A Model for Abundances in Metal-poor Stars

A model is presented that seeks to explain quantitatively the stellar abundances of r-process elements and other elements associated with the r-process sites. It is argued that the abundances of all these elements in stars with -3 ≲ [Fe/H] < -1 can be explained by the contributions of three sources. The sources are the first generations of very massive (≳100 M_☉) stars that are formed from big bang debris and are distinct from Type II supernovae (SNe II) and two types of SNe II, the H and L events, which can occur only at [Fe/H] ≳ -3. The H events are of high frequency and produce dominantly heavy (A > 130) r-elements but no Fe (presumably leaving behind black holes). The L events are of low frequency and produce Fe and dominantly light (A ≲ 130) r-elements (essentially none above Ba). By using the observed abundances in two ultra-metal-poor stars and the solar r-abundances, the initial or prompt inventory of elements produced by the first generations of very massive stars and the yields of H and L events can be determined. The abundances of a large number of elements in a star can then be calculated from the model by using only the observed Eu and Fe abundances. To match the model results and the observational data for stars with -3 < [Fe/H] < -1 requires that the solar r-abundances for Sr, Y, Zr, and Ba must be significantly increased from the standard values. No such changes appear to be required for all other elements. If the changes in the solar r-abundances for Sr, Y, Zr, and Ba are not permitted, the model fails at -3 < [Fe/H] < -1 but still works at [Fe/H] ≈ -3 for these four elements. By using the corrected solar r-abundances for these elements, good agreement is obtained between the model results and data over the range -3 < [Fe/H] < -1. No evidence of s-process contributions is found in this region, but all the observational data in this region now show regular increases of Ba/Eu above the standard solar r-process value. Whether the solar r-components of Sr, Y, Zr, and Ba used here to obtain a fit to the stellar data can be reconciled with those obtained from solar abundances by subtracting the s-components calculated from models is not clear.