1. Introduction

Classical optical microscopes are restricted in resolution by Abbe’s diffraction limit, as waves carrying sub-diffraction details, which are present in the near-field, are evanescent and disappear in the far-field [1–3]. Recent reports provided evidence of transmission of near-field information by placing directly on top of the sample nano-scale lenses [4,5], polymer microdroplets [6] or dielectric microspheres [7–12]. Such methods, in which a micro- or nano-object is transmitting near-field information into the far-field, where it is collected by a conventional optical microscope, are called super-resolution microscopies, as features with a size of λ/3-λ/7 [11,13], where λ is the wavelength of the illumination, were visualized. However, a drawback of these systems is the limited field-of-view that cannot exceed the size of the applied micro- or nano-object placed on the sample. Another limiting factor is that, in these studies, the microspheres are placed randomly on the sample, which means that only a certain part of it can be imaged and the position cannot be changed after the initial placement. On the other hand, using laser light sources combined with fluorescent labeling, conventional microscopes can be upgraded to enable confocal imaging, but the instrumentation costs and the imaging time become important [14,15]. An elegant technique, utilizing both a laser light source and fluorescent samples, called structured illumination microscopy (SIM), can achieve super-resolution by applying a rotating linear grating [16]. During imaging, a set of pictures are recorded meanwhile the grating is turned around, which is followed by a Fourier analysis-based image processing step that generates the final image. A technique, called stimulated emission depletion microscopy (STED), is based on adding a second laser to the imaging system, which modifies the point-spread function and therefore increases the resolution by suppressing signals located far from the center of the imaging area. This increment is proportional to the square route of the power of the stimulating beam [17,18]. To image highly dense population of molecules, special fluorescent imaging systems like photo-activated localization microscopy (PALM) [19] and stochastic optical reconstruction microscopy (STORM) [20] were created. These systems are based on switching different fluorescent probes on and off, while mapping their individual positions. Although they provide excellent resolution, only fluorescent samples can be imaged in an area that is typically below ∼50 × 50 μm². Another improvement was made with the invention of light-sheet based microscopy: applying a different pathway for illumination and detection enabled a higher resolution, better signal-to-noise ratio and larger imaged surface at the cost of increasingly complex instrumentation and a need for large data processing [21–23].

Scanning tip-based microscopies, like atomic force microscopy (AFM) or scanning tunneling microscopy (STM), can image samples over micrometer ranges, meanwhile providing a resolution deeply below the optical diffraction limit [24–27]. Recently it was reported, that AFM can be combined with optical fluorescent imaging to have enhanced picture quality of biological samples [28]. Scanning near-field optical microscopy (SNOM) and tip-enhanced near-field optical microscopy (TENOM) also achieved super-resolution, while imaging up to ~25 μm² surfaces [29–32]. Others achieved super-resolution by replacing the point of an AFM tip with a so-called microsphere superlens, which could produce extended super-resolution images by scanning of the AFM head over the sample [33]. While benefitting from established AFM-positioning control schemes, these systems are less straightforward to operate, because of the intrinsic sensitivity of the detection principle on vibrations, which can hinder experimentation. A further advantage of these established techniques is that the image processing is already implemented, meaning that the acquired data is processed automatically and shown to the user as a final picture. On the other hand, many image processing algorithms were recently developed and applied in various research projects [34–38]. It is possible to implement these in customized microscope setups, therefore, if the microsphere scanning principle could be combined in a robust way with observation trough a normal optical microscope objective, this would be of great value to microscopy, as such simple add-on tool would potentially upgrade any classical microscope to a super-resolution one.

2. Materials and Methods

In this paper, we present a novel method called microsphere-based super-resolution scanning optical microscopy (MS-SOM), in which we combined the advantages of a classical easy-to-operate optical microscope with a surface scanning principle. The basis of our microscopy system is a dielectric microsphere with refractive index n_sphere = 1.95 that is placed on top of the sample and surrounded by a medium with refractive index n_medium = 1.56 [Fig. 1(a)]. We have chosen these materials and the corresponding values based on our previous studies [8,39]. A super-resolution image is created due to two factors [39]. First, the microsphere acts as a solid immersion lens and increases the numerical aperture of the system locally. Since the sample is placed within the focal length of this lens, it projects a virtual magnified imagebelow the sample plane. The second factor is the development of the photonic nanojet [40–43], which is a narrow light beam with high optical density emerging over a length ~2λ away from the microsphere and with full-width-at-half-maximum (FWHM) of ~λ/3 [39]. While a dielectric microsphere is capable of creating super-resolution images, its field-of-view is restricted by its diameter. To overcome this limitation, we have first fixed 40 μm diameter barium titanate (BTG) microspheres (Cospheric, Santa Barbara, CA, USA) on a 150 μm-thick glass microscope slide using Norland Optical Adhesive 63 (NOA63) glue, which was subsequently attached to the microscope objective via a metal frame [Fig. 1(b)]. The latter was connected to a fixation holder via four metal rods, by which the frame can slide with respect to the microscope objective. A motorized microscope stage carrying the sample can move independently from the fixed microsphere, in the x-, y- and z-directions.

 

Fig. 1 Principle of operation of the super-resolution scanning optical microscope. a) A dielectric microsphere with refractive index n_sphere and situated in an optical medium with refractive index n_medium is placed on top of the sample and is observed by a classical optical microscope objective. This generates a virtual image below the sample plane, which shows super-resolution. b) Schematic cross-section of the experimental setup, allowing scanning of the sample with respect to the microsphere-objective system. 1: microscope objective; 2: fixation holder to the objective; 3: four metal rods are used for fixing element number 4, i.e. a metal frame with circular opening, to which a coverglass is glued (element number 5), to which a dielectric microsphere (element number 7), typically a BTG microsphere of 10-40 μm diameter, is fixed via a thin layer of NOA 63 glue (element number 6); 8: sample to be imaged, which is mounted on a motorized microscope stage. An initial vertical scan of the stage (Z_init) is performed, during which the frame is translated with respect to the microscope objective by a frictional slide along the four rods, to find the optical image plane. c) Three-dimensional exploded view of the scanning system.

Download Full Size | PDF

A Zeiss Axio Imager M2m upright optical microscope equipped with HAL100 halogen light source, an AxioCam MRm camera and a 63x oil immersion objective with a numerical aperture (NA) of 1.4 and cover glass correction up until 170 μm thickness was used. The initial step of the operation is positioning the sample within the focus of the microscope objective [Fig. 1(c)]. This is followed by a search for the virtual image plane, which is below the sample and where the highest magnification and the biggest contrast is present. The process involves raising the stage step-by-step causing a frictional slide on the metal rods, thereby focusing the objective on lower-and-lower planes beneath the sample (z-axis). After finding the imaging plane, the movement of the stage is stopped and the obtained z-axis value is stored as initial value. The fully automated x-y scan starts from this point. Although the software provided by the manufacturer has the capability for automatic stitching of tiles, we could not utilize this function, as it only allows stitching together the full field-of-view tiles recorded by the mounted camera. Since we could only use a fraction of the area of a tile, namely the regions underneath a microsphere where super-resolution occurs, a new procedure had to be developed. A custom algorithm, written in Visual Basic, which runs within the Zeiss AxioVision software, was created to control all stage movements and the camera. A single step in the scanning process involves first taking a picture through the microscope, after which the stage moves 5 μm down along the z-axis and makes one step along either the x- or the y-axis. Hereafter, the stage raises back over 5 μm, ready for taking the next picture. The extra lifting along the z-axis was implemented to avoid scratching the sample surface with the microsphere and to reduce the shear stress. Our software interface allows setting freely the scanning step size and choosing the area to be imaged; also it saves line-by-line the pictures for subsequent analysis.

3. Results and Discussion

As imaging test structures, we used line patterns, like the ones shown in the insets of Fig. 2(a) and characterized by a line width L between 0.13 and 0.18 μm and a pitch P between 0.26 and 0.36 μm, as present on a MetroChip microscope calibration target (Pelco, Redding, CA, USA). Before performing the super-resolution imaging experiments, various samples were imaged by the microscope objective without use of a microsphere [Fig. 2(a) inset and Fig. 2(c)]. Theoretically, the resolution d of a classical microscope is determined by Abbe’s diffraction limit: d = λ/(2*NA). The halogen light source typically emits light in the λ = 400-700 nm spectral range, with maximum peak irradiance at λ = 600 nm (data obtained from Zeiss), at which a resolution d ∼215 nm is calculated. Experimentally, without use of a microsphere, a line pattern with 360 nm pitch distance could be imaged, while the one with 300 nm pitch could not be resolved [Fig. 2(c)]. This was confirmed by gray-scale analysis using ImageJ software. During acquisition of the image tiles, the camera was set to take the pictures with the highest possible contrast to make imaging of all tiles comparable. Subsequently, the regions of interest (ROI) were cropped from the pictures and the gray values within the ROI were analyzed along a defined line (yellow line on the figures). When plotting these values,black color is marked with 0 and white with 255. As the sample had 11 dark lines, the plot should show 11 peaks pointing downwards. When placing a 40 μm diameter BTG microsphere on the samples, the line pattern could be clearly resolved [Figs. 2(a) and 2(d)]. Interestingly, imaging periodic nanometric line patterns with visible light tends to be a more challenging task, than imaging standalone objects [44], non- equal sized objects and spaces [12,45], laser light source [9], fluorescent samples [17,19,21] or any combination of these. Indeed, it is easy to realize that a signal coming from a standalone nanometric object can be much weaker and still be detectable, since it will have a bigger contrast with respect to the background than when imaging a periodic nanometric structure. The same applies for periodic structures, where the gap between nanometric objects (particles or lines) is larger than the size of the particles or lines themselves. This phenomenon also can be exploited vice versa: having non-diffraction-limited objects combined with diffraction-limited gaps will be beneficial to the actual resolution of the system. Also applying a laser light source highly increases the illumination power and therefore increases the reflected light’s intensity. If the laser is combined with a fluorescent sample, this further increases the lowest detectable signal, because in this case the nanometric sample acts as a virtual light source, and the background is almost fully dimmed. Therefore, when imaging line patterns, although the pitch distance is a crucial parameter of the system, we consider the actual line width to be the resolution of the system to have fair comparison with other systems.

 

Fig. 2 Image obtained after a single scan step and analysis of the super-resolution effect. a) Super-resolution imaging of the sample represented in the insets by positioning a 40 μm diameter BTG microsphere on top of it, as observed by a regular microscope objective (63x, oil immersion, NA = 1.4). The dotted red circle indicates the total field of view of the microsphere, as defined by its radius r_sphere. The green dotted circle with radius r_sup.res. marks the area where super-resolution imaging occurs. Inset top left: Schematic of a typical sample imaged in this study, consisting of 11 line patterns with a pitch of 0.28 µm and a line width of 0.14 µm. Inset bottom left: Sample imaged without the use of a microsphere. b) Three- and two-dimensional (inset) schematics of the imaging by a microsphere. The light waves in the center of the microsphere (green) carry the super-resolution information; r_sup.res. implicitly defines the vertical distance h between sample and microsphere surface, below which super-resolution imaging is enabled. c) Samples with a pitch of 0.36 µm, 0.3 µm, 0.26 µm, respectively, and a line:interspace ratio 1:1, as imaged with the same microscope objective as used in a) without the use of a microsphere. Gray-scale analysis along the marked yellow line is showed on the right of every image, respectively. The 11 down pointing spikes corresponding to the 11 lines of the samples can only be seen on the top picture, and in the middle but in the bottom they are not resolved. Scale bar: 1 µm. d) Improved imaging of the same samples as in c) by using a microsphere on top. Scale bar: 1 µm.

Download Full Size | PDF

The exploitable field-of-view for super-resolution was limited to the central part of the microsphere [Fig. 2(a)], marked by the radius r_sup.res.. During experiments it was observed, that the position r_sup.res. corresponds to a distance h between the sample and the microsphere of ∼1 μm, which is a critical distance above which super-resolution imaging is no longer possible [Fig. 2(b)].

Theoretically, Maxwell’s wave equation applies in a medium with refractive index n_medium [Eq. (1)]:

To further investigate which parameters influence the h value, finite element simulations (COMSOL Multiphysics) of the light propagation from a multiple line source with wavelength of λ = 650 nm positioned on a silicon substrate (n_silicon = 3.48) through the BTG microsphere (n_sphere = 1.95) and surrounding dielectric medium (n_medium = 1.56) were carried out. A scalar equation [Eq. (4)] was used to study transverse electric waves in a two-dimensional model,

Figure 3(a) and 3(c) are simulations of a 450 nm wide multiple line pattern with S_offset = 1 nm, i.e. the microsphere is practically in contact with the substrate. Since, during an imaging experiment, the reflected light from a grating structure is investigated, we represent in the simulation a line of the grating as a light source. The simulations show the interference pattern between eleven line sources positioned at the bottom of the simulation area, propagating towards the microscope objective (not shown in this simulation). The modulation can be clearly observed, even if the distance increases to S_offset = 1000 nm [Figs. 3(b) and 3(d)]. Then we changed the width of the lines to 150 nm, which is below the diffraction limit. Figure 3(e) shows that the modulation was only present when S_offset = 1 nm. The simulation of the S_offset = 1000 nm shows that the modulation is lost [Fig. 3(f)], which agrees with Abbe’s diffraction limit for resolving sub-diffraction features, i.e. the wave with spatial frequency of 1/(300 nm) is non-propagating in the medium.

 

Fig. 3 Simulated electric fields of light propagation originating from a line pattern sample source placed at varying distances beneath a microsphere. a, b) A 40 µm BTG sphere is placed 1 nm and 1000 nm distance, respectively, above a 450 nm wide line pattern. Scale bar 10 µm. c, d) Zoom on the region near the sample source of a) and b), respectively. Modulation of the electrical field can be clearly observed in both cases. Scale bar 1 µm. e, f) Zoom of the same region as in c) and d), but taking a 150 nm wide line pattern. While the modulation can be still observed in e), it is not any more present in f), indicating that a too large propagation distance for the light in the medium provokes loss of nanometric feature information. Scale bar 1 µm.

Download Full Size | PDF

To map how the value of the modulation changes as function of these two parameters, an extended electric field simulation study was carried out. The parameter W was varied between 120 nm and 450 nm with 1.3 times multiplication steps. S_offset was set to 1, 10, 100 and 1000 nm. All combinations of these parameters resulted in 24 different cases [Table 2 in the Appendix]. For analysis, three lines for the evaluation of the normalized electrical field were introduced [Fig. 4(a)]. The first, at Z₀, was placed on the light source. The second, at Z₁, was placed at a distance S_offset/2. The third, at Z₂, was placed at a distance S_offset × (1.5 ± ε), where ε was such that the line was evaluating the first full wavefront within the microsphere. The normalized electrical field was plotted along the three lines [Fig. 4(b)] and these data were analyzed afterwards by considering the averaged maximum and minimum normalized electrical fields that together define the modulation. The 72 final modulation values are summarized on Fig. 4(c). The result shows that, for the smaller W values, the modulation is proportional to W and is even absent for the larger S_offset. For the larger W values, the modulation is maintained, i.e. the wave is fully propagating, whatever the value of S_offset is. These results show how the presence of a microsphere improves the imaging, as sub-diffraction values of W have non-zero modulation values for smaller S_offset values, meaning that, when the microsphere is in close contact to the sample, imaging at a sub-diffraction scale is possible.

 

Fig. 4 Analysis of the simulation results. a) Example of the three measurement lines that were used for evaluation of the modulation patterns in the electric field simulations. Z₀ was placed at the light source, Z₁ at the distance S_offset/2 and Z₂ at the distance S_offset × (1.5 ± ε), where ε was chosen such that the line is evaluating the first full wavefront within the microsphere. W = 260 nm, S_offset = 1000 nm. b) Plot of the electrical field simulated in a) along the three measurement lines. c) Summary of the modulations that were calculated for a range 10⁰ nm < S_offset < 10³ nm and 120 nm < W < 450 nm, based on the electrical field calculations, like the one shown in b).

Download Full Size | PDF

Figure 2(a) showed already that the field-of-view for super-resolution was limited to the central part of the microsphere, marked by the radius r_sup.res.. For each of the image tiles that was stored during the scanning process, our stitching algorithm exploited the central square comprised within the circle of radius r_sup.res. [Fig. 5(a)]. This was possible, since the ROI are positioned exactly at the same place on all image tiles. Our algorithm created a new virtual canvas, to which the ROI were mapped, forming a mosaic image. To maximize the useful area per picture, the overlap between the tiles was not implemented in the stitching algorithm, but in the scanning process itself. This approach was chosen because the pictures are not stitched after they are saved by the microscope-mounted camera, like in traditional mosaic imaging. Instead, only a central region of interest of a picture is used, namely exactly that where super-resolution occurs, i.e. in the center region of the microsphere. Since this area is relatively small compared to the field-of-view of the camera and even to the diameter of the microsphere, every pixel is important. To not lose any, a 5 µm step size was used during the scanning (along both x- and y-axis) which resulted in an overlap that was monitored in the non-superresolution part of the image created by the microsphere. Since this part of the picture is cropped out from the final stitch, at the end it looks like an overlap-free stitching, though it is not strictly speaking. This way our program, which is written in Visual Basic and which runs independently from Zeiss AxioVison software, stitched the square tiles and enabled to reconstruct the big field-of-view image, with high magnification and with super-resolution. Figures 5(b) and 5(c) show a reconstructed image with and without showing the individual tiles, respectively. An advantage of the hybrid stitching presented above is that it is possible to adjust the stitching algorithm in a way that it can take into consideration a systematic error when scanning the frame over a large distance, for example when going back to the next line of the scan [Figs. 5(d) and 5(e)]. The total area imaged in Fig. 5(e) is ∼2500 μm² and the scanning time was less than 1 minute.

 

Fig. 5 Scanning and image reconstruction. a) Image obtained after a single scan step. Only the area in the green central square that fits into the circle with radius r_sup.res. is used in the image reconstruction. b) Composed image obtained by stitching all square regions recorded during the scanning process. c) Same image as in B without indication of the tiles. d) Demonstration of the scanning super-resolution imaging of a larger sampling area, containing patterns with line width of 0.13 µm and width:interspacing ratio of 1:1.2 (left) and 1:1.4 (right), respectively. e) Same image as in d) without indication of the tiles. The zoom in the blue circle shows that 130 nm lines with 156 nm interspacing can be resolved indeed.

Download Full Size | PDF

To further demonstrate the imaging capabilities of our system, the surface of a Blu-ray disk was imaged. Figure 6(a) shows a scanning electron microscope (SEM) picture of a single layer 25 GB capacity Blu-ray disk with indication of typical feature sizes. The disk contains embossed concentric grooves with a pitch of 320 nm and data bits that have a minimum length of 150 nm [46]. To enable imaging these features, the 100 μm-thick protective film was removed from the disk, so that the sub-diffraction features could be in direct contact with the microsphere. After this step, the imaging process described before was started and super-resolution images were recorded [Figs. 6(b) and 6(c)]. Each tile had a brighter spot in the middle, the origin of which was the internal reflective layer of the Blu-ray disk, but this did not hinder the imaging. The result of the stitching operation of the scanned images is shown in Fig. 6(d). For Fig. 6(di), we found that the microsphere was slightly displaced with respect to the frame during scanning, due to friction forces on the glue, which caused a shadowing effect. However, our algorithm perfectly permitted to compensate for these small systematic position errors (except the dark-edge effect), resulting in a coherent composite image. For comparison, we conducted the same stitching with a seamless stitching method in ImageJ. Results showed that the other algorithm can mask the dark-edge error nicer, though not correct completely the shadowing effect. Therefore, it was concluded, that the scanning process has a much higher impact on the quality of the final image, than the chosen stitching algorithm.

 

Fig. 6 Scanning and image reconstruction of a Blu-ray disk surface. a) SEM picture of the surface of a Blu-ray disk with indication of the typical size of embossed features. b) Image obtained after a single scan step using a 26 μm size microsphere. For imaging, the 100 μm thick protective coating layer was removed from the disk. c) Image obtained after a single scan step using a 40 μm size microsphere. Only the area in the green central square will be used in the image reconstruction. d) (i and ii) Composed images obtained by stitching different square regions recorded during the scanning process.

Download Full Size | PDF

In Table 1, we present an overview on the performance of existing super-resolution imaging techniques in comparison with our MS-SOM system. A classical optical microscope can be considered as the basic system, which is cost-effective, has a high field-of-view, but is diffraction-limited in resolution. One direction of development is going for confocal microscopy and the advanced techniques that are based on it, such as STED, PALM and STORM. They have an excellent lateral resolution performance, at the cost of increasing imaging time and instrumentation complexity. Tip-based scanning techniques can also achieve very high resolution, even in the sub-nanometer range, but their field-of-view is smaller and they are less robust than lens-based microscopes. On the other hand, micro- and nano-sized refractive structures offer an alternative for sub-diffraction imaging at low cost, providing fast imaging that, however, is limited by a small field-of-view. The main advantage of our system is that it has all the good properties of a classical optical microscope (field-of-view, cost, imaging time), but it can produce images with super-resolution resolution, paving the way for generalized use by upgrading classical optical microscopes.

Table 1. Performance evaluation of super-resolution imaging microscope systems. Lateral resolution (LR): ▪= (LR > 200 nm); ▪▪= (100 nm < LR < 200 nm); ▪▪▪= (LR < 100 nm). Size of imaged area (SA): ●= (SA < 102 μm2); ●●= (102 μm2 < SA < 108 μm2); ●●●= (SA > 108 μm2). Estimated cost of system (EC): $ = (EC < 50 000 USD); $$ = (50 000 USD< EC < 250 000 USD); $$$ = (EC > 250 000 USD). Estimated time of imaging (ET): + = (ET < 1 min); + + = (1 min < ET < 10 min); + + + = (ET > 10 min).

View Table | View all tables in this article

4. Conclusion

We introduced super-resolution scanning optical microscopy, using a transparent dielectric microsphere that was translated over a sample surface using a conventional microscope objective. By performing extensive finite element simulations of the light propagation within the sample/microsphere/medium system, we pointed out a critical separation distance between the sample and the microsphere, below which super-resolution imaging was enabled. Therefore, such imaging of a sample placed beneath the microsphere was only possible within a very restricted area of ∼10 μm² near the contact of the substrate and the microsphere. Combining the super-resolution imaging capability of a microsphere with customized scanning and image reconstruction algorithms allowed creating super-resolution images over the full field-of-view of the microscope objective. Our proof-of-concept device was tested on linear calibration samples with nanometric features. Sample areas in the ∼10⁴ μm² range with features varying between 130 and 160 nm were imaged at a speed of 2 tiles/second. Compared to other techniques, our microscopy system allows super-resolution imaging in an affordable way, gaining almost 20% resolution compared a classical optical microscope meanwhile benefiting from all of its advantages. We think our findings may be at the basis of a future generalized and affordable scanning super-resolution optical microscopy with huge potential in many research fields.

Appendix

Table 2. Electric field simulation results of multiple line patterns with different width (Width) and different sphere - sample distance (S_offset). Smaller Widths and bigger S_offset distances significantly can reduce the modulation. Scale bar: 10^-6 m.

View Table | View all tables in this article

Funding

Swiss National Science Foundation Grant (200021-152948); European Research Council Grant (ERC-2012-AdG-320404).

Acknowledgments

The authors would like to express their gratitude to dr. Axel Hochstetter and to dr. Arne Seitz for their helpful thoughts.

References and links

1. J. B. Pendry, “Negative refraction makes a perfect lens,” Phys. Rev. Lett. 85(18), 3966–3969 (2000). [CrossRef]   [PubMed]  

2. N. Fang, H. Lee, C. Sun, and X. Zhang, “Sub-Diffraction-Limited Optical Imaging with a Silver Superlens,” Science 308(5721), 534–537 (2005). [CrossRef]   [PubMed]  

3. H. Cang, A. Salandrino, Y. Wang, and X. Zhang, “Adiabatic far-field sub-diffraction imaging,” Nat. Commun. 6, 7942 (2015). [CrossRef]   [PubMed]  

4. J. Y. Lee, B. H. Hong, W. Y. Kim, S. K. Min, Y. Kim, M. V. Jouravlev, R. Bose, K. S. Kim, I.-C. Hwang, L. J. Kaufman, C. W. Wong, P. Kim, and K. S. Kim, “Near-field focusing and magnification through self-assembled nanoscale spherical lenses,” Nature 460(7254), 498–501 (2009). [CrossRef]  

5. M.-S. Kim, T. Scharf, M. Brun, S. Olivier, S. Nicoletti, and H. P. Herzig, “Advanced optical characterization of micro solid immersion lens,” in C. Gorecki, A. K. Asundi, and W. Osten, eds. (2012), p. 84300E–84300E–10.

6. D. Kang, C. Pang, S. M. Kim, H. S. Cho, H. S. Um, Y. W. Choi, and K. Y. Suh, “Shape-Controllable Microlens Arrays via Direct Transfer of Photocurable Polymer Droplets,” Adv. Mater. 24(13), 1709–1715 (2012). [CrossRef]   [PubMed]  

7. L. Li, W. Guo, Y. Yan, S. Lee, and T. Wang, “Label-free super-resolution imaging of adenoviruses by submerged microsphere optical nanoscopy,” Light Sci. Appl. 2(9), e104 (2013). [CrossRef]  

8. H. Yang, N. Moullan, J. Auwerx, and M. A. M. Gijs, “Fluorescence Imaging: Super-Resolution Biological Microscopy Using Virtual Imaging by a Microsphere Nanoscope,” Small 10(9), 1876 (2014). [CrossRef]  

9. Y. Yan, L. Li, C. Feng, W. Guo, S. Lee, and M. Hong, “Microsphere-coupled scanning laser confocal nanoscope for sub-diffraction-limited imaging at 25 nm lateral resolution in the visible spectrum,” ACS Nano 8(2), 1809–1816 (2014). [CrossRef]   [PubMed]  

10. A. Darafsheh, N. I. Limberopoulos, J. S. Derov, D. E. Walker Jr, and V. N. Astratov, “Advantages of microsphere-assisted super-resolution imaging technique over solid immersion lens and confocal microscopies,” Appl. Phys. Lett. 104(6), 061117 (2014). [CrossRef]  

11. H. Yang and M. A. M. Gijs, “Optical microscopy using a glass microsphere for metrology of sub-wavelength nanostructures,” Microelectron. Eng. 143, 86–90 (2015). [CrossRef]  

12. F. Wang, L. Liu, P. Yu, Z. Liu, H. Yu, Y. Wang, and W. J. Li, “Three-dimensional super-resolution morphology by near-field assisted white-light interferometry,” Sci. Rep. 6(1), 24703 (2016). [CrossRef]   [PubMed]  

13. D. R. Mason, M. V. Jouravlev, and K. S. Kim, “Enhanced resolution beyond the Abbe diffraction limit with wavelength-scale solid immersion lenses,” Opt. Lett. 35(12), 2007–2009 (2010). [CrossRef]   [PubMed]  

14. X. Chen, Z. Zeng, H. Wang, and P. Xi, “Three-dimensional multimodal sub-diffraction imaging with spinning-disk confocal microscopy using blinking/fluctuating probes,” Nano Res. 8(7), 2251–2260 (2015). [CrossRef]  

15. J. Huff, “The Airyscan detector from ZEISS: confocal imaging with improved signal-to-noise ratio and super-resolution,” Nat. Methods12, (2015).

16. M. G. Gustafsson, “Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy,” J. Microsc. 198(2), 82–87 (2000). [CrossRef]   [PubMed]  

17. S. W. Hell and J. Wichmann, “Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy,” Opt. Lett. 19(11), 780–782 (1994). [CrossRef]   [PubMed]  

18. L. Kastrup, H. Blom, C. Eggeling, and S. W. Hell, “Fluorescence fluctuation spectroscopy in subdiffraction focal volumes,” Phys. Rev. Lett. 94(17), 178104 (2005). [CrossRef]   [PubMed]  

19. E. Betzig, G. H. Patterson, R. Sougrat, O. W. Lindwasser, S. Olenych, J. S. Bonifacino, M. W. Davidson, J. Lippincott-Schwartz, and H. F. Hess, “Imaging intracellular fluorescent proteins at nanometer resolution,” Science 313(5793), 1642–1645 (2006). [CrossRef]   [PubMed]  

20. M. J. Rust, M. Bates, and X. Zhuang, “Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM),” Nat. Methods 3(10), 793–796 (2006). [CrossRef]   [PubMed]  

21. C. J. Engelbrecht and E. H. Stelzer, “Resolution enhancement in a light-sheet-based microscope (SPIM),” Opt. Lett. 31(10), 1477–1479 (2006). [CrossRef]   [PubMed]  

22. E. G. Reynaud, U. Kržič, K. Greger, and E. H. K. Stelzer, “Light sheet-based fluorescence microscopy: More dimensions, more photons, and less photodamage,” HFSP J. 2(5), 266–275 (2008). [CrossRef]   [PubMed]  

23. E. G. Reynaud, J. Peychl, J. Huisken, and P. Tomancak, “Guide to light-sheet microscopy for adventurous biologists,” Nat. Methods 12(1), 30–34 (2015). [CrossRef]   [PubMed]  

24. T. Uchihashi, H. Watanabe, S. Fukuda, M. Shibata, and T. Ando, “Functional extension of high-speed AFM for wider biological applications,” Ultramicroscopy 160, 182–196 (2016). [CrossRef]   [PubMed]  

25. S. Fukuda, T. Uchihashi, R. Iino, Y. Okazaki, M. Yoshida, K. Igarashi, and T. Ando, “High-speed atomic force microscope combined with single-molecule fluorescence microscope,” Rev. Sci. Instrum. 84(7), 073706 (2013). [CrossRef]   [PubMed]  

26. P. Hapala, G. Kichin, C. Wagner, F. S. Tautz, R. Temirov, and P. Jelínek, “Mechanism of high-resolution STM/AFM imaging with functionalized tips,” Phys. Rev. B 90(8), 085421 (2014). [CrossRef]  

27. B. Knoll and F. Keilmann, “Near-field probing of vibrational absorption for chemical microscopy,” Nature 399(6732), 134–137 (1999). [CrossRef]  

28. M. Cascione, V. de Matteis, R. Rinaldi, and S. Leporatti, “Atomic force microscopy combined with optical microscopy for cells investigation: AFM Combined with Optical Microscopy,” Microsc. Res. Tech. 80(1), 109–123 (2016).

29. A. Lewis, H. Taha, A. Strinkovski, A. Manevitch, A. Khatchatouriants, R. Dekhter, and E. Ammann, “Near-field optics: from subwavelength illumination to nanometric shadowing,” Nat. Biotechnol. 21(11), 1378–1386 (2003). [CrossRef]   [PubMed]  

30. B. Hecht, B. Sick, U. P. Wild, V. Deckert, R. Zenobi, O. J. Martin, and D. W. Pohl, “Scanning near-field optical microscopy with aperture probes: Fundamentals and applications,” J. Chem. Phys. 112(18), 7761–7774 (2000). [CrossRef]  

31. Z. B. Wang, N. Joseph, L. Li, and B. S. Luk’yanchuk, “A review of optical near-fields in particle/tip-assisted laser nanofabrication,” Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 224(5), 1113–1127 (2010). [CrossRef]  

32. N. Mauser and A. Hartschuh, “Tip-enhanced near-field optical microscopy,” Chem. Soc. Rev. 43(4), 1248–1262 (2014). [CrossRef]   [PubMed]  

33. F. Wang, L. Liu, H. Yu, Y. Wen, P. Yu, Z. Liu, Y. Wang, and W. J. Li, “Scanning superlens microscopy for non-invasive large field-of-view visible light nanoscale imaging,” Nat. Commun. 7, 13748 (2016). [CrossRef]   [PubMed]  

34. T. Xiang, G.-S. Xia, X. Bai, and L. Zhang, “Image Stitching by Line-guided Local Warping with Global Similarity Constraint,” ArXiv Prepr. ArXiv170207935 (2017).

35. L. Juan and G. Oubong, “SURF applied in panorama image stitching,” in Image Processing Theory Tools and Applications (IPTA),20102nd International Conference on (IEEE, 2010), pp. 495–499. [CrossRef]  

36. J. Jia and C.-K. Tang, “Image Stitching Using Structure Deformation,” IEEE Trans. Pattern Anal. Mach. Intell. 30(4), 617–631 (2008). [CrossRef]   [PubMed]  

37. A. Zomet, A. Levin, S. Peleg, and Y. Weiss, “Seamless image stitching by minimizing false edges,” IEEE Trans. Image Process. 15(4), 969–977 (2006). [CrossRef]   [PubMed]  

38. R. Szeliski and H.-Y. Shum, “Creating full view panoramic image mosaics and environment maps,” in Proceedings of the 24th Annual Conference on Computer Graphics and Interactive Techniques (ACM Press/Addison-Wesley Publishing Co., 1997), pp. 251–258. [CrossRef]  

39. H. Yang, R. Trouillon, G. Huszka, and M. A. M. Gijs, “Super-resolution imaging of a dielectric microsphere is governed by the waist of its photonic nanojet,” Nano Lett. 16(8), 4862–4870 (2016). [CrossRef]   [PubMed]  

40. M.-S. Kim, T. Scharf, S. Mühlig, C. Rockstuhl, and H. P. Herzig, “Engineering photonic nanojets,” Opt. Express 19(11), 10206–10220 (2011). [CrossRef]   [PubMed]  

41. A. V. Itagi and W. A. Challener, “Optics of photonic nanojets,” J. Opt. Soc. Am. A 22(12), 2847–2858 (2005). [CrossRef]   [PubMed]  

42. P. Ferrand, J. Wenger, A. Devilez, M. Pianta, B. Stout, N. Bonod, E. Popov, and H. Rigneault, “Direct imaging of photonic nanojets,” Opt. Express 16(10), 6930–6940 (2008). [CrossRef]   [PubMed]  

43. A. Heifetz, S.-C. Kong, A. V. Sahakian, A. Taflove, and V. Backman, “Photonic Nanojets,” J. Comput. Theor. Nanosci. 6(9), 1979–1992 (2009). [CrossRef]   [PubMed]  

44. S. Lee, L. Li, Y. Ben-Aryeh, Z. Wang, and W. Guo, “Overcoming the diffraction limit induced by microsphere optical nanoscopy,” J. Opt. 15(12), 125710 (2013). [CrossRef]  

45. Z. Wang, W. Guo, L. Li, B. Luk’yanchuk, A. Khan, Z. Liu, Z. Chen, and M. Hong, “Optical Virtual imaging at 50 nm lateral resolution with a white-light nanoscope,” Nat. Commun. 2, 218 (2011). [CrossRef]   [PubMed]  

46. T. R. Kumar and R. V. Krishnaiah, “Optical Disk with Blu-Ray Technology,” ArXiv Prepr. ArXiv13101551 (2013).