DFPT phonon with soc problem:rhoijim must be allocated!

Moderators: torrent, jzwanzig

ebousquet
Posts: 433
Joined: Tue Apr 19, 2011 11:13 am
Location: University of Liege, Belgium

Re: DFPT phonon with soc problem:rhoijim must be allocated!

Post by ebousquet » Sun Feb 23, 2020 9:30 pm

Dear Andy,
OK, soc and magnetism problem are different things.
In the latest Abinit school, Marc Torrent presented the new implementation of DFPT for non-linear response that should work for magnetism (and +U if PAW), see slide 13:
http://palata.fzu.cz/abinitschool/download/Th2_DFPT.pdf
Which method did you use? (the new method might not be updated on the tuto page, let me ask).
Best wishes,
Eric

andyamygto
Posts: 35
Joined: Tue Nov 26, 2019 12:50 am

Re: DFPT phonon with soc problem:rhoijim must be allocated!

Post by andyamygto » Mon Feb 24, 2020 8:42 am

Dear Eric,
I use the FULL DFPT which is the new implementation.
I have read slide just after it was uploaded. That's why i use the "FULL DFPT" since it says that it can deal the magnetism system.(slide 19)
and i met the problem i mentioned before.

Regards,
Andy

baguetl
Posts: 4
Joined: Mon Oct 05, 2015 9:06 am

Re: DFPT phonon with soc problem:rhoijim must be allocated!

Post by baguetl » Mon Feb 24, 2020 9:56 am

Dear Andy,

The error you see says that in the case of non-linear computations (optdriver=5) with the "FULL DFPT" method (usepead=0) the variable "occopt" should be 1 (0 should be fine too, I think).

Indeed this routine has been validated only in the case of insulators. It's true that in principle one can use metallic occupations with a very small smearing to describe an insulator, but the non-linear implementation has not been checked and validated in that case yet.

There is still some work to do in the non-linear implementation to make it available for more cases. We are working on it, and this specific problem will be taken into consideration. However it won't be released soon...

Best wishes,
Lucas Baguet

andyamygto
Posts: 35
Joined: Tue Nov 26, 2019 12:50 am

Re: DFPT phonon with soc problem:rhoijim must be allocated!

Post by andyamygto » Mon Feb 24, 2020 3:29 pm

Dear Lucas,

Thanks for the info.
So, seems like the magnetism system (At least ferromagnetic) with non-linear implementation is not available yet?
Or is there another way to achieve the such calculation?

Best wishes,
Andy

baguetl
Posts: 4
Joined: Mon Oct 05, 2015 9:06 am

Re: DFPT phonon with soc problem:rhoijim must be allocated!

Post by baguetl » Tue Feb 25, 2020 9:36 am

Dear Andy,

What I can say for sure is that the nonlinear implementation works for nspinor=1 and nsppol=1 or 2.

For now Abinit doesn't provide tools to compute Raman intensities without using DFPT...

Best wishes,
Lucas Baguet

andyamygto
Posts: 35
Joined: Tue Nov 26, 2019 12:50 am

Re: DFPT phonon with soc problem:rhoijim must be allocated!

Post by andyamygto » Wed Feb 26, 2020 3:20 am

Dear Lucas,

Okay. Anyway, thanks for the help and the info.
Hope this issue will make Abinit better.
I am looking forward!

I also want to mentioned that for a response function run, spinmagntarget is required to be 0.0d0 or the default value.
this issue also interrupt me when I try to do the calculation for the ferromagnetic system with DFPT.

Best wishes,
Andy

Post Reply