On the Origin of Sulfur

We present our work on the halo evolution of sulfur, based on observations of the S I lines around 9220 Å for ten stars for which the S abundance was obtained previously from much weaker S I lines at 8694 Å. We cannot confirm the rise and the high [S/Fe] abundances for low [Fe/H], as claimed in the literature from analysis of the 8694 Å lines. The reasons for claims of an increase in [S/Fe] with decreasing [Fe/H] are probably twofold: uncertainties in the measurements of the weak 8694 Å lines, and systematic errors in metallicity determinations from Fe I lines. The near-infrared sulfur triplet at 9212.9, 9228.1, and 9237.5 Å are preferred for an abundance analysis of sulfur for metal-poor stars. Our work was presented in full by Ryde & Lambert (2004). There has recently been a debate in the literature on the chemical evolution of sulfur in the halo phase of the Milky Way ([Fe/H] ≤ −1). Israelian & Rebolo (2001) and Takada-Hidai et al. (2002) claim a monotonic rise in [S/Fe] for decreasing [Fe/H] based on an analysis using weak sulfur lines at 8694 Å measured in spectra of approximately ten stars each. This rise has consequences for our understanding of the sites of formation of sulfur in the Universe. Is there a rise or not? In order to answer this question, we have reinvestigated ten of the stars examined previously, by analysing the infrared triplet lines of neutral sulfur at 9220 Å instead of the weaker 8694 Å lines used before. The lines are indicated in the Grotrian diagram shown in Figure 1. The lower levels of the 8694 Å lines are the upper levels of the 9220 Å lines and hence the excitation potential difference is 1.34 eV. We observed the stars with the 2dCoudé spectrometer (Tull et al. 1995) at McDonald Observatory in 2001. Our stars are a mix of dwarfs and giants spanning a temperatures range of 4200 − 6300 K, and a metallicity range from [Fe/H]= −0.7 down to [Fe/H]= −2.9. Exposure times ranged from half-an-hour to three hours per star. In our analysis, we use the model atmosphere parameters used by Israelian & Rebolo (2001) and Takada-Hidai et al. (2002) with the exception that the metallicity is re-determined from Fe II lines, which are not much affected by non-LTE effects, instead of the more uncertain determination of the metallicity based on an non-LTE analysis of Fe I lines.