Monitoring a single-pot granulator using near-infrared spectroscopy

Significance of the Topic

Real-time monitoring of residual solvents and moisture levels in high-shear granulation is essential for ensuring product purity, reducing off-line testing, and improving process control in pharmaceutical manufacturing.

Objectives and Study Overview

This application study aims to demonstrate the feasibility of using a novel angled near-infrared (NIR) fiber-optic probe integrated with a Metrohm NIRS XDS Process Analyzer to predict residual methanol and moisture during the drying phase of a Bohle VMA 70 single-pot high-shear granulator. Key goals include:

• Minimizing density fluctuations in the granulator to enable accurate in-line NIR measurements.
• Developing multivariate models for solvent quantification without sample preparation.
• Comparing NIR results to conventional off-line gas chromatography methods.

Methodology and Instrumentation

A single-pot granulation of 24 kg blend (91% lactose, 9% starch) was performed, with moisture, methanol, and trace dye added, followed by a 6-minute wet agglomeration and a 43-minute drying cycle under vacuum (200 mBar), jacket heating, and nitrogen purge (1.0 Nm3/h) at 200 rpm. Continuous NIR spectra (800–2100 nm, reflectance mode, 0.5 nm intervals, 32 co-added scans) were acquired at approximately two spectra per minute, totaling ~150 spectra per batch. Three calibration and three validation batches were analyzed using Vision® software. The instrument setup included:

• Metrohm NIRS XDS Process Analyzer
• Angled fiber-optic probe designed to maintain quasi-laminar powder flow and constant density
• Bohle VMA 70 single-pot high-shear granulator

Main Results and Discussion

• Second derivative preprocessing in the 1800–2100 nm region was used to build a partial least squares (PLS) model.
• The calibration model achieved R2=0.9823, SEC=0.44% methanol, and SECV=0.62% methanol.
• Smooth, real-time trend charts captured methanol removal profiles, revealing initial rapid solvent loss followed by slight increases attributed to granule case-hardening.
• Model robustness was limited by fewer mid-level data points; additional calibration or GC reference data could improve accuracy.

Benefits and Practical Applications

• Eliminates delays and labor associated with off-line GC analysis.
• Enables continuous in-line monitoring of residual solvents and moisture without sample preparation.
• Supports FDA Process Analytical Technology (PAT) initiatives for real-time quality assurance and end-point determination.

Future Trends and Application Possibilities

• Integration of enhanced probe designs for other turbulent process equipment.
• Expansion to monitor additional critical quality attributes (e.g., API content, particle size).
• Advanced multivariate algorithms and machine learning to improve model accuracy and robustness.

Conclusion

The study confirms that in situ NIR spectroscopy with an angled probe can reliably predict residual methanol levels during high-shear granulation and drying. Maintaining a consistent sample density at the probe interface is crucial for model performance, paving the way for broader PAT implementation in pharmaceutical continuous manufacturing.

References

• R. A. Mattes, R. Schroeder, V. Dhopeshwarkar, R. Kowal, W. Randolph, Monitoring a Granulation Drying Process Using Near-Infrared Spectroscopy for In Situ Analysis of Residual Moisture and Methanol, Pharm. Tech., Sept. 2004.