We introduce a new, easily implementable sub-diffraction-limit microscopy technique utilizing the optical AND-gate property of fluorescent nanodiamond (FND). We demonstrate that when FND is illuminated by two spatially-offset lights of different wavelengths, emission comes only from the region of their overlap, which is used to reduce the effective point spread function from ~300 nm to ~130 nm in lateral plane, well below the diffraction limit.
©2012 Optical Society of America
Far-field fluorescence microscopy is widely used in biology as a vital tool to image the interior of living cells, often with molecular specificity. However, the spatial resolution of an optical microscope is severely constrained by the diffraction limit of visible light , which frequently makes it difficult to directly observe the structures of biomolecules and their interactions. In order to overcome the diffraction limit, sub-diffraction-limit optical microscopy techniques have been introduced in the last two decades and improved the spatial resolution at least by an order of magnitude. They take advantage of reversible fluorescence transition to a dark state to break the diffraction limit by using the metastable non-fluorescent states of fluorophores. Two subgroups of technique have been established: one based on reducing the volume where the fluorophores remain in bright state and the other attempting localization of single fluorophores. The former uses spatially defined population depletion of a bright state, a concept generally called reversible saturable optical fluorescence transition , which includes stimulated emission depletion (STED) microscopy , ground-state depletion (GSD) microscopy , and saturated structured illumination microscopy . The latter method uses stochastic localization of a single fluorophore to distinguish it from surrounding fluorophores by fluorescence switching, and subsequent reconstruction of the final image using their locations. This method usually requires photo-switchable fluorophores such as Cy5 in stochastic optical reconstruction microscopy  or photo-switchable green fluorescent proteins in (fluorescence) photoactivation localization microscopy [7,8]. These super-resolution microscopy techniques enable 3-dimensional [9–11], multi-color [12–14] imaging of living cells [15–18], and thus have played a crucial role in elucidating biological problems that would have been otherwise impossible with diffraction-limited confocal microscopy.
In this report, we introduce a new, easily implementable sub-diffraction-limit microscopy technique that utilizes the optical AND-gate property of fluorescent nanodiamond (FND). FND is a nanometer-size diamond grain containing color centers that consist of an impregnated nitrogen atom and a neighboring, negatively charged vacancy defect (called nitrogen-vacancy center or NV− center) [19–21] (Fig. 1(a) ). FND has optical absorption centered at 560 nm due to its transition from the ground state (3A) to an excited state (3E), which has a broad emission band from 600 to 850 nm with a lifetime of 11.6 ns. A notable optical property of NV− center is its photo-stability with no photo-bleaching and little or no photo-blinking. This feature has enabled STED and GSD microscopy techniques with about ten nanometer resolution [22–24]. Another important feature of NV− center, which is of particular value for our technique, is its all-optically controlled photo-switching behavior  (Fig. 1(b)). Illuminating the NV− center with a red light at 638 nm excites it first to a bright, luminescent state then causes a transition (in a process called “shelving”) from the bright state to a dark state that persists with a lifetime of τd = 150 s, whereas a blue light at 473 nm induces a reverse transition (“de-shelving”). Both shelving and de-shelving processes occur in the sub-microsecond time scale with no photo-bleaching and little or no photo-blinking for over 800 seconds. Since the absorbance of NV− center at 473 nm is about one eighth that at absorption maxima (560 nm), the fluorescence intensity from NV− center under continuous 473 nm illumination is maintained at low level. Modulating one of the two beams modulates the emission intensity and clearly demonstrates that NV− center functions as an optical AND-gate, yielding significant light emission only when illuminated by both the blue and red lights simultaneously (Fig. 1(c)).
We propose a new method to overcome the diffraction limit of far-field fluorescence microscope by using the AND-gate property of FND. This method can be easily implemented and provide enhanced spatial resolution while keeping the advantages of confocal microscopy.
2. Design and experimental details
Concept of dual-point illumination
Our strategy for achieving sub-diffraction-limit spatial resolution is to reduce the effective point spread function (PSF) by spatially confining the emission from FND by illuminating it with two spatially-offset lights of different wavelengths for shelving and de-shelving, which leads to light emission only from the region of their overlap because of the optical AND-gate property of FND. A similar approach has been proposed by Sakai and Fujii using transient time-resolved IR spectroscopy based on two-color IR-UV/Vis double resonance spectroscopy [25,26]. The resonant IR light induces a vibrational transition and only the vibrationally excited molecule can absorb the UV/Vis light, yielding a transient fluorescence signal. In this scheme, fluorescence comes only from the region of overlap between the IR and UV/Vis lights, which is smaller than the diffraction limit of the IR light, but not of the UV/Vis light.
Our new method, named “dual-point illumination AND-gate microscopy on fluorescent nanodiamond” (DIAMOND in acronym), can readily reduce the effective PSF below the diffraction limit of visible light. If we were to apply dual-point illumination to a typical fluorescent dye in the scheme of Fig. 2(a) , fluorescence will result from all regions illuminated by either light, yielding an effective PSF in the shape of a dumbbell lying sideways. If, on the other hand, we adopt an optical AND-gate material such as FND, fluorescence will come only from the region of overlap between the two spatially-offset lights, resulting in an oval-shaped effective PSF with a greatly reduced lateral dimension. With this concept, one can easily achieve sub-diffraction-limited spatial resolution based on scanning optical microscope, without using a complex optical arrangement.
The full-width at half maximum (FWHM) of spatial overlap can be estimated by a simple calculation, assuming Gaussian-like profiles for each laser focal spot described by and , where P and F are respectively the laser power and FWHM of each laser (b for blue and r for red) and d is the inter-focal distance. We used an FWHM of 240 nm for the blue laser (473 nm) and 320 nm for the red laser (640 nm). The position of overlapped intensity maximum, xc, is a point of intersection and can be calculated as , where k is . The exact numerical value of maximum intensity can be calculated by using PSFb(xc) and PSFr(xc), and they have to give the same result, PSFmax. The FWHM can be determined by using the inverse functions of PSFb and PSFr, and , and the half of its maximum intensity, . Because we use the positive inter-focal distances for the calculations, the FWHM should be the difference between the rising edge of and falling edge of , which is given by . In this case, the spatial offset, i.e., the distance between the two illumination focal points, governs the FWHM of the effective PSF, leading to higher spatial resolution at larger spatial offset for an ideal AND-gate fluorophore and a Gaussian-like PSF of microscope (Fig. 2(b)).
The DIAMOND technique is entirely confocal-based and thus inherits all the advantages of confocal microscopy, not the least of which is its simplicity. Therefore, one can achieve sub-diffraction-limit spatial resolution in a quick, easy, convenient, and inexpensive way with no technical complexities often encountered in other such methods.
FND samples of 30 and 140 nm diameters were obtained from the Huan-Cheng Chang group of the Institute of Atomic and Molecular Sciences, Taiwan, and placed on the surface of a cover glass. Immobilization of FND was performed using the electrostatic interaction between poly-L-lysine (P8920, Sigma-Aldrich) and the carboxyl group on the surface of FND. The cover glass (BB024024A1, Menzel-Gläser) was washed with methanol (A452-4, Fisher Scientific), acetone (A949-4, Fisher Scientific), and deionized (DI) water and dried with nitrogen gas. 10 μL of poly-L-lysine was incubated on the surface of the cover glass for 3 minutes and washed with DI water following nitrogen gas drying. 10 μL of FND sample went through the same incubation and washing processes. An immersion oil (10976, Fluka) was used as a mounting medium between cover glass and slide glass (12-544-2, Fisher Scientific) to match the refractive index with the microscope.
The PSF of each focused laser spot and the distance between the two spatially-offset lights were measured by using a sample of 80-nm gold beads (EM.GC80, BB International), which were immobilized on the surface of a cover glass using their electrostatic interaction with poly-L-lysine. The immersion oil was used as a mounting medium. An identical preparation method was employed for the gold beads with that for the FND sample.
A home-built fluorescence confocal microscopy system was used to demonstrate the DIAMOND concept (Fig. 3 ). Lights from three continuous-wave excitation lasers (473 nm: 35-LAP-321, CVI Melles-Griot; 532 nm: SambaTM 532, Cobolt; 640 nm: TECRL-25G-635, World Star Tech) were transmitted through a single-mode fiber (Φ = 3 – 5 μm, P1-488-PM-FC and P1-630PM-FC, Thorlabs) coupling, and combined through dichroic mirrors (z473rdc, z532rdc and z647rdc, Chroma). Lasers were circularly polarized using achromatic λ/4 and λ/2 retarders (RAC 3.4.15, RAC 3.2.15 and RAC 4.2.15, Bernhard Halle Nachfl.) to efficiently illuminate the sample through an oil-immersion objective (NA 1.45, 60x, f = 3 mm, Olimpus). The sample was placed on a piezo-electric scanning system (NanomaxTM Max311, Thorlabs) with its own controller (BPC203, Thorlabs) to scan the imaging area with 5 nm accuracy. The fluorescence signal from FND was filtered through an emission filter (HQ700/75m, Chroma) and detected by an avalanche photodiode (SPCM-AQR-14FC, Perkin Elmer) through a multi-mode fiber (Φ = 62.5 μm, M31L02, Thorlabs) to enhance the signal-to-background (S/B) ratio, as does a pinhole in a confocal microscope. The detection window was allowed to span a tightly confined region of the intersection by this multi-mode fiber and detection lens (f = 1000 mm) to reduce the background from out of the intersection region. We used a multi-channel analyzer (P7882 photon counter, Fast ComTec) to convert the analogue signal to photon counts, which were then visualized by an imaging program, Imspector. All measurements were performed on 128x128 pixels covering various imaging areas, which are shown in each figure by a scale bar.
A key factor for DIAMOND in achieving the sub-diffraction-limit resolution is to enable precise spatial control of the illumination lasers. Two dichroic mirrors that reflect the 640-nm and 473-nm lasers were used to accomplish such a goal. By carefully adjusting the reflection angle of the dichroic mirrors, we were able to control the focal positions of the lasers to result in a spatial overlap with an arbitrary distance between the two focal points. The spatial offset and the PSF were identified by observing the scattered light from an 80-nm gold bead sample, which was detected by photomultiplier tube without multi-mode fiber coupling to clearly display the PSF at various values of spatial offset. We then observed the FND samples by confocal microscopy (532 nm) and DIAMOND (473 nm and 640 nm) using a tightly defined detection window to enhance the S/B ratio by reducing the fluorescence from out of the intersection region.
3. Results and discussion
DIAMOND with 140-nm FND sample
In order to demonstrate the feasibility of DIAMOND, we investigated whether the effective PSF is actually reduced using FNDs of different sizes. Since the position of the two focal points can be controlled in the imaging plane in both the x- and y- directions by adjusting dichroic mirrors, we can choose to have the spatial offset along one of the axes, in which case the effective PSF shrinks along that axis but not along the other axis (Fig. 4(d) ). The resulting effective PSF is an oval shape, whose short axis can be further reduced by increasing the spatial offset. The DIAMOND images shown in Fig. 4 (and those in subsequent figures) were obtained by continuously illuminating the two spatially-offset focal points with blue (473 nm, 40 μW) and red (640 nm, 3 mW) lasers, while the confocal microscopy images were taken by continuous green (532 nm, 160 μW) illumination. Since the NV− center is not an ideal AND-gate and thus emits weak fluorescence under blue or red illumination, these spurious signals contribute to the background and cause mild blurring of the image. We can readily subtract these spurious signals to improve the S/B ratio from 1: 0.6 to 1: 0.05.
Figure 4 shows that the effective PSF measured from single 140-nm FND is clearly reduced along one axis as we go from regular confocal microscopy (Fig. 4(a)) to DIAMOND (Fig. 4(d)). At the wavelength of confocal microscopy (532 nm), shelving and de-shelving processes occur simultaneously to result in a strong fluorescence emission. When we employ the DIAMOND scheme but use only the 640-nm light, shelving of the excited population to the dark state greatly reduces the fluorescence intensity to the level of background signal (Fig. 4(b)). On the other hand, when we use only the 473-nm light, de-shelving from the dark state can keep a significant population in the bright state, but the absorption cross section at this wavelength is so small in the first place that we have little fluorescence intensity (Fig. 4(c)). Finally, using both the 640-nm and 473-nm lights induces optical AND-gate emission from the region of their spatial overlap, yielding a typical DIAMOND image with an oval-shaped effective PSF (Fig. 4(d)). The intensity profiles also show that the FWHM of confocal PSF (302 nm, measured along the yellow dashed line) is significantly reduced to a sub-diffraction-limit value (195 nm) with DIAMOND (Fig. 4(e) and 4(f)).
The enhanced resolving power also allows unambiguous, discrete imaging of multiple particles in close proximity. For example, the particles along the yellow dashed lines of Fig. 5(a) and 5(b) are not totally distinguishable by confocal microscopy, but we can readily verify that they consist of two particles at a clear separation by DIAMOND imaging (Fig. 5(c)). In addition, all particles detected by confocal microscopy are also faithfully observed in DIAMOND without any missing or spurious additional images.
DIAMOND with 30-nm FND sample
The effective PSF can be further reduced if we employ smaller fluorescent particles. Figure 6(a) and 6(b) show the confocal and DIAMOND images of 30-nm FND single particle showing that the effective PSF of DIAMOND is 133 nm (FWHM), which is much smaller than that of confocal microscopy, 301 nm, and obviously under the diffraction limit (Fig. 6(e) and 6(f)). Such an enhanced effective PSF enables distinction of two adjacent FND particles (Fig. 6(d) and 6(g), at a separation of 280 nm), even when the fluorescence signal of one particle is much smaller than that of the other. In contrast, confocal microscopy gives only a blurred image and we cannot even figure out how many particles exist in the area (Fig. 6(c)).
2D-DIAMOND for symmetric images
Although DIAMOND offers many advantages such as instrumental simplicity and low demand for laser power over other sub-diffraction-limit microscopy techniques, the drawback for DIAMOND so far is its oval-shaped effective PSF due to the one-dimensional resolution enhancement that causes image distortion of the sample. In order to overcome this problem, we introduce 2D-DIAMOND imaging that gives a symmetric effective PSF (Fig. 7(a) and 7(b)) by convoluting two independent DIAMOND images obtained along two perpendicular axes (Fig. 7(c) and 7(d)). It is straightforward to control the offset direction of the focal points by simply adjusting the mirror mounts. The emitted photon count at a pixel in one offset was saved and compared with that of the other offset in the perpendicular direction, and the smaller photon count was selected to generate a new image, which is now symmetric. Since the fluorescence signal from FND is sufficiently larger than the thermal noise, signals can be distinguished from the background when we select the lower photon count, which means that this process does not allow missing any fluorescent particle, if we were to accept a lower S/B ratio. The experimentally measured effective PSF of 2D-DIAMOND with 30-nm FND has a FWHM of 154 nm in one direction and 159 nm in the other (Fig. 7(e) and 7(f)). As already mentioned, since our data processing protocol is optimized at not missing a fluorescent particle at the expense of a lower S/B ratio, these values are slightly larger than the FWHM of the effective PSF in 1D-DIAMOND images.
Figure 8(a) and 8(b) show that 2D-DIAMOND can also provide enhanced resolving power over confocal microscopy for a sample with multiple features, which in this case is a set of adjacent FND particles lying along the y axis. Using 1D-DIAMOND with a spatial offset along the x axis (Fig. 8(c)), the resolution enhancement appears only along that axis and thus the two particles remain unresolved, whereas an orthogonal spatial offset along the y axis reveals a completely resolved image for the two particles (Fig. 8(d)). For particles of any spatial arrangements, one can employ 2D-DIAMOND to enhance spatial resolution with a symmetric effective PSF, while keeping the resolving power in both dimensions. The fluorescence intensity profiles also clearly shows the reduced effective PSF and enhanced resolving power of 2D-DIAMOND (Fig. 8(e)).
Because we make a comparison of the fluorescence intensities at a given pixel of two 1D-DIAMOND images to generate a symmetric (i.e., circular, not oval-shaped) image, there is a possibility for a crosstalk from an off-center region during data processing. Although the fluorescence intensity due to the crosstalk should often be distinguishable because of its lower intensity than that of the fluorescent particles, the spurious signal from the crosstalk remains a challenging problem with the current 2-D scanning method.
In conclusion, we propose a new concept and experimental platform to overcome the diffraction limit in far-field optical microscopy by employing two laser beams with a spatial offset that illuminate an optical AND-gate material. We demonstrated its feasibility and sub-diffraction-limit nature by using different sizes of FND. With this technique, sub-diffraction-limit microscopy can be implemented in a quick, easy, convenient, and inexpensive way with no technical complexities often encountered in other methods. Furthermore, since FND is an ideal fluorescent material with high photo-stability, this new method may find its use in dynamic imaging over a long duration of time.
We would like to thank Professor Stefan W. Hell of Max Planck Institute for Biophysical Chemistry for kindly lending us an imaging program, Imspector. This work was supported by the National Research Foundation of Korea through the World Class University Program (R31-2010-100320), the Star Faculty Program (KR-2005-084-C00017), the Chemical Genomics Grant (M10526020002-08N2602-00210), and the Global Frontier R&D Program on Center for Multiscale Energy System.
References and links
1. E. Abbe, Gesammelte Abhandlungen (G. Fisher, Jana, 1904).
3. S. W. Hell and J. Wichmann, “Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy,” Opt. Lett. 19(11), 780–782 (1994). [CrossRef] [PubMed]
4. S. W. Hell and M. Kroug, “Ground-state depletion fluorescence microscopy: a concept for breaking the diffraction resolution limit,” Appl. Phys. B 60(5), 495–497 (1995). [CrossRef]
5. M. G. L. Gustafsson, “Nonlinear structured-illumination microscopy: wide-field fluorescence imaging with theoretically unlimited resolution,” Proc. Natl. Acad. Sci. U.S.A. 102(37), 13081–13086 (2005). [CrossRef] [PubMed]
7. E. Betzig, G. H. Patterson, R. Sougrat, O. W. Lindwasser, S. Olenych, J. S. Bonifacino, M. W. Davidson, J. Lippincott-Schwartz, and H. F. Hess, “Imaging intracellular fluorescent proteins at nanometer resolution,” Science 313(5793), 1642–1645 (2006). [CrossRef] [PubMed]
11. M. F. Juette, T. J. Gould, M. D. Lessard, M. J. Mlodzianoski, B. S. Nagpure, B. T. Bennett, S. T. Hess, and J. Bewersdorf, “Three-dimensional sub-100 nm resolution fluorescence microscopy of thick samples,” Nat. Methods 5(6), 527–529 (2008). [CrossRef] [PubMed]
12. G. Donnert, J. Keller, C. A. Wurm, S. O. Rizzoli, V. Westphal, A. Schönle, R. Jahn, S. Jakobs, C. Eggeling, and S. W. Hell, “Two-color far-field fluorescence nanoscopy,” Biophys. J. 92(8), L67–L69 (2007). [CrossRef] [PubMed]
14. H. Shroff, C. G. Galbraith, J. A. Galbraith, H. White, J. Gillette, S. Olenych, M. W. Davidson, and E. Betzig, “Dual-color superresolution imaging of genetically expressed probes within individual adhesion complexes,” Proc. Natl. Acad. Sci. U.S.A. 104(51), 20308–20313 (2007). [CrossRef] [PubMed]
15. B. Hein, K. I. Willig, and S. W. Hell, “Stimulated emission depletion (STED) nanoscopy of a fluorescent protein-labeled organelle inside a living cell,” Proc. Natl. Acad. Sci. U.S.A. 105(38), 14271–14276 (2008). [CrossRef] [PubMed]
17. H. Shroff, C. G. Galbraith, J. A. Galbraith, and E. Betzig, “Live-cell photoactivated localization microscopy of nanoscale adhesion dynamics,” Nat. Methods 5(5), 417–423 (2008). [CrossRef] [PubMed]
19. A. Gruber, A. Dräbenstedt, C. Tietz, L. Fleury, J. Wrachtrup, and C. von Borczyskowski, “Scanning confocal optical microscopy and magnetic resonance on single defect centers,” Science 276(5321), 2012–2014 (1997). [CrossRef]
20. F. Jelezko and J. Wrachtrup, “Single defect centres in diamond: a review,” Phys. Status Solidi., A Appl. Mater. Sci. 203(13), 3207–3225 (2006). [CrossRef]
21. C. C. Fu, H. Y. Lee, K. Chen, T. S. Lim, H. Y. Wu, P. K. Lin, P. K. Wei, P. H. Tsao, H. C. Chang, and W. Fann, “Characterization and application of single fluorescent nanodiamonds as cellular biomarkers,” Proc. Natl. Acad. Sci. U.S.A. 104(3), 727–732 (2007). [CrossRef] [PubMed]
22. E. Rittweger, K. Y. Han, S. E. Irvine, C. Eggeling, and S. W. Hell, “STED microscopy reveals crystal colour centres with nanometric resolution,” Nat. Photonics 3(3), 144–147 (2009). [CrossRef]
23. E. Rittweger, D. Wildanger, and S. W. Hell, “Far-field fluorescence nanoscopy of diamond color centers by groud state depletion,” Europhys. Lett. 86(1), 14001 (2009). [CrossRef]
24. K. Y. Han, S. K. Kim, C. Eggeling, and S. W. Hell, “Metastable dark States enable ground state depletion microscopy of nitrogen vacancy centers in diamond with diffraction-unlimited resolution,” Nano Lett. 10(8), 3199–3203 (2010). [CrossRef] [PubMed]
25. M. Sakai, Y. Kawashima, A. Takeda, T. Ohmori, and M. Fujii, “Far-field infrared super-resolution microscopy using picosecond time-resolved transient fluorescence detected IR spectroscopy,” Chem. Phys. Lett. 439(1-3), 171–176 (2007). [CrossRef]
26. K. Inoue, N. Bokor, S. Kogure, M. Fujii, and M. Sakai, “Two-point-separation in a sub-micron nonscanning IR super-resolution microscope based on transient fluorescence detected IR spectroscopy,” Opt. Express 17(14), 12013–12018 (2009). [CrossRef] [PubMed]