BSE calculation

Run-time issues concerning Yambo that are not covered in the above forums.

Moderators: myrta gruning, andrea marini, Daniele Varsano, Conor Hogan

8813204602
Posts: 58
Joined: Tue Jun 15, 2021 5:53 pm

BSE calculation

Post by 8813204602 » Sun Feb 20, 2022 9:09 pm

Dear Yambo community forum

I have already done the full GW calculation for a 2D magnetic material and now I want to do the BSE calculations considering the truncated coulomb potential. I have some questions in this regard:
1) If the truncated coulomb potential is used in the calculation of BSE, is it correct to plot the imaginary part of polarizability alpha to examine the convergence of the BSEbands ,BSENGBlk and other parameters?
2) According to GW calculation output, the amount of DFT gap correction varies for up and down spin channels:
GW gap-DFT gap for spin-up= 1.93949
GW gap-DFT gap for spin-down= 2.138687

Which value should be considered as the value of scissor in setting parameter kfnQP-E?
3) To plot the optical absorption spectrum, is it correct to use the equation ε =1+4π*α2D/d to obtain the imaginary part of the dielectric function?
4) Is it correct to calculate the static screening using the input file created by command: yambo -r -b and to adjust the NGsBlkXs,BndsRnxs, RandQpts,RandGvec and CUTBOX parameters according to the same convergence values obtained from the GW calculation?

Thanks and Regards,
Mitra Helmi
Mitra Helmi,Ph.D student
Ferdowsi university of Mashhad
Mashhad,Iran

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: BSE calculation

Post by Daniele Varsano » Mon Feb 21, 2022 9:49 am

Dea Mitra,

1) Yes
2) Both, if you. add to your input file -V qp (or -V all) you sill see in your input the variables:

Code: Select all

% KfnQP_up_E
% KfnQP_dn_E
3) It depends on what you want to look at, you can plot directly the polarizability alpha
4) Yes, but you can avoid this calculation by using the ndb.pp file already calculated for the GW correction. By using the -p p string in your input. Yambo will read the static part of the dielectric matrix from the database already calculated.


Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

8813204602
Posts: 58
Joined: Tue Jun 15, 2021 5:53 pm

Re: BSE calculation

Post by 8813204602 » Wed Mar 02, 2022 1:03 pm

Dear Daniele,

I want to do BSE calculation on the top of the quasi-particle energies obtained from GW calculation. I used Qpkrange: 1| 221| 12| 21 in the calculation of GW, while the value of the convergence obtained for the BSEBAnds parameter in BSE calculation is:
BSEBands: 8|26 |
Given that ‌BSEBands parameter contains more bands than Qpkrange bands, can I do full GW + BSE calculation or I need to repeat the GW calculation again with:
Qpkrange: 1 |221 | 8| 26


Thanks and Regards,
Mitra Helmi
Mitra Helmi,Ph.D student
Ferdowsi university of Mashhad
Mashhad,Iran

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: BSE calculation

Post by Daniele Varsano » Thu Mar 03, 2022 1:04 pm

Dear Mitra,

yes you can do it and Yambo will automatically take care of interpolating linearly the GW energies of the missing points.
You can verify anyway that linear interpolation is reasonable by plotting the QP energy in function of KS energy (column ($3+$4) vs column ($3)) and see if the show a linear behaviour.

See e.g. step 3 in this tutorial:
http://www.yambo-code.org/wiki/index.ph ... 21_version

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

8813204602
Posts: 58
Joined: Tue Jun 15, 2021 5:53 pm

Re: BSE calculation

Post by 8813204602 » Wed Mar 30, 2022 10:20 am

Hi,

I have some questions about BSE calculations:

1) In BSE calculation with Haydock solver,when I set ‌BSHayTrs = +0.02 in the input file, the value of this parameter is reported zero at output:

Haydock |Accuracy (requested) [o/o]: 0.000000
|Accuracy (reached) [o/o]: 0.000000
|Iteration :1000

Is this normal and does not cause a problem in the final result?

2) To calculate zz component of the dielectric function for a 2D material, is it correct to set the LongDrXs and BLongDir parameters as follows?

LongDrXs: 1| 1 | 0 |
BLongDir: 0| 0 | 1 |

Thanks and Regards,
Mitra Helmi
Mitra Helmi,Ph.D student
Ferdowsi university of Mashhad
Mashhad,Iran

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: BSE calculation

Post by Daniele Varsano » Thu Mar 31, 2022 8:43 am

Dear Myrta,

1) most probably the "+" sign is disturbing the parse. Anyway, yambo should assign the default value that is safe. You can check it by inspecting the output, in the last columns you have the results for the n-1 iteration to be compared with the first columns (n iteration).

2) Yes that's correct. Anyway, note that the dielectric function itself will go to (1,0) when the supercell goes to infinity, ie it depends on the Lz size of your supercell. When using the cutoff coulomb potential you will have essentially (1,0) for the epsilon and the quantity to check is the polarizability.

Best,

Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

8813204602
Posts: 58
Joined: Tue Jun 15, 2021 5:53 pm

Re: BSE calculation

Post by 8813204602 » Mon Apr 04, 2022 12:42 pm

Dear Daniele,

I did the BSE calculation for a 2D magnetic material, but as you can see in the figure, the spectrum obtained from the BSE calculation using a simple scissor operator is completely different from the spectrum obtained from full GW, especially in the energy range 4-8 eV. Is this difference normal or is it due to a wrong in calculations?
I did the convergence tests completely for BSE calculation with a simple scissor operator. Is it correct to use the same convergent values to calculate the BSE with full GW ?

Thanks a lot for your support.

Best,
Mitra
You do not have the required permissions to view the files attached to this post.
Mitra Helmi,Ph.D student
Ferdowsi university of Mashhad
Mashhad,Iran

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: BSE calculation

Post by Daniele Varsano » Fri Apr 08, 2022 9:51 am

Dear Mitra,
this is reasonable, it means that a scissor operator cannot be applied.
Indeed, you have similar results for the onset but then for higher bands, the QP corrections do not follow a simple scissor behaviour.
You can inspect your QP wrt the KS energies and realize what's the behaviour. Usually, also stretching is needed to mimic the QP corrections, but
it can also happen that you cannot fit with scissor/stretching.
You may have a look at the plot in Step 3 of this tutorial:
http://www.yambo-code.org/wiki/index.ph ... 21_version

and see if scissor/stretching applies to your case.

Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

8813204602
Posts: 58
Joined: Tue Jun 15, 2021 5:53 pm

Re: BSE calculation

Post by 8813204602 » Sat Apr 09, 2022 9:30 am

Dear Daniele,

Thanks for your reply.

I did the convergence tests completely for BSE calculation with a simple scissor operator.Given that you said the difference between BSE calculations with a simple scissor operator and full GW means that a scissor operator cannot be applied, can I use the same convergent values for full GW calculation or do I have to do the convergence tests again from the beginning to calculate the full GW?

Best,
Mitra
Mitra Helmi,Ph.D student
Ferdowsi university of Mashhad
Mashhad,Iran

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: BSE calculation

Post by Daniele Varsano » Sat Apr 09, 2022 10:04 pm

Dear Mitra,
the convergence of the BSE kernel does not depend on the QP energies which enter in the diagonal part, you do not need to repeat convergences.
Best,
Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Post Reply