Page 1 of 1

fatband file and its Fermi level?

Posted: Mon Jun 06, 2016 7:59 am
by pipidog
Hello everyone,

I'm using "PAWFATBAND" to extract the components on band structure in systems with spin-orbit coupling (SOC).
I have three question:

1. the output files are named as:
xxxx_o_DS3_FATBANDS_at0001_Al_is1_l0_m+0

I believe "is1" means "spin up". However, there is always "is1" files. My abinit (8.0.7) doesn't output any "is2" files.
Is there anything wrong? (Note: I'm running a SOC calculation)

2. I tried to plot the data by myself. Unlike normal band structure calculation where the Fermi level has to be
read from the output file of SCF. The fatband file seems already shifted Ef to zero and always in unit of "eV".
Is that correct? (I tested on many cases and I always need to assume Ef=0 in the fatband output files to get
correct band structure. )

3. A final check, the fatband output file has two columns. The first corresponds the the eigenvalue and the
second corresponding to the square of the absolute values of the band state on a that component, i.e.
|<atom, l,m | psi_{En,k} >|^2, right?

Thanks,
Shu-Ting Pi
University of California, Irvine

Re: fatband file and its Fermi level?

Posted: Mon Jun 06, 2016 9:57 am
by pipidog
BTW, I've tested 7.10.5 and the Fermi level behaves as usual. There is no such shift in fatband output.

Re: fatband file and its Fermi level?

Posted: Tue Sep 13, 2016 9:45 am
by amadon
Hello,

1 For non collinear calculations (SOC requires a non collinear calculation), ABINIT uses spinors, so there is no more up and down wavefunctions. So output
files do all contain is1 in their names.

2. Yes in 8.0, the band structure is shifted such that the Fermi level is at 0 for convenience.

3. Yes

Best regards
Bernard