No abstract

We argue that extreme metal-poor stars show a high dispersion in metallicity, because their abundances are the outcome of very few supernova events. Abundance anomalies should appear because of the discrete range of progenitor masses. There is a natural metallicity threshold of Z/Z ⊙ ∼ 10 −4 below which one would expect to find very few, if any, halo stars. Similar reasoning is applied to lower mass systems, such as metal-poor compact blue galaxies and Lyman alpha absorption line clouds seen towards high redshift quasars, where a somewhat higher threshold is inferred.

Motivated by the WMAP results indicating an early epoch of reionization, we consider alternative cosmic star formation models which are capable of reionizing the early intergalactic medium. We develop models which include an early burst of massive stars (with several possible mass ranges) combined with standard star formation. We compute the stellar ionizing flux of photons and we track the nucleosynthetic yields for several elements : D, 4 He, C, N, O, Si, S, Fe, Zn. We compute the subsequent chemical evolution as a function of redshift, both in the intergalactic medium and in the interstellar medium of forming galaxies, starting with the primordial objects which are responsible for the reionization. We apply constraints from the observed abundances in the Lyman α forest and in Damped Lyman α clouds in conjunction with the ability of the models to produce the required degree of reionization. We also consider possible constraints associated with the observations of the two extremely metal-poor stars HE 0107-5240 and CS22949-037. We confirm that an early top-heavy stellar component is required, as a standard star formation model is unable to reionize the early Universe and reproduce the abundances of the very metal-poor halo stars. A bimodal (or topheavy) IMF (40 -100 M ⊙ ) is our preferred scenario compared to the extreme mass range ( 100 M ⊙ ) often assumed to be responsible for the early stages of reionization. A mode of even more extreme stellar masses in the range (≥ 270 M ⊙ ) has also been considered. All massive stars in this mode collapse entirely into black holes, and as a consequence, chemical evolution and reionization are de-correlated. The ionizing flux from these very massive stars can easily reionize the Universe at z ∼ 17. However the chemical evolution in this case is exactly the same as in the standard star formation model, and the observed high redshift abundances are not reproduced. We show that the initial top-heavy mode, which originally was introduced to reionize the early Universe, produces rapid initial metal pollution. The existence of old, C-rich halo stars with high [O/Fe] and [C/Fe] ratios is predicted as a consequence of these massive stars. The recently observed abundances in the oldest halo stars could trace this very specific stellar population. The extreme mass range is disfavored and there is no evidence, nor any need, for a hypothesised primordial population of very massive stars in order to account for the chemical abundances of extremely metal-poor halo stars or of the intergalactic medium.The combined population of early-forming, normal (0.1 -100 M ⊙ ) and massive (40 -100 M ⊙ ) stars can simultaneously explain the cosmic chemical evolution and the observations of extremely metal-poor halo stars and also account for early cosmological reionization.

The primordial abundances of Deuterium, 4 He, and 7 Li are crucial to determination of the baryon density of the Universe in the framework of standard Big Bang nucleosynthesis (BBN). 6 Li which is only produced in tiny quantities and it is generally not considered to be a cosmological probe. However, recent major observational advances have produced an estimate of the 6 Li/ 7 Li ratio in a few very old stars in the galactic halo which impacts the question whether or not the lithium isotopes are depleted in the outer layers of halo stars, through proton induced reactions at the base of (or below) the convective zone. Here, we use i) an empirical relation, independent of any evolutionary model, to set an upper limit on the 6 Li rise compatible with the very existence of the Spite's plateau (i.e. the flat lithium abundance measured in very old stars of the halo of our Galaxy of different iron content) and ii) a well founded evolutionary model of light elements based on spallation production .Indeed, 6 Li is a pure product of spallation through the major production reactions, fast oxygen and alphas interacting on interstellar H, He (especially in the early Galaxy). The rapid nuclei are both synthesized and accelerated by SN II. In this context, the 6 Li evolution should go in step with that of beryllium and boron, recently observed by the Keck and HST telescopes. 6 Li adds a new constraint on the early spallation in the Galaxy. In particular, if confirmed, the 6 Li/ 9 Be ratio observed in two halo stars (HD 84937, BD +26 • 3578 = HD 338529) gives strong boundary conditions on the composition and the spectrum of the rapid particles involved.Both methods converge to show that 6 Li is essentially intact in halo stars, and a fortiori 7 Li, which is more tightly bound. Moreover, extrapolating empirical and theoretical evolutionary curves to the very low metallicities, we can define a range of the 6 Li abundance in the very early Galaxy consistent with Big Bang nucleosynthesis (5.6 10 −14 to 3. 10 −13 ) . Following the evolution at increasing metallicity, we explain the abundance in the solar system within a factor of about 2. The whole evolution from Big Bang to present is reasonably reproduced, which demonstrates the general consistency of the present analysis of 6 Li.The baryonic density derived from both lithium isotopes is between 1.5 to 3.5 per cent of the critical one, in good agreement with the determination based on independent analyses. Consequently, thanks to these new data and theoretical developments, we show that 6 Li can be used to establish stellar 7 Li abundances -3as a valid tracer due to the fact that it allows to reinforce the Spite's plateau as a primordial test of BBN; on the other hand, its early evolution can be used to corroborate the calculated BBN abundances. In the framework of this work, a pregalactic α + α process producing 6 Li is not necessary. Finally, thanks to 6 Li, the physics of spallative production of light elements should be more easily mastered when more data will become available.