I am confused about when should we choose ISYM=0.
When I do geometry optimization for supercell with and without doped ions, should I choose ISYM=0?
Cause compared with pure supercell, symmetry of supercell with doped ion is different. In this case, if I
want to compare their structures and energies, should I switch off the symmetry?
Could somebody give me some instructions? Thank you very much.
Question about ISYM
Moderators: Global Moderator, Moderator
-
hongbiao
- Newbie
- Posts: 16
- Joined: Fri Dec 11, 2015 10:28 pm
- License Nr.: 5-1445
-
Neutrino
Re: Question about ISYM
Dear hongbiao,
Allowing symmetry change by using ISYM=0 is recommended in defect calculations. This helps in getting the lowest energy possible of the defective cell and also eliminate some artificial imaginary vibrational frequency. The drawback is that it is computationally more expensive (K-point grid is larger, more memory needed, ....).
Mostafa
Allowing symmetry change by using ISYM=0 is recommended in defect calculations. This helps in getting the lowest energy possible of the defective cell and also eliminate some artificial imaginary vibrational frequency. The drawback is that it is computationally more expensive (K-point grid is larger, more memory needed, ....).
Mostafa
-
hongbiao
- Newbie
- Posts: 16
- Joined: Fri Dec 11, 2015 10:28 pm
- License Nr.: 5-1445
Re: Question about ISYM
Dear Neutrino,
Thank you for your reply. I have ever tried to switch off symmetry (ISYM=0) for optimization of 2*2*2 supercell with doped ion.
The doping concentration is relatively low. Finally, in the CONTCAR, lattice constants of the supercell a≠b≠c. I am not sure about
its reasonability. What's your opinion on this?
Thank you.
Thank you for your reply. I have ever tried to switch off symmetry (ISYM=0) for optimization of 2*2*2 supercell with doped ion.
The doping concentration is relatively low. Finally, in the CONTCAR, lattice constants of the supercell a≠b≠c. I am not sure about
its reasonability. What's your opinion on this?
Thank you.
-
Neutrino
Re: Question about ISYM
Dear Hongbiao,
It seems you are using ISIF=3. It is common to fix cell size and shape (ISIF=2) if one is interested in defects in the dilute limit. The idea is that diulte defect concentration should not change the volume or shape of the simulation cell. In that case using ISYM=0 is also appropriate.
IF you want to simulate high concentration which also means that your result is valid only for the single concentration you are simulating in the supercell, then you have to use ISIF=3. But then it is up to you to decide maintaining the symmetry of the crystal (ISYM=2) at the expense of not achieving the lowest formation energy, or sacrificing the crystal symmetry (ISYM=0) but obtain lowest formation energy. Which one is better, you have to decide.
It seems you are using ISIF=3. It is common to fix cell size and shape (ISIF=2) if one is interested in defects in the dilute limit. The idea is that diulte defect concentration should not change the volume or shape of the simulation cell. In that case using ISYM=0 is also appropriate.
IF you want to simulate high concentration which also means that your result is valid only for the single concentration you are simulating in the supercell, then you have to use ISIF=3. But then it is up to you to decide maintaining the symmetry of the crystal (ISYM=2) at the expense of not achieving the lowest formation energy, or sacrificing the crystal symmetry (ISYM=0) but obtain lowest formation energy. Which one is better, you have to decide.