SFR = const
Figure 4.10. Left: Observed Lya equivalent-width distribution of z = 4.5 sources from the LALA survey. From Malhotra & Rhoads (2002). Right: Predicted Lya equivalent width for star-bursts at various metallicities (from Solar to Population III). Normal metallicities (Z > (1/50)Z0) are shown by the pale dashed lines. The maximum value predicted in this case is W(Lya) — 300 A. From Schaerer (2003).
et al. (2006) argue that LAEs are limited to a relatively narrow mass range around M+ — 1O9M0. Further studies will be necessary for a proper understanding of the connections between LBGs and LAEs and the evolution of the two populations with redshift.
4-4-2.1 Population III signatures in LAEs?
The Large Area Lyman-a (LALA) survey by Rhoads and collaborators, carried out on 4-m-class telescopes, has been one of the first to find a significant number of LAEs at high redshift (z = 4.5 and 5.7, and later also objects at z = 6.5). Among the mostinteresting results from LALA is the finding of a large fraction of LAEs with an apparently high median Lya equivalent width, compared with expectations from normal stellar populations (see Figure 4.10). Indeed, half of their z = 4.5 candidates have W(Lya) in excess of - 200-300 A (Malhotra & Rhoads 2002), a value expected only for very young starbursts, populations with extreme IMFs, or very metal-poor (or Population III) stars (cf. Schaerer 2003). Malhotra & Rhoads (2002) suggested that these could be AGNs or objects with peculiar top-heavy IMFs and/or Population III-dominated. In this context, and to explain other observations, Jimenez & Haiman (2006) also advocate a significant fraction of Population III stars, even in z — 3-4 galaxies. Recently Hansen & Oh (2006), reviving an idea of Neufeld (1991), have suggested that the observed W(Lya) could be "boosted" by radiation-transfer effects in a clumpy ISM.
Follow-up observations of the LALA sources have allowed one to exclude the narrowline AGN "option" (Wang et al. 2004), but have failed to provide further explanations of this puzzling behaviour. About 70% of the LALA LAEs have been confirmed spectro-scopically; some high equivalent-width measurements could also be confirmed spectro-scopically.t Deep spectroscopy aimed at detecting other emission lines, including the He ii A1640 line indicative of a Population III contribution (see Section 4.2.3), have been unsuccessful (Dawson et al. 2004), although the achieved depth (Heii A1640/Lya < 13%-20%
t But aperture effects may still lead to an overestimate of W(Lya).
at (2—3)a and W(HeII A1640) < 17-25 A) might not be sufficient. The origin of these high W(Lya) remains thus unknown.
However, there is some doubt about the reality of the LALA high equivalent widths measured from NB and broad-band imaging, or at least regarding their being so numerous even at z = 4.5. First of all the objects with the highest W(Lya) have very large uncertainties since the continuum is faint or not detected. Second, the determination of W(Lya) from a NB and a centred broad-band filter, R-band in the case of Malhotra & Rhoads (2002), may be quite uncertain, due to unknowns in the continuum shape, the presence of a strong spectral break within the broad-band filter etc.; see Hayes & Oestlin
(2006) for a quantification and Shimasaku et al. (2006). Furthermore, other groups have not found such high-W objects (Hu et al. 2004; Ajiki et al. 2003), suggesting also that this may be related to insufficient depth of the LALA photometry.
More recently larger samples of LAEs were obtained, e.g. at z = 5.7; e.g. Shimasaku et al. (2006) has 28 spectroscopically confirmed objects. Although their observed restframe equivalent widths WObSt (Lya) (median value and W distribution) are considerably lower than those of Malhotra & Rhoads at z = 4.5, and only a few objects (1-3 out of 34) have Woref(Lya) > 200 A, it is possible that in several of these objects the maximum Lya equivalent width of normal stellar populations is indeed exceeded. This would clearly be the case if the IGM transmission at this redshift were Ta > 0.3-0.5 (cf. Shimasaku et al. 2006; Dijkstra et al. 2006c), which would imply that the true intrinsic Wrest = 1/Ta x WObf is >2-3 times higher than the observed one. Shimasaku et al. estimate that >30%-40% of their LAEs have Wrest(Lya) > 240 A and suggest that these may be young galaxies or again objects with Population III contribution. Dijkstra & Wyithe
(2007), on the basis of Lya-LF and W (Lya) modelling, also argue for the presence of Population III stars in this z = 5.7 LAE sample.
Another interesting result is the increase of the fraction of large W (Lya) LBGs with redshift, e.g. from >2% of the objects with Wrest(Lya) > 100 Ä at z > 3 to >80% at redshift 6, which is tentatively attributed to lower extinction, younger ages or an IMF change (Shimasaku et al. 2006; Nagao et al. 2007).
Despite these uncertainties it is quite clear that several very strong LAE emitters are found and that these objects are probably the most-promising candidates with which to detect direct in situ signatures of Population III at high redshift (see also Scannapieco et al. 2003). Searches are therefore going on (e.g. Nagao et al. 2005) and the first such discovery may be "just around the corner", or may need more-sensitive spectrographs and multi-object near-IR spectroscopy (see Section 4.4.4).
Although there are indications that LAEs selected through their Lya emission are mostly young and relatively dust-free objects (e.g. Shimasaku et al. 2006; Pirzkal et al. 2006; Gawiser et al. 2006), it is of great interest to search for signatures of dust in distant/primeval galaxies.1 Furthermore, some models predict a fairly rapid production and the presence of significant amounts of dust at high z (Mao et al. 2006). LAEs have the advantage of being at known redshift and of indicating the presence of massive stars. SED fits of such objects must therefore include populations of age <10Myr, providing thus an additional constraint on modelling.
Recently the stellar populations of some high-z LAEs have been analysed with such objectives in mind. For example the z = 6.56 gravitationally lensed LAE discovered by
1 Remember that e.g. sub-millimetre-selected galaxies - i.e. very dusty objects - or at least a subsample of them exhibit also Lya emission (Chapman et al. 2003).
Was this article helpful?