search

zerothi / sisl

Electronic structure Python package for post analysis and large scale tight-binding DFT/NEGF calculations
https://zerothi.github.io/sisl
Mozilla Public License 2.0
173 stars 57 forks source link

fhi-aims hamiltonian #694

Open zerothi opened 4 months ago

zerothi commented 4 months ago

Describe the feature

It would be great to allow reading in fhi-aims hamiltonians. In this way we could have an all-electron Hamiltonian access point.

@ialcon here your help could be really valuable, trying to get this up and running requires knowledge on the aims side (which I don't have), and also there is plenty of code in sisl to see how to read stuff from.

ialcon commented 4 months ago

Yes, this would of course be a fantastic extra tool for sisl. I am worried though of a couple of things. First (maybe not that important) is about FHI-AIMS being a private code. Do you know if there would be any issues by implementing these tools without their collaboration/permission? Probably this is not a problem, as sisl would only be used for these purposes by FHI-AIMS licensed users after all, but I prefer to get your opinion on this - as far as I recall, you already contacted the FHI-AIMS developing team time ago, but that was not a very successful attempt, right? Secondly, and this worries me a bit more (at least as of know), I have not directly worked with hamiltonian objects from FHI-AIMS - my experience (and that of the regular user of FHI-AIMS, I believe) is based on plotting band-structures and DOS plots (or real-space wave-function distributions, etc) which are directly created from FHI-AIMS via specific parameters within the control.in input file. That being said, I will take a look at the manual, as it should be possible to get the hamiltonian printed in some file format.

zerothi commented 4 months ago

I have already talked with Volker about this, there should be no issues. :)
sisl will post-process fhi-aims output. So there should be no licensing issues (as long as one does not copy code from aims).

As for the Hamiltonian output, yeah, that was the main problem. It was not that they weren't interested, however, they of course didn't have time ;) So if the manual tells something about writing out the Hamiltonian, that would be ideal!

ialcon commented 4 months ago

Ok, @zerothi , I will get into it from next Monday/Tuesday. Once I get some useful info from the manual I will let you know, and will start working on it.

zerothi commented 3 months ago

this is a duplicate of #580