A knowledge of the thermophysical and phase equilibria properties of fluid mixtures is essential in designing and optimizing chemical processes. For instance, most industrial applications include separation processes (distillation, absorption etc.) where phase equilibrium data are required from the early stages of process design. However, the experimental data needed to build thermodynamic models are scarce for complex molecules and experimental measurements are extremely costly and time consuming. Thus, predictive models such as the popular UNIFAC group contribution technique [1] have an important role to play in process design. Although UNIFAC is applicable to a very large variety of mixtures, it does not allow an accurate description of large multifunctional molecules. Furthermore, it does not give access to derived mixture properties such as caloric values and it is not suitable for high-pressure applications as no pressure dependence is included in the model. A group contribution approach based on the SAFT equation of state [2] has been presented by Tamouza et al. [3]. Despite the fact that this method gives good results for several classes of compounds, it models the molecules as chains of homonuclear segments, which does not reflect the real molecular structure of compounds; furthermore, the averaging procedure employed in developing the group contribution expressions for the intermolecular parameters means that the identity of the group is lost within the final molecule and the approach is difficult to apply to mixtures without the use of additional binary interaction parameters. To address this issue, we have developed a predictive equation of state by modifying the molecular-based SAFT-VR equation of state, extending it to treat model molecules which are formed by heterogeneous segments. In our generalised SAFT-gamma approach, molecules are modelled from chemical groups. Each group is characterized by (different) size parameters, and range and energy parameters of the attractive dispersion interaction. In the case of associating compounds the groups are also characterised by a number of bonding sites with additional association energy and association range parameters. We have investigated and implemented the changes within the SAFT-VR framework in order to account for the different molecular groups. A key feature of the new SAFT-gamma equation of state is that the fitting of the group binary interaction parameters does not require any mixture data, provided data are available for molecules containing all combinations of the groups of interest. To obtain group parameters, we have interfaced our equation of state implementation and an in-house vapour-liquid phase equilibria routine to a commercial parameter estimation tool (gPROMS). Parameters for a number of groups have been fitted to an extensive database of VLE data. The homologous series that can currently be treated by the investigated groups include n-alkanes, branched alkanes, aromatics, alcohols, alkenes, etc. It is shown that not only is a good fit obtained for the compounds included in the database, but also that satisfactory results arise when we extrapolate to compounds outside the database. In addition, the parameters obtained from pure component fitting are applied in a predictive manner to treat selected binary mixtures and excellent results are obtained.

[1] A. Fredenslund, R.L. Jones and J.M. Prausnitz. Group contribution estimation of activity coefficients in nonideal liquid mixtures. AIChE Journal, 21: 1086-1100, 1975.

[2] A. Gil-Villegas, A. Galindo, P. J. Whitehead, S. J. Mills, G. Jackson, and A. N. Burgess. Statistical associating fluid theory for chain molecules with attractive potentials of variable range. Journal of Chemical Physics, 106:4168-4186, 1997.

[3] S. Tamouza, J. Philippe Passarello, P. Tobally, and J. Charles de Hemptinne. Group contribution method with SAFT EOS applied to vapor liquid equilibria of various hydrocarbon series. Fluid Phase Equilibria, 222-223:67-76, 2004.