In this research the growth of an initial population is studied for seeded batch crystallization of ammonium sulphate in water. The system is analysed in a comparison with so-called ideal growth behaviour. Ideal growth is a situation where the number of crystals remains constant during the batch. Deviation from ideal growth can be caused by partial dissolution of the seeds or by nucleation. The focus of this work is on the influence of seed quality on sensitivity for dissolution and on the influence of crystal-impeller collisions on nucleation.

Three different types of seeds are used for seeded experiments in a 75-l draft tube (DT) crystallizer operated in evaporative fed-batch mode. The results show that seeding with ground seeds can improve product quality and reproducibility compared to unseeded operation. However, partial dissolution of the seed crystals was detected especially at a low seed mass. Seeds that are prepared from product slurry do not suffer from this drawback. However, the large median size and broad CSD of this seeding material result in excessive nucleation and a broad final product CSD. To combine favourable properties of previous seeding methods, primary nucleation is used to produce seed crystals in a separate seeding vessel by addition of an anti-solvent. Large amounts of nuclei in the seeding vessel are generated upon addition of anti-solvent. The seeds did not dissolve at a low seed mass, but agglomeration of the small particles reduces the surface area available for growth in the 75-l DT crystallizer.

In all experiments in the 75-l DT crystallizer deviations from ideal growth is observed due to the onset of nucleation early in the batch. To investigate the cause of this nucleation, some seeded batch cooling crystallization experiments are carried out in a bubble column using the same system. The bubble column does not have any moving mechanical parts. The mixing is provided by the upward velocity of the air bubbles. The results show that ideal growth is already achieved for a low amount of seeds compared to an agitated crystallizer. It demonstrates that attrition due to crystal-impeller collisions is the main source of nucleation in agitated crystallizers.