Open Access
Add to BibSonomy
Abstract
The 1600-1700-nm ultrafast fiber lasers attract great interests in the deep multiphoton microscopy, due to the reduced levels of the tissue scattering and absorption. Here, we report on the 86.7-MHz, 717-mW, 91.2-fs, all-fiber laser located in the spectral range from 1600 nm to 1700nm. The soliton self-frequency shift (SSFS) was introduced into the Er:Yb co-doped fiber amplifier (EYDFA) to generate the high-power, 1600-1700-nm Raman soliton. Detailed investigations of the nonlinear fiber amplification process were implemented in optimizing the generated Raman soliton pulses. The miniature multiphoton microscopy was further realized with this home-built laser source. The clearly imaging results can be achieved by collecting the generated harmonic signals from the mouse tail skin tissue with a penetration depth of ∼500 µm. The experimental results indicate the great potential in utilizing this 1600-1700-nm fiber laser in the deep multiphoton microscopy.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
In recent years, the multiphoton microscopy (MPM) has attracted wide interests in the field of biomedical science, based on the high imaging resolution, high signal-to-background ratio (SBR), and three-dimensional imaging capabilities [1,2]. Utilizing the 1600-1700-nm femtosecond fiber lasers in the MPM was proved to be an efficient approach in realizing the multiphoton microscopy with deep penetration depth [3–5]. The reduced tissue scattering and water absorption in this wavelength region can significantly improve the SBR of the generated imaging signals [6]. Recently, multiple 1600-1700-nm femtosecond fiber lasers were realized by introducing the SSFS into the additional nonlinear propagation stage with the 1560-nm front-end laser [7–11]. K. Wang et al. reported the generation of the 15-nJ, 69.4-fs, 1700-nm Raman soliton by coupling the 1550-nm femtosecond laser into the large-mode-area (Thorlabs, LMA35) fiber. This laser was utilized in the third-harmonic generation (THG) imaging system of the mouse brain in vivo with 1-mm penetration depth [4]. X. He et al. realized the ∼35-mW, 384-fs, 102-MHz, 1700-nm fiber laser, utilizing the combination of the Er-doped fiber laser and the standard telecom fiber [12]. H. Liu et al. reported the 1600-1840-nm femtosecond solitons, based on the large-mode-area fiber (NKT Photonics, LMA-PM-35) pumped by a 1550-nm laser. This laser source was utilized to illustrate the wavelength dependent 3-photon action cross-sections of the red fluorescent proteins [13]. Further investigations in generating the high-power, 1600-1700-nm femtosecond fiber lasers without the additional nonlinear stage are still necessary in realizing the MPM laser sources with higher compactness and less costs [14,15].
Generating the frequency-shifted Raman soliton during the 1560-nm femtosecond fiber laser amplification process can be an effective approach [14,15]. The generated broadband Raman spectral contents of the amplified 1560-nm signal pulses can be transferred to even longer wavelength region, based on the in-band Raman gain process and the conjugation between the negative group velocity dispersion (GVD) of the gain fiber and the accumulated nonlinear phase [16,17]. Therefore, the high-power, all-fiber, 1600-1700-nm femtosecond lasers can be realized by carefully managing the nonlinear evolution process of the amplified 1560-nm signal lasers. However, such kinds of all-fiber, 1600-1700-nm fiber laser sources were rarely reported.
In this letter, we report on a high-power, sub-100-fs, all-fiber, 1600-1700-nm femtosecond laser, generated by introducing the SSFS into the EYDFA. The 86.7-MHz, 1573-nm mode-locking signal laser was delivered by the home-built polarization-maintaining (PM) fiber oscillator. The dispersion and the accumulated nonlinear phase of the all-fiber laser system were carefully managed in directly generating the 86.7-MHz, 717-mW, 91.2-fs, 1600-1700-nm Raman soliton pulses inside the EYDFA. Detailed investigations of the fiber laser system were also implemented in optimizing the generated Raman soliton pulses. The miniature multiphoton microscopy was further built with this compact 1600-1700-nm fiber laser. The clearly imaging results can be achieved by collecting the generated harmonic signals from the mouse tail skin tissue with a penetration depth of ${\sim} $500 µm. The experimental results indicate the great potential in utilizing this high-power, sub-100-fs, 1600-1700-nm fiber laser in the deep multiphoton microscopy.
2. Laser setup and generation of Raman solitons
The schematic design of the 1600-1700-nm fiber laser system is shown in Fig. 1. This fiber laser system consists of the all-PM fiber seed laser, the fiber stretcher, the all-PM fiber pre-amplifier, and the non-PM power amplifier. The seed laser is a home-made, Er-doped mode-locking fiber laser based on the soliton mode-locking scheme. The all-PM-fiber nonlinear amplified loop mirror (NALM) was utilized in realizing the mode-locking state. The 86.7-MHz, 2.8-mW, 1573-nm signal pulses can be delivered by this fiber oscillator with a net-cavity dispersion of -28228.5 fs2. The pulse duration of the mode-locking signal pulse was 0.7 ps based on the Gaussian assumption. The optical spectrum is illustrated as the black curve in Fig. 2(a). The corresponding spectral bandwidth was 13 nm. The signal pulses were stretched by the 6-m dispersion compensation fiber (DCF), after propagating through the all-PM fiber isolator with the 0.5-m PM1550 fiber pigtails. The corresponding group velocity dispersion (GVD) provided by the DCF and the PM1550 fiber is 36.98 fs2/mm and −22.96 fs2/mm, respectively. Therefore, the nonlinear chirped-pulse amplification configuration can be realized inside the first-stage fiber amplifier [18,19]. The first-stage fiber amplifier consists of the 3.6-m Er-doped gain fiber (DHB1500, Fibercore), two all-PM fiber wavelength division multiplexers (WDMs) with 0.5-m PM1550 fiber pigtails, and two 976-nm single mode diode lasers. The bidirectionally core-pumping scheme was utilized in scaling up the average power from 2 mW to 200 mW. The GVD provided by the gain fiber is 30.61 fs2/mm. Therefore, the self-phase modulation (SPM), the positive GVD, and the flat optical gain contributed to the signal pulse evolution during the first-stage fiber amplification process. The pulse duration of the positively chirped signal pulse was further stretched by the normal dispersion of the gain fiber. The amplified optical spectrum can be slightly broadened by the SPM effect [20]. There was the fiber isolator with 2-m PM1550 fiber pigtails spliced after the fiber WDM to isolate the potential back reflection. Figure 2 illustrates the optical spectrum (the red curve) and the corresponding auto-correlation trace measured at the input port of the fiber combiner. The dual peak construction of the spectrum was caused by the conjugation between the SPM and the negative GVD inside the PM1550 fiber. The corresponding spectral bandwidth was 20 nm. The pulse duration of the signal pulse was compressed to 1.48 ps by the PM1550 fiber, based on the Gaussian assumption.
Fig. 1. The schematic design of the high-power, all-fiber, 1600-1700-nm femtosecond laser system. DCF, dispersion compensation fiber; WDM, wavelength division multiplexer; EDF, Erbium-doped fiber; ISO, isolator; PC, polarization controller; EYDF, Erbium-Ytterbium co-doped fiber; LPF, long pass filter
Fig. 2. (a) Pulse spectrum before being injected into the power amplifier; (b) pulse auto-correlation trace before being injected into the power amplifier;
The 86.7-MHz, 200-mW, pre-amplified signal pulses were coupled into the 2.6-m Er:Yb co-doped double-cladding fiber (2CF-EY-O-17-130-L2, iXblue) by the non-PM fiber combiner. The input signal fiber of the fiber combiner was 0.2-m SMF-28e single mode fiber. There was the in-line fiber polarization controller (CPC250, Thorlabs) utilized with the non-PM signal fiber to control the initial polarization of the 1573-nm signal pulses. The corresponding GVD of the SMF-28e and the 0.2-m passive output fiber of the fiber combiner were −22 fs2/mm and −27.1 fs2/mm, respectively. The dual peak construction of the positively chirped signal pulse can be enhanced at the beginning part of the power amplifier, due to the conjugation between the negative GVD of the gain fiber and the SPM effect. The positively chirped pulse duration was kept being compressed by the negative GVD of the gain fiber, enhancing the SPM effect inside the fiber power amplifier. The longer wavelength tail generated by the SPM effect can be amplified by the optical Raman gain provided by the stimulated Raman scattering (SRS) effect during the nonlinear amplification process. The amplified Raman spectral contents can evolve into the Raman soliton inside the gain fiber with negative GVD. Further, the longer wavelength tail of the Raman soliton can experience the Raman gain at the expense of the energy of the shorter wavelength contents during the nonlinear amplification process, leading to the Raman soliton self-frequency shift [21]. As is well-known, the pulse energy of the fundamental soliton is proportional to the effective mode area of the fiber [21,22]. The commercialized EYDF with a large core diameter of 17 µm was utilized to generate the Raman soliton with high pulse energy. No more delivery fiber was spliced after the Er:Yb co-doped gain fiber. The output port of the gain fiber was cleaved with an 8-degree angle to inhibit the surface reflection. An aspheric lens with 4.5-mm focal length was used to collimate the output beam. The gain fiber was coiled with a metal cylindrical heat sink with a diameter of 8 cm. The long-wavelength-pass filter (LPF, #84672, Edmund Optics) with the cut-off wavelength of 1600 nm was utilized to remove the residual 1.56-µm spectral contents. Therefore, the clean soliton pulses with the central wavelength located between 1600 nm and 1700nm can be obtained for further investigations.
The central wavelength of the shifted Raman soliton can be simply modified by managing the injected pump power. Figure 3 illustrates the measured spectral evolution of the self-frequency shifted Raman soliton with the increased pump powers and the corresponding auto-correlation traces, under the optimized initial polarization state of P2 shown in Fig. 4(a) and (b). The optimized initial polarization of the injected 1573-nm signal pulses was achieved by realizing the generated 1600-1700-nm Raman soliton with the highest average power. The faster nonlinear spectral broadening process can be introduced by the amplified 1573-nm signal pulses with higher pump power at the beginning part of the fiber power amplifier. The corresponding Raman spectral contents can be generated earlier during the power amplification process, leading to the longer Raman scattering interaction distance inside the fiber power amplifier. Therefore, the generated Raman soliton can be shifted to longer wavelength range. The in-band pump scheme was also applied to transfer the optical power from the shorter wavelength contents to the generated Raman spectral contents, before the Raman soliton was shifted outside the optical gain wavelength range of the EYDF. Therefore, as is shown in the red curve of the Fig. 4(b), the highest average power of the generated 86.7-MHz Raman soliton pulses was 717 mW after increasing the pump power to 12 W, under the initial polarization state of P2. The corresponding central wavelength of the Raman soliton pulse was 1640 nm. Due to the inelastic scattering energy loss [23], the average power of the further shifted Raman soliton pulses experienced a slight decrease with the further increased pump power. The optical spectrum of the 717-mW, 86.7-MHz Raman soliton pulses was illustrated in Fig. 4(c). The corresponding pulse duration of the measured auto-correlation trace was 91.2 × 1.414 fs based on the Gaussian assumption, shown in Fig. 4(d).
Fig. 3. The process of (a) the soliton self-frequency shifting with the increased pump power. The black curves represent the output spectra without the LPF. The red curves represent the optical spectra of the Raman solitons filtered out by the LPF. (b) The corresponding auto-correlation traces of the generated Raman solitons.
Fig. 4. (a) The optical wavelength variations as the functions of the pump power with the initial polarization states of P1 and P2. (b) The average power of the generated Raman solitons as the functions of the pump power with the initial polarization states of P1 and P2. (c) The optical spectrum of the amplified signal laser with 12-W pump power (the red curve), and the corresponding optical spectrum of the 717-mW Raman soliton filtered out by the LPF. (d) The measured auto-correlation trace of the 1640-nm, 717-mW Raman soliton.
Figure 4(a) illustrates the central wavelength variations of the generated Raman solitons with different initial polarization states as a function of the increased the pump power. Optimizing the polarization states of the 1573-nm signal pulses can influence the Raman frequency conversion process [24,25]. The accumulated nonlinear phase and the integrated optical dispersion can be well adjusted by managing the initial polarization of the signal pulses, based on the optical birefringence of the non-PM double cladding fiber. The red curve in Fig. 4 plots the optimum results in achieving the highest average power of the Raman soliton, after carefully optimizing the initial polarization. Further adjusting the initial polarization states led to the lower nonlinear frequency conversion efficiency, shown as the black curve in Fig. 4. The output powers as the function of the increased pump power were illustrated in Fig. 4(b). The output average power stability is illustrated in Fig. 5. The measured root mean squared error of the output average power at 670 mW is less than 1.7% within 40 minutes, demonstrating the laser operated with good stability.
3. MPM imaging with the generated Raman soliton pulses
Multiphoton microscopy is especially advantageous in the imaging of living cells and other objects [26]. The 1600-1700-nm femtosecond pulses were further routed into a miniature multiphoton microscope probe-head to verify its imaging capability. The imaging setup is depicted in Fig. 6. The collimated beam was coupled into the probe head by the 1.5-m anti-resonant fiber (ARF), after propagating through the LPF. The core-diameter of the utilized ARF is 50 $\mathrm{\mu}\textrm{m}$ with the transmission loss of 50 dB/km at 1550 nm. There was no significant optical chirp introduced by this 1.5-m ARF into the femtosecond pulses. Therefore, no additional dispersive elements were applied to further manage the optical dispersion. The combination of the wave-plates and the PBS illustrated in Fig. 6 was utilized to manage the injected optical power of the 1600-1700-nm femtosecond pulses.
Fig. 6. The schematic construction of the home-built miniature multiphoton microscopy. HWP, half waveplate; QWP, quarter waveplate; PBS, polarization beam splitter; MEMS, microelectromechanical system; SL, scan lens; DM, dichroic mirror; PMT, photomultiplier tube.
The probe-head consists of a MEMS scanner, a scan lens, a dichroic mirror, an objective and a collecting lens, shown in Fig. 6. The operation frequency of the MEMS was set to be 2400 Hz, resulting a frame rate of 9 fps with a frame size of 180 µm × 160 µm. The excitation light was focused by the 9x microscopic objective (N.A. = 0.9) with an average power of ∼180 mW. The generated harmonic signals were collected by the objective, filtered by the dichroic mirror and delivered to a photomultiplier tubes (PMTs) by a silica fiber. A bandpass filter (FF01-819/44, Semrock) was put in front of the PMT in the second harmonic generation (SHG) channel. Likewise, another bandpass filter (FF01-550/49, Semrock) was put in the third harmonic generation (THG) channel. The mouse tail skin tissue can generate intense SHG signals since it contains much muscle. The imaging results of the mouse tail skin tissue with the generated Raman soliton pulses are shown in Fig. 7. Myofibers and thick bundles of collagens in the dermis layer can be clearly visualized with the SHG signals. Loose connective tissue and adipocytes with associated blood vessels can be resolved with the THG signals. The corresponding penetration depth was ∼500 µm under the skin surface [8,27]. The imaging process was stain-free, indicating this home-built system can be utilized in the multiphoton microscopy with deep penetration depth. Further multiphoton excited fluorescence imaging experiments can also be implemented after decreasing the repetition rate of this Raman soliton laser to ∼1 MHz [3,4,5].
Fig. 7. The measured MPM images of the mouse tail skin, utilizing the 1600-1700-nm femtosecond pulses as the excitation source. Red: SHG signals. Green: THG signals. Frame size: 180 µm × 160 µm. Scale bar is 50 µm.
4. Conclusion
We have demonstrated the high-power, all-fiber, 1600-1700-nm femtosecond laser by introducing the soliton self-frequency shift into the EYDFA. The 86.7-MHz, 717-mW, 91.2-fs, 1600-1700-nm Raman soliton pulses can be delivered. Detailed investigations of the utilized nonlinear fiber amplifier were also implemented in optimizing the generated Raman soliton pulses. The generated Raman soliton pulses were further utilized as the excitation source for the home-built miniature multiphoton microscopy. The imaging results of the mouse skin tissue were achieved with the penetration depth of ${\sim} $500 µm. The myofibers and thick bundles of collagens in the dermis layer can be clearly visualized with the SHG signals. Loose connective tissue and adipocytes with associated blood vessels can be resolved with the THG signals. The achieved stain-free imaging results indicate the great potential in utilizing this 1600-1700-nm Raman soliton pulses in the deep multiphoton microscopy.
Funding
National Key Research and Development Program of China (2020YFB1312802); Natural Science Foundation of Shandong Province (ZR2021QF054); Chinese Academy of Science Pioneer Hundred Talents Program (E1Z1D10900).
Disclosures
The authors declare no conflicts of interest.
Data availability
All data are available in the main text and from the corresponding author upon reasonable request.
References
1. W. R. Zipfel, R. M. Williams, and W. W. Webb, “Nonlinear magic: multiphoton microscopy in the biosciences,” Nat. Biotechnol. 21(11), 1369–1377 (2003). [CrossRef]
2. K. König, “Multiphoton microscopy in life sciences,” J. Microsc. 200(2), 83–104 (2000). [CrossRef]
3. N. G. Horton, K. Wang, D. Kobat, C. G. Clark, F. W. Wise, C. B. Schaffer, and C. Xu, “In vivo three-photon microscopy of subcortical structures within an intact mouse brain,” Nat. Photonics 7(3), 205–209 (2013). [CrossRef]
4. W. Ke, N. G. Horton, K. Charan, and C. Xu, “Advanced Fiber Soliton Sources for Nonlinear Deep Tissue Imaging in Biophotonics,” IEEE J. Sel. Top. Quantum Electron. 20(2), 50–60 (2014). [CrossRef]
5. L.-C. Cheng, N. G. Horton, K. Wang, S.-J. Chen, and C. Xu, “Measurements of multiphoton action cross sections for multiphoton microscopy,” Biomed. Opt. Express 5(10), 3427–3433 (2014). [CrossRef]
6. K. Wang, R. Liang, and P. Qiu, “Fluorescence Signal Generation Optimization by Optimal Filling of the High Numerical Aperture Objective Lens for High-Order Deep-Tissue Multiphoton Fluorescence Microscopy,” IEEE Photonics J. 7(6), 1–8 (2015). [CrossRef]
7. K. Wang, Y. Wei, D. Kobat, and C. Xu, “Tunable high-energy soliton pulse generation from a large-mode-area fiber pumped by a picosecond time-lens source,” in CLEO: 2011 - Laser Science to Photonic Applications (2011), pp. 1–2.
8. K. Wang, W. Wen, Y. Wang, K. Wang, J. He, J. Wang, P. Zhai, Y. Yang, and P. Qiu, “Order-of-magnitude multiphoton signal enhancement based on characterization of absorption spectra of immersion oils at the 1700-nm window,” Opt. Express 25(6), 5909–5916 (2017). [CrossRef]
9. B. Li, C. Wu, M. Wang, K. Charan, and C. Xu, “An adaptive excitation source for high-speed multiphoton microscopy,” Nat. Methods 17(2), 163–166 (2020). [CrossRef]
10. T. N. Nguyen, K. Kieu, D. Churin, T. Ota, M. Miyawaki, and N. Peyghambarian, “High Power Soliton Self-Frequency Shift With Improved Flatness Ranging From 1.6 to 1.78 $\mathrm {\mu} \textrm{m}$,” IEEE Photonics Technol. Lett. 25(19), 1893–1896 (2013). [CrossRef]
11. H. Cheng, S. Tong, X. Deng, H. Liu, Y. Du, C. He, P. Qiu, and K. Wang, “Deep-brain 2-photon fluorescence microscopy in vivo excited at the 1700 nm window,” Opt. Lett. 44(17), 4432–4435 (2019). [CrossRef]
12. X. He, Q. Lin, H. Guo, J. Sun, J. Bai, L. Hou, and K. Wang, “Robust 1.7-µm, all-polarization-maintaining femtosecond fiber laser source based on standard telecom fibers,” Appl. Phys. Express 12(7), 072007 (2019). [CrossRef]
13. H. Liu, J. Wang, X. Peng, Z. Zhuang, P. Qiu, and K. Wang, “Ex and in vivo characterization of the wavelength-dependent 3-photon action cross-sections of red fluorescent proteins covering the 1700-nm window,” J. Biophotonics 11(7), e201700351 (2018). [CrossRef]
14. N. Nishizawa and T. Goto, “Compact system of wavelength-tunable femtosecond soliton pulse generation using optical fibers,” IEEE Photonics Technol. Lett. 11(3), 325–327 (1999). [CrossRef]
15. K. Wang, N. Horton, C. C. Wang, D. Ouzounov, and C. Xu, “Tunable megawatt soliton pulse generation covering the optimum wavelength window for tissue penetration,” in CLEO: 2013 (2013), pp. 1–2.
16. M. N. Islam, G. Sucha, I. Bar-Joseph, M. Wegener, J. P. Gordon, and D. S. Chemla, “Femtosecond distributed soliton spectrum in fibers,” J. Opt. Soc. Am. B 6(6), 1149–1158 (1989). [CrossRef]
17. J. H. Lee, J. van Howe, C. Xu, and X. Liu, “Soliton Self-Frequency Shift: Experimental Demonstrations and Applications,” IEEE J. Sel. Top. Quantum Electron. 14(3), 713–723 (2008). [CrossRef]
18. L. Kuznetsova, A. Chong, and F. W. Wise, “Interplay of nonlinearity and gain shaping in femtosecond fiber amplifiers,” Opt. Lett. 31(17), 2640–2642 (2006). [CrossRef]
19. H. Yu, P. Zhang, X. Wang, P. Zhou, and J. Chen, “High-Average-Power Polarization Maintaining All-Fiber-Integrated Nonlinear Chirped Pulse Amplification System Delivering Sub-400 fs Pulses,” IEEE Photonics J. 8(2), 1–7 (2016). [CrossRef]
20. Y. Qin, O. Batjargal, B. Cromey, and K. Kieu, “All-fiber high-power 1700nm femtosecond laser based on optical parametric chirped-pulse amplification,” Opt. Express 28(2), 2317–2325 (2020). [CrossRef]
21. F. M. Mitschke and L. F. Mollenauer, “Discovery of the soliton self-frequency shift,” Opt. Lett. 11(10), 659–661 (1986). [CrossRef]
22. J. P. Gordon, “Theory of the soliton self-frequency shift,” Opt. Lett. 11(10), 662–664 (1986). [CrossRef]
23. A. A. Voronin and A. M. Zheltikov, “Soliton self-frequency shift decelerated by self-steepening,” Opt. Lett. 33(15), 1723–1725 (2008). [CrossRef]
24. K. Masao, “Wavelength-tunable multicolor Raman soliton generation using an ellipse polarized pump pulse and highly birefringent optical fibers,” J. Lightwave Technol. 24(2), 805–809 (2006). [CrossRef]
25. N. Nishizawa, R. Okamura, and T. Goto, “Simultaneous generation of wavelength tunable two-colored femtosecond soliton pulses using optical fibers,” IEEE Photonics Technol. Lett. 11(4), 421–423 (1999). [CrossRef]
26. W. Denk, J. H. Strickler, and W. W. Webb, “2-PHOTON LASER SCANNING FLUORESCENCE MICROSCOPY,” Science 248(4951), 73–76 (1990). [CrossRef]
27. Y. Wang, K. Wang, W. Wen, P. Qiu, and K. Wang, “Comparison of Signal Detection of GaAsP and GaAs PMTs for Multiphoton Microscopy at the 1700-nm window,” IEEE Photonics J. 8(6), 1–8 (2016). [CrossRef]
