2584
OPTICS LETTERS / Vol. 40, No. 11 / June 1, 2015
Tunable slow light via stimulated Brillouin scattering at 2 μm based on Tm-doped fiber amplifiers Xiong Wang, Pu Zhou,* Xiaolin Wang, Hu Xiao, and Zejin Liu College of Optoelectronic Science and Engineering, National University of Defense Technology, Changsha 410073, China *Corresponding author: [email protected] Received March 4, 2015; revised May 4, 2015; accepted May 11, 2015; posted May 11, 2015 (Doc. ID 235427); published May 28, 2015 We present a slow light system based on stimulated Brillouin scattering (SBS) at 2 μm. A single-frequency fiber laser with Tm-doped fiber amplifiers was used to generate the SBS signal laser and the Brillouin pump light at 1.971 μm. The maximum delay time reaches 16 ns for pulses with 43-ns width, and the pulse width is broadened to 56.4 ns. The maximum delay time for 57-ns pulses reaches 33.4 ns, and the pulse width is broadened to 77.6 ns. The relative delays are 0.37 and 0.58 for 43 and 57 ns pulses, respectively. This is the first demonstration, as far as we know, on a slow light system at 2 μm, which may be substantial for future optical communications and LIDAR systems employing laser sources near 2-μm band. © 2015 Optical Society of America OCIS codes: (140.3070) Infrared and far-infrared lasers; (140.3510) Lasers, fiber; (190.4370) Nonlinear optics, fibers; (290.5900) Scattering, stimulated Brillouin. http://dx.doi.org/10.1364/OL.40.002584
The broadest gain spectrum around ∼1700–2200 nm among rare-earth-doped fiber lasers enables Tm-doped fiber lasers (TDFLs) significant laser sources in various applications in recent decades [1–3]. One of the valuable potential applications is optical communications [4,5]. The exponentially increasing volume of internet traffic drives the current networks using 1.55-μm fiber lasers to their capability limits to a large extent. Thus the TDFLs with much broader emission band may become the prospective laser sources for future optical communications, and the low loss of hollow-core photonic bandgap fibers near 2 μm further enhances the availability and practicability of prospective networks operating near 2 μm [6]. Another meaningful application is the light radar (LIDAR) [7–10]. The emission band of TDFLs covers the transmission window of atmosphere around 2 μm, which also contains the absorption lines of various important atmosphere gases. So TDFLs are admirable sources for LIDAR in remote sensing and detecting. Thus lasers sources based on Tm-doped fiber have been investigated intensively in the last decades [11,12]. It should be noted that the two applications above have one resemblance, i.e., the control of group delay of light. In fact, all-fiber-based buffer is one crucial topic for optical communications, and the tunable delay of optical signals in the futuristic fiber networks is an interesting approach for improving the data traffic, as well as enhancing the capability of optical signal processing [13–15]. In addition, the tunable delay of optical pulses is a favorable method to effectively and rapidly correct the optical group delay mismatches among all the apertures in a LIDAR system [9,10]. Tunable delay of optical pulses has been demonstrated and investigated vastly in recent years, and one of the most convenient methods is to realize fast and slow light via stimulated Brillouin scattering (SBS) [16–24]. The SBS effect can cause large normal dispersion and increase the group index in a narrow spectral range, which will reduce the group velocity of optical pulses in fibers [17,23]. The tunable delay of optical fiber pulses based on SBS possesses some excellent advantages, such as low operating pump power and linearly tunable delay [16,21,22]. It is also convenient for 0146-9592/15/112584-04$15.00/0
the delay system being integrated into the optical networks nearly perfectly [13–15]. Thus, slow light based on SBS stimulated tremendous research passion, and the objects are all mainly focused on fiber lasers near 1.55 μm, since it is the current practical wavelength for optical communications and LIDAR. However, as aforementioned, TDFLs near 2 μm may provide an advanced optical communication method and improve the performance of LIDAR. Although previous investigations state that the slow light can be generated in the current communication window (∼1.55 μm), there has been no demonstration on slow light at 2 μm so far, to the best of our knowledge. In this Letter, we experimentally demonstrated the tunable slow light based on SBS at 2 μm. A single-frequency (SF) fiber laser was used as the seed laser to generate the pump laser and SBS signal. A piece of 450-m single mode fiber (SMF) was employed as the delay fiber. The delay time can be tuned by adjusting the pump power. The maximum delay time can reach 33.4 ns when using pulses with width of 57 ns, and the delay time can reach 16 ns when pulses with 43-ns width were used. This is the first demonstration, as far as we know, on slow light based on SBS at 2 μm. The experimental setup is shown in Fig. 1. A SF fiber laser at 1971 nm based on ultra-short cavity with a linewidth of less than 100 kHz and an output power of ∼40 mW was used as the seed laser [25]. A 50/50 fiber coupler split the laser into two parts. One part of the laser was amplified by a home-made single-mode Tm-doped
Fig. 1. Experimental setup of the slow light system based on SBS using TDFAs. © 2015 Optical Society of America
June 1, 2015 / Vol. 40, No. 11 / OPTICS LETTERS
fiber amplifier (TDFA) to serve as the Brillouin pump (BP) light. The power of the BP light can reach ∼500 mW with a continual tuning range by adjusting the current of the pump laser diodes. The other part of the laser was amplified by a similar single-mode TDFA, then the output laser was launched into a length of 450-m SMF (SMF-28 fiber, SMF 1) via a three-port circulator. The SBS light was generated in the SMF 1 and backward propagated into the circulator. The continuous wave SBS light was then modulated to pulses by an acousto-optic modulator (AOM), which had a rise time of ∼70 ns. The BP light and the SBS light encountered each other in another piece of 450-m SMF (SMF-28 fiber, SMF 2). A polarizationinsensitive isolator (ISO) was inserted between the SMF 2 and the AOM to stop the residual BP light. The SBS pulses were delayed in the SMF 2 and launched out through the second circulator. The output pulses were measured by a high-speed photodetector and a digital oscilloscope with bandwidth of 1 GHz. A unique advantage of this system is that the frequency difference of the BP light and SBS light is essentially guaranteed to be near the Brillouin frequency shift with only one seed laser. The spectra of the BP light and SBS light were monitored by an optical spectrum analyzer (OSA) with resolution of 0.05 nm, as shown in Fig. 2. The Brillouin frequency shift at 2 μm is conformed in the results. The repetition rate of the AOM was set to be 100 kHz, and the widths of the SBS pulse were set to be 44 and 60 ns, respectively. The SBS signal’s power was maintained at ∼150 μW. When the power of the BP light was increased gradually, the SBS pulse was delayed, as shown in Fig. 3. The maximum delay time reaches 11.2 ns for pulses with a width of 44 ns, and the pulse width is slightly broadened to 46.8 ns, corresponding to a pulse-broadening factor of 1.07. The maximum delay time reaches 20 ns for pulses with a width of 60 ns, and the pulse width is slightly broadened to 64.8 ns, corresponding to a pulse
Fig. 2. Spectra of the BP light and SBS light of the slow light system.
Fig. 3. Observation of slow light via SBS at 2 μm. (a) Input pulse width is 44 ns and (b) input pulse width is 60 ns.
2585
broadening factor of 1.08. Both the broadening factors are rather low, thus the pulse shapes are well maintained during the delay process. Figure 4 shows the delay time as a function of the Brillouin gain. The delay time via SBS can be expressed as ΔT ≈ G∕ΓB , where ΓB is the Brillouin linewidth. The Brillouin gain factor is G g0 I p L, where g0 is the line center gain factor, L is the fiber length, I p is the BP power, and the exponential of G is the small-signal gain [17]. Thus the delay time can be continually and linearly tuned via adjusting the BP power. The maximum relative delay is ΔT∕tp 0.33 and 0.25 for pulse with width of 60 and 44 ns, respectively, where tp represents the pulse width. According to the expression ΔT ≈ G∕ΓB , the delay time at 2 μm will be higher than the delay time at 1.55 μm with a fixed Brillouin gain, due to the narrower Brillouin gain linewidth (∼10 MHz to ∼30 MHz) [17,26]. However, the relatively high background loss of current commercially available SMF-28 fiber at 2 μm may reduce the Brillouin gain and delay time, and employing of improved fibers with lower loss at 2 μm will mitigate the limit. In fact, the curves in Fig. 4 show uneven increase of the delay time versus the Brillouin gain. We attribute this deficiency to the instability of the SBS light generated from the SMF 1, since the randomness character of SBS determines that the SBS power may fluctuated in a certain extent [27]. A natural method to solve this problem is to employ Brillouin amplifier based on a TDFA [28]. However, the additional amplified spontaneous emission (ASE) noise introduced by utilizing a TDFA will degrade the signal-to-noise ratio (SNR) of the slow light system. So we improved the system by employing Brillouin fiber laser (BFL) configuration, as shown in Fig. 5. A 10/90 fiber coupler was inserted before the AOM, then 10% of the SBS light was launched back into the SMF 1. Thus a BFL cavity was built up by a circulator,
Fig. 4. Delay time versus the Brillouin gain for different pulse widths.
Fig. 5. BFL.
Experimental setup of the slow light system based on
2586
OPTICS LETTERS / Vol. 40, No. 11 / June 1, 2015
a coupler, and a piece of SMF. The 90% port of the coupler launched the SBS laser into the AOM to generate the pulsed signal laser. Using this improved configuration, the output of the SBS pulsed laser is much stabler, and the SNR of the SBS pulse is improved, so the performance of the slow light system is enhanced. Furthermore, the output power of the BFL is higher than that of SBS light without the coupler, which means the pump power of the TDFA 1 can be also reduced to make the system more feasible in practical applications. The results of slow light using the modified configuration are shown in Fig. 6. The power of the SBS pulses was adjusted to be the same with that of previous experiments by turning down the pump power of the TDFA 1. The initial pulse widths are slightly narrower than that of the previous results due to the more stable and more precise measurement. The maximum delay time for pulses with a width of 43 ns reaches 16 ns, but the pulse width is also broadened to 56.4 ns, corresponding to a broadening factor of 1.31. The maximum delay time for pulses with a width of 57 ns reaches 33.4 ns, and the pulse width is broadened to 77.6 ns, corresponding to a broadening factor of 1.35. The relative delays are 0.37 and 0.58 for 43 and 57 ns pulses, respectively. From the linear fit of the experimental data curves in Fig. 6, we conclude that the delay efficiency is 5 and 11.7 ns/dB for 43 and 57 ns pulses, respectively. The experimental curves in Fig. 6 show slight nonlinearity and fluctuations. One reason may be that the Brilloiin gain linewidth at 2 μm is narrower than that at 1.5 μm, so the linear tuning range for time delay will be shorter when the pulse width is relatively short [29]. This can be conquered via broadening the BP wave power spectral density, by which shorter pulses can be delayed effectively and linearly [23]. Since the polarization states will also influence the delay process, the environment vibration may introduce fluctuations, which can be mitigated by enhancing the stability of the system and the polarization states. Actually, the first obstacle one must face is the considerable background loss of silica fiber when generating SBS slow light at 2 μm (about 20 dB/km with common telecommunication fibers). In the previous theoretical demonstrations, the loss coefficient of the fiber is either neglected or essentially tiny (∼0.2 dB∕km) [16,30,31], since the operation wavelength is around ∼1.5 μm. Thus the delay of SBS slow light almost starts when the Brillouin gain is slightly more than zero theoretically.
Fig. 6. Experimental and simulated delay time data versus the Brillouin gain data for different pulse widths with improved configuration. (a) Input pulse width is 43 ns and (b) input pulse width is 57 ns.
However, the loss of fiber plays an obvious role on the “starting point” of the delay time versus Brillouin gain, i.e., the delay begins when the gain is ∼6 [17] or ∼5 dB [20], though the loss coefficient is rather low. As for slow light at 2 μm, the seriously increasing loss coefficient makes the “starting point” much higher compared with that of slow light at 1.5 μm band, as shown in Figs. 4 and 6. One should also note that the AOM inevitably introduces frequency shift to the generated signal pulses (80 MHz in this experiment). The reason we use AOM instead of electro-optic modulator (EOM) is that the EOM at 2 μm in our lab cannot provide perfect intense modulation with high SNR for the time being. Thus the frequency of the SBS pulses is 80 MHz away from the center of Brillouin resonance, which means the delay results can be further improved when making the frequency of the pulses in the center of the resonance. Furthermore, the detune of the SBS pulse from the resonance center indicates that the actual Brillouin gain should be modified, provided that the practical gain factor is practically measured. Furthermore, the difference of pulse widths in our experiment is small, owing to the limited bandwidth of the AOM for generating narrower pulse width. Thus the curves of delay time with two different pulse widths even overlap to a certain extent. Due to the high loss coefficient of fiber at 2 μm and the large detune introduced by the AOM, the gain is far below the saturation value with SBS signal’s power of ∼150 μW. Actually, the saturation value may be even more than 30 dB at 1.5 μm band [20]. Thus the saturation effects have not occurred seriously in our experiment yet. We simulated the slow light at 2 μm based on threewave coupled equations [29], and the results are also shown in Fig. 6. The simulated maximum delay time for 43 ns pulses is 15.8 ns, and that for 57 ns pulses is 33.7 ns. The simulated broadening factor when the gain reaches the maximum value is 1.29 and 1.37 for 43 and 57 ns pulses, respectively. The simulated curves of delay time data versus gain data have the similar increasing trend with those of the experimental results. From the above comparison, one can see that the simulated and experimental results matched well with each other in a certain extent, which indicates that realization of slow light at 2 μm is feasible experimentally and theoretically. The simulation does not include the processes of selfphase modulation and cross-phase modulation for simplification. By the way, the spectral part of the signal pulses near the resonance center of SBS can be amplified when the pump power is intense enough, which will result in the decrease of the frequency detune between the pump and signal lights. Thus, the delay time might be even higher when the pump power is increasing but well below the saturation threshold. We also note that when the Brillouin gain is lower than 19 dB, the increase of the delay time is slightly nonlinear, but when the Brillouin gain is higher than 19 dB, the delay time increases more linearly. Meanwhile, the broadening factor is suddenly increased from 1.15 to 1.35 when the Brillouin gain surpasses 19 dB. A detailed simulation will be performed to further investigate this phenomenon. However, the experimental and simulated results demonstrate the first realization of slow light at 2-μm band,
June 1, 2015 / Vol. 40, No. 11 / OPTICS LETTERS
which may be rather meaningful for the future optical networks and LIDAR systems employing 2-μm lasers. In conclusion, we demonstrate the first tunable slow light system at 2 μm based on TDFAs. BFL was used to generate SBS signal laser with the Brillouin shfit compared with that of the BP laser. The maximum delay time of 43-ns pulses reaches 16 ns with a broadening factor of 1.31, and the maximum delay time of 57-ns pulses reaches 33.4 ns with a broadening factor of 1.35. The performance of the slow light system can be further improved by utilizing BP light with broader power spectral density. This experiment shows the availability of slow light laser system at 2-μm band, which will be futuristic and significant for the prospective optical communications and LIDAR systems. This work was supported by the Graduate Student Innovation Foundation of National University of Defense Technology (Grant No. B130704), National Natural Science Foundation of China (Grant No. 11274386), and Program for New Century Excellent Talents in University and Hunan Provincial Innovation Foundation for Postgraduate. We thank Haibin Lv for his helpful discussion and support when performing the simulation. References 1. S. D. Jackson, Nat. Photonics 6, 423 (2012). 2. J. Geng, Q. Wang, and S. Jiang, Proc. SPIE 8164, 816409 (2011). 3. P. F. Moulton, G. A. Rines, E. V. Slobodtchikov, K. F. Wall, G. Frith, B. Samson, and A. L. G. Carter, IEEE J. Sel. Top. Quantum Electron. 15, 85 (2009). 4. S. U. Alam, Z. Li, J. M. O. Daniel, Y. Jung, A. M. Heidt, and D. J. Richardson, OptoElectronics and Communications Conference Held Jointly with 2013 International Conference on Photonics in Switching (OECC/PS) (Optical Society of America, 2013), paper WS1_3. 5. Z. Li, A. M. Heidt, N. Simakov, Y. Jung, J. M. O. Daniel, S. U. Alam, and D. J. Richardson, Opt. Express 21, 26450 (2013). 6. M. N. Petrovich, F. Poletti, J. P. Wooler, A. M. Heidt, N. K. Baddela, Z. Li, D. R. Gray, R. Slavík, F. Parmigiani, N. V. Wheeler, J. R. Hayes, E. Numkam, L. Grűner-Nielsen, B. Pálsdóttir, R. Phelan, B. Kelly, M. Becker, N. Macsuibhne, J. Zhao, F. C. G. Gunning, A. D. Ellis, P. Petropoulos, S. U. Alam, and D. J. Richardson, European Conference and Exhibition on Optical Communication (Optical Society of America, 2012), paper Th.3.A.5.
2587
7. S. W. Henderson, P. J. M. Suni, C. P. Hale, S. M. Hannon, J. R. Magee, D. L. Bruns, and E. H. Yuen, IEEE Trans. Geosci. Remote Sens. 31, 4 (1993). 8. T. J. Carrig, A. K. Hankla, G. J. Wagner, C. B. Rawle, and I. T. Mckinnie, Proc. SPIE 4723, 147 (2002). 9. A. Schweinsberg, Z. Shi, J. E. Vornehm, and R. W. Boyd, Opt. Express 19, 15760 (2011). 10. A. Schweinsberg, Z. Shi, J. E. Vornehm, and R. W. Boyd, Opt. Express 37, 329 (2012). 11. G. D. Goodno, L. D. Book, and J. E. Rothenberg, Opt. Lett. 34, 1204 (2009). 12. X. Wang, P. Zhou, H. Zhang, X. Wang, H. Xiao, and Z. Liu, Opt. Lett. 39, 4329 (2014). 13. Y. A. Vlasov, M. O. Boyle, H. F. Hamann, and S. J. Mcnab, Nature 438, 65 (2005). 14. M. S. Bigelow, N. N. Lepeshkin, and R. W. Boyd, Science 301, 200 (2003). 15. R. S. Tucker, P. Ku, and C. J. Chang-Hasnain, J. Lightwave Technol. 23, 4046 (2005). 16. K. Y. Song, M. G. Herráez, and L. Thévenaz, Opt. Express 13, 82 (2005). 17. Y. Okawachi, M. S. Bigelow, J. E. Sharping, Z. Zhu, A. Schweinsberg, D. J. Gauthier, R. W. Boyd, and A. L. Gaeta, Phys. Rev. Lett. 94, 153902 (2005). 18. C. Jáuregui, P. Petropoulos, and D. J. Richardson, Opt. Express 15, 5126 (2007). 19. T. Schneider, M. Junker, and K. Lauterbach, Opt. Lett. 32, 220 (2007). 20. G. K. W. Gan, Y. G. Shee, K. S. Yeo, G. A. Madhiraji, F. R. M. Adikan, and M. A. Mahdi, Opt. Lett. 39, 5118 (2014). 21. A. Zadok, A. Eyal, and M. Tur, Appl. Opt. 50, E38 (2011). 22. Z. Shi, A. Schweinsberg, Jr., J. E. Vornehm, M. A. M. Gámez, and R. W. Boyd, Phys. Lett. A 374, 4071 (2010). 23. L. Thévenaz, Nat. Photonics 2, 474 (2008). 24. T. Schneider, Opt. Lett. 33, 1398 (2008). 25. J. Geng, Q. Wang, T. Luo, S. Jiang, and F. Amzajerdian, Opt. Lett. 34, 3493 (2009). 26. G. D. Goodno, L. D. Book, J. E. Rothenberg, M. E. Weber, and S. B. Weiss, Opt. Eng. 50, 111608 (2011). 27. R. G. Harrison, J. S. Uppal, A. Johnstone, and J. V. Moloney, Phys. Rev. Lett. 65, 167 (1990). 28. L. Xing, L. Zhan, S. Luo, and Y. Xia, IEEE J. Quantum Electron. 44, 1133 (2008). 29. G. P. Agrawal, Nonlinear Fiber Optics, 4th ed. (Academic, 2007). 30. V. P. Kalosha, L. Chen, and X. Bao, Opt. Express 14, 12693 (2006). 31. Z. Zhu, D. J. Gauthier, Y. Okawachi, J. E. Sharping, A. L. Gaeta, R. W. Boyd, and A. E. Willner, J. Opt. Soc. Am. B 22, 2378 (2005).
OPTICS LETTERS / Vol. 40, No. 11 / June 1, 2015
Tunable slow light via stimulated Brillouin scattering at 2 μm based on Tm-doped fiber amplifiers Xiong Wang, Pu Zhou,* Xiaolin Wang, Hu Xiao, and Zejin Liu College of Optoelectronic Science and Engineering, National University of Defense Technology, Changsha 410073, China *Corresponding author: [email protected] Received March 4, 2015; revised May 4, 2015; accepted May 11, 2015; posted May 11, 2015 (Doc. ID 235427); published May 28, 2015 We present a slow light system based on stimulated Brillouin scattering (SBS) at 2 μm. A single-frequency fiber laser with Tm-doped fiber amplifiers was used to generate the SBS signal laser and the Brillouin pump light at 1.971 μm. The maximum delay time reaches 16 ns for pulses with 43-ns width, and the pulse width is broadened to 56.4 ns. The maximum delay time for 57-ns pulses reaches 33.4 ns, and the pulse width is broadened to 77.6 ns. The relative delays are 0.37 and 0.58 for 43 and 57 ns pulses, respectively. This is the first demonstration, as far as we know, on a slow light system at 2 μm, which may be substantial for future optical communications and LIDAR systems employing laser sources near 2-μm band. © 2015 Optical Society of America OCIS codes: (140.3070) Infrared and far-infrared lasers; (140.3510) Lasers, fiber; (190.4370) Nonlinear optics, fibers; (290.5900) Scattering, stimulated Brillouin. http://dx.doi.org/10.1364/OL.40.002584
The broadest gain spectrum around ∼1700–2200 nm among rare-earth-doped fiber lasers enables Tm-doped fiber lasers (TDFLs) significant laser sources in various applications in recent decades [1–3]. One of the valuable potential applications is optical communications [4,5]. The exponentially increasing volume of internet traffic drives the current networks using 1.55-μm fiber lasers to their capability limits to a large extent. Thus the TDFLs with much broader emission band may become the prospective laser sources for future optical communications, and the low loss of hollow-core photonic bandgap fibers near 2 μm further enhances the availability and practicability of prospective networks operating near 2 μm [6]. Another meaningful application is the light radar (LIDAR) [7–10]. The emission band of TDFLs covers the transmission window of atmosphere around 2 μm, which also contains the absorption lines of various important atmosphere gases. So TDFLs are admirable sources for LIDAR in remote sensing and detecting. Thus lasers sources based on Tm-doped fiber have been investigated intensively in the last decades [11,12]. It should be noted that the two applications above have one resemblance, i.e., the control of group delay of light. In fact, all-fiber-based buffer is one crucial topic for optical communications, and the tunable delay of optical signals in the futuristic fiber networks is an interesting approach for improving the data traffic, as well as enhancing the capability of optical signal processing [13–15]. In addition, the tunable delay of optical pulses is a favorable method to effectively and rapidly correct the optical group delay mismatches among all the apertures in a LIDAR system [9,10]. Tunable delay of optical pulses has been demonstrated and investigated vastly in recent years, and one of the most convenient methods is to realize fast and slow light via stimulated Brillouin scattering (SBS) [16–24]. The SBS effect can cause large normal dispersion and increase the group index in a narrow spectral range, which will reduce the group velocity of optical pulses in fibers [17,23]. The tunable delay of optical fiber pulses based on SBS possesses some excellent advantages, such as low operating pump power and linearly tunable delay [16,21,22]. It is also convenient for 0146-9592/15/112584-04$15.00/0
the delay system being integrated into the optical networks nearly perfectly [13–15]. Thus, slow light based on SBS stimulated tremendous research passion, and the objects are all mainly focused on fiber lasers near 1.55 μm, since it is the current practical wavelength for optical communications and LIDAR. However, as aforementioned, TDFLs near 2 μm may provide an advanced optical communication method and improve the performance of LIDAR. Although previous investigations state that the slow light can be generated in the current communication window (∼1.55 μm), there has been no demonstration on slow light at 2 μm so far, to the best of our knowledge. In this Letter, we experimentally demonstrated the tunable slow light based on SBS at 2 μm. A single-frequency (SF) fiber laser was used as the seed laser to generate the pump laser and SBS signal. A piece of 450-m single mode fiber (SMF) was employed as the delay fiber. The delay time can be tuned by adjusting the pump power. The maximum delay time can reach 33.4 ns when using pulses with width of 57 ns, and the delay time can reach 16 ns when pulses with 43-ns width were used. This is the first demonstration, as far as we know, on slow light based on SBS at 2 μm. The experimental setup is shown in Fig. 1. A SF fiber laser at 1971 nm based on ultra-short cavity with a linewidth of less than 100 kHz and an output power of ∼40 mW was used as the seed laser [25]. A 50/50 fiber coupler split the laser into two parts. One part of the laser was amplified by a home-made single-mode Tm-doped
Fig. 1. Experimental setup of the slow light system based on SBS using TDFAs. © 2015 Optical Society of America
June 1, 2015 / Vol. 40, No. 11 / OPTICS LETTERS
fiber amplifier (TDFA) to serve as the Brillouin pump (BP) light. The power of the BP light can reach ∼500 mW with a continual tuning range by adjusting the current of the pump laser diodes. The other part of the laser was amplified by a similar single-mode TDFA, then the output laser was launched into a length of 450-m SMF (SMF-28 fiber, SMF 1) via a three-port circulator. The SBS light was generated in the SMF 1 and backward propagated into the circulator. The continuous wave SBS light was then modulated to pulses by an acousto-optic modulator (AOM), which had a rise time of ∼70 ns. The BP light and the SBS light encountered each other in another piece of 450-m SMF (SMF-28 fiber, SMF 2). A polarizationinsensitive isolator (ISO) was inserted between the SMF 2 and the AOM to stop the residual BP light. The SBS pulses were delayed in the SMF 2 and launched out through the second circulator. The output pulses were measured by a high-speed photodetector and a digital oscilloscope with bandwidth of 1 GHz. A unique advantage of this system is that the frequency difference of the BP light and SBS light is essentially guaranteed to be near the Brillouin frequency shift with only one seed laser. The spectra of the BP light and SBS light were monitored by an optical spectrum analyzer (OSA) with resolution of 0.05 nm, as shown in Fig. 2. The Brillouin frequency shift at 2 μm is conformed in the results. The repetition rate of the AOM was set to be 100 kHz, and the widths of the SBS pulse were set to be 44 and 60 ns, respectively. The SBS signal’s power was maintained at ∼150 μW. When the power of the BP light was increased gradually, the SBS pulse was delayed, as shown in Fig. 3. The maximum delay time reaches 11.2 ns for pulses with a width of 44 ns, and the pulse width is slightly broadened to 46.8 ns, corresponding to a pulse-broadening factor of 1.07. The maximum delay time reaches 20 ns for pulses with a width of 60 ns, and the pulse width is slightly broadened to 64.8 ns, corresponding to a pulse
Fig. 2. Spectra of the BP light and SBS light of the slow light system.
Fig. 3. Observation of slow light via SBS at 2 μm. (a) Input pulse width is 44 ns and (b) input pulse width is 60 ns.
2585
broadening factor of 1.08. Both the broadening factors are rather low, thus the pulse shapes are well maintained during the delay process. Figure 4 shows the delay time as a function of the Brillouin gain. The delay time via SBS can be expressed as ΔT ≈ G∕ΓB , where ΓB is the Brillouin linewidth. The Brillouin gain factor is G g0 I p L, where g0 is the line center gain factor, L is the fiber length, I p is the BP power, and the exponential of G is the small-signal gain [17]. Thus the delay time can be continually and linearly tuned via adjusting the BP power. The maximum relative delay is ΔT∕tp 0.33 and 0.25 for pulse with width of 60 and 44 ns, respectively, where tp represents the pulse width. According to the expression ΔT ≈ G∕ΓB , the delay time at 2 μm will be higher than the delay time at 1.55 μm with a fixed Brillouin gain, due to the narrower Brillouin gain linewidth (∼10 MHz to ∼30 MHz) [17,26]. However, the relatively high background loss of current commercially available SMF-28 fiber at 2 μm may reduce the Brillouin gain and delay time, and employing of improved fibers with lower loss at 2 μm will mitigate the limit. In fact, the curves in Fig. 4 show uneven increase of the delay time versus the Brillouin gain. We attribute this deficiency to the instability of the SBS light generated from the SMF 1, since the randomness character of SBS determines that the SBS power may fluctuated in a certain extent [27]. A natural method to solve this problem is to employ Brillouin amplifier based on a TDFA [28]. However, the additional amplified spontaneous emission (ASE) noise introduced by utilizing a TDFA will degrade the signal-to-noise ratio (SNR) of the slow light system. So we improved the system by employing Brillouin fiber laser (BFL) configuration, as shown in Fig. 5. A 10/90 fiber coupler was inserted before the AOM, then 10% of the SBS light was launched back into the SMF 1. Thus a BFL cavity was built up by a circulator,
Fig. 4. Delay time versus the Brillouin gain for different pulse widths.
Fig. 5. BFL.
Experimental setup of the slow light system based on
2586
OPTICS LETTERS / Vol. 40, No. 11 / June 1, 2015
a coupler, and a piece of SMF. The 90% port of the coupler launched the SBS laser into the AOM to generate the pulsed signal laser. Using this improved configuration, the output of the SBS pulsed laser is much stabler, and the SNR of the SBS pulse is improved, so the performance of the slow light system is enhanced. Furthermore, the output power of the BFL is higher than that of SBS light without the coupler, which means the pump power of the TDFA 1 can be also reduced to make the system more feasible in practical applications. The results of slow light using the modified configuration are shown in Fig. 6. The power of the SBS pulses was adjusted to be the same with that of previous experiments by turning down the pump power of the TDFA 1. The initial pulse widths are slightly narrower than that of the previous results due to the more stable and more precise measurement. The maximum delay time for pulses with a width of 43 ns reaches 16 ns, but the pulse width is also broadened to 56.4 ns, corresponding to a broadening factor of 1.31. The maximum delay time for pulses with a width of 57 ns reaches 33.4 ns, and the pulse width is broadened to 77.6 ns, corresponding to a broadening factor of 1.35. The relative delays are 0.37 and 0.58 for 43 and 57 ns pulses, respectively. From the linear fit of the experimental data curves in Fig. 6, we conclude that the delay efficiency is 5 and 11.7 ns/dB for 43 and 57 ns pulses, respectively. The experimental curves in Fig. 6 show slight nonlinearity and fluctuations. One reason may be that the Brilloiin gain linewidth at 2 μm is narrower than that at 1.5 μm, so the linear tuning range for time delay will be shorter when the pulse width is relatively short [29]. This can be conquered via broadening the BP wave power spectral density, by which shorter pulses can be delayed effectively and linearly [23]. Since the polarization states will also influence the delay process, the environment vibration may introduce fluctuations, which can be mitigated by enhancing the stability of the system and the polarization states. Actually, the first obstacle one must face is the considerable background loss of silica fiber when generating SBS slow light at 2 μm (about 20 dB/km with common telecommunication fibers). In the previous theoretical demonstrations, the loss coefficient of the fiber is either neglected or essentially tiny (∼0.2 dB∕km) [16,30,31], since the operation wavelength is around ∼1.5 μm. Thus the delay of SBS slow light almost starts when the Brillouin gain is slightly more than zero theoretically.
Fig. 6. Experimental and simulated delay time data versus the Brillouin gain data for different pulse widths with improved configuration. (a) Input pulse width is 43 ns and (b) input pulse width is 57 ns.
However, the loss of fiber plays an obvious role on the “starting point” of the delay time versus Brillouin gain, i.e., the delay begins when the gain is ∼6 [17] or ∼5 dB [20], though the loss coefficient is rather low. As for slow light at 2 μm, the seriously increasing loss coefficient makes the “starting point” much higher compared with that of slow light at 1.5 μm band, as shown in Figs. 4 and 6. One should also note that the AOM inevitably introduces frequency shift to the generated signal pulses (80 MHz in this experiment). The reason we use AOM instead of electro-optic modulator (EOM) is that the EOM at 2 μm in our lab cannot provide perfect intense modulation with high SNR for the time being. Thus the frequency of the SBS pulses is 80 MHz away from the center of Brillouin resonance, which means the delay results can be further improved when making the frequency of the pulses in the center of the resonance. Furthermore, the detune of the SBS pulse from the resonance center indicates that the actual Brillouin gain should be modified, provided that the practical gain factor is practically measured. Furthermore, the difference of pulse widths in our experiment is small, owing to the limited bandwidth of the AOM for generating narrower pulse width. Thus the curves of delay time with two different pulse widths even overlap to a certain extent. Due to the high loss coefficient of fiber at 2 μm and the large detune introduced by the AOM, the gain is far below the saturation value with SBS signal’s power of ∼150 μW. Actually, the saturation value may be even more than 30 dB at 1.5 μm band [20]. Thus the saturation effects have not occurred seriously in our experiment yet. We simulated the slow light at 2 μm based on threewave coupled equations [29], and the results are also shown in Fig. 6. The simulated maximum delay time for 43 ns pulses is 15.8 ns, and that for 57 ns pulses is 33.7 ns. The simulated broadening factor when the gain reaches the maximum value is 1.29 and 1.37 for 43 and 57 ns pulses, respectively. The simulated curves of delay time data versus gain data have the similar increasing trend with those of the experimental results. From the above comparison, one can see that the simulated and experimental results matched well with each other in a certain extent, which indicates that realization of slow light at 2 μm is feasible experimentally and theoretically. The simulation does not include the processes of selfphase modulation and cross-phase modulation for simplification. By the way, the spectral part of the signal pulses near the resonance center of SBS can be amplified when the pump power is intense enough, which will result in the decrease of the frequency detune between the pump and signal lights. Thus, the delay time might be even higher when the pump power is increasing but well below the saturation threshold. We also note that when the Brillouin gain is lower than 19 dB, the increase of the delay time is slightly nonlinear, but when the Brillouin gain is higher than 19 dB, the delay time increases more linearly. Meanwhile, the broadening factor is suddenly increased from 1.15 to 1.35 when the Brillouin gain surpasses 19 dB. A detailed simulation will be performed to further investigate this phenomenon. However, the experimental and simulated results demonstrate the first realization of slow light at 2-μm band,
June 1, 2015 / Vol. 40, No. 11 / OPTICS LETTERS
which may be rather meaningful for the future optical networks and LIDAR systems employing 2-μm lasers. In conclusion, we demonstrate the first tunable slow light system at 2 μm based on TDFAs. BFL was used to generate SBS signal laser with the Brillouin shfit compared with that of the BP laser. The maximum delay time of 43-ns pulses reaches 16 ns with a broadening factor of 1.31, and the maximum delay time of 57-ns pulses reaches 33.4 ns with a broadening factor of 1.35. The performance of the slow light system can be further improved by utilizing BP light with broader power spectral density. This experiment shows the availability of slow light laser system at 2-μm band, which will be futuristic and significant for the prospective optical communications and LIDAR systems. This work was supported by the Graduate Student Innovation Foundation of National University of Defense Technology (Grant No. B130704), National Natural Science Foundation of China (Grant No. 11274386), and Program for New Century Excellent Talents in University and Hunan Provincial Innovation Foundation for Postgraduate. We thank Haibin Lv for his helpful discussion and support when performing the simulation. References 1. S. D. Jackson, Nat. Photonics 6, 423 (2012). 2. J. Geng, Q. Wang, and S. Jiang, Proc. SPIE 8164, 816409 (2011). 3. P. F. Moulton, G. A. Rines, E. V. Slobodtchikov, K. F. Wall, G. Frith, B. Samson, and A. L. G. Carter, IEEE J. Sel. Top. Quantum Electron. 15, 85 (2009). 4. S. U. Alam, Z. Li, J. M. O. Daniel, Y. Jung, A. M. Heidt, and D. J. Richardson, OptoElectronics and Communications Conference Held Jointly with 2013 International Conference on Photonics in Switching (OECC/PS) (Optical Society of America, 2013), paper WS1_3. 5. Z. Li, A. M. Heidt, N. Simakov, Y. Jung, J. M. O. Daniel, S. U. Alam, and D. J. Richardson, Opt. Express 21, 26450 (2013). 6. M. N. Petrovich, F. Poletti, J. P. Wooler, A. M. Heidt, N. K. Baddela, Z. Li, D. R. Gray, R. Slavík, F. Parmigiani, N. V. Wheeler, J. R. Hayes, E. Numkam, L. Grűner-Nielsen, B. Pálsdóttir, R. Phelan, B. Kelly, M. Becker, N. Macsuibhne, J. Zhao, F. C. G. Gunning, A. D. Ellis, P. Petropoulos, S. U. Alam, and D. J. Richardson, European Conference and Exhibition on Optical Communication (Optical Society of America, 2012), paper Th.3.A.5.
2587
7. S. W. Henderson, P. J. M. Suni, C. P. Hale, S. M. Hannon, J. R. Magee, D. L. Bruns, and E. H. Yuen, IEEE Trans. Geosci. Remote Sens. 31, 4 (1993). 8. T. J. Carrig, A. K. Hankla, G. J. Wagner, C. B. Rawle, and I. T. Mckinnie, Proc. SPIE 4723, 147 (2002). 9. A. Schweinsberg, Z. Shi, J. E. Vornehm, and R. W. Boyd, Opt. Express 19, 15760 (2011). 10. A. Schweinsberg, Z. Shi, J. E. Vornehm, and R. W. Boyd, Opt. Express 37, 329 (2012). 11. G. D. Goodno, L. D. Book, and J. E. Rothenberg, Opt. Lett. 34, 1204 (2009). 12. X. Wang, P. Zhou, H. Zhang, X. Wang, H. Xiao, and Z. Liu, Opt. Lett. 39, 4329 (2014). 13. Y. A. Vlasov, M. O. Boyle, H. F. Hamann, and S. J. Mcnab, Nature 438, 65 (2005). 14. M. S. Bigelow, N. N. Lepeshkin, and R. W. Boyd, Science 301, 200 (2003). 15. R. S. Tucker, P. Ku, and C. J. Chang-Hasnain, J. Lightwave Technol. 23, 4046 (2005). 16. K. Y. Song, M. G. Herráez, and L. Thévenaz, Opt. Express 13, 82 (2005). 17. Y. Okawachi, M. S. Bigelow, J. E. Sharping, Z. Zhu, A. Schweinsberg, D. J. Gauthier, R. W. Boyd, and A. L. Gaeta, Phys. Rev. Lett. 94, 153902 (2005). 18. C. Jáuregui, P. Petropoulos, and D. J. Richardson, Opt. Express 15, 5126 (2007). 19. T. Schneider, M. Junker, and K. Lauterbach, Opt. Lett. 32, 220 (2007). 20. G. K. W. Gan, Y. G. Shee, K. S. Yeo, G. A. Madhiraji, F. R. M. Adikan, and M. A. Mahdi, Opt. Lett. 39, 5118 (2014). 21. A. Zadok, A. Eyal, and M. Tur, Appl. Opt. 50, E38 (2011). 22. Z. Shi, A. Schweinsberg, Jr., J. E. Vornehm, M. A. M. Gámez, and R. W. Boyd, Phys. Lett. A 374, 4071 (2010). 23. L. Thévenaz, Nat. Photonics 2, 474 (2008). 24. T. Schneider, Opt. Lett. 33, 1398 (2008). 25. J. Geng, Q. Wang, T. Luo, S. Jiang, and F. Amzajerdian, Opt. Lett. 34, 3493 (2009). 26. G. D. Goodno, L. D. Book, J. E. Rothenberg, M. E. Weber, and S. B. Weiss, Opt. Eng. 50, 111608 (2011). 27. R. G. Harrison, J. S. Uppal, A. Johnstone, and J. V. Moloney, Phys. Rev. Lett. 65, 167 (1990). 28. L. Xing, L. Zhan, S. Luo, and Y. Xia, IEEE J. Quantum Electron. 44, 1133 (2008). 29. G. P. Agrawal, Nonlinear Fiber Optics, 4th ed. (Academic, 2007). 30. V. P. Kalosha, L. Chen, and X. Bao, Opt. Express 14, 12693 (2006). 31. Z. Zhu, D. J. Gauthier, Y. Okawachi, J. E. Sharping, A. L. Gaeta, R. W. Boyd, and A. E. Willner, J. Opt. Soc. Am. B 22, 2378 (2005).
