Batch crystallization is widely used in the chemical and pharmaceutical industries for the manufacture of a wide range of high-value added chemicals and for purification of materials. Batch crystallizers are generally useful in small-scale operations. They are flexible in operations, requiring less capital investments than those for continuous crystallizers. However, optimal control of a batch crystallizer is a difficult task, because batch crystallization is a complicated unsteady-steady state process, where temperature, supersaturation, crystal size and population density of crystals are all changing with time.

　Control of crystal size distribution (CSD) is important in batch crystallization from the following reasons. When it is used as a separation technique, the CSD determines operabilities of down-stream processes such as filtration, centrifuge and drying, and hence it is closely related to the manufacturing cost and the product purity. Crystals with a large average size are preferable to reduce the load for such down-stream processes. As a particle production technique, control of CSD is related to powder characteristics of crystalline products such as flowability and bulk density.

　In the present study, the usefulness of seeding for control of CSD in batch cooling crystallization is investigated extensively. This thesis consists of the following 8 chapters.

　Chapter 1 describes outline of the present study and briefly reviews previous studies in literature related to the present work.

　In chapter 2, effects of some operational conditions on CSD were examined for the potassium alum system under different batch cooling crystallization conditions. In seeded batch cooling crystallization, the crystal size distribution changed significantly, depending on the condition. With enough seed loading above a critical seed concentration, uni-modal product crystals of grown seeds with virtually no nucleation were obtained regardless of the cooling mode. High suspension density caused more agglomeration. For generation of uni-modal product crystals, three factors of (1)low secondary nucleation rate caused by low supersaturation peak, (2)short period of nucleation caused by short time of high supersaturation, and (3)agglomeration of nuclei to the seed crystals were shown to work simultaneously. Homogeneity of crystal suspension is a condition for producing product crystals of uni-modal distribution.

　In chapter 3, the critical seed concentration was correlated with seed size for batch cooling crystallization of the potassium alum system by

Cs^*=1.79x10^-6・Ls^{2　　　　　(1)}

where Cs^* is the critical seed concentration, defined as the ratio of the theoretical crystal yield with the mass of the seed crystals added, and Ls is the mean mass size of the original seeds. Eq.(1) enables us to determine the amount of seeds to be added for a given size of seed crystals and also enables us to estimate the mean mass size of product crystals with the help by the mass balance equation assuming no nucleation, no breakage, no agglomeration and no crystal shape change with no residual supersaturation. Further, the volume of a crystallizer could be easily calculated. These calculations could be also made graphically. The physical meaning of Eq.(1) was discussed. To suppress nucleation effectively for larger seeds, supersaturation must be lowered more by adding seed crystals more than the amount needed for the constant surface area.

　Then, in chapter 4, the multi-quenching crystallization method was proposed to overcome the limitation of the one-stage quenching crystallization discussed in preceding chapters. The method for determining quenching strategy using the design diagram was proposed. The multi-quenching crystallization was shown to work successfully for the model system of potassium alum. The mean mass product size was closely in agreement with the predicted average size by the mass balance equation. Any seeds of a given size were expected to be grown to product crystals of any size by this multi-quenching technique.

　Chapter 5 describes the scale-effect on CSD in seeded batch cooling crystallization. With enough seed loading over a critical seed concentration under the well-mixed condition, uni-modal product crystals of grown seeds were obtained under one-stage quenching (natural cooling) mode even in a large pilot-scale crystallizer. No scale-up effect on CSD was observed if enough seeds were loaded over the critical seed concentration. Simple scale-up design calculations were made.

　In chapter 6, numerical simulation using a computer was performed for seeded batch cooling crystallization of potassium alum. Simulation results were fairly in agreement with the results from the corresponding experiments. This means that batch cooling crystallization proceeds mainly with ordinary nucleation and growth mechanisms.

　In chapter 7, seeding effect was discussed for difficult-to-handle systems. Sodium sulfate decahydrate and monosodium L-glutamate were crystallized as typical difficult-to-handle systems. These two were crystallized successfully by seeding as similarly as in the potassium alum system -a typical easy-to-handle system-, discussed in previous chapters. Scale formation was completely avoided for sodium sulfate decahydrate crystallization.

　Finally in chapter 8, the present work was summarized and the conclusions were drawn.