We present the results of a deep Hubble Space Telescope ( HST ) exposure of the nearby globular cluster NGC 6397 , focussing attention on the cluster ’ s white dwarf cooling sequence .
This sequence is shown to extend over 5 magnitudes in depth , with an apparent cutoff at magnitude F 814 W \sim 27.6 .
We demonstrate , using both artificial star tests and the detectability of background galaxies at fainter magnitudes , that the cutoff is real and represents the truncation of the white dwarf luminosity function in this cluster .
We perform a detailed comparison between cooling models and the observed distribution of white dwarfs in colour and magnitude , taking into account uncertainties in distance , extinction , white dwarf mass , progenitor lifetimes , binarity and cooling model uncertainties .
After marginalising over these variables , we obtain values for the cluster distance modulus and age of \mu _ { 0 } = 12.02 \pm 0.06 and T _ { c } = 11.47 \pm 0.47 Gyr ( 95 % confidence limits ) .
Our inferred distance and white dwarf initial-final mass relations are in good agreement with other independent determinations , and the cluster age is consistent with , but more precise than , prior determinations made using the main sequence turnoff method .
In particular , within the context of the currently accepted \Lambda CDM cosmological model , this age places the formation of NGC 6397 at a redshift z \sim 3 , at a time when the cosmological star formation rate was approaching its peak .