calculation of spin hall conductivity (SHC) & spin berry curvature

[nilcm90]

Dear developers

I am trying to reproduce the results for spin hall conductivity (SHC) & spin berry curvature of Pt as given in arXiv:1907.09788 (2019) using github example29.

I am using same Pt.win & others file as given in example29.

But unfortunately, during post processing I am getting an error of "value of berry_task not recognised in param_read".
Can you please kindly guide me how to resolve this issue ?? I will be grateful to you for your kind help.
Thank you.

[qiaojunfeng]

Which version of Wannier90 are you using? What do you set at berry_task? It should be berry_task = eval_shc for calculating SHC.

[nilcm90]

I am using wannier90, version 3.0.0. I have set berry_task = eval_shc for calculating SHC as given in example29 for Pt.
I have attach the .win and .wpout file here. Please have a look.

Pt.zip

[qiaojunfeng]

The SHC code is available in the develop branch, not in the latest release. Your uploaded Pt.zip file is empty, so I can't find out if other problems exist in your input file. Can you please rerun your calculation with the latest develop branch?

[sfragkos]

Hello people!
Is there any version where I can calculate SHC and spin Berry curvature?
Thank you

[qiaojunfeng]

Hello people!
Is there any version where I can calculate SHC and spin Berry curvature?
Thank you

Try the develop branch:
https://github.com/wannier-developers/wannier90/archive/develop.zip
and have a look at the directory examples/example29 and the corresponding tutorial at doc/compiled_docs/tutorial.pdf

[sfragkos]

Hello people!
Is there any version where I can calculate SHC and spin Berry curvature?
Thank you

Try the develop branch:
https://github.com/wannier-developers/wannier90/archive/develop.zip
and have a look at the directory examples/example29 and the corresponding tutorial at doc/compiled_docs/tutorial.pdf

Thank you very much!
I tried it but I get the following error "value of berry_task not recognised in param_read"
I have tried both berry_task = eval_shc and berry_task = shc
Do you know how can I solve this problem?

[qiaojunfeng]

Hello people!
Is there any version where I can calculate SHC and spin Berry curvature?
Thank you

Try the develop branch:
https://github.com/wannier-developers/wannier90/archive/develop.zip
and have a look at the directory examples/example29 and the corresponding tutorial at doc/compiled_docs/tutorial.pdf

Thank you very much!
I tried it but I get the following error "value of berry_task not recognised in param_read"
I have tried both berry_task = eval_shc and berry_task = shc
Do you know how can I solve this problem?

You should download the link that I pasted and compile it. The SHC code is available only in this version. The correct input parameter is berry_task = eval_shc.

[sfragkos]

Hello people!
Is there any version where I can calculate SHC and spin Berry curvature?
Thank you

Try the develop branch:
https://github.com/wannier-developers/wannier90/archive/develop.zip
and have a look at the directory examples/example29 and the corresponding tutorial at doc/compiled_docs/tutorial.pdf

Thank you very much!
I tried it but I get the following error "value of berry_task not recognised in param_read"
I have tried both berry_task = eval_shc and berry_task = shc
Do you know how can I solve this problem?

You should download the link that I pasted and compile it. The SHC code is available only in this version. The correct input parameter is berry_task = eval_shc.

Ok problem fixed. I guess the .spn file is generated from the pw2wannier90 interface, correct?
Is it possible for VASP users to generate that file?

[qiaojunfeng]

Hello people!
Is there any version where I can calculate SHC and spin Berry curvature?
Thank you

Try the develop branch:
https://github.com/wannier-developers/wannier90/archive/develop.zip
and have a look at the directory examples/example29 and the corresponding tutorial at doc/compiled_docs/tutorial.pdf

Thank you very much!
I tried it but I get the following error "value of berry_task not recognised in param_read"
I have tried both berry_task = eval_shc and berry_task = shc
Do you know how can I solve this problem?

You should download the link that I pasted and compile it. The SHC code is available only in this version. The correct input parameter is berry_task = eval_shc.

Ok problem fixed. I guess the .spn file is generated from the pw2wannier90 interface, correct?
Is it possible for VASP users to generate that file?

Yes, the .spn file is generated by pw2wannier90. Maybe you can consult VASP developers to see if it can generate .spn file or not.

[stepan-tsirkin]

dear sfragkos (and others interested)

I sometimes produce .spn files from WAVECAR with the following script:
https://github.com/stepan-tsirkin/wannier90/blob/develop/utility/vaspspn.py

Note, that it uses only the plane-wave part of the pseudo-wavefunction (no PAW projectors aded), I am not sure if it can make serious problems, but so far it gave no issues to me.

[sfragkos]

Dear stepan-tsirkin and others,

1. Thank you very much , this was very helpfull. Do you have any script for .dmn file also?

2. kubo_smr_fixed_en_width tag is need to avoid near zero number in the denominator of the kubo formula. But what if i want to study a four-folded crossing point? How much should I put the tag in this case?
   What is the expected value of spin Ω at a four-folded crossing (Dirac) point ?
   Is it zero, because the spin Ω is opposite for the two bands that cross each other? Or, spin Ω is infinite since the denominator is zero?

[giovannipizzi]

Dear all, I think the original question has been answered. I also mention that the feature will be available in the upcoming release 3.1, that should be published in the next few days.

If you want to continue the discussion on the additional questions:

• if it's a more physical question, we suggest to continue on the mailing list (and maybe report here the final answer to the questions, possibly)
• if it's more of a code issue, I suggest to open a new issues focused on the problem

Thanks a lot!

[Simba2805]

Hello people!
Is there any version where I can calculate SHC and spin Berry curvature?
Thank you

Try the develop branch:
https://github.com/wannier-developers/wannier90/archive/develop.zip
and have a look at the directory examples/example29 and the corresponding tutorial at doc/compiled_docs/tutorial.pdf

Thank you very much!
I tried it but I get the following error "value of berry_task not recognised in param_read"
I have tried both berry_task = eval_shc and berry_task = shc
Do you know how can I solve this problem?

You should download the link that I pasted and compile it. The SHC code is available only in this version. The correct input parameter is berry_task = eval_shc.

Ok problem fixed. I guess the .spn file is generated from the pw2wannier90 interface, correct?
Is it possible for VASP users to generate that file?

Yes, the .spn file is generated by pw2wannier90. Maybe you can consult VASP developers to see if it can generate .spn file or not.

Dear all,

I am a VASP user. Using the wannier90.win file which is available in example29, as mentioned above I am also facing the same issue (the issue with which this thread initially started "param_read"). But when I compiled again my wannier90 with the link given in this thread for SHC I am getting the same but manipulated error "Error: berry_task=shc and shc_method is not set". After compilation, I am also getting this w90spn2spn.x. Do I need to use that?

This is what I'm trying to do:

1. First I am creating a WAVECAR file from a regular SCF VASP run.
2. I copy WAVECAR to another directory along with other VASP input files and INCAR I use LWANNIER90 = .TRUE. tag to create necessary files.
3. Here I create a wannier90.win file and then add necessary commands given in exampple29 to calculate SHC for my system.
4. I make another directory and copy all wannier90.* files and run the usual "wannier90.x wannier" to calculate SHC. But it returns the error I mentioned above.

[stepan-tsirkin]

@Simba2805 ,

first, you can combine 1 and 2 together

second, you need a wannier90.win file already at your step 2, in order to specify the starting projections.

you also need to produce a .spn file, which can be done using a script which is now part of the WannierBerri project.

After compilation, I am also getting this w90spn2spn.x. Do I need to use that?

I do not think so :

wannier90/src/w90spn2spn.F90

Lines 15 to 17 in 528a67f

	module w90_conv_spn
	!! Module to convert spn files from formatted to unformmated
	!! and vice versa - useful for switching between computers

I am getting the same but manipulated error "Error: berry_task=shc and shc_method is not set"

did you make sure the following line is in your .win file?

wannier90/examples/example29/Pt.win

Line 5 in 528a67f

shc_method = qiao