2.3. The Metallicity of DLAs
Even before the advent of 8-10m class telescopes, it was realised that the metal and dust content of DLAs could be investigated effectively by targeting a pair of (fortuitously) closely spaced multiplets, Zn II 2025, 2062 and Cr II 2056, 2062, 2066 (Meyer, Welty, & York 1989; Pettini, Boksenberg, & Hunstead 1990). The key points here are that while Zn is essentially undepleted in local diffuse interstellar clouds, Cr is mostly locked up in dust grains (Savage & Sembach 1996). Consequently, the ratio N(Zn II) / N(H I) observed in DLAs, when compared with the solar abundance of Zn, yields a direct measure of the degree of metal enrichment (in H I regions Zn is predominantly singly ionised, and the ratios Zn I / Zn II and Zn III / Zn II are both << 1). On the other hand, a deficit - if one is found - of N(Cr II) / N(Zn II) compared to the solar relative abundances of these two elements would measure the extent to which refractory elements have condensed into solid form in the interstellar media traced by DLAs and, by inference, be an indication of the presence of dust in these early galaxies (Cr is also singly ionised in H I gas).
From the point of view of chemical evolution, both Zn and Cr closely trace Fe in Galactic stars with metallicities [Fe/H] between 0 (i.e. solar) and -2 (1/100 of solar; Sneden, Gratton, & Crocker 1991, McWilliam et al. 1995). Additional advantages are the convenient rest wavelengths of the Zn II and Cr II transitions, which at z = 2 - 3 (where DLAs are most numerous in current samples) are redshifted into a easily observed portion of the optical spectrum, and the inherently weak nature of these lines which ensures that they are nearly always on the linear part of the curve of growth, where column densities can be derived with confidence from the measured equivalent widths (e.g. Bechtold 2002).
All in all, a small portion of the red spectrum of a QSO with a damped Ly system has the potential of providing some important chemical clues on the nature of these absorbers and the evolutionary status of the population of galaxies they trace. One of the first detections of Zn and Cr in a DLA is reproduced in Figure 6. However, data of sufficiently high signal-to-noise ratio to detect the weak absorption lines of interest (or to place interesting upper limits on their equivalent widths) typically require nearly one night of observation per QSO with a 4m class telescope. Consequently, it was necessary to wait until the mid-1990s before a sufficiently large sample of Zn and Cr measurements in DLAs could be assembled.
Figure 6. Portion of the Palomar spectrum of the bright QSO PHL 957 recorded by Pettini et al. (1990), encompassing the region of the Zn II and Cr II absorption lines in the zabs = 2.3091 DLA. The vertical tick marks indicate the positions of the lines as follows. Line 1: Zn II 2025.483; line 2: Cr II 2055.596; line 3: Cr II 2061.575 + Zn II 2062.005 (blended); and line 4: Cr II 2065.501. The spectrum has been normalised to the underlying QSO continuum and is shown on an expanded vertical scale.
Figure 7 shows the current data set from the survey by our group (e.g. Pettini et al. 1997b; 1999). There are several interesting conclusions which can be drawn from these results.
Figure 7. Plot of the abundance of Zn against redshift for the full sample of 41 DLAs from the surveys by Pettini and collaborators. Abundances are measured on a log scale relative to the solar value shown by the broken line at [Zn/H] = 0.0; thus a point at [Zn/H] = -1.0 corresponds to a metallicity of 1/10 of solar. Upper limits, corresponding to non-detections of the Zn II lines, are indicated by downward-pointing arrows. Upward-pointing arrows denote lower limits in two cases where the Zn II lines are sufficiently strong that saturation may be important.
(a) Damped Ly systems are generally metal poor at all redshifts sampled. Evidently DLAs arise in galaxies at early stages of chemical evolution, since nearly all of the points in Figure 7 lie below the line of solar abundance. The statistic of relevance here is the the column density-weighted mean abundance of Zn
and the summations in eq. (2.5) are over the n DLA systems in a given sample. In this way, by counting all the Zn atoms per unit cross-section (cm-2) and dividing by the total column density of neutral hydrogen we find that DLAs have a typical metallicity of only ~ 1/13 of solar ([<Zn / HDLA>] = - 1.13).
(b) There appears to be a large range in the values of metallicity reached by different galaxies at the same redshift. Values of [Zn/H] in Figure 7 span nearly two orders of magnitude, pointing to a protracted `epoch of galaxy formation' and to the fact that chemical enrichment probably proceeded at different rates in different DLAs. The wide dispersion in metallicity goes hand in hand with the diverse morphology of the DLA galaxies which have been imaged at z < 1, as discussed earlier (Section 2.1).
When the metallicity distribution of damped Ly systems is compared with those of different stellar populations of the Milky Way, we find that is broader and peaks at lower metallicities than those of either thin or thick disk stars (Figure 8). At the time when our Galaxy's metal enrichment was at levels typical of DLAs, its kinematics were closer to those of the halo and bulge than a rotationally supported disk. This finding is at odds with the proposal that most DLAs are large disks with rotation velocities in excess of 200 km s-1, put forward by Prochaska & Wolfe (1998).
(c) There is little evidence from the data in Figure 7 for any redshift evolution in the metallicity of DLAs. This question has also been addressed using the abundance of Fe which, thanks to its rich absorption spectrum, can be followed to higher redshifts and lower metallicities than Zn (Prochaska, Gawiser, & Wolfe 2001). Once allowance is made for the fraction of Fe in dust grains, Vladilo (2002b) finds a gradient of -0.32 in the linear regression of [Fe/H] vs. zabs. However, while Figure 7 suggests that the chances of finding a DLA with [Zn/H] > - 1.0 are possibly greater at z < 1, the column density-weighted metallicities - which measure the density of metals per comoving volume - are consistent with no evolution over the range of redshifts probed so far, irrespectively of whether Zn or Fe are considered (see Figure 9). Evidently, the census of metals at all redshifts is dominated by high column density systems of low metallicity.
Figure 9. Column density-weighted metallicities of DLAs in different redshift intervals, from the surveys by Pettini et al. (1999) for Zn, and Prochaska & Wolfe (2002) for Fe. The lower abundance of Fe relative to Zn probably reflects the presence of moderate amounts of dust in most DLAs (Vladilo 2002b).
The lack of evolution in both the neutral gas and metal content of DLAs was unexpected and calls into question the notion that these absorbers are unbiased tracers of these quantities on a global scale. On the other hand, the paucity of data at redshifts z < 1, that is over a time interval of more than half of the age of the universe (Table 1), makes it difficult to draw firm conclusions and it may yet be possible to reconcile existing measurements with models of cosmic chemical evolution (Pei, Fall, & Hauser 1999; Kulkarni & Fall 2002).