Open Access
Add to BibSonomy
Abstract
The spectral filtering effect is essential to dissipative dynamics in an all-normal-dispersion (ANDi) mode-locked fiber laser. In this study, we numerically and experimentally demonstrate the spectral filtering process of a nonlinear optical loop mirror (NOLM). Taking advantage of the 40/60 NOLM’s spectral filtering ability, we designed a novel all-polarization-maintaining ANDi mode-locked fiber laser without using a separate spectral filter. The NOLM functions as an artificial saturable absorber and a spectral filter in an ANDi cavity. During mode locking, we observed that the NOLM decreased the spectral width of the pulse from 5.46 to 4.38 nm. The fiber laser generated 509-fs compressed pulses at the repetition rate of 13.4 MHz. Our work provides a promising novel and compact ANDi fiber laser for ultrafast photonic applications.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Ultrafast mode-locked lasers are applicable in multiple areas such as microprocessing [1–4], biomedical applications [5], optical metrology [6], and nonlinear microscopy [7,8]. All-normal-dispersion (ANDi) mode-locked fiber lasers have attracted great attention in the past decade [9–11]. The combination effects of self-phase modulation (SPM) and normal dispersion in active fibers inside ANDi lasers support parabolic amplification, generating high-energy and broad-spectrum pulses. Recently, an ANDi Mamyshev oscillator with a large-mode-area fiber amplifier that delivers pulses with energy more than 1 µJ and compresses the pulse duration of 41 fs was demonstrated [12]. Moreover, dissipative solitons generated by ANDi lasers usually have a massive linear chirp and can tolerate large nonlinearity, which is highly suitable to seed a chirped-pulse amplifier [13–15].
Since no anomalous element is inserted into an ANDi laser cavity, the dissipative process relies on a spectral filter (SF) that allows it to reduce SPM-induced broad spectra to their original widths. Volume Bragg gratings with low insertion loss and high efficiency can realize efficient narrowband filtering [16]. Multiple-microring cavity is the type of SF that provides a high signal-to-noise ratio between resonant and out-of-band light, which can be used in optical communication systems [17,18]. Moreover, a Lyot filter includes two polarizers with a birefringent plate placed at the angle of 45°, permitting a cosine-square profile transmittance against the phase difference of the two orthogonal states in a polarization-maintaining (PM) fiber [19]. Similarly, another SF composed of two segments of PM fibers with different lengths and two polarization controllers was proposed [20]. An all-PM ANDi mode-locked configuration that utilizes a SESAM was reported, in which a tilted fiber Bragg grating functioned as a SF [21]. Nonlinear optical loop mirrors (NOLMs) and nonlinear amplifying loop mirrors have been widely used as saturable absorbers (SAs) since they have been reported to have intensity-dependent transmission [22–26]. However, the spectral filtering effect of a NOLM has never been explored.
In this study, the SF effect of a NOLM is numerically and experimentally demonstrated. The impacts of the average power of the input pulses, fiber length, and coupling ratio of the NOLM’s SF effect are studied. Taking advantage of the spectral filtering effect, an all-PM ANDi mode-locked fiber laser without a separate SF is designed for the first time and to the best of our knowledge.
2. Experimental setup and results
The experimental configuration for analyzing the NOLM’s spectral filtering effect is shown in Fig. 1(a). A home-made ANDi mode-locked oscillator is cascaded with a 5-m-long Yb-doped fiber amplifier. The pulses output from the amplifier have a 3-dB spectral width (SW) of 9.9 nm and a 3-dB pulse width of 33 ps (Gaussian pulse is assumed), which indicates that the pulses have great chirp. The spectrum and autocorrelation of the input pulses are illustrated in Figs. 1(b) and (c), respectively. A laser source delivers pulses at the repetition rate of 19.3 MHz with 48 mW maximum average power. A fiber attenuator is spliced into the laser source to adjust the power input to the NOLM. A 10/90 coupler spliced between the attenuator and the NOLM, extracting 10% of the input power is used to characterize input pulses of the NOLM. All the 2 × 2 filter couplers used in our experiment are coated using a thin film. Figure 1(d) shows the wavelength-dependent coupling ratio of a 2 × 2 40/60 coupler; it can be seen from the figure that the coupling ratio of the coupler is insensitive to wavelengths within the range of 1030–1060 nm and has a 3-dB bandwidth, larger than 50 nm. Moreover, the inset of Fig. 1(d) shows the spectra of the incident ASE source, passed port, and reflected port. All the fibers used in the experiment are PM fibers.
Fig. 1. (a) Schematic of the experimental configuration—LD: laser diode, WDM: wavelength-division multiplexer; (b) spectrum of the input pulses in logarithmic scale and linear scale (inset); (c) autocorrelation of the input pulses; and (d) wavelength coupling ratio of a 40/60 coupler, spectra of the injected ASE source, passed port, and reflected port (inset)
The average power of input pulses, loop length, and coupling ratio of the NOLM are the key factors that greatly impact its filtering effect. These three factors can be tuned independently in our experiment. First, the impact of input power is investigated by rotating the mechanical knob of the attenuator. A 10-m segment of fiber is spliced into the 40/60 coupler. The detailed spectra of the output pulses from the NOLM are plotted in Fig. 2(a) with a logarithmic scale and a linear scale. The SW of the output pulses from the NOLM as the function of input power is plotted in Fig. 2(b). To show the exact rule of the spectral filtering effect from the experimental results, 3-dB SW and 10-dB SW are measured simultaneously. It can be seen that both the 3-dB SW and 10-dB SW tend to decrease as the input power increases. The strongest spectral filtering effect occurs at the input power of 48 mW, corresponding to the spectral filtering ratio of 0.57, which is defined as the ratio of output SW to the input SW at 3-dB SW. The relationship of the NOLM’s transmission over input power is plotted in Fig. 2(c), which shows that the transmission of the NOLM does not attain saturation. The largest nonlinear phase difference (NPD) with the input power of 48 mW is calculated as 0.87 rad.
Fig. 2. (a) Output spectrum of different input power—48 mW in green, 32 mW in blue, 16 mW in magenta, and 8 mW in red; (b) output spectral width (SW) as the function of different input power—10-dB SW in magenta and 3 dB SW in red (the dashed line is for ease of observation); and (c) transmission as a function of input power (the dashed line is for ease of observation) for the 40/60 NOLM with the fiber length of 10 m
The detailed spectra of the output pulses from the NOLM with the loop length of 100 m for different input power are plotted in Fig. 3(a). The SWs of output pulses as the function of input power are depicted in Fig. 3(b), which shows that the SW does not monotonically decrease as the input power increases. The SW reaches the minimum at the input power of 23 mW. The NOLM’s transmission as a function of input power is shown in Fig. 3(c). Coincidentally, the curve has a maximum slope at the input power of 23 mW. A cubic polynomial (magenta dotted line) fits the experimental data perfectly, and the inflection point is derived from the polynomial form at the input power of 22.6 mW, which is consistent with the experimental results. The strongest spectral filtering effect occurs at the input power of 23 mW, which leads to the 3-dB spectral filtering ratio of 0.47.
Fig. 3. (a) Output spectra of different input power—48 mW (in green), 32 mW (in blue), 16 mW (in magenta), and 8 mW (in red); (b) output SW as the function of different input power—10-dB SW (in magenta) and 3-dB SW (in red, the dashed line is for ease of observation); and (c) transmission as the function of input power (the dashed line is for ease of observation) where the dotted line in magenta is the fitting curve
Second, the spectral filtering effect also depends on the loop length of the NOLM. Figure 4(a) shows the 3-dB and 10-dB SWs of the output spectra for different lengths of the NOLM with the input power of 16 and 32 mW, respectively. The 3-dB and 10-dB SWs both decrease as the length of the NOLM increases. Finally, the impact of the coupling ratio on the NOLM’s spectral filtering effect is investigated. For this investigation, the loop length of the NOLM is 10 m. Output SWs as the function of input power with the coupling ratios 40/60 (red) and 45/55 (magenta) are shown in Fig. 4(b). The 3-dB and 10-dB SWs of the output spectrum with the 40/60 NOLM is smaller than the 45/55 NOLM at the same input power, which validates that the spectral filtering ability of the NOLM with the 40/60 coupling ratio is stronger than that with the 45/55 coupling ratio.
Fig. 4. (a) Output SW as the function of different loop lengths with the coupling ratio of 40/60 and the input power of 16 and 32 mW; and (b) output SW as the function of input power for the 10-m NOLM with the coupling ratios of 40/60 and 45/55
The aforementioned experimental results show that the NOLM functions as the SF for positively chirped pulses. Using the spectral filtering effect of the NOLM, we design a compact laser without using a separate SF, the schematic of which is shown in Fig. 5. The laser can be nominally considered as consisting of two parts. The first part comprises a coupler with a 10% output ratio, an isolator to ensure single direction propagating, and an amplifier. The amplifier is composed of a wavelength-division multiplexer and a gain medium with a 1-m-long Yb3+-doped fiber (Nufern: PM-YSF-HI) pumped by a 976-nm single-mode laser diode (LD) of 600-mW maximum power. The second part is a NOLM formed by a 5-m-long passive fiber. The two parts are connected using a 2 × 2 coupler with the splitting ratio of 40/60. Another segment fiber in the cavity is the standard 9-m PM single-mode fiber (Nufern: PM 980). It is worth noting that our laser setup has no separate SF to reset the temporal and spectral width from one round-trip to another inside the cavity, which is different from traditional ANDi mode-locking lasers. Each element used in the cavity has a SW much larger than 30 nm, centering at 1040 nm, which means no spectral filtering occurs owing to the limited bandwidth. A 1 × 2 coupler with 1% output was inserted at the NOLM’s input and transmitted port in sequence to measure the SW. Note that the mode-locked state was not changed when the coupler was inserted; however, the output power decreased slightly.
The mode-locked state is self-started when the pump power increases to 500 mW. However, at this high pump power, the mode-locked state was noise-like. By slowly reducing the pump power to 280 mW, the mode-locked state became stable. The spectra of these two output ports are plotted in Fig. 6(a) using linear and logarithmic scales. The SWs of pulses are decreased from 5.46 to 4.38 nm, as is predicted by the aforementioned study of spectral filtering effect. Furthermore, we depict the temporal profile of these two ports in Fig. 6(b), which is measured using a fast photodiode (Alphalas, UPD-15-IR2-FC, risetime < 15 ps) in conjunction with a broadband oscilloscope (Agilent 86100A, 20GHz). The pulse durations of the NOLM pulse input and output are 80.3 ps (solid black) and 61.2 ps (dotted blue), respectively, which indicates a temporal narrowing similar to the spectral filtering occurring in the NOLM.
Fig. 6. (a) spectral and (b) temporal profiles before (solid black) and after (dotted blue) the NOLM; (c) output pulse train; and (d) autocorrelation trace of the output pulse compressed by a pair of gratings
The pulse train is shown in Fig. 6(c), and its repetition rate is 13.4 MHz. The output average power of 10.4 mW corresponds to the energy of 0.78 nJ. A pair of 1000 lines/mm gratings was used to compress the pulses. The compressed pulse duration is measured to be 509 fs using an APE autocorrelator when the Gaussian profile is assumed, as shown in Fig. 6(d). The pulse is 1.43 times its transform limit due to the uncompressed nonlinear chirp that was induced by the long fiber and gratings.
3. Numerical simulations
The spectral filtering effect of a NOLM is studied in this section by numerical methods. Numerical simulations for a passive fiber in the NOLM are based on the commonly used generalized nonlinear Schrödinger equation, which is solved by the split-step Fourier method. Our model is simple and includes only second-order dispersion and SPM:
Modeling of the 2 × 2 coupler can be described using the equation below:
Ac and Acc refer to the transmitted field of the input pulses that propagate clockwise and counterclockwise, respectively. r is the splitting ratio of the NOLM. The input pulses take the form of a chirped parabola for convenience due to a linear chirp, which is defined as follows:The input pulses have the 3-dB pulse width of 35 ps and the 3-dB SW of 5.7 nm. The coupling ratio of the NOLM is 40/60. Figure 7 shows the temporal and spectral profiles transmitting from the NOLM. Figures 7(a) and (b) show the output pulse shape and spectra changing with different loop-fiber lengths for the input pulse energy of 3 nJ. The NOLM with a 10-m-long loop fiber has a stronger filtering effect than the one with a 5-m-long loop fiber. The NPD of the two coupling ports for the NOLM with the 10-m-long loop fiber is 1.1 rad. This trend closely corresponds with our experimental results at such a low NPD. When the loop length increases, both the temporal regime and spectrum start to split, and energy is transferred to the side lobe of the pulses. However, for the NOLM with a 100-m-long loop fiber, the spectral filtering on the central peak can be observed. The NPDs between the two coupling ports are 5.6 and 8.4 rad for the NOLM with 60-m- and 100-m-long loop fibers, respectively. Figures 7(c) and (d) illustrate the output pulse shape and spectra changing with different input pulse energy for the given loop length of 5 m. These results are similar to the loop length varying case for low input energy, and both the pulse duration and SW reduce as the pulse energy increases, while new peaks rise at high input energy of 30 nJ. The NPDs for 1, 2, 35, and 40 nJ are 0.2, 0.4, 5.5, and 8.3 rad, respectively. Moreover, the spectra of the magenta and yellow curves in Figs. 7(b) and (d) are similar, which indicates that the NOLM’s spectral filtering effects may be the same at the same NPD.
Fig. 7. Simulation results for the spectral filtering effect of the NOLM—(a) temporal and (b) spectral profiles of the NOLM output pulses for different nonlinear loop lengths with the input energy of 3 nJ; (c) temporal and (d) spectral profile of the NOLM for different input energy with the nonlinear loop length of 5 m; the coupling ratio of the NOLM is 40/60
4. Conclusions
In conclusion, we numerically and experimentally demonstrated the spectral filtering effect of a NOLM. Impacts of three parameters, including different input pulse power, loop length, and coupling ratio, on the NOLM’s spectral filtering ability were investigated. For a NOLM with 10-m loop length, the higher is the input power or the longer is the loop length, the stronger is the NOLM’s spectral filtering ability. Besides, the spectral filtering ability of the NOLM with the coupling ratio of 40/60 is better than the one with the coupling ratio of 45/55. However, for the NOLM with 100-m loop length, its spectral filtering ability is much more complex, and the minimum spectral filtering ratio of 0.47 is obtained. Based on the experimentally verified spectral filtering effect of the 40/60 NOLM, we designed an all-PM ANDi mode-locked laser scheme with the NOLM but without a separate SF. The simulation results showed that the NOLM not only functioned as a SA but also served as a SF because both the SW and pulse width were reduced by the NOLM. The laser delivers pulses at the repetition rate of 13.4 MHz and compresses the pulse duration of 509 fs. We also revealed the spectral filtering ability of the NOLM by numerical methods. We believe our compact all-PM ANDi mode-locked fiber laser has potential applications in ultrafast photonics.
Funding
South China Normal University (19KJ13); Young Innovative Talents Project in Universities of Guangdong Province (2018KQNCX057); Key-Area Research and Development Program of Guangdong Province (2020B090922006); Guangdong Science and Technology Department (2018B030323017); National Key Research and Development Program of China (2017YFB1104500).
Acknowledgments
We would like to thank professor Zhigang Zhang for the instruction on our work.
Disclosures
The authors declare no conflicts of interest.
References
1. C. Kerse, H. Kalaycıoğlu, P. Elahi, B. Çetin, D. K. Kesim, O. Akçaalan, S. Yavaş, M. D. Aşık, B. Öktem, H. Hoogland, R. Holzwarth, and F. O. Ilday, “Ablation-cooled material removal with ultrafast bursts of pulses,” Nature 537(7618), 84–88 (2016). [CrossRef]
2. M. Erdogan, B. Oktem, H. Kalaycoglu, S. Yavas, P. K. Mukhopadhyay, K. Eken, K. Ozgoren, Y. Aykac, U. H. Tazebay, and F. O. Ilday, “Texturing of Titanium medical implant surfaces with MHz repetition rate femtosecond and picosecond fiber lasers,” Opt. Express 19(11), 10986–10996 (2011). [CrossRef]
3. P. Zhou, C. Liao, Z. Li, S. Liu, and Y. Wang, “In-Fiber Cascaded FPI Fabricated by Chemical-Assisted Femtosecond Laser Micromachining for Micro-Fluidic Sensing Applications,” J. Lightwave Technol. 37(13), 3214–3221 (2019). [CrossRef]
4. D. Yu, R. Huang, and W. H. Knox, “Femtosecond laser micromachining in ophthalmic hydrogels: spectroscopic study of materials effects,” Opt. Mater. Express 9(8), 3292–3305 (2019). [CrossRef]
5. C. Kong, C. Pilger, H. Hachmeister, X. Wei, T. H. Cheung, C. S. W. Lai, T. Huser, K. K. Tsia, and K. K. Y. Wong, “Compact fs ytterbium fiber laser at 1010 nm for biomedical applications,” Biomed. Opt. Express 8(11), 4921–4932 (2017). [CrossRef]
6. J. L. Hall, J. Ye, S. A. Diddams, L.-S. Ma, S. T. Cundiff, and D. J. Jones, “Ultrasensitive spectroscopy, the ultrastable lasers, the ultrafast lasers, and the seriously nonlinear fiber: a new alliance for physics and metrology,” J,” Quantum Electron. 37(12), 1482–1492 (2001). [CrossRef]
7. A. Zach, M. Mohseni, C. Polzer, J. W. Nicholson, and T. Hellerer, “All-fiber widely tunable ultrafast laser source for multimodal imaging in nonlinear microscopy,” Opt. Lett. 44(21), 5218–5221 (2019). [CrossRef]
8. D. A. Plemmonsa, S. T. Parkb, A. H. Zewailb, and D. J. Flannigana, “Characterization of fast photoelectron packets in weak and strong laser fields in ultrafast electron microscopy,” Ultramicroscopy 146, 97–102 (2014). [CrossRef]
9. A. Chong, J. Buckley, W. Renninger, and F. Wise, “All-normal dispersion femtosecond fiber laser,” Opt. Express 14(21), 10095–10100 (2006). [CrossRef]
10. M. Tokurakawa, H. Sagara, and H. Tünnermann, “All-normal-dispersion nonlinear polarization rotation mode-locked Tm: ZBLAN fiber laser,” Opt. Express 27(14), 19530–19535 (2019). [CrossRef]
11. S. D. Chowdhury, A. Pal, S. Chatterjee, R. Sen, and M. Pal, “Multipulse Dynamics of Dissipative Soliton Resonance in an All-Normal Dispersion Mode-Locked Fiber Laser,” J. Lightwave Technol. 36(24), 5773–5779 (2018). [CrossRef]
12. W. Liu, R. Liao, J. Zhao, J. Cui, Y. Song, C. Wang, and M. Hu, “Femtosecond Mamyshev oscillator with 10-MW-level peak power,” Optica 6(2), 194–197 (2019). [CrossRef]
13. D. N. Schimpf, J. Limpert, and A. Tünnermann, “Controlling the influence of SPM in fiber-based chirped-pulse amplification systems by using an actively shaped parabolic spectrum,” Opt. Express 15(25), 16945–16953 (2007). [CrossRef]
14. H. Chen, J. Lim, S. 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]
15. H. Chang, Z. Cheng, R. Sun, Z. Peng, M. Yu, Y. You, M. Wang, and P. Wang, “172-fs, 27-µJ, Yb-doped all-fiber-integrated chirped pulse amplification system based on parabolic evolution by passive spectral amplitude shaping,” Opt. Express 27(23), 34103–34112 (2019). [CrossRef]
16. D. Ott, M. SeGall, I. Divliansky, G. Venus, and L. Glebov, “High-contrast filtering by multipass diffraction between paired volume Bragg gratings,” Appl. Opt. 54(31), 9065–9070 (2015). [CrossRef]
17. B. E. Little, S. T. Chu, H. A. Haus, J. Foresi, and J. Laine, “Microring resonator channel dropping filters,” J. Lightwave Technol. 15(6), 998–1005 (1997). [CrossRef]
18. Y. Akahane, T. Asano, H. Takano, B. Song, Y. Takana, and S. Noda, “Two-dimensional photonic-crystal-slab channel-drop filter with flat-top response,” Opt. Express 13(7), 2512–2530 (2005). [CrossRef]
19. K. Özgören and FÖ İlday, “All-fiber all-normal dispersion laser with a fiber-based Lyot filter,” Opt. Lett. 35(8), 1296–1298 (2010). [CrossRef]
20. Y. S. Fedotov, A. V. Ivanenko, S. M. Kobtsev, and S. V. Smirnov, “High average power mode-locked figure-eight Yb fibre master oscillator,” Opt. Express 22(25), 31379–31386 (2014). [CrossRef]
21. J. Lecourt, C. Duterte, F. Narbonneau, D. Kinet, Y. Hernandez, and D. Giannone, “All-normal dispersion, all-fibered PM laser mode-locked by SESAM,” Opt. Express 20(11), 11918–11923 (2012). [CrossRef]
22. N. J. Doran and D. Wood, “Nonlinear-optical loop mirror,” Opt. Lett. 13(1), 56–58 (1988). [CrossRef]
23. J. Shi, Y. Li, S. Gao, Y. Pan, G. Wang, R. Ji, and W. Zhou, “All-polarization-maintaining figure-eight Er-fiber ultrafast laser with a bidirectional output coupler in the loss-imbalanced nonlinear optical loop mirror,” Chin. Opt. Lett. 16(12), 121404 (2018). [CrossRef]
24. C. Aguergaray, N. G. R. Broderick, M. Erkintalo, J. S. Y. Chen, and V. Kruglov, “Mode-locked femtosecond all-normal all-PM Yb-doped fiber laser using a nonlinear amplifying loop mirror,” Opt. Express 20(10), 10545–10551 (2012). [CrossRef]
25. Y. Yu, H. Teng, H. Wang, L. Wang, J. Zhu, S. Fang, G. Chang, J. Wang, and Z. Wei, “Highly-stable mode-locked PM Yb-fiber laser with 10 nJ in 93-fs at 6 MHz using NALM,” Opt. Express 26(8), 10428–10434 (2018). [CrossRef]
26. M. Erkintalo, C. Aguergaray, A. Runge, and N. G. R. Broderick, “Environmentally stable all-PM all-fiber giant chirp oscillator,” Opt. Express 20(20), 22669–22674 (2012). [CrossRef]
