[Wien] BJ vs. mBJ

[delamora]

Thank you Dr. Tran,
          In the setting of the mBJ I used;
init_mbj

Which has different parametrizations of mBJ:

0:    Original mBJ values (Tran,Blaha PRL102,226401) (default)
1:    New parameterization (Koller etal, PRB85, 155109)
2:    New parameterization for semiconductors (gaps up to 7 eV)
3:    Unmodified BJ potential (Becke,Johnson J.Chem.Phys 124,221101

       I used 
                      0:    Original mBJ values
                      3:    Unmodified BJ potential
So my question was; in option '3:' it is still a potential and not a functional and it does not work if :FER is not in the gap?
         This is because I have a system with Sm which has a peak in the gap and :FER is in that peak, so if the "Unmodified BJ potential" was more "flexible" than the "Original mBJ values" then I could use this option.

I did the calculations again and I got for the gap:
without BJ    5.05eV
with mBJ       8.5eV
with unmodified BJ    5.8eV

but this time I did the things right and, for NaCl, I got :FER at the top of the valence band, which in earlier WIEN2k versions would move into the gap.

     So, my second question is; Which is better mBJ or 'unmodified BJ'?

Pablo
________________________________________

If a system has a gap, then the Ef (:FER) is set at the top of the
valence band maximum. Since there is for sure a gap in NaCl, whatever
is the used potential, then I don't understand your statement about
unmodified BJ. The value of :FER in case.scf should be equal to
the valence band maximum (grep for :BAN00008: in case.scf). If not
there is something wrong in your calculations.

F. Tran

On Thu, 14 Jan 2016, delamora wrote:

>
> WIEN2k users,
>
>         I am trying the mBJ potential and the Unmodified BJ potential, so I took NaCl as an example, so I get for the gap;
>
> without BJ:      5eV
>
> mBJ:                 8.2eV, and Ef is at the bottom of the gap
>
> Unmodified BJ: 6eV, Ef is not at the bottom of the gap
>
>         With these 3 very different values which one is the best; mBJ or Unmodified BJ?
>
>
>         I suppose that the Unmodified BJ is still a potential and cannot be used when Ef is not in the gap.
>
>
>             Pablo
>
> I am using WIEN2k 14.2
>
>
>