A New cw-NMR Q-meter for Dynamically Polarized Targets for Particle Physics

J. Maxwell, James Brock, C. Cuevas, Hai Dong, C. Keith, J. PiercearXiv 2025 · DOI · arXiv

Abstract

Polarized solid targets produced via Dynamic Nuclear Polarization rely on Continuous-Wave Nuclear Magnetism Resonance (cw-NMR) measurements to accurately determine the degree of polarization of bulk samples polarized to nearly 100%. Since the late 1970s, phase-sensitive detection methods have been utilized to observe the magnetization of a sample as a small change in inductance under RF excitation near the Larmor frequency of the nuclear species of interest, using a device known as a Q-meter. Liverpool Q-meters, produced in the UK in the 80s and 90s, have been the workhorse devices for these targets for decades; however, their age and scarcity have meant new systems are needed. We describe a Q-meter system designed and built at Jefferson Lab in the Liverpool style to have comparable electronic performance with several improvements to update and adapt the devices for modern use.

Introduction

In high-energy particle physics, experiments often utilize dynamically polarized solid targets, such as protons or deuterons, to investigate the spin-dependent properties of matter. Achieving and measuring near 100% polarization is critical. The standard instrument for this measurement has long been the "Liverpool Q-meter," which detects minute changes in a coil's inductance as it sweeps through the Larmor frequency of the target nuclei.

However, these legacy devices are no longer manufactured and rely on obsolete components. A significant technical challenge for any replacement is maintaining extreme linearity over a vast dynamic range: the thermal equilibrium calibration signal is often 300 times smaller than the signal from a fully polarized target. The Jefferson Lab team sought to develop a modern successor that preserves the proven sensitivity of the Liverpool design while integrating 21st-century electronics and control systems.

Design and Methods

The new system retains the core frequency down-mixing analog architecture of the original Liverpool design but implements it with modern hardware and modularity. Key improvements include:

  • Modern RF Components: Decades-old amplifiers were replaced with surface-mount RF gain blocks (e.g., ADL5531), providing superior linearity and thermal stability.
  • Modular Shielding: To minimize cross-talk, the circuit is divided into EMI-shielded modules, including the high-frequency mixer, low-frequency amplifier, and reference conditioner.
  • Electronic Tuning: Unlike legacy systems that required manual adjustment of rotary capacitors and coaxial cable lengths, the new design utilizes voltage-tunable varactor diodes and electronic phase shifters. This allows for complete software-based remote tuning.
Comparison of the traditional Liverpool Q-meter circuit (left) and the modernized Jefferson Lab design (right).
  • Advanced DAQ and Software: A custom FPGA-based data acquisition (DAQ) system with 24-bit Analog-to-Digital Converters (ADCs) was developed. This high resolution eliminates the need for secondary analog amplification stages. The control software was also modernized, moving from proprietary LabVIEW to a multi-threaded Python interface.

Key Results

The modernized Q-meter was validated during a nine-month physics run at Jefferson Lab, successfully capturing over 180,000 events. The integration of FPGA control for the RF signal generator reduced frequency hopping settling times to 10 µs, resulting in a data-taking rate more than three times faster than previous systems.

Analysis of an enhanced proton NMR signal showing raw data, baseline subtraction, and the final extracted signal.

Direct A/B comparisons with legacy Liverpool hardware confirmed the new system's performance. The device maintained strict linearity up to approximately 65% proton polarization. The signal processing chain effectively extracts the final NMR signal from a large background "Q-curve," as shown in the analysis of enhanced proton signals.

Discussion and Future Work

While electronic tuning offers significant flexibility, it introduces the risk of "false tunes" if not carefully managed. Furthermore, the use of a mechanical RF switch for A/B testing introduced additional noise, which slightly limited pristine signal-to-noise comparisons during validation.

Future developments include adding a second mixer for simultaneous quadrature measurement (capturing both absorptive and dispersive signal components) and exploring "algorithmic down-mixing" to directly digitize high-frequency RF signals. Additionally, the team is investigating machine learning techniques to improve the polynomial fitting of background signals, which would further enhance the accuracy of polarization measurements under varying experimental conditions.

Summary updated

Abstract

Polarized solid targets produced via Dynamic Nuclear Polarization rely on Continuous-Wave Nuclear Magnetism Resonance measurements to accurately determine the degree of polarization of bulk samples polarized to nearly 100%. Since the late 1970's phase sensitive detection methods have been utilized to observe the magnetization of a sample as a small change in inductance under RF excitation near the Larmor frequency of the nuclear species of interest, using a device known as a Q-meter. Liverpool Q-meters, produced in the UK in the 80's and 90's, have been the workhorse devices for these targets for decades, however their age and scarcity has meant new systems are needed. We describe a Q-meter system designed and built at Jefferson Lab in the Liverpool style to have comparable electronic performance with several improvements to update and adapt the devices for modern use.
BibTeX
@misc{maxwell2025new,
  author = {J. Maxwell and James Brock and C. Cuevas and Hai Dong and C. Keith and J. Pierce},
  title = {{A New cw-NMR Q-meter for Dynamically Polarized Targets for Particle Physics}},
  year = {2025},
  eprint = {2506.14637},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2506.14637},
  howpublished = {arXiv},
}
Figure