Coster-Kronig transitions are a special type of Auger-Meitner emission – in which the core-hole and relaxing electron come from the same shell. The relevant outcome from this process is that the lifetime of the lower energy core-hole will be shorter than its higher state counterpart (reminder that core-hole lifetimes impact XPS peak widths). In the example in the figure below, this would result in a decreased lifetime for the L₂ state as it is quickly filled by the decaying L₃ state after formation—resulting in the release of the Coster–Kronig Auger–Meitner electron.(1)

In a simplistic model, ignoring relaxation and screening effects, in order to emit an Auger–Meitner electron, we must have enough energy to overcome the ionisation energy of the electron in question. For conventional Auger–Meitner emissions, this is relatively trivial, since the difference between the original core-hole and the decaying electron tends to be quite large. For example, for a Ni 2p (L_2,3) core-hole, the hole is filled by a 3d (M_4,5) valence electron—with an energy difference of around 850 eV.

By contrast, the determining factor of the Coster–Kronig Auger–Meitner electron for the Ni 2p core-hole will be the doublet separation between the L₃ and L₂ states—since the process in question is an Ni L₃L₂M_4,5 Auger–Meitner emission.

Given this energy gap is the source of the exciting energy for the subsequent photoemission process, if we have a scenario where the doublet separation is quite small (e.g., molybdenum, figure (b) below) then for the cases of localised electron energy states (e.g., nonmetals), this energy may be insufficient to overcome the binding energy of the final Auger–Meitner electron, and therefore, this Coster–Kronig process is forbidden. In Figure (a), we report the measured full-width at half maximum (FWHM) for a titanium metal and oxide sample (doublet separation energy > ionisation potential for both system) and (b) below, a MoTe chalcogenide and MoO₃ oxide (doublet separation energy > ionisation potential for chalcogenide, doublet separation energy < ionisation potential for oxide), which highlights the forbidden process once the ionisation potentially exceeds the doublet separation of the initial core-hole orbital. The valence regions of the MoTe and MoO₃ are presented in (c) below, with the summary of the energetics involved in (d) below. From this figure, we can see that the only case where our valence band maximum (VBM) energy > our doublet separation, the peak widths are equivalent.

The degree of broadening, dependent on the rate of the Coster–Kronig transitions, is influenced by the kinetic energy of the emitted electron (higher kinetic energies = higher rate), number of electrons capable of participating in the Coster–Kronig process and screening effects

The degree of broadening for specific orbitals can be found in literature,(2, 3) and these broadenings can be used within modelling functions by adding a fixed value on to the measured FWHM of the higher energy (unbroadened) peak—though there are often other factors affecting the recorded spectrum such as plasmons or satellites, and as such, it might prove easier to model standard reference data before incorporating the developed peak models into your unknown dataset.

References #