Page 1 of 1

BSE calculations in magnetic system

Posted: Wed Dec 02, 2020 1:14 pm
by yuanchang_li
Dear VASP users,
I am doing a g0w0+bse fatband calculation in magnetic system. I got the exciton origination in the BSEFATBAND file, but I don't know whether the spin-polarization is considered and the serial number of bands represent spin up or spin down. Could anybody please give me some suggestions? Here is my input file and should I add some additional tags in doing bse calculations for treating magnetic system? Thanks in advance.
ISTART=1
ICHARG=11
PREC=Accurate
ENCUT=400
EDIFF = 1E-08
EDIFFG = -0.01
#NSW=500
IBRION=-1
#ISIF=2
#ISYM = 2
ISMEAR=0
LREAL = .F.
#NPAR = 4
ISPIN = 2
MAGMOM = 12*0 4 4 4 4
SIGMA=0.05
LORBIT = 11
LWAVE = .T.
LCHARG = .T.
IVDW = 11
NBANDS = 480
PRECFOCK = FAST
LOPTICS = T
ALGO = BSE
NBANDSO = 10
NBANDSV = 6
NOMEGA = 96
ENCUTGW = 100
LSPECTRAL = T
NELM = 1
NEDOS = 2000
OMEGAMAX = 7
NBSEEIG = 10

Re: BSE calculations in magnetic system

Posted: Thu Dec 03, 2020 8:07 am
by martin.schlipf
BSE and spin polarized calculation are implemented, but the same NBANDSO and NBANDSV are used for both spin channels. To verify that what you are doing is correct, there are a few tests that you can do
wiki/index.php/BSE_calculations#Consistency_checks
Please also read the rest of that page, in particular the paragraph concerning the treatment of the local exchange-correlation to understand why spin-polarized BSE calculation may be a bit less accurate.