convergence test issue

[vasp-user]

hi

doing a convergence test for a system with about 220 atoms.

the ENCUT Vs energy gave the graph as attached, which is not correct. can any one suggest what is possible mistake? 

[Hitanshu Sachania]

Hello, did you keep k-points sufficiently high for all of these calculations? If not then changing the energy cut-off might lead to inconsistency. I'm not completely sure though. This is interesting.

As an aside, the Matter Modeling Stack Exchange has become quite an active resource to get any answers related to materials modelling queries. I would recommend you to ask this question there.

Regards,

Hitanshu Sachania

Indian Institute of Technology Madras

[KK M]

thanks for the reply. i wanted to do KPOINTS convergence after fixing ENCUT.

the supercrystal lattice values are

   a        b        c       alpha    beta     gamma
34.08380 17.04190 21.51710  90.0000  90.0000 120.0004

used [ 1x2x2]  as a starting KPOINTS.

will see the stack exchange site also

--
You received this message because you are subscribed to the Google Groups "VASP-SimulationPackage-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to vasp-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/vasp-users/79df17cc-4158-4bfb-a942-d3b3cc1eec9fn%40googlegroups.com.

[Hitanshu Sachania]

What is your ISIF value? It may be so that if you're using ISIF = 2, the system isn't relaxing at all and the first ionic step ends up giving inconsistent values.

To view this discussion on the web visit https://groups.google.com/d/msgid/vasp-users/CADsoCrVd_6voz9_2DMF%3D%2B0wZ7nX%3D%2BTao7R5gokv9bxs8oF1AcA%40mail.gmail.com.

[KK M]

oooh. No ISIF tag used. actually its long story. the system WAS giving lot of issues for SCF convergence, so as suggested in a forum, i started with minimum tags first to converge scf. so SCF is done with minimum tags. Now, started convergence without any tags. ISIF must have been considered as default by VASP. 

To view this discussion on the web visit https://groups.google.com/d/msgid/vasp-users/CAA8UMw_19WF9WGA8rCFO%3DzZsqBXX2sn6h5R5MvEAXA2VL%2BAW9g%40mail.gmail.com.

[Hitanshu Sachania]

The default for ISIF is 2. Though 2 is generally what I use for my convergence as well and I've never found the kind of variation that you get. If possible just recheck if all of your calculations are consistent. Also, to get a wider perspective, ask it at the matter modeling SE.

To view this discussion on the web visit https://groups.google.com/d/msgid/vasp-users/CADsoCrXuLb5mX_ZBor1DfhdnRi_-CCHrj72%3DF0UHQwmZ0DQA1w%40mail.gmail.com.