Letter
Vol. 42, No. 23 / December 1 2017 / Optics Letters
4885
Improved lateral resolution with an annular vortex depletion beam in STED microscopy BIN WANG,1,2 JINMENG SHI,1,2 TIANYUE ZHANG,2,3 XIAOXUAN XU,1 YAOYU CAO,2
AND
XIANGPING LI2,4
1
The Key Laboratory of Weak-Light Nonlinear Photonics, Ministry of Education, School of Physics, NanKai University, Tianjin 300071, China Guangdong Provincial Key Laboratory of Optical Fiber Sensing and Communications, Institute of Photonics Technology, Jinan University, Guangzhou 510632,China 3 e-mail:
[email protected] 4 e-mail:
[email protected] 2
Received 27 September 2017; revised 31 October 2017; accepted 31 October 2017; posted 1 November 2017 (Doc. ID 308122); published 22 November 2017
We report on the experimental demonstration of improved lateral resolution in stimulated emission depletion (STED) microscopy using an annular depletion beam configuration. A tight and finely tuned doughnut focal spot can be created by annular vortex illumination. Its application in STED microscopy for enhanced lateral resolution is systematically investigated by imaging 40 nm fluorescent beads. An improved resolution with more than 20% reduced effective point spread function of the imaging system determined by the full width at half-maximum compared to that of the conventional STED is achieved. The proposed scheme also demonstrates its resolving capability for biological samples. The principle holds great potential in the research fields of biological superresolution imaging as well as STED-based nanolithography and high-density optical data storage. © 2017 Optical Society of America OCIS codes: (100.6640) Superresolution; (180.2520) Fluorescence microscopy. https://doi.org/10.1364/OL.42.004885
Stimulated emission depletion (STED) microscopy as a powerful imaging method with the capability to break the diffraction limit has experienced rapid growth in recent years [1–4]. It has been widely used to resolve and visualize sophisticated details of structures, which benefits research fields such as biology and material science [5–7]. STED microscopy creates superresolution images by selectively deactivating the fluorophores, minimizing the area of effective point spread function (PSF) of the imaging system. Typically, a doughnut-shaped depletion beam is utilized overlapping the excitation laser beam in order to inhibit the fluorescence at the periphery of the excitation focus through stimulated emission, leaving highly confined central region active to emit fluorescence, thus compressing the actual size of the PSF [8,9]. STED microscopy uses PSF engineering to achieve high spatial resolution; therefore, strategies which involve generating a perfect doughnut profile and shrinking 0146-9592/17/234885-04 Journal © 2017 Optical Society of America
the zero-intensity central region of the depletion beam are essential to improve the performance of STED imaging. On the other hand, the application of annular apertures in light microscopy has been nicely performed in many early studies for engineering PSF [10–16]. This is primarily due to two major effects of the annular apertures on the PSF of the lens: the main maximum central lobe of the PSF is sharpened in the focal plane, and the axial extent of the PSF is elongated along the optical axis [10]. Both effects depend on the size of the annulus with respect to the full aperture. Such a concept can be applied to STED microscopy, providing a simple and efficient way to tailor the profile of the focus of the depletion beam. The sharpening of the lateral PSF increases the lateral resolution of the microscope, while the increased axial PSF enlarges the focal depth, which is particularly beneficial in STED deep imaging [17]. Various studies have presented the theoretical groundwork on focusing properties of vectorial light beams with different polarization conditions to form an ideal doughnut beam [18–26]. However, the experimental implementation of manipulating the focal spot to engineer the PSF of a STED microscope is still very lacking. In this Letter, we present the experimental demonstration of the improved lateral resolution of STED utilizing an annular amplitude mask to tailor the depletion beam profile. Because of the narrowed doughnut focal spot as a result of center-blocking annular vortex illumination, the effective PSF of the system is compressed with its full width at half-maximum (FWHM) reduced by more than 20% compared to that of the conventional STED. The ability to enhance the spatial resolution with annular depletion illumination is verified by imaging 40 nm fluorescent beads and also biological specimens. The scheme of our STED imaging system with amplitude manipulation is illustrated in Fig. 1(a). A pulsed diode laser (Göttingen, Germany) at 640 nm was employed for Gaussian excitation. The excitation spot was overlaid with a timedelayed, doughnut-shaped depletion beam at the wavelength of 775 nm for molecular de-excitation. The depletion laser beam was converted into a vortex beam after passing through a phase plate (RPC Photonics, VPP-1a) and then converted into an annular beam by an opaque mask before being focused
4886
Vol. 42, No. 23 / December 1 2017 / Optics Letters
Letter
Fig. 1. (a) Schematic illustration of modulating the depletion beam into the annular illumination configuration in a STED system. (b) Comparison of calculated normalized line profiles of the effective PSF in STED and A-STED imaging systems. The normalized line profiles of Gaussian excitation PSF and the depletion PSF for A-STED/STED on the focal plane are also plotted. The inset shows the electric field distributions for the focused depletion beam.
by the objective lens (Olympus, 100×, oil immersion, NA 1.4). The ratio of inner to outer radius of the annular depletion beam is defined as ε r 1 ∕r 2 , as depicted in the inset in Fig. 1(a). By varying the radius parameter ε, the influence of amplitude modulation on the STED imaging performance can be evaluated systematically. Such annular vortex depletion STED configuration (referred to as A-STED for short) can be easily restored to standard STED microscopy by translating out the amplitude mask. We first theoretically investigate the focusing behavior of the annular vortex depletion beam. A circularly polarized vortex beam is a circularly polarized beam with a spiral phase wavefront. We consider p a ffiffiffileft-handed circularly polarized beam E LHC e x ie y ∕ 2 transmitting through a vortex 0–2π phase plate with the phase modulation of e iφ . When focused by a high NA objective, based on the Richards–Wolf vectorial diffraction theory, all vectorial light fields interfere destructively at the focal point and this creates the minimum field at the center of the focus leaving a doughnut pattern. The focal field distribution can be expressed as [27,28] Z α pffiffiffiffiffiffiffiffiffiffi 2iφ s E ρ ρs ; φs ; z s −Ce cos θ sin θ expikz s cos θ ×
β
cos θJ 3 kρs sin θ − J 1 kρs sin θ dθ; −J 3 kρs sin θ J 1 kρs sin θ (1)
Z E φ ρs ; φs ; z s iCe 2iφs ×
α pffiffiffiffiffiffiffiffiffiffi
cos θ sin θ expikz s cos θ
β
cos θJ 3 kρs sin θ J 1 kρs sin θ dθ; −J 3 kρs sin θ − J 1 kρs sin θ (2)
Z E z ρs ; φs ; z s 2iCe 2iφs
α β
pffiffiffiffiffiffiffiffiffiffi cos θ sin2 θ expikz s cos θ
× J 2 kρs sin θdθ:
(3)
When an annular amplitude modulation is introduced, this is considered in the mathematical treatment by setting the lower limit of the diffraction integral to a value β arc sinε × NA∕n, where n is the refractive index of the medium. The parameters used in calculations are kept the same
to that used in the experiment, with n 1.52 for the dielectric environment, NA 1.4 for the objective lens, and 640 nm/ 775 nm for the excitation and the depletion laser wavelength, respectively. From the numerical calculations shown in Fig. 1(b), it is evident that the annular vortex depletion beam maintains intensity-zero at the center of the focus and a tighter doughnut focal spot is formed compared to that of the unmodulated beam. Figure 1(b) shows the intensity profiles of the excitation PSF, the depletion PSF with and without annular modulation for comparison. The electric field distributions for the focused depletion beam are shown in the inset. The effective PSF of the STED microscope h_eff is given by [9,29] h_eff h_exc exp−σh_dep Φmax ;
(4)
with h_exc and h_dep describing the excitation PSF and depletion PSF. σ is the is the cross section for stimulated emission at the used wavelength. The flux Φmax gives the number of photons per area per pulse found at the doughnut crest. The equation above explicitly emphasizes that the quality of the doughnut pattern of the depletion beam which overlaps with the excitation beam determines the ultimate resolution of the system. Our results confirm that in the presence of the amplitude mask, a tighter doughnut pattern which creates a shaper intensity gradient around the central region can turn off the fluorescence more efficiently, resulting in a smaller effective PSF in the lateral dimension. In the experiments, we visualized the annular vortex depletion PSF by imaging gold nanoparticles with the size of ∼150 nm. To check the beam profiles, we scanned gold nanoparticles dispersed onto a microscope coverslip and imaged the backscattered signal on a photomultiplier tube. To avoid fluctuations among single nanoparticles with slightly variant sizes and shapes, the PSF images were averaged over tens of gold nanoparticles. Different annular masks were applied, with corresponding ε equaled to 0.25, 0.6, and 0.75. The central minimum FWHM of the h_dep profile was served as the characteristic of the doughnut. As shown in Fig. 2, the FWHM of the line profile was remarkably suppressed from 222 nm to 196 nm when ε increased from 0 to 0.75. Our measurements demonstrate that increasing the inner radius of annular beam leads to reduced lateral doughnut size, which is in good agreement with numerical calculations [Fig. 2(b)]. For superresolution imaging, the spatial overlapping between the depletion beam and the excitation beam were checked by PSF images of gold nanoparticles, as shown in Figs. 3(a)–3(c). The lateral resolution of our A-STED setup is thus determined by the effective fluorescence emitting area suppressed by the doughnut-shaped annular vortex depletion beam. To examine the performance of A-STED, 40 nm sized crimson fluorescent beads immobilized on a coverslip were used. For fluorescence imaging, the signal was collected by the same objective [shown in Fig. 1(a)] passed through a bandpass filter (Thorlabs, ET685/70M) before being focused into an 8 μm diameter single-mode fiber as a confocal pinhole connected to a photon-counting avalanched photodiode (Abberior Instruments). Confocal, conventional STED, and A-STED were carried out to record the same imaging area for intuitive comparison. Figures 3(d) and 3(e) reveal an impressive resolution enhancement with A-STED (ε 0.75) under various depletion powers. When we zoom in the rectangle area and
Letter
Vol. 42, No. 23 / December 1 2017 / Optics Letters
4887
advantage over conventional STED that a better resolution is achieved under the same depletion peak illuminance. This is particularly useful when imaging biological specimens which are sensitive to the photobleaching effect. To demonstrate this advantage, we performed the experiments on STAR635P labeled vimentin (Abberior) under the relative low depletion intensity. Figure 4 shows that although the STED image can render much more detail than the vague confocal image, more features have been resolved in the A-STED image through the amplitude modulation. The effect is also clearly visible in the cross sections plotted in Figs. 4(b) and 4(c), where two represented areas are chosen. We measured the apparent
Fig. 2. PSF of depletion beams visualized by imaging 150 nm gold nanoparticles. (a) Intensity distributions and their cross sections for ε 0, 0.25, 0.6, and 0.75, respectively. The scale bar is 2 μm. (b) The central minimum FWHM of the doughnut as a function of ε with the solid line and discrete point data plotting the simulation and measured results, respectively.
have a closer look, this improvement is even more obvious. It is clear that the fluorescent beads that are blurred out by the diffraction limit in the confocal imaging, can be individually distinguished in the STED imaging. The resolution is further improved by A-STED, which is evident from measured FWHM values averaged from several beads, yielding 70 9 nm for STED and 51 6 nm for A-STED imaging with the depletion power at the radiation peak intensity of 33 GW∕cm2 . It is noteworthy that a tighter lateral confinement is achieved by the annular illumination that blocks the center portion of the depletion beam. However, this is obtained at the expense of increased sidelobes. And the sidelobes become more pronounced along with the increase of ε, as already shown in Fig. 2. Therefore, to compensate the power loss due to the beam blocking, we raised the depletion laser output power to make sure the measured peak intensity of doughnut main maximum at the focal plane in A-STED was the same value compared to that of regular STED. This treatment is reasonable as the resolution of STED microscopy scales inversely with the square root of the ffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi depletion intensity, described by d ≈ λ∕2 NA 1 I ∕I sat , where I sat is the effective saturation intensity at which the fluorescence has dropped to 1∕e of the initial value [8,29]. Applying a greater power further squeezes the spot diameter (i.e., FWHM) to subdiffraction dimensions following the above square-root law. The features demonstrated by the experimental results are consistent with the calculation data displayed by Fig. 3(e). More measurements were taken to evaluate the resolution enhancement for other ε conditions, as shown in Fig. 3(f ). To this end, we confirm that under the same illumination intensity of the depletion laser, A-STED exhibits higher resolving ability in comparison with regular STED. It is well known that the resolution of STED microscopy can be enhanced by increasing the intensity of the STED depletion beam, which inevitably increases the risk of photodamage to the sample. Therefore, A-STED exhibits its
Fig. 3. (a)–(c) Spatial overlapping between the depletion and the excitation beam checked by PSF images of gold nanoparticles. The line profiles are also plotted for the marked nanoparticle. (d) Imaging of 40 nm sized crimson fluorescent beads via confocal, STED, and A-STED with the excitation at 640 nm, the depletion at 775 nm under the depletion intensity at 20 GW∕cm2 . The scale bar is 500 nm. (e) Averaged resolution (FWHM of the fluorescence spot) of 15 beads via STED and A-STED imaging as a function of the depletion beam radiation peak intensity. Solid lines are the corresponding calculated resolution (FWHM of the effective PSF h_eff ). The inset shows the compared zoom-in images for STED and A-STED. (f ) The influence of various annular parameters ε on the resolution in A-STED configuration.
4888
Vol. 42, No. 23 / December 1 2017 / Optics Letters
Letter of Guangdong Province (2017A030313006); Innovation and Entrepreneurship Team Project (2016ZT06D081); Tianjin science and technology plan project (15ZCZDGX00780). REFERENCES
Fig. 4. (a) Images of the biological sample vimentin taken with confocal, STED, and A-STED microscope, respectively. The scale bar is 2 μm. Two represented areas are chosen to make the line profile analysis. (b) Comparison of cross sections for an individual vimentin structure imaged by STED and A-STED. Discrete circle points are measured intensity profile and the solid lines are Gaussian fit. (c) STED and A-STED image-based cross sections for vimentin filaments located closely in space.
width of an individual vimentin structure in the image [indicated by line 1 in Fig. 4(a)] by performing Gaussian fit on its cross section and found the FWHM to be 304 nm for the confocal image, 180 nm for STED, and 137 nm for the A-STED image [Fig. 4(b)] under the depletion intensity at 13.5 GW∕cm2 . Figures 4(a) and 4(c) manifest that for vimentin filaments in close proximity to each other (area 2), A-STED provides higher resolving ability to separate the individual thin vimentin. In conclusion, we have demonstrated the enhanced resolution of STED imaging achieved by an annular vortex depletion beam theoretically and experimentally. Applying an annular mask provides a straightforward and easy-to-implement method to finely reshape the depletion beam. Taking advantage of the tighter doughnut pattern, the effective PSF of STED microscopy is further compressed. The ability to enhance the lateral resolution with the annular depletion illumination is fully investigated by imaging 40 nm sized fluorescent beads. The proposed scheme also demonstrates its resolving capability for biological specimens. Combined with other techniques, such as phase and polarization modulation, more advanced focal spot manipulation and PSF engineering can be realized, which holds the potential for applications such as STED-based nanofabrication and optical storage [15,16,30–33]. Funding. National Natural Science Foundation of China (NSFC) (61432007, 61522504); Natural Science Foundation
1. S. W. Hell and J. Wichmann, Opt. Lett. 19, 780 (1994). 2. S. W. Hell, Science 316, 1153 (2007). 3. J. G. Danzl, S. C. Sidenstein, C. Gregor, N. T. Urban, P. Ilgen, S. Jakobs, and S. W. Hell, Nat. Photonics 10, 122 (2016). 4. H. Blom and J. Widengren, Chem. Rev. 117, 7377 (2017). 5. K. I. Willig, S. O. Rizzoli, V. Westphal, R. Jahn, and S. W. Hell, Nature 440, 935 (2006). 6. L. Schermelleh, R. Heintzmann, and H. Leonhardt, J. Cell. Biol. 190, 165 (2010). 7. E. Rittweger, K. Y. Han, S. E. Irvine, C. Eggeling, and S. W. Hell, Nat. Photonics 3, 144 (2009). 8. V. Westphal and S. W. Hell, Phys. Rev. Lett. 94, 143903 (2005). 9. K. Willig, J. Keller, M. Bossi, and S. Hell, New J. Phys. 8, 106 (2006). 10. S. W. Hell, P. E. Hänninen, A. Kuusisto, M. Schrader, and E. Soini, Opt. Commun. 117, 20 (1995). 11. L. Liu, X. Deng, L. Yang, G. Wang, and Z. Xu, Opt. Lett. 25, 1711 (2000). 12. M. A. A. Neil, R. Juškaitis, T. Wilson, Z. J. Laczik, and V. Sarafis, Opt. Lett. 25, 245 (2000). 13. M. Martínez-Corral, A. Pons, and M. A.-T. Caballero, J. Opt. Soc. Am. A 19, 1532 (2002). 14. R. Dorn, S. Quabis, and G. Leuchs, Phys. Rev. Lett. 91, 233901 (2003). 15. X. Li, Y. Cao, and M. Gu, Opt. Lett. 36, 2510 (2011). 16. X. Li, T. H. Lan, C. H. Tien, and M. Gu, Nat. Commun. 3, 998 (2012). 17. W. Yu, Z. Ji, D. Dong, X. Yang, Y. Xiao, Q. Gong, P. Xi, and K. Shi, Laser Photon. Rev. 10, 147 (2016). 18. N. Bokor and N. Davidson, Opt. Lett. 31, 149 (2006). 19. X. Hao, C. Kuang, T. Wang, and X. Liu, J. Opt. 12, 115707 (2010). 20. Y. Xue, C. Kuang, S. Li, Z. Gu, and X. Liu, Opt. Express 20, 17653 (2012). 21. X. Li, P. Venugopalan, H. Ren, M. Hong, and M. Gu, Opt. Lett. 39, 5961 (2014). 22. Q. Zhan, Opt. Lett. 31, 867 (2006). 23. Q. Zhan, Adv. Opt. Photon. 1, 1 (2009). 24. P. Zhang, P. Goodwin, and J. Werner, Opt. Express 22, 12398 (2014). 25. X. Hao, E. S. Allgeyer, M. J. Booth, and J. Bewersdorf, Opt. Lett. 40, 3627 (2015). 26. Y. Li, L. Lai, C. Rui, and L. Wang, Opt. Commun. 372, 132 (2016). 27. M. Gu, Advanced Optical Imaging Theory (Springer, 2000), pp. 143– 176. 28. Y. Jiang, X. Li, and M. Gu, Opt. Lett. 38, 2957 (2013). 29. B. Harke, J. Keller, C. K. Ullal, V. Westphal, A. Schönle, and S. W. Hell, Opt. Express 16, 4154 (2008). 30. X. Li, Y. Cao, N. Tian, L. Fu, and M. Gu, Optica 2, 567 (2015). 31. A. K. Thomas, W. Richard, and J. Jaroslaw, Phys. Scr. T162, 014049 (2014). 32. J. Fischer, J. B. Mueller, A. S. Quick, J. Kaschke, C. Barner-Kowollik, and M. Wegener, Adv. Opt. Mater. 3, 221 (2015). 33. B. Buchegger, J. Kreutzer, B. Plochberger, R. Wollhofen, D. Sivun, J. Jacak, G. J. Schütz, U. Schubert, and T. A. Klar, ACS Nano 10, 1954 (2016).