The utility of modulating rotation speed in tumble mixing and its mechanistic interplay with particle size distribution (PSD) variability in excipients remain underexplored. They were investigated in this study. For the present purpose, PSD of a commercial grade lactose was modified to reflect commercially relevant variations; and mixed with microcrystalline cellulose and chlorpheniramine in a double-cone blender, at various rotation speeds. Model of mixing was constructed using avalanche rheological properties and was also rendered as quantifiable visual models using avalanche rheological visual metric (ARVM), to uncover mechanistic relationships. ARVM was derived through multivariate image analysis of avalanche flow. It was observed that increasing rotation speed reduced the number of rotations needed to achieve blend homogeneity by 30-33% for PSD variants with 16-20% fines, while increasing the number of rotations by 134% in PSD variants with less than 15% fines (p ≈ 0.00). ARVM successfully modelled (root mean square error of external validation = 2.46%) and revealed the mechanistic interplay. With increased proportion of fines, lactose exhibited quasi-parabolic motion with disaggregation of soft agglomerates and improved mixing. With decreased proportion of fines, lactose flowed as coherent wave-like masses with imperceptible dispersive tendency and increased dilation, which impeded mixing. In conclusion, this study contributes to process understanding and ideas for designing robust mixing operations. It showcases the usefulness of a quantitative visual approach, exemplified by the ARVM, to evaluate material variability and uncover its mechanistic impact on processing.
Keywords: Avalanche flow; Excipient variability; Multivariate image analysis; PLS; Particle size distribution; Powder rheology.
Copyright © 2020 Elsevier B.V. All rights reserved.