A novel improved method for analysis of 2D diffusion-relaxation data--2D PARAFAC-Laplace decomposition

Area of Science:

• Nuclear Magnetic Resonance (NMR) Spectroscopy
• Multivariate Data Analysis
• Food Science

Background:

• Analyzing complex diffusion-relaxation correlation NMR spectra is challenging.
• Existing methods struggle with interpretation and quantification of components.
• Advanced data decomposition techniques are needed for accurate analysis.

Purpose of the Study:

• To demonstrate the utility of multi-linear PARAFAC modeling prior to 2D-Laplace inversion for NMR spectra.
• To improve the interpretation and quantification of diffusion-relaxation components.
• To accurately determine oil-to-water ratios in complex mixtures.

Main Methods:

• Utilized a multi-linear PARAFAC model for decomposing 2D diffusion-relaxation NMR spectra.
• Employed a 300 MHz NMR spectrometer with PGSTE and CPMG pulse sequences.
• Applied 2D-Laplace inversion to transform data to the T(2)-D domain.
• Analyzed seventeen mixtures of wheat flour, starch, gluten, oil, and water.

Main Results:

• PARAFAC decomposition extracted two unique diffusion-relaxation components, explaining 99.8% of data variation.
• Successfully transformed components to the T(2)-D domain and assigned them to oil and water.
• Identified distinct diffusion and relaxation times for oil and water populations.
• Demonstrated accurate oil-to-water ratio determination, correlating perfectly with known values.
• Showcased effective filtering of artefacts from Laplace transformation.

Conclusions:

• The PARAFAC-enhanced 2D-Laplace inversion method offers superior potential for diffusion-relaxation spectra analysis.
• This approach significantly improves both the interpretation and quantification of spectral data.
• The method provides accurate insights into the composition of complex mixtures, such as food products.