LAUSR

Levulinic acid was esterified with methanol, ethanol, and 1-butanol with the final goal to predict the maximum yield of these equilibrium-limited reactions as function of medium composition. In a first step, standard reaction data (standard Gibbs energy of reaction ΔRg0) were determined from experimental formation properties. Unexpectedly, these ΔRg0 values strongly deviated from data obtained with classical group contribution methods that are typically used if experimental standard data is not available. In a second step, reaction equilibrium concentrations obtained from esterification catalyzed by Novozym 435 at 323.15 K were measured, and the corresponding activity coefficients of the reacting agents were predicted with perturbed-chain statistical associating fluid theory (PC-SAFT). The so-obtained thermodynamic activities were used to determine ΔRg0 at 323.15 K. These results could be used to cross-validate ΔRg0 from experimental formation data. In a third step, reaction-equilibrium experiments showed that equilibrium position of the reactions under consideration depends strongly on the concentration of water and on the ratio of levulinic acid: alcohol in the initial reaction mixtures. The maximum yield of the esters was calculated using ΔRg0 data from this work and activity coefficients of the reacting agents predicted with PC-SAFT for varying feed composition of the reaction mixtures. The use of the new ΔRg0 data combined with PC-SAFT allowed good agreement to the measured yields, while predictions based on ΔRg0 values obtained with group contribution methods showed high deviations to experimental yields.