Crystallization is an important industrial separation and purification process. The purpose of any crystallization process is the separation of solute and solvent upon freezing. The separations obtained should conform to those predicted by the equilibrium phase diagram. The degree of separation achieved, however, is often much less than predicted. Wilcox (1964) and Cheng et al. (1967) postulated that these nonequilibriuin compositions could be the result of microscopic pockets of solvent being trapped inside the crystal as a second phase. These second phase impurities are called occlusions (inclusions).The presence of occlusions in crystals grown from solution and the melt has been widely reported in the literature (Belyustin and postulated that occlusion formation (hence solvent trapping) was the result of the crystal interface becoming unstable and growing dendritically with the dendrite arms impinging on one, another in such a way as to form the occlusions. Edie and Kirwan (1973) used results from the literature on the radius and spacing of dendrites in metal systems to develop a quantitative relation between solvent trapping and various process variables (i.e., crystal growth velocity, interfacial temperature gradient). This relation was refined by Myerson and Kinvan (1977b), who obtained a relation that predicted increased trapping with increased crystal growth velocity and a decrease in the absolute value of the interfacial temperature gradient. This relation was shown (Myerson and Kinvan (1977a)) to adequately correlate experimental data for the growth of single crystals of sodium chloride, potassium aluminum sulfate dodecahydrate, and calcium sulfate dihydrate grown from aqueous solution, as well as literature data on melt crystallization.Nidel(l977) attempted to reproduce experimental results on occlusion formation in single crystals in a continuous flow, MSMPR (mixed suspension, mixed product removal) crystallizer. In studies of the growth of potassium aluminum sulfate dodecahydrate in a suspension crystallizer at growth velocities employed in single crystal studies, no impurity trapping was observed. The occlusion-free crystals obtained in the MSMPR svstern, however, were significantlv smaller than those with Parll S1,iminhn 15 currentl! with E . 1. DuPont de Swnntir\ C o , Xtagargi F.ills, NY. Cnrre\)mndence concerning this paper shotild I)r w a t t to A S. !d!ersm. ooO1-1.541-81-0294-1029-$2.00. 0 The American Institute of Chemical Engineers, 1981 occlusions reported in single crystal studies. Previous investigators (Denbigh and White (1966), Brooks et al. (1969) have reported qualitative observations indicating that occlusion formation is a function of crystal size. Brice and Bruton (1974) developed a simple model which predicts that interfacial breakdown (hence unstable growth and occlusion formation) could only occur when the size of a crystal face exceeded some critical size, which is a function of the growth parameters. The model could explain the lack of trapping observed in Nidel's (1977) wo...