Optimizing Savitzky-Golay parameters for improving spectral resolution and quantification in infrared spectroscopy

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Chemometrics

Background:

• The Savitzky-Golay (SG) filter is widely used for spectral data preprocessing.
• Derivative calculations are essential for resolving overlapping signals and enhancing spectral features.
• Instrumental and sample variations can introduce anomalies like noise and baseline drift.

Purpose of the Study:

• To investigate the impact of SG filter parameter optimization on spectral derivative calculations.
• To evaluate the performance of SG filtering on simulated and measured infrared spectral data.
• To determine the optimal order of applying SG filtering and Extended Multiplicative Signal Correction (EMSC).

Main Methods:

• Simulated infrared spectral data generation with various anomalies.
• Application of the Savitzky-Golay (SG) numerical algorithm for derivative calculation.
• Partial least-squares regression analysis.
• Comparison of SG filter performance with different window sizes and polynomial orders.
• Evaluation of the combined SG and EMSC procedures.

Main Results:

• Optimization of SG parameters (window size, polynomial order) is critical for accurate derivative spectra.
• Optimal SG window size is component-specific, necessitating careful selection.
• Spectral noise level can serve as a guideline for assessing window size.
• The SG derivative calculation must precede EMSC normalization for effective data correction.

Conclusions:

• Proper optimization of Savitzky-Golay parameters significantly improves spectral derivative quality.
• The order of applying SG and EMSC is crucial for successful spectral data preprocessing.
• This study provides insights into optimizing SG filter application for infrared spectral analysis.