Using the Thermo Scientific MarqMetrix All-In-One Process Raman Analyzer for real-time monitoring of a hot-melt extrusion process

Significance of the topic

Hot-melt extrusion (HME) is a widely used pharmaceutical processing route to solubilize poorly water-soluble active pharmaceutical ingredients (APIs), especially Biopharmaceutics Classification System class IV compounds. Real-time monitoring of HME is important to ensure correct API distribution, control polymorphism and crystallinity, and to provide documented, reproducible data required by GMP and regulatory frameworks. Inline vibrational spectroscopy such as Raman offers a nondestructive, sample-preparation-free pathway to acquire chemical and solid-state information during processing, enabling timely corrective actions and reducing the risk of producing out-of-specification batches.

Objectives and overview of the study

This application note evaluated the feasibility of using the Thermo Scientific MarqMetrix All-In-One Process Raman Analyzer for online, real-time monitoring of API concentration during a hot-melt extrusion run. The study aimed to build and test chemometric calibration models correlating Raman spectra acquired at the extruder die to API concentration determined by off-line HPLC, and to demonstrate the practical benefits of inline Raman as a Process Analytical Technology (PAT) tool for HME.

Methodology

• Processing: A model API was blended into a polymer matrix and processed using a Thermo Scientific Pharma 11 twin-screw extruder with two gravimetric feeders to independently control polymer and API feed rates and keep a constant total throughput.
• Inline sampling: A ball-probe extruder sampling optic was mounted in the extruder die and coupled via fiber to the MarqMetrix All-In-One Process Raman Analyzer to collect spectra from the melt stream.
• Spectral acquisition: Spectra were recorded at one-minute intervals during concentration adjustments across a range of nominal API loadings (15–60%). Continuous monitoring used 800 ms integration, 10 averages, 300 mW laser power, yielding ~16 s per scan.
• Reference analysis: Parallel extrudate samples (pellets ~1 mm × 1 mm) were collected and analyzed by HPLC to provide ground-truth API concentrations. Two sample masses were used for validation checks (~2 mg and ~30 mg), with duplicate HPLC analyses per point.
• Data processing and modeling: Spectral region 800–1800 cm−1 was selected. Preprocessing included first derivative (second order, 15-pt window, polynomial interpolation for ends), standard normal variate (SNV) and mean centering. Partial least squares (PLS) regression was used to build calibration models against the two separate HPLC result sets.

Instrumentation used

• Thermo Scientific Pharma 11 twin-screw extruder with two gravimetric feeders
• Ball-probe sampling optic (Dynisco-style extruder probe) mounted in extruder die
• Thermo Scientific MarqMetrix All-In-One Process Raman Analyzer (fiber-coupled)
• HPLC system for off-line reference quantification

Main results and discussion

• HPLC verification: Off-line HPLC results confirmed overall homogeneity of extrudate samples. Two mass loadings (2 mg and 30 mg) produced comparable results, supporting sample consistency. Some discrepancies between targeted feed and measured API content occurred at the lowest and highest target concentrations — attributed to very low dosing rates at the low end and limitations in dosing mass at the high end.
• Spectral region: The 800–1800 cm−1 window contained dominant spectral markers for both polymer and API and was sufficient for quantitative modeling in this system.
• Calibration models: Two PLS models were developed using the two sets of HPLC reference values. Preprocessing (1st derivative + SNV + mean centering) improved model behavior. Both models produced usable quantification of API content; the model based on the first HPLC round showed slightly lower RMSE for calibration, cross-validation and prediction than the second-round model.
• Operational performance: Continuous Raman probing at the die provided rapid, non-contact measurements and clear time stamps that could be correlated with HPLC samples. The analytical throughput and response time are compatible with process control in HME.
• Limitations: The models require independent external validation with additional unknown datasets prior to deployment. The study also highlighted practical dosing limitations at concentration extremes that influenced reference accuracy and therefore calibration quality.

Benefits and practical applications

• Real-time control: Inline Raman enables immediate detection of deviations in API concentration, allowing corrective actions before defective batches are produced.
• Non-destructive and waste-minimizing: No sample preparation or consumables are required for the Raman measurement, reducing waste and handling time.
• Multivariate capability: A single Raman scan can yield information about API content, crystallinity/polymorphism and homogeneity concurrently, reducing the need for multiple offline tests.
• Regulatory documentation: Automatic logging of spectral data and time-resolved measurements supports GMP recordkeeping and PAT-driven process understanding.
• Flexible deployment: Fiber coupling and probe-based sampling permit separation between detector and process, facilitating safe, remote installation in production environments.

Future trends and potential uses

• Model robustness and transfer: Emphasis on robust calibration transfer strategies, periodic model maintenance, and validation with broader process variability to enable scale-up and manufacturing deployment.
• Advanced chemometrics: Integration of non-linear methods, adaptive models and real-time anomaly detection will improve sensitivity to subtle solid-state changes and low-level content variations.
• Expanded PAT networks: Combining inline Raman with complementary sensors (NIR, terahertz, process temperature/torque) and process control systems will enable closed-loop control strategies for HME.
• Regulatory acceptance: As PAT becomes more widespread, harmonized guidelines and case studies will accelerate acceptance of inline Raman for lot release and continuous manufacturing workflows.
• Solid-state monitoring: Greater focus on polymorphism/crystallinity and residual solvent/moisture measurement by Raman for optimizing downstream performance and stability.

Conclusion

The study demonstrates that the MarqMetrix All-In-One Process Raman Analyzer, when coupled via a die-mounted ball probe to a twin-screw extruder, can provide fast, reliable inline monitoring of API concentration during hot-melt extrusion. PLS-based chemometric models built on the 800–1800 cm−1 spectral range delivered quantitative predictions that align with HPLC references, although calibration quality depends on accurate reference dosing and requires external validation. Inline Raman offers clear operational and regulatory advantages for PAT implementation in HME, enabling better process understanding, reduced waste and more consistent product quality.

Reference

• Application note 1498: Using the Thermo Scientific MarqMetrix All-In-One Process Raman Analyzer for real-time monitoring of a hot-melt extrusion process. Authors: Ceren Yüce, Dirk Hauch, Linxi Chen. Thermo Fisher Scientific, 2025.