The ABINIT Users Mailing List ( CLOSED )

["Yu Zhou" <zhouy6@rpi.edu>]

Hi Matteo,

Thank you very much for telling me this, sorry for my carelessness.
I'm new to this, and I'm still having some trouble in writing the
shiftk. From the
tutorial(http://www.abinit.org/Infos_v5.5/input_variables/varbas.html#shiftk),
I know for FCC structure, if I want to sample the Gamma point, shiftk
should be
shiftk  0.0 0.0 0.0
         0.0 0.5 0.5
         0.5 0.0 0.5
         0.5 0.5 0.0
But I got error if I want to shift to another K-Point, say (kx,ky,kz)
Would you please tell me how to write it correctly?

(Actually I also have some difficulty in understanding how to write
"Shiftk" in the manual, I simply write
nshiftk2 1
shiftk2 1/ngkpt(1)*kx, 1/ngkpt(2)*ky, 1/ngkpt(3)*kz   ## they values
have been calculated, not in formula form
the error is still sth like:  cchi0 : ERROR -  kp=k-q-G0 not found
)

Thanks a lot!

Yu


On Mon, Jun 9, 2008 at 4:38 PM, Matteo Giantomassi
<Matteo.Giantomassi@uclouvain.be> wrote:
>
>
>> Hi dear all,
>>
>
>  Dear Yu,
>
>  The answer to your question is reported in the first GW tutorial
>  at the end of Section 9:
>
> "Finally, it is in principle possible to calculate a full band plot of a
> system. This is a very cumbersome calculation, above all for the fact that
> there is still not enough automatization at the actual state of the code.
> The way to perform a full band plot is by shifting the k-point grid where
> the convergence has been achieved, such as to cover all the desired
> k-point positions to be reported in the band plot. A simplification can
> also be represented by the fact that the GW corrections are quite linear
> with the energy."
>
>  Regards,
>  Matteo Giantomassi
>
>> I want to plot the band structure through a specified direction in
>> k-space. but
>> in the 2nd step. If I give the k-points along this direction manually in
>> the
>> 2nd step (Calculation of KSS file), it will report error ( cchi0 : ERROR -
>> kp=k-q-G0 not found). If I use a shifted mesh, there would be no problem.
>> But
>> if I want to scan more k-points in some region, I have to increase the
>> mesh
>> grid to 32x32x32 in order to include all the k-points I'm interested in,
>> but
>> it will make the calculation too slow.
>>
>> So my question is how I can plot the band-structure with GW?
>>
>> Thanks a lot!!
>>
>> Yu Zhou
>> Dept. of Phys.
>> Rensselaer Polytechnic Institute
>>
>
>
>