3.1.3. Morphologies of Lyman-limit systems considered dangerous
Deep HST imaging of galaxies selected on the basis of colour to be at redshifts z > 2 has provided an important breakthrough in our understanding of high-redshift morphology [81, 42]. Such surveys probe systems seen at ages corresponding to 10-20% of the present age of the Universe. So far the results appear somewhat contradictory. Early studies focussed on the compact cores of these systems, which appear well-modelled by r1/4 law profiles, suggestive of proto-ellipticals (Giavalisco et al. 1996). Deeper observations have revealed morphological characteristics among the most bizarre yet seen on deep HST data (Fig. 5). Whether the morphologies of Lyman limit systems are interpreted as fairly regular or totally bizarre is a strong function of the limiting surface brightness of the observations, and whether the complex structures linking the bright knots have been resolved. Thus any interpretation of the morphologies of Lyman limit galaxies remains speculative. In the context of these lectures perhaps the best that can be done is to present explicitly the various factors that make the interpretation of the z > 1.5 morphological data so subtle.
Figure 6. Predicted I-band surface brightness as a function of age for a stellar population with a projected surface mass density of 108 M kpc-2. The three curves shown correspond to stellar populations with exponential star formation histories (with e-folding timescales of 1 Gyr) seen at redshifts z = 0.5, z = 1.0, and z = 2.0. Calculations are based on the predictions of the GISSEL96 spectral synthesis package (Bruzual & Charlot 1996). Also shown is the approximate I-band (F814W) isophotal detection limit for the Hubble Deep Field. Note the precipitous decline in visibility of older stellar populations with redshift. Curves such as these suggest that beyond z > 1.5 substantial mass could be locked in ``old'' (ages greater than one or two Gyr) stellar populations that would be undetectable even in red optical bands.
The most obvious complication is that Lyman break systems are being observed at rest wavelengths so far in the ultraviolet that even the relatively sparse ground-based U-band data of local systems provide a poor reference standard for comparison. (In fact, the handful of local systems observed with the Ultraviolet Imaging Telescope on the Astro-1 and Astro-2 missions provide the best calibration reference.) In any case, Lyman limit systems are at redshifts so high that the effects of evolution must be incorporated explicitly when making meaningful comparisons. At the epoch being observed, the Universe may simply be too young to have evolved the old stellar populations that play an important role in defining local morphology. Also, extraordinarily strong surface-brightness selection effects bias against the detection of even intermediate-aged stellar populations. These latter points are perhaps best understood from plots such as that shown in Figure 6, which illustrates the surface brightness of a surface mass density as a function of population age and observed redshift. Stellar populations with ages greater than 1-2 Gyr are strongly biased against in observations at z > 1.5, and only the youngest stellar populations (the ``tip of the iceberg'' in terms of stellar mass) are detectable. On the other hand, if is large then the Universe at z ~ 2-3 may be sufficiently young that the populations biased against would not have time to have formed, so a large proportion of the total mass would be detectable.