455229 Linking Particle Size and Shape Distributions to Filtration Performance Using Measurements and Models

Thursday, November 17, 2016: 10:35 AM
Cyril Magnin I (Parc 55 San Francisco)
Giulio Perini1,2, Fabio Salvatori2, David R. Ochsenbein2, Marco Mazzotti2 and Thomas Vetter1, (1)School of Chemical Engineering and Analytical Science, University of Manchester, Manchester, United Kingdom, (2)Institute of Process Engineering, ETH Zurich, Zurich, Switzerland

Chemical products sold as solids typically undergo at least one crystallization step in their respective production process[1]. Crystallization processes are primarily used for purification, but also as a means to establish the downstream processing properties of the generated crystals, for instance their filterability, flowability and ultimately – if relevant, such as for pharmaceuticals – their tabletability. These properties are thus of vital importance when moving from a suspension towards a dry product.

It is widely accepted that the particle size and shape distribution (PSSD) originating from a crystallization process is chiefly responsible for a product’s downstream processing properties[2]. Considering the processing step immediately following crystallization, filtration, the PSSD affects the efficiency and effectiveness of the filtration step, i.e., how fast and how complete the separation of solids from the liquid is achieved. For example, there is ample experimental evidence that small particles and wide particle size distributions lead to longer filtration times (lower filtration efficiency)[3], while it is generally observed that non-isotropic particle shapes (needles, platelets) lead to higher mother liquor retention (lower filtration effectiveness)[4].

Current state of the art approaches describing the relationship between particle size and shape and filtration performance make strong simplifications with respect to the particles’ size and shape, i.e., the crystals are often assumed to be monodispersed, all of the same shape or both. In this work, we investigate whether (and when) more accurate information about the particles, such as the full PSSD, is beneficial to predict filtration performance. To this end, several different particle size and shape distributions of β-L-glutamic acid crystals were prepared by different crystallization, milling and sieving operations. The particles were generally needle-like, but with varying aspect ratios and sizes. On these materials two types of measurements were carried out: first, we report measurements of complete, quantitative PSSDs obtained with a custom-built stereoscopic measurement device[5]. Second, we report measurements of the mother liquor flow rate through cakes of these materials at constant pressure drops up to 5 bar. By correlating the full PSSDs as well as simplified particle characterizations (such as 1D particle size distributions and their moments/quantiles), with filtration performance, we aim at providing insight which level of particle characterization allows for a (reasonably) quantitative link.


[1] Variankaval, N., Cote, A.S., Doherty, M.F., “From form to function: Crystallization of active pharmaceutical ingredients”, AIChE J., 2008, 54, 1682-1688.

[2] Wakeman, R., Tarleton, S., “Solid Liquid Separation: Principles of Industrial Filtration”, Elsevier, Amsterdam, Netherlands, 2005.

[3] Wakeman, R., “The influence of particle properties on filtration”, Sep. Purif. Technol., 2007, 58, 234–241.

[4] Garside, J. (ed.), “Separation Technology: The next ten years”, Institution of Chemical Engineers, Rugby, UK, 1994, p. 81.

[5] Schorsch, S., Ochsenbein, D.R., Vetter, T., Morari, M., Mazzotti, M., “High accuracy online measurement of multidimensional particle size distributions during crystallization”, Chem. Eng. Sci., 2014, 105, 155-168.

Extended Abstract: File Not Uploaded