The ABINIT Users Mailing List ( CLOSED )

[Xavier Gonze <gonze@pcpm.ucl.ac.be>]

homegchiche@yahoo.fr wrote:
> dear users of Abinit,
> Actualy I'm trying to studies the physicals properties of nickel using abinit code with LDA 
> pseudopotential generated with the FHI98. the result of my calculation for 
the acell is 6.36 a.u
> experimental is 6.65 a.u the precison is 4% wicth not tolerable!! and 
0.52U.B,
> the experimental is 0.60 U.B.
> If any one can help me in pseudopotential ( the cut-off radius for s, p, d and the radius 
> of core correction to use to generate the pseudopotential) and also what 
orbital we take as local?

Please have a look at Moroni et al Phys.Rev.B56, 15629 (1997). The LDA
error for the acell is 3%. You get 4% discrepancy, but the difference with
Moroni is likely due to convergence parameters or different LDA and/or 
pseudopotential.
So, despite the fact that you do not like a discrepancy of 3%-4%, it is the 
LDA
answer ! GGA is much better.
(Gian-Marco Rignanese gave me the ref)
Xavier