Incorreect Im(alpha)

[shench]

Dear Developer,

I have been utilizing your software for some calculations and am now encountering an issue that I believe requires your expertise and assistance.

In my calculations, I have noticed that the values of Im(alpha) seem to be incorrect, while the results for Im(alpha`) appear to be accurate.

To facilitate a thorough analysis, I have attached the relevant files. We would greatly appreciate it if you could review them and help me identify any potential errors that may have occurred.

Thank you very much for your attention to this matter. We look forward to your response and are eager to resolve this discrepancy.

Best regards,

Chen Shen

[Daniele Varsano]

Dear Chen Shen,
please sign your post with your affiliation, this is a rule of the forum, and you can do once for all by filling your signature in the user profile.

Can you please explain what you are aiming to calculate?

In your input file I can see you set many variables, and it is not clear what kind of calculation you want to do, e.g.
why are you using the inversion solver?
Are you sure you are interested in dichroism?
Why are you adding QP corrections in the screening?

Best,

Daniele

[shench]

Dear Daniele Varsano,

Firstly, I would like to express my gratitude for your prompt response.

I am currently conducting calculations aimed at obtaining the absorption spectrum and exciton binding energy of materials. In my computations, I have employed a double grid technique, which necessitated the use of the inversion solver.
I have noticed that the dichromism setting appears to be incorrect, as the inversion solver does not support the output of dichromism spectra.
Additionally, I would like to calculate the exciton effect under the quasiparticle (QP) correction. I have introduced the QP correction, but I am uncertain if this configuration is accurate. Could you please advise if the setting should be adjusted to “E > GW0/ndb.QP”?

Thank you very much for your assistance and guidance.

Best regards,
Chenhai Shen

[Daniele Varsano]

Dear Chenhai,

If you need to calculate absorption spectrum and binding energy, I suggest you to keep only BSEprop= "abs".
Yes, the QP correction is taken into account by setting: KfnQPdb= "E < GW0/ndb.QP"

Next, I suggest you as first step to not use the k grid at the beginning, start your test with a single grid using a diagonalization or iterative solver.
Then, you can switch to a double grid in case it is needed.

I'm not an expert on double grid calculation, maybe others more familiar can tell you what is going wrong.

Best,
Daniele

[Davide Sangalli]

Few more info.

The double grid technique can be used also with the Haydock solver (one of the iterative solvers suggested by Daniele). You can refer to this paper for details:
- https://www.frontiersin.org/journals/ch ... 63946/full

For dichroism, in case you are interested, please keep in mind that the implemented version has been tested only in isolated systems. Please refer to these papers:
- https://iopscience.iop.org/article/10.1 ... 12005/meta
- https://www.sif.it/riviste/sif/ncc/econ ... article/18

Also, dichroism beyond the independent-particles approximation will be available only in version 5.3, which we are finalizing:
- https://github.com/yambo-code/yambo/releases/tag/5.3.0

Best,
D.

[shench]

Dear developers,

I am writing to share my recent findings regarding the recalculated absorption spectrum. In this iteration, I decided not to use the double-grid technique and only kept BSEprop= "abs". However, I have noticed a significant difference between the absorption spectra obtained at the GW0+BSE level and the GW0 level. I am unsure whether this discrepancy is due to an error in my input file or if it is an expected outcome of the calculations.

I have attached the relevant files. Thank you very much for your help.

Best regards,
Chenhai Shen

[Daniele Varsano]

Dear Chenhai,

the input file, at first glance, seems to be ok.
Can you be more explicit about the discrepancy you are observing?
BSE gap (optical gap= and GW gap (electronic gap) are in general different in presence of bound excitons due to electron-hole attraction.

Best,
Daniele

[shench]

Dear Daniele Varsano,

Thank you once again.

I have attached the images of the absorption spectra. There is a significant difference in the shapes between the BSE absorption spectrum and the GW0 absorption spectrum.
Given your wealth of experience in this field, I was wondering if you could share your thoughts on whether these variations in spectral shapes might be considered normal.

Thank you very much for your assistance.

Best,
Chenhai

[Daniele Varsano]

Dear Chenhai,

variations in the spectral shape are usual. The fact you have a large peak in BSE where the IP spectrum is practically zero is instead a bit unusual.
I suggest you to carefully check convergences in the various parameters.
In particular, I suggest you to make a systematic check in the bands included in the BSE calculation. In order to speed up the calculation you can use the haydock solver.

Best,
Daniele

[shench]

Dear Daniele Varsano,

Thank you so much for your assistance. I have one final question that I hope you can help me with.

In the plot I created using the second (black line), fourth (red line), and sixth (blue line) columns from the output file in my first post, I noticed a significant discrepancy between the second and sixth columns. As I understand it, the sixth column, which represents Im(alpha'), should reflect the data from the penultimate step of the inversion's iterative process. On the other hand, the second column shows the data at the end of the iteration process.

If my understanding is correct, I am unsure why there is such a notable difference between these two columns. Could you please help me understand this?

Thank you once again for your support.

Best,
Chenhai