Astron. Astrophys. 355, 949-965 (2000)

1. Introduction

The nature and origin(s) of Galactic Bulges are key aspects of any galaxy formation model, and of the Hubble galaxy classification sequence. However, the present-day properties of bulges in spiral galaxies are not well known (e.g. Silk & Wyse 1993; Wyse et al. 1997). Do bulges form late, early or continuously? Are bulges related to halos? To disks? Are they single stellar populations? The existence of smooth and/or discontinuous gradients in age and/or metallicity in the stellar population(s) in the Galactic Bulge can help to discriminate between these different scenarios, and is the topic of this paper.

Is there evidence for the widely repeated assumption that the Galactic Bulge is old? Recently published colour magnitude diagrams from HST/WFPC2 and HST/WFC1 of the Galactic Bulge and the bulge globular clusters (Vallenari et al. 1996; Ortolani et al. 1995; Holtzman et al. 1993) while dominated by fairly old stars, show a substantial population of stars above the dominant old turnoff. Are these foreground disk stars, or is there a minority very young bulge population? Since even a minority young bulge population is of interest, we examine here the limits on young stars in the bulge windows.

One common approach to determine the properties of the Bulge is to study the so called bulge globular clusters (see e.g. Zinn 1996; Ortolani et al. 1995; Minniti 1996). While primarily motivated by observational convenience, the rationale behind this approach is the assumption that these clusters may be valid tracers of the stellar population(s) of the Bulge (see however Zinn 1996 and Harris 1998) Formation scenarios relevant to this approach include the possibility that the Bulge has been assembled from numerous such clusters and these are the last surviving (Gnedin & Ostriker 1997), or that the clusters formed a system associated with the Bulge rather than with the rest of the spheroidal component(s) of the Galaxy. Thus the idea is that we may be able to infer the age and/or metallicity of the Bulge stellar population either directly from studies of these clusters, or through differential studies of the clusters and field stars, under the assumption of similar metallicity. Since there is no ab initio understanding of the formation of either galactic bulges or globular clusters, and the age range of the globular cluster system remains a topic of active debate, such analyses merit close scrutiny. The most recent and extensive such analysis is that of Ortolani et al. (1995, 1996) who observed two such globular clusters, NGC6553 and NGC6528, with HST/WFPC2, and deduced that the Bulge has the same age as the Halo.

Analysis of suitably-chosen globular clusters introduces several possible complications. The first is the major problem of defining a proper population of `bulge' globular clusters. Some of the clusters used, e.g. Ter7, have recently been shown to be associated with the satellite dwarf galaxy Sgr dSph, rather than with the Galaxy itself. They may therefore not be representative of the Galactic Bulge. The method of comparing ridge-lines of globular clusters to infer relative ages requires the clusters in question to have similar metallicities, and relative chemical abundances of the alpha-elements, to avoid an age-metallicity degeneracy (Stetson et al. 1996; VandenBerg et al. 1990, 1996). New results by Cohen et al. (1999) show that NGC6553 may be as much as dex more metal-rich than 47 Tuc, illustrating the potentially large effects of metallicity range. The impressive recent study of Rosenberg et al. (1999), indicating a dispersion in ages for the intermediate metallicity globular clusters, and a large systematic age difference between the metal-rich and metal-poor clusters illustrates the complexity.

Another potential uncertainty is that the metallicity distribution function of the stars in the galactic Bulge is more similar to that in the solar-neighbourhood (cf. Wyse & Gilmore 1995) than to that for clusters within 5 degrees of the Galactic centre (Minniti 1996; Barbuy et al. 1998). The bulge cluster distribution is both more narrow and less metal-rich than the bulge field stars, complicating any direct comparison.

Clearly, if one wishes to know the age of the bulge field stars, it is desirable to observe the stars in the Bulge directly. Direct studies of the Bulge are difficult due to the severe crowding towards the central regions of the Galaxy and the large, patchy, reddening along the line of sight. Several detailed studies of the outer Galactic Bulge exist, providing kinematics and chemical abundance distribution functions (Ibata & Gilmore 1995a, 1995b; Minniti et al. 1995; see also Wyse et al. 1997). For the inner Bulge several analyses of the low reddening Baade's window are available (e.g. Ortolani et al. 1995; Vallenari et al. 1996; Holtzman et al. 1993; Terndrup 1988; Ng et al. 1996), with direct studies of the inner bulge field stars in the near-IR recently also becoming available (Frogel et al. 1999), and even mid-IR ISO photometry (Omont et al. 1999; Glass et al. 1999). Additionally, many recent studies have emphasized the high continuing rate of star formation in the inner bulge/disk. Do these stars diffuse with time the few hundred parsecs into the Sgr and Baade's windows?

The interpretation of extant data is unclear, with a variety of contradictory results. Vallenari et al. (1996) use a mixture of WF/PC-1 and NTT data while Holtzman et al. (1993) rely exclusively on (the same) WF/PC-1 data. Both groups reached the conclusion that the Bulge is dominated by a significant young stellar populations. Ortolani et al. (1995) using similar NTT data for Baade's Window found the Bulge to be as old as the Halo.

The confusion among the results from the space based observations should be contrasted with ground based optical observations which find little evidence for a substantial young stellar population(s) in the Galactic Bulge (eg Terndrup 1988), albeit rather far from the centre. Note however that these observations do not cover the main-sequence turn-off and the results are based on the giant branch. The main-sequence turn-off is more sensitive to detection of a significantly younger stellar populations. Further complication is provided by studies of OH/IR stars, suspected to be of intermediate age, which are common in the inner bulge, or disk (Sevenster et al. 1997).

The distribution function of chemical abundances is a key parameter defining a stellar population. In the outer Galactic Bulge Ibata & Gilmore (1995b) derived the relevant distribution function from spectroscopy of K giants. They found a mean abundance of dex, with a very wide dispersion. In Baade's window McWilliam & Rich (1994) and Sadler et al. (1996) provided similar results. Sadler et al. (1996) found for 400 K giants a mean abundance of [Fe/H] dex, with more than half the sample in the range . This is similar to the results from the detailed spectroscopic analysis of McWilliam & Rich (1994). A metallicity gradient has been suspected for fields outside Baade's Window, Terndrup (1988) and Minniti et al. (1995). The important conclusion from the spectroscopic analyses is that the stellar populations of the inner Galactic Bulge are complex, and that their analysis requires careful consideration of projected disk and other populations.

In this paper we study colour-magnitude diagrams, derived from archival HST/WFPC2 images, for 4 fields and two clusters towards the Galactic Bulge and one "disk" cluster. We perform a purely differential study of the properties of the Galactic Bulge population(s), quantifying any systematic offsets and/or gradients in age and/or metallicity in the field population(s).

The paper is organized as follows; in Sect. 2 we detail the observations used and in Sect. 3 describe how we derive the photometric magnitudes from the images. Sect. 4 discusses reddening and distances for the individual fields and clusters, and the utility of the cluster data. Sect. 5 discusses the age of the Bulge, while Sect. 6 asks the question whether a metallicity gradient might be present in the inner Bulge. Sect. 7 includes a summarizing discussion which puts our results into the context of other studies. A brief summary is found in Sect. 8.

© European Southern Observatory (ESO) 2000

Online publication: March 21, 2000
helpdesk.link@springer.de