An experimental validation of a Bayesian model for quantification in NMR spectroscopy

Area of Science:

• Analytical Chemistry
• Spectroscopy

Background:

• Traditional peak integration in NMR spectroscopy struggles with overlapping peaks and noise.
• Model-based approaches offer advanced quantification and automation for NMR data analysis.

Purpose of the Study:

• To present a general, principled model for NMR signals accounting for various spectral imperfections.
• To validate the model's performance against established methods using simulations and experiments.

Main Methods:

• Development of a general NMR signal model incorporating chemical shifts, relaxation, lineshape, phasing, and baseline distortions.
• Testing the model on simple spectra with well-resolved peaks via simulations and experimental data.

Main Results:

• The model-based approach achieves high accuracy (≥0.01 mol/mol) at high signal-to-noise ratios (>40dB).
• Demonstrates successful quantification (0.05-0.1 mol/mol accuracy) even at low signal-to-noise ratios (<20dB) where phasing is difficult.

Conclusions:

• The presented model-based approach offers a robust and accurate method for quantitative NMR analysis.
• It significantly outperforms traditional methods, especially in challenging spectral conditions with noise and overlapping signals.