My Community

hi,
i have a question about calculating a DOS using HSE method. The way I did it, I ran it with ALGO=All and ISMEAR=0 and took the total DOS from the DOSCAR file as my final result. Now, I was told that one should run it with ALGO=53 (Damped) and ISMEAR=-5 after pre-converging with ISMEAR=0. Is this step necessary? How would that affect (improve) my results from the ALGO=All run? Are my results still correct?
I would appreciate your quick response.
THANKS.

Dear admin,
would you please comment on the warning message I receive when running HSE jobs:| |
| ALGO=A and IALGO=5X tend to fail with the tetrahedron method |
| (e.g. Bloechls method ISMEAR=-5 is not variational) |
| please switch to IMSEAR=0-n, except for DOS calculations |
| For DOS calculations use IALGO=53 after preconverging with ISMEAR>=0 |
| I HOPE YOU KNOW, WHAT YOU ARE DOING ||

I am interested in running a DOS calculation. If I run with ALGO=All and ISMEAR=0, I don't have any warning messages. But in light of the comment (For DOS calculations use IALGO=53 after preconverging with ismear=>0....), is my approach correct? Or, in other words, can I just take the resultant DOSCAR file for my DOS plots? If I try IALGO=53 with ISMEAR=-5 from my converged parameters, it does not work (the message reappears).
I would appreciate your explanation of the warning message.
Thanks.

With HSE functionals, you can't use tetrahedron method ISMEAR=-5. It fails as it's a non-variational method. The only exception is when you do a postprocessing e.g. DOS calculation. I'm wondering If you have set ICHARG=11 in your INCAR for DOS calculation.

Thanks for your reply. No, I haven't set ICHARG=11. Should I? I will try.

No, you cannot use local charges (i.e. ICHARGE =11), as exact exchange is rather a functional of KS orbitals.