Next Contents Previous

6. FIVE UNSOLVED PROBLEMS IN GALAXY FORMATION

1. What does a protogalaxy look like? The term protogalaxy has been used loosely here and in the literature to describe young galaxies at high redshift. Part of the difficulty is that once an object has sufficient stars to be observed in rest-frame UV or optical radiation, we consider it a galaxy. But before this time it is either unobservable or only observable in absorption (e.g. DLAs), X-ray emission from a supermassive black hole (quasars/AGN), or in far-infrared radiation from dust which could be enshrouding either a powerful AGN or rapid star formation. If dark matter halo collapse, initial star formation and supermassive black hole formation all occur simultaneously, the formation epoch of the galaxy is well-defined, and the picture is simple. But it is possible that many collapsed halos remain quiescent clouds of neutral gas until star formation is triggered by later mergers; these objects could comprise the half of DLAs that fail to show significant cooling in the [CII] 158 micron line (Wolfe et al. 2004). The distribution of lag times between dark matter halo collapse, supermassive black hole formation, and rapid star formation remain uncertain.

2. When/how did each component of the Galaxy form? Observations indicate that the thin disk formed at z appeq 2, but simulations have trouble creating disk galaxies. One area now receiving attention is the manner in which angular momentum coupling between dark matter and baryons affects bar/disk formation and the cuspiness of bulges. It is still not clear if the globular clusters should be considered Galactic components or were all formed earlier and captured, despite evidence that some globular clusters are captured dwarf galaxies. Could globular clusters have formed in the same low-mass halos that met the Jeans threshold for collapse after recombination and hosted the Population III stars?

3. When/how did galaxy sequences evolve? HST observations of morphologies of galaxies at z > 2 imply that the Hubble sequence was not yet present. This is somewhat subtle, as cosmological surface brightness dimming would make a face-on spiral appear very different at high redshift, but most objects display irregular morphology and even the most promising edge-on disk candidates show spectroscopic kinematics inconsistent with the presence of disks (Erb et al. 2004). However, in the low-redshift (z < 1) universe we see a clear bimodality in the distribution of galaxy properties, the so-called red and blue sequences (e.g. Bell et al. 2004, Kannappan 2004). Such bimodalities are unlikely to arise from cosmological structure formation but are presumably caused by baryonic physics and appear directly linked to the "downsizing" behavior discussed in Section 1.1.

4. What role did feedback play? The non-linear baryonic physics of star formation leads to highly energetic processes (ultraviolet radiation, stellar winds, supernova explosions) that can ionize or expel neutral gas that would otherwise participate in further star formation. It is now clear that the processes of galaxy and supermassive black hole formation are intimately connected, as evidenced by striking correlations between the masses of black holes and the velocity dispersions (or masses) of bulges in which they are embedded (Gebhardt et al. 2000, Ferrarese & Merritt 2000, Kormendy & Richstone 1995, Magorrian et al. 1998). Possible explanations include simultaneous hierarchical growth of galaxies and their central black holes through mergers (Haehnelt & Kauffmann 2000, Di Matteo et al. 2005), a strong coupling between black hole accretion and star formation in proto-disks at high redshift (e.g. Burkert & Silk 2001), and the effects of AGN feedback on the surrounding intergalactic medium (Scannapieco et al. 2005). One way or another, it appears that feedback from AGN, supernovae, and galactic winds must regulate the joint formation of the bulge and central black hole. Feedback may also play a role in determining the cuspiness of the dark matter halos, which does not appear consistent with profiles predicted from N-body simulations (Silk 2004). The galactic winds play a critical role in metal enrichment of the intergalactic medium and probably play a lesser role in ejecting neutral gas from the galaxies. As mentioned above, supernova feedback may explain the apparent minimum galaxy mass.

5. When/how was the universe reionized? A major area of ongoing investigation is the reionization epoch when the intergalactic medium was ionized. Slightly inconsistent results have been reported for the reionization redshift from WMAP observations of the temperature-E-mode cross-power-spectrum (zr = 20 ± 9, Bennett et al. 2003) and the apparent end of reionization where the neutral hydrogen fraction dropped to 0.01 as seen in SDSS quasar spectra at z appeq 6.3 (Fan et al. 2002). It seems premature to hypothesize bimodal models of reionization where separate families of sources produce the "early" ionization seen by WMAP and the completion of reionization seen by SDSS. Nonetheless, it is unclear at present which sources reionized the universe, and the leading candidates are the first generation of zero-metallicity stars (Population III) and starbursting galaxies including LBGs and LAEs. The quasars have very hard, ionizing spectra but were not numerous enough to reionize the universe at z > 6; they appear to dominate HeII reionization at z ~ 3. Significant uncertainties exist regarding the nature of the Population III stars: did they form in 106 Modot dark matter halos that collapsed after recombination, or in larger galaxies later on? A top-heavy initial mass function (IMF) is presumed for Population III, but what was the exact mass range and nature of stellar death? Did multiple stars occur per halo, or did the death of the first very massive star prevent further star formation or cause sufficient metal enrichment to generate Population II stars?

Next Contents Previous