Hello,
I am trying to optimize a zeolite cell and I have several questions:
_ First I am optimizing only the zeolite cell using ISIF=3, prec = high, my question is untill which value of DIFFG should I go. After 50 step DE = 7 E-3 and after 300 steps DE= 4 E-4
After optimizing the Cell (shape/volume) and ions at the same time, I heard that a other turn only for the ions is needed (isif = 2, prec = normal) is it true ?
_ Does the optimization with force criteria is more reliable ? So which threshold should i use ?
_ I want to compute different conformation of a molecule in the channels. Should I re-optimize the cell( volume/shape) for the different conformation or can I use the same cell parameters for the different conformation of my moleucle ?
Thank you in advance,
Pierre.
Geometry Relaxation
Moderators: Global Moderator, Moderator
-
pmignon
- Newbie
- Posts: 29
- Joined: Mon Jul 03, 2006 11:19 am
Geometry Relaxation
Last edited by pmignon on Tue Dec 19, 2006 10:22 am, edited 1 time in total.
-
admin
- Administrator
- Posts: 2921
- Joined: Tue Aug 03, 2004 8:18 am
- License Nr.: 458
Geometry Relaxation
--) I would rather optimize with respect to the forces and stress tensors (EDIFFG0).
--) please check how large the relaxation effects were: if shape and volume change significantly, even re-start a few optimization steps (ISIF=3) from CONTCAR (but with a new WAVECAR), because the FFT-basis which was calculated for the initial geometry guess might be not appropriate (NGX,,....are not adjusted in a single vasp run). If the V or shape changes were only small, just re-run for the ionic positions at the end (ISIF=2) to obtain CHGCAR (and WAVECAR) with appropriate FFT-meshes for all post-processing steps (like DOS,...)
--) if you choose the forces as convergence criterium, | EDIFFG | should be ca 10xEDIFF
--) this depends of course on the (size of the) molecule, but usually it is sufficient to use the converged cell parameters of the pure zeolite (cell V and shape) and just relax the atoms positions.
--) please check how large the relaxation effects were: if shape and volume change significantly, even re-start a few optimization steps (ISIF=3) from CONTCAR (but with a new WAVECAR), because the FFT-basis which was calculated for the initial geometry guess might be not appropriate (NGX,,....are not adjusted in a single vasp run). If the V or shape changes were only small, just re-run for the ionic positions at the end (ISIF=2) to obtain CHGCAR (and WAVECAR) with appropriate FFT-meshes for all post-processing steps (like DOS,...)
--) if you choose the forces as convergence criterium, | EDIFFG | should be ca 10xEDIFF
--) this depends of course on the (size of the) molecule, but usually it is sufficient to use the converged cell parameters of the pure zeolite (cell V and shape) and just relax the atoms positions.
Last edited by admin on Tue Dec 19, 2006 2:35 pm, edited 1 time in total.