A laboratory study of gaseous sulfur hexafluoride has found that its sulfur core-valence spectra resemble the molecule’s valence spectrum in the positions of features and their intensity distribution, once differences in resolution are taken into account. The fluorine core-valence spectrum looks different: it carries additional features from 30.4 electronvolts upward, with intensity at 25 electronvolts and above possibly involving the 5a valence orbital.
That contrast also shows where the main calculation runs into trouble. Calculated spectra agree with the measurements in the outer-valence region, but deviations appear in the inner-valence region, especially for fluorine. The results are consistent with molecular-orbital breakdown beyond a model built around a doubly charged molecular state, but the paper presents that as an interpretation of the spectra and calculations.
How the researchers compared the edges
Researchers used time-of-flight photoelectron–photoelectron coincidence spectroscopy at PETRA III/DESY and in Gothenburg. The home laboratory instrument was a 5.6-metre magnetic-bottle spectrometer.
The gaseous sample was commercially obtained from Air Liquide, with a stated purity of 99.9 per cent. It was fed through 1-millimetre stainless-steel needles to produce effusive molecular beams.
The study covered the S2p, F1s and S1s core edges, using photon energies above each edge; comparison valence spectra were recorded at 21 and 41 electronvolts.
For interpretation, the researchers combined the spectra with quantum-chemical calculations. Their intensity model included photoionization, dipolar and monopolar coupling, shake-off, and spin coupling between dicationic states and two continuum electrons. The electronic structures were calculated with OSRHF-level RASSCF in OpenMOLCAS, using the cc-pVDZ basis set.
At sulfur, the valence pattern persists
Across the sulfur edges, the resemblance concerned the arrangement of the features rather than an identical energy range. The S1s and S2p outer-valence spectra extend about 4 electronvolts further down in energy than the valence spectrum recorded with 21-electronvolt photons. The sulfur core-valence patterns therefore retain the valence spectrum’s feature positions and intensity distribution while covering a wider interval.
That pattern appeared in both sulfur core-valence spectra after resolution differences were considered. The comparison is qualitative: the analysis reports a resemblance in feature positions and intensity distribution, rather than an exact overlay.
Fluorine is where the model strains
The fluorine edge produced the clearest departure. The F1s spectrum differed from the other core-valence spectra and showed extra features at 30.4 electronvolts and higher. The analysis says intensity from 25 electronvolts upward may involve the 5a valence orbital, so that assignment remains possible rather than settled.
In the 30–40 electronvolt range, the OSRHF calculation significantly underestimated the observed high intensity. A frozen-orbital reference supplied intensity after a correction for relaxation energy, but the comparison still leaves the source of the excess high-energy signal unresolved.
Calculations also reported a localized F1s core hole and symmetry breaking, while corresponding antisymmetric vibronic coupling was not reported after ionization at central sulfur. These are conclusions from calculated states and spectral analysis, not separately quantified measurements of symmetry breaking.
Spin changes the sulfur 2p picture
The calculations show that singlet–triplet splitting—the energy gap between two possible spin arrangements—depends on both the core edge and the valence orbital. It is negligible for outer molecular orbitals, rises toward higher binding energy, and is generally larger for S2p than for S1s. For the inner F1s 5a orbital, the calculated splitting reaches 2.5 electronvolts.
The S2p spectrum adds a distinct spin-orbit pattern: two parallel progressions separated by 1.2 electronvolts. Because the S2p spin-orbit splitting is much higher than at the 1s sites, the analysis uses an intermediate coupling regime that tracks j, the total-angular-momentum label, rather than spin alone.
A related deviation in the intensity ratio could not be resolved under the reported conditions.
Where the evidence stops
Taken together, the comparison places a clear boundary on the calculation. It works best in the outer-valence region; the model centers on shake-off and leaves out knock-out contributions, which are expected to matter more near threshold. OSRHF also excludes electron-correlation effects that may become more important for deeper core-valence states. Finite experimental resolution, state lifetimes and vibrational broadening limit some tests of peak separation and intensity ratios.
The evidence is limited to laboratory electron spectroscopy and quantum-chemical modeling of gaseous sulfur hexafluoride, one molecular system. The study does not report molecule counts, detected-event counts, replicate numbers or inferential statistical tests and formal uncertainty intervals.
Further tests would require higher-resolution measurements and calculations with explicit electron correlation to examine the proposed F1s localization, symmetry breaking and molecular-orbital breakdown. Measurements at lower excess photon energies could help assess the omitted knock-out contribution.
Paper data and sources
Original title: Core-valence double ionization of SF6 involving S2p, F1s and S1s inner shells
Authors: Veronica Daver Ideböhn, Daniel M. Pereira, Lucas M. Cornetta et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text