Next Contents Previous

4.2. Enhanced Light-to-Heavy-Element Ratios

The light-element 6 enhancement can be seen in the case of [Mg/Fe] by plotting a magnesium feature index (Mg2) versus an average iron feature (<Fe> is the arithmetic average of indices Fe5270 and Fe5335). The age sensitivities of these indices are about the same (approx 3/2!), so models of different ages and metallicities should lie on top of one another. They do, as seen in Figure 12, but at high Mg strength the galaxies follow another distinct trajectory entirely with some galaxies trending toward strong Mg2 strength at nearly constant <Fe> and hence, we infer, relatively enhanced Mg abundance.

Figure 12

Figure 12. Models and galaxies are shown in the <Fe> vs. Mg2 diagram. <Fe> is the arithmetic average of the Fe5270 and Fe5335 indices. Each dotted line connects single-age models of [Z/H] = -0.25, 0.0, and 0.25 dex, lower left to upper right in order. The three lines are (from weak to strong Mg2 index strength) for ages 2, 9, and 18 Gyr. Composite combinations of age and metallicity add approximately as vectors. The models are the same in the three panels. Spiral nucleus/bulge data appear in the top panel, S0 galaxy data in the middle panel, and elliptical galaxy data in the bottom. Solid lines trailing from the nuclear data points are trends with galaxy radius in the studies that include radial gradient information. The Galactic bulge (BW is "Baade's window") is included in the top panel. The tendency of most elliptical nuclei to lie to the right (to higher Mg2 strength) of the model lines is interpreted as a real overabundance of Mg compared with Fe. Spiral bulges show almost no [Mg/Fe] enhancement, but Es show a substantial amount, with S0 galaxies intermediate. Gradients tend to align with the model age-metallicity direction rather than the relation linking nuclei, suggesting that [Mg/Fe] is more than a nucleus-only phenomenon. The references are Proctor, Sansom, & Reid (1999), Jablonka, Martin, & Arimoto (1996), Vazdekis et al. (1996), Deslisle (1998), Carollo & Danziger (1994a, 1994b), "FFI" is Fisher et al. (1996), Davies, Sadler, & Peletier (1993), and "Lick" is Trager et al. (1998).

The models are scaled-solar since they are built from local stars, so they cannot track altered abundances. Composite populations add approximately like vectors, so any combination of ages and metallicities still lands on the same model locus. Different models built by different authors have a spread of something like ± 0.5 dex at constant index strength, but all models follow almost exactly the same slope in the Figure 12 diagrams, so that the inferred [Mg/Fe] for the high-Mg2 group of elliptical galaxies is in the range [Mg/Fe] = 0.3-0.5 dex.

A crucial thing to notice is that velocity dispersion tracks Mg2 very tightly, so the high Mg2 galaxies are also the largest galaxies (or, more precisely, the "dynamically hottest"). The Mg2-sigma relation is shown in Figure 13; it is one of the tighter scaling relations known, much tighter than, say, the <Fe>-sigma diagram, which is almost a scatter plot. With the existence of the Mg-sigma relation, [Mg/Fe] increases with galaxy size, where a velocity dispersion of sigma approx 200 km s-1 seems to mark the beginning of noticeable Mg enhancement.

Figure 13

Figure 13. The <Mg2>-sigma relation, data from Faber et al. (1989) and Bender et al. (1993). A regression line of <Mg2> = -0.166 + 0.20 log sigma determined by Bender et al. (1993) is shown. Note that the <Mg2> index differs from the standard definition (found in Worthey et al. 1994) of Mg2. See Trager et al. (1998) to convert between the two systems. This relatively tight scaling relation is useful in galaxy evolution studies because younger-aged galaxies drift toward weaker Mg2 at constant velocity dispersion.

Figure 12 shows separate diagrams for spiral bulges, S0 galaxies, and elliptical galaxies. No marked difference between Hubble types is seen except that already ascribed to velocity dispersion. That is, spiral bulges hover near the solar ratio area, but only two bulges have sigma > 200 km s-1 (and those are on the high-Mg side of the distribution). Elliptical galaxies possess both the most extreme velocity dispersions and the most extreme Mg enhancement.

The gradient vectors shown in Figure 12 tend to parallel the age-metallicity direction traced by the various models rather than the more horizontal slope defined by the nuclei. This would suggest that the Mg enhancement is global throughout the galaxy rather than concentrated only at the nucleus. Is this a hint that the enrichment mechanism (presumably supernova) spreads enriched gas 10 or 20 kpc from its origin, or does it merely imply effective mixing?

The [Mg/Fe] data suggest a variation in enrichment from Type Ia (mostly Fe) and Type II (all elements) supernovae passing from small galaxies or bulges to large ones, in the sense that the large galaxies have more Mg and hence comparatively more Type II enrichment. Figure 14 shows some corroborating evidence from Trager et al. (1998) nuclear data in that Ca appears to track Fe, while Na and N are enhanced in a way similar to that of Mg; only in the larger galaxies or bulges.

Figure 14

Figure 14. Lick index models and E+S0 galaxy nuclei are shown with model lines of ages 9 and 18 Gyr, connecting [Z/H] = -0.5, -0.25, 0.0, 0.25, and 0.5 dex. In the cases of CN1, Mg2, and Na D, large elliptical galaxies (solid symbols) lie significantly off the model sequence, indicating light-element enhancement. Ca4455, on the other hand, tracks Fe in a scaled-solar fashion (Worthey 1998).

The mechanism for modulating Type Ia/Type II enrichment is not known. It could be due to a time delay in Type Ia metal production, or could be some other connection to velocity dispersion like a variable IMF or a variable fraction of binary stars. (Worthey, Faber, & González 1992; Weiss, Peletier, & Matteucci 1995).

6 In this subsection we make a distinction between "light" and "heavy" elements, divided at the fourth row of the periodic table, so that Ca and Fe are heavy, but Na, Mg, and N are light. The oft-standard terminology is to speak of "alpha" elements, but "alpha" usually includes Ca and excludes N, which makes little sense for the abundance pattern seen in massive elliptical galaxies. Back.

Next Contents Previous