Precipitation processes are widely used for the production of pharmaceuticals and fine chemi-cals. However, the kinetics of particle formation is often unknown. When trying to tailor the parti-cle size this knowledge is essential. In this work we present our approach for characterizing the key phenomena, i.e. nucleation, growth and agglomeration.

An organic compound, L-glutamic acid, was chosen as a model system. The metastable α-form of L-glutamic acid was precipitated from monosodium glutamate via pH-shift in aqueous solu-tion. The precipitation of the α-polymorph was confirmed using Raman spectroscopy [1]. Growth kinetics was determined in seeded batch experiments by employing ATR-FTIR spectroscopy for the measurement of the desupersaturation profile. Nucleation kinetics was quantified by induc-tion time measurements using the Focused Beam Reflectance Measurement (FBRM) [2]. Ag-glomeration could be observed via optical microscopy over a wide range of supersaturations. The kinetics of agglomeration was determined based on seeded batch experiments. FBRM monitoring was used to confirm that nucleation is negligible. The particle size distribution (PSD) was measured using the Beckman Coulter Multisizer and the desupersaturation profile was de-termined employing ATR-FTIR. The experiments are conducted in a wide range of different op-erating conditions, i.e. initial supersaturation, particle size of seed crystals, suspension density and stirring rate. Based on these results, different modeling approaches for the description of agglomeration are applied and assessed. The simple model only takes into account the influ-ence of the supersaturation on the agglomeration process. In the more rigorous modeling ap-proaches the agglomeration rate coefficient is decomposed into a size-dependent collision ker-nel, describing the collision of particles in turbulent flow, and a sticking probability, taking into account that not every collision leads to a stable agglomerate [3, 4]. Computational fluid dynam-ics (CFD) is used to model the turbulent flow in the stirred vessel and to extract information about the shear rate distribution, which in turn can be used to incorporate the dependence of the agglomeration kernel on the local shear rate. The aforementioned experiments were utilized to determine the empirical parameters in the agglomeration models using an integral parameter estimation technique.

A population balance model accounting for the nucleation, growth and agglomeration has been developed to simulate the precipitation of L-glutamic acid in stirred vessels. Finally, the accu-racy of the kinetics correlations is assessed by comparison of process simulations and experi-mental results.

[1] J. Schöll, D. Bonalumi, L. Vicum, M. Mazzotti: In Situ Monitoring and Modeling of the Solvent-Mediated Polymorphic Transformation of L-Glutamic Acid, Cryst. Growth. Des., 6(4), 881-891, 2006.

[2] J. Schöll, L. Vicum, M. Müller, M. Mazzotti: Precipitation of L-glutamic acid: Determination of nuclea-tion kinetics, Chem. Eng. Technol., 29 (2), 257-264, 2006.

[3] D. Ilievski, I. Livk: An agglomeration efficiency model for gibbsite precipitation in a turbulent stirred vessel, Chem. Eng. Sci., 61 (6), 2010-2022, 2006.

[4] J. Bałdyga, M. Jasińska, A. Krasińksi, A. Rożeń: Effects of fine scale turbulent flow and mixing in ag-glomerative precipitation, Chem. Eng. Technol., 27 (3), 315-323, 2004.

* Author to whom correspondence should be addressed. Phone: +41-44-632 24 56. Fax: +41-44-632 11 41. Email: marco.mazzotti@ipe.mavt.ethz.ch