Results and discussion

One way to determine the evolutionary paths of early-type galaxies is to study the relation between the metallicity gradients and other global properties of these systems, since different physical processes are expected to lead to different correlations. For example, dissipational processes are believed to create steeper gradients in more-massive galaxies (Larson 1974; Bekki & Shioya 1999), although this is sensitive to the feedback prescription adopted in the simulations (e.g. Bekki & Shioya 1999). Dissipationless mergers of galaxies, in contrarst, are expected to produce some dilution of the gradients in galaxies (White 1980), deleting or producing an inverse correlation among stellar population gradients and mass. Figure 41.1 shows the correlation of the metallicity with the central velocity dispersion (which is a proxy for virial mass) for our sample of galaxies. Although the sample is not very large, we confirm the lack of correlation previously noted by other authors using line-strength indices (e.g. Gorgas et al. 1990; Mehlert et al. 2003). It is not galaxies

,4472 |I600


4551 «3379

T *1700


m, i,

□ 4387 [TTL600


64 * 4555s4472






Figure 41.1. The relations between the metallicity gradients and the central velocity dispersion (left panel) and between the metallicity gradient and the metallicity (right panel).

with steeper gradients that are the most massive, but rather the ones with o ~ 120

kms-1 (corresponding roughly to galaxies with MB ---20.5, M ~ 1011 M0). In the figure we also distinguish between galaxies with different shapes of the inner profile. It can be seen there that power-law and core galaxies follow different trends in this plane. The trends could be explained by assuming that more-massive galaxies formed via mergers without dissipation (and that more-massive galaxies have suffered more mergers) whereas less-massive ones have grown by gas-rich mergers.

The right panel in Figure 41.1 shows the relation between the central metallicity and the metallicity gradient. We have also separated the galaxies as a function of their central ages. There exists a trend for galaxies with higher central metallicities also to have steeper metallicity gradients, although there is considerable scatter in the relation. We also find a relation between the metallicity gradient and the central age, whereby galaxies with a younger age in their centres also have steeper metallicity gradients.

Finally, we study the relations between the metallicity gradient and some structural parameters of the galaxies. Figure 41.2 shows the relations between the metallicity gradient and the rotational velocity, the anisotropy parameter (log(v/o)*), as defined by Bender (1990), and the (a4/a) x 100 parameter, which measures the deviation of the shape of the isophotes from a perfect ellipse. We could not find reference values of a4 for three of our galaxies (NGC 3384, NGC 4458 and NGC 4464) and, therefore, they are not included in Figure 41.2. A non-parametric Spearman rank-order test gives a probability lower than 0.5% that the correlation observed in the figures could have been produced by chance for the three relations, although, in the case of the relation with the anisotropy parameter, the correlation is mainly driven by two galaxies (NGC 1600 and NGC 2865). The correlation between grad[Z/H] and (a4/a) x 100 (disky galaxies have stronger gradients and












i i i i i i i i i

I 1 1 _


i 4387




" t 4458


" 3379





f §1700


a 3384





^865 1 ■ ■ -



_i I i i i i

1 1 1 1 I 1 1 1 1 I 1

1 1 J




I_ 11600








£55*1 i4464





■ 3384







I 1 i

£1600 ,4472



¿3379 ^700

— I I

¿3372 1 , -

Figure 41.2. Relations between the metallicity gradient and other structural parameters. Left panel: the relation between the metallicity gradient and the rotational velocity. Central panel: the relation between the metallicity gradient and the anisotropy parameter normalised with respect to an isotropic rotator. Right panel: the relation between the metallicity gradient and the (a^a) xlOO parameter, extracted from Bender et al, (1989).

NGC 1600 "


NGC 1600 "

NGC 2865


NGC 2865



Figure 41.3. Examples of the [E/Fe] profile from the sample of galaxies presented here. The lines represent a linear fit to the data. Open symbols represent the points excluded from the fit.

the gradient decreases with the boxiness of the isophotes) is somehow surprising, because the a4 parameter measured for simulated galaxies depends on the projection effects and, therefore, the same galaxy can have boxy and disky isophotes depending on the viewing angle (e.g. Stiavelli et al. 1991; Governato et al. 1993).

It has been suggested in the literature that the age spread in the stellar population of nearby galaxies found in several studies (e.g. Gonzalez 1993; Trager et al. 2000; Sanchez-Blazquez et al. 2006) could be a consequence of a small frosting of young stars in the centre of the galaxies, but not connected with the main mechanism of galaxy formation. However, the results presented here show that the central populations are related to the metallicity gradients, which are related to the structural parameters and, therefore, probably connected with the formation mechanism of the galaxy.

Bekki & Shioya (1998) performed numerical simulations of mergers between gas-rich galaxies, studying the effect of star formation on the structural parameters of the remnant. They found that the rapidity of gas consumption by star formation greatly affects the isophotal shape of the merger remnant. Mergers with gradual star formation are more likely to form elliptical galaxies with disky isophotes, whereas those for which the star formation is more rapid are more likely to form boxy ellipticals (although this depends on the viewing angle and, therefore, they can be seen as disky too). This scenario could explain the relation between the metallicity gradient and the shape of the isophotes found here.

The variations in chemical-abundance ratio with radius give information about the timescales of star formation within the galaxies.1 In the classical models of monolithic collapse stars form in essentially all regions during the collapse and remain in their orbits with little inward migration, whereas the gas, being continuously enriched by the evolving stars, dissipates inwards. In these scenarios the star formation is halted by the galactic winds, which initiated beforehand in the external parts of the galaxies due to their shallower potential well. This mechanism creates positive [a/Fe] gradients. The resultant [a/Fe] gradient of a merger remnant is more difficult to predict, since it depends on the initial gradient and gas fraction of the interacting galaxies and also on whether star formation occurs.

Figure 41.3 shows some examples of [E/Fe] for our sample of galaxies. As can be seen, there are both positive and negative cases. This rules out a simple inside-out or outside-in mechanism for all the galaxies. Furthermore, they are considerably shallower than the relationship predicted by simple models of gaseous collapse (e.g. Pipino et al. 2006).

Was this article helpful?

0 0

Post a comment