Perturbative SOC

[Bruno]

Dear developers,

Is it possible to include the spin-orbit coupling perturbatively in Yambo? If that is the case is it possible to do it in both GW and BSE calculations and how would one do it?

Regards,

[Davide Sangalli]

Dear Bruno,
yes, it is possible.
Details here in section IV: https://doi.org/10.48550/arXiv.2103.02266 ( https://arxiv.org/abs/2103.02266 )

All you need is to compute a SOC mapping database (eq 68).

This is done running ypp

Code: Select all

ypp -F map_wfs.in -J SOC_map -C SOC_map

with the following input (this is just an example)

Code: Select all

#
WFs_map                   # [R] WF map WFs with SOC to present WFs
WFs_SOC_map                   # [R] WF map WFs with SOC to present WFs
SOC_DBs= "SOC_DBs"        # Path to the folder containing the DB with SOC
SOC_splitting= 0.01 eV    # SOC splitting between the levels
% bands_to_map
  5 | 16 |                # Number of bands to map
%
bands_buffer=1

where you run ypp from the folder containing the SAVE without SOC, while the SAVE with SOC is inside the another folder specified by the input variable "SOC_DBs".

After that you just run a standard BSE (eventually including QP corrections), allowing yambo to read the mapping DB

Code: Select all

yambo -F bse.in -J "BSE_run,SOC_map" -C BSE_run

Best,
D.

[Bruno]

Dear Davide,

Thank you very much! I will take a good look at the paper.

Best regards,

[Bruno]

Hello,

I manage to do the SOC mapping with no problems using "ypp -F map_wfs.in -J SOC_map -C SOC_map" as instructed. The database is fine and the output is alright. Now I'm trying to run a G0W0 calculation using "yambo -F bse.in -J "BSE_run,SOC_map" -C BSE_run" but yambo doesn't seems to read the SOC database. The GW calculations finishes fine but there is not indication on the outputs that the database was read, and even when I change the value of "SOC_splitting" during the mapping, I still obtain the exact same GW band-gap. Is there something that needs to be added directly on the GW input to read the database? If it's necessary to provide some files let me know. Thank you!

Regards,
Bruno

[Davide Sangalli]

The databse can be used in a BSE run (as the name of the input file suggests). The GW calculation is not affected.
Indeed the assumption is simply that the KS SOC splitting can be applied to the GW band structure.
This is done automatically by the code when loading the QP corrections in the BSE simulation.

Best,
D.