Esterification with cation exchange resin catalysts: The concept of an active surface layer
Advancing the chemical engineering fundamentals
Catalysis (T2-13P)
Keywords: esterification, cation exchange resin, catalysis, kinetic modeling
Esterification of acetic acid with methanol was performed in a batch reactor over a wide range of initial reaction mixture concentrations at 50°C and 100°C respectively. All the experiments were repeated with two different catalysts: Amberlyst 15, a fully sulphonated cation exchange resin and Amberlyst 46 which is only surface sulphonated (i.e. negligible amount of acid sites in the gel phase compared to those on the surface of the micro beads). Kinetic modeling of the experimental data shows that the reaction rate – based on the number of acid sites – of the Amberlyst 15 catalysed experiments is always only 15% of those catalysed by Amberlyst 46 at any given mixture composition. Since the reactions catalysed by Amberlyst 46 were proven to represent intrinsic reaction kinetics, the data indicates a constant effectiveness factor for Amberlyst 15 for this specific reaction. This observation indicates that internal diffusion in the micro beads of the catalyst is negligible. Therefore only the acid sites at the surface of the micro beads participate in the reaction, while the reactants have no access to most, if not all, of the acid sites in the gel phase – regardless of the mixture composition. This despite the fact that significant differences in resin swelling have been observed in different reaction mixtures, suggesting that access to the active gel phase may be influenced by the composition of the reaction mixture. The concept of a narrow active layer at the surface of the micro beads is further supported by the fact that the same activation energy (≈ 51 kJ/mol.K) is calculated for the two different catalysts – a definite indication of the absence of internal diffusion effects. It is also shown that as long as an applicable thermodynamic model is used to describe the activities of the components in the reaction mixture, there is no need to model the reaction kinetics with a more complicated model than the pseudo-homogenous one. With the NRTL activity coefficient model, a pseudo-homogenous rate expression represented the experimental data with an AARE of less than 5% which is comparable to errors obtained with Langmuir-Hinshelwood rate expressions for this system. More complicated models that take the absorption of reactants and products and subsequent swelling of the resin phase into account can therefore not be justified.
Presented Wednesday 19, 13:30 to 15:00, in session Catalysis (T2-13P).
