OccTresh

[amolina]

Hello,

I would like to remove fractional occupations. Can this be done with OccTresh? How?

Best,

Alejandro.

[Daniele Varsano]

Dear Alejandro,
you can try to raise the OccTresh, but I do not know if you arrive to remove all the fractional occupancies, may some more experts dealing metals can help you in a better way. Other solution, is playing by reducing the smearing in the ground state calculation.
Best,
Daniele

[amolina]

In connection with my previous question. I am trying to do a Bethe-Salpeter calculation in a metal-semiconductor structure. How can I get avoid the error message:
"Fractional e/h occupations found. BS coupling forced."
Is it possible to select which states I want to be in the BS equation? For example, transitions between states 11,12 and 15. Or better, is it possible to tell yambo to do BS calculations only between fully occupied and fully empty states?
Thanks for the help.
Alejandro.

[Daniele Varsano]

Dear Alejandro,
ok, please note that ""Fractional e/h occupations found. BS coupling forced" it is not an error but the coupling term has to bee included in this case.
Next, a way to avoid to include that states could be using the postprocessing executable:
ypp -q
in this way you can modify the QP database, shifting in energies by hand the bands (13,14) you want to eliminate from the calculations. Shifting them in energy by huge amount they should not partecipate in the low lying energy part of the spectrum anymore.
Best,
Daniele

[amolina]

In connection with my calculations of BSE in metals (or my try to do it), I have achieved to obtain a BSE spectra. However, if I try to use the diagonalization to plot the wave-functions, Yambo stops without any results... Moreover, it seems that the k-grid is an issue, and for some grids calculations are done and for others not...
Has anyone tried a BSE calculation in a metal-semiconductor?
Cheers,
Alejandro.

[Daniele Varsano]

Dear Alejandro,
can you post what is exactly the problem with the wave-function? At what stage yambo stops? please post your input/output/report and any other information that can help to spot the problem.
Best,
Daniele

[amolina]

Dear Daniele,

you have here a sample of my input-output files. As you will see, even I choose the option "d" for the BSE solver, yambo does not write the wfs... The thing is that the BS spectra is done.

Cheers,

Alejandro.

[Daniele Varsano]

Dear Alejandro,
in order to plot the exciton wavefuntion yambo needs to weite the eigenvector of the excitonic Hamiltonian. This could be disk space consuming and it is not done by default. In order to do that you need to activate the WRbsWF variable. As I can see from your yambo.in file it is commented:

Code: Select all

#WRbsWF  

So in order to have it you need to uncomment that variable and diagonalize again the BSE hamilotonian. In order to do that, may be you need to remove the (n)db.BS_diago file from the SAVE directory.

Best,

Daniele