Scalable, nanoscale positioning of highly coherent color centers in prefabricated diamond nanostructures

Nanophotonic devices in color center-containing hosts provide efficient readout, control, and entanglement of the embedded emitters. Yet control over color center formation - in number, position, and coherence - in nanophotonic devices remains a challenge to scalability. Here, we report a controlled creation of highly coherent diamond nitrogen-vacancy (NV) centers with nanoscale three-dimensional localization in prefabricated nanostructures with high yield. Combining nitrogen $δ$-doping during chemical vapor deposition diamond growth and localized electron irradiation, we form shallow NVs registered to the center of diamond nanopillars with wide tunability over NV number. We report positioning precision of ~ 4 nm in depth and 46(1) nm laterally in pillars (102(2) nm in bulk diamond). We reliably form single NV centers with long spin coherence times (average $T_2^{Hahn}$ = 98 $μs$) and 1.8x higher average photoluminescence compared to NV centers randomly positioned in pillars. We achieve a 3x improved yield of NV centers with single electron-spin sensitivity over conventional implantation-based methods. Our high-yield defect creation method will enable scalable production of solid-state defect sensors and processors.

💡 Research Summary

The authors present a scalable method for creating highly coherent nitrogen‑vacancy (NV) centers in pre‑fabricated diamond nanostructures with nanometer‑scale three‑dimensional positioning. The approach combines nitrogen δ‑doping during chemical vapor deposition (CVD) growth with localized 200 keV electron‑beam irradiation (referred to as δ‑electron irradiation) to generate vacancies that later pair with the doped nitrogen atoms to form NV centers.

First, a 53 nm‑deep layer of ^15N is incorporated into a high‑purity CVD diamond. Nanopillars of 280 nm and 480 nm diameter, ~1 µm tall, are defined by electron‑beam lithography and inductively coupled plasma reactive‑ion etching. Using a commercial 200 keV electron‑beam lithography system, a 20 nm‑diameter electron spot is precisely aligned to the centre of each pillar. The electron beam creates a narrow column of carbon monovacancies extending up to ~50 µm depth; the vacancy density is tuned by varying the electron dose from 1.6 × 10¹⁹ to 4.8 × 10²¹ e⁻ cm⁻².

Subsequent annealing at 850 °C for 11 min in vacuum allows monovacancies to diffuse and become captured by the δ‑doped nitrogen atoms, forming NV centers. By adjusting the dose, the average number of NVs per pillar can be continuously varied from ≈0.05 to ≈10. Monte‑Carlo simulations of a simple diffusion‑capture model yield a vacancy diffusion coefficient D_V of 17–21 nm² s⁻¹, consistent with literature values.

Lateral positioning precision is quantified in two ways. In unpatterned mesas, the standard deviation of NV positions relative to the intended spot is σ_loc ≈ 102 nm, dominated by vacancy diffusion during annealing. In the pillars, the photonic mode distorts direct imaging, so simulations are used; the effective lateral spread is σ_pillar_loc = 46 nm for 280 nm pillars and 72 nm for 480 nm pillars—significantly tighter than the mesa value. The improvement is attributed to vacancy capture at the pillar sidewalls, which limits lateral diffusion. Depth confinement is ~4 nm, as measured by the δ‑doped layer thickness.

Spin coherence is evaluated using Hahn‑echo measurements on single NVs formed with a dose of 1.6 × 10²⁰ e⁻ cm⁻². The average T₂^Hahn is 98 µs (σ = 37 µs), comparable to the limit set by the surrounding P1 nitrogen bath and markedly better than the <50 µs typical of 15 keV ion‑implanted NVs. Higher doses (>4.8 × 10²⁰ e⁻ cm⁻²) lead to a modest reduction in T₂, likely due to vacancy clustering. Optical readout contrast (spin‑dependent PL) reaches 18 % (C_Rabi), and the saturation count rates are 0.79 Mcps (280 nm pillars) and 1.06 Mcps (480 nm pillars). Compared with non‑irradiated pillars containing randomly distributed NVs, the localized NVs exhibit a 1.8‑fold increase in PL brightness for the larger pillars, confirming the benefit of positioning NVs at the optical mode maximum. Finite‑difference time‑domain (FDTD) simulations corroborate that lateral localization enhances collection efficiency.

By selecting appropriate doses (≈1.6 × 10²⁰ e⁻ cm⁻² for 280 nm pillars and ≈3.0 × 10²⁰ e⁻ cm⁻² for 480 nm pillars), the authors achieve an average of one NV per pillar with a three‑fold higher yield of single‑spin‑sensitive magnetometers compared with conventional ion‑implantation methods. The entire workflow—CVD growth, standard lithography, electron‑beam irradiation, vacuum annealing, and acid cleaning—relies on commercially available tools, making it amenable to high‑throughput manufacturing.

In summary, the paper demonstrates a high‑yield, low‑damage, and precisely controllable technique for embedding NV centers within diamond nanophotonic structures. The method simultaneously delivers nanometer‑scale lateral and depth positioning, long spin coherence, and enhanced optical performance, thereby addressing a critical bottleneck for scalable quantum sensing and quantum information processing platforms based on solid‑state spin defects.