Testing Metal-poor Stellar Models and Isochrones with HST Parallaxes of Metal-poor Stars

Hubble Space Telescope (HST) fine guidance sensor observations were used to obtain parallaxes of eight metal-poor ([Fe/H] < -1.4) stars. The parallaxes of these stars determined by the new Hipparcos reduction average 17% accuracy, in contrast to our new HST parallaxes, which average 1% accuracy and have errors on the individual parallaxes ranging from 85 to 144 mu as. These parallax data were combined with HST Advanced Camera for Surveys photometry in the F606W and F814W filters to obtain the absolute magnitudes of the stars with an accuracy of 0.02-0.03 mag. Six of these stars are on the main sequence (MS) (with -2.7 < [Fe/H] < -1.8) and are suitable for testing metal-poor stellar evolution models and determining the distances to metal-poor globular clusters (GCs). Using the abundances obtained by O'Malley et al., we find that standard stellar models using the VandenBerg & Clem color transformation do a reasonable job of matching five of the MS stars, with HD 54639 ([Fe/H] = -2.5) being anomalous in its location in the color-magnitude diagram. Stellar models and isochrones were generated using a Monte Carlo analysis to take into account uncertainties in the models. Isochrones that fit the parallax stars were used to determine the distances and ages of nine GCs (with -2.4 <= [Fe/H] <= -1.9). Averaging together the age of all nine clusters led to an absolute age of the oldest, most metal-poor GCs of 12.7 +/- 1.0 Gyr, where the quoted uncertainty takes into account the known uncertainties in the stellar models and isochrones, along with the uncertainty in the distance and reddening of the clusters.