pubs.acs.org/crystal Published on Web 09/24/2010 r 2010 American Chemical Society DOI: 10.1021/cg100305w 2010, Vol. 10 4728–4740 Evaluation of the Effect of Seed Preparation Method on the Product Crystal Size Distribution for Batch Cooling Crystallization Processes E. Aamir, Z. K. Nagy,* and C. D. Rielly Loughborough University, Loughborough, Leicestershire LE11 3TU, United Kingdom Received March 7, 2010; Revised Manuscript Received August 28, 2010 ABSTRACT: This paper provides an experimental and simulation based analysis of the effect of seed quality on the shape of the product crystal size distribution of a seeded batch cooling crystallization process. Various seeds were prepared using different protocols, involving milling, washing, and sieving. The cooling batch crystallization processes of potassium dichromate in water with the different seeds were monitored using process analytical technologies (PAT), such as attenuated total reflectance (ATR) UV/vis spectroscopy, focused beam reflectance measurement (FBRM), and online laser diffraction for real-time crystal size distribution measurement. A population balance model with apparent size-dependent growth, which incorporates the effect of growth rate dispersion, is used to simulate the evolution of the crystal size distribution (CSD) using the different seed distributions as initial conditions. The simulation results were in good agreement with the experimental product CSD, when the good quality crystalline seed was used with no fines. However, the mean crystal size of the product was overpredicted by the growth-only model, when milled seeds were used with different fine contents. This was caused mainly by the excessive initial breeding, due to the different surface properties resulting from the preparation method, the Ostwald ripening promoted by the fines, and the pronounced agglomeration observed in these cases during the experiments. The simulation and experimental results provide evidence of the importance of consistent and well-defined seed quality and suitable preparation procedures for high quality crystalline products. 1. Introduction Crystallization is a widely used separation technique used in many chemical, petrochemical, food, pharmaceutical, and microelectronics industries. 1-4 Crystallization has a wide range of applications as a separation technique, but the process development and scale-up requires extensive time and effort due to the complex kinetics and hydrodynamics of the process. 5-9 Considerable time and effort is invested in the development of batch crystallization processes to obtain consistent crystal prop- erties, such as purity, shape, size, habit, morphology, and size distribution as well as to improve the performance of down- stream processes, e.g. to reduce the filtration time after the batch. 1,8,10-12 One of the main difficulties in batch crystal- lization is to accomplish a uniform and reproducible crystal size distribution (CSD). 2,9 Poor control of the CSD can pre- vent the plant from running at full capacity and reduce profits. The loss can be considerable for pharmaceutical companies due to strong regulatory constraints and limited patent life for drug compounds. The shape of a crystal size distribution pro- duced from crystallization not only affects the efficiency of the downstream operations, such as filtration, drying, and wash- ing, but may have considerable impact on the bioavailability of the active pharmaceutical ingredients (API). Most of the product qualities (e.g., dissolution rate, bulk density, flow- ability, and packing properties, etc.) are also directly related to the crystal size distribution. 6 Therefore, controlling the size and shape of the distribution is generally the primary bottle- neck for the processes which use crystallization as a separation technique. Usually, most batch crystallization processes involve seed- ing. Seed loading varies from as low as 0.5% to as high as 10%, depending on the size and volume of the batch crystal- lizer. Seeding has been known for a long time as an effective technique to stabilize batch crystallization processes. 1 In seeded crystallization, the supersaturation is maintained at a low value away from the metastable limit, by slow cooling or optimal cooling (or antisolvent addition), or, in more recent systems, at a desired constant value throughout the entire batch by application of properly designed control algorithms, using either model-based optimization 7,10,13-21 or model-free approaches based on supersaturation control 5,22,23 and direct design concepts. 24,25 These approaches can be implemented in an open-loop or closed-loop structure with respect to the product property. Although product property-based closed- loop implementation will show a certain level of inherent robustness to uncertainties, 26 and robust open-loop control strategies have also been developed, 27 these advanced control approaches are very seldom applied in practice due to their increased implementation complexity. In the vast majority of cases, crystallization processes are controlled by tracking operating trajectories determined off-line by nominal open- loop optimization or trial-and-error procedures. In these cases, the properties of the seed play an important role and strongly affect the quality of the crystal size distribution obtained at the end of the batch. Although several techniques have been proposed for the in situ generation of seeds via controlled nucleation/dissolution events, 24,25 seeded crystallization is still predominantly applied in the chemical and pharmaceutical industries using seeds generated from the crystallization product. Generally, many steps are involved during the preparation of seed, such as milling, blending, grinding, sieving, and washing. 11,28-31,33 All these processes affect the quality and properties of the seed, and significant variations in seed quality may be observed based on the method used to produce the seed. 29,31,32 In addition, quantitative information on the *Corresponding author. E-mail: z.k.nagy@lboro.ac.uk.