6.2. Luminosity and Metallicity
Mean galaxy metallicity and mean galaxy luminosity are well correlated, as has been known for a long time. For dIrrs, one usually considers present-day oxygen abundances, and when comparing them to other galaxy types without ionized gas (such as dSphs), stellar [Fe/H] values are converted into what is assumed to be the corresponding nebular abundance. This conversion comes with a number of uncertainties. Grebel et al. (2003) therefore used the stellar (red giant) metallicities of Local Group dwarf galaxies of all types to directly compare the properties of their old populations. A plot of V-band luminosity LV versus <[Fe/H]> (Fig. 5, left panel) shows a clear trend of increasing luminosity with increasing mean red giant branch metallicity. However, different galaxy types (gas-rich dIrrs and gas-deficient dSphs) are offset from each other in that the dIrrs are more luminous at the same metallicity.
Figure 5. V-band luminosity (left panel) and baryonic luminosity (right panel, corrected for baryon contribution of gas not yet turned into stars) versus mean metallicity of red giants. The symbols are the same as in Fig. 4. The error bars in metallicity indicate the metallicity spread in the old populations, not the uncertainty of the metallicity. DIrrs are more luminous at equal metallicity than dSphs. Or, in other words, dSphs are too metal rich for their low luminosity. However, several dIrr/dSph transition-type galaxies coincide with the dSph locus. These objects are indistinguishable from dSphs in all their properties except for gas content. (Figure from Grebel et al. 2003.)
In other words, the dIrrs have too low a metallicity for their luminosity as compared to dSphs. Thus, dSphs, most of which have been quiescent over at least the past few Gyr, must have experienced chemical enrichment faster and more efficiently than dIrrs, which continue to form stars until the present day (Grebel et al. 2003). It is tempting to speculate that environment may once again have affected this in the sense of a denser environment leading to more vigorous early star formation rates.
When plotting the baryonic luminosity (Milgrom & Braun 1988; Matthews, van Driel, & Gallagher 1998) against metallicity (Fig. 5, right panel), the locus of the dSphs remains unchanged while the dIrrs move to higher luminosities as compared to the dSphs. Thus, if star formation in present-day dIrrs were terminated when all of their gas was converted into stars, then these fading dIrrs would be even further from the dSph luminosity-metallicity relation. For a discussion of the amount of fading, time scales, angular momentum loss, etc. required for converting a dIrr into a dSph, see Grebel et al. (2003). Here we simply want to emphasize that dIrrs follow a metallicity-luminosity relation that requires a different evolutionary path than in other types of dwarf galaxies. In particular, it seems that dIrrs are an intrinsically different type of galaxy than dSphs. We note in passing that for dIrrs not only do metallicities correlate well with luminosities, but also with surface brightness.