[Pw_forum] pbe normconserving pseudopotentials for Fe

[Ludwig, Stephan]

Hi,

i want to do scf calculations on BaFe2As2. In order to find a suitable ecutwwc I look for a convergency for 

the total energy with respect to the ecutwfc.

Using an ultrasoft pseudopotential for iron (Fe.pbe-nd-rrkjus.UPF) I achieved convergency without problems. 

Unfortunately I whant to use the Yambo program to calculate linear response. Therefore I need to use Normconserving Pseudos.

By using the normconserving pseudo Fe.pbe-sp-mt_gipaw.UPF I can't achieve convergency, even not at ecutwfc of 460 (although I use the kgrid 8 8 8).

I used this Pseudopotential already to simulate CaFe2As2. In this case I achieved convergency (but it was on another computer) but with a huge ecutwfc of 260.

What's wrong with this Pseudopotential? Why is it not suitable?



Whould it be a better idea to use the Normconserving pseudo Fe.pbe-mt_fhi.UPF? The Pseudopotential Fe.pbe-sp-hgh.UPF would not be suitable since also I whant to have PDOS.

Here you see my input-file





&CONTROL
 title = 'BaFe2As2_20K' ,
 calculation = 'scf' ,
 wf_collect = .false. ,
 outdir = './' ,
 wfcdir = './',
 pseudo_dir = '/home/espresso-5.1/pseudo/' ,
 prefix = 'BaFe2As2_20K' ,
 
 /
 &SYSTEM
 ibrav = 10,
 A = 5.5742 ,
 B = 5.6146 ,
 C = 12.9453 ,
 cosAB = 0 ,
 cosAC = 0 ,
 cosBC = 0 ,
 nat = 5,
 ntyp = 3,
 ecutwfc = 460 ,
 occupations = 'smearing' ,
 degauss = 0.02 ,
 smearing = 'gaussian' ,
 exxdiv_treatment = 'gygi-baldereschi' ,
 /
 &ELECTRONS
 conv_thr = 1.0D-8 ,
 /
ATOMIC_SPECIES
 Ba 137.33000 Ba.pbe-mt_fhi.UPF 
 Fe 55.84700 Fe.pbe-sp-mt_gipaw.UPF 
 As 74.92200 As.pbe-hgh.UPF 
ATOMIC_POSITIONS angstrom 
 Ba 0.000000000 0.000000000 0.000000000 
 Fe 1.393550000 1.403650000 3.236325000 
 Fe 1.393550000 4.210950000 3.236325000 
 As 0.000000000 0.000000000 4.580047140 
 As 0.000000000 0.000000000 8.365252860 
K_POINTS automatic 
 8 8 8 0 0 0 





Thanks and Regards



Stephan 




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20150415/8566338e/attachment.html>