How to relate k point index to coordinates

Anything regarding the post-processing utility (e.g. excitonic wavefunction analysis) is dealt with in this forum.

Moderators: Davide Sangalli, andrea marini, Daniele Varsano

yhypku
Posts: 12
Joined: Wed Sep 02, 2020 11:10 am

How to relate k point index to coordinates

Post by yhypku » Sun Dec 06, 2020 6:46 am

Dear all,
I ran a 36x36 BSE of 2D BN on yambo and follow the advice of Daniele in poster: viewtopic.php?f=9&t=1742&p=9073&hilit=weight#p9073
to print the real and imaginary part of A(cv)_lamda. However, I only get the index of the k points('ik' in the code), not the exact coordinates. For my case, the maximum index is 1296(36x36), and I would like to know how to deal with them properly.
Thanks in advance for your hard work for the community of Yambo.
Best,
Huaiyuan
Huaiyuan Yang
Peking University
Beijing, China

User avatar
Daniele Varsano
Posts: 2997
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: How to relate k point index to coordinates

Post by Daniele Varsano » Sun Dec 06, 2020 8:08 am

Dear Huaiyuan,
in each report, you have the k points coordinate corresponding to the indexes (check for Energies [ev] & Occupations section):
you find something like:

Code: Select all

*X* K [1] : 0.000000  0.000000  0.000000 ( cc) * Comp.s 160123 * weight   0.05556
              0.000000  0.000000  0.000000 (rlu)
  E -44.84562 -44.82319 -44.79297 -44.78954 -44.77050 -44.76947 -44.71061 -44.70914
....
*X* K [9] :-0.275149  -0.05404  0.000000 ( cc) * Comp.s 160264 * weight   0.05556
             -0.500000 -.4470E-7  0.000000 (rlu)
  E -44.84553 -44.82310 -44.79297 -44.78954 -44.77050 -44.76946 -44.71061 -44.70914
....
Between [ik] is the index, next there are the coordinates:
cc : cartesian coordinates
rlu : reciprocal lattice units

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/

yhypku
Posts: 12
Joined: Wed Sep 02, 2020 11:10 am

Re: How to relate k point index to coordinates

Post by yhypku » Sun Dec 06, 2020 9:03 am

Dear Daniele:
Thanks for your reply. I'm sorry that I didn't explain my question clearly. The k points in your reply is the IBZ kpoints, but I would like to know all the kpoints information, cause in the output 'ik' could get 1296(36x36). Is there a convenient way to do that?
Best,
Huaiyuan
Huaiyuan Yang
Peking University
Beijing, China

User avatar
Daniele Varsano
Posts: 2997
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: How to relate k point index to coordinates

Post by Daniele Varsano » Sun Dec 06, 2020 9:39 am

Dear Huaiyuan,
sorry, I didn't understand,
Anyway, if I have understood now if you want information on k points in the whole BZ, you can add these two lines (see e.g. lines from 158 and beyond) after the line 84 indicated in the post you mentioned:

Code: Select all

ikibz = Xk%sstar(ikbz,1)
 is = Xk%sstar(ikbz,2)
and print ikibz and is.

In this way, for each k points in the whole BZ, you have the corresponding k point in the IBZ and the related symmetry operation.
Symmetry operations are indexed and reported in the report file.

Hope this answer to your question,
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/

yhypku
Posts: 12
Joined: Wed Sep 02, 2020 11:10 am

Re: How to relate k point index to coordinates

Post by yhypku » Sun Dec 06, 2020 10:24 am

Thanks, it helps me a lot. :)
Huaiyuan Yang
Peking University
Beijing, China

yhypku
Posts: 12
Joined: Wed Sep 02, 2020 11:10 am

Re: How to relate k point index to coordinates

Post by yhypku » Wed Dec 09, 2020 9:06 am

Dear Daniele:
I followed your suggestion and print out the Acv_lamda in the whole BZ. The system is 2D BN, using 36*36 kgrid to run the BSE. What confuses me is that the exciton of index 1 and index 2 have the same Acv_lamda in the whole BZ, but they have different oscillating strength. Could you please tell me what's the reason for that?
Here are my E_sorted file.
# Maximum Residual Value = .49576E+00
# E [ev] Strength Index
#
5.473043918610 1.000000000000 1.000000000000
5.473090171814 0.910880088806 2.000000000000
(Acv_lamda for 1 and 2 are too large to add into attachments. They are just the same.)
Best,
Huaiyuan
Huaiyuan Yang
Peking University
Beijing, China

User avatar
Daniele Varsano
Posts: 2997
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: How to relate k point index to coordinates

Post by Daniele Varsano » Wed Dec 09, 2020 10:14 am

Dear Huaiyuan,
you have two degenerate excitons, most probably coming from degenerate transitions.
Note that the strength is normalized to its maximum reported in the output ( 0.49576), so the two strength differs for about 1%.
When you say they have the same A_cv_lambda you refer to their modulus? I would say that A_cv_lambda for two eigenvalues cannot be the same.

Important: when analyzing excitons, it is recommended to see to zero the degeneracy threshold, otherwise, two degenerate states are merged together. In order to do that you need to set: Degen_Step=0.0 eV in your ypp input file, as by default it is set to 10^-3 eV. This could be the reason why you have two identical eigenvectors.

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/

yhypku
Posts: 12
Joined: Wed Sep 02, 2020 11:10 am

Re: How to relate k point index to coordinates

Post by yhypku » Wed Dec 09, 2020 1:10 pm

Dear Daniele:
I'm afraid that their Acv_lamdas are the same too. The whole file is too large to add into attachment so I just post part of it. The outputs represent i_lamda, ikbz, ikibz, is, iv, ic, ReA, ImA in order. File A1 is the part of the Acv_lamda of the first exciton and A2 is that of the second exciton. I really don't know the reason.
Best,
Huaiyuan
You do not have the required permissions to view the files attached to this post.
Huaiyuan Yang
Peking University
Beijing, China

User avatar
Daniele Varsano
Posts: 2997
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: How to relate k point index to coordinates

Post by Daniele Varsano » Wed Dec 09, 2020 1:17 pm

Dear Huaiyuan,
have you seen my comment above about degeneracy step?
Can you check the normalization of each lambda \sum_cv |A_cv|^2

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/

yhypku
Posts: 12
Joined: Wed Sep 02, 2020 11:10 am

Re: How to relate k point index to coordinates

Post by yhypku » Wed Dec 09, 2020 1:47 pm

Dear Daniele:
I set degen_step = 0 and the eigenvectors become different. Thanks for your help.
Best,
Huaiyuan
Huaiyuan Yang
Peking University
Beijing, China

Post Reply