Skip to content

Discovery of ST2 centers in natural and CVD diamond

FieldValue
Journalnpj Quantum Information
Published2026-03-16
AuthorsJonas Foglszinger, Andrej Denisenko, G. V. Astakhov, Lev Kazak, Petr Siyushev
DOI10.1038/s41534-025-01116-8

ST2 centers retained significant ODMR contrast at 30 mT across nearly the entire 4π unit sphere. At 30 mT, the NV center fell below 3% contrast when the field tilted more than 10° from its symmetry axis. Carbon implantation and annealing created the centers in CVD diamond, but the yield remained about 2 × 10⁻⁸ centers per implanted ion. Natural-diamond ST2 centers showed no photoinstability during continuous excitation over several weeks.

Researchers first found ST2 in a natural diamond of unknown origin and treatment history. Identification used a 446 nm zero-phonon line, then confirmed each defect by continuous-wave ODMR. The confirmed protocol implanted 10 keV ¹²C ions into the (100) plane of CVD diamond. The dose was 2.5 × 10¹⁵ ions/cm², slightly above the graphitization threshold. Annealing at 1200 °C for 1 h followed implantation.

Centers were absent immediately after implantation and appeared only after annealing at 1100 °C or above. Production peaked near 1300 °C, with about 35 centers in a 10 μm × 10 μm area. Surface graphitization above 1300 °C complicated yield assessment, and the 1400 °C comparison used limited data. A 1600 °C trial used higher implantation energy, so it did not isolate the annealing effect.

The conversion rate was about 2 × 10⁻⁸ centers per implanted ion at 10 keV. At that energy, the estimated local density limit was 6 × 10⁴ centers per cubic micron. This equals roughly 120 centers in a 10 μm × 10 μm field. Optimization reached roughly 50 centers in 100 μm², about 50% of the estimated maximum.

Incremental oxygen-plasma etching showed that the ST2 depth profile followed the total vacancy profile. An SRIM-TRIM simulation reproduced that profile, while the implanted-carbon profile did not. Carbon-only implantation and the vacancy correlation make an intrinsic lattice defect plausible. The authors could not determine whether the defect contains vacancies, interstitial sites, or both. The chemical structure of ST2 remains unknown.

All optical experiments used linearly polarized 410 nm excitation in a home-built confocal microscope. Emission was collected from 450 to 525 nm with a single-photon detector. The 446 nm zero-phonon line provided the primary spectral identification. ODMR provided final confirmation through two resonances at 495.3 MHz and 2267.5 MHz. Microwaves were delivered through integrated Au contacts.

QuantityResultBasis or condition
Excited-state lifetime1.4 nsLow-power antibunching
Zero-field splittingD = 1636.6 MHz; E = 896.6 MHzDeduced from ODMR
CPT state lifetimeAt least 10 μsTwo long-lived states
Triplet lifetimesτₓ = 27 μs; τᵧ = 34 μs; τ_z = 2.6 μsPopulation-transfer decay

The antibunching dip fell below 0.5, confirming single-photon emission from the defect. Optical saturation exposed a bunching shoulder associated with a metastable triplet. An external magnetic field changed saturated brightness by up to twofold.

The data support one singlet ground state, one singlet excited state, and a metastable triplet. Continuous-wave ODMR resolved transitions between triplet sublevels because their lifetimes differ. Coherent population trapping confirmed two long-lived states.

The missing third ODMR resonance remains partly unresolved. Population imbalance may cancel the Tₓ-Tᵧ signal, although the cited simulation instead yielded γₓ = γᵧ.

A 10 × 10 × 10 mm³ NdFeB magnet supplied a 1.4 T magnetization perpendicular to the sample surface. Stepper motors varied the field direction while fluorescence and ODMR were recorded. A spin-1 Hamiltonian with g = 2 reproduced the measured magnetic maps. The model also used the measured transition rates and level structure.

The ST2 z-axis followed the diamond [111] direction. The y-axis lay in the plane formed by two adjacent σ-bonds. The orientation analysis predicted twelve inequivalent ST2 orientations arranged into four triples. Each triple combined one σ-bond direction with the lattice’s threefold rotational symmetry. The twelve magnetic orientations aligned with six optical dipole directions, with two magnetic orientations per dipole.

This geometry indicates low defect symmetry. The absence of electric-field shifts supports inversion symmetry over one-mirror-plane symmetry.

Simulations at 30 mT predicted significant ODMR contrast over nearly the full 4π unit sphere. This makes ST2 suitable for strong fields whose directions cannot be aligned with a known sensor axis. The shot-noise calculation used 5% ODMR contrast, 10 kcps fluorescence, and a 0.65 MHz linewidth. It also used an ODMR resonance shift of 28 GHz/T.

ST2 loses magnetic sensitivity when the field-induced shift is comparable to or below the zero-field splitting. The large E value produces clock-like transitions with no linear magnetic-field dependence in that regime. The NV center has the opposite limitation because very weak fields cause little state mixing. ST2 complements the NV center rather than replacing it across the full field range.

TR12 has a similar level structure and broad acceptance angle, but insufficient photostability. ST1 and TR12 gave similar sensitivity results, while ST2 appeared more photostable. The authors describe that photostability advantage as qualitative and call for detailed confirmation.

No ODMR frequency shift was detected at electric fields up to 2 × 10⁶ V/m. This result supports the proposed inversion symmetry and predicts no first-order electric sensitivity.

Temperature measurements covered 40 °C to 60 °C with less than 0.1 K control precision. Each temperature was held for at least 5 min before measurement. The text reports ODMR frequency shifts of 9 kHz/K and −30 kHz/K. The Figure 3 caption instead gives D and E shifts of 10.5 kHz/K and −19.5 kHz/K. The paper does not present one consistent calibration for these two parameters.

The authors describe these shifts as two to three times lower than the reported NV shift of −74 kHz/K. They therefore see ST2 as a local thermometry option, not the preferred primary temperature sensor.

The low production yield makes only single-defect uses currently viable. Statistical variation and implantation damage also complicate deterministic device placement. The unknown atomic structure limits defect engineering, even though the electronic level structure is well constrained.

Current work is transferring NV-style pillar and cantilever approaches to ST2. The authors report encouraging preliminary results but do not provide a completed device demonstration. Proposed uses include imaging internal magnetic textures such as skyrmions with randomly oriented fields. Scaling production is the main barrier and may also advance electrical readout and electroluminescence studies.

ST2 is therefore a specialized option for strong, arbitrarily oriented fields rather than a general replacement for NV centers.

The experiments require a (100)-oriented CVD diamond host, carbon implantation and annealing, and Au-contact metallization. The 6CCVD diamond materials catalog lists MPCVD-grown single-crystal diamond and Au metallization, covering the host and contact materials. Questions about preparing and integrating these materials can be directed through engineering support.

View Original Abstract

Abstract The ST2 center is an optically addressable point defect in diamond that facilitates spin initialization and readout at room temperature. However, while this study presents the discovery of photostable ST2 centers first observed in a natural diamond and provides a reliable technique for artificially creating them, its chemical structure remains unknown. To assess the potential of ST2, we map out its basic optical characteristics, reveal its electronic level structure, and quantify the intrinsic transition rates. Furthermore, we investigate its response to microwaves, static magnetic fields, and the polarization of excitation laser light, revealing twelve inequivalent orientations of the ST2 center. Simultaneous exposure to microwaves and static magnetic fields also reveals an exceptionally wide acceptance angle for sensing strong magnetic fields, unlike the well-established NV center, which is sensitive only within a narrow cone aligned with its symmetry axis. This finding establishes the ST2 center as a highly promising candidate for nanoscale quantum sensing.