Presentation Information
[AP1-03]Considerations on 1/f Flux Noise in High Temperature Superconductor based Magnets used for (Ultra) Low Field Magnetic Resonance Imaging
*Pavel Povolni1, Sergej Maltsev1, Farnaz Tahouni Bonab2, Friedemann Bullinger1, Julius Steiglechner1,3, Dieter Koelle2, Reinhold Kleiner2, Klaus Scheffler1,3, Kai Buckenmaier1 (1. High Field Magnetic Resonance Center - Max Planck Institute for Biological Cybernetics - 72076 Tübingen (Germany), 2. Physikalisches Institut - Center for Quantum Science (CQ) and LISA+ - University of Tübingen - 72076 Tübingen (Germany), 3. Department for Biomedical Magnetic Resonance - University of Tübingen - 72076 Tübingen (Germany))
Keywords:
1/f noise,Low Field MRI,HTS
Purpose
Low-field MRI (LFMRI) is becoming an important tool in medical imaging due to its affordability and accessibility. Since solutions based on permanent magnets or ohmic electromagnets have usage limitations due to field drift or power losses, high-temperature superconductors (HTS) cooled with liquid nitrogen are of particular interest for LFMRI. However, due to the low intrinsic signal-to-noise ratio (SNR) in LFMRI, it must be determined whether HTS-related noise will degrade image quality and reduce the overall SNR. This evaluation is crucial prior to the design of any such device.
Method
The sample was cooled using liquid nitrogen. A Helmholtz coil is employed to generate magnetic flux through the HTS during its cooling process to 77 K. The noise is measured using a Nb SQUID (coupled with a second-stage gradiometer and cooled to 4.2 kelvin) and subsequently fitted during post-processing.To evaluate the impact of HTS noise on an LFMRI, an existing copper-based LFMRI was modeled using HTS. The measured noise characteristics were then extrapolated to the modeled LFMRI and compared to the typical Nyquist noise of an exemplary receiver chain. For simplification, the upper limit of the expected noise level was calculated.
Results
The measured noise characteristics showed that trapped Abrikosov vortices cause significant excess noise at low frequencies. However, the measured spectral density of field noise scales with 1/f².The modeled noise transfer to the LFMRI shows that, for this specific LFMRI magnet, the expected HTS-related noise is lower than the receiver's Nyquist noise.
Conclusion
Because 1/f² behavior was observed, it can be assumed that random flux creep dominates the measurement and creates Brownian noise. The modeled upper limit of the expected HTS noise is significantly smaller than the Nyquist noise of the receiver and therefore neglectable. Therefore, for a construction of a HTS-based liquid Nitrogen cooled LFMRI-magnet, the SNR optimization should be focused on the receiver and sample noise itself.
Low-field MRI (LFMRI) is becoming an important tool in medical imaging due to its affordability and accessibility. Since solutions based on permanent magnets or ohmic electromagnets have usage limitations due to field drift or power losses, high-temperature superconductors (HTS) cooled with liquid nitrogen are of particular interest for LFMRI. However, due to the low intrinsic signal-to-noise ratio (SNR) in LFMRI, it must be determined whether HTS-related noise will degrade image quality and reduce the overall SNR. This evaluation is crucial prior to the design of any such device.
Method
The sample was cooled using liquid nitrogen. A Helmholtz coil is employed to generate magnetic flux through the HTS during its cooling process to 77 K. The noise is measured using a Nb SQUID (coupled with a second-stage gradiometer and cooled to 4.2 kelvin) and subsequently fitted during post-processing.To evaluate the impact of HTS noise on an LFMRI, an existing copper-based LFMRI was modeled using HTS. The measured noise characteristics were then extrapolated to the modeled LFMRI and compared to the typical Nyquist noise of an exemplary receiver chain. For simplification, the upper limit of the expected noise level was calculated.
Results
The measured noise characteristics showed that trapped Abrikosov vortices cause significant excess noise at low frequencies. However, the measured spectral density of field noise scales with 1/f².The modeled noise transfer to the LFMRI shows that, for this specific LFMRI magnet, the expected HTS-related noise is lower than the receiver's Nyquist noise.
Conclusion
Because 1/f² behavior was observed, it can be assumed that random flux creep dominates the measurement and creates Brownian noise. The modeled upper limit of the expected HTS noise is significantly smaller than the Nyquist noise of the receiver and therefore neglectable. Therefore, for a construction of a HTS-based liquid Nitrogen cooled LFMRI-magnet, the SNR optimization should be focused on the receiver and sample noise itself.