Understanding the origins of the cosmic background radiation, from the X-ray to the radio, has led to detailed analysis of distant galaxy populations and their formation and evolution. While astronomers have made significant progress in studying the stellar mass content of galaxies, the growth and assembly of stellar populations, the physics surrounding supermassive black holes in galactic centers, and the large number of luminous galaxies in the early Universe - all from optical, X-ray and radio observations - it has become increasingly evident that a significant portion of galaxies' light is emitted at far-infrared and submillimeter wavelengths.
In the early 1990s, the Far-InfraRed Absolute Spectrophotometer (FIRAS) aboard the space-based Cosmic Background Explorer (COBE) measured the absolute energy spectrum of the Universe at far-infrared and sub-millimeter wavelengths above 150 µm. These measurements, along with prior observations of nearby galaxies with IRAS in the 1980s, showed for the first time that the Universe emits a comparable energy density at infrared and sub-millimeter wavelengths as it does in the more traditionally studied optical and ultraviolet. The implications of this are significant: optical and ultraviolet observations alone will miss roughly half of the star formation activity in the Universe. More troublesome is that this problem is exacerbated at high-redshift, where the bulk of the cosmic star formation activity takes place. The original COBE measurements of the cosmic infrared background (CIB), combined with galaxy surveys at optical wavelengths, showed that there must be either a population of galaxies that are enshrouded in dust and/or numerous dust-enshrouded regions within optically-detected galaxies where newly born massive stars are likely born. The radiation coming from stars in dusty regions heat the dust and the thermal emission from the dust at far-infrared and submm wavelengths (from roughly 10 - 1000 µm) leaves the signature of their presence.
At these long wavelengths, one encounters multiple observational limitations: the water vapor in the atmosphere forces sensitive observations to be made from either high and dry locations on Earth or from space. Another challenge has been the hitherto poor sensitivity of detectors operating at these long infrared wavelengths. Still, since the original measurements of the IR background by COBE, the field has seen an explosion of interest. In the late 1990s, deep blank field pointings with the Submillimeter Common-User Bolometer Array at 850 µm on the James Clerk Maxwell Telescope (JCMT) directly detected, for the first time, populations of high-redshift galaxies that are extremely bright at far-IR/submm wavelengths but are nearly invisible in the optical. These ground-breaking studies revolutionized the field of galaxy formation, and turned the study of high-z dusty galaxies into one of the fastest growing areas of extragalactic astronomy. Following surveys by the Spitzer Space Telescope at mid- and far-IR wavelengths, as well as by both balloon and ground-based submillimeter (submm) and millimeter (mm) single dish radio facilities, have since detected galaxies with a wide range of inferred star formation rates, stellar masses, and black hole luminosities. Some of the most significant and recent developments have come from the Herschel Space Observatory which, during its 4 year operations between 2009 and 2013, mapped more than 1300 deg2 of the sky between 100 and 500 µm and detected more than a million galaxies bright at far-infrared and submm wavelengths (Figure 1), and the South Pole Telescope, which has mapped ~ 2500 deg2 at 1.4-2.0 mm, detecting hundreds of brightly lensed dusty galaxies. The increases afforded by these facilities have allowed for large statistical studies of a cosmologically crucial populations of galaxies which is a marked improvement over the original samples of ~ 200 galaxies discovered in the original submm surveys with SCUBA in the late 1990s. With the recent commissioning of Scuba-2 on the JCMT, as well as the Atacama Large Millimeter Array (ALMA) and soon-to-be construction of the Cerro Chajnantor Atacama Telescope (CCAT) in Chile, unprecedented statistics and detailed physical characterization of precisely-identified submm bright galaxies at high-z are imminent.
Figure 1. The false-color image of Herschel-Spire instrument map of the GOODS-N region of the sky. The three panels to the left show the sky at 250, 350 and 500 µm in blue, green, and red, respectively. At right, the combined false-color image shows galaxies that are brighter at 250 µm (bluer) vs those that are brighter at 500 µm (redder). This color change could come from either differences in the thermal dust temperature or due to differences associated with the redshifting of the thermal spectrum as a function of the redshift. The red color then indicates galaxies that contain colder dust or are at higher redshifts. As we discuss later, follow-up observations have shown the latter case to be the primary reason for the color changes. The image spanning 30 × 30 arcmin2 contains close to a few hundred individually detected galaxies that are brighter at these submm wavelengths and makes up less than a thousandth of the area surveyed by Herschel. This image was done with 30 orthogonal scans of the Herschel-Spire instrument. Most of the extragalactic sky area covered by Herschel-Spire involves two orthogonal scans which effectively reaches the same confusion-limited depth as this data but is less useful for advanced statistical test (for an example of a map with two scans see Fig. 34).
Since their initial discovery, dusty star-forming galaxies (DSFGs 1) at high-z have become a critical player in our understanding of cosmic galaxy formation and evolution. The most luminous of these systems are the brightest galaxies in the Universe, and are seen back to just ~ 800 Myr after the Big Bang. These DSFGs are the most intense stellar nurseries in the Universe, and have inferred star formation rates (SFRs) of as much as a few thousand solar masses per year compared to the Milky Way's paltry ~ 2 M⊙yr-1 (Robitaille & Whitney, 2010, Chomiuk & Povich 2011). Much of this is happening in a spatial extent so compact, that the star formation rate surface densities of these galaxies are among the largest known. These DSFGs provide a unique laboratory for investigating the physics of star formation in environments far more extreme than can be found in our own galaxy 2.
A sub-sample of DSFGs are known to harbor heavily dust-enshrouded supermassive black holes. A significant fraction of these galaxies' infrared luminosity can actually be dominated by AGN heating mechanisms, rather than star formation processes, as is the case for the recently characterized WISE-selected galaxies (e.g. Blain et al., 2013). In fact, a variety of lines of evidence suggest that these galaxy may serve as precursors to luminous quasars, and serve as the site of a rapid growth phase of central black holes, as they approach nearly a billion solar masses. At the other extreme, fainter galaxies just barely detected in the far-IR/submm form stars at rates of tens to hundreds of solar masses per year, and contribute principally to a cosmic infrared background that is comparable to the energy density of all direct starlight from all galaxies in the ultraviolet and optical wavelength regimes.
The discovery of copious numbers of submm and infrared emitting galaxies at high-z has proven to be a significant challenge for theoretical models of galaxy formation. As we will discuss later, cosmological models of structure formation and galaxy evolution have had a difficult time understanding the origin and evolutionary destiny of these heavily star-forming systems utilizing simulations, especially those designed ab initio. Whether or not they are simply scaled up versions of local extreme galaxies (such as UltraLuminous Infrared Galaxies, ULIRGs), or different beasts all together is still a heavily debated topic today.
Extragalactic infrared-based astronomy is at a period of great growth. With the development of the James Webb Space Telescope (JWST; space-based near and mid-IR), continued development and operations of the Atacama Large Millimeter Array (ALMA; ground-based far-IR and submm interferometer) and the planning of CCAT, the community is investing heavily in facilities that will enable the detection of and physical characterization of dusty systems out to the Universe's earliest epochs.
In this review, we summarize what has been learned about high-z dusty star-forming galaxies over the past decade, from the characterization of original submm sources detected by Scuba, to newer DSFGs found by AzTEC, Herschel, Scuba-2, SPT and others. We will summarize both the population statistics—number counts, redshift distribution, luminosity functions—as well as detailed physical properties of these dusty, star-forming galaxies. We will review the contribution of these galaxies to the cosmic star-formation rate density, the stellar mass build-up of the Universe, the formation of massive early-type galaxies, and properties of DSFGs' star-forming regions. We will also present results related to source clustering and the anisotropies of the background. Finally, we review theoretical attempts over the last decade to understand the origin and evolution of DSFGs in a cosmological context.
In this review, we summarize what has been learned about high-z dusty star-forming galaxies over the past decade, from the original days of Scuba, the the flourishing, diverse datasets and simulations we have today. This review is organized as follows. In § 2 we summarize basic properties of various galaxy populations that are characterized as DSFGs and the observational programs at submm and far-IR wavelengths. In § 3 we review existing measurements related to the number counts of DSFGs at a variety of wavelengths, including methods for analyzing counts in submm maps to gravitationally lensed counts at the bright-end of submm/mm flux densities. In § 4 we review the redshift distributions of DSFGs, how to fit spectral energy distributions to their far-infrared data, and the implied luminosity functions and measurements of the cosmic star formation rate density. The internal physical characterization of DSFGs, including multi-wavelength properties, morphologies, and dynamics are reviewed in § 5. Although § 5 discusses the role of AGN in DSFGs, we note here this review focuses primarily on extreme star-forming galaxies and does not explicitly focus on dusty luminous galaxies for which luminous black holes are thought to dominate the bolometric luminosity. In § 6 we review some of the basic physical properties of individual DSFGs that are studied in detail in the literature. The spatial distribution of DSFGs and galaxy clustering and environmental effects are in § 7. In § 8 and § 9 we review the molecular gas, mainly CO and dense gas tracers such as HCN, and ionized gas, such as [CII], properties of star-forming galaxies, respectively. § 10 presents a review of theoretical models related to DSFG formation and evolution, from numerical and hydrodynamical simulations to semi-analytical recipes in the literature. We conclude our review with a summary of outstanding scientific questions for future research programs in § 11. When quoting results, as needed, we assume a general cosmological model consistent with Planck data (Plank Collaboration et al., 2013b). We state our initial mass function (IMF) assumptions when appropriate, and throughout, address how changes in the IMF will alter select results. Similarly, when appropriate, we discuss the issue of AGN dust heating and how that impacts estimated star formation rates and the physical interpretation of certain DSFGs.
1 Hereafter, we generalize all galaxies at high-z that have been originally selected at infrared or sub-millimeter wavelengths as dusty star-forming galaxies (DSFGs). This encompasses a diverse zoo of galaxies, that we discuss in greater detail in § 2. Back.
2 While certain regions of the Milky Way have star formation surface densities as high, distant DSFGs are unique in that the high density environment encompasses the entire ISM (Wu et al., 2009). Back.