Scalable integration of silicon carbide color centers into nanophotonic structures

Abstract

Color centers in silicon carbide are a promising platform for quantum technologies, offering long-coherence electron and nuclear spins for storing and manipulating quantum information, as well as single-photon emission for quantum communication. Low photon collection efficiency is commonly addressed by integrating color centers into nanophotonic devices. Here, we report a scalable method for the deterministic integration of color centers into silicon carbide nanopillars, using the same nanoscale mask for both implantation of oxygen-related color centers (PL5, PL6) and subsequent nanopillar fabrication via dry etching. This approach solves the issue of low color center creation yield in nanostructures, with figures of merit comparable to bulk samples. We demonstrate count rate enhancements of up to one order of magnitude for PL5 and four times for PL6, while preserving spin coherence times. Our method is directly applicable to other solid-state platforms, offering a scalable route to efficiently integrate color centers into nanostructures.

The paper addresses how to integrate single spin-active color centers into SiC nanopillars without separate alignment of implantation sites and photonic structures. A single patterned PMMA mask defines both ion-implantation windows and, through nickel deposition and lift-off, the subsequent dry-etch mask. This registers implanted regions to pillars, but defect formation and lateral placement within each pillar remain probabilistic. PL5 and PL6 centers are characterized at room temperature in bulk SiC, conventional pillars, and pillars on 30–40 $\mu$m membranes with backside collection. Oxygen implantation gives pillar-site yields of 31.6% for PL5 and 22.7% for PL6, comparable to bulk-site yields of 19.2% and 28.5%, respectively, and substantially above carbon or nitrogen implantation. ODMR identifies defect types, while photon autocorrelation verifies single-emitter behavior with $g^{(2)}(0)<0.5$. Mean photon-count enhancements relative to bulk are $4.1\pm1.1$ and $2.1\pm0.7$ for PL5 and PL6 in conventional pillars, increasing to $6.0\pm2.0$ and $2.8\pm0.7$ in membrane pillars; individual enhancements exceed tenfold for PL5 and approach fourfold for PL6. Hahn-echo coherence times average $18.8\pm6.8\,\mu$s in conventional pillars and $12.2\pm5.6\,\mu$s in membrane pillars, compared with $21.5\pm5.7\,\mu$s for oxygen-implanted bulk defects. CW-ODMR parameters imply a best magnetic sensitivity of $140\,\mathrm{nT}/\sqrt{\mathrm{Hz}}$, approximately threefold better than bulk. Measured optical enhancements fall below FDTD predictions; the paper attributes this to averaging over pillar diameters, emitter-position variation, and fabrication inhomogeneity. Applicability to other hosts and photonic geometries is proposed rather than demonstrated.