From sensors to reasons : fluidized bed spray granulation process monitoring by acoustic emission
Year: 2013
Type: dissertation
Abstract: Wet granulation is an integral unit operation in the manufacturing of many pharmaceutical solid forms.Despite its importance, the process has long been empirical and the endproduct quality has relied on the experience of the operator.In 2003, the U. S. Food and Drug Administration launched an initiative with the aim of modernizing the regulation and quality of pharmaceutical manufacturing.One reason for this initiative was to encourage a paradigm shift towards Quality by Design (QbD).In-line process analyzers are considered essential for successful implementation of QbD.The objective of this study was to develop an in-line process analyzer based on a passive acoustic emission (AE) technique for a laboratory scale fluidized bed spray granulation (FBSG) process.The specific aims included model calibration to match the AE data with granule size distribution (GSD) and granule moisture content.First, the suitability of the AE technique for determination of granule water content and mean particle size during FBSG process was evaluated.By combining AE data and the information from process variable measurements, granule moisture content could be determined with a PLS regression.In addition, dry granules of different sizes could be qualitatively classified from their AE spectra.Moreover, the mean particle size of blends could be estimated.Next, a predictive N-PLS model for the 25% quartile of cumulative endproduct GSD was developed using the spectra recorded during the nucleation.The length of the nucleation phase was determined and in addition, covariance between granule structure and drug content was assessed.Drug content inhomogeneity at low saturation levels was caused by segregation of the primary particles.After a certain saturation level was reached, the granule growth mechanism changed, altering the intragranular structure.This change promoted a more homogenous distribution of the drug.In the study, inhomogeneity by solute migration was shown to be caused by the product drying rate.The solute migrated towards the granule outer crust at low drying rates, which also affected the dry granule strength.Drying at higher rate halted solute migration, or inverted the flow towards the granule core.Finally, the information regarding the duration of the nucleation period and different granule growth mechanisms was used in the PLS calibration from the AE data.The GSD could be predicted with good accuracy and precision.In addition, earlywarning of process deviations could be obtained.In conclusion, the AE technique could be used to predict GSD and granule moisture content, and applied as an early-warning system for process deviations.However, off-line analytics were essential for the improvement of process understanding in order to allow proper model calibration.
Author: Sami Poutiainen
Language: English
Subfield: Computational Mechanics
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3096576313