soc+gw+bse

[shan dong]

Dear all,

I have done a GW+SOC calculation with Yambo(QE6.2 and yambo4.4). After the calculation, I found that the VBM and CBM had reversed and overlapped, and the system showed metallic properties.Can I perform the BSE calculation by using the GW band structure(set KfnQPdb)? Will the number of GW bands affect the result of BSE ?

Waiting for your replay. Thank you.

[Daniele Varsano]

Dear Shan Dong,
if your GW calculation inverts the order of the bands turning in metal they cannot be modelled with a scissor+stretching, so you need to read the calculated GW bands (KfnQPdb) as you argued. The BSE run will read the calculated QP and will fit the correction for the rest of the bands included in the BSE and not explicitly calculated. The result of the fit can depend on the calculated GW corrections and this is something you need to check.
Note that the result of the fit will be reported as output (o-data, o-fit) and you can check if it is reasonable or not.

Best,
Daniele

[shan dong]

Dear Daniele,

Thanks for your reply. Through researching the literature, I found that the two-dimensional AlSb's soc + gw should have a band gap, but the results I calculated show that the bands have reversed and overlapped. I don't know where the problem is. Could you help me solve it?

input.zip

[Daniele Varsano]

Dear Shan-Dong,
I suggest you check all your convergence parameters.
Some suggestions:
1) In doing convergence calculation I suggest you to not calculate the corrections for many bands in the whole BZ (%QPkrange), but just for few of them near the Fermi level (e.g. VBM and CBM only). This will speed up a lot your calculation.

2) From your k point sampling, it seems you want to study a 2D system, but from your input, you are not isolating the system, so in practice, you are considering a layered system. Consider to use coulomb cutoff potential (activating the runlevel adding yambo -r in our command line):
rim_cut
RandQpts=1000000 # [RIM] Number of random q-points in the BZ
RandGvec= 1 RL # [RIM] Coulomb interaction RS components
CUTGeo= "box z" # [CUT] Coulomb Cutoff geometry: box/cylinder/sphere
% CUTBox
0.00000 | 0.00000 | zcut | # [CUT] [au] Box sides
%

where zcut (in bohr) is a bit smaller than your cell side: ~37
You can have a look at this tutorial:
http://www.yambo-code.org/wiki/index.ph ... al_systems

Best,
Daniele