Real-time NIR Monitoring of a Pharmaceutical Blending Process through Multivariate Analysis-derived Models

Abstract:- Process analytical technology guidance by the US Food & Drug Administration lately became the major driver of pharmaceutical process optimization. The majority of these processes are complex and consequently multivariate. Although new insights have improved knowledge of the phenomena taking place, it is not usually possible to develop deterministic modeling. Processes involving powder handling, such as multi-component pharmaceutical formulation blending, are common and the real-time monitoring of their physico-chemical attributes is challenging. In this work, we propose multivariate analysis of a V-blender mixing unit operation with an in-line near-infra red (NIR) measurement technique. The NIR measurement system used in this study consists of a micro-electro-mechanical system (MEMS)-based spectrometer connected to an IP-65-encased optical measuring head (sampling probe) through a 1-meter length umbilical wire cord. It deploys diffuse reflectance sampling technology, providing 40-mm spot size with a spectral range of 1,350 to 1,800 nm. The methodology includes the following steps: (1) modification of a nominal 1-ft (30-L) V-blender unit to accommodate Axsun’s NIR spectroscopy system; (2) 3 experimental runs, each with a different mixing time, while monitoring powder homogeneity with NIR spectroscopy; (3) acquisition of 10 powder samples after each run from pre-determined locations in the V-blender, evaluated with current Quality Assurance (QA)/Quality Control (QC) lab methods, to determine reference mixing endpoints; and (4) NIR data analysis by SIMPA-P+ an