how would I relax the lattice structure in x and y direction and keep the strain in z direction?
Thank you.
relaxiation of the lattice
Moderators: Global Moderator, Moderator
-
hzhao
- Newbie
- Posts: 1
- Joined: Wed Sep 15, 2010 7:39 pm
- License Nr.: 5-90
relaxiation of the lattice
Last edited by hzhao on Fri Sep 17, 2010 12:57 pm, edited 1 time in total.
-
maartendft
- Newbie
- Posts: 38
- Joined: Sun Feb 07, 2010 7:17 pm
- Location: The Hague, Netherlands
relaxiation of the lattice
Hi,
Take your initial structure and compute the stress tensor acting on it, but without relaxing any lattice parameters (use e.g. ISIF=2). Subsequently, retrieve the stresses in x and y-direction from your OUTCAR.
Then, construct a new POSCAR with the c-lattice parameter as before but with a and b scaled according to a certain scalar value. You need to adapt a and b according to the stresses computed in the previous step, e.g. make the lattice vector b bit smaller if you found compressive stresses in the y-direction.
Iterate until you have reached a certain desired accuracy. I believe people have posted bash scripts on this forum before that automate this procedure.
<span class='smallblacktext'>[ Edited Fri Sep 17 2010, 06:42PM ]</span>
Take your initial structure and compute the stress tensor acting on it, but without relaxing any lattice parameters (use e.g. ISIF=2). Subsequently, retrieve the stresses in x and y-direction from your OUTCAR.
Then, construct a new POSCAR with the c-lattice parameter as before but with a and b scaled according to a certain scalar value. You need to adapt a and b according to the stresses computed in the previous step, e.g. make the lattice vector b bit smaller if you found compressive stresses in the y-direction.
Iterate until you have reached a certain desired accuracy. I believe people have posted bash scripts on this forum before that automate this procedure.
<span class='smallblacktext'>[ Edited Fri Sep 17 2010, 06:42PM ]</span>
Last edited by maartendft on Fri Sep 17, 2010 1:30 pm, edited 1 time in total.