hi,
I am trying to optimize crystal structure of a monoclinic oxide. While the a, b, c, and the angle do not vary as I increase the cutoff energy from 400 to 500 and lastly to 550, or when I use more dense k-points, the total energy changes (goes down) by 0.1eV when using these more stringent parameters. Should I expect the total energy to converge at some point (for some high cutoff E and k-points)? Is that where I get the optimized lattice parameters?
Thanks.
convergence
Moderators: Global Moderator, Moderator
-
apple
- Jr. Member
- Posts: 99
- Joined: Tue Feb 16, 2010 7:09 pm
convergence
Last edited by apple on Fri Nov 12, 2010 12:04 am, edited 1 time in total.
-
alex
- Hero Member
- Posts: 586
- Joined: Tue Nov 16, 2004 2:21 pm
- License Nr.: 5-67
- Location: Germany
convergence
Hi apple,
the short answers are: a) yes and b) yes.
However, if you change your ENCUT by only 20% do not expect too much change. Double it. Then you'll see, if the first set-up was ok.
Do you use O or O_s PAW? In case of O the cut-off is too small anyway (Pulay stress ...).
Have you wondered about raising total energy with higher cut-off energies? There is some compensation built in ('kinetic energy error for atom') ...
Hth
alex
the short answers are: a) yes and b) yes.
However, if you change your ENCUT by only 20% do not expect too much change. Double it. Then you'll see, if the first set-up was ok.
Do you use O or O_s PAW? In case of O the cut-off is too small anyway (Pulay stress ...).
Have you wondered about raising total energy with higher cut-off energies? There is some compensation built in ('kinetic energy error for atom') ...
Hth
alex
Last edited by alex on Fri Nov 12, 2010 8:12 am, edited 1 time in total.
-
apple
- Jr. Member
- Posts: 99
- Joined: Tue Feb 16, 2010 7:09 pm
convergence
Thanks Alex. I will play with the Ecutoffs.
Last edited by apple on Fri Nov 12, 2010 8:42 pm, edited 1 time in total.