4.7. Spiral Bulges
Abundances in bulges generally are obtained the same way as for ellipticals, by spectroscopy and modeling of absorption line indices from the integrated stellar population, and so suffer from the same uncertainties. Most of the line indices in the Lick system vary in the same way with age and metallicity, and so poorly distinguish between the two in model grids. A few, such as H, Mg b, Fe5270, and Fe5335, do provide some ability to separate age and metallicity, and have been used to obtain estimates of ages, [Fe/H], and [Mg/Fe] in spheroidal systems (Worthey 1998).
Most work on abundances in bulges have come from spectroscopy of the central regions. Jablonka, Martin, & Arimoto (1996) found that Mg2 correlated with both bulge luminosity and stellar velocity dispersion in spirals with T = 0-5, but that the Fe5270 feature did not. Comparing with a grid of synthetic spectra with non-solar [/Fe] they inferred that [Mg/Fe] increased with bulge luminosity as well. Similar results are obtained for ellipticals (Worthey 1998). Maps of spectral line indices obtained with integral field units (Peletier et al. 1999, de Zeeuw et al. 2002) show hints that Mg/Fe varies with radius, although only a small number of galaxies have been analyzed so far.
High-resolution spectroscopy of red giants in our own Galactic bulge recommends caution in interpreting line indices in integrated spectra of ellipticals and bulges. Giants in the Baade's Window region show high /Fe ratios and a mean [Fe/H] -0.25, similar to [Fe/H] for solar neighborhood stars (McWilliam & Rich 1994). The mean metallicity is lower by about 0.3 dex compared to low-resolution spectroscopic and photometric determinations. This has several implications. Enhanced Ti/Fe ratios make the spectral types of the bulge giants later for the same IR colors compared to stars with solar abundance ratios. Enhanced /Fe alter both stellar line indices and the location of isochrones in population synthesis models, which have been computed using solar abundance ratios so far. Thus, metallicities derived for spheroids may be overestimated by a factor of two or so. Most population synthesis Trager et al. (2000a, b) attempt to correct the models for the effects of non-solar element abundances ratios. Such corrections are non-trivial, as both isochrones and line strengths are affected. No such corrections have been applied to bulges yet.
The formation of bulges is still mysterious. The candidate mechanisms are: monolithic collapse with rapid star formation (Eggen, Lynden-Bell, & Sandage 1962); mergers of roughly equal mass objects in hierarchical clustering models for galaxy formation (Baugh, Cole, & Frenk 1996: Kauffmann 1996); and secular growth from disk material, for example by mass transfer via bars (Combes et al. 1990; Hasan, Pfenniger, & Norman 1993). High /Fe (that is, higher than the solar ratio) would tend to favor the models with rapid bulge formation from relatively metal-poor gas, because most of the Fe is expected to come from Type Ia supernovae. Solar or less /Fe would tend to favor secular evolution from already enriched material, or star formation extended over times cales greater than 1 Gyr. The correlation of Mg/Fe with bulge luminosity and velocity dispersion suggests that a mix of formation mechanisms are at work; moreover, Andredakis et al. (1995) have found that bulge structure parameters correlate with Hubble type and bulge luminosity, such that large, luminous bulges tend toward R1/4 profiles similar to ellipticals, while small bulges tend to have exponential profiles (although de Jong 1996 argues that bulges have exponential profiles in general). The trends in Mg/Fe and bulge shape together suggest that large bulges formed rapidly at early times, while small bulges may have formed more slowly via secular processes.