Writing the data analysis section
The section titled "Please explain how you intend to analyse the data you collect" gives you an opportunity to convince reviewers that you have thought carefully about your proposed experiment, all the way through to data analysis. It also helps the beamline team understand how you intend to use the data, enabling them to make better recommendations regarding sample preparation and measurement conditions.
A good way to begin is with a clear statement of what you want to obtain from the analysis, followed by the method by which you will achieve it. Examples include:
The aim of the data analysis is to differentiate oxidation states of Fe within the sample suite. We will achieve this by collecting XANES spectra and monitoring the edge shift.
The aim of the data analysis is to identify the speciation of Cu in our homogenised plant material. We will use linear-combination fitting of Cu K-edge XANES spectra based on a suite of reference spectra collected from known compounds.
The aim of the data analysis is to identify the coordination environment of Zn in a novel protein. We have two candidate environments in mind, and will discriminate between them via EXAFS fitting using both candidate structures.
The aim of the data analysis is to demonstrate that Co in our catalyst does not exhibit the same spectrum as metallic Co, supporting the interpretation that it exists as a single atom. We will achieve this by qualitative comparison of the R-space spectra of our sample with those of a Co foil.
With a clear statement of aim, reviewers can assess whether your proposed analysis approach is appropriate.
Most XAS data processing follows a common sequence of steps:
Raw data are imported into an analysis package (commonly Athena/Demeter or xraylarch); multiple scans are aligned and averaged to improve signal-to-noise.
Pre- and post-edge lines are fitted to the data and used to normalise the XAS spectrum, placing the edge step at unity.
For EXAFS analysis, a spline background is fitted to the normalised data to isolate the EXAFS oscillations χ(k), which are then weighted by k², k², or k³ as appropriate.
The weighted χ(k) is Fourier transformed to produce the R-space spectrum, which approximates a pseudo-radial distribution function around the absorbing atom.
For XANES analysis, features of interest — edge position, white-line intensity, and pre-edge features — are extracted from the normalised spectrum, and linear combination fitting or fingerprint comparison is performed as appropriate.
For EXAFS fitting, theoretical scattering paths are calculated using FEFF and fitted to R-space (or k-space) data in Artemis or an equivalent program, refining structural parameters such as coordination number, interatomic distances, and disorder (σ²).
Other aspects worth addressing:
Your strategy for monitoring and correcting energy drift during the experiment.
Your approach to aligning spectra from different samples in energy space.
Your approach to absolute energy calibration.
If you are performing EXAFS fitting, how do you intend to determine the amplitude reduction factor S₀²? If you have thought this through, it is well worth communicating here.