[Pw_forum] Volume Optimisation :-Reg

[Ari P Seitsonen]

Dear A Suresh,

   My standard answer, sorry for that - did you check that the value of 
'ecutwfc' is sufficient? In the Good Old Days (at FHI) we used to use 
something more like 80-100 Ry for FHI-pseudos of oxygen, and that in 
ionc relaxations-only: The 'vc-relax' requires yet-a higher value. Did you 
try to start a new calculation from the lattice+ionic geometry that you 
obtained, _without_ restarting the electronic structure (ie copying the 
new values lattice vectors and ionic coordinates into the input file and 
keeping the "restart_mode = 'from_scratch'"): If the lattice constants 
start to change a lot, the cut-off energy is too low. Or you look at the 
value of the stress at the end of your previous calculation: The final 
step is done with re-generated plane waves, so if your lattice changed a 
lot and the cut-off was too low, there suddenly is a large value of 
stress.

   On the other hand I do not know about the accuracy of the pseudo 
potential for Ti: I remember there were some problems with a 
norm-conserving one, at least if the number of valence electrons is four. 
Maybe this one would work better, I do not know this one, but I would 
start checking the value of the cut-off energy.

     Greetings from Sunny Paris,

        apsi

PS Why this "exxdiv_treatment" in your input...?

-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
   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 Tue, 16 Feb 2016, Suresh A wrote:

> Dear All,
>             I have optimized the structure for anatase simple tetragonal. The optimiized lattice parameters are a= 3.754336A. c=9.193631522694A. but the experimental values are
> a= 3.785A, c=9.514A. There is big difference in c value for optimised strucutre.But the same calculation for GGA pseudo potential gives the lattice parameter values which are
> good with  experimental values. In literature  previous results for LDA has good lattice parameter values.
>  Which parameter I have to  change in order to get good results for LDA?
>   Below is my input and the version i have used is 5.2.1
> 
>  &CONTROL
>                        title = 'anatase' ,
>                  calculation = 'vc-relax' ,
>                 restart_mode = 'from_scratch' ,
>                   pseudo_dir = '/home/suresh/GN2/',
>                       outdir ='/home/suresh/Desktop/ashahicheck/asahiopt/tmp/',
>                       prefix = 'anatase' ,
>                      tstress = .true. ,
>                      tprnfor = .true. ,
>  /
>  &SYSTEM
>                        ibrav = 6,
>                            A = 3.692,
>                            C = 9.471 ,
>                          nat = 12,
>                         ntyp = 2,
>                      ecutwfc = 60 ,
>             exxdiv_treatment = 'none' ,
>  /
>  &ELECTRONS
>  /
>  &IONS
>                   ion_dynamics = 'bfgs' ,
>  /
>  &CELL
>                cell_dynamics = 'bfgs' ,
>  /
> ATOMIC_SPECIES
>    Ti   47.86700  Ti.pz-mt_fhi.UPF
>     O   15.99940  O.pz-mt_fhi.UPF
> ATOMIC_POSITIONS (crystal)
> Ti      0.0000000   0.0000000   0.0000000
> Ti      0.5000000   0.0000000   0.7500000
> Ti      0.0000000   0.5000000   0.2500000
> Ti      0.5000000   0.5000000   0.5000000
> O       0.0000000   0.0000000   0.2066000
> O       0.5000000   0.0000000   0.5434000
> O       0.0000000   0.5000000   0.4566000
> O       0.5000000   0.0000000   0.9566000
> O       0.0000000   0.5000000   0.0434000
> O       0.0000000   0.0000000   0.7934000
> O       0.5000000   0.5000000   0.7066000
> O       0.5000000   0.5000000   0.2934000
> K_POINTS automatic
> 4 4 2 1 1 1
> 
> 
> Thanks in advance.
> 
> 
>                   With Regards,
> A.Suresh,
> Research Scholar,
> Madurai Kamaraj University,
> Madurai.
> 
>