BSE and IPA

[niloufar]

Hello dear developers,
I just did the BSE calculation to get the absorption spectra. The first peak of dielectric function has a huge shift in compare to IPA around 10 ev or more and I'm not sure if it is acceptable or not.
I want to know that this much of shift is possible or not.
Thanks for helping in advance.

[Daniele Varsano]

Dear Niloufar,

not easy to answer with so few details, can you provide more info?, bettere if you post the output and report of your caluculation.
Best,
Daniele

[niloufar]

Hi again and thanks for your reply.
I attached the files, I should mention that I didn't converge the parameters because I thought it would change the result around 1 ev and not more...
But I faced another problem with bands too, and because they are the same compound and they are related, I attach them here, I'm sorry.

I'm completely confused and really looking forward to your help. My band structure doesn't seem normal and I don't know what the problem is. I converged ecut and Kpoints in scf calculation and I did all convergency parts in yambo before I got the band structure. I attach the files and I would be so over the moon if you help me. thanks in advance.

[Daniele Varsano]

Dear Niloufar,
let's try to go step by step.
1) Band structure, can you repeat the calculation using INTERP_mode= "BOLTZ" and see if you get the expected band structure?
2) The BSE spectrum is indeed very weird, the problem I can see is that the screening has been not read. Note, the report for the screening indicates that the calculation stopped for some reason. Actually, you can either recalculate the screening (static) or use the one you have already calculated it for the GW part (ndb.pp contains also the static part). You can use that screening by generating an input as:

Code: Select all

 yambo -o b -k sex -X p -y d

If you want to recalculate the screening instead (e..g if you want to use different parameters):

Code: Select all

 yambo -o b -k sex -X s -y d

Best,
Daniele

[niloufar]

Thanks a lot for your time and help, it is so precious. I will try them and post the result.

[niloufar]

Dear Daniele,
thanks a lot for your help, both problems were resolved and I'm so grateful.
wish you the best

[niloufar]

Hello dear Daniele,
First I'm thankful again for all your help, and second I have a new problem!
I considered SOC in my QE calculation, but in yambo setup file it identifies my semiconductor as a metal, I think DFT underestimated the gap and it is zero or near zero with soc so this happened. How can I solve this problem? By the way, I didn't do the rest of calculations in yambo because I thought it won't be fixed, can it?

[Daniele Varsano]

Dear Niloufar,

you can try to repeat your setup setting ElecTemp=0.0 eV , before you need to remove your ndb* files from the SAVE directory.
if the DFT calculation has a small gap this could solve the problem.
Otherwise, you can ignore the fact the system is metallic and see if a GW calculation opens the gap ignoring the warning about metallic system and Drude not included.
Best,
Daniele

[niloufar]

OK dear Daniele, I will do that, thanks a million @>-

[niloufar]

Hello dear Daniele,
thanks a zillion again for all your help.

I did a part of the exciton calculation and got the first bright exciton amplitude, and now I'm willing to show the path on the band structure, I mean show where it happens but I have a question (Sorry if it is stupid). How I can turn the Kpoint coordinate to a.u. values in the band structure axis? like 0.0 0.25 0.25 is what number?