The key step was resolving the lines
To learn chemistry from electrons that photons knock out of matter, an instrument must resolve small differences in their energy; a broad envelope of emitted electrons is not enough. The width of the source line, the electron optics, stray fields and the detector response all shape the measured line. Collecting more counts cannot sharpen an energy distribution the instrument has already blurred.
Kai Siegbahn’s lecture traces the path from electron spectroscopy in nuclear physics to sharp photoelectron spectra and chemical shifts. The 1957 paper on the precision method lists Carl Nordling, Evelyn Sokolowski and Kai Siegbahn as authors. This work is separate from Manne Siegbahn’s X-ray spectroscopy, recognized in 1924, and from the laser spectroscopy honoured in the same 1981 prize year.
Evidence and definitions: K. Siegbahn, Electron spectroscopy for atoms, molecules and condensed matter (Nobel lecture, 1981)C. Nordling, E. Sokolowski & K. Siegbahn (1957), Precision Method for Obtaining Absolute Values of Atomic Binding Energies
Know which particle the detector receives
In X-ray photoelectron spectroscopy, X-rays go in and the energy analyzer collects the electrons that come out of the specimen. In energy-dispersive X-ray analysis, the detector collects emitted X-ray photons instead. Both can identify elements, but the signals travel different paths and keep different information about the chemical environment.
An electron spectrum is a count of electrons at each analyzer setting. Before assigning a peak, identify the photon source, the energy reference and any photoelectron or Auger features that might overlap it. An element label on a simulated peak is an interpretation someone supplied, not a measurement of chemical identity.
Evidence and definitions: K. Siegbahn, Electron spectroscopy for atoms, molecules and condensed matter (Nobel lecture, 1981)Physical Electronics (2026), Why does the sample work function drop out of the energy accounting for photo-electrons in XPS?
Binding energy needs a stated zero
When a conducting specimen is electrically connected to a calibrated analyzer, the two share a Fermi level, and the relation below measures binding energy from it. The kinetic energy is the value at the analyzer, and the work function is the analyzer’s. Subtracting the specimen’s work function as well would mix two reference conventions.
The relation assumes electrical equilibrium, a known photon energy and no uncorrected sample bias. Gas-phase ionization energies are measured from the vacuum level, which is a different convention. An insulating specimen can charge up, and then the simple shared reference no longer holds. The equation applies under these conditions; it is not a universal correction for every surface.
Evidence and definitions: Physical Electronics (2026), Why does the sample work function drop out of the energy accounting for photo-electrons in XPS?
Same element, different local environments
The chemically useful question is whether one core level splits into distinguishable components when atoms of the same element sit in different environments. Siegbahn’s lecture discusses the two sulfur environments in thiosulfate and, in Fig. 3, shows the carbon spectrum of ethyl trifluoroacetate. Such splittings say far more than the fact that different elements have different core energies.
A chemical shift reflects bonding, screening and final-state relaxation together; it is not a voltmeter reading of an isolated atom’s charge. The measured energy needed to remove an electron includes how the remaining electrons respond to the core hole. Assign peaks using the whole spectrum and suitable references, and do not turn each fitted component into an exact oxidation number.
Evidence and definitions: K. Siegbahn, Electron spectroscopy for atoms, molecules and condensed matter (Nobel lecture, 1981)
Charging can imitate chemistry
A charged insulator can shift peaks even when bonding is unchanged. If emission leaves an insulating region positively charged, escaping electrons lose kinetic energy and appear at higher binding energy on the usual scale. Uniform charging can shift many lines together; uneven charging can shift or broaden components by different amounts. Neither effect, by itself, shows a change in bonding.
Before claiming a chemical shift, check the electrical contact, the neutralization conditions and whether the energy reference is sound. A steady spacing between related peaks can point to a common offset, but this test does not settle every case. A fit can match a charged spectrum very closely and still give it the wrong physical meaning.
Evidence and definitions: Thermo Fisher Scientific, Charging Effects in X-Ray Photoelectron Spectroscopy
Signal fades with depth instead of coming from a fixed slice
An electron that loses energy on its way out may drop out of the sharp photoelectron peak and into the background of energy losses. For a flat, uniform layer, a simple model with straight escape paths makes the signal fall off exponentially with depth. In the expression below, depth is measured inward from the surface and the emission angle from the outward surface normal.
The attenuation length depends on the electron energy and the material; in practice the effective length also depends on the application and geometry. NIST distinguishes these lengths from the inelastic mean free path and other depth measures. Collecting electrons at a grazing angle weights the signal more toward the surface, but one spectrum cannot uniquely recover an arbitrary depth profile, and it does not count every atom in a slice of fixed thickness.
Evidence and definitions: C. J. Powell & A. Jablonski (2020), Effective attenuation lengths for different quantitative applications of XPS — NIST
Peak areas come with assumptions
A peak area depends on how many atoms are present, the photoionization probability, the analyzer transmission, attenuation and how the background is subtracted. If the specimen is assumed uniform, areas corrected by sensitivity factors can estimate relative composition. An oxide layer on a substrate is weighted differently with depth from the substrate itself, so the ratio of their raw peak areas is not automatically a bulk atomic fraction.
In the Lab, first tell a shift in peak position apart from a change in relative signal. Then check charging and take-off angle before calling anything a chemical change. High-resolution spectroscopy made these comparisons possible. It did not remove the need for references, control of the specimen and a model of how electrons reach the detector.
Evidence and definitions: C. J. Powell & A. Jablonski (2020), Effective attenuation lengths for different quantitative applications of XPS — NISTThermo Fisher Scientific, Charging Effects in X-Ray Photoelectron SpectroscopyK. Siegbahn, Electron spectroscopy for atoms, molecules and condensed matter (Nobel lecture, 1981)