Application of the zeroth approximation of the DISQUAC model to cyclohexane + n-alkane mixtures using different combinatorial entropy terms

Abstract

Literature data on molar excess functions, Gibbs energy G E, enthalpy H E, and heat capacity C p E, on activity coefficients γ i ∞, and partial molar excess enthalpies H i E,∞, at infinite dilution and on solid-liquid equilibria, SLE, of the cyclohexane + n-alkane mixtures are examined on the basis of the zeroth approximation of the DISQUAC group contribution model. The model provides a quite satisfactory description of the thermodynamic properties for the mixtures under study, although the symmetry of the calculated excess functions differs from the experimental one for systems containing long-chain n-alkanes. This may be due to the so-called Patterson effect. The influence of different combinatorial entropy terms (Flory-Huggins, Stavermann-Guggenheim or Kikic equations) on the prediction of thermodynamic properties such as G E, ln γ i ∞ and SLE is also examined. H E, C P E or H i E,∞ are represented by an interactional term only. The results calculated using the Flory-Huggins term are slightly better than those obtained applying the Stavermann-Guggenheim equation. Results based on the Kikic expression are poorer than those given by Flory-Huggins, particularly at high concentration of cyclohexane in systems containing the longer n-alkanes. So, the Kikic equation leads to poorer results for ln γ 2 ∞ for these systems. SLE predictions are determined mainly by the physical constants of the pure compounds. So, essentially they do not depend on the combinatorial term used. A comparison between the zeroth approximation of DISQUAC and the modified UNIFAC model (Lyngby version) is also presented. Such comparison shows that both methods lead to similar results; although the latter gives poorer predictions on the temperature dependence of the excess functions than the former. On the other hand, the number of interaction parameters needed in modified UNIFAC is larger than when the zeroth approximation of DISQUAC is applied and, more important, they change with the number of carbon atoms of the n-alkane in a rather erratic way for the first members of the series. This makes the predictive task of UNIFAC more difficult.