symmetry change upon ISIF=3

Message

brsahu · #1 Post by **brsahu** » Thu Mar 02, 2006 7:37 pm

Dear VASPusers,

I used ISIF=3 tag for simultaneously relaxing the lattice constant and angles for a trigonal(rhombohedral) cell but at the end of the relaxation run, the symmetry changes to monoclinic.

Is there a way to keep the symmetry of the original cell fixed that is trigonal or rhombohedral but
simultaneouly relax the angles and the lattice constant with the
use of ISIF or any other tag in VASP so that at the end of the relaxation my cell would be still trigonal or rhombohedral but angles and lattice constant relaxed.

Sahu

lcyin · #2 Post by **lcyin** » Fri Mar 03, 2006 3:10 am

I met the same question, does any master know how to deal with it?

Veronika · #3 Post by **Veronika** » Fri Mar 03, 2006 8:16 am

If the cell changes to monoclinic than your initial symmetry due to the setup (atoms, magnetic moments, etc. ) was lower than for a rhombohedral cell. Since vasp determines the symmetry at the beginning of the calculation and does not change afterwards, you will run in this (monoclinic) minimum using ISIF = 3.
Probably you can use one of the flags 4, 5, 6 or 7 not relaxing everything within one run. (http://cms.mpi.univie.ac.at/vasp/vasp/node103.html)
<span class='smallblacktext'>[ Edited Fri Mar 03 2006, 10:36AM ]</span>

brsahu · #4 Post by **brsahu** » Fri Mar 03, 2006 8:56 pm

veronika,

VASP reads correctly reads the trigonal or rhombohedral symmetry in the begining and identifies it. It is only at the end of relaxation (ISIF =3) does it changes to monoclinic.

This can be checked by copying final CONTCAR to POSCAR and re-running VASP again.

So the question is how to keep the initial symmetry for ISIF=3?

Sahu

rita · #5 Post by **rita** » Sun Mar 19, 2006 6:41 am

Sahu,

If look at the running process, there are some warning or message writen in different step. May be can help to find the reason induced in symmetry broken.

Rita

#6 Post by **admin** » Tue Mar 28, 2006 9:38 am

how did you set ISYMM?