1. INTRODUCTION

Discovering the process by which the dense, gravitationally bound galaxies formed in the Universe from an initially almost uniform gas, and understanding the way their constituent populations of stars were born is a key goal of modern physical cosmology. A wide range of well understood physical processes are involved; including general relativity, gas dynamics and cooling physics, nuclear reactions and radiative transfer. However, the range of possible initial conditions and the non-linear nature of most of the events, starting with the collapse of primordial density perturbations, ensure that these intimately connected processes can generate a very wide range of possible scenarios and outcomes. Galaxy formation can be studied by attempting to reproduce the observed Universe via analytical models and numerical simulations. The information required to constrain these models is provided by both forensic studies of the current constituents of the Universe, including stellar ages, chemical abundances and the sizes and shapes of galaxies, and by direct observations of the galaxy formation process taking place in the young Universe at great distances. Direct observations exploit both the light emitted by distant galaxies, and the signature of absorption due to intervening structures along the line of sight, and began almost 50 years ago using sensitive optical and radio telescopes. Astronomers must now use all available frequencies of radiation to probe the properties of the Universe, from the lowest energy radio waves to the highest-energy -rays. It is vital to combine the complementary information that can be determined about the constituents of the Universe at different wavelengths in order to make progress in our understanding.

This review discusses the results of a new type of direct observation of the galaxy formation process, made possible by the development of powerful new radiation detectors sensitive to wavelengths in the range 200 µm to about 1 mm: the submillimeter (submm) waveband. The detection of submm radiation from distant galaxies is one of the most recent developments in observational cosmology, and has finally brought this region of the electromagnetic spectrum into use for making cosmological observations not directly connected with the cosmic microwave background (CMB; Partridge and Peebles, 1967). With the possible exception of the hardest X-ray wavebands, studies of distant galaxies in the submm waveband remained elusive for the longest period. We will also discuss some observations at the mid- and far-infrared(IR) wavebands that bound the submm waveband at short wavelengths, usually defined as the wavelength ranges from about 5-40 and 40-200 µm, respectively.

The most significant reason for the late flowering of submm cosmology is the technical challenge of building sensitive receivers that work efficiently at the boundary between radio-type coherent and optical-like incoherent detection techniques. In addition, atmospheric emission and absorption permits sensitive submm observations from only high mountain sites, and only in specific atmospheric windows. The zenith opacity from the best sites in the clearest submm atmospheric window at 850 µm is typically about 0.1. Furthermore, the long wavelength of submm radiation limits spatial resolution unless very large filled or synthetic apertures are available. The largest single apertures available at present are in the 10-30 m class, providing spatial resolution of order 10 arcsec. This resolution is much coarser than the sub-arcsec resolution of optical and near-IR observations. The appearance of the same region of sky at optical and submm wavelengths is compared in Fig. 1 to illustrate this point: the multicolor optical image was obtained using the Hale 5-m telescope at Mt. Palomar, while the 850-µm submm image was obtained using the 15-m James Clerk Maxwell Telescope (JCMT) on Mauna Kea. Interferometers can dramatically enhance the resolution of images, but so far have only operated at longer mm wavelengths. The commissioning of the 8-element Sub-Millimeter Array (SMA; Ho, 2000) ² on Mauna Kea in Hawaii with baselines of up to about 500 m, the first dedicated submm-wave interferometer, will provide images with subarcsecond resolution. The much larger 64-element Atacama Large Millimeter Array (ALMA; Wootten, 2001) ³ will be in service at the end of decade.

Figure 1. A comparison of deep optical and submm views of the sky. The background image is a 3-color optical image of the rich cluster of galaxies Abell 1835 at the low/moderate redshift z = 0.25 (Smail et al., 1998b) taken using the 5-m Hale telescope, overlaid with the 14-arcsec resolution contours of a SCUBA 850-µm submm-wave image of the same field (Ivison et al., 2000a). North is up and East to the left. The brightest SCUBA galaxies at (-45,-15), (65,0) and (20,-60), and the central cD galaxy (Edge et al., 1999), all have clear radio detections at a frequency of 1.4 GHz in images with higher spatial resolution than the SCUBA contours, obtained at the Very Large Array (VLA), supporting their reality. The bright SCUBA galaxy at (-45, -15) is associated with SMM J14011+0253, an interacting pair of galaxies at redshift z = 2.56 in the background of the cluster (Frayer et al., 1999). Spectacular fragmented structure appears in the Easterly red component of this galaxy in Hubble Space Telescope (HST) images (Fig. 18).

A key development was the commissioning of the Submillimetre Common-User Bolometer Array (SCUBA) camera at the JCMT in 1997 (Holland et al., 1999). SCUBA images the sky in the atmospheric windows at both 450 and 850 µm in a 2.5-arcmin-wide field, using hexagonal close-packed arrays of 91 and 37 bolometer detectors at the respective wavelengths. SCUBA provided a dramatic leap forward from the pre-existing single-pixel or one-dimensional array instruments available. The combination of field of view and sensitivity was sufficient to enable the first searches for submm-wave emission from previously unknown distant galaxies. The Max-Planck Millimetre Bolometer Array (MAMBO; Kreysa et al., 1998) is a 1.25-mm camera with similar capabilities to SCUBA, which operates during the winter from the Institut de Radio Astronomie Millimétrique (IRAM) 30-m telescope on Pico Veleta in Spain. A similar device - the SEST Imaging Bolometer Array (SIMBA) - designed at Onsala in Sweden is soon to begin operation on the 15-m Swedish-ESO Submillimetre Telescope (SEST) in Chile, providing a sensitive submm imaging capability in the South. The capability of mm and submm-wave observatories is not standing still: a number of larger, more sensitive mm- and submm-wave cameras are under construction, including the SHARC-II (Dowell et al., 2001), BOLOCAM (Glenn et al., 1998) and SCUBA-II instruments. ⁴ Bolometer technology continues to advance. The advent of extremely stable superconducting bolometers that require no bias current and can be read out using multiplexed cold electronics, should ultimately allow the construction of very large submm detector arrays of order 10^4-5 elements (for example Benford et al., 2000). SCUBA-II is likely to be the first instrument to exploit this technology, providing a 8 × 8-arcmin² field of view at the resolution limit of the JCMT.

The first extragalactic submm/mm surveys using SCUBA and MAMBO revealed a population of very luminous high-redshift galaxies, which as a population, were responsible for the release of a significant fraction of the energy generated by all galaxies over the history of the Universe (Blain et al., 1999b). Almost 200 of these galaxies are now known (Smail et al., 1997; Barger et al., 1998; Hughes et al., 1998; Barger et al., 1999; Eales et al., 1999, 2000; Lilly et al., 1999; Bertoldi et al., 2000; Borys et al., 2002; Chapman et al., 2002a; Cowie et al., 2002; Dannerbauer et al., 2002; Fox et al., 2002; Scott et al., 2002; Smail et al., 2002; Webb et al., 2002a). There is strong evidence that almost all of these galaxies are at redshifts greater than unity, and that the median redshift of the population is likely to be of order 2-3 (Smail et al., 2000, 2002). However, only a handful of these objects have certain redshifts and well-determined properties at other wavelengths (Frayer et al., 1998, 1999; Ivison et al., 1998, 2001; Kneib et al., 2002). The results of these mm/submm surveys provide complementary information to deep surveys for galaxies made in the radio (Richards, 2000), far-IR (Puget et al., 1999), mid-IR (Elbaz et al., 1999) and optical (Steidel et al., 1999) wavebands. Submm observations are a vital component of the search for a coherent picture of the formation and evolution of galaxies, which draws on data from all wavebands where the distant Universe can be observed.

In this review, we describe the key features of the submm emission processes in galaxies. We summarize the current, developing state of submm-wave observations of distant galaxies, including the results of both blank-field surveys, and targeted observations of known high-redshift galaxies, including radio-galaxies, optically-selected quasars/QSOs, X-ray detected active galactic nuclei (AGNs) and optically-selected Lyman-break galaxies (LBGs). Submm-wave surveys are not immune to selection effects, and we discuss their strengths and weaknesses. We describe the properties of the class of submm-luminous galaxies, and discuss the key results that are required to make significant progress in understanding them. We consider the relationship between the submm-selected galaxies and other populations of high-redshift galaxies, and describe models that can account for the properties of submm-selected galaxies. We introduce the unusually significant effects of the magnification of distant submm-selected galaxies due to gravitational lensing (Schneider et al., 1992). Finally, we recap the key developments that are keenly awaited in the field, and describe some of the exciting science that will be possible in the next decade using future instruments.

The cosmological parameter values assumed are generally listed where they appear. We usually adopt a flat world model with a Hubble constant H₀ = 65 km s^-1 Mpc^-1, a density parameter in matter _m = 0.3 and a cosmological constant = 0.7.

² http://sma-www.harvard.edu/ Back.

³ http://www.alma.nrao.edu Back.

⁴ Details can be found in Table 3. The next-generation SCUBA-II camera for the JCMT is under development at the United Kingdom Astronomy Technology Centre (UKATC). See http://www.jach.hawaii.edu/JACpublic/JCMT/Continuum_ observing/SCUBA-2/home.html. Back.