The Trager et al. Sample: Intermediate ages and $\alpha$-Enhancement in Giant Early-Type Galaxies

We start by displaying in Figure 27 the data from Trager et al. (2000) for nearby early-type galaxies on top of our model grids in the $<Fe>$ vs. $H\beta$ and Mg $b$ vs $H\beta$ planes. In the lower panels we compare the data with solar-scaled models, computed adopting the solar-scaled Padova isochrones. Same-[Fe/H] lines are labeled according to the abundances of iron, [Fe/H], (left panel) and magnesium, [Mg/H], (right panel), which of course are the same for these [Mg/Fe]=0 models. Model ages are indicated in the lower left panel and are the same on all the other panels. As discussed in previous sections, both $<Fe>$ and $H\beta$ are very insensitive to abundance ratios, so they can be used to estimate the mean [Fe/H] and age of the stellar populations in galaxies. Therefore, according to the lower left panel of Figure 27, the bulk ($\sim$ 3/4) of the Trager et al. sample has roughly solar [Fe/H] and ages between $\sim$ 7 and 14 Gyr. The remaining 1/4 of the sample has mean ages lower than $\sim$ 7 Gyr, with some galaxies reaching ages of the order of 2.5 Gyr. We focus here on the 3/4 of the sample with ages older than $\sim$ 7 Gyr. In the lower right panel, the same models are compared with data on the Mg $b$-$H\beta$ plane. In this panel, we can see that the same models that match $<Fe>$ data under-predict Mg $b$ by $\sim$ 1 ${\rm\AA}$. This is a well-known result, which is telling us that stars in giant early-type galaxies are magnesium-enhanced ([Mg/Fe] $>$ 0). In the top panels we compare the same data with models computed adopting [Mg/Fe]=+0.3 and keeping all other abundance ratios solar. These models are computed adopting the solar-scaled Padova isochrones (see Section 6.1.1). Note that the models plotted in the upper panels have the same [Fe/H] values as in the lower panels, but the values for [Mg/H] are +0.3 dex higher in the upper panels. The plots in the upper panels show that the [Mg/Fe]=+0.3 models match the $<Fe>$ and $H\beta$ data for the same ages and [Fe/H] as the solar-scaled models, with the oldest 3/4 of the sample having roughly solar [Fe/H]. On the other hand, the Mg-enhanced models with [Fe/H]=0 are a much better match to the Mg $b$ data, indicating that the old galaxies in the Trager et al. sample have mean [Mg/Fe] ([Mg/H]) of the order of +0.3 (+0.3). Therefore, we reproduce the results by Worthey et al. (1992), who found that giant early-type galaxies have [Mg/Fe] higher than solar. The models employed in Figure 27 are presented in Tables A through A in the Appendix.

We also confirm previous results indicating that early-type galaxies have a large spread in mean ages, hinting at the presence of an intermediate age component in their stellar populations (e.g., Trager et al. 2000, Kuntschner 2000, Caldwell, Rose & Concannon 2003, Denicoló et al. 2005, Thomas et al. 2005, Mendes de Oliveira et al. 2005). We note that a pattern of younger galaxies having higher [Fe/H] and lower [Mg/Fe] is also found, which is also in agreement with the findings by previous authors. The latter result is more clearly seen in Figure 28 where models and data are compared in $<Fe>$-Mg $b$ space. In the left panel, all galaxies are over-plotted on solar-scaled and [Mg/Fe]=+0.3 models for 8 Gyr and older models. Galaxies with ages younger than $\sim$ 7 Gyr ($H\beta > 2$ in Figure 27) are plotted with open squares, while older galaxies are represented by solid symbols. There is a clear trend in the sense that younger galaxies tend to be closer to the solar-scaled models, whereas older galaxies lie closer to the Mg-enhanced lines. In the right panel, only younger galaxies are compared with model predictions for single stellar populations with comparable ages, showing that the trend is confirmed.

Finally, we point out that there are no galaxies in Figure 27 with mean stellar ages older than $\sim$ 14 Gyr, which means no galaxies older than the universe (Spergel et al. 2003), demonstrating that stellar population synthesis models are reaching a state of maturity whereby mean stellar ages obtained from comparison with high quality data are meaningful in an absolute sense.