|Annu. Rev. Astron. Astrophys. 2005. 43:
Copyright © 2005 by Annual Reviews. All rights reserved
A fraction of the stellar radiation produced in galaxies is absorbed by dust and re-radiated from mid-infrared to millimeter wavelengths. Understanding dust properties and the associated physics of the absorption and emission are thus essential. These determine the Spectral Energy Distribution (SED) of the galaxies.
2.1. Dust Particles
Small dust particles with sizes ranging from a nanometer to a fraction of micrometer are ubiquitous in the interstellar medium. They result from natural condensation in cool stellar atmospheres, supernovae, and the interstellar medium of the heavy elements produced by the nucleosynthesis in stars and released to the diffuse medium by late type stars and supernovae explosions. Interstellar grain models have been improved for 30 years in order to fit all observational constraints: elemental abundances of the heavy elements, UV, visible and infrared absorption and scattering properties, infrared emission, polarization properties of the absorbed and emitted light. The models include a mixture of amorphous silicate grains and carbonaceous grains, each with a wide size distribution ranging from molecules containing tens of atoms to large grains 0.1 µm in diameter that can be coated with ices in dense clouds and/or organic residues (e.g., Désert et al. 1990; Li & Draine 2001). It is now widely accepted that the smallest carbonaceous grains are Polycyclic Aromatic Hydrocarbons (PAHs) that emit a substantial fraction of the energy in a set of features between 3 and 17 µm (3.3, 6.2, 7.7, 8.6, 11.3, 12.7, 16.3, 17 µm for the main ones) that used to be known as the UIB for Unidentified Infrared Bands. These features result from C-C and C-H stretching/bending vibrational bands excited by the absorption of a single UV or optical photon and are a good tracer of normal and moderately active star formation activity in spiral and irregular galaxies (e.g., Helou et al. 2000; Peeters et al. 2004). For radii a 50 Å, the carbonaceous grains are often assumed to have graphitic properties. The so-called very small grains of the interstellar medium are small enough to have very low heat capacity, so their temperature are significantly affected by single-photon absorption. In the diffuse ISM of our Galaxy, they dominate the infrared emission for wavelengths smaller than about 80 µm. At longer wavelengths, the infrared spectrum is dominated by the emission of the larger grains at their equilibrium temperatures. Considering the energy density of the radiation in a galactic disc like ours, the temperature of the larger grains is rather low; 15 to 25 K. For these grains, the far-infrared emissivity decreases roughly as the square of the wavelength. This in turn makes the temperature dependence on the radiation energy density u very weak (T u1/6). For a galaxy like the Milky Way, the infrared part of the SED peaks at 170 µm whereas for an Ultra Luminous Infrared Galaxy (ULIRG) it peaks at about 60 µm: a factor 3 in temperature for a factor 103 in energy density or luminosity. At long wavelengths in the submillimeter and millimeter, the intensity should increase like I 4.
In our Galaxy, the extinction curve of the diffuse ISM has been known for a few decades. The average optical depth perpendicular to the disk of our Galaxy in the solar vicinity is small (Av 0.2) and typical of spiral galaxies. The average optical depth increases to a few in large molecular cloud complexes. It can become very large in galactic nuclei. Finally it should also be remembered that the optical depth in the UV is typically 10 times larger than that in optical wavelengths. The conversion of star light into infrared radiation will thus depend strongly on the location of the stars and their spectral types.
In external galaxies, modeling the extinction is very hard because it strongly depends on the geometric distribution of the ISM and of the chemical abundances. Simple models have been used to take this into account to first order. Galaxies can be modeled as an oblate ellipsoid where absorbers (dust) and sources (stars) are homogeneously mixed; the dust absorption can be computed in a "screen" or "sandwich" geometry (dust layers in front of the stars or sandwiched between two star layers). As a consequence, the reddening curve average over a whole galaxy appears to vary within a class of objects and between the different classes, from normal star-forming galaxies to highly concentrated starburst. It is thus very difficult to derive the total dust optical depth (e.g., Calzetti et al. 1994). In the local Universe, the average extinction per galaxy is quite low. About one third of the bolometric flux is emitted in the far-infrared, and this is typical of our Galaxy. In more actively star-forming galaxies, up to 70% of the bolometric flux is emitted in the far-infrared. In some of these, the starburst activity is mostly in the disk (like in M51). For a given total luminosity, the radiation energy density is lower than in the case of a starburst concentrated in a small volume in the nucleus. In this case, the dust will be hotter due to the larger energy density and the conversion of stellar light to infrared will be more efficient. Some ULIRGs emit more than 95% of their energy in the far-infrared (e.g., Arp 220). Such galaxies are very compact, dusty starbursts where dust optical depths are very large. In such galaxies, fine structure and recombination line ratios imply an equivalent "screen" dust extinction between Av ~ 5 and 50. The result is that the SED is significantly distorted in the opposite way from the higher dust temperature (less mid-infrared emission). In the following, we will refer to "infrared galaxies" and to "optical galaxies" to designate galaxies in which the infrared emission, respectively optical emission dominates. Different typical spectra of galaxies are shown in Figure 1 from the UV to the millimeter. We clearly see the variation of the optical to infrared energy ratio as starburst activity increases.
Figure 1. Spectral energy distributions of galaxies from UV to the millimeter. The ULIRG is observed at redshift z = 0.66 and is represented here in the rest-frame (from Galliano 2004).
2.3. Local Infrared Galaxies
A few very luminous infrared galaxies were observed in the seventies (Rieke & Lebofsky 1979). Then IRAS satellite, launched in 1983 gave for the first time a proper census of the infrared emission of galaxies at low redshift. The Luminosity Function (LF) at 60 and 100 µm is dominated by L⋆ spiral galaxies as could be expected - the reradiated stellar luminosity absorbed by dust. In addition, a high-luminosity tail of luminous galaxies was found (e.g., Sanders & Mirabel 1996). This high-luminosity tail can be approximated by a power-law, (L) LIR2.35, which gives a space density for the most luminous infrared sources well in excess of predictions based on the optical LF. These sources comprise the Luminous Infrared Galaxies, LIRGs, and the ULIRGs with luminosities 11 < log(LIR / L) < 12 and log(LIR / L) > 12, respectively. These galaxies are often associated with interacting or merging, gas-rich disks. The fraction of strongly interacting/merger systems increases from ~ 10% at log(LIR / L) = 10.5-11 to ~ 100% at log(LIR / L) > 12. LIRGs are the site of intense starburst activity (about 10-100 M year-1) induced by the interaction and/or strong spiral structure. The ULIRG phase occurs near the end of the merging process when the two disks overlap. Such galaxies may be the precursors of Quasi Stellar Objects (QSOs; Sanders et al. 1988a, 1988b; Veilleux et al. 1995; Lutz et al. 1999). These objects have been the subject of intense debate concerning the nature of the dominant source of emission: starburst versus dust-enshrouded AGN (e.g., Filipenko 1992; Sanders & Mirabel 1996; Joseph 1999). Indeed, spectra show evidence of extremely large optical depth (heavily reddened continuum and large Balmer decrement) but also exhibit AGN-like high excitation fine-structure lines. We had to wait for ISO to clearly determine the power sources of ULIRGs. The difference between the mid-infrared spectra of starburst and AGNs is striking. Starburst are often characterized by strong, low-excitation, fine-structure lines, prominent PAH features and a weak 10 µm continuum whereas AGNs display a highly excited emission line spectrum with weak or no PAH features, plus a strong mid-infrared continuum. It has been thus possible to build mid-infrared diagnostic diagrams (e.g., Genzel et al. 1998; Laurent et al. 2000) that clearly separates starburst-dominated galaxies from AGN-dominated galaxies. These diagrams demonstrate that ULIRGs appear to be composite objects, but star formation dominates in most objects. That is on average, 70% of the reradiated energy comes from starbursts and 30% comes from AGNs (Genzel et al. 1998; Lutz et al. 1998). However the fraction of AGN-powered objects increases with luminosity. About 15% of ULIRGs at luminosities below 2 × 1012 L are mostly AGN powered, but this fraction increases to about half at higher luminosity.
All these well-studied LIRGs and ULIRGs are at low redshift. They do not dominate the energy production locally. As an example, the total infrared luminosity from these galaxies in the IRAS Bright Galaxy Sample accounts for only ~ 6% of the infrared emission in the local Universe (Soifer & Neugeubauer, 1991). As we will see, the situation changes dramatically at higher redshift where these galaxies fully dominate the infrared energy output.