Dependence on QPkrange of electron-phonon self-energy for one state

[minyez]

Dear developers,

Greetings. My question is about the tutorial on electron-phonon coupling. In doing the "Optical properties at finite temperature" tutorial, I changed the `QPkrange` from "1|1|4|5|" in the previous EPC tutorial to "1|8|2|7|" such that electron-phonon self-energy (EPSE) is calculated for more k-points and bands. I found that this will lead to change in EPSE of |n,k>=|4,1> and |5,1>. With "1|1|4|5|", the o-T0.qp output is

Code: Select all

#    K-point            Band               Eo [eV]            E-Eo [eV]          Sc|Eo [eV]         Width [meV]
#
        1                   4                 0.000000           0.044131           0.045563           9.431183
        1                   5                 2.511291           0.003103           0.001955          -19.76448

While with "1|8|2|7|", o-T0.qp looks

Code: Select all

#    K-point            Band               Eo [eV]            E-Eo [eV]          Sc|Eo [eV]         Width [meV]
#
        1                  3                  0.000000           0.045297           0.046870           9.375368
        1                   4                 0.000000           0.045297           0.046870           9.375368
        1                   5                 2.511291           0.004694           0.002486          -25.78777
        1                   6                 2.511291           0.004694           0.002486          -25.78777
        1                   7                 2.511291           0.004694           0.002486          -25.78777
...

One can note slight difference in E-Eo, Sc|Eo and the width. If I understand correctly, for a particular |nk>, when the state is included in `QPkrange`, its EPSE is calculated exactly by the formula instead of interpolation. Therefore, as long as the state is included, change of `QPkrange` should not change its EPSE. Is my understanding correct?

Thank you very much!

Best Regards,
Min-Ye

Version:
- QE 6.8
- Yambo 5.0.4

[claudio]

Dear Min-Ye Zhang

thank you for reporting this issue, it seems a bug, I will check it and let you know

best
Claudio

[claudio]

Dear Min-Ye Zhang

your problem is due to the fact that Yambo averages the el-ph correction on the degenerate states.

Please include all degenerate states in your calculation in order to get the correct result.

best
Claudio

ps I added a warning in the tutorial

[minyez]

Dear Claudio

Thank you very much for the clarification. The results of the top 3 valence bands and bottom 3 empty bands are indeed the same using "1|1|1|8" and "1|1|2|7".

Best Regards,
Min-Ye