I was going through the tutorial handsout and I was running 3_4 and 3_6, in which the top layer of Ni(111) is relaxed. I have two questions about this.
1) Although I have copied the files from the file server, I get different numbers for the energies. It is also not clear to me which energies I should use. I have read on this forum that I should use the energy without entropy or the sigma --> 0 value, but they are quite different from each other. Can anyone explain?
2) In 3_4, the change in z-position of the two top layers are shifted quite a lot, but in 3_6, there is no change in z-position in the first 8 significant digits. All the input files are identical, the only difference being ENMAX = 240 in 3_4 and ENMAX = 400 in 3_6. I do not trust the latter value, although it should be more accurate. Can anyone explain?
Relaxation at different energy cutoffs
Moderators: Global Moderator, Moderator
-
physicalattraction
- Newbie
- Posts: 44
- Joined: Wed May 21, 2008 1:47 pm
- License Nr.: 916
- Location: Eindhoven, The Netherlands
Relaxation at different energy cutoffs
Last edited by physicalattraction on Thu Aug 14, 2008 12:33 pm, edited 1 time in total.
-
admin
- Administrator
- Posts: 2921
- Joined: Tue Aug 03, 2004 8:18 am
- License Nr.: 458
Relaxation at different energy cutoffs
1) the difference is due to the rather large smearing with, it should go to 0 as SIGMA is decreased.
2) these examples were given
-- to show the influence of the cutoffs on the quality of the basis set. Please have a look at the total drifts written after the forces on the atoms, they should be as small as possible: the results with the increased cutoffs (ENCUT=400) give lower drifts and are thus better
-- furthermore a run with the same cutoff as the CO/Ni calculation is needed to obtain the adsorption energies (example 3_5, the cutoffs of C and O are 400 eV). Therefore you are of course correct, the result of 3_4 looks much more reliable
!!!! However, please note that in 3_6 no ionic relaxation step is done (NSW=0), therefore it is just a static run with the same input geometry as 3_4, whereas ionic relaxation is done in 3_4 (NSW=100, 4 ionic steps are done to reach the ionic convergence criterium of E(N) - E(N-1) < 0.001
2) these examples were given
-- to show the influence of the cutoffs on the quality of the basis set. Please have a look at the total drifts written after the forces on the atoms, they should be as small as possible: the results with the increased cutoffs (ENCUT=400) give lower drifts and are thus better
-- furthermore a run with the same cutoff as the CO/Ni calculation is needed to obtain the adsorption energies (example 3_5, the cutoffs of C and O are 400 eV). Therefore you are of course correct, the result of 3_4 looks much more reliable
!!!! However, please note that in 3_6 no ionic relaxation step is done (NSW=0), therefore it is just a static run with the same input geometry as 3_4, whereas ionic relaxation is done in 3_4 (NSW=100, 4 ionic steps are done to reach the ionic convergence criterium of E(N) - E(N-1) < 0.001
Last edited by admin on Mon Aug 18, 2008 12:41 pm, edited 1 time in total.