1.3 Galaxy Activity at Infrared Wavelengths

The build up of stellar populations in galaxies is most usually investigated by looking at the optical/near-IR emission from already formed stars in distant galaxies. The complementary approach, less frequently used, is to look at the diffuse media - atomic and molecular gas and dust - and their progressive transformation into stars. While observations of starlight emission in the optical/near-IR can exploit large telescopes on ground and very efficient photon detectors, reliable probes of the diffuse media require longer-wavelength observations in the mid-to-far-IR and sub-millimeter, where a large variety of lines from atomic species and molecules in the Inter-Stellar Medium (ISM) at all ionization levels are observable. Another fundamental component of the ISM, dust grains present in all astrophysical settings ranging from planetary disks to nuclear accretion tori around quasars, have the property to emit at these wavelengths, typically between a few $\mu$m to 1 mm. Observations at long wavelengths are then essential to study diffuse media in galaxies and are particularly suited (and needed) to study the early phases in galaxy evolution, when a very rich ISM is present in the forming system.

Following this line of reasoning, by the generic term of galaxy activity we refer to transient phases in the secular evolution of a galaxy during which the various transformations of baryons undergo a significant enhancement with respect to their average rate, for reasons to be ascertained. As it turns out, these phenomena concern both enhanced rates of conversion of the ISM gas into stars, the Starburst phenomenon, and phases of increased activity of the nuclear emission following an event of fast accretion of gas onto the super-massive black hole, the Active Galactic Nucleus phase, reaching parossistic levels of photon production of up to $> 10^{50}$ erg/s in some high-$z$ quasars.

Conversely, by quiescent (or non-active) galaxies we indicate those galaxies in which current and/or recent star formation and AGN activity contribute an insignificant fraction of bolometric luminosity. Ellipticals and lenticulars are generally regarded as the most quiescent among galaxies. They have relatively small amounts of interstellar matter and therefore little fuel to feed either star formation or AGN activity. However, they are not devoid of circumstellar and interstellar gas and dust, as IR observations indicate.

IR and sub-mm wavelengths provide a privileged viewpoint to investigate galaxy activity in general, for three main reasons: (a) in many cases this wavelength range includes a dominant fraction of the whole bolometric output of active objects; (b) at long wavelengths the screening effect of diffuse dust, present in large amounts in active galaxies, is not effective anymore, and a free access to even the most extreme column-density regions is possible; (c) the infrared is a domain so rich in spectral lines, less subject to absorption than optical ones, that infrared spectra allows a clearer physical characterization of the sources under exam.

In starburst galaxies, the near-IR continuum emission, especially at 1-3 $\mu$m, is also usually dominated by photospheric radiation from red giant and supergiant stars, with smaller detectable contributions often arising from ionized gas, very hot dust and blue stars. Somewhere between 3 and 10 $\mu$m, depending on the star formation intensity, the continuum begins to rise with increasing wavelength, as the reradiated emission from hot dust becomes more important. The far-IR dust distribution reaches a maximum in the 80-100 $\mu$m region, falling off rapidly with wavelength thereafter (see Figure 1.3).

The infrared also plays a particular role in studies of the AGN phenomenon. In recent years, multi-wavelength observations of AGN have provided support to the AGN unified scheme, according to which these objects are intrinsically similar, with various scales, but are viewed under different angles. An energy source at their center is surrounded by a torus of obscuring gas and dust, with a radius from a few parsecs to tens or a hundred parsecs. The observational properties of the objects depend not only of the energy source, but also on the angle at which they are observed with respect to the torus plane. In this scheme, radio quiet, steep spectrum quasars and Seyfert 1 galaxies are seen at intermediate angles, flat spectrum quasars and BL Lacs are pole-on, and radio-galaxies and Seyfert 2 edge-on. In the infrared, it is possible to observe the direct emission of the torus. Models indicate that tori are still optically thick in the MIR. But in the FIR a simple prediction of the unification models is that sources of similar energy should emit the same amount of thermal radiation from torus-heated dust.

Given that both Starbursts and AGN emit substantial amounts of energy at long wavelengths, but also that spectral discrimination between these two forms of galaxy activity is extremely difficult and demanding in terms of telescope time, one would like to be able to distinguish them on the basis of multi-colour broad-band infrared photometry.

Figure 1.4: IRAS Two-Colour Diagram for Normal (i.e. Non-AGN) Galaxies. From Helou (1986).

Helou (1986) showed that IRAS colours of non-AGN galaxies are reasonably well defined. Figure 1.4 shows Helou's colour-colour plot for unresolved IRAS galaxies with high quality fluxes. The straightforward interpretation suggested by Helou for this locus of colours is that it represents a mixing line, with cirrus emission dominating the colours to the lower right, starburst emission dominating in the upper left, and a mixture of these two determining a galaxy's specific colours between these two extremes. The cirrus dust emitting the far-IR radiation is very cool (low 60-100 $\mu$m ratio) because the intensity of the ISRF is low, and the mid-IR emitting cirrus is hot (high 12-25 $\mu$m ratio) because those grains are very small and transiently heated. The higher radiative energy densities in starbursts heat up the larger grains, thereby increasing their 60-100 $\mu$m colour temperature. The flux from those grains increases dramatically with temperature ( $\sigma T^4$), and the energy distribution shifts to shorter wavelengths. Emission from these larger grains, which are cooler than the small transiently heated ones, begin to dominate the mid-IR, thus resulting in a lower 10-25 $\mu$m colour temperature. The distribution of galaxy colours in the two colour plot can also be understood as the locus generated as the energy density that heats the dust is increased from cirrus values (lower right) to starburst values (upper left).

Multi-colour infrared photometry also helps in employing IR SEDs as a diagnostic tool between Starbursts and AGN. The reliability of the assessment depends on where an object's colour fall in colour-colour plots, since there are some regions where different classes overlap more. Rowan-Robinson (1989) have made an effort to separate the contributions from starburst, cirrus and AGN to the IR emission from a large sample of IRAS galaxies. They assumed canonical distributions for each class, and they showed that generally these separations are straightforward and plausible.