand). PAW and LDA (VASP) gives GaN bulk
> l.p.'s within 0.5 % of expt.

I totally agree with you. The problem is that Abinit cannot use any uspp 
and the paw is in the implementation phase. And I need results as soon as 
possible.

> This notwithstanding, you can get around it with ncpp,
> occasionally even with non-linear corrections. This is
> testified by the excellent past and current work by
> Joerg Neugebauer, now at FHI Berlin. He is the one who
> is in the know about nitrides with ncpp's, and he may not
> be on this forum...

I think that Martin Fuchs is quite close to him and I also think that the 
group of Berlin know how to do these things with ncpp's. I asked M. Fuchs 
for the potentials and then I will make some calculations inorder to look 
if the accuracy of the pp's are sufficent. Well, I won't calculate any 
bandgaps with GW, so I hope that for my purposes the pp's will be sufficent.

Best wishes,
Thomas


--
__________________________________________________________________________

Thomas Maxisch
EPFL FSB-ITP/CPNMC
PHB-Ecublens
CH-1015 Lausanne
Switzerland
Tel : +41 (0) 21 693 45 16
Fax : +41 (0) 21 693 44 44
Email : Thomas.Maxisch@epfl.ch
__________________________________________________________________________