5
$\begingroup$

Do I do rest, scf, dos steps when calculating DOS like PBE in HSE06? Can anyone help me with this?

$\endgroup$
3
  • $\begingroup$ This question made it into Hot Network Questions, but those acronyms are unknown to readers who might be casually interested and wondering what this might be about. I have no idea what any of this is about; I was expecting it to be about denial-of-service attacks on one of the security stack exchange sites, or MS-DOS (Disk Operating System) or something on retrocomputing. Matter Modeling was a surprise. IDK if there's anything you can do to give random readers a clue. $\endgroup$ Nov 1, 2022 at 3:31
  • 1
    $\begingroup$ @PeterCordes I totally agree that this question needs a lot of improvement. For future reference DOS here means density-of-states. Unfortunately if you search DOS on this site you'll see it in dozens of titles, so we can't stop people from using it like that without spelling it out, but I agree that the rest of the question needs significant improvement! $\endgroup$ Nov 2, 2022 at 13:08
  • $\begingroup$ DOS (Density of States). Physical understanding. $\endgroup$ Nov 2, 2022 at 17:24

1 Answer 1

5
$\begingroup$

You will often want to relax the structure first. This can be done with PBE or HSE (you can use PBE if HSE is too slow). I suppose the “can” is vague. The code will allow you to calculate the DOS with HSE on any structure you give it, so in that since, you can relax with PBE or HSE, or not relax at all and then do a HSE calculation. In terms of what is good practice, it depends a lot more on the system and what you want to calculate. If you’re looking at a system with a well established crystal structure, then it might be reasonable to not relax at all and do an HSE DOS. Sometimes the PBE and HSE structures will be similar, other times they will be different. Sometimes two sufficiently similar structures will yield basically the same DOS. If you’re concerned with the impact of these different choices on your results, you can calculate the DOS for different choices and see how much it changes. In some cases this won’t be computationally feasible, and you’ll have to hope that you can get good results with whatever choice you make, but you should try to understand what sort of impact these different choices regarding relaxing or not typically has in similar systems. It might be that there isn’t a computational feasible way to generate data that should be accurate.

Regarding the scf calculation before the DOS, typically with HSE calculations, one doesn’t do something like with PBE where you do a coarse k grid calculation to generate the charge density and then do a much finer k grid calculation non-self-consistently. Hybrid calculations require calculating terms involving wave functions and eigenvalues at different k points, so you can’t just do a nscf calculation like you can with non-hybrids. One often just has to do a scf calculation with a larger k mesh, or do something like Wannier interpolation (which can also be used for band structures).

$\endgroup$
4
  • $\begingroup$ I optimized the structure in PBE. Now can you guide me on HSE06 Material Optics and DOS calculation. I would appreciate it if you could write step by step. $\endgroup$ Oct 31, 2022 at 7:43
  • $\begingroup$ vasp.at/wiki/index.php/Fcc_Ni_DOS_with_hybrid_functional you just run an scf calculation as shown here. They also do a band structure in one of the tutorials $\endgroup$
    – AGS
    Nov 1, 2022 at 3:40
  • $\begingroup$ Can I calculate the DOS in HSE06 from the scf step as well? $\endgroup$ Nov 10, 2022 at 16:49
  • $\begingroup$ Yes, you can always calculate it, it’s just you might need a denser k grid $\endgroup$
    – AGS
    Nov 12, 2022 at 4:32

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .