Quadrature-detected zero-to-ultralow-field nuclear magnetic resonance

AI summary

Overview

Research area: Quantum sensing and magnetic resonance — specifically zero-to-ultralow-field nuclear magnetic resonance (ZULF NMR) detected with atomic optically pumped magnetometers.

Technical level: Intermediate. The paper combines a compact analytical spin-dynamics treatment with experimental spectra and density-matrix simulations, so readers should be comfortable with basic NMR concepts (J-coupling, spin-1/2 systems, Fourier transforms), but the core idea is geometrically simple.

Scope (one sentence): The paper introduces and experimentally demonstrates quadrature-detected ZULF NMR, in which two orthogonal dual-axis atomic magnetometers are combined gradiometrically into a complex signal that distinguishes positive from negative resonance frequencies in the laboratory frame, illustrated on thermally polarized [¹³C]-formic acid (an XA spin system) and [¹³C]-methanol (an XA₃ spin system).

What This Paper Is About

In conventional high-field NMR, signals are referenced to a carrier frequency, so only relative frequencies are measured; in ZULF NMR, signals are referenced to zero frequency in the laboratory frame, so absolute frequencies are in principle accessible — yet a single-axis sensor still cannot tell which direction the magnetization is rotating, and most ZULF systems built to date have not been designed to determine the absolute sign of resonance frequencies. The authors solve this by recording two orthogonal components of the oscillating sample magnetization simultaneously and recombining them into a complex (quadrature) signal, which carries the direction of rotation and therefore the sign of the resonance frequency. The goal is to demonstrate this scheme experimentally, validate it against theory and simulation, and show that it also reduces spectral crowding.

Key Contributions

1. Concept and first demonstration of QD ZULF NMR. Two atomic optically pumped magnetometers, each with dual-axis sensitivity, are placed orthogonally in the x–y plane at equal distance from the sample; paired gradiometric signals along x̂ and ŷ are combined into a complex quadrature signal, enabling phase-sensitive detection in the laboratory frame.

2. A theoretical framework for QD ZULF signals. The authors derive closed-form expressions for the single-axis signals Sx and Sy, the gradiometric subtraction, and the resulting complex signal S(t) = e^{iφ}[Sx(t) − i·Sy(t)] = 1.5B·e^{i(ω₀t+φ)−t/T₂}, whose Fourier transform is a complex Lorentzian centered at ω = ω₀, where ω₀ may be positive or negative.

3. Experimental validation on two spin systems. J-spectra of thermally polarized [¹³C]-formic acid (XA) and [¹³C]-methanol (XA₃) at Bz = 230 nT show counter-rotating J-components separated into positive- and negative-frequency space, in agreement with Spinach (MATLAB) simulations that include dipolar relaxation and realistic experimental parameters.

4. Quantified signal and resolution advantages. Quadrature detection gives a factor of 3 signal-amplitude enhancement over single-axis measurement (3/2 from the gradiometric construction, 2 from the quadrature construction), removes the symmetry about zero frequency, decrowds multiplets, and eliminates the phase-distortion artifact where one J-peak appears absorptive while its mirror appears dispersive.

Main Findings

• Absolute frequency sign is recovered. Because ω₀ = −γBz is an absolute frequency in the laboratory frame, the quadrature signal reveals the absolute sign of γ with respect to a known reference; single-channel or gradiometric-only Fourier transforms are symmetric about zero and lose this information.

• Formic acid J-peaks split predictably. For the XA system with J = 222 Hz, the J-peaks split into one positive-frequency peak at J + Σ/2 and one negative-frequency peak at −J + Σ/2, where Σ = (γ_H + γ_C)Bz, with γ_H/2π = 42.6 Hz/μT and γ_C/2π = 10.7 Hz/μT.

• nZF peaks appear on one side only. Near-zero-frequency peaks (from initial magnetization ∝ (Î₁ₓ + Î₂ₓ)) appear only on one side of the quadrature spectrum depending on γ, while J-peaks (from antisymmetric terms ∝ (Î₁ₓ − Î₂ₓ)) split into positive and negative components.

• Excellent spectral quality. For an aqueous ~25.1 M solution of ¹³C-labeled formic acid at Bz = 230 nT, with 4096 averaged scans, the linewidth of the J-peaks around ±230 Hz was ~100 mHz with an SNR of over 3000 (before lineshape broadening). A small feature at ±J_Hz, where J_Hz = 222.4 Hz, is attributed to stray x–y fields.

• Methanol shows richer, still-consistent behavior. For 8 M ¹³CH₃OH in H₂O (1.5 mL) at Bz = 230 nT, with 4096 scans, the 2J-region sextet divides into two groups: peaks with absolute frequency >2J appear only at positive frequencies and those <2J only at negative frequencies; experimental +2J peaks appear at 288.6 Hz, 292.5 Hz, and 296.2 Hz. Linewidths were ~200 mHz for J-peaks around ±140 Hz and ~600 mHz for 2J-peaks around ±280 Hz, with J_Hz = 142.3 Hz.

• Sensor matching is practical but imperfect. Relative gains of the gradiometric channels required adjustment of 5% (formic acid) and 7% (methanol); the authors state that variability in gain and frequency response of commercial atomic magnetometers typically necessitates adjustments of 0–10%. Sensors are held in a single 3D-printed structure to minimize misalignment.

• Shimming becomes simpler. Stray DC fields normally must be painstakingly removed to avoid splitting overlapping J-peaks in non-quadrature zero-field spectra; in the quadrature case the ±J peaks are already resolved by their opposite frequency signs, so this is unnecessary.

Methodology in Plain English

The experiment prepolarizes each sample in a 1 T magnet for 10 s, then mechanically shuttles it 36 cm within 100 ms into the measurement region. A 30 μT DC magnetic-field pulse along the y-axis initializes the desired spin state (200 μs for the formic-acid nZF spectrum and 550 μs for its J-spectra; 150 μs for the methanol nZF spectrum and 550 μs for its J- and 2J-spectra). Spin evolution in the x–y plane is then detected for up to 10 s by two dual-axis atomic magnetometers placed orthogonally at equal distance from the sample, inside a piercing solenoid that applies roughly 230 nT along z (a 10 Hz proton precession frequency) to the sample and not to the sensors. The ~20 s duty cycle is repeated many times (4096 scans) to average down noise such as 50 Hz power-line noise and harmonics.

Each sensor reports two channels, giving Sx1, Sy1, Sx2, Sy2. The transverse channels read half the field of the radial channels because of the dipolar field geometry — hence the factor of 1/2 on Sx1 and Sy2. Gradiometric channels are formed by subtraction (Sx = Sx1 − Sx2, Sy = Sy1 − Sy2) to cancel common-mode noise, and the quadrature signal is built as Sx − i·Sy. Because the sine component changes sign when the frequency changes sign, the resulting complex signal rotates one way or the other depending on the sign of ω₀, and its Fourier transform is a single complex Lorentzian instead of a symmetric pair. Simulations used the Spinach package in MATLAB with realistic parameters, including the z-field profile during shuttling and the y-pulse applied in the presence of Bz.

Why This Matters

Impact on research: QD ZULF NMR restores a piece of information — the direction of spin precession and thus the absolute sign of γ and of scalar couplings J — that had been largely inaccessible in zero-field experiments, and does so while simultaneously simplifying spectra. This aids spin-system assignment, symmetry analysis of spin-state evolution, and interpretation of nutation and relaxation behavior of individual lines. It also turns a practical nuisance (meticulous zero-field shimming) into a non-issue for ±J peak resolution.

Potential applications (as framed by the paper):

• Streamlined assignment of spin systems in small organic molecules without the need for careful stray-field shimming.
• Analysis of more complex molecules and chemical mixtures, extending the direction of in silico ZULF-NMR screening of more than 200 molecules.
• A foundation for hyperpolarized zero-field NMR and multidimensional zero-field methods.
• Studies of relaxation mechanisms of distinct resonance lines and of nutation dynamics under variable pulse conditions.

Industry relevance: The scheme relies on commercially available atomic optically pumped magnetometers and a compact gradiometric geometry built from a single 3D-printed structure, which is compatible with shuttled, repeatable-scan benchtop workflows rather than large superconducting magnets. The authors note that gain and frequency-response variability of commercial sensors (requiring 0–10% gain adjustment) is a practical engineering consideration for anyone building such a system. Hyperpolarization is explicitly stated to be beyond the scope of this work.

Future Directions

• Extending beyond thermal polarization. The authors state that hyperpolarization is beyond the scope of this work because more complicated spin states are created, but note that extension of the theoretical description is possible.
• Toward multidimensional and hyperpolarized zero-field methods, as explicitly proposed in the conclusion.
• Analyzing nutation behavior and relaxation. The paper points to using QD ZULF NMR for nutation analysis under variable pulse conditions and for studying relaxation mechanisms of different resonance lines.
• Tackling complex mixtures and larger molecules, building on the authors' suggestion that quadrature detection will help with systems such as those in the in silico study of more than 200 molecules.

Target Audience

Researchers and students in magnetic resonance, atomic magnetometry, and quantum sensing who already know basic NMR but want an accessible entry point into phase-sensitive zero-field spectroscopy; also method developers working on portable or magnet-free NMR, and chemists interested in spin-system assignment and spectral simplification without high-field instrumentation.

Authors’ abstract

We present the concept of quadrature-detected zero-to-ultralow-field nuclear magnetic resonance (QD ZULF NMR). Two orthogonal sensors, atomic optically pumped magnetometers, are used to separately detect components of oscillating sample magnetization, which are then combined into a quadrature signal. This approach enables the distinction of positive and negative resonance frequencies with respect to the laboratory frame---such that absolute signs of gyromagnetic ratios can be determined. In addition, QD ZULF NMR reduces peak density and overlap compared to ZULF NMR with single-axis detection, while providing richer information about nuclear-spin dynamics. As an illustration, we present experimental $J$-spectra for thermally polarized [$^{13}$C]-formic acid and [$^{13}$C]-methanol, compounds commonly used in ZULF NMR which represent XA and XA$_3$ spin systems, respectively. A theoretical framework describing QD ZULF NMR signals is provided.