[Pw_forum] The band gap problem

[Paolo Giannozzi]

On Thursday 03 November 2005 04:49, #TAO JUNGUANG# wrote:

>  I am simulating the band structure of TiO2, at first ,I use nscf to
> calculate the band structure of TiO2 ,and got the band gap of 3.3eV
> ,which is very close to the experiment value.

and this is very strange: calculated gaps in DFT are typically
20 to 50% smaller than experimental gaps

> But as the following, I use the input file like this to calculate 
> the DOS, 
> &inputpp
>    outdir='/home/tao/NEW_TiO2/band/new/test/'
>     prefix='TiO2'
>     fildos='TiO2.dos'
>     Emin=3.0, Emax=25.0, DeltaE=0.1
>
> /
>      And my dos figure show that the band gap is about 2eV.

if you did not specify a broadening, or tetrahedra, in the input data 
file, the value of DeltaE (0.1 eV) is used as broadening. This should
be printed in the output.

P.
-- 
Paolo Giannozzi             e-mail:  giannozz at nest.sns.it
Scuola Normale Superiore    Phone:   +39/050-509876, Fax:-563513 
Piazza dei Cavalieri 7      I-56126 Pisa, Italy