Setting the stage for cosmic chronometers I. Minimizing frosting with an optimized selection of cosmic chronometers

The expansion history of the Universe can be constrained in a cosmology-independent way by measuring the differential age evolution of cosmic chronometers. This yields a measurement of the Hubble parameter as a function of redshift. The most reliable cosmic chronometers known so far are extremely massive and passively evolving galaxies. Age-dating these galaxies is, however, a difficult task, and even a small contribution of an underlying young stellar population ("frosting") could, in principle, affect the age estimate and its cosmological interpretation. We present several spectral indicators to detect, quantify and constrain such a young component in old galaxies, and study how their combination can be used to maximize the purity of cosmic chronometers selection. In particular, we analyze the CaII H/K ratio, and the presence (or absence) of H$\alpha$ and [OII] emission lines, higher order Balmer absorption lines, and UV flux; each indicator is especially sensitive to a particular age range, allowing us to detect young components ranging between 10 Myr and 1 Gyr. The combination of these indicators minimizes the contamination below the 1% level, and offers a way to control the systematic error it could introduce in the Hubble parameter determination. We show that for our previous measurements this effect is well below the current errors. We envision that these indicators will be instrumental in strengthening the selection criterion of cosmic chronometers, paving the road for a robust and reliable dating of the old population and its cosmological interpretation.