Determination of moisture content in freeze-dried materials by FT-NIR spectroscopy

Importance of the topic

Accurate determination of residual moisture in freeze-dried pharmaceutical products is critical because water content strongly influences chemical stability, physical integrity, and shelf life. Traditional gravimetric or Karl Fischer methods are accurate but destructive, reagent-dependent, and slow. Near-infrared (NIR) spectroscopy—especially Fourier transform NIR (FT-NIR)—offers rapid, non-destructive, reagent-free analysis suitable for at-line or near-line monitoring during lyophilization, supporting process control and reducing material waste.

Objectives and overview of the study

The study evaluated the feasibility of FT-NIR spectroscopy to quantify moisture in lyophilized folinic acid calcium salt (Leucovorin Ca). Main aims were to demonstrate a non-destructive method capable of measuring residual water across typical product batches and to build a calibration model for rapid moisture determination as a potential alternative to routine Karl Fischer or gravimetric assays.

Methodology

Eight batches of lyophilized Leucovorin Ca with gravimetrically determined water contents spanning 2.65% to 8.04% were analyzed. Key procedural points:

• Samples remained sealed in their original clear glass vials to avoid hygroscopic changes and were measured through the vial wall.
• Diffuse reflectance FT-NIR spectra were collected from 12,000 to 4,000 cm-1 with 4 cm-1 resolution and 1 minute acquisition time; multiple spectra per vial were recorded to account for measurement variability.
• Spectral preprocessing included second-derivative calculation to remove baseline offsets and slope.
• Stepwise multiple linear regression (SMLR) was used to select spectral regions and build the quantitative calibration model; two component regions were included by the algorithm.
• Leave-one-out cross-validation was applied to evaluate model robustness.

Instrumentation used

The measurements and model development used:

• Thermo Scientific Antaris FT-NIR Analyzer equipped with an Integrating Sphere module (diffuse reflectance geometry).
• Internal gold reference flag as background.
• Thermo Scientific TQ Analyst Method Development software for calibration (SMLR) and model building.

Main results and discussion

Principal analytical outcomes and interpretations:

• Spectral changes correlated strongly with moisture, especially in O–H overtone and combination regions; baseline shifts likely reflected vial and sample morphology differences.
• Calibration performance over the 2.65–8.04% moisture range: correlation coefficient (R2) = 0.9996 and RMSEC = 0.0610%.
• Cross-validation (leave-one-out) results: R2 = 0.9989 and RMSECV = 0.0987%.
• The particle size and composition were sufficiently uniform across samples so that pathlength or scattering correction was not required for this dataset.
• Using sealed vials minimized hygroscopic alteration but implies the method relies on sufficient NIR transmission/reflectance through vial walls; vial variability contributed to baseline differences.

The high R2 and low RMSE values indicate that FT-NIR with SMLR can quantify residual moisture in this lyophilized product with precision comparable to reference methods, enabling near-real-time assessment of lyophilization end-points.

Benefits and practical applications

Practical advantages demonstrated or implied by the study:

• Non-destructive, reagent-free measurements preserve expensive pharmaceutical material and avoid chemical waste.
• Rapid analysis (about 1 minute per spectrum) supports at-line or near-line monitoring of lyophilization, allowing earlier decision-making and potential reduction of over-drying or batch loss.
• Capability to analyze through clear glass vials simplifies sampling logistics for production environments.
• Multicomponent sensitivity of FT-NIR opens potential extension to simultaneous monitoring of other excipients or product attributes.

Future trends and potential applications

Potential directions for broader implementation and method improvement:

• Adoption of modern FT-NIR hardware (e.g., Antaris II and later instruments) will improve speed, S/N, and robustness for production use.
• Wider calibrations employing larger, more diverse sample sets and multivariate methods such as PLS regression could increase model transferability across lots, vial types, and instruments.
• Implementation within PAT frameworks: inline or at-line sensors, automated model application, and feedback control to stop lyophilization at target moisture.
• Data transfer and model standardization approaches (standardization algorithms, transfer learning) to enable deployment across sites and instruments.
• Integration with process analytics, hyperspectral mapping, and advanced preprocessing to handle scattering, morphology changes, and container effects.

Conclusion

The feasibility study demonstrates that FT-NIR spectroscopy, measured by diffuse reflectance through sealed vials and modeled by SMLR, provides a fast, accurate, and non-destructive method to quantify residual moisture in lyophilized Leucovorin Ca within the tested range (2.65–8.04%). High calibration and cross-validation statistics indicate the approach is suitable for at-line monitoring of lyophilization. For routine production deployment, expanded calibration sets, consideration of vial and morphology variability, and modern FT-NIR instrumentation are recommended to ensure robustness and transferability.

References

The source material is an application note describing FT-NIR determination of moisture in freeze-dried Leucovorin Ca using a Thermo Scientific Antaris FT-NIR analyzer and TQ Analyst software (Thermo Fisher Scientific application note AN50780).