Next Contents Previous

2.4. Dust Morphology

A major complication for interpreting SEDs of gas- and dust-rich galaxies is the potentially large uncertainty introduced by dust obscuration. Star-forming galaxies are associated with large molecular gas densities. This immediately suggests that dust must be plentiful since molecular hydrogen forms by adsorption on dust grains. Moreover, star formation comes together with metal production. Therefore star formation itself will rapidly enrich the ISM in heavy elements and dust. This occurs over rather short time scales. The time scales for the formation of dust are only a few hundred Myr ([7]) so that obscuration effects can set in fairly early in the history of a galaxy.

Dust efficiently scatters and absorbs UV radiation. In order to quantify the interaction between dust and photons, assumptions on the amount of dust must be made, as well as on its geometry and chemical composition. Three processes play a role. (i) Dust obscures matter, i.e., dust conceals stars and gas from view by covering them wholly or in part. (ii) Dust attenuates light, i.e., it diminishes the amount of light seen by an observer. This leaves open the possibility of either absorption or scattering. (iii) Dust absorbs photons, i.e., it transforms the photons energy. The significance of these processes becomes obvious by recalling that in the Local Group the UV extinction law varies from galaxy to galaxy and within galaxies. The extinction is due to a combination of absorption and scattering of photons. The geometry is quite different in more distant galaxies where individual stars cannot be isolated. In such cases much of the light may be scattered into the line of sight, and the properties of the dust become paramount for the interpretation of the observed photon distribution ([4]).

These dust effects are of course always accounted for before making comparisons with observations. What is often not appreciated, is the subtle interplay of the evolution of the stellar SED and the spatial morphology of the dust. In Fig. 5 I have plotted models for the important C IV lambda1550 line, which assume time-dependent dust attenuation. This line is formed by a population of OB stars, with the emission and absorption coming from the most massive O stars, and the continuum from less massive O, and some B stars. The reason for the "dilution" of the line profile over time is a gradual decrease of the dust reddening with time. The astrophysical justification is that we are seeing the less massive stars through a hole where the energetic O-star winds have blown out the natal cocoon veiling the most massive stars. Very young clusters are known to remain embedded in their natal material until energetic stellar winds from evolving massive stars blow out the surrounding gas and dust ([50]). If younger populations are preferentially more attenuated, the equivalent widths of C IV lambda1550 and other lines become skewed towards smaller values. If the ISM is inhomogeneous, a fundamental property of a spectral line associated with a single star changes: its equivalent width becomes reddening dependent.

Figure 5

Figure 5. C IV lambda1550 for a model of a 50 Myr old stellar population forming stars constantly. The different line types denote different dust dispersal time scales. Solid line: no time dependence; dotted line: 5 Myr; dashed line: 15 Myr; dash-dotted line: 50 Myr ([29]).

This ISM structure invalidates the frequently made assumption of isotropy and homogeneity. If different stellar phases are associated with different dust columns, the composite SED becomes dependent on the dust morphology and its evolution, even for the assumption of a simple foreground screen model.

Next Contents Previous