Modeling of poly(ethylene terephthalate) reactors. VIII. A modified continuous transesterification process

Recently, Ravindranath and Mashelkar' developed a mathematical model for a continuous transesterification process.The influence of various operating and processing variables (such as residence time distribution, temperature distribution, and the number of reactors) was investigated.In particular, their influence on the dimethyl terephthalate (DMT) conversion and the side products formation was examined critically.The mathematical model developed by us is valid only when a single fractionating column is used for all the reactors and when ethylene glycol (EG) is refluxed completely to the first reactor (see Fig. 1).This is the industrial practice in new continuous p l a n k 2 However, there appear to be a number of instances when the old semibatch plants are being converted into continuous plants where the practice differs3 (see Fig. 2).In such cases, each reactor has a separate fractionating column, and EG will be refluxed to the individual reactor immediately.We shall refer to this as the "modified process."It is obvious that these two different arrangements are likely to produce different concentration profiles, thereby changing the levels of conversion as well as those of side products formation.It is not very clear a priori as to what changes such difference in industrial practice should bring in, and yet the problem seems to be of immediate industrial importance.We therefore decided to extend the model built previously to answer this question.During the course of this investigation, we have also examined the influence of different process and operating variables on DMT conversion and side products formation in the modified process, since our previous work does not provide this information.The framework of the mathematical model to be used has been described in detail by Ravindranath and Mashelkar.'The reactors in the train are assumed to be completely backmixed.The kinetic parameters, equilibrium constants, and the vapor pressure data for volatile components were taken from our earlier work.'We shall not provide the details of the mathematical model, except to state the major difference between this work and our earlier work.'These lie mainly in the calculation of the vapor flow rates and the composition of the reaction mixture.The vapors emerging from the fractionating column in the modified plants contain mainly methanol, water, and acetaldehyde, since EG is refluxed immediately.No reactions occur in the fractionating column.Furthermore, it is assumed that complete separation of the mixture occurs in the fractionation column.Therefore, eqs.( 15), ( 16), (19), and (21) in Ref. 1 are now modified as follows:*