Raman-scattering-assisted noise-like pulse (NLP) generation was achieved by using an appropriate segment of high nonlinearity fiber in an erbium-doped fiber laser. Broadband spectrum with 203 nm 3-dB bandwidth was obtained, which, to the best of our knowledge, is the broadest bandwidth achieved for NLPs. The broadband operation is the result of tailored cavity design, which optimizes various effects including the Raman scattering effect to maximize the bandwidth of NLPs. Further broadening the NLP spectrum up to 294 nm was achieved by using spectral filtering outside the cavity with a polarization beam splitter.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Ultrashort-pulsed fiber lasers are attractive alternative for all-solid-state lasers, because of their potential for alignment and maintaining free operation, as well as for their good beam quality and various means of improving pulsed performance. Fiber lasers are also a perfect platform for fundamental research in ultrafast optics and nonlinear optics because of the feasibility of easy controlling of laser parameters and long interaction distance between light and fibers forming the cavity.
Apart from generation of conventional ultrashort pulses, there exist a lot of pulsing patterns in fiber lasers, such as dark solitons , domain-wall solitons  and noise-like pulses (NLPs) [3–15]. NLPs are a kind of mode-locked pulses in fiber lasers with the following features : 1) long pulse duration (from picosecond to nanosecond) with stochastically developing femtosecond intra-pulse structure; 2) smooth and broad bandwidth if averaged; 3) low temporal coherence. Based on their unique features, NLPs are very useful in real-time interrogation of fiber Bragg gratings , super-continuum generation  and laser machining . So far NLPs have been reported in various fiber lasers operating in the anomalous dispersion regime  and in the normal-dispersion-based cavities [9,10]. NLPs have been observed in fiber lasers using different mode locking mechanisms, for example, by using the nonlinear polarization rotation (NPR) technique , the nonlinear fiber loop mirror (NOLM) , or the saturable absorber made of carbon nanotubes . NLPs are also achieved at different central wavelengths: 1030 nm , 1550 nm  and 2000 nm . In general, NLPs in fiber lasers are one of the intrinsic mode-locking states and can be easily produced.
Research in the area of NLP properties gradually focused on controlling of the pulse energy and pulse bandwidth. As NLPs could be breaking-free, the achievable pulse energy and the pulse energy limitation are the important factors to consider in designing of practical systems. For example, the potential to increase the pulse energy from 15 nJ  to 135 nJ  in the Erbium-doped fiber (EDF) lasers has been experimentally validated. 249 nJ pulse energy has been achieved in an all-fiber Tm-doped NOLM-based oscillator operating at wavelength of 2 µm . Limited by the available pump power, the upper limit of pulse energy is still under investigation. Another interesting feature of NLPs is the spectral width. NLPs are a wave packet with stochastically developing intra-pulse-structure. Therefore, there are no two identical NLPs in a fiber laser [14,15]. However, we can still obtain the spectral measurement of NLPs by using an optical spectrum analyzer (OSA). It is because that the scanning time of the OSA is generally much slower than the cavity roundtrip time of a fiber laser. Consequently, the optical spectrum of NLPs measured by the OSA is an averaged result. Nevertheless, we could still retrieve information from the measured spectrum such as the pulse energy distribution, pulse energy density and equivalent pulse bandwidth. In the potential applications of super-continuum generation and those of optical coherent tomography , NLPs with broad bandwidth are favorable. Various schemes are adapted to achieve NLPs with broad bandwidth, which includes taking into account four-wave-mixing and the soliton self-frequency shift effects , as well as Raman scattering [10,17,18]. However, so far, the broadest bandwidth achieved was 165 nm in a dumbbell-shaped Yb-doped fiber laser . In this paper, we revisited the possible schemes to broaden the NLP bandwidth. Raman scattering effect is strengthened by tailoring the fiber laser cavity. 203 nm bandwidth is achieved, which is, to the best of our knowledge, the broadest bandwidth of NLPs. By using an external polarization beam splitter which functions as both the spectral filter and the polarization filter, the feasibility for obtaining 294 nm of spectral width is experimentally demonstrated.
2. Experimental results and discussion
Figure 1 shows the developed fiber laser setup. The cavity length is about 81.5 meters. A segment of 2-m EDF with dispersion of -48 (ps/nm)/km at 1560 nm is utilized as the gain medium. Pump power at 980 nm is introduced in the laser through a 980/1550nm wavelength-division multiplexer. A 5% output coupler is used to output pulses. Two polarization controllers (PCs) are applied to achieve mode-locking. A fiber pigtailed isolator is inserted in the laser to force unidirectional operation. To enhance the cavity nonlinearity, a segment of 3.9-m high nonlinear fiber (HNLF) (γ=10 W-1/km) is deployed after the EDF. The HNLF is selected as it is sufficient for enhancing self-phase modulation. Two segments of 10-cm ultrahigh numerical aperture fiber (NA = 0.41) are spliced at both ends of the HNLF as a bridge fiber to connect with other fibers. The UHNAF is chosen to reduce the splicing loss. An OSA (Yokogawa AQ6375B), a commercial auto-correlator (Femtochrome 103XL), a radio-frequency analyzer (N9320B, Agilent Technologies, Inc.), and a digital oscilloscope (DSO9104A, Agilent Technologies, Inc.) are used to monitor the laser output.
NLPs with bandwidth less than 100 nm are readily obtained. In general, contribution of strong nonlinear effects such as self-phase modulation corresponds to enhancement of bandwidth. Obtaining NLPs with broad bandwidth benefits from enhancing nonlinearity in the system. There also exist many spectral filtering effects in a fiber laser, which will limit pulse bandwidth, and special concerns should be taken to weaken this kind of spectral limitation.
In our experiments, several measures were taken to achieve spectrally broad NLPs: 1) the NPR technique was used, because it enables periodic spectral filtering  which could facilitate Raman scattering, if the spectral period matches Raman shift; 2) the nonlinear response was enhanced by limiting the output coupling to 5%, so that most of the pulse energy would remain in the cavity; a segment of HNLF was exploited, and a bridge fiber was used to reduce the transmission loss between the HNLF and the single mode fiber (SMF); 3) cavity engineering was carried out including placing the HNLF after the EDF along the pulse propagation direction, optimized the gain fiber length and matched HNLF length were used, and appropriate passive fiber length was chosen to make the artificial spectral filter period match the Raman shift. As expected, NLPs with broader bandwidth are gradually obtained after tailoring the fiber laser and appropriate polarization setting are made.
The NLP bandwidth can be increased with pump power. However, it is also subject to the polarization setting. In the experiment, different length of the passive fiber from 100 m to 10 m was adopted in the fiber cavity to enhance the generation of the spectrum induced by the Raman scattering effect. The length of HNLF was tailored from 10 m to 1 m, and experimental results indicates a segment of 3.9 m is in favor of broad NLP spectrum. Figures 2(a)-2(d) shows an optimized NLP when the pump power is maximized at 1.5 W. As shown in Fig. 2(a), the optical spectrum has a 3-dB bandwidth of 203.3 nm, distributing from 1545.4 nm to 1748.7 nm. The output power is 4.2 mW. There are two peaks at around 1557 nm and 1668 nm, respectively. The spacing between them matches the Raman shift relationship  well so the second peak is the Raman scattering of the first peak. The Raman component combined with the main pulse forms the wave packet of the NLP. The spectrum is not smooth around 1350 nm and 1800 nm, which is caused by water absorption. We note that the 10-dB bandwidth is 474.9 nm from 1417.0 nm to 1891.9 nm while the 20-dB bandwidth is 677.9 nm covering from 1315.5 nm to 1993.4 nm, which are comparable to super-continuum generation. It is worth noting that the smooth spectrum of the NLPs is a result of averaging over a large number of pulses. Moreover, the broadband spectrum of NLP is mainly contributed by a joint effect of broad NLP spectral component generated due to high nonlinearity and the resultant new spectral components. We also measured the pulse temporal features using a photodetector with a bandwidth of 2 GHz. The zoom-in of the NLP is shown in Fig. 2(b), where the pulse duration under the low-bandwidth measurement is 0.75 ns. As shown in Fig. 2(c), the pulse separation is 407.0 ns, which matches well with the cavity length. A typical autocorrelation trace of NLPs was shown in Fig. 2(d), the coherent peak sits on the top of a broad pedestal, which exhibits a broad wave packet comprised of randomly varying fine temporal structures. The sharp steps in Fig. 2d are due to the narrow scanning range of the auto-correlator compared with the pulse duration of the NLPs.
RF spectrum were shown in Figs. 3(a)–3(b), where the signal-to-noise-ratio of the fundamental frequency is about 57.7 dB together with a 7-dB difference between the 6th order and the fundamental one. It suggests that NLPs are a relative stable mode locking state although they are evolving in the fiber laser.
The influence of pump power to the NLP bandwidth was shown in Fig. 4, where the pump power increase broadens the NLP bandwidth from 146 to 192 nm. The broadening is monotonously increasing, but the increase is not a linear fit, showing some saturation tendency above roughly 1000 mW of pump power. Further increasing of pump power would therefore have only finite potential to spectrally broaden the NLPs. SMF with different lengths at the output is used to investigate the possible influence to the pulse spectrum. Due to the fs intra-pulse structure of the NLP, spectral broadening could be observed if the pulse energy was strong enough. We compared the measured bandwidth with pigtail fiber length from 10 m to 100 m with a 10 m step. The power of input NLPs is 4.2 mW. The variation of the 3-dB bandwidth is ± 2.5 nm @ 202 nm. Consequently, the pigtail length plays a weak influence on the NLP bandwidth.
The two-peak spectral appearance together with the dip of the NLP would be detrimental to potential applications. Therefore, we connected the fiber laser to a polarization projection system as shown in Fig. 5 to filter the spectrum. It is made of a PC and a polarization beam splitter (PBS). The two output branches of the PBS are polarization maintaining.
Adjusting the PC, the 3-dB bandwidth can be broadened up to 298.5 nm from output port 1, while that from the output port 2 is 170.5nm, as shown in Figs. 6(a)–6(b). NLPs could be considered as a bunch consisting of large number of narrow pulses with random peak power, pulse duration, and pulse separation. The different spectral components of the NLP have different polarization orientations. Therefore, it is possible to project different spectral components along different polarization axis of the PBS. Consequently, the polarization projection system also functions as a spectral filter. It is feasible to compress or filter out the first peak of the output NLPs as shown in Fig. 6(a). Thus, NLPs with dramatically broadened bandwidth, in our case from 203.3 nm jumping to 298.5 nm, could be expected.
NLPs are a special mode-locking state existing in fiber lasers. Various effects could affect the NLP bandwidth. To achieve bandwidth enhancement, combined methods should be taken. First, depending on the available pump power, a segment of 2 m gain fiber was chosen to fully absorb the pump energy; second, boosting of nonlinearity, including especially such as self-phase-modulation, should be taken care of by including a suitable HNLF in the cavity design; third, it is necessary to avoid spectral filtering responsible for terminating of spectral broadening on the one hand, while on the other it is important to take advantage of the second or even third transmission peak of a periodic spectral filter. In our experiment, enhanced nonlinearity together with Raman-scattering effect matched with the invisible spectral filter dramatically broadened the NLP bandwidth. Although our results revealed that the boost of the pump power would not allow to endlessly broaden the bandwidth, the recorded 3-dB bandwidth achieved in our work exceeded 200 nm. By using an external polarization projection system, this could be further broadened up to almost 300 nm.
National Natural Science Foundation of China (11674133, 11911530083, 61575089); Russian Foundation for Basic Research (19-52-53002); Natural Science Research of Jiangsu Higher Education Institutions of China (17KJA416004); Royal Society (IE161214); Protocol of the 37th Session of China-Poland Scientific and Technological Cooperation Committee (37-17); Horizon 2020 Framework Programme (790666).
We acknowledge support from Jiangsu Overseas Visiting Scholar Program for University Prominent Young & Middle-aged Teachers and Presidents and Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD). Mariusz Klimczak acknowledges support from Fundacja na rzecz Nauki Polskiej (FNP) in scope of First TEAM/2016-1/1.
1. H. Zhang, D. Y. Tang, L. M. Zhao, and X. Wu, “Dark pulse emission of a fiber laser,” Phys. Rev. A 80(4), 045803 (2009). [CrossRef]
2. H. Zhang, D. Y. Tang, L. M. Zhao, and X. Wu, “Observation of polarization domain wall solitons in weakly birefringent cavity fiber lasers,” Phys. Rev. B 80(5), 052302 (2009). [CrossRef]
3. M. Horowitz, Y. Barad, and Y. Silberberg, “Noise-like pulses with a broadband spectrum generated from an erbium-doped fiber laser,” Opt. Lett. 22(11), 799–801 (1997). [CrossRef]
4. J. W. Chen and L. M. Zhao, “Noise-like pulsed fiber lasers,” Laser & Optoelectronics Progress 54(11), 110002 (2017). [CrossRef]
5. S. Keren and M. Horowitz, “Interrogation of fiber gratings by use of low-coherence spectral interferometry of noise-like pulses,” Opt. Lett. 26(6), 328–330 (2001). [CrossRef]
6. Y. Takushima, K. Yasunaka, Y. Ozeki, and K. Kikuchi, “87-nm bandwidth noise-like pulse generation from erbium-doped fiber laser,” Electron. Lett. 41(7), 399–400 (2005). [CrossRef]
7. K. Ozgören, B. Oktem, S. Yilmaz, and F.Ö. Ilday, “83 W, 3.1 MHz, square-shaped, 1 ns-pulsed all-fiber-integrated laser for micromachining,” Opt. Express 19(18), 17647–17652 (2011). [CrossRef]
8. L. M. Zhao and D. Y. Tang, “Generation of 15-nJ bunched noise-like pulses with 93-nm bandwidth in an erbium-doped fiber ring laser,” Appl. Phys. B: Lasers Opt. 83(4), 553–557 (2006). [CrossRef]
9. L. M. Zhao, D. Y. Tang, J. Wu, X. Q. Fu, and S. C. Wen, “Noise-like pulse in a gain-guided soliton fiber laser,” Opt. Express 15(5), 2145–2150 (2007). [CrossRef]
10. D. J. Li, D. Y. Shen, L. Li, H. Chen, D. Y. Tang, and L. M. Zhao, “Raman-scattering-assistant broadband noise-like pulse generation in all-normal-dispersion fiber lasers,” Opt. Express 23(20), 25889–25895 (2015). [CrossRef]
11. J. F. Li, Z. X. Zhang, Z. Y. Sun, H. Y. Luo, Y. Liu, Z. J. Yan, C. B. Mou, L. Zhang, and S. K. Turitsyn, “All-fiber passively mode-locked Tm-doped NOLM-based oscillator operating at 2-µm in both soliton and noisy-pulse regimes,” Opt. Express 22(7), 7875–7882 (2014). [CrossRef]
12. Q. Q. Wang, T. Chen, M. S. Li, B. T. Zhang, Y. F. Lu, and K. P. Chen, “All-fiber ultrafast thulium-doped fiber ring laser with dissipative soliton and noise-like output in normal dispersion by single-wall carbon nanotubes,” Appl. Phys. Lett. 103(1), 011103 (2013). [CrossRef]
13. X. W. Zheng, Z. C. Luo, H. Liu, N. Zhao, and Q. Y. Ning, “High-energy noise-like rectangular pulse in a passively mode-locked figure-eight fiber laser,” Appl. Phys. Express 7(4), 042701 (2014). [CrossRef]
14. L. M. Zhao, D. Y. Tang, T. H. Cheng, H. Y. Tam, and C. Lu, “120 nm Bandwidth noise-like pulse generation in an erbium-doped fiber laser,” Opt. Commun. 281(1), 157–161 (2008). [CrossRef]
15. A. F. J. Runge, C. Aguergaray, N. G. R. Broderick, and M. Erkintalo, “Coherence and shot-to-shot spectral fluctuations in noise-like ultrafast fiber lasers,” Opt. Lett. 38(21), 4327–4330 (2013). [CrossRef]
16. D. Huang, E. A. Swanson, C. P. Lin, J. S. Schuman, W. G. Stinson, W. Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafito, and J.G. Fujimoto, “Optical coherence tomography,” Science 254(5035), 1178–1181 (1991). [CrossRef]
17. L. A. Vazquez-Zuniga and Y. Jeong, “Super-broadband noise-like pulse erbium-doped fiber ring laser with a highly nonlinear fiber for Raman gain enhancement,” IEEE Photonics Technol. Lett. 24(17), 1549–1551 (2012). [CrossRef]
18. W. C. Chang, J. H. Lin, T. Y. Liao, and C. Y. Yang, “Characteristics of noise-like pulse with broad bandwidth based on cascaded Raman scattering,” Opt. Express 26(24), 31808–31816 (2018). [CrossRef]
19. H. Chen, S. P. Chen, Z. F. Jiang, and J. Hou, “0.4 µJ, 7 kW ultra-broadband noise-like pulse direct generation from an all-fiber dumbbell-shaped laser,” Opt. Lett. 40(23), 5490–5493 (2015). [CrossRef]
20. L. M. Zhao, D. Y. Tang, X. Wu, and H. Zhang, “Dissipative soliton generation in an Yb-fiber laser with an invisible intra cavity band-pass filter,” Opt. Lett. 35(16), 2756–2758 (2010). [CrossRef]
21. G. P. Agrawal, Nonlinear Fiber Optics, 4th ed. (Academic, 2007).