4.2.1 Chemical Abundances of Local Group dE/dSph Galaxies
Aaronson (1986) showed that dEs seemed to follow a tight relation between metallicity ([Fe/H]) and optical luminosity. This has later been confirmed and a rather tight luminosity metallicity relation for dwarf ellipticals is now quite well established (Caldwell et al. 1998). Metallicity data for dEs based on CMD analysis, can be found at several places in the literature (see Mateo 1998 for a recent compilation). These data also reveal that most dEs have a considerable spread in metallicity as measured from the width of the RGB.
Stellar spectroscopic abundances exist for several of the Milky Way satellites, e.g. the Sextans dSph (Suntzeff et al. 1993). Another well studied galaxy is Draco, where Shetrone et al. (1998) found -3.0 < [Fe/H] -1.5 confirming earlier findings of a substantial metallicity spread (Lehnert et al. 1992). The metallicity derived from a CMD is [Fe/H] = -2.0 ± 0.15.
Richer and McCall
(1995)
made a spectroscopic study of planetary
nebulae (PNe) in local dEs, including data from the literature on the
Fornax dwarf.
They also revisit the metallicity-luminosity relation for dwarf irregulars
by Skillman et
al. (1989),
using new determinations of the distance and metallicity.
Their PN spectra have rather low signal to noise requiring empirical relations
between [O III]
5007 and
H
to estimate lower
bounds on the oxygen abundances
in the three dEs, from which they make a transformation
to a ``mean oxygen abundance''. They claim these to be systematically
higher in dEs than in dIs of the same luminosity. Although based on a sound
line of reasoning their derivation of the ISM mean oxygen abundance
involves several steps, and must be regarded as quite uncertain. Thus, while
tentatively very important, these results need to be quantitatively
confirmed for a
larger sample of PNe in more galaxies, with higher S/N and better spectral
coverage. Recently, PNe were studied also in the Sagittarius dwarf
(Walsh et al. 1997).
Richer and McCall (1995) also claim that [O/Fe] is systematically higher in dEs than dIs, where the latter class was represented by the Magellanic Clouds. The Fe abundance adopted for LMC and SMC are based on young supergiants, and thus do not measure the same stellar generation as in the dEs. To compensate for this, they modify the abundance ratios.
If one instead compares the Magellanic Cloud O-abundances with [Fe/H] estimates from field stars with ages comparable to those of the red giants in the dEs (e.g. Hilker et al. 1995) the Magellanic Cloud [O/Fe] values increase and thus the difference compared to the dEs decreases. In Fig. 1 we plot [Fe/H] vs. [O/H] for all nearby dwarfs with both elements measured. It is clear that the overabundance over oxygen with respect to iron seems to be a general feature of local dwarfs (see also Fig. 7 in Mateo 1998) as it is for giant halo stars (Barbuy 1988) and metal-poor unevolved halo stars in our Galaxy (Israelian et al. 1998). There is no indication that [O/Fe] is significantly higher in dEs than dIs. In a later work, Richer et al. (1998) find [O/Fe] to be systematically higher in dEs than e.g. in M32 and the Galactic bulge, but similar to the galactic halo, and they suggest that the star formation timescale has been shorter in dEs than ellipticals and spiral bulges.
|
Figure 1. [Fe/H] vs. 12 + log(O/H) for all local dwarf galaxies where both iron and oxygen abundances have been obtained (data taken from Mateo 1998), plus the Magellanic clouds, a BCG, the Sun and the Milky Way halo. dIs are denoted by filled squares, dI/dE intermediate type (Pegasus) by a cross, and dE (Sagittarius) by an asterisk; stars are the dEs studied by Richer & McCall with oxygen abundances from planetary nebulae. The triangles joined by dashed line represent the Magellanic clouds, where for the open symbols the [Fe/H] are based on young supergiants, while for the filled symbols the [Fe/H] have been determined from Strömgren photometry of the field population. The open diamond indicates the location of the nearby BCG VII Zw403 (see Sect. 4.4.7). The location of Milky Way halo stars is indicated with ``MW'' (from Richer and McCall 1995), and that of the Sun with ``Sun''. The dotted line corresponds to solar [O/Fe]. In general (except for the open triangles) Fe/H traces an older stellar population than O/H to some extent explaining the apparent general deficiency of iron, with respect to oxygen. The typical errors are around 0.2 dex for both quantities, as illustrated by the error bars in the lower right corner. |
Now, how metal-poor are the Local Group dEs? As was discussed above,
[Fe/H] measured
from CMDs tends to be systematically smaller than [O/H] measured from
H II regions with
up to 0.5 dex, making these two quantities difficult to compare.
Of the Local Group dEs, a dozen have [Fe/H] 
-1.5, and the lowest value, [Fe/H] = -2.2,
is found for the Ursa Minor dSph, see Table 2. In
summary, many Local Group dEs are very metal
poor, which probably is related to their intrinsic faintness. Taking the
average offset between [Fe/H] and [O/H] into account,
Mateo (1998)
concludes that dEs appear to be more metal rich in view of their luminosities,
in agreement with
Richer and McCall
(1995).
| Galaxy name | [Fe/H] | 12 + log(O/H) | MV | log(
|
| Ursa Minor | -2.2 | -8.9 | 7.36 | |
| Draco | -2.0 | -8.8 | 7.34 | |
| Carina | -2.0 | -9.3 | 7.11 | |
| Andromeda III | -2.0 | -10.3 | ||
| Leo II | -1.9 | -9.6 | 6.99 | |
| Phoenix | -1.9 | -10.1 | 7.52 | |
| LGS 3 | -1.8 | -10.5 | 7.11 | |
| Antlia | -1.8 | -10.8 | 7.08 | |
| Sculptor | -1.8 | -11.1 | 6.81 | |
| Sextans dSph | -1.7 | -9.5 | 7.28 | |
| Tucana | -1.7 | -9.6 | ||
| Andromeda II | -1.6 | -11.1 | ||
| Andromeda I | -1.5 | -11.9 | ||
| Leo I | -1.5 | -11.9 | 7.34 | |
| Fornax | -1.3 | 8.0 | -13.2 | 7.83 |
| NGC 185 | -1.2 | 8.2 | -15.5 | 8.11 |
| NGC 147 | -1.1 | -15.5 | 8.04 | |
| M32 | -1.1 | -16.7 | 9.33 | |
| Sagittarius | -1.0 | 8.3 | -13.4 | |