(257al) Thermodynamic Model for the Tri-n-Butilcitrate, Butanol, Citric Acid and Water System

Authors: 
Orjuela, A., Universidad Nacional de Colombia
Santaella, M., Universidad Nacional de Colombia
Suaza, A., Universidad Nacional de Colombia
Rivera, J. L., Universidad Nacional de Colombia
Berdugo, C., Universidad Nacional de Colombia

Some petrochemical plasticizers, mainly phthalates, have been recognized to be harmful for the human health. Although the plasticizer market continues to grow, traditional phthalic acid esters have started to be replaced by some biobased plasticizers which are generally recognized as safe (GRAS) and derived from renewable feedstock. Citric acid esters (with or without acetylation) belong to this group of plasticizers and they are already used in the fabrication of toys, medical devices, food packaging, cosmetics, and many other applications.

Several difficulties with the current production process of commercial citrates have been identified: the reaction time is long, large excess of alcohol is required, and the separation processes are energy intensive. Therefore, the explorations of new production technologies through process intensification have been proposed to reduce production costs. In order to evaluate process feasibility, computer simulation is required, and good thermodynamic models for reaction and separations design are needed

In this work the phase equilibria in mixtures with tri-n-butilcitrate, butanol, citric acid and water has been evaluated. Isothermal equilibrium data from Liquid-Liquid, Liquid-Solid, Liquid-Liquid-Solid and Vapor-Liquid have been experimentally determined. Quantification in samples was performed by HPLC, GC and Karl-Fisher methods. Wisniak test was used for the Liquid Vapor data consistency, while the Othmer-Tobias correlation was used for ternary Liquid-Liquid analysis. Complementary equilibrium data has been found from a literature review.

Uniquac and NRTL binary parameters for the complete system were obtained by data regression. Fitting has been done using a genetic algorithm and has been compared against the Aspen Plus fitting tools. Different set of parameters have been found to fit better specific experimental data. Nevertheless the final parameters reproduce adequately the altogether experimental observations and may be used in the design of integrated units for citric acid ester production.