strange behavior due to the GW correction

[rezuzu]

Dear all,

After a GW correction, the comparison betwen the obtained band structure and the DFT one (attached file) shows a shift of the valence bands
to higher energies (strange behavior).

Can you help to understand the origin of this unusual behavior. I this due to a possible error in the gw calculation or it is a sign of the rise of a new physics??

Karim

[Daniele Varsano]

Dear Karim,
can you post your input/report/output file? We do not know on which system you calculated the band structure and with just the figure it is really hard to understand.

Best,
Daniele

[rezuzu]

Dear Daniele,

Please find enclosed what you asked me for.

Best,

Karim

[rezuzu]

Dear Daniele,

The system under study consists of a graphene sheet co-decorated with H and F (ie, hydrofluorinated graphene) in a chair-like configuration .

Best;

Karim

[Daniele Varsano]

Dear Rezuzu,

Sorry, but I can't recognize the band structure you plor with the o.qp file. Looking to the QP file I can see a GW corrections E-Eo, always negative for the occupied bands and positive for the empty bands. Something has to be happened in the post-processing.procedure. Most probably ypp realigned the Fermi level at 0, so you do not have a common reference for the DFT and GW band structure. I need to c
heck it. In the meantime you can check if it is really the case and shift you GW bands by the Delta E at Gamma, ie, DeltaE=-1.321478

Anyway let me give you a couple of suggestions:

You have a 2D system, which need some care..
1) Cutoff Box, you are using a very small Zcut, this is not your fault, but our fault, we still do not have corrected this part in the input file, but the cutoff effectively used by the code turns to be
Zcut/2. Please search discussion on the topic in the forum. So I suggest you to use a Zcut=32.

Next, the convergence in GW with respect the bands in the Green Function is very slowi. Moreover it is a sum-over-state, and you shouyld include in the summation all the occupied bands (I think that yambo reset it to do the work, you should have a warning in the standard output, but I'm not sure, I'm travelling and I do not have the source now with me). Anyway put the GbndRnge start from 1 and check if 15 bands are enough to have converged results.

Hope it helps,
Daniele

[rezuzu]

Dear Daniele,

Thank you so much for your valuable advices.

Karim