Development and modelling of a new catalytic distillation process

Abstract

Catalytic Distillation (CD) is a relatively new process used by the petrochemical industry. The process involves placing a heterogeneous catalyst inside of a distillation column so that a chemical reaction and product separation can take place simultaneously in the same piece of equipment. The first commercial application of CD was introduced in 1980 for the production of methyl tertiary-butyl ether (MTBE), and the success of this process has created a great deal of interest in using CD as a more generally applied reaction technique.

This thesis describes a new application of CD for the production of diacetone alcohol (DAA) via the base-catalyzed aldol condensation of acetone. This reaction is equilibrium limited to only 4.3 wt% conversion of acetone to DAA at 54*C. As well, DAA can dehydrate to mesityl oxide (MO). MO can continue to react with acetone to form heavier byproducts. DAA is a very useful chemical intermediate and an environmentally friendly cleaning solvent. Therefore, this reaction is commercially relevant, while providing an opportunity to study product selectivity and verify that CD can increase the conversion of acetone beyond the equilibrium limit.

An important goal of this thesis is to develop a new model of CD that takes into account mass transfer between the catalyst and the liquid phase. It is well known that mass transfer can influence the reaction rate and product selectivity of a heterogeneously catalyzed reaction. Yet, as the literature currently exists, no models of CD take mass transfer into account.

In order to develop a model of CD, an experimental program was completed to obtain basic data for reaction kinetics and mass transfer. The rate of DAA formation over the anion exchange resin catalyst was found to be strongly limited by pore diffusion, having an estimated Thiele modulus of 11. However, reaction rates in the batch reactor were not influenced by external mass transfer to the catalyst surface. The reaction rate was also found to be extremely sensitive to the background concentration of water in the reaction mixture. Although high water concentrations slow down the rate of DAA formation, they improve the selectivity of the reaction toward DAA and improve the lifetime of the catalyst.

Mass transfer experiments were also conducted. Mass transfer rates between the catalyst structure and the liquid flowing through the column and mass transfer coefficients for rate-based distillation were measured. The rate of mass transfer between the liquid and the catalyst, however, was found to be extremely sensitive to the flow pattern which is established in the column. This created certain difficulties since flow patterns in packed distillation columns establish themselves in a random way. Thus, it is impossible to make the model of CD completely predictive. Certain steps may be taken, however, such as using liquid distributors, to improve the reproducibility and the predictability of the CD process.

Several CD experiments were conducted in the bench scale and pilot scale CD facilities at the University of Waterloo. These experiments provided operating data for a variety of liquid flow rates and catalyst bed heights. It was also verified that the CD process can exceed the equilibrium conversion of acetone, as a product containing 50 wt% DAA was easily obtained.

A differential rate-based model was developed which describes the reaction zone of the CD column. The model takes into consideration the kinetic data obtained in the absence of external mass transfer and modifies the kinetic equations to incorporate the various types of external mass transfer which exist in the CD column. The model was fitted to the CD operating data so that the rate of DAA production predicted by the model matched the experimental data. Agreement between the model and other experimental variables which were not fitted was very good. The predicted flow patterns were in agreement with the earlier work of Porter and Templeman (1968), and the model was also able to provide a good estimate of the expected product selectivity and MO production rate. Several samples from within the catalyst bag were also taken, and they confirmed that the predicted concentration profiles in the catalyst bag were close to measured values.

Several experiments were conducted with a proper liquid distributor located just above the catalyst zone of the column. Improving the liquid distribution over the catalyst bag resulted in 45-240% gains in DAA production and 112-343% gains in product selectivity (measured as mol DAA/mol MO). The distributor also reduced the variability of the system. Clearly, liquid distribution and mass transfer are two of the most important issues in catalytic distillation.

A major problem encountered with this process is catalyst deactivation. If a catalyst can be found that has improved lifetime characteristics, the production of DAA with Catalytic Distillation could easily be commercialized.