Sampling practices in powder blending

Abstract

Blending and sampling of granular materials are discussed in the context of pharmaceutical operations. Several experiments are used to compare the state of the art with regulatory expectations, especially those highlighted in a recent court ruling generally known as the `Barr Decision'. It is shown that samples obtained using thief probes can exhibit considerable errors, possibly leading to misleading results. The effects of sample size and number of samples are examined using an alternative characterization technique in which a powder mixture is solidified through infiltration with a polymer solution, sliced, and examined using image analysis. It is shown that number of samples is the most important parameter controlling the accuracy of the characterization. While a homogeneous mixture can be accurately characterized using just 30 samples, poorly mixed systems require hundreds of samples. It is also shown that an accurate characterization of a mixture inside a blender may not be indicative of the final state of the mixture after additional processing steps.