Atomic structure of the PL5 defect in silicon carbide revealed by single-spin spectroscopy and oxygen implantation

TL;DR

The paper resolves the atomic structure of PL5/PL6 centers in 4H-SiC by testing competing models with single-defect imaging, ODMR spectroscopy, and ab initio calculations. It rules out the stacking-fault–driven VV-SF model and provides strong evidence that PL5 is the neutral oxygen-substituted silicon vacancy OC_V_Si in the kh configuration, aided by a dramatic yield increase under oxygen implantation. The work delivers a clear defect identity and a scalable method for high-yield generation of these centers, enabling high-sensitivity quantum sensing in two and three dimensions.

Abstract

PL5 and PL6 centers in 4H-SiC are promising for quantum applications due to their superior charge stability and optically detected magnetic resonance (ODMR) properties at room temperature. However, their atomic structures remain unresolved, with ongoing controversy regarding their potential association with stacking faults. Previous measurements relying on spin ensemble detection were insufficient to draw definitive conclusions. In this work, we conduct correlative imaging of stacking faults and PL5/PL6 at the single-defect level, definitively ruling out any spatial correlation and demonstrating that these centers are not associated with stacking faults. Furthermore, we find that substituting oxygen for nitrogen in ion implantation enhances the yields of PL5 and PL6 by more than $11$-fold and $23$-fold, respectively. Single-spin ODMR spectroscopy of PL5 reveals six distinct orientations, determines the transverse zero-field splitting parameter $E$, and characterizes the hyperfine coupling. Combined with our ab initio calculations, these results provide compelling evidence for the assignment of PL5 as an OV($kh$) defect, consisting of an oxygen atom occupying the C($k$) site as the nearest neighbor to a Si($h$) vacancy. The structural analysis together with the demonstrated defect yield enhancement lays the foundation for fabricating high-sensitivity, high-contrast ensemble quantum sensors in two and three dimensions.

Figures (6)

• Figure 1: (a) Schematic illustration of the VV-SF model. (b) Predicted distribution of PL5--6 color centers based on the VV-SF model. (c) ODMR spectra of PL5 and PL6 measured in the sample. (d) Confocal images of the stacking fault edge region: the top image displays the integrated photoluminescence intensity from 410-430 nm under 325 nm laser excitation, while the bottom image shows the photoluminescence intensity of PL5 and PL6 under 914 nm laser excitation in the same area. Circles denote the locations of PL5, and squares indicate the locations of PL6.
• Figure 2: (a) Photoluminescence (PL) image of a nitrogen-implanted sample. Implantation was performed at 15 keV with a dose of $10^{11}\,\text{cm}^{-2}$, followed by annealing at $1050\,^\circ\mathrm{C}$ for 30 minutes. (b) PL image of an oxygen-implanted sample under identical implantation and annealing conditions. In both panels, PL5 and PL6 centers identified by ODMR (see Supplementary Material SupplementalMaterial) are marked with blue squares and red circles, respectively.
• Figure 3: (a) Schematic of six PL5 directions D1--D6 in 4H-SiC lattice. (b) Schematic of six PL5 directions, coordinate axes, and the magnetic field for measurements in (d-e). The magnitude of the magnetic field was set to 25 Gauss. (c) ODMR spectrum of a D1 direction PL5 under $\phi=30^{\circ}$. (d)(e) Polar plots of ODMR splittings of PL5 of different orientations as functions of the magnetic orientation $\phi$.
• Figure 4: (a) Schematic of PL5, coordinate axes and the magnetic field. The magnitude of the magnetic field was set to 25 Gauss. The coordinate axes are defined such that the direction of axially symmetric splitting is selected as the z-axis. The coordinate system is chosen such that the c-axis lies within the xz plane. (b) Polar plot of the measured detuning as a function of the magnetic orientation $\varphi_{B}$. The detuning is set from $f_0= 1375.3\ \text{MHz}$. Fitting using \ref{['eq:omega_pm']} yields $\varphi_E = 182.2\pm 3.5 ^{\circ}$.
• Figure 5: (a) ODMR spectra of PL5 coupled to a $^{13}\text{C}_{\text{I}}$ nuclear spin adjacent to the $\text{V}_{\text{Si}}$ in PL5. Blue(Red) line represents the fit for the data of PL5 coupled with $^{13}\text{C}_{\text{Ia}} (^{13}\text{C}_{\text{Ib}})$. All spectra were acquired under an applied magnetic field of 185 Gauss. (b) Hyperfine splitting $\delta$ induced by the interaction with $^{13}\text{C}_{\text{Ia}}$(blue) and $^{13}\text{C}_{\text{Ib}}$(red), as derived from different PL5. (c) The splitting difference $\delta(\text{C}_{\text{Ia}})-\delta(\text{C}_{\text{Ib}})$ for PL3--5. Experimental data for PL3 and PL4 are reproduced from sonDivacancy4HSiC2006.