The magneto-optical Kerr effect

[uguryorulmaz]

Dear all,

We are trying to calculate magneto-optical properties of monolayer CrI3 like in paper of Alejandro Molina-Sánchez and et all. (https://doi.org/10.1039/D0TC01322F)

We used Quantum Espresso (QE) code version v7.1. Our calculations include noncollinear, SOC and Hubbard completely same in paper of Alejandro Molina-Sánchez and et all. (https://doi.org/10.1039/D0TC01322F)

After passing QE to Yambo code (v5.0.4) with p2y, we constructed input files using this command like in tutorial at Yambo wiki website (https://www.yambo-code.eu/wiki/index.ph ... err_effect)

Code: Select all

yambo -optics bse -coulomb -V resp -F yambo_kerr.in

And then we calculated both the magneto-optical properties and standard absorption, but we do not obtain any results about absorption or Kerr effect. The results which we obtained are given in zip file.

My Questions:

1- There is no database in tutorial webpage (https://www.yambo-code.eu/wiki/index.ph ... err_effect). May you share that database?
2- Why did not we obtain results? Which step did we do wrong?

[Davide Sangalli]

Dear Uğur Yorulmaz,

for the simulations, simply you are not specifying the kind of solver you would like to use, e.g. the command to generate the input file should be

Code: Select all

yambo -optics bse -coulomb -Ksolver h -V resp -F yambo_kerr.in

To see the available options please use

Code: Select all

yambo -h Ksolver

After that you should obtaine the absorption and Kerr output files.

For the tutorial, unfortunately it has not yet been fully restored, after the migration of the tutorials to the wiki.
As you see below the tutorial title

Code: Select all

WARNIGN: This page is presently under construction

It is in our "to do list". Hope it will be ready soon. I'll notify you when it is ready.

Please notice that working on top of LDA+U there some technical aspects to be considered, e.g.
- when doing GW the <U> term is not removed together with Vxc, this leads to a partial double counting since GW contains in part the physics captured by U
- the standard approach to compute the dipoles ignores the U, which is a non local term in the hamiltonian. You can use other approaches to the yambo dipoles

Also, since you include SOC, you might want to check the effect of a recent fix we did in the code.
To this end I'd suggest you to download the latest version of the code (5.2.0)
https://github.com/yambo-code/yambo/wik ... gz-format)
You might encounter a barrier for LDA+U conversion. You can just switch this off

Code: Select all

grep "+U" interfaces/p2y/*.F

and comment the corresponding lines before running p2y.

Best,
D.