Mechanistic prediction of powder flowability through Bond number: Model calibration and estimation of flow enhancement after dry coating

The granular Bond number, estimated using Chen's multi-asperity particle adhesion model, has been shown as a basis for powder flowability prediction for both as-received and dry-coated powders. Here, the prediction model calibrated solely using as-received powders is tested by estimating the expected flowability enhancement after dry coating with two different nano-silicas. Prediction accuracy for 9 APIs and 18 excipients is compared for three characteristic particle size parameters for the Bond number; median particle size (d50), Sauter mean diameter (d3,2), and the size class-dependent (SCD) metric utilizing entire particle size distribution (PSD). A new sigmoid-function based flowability prediction model is proposed and fitted for as-received powders. Both d3,2 and SCD lead to better predictive accuracy with narrower prediction intervals as compared to d50, suggesting d3,2 captures the same surface-area-dependent PSD effects as SCD without added computational burden. The model could accurately predict flow category for all but one as-received powder. The model calibrated through as-received uncoated powders can predict enhancement of one to three flow categories for most dry coated powders with high accuracy. Notable exceptions are materials with SSA above ∼2 m²/g whose experimental SSA deviates significantly from PSD-estimated theoretical SSA (high ΔSSA), indicating surface roughness or adherent debris that restricts dry coating effectiveness. Materials with high SSA but low ΔSSA, where high surface area is simply a consequence of fine particle size, respond well to dry coating as predicted. In summary, the proposed unifying model can be calibrated using small samples of as-received powders to estimate expected flowability improvements after dry coating and, importantly, to identify dry coated powders likely to underperform.