the LRO and SRO theories remained for 40 years until in 1971 the pioneering work by Sayers, Stern,
and Lytle performed the k-space Fourier transform to separate peaks corresponding to various atomic
shells. [5] After that it is widely recognized that the SRO theory is more appropriate in interpreting the
origin of XAFS. Another key development in the XAFS theory is that of an accurate treatment of
multiple-scattering (MS) effects. In 1975, Lee and Pendry first proposed the theoretical treatment of
multiple-scattering EXAFS in terms of effective path length. [6] This approach takes the advantage of
the Fourier transform in the EXAFS data analysis. A major advance in the path by path XAFS theory,
which overcomes all the major computational difficulties, is based on a separable representation of the
Green’s-function propagators developed by Rehr and Albers. [7]-[8] With the use of the second
generation machine, XAFS theory was improved gradually [9]-[12] adapted to better resolved fine
structures in the experimental spectra. Nowadays, nearly all parameters (e.g. temperature, [13] angular
[7], [14], [15] and spin [16] dependence) are fixed by ab initio starting conditions, except the treatment
of multi-electron excitations (MEEs) which is phenomenologically described by an amplitude
reduction factor
2
0
S (see equation (1) in the next section) weakly depending on energy only and
approximated by a constant in practical calculations.[17] So far, various advancements in the
theoretical models have been obtained such as full multiple-scattering (FMS),[18] self-consistent field
and fully relativistic calculations [19] especially for a quantitative analysis of the near-edge regime.
Different approaches also address the interaction of the core hole with the photoelectron and thereby
pave the way for the theoretical description of an excited state spectroscopy.[19]-[20] Recently,
theoretical standards eliminate the arbitrariness in the experimental analysis as it was done in the early
days by means of tabulated standards.[21] The EXAFS theory is implemented in various ab initio codes
like CONTINUUM, [22] EXCURV, [23] FEFF, [19] GNXAS, [24] WIEN2k, [25] the Munich
SPRKKR package, [26] NSRL-XAFS[27] and others. [12]
It should be remarked that the controversy between the early XAFS theory, i.e., SRO and LRO, actually
lies in the energy-dependent competition between scattering strength and inelastic losses. The excited
high-energy photoelectron state is not infinitely long lived, but must decay as a function of time and
distance and hence cannot probe long-range effects. This decay is due to the inelastic losses (extrinsic
losses) as it traverses the material, and the intrinsic lifetime of the core-hole state (intrinsic losses).
Therefore, the excited photoelectron outgoing wave dies away as it moves further away from the
absorbing atom and becomes too weak to significantly reflect any waves off of distant atoms. The net
effect is that XAFS can only measure the local atomic structure within a range limited by the effective
mean free path of the excited photoelectron. This range is typically on the order of tens of angstroms.
Furthermore, it turns out that a fully quantitative treatment of the near-edge structures is still
challenging. This is due to various many-body effects, such as the photoelectron-core hole interaction,
multiplet effects, the photoelectron self-energy and inelastic losses. Various theoretical treatments are
based on the final state rule in a one-electron approach. A different approach for the calculation of the
near-edge structures is the atomic multiplet theory.[28] Unfortunately, both approaches have their
specific drawbacks.[29] Recently, in some of works a GW approach is used.[30] Another approach is
based on a time-dependent density-functional theory (TDDFT), as it has been successfully applied to
optical excitations in atoms.[31] The application of this approach to x-ray absorption spectra at the L2,3
edges for the 3d transition metal series has been demonstrated recently.[32]
As for the experimental development, from 1930’s to 1960, XAFS spectra were taken in transmission
with low-power conventional X-ray tubes. Over the last three decades, the technique of XAFS has
experienced a rapid development. This is primarily due to the application of synchrotron radiation (SR)
which has a number of unique and excellent properties, such as high brilliance, wide and continuous
wavelength, polarizability and low emittance. In the 1970s, with the advent of the first dedicated
synchrotron radiation sources using the ‘white’ synchrotron radiation of bending magnets[33] (i.e.
SSRL), the data quality and the acquisition times for the XAFS spectra were greatly improved, although
those spectra were allowed only for the investigation of the nearest neighbor distance and the local
bonding geometry.[34] The first XAFS spectrum recorded on synchrotron radiation was measured by
Stern et al in 1974,[35] which signaled the beginning of rapid development in measurement techniques.
A transmission technique utilizing a channel-cut monochromator and ion chamber was developed,[36]
http://www.paper.edu.cn