Electrical Tuning of Single Nitrogen-Vacancy Center Optical Transitions Enhanced by Photoinduced Fields

We demonstrate precise control over the zero-phonon optical transition energies of individual nitrogen-vacancy (NV) centers in diamond by applying multiaxis electric fields, via the dc Stark effect. The Stark shifts display surprising asymmetries that we attribute to an enhancement and rectification of the local electric field by photoionized charge traps in the diamond. Using this effect, we tune the excited-state orbitals of strained NV centers to degeneracy and vary the resulting degenerate optical transition frequency by $>10\mathrm{GHz}$, a scale comparable to the inhomogeneous frequency distribution. This technique will facilitate the integration of NV-center spins within photonic networks.

Figure 1

(a) Simplified energy-level diagram (not to scale) showing only the $|S_z⟩$ levels of the NV-center ground and excited states, with resonant excitation (red arrows) and redshifted emission (gray arrow) marked. (b) PLE spectrum (points) with no applied bias, marked by dashed line in (d), with a two-Lorentzian fit (solid curve). (c) Micrograph and photoluminescence image (center) of device $A$, with electrical connections marked. (d) PLE spectra of the $6\mathrm{\text{−}}\mu \mathrm{m}$-deep NV center marked by an arrow in (c) as a function of lateral bias applied symmetrically to the $X$ (left panel) or $Y$ (right panel) gate pairs, with $V_{\mathrm{dc}}=0\mathrm{V}$. In all PLE spectra, the origin of the relative frequency axis is arbitrary.

Figure 2

(a) Stark-shift hysteresis loop for a NV center $13\mu \mathrm{m}$ below a transparent top gate, as a function of top-gate voltage. Points mark the transition frequencies from a two-Lorentzian fit to a PLE spectrum at the corresponding voltage, and color-coded arrows indicate the sweep direction. (b) Schematic of the local electric fields in a lateral geometry. Photoionized charge traps in the illuminated volume contribute a rectified field $\mathbf{R}$ predominantly in the $+Z$ direction, which adds to the dielectric field $\mathbf{E}$ to shift the direction of the local field $\mathbf{F}$. (c) dc Stark components $\overline{\nu }$ and $\delta$ (points) extracted from fits of the PLE spectra in Fig. 1d, with a combined fit according to the model described in the text (solid curves). (d) dc Stark components (points) and fits (solid curves) measured both with (green squares) and without (orange circles) the 532 nm repump excitation. Inset: Micrograph and photoluminescence image of device $C$, with electrical connections marked. The $7\mathrm{\text{−}}\mu \mathrm{m}$-deep NV center measured in (d) is circled. In all cases, marker sizes slightly exceed measurement uncertainties.

Figure 3

Tuning diagram for the NV center marked in Fig. 1c. Degeneracy is achieved at different frequencies by setting the applied biases ($V_{\mathrm{dc}},V_X,V_Y$) as marked in the upper panel; in the lower panel we show PLE spectra as a function of $V_X$ around each of these points for fixed $V_{\mathrm{dc}}$ ($V_Y$ is also varied to keep the ratio $V_X/V_Y$ constant). We can shift a second NV center to degeneracy within the tuning range of the first, as shown by the PLE spectra outlined in blue (lower panel) and the corresponding blue line (upper panel) marking the degenerate frequency.