Robust prediction of chemical equilibria by group contribution

FRANCISCO A. SÁNCHEZ; SELVA PEREDA

Campinas

Conferencia; X Iberoamerican Conference on Phase Equilibria and Fluid Properties for Process Design (EQUIFASE 2022); 2022

Universidad Estatal de Campinas (UNICAMP)

Nowadays, biorefineries that produce first-generation biofuelsrequire public support from governments in most countries, if notall. Therefore, the integral use of biomass and the consequentdiversification of products is key to improve the economy of theexistent productive sites. This implies the production ofsecond-generation biofuels, as well as solvents and other value-addedcompounds. Optimal design of new production roadmaps requiresmathematical models able to assess yield and the technical-economicfeasibility of synthesis and purification processes. Unfortunately,at present, there are still plenty of under-characterized routes inthe milieu of biomass processing. For example, even though a greatdeal of experimental work has been done on esterification andtransesterification reactions, which play an important role inbiomass upgrading, still there are no robust models to fully predictsimultaneous chemical and phase behavior. As it is well-known, thiskind of reactions are reversible, hence maximizing yields requiresgood control of the effect of operating conditions on their chemicalequilibrium.Predictingchemical equilibrium requires, besides a tuned thermodynamic model,either knowledge of formation properties of all components involvedin the reaction, or experimental data on the chemical equilibriumthermodynamic constant (K). In many cases the formationproperties available, even in well-known thermodynamic databases, butshow important deviations. Hence, the most frequent approach,regardless the formation properties are available or not, it is theexperimental measurement of K and its correlation with athermodynamic model. It is remarkable that, in many cases,incongruent results are observed between the correlated Kvalues reported in the literature for the same reaction, which meansthat the K values reported are model-dependent. In this work, we challenge the GroupContribution with Association Equation of State (GCA-EOS), previouslytuned to only non-reactive phase equilibrium data, to predict thesimultaneous chemical and phase equilibria of n-alkanolacetylation reactions. We also show the importance of assessing thetrends of the formation properties within the homologous series todevelop a robust model. Our approach also allows predicting chemicalequilibrium when not all the formation properties are available.