Electrically driven single-photon source at room temperature in diamond

Single-photon sources that provide non-classical light states on demand have a broad range of applications in quantum communication, quantum computing and metrology1. Singlephoton emission has been demonstrated using single atoms2, ions3, molecules4, diamond colour centres5,6 and semiconductor quantum dots7–11. Significant progress in highly efficient8,11 and entangled photons9 sources has recently been shown in semiconductor quantum dots; however, the requirement of cryogenic temperatures due to the necessity to confine carriers is a major obstacle. Here, we show the realization of a stable, room-temperature, electrically driven single-photon source based on a single neutral nitrogen-vacancy centre in a novel diamond diode structure. Remarkably, the generation of electroluminescence follows kinetics fundamentally different from that of photoluminescence with intra-bandgap excitation. This suggests electroluminescence is generated by electron–hole recombination at the defect. Our results prove that functional single defects can be integrated into electronic control structures, which is a crucial step towards elaborate quantum information devices. Single defects in diamonds5, particularly single nitrogen-vacancy (NV) centres6 (Fig. 1a), have been used as single-photon sources for quantum cryptography12 and single-photon interference13 because of their outstanding photostability at room temperature. The NV centre also exhibits excellent spin characteristics, including long coherence time14,15 and fast manipulation rates16, and allows implementation of few-qubit quantum registers17,18. Coherent coupling between spin and photon19 and spin–photon entanglement20 have been reported. However, electrical excitation of the NV centre has not been realized. By demonstrating electroluminescence from a single neutral NV centre we have added an additional central element to the quantum toolbox of diamond defects. This provides new opportunities for integrating single-photon sources based on diamond defects into electronic control circuitry and for spintronic applications for quantum communication and processing. The efficient generation of electroluminescence from diamond single defects requires the synthesis of electron (n-type) and hole (p-type) conducting materials as well as an ultrapure intrinsic (i) layer in a p–i–n diode structure. In the last decade, techniques for diamond semiconductor synthesis including the realization of n-type formation have been developed using microwave plasmaenhanced chemical vapour deposition (CVD)21,22. Diamonds are doped with a large amount of boron and phosphorus to create semiconducting properties21,22; however, this leads to defects that emit electroluminescence in the visible spectral range21–23. We therefore introduced an extremely high-quality undoped region to form a p–i–n diamond diode. In previous work we achieved CVD growth of ultrapure intrinsic diamonds, including an isotopically engineered diamond14,15,18, where the concentration of colour centres was reduced to ≪0.1 ppb (1× 10 cm), well below the upper limit for single-colour-centre microscopy. In the present work, iand phosphorus-doped n-type layers were independently grown by CVD on (001) p-type diamond (see Methods). The doping concentrations of phosphorus and boron were 1× 10 and 1× 10 cm, respectively. The electron and hole Hall mobilities were 150 and 10 cm V s at room temperature, respectively. After growth, round mesa structures were fabricated (Fig. 1b) and nitrogen was subsequently implanted, because the i-layer without implantation proved to be too pure to detect native NV centres (concentration of the NV centre is ≪1× 10 cm). Measurements were carried out on the NV centres in the i-layer near the edge of the mesa structures (Fig. 1b). The p–i–n diamond, showing an ideal diode characteristic with a rectification ratio of 1× 10 at +30 V (Fig. 1c), was investigated for two different surface terminations (hydrogenated and oxidized) of the i-layer24. A homebuilt confocal microscope was used to address single defect centres. All experiments were conducted under ambient conditions. Photoluminescence raster scans of the intrinsic area were recorded first (Fig. 2a). The photoluminescence spectra indicate the presence of NV centres in a neutral charge state (NV) in the hydrogenated i-layer surface (Fig. 2d) and in a negatively charged state (NV) in the oxidized i-layer surface (Fig. 2c). Antibunching measurements were performed on several NV centres, shown in Fig. 2a and in other regions. Figure 3a shows a typical result of NV photoluminescence. The second-order autocorrelation function g(t) at t1⁄4 0 was less than 0.5, indicating that most NV photoluminescence in Fig. 2a originated from single centres. The non-zero value of g(0) is considered to be caused mainly by residual photoresist on the surface. In our analysis, the background signal contribution to g(0) was not subtracted. After current injection, electroluminescence was observed from the same region as that of photoluminescence (Fig. 2b) in areas with hydrogenated surfaces. A comparison of Fig. 2a and b reveals that a significant number of NV centres show photoluminescence and electroluminescence. Note that only NV electroluminescence was observed, even for NV centres that show NV in photoluminescence. The electroluminescence intensity and temporal stability were sufficiently high to record electroluminescence antibunching (Fig. 3b). A value of g(0)1⁄4 0.45