Bond lengths and coordination numbers fromL2,3-edge versusK-edge surface extended x-ray-absorption fine structure

The interference between final states of s and d symmetry arising from anisotropic $L_2$,3-edge absorption is explicitly considered. Previous arguments by others that such an interference can significantly affect the amplitude and phase of the $L_2$,3-edge surface extended x-ray-absorption fine structure (SEXAFS) had been based on calculated phase shifts and analysis of simulated data. Here, empirical methods are developed to isolate the effects of the interference. For the case of $L_2$,3-edge absorption from iodine it is shown that the calculated s- and d-state phase-shift differences are about a factor of 2 larger than the experimental value. This result accounts for the success of earlier $L_2$,3-edge SEXAFS amplitude analyses which approximate the interference by simple addition of the d and coupled s-d absorption terms. The previously argued effect on the SEXAFS phase correction is also shown to be too large by about a factor of 2 as a result of the analysis procedure used in that work. The effects of the interference are explored as a function of absorbing adatom, adsorption site, and accuracy of the calculated s and d-state phase-shift differences. With the assumption that these calculated values are reliable it is shown that only the atop configuration should manifest a measurable phase correction (corresponding to ∼0.01 Å in the bond length) and that the additive approximation used in the earlier SEXAFS studies of I and Te should be valid for all atoms. In the absence of data from other elements to test the accuracy of the calculated phase shifts, a general and theory-independent procedure was developed for analyzing $L_2$,3-edge data. This procedure removes the effects of the interference and allows bond lengths and adsorption sites to be determined with a reliability comparable to that obtained from K-edge SEXAFS data.