While the enrichment of the primordial gas by metals ejected in the first supernovae is likely to preclude primordial star formation in a large fraction of the first galaxies (Johnson et al. 2008; Wise & Abel 2008; Greif et al. 2010; Maio et al. 2010), it is also not likely that metal enrichment abruptly ends the epoch of Pop III star formation after the formation of the first stars. As discussed in Section 2.2, it is possible for the photodissociating background radiation established by early generations of stars to slow the collapse of the primordial gas, potentially delaying a large fraction of Pop III star formation and metal enrichment until later times. Also, as discussed in Section 3.2, the mixing of the first metals with the primordial gas, especially within minihalos, may not occur efficiently. Therefore, it is a distinct possibility that Pop III star formation continues well after the formation of the first stars (e.g. Scannapieco et al. 2003; Tornatore et al. 2007; Trenti et al. 2009; Maio et al. 2010), and that substantial primordial star formation may be detectable in the first galaxies.
It is therefore critical to predict observable signatures of Pop III star formation, in order that it can be identified in high redshift galaxies (e.g. Zackrisson et al. 2011). Some distinctive signatures derive from the high surface temperatures of primordial stars, which arise due to a relatively low opacity in the stellar interior. This low opacity translates into a smaller radii R∗ for primordial stars than for their metal-enriched counterparts. In turn, because stellar luminosity scales as L∗ ∝ R∗2 T∗4, for a given luminosity the surface temperature T∗ of a primordial star will be higher than a metal-enriched star (e.g. Siess et al. 2002; Lawlor et al. 2008). For very massive primordial stars, the surface temperature is very high, roughly ∼ 105 K (Bromm et al. 2001; Schaerer 2002). Owing to this high temperature, primordial stars emit copious high energy radiation, a relatively large fraction of which is able to ionize not only hydrogen (H i), but also helium (He i and He ii).
As a substantial portion of the ionizing photons emitted from stars in early galaxies are absorbed by the relatively dense gas in the interstellar medium before escaping into the IGM (e.g. Wood & Loeb 2000; Gnedin et al. 2008; Wise & Cen 2009; Razoumov & Sommer-Larsen 2010; Paardekooper et al. 2011; Yajima et al. 2011), the energy in these photons is reprocessed into emission lines arising from the recombination of the ionized species (e.g. Osterbrock & Ferland 2006). For the case of primordial stars, because a relatively large fraction of the emitted radiation ionizes He ii, the photons emitted during the recombination of He iii to He ii produce strong emission at characteristic wavelengths. The most prominent recombination line emitted from such He iii regions, with a wavelength of 1640 Å, emerges from the radiative decay of the lone electron in this ion from the n = 3 to the n = 2 state 12. The most prominent emission lines from the recombination of ionized hydrogen in the H ii regions surrounding primordial stars are the same as expected from metal-enriched stars, Lyα and Hα, which arise from the radiative decay from the n = 2 → 1 and n = 3 → 2 energy levels of hydrogen, respectively. The key observational signature of primordial star formation, as opposed to metal-enriched star formation, is a relatively large ratio of the luminosity emitted in the helium line, He ii λ1640, to that emitted in the hydrogen lines (see e.g. Tumlinson et al. 2001; Oh et al. 2001; Schaerer et al. 2003; Raiter et al. 2010).
 |
Figure 18. The luminosity of Pop III star clusters in a first galaxy at z ∼ 12, as a function of the time from their formation, in three recombination lines: Lyα (dot-dashed blue), Hα (solid red), and He II λ1640 (dashed black). The four panels correspond to four different combinations of stellar IMF and total stellar mass; these are, clockwise from top-left: twenty-five 100 M⊙ stars, two hundred fifty 100 M⊙ stars, one thousand 25 M⊙ stars, and one hundred 25 M⊙ stars. The relative luminosities of detected He ii and H i recombination lines can provide information about the stellar metallicity and IMF; more massive and more metal-poor stars emit more high energy radiation that can ionize He ii, which leads to strong He ii λ1640 emission relative to Hα and Lyα. From Johnson et al. (2009). |
Figure 18 shows the luminosity emitted in each of the three recombination lines mentioned above from an instantaneous burst of Pop III star formation in a first galaxy formed in a halo of mass Mh ∼ 108 M⊙ at z ∼ 12, as gleaned from cosmological radiative transfer simulations (Johnson et al. 2009). Each of the panels shows the line luminosities for a different combination of the characteristic stellar mass of the stars (either 25 or 100 M⊙) and of the total stellar mass (either 2,500 or 25,000 M⊙). Even assuming such large characteristic masses for Pop III stars and that such a large fraction (either ∼ 1 or ∼ 10 percent) of the gas in the first galaxies is converted into stars, the luminosities of the recombination lines are likely to be too dim to detect with telescopes in the near future. To see this, we can estimate the total flux F that would be in these lines at z = 0 as
 |
(44) |
where L is the luminosity in a given line and DL is the luminosity distance to redshift z. At z ≥ 10, even the most luminous line, Lyα, would be seen at z = 0 with a flux of of ≤ 4 × 10−20 erg s−1 cm−2, which is well below the flux limit of ∼ 2 × 10−19 erg s−1 cm−2 of surveys planned for the JWST (Gardner et al. 2006; Windhorst et al. 2006).
Instead of the first galaxies, hosted in halos with masses of ∼ 108 M⊙ at z ≥ 10, it thus appears likely that observations in the next decade may reveal somewhat more developed galaxies hosted in more massive halos (e.g. Barkana & Loeb 2000; Ricotti et al. 2008; Johnson et al. 2009; Pawlik et al. 2011), although there is the possibility of detecting less developed galaxies if their flux is magnified by gravitational lensing (see e.g. Zackrisson 2011). As shown in Figure 19, the JWST is predicted to be capable of detecting both He ii λ1640 and Hα from metal-free starbursts in halos with masses ≥ 3 × 108 M⊙, if the IMF is very top-heavy. Though, if the typical mass of Pop III stars is < 50 M⊙, it is likely that He ii λ1640 will only be detectable from significantly more massive stellar clusters, expected to form in similarly more massive halos. However, because more massive halos are formed from the mergers of smaller halos which themselves may have hosted star formation, it may be predominantly metal-enriched Pop II stars that form in the galaxies which will be detected by the JWST (e.g. Johnson et al. 2008). 13
 |
Figure 19. Stellar masses M∗,min of the lowest mass starburst observable through the detection of the Hα line (solid curves), the He II λ1640 line (dashed curves), or the continuum at 1500 Å(dash-dotted curves) with JWST, assuming an exposure time of 106 s and a signal-to-noise ratio of S/N = 10. Stellar masses derived from the Schaerer (2003) zero-metallicity starbursts with a standard Salpeter-like IMF, zero-metallicity starbursts with a top-heavy IMF, and low-metallicity starbursts, are shown in blue, red, and black, respectively. The right axis shows the masses Mmin = 103 M∗,min of halos expected to host a starburst with stellar mass M∗,min. The dotted curve shows the mass of a first galaxy-sized halo, with virial temperature Tvir = 104 K. From Pawlik et al. (2011). |
That said, there is the possibility that substantial Pop III star formation takes place even well after the epoch of the first galaxies (i.e. at z < 10), either due to inefficient mixing of primordial and metal-enriched gas (e.g. Jimenez & Haiman 2006; Pan & Scalo 2007; Wyithe & Cen 2007; Dijkstra & Wyithe 2007; Cen 2010) or to the collapse of primordial gas into late-forming atomic cooling halos (e.g. Tornatore et al. 2007; Trenti et al. 2009; Johnson 2010). Pop III star formation at such late times could be detected more easily, in large part because the emission line flux increases strongly with decreasing redshift, as shown in equation (44). Population III However, at such low redshifts the background ionizing radiation field that builds up during reionization can strongly inhibit the infall of primordial gas into halos, limiting the amount of Pop III star formation that can occur even in metal-free galaxies (e.g. Efstathiou 1992; Gnedin 2000; Tassis et al. 2003; Dijkstra et al. 2004).
12 While photons are also emitted in transitions to the n = 1 state, the IGM is optically thick to these photons before reionization due to absorption by neutral hydrogen, and so they are not expected to be observable from the first galaxies. Back.
13 It is also likely that other, complementary next generation facilities, such as the Atacama Large Millimeter Array (e.g. Combes 2010), will detect only metal-enriched star-forming galaxies. Back.