MOQUAC, a new expression for the excess Gibbs free energy based on molecular orientations

The three-dimensional structure of molecules determines if effects like steric hindrance or multiple contact points upon a molecular contact occur. These effects are especially important for molecules with several strongly interacting functional groups, since contacts between such groups have a strong influence on the behavior of liquid systems. The three-dimensional structure of molecules also needs to be considered if stereoisomers are to be distinguished. Bio-based molecules often have several strongly interacting functional groups and show enantiomerism. Because of the gradual shift towards greater use of renewable resources in the chemical industry, the consideration of the molecular geometry in GE-models will become increasingly important. However, for all state-of-the-art GE-models, assumptions are made that lead to the loss of information about the three-dimensional molecular structure. Therefore, in this work a new model is derived that does not need such assumptions. The new model MOQUAC described in this thesis is based on a quasi-chemical approach. For this model the orientation of molecules upon a molecular contact is considered, which enables consideration of the three-dimensional structure of molecules. By comparison to results from lattice simulations, it is shown that MOQUAC can describe systems with coupled interactions. Additionally, a model to predictively describe the interaction energy between real molecules as a function of their orientation is derived. It is shown that MOQUAC can be used together with this interaction-energy model to predict the behavior of systems of real components. In addition to a new expression for the residual contribution to the Gibbs energy, MOQUAC consists of a physically-founded improvement of the UNIQUAC combinatorial term. For the combinatorial contribution, a standard segment is used to determine the structural parameters r and q. It is generally assumed that the choice of the standard segment and the influence of the absolute value of q are negligible. The standard segment area, however, does not cancel out in the model equation and is therefore a model parameter. The improvement of the combinatorial term consists of a fit of the size of the standard segment to carefully selected experimental data. It is shown that the new standard segment significantly improves the performance of the UNIQUAC combinatorial term and that the physically founded improved model performs at least as well as empirical modifications of the original term.