Diamond colour centres as a nanoscopic light source for scanning near-field optical microscopy

Transcription

1 Journal of Microscopy, Vol. 202, Pt 1, April 2001, pp. 2±6. Received 28 August 2000; accepted 11 October 2000 Diamond colour centres as a nanoscopic light source for scanning near-field optical microscopy S. KUÈ HN, C. HETTICH, C. SCHMITT, J-PH. POIZAT* AND V. SANDOGHDAR Fachbereich Physik & Optik-Zentrum Konstanz, UniversitaÈt Konstanz, Fach M696, D Konstanz, Germany *Institut d'optique, B.P. 147, F Orsay, France Key words. Diamond, SNOM. Summary Recently it was shown that a single molecule at cryogenic temperatures could be used as a local light source for illumination of a sample in the near field. Conventional light-emitting systems such as dye molecules and semiconductor quantum dots could also be used for this purpose, but they suffer from lack of photostability. However, colour centres in diamond have been found to be remarkably stable against bleaching and blinking effects. Here we present the first SNOM images taken with nanoscopic diamond crystals as a light source. Introduction Since its debut in the early 1980s (Lewis et al., 1984; Pohl et al., 1984) scanning near-field optical microscopy (SNOM) has witnessed many breakthroughs and novel configurations aimed at achieving an ever higher resolution. In the most common form of SNOM one illuminates the sample by a subwavelength opening in a metallic film around a very sharp dielectric probe such as an optical fibre tip. In theory one could achieve an unlimited high resolution if only one could confine the illumination to an arbitrarily small region by making this aperture as small as possible. In this way one would create correspondingly high spatial frequencies for imaging the sample. Unfortunately, the finite skin depth of optical radiation in realistic metals prevents us from arriving at arbitrarily small apertures (Novotny et al., 1995) and therefore one cannot hope to reach the interesting realm of nanometre resolution. An intuitive extension of an opening in an otherwise opaque screen is to use a nanoscopic source of light suspended in front of the sample. The advantage of this scheme is that such a light-emitting medium could be made as small as one single atom or molecule, pointing to the Correspondence to: V. Sandoghdar. Tel.: ; fax: ; vahid.sandoghdar@uni-konstanz.de possibility of obtaining molecular resolution in optics if only the source molecule could be brought nearly in contact with the sample. The use of a single atom or molecule as a probe was suggested in the literature as early as 1991 (Lewis & Lieberman, 1991; Sekatskii & Letokhov, 1996) and was very recently demonstrated in our laboratory (Michaelis et al., 2000). The reasons for this long delay between proposal and realization are many-fold. First, optical detection of single molecules itself is a very young field and has been maturing in this period (Bache et al., 1997). Second, under most conditions single molecules, and in fact even single quantum dots, have proven to suffer from photobleaching and an intermittent emission known as blinking (Dickson et al., 1997). Common estimates amount to the emission of only photons before the molecule undergoes an irreversible bleaching at room temperature (Soper et al., 1993). Third, it is not trivial to place a single molecule close to a sample surface to within a few nanometres. In our recent experiment we used fluorescence excitation spectroscopy (Orrit & Bernard, 1990) at T ˆ 1.4 K to identify a single terylene molecule in a microscopic crystal glued to the end of a sharpened optical fibre. This probe could be positioned against a surface using shear-force distance regulation. Under these conditions photostability was not a problem, and one could perform continuous measurements with the same single molecule. Given the typical resolution of 50±150 nm obtained in the laboratory using aperture probes, closing the bridge between this regime and the single molecule level would be also of great importance. At a first glance using a finite number of emitters in a small volume might seem to relieve the problems discussed above. Indeed, several groups have attempted fabrication and demonstration of probes containing a finite nanoscopic amount of emitting material at the end of a fibre probe (Lewis & Lieberman, 1991; GoÈttlich & Heckl, 1995; Kurihara et al., 1996; Kramper et al., 1999). However, even a volume as large as 50 nm in diameter 2 q 2001 The Royal Microscopical Society

2 NANOSCOPIC LIGHT SOURCE FOR SNOM 3 contains a small number of emitters, and therefore the issue of photobleaching remains a major obstacle. In principle lithographically fabricated single quantum dots that are capped with a protecting layer can satisfy these requirements (Sandoghdar & Mlynek, 1999), but to our knowledge due to fabrication difficulties this has not been realized yet. It has been shown that single nitrogen-vacancy colour centres in diamond emit quite efficiently and show a remarkable photostability (Gruber et al., 1997). As a result they have been proposed as sources of single photons (Brouri et al., 2000; Kurtsiefer et al., 2000) and as probes for SNOM (Martin et al., 1999). In this paper we present what is to our knowledge the first optical images recorded using the emission of colour centres in diamond nanocrystals. We discuss our experimental procedure and give an outlook for future work. Results and discussion We used type 1b submicrometre-sized powder diamond that had been irradiated at a dose of ecm 22 with 2 MeV electrons and then annealed for 2 h at 850 8C in vacuum in order to create defects. In such a process carbon vacancies are created which diffuse through the sample and combine with nitrogen impurities (Martin et al., 1999). This Fig. 1. Prior to the tip preparation confocal images of the diamond crystals were taken in the fluorescence (a) and backscattering, (b) regime; (c) shows a scan of a smaller region around a pre-selected diamond. The FWHM of the cross-section in image (d) taken along the line drawn in image (c) is 450 nm. This shows that the emission centre is smaller than the resolution of the microscope which is about 400 nm.

3 4 S. KUÈ HN ET AL. results in defects known as the nitrogen-vacancy (N-V) colour centres. These centres can be excited using the light of an argon ion laser at l, 514 nm and yield a fairly broad emission around l, 650 nm (Gruber et al., 1997). After spin coating the powder particles on a cover glass they were next examined using a fluorescence confocal microscope with a sample scanning stage. Figure 1 shows the images recorded when detecting the back-scattered excitation light and the fluorescence signal. We note that some particles do not fluoresce. However, those that do fluoresce show no signs of photobleaching. Once a particle has been selected it is approached with an optical fibre tip that has been treated to contain a thin layer of polyethylenimine (Kalkbrenner et al., 2001). This approach takes place under the control of shear-force regulation using a quartz tuning fork. This is a part of the SNOM head unit that fits onto the inverted optical microscope so that fluorescence and shear-force measurements can be performed simultaneously. By monitoring this process on line we can pick up a nanocrystal with a very good control. Next the SNOM head including this tip is removed, a test sample is placed on the scanner stage, the tip is replaced and then approached. The sample used is a thin film of gold deposited on a cover glass in a way to contain holes of diameters 6 mm (Kalkbrenner et al., 2001). Excitation of the diamond nanocrystal at the end of the tip can be achieved by illumination through the fibre holding it, from the side using conventional lenses, or through the microscope objective from below. In this first experiment we have chosen to do the latter. In this arrangement we position the tip on the optical axis of the microscope and then scan the sample so that the relative position of the tip and the excitation does not change during the experiment. The fluorescence light is collected by the same microscope immersion oil objective with a numerical aperture of 1.25 and is sent to a photomultiplier tube after passing through a holographic notch filter and a band pass filter. Figure 2(a) displays a SNOM scan taken with the fluorescence light of the tip. A section of a hole in the gold film is clearly imaged with a very good signal-to-noise ratio. Typical count rates and integration times were about Fig. 2. Image (a) was taken with the fluorescence of a confocally excited diamond probe scanned in constant-gap mode across a section of a gold mask. In addition to the fluorescence intensity the shear-force and back-scattering signals were recorded. Images ((b), (c) and (d) show cross-sections of these signals along the line drawn in image (a). By fitting a tanh[(x 2 x 0 )/l] step to the data and taking a 10±90% criterion we obtained an estimate for the edge-sharpness for each of the imaging signals. As is known from AFM measurements the gold edge-sharpness is not limiting all of the above resolutions. A shift in the position of the edges between the fluorescence and confocal images could stem from a slight misalignment between the diamond and the intensity maximum of the Gaussian illumination spot. The shift between the fluorescence and the shear-force images can be explained by the diamond being attached rather to the side of the tip.

4 NANOSCOPIC LIGHT SOURCE FOR SNOM 5 Fig. 3. The diamond probe was alternatively positioned above the clear cover glass (a) and the evaporated metal film (b) and excited confocally at 514 nm. The figure shows the fluorescence spectra at the two sites and the remains of the strongly rejected excitation light. In a rough estimate the difference in amplitude agrees with the transmittivity of a 30-nm gold film passed twice photons s 21 and 100ms pixel 21, respectively. In Fig. 2(b,c and d) we show fluorescence, confocal (excitation light) and shear-force topography cross-sections corresponding to the indicated line in this scan. The topography profile is shifted with respect to the SNOM image, indicating that most probably the crystal was mounted on one side of the tip whereas the topography sensing was done with the other side. A fit using a hyperbolic tangent tanh[(x 2 x 0 ) l 21 ] is used to quantify the edge sharpness in these cuts. For the SNOM image in Fig. 2(b) we read off an edge sharpness of about 300 nm. For the confocal scan the edges are as broad as 2.2 mm because in this measurement the excitation beam was de-focused in order to decouple the spatial extent of the excitation beam from that of emission. The intensity oscillations in the image 2(a) represent the modulations of the excitation light due to diffraction from the hole edges, and we have seen that they can be eliminated if the illumination is tightly focused. When the excitation beam was optimally focused we obtained an edge sharpness of about 500 nm in confocal images. In order to check further that the light falling on the photomultiplier tube is indeed due to the emission from diamond N-V centres, we have recorded spectra of the emission. Figure 3 shows spectra taken with the tip in front of the hole and in front of the metal part of the sample. The spectra are typical of the N-V centres in diamond. We have performed such measurements with two different tips and have obtained similar results. Conclusion We have presented the first SNOM images taken with colour centres as nanoscopic sources. The N-V defect centres in diamond used in this experiment have been shown to be extremely photostable, displaying no sign of ageing or bleaching at room temperature and under ambient conditions. This opens the door to practical applications of SNOM using active sources of light. By examining the size of the powder particles in topography scans in addition to the fluorescence measurements reported here, we plan to fabricate tips consisting of smaller nanocrystals in the range of 20±50 nm. Reaching this size regime also implies approaching the single colour centre limit (Gruber et al., 1997), offering a room temperature analogue for the single molecule SNOM source (Michaelis et al., 2000). Access to single defect centres in our arrangement will be also of great interest for a range of experiments such as manipulation and efficient coupling of single photons (Kurtsiefer et al. 2000). Acknowledgements We gratefully acknowledge the financial support by the European Union. We thank T. Kalkbrenner for help with tip preparation, J. Michaelis for fruitful discussions and J. Mlynek for continuous support. References BacheÂ, T., Moerner, W.E., Orrit, M. & Wild, U.P (1997) Single Molecule Optical Detection, Imaging and Spectroscopy. VCH, Weinheim, Germany. Brouri, R., Beveratos, A., Poizat, J.-P. & Grangier, P. (2000) Photon antibunching in the fluorescence of individual color centres in diamond. Opt. Lett. 17, 1294±1296. Dickson, R.M., Cubitt, A.B., Tsien, R.Y. & Moerner, W.E. (1997) On/ off blinking and switching behaviour of single molecules of green fluorescent protein. Nature, 388, 355±358. GoÈttlich, H. & Heckl, W.M. (1995) A novel probe for near field optical microscopy based on luminescent silicon. Ultramicroscopy, 61, 145±153. Gruber, A., DraÈbenstedt, A., Tiez, C., Fleury, L., Wrachtrup, J. & von Broczyskowski, C., (1997) Scanning Confocal Optical Microscopy and Magnetic Resonance on Single Defect Centres. Science, 276, 2012±2014. Kalkbrenner, T., Ramstein, M., Mlynek, J. & Sandoghdar, V. (2001) A single gold particle as a probe for apertureless scanning nearfield optical microscopy. J. Microsc. 202, 72±76. Kramper, P., Jebens, A., MuÈ ller, T., Mlynek, J. & Sandoghdar, V. (1999) A novel fabrication method for fluorescence-based apertureless scanning near-field optical microscope probes. J. Microsc. 194, 340±343. Kurihara, K., Watanabe, K. & Ohtsu, M. (1996) Photon scanning tunneling microscopy with light-emitting probes. Proc. OFS-11, 694±697. Kurtsiefer, C., Mayer, S., Zarda, P. & Weinfurter, H. (2000) Stable Solid- State Source of Single Photons. Phys. Rev. Lett. 85, 290±293. Lewis, A. & Lieberman, K. (1991) Near-field optical imaging with a non-evanescently excited high-brightness light source of subwavelength dimensions. Nature, 354, 214±216.

5 6 S. KUÈ HN ET AL. Lewis, A., Isaacson, M., Harootunian, A. & Murray, A. (1984) Development of a 500 Angstrom Spatial Resolution Light Microscope. Ultramicroscopy, 13, 227±231. Martin, J., Wannemacher, R., Teichert, J., Bischoff, L. & KoÈhler, B. (1999) Generation and detection of fluoresence color centers in diamond with submicron resolution. Appl. Phys. Lett. 75, 3096± Michaelis, J., Hettich, C., Mlynek, J. & Sandoghdar, V. (2000) Optical microscopy with a single-molecule light source. Nature, 405, 325. Novotny, L., Pohl, D.W. & Hecht, B. (1995) Scanning near-field optical probe with ultrasmall spot size. Opt. Lett. 20, 970±972. Orrit, M. & Bernard, J. (1990) Single pentacene molecules detected by fluorescence excitation in a p-terphenyl crystal. Phys. Rev. Lett. 65, 2716±2719. Pohl, D.W., Denk, W. & Lanz, M. (1984) Optical stethoscopy: Image recording with resolution lambda/20. Appl. Phys. Lett. 44, 651± 653. Sandoghdar, V. & Mlynek, J. (1999) Prospects of apertureless SNOM with active probes. J. Opt. A: Pure Appl. Opt. 1, 523±529. Sekatskii, S.K. & Letokhov, V.S. (1996) Single fluorescence centers on the tips of crystal needles: First observation and prospects for application in scanning one-atom fluorescence microscopy. Appl. Phys. B, 63, 525±530. Soper, S., Nutter, H.L., Keller, R., Davis, L.M. & Shera, E.B. (1993) The Photophysical Constants of Several Fluorescent Dyes Pertaining to Ultrasensitive Fluorescence Spectroscopy. Photochem. Photobiol. 57, 972±977. StuÈ rmer, H., KoÈhler, J. & Jovin, T. (1998) Microstructured polymer tips for scanning near-field optical microscopy. Ultramicroscopy, 71, 107±110.