Investigation of Different Sampling Techniques for the Analysis of Capsule Contents by Fourier Transform Near-Infrared Spectroscopy

Significance of the topic

The reliable analysis of hard gelatin capsules is a frequent challenge in pharmaceutical development and quality control. Capsules account for a large share of solid oral dosage forms and demand rapid, robust analytical methods for formulation identification, assay and content uniformity. Fourier transform near-infrared (FT-NIR) spectroscopy offers a fast, low-preparation alternative to chromatographic assays, enabling non-destructive or minimally destructive testing that can support at-line and in-process control.

Objectives and overview of the study

This application note evaluates FT-NIR (Thermo Scientific Antaris MDS) for both qualitative and quantitative analysis of conventional hard gelatin capsules. Two sampling workflows were compared: intact whole-capsule reflectance (non-destructive) and reflectance of the emptied powder contents (destructive but no extraction solvents). Three sample sets were used as model cases: 1) qualitative discrimination among four formulations sharing a common active, 2) quantitative distinction of multiple clinical supplies of common formulations, and 3) low-level assay calibration around a marketed formulation (nominal 5 mg active per capsule; 90–110% range). The study assesses viability, accuracy and precision of each sampling approach and reports chemometric models and performance metrics.

Methodology and sampling

- Samples: standard hard gelatin capsules filled with powders; three sets varying in formulation composition and active levels (including sustained-release variants).
- Sampling modes compared:

• Whole-capsule reflectance: capsules placed intact on the integrating-sphere window (non-destructive).
• Powder reflectance: capsule contents emptied into micro sample cups with glass windows or small glass vials and analyzed through the glass (destructive but avoids solvent extraction).

- Spectrometer and data: Thermo Scientific Antaris Method Development Sampling (MDS) FT-NIR system with an integrating sphere; internal gold flag used as background reference. Data acquisition and chemometric modeling used RESULT and TQ Analyst software.
- Spectral pretreatment: Norris second derivative (segment 11, gap 0) was applied for quantitative models; derivative not required for some qualitative discrimination tasks.

Used Instrumentation

- Thermo Scientific Antaris FT-NIR Method Development Sampling (MDS) analyzer (integrating sphere, transmission compartment, fiber probe options).
- RESULT software for data collection and ValPro system qualification for instrument performance verification.
- TQ Analyst for multivariate modeling (Discriminant Analysis and Partial Least Squares regression).

Data analysis approach

- Qualitative classification (sample set 1): Discriminant Analysis with Mahalanobis distance as matching metric across the 4000–10,000 cm-1 region. A threshold of ~3 Mahalanobis units was used to indicate reasonable class separation.
- Quantitative calibration (sample sets 2 and 3): Partial Least Squares (PLS) regression with RMSEC and RMSECV used to evaluate calibration and cross-validation performance. Correlation coefficients were reported to assess fit.

Main results and discussion

- Sample set 1 (formulation discrimination): Both whole-capsule reflectance and powder reflectance allowed unambiguous distinction among the four dosage forms. The DA model produced correct classification for all capsules in the calibration set, with best-match Mahalanobis distances well below 3 and next-best matches >3 in most cases, indicating clear separation. For this qualitative task, non-destructive whole-capsule measurement was adequate.

- Sample set 2 (quantitative distinction of clinical supplies): Reasonable PLS calibrations were obtained from both sampling modes. Key metrics:

• Whole capsule (PLS, 3 factors): correlation 0.9991, RMSEC 0.143 mg/capsule, RMSECV 0.742 mg/capsule.
• Powder (PLS, 2 factors): correlation 0.9993, RMSEC 0.126 mg/capsule, RMSECV 0.140 mg/capsule.

Powder-based models gave substantially better cross-validation performance (lower RMSECV), indicating higher predictive robustness when the shell contribution is removed.

- Sample set 3 (assay of low-level active, nominal 5 mg per capsule): This is a demanding case because the active is only ~4% of the formulation. Results show:

• Whole capsule (PLS, 2 factors): correlation 0.8686, RMSEC 0.152 mg/capsule, RMSECV 0.302 mg/capsule.
• Powder (PLS, 2 factors): correlation 0.9960, RMSEC 0.0274 mg/capsule, RMSECV 0.151 mg/capsule.

Here the destructive powder analysis was clearly superior and necessary to obtain acceptable assay performance. Even so, the study notes that calibration quality could be improved with larger standard sets and optimized sample containment (smaller vials/cups to increase signal uniformity).

- General observations: The capsule shell introduces spectral interference that commonly reduces quantitative accuracy. For many routine discrimination tasks, whole-capsule (non-destructive) FT-NIR is sufficient and offers logistical advantages. For low-level assay and highest accuracy, removing the fill and measuring powder reflectance yields better results. All FT-NIR approaches tested required minimal sample preparation, no solvent extraction, and short measurement times (<1 minute per capsule).

Benefits and practical applications

- Rapid, high-throughput screening for formulation identification and supply chain verification without destroying the unit (when suitable).
- At-line or near-patient testing: portable/benchtop FT-NIR enables testing outside central labs.
- Reduced sample preparation and elimination of solvents lower operator exposure and waste handling compared with HPLC-based assays.
- Powder reflectance measurements enable quantitative assay and content-uniformity evaluations when shell interference prevents non-destructive measurement.

Future trends and potential applications

- Integration of FT-NIR into PAT (process analytical technology) workflows for real-time capsule quality control and line-side monitoring.
- Improved chemometric strategies (robust preprocessing, variable selection, model transfer methods) to reduce shell interference and increase model transferability across instruments and sites.
- Automated sample presentation and probe-based or conveyorized sampling for true in-line capsule screening.
- Combining FT-NIR with other spectroscopic or imaging modalities for enhanced discrimination of complex formulations (e.g., coated or multi-layer capsules).
- Regulatory acceptance pathways and validation packages for FT-NIR methods tailored to content uniformity and release testing.

Conclusions

FT-NIR spectroscopy using the Antaris MDS platform is a viable tool for both qualitative and quantitative analysis of hard gelatin capsules. Non-destructive whole-capsule reflectance frequently provides adequate discrimination for formulation identification and can be used when preserving sample integrity is required. However, the gelatin shell is a non-negligible source of spectral interference that commonly degrades quantitative performance; measuring powder fills (destructive but solvent-free) yields markedly better calibration accuracy for demanding assay tasks, especially when the active is present at low percentage. Overall, FT-NIR enables rapid, low-preparation testing that can reduce reliance on time-consuming chromatographic methods in many routine capsule-analytical workflows.

References

Thermo Fisher Scientific. Application Note 51595: Investigation of Different Sampling Techniques for the Analysis of Capsule Contents by Fourier Transform Near-Infrared Spectroscopy. AN51595_E, April 2008. Antaris FT-NIR Method Development Sampling (MDS) system; RESULT and TQ Analyst software.