Spectral dating of high-redshift stellar populations

Abstract

Age derivation techniques for unresolved stellar populations at high redshifts are explored using the near-ultraviolet (NUV) spectrum of LBDS 53W091 (z = 1.55) and LBDS 53W069 (z = 1.43). The photometry and morphology of these galaxies - which are weak radio sources -suggest they are early-type systems, a feature that makes them ideal testbeds for the analysis of their ages and metallicities with the use of population-synthesis models. In the analysis that is based on simple stellar population models, we find a significant degeneracy between the derived ages and metallicities both in optical+near-infrared (NIR) photometric and NUV spectroscopic analyses. This degeneracy is not so strong for LBDS 53W069. However, even in this case the stellar age cannot be constrained better than to a range roughly encompassing one-third of the age of the Universe at z = 1.43 (90 per cent confidence level). We have explored several independent population-synthesis models and consistently found similar results. Broadband photometry straddling the rest-frame 4000-A break is also subject to a strong age-metallicity degeneracy. The use of realistic chemical enrichment assumptions significantly helps in disentangling the degeneracy. Based on this method, we derive the average stellar age for both galaxies around ∼ 3.6-3.8 Gyr with better constraints on the youngest possible ages (∼3 Gyr at the 90 per cent confidence level). The comparison with simple stellar population models suggest subsolar metallicities (log Z/Z ○. = -0.2). A composite model using chemical enrichment gives slightly higher metallicities in both galaxies (log Z/Z ○. = - 0.1). Given that the stellar component in galaxies forms over times which are larger than a typical chemical enrichment time-scale, we conclude that composite stellar populations must be used in all photospectroscopic analyses of galaxies. From the observational point of view, the most efficient (and feasible) way to set limits on unresolved stellar populations comprises a combination of Balmer absorption lines along with either low signal-to-noise ratio (SNR) rest-frame NUV spectroscopy or accurate optical and NIR photometry.