[Pw_forum] optimization using fully relativistic pp

[Ari P Seitsonen]

Dear Sohail Ahmad,

   Which convergence do you mean: Electronic self-consistency or the 
geometry (ions and cell)? About the former I do not know much - I tried 
your input, with few k points it seems to converge; only that the value of 
'degauss' is quite small - but some notes on the input otherwise:

   You have chosen a _very_ high cut-off 'ecutwfc'; considering that even 
if you are doing the relaxation of the cell and thus need a well converged 
stress tensor, this value look very high. Your system looks 
nearly-hexagonal but is not just quite - is this intentional? If you 
decide to go precisely hexagonal cell I would recommend using a k point 
set containing the later Gamma-point (shift "0 0" of the mesh along the 
first two reciprocal directions), as the original Monkhorst-Pack k point 
mesh (corresponding to "1 1") yields a lower symmetry than the hexagonal 
three-fold one (discussed eg in the Appendix of my PhD thesis). But what I 
would change in any case is to use the keyword 'cell_dofree' in your input 
in order to avoid that the cell vector is modified between the layers, or 
along the [001] direction.

     Greetings from Montrouge/France,

        apsi

-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
   Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
   Ecole Normale Supérieure (ENS), Département de Chimie, Paris
   Mobile (F) : +33 789 37 24 25    (CH) : +41 79 71 90 935


On Sat, 5 Dec 2015, Sohail Ahmad wrote:

> Dear QE users
> I am finding difficulty in getting convergence while optimizing structure by vc-relax 
> using fully relativistic potential.
> Is there any special trick while using fully relativistic pseudo potential ?
>
> ---------------------------------------------------------------------------
> &control
>  calculation = 'vc-relax',
>  restart_mode = 'from_scratch',
>  prefix = 'PtS',
>  pseudo_dir = '/home/sohail/Espresso/espresso-5.0.1/pseudo',
>  outdir = './OUT',
> /
> &system
>  ibrav = 0,
>  nat = 3, ntyp = 2,
>  ecutwfc = 150,
>  ecutrho = 600,
>  nbnd = 25,
>  occupations = 'smearing', smearing = 'gaussian', degauss = 0.001,
>  noncolin = .true.,
>  lspinorb = .true.,
>  starting_magnetization(1) = 0.1d0,
> /
> &electrons
>  mixing_beta = 0.3,
>  conv_thr = 1.0d-9,
> /
> &ions
> ion_dynamics = 'bfgs',
> /
> &cell
> cell_dynamics = 'bfgs',
> press = 0.d0,
> press_conv_thr = 1.2d0,
> /
> ATOMIC_SPECIES
> Pt  195.08  Pt.rel-pbe-n-kjpaw_psl.0.1.UPF
> S   32.06  S.rel-pbe-n-kjpaw_psl.0.1.UPF
> ATOMIC_POSITIONS  angstrom
> Pt  0.00000000  0.00000000  0.00000000
> S   1.80374872  1.02665089  1.24494264
> S  -1.80374872 -1.02665089 -1.24494264
> CELL_PARAMETERS {hexagon}
> 3.37036749  5.86580486  0.00000000
> -6.81427354 0.03307209  0.00000000
> 0.00000000  0.00000000  28.34590000
> K_POINTS AUTOMATIC
> 12 12 1 1 1 1
>
>
>
> --------------------------------------------------------------------- 
>
> Sohail Ahmad
> King Khalid University
> Abha, Saudi Arabia 
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum