Predicting pharmaceutical powder segregation from particle property disparity and blend flowability.

Powder segregation is a critical quality risk in solid dosage manufacturing, yet current risk assessment is resource-intensive hence poorly suited to early formulation screening. This work addresses that by introducing a mechanistic predictive framework linking constituent particle properties and blend flowability to segregation risk. Binary blends of two ibuprofen grades (IBU25, IBU70) and micronized acetaminophen (mAPAP) with eight pharmaceutical excipients (n = 24 total) were characterized for particle size, shape, density, and flow function coefficient (FFC). Segregation was quantified by NIR spectroscopy (SPECTester) using Segregation Intensity (SI) and Cumulative Segregation Area (CSA), a composition-integrated metric. Particle size-density disparity between API and excipient establishes the mechanical driving force for segregation, while blend flowability determines how fully that potential is expressed; neither factor alone is sufficient, but together they could be predictive. Consequently, a dimensionless Particle Mechanical Property Factor (PMPF), combining D3,2 ratio and bulk density ratio, was introduced as a pre-formulation disparity descriptor, leading to a novel parameter called the Segregation Propensity Number (SPN) as the product PMPF × FFC. It was used in a power-law model y = k(SPN)n, trained on IBU25 and mAPAP blends (n = 16) and validated on held-out IBU70 blends (n = 8) achieved R2_ext = 0.95 for CSA and R2_ext = 0.84 for SI, with CSA showing superior generalizability and minimal systematic bias. This framework is expected to enable a priori segregation risk assessment from routinely collected particle data paired with a single blend flowability measurement, supporting QbD-driven excipient selection and early formulation development.