5.2. The Metallicity-Luminosity Relationship

As noted before, Garnett & Shields (1987) showed that spiral disk abundances correlated very well with galaxy luminosity, which was later confirmed by VCE and ZKH. Figure 17 shows both the central oxygen abundance extrapolated to zero radius from the abundance gradient, and the characteristic mean abundance at one disk scale length for the sample of well studied spirals assembled by Garnett et al. (1997b). This metallicity-luminosity correlation extrapolates smoothly to include low luminosity irregular galaxies all the way down to M_B -10 (Skillman, Kennicutt, & Hodge 1989, ZKH, see Section 4.2). This remarkable correlation tells us that galaxy gravitational potential plays a key role in determining the subsequent evolution of a system. The characteristic metallicity of a galaxy shows a mild correlation with Hubble type, reflecting the tendency for early T-types to be more massive than late T-types (ZKH).

Figure 17. Fiducial interstellar O/H versus galaxy luminosity, for selected spiral galaxies. The upper panel shows the central abundance extrapolated from the abundance gradient; lower panel shows the characteristic O/H at one disk scale length from the nucleus. (From Garnett et al. 1997b).

Is there more to the luminosity-metallicity relationship? McCall (1982) noted a correlation between the stellar mass surface density, (_d), and oxygen abundance in spiral galaxies. Edmunds & Pagel (1984) reiterated this point and showed an impressive correlation between _d and O/H for late type (Scd) spiral galaxies. Garnett et al. (1997b) compared O/H vs. stellar mass surface density for both NGC 2403 and M33 found their gas abundances follow the same metallicity-surface density correlation as seen in other Scd spirals by Edmunds & Pagel. This may reflect fundamental similarities in the evolution of these galaxies. For example, Ryder (1995) argued that the surface brightness-metallicity correlation can be understood in terms of a galaxy evolution model which includes self-regulating star formation (see also Phillips & Edmunds 1991). In these models energy input from newly formed stars as well as older stars in the galaxy is assumed to feed energy back into the interstellar medium so as to inhibit further star formation. Such models appear to reproduce the correlation between current star formation rate and surface brightness (Dopita & Ryder 1994), and between metallicity and surface mass density (Phillips & Edmunds 1991; Ryder 1995), better than models based on a simple Schmidt (1959) law for star formation.

Figure 18. Characteristic abundance at a given value of surface brightness versus galaxy luminosity. Filled circles represent galaxies observed by Ryder (1995), with I-band surface brightness measurements. Unfilled circles are galaxies observed by ZKH, with photograph F-band surface brightness data. All data except for NGC 2403 and M33 were taken from Ryder (1995). (From Garnett et al. 1997b).

Is this relation universal? To test this, Don compared the abundances in different galaxies at similar disk optical surface brightness. The results are shown in Figure 18. NGC 2403 and M33 have been added to galaxies studied by Ryder and ZKH. The result is rather remarkable. There is a clear correlation between the abundance at a given surface brightness and the total luminosity of the galaxy. Figure 18 suggests that massive spirals are more efficient at enriching their ISM at fixed surface brightness than low mass spirals. Note however, that the sample is small, and probably contains a limited range in overall galaxy surface brightness. Nonetheless, I find this result very exciting and hope that there will be more follow-up work. This diagram must be telling us something fundamental about galaxy evolution. Perhaps we are closer to defining a fundamental set of parameters which govern spiral galaxy evolution (e.g., Mollá, Ferini, & Díaz 1996, and references therein).