The solar sulphur abundance in view of large-scale atomic structure calculations and 3D non-LTE models

TL;DR

This work revisits the solar sulphur abundance by analyzing seven S I lines with a fully consistent 3D non-LTE framework, employing independent oscillator-strength data and state-of-the-art 3D radiative-transfer post-processing. By combining disc-centre and disc-integrated observations with a detailed comprehensive and a reduced model atom, the study quantifies 3D and non-LTE abundance corrections, concluding with a preferred $A(\mathrm{S})=7.06\pm0.04$. The results challenge earlier higher abundances (e.g., $A(\mathrm{S})\approx7.12$–$7.16$) and support a systematic difference between the solar photosphere and CI chondrites correlated with $50\%$ condensation temperature. The findings underscore the critical role of precise oscillator strengths and 3D non-LTE modelling, and call for independent cross-checks with 3D magnetohydrodynamic photosphere models to solidify implications for solar-system chemical evolution.

Abstract

The solar chemical composition is a fundamental yardstick in astrophysics and the topic of heated debate in recent literature. We re-evaluate the abundance of sulphur in the photosphere by studying seven S I lines in the solar disc-centre intensity spectrum. Our analysis considers independent sets of experimental and theoretical oscillator strengths together with, for the first time, three-dimensional non-local thermodynamic equilibrium (3D non-LTE) S I spectrum synthesis. Our best estimate is $A(\mathrm{S})=7.06\pm0.04$, which is $0.06$ dex to $0.10$ dex lower than that in commonly-used compilations of the solar chemical composition. Our lower solar sulphur abundance deviates from that in CI chondrites, and thereby supports the case for a systematic difference between the composition of the solar photosphere and of CI chondrites that is correlated with $50\%$ condensation temperature. We suggest that precise laboratory measurements of S I oscillator strengths and abundance analyses using 3D magnetohydrodynamic models of the solar photosphere be conducted to further substantiate our conclusions.

Figures (5)

• Figure 1: Gas temperature distribution with vertical logarithmic optical depth in the 3D model atmosphere (blue). The $T-\log\tau$ relation of the corresponding 1D model atmosphere is also shown (orange). The full widths at half maximum of the line-profile integrated 1D non-LTE contribution functions to the line depressions 2015MNRAS.452.1612A are shown for the diagnostic SI lines (the line-averaged result is shown for the SI $1045\,\mathrm{nm}$ triplet). These are presented for both the disc-centre intensity (black) and disc-integrated flux (red) with circles indicating the positions of the peaks.
• Figure 2: Grotrian diagram for the comprehensive (left) and reduced (right) model atoms used in this work. Terms for which fine-structure are unresolved are shown as short blue horizontal lines. Super levels in the reduced model atom are shown as long blue horizontal lines. Transitions shown as slanted grey lines; those used as abundance diagnostics are shown as coloured slanted lines in the right panel and labelled in the legend.
• Figure 3: Abundances inferred from different sulphur lines as a function of logarithmic reduced equivalent width. Upper row shows results from the disc-centre intensity, and lower row shows results from the disc-integrated flux. Left panel shows results for different sets of oscillator strengths based on the 3D non-LTE model. Right panel shows results for different models based on GRASP oscillator strengths. Black squares are the same in the left and right panels of a given row. Error bars reflect the estimated $\pm5\%$ uncertainty on the measured equivalent widths. Weighted linear regressions to different line groups shown, with the SI $675\,\mathrm{nm}$ triplet (not present in the CIV3 set; open symbols) given zero weight in the fits. The shaded rectangle shows the solar sulphur abundance given in 2021AA...653A.141A of $A(\mathrm{S})=7.12\pm0.03$.
• Figure 4: Disc-integrated flux to disc-centre intensity ratios for the SI $1045\,\mathrm{nm}$ triplet. Observations from the Hamburg atlas. The thickness of the plotted lines corresponds to the $\pm1\sigma$ uncertainty in the 3D non-LTE and 3D LTE abundances illustrated in Figure \ref{['fig:abundances']}, that results from a $\pm5\%$ uncertainty in the measured equivalent widths. Instrumental broadening of $R=3.5\times10^{5}$ included for the synthetic intensities and fluxes, and rotational broadening of $\mathrm{V_{\mathrm{rot}}\sin{\iota}}=2\,\mathrm{km\,s^{-1}}$ included for the synthetic flux.
• Figure 5: Photospheric versus CI chondrite abundance differences as a function of $50\%$ condensiation temperature from 2019AmMin.104..844W. Photospheric data AAG21, CL11, and MB22 are from 2021AA...653A.141A, 2011SoPh..268..255C, and 2022AA...661A.140M respectively, the latter restricted to their derivations based on a horizontally- and temporally-averaged 3D model. CI chondrite data is from 2021SSRv..217...44L, converted to the solar scale with $A(\mathrm{Si})=7.51$ from 2021AA...653A.141A. Only elements with combined uncertainties less severe than $\pm0.1\,\mathrm{dex}$ are included, and the reference element silicon is omitted from those data sets where it was available. Sulphur, with $T_{\mathrm{cond}}=672\,\mathrm{K}$, has been circled, and the new solar sulphur abundance found in this work is used to update AAG21, forming the data set AAG21/new S. Weighted linear regressions are overplotted.