Modelling the ultraviolet-to-infrared spectral energy distributions of galaxies

The spectral energy distributions of galaxies at ultraviolet, optical and infrared wavelengths are clues to both the radiation produced by stellar populations and the effect of gas and dust in the interstellar medium (ISM) on this radiation. Spectral interpretation studies are therefore crucial to understand how galaxies formed and evolved. Observationally, combined ultraviolet, optical and infrared data are now becoming available for large samples of galaxies. So far, these have focused mainly on the local galaxy population, but future deep surveys will provide observations of large galaxy samples at higher redshifts. To extract constraints on the stellar populations and ISM of galaxies from these multi-wavelength observations requires the consistent modelling of the emission by stars, gas and dust.
In this thesis, we present a simple, largely empirical but physically motivated model, which is designed to interpret consistently multi-wavelength observations from large samples of galaxies in terms of physical parameters, such as star formation rate, stellar mass and dust content. This model relies on an existing angle-averaged prescription to compute the absorption of starlight by dust in stellar birth clouds and in the ambient interstellar medium (ISM) in galaxies. We compute the spectral energy distribution of the power re-radiated by dust in stellar birth clouds as the sum of three components: a component of polycyclic aromatic hydrocarbons (PAHs); a mid-infrared continuum characterising the emission from hot grains at temperatures in the range 130–250 K; and a component of grains in thermal equilibrium with adjustable temperature in the range 30–60 K. In the ambient ISM, we fix for simplicity the relative proportions of these three components to reproduce the spectral shape of diffuse cirrus emission in the Milky Way, and we include a component of cold grains in thermal equilibrium with adjustable temperature in the range 15–25 K.
Our model is both simple and versatile enough to allow the derivation of statistical constraints on the star formation histories and dust contents of large samples of galaxies using a wide range of ultraviolet, optical and infrared observations. We illustrate this by deriving median-likelihood estimates of a set of physical parameters describing the stellar and dust contents of local star-forming galaxies from three different samples. The model reproduces well the observed spectral energy distributions of these galaxies across the entire wavelength range from the far-ultraviolet to the far-infrared, and the star formation histories and dust contents of the galaxies are well constrained.
A main advantage provided by this model is the ability to study the relation between different physical parameters of observed galaxies in a quantitative and statistically meaningful way. Our analysis of star-forming galaxies from the Spitzer Infrared Nearby Galaxy Sample (SINGS) and the Spitzer-SDSS-GALEX Spectroscopic Survey (SSGSS) reveals that the mid- and far-infrared colours of galaxies correlate strongly with the specific star formation rate, as well as with other galaxy-wide quantities connected to this parameter, such as the ratio of infrared luminosity between stellar birth clouds and the ambient ISM, the contributions by PAHs and grains in thermal equilibrium to the total infrared emission, and the ratio of dust mass to stellar mass. These correlations provide important insight into the link between star formation and ISM properties in galaxies.
We investigate further the relation between star formation activity and dust content in galaxies by assembling a large sample of 3321 galaxies with available observations at ultraviolet (GALEX), optical (SDSS) and infrared (IRAS) wavelengths. We find that the star formation rate correlates remarkably well with galaxy dust mass over four orders of magnitude in both quantities. This allows us to provide a simple empirical recipe to estimate the total dust mass of galaxies from the star formation rate. We compare our findings with the predictions from recent models of the chemical evolution of galaxies. We also compare the relations between the specific star formation rate, the ratio of dust luminosity to stellar mass and the ratio of dust luminosity to star formation rate obtained from our analysis with those predicted by state-of-the-art cosmological simulations of galaxy formation.
The model presented in this thesis can be straightforwardly applied to interpret ultraviolet, optical and infrared spectral energy distributions from any galaxy sample.