Commonly used P-V-T equations of state (EOS) are continuous functions, allowing the simultaneous description of some (or most) thermodynamic properties for both vapor and liquid phases. Such EOS's are well-suited to modeling high pressure vapor-liquid equilibrium (VLE). Generically speaking, the PVT form of an EOS enables the description of high pressure behavior, while the mixing rules and interaction parameters control the description of mixing. In reality, successful implementation of an EOS requires that its mathematical (PVT) form be commensurate with the corresponding mixing rules for its parameters.

Simple two-parameter cubic EOS's are popular choices since they permit fast process calculations. However, they are limited in the number of interaction parameters that they can accommodate which, in turn, restricts their ability to describing non-ideal systems. Consequently, a majority of systems involving organic substances at low pressure, especially those exhibiting vapor-liquid-liquid equilibria (VLLE), are correlated with liquid activity coefficient models (along with a simplified EOS model for the vapor phase).

This study involves the description of non-ideal phase equilibria, and recognizes that, in a significant number of systems, non-ideality is concentrated in the liquid phase. It combines this notion of liquid phase non-ideality along with the dual phase design of P-V-T EOS's, to investigate the feasibility of correlating VLE and VLLE in non-ideal systems. Furthermore, it is the belief of this author that this objective necessitates that any EOS employed for this purpose must predict pure component and mixture liquid densities correctly, without sacrificing the accuracy of other properties (e.g., vapor pressures). This requirement precludes the use of most two-parameter cubic EOS formulations. An essential element of this investigation, therefore, is the introduction of a “higher-than-cubic” multi-parameter EOS with corresponding higher-order virial-based mixing rules. This is a proof-of-concept study which does not focus on developing this EOS or mixing rules, but simply employs these as vehicles to correlate VLE and VLLE of non-ideal systems hitherto non amenable for correlation via EOS's.

Several binary and ternary systems are investigated, and their predictions are compared with experimental data. The results are quite encouraging, and, in some cases, very good. With some fine-tuning, this could be considered a viable approach to correlating the large body of non-ideal VLE and VLLE data available in the literature.

An extension of the above concept is the potential simultaneous correlation of phase equilibrium, excess volume and excess enthalpy data. Some systems with only a few VLE / VLLE data points are simultaneously correlated with mixture enthalpy / volumetric data to reliably extend VLE predictions over a wider concentration range.