Dear all,
By using ISYM tag it is possible to switch on/off symmetry for relaxations. but I wonder if it is possible to constrain some of the symmetries but relaxing the others. (Such as imposing a symmetry by giving space group which is different than the group detected by VASP automatically or in an other way?)
Thank you in advance...
calculations with constrained symmetry
Moderators: Global Moderator, Moderator
-
edurgun
- Newbie
- Posts: 10
- Joined: Thu Sep 29, 2005 11:52 am
- License Nr.: 5-725
calculations with constrained symmetry
Last edited by edurgun on Mon May 26, 2008 8:20 pm, edited 1 time in total.
-
admin
- Administrator
- Posts: 2922
- Joined: Tue Aug 03, 2004 8:18 am
- License Nr.: 458
calculations with constrained symmetry
in principle, you can only turn off symmetry completely by setting ISYM=0.
Then, the forces are not symmetrized and relaxations are not symmetry-restricted at all.
If the space group you want to impose is lower than the space group detected by vasp, you can in principle try one of the following work-arounds:
1) you can always define equivalent atoms of the high-symmetry cell as inequivalent by spitting their number in POSCAR and simply doubling the PP of the respective atom type in POTCAR. ( like splitting 3 atoms A to 2 1 and copying the PP of A twice to POTCAR)
2) or (if the reduction of the symmetry has the effect that certain atoms only relax along certain directions in space), please choose ISYM=2 in INCAR, but 'selective dynamics' in POSCAR and set the relaxation flags (T/F) for each atom appropriately after the atomic position : i.e. if due to the different space group one atoms relaxes along z only instead of relaxation in xyz , simply set x y z F F T instead of x y z T T T
Then, the forces are not symmetrized and relaxations are not symmetry-restricted at all.
If the space group you want to impose is lower than the space group detected by vasp, you can in principle try one of the following work-arounds:
1) you can always define equivalent atoms of the high-symmetry cell as inequivalent by spitting their number in POSCAR and simply doubling the PP of the respective atom type in POTCAR. ( like splitting 3 atoms A to 2 1 and copying the PP of A twice to POTCAR)
2) or (if the reduction of the symmetry has the effect that certain atoms only relax along certain directions in space), please choose ISYM=2 in INCAR, but 'selective dynamics' in POSCAR and set the relaxation flags (T/F) for each atom appropriately after the atomic position : i.e. if due to the different space group one atoms relaxes along z only instead of relaxation in xyz , simply set x y z F F T instead of x y z T T T
Last edited by admin on Tue May 27, 2008 1:47 pm, edited 1 time in total.