2026A&A...707A.141L

Accurate and comprehensive atomic data are essential for the modelling of stellar spectra. Uncertainties in the oscillator strengths of specific lines used for abundance analyses directly translate into uncertainties in the derived elemental abundances; incomplete or biased atomic data sets can impart significant errors in non-local thermodynamic equilibrium (non-LTE) modelling. Theoretical calculations of atomic data are therefore crucial to supplement the limited experimental results. In this work, we present extensive atomic data, including oscillator strengths, transition rates, and lifetimes for 1730 electric-dipole (E1) transitions among 107 levels in neutral sulphur (S I) using the multi-configuration Dirac-Hartree-Fock (MCDHF) and relativistic-configuration-interaction (RCI) methods. These levels belong to the configurations 3p³np (n = 3 - 7), 3p³nf (n = 4, 5), 3s3p⁵, 3p³ns (n = 4 - 7), and 3p³nd (n = 3 - 6). The accuracy of the computed transition rates is assessed by combining the comparison of the differences in transition rates between the Babushkin and Coulomb gauges with a cancellation-factor (CF) analysis. Approximately 16% of the ab initio results achieved an accuracy classification of A-B, corresponding to uncertainties within 10%, as defined by the Atomic Spectra Database of the National Institute of Standards and Technology (NIST ASD). Applying a fine-tuning technique was found to significantly improve the accuracy of the results in the Coulomb gauge, thereby improving the consistency between the Babushkin and Coulomb gauges; about 24% of the fine-tuned transition data are assigned to the accuracy classes A-B.