|Annu. Rev. Astron. Astrophys. 2000. 38: 761-814 |
Copyright © 2000 by Annual Reviews. All rights reserved
In the local Universe surveyed by IRAS, only 30% of the total energy output of galaxies emerges in the mid- and far-IR (Soifer & Neugebauer 1991). If the same were true at high redshifts, optical/UV observations alone would be sufficient for determining the cosmic star formation history. Spectacular (rest frame) UV observations of star-forming galaxies at redshifts 3 have become possible in the last few years (Steidel et al 1996, Lilly et al 1996). The inferred star formation rate per (comoving) volume element increases from z = 0 by more than an order of magnitude to a peak at z ~ 1-3 (Madau et al 1996, Pettini et al 1998, Steidel et al 1999).
A significant new element in the picture of high-z star formation has emerged with the COBE detection of an extragalactic submillimeter background (Puget et al 1996). This diffuse background has an integrated intensity that is comparable to or larger than that of the integrated UV/optical light of galaxies (Dwek et al 1998, Hauser et al 1998, Lagache et al 1999). The implication is that there likely exists a very significant contribution of dust-obscured star formation at high redshifts. Clearly, the next major step is to detect these high-redshift infrared galaxies directly. With the exception of a few hyperluminous and/or lensed objects (such as F10214+4724; Rowan-Robinson et al 1991), IRAS was only able to detect infrared galaxies to moderate redshifts (z ~ 0.3). In the mid-infrared, ISOCAM is ~ 103 times more sensitive and has 60 times higher spatial resolution than IRAS. In the far-IR with ISOPHOT, the improvement in sensitivity is modest, but the extension to longer wavelength (175 µm) is of substantial benefit. Exploiting these improvements in deep surveys, ISO has been able to provide for the first time a glimpse of the infrared emission of galaxies at z 0.5.