On-line monitoring of blend uniformity in continuous drug product manufacturing process—The impact of powder flow rate and the choice of spectrometer: Dispersive vs. FT

[Display omitted] •The use of a partial least squares (PLS)-based spectral pretreatment to counter the impact of flow rate on spectroscopy-based process monitoring of blend uniformity.•Equivalent error statistics between a FT-NIRS and a dispersive NIRS method given appropriate co-adds used in both instruments.•Advantageous sensitivity and specificity of a FT-NIRS method in comparison to a dispersive NIRS method.•NIRS method performance on monitoring blend uniformity on flowing powder in continuous manufacturing of drug product shall be evaluated with respect to both multivariate FOM and measurement time consideration. One of the commonly acknowledged issues in continuous manufacturing of drug products is how to provide a representative sampling on flowing powder to assure its blend uniformity. An investigation was conducted to improve understanding on the impact of powder flow rate under different continuous manufacturing conditions and the impact of optical parameters (such as resolution, co-adds, and integration time) on NIR spectral quality with respect to a dispersive and a Fourier transform instrument. A partial least squares (PLS)-based spectral pretreatment was found useful to tackle the impact of different flow rates on NIR spectral signals. Multivariate figures of merit (FOM) were used to evaluate performances across different instruments and optical settings and discover the advantageous selectivity and sensitivity on the Fourier transform than the dispersive instrument regardless of the use of co-adds.