Star formation history in early-type galaxies - I. The line absorption indices diagnostics

To unravel the formation mechanism and the evolutionary history of elliptical galaxies (EGs) is one of the goals of modem astrophysics. In a simplified picture of the issue, the question to be answered is whether they have formed by hierarchical merging of pre-existing substructures (maybe disc galaxies) made of stars and gas, with each merging event probably accompanied by strong star formation, or conversely, whether they originated from the early aggregation of lumps of gas turned into stars in the remote past via a burst-like episode ever since followed by quiescence so as to mimic a sort of monolithic process. Even if the two alternatives seem to oppose each other, actually they may both contribute to shaping the final properties of EGs as seen today. Are there distinct signatures of the underlying dominant process in the observational data? To this aim we have examined the line absorption indices on the Lick system of the normal, field EGs of Trager and the interacting EGs (pair- and shell-objects) of Longhetti et al. The data show that both normal, field and interacting galaxies have the same scattered but smooth distribution in the Hbeta versus [MgFe] plane even if the interacting ones show a more pronounced tail toward high Hbeta values. This may suggest that a common physical cause is at the origin of their distribution. There are two straightforward interpretations of increasing complexity. (i) EGs span true large ranges of ages and metallicities. A young age is the signature of the aggregation mechanism, each event accompanied by metal enrichment. This simple scheme cannot, however, explain other spectro-photometric properties of EGs and has to be discarded. (ii) The bulk population of stars is old but subsequent episodes of star formation scatter the EGs in the diagnostic planes. However, this scheme would predict an outstanding clump at low Hbeta values, contrary to what is observed. The model can be cured by supposing that the primary star formation activity lasted for a significant fraction of the Hubble titne (5 less than or equal to T less than or equal to 13 Gyr) accompanied by global metal enrichment. The 'younger' galaxies are more metal-rich. The later burst of star formation should be small otherwise too many high-Hbeta objects would be observed. Therefore, the distribution of normal, pair- and shell-galaxies in the Hbeta versus [MgFe] plane is due to global metal enrichment. Even though the above schemes provide a formal explanation, they seem to be too demanding because of the many ad hoc ingredients that have to be introduced. Furthermore, they neglect the observationally grounded hint that the stellar content of EGs is likely to be enhanced in alpha-elements with [alpha/Fe] ranging from 0.1 to 0.4 dex. Here we propose a new scheme, in which the bulk dispersion of galaxies in the Hbeta versus [MgFe] plane is caused by a different mean degree of enhancement. In this model, neither the large age ranges nor the universal enrichment law for the old component are required and the observed distribution along Hbeta is naturally recovered. Furthermore, later bursts of stellar activity are a rare event, involving only those galaxies with very high Hbeta (roughly >2.5). Finally, simulations of the scatter in broad-band colours of EGs seem to confirm that the bulk stars have formed in the remote past, and that mergers and companion star formation