The SXAS formulae are short. The time goes into beamtime, vacuum, and the sample. What follows is the shape of a typical synchrotron soft X-ray experiment. It is a process note, not a claim that a particular beamline has already been measured.

End station of a soft X-ray beamline. Sample, detectors, and the vacuum chamber are packed into this section. Image: Wikimedia Commons / Australian Synchrotron.

1. Write the proposal, then wait for time

A synchrotron station is not walk-up. Users submit a proposal: the scientific question, why this edge is required, the sample form, the polarisation, temperature, or in-situ environment, and how many edges will be scanned. Approved time is beamtime, counted in shifts, often blocks of about 8 hours.

Before that, decide whether the measurement is surface or bulk, whether it needs in-situ charge–discharge or a gas, whether the sample conducts, and whether it survives the beam. Those choices decide TEY versus TFY, and whether a soft X-ray station can do the experiment at all.

2. The sample: vacuum comes before a pretty picture

Soft X-rays almost always mean vacuum. Powders are pressed into pellets or spread on conductive carbon tape or indium foil. Films need a thought-through substrate and a ground path. Samples that outgas, contain water, or still hold solvent may need a pretreatment; otherwise the chamber never reaches vacuum and the whole line waits.

Insulators — oxide or sulfide solid electrolytes, polymer separators — readily charge under total-electron-yield detection. The spectrum then shifts and distorts. Before the shift, plan a close conductive-carbon contact, a very thin sample, or a fluorescence-yield fallback.

Bring references: metal foils, known oxides, graphite, carbonates. Without them, the trip home is one spectrum and no scale.

3. On the floor: from undulator to sample

Electrons in the storage ring pass an undulator and emit bright, nearly monochromatic light. The beamline then uses a grating or crystal monochromator to park that light on the absorption edge. At the end station the experimenter sees a photon energy that steps in eV, and an $I_0$ monitor for the incident intensity — often a gold mesh, a thin film, or a photodiode.

An undulator: a row of magnets with alternating polarity forces the electrons onto a snaking path and radiates bright, tunable light. Image: Wikimedia Commons / Australian Synchrotron.

Alignment usually means checking the energy calibration against a known edge or a reference, setting slits and spot size, and choosing the incidence angle. Grazing incidence shortens the probing depth; normal incidence averages more of the sample. The angle is part of the data and belongs in the notebook.

4. The scan: TEY, TFY, and the $I_0$ that gets forgotten

An edge scan is not a raw sample signal plotted against energy. The relative absorption comes from normalisation:

$$\mu(E) \propto I_{\mathrm{sample}}(E) / I_0(E)$$

Do not plot raw counts. Ring current decays, and grating efficiency moves with energy. Without dividing by $I_0$, a bump in the spectrum can be the source rather than the sample.

Collecting TEY and TFY together is common. If they disagree, do not average them yet. Surface oxide, adsorbed water, and carbon contamination make TEY look dirtier, while TFY can flatten the white line by self-absorption. Keeping both is more honest than handing in only the prettier curve.

Step size near the edge can be on the order of 0.1 eV and coarser before and after the edge. Dwell time is a trade between signal and damage. Organics, hydrogels, and some lithium-battery electrodes change shape while they are scanned. That is the beam altering the sample. Move to a fresh spot, reduce the flux, cool the sample, or accept that the edge is only qualitative.

5. After the shift: normalisation is not cosmetics

The usual processing is another energy calibration, a linear background, and an edge-jump normalisation on a region above the edge so different samples can be compared. XANES is overlaid on references. EXAFS, if it is actually being done, then goes into $k$-space and a Fourier transform.

A careful note or paper supplement says which beamline, which detection mode, the vacuum, the incidence angle, how the energy resolution was estimated, and what the references were. Without those, a reader cannot tell chemistry from an artefact.

Magnets and vacuum pipe in a storage ring. A beamline only borrows a beam from here. The measurement still depends on the sample and on normalisation. Image: Wikimedia Commons.

6. How this sits with the earlier notes

  • The synchrotron supplies tunable, bright, polarised soft X-rays.
  • SXAS uses that light to ask about the unoccupied states of one element.
  • The near edge (XANES / NEXAFS) reads valence and fingerprint. The extended region (EXAFS) reads bond length, if the energy window is long enough.
  • The experiment turns those ideas into TEY / TFY curves.

These notes are an entry map. Reading a real spectrum means comparing it with references from the same field and with the technical manual of the beamline. Machines are upgraded every year. The picture is more stable: electrons turn and emit light, the light sits on an absorption edge, and the sample reports how much it absorbed through electrons or fluorescence.

How to cite:

Tianlu Pang. How an SXAS experiment runs: from sample to spectrum[EB/OL]. PANG Blog, 2026. https://blog.pang.work/en/2026/08/22/sxas-experiment-from-sample/