Sources for Metals in the Intergalactic Medium

Abstract

We present a discussion of possible sources of C, O, and Si in the intergalactic medium (IGM) using the yields of very massive stars (VMSs) and Type II supernovae (SNe II). The chemical evolution of the IGM is considered on the basis of analytical phenomenological models of hierarchical structure formation and the required metal production rates. Two regimes are considered: one for gas expulsion by VMSs and SNe II in low-mass halos prior to dissociation of H_2 molecules and reionization, and the other for later SN II-driven outflows from intermediate-mass halos. We use recent data on the abundances of C, O, and Si in the IGM inferred from two UV background (UVB) models. We show that the values of [C/O]_(IGM) and [Si/C]_(IGM) inferred from the UVB model using quasars only cannot be explained by any existing stellar models. To account for the elemental ratios in the IGM by stellar models requires that a softer UVB model be used in calculating the elemental abundances from the data on ionic species. This then permits a solution using the yields of VMSs and SNe II. The results, in particular [Si/C]_(IGM), show that VMSs must provide ≳15% of the C in the IGM. The preferred scenario is that VMSs in low-mass halos provided between 15% and 60% of the C in the regime of very high redshift (z ≳15) and that galactic outflows provided the remainder during later epochs (4 < z ≾6). Thus, there is a large gap in z between metal production by VMSs very early in the chemical evolution of the IGM and subsequent contributions from galactic outflows. It is not possible to explain the observations by just SN II sources associated with galactic outflows. The observational estimate of [O/H]_(IGM) implies a high [Fe/H]_(IGM) ≥ -3 for all cases considered (including the case of a pure VMS source). This raises problems with regard to observations of metal-poor stars in the Galaxy. In addition, the [O/Fe] values for SN II models are in conflict with stellar observations, which indicate that the calculated average Fe yields of SNe II are a factor of ~2 too high. These issues and the general problem of relating abundances in the IGM to those in metal-poor stars remain to be investigated.