In-line NIR spectroscopy for comprehensive process understanding of semi-continuous blending in continuous production...

In continuous manufacturing, Process Analytical Technology (PAT) tools are part of the control strategy to ensure the intended quality of an intermediate or final product. This study investigates the implementation of in-line Near-Infrared (NIR) spectroscopy into a semi-continuous blending process, aiming to detect blend uniformity across varying process conditions. A low-dose, binary formulation, containing 2% (w/w) caffeine as Active Pharmaceutical Ingredient was used. Two spectral analysis methods, including Partial Least Square (PLS) regression and the Moving Block Standard Deviation (MBSD) method were applied as endpoint detection procedures. These approaches enabled the assessment of both convective and dispersive mixing performance across the different blender settings. Results showed that increasing impeller speed significantly reduced convective mixing times, while larger powder volumes had the opposite effect. For dispersive mixing, similar trends were observed, though higher fill levels extended the blending time only at lower impeller speeds. Overall, the study demonstrates that real-time Near-Infrared spectroscopy (NIR) monitoring can significantly improve process understanding of a semi-continuous blending process.