Luminous red galaxies in simulations: cosmic chronometers?

There have been a number of attempts to measure the expansion rate of the Universe at high redshift using luminous red galaxies (LRGs) as 'chronometers'. The method generally assumes that stars in LRGs are all formed at the same time. In this paper, we quantify the uncertainties on the measurement of H(z) which arise when one considers more realistic, extended star formation histories. In selecting galaxies from the Millennium simulation for this study, we show that using rest-frame criteria significantly improves the homogeneity of the sample and that H(z) can be recovered to within 3 per cent at z similar to 0.42 even when extended star formation histories are considered. We demonstrate explicitly that using single stellar populations (SSPs) to age-date galaxies from the semi-analytical simulations provides insufficient accuracy for this experiment but accurate ages are obtainable if the complex star formation histories extracted from the simulation are used. We note, however, that problems with SSP fitting might be overestimated since the semi-analytical models tend to overpredict the late-time star formation in LRGs. Finally, we optimize an observational programme to carry out this experiment.