[Pw_forum] XSpectra - tail of the spectrum falling relatively quickly

[Yevgen Melikhov]

Dear all users of
QE,

I have recently
started using XSpectra package to calculate an X-ray absorption
spectrum (XANES to be more specific) for GaMnAs material K-edge of
Mn. Currently, I work with the simplest model: cubic structure, 8
atoms in total: 1 Mn, 3 Ga and 4 As, no core-holes, etc.

After setting up the
model, I encountered a feature in the spectrum, which I am puzzled
with:

It seems that the
spectrum is more-or-less OK near fermi level (I can cut occupied
states). Then, from 0 eV up to about 15-17 eV, the spectrum has
correct positions of 2 peaks at around 9 and 13.5 eV (the ratio
between them is 1.0, however, as oppose to the experimental value of
1.2; still I can leave with that and maybe correct it by changing
gamma from 1.6 to slightly higher values).

However, after
reaching the maximum at 13.5 eV, the spectrum falls down steadily
(like 1/x) and relatively quickly. The maximum value at 13.5 eV was
0.0013 and at 40 eV it is less than 0.0001. The experimental data
instead would oscillate around value 0.0012 (which is a usual
behavior for a spectrum).

I tried several
things (just in case):
- I increased
number of k-points in SCF calculations;
- I increased
number of k-points in XSpectra calculations;
- I increased
number of bands in SCF calculations;
In all these cases
the spectra were the same (in numerical sense, of course)!!!

I also changed the
gamma coefficient from very low (0.01), to reasonable values (0.7 and
1.6), to very high (5). The gamma coefficient affects the
calculations as expected, but could not “lift up” the tail at all. At the
very high value the spectrum becomes broad one with one maximum only,
i.e. as expected, but the tail keeps falling down...


So, my question is
whether anybody experienced such an effect before and what should I do in
order to correct it. 

Thank you in advance
for any hints/comments!

Yevgen Melikhov

Lab of X-ray and Electron Microscopy Research

Institute of
Physics, PAS
Warsaw, Poland
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20140710/95f1f569/attachment.html>