DFT-assisted natural abundance 13C zero-field NMR via optical magnetometry

This paper demonstrates a breakthrough in natural-abundance 13C zero-field NMR spectroscopy by combining a compact optical magnetometer with vibrationally corrected DFT predictions to achieve high-sensitivity, isotopomer-resolved molecular identification and the extraction of transient solution-state structural information without requiring hyperpolarization or large magnetic fields.

The Gist

Imagine you are trying to identify a specific person in a crowded room. Usually, you need a giant, expensive spotlight (a high-field magnet) to make them stand out clearly. But what if you could identify them just by listening to the unique rhythm of their heartbeat, even in a dark, noisy room?

This paper describes a breakthrough in doing exactly that for molecules, using a technique called Zero-Field NMR.

Here is the story of what the researchers achieved, broken down into simple concepts:

1. The Problem: The "Giant Magnet" Bottleneck

Traditional chemical analysis (NMR) is like a high-end concert hall. It requires massive, super-cooled magnets to create a strong, uniform magnetic field. This makes the equipment huge, incredibly expensive, and sensitive to any interference (like a metal door or a power line). Because of this, it's hard to use for quick, everyday chemical checks, especially for liquids that conduct electricity or are in metal containers.

2. The Solution: Listening in the Dark (Zero-Field)

The researchers turned the lights off. Instead of using a giant magnet, they turned the magnetic field down to almost zero.

• The Analogy: Imagine a choir. In a normal room (high field), everyone sings slightly different notes based on where they stand (chemical shifts). In a silent room (zero field), the only thing you hear is how the singers are holding hands and tapping each other's shoulders (scalar J-couplings).
• The Benefit: Without the giant magnet, you can put your sample inside a metal box or a conductive container, and the "music" still comes through clearly. It's cheap, portable, and flexible.

3. The Challenge: The "Static" Noise

There were two big problems with this "silent room" approach:

1. It was too quiet: Natural molecules have very weak signals. Usually, you needed to "supercharge" the molecules (hyperpolarization) or use expensive, rare versions of them (isotopically enriched) to hear anything.
2. It was too messy: Without the big magnet, the "music" becomes a chaotic jumble of thousands of overlapping notes. It's like hearing a thousand people talking at once; without a map, you can't tell who is saying what.

4. The Breakthrough: The "Super-Listener" and the "Translator"

The team solved both problems by combining a better microphone with a smart computer translator.

A. The Super-Listener (Hardware)
They used a compact, commercial device called an Optically Pumped Magnetometer (OPM). Think of this as a super-sensitive ear that can hear the faintest whisper of a molecule.

• They improved the setup so the machine could listen for over a week without getting tired or drifting out of tune.
• This stability allowed them to listen to natural abundance samples. They didn't need rare, expensive chemicals. They just grabbed bottles of common liquids (like benzaldehyde or formic acid) off the shelf, put them in a vial, and listened.
• The Result: They could even hear the "whispers" of extremely rare molecules (doubly labeled carbon atoms) that appear only once in every 8,000 molecules.

B. The Translator (Software/DFT)
To make sense of the chaotic jumble of notes, they used Density Functional Theory (DFT).

• The Analogy: Imagine you have a sheet of music for a complex song, but you don't know what the song sounds like. The computer acts as a translator that predicts exactly what the song should sound like based on the molecule's shape.
• They added a special "vibration correction" to the computer model. Molecules aren't frozen statues; they wiggle and vibrate. The computer now accounts for this wiggling, making its predictions incredibly accurate (within a few Hertz).

5. The Magic: When the Prediction is Wrong, It Tells a Story

Usually, if a computer prediction doesn't match the experiment, you think the computer is wrong. But here, the researchers found that the difference between the prediction and the real measurement is actually useful information.

• The Metaphor: Imagine you predict how a rubber band stretches in a vacuum. Then you stretch it in water. If it stretches differently, that difference tells you about the water's resistance.
• The Application: By comparing their "vacuum prediction" with the "real-world measurement," they could see how molecules were interacting with their environment.
  • They could detect hydrogen bonding (molecules holding hands).
  • They could see hydration (molecules surrounded by water).
  • They could spot ion pairing (charged molecules sticking together in salty solutions).

Summary

This paper demonstrates a new way to identify chemicals without needing giant, expensive magnets.

1. They built a stable, portable "ear" that can hear natural chemicals clearly.
2. They built a "translator" (computer model) that predicts what those chemicals should sound like.
3. By comparing the two, they can not only identify the molecule but also see how it is behaving in a solution (like how it interacts with water or salt).

This turns Zero-Field NMR from a niche physics experiment into a practical tool that could one day be used to analyze chemicals in metal containers, salty solutions, or even in the field, without needing a massive laboratory.

Technical Summary

1. Problem Statement

Zero-field (ZF) Nuclear Magnetic Resonance (NMR) spectroscopy offers a unique alternative to conventional high-field NMR by detecting scalar $J$-couplings instead of chemical shifts. This approach allows for:

• Operation without large, expensive superconducting magnets.
• Detection through conductive or metallic enclosures.
• High sensitivity to molecular topology and connectivity.

However, the widespread adoption of ZF NMR for chemical analysis has been hindered by two primary challenges:

1. Sensitivity: Conventional ZF NMR relies on optically pumped magnetometers (OPMs), which are less sensitive than high-field inductive detectors. Consequently, previous studies required isotopically enriched samples (e.g., $>99\%$ $^{13}$C) and extensive sample preparation (e.g., oxygen removal via freeze-pump-thaw), making it impractical for routine analysis of natural abundance (1.1% $^{13}$C) samples.
2. Spectral Interpretation: ZF spectra consist of dense, complex multiplets arising from untruncated scalar couplings. Unlike the simpler patterns in high-field NMR, these spectra are difficult to interpret without computational support. Furthermore, predicting these spectra ab initio is challenging because small errors in calculated coupling constants ($J$) lead to large discrepancies in the predicted multiplet patterns.

2. Methodology

The authors addressed these challenges by integrating advanced instrumental engineering with high-accuracy quantum chemical calculations.

A. Instrumental Advances

• Apparatus: The system utilizes a compact, commercial $^{87}$Rb optically pumped magnetometer (QuSpin) housed in a mu-metal shield.
• Polarization: Samples are pre-polarized in an inexpensive, inhomogeneous 9.4 T magnet, then shuttled (~1 s) into the zero-field region.
• Stability: The system achieves exceptional long-term stability (weeks), allowing for continuous signal averaging over many days without spectral degradation.
• Sample Handling: Experiments were conducted on "off-the-shelf" liquids at natural abundance without oxygen degassing or special conditioning.

B. Computational Framework (DFT)

• Prediction: The authors developed a workflow using Density Functional Theory (DFT) with the PBE0 functional to predict scalar couplings.
• Vibrational Corrections: Crucially, they incorporated vibrationally corrected calculations (using second-order vibrational perturbation theory, VPT2). This accounts for the fact that measured couplings are rovibrationally averaged, significantly reducing systematic errors.
• Spin Dynamics: The predicted coupling networks were propagated through spin-dynamics simulations (using the Spinach library) to generate full ZF spectra for direct comparison with experiment.

C. Experimental Workflow

• A library of 13 diverse molecules (including benzaldehyde, benzene, acetone, and formic acid) was measured.
• Specific attention was paid to detecting rare isotopomers (e.g., doubly $^{13}$C-labeled species) and analyzing solution-state effects (hydration, ion pairing) by comparing experimental shifts against isolated-molecule DFT predictions.

3. Key Contributions

1. First Natural Abundance ZF NMR Library: The study demonstrates, for the first time, natural abundance ($1.1\%$) $^{13}$C ZF spectroscopy on a library of commercial liquids without hyperpolarization or isotopic enrichment.
2. Detection of Rare Isotopomers: Leveraging week-long stability, the team detected the rare (0.0121%) doubly $^{13}$C-labeled species in acetic acid, a feat previously thought impossible without enrichment.
3. High-Accuracy Spectral Prediction: The authors demonstrated that vibrationally corrected DFT can predict complete ZF multiplets with few-Hz accuracy, enabling the construction of in silico reference libraries for molecular identification.
4. Chemical Insight from Residuals: The paper establishes that deviations between experiment and vacuum-state DFT are not merely errors but are chemically informative. These residuals report on specific non-covalent interactions, such as hydrogen bonding, hydration numbers, and contact ion pairing.

4. Key Results

• Sensitivity and Resolution: The system achieved linewidths of $\le$ 250 mHz and demonstrated a signal-to-noise ratio (SNR) scaling linearly with $\sqrt{N_{scans}}$ over >10 days. This represents a ~3-fold sensitivity gain over previous benchmarks, enabling the detection of 100 mM formic acid at natural abundance.
• Isotopomer Resolution: The ZF spectra successfully resolved distinct isotopomers based on their local spin topology (e.g., $XA$, $XA_2$, $XA_3$ motifs corresponding to CH, CH$_2$, and CH$_3$ groups), providing a "fingerprint" richer than standard DEPT experiments.
• Solvent and Ion Effects:
  • Benzene: Solvent changes (C$_6$D$_6$ vs. cyclohexane) induced measurable line shifts (100–500 mHz), demonstrating the technique's sensitivity to the chemical environment.
  • Formic Acid/Formate: The study quantified massive effective chemical shift changes (up to $10^5$ ppm) in the $^1J_{CH}$ coupling upon dilution and deprotonation.
  • Hydration & Ion Pairing: By matching experimental $J$-shifts to DFT models, the authors determined the average hydration number of formate to be ~5 and resolved counterion-specific shifts (Li$^+$, Na$^+$, K$^+$) in concentrated (5 M) solutions, confirming contact ion pairing.
• Water Suppression: The method intrinsically suppresses water signals (as H$_2$O lacks a heteronuclear pair), allowing high-resolution analysis of aqueous and salt-rich samples that are difficult to study with high-field NMR.

5. Significance

This work transforms ZF NMR from a niche fundamental physics tool into a general platform for chemical analysis.

• Accessibility: By removing the need for superconducting magnets and isotopic enrichment, it lowers the barrier to entry for NMR spectroscopy, enabling portable and field-deployable molecular identification.
• Structural Biology & Drug Discovery: The ability to extract transient solution-state structures (via $J$-coupling observables) and detect non-covalent interactions (hydrogen bonding, ion pairing) suggests applications in studying protein-ligand binding and small-molecule drug interactions.
• Complex Media: The technique's immunity to magnetic susceptibility broadening makes it uniquely suited for analyzing ionic liquids, concentrated electrolytes, and samples inside conductive containers.
• Future Outlook: The combination of DFT-assisted prediction and high-stability OPMs paves the way for real-time, in situ monitoring of industrial chemical processes and the analysis of complex natural products where stereochemical assignment is challenging.