Study On Evaluating Application Of UNIFAC And Its Modified Models

Data of vapor-liquid equilibrium is important in chemical base data, it is vital for the study of chemical theory and practical application, it is also convenient to calculate other physical properties. With the increase of compound's species and quantity, it is impossible to obtain all the needed data by experiment, thus the idea of estimation was presented. The group contribution method UNIFAC model was proposed for that. In order to promote the improvement and development of this model, and meet the different requests in thermodynamic properties calculation of all kinds of fluids, a more systematic study and evaluation of UNIFAC and its modified models was made.In this paper, the methods basing on group contribution to estimate the vapor- liquid equilibrium in literature were summarized. Then the widely used methods with high estimation accuracy were selected, which suited for the system of medium and small molecules, they were original UNIFAC model, UNIFAC(Larsen) model and UNIFAC(Dortmund) model. The thermodynamic properties of systems were estimated, which contained water, aromatic compound, heterocyclic compound, halogenated hydrocarbon, xylene isomer, polar compound and the systems at different temperatures, the results were compared with the data from literature, three models were evaluated.For the selected system, original UNIFAC model was more suited for the systems containing aromatic compound, xylene isomer, tetrachloromethane/benzene, benzene /furfural, p-xylene/furfural, thiophene/hexane, 1-methyl-2-pyrrolidinone/2-propanol. The estimating results are consistent with the experiment data. Among these systems, the result of system containing tetrachloromethane/benzene is best, the deviation of vapour-phase composition is 0.46%; this model is more accurate than other models for systems containing water/pyridine, ethyl acetate/water, hexane/fluorobenzene, trichlor- omethane/formic acid, acetaldehyde/furfural, ethyl formate/tert-butyl acohol, tetrarch- loroethylene/1-pentanol but the deviation is large, and the deviation is larger than 3%. For the selected system, due to lack of group parameters, UNIFAC(Larsen) modelcan't estimate the systems containing acetaldehyde/furfural, acetone/formic acid, ethyl formate/tert-butyl acohol, thiophene/hexane, 1-methyl-2-pyrrolidinone/2-propanol; This model is more suited for systems containing ethanol/benzene, hexane/benzene, heptane/benzene, 1-hexene/o-xylene, the system of tetrachloro-methane/benzene at 40℃and the system of methanol/benzene at 90℃, the results are acceptable, the deviations are 0.56%and 0.73%.For the selected systems, UNIFAC(Dortmund) model is more suited for the systems containing methanol/benzene, benzene/2-propanol, acetone/benzene, heptane /furfural, pyridine/decane, heptane/ethyl bromide, methyl iodide/tertrachloromethane, heptane/iodoethane, acetone/ethanol, dibromomethane/ethyl acetate, ethyl tertbutyl ether/ethanol, 2-methyl-1-propanol/octane. The deviation is within 3%, and the system of tetrachloromethane/benzene at 60℃, which deviation is 0.39%. The system of tetrachloromethane/benzene at 90℃and the system of methanol/benzene at 40℃, the results are better than others, but the deviation is large; this model can't estimate the systems containing thiophene/hexane, 1-methyl-2-pyrrolidinone/2-propanol, due to lack of group parameters.By comparison, we know that all the models have their merits and drawbacks. With the supplement of parameter matrix, the application of UNIFAC and its modified models is wider and wider, which are good at estimating nonpolar systems containing alkanes, symmetrical compound at the condition of low and moderate temperature, for some systems containing polar compounds, halogenated hydrocarbon, and isomer, the results are good, but it's bad for systems containing strong polar compounds and at the condition of high temperature. In addition, the distinguish isn't obvious for system containing isomer, thus, it's necessary to further modify the model, introduce new group, modify parameter and extend parameter matrix.