Next Contents


By allowing us a detailed view of the full 2-200 µm spectral range, ISO has dramatically increased our ability to investigate the processes giving rise to the spectral energy distribution (SED) of galaxies. Yet before actually starting this review, let us use well-known Galactic sources to exemplify the links between SEDs and the physical state of the objects in which they occur.

1.1. Mid-infrared bands and transient heating

The mid-infrared (MIR, 5-40 µm) is distinct from the far-infrared (FIR) in the sense that it shows a large number of broad spectral bands (especially in the 5-15 µm range) and that most of the dust is out of thermal equilibrium, in a regime of transient heating. The interesting property of transient heating is that, for a range of energy densities, the MIR flux scales linearly with the heating flux. This is clearly shown by Boulanger et al. (1998a) where they compare the MIR spectrum obtained on the peak of the NGC 7023 region (exposed to a B star radiation field) to that obtained on a diffuse cloud in the Chameleon region: the radiation fields differ by more than 3 orders of magnitude yet the MIR spectra are virtually identical. More quantitatively, they have shown that this linear scaling remains valid for radiation fields lower than a few 103 times the solar neighborhood value.

Without expanding on the nature of the dust giving rise to this family of bands, it is worthwhile to list a few properties that ISO has now firmly established. Even at the high resolution of the ISOSWS, the bands do not break up into a family of lines (as suggested e.g. by [Léger et al. 1989]). In fact, in most regions of the ISM, the band profiles are very constant, and much better represented by Lorentzian than by Gaussian functions (see e.g. [Boulanger et al. 1998b]; [Mattila et al. 1999]). Given the broad wings of a Lorentzian, in regions where the bands are prominent, most of the detected flux actually comes from the band carriers. As the debate still continues on the exact nature of the carriers, we will refer to them as the "infrared bands".

Studies of Galactic regions also help to pinpoint the major sites of emission: although they are detected in diffuse cirrus clouds ([Boulanger et al. 1996]), most of the infrared band emission originates in the interface between HII regions and molecular clouds, the photo-dissociation regions (PDR, see e.g. [Cesarsky et al. 1996]; [Verstraete et al. 1996]).

1.2. Continuum emission

From 10-15 µm to the submillimeter, the SEDs of most sources consist of a broad continuum. This continuum cannot be fitted by a single back-body curve. On the long wavelength side, this is due to the existence of more than one component of dust in thermal equilibrium. On the short wavelength side, this is due to the transition from the transient heating regime to the thermal equilibrium regime: below a size threshold fixed by grain properties and the radiation field intensity, dust grains still undergo noticeable temperature fluctuations. Depending on the heating radiation field, continuum emission will start to dominate over the infrared bands in the 12-15 µm range (e.g. in the PDR of M 17-SW, [Cesarsky et al. 1996]), or even over the whole MIR range (e.g. in the HII region of M 17-SW, [Cesarsky et al. 1996], see also the evolution of compact HII region spectra in [Cox et al. 1999]). One should note however that except in regions with a particularly hard radiation field, the MIR spectrum is generally dominated by the infrared bands.

1.3. infrared lines

The IR domain also gives access to very important diagnostic lines. These allow an almost extinction-free measurement of the intrinsic ionizing spectrum, the nature of the energy source, or the energetics of the interstellar medium (e.g. through the [CII]158 µm or [OI]63 µm lines).

The great advantage of ISO in this area is the possibility to observe the full set of lines from an object, free of any foreground emission, in a single fixed aperture (see e.g [Colbert et al. 1999], for M82).

1.4. Active sources

The three previous sections referred to emission processes occurring in the interstellar medium (ISM) of galaxies. However, the IR is also a wavelength range where emission from active galactic nuclei (AGN) can be detected. This can take two forms: (1) thermal emission from dust in the torus around the AGN, in which case we expect a very hot continuum as grains probably reach their sublimation temperatures, and (2) synchrotron emission from charged particles in the AGN strong magnetic field, in which case the emission takes the form of a featureless spectrum increasing with frequency.

Next Contents