2015MNRAS.452.1045F

We present a novel, observationally-based framework for the formation epochs and sites of globular clusters (GCs) in a cosmological context. Measuring directly the mean ages of the metal-poor and metal-rich GC subpopulations in our own Galaxy, and in other galaxies, is observationally challenging. Here we apply an alternative approach utilizing the property that the galaxy mass-metallicity relation is a strong function of redshift (or look-back age) but is relatively insensitive to galaxy mass for massive galaxies. Assuming that GCs follow galaxy mass-metallicity relations that evolve with redshift, one can estimate the mean formation epochs of the two GC subpopulations by knowing their mean metallicities and the growth in host galaxy mass with redshift. Recently, the SAGES Legacy Unifying Globulars and GalaxieS (SLUGGS) survey has measured the spectroscopic metallicities for over 1000 GCs in a dozen massive early-type galaxies. Here we use these measurements, and our new metallicity matching method, to infer a mean age for metal-rich GCs of 11.5 Gyr (z = 2.9) and a range of 12.2-12.8 Gyr (4.8 < z < 5.9) for the metal-poor GCs, depending on whether they mostly formed in accreted satellites or in situ within the main host galaxy. We compare our values to direct age measurements for Milky Way GCs and predictions from cosmological models. Our findings suggest that reionization preceded most GC formation, and that it is unlikely to be the cause of GC bimodal metallicity distributions.