Correction of temperature-induced spectral variation by continuous piecewise direct standardization

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Chemometrics

Background:

• Process analytical applications face challenges with constant measurement conditions.
• External variables like temperature significantly impact measurement results, causing biased predictions in multivariate calibration models (e.g., PLS).
• Nonlinear temperature effects on Near-Infrared (NIR) spectra are a key concern.

Purpose of the Study:

• To develop and evaluate a new method for correcting external influences, specifically temperature effects, on spectral data.
• To generalize the discrete Piecewise Direct Standardization (PDS) method for continuous nonlinear influences.
• To improve the accuracy of multivariate calibration models in process analytical applications.

Main Methods:

• Developed Continuous Piecewise Direct Standardization (CPDS) as a generalization of PDS.
• Applied CPDS to shortwave NIR spectra of ethanol/water/2-propanol mixtures.
• Measured spectra across a temperature range of 30-70 degrees C.

Main Results:

• CPDS effectively removed nonlinear temperature effects from NIR spectra.
• The method achieved near-complete correction of temperature-induced spectral variations.
• Predictions of chemical component mole fractions closely matched results obtained under constant temperature conditions.

Conclusions:

• CPDS is a robust method for correcting continuous nonlinear influences, such as temperature, in process analytical applications.
• The developed method significantly enhances the reliability of multivariate calibration models when measurement conditions fluctuate.
• CPDS offers a valuable solution for maintaining prediction accuracy in real-world process monitoring.