(686c) Determination of Minimum Catalyst Mass Requirements for Heterogeneous Reactive Distillation Columns

The method presented determines in the early design stage the minimum amount of catalyst needed to operate a heterogeneous reactive distillation column based on detailed kinetic data. The amount of catalyst needed is among others important for feasibility evaluation.

Reactive distillation has matured during the past four decades and may be viewed as conventional technology for equilibrium limited reactions within the chemical industry, with over 150 applications known to be in operation at an industrial scale. Research on RD has focused on the feasibility, design/simulation and optimization of specific processes, using rigorous simulation models.

Obtaining a good estimate for the required catalyst loading is rarely discussed and methods are lacking, while inappropriate initial guesses may yield convergence issues. Short-cut methods, such as those of distillation for reflux (1.2 – 1.5 * R_min­) and number of stages (2 * N_min) are therefore needed. An analytical method has been published by Cheng, Ward & Yu (2010) for binary systems using an analogy with an infinite amount of CSTR’s, comparable to a column with infinite stages, for a specific reflux ratio. The recommended amount of catalyst is then estimated as double the minimum catalyst mass. For multicomponent systems a method is proposed by Subawalla & Fair (1999) via simulation of isothermal reaction and ideal separation in series, but this is not shown to be a minimum.

The situation at infinite, or very high, reflux ratios is explored in this paper. Under these circumstances, the amount of catalyst can be minimized and is a function of the required column turnover and local compositions. Component concentrations with respect to column length are no longer affected by feeds or draws at the very high reflux ratios, due to large internal streams.

Method

The generic reaction under consideration is the quaternary A + B <-> C + D reaction, described by Luyben & Yu (2008), which can be described using power law kinetics and is formulated such that relative volatility is not temperature dependent. Both reactants are middle boilers, while products are the highest and lowest boilers.

Minimum amount of catalyst required to achieve a certain turnover under column conditions was estimated without the use of rigorous column modeling in Aspen Plus by calculating the maximum reaction rate achievable by the catalyst, assuming no mass transfer limitations, obtained from the liquid boiling point of mixtures in top and bottom of the column. No mixing of top and bottom products was assumed, thus the backwards reaction may be neglected, and the mixtures may be approximated as a ternary mixture (ABC – ABD).

Aspen Plus was used to simulate the reactive distillation column at the specified turnover and the minimum catalyst amount was determined by calculating the required reflux ratio while reducing the amount of catalyst per stage.

Results

The described method yielded a ternary diagram for top and bottom section showing the maximum rates that can be achieved with respect to column conditions. Figure 1 shows the combination of top and bottom sections and resulted in a diamond shaped diagram showing the composition with the highest forward rate and thus the lowest catalyst requirement assuming an effectiveness factor of 1. For the case described this was a minimum catalyst amount of 9.65 kg at a feed of 50 kmol/hr A and 95% conversion to C.

Simulations in Aspen Plus showed that reducing the amount of catalyst in the column lead to an increase in reflux ratio, eventually reaching an asymptote as can be seen in figure 2. The location of this asymptote was determined partly by the distribution of the catalyst in the column, with a single reactive tray yielding the lowest value. The Aspen Plus result for the described case was a minimum of 10.48 kg of catalyst and coincides with the composition found in the diamond diagram. This composition may be found using the concentration profile over the column at infinite reflux.

A second asymptote may be identified at the point of minimum reflux for high catalyst amounts. This can be seen in figure 3. One asymptote is at infinite reflux and minimum catalyst, where the amount of catalyst is minimized due to the high degree of separation imposed by the large reflux. The other asymptote lies at infinite catalyst and minimum reflux, where the reaction reaches chemical equilibrium on every reactive stage.

Analysis of a second case: methyl lactate hydrolysis yielded an estimated minimum of 98.8 kg of catalyst and an Aspen Plus result of 116.2 kg of catalyst for 95% conversion of A at a feed of 50 kmol/hr.

Conclusions

A method has been presented to determine in an early stage the minimum amount of catalyst required for a specific column turnover at the condition of infinite reflux, without the need for rigorous simulation of the RD column. This situation has been validated using Aspen Plus simulations at high reflux ratios and show results in line with the estimates.

Two asymptotes may be identified for the relation of reflux ratio and catalyst amount, yielding one associated with chemical equilibrium and minimum reflux, and one with high separation power and minimum catalyst requirement.

A minimum catalyst amount can be found on a single reactive stage, the optimum location is identified using the concentration profile over the column at infinite reflux and the ideal composition from the diamond diagrams.