[Wien] Need advice for modeling the band gap in a charge transfer insulator

[Peter Blaha]

Are you sure, your LDA+U calculations are done correctly ?
You have to make spin polarized calculations with the correct magnetic order.
Often a FM state is metallic, while the proper AFM state is insulating.

The full hybrid-DFT method will be available in the next release.
However, this is an expensive method and you cannot apply it to large
systems. At least you will need a powerful cluster with many nodes ...

PS: Have you done mBJ calculations ? In our experience this is at least as
accurate as hybrid DFT at a fraction of cpu-time when you need only the
bands/DOS.
It also works well for "correlated" electrons and in addition it does not have
a parameter (like the alpha in hybrid-DFT or the U in LDA+U).


Am 25.10.2011 12:14, schrieb David Tompsett:
> Hi Shamik,
>
> DFT+U can sometimes give a reasonable description of materials with some charge transfer (possibly also some Mott) insulating character eg. NiO. But it seems that in your case
> itinerant states are important to the band gap and in that case using Hybrid functionals that include exchange between all states (including itinerant ones) often helps eg. MnO. I
> think a full hybrid functional set will come out in an upcoming version of Wien2k?
>
> Some references:
> Tran et al.
> PHYSICAL REVIEW B83,235118 (2011)
>
> Best,
> David.
>
> 2011/10/25 shamik chakrabarti <shamikphy at gmail.com <mailto:shamikphy at gmail.com>>
>
>
>     Dear wien2k users,
>
>                                  We all know that LSDA (or GGA)+U can be used successfully for a Mott-Hubbard insulator. Anisimov et al. [PRL 99, 156404 (2007)] have described
>     that for a charge transfer insulator we need to use LDA+DMFT for modelling their band structure properly. Now my question is :
>
>     (1) Is their any option in wien2k by which we can use LDA+DMFT?
>
>     (2) In our calculation after applying GGA+U (U for d orbital of TM) we saw that although there are a band gap between spin up & spin dn d DOS of transition metal, there are
>     ligand (O p) DOS at Fermi Energy which is making our material metallic while experimentally it is an insulator [conductivity ~ 10^(-7)]. Here should we apply some value of U
>     (say 0.15 Ry) for O p orbital also?
>
>     (3) Any other advice regarding to this problem will be highly appreciated.
>
>     Thanks in advance.
>
>     yours faithfully,
>     --
>     Shamik Chakrabarti
>     Senior Research Fellow
>     Dept. of Physics & Meteorology
>     Material Processing & Solid State Ionics Lab
>     IIT Kharagpur
>     Kharagpur 721302
>     INDIA
>
>     _______________________________________________
>     Wien mailing list
>     Wien at zeus.theochem.tuwien.ac.at <mailto:Wien at zeus.theochem.tuwien.ac.at>
>     http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
>
>
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

-- 

                                       P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------