Cosmic Star Formation, Reionization, and Constraints on Global Chemical Evolution

Abstract

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.

Gravitational waves from the first stars

“…However, as discussed in Section 3, Model 0 alone could not provide a sufficient flux of ionizing photons for reionization at high redshift. As was mentioned in Section 2, the shape of the spectrum of gravitational waves depends on the parameters a and b in Equation 6. While any variation in these parameters can have a dramatic effect on the resultant spectrum when the anisotropy parameter and neutrino luminosity are specified, the effects are strongly suppressed by our normalization scheme in which we specify the efficiency of gravitational wave production, ǫ, for stars that collapse to black holes.…”

“…[7], as will be discussed presently. Both the normal and massive components can contribute to the chemical enrichment of galaxy forming structures and the IGM, though the normal mode is not sufficient for accounting for the early reionization of the IGM [6]. Here, we restrict our attention to the best fit hierarchical model in [7] in which the minimum mass for star formation is 10 7 M ⊙ .…”

Hierarchical Growth and Cosmic Star Formation: Enrichment, Outflows, and Supernova Rates

“…The type II supernova rate is also well fitted in the same range of redshifts in our models and it is directly correlated to the cosmic SFR. We incorporate an improved treatment of structure formation compared to our previous work (Daigne et al 2004) that leads to new insights into the initial mass function (IMF) of the population III stars at high redshift. We compare three possible mass ranges: 40-100 M ⊙ (normal supernovae), 140-260 M ⊙ (pair-instability supernovae) and 270-500 M ⊙ .…”

“…Thus, as a consequence of the SN Ia time delay, we predict that while ∼ 50 % of iron in structures is produced by type Ia supernovae, this fraction is only ∼ 10 % in the IGM. Daigne et al (2004) that Model 0 alone is not capable of explaining the observed abundance patterns in the extremely iron-poor stars, CS 22949-037 (Depagne et al 2002;Israelian et al 2004), HE 0107-5240 Bessell et al 2004), HE 1327-2326 (Frebel et al (2005)) and G 77-61 (Plez & Cohen (2005)). …”

“…However, there is an active debate as to the exact nature of the massive mode. As in Daigne et al (2004), we consider three possibilities for the massive mode. 1) stars with masses in the range 40 -100 M ⊙ , Model 1; these stars terminate as type II supernovae.…”