|Annu. Rev. Astron. Astrophys. 2000. 38: 289-335
Copyright © 2000 by Annual Reviews. All rights reserved
5.1. The Physical Nature of Groups
Simulations of local large-scale structure suggest that a significant fraction of the groups identified in redshift surveys are not real, bound systems (Frederic 1995, Ramella et al 1997). The existence of diffuse X-ray emitting gas is often cited as evidence that a group is real. This is not necessarily the case, however. Hernquist et al (1995) noted that primordial gas may be shock-heated to X-ray emitting temperatures along filaments. When these filaments are viewed edge-on, a "fake" group with an X-ray halo could be observed. Ostriker et al (1995) proposed a test of the Hernquist et al (1995) filament model by defining an observable quantity Q, that is proportional to the axis ratio of the group. Applying this test to the early ROSAT observations of HCGs, Ostriker et al (1995) found that the Q values for most HCGs are consistent with their being frauds. However, the low Q values for groups can also be explained if the ratio of gas mass to total mass is smaller in groups than in rich clusters. Both ROSAT observations (David et al 1995, Pildis et al 1995, Mulchaey et al 1996a) and simulations (Diaferio et al 1994, Pildis et al 1996) of X-ray groups are in fact consistent with this idea, suggesting that the Ostriker et al test may in the end not be very useful.
Several arguments support the idea that at least some X-ray groups are real, bound systems and that the X-ray gas is virialized. In the most X-ray luminous groups, the diffuse gas extends on scales of hundreds of kiloparsecs and appears smooth. This is consistent with what one expects for a "smooth" group potential. The gas temperature in these cases agrees fairly well with the temperature expected, based on the velocity dispersion of the groups. Furthermore, most of these groups show evidence for cooling flows in their centers, suggesting that the gas is in an equilibrium state and has probably existed for at least several gigayears.
Ironically, perhaps the best evidence for the reality of the X-ray luminous groups has come from optical studies of these systems. Zabludoff & Mulchaey (1998) used multifiber spectroscopy to study the faint galaxy population in a small sample of groups and found large differences in the number of faint galaxies in X-ray detected and non-detected groups. All of the X-ray detected groups in the Zabludoff & Mulchaey (1998) sample contain at least 20-50 group members (down to magnitudes as faint as MB ~ -14 + 5 log10 h100). Even down to these relatively faint magnitude limits, many of the X-ray detected groups have very high early-type fractions (nearly 60% in some cases). The large number of group galaxies argue that these X-ray groups must be real, physical systems and not radial superpositions. There are also strong correlations between dynamical measures of the gravitational potential (i.e. velocity dispersion/gas temperature) and the early-type fraction of the group (Zabludoff & Mulchaey 1998, Mulchaey et al 1998). These correlations imply either that galaxy morphology is set by the local potential at the time of galaxy formation (Hickson et al 1988) or that the potential grows as the group evolves (Diaferio et al 1993). Either scenario requires that most X-ray luminous groups be real, bound systems.
However, it is likely that some X-ray detected groups are not virialized systems. In particular, low-luminosity, low-temperature groups tend to have irregular X-ray morphologies with the X-ray emission distributed in the immediate vicinity of individual galaxies. These X-ray morphologies suggest that these groups are still dynamically evolving. In some cases, such as HCG 92, gas has apparently reached X-ray emitting temperatures by other mechanisms such as shocks. Therefore, X-ray detection alone does not indicate that a system is virialized.