Our multiwavelength imaging and spectroscopic study of the 4 UV brightest Seyfert 2 clearly shows that nuclear (radii of 83#83 pc) starbursts are an important component in the energetics of these nuclei. They are the primary source of the so-called `featureless-continuum' in the UV-through-near-IR regime in these objects. In fact, this spectral component is not featureless. We detect the direct spectroscopic signature of massive stars: 1) stellar wind and photospheric lines in our HST GHRS data, 2) HeI and high-order Balmer absorption-lines in the near-UV and visible and 3) CaII triplet absorption in the near-IR due to red supergiants.
The estimated bolometric luminosity of the nuclear starburst is roughly 1010 L2#2 in NGC 5135, IC 3639, and NGC 7130, and about 4 84#84 L2#2 in Mrk 477. The ratios between the starburst luminosity and the radio and [OIII] luminosities in these galaxies are similar to those found in Mrk 477. The difference found between the power of the starburst in Mrk 477 and that in NGC 7130, NGC 5135 and IC 3639 reflects the diference in one order of magnitud between nebular [OIII] emission line and the nonthermal radio continuum power of Mrk 477 with respect to the other three Seyfert 2 nuclei. It seems that more powerful active nuclei are associated with more vigorous nuclear starbursts (at least in this small sample). In fact, we estimate that the bolometric luminosity of nuclear starburst is similar to that of the AGN itself (the hidden type 1 Seyfert nucleus). In the case of Mrk 477, the nuclear starburst plus the AGN may actually dominate the overall energy output of the entire galaxy. In the other objects, the sum of the nuclear starburst and AGN may contribute roughly 10 to 20% to the overall galaxy bolometric luminosity (which in all four cases is primarily emitted in the mid/far-IR). The UV spectra taken through the relatively large aperture of IUE imply that the most of the IR emission not due to the AGN plus nuclear starburst is produced by a larger (several kpc-scale) dusty starburst.
The WFPC2 and FOC images suggest that the nuclear starburst has formed as a consequence of the dynamics of the host galaxy. Bars (in NGC 7130, NGC 5135 and IC 3639) and/or interaction (Mrk 477 and NGC 7130) provide a very efficient mechanism to drive the gas from the outer parts of the galaxy to the nuclear region. Knots of star formation have been detected along the leading edges of the bars and in spiral-like structures that extend into the nucleus in these Seyferts 2 galaxies (as well as in other low luminosity active galaxies - Colina et al 1997). This flow of gas toward the inner core may trigger star formation along the way, and probably forms a molecular torus that acts as a gas reservoir that obscures and finally feeds the nucleus. There, as has been suggested by Cid Fernandes & Terlevich (1995) and Heckman et al (1997), a nuclear starburst can naturally form from the reservoir of molecular gas.
The four cases studied in this paper and in Heckman et al (1997) have been selected on the basis of the brightness of the nucleus in the UV (to enable spectroscopy with the GHRS). This criterion may favor the selection of galaxies with more vigouros nuclear starbursts. Thus, an obvious and important question is `Can this result be generalized to all the Seyfert 2 galaxies?' Our preliminary analysis of ground-based near-UV and optical spectra of the brightest 20 Seyfert 2 nuclei indicates that massive stars are definitely present in the innermost arcsec of at least a-third of the nuclei. These are just those cases where the strength of the young stellar continuum is greatest relative to the light from the old underlying stellar population and from the nebular gas (e.g. in the cases where the starburst continuum is most easily studied). Thus, we do not yet know whether nuclear starbursts are present in all Seyfert nuclei or only in some. If the latter, is there some key evolutionary connection between the starburst and the Seyfert nucleus? The availibility of NICMOS and STIS on HST, together with complementary multi-wavelength and ground-based data may soon allow us to answer these fundamental questions.
Acknowledgments
We are grateful to Roberto Terlevich and Enrique Pérez for many stimulating discussions and helpful suggestions, and to the staff at STScI and at KPNO for their help in obtaining the data presented in this paper. This work was supported by HST grants AR-05804.01-94A, GO-5944, and GO-6539 from the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555. It was also supported in part by the NASA LTSA grant NAGW-3138.