Hi!
I have some Diffraction Anomalous Fine Structure (DAFS) data and ideally I want to use it to obtain f' and f''. I can fit the data very well by fixing f' and f'' to values obtained from XANES, and can then even fit the "fine structure" observable in the diffracted intensities. Allowing f' and f'' to refine "freely" (I use a penalty on each value so it does not deviate too much from the values at the previous energy, and I have fixed the values of the first energy to expected values), I also get a good fit, but the f' and f'' curves looks nonphysical.
I expect there to be strong correlations between f' and f'', and also with other parameters in the refinement like scale factor, thermal parameters and adsorption correction. When refining just one overall scale factor for 100 diffraction patterns, fixing the thermal parameters and calculating mu (and thus muR for absorption correction) from f'', I am still not able to obtain more reasonable f' and f'' curves (especially f' is problematic), although the fits are still very good! At this point, the only thing I am refining individually for each xdd is 7 background parameters, zero error, f', f'' and energy/wavelength (the first wavelength is fixed).
With this in mind, I thought I could try to link f' and f'' through their Kramers-Kronig relation in the refinement! Maybe stupid, but worth a try? Would this be possible to do, and if so, how would I do it?
Cheers
Ola