Energy Range Problem in BSE XAS Calculation

[huiyao_kuang]

Dear VASP team,

I am using VASP 6.6.0 to calculate the Eu L-edge XAS using the BSE method for an EuO system. After a G0W0 calculation, I performed the BSE step with the following key settings:

Code: Select all

   ENCUT = 530
   ENCUTGW = 353.3
   PREC = Accurate
   NBANDS = 192
   ISMEAR = 0
   SIGMA = 0.05
   EDIFF = 1E-8
   LREAL = .FALSE.
   ALGO = BSE
   IBSE = 2
   ANTIRES = 0
   LADDER = .TRUE.
   LHARTREE = .TRUE.
   NBANDSO = 0
   NBANDSV = 64
   ICORELEVEL = 2
   CLNT = 1
   CLN = 2
   CLL = 1
   CLZ = 1.0
   OMEGAMIN = 6950
   OMEGAMAX = 7200
   NEDOS = 999999
   CSHIFT = 0.2
   CH_AMPLIFICATION = 10
   NCORE = 1
   KPAR = 8
   NSIM = 1

The imaginary part of the dielectric function spans 0–7200 eV, with all spectral weight crowded near ~7100 eV (the Eu 2p binding energy region). Even with NEDOS = 999999, the spectrum in the region of interest (~7050–7150 eV) is extremely coarse (see attachment) because nearly all grid points are wasted on the 0–7000 eV region.

I attempted the following to restrict the output energy range, without success:

1. OMEGAMAX = 300 with CLE = -7000: produced a completely flat (zero) spectrum.
2. CLE = -7000 with OMEGAMAX = 7200: no change in the energy axis, suggesting CLE is ignored in the BSE code path. I noticed that CLE does not appear in the related tags listed on the "Bethe-Salpeter equation for core excitations" wiki page. I wonder whether CLE is supported for BSE XAS calculations.
3. EMAX/EMIN: no effect on the BSE dielectric function output range.

My questions are:

1. Is there a supported way to restrict the output energy axis of the BSE dielectric function to the physically relevant window near the edge onset (e.g., 7050–7150 eV)?
2. Is CLE applicable to BSE XAS?
3. If no such tag exists, is the only solution to post-process the output and extract the relevant window?

Thank you for your help.

Best,
Huiyao