Pre-chirp technique was used in an Nd-doped fiber amplifier to optimize high-quality 910 nm pulses with the pulses width of 114 fs and pulse energy of 4.4 nJ. The in vivo zebrafish imaging results from our totally home-made microscopy proves our femtosecond Nd fiber laser an ideal source in two-photon microscopic imaging.
© 2016 Optical Society of America
The rapid development of two-photon microscopy (TPM)  in biological and medical researches urged a lower-cost portable femtosecond laser source. Compared to solid-state laser systems, like conventional Ti:sapphire lasers, fiber laser shows apparent advantages , such as low cost, outstanding thermo-optical behavior, compactness, robustness and simplicity of operation, especially for applications requiring specific laser wavelengths.
Even though Ti:sapphire lasers can offer broad and tunable wavelengths (from 690 to 1050 nm) , the most popular wavelengths desired by TPM focus on ~900 nm, corresponding to exciting wavelengths of green fluorescent proteins (GFP), for it can be attached to virtually any protein of interest . Furthermore, the use of GFP has been stimulated by the engineering of mutant GFPs with improved properties. For example, insertion of circularly permuted GFP between CaM and M13 yields Ca2+ indicators that are known as GCaMP , which is a powerful and commonly used tool for systems neuroscience. Therefore, the development of ~900 nm femtosecond fiber lasers has tremendous potential applications as laser sources for TPM.
Two major techniques have been used to build the femtosecond 900 nm pulses: the direct generation from the neodymium doped fiber or the nonlinear frequency conversion or shift. The direct generation is one of the potential ways, since its three-level transition (4F3/2 − 4I9/2) falls in the wavelength bands of ~900 nm. The nonlinear wavelength shift in a photonic crystal fiber (PCF) requires the high-power femtosecond Yb fiber laser [6,7]. Compared with nonlinear frequency conversion, the direct generation of 900 nm laser by Nd-doped fiber is preferred due to its robustness and simplicity. Besides, M. Lang  used two color lasers at 780 nm and 1030 nm to realize 2 photon excitation of GFP, which is novel but also relatively complex.
However, lasing and amplifying at this three-level transition is neither easy because of ground-state absorption and the undesired competition with the four-level transition 4F3/2-4I11/2 (1060-1100 nm) when pumped at 808 nm . To keep the fluorescence intensity high in 4F3/2 − 4I9/2 lasing transition, other elements have to be co-doped in the gain fiber [10,11], plus some specified fiber structures to suppress the four-level transition wavelength of the Nd doped fiber, such as W-type index profile fiber [12,13] and photonic bandgap fiber .
In our previous papers , we have managed to obtain mode locked pulses at 920 nm with such fiber. The output pulse is 217 fs at pulse energy of 1.5 nJ. With such pulses, we have made a pollen-imaging test in a homemade two-photon microscope.
However, limited by the low pulse energy and the long pulse width, that laser is not able to achieve in-vivo biological imaging. Following the analysis of Xu et al. , the time-averaged fluorescence detected photo flux can be expressed as
Due to the limited available power of single-mode fiber coupled 808 nm laser diodes, it is not possible to increase the average output power of the mode locked fiber laser. On the other hand, the pulse width may be shortened by the interplay of the nonlinearity and the dispersion.
We adopted the technique of negative pre-chirp  and post-dechirp amplification . By inserting a grating pair before amplification, we introduced negative dispersion to the pulses. Due to the positive group velocity dispersion (GVD) in the amplification fibers, the negatively pre-chirped pulse will be gradually compressed in the fiber amplifier to reach the high peak power and increase the nonlinear effects, such as the self-phase modulation (SPM), which will cause the broadening of the pulse spectrum. After that the pulse will expand again due to the positive dispersion, and can be compressed to a shorter pulse by using another grating pair.
In this paper, we demonstrate 4.4 nJ, 114 fs pulses at ~910 nm with such a fiber amplifier. Compared to our previous work, the pulse energy increases from 1.5 to 4.4 nJ, which is a factor of 2.93, and the pulse width reduced by the factor of 0.52, therefore the peak power increased by the factor of 5.58 correspondingly. We conducted the in-vivo imaging of a zebrafish, which is labeled with EGFP, by this TPM with such pulses. To our knowledge, this is the first time to study in-vivo biological sample imaging at ~910 nm with a femtosecond Nd:fiber laser as the light source.
2. Experimental setup
The schematic of the experiment is illustrated in Fig. 1. The system consists of the 910 nm femtosecond Nd:fiber laser amplifier and the two-photon microscope.
The femtosecond Nd:fiber laser oscillator was the same as described in Ref.13 with an average power of 46 mW serves as a seed source, and the repetition rate was 27.5 MHz. Two polarization beam splitters (PBS), a Faraday rotator (FR) and a half-wave plate (HWP) compose an isolator placed between the oscillator and the amplifier. We used two sets of polarization-combined 808 nm laser diodes as the pump source from both directions in the fiber amplifier. The gain fiber is 7.2 m long with the absorption of ~8 dB/m at 808 nm and the total input pump power is 1.02 W. The rest fiber in the amplifier is a single mode fiber (HI 780).
A grating pair with a groove density of 1851 lines/mm operating at the incident angle of 58° was used to introduce a negative chirp to the pulses. The pre-chirped pulses are coupled into the amplifier fibers with >70% efficiency. After the amplifier, pulses were de-chirped by a subsequent grating compressor before it was launched into our microscopy.
The microscope consists of a scanning system (6215H, Cambridge Technology), a high-NA microscope objective (Nikon NIR Apo 40x 0.8NA), a 730-nm long-pass dichroic mirror (Semrock), a 535/40 nm band-pass filter (Semrock), a detector (H7422p-40, Hamamatsu) and some lenses. Lenses L1 and L2 (Thorlabs) compose a telescope to collimate and expand the laser beam by 2.5 times. After raster scanned by a galvanometer, the beam passes through the scan lens (LSM04-BB, Thorlabs) and tube lens (AC508-200-B, Thorlabs) and then reached to the objective. The signal photons are collected by the objective and directed to a PMT. A data acquisition card (NI PCI-6110, National Instrument) at sampling rate of 5 MHz is used to realize analogue-to-digital conversion. The lateral resolution of the imaging system is about 620 nm.
3. Experiment results and discussions
The output spectral bandwidth of our laser oscillator is about 10 nm, shown as the black line in Fig. 2. The Fourier transform limited pulse width is 218 fs. The GVD of the gain fiber and single mode fiber is estimated to be similar and about 27.8 fs2/mm, so that the net group delay dispersion (GDD) of the oscillator cavity is around 1.97 × 105 fs2. We tuned the pre-chirp by changing the distance of the first grating pair and then optimized the compressor grating pair accordingly for the shortest duration. By repeating this procedure, we determined the best combination of the grating pairs. In this case, the amount of GDD was calculated by using the distance of the pre-chirp grating pair and shown in the legend. The pulse spectra before the compressor (Fig. 2) and intensity autocorrelation traces after the compressor (Fig. 3) were record under the same values of pre-chirp.
Due to the dispersion introduced in the oscillator, the amplifier operated in a low-nonlinearity regime without the pre-chirp, and the spectrum of the amplified pulses experienced almost no changes after the amplifier. With the increase of the amount of the negative pre-chirp, the SPM went stronger, so that the spectrum got broadened. When the pre-chirp GDD reached −6.1 × 105 fs2, the spectrum broadening process stopped at 20 nm.
Besides, we investigated the dependence of the pulse shaping on pre-chirp after the amplifier. Figure 3 shows that the pulse width decreases with the increasing negative pre-chirp, indicating the growth of the bandwidth of spectrum. The pulse width was shortened from 208 fs, to 95 fs, by –5.4 × 105 fs2 pre-chirp GDD. From that point, the pedestals appear in the pulse profile due to the third order dispersion (Fig. 3(e), 3(f)). The shortest pulse with the lowest third order dispersion (TOD) was 114fs, close to the calculated Fourier transform limited pulse of 108 fs from the pulse spectrum, with –4.9 × 105 fs2 pre-chirp (Fig. 3(d)).
The output power at 910 nm as a function of the pump power is depicted in Fig. 4. The output power reaches 121 mW after compression. The corresponding pulse energy is 4.4 nJ.
We integrated the two-photon microscope and recorded the in-vivo imaging of a bio sample with such the laser amplifier. Figure 5 shows the in vivo 3D-reconstructed imaging of the blood vessel of a two-day-old zebrafish. The vascular epithelial cells are labeled with EGFP in the Tg (kdrl:EGFP) transgenic fish line. For imaging, the fish was raised in embryo medium containing 0.002% phenylthiourea (PTU, Sigma) to suppress pigmentation synthesis. Prior to live imaging, the fish was anaesthetized with 0.01% tricaine (Sigma), to minimize movement artifacts during imaging. After that, the anaesthetized zebrafish was embedded in a 1% ultra-pure agarose (Invitrogen) and immersed in E3 medium containing 0.01% tricaine. The blur in the bottom right area in Fig. 5 was the motion artifacts caused by the beating heart. The laser average power out of the objective we used was about 20 mW. The image stack consists of 580 x-y images (512 × 512 pixels, 1 fps) with the step size of 1 μm, and the field of view (FOV) for the image is 600 μm × 600 μm. Considering the dispersion introduced by the optics in the microscopy, we adjusted the compressor to obtain the best imaging quality.
The clear structure of the vessel can prove the capability of our system to do deep-tissue in vivo imaging. It will be promising for the following integrated portable TPM, which can be broadly used in biological and medical application fields.
We have demonstrated an optimized femtosecond Nd:fiber laser amplifier and applied to the TPM. With the best pre-chirp and post-dechirp, the amplifier produced 114 fs pulses at an output power of 220 mW at 920 nm and the pulse energy of 4.4 nJ. The interplay between the nonlinearity and the dispersion broadened pulse spectrum and reduced the high order dispersion. With this novel fiber laser, we obtained an in vivo two photon fluorescence-image. Our laser system can also be integrated into a 35 × 35 × 9 cm3 box. This Nd:fiber laser shows great application potentials in miniature in vivo two-photon absorption fluorescence spectroscopy and imaging.
National Natural Science Foundation of China (61475008,31327901).
The authors are grateful to Heping Cheng, and Zhuan Zhou from the Institute of Molecular Medicine, Peking University for their assistance in the TPM imaging experiments.
References and links
2. A. C. Millard, P. W. Wiseman, D. N. Fittinghoff, K. R. Wilson, J. A. Squier, and M. Müller, “Third-harmonic generation microscopy by use of a compact, femtosecond fiber laser source,” Appl. Opt. 38(36), 7393–7397 (1999). [CrossRef] [PubMed]
3. C. Xu and W. W. Webb, “Measurement of two-photon excitation cross sections of molecular fluorophores with data from 690 to 1050 nm,” J. Opt.Soc.Am. B 13(3), 481–491 (1996). [CrossRef]
4. J. L. Schwartz, N. A. Bonnet, and G. H. Patterson, “Photobleaching and photoactivation: following protein dynamics in living cells,” Nature Cell Biology 9, 7–14 (2003).
6. J. R. Unruh, E. S. Price, R. G. Molla, L. Stehno-Bittel, C. K. Johnson, and R. Hui, “Two-photon microscopy with wavelength switchable fiber laser excitation,” Opt. Express 14(21), 9825–9831 (2006). [CrossRef] [PubMed]
7. H. W. Chen, Z. Haider, J. Lim, S. Xu, Z. Yang, F. X. Kärtner, and G. Chang, “3 GHz, Yb-fiber laser-based, few-cycle ultrafast source at the Ti:sapphire laser wavelength,” Opt. Lett. 38(22), 4927–4930 (2013). [CrossRef] [PubMed]
8. M. Lang, “Multi-line fibre laser for multi-photon microscopy,” http://www.toptica.com/fileadmin/_migrated/content_uploads/toptica_lang_Photonik_International.pdf.
9. B. J. Ainslie, S. P. Craig, S. T. Davey, D. J. Barber, J. R. Taylor, and A. S. L. Gomes, “Optical and structural investigation of Nd3+ in silica-based fibres,” J. Mater. Sci. Lett. 6(11), 1361–1363 (1987). [CrossRef]
10. K. Arai, H. Namikawa, K. Kumata, T. Honda, Y. Ishii, and T. Handa, “Aluminum or phosphorus codoping effects on the fluorescence and structural properties of neodymium doped silica glass,” J. Appl. Phys. 59(10), 3430–3436 (1986). [CrossRef]
11. P. D. Dragic and G. C. Papen, “Efficient Amplification Using the 4F3/2 − 4I9/2 Transition in Nd-Doped Silica Fiber,” IEEE Photonics Technol. Lett. 11(12), 1593–1595 (1996). [CrossRef]
12. S. Yoo, D. B. S. Soh, J. Kim, Y. Jung, J. Nilsson, J. K. Sahu, J. W. Lee, and K. Oh, “Analysis of W-type waveguide for Nd-doped fiber laser operating near 940 nm,” Opt. Commun. 247(1-3), 153–162 (2005). [CrossRef]
15. X. Gao, W. Zong, B. Chen, J. Zhang, C. Li, Y. Liu, A. Wang, Y. Song, and Z. Zhang, “Core-pumped femtosecond Nd:fiber laser at 910 and 935 nm,” Opt. Lett. 39(15), 4404–4407 (2014). [CrossRef] [PubMed]
16. H. W. Chen, J. Lim, S. W. Huang, D. N. Schimpf, F. X. Kärtner, and G. Chang, “Optimization of femtosecond Yb-doped fiber amplifiers for high-quality pulse compression,” Opt. Express 20(27), 28672–28682 (2012). [CrossRef] [PubMed]
17. B. Chen, T. Jiang, Z. Zhang, and A. Wang, “4.4 nJ, 114 fs Nd-doped fiber laser pulses at 920nm for in vivo two-photon microscopic imaging”, Conference on Laser and Electro-Optics (CLEO,2015), paper AM1J.4. [CrossRef]