Identification of pharmaceutical excipients using NIR spectroscopy and SIMCA

Area of Science:

• Analytical Chemistry
• Pharmaceutical Sciences
• Chemometrics

Background:

• Near-infrared (NIR) spectroscopy is crucial for pharmaceutical analysis.
• Soft Independent Modelling of Class Analogy (SIMCA) is a pattern recognition technique.
• Excipient characterization requires robust identification methods.

Purpose of the Study:

• To evaluate the performance of the SIMCA method for identifying NIR spectra of pharmaceutical excipients.
• To assess SIMCA's effectiveness with non-homogeneous spectral data from various batches and suppliers.
• To compare SIMCA performance across different data pre-processing techniques and confidence levels.

Main Methods:

• Application of Soft Independent Modelling of Class Analogy (SIMCA) to NIR spectral data.
• Collection of at least 15 samples per excipient class, including variations in batches and suppliers.
• Evaluation of SIMCA at 95% and 99% confidence levels.
• Comparison of results using original data, Standard Normal Variate (SNV) transformation, and second derivative pre-processing.

Main Results:

• SIMCA successfully identified NIR spectra of ten pharmaceutical excipients.
• No spectral objects were assigned to an incorrect class, indicating high specificity.
• The study analyzed reasons for rejection rates within the SIMCA model.
• Performance varied slightly based on pre-processing methods and confidence levels.

Conclusions:

• SIMCA is a reliable method for classifying pharmaceutical excipients based on NIR spectra.
• The method maintains high accuracy even with heterogeneous sample data.
• Pre-processing techniques and confidence levels influence SIMCA model performance, warranting careful selection.