Quantum sensors in diamonds for magnetic resonance spectroscopy: Current applications and future prospects Open Access

Whether it is a gas, a liquid, a solid, a powder, or a single crystal, at room or cryogenic temperatures, a material composed of spin-bearing particles can be a potential sample for magnetic resonance (MR) spectroscopy.^1–3 In the case of nuclear spins, a look at the periodic table of elements will tell us that almost all elements have spin-carrying isotopes that can be detected. This highlights the countless scenarios where nuclear magnetic resonance (NMR) can play a critical role, utilizing its multitude of methods and variants, including cellular and molecular structural biology,^4,5 drug discovery,⁶ materials science and organic synthesis,^7,8 fundamental physics,⁹ energy conversion research,¹⁰ food processing,¹¹ agricultural and environmental sciences,¹² or quantum information research.^13–15 Although unpaired electrons are not as ubiquitous as spin-carrying nuclei in molecules, the application of electron spin resonance (ESR, or EPR, electron paramagnetic resonance) is multifaceted: it provides unique information on structure, reactivity, and dynamics of paramagnetic systems.^16–20 Traditionally applied to solid-state physics and chemistry, the combination with spin-labeling techniques has made it one of the prominent biophysical methods for the study of biological structures.²¹ Furthermore, ESR has lately become a novel tool in the research on novel 2D materials,^22,23 quantum technologies,²⁴ and in the study of unconventional magnetic properties and phenomena.^25,26 However, especially for NMR, a well-known drawback is the low sensitivity that is primarily caused by the poor spin polarization under typical experimental conditions. This fundamental issue has been recognized since the early days of MR spectroscopy and often restricts the detection of MR signals to macroscopic sample volumes.²⁷ 

Recent discoveries in the field of solid-state physics have led to the development of atomic-sized magnetic field sensors, which seem to be ushering in a new era of MR spectroscopy on nano- and microscale sample volumes. The most established sensor is a spin defect in diamond called a nitrogen-vacancy (NV) center that can be introduced with various densities and at different depths within the diamond crystal and exhibits a distinct combination of spin and optical properties. Under green light illumination, the NV spin state can be first initialized, then coherently manipulated by microwave (MW) irradiation, and finally optically read out through its spin-state-dependent red fluorescence intensity.²⁸ Importantly, all of these processes can be conducted under ambient conditions. These properties underpin its capability as an ultra-sensitive atomic-scale magnetometer,^29,30 enabling the detection of even the minuscule magnetic fields generated by a nearby single nuclear spin.^31,32

We consider the NV center in diamond as an extremely powerful MR sensor for two main reasons:

1. Proximity—the NV centers can be placed a few nanometers away from the sample, resulting in a strong magnetic interaction between the sensor and the sample.

2. Optical readout—the NV quantum state, encoding the information about the sensed interaction, is detected optically, which allows for single NV-detection and (wide-field) imaging of NV-ensembles.

These distinctive properties allow for MR spectroscopy with high sensitivity and, crucially, spatial resolution.

Several excellent review papers on NV quantum sensing have been published in recent years.^33–40 Recently, Allert et al. provided an extensive review focusing on NV-NMR spectroscopy, in particular selecting topics that could be of particular interest for the chemistry community.⁴¹ In this article, we will provide a perspective on the main scientific fields where we anticipate NV-NMR and NV-ESR to have a significant impact in the future. On the one hand, we see the most promising direction in the life sciences with applications in single molecule and cell studies, lab-on-a-chip (LOC) applications, and radical or ion sensing. On the other hand, we foresee a vast employment of these techniques in the material sciences, including surface and interfaces, catalysis, 2D materials, materials under extreme conditions, and quantum technologies. To provide context for the many applications of MR spectroscopy using NV sensors in diamond, we will first give a brief overview of the fundamental principles of MR spectroscopy and provide an introduction to NV centers in diamond.

When an electron or proton spin (⁠ $S , I = 1 / 2$⁠) is exposed to a magnetic field B₀, its spin states $| m S , I = + 1 / 2 ⟩$ and $| m S , I = − 1 / 2 ⟩$ are split by the energy $Δ E = | γ | ℏ B 0$⁠, where γ is the gyromagnetic ratio. This splitting corresponds to the precession frequency of the spin's magnetic moment in the B₀ field (Larmor frequency $ω L = | γ | B 0$⁠), which can be probed spectroscopically. Although ESR and NMR spectroscopy share the same principles, the involved energies, time scales, and Larmor frequencies (GHz vs MHz) greatly distinguish the two techniques both from the methodological and the instrumental point of view. One key advantage of NMR spectroscopy is its high-spectral resolution (sub-Hz or ppm linewidths), allowing for the detection of weak interactions such as spin–spin couplings or chemical shifts that provide information about the nuclear local environment and molecular connectivity at atomic resolution. For this reason, NMR spectroscopy is one of the key methods for molecular structural analysis, ranging from small molecules to complex structures, such as proteins. ESR spectroscopy offers distinct and often complementary information in comparison with NMR spectroscopy. It can uniquely characterize paramagnetic systems and their molecular environments, determining parameters such as g-factors, hyperfine coupling constants, spin–spin couplings, or zero-field splitting constants. In addition, ESR can measure distances and relative orientations between two or more electron spin labels by exploiting the long-range dipolar coupling (up to $∼ 10 nm$⁠). This allows for insight into the structure and dynamics of large biomolecular systems.⁵² 

An NV center is a point defect consisting of a nitrogen atom and an adjacent vacancy that replace two carbon atoms in the tetrahedral crystal structure of diamond. The negatively charged NV center constitutes the actual sensor and has an electronic structure, where six electrons form a spin triplet ground state (S = 1) consisting of the $| m s = 0 ⟩$ (⁠ $| 0 ⟩$⁠) and $| m s = ± 1 ⟩$ (⁠ $| ± 1 ⟩$⁠) states.^53–55 The excitation of the NV ground state by green light results in a red spin-state-dependent fluorescence intensity, which can be utilized for discriminating the $| 0 ⟩$ and $| ± 1 ⟩$ states. Furthermore, the optical pumping of the electronic transitions and the unique spin relaxation properties lead to almost full polarization of the system into the $| 0 ⟩$ state. The spin states can be coherently manipulated with microwave pulses with high fidelity, even under ambient conditions.^56,57

NV-based MR spectroscopy can be categorized based on two key factors: (i) whether the objective is to detect nuclear or electron spins, resulting in NV-NMR and NV-ESR, respectively, and (ii) the scale of the sensing volume being targeted, which can range from microscale or picoliters (pL) to nanoscale or zeptoliters (zL). As the sensing length scale is defined by the NV-target distance, in the former case, the typical strategy involves the use of thick (⁠ $∼ μ$m) layers of dense NV ensembles. In the latter case, shallow NV centers (either single or in ensembles) are usually employed, and these are implanted a few nanometers beneath the diamond surface. The NV sensor detects magnetic fields primarily from spins located within a hemisphere whose radius roughly corresponds to the NV center's depth.^42,58,59 In a typical nanoscale NV-NMR experiment, the signal is detected from a few spins to a few thousand spins on top of the diamond. In this regime, statistical polarization (SP) or spin noise is much stronger than thermal or Boltzmann polarization (BP).^41,60–63 SP arises from the incomplete cancellation of spin magnetic moments leading to magnetic noise at the Larmor frequency with stochastically varying phases and amplitudes. In contrast to BP (BP $∝ B 0 / T$⁠), SP does not require RF or MW excitation for spin detection and it does not directly scale with the applied magnetic field but with the number of spins N according to $1 N$⁠. SP reaches polarization levels of a few percent within nanoscale sample volumes.^63,64 In contrast, if the number of detected spins is largely increased by using μm-thick NV-doped layers (“microscale NV experiments”), the signal is dominated by the BP, and the experiments resemble conventional MR: the MR signal is proportional to the external magnetic field B₀, and its detection requires RF excitation of the sample.

An NV-MR experiment typically starts with the NV spin state being in a superposition state that is exposed for a specific duration to the magnetic signals arising from electron or nuclear sample spins. This will lead to the accumulation of a relative phase during the evolution of the superposition, which encodes the relevant spectroscopic information. Finally, the phase information is transposed to a spin-population difference, resulting in a detectable difference in fluorescence intensity. Often, NV-MR experiments rely on dynamical decoupling schemes, which form a narrow-band magnetic field detector around the MR frequency to be sensed.³³ 

At the nanoscale, the NMR signals exhibit random amplitude and phase, due to the stochastic nature of spin noise occurring at the Larmor frequency. NV-NMR experiments typically rely on dynamical decoupling pulsed schemes that are designed to detect the spin noise by measuring the variance rather than the magnitude of the nuclear magnetic fields.^{33,41,42,47,59} These methods offer great sensitivity; however, they have limitation in terms of spectral resolution, as discussed in Sec. III A. At the microscale, where BP dominates, the NMR signals can be synchronized to the dynamical decoupling sequences [coherently averaged synchronized readout (CASR)], enabling high spectral resolution.⁴⁷ According to the same principles, a sub-variant of NV-NMR is NV-based nuclear quadrupole resonance (NV-NQR) spectroscopy that relies on the detection of quadrupolar nuclear species (I > 1∕2). Here, the interaction of the electric field gradients generated by the non-spherical charge distribution of the molecular bonds with the nuclear quadrupolar moment produces an observable energy splitting of the nuclear spin states. This results in a MR technique that does not require the application of an external magnetic field and, importantly, in a spectroscopic method with very high chemical specificity in solids.^65–68 

NV-ESR experiments have been typically performed on the nanoscale, where statistical polarization dominates. This is accomplished mainly using two approaches: (i) double electron–electron resonance (DEER). The key idea is to detect a change in the magnetic field at the NV position (due to dipolar coupling) upon flipping the sample electron spin with a microwave pulse. By sweeping the frequency of the MW pulse, the ESR spectrum of the paramagnetic species can be recorded. (ii) NV-relaxometry. In contrast to the general detection strategies discussed previously, NV-relaxometry relies on monitoring the impact of paramagnetic species in the sample on the shortening of the NV spin lifetime T₁. This change can be accurately tracked, and, by controlling the external magnetic field, the NV system can be tuned to resonate with the transitions of the target electronic spin system, enabling the complete reconstruction of the sample's EPR spectrum.

In the following sections, we will review recent advances and discuss the potential applications of NV-based MR spectroscopy in highly promising fields, ranging from life sciences to materials sciences.

The ultimate goal of nanoscale NV-NMR is the structural analysis of individual molecules. Although this technique is still in its infancy, key milestones have been reached. Two groundbreaking studies by Mamin et al.⁵⁹ and Staudacher et al.⁴² marked the birth of nanoscale NV-NMR (see Fig. 1). The authors demonstrated the detection of NMR signals from nanoscale samples under ambient conditions, using a single NV center located just a few nanometers below the diamond surface. Subsequent investigations have confirmed the potential of NV centers to reach the ultimate sensitivity of a single proton spin.^31,32,69 Magnetic resonance signals from a single protein were successfully detected by Lovchinsky et al.⁴⁴ using a single NV spin defect. By utilizing quantum logic to improve the readout fidelity and surface chemistry to increase the coherence of near-surface NV centers, the authors achieved enough sensitivity to detect nuclear spins from a single immobilized Ubiquitin protein on the diamond surface [Fig. 2(a)]. These impressive results have established NV-NMR as a promising platform to probe the structure and dynamics of biological and biochemical systems not only at the nanoscale but even down to the level of a single molecule. However, this exceptional result comes with the main drawback—the lack of chemical resolution. Aslam and co-workers⁴⁵ addressed this issue, albeit not at the single-molecule level. They performed their experiments at much higher magnetic fields (3 T) [Fig. 2(b)], to improve the chemical resolution and used the ancillary ¹⁵N nucleus as a quantum memory to increase the spectral resolution of the NV-sensor. This allowed them to detect ¹H and ¹⁹F NMR signals from zeptoliters sample volumes with chemical shift resolution. Despite this improvement, achieving high spectral resolution at the nanoscale for both solid and liquid-state samples remains a challenge. In the solid state, the spectral resolution is mainly limited by dipolar broadening. Aslam et al. mitigated this effect by applying decoupling protocols, which decreased the resonance linewidth from $26 kHz$ (206 ppm) down to $1.5 kHz$ (12 ppm) [see Fig. 2(c)]. In this case, the linewidth is still limited by the finite pulse length in the homonuclear decoupling scheme. Higher RF power through optimized RF delivery could improve these results even further. High spectral resolution in NMR spectroscopy is typically achieved in liquid samples, because dipolar interactions are averaged out due to molecular motions. However, nanoscale NV-NMR experiments detecting statistical polarization face another issue here: molecular diffusion in and out of the nanoscale sensing volume limits the time the NV center can interact with the nuclear spin. This results in dramatic line broadening and restricts the detection to viscous samples.^45,70,71 By increasing the NV-sample distance, Aslam et al. could exploit larger detection volumes and consequently longer diffusion times, being able to show chemical shift resolution of viscous samples [see Fig. 2(d)]. To overcome diffusional line broadening and improve the interaction time between the NV center and the nuclear spins, immobilizing the sample molecules through confinement structures, such as nanoporous materials, liposomes, micelles, and polymer shells, has been proposed.⁷⁰ Recently, Liu et al. demonstrated this approach by functionalizing a NV-diamond chip with a thin layer of a porous material based on a metal-organic framework (MOF) architecture.⁷² By effectively confining the molecules in its angstrom-sized-pores, the MOF layer is able to strongly reduce molecular diffusion, greatly extending the sensor–sample interaction time. Using ensembles of NV centers, the authors showed that, in the absence of the MOF layer, a liquid sample cannot be detected; however, with the MOF caging structure, a clear NV-NMR signal is unambiguously detected. A great advantage of using MOFs is that their reticular synthesis allows for great versatility in terms of tailored pore size and physico-chemical properties.^73,74 This method appears to be broadly applicable for the detection of a variety of small molecules and desired analytes. Moreover, a careful line shape analysis can also improve the spectral resolution, which has been demonstrated recently.^70,71,75 The combination of these approaches with colocalization of the pores with single NV centers is highly promising and could pave the way for structural analysis of individual molecules.⁷² 

In the solid state, dipolar broadening can be tackled, for instance, by more effective decoupling experiments or by the implementation of magic angle spinning. However, the latter is a non-trivial engineering task. Alternatively, NV-NQR at the single-molecule level holds promise, and, with the already available performance in terms of sensitivity and spectral resolution, it could provide structural information and be used to investigate chemical environments within individual molecules. Although proposed,⁴⁴ an experimental demonstration for a single molecule level of this highly promising approach is still lacking. However, NV-NQR has been successfully demonstrated for 2D materials,⁴⁶ which will be discussed later in this perspective.

Similarly, NV-ESR has been demonstrated to be a very promising tool for the analysis of single (bio)molecules. A significant breakthrough in this direction was achieved by Shi et al.,⁷⁶ where they immobilized a mitotic arrest deficient-2 (MAD2) protein spin-labeled with a nitroxide radical onto a diamond surface [see Figs. 3(a) and 3(b)]. The authors were able to not only detect the ESR signal from a single spin-labeled protein under ambient conditions, but even achieved to extract some structural and dynamical information. Subsequently, Shi et al. demonstrated the detection of NV-ESR signals from individually spin-labeled DNA molecules⁷⁷ [see Figs. 3(c) and 3(d)]. These results were crucial in demonstrating the feasibility of single-molecule NV-ESR. This technique holds immense potential for probing the structural characteristics of individual biomolecules in their native-like state.

As discussed in Sec. II, NV-layers that are μm-thick enable the detection of picoliter sample volumes where thermal polarization exceeds statistical polarization. This results in a coherent NMR signal that is not affected by molecular diffusion and is comparable to conventional NMR spectroscopy. This microscale NV-NMR has the potential to establish a novel platform for lab-on-a-chip (LOC) applications, which could greatly enhance the study of biological and chemical systems.⁷⁹ The first pioneering study by Glenn et al.⁴⁷ demonstrated the detection of NMR signals with Hertz spectral resolution from a sample volume of around 10 picoliters. The authors developed a pulse sequence called coherently averaged synchronized readout (CASR), which consists of dynamical decoupling sub-sequences that are synchronized to the free nuclear decay induced by a $π / 2$-pulse applied to the nuclear sample spins. With this approach, chemical shift and J-coupling of organic compounds could be resolved [see Fig. 4(a)]. Although thermal polarization and chemical information would increase by working at higher magnetic fields, so far the microscale NV-NMR experiments have been performed at low magnetic fields (below 0.2 T).^{41,47,48,78,80} This is because current quantum sensing protocols work most efficiently for NMR signals at a few megahertz. Higher magnetic fields would result in higher resonance frequencies of both sample and NV spins, strongly increasing the complexity of the required microwave equipment. Overcoming this engineering challenge and performing microscale NV-NMR at magnetic fields exceeding 1 Tesla is critical for obtaining chemical information necessary for real world applications. Novel pulse sequences may also mitigate the MW requirements and enable microscale NV-NMR even at higher fields.^81,82 We would also like to point out that in contrast to nanoscale NV-NMR, where the statistical polarization determines a weak angular and strong dependence of the signal on the NV centers depth, microscale NV-NMR signals are significantly influenced by the sample geometry, as investigated in detail by the work of Bruckmaier et al.⁵⁸ In 2019, Smits et al. made the first steps toward integrating NV-NMR into microfluidics,⁴⁸ a crucial development for lab-on-a-chip applications. They employed a tape-based straight channel that allowed in-line measurements of one- and two-dimensional NMR spectra of liquid samples. More recently, Allert et al. demonstrated several NV center-based MR experiments utilizing a novel, fully integrated microfluidic quantum sensing platform⁵⁰ (see Fig. 5). The biocompatible platform enabled microscopic sample control within a broad range of microfluidic channel geometries while maintaining complete quantum sensing capabilities. The combination of microfluidic sample manipulation and sensor miniaturization enables many complex experiments, such as organs-on-a-chip,^83,84 single-cell,⁸⁵ or electrochemical studies.⁸⁶ Chemical analysis in microfluidics is often restricted to optical spectroscopy methods, which provide only limited information.⁸⁴ NMR spectroscopy has the potential to enable noninvasive, quantitative, and label-free analysis,⁸⁷ but conventional NMR suffers from poor sensitivity. However, the small sensing volumes (10–500 pl) of microscale NV-NMR correspond to the length scales of most microfluidic operations, offering the potential for high-throughput analysis in chemistry and biotechnology, as well as NMR-based droplet sorting.

The long-standing goal of probing single cells with MR is within reach with NV centers. While NV-NMR has demonstrated the ability to detect resolved chemical shifts and scalar couplings in picoliter sample volumes (corresponding to single cell volumes), its poor concentration sensitivity precludes the detection of biologically relevant molecules, such as metabolites. However, hyperpolarization methods, commonly used in conventional NMR spectroscopy, can significantly increase the sensitivity.^88–90 In a recent study by Bucher et al., Overhauser Dynamic Nuclear Polarization (DNP) was shown to be compatible with NV-NMR detection, providing signal enhancements of up to two orders of magnitude for ¹H-NMR.⁸⁰ In this case, the electronic spin polarization of organic radicals added to the sample is transferred to the nuclear spins via microwave (MW) irradiation. With this approach, femtomole sensitivity in a 10 pl volume (which corresponds to ∼5 mM) was demonstrated for a broad range of small molecules. Nevertheless, Overhauser DNP has some drawbacks to overcome. For instance, the enhancement factors vary significantly between molecules⁸⁰ and even between nuclei within the same molecule,^91,92 complicating quantitative NMR studies and making predictions difficult, which hinders broad applicability. In addition, technical problems, such as undesired heating effects due to long MW irradiation or the toxicity of the added radicals, may discourage the application of Overhauser-DNP methods for single-cell studies. Much higher signal enhancements can be accomplished by means of (parahydrogen induced polarization) PHIP methods, where the nuclear singlet state of parahydrogen is used as a polarization source.^93–96 Arunkumar et al.⁷⁸ combined this technique with microscale NV-NMR and achieved 0.5% proton spin polarization at 6.6 mT, surpassing the thermal polarization by approximately five orders of magnitude. This method enabled detection of concentrations as low as one millimolar with a signal-to-noise ratio of 50 after 300 s of averaging time within 10 pl sample volumes [see Fig. 4(b)]. Further improvements are needed to detect sub-millimolar concentrations, which would pave the way for single cell metabolomics. A detailed perspective on single cell NV-NMR is given in the work of Neuling et al.⁹⁷ Recently, Bruckmaier and colleagues⁵¹ demonstrated microscale molecular diffusion measurements using NV-NMR spectroscopy. They combined pulsed magnetic field gradients with microscale NV-NMR detection, which allowed them to image local water diffusion within microstructures. This technique paves the way for probing water diffusion within tissues with single cell resolution. We envision microscale NV-NMR as a powerful tool for single cell biology, enabling a vast scenario of applications in fundamental biological and applied medical research.

As introduced in Sec. II, a highly promising approach to detect electron spins (NV-ESR), particularly for biophysical applications, relies on monitoring the NV longitudinal relaxation time T₁ to sense magnetic ions, radicals, or other paramagnetic species. Steinert et al.⁹⁸ developed an NV-based relaxometry technique that was able to detect magnetically labeled cellular structures under ambient conditions (see Fig. 6). They used a microfluidic chip to optimize the sample volumes and a CCD camera for widefield fluorescence detection. The method demonstrated a spatial resolution below 500 nm with a sensitivity of around 1000 statistically polarized spins, making it promising for subcellular magnetic imaging. Ziem et al.⁹⁹ also used a wide-field relaxation-based approach to detect paramagnetic systems of biological interest under physiological conditions, such as manganese (II) ions and ferritin proteins adsorbed on the diamond surface. Hall et al.¹⁰⁰ proposed a relaxometry-based approach to detect the spectrum of paramagnetic impurities in diamond, such as nitrogen electron-donor defects (P1 centers). By controlling the external magnetic field, the NV system could be adjusted into resonance with the target electronic system's transitions, allowing full reconstruction of the ESR spectrum. Building upon this approach, Simpson et al.¹⁰¹ developed what they called a quantum probe relaxation imaging method working under biologically compatible conditions and able to provide diffraction-limited images of paramagnetic species (e.g., hexaaqua-Cu²⁺ complexes in solution). This technique offers a spin sensitivity of 100 zmol and can monitor the redox reaction kinetics of Cu²⁺ species in the presence of a reducing agent.

Furthermore, nanodiamonds containing NV centers play an important role for studying biological systems. Such systems can work as magnetic resonance nanosensors that can be internalized by living biological cells, and their uptake can even be guided to subcellular targets upon proper nanodiamond functionalization. Such fluorescent nanoprobes can be used, for instance, to monitor the production of free radicals by the cellular metabolism with subcellular resolution.^102–104 

Finally, a recent work demonstrated that diamagnetic ions can also be detected by monitoring the NV-T₁ relaxation time using relaxometry experiments.¹⁰⁵ Specifically, the presence of diamagnetic ions (e.g., Na⁺, Li⁺, and K⁺) at millimolar concentrations in aqueous solutions was found to perturb the electric field distribution at the solid–liquid interface, resulting in a significant extension of the T₁ time of near-surface NV-ensembles. This observation could expand the application of NV-sensors to electrolyte sensing in cell biology and neuroscience or to probe electrochemical interfaces. Overall, the relaxometry-based techniques described above show great promise, offering high sensitivity and spatial resolution while being relatively straightforward to implement (without requiring complex microwave manipulations). Beyond biophysical applications, we anticipate that these methods will find broad use in areas such as electrochemistry and catalysis for investigating nanoscale chemical reactions or electron transfer processes.¹⁰⁶ 

Although MR techniques are rather ubiquitous in all fields of chemical research, one of the few areas in which NMR has never played a major role is surface science. This is due to the low sensitivity of conventional NMR spectroscopy compared to other available surface spectroscopy techniques, which has made it difficult to detect the small number of spins located at surfaces or interfaces. Using quantum sensors in diamond, Liu and coworkers have developed a general platform to perform surface-sensitive NMR spectroscopy at ambient conditions with sub-monolayer sensitivity and microscopic spatial resolution.^49,107 To reach this goal, the authors used near surface NV ensembles, which were implanted with an average depth of less than 5 nm [see Fig. 7(a)]. Their diamond sensors comprised a large number of NV centers (⁠ $∼ 10 5$⁠), providing a significant sensitivity advantage and higher robustness compared to single NV experiments. In their proof of concept study, they detected ¹⁹F and ³¹P nuclei from a self-assembled monolayer (SAM) anchored by phosphonate chemistry on a 1 nm-thick Al₂O₃ substrate deposited on the diamond surface [as depicted in Fig. 7(b)]. The NV-NMR measurements provided a calibration-free approach for determining the molecular coverage of the SAM layer. The optical readout was used to locally probe the diamond surface by moving the laser spot across the diamond. Alternatively, the scanning approach can be replaced by using a camera to perform surface NV-NMR imaging on a widefield.⁴³ Additionally, the authors demonstrated the real-time anchoring reaction of the molecules on Al₂O₃ and measured the reaction kinetics by detecting the growth of the ¹⁹F signal size during monolayer formation [Fig. 7(c)]. As highlighted in the previous chapter on single-molecule studies, also in the case of surface NV-NMR, one major limitation is the low chemical information that it provides currently, owing to broad resonance lines with a width of a few kilohertz. However, this issue can potentially be addressed in the future by using effective decoupling pulsed protocols.⁴⁵ Furthermore, the detection of NQR nuclei^46,108 could already be a task that can be well performed by this method and provide important chemical information. In the future, we see this technique as a valuable alternative to commonly used surface-sensitive methods such as x-ray photoelectron spectroscopy (XPS), secondary ion mass spectrometry (SIMS), and SFG/SHG spectroscopy.⁴⁹ While these methods can perform chemical analysis, they often require demanding experimental conditions, including ultra-high vacuum. By contrast, we envision NV-NMR as an ambient-condition surface-sensitive technique with microscale spatial resolution, which will open up probing surfaces, interfaces, and thin films in areas such as bioanalytics, catalysis, and energy conversion research.

In the field of solid-state physics and materials science, the discovery of new properties and physical phenomena that can occur when materials are organized as atomically thin layers has led to an exponential growth of research groups dedicated to investigating two-dimensional (2D) materials and their heterostructures.^109–111 Structural characterization of low-dimensional materials requires highly sensitive methods that conventional NMR has not yet been achievable. In this regard, Lovchinsky et al.⁴⁶ demonstrated the potential of nanoscale NMR based on quantum sensors by implementing nuclear quadrupole resonance (NQR) spectroscopy^112,113 with single NV defects in diamond. In this study, single NV defects were used to detect nuclear spins (down to only 30 spins) in a 2D layer of hexagonal boron nitride (hBN). As shown in Fig. 8(a), a target hBN flake was placed on the diamond surface, and a single shallow NV center was selected as a nanoscale NQR sensor. Dynamical decoupling experiments allowed for NQR spectroscopy on ¹¹B [Fig. 8(b)], ¹⁰B, and ¹⁴N nuclei from a 30 nm-thick hBN layer, from which it was possible to extract the quadrupolar constants. The authors observed a shift in the resonance frequency as the dimensionality of the h-BN crystal decreased [see Figs. 8(c) and 8(d)]. Through the use of DFT simulations, they were able to attribute this effect to changes in the electrostatic and magnetic environment resulting from the reduction in the number of layers. These findings demonstrate that NV-NQR can provide unprecedented insight into these systems. Recently, an NV-ensemble has been used to perform NQR on hBN, with increased sensitivity.¹⁰⁹ We see NV-NQR as an effective tool for investigating local properties and structures of atomically thin materials. Since around three-quarters of the nuclei in the periodic table possess a quadrupole moment, and such nuclear species are common in 2D materials, NV-NQR may become a standard characterization platform for low-dimensional structures. Moreover, promising alternatives to the NV center, such as the boron vacancy (⁠ $V B −$⁠) centers in hBN, could provide a novel platform for quantum sensing of 2D materials. These ultra-thin systems have recently shown the ability to perform sensing even in ambient conditions, with potential applications in quantum technology, nanoelectronics, and spintronics, as demonstrated in recent studies.^22,114–118

MR based on NV centers in diamond has the potential to be a robust and effective tool for analyzing processes that occur under extreme conditions. While actual NV-MR experiments have not been conducted yet, a few recent proof-of-principle studies have laid the groundwork. In a pioneering study, Doherty et al.¹¹⁹ demonstrated optically detected MR (ODMR) at pressures of up to 60 GPa. In their work, they placed a single-crystal chemical vapor deposition (CVD)-grown diamond inside a diamond anvil cell (DAC), a device able to continuously apply pressures of several hundreds of gigapascals and measured the pressure-dependent shift of the optical zero phonon line and ground state electron spin resonance. The authors of three back-to-back publications built upon this concept by interfacing NV centers with DACs, allowing them to perform vector magnetometry at extreme pressures and cryogenic temperatures.¹²⁰ Hsieh et al.¹²¹ demonstrated the combination of NV-based magnetometry with a DAC by implanting NV centers ∼50 nm beneath the surface of the diamond anvil culet and using scanning confocal microscopy to obtain two-dimensional ODMR maps with diffraction-limited optical resolution. They monitored the change in magnetization of a polycrystalline iron pellet undergoing pressure-driven phase transition up to 22 GPa. Lesik et al.¹²² took a similar approach but used a camera to perform wide-field ODMR with micrometer spatial resolution in their hybrid system, as shown in Fig. 9(a). Yip et al.¹²³ employed a different approach by spreading diamond particles (approximately ∼1 μm in size) on a single-crystal type II superconductor and placing it in a pressure cell. They used a confocal microscopy setup to probe the diamond particles at various locations with different distances to the sample, achieving micrometer spatial resolution. They were able to investigate the superconducting transition temperature, magnetic field profile, and other high-pressure features by probing the diamond particles at different distances to the sample using a confocal microscopy setup. The experiments were conducted at cryogenic temperatures and pressures up to 6 GPa. More recently, NV-based magnetic field sensing at pressures higher than 100 GPa has been demonstrated.¹²⁴ The NV sensor's ability to operate over a wide range of temperatures and pressures, its high sensitivity, and spatial resolution make it an ideal tool for investigating exotic magnetic phenomena. As such, we anticipate that the implementation of NV-MR in these experiments will serve as a distinct tool to explore structures under extreme conditions and uncover new phases or material properties.

A field in which NV-NMR/ESR may play a key role in the future is the study and development of novel technologies for quantum information processing. Coherently coupled electron spins can, in fact, be ideal candidates as elementary constituents of quantum information processors and quantum networks due to their fast gate times and long-range interactions. The work of Schlipf et al.¹²⁵ demonstrated the possibility of utilizing spin-labeled molecules as suitable building blocks that can be deterministically engineered in a scalable way. The potential of NV-based NMR in investigating solid-state spin qubits for use in quantum computation and quantum networks is exemplified by the work of Abobeih et al.¹²⁶ Here, the authors realized atomic-scale imaging using a single NV center in diamond and were able to measure and disentangle nuclear–nuclear spin interactions with high spectral resolution and accuracy. With this method, the three-dimensional structure of a cluster composed of 27 ¹³C spins surrounding the NV defect was reconstructed with sub-angstrom resolution. This study provided crucial information that enabled the authors to subsequently use the same spin system as a diamond-based quantum processor capable of implementing fault-tolerant quantum operations, which are essential for advancing large-scale quantum information processing.¹²⁷ 

The last decade has seen the outstanding capabilities of NV centers in diamond as nanoscale detectors for magnetic signals. First important milestones, such as sensing and imaging of MR signals from single biomolecules, microscale NV-NMR in microfluidics or NV-NMR at surfaces have already been achieved as summarized in Fig. 1. Despite the impressive proof-of-concept studies demonstrated by NV-based magnetic resonance in recent years, several challenges need to be overcome to make this technique viable for real-world applications in the coming years. One of the main future goals for nanoscale NV-NMR is to improve spectral resolution and consequently increase the achievable chemical information for single molecule studies. Micronscale NV-NMR experiment suffer so far from limited concentration sensitivity, which can be overcome by hyperpolarization technique, which may enable single-cell studies in the future. The integration of these techniques with microfluidics can enable screening and cell sorting, with applications in metabolomics, drug discovery, and diagnosis. Additionally, NV-detected magnetic resonance spectroscopy has been shown to be useful in characterizing thin film materials and surface chemistry with micro- to nanoscale spatial resolution under chemically relevant conditions. Improving the chemical information, for instance, by using NV-NQR spectroscopy, will have significant implications for material, energy, or surface sciences. Finally, the possibility of detecting both nuclear and electron spins (NV-NMR and ESR) as well as magnetic fields with the same experimental platform makes NV-detected MR particularly attractive as a multimodal sensor and microscopy platform. This enables the development of different strategies adapted to the complexity of the system under investigation with relative ease of implementation. On the basis of what has been highlighted in this perspective, there is great potential for NV-based MR techniques to make significant advancements in the future, paving the way for the development of novel and powerful tools in the fields of life and materials sciences.

This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (Grant Agreement No. 948049) and from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)—No. 412351169 within the Emmy Noether program. R.R. acknowledges support from the DFG Walter Benjamin Programme (Project No. RI 3319/1-1). The authors acknowledge support by the DFG under Germany's Excellence Strategy–EXC 2089/1‐390776260 and the EXC-2111 390814868 as well as by the Bayerisches Staatministerium für Wissenschaft und Kunst through project IQSense via the Munich Quantum Valley (MQV).

The authors have no conflicts to disclose.

Roberto Rizzato: Conceptualization (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Nick R. von Grafenstein: Conceptualization (equal); Visualization (equal); Writing – review & editing (equal). Dominik B. Bucher: Conceptualization (equal); Supervision (lead); Validation (equal); Visualization (equal); Writing – review & editing (equal).

Data sharing is not applicable to this article as no new data were created or analyzed in this study.