In this study, a microscope based on spatiotemporal focusing offering widefield multiphoton excitation has been developed to provide fast optical sectioning images. Key features of this microscope are the integrations of a 10 kHz repetition rate ultrafast amplifier featuring high instantaneous peak power (maximum 400 μJ/pulse at a 90 fs pulse width) and a TE-cooled, ultra-sensitive photon detecting, electron multiplying charge-coupled camera into a spatiotemporal focusing microscope. This configuration can produce multiphoton images with an excitation area larger than 200 × 100 μm2 at a frame rate greater than 100 Hz (current maximum of 200 Hz). Brownian motions of fluorescent microbeads as small as 0.5 μm were observed in real-time with a lateral spatial resolution of less than 0.5 μm and an axial resolution of approximately 3.5 μm. Furthermore, second harmonic images of chicken tendons demonstrate that the developed widefield multiphoton microscope can provide high resolution z-sectioning for bioimaging.
© 2012 OSA
Since 1990, multiphoton excited (MPE) fluorescence microscopy has been extensively used for biological imaging . With superior features such as minimum invasiveness, low photobleaching, and deep penetration depth, multiphoton microscopy has proven to be particularly suitable for imaging thick tissues and living animals . Furthermore, second harmonic generation (SHG), another phenomenon of nonlinear optics, can be effectively employed to acquire non-centrosymmetric contour information directly within specimens without labeling, such as collagen within tissue . As has been documented, a high numerical aperture (NA) objective lens and an ultrafast laser for spatially and temporally generating an extremely high electromagnetic field are needed for the excitation of both two-photon excited fluorescence (TPEF) and SHG. TPEF and SHG occur only in a small region near the focal point of the objective lens; hence, high spatial signal-to-noise ratio (SNR) signals for three-dimensional (3D) imaging of specimens can be obtained via pixel by pixel beam scanning. In order to increase the multiphoton imaging frame rate, techniques such as the line scanning system and multifocal multiphoton microscope have been developed [3–5].
Recent studies have shown that simultaneous spatial and temporal focusing techniques can provide widefield and axial-resolved multiphoton excited imaging [6–12]. The spatiotemporal focusing microscope typically consists of a diffraction grating, a collimating lens, and a high NA objective lens . When the laser pulse impinges on the grating, spatial separation of the pulse spectrum is attained. The pulse spectrum are then collected by the collimating lens, propagate along the optical axis, and focused into the sample from different angles by the high NA objective lens. Only in the focal plane do the different frequency components construct in phase and produce a short, high-peak power pulse, allowing effective multiphoton excitation to occur [6,7,9–12]. Compared to conventional beam scanning multiphoton excitation microscopy, widefield multiphoton excitation microscopy using the temporal focusing technique detects the overall fluorescence and harmonic generation signal of the entire illumination area, which depends on the laser beam spot size and the magnification of the microscope . The advantage of widefield microscopy is that less time is required to capture one frame, enabling a fast frame rate for capturing dynamic events. A video-rate multiphoton microscope has been realized by the combination of temporal focusing and mechanical line scanning . However, the disadvantage of the mechanical scanning technique is that the entire frame is not illuminated simultaneously, limiting its maximum frame rate. When a fast, high-sensitivity camera and an ultrahigh peak power laser are employed in widefield multiphoton excitation microscopy, an image rate of a few hundred frames per second can be achieved. However, there is inherent difficulty for the camera-based detection to reach the same SNR, spatial resolution, and penetration depth as the photomultiplier tube-based single photon counting (SPC) technique, which effectively reduces the background noise and the scattering effect of emission light [13,14].
Differing from fixed biosample studies, interactions between living cells, the migration of cells, and the signal transition of neuron cells all require real-time microscopic imaging [15,16]. Examples of this include flagellated cells and the beating heart of embryonic zebra fish, both of which require high frame-rates per second for observation. Although widefield multiphoton excitation microscopy based on spatiotemporal focusing can in principle observe samples in real-time, factors such as the emission efficiency of the excited fluorophores, the peak power of the ultrafast laser in the focal plane, and the sensitivity of the camera can all affect the frame rate. In this paper, a 4.0 W titanium-sapphire (ti-sa) ultrafast amplifier with a 10 kHz repetition rate was used to enhance peak power. In addition, an electron multiplying charge-coupled device (EMCCD) was employed, and by cooling the EMCCD chip to −80 þC, background noise and dark current are reduced. Furthermore, by shrinking the number of pixels or binning the pixels, the frame rate is increased to 100 Hz with a pixel number of 256 × 256. Since typical fluorophores have a finite number of excitation emission cycles before photobleaching, shorter exposure times can provide longer fluorescence observations [17,18]. By reducing the exposure time with a mechanical shutter to 20 ms the photobleaching issue was overcome. In the experimental results, Brownian motions of fluorescent microbeads, with sizes ranging from 1.0 μm to 0.5 μm, were successfully observed with a lateral spatial resolution of less than 0.5 μm and an axial resolution of 3.5 μm at the 100 Hz frame rate. Additionally, SHG images of chicken tendon at 100 Hz frame rate were obtained, demonstrating that the widefield multiphoton microscope can provide high resolution z-sectioning for bioimaging in real-time.
2. Optical setup
2.1. Widefield multiphoton microscope
Figure 1 illustrates a schematic of the developed widefield multiphoton microscope based on spatiotemporal focusing. Key components include a ti-sa ultrafast amplifier (Spitfire Pro., Newport, USA), a ti-sa ultrafast oscillator (Tsunami, Spectra-Physics, USA) as the seed beam of the amplifier, an upright optical microscope (Axio imager 2, Carl Zeiss, Germany), a triple-axis sample positioning stage (H101A ProScanTM, Prior, UK), an Andor EMCCD camera (iXonEM + 885 EMCCD, Andor, UK), and a data acquisition (DAQ) card with a field-programmable gate array (FPGA) module (PCI-7831R, National Instruments, USA). The ultrafast amplifier has a peak power of 400 μJ/pulse, with a pulse width of 90 fs and a repetition rate of 10 kHz.
A half-wave plate and a polarizer adjust the polarization and power of the amplifier. The beam is spatially dispersed via a grating with 1200 lines/mm. The incident angle of the grating was adjusted to ensure the central frequency follows the optical axis and propagates through the 4f setup, which comprises the collimating lens and the objective lens (W Plan-Apochromat 40X/ NA 1.0, Carl Zeiss, Germany). By filtering the collected signal through a dichroic mirror and a short-pass filter, only nonlinear optical signals are collected through the objective lens and imaged onto the EMCCD camera. By controlling the motorized stage in the z-axis via the FPGA, sequential sectioning images at different depths can be obtained, and then assembled to reconstruct a 3D image.
After diffraction from the grating at an oblique incident angle of 69.4°, the laser’s beam attains an elliptical cross section with the two axes at 200 and 100 μm in the image plane. However, refractive optical elements, such as the objective lens and collimating lens, can lead to additional group velocity dispersion disturbing the overall system. To rectify this problem, the built-in grating compressor of the amplifier was adjusted to approach the optimal pulse width (<120 fs). Consequently, the efficiency of temporal focusing for excited nonlinear signals can be enhanced, allowing for a nonlinear image of a specimen to be attained with ten laser pulses at 1 ms exposure time. And as aforementioned, the 20 ms mechanical shutter speed reduces photobleaching.
2.2. System calibration
The axial resolution of the widefield multiphoton microscopy based on the optical parameters in full-width at the half maximum (FWHM) can be expressed as :Fig. 2(a) . It can be seen that the axial resolutions are 2.5, 2.8, 3.1, and 3.4 μm at the respective laser powers of 10, 20, 30, and 40 mW. When the laser power was increased from 10 to 40 mW, the axial resolution degraded slightly; however, an axial resolution of less than 3.4 μm can be provided as the maximum laser power is less than 40 mW.
The NA of the objective lens, the excitation wavelength, and the multiphoton mechanism together determine the spatial resolution . According to the optical parameters used in this study, the theoretical lateral resolution is approximately 0.4 μm. Fluorescent beads with a diameter of 0.2 μm were used to examine the lateral resolution. Figure 2(b) shows that the point spread function (PSF) of a single bead is approximately 0.5 μm. By taking the average PSF of 10 beads, the FWHM lateral resolution of the microscope was determined to be 0.5 μm, which is close to that of the theoretical result given the size of the fluorescent beads.
3. Experimental results and discussions
Photobleaching of fluorophores causes decay of their emission intensity. One particular concern here is photobleaching when exposed to a high peak intensity light field for a long period of time. Figure 3 shows the behavior of TPEF for the R6G thin film described in Sec. 2.2 when continuously illuminated for 1 minute by the ti-sa ultrafast oscillator and the ultrafast amplifier. By using the ultrafast oscillator at powers of 40, 50, and 58 mW, the florescent intensity does not decay and reveals no photobleaching. However, fluorescent intensity dramatically decayed and the photobleaching phenomenon occurred when the ultrafast amplifier with excitation powers greater than 10 mW was used. The average power of the ti-sa ultrafast oscillator is higher than that of the ultrafast amplifier; however, the instantaneous pulse energy of the ultrafast amplifier is 8,000 times greater than that of the ultrafast oscillator and consequently induces the photobleaching effect. Therefore, high excitation peak power with long-term exposure should be avoided in multiphoton excited fluorescence. Under the same excitation power of 10 mW, the exposure times for grabbing similar image intensity are 16.5 ms and 60 s for the ultrafast amplifier and the ultrafast oscillator, respectively. The exposure time by using the ultrafast amplifier is 3,600 times faster than that by the ultrafast oscillator and hence the ultrafast amplifier provides high efficiency for the TPEF excitation.
3.2. Brownian motions of fluorescent microbeads
The high frame-rate capability of the developed microscope was validated by observing the Brownian motions of fluorescent microbeads. Brownian motion is a random process involving small particles moving in arbitrary directions due primarily to particle collisions and thermal perturbations. According to the Stokes-Einstein Eq , root mean square (RMS) displacement of a small particle in 3D space between a time interval can be evaluated by:Eq. (2) with a diffusion coefficient for a 1.0 μm diameter fluorescent bead , the RMS displacement at a frame rate of 12 Hz is calculated as 496 nm, which is less than the bead’s radius of 500 nm. As such, a frame rate of 12 Hz should be necessary to observe the Brownian motion of 1.0 μm fluorescent beads. However, since particles will move in random directions (lateral or axial directions) and is only slightly less than the bead’s radius, a higher frame rate would be a better choice. For example, the RMS displacement at a frame rate of 100 Hz based on Eq. (2) is 172 nm for a 1.0 μm fluorescent bead. Beads with diameters of 1.0 and 0.5 μm immersed in DI water were sealed within the cavity of a slide measuring 18 mm in diameter and 0.8 mm in depth. Figure 4(a) shows an image of the Brownian motions for the 1.0 μm diameter fluorescent beads using the higher frame rate of 100 Hz, where the excitation laser power is 40 mW, the field of view is 50 x 50 μm2, and the exposure time is 9 ms per frame. A weighted algorithm performed on the images of the 1.0 μm fluorescent beads can provide sub-pixel resolution for centroid determination. Figure 4(b) shows the displacements of the bead at the top of Fig. 4(a) along the x and y axes as a function of time at 10 ms interval from Media 1. The maximum displacements along the x and y axes are about 140 and 150 nm, respectively, which is comparable to the simulation result of 172 nm.
A video featuring the 100 Hz frame rate clearly demonstrates the Brownian motion trajectories in the observed plane. Since the smaller the fluorescent bead’s diameter is, the faster the Brownian motion is; thus, for smaller particles, a faster frame rate is needed. For a 0.5 μm diameter fluorescent bead, the RMS displacement at a frame rate of 94 Hz is 250 nm, which is equal to the bead’s radius, and so the frame rate must be 94 Hz or faster for observing 0.5 μm fluorescent beads in 3D space without losing trajectory information. The fluorescent volume of a 0.5 μm fluorescent bead is also reduced by 23 compared to that of a 1.0 μm fluorescent bead. Figure 5(a) shows an image of the Brownian motion at 100 frames per second with an increased excitation laser power of 60 mW to obtain brighter fluorescence images. Figure 5(b) shows 4 sequential shots of the 0.5 μm fluorescent beads at a frame rate of 100 Hz. If the frame rate in Fig. 5(b) were lower at 33 frames/s, only the leftmost and the rightmost shots could be observed, and pertinent information between shots would be lost. The average displacement of the 0.5 μm fluorescent beads in Fig. 5(b) is 185 nm. Based on the simulation, the displacement of a 0.5 μm fluorescent bead in 3D space projected into a two-dimensional observation plane is 196 nm when the probabilities of the displacement in three directions are assumed to be equal, which is in close agreement with our observation.
3.3. Second harmonic generation imaging
Collagen possesses a triclinic microcrystalline structure, which can produce SHG signals efficiently [21,22]. The SHG signal is proportional to the square of the excitation intensity and can be collected based on forward (transmission) and backward (backscattering) schemes. Superior image quality is typically attained from forward rather than backward schemes ; however, backscattering is preferable for observing thick biosample slices or live animals. It should be noted that intensities of label-free signals of biosamples, such as SHG from collagens and TPEF from autofluorescence dyes, are generally weak, resulting in degraded image quality and reduced frame rate. Harmonic generation with temporal focusing has been realized and theoretically analyzed . This study targeted future applications of live label-free animals at a high frame rate; hence, a backward scheme with a high excitation laser peak power and fast frame rate were used.
The SHG images of chicken tendon in a phosphate-buffered saline solution were acquired at a high frame rate and illuminated by the ultrafast amplifier at a wavelength of 780 nm. Generated SHG signals were collected by the EMCCD camera through a dichroic mirror, a 680 nm short-pass filter, and an additional 390 nm band-pass filter. As before, the EMCCD chip was cooled to −80 þC, leading to reduce the noise level. The SHG images of the chicken tendon are shown in Fig. 6 . Figure 6(a) shows chicken tendon collagen imaged at the laser power of 30 mW. The pixel number is 1000 × 1000 pixels (illumination area of approximately 200 × 200 μm2), with an exposure time of 100 ms, demonstrating a larger field of view is possible, but the frame rate then decreases to only 10 Hz. Using the binning process of the EMCCD camera, a faster frame rate and a reduced sample scattering effect can be obtained with a larger sensing area per pixel. However, in doing so, spatial resolution is sacrificed. The pixels of the EMCCD camera were binned 4 × 4, i.e. a total pixel number was reduced to 250 × 250 pixels per image, but the 200 × 200 μm2 view field remained unchanged. The exposure time was 9 ms with a frame rate up to 100 Hz. Figure 6(b) shows the chicken tendon collagen of the same region imaged at 100 frames per second under the binning setting. The SHG image of the chicken tendon collagen shown in Fig. 6(c) was taken with the pixel number reduced to 256 × 256 pixels without binning but with a reduced field of view near the red square region indicated in Fig. 6(a), at a frame rate of 100 Hz, and an exposure time of 9 ms. Figure 6(d) shows ten shots of the same red square region accumulated in one frame. Compared with Fig. 6(c), Fig. 6(d) provides better image quality with less speckle noise. Because the coherent nature of SHG can produce artifacts in performing widefield coherent microscopy , a high NA objective could be used to reduce the artifacts. For observing label-free samples, the developed widefield multiphoton microscopy, with more than 3 μJ/pulse of ultrafast, amplified peak energy and the use of a highly sensitive, TE-cooled EMCCD camera allows high-speed, multiphoton imaging to be achieved.
The first bioimage, cells stained with DAPI and captured by spatiotemporal focusing-based widefield multiphoton microscopy with a ti-sa ultrafast oscillator as light source, had an exposure time of 30 seconds for one image within a 140 × 140 μm2 area . The excitation power was too low to excite fluorescence with such a large area, and hence the advantage of fast imaging of widefield microscopy wasn’t realized. In terms of excitation, multiphoton microscopy with spatiotemporal focusing through deep scattering tissue can maintain its axial sectioning ability . However, the artifacts of widefield coherent harmonic generation and sample scattering will blur or distort the image when imaging the nonlinear emission to the detector. To overcome the artifacts and achieve an image quality near the diffraction limit or above, i.e. super resolution, while simultaneously maintaining fast imaging characteristics the widefield multiphoton microscopy based on spatiotemporal focusing could integrate image improvement techniques such as the adaptive optics system , structure illumination microscopy (SIM) [27,28], photoactivated localization microscopy (PALM), stochastic optical reconstruction microscopy (STORM) [29–31], and background rejection HiLo microscopy . Moreover, the spatiotemporal focusing-based widefield multiphoton microscopy with a high peak power amplifier can be integrated with patterned generation devices to achieve fast multiphoton microfabrication of freeform polymer microstructures [33,34]. Compared with the beam scanning process, this approach improves fabrication speed by two or three orders .
In this study, we have developed a widefield, multiphoton excitation microscope based on spatiotemporal focusing. In addition to capturing both TPEF and SHG images at a frame rate of over 100 Hz, high axial resolution was also achieved. The photobleaching issue was overcome by reducing the exposure time with a mechanical shutter to 20 ms. Brownian motion of fluorescent microbeads with diameters as small as 0.5 μm was observed in real-time for determination of their RMS displacements. Moreover, fast SHG images of chicken tendon demonstrated that our widefield multiphoton excitation microscope can provide high resolution sectioning for bioimaging. The results achieved in this study suggest that our technique may be extended to label-free, high-speed imaging in biological samples.
This work was supported by the National Science Council (NSC) in Taiwan with grant numbers NSC 99-2627-B-006-017, NSC 99-3111-B-006-004, and NSC 100-2623-E-006-016-D.
References and links
3. G. J. Brakenhoff, J. Squier, T. Norris, A. C. Bliton, M. H. Wade, and B. Athey, “Real-time two-photon confocal microscopy using a femtosecond, amplified Ti:sapphire system,” J. Microsc. 181(3), 253–259 (1996). [CrossRef] [PubMed]
4. A. H. Buist, M. Müller, J. Squier, and G. J. Brakenhoff, “Real time two-photon absorption microscopy using multipoint excitation,” J. Microsc. 192(2), 217–226 (1998). [CrossRef]
9. A. Vaziri, J. Tang, H. Shroff, and C. V. Shank, “Multilayer three-dimensional super resolution imaging of thick biological samples,” Proc. Natl. Acad. Sci. U.S.A. 105(51), 20221–20226 (2008). [CrossRef] [PubMed]
11. E. Papagiakoumou, F. Anselmi, A. Bègue, V. de Sars, J. Glückstad, E. Y. Isacoff, and V. Emiliani, “Scanless two-photon excitation of channelrhodopsin-2,” Nat. Methods 7(10), 848–854 (2010). [CrossRef] [PubMed]
12. O. D. Therrien, B. Aubé, S. Pagès, P. D. Koninck, and D. Côté, “Wide-field multiphoton imaging of cellular dynamics in thick tissue by temporal focusing and patterned illumination,” Biomed. Opt. Express 2(3), 696–704 (2011). [CrossRef] [PubMed]
15. P. J. Keller and E. H. K. Stelzer, “Quantitative in vivo imaging of entire embryos with digital scanned laser light sheet fluorescence microscopy,” Curr. Opin. Neurobiol. 18(6), 624–632 (2008). [CrossRef] [PubMed]
17. J. R. Lakowicz, Principles of Fluorescence Spectroscopy (Springer, 2006).
18. A. Van Orden, N. P. Machara, P. M. Goodwin, and R. A. Keller, “Single-molecule identification in flowing sample streams by fluorescence burst size and intraburst fluorescence decay rate,” Anal. Chem. 70(7), 1444–1451 (1998). [CrossRef] [PubMed]
20. R. M. Mazo, Brownian Motion: Fluctuations, Dynamics, and Applications (Oxford University Press, 2009).
21. T. A. Theodossiou, C. Thrasivoulou, C. Ekwobi, and D. L. Becker, “Second harmonic generation confocal microscopy of collagen type I from rat tendon cryosections,” Biophys. J. 91(12), 4665–4677 (2006). [CrossRef] [PubMed]
23. D. Oron and Y. Silberberg, “Harmonic generation with temporally focused ultrashort pulses,” J. Opt. Soc. Am. B 22(12), 2660–2663 (2005). [CrossRef]
24. D. Oron and Y. Silberberg, “Third-harmonic generation with cylindrical Gaussian beams,” J. Opt. Soc. Am. B 21(11), 1964–1968 (2004). [CrossRef]
25. E. Papagiakoumou, A. Be`gue, O. Schwartz, D. Oron, and V. Emiliani, “Shaped two-photon excitation deep inside scattering tissue,” submitted for publication.
26. C.-Y. Chung, K.-C. Cho, C.-C. Chang, C. H. Lin, W. C. Yen, and S. J. Chen, “Adaptive-optics system with liquid-crystal phase-shift interferometer,” Appl. Opt. 45(15), 3409–3414 (2006). [CrossRef] [PubMed]
27. M. A. A. Neil, R. Juškaitis, and T. Wilson, “Method of obtaining optical sectioning by using structured light in a conventional microscope,” Opt. Lett. 22(24), 1905–1907 (1997). [CrossRef] [PubMed]
30. 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]
34. Y.-C. Li, L.-C. Cheng, C.-H. Lien, C.-Y. Chang, N.-S. Chang, P. J. Campagnola, C. Y. Dong, and S.-J. Chen, “Fast multiphoton microfabrication of freeform polymer microstructures by spatiotemporal focusing and patterned generation,” submitted for publication.