Broadband, sub-10-fs pulses, can be propagated through polarization-maintaining single mode fiber (PMF) for use in nonlinear optical microscopy (NLOM). We demonstrate delivery of near transform-limited, 1 nJ pulses from a Ti:Al2O3 (75 MHz repetition rate) oscillator via PMF to the NLOM focal plane while maintaining 120 nm of bandwidth. Negative group delay dispersion (GDD) introduced to pre-compensate normal dispersion of the optical fiber and microscope optics ensured linear pulse propagation through the PMF. The minimized time-bandwidth product of the laser pulses at the NLOM focus allowed the nonlinear excitation of multiple fluorophores simultaneously without central wavelength tuning. Polarization sensitive NLOM imaging using second harmonic generation in collagen was demonstrated using PMF delivered pulses. Two-photon excited fluorescence spectra and second harmonic images taken with and without the fiber indicates that the fiber based system is capable of generating optical signals that are within a factor of two to three of our traditional NLOM.
© 2008 Optical Society of America
Nonlinear optical microscopy (NLOM) utilizing femtosecond laser pulses is a proven tool for imaging living tissues. However, further characterization of tissues in their native environment requires miniaturization of bench top microscopy systems into portable, optical fiber based imaging systems[2, 3]. Optical fiber pulse delivery systems[4–7] provide a simple and efficient platform on which to develop miniaturized microscopes[8–10] and microendoscopes[11–13] for tissue interrogation without excision. A difficult challenge in developing these imaging systems is preserving the temporal and spectral characteristics of femtosecond laser pulses at the focal plane. The susceptibility of ~100 fs pulses, common in current systems, to severe nonlinear broadening limits pulse energies delivered to the specimen of tens of picojoules or requires complex optical systems to control. Here, we report the optical fiber delivery of high energy (≥1 nJ), dispersion minimized sub-10-fs pulses to NLOM focus for imaging using simple dispersion compensation.
Femtosecond pulse propagation through single mode optical fiber (SMF) is governed by the interplay of linear (dispersion) and nonlinear (i.e., self-phase modulation, SPM) effects arising within the core. Group velocity dispersion (GVD) will broaden the pulse over a characteristic length (factor of 1.4 for Gaussian pulses) given by, , where T o and β 2 (40 ps2/km) are the laser (transform-limited) pulse duration and GVD coefficient, respectively. It is estimated that LD<3 mm in SMF for a sub-10-fs pulse. Nonlinearities in SMF will similarly broaden the pulse by SPM over a nonlinear interaction length, , where P o and γ (5 W-1km-1) are the input peak power and SMF nonlinearity coefficient, respectively. For 1 nJ pulse energy, significant pulse broadening is estimated to occur within a few millimeters for a sub-10-fs pulse. The ratio, , indicates that nonlinear pulse propagation is mitigated by dispersion mediated pulse broadening, and hence, peak intensity reduction, occurring on the same length scale.
Fiber nonlinear effects have been minimized by either manipulating fiber properties or pulse shaping methods. Large mode area fibers have been used to decrease, but not eliminate, nonlinearity[4, 6, 15]. A novel high dispersion, few mode fiber has been developed where the dispersion interaction length is sufficiently shorter than the nonlinear interaction length allowing dispersion dominated pulse propagation for 150 fs pulses at 1 nJ. Reconstruction of 100 fs pulses following sequential propagation through two SMFs has been demonstrated. Self-phase modulation in the first SMF spectrally (and temporally) broadened the pulse. A negatively dispersive element compensated for normal dispersion in both fibers and SPM of pulses with negative group delay dispersion (GDD) in the second SMF narrowed the pulse spectrum resulting in pulses of the same duration as the input. Coherent pulse shaping methods[18, 19] relying on spatial light modulators have been used to compensate for nonlinear pulse distortion, however, high system losses limit power within the fiber. Highly chirped 50 fs pulses have been delivered through short pieces of SMF at moderate power levels, although, significant post fiber normal dispersion was required and spectral narrowing was still observed.
Hollow-core photonic crystal fibers (HC-PCF’s) have allowed the distortion free delivery of 100 fs pulses propagating at and near the zero dispersion wavelength of the fiber. The steep dispersion slope, limited bandgap (~60–80 nm), and zero dispersion wavelength in the middle of the bandgap make these fibers unsuitable for broadband pulse propagation. Additionally, 100 fs pulse systems relying on HC-PCF’s require adjustment of dispersion compensation for each new central wavelength, which is critical for fiber lengths longer than a few centimeters. This complicates the imaging of samples labeled with multiple fluorophores that do not have well overlapping excitation spectra.
A common method to deliver pulses with net-zero GDD to the focus of an optical fiber based NLOM is to pre-compensate normal dispersion from the SMF and downstream microscopy optics with negative GDD. Following the SMF, the remaining negative GDD necessary to compensate dispersion of strong focusing optics is approximately 1,000 fs2. The peak intensity of 100 fs pulses having 1,000 fs2 GDD is reduced by <4%, and is not sufficient to reduce fiber nonlinearity. This is consistent with previous observations of nonlinear propagation of 100 fs pulses in optical fiber causing spectral narrowing of the pulse and a corresponding increase in pulse duration detrimental to efficient optical signal generation for NLOM. In contrast, this amount of GDD reduces peak intensity of sub-10-fs pulses by a factor ≥25, suggesting dispersion dominated propagation through SMF in microscopy applications. The high susceptibility of sub-10-fs pulses to dispersion will mitigate nonlinearity in optical fiber delivery at pulse energies in excess of 1nJ without relying on expensive, exotic waveguides or complex optical systems. Furthermore, the broad pulse spectrum of the delivered ultrashort laser pulse may be used for multiple fluorophore excitation without laser tuning and for efficient signal generation. Our system represents a significant improvement over those previously reported in terms of simplicity and reduced cost while increasing imaging efficiency and versatility.
2. PMF delivery of femtosecond pulses
In our design, sub-10-fs pulses originating from a Ti:Al2O3 oscillator (Femtosource, Femtolasers) were stretched to ~8.5 ps (~22,000 fs2 negative GDD) using three pairs of dispersion compensation mirrors (Femtolasers) before being coupled into ~400 mm of bare PMF (PM780 HP, Nufern). Normal dispersion of the PMF temporally recompressed the pulse to ~1.5 ps. The peak intensity of a 1 nJ pulse broadened to 1.5 ps in duration was not sufficient ( ) to produce nonlinear pulse broadening within the PMF. Final recompression occurred while passing through the optics of the NLOM (~4000 fs2) to deliver net zero GDD at the focal plane, schematically illustrated in Fig. 1. An achromatic zero-order wave plate was used to align the laser polarization with an axis of the PMF. A near-infrared corrected achromat was used to couple the pulses (2 nJ) into PMF with ~50% efficiency. Fiber length was limited by the amount of negative GDD imparted by the dispersion compensation mirrors and throughput efficiency of the optical system.
Pulse spectra preceding and following the PMF, shown in Fig. 2(a), were measured by inserting a portable spectrometer (USB2000, Ocean Optics) into the beam path following the fiber collimating lens. The post fiber spectrum exhibited a modulated structure that most likely resulted from modal interference from the elliptical beam. Signal generated in a GaAsP (two-photon) photodiode was measured with respect to PMF laser input energy and exhibited a quadratic dependence indicating linear pulse propagation through the fiber as shown in Fig. 2(b). The modulations in the pulse spectrum remained with <60pJ coupled into the PMF. These same characteristics were observed when using isotropic (non-polarization maintaining) SMF.
2.1 Femtosecond pulse characterization for nonlinear optical microscopy
Pulse duration was measured by interferometric autocorrelation at the focal plane of the NLOM imaging system. An Interferometric autocorrelation of the pulse delivered through PMF at the NLOM focal plane is shown in Fig. 3(a). The collimated beam was directed through a Michelson interferometer with collinear beams steered to the microscope passing through a 1.5X telescope and reflected off the primary dichroic mirror (635dcspxruv3p, Chroma) to a Zeiss 40X Achroplan (0.8NA) water immersion objective. The objective focused the beam through a water column onto a GaAsP photodiode with a 1 mm thick borosilicate window. Fine dispersion control was provided by inserting BK7 glass windows (1 mm thickness increments) into the beam path after the fiber.
The measured pulse duration of 12.7 fs was calculated from the central peak FWHM assuming a sech2 pulse shape. Oscillations at long delays of the autocorrelation were the result of residual chirp and modulations in the post fiber pulse spectrum, see inset Fig. 3(a). The width of the central peak was identical to a previous measurement for this objective, shown in Fig. 3(b), suggesting optimal dispersion management through PMF. Estimates of β 2 = 40 fs2/mm and β 3 = 30 fs3/mm (third-order dispersion, TOD) for PMF were based on our measurements and for silica. TOD mismatch in the fiber arose from TOD/GDD ≈ 0.68 for our dispersion compensation mirrors. Noticeably absent from the autocorrelation in Fig. 3(a) were side lobes resulting from chromatic and spherical aberration of the objective lens as observed in Fig. 3(B). Side lobes were not present in Fig. 3(a) because the final beam diameter filled 35% of the back focal aperture compared to 70% without the fiber to minimize beam clipping in the interferometer. It should be noted that the effects of spherical aberration, chromatic aberration and a radially dependent dispersion profile of the objective ultimately limit the pulse duration at the sample. Minimizing these effects requires well corrected objective lenses for good pulse maintenance making the now common practice of using gradient index (GRIN) objectives for in-vivo nonlinear microscopic imaging incompatible with laser pulses of such short duration.
3. Assessment of PMF delivered pulses to NLOM
Imaging performance was measured by comparing two-photon excited fluorescence (TPF) of three common biological fluorophores Indo-1 (Molecular Probes), FITC (Sigma) and TRITC (Sigma), and image analysis of second harmonic generation (SHG) in collagen from rat skin. Fluorescent dyes were dissolved in their appropriate solvents to a concentration of 100 µM. TPF was generated by focusing laser pulses into individual dye solutions using the 40X Achroplan objective and detected in the backscattered direction by the focusing objective. Fluorescence signal was coupled, non-descanned, into a multimode optical fiber connected to a spectrometer (SpectraPro 2300i, Roper Scientific) shown in Fig. 1. Back scattered laser light was filtered using 3 mm BG-39 glass. Relative excitation efficiency was determined by comparing the emission intensity from each dye excited with the laser delivered through air and by the PMF. A longer focal length collimating lens for the fiber was used to ensure matched objective NA.
where E(ω) is the Fourier transform of electric field E(t) and Ω is an iterative variable that ensures integration over non-degenerate (sum frequency) and degenerate (second harmonic) frequency combinations. Assuming transform limited pulses, calculated T(ω) using the laser and post fiber spectrum are shown in Fig. 4(a). The probability of non-resonant two-photon absorption is proportional to the overlap integral[26, 27],
where γ(ω) is the molecular two-photon absorption profile. γ(ω) for the three dyes was estimated by measuring TPF intensity as a function of central wavelength (two-photon photoluminescence excitation spectrum) of a narrowband, 170 fs Ti:Al2O3 laser (Mira 900F, Coherent). It was assumed that γ(ω) of the dyes was directly proportional with their two-photon photoluminescence excitation spectra and are shown in Fig. 4(b). The laser Τ(ω) is shown for reference and indicates that our broadband pulse was well suited to excite the different fluorophores simultaneously.
PMF, with respect to air, delivered pulses generated TPF from individual Indo-1, FITC and TRITC solutions to within 43.7%, 42.2% and 36.0%, respectively. Using Equation 2, it was calculated PMF delivered pulses should generate TPF signal within 10.4%, 10.1% and 13.5% for Indo-1, FITC and TRITC, respectively, compared to air delivered pulses. The measured and calculated TPF comparisons were brought into closer agreement by including residual chirp in the PMF delivered pulse of <30 fs2 (~.8 mm fused silica) and 1,200 fs3 (estimated TOD mismatch). This calculation does not take into account focal volume differences from slightly different degrees of collimation of the incident beams. All three dyes were combined and their simultaneous TPF emission profiles are shown in Fig. 5 excited using PMF and air delivered pulses. TPF emission spectra were normalized to the TRITC emission peak. The TPF spectra indicate a disproportionate decrease in the high energy side of the PMF delivered two-photon excitation power spectrum more than likely resulting from (additional) residual chirp in the pulse.
NLOM images of rat skin using SHG in collagen are shown in Fig. 6 obtained with air (a) and PMF delivered pulses (b and c). Skin was excised from recently sacrificed 6 week old Sprague-Dawley rats. NLOM images were acquired at 0.0625 Hz frame rate and averaged over four frames from the dermal side at a depth of 15µm. SHG signal is collected in the backscattered direction and directed non-descanned onto a PMT depicted in Fig. 1. A 430dcxru long pass dichroic mirror (Chroma) and HQ405/40 bandpass filter (Chroma) are placed in front of the PMT. Intensity analysis was performed by averaging pixel intensity over each set of images. The incident laser polarization orientations are indicated by double headed arrows. The images were from the same focal plane and showed similar collagen morphology and resolution. Intensity analysis of NLOM images indicated PMF delivered pulses were capable of generating second harmonic signals to within a factor of 2 to 3. The measured polarization extinction ratio of the laser following PMF was 170:1 compared to 500:1 for the oscillator. A dominant uniaxial component of collagen second-order susceptibility has been measured aligned along the fiber axis. From Fig. 6(b) to (c), the incident polarization was rotated 90° to demonstrate spectroscopic contrast was maintained with PMF delivered pulses. It was observed that fibers aligned along the incident polarization were preferentially highlighted in the SHG images.
We have demonstrated sub-10-fs pulses can be delivered via 400 mm of PMF to the focus of an NLOM system with minimal spectral and temporal distortion. Our data suggest that longer PMF lengths can be used with proportional addition of dispersion compensation (cf. Fig. 2(b)). Nonlinear optical signal generation, particularly TPF and SHG, by PMF delivered pulses was shown to be within a factor of 2 to 3 of air delivered pulses. The bandwidth of PMF delivered pulses was sufficiently wide to excite multiple fluorophores simultaneously eliminating the need for central wavelength tuning and concomitant adjustments in dispersion compensation. Furthermore, polarization was maintained providing an additional mechanism of image contrast for optical fiber based NLOM. Our design provides a simple, efficient, and experimentally flexible platform to develop miniaturized nonlinear microscopy and microendoscopy imaging systems motivated by in vivo multi-molecular imaging studies in small animals.
We wish to thank Jason Hirshburg for providing rat skin to image, Chao Wang for collecting the dye two-photon excitation photoluminescence spectra, Professor Kenith Meissner for use of the Mira and SpectraPro 2300i spectrometer, Femtolasers and Microscopy & Imaging Center at Texas A & M University for providing additional chirped mirrors, and NuFern for providing the optical fibers. This work was funded by a NSF Faculty Early Career Development (CAREER) Award.
References and links
1. A. T. Yeh, H. Gibbs, J.-J. Hu, and A. M. Larson, “Advances in nonlinear optical microscopy for visualizing dynamic tissue properties in culture,” Tissue Eng. Part B Rev. 14, 119–131 (2008). [CrossRef] [PubMed]
5. S.-P. Tai, M.-C. Chan, T.-H. Tsai, S.-H. Guol, L.-J. Chen, and C.-K. Sun, “Two-photon fluorescence microscope with a hollow-core photonic crystal fiber,” Opt. Express 12, 6122–6128 (2004). [CrossRef] [PubMed]
8. F. Helmchen, M. S. Fee, D. W. Tank, and W. Denk, “A miniature head-mounted two-photon microscope: High resolution brain imaging in freely moving animals,” Neuron 31, 903–912 (2001). [CrossRef] [PubMed]
9. W. Gobel, J. N. D. Kerr, A. Nimmerjahn, and F. Helmchen, “Miniaturized two-photon microscope based on a flexible coherent fiber bundle and a gradient-index lens objective,” Opt. Lett. 29, 2521–2523 (2004). [CrossRef] [PubMed]
10. C. J. Engelbrecht, R. S. Johnston, E. J. Seibel, and F. Helmchen, “Ultra-compact fiber-optic two-photon microscope for functional fluorescence imaging in vivo,” Opt. Express 16, 5556–5564 (2008). [CrossRef] [PubMed]
11. B. A. Flusberg, J. C. Jung, E. D. Cocker, E. P. Anderson, and M. J. Schnitzer, “In vivo brain imaging using a portable 3.9 gram two-photon fluorescence microendoscope,” Opt. Lett. 30, 2272–2274 (2005). [CrossRef] [PubMed]
12. L. Fu, A. Jain, H. Xie, C. Cranfield, and M. Gu, “Nonlinear optical endoscopy based on a double-clad photonic crystal fiber and a MEMS mirrors,” Opt. Express 14, 1027–1032 (2006). [CrossRef] [PubMed]
14. G. P. Agrawal, Nonlinear Fiber Optics (Acadmic Press, San Diego, 2007).
15. D. G. Ouzounov, K. D. Moll, M. A. Foster, W. R. Zipfel, W. W. Webb, and A. L. Gaeta, “Delivery of nanojoule femtosecond pulses through large-core microstructured fibers,” Opt. Lett. 27, 1512–1515 (2002). [CrossRef]
16. S. Ramachandran, M. F. Yan, J. Jasapara, P. Wisk, S. Ghalmi, E. Monberg, and F. V. Dimarcello, “High energy (nanojoule) femtosecond pulse delivery with record dispersion higher-order mode fiber,” Opt. Lett. 30, 3225–3227 (2005). [CrossRef] [PubMed]
17. S. W. Clark, F. O. Ilday, and F. W. Wise, “Fiber delivery of femtosecond pulses from a Ti:sapphire laser,” Opt. Lett. 26, 1320–1322 (2001). [CrossRef]
18. S. H. Lee, A. L. Cavalieri, D. M. Fritz, M. Myaing, and D. A. Reis, “Adaptive dispersion compensation for remote fiber delivery of near-infrared femtosecond pulses,” Opt. Lett. 29, 2602–2604 (2004). [CrossRef] [PubMed]
19. M. Lelek, E. Suran, F. Louradour, A. Barthelemy, B. Viellerobe, and F. Lacombe, “Coherent femtosecond pulse shaping for the optimization of a non-linear micro-endoscope,” Opt. Express 15, 10154–10162 (2007). [CrossRef] [PubMed]
20. B. Rozsa, G. Katona, E. S. Vizi, Z. Varallyay, A. Saghy, L. Valenta, P. Maak, J. Fekete, A. Banyasz, and R. Szipocs, “Random access three-dimensional two-photon microscopy,” Appl. Opt. 46, 1860–1865 (2007). [CrossRef] [PubMed]
21. W. Gobel, A. Nimmerjahn, and F. Helmchen, “Distortion-free delivery of nanojoule femtosecond pulses from a Ti:sapphire laser through a hollow-core photonic crystal fiber,” Opt. Lett. 29, 1285–1287 (2004). [CrossRef] [PubMed]
22. J. K. Ranka, A. L. Gaeta, A. Baltuska, M. S. Pshenichnikov, and D. A. Wiersma, “Autocorrelation measurement of 6-fs pulses based on the two-photon-induced photocurrent in a GaAsP photodiode,” Opt. Lett. 22, 1344–1346 (1997). [CrossRef]
24. D. Meshulach and Y. Silberberg, “Coherent quantum control of two-photon transitions by a femtosecond laser pulse,” Nature 396, 239–242 (1998). [CrossRef]
25. J. P. Ogilvie, K. J. Kubarych, A. Alexandrou, and M. Joffre, “Fourier transform measurement of two-photon excitation spectra: applications to microscopy and optimal control,” Opt. Lett. 30, 911–913 (2005). [CrossRef] [PubMed]
26. D. Meshulach and Y. Silberberg, “Coherent quantum control of multiphoton transitions by shaped ultrashort optical pulses,” Phys. Rev. A. 60, 1287–1292 (1999). [CrossRef]
27. K. A. Walowicz, I. Pastirk, V. V. Lozovoy, and M. Dantus, “Multiphoton Intrapulse Interference. 1. Control of multiphoton processes in condensed phases,” J. Phys. Chem. A 106, 9369–9373 (2002). [CrossRef]
28. P. Stoller, B.-M. Kim, A. M. Rubenchik, K. M. Reiser, and L. B. Da Silva, “Polarization-dependent optical second-harmonic imaging of rat-tail tendon,” J. Biomed. Opt. 7, 205–214 (2002). [CrossRef] [PubMed]