EELS calculation, implemented broadening functions

[tomohisa_ogawa]

I have one technical questions on EELS calculation.
A calculated EELS spectrum is smeared/broaden so that one compares it with experimental results.

I understand three kinds of broadening are known in EELS (referring to Unoccupied Electronic States, Fundamentals for XANES, EELS, IPS and BIS, Textbook, 1992)

1). instrumental broadening, which is expressed in Gaussian function with FWHM determined by the device
2). core-hole lifetime broadening, which is probably hundled with Lorentzian smearing scheme
3). excited-state lifetime broadening, which requires certain energy-dependent broadening

I found the comments in XAS tutorial like
“The broadening of the core electron absorption spectrum is controlled by the tag CH_SIGMA. Usually it is good practice to set this value low and broaden the spectrum in post processing.”
“Another important issue is the broadening. Because the observed broadening is driven by many factors depending on the particular experimental setup, ....”

Does CH_SIGMA correspond to No.2) ?
If I’d like to apply No.1), Should I use SIGMA in VASP tag?
Is No.3) implemented in VASP? how can I control it?

I may have missed some descriptions, but I would appreciate your comments.
Best regards,Tomohisa Ogawa

[martin.schlipf]

Just to be clear, you are only interested in core absorption spectra, aren't you? You can obtain absorption spectra i.e. imaginary parts of the dielectric function with many different methods in VASP so you need to be sure that your technique matches the desired experimental reference.

For core spectroscopy specifically, we do not compute lifetimes so all peaks would be infinitely sharp. To produce a plot, you can introduce a finite broadening CH_SIGMA to the spectrum. In practice, you would take the resulting spectrum and convolute it with another broadening such that the width of the peaks matches the experiment.

[tomohisa_ogawa]

Dear Martin
Thank you for the useful comment

The mentions in the XAS tutorial or your message,
such as “… and broaden the spectrum in post processing” or “, you would take the resulting spectrum and convolute it with another broadening”.

This means that one has to prepare certain broadening function by oneself to recover the experimental spectrum from the VASP-EELS spectrum. (VASP has not yet implemented a function to realize the lifetime-broadning.)

My undestanding is correct??
Best regards,
Tomohisa Ogawa

[martin.schlipf]

Yes, VASP does have a broadening function where you control the width of the broadening with CH_SIGMA. However, if you use this feature you need to rerun the calculation to change the width of the broadening. For this reason, we recommend using a very small broadening and then using your own broadening function as a postprocessing.

PS: Sometimes it may also be necessary to have a energy dependent smearing function because the experimental apparatus smears higher energy features more than lower energy ones.

[tomohisa_ogawa]

OK, I understood the situation above,
Thank you for the kind comments. If new implementation is made for the smearing, please teach me about it
Best regards,
Tomohisa