(M1430-06-33) Use of PAT Tools in the Development, Monitor and Control of Continuous Manufacturing Processes

Purpose: Applying a Quality by Design concept in the development of a continuous process, allows process parameters to be more flexible while ensuring a quality product is obtained. Along with line and soft sensors, Process Analytical Technology (PAT) tools are critical to ensure the process stays under the “state of control” whereas tracking and diversion of non-conforming product based on Residence Time Distribution (RTD) are part of the automation system of the equipment and must be defined in the development stage. Moreover, these monitoring and control mechanisms opens the possibility for Real Time Release (testing), which implies the PAT location is critical, providing it can act on a representative sample. The aim of this work is to explore the relevance of PAT probe placement in a continuous direct compression process with focus on the two most common options used with continuous lines and options for method development in the lab for both.

Methods: A standard tablet formulation was used for these trials with a commonly used poorly flowing API surrogate. NIR monitoring was performed with a Light House Probe, Sentronic, placed before the tablet press and a DRLS probe, Sentronic, installed in the feedframe of the press. Off-line NIR tablet characterization was performed with a Bruker MPA.

Results: In a GEA Continuous Direct Compression system, the choice between a Light House Probe (LHP) and the DRLS installed in the feedframe to monitor blend potency (image attached), must take into consideration the intended overall control strategy. In this sense, whereas the LHP monitors the powder before the final mixing (which occurs in the feedframe) and therefore the RTD must be taken into consideration to initiate diversion (at the press level), the DRLS is more prone to sample presentation deviations (when powder properties slightly change during the batch). For flexibility purposes, the development of chemometric models (calibration) can be initiated in the laboratory in a setup that mimics the original environment (PAT stack rig for LHP and feedframe simulators for DRLS). In the laboratory, PAT tools are also widely used in the determination of RTD, whereas method configuration of the PAT Management System can also be optimized and used to support the development activities. If needed in the control strategy or for development purposes only, and if proved feasible, tablet assay determination in transmittance by Near Infrared or Reflectance by Raman, can also be developed at this stage. Applying these models at scale, the influence of mixing inside the feed frame seems to be more important than the delay between LHP and DRLS as the RTD model matches the DRLS values (image attached). Modification of powder properties during step changes led to a change in the hold up mass of blenders which affected the volume of powder inside the feed frame. This led to a different sample presentation to the DRLS (distance between probe and powder) but did not affect the LHP as it is placed above the press the blender. Although the prediction is not affected, DRLS probe installation optimization must be performed.

Conclusion: Powder properties affect sample presentation (higher variability in DmodX) to the DRLS due to the different volume of powder inside the feed frame throughout the batch, however, API surrogate prediction is not affected. The short lag between LHP and DRLS indicates that limited mixing of material inside the feed frame is occurring, but prediction of API surrogate obtained by DRLS inside the feed frame is more representative of final product concentration.

Location of NIR probes



LHP vs DRLS profile