[Pw_forum] Convergence problem of SnO2 unit cell

[Ari P Seitsonen]

Dear A Bahrami,

   Some observations:

  - You are not optimising bulk SnO2 but a monolayer of SnO2, right?

  - Is the system an insulator? Otherwise you can the broadening of the 
occupation numbers, for example

   occupations = 'smearing'
   smearing    = 'fermi-dirac'
   degauss     = 0.0036749326

in the namelist '&SYSTEM'

  - The "mt_fhi" refers to the Troullier-Martins pseudo potentials (and 
Fritz-Haber-Institut, where I did my PhD, partially in the same office 
with Martin Fuchs who was our "maestro of pseudo potentials"): They are 
norm-conserving, so no need to specify both 'ecutwfn' and 'ecutrho'; the 
latter can remain at its default == 4 * ecutwfn

  - The value for 'ecutwfn' is far too low: I would start from something 
like 80-120 Ry

  - In systems with vacuum I usually get faster convergence with the mixing 
algorithm 'local-TF' than the dfault

  - I would never increase the number of 'electron_maxstep' beyond 200 if I 
do not know that the system has some difficult electronic structure 
(magnetism etc), and usually if the default value of 100 is reached there 
is something wrong in the input

     Greetings from Montrouge,

        apsi

PS My personal opinion: Why use "CELL_PARAMETETERS"/"ibrav = 0" when the 
cell is so simple? "ibrav = 8" and "a"/"b"/"c" would do the same job

-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
   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 Sun, 15 Feb 2015, Aidin Bahrami wrote:

> 
> Dear all,
> 
> I want to optimize the geometry of an SnO2 unit cell ( The geometry and cell
>  dimensions). But, it doesn't converge even after 1000 steps. Could
>  you give me some suggestions? Any suggestions are welcome!
> 
> I am very thankful any to solve this problem.
> 
> Attached below is the input file:
> &control
> calculation='relax'
>  prefix = 'SnO2UC' ,
>  pseudo_dir  = "/home/aidin/Work/Pseudo",
>  outdir      = "/home/aidin/Work/tempdir",
>  /
>  &system
>  ibrav=  0,
>  nat=  6,
>  ntyp= 2,
>  ecutwfc =33.0,
>  ecutrho  =300.0,
>  occupations='fixed',
>  /
>  &electrons
>  diagonalization='david',
>   conv_thr    = 1.D-7,
>   mixing_beta = 0.3,
>   electron_maxstep=5000,
> /
> &IONS
>   ion_dynamics      = 'bfgs',
> /
> ATOMIC_SPECIES
>  Sn 195.08 Sn.pbe-mt_fhi.UPF
>  O  15.9994 O.pbe-mt_fhi.UPF.UPF
> CELL_PARAMETERS {angstrom}
> 2.64660710    0.00000000    0.00000000
> 0.00000000    6.38890841    0.00000000
> 0.00000000    0.00000000   15.00000000
> ATOMIC_POSITIONS {angstrom}
>  Sn                   0.51714893    3.87377899    1.80974824
>  O                     2.11190545    3.86421764    3.11002137
>  O                     0.53756837    1.81823266    1.75930720
>  O                     0.49229597    5.92499155    1.85643211
>  O                     2.11263443    3.92197449    0.52041771
>  Sn                   2.14332255    0.54183064    1.73115747
> K_POINTS {automatic}
>  10 10 1 0 0 0 
> 
> Thanks in advance
> 
> A.Bahrami
> 
> Ph.D of chemistry
> University of Tarbiat modares
> Tehran
> Iran
> 
>