How well do we know the age and mass distributions of the star cluster system in the Large Magellanic Cloud?

Abstract

[ABRIDGED] The LMC star cluster system offers the unique opportunity to independently check the accuracy of age and mass determinations based on a number of complementary techniques, including isochrone analysis. Using our sophisticated tool for star cluster analysis based on broad-band spectral energy distributions (SEDs), we reanalyse the Hunter et al. (2003) LMC cluster photometry. Our main aim is to set the tightest limits yet on the accuracy of ABSOLUTE age determinations based on broad-band SEDs, and therefore on the usefulness of such an approach. Our broad-band SED fits yield reliable ages, with statistical absolute uncertainties within Delta[log(Age/yr)] = 0.4 overall. The systematic differences we find with respect to previous age determinations are caused by conversions of the observational photometry to a different filter system. The LMC's cluster formation rate (CFR) has been roughly constant outside of the well-known age gap between ~3 and 13 Gyr, when the CFR was a factor of ~5 lower. We derive the characteristic cluster disruption time-scale, log(t_4^dis/yr) = 9.9 +- 0.1, where t_dis = t_4^dis (M_cl/10^4 Msun)^0.62. This long characteristic disruption time-scale implies that we are observing the INITIAL cluster mass function (CMF). We conclude that the youngest mass and luminosity-limited LMC cluster subsets show shallower slopes than the slope of alpha = -2 expected (at least below masses of a few x 10^3 Msun), which is contrary to dynamical expectations. This may imply that the initial CMF slope of the LMC cluster system as a whole is NOT well represented by a power-law, although we cannot disentangle the unbound from the bound clusters at the youngest ages.