The ages, masses and star formation rates of spectroscopically confirmed z ∼ 6 galaxies in CANDELS
Dickinson, M.; Ferguson, H. C.; Finkelstein, S. L.; McLure, R. J.; Dunlop, J. S.; Koekemoer, A. M.; Fazio, G. G.; Fontana, A.; Vanzella, E.; Faber, S. M.; Cirasuolo, M.; Ashby, M. L. N.; Hartley, W. G.; Willner, S. P.; Grogin, N. A.; Curtis-Lake, E.; Kocevski, D. D.; Ellis, R. S.; Almaini, O.; Lai, K.; Rogers, A. B.; Schenker, M.; Robertson, B. E.; Bradshaw, E. J.; Targett, T.
United Kingdom, United States, Italy
Abstract
We report the results of a study exploring the stellar populations of 13 luminous (L > 1.2L*), spectroscopically confirmed, galaxies in the redshift interval 5.5 < z < 6.5, all with Hubble Space Telescope (HST) Wide Field Camera 3/infrared and Spitzer Infrared Array Camera imaging from the HST/Cosmic Assembly Near-infrared Deep Survey and Spitzer Extended Deep Survey. Based on fitting the observed photometry with galaxy spectral energy distribution (SED) templates covering a wide range of different star formation histories, including exponentially increasing star formation rates and a self-consistent treatment of Lyα emission, we find that the derived stellar masses lie within the range of 109< M* < 1010 M⊙ and are robust to within a factor of 2. In contrast, we confirm previous reports that the ages of the stellar populations are poorly constrained. Although the best-fitting models for 3/13 of the sample have ages of ≳300 Myr, the degeneracies introduced by dust extinction mean that only two of these objects actually require a ≳300 Myr old stellar population to reproduce the observed photometry. We also explore SED fitting with more general, two-component models (burst plus ongoing star formation), thereby relaxing the requirement that the current star formation rate and assembled stellar mass must be coupled, and allow for nebular line+continuum emission. On average, the inclusion of nebular emission leads to lower stellar mass estimates (median offset 0.18 dex), moderately higher specific star formation rates, and allows for a wider range of plausible stellar ages. However, based on our SED modelling, we find no strong evidence for extremely young ages in our sample (i.e. <50 Myr). Finally, considering all of the different star formation histories explored, we find that the median best-fitting ages are of the order of ≃200-300 Myr and that the objects with the tightest constraints indicate ages in the range of 50-200 Myr.