Extracting chemical information from spectral data with multiplicative light scattering effects by optical...

Area of Science:

• Spectroscopy
• Chemometrics
• Analytical Chemistry

Background:

• Spectroscopic analysis of complex mixtures faces challenges due to sample variability.
• Variations in optical path length cause multiplicative light scattering, perturbing spectral data.
• Existing algorithms for light scattering correction often require unavailable auxiliary information.

Purpose of the Study:

• To introduce a novel algorithm, optical path-length estimation and correction (OPLEC), for addressing multiplicative light scattering.
• To eliminate the need for additional spectral information in light scattering correction.
• To evaluate OPLEC's performance against established methods like EMSC and EISC.

Main Methods:

• Development and application of the optical path-length estimation and correction (OPLEC) algorithm.
• Testing OPLEC on near-infrared transmittance spectral data from powder mixtures and wheat kernels.
• Comparative analysis of OPLEC with extended multiplicative signal correction (EMSC) and extended inverted signal correction (EISC).

Main Results:

• The EMSC algorithm was unsuitable for wheat kernel data due to missing information.
• EISC applied to powder mixture data yielded inaccurate conclusions and an unacceptable calibration model.
• OPLEC effectively mitigated light scattering effects and enhanced calibration model prediction accuracy for both datasets.

Conclusions:

• OPLEC offers a robust solution for multiplicative light scattering in spectroscopic analysis.
• The algorithm performs effectively without requiring additional information about calibration samples.
• OPLEC significantly improves the accuracy of predictive models for complex mixtures with inherent variability.