Neutral Silicon Vacancy Centers in Undoped Diamond via Surface Control

Neutral silicon vacancy centers (${\mathrm{SiV}}^0$) in diamond are promising candidates for quantum applications; however, stabilizing ${\mathrm{SiV}}^0$ requires high-purity, boron-doped diamond, which is not a readily available material. Here, we demonstrate an alternative approach via chemical control of the diamond surface. We use low-damage chemical processing and annealing in a hydrogen environment to realize reversible and highly stable charge state tuning in undoped diamond. The resulting ${\mathrm{SiV}}^0$ centers display optically detected magnetic resonance and bulklike optical properties. Controlling the charge state tuning via surface termination offers a route for scalable technologies based on ${\mathrm{SiV}}^0$ centers, as well as charge state engineering of other defects.

Figure 1

Charge state tuning of SiV centers via surface termination. (a) Energy band diagram of diamond under different surface terminations. Under O termination, the Fermi level is dominated by the nitrogen donors (with ionization energy of 1.7 eV) in the bulk. Under H termination, upward band bending occurs, and the Fermi level near the diamond surface is pinned to the valence band maximum ($E_V$). $E_C$ and $E_F$ represent the positions of the conduction band minimum and Fermi level, respectively, of diamond. The ${\mathrm{SiV}}^{+/0}$ charge transition level ($E_V+0.25\mathrm{eV}$) is shown as the blue curve, and the ${\mathrm{SiV}}^{0/-}$ charge transition level ($E_V+1.43\mathrm{eV}$) is shown as the red curve. (b) Scanning electron microscope (SEM) image of the PRs hosting the SiV centers (80° tilt). Inset: high-resolution SEM of PRs. (c) Characteristic emission spectrum of SiV centers in a representative PR under different terminations. ${\mathrm{SiV}}^0$ spectra were taken using 10 mW 857 nm excitation at 70 K. ${\mathrm{SiV}}^{-}$ spectra were taken with 0.1 mW 561 nm excitation at room temperature.

Figure 2

Characterization of the surface termination process. (a) Schematic illustration of the surface termination process. The O-terminated surface is prepared using tri-acid cleaning. The H-terminated surface is prepared by annealing the sample in forming gas at $800°\mathrm{C}$ for 72 h or in hydrogen at $750°\mathrm{C}$ for 6 h. (b) XPS of the oxygen $1s$ peak in a test sample after the different processes. Hydrogen cannot be detected using XPS. Therefore, the absence of the oxygen $1s$ peak and any peaks other than the carbon $1s$ peak indicates the presence of H termination. We measure XPS of the sample immediately after the processes to minimize contribution from contamination on the surface. (c) AFM scans of the surface of a test diamond before and after the annealing. The AFM scans were filtered to remove instrumental noise (Fig. S3 [28]). The surface morphology is unaffected by the annealing process.

Figure 3

Statistics of SiV emission under different surface terminations. (a) Left: ${\mathrm{SiV}}^0$ emission intensity distribution under H and O termination measured at 10 K with 11 mW 857 nm excitation. Right: ${\mathrm{SiV}}^{-}$ emission intensity distribution under H and O termination measured in air with 0.1 mW 561 nm excitation. The intensities are extracted from Lorentzian fits of zero-phonon line intensities. Intensity for PRs without observable SiV emission is set to zero. (b) Occurrence (ratio of the number of PRs with SiV emission above the background to total number of PRs) of ${\mathrm{SiV}}^0$ and ${\mathrm{SiV}}^{-}$ centers under different surface terminations. Orange denotes O termination, while blue denotes H termination. Inset: the order of surface termination under study. (c) Average intensity of ${\mathrm{SiV}}^0$ and ${\mathrm{SiV}}^{-}$ centers under different surface terminations. ${\mathrm{SiV}}^0$ emission is suppressed under O termination, while ${\mathrm{SiV}}^{-}$ emission is suppressed under H termination. The PRs where no emission was identified are excluded from the analysis. ${\mathrm{SiV}}^{-}$ centers were measured using 1.88 mW 561 nm excitation for the first O-terminated surface, while the power was kept at 0.1 mW for the rest of the measurements. The intensities are scaled by power.

Figure 4

PLE and ODMR of ${\mathrm{SiV}}^0$ under H termination. (a) PLE of the bound exciton transition on ${\mathrm{SiV}}^0$ in three H-terminated PRs (orange) and a bulk-doped sample (blue). The excitation power is $\sim 10\mathrm{mW}$. (b) PLE of the zero-phonon line transition of ${\mathrm{SiV}}^0$ in three H-terminated PRs (orange) and a bulk-doped sample (blue). The PLE curves are background subtracted. The excitation power is $\sim 0.3\mathrm{mW}$. The curves in (a) and (b) are offset from each other for clarity. (c) Continuous-wave ODMR of ${\mathrm{SiV}}^0$ centers in a PR measured by exciting at 946.45 nm at 10 K. The orange curve shows a Lorentzian fit to the data. (d) Wavelength dependence of ODMR contrast (orange curve). The blue curve shows the PLE spectrum for comparison. ODMR was measured after H termination and three piranha cleaning steps with the excitation power set to $\sim 25\mathrm{\mu }\mathrm{W}$.