# Fileset

[PDMRReviewVer5.pdf](https://mdr.nims.go.jp/filesets/f9889b09-97a3-46b2-81cb-856495ea36e6/download)

## Creator

Hiroki Morishita, Naoya Morioka, Eikichi Kimura, Keigo Arai, Yuichi Yamazaki, Toshu An, Shigemi Mizukami, Norikazu Mizuochi

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Device engineering for photocurrent detected magnetic resonance and scanning probes using solid-state spin defects](https://mdr.nims.go.jp/datasets/3590d66d-7c0f-4fde-b2e4-60bc484bbcca)

## Fulltext

STAMDevice engineering for photocurrent detected magnetic resonanceand scanning probes using solid-state spin defectsHiroki Morishita, a,b Naoya Morioka, c,d Eikichi Kimura, e Keigo Arai, e YuichiYamazaki, f Toshu An, g Shigemi Mizukami, b,a and Norikazu Mizuochi c,daCenter for Science and Innovation in Spintronics, Tohoku University, 2-1-1, Katahira,Aoba-ku, Sendai 980-8577, Japan ; bWPI Advanced Institute for Materials Research, TohokuUniversity, 2-1-1, Katahira, Aoba-ku, Sendai, Miyagi 980-8577, Japan; c Institute forChemical Research, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan; d Center forSpintronics Research Network, Institute for Chemical Research, Kyoto University, Gokasho,Uji, Kyoto 611-0011, Japan; e School of Engineering, Institute of Science Tokyo, 4259,Nagatsuta-cho, Midori-ku, Yokohama, Kanagawa 226-8501, Japan; f National Institutes forQuantum Science and Technology, 1233 Watanukimachi, Takasaki, Gunma, 370-1292, Japan;g School of Materials Science, Japan Advanced Institute of Science and Technology, 1-1Asahidai, Nomi, Ishikawa 923-1292, JapanARTICLE HISTORYCompiled June 11, 2026ABSTRACTSolid-state spin defects provide a versatile platform for quantum sensing withnanoscale spatial resolution and room-temperature operation. Spin defects in di-amond have enabled mature scanning-probe devices, while related defects in siliconcarbide and hexagonal boron nitride are being actively explored for scalable sensingplatforms. However, the sensitivity of practical and scanning-probe devices remainsbelow that of optimized bulk systems. Although optical fluorescence detection iswidely used, practical performance is often constrained by limitations in signal ac-quisition and readout efficiency, motivating continued efforts to improve readouttechnologies. This review surveys material platforms and optical and photoelec-trical readout technologies for solid-state spin defects. We compare fluorescence-and photoelectric-based detection schemes in terms of readout fidelity, sensitivity,and scalability, and discuss how materials properties and carrier transport influencepractical performance. These perspectives provide guidelines for improving readoutefficiency and advancing high-sensitivity quantum sensors and scanning probes.KEYWORDSSolid-State Quantum Sensor; Quantum Spin; Diamond; Silicon Carbide;Photoelectrical detection; Quantum Scanning Probe;1. IntroductionQuantum sensing has emerged as a powerful approach for detecting extremely weakphysical signals by exploiting quantum mechanical resources such as the quantumcoherence of spin states [1]. Owing to their high sensitivity and non-invasive opera-tion, quantum sensors enable measurements that are difficult or impossible with con-CONTACT Hiroki Morishita. Email: hiroki.morishita.d8@tohoku.ac.jpventional classical probes, particularly when nanoscale spatial resolution and room-temperature functionality are simultaneously required. Among the various solid-stateimplementations, spin defects in wide-bandgap materials are especially attractive be-cause of their robustness, scalability, and reliable operation under ambient conditions.Here, scalability mainly refers to ensemble scaling, namely the ability to increase thenumber of sensing defects or sensing elements contributing to the detected signalthrough materials growth and device fabrication. In this review, scalability refers to thepossibility of increasing the number of sensing defects or sensing elements contributingto the detected signal through materials growth and device fabrication. Their atomic-scale dimensions enable highly localized measurements with nanometer spatial reso-lution, making them well suited for probing microscopic and spatially heterogeneousphenomena in condensed-matter systems. Nitrogen-vacancy (NV) centers in diamondrepresent one of the most mature and versatile platforms [1–3]. The NV electron spincan be optically initialized and read out while maintaining long coherence times evenat room temperature, enabling sensitive detection of magnetic fields, electric fields,temperature, and strain [4]. When incorporated into thin membranes or nanostruc-tures, NV centers further function as scanning quantum probes that bring the sensorinto close proximity with a target sample [5–7]. Such platforms enable nanoscale mag-netometry and current-density imaging with unprecedented spatial resolution [7–10].Despite these advances and the recent commercialization of NV-based probes, thesensitivity of practical scanning implementations remains significantly below the lim-its demonstrated in optimized bulk materials. Commercial devices typically operateat sensitivities on the order of µT/√Hz, whereas optimized sensors range from thenT/√Hz regime for single-spin devices down to the fT/√Hz regime for bulk ensem-ble systems under optimized conditions [11–13]. This performance gap persists eventhough both bulk and scanning implementations rely on optical fluorescence readout.In practice, device performance is influenced by several factors, including system-levelconstraints in miniaturized geometries and limitations in signal acquisition. Althoughfluorescence-based readout remains widely used and versatile, practical optical losses,background signals, and experimental overheads can reduce signal-to-noise ratios andscalability. Consequently, improving readout efficiency has become an important focusfor enhancing the sensitivity of practical quantum sensors.Beyond diamond, alternative host materials are being explored to address scala-bility and device-integration challenges. Silicon carbide (SiC) provides a particularlyattractive semiconductor platform compatible with wafer-scale growth and establishedmicrofabrication processes, enabling large-area devices and straightforward electri-cal integration. In contrast, hexagonal boron nitride (hBN) offers an emerging two-dimensional host that allows atomically thin sensors and extreme sensor-sample prox-imity, opening opportunities for surface-sensitive measurements. However, comparedwith diamond and SiC, hBN-based spin defects remain at an earlier stage of tech-nological maturity, particularly with respect to reproducible defect engineering andelectrical readout implementations. Because this review emphasizes readout engineer-ing for quantum sensors and scanning applications, we focus on the more establisheddiamond and SiC platforms while discussing hBN as a promising future direction. Toaddress the limitations of fluorescence detection, increasing attention has been directedto alternative readout mechanisms that convert spin information into electrical signals.One such approach is photoelectric detection of magnetic resonance (PDMR), in whichspin-dependent photoionization processes generate a measurable photocurrent [14]. Inthis technique, optical excitation is used to drive the photoionization process, whilethe resulting signal is detected electrically. Thus, PDMR combines optical excitation2with electrical readout. This mechanism is distinct from electrically detected magneticresonance (EDMR), where spin-dependent transport or recombination processes aremonitored electrically [15]. PDMR-based photocurrent detection can be implementedusing on-chip electrodes and compact electronics, enabling efficient signal collection,reduced optical complexity, and potential compatibility with integrated and scanning-probe architectures, depending on device implementation. Such approaches provide apathway toward scalable, chip-level quantum sensors with improved signal acquisitionand system robustness.Here we distinguish ODMR, EDMR, and PDMR to avoid confusion among relatedspin readout techniques. A representative comparison relevant to spin-defect quan-tum sensors is summarized in Table 1. ODMR detects magnetic resonance throughspin-dependent optical observables, such as fluorescence, absorption, or polarizationsignals [4,16,17]. EDMR is used here in a broader sense to denote resonance detectionthrough spin-dependent electrical observables, including current [18–22], photocurrent[23–26], capacitance [27], and related responses [28]. PDMR refers to a photoelectricreadout scheme in which optical excitation drives spin-dependent photoionization;the resulting signal is detected electrically, typically as a photocurrent [14]. Althoughthese approaches may share experimental components such as optical excitation ormicrowave driving depending on the material platform, they are distinguished here bytheir dominant readout mechanisms. In this review, we primarily focus on fluorescence-based ODMR and PDMR as representative readout schemes for solid-state spin-defectquantum sensors, while EDMR is included mainly to clarify the broader landscape ofelectrical spin readout techniques.This review surveys material platforms and readout technologies for solid-state spindefects, with an emphasis on engineering optical and photoelectrical signal acquisi-tion for compact, scanning-compatible devices. It is organized as follows. Section 2introduces the material platforms, including NV centers in diamond, silicon vacancy(VSi) defects in SiC, and boron vacancy (VB) centers in hBN. Section 3 describesthe operating principles of spin-defect sensors, including spin initialization, manipu-lation, optical readout, and sensing performance. Section 4 discusses photoelectricaldetection techniques, carrier transport, contact engineering, and charge-state stabil-ity. Section 5 presents scanning-probe fabrication methods and associated limitations.Section 6 compares optical and photoelectrical readout schemes in terms of readoutfidelity, sensitivity, scalability, and practical constraints. Section 7 summarizes repre-sentative demonstrations in diamond and SiC. Finally, Section 8 concludes the reviewand outlines future directions.2. Material Platforms of Spin-Defect Quantum SensorsQuantum sensors are generally classified into three types according to Degen et al. [1]:(1) Sensors that employ quantum objects or quantum states,(2) Sensors that exploit quantum coherence,(3) Sensors that utilize quantum entanglement.While type-3 sensors may ultimately surpass classical sensitivity limits, this reviewfocuses on type-1 and type-2 sensors, which are more practical with current solid-state systems.The operation of all solid-state quantum sensors is predicated on three fundamentalsteps: (1) The preparation of a well-defined quantum state (e.g. spin initialization via3thermal polarization or optical pumping), (2) sensing via spin manipulation interact-ing with external fields and (3) readout of the final quantum state (Fig. 1a, 1b). Spincontrol frequently utilizes magnetic resonance techniques. For NV centers in diamondand defects such as VSi in SiC, the readout is typically optical or photoelectrical.Fig. 1c illustrates representative quantum sensing implementations based on opticaland photoelectrical detection schemes using confocal laser scanning microscopy andscanning photocurrent microscopy, respectively. A focused laser is used for spin initial-ization and readout, either optically via fluorescence or electrically via photocurrent.The objective lens with the magnitude of 50 and NA of 0.7 focuses the excitationbeam onto a single or small ensemble of defects with a nearly diffraction-limited spotdiameter of ∼ 0.5 µm. The focal position is raster-scanned using a piezoelectric stage,providing a typical travel range of up to 100 µm along the x, y, and z directions. Theupper left region schematically shows integrated electrodes for electrical readout anda microwave antenna for coherent spin manipulation. The gap between the electrodesare ∼ 10 µm. In the optical detection scheme, fluorescence emitted from the defectis collected through a pinhole (diameter ∼ 50 µm) and detected using an avalanchephotodiode (APD). In contrast, for photoelectrical detection, a constant bias voltageis applied across the electrodes, and the laser-induced photocurrent is measured usinga transimpedance amplifier followed by lock-in detection.Diamond is known to host a variety of defects and impurities that are advantageousfor quantum devices. Examples of such centers include the NV center [4], the ST1 cen-ter [29] and the NiV center [30]. Among these candidates, the NV center is consideredthe most promising for solid-state quantum sensing due to the extremely long coher-ence times among solid-state electron spins [12,31]. To comprehend the operationalmechanisms of the NV quantum center, the spin Hamiltonian ((H)) can be delineatedby the following equation:H = S⊤ ·D · S + geµBS ·B0, (1)The first term is described as the zero-field splitting (ZFS) term of the NV electronspin (S = 1), which depends on the electric field, temperature and pressure [4]. theZFS parameters of D and E are defined as D ≡ 2Dzz/3 ∼ 2.87 GHz , and E ≡(Dxx +Dyy) /2, respectively. The second term describes the Zeeman interaction of NVelectron spins with a magnetic field, ge ∼ 2.003 and µB are g-factor of the NV electronspin and the Bohr magneton, respectively. Consequently, the diamond NV quantumsensor has the capacity to measure magnetic fields, electric fields, temperature andpressure.SiC is known to harbor a variety of defects that are for utilization in quantumdevices [32,33]. Among of them, VSi is the most extensively studied defect. The defecthas spin values of S = 3/2 (VSi) with its respective spin Hamiltonian exhibiting acomparable character to that of the NV center in diamond [34,35]. The effect of spinvalue on magnetic field measurement is discussed in Sec. 3.3 Since VSi is a singlevacancy, unlike NV centers which possess four quantization axes, its response to themeasured magnetic field is simpler. However, vector measurements require complexprotocols [36]. A significant discrepancy has been observed in their ZFS values: VSihas a ZFS of approximately 35 MHz [37]. Although not discussed in this paper, inmagnetic field measurements using level anti-crossing, this small ZFS also means thatthe required external magnetic field can be reduced by at least one order of magnitudecompared to NV center [38].In addition to bulk hosts such as diamond and SiC, color centers in hBN have re-4cently attracted considerable interest as an alternative platform for solid-state quan-tum sensing. Owing to its atomically thin, van der Waals layered structure, hBN en-ables defect spins to be positioned within a few nanometers—or even sub-nanometerdistances—from target systems. Such extreme proximity is particularly advantageousfor sensing weak magnetic or electric fields from low-dimensional materials, molecularsystems, and surface-bound phenomena, where the detectable signal strength scalesstrongly with sensor-sample distance. Several optically addressable spin defects, includ-ing boron-vacancy (VB) centers, have demonstrated room-temperature spin polariza-tion, microwave manipulation, and ODMR-based readout. The two-dimensional ge-ometry further offers opportunities for integration with heterogeneous material stacksand nanofabricated devices, suggesting potential routes toward ultrathin and flexi-ble quantum sensors. At present, however, the technological maturity of hBN-basedspin defects remains lower than that of diamond and SiC. In particular, reproducibledefect creation, spectral stability, and spin coherence times are still under active de-velopment, and electrical readout schemes such as PDMR have not yet reached thesame level of implementation as in bulk semiconductor hosts. Consequently, while hBNrepresents a promising future platform for surface-proximal and two-dimensional sens-ing, the present review focuses primarily on diamond- and SiC-based systems, whereestablished optical and photoelectrical readout techniques already enable practicalscanning-probe operation.Table 2 summarizes representative material properties of solid-state spin-defect plat-forms, including the NV center in diamond, the VSi in SiC, and the VB center inhBN. These parameters collectively determine the achievable spin-readout contrast,coherence time, optical efficiency, and ultimately the magnetic-field sensitivity attain-able in bulk crystals and device implementations. Among these platforms, diamondNV centers currently provide the most balanced overall performance. Although theDebye-Waller factor is relatively small (∼ 3 %) [39], NV centers exhibit high fluores-cence count rates, large ODMR contrast (up to ∼ 40 %), and long room-temperaturecoherence times [12]. In isotopically purified diamond, millisecond-scale T2 times havebeen achieved, and further improvements can be realized through optimized mate-rial engineering and donor control [12]. Together with excellent thermal stability andmature single-crystal growth technology, these attributes establish diamond as the cur-rent benchmark platform for high-sensitivity quantum sensing and a widely adoptedmaterial for scanning-probe implementations.SiC hosts silicon-vacancy centers and offers a technologically mature semiconduc-tor platform. The availability of large-diameter wafers and relatively low dislocationdensities make SiC promising for mass production. However, lower photon count rates[40] and reduced ODMR contrast presently result in lower spin-readout performancecompared with diamond-based systems [41]. These limitations mainly arise from theintrinsic optical and spin-dependent properties of VSi centers. Crystal can be an im-portant factor for coherence properties such as T ∗2 . However, the transition between|ms⟩ = |+1/2⟩ and |−1/2⟩ of VSi centers can be exploited as a strain-insensitive spinbasis that exhibits T ∗2 much longer than the transitions between |±1/2⟩ and |±3/2⟩and less sensitive to the electron-irradiation dose, indicating a reduced sensitivity toinhomogeneous broadening induced by local strain and crystal inhomogeneity [42]. On-going improvements in material purity, defect engineering, and photonic integrationare expected to mitigate these limitations and enhance overall sensing performance.Two-dimensional hBN has recently emerged as a promising host material due to itsatomically thin geometry, which enables extreme proximity between the spin sensorand the target magnetic source. This feature is particularly advantageous for nanoscale5and surface-sensitive scanning applications. At the same time, quantitative perfor-mance metrics—such as coherence time, ODMR contrast, and sensitivity—remain lessmature than those of diamond and SiC, as reflected by the shorter coherence timesand lower readout contrast summarized in Table 2. Defect reproducibility, spectralstability, and material uniformity therefore remain active research challenges.Overall, the selection of the host material involves trade-offs among spin coherence,optical readout efficiency, fabrication scalability, and thermal robustness. Diamondremains the benchmark platform for high-sensitivity scanning magnetometry owing toits superior combination of coherence and optical readout performance, while SiC offersstrong advantages for scalable semiconductor integration and hBN provides uniqueopportunities for ultrathin, surface-proximal sensing architectures.3. Material-Dependent Spin Initialization, Manipulation, and OpticalReadout3.1. InitializationSolid-state quantum sensors utilize electron and/or nuclear spins, with the initial stepbeing spin polarization (initialization). The simplest and most fundamental spin po-larization mechanism is thermal polarization, where the spin population follows theBoltzmann distribution. Although this mechanism describes the equilibrium spin pop-ulation, it yields only below 1-% polarization at room temperature and typical mag-netic fields. Therefore, thermal polarization is insufficient for high-sensitivity quantumsensing. Consequently, most modern solid-state quantum sensors, such as NV centersin diamond or VSi in SiC, employ active initialization methods of optical pumping toachieve near-unity polarization.In contrast, the proposed and demonstrated quantum sensors based on NV centersin diamond and VSi in SiC can be initialized through optical pumping processes. Thisrepresents a substantial benefit for solid-state quantum sensors, as they are capable offunctioning under ambient conditions over a broad temperature range. The subsequentsection elucidates the mechanism of spin initialization for NV and VSi electron spinsvia optical pumping [4]. As illustrated in Fig. 2a, the energy levels of an NV centerare depicted without a static magnetic field. Following laser excitation, an electron inthe ground state (GS) (3A2) is excited to the excited state (ES) (3E). The electronsubsequently undergoes a relaxation process, either radiatively, returning to the 3A2state (illustrated by red solid arrows) or non-radiatively, via a metastable state (1A1)(depicted by black dashed arrows). These relaxation pathways are spin dependent:when the NV electron spin is |ms⟩ = |0⟩, relaxation is primarily radiative. Conversely,for |ms⟩ = |±1⟩, the electron preferentially relaxes through the non-radiative pathway.Once an electron relaxes to the 1A1 state, it predominantly relaxes to the |0⟩ in the3A2 state. The electron spin can be initialized into the |0⟩ state with a high degree offidelity (i.e. 90 % or more) by repeating the cycle of photoexcitation and relaxation.VSi is classified as a spin defect with S = 3/2 [34] and it has been observed that GSand ES possess spin sublevels |ms⟩ = |±1/2⟩ and |±3/2⟩, respectively. The transitiondiagram of VSi (V2) is illustrated in Fig. 2b [43]. As observed in the NV center indiamond, following photoexcitation (O1 or O2), an electron undergoes either a directtransition with photoemission or a non-radiative transition via metastable states (MS1or MS2) to the GS. For initialization by laser irradiation, the transition probability ofeach transition, including the metastable states, is important. The lifetimes (transition6rates) were measured at 5.5 K, as shown in Fig. 2b [43]. The lifetimes from MS2 tothe GS (γ′3, γ′4) are very long compared to those from MS1 (γ3, γ4) and the formershow laser power dependence. Because of 1/γ3 < 1/γ4, the transition tends to bepreferentially towards |±1/2⟩. Since γ′3 (= γ′4) becomes large under high-laser-powerconditions, the spin polarization after initialization shows laser power dependence.3.2. Optical ReadoutOptically detected magnetic resonance (ODMR) of NV centers is a well-establishedtechnique in which spin-dependent intersystem crossing leads to spin-dependent flu-orescence contrast under microwave excitation. The basic energy-level structure andreadout mechanism have been extensively reviewed elsewhere [4] and are not repeatedhere. Instead, we focus on how materials properties influence the achievable ODMRcontrast. The ODMR contrast of single NV centers directly reflects the intrinsic spinand charge properties of the defect, whereas ensemble measurements are additionallyaffected by dipolar interactions, inhomogeneous broadening, and orientation averaging.Consequently, single-defect contrast provides a clearer metric for evaluating materialquality. NV centers can be created either by nitrogen ion implantation followed byannealing or during CVD growth. Ion implantation typically introduces vacancy clus-ters and complex defect structures near the NV center, which degrade spin coherenceand charge stability, resulting in reduced ODMR contrast compared with high-purityCVD-grown material [44,45]. Charge-state engineering further plays a critical role. In-corporation of phosphorus donors during CVD growth stabilizes the NV−charge state,increasing the steady-state NV− population under optical excitation. Under such con-ditions, nearly 100% NV− occupancy has been reported [46], compared with ∼ 50-70 %in intrinsic diamond [47]. Using n-type phosphorus-doped diamond, single-NV ODMRcontrasts as high as ∼ 37 % have been achieved [12]. Surface termination and near-surface defects also strongly influence ODMR performance by affecting both chargestability and spin coherence through electric and magnetic noise sources [48]. Theseobservations highlight that ODMR contrast is governed not only by the intrinsic defectphysics but also by materials quality and device processing.The principle of ODMR for the SiC VSi defect is analogous to that of the NV center,but its S = 3/2 spin makes the situation more complex. The application of a magneticfield in the direction of the c-axis gives rise to the splitting of each Kramers doubletbecause of the Zeeman effect, thereby yielding two observable transitions (|−1/2⟩ ↔|−3/2⟩ and |+1/2⟩ ↔ |+3/2⟩) [49], where. the ODMR contrast is reduced by half dueto the presence of two distinct populations. In the case of fields misaligned with thec-axis, up to six ODMR signals can be detected.The decrease in the ODMR contrast in the presence of a magnetic field is at-tributable to the S = 3/2 configuration. This is not exclusive to the contrast andthe signal intensity also diminishes in pulse measurements such as Rabi, T1 and T2measurements. To overcome this disadvantage, the measurements were performed oneach of the two split signals (in the case of a parallel magnetic field) and the signalsobtained were summed. This process has been shown to restore the signal intensity toits original value in the absence of a magnetic field [43].73.3. Manipulation3.3.1. Continuous wave magnetic resonanceThe continuous-wave magnetic resonance, including the ODMR and PDMR protocols,exhibits sensitivity to magnetic fields with a frequency in the DC ∼ kHz range. In thissection, the NV quantum sensor is employed as a case study, given its analogousfunctionality to the SiC quantum sensor. The Hamiltonian of the NV center, in thepresence of a magnetic field Bz along the NV axis isHI = DGSS2z + γeBzSz. (2)where DGS is the ZFS, Sz is the z component of the electronic spin and γe = geµBis the gyromagnetic ratio of the NV center. The resulting Zeeman splitting leads toresonance frequenciesω± = DGS ± γeBz. (3)The lifted energy levels are measured as follows: the continuous illumination of a laserexcites the electronic state, which subsequently undergoes relaxation to the |0⟩ state.Simultaneous irradiation with the continuous microwaves whose magnetic field com-ponent is perpendicular to the NV axis has been shown to induce transitions betweenthe |0⟩ and |±1⟩ states, provided that the microwave frequency, ω, matches either ofthe resonant frequencies, ω± (see further details in Sec. 3.3.2). Consequently, the res-onant frequencies, ω±, can be determined by observing the reduction of fluorescencein ODMR or the change in the photocurrent in PDMR.While the basic detection mechanism is well established, the achievable CWmagnetic-field sensitivity is strongly governed by material properties. To first order,the sensitivity can be approximated asηCW ∝ ∆νC√R, (4)where ∆ν is the resonance linewidth, C is the ODMR contrast, and R is the detectedphoton rate [1,2]. All three parameters are highly sensitive to crystal growth method,defect density, strain homogeneity, and charge-state stability.The linewidth ∆ν is fundamentally limited by the inhomogeneous dephasing timeT ∗2 , but in ensemble measurements it is often dominated by spatial variations in lat-tice strain and defect-induced electric-field fluctuations. Recent advances demonstratethat CW-based NV magnetometry can approach the pT/√Hz regime when employ-ing a narrow-linewidth ensemble (∼ 28 kHz) with long dephasing time (T ∗2 ∼ 8.5µs) and optimized optical readout [50]. Such narrow linewidths are achievable only inhigh-quality CVD-grown diamond with excellent strain homogeneity and low residualdefect density. In contrast, strain inhomogeneity significantly broadens the ODMRlinewidth. For example, in 10-MeV electron-irradiated HPHT diamond, irradiation-induced vacancy/interstitial complexes generate spatially varying lattice strain, lead-ing to MHz-scale broadening; a maximal ODMR splitting of ∼ 6 MHz was observedat high fluence (≈ 1018 cm−2) [51]. These results directly illustrate that CW-ODMRperformance is strongly governed by crystal growth conditions and defect engineering.In SiC, similar material-dependent behavior has been observed for VSi defects. ForDC magnetic-field sensing, the nuclear-spin environment strongly limits the inhomo-8geneous dephasing time T ∗2 . Isotope purification (0.15% 13C and 0.01% 29Si) has beenshown to increase T ∗2 at room temperature by approximately one order of magnitude,from ∼0.4 µs to ∼4 µs in the {±1/2,±3/2} spin basis [42]. Furthermore, the silicon-vacancy center in SiC provides an alternative spin basis {−1/2,+1/2} that is largelydisentangled from the zero-field splitting and therefore relatively insensitive to strain.In isotopically purified material, this strain-insensitive basis has demonstrated T ∗2 ≈20 µs [42]. These material improvements have enabled DC magnetic-field sensitivitiesdown to ∼4 nT/√Hz [41], highlighting the critical role of isotope engineering anddefect-basis selection in determining CW sensing performance.3.3.2. Rabi oscillationThe Rabi measurement quantifies the amplitude and frequency of oscillating magneticfields [52,53]. The Hamiltonian with static Bz and oscillating Bx with a frequency ωisHII = DgsS2z + γeBzSz + γeBx cos (ωt)Sx, (5)where Sx is the x component of the electronic spin. When the resonant conditionω ∼ ω+ is met, the Hamiltonian on the rotating frame can be simplified to the followingform:HIIrot =Ω2(|0⟩ ⟨−1|+ |−1⟩ ⟨0|) , (6)where γeBx is denoted as Ω. The time evolution of the initial state |0⟩ is expressed as∣∣ψII(t)〉= cos(Ω2t)|0⟩+ i sin(Ω2t)|1⟩ , (7)which leads to the oscillating probability observing |0⟩:pII|0⟩(t) =12+12cos (Ωt) . (8)This equation suggests that both the amplitude Bx and frequency ω of the magneticfield can be obtained when the resonant frequency ω is tuned to ∼ ω. The pulse whichbrings |0⟩ into |+⟩ = (|0⟩+ |1⟩)√2 is called a π/2-pulse, whereas the pulse whichbrings |0⟩ into |1⟩ is called a π-pulse. These pulses are used in the following sensingprotocols.In practice, Rabi fringes decay over time, and both the achievable contrast and theeffective decay time are strongly influenced by material properties. The Rabi contrastdepends critically on stable spin initialization and readout. Donor doping can stabi-lize the negative charge state and suppress paramagnetic defect formation, therebyenhancing the oscillation amplitude. For example, a Rabi contrast as high as 37% hasbeen reported in donor-engineered diamond [12], demonstrating the direct impact ofdefect and charge-state control on coherent manipulation fidelity. In ensemble mea-surements, the apparent Rabi decay is often dominated not by intrinsic relaxationbut by a spatial distribution of the Rabi frequency Ω across the sensing volume. Suchinhomogeneity arises from variations in microwave-field strength, defect density, and9strain. This effect becomes particularly significant in GHz-field sensing, where ensem-ble inhomogeneous linewidth and microwave-field gradients directly limit performance.Heterodyne detection schemes have been proposed to mitigate this limitation and effec-tively extend the observable Rabi decay time [54,55]. In bulk diamond containing fourcrystallographic NV orientations, only a subset of centers contributes constructivelyunder a given bias field configuration. Preferential alignment during CVD growth re-duces orientation averaging and significantly improves ensemble Rabi contrast [56–58].These observations indicate that Rabi-based sensing performance is governed not onlyby intrinsic spin coherence but also by charge engineering, microwave-field uniformity,and crystallographic control.3.3.3. Ramsey InterferometryThe Ramsey interferometry is sensitive to the energy splitting ω because of the accu-mulation of the relative phase between |0⟩ and |1⟩ [59]. In the first step of the protocol,the state is prepared to be the sensing state∣∣ψIIIs〉upon the application of a π/2-pulse(Rπ/2) to the initial state of |0⟩:∣∣ψIIIs〉= Rπ/2 |0⟩ = |+⟩ = |0⟩+ |1⟩√2, (9)Subsequently, the sensing state acquires a phase under the Hamiltonian HI within aduration of t:∣∣ψIIIs′〉=|0⟩+ e−iω−t |1⟩√2. (10)Using a second π/2-pulse, the relative phase is mapped onto the measurable populationdifference.|ψm⟩ = Rπ/2∣∣ψIIIs′〉=(1 + e−iω−t)|0⟩+(1− e−iω−t)|1⟩√2. (11)The probability of observing |0⟩ ispIII|0⟩(t) =12+12cos (ωt) , (12)from which the amplitude of the target magnetic field can be obtained as the sinefringe frequency.In Ramsey magnetometry, the DC magnetic-field sensitivity scales approximatelyasηRamsey ∝ 1γeC√NT ∗2, (13)indicating that improvements in the inhomogeneous dephasing time T ∗2 directly trans-late into enhanced sensitivity [1,2].10In diamond, the Ramsey linewidth is primarily governed by the 13C nuclear spinbath and paramagnetic impurities. Its dependence on 13C concentration has been ex-tensively studied, showing that T ∗2 can reach ∼ 30 µs at a reduced 13C abundance of0.03% [60,61]. Further improvement can be achieved through controlled phosphorusdonor doping, where donated electrons suppress the formation of paramagnetic va-cancy complexes; under optimized conditions, T ∗2 values approaching ∼ 1.5 ms havebeen reported [12]. For ensembles, T ∗2 is often limited by nitrogen-related paramag-netic defects. Systematic studies indicate that T ∗2 spans approximately 0.1 - 10 µs fornitrogen concentrations in the 1-50 ppm range at room temperature [62]. For shallowNV centers, additional degradation arises from surface band bending and terminationchemistry, which affect charge stability and introduce electric and magnetic noise. Withappropriate donor doping, near-surface NV− centers (∼ 5 nm depth) with T ∗2 ≈ 0.81µs have been demonstrated [63].Similar material-dependent limitations are observed in SiC-based defects. In VSi cen-ters, the nuclear spin environment strongly influences T ∗2 . Isotope purification (0.15%13C and 0.01% 29Si) increases T ∗2 at room temperature by approximately one orderof magnitude, from ∼ 0.4 µs to ∼ 4 µs in the {±1/2,±3/2} basis [42]. Moreover, thestrain-insensitive {−1/2,+1/2} spin basis enables further improvement to T ∗2 ≈ 20µs in isotopically purified samples [42]. These advances have enabled DC magnetic-field sensitivities down to ∼ 4 nT/√Hz [41]. Overall, Ramsey magnetometry directlyconverts materials improvements in nuclear-spin purification, defect control, and sur-face engineering into enhanced DC sensitivity, making T ∗2 a key figure of merit forcomparing diamond and SiC quantum sensors.3.3.4. Spin echoThe fundamental premise of the spin echo protocol is the cancellation of the DCmagnetic field noise by applying a π-pulse in the middle of the Ramsey measurement.A DC magnetic field is canceled over the echo sequence because the first and secondphase accumulations have opposite signs. In other words, if the magnetic field signalis synchronized with the sequence, the phase proportional to Bz is accumulated.The spin-echo protocol suppresses quasi-static magnetic-field noise and reveals theintrinsic coherence time T2 of the defect spin. The achievable AC magnetic-field sen-sitivity therefore scales approximately asηEcho ∝1γeC√NT2, (14)making T2 the key materials-dependent figure of merit in echo-based sensing [1,2].In diamond, the echo decay is governed primarily by coherent coupling between theNV electron spin and the surrounding 13C nuclear-spin bath [64]. Systematic studiesshow that T2 spans from ∼10 µs to 1.8 ms as the 13C concentration is reduced from∼10% to ∼0.3%, exhibiting an approximately inverse dependence on nuclear-spin den-sity [60,61]. Phosphorus donor doping can further extend T2 to ∼2.4 ms by suppressingparamagnetic vacancy complexes [12]. For near-surface NV centers, electric and mag-netic surface noise becomes increasingly important; however, recent work demonstratesthat surface-induced strain engineering combined with weak DC magnetic fields cansignificantly enhance coherence even for ∼1 nm-deep NV centers in 12C-enriched di-amond [65]. These results illustrate that nuclear-spin purification, defect control, andsurface engineering directly determine the achievable echo coherence.11Similar material-dependent behavior is observed in SiC-based spin defects. Innatural-abundance 4H-SiC, single VSi centers exhibit echo signals persisting for atleast 200 µs at room temperature [40]. Ensemble measurements report T2 = 47± 8 µs[66]. Theoretically, millisecond-scale T2 is expected in SiC due to reduced nuclear-spinflip-flop interactions arising from the binary lattice structure and the energetic sepa-ration of 29Si and 13C spin baths at moderate magnetic fields (B ≳ 300 G) [67,68].Consistent with these predictions, T2 values of 1.2 - 1.3 ms have been experimentallydemonstrated for ensemble divacancies in natural-abundance SiC at 20 K [69].Overall, echo-based sensing directly converts improvements in nuclear-spin environ-ment and defect engineering into extended T2, thereby enhancing AC magnetic-fieldsensitivity. Comparison of diamond and SiC indicates that both platforms can reachmillisecond-scale coherence under optimized conditions, though via distinct materialsmechanisms.3.3.5. Dynamical decouplingAs an extension of the echo sequence, a narrower bandwidth of the target signalfrequency can be obtained using dynamic decoupling (DD) protocols [70]. The DDsequence is comprised of a train of π-pulses which are equally spaced by τ , therebysuppressing noises at all frequencies except for the harmonics of τ . The DD sequencehas been subject to several variations, including Carr-Purcell-Meiboom-Gill (CPMG)[71,72], periodic dynamical decoupling (PDD) [73] and XY8 [74]. Because these se-quences are insensitive to magnetic noise other than the harmonics of τ , the coherencetime can be extended significantly. However, the achievable effective coherence timeto TDD2 remains fundamentally limited by materials-dependent factors. Pulse imper-fections arising from microwave-field inhomogeneity, dielectric loss, and heating be-come increasingly important at large pulse numbers. In high-defect-density samples,residual magnetic noise from paramagnetic impurities limits the ultimate extensionof coherence. Thus, improvements in nuclear-spin purification, defect density reduc-tion, and microwave engineering are essential for realizing the full potential of DDprotocols. Moreover, the tunable frequency response of DD sequences enables noisespectroscopy, providing a powerful tool to characterize spin impurities and materialsquality in diamond and SiC [66,75].3.3.6. Quantum heterodyne detectionQuantum heterodyne detection (Qdyne) has been demonstrated to enable ultrahigh-frequency resolution that exceeds the limit set by the coherence time of the colorcenter. Qdyne utilizes a stroboscopic (under-sampling) principle, whereby the phaseof the fluorescence signal is recorded at fixed, equally spaced time intervals. This sam-pling technique has been shown to cause aliasing, thereby enabling the accumulationof phase shifts to accurately determine the frequency of the target oscillating sig-nal [76,77]. Although Qdyne overcomes the coherence-limited frequency resolution ofconventional pulsed sensing, the signal-to-noise ratio per unit time remains governedby materials-dependent parameters such as the contrast C, coherence time T2, andspin-relaxation time T1. Long-term frequency stability further requires suppression ofcharge-state fluctuations, thermal drift, and defect-induced noise. Consequently, evenin advanced protocols such as Qdyne or coherently averaged synchronized readout(CASR), ultimate performance is bounded by spin coherence, relaxation, and materi-als stability. These considerations highlight that while sophisticated control protocols12can surpass apparent coherence limits in frequency resolution, their practical sensitiv-ity and stability remain intrinsically tied to materials engineering.3.4. Sensing performanceIn the domain of quantum sensing, critical performance metrics encompass the targetsignal frequency, sensitivity, frequency resolution and dynamic range. It is generallyaccepted that protocols with a higher target signal frequency offer enhanced sensitivityyet concomitantly exhibit diminished frequency resolution and dynamic range. There-fore, it is imperative to select an appropriate sensing protocol based on the targetsignal.The signal whose frequency exceeds the optical transition rate of the NV centeris potentially difficult to detect, but a higher frequency is still detectable by pulsedsequences. The CW ODMR can detect a DC ∼ kHz signal. The Rabi can detect aDC ∼ GHz signal because the resonant condition occurs within the range of the ZFSorder. The Ramsey can detect DC ∼ 100 kHz signals because the initialization takesa few microseconds in each repetition. The spin echo and DD have the capacity todetect frequencies of an ideal nature that are very high frequency, but in fact, this islimited to tens of MHz because the π-pulse takes ∼ 100 ns.The sensitivity, defined in units of T/√Hz, is a measure of the minimum detectablemagnetic field within a given time. It depends on noise and the slope of the fluorescencesignal with respect to the magnetic field. The noise is fundamentally limited by thephoton shot noise and the spin projection noise. The slope of the curve depends onthe signal contrast and linewidth. The signal contrast is higher in the pulsed sequencethan in the CW-ODMR. In the CW ODMR spectra, the linewidth is limited by thecoherence time T ∗2 but usually deteriorates with the MW power [78]. In the fringe ofRamsey, the linewidth is limited by only T ∗2 . In the echo sequence, the coherence time isextended to T2 , resulting in better sensitivity than that of Ramsey [79]. DD sequencealso extends the coherence time to TDD2 , which gives an even better sensitivity [31,80].The frequency resolution is converted to the bandwidth of the window functions.This is expressed in the following form [81] for the window function of the DD sequencethat has equally τ spaced n pulses:WDD(ω) =sin(ω2nτ)ω2nτtan(ω2τ)sin(ω2nτ). (15)The first sinc term is dominant at the frequency of 1/(2τ) and the peak has a band-width of 1/(nτ). Although increasing n narrows the bandwidth, the total interrogationtime nτ is bounded by the coherence time TDD2 , which is ultimately materials limited.Therefore, improvements in nuclear-spin environment and defect engineering directlyenhance spectral resolution.Consequently, increasing the pulse number n enables the DD sequence to achieveenhanced frequency resolution. In practice, however, the total sensing time nτ is fun-damentally restricted by the coherence time discussed above, which in turn limits theachievable bandwidth. To circumvent this limitation, correlation spectroscopy probesthe correlations between two DD sequences, allowing the bandwidth to reach 1/T1 [82].Because T1 is governed by phonon interactions and defect-related relaxation processes,materials engineering directly determines the achievable resolution. For example, T1can be extended from several milliseconds to the order of one second by cooling the13system from room temperature to 77 K [31]. Ultimately, even narrower bandwidthsapproaching 1 Hz can be achieved in the CASR scheme [83], in which magnetometrysequences with spin readouts are synchronized to an external clock. Nevertheless, thelong-term performance of such advanced schemes remains limited by spin-relaxationprocesses and materials stability.The dynamic range of the detectable signal amplitude is subject to an trade-offwith sensitivity. In CW-ODMR, relatively large magnetic fields can be detected di-rectly. In other pulsed sequences, the oscillating signal fringe possesses a 2π ambiguity,thereby limiting the detectable range to one period of the fringe. Extending the sensingtime improves sensitivity but reduces dynamic range, whereas shortening the sensingtime increases range at the expense of precision. Since the maximum sensing time isfundamentally bounded by coherence, this trade-off is ultimately materials limited.The trade-off between sensitivity and measurement range in the conventional pulsedsequences can be resolved by combining pulse sequences with different interpulse in-tervals and optimizing them via a Bayesian algorithm [84]. This approach enabled adynamic range of nearly seven orders of magnitude for a single NV center at roomtemperature while preserving high sensitivity, which is the largest dynamic range everdemonstrated for NV centers. [84]. Even in such optimized schemes, the achievable per-formance remains constrained by coherence time, readout contrast, and spin-relaxationprocesses determined by materials quality. Overall, while protocol choice determinesthe nominal frequency regime, the achievable sensitivity, frequency resolution, anddynamic range are fundamentally constrained by materials-dependent parameters in-cluding T ∗2 , T2, T1, microwave homogeneity, strain, and defect density.4. Photoelectrical Readout of Spin Defects4.1. Photoelectrical ReadoutPDMR detects the NV spin states by measuring the changes in the photocurrentgenerated under magnetic resonance. Under green laser excitation, electron-hole pairsare produced as the NV center undergoes charge-state conversion between NV− andNV0. In contrast to the SCC, the PDMR monitors spin-dependent ionization via pho-tocurrent [14,85,86]. The overall photoelectrical detection process can be regarded as asequence of consecutive steps comprising photoionization, transport and collection ofthe generated charge carriers, and electrical readout. At a conceptual level, for a fixednumber of defects and excitation cycles, the number of detected charges Ndetectioncanbe expressed asNdetection ∝ Pionization × ηcollection × ηreadout, (16)where Pionization is the photoionization probability, ηcollection is the efficiency with whichthe generated carriers are transported to and collected by the electrodes, and ηreadoutrepresents the effective electrical readout efficiency, including losses and bandwidthlimitations in the readout circuitry. For PDMR, however, a large total number ofdetected charges does not necessarily correspond to a large magnetic-resonance signal.The spin-dependent component (∆NPDMR) can be described conceptually as∆NPDMR ∝ Cspin ×Ndetection, (17)14where Cspin represents the spin-dependent ionization contrast. Thus, even when thecharge-collection efficiency approaches unity under appropriate bias and device con-ditions, the practical PDMR performance can remain limited by the photoionizationprobability and spin-dependent ionization contrast. The signal-to-noise ratio is furtheraffected by dark current, spin-independent background photocurrent, current-amplifiernoise, laser-intensity noise, charge-state fluctuations, and the local charge environmentnear the target defects. To describe these processes in more detail, we first considerthe photocarrier-generation mechanism of the NV center under green-laser excitation,as illustrated in Fig. 3a. Furthermore, it elucidates the laser excitation and relax-ation rates (illustrated by the solid and dashed arrows, respectively), as previouslydiscussed in Ref. [86]. In this process, an electron in the NV− 3A2 state is 3E statesand then to the conduction band (CB). Alternatively, an electron in 1A1 state can beexcited to the CB after relaxation from 3E to 1A1 state. This process of ionizationproduces a conduction electron and converts the NV− to the NV0 state. The NV−state can emit a hole into the valence band and recapturing an electron, thereby cy-cling back to NV−. This continuous cycle maintains steady photocarrier generation,enabling direct photoelectrical readout of the spin resonance. Although photoioniza-tion is accompanied by charge-state conversion during the readout process, the defectis continuously reinitialized under optical excitation, allowing repeated measurementcycles and signal accumulation. Consequently, coherent spin manipulation [86–92],spin-echo measurements [86,90–93], and nuclear-spin-related measurements [91,93,94]can be performed using PDMR in a manner analogous to ODMR. As outlined inSec. 3.1, the NV electron spin can be initialized into the |0⟩ state through the uti-lization of optical pumping. The occurrence of magnetic resonance results in the spinpopulation transitioning to the |±1⟩ state, impacting the way photocarriers are gen-erated by the green laser. If the photoelectrons are predominantly excited from the3E state, the photocurrent undergoes a decrease under resonance, thereby yieldinga negative contrast. Conversely, when excitation primarily occurs from the compar-atively slower excitation from the long-lived 1A1 state, the photocurrent increases,resulting in positive contrast. (Fig. 3b). The predominant pathway is contingent onfactors such as the excitation laser power, which exerts a substantial influence on thecontrast sign [86] (Fig. 3c). As demonstrated by Bourgeois et al. [14], the presenceof acceptor-type defects has been shown to cause an increase in photocurrent duringresonance, or positive contrast. Increasing the excitation power generally enhances thephotoionization probability and the total photocurrent. However, this increase doesnot necessarily produce a proportional improvement in the spin-dependent PDMRsignal. If the resonance-on and resonance-off photocurrents increase by the same rela-tive factor, the PDMR contrast remains unchanged, while the photocurrent differencebetween the two conditions increases. Under shot-noise-limited conditions, this canimprove the signal-to-noise ratio. More generally, however, the power dependence ofthe PDMR signal-to-noise ratio is determined by the relative power dependences ofthe spin-dependent signal, background photocurrent, and noise contributions arisingfrom photocurrent shot noise, electronic readout noise, and laser-intensity fluctuations.Laser-intensity noise is not unique to PDMR and can also affect ODMRmeasurements.Consequently, to optimize PDMR sensitivity, it is imperative to meticulously manip-ulate the excitation conditions, laser wavelength and impurity or defect concentrationsin the diamond. In comparison with optical techniques, photoelectrical detection rep-resents a pivotal technology for the integration of quantum sensors, with the potentialto achieve enhanced collection efficiency, thereby circumventing the optical limitationsimposed by the high refractive index of diamond [86,89]. Utilization of PDMR, the de-15tection of NV electron spin and coupled nuclear spins has been demonstrated [89,93],as has DC and AC magnetic sensing, which exhibit sensitivities of 100 nT/√Hz and29 nT/√Hz, respectively [86,95]. It is noteworthy that Hrubesch et al. [88] estimatedthat PDMR-based magnetometers could be approximately three times more sensitivethan conventional ODMR, thereby highlighting the technique’s potential to advancethe performance of NV quantum sensors.SiC is a material that has been extensively developed in the field of electronics.Due to its advanced device technologies, it is regarded as a promising platform forthe sophisticated integration of quantum technologies and electronics. Consequently,the utilization of spin detection via PDMR in SiC is regarded as a highly promisingtechnique. In SiC, PDMR has been demonstrated for the V2 electronic spin ensem-ble [90] and subsequently, the ENDOR-based detection of 29Si nuclear spins locatedat the second-nearest neighbor sites of the V2 center has also been reported [91]. Re-cently, the photoelectrical imaging and the single-spin detection of the single V2 centerhave been realized [92], as demonstrated in Fig. 5a-c. The realization of single-defectPDMR represents an important milestone for photoelectrical spin detection. Single-defect PDMR detection enables access to various fundamental properties necessary tooptimize PDMR performance, such as signal intensities from a single defect, power-dependent ionization efficiency, the effects of other defects and background, and se-quence strategies to reduce signals. This information deepens physical understanding,which is crucial to designing highly sensitive sensors in the future. Furthermore, single-defect PDMR can be combined with a scanning-probe sensing technique, and thus thesingle-defect PDMR is also important for practical quantum sensing applications. Themechanism of PDMR of the V2 center, as illustrated in Fig. 4, is analogous to that ofthe NV center in diamond. The V2, the negatively charged state, absorbs two photons,resulting in the emission of a charge carrier into the energy bands. The question ofwhether the second photo absorption occurs while the system is in the excited-statequartet or in the metastable doublet state remains unresolved [90]. However, the obser-vation of the positive contrast sign for single defects, the same as in ODMR, indicatesthat ionization predominantly occurs while the system resides in the quartet ES [92].The combination of a properly biased back-to-back Schottky contact device (Fig. 5d;see Sec. 4.3) with laser pulse engineering [88,92] has been demonstrated to achievea signal-to-noise ratio approximately 1.7 times higher than that of shot-noise-limitedODMR for the same defects (Fig. 5e). This improvement arises from efficient chargecollection in the photoelectrical scheme, compared with the limited photon collectionefficiency inherent to ODMR. This enhancement approaches the ODMR enhancementachieved through the utilization of solid-immersion lenses; nevertheless, the adoptionof more efficient ionization would further augment the SNR in PDMR. The improve-ment in PDMR sensitivity has enabled the use of coherent detection techniques forthe readout of single defects, such as Ramsey interferometry (Fig. 5f) and Hahn-echo,thereby evidencing clear nuclear spin coupling. Although the application of PDMR inSiC has been primarily focused on the detection of silicon vacancies, there is a possi-bility that its applicability may extend to a broader range of defects. While the opticaltechnique is inherently limited by the sensitivity of photodetectors, the photoelectricalmethod provides a non-optical pathway that may allow for the detection of defectsspanning a broader range of spectral characteristics, as seen in a recent study [92],revealing an additional unidentified feature in the photocurrent image (X in Fig. 5b)that was not visible in the fluorescence image (Fig. 5a). For such photocurrent-activedefects, photoluminescence-based characterization can be difficult when the defectsare non-emissive or weakly emissive, emit outside the efficient detection range of avail-16able photodetectors, or cannot be spectrally isolated due to overlapping emission fromother centers. In such cases, wavelength-dependent photoionization responses can pro-vide spectroscopic fingerprints that are electrically accessible even down to the single-defect level [96]. When combined with first-principles calculations, photocurrent spec-tra may further help narrow down candidate microscopic defect models [97]. Suchcharacterization may help assess whether these photocurrent-active defects should besuppressed as background-current sources or explored as potential candidates for elec-trically detected sensing. Furthermore, wavelength-dependent ionization processes canprovide photoelectrical detection selectivity. For spin-active defects, this wavelengthselectivity can be useful for improving the selectivity of magnetic-resonance signalstoward target defects, even when spectral filtering of photoluminescence is ineffectivein ODMR [96,98,99]. Further, PDMR can provide detection selectivity through thewavelength dependence of ionization processes [98], which can be useful for improvingsignal contrast even when spectral filtering of the photoluminescence is ineffective.Overall, these results demonstrate the potential for highly sensitive spin detection, aswell as its applicability to a wider variety of defect types. The potential for highlysensitive spin detection, as well as its applicability to a wider variety of defect types, isdemonstrated. This highlights the prospects of PDMR in SiC as a compelling avenuefor future investigation and development in spin-based quantum technologies.4.2. Doping & Carrier TransportFor photoelectrical spin detection, the properties of the host semiconductor directly de-termine the achievable signal-to-noise ratio and therefore the ultimate readout fidelity.Ideally, the material should exhibit (i) low dark current in the absence of illumination,(ii) a low background free-carrier density, and (iii) efficient carrier collection. A lowdark current sets the electrical noise floor of the measurement, while a low equilib-rium carrier density suppresses leakage pathways that reduce photocurrent contrast.Wide-bandgap semiconductors such as diamond and SiC are therefore particularlysuitable for PDMR. Their large bandgaps strongly suppress thermally excited car-riers, which minimizes leakage current even at room temperature. In addition, bothmaterials possess indirect bandgaps that suppress radiative recombination and leadto comparatively long carrier lifetimes. These long lifetimes enhance charge-collectionefficiency and constitute a key intrinsic materials advantage for photoelectrical spinreadout. To fully exploit these intrinsic benefits, however, careful control of mate-rial doping is required. Heavy n-type or p-type doping introduces shallow donors oracceptors that supply free carriers even without optical excitation. These carriers in-crease the dark current and reduce the relative change in photocurrent induced byspin-dependent photoionization, thereby degrading the readout contrast [85,93]. Con-sequently, high-purity or lightly doped materials are generally favored. When dopingis employed, it is primarily used to tune the Fermi level and stabilize the desiredcharge state of the color center rather than to increase conductivity, since excessiveconductivity is detrimental to sensitive photocurrent measurements. In practice, thisideal behavior can further be compromised by recombination centers, extended defects,and trap states that act as carrier sinks and introduce additional fluctuations. Suchdefect-assisted trapping and charge-exchange processes reduce the collected currentand increase low-frequency noise, thereby degrading PDMR contrast and sensitivity,as directly demonstrated in photoelectric NV readout experiments [14]. Recent studiesby Todenhagen et al. [100] and Le et al. [101] further show that carrier transport and17device geometry strongly influence the overall detection efficiency.4.3. Contact EngineeringForming low-resistance Ohmic contacts on wide-bandgap semiconductors such as dia-mond and SiC is generally challenging because their large bandgaps and low intrinsiccarrier densities hinder efficient carrier injection. Achieving Ohmic behavior typicallyrequires specific metallization schemes and post-deposition treatments, for exampleTi/Pt/Au multilayers combined with high-temperature annealing and careful surfacepreparation to reduce interface barriers [102,103]. For PDMR measurements, however,minimizing contact resistance is often less critical than minimizing dark current, asthe dark current directly determines the electrical noise floor and limits the achievablesignal-to-noise ratio. Consequently, photoelectrical spin readout does not necessarilyrequire fully Ohmic contacts. Instead, Schottky-type contacts can be advantageousbecause the Schottky barrier suppresses thermionic leakage and reduces backgroundcurrent, thereby improving photocurrent contrast. In the low-current regime relevantto PDMR, carrier extraction rather than carrier injection governs the detection ef-ficiency. Photogenerated carriers can therefore be efficiently collected by drift in thedepletion region formed by back-to-back Schottky junctions when biased above theflat-band condition, enabling high charge-collection efficiency even without strictlyOhmic behavior [92,101].Practical electrode engineering must also consider material-specific constraints.Nickel is a material of choice for Ohmic and Schottky contacts in 4H-SiC [104], but itsmagnetization can cause inhomogeneous broadening of the PDMR linewidth [90]. Theutilization of non-magnetic materials, such as titanium, has been demonstrated to re-sult in the attainment of narrower linewidths [91]. Leakage current represents anotherkey difference between host materials. In high-purity diamond, the extremely highresistivity (∼ 1015 Ω·cm) allows Ohmic contacts to be used with minimal backgroundcurrent [85,89]. In contrast, SiC has a narrower bandgap and shallower impurities, lim-iting its resistivity to 1-104 Ω·cm even in high-purity 4H-SiC, making leakage currentmore problematic, especially for single-spin detection [104]. In recent demonstrations,leakage currents have been suppressed to the femtoampere level using back-to-backgold Schottky diodes (Fig. 5d), which provide high barrier heights for both elec-trons and holes. Furthermore, Schottky contacts have been demonstrated to facili-tate charge-collection efficiencies approaching unity under appropriate bias conditions[80,89,93], when the carrier transit time is sufficiently shorter than the trapping or re-combination time constants. Indeed, near 100 % charge-collection efficiency has beenreported in semiconductor radiation detectors of SiC devices [105–107] and a single-crystal CVD diamond detector [108] and holes in a CVD diamond detector [109]. Inthe SiC PDMR device reported in Ref. [92], with an estimated carrier lifetime of about1 µs and a carrier transit time on the order of nanoseconds, photocurrent saturationabove the designed flat-band condition was reported, supporting the assumption ofnear-unity charge-collection efficiency in the device.The ability to operate efficiently without strictly Ohmic contacts offers additionaladvantages for defect and device engineering. Forming Ohmic contacts in SiC typicallyrequires high-dose ion implantation and high-temperature annealing [104], processesthat can introduce damage and degrade spin coherence. Schottky-based architecturestherefore provide a practical and low-damage alternative for photoelectrically detectedspin readout, offering both reduced leakage and improved compatibility with sensitive18quantum defects.4.4. Charge State StabilityPhotoelectrical readout in PDMR is generally based on spin-dependent photoioniza-tion processes that involve charge-state conversion of the defect under optical illumi-nation. The generated photocurrent originates from these photoinduced capture andemission events. Therefore, controlled charge-state conversion dynamics are essentialfor producing a photoelectrical signal [100]. This requirement differs from ODMR,where a stable defect charge state is typically desired because ionization reduces flu-orescence contrast and degrades the readout fidelity [12,46]. In PDMR, by contrast,the charge-state transition itself forms the signal mechanism. Nevertheless, completecharge instability is also detrimental, since excessive ionization suppresses the avail-able spin-active population and reduces the photocurrent contrast. Consequently, bothmodalities require an optimal balance between charge stability and controllable con-version dynamics. Although the readout process occurs under nonequilibrium opticalexcitation, the equilibrium charge-state distribution is influenced by the Fermi level po-sition, which is governed by impurity doping and surface termination. The Fermi leveldetermines the initial population of spin-active defects and the availability of carriers,thereby affecting both fluorescence and photocurrent generation. These considerationsapply broadly to different material platforms, including NV center in diamond andVSi in SiC, although the optimal doping and surface conditions depend on the specificdefect levels in each host. Thus, charge-state engineering is a central design parameterfor both diamond- and SiC-based quantum sensors, albeit with different optimizationcriteria for optical and photoelectrical detection.5. Quantum Scanning Probe Implementations5.1. Tip FabricationThe fabrication of monolithic diamond scanning probes typically relies on plasma-based etching processes, such as reactive ion etching (RIE) or inductively coupledplasma (ICP) etching, to define high-aspect-ratio nanopillars and sharp tips [110–113]. While these techniques enable precise nanostructuring, plasma exposure can in-troduce sub-surface damage, including vacancy complexes, amorphized layers, surfacegraphitization, and charge traps near the diamond surface [114–117]. Such defects areparticularly detrimental for shallow NV centers, whose spin properties are stronglyinfluenced by the near-surface environment [118]. Experimental studies have demon-strated that oxygen plasma treatment can substantially modify the structural andelectronic environment of shallow NV centers [119]. In particular, plasma exposurehas been reported to cause the disappearance or modification of pre-existing shal-low NVs, indicating significant alteration of the near-surface defect landscape [115].Such observations are consistent with plasma-induced displacement damage and re-structuring of vacancy-related complexes in the sub-surface region. Oxygen plasmaprocessing can introduce or activate electrically active defects and charge traps nearthe diamond surface [115,116]. Because shallow NV centers reside within the regionmost strongly affected by surface modification, ion-induced defect formation can per-turb their local electrostatic and magnetic environment. Near-surface defect statesand charge traps can modify the local band bending and influence the occupation of19the NV charge-transition level (NV0/−), thereby affecting the stability of the NV−charge state [118,119]. Moreover, increased densities of surface and sub-surface de-fects are known to contribute to electric-field noise and magnetic noise that coupleto shallow NV spins [120–122]. Because the strength of surface-induced decoherencescales strongly with NV depth , plasma-induced near-surface damage can reduce spincoherence times (T2), broaden ODMR linewidths, and degrade magnetic sensitivity[121,123,124]. Importantly, the extent of these effects depends strongly on plasmaparameters, and carefully optimized low-damage ICP processes have been shown tomitigate such degradation [116]. Therefore, from a materials engineering perspective,controlling plasma-induced structural and electronic modifications of the near-surfaceregion is essential for achieving reproducible shallow NV coherence. Optimization re-quires balancing nanostructuring fidelity with minimization of displacement damage,defect complex formation, and charge-trap density in the surface regime.5.2. Surface TerminationShallow NV are essential for nanoscale magnetic imaging in scanning NV microscopywhere spatial resolution scales with the sensor-sample distance. However, at suchshallow depths, NV properties become strongly influenced by surface states. Surfacetermination plays a critical role in determining charge stability, spin coherence, andelectric-field noise, thereby directly impacting magnetic sensitivity and measurementreliability. The charge state of the NV is governed by the local band structure nearthe diamond surface [118,125]. Surface termination modifies the electron affinity (EA)and induces band bending through surface dipoles and charge transfer, thereby shiftingthe Fermi level relative to the NV charge-transition level (NV0/−). These effects canstabilize or destabilize NV− state required for quantum sensing [2]. Oxygen termina-tion typically yields a positive electron affinity (PEA) and moderate downward bandbending at the surface [126]. In this configuration, the near-surface conduction bandedge is lowered relative to vacuum, and the Fermi level remains sufficiently high withinthe band gap to favor occupation of the NV− state. As a result, oxygen-terminateddiamond generally supports stable NV− under ambient conditions and has become thestandard surface preparation for shallow-NV experiments [119]. In contrast, hydrogentermination produces negative electron affinity (NEA) and typically induces upwardband bending, often accompanied by a surface transfer-doping mechanism that formsa two-dimensional hole accumulation layer (2DHG) [125]. The upward band bendingshifts the near-surface Fermi level toward the valence band maximum, reducing elec-tron availability at shallow depths. Consequently, shallow NV− can become unstable,favoring conversion to NV0, which leads to reduced ODMR contrast and charge-stateinstability [127]. Fluorine termination introduces strong surface dipoles due to thehigh electronegativity of fluorine, resulting in enhanced positive electron affinity andcan induce strong downward band bending depending on surface chemistry [128,129].Importantly, band bending not only determines the equilibrium charge state but alsomodulates the occupation of surface trap states and fluctuating charge centers. Theresulting noise sources scale strongly with NV depth and become dominant for shallowNV centers [120,122]. Therefore, surface termination engineering is central to control-ling both NV− charge stability and noise suppression in shallow NV-based magnetom-etry.205.3. Limitations of Quantum Scanning ProbeScanning NV probe microscopy has enabled quantitative imaging of magnetic andelectric fields with nanoscale spatial resolution under ambient conditions and has be-come one of the most powerful solid-state quantum sensing platforms. In most currentimplementations, a single NV center located near the apex of a diamond tip is op-tically initialized and read out using confocal microscopy, while microwave fields areapplied for coherent spin control. This configuration has demonstrated high sensi-tivity and has been widely used for nanoscale magnetometry and current imaging.Nevertheless, several practical constraints arise in scanning-probe implementations.First, ODMR requires both optical excitation for spin initialization and microwaveirradiation for spin manipulation. More importantly, the readout signal relies on fluo-rescence detection. Because the sensing signal is carried by photons, efficient detectionrequires high-numerical-aperture objectives, optical filters, and photon-counting de-tectors. These free-space optical components typically dominate the system size andcomplexity, making miniaturization or on-chip integration challenging. In addition,the high refractive indices of diamond and SiC lead to strong total internal reflection,which fundamentally limits photon extraction efficiency even with optimized opticaldesigns. As a result, only a small fraction of the emitted photons contributes to thedetected signal. These photon-collection constraints complicate probe integration andscalability. Consequently, while ODMR-based scanning probes are mature and highlyeffective, their scalability and device integration are often limited by practical opticalconsiderations rather than by the intrinsic spin physics alone.Photoelectrical readout based on PDMR replaces fluorescence collection with cur-rent detection. Although optical excitation and microwave control are still required,the sensing signal is carried by charge and can be extracted directly through electrodesand processed using integrated electronics. Consequently, bulky photon-collection op-tics are no longer required, enabling more compact device architectures and improvedcompatibility with microfabricated probes [130,131]. In contrast, PDMR performanceis instead limited by carrier transport properties and electronic noise, highlightinga different set of engineering trade-offs. Therefore, ODMR and PDMR should beregarded as complementary techniques rather than mutually exclusive alternatives.ODMR offers mature, often shot-noise-limited optical detection, whereas PDMR pro-vides advantages in electrical integration and potential miniaturization. These comple-mentary characteristics are particularly relevant for next-generation scanning quantumprobes, where compactness and device-level integration are critical. These limitationsmotivate the exploration of alternative readout strategies that reduce the dependenceon fluorescence collection, such as photoelectrical detection based on PDMR.The single-defect PDMR may in principle be combined with a scanning-probe sens-ing technique; however, such an implementation remains an open technical challenge.Scanning-probe ODMR sensors based on near-surface NV centers in diamond nanopil-lars have been widely developed [110,111], and the preservation of spin propertiesafter optimized nanofabrication has been demonstrated [132]. In addition, photocur-rent detection using metal/diamond electrode structures (e.g., Ti/Au contacts) hasbeen reported in planar PDMR devices [85,89,130]. These results indicate that theindividual technological components required for scanning PDMR probes have beenestablished. Nevertheless, their simultaneous integration into a scanning-probe ge-ometry has not yet been achieved. Nanopillar fabrication processes such as etching,focused ion beam processing, or nanostructuring can introduce crystal damage andnear-surface defects, leading to reduced spin coherence and charge-state instability21[118]. Furthermore, implementing nanoscale electrodes on submicron-scale pillars isnontrivial. Metallic electrodes placed within tens to hundreds of nanometers of thespin defect may affect the local electrostatic environment, induce band bending, in-crease background current, and enhance spin relaxation through magnetic Johnsonnoise [120]. These effects are particularly relevant for pillars with diameters on theorder of ∼ 200 nm, where the defect–metal distance can be ∼ 100 nm or less. Whilenanoscale metallization of sharp probe structures has been demonstrated in othersystems, such as SQUID-on-tip devices [133,134], these approaches involve differentmaterials and do not require preservation of spin-defect properties. Therefore, theyprovide only a limited structural reference and do not directly establish the feasibil-ity of scanning PDMR probes. Taken together, although the key elements of scanningPDMR probes have been demonstrated individually, their integration remains a signif-icant technical challenge. Thus, scanning-probe PDMR should be regarded as a futureresearch direction requiring further advances in nanofabrication, surface engineering,and electrode integration.5.4. PDMR & Scanning ProbeA key consideration for integrating spin-defect sensors into scanning-probe and de-vice architectures is the local electromagnetic and material environment in which thesensor operates. In many nanoscale measurement scenarios, such as probing local cur-rent distributions or electronic transport, the sensing element must be positioned inclose proximity to metallic electrodes, conductive layers, or nanostructured devices. Insuch environments, fluorescence-based ODMR readout can be significantly affected byoptical quenching mechanisms, including nonradiative decay channels and plasmonicinteractions near metal surfaces [111,135–137]. These effects can reduce the detectablefluorescence signal and degrade readout reliability, particularly for near-surface de-fects or scanning-probe configurations, although plasmonic enhancement is typicallyrestricted to carefully engineered configurations [138] and is not generally applicableto practical sensing environments. In contrast, PDMR employs spin-dependent pho-toionization processes in which optical excitation generates charge carriers that aredetected electrically as a photocurrent. Because the readout signal is electrical ratherthan optical, PDMR is less directly dependent on efficient fluorescence collection andphoton-collection limitations. Although optical excitation remains necessary, electricalreadout may provide an alternative pathway for accessing spin information throughelectrically contacted device structures. Such an approach may be particularly relevantin extreme near-field sensing geometries where conductive materials are intentionallyplaced in close proximity to spin defects and fluorescence-based readout becomes in-creasingly challenging due to quenching effects. Importantly, this does not imply thatPDMR universally outperforms ODMR. Rather, PDMR may offer distinct advantagesin specific scenarios, particularly in electrically integrated architectures and scanning-probe geometries where optical readout is compromised. Nevertheless, the relative ad-vantages and limitations of ODMR and PDMR in such environments remain an activetopic of investigation, and the superiority of either approach has not yet been estab-lished. Realizing scanning-probe PDMR devices remains technically challenging. Whileindividual components―such as ODMR-based scanning probes [110,111], photocur-rent detection in planar PDMR devices [14], and nanoscale metallization techniques―have been demonstrated, their integration into a single platform is still an open prob-lem. Fabrication processes can introduce surface damage and defects that degrade spin22coherence and charge stability [118], and the presence of nearby metallic electrodesmay influence the local electrostatic environment and induce additional noise sources[120]. In addition, conductive structures may modify the local electrostatic landscapethrough surface band bending, charge trapping, and Fermi-level shifts, potentiallyaffecting defect charge-state stability and spin-dependent photoionization processes.Increased background photocurrent, leakage current, and low-frequency electrical noisemay further reduce the achievable spin-readout contrast and sensitivity. We thereforeemphasize that scanning-probe PDMR should be regarded as a promising but techni-cally demanding direction, rather than a fully established technique. Future progresswill require coordinated advances in defect engineering, electrode design, and noisemitigation strategies.6. Comparison of Optical and Photoelectrical Readout6.1. Technical Features and Device RequirementsOptical readout of solid-state spin defects is most commonly implemented using a con-focal microscope configuration, which also forms the basis of scanning-probe NV mag-netometry. In confocal detection, a tightly focused laser excites a diffraction-limitedvolume and the emitted fluorescence is collected through the same objective lens anddetected using single-photon detectors such as avalanche photodiodes (APDs). Spatialfiltering with a pinhole suppresses out-of-focus background light, enabling high signalcontrast and near shot-noise-limited photon counting [139]. This configuration providesseveral key advantages that are particularly important for scanning-probe sensing:high spatial resolution, efficient background rejection, and compatibility with single-defect measurements at the tip apex. Consequently, confocal detection has becomethe standard platform for single-NV magnetometry and nanoscale imaging, whereprecise optical addressing and low technical noise are essential. For ensemble sensing,wide-field approaches are often adopted to increase the detected photon rate. Thesemay use camera-based imaging for parallel, spatially resolved readout or integratingphotodiodes that collect the total fluorescence intensity [50,83,140,141]. In the lattercase, the analog signal is amplified using a transimpedance amplifier and frequentlyprocessed using differential or lock-in detection to suppress laser intensity noise. Al-though these electronics introduce additional technical noise, the dominant limitationtypically remains photon shot noise, and the sensitivity scales favorably with the num-ber of emitters. Electrical detection provides an alternative approach in which spin-dependent processes are converted into measurable electrical signals. Such techniquesinclude EDMR and PDMR. In PDMR, optical excitation is used to generate carri-ers through photoionization while the resulting signal is detected electrically throughdevice electrodes (Fig. 1c), effectively replacing photon counting with photocurrentmeasurement. Because the signal is an photoelectrical current rather than discretephoton events, low-noise amplification electronics are inherently required. The pho-tocurrent is commonly amplified using a transimpedance amplifier and detected usinglock-in [85,89], DC [88], or digitizer-based schemes [93] depending on the protocol. Inaddition to intrinsic carrier shot noise, the measurement can be affected by amplifiernoise, contact resistance fluctuations, carrier recombination losses, and low-frequencytransport (1/f) noise. These device- and electronics-dependent contributions oftendetermine the practical noise floor and can exceed the fundamental shot noise. Froman integration perspective, the two modalities also differ substantially. Optical detec-23tion typically relies on optical components such as objective lenses, filters, and photondetectors. While parts of the optical path can be fiber-coupled or integrated using com-pact photonic components, optical addressing of the defect generally requires opticalaccess near the sensing region, which can contribute to system-level integration con-straints depending on the device geometry. In contrast, PDMR detects spin-dependentphotoionization as a photocurrent at the device electrodes, which can, in principle, beinterfaced with on-chip electrodes and compact electronics [131], although practical im-plementation requires careful device engineering and low-noise electronic design. Thiscompatibility with semiconductor processing suggests potential for more compact orintegrated sensor architectures, although significant technical challenges remain. Nev-ertheless, it is important to emphasize that PDMR still requires optical excitation forspin initialization and photoionization, and therefore does not eliminate the need foroptical access, but rather replaces only the fluorescence detection pathway. Overall,confocal ODMR provides robust, near shot-noise-limited performance and remains thedominant technique for nanoscale and scanning-probe sensing, whereas PDMR-basedphotocurrent detection offers potential advantages in device-level signal collection andintegration, but its performance remains sensitive to materials quality and electronicnoise. Continued advances in materials quality, contacts, and low-noise electronics aretherefore essential to fully exploit the potential of PDMR-based spin readout.6.2. Readout FidelityIn the quantum sensing community, spin-readout performance is commonly character-ized using the readout-noise factor, σR, introduced by Barry et al. [2]. In this work,we refer to its inverse of read-out factor of σ−1R as the readout fidelity. This quan-tity characterizes the excess noise associated with spin-state readout relative to thespin-projection-noise limit and is distinct from quantum-state-overlap fidelity. Sincethe readout-noise factor is satisfied the condition of σR ≥ 1, the readout fidelity sat-isfies 0 < σ−1R ≤ 1. To evaluate the same readout metric in the presence of arbitraryexperimental noise, we consider two spin-dependent mean signals, S0 and S1, with cor-responding conditional variances σ20 and σ21. At the maximum-slope operating point,the inverse readout-noise factor is expressed asσ−1R =[1 +2(σ20 + σ21)(S0 − S1)2]−1/2. (18)This generalized expression permits σ−1R to be evaluated using experimentally mea-sured noise, including technical and electronic noise. By contrast, when the detectionis assumed to be limited under shot-noise-limited detection conditions, σ2i = Si. Underthis shot-noise-limited assumption, σ−1R reduces toσ−1R =[1 +1C2S̄]−1/2, (19)where C = |S0−S1|S0+S1is the symmetric readout contrast and S̄ = S0+S12 is the mean num-ber of detected photons or carriers per readout cycle. In the low-readout-fidelity limit,C2S̄ ≪ 1, this expression becomes σ−1R ≃ C√S̄. When a detection rate R is reported,the detected number per readout cycle is obtained as S̄ = Rtro, where tro is the read-out duration. Based on this common definition, Table 3 summarizes values of σ−1R for24optical and photoelectrical readout of NV centers in diamond and VSi centers in SiC.The tabulated values were obtained by taking the reciprocal of a reported σR, recalcu-lating σ−1R from reported contrast and detection parameters, or converting a reportedor experimentally measured SNR to a per-readout-cycle quantity using the generalizedexpression above. The evaluation procedure used for each entry is explicitly indicatedin the table. It should be noted that Table 3 is intended to provide a quantitativecomparison of reported readout metrics under a unified definition and does not implythat PDMR currently outperforms ODMR. Table 3 shows that the reported inversereadout-noise factors vary substantially among both single-defect and ensemble mea-surements. The ensemble values should not be interpreted as exhibiting a universal√Nenhancement of σ−1R . In the sensitivity formalism, the number of sensing spins, N , andthe readout-noise factor, σR, represent distinct contributions. Although the total mea-surement SNR may improve approximately as√N , the per-readout-cycle fidelity alsodepends on the collection or carrier-detection efficiency, readout contrast, backgroundsignal, readout duration, and technical noise. For optical detection, many of the tab-ulated values were obtained under a photon-shot-noise-limited assumption. PracticalODMR measurements may nevertheless also be affected by laser-intensity fluctuations,microwave instability, detector noise, and background fluorescence. Similarly, practicalPDMR measurements can be influenced by carrier-transport and recombination pro-cesses, contact resistance, interface instabilities, transimpedance-amplifier noise, lock-in-amplifier noise, and low-frequency electrical fluctuations. The spread of the reportedvalues therefore reflects not only differences between optical and electrical detection,but also differences in readout protocols, measurement durations, device structures,experimental conditions, and estimation procedures. These considerations are particu-larly important for PDMR because its readout performance depends strongly on mate-rials engineering, carrier transport, electrical-contact design, and readout electronics.These considerations are particularly important for PDMR because its readout perfor-mance depends strongly on materials engineering, carrier transport, electrical-contactdesign, and readout electronics. Theoretical analyses have suggested that photoelectri-cal readout may achieve competitive magnetic-field sensitivity under optimized noiseand device conditions [88]. However, current experimental implementations remain af-fected by excess noise associated with carrier transport, interfaces and contacts, andreadout electronics [86,88,142]. Further device and electronic optimization is there-fore required to establish whether the predicted sensitivity advantages can be realizedexperimentally. Accordingly, Table 3 should be interpreted as a summary of protocol-specific reported readout metrics rather than as a direct ranking of ODMR and PDMRor as evidence that photoelectrical readout presently provides superior overall sensingperformance.6.3. SensitivityA fair comparison of magnetic-field sensitivity between PDMR and ODMR is inher-ently complex. As discussed in Secs. 6.1 and 6.2, ODMR measurements frequentlyapproach the photon shot-noise limit, which originates from the Poisson statistics ofdetected photons, whereas PDMR relies on electrical detection of photocurrent andits sensitivity in practical experiments is often constrained by additional noise sourcesarising from carrier transport within the host diamond and from the detection elec-tronics. These noise sources frequently dominate over the intrinsic current shot noiseand therefore prevent the measurement from reaching the current shot-noise limit.25Consequently, the two modalities are not always directly comparable under identicalexperimental conditions. Nevertheless, several recent studies have performed directside-by-side comparisons using identical excitation conditions and the same diamondsamples while varying only the detection method. Zheng et al. [95] demonstratedmicrowave-free magnetometry at the ground-state level anticrossing (∼102.4 mT).They reported that the noise floor of PDMR was approximately four times higherthan that of ODMR. However, the sensitivity normalized to the excitation volumewas reported to be higher for PDMR, which is attributable to its more localizedeffective detection volume [130,142]. In fluorescence-based confocal ODMR, the effec-tive detection volume is determined not only by the excitation profile but also by thefluorescence collection process, including the finite pinhole size and the emission point-spread function. For color centers such as NV and VSi, the Stokes shift further broadensthe effective detection volume because the fluorescence wavelength is longer than theexcitation wavelength. In contrast, PDMR directly detects photocurrent generated atthe excitation site and is therefore not affected by fluorescence collection broadening.Consequently, PDMR approaches the spatial resolution expected for an ideal nonlin-ear optical microscopy technique and can provide a smaller effective sensing volumethan confocal ODMR under comparable excitation conditions. Recent work by oneof the authors directly compared the spatial resolution of photoelectric and confocaloptical readout using a single NV center and reported a narrower image full widthat half maximum for photoelectric detection under identical excitation conditions,supporting this interpretation[130]. This quantity should not, however, be interpretedas evidence that PDMR intrinsically generates a larger magnetic-field signal from anidentical sensing volume. Rather, it reflects the ability to achieve comparable sens-ing performance while probing a smaller effective sensing volume. Similarly, Hruby etal. [142] performed simultaneous continuous-wave measurements using both PDMRand ODMR on the same sample. By employing a yellow-green laser (561 nm) forPDMR to suppress background photoexcitation from other defects such as P1 centers,they reported that, under certain experimental conditions, photoelectric readout canachieve a shot-noise-limited magnetic-field sensitivity that is improved compared withODMR. These results indicate that PDMR-based photoelectrical readout should beregarded as a complementary approach for quantum sensing rather than a universallysuperior alternative to ODMR. Both ODMR and PDMR require optical excitation;however, PDMR-based photoelectric detection may simplify certain chip-scale imple-mentations by removing the need for fluorescence collection optics, while introducingadditional design constraints such as electrode structures and routing. In addition,spatially confined carrier generation may be advantageous in localized probing, spa-tially selective sensing, or integrated sensing architectures where the effective sensingvolume must remain small. In such situations, sensitivity normalized to sensing vol-ume provides a useful complementary figure of merit because absolute magnetic-fieldsensitivity alone does not allow a fair comparison between sensing platforms operatingwith substantially different sensing volumes or numbers of addressed color centers.6.4. Scalability and Sensitivity EnhancementA primary route to improving sensitivity in solid-state quantum sensors is to increasethe number of active color centers participating in the measurement. In ensemble-based quantum sensing using color centers, the detected signal generally scales ap-proximately linearly with the number of emitters N , while the fundamental shot noise26scales as√N . In ODMR, for example, the detected fluorescence rate follows this scal-ing behavior. In ODMR, the detected fluorescence rate scales approximately linearlywith the number of emitters N , while photon shot noise scales as√N . Consequently,the readout efficiency follows the well-known scaling σ−1R ∝ C√R ∝√N , enablingstraightforward sensitivity enhancement through ensemble measurements [2,59]. Thisfavorable statistical scaling has motivated the widespread use of wide-field or bulkoptical detection, where large sensing volumes containing many defects can be in-terrogated simultaneously. In practice, however, ensemble scaling in ODMR is notpurely statistical. Increasing the defect density can introduce additional dephasingmechanisms such as dipolar spin-spin interactions, strain inhomogeneity, and opticalpower broadening, which shorten coherence times and reduce contrast [62,143,144].Therefore, an optimal balance between emitter number and material quality is re-quired to fully benefit from the ideal√N scaling. As discussed in Sec. 3.3, sensitivityenhancement is not limited to increasing the number of emitters. Pulse-sequence en-gineering can extend the effective T2, thereby improving sensitivity through longerphase accumulation. Correlation-based approaches such as Qdyne further enable fre-quency estimation beyond the conventional T2 limit by coherently combining repeatedmeasurements. Moreover, the use of correlated or entangled spin states provides aroute beyond the standard quantum limit. Techniques including spin squeezing andentanglement-assisted sensing have recently demonstrated experimentally improvedsensitivity or enhanced spatial resolution in solid-state ensembles, approaching scalingbeyond the conventional 1/√N behavior toward the Heisenberg limit (1/N) [145,146].These results indicate that sensitivity gains can arise not only from statistical averag-ing but also from quantum control and many-body correlations.For PDMR, a similar√N scaling is expected in principle because the photocurrentincreases proportionally with the number of photoionization events and thus with thenumber of defects. In practice, however, the signal is an analog current that must betransported through the host semiconductor before reaching the external detectionelectronics. During this transport process, carriers may recombine, become trappedat defects, or be lost due to finite carrier lifetime, reducing the collection efficiencyand introducing additional fluctuations. Consequently, the achievable signal is stronglyinfluenced by host-material properties such as carrier lifetime, mobility, and trap den-sity. In addition, photoelectrical readout requires amplification using transimpedanceamplifiers and other circuitry, which introduces further electronic noise as discussedin Sec. 6.1. As a result, both carrier-transport losses inside the material and electronicnoise in the measurement chain can exceed the intrinsic shot noise. These effects oftenprevent ideal√N scaling and make the achievable sensitivity dependent not only ondefect number but also on material quality, device structure, and readout electronics.Therefore, improving PDMR sensitivity relies on both increasing the number of colorcenters and optimizing the host material and device design, including longer carrierlifetimes, reduced trap densities, efficient carrier extraction, stable contacts, and low-noise amplification electronics. With appropriate materials optimization and deviceengineering, ensemble PDMR may approach similar statistical scaling behavior. How-ever, the realization of such scaling remains strongly dependent on carrier transportefficiency, defect density control, and low-noise electronics.276.5. Inherent Limitations and Practical ConsiderationsAlthough photoelectrical readout replaces photon detection with current measure-ment, it does not eliminate the need for optical excitation. PDMR still relies on laserillumination to initialize the spin state and to induce spin-dependent photoionizationand charge-state conversion. Consequently, optical access remains necessary in bothmodalities, and detection substitutes only the fluorescence readout pathway ratherthan the entire optical infrastructure. The two approaches also differ fundamentally intheir detection physics. ODMR detects discrete fluorescence photons immediately afteremission, and the signal is therefore governed primarily by photon collection efficiencyand photon shot noise. In contrast, PDMR measures an analog photocurrent generatedby charge carriers that must be transported through the host semiconductor beforereaching the electrodes. As a result, the readout becomes sensitive to carrier recombi-nation, trapping, finite carrier lifetime, and background photocurrents, in addition tonoise introduced by the amplification electronics. These transport- and device-relatedprocesses can reduce collection efficiency and introduce excess fluctuations beyond theintrinsic shot noise. From a practical standpoint, ODMR typically provides stable,near shot-noise-limited performance with mature and well-established optical hard-ware. Photoelectrical detection, while offering potential advantages in certain devicearchitectures, remains strongly dependent on host-material quality, device engineer-ing, and low-noise electronics. The realization of fully optimized PDMR-based sensorstherefore continues to require significant advances in materials control and electronicdesign.7. Demonstrations of the Quantum Sensor7.1. Diamond NV quantum sensorThe basis of the magnetometry using scanning NV-center microscopy was reported in2008 [5,8,59], which enables the characterization of the stray magnetic field with highsensitivity and atomic-size spatial resolution, without any back-reaction on the sample.The technique has been demonstrated to be operational under ambient conditions, aswell as at cryogenic temperatures [147]. The merits of this technique are twofold, incomparison to both traditional magnetic force microscopy and magnetic resonance mi-croscopy, and it has been demonstrated to engender significant advances in understand-ing the various magnetic natures of materials and nanostructures. In 2012, the firstdocumented instance of stray field imaging in magnetic storage media was reported[11,111]. Subsequently, the magnetic imaging of living cells, i.e. magnetotactic bacte-ria, was achieved using a single electron spin [148]. While Faraday- and Kerr-basedmagneto-optical techniques can visualize magnetization patterns in thin films, theyprobe magnetization indirectly through material-dependent optical responses. In con-trast, solid-state quantum sensors such as NV centers measure the stray magnetic fielditself through Zeeman shifts of the spin resonance frequency, enabling direct and quan-titative magnetic field imaging with nanoscale spatial resolution [149]. This distinctionis particularly important when quantitative reconstruction of stray-field distributionsis required. Scanning NV microscopy is now commercially available, with integratedsystems offered by several companies. Nevertheless, its widespread adoption remainslimited at the current stage. This is primarily due to the relatively high system cost,the limited robustness and lifetime of NV probes, and the comparatively low imagingthroughput inherent to point-by-point scanning and ODMR acquisition. Furthermore,28the requirement of microwave excitation, precise optical alignment, and advanced dataanalysis increases experimental complexity [7,150]. Consequently, although scanningNV microscopy represents a powerful platform for quantitative nanoscale magneticfield imaging, it is presently more suited to specialized investigations than to routinemagnetic characterization.To date, stray-field imaging techniques have been applied to a wide range of mag-netic systems, including bulk materials, thin films and patterned structures with fer-romagnetic, ferrimagnetic and antiferromagnetic metals, oxides and 2D materials. Asdemonstrated in seminal studies, the imaging of vortex cores in patterned Ni-Fe alloyswith in-plane magnetization [151] and domain walls in perpendicularly magnetizedTa/CoFeB/MgO films [152] was a significant milestone in the field. The technique hasalso been employed to reveal the internal structures of the domain walls in insulatingferrimagnetic Tm3Fe5O12 films [153] and is now widely used to characterize stray fieldpatterns in devices [154,155]. In non-collinear magnets, the local dynamics of topolog-ical defects were mapped in bulk FeGe [156] and stray fields were reported for BiFeO3films [157]. The existing body of research on collinear antiferromagnets encompassesthe following studies: first, the domain wall imaging of synthetic and bulk Cr2O3 [158]and second, the investigation of topological spin textures in Fe2O3 [159]. The techniquehas recently become popular for imaging stray fields in patterned antiferromagneticstructures [160–162] and in 2D van der Waals (vdW) magnets [163]. The techniquehas been demonstrated to be a highly effective state-of-the-art magnetometry for theinvestigation of emergent magnetism in moiré magnetic superlattices [164,165].7.2. SiC quantum sensorFor VSi-based sensors, the measurable physical quantities demonstrated to date us-ing ODMR-based techniques include magnetic field, temperature, and electric field.Reported sensitivities include an ensemble DC sensitivity of 4 nT/√Hz [41], a single-defect AC sensitivity of 358 µT/√Hz [166], and an ensemble AC sensitivity of 1.4µT/√Hz [167]. Proof-of-principle stray-field measurements of magnons using V2 de-fects [168] have been reported; however, systematic sensitivity benchmarking for theseimplementations has not yet been fully established. Experimental demonstrations fur-ther support the feasibility of direct sensing within SiC device structures. This isachieved by VSi quantum sensors formed at specific locations within the device afterdevice fabrication process. Since VSi formation will degrade device performance, min-imizing device degradation is critical for accurate sensing. Therefore, there is a needfor VSi selective formation techniques such as particle beam writing [169,170] or fem-tosecond laser irradiation [171], and for controlling the amount of VSi to a minimumalthough this reduces sensor sensitivity. In this respect, this measurement methodis not non-invasive. Furthermore, in power devices, it is necessary to maintain theintended structure to prevent device failure caused by high electric field concentra-tion. For example, it is not possible to introduce a SIL structure to improve photon-collection efficiency. Therefore, improvements of sensor sensitivity achieved throughquantum operations become even more important [172]. For ODMR, florescence fromspin defects is used for sensing. For this reason, measurements cannot be taken underelectrodes, which limits the areas that can be measured. As a device structure becomesmore complex, for example, in trench-type or super-junction-type devices, structuralfactors may further limit the measurable region. On the other hand, PDMR may en-able sensing under the electrode if quantum sensors can be selectively formed in such29regions, which is achievable by adjusting the ion energy. There are no physical con-straints on excitation methods as strict as those on detection methods. It is possibleto choose a method where excitation light is incident from the side of the device, aswas demonstrated with NV center-based quantum sensor [173]. This eliminates thedifficulty of optically exciting defects located directly beneath the electrode. However,since PDMR detects current as a signal, it becomes difficult to identify the origin ofthe signal if there are other current sources (noise) contributing to the detected signal.To improve the spatial resolution in PDMR, it is necessary to reduce noise sourcesas much as possible and separate noise from the signal. Despite these limitations, thedirect sensing techniques using ODMR and PDMR can still be usable because it coverssome of the areas that are critical for power devices.Temperature monitoring has been achieved in a planar-type SiC pn diode, where thetemperature rise induced by current flow was detected using VSi defects via an ODMR-based technique [174]. More recently, electric-field detection has been demonstrated ina vertical pn diode configuration, also based on ODMR readout [175]. These resultsindicate that both thermal and electric field quantities can be accessed within function-ing SiC device structures, providing experimental evidence supporting the feasibilityof in-device sensing using VSi defects. In these papers, ODMR was used. Overall, manySiC-based sensing demonstrations remain at the proof-of-principle stage, and system-atic quantitative performance optimization is an important direction for future work.Beyond sensing performance, the materials platform itself offers distinctive advan-tages. As shown in Table 2, 8-inch SiC wafers are commercially available, and researchscale demonstrations have extended wafer diameters to 12 inches. Furthermore, we canutilize a matured device fabrication technique. In electrically detected sensing methodssuch as PDMR and EDMR, these technologies can be directly leveraged. Specifically,lithography enables patterned electrode formation, metal deposition is used for electri-cal contacts, doping and epitaxial growth allow carrier transport control, and thermalprocessing is used to form low-resistance ohmic contacts. These processes are essen-tial for controlling device characteristics and play a key role in determining sensorperformance and reproducibility. In contrast, ODMR-based sensing relies on opticalexcitation and fluorescence detection and therefore does not require semiconductordevice fabrication processes such as electrode formation or junction engineering. Itsimplementation is largely independent of device processing, except for material prepa-ration steps such as the defect formation of VSi [171]. Currently, for ODMR, VSi-basedquantum sensors exhibit lower sensitivity compared to NV-based quantum sensors dueto intrinsic properties such as lower contrast. However, SiC has advantages in applica-tions that exploit the unique properties of VSi (for example, their small ZFS enableslevel anticrossing at low magnetic fields, allowing all-optical magnetic sensing [38],which is being considered for planetary exploration [22]), as well as in sensing meth-ods that require electrical circuits such as EDMR and PDMR, where device fabricationstrongly influences sensor performance. From an industrial perspective, compatibilitywith semiconductor fabrication processes enables integration with electronic circuits,device miniaturization, and wafer-level manufacturing for large-scale production.8. Conclusions and Future DirectionsIn this review, we have surveyed solid-state spin defects as versatile quantum sen-sors and scanning probe with PDMR-based photocurrent readout techniques. We dis-30cussed three representative material platforms—NV centers in diamond, VSi in SiC,and VB centers in hBN—and compared their respective advantages. NV centers indiamond currently provide the most mature and well-balanced performance for room-temperature sensing, while SiC offers wafer-scale fabrication and partial compatibilitywith conventional semiconductor processing. Meanwhile, hBN represents an emergingtwo-dimensional platform that enables close proximity between the sensor and targetsystems, opening opportunities for ultrathin and surface-sensitive quantum probes. To-gether, these hosts provide a broad materials foundation for scalable solid-state quan-tum sensing platforms based on ensembles or reproducible defect fabrication. Scanningimplementations based on these defects enable local measurements with nanometer-scale spatial resolution and have established quantum probes as powerful tools forcondensed-matter physics. Despite these advances, however, the sensitivity of prac-tical scanning probes remains significantly below the intrinsic limits predicted fromspin coherence and photon statistics. In particular, commercially available NV-basedprobes typically operate at sensitivities on the order of µT/√Hz, whereas isolateddefects in optimized bulk materials can reach the nT/√Hz regime. This performancegap originates from two closely connected challenges. First, bringing defects close tosurfaces—an essential requirement for high spatial resolution—often degrades spin co-herence and charge-state stability due to surface damage, paramagnetic impurities, andelectric-field noise. Second, and equally important, conventional fluorescence-based op-tical readout is often limited by photon-collection efficiency, as only a small fraction ofemitted photons can be collected due to limited numerical aperture, optical losses, andrefraction at the high-refractive-index interfaces of host materials such as diamond andSiC. Optical engineering approaches such as solid-immersion lenses (SILs), nanopil-lars, and other nanophotonic structures can significantly improve photon-collectionefficiency [176–180]. For example, SIL-based optical structures have demonstrated sub-stantial fluorescence enhancement in defect-based quantum systems [176,179]. How-ever, such approaches typically require nanofabrication in the immediate vicinity of thedefect to achieve efficient optical coupling. This can introduce surface damage, strain,and charge noise that may affect the spin and charge properties of near-surface de-fects, which are particularly critical for sensing applications. In addition, fluorescence-based readout fundamentally relies on photon collection through optical systems, andtherefore remains subject to practical constraints such as numerical aperture, opticalalignment, and system complexity. In this context, PDMR provides a complementaryreadout strategy by converting spin-dependent photoionization directly into an elec-trical signal without relying on photon collection optics. Recent experiments in SiChave demonstrated that photocurrent signals obtained through PDMR can exhibitsignal amplification comparable to that achieved using solid-immersion lenses in op-tical detection schemes [92]. To address this readout bottleneck, we have highlightedPDMR based on spin-dependent photoionization. By converting spin-dependent pho-toionization processes directly into a photocurrent detected through device electrodes,PDMR provides an alternative readout pathway that reduces the reliance on bulkyoptical fluorescence collection optics and enables signal readout via electrically con-tacted device structures. Such configurations include planar semiconductor devicesincorporating spin defects within photoconductive or diode-like structures, as well asscanning-probe geometries in which the sensing element is integrated with nanoscaleelectrodes. In these device architectures, metallic or conductive components are oftenpresent in close proximity to the defects, which can significantly suppress fluores-cence signals due to optical quenching mechanisms. As a result, fluorescence-basedODMR can be strongly limited in such environments, although plasmonic enhance-31ment is typically restricted to carefully engineered configurations and is not generallyapplicable to practical sensing environments [138]. In contrast, PDMR enables directelectrical readout that remains operable even in the presence of nearby conductive ormetallic structures, where optical readout may become inefficient or impractical. Thisphotocurrent-based detection therefore offers several potential advantages in electri-cally integrated platforms, including high carrier-collection efficiency under optimizedconditions, simplified optical requirements, compact device architectures, and possibleintegration with scalable semiconductor fabrication [88,92,93,181]Looking forward, the integration of PDMR-based photocurrent detection schemeswith scanning-probe geometries represents a promising direction for next-generationquantum sensors, although practical implementation remains technically challenging.In such architectures, PDMR may offer opportunities for closer local integration ofsensing and readout components, potentially supporting compact probe implementa-tions. Realizing this vision will require coordinated advances in several areas, includingthe creation of high-quality near-surface defects with long coherence times and stablecharge states, low-noise electrical amplification, optimized electrode geometries, andmicrowave delivery schemes compatible with scanning operation. In parallel, improvedsensing protocols that balance contrast, bandwidth, and excitation power will be es-sential for maximizing performance. With continued progress in these areas, scanningprobes employing PDMR-based photocurrent detection based on solid-state spin de-fects may provide a viable pathway toward miniaturized and scalable quantum sensors.In particular, PDMR may offer opportunities for sensing near conductive or metallicinterfaces, where fluorescence-based ODMR can be strongly limited by optical quench-ing at graphene or metal surfaces [111,135–137], although plasmonic enhancement istypically restricted to carefully engineered configurations and is not generally appli-cable to sensing environments [138]. Such capabilities may broaden opportunities forquantum sensing in condensed-matter systems, conductive heterostructures, and re-lated device platforms.FundingH.M. acknowledges the support from the JST PRESTO Grant Number JPMJPR21B9,3rd SIP “Quantum”, the research grant from the Asahi Glass Foundation, JST FOR-EST Grant Number JPMJFR244B, Murata Science and Education Foundation, andKAKENHI Grant Number JP25K22194. N.Mo. acknowledges the support from KAK-ENHI Grant Numbers JP25K01262 and JP23K22796, JST PRESTO Grant NumberJPMJPR245C and research grants from The Mazda Foundation, Murata Science andEducation Foundation and Asahi Glass Foundation. K.A. acknowledges the supportfrom KAKENHI Grant numbers JP23K26528 and JP24K21730 and JST K ProgramGrant Number JPMJKP24F3. Y.Y acknowledges the support from MEXT QuantumLeap Flagship Program (MEXT Q-LEAP) Grant Number JPMXS0118067395 and3rd SIP “Quantum”. T.A. acknowledges the support from KAKENHI Grant Num-ber JP24K01286. S.M. acknowledges the support from KAKENHI Grant NumberJP24K21234 and JP21H05000 and MEXT X-NICS Grant Number JPJ011438. N.Miacknowledges the support from MEXT Quantum Leap Flagship Program (MEXT Q-LEAP) Grant Numbers JPMXS0118067395 and JPXS0120330644 and JST CRESTGrant Number JPMJCR23I5. H.M. and S.M. also thank CSRN of CSIS at TohokuUniversity and the Spintronics Research Network of Japan (Spin-RNJ).32References[1] Degen CL, Reinhard F, Cappellaro P. Quantum sensing. Rev Mod Phys. 2017;89:035002.[2] Barry JF, Schloss JM, Bauch E, et al. Sensitivity optimization for NV-diamond magne-tometry. Rev Mod Phys. 2020;92:015004.[3] Du J, Shi F, Kong X, et al. Single-molecule scale magnetic resonance spectroscopy usingquantum diamond sensors. Rev Mod Phys. 2024;96:025001.[4] Doherty MW, Manson NB, Delaney P, et al. The nitrogen-vacancy colour centre indiamond. Phys Rep. 2013;528:1.[5] Degen CL. Scanning magnetic field microscope with a diamond single-spin sensor. ApplPhys Lett. 2008;92(24):243111.[6] Rondin L, Tetienne JP, Hingant T, et al. Magnetometry with nitrogen-vacancy defectsin diamond. Rep Prog Phys. 2014;77(5):056503.[7] Casola F, van der Sar T, Yacoby A. Probing condensed matter physics with magnetom-etry based on nitrogen-vacancy centres in diamond. Nat Rev Mater. 2018;3(1):17088.[8] Balasubramanian G, Chan IY, Kolesov R, et al. Nanoscale imaging magnetometry withdiamond spins under ambient conditions. Nature. 2008;455:648.[9] van der Sar T, Casola F, Walsworth R, et al. Nanometre-scale probing of spin wavesusing single electron spins. Nat Commun. 2015;6:7886.[10] Palm M, Huxter W, Welter P, et al. Imaging of submicroampere currents in bilayergraphene using a scanning diamond magnetometer. Phys Rev Appl. 2022;17:054008.[11] Rondin L, Tetienne JP, Spinicelli P, et al. Nanoscale magnetic field mapping with asingle spin scanning probe magnetometer. Appl Phys Lett. 2012;100(15):153118.[12] Herbschleb E, Kato H, Maruyama Y, et al. Ultra-long coherence times amongst room-temperature solid-state spins. Nat Commun. 2019;10:3766.[13] Barry JF, Steinecker MH, Alsid ST, et al. Sensitive ac and dc magnetometry withnitrogen-vacancy-center ensembles in diamond. Phys Rev Appl. 2024;22:044069.[14] Bourgeois E, Soucek J, Hruby J, et al. Photoelectric detection of nitrogen-vacancy centersmagnetic resonances in diamond: Role of charge exchanges with other optoelectricallyactive defects. Adv Quantum Technol. 2022;5(5):2100153.[15] Brandt MS, Goennenwein STB, Graf T, et al. Spin-dependent transport in elemental andcompound semiconductors and nanostructures. Phys Sta Sol (c). 2004;1(8):2056–2093.[16] Tarasenko SA, Poshakinskiy AV, Simin D, et al. Spin and optical properties of siliconvacancies in silicon carbide - a review. Phys Status Solidi (b). 2018;255(1):1700258.[17] Suter D. Optical detection of magnetic resonance. Mag Res. 2020;1(1):115–139.[18] Hornmark E, Lyon S, Poindexter E, et al. New features of electrically detected magneticresonance in silicon p-n diodes. Solid State Commun. 2000;116:279–282.[19] Graf T, Brandt M, Nebel C, et al. Electrically detected magnetic resonance studies ofphosphorus doped diamond. Phys B. 2001;308-310:593–597.[20] Lupton JM, McCamey DR, Boehme C. Coherent spin manipulation in molecular semi-conductors: Getting a handle on organic spintronics. Chem Phys Chem. 2010;11:3040–3058.[21] Lo CC, Li J, Appelbaum I, et al. Microwave manipulation of electrically injected spin-polarized electrons in silicon. Phys Rev Appl. 2014;1:014006.[22] Cochrane CJ, Blacksberg J, Anders MA, et al. Vectorized magnetometer for space appli-cations using electrical readout of atomic scale defects in silicon carbide. Sci Rep. 2016;6:37077.[23] Boehme C, Lips K. Theory of time-domain measurement of spin-dependent recombina-tion with pulsed electrically detected magnetic resonance. Phys Rev B. 2003;68:245105.[24] Stegner AR, Boehme C, Huebl H, et al. Electrical detection of coherent 31p spin quantumstates. Nat Phys. 2006;2:835.[25] Morishita H, Vlasenko LS, Tanaka H, et al. Electrical detection and magnetic-field con-trol of spin states in phosphorus-doped silicon. Phys Rev B. 2009;80:205206.[26] Schnegg A, Behrends J, Fehr M, et al. Pulsed electrically detected magnetic resonance33for thin film silicon and organic solar cells. Phys Chem Chem Phys. 2012;14:14418–14438.[27] Brandt MS, Neuberger RT, Stutzmann M. Spin-dependent capacitance of silicon field-effect transistors. Appl Phys Lett. 2000;76:1467.[28] Brandt MS, Goennenwein STB, Stutzmann M. Spin-dependent electronic noise. PhysicaE. 2001;10:67.[29] Balasubramanian P, Metsch MH, Reddy P, et al. Discovery of ST1 centers in naturaldiamond. Nanophoto. 2019;8(11):1993–2002.[30] Nazare MH, Mason PW, Watkins GD, et al. Optical detection of magnetic resonanceof nitrogen and nickel in high-pressure synthetic diamond. Phys Rev B. 1995;51:16741–16745.[31] Bar-Gill N, Pham L, Jarmola A, et al. Solid-state electronic spin coherence time ap-proaching one second. Nat Commun. 2013;4:1743.[32] Weber JR, Koehl WF, Varley JB, et al. Quantum computing with defects. Proc NatlAcad Sci USA. 2010;107(19):8513–8518.[33] Son NT, Anderson CP, Bourassa A, et al. Developing silicon carbide for quantum spin-tronics. Appl Phys Lett. 2020;116(19):190501.[34] Mizuochi N, Yamasaki S, Takizawa H, et al. Continuous-wave and pulsed epr study ofthe negatively charged silicon vacancy with S = 32 and C3v symmetry in n-type 4H−SiC.Phys Rev B. 2002;66:235202.[35] Falk AL, Buckley BB, Calusine G, et al. Polytype control of spin qubits in silicon carbide.Nat Commun. 2013;4:1819.[36] Lee SY, Niethammer M, Wrachtrup J. Vector magnetometry based on s = 32 electronicspins. Phys Rev B. 2015;92:115201.[37] Fuchs F, Stender B, Trupke M, et al. Engineering near-infrared single-photon emitterswith optically active spins in ultrapure silicon carbide. Nat Commun. 2015;6:7578.[38] Simin D, Soltamov VA, Poshakinskiy AV, et al. All-optical dc nanotesla magnetometryusing silicon vacancy fine structure in isotopically purified silicon carbide. Phys Rev X.2016;6:031014.[39] Alkauskas A, Buckley BB, Awschalom DD, et al. First-principles theory of the lumi-nescence lineshape for the triplet transition in diamond NV centres. New J Phys. 2014;16(7):073026.[40] Widmann M, Lee SY, Rendler T, et al. Coherent control of single spins in silicon carbideat room temperature. Nat Mater. 2014;14:164.[41] Lekavicius I, Carter S, Pennachio D, et al. Magnetometry based on silicon-vacancy cen-ters in isotopically purified 4H-SiC. Phys Rev Appl. 2023;19:044086.[42] Lekavicius I, Myers-Ward R, Pennachio D, et al. Orders of magnitude improvement incoherence of silicon-vacancy ensembles in isotopically purified 4H-SiC. PRX Quantum.2022;3:010343.[43] Liu D, Kaiser F, Bushmakin V, et al. The silicon vacancy centers in SiC: determinationof intrinsic spin dynamics for integrated quantum photonics. npj Quantum Inf. 2024;10:72.[44] van Dam SB, Walsh M, Degen MJ, et al. Optical coherence of diamond nitrogen-vacancycenters formed by ion implantation and annealing. Phys Rev B. 2019;99:161203.[45] Achard J, Jacques V, Tallaire A. Chemical vapour deposition diamond single crystalswith nitrogen-vacancy centres: a review of material synthesis and technology for quantumsensing applications. J Phys D. 2020;53(31):313001.[46] Doi Y, Fukui T, Kato H, et al. Pure negatively charged state of the NV center in n-typediamond. Phys Rev B. 2016;93:081203(R).[47] Aslam N, Waldherr G, Neumann P, et al. Photo-induced ionization dynamics of thenitrogen vacancy defect in diamond investigated by single-shot charge state detection.New J Phys. 2013;15(1):013064.[48] Yuan Z, Fitzpatrick M, Rodgers LVH, et al. Charge state dynamics and optically detectedelectron spin resonance contrast of shallow nitrogen-vacancy centers in diamond. PhysRev Res. 2020;2:033263.34[49] Simin D, Fuchs F, Kraus H, et al. High-precision angle-resolved magnetometry withuniaxial quantum centers in silicon carbide. Phys Rev Appl. 2015;4:014009.[50] Zhang C, Shagieva F, Widmann M, et al. Diamond magnetometry and gradiometrytowards subpicotesla dc field measurement. Phys Rev Appl. 2021;15:064075.[51] Levchenko AO, Vasil’ev VV, Zibrov SA, et al. Inhomogeneous broadening of opticallydetected magnetic resonance of the ensembles of nitrogen-vacancy centers in diamondby interstitial carbon atoms. Appl Phys Lett. 2015;106(10):102402.[52] Aiello CD, Hirose M, Cappellaro P. Composite-pulse magnetometry with a solid-statequantum sensor. Nat Commun. 2013;4:2375.[53] Fedder H, Dolde F, Rempp F, et al. Towards T1-limited magnetic resonance imagingusing rabi beat. Appl Phys B. 2011;102:497.[54] Wang Z, Kong F, Zhao P, et al. Picotesla magnetometry of microwave fields with dia-mond sensors. Sci Adv. 2022;8(32):eabq8158.[55] Meinel J, Vorobyov V, Yavkin B, et al. Heterodyne sensing of microwaves with a quantumsensor. Nat Commun. 2021;12(1):2737.[56] Michl J, Teraji T, Zaiser S, et al. Perfect alignment and preferential orientation ofnitrogen-vacancy centers during chemical vapor deposition diamond growth on (111)surfaces. Appl Phys Lett. 2014;104(10):102407.[57] Lesik M, Tetienne JP, Tallaire A, et al. Perfect preferential orientation of nitrogen-vacancy defects in a synthetic diamond sample. Appl Phys Lett. 2014;104(11):113107.[58] Fukui T, Doi Y, Miyazaki T, et al. Perfect selective alignment of nitrogen-vacancy centersin diamond. Appl Phys Express. 2014;7(5):055201.[59] Taylor JM, Cappellaro P, Childress L, et al. High-sensitivity diamond magnetometerwith nanoscale resolution. Nat Phys. 2008;4:810.[60] Mizuochi N, Neumann P, Rempp F, et al. Coherence of single spins coupled to a nuclearspin bath of varying density. Phys Rev B. 2009;80:041201.[61] Balasubramanian G, Neumann P, Twitchen D, et al. Ultralong spin coherence time inisotopically engineered diamond. Nat Mater. 2009;8:383–7.[62] Bauch E, Singh S, Lee J, et al. Decoherence of ensembles of nitrogen-vacancy centers indiamond. Phys Rev B. 2020;102:134210.[63] Ohashi K, Rosskopf T, Watanabe H, et al. Negatively charged nitrogen-vacancy centersin a 5 nm thin 12C diamond film. Nano Lett. 2013;13(10):4733–4738. PMID: 24020334.[64] Childress L, Gurudev Dutt MV, Taylor JM, et al. Coherent dynamics of coupled electronand nuclear spin qubits in diamond. Science. 2006;314(5797):281.[65] Pershin A, Tarkanyi A, Verkhovlyuk V, et al. A coherence-protection scheme for quantumsensors based on ultra-shallow single nitrogen-vacancy centers in diamond. Nat Commun.2025;16(1):9797.[66] Simin D, Kraus H, Sperlich A, et al. Locking of electron spin coherence above 20 ms innatural silicon carbide. Phys Rev B. 2017;95:161201.[67] Yang LP, Burk C, Widmann M, et al. Electron spin decoherence in silicon carbide nuclearspin bath. Phys Rev B. 2014;90:241203.[68] Seo H, Falk AL, Klimov PV, et al. Quantum decoherence dynamics of divacancy spinsin silicon carbide. Nat Commun. 2016;7:12835.[69] Christle DJ, Falk AL, Andrich P, et al. Isolated electron spins in silicon carbide withmillisecond coherence times. Nat Mater. 2015;14:160.[70] Viola L, Lloyd S. Dynamical suppression of decoherence in two-state quantum systems.Phys Rev A. 1998;58:2733–2744.[71] Naydenov B, Dolde F, Hall LT, et al. Dynamical decoupling of a single-electron spin atroom temperature. Phys Rev B. 2011;83:081201.[72] Gullion T, Baker DB, Conradi MS. New, compensated carr-purcell sequences. Journalof Magnetic Resonance (1969). 1990;89(3):479–484.[73] Khodjasteh K, Lidar DA. Fault-tolerant quantum dynamical decoupling. Phys Rev Lett.2005;95:180501.[74] Ryan CA, Hodges JS, Cory DG. Robust decoupling techniques to extend quantum co-35herence in diamond. Phys Rev Lett. 2010;105:200402.[75] Bar-Gill N, Pham L, Belthangady C, et al. Suppression of spin-bath dynamics for im-proved coherence of multi-spin-qubit systems. Nat Commun. 2012;3:858.[76] Boss JM, Cujia KS, Zopes J, et al. Quantum sensing with arbitrary frequency resolution.Science. 2017;356:837.[77] Schmitt S, Gefen T, Stürner FM, et al. Submillihertz magnetic spectroscopy performedwith a nanoscale quantum sensor. Science. 2017;356:832.[78] Dréau A, Lesik M, Rondin L, et al. Avoiding power broadening in optically detectedmagnetic resonance of single NV defects for enhanced dc magnetic field sensitivity. PhysRev B. 2011;84:195204.[79] Stanwix PL, Pham LM, Maze JR, et al. Coherence of nitrogen-vacancy electronic spinensembles in diamond. Phys Rev B. 2010;82:201201.[80] de Lange G, Wang ZH, Risté D, et al. Universal dynamical decoupling of a single solid-state spin from a spin bath. Science. 2010;330(6000):60–63.[81] Hirose M, Aiello CD, Cappellaro P. Continuous dynamical decoupling magnetometry.Phys Rev A. 2012;86:062320.[82] Laraoui A, Dolde F, Burk C, et al. High-resolution correlation spectroscopy of 13C spinsnear a nitrogen-vacancy centre in diamond. Nat Commun. 2013;4:1651.[83] Glenn DR, Bucher DB, Lee J, et al. High-resolution magnetic resonance spectroscopyusing a solid-state spin sensor. Nature. 2018;555:351.[84] Herbschleb ED, Kato H, Makino T, et al. Sensitivity optimization for NV-diamondmagnetometry. Nat Commun. 2021;12:306.[85] Bourgeois E, Jarmola A, Siyushev P, et al. Photoelectric detection of electron spin res-onance of nitrogen-vacancy centres in diamond. Nat Commun. 2015;6:8577.[86] Morishita H, Morioka N, Nishikawa T, et al. Spin-dependent dynamics of photocarriergeneration in electrically detected nitrogen-vacancy-based quantum sensing. Phys RevAppl. 2023;19:034061.[87] Gulka M, Bourgeois E, Hruby J, et al. Pulsed photoelectric coherent manipulation anddetection of N−V center spins in diamond. Phys Rev Appl. 2017;7:044032.[88] Hrubesch FM, Braunbeck G, Stutzmann M, et al. Efficient electrical spin readout ofNV− centers in diamond. Phys Rev Lett. 2017;118:037601.[89] Siyushev P, Nesladek M, Bourgeois E, et al. Photoelectrical imaging and coher-ent spin-state readout of single nitrogen-vacancy centers in diamond. Science. 2019;363(6428):728–731.[90] Niethammer M, Widmann M, Rendler T, et al. Coherent electrical readout of defectspins in silicon carbide by photo-ionization at ambient conditions. Nat Commun. 2019;10:5569.[91] Nishikawa T, Morioka N, Abe H, et al. Electrical detection of nuclear spins via siliconvacancies in silicon carbide at room temperature. Appl Phys Lett. 2022;121(18):184005.[92] Nishikawa T, Morioka N, Abe H, et al. Coherent photoelectrical readout of single spinsin silicon carbide at room temperature. Nat Commun. 2025;16:3405.[93] Morishita H, Kobayashi S, Fujiwara M, et al. Room temperature electrically detectednuclear spin coherence of NV centres in diamond. Scientific Report. 2020;10(1):792.[94] Gulka M, Wirtitsch D, Ivady V, et al. Room-temperature control and electrical readoutof individual nitrogen-vacancy nuclear spins. Nat Commun. 2021;12(1):4421.[95] Zheng H, Hruby J, Bourgeois E, et al. Electrical-readout microwave-free sensing withdiamond. Phys Rev Appl. 2022;18:024079.[96] Okajima K, Nishikawa T, Abe H, et al. Photoionization current spectroscopy of individ-ual silicon vacancies in silicon carbide. Nano Lett. 2026;26(17):5636.[97] Londero E, Bourgeois E, Nesladek M, et al. Identification of nickel-vacancy defects bycombining experimental and ab initio simulated photocurrent spectra. Phys Rev B. 2018;97:241202.[98] Morioka N, Nishikawa T, Abe H, et al. Photoelectrical detection and characterization ofdivacancy and PL5-PL7 spins in silicon carbide. Phys Rev B. 2026;113:104426.36[99] Zappacosta A, Haylock B, Fisher P, et al. Wavelength-dependent electrical readout ofspin ensembles in a thin-film sic-on-insulator platform. Nano Lett. 2026;26(17):5628.[100] Todenhagen LM, Brandt MS. Optical and electrical readout of diamond NV centers independence of the excitation wavelength. Appl Phys Lett. 2025;126(19):194003.[101] Le XP, Mayer L, Magaletti S, et al. Field-effect detected magnetic resonance of nitrogen-vacancy centers in diamond based on all-carbon schottky contacts. Commun Eng. 2025;4(1):209.[102] Moazed K, Nguyen R, Zeidler J. Ohmic contacts to semiconducting diamond. IEEEElectron Device Lett. 1988;9(7):350–351.[103] Takeuchi D, Yamanaka S, Watanabe H, et al. Device grade B-doped homoepitaxialdiamond thin films. Phys Status Solidi A. 2001;186(2):269–280.[104] Kimoto T, Cooper JA. Fundamentals of silicon carbide technology: Growth, characteri-zation, devices, and applicationse. John Wiley & Sons, Ltd; 2014.[105] Nava F, Wagner G, Lanzieri C, et al. Investigation of Ni/4H− SiC diodes as radiationdetectors with low doped n-type 4H − SiC epilayers. Nucl Instrum Methods Phys ResA. 2003;510:273–280.[106] Nava F, Castaldini A, Cavallini A, et al. Radiation detection properties of 4H − SiCschottky diodes irradiated up to1016 n/cm2by 1 MeV neutrons. IEEE Trans Nucl Sci.2006;53:2977–2982.[107] Bruzzi M, Lagomarsino S, Nava F, et al. Characterisation of epitaxial SiC schottkybarriers as particle detectors. Diam Relat Mater. 2003;12:1205–1208.[108] Abdel-Rahman MA, Lohstroh A, Bryant P. Alpha spectroscopy and X-ray induced pho-tocurrent studies of a SC CVD diamond detector fabricated with PLD contacts. RadiatPhys Chem. 2019;164:108357.[109] Kaneko JH, Fujita F, Konno Y, et al. Growth and evaluation of self-standing CVDdiamond single crystals on off-axis (001) surface of HP/HT type IIa substrates. DiamRelat Mater. 2012;26:45–49.[110] Appel P, Neu E, Ganzhorn M, et al. Fabrication of all diamond scanning probes fornanoscale magnetometry. Rev Sci Instru. 2016;87(6):063703.[111] Maletinsky P, Hong S, Grinolds MS, et al. A robust scanning diamond sensor fornanoscale imaging with single nitrogen-vacancy centres. Nat Nanotech. 2012;7:320.[112] Castelletto S, Rosa L, Blackledge J, et al. Advances in diamond nanofabrication forultrasensitive devices. Microsyst Nanoeng. 2017;3:17061.[113] Kainuma Y, Hayashi K, Tachioka C, et al. Scanning diamond nv center magnetometerprobe fabricated by laser cutting and focused ion beam milling. J Appl Phys. 2021;130(24):243903.[114] Uzan Saguy C, Cytermann C, Brener R, et al. Damage threshold for ion beam inducedgraphitization of diamond. Appl Phys Lett. 1995;67(9):1194–1196.[115] Kim M, Mamin HJ, Sherwood MH, et al. Effect of oxygen plasma and thermal oxidationon shallow nitrogen-vacancy centers in diamond. Appl Phys Lett. 2014;105(4):042406.[116] Fávaro de Oliveira F, Momenzadeh SA, Wang Y, et al. Effect of low-damage inductivelycoupled plasma on shallow nitrogen-vacancy centers in diamond. Appl Phys Lett. 2015;107(7):073107.[117] Ruf M, IJspeert M, van Dam S, et al. Optically coherent nitrogen-vacancy centers inmicrometer-thin etched diamond membranes. Nano Lett. 2019;19(6):3987–3992. PMID:31136192.[118] Janitz E, Herb K, Völker LA, et al. Diamond surface engineering for molecular sensingwith nitrogen-vacancy centers. J Mater Chem C. 2022;10:13533–13569.[119] Ohno K, Joseph Heremans F, Bassett LC, et al. Engineering shallow spins in diamondwith nitrogen delta-doping. Appl Phys Lett. 2012;101(8):082413.[120] Myers BA, Das A, Dartiailh MC, et al. Probing surface noise with depth-calibrated spinsin diamond. Phys Rev Lett. 2014;113:027602.[121] Romach Y, Müller C, Unden T, et al. Spectroscopy of surface-induced noise using shallowspins in diamond. Phys Rev Lett. 2015;114:017601.37[122] Sangtawesin S, Dwyer BL, Srinivasan S, et al. Origins of diamond surface noise probedby correlating single-spin measurements with surface spectroscopy. Phys Rev X. 2019;9:031052.[123] Rosskopf T, Dussaux A, Ohashi K, et al. Investigation of surface magnetic noise byshallow spins in diamond. Phys Rev Lett. 2014;112:147602.[124] Kim M, Mamin HJ, Sherwood MH, et al. Decoherence of near-surface nitrogen-vacancycenters due to electric field noise. Phys Rev Lett. 2015;115:087602.[125] Ristein J. Surface science of diamond: Familiar and amazing. Sur Sci. 2006;600(18):3677–3689.[126] Yamano H, Kawai S, Kato K, et al. Charge state stabilization of shallow nitrogenvacancy centers in diamond by oxygen surface modification. Jpn J Appl Phys. 2017;56(4S):04CK08.[127] Hauf MV, Grotz B, Naydenov B, et al. Chemical control of the charge state of nitrogen-vacancy centers in diamond. Phys Rev B. 2011;83:081304.[128] Cui S, Hu EL. Increased negatively charged nitrogen-vacancy centers in fluorinated di-amond. Appl Phys Lett. 2013;103(5):051603.[129] Rietwyk KJ, Wong SL, Cao L, et al. Work function and electron affinity of the fluorine-terminated (100) diamond surface. Appl Phys Lett. 2013;102(9):091604.[130] Nakamura S, Morioka N, Mizuochi N, et al. Scanning photocurrent imaging resolutionof single NV center in diamond. Appl Phys Lett. 2025;127(22):224003.[131] Ashok M, Hu Y, Wang H, et al. Heterogeneously integrated nitrogen-vacancy sensing forreal-time cmos security threat detection. IEEE Trans VLSI Syst. 2025;33(11):3145–3155.[132] Neu E, Appel P, Ganzhorn M, et al. Photonic nano-structures on (111)-oriented diamond.Appl Phys Lett. 2014;104:153108.[133] Finkler A, Segev Y, Myasoedov Y, et al. Self-aligned nanoscale squid on a tip. NanoLett. 2010;10(3):1046–1049.[134] Anahory Y, Naren HR, Lachman EO, et al. Squid-on-tip with single-electron spin sensi-tivity for high-field and ultra-low temperature nanomagnetic imaging. Nanoscale. 2020;12:3174–3182.[135] Anger P, Bharadwaj P, Novotny L. Enhancement and quenching of single-molecule flu-orescence. Phys Rev Lett. 2006;96:113002.[136] Brenneis A, Gaudreau L, Seifert M, et al. Ultrafast electronic readout of diamondnitrogen-vacancy centres coupled to graphene. Nat Nanotech. 2015;10:135.[137] Leibold JP, Todenhagen LM, Althammer M, et al. Influence of platinum thin films onthe photophysical and quantum properties of near-surface nv centers. Adv Opt Mater.2026;14(7):e03544.[138] Hapuarachchi H, Campaioli F, Jelezko F, et al. Plasmonically engineered nitrogen-vacancy spin readout. Opt Express. 2024;32(13):22352–22361.[139] Gruber A, Dräbenstedt A, Tietz C, et al. Scanning confocal optical microscopy andmagnetic resonance on single defect centers. Science. 1997;276(5321):2012–2014.[140] Scholten SC, Healey AJ, Robertson IO, et al. Widefield quantum microscopy withnitrogen-vacancy centers in diamond: Strengths, limitations, and prospects. J Appl Phys.2021;130(15):150902.[141] Wolf T, Neumann P, Nakamura K, et al. Subpicotesla diamond magnetometry. PhysRev X. 2015;5:041001.[142] Hruby J, Gulka M, Mongillo M, et al. Magnetic field sensitivity of the photoelectricallyread nitrogen-vacancy centers in diamond. Appl Phys Lett. 2022;120(16):162402.[143] Hayashi K, Matsuzaki Y, Ashida T, et al. Experimental and theoretical analysis of noisestrength and environmental correlation time for ensembles of nitrogen-vacancy centersin diamond. J Phys Soc Japan. 2020;89(5):054708.[144] Blinder R, Mindarava Y, Tran TH, et al. Reducing inhomogeneous broadening of spinand optical transitions of nitrogen-vacancy centers in high-pressure, high-temperaturediamond. Com Mater. 2024;5:224.[145] Rovny J, Kolkowitz S, de Leon NP. Multi-qubit nanoscale sensing with entanglement as38a resource. Nature. 2025;647(8091):876–882.[146] Zhou X, Wang M, Ye X, et al. Entanglement-enhanced nanoscale single-spin sensing.Nature. 2025;647(8091):883–888.[147] Pelliccione M, Jenkins A, Ovartchaiyapong P, et al. Scanned probe imaging of nanoscalemagnetism at cryogenic temperatures with a single-spin quantum sensor. Nat Nanotech.2016;11:700.[148] Sage DL, Arai K, Glenn DR, et al. Optical magnetic imaging of living cells. Nature.2013;496:486.[149] Chen J, Zhou T, Ho KO, et al. Widefield in situ nv-moke microscopy for magnetometry.Revi Sci Instrum. 2025;96(12):123711.[150] Hong S, Grinolds MS, Pham LM, et al. Nanoscale magnetometry with nv centers indiamond. MRS Bulletin. 2013;38:155.[151] Rondin L, Tetienne JP, Rohart S, et al. Stray-field imaging of magnetic vortices with asingle diamond spin. Nat Commun. 2013;4:2279.[152] Tetienne JP, Hingant T, Kim JV, et al. Nanoscale imaging and control of domain-wallhopping with a nitrogen-vacancy center microscope. Science. 2014;344(6190):1366–1369.[153] Vélez S, Schaab J, Wörnle MS, et al. High-speed domain wall racetracks in a magneticinsulator. Nat Commun. 2019;10:4750.[154] Zhang X, Cai W, Wang M, et al. Spin-torque memristors based on perpendicular mag-netic tunnel junctions for neuromorphic computing. Adv Sci. 2021;8(10):2004645.[155] McLaughlin NJ, Li S, Brock JA, et al. Local control of a single nitrogen-vacancy centerby nanoscale engineered magnetic domain wall motion. ACS Nano. 2023;17(24):25689–25696. PMID: 38050827.[156] Dussaux A, Schoenherr P, Koumpouras K, et al. Local dynamics of topological magneticdefects in the itinerant helimagnet fege. Nat Commun. 2016;7:12430.[157] Gross I, Akhtar W, V Garcia amd L J Mart́ınez SC, et al. Real-space imaging of non-collinear antiferromagnetic order with a single-spin magnetometer. Nature. 2017;549:252.[158] Hedrich N, Wagner K, Pylypovskyi OV, et al. Nanoscale mechanics of antiferromagneticdomain walls. Nat Phys. 2021;17:574.[159] Tan AKC, Jani H, Högen M, et al. Revealing emergent magnetic charge in an antiferro-magnet with diamond quantum magnetometry. Nat Mater. 2024;23:205.[160] Yan GQ, Li S, Lu H, et al. Quantum sensing and imaging of spin-orbit-torque-driven spindynamics in the non-collinear antiferromagnet Mn3Sn. Adv Mater. 2022;34(23):2200327.[161] Li S, Huang M, Lu H, et al. Nanoscale magnetic domains in polycrystalline Mn3Snfilms imaged by a scanning single-spin magnetometer. Nano Lett. 2023;23(11):5326–5333. PMID: 37219013.[162] Rickhaus P, Pylypovskyi OV, Seniutinas G, et al. Antiferromagnetic nanoscale bit ar-rays of magnetoelectric Cr2O3 thin films. Nano Lett. 2024;24(42):13172–13178. PMID:39387710.[163] Thiel L, Wang Z, Tschudin MA, et al. Probing magnetism in 2D materials at thenanoscale with single-spin microscopy. Science. 2019;364(6444):973–976.[164] Song T, Sun QC, Anderson E, et al. Direct visualization of magnetic domains and moirémagnetism in twisted 2D magnets. Science. 2021;374(6571):1140–1144.[165] Li S, Sun Z, McLaughlin NJ, et al. Observation of stacking engineered magnetic phasetransitions within moiré supercells of twisted van der waals magnets. Nat Commun.2024;15:5712.[166] Fisher P, Zappacosta A, Fuhrmann J, et al. High-resolution nanoscale ac quantum sens-ing in cmos compatible sic. Nano Lett. 2025;25(30):11626–11631. PMID: 40693560.[167] Tahara K, i Tamura S, Toyama H, et al. Quantum sensing with duplex qubits of siliconvacancy centers in SiC at room temperature. npj Quantum Inf. 2025;11:58.[168] Bejarano M, Goncalves FJT, Hache T, et al. Parametric magnon transduction to spinqubits. Sci Adv. 2024;10(12):eadi2042.[169] Ohshima T, Satoh T, Kraus H, et al. Creation of silicon vacancy in silicon carbide byproton beam writing toward quantum sensing applications. J Phys D. 2018;51:333002.39[170] Kraus H, Simin D, Kasper C, et al. Three-dimensional proton beam writing of opticallyactive coherent vacancy spins in silicon carbide. Nano Lett. 2017;17(5):2865–2870.[171] Chen YC, Salter PS, Niethammer M, et al. Laser writing of scalable single color centersin silicon carbide. Nano Lett. 2019;19(4):2377–2383.[172] Yamazaki Y, Masuyama Y, Kojima K, et al. Highly sensitive temperature sensing usingthe silicon vacancy in silicon carbide by simultaneously resonated optically detectedmagnetic resonance. Phys Rev Appl. 2023;20:L031001.[173] Sekiguchi N, Fushimi M, Yoshimura A, et al. Diamond quantum magnetometer with dcsensitivity of sub-10 pT Hz −1/2 toward measurement of biomagnetic field. Phys RevAppl. 2024;21:064010.[174] Hoang TM, Ishiwata H, Masuyama Y, et al. Thermometric quantum sensor using excitedstate of silicon vacancy centers in 4H-SiC devices. Appl Phys Lett. 2021;118(4):044001.[175] Scheller D, Hrunski F, Schwarberg J, et al. Quantum-enhanced electric field mappingwithin semiconductor devices. Phys Rev Appl. 2025;24:014036.[176] Hadden JP, Harrison JP, Stanley-Clarke AC, et al. Strongly enhanced photon collectionfrom diamond defect centers under microfabricated integrated solid immersion lenses.Appl Phys Lett. 2010;97(24):241901.[177] Aharonovich I, Neu E. Diamond nanophotonics. Adv Opt Mater. 2014;2:911–928.[178] Schröder T, Mouradian SL, Zheng J, et al. Quantum nanophotonics in diamond. J OptSoc Am B. 2016;33(4):B65–B83.[179] Bekker C, Arshad MJ, Cilibrizzi P, et al. Scalable fabrication of hemispherical solidimmersion lenses in silicon carbide through grayscale hard-mask lithography. Appl PhysLett. 2023;122(17):173507.[180] Hao ZH, Zhou JY, Li Q, et al. Photon collection enhancement of shallow single spindefects in silicon carbide. ACS Photo. 2024;11(11):4725–4732.[181] Reinhardt D, Heupel J, Popov C, et al. Laser beam induced charge collection for defectmapping and spin state readout in diamond. Adv Quantum Technol. 2024;7:2470035.[182] Shang Z, Hashemi A, Berencén Y, et al. Local vibrational modes of Si vacancy spinqubits in SiC. Phys Rev B. 2020;101:144109.[183] Udvarhelyi P, Thiering GmH, Morioka N, et al. Vibronic states and their effect on thetemperature and strain dependence of silicon-vacancy qubits in 4H-SiC. Phys Rev Appl.2020;13:054017.[184] Gottscholl A, Diez M, Soltamov V, et al. Room temperature coherent control of spindefects in hexagonal boron nitride. Sci Adv. 2021;7(14):eabf3630.[185] Gong R, Du X, Janzen E, et al. Isotope engineering for spin defects in van der waalsmaterials. Nat Commun. 2024;15:104.[186] Fan JW, Guo SW, Lin C, et al. Quantum coherence control at temperatures up to 1400k. Nano Lett. 2024;24(46):14806–14811. PMID: 39529435.[187] Castelletto S, Johnson BC, Ivády V, et al. A silicon carbide room-temperature single-photon source. Nat Mater. 2014;13(2):151–156.[188] Karsthof R, Bathen ME, Galeckas A, et al. Conversion pathways of primary defects byannealing in proton-irradiated n-type 4H-SiC. Phys Rev B. 2020;102:184111.[189] Suzuki T, Yamazaki Y, Taniguchi T, et al. Spin property improvement of boron va-cancy defect in hexagonal boron nitride by thermal treatment. Appl Phys Express. 2023;16(3):032006.[190] Schreck M, Gsell S, Brescia R, et al. Ion bombardment induced buried lateral growth:the key mechanism for the synthesis of single crystal diamond wafers. Sci Rep. 2017;7.[191] Tallaire A, Collins A, Charles D, et al. Characterisation of high-quality thick single-crystal diamond grown by CVD with a low nitrogen addition. Dia Rel Mater. 2006;15(10):1700.[192] Coherent C. SILICONCARBIDE(SIC)SUBSTRATES ; 2023. Available from:https://www.coherent.com/resources/datasheet/materials/sic-substrates-ds.pdf.[193] Omote K. Crystal defects in SiC wafers and a new X-ray topography system; 2013.Available from: https://api.semanticscholar.org/CorpusID:30917668.40[194] Li J, Yang G, Liu X, et al. Dislocations in 4h silicon carbide. J Phys D: Appl Phys. 2022;55(46):463001.[195] Kranert C, Wimmer P, Drouin A, et al. Comparative study of methods for counting ofdislocations in 4H-SiC. Mater Sci Semicond Process. 2024;170:107948.[196] Singh H, Hollberg MA, Ghezellou M, et al. Characterization of single shallow silicon-vacancy centers in 4H−SiC. Phys Rev B. 2023;107:134117.[197] Gu H, Nakamura Y, Sasaki K, et al. Multi-frequency composite pulse sequences forsensitivity enhancement in hexagonal boron nitride quantum sensor. Appl Phys Express.2023;16(5):055003.[198] Zhou F, Jiang Z, Liang H, et al. Dc magnetic field sensitivity optimization of spin defectsin hexagonal boron nitride. Nano Lett. 2023;23(13):6209–6215. PMID: 37364230.[199] Patrickson CJ, Baber S, Gaál BB, et al. High frequency magnetometry with an ensembleof spin qubits in hexagonal boron nitride. npj Quantum Inf. 2024;10:5.[200] Shields BJ, Unterreithmeier QP, de Leon NP, et al. Efficient readout of a single spinstate in diamond via spin-to-charge conversion. Phys Rev Lett. 2015;114:136402.[201] Lovchinsky I, Sushkov AO, Urbach E, et al. Nuclear magnetic resonance detection andspectroscopy of single proteins using quantum logic. Science. 2016;351(6275):836–841.[202] Fang K, Acosta VM, Santori C, et al. High-sensitivity magnetometry based on quantumbeats in diamond nitrogen-vacancy centers. Phys Rev Lett. 2013;110:130802.[203] Hopper DA, Grote RR, Exarhos AL, et al. Near-infrared-assisted charge control andspin readout of the nitrogen-vacancy center in diamond. Phys Rev B. 2016;94:241201.[204] Jaskula JC, Shields B, Bauch E, et al. Improved quantum sensing with a single solid-statespin via spin-to-charge conversion. Phys Rev Appl. 2019;11:064003.[205] Neumann P, Beck J, Steiner M, et al. Single-shot readout of a single nuclear spin. Science.2010;329(5991):542–544.411) Initialization 2) Sensing 3) Readout0 or 1 withthe probability; or Laser AOMAPDPiezo stage Voltage sourceTransimpedanceAmp.Lock-in Amp.Pin hole~50 µm・Diamond・SiC・ hBNDicroic MirrorMirrorOptical detectionElectrical DetecitonI/VbcInitializaiton・LaserCoherent Manipulation・Magentic fieldali ittnt Maniipuullaattiioonnalizaaiittonalizaaitonalizzaaitooonnneerr Defectin SpinaReadout・PhotonObjective lens 50x NA 0.7Spot size ~ 0.5 µmElectrodes and MW antenna200 µmFigure 1. a Operation of a solid-state quantum sensor, encompassing the initialization of the spin, coherentmanipulation and the readout of a defect spin. b Typical quantum sensing protocol. c Schematic illustrationsof confocal laser microscopy and scanning photocurrent microscopy, demonstrating quantum sensing usingoptical and photoelectrical detection, respectively. The upper-right panel shows a photograph of the U-shapedmicrowave antenna for spin manipulation and two opposing electrodes for photoelectrical detection, separatedby a gap of ∼ 10 µm. Scale bar: 200 µm.Table 1. Comparison of detection methods of ODMR, EDMR, and PDMR for quantum spin in solid.Method Spin-dependent pro-cessTypical excitation Key features Practical constraintsODMR Optical contrast Optical + MW Optical observables (e.g., photo-luminescence, absorption, polar-ization)Optical access and signal col-lection requiredEDMR Electrical responses Electrical and/oroptical + MWElectrical observables (e.g., cur-rent, photocurrent, capacitance,noise)Observable and implementa-tion depend on material plat-form and device structurePDMR Photoionization Optical + MW Photocurrent-based electricalreadoutRequires optical excitation andcharge collection structures42O₁O₂= (17.84 ns)-1(15.8 ns)-1(15.9 ns)-1(12.5 ns) -1(76.9 ns) -1(463 ns)-1(289 ns)-1= (56.75 ns)-1= (130.59 ns)-1= (41.02 ns)-1= (250.72 ns)-1MS₁MS₂= (1035.35 ns)-1ππResonant excitation Power (mW) (=    ) (ns)-16 5928.7310 4377.8515 2170.8020 1481.69Laser PL3A23E1A1abFigure 2. a Transition diagram of NV−in diamond with transition rates. The solid and dashed down arrowsrepresent the radiative and non-radiative transitions, respectively (see Ref. [4]). b (Top) The transition diagramof VSi (V2) with lifetimes (transition rates) measured at 5.5 K. (Bottom) the lifetime of (γ′3’and (γ′4’ as afunction of the resonant excitation power (details are discussed in Ref. [43]). Figure and Table were adaptedand modified from Ref. [43], CC BY 4.0 [https://creativecommons.org/licenses/by/4.0/)].433A23E1A12A22EaCharge-statetrasnsitionNV – an =  1.1 × 10-3 W-1ns-1an = 1.1 × 10-3 W-1ns-11.3 an4.4 × 10-3 an0.08 an0.037 an0.037 an(15.8 ns)-1(15.9 ns)-1 (20.0 ns)-1 (463 ns)-1 (289 ns)-1 (76.9 ns) -1(12.5 ns) -1NV0Valance bandConduction band e–h+NV– to NV0  in the dark: (100 µs)-1NV0 to NV–  in the dark:  (37 µs)-1MW Frequency (MHz)1.02840 2860 2880 29000.80.60.40.20.0ΔQ (fC)Init. Readoutπb Laser LaserMW10"10# 10$Laser Power Density (W/cm#)0.05200 ns300 ns600 ns1 µs5 µs50 µs0.00-0.05-0.10-0.15cPDMR ContrastFigure 3. a Laser excitation and relaxation rates under the 532-nm laser illumination with the charge-statetransition between NV−and NV0. b PDMR spectrum of the NV center exhibits a positive contrast with thepulse sequence at the top. c PDMR contrast as a function of the laser power density with six different integrationtimes under laser illumination. a, b, and c Adapted with permission and modified from: Reprinted figure withpermission from [H. Morishita , N. Morioka, T. Nishikawa, H. Yao, S. Onoda, H. Abe, T. Ohshima, and N.Mizuochi. Spin-dependent dynamics of photocarrier generation in electrically detected nitrogen-vacancy-basedquantum sensing. Phys. Rev. Appl. 19, 034061 (2023).] Copyright (2023) by the American Physical Society.44Figure 4. PDMR mechanism of the negatively-charged silicon vacancy.Figure 5. Photoelectrical detection of single silicon vacancies [92]. a Fluorescence and b photocurrent scan-ning images of the same area in a 4H-SiC PDMR device. Defect A is identified as a single V2 center and defectB is assumed to be a single V1 center. Spot X is an unidentified defect only observable using photoelectricalimaging. c Magnetic-field sweep PDMR spectra of a V2 center at a fixed RF frequency of 199 MHz. The insetbar represents a signal contrast of 1 %. d Schematic representation of a back-to-back Schottky diode underflat-band bias, facilitating concurrent efficient leakage blocking and photocurrent collection. e Comparison ofthe SNR between ODMR and PDMR for Defect A. The dashed black line indicates the shot-noise-limitedSNR for ODMR. f PDMR-based Ramsey interferometry of defect A, demonstrating hyperfine coupling to twofourth-nearest-neighbor 29Si nuclear spins (upper panel: simulation, lower panel: experimental data). All panelswere adapted from Ref. [92] under CC BY 4.0 [https://creativecommons.org/licenses/by/4.0/]; panels c and ewere replotted for this review.45Table 2. Comparison of various material properties of color center in diamond, SiC, and hBN.Materials NV center in Diamond VSi in SiC VB in hBNZero-Phonon Line fraction(Debye-Waller factor) (%)at low temperature ∼ 3 [39] 6 - 9 [182,183]Fluorescencecount rate (kcps) ∼ 250 [12] ∼ 10 [40]T2 at room temperatureSingle in Natural Abundance (µs) ∼ 650 [60] > 200 [40]Single in isotope pure (ms) ≈ 2.4 [12]Ensemble in natural abundance (µs) ≈ 700 [62] ∼ 50 [66] 2 [184]Ensemble in isotope pure (µs) ≈ 700 [62] ∼ 100 [42] 0.1 - 0.2 [185]Thermal Stability of defects (◦C) ∼ 1130 [186] ∼ 600 [42,187,188] ∼ 500 [189]Maturity of crystal growth8 inch (commercialized)Wafer size 92 mm diameter [190] 12 inch (development)Dislocation Density (cm−2) 103 - 105 [190,191] 103 − 104 [192–195]10Rocking Curve FWHM (arcsec) 5 - 300 [190] Reaching the theoretical limit [193]2 - 6Rabi Contrast (%) ∼ 37 [12] (including ODMR Contrast) [40,92,196] 0.5 - 3 [184,197]Shot-noise limited sensitivitySingle DC (T/√Hz) ∼ 6 × 10−9 [12]Ensemble DC (T/√Hz) 460 × 10−15 [13] 4 × 10−9 [41] 2.87 ×10−6 [198]Single AC (T/√Hz) ∼ 9.1 × 10−9 [12] 358 × 10−6 [166]Ensemble AC (T/√Hz) 210 × 10−15 [13] (16 - 25) × 10−9 [41] 1 ×10−6 [199]Table 3. Comparison of optical and electrical detection methods of NV centers in diamond and VSi in SiC. Readout-fidelity values(σ−1R ) are presented following the formalism of Barry et al. [2]. The column“Evaluation Basis” indicates whether each value wasobtained from reported σ−1R or derived from experimentally measured SNRs (σ−1REx).Material DetectionMethodReadout Protocol Single/Ensemble σ−1R σ−1REx EvaluationBasisReferenceNV center Optical detec-tionODMR Single 9.4 ×10−2 ReportedσR = 10.6[200]NV center Optical detec-tionSpin-to-charge con-versionSingle 3.6 ×10−1 ReportedσR = 2.76[200]NV center Optical detec-tionODMR Single 2.9 ×10−2 ReportedσR = 35[201]NV center Optical detec-tionODMR Single 1.3 ×10−2 ReportedσR = 80[202]NV center Optical detec-tionODMR Single 2.1 ×10−2 ReportedσR = 48[203]NV center Optical detec-tionODMR Single 1.9 ×10−2 ReportedσR = 54[204]NV center Optical detec-tionAncilla-assistedrepetitive readoutSingle 9.1 ×10−1 ReportedσR = 1.1[205]NV center Optical detec-tionODMR Ensemble 1 ×10−3 ReportedσR ∼1000[141]NV center PhotoelectricaldetectionPDMR Ensemble 1.3 ×10−4 Contrast +detectionrate[85]NV center PhotoelectricaldetectionPDMR Ensemble 2 ×10−2 Contrast +detectionrate[86]NV center Optical detec-tionODMR Single 1.6 ×10−3 MeasuredSNR[130]NV center PhotoelectricaldetectionPDMR Single 1.5 ×10−3 MeasuredSNR[130]VSi Optical detec-tionODMR Single 8.0 ×10−4 MeasuredSNR[92]VSi PhotoelectricaldetectionPDMR Single 1.3 ×10−3 MeasuredSNR[92]46