We demonstrate a simple, in-line scheme for determining the duration of ultrafast pulses in the focal region of a high-numerical-aperture microscope objective. Photocurrent generated in a GaAsP photodiode by two-photon absorption of orthogonally-polarized laser beams that meet at a slight angle is used to autocorrelate lasers non-interferometrically. Crosscorrelation between two lasers is also demonstrated. This setup, which can be built readily by a microscope user who is not an optics expert, allows for the rapid characterization of pulses that can be hundreds of fs long while making it possible for all of the laser intensity to be employed for nonlinear optical microscopy after the pulse duration has been measured.
© 2006 Optical Society of America
Ultrafast lasers are finding increasing use for a range of applications in nonlinear optical microscopy [1–6] and fabrication [7–9]. Nonlinear processes are highly dependent upon pulse length, and at the same time high-numerical-aperture (NA) optics contain many optical elements and so can broaden pulses significantly. As a result, in recent years a number of sophisticated and powerful techniques have been developed that allow for the detailed characterization and optimization of laser pulses as short as a few optical cycles at or near the focus of a microscope objective.
There is a large and rapidly growing group of users of systems that couple ultrafast lasers with microscopes who are experts in neither optics nor lasers. It would be helpful for such users to be able to characterize (and optimize) laser pulses in the focal region of their microscopes. However, for a number of reasons such measurements are rarely made by nonexpert users. First, the construction of a sophisticated pulse characterization setup may require considerable expertise in optics. Second, for most users the pulse characterization setup must be compatible with a microscope that employs epifluorescence detection. Third, most schemes for characterization of pulses at the focal plane of an objective are not in-line, and are therefore less likely to be used on a routine basis. Fourth, non-expert users are likely to employ lasers that have pulse durations of hundreds (rather than tens) of fs and to have laboratories that are not interferometrically stable; such conditions are not suitable for the majority of existing characterization techniques for tightly focused pulses, many of which are based on inteferometric autocorrelation (IA).
Two major approaches have been used for microscope autocorrelation to date. One of these approaches is frequency-resolved optical gating (FROG) [10, 11]. FROG is a powerful technique that allows for measurement of both the instantaneous magnitude and phase of the electric field of a laser pulse . FROG requires the measurement of the spectrum of the signal from a nonlinear optical process as a function of the delay between two pulses. In the context of microscopy, a practical consequence of this requirement is that a second objective must be used to collect the signal for spectral analysis. Thus, FROG is not ideally suited for use in microscopes in which excitation and signal collection are accomplished by the same objective.
The other major approach that has been used for characterizing pulses after microscope objectives is IA [13–19]. In IA, the incoming light generally traverses a Michelson interferometer, the output of which is focused through the objective and into a nonlinear medium such as a second-harmonic crystal or a two-photon-absorbing material. To characterize the pulses, the nonlinear-optical signal is collected as a function of the delay between the two arms of the interferometer. There are many detection schemes that can be employed for IA that allow a single objective to be employed, such as epidetection of twophoton-absorption (TPA) induced fluorescence. However, IA requires interferometric stability and the step size must be considerably smaller than the wavelength of light in order to collect a high-quality interferogram. As a result, IA is best suited for the characterization of pulses that are at most a few tens of fs long. In addition, while IA can be used for in-line autocorrelation (in which the autocorrelation optics are a permanent part of the excitation path), one arm of the interferometer generally must be blocked when the laser beam is to be used for microscopy, leading to the loss of 75% of the laser power. Thus, in many implementations of IA for microscope pulse characterization the autocorrelation setup is not in-line.
Ideally a microscope autocorrelator for the non-expert user should be an in-line system that can be used to characterize laser pulses that are anywhere from a few fs to hundreds of fs in duration while allowing all of the laser power to reach the microscope. For such users, pulse characterization need not extend beyond measuring the duration, but the technique must be straightforward and rapid. Here we introduce a non-interferometric, in-line setup for microscopes that allows for rapid autocorrelation of ultrafast laser pulses and for crosscorrelation between pulses from two sources without sacrificing any laser intensity.
2. Experimental setup
Photodiodes that have a bandgap that is larger than the energy of a single photon of an ultrafast laser pulse but that generate photocurrent efficiently upon multiphoton absorption (MPA) have proven to be excellent nonlinear media for performing autocorrelations [14, 20–22]. With simple modifications, such photodiodes can be used readily for autocorrelation in the focal region of even a high-NA, oil-immersion, microscope objective .
The vast majority of pulse characterization schemes using MPA-induced photocurrent have been interferometric in nature. In fact, due to the crystal symmetry of many photodiode materials, even orthogonally-polarized laser beams can exhibit strong interference upon MPA in these media . Thus, although using a Mach-Zender interferometer with polarizing beam splitters is a common trick for combining two laser pulses without loss, such a scheme still results in an interferometric autocorrelation with such diodes.
In order to perform non-interferometric autocorrelation, we employ a polarizing Mach-Zender interferometer that is slightly misaligned such that the beams are not collinear but rather meet at a small angle. In the crossing region of the beams there is a spatially-varying polarization pattern that can be envisioned as the sum of four single-polarization interference patterns: one that is linearly polarized at 45°, one that is right circularly polarized and is shifted spatially by a quarter period, one that is linearly polarized at -45° and is shifted spatially by another quarter period, and one that is left circularly polarized and is shifted spatially by an additional quarter period [24, 25]. Unequal efficiency for the MPA of linearlyand circularly-polarized light leads to different photocurrents for these two polarizations (which is why interference can be seen in the MPA-induced photocurrent for orthogonallypolarized beams). However, so long as the crossing angle is large enough to create a few fringes of the polarization pattern, each polarization will be represented equally in the crossing region, and the photocurrent will be insensitive to phase shifts between the beams.
Our experimental setup is shown in Fig. 1. A Ti:sapphire laser that is tuned to roughly 800 nm passes through a polarizer and a half-wave plate (HWP) before entering a polarizing beam cube (PBC). The HWP is adjusted so that, upon exiting the PBC, roughly equal intensities are created of vertically-polarized light traveling in one direction and horizontallypolarized light traveling in the other direction. One beam traverses a computer-controlled delay line, and the two beams are recombined, with a small angle between them, in a second PBC. Before recombination, the two beams are chopped at different frequencies by two rings of a chopper wheel. The beams are magnified with a telescope before being introduced into the microscope and sent through a 1.45 NA, 100× objective (Zeiss α Plan-FLUAR). A GaAsP photodiode (Hamamatsu G1117 with the resin coating removed) is placed in the focal region of the objective, and is connected to a lock-in amplifier that is referenced to the sum of the chopping frequencies. Autocorrelations are obtained by measuring the TPA-induced photocurrent as a function of the delay-line position. When autocorrelation is complete, the photodiode is removed and a sample is placed on the microscope stage. The leg of the setup for which the output beam does not pass down the optical axis of the objective is blocked. The input HWP is used to control the intensity of the light that traverses the leg of the setup that passes directly down the optical axis of the objective to reach the sample. The HWP is capable of directing all of the laser intensity to the sample stage if desired.
To perform autocorrelations, mirror FM2 (Fig. 1) is removed from the optical path. If mirror FM1 is in the optical path then laser 1 is autocorrelated, whereas if FM1 is not in the optical path laser 2 is autocorrelated. To perform cross-correlations, mirror FM1 is removed from the beam path and mirror FM2 is inserted. The HWPs are set so that the beam from laser 2 passes directly through both PBCs and the beam from laser 1 is reflected from second PBC. In this case the beams are aligned so as to be collinear upon exiting the second PBC. For cross-correlations, active stabilization was achieved using a Coherent SynchroLock apparatus. Laser 1 (Coherent Mira Basic) was slaved to laser 2 (Coherent Mira 900-F). Typical average laser powers at the photodiode for auto- and cross-correlation were on the order of 1 mW or less.
To find the optimum crossing angle between the beams, autocorrelations of laser 1 were obtained as a function of the number of fringes in the beam overlap region. The number of fringes was measured by inserting a polarizer set to pass light polarized at 45° into the beam path after the second PBC. The fringes in the overlap region were counted, and then the polarizer was removed.
Shown in Fig. 2 are autocorrelations obtained with different numbers of fringes. The highest-quality data were obtained with the photodiode placed slightly beyond the focal spot of the objective, where the overlap of the two beams is greater when they are not collinear. The data for collinear beams are in good agreement with previously-reported results for autocorrelation of orthogonally-polarized beams with a GaAsP photodiode . The modulation depth of the autocorrelation fringes is roughly 1/3 of the total amplitude under these conditions. The 100-nm step size of our delay line is clearly insufficient to resolve the autocorrelation fringes fully. Even with a delay line with considerably higher resolution, the lack of interferometric stability would make it impossible to obtain acceptable IA data. The modulation depth of the autocorrelation fringes is reduced considerably when there are two complete fringes in the crossing region, and by the time there are six fringes in the crossing region the modulation amplitude reaches an insignificant fraction of the total autocorrelation amplitude. All further experiments were performed with six fringes in the crossing region.
Shown in Fig. 3 are single-scan autocorrelations of lasers 1 and 2. For laser 1 the step size used in the autocorrelation was 7.5 µm (50 fs) and for laser 2 the step size used was 10 µm (66.7 fs). No attempt was made to compress the output of either laser, so that we could judge the time required to obtain high-quality autocorrelations for pulses that are typical of the longest that might be measured from a Ti:sapphire oscillator of modest bandwidth after passing through optics and a microscope objective. The scans shown here were collected in 3 to 4 sec, which is fast enough that adjustments to a laser or a pulse compressor can be made in real time to improve the pulse duration. Both autocorrelations fit well to Gaussian profiles, with a 240 fs FWHM pulse length for laser 1 and a 550 fs FWHM pulse length for laser 2. In both cases the bandwidth can support considerably shorter pulses. Note that for these autocorrelations to have been performed interferometrically, the step sizes used would have had to have been more than two orders of magnitude smaller, and the data acquisition time would have been correspondingly longer.
Also shown in Fig. 3 is a cross-correlation of lasers 1 and 2. Due to timing jitter between the two lasers, which occurs on multiple time scales, the cross-correlation is considerably noisier than the autocorrelation even though the pulses cannot interfere with one another. The data shown in Fig. 3 are the result of averaging 200 scans, which took approximately 35 minutes at a step size of 25.0 µm (167 fs). In a system with less timing jitter the crosscorrelation could be obtained considerably more rapidly. There is usually no adjustment made to an experimental setup based on the width of a cross-correlation, and so a longer scan time is not as much of a problem as it would be for autocorrelations.
The FWHM of the cross-correlation scan in Fig. 3 is 1.65 ps, as compared to the ideal value of 600 fs predicted from the autocorrelations (red trace). These results imply that the RMS timing jitter between the pulses from lasers 1 and 2 is 650 fs.
We have demonstrated an in-line, non-interferometric auto- and cross-correlator for characterizing ultrafast laser pulses in the focal region of a microscope objective. The technique is simple and can be implemented readily by a non-expert user for the real-time measurement and optimization of pulse duration. We have used this device to autocorrelate laser pulses that are greater than 100 fs in duration, which is typical for many of the commercial ultrafast laser sources used in microscopy. Even for such long laser pulses, autocorrelations can be obtained in a few seconds in a laboratory that is not interferometrically stable. This technique can also be used to measure the duration of considerably shorter pulses, although it does not provide as much information as do more sophisticated techniques. In addition, the use of a polarizing Mach-Zender interferometer allows for the full intensity of one or both lasers to be delivered to the microscope objective after pulse characterization.
This work was supported by the Air Force Office of Scientific Research, Grant F49620-01-1-0455. We thank Prof. Thomas Murphy for helpful discussions and Amity Ziegler and Juliet Znovena for their contributions to this research.
References and links
2. L. Moreaux, O. Sandre, M. Blanchard-Desce, and J. Mertz, “Membrane imaging by Simultaneous Second- Harmonic Generation and Two-Photon Microscopy,” Opt. Lett. 25, 320–322 (2000). [CrossRef]
4. J. X. Cheng, A. Volkmer, L. D. Book, and X. S. Xie, “An Epi-Detected Coherent Anti-Stokes Raman Scattering (E-CARS) Microscope with High Spectral Resolution and High Sensitivity,” J. Phys. Chem. B 105, 1277–1280 (2001). [CrossRef]
5. T. A. Klar, S. Jakobs, M. Dyba, A. Egner, and S. W. Hell, “Fluorescence microscopy with diffraction resolution barrier broken by stimulated emission,” Proc. Natl. Acad. Sci. USA 97, 8206–8210 (2000). [CrossRef]
6. E. O. Potma, W. P. de Boeij, and D. A. Wiersma, “Femtosecond Dynamics of Intracellular Water Probed with Nonlinear Optical Kerr Effect Microspectroscopy,” Biophys. J. 80, 3019–3024 (2001). [CrossRef]
7. T. Baldacchini and J. T. Fourkas, “Three-dimensional nanofabrication using multiphoton absorption,” in Encyclopedia of Nanoscience and Nanotechnology, J. A. Schwarz, C. I. Contescu, and K. Putyera, eds. (Marcel Dekker, New York, 2004), pp. 3905–3915.
8. D. Yang, S. J. Jhaveri, and C. K. Ober, “Three-dimensional microfabrication by Two-Photon Lithography,” MRS Bull. 30, 976–982 (2005). [CrossRef]
9. H. B. Sun and S. Kawata, “Two-Photon Laser Precision Microfabrication and its applications to micronano devices and systems,” J. Lightwave Technol. 21, 624–633 (2003). [CrossRef]
10. D. N. Fittinghoff, J. A. Squier, C. P. J. Barty, J. N. Sweetser, R. Trebino, and M. Muller, “Collinear Type II Second-Harmonic-Generation frequency-resolved Optical Gating for use with high-numerical-aperture objectives,” Opt. Lett. 23, 1046–1048 (1998). [CrossRef]
11. I. Amat-Roldan, I. G. Cormack, P. Loza-Alvarez, and D. Artigas, “Starch-based Second-Harmonic- Generated Collinear Frequency-Resolved Optical Gating Pulse Characterization at the Focal Plane of a High-Numerical-Aperture Lens,” Opt. Lett. 29, 2282–2284 (2004). [CrossRef]
12. R. Trebino, Frequency-Resolved Optical Gating: The Measurement of Ultrashort Laser Pulses (Springer, Norwell, MA, 2002). [CrossRef]
13. M. Muller, J. Squier, and G. J. Brakenhoff, “Measurement of Femtosecond Pulses in the Focal Point of a High-Numerical-Aperture Lens by 2-Photon Absorption,” Opt. Lett. 20, 1038–1040 (1995). [CrossRef]
14. A. C. Millard, D. N. Fittinghoff, J. A. Squier, M. Muller, and A. L. Gaeta, “Using GaAsP Photodiodes to Characterize Ultrashort Pulses Under High Numerical Aperture Focusing in Microscopy,” J. Microsc.- Oxford 193, 179–181 (1999). [CrossRef]
15. K. S. Wong, T. Sun, B. K. K. Fung, I. K. Sou, and G. K. L. Wong, “Visible to ultraviolet femtosecond autocorrelation measurements based on two-photon absorption using ZnSSe photodetector,” J. Cryst. Growth 227, 717–721 (2001). [CrossRef]
16. D. N. Fittinghoff, J. A. der Au, and J. Squier, “Spatial and temporal characterizations of Femtosecond Pulses at High-Numerical Aperture using Collinear, background-free, Third-Harmonic Autocorrelation,” Opt. Commun. 247, 405–426 (2005). [CrossRef]
18. F. Quercioli, A. Ghirelli, B. Tiribilli, and M. Vassalli, “Ultracompact autocorrelator for Multiphoton Microscopy,” Microscopy Research and Technique 63, 27–33 (2004). [CrossRef]
19. F. Cannone, G. Chirico, G. Baldini, and A. Diaspro, “Measurement of the Laser Pulse Width on the Microscope Objective Plane by Modulated Autocorrelation Method,” J. Microsc. 210, 149–157 (2003). [CrossRef]
20. A. M. Streltsov, K. D. Moll, A. L. Gaeta, P. Kung, D. Walker, and M. Razeghi, “Pulse autocorrelation measurements based on Two- and Three-Photon Conductivity in a GaN Photodiode,” Appl. Phys. Lett. 75, 3778–3780 (1999). [CrossRef]
21. P. Langlois and E. P. Ippen, “Measurement of Pulse Asymmetry by Three-Photon-Absorption Autocorrelation in a GaAsP Photodiode,” Opt. Lett. 24, 1868–1870 (1999). [CrossRef]
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]
23. S. Santran, M. E. Martinez-Rosas, L. Canioni, and L. Sarger, “Characterization of Optical Nonlinearity in semiconductor photodiodes using cross-polarized autocorrelation,” IEEE J. Quantum Electron. 40, 1687–1694 (2004). [CrossRef]
24. H. J. Eichler, P. Günter, and D. W. Pohl, Laser-Induced Dynamic Gratings (Springer-Verlag, New York, 1986).
25. J. T. Fourkas, R. Trebino, and M. D. Fayer, “The Grating Decomposition Method: A new approach for understanding polarization-selective transient grating experiments. I. Theory,” J. Chem. Phys. 97, 69–77 (1992). [CrossRef]