DescriptionThe objectives of this study consisted of: (1) determining the optimal instrument set-up of a Near Infrared (NIR) spectrometer and, (2) monitoring and quantifying critical quality attributes (CQAs) of a continuous mulling process. For this, off-line and in-line calibration standards of alumina/nitric acid granules with known water concentrations were prepared. Subsequently, the relationship between sample-detector distance and prediction accuracy was investigated for a JDSU microNIR spectrometer. It was found that the optimal distance was around 15 mm, 15mm of sample thickness and 25 of scan number. Additionally, other set-ups which could be used for in-line monitoring have been investigated including monitoring from the side and/or bottom surface through glass or plastic. Next, after ensuring adequate method accuracy, NIR prediction models were built for the in-line measurement of CQAs: granule water content and agglomerate size.