Where Can We Really Find the First Stars’ Remnants Today?

Abstract

A number of recent numerical investigations concluded that the remnants of rare structures formed at very high redshift, such as the very first stars and bright redshift z % 6 QSOs, are preferentially located at the center of the most massive galaxy clusters at redshift z ¼ 0. In this paper we readdress this question using a combination of cosmological simulations of structure formation and extended Press-Schechter formalism, and we show that the typical remnants of Population III stars are instead more likely to be found in a group environment, that is, in dark matter halos of mass P2 ; 10 13 h À1 M . Similarly, the descendants of the brightest z % 6 QSOs are expected to be in medium-sized clusters (mass of a few 10 14 h À1 M ), rather than in the most massive superclusters (M > 10 15 h À1 M ) found within the typical 1 Gpc 3 cosmic volume where a bright z % 6 QSO lives. The origin of past claims that the most massive clusters preferentially host these remnants is rooted in the numerical method used to initialize their numerical simulations: Only a small region of the cosmological volume of interest was simulated with sufficient resolution to identify low-mass halos at early times, and this region was chosen to host the most massive halo in the cosmological volume at late times. The conclusion that the earliest structures formed in the entire cosmological volume evolve into the most massive halo at late times was thus arrived at by construction. We demonstrate that, to the contrary, the first structures to form in a cosmological region evolve into relatively typical objects at later times. We propose alternative numerical methods for simulating the earliest structures in cosmological volumes.

ΛCDM predictions for galaxy protoclusters - I. The relation between galaxies, protoclusters and quasars at

“…In the top row of Fig. 20 we show the mass of the most massive progenitors at z = 5.7 (left‐hand panels), z = 6.2 (middle panels) ad z = 6.7 (right‐hand panels) of haloes selected at z = 0 (see also Trenti et al 2008). The dotted line indicates the threshold corresponding to massive galaxy clusters at z = 0.…”

“…Although the progenitor mass increases systematically with increasing local mass, the dispersion in the mass of the most massive z ∼ 6 progenitors is about or over one order of magnitude, and this is true even for the most massive clusters. As explained in detail by Trenti et al (2008) this observation leads to an interesting complication when using the refinement technique often used to simulate the most massive regions in the early Universe by resimulating at high redshift the most massive region identified at z = 0 in a coarse grid simulation. In the bottom panels of Fig.…”

“… The correspondence between the most massive haloes selected at z = 0 and 6 (see also Trenti et al 2008). In the top row of panels we plot the mass of the most massive progrenitor of haloes selected at z = 0 for snapshots at z = 5.7 (left‐hand panels), z = 6.2 (middle panels) and z = 6.7 (right‐hand panels).…”

“…Because the descendants of the most massive haloes at z ∼ 6 may evolve into haloes of >10 15 M ⊙ at z = 0 in a Lambda cold dark matter (ΛCDM) model (e.g. Springel et al 2005; Suwa, Habe & Yoshikawa 2006; Li et al 2007, but see De Lucia & Blaizot 2007; Trenti, Santos & Stiavelli 2008 and of this paper), it is believed that the QSOs trace highly biased regions that may give birth to the most massive present‐day galaxy clusters. If this is true, the small‐scale environment of z ∼ 6 QSOs may be expected to show a significant enhancement in the number of small, faint galaxies.…”

The journey of QSO haloes from z ∼ 6 to the present

“…Previous studies in this area have relied on analytic models calibrated using simulations of less extreme objects (e.g. Trenti, Santos & Stiavelli ) or resimulation techniques applied to a limited number of systems (Li et al ; Sijacki, Springel & Haehnelt ; Romano‐Diaz et al ). In this paper, we present an extremely large N ‐body calculation, which meets the computational challenges directly by simulating a very large volume at adequate resolution within the ΛCDM paradigm.…”