No Absorption with BSE calculation on monolayer MoS2

[pandachang97]

Hello Daniele,

I have new problems when calculating the absorption spectra of monolayer MoS2. I only include 3 atoms (just one MoS2) in the cell. All the PW calculation is done with QE 7.0 at PBE level. The pseudo potentials are all included in the folder, which are recognized by p2y. I have included the r-setup in the attachment. Then, I followed the tutorials on the Yambo wiki to run the static screening with the commands:

Code: Select all

yambo -X s -F 01_3D_BSE_screening.in  
yambo -F 01_3D_BSE_screening.in -J 3D_BSE  

.
The output is also included in the attachment.
Then, I go to the next step, Bethe-Salpeter kernel, in the same folder, as below:

Code: Select all

yambo -o b -k sex -F 02_3D_BSE_kernel.in -J 3D_BSE
yambo -F 02_3D_BSE_kernel.in -J 3D_BSE 

Finally, I run the third step, Bethe-Salpeter solver: diagonalization, in the same folder, as below:

Code: Select all

cp 02_3D_BSE_kernel.in 03_3D_BSE_diago_solver.in
yambo -F 03_3D_BSE_diago_solver.in -y d -V qp -J 3D_BSE
yambo -F 03_3D_BSE_diago_solver.in -J 3D_BSE  

Then, when I check the output (also included in the attachment), the imaginary part of dielectric constant is all zero (the real part is all 1, definitely). It seems there must be some errors in the input paras. Please help me check the input and feel free to let me know what the the problems are.

[Daniele Varsano]

Dear Xin,

looking at the BSE report, it seems that the job did not finish properly, at least the report is truncated. Please check it.
Next, if you are dealing with a 2D system, you need to look at the polarizability alpha and not at epsilon (which depends on the volume of the supercell and tends to (1,0) for large cell volume, in order to have the alpha you need to activate a truncated coulomb potential technique (yambo -r -).

Finally, not related to your problem, usually few bands across the gap are needed to get a converged BSE results, and you do not need to include all the occupied bands.

Best,
Daniele

[pandachang97]

Thank you so much for your advises, Daniele. I will include less bands in the calculations. Best regards,
Xin

[Reza_Reza]

Dear Daniele,

I am currently conducting BSE calculations to compute optical absorption. While attempting to converge the BSEbands, I noticed that the results have changed significantly when adjusting the number of bands. Consequently, I am unable to find the converged value. Could you please take a look at the results that I have attached here?

Thank you.

[Daniele Varsano]

Dear Reza,

it seems to me that the first part of the spectrum it is converged.
Of course, the larger the energy range you want to converge you will need to add bands in order to consider higher energy transitions.
Usually, one it is interested in the first excitations, where bound excitons lies or just above the gap.
I noticed you are not including QP corrections in the diagonal part of the BSE so you are doing BSE on top of KS energies, I do not know if this is intentional or not.

Best,
Daniele

[Reza_Reza]

Dear Daniele,

Thank you.
Do I should keep the value of the valence band to 61 or 62 which I converged, and increase the value of the conduction band?
In the case of including QP corrections, since I am converging based on the Yambo tutorial, it seems I can include QP corrections after the convergence of BSENGblk, BSENGexx, and BSEbands.
Or should i include QP?


Best,
Reza

[Daniele Varsano]

Dear Reza,

what matters is the electron-hole energy range of the bands you are including, usually few bands across the gap are enough to converge the first part of the spectrum. Yes, you can apply the QP correction after convergence tests.

Best,
Daniele

[Reza_Reza]

Hello,

I have calculated the G0W0 band structure and the QP optical absorption for a structure under strain. In the figure of energy-optical absorption, the positions of peaks are not the same as the gaps in the band structure figures. Is there a problem with the calculations?

I have attached all the related files. (The Yambo report for the figure named "2" in the jpeg file is separated into two parts due to the limitation in size.)


Best regards,

Reza

[Daniele Varsano]

Dear Reza,

it seems you are doing a BSE calculation (i.e. including excitonic effects), so in presence of bound excitons absorption peaks are found inside the gap. If you want to compare with the band structure, you need to plot the Independent Particle calculation that you found in the same output in column 4.

Next, I can see you are calculating spectra for many Q points, so I do not know if you are plotting the optical spectrum q-->0 limit, iq=1.

Best,
Daniele

[Reza_Reza]

Dear Daniele,

Thanks for your response, I plotted the figure output in column 4 (gnuplot: ........ u 1:4 w l) and attached the result (1_4).
If you need some more information, I can provide it.

regarding "plotting the optical spectrum q-->0 limit, iq=1.", I don't know how should I plot this.
It would be great if you could explain.

Best,

Reza