The influence of residence time distribution and backmixing

In most experimental research studies, polymerizations are carried out batchwise. For operations on a commercial scale, continuous polymerization may be attractive. This applies specifically to the free radical polymerization of ethylene under very high pressure. Of course, compressors can only operate continuously, and the very short residence time in the reactor (on the order of one minute) is only practical in continuous operation. 

It has been demonstrated both by calculations and by experiments, that polymers produced in batch or in continuous process may have a different molecular mass distribution for a given average molecular mass. The molecular mass distribution is expressed as the dispersion index D, that is the ratio of the mass (weight) average and the number average values of the molecular mass of the polymer  

When all polymer molecules would have the same length, D = 1. In practice D 1. The larger D is, the broader the molecular mass distribution, which may be desirable for the processability, but undesirable for the mechanical strength. For each polymer, optimum values of both D and M are desired for each application. 

Both graphs are based on the models of Biesenberger and Sebastian (1983). m = perfectly micro-mixed, s = fully segregated (CSTR s with perfect macro-mixing). 

In the case of free radical polymerization (figure 13.2) the dispersion index D at low degrees of conversion appears to be 1.5. In a batch or plug flow reactor D increases as the degree of conversion goes up. The reason is that the propagationrinitiation ratio decreases as the monomer is consumed. For a segregated CSTR the effect is enhanced by the residence time distribution. For a well mixed CSTR D remains constant and low. This is explained by the fact that all polymer molecules are made under identical conditions. 

---