The self-assembly and structure of carbonaceous particles are investigated using molecular modelling methods. This provides a deeper understanding of molecular interactions relevant to pollutant formation and growth in combustion processes and other carbon-based applications. The existing soot particle model, a cluster containing planar pericondensed polycyclic aromatic hydrocarbons (PAHs), is extended to include PAHs of varying sizes. The resulting nanostructures show that the classic core-shell morphology reported experimentally for mature soot particles is not energetically feasible if only considering physical interactions between PAHs. It is proposed that young soot particles present the inverse molecular size partitioning. A detailed survey of the surface properties of heterogeneous PAH clusters is conducted, identifying composition-, size- and temperature-dependent behaviours. A novel stochastic global optimisation method, the Sphere Encapsulated Monte Carlo method, is also developed to allow minimum energy structures of large aromatic systems to be determined at considerably less computational expense than existing methods. The properties of curved PAH molecules are then investigated, and it is hypothesised that their enhanced electronic interactions could play a role in soot particle nucleation. A new intermolecular potential, curPAHIP, is developed to allow the simulation of curved PAHs. Subsequent dynamic clustering studies show that there is a significant increase in particle formation for systems containing curved PAHs and cations, suggesting the importance of these interactions in combustion processes. Further work investigates the structure of clusters containing curved PAHs, and the corresponding influence of cluster size, molecule size and curvature, molecular ratio, and presence of ions. This work develops computational tools useful for examining large systems of aromatic molecules as well as those containing curved species. Detailed studies on nanoparticle nucleation, structure, and surface properties provide valuable information on self-assembly processes crucial to understanding the production and properties of carbonaceous nanoparticles.