Embedding Center Localization in Silicon Photonics

• Embedding center localization is a deterministic method for nanoscale positioning of single quantum emitters in nanophotonic devices, ensuring optimal emitter-cavity coupling.
• The process integrates fluorescence localization and advanced nanofabrication, achieving <20 nm alignment tolerance and >10× Purcell enhancement.
• This technique overcomes random emitter distribution challenges, yielding a 15 nm accuracy that boosts photoluminescence by 30× for scalable quantum photonic applications.

Embedding center localization refers to the deterministic, nanoscale positioning of optically active defect centers—specifically single G centers in silicon—within nanophotonic structures such as optical cavities. This process is fundamental for scalable integration of quantum light sources, as it enables in situ enhancement of emission properties by optimally coupling quantum emitters to optical cavities. The technique overcomes challenges associated with the stochastic distribution of emitters and the resultant low probability of their placement at regions of maximal field intensity within nanophotonic devices. Fluorescence-localization techniques (FLT) are a central methodology for achieving this nanometer-scale accuracy, integrating confocal microscopy with robust coordinate transfer for subsequent cavity fabrication (Ma et al., 15 Mar 2025).

1. Theoretical Basis of Localization Precision

The localization of single G centers in silicon leverages the diffraction-limited imaging capabilities of confocal microscopy. The microscope’s point-spread function (PSF) in the focal plane is approximated by a circular Gaussian: $$I(x,y)=I_b + I_0\exp\left(-\frac{(x-x_0)^2+(y-y_0)^2}{2s^2}\right)$$ where $I_b$ is the background, $I_0$ the emitter peak intensity, $(x_0,y_0)$ the center coordinates, and $s \approx 0.21\,\lambda/\mathrm{NA}$ the Gaussian width parameter. Practical fitting employs a two-dimensional Lorentzian model to accommodate PSF deviations due to long-wavelength tails. Localization precision is characterized via the Cramér–Rao lower bound: $$\sigma_{\rm loc} \approx \sqrt{\frac{s^2}{N} + \frac{a^2}{12\,N} + \frac{8\pi s^4 b^2}{a^2 N^2}}$$ where $N$ is the total collected photon count from the emitter, $a$ is the pixel size, and $b$ the per-pixel background. For negligible background ($b \ll I_0$), this reduces to $\sigma_{\mathrm{loc}} \approx s/\sqrt{N}$. This framework directly links localization accuracy to photon statistics and imaging parameters, indicating improvements with increased photon count and decreased pixel size (Ma et al., 15 Mar 2025).

2. Experimental Workflow: Fluorescence-Localization Technique (FLT)

The FLT workflow comprises a sequence of nanofabrication and optical characterization stages:

1. Sample Preparation: Starting from an SOI wafer (220 nm top Si, 3 µm oxide), $^{12}$C ions are implanted (30 keV, dose $5\times10^{13}\,\mathrm{cm}^{-2}$), followed by a rapid thermal anneal at 1000 °C for 20 s to create isolated G centers.
2. Alignment Marker Fabrication: Positive PMMA resist is spin-coated and patterned via photolithography to define fiducial markers (crosses and square frames), followed by 100 nm Au evaporation and lift-off.
3. Low-Temperature Confocal Imaging: Samples are cooled to 7 K. Off-resonant excitation (532 nm CW laser, 100–200 µW) through a ×100, NA=0.85 objective allows for a two-dimensional raster scan (∼50 nm steps) monitored by InGaAs APDs, specifically detecting G-center emission at 1278 nm.
4. Coordinate Extraction: Local PL maxima are fit with a 2D Lorentzian function:

$$L(x,y) = A\left[1 + \left(\frac{x-x_0}{w_x}\right)^2 + \left(\frac{y-y_0}{w_y}\right)^2\right]^{-1} + B$$

yielding the positions $(x_0, y_0)$ relative to gold markers. Localization dispersion across >90 centers is $\sigma_{x,y}\approx15$ nm.

1. E-beam Lithography Transfer: The determined coordinates are transformed into e-beam write-field positions, with sub-20 nm overlay alignment tolerance to the gold markers (Ma et al., 15 Mar 2025).

3. Nanophotonic Cavity Fabrication and Alignment

Deterministic embedding of the localized G center is realized by in situ fabrication of a circular Bragg grating (CBG) cavity:

• Design Parameters: The CBG cavity comprises a central disk (diameter 1100 nm), grating period 470 nm, ring width 125 nm, and five concentric rings in the 220 nm Si layer, as determined by FDTD optimization.
• Nanofabrication: A negative-tone resist (e.g., HSQ) is spun and patterned via e-beam lithography, aligned to the gold markers. Developed patterns are then transferred into the silicon layer by reactive ion etching (SF$_6$/C$_4$F$_8$); resist stripping exposes the completed CBG.
• Alignment Tolerance: Optical simulations confirm a Purcell factor that falls off substantially outside a central ∼200×100 nm$^2$ elliptical region. The 15 nm-scale localization ensures an observed >10× enhancement in 60% (25/40) of fabricated devices. In contrast, the probability of random overlap is only ≈0.25%, yielding a 240-fold greater success probability using FLT.

Step	Key Precision/Parameter	Observed Value
Localization	$\sigma_{x,y}$	≈15 nm
CBG placement	Overlay error	<20 nm
Cavity enhancement	>10× Purcell in (%)	60% (FLT) vs 0.25% (random)

4. Quantitative Enhancement and Purcell Factor Assessment

Coupling localized G centers to CBG cavities enables quantifiable enhancements in both photoluminescence and emission rate:

• Intensity Enhancement: Saturated PL intensity increases from $I_{sat,\,\rm off}=8.6$ kHz (unpatterned Si) to $I_{sat,\,\rm on}=259.3$ kHz (CBG-coupled), a 30-fold improvement.
• Lifetime Measurements: Time-resolved PL under 5 MHz pulsed excitation yields $\tau_{\rm off}=10.27\pm0.19$ ns (outside cavity), $\tau_{\rm on}=4.13\pm0.01$ ns (within CBG), reflecting a rate acceleration $R=2.5$.
• Purcell Factor Lower Bound: Accounting for nonradiative and phonon-sideband decay, the conservative lower bound is

$F_p > \frac{R-1}{D} + 1$

where $D\approx0.15$ is the Debye–Waller factor. Using the rate acceleration, $F_p > 11$.

5. Statistical Performance and Yield

Quantitative assessment across an ensemble of devices demonstrates the deterministic advantage of embedding center localization:

• Localization Accuracy: Standard deviation $\sigma_{x,y}\approx15$ nm (1σ).
• Yield: 60% of CBG cavities fabricated using FLT demonstrate >10× enhancement, compared to 0.25% for random placement.
• Intensity and Rate Gains: PL intensity displays a 30× enhancement, emission rate a 2.5× increase, and the Purcell factor exceeds 11 as a lower bound (Ma et al., 15 Mar 2025).

6. Implications for Quantum Photonic Integration

The integration of embedding center localization with in situ cavity fabrication enables scalable quantum photonic device production on silicon. The deterministic nature of the FLT workflow supports high-yield placement and repeatable enhancement of single quantum emitters, necessary for networks reliant on indistinguishable photons in the telecommunication band. The capacity to localize and embed individual G centers within photonic nanostructures at the 15 nm scale paves the way for large-scale integration of quantum light sources and could broadly impact the development of silicon-based quantum networks (Ma et al., 15 Mar 2025).