[Ifeffit] Re: from I.Demchenko - Institute of Physics PAS (question about FEFF6, 8)

Bruce Ravel bravel at anl.gov
Wed May 17 17:55:50 CDT 2006


Hi Iraida,

I am sending my answer to the mailing list -- in part because it is an
interesting and relevant question and in part because I am not sure I know the
correct answer to your question.

On Wednesday 17 May 2006 08:02, you wrote:
> Dear Bruce,
>
> After reading on site http://millenia.cars.aps.anl.gov/pipermail/ifeffit/
> about possible differences in calculations by FEFF6 and FEFF8 I've tested
> my models by both FEFF programs. And now, perhaps I'm making a stupid
> mistake in atom.inp file and next in feff.inp files.or I have a problem. I
> need to examine GaxMn1-xAs samples with the concentration of Mn about 1-2%
> (for example). The measurements were done for K edge of Mn. Within the
> initial model Ga atoms are replaced by Mn. My atom.inp file is present
> below:
>
> title = GaAs:Mn (Mn_Ga model)
> space = F -4 3 M
> a =     5.65330
> core =      Ga
> edge =      K
> rmax =        6.0
> out = feff.inp
> atoms
> ! elem   x          y          z     tag
>   Ga    0.00000    0.00000    0.00000
>   As    0.25000    0.25000    0.25000
>
> Then in the feff.inp file in "potentials" card I change the central Ga atom
> to Mn (in connection with the small concentration of Mn (1-2%) I didn't add
> any Mn atoms in the second and third coordination shells). For this model
> in the elementary cell we have: 48.4372% of Ga atoms; 50% of As atoms and
> 1.56% of Mn atoms. Our cluster is limited to 6 Å (rmax =        6.0).
>
> POTENTIALS
>  *   ipot   z [ label   l_scmt  l_fms  stoichiometry ]
>        0   25    Mn     -1      -1       0
>        1   33    As     -1      -1       1
>        2   31    Ga     -1      -1       1
>
>  ATOMS
>    0.00000     0.00000     0.00000    0   Mn               0.00000
>   -1.41332    -1.41332     1.41332    1   As               2.44795
>   -1.41332     1.41332    -1.41332    1   As               2.44795
>    1.41332    -1.41332    -1.41332    1   As               2.44795
>    1.41332     1.41332     1.41332    1   As               2.44795
>   -2.82665     0.00000    -2.82665    2   Ga               3.99749
>    2.82665    -2.82665     0.00000    2   Ga               3.99749
>    0.00000    -2.82665     2.82665    2   Ga               3.99749
>   -2.82665    -2.82665     0.00000    2   Ga               3.99749
>    0.00000     2.82665     2.82665    2   Ga               3.99749
>    0.00000     2.82665    -2.82665    2   Ga               3.99749
>    2.82665     0.00000     2.82665    2   Ga               3.99749
>   -2.82665     0.00000     2.82665    2   Ga               3.99749
>    0.00000    -2.82665    -2.82665    2   Ga               3.99749
>   -2.82665     2.82665     0.00000    2   Ga               3.99749
>    2.82665     0.00000    -2.82665    2   Ga               3.99749
>    2.82665     2.82665     0.00000    2   Ga               3.99749
> ..........
> ..........
>
> Then I do the calculations by FEFF6 and FEFF8 programs, respectively.
>
> In the table you'll find the differences in Amp for calculations done by
> FEFF6 and FEFF8 programs (there are values only for the first three single
> scat.).
>
>       #      degen      reff      Amp      Scat. path
>       1      4       2.448      100    feff6    Mn-As
>                                 100    feff8
>       2      12      3.998      88.07  feff6    Mn-Ga
>                                 101.44 feff8
>       5      12      4.688       64.42 feff6    Mn-As
>                                  67.22 feff8
>
>
>
> My question is: why do we have such big differences in Amp? Should I prefer
> the calculations done by FEFF6 program?

Asimilar question to this came up recently on the mailing list:

  http://millenia.cars.aps.anl.gov/pipermail/ifeffit/2006-March/002368.html

And here is my response, explaining my best memory of how that amplitude term
is computed

  http://millenia.cars.aps.anl.gov/pipermail/ifeffit/2006-March/002405.html

Jeremy and Feng also give useful responses to the original question in that
thread.

The bottom line is, I think, that the amplitude term is a very imprecise
measure of the actual size and importance of a path.  What's more, the exact
details of how it is inaccurately calculated may be different in Feff6 and
Feff8 (although I do not know that for a fact).

What seems like a much more relevant comparison between the two is how they
actually compare when plotted.  So, import both the feff6 and feff8
calculations into Artemis and compare the two calculations path by path in
k-space AND in R-space.  When you do that, how different are they?  And where
in k- and R-space are the differences most pronounced?  That will, I think, be
a much more accurate comparison of the two calculations.

Hope that helps,
B


-- 
 Bruce Ravel  ---------------------------------------------- bravel at anl.gov

 Molecular Environmental Science Group, Building 203, Room E-165
 MRCAT, Sector 10, Advanced Photon Source, Building 433, Room B007

 Argonne National Laboratory         phone and voice mail: (1) 630 252 5033
 Argonne IL 60439, USA                                fax: (1) 630 252 9793

 My homepage:    http://cars9.uchicago.edu/~ravel 
 EXAFS software: http://cars9.uchicago.edu/~ravel/software/





More information about the Ifeffit mailing list