The Raman(NLO) calculation for the ferromagnetic system

Moderators: mverstra, aromero

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

The Raman(NLO) calculation for the ferromagnetic system

Post by andyamygto » Sun Dec 15, 2019 4:42 pm

Dear Abinit users,

I was trying to performing the FULL DFPT calculation to get the Raman tensor for some ferromagnetic system.
In my experiment , I let the code determine the total spin-polarization, by using a metallic value for occopt(e.g. 4 7)
But in the calculation i got the ERROR message:

chkint_prt: ERROR -
Context: the value of the variable usepead is 0.
Context: the value of the variable optdriver is 5.
The value of the input variable occopt is 4, while it must be
equal to 1
Action: you should change one of the input variables occopt,
usepead or optdriver.

So I try to determine the spin by using the spinmagntarget variable
But again, the code send me an ERROR message:

--- !ERROR
src_file: m_chkinp.F90
src_line: 2888
mpi_rank: 0
message: |
spinmagntarget was input as 4.00
For a response function run, spinmagntarget is required to be 0.0d0 or the default value.
Action: modify value spinmagntarget or nsppol in input file.
...


Dose that means i cannot calculate the non-linear properties for ferromagnetic system ?

By the way i can finish the calculation if i set the spinmagntarget to 0 .
But it will be the anti-ferromagnetic.

I'll be glad to hear any kind of suggestion.
Thank you!

Andy

mverstra
Posts: 652
Joined: Wed Aug 19, 2009 12:01 pm

Re: The Raman(NLO) calculation for the ferromagnetic system

Post by mverstra » Thu Dec 26, 2019 3:15 pm

NL/Raman will only work for insulators. If your FM system is really insulating, you should be able to run with occopt 1. Recent versions of abinit should allow for metallic occopt as well (presuming the user knows what she is doing), but I'm not sure this was merged into the public branch yet. Which version are you using, and have you tried the latest public branch 8.10.3? The basic problem which had to be solved was that not only are the occupied number of bands strictly limited for an insulator, they also have to be different for the 2 spin channels.
Matthieu Verstraete
University of Liege, Belgium

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

Re: The Raman(NLO) calculation for the ferromagnetic system

Post by andyamygto » Thu Dec 26, 2019 6:55 pm

Thanks for the reply

I am using the latest public branch 8.10.3 .
I have tried to run the jobs with occopt 1 and it sends the ERROR


--- !ERROR
src_file: m_chkinp.F90
src_line: 2012
mpi_rank: 0
message: |
This is a calculation with spin-up and spin-down wavefunctions, ... nsppol= 2
in which the occupation numbers are to be determined automatically. ... occopt= 1
However, in this case, the target total spin magnetization
must be specified, while the default value is observed. ... spinmagntarget= -99.99
Action: if you are doing an antiferromagnetic calculation, please use nsppol=1 with nspden=2 ;
on the other hand, if you are doing a ferromagnetic calculation, either specify your own spinmagntarget,
or let the code determine the total spin-polarization, by using a metallic value for occopt (e.g. 7 or 4 ...).

I am not sure how to set the occupied number of bands to be different for the 2 spin channels.
Is that means i have to set the occ by myself?

Regard,
Andy Hsu

...

mverstra
Posts: 652
Joined: Wed Aug 19, 2009 12:01 pm

Re: The Raman(NLO) calculation for the ferromagnetic system

Post by mverstra » Thu Dec 26, 2019 10:55 pm

Use spinmagtarget, it's easiest.

Do you confirm your system is insulating? Otherwise there is no point...
Matthieu Verstraete
University of Liege, Belgium

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

Re: The Raman(NLO) calculation for the ferromagnetic system

Post by andyamygto » Fri Dec 27, 2019 1:30 am

Dear mverstra
Thanks for the reply

I am pretty sure that my system in insulating ( I have the reference and my own calculated band structure)
and i have used the spinmagtarget as well.
As i mentioned in the topics , i got the ERROR message when using spinmagtarget > 0

--- !ERROR
src_file: m_chkinp.F90
src_line: 2888
mpi_rank: 0
message: |
spinmagntarget was input as 4.00
For a response function run, spinmagntarget is required to be 0.0d0 or the default value.
Action: modify value spinmagntarget or nsppol in input file.

It's seems like i can do the anti-ferromagnetic calculation only.
Or do i miss something else?

Regards,
Andy Hsu

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

Re: The Raman(NLO) calculation for the ferromagnetic system

Post by andyamygto » Tue Mar 31, 2020 4:35 am

mverstra wrote:
Thu Dec 26, 2019 10:55 pm
Use spinmagtarget, it's easiest.

Do you confirm your system is insulating? Otherwise there is no point...
Hi

I just noticed that abinit release the new version.
Just wonderIng is there any other sight or solution to the problem I met?
I haven’t try the new version(9.0) yet
But I want to make sure is the problem I met is a bug or I just miss something while doing the calculation?

Cheers,
Andy Hsu

Post Reply