* dantus@msu.edu; phone 1 517 355 9715x314; fax 1 517 353 1793

Two-Photon Laser Scanning Microscopy with Ultrabroad Bandwidth 110 nm FWHM Femtosecond Pulses

Peng Xi, Lindsay R. Weisel, Yair Andegeko, Vadim V. Lovozoy, Marcos Dantus*

Department of Chemistry, Michigan State University, MI 48824 USA

ABSTRACT

Shorter pulses, in theory, should be favorable in nonlinear microscopy and yield stronger signals. However, shorter pulses are much more prone to chromatic dispersion when passing through the microscope objective, which significantly broadens its pulse duration and cancels the expected signal gain. In this paper, multiphoton intrapulse interference phase scan (MIIPS) was used to compensate chromatic dispersion introduced by the 1.45 NA objective. The results show that with MIIPS compensation, the increased signal is realized. We also find that third and higher order dispersion compensation, which cannot be corrected by prism pairs, is responsible for an additional factor of 4.7 signal gain.

Keywords: Two-photon, pulse characterization, third-order dispersion, group velocity dispersion

1. INTRODUCTION Two-photon microscopy has been widely applied in the field of biomedical imaging after its first introduction in 1990 by Webb et al. [1]. The laser pulses used for that work had a pulse duration of ~100 fs, their short pulse duration required a high peak power density to induce the nonlinear optical excitation. The spectral width for such a pulse is typically ~10 nm. Given that two-photon excitation require high peak intensity, shorter pulses should be preferable to improve two-photon imaging. However, currently most research groups and two-photon microscopy vendors are still using 100 fs laser. This is despite great advances in ultrashort laser technology, which has already achieved single-digit fs pulses in compact, commercially available sources [2]. The major barrier to the use of ultrashort laser pulses is the chromatic dispersion introduced by the microscope objective. Dispersion can cause a femtosecond pulse to broaden by up to 100 times. Efforts to reduce second-order dispersion have been quite successful; however, higher-order dispersion reduction remains a challenge. Recently, our group reported the successful use of a 10 fs, 110 nm FWHM spectral width pulse with full dispersion compensation, to achieve higher excitation efficiency [3]. With such an ultrashort pulse, a number of improvements to two-photon microscopy become possible.

At the heart of our research is achieving two-photon excitation with maximum efficiency, while minimizing deleterious processes such as photobleaching and photodamage. Optimization of single-photon excitation requires tuning the excitation source to the maximum of the absorption spectrum. Photons with energy too low or too high are not as efficient and may lead to photodamage. Two-photon excitation has a similar wavelength dependence as single-photon excitation, however, there is an additional dependence on peak intensity required to achieve the simultaneous absorption of two photons. Because peak intensity is proportional to the inverse of the pulse duration, the shorter the pulses the greater the efficiency. Interestingly, the shorter the pulses the broader the bandwidth of the pulse (a consequence of the uncertainty principle). Therefore, ultrashort pulses provide more possibilities for improving two-photon excitation, however it isn’t obvious how to achieve those advantages given the possibility of enhancing unwanted processes.

Photobleaching is a common phenomenon in confocal and multiphoton microscopy, in which the fluorescence intensity decreases along the continuous imaging [4, 5]. Photobleaching is the irreversible damage to the fluorophores during imaging [5, 6]. One major advantage of multiphoton microscopy over confocal microscopy is the limit of photobleaching in the entire region of illumination, as multiphoton microscopy employs an excitation wavelength far from UV region and thus has a higher ratio of transmission and less damage to the sample [7-9]. However, on the focal plane, the photobleaching rate for two-photon excitation can be much higher than that of single-photon excitation [7]. This is because the peak intensity in two-photon microscopy is generally high enough to generate higher-order nonlinear

Multiphoton Microscopy in the Biomedical Sciences VIII, edited by Ammasi Periasamy, Peter T. C. So,Proc. of SPIE Vol. 6860, 68601U, (2008) · 1605-7422/08/$18 · doi: 10.1117/12.763698

Proc. of SPIE Vol. 6860 68601U-12008 SPIE Digital Library -- Subscriber Archive Copy

effects, which are usually considered as the major reason for photobleaching [10]. Similar to photobleaching, photodamage is also an important problem that limits the capabilities of tissue imaging in two-photon microscopy, especially when imaging living samples where minimal perturbation is desired. Thus, how to efficiently cause two-photon excitation while avoiding photobleaching and photodamage is a critical issue to be solved in order to significantly improve two-photon microscopy.

In this work, the 10 fs ultrashort laser pulse was achieved through chromatic dispersion compensation. A custom-built two-photon laser scanning microscope was used to obtain the fluorescence signal. The images with both transform-limited (TL) and group delay dispersion (GDD)-compensation were compared in both tri-stained mouse kidney sample and a carrot sample. As the MIIPS compensated TL pulse can give more efficient excitation, less photobleaching can be obtained through the decrease of the TL pulse excitation power.

2. ULTRASHORT LASER PULSE CHARACTERIZATION AND COMPENSATION WITH MIIPS

2.1 Linear relationship between pulse duration and spectral width

Unlike a continuous-wave laser source, the ultrashort laser pulse has multiple frequencies. The short pulse duration is a fact of the enhancing and canceling of the sum of the spectral complex power density. This can be described as a Fourier transformation:

2

)]([)()( ωω ωϕω deEtI ti∫ += (1)

where )(tI is the temporal intensity of the pulse, )(ωE denotes the spectral intensity, and )(ωϕ describes the phase of the frequency component ω . A TL pulse has zero phase dispersion. Practically, for a TL pulse at central wavelength of 800-nm with a Gaussian spectral intensity slope, 44.0≈∆∆ τω . However, after transmission through a thick optical media such as a multi-element high-NA objective, the phase dispersion becomes dominant.

2.2 Nonlinear signal intensity

The second-order nonlinear excitation can be expressed in terms of multiphoton intrapulse interference (MII) theory [11]. The intensity of )2()2( ωI , can be expressed as

[ ] 22)2()2( )()(exp)()()2()2( ΩΩ−+Ω+Ω−Ω+∝= ∫ diEEEI ωϕωϕωωωω (2)

The signal intensity is proportional to the integral of the product of the amplitude of the field times a complex quantity that is dependent upon the spectral phase of the laser pulses )(ωϕ . Here, Ω is a dummy integration variable that plays the role of spectral de-tuning away from the fundamental frequency, ω . It should be noted that this is based on an ideal upconversion, whereas in most fluorophores the cross-section has a dependency with wavelength [12].

2.3 MIIPS characterization and compensation

Consider a Taylor expansion of )(ωφ around ∆ we can find that

nn

n n))((

!1|)( )( Ω−Ω= ∑Ω ωφωφ (3)

The first term is the common phase factor, and the second term )(' Ωφ denotes group delay of the pulse. From Eq. (1) we can see that, both of them have no effect to the pulse duration. Thus, in Eq. (2) we can find that the maximum of

)2()2( ωI appears at 0)('')('' =Ω−+Ω+ ωϕωϕ . If a known phase function is introduced so that the resulting pulse phase becomes )()()( ωωϕωφ f+= , then the local SHG spectral maximum locates at )(")(" ωωϕ f−= .

Proc. of SPIE Vol. 6860 68601U-2

UltrashortPulse Shaper

fl cH-1H

MIIPS : Two-photon Laser Scanning Microscope

Here we use the sinusoidal function )sin()( ∆+= γωαωf , for which )sin()(" 2 ∆+−= γωαγωf . Thus, from

the matrix of ∆~)( )2(IMaxω , )(" ωϕ can be obtained. We can then integrate )(" ωϕ to get the actual phase. It should

be noted that, the second-derivative )(" ωϕ contains not only the second-order dispersion (SOD) but also all the high-order dispersions (HOD). Because a SLM is used, the inverse phase can be conveniently applied to compensate for the pulse.

3. SYSTEM SETUP The schematic configuration of the microscope is shown in Fig. 1. An ultrashort oscillator with pulse duration of 10fs and repetition rate of 80MHz was used as the excitation source (K&M Labs). The beam was expanded spatially before illuminated on the 300-line/mm grating. A curve-mirror with a focal length of 350mm was used to focus the spectrally expanded beam onto a liquid-crystal SLM. A pair of SF11 prisms was used to introduce negative dispersion for the compensation of the GDD of the system. After passing through the two-photon laser scanning system (described below), a BBO crystal was placed at the focal plane of the objective when collecting the MIIPS SHG signal. An f=75mm lens was used to focus the SHG signal for collection with a fiber based compact spectrometer (Ocean Optics, USB4000).

The two-photon microscopy system used was a custom-made laser scanning two-photon microscopy system based on a Nikon TE200 inverted microscope. After the prism pair, the beam was reflected through a dichroic filter (DC) (Chroma Technologies) onto a pair of galvanometers. The galvanometer scans the beam through a 3X telescope system to the microscopic objective. The telescope was composed of one f=100mm scan lens on the galvanometer side and one f=300mm tube lens on the objective side, to collimate and expand the beam for the infinity-corrected objective. In between the telescope, a periscope was installed to raise the beam height to the mercury lamp port. The mercury lamp port was used to introduce the laser directly to the microscopic objective with no additional optical media presented. The incident beam diameter was 8 mm, so after the telescope the beam diameter is 24 mm, which fills the numerical aperture of the objective.

A

Proc. of SPIE Vol. 6860 68601U-3

CM

G

M

SLM

Fig. 1 10 fs ultrashort two-photon laser scanning microscope. A: schematic diagram of the system; B: pulse shaper diagram; and C: the scanning and detection unit of the two-photon laser scanning microscope.

The galvanometer system (Nutfield Technologies) was controlled with NI PCI-6250 DAQ board, and a BNC2110 adapter. In this work bi-direction scan mode was used, in which the galvanometer was commanded to scan back and forth at the same speed. A driving voltage of +/-1 V is needed for scanning the 100µm with 1V per optical degree. Because the maximum scan speed of the galvanometer for this angle is ~400Hz, we used 256 Hz for the fast axis (line scan) and 0.5 Hz for the slow axis. This allows a 512x512 @1fps, and the pixel rate is 0.26MHz. Since the digitization speed of the PCI-6250 is 1.25MHz, we read out four pixels during each scan pixel and average them to reduce noise.

4. LINEAR DEPENDENCE OF TWO-PHOTON SIGNAL WITH LASER PULSE BANDWIDTH

The first advantage that should result from the use of shorter pulses is that two-photon excitation fluorescence (TPEF) signal should be greatly increased. For TL pulses, TPEF signal should linearly increase with the bandwidth of the pulse. Interestingly, this linear relationship has not been observed when GDD compensation is accomplished using a prism pair arrangement. We compared prism-pair compensation, which corrects only second order dispersion, with full dispersion compensation using MIIPS. The TPEF intensity as a function of laser pulse bandwidth is plotted in Fig. 2, for both conditions. For GDD compensated pulses no advantage is found when the pulses are 50 fs in duration or shorter. However, for TL pulses, the linear relation with bandwidth is observed and shorter pulses lead to greater signal. As the bandwidth increases from 10 nm to 80 nm, TPEF signal increases by a factor of ~8. In both experiments, the excitation intensities were kept constant for all bandwidths at 2.28 mW, the only difference between the two is caused by HOD. For TPEF generation, a red fluorescence standard slide (Chroma Technologies) was employed.

5. GREATER INTENSITY AND REPRODUCIBILITY WITH 10 FS TL PULSES Unlike GDD, which reduces the probability of two-photon excitation, HOD causes multiphoton intrapulse interference (MII) among different frequency components [13]. Destructive interference can lead to complete suppression of TPEF at certain frequencies within the bandwidth of the pulse. The effect of HOD introduced by our 1.45 NA microscope objective is shown in Fig. 3, where we measured the second harmonic spectrum of the TL pulse (black dots), after MIIPS compensation, and the spectrum for a GDD compensated pulse (red dots). Second harmonic generation allows us to measure the second-order component of electric field spectrum,

2(2) ( )E ω , responsible for TPEF signal [13]. Clearly, the GDD compensated pulses have significantly less integrated intensity; they are also shifted in wavelength, another consequence of MII.

B C

Proc. of SPIE Vol. 6860 68601U-4

0 20 40 60 800.0

0.2

0.4

0.6

0.8

1.0

MIIPS GDD-only

TPE

F In

tens

ity [A

. U.]

Spectral Width (nm)

Fig. 2 The effect of increasing bandwidth on integral TPEF signal using TL and GDD compensated pulses.

380 390 400 410 420 430 4400

200

400

600

800

1000

1200

1400

1600

SH

G In

tens

ity [A

. u.]

Wavelength (nm)

MIIPS GDD-only

Fig. 3 The measured second harmonic generation spectra for (a) GDD compensated and (b) MIIPS compensated pulses.

The narrowed 2(2) ( )E ω of the GDD compensated pulses can have detrimental effects in biomedical imaging that go

beyond lower signal intensity. In Fig. 4 we show an image of a mouse kidney (Molecular Probes, F-24630), obtained with a Nikon TIRF PlanApo 60x NA 1.45 oil-immersion objective. As can be seen, only the filamentous actin (red) can be observed when imaging with GDD-only compensated pulses. When the pulses are compensated with MIIPS, all the three dyes are correctly excited, revealing richer biological information. Given that chromatic dispersion is a parameter

Proc. of SPIE Vol. 6860 68601U-5

that varies with the laser and the microscope objectives, it is important to have HOD compensation to ensure reproducible results.

(a) (b)

Fig. 4 Mouse kidney stained with DAPI (nuclei), Alexa-488 (glomeruli and convoluted tubules), and Alexa-568

(filamentous actin), and excited with: (a) GDD compensation; (b) MIIPS compensation. Scale bar: 20 µm.

6. GREATER SIGNAL IN DEPTH-RESOLVED TWO-PHOTON MICROSCOPY To study the signal as a function of penetration depth in a biological sample, the depth-resolved autofluorescence images from a carrot sample were collected. A Nikon Plan 20x 0.6 NA objective was used. Images obtained for TL and GDD compensated pulses are shown in Fig. 5. The signal intensity and the signal-to-noise ratio for TL pulses is a factor of 3 times greater than for GDD compensated pulses.

0µm

0µm

40µm

40µm

80µm

80µm

130µm

130µm

Fig. 5 Images of a transverse slice of carrot obtained at different depths are shown for TL pulses (top) and GDD

compensated pulses (bottom). Scale bar: 50 µm.

Proc. of SPIE Vol. 6860 68601U-6

7. REDUCED PHOTOBLEACHING IN TWO-PHOTON MICROSCOPY The green fluorescence standard (Chroma Technologies) was used to study photobleaching effects because it provided a homogenous medium with fixed chromophores. The laser was focused using the Nikon 60x 1.45 NA objective. For GDD compensation, the excitation power was 9 mW. The excitation power was attenuated to 3 mW in Fig. for MIIPS compensated pulses to keep the fluorescence intensity the same as that of GDD compensation. For these experiments, we continuously scanned a 5 µm line on the sample at a 4 ms/line rate. We find that when the initial fluorescence intensity is kept constant, for MIIPS compensated pulses the timescale is always longer than that of GDD compensation, and the components with longer time scales have a larger value than that of GDD-compensation. When the same excitation power is used, MIIPS compensated pulses give greater signal than those of GDD-compensation even after an extended period of time, which makes the signal with MIIPS compensation always greater than that of GDD-compensation excitation.

0 20 40 60 80 1000.0

0.2

0.4

0.6

0.8

1.0 MIIPSGDD-only

Inte

nsity

[A. U

.]

Time (s)

Fig. 6 Photobleaching curves for MIIPS compensated and GDD compensated pulses at the same initial fluorescence

intensity.

8. CONCLUSION For the past decades two-photon microscopy is done with ~ 100 fs pulses, although the shorter pulse has been already available and can generate more nonlinear signal for two-photon biomedical imaging. In this work, we show that through proper compensation of HOD, the use of 10 fs pulses results in significant advantages in two-photon microscopy. Also a significant increase in the signal intensity and the ability to ensure reproducible excitation of different chromophores was observed. In addition, deeper penetration, and higher signal-to-noise ratio were achieved. We also found that reduced photobleaching when comparing TL versus GDD-compensated pulse excitation, when the laser intensity is adjusted so that both produce the same amount of signal.

Proc. of SPIE Vol. 6860 68601U-7

ACKNOWLEDGEMENTS

We gratefully acknowledge the funding support from the National Science Foundation, Major Research Instrumentation grant CHE-0421047, and single investigator grant CHE-0500661. This work benefited from the participation of Rebekah M. Martin and Daniel Schlam through an undergraduate research opportunity.

REFERENCES

1. W. Denk, J. H. Strickler, and W. W. Webb, "Two-photon laser scanning fluorescence microscopy," Science 248(4951), 73-76 (1990).

2. U. Keller, "Recent developments in compact ultrafast lasers," Nature 424, 831-838 (2003). 3. P. Xi, Y. Andegeko, L. R. Weisel, V. V. Lozovoy, and M. Dantus, "Greater signal, increased depth, and less

photobleaching for two-photon microscopy with 10 Femtosecond Pulses," Optics Communications, (In Press) (2007).

4. T. Hirschfeld, "Quantum efficiency independence of the time integrated emission from a fluorescent molecule," Applied Optics 15(12), 3135-3139 (1976).

5. N. Klonis, M. Rug, I. Harper, M. Wickham, A. Cowman, and L. Tilley, "Fluorescence photobleaching analysis for the study of cellular dynamics," European Biophysics Journal 31(1), 36-51 (2002).

6. P. S. Dittrich and P. Schwille, "Photobleaching and stabilization of. fluorophores used for single-molecule analysis. with one-and two-photon excitation," Applied Physics B: Lasers and Optics 73(8), 829-837 (2001).

7. P. T. C. So, C. Y. Dong, B. R. Masters, and K. M. Berland, "Two-Photon Excitation Fluorescence Microscopy," Annual Review of Biomedical Engineering 2, 399-429 (2000).

8. K. Svoboda and R. Yasuda, "Principles of Two-Photon Excitation Microscopy and Its Applications to Neuroscience," Neuron 50(6), 823-839 (2006).

9. P. T. C. So, H. Kim, and I. Kochevar, "Two-Photon deep tissue ex vivo imaging of mouse dermal and subcutaneous structures," Optics Express 3(9), 339-350 (1998).

10. G. Patterson and D. Piston, "Photobleaching in Two-Photon Excitation Microscopy," Biophysical Journal 78(4), 2159-2162 (2000).

11. V. V. Lozovoy and M. Dantus, "Systematic control of nonlinear optical processes using optimally shaped femtosecond pulses," Chem. Phys. Chem 6, 1970-2000 (2005).

12. C. Xu, W. Zipfel, J. B. Shear, R. M. Williams, and W. W. Webb, "Multiphoton fluorescence excitation: New spectral windows for biological nonlinear microscopy," Proceedings of the National Academy of Sciences 93(20), 10763-10768 (1996).

13. V. V. Lozovoy, I. Pastirk, and M. Dantus, "Multiphoton intrapulse interference IV. Ultrashort laser pulse spectral phase characterization and compensation," Optics Letters 29(7), 775-777 (2004).

Proc. of SPIE Vol. 6860 68601U-8