The Sun, stellar-population models, and the age estimation of high-redshift galaxies

Abstract

Given sufficiently deep optical spectroscopy, the age estimation of high-redshif t ($z > 1$) galaxies has been claimed to be a relatively robust process (e.g. Dunlop et al. 1996) due to the fact that, for ages $< 5$Gyr, the near-ultraviolet light of a stellar population is expected to be dominated by `well-understood' main-sequence (MS) stars. Recently, however, the reliability of this process has been called into question by Yi et al (2000), who claim to have developed models in which the spectrum produced by the main sequence reddens much more rapidly than in the models of Jimenez et al (2000a), leading to much younger age estimates for the reddest known high-redshift ellipticals. In support of their revised age estimates, Yi et al cite the fact that their models can reproduce the spectrum of the Sun at an age of 5 Gyr, whereas the solar spectrum is not reproduced by the Jimenez et al models until $\simeq 10$ Gyr. Here we confirm this discrepancy, but point out that this is in fact a {\it strength} of the Jimenez et al models and indicative of some flaw in the models of Yi et al (which, in effect, imply that the Sun will turn into a red giant any minute now). We have also explored the models of Worthey (1994) (which are known to differ greatly from those of Jimenez et al in the treatment of post-MS evolution) and find that the main-sequence component of Worthey's models also cannot reproduce the solar spectrum until an age of 9-10 Gyr. We conclude that either the models of Yi et al are not as main-sequence dominated at 4-5 Gyr as claimed, or that the stellar evolutionary timescale in these models is in error by a factor possibly as high as two. (abridged)