Ultra-Luminous IR Galaxies at Low and High Redshift

In the luminous infrared galaxies, nuclear starbursts and active nuclei are fueled by extraordinarily large masses of gas and dust concentrated at radii of a few hundred pc by galactic merging and viscous accretion. The nearby ULIRGS are probably excellent analogs of galaxies seen at high redshift during the epoch of galaxy formation and growth. We summarize results from the NICMOS-GTO survey of 24 ultra-luminous IR galaxies together with mm-interferometry of the molecular gas in these galaxies and in selected high redshift submm sources (detected initially with SCUBA). Eight of the 24 galaxies imaged with NICMOS have significant nuclear point sources and eleven have double nuclei; in point-nucleus systems it is possible that AGNs contribute significantly to the high infrared luminosity output. Indeed, most of the eight exhibit emission line spectra indicative of a hard ionization source. For the remainder of the sample, most of the near-infrared flux clearly originates outside the central 50–200 pc and is thus stellar (probably starburst) in origin. Nine of the 24 systems are fit better by an r^(1/4) law (rather than an exponential disk), suggesting that the young starburst population can relax rapidly in violent mergers. Whether these star systems eventually bcome massive central bulges or giant elliptical galaxies will depend on if the present ISM is converted into stars at high efficiency. In the IR galaxies much of the enhancement in the star formation probably occurs via the collision of massive clouds since, often, large numbers of bright clusters may be found in the overlap regions of the colliding galaxies. Based on these recent high resolution NICMOS and mm-wave interferometric imaging on local luminous infrared galaxies (e.g., Arp 220), we expect that the optical/UV emission of high redshift starburst galaxies will be highly extincted by extremely large dust concentrations. The most luminous systems are very likely under-represented in optical surveys, and the empirical corrections to the UV/optical fluxes, based on observed SED slopes, are probably inappropriate for the most luminous and most obscured systems