In-line quantification of drug and excipients in cohesive powder blends by near infrared spectroscopy

This work was aimed at investigating the utility of near infrared (NIR) spectroscopy for simultaneous in-line quantification of drug and excipients in cohesive powder blends in a bin blender. A model formulation containing micronized chlorpheniramine maleate (mu CPM), lactose, microcrystalline cellulose (MCC) and magnesium stearate (MgSt) was selected for the blending study. An optical head comprising a sapphire window mounted on the lid of the bin was used to collect in-line NIR spectral data of the powder blends. Validated partial least square (PLS) calibration models were used to quantify each component from the NIR spectra of the blends. Additionally, effects of premixing by sieving and high shear mixing and use of an internal prism fixed within the bin on the mixing performance of each component were studied. The statistical results obtained for PLS calibration models and their validation showed the sensitivity of NIR for accurate quantification of blend components. The blend prepared with high shear premixing and with prism achieved uniformity more rapidly than that with high shear premixing but without prism during blending in a bin blender. Premixing using sieving proved to be inadequate for uniform mixing of the blend components as none of the components except MgSt achieved uniform distribution after the preset blending time when blended in the bin blender. This study demonstrated that by high speed sampling and rapid spectral acquisition, distribution of individual blend components can be assessed with high accuracy during blending. Furthermore, high shear premixing facilitated rapid distribution and uniformity achievement of blend components. This technique may be used to monitor the relative distribution of individual blend components in real time and thus, to assess the performance of a bin blender for mixing of cohesive multi-component powder blends during development and production. (C) 2009 Published by Elsevier B.V.