Next Contents Previous

8.3 Star Forming Galaxies at High Redshift

It has been conjectured that primeval galaxies would be nearly dust free and devoid of metals at their early stages ( but see Puget and Guiderdoni 1999). The search for ``normal'' galaxies (i.e. where light is generated by stars) at high redshifts has seen a tremendous progress in the last years thanks to programs employing the Lyman break technique (cf. Steidel et al. 1995, 1996, 1999), and projects like the Hubble Deep Field (HDF, Williams et al. 1996). Steidel and collaborators (1996) have shown that on the order of a few percent of the galaxies at faint optical magnitudes (curlyR ~ 25) are actively star forming galaxies at redshift greater than 3. With the assumption that the bulk of emission is produced by photoionisation of stars, their star formation rate is found to range from a few up to a thousand solar masses per year, hence much larger than in local metal-poor, low mass, galaxies such as IZw18. Morphologically many appear to be star-forming spheroids smaller than the bulge of a spiral galaxy. They may be part of the reservoir from which many of today's luminous galaxies were formed through hierarchical merging. Presently, the true mass and the metallicity of these galaxies is poorly known. Only a few have been followed up using near-infrared spectroscopy (Pettini et al. 1998) showing that their dust content (as obtained from rest-frame visible light) does not require a major reassessment for the star formation rates derived from the rest frame UV. Masses obtained from line-widths lead to masses of 1010 Msun, hence larger than that of the local dwarf galaxies we have reviewed throughout this paper, and furthermore these may even be underestimated (Pettini et al. 1998). There are indications that energetic outflows are taking place, similar to what is now observed in local starbursts (Kunth et al. 1998). Broadened interstellar lines are most likely the result of mechanical input from winds and supernova. Metallicity estimates using synthetic stellar-wind profiles calibrated against a local star-forming galaxy sample suggest Magellanic Cloud-like composition (Leitherer 1999). Although many galaxies are found to be extremely dusty even at large redshift (Puget and Guiderdoni 1999), it is possible that there exists an intrinsically fainter population of dust-free star-forming galaxies that could represent the earliest phases of galaxy formation. New Lyalpha emitters are now found at high-redshift from surveys using large telescopes with narrow-band filters (Hu et al. 1998, Pascarelle et al. 1998). Limits down to a few 10-18 erg/cm2/s are now reachable and give access in principle, to galaxies with fainter continuum magnitudes than the Lyman break galaxies (see Fig. 12). On the other hand local starbursts (Kunth et al. 1998) indicate that an unknown fraction of the youngest galaxies may not be Lyalpha emitters. The systematic search and discovery of this kind of objects should offer the opportunity of studying processes of star and galaxy formation and evolution at a substantial cosmological look-back time. Unfortunately even the brightest emission lines objects are difficult to follow up, and only the advent of 8m class telescopes with near-IR spectrographs will allow to tackle the difficult task of measuring metal abundances.

Deep optical galaxy counts have shown a strong excess of faint blue galaxies, which however is absent in near-infrared K-band surveys (Tyson 1988). Redshift surveys find that the bulk of this faint population is at intermediate redshifts (z < 0.6; see, e.g., Broadhurst et al. 1988, Colles et al. 1990, Lilly, Cowie & Gardner 1991). It has been proposed (see also Cowie et al. 1991, Babul and Rees 1992), that the faint blue excess could be explained by a population of star bursting dwarf galaxies at intermediate redshifts. It has been demonstrated that the blue excess can be identified with a population of irregular/peculiar galaxies (e.g. Glazebrook et al. 1995, see Ellis 1997 for a review).

Spectroscopic follow up of compact galaxies in the flanking fields around the Hubble Deep Field (HDF) have shown these to be active star forming galaxies at 0.4 < z < 1.4 with narrow emission line. In many respects these objects are similar to the most luminous local BCGs and H II galaxies (e.g. Guzmán et al. 1997). Their metallicities are largely unknown but do not seem to be very low. Similar high redshift blue compact galaxies were found in the Canada-France redshift survey (Schade et al. 1995, Hammer et al. 1997). Kobulnicky and Zaritsky (1999) studied rather luminous emission line galaxies at redshift 0.1 to 0.5, and found oxygen abundances from 1/5 Zsun to Zsun, and N/O values consistent with local galaxies of similar metallicity.

If, as in the local Universe, low metallicity galaxies at high redshift would be dwarf galaxies of low or moderate luminosity, they would be missing in existing surveys. Indeed, the luminosity function of LBGs at faint magnitudes is very steep, indicating a large population of intrinsically fainter objects lurking below the current detection threshold (Steidel et al. 1999)

Figure 12

Figure 12. The relation between absolute blue magnitude and equivalent with of Hbeta for a sample of H II galaxies (Terlevich et al. 1991). The apparent trend may in part be due to selection effects. (Courtesy R. Terlevich)

Next Contents Previous