Globular Cluster Formation

The discovery of young globular clusters in merging galaxies and other environments provides an opportunity to study directly the process of globular cluster formation. Empirically it appears that globular cluster formation occurs preferentially in regions in which star formation occurs at a high rate and efficiency. Further, the interstellar medium in such regions is likely to be at a higher pressure than less active star-forming environments. An additional observational clue to the globular cluster formation process is that young globular clusters have little or no mass-radius relationship. In this paper I argue that high pressure and high star-formation efficiency are responsible for current globular cluster formation. I suggest that the precursors to globular clusters are molecular clouds and that the mass-radius relationship exhibited by such clouds is wiped out by a variable star formation efficiency. 1 Empirical Foundations Early models of globular cluster formation were largely motivated by two observational results: Milky Way globular clusters are old and massive. Consequently, these models tended to exploit physical conditions unique to the early universe that might give rise to bound clusters of stars with masses around 10 M . Over the last decade or so, observations of extragalactic globular cluster systems and the discovery of young globular clusters have dramatically expanded the empirical basis of globular cluster formation theories. Perhaps most importantly, young globular cluster systems allow the formation process to be probed directly. In this section I describe the observations that are useful in investigating and constraining the process of globular cluster formation. 1.1 What do we know? The Milky Way globular cluster system is comprised of at least two distinct populations (e.g. Armandroff and Zinn 1988 and references therein). The more numerous metal-poor clusters are distributed in a spherical halo, whereas the metal-rich clusters have spatial and kinematic properties similar to the bulge or thick disk. Despite the marked distinction in these properties between the two populations, the mass distributions of the metal-poor and metal-rich clusters are indistinguishable. Other spiral galaxies also show evidence for similar metal-rich and metal-poor populations of globular clusters, the most compelling case being M31 (Ashman and Bird 1993; Barmby et al 2001; Perrett et al 2002).