Indirect Overgrowth as a Synthesis Route for Superior Diamond Nano Sensors

Nitrogen-vacancy (NV−) centers in diamond are prominent room-temperature quantum sensors, sensitive to temperature, strain, electric and magnetic fields, with millisecond-scale spin coherence enabling nanoscale metrology. For strong coupling to external spins, NV− centers must be within a few nanometers of the surface, but shallow centers suffer from decoherence and charge instability due to surface-related defects. Increasing depth mitigates surface noise and extends coherence times, creating a trade-off between sensitivity and stability, with an optimal intermediate depth around 10–30 nm. CVD diamond overgrowth on implanted NV− centers can bury them deeper, enhancing coherence while maintaining surface sensitivity. However, substantial losses of NV− fluorescence following overgrowth have been reported, with the prevailing mechanism unclear (etching vs hydrogen passivation to NVH). This study investigates the origin of NV− losses during overgrowth, tests and validates an indirect overgrowth protocol (overgrow before annealing) to mitigate passivation, and targets fabrication of depth-confined NV− nanosensors with enhanced T2 and T2* while preserving near-surface sensing capability. The work also explores how implantation dose and energy influence passivation kinetics and hydrogen diffusion beneath the surface.

Previous work showed that low-energy implantation followed by CVD overgrowth can enhance shallow NV− coherence (Staudacher et al.). Yet, hydrogen plasma necessary for CVD can either etch the surface or passivate NV− centers (Lesik et al.; Stacey et al.). NVH complexes are known in CVD diamond and hydrogen is believed to diffuse tens of microns into intrinsic diamond and can interact with NV−, but direct confirmation of NV−→NVH conversion for shallow layers via EPR is impractical due to small active volumes. Optical methods allow NV− counting in thin layers. Prior studies established depth-dependent decoherence from surface noise and benefits of increased depth, as well as drawbacks of higher implantation energies (broader depth distribution, more damage and paramagnetic defects). Overgrowth offers increased depth without additional straggle. Reports also note possible growth-interface defects that might affect spin properties. Collectively, the literature motivates a controlled overgrowth approach and careful disentangling of etching from hydrogen passivation mechanisms.

Two fabrication sequences were compared: (1) Direct overgrowth: 15N+ implantation into 99.999% 12C-enriched diamond, UHV annealing at 1000 °C for 3 h to form NV−, followed by CVD overgrowth of a nanometer-scale capping layer; (2) Indirect overgrowth: identical but with CVD overgrowth performed before the post-implant annealing. 15N implant energies were 2.5 and 5 keV; implantation doses were 1×10^9 (low), 1×10^11 (medium), and 1×10^12 (high) 15N+/cm^2. CVD growth used a home-built reactor: for buffer layers, 99.999% 12CH4 at 0.2% in H2, 600 sccm H2 flow, 22.5 mbar, 1.2 kW microwave power; for overgrowth steps, 12CH4 concentration reduced to 0.05% in H2. The holder was heated to ~700 °C, samples exposed to H2 plasma for 5 min, then capping layers grown at 900 °C. Process gases were purified. UHV annealing was performed at 1000 °C for 3 h (<1×10^−7 mbar). Confocal microscopy with a 519 nm pulsed laser and oil immersion objective captured NV− fluorescence; pulsed ODMR measured hyperfine splitting to verify 15N-origin NV−. Depths were determined using nano-NMR of 1H spins in immersion oil, with NV electron spins as sensors, and compared to C-TRIM simulations (simulated distributions shifted by capping thickness). NV− yields were quantified by analyzing fluorescence profiles to estimate NV− density per confocal volume and normalizing to implanted dose. Coherence properties (T2 via Hahn-echo, T2* via Ramsey) were measured under aligned magnetic fields; growth rates were extracted from linear increase of average NV depth with overgrowth time, giving NV-calibrated rates. Control experiments included direct overgrowth tests to assess etching/passivation and comparison across energies and doses.

• Direct overgrowth under the implemented conditions caused implanted NV− spots to disappear, indicating losses inconsistent with dominant etching of the diamond surface layer.
• Indirect overgrowth (overgrow before annealing) preserved 15NV− centers for medium and high doses even after 2 h overgrowth; low-dose regions still vanished.
• Depth analysis via 1H nano-NMR showed the entire NV− depth distribution uniformly shifted deeper after overgrowth, without narrowing at shallow depths. This rules out significant hydrogen-induced etching of the first nanometers and supports hydrogen passivation as the main loss mechanism.
• Average NV depth increased linearly with overgrowth time. NV-calibrated growth rates: ~5 nm h−1 (2.5 keV) and ~8 nm h−1 (5 keV); average used ~6.5 nm h−1.
• Overgrowth improved coherence: burying by ~13 nm yielded up to fivefold T2 enhancement for 2.5 keV implants and threefold for 5 keV. Single NVs showed T2 up to ~100 μs; T2* reached ~20 μs. Across samples, T2* remained ~15–20% of T2 before and after overgrowth, indicating no additional spin baths beyond reduced surface influence.
• Despite capping layers up to ~13 nm, NVs retained sensitivity to surface protons (1H), enabling nano-NMR of immersion oil.
• NV− yield after overgrowth and subsequent annealing decreased rapidly with overgrowth duration for all energies and doses, consistent with passivation to NVH. Energy (2.5 vs 5 keV) had minor effect on passivation rate at high dose, but implantation dose had a strong impact: high dose slowed passivation markedly compared to medium dose (yield curves flatten after ~1 h for high dose).
• Comparing yields before and after annealing showed that during early overgrowth (<~30 min), most NVs form during the final anneal, implying NV formation (step I) is slower than passivation (step II). A two-step mechanism is proposed: (I) NV formation from implanted nitrogen and mobile vacancies; (II) rapid hydrogen passivation of formed NVs from the plasma.
• Interpretation: higher implanted nitrogen concentration may reduce hydrogen mobility (via charge-state effects), slowing passivation kinetics and thereby improving NV survival during overgrowth.

The study resolves the origin of NV− losses during CVD overgrowth for shallow implants: rather than surface etching, hydrogen passivation to NVH predominantly depletes NV− centers. By rearranging the sequence (indirect overgrowth), overgrowth occurs while most nitrogen remains unconverted, significantly mitigating passivation. Depth distributions shift uniformly without increased straggle or additional damage, confirming that overgrowth tunes depth independently of implantation energy. Enhanced T2 and T2* result primarily from increased distance to the surface, while preserving sensitivity to surface nuclear spins. Dose-dependent passivation kinetics suggest a role of nitrogen in modulating hydrogen diffusion/charge state beneath the surface, offering a handle to stabilize NVs during growth. These insights enable fabrication of depth-confined, high-coherence NV nanosensors not attainable by higher-energy implantation alone, and inform process windows (dose, overgrowth duration) to balance NV formation and passivation.

Indirect overgrowth—low-energy 15N implantation, CVD overgrowth prior to annealing—provides a reliable route to engineer shallow, depth-confined NV− centers with substantially improved coherence times while maintaining surface sensitivity. Overgrowth increases NV depth without broadening the distribution or adding significant spin noise, yielding T2 up to ~100 μs and T2* ~20 μs for ~13 nm capping layers. NV− losses during overgrowth are attributed to hydrogen passivation rather than etching, with passivation rates strongly reduced at higher implantation doses. The two-step formation–passivation mechanism explains the yield evolution with growth time. This approach advances deterministic NV engineering for quantum sensing and opens avenues to probe NV formation and hydrogen passivation dynamics, guiding future optimization of defect generation and stabilization in diamond.

• NVH formation was not directly measured in the shallow implanted regions (EPR sensitivity insufficient for near-surface layers); passivation is inferred from NV− loss patterns and depth-distribution arguments.
• Growth-rate calibration via NV depth sensing is valid only while NV–1H coupling remains detectable (tens of nanometers) and for sufficiently long T2.
• Direct overgrowth results may be process-specific; hydrogen plasma conditions and temperatures could alter etching/passivation balance.
• Only two implantation energies (2.5, 5 keV) and selected doses were explored; broader parameter sweeps could refine kinetics models.
• The linear T2–depth trend beyond ~8 nm is observed but not fully explained; potential growth-interface contributions were not exhaustively characterized.