An error occurred when using truncated Coulomb potential for BSE calculation

[202321070091]

Dear all,
I want to calculate the exciton binding energy of two-dimensional materials, so I started working on low dimensional systems. When I reached the BSE calculation stage, I encountered an error "[WARNING] Exchange Kernel FFT size is too big. NG_X reduced 110983 --> 18183 (35939 mHa),[WARNING] Bigger FFT discarded to avoid slow computation for corr part of Kernel,[ERROR] STOP signal received while in[09.01.05] BSE Kernel @q1 (Resonant CORRRELATION EXCHANGE),[ERROR] Reading File ./SAVE//ndb.kindx; Variable Bindx; NetCDF: Variable not found". How can I solve this problem? Additionally, I have learned a tutorial on how to analyze excitons. May I ask if the exciton binding energy is in the o-2D-WR_WC. exc_weights at_1 file in the tutorial example h-BN?
Yours sincerely, WangQi

[Daniele Varsano]

Dear WangQi,

I do not think that problem is related with the use of truncated Coulomb potential. Here something happened to the ndb.kindx file.
You can try to delete it in the SAVE directory and rerun your calculation.
If this does not solve the problem please post your input/report file.

The binding energy it is not reported and need to be calculated as the difference between the electronic GW gap and the excitation you want to analyse.
The excitation values are reported in o.*E_sorted

Best,
Daniele

[202321070091]

Dear Daniele,
Thank you for your continuous help. I'm sorry I haven't resolved this issue yet. In my initial report,"X grid is uniform" displayed "yes", but when it came to BSE calculation, it displayed "no", and there was an error in the output report of BSE. Then I would like to ask if the calculation of GW gap is the same as the calculation of G0W0 when dealing with low dimensional systems.
Yours sincerely, WangQi.

[Daniele Varsano]

Dear WangQi,

you can try the following.
1) Delete the ./SAVE/ndb.kindx file
2) rerun the setup adding in your input the flag NoDiagSC. A simple input file containing:

setup
NoDiagSC

3) Then redo the BSE calculation.

Then I would like to ask if the calculation of GW gap is the same as the calculation of G0W0 when dealing with low dimensional systems.

Sorry, but I do not get your question. What do you exactly mean? The procedure for 2D is the same and you can use cutoff coulomb potential to remove spurious interaction among replica. Usually k point convergence is slower for 2D, but you can accelerate using the method described in: NPJ Computational Materials 9, 44 (2023)

Here you can find a tutorial to apply the method:
https://www.yambo-code.eu/wiki/index.ph ... 2D_systems

Best,
Daniele

[202321070091]

Dear Daniele,
Thank you for your answer. I executed the command "yambo - i", added the parameter NoDiagSC, and then executed "yambo -F yambo.in". In the report "r-2D_WR_WC_rim_cut_optics_dipoles_bss_bse_em1d_ppa_01", "X grid is uniform" is still no. I deleted "./SAVE//ndb.kindx" and executed the command "yambo -J 2D_WR_WC -F yambo_BSE.in". Finally, there is still an error message "[WARNING] Exchange Kernel FFT size is too big. NG_X reduced 171131 --> 27083 (29260 mHa),[WARNING] Bigger FFT discarded to avoid slow computation for corr part of Kernel,[ERROR] STOP signal received while in[09.01.05] BSE Kernel @q1 (Resonant CORRRELATION EXCHANGE),[ERROR] Reading File ./SAVE//ndb.kindx; Variable Bindx; NetCDF: Variable not found". Also, I am very sorry that due to language issues, I often cannot describe my problems clearly. The question I would like to ask is, I want to calculate the exciton binding energy, which requires the value of the GW bandgap. Can I use the bandgap value of G0W0 in report "o-2D_WR_WC.qp" instead of the GW bandgap value.
Yours sincerely, WagQi.

[Daniele Varsano]

Dear WangQi,

can you send the report files r_setup with and without the NoDiagSC keyword? Please post also the nscf.in file.
Please delete the ndb.kindx file before running the two setups.

Can I use the bandgap value of G0W0 in report "o-2D_WR_WC.qp" instead of the GW bandgap value.

This should be the same if the GW correction has been calculated for all the k points.
From the report file I can see that indeed they are.

Code: Select all

[K+QP] Direct Gap                                 :  5.916764 [eV]

Eqp @ K [ 3]
... -0.094440  5.822323...

Best,
Daniele

[202321070091]

Dear Daniele,
Thank you for your answer. May I ask in which report can the results of GW correction be viewed? Is this the link(https://www.yambo-code.eu/wiki/index.ph ... 21_version) for calculating GW. The file r_setup in the attachment does not include the keyword NoDiagSC.
Yours sincerely, WagQi.

[202321070091]

Dear Daniele,
Thank you for your answer. May I ask in which report can the results of GW correction be viewed? Is this the link(https://www.yambo-code.eu/wiki/index.ph ... 21_version) for calculating GW. The file r_setup in the attachment does not include the keyword NoDiagSC.
Yours sincerely, WagQi.