Star Cluster (SC) formation is major/dominant mode of all star formation (SF) (deGrijs+03) and occurs in very different environments. This raises the question whether young SCs (YSCs) are similar or different in different environments, individually or as a population.
YSCs trace SF distribution (& its history) within a galaxy, old SCs - Globular Clusters (GCs) - trace violent SF phases in their parent galaxy over a Hubble time. But SCs also fade and dissolve. The youngest SCs can still be embedded in their native cocoon while part of the older SCs already gone. It is, hence, important to account for all these processes when comparing SC populations in different galaxies. They all depend on the initial properties of individual SCs (masses, radii, abundances, IMF) and of the SC population (luminosity function LF, mass function MF, distribution of radii, ages, …)

In the Tadpole and Mice galaxies, we found >35% of all SF to go into the formation of YSCs. This analysis needs to be extended to different types of galaxies (starburst/non-starburst, dwarf/normal, gas-rich/gas-poor, interacting/non-interacting, in various stages of interaction) to explore the systematics. There are indications that the ratio between the SFR that goes into the formation of massive long-lived SCs and the SFR that goes into low-mass, short-lived SCs increases with increasing burst strength or overall SFE.

Detection and accurate photometry of SCs is possible to large distances (<= 100 Mpc). We have developed an analysis tool AnalySED that largely automatically compares observed multi-band spectral energy distributions (SEDs) of SCs with a grid of GALEV models and determines their ages, metallicities, extinction values, and masses (Anders+04a). Artificial star cluster tests with model and observational uncertainties have shown that a UV- or U-band is important for age dating of YSCs, while one NIR-band is important for their metallicities. For YSCs (in dusty galaxies), 4 passbands (UV/U,..., H or K) are required, for old GCs (in dustfree galaxies) 3 passbands (U/B,..., H or K) are enough to disentangle ages and metallicities and get ages to Delta age/age < 0.3, metallicities to +/- 0.2 dex (Anders+04a, de Grijs+03). We found the isolated dwarf starburst galaxy NGC 1569 has not formed (many) new GCs (ESA/NASA Press Release 0406), while the starburst in the massive gas-rich spiral-spiral merger NGC 7252 formed many new GCs. The ongoing starburst in NGC 4038/39 forms a rich YSC system, its luminosity function (LF) features a turnover to 99.5\% significance (Anders+07), different from the LF of YSCs in dwarf, spiral, or isolated starbursts, and different from the mass function of molecular clouds and molecular cloud cores in undisturbed galaxies. Due to problems with the translation of a completeness limit in luminosity into a completeness limit in mass we could not yet determine the MF. If the turnover in the LF would reflect a turnover in the MF, this would tie in nicely with Parmentier \& Gilmore's  05, 07 result that the Milky Way GC system initially had a mass spectrum with turnover around 10^5 Msun and indicate that the MF of the molecular clouds in the massive gas-rich Antennae merger (LIRG) is different from what it is in undisturbed spirals, dwarf galaxy starbursts, as expected due to pressure effects. GC formation,  destruction, and the MF of SC systems are being investigated in collaboration with P. Anders and H. Lamers (Utrecht), R. de Grijs (Sheffield) and G. Parmentier (Bonn). A worrisome result showed up recently in our analysis of SC observations in apparently undisturbed actively SFing spirals observed by S. Larsen: quite a number of these YSCs have masses and radii in the range of GCs! Either these spirals are very hostile to their SCs and will destroy even the strongly bound ones soon or one would expect a continuous age distribution of GCs. In collaboration with S. Larsen, P. Anders, H. Lamers, R. de Grijs we began to explore this further.

The condition for the formation of star clusters, and of massive long-lived clusters in particular, as well as the survival and destruction during secular evolution are important to explore in view of the potential of GCs to trace the formation history of their parent galaxies.

To explore in detail the feedback from strongly clustered violent SF as opposed to more diffuse lower level SF is important in view of a future coupling of GALEV with a dynamical code. E.g. has the PhD thesis by Marc Westmoquette (UCL), that I coadvised, shown, combining integrated field spectroscopy with various resolutions/field sizes, that massive star clusters can be the starting points of a major galactic wind and that this wind entrains a substantial amount of neutral gas.  

Most elliptical and S0 galaxies feature two peaks in their optical GC colour distributions with a universal blue peak of old and metal-poor GCs and a variable (in colour and relative height) red peak, the origin of which is still controversial. Different hypotheses for the formation of the red GC subpopulation, ranging from in situ formation in a 2nd step within their galaxy to formation during a major gas-rich merger and hierarchical accretion of a number of dwarf galaxies, predict different age, metallicity, and spatial distributions for the red peak GCs. We have embarked on a campaign to investigate the nature of the GC population(s) in this red peak. First results from a feasibility study on a Virgo cluster elliptical for which we supplemented existing B- and I-imaging with own deep K-data showed that the optical red peak splits up into several subpopulations of GCs with different age-metallicity combinations, the youngest being only ~1 Gyr old. We will investigate more elliptical galaxies like this to find out their violent SF histories. At the same time, we will test the luminosity functions of all GCs and of the blue-peak GCs only in their accuracies for distance determination. We anticipate that using the turn-over in the luminosity function of the universal blue peak GCs only will allow to significantly improve this method of distance determination.

References: