Development of a univariate calibration model for pharmaceutical analysis based on NIR spectra

Area of Science:

• Analytical Chemistry
• Pharmaceutical Analysis
• Spectroscopy

Background:

• Near-infrared spectroscopy (NIRS) is vital for real-time pharmaceutical quality control.
• Traditional NIRS requires complex multivariate models to address spectral collinearity and poor selectivity.
• Developing simpler calibration models for NIRS is crucial for broader application.

Purpose of the Study:

• To develop a novel, simplified methodology for constructing near-infrared spectroscopy (NIRS) calibration models.
• To enable accurate quantification of active principle ingredients (APIs) in pharmaceuticals with varying physical properties.
• To offer an alternative to complex multivariate calibration methods.

Main Methods:

• A new calibration method based on correlating analyte spectra with sample spectra using a correlation coefficient.
• Least-squares regression is applied to relate correlation coefficients to sample concentration.
• API concentration is predicted by interpolating correlation coefficients of validation samples.

Main Results:

• The proposed method successfully quantifies APIs in diverse pharmaceutical forms (granules, coated/non-coated tablets).
• Achieved error values below 1-2% across different sample types using a single model.
• Results are comparable to traditional PLS1 multivariate models requiring separate calibrations per sample type.

Conclusions:

• The developed methodology provides a straightforward and effective approach for NIRS calibration model construction.
• This method allows for rapid prediction of API concentration in samples with varying physical characteristics.
• It serves as a valuable alternative to multivariate calibration, simplifying pharmaceutical analysis.