Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 345, 419-429 (1999)

Previous Section Next Section Title Page Table of Contents

5. The blue indices diagnostic

In this section we compare the results of our simulations with the data for real galaxies making use of two diagnostic planes. i.e. H+K(CaII) vs. H[FORMULA]/FeI and H+K(CaII) vs. [FORMULA]4000, displayed in Figs. 5 and 6, respectively.

[FIGURE] Fig. 5. H+K(CaII) vs. H[FORMULA]/FeI diagram for shell galaxies (top panel ) and pair members (bottom panel ). The dashed lines represent models of post-star-burst galaxies with [FORMULA]. Each line is labeled by the burst age. See the text for more details. In each panel the average error-bars are shown

[FIGURE] Fig. 6. Same as Fig. 5 but for the H+K(CaII) and [FORMULA]4000 indices

The galaxy data are from the catalog of Longhetti et al. (1998a) further selected by choosing only those objects with index H+K(CaII) measured with less than 0.25 units of error. Due to the low S/N in the blue part of the spectrum of some galaxies, only 18 shell galaxies and 16 pair members pass the scrutiny. The basic data for this sub-sample of objects is presented in Table 4.


Table 4. Sample of shell and pair galaxies with best measured H+K(CaII), [FORMULA]4000 and H[FORMULA]/FeI

The models (dashed lines) superposed to data in Figs. 5 and 6 refer to the solar metallicity and show the expected values for the three indices in post-star-burst galaxies. All lines start from a common origin which is the locus in these diagrams of the host quiescent galaxy (15 Gyr old and solar metallicity). The indices of this reference galaxy are H+K(CaII)[FORMULA]1.21, H[FORMULA]/FeI[FORMULA]1.12 and [FORMULA]4000[FORMULA]2.32. Each line refers to a specific age of the burst event as annotated along the curves. Moving from the common origin, along each line what is varying is the percentage of mass turned into stars during the burst event. The percentage increases moving from the origin to the blue end of each curve. This in fact represents the case of a 100% young galaxy with the age indicated.

In these diagrams, an old quiescent galaxy suffering from a burst of star formation would soon leave the original place and move toward the blue corner up to a maximum location dictated by the percentage of newly born stars. As soon as the burst is over and the new stellar component start aging, the galaxy goes back to it original place. Once about 3 Gyr have elapsed, the current position is virtually indistinguishable from the original one (see also the run of the three indices as a function of the age in Fig. 3).

The remarkable novelty showing up both in Figs. 5 and 6 is the different distribution of the shell and pair galaxies. In fact, while shell galaxies fall in the region covered by the theoretical models, pair galaxies have a much broader distribution.

The large error bars affecting the data (in particular the index H[FORMULA]/FeI) do not allow us to derive any specific conclusion about the age and intensity of the secondary episode of star formation for individual galaxies. The diagnostic diagrams have to be interpreted only in a statistical sense .

Recalling that old galaxies suffering from a secondary burst of star formation perform the loop in the diagnostic diagrams away from the quiescence position on a rather short time scale (which in turn depends on the amount of mass engaged in the burst), the different behavior of shell and pair galaxies can be explained if the bursting episode in the former is statistically older than in the latter.

This is also somehow hinted by the large value of H+K(CaII), larger than 1.2, that some pair galaxies reach. Such high values of H+K(CaII) are expected only if the index is contaminated by the H[FORMULA] emission line (see below for more details). Therefore, the distribution of pair galaxies in the two diagnostic planes is statistically indicative of ongoing star formation . In contrast, the distribution of shell galaxies seem to suggest that the secondary star forming episode is on the average older than 1 Gyr .

We have checked that the above conclusions are not affected by the age adopted for the host galaxy at quiescence. Indeed similar results hold assuming for this object ages of 12 and 18 Gyr.

5.1. Effect of line emission

In order to support our preliminary suggestion that values of H+K(CaII) higher than 1.2 are indicative of contamination by the H[FORMULA] emission line and hence ongoing star formation, we have analyzed the spectroscopic atlas by Kennicutt (1992).

It contains 55 local galaxies, well distributed among the different morphological types. Table 5 reports the results of our measurements of the H+K(CaII) and [FORMULA]4000 indices for this sample. All peculiar and irregular galaxies have been excluded from the analysis.


Table 5. Data from the atlas of Kennicutt (1992)

The resolution of the Kennicutt (1992) spectra is between 5 Å and 8 Å FWHM, slightly better than the one in usage here. While [FORMULA]4000 is not expected to depend on the spectral resolution and/or velocity dispersion, H+K(CaII) tends to increase as the spectral lines get wider. However, the expected increase in H+K(CaII), passing from the spectral resolution of Kennicutt (1992) data to ours, is very small and within statistical errors.

From the entries of Table 5, the mean value of the index H+K(CaII) for early-type galaxies (E+S0) is 1.19 ([FORMULA] 0.16). In galaxies of later types the index gets smaller. H+K(CaII) indeed decreases in stellar systems richer in A0-type stars, i.e. stellar systems with active star formation. Interestingly enough, the value of H+K(CaII) in post-star-burst models with solar (Figs. 5 and 6) and non solar metallicities (not shown here for the sake of brevity) agrees with the maximum observational value [FORMULA].

In contrast, the bottom panels of Figs. 5 and 6 show some pair members (RR210b, RR381a, RR387a, RR387b and RR397b) characterized by values of H+K(CaII)[FORMULA] 1.3. This means that the strength of the blend of H(CaII) with the Balmer line H[FORMULA] is lower than expected.

The only possible explanation of this fact is to consider the contamination of the H[FORMULA] absorption line by the corresponding emission line.

It is worth noticing that none of the above 5 galaxies shows a clear H[FORMULA] emission line. Although the H[FORMULA] emission line is expected to have a lower absolute intensity with respect to that of the H[FORMULA] line, its influence on the H+K(CaII) index could be greater than that of H[FORMULA] on the corresponding absorption feature. Then a small variation in the intensity of the H[FORMULA] absorption line caused by the corresponding emission component could be more easily revealed in the H+K(CaII) index, thus making it a good indicator of star formation.

5.2. Effects of the metallicity

We have also calculated two grids of post-star-burst models with metallicities different from the solar value. The results are shown in Fig. 7. As explained in Sect. 3, models of non solar metallicity represent only a first order approximation because of the incomplete dependence of our simulations on this parameter (only in the isochrones and not in the library of stellar spectra). The new models are, however, much similar to the case of solar metallicity presented above, even though they differ in some details.

[FIGURE] Fig. 7. Effects of metallicity on the diagnostic diagrams H+K(CaII) vs. H[FORMULA]/FeI and H+K(CaII) vs. [FORMULA]4000. The dotted and solid lines are for [FORMULA] and [FORMULA], respectively.

(i) The reference galaxy of 15 Gyr age and the loci of the bursting galaxies shift to bluer and redder indices passing from solar to [FORMULA] and [FORMULA], respectively. In more detail, for [FORMULA] the reference galaxy has indices H+K(CaII)[FORMULA]1.17, H[FORMULA]/FeI[FORMULA]0.82 and [FORMULA]4000[FORMULA]2.05, whereas for [FORMULA] it has H+K(CaII)[FORMULA]1.23, H[FORMULA]/FeI[FORMULA]1.18 and [FORMULA]4000[FORMULA]2.55. All the paths of the bursting galaxies are accordingly shifted.

(ii) In the H+K(CaII) vs. H[FORMULA]/FeI plane, different metallicities produce almost the same loci for the post-star-burst galaxies.

(iii) A significantly larger variation is seen in the H+K(CaII) vs. [FORMULA]4000 plane.

(iv) As compared to the observational data, the [FORMULA] models are too blue, whereas the [FORMULA] ones are too red but still marginally compatible.

Given the adopted approximation and the uncertainty in the data (see the error-bars) we cannot contrive the mean metallicity of galaxies from their location in the diagnostic planes H+K(CaII) vs. H[FORMULA]/FeI and H+K(CaII) vs. [FORMULA]4000. Nevertheless, the results shown in Figs. 5, 6 and 7 hint that the solar metallicity is perhaps suited to represent the mean value in early-type galaxies. Similar conclusions have been reached by Bressan et al. (1996), Greggio (1996) and Longhetti et al. (1998c) analyzing the position of early-type galaxies in the H[FORMULA] vs. [MgFe] diagram.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1999

Online publication: April 19, 1999