[Pw_forum] DFT+U

[Suza W]

Hi Emine,

Thanks a ton for your kind reply.

For lead (Pb), if one gives Hubbard_5d = y eV, is this enough ?
Or, one also has to have a Pb pseudopotential where 10 electrons
of 5d-state must be treated as valence states ?

I mean, is this acceptable,  if one has a Pb pseudo with  6s2 6p2
(only four electrons in valence)  and then she/he wants to apply
Hubbard_5d = y eV for the 5d-orbital which is fully frozen inside the core ?

Furthermore, is this at all permitted to give Hubbard U_f  and Hubbard U_d
together for the same element ? In some cases in the literature, the
Hubbard U_p
and Hubbard U_d together have been used for the same Titanium
(Ti) element.  Is this physically meaningful ?

Regards,
Suza

On Sat, Apr 27, 2013 at 11:16 PM, Kucukbenli Emine <emine.kucukbenli at epfl.ch
> wrote:

>  Hi Suza,
>
>  there are two places in the code which you would need to modify to be
> able to use a different angular momentum than the default.  Dont be scared
> though, they are fairly easy to understand and modify:
>
>  espresso/flib/set_Hubbard_l.f90
> tells which angular momentum to use for each atom type (very transparent:
> Hubbard_l 2 is for d and 1 is for p)
>
>  espresso/PW/src/tabd.f90
> tells how to distribute initial occupation on the Hubbard projector states
>
>  Change these files to your wish and re-make pw and you will be set.
>
>  best,
> emine kucukbenli, postdoc at theos, epfl, switzerland
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20130428/1be20229/attachment.html>