Hi Yang,
That's an interesting report. One possible explanation is that Larix / XAS
Viewer and Larch use Feff8 by default, while Artemis and (and maybe other
apps?) use Feff6 by default. That's basically because of when these
versions of Feff were released for us to distribute.
I believe that the consensus view is (and Bruce Ravel has demonstrated this
nicely) that there is very little difference between Feff6 and Feff8 for
EXAFS, with possible exceptions for very heavy elements (say, Z>80). For
example, self-consistent potentials are shown to be "mostly not very
important" for EXAFS results.
One common observation is that Feff8 is better at allowing hydrogen in
structures -- not so much that Metal-H scattering is important, but just
that the presence of hydrogen often messes up the calculations of the
atomic potentials in Feff6, and Feff8 seems to be noticeably better at that.
But, I am not sure whether Metal-Li has been carefully compared between
Feff6 and Feff8. Maybe someone here has tried that? You could try using
Feff6 from the CIF / Feff calculation page in Larix and see if there is a
difference. I think that would be interesting to document somewhere, maybe
in a paper and/or Jupyter notebook.
--Matt
On Fri, Mar 29, 2024 at 7:54 AM Hu, Yang (HIU)
Dear ifeffit members,
I would like to ask a few questions about the FEFF calculation and fitting of scattering paths between the light lithium (scatters) and transition metals (central abs.). For materials containing Li, there are Li scattering paths in FEFF calculations, but usually with low “rank” or low “importance”.
Through my limited experience, I often ignored those paths when treating cations-disordered materials, where Li share the same crystallographic site with other TMs. For the pristine materials, if I include the Li paths, the fitting got unstable especially for the paths degeneracy. Things became more complicated with Li extraction/insertion in cycled cathode materials.
Recently I am rechecking previous data (same cif, before XASViewer and now Larix), and I could “fit” some spectra with Li paths (the fitted parameters were not so crazy), which I never managed before (I tried a lot)… I am wondering if there have been developments in FEFF or software that would allow better treats of Li scattering paths in shell fitting?
The behaviour of Li and their influence on the TM local structure are quite interesting and important for our works, but I rarely see literature involving Li scattering paths. I just don’t know if it is worth continuing to invest time and efforts on this...... I really appreciate if you can provide some suggestions and opinions J
Thanks very much,
Best regards,
Yang
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov https://urldefense.us/v3/__http://millenia.cars.aps.anl.gov/mailman/listinfo... Unsubscribe: https://urldefense.us/v3/__http://millenia.cars.aps.anl.gov/mailman/options/...
-- --Matt Newville <newville at cars.uchicago.edu> 630-327-7411