Absorption line systems

• A potentially powerful (and easier) way of studying galaxies at high redshift is to look at the absorption they produce when they lie between us and QSOs (sketch). This is interesting from several perspectives:
  • identifying truly primordial galaxies, i.e. those which have not yet converted gas into stars,
  • identifying a sample of formed galaxies by detection of their ISM, a selection criterion totally distinct from that, e.g., of apparent magnitude - this allows an independent test of conclusions made from the deep redshift surveys.

• problem with doing galaxy evolution with these without optical identification is difficulty of going from line of sight measurement to a global property of an object, but this is to some degree handled statistically with a large sample of absorbers.

• What is the effect of intervening gas clouds on QSO emission? The presence of neutral hydrogen can create a variety of effects, depending on the column density through which the light passes.
  • At the lowest column densities, light at Ly $\alpha$ is resonantly scattered, creating an absorption feature at 1216 A in the rest wavelength of the absorber. This leads to the Lyman alpha forest. These lines vary in strength depending on the column density, but have been detected for systems with $\Sigma$ as low as $10^{13} cm^{-2}$.
  • At higher column densities, $\Sigma > \sim 10^{17} cm^{-2}$, the bound-free transitions become noticeable, creating Lyman limit systems; these create distinctive breaks in the spectra of the background QSOs, which have the shape expected for b-f.
  • Finally, at the highest column densities, $\Sigma > 10^{20} cm^{-2}$, the Lyman alpha becomes so strong that the damping wings are apparent, creating damped Lyman alpha systems; (review curve of growth.

• The latter are particularly interesting because the have surface densities comparable to those of present-day spirals. Damped systems are not super common - several dozens known to date, but still they contain a significant fraction of the netural H observed in absorption. Some typical spectra: (Charlton and Churchill Fig 1). (Storrie Lombardi II).

• Note that if a significant fraction IGM was neutral, there would be complete absorption - the so-called ``Gunn-Peterson'' trough. This has been observed in redshift $>6$ quasars (e.g. Becker et al Fig 1, suggesting that this era was the end of the reionization era.

• A related set of objects are the metal line absorption systems. These occur in a similar way, but with absorption from elements other than H. These generally appear only in systems with higher surface density, and obviously, only in systems with heavy elements. Because of the latter, they are highly likely to be identifiable with galalxies, and in fact, have been imaged as such. Metal-line systems have been more extensively discussed in the context of intermediate redshift objects because for such objects, the relevant resonance lines fall in the optical (Januzzi Figure 1); table of common absorbers.

• Deep images suggest that at least the high column density systems are associated with galaxies; a galaxy is usually found with a few tens of kpc of the quasar at the appropriate redshift.

• Metal-line systems are generally associated with higher column density HI systems, although there is now a class of weak MgII systems that may not be associated with galaxies.

• Consider the distribution of absorbers as a function of column density. Generally, this may also be a function of redshift as well. Common practice has been to fit a power law distribution of the form $f(N) = k N^{-\beta}$. Typical derived values have been in the range 1.4 to 1.7 for $\beta$. Note that derived values are sensitve to range of N sampled, choice of binning, also that these power laws don't converge. However, possible better fit obtained with Schecter-like function (SL III Figs 3 and 6).

• We can also consider whether there is evidence for evolution in the population of absorbers, as we looked for in galaxies. As for galaxies, cosmological geometry plays a large role. The predicted number densities , one gets:
  
  $$N(z) = \Phi(0)\pi R_0^2 c H_0^{-1} (1+z) (1+2q_0z)^{-1/2}$$
  
  
  which for no evolution ($\Phi$ and $R_0$ constant in time), one gets $n\propto (1+z)^\gamma$ with $\gamma$ equal 1 for $q_0=0$ or 0.5 for $q_0=0.5$. Note that absorber abundance as a function of redshift also will depend on structure formation and ionization background, since increased ionization background will ionize clouds and reduce the overall number of neutral systems.
  • An evolutionary effect is immediately obvious (Januzzi figs 2-4) for Lyman alpha systems; small, neutral gas clouds were much more abundant at high redshift. However, the relation of these to galaxies in uncertain.

  • For Lyman limit and damped lyman alpha systems, the situation may be less less clear. The distribution as a function of column density is a function of redshift (SL III, Fig 5). There will be effects both from the growth of systems, but also from the conversion of gas into stars. The data is marginally consistent with no evolution for low density universe, but probably requires evolution in higher density universe such that more clouds were around in the past out to intermediate redshift. However, highest redshift data suggests a drop in density past $z\sim 3.5$ for the highest density systems (SL III Fig 9). Potential interpretation: at highest redshift, we are seeing epoch when large clouds got assembled, since then, gas clouds gradually get formed into stars, i.e. galaxies!

• Considering implications for intermediate redshift evolution of galaxies (not clouds) - Steidel et al have identified galaxies selected by absorption strength, and find essentially no evolution with redshift at all in absolute magnitude; however, if small sample is split by color (CFRS VI), then possible evolution is seen for blue galaxies - consistent with results of deep redshift surveys.

• One can estimate the mean density of neutral gas as a function of redshift:
  
  $$<\Omega_g> = {H_0 \mu m_H\over c \rho_{crit}} \int_{Nmin}^{\infty} N f(N,z) dN$$
  
  
  and one finds it increases to higher z, but possibly rolls over at highest z, consistent with above interpretation (SL IV Fig 1). This also shows plausible identification of damped systems with galaxies, since density in high z systems is comparabel to that observed in stars today.

• One can probe SF history in these clouds by measuring metallicities. One finds at $z\sim 2$, typical metallicities are about 0.1 solar in high column density systems, but with wide variation. Evidence that some refractory elements have been converted into dust, but possibly with dust to gas ratio of only about 0.1 of current day value. Issue of current interest: is the observed decrease in $\Omega_g$ with time consistent with observed metallicity increase assuming simple closed-box type model? Depends on SF history as well as on presence of dust (Pei & Fall ApJ 454, 69) - indications are that SF peaked at Z=1-2, which seems to fit with observations of high redshift galaxies as discussed previously, although both are very crude!

• With high resolution observations, can now get detailed abundances of DLA systems. This allows one to probe the abundance and ionization structure of the clouds, e.g. identifying multiphase media in the objects.

• Find relatively large line widths from low res observations. High spectral resolution also allows one to probe velocity structure in DLAs - are these rotating disks, or assembling protogalactic clumps? (cartoon and data,) data + toy models)

• Other issue of interest more relevant to cosmology/large scale structure than to galaxy formation: are absorption systems clustered, and if so, what is clustering amplitude as f(z)?