(169a) Application of Seeding as a Process Actuator in a Model Predictive Control Framework for Fed-Batch Crystallization of Ammonium Sulphate

Batch crystallization is often used for production of high purity, high added-value materials with tight specifications on crystal properties (size, purity, morphology). In order to meet these requirements, an effective control strategy is needed. With the current progress in crystallization modeling, rigorous models having predictive capabilities become available. In principle it is possible to use these models in a Model Predictive Control (MPC) framework.

In this work, seeding is discussed from the viewpoint of process actuation. It is considered here as one of the (discrete) manipulated variables of MPC that is active in the start-up phase of a fed-batch evaporative crystallization process. Other manipulated variables (heat input, impeller frequency, and fines removal rate) can be used for steering in the course of the fed-batch process.

Numerous unseeded fed-batch evaporative crystallization experiments carried out in a 75-l draft-tube crystallizer and in an 1100-l draft-tube-baffle crystallizer show a wide range of values for supersaturation at which primary nucleation occurs. As a consequence, it is not possible to obtain a reproducible start-up behavior for the batches with the same operating conditions. It is clear that the outcomes of these irreproducible batches can differ significantly.

The used crystallization model is a first-principle distributed-parameter mathematical description of the crystallization systems. Crystal Size Distribution (CSD) is modeled using a Population Balance Equation (PBE). Secondary nucleation is described using modeling framework of Gahn and Mersmann ([1] and [2]). However, the model is lacking description of primary nucleation and therefore, cannot accurately predict influence of variations during the start-up on the final product quality. This serves as a main incentive to investigate influence of the seed characteristics on the product quality.

As it becomes evident from the recent study of the fed-batch crystallization of ammonium sulphate [3], evolution of CSD during a batch is strongly dependent on the initial CSD. For this particular substance having crystals with a relatively high value of hardness and low density (with respect to the density of saturated solution) this influence stretches beyond one half of the total batch time. Towards the end of the batch, changing the impeller frequency can influence the CSD to a certain limited extent. The particles being subject to attrition upon impeller-particle collisions give rise to secondary nucleation. As secondary nucleation is undesired from viewpoint of crystal size and purity, one should strive for minimizing the impact of secondary nucleation on the product quality.

A group of authors ([4] and references therein) proposed a seeding strategy that allows minimizing secondary nucleation and achieving a unimodal product CSD irrespective of the cooling profile employed in batch crystallization. According to their results, supersaturation can be kept at a low level within the metastable zone at all times by usage of an excessive amount of seeds. With the help of their findings, some seeded experiments were carried out with ammonium sulphate in the mentioned crystallizers. The main emphasis was put on getting reproducible results rather than on striving for a narrower CSD of the product crystals. The seeded experiments were done at different operating conditions (different impeller frequencies, heat inputs, and seed sizes) and compared to the corresponding unseeded batches.

A special attention is paid to the seed preparation. As an extensive amount of seeds is required for pilot-scale experiments, ground seeds were chosen due to ease in their preparation. It is known that ground seeds induce breeding when introduced dry and without pretreatment. The main cause of breeding is small particles adhering on the surface of seed crystals. When released in supersaturated solution, these dust particles act as nuclei whereby the CSD gets wide and often bimodal. It has been shown [5] that seeds fed as a suspension result in the product having characteristics similar to batches seeded with grown seed crystals. Therefore, a study has been performed to investigate how milled seeds are changing when introduced in a relatively small volume of saturated solution and left for some time. The CSD is monitored using a laser diffraction instrument (Mastersizer-X, Malvern Instruments, United Kingdom). In addition, a number of samples is taken at different times and analyzed with Scanning Electron Microscope (SEM). The study shows that dust particles adhering to the seeds crystals rapidly dissolve within a few minutes. The damaged seed crystals partially dissolve to release stress and then grow due to a slight supersaturation caused by dissolution of fines. This observation is in line with the results of a more profound study on behavior of a single secondary nucleus of ammonium sulphate in supersaturated solutions [6].

An analysis of the seeded experiments reveals a more reproducible behavior of the crystallizers than that of the unseeded operation. An application of seeding also enables a better control of the product CSD as the final product CSD is narrower than in the unseeded runs. The experimental results are in line with the ones obtained using rigorous crystallizer models. A recent study [7] proved efficiency of the impeller frequency as a process actuator. The presented results assure that seeding is another useful means of control that can be used in MPC.

References

[1] Gahn, C. and A. Mersmann (1999a). Brittle fracture in crystallization processes. Part A. Attrition and abrasion of brittle solids. Chem. Eng. Sci. 54, 1273-1282.

[2] Gahn, C. and A. Mersmann (1999b). Brittle fracture in crystallization processes. Part B. Growth of fragments and scale-up of suspension crystallizers. Chem. Eng. Sci. 54, 1283-1292.

[3] Kalbasenka, A., A. Huesman, H. Kramer, and O. Bosgra (2005). Controllability analysis of industrial crystallizers. In M. Jones and J. Ulrich (Eds.), the 12th International Workshop on Industrial Crystallization, September 7-9, Halle (Saale), Germany, pp. 157-164.

[4] Doki, N.; Kubota, N.; Yokota, M. and Chianese, A. (2002). Determination of Critical Seed Loading Ratio for the Production of Crystals of Uni-Modal Size Distribution in Batch Cooling Crystallization of Potassium Alum. J. Chem. Eng. Japan 35, pp. 670-676.

[5] Warstat, A. and J. Ulrich (2005). Seeding during batch cooling crystallization ? an initial approach to heuristic rules. In VDI-Gesellschaft Verfahrenstechnik und Chemieingenieurwesen (Ed.), The 16th International Symposium on Industrial Crystallization, Volume VDI-Berichte 1901.2, September 11-14, Dresden, Germany, pp. 1033-1038.

[6] Virone, C., J. ter Horst, H. Kramer, and P. Jansens (2005). Growth rate dispersion of ammonium sulphate attrition fragments. Journal of Crystal Growth 275, e1397-e1401.

[7] Kalbasenka, A.N.; Huesman, A.E.M. and Kramer, H.J.M. (2004). Impeller frequency as a process actuator in suspension crystallization of inorganic salts from aqueous solutions. Proceedings of the International Workshop on Industrial Crystallization, GyeongJu, South Korea, September 15-17, 2004.