Confocal laser scanning fluorescence microscopy is demonstrated using a photonic crystal fiber-based excitation source. A 38 cm-long section of photonic crystal fiber is pumped with femtosecond pulses from a Ti:sapphire laser, and the resultant visible continuum is selectively filtered to provide the peak excitation wavelengths required for a range of fluorescently labeled biological tissue.
©2004 Optical Society of America
Confocal laser scanning microscopy (CLSM) has rapidly become an essential tool in the life sciences laboratory, enabling minimally intrusive optical sectioning of fluorescently prepared cells and tissue at sub-micron resolutions . The majority of current commercially available CLSM systems employ a gas laser, e.g. a helium-neon or argon ion laser, to provide the excitation radiation. However, these lasers have several shortcomings, including the maintenance requirements, short lifetimes, heat generation and high noise levels. Solid-state sources can provide a more robust alternative  but like gas lasers, they are not wavelength flexible and hence limit the range of useful fluorophores that can be excited.
Multi-photon laser scanning microscopy (MPLSM) is increasingly becoming a complementary method to CLSM . For example, excitation of DAPI, which is an ultraviolet-absorbing nucleic acid-specific fluorophore, can be performed efficiently by two-photon excitation while a longer wavelength source such as a Kr/Ar laser can excite alternatively labelled structures, such as nerves or muscles. In MPLSM, typically an ultrashort pulsed infrared emitting laser source provides the high peak power densities required to allow the simultaneous absorption of multiple photons by a fluorophore molecule . Therefore, for a truly comprehensive imaging workstation, several laser sources are often required .
To overcome these concerns, I demonstrate a flexible laser source for CLSM that is based on a laser platform that is ideally suited to MPLSM. I have used a tunable, ultrashort-pulsed Ti:sapphire laser as a pump for anomalously dispersive photonic crystal fiber (PCF) to generate a continuum in the visible range of the spectrum. The wavelength range desired to efficiently match the absorption profile of the fluorophore is then selected from the visible continuum using conventional and inexpensive optical bandpass filters.
This optional modular adaptation means that the same platform source can be applied for both CLSM and MPLSM, reducing the cost and complexity of instrumentation while increasing task flexibility. Potentially, this system could be used for accessing the whole range of fluorophores currently used in CLSM, with the added advantage of using a laser source that can also be used to perform MPLSM, thereby providing a complete laser scanning excitation solution.
2. Photonic crystal fibers
Advances in optical fiber technology through the development of PCF  have transformed the fields of nonlinear fiber optics and laser source development. PCFs are microstructured fibers, where the incident radiation is guided by periodically arranged air holes surrounding a solid silica core that extends the length of the fiber. This design produces a photonic bandgap in the transverse direction that results, for instance, in fibers that are continuously single-mode throughout the visible range . This improved guiding property also enables a reduction in the core diameter down to a few microns. This leads to a significant increase in the propagating peak intensity that is enhanced with the application of ultra-short pulsed radiation. This peak intensity increase is of obvious benefit in the study of nonlinear effects, but also significant is that the exact nature of the microstructure determines the group velocity dispersion (GVD) of the fiber. Typically, the zero-dispersion point, λ 0, in a 1–2 µm core diameter PCF is shifted from the bulk silica value of around λ=1270 nm down to λ=600-800 nm [8, 9]. This means that it is now possible to have a fiber with anomalous dispersion at a convenient input wavelength to accommodate a wider range of commercially available laser sources, such as the Ti:Sapphire laser.
Easily the most impressive manifestation of the intrinsic high nonlinearity in an anomalously dispersive PCF is continuum and white light supercontinuum generation, which can extend well over an optical octave . The specific mechanisms involved in the supercontinuum generation are complex [10, 11] but the key is the ability to form solitons at wavelengths above the zero-point for the group velocity dispersion. The associated and well-known effects of soliton self-frequency shift and shedding of energy to shorter wavelengths due to third-order dispersion provide the broadening, while four-wave mixing tends to fill in any remaining gaps . The notable feature of the PCF is that the small mode area brings extreme prominence to these otherwise often subtle effects.
3. Experimental procedure
A schematic of the experimental configuration is shown in Fig. 1(a). A commercial Ti:Sapphire laser (Coherent, Mira 900-F) with an emission wavelength of λ=750 nm was used as the platform source. This single-mode laser emitted a 76 MHz train of pulses, with the pulse width previously measured to be approximately 250 fs with an interferometric autocorrelation trace consistent with a sech2 pulse shape . This pump radiation was propagated through a Faraday isolator to reduce feedback from the PCF facets and was subsequently focused into a 38 cm length of PCF using an aspheric anti-reflection coated f=+4.5 mm lens with a numerical aperture of 0.4.
The PCF was made of pure silica and had a hexagonal structure of air holes to guide the light in a 1.7 µm core. This arrangement gave rise to a zero dispersion wavelength at λ=670 nm. The fiber therefore had a low and positive dispersion throughout the normal Ti:Sapphire operating range (λ=720-890 nm) and was therefore suitable for continuum generation for pump wavelengths within this range.
The radiation transmitted by the fiber was collimated using an identical lens and was propagated into a commercial scan-head (Bio-Rad 1024ES) coupled to an inverted microscope (Nikon, TE300), as shown in Fig. 1(b). The light entering the scan-head was reflected and manipulated by scanning mirrors towards the microscope. A 40x/1.3 numerical aperture oil-immersion microscope objective lens was used to focus the radiation onto the chosen fluorescently stained sample. Fluorescence resulting from confocal excitation was collected by the same objective lens and propagated through an optical bandpass filter to reject reflected light from the exciting source. The fluorescence was then relayed to a sensitive photomultiplier tube. This signal was used, along with image capture software, to visualise fluorescently stained regions of the sample.
In order to experimentally quantify the spectral output from the fiber, an optical spectrum analyzer with a resolution of 1 nm across the wavelength range investigated was used. Figure 2 shows a typical example of the recorded visible continuum spectrum transmitted through the fiber at a measured average output power of 51 mW. The length of the fiber was chosen to maximise the generation of radiation at visible wavelengths in order to match the single-photon excitation wavelengths of many fluorophores. Spectral flatness is of significance for simultaneous or sequential multiple fluorophore excitation through the CLSM, but in these experiments only one fluorophore was applied to each sample in order to demonstrate the application and hence this effect can be neglected at present. Instead, the convenient broad intensity peaks across the continuum can be selectively filtered to provide ample average power at useful wavelengths for CLSM.
To demonstrate the application of a filtered visible continuum to CLSM, a number of biological samples and fluorophores were imaged. The first subject imaged was a thick fixed tissue sample of guinea pig detrusor (bladder smooth muscle layer) labeled with anti-PGP 9.5 and Alexa 488. Isolated smooth muscle cells from the guinea pig bladder loaded with 2µM of Fluo-4 AM were also imaged. Alexa 488 and Fluo-4 AM have peak absorption wavelengths at 490 nm and 494 nm respectively, therefore in order to excite the applied fluorophores in each sample, a 488±5 nm optical bandpass filter with a transmission of 82% was applied to the visible continuum to select this wavelength range. This wavelength range encompasses a spectral intensity peak in the continuum. An average power of 1.26 mW was measured across this spectral range that was attenuated with a neutral density filter of ND=0.4 and then coupled into the scanhead, as previously described. The average power at the sample was measured to be 620 µW, which is comparable to the average power used from a standard Kr/Ar laser for the purpose of CLSM. The images were taken at a capture rate of 0.95 Hz with a 512×512 pixels box size and were averaged over six consecutive scans.
Figure 3(a) shows a typical confocal fluorescence xy cross-section of the guinea pig detrusor labeled with anti-PGP 9.5 and Alexa 488 obtained at a depth within the sample of approximately 41 µm. Figure 3(b) displays the transmission image of the same region of the sample. In the fluorescence image, it is possible to observe the fluorescently labeled nerves wrapping around a blood vessel over a partial depth. CLSM of this sample was possible over a total depth of 59 µm.
Over a typical imaging period of several hours, no photobleaching or tissue damage was observed with this thick sample, as with conventional Kr/Ar laser confocal excitation. A high contrast ratio was routinely observed, with a signal-to-background ratio that typically exceeded 100:1, which is comparable to excitation via a Kr/Ar laser system.
Figure 4(a) shows a superimposed fluorescence and transmission image of an isolated smooth muscle cell from the guinea pig bladder loaded with 2µM of Fluo-4 AM.
Once more, following continual exposure to radiation over several hours, no cell damage (i.e. membrane blebbing or shape change) or pronounced photobleaching was observed. In 10 separate samples, a high contrast ratio was measured, as indicated by a comparison of the fluorescence signal from the fluo-4 loaded and control (unloaded) cells (Fig. 4(b)).
In conclusion, CLSM using a filtered visible continuum was demonstrated using a commercial ultra-short pulsed Ti:sapphire laser as the platform pump source for an anomalously dispersive PCF. CLSM of both living cells and thick sections of fixed tissue was performed at excitation wavelengths around λ=490 nm using an average power of 620 µW at the sample. This power level is comparable to the average powers applied from a Kr/Ar laser for CLSM.
This simple modular accompaniment to a laser suited to MPLSM means that both CLSM and MPLSM can be performed using the same laser scanning system and microscope, which reduces the cost, maintenance requirements and complexity of instrumentation while increasing the range of existing fluorophores that can be excited. Potentially, this system could be used for accessing the whole range of fluorophores currently used in CLSM, with the added advantage of using a laser source that can also be used to perform MPLSM, thereby providing a complete laser scanning excitation solution.
One potential limitation in applying a continuum source for fluorescence microscopy is the fundamental amplitude noise limit, arising from the amplification of input shot noise. For example, where sensitive detection is employed, such as in photon-counting methods, amplitude noise may affect system stability and hence influence the reproducibility of images . This effect is known to be aggravated with increasingly short pulses and high input power. Investigations into the effect of noise fluctuations on CLSM image reproducibility will be made in due course.
Future studies will focus on exciting alternative fluorescent labels with different absorption properties and exploring the effects of applying a spectrally flatter continuum for CLSM through an improved choice of PCF.
The author is indebted to Ross Davidson and Karen McCloskey (Centre for Biophotonics, University of Strathclyde) for supplying and preparing the biological samples used in this investigation. The author also acknowledges financial support from the Royal Society of Edinburgh.
References and links
1. J.B. PawleyHandbook of Biological Confocal Microscopy, 2nd (Plenum Press, New York, 1995). [CrossRef]
2. J.M. Girkin, A.I. Ferguson, D.L. Wokosin, and A.M. Gurney, “Confocal microscopy using an InGaN violet laser diode at 406 nm,” Opt. Express. 7336–341 (2000); http://www.opticsexpress.org/abstract.cfm?URI=OPEX-7-10-336 [CrossRef] [PubMed]
3. D.L. Wokosin, V. Centonze, J.G. White, D. Armstrong, G. Robertson, and A.I. Ferguson, “All-solid-state ultrafast lasers facilitate multiphoton excitation fluorescence imaging,” IEEE J. Sel. Top. Quantum Electron. 21051–1065 (1996). [CrossRef]
4. W. Denk, J. Strickler, and W.W. Webb, “Two-photon laser scanning fluorescence microscopy,” Science 1073–75 (1990). [CrossRef]
5. D.L. Wokosin, J.M. Squirrell, K.W. Eliceiri, and J.G. White, “Optical workstation with concurrent, independent multiphoton imaging and experimental laser microbeam capabilities,” Rev. Sci. Instrum. 74193–201 (2003). [CrossRef]
8. J..K. Ranka, R.S. Windeler, and A.J. Stentz, “Visible continuum generation in air-silica microstructure optical fibers with anomalous dispersion at 800 nm,” Opt. Lett. 2525–27 (2000). [CrossRef]
9. J.K. Ranka, R.S. Windeler, and A.J. Stentz, “Optical properties of high-delta air silica microstructure optical fibers,” Opt. Lett. 25796–798 (2000). [CrossRef]
11. K.M. Hilligsøe, H.N. Paulsen, J. Thøgersen, S. R. Keiding, and J. J. Larsen, “Initial steps of supercontinuum generation in photonic crystal fibers,” J. Opt. Soc. Am. B. 201887–1893 (2003). [CrossRef]
12. G. McConnell and E. Riis, “Ultra-short pulse compression using photonic crystal fibre,” App. Phys. B. 78557–563 (2004). [CrossRef]
13. K.L. Corwin, N.R. Newbury, J.M. Dudley, S. Coen, S.A. Diddams, K. Weber, and R.S. Windeler, “Fundamental noise limitations to supercontinuum generation in microstructure fiber,” Phys. Rev. Lett. 90113904-1-4 (2003). [CrossRef]