Spin-polarized band interpolation

[kle2797]

Dear Developers,

I used ypp to interpolate the bands from QE calculation for a spin polarized system and got some weird output compared to the band structure I got from QE's `bands.x` and `plotbands.x`.

fig.pdf

What I did was simply generate the SAVE database using `p2y` from a nscf calculation, and then used 'ypp -s b' to get the input file. The spin up channel seems good, but the spin down is quite off. So I wonder whether I generate the SAVE database incorrectly which may affect my other calculations.

I also attached my nscf and ypp input file here for you inspection.

Thank you!

[Davide Sangalli]

Dear Khoa,

the issue is that there is a missing conversion factor in the values of the spin down channel.
Just multiply the energies by 27.2114 (the conversion factor from Hartree to eV) and you should get the correct value.

Best,
D.

P.S.: this could give problems

Code: Select all

%BANDS_kpts                      # K points of the bands circuit
0.000000  |0.000000  |0.000000  |  !G
0.500000  |0.000000  |0.000000  |  !M
0.333333  |0.333333  |0.000000  |  !K
0.000000  |0.000000  |0.000000  |  !G
%

can you try changing it to

Code: Select all

%BANDS_kpts                      # K points of the bands circuit
0.000000  |0.000000  |0.000000  |  #G
0.500000  |0.000000  |0.000000  |  #M
0.333333  |0.333333  |0.000000  |  #K
0.000000  |0.000000  |0.000000  |  #G
%

[kle2797]

Dear Khoa,

the issue is that there is a missing conversion factor in the values of the spin down channel.
Just multiply the energies by 27.2114 (the conversion factor from Hartree to eV) and you should get the correct value.

Best,
D.

P.S.: this could give problems

Code: Select all

%BANDS_kpts                      # K points of the bands circuit
0.000000  |0.000000  |0.000000  |  !G
0.500000  |0.000000  |0.000000  |  !M
0.333333  |0.333333  |0.000000  |  !K
0.000000  |0.000000  |0.000000  |  !G
%

can you try changing it to

Code: Select all

%BANDS_kpts                      # K points of the bands circuit
0.000000  |0.000000  |0.000000  |  #G
0.500000  |0.000000  |0.000000  |  #M
0.333333  |0.333333  |0.000000  |  #K
0.000000  |0.000000  |0.000000  |  #G
%

Hi Davide,

Thank you for your reply. Multiply by the conversion factor does give the good bands. However, I tried your suggestion to the ypp input, but it did not fix it.

It seems to be just about 'ypp' interpolation, so do you think the database is good to do GW-BSE for the spin-flip channel?

Best,

[Daniele Varsano]

Dear Khoa Le

the missing factor is not related to the input file.

Code: Select all

so do you think the database is good to do GW-BSE for the spin-flip channel?

You can safely proceed with GW/BSE calculations, however, it is not clear to me what do you mean by spin-flip channel.

Best,
Daniele

[kle2797]

You can safely proceed with GW/BSE calculations, however, it is not clear to me what do you mean by spin-flip channel.

Hi Daniele,

Thank you for the reply and my apologize for the late respond.
What I mean is the calculation of magnon using BSEprop = "magn". So can I assume it is still safe in this case then?

Khoa

[Daniele Varsano]

Dear Khoa,

most probably yes, even if I'm not an expert in this kind of calculation.

Best,
Daniele

[kle2797]

Dear Khoa,

most probably yes, even if I'm not an expert in this kind of calculation.

Best,
Daniele

Thank you for your help!