next up previous
Next: CONCLUSION Up: DISCUSSION Previous: Co-refinement of the EDAFS

DAFS on non-crystalline materials

The crystallographic and diffraction sensitivities of DAFS that enhance XAFS come from the Fourier transform nature of diffraction. This allows DAFS to pick out specific Fourier components of the density. How much of this crystallographic utility persists when DAFS is applied to less-ordered materials? Some of the interesting systems are quasicrystals, amorphous materials and liquids. Even in these systems the diffraction condition still selects a specific wavevector component of the density. So, roughly speaking, DAFS will provide the XAFS-like radial distribution function of the atoms surrounding the resonantly selected element type, weighted by the contribution of the resonantly scattering atom to the specific Fourier component of the density selected by the diffraction condition.

Standard crystallography cannot fully solve the structure of quasicrystals because the necessary diffraction information is distributed throughout reciprocal space in infinitely many weak ``quasi-Bragg'' peaks. This provides the motivation for pursuing further structural information with DAFS. Will all of the strong quasi-Bragg DAFS intensities look the same, or will they look different? The same question arises in amorphous materials and liquids: Does the EDAFS radial distribution function depend on the diffraction wavevector? It is known that the differential anomalous scattering (DAS) does depend on the diffraction wavevector for compositionally modulated glasses [43]. In this case, the different atom types appear preferentially at specific diffraction wavevectors, and consequently the DAS signals vary with the diffraction momentum transfer. What new structural and spectroscopic information can be obtained from EDAFS and DANES studies of non-crystalline materials?

The formal reason that DAFS can be used, in principle, to obtain more information about disordered and non-crystalline materials than conventional anomalous scattering, is that DAFS is sensitive to the pair and multiparticle correlation functions, while conventional anomalous diffraction is only sensitive to the pair correlation functions. DAFS is sensitive both to the pair and to the multiparticle correlation functions for several reasons: (1) Some of the oscillatory fine structure is produced by the interference effects of multi-leg photoelectron paths which include many (3, 4, 5, ...) atoms; the analogous sensitivity of XANES spectra to higher order correlations has been studied for over ten years [44]. (2) DAFS selects one wavevector with its diffraction condition (which is sensitive to the pair correlation functions) and then probes the associated multiparticle correlation functions with its fine structure signals. (3) Polarization analyzed DAFS probes the initial and final legs of the photoelectron path with more sensitivity than polarization analyzed XAFS because both the incident and the scattered photon polarizations, tex2html_wrap_inline2132 and tex2html_wrap_inline2134, can be varied independently. There has been a recent theoretical discussion of the potential application of some of these sensitivities to studies of the ternary correlation functions of amorphous systems [45].


next up previous
Next: CONCLUSION Up: DISCUSSION Previous: Co-refinement of the EDAFS