Evaluation of in-line spatial filter velocimetry as PAT monitoring tool for particle growth during fluid bed granulation

• Published in: European journal of pharmaceutics and biopharmaceutics 2010-09, Vol.76 (1), p.138-146
• Main Authors: Burggraeve, A., Van Den Kerkhof, T., Hellings, M., Remon, J.P., Vervaet, C., De Beer, T.
• Format: Article
• Language: English
• Subjects: Biological and medical sciences; Chemistry, Pharmaceutical; Equipment Design; Excipients - chemistry; Feasibility Studies; Fluid bed granulation; General pharmacology; Glucose - chemistry; Hypromellose Derivatives; In-line measurements; Materials Testing; Medical sciences; Methylcellulose - analogs & derivatives; Methylcellulose - chemistry; Models, Statistical; Particle Size; Particle size (distribution); Pharmaceutical technology. Pharmaceutical industry; Pharmacology. Drug treatments; Polysorbates - chemistry; Powders; Process analytical technology (PAT); Process understanding; Rheology - instrumentation; Spatial filter velocimetry (SFV); Starch - chemistry; Technology, Pharmaceutical - instrumentation; Temperature
• ISSN: 0939-6411; 1873-3441
• DOI: 10.1016/j.ejpb.2010.06.001

Picture of the laboratory scale fluid bed granulator (GPCG 1, Glatt, Binzen, Germany) used during the study. A top spray nozzle was installed at a height of 26 cm from the distributor plate. In-line particle size measurements were performed with a spatial filter velocimetry probe (Parsum IPP 70), installed at a height of 20 cm and connected to an air supply system. In this study, the feasibility of spatial filter velocimetry (SFV) as process analytical technology tool for the in-line monitoring of the particle size distribution during top spray fluidized bed granulation was examined. The influence of several process (inlet air temperature during spraying and drying) and formulation variables (HPMC and Tween 20 concentration) upon the particle size distribution during processing, and the end product particle size distribution, tapped density and Hausner ratio was examined using a design of experiments (DOE) (2-level full factorial design, 19 experiments). The trend in end granule particle size distributions of all DOE batches measured with in-line SFV was similar to the off-line laser diffraction (LD) data. Analysis of the DOE results showed that mainly the HPMC concentration and slightly the inlet air temperature during drying had a positive effect on the average end granule size. The in-line SFV particle size data, obtained every 10 s during processing, further allowed to explain and better understand the (in)significance of the studied DOE variables, which was not possible based on the LD data as this technique only supplied end granule size information. The variation in tapped density and Hausner ratio among the end granules of the different DOE batches could be explained by their difference in average end granule size. Univariate, multivariate PLS and multiway N-PLS models were built to relate these end granule properties to the in-line-measured particle size distribution. The multivariate PLS tapped density model and the multiway N-PLS Hausner ratio model showed the highest R 2 values in combination with the lowest RMSEE values ( R 2 of 82% with an RMSEE of 0.0279 for tapped density and an R 2 of 52% with an RMSEE of 0.0268 for Hausner ratio, respectively).