Development of Bio-UNIFAC Group Contribution Model to Predict the Phase Equilibria of Bio-Oil-Related Mixtures

Group contribution models such as UNIFAC-DMD are widely used to predict phase equilibria by tuning their interaction parameters with various thermodynamic data. However, because vapor–liquid equilibrium (VLE) data disproportionately dominate optimization, these models naturally predict VLE more accurately than liquid–liquid equilibrium (LLE) or solid–liquid equilibrium (SLE). To ensure reliable predictions across data types, distinct interaction parameters must be derived for each equilibrium type. Accordingly, the Bio-UNIFAC model was developed for emerging bio-oil-relevant compounds, with interaction parameters optimized using high-quality binary VLE and LLE data involving 296 molecules selected on the basis of GC–MS data from pyrolyzed biomass samples. A new main group, AC2H2O (G100), with AC2H2O, AC2HO, and AC2O subgroups, was introduced for furan-containing bio-oil mixtures. Parameter optimization used 230 572 binary VLE points from 7356 literature data sets. The model was tested using binary and ternary VLE and LLE data containing bio-oil components, demonstrating its suitability for the reliable phase-equilibrium prediction of bio-oil-based mixtures.