The ABINIT Users Mailing List ( CLOSED )

[David Clatterbuck <clatterb@uclink4.berkeley.edu>]

I had previously done some testing of Cu using VASP and the supplied 
ultrasoft pseudopotential (both LDA and GGA). The numbers below are from fits 
of the energy to the murnahan equation of state with 9 data points. They are 
are well converged with respect to kpoint sampling, and I used the 
tetrahedral method so there is no smearing parameter. The ecut was set to 
130% of the value suggested for the pseudopotential.  

Cu - Bulk Modulus
LDA: 189 GPa
GGA: 142 GPa

So it doesn't appear that using ultrasoft-pseudopotentials will help. Based 
on these I would also guess that the other calculation you reference for Cu 
used GGA rather than LDA. As you may know the LDA tends to overbind and 
produces elastic constants larger than experiment while the GGA often 
corrects this (though it can overcorrect in some cases).  

David Clatterbuck
Lawrence Berkeley National Lab



On Tuesday 25 February 2003 03:11 am, you wrote:
> Dear ab init users,
>
> I interested in calculating the elastic constants of Cu and Au (both simple
> fcc metals). For that I used the Troullier-Martins pseudopotentials
> available on the web, a cutoff of 40 Ry, a grid of 8 x 8 x 8 k points and
> the input file given below.
> The results for the bulk moduluae I have got (in GPa) are:
>
>                  Cu    Au
> My calculation  180   188
> Experiment      142   180
> Other calc.     144   180
>
> So I am getting a value of B for Cu that is much higher than previous ab
> initio calculations.
>
> Any ideas about what am I doing wrong?
>
> Thank you very much
>
>
> Gustavo Barrera
> ISIS, Rutherford Appleton Laboratory