This thesis will concentrate on the effect that dust grains, within the interstellar medium of galaxies, have, on how we perceive the light emitted from these galaxies and how this effect can be modelled. In particular improvements and extensions made to, and applications made with, a popular spectrophotometric galaxy model, GRASIL, which includes the effects of dust, will be presented. We have interfaced the GRASIL model (Silva et al. 1998) with a detailed chemical evolution model, which also incorporates the evolution of the dust component (Calura et al. 2008), in order to investigate in detail how an evolving dust component affects the spectral energy distribution of different types of galaxies. Usually the chemical composition of dust within the GRASIL code, as well as in other spectrophotometric codes, is fixed to that observed in the Milky Way and the mass of dust calculated using a simple assumption. Consequently by studying the two approaches, we have assessed the importance of accurately following the evolution of dust and the possible errors if this is not done. In a related effort, we have also integrated an artificial neural network into the GRASIL model in order to reduce the computational time required for
the code. The model has already been used extensively to calculate the spectral energy distributions of semi-analytical galaxy formation codes, however, the computational time required for GRASIL to calculate the spectral energy distribution, is quite often the bottleneck of the whole operation, limiting the number of mock galaxies that can be processed in realistic applications to ∼ 104 . We present first results from this integration of the artificial neural network which should allow, in the near future, an efficient comparison of the spectral energy distribution of semi analytic models, which include a proper dust treatment, to large observational galaxy surveys. Finally we introduce an entirely new photometric galaxy dust model, GRASIL-3D. This model builds on the strengths of the GRASIL model, but allows for any arbitrary geometry for the stars and gas which make up the galaxy. This will allow our new model to be applied to galactic hydrodynamic simulations, which calculate spatially resolved structural and dynamic properties of galaxies, without substantial loss of information. Our code has been integrated with a smooth particle hydrodynamics code, P-DEVA and the method and results are discussed and necessary future improvements are highlighted.
