Figure 4.11. Illustrations of SED fits to z = 5.7 LAEs from Lai et al. (2007). Left: x2 contour plots showing the best solution for one object and degeneracies in the fitting parameter. Right: Comparison of best-fit SEDs with constant SFR with observations for three LAEs. These results show indications for the presence of dust in z = 5.7 LAEs. See the text for discussion.
Hu et al. (2002) has recently been analysed by Schaerer & Pello (2005), who find that a non-negligible extinction (AV ~ 1.) may be necessary to reconcile the relatively red UV-rest-frame SED and the presence of Lya. Later this interpretation was supported by Chary et al. (2005), including by longer-wavelength photometry obtained with the Spitzer telescope. Three NB-selected LAEs at z = 5.7 detected in the optical and with the Spitzer telescope at 3.6 and 4.5 |im have recently been analysed by Lai et al. (2007). Overall they find SED fits degenerate in age, extinction, metallicity and SF history with stellar population ages up to 700 Myr. Most solutions require some dust extinction (see Figure 4.11). If the need for Lya emission, i.e. for the presence of young (massive) stars, is taken into account, it seems that a constant SFR scenario is likely, together with an extinction of EB—v ~ 0.1-0.2.
Although the evidence is not yet conclusive, these four z ~ 5.7-6.6 galaxies provide what are currently, to the best of my knowledge, the best indications for dust in "normal" galaxies about 1 Gyr after the Big Bang.t As already mentioned, these objects are probably not representative of the typical high-z LAEs, but they may be of particular interest for direct searches for high-z dust. In any case, the first attempts undertaken so far to detect dust emission from z ~ 6.5 galaxies in the sub-millimetre range (Webb et al. 2007; Boone et al. 2007) have provided upper limits on their dust masses of the order of ~(2 — 6) x 108M0. Future observations with more sensitive instruments and targeting gravitationally lensed objects should soon allow progress in this field.
t Dust emission has been observed in quasars out to z ~ 6, as discussed briefly in Section 4.2.6.
4.4.3 Lyman-break galaxies In general Lyman-break galaxies (LBGs) are better known than the galaxies selected by Lya emission (LAEs) discussed above. There is a vast literature on LBGs, summarized in an annual review paper in 2002 by Giavalisco (2002). However, progress being so fast in this area, frequent "updates" are necessary. In this last part, I shall give an overview of the current knowledge about LBGs at z > 6, trying to present the main methods, results, uncertainties and controversies, and finally to summarize the main remaining questions. A more-general overview about galaxies across the Universe and out to the highest redshift is given in the lectures of Ellis (2007). Recent results from deep surveys including LBGs and LAEs can be found in the proceedings from "At the Edge of the Universe" (Afonso et al. 2007). Chapter 2 of this volume also covers in depth galaxy surveys.
The general principle of the LBG selection has already been mentioned above. The numbers of galaxies identified so far are approximately as follows: 4000 z ~ 4 galaxies (B-dropout), 1000 z ~ 5 galaxies (U-dropout) and 500 z ~ 6 galaxies (¿-dropout) according to the largest dataset compiled by Bouwens and collaborators (cf. Bouwens & Illingworth 2006). The number of good candidates at z > 7 is still small (see below).
Typically two different selections are applied to find z ~ 6 objects: (1) a simple (i — z)ab > 1.3-1.5 criterion establishing a spectral break plus optical non-detection and (2) (i — z)ab > 1.3 plus a blue UV (rest-frame) slope to select actively star-forming galaxies at these redshifts. The main samples have been found thanks to deep HST imaging (e.g. in the Hubble Ultra-Deep Field and the GOODS survey) and with SUBARU (Stanway et al. 2003, 2004; Bunker et al. 2004; Bouwens et al. 2003; Yan et al. 2006).
In general all photometric selections must avoid possible "contamination" by other sources. For the i-dropouts possible contaminants are L- or T-dwarfs, z ~ 1-3 extremely red objects (EROs) and spurious detections in the z band. Deep photometry in several bands (ideally as many as possible!) is required in order to minimize the contamination. The estimated contamination of i-dropout samples constructed using criterion (1) is somewhat controversial and could reach up to 25% in GOODS data, according to Bouwens et al. (2006) and Yan et al. (2006). Follow-up spectroscopy has shown quite clearly that L-dwarfs contaminate the bright end of the i-dropout samples, whereas at fainter magnitudes most objects appear to be truly at high-z (Stanway et al. 2004; Malhotra et al. 2005).
The luminosity function (LF) of z ~ 6 LBGs has been measured and its redshift evolution studied by several groups. Most groups find an unchanged faint-end slope of a ~ —1.7 from z ~ 3 to 6. Bouwens et al. (2006) find a turnover at the bright end of the LF, which they interpret as being due to hierarchical build-up of galaxies. However, the results on M* and a remain controversial. For example Sawicki & Thompson (2006) find no change of the bright end of the LF but an evolution of its faint end on going from z ~ 4 to z ~ 2, while other groups (e.g. Bunker et al. 2004, Yoshida et al. 2006, Shimasaku et al. 2006) find similar results to those of Bouwens et al. The origin of these discrepancies remains to be clarified.
The luminosity density of LBGs and the corresponding SFRD have been determined by many groups up to redshift z ~ 6. Most of the time this is done by integration of the LF down to a certain reference depth, e.g. 0.3L*(z = 3), and at high z generally no extinction corrections are applied. Towards high z, the SFRD is found to decrease somewhat on
Was this article helpful?