Pixel-based Raman hyperspectral identification of complex pharmaceutical formulations.

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Chemometrics

Background:

• Hyperspectral imaging (HSI) is crucial for chemical compound identification.
• Existing chemometric tools often struggle with large datasets and minor compounds due to reliance on factorial decomposition.

Purpose of the Study:

• To introduce a novel pixel-based identification (PBI) approach for Raman hyperspectral imaging data.
• To enable rapid and reliable spectral signature identification, especially for complex pharmaceutical formulations and impurity detection.

Main Methods:

• Developed a pixel-based identification (PBI) strategy using essential spectral pixels (ESP) identified via convex hull calculation.
• Reduced spectral data to purest signatures for efficient database matching.
• Evaluated PBI on genuine and falsified pharmaceutical tablets of varying complexity and map sizes.

Main Results:

• PBI successfully analyzed complex pharmaceutical formulations, detecting impurities down to 0.1% (w/w) in large map sizes (<2 min).
• The method identified all compounds in falsified medicines, including those in trace amounts and varying proportions.
• Enabled identification of spectral signatures not present in the reference database for further investigation.

Conclusions:

• Raman hyperspectral imaging combined with PBI offers a rapid and reliable method for spectral identification in complex pharmaceutical analysis.
• PBI overcomes limitations of traditional methods, proving effective for impurity detection and counterfeit medicine analysis.