Fail to reproduce DFT bandstructure

[Davide Sangalli]

Ok.
So if you are still not able to use the svn version what you can do is modify by hand a2y.

Open the file "interfaces/int_modules/mod_com2y.F" and look for the lines (around line 93)

Code: Select all

       i_time_rev=1
       if(l_spin_orbit.and.n_spin_den==4) i_time_rev=0
       double_symmetries=(i_time_rev==1)

and manually change the first one to

Code: Select all

i_time_rev=0

Then do a make clean_all and recompile the interfaces.
a2y will now set the time-reversal to false by definition.
Remember to use this a2y only for KSS files generated without symmetries.

Best,
D.

[JavadH]

Hello Davide,
Thanks, I did that and a2y works now, although very slowly. When I run yambo -N after that I get:

Code: Select all

[WARNING]Spatial Inversion not found among the given symmetry list

And the band structure still differs with the one of DFT. Can this difference be related to something else than the symmetries?
These figures show both band structures. The subtle splits around C k-point (~42 in yppBAnd) is not reproduced correctly.

AbinitBand.png

yppBand.png

Thanks again for the help,
Regards,
Javad

[Davide Sangalli]

I guess you are using the ypp interpolation scheme.

In this case the bands are not computed but interpolated using sinusoidal functions from the regular mesh to the path of your choice in the BZ.
Crossing points / degeneracy points are hardly well reproduced. You should probably start from a regular grid which includes C to obtain something reasonable.

Best,
D.

[JavadH]

OK, is there any way to calculate the bands instead of interpolation? I need a more accurate band structure here.


Thanks again,
Best,
Javad