[Ifeffit] Odd feff result

[Scott Calvin]

Hi all,

I just ran a feff calculation using feff6L and got a result that 
surprised me a little. I was modelling a ferrite with a nickel and 
boron substituted for iron in a small fraction of sites. Here's part 
of files.dat:

     file        sig2   amp ratio    deg    nlegs  r effective
  feff0010.dat 0.00000     2.706     2.000     3   3.7141
  feff0011.dat 0.00000    26.472    20.000     3   3.7141
  feff0012.dat 0.00000     4.038     2.000     3   3.7141

and the corresponding part of the paths.dat:

     10    3   2.000  index, nleg, degeneracy, r=  3.7141
       x           y           z     ipot  label      rleg      beta        eta
    -3.126710    1.042240    1.042240   1 
'Ni    '     3.4567  150.5039    0.0000
    -1.146460    1.146460    1.146460   3 
'O     '     1.9857   58.9923    0.0000
     0.000000    0.000000    0.000000   0 
'Fe    '     1.9857  150.5039    0.0000
     11    3  20.000  index, nleg, degeneracy, r=  3.7141
       x           y           z     ipot  label      rleg      beta        eta
     1.042240   -3.126710    1.042240   2 
'Fe    '     3.4567  150.5039    0.0000
     1.146460   -1.146460    1.146460   3 
'O     '     1.9857   58.9923    0.0000
     0.000000    0.000000    0.000000   0 
'Fe    '     1.9857  150.5039    0.0000
     12    3   2.000  index, nleg, degeneracy, r=  3.7141
       x           y           z     ipot  label      rleg      beta        eta
    -1.042240    3.126710    1.042240   4 
'B     '     3.4567  150.5039    0.0000
    -1.146460    1.146460    1.146460   3 
'O     '     1.9857   58.9923    0.0000
     0.000000    0.000000    0.000000   0 
'Fe    '     1.9857  150.5039    0.0000

So here's the question. These three paths are identical except for 
the identity of one of the scattering atoms. The degeneracy of the 
path with boron and the path with nickel is the same. Why is the feff 
amplitude ratio for the boron-containing path higher than that of the 
nickel-containing path? Shouldn't nickel scatter more strongly? And 
yes, I double-checked...the potentials are defined correctly in the 
feff.inp file.

--Scott Calvin
Sarah Lawrence College