VASP Forum

Hi,

I am trying to apply a broadening to my XAS spectra dependent on the energy relative to the fermi energy. Therefore, I would like to know the corresponding value of the fermi energy in reference to the energies in the printed dielectric function output. On the wiki, I found "The energies of the excitations are with respect to the energy levels of the core electron of interest," however, when I take the difference between the fermi energy and the core-state eigenenergy I noticed there is still ~7-10 eV offset between this value and the absorption onset of the XAS spectrum, so I am still missing a factor. How do I directly calculate the corresponding band energy values in reference to the outputted dielectric function energies?

Thanks!

Dear nsolan,

Could you please provide the input and output files for your calculations so that I can better understand the context?

In a semiconducting system, the core-state excitation occurs into the lowest unoccupied bands, so the offset you mentioned could simply be the energy difference between the Fermi level and the lowest conduction-band state.

Best,
Alexey

Hi Alexey,

I am calculating the XAS of Li2S with the XCH approach and have attached the files. The peak onset in the imaginary part of the frequency-dependent dielectric function occurs at ~2397.5 eV which I believe should correspond to the band energy of the lowest conduction band (around 4.5 to 4.6 eV). I understand the energies in the dielectric function are referenced to the core electron energy, but I was wondering if you could provide a bit more information on how the energies printed in the dielectric function are specifically determined.

Thanks!
Nick

Hi Nick,

Thank you for providing the files.

In the XCH approach, the excited electron is simulated by adding one extra electron to the conduction bands and performing an SCF calculation. After the SCF cycle, this extra electron is removed, and the X-ray absorption spectrum is calculated. Therefore, the lowest-energy transition in your spectrum should correspond to the energy difference between the core state and the lowest unoccupied band after the extra electron has been removed.

In your calculation, after removing the extra electron, the lowest available state for excitation should be band 129:

 125       0.6716      1.00000
 126       0.8861      1.00000
 127       0.9954      1.00000
 128       0.9954      1.00000
 129       4.3787      1.00000
 130       4.5455      0.56060
 131       4.5819      0.18983
 132       5.0002      0.00000

Therefore, if we take the core-state energy from the POTCAR file, we obtain

E = 2393.1495 + 4.3787 = 2397.5282 eV

Best,
Alexey