Zero- to ultralow-field J-spectroscopy with a diamond magnetometer
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Communications Chemistry |
| Published | 2026-03-06 |
| Authors | Muhib Omar, Jingyan Xu, Raphael Kircher, Pouya Sharbati, Shaowen Zhang |
| DOI | 10.1038/s42004-026-01962-3 |
A nitrogen-vacancy center ensemble in diamond detected hyperpolarized acetonitrile at zero field with 13 pT/√Hz sensitivity above 5 Hz. The magnetometer resolved the expected 1.7 Hz and 3.4 Hz J-coupling features. It also recorded NMR signals up to 580 Hz under applied bias fields.
Zero-field detection at 1.7 and 3.4 Hz
Section titled “Zero-field detection at 1.7 and 3.4 Hz”The sample was [¹⁵N]-acetonitrile hyperpolarized by signal amplification by reversible exchange (SABRE). At zero field, its -1.69 Hz scalar J coupling produced one transition at the J frequency and another at twice that frequency. The measured peaks at 1.7 Hz and 3.4 Hz matched this pattern.
The authors also recorded spectra at 10 nT and 40 nT. Measured patterns agreed with numerical simulations, including more complex allowed transitions as the bias field approached the J-coupling strength.
Both the commercial rubidium optically pumped magnetometer (OPM) and the NV-diamond magnetometer resolved the zero-field pattern.
How zero-bias NV magnetometry works
Section titled “How zero-bias NV magnetometry works”Optically detected magnetic resonance (ODMR) converts a magnetic field into a change in diamond photoluminescence. Optical pumping prepares the NV electron spin in mₛ = 0. Microwave drive transfers population into the mₛ = ±1 states, whose photoluminescence is lower.
At zero bias, the mₛ = ±1 transitions remain degenerate. A magnetic field broadens the resonance to first order instead of shifting it, so conventional ODMR is insensitive inside a shield.
The sensing diamond contained 3.7 ppm NV centers and less than 10 ppm nitrogen. It was a truncated pyramid 180 µm high with a 500² µm² base.
The authors modulated the microwave at 140 kHz with 100 kHz amplitude to create a dispersive lock-in error signal. A proportional-integral-derivative loop held it on the temperature-dependent zero-field feature. Its frequency changes with temperature, so the lock stabilizes the operating point.
Field modulation at 6.12 kHz and approximately 3 µT followed the diamond [100] direction. It produced a second signal linear in field near zero. Separate demodulation frequencies avoided cross-modulation. This protocol did not require circularly polarized microwaves.
SABRE sample and acquisition sequence
Section titled “SABRE sample and acquisition sequence”The sample contained [Ir(IMes)(COD)Cl] at 5 mM in acetonitrile with 125 mM benzylamine. The precursor became the active catalyst after parahydrogen was dissolved at 7 bar, 15 minutes before detection. Parahydrogen then bubbled through the liquid at 20 scc min−1.
Acetonitrile was enriched to 25% ¹⁵N, compared with about 0.4% natural abundance. A 12.5 mm diameter glass sphere held the liquid inside an MS-2 magnetic shield.
A piercing solenoid supplied bias fields, while a Helmholtz pair applied an 860 µs, 6 µT pulse along the x axis. The pulse initiated free nuclear magnetization decay.
Each diamond acquisition lasted 40 s and used 50 time-domain averages. The lock-in amplifier sampled at 14 k samples per second. Its low-pass bandwidth was 100, 350, or 600 Hz, depending on the expected signal.
Fourier transformation produced a single-sided amplitude spectrum, scaled so a 100 pT sinusoid produced a peak of equal amplitude.
Sensitivity and magnetic noise
Section titled “Sensitivity and magnetic noise”The magnetically insensitive noise floor averaged 13 pT/√Hz below 50 Hz. A separate estimate over 20 to 30 Hz gave the same value. The sensitive trace followed a 1/f profile below 5 Hz and matched the insensitive floor above 5 Hz.
Laser noise and electronic noise did not explain the low-frequency excess. The authors excluded environmental magnetic noise using OPM measurements and reduced temperature drift through microwave locking. They considered the field modulation source the likely cause.
Calibration used a known test field. For a 9 s Hann-windowed trace and 50 averages, the expected RMS amplitude was 0.8 pT. The measured spectrum histogram gave σ = 1.4 pT and an RMS of about 2 pT. The authors linked the difference to the 1/f tail.
The demonstrated noise floor was 13 pT/√Hz, while the stated photon-shot-noise limit was 2 pT/√Hz. The authors suggest better laser cancellation and a high-pass-filtered modulation current could approach that limit.
Bandwidth and comparison with other sensors
Section titled “Bandwidth and comparison with other sensors”The diamond magnetometer trades sensitivity for proximity and bandwidth. The paper’s table lists 13 pT/√Hz for NV diamond, compared with 10 fT/√Hz for the QuSpin OPM. The NV stand-off scale was below 0.2 mm, while the OPM stand-off exceeded 5 mm.
| Sensor | Sensitivity | Bandwidth | Stand-off scale | Dynamic range |
|---|---|---|---|---|
| NV diamond | 13 pT/√Hz | ≈3 kHz | <0.2 mm | ≈10 µT |
| QuSpin OPM | 10 fT/√Hz | 150 Hz | >5 mm | ±5 nT |
| SQUID | 2.5 fT/√Hz | DC to MHz | 30 mm | ≈100 nT |
| Induction coil | 20 fT/√Hz | 3 to 10 kHz | 20 mm | Unspecified |
| Magnetoresistive | ≈1 pT/√Hz | 142.5 Hz | ≈5 mm | ≈100 µT |
Measured NMR features reached 580 Hz with the diamond sensor. The OPM did not respond above 500 Hz because of its sixth-order hardware low-pass filter. Its specified bandwidth was 150 Hz.
The SQUID option provides 2.5 fT/√Hz across DC to MHz, but requires cryogenic cooling. The induction coil lists 20 fT/√Hz and 3 to 10 kHz, with a 20 mm coil diameter.
At 0.8 and 7 µT, ¹⁵N precession lines reached about 1 pT in OPM data. The J-coupling spectra peaked near 35 pT. The NV sensor did not resolve the ¹⁵N lines within 50 averages. The OPM data were acquired without averaging.
Fitted field slopes were 41.1 ± 0.1 Hz/µT for ¹H and -4.1 Hz/µT for ¹⁵N. These matched free-nucleus values of 42.57 and -4.3 Hz/µT.
Distance scaling and line broadening
Section titled “Distance scaling and line broadening”The distance series normalized integrated 2J-peak strength to its value at 14.3 mm. The 2J signal was integrated from 2.5 to 6 Hz. A 6.2 to 9.7 Hz interval supplied the subtracted baseline.
The closest center-to-center distance was 6.8 mm because the piercing solenoid wall limited access. The paper separately reports a possible sensor-to-sample stand-off below 0.2 mm.
Signal strength followed inverse-cubic dipole scaling at larger separations. This behavior was observed from 9.3 mm between the sensor-volume and sample-volume centers.
The maximum normalized diamond signal reached three times the 14.3 mm value. The closest data then fell below the simple dipole extrapolation.
Both sensors showed broadening at short distances, but the effect was stronger with diamond. The authors linked the broadening to transverse field variation, field gradients, and slightly magnetic sensor components. Position-by-position shimming with the OPM compensated for these effects.
Engineering implications and unresolved issues
Section titled “Engineering implications and unresolved issues”Zero field does not remove the need for polarization. The signals came from SABRE-hyperpolarized molecules, so the experiment did not detect unpolarized samples directly.
The sensor still used a modulated field of about 3 µT at 6.12 kHz. It also required microwave locking and calibration against a known field.
The authors propose high-pass filtering the modulation current and reducing magnetic components to control broadening. They also propose improved laser-noise cancellation toward 2 pT/√Hz.
The broad bandwidth may suit molecules with larger J couplings. The authors cite calculated platinum couplings in the tens of kHz and PF₆ couplings of several multiples of 700 Hz. Those bands were not measured here.
Possible polarization methods include high-field prepolarization with sample shuttling, dynamic nuclear polarization, photo-induced dynamic nuclear polarization, and other parahydrogen-based approaches. These are proposed extensions, not demonstrated results.
The authors argue that smaller samples can be measured without fundamental sensitivity loss when magnetization behaves as a classical dipole. This remains a scaling argument, not a measured miniaturized-sample result.
The work suggests portable noninvasive diagnostics and sensing through metal enclosures, but did not demonstrate either application. The demonstrated result is a 13 pT/√Hz sensor that resolves 1.7 Hz and 3.4 Hz features at zero DC bias.
Materials and support
Section titled “Materials and support”For this magnetometer application, the 6CCVD diamond materials catalog lists MPCVD-grown single-crystal and polycrystalline diamond plates and wafers with thickness from 0.1 to 500 µm. The paper needs an NV-center element with 3.7 ppm NV, less than 10 ppm nitrogen, and a 180 µm-high shape with a 500² µm² base. The linked engineering support page can assess custom shaping and material suitability, although the catalog does not state the required NV concentration.
View Original Abstract
Abstract Nuclear magnetic resonance (NMR) is a powerful tool for probing molecular structure and dynamics, but conventional high-field systems are bulky and suffer from field inhomogeneities. Zero- to ultra-low-field (ZULF) NMR overcomes these limits by exploiting internal spin interactions in a magnet-free, shielded environment. When combined with nitrogen-vacancy centers in diamond, it enables a compact, portable platform with high spatial resolution and broad bandwidth for noninvasive chemical sensing in microscopic volumes and real-world settings. We report detection of zero- to ultralow-field nuclear magnetic resonance (ZULF NMR) signals at frequencies of a few hertz using a diamond magnetometer. The sensing diamond is a truncated pyramid with 180 μ m height and a 500 2 μ m 2 base. The minimum stand-off distance is <1 mm, and the sensor sensitivity is 13 pT/ $$\sqrt{Hz}$$ <mml:math xmlns:mml=“http://www.w3.org/1998/Math/MathML”> <mml:msqrt> <mml:mrow> <mml:mi>H</mml:mi> <mml:mi>z</mml:mi> </mml:mrow> </mml:msqrt> </mml:math> at frequencies f above 5 Hz with 1/ f -like behavior at lower frequencies. NMR signals were generated via signal amplification by reversible exchange (SABRE) parahydrogen-based hyperpolarization resulting in zero-field signals at 1.7 Hz and 3.4 Hz corresponding to the expected hetero-nuclear J -coupling pattern of acetonitrile. This work demonstrates a magnet-free platform for detecting chemically specific NMR signals paving the way for portable noninvasive diagnostics in microscopic sample volumes for biomedicine, industrial sensing through metal enclosures.
High Concentration Quantum Grade Single Crystal Diamond
View Specs →Products
View Specs →
Single Crystal Cvd Diamond Plate
View Specs →References we cover
Section titled “References we cover”- Scanning nitrogen-vacancy magnetometry down to 350 mK
- Multiplexed Sensing of Magnetic Field and Temperature in Real Time Using a Nitrogen-Vacancy Ensemble in Diamond
- Nuclear Spin Gyroscope based on the Nitrogen Vacancy Center in Diamond
- Microwave-Free Vector Magnetometry with Nitrogen-Vacancy Centers along a Single Axis in Diamond
- Sensitivity optimization for NV-diamond magnetometry