Critical Assessment of the Thermodynamics in Acidic Resin-Catalyzed Esterifications

Abstract

Resins are interesting catalysts for organic reactions, which are characterized by a remarkable swelling behavior upon exposure to liquids, typically resulting in selective sorption in the case of mixtures. Hence, when constructing kinetic models for reactions catalyzed by such resins, the component partitioning between the bulk liquid and polymer liquid and its effect on their thermodynamic activities have to be properly taken into account. A kinetic model is developed for acidic resin-catalyzed esterification, while comparing two scenarios for capturing the thermodynamics within the resin pores. For the specific case study on a single resin with a specific solvent, compensation effects between the calculated thermodynamic activities in the resin and the estimated rate coefficients render only minor performance differences between the two scenarios. Still, aiming at extending the scope toward other solvents and/or resins, it is deemed crucial to explicitly account for the resin's selective sorption and swelling behavior.

First principles prediction of NRTL binary interaction parameters for furfural derivatives

“…7,43 The NRTL equation is based on the two-liquid model of Scott and assumes 44 a non-randomness in liquid mixtures, similar to the assumption used by Wilson to describe the distribution of components at the local molecular level. [7][8][9] In a multicomponent mixture the activity coefficient of a component is calculated using Equations ( 2)-( 4). Multiplying the activity coefficient γ i with the mole fraction of component i in the binary mixture, gives the thermodynamic activity a i of that component, which is, among other things, used to simulate VLE's.…”

“…Hereto, the sum of squared residuals between the targeted activity coefficients and those calculated by the model was minimized using Athena Visual Studio. 7,43 The NRTL equation is based on the two-liquid model of Scott and assumes 44 a non-randomness in liquid mixtures, similar to the assumption used by Wilson to describe the distribution of components at the local molecular level. [7][8][9] In a multicomponent mixture the activity coefficient of a component is calculated using Equations ( 2)-( 4).…”

“…Therefore, the value of this parameter has been fixed to 0.3 as recommended by Renon and Prausnitz, and only a ij and b ij are obtained from the regression. 7,9 The regression was evaluated via the F-test for the 95% global significance of the regression, the 95% confidence interval for the statistical significance of the parameter estimates. Moreover, the binary correlation coefficients between the estimated parameters were calculated and evaluated.…”

“…The boiling temperature is measured using a thermocouple (3) located at the top of the heating chamber. The vapor rises further and is condensed (7) to be recirculated to the mixing chamber ( 6), whereas the liquid is immediately recirculated to the mixing chamber. This process continues throughout the experiment.…”

“…Kinetic modeling in turn allows for reaction condition optimization or catalyst and/or reactor design. [5][6][7] The thermodynamic non-ideality of mixtures can be described via several methods, that is, activity coefficient models such as Wilson,8 NRTL (non-random two-liquid) 9 or UNIQUAC (universal quasichemical), 10 equations of states like Peng-Robinson 11 or Hayden-O'Connell 12 and group contribution methods such as UNIFAC (UNIQUAC functionalgroup activity coefficients) 13 or GC-PC-SAFT (group-contribution perturbed-chain statistical associating fluid theory). 14 Most of these methods are implemented in commercial software packages like Aspen Plus 15 to allow straightforward process design.…”