Extending the use of Earth's Field NMR using Bayesian methodology: application to particle sizing

Area of Science:

• Physics
• Chemistry
• Biophysics

Background:

• Low-field magnetic resonance imaging (MRI) offers potential for portable and cost-effective structural characterization.
• Conventional MRI techniques struggle with insufficient signal-to-noise ratios (SNR) at low magnetic fields.
• Developing methods to extract spatial information from low-field MRI data is crucial for broader applications.

Purpose of the Study:

• To apply a Bayesian magnetic resonance approach for object sizing using low magnetic field measurements.
• To overcome signal-to-noise limitations inherent in low-field MRI for structural analysis.
• To validate the method's accuracy and uncertainty estimation for object dimension determination.

Main Methods:

• Utilized an Earth's Field Nuclear Magnetic Resonance (EFNMR) spectrometer with pre-polarization for measurements.
• Applied a Bayesian inference framework to analyze low-field magnetic resonance data.
• Conducted numerical simulations to optimize k-space sampling and assess performance across varying SNRs.

Main Results:

• Successfully demonstrated object sizing, specifically determining the radius of spheres.
• Achieved an accuracy of ±1mm in determining the sphere radius (9.5mm).
• Confirmed that the Bayesian posterior distribution accurately estimates measurement uncertainty.

Conclusions:

• The Bayesian magnetic resonance approach is effective for object sizing in low-field NMR, even with low SNR.
• This method provides a reliable way to obtain structural information where conventional imaging fails.
• The technique offers accurate dimensional measurements with reliable uncertainty quantification.