[Pw_forum] What's wrong in my calculation of epsilon of fcc Au?

gbliu goodluck_1982 at 163.com
Fri Jul 8 18:14:41 CEST 2011


Hi Tao,

In my scf run with shifted kmesh, it also gives out P=-1708.63 kbar, 
Etot=-65.80769644 Ry.
As for the stress (or pressure), I'm not clear whether this value is too 
large. In my opinion, the stress comes from that I just use the 
experimental lattice constant a=4.08Angstrom=7.71bohr and no volume 
optimization is done.

Because the band structure is as good as the result from wien2k, I think 
the pseudopotential file is just ok and the energy cutoff 25Ry also 
should be ok.

About the k-mesh, I just calculated shifted 15x15x15 and shifted 
30x30x30 (total 27000 kpoints should be enough, I think). But the 
results are similar, hence not good either. I don't know why.

BTW: Mailing to the adress pw_forum at pwscf.org is just ok. Then the mail 
will appear in the mailing list.

Best,

Gui-Bin

于 2011/7/8 14:40, Tao Sun 写道:
> Hi Gui-Bin,
>
> Sorry about that...
>
> I just used your nscf file as an input to do a scf calculation.  I got
> a total energy of -65.80738250 Ry.  The pressure I got is more than
> -170 GPa, which is not right. Did you get the same value in your
> shifted mesh scf?  I hope it is not because I am running from a mac...
> If the pressure is wrong, then maybe 25 Ry is a bit small for a
> norm-conserved pseudopotential, or the unshift 15x15x15 mesh is not
> dense enough.  This is just my guess. It would be a bit surprising to
> me if the problem is due to pseudo versus lapw, or there is a bug in
> epsilon.x.
>
> Best
>
> Tao






More information about the users mailing list