- Email: [email protected]
Super-resolution microscopy based on fluorescence emission difference of cylindrical vector beams
Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Q1 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Contents lists available at ScienceDirect
Optics Communications journal homepage: www.elsevier.com/locate/optcom
Super-resolution microscopy based on fluorescence emission difference of cylindrical vector beams Zihao Rong a, Cuifang Kuang a,n, Yue Fang a, Guangyuan Zhao a, Yingke Xu b, Xu Liu a a b
State Key Laboratory of Modern Optical Instrumentation, Department of Optical Engineering, Zhejiang University, Hangzhou, 310027, China Key Laboratory of Biomedical Engineering of Ministry of Education, Department of Biomedical Engineering, Zhejiang University, Hangzhou 310027, China
art ic l e i nf o
a b s t r a c t
Article history: Received 18 March 2015 Received in revised form 22 May 2015 Accepted 24 May 2015
We propose a novel fluorescence emission difference microscopy (FED) system based on focusing cylindrical vector beams. In conventional FED, a Gaussian beam and a 0–2π vortex phase plate are used to generate solid and hollow spots. We focus radially polarized and azimuthally polarized cylindrical vector beams to obtain an expanded solid spot and a shrunken hollow spot, taking advantage of the optical properties of cylindrical vector beams to improve the conventional FED performance. Our novel method enhances FED performance because the hollow spot size determines the FED resolution and an expanded solid spot effectively reduces negative side-lobe emergence during image processing. We demonstrate improved performance theoretically and experimentally using an in-house built FED. Our FED achieved a resolution of less than λ/4 in test images of 100 nm nanoparticles, better than the confocal image resolution by a factor of approximately 1/3. We also discuss detailed simulation analyses and FED imaging of biological cells. & 2015 Published by Elsevier B.V.
Keywords: Super-resolution Fluorescence imaging FED microscopy Cylindrical vector beam
1. Introduction Owing to its simplicity, versatility, and noninvasiveness, farfield fluorescence microscopy has long been applied in biological and medical science to observe and investigate microstructures and their movements [1]. However, the diffraction barrier [2] confines the resolving ability of conventional far-field fluorescence microscopy above about half the illumination light wavelength, restricting observation of microstructures featuring length scales less than 100 nm, such as vimentin fibers [3], vesicles, and microtubules. Confocal scanning laser microscopy (CSLM) [4] can enhance spatial resolution by a factor of 2 and improve micrograph contrast by spatial filtering with a pinhole, which also endows CSLM with optical sectioning ability [5]. Nonetheless, CSLM resolution is still limited by the diffraction barrier. Fluorescence emission difference (FED) [15,16] microscopy was recently reported as a novel super-resolution technique based on intensity subtraction [17] between two images acquired under different illumination patterns. It joins a long list of validated super-resolution fluorescence microscopy methods, e.g., photoactivated localization microscopy (PALM) [6], stochastic optical reconstruction microscopy (STORM) [7], stimulated emission depletion microscopy (STED) [8], structured illumination microscopy n
Corresponding author. E-mail address: [email protected] (C. Kuang).
(SIM) [9], super-resolution optical fluctuation imaging (SOFI) [10], total internal reflection fluorescence microscopy (TIRF) [11], and others [12–14]. Super-resolution techniques can be classified into two categories according to underlying principles and resolving abilities: diffraction unlimited and limited. Diffraction-unlimited techniques, including PALM, STORM, and STED, can yield resolution far beyond the diffraction barrier and overcome the barrier in true sense. Diffraction-limited techniques, including SIM, FED, SOFI, and TIRF, are fundamentally limited by the diffraction barrier. SIM, SOFI, and FED can enhance the resolution by a factor of 2 at most, while TIRF achieves super-resolution in the axial direction, but does not enhance lateral resolution. Subtractive imaging was invented decades ago to enhance image resolution by subtracting confocal signals taken at different collection pinhole sizes under the same illumination pattern [18,19]. However, its signal-to-noise ratio (SNR) is relatively low since final resolution-enhanced image construction uses signals emitted from the periphery of the excitation spot. To overcome this, fluorescence emission difference microscopy (FED) [15] and switching laser mode microscopy (SLAM) [12,20] were proposed as subtractive imaging alternatives that apply different illumination patterns with the same collection pinhole size. In FED, the sample is alternately scanned by a solid excitation spot and a doughnut-shaped excitation spot [21], and central signals are extracted by eliminating peripheral signals through subtraction [22]. The resolution and SNR of FED can be further enhanced through beam modulation [23–25] to decrease the solid and hollow spot
http://dx.doi.org/10.1016/j.optcom.2015.05.057 0030-4018/& 2015 Published by Elsevier B.V.
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
sizes. However, image subtraction inevitably produces negative side-lobes in the FED point spread function (PSF) under any subtractive factor, since the PSF profiles of the two illumination modes do not perfectly match [26]. This mismatch may introduce artifacts and deteriorate image quality. To reduce negative side-lobes and eliminate image deformation, the PSF profiles of the two illumination patterns were approximately matched by beam modulation [25,26]. Kobayashi et al. numerically studied the FED subtractive factor r, offering considerable selection criteria for it [27]. Moreover, image subtractive techniques can be combined with other super-resolution methods such as TIRF [28] and STED [17] to further enhance their performance. Since FED is, in principle, built upon CSLM, its attainable resolution is determined by the size of the two micro-imaging mode PSFs, not the diffraction barrier [23]. Moreover, FED achieves image resolution beyond the diffraction barrier at low laser intensity and relatively high imaging speed in a simple and compact system, which undoubtedly adds to its availability in practical scientific research [29]. Cylindrical vector beams, i.e., radially polarized beams and azimuthally polarized beams, have attracted a lot of interest due to their unique polarization states and prominent beam properties, and have prospective applications in technologies such as laser cutting, particle acceleration and optical microscopy [30]. Research show that a radially polarized field can be focused to a spot size of 0.16λ, significantly smaller than 0.26λ for a linear polarized field [31]; an azimuthally polarized field can similarly be focused to a smaller spot size than the hollow spot created by modulating a linear polarized field through a 0–2π vortex phase plate. However, tightly focusing a radially polarized beam by a high NA objective lens enlarges the solid spot size because of a focal lateral doughnut-shaped component that broadens the entire radially polarized beam PSF. We establish a novel FED system that capitalizes on these features of cylindrical vector beams to enhance performance. We established a cylindrical vector beam FED system in which a radially polarized beam and an azimuthally polarized beam create an extended solid spot and a shrunken hollow spot, respectively, at the focal plane by tight focusing. Simulations demonstrated significant performance enhancement by cylindrical vector beam FED and experiments achieved a resolution less than 120 nm.
2. Theory In FED, two different confocal scanning images are required to obtain the final FED image: the confocal image is acquired under the solid excitation pattern and the negative confocal image is acquired under the hollow excitation pattern [15]. The excited fluorescence is filtered by a pinhole and detected by a photomultiplier tube (PMT) to form both images. The final FED superresolution image is constructed by mathematical intensity subtraction of the two images:
IFED = Isolid − r × Ihollow
(1)
Here, IFED , Isolid , and Ihollow are the normalized intensity distributions of the FED, confocal, and negative confocal images, respectively, and r is the subtractive factor. Negative intensity differences, which are inevitable after subtraction, are excluded from the final image for better image quality. The resolving ability of FED is determined by the PSFs of the two illumination modes, which can be described by the Debye integral [23],
E (r2, φ 2, z 2 ) = iC
⎡ px ⎤ ⎢ ⎥
∫ ∫Ω sin θ⋅A1 (θ, φ)⋅A2 (θ, φ)⋅⎢ py ⎥⋅exp [if (θ, φ)] ⎢p ⎥ ⎣ z⎦
⎡ exp ⎣ikn z 2 cos θ + r2 sin θ cos θ φ−φ 2
(
(
))
⎤ ⎦ dθ dφ
(2)
E (r2, φ2, z2 ) is the electric field vector at the point (r2, φ2, z2 ) in cylindrical coordinates relative to an origin at the focal point of the objective lens. (θ , φ) represents angular position on the wave front of the incident beam, where θ is the angle between the ray direction and the optical axis and φ is the azimuthal angle. Ω represents the effective incident aperture of the beam, A1 (θ , φ) is the amplitude function of the input light, and A2 (θ , φ) is the aberration function determined by the structure of the objective lens. ⎡p , p , p ⎤T is the polarization state of the incident beam and ⎣ x y z⎦ f (θ , φ) represents the phase modulation function applied to the input light. From Eq. (2), tuning the polarization state, phase distribution, and amplitude distribution can optimize the PSF size of both imaging modes to improve the FED performance [23]. We used cylindrical vector beams to generate two patterns of illumination. Unlike conventional linearly polarized beams or circularly polarized beams, cylindrical vector beams feature axisymmetric and anisotropic polarizations. Axisymmetric and anisotropic polarization have invariant angles between the electric field vector direction and the radial direction throughout the beam cross-section [32]. Fig. 1(a) and (b) shows that the electric field vector directions of radially polarized and azimuthally polarized beams are always parallel and perpendicular to the radial direction, respectively. These polarization properties cause a tightly focused radially polarized beam or azimuthally polarized beam to create a solid spot or hollow spot on the focal plane, respectively. When concentrated by a high NA objective lens, radial polarization direction is deflected, generating a longitudinal electric field component. The lateral electric field component then coexists with a strong longitudinal electric field component at the focal area. The intensity distributions of both components' focal areas follow Eq. (2). The lateral electric component appears as a hollow spot with a central valley, and the longitudinal electric component presents as a solid spot with a central peak. The hollow lateral component expands the solid PSF wider than a Gaussian beam; this expansion reduces the negative values resulting from image subtraction. A characteristic that we do not utilize here is the high strength of the longitudinal electric component relative to the lateral one, and its smaller spot size relative to linearly or circularly polarized beams [31]; a sharper solid focal spot can be obtained by extracting the lateral component. The azimuthal polarization directions remain unchanged by focusing with a high NA objective lens, therefore no longitudinal electric component is generated. According to vector beam diffraction theory, the azimuthally polarized beam only possesses a lateral electric component at the focal area and forms a hollow spot smaller than a Gaussian beam modulated through a 0–2π vortex phase plate [30]. Our FED PSF is established by subtracting the hollow spot generated by the azimuthally polarized beam scaled by a subtractive parameter from the solid spot generated by the radially polarized beam, as in Eq. (1).
3. Simulation We first demonstrate the resolving ability and performance enhancement of our cylindrical vector beam FED by simulating both the conventional FED PSF and the new cylindrical vector beam FED PSF. The conventional FED PSF is generated by subtracting a hollow excitation pattern modulated by a 0–2π vortex phase plate from a solid confocal excitation pattern of a focused
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Fig. 1. Schematic of cylindrical vector beam polarizations and corresponding focal spots. (a) Radially polarized beams and (b) azimuthally polarized beams form (c) solid and (d) hollow focal spots, respectively, after tightly focusing.
3
Fig. 2(d) and the conventional FED PSF profile in Fig. 2(f) share the same minimal negative value of 0.27, but the cylindrical vector beam FED features a narrower PSF than the conventional FED, demonstrating its theoretical super resolving ability. We performed imaging simulations of a 3 3 point lattice sample to verify the performance enhancement. As shown in Fig. 2 (g), the sample is composed of nine points in a square distribution, where the four points at the corners (colored in white) have twice the intensity of the other points, and vertically and horizontally adjacent points are separated by a distance of 0.3λ. In the simulated confocal image of Fig. 2(h), the four corner points are discernible, but the other points are undetectable against the periphery signal of the corner points. In the conventional FED image (r ¼0.7) of Fig. 2(i), the four corner points appear more distinct, but the other five points are nearly missing due to excessive subtraction. However, in the cylindrical vector beam FED image of Fig. 2 (j), all nine points are distinguishable, despite some acceptable artifacts. Defocusing may cause inevitable artifacts in FED images. However, the out-of-focus dispersed solid spot shrinks when a similarly dispersed hollow spot is subtracted; thus, defocusing artifacts can be reduced. Therefore, FED suffers from fewer defocusing artifacts than confocal microscopy.
4. Experiment Gaussian left-handed circular beam. The new FED PSF is generated by subtracting the hollow pattern of a focused azimuthally polarized beam from the solid pattern of a radially polarized beam focused by a high NA objective. A subtractive parameter is needed in both situations. The PSFs are calculated by Debye integrals using self-developed Java software and the following integral parameters: the excitation light wavelength λ is 488 nm; the objective lens NA is 1.4; the pinhole is considered infinitely small; and the subtractive factor r is 0.7. We study a 2λ 2λ lateral field of view centered at the focus and normalize intensity for simplicity. Fig. 2(a–f) illustrates the PSFs of conventional FED and cylindrical vector beam FED, and demonstrates cylindrical vector beam FED feasibility and image enhancement relative to confocal microscopy and conventional FED. In Fig. 2(a), the red solid curve represents the Gaussian beam PSF, while the radially polarized beam PSF and its lateral and longitudinal components are denoted by the blue solid, dotted, and dashed curves, respectively. The final solid PSF results from summing the lateral component and the longitude component, which extends the PSF’s size, thus reducing the negative value area created by subtracting the hollow PSF from this solid PSF to eliminate deformations. As shown in Fig. 2(b), the sizes of the 0–2π vortex beam PSF and the azimuthally polarized beam PSF were 0.28λ and 0.24λ, respectively. Thus, the PSF of the azimuthally polarized beam provides a 14% smaller hollow pattern than the 0–2π vortex beam does. As shown in Fig. 2(c–f), using different subtractive factors to subtract 0–2π vortex beam PSFs from Gaussian beam PSFs, in conventional FED, and azimuthally polarized beam PSFs from radially polarized beam PSFs, in the new FED, can give narrower FED PSFs and stronger resolving ability relative to confocal microscopy. Moreover, using cylindrical vector beams reduces the minimal value of negative side-lobes in the new FED by a factor of approximately 1/3 for r ¼0.9, and even more for lower r. In Fig. 2(f), the new FED PSF has no negative values for r ¼0.4, whereas the conventional FED PSF has obvious negative side-lobes. While the new FED PSF FWHM does not dramatically narrow in comparison with that of the conventional FED PSF for the same subtractive factor, the new FED can enhance the image resolution for the same minimal negative values. The new FED PSF profile in
4.1. Cylindrical vector beam generation We use an S-waveplate (Altechna, Lithuania) to generate cylindrical vector beams. An S-waveplate is a super-structured polarization converter that converts linear polarizations to radial or azimuthal polarizations. Typically, the center of the polarization converter should be aligned with the optical axis of the incident linearly polarized beam. An alignment mark on the S-waveplate is aligned parallel to the incident linear polarization orientation for radial polarization, or perpendicular for azimuthal polarization. Thus, the incident light polarization can be chosen by rotating a λ/2 waveplate, and the polarization state of the output beam can be controlled by rotating the converter placed after the waveplate. However, this method suffers from human error and inefficiency. Therefore, we split the incident laser into two arms featuring linear polarizations parallel (p-polarization) and perpendicular (s-polarization) to the experimental platform, respectively, so that the polarization state of the output beam can be controlled by switching the incident beam with shutters. 4.2. Experimental schematic We built a cylindrical vector beam FED experimental setup, as illustrated in Fig. 2. A collimating lens (CL1) collimates the 488 nm wavelength laser beam emitted by a laser diode (Coherent OBIS 488) and transmitted through a single-mode fiber. The collimated beam is split by a polarizing beam splitter (PBS, B. Halle GmbH, Berlin, Germany) into two paths, a p-polarization beam path and an s-polarization beam path. A λ/2 waveplate (λ/2 WP) tunes the intensity distribution between the p-polarization beam and the s-polarization beam. A second PBS reintegrates the split beams. Shutters inserted into each light path allow for the p-polarization beam and the s-polarization beam to alternately pass through the key S-waveplate (Altechna RPC-488-06). We set the alignment mark direction of the S-waveplate perpendicular to the experimental platform so that S-waveplate modulates the s- and p-polarization beams into radially and azimuthally polarized beams, respectively. After polarization modulation, the cylindrical vector
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 Q2 59 60 61 62 63 64 65 66
Fig. 2. Simulation PSFs of conventional FED and cylindrical vector beam FED. (a) PSFs of confocal illumination patterns (RB ¼radially polarized beam). (b) PSFs of negative confocal illumination patterns. (c–f) PSFs of confocal microscopy and conventional and cylindrical FED using subtractive factors r¼ 0.9, 0.7, 0.5 and 0.4, respectively. (g) Sample 3 3 point lattice. (h) Simulated confocal image. (i) FED image with r¼ 0.7. (j) Cylindrical vector beam FED image with r¼ 0.7.
beam is reflected by a galvo mirror (Thorlabs GVS002) controlled by a DAQ card (NI USB-6259) to realize two-dimensional fast scans of the sample. The scanning lens (SL) creates an intermediate image before the tube lens (TL), parallelizing the laser light. A high NA objective lens (Olympus Uplan SApo 100 /1.4 Oil) strongly focuses the cylindrical vector beam and collects the fluorescence emitted from the sample. The dichroic mirror (Di02-R488-25*36, Semrock) reflects the illumination while transmitting the fluorescence, which is then spatially filtered by a pinhole, coupled into a multi-mode fiber (M31L02, Thorlabs), and finally detected by an avalanche photon diode (APD) ( Fig. 3). 4.3. Results Imaging of both 100 nm fluorescent nanoparticles and microtubule networks verify the super-resolution ability of our FED system. The final FED image is obtained by processing two acquired images according to Eq. (1) using a self-developed MATLAB
software. Only confocal and cylindrical vector beam FED images were acquired; we do not experimentally verify the improved performance of our new FED relative to conventional FED here. The value of r, is adjusted from 0 to 1 based on our empirical judgment of the best quality FED images. 4.3.1. Nanoparticle imaging We test the resolution of cylindrical vector beam FED by imaging fluorescent nanoparticles (100 nm, yellow-green FluoSpheres, Molecular Probes). Fig. 4(a) and (b) presents the confocal and FED images of a 10 μm*10 μm area of fluorescent nanoparticles. In Fig. 4(c), we obtain an FED þ image using the Richardson–Lucy algorithm to deconvolve the FED image. Fig. 4(d) shows a magnified view of the region indicated by the dashed boxes in Fig. 4(a), (b), and (c). Two nanoparticles on the left in confocal images appear stuck together and can barely be distinguished, but they are clearly distinguished in the FED image. Moreover, the resolution and SNR are further improved in the FED þ image. Fig. 4
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
5
Fig.3. Schematic of the cylindrical vector beam FED microscopy system. CL1-collimating lens, λ/2 WP-λ/2 waveplate, PBS-polarizing beam splitter, R-reflecting mirror, S-shutter, DM-dichroic mirror, SL-scanning lens, TL-tube lens, OL-objective lens, CL2-convergent lens, APD-avalanche photon diode.
(e) presents the normalized intensity profiles along the dashed lines marked in Fig. 4(d); the red, blue, and yellow lines represent the intensities in the confocal, FED, and FED þ images, respectively. The confocal profile features only one obvious peak with a broad flat top, while the FED and FED þ profiles have three distinct peaks separated by lower intensity valleys. Fig. 4(f) gives a magnified view of the region indicated by the boxes in Fig. 4(a), (b), and (c), in which the two particles that overlap each other in the confocal image are readily discernible in the FED and FED þ images. Fig. 4(g) shows the normalized intensity profiles along the dashed line in Fig. 4(f), using the same color scheme as in Fig. 4(e). In the confocal profile, the two particle profiles appear as one peak due to limited resolving ability, whereas, in the FED profile, the two peaks are resolved separated at a distance of about 140 nm. The FED image achieves a FWHM of about 110 nm, demonstrating significantly improved resolving ability of our cylindrical vector beam FED system over confocal. The result in Fourier domain are shown in Fig. 4(h)–(j) corresponding to (a)–(c), respectively. Compared with the confocal microscopy, our new FED can yield the improved resolution by a factor of 1.3125, while FED after deconvolution (FED þ), as shown in Fig. 4(j), contains more information and can achieve enhanced resolution by a factor of 1.875. 4.3.2. Biological cells imaging We imaged biological cells (U373 human astrocytes) to testify the super-resolution of our FED system. The U373 human astrocytes were cultured in 10% FBS with 1% penicillin/streptomycin and 5% CO2 at 37 °C and then immersed in sealing liquid of 3% FBS and 1% BSA in PBS for 1.5–2 h. Primary antibody solution was prepared with mouse anti-tubulin in sealing liquid at a ratio of 1:200, and secondary antibody solution was prepared with goat anti-mouse in sealing liquid at a ratio of 1:400. We immersed the sample upside-down in the primary antibody solution overnight at 4 °C. After that, we immersed the sample upside-down in secondary antibody solution for 1.5–2 h in the dark. The sample was single-photon excited at a wavelength of 488 nm. Experimental results of the microtubule networks of cultured U373 human astrocytes are shown in Fig. 5; Fig. 5(a), (b), and (c) shows the confocal, FED (r ¼0.75) and FED þ images of a 20 μm*20 μm area of the microtubule networks, respectively, with dashed boxes corresponding to magnified views in Fig. 5(d), (e), and (f). The FED and FED þ images reveal more details of the
microtubules than the confocal images. Fig. 5(g) shows normalized intensity profiles along the arrows indicated in (d), (e), and (f). The FED and FED þ intensity profiles feature higher peak values and lower valley values than the confocal profile, indicating greater apparent lateral resolution in the FED system.
5. Discussion In our experiment, the confocal pinhole is replaced by a multimode optical fiber (M31L02, Thorlabs) with a pinhole size S pinhole ¼ 62.5 mm. The Airy spot size is calculated on the focal plane according to 1.22λ/NA (NA ¼1.4), assuming a 500 nm average of the excitation and fluorescence wavelengths. The fluorescence is collected by an objective lens (f ¼1.8 mm) and then successively passed through a tube lens (f ¼200 nm), a scanning lens (f ¼70 mm) and a convergent lens (f ¼200 mm). This yields an Airy spot of size S Airy = 138 μm imaged on the pinhole plane, and S pinhole = 0.5S Airy . For such a pinhole size, we have considered both the resolution and SNR of the FED images. Before imaging with our system, we set the alignment mark direction of the S-waveplate perpendicular to the sample platform so that the s-polarization beam and the p-polarization beam are converted into a radially polarized beam and an azimuthally polarized beam, respectively. However, this alignment is accomplished manually, resulting in a nonzero angle between the alignment mark direction and the s-polarization direction. Thus, at the S-waveplate, the incident beam consists of components parallel and perpendicular to the mark direction, and the output is not a purely radially polarized or azimuthally polarized beam. Thus the solid spot generated by focusing the radially polarized beam is further extended and the hollow spot generated by focusing the azimuthally polarized beam has a nonzero central minimal intensity; this may lower the SNR by reducing the effective signal intensity. On the other hand, we can use this shortcoming by tuning the intensity distribution between the radially polarized beam and azimuthally polarized beam to form appropriately extended spots, and thus realize the non-artifact FED. In our method, the sample is scanned frame by frame with a galvo mirror, meaning the sample is scanned continuously with the solid illumination pattern until a confocal image frame is obtained and then with the hollow illumination pattern until a
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
6
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 Q3 56 57 58 59 60 61 62 63 64 65 66
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Fig.4. Fluorescent nanoparticle imaging by FED compared to confocal. (a) Confocal image. (b) FED image with a subtractive factor r¼ 0.7. (c) Deconvoluted FED image. (d) Magnified views of regions indicated by the dashed boxes in (a), (b), and (c). (e) Normalized intensity profiles along the dashed line labeled in (d). (f) Magnified views of regions indicated by the boxes in (a), (b), and (c). (g) Normalized intensity profiles along the dashed line labeled in (f). (h)–(j) Fourier domain images corresponding to (a)–(c), respectively. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
negative confocal image frame is obtained. Sample mismatch may cause pixel mismatch during scanning, which will deteriorate image quality. To solve this problem, we scan the sample pixel by pixel, i.e., we scan the sample at a single point with the solid excitation mode, then switch to the hollow mode and scan the same point, and repeat for each point until a whole frame of image is acquired. This switching is realized by automated shutter control.
6. Conclusion We proposed a novel FED system that uses cylindrical vector beams. In the system, a radially polarized beam and an azimuthally polarized beam serve as confocal and negative confocal illuminations to generate an expanded solid and shrunken hollow focal spot, respectively. We demonstrated performance
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 Q4 58 59 60 61 62 63 64 65 66
7
Fig.5. Biological cells imaged by FED. (a) Confocal image. (b) FED image with a subtractive factor r¼ 0.75. (c) FED image after deconvolution. (d), (e), and (f) Corresponding magnified views of the region indicated by the dashed boxes in (a), (b), and (c), respectively. (g) Normalized intensity profiles along the arrows indicated in (d), (e), and (f).
enhancement of the new FED method theoretically using Debye integral calculations and MATLAB simulations. We built an FED system based on our proposal and compared its performance imaging 100 nm nanoparticles and biological cells with that of conventional confocal microscopy. The new FED system has a superior resolving ability, with lateral resolution less than 120 nm. In prospect, further system optimization should be pursued.
Acknowledgment This work was financially supported by grants from the National Basic Research Program of China (973 Program) (No. 2015CB352003), the National Natural Science Foundation of China (Nos. 61377013, 61205160, 61378051, 61335003 and 61427818), and the Open Foundation of the State Key Laboratory of Modern Optical Instrumentation.
Reference [1] T.A. Klar, S.W. Hell, Subdiffraction resolution in far-field fluorescence
microscopy, Opt. Lett. 24 (1999) 954–956. [2] H. Köhler, On Abbe's theory of image formation in the microscope, J. Mod. Opt. 28 (1981) 1691–1701. [3] F. Bergermann, L. Alber, S.J. Sahl, J. Engelhardt, S.W. Hell, -fold parallelized dual-color STED fluorescence nanoscopy, Opt. Express 23 (2015) (2000) 211. [4] T. Wilson, Confocal Microscopy, Academic Press, London (1990), p. 1–64. [5] T. Wilson, Resolution and optical sectioning in the confocal microscope, J. Microsc. 244 (2011) 113–121. [6] S.T. Hess, T.P. Girirajan, M.D. Mason, Ultra-high resolution imaging by fluorescence photoactivation localization microscopy, Biophys. J. 91 (2006) 4258–4272. [7] M.J. Rust, M. Bates, X. Zhuang, Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM), Nat. Methods 3 (2006) 793–796. [8] S.W. Hell, J. Wichmann, Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy, Opt. Lett. 19 (1994) 780–782. [9] L. Shao, P. Kner, E.H. Rego, M.G. Gustafsson, Super-resolution 3D microscopy of live whole cells using structured illumination, Nat. Methods 8 (2011) 1044–1046. [10] T. Dertinger, R. Colyer, R. Vogel, J. Enderlein, S. Weiss, Achieving increased resolution and more pixels with superresolution optical fluctuation imaging (SOFI), Opt. Express 18 (2010) 18875. [11] M.J. Franken, C. Poelma, J. Westerweel, Nanoscale contact line visualization based on total internal reflection fluorescence microscopy, Opt. Express 21 (2013) 26093–26102. [12] H. Dehez, M. Piché, Y. De Koninck, Resolution and contrast enhancement in laser scanning microscopy using dark beam imaging, Opt. Express 21 (2013) 15912–15925. [13] S.W. Hell, M. Kroug, Ground-state-depletion fluorscence microscopy: a concept for breaking the diffraction resolution limit, Appl. Phys. B 60 (1995)
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
8
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
495–497. [14] W. Yifan, K. Cuifang, L. Chenge, L. Wei, C. Huanqing, H. Xiang, L. Xu, Sub-diffraction Imaging with Confocal Fluorescence Microscopy by Stochastic Photobleaching, 2013. [15] C. Kuang, S. Li, W. Liu, X. Hao, Z. Gu, Y. Wang, J. Ge, H. Li, X. Liu, Breaking the diffraction barrier using fluorescence emission difference microscopy, Sci. Rep. 3 (2013) 1441. [16] Z. Rong, S. Li, C. Kuang, Y. Xu, X. Liu, Real-time super-resolution imaging by high-speed fluorescence emission difference microscopy, J. Mod. Opt. 61 (2014) 1364–1371. [17] Y. Fang, Y. Wang, C. Kuang, X. Liu, Enhancing the resolution and contrast in CW-STED microscopy, Opt. Commun. 322 (2014) 169–174. [18] R. Heintzmann, V. Sarafis, P. Munroe, J. Nailon, Q.S. Hanley, T.M. Jovin, Resolution enhancement by subtraction of confocal signals taken at different pinhole sizes, Micron 34 (2003) 293–300. [19] T. Wilson, D. Hamilton, Difference confocal scanning microscopy, J. Mod. Opt. 31 (1984) 453–465. [20] H. Dehez, M. Piché, Y. De Koninck, High Resolution Imaging with TM01 Laser Beams, Photonics North International Society for Optics and Photonics, 2009, pp. 738605–738606. [21] N. Wang, J. Miyazaki, J. He, K. Seto, T. Kobayashi, Sub-diffraction-limit imaging using mode multiplexing, Opt. Commun. 343 (2015) 28–33. [22] S. You, C. Kuang, Z. Rong, X. Liu, Z. Ding, Isotropic superresolution imaging for fluorescence emission difference microscopy, Appl. Opt. 53 (2014) 7838–7844. [23] S. Li, C. Kuang, X. Hao, Y. Wang, J. Ge, X. Liu, Enhancing the performance of
[24] [25] [26] [27] [28]
[29]
[30] [31] [32]
fluorescence emission difference microscopy using beam modulation, J. Opt. 15 (2013) 125708. S. Segawa, Y. Kozawa, S. Sato, Resolution enhancement of confocal microscopy by subtraction method with vector beams, Opt. Lett. 39 (2014) 3118–3121. S. Segawa, Y. Kozawa, S. Sato, Demonstration of subtraction imaging in confocal microscopy with vector beams, Opt. Lett. 39 (2014) 4529–4532. S. You, C. Kuang, Z. Rong, X. Liu, Eliminating deformations in fluorescence emission difference microscopy, Opt. Express 22 (2014) 26375–26385. N. Wang, T. Kobayashi, Numerical study of the subtraction threshold for fluorescence difference microscopy, Opt. Express 22 (2014) 28819–28830. Y. Ma, C. Kuang, Y. Fang, B. Ge, X. Mao, S. Shen, Y. Xu, X. Liu, Particle localization with total internal reflection illumination and differential detection, Opt. Commun. 339 (2015) 14–21. Y. Wang, C. Kuang, Z. Gu, X. Liu, Image subtraction method for improving lateral resolution and SNR in confocal microscopy, Opt. Laser Technol. 48 (2013) 489–494. Q. Zhan, Cylindrical vector beams: from mathematical concepts to applications, Adv. Opt. Photonics 1 (2009) 1. R. Dorn, S. Quabis, G. Leuchs, Sharper focus for a radially polarized light beam, Phys. Rev. Lett. 91 (2003). M. Born, E. Wolf, Principles of Optics: Electromagnetic Theory of Propagation, Interference and Diffraction of Light, CUP Archive, 1999.
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Q1 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Contents lists available at ScienceDirect
Optics Communications journal homepage: www.elsevier.com/locate/optcom
Super-resolution microscopy based on fluorescence emission difference of cylindrical vector beams Zihao Rong a, Cuifang Kuang a,n, Yue Fang a, Guangyuan Zhao a, Yingke Xu b, Xu Liu a a b
State Key Laboratory of Modern Optical Instrumentation, Department of Optical Engineering, Zhejiang University, Hangzhou, 310027, China Key Laboratory of Biomedical Engineering of Ministry of Education, Department of Biomedical Engineering, Zhejiang University, Hangzhou 310027, China
art ic l e i nf o
a b s t r a c t
Article history: Received 18 March 2015 Received in revised form 22 May 2015 Accepted 24 May 2015
We propose a novel fluorescence emission difference microscopy (FED) system based on focusing cylindrical vector beams. In conventional FED, a Gaussian beam and a 0–2π vortex phase plate are used to generate solid and hollow spots. We focus radially polarized and azimuthally polarized cylindrical vector beams to obtain an expanded solid spot and a shrunken hollow spot, taking advantage of the optical properties of cylindrical vector beams to improve the conventional FED performance. Our novel method enhances FED performance because the hollow spot size determines the FED resolution and an expanded solid spot effectively reduces negative side-lobe emergence during image processing. We demonstrate improved performance theoretically and experimentally using an in-house built FED. Our FED achieved a resolution of less than λ/4 in test images of 100 nm nanoparticles, better than the confocal image resolution by a factor of approximately 1/3. We also discuss detailed simulation analyses and FED imaging of biological cells. & 2015 Published by Elsevier B.V.
Keywords: Super-resolution Fluorescence imaging FED microscopy Cylindrical vector beam
1. Introduction Owing to its simplicity, versatility, and noninvasiveness, farfield fluorescence microscopy has long been applied in biological and medical science to observe and investigate microstructures and their movements [1]. However, the diffraction barrier [2] confines the resolving ability of conventional far-field fluorescence microscopy above about half the illumination light wavelength, restricting observation of microstructures featuring length scales less than 100 nm, such as vimentin fibers [3], vesicles, and microtubules. Confocal scanning laser microscopy (CSLM) [4] can enhance spatial resolution by a factor of 2 and improve micrograph contrast by spatial filtering with a pinhole, which also endows CSLM with optical sectioning ability [5]. Nonetheless, CSLM resolution is still limited by the diffraction barrier. Fluorescence emission difference (FED) [15,16] microscopy was recently reported as a novel super-resolution technique based on intensity subtraction [17] between two images acquired under different illumination patterns. It joins a long list of validated super-resolution fluorescence microscopy methods, e.g., photoactivated localization microscopy (PALM) [6], stochastic optical reconstruction microscopy (STORM) [7], stimulated emission depletion microscopy (STED) [8], structured illumination microscopy n
Corresponding author. E-mail address: [email protected] (C. Kuang).
(SIM) [9], super-resolution optical fluctuation imaging (SOFI) [10], total internal reflection fluorescence microscopy (TIRF) [11], and others [12–14]. Super-resolution techniques can be classified into two categories according to underlying principles and resolving abilities: diffraction unlimited and limited. Diffraction-unlimited techniques, including PALM, STORM, and STED, can yield resolution far beyond the diffraction barrier and overcome the barrier in true sense. Diffraction-limited techniques, including SIM, FED, SOFI, and TIRF, are fundamentally limited by the diffraction barrier. SIM, SOFI, and FED can enhance the resolution by a factor of 2 at most, while TIRF achieves super-resolution in the axial direction, but does not enhance lateral resolution. Subtractive imaging was invented decades ago to enhance image resolution by subtracting confocal signals taken at different collection pinhole sizes under the same illumination pattern [18,19]. However, its signal-to-noise ratio (SNR) is relatively low since final resolution-enhanced image construction uses signals emitted from the periphery of the excitation spot. To overcome this, fluorescence emission difference microscopy (FED) [15] and switching laser mode microscopy (SLAM) [12,20] were proposed as subtractive imaging alternatives that apply different illumination patterns with the same collection pinhole size. In FED, the sample is alternately scanned by a solid excitation spot and a doughnut-shaped excitation spot [21], and central signals are extracted by eliminating peripheral signals through subtraction [22]. The resolution and SNR of FED can be further enhanced through beam modulation [23–25] to decrease the solid and hollow spot
http://dx.doi.org/10.1016/j.optcom.2015.05.057 0030-4018/& 2015 Published by Elsevier B.V.
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
sizes. However, image subtraction inevitably produces negative side-lobes in the FED point spread function (PSF) under any subtractive factor, since the PSF profiles of the two illumination modes do not perfectly match [26]. This mismatch may introduce artifacts and deteriorate image quality. To reduce negative side-lobes and eliminate image deformation, the PSF profiles of the two illumination patterns were approximately matched by beam modulation [25,26]. Kobayashi et al. numerically studied the FED subtractive factor r, offering considerable selection criteria for it [27]. Moreover, image subtractive techniques can be combined with other super-resolution methods such as TIRF [28] and STED [17] to further enhance their performance. Since FED is, in principle, built upon CSLM, its attainable resolution is determined by the size of the two micro-imaging mode PSFs, not the diffraction barrier [23]. Moreover, FED achieves image resolution beyond the diffraction barrier at low laser intensity and relatively high imaging speed in a simple and compact system, which undoubtedly adds to its availability in practical scientific research [29]. Cylindrical vector beams, i.e., radially polarized beams and azimuthally polarized beams, have attracted a lot of interest due to their unique polarization states and prominent beam properties, and have prospective applications in technologies such as laser cutting, particle acceleration and optical microscopy [30]. Research show that a radially polarized field can be focused to a spot size of 0.16λ, significantly smaller than 0.26λ for a linear polarized field [31]; an azimuthally polarized field can similarly be focused to a smaller spot size than the hollow spot created by modulating a linear polarized field through a 0–2π vortex phase plate. However, tightly focusing a radially polarized beam by a high NA objective lens enlarges the solid spot size because of a focal lateral doughnut-shaped component that broadens the entire radially polarized beam PSF. We establish a novel FED system that capitalizes on these features of cylindrical vector beams to enhance performance. We established a cylindrical vector beam FED system in which a radially polarized beam and an azimuthally polarized beam create an extended solid spot and a shrunken hollow spot, respectively, at the focal plane by tight focusing. Simulations demonstrated significant performance enhancement by cylindrical vector beam FED and experiments achieved a resolution less than 120 nm.
2. Theory In FED, two different confocal scanning images are required to obtain the final FED image: the confocal image is acquired under the solid excitation pattern and the negative confocal image is acquired under the hollow excitation pattern [15]. The excited fluorescence is filtered by a pinhole and detected by a photomultiplier tube (PMT) to form both images. The final FED superresolution image is constructed by mathematical intensity subtraction of the two images:
IFED = Isolid − r × Ihollow
(1)
Here, IFED , Isolid , and Ihollow are the normalized intensity distributions of the FED, confocal, and negative confocal images, respectively, and r is the subtractive factor. Negative intensity differences, which are inevitable after subtraction, are excluded from the final image for better image quality. The resolving ability of FED is determined by the PSFs of the two illumination modes, which can be described by the Debye integral [23],
E (r2, φ 2, z 2 ) = iC
⎡ px ⎤ ⎢ ⎥
∫ ∫Ω sin θ⋅A1 (θ, φ)⋅A2 (θ, φ)⋅⎢ py ⎥⋅exp [if (θ, φ)] ⎢p ⎥ ⎣ z⎦
⎡ exp ⎣ikn z 2 cos θ + r2 sin θ cos θ φ−φ 2
(
(
))
⎤ ⎦ dθ dφ
(2)
E (r2, φ2, z2 ) is the electric field vector at the point (r2, φ2, z2 ) in cylindrical coordinates relative to an origin at the focal point of the objective lens. (θ , φ) represents angular position on the wave front of the incident beam, where θ is the angle between the ray direction and the optical axis and φ is the azimuthal angle. Ω represents the effective incident aperture of the beam, A1 (θ , φ) is the amplitude function of the input light, and A2 (θ , φ) is the aberration function determined by the structure of the objective lens. ⎡p , p , p ⎤T is the polarization state of the incident beam and ⎣ x y z⎦ f (θ , φ) represents the phase modulation function applied to the input light. From Eq. (2), tuning the polarization state, phase distribution, and amplitude distribution can optimize the PSF size of both imaging modes to improve the FED performance [23]. We used cylindrical vector beams to generate two patterns of illumination. Unlike conventional linearly polarized beams or circularly polarized beams, cylindrical vector beams feature axisymmetric and anisotropic polarizations. Axisymmetric and anisotropic polarization have invariant angles between the electric field vector direction and the radial direction throughout the beam cross-section [32]. Fig. 1(a) and (b) shows that the electric field vector directions of radially polarized and azimuthally polarized beams are always parallel and perpendicular to the radial direction, respectively. These polarization properties cause a tightly focused radially polarized beam or azimuthally polarized beam to create a solid spot or hollow spot on the focal plane, respectively. When concentrated by a high NA objective lens, radial polarization direction is deflected, generating a longitudinal electric field component. The lateral electric field component then coexists with a strong longitudinal electric field component at the focal area. The intensity distributions of both components' focal areas follow Eq. (2). The lateral electric component appears as a hollow spot with a central valley, and the longitudinal electric component presents as a solid spot with a central peak. The hollow lateral component expands the solid PSF wider than a Gaussian beam; this expansion reduces the negative values resulting from image subtraction. A characteristic that we do not utilize here is the high strength of the longitudinal electric component relative to the lateral one, and its smaller spot size relative to linearly or circularly polarized beams [31]; a sharper solid focal spot can be obtained by extracting the lateral component. The azimuthal polarization directions remain unchanged by focusing with a high NA objective lens, therefore no longitudinal electric component is generated. According to vector beam diffraction theory, the azimuthally polarized beam only possesses a lateral electric component at the focal area and forms a hollow spot smaller than a Gaussian beam modulated through a 0–2π vortex phase plate [30]. Our FED PSF is established by subtracting the hollow spot generated by the azimuthally polarized beam scaled by a subtractive parameter from the solid spot generated by the radially polarized beam, as in Eq. (1).
3. Simulation We first demonstrate the resolving ability and performance enhancement of our cylindrical vector beam FED by simulating both the conventional FED PSF and the new cylindrical vector beam FED PSF. The conventional FED PSF is generated by subtracting a hollow excitation pattern modulated by a 0–2π vortex phase plate from a solid confocal excitation pattern of a focused
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Fig. 1. Schematic of cylindrical vector beam polarizations and corresponding focal spots. (a) Radially polarized beams and (b) azimuthally polarized beams form (c) solid and (d) hollow focal spots, respectively, after tightly focusing.
3
Fig. 2(d) and the conventional FED PSF profile in Fig. 2(f) share the same minimal negative value of 0.27, but the cylindrical vector beam FED features a narrower PSF than the conventional FED, demonstrating its theoretical super resolving ability. We performed imaging simulations of a 3 3 point lattice sample to verify the performance enhancement. As shown in Fig. 2 (g), the sample is composed of nine points in a square distribution, where the four points at the corners (colored in white) have twice the intensity of the other points, and vertically and horizontally adjacent points are separated by a distance of 0.3λ. In the simulated confocal image of Fig. 2(h), the four corner points are discernible, but the other points are undetectable against the periphery signal of the corner points. In the conventional FED image (r ¼0.7) of Fig. 2(i), the four corner points appear more distinct, but the other five points are nearly missing due to excessive subtraction. However, in the cylindrical vector beam FED image of Fig. 2 (j), all nine points are distinguishable, despite some acceptable artifacts. Defocusing may cause inevitable artifacts in FED images. However, the out-of-focus dispersed solid spot shrinks when a similarly dispersed hollow spot is subtracted; thus, defocusing artifacts can be reduced. Therefore, FED suffers from fewer defocusing artifacts than confocal microscopy.
4. Experiment Gaussian left-handed circular beam. The new FED PSF is generated by subtracting the hollow pattern of a focused azimuthally polarized beam from the solid pattern of a radially polarized beam focused by a high NA objective. A subtractive parameter is needed in both situations. The PSFs are calculated by Debye integrals using self-developed Java software and the following integral parameters: the excitation light wavelength λ is 488 nm; the objective lens NA is 1.4; the pinhole is considered infinitely small; and the subtractive factor r is 0.7. We study a 2λ 2λ lateral field of view centered at the focus and normalize intensity for simplicity. Fig. 2(a–f) illustrates the PSFs of conventional FED and cylindrical vector beam FED, and demonstrates cylindrical vector beam FED feasibility and image enhancement relative to confocal microscopy and conventional FED. In Fig. 2(a), the red solid curve represents the Gaussian beam PSF, while the radially polarized beam PSF and its lateral and longitudinal components are denoted by the blue solid, dotted, and dashed curves, respectively. The final solid PSF results from summing the lateral component and the longitude component, which extends the PSF’s size, thus reducing the negative value area created by subtracting the hollow PSF from this solid PSF to eliminate deformations. As shown in Fig. 2(b), the sizes of the 0–2π vortex beam PSF and the azimuthally polarized beam PSF were 0.28λ and 0.24λ, respectively. Thus, the PSF of the azimuthally polarized beam provides a 14% smaller hollow pattern than the 0–2π vortex beam does. As shown in Fig. 2(c–f), using different subtractive factors to subtract 0–2π vortex beam PSFs from Gaussian beam PSFs, in conventional FED, and azimuthally polarized beam PSFs from radially polarized beam PSFs, in the new FED, can give narrower FED PSFs and stronger resolving ability relative to confocal microscopy. Moreover, using cylindrical vector beams reduces the minimal value of negative side-lobes in the new FED by a factor of approximately 1/3 for r ¼0.9, and even more for lower r. In Fig. 2(f), the new FED PSF has no negative values for r ¼0.4, whereas the conventional FED PSF has obvious negative side-lobes. While the new FED PSF FWHM does not dramatically narrow in comparison with that of the conventional FED PSF for the same subtractive factor, the new FED can enhance the image resolution for the same minimal negative values. The new FED PSF profile in
4.1. Cylindrical vector beam generation We use an S-waveplate (Altechna, Lithuania) to generate cylindrical vector beams. An S-waveplate is a super-structured polarization converter that converts linear polarizations to radial or azimuthal polarizations. Typically, the center of the polarization converter should be aligned with the optical axis of the incident linearly polarized beam. An alignment mark on the S-waveplate is aligned parallel to the incident linear polarization orientation for radial polarization, or perpendicular for azimuthal polarization. Thus, the incident light polarization can be chosen by rotating a λ/2 waveplate, and the polarization state of the output beam can be controlled by rotating the converter placed after the waveplate. However, this method suffers from human error and inefficiency. Therefore, we split the incident laser into two arms featuring linear polarizations parallel (p-polarization) and perpendicular (s-polarization) to the experimental platform, respectively, so that the polarization state of the output beam can be controlled by switching the incident beam with shutters. 4.2. Experimental schematic We built a cylindrical vector beam FED experimental setup, as illustrated in Fig. 2. A collimating lens (CL1) collimates the 488 nm wavelength laser beam emitted by a laser diode (Coherent OBIS 488) and transmitted through a single-mode fiber. The collimated beam is split by a polarizing beam splitter (PBS, B. Halle GmbH, Berlin, Germany) into two paths, a p-polarization beam path and an s-polarization beam path. A λ/2 waveplate (λ/2 WP) tunes the intensity distribution between the p-polarization beam and the s-polarization beam. A second PBS reintegrates the split beams. Shutters inserted into each light path allow for the p-polarization beam and the s-polarization beam to alternately pass through the key S-waveplate (Altechna RPC-488-06). We set the alignment mark direction of the S-waveplate perpendicular to the experimental platform so that S-waveplate modulates the s- and p-polarization beams into radially and azimuthally polarized beams, respectively. After polarization modulation, the cylindrical vector
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 Q2 59 60 61 62 63 64 65 66
Fig. 2. Simulation PSFs of conventional FED and cylindrical vector beam FED. (a) PSFs of confocal illumination patterns (RB ¼radially polarized beam). (b) PSFs of negative confocal illumination patterns. (c–f) PSFs of confocal microscopy and conventional and cylindrical FED using subtractive factors r¼ 0.9, 0.7, 0.5 and 0.4, respectively. (g) Sample 3 3 point lattice. (h) Simulated confocal image. (i) FED image with r¼ 0.7. (j) Cylindrical vector beam FED image with r¼ 0.7.
beam is reflected by a galvo mirror (Thorlabs GVS002) controlled by a DAQ card (NI USB-6259) to realize two-dimensional fast scans of the sample. The scanning lens (SL) creates an intermediate image before the tube lens (TL), parallelizing the laser light. A high NA objective lens (Olympus Uplan SApo 100 /1.4 Oil) strongly focuses the cylindrical vector beam and collects the fluorescence emitted from the sample. The dichroic mirror (Di02-R488-25*36, Semrock) reflects the illumination while transmitting the fluorescence, which is then spatially filtered by a pinhole, coupled into a multi-mode fiber (M31L02, Thorlabs), and finally detected by an avalanche photon diode (APD) ( Fig. 3). 4.3. Results Imaging of both 100 nm fluorescent nanoparticles and microtubule networks verify the super-resolution ability of our FED system. The final FED image is obtained by processing two acquired images according to Eq. (1) using a self-developed MATLAB
software. Only confocal and cylindrical vector beam FED images were acquired; we do not experimentally verify the improved performance of our new FED relative to conventional FED here. The value of r, is adjusted from 0 to 1 based on our empirical judgment of the best quality FED images. 4.3.1. Nanoparticle imaging We test the resolution of cylindrical vector beam FED by imaging fluorescent nanoparticles (100 nm, yellow-green FluoSpheres, Molecular Probes). Fig. 4(a) and (b) presents the confocal and FED images of a 10 μm*10 μm area of fluorescent nanoparticles. In Fig. 4(c), we obtain an FED þ image using the Richardson–Lucy algorithm to deconvolve the FED image. Fig. 4(d) shows a magnified view of the region indicated by the dashed boxes in Fig. 4(a), (b), and (c). Two nanoparticles on the left in confocal images appear stuck together and can barely be distinguished, but they are clearly distinguished in the FED image. Moreover, the resolution and SNR are further improved in the FED þ image. Fig. 4
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
5
Fig.3. Schematic of the cylindrical vector beam FED microscopy system. CL1-collimating lens, λ/2 WP-λ/2 waveplate, PBS-polarizing beam splitter, R-reflecting mirror, S-shutter, DM-dichroic mirror, SL-scanning lens, TL-tube lens, OL-objective lens, CL2-convergent lens, APD-avalanche photon diode.
(e) presents the normalized intensity profiles along the dashed lines marked in Fig. 4(d); the red, blue, and yellow lines represent the intensities in the confocal, FED, and FED þ images, respectively. The confocal profile features only one obvious peak with a broad flat top, while the FED and FED þ profiles have three distinct peaks separated by lower intensity valleys. Fig. 4(f) gives a magnified view of the region indicated by the boxes in Fig. 4(a), (b), and (c), in which the two particles that overlap each other in the confocal image are readily discernible in the FED and FED þ images. Fig. 4(g) shows the normalized intensity profiles along the dashed line in Fig. 4(f), using the same color scheme as in Fig. 4(e). In the confocal profile, the two particle profiles appear as one peak due to limited resolving ability, whereas, in the FED profile, the two peaks are resolved separated at a distance of about 140 nm. The FED image achieves a FWHM of about 110 nm, demonstrating significantly improved resolving ability of our cylindrical vector beam FED system over confocal. The result in Fourier domain are shown in Fig. 4(h)–(j) corresponding to (a)–(c), respectively. Compared with the confocal microscopy, our new FED can yield the improved resolution by a factor of 1.3125, while FED after deconvolution (FED þ), as shown in Fig. 4(j), contains more information and can achieve enhanced resolution by a factor of 1.875. 4.3.2. Biological cells imaging We imaged biological cells (U373 human astrocytes) to testify the super-resolution of our FED system. The U373 human astrocytes were cultured in 10% FBS with 1% penicillin/streptomycin and 5% CO2 at 37 °C and then immersed in sealing liquid of 3% FBS and 1% BSA in PBS for 1.5–2 h. Primary antibody solution was prepared with mouse anti-tubulin in sealing liquid at a ratio of 1:200, and secondary antibody solution was prepared with goat anti-mouse in sealing liquid at a ratio of 1:400. We immersed the sample upside-down in the primary antibody solution overnight at 4 °C. After that, we immersed the sample upside-down in secondary antibody solution for 1.5–2 h in the dark. The sample was single-photon excited at a wavelength of 488 nm. Experimental results of the microtubule networks of cultured U373 human astrocytes are shown in Fig. 5; Fig. 5(a), (b), and (c) shows the confocal, FED (r ¼0.75) and FED þ images of a 20 μm*20 μm area of the microtubule networks, respectively, with dashed boxes corresponding to magnified views in Fig. 5(d), (e), and (f). The FED and FED þ images reveal more details of the
microtubules than the confocal images. Fig. 5(g) shows normalized intensity profiles along the arrows indicated in (d), (e), and (f). The FED and FED þ intensity profiles feature higher peak values and lower valley values than the confocal profile, indicating greater apparent lateral resolution in the FED system.
5. Discussion In our experiment, the confocal pinhole is replaced by a multimode optical fiber (M31L02, Thorlabs) with a pinhole size S pinhole ¼ 62.5 mm. The Airy spot size is calculated on the focal plane according to 1.22λ/NA (NA ¼1.4), assuming a 500 nm average of the excitation and fluorescence wavelengths. The fluorescence is collected by an objective lens (f ¼1.8 mm) and then successively passed through a tube lens (f ¼200 nm), a scanning lens (f ¼70 mm) and a convergent lens (f ¼200 mm). This yields an Airy spot of size S Airy = 138 μm imaged on the pinhole plane, and S pinhole = 0.5S Airy . For such a pinhole size, we have considered both the resolution and SNR of the FED images. Before imaging with our system, we set the alignment mark direction of the S-waveplate perpendicular to the sample platform so that the s-polarization beam and the p-polarization beam are converted into a radially polarized beam and an azimuthally polarized beam, respectively. However, this alignment is accomplished manually, resulting in a nonzero angle between the alignment mark direction and the s-polarization direction. Thus, at the S-waveplate, the incident beam consists of components parallel and perpendicular to the mark direction, and the output is not a purely radially polarized or azimuthally polarized beam. Thus the solid spot generated by focusing the radially polarized beam is further extended and the hollow spot generated by focusing the azimuthally polarized beam has a nonzero central minimal intensity; this may lower the SNR by reducing the effective signal intensity. On the other hand, we can use this shortcoming by tuning the intensity distribution between the radially polarized beam and azimuthally polarized beam to form appropriately extended spots, and thus realize the non-artifact FED. In our method, the sample is scanned frame by frame with a galvo mirror, meaning the sample is scanned continuously with the solid illumination pattern until a confocal image frame is obtained and then with the hollow illumination pattern until a
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
6
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 Q3 56 57 58 59 60 61 62 63 64 65 66
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Fig.4. Fluorescent nanoparticle imaging by FED compared to confocal. (a) Confocal image. (b) FED image with a subtractive factor r¼ 0.7. (c) Deconvoluted FED image. (d) Magnified views of regions indicated by the dashed boxes in (a), (b), and (c). (e) Normalized intensity profiles along the dashed line labeled in (d). (f) Magnified views of regions indicated by the boxes in (a), (b), and (c). (g) Normalized intensity profiles along the dashed line labeled in (f). (h)–(j) Fourier domain images corresponding to (a)–(c), respectively. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
negative confocal image frame is obtained. Sample mismatch may cause pixel mismatch during scanning, which will deteriorate image quality. To solve this problem, we scan the sample pixel by pixel, i.e., we scan the sample at a single point with the solid excitation mode, then switch to the hollow mode and scan the same point, and repeat for each point until a whole frame of image is acquired. This switching is realized by automated shutter control.
6. Conclusion We proposed a novel FED system that uses cylindrical vector beams. In the system, a radially polarized beam and an azimuthally polarized beam serve as confocal and negative confocal illuminations to generate an expanded solid and shrunken hollow focal spot, respectively. We demonstrated performance
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 Q4 58 59 60 61 62 63 64 65 66
7
Fig.5. Biological cells imaged by FED. (a) Confocal image. (b) FED image with a subtractive factor r¼ 0.75. (c) FED image after deconvolution. (d), (e), and (f) Corresponding magnified views of the region indicated by the dashed boxes in (a), (b), and (c), respectively. (g) Normalized intensity profiles along the arrows indicated in (d), (e), and (f).
enhancement of the new FED method theoretically using Debye integral calculations and MATLAB simulations. We built an FED system based on our proposal and compared its performance imaging 100 nm nanoparticles and biological cells with that of conventional confocal microscopy. The new FED system has a superior resolving ability, with lateral resolution less than 120 nm. In prospect, further system optimization should be pursued.
Acknowledgment This work was financially supported by grants from the National Basic Research Program of China (973 Program) (No. 2015CB352003), the National Natural Science Foundation of China (Nos. 61377013, 61205160, 61378051, 61335003 and 61427818), and the Open Foundation of the State Key Laboratory of Modern Optical Instrumentation.
Reference [1] T.A. Klar, S.W. Hell, Subdiffraction resolution in far-field fluorescence
microscopy, Opt. Lett. 24 (1999) 954–956. [2] H. Köhler, On Abbe's theory of image formation in the microscope, J. Mod. Opt. 28 (1981) 1691–1701. [3] F. Bergermann, L. Alber, S.J. Sahl, J. Engelhardt, S.W. Hell, -fold parallelized dual-color STED fluorescence nanoscopy, Opt. Express 23 (2015) (2000) 211. [4] T. Wilson, Confocal Microscopy, Academic Press, London (1990), p. 1–64. [5] T. Wilson, Resolution and optical sectioning in the confocal microscope, J. Microsc. 244 (2011) 113–121. [6] S.T. Hess, T.P. Girirajan, M.D. Mason, Ultra-high resolution imaging by fluorescence photoactivation localization microscopy, Biophys. J. 91 (2006) 4258–4272. [7] M.J. Rust, M. Bates, X. Zhuang, Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM), Nat. Methods 3 (2006) 793–796. [8] S.W. Hell, J. Wichmann, Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy, Opt. Lett. 19 (1994) 780–782. [9] L. Shao, P. Kner, E.H. Rego, M.G. Gustafsson, Super-resolution 3D microscopy of live whole cells using structured illumination, Nat. Methods 8 (2011) 1044–1046. [10] T. Dertinger, R. Colyer, R. Vogel, J. Enderlein, S. Weiss, Achieving increased resolution and more pixels with superresolution optical fluctuation imaging (SOFI), Opt. Express 18 (2010) 18875. [11] M.J. Franken, C. Poelma, J. Westerweel, Nanoscale contact line visualization based on total internal reflection fluorescence microscopy, Opt. Express 21 (2013) 26093–26102. [12] H. Dehez, M. Piché, Y. De Koninck, Resolution and contrast enhancement in laser scanning microscopy using dark beam imaging, Opt. Express 21 (2013) 15912–15925. [13] S.W. Hell, M. Kroug, Ground-state-depletion fluorscence microscopy: a concept for breaking the diffraction resolution limit, Appl. Phys. B 60 (1995)
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
8
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Z. Rong et al. / Optics Communications ∎ (∎∎∎∎) ∎∎∎–∎∎∎
495–497. [14] W. Yifan, K. Cuifang, L. Chenge, L. Wei, C. Huanqing, H. Xiang, L. Xu, Sub-diffraction Imaging with Confocal Fluorescence Microscopy by Stochastic Photobleaching, 2013. [15] C. Kuang, S. Li, W. Liu, X. Hao, Z. Gu, Y. Wang, J. Ge, H. Li, X. Liu, Breaking the diffraction barrier using fluorescence emission difference microscopy, Sci. Rep. 3 (2013) 1441. [16] Z. Rong, S. Li, C. Kuang, Y. Xu, X. Liu, Real-time super-resolution imaging by high-speed fluorescence emission difference microscopy, J. Mod. Opt. 61 (2014) 1364–1371. [17] Y. Fang, Y. Wang, C. Kuang, X. Liu, Enhancing the resolution and contrast in CW-STED microscopy, Opt. Commun. 322 (2014) 169–174. [18] R. Heintzmann, V. Sarafis, P. Munroe, J. Nailon, Q.S. Hanley, T.M. Jovin, Resolution enhancement by subtraction of confocal signals taken at different pinhole sizes, Micron 34 (2003) 293–300. [19] T. Wilson, D. Hamilton, Difference confocal scanning microscopy, J. Mod. Opt. 31 (1984) 453–465. [20] H. Dehez, M. Piché, Y. De Koninck, High Resolution Imaging with TM01 Laser Beams, Photonics North International Society for Optics and Photonics, 2009, pp. 738605–738606. [21] N. Wang, J. Miyazaki, J. He, K. Seto, T. Kobayashi, Sub-diffraction-limit imaging using mode multiplexing, Opt. Commun. 343 (2015) 28–33. [22] S. You, C. Kuang, Z. Rong, X. Liu, Z. Ding, Isotropic superresolution imaging for fluorescence emission difference microscopy, Appl. Opt. 53 (2014) 7838–7844. [23] S. Li, C. Kuang, X. Hao, Y. Wang, J. Ge, X. Liu, Enhancing the performance of
[24] [25] [26] [27] [28]
[29]
[30] [31] [32]
fluorescence emission difference microscopy using beam modulation, J. Opt. 15 (2013) 125708. S. Segawa, Y. Kozawa, S. Sato, Resolution enhancement of confocal microscopy by subtraction method with vector beams, Opt. Lett. 39 (2014) 3118–3121. S. Segawa, Y. Kozawa, S. Sato, Demonstration of subtraction imaging in confocal microscopy with vector beams, Opt. Lett. 39 (2014) 4529–4532. S. You, C. Kuang, Z. Rong, X. Liu, Eliminating deformations in fluorescence emission difference microscopy, Opt. Express 22 (2014) 26375–26385. N. Wang, T. Kobayashi, Numerical study of the subtraction threshold for fluorescence difference microscopy, Opt. Express 22 (2014) 28819–28830. Y. Ma, C. Kuang, Y. Fang, B. Ge, X. Mao, S. Shen, Y. Xu, X. Liu, Particle localization with total internal reflection illumination and differential detection, Opt. Commun. 339 (2015) 14–21. Y. Wang, C. Kuang, Z. Gu, X. Liu, Image subtraction method for improving lateral resolution and SNR in confocal microscopy, Opt. Laser Technol. 48 (2013) 489–494. Q. Zhan, Cylindrical vector beams: from mathematical concepts to applications, Adv. Opt. Photonics 1 (2009) 1. R. Dorn, S. Quabis, G. Leuchs, Sharper focus for a radially polarized light beam, Phys. Rev. Lett. 91 (2003). M. Born, E. Wolf, Principles of Optics: Electromagnetic Theory of Propagation, Interference and Diffraction of Light, CUP Archive, 1999.
Please cite this article as: Z. Rong, et al., Optics Communications (2015), http://dx.doi.org/10.1016/j.optcom.2015.05.057i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132