Sensor Configuration and Algorithms for Power-Line Interference Suppression in Low Field Nuclear Magnetic Resonance

Area of Science:

• Nuclear Magnetic Resonance Spectroscopy
• Quantum Sensing

Background:

• Low-field (LF) NMR offers advantages for studying heteronuclear J-coupling and material fine structure.
• Interference from power-line harmonics and fixed-frequency noise can degrade LF-NMR spectra, especially in open or shielded environments.
• Disturbed J-coupling spectra hinder detailed analysis of matter at LF.

Purpose of the Study:

• To develop and demonstrate effective noise suppression algorithms for LF-NMR spectroscopy.
• To improve the signal-to-noise ratio (SNR) of J-coupling spectra recorded at LF.
• To address challenges posed by low SNR (<1) in LF-NMR data acquisition.

Main Methods:

• Utilized a multi-channel superconducting quantum interference device (SQUID) sensor configuration.
• Employed discrete wavelet analysis (DWA) combined with least squares method (LSM) and gradient descent (GD) for noise suppression.
• Leveraged spatial correlation of interferences across superconducting sensors for de-noising.

Main Results:

• Successfully measured the J-coupling spectrum of 2,2,2-trifluoroethanol.
• Demonstrated experimental validation of the proposed noise suppression algorithms.
• The DWA-LSM algorithm significantly reduced noise, recovering SNR > 1 for most signal peaks.
• The DWA-GD algorithm further improved SNR but required more computational time.

Conclusions:

• The developed DWA-based algorithms effectively suppress noise in LF-NMR spectra.
• The choice between DWA-LSM and DWA-GD depends on the specific application's need for accuracy versus computational speed.
• These methods enhance the feasibility of detailed J-coupling analysis in LF-NMR.