Hi Paul,

Actually, the inequivalent sites do interfere coherently. EXAFS chi's aren't like quantum mechanical wave functions, for which only relative phase matters. The phase gives information about E0, and that's a physically meaningful quantity. In other words, if one of the sites is producing a peak at 7200 eV and another is producing a trough, they will tend to cancel.

So yes, Artemis and Ifeffit do the right thing, but the right thing is to add everything up coherently regardless of which feff calculation it comes from.

A few other notes on this kind of fit: you must, of course, weight the amplitudes of each feff  calculation by the relative number of sites of that type. Also, it is quite possible that you'll need different e0's for the different sites.

--Scott Calvin
Sarah Lawrence College



  A simple question.  I am have been doing multiple edge fits of different materials for a while in artemis (diana).  Now, I would like to analyze a material with inequivalent sites for one of the edges.  I understand how to read in the data and have read in the different feff calculations (for each inequivalent site), but what I just want to confirm that artemis/diana/ifeffit will do the "right" thing, namely coherently add the contributions *within* each feff calculation and add the intensities of each feff calculation -- e.g. I don't think there is any interference between the XAFS signal on different crystallographic sites -- or even if there is it is ignorable at least for first neighbor interactions.  If there is anyone out there with experience in doing this (or is the right thing done automatically?), drop me (or the list) a line.  Thanks a lot!