G0W0 band plot

[xjxiao]

Dear all,
I have some questions about plotting G0W0 band structures and would greatly appreciate your guidance.
1) Can Yambo perform Wannier interpolation of G0W0 scatter-point data?
2) Can energy ranges be specified for band plots (ypp output)?
Thanks for your help!!
Xiao

[Daniele Varsano]

Dear Xiangjun,

1) Yes, here you can find a tutorial:
https://github.com/wannier-developers/w ... rial_23.md

2) The energy range is determined by the number of bands you are interpolating.

Best,

Daniele

[xjxiao]

Dear Daniele,

1) Yes, here you can find a tutorial:
https://github.com/wannier-developers/w ... rial_23.md

Thank you for your response. I have carefully studied the examples in the provided links. However, while working on my own calculations, I encountered an issue and would appreciate your assistance.

My system has as many as 26 valence bands. I used Quantum ESPRESSO (QE) to calculate 64 PBE bands and successfully performed Wannier interpolation for the 8 bands near the Fermi level. Then, I used Yambo to compute GW corrections for the 10 bands closest to the Fermi level, hoping to interpolate 8 of them. However, I found that wannier90.x seems to require computing all valence bands during the fitting process; otherwise, it cannot correctly locate these 8 bands within the full set of 64 bands (more precisely, the software appears to directly apply the GW corrections of the computed 10 bands to the first 10 PBE bands).

Since GW calculations in Yambo are extremely time-consuming, I would like to compute corrections only for the bands near the Fermi level and still achieve Wannier interpolation. Is there a way to accomplish this?

[AntimoMarrazzo]

Dear Xiangjun,

In short, you need to calculate QP corrections for all the bands that you used in the DFT Wannierization, i.e., all the NSCF bands minus those possibly removed with the "exclude_bands" option in Wannier90.

An advantaged with respect to the DFT calculation is that you only need to calculate GW QP corrections on fewer k-points in the irreducible BZ and then use YPP to unfold them to the full BZ, preparing a uniform k-grid as required by Wannier90.

Hope that helps.

[xjxiao]

Dear Xiangjun,

In short, you need to calculate QP corrections for all the bands that you used in the DFT Wannierization, i.e., all the NSCF bands minus those possibly removed with the "exclude_bands" option in Wannier90.

An advantaged with respect to the DFT calculation is that you only need to calculate GW QP corrections on fewer k-points in the irreducible BZ and then use YPP to unfold them to the full BZ, preparing a uniform k-grid as required by Wannier90.

Hope that helps.

Dear AntimoMarrazzo,
Thank you very much for your answer. Adding the `exclude_bands` parameter when fitting the DFT results with Wannier90 can indeed reduce the number of GW bands that need to be calculated.
The tutorial provides a `ypp.in` file with the following parameters. May I ask which command should be used to generate this input file?

Code: Select all

wannier                      # [R] Wannier 90 I/O
Seed= "alsb"                     # Seed name
WriteAMU= ""                 # What to write (A, M, U)

[Daniele Varsano]

Dear Xiangjun,

the command line is :

Code: Select all

ypp -wannier

in general

Code: Select all

ypp -h 

provides you a command line help.

Best,

Daniele