Introduction

Globular clusters (GCs) are generally regarded as the oldest objects in the universe. They can provide the most stringent lower limit of the age of the universe, and hence, a fundamental constraint on cosmological models. Moreover, GCs are of great significance because they contain the earliest information of galaxy formation, besides, they serve as extragalactic distance ladder by providing means of calibrating the absolute magnitudes of secondary distance indicator, such as RR Lyraes.

There is discrepancy between previously estimated age of GCs and the age of the universe with the bottom line that GCs are formed some time after the birth of the universe. The delay could be at least 0.5 Gyr, for example, but less than 1 Gyr because considerable evolution has already occured in the quasar at redshift of 5 (Sandage 1993). On the one hand, the most recent determinations of the Hubble constant are in the range of kms Mpc (see e.g. Freedman et al. 1997; Hamuy et al. 1996). In an Einstein-de Sitter model, where , this range gives an upper constraint of 11.9Gyr, while in a flat universe with and , an upper limit of age is Gyr. On the other hand, while the ages of globular clusters derived by different groups are diverse before the Hipparcos data, they are generally comparable or even larger than the age of the universe from the cosmological models. Most previous studies have agreed in deriving cluster age of 14 Gyr or more. For example, Bolte & Hogan (1995) gives Gyr, while Chaboyer (1995) suggested Gyr, and VandenBerg et al. (1996) derived Gyr from main sequence fitting to M92.

One way to solve the discrepancy is to check the age of globular clusters, an area which is already extensively studied. A good review on this is given by VandenBerg et al. (1996). While the age determination of clusters rests mainly on the accuracy of their distances, large uncertainties always remain in obtaining them. Basically there are five ways toward GC distances. Astrometric distances are obtained through a direct comparison of the proper motion and radial velocity dispersions within a cluster. This method is independent of reddening (Cudworth 1979), though, it requires a dynamical model of a cluster, and is of relatively low precision. A recent observation is done by Rees (1996), who gives the distances of eight globular clusters. The basic idea of white dwarf sequence fitting (Fusi Pecci & Renzini 1979) is to use WDs as samples, and to fit the WD cooling sequence of a globular cluster to an appropriate empirical cooling sequence constructed using local WDs with well-determined trigonometric parallaxes. This method does not depend on metallicity, which introduces a lot of uncertainties, nor does it depend on convection in stellar models. Moreover, local white dwarfs are lot more abundant than subdwarfs. However, WDs are very faint in globular clusters, even in the nearest ones (De Marchi et al. 1995; Richer et al. 1995). HST observations are made to derive the distance of NGC 6752 (Renzini et al. 1996), leading an age of Gyr. There are also uncertainties in reddening. Furthermore, there is a question of whether the calibrating WDs have the same mass, and thus the same parent popultions with those in clusters.

Subdwarf main sequence fitting envisions the idea of comparing the cluster MS stars with a suitable template (Sandage 1970). Since old globular clusters are population II objects, and are usually metal-poor, subdwarfs with low metallicity are better candidates for calibration than ZAMS with solar metallicity because theoretical models predict that the location of the ZAMS is a sensitive function of metallicity. To estimate the distances, the observed cluster color-magnitude diagram can either be compared empirically against the subdwarf main- sequence of appropriate metallicity, each of the stars having accurate trigonometric parallax measurements, or be matched to theoretical isochrones calibrated against the local subdwarfs. The latter needs to transform the theoretical HR diagram to the observational color-magnitude diagram with accurate bolometric corrections and temperature-color calibrations. Finally, one thing to bear in mind is that cluster abundances are usually measured using red giant stars while the field subdwarfs are unevolved.

The fourth way of estimating globular cluster distances is through the calibration of RR Lyraes absolute magnitudes via LMC, by comparing them with thoses in clusters. However, as Chaboyer et al. (1997) stated, the distance moduli of LMC measured by different groups give quite diverse results, apart from the fact that M(RR) depends on both the age of the stars and metallicity. After encompassing the present results, Chaboyer et al. (1997) give a value of M(RR) at [Fe/H]=-1.9. Theoretical HB models is another way to predict M(RR) at different metallicity and helium abundances. With the understanding in the basic physics of stellar evolution, synthetic HB models for various clusters can be constructed.

Main sequence fitting of isochrones with subdwarfs is the most commonly used, though, there is always a shortage of local subdwarf samples with sufficient precision. The Hipparcos data of subdwarfs has improved markedly the accuracy in parallaxes, both for the "classical" subdwarf calibrators, and for metal-poor samples with increased number so that one can isolate sufficient subdwarfs within a relatively narrow abundance range.

Using the newly released Hipparcos subdwarf data with typical accuracies of mas, better distance of globular clusters are estimated by several groups (Reid 1997; Gratton et al. 1997; Chaboyer et al. 1997). Different techniques are adopted in these estimates, though, they all tend to give larger distance moduli for their sample Galactic globular clusters and thus younger ages for the oldest GCs by upto than previous values. A typical value given by Gratton et al. (1997) is Gyr. In section 2, we introduce the techniques and methods adopted by these groups and their results in GC age will be presented in §3. A discussion on the impact of younger globular clusters is in section 4.