4
$\begingroup$

How to carry out Spin-projected band structure using VASP? I couldn't find any good source for this kind of plotting, is there a code for this kind of plotting ?

$\endgroup$
1
  • 1
    $\begingroup$ Are you looking for help doing the calculation with ISPIN = 2, or are you looking for plotting the results? $\endgroup$ Commented Jan 3, 2023 at 6:37

1 Answer 1

3
$\begingroup$

To perform spin-projected band structure calculations in VASP, you typically need to perform a self-consistent field (SCF) calculation first to generate CHGCAR file. The CHGCAR file is not directly involved in the spin projection process, but it is a necessary input file for the band structure calculation.

You should include the following tags in your INCAR file for both SCF and bandstructure calculations:

ISPIN = 2    ! Set ISPIN to 2 to enable spin polarization.
LSORBIT = .TRUE.    ! Enable spin-orbit coupling if needed.
LNONCOLLINEAR = .TRUE.    ! Enable non-collinear spin calculations if needed.

Assuming you have a relaxed structure and did convergence tests for total energy of the system. The following is a sample file for SCF calculation:

ISTART = 0
ICHARG = 2
NELM = 100
EDIFF = 1E-8
ENCUT = 400
ALGO = Fast
IBRION = -1
ISIF = 2
POTIM = 0.5
NSW = 0
EDIFFG = -0.01
ISMEAR = -5
SIGMA = 0.2
PREC = Accurate
ISPIN = 2 ! spin-projection
LREAL = .F.
LWAVE = .T.
LCHARG = .T.  !write the CHGCAR file
LAECHG = .TRUE.
GGA = PE !GGA-PBE functional
IVDW = 11 !vdW correction
ADDGRID = .TRUE.
LASPH = .TRUE.
LORBIT = 11

You need to check VASP documentation for further details. After getting the CHGCAR file from the SCF calculation, copy that file into a new directory where you intend to do the bandstructure calculation, remember to set the tags for reading the CHGCAR file. The following is a sample file for such calculation:

ISTART = 0
ICHARG = 11 ! Read CHGCAR file
NELM = 100
EDIFF = 1E-8
ENCUT = 400
ALGO = Fast
IBRION = -1
ISIF = 2
POTIM = 0.5
NSW = 0
EDIFFG = -0.01
ISMEAR = 0
SIGMA = 0.2
PREC = Accurate
ISPIN = 2
LREAL = .F.
LWAVE = .F.
LCHARG = .F.
LAECHG = .F.
GGA = PE
IVDW = 11
ADDGRID = .TRUE.
LASPH = .TRUE.
LORBIT = 11
NBANDS = 80

I have to mention that this is just an example and you need to understand the purpose of each of those tags for your own situation or structure.

Now assuming that you have your spin-projected bandstructure calculations done, VASP will write an .xml file called vasprun.xml. You may use this file to do the plotting using sumo-bandplot.

In a directoty where you have your vasprun.xml file along with the KPOINTS used for the bandstructure calculation, you can just run

sumo-bandplot --filenames vasprun.xml --spin 1

--spin will determines whether to plot the up spin (1) or down spin (-1). Further options such as projection of atomic contributions can be found in the documentation.

$\endgroup$
4
  • $\begingroup$ I want to upvote this, but are you able to go into more detail and make this a more substantial answer? I encourage you to look through the other answers on this site and see how many answers you can find (other than yours) that were this short. $\endgroup$ Commented Jan 3, 2023 at 4:49
  • 1
    $\begingroup$ That was long time ago, I will update the answer with detailed explanation $\endgroup$ Commented Dec 16, 2023 at 2:21
  • $\begingroup$ Thanks. We can turn off community wiki now. $\endgroup$ Commented Dec 16, 2023 at 4:34
  • $\begingroup$ Doesn't LSORBIT force ISPIN = 1, since spin is no longer a good quantum number and can't really be separated into spin up/down (though it can be projected on x, y, z)? That was my understanding, also seen in the VASP forum or Eric Welch's answer to this question on researchgate $\endgroup$ Commented Sep 16 at 23:40

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .