[Pw_forum] Comparing xspectra to experimental data: absolute energy shift

[Guido Fratesi]

Dear Hud,

I potentially see two sub-questions in your message:

1) if I have a system with various C atoms, concurring to the same 
K-edge but possibly slightly misaligned, how to align them with respect 
to each other? Let me call this an "internal alignment" of the spectrum.

Here you have to consider that various C atoms may have different C1s 
binding energy (that is, the 1s energy wrt vacuum level), as well as a 
different alignment of the empty energy levels with respect to the 
vacuum level. Since the photon energy is the difference between the two 
quantities, both the core-level-shift and such alignment have to be 
taken into account as a shift of the xanes.dat files before summing 
them. This is what we do for aromatic molecules and we described it in 
J. Phys. Chem. C, 2013, 117 (13), pp 6632–6638 DOI: 
http://dx.doi.org/10.1021/jp312569q
If yours is a bulk system, the experimental C1s energy may be referred 
to the Fermi energy rather than the vacuum level, and so in the 
calculations.

2) once a total spectrum is determined (with arbitrary energy 
reference), or if you have just one C species, how to align this to 
experiments?

The spectra "shifted internally" are referred to an unknown quantity, 
that is however a constant, that is the average core level binding 
energy for your C atoms (see E_i^avg in Eq.4 in that JPCC paper), that 
for C amounts to about 285eV (that is indeed the shift you need to make 
a sensible comparison). So you are left with few options, at least:

- take such average core level binding energy from experiments (C1s-XPS)
- evaluate that with some theoretical approach (e.g., all-electron / PAW)
- assume an empirical shift (that would be the only empirical parameter 
in your calculation)

Kind regards,
Guido


On 05/08/2016 00:57, Hud Wahab wrote:
>
> Hello QE community
>
> I am trying to compare the xspectra (5.3.0) and experimental 
> absorption data at the carbon K-edge, that is, the measured energies 
> for carbon K-edge between 285 eV and 320 eV.  How does one shift the 
> xspectra simulation to the absolute energy scale of the experiment?  I 
> think that means getting an element-specific absolute value of the 
> energy shift.  I understand that the calculated Fermi energy is one 
> quantity we need to shift by – but this value is nowhere near the ~285 
> eV needed to make a sensible comparison.
>
> And, if the absolute value was element-specific, in this case for 
> carbon, it presumably comes from the generation of the pseudopotential 
> and is hopefully recorded in the pseudopotentil file (e.g.
>
> C.star1s-pbe-mt_gipaw.UPF).  What are the units of the energy given in 
> this particular file?
>
> “<PP_GIPAW_CORE_ORBITAL>
>
>    1    0     N  L 1S     eig:  -25.48712189”
>
> -Hud
>
> UNSW Canberra
>

Guido Fratesi
Dipartimento di Fisica
Universita` degli Studi di Milano
Via Celoria 16, 20133 Milano
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20160811/d1ce4118/attachment.html>