Quantum Silicon Carbide Microscope

• Quantum SiC Microscope (QSiCM) is a silicon-carbide–based quantum imaging tool that uses optically addressable color centers for local field sensing.
• It integrates wide-field magnetometry, operando electric-field mapping, and noise spectroscopy to achieve nanoscale sensitivity.
• Leveraging engineered SiC defects and advanced photonic architectures, QSiCM enables scalable, multiplexed quantum imaging under ambient conditions.

Quantum Silicon Carbide Microscope (QSiCM) denotes a class of silicon-carbide-based quantum imaging and sensing instruments in which optically addressable color centers in SiC serve as local probes, photonic pixels, or both. In its realized form, QSiCM includes a wide-field quantum magnetometer built on a 4H-SiC wafer for imaging current-induced magnetic fields, while the broader concept also encompasses operando electric-field microscopes based on single silicon vacancies in SiC diodes and integrated nanophotonic or scanning architectures that use SiC defects for nanoscale sensing and readout (Suhana et al., 18 Sep 2025, Scheller et al., 2024).

1. Defect physics and material basis

The material foundation of QSiCM is predominantly 4H-SiC, a wide-bandgap semiconductor with bandgap $\approx 3.2\ \text{eV}$, refractive index $n \approx 2.6$ at near-IR wavelengths, mature wafer-scale growth and processing, and biocompatibility. These properties support both stable color centers and nanophotonic confinement, which is why 4H-SiC recurs across QSiCM proposals and demonstrations as host, sensor substrate, and photonic medium (Radulaski et al., 2016).

The principal sensing defects are silicon vacancies. For the 4H-SiC V2 center, the ground-state electronic spin is $S=3/2$ with zero-field splitting $D \approx 70\ \text{MHz}$, and the zero-phonon line lies near $918$–$916\ \text{nm}$. Under continuous $730\ \text{nm}$ excitation, spin-dependent transitions via a metastable state produce optical spin polarization, and spin readout proceeds by optically detected magnetic resonance (ODMR). Previous work on single $V_\text{Si}$ in 4H-SiC reports room-temperature spin coherence times $T_2 \gtrsim 160\ \mu\text{s}$, while later surveys cite millisecond $T_2$ under dynamical decoupling for $n \approx 2.6$0 and up to $n \approx 2.6$1 for divacancy-based nuclear-memory registers. The same materials platform also hosts neutral divacancies, NV-like defects, and transition-metal centers such as Cr$n \approx 2.6$2, V$n \approx 2.6$3, Mo$n \approx 2.6$4, W$n \approx 2.6$5, and Er$n \approx 2.6$6, spanning ZPLs from the near-IR to telecom-adjacent bands and offering different Debye–Waller factors, lifetimes, and spin manifolds (Majety et al., 2021).

Within the silicon-vacancy family, the V1 center in 4H-SiC is notable for two sharp ZPLs at $n \approx 2.6$7 and $n \approx 2.6$8, a spin-quartet ground state with $n \approx 2.6$9, and a measured Debye–Waller factor $S=3/2$0 at $S=3/2$1. Under resonant V1 excitation, ODMR contrast reaches $S=3/2$2, and dynamical decoupling extends ensemble coherence to $S=3/2$3. These properties single out V1 as a strong spin–photon interface candidate, whereas V2 is more established for room-temperature sensing and imaging (Nagy et al., 2017).

A distinct room-temperature sensing branch uses PL5 centers in 4H-SiC. PL5 is a spin-1 defect with zero-field splitting $S=3/2$4, transverse strain term $S=3/2$5, and a giant Stark effect with transverse Stark coefficient $S=3/2$6 and longitudinal coefficient $S=3/2$7. In practice, this makes PL5 a broadband probe of local electric and magnetic noise and extends the QSiCM concept from field imaging to defect-resolved noise microscopy (Liu et al., 27 Dec 2025).

A major microscopic controversy concerned whether the room-temperature Si-vacancy qubits in 4H-SiC are isolated negatively charged silicon vacancies or vacancy complexes. High-precision first-principles calculations together with high-resolution electron spin resonance identified the relevant centers as isolated negatively charged silicon vacancies, with V1 assigned to the $S=3/2$8 site and V2 to the $S=3/2$9 site in 4H-SiC (Ivády et al., 2017).

2. Device architectures and photonic embodiments

One architectural line realizes QSiCM sensing pixels as SiC nanopillars containing single silicon vacancies. In 4H-SiC, a high-purity semi-insulating wafer is irradiated with $D \approx 70\ \text{MHz}$0 electrons at fluence $D \approx 70\ \text{MHz}$1, then patterned into $D \approx 70\ \text{MHz}$2 tall nanopillars with diameters from $D \approx 70\ \text{MHz}$3 to $D \approx 70\ \text{MHz}$4. Arrays are fabricated as $D \approx 70\ \text{MHz}$5 blocks with $D \approx 70\ \text{MHz}$6 pitch, and finite-difference time-domain simulations show $D \approx 70\ \text{MHz}$7–$D \approx 70\ \text{MHz}$8 orders of magnitude improvement in collection efficiency relative to bulk SiC, with optimal diameters near $D \approx 70\ \text{MHz}$9 and $918$0. Experimentally, single defects in $918$1 pillars reach saturation count rates up to $918$2, about $918$3–$918$4 brighter than bulk single $918$5, while preserving antibunching and spin polarization. Approximately $918$6 out of $918$7 pillars in a tested array host single, well-placed $918$8 centers with enhanced emission, and about $918$9 of bright single emitters show ODMR signals clearly (Radulaski et al., 2016).

A second line uses suspended 4H-SiC membranes. A monolithic process based on photoelectrochemical undercutting yields suspended, undoped $916\ \text{nm}$0 membranes that are then patterned into 1D photonic crystal cavities, fiber-coupled photonic crystal cavities with tapered waveguide interfaces, and lithium-niobate-on-4H-SiC acoustic cavities. In this platform, free-space 1D photonic crystal cavities show measured quality factors of a few thousand, fiber-coupled versions show $916\ \text{nm}$1 in reflection, and the membrane-first process supports high-temperature annealing and heterogeneous integration. This provides a direct route to QSiCM architectures that combine optical readout, phononic control, and defect integration in the same suspended layer (Xie et al., 14 Aug 2025).

Triangular-cross-section nanodevices form a third photonic architecture. In 4H-SiC, angle-etched triangular waveguides and nanobeam photonic crystal cavities were modeled with finite-difference eigensolver and finite-difference time-domain methods. For a representative triangular nanobeam cavity, the fundamental mode lies near $916\ \text{nm}$2 with simulated $916\ \text{nm}$3 and mode volume $916\ \text{nm}$4, implying $916\ \text{nm}$5 for ideal overlap. The same framework extends to photonic crystal molecules, where inter-cavity coupling $916\ \text{nm}$6 supports collective polariton and subradiant modes. This suggests cavity-enhanced QSiCM nodes in bulk SiC, especially for telecom-compatible defects (Majety et al., 2020).

Inverse-designed Fabry–Perot cavities in 4H-SiC-on-insulator provide another route. These are straight 4H-SiC waveguides terminated by inverse-designed reflectors, fabricated in $916\ \text{nm}$7 SiCOI. Across 65 devices and 1069 resonances, the measured waveguide propagation loss is $916\ \text{nm}$8, and the average reflector reflectivity is $916\ \text{nm}$9. The measured integrated dispersion parameters include $730\ \text{nm}$0 and $730\ \text{nm}$1, enabling spontaneous four-wave mixing, optical parametric oscillation at threshold $730\ \text{nm}$2, and simultaneous visible-frequency generation via phase-matched SHG and SFG. For QSiCM, these cavities are relevant less as sensors themselves than as on-chip sources, converters, and routing elements for quantum illumination and multiplexed readout (Yang et al., 2023).

Earlier 3C-SiC microdisk resonators established that whispering-gallery cavities in SiC can operate in the visible and near-IR with measured $730\ \text{nm}$3 up to $730\ \text{nm}$4 and mode volume $730\ \text{nm}$5. Although these devices were aimed at ensemble impurities in 3C-SiC, they represent an early photonic cavity lineage relevant to cavity-enhanced SiC microscopy and sensing (Radulaski et al., 2014).

3. Readout physics, control protocols, and metrological observables

QSiCM implementations use a common toolbox of single-photon characterization, resonant or off-resonant optical pumping, ODMR, and phase-sensitive detection. In silicon-vacancy nanopillars, the second-order correlation function is measured in a Hanbury–Brown–Twiss geometry, with background-corrected $730\ \text{nm}$6 confirming antibunching through $730\ \text{nm}$7. Continuous-wave ODMR is driven by RF delivered through a copper wire across the sample; for V2, the central resonance near $730\ \text{nm}$8 identifies the $730\ \text{nm}$9 ground state, and the room-temperature ODMR contrast is in the range $V_\text{Si}$0–$V_\text{Si}$1 (Radulaski et al., 2016).

A more selective control modality is the double radio-optical resonance demonstrated for silicon vacancies in 6H-SiC. There, a tunable diode laser with linewidth below $V_\text{Si}$2 addresses a particular zero-phonon line while X-band ESR at $V_\text{Si}$3 addresses the spin transition. The resonance condition is

$V_\text{Si}$4

with extracted zero-field splittings $V_\text{Si}$5 for $V_\text{Si}$6 and $V_\text{Si}$7 for $V_\text{Si}$8. Optical pumping creates a non-Boltzmann spin population, the ESR amplitude follows

$V_\text{Si}$9

with $T_2 \gtrsim 160\ \mu\text{s}$0, and pulsed recovery measurements yield $T_2 \gtrsim 160\ \mu\text{s}$1 at low temperature. This protocol is relevant to QSiCM because it shows how spectral addressability and spin selectivity can be combined in SiC for site-specific sensing (Riedel et al., 2012).

For electric-field microscopy, the key observable is the Stark shift of optical or spin transitions. In an operando 4H-SiC pin-diode containing single VSi centers in the intrinsic region, the A1 optical transition shifts as

$T_2 \gtrsim 160\ \mu\text{s}$2

with measured values $T_2 \gtrsim 160\ \mu\text{s}$3 and $T_2 \gtrsim 160\ \mu\text{s}$4 for a representative center. The ground-state Stark-shifted spin Hamiltonian is

$T_2 \gtrsim 160\ \mu\text{s}$5

and the Lorentz local-field correction gives

$T_2 \gtrsim 160\ \mu\text{s}$6

which is approximately $T_2 \gtrsim 160\ \mu\text{s}$7 in 4H-SiC. These relations allow QSiCM-style devices to infer local electric field, depletion width, and carrier distributions from either PLE or ODMR (Scheller et al., 2024).

For broadband noise microscopy, PL5 centers extend the readout space. ODMR peak tracking yields time-series $T_2 \gtrsim 160\ \mu\text{s}$8 with random telegraph noise, while dynamical decoupling extends noise spectroscopy from near-DC to MHz frequencies and $T_2 \gtrsim 160\ \mu\text{s}$9 relaxometry probes MHz-GHz noise. The spin Hamiltonian

$T_2$0

makes the defect simultaneously sensitive to electric and magnetic noise channels, enabling a QSiCM to function as a nanoscale noise spectrometer as well as a field imager (Liu et al., 27 Dec 2025).

The wide-field QSiCM realization adds two system-level protocols. For ODMR magnetometry with silicon vacancies in 4H-SiC, the main spin transition frequencies under a bias field along the $T_2$1 axis are

$T_2$2

A dual-frequency sensing protocol drives both transitions to enhance magnetic sensitivity while suppressing strain- and temperature-induced common-mode shifts. A microwave-free protocol instead uses the ground-state level anticrossing at $T_2$3, where PL changes sharply with magnetic field and can be read out without microwave delivery (Suhana et al., 18 Sep 2025).

4. Demonstrated imaging and sensing modalities

QSiCM has moved from a proposed architecture to an experimentally realized instrument. The reported wide-field implementation images magnetic fields generated by electrical currents using a 4H-SiC wafer containing V2 centers. It employs a dual-frequency sensing protocol, a lock-in camera, and a $T_2$4 excitation beam over a $T_2$5 spot, and demonstrates spatial imaging of current-induced magnetic fields with a field of view of $T_2$6 virtual pixels, temporal resolution of $T_2$7, spatial resolution of $T_2$8, and sensitivity of about $T_2$9 per pixel. The same platform also performs microwave-free imaging near GSLAC-2 at $n \approx 2.6$00 and quantitatively reconstructs the magnetic field of a wire carrying electrical current (Suhana et al., 18 Sep 2025).

Single-defect operando electric-field microscopy has also been demonstrated in 4H-SiC power-device structures. In a pin-diode with VSi centers at known positions inside the intrinsic region, PLE Stark-shift measurements reveal the expansion of the depletion zone and permit extraction of the intrinsic doping concentration as

$n \approx 2.6$01

in agreement with capacitance–voltage measurements giving $n \approx 2.6$02. From time-series measurements at the point of maximum PLE slope, the electric-field sensitivity reaches

$n \approx 2.6$03

The same measurements additionally infer local free-carrier concentrations from linewidth changes associated with the presence or absence of mobile charge carriers (Scheller et al., 2024).

Modality	Observable	Reported result
Wide-field magnetic imaging	Current-induced magnetic field	$n \approx 2.6$04 virtual pixels; $n \approx 2.6$05; $n \approx 2.6$06; about $n \approx 2.6$07 per pixel
Operando electric-field mapping	Local static electric field, depletion width, dopants	$n \approx 2.6$08; $n \approx 2.6$09
Room-temperature noise microscopy	Single-charge tunneling and broadband noise	Real-time single-charge tunneling; near-DC to MHz DD spectroscopy; MHz-GHz $n \approx 2.6$10 relaxometry

Another branch focuses on the local electronic environment rather than direct field imaging. In a 4H-SiC quantum optoelectronic p–i–n device, individual V2 centers undergo charge-state switching under applied bias, with the bright $n \approx 2.6$11 state emitting in the $n \approx 2.6$12–$n \approx 2.6$13 band and the neutral state acting as a dark state in the experimental window. The switching bias depends on emitter depth, and under specific optical excitation conditions the photon rate is enhanced near the switching bias. The reported saturation-count parameters for one emitter change from $n \approx 2.6$14 and $n \approx 2.6$15 at $n \approx 2.6$16 to $n \approx 2.6$17 and $n \approx 2.6$18 at $n \approx 2.6$19. This device-level behavior turns single color centers into local probes of quasi-Fermi-level position, depletion-region structure, and defect-assisted carrier cycling (Widmann et al., 2019).

The PL5 work generalizes the microscope concept further by showing room-temperature broadband quantum noise spectroscopy of nanoscale charge defects in commercial 4H-SiC. It reports the first real-time, nanoscale observation of single-charge tunneling dynamics in a commercial semiconductor at room temperature, electrical noise imaging showing distinct noise variations across different wafer substrates, dynamical decoupling spectroscopy from near-DC to MHz frequencies, and $n \approx 2.6$20 relaxometry that identifies the origin of MHz-GHz noise through a nanoscale electron paramagnetic resonance fingerprint of charge defects in SiC (Liu et al., 27 Dec 2025).

5. Quantum photonics, nonlinear optics, and quantum-imaging extensions

Beyond direct sensing, QSiCM is embedded in the larger program of integrated SiC quantum photonics. Reviews of the field identify three application domains—quantum networking, simulation, and computing—and emphasize that SiC hosts color centers with optical interfacing, long coherence times, spin–photon and spin–spin entanglement, and scalability on quantum-grade wafers. Within this framework, a QSiCM pixel can be interpreted as a defect–photonics node: a local quantum sensor whose fluorescence or resonance is routed through waveguides, resonators, or detector-integrated circuits. The same literature cites divacancy gate fidelities of $n \approx 2.6$21 for single-qubit operations, CNOT fidelity of $n \approx 2.6$22, entangled-state fidelity of $n \approx 2.6$23, and spin-photon entanglement with silicon vacancy centers at Hong–Ou–Mandel visibility $n \approx 2.6$24, all of which indicate that microscopy-grade readout and network-grade photonic interfacing can coexist in SiC (Majety et al., 2021).

The integrated-photonics perspective also clarifies what remains technically difficult. In quantum-grade 4H-SiCOI, intrinsic absorption has been measured as low as $n \approx 2.6$25, microring resonators reach $n \approx 2.6$26, and photonic crystal cavities reach $n \approx 2.6$27 without emitters. Yet defect-integrated quantum devices still lag behind classical structures in $n \approx 2.6$28 and loss. The gap matters directly for QSiCM because cavity-enhanced readout, narrow-linewidth spectroscopy, and large-scale multiplexing all depend on closing it (Lukin et al., 2020).

Nonlinear photonics extends QSiCM toward quantum illumination and frequency-converted imaging. In inverse-designed 4H-SiC Fabry–Perot cavities, spontaneous four-wave mixing yields a quantum comb with coincidence-to-accidental ratio 110 and 275 for different mode pairs and pair generation rates of $n \approx 2.6$29 and $n \approx 2.6$30. The same devices realize optical parametric oscillation across the C-band and, through SHG and SFG, generate visible comb light around $n \approx 2.6$31. A plausible implication is that QSiCM could combine SiC defect sensing with on-chip generation of correlated or frequency-converted light, using telecom pumps while collecting or displaying signals in the visible (Yang et al., 2023).

Quantum imaging methods developed for semiconductor inspection point in the same direction. A nonlinear interferometer based on nondegenerate SPDC probes a silicon chip at $n \approx 2.6$32 while detection occurs at $n \approx 2.6$33, with the signal intensity

$n \approx 2.6$34

and the phase-scanned variance satisfying

$n \approx 2.6$35

This suggests that future QSiCM implementations could pair SiC quantum sensors with indirect IR microscopy, allowing buried-structure imaging at an SiC-compatible probe wavelength while retaining visible-range detection hardware (Paterova et al., 2020).

6. Constraints, controversies, and forward development

The central limitations are now clear. In nanopillar arrays, the emitter yield is still modest, with only about $n \approx 2.6$36 of pillars hosting single, well-placed $n \approx 2.6$37 centers and only about $n \approx 2.6$38 of bright single emitters showing clear ODMR. Room-temperature ODMR contrast in that architecture is $n \approx 2.6$39–$n \approx 2.6$40, which is adequate for sensing but not yet optimal for high-throughput microscopy. Collection efficiency depends sensitively on emitter depth and radial position, and fabrication imperfections reduce experimental brightness below simulated limits (Radulaski et al., 2016).

At the integrated-photonics level, low Debye–Waller factors for many SiC defects, non-radiative recombination, spectral diffusion, inhomogeneous broadening, and charge-state control remain persistent bottlenecks. Reviews emphasize that defect-integrated quantum devices still trail classical SiC photonics by orders of magnitude in achievable $n \approx 2.6$41, and that surface charges, strain from fabrication, and implantation damage complicate both coherent optics and uniform device scaling. These are not minor engineering details: they determine whether a QSiCM behaves as a calibrated quantum imager or as an ensemble of inhomogeneous local probes (Lukin et al., 2020, Majety et al., 2021).

Materials and fabrication developments address several of these points. The suspended 4H-SiC membrane platform specifically targets robust thin-film processing, high-temperature annealing compatibility, and heterogeneous material integration; this is important because many promising transition-metal emitters and charge-control structures require aggressive post-processing that is hard to reconcile with conventional bonded thin films. At the instrument level, the demonstrated wide-field QSiCM already identifies two near-term sensitivity levers: isotopically purified SiC and improved light collection in crystallographically optimized wafer orientations (Xie et al., 14 Aug 2025, Suhana et al., 18 Sep 2025).

The forward path therefore has a relatively coherent structure. Controlled defect placement, wafer-scale cavity and waveguide integration, on-chip RF or microwave delivery, and multiplexed photonic readout are the enabling technologies for dense QSiCM arrays. Wide-field magnetic imaging, operando electric-field mapping, and nanoscale charge-noise spectroscopy already establish that SiC can host quantum microscopes that operate in technologically relevant conditions, including room temperature and working electronic devices. What remains is to unify these capabilities—spin sensing, charge-state control, low-loss photonics, and scalable nanofabrication—into instruments in which each SiC defect or photonic node functions as a calibrated quantum pixel rather than an isolated proof-of-principle device.