pubs.acs.org/NanoLett
Fiber-Integrated Diamond-Based Single Photon Source Tim Schro¨der,*,† Andreas W. Schell,† Gu¨nter Kewes, Thomas Aichele, and Oliver Benson Institute of Physics, Humboldt-Universita¨t zu Berlin, Newtonstraße 15, D-12489 Berlin, Germany ABSTRACT An alignment free, micrometer-scale single photon source consisting of a single quantum emitter on an optical fiber operating at room temperature is demonstrated. It easily integrates into fiber optic networks for quantum cryptography or quantum metrology applications.1 Near-field coupling of a single nitrogen-vacancy center is achieved in a bottom-up approach by placing a preselected nanodiamond directly on the fiber facet. Its high photon collection efficiency is equivalent to a far-field collection via an objective with a numerical aperture of 0.82. Furthermore, simultaneous excitation and re-collection through the fiber is possible by introducing a fiber-connected single emitter sensor. KEYWORDS Single photon source, single emitter sensor, fiber integration, nitrogen vacancy, nanomanipulation, nanodiamond
R
ecently, important steps have been made in the design and fabrication of efficient and bright solidstate single photon sources.2,3 Such sources are of particular interest to build quantum information devices and to perform quantum cryptography experiments.4 The nitrogen-vacancy (N-V) center in diamond has been proven to be a stable and bright single photon emitter (SPE), even at room temperature.5 It can also be used for quantum storage6 as it provides a triplet ground state with electron spin decoherence times in the millisecond range that can be controlled and read out optically.7 The efficient and smallscale coupling of a SPE to a photonic device or a fiber network is an essential requirement to build complex quantum systems.8 Efforts have been made by growing diamond crystals directly on fiber facets,9 by depositing preselected nanodiamonds on specially prepared fibers,10 or by indirectly coupling N-V centers via a toroidal resonator to a fiber.11 However, a direct near-field coupling of single photons from a SPE to guided optical modes of a fiber has not been achieved. The approach taken in this Letter is a bottom-up assembly of a single emitter-fiber system via direct pick-and-place manipulation of a nanodiamond by an atomic force microscope (AFM)12 (see Methods). This bottom-up approach allows preselection of an emitter that features the required properties such as emission wavelength, brightness, type of defect, and other physical characteristics for a specific application. Preselection is a fundamental requirement if complex systems are designed to guarantee that all interacting components have matching properties.8 We selected a 30 nm sized nanodiamond containing just one single N-V center which emits single photons after optical * To whom correspondence physik.hu-berlin.de.
should
be
addressed,
excitation. Single photon generation was verified via a measurement of the normalized intensity autocorrelation function
g(2)(τ) ) 〈I(t)I(t + τ)〉/〈I(t)〉2 in a Hanbury Brown and Twiss (HBT) setup.13 Finding a value of g(2)(0) < 0.5 proves emission from a single N-V center. The nanodiamond was deposited via the pick-and-place manipulation technique in the core region of a commercial photonic crystal fiber (PCF) with a diameter of 90 µm, a core region of 1.5 µm, and a length of about 10 cm. Figure 1 shows a picture of the fiber facet taken with an electron microscope as well as an AFM surface scan of the fiber core featuring the nanodiamond on the facet. The nanodiamond was placed close to the center of the core where a maximum coupling efficiency is expected from finite difference time domain (FDTD) simulations (Figure 3b). The deposited nanodiamond emitter was optically excited on the fiber facet by continuous wave (CW) laser radiation of 532 nm with a typical excitation power of 40 µW in the focus of a homemade scanning confocal microscope with a numerical aperture (NA) of 0.9. Figure 2b depicts a fluorescence image taken with a charge coupled device camera while filtering the excitation laser light. This figure clearly shows the structure of the fluorescence modes of the PCF. However, the N-V center’s emission is outshined by strong background fluorescence below 670 nm as indicated in Figure 2c. With spectral filtering this background light stemming from inelastic Raman scattering and fluorescence emitted from fiber impurities can be significantly suppressed. Panels e and f of Figure 2 show the filtered fluorescence intensity collected by the confocal microscope and the fiber, respectively. In Figure 2e excitation and scanning confocal detection were both performed on the diamond-loaded side of the fiber, whereas in Figure 2f the excitation laser was scanned across the diamond-loaded facet of the fiber while the fluorescence was detected through the fiber (see configurations I and II in the schematic in Figure 2a). The N-V center’s emission appears as a
tim.schroeder@
† Contributed equally to this work. Received for review: 09/29/2010 Published on Web: 12/07/2010
© 2011 American Chemical Society
198
DOI: 10.1021/nl103434r | Nano Lett. 2011, 11, 198–202
FIGURE 1. Fiber facet with a single mounted nanodiamond. (a) Scanning electron microscope pictures of a photonic crystal fiber with a diameter of 90 µm. The white box indicates a zoom in on the fiber facet. (b) Atomic force microscope (AFM) image of the fiber core region. The diamond in the middle of the core (marked by an arrow) has a diameter of about 30 nm and was placed via an AFM pick-and-place technique. It contains a single N-V center as single photon emitter.
FIGURE 2. Excitation and detection scheme and fluorescence light characteristics. (a) Schematic showing the different experimental configurations, i.e., (I) excitation and detection on the diamond-loaded side of the fiber, (II) excitation on the diamond-loaded side of the fiber and detection through the fiber, and (III) detection and excitation through the fiber. (b) CCD-camera image of the light collected from the fiber under 532 nm excitation. (c) Spectrum of this light (black line) and the spectrum after adding a 650 nm long-pass filter (red line). The inset displays a section of the entire spectrum as indicated by the dashed lines. (d) Overlay image of the fiber core consisting of a scanning electron and a corresponding confocal microscopy image. The white dot represents the fluorescence of the N-V center. (e, f) Scanning confocal microscopy images in geometries I and II, respectively.
fluorescence.14 One of the two dipoles is inclined strongly relative to the optical axis and therefore only weakly contributes to the fluorescence guided by the fiber. The numerical simulation (Figure 3b-d) confirms this behavior. The measurement implies that coupling of the emitter to the fiber modes and thus photon collection efficiency can be further improved by optimizing the orientation of the nanodiamond. Such optimization of the dipole orientation could be achieved in an in situ feedback process by simultaneously manipulating the nanodiamond and maximizing the fluorescence intensity coupled into the fiber.
diffraction limited spot with high intensity (about 9 and 20 kilocounts/s for configurations I and II, respectively) while the holes of the PCF emit a 6 times lower intensity. This proves that the N-V center emission couples directly to optical modes of the fiber and can be separated from background fluorescence. Figure 3a shows the measured polarization properties of the light after the fiber (measured in configuration II). The C3v symmetry of the N-V center determines an emission pattern resembling that of two orthogonal dipole emitters. Their relative alignment to the fiber facet is deduced by analyzing the polarization of the © 2011 American Chemical Society
199
DOI: 10.1021/nl103434r | Nano Lett. 2011, 11, 198-–202
FIGURE 3. Diamond-fiber coupling. (a) Polarization properties of the N-V center fluorescence detected through the fiber. The observed modulation accounts for the symmetry of the N-V center resembling an emission pattern of two orthogonal dipole emitters. The visibility V ) (Imax - Imin)/(Imax + Imin) of 0.48 indicates that one of the dipoles is strongly inclined relative to the optical axis.14 (b-d) Results of a FDTD simulation showing the coupling efficiency of a single dipole emitter to the fiber modes versus the dipole’s position at a fixed distance of 12.5 nm to the fiber facet. In (b) and (c) two-dimensional position images with the dipole axis orthogonal and parallel to the optical axis of the fiber, respectively, are plotted while (d) shows the coupling along the black line in (b) and (c). Crosses represent dipole orientation in the y direction, circles in the x direction, and × marks orientation perpendicular to the x-y plane. Maximum coupling efficiency is achieved in the center of the core.
FIGURE 4. Fluorescence emission characterization. (a, b) Autocorrelation function g(2)(τ) of the N-V center’s fluorescence measured in configurations I and II of Figure 2a, respectively, under continuous wave excitation. g(2)(0) < 0.5 demonstrates the single photon character. No background was subtracted. (c, d) Saturation measurements of the N-V center fluorescence in configurations I and II. The maximum count rate is 52.6 kilocounts/s for configuration I where light was collected via an objective with NA of 0.9. In configuration II where light was collected via the fiber, the maximum count rate is 43.2 kilocounts/s. This corresponds to an effective NA of about 0.82 of the fiber. (e, f) Autocorrelation function g(2)(τ) of the N-V center fluorescence in configurations II and III under pulsed excitation. The data were postprocessed to imitate gating with an offset of 3.5 ns (see Methods). Red curves are fits to the data.
for configuration I and Rinf ) 43.2 kilocounts/s for configuration II (Figure 4c,d). Photon collection efficiency could be further improved by coating a Bragg mirror on top of the diamond loaded facet suitable to reflect single photon emission into the core of the fiber. An even larger emission rate of single photons could be obtained by choosing a brighter defect center, such as a Si- or Cr-related defect.15 A true on-demand single photon source requires pulsed excitation which confines the single photon in a time interval determined by the spontaneous lifetime. Pulsed excitation of our fiber-integrated SPE with 532 nm results in a g(2)(0) of about 0.5. However, due to the larger power density of the pulsed excitation, single photon emission is accompanied by strong
The light emitted by the N-V center on the fiber facet shows strong single photon character as indicated by the measured autocorrelation functions at zero time delay g(2)(0) (Figure 4a,b). The photons measured on the diamond-loaded side in configuration I and collected with a NA ) 0.9 objective have g(2)(0) ) 0.45 at 40 µW excitation power in the focus. The light collected by the fiber in configuration II has g(2)(0) ) 0.36 at 49 µW excitation power. No background correction was applied to the data. To further improve the single photon character experimentally a Bragg mirror could be coated on the fiber facet to reflect laser excitation light and thus prevent fiber fluorescence. Maximum single photon emission at saturation is calculated to Rinf ) 52.6 kilocounts/s © 2011 American Chemical Society
200
DOI: 10.1021/nl103434r | Nano Lett. 2011, 11, 198-–202
FIGURE 5. A variety of optical fibers with single nanodiamonds placed on the fiber core. (a-c) Atomic force microscope images of different fiber facets. The arrows indicate the nanodiamond position. The white boxes indicate a zoom-in on details. A cloverleaf-shaped suspended core fiber is shown in (a) before and after placing of the diamond. In (b) a step index fiber is displayed. The dashed circle indicates the core position, which is not clearly visible in the topography image. A suspended core fiber with wagon wheel design is depicted in (c), before and after placing of the diamond.
background fluorescence with lifetimes below 3.5 ns, which are much shorter than the lifetime of the N-V center of about τ ) 25 ns. Therefore, temporal filtering is possible to suppress the background fluorescence. Here we performed postprocessing of the data to mimic a physical temporal filter, e.g., a fast shutter opening 3.5 ns after the excitation pulse (see Methods). As expected, the autocorrelation function g(2)(τ) calculated from the postprocessed data shows a deeper antibunching of g(2)(0) ) 0.21 (red curve in Figure 4e). A comparison of the measured intensities at saturation in configurations I and II reveal an effective numerical aperture NAeff ) 0.82 of the fiber assuming isotropic light radiation. Simulations considering dipole emission patterns give a NAeff of 0.77 and 0.68 for dipole emitters orthogonal and parallel to the optical axis of the fiber, respectively (see Methods). Compared to the far field NA of 0.45 of the fiber, this is a tremendous improvement caused by the near-field emitter-fiber coupling and the refractive index n ) 1.46 at 638 nm of the fiber.16 This near-field coupling improves the collection in a similar way as a solid immersion lens17,18 and could be further increased using fiber materials with a higher index of refraction.18 The robust on-facet integration of a single defect nanodiamond has another thriving feature. It was also possible to excite the N-V center through the fiber while its emission © 2011 American Chemical Society
is simultaneously collected through the same fiber (configuration III in Figure 2a). The collected light still shows single photon character with g(2)(0)pulsed ) 0.23 if temporal filtering is applied (Figure 4f). The fiber-integrated SPE thus represents a versatile quantum light source for optical scanning probe imaging19 or as local source of quantum light, e.g., to launch single photons in the evanescent field of other optical components such as microcavities.20 Moreover, the optical read-out of the spin state of the single defect center at the end of a fiber tip resembles a small and robust magnetic field probe.21,22 In all cases operation is at room temperature and all optical components are fiber coupled. Our pick-and-place approach based on AFM manipulation described in Methods is not limited to assembly of a specific diamond-PCF fiber system. We applied it to a variety of optical fibers with different glass compositions and designs (Figure 5a-c). The high positioning precision of the pick-and-place technique is a crucial requirement as the cores of these fibers have dimensions down to diameters of 605 nm (Figure 5c). This proves the broad suitability of the presented manipulation technique which is also not limited to optical fibers. In conclusion, we built a fully integrated single photon source showing a pronounced antibunching dip with g(2)(0) < 0.21 by combining a diamond nanocrystal with a commercial optical fiber using an atomic force microscope 201
DOI: 10.1021/nl103434r | Nano Lett. 2011, 11, 198-–202
manipulation technique. This technique is independent of emitter type and sample geometry and thus not limited to optical fibers. The assembled system reaches an effective numerical aperture of 0.82 while having only micrometer dimensions. It therefore allows integration into ultracompact quantum photonic devices or fiber networks. Since the direct fiber-emitter coupling is ultrastable, these devices would be maintenance-free, and scalability to more complex systems is feasible. Such fiber-based sources can be bundled to launch a larger, yet precisely defined number of photons into on-chip photonic structures for integrated quantum optical technology. The possibility to simultaneously excite the N-V center and re-collect its single photon emission through the fiber will open the way toward new devices where single emitters need to be coupled directly to photonic structures or where they are used as local quantum sensors. Materials and Methods. Pick-and-Place. Nanodiamond samples were prepared by spin coating (2500 rpm) of centrifuge cleaned nanodiamonds (synthesized nonirradiated diamond type Ib, median size 25 nm) in a solution of water and 0.02% polyvinyl alcohol (PVA) on a coverslip. Precharacterization was performed on an inverted optical microscope using a home-built confocal setup equipped with a Hanbury Brown and Twiss (HBT) correlator and atomic force microscope (AFM) access. The nanodiamond was picked up using an AFM (JPK Instruments) by repeatedly pressing the tip down on the nanodiamond with a force on the order of magnitude of 0.1 µN. After the AFM tip was brought manually in a position atop the fiber cladding, the nanodiamond was placed using the same technique. In a second step the fiber core was identified by scanning in intermittent contact mode and the nanodiamond was transferred to the core via the previously described technique. Autocorrelation Measurements. A Hanbury Brown and Twiss setup consisting of two avalanche photodiodes (PerkinElmer SPCM) and a time-correlated single photon counter (PicoQuant TimeHarp 200 or PicoHarp 300) equipped with a suitable electric delay in one of its channels was used for the measurement of the autocorrelation function. Continuous wave excitation measurements were performed by recording the time intervals between the photon detection events. For pulsed excitation measurements the time-tagged time-resolved mode of the PicoHarp 300 was used. This mode records the excitation pulses and the arrival times of all photons with a time stamp. Calculation of the autocorrelation function is done afterwards. To suppress fast decaying background fluorescence photons in an interval of 3.5 ns after the laser pulse were not included in the correlation calculations. Numerical Calculations. Finite difference time domain simulations were carried out using a commercial software package (Lumerical FDTD Solutions). The simulated geometry consists of a section of 10 µm × 10 µm × 15 µm of the fiber with a dipole emitter placed 12.5 nm above the fiber facet. The dipole x-y position and the dipole orientation
© 2011 American Chemical Society
werevaried.Toobtainaquantitativeresultforthedipole-fiber coupling light arriving at the end of the fiber within a circle of radius 1.65 µm from the center was considered to be guided by the fiber. Acknowledgment. Financial support of the BMBF (KEPHOSI) and DFG (AI 92/3) is acknowledged. The authors thank Mr. Jens Kobelke, IPHT-Jena, and Mr. Alexandre Francois, University of Adelaide, for providing the suspended core fiber with cloverleaf design and the suspended core fiber with wagon wheel design, respectively. REFERENCES AND NOTES (1) (2) (3) (4) (5) (6)
(7)
(8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21)
(22)
202
O’Brien, J. L.; Furusawa, A.; Vuckovic´, J. Nat. Photonics 2009, 3, 687–695. Claudon, J.; Bleuse, J.; Malik, N. S.; Bazin, M.; Jaffrennou, P.; Gregersen, N.; Sauvan, C.; Lalanne, P.; Ge´rard, J. Nat. Photonics 2010, 4, 174–177. Babinec, T. M.; Hausmann, B. J. M.; Khan, M.; Zhang, Y.; Maze, J. R.; Hemmer, P. R.; Loncar, M. Nat. Nanotechnol. 2010, 5, 195– 199. Beveratos, A.; Brouri, R.; Gacoin, T.; Villing, A.; Poizat, J.; Grangier, P. Phys. Rev. Lett. 2002, 89, 187901. Kurtsiefer, C.; Mayer, S.; Zarda, P.; Weinfurter, H. Phys. Rev. Lett. 2000, 85, 290–293. Gaebel, T.; Domhan, M.; Popa, I.; Wittmann, C.; Neumann, P.; Jelezko, F.; Rabeau, J. R.; Stavrias, N.; Greentree, A. D.; Prawer, S.; Meijer, J.; Twamley, J.; Hemmer, P. R.; Wrachtrup, J. Nat. Phys. 2006, 2, 408–413. Balasubramanian, G.; Neumann, P.; Twitchen, D.; Markham, M.; Kolesov, R.; Mizuochi, N.; Isoya, J.; Achard, J.; Beck, J.; Tissler, J.; Jacques, V.; Hemmer, P. R.; Jelezko, F.; Wrachtrup, J. Nat. Mater. 2009, 8, 383–387. O’Brien, J. L. Science 2007, 318, 1567. Rabeau, J. R.; Huntington, S. T.; Greentree, A. D.; Prawer, S. Appl. Phys. Lett. 2005, 86, 134104. Ampem-Lassen, E.; Simpson, D. A.; Gibson, B. C.; Trpkovski, S.; Hossain, F. M.; Huntington, S. T.; Ganesan, K.; Hollenberg, L. C.; Prawer, S. Opt. Express 2009, 17, 11287. Gregor, M.; Henze, R.; Schro¨der, T.; Benson, O. Appl. Phys. Lett. 2009, 95, 153110. Schietinger, S.; Barth, M.; Aichele, T.; Benson, O. Nano Lett. 2009, 9, 1694–1698. Brown, R. H. Nature 1956, 177, 27–29. Alegre, T. P. M.; Santori, C.; Medeiros-Ribeiro, G.; Beausoleil, R. G. Phys. Rev. B 2007, 76, 165205. Aharonovich, I.; Castelletto, S.; Simpson, D. A.; Stacey, A.; McCallum, J.; Greentree, A. D.; Prawer, S. Nano Lett. 2009, 9, 3191–3195. Malitson, L. H. J. Opt. Soc. Am. 1965, 55, 1205. Mansfield, S. M.; Kino, G. S. Appl. Phys. Lett. 1990, 57, 2615– 2616. Hadden, J. P.; Harrison, J. P.; Stanley-Clarke, A. C.; Marseglia, L.; Ho, Y. D.; Patton, B. R.; O’Brien, J. L.; Rarity, J. G. arXiv: 1006.2093v2, 2010. Michaelis, J.; Hettich, C.; Mlynek, J.; Sandoghdar, V. Nature 2000, 405, 325–328. Barclay, P. E.; Fu, K.; Santori, C.; Beausoleil, R. G. Appl. Phys. Lett. 2009, 95, 191115. Balasubramanian, G.; Chan, I. Y.; Kolesov, R.; Al-Hmoud, M.; Tisler, J.; Shin, C.; Kim, C.; Wojcik, A.; Hemmer, P. R.; Krueger, A.; Hanke, T.; Leitenstorfer, A.; Bratschitsch, R.; Jelezko, F.; Wrachtrup, J. Nature 2008, 455, 648–651. Maze, J. R.; Stanwix, P. L.; Hodges, J. S.; Hong, S.; Taylor, J. M.; Cappellaro, P.; Jiang, L.; Dutt, M. V. G.; Togan, E.; Zibrov, A. S.; Yacoby, A.; Walsworth, R. L.; Lukin, M. D. Nature 2008, 455, 645– 648.
DOI: 10.1021/nl103434r | Nano Lett. 2011, 11, 198-–202