IFPAC 2011

Middleton Research is presenting Characterization of Blending Processes Using Hyperspectral Imaging

Abstract: The mixing of solid organic particles is a complex process that is difficult to characterize with the single metric of the standard deviation of the overall spectra. Near-infrared chemical imaging provides the necessary chemical and spatial information to characterize the entire process of blending, not just the end-point. A laboratory scale rotating drum blender was monitored with a SWIR push-broom camera system to provide images of the pharmaceutical blend after each rotation of the drum. The hypercube of spectral and spatial information for each rotation was processed using PLS and SBC models to illustrate the evolution of the mixing of each ingredient during the process. The imaging allows a more complex mathematical description of the uniformity of the blend. The degree of mixedness within each image is a more informative metric than the single spectral variance calculated from rotation to rotation by current the blend monitors. Spatial uniformity, i.e., the size, shape and localization of the remaining aggregates of each ingredient, provides additional relevant information for the characterization of this unit operation. Possible new metrics, combining the standard deviation of the measured composition and the spatial distribution of the ingredients, for improved monitoring of the blending process will be proposed.

Request Presentation

Back to What's New