An improved cubic equation of state for accurate modelling of phase equilibria in nonideal mixtures

Numerous improvements have been proposed for common cubic equations of state such as Peng-Robinson and Redlich-Kwong. These modifications are introduced to improve phase equilibrium calculations of the equation of state by providing a better representation of pure-component vapor pressures or nonideal mixture properties. The former improvement is usually achieved by modifying the temperature dependence of the attractive pressure term. The latter improvement is attained by modifying the conventional quadratic mixing rules.^ In this work, the correlation of vapor pressures have been substantially improved by introducing a new form for the temperature dependence of the attractive pressure term; an exponential function which is fine-tuned by a ratio of two simple polynomials. This model has been compared with six other temperature expressions reported in the literature. The comparison shows that the proposed temperature dependence, with two or three adjustable parameters, gives a remarkable improvement in the representation of pure-component vapor pressures.^ The adjustable parameters of the proposed model are empirically estimated from pure-component P-T data. This model also provides reliable vapor pressure predictions when extrapolated to very low reduced temperatures and gives physically reasonable values for the attractive term at high reduced temperatures and, hence, reliable prediction of fluid properties.^ In addition to the introduction of the new temperature dependence, a new mixing rule has been proposed to enable the equation of state to accurately predict vapor-liquid and liquid-liquid equilibria in ideal as well as highly nonideal mixtures. The adjustable parameters of the proposed mixing rule are empirically determined from either vapor-liquid binary data or mutual solubility data of partially miscible liquids. The mean and maximum deviations from experimental pressure and vapor composition data are compared with those calculated with standard activity-coefficient models, such as Margules, Van Laar, Wilson, NRTL, and UNIQUAC. The results of these comparisons show that, even for highly nonideal binary and ternary systems, the improved equation of state is superior to, or at least as good as, the activity coefficient models in predicting phase equilibria.