Phase equilibrium engineering in biorefinery reactive systems: n-alkanol acety-lation

Abstract

In recent years, the need to diversify the demand for liquid fuels has promoted the research in new and advanced biofuels. In addition, the valorisation of available renewable resources also calls for other products like solvents and value-added chemicals, to make the whole process competitive in today’s market. In this context, the synthesis paths of various biobased value-added products involve reversible chemical reactions; therefore, process development and optimization require modelling the chemical equilibrium (CE) of these systems, as well as simultaneous chemical and phase equilibrium (CPE). There are many other applications of such models, for example in the biphasic dehydration and/or hydrogenation reactors or in the well-developed reactive distillation units, where phase separation is design to improve the reaction yield. In this sense, esterification and transesterification reactions are found in many productive pathways of value-added products; such as, biodiesel, glycerol acetates, or valeric biofuels, to name a few. In any case, CE calculation requires, a priori, the formation Gibbs energy of each component involved in the reaction (Δgf) or the equilibrium constant (K), which is equivalent to the former. However, this information is not always available and hence, K is frequently correlated to CE experimental data (Bucalá et al., 2006; Schmid et al., 2008). There are other examples that, even though K is available, it is disregarded due to parametric sensitivity and again K is correlated with CE experimental data using a thermodynamic model tuned to phase equilibrium data (Grob & Hasse, 2014; Riechert et al., 2015). In particular, Riechert et al. discuss the influence of the physical formalism of the thermodynamic model, and conclude that those models that take into account the molecular phenomena that occur in the multicomponent mixture achieve a better K correlation. However, they do not contrast their results against literature values of K or Δgf. In this work, we challenge the Group Contribution with Association Equation of State (Sánchez et al., 2011) (GCA EOS) to predict the CPE of various acetylation reactions of nalkanols based on Δgf ig reported in literature and experimental databases (Rowley et al., 2003), i.e. without correlating CE data. We select this model system because of the large number of experimental data available and the advantage of using a group contribution model to assess homologous series. Despite of the parametric sensitivity of these systems CE, we show that the GCA EOS can predict CPE of this homologous series.