One of the most universally accepted facts used in EXAFS analysis is that the amplitude reduction factor S02 is chemically transferable.

I've been trying to find a good reference for this--either a key study establishing it, or a review article asserting it. I know there has been plenty of good work in recent years trying to make theoretical calculations of S02, and that they are gradually becoming more accurate. But these recent advances can't be the reason we treat S02 as transferable, because we've been doing it for decades.

How was this established? It seems devilishly difficult to do experimentally with good accuracy, because S02 will correlate to some extent with other parameters.  Those correlations can be broken to some extent by using k dependence, but it seems the uncertainties would still be somewhat high. On the other hand, were there compelling theoretical reasons back in the 80's for believing transferability to hold?

Currently, how good do we think the assumption of transferability really is? Good to 5% for any compound at that edge? Good to 1% for compounds with similar local environments at that edge? Better than that? (I'm asking about the EXAFS region; say, more than two inverse angstroms above the edge.)

I eagerly await your collective wisdom, knowledge, and humorous anecdotes.

--Scott Calvin

Sarah Lawrence College