Calibration in non-linear NIR spectroscopy using principal component artificial neural networks.

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Chemometrics

Background:

• Non-destructive analysis of pharmaceutical tablets is crucial for quality control.
• Near-infrared (NIR) spectroscopy offers a rapid and efficient analytical technique.
• Developing robust chemometric models is essential for accurate quantitative analysis.

Purpose of the Study:

• To develop and validate a simultaneous, non-destructive method for analyzing antipyriine and caffeine citrate tablets using NIR spectroscopy.
• To compare the performance of principal component artificial neural networks (PC-ANNs) with conventional artificial neural networks (ANNs).
• To evaluate the impact of spectral pretreatment methods on model robustness and accuracy.

Main Methods:

• Near-infrared (NIR) spectroscopy was employed for spectral acquisition.
• Principal Component Analysis (PCA) was used to reduce spectral data dimensionality.
• Artificial Neural Networks (ANNs), including PC-ANNs, were constructed for quantitative analysis.
• Spectral data underwent various pretreatments: first-derivative, second-derivative, Standard Normal Variate (SNV), and Multiplicative Scatter Correction (MSC).
• External validation using a testing set was performed to assess model performance.

Main Results:

• PC-ANNs models demonstrated improved robustness and simplification compared to conventional spectra.
• The Standard Normal Variate (SNV) pretreatment, when used with PC-ANNs, resulted in the lowest training and prediction errors.
• The degree of approximation concept was successfully applied for selecting optimal network parameters.
• PC-ANNs models outperformed standard ANNs models in terms of accuracy and predictive capability.

Conclusions:

• The developed PC-ANNs method, particularly with SNV pretreatment, provides a highly accurate and reliable approach for the simultaneous, non-destructive analysis of antipyriine and caffeine citrate tablets.
• NIR spectroscopy coupled with advanced chemometric techniques offers a powerful tool for pharmaceutical quality control.
• Spectral data preprocessing significantly enhances the performance of chemometric models in quantitative analysis.