THE STELLAR AGES AND MASSES OF SHORT GAMMA-RAY BURST HOST GALAXIES: INVESTIGATING THE PROGENITOR DELAY TIME DISTRIBUTION AND THE ROLE OF MASS AND STAR FORMATION IN THE SHORT GAMMA-RAY BURST RATE

We present multi-band optical and near-infrared observations of 19 short gamma-ray burst (GRB) host galaxies, aimed at measuring their stellar masses and population ages. The goals of this study are to evaluate whether short GRBs track the stellar mass distribution of galaxies, to investigate the progenitor delay time distribution, and to explore any connection between long and short GRB progenitors. Using single stellar population models we infer masses of log(M-*/M-circle dot) approximate to 8.8-11.6, with a median of < log(M-*/M-circle dot)> approximate to 10.1, and population ages of tau(*) approximate to 0.03-4.4 Gyr with a median of <tau(*)> approximate to 0.3 Gyr. We further infer maximal masses of log(M-*/M-circle dot) approximate to 9.7-11.9 by assuming stellar population ages equal to the age of the universe at each host's redshift. Comparing the distribution of stellar masses to the general galaxy mass function, we find that short GRBs track the cosmic stellar mass distribution only if the late-type hosts generally have maximal masses. However, there is an apparent dearth of early-type hosts compared to the equal contribution of early-and late-type galaxies to the cosmic stellar mass budget. Similarly, the short GRB rate per unit old stellar mass appears to be elevated in the late-type hosts. These results suggest that stellar mass may not be the sole parameter controlling the short GRB rate, and raise the possibility of a two-component model with both mass and star formation playing a role (reminiscent of the case for Type Ia supernovae). If short GRBs in late-type galaxies indeed track the star formation activity, the resulting typical delay time is similar to 0.2 Gyr, while those in early-type hosts have a typical delay of similar to 3 Gyr. Using the same stellar population models, we fit the broadband photometry for 22 long GRB host galaxies in a similar redshift range and find that they have significantly lower masses and younger population ages, with < log(M-*/M-circle dot)> approximate to 9.1 and <tau(*)> approximate to 0.06 Gyr, respectively; their maximal masses are similarly lower, < log(M-*/M-circle dot)> approximate to 9.6, and as expected do not track the galaxy mass function. Most importantly, the two GRB host populations remain distinct even if we consider only the star-forming hosts of short GRBs, supporting our previous findings (based on star formation rates and metallicities) that the progenitors of long and short GRBs in late-type galaxies are distinct. Given the much younger stellar populations of long GRB hosts (and hence of long GRB progenitors), and the substantial differences in host properties, we caution against the use of Type I and II designations for GRBs since this may erroneously imply that all GRBs which track star formation activity share the same massive star progenitors.