The ABINIT Users Mailing List ( CLOSED )

[Fabien Bruneval <fabien.bruneval@polytechnique.fr>]

I ran your calculation on my computer and I found a band gap of 0.58 eV 
at Gamma, which is a very reasonable value for the LDA band gap of silicon.
In the conventional cell of silicon, there are 16 occupied bands. 
Therefore, the band gap occurs between the 16th and the 17th band (and 
not between the 4th and 5th states as in the primitive cell).

I further suggest you to define the input variable "ngkpt" in the first 
dataset.
For instance, ngkpt1 2 2 2

Good luck.


Fabien


ajlu@ustc.edu wrote:
> Thanks a lot. But I need furhter help yet.
>
> 1. I used TM psp download from website. It works well for silicon band structure calculation with 2 atoms (primitive cell). So I think it's okey to be used here. 
>
> 2. I did not pay attention to the special point label in the reciprocal 
> lattice. Thanks for your suggestion. In fact, I care for the Gamma point 
> mainly, but the problem just occurs at this point.
>
> 3. 32 k points are calculated, and there are no warnings in the log file. So, 
> I can not find other useful clues to revise my input file.
>
> Thanks again.