Vapor−liquid equilibria (VLE), liquid−liquid equilibria (LLE), and vapor−liquid−liquid equilibria (VLLE) for systems involving highly nonideal components, namely, water, alcohols, alkanes, ketones, aldehydes, esters, and ethers, were investigated to evaluate the perturbedchain statistical associating fluid theory equation of state (PC-SAFT EOS) and two widely used activity coefficient models, that is, the universal quasichemical (UNIQUAC) and UNIQUAC functional-group activity coefficients (UNIFAC). Parameters used for the PC-SAFT EOS were taken from literature or estimated in this work, while those for UNIQUAC and UNIFAC were from commercial process simulator Aspen plus 8.4. It was found that all the three models yield reliable correlations/predictions for VLE calculations. However, UNIQUAC and UNIFAC were observed to be unreliable for LLE and VLLE calculations despite successful reproductions of experimental data in some cases. The calculated results deviate significantly from experimental data in many cases. Particularly, both models predict artificial liquid− liquid phase splitting for a number of miscible mixtures. Nonetheless, PC-SAFT EOS with the use of a single set of parameters reproduces experimental data quantitatively in most cases and provides reasonably accurate results in all other cases. This remarkable performance of PC-SAFT EOS potentially eliminates the need for various thermodynamic models and consequently the need for selecting a thermodynamic model when performing phase equilibria calculations using commercial software. This is important for practitioners, since (1) it remains unclear to select an appropriate model from the available models of a process simulator or thermodynamic package for a given phase equilibria calculation despite the presence of some type of rule of thumb and (2) it is also likely that none of the existing models is sufficiently accurate. In addition, it was shown that both purecomponent parameters and binary interaction parameters for the PC-SAFT EOS are well-behaved for a homologous series, which allows for parametrization for weakly characterized components by interpolation or extrapolation, and consequently, facilitates the development of a practical tool for phase equilibria calculations.

As an alternative route to produce liquid fuels, Fischer–Tropsch synthesis (FTS) has received considerable attention. Phase equilibria have been one of the key issues in the design and operation of not only an FTS reactor, but also relevant downstream processing, e.g., stepwise condensation of FTS products, wastewater treatment, and the recovery of α-olefins. These issues are extremely complex due to a wide spectrum of FTS products and the presence of noncondensable gas, nonpolar components, associating components, and highly polar components. To address these issues, the perturbed-chain statistical associating fluid theory (PC-SAFT) equation of state (EoS) was evaluated using phase equilibrium data of binary and ternary mixtures containing water, alcohols, hydrocarbons, and noncondensable gases. The parameter values are available via a three-step parametrization strategy established in our previous work [Ind. Eng. Chem. Res.2018573014]. A couple of popular benchmark models, namely the Soave–Redlich–Kwong (SRK) EoS with the modified Huron–Vidal second-order mixing rule (SRK/MHV2) and the predictive SRK (PSRK) EoS, were selected for comparison to put the performance of the PC-SAFT EoS into perspective. It was found that the PC-SAFT EoS gives reliable correlations/predictions in all the studied cases, whereas the other two models fail to do so. The three models were also employed to simulate hot and cold traps where the stepwise condensation of the reactor effluent occurs. It was found that all three models yield satisfactory results for the paraffin distribution of the liquid stream leaving the hot trap, while the SRK/MHV2 and PSRK models are less accurate for the olefin distribution, particularly at high temperature. It was also found that all three models are capable of predicting the composition of the two liquid streams (namely organic and aqueous) leaving the cold trap except that the SRK/MHV2 model fails to accurately reproduce the distribution of alcohols in the two liquid streams. This study advocates the practical use of the PC-SAFT EoS by its application to a complex, industrially important FTS process and a comprehensive evaluation, together with the two popular benchmark models.