We present a comparison between an in-field and in-laboratory 50 km ultralong Erbium fiber lasers actively mode-locked with repetition rate varying from 1 to 10 GHz generating pulses from 35.2 to 68.7 ps. The pulse widths generated at higher frequencies are in agreement with Kuizenga-Siegman theory. However, for lower frequencies the pulses have higher intracavity peak power which allows the soliton effect to take place. Depending on the pump power level, the repetition rate and the cavity length both lasers can operate in active mode-locking or under the influence of the soliton regime that locks the pulse duration according to the dispersion and cavity length. Due to the soliton robustness, this condition eliminates most of the environmental influence in the in-field mode-locking regime and makes both lasers very similar.
©2012 Optical Society of America
Studies of the capacity assessment show the necessity of new concepts development in lightwave systems to support the huge global network traffic demand . Improved laser characteristics such as high repetition rate, generation of ultrashort pulses and high output power are some of these challenges.
The study of long cavities started in 1982 by Nakazawa et al.  who demonstrated a fiber laser with 101 meters long and emission wavelength of 1150 nm. Recent publications reports advances in a new type of lasers, the ultralong fiber lasers. Ania-Castañón  first proposed and theoretically studied the ultralong cavity by using fiber Bragg gratings for Raman amplification which were applied experimentally to demonstrate a quasi-lossless transmission in cavity length with 75 km  and 270 km . In these works, the optical transmission link and the cavity are the same medium and the gain was provided by distributed Raman amplification with bidirectional pumping. Analysis of mode structure in 84 km ultralong Raman fiber laser  and the impact of nonlinear spectral broadening  are complementary studies of this kind of laser.
Actively mode-locked Erbium doped fiber lasers with cavity length of 1 and 50 km were used to demonstrate localization behavior in their frequency domain . Another application of these giant lasers was shown by Scheuer et al. for secure key distribution by using two Erbium doped fiber amplifiers (EDFA), 10 km , 50 km  and 200 km  of standard telecom fiber and also, two mirrors in each end of the link to set the peak reflectivity frequencies representing a key bit. Also, cavity lengths of kilometers have been developed to generate very low repetition rates and very high pulse energies in passive mode-locked regime .
In recent work, we investigated the mechanisms responsible for pulse formation and evolution in Erbium-doped fiber lasers with cavity lengths varying from 16.4 m to 100.8 km actively mode-locked. It was possible to distinguish three operation regimes which depend on the ratio between the cavity length and the dispersion and nonlinear lengths . Another important work is the demonstration of an ultralong fiber laser with a ring configuration, generating pulses inside the communication link . The polarization controllers are commonly used for the generation and control of ultrashort pulses  and they were tuning to obtain the condition of stable pulses. The balance between the anomalous dispersion of the SMF with the non-linearity generated by amplification creates an environment for the emergence of solitons which maintain the ideal conditions for the least deterioration of the pulse along its propagation . We identified specific conditions of intracavity peak power and modulation frequency that contributes to the generation and maintenance of soliton pulses.
In this paper, we investigate the generation of solitons in two distinct (in-field and in-laboratory) 50 km ultralong Erbium-doped fiber lasers that were designed to operate in actively mode-locked regime with a modulation frequency from 1 to 10 GHz. A comparison between the output pulse widths with Kuizenga and Siegman theory  demonstrated a soliton regime with intracavity pulse compression.
2. Experimental setup
Figure 1(a) shows the in laboratory experimental setup of the ultralong fiber laser. It consists of 2 m of Erbium-doped fiber with a co-propagating pump laser of 980 nm with a maximum pump power of 400 mW, an optical isolator with 50 dB of isolation and insertion loss of 0.07 dB at 1550 nm, a polarization controller, an amplitude modulator, an output coupler of 15.3%, standard single mode fiber (SMF) which is designed by using two spools of 25.3 km each and a 1% monitoring coupler after the first SMF spool. Moreover, another setup was carried in a geographically-distributed optical network called KyaTera that is a Brazilian test-bed for studying advanced communication technologies . Figure 1(b) shows the optical link section which connects two universities in downtown Sao Paulo, one of the largest city of the world.
The total length of the cavity is approximately 50 km with attenuation of 18.8 dB for the in-field laser and 13.2 dB for the in-lab setup. For these ultralong fiber lasers the fundamental repetition rate is about 4 kHz.
3. Results and discussion
3.1 Continuous wave operation
Considering the total pump power of 400 mW, the average output power for the ultralong Erbium fiber laser was 10 mW when the spools were used. For the in-field setup, we obtained 8 mW for the same total pump power. Figure 2(a) shows the output characteristics of both laser setups.
The high loss of the ultralong cavity results in low conversion efficiency: 2.65% for in-lab laser and 2.25% for in-field setup. The threshold is approximately 27 mW of pump power for the in-lab configuration and 40 mW for the other case.
The output spectra in both cases are presented in Fig. 2(b). In the installed link, the peak wavelength was around 1556.8 nm with OSNR of 34.3 dB. Using the spools, we observed an improvement of OSNR to 45.5 dB and a slightly shift in the emission wavelength to 1557.8 nm because the difference of polarization adjustment. Note that the spectra shown in the in-field laser present a large incoherent pedestal, due to the huge attenuation within the cavity.
3.2 Active mode-locking operation
For the mode-locking operation, an amplitude modulator was inserted in the cavity to establish modulation frequencies from 1 to 10 GHz. Figure 3(a) shows the output pulse widths in both configurations at a repetition rate of 1 GHz. The measurements of the pulse were 59.3 ps for in-laboratory fiber laser and 63.0 ps for in-field setup. Besides the wider duration, the KyaTera fiber laser presents a noisy oscilloscope shape.
Figure 3(b) shows the comparison of output spectrum. For in-laboratory setup, the spectral width was 0.072 nm that results in a time-bandwidth product of 0.5286. The FWHM was 0.068 nm for in-field fiber laser with small shift in the peak emission wavelength because of the adjustment of the incoming polarization. The time-bandwidth product is 0.5315 that indicates an excess of bandwidth which means the pulses are not transform-limited. At the inset, in the log scale it is observed the soliton spectral sidebands  which explain the similar behavior of these lasers. According to Ref , both lasers are in dispersion dominant regime and the cavities are long enough to allow the soliton effect to manifest and determine the pulse duration and evolution.
Figure 4 shows the measured output pulse width as a function of modulation frequency from 1 to 10 GHz, with pump power of 125 mW for in-laboratory laser (blue square curve) which is designed by two spools of 25.3 km of SMF. As predicted by Kuizenga and Siegman theory , the pulse width produced by active mode-locking is inversely proportional to the square root of modulation frequency as where is the pulse width and is the modulation frequency. For frequencies from 5 to 10 GHz, the average pulse duration was 51.9 to 35.2 ps which were according to the theory. However, the pulse widths at the modulation frequencies of 1 to 4 GHz (57.7 to 47.3 ps), deviate from the theoretical curve and are shorter than expected.
We performed measurements in the second setup which uses the KyaTera network keeping the same pump power level, as shown in Fig. 4 (red circle curve). For frequencies from 7 to 10 GHz, the duration varied from 50.0 to 39.5 ps. The pulse widths are slightly larger than expected because the instabilities of the huge number of modes and higher attenuation of in-field setup. However, for modulation frequencies of 1 GHz (63.0 ps) to 5 GHz (46.4 ps) the pulse widths deviate from the theoretical curve that are shorter than expected. This is due to the intracavity soliton regime.
3.3 Soliton and intracavity peak power analysis
With an average dispersion of 17 ps/nm.km for the fiber in the ring, the theoretical peak soliton power for the lasers was ~25 mW with a ~50 ps pulse, which was calculated by Eq. (1):20].
Figure 5(a) shows the measured intracavity peak powers (blue square curve) for the in-laboratory laser and the calculated first order soliton peak power (black triangle curve). For frequencies below 8 GHz the intracavity peak power is higher than the theoretical peak soliton power. Considering for instance the modulation frequency of 2 GHz, the intracavity pulse has energy equivalent to a soliton of order 2 (N = 2). At 8 GHz we obtained N = 1 which is the soliton effect threshold.
We performed the same analysis in the second configuration which used the link of KyaTera Network. Figure 5(b) shows the values of intracavity peak power (red circle curve) that were measured considering the same conditions of pump power, modulation depth and modulation frequency in each case. The intracavity peak power is higher than the theoretical peak soliton power (black triangle curve) for frequencies below 6 GHz, where we obtained N = 1 which means that below this frequency, the soliton regime is present. For both setups, there is no soliton regime at higher frequencies. The difference of the frequency threshold in these configurations is due to the higher intracavity loss of the in-field setup.
3.4 Output pulse spectrum analysis
The mode-locked pulse spectra of the ultralong Erbium-doped fiber laser for the in-laboratory setup are shown in Fig. 6 . Output spectra at modulation frequencies of 1 to 7 GHz (blue graphics) show spectral shape corresponding to the soliton-like nature of the pulses. In comparison with the Fig. 5(a), the soliton condition is determined by the intracavity peak power that is higher than the fundamental soliton power. For modulation frequencies of 8 to 10 GHz, there are different spectral profiles with multiple peaks with distance between them related to the modulation frequency for each case.
In addition, we plot the output of the ultralong Erbium-doped fiber laser of the KyaTera Network in Fig. 7 , that show the measurements with sidebands typically obtained when there is soliton propagation (red graphics). Figure 5(b) shows the intracavity peak powers for modulation frequencies below 5 GHz that are higher than the fundamental soliton power. Also, the spectra can confirm the soliton conditions. For frequencies higher than 6 GHz, the pulse formation is determined by the active modulator without soliton effects.
By comparing Figs. 5(a) and 5(b) with Fig. 6 and Fig. 7, respectively, we observe that their respective spectra are a confirmation of the threshold of intracavity soliton regime at 8 GHz for in-laboratory and 6 GHz for in-field fiber lasers.
3.5 Intracavity pulse compression
In order to investigate the intracavity propagation, a comparison was made between output characteristics and the measurements after propagation in 25 km of SMF. These measurements were made in the in-laboratory setup with 1% output coupler between the two fiber spools. As shown in Fig. 3(a), the output pulse width has duration of 59.3 ps and the corresponding optical spectrum has FWHM of 0.072 nm at 1 GHz. The time-bandwidth product is 0.5286.
Figure 8(a) shows the pulse width of 42.4 ps measured in the 1% monitoring coupler after the first SMF spool at the same frequency of 1 GHz. This measurement demonstrated that the pulse width is compressed along the intracavity propagation. In addition, we can observe in the Fig. 8(b) the corresponding optical spectrum with FWHM of 0.069 nm. At the inset, in the log scale it is observed the spectral sidebands . The time-bandwidth product in this case is 0.3623, close to the reference value of 0.315 for hyperbolic secant profile.
The positive dispersion allows the fiber itself to act as a distributed compressor which contributed to the degree of pulse narrowing. This pulse narrowing is defined by two independent parameters, the soliton order N and the soliton period Z0. For our 1 GHz experiment conditions, the peak intracavity power is Pin = 233 mW. Assuming the peak power for fundamental soliton is P0 = 18.5 mW, it can obtain N = (233/18.5)1/2 = 3.5. The soliton period is Z0 = 77 km that is around 1.5 times longer than the cavity length, obtained by the Eq. (2).
The existence of a soliton period refers to the fact that, for N > 1 (corresponding to higher-order solitons), the pulse behavior is periodic with propagation and for this ultralong fiber laser, it matches the cavity length. Z0 is independent of N and it can be a scale factor to the compression .
As shown in the Fig. 8(a), the compression factor was 1.4X (59.3/42.4). This result shows that very long cavity lengths can contribute to development of soliton pulse shortening.
The comparison between in-field and in-laboratory 50 km ultralong Erbium fiber lasers actively mode-locked with repetition rate varying from 1 to 10 GHz show small differences in the pulse duration. The pulse widths generated at higher frequencies are in agreement with Kuizenga-Siegman theory. However, for lower frequencies the pulses have higher intracavity peak power which allows the soliton effect to take place. Depending on the pump power level, the repetition rate and the length of the cavity both lasers can operate actively mode-locked or under the influence of the soliton regime that locks the pulse duration according to the dispersion and cavity length . Due to the soliton robustness, this condition eliminates most of the environmental influence in the mode-locking regime and makes both lasers very similar.
This work is supported by Fundo Mackenzie de Pesquisa, Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) and Instituto Nacional de Ciência e Tecnologia Fotônica para Comunicações Ópticas (INCT FOTONICOM).
References and links
1. E. B. Desurvire, “Capacity demand and technology challenges for lightwave systems in the next two decades,” J. Lightwave Technol. 24(12), 4697–4710 (2006). [CrossRef]
2. M. Nakazawa, M. Tokuda, and N. Uchida, “Continuous-wave laser oscillation with an ultralong optical-fiber resonator,” J. Opt. Soc. Am. 72(10), 1338–1344 (1982). [CrossRef]
4. T. J. Ellingham, J. D. Ania-Castañón, R. Ibbotson, X. Chen, L. Zhang, and S. K. Turitsyn, “Quasi-Lossless Optical Links for Broad-Band Transmission and Data Processing,” IEEE Photon. Technol. Lett. 18(1), 268–270 (2006). [CrossRef]
5. S. K. Turitsyn, J. D. Ania-Castañón, S. A. Babin, V. Karalekas, P. Harper, D. Churkin, S. I. Kablukov, A. E. El-Taher, E. V. Podivilov, and V. K. Mezentsev, “270-km ultralong Raman fiber laser,” Phys. Rev. Lett. 103(13), 133901 (2009). [CrossRef] [PubMed]
6. S. A. Babin, V. Karalekas, P. Harper, E. V. Podivilov, V. K. Mezentsev, J. D. Ania-Castañón, and S. K. Turitsyn, “Experimental demonstration of mode structure in ultralong Raman fiber lasers,” Opt. Lett. 32(9), 1135–1137 (2007). [CrossRef] [PubMed]
7. V. Karalekas, J. D. Ania-Castañón, P. Harper, S. A. Babin, E. V. Podivilov, and S. K. Turitsyn, “Impact of nonlinear spectral broadening in ultra-long Raman fibre lasers,” Opt. Express 15(25), 16690–16695 (2007). [CrossRef] [PubMed]
8. B. Fischer, B. Vodonos, S. Atkins, and A. Bekker, “Experimental demonstration of localization in the frequency domain of mode-locked lasers with dispersion,” Opt. Lett. 27(12), 1061–1063 (2002). [CrossRef] [PubMed]
10. A. Zadok, J. Scheuer, J. Sendowski, and A. Yariv, “Secure key generation using an ultra-long fiber laser: transient analysis and experiment,” Opt. Express 16(21), 16680–16690 (2008). [CrossRef] [PubMed]
11. O. Kotlicki and J. Scheuer, “Secure key distribution over a 200 km long link employing a novel Ultra-long Fiber Lasers (UFL) scheme,” in Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science and Photonic Applications Systems Technologies, Technical Digest (CD) (Optical Society of America, 2010), paper ATuA4.
12. H. G. Rosa and E. A. de Souza, “58 kHz Ultra-low Repetition Rate Ultralong Erbium-Doped Fiber Laser Mode-Locked by Carbon Nanotubes,” in Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science and Photonic Applications Systems Technologies, Technical Digest (CD) (Optical Society of America, 2011), paper JWA29.
13. L. A. M. Saito and E. A. Thoroh de Souza, “Identifying the mechanisms of pulse formation and evolution in actively mode-locked Erbium fiber lasers with meters and kilometers-long,” Opt. Express (Submitted for publication).
14. L. A. M. Saito, M. A. Romero, and E. A. Souza, “48.8 km Ultralong Erbium fiber laser in active mode-locking operation,” Opt. Rev. 17(4), 385–387 (2010). [CrossRef]
15. E. S. Boncristiano, L. A. M. Saito, and E. A. De Souza, “396 fs pulse from an asynchronous mode-locked Erbium fiber laser with 2.5-12 GHz repetition rate,” Microw. Opt. Technol. Lett. 50(11), 2994–2996 (2008). [CrossRef]
16. M. Alcon-Camas, A. E. El-Taher, H. Wang, P. Harper, V. Karalekas, J. A. Harrison, and J.-D. Ania-Castañón, “Long-distance soliton transmission through ultralong fiber lasers,” Opt. Lett. 34(20), 3104–3106 (2009). [CrossRef] [PubMed]
17. D. J. Kuizenga and A. E. Siegman, “FM and AM mode locking of the homogeneous laser - Part I: Theory,” IEEE J. Quantum Electron. 6(11), 694–708 (1970). [CrossRef]
18. M. Silva, T. C. Carvalho, R. M. Silveira, G. M. Ferreira, W. V. Ruggiero, H. L. Fragnito, H. Waldmann, C. Ruggiero, and L. F. Lopez, “Fiber-based testbed architecture enabling advanced experimental research,” in 2nd International Conference on Testbeds and Research Infrastructures for the Development of Networks & Communities, IEEE, 246–253 (2006).
19. S. M. J. Kelly, “Characteristic sideband instability of periodically amplified average soliton,” Electron. Lett. 28(8), 806–807 (1992). [CrossRef]
20. G. P. Agrawal, Nonlinear Fiber Optics, 4th ed. (Academic Press 2007).
21. L. F. Mollenauer, R. H. Stolen, J. P. Gordon, and W. J. Tomlinson, “Extreme picosecond pulse narrowing by means of soliton effect in single-mode optical fibers,” Opt. Lett. 8(5), 289–291 (1983). [CrossRef] [PubMed]