AbstractHigh Shear wet granulation is traditionally considered a highly unrepeatable solids processing step, with the considerable domination of art over science. In The United States, the Food and Drug Administration, a major regulator of pharmaceutical processes is encouraging pharmaceutical companies to understand the science and engineering aspects of their process, and move from the current qualitative understanding to a quantitative and mechanistic understanding of their processes.
In order to examine the intrinsic variability of granulation, many carefully controlled repeat granulations were performed in a high shear mixer and the size distributions of the resulting granules were analysed, along with traditional on-line process data surrounding current consumption of the mixer, and temperature of the granulating mass. The raw experimental results were combined and analysed with the use of multivariate techniques to examine the causes of variability. It is seen that none of the measured variables fully explain the variability observed and that despite the care taken to repeat exact conditions, the granules still produced were variable.
Final analysis of the data shows that batches can be separated into two different sets, distinguished by different operators showing that minute unrecorded changes to the experimentation process, affect the numerical outcome of the procedures. The current study shows the power of multivariate techniques to analyse data, and provide insight into the variability of the underlying process.
Keywords: Granulation; Variability; Six Sigma; Chemometrics