Using NV centers in diamond to detect DC to very-low frequency magnetic fields

TL;DR

This work addresses detecting DC to very-low frequency magnetic fields with a compact NV-diamond magnetometer. It introduces a dual-resonance ODMR approach that suppresses temperature drifts while preserving Zeeman sensitivity, enabling accurate single-axis current sensing and potential magnetic communication. The authors demonstrate both close-range current monitoring and meter-scale remote sensing, achieving a noise floor of about 2.3 nT/√Hz and a shot-noise-limited sensitivity near 585 pT/√Hz, with practical performance limited by MW delivery and temperature control. They also explore low-frequency magnetic communication by encoding information with discrete FM (FSK) signals and demonstrate robust symbol recovery under realistic conditions, highlighting the method's relevance for stable, portable magnetic sensing and low-frequency communication in challenging environments.

Abstract

In this work we present a compact and portable tabletop magnetometer that utilizes negatively charged nitrogen-vacancy (NV) centers in diamond. The magnetometer is operated using a dual microwave resonance detection approach in combination with an optically detected magnetic resonance (ODMR) technique (mitigating drifts in results due to changes of the diamond temperature), capable of simultaneously exciting and registering two ODMR transitions. The experimentally measured magnetic field noise-floor is $\approx 2.3~\textrm{nT}\sqrt{\textrm{Hz}}$ while the calculated shot-noise-limited magnetic field sensitivity is $\approx 585~\textrm{pT}\sqrt{\textrm{Hz}}$ when excited with a continuous wave laser at 0.5~W. These results pave the way for realizing a simple set-up magnetometer for precise single axis magnetic field measurements for example for accurate electric current measurements for stabilization purposes and magnetic communication applications.

Figures (12)

• Figure 1: Experimental setup. The microwave frequency was set by changing the voltage that was sent to the VCOs.
• Figure 2: ODMR spectrum with no electrical current in the test wire. The blue curve shows some noise in the $\vert m_S = 0 \longrightarrow m_S = +1\rangle$ transition frequency range, this is attributed to the quality and performance of the VCOs and the electrical boards that they were mounted on that were used in the experiment.
• Figure 3: ODMR spectrum with an electrical current of 50 A in the test wire. Distortions in the ODMR spectra compared to the 0 A situation can be attributed to such effects as thermal changes due to rapid heating of the test wire from the electric current and as the bias field is partially formed by the current, it is also dependent on the current stability (increased magnetic-field inhomogeneity across the illuminated NV ensemble). The limits of the VCO operating range and efficiency also impact this.
• Figure 4: Magnetic field over time (in seconds) for multiple electrical current values.
• Figure 5: Typical noise-floor (2.26 nT/$\sqrt{\textrm{Hz}}$) spectra with no additional time varying magnetic field showing signals at 50 Hz (from the surrounding electronics) and its harmonics. The time constant (1 ms) in combination with the roll-off settings (24 dB/oct) for the lock-in amplifier limits the actual sensing range to slightly below 1 kHz as higher frequency signals are increasing being filtered out from the measured magnetic field (see inset with log-log axis).