Massive star formation in galaxies: radiative transfer models of the UV to millimetre emission of starburst galaxies

Abstract

We present illustrative models for the UV to millimetre emission of starburst galaxies which are treated as an ensemble of optically thick giant molecular clouds (GMCs) centrally illuminated by recently formed stars. The models follow the evolution of the GMCs owing to the ionization‐induced expansion of the H ii regions and the evolution of the stellar population within the GMC according to the Bruzual & Charlot stellar population synthesis models. The effect of transiently heated dust grains/PAHs on the radiative transfer, as well as multiple scattering, is taken into account. The expansion of the H ii regions and the formation of a narrow neutral shell naturally explain why the emission from PAHs dominates over that from hot dust in the near‐ to mid‐infrared, an emerging characteristic of the infrared spectra of starburst galaxies. The models allow us to relate the observed properties of a galaxy to its age and star formation history. We find that exponentially decaying 107–108 yr old bursts can explain the IRAS colours of starburst galaxies. The models are also shown to account satisfactorily for the multiwavelength data on the prototypical starburst galaxy M82 and NGC 6090, a starburst galaxy recently observed by ISO. In M82 we find evidence for two bursts separated by 107 yr. In NGC 6090 we find that at least part of the far‐infrared excess may be due to the age of the burst (6.4×107 yr). We also make predictions about the evolution of the luminosity of starbursts at different wavelengths which indicate that far‐infrared surveys may preferentially detect older starbursts than mid‐infrared surveys.

Dusty star-forming galaxies at high redshift

“…When discussing these models, it's important to keep in mind that the accuracy or applicability of these models cannot be tested with data since it simply does not exist in enough detail to disentangle effects of geometry, distribution, optical depth, etc. It's also important to note that many have done work in this area, particularly modeling radiative transfer in local starburst populations to generate SEDs (Efstathiou & Rowan-Robinson, 1995;Efstathiou et al, 2000;Efstathiou & Siebenmorgen, 2009;Nenkova et al, 2002;Dullemond & van Bemmel, 2005;Piovan et al, 2006;Nenkova et al, 2008;Takagi et al, 2003;Fritz et al, 2006;Hönig et al, 2006;Schartmann et al, 2008), but here we try to focus on the techniques which have been most commonly employed for SED fitting of high-z dusty starbursts 11 . Silva et al (1998) developed the Grasil code to model galaxy emission by explicitly accounting for dust absorption and emission from the ultraviolet through to the far-infrared.…”

Panchromatic radiation from galaxies as a probe of galaxy formation and evolution

“…Rowan-Robinson & Efstathiou (2009) have shown that the suite of starburst models of Efstathiou et al (2000) and AGN dust tori models of Efstathiou & Rowan-Robinson (1995) successfully explain the distribution of starbursts and AGN in the diagnostic diagram proposed by Spoon et al (2007), 9.7 µm silicate depth versus 6.2 µm PAH equivalent width (Fig 1). While there is aliasing of the models in this diagram, since different combinations of starburst and AGN dust torus models can predict the same point in the diagram, the model sequences do populate the distribution of observed points well.…”

“…Krugel & Siebenmorgen 1994, Silva et al 1998). Efstathiou, Rowan-Robinson & Siebenmorgen (2000) made such models slightly more realistic by following the evolution of a spherically symmetric HII region within a dense molecular cloud. They assume an embedded phase, lasting for 10 7 years.…”

“…At 10 8 years the emergent spectrum is virtually indistinguishable from that of the general optically thin interstellar medium (the 'cirrus'). The time-sequence of SEDs calculated by Efstathiou et al (2000) poses the interesting question: can we identify starbursts of different ages observationally (see section 8 below)? Subsequent work on radiative transfer models for starbursts includes the models of Takagi et al (2003), which includes distributed stars and dust satisfying a King law; the programme of Dopita et al (2005, Groves et al (2008) which incorporate detailed HII region physics; and the large library of parameterised models generated by Siebenmorgen & Krugel (2007).…”

Should we believe the results of ultraviolet–millimetre galaxy spectral energy distribution modelling?

“…One method is to perform radiative transfer calculations assuming some simple galaxy geometry (e.g. Silva et al 1998;Efstathiou et al 2000;Granato et al 2000;Popescu et al 2000;Tuffs et al 2004;Siebenmorgen & Krügel 2007;Groves et al 2008;Michałowski et al 2010a,b). Alternatively, one can take a more empirical approach and treat the FIR SED as a sum of modified blackbodies (e.g.…”