Multi-objective optimization for an automated and simultaneous phase and baseline correction of NMR spectral data

Area of Science:

• Chemometrics
• Spectroscopy
• Data Analysis

Background:

• Spectral data preprocessing is crucial for chemometric analyses.
• Nuclear Magnetic Resonance (NMR) spectra often require phase and baseline correction.
• Automated and fast preprocessing is desirable for high-resolution spectral series.

Purpose of the Study:

• To develop a novel, automated method for simultaneous phase and baseline correction of NMR spectra.
• To improve preprocessing efficiency and accuracy in chemometric workflows.
• To offer an alternative to manual or consecutive correction methods.

Main Methods:

• A multi-objective optimization (Pareto optimization) approach is employed.
• An objective function with strong penalty constraints and weaker regularization is utilized.
• A modified Whittaker smoother is used for baseline correction, including zero baseline region detection.

Main Results:

• The new method achieves simultaneous, automated phase and baseline correction without manual input.
• Pareto optimization yields superior results compared to consecutive correction steps.
• Experimental NMR spectra demonstrate the method's functionality, verified against gravimetric data.

Conclusions:

• The developed simultaneous correction method is efficient and accurate for NMR spectral preprocessing.
• This automated approach offers significant advantages over traditional consecutive correction techniques.
• The method is a valuable tool for chemometric analysis of NMR data.