LDA+U could also be of help - and it's implemented in Siesta, as opposed to
GW and hybrid functionals -  but then U should be chosen carefully, such
that the underlying physics comes out reliably.

Marcos

On Tue, Jun 15, 2010 at 6:27 PM, Nichols A. Romero <[email protected]>wrote:

> If your goal is to obtain a more accurate band gap, you will need DFT
> with hybrid functions and or GW.
>
> On Tue, Jun 15, 2010 at 11:18 AM, Koch, Robert J <[email protected]> wrote:
> > Hello All,
> >
> > I'd greatly appreciate any advice on the following:
> >
> > I'm trying to calculate the band gap of TIO2 (Rutile).  I have attached
> my fdf file, my pseudo .psf files as well as their input .inp files.
> >
> > When I run this and use the eig2dos utility to plot density of states, I
> see a band gap of about 1.4 eV which is far too low (should be about 3.3
> eV).  I'm aware that typically DFT calculations underestimate the band gap,
> but this underestimation seems like a little much.  Can anyone make any
> recemendations as to how to improve my calculation?  I'm quite stuck and not
> sure how to proceed.
> >
> > Thanks in advance!
> > Rob Koch
> >
>
>
>
> --
> Nichols A. Romero, Ph.D.
> Argonne Leadership Computing Facility
> Argonne, IL 60490
> (630) 252-3441 (O)
> (630) 470-0462 (C)
>

Responder a