Summary and conclusions

In this contribution, we have confirmed that starburst galaxies can be effectively detected and discriminated from normal galaxies and AGNs using criteria based on their FIR emission.

In comparison with normal galaxies, the starbursts typically show an excess of FIR luminosity and distinct colors. These differences are explained by a higher quantity of hot dust due to the higher star formation rates in starbursts as compared to normal galaxies. In the FIR, the starbursts form a continuous sequence with the normal galaxies, but can be distinguished based on a color alpha(60,25) < -2.5. This color cut-off also marks a variation in the dominant morphological types of the galaxies: normal IRAS galaxies are mostly late-type (Sb and later) spirals, while starbursts are more numerous among the early-type ones (Sb and earlier). This result is consistent with the observations made by Devereux & Hameed (1997), and suggests that many early-type spiral galaxies are still actively forming stars. However, this preference of massive starburst for early-type morphologies is not new, but a trait of SBNGs (Coziol et al. 1997a).

Like for other samples of SBNGs, no preference for barred galaxies is found among the PDS starbursts. The role of the bar in starbursts is not well established. The fact that the number of barred starbursts increases towards the late-type morphologies may suggest that the bar is important only in the late-type starbursts. Alternatively, the bar could also have a shorter life time in a galaxy with a stronger bulge.

The PDS starbursts are mostly massive (M_B = -20.0 +- 1.0). This result is consistent with the fact that the small-mass H II galaxies are relatively deficient in dust. In the H II galaxies the dust may have been swept away by starburst winds, or these galaxies may be too young to have produced enough dust. The FIR colors of the H II galaxies suggest that their dust temperatures are higher than those of the SBNGs. This phenomenon probably has something to do with the low metallicities of these galaxies, but also suggests a larger number of young stars than in the SBNGs and, consequently, younger age for the bursts in the H II galaxies than in the SBNGs.

In general, there seems to be no difference between the FIR characteristics of the UV-bright starbursts and those selected in the FIR. The PDS starbursts simply correspond to the FIR luminous branch of the UV-bright SBNGs with a mean FIR luminosity log(L_IR/L_solar) = 10.3 +- 0.5 and redshift z <0.1. The PDS starbursts, in particular, are limited in flux in the FIR at 10^-10 erg/(cm^2 s) while the UV-bright starbursts are limited in flux at 10^-11 erg/(cm^2 s).

Another interesting result of our analysis is the very few starbursts detected in X-rays. Only 4% of the 200 PDS starbursts were detected by ROSAT. Furthermore, the nature of 3 of these X-ray starbursts for which we obtained a spectrum was found to be ambiguous, the galaxies showing spectral characteristics intermediate between those of starbursts and LINERs. This result, and the fact that a very high number of Sy1 (71%), but few Sy2 (22%) were detected in X-rays, cautions against the utilization of the FIR, without other means of discrimination, to select starburst galaxies in order to study their X-ray properties. Our observations suggest, instead, that the contribution by starbursts to the cosmic X-ray background, for example, could be negligible (Hasinger 1998).

The relatively high fraction (38%) of AGNs in our sample is consistent with observations which suggest that the probability of finding an AGN increases with the FIR luminosity (de Grijp et al. 1987; Veilleux et al. 1995). In our sample, 62% of the AGNs are Sy2, 34% are Sy1 and the remaining 4% are LINERs. The higher number of Sy2 encountered is consistent with the idea that these galaxies are slightly richer in dust than the Sy1 (Malkan et al. 1998). The low fraction of LINERs, on the other hand, is consistent with the idea that these galaxies are low luminosity AGNs which are not in a starburst phase (Coziol 1996). Only 10% of the PDS Sy1 and 38% of the PDS Sy2 have a spectral index in the range -2.5 <= alpha(60,25) <= - 1.9. But the most striking result of our analysis is the fact that only 1% of the PDS starbursts have a spectral index alpha(60,25) > -1.9. The few starbursts which passed this limit are at the extreme of the sequence traced by the the starburst galaxies and show spectral characteristics at the borderline between those of starbursts and LINERs.

Our analysis clearly shows that Seyfert galaxies have distinct FIR colors from starburst galaxies. This means that the active nucleus in AGNs must contribute significantly to the FIR excess emission observed in these objects. Now, because this excess is, on average, barely equal to that observed in SBNGs, it suggests that the level of star formation in Seyfert galaxies may be different from that in starbursts. Taken at face value, our results imply that only a small fraction of the Seyfert galaxies, maybe about 40% of the Sy2 and about 10% of the Sy1, could be dominated by star formation. It is remarkable to find these fractions roughly consistent with those determined by González Delgado et al. (1997) based on their discovery of different star formation properties between Sy2 and Sy1 (for similar results see also Glass & Moorwood 1985; Maiolino et al. 1995; Hunt et al. 1997 and more recently Malkan et al. 1998). A better determination of the contribution of the active nucleus to the FIR is necessary in order to understand the relation of AGNs with starbursts.

Acknowledgments:

The electronic version of the tables for the different samples of galaxies defined in this paper (including the 187 IRAS normal galaxies and the Seyfert galaxies) are available upon request to the first author. We thank an anonymous referee for useful comments which allow us to improve the quality of our paper. R. C. acknowledges the CNPq for research fellowship. This research has made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet Propulsion Laboratory, Caltech, under contract with the National Aeronautics and Space Administration.