The ABINIT Users Mailing List ( CLOSED )

[Michel Côté <michel.cote@umontreal.ca>]

Title: Re: [abinit-forum] Re: band structure & Fermi Energi, (my email to forum abinit rejected again)

Dear Pierre-Matthieu and Ariesto,

Concerning the Fermi energy, usually the Fermi energy is computed with a Monkhost-Pack grid which is shifted form the origin. In most materials, the highest valence state is a symmetry point so the grid misses it and the code return a Fermi energy corresponding to the highest occupied state. When the band structure is computed after, it usually includes high symmetry points which could be slightly higher in energy. In this case, you should shift the whole band structure to fit the physical Fermi energy. This “problem” is mentioned in the presentation of the computation of the DOS that was done in Santa Barbara summer school (http://www.abinit.org/summer_05/dos_charge_wavefn.pdf). By the way, the Fermi energy for a semiconductor at T=0K could be anywhere in the gap, we usually put it at the top of the valence band. It is only at finite temperature that it is located near the middle of the gap but not exactly in the middle!

Concerning the overall shape, as Pierre-Matthieu indicated, it is usually fairly correct, however, you could have cases that require more attention. A GW calculation will help to get the correct band structure. Particularly, LDA tends to overestimate the energy width of the valence bands, GW corrects that. Also, conduction bands could be not as well described as the valence bands and it could depend on the pseudopotentials used. I have seen such cases in GaAs.

I hope this help.

Michel

Le 28/02/07 22:52, « Anglade Pierre-Matthieu » <anglade@gmail.com> a écrit :

> Hi all,
>
> For as much as I know, Band structures are known to have almost
> correct shapes but with large discrepancies in term of band gap.
> Ariesto find that the shape is also not correct does anyone knows
> about that ?
>
> It happens that some people shift only there computed band structure
> to reach experimental band gap without much other changes. So I guess
> that you should usually find almost correct shape for your valence
> band.
>
> regards
>
> PMA
>
>
>
> On 3/1/07, 아리에스토 <ariesto@postech.ac.kr> wrote:
> >
> >
> >
> > Dear DR.Pierre-Matthieu Anglade,
> >
> >
> >
> > Sorry for sending question personaly, because my  email to forum was rejected again as before,although I login  properly
> >
> >
> >
> >
> >
> > I have aquestion,
> >
> > It is already  mentioned in tutorial 3.5 that DFT calculation produces underestimate value of  bandgap.
> >
> > I also found ,after figuring out t35.out file, the fermi Energi  is not in the midle of band gap. the fermi energy is located below the top of  valence band in Gamma point.
> >
> > Now, in graphene system , I also found the  same thing.
> >
> > Is the problem also well known as band strucutre DFT  problem?
> >
> > Best regards,
> > Ariesto
> >
> >
> >
> >
> >
> >
> >
> >


-- 
*************************************************************** 
  Michel Cote                          tel: +1 (514) 343-5628     
  Professeur agrégé                   fax: +1 (514) 343-2071               
  Département de physique 
  Université de Montréal 
  C.P. 6128, succ. Centre-ville 
  Montréal (Québec) H3C 3J7          Michel.Cote@umontreal.ca 
  Canada                   www.phys.umontreal.ca/~michel_cote 
***************************************************************