Astron. Astrophys. 329, 937-942 (1998)

2. The reference metallicity scale

Traditionally, all indices have been tied up to now to the metallicity scale defined by the Zinn's group (Zinn & West 1984, Armandroff et al. 1992, Da Costa & Armandroff 1990; hereinafter ZW on the whole), systematically ignoring any later result from high dispersion spectroscopy. ZW's scale is a compilation of metallicities from several different parameters, all referred to the integrated parameter Q₃₉, and tied to a high resolution spectroscopic scale based on old photographic echelle spectra (see Zinn & West 1984). Moreover, there have been in the past years several claims (e.g. Manduca 1983; Frogel et al. 1983) that the integrated light measurements, upon which the ZW scale was primarily based, are likely to underestimate the true metallicities of clusters with exceptionally blue horizontal branches (HBs), since the HB morphology affects the determination of the Q₃₉ index.

Very recently, Carretta & Gratton (1997; CG97) used high dispersion, high signal-to-noise spectra of more than 160 red giants in 24 clusters to derive a new metallicity scale based on direct detailed abundance analysis, coupled with the most recent and upgraded model atmospheres (Kurucz 1992). The observational material consisted in equivalent widths (EWs) measured on high quality CCD echelle spectra. Different sets of EWs taken from literature were carefully checked and, if necessary, brought on a common, homogeneous system, tied to EWs from the highest resolution spectra. The same set of atomic line parameter was used for all stars, with model atmospheres from the Kurucz grid. Also the reference value for the solar [Fe/H] was obtained from the same set of atomic parameter and the solar model extracted from the same grid used for giant stars. Input atmospheric parameters (effective temperature and gravity) for all stars were obtained from the Frogel et al. papers (e.g. Frogel et al. 1983, but see detailed references in CG97), mostly based on accurate infrared colours and magnitudes. The choice of a temperature scale has an impact on the derived metallicities: CG97 estimate that the adopted one cannot be systematically incorrect by more than about 50 K, given the small differences found in abundances derived from FeI and FeII. They also estimate that random errors, due to uncertainties in individual star colours and cluster reddenings, are of the same order of magnitude.

The average internal uncertainty in metal abundance on the CG97 scale is 0.06 dex, resulting in a precise ranking of cluster metallicities. CG97 also demonstrated that ZW's scale is clearly non-linear, in comparison to their improved scale. All [Fe/H] values on the ZW's scale were then translated to the CG97 scale by means of a quadratic interpolating relation, covering the range in [Fe/H] spanned by the 24 calibrating clusters ( [Fe/H] ).

In the present paper we recalibrate some of the most used metal abundance indicators to the new CG97 scale, concentrating on those based on the morphology and position of the red giant branch (RGB) in the colour-magnitude diagram (CMD). We will present calibrations for the indices , , 1.1 in the plane, and the analogous in the one, as defined in Sect. 3.

Such a revised calibration is needed, since the advent of sophisticated high resolution imaging facilities outside the atmosphere, like the Hubble Space Telescope, allows the observations of clusters in the whole Galaxy, providing CMDs with giant branches well defined also for very distant or obscured objects. Precise photometry of extragalactic clusters (in M31, in the Magellanic Clouds and in Fornax) is also feasible (see e.g. Fusi Pecci et al. 1996), and it is possible to derive for them quite accurate metallicities, provided that a good calibration from nearby clusters is available.

© European Southern Observatory (ESO) 1998

Online publication: December 16, 1997
helpdesk.link@springer.de