(321at) Understanding the Phase Behavior of Aqueous Hydrogen Fluoride Mixture by Incorporating Self and Cross Association Patterns

The phase behavior of aqueous hydrogen fluoride (HF) mixture has been studied from the 1940s. Munter and co-workers reported an azeotrope with a composition of 38.26 wt % HF and a maximum boiling temperature of 385 K [1]. The presence of this extreme negative azeotropy indicates the extent of the large non-ideality of this system [2].The corrosive nature of HF and its complex interactions with water have limited the development of experimental studies as well as robust thermodynamic models. Earlier studies have failed to provide a reasonable description of this mixture because of the complex association interactions between these compounds, which were not adequately modeled. We explore these interactions in order to deliver a new model developed on the basis of association patterns. Pure HF is modeled using more than 12 models that allow formations of different physically meaningful oligomers with different distribution schemes (1-2, 1-6, 1-2-6 etc), where the 1-2 model allows formation of monomers and dimers and likewise. The parameters for these pure component models were developed by correlating the phase co-existence properties of pure HF. Several other pure component properties such as ΔH_vap, C_P, ρ etc., are predicted using these models. Spectroscopic studies indicate the predominance of hexamer in HF and as one expects the preliminary results indicate that the models that allows formation of hexamers shows high predictive ability (1-6, 1-2-6, 1-2-6-8 etc). However, in HF it is not only the existence of these oligomers but their distribution is important as well. The dominance of these association patterns and their distribution are understood based on the predictive ability of these models.

These pure component models are extended to the binary mixture with water. During the extension to mixtures, initially, the phase co-existence properties are correlated using different association patterns for the pure components (HF 1-2, 1-2-6, 1-2-6-8 etc and water 1-2) with no considerations for the strong association between them. Preliminary results indicate that these self association models improves the binary interaction parameter value that was reported in the earlier work (AEOS-VK) [3] that did not include association for water in the mixture. Additionally, the double azeotrope that was predicted by the AEOS-VK model was not present in these self association models. After this, the cross association is included via different association schemes. The association model that allows infinite self and cross association did not provide a significant improvement from the models with no cross association. However, in this mixture, as in the pure components, it is important to have both the existence of physically meaningful oligomers as well as the correct distribution schemes. The cross association patterns that are adopted in this work are both similar as well as different from the association patterns of the pure components. These models with self and cross association are used to predict properties such as bubble point curve, mixture density etc. The significance of these self and cross association patterns are studied and understood based on the binary interaction parameter values as well as the predictive ability of the model.

References:

(1) P. A. Munter, O. T. Aepli and R. A. Kossatz, "Hydrofluoric acid-Water and Hydrofluoricsilicic acid-Water", Ind.Eng.Chem, 39, 427, (1947)

(2) A. Galindo, P. J. Whitehead and G. Jackson, "Predicting the Phase Equilibria of Mixtures of Hydrogen Fluoride with Water, Difluoromethane (HFC-32) and 1,1,1,2-Tetrafluoroethane (HFC-134a) Using a Simplified SAFT Approach", J.Phys.Chem. B, 101, 2082, (1997)

(3) B. Baburao and D. P. Visco, "VLE/VLLE/LLE Predictions for Hydrogen Fluoride Mixtures Using an Improved Thermodynamic Equation of State", Ind. Eng. Chem. Res., 41, 4863-4872, (2002)