Hi David, Sure thing. I'll send you a Pt file as a separate email; I'll also send you a copy of the paper we're currently preparing for publication. The final versions we used we're big Ifeffit scripts rather than Artemis files, but I'm sure I can scare up a representative Artemis file from earlier in the process. Just a brief comparison of the two methods I've seen out there: 1) Fit the coordination number of each shell. Anatoly Frenkel has several publication related to this. 2) Assume the crystallites are "roughly" spherical and constrain the coordination numbers in a fit to reflect this. The only size-related fitting parameter is then R, the radius of the crystallite. I have written a few papers related to this method. Method 1 has the advantage that it can in principle handle any morphology; Anatoly has used it with supported catalysts (they may even have been Pt, but I don't recall offhand) and distinguished between shapes like rafts, hemispheres, etc.. Method 2 has the advantage that is uses only 1 size-related parameter. For a complicated fit or one with limited data quality, that's nice. On the other hand, if the morphology is far from spherical (e.g. needles), then it doesn't work very well. This method can even be used with an empirical standard (as opposed to a theoretical standard like FEFF) to give a (rough) model-independent estimate of size. So I tend to think method 1 is better when you have a very uniform sample, as can happen with supported catalysts, while method 2 is often appropriate when you have a "messier" sample with roughly spherical particles, as is often the case when wet chemical methods are used. It should also be noted thatusing both techniques there have been many reports that for polydispersed samples EXAFS yields a mean smaller than that found by other methods (e.g. Scherrer analysis). Part of that can be explained by the different weightings inherent in the different techniques (EXAFS reads a 10 nm and 10 micron crystallite as essentially identical, but "notices" a 1 nm crystallite, while XRD would find the 10 micron signal dominant), but I am not yet convinced that there isn't something else at work as well. My group is currently planning a follow-up study of Pt to resolve this question. --Scott Calvin Sarah Lawrence College
Ifeffit Community:
Over the past year or so I have seen several discussions of estimation of metal cluster size (collections of several to hundreds of metal atoms) using coordination numbers derived from EXAFS fitting. I was wondering if anyone on the list would be willing to share an example of their Artemis project file where this analysis was done successfully for Pt and/or Au particle sizes. As a first pass I am just interested in roughly estimating an average particle size with future considerations for estimations of particle size distributions if data quality and theory allows.
Thanks in advance to anyone willing to share their project files. Given the number of experts in the field reading this list it just seemed to make sense to build upon the back giants.