The separation of long chain isomers or mono- and diester is a special challenge as distillation is very elaborative because of the small separation factor and extraction processes do not match the high purity demands. In principal this separation task can be achieved by processes based on crystallization or liquid chromatography. For the design of these processes thermodynamic data is necessary, but this data cannot be measured as the branched molecules are not commercially available in the needed purity. Furthermore a lot of experiments are necessary to find an appropriate adsorption material for the liquid chromatography.The goal is to develop the physical knowledge of a process model for the separation of long chain molecules with different architectures. To consider the influence of branching on thermodynamic phase behavior the Lattice Cluster Theory (LCT) is used to consider architectural influence of linear and branched molecules. This framework forms the basis for the modeling of crystallization of linear and branched molecules. The results are verified by selected experiments. The combination of the LCT with the density functional theory aims to the development of a method to predict adsorption isotherms of mixtures based on adsorptions isotherms of pure components.