Abstract

Interstellar dust plays a major role in the transfer of radiation in galaxies, both through its microscopic optical properties and its macroscopic spatial distribution. Dust absorbs and scatters the stellar light, and re-emits it in the far-infrared (FIR) and sub-millimetre (submm) wavelength range, redistributing both the direction of the propagating photons, and their energy. The complexity of these processes is best described by radiative transfer (RT) models. In particular these models have been used to derive the intrinsic distributions of stars and dust in galaxies by decoding the panchromatic imaging observations of galaxies.

While RT models have been remarkably successful to date in performing this task, they have also encountered an energy balance problem between direct stellar light and dust re-emission, particularly in edge-on galaxies. This work attempts to resolve this long-standing issue by incorporating a treatment for the clumpy structure of the interstellar medium (ISM). Because the scale of individual clumps (∼ 1 pc) lies below the resolution of galaxy-scale models, a subgrid approach is adopted, treating the quiescent dust clouds as pseudo-dust grains, with equivalent optical and thermal emission properties. Dedicated RT calculations are performed to derive key quantities such as the absorption, scattering and extinction cross-sections. This enables the virtualisation of the macroscopic clump into a microscopic pseudo-grain that can be included alongside the existing dust model constituents. The addition of the pseudo-grain results in a flatter extinction curve. A library of clump emission spectral energy distributions (SEDs) is constructed for radiation fields of various colours and intensities.

With the full incorporation of the extinction and emission properties of clumps into the RT models, this new clumpy model is applied to the edge-on galaxy NGC 891. For the first time, the clumpy model is able to achieve a good energy balance, simultaneously fitting both the submm and Near Infrared (NIR) data.

The clumpy model is further applied to a small sample of eight galaxies, and the results are compared with those from the purely diffuse models. Overall the clumpy models produced improved fits and provide a consistent solution for both face-on and edge-on galaxies.

The clumpy models are characterised by a reduction in dust opacity—and therefore attenuation—compared to their purely diffuse counterparts. Thus, the maximum face-on optical depth in the B-band, max(τfB ), derived from the clumpy models is found to be lower by factors ranging from 1.3 to 2.8. Of the eight galaxies, two are found to be optically thick in their centres, two are found to be moderately optically thick, and four are found to be optically thin. As a consequence of the decreased attenuation, the clumpy models also predict lower star-formation rates (SFRs), reduced by factors between 1.2 and 2.0.