Monitoring a single-pot granulator using near-infrared spectroscopy

Importance of the Topic

Near-infrared spectroscopy (NIRS) enables rapid, nondestructive in-line monitoring of solvent and moisture levels in pharmaceutical granulation processes, facilitating real-time process understanding, quality control and compliance with Process Analytical Technology (PAT) initiatives.

Objectives and Overview

The study investigates a novel angled optical probe design integrated with a NIRS XDS Process Analyzer to reduce sample density fluctuations in a Bohle VMA 70 single-pot high-shear granulator and to accurately predict residual methanol during the drying phase.

Methodology

A blend of 91 % lactose anhydrous and 9 % starch was granulated with water, methanol and trace dye, then subjected to an agglomeration cycle followed by controlled drying under vacuum (200 mbar) and nitrogen purge (1.0 Nm³/hr) at 200 RPM. NIR spectra (800–2100 nm, 0.5 nm intervals, 32 co-added scans) were collected continuously at approximately two spectra per minute over blending, granulation and a 43-minute drying period. Three calibration and three test batches were analyzed. Spectral preprocessing included second derivative transformation and subtraction of moisture contributions to focus on methanol absorbances.

Instrumentation Used

• NIRS XDS Process Analyzer (Metrohm)
• Angled-tip fiber-optic probe custom-designed to stabilize powder flow
• Bohle VMA 70 single-pot high-shear granulator
• Vision® software for spectral acquisition and chemometric modeling

Main Results and Discussion

A partial least squares (PLS) model over 1800–2100 nm achieved R² = 0.9823, SEC = 0.443 % MeOH and SECV = 0.622 % MeOH. Trend analysis showed smooth, realistic predictions of methanol depletion, capturing initial rapid loss and a slight rebound (attributed to granule case-hardening) before final drying. One spectrum was excluded as an outlier by library qualification criteria. The angled probe maintained quasi-laminar powder flow and minimized density variance, critical for reproducible in situ measurements.

Benefits and Practical Applications

• Real-time, non-destructive monitoring of residual solvents
• Elimination of time-consuming offline GC analyses and faster end-point determination
• Enhanced process control and compliance with PAT guidelines
• Improved product quality assurance and reduced batch release times

Future Trends and Opportunities

Further developments may include integration of GC reference data for enhanced model robustness, extension to additional solvents and formulations, application of advanced chemometric and machine learning techniques, and seamless automation within manufacturing execution systems.

Conclusion

The study demonstrates that an angled NIR probe in a high-shear granulator enables accurate, in situ prediction of residual methanol levels. By stabilizing sample density and leveraging PLS modeling, this approach supports rapid process monitoring and robust quality control in pharmaceutical manufacturing.

References

Mattes R.A., Schroeder R., Dhopeshwarkar V., Kowal R., Randolph W. Monitoring a Granulation Drying Process Using Near-Infrared Spectroscopy for In Situ Analysis Of Residual Moisture and Methanol. Pharmaceutical Technology, September 2004.