Photoluminescence excitation and spectral hole burning spectroscopy of silicon vacancy centers in diamond

Silicon-vacancy (SiV) centers in diamond are promising systems for quantum information applications due to their bright single-photon emission and optically accessible spin states. Furthermore, SiV centers in low-strain diamond are insensitive to perturbations of the dielectric environment; i.e., they show very weak spectral diffusion. This property renders ensembles of SiV centers interesting for sensing applications. We here report on photoluminescence excitation (PLE) spectroscopy on an SiV ensemble in a low strain, chemical vapor deposition-grown high-quality diamond layer, where we measure the fine structure with high resolution and obtain the line widths and splittings of the SiV centers. We investigate the temperature dependence of the width and position of the fine structure peaks. Our measurements reveal line widths of about 10 GHz as compared to a lifetime limited width on the order of 0.1 GHz. This difference arises from the inhomogeneous broadening of the transitions caused by residual strain. To overcome inhomogeneous broadening we use spectral hole burning spectroscopy, which enables us to measure a nearly lifetime limited homogeneous line width of 279 MHz. Furthermore, we demonstrate evidence of coherent interaction in the system by driving a $\mathrm{\Lambda }$ scheme. Additional measurements on single emitters created by ion implantation confirm the homogeneous line widths seen in the spectral hole burning experiments and relate the ground-state splitting to the decoherence rate.

Figure 1

(a) Structure of the SiV center, consisting of a split vacancy (transparent) and an interstitial silicon atom (yellow) oriented along the [111] direction. Next-neighbor carbon atoms are shown in green. (b) The SiV ensemble sample is a commercial HPHT diamond overgrown with a thin homoepitaxial film in a CVD process. SiV centers are located in the thin film. (c) Electronic structure of the SiV center with split ground and excited states. The four possible ZPL transitions are denoted as A–D.

Figure 2

PLE spectrum of the SiV ensemble at 5K. It is dominated by four strong peaks with individual line widths of about 10 GHz arising from the electronic structure of the SiV center. Additional peaks are attributed to subensembles created by different silicon isotopes.

Figure 3

Temperature-dependent PLE spectrum of the SiV ensemble. The four-line fine structure only develops below 60 K. The peak positions are blue-shifted at lower temperatures. Oscillations in the spectra at higher temperatures are caused by residual laser power fluctuations in one mode-hop-free scanning range of 30 GHz. All spectra have been corrected for laser-power fluctuations throughout the scans.

Figure 4

(a) Temperature-dependent line width of the SiV ensemble. At 130 K, peaks B and C merge and the line width is measured using the spectrometer. Data from peak B is fitted using several dependencies on T. (b) Temperature-dependent line shift analysis is performed using peak C.

Figure 5

Spectral hole burning signal of transition C and corresponding level scheme. Probe laser frequency is shown in respect to the detuning from the pump-laser frequency. The inset shows the power-broadened spectral hole with a line width of 346 MHz. Data points are fitted using a bi-Lorentzian function.

Figure 6

(a) Autocorrelation measurement of an emitter in the SIL array with a $g^{\left(2\right)}\left(0\right)$ of 0.29. The data is fitted with the model for a three-level system. (b) Spectral hole burning measurement on the same emitter. When shining in the pump laser, the peak vanishes, indicating the absence of inhomogeneous broadening.

Figure 7

Power broadened PLE spectrum of a single SiV center in the SIL array. We extrapolate zero excitation power line widths of 270 resp. 575 MHz, thereby showing line widths close to the lifetime limit. We tentatively attribute the double peak structure of transition D to the presence of a second emitter in close spatial and spectral proximity.

Figure 8

Line width of transition C as a function of the ground-state splitting for 12 single SiV centers. The solid data points are fitted with the relation given in the text. The data point labeled with the open circle is omitted from the fit (see text). The data point labeled with the red dot represents a strained single emitter for which a CPT measurement was performed (see Fig. 10).

Figure 9

(a) $\mathrm{\Lambda }$ scheme addressed by the two lasers on transitions C and D. (b) Fluorescence intensity of the SiV ensemble as a function of the two-laser-detuning. A dip is visible when both lasers are in resonance with the respective transition.

Figure 10

CPT measurement for a single strained emitter (data point labeled with red dot in Fig. 8). A $\mathrm{\Lambda }$ scheme between transitions C and D is addressed. The CPT dip that is observed for emitters with smaller ground state splitting [14] completely vanishes due to the increased decoherence rate.