[Ifeffit] difference of feff interpretation generated by different version of artemis

[Matt Newville]

Hi Qiang,

On Wed, Oct 17, 2018 at 9:22 AM woschangqiang at 126.com <woschangqiang at 126.com>
wrote:

> Hello, everyone. I recently ran into an issue in using Artemis software.
>
> I  used two  versions of Artemis  (Artemis 0.8.012 and Demeter 0.9.25) to
> generate the paths from the same input file (FeS2.inp), but there is big
> difference between the two path results.
>

Well, maybe not a "big difference".  But, it might be reasonable to expect
slightly different results from different versions of the code.


> e.g. 1, the path4 and path6 generated in Demeter 0.9.25  are not
> generated in Artemis 0.8.012, though they have a very low applitude.
>


Feff has an option to not generate path data from paths with "too low an
amplitude".  What that means in detail depends slightly on the versions of
the program used.  And Artemis may have changed the settings it passes to
Feff as well.

You attached the input to Atoms, and the results from Feff, but not the
input to Feff.  Those are probably slightly different in the two versions.


> e.g. 2, the path7 in Demeter 0.9.25 is likely identical to path5 in Artemis 0.8.012
> judging from the degenracy, reff, and amplitude, but their scattering name
> are different: (+)-S1-S2-(+) in Demeter 0.9.25 and (+)-S2-S1-(+) in Artemis 0.8.012.
>   In my opinion, the scattring sequence of photoelectron in multiple
> scattering affect the paramters of paths. Why do these two paths have
> different scattering sequence? Thanks very much for your help.
>


Scattering paths that are simply reversed (Absorber-> S1 -> S2 -> Absorber
vs Absorber -> S2 -> S1 > Absorber) are equivalent.

--Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20181017/77f81f4a/attachment.html>