1. Introduction

Diamond has a range of readily produced optical centres that can be utilized as single photon sources (SPSs). Diamond based SPSs offer the unique advantage of long-term photo stability at room temperature which has generated significant interest in areas such as quantum information processing (QIP) [1–3]. Among the single photon emitters identified to date [4–9], the nitrogen vacancy (N-V) centre remains the most studied and applicable for QIP. N-V centres have been identified in natural diamonds and can also be created synthetically using the microwave plasma enhanced chemical vapour deposition (MPECVD) technique [10] with single NV centres found to reside in diamond nano-crystals as small as 50 nm in size [11]. An alternative fabrication technique, involving the direct implantation of protons into commercial grade diamond nano powder rather than nitrogen ions [12] or electrons [13], has now realized single NV emitters in 25 nm sized nano-crystals [14]. The ability to readily fabricate these single defect centres has in recent times seen the commercialization of the first SPS based on the N-V defect in diamond [15]. However, fabricating these sources in desired locations and on particular structures remains difficult due to the statistical nature of creating the atomic defect. The demonstration of nano-scale magnetometry [16–19] and the possibility of decoherence imaging using single N-V centres opens up exciting opportunities in quantum imaging (QI) [20]. However, if these types of opportunities are to be explored, the manipulation of single quantum systems and sources is required. The manipulation of nano-particles has previously been reported through the use of an atomic force microscope (AFM) [21] and a scanning confocal microscope/AFM system [16]. The combined technique allows the position and optical characteristics of the emitter to be measured simultaneously provided the focused laser spot from the confocal microscope coincides with the AFM cantilever tip. Such systems have been shown to be an effective tool in terms of attaching single quantum emitters onto the AFM tip. However, the technique is limited to substrates which are transparent and optically thin. Scanning near field optical microscopes (SNOMs) have delivered similar outcomes whereby the nano-particle is attached to a tapered optical probe containing a thin layer of glue or PMMA [22] or by scanning the probe close to the substrate with the aim to attract the particle to the tip [23]. However, the yield and ability to select a desired nano-particle remains low and difficult at this point in time. Nevertheless, the manipulation of nano-crystals down to a size of ~25 nm has been demonstrated.

In this paper we present a technique for characterizing and manipulating isolated diamond nano-particles containing single photon emitters. The nano-manipulation of single diamond nano-crystals is achieved using a custom tapered optical fibre probe in a standard scanning electron microscope (SEM) with a nano-manipulator. The technique offers the unique advantage of being able to inspect and manipulate single diamond nano-crystals onto varied substrates and waveguiding devices with nano-scale precision.

2. Experimental details

To correlate the position of a single diamond nano-crystal containing a single defect centre in the SEM environment a marked silicon substrate was prepared using a focused ion beam (FIB). The markers were milled using 30 keV Ga⁺ ions with a current of 7 nA to a depth of ~3 µm to enable detection in the confocal microscope. The marked silicon substrate was then ultrasonically seeded in a solution of methanol and commercial grade diamond powder 0–500 nm (Metallographics Australia) known to contain appreciably high numbers of (N-V) defect centres. The diamond nano-particles are subsequently rinsed in acetone, methanol and deionised water to remove the unwanted larger and coagulated particles.

The marked region of the silicon substrate was characterised optically with an in-house scanning confocal microscope. The room temperature luminescent properties of the emitting centres were studied under continuous wave (CW) excitation at 532 nm with a 100×0.95 NA microscope objective, providing a spatial resolution of ~400 nm. The photo-luminescence from the emitting crystals captured by the microscope objective was coupled into a standard 62.5/125 µm multimode optical fibre to provide the confocallity of the microscope. A 560 nm long-pass and 650–750 nm band-pass filter was used to remove the unwanted pump excitation before being coupled into the fibre. The photon statistics of the diamond emitters were characterised using a fibre based version of the Hanbury Brown and Twiss (HBT) interferometer [24]. The characteristic dip in the second order correlation function at zero delay time enabled single N-V centres to be identified [25]. The confocal images of the marked silicon substrate were then used to correlate the position of the single N-V centres in the SEM. The SEM was operated with a nano-manipulator containing a custom tapered optical fibre probe with a tip size of ~50 nm to manipulate the diamond nano crystals [26]. The probe was fabricated using a custom CO₂ laser based fibre pulling system.

3. Results and discussion

The FIB marked silicon substrate used to correlate the position of the single photon emitters is shown in Fig. 1.

 

Fig. 1. Focused ion beam (FIB) marked silicon substrate prior to diamond seeding. Note: the depth of the milled markers was ~3 µm. (Media 1)

Download Full Size | PDF

A fluorescence intensity confocal map of a 70×70 µm region of the seeded marked substrate is shown in Fig. 2(a). The photon statistics of each fluorescent emitter were studied to identify a single N-V centre. The fluorescent emitter highlighted in Fig. 2(a) was found to exhibit the characteristic dip in the second order correlation function at the zero delay time (t=0) (see Fig. 2(d)), with a minimum g⁽²⁾(0) of 0.16 observed under 500 µW of excitation power at 532 nm. The measured single photon emission rate was ~270 kHz under the same excitation conditions. After the initial optical characterization, the sample was placed into the SEM for inspection and manipulation. The properties of silicon are such that the substrate does not experience significant charging in the SEM when operated with 10 keV electrons at 50 pA. The fabricated optical probe however does suffer from charging due to insulating nature of silica. Therefore, to avoid the charge related issues, the optical probe was coated with a thin layer of carbon. The carbon coating was deposited onto the probe to a thickness of ~10 nm using an Emitech K950X carbon coater. The areas of the marked substrate were compared to the confocal images to correlate the position of the single N-V emitter. Figure 2(b) shows the SEM image of the same 70×70 µm marked region of the substrate.

 

Fig. 2. (a) Confocal map of a 70×70 µm area of the marked and seeded silicon substrate. (b) SEM image of the same 70×70 µm region of a marked and seeded silicon substrate (c) The photo-luminescence spectrum of the fluorescence emitter identified in (a). (d) The photon statistics of a single N-V centre identified in (a) measured at room temperature with 500 µW excitation power at 532 nm. (Media 2)

Download Full Size | PDF

Once the position of the single N-V centre was identified, the tapered probe was used to remove the diamond nano-crystal from the substrate. When approaching the nano-crystal, the particle appears to energetically favor the tip and in some cases “jumps” from the substrate onto the tip with a high yield of ~90%. Although the silicon substrate conducts electrons well enough to reduce the charging effects from the electron beam, it may be that the charge path from the carbon coated tip provides a stronger attraction than the untreated silicon substrate. Once attached to the probe, the attraction remains strong enough for the probe to be manipulated and re-positioned to other locations of interest. It should be noted that several steps had to be taken to manipulate the single emitter onto a waveguiding structure in the form of a single mode optical fibre (3M FS-SN 3224). Firstly, in order to identify the core region of an optical fibre in the SEM, the end face was etched in 25% hydrofluoric (HF) acid for 2 minutes [27]. Secondly, the bulk of the optical fibre is made up of SiO₂ which suffers from significant charging in the SEM. To eliminate this effect, the optical fibre was coated with a thin layer of carbon to a thickness of ~20 nm. When manipulating the probe to the surface of the fibre endface, the charge path provided by the carbon coated endface was significant enough to allow the particle to be manipulated onto the surface once contact was made between the nano-particle and substrate.

It is evident that the process to remove the nano-particle from the probe is not yet optimized. The ideal surface preparation of the fibre end face, which maximizes the probability of successful transfer, is yet to be determined. Nevertheless, the process is successful ~70% of the time. The unsuccessful attempts to manipulate the nano-particles were due to the particle being stuck to the tip and occasionally the particle was repelled from both the probe and substrate. The single emitter in this particular example was manipulated to a region outside the core of the optical fibre to avoid the unwanted fluorescence background from co-dopants such as germanium and fluorine [27]. The core fluoresce is significant enough to degrade the sub-poissonian statistics of the single emitter. The ability to excite the nano-crystal on the core of the fibre without generating significant background fluorescence needs to be overcome before such waveguiding structures can be used as efficient collection devices. Figures 3(a)–(c) illustrate a sequence of SEM images taken before and after the manipulation of the SPS, with the corresponding confocal image of the endface of the optical fibre shown in Fig. 3(d).

 

Fig. 3. (a) SEM image of a single diamond nano-crystal (~300 nm) being removed from the marked silicon substrate. (b) SEM image of the tapered probe with the diamond nano-crystal attached and positioned above the optical fibre endface. (c) SEM image of the single emitting diamond nano-crystal positioned in the region outside the fibre core. (d) Confocal image of the fibre endface after manipulation and 1 hour anneal at 650 °C in air. (Media 3)

Download Full Size | PDF

After manipulation, the thin layer of carbon coating applied to the endface of the optical fibre was removed by annealing the fibre at 650 °C in air using an 1100 Quartz Tube furnace for 1 hour. To confirm that the manipulated N-V centre was not damaged or degraded in the process, the centre was characterized again using the same experimental configuration and compared with the results in Fig. 2 obtained on the silicon substrate. The confocal map of the post annealed fibre endface shown in Fig. 3(d) clearly identifies the single N-V emitter outside the core region of the fibre. It is worth mentioning here that the fluorescence from the core region of the fibre is not observed in the confocal map as the etched region of the core is beyond the depth of field (DOF) of the microscope objective. The most striking feature of the confocal image is the contrast between the SPS and its surroundings. In the 17×17 µm scan, the single N-V centre is the most predominant fluorescing object, allowing it to be easily identified. The photon statistics from the single centre after manipulation are shown in Fig. 4(a) and the measured single photon emission rate as a function of incident pump power is shown in Fig. 4(b). The performance characteristics of the centre are in excellent agreement with those obtained on the original silicon substrate. The autocorrelation function minimum at the zero delay time, g⁽²⁾(0), was measured to be 0.16 before and after manipulation and the measured single photon emission rate was within 5% of the rate obtained on the original substrate. These results confirm that the manipulation technique does not degrade the performance of the single photon source.

 

Fig. 4. (a) The photon statistics of the single N-V centre on the endface of the optical fibre measured at room temperature with 500 µW excitation power at 532 nm. (b) Single photon emission rate as a function of incident pump power. (Media 4)

Download Full Size | PDF

The minimum sized nano-crystal which can be manipulated using this technique is in principle limited to the resolution of the scanning electron microscope (~ few nm). In this work, the size of the nano-particle was governed by the size of the nano-crystal hosting a single NV centre. Recent work using a similar technique has demonstrated the manipulation of <100 nm sized nano-crystals, however the manipulation was limited to a distance of ~30 µm on the same substrate [28].

The nano-manipulation technique described here opens up the possibility to manipulate single photon sources onto structures such as mirrors, nano-wires and cavities which can all act to enhance the emission properties [29–32]. Furthermore, the ability to nano-manipulate single quantum systems could prove advantageous in realizing QIP and QI applications. Although the sample preparation can be laborious, economies of scale can be achieved since a single seeded substrate contains numerous single N-V centres.

4. Conclusion

A novel technique for manipulating single diamond nano-crystals, containing single N-V centres, was presented. The technique maintains the performance characteristics of a single N-V centre in a 300 nm diamond crystal before and after manipulation. The ability to inspect the nano-particles throughout manipulation process provides an important adjunct to existing methods for analyzing and manipulating single photon sources. Manipulating these types of quantum systems onto nanostructures and other plasmonic and cavity devices is the goal of future work.