Annu. Rev. Astron. Astrophys. 2005. 43: 861-918 Copyright © 2005 by Annual Reviews. All rights reserved

7. GALAXY IDENTIFICATIONS

In this section we describe the results of searches for galaxies physically associated with damped Ly systems.

7.1. Galaxies with z 1.6

Because of the presence of bright, B ~ 18.5, background QSOs, surveys for galaxies associated with high-z damped Ly systems are more challenging than surveys for randomly selected galaxies. To illustrate this point consider a sightline passing through an L_* galaxy ⁶ at a reasonable impact parameter of 10 kpc. At z = 3 the galaxy will have an AB magnitude of 24.7 and an impact parameter of 1.3 arcsec. Detection of the galaxy against the QSO PSF with ground-based telescopes would be exceedingly difficult even under the best seeing conditions and has even proven difficult with broadband-space imaging. Using the NICMOS IR camera on the HST, Colbert & Malkan (2002) surveyed 22 damped Ly systems and detected only one candidate counterpart down to H_AB = 23.5, implying that most damped Ly systems are not drawn from the luminous end of the Lyman Break Galaxy luminosity function. Warren et al. (2001) probed even deeper, to H_AB = 25, and found 41 candidate counterparts near 18 high-redshift damped Ly systems. Broadband imaging to the necessary depth is thus limited by source confusion within reasonable impact parameters.

The most widely used techniques are searches for Ly emission lines at the absorption redshift. The advantage of this method is that the wavelength of Ly emission is located at the bottom of the damped Ly absorption trough, which blocks the bright light of the background QSO. As a result, background night sky emission is the only source of external noise. Using slit spectroscopy, Foltz, Chaffee & Weymann (1986) failed to detect Ly emission with a 3 upper limit of F ~ 10^-16 ergs cm^-2 s^-1 for an unresolved object. While Hunstead, Fletcher & Pettini (1990) claimed detection of Ly emission from a compact source coinciding with DLA0836 + 11 at z = 2.466, this feature was not confirmed in spectra acquired by Wolfe et al. (1992) and Lowenthal et al. (1995). Nor was Ly emission detected in imaging surveys using narrow-band interference filters or Fabry-Perot interferometers. In this case the QSO light is blocked because the bandwidth of the filter is centered on the damped Ly line but has a narrower FWHM. This technique is ideal for impact parameters large compared to the seeing radius, since an emitter located outside a slit could still be detected in the narrow-band image. Smith et al. (1989), Deharveng, Bowyer & Buat (1990) and Wolfe et al. (1992) carried out narrow-band surveys for Ly emission. Deharveng, Bowyer & Buat (1990), Wolfe et al. (1992), and Lowenthal et al. (1995) carried out Fabry-Perot surveys. No detections to limiting fluxes of F 5 × 10^-17 ergs cm^-2 s^-1 for unresolved objects were reported. The extended Ly emitter associated with DLA0836 + 11 at z = 2.466 claimed by Wolfe et al. (1992) is more likely to be a galaxy associated with a lower redshift Mg II absorption system (Lowenthal et al. 1995).

The failure to detect Ly emission could result from the destruction of resonantly trapped photons by even a small amount of dust (Charlot & Fall 1991). For this reason several groups attempted to detect damped Ly systems in H emission since H photons are not resonantly trapped. Bunker et al. (1999) used IR detectors on a 4 m class ground-based telescope to search for H in five damped Ly systems, but none was detected (see also Mannucci et al. 1998). Of course, the failure to detect H emission might also be due to small impact parameters, since an emitting region located within the PSF of the QSO would not be detected from the ground. However, Kulkarni et al. (2000, 2001) used NICMOS to search for H emission from two damped Ly systems and none was found (see Table 1).

TABLE 1. Measurements of emission from damped Ly systems

DLA	zDLA a	zLyb	b c	bd	F(Line)e	SFR	SFR	Refg
			arcsec	kpc	10-17(cgs)	Diagnosticf	M yr-1
0458-02	2.0395	2.0396	0.3 ± 0.3	2.5 ± 2.5	5.40.20.8	Ly	>1.5	1
0953+47A	3.407	3.415	<0.5	<3.7	0.7 ± 0.2	Ly, C	0.8 7.0	2
2206-19A	1.9205	1.9229	1	8.4	26 ± 3.0	C	26 50	3
1210+17	1.8918	......	0.25	2.1	< 2.5	H	<5.0	4
1244-34	1.8590	......	0.16-0.24	1.4-2.0	<0.8	H	<1.6	5
8 DLAs	2.095-2.615	......	1.5	10.0	< 9.0	H	<30	6
aAbsorption redshift of DLA b Ly emission redshift c Displacement angle of emitter (or candidate) from QSO d Displacement distance of emitter (or candidate) from QSO e Line flux in units of ergs cm-2 s-1. First 3 entries are Ly and second 3 are H f Diagnostic for obtaining SFR: Ly is Ly emission, C is FUV continuum, and H is from H emission. g Reference:(1) Møller et al.(2004); (2) A. Bunker (2004), priv. comm; (3) Møller et al.(2002); (4) Kulkarni et al. (2001); (5) Kulkarni et al. (2000); (6) Bunker et al. (1999)

With the recent detections of Ly emission from at least 2 out of 18 damped Ly systems surveyed using 8- to 10-m class telescopes (Møller et al. 2002; Møller, Fynbo & Fall 2004), it is evident that the previous null detections of Ly were largely due to the lower sensitivity of 4-m class telescopes. The results, summarized in Table 1, indicate that two of the three known Ly emitters, DLA0458 - 02 and DLA0953 + 47A, would not have been detected in the earlier surveys. While the sample is too small to draw general conclusions, the results are interesting for the following reasons. First, damped Ly systems resemble randomly selected Ly emitters by the similarity in Ly luminosity and in the compact size of the emission regions. Second, the small impact parameters for DLA0458 - 02 and DLA0953 + 47A suggest that the H I absorbing layers are smaller than ~ 5 kpc. However, since the H I content of DLA0458 - 02 is known to extend over linear scales exceeding 17 kpc (Briggs et al. 1989), the sightline to the QSO must pass close to a compact star forming region, which is embedded in a much larger layer of H I. Third, no continuum emission has been detected from the same two damped Ly systems. The B > 27 limit on DLA0953 + 47A places this damped Ly system near the faint end of the known luminosity function of Lyman Break Galaxies.

By contrast, DLA2206 - 19A is a luminous Lyman Break Galaxy. This emitter was first located in IR images obtained with NICMOS (Warren et al. 2001). The STIS image (Møller et al. 2002) shows rest-frame FUV stellar emission extending 1 arcsec between a bright knot and the QSO sightline. Ly emission at the redshift of the damped Ly system was detected from the knot with spectra obtained with the VLT (Møller et al. 2002). The magnitude integrated over the object is V = 23 (P. Møller 2004, priv. comm.), which places this damped Ly system at the bright end of the Lyman Break luminosity function.

As a result, there is little overlap between the luminosity functions of damped Ly systems and the R < 25.5 spectroscopic sample of Lyman Break Galaxies (see Møller et al. 2002, Schaye 2001). Efforts to detect the clustering of damped Ly systems with neighboring Lyman Break Galaxies have so far only yielded upper limits (Adelberger et al. 2003, Gawiser et al. 2001), providing further evidence that the damped Ly systems are nearly disjoint from the R < 25.5 spectroscopic sample of Lyman Break Galaxies, with DLA2206 - 19A a clear exception. Based on the UV continua implied by the Ly luminosities of the other two damped Ly systems detected in emission, there is growing evidence that at least some damped Ly systems overlap with the dimmer "photometric" sample of Lyman Break Galaxies at R < 27 seen in the Hubble deep fields.

7.2. Galaxies with z < 1.6

While the nature of the galaxies associated with damped Ly systems at z > 1.6 is still unclear, at lower redshifts there should be a close resemblance to objects drawn from the population of normal galaxies if damped Ly systems trace the star-formation history of normal galaxies. However, the low metallicities inferred for most low-z damped Ly systems (Kulkarni et al. 2004, Pettini & Steidel 1999) has cast doubt on this idea and has led to the suggestion that low-z damped Ly systems are metal-poor objects such as dwarf galaxies (Calura, Matteucci & Vladilo 2003) or low surface-brightness galaxies (Jimenez, Bowen & Matteuci 1999). As a result, identification of galaxies associated with damped Ly systems at z < 1 is of vital importance.

There are 23 damped Ly systems known at z < 1.6 (as of October 2004). Thirteen of these have been identified using spectroscopic or photometric redshift techniques (Chen, Kennicutt & Rauch 2004; Chen & Lanzetta 2003) and possible host galaxies for additional eight systems have been found in imaging surveys (Le Brun et al. 1997, Rao et al. 2003). Chen & Lanzetta (2003) analyzed an unbiased subset of nine galaxies with redshifts and concluded that the resultant luminosity distribution was dominated by luminous galaxies with L / L_* > 0.1, i.e., luminous galaxies dominate the neutral gas cross-section at z < 1. More specifically, they used a maximum-likelihood technique to determine the dependence of H I cross-section on luminosity and showed that a cross-section weighted Schechter function with typical parameters for normal galaxies provided a good fit to the data.

On the other hand, Rao et al. (2003) analyzed a heterogeneous sample of 14 damped Ly system host galaxy candidates, including objects without confirmed redshifts and concluded that the neutral gas cross-section at z < 1 was dominated by dwarf galaxies. Since Rao et al. (2003) did not quantify their result, this conclusion is difficult to evaluate. However, comparison with the sample of Chen & Lanzetta (2003) shows a larger fraction of galaxies with lower values of L/L_* in the Rao et al. (2003) sample. In some cases these are galaxies without confirmed redshifts at small angular separations from the damped Ly sightline. In other instances, where two or more galaxies are found to be associated with the damped Ly system, Rao et al. (2003) chose the lowest luminosity galaxy for the analysis because it had the smallest impact parameter. However, Chen & Lanzetta (2003) argue that in these cases, an entire group of galaxies is responsible for the absorption profile, so the object with lowest impact parameter may not be the appropriate choice. Given the uncertainties arising from the small number of objects, we conclude that the current data are consistent with the galaxies associated with low-redshift damped Ly systems being drawn from the population of normal galaxies.

⁶ As defined in the z = 3 Lyman Break Galaxy luminosity function of Steidel et al. (1999). Back.