Open Access
Add to BibSonomyAbstract
We demonstrate a multiwavelength laser at 2 µm based on a hybrid gain scheme consisting of a Brillouin gain medium and a thulium-doped fiber. The laser has switchable frequency spacing, corresponding to the single and double Brillouin frequency shifts. In the 20 dB bandwidth, seven lasing channels with a frequency spacing of 0.1 nm (7.62 GHz) and eleven channels with a double-spacing of 0.2 nm (15.24 GHz) are obtained. A wavelength tunability of 1.3 nm is realized for both laser configurations by shifting the pump wavelength. Strong four wave mixing is observed in the double-spacing laser resulting in an improved performance: larger number of channels and better temporal stability.
© 2014 Optical Society of America
1. Introduction
Multiwavelength fiber lasers in 1550 nm wavelength regime have been applied to optical instruments testing, fiber sensing, microwave generation, and fiber-optic wavelength division multiplexing (WDM) in communication networks [1–4]. In recent years, the 2 µm wavelength regime has attracted significant attention from the community. The eye-safe property and strong atmospheric absorption for certain gases, like carbon dioxide, water, and ammonia, ensure its potential in atmospheric spectroscopy and coherent LIDAR applications [5,6]. High-speed WDM data transmission at 2 µm has also been studied as a solution to increase the capacity of fiber-optic communication [7,8]. These applications suggest future demand for 2 µm multiwavelength lasers.
A few schemes have been proposed to achieve multiwavelength laser sources at 2 µm wavelength range. Zhao et.al demonstrated a thulium-doped fiber laser (TDFL) based on in-cavity comb filters [9]. Wang et.al presented a multiwavelength TDFL based on nonlinear polarization rotation (NPR) and four wave mixing (FWM) [10]. Peng et.al reported a TDFL based on a nonlinear amplifier loop mirror (NALM) [11]. However, in order to suppress the mode competition in a broadband gain rare-earth doped-fiber, the realizations of stable outputs are highly dependent on precise polarization control, which induces system complexity and limits its potential in practical uses.
On the other hand, cascaded stimulated Brillouin scattering (SBS) is a nonlinear effect that is promising for 2 µm multiwavelength laser realization [12,13] for its ability of the generation of multiple narrow gain bands (Stokes waves) with an equal spacing. Once generated, Stokes waves can be amplified in a thulium-doped fiber, increasing the cavity gain, helping to overcome the high loss of silica fibers at 2 µm. By introducing these two gain mechanism, a hybrid-gain multiwavelength laser could be realized [14,15]. The Brillouin-thulium hybrid multiwavelength laser was reported recently by Wang et.al, using a figure-of-eight cavity design assisted by an external optical amplifier [12]. Strong 793 nm pumps were used to achieve a fixed 1.97 µm Brillouin pump (BP) with a power of 3.2 W. Five Stokes waves with a wavelength spacing of 0.105 nm and a power fluctuation of 0.5 dB were generated. However, this wavelength spacing is too small for WDM standards and likely to cause difficulties in the signal de-multiplexing. One of the better options would be to increase the frequency spacing by choosing only even components with double Brillouin frequency (~15 GHz / 0.2 nm) spacing. Such lasing channels could be easily split with existing 2 µm fiber Bragg grating (FBG) technology. A larger spacing is also more suitable for WDM applications, close to the ultra-dense WDM channel spacing of 12.5 GHz [16,17].
In this paper, we demonstrate a multiwavelength laser at 2 µm based on a hybrid gain scheme consisting of a Brillouin gain medium and a thulium-doped fiber. The laser has switchable frequency spacing, corresponding to a single and double Brillouin frequency shifts. In the 20 dB bandwidth, seven lasing channels with a frequency spacing of 0.1 nm (7.62 GHz) and eleven channels with a double-spacing of 0.2 nm (15.24 GHz) are obtained. A wavelength tunability of 1.3 nm is realized for both laser configurations by shifting the pump wavelength. Strong four wave mixing is observed in the double-spacing laser resulting in an improved performance: larger number of channels and better temporal stability. We believe the laser’s potential in application like WDM fiber-optic communications can be extended with these characteristics in conjunction with the switchable wavelength spacing.
2. Experimental setup
The experimental configuration of the MW-BTFL with a single-Brillouin-frequency spacing is shown in Fig. 1(a). A pre-amplified continuous-wave (CW) 2 µm diode laser (Eblana EP2000-DM-B) is followed by an isolator (ISO) to prevent the damage from the back-reflected light. 30% of the 2 µm power is launched into the main cavity by a 70/30 optical coupler (OC), serving as the Brillouin pump (BP) and the signal laser. A thulium-doped fiber amplifier (TDFA), which consists of a 1560 nm laser pump, a 1560/1950 wavelength division multiplexer (WDM), and a piece of thulium-doped fiber (TmDF, TmDF 200, OFS), is inserted in the cavity to generate enough gain to overcome the cavity loss [18]. The wavelength of 1560 nm is chosen specifically to achieve higher amplification efficiency in the TDFA. Through port 1 to port 2 of the circulator (CIR #1), the amplified BP is sent to a 1 km long highly nonlinear fiber (HNLF, OFS), which serves as the Brillouin gain medium. The HNLF is chosen over single mode fiber (SMF) [12] for its smaller mode effective area, thus maximizing the Brillouin gain GB ~(Aeff)-1 [19]. The loss of the HNLF at 2 µm was measured to be approximately 0.01 dB/m. The effective length Leff is calculated to be 0.59 km [20]. A polarization controller (PC) is located before the HNLF to adjust the polarization state of the light sent into the HNLF, optimizing the laser performance. Another circulator (CIR #2) is placed at the end of the HNLF as a reflecting mirror by using port 2 as the input port, and connecting port 1 and port 3, which ensures a closed-loop cavity. Port 3 of CIR #1 is connected to the 70/30 coupler, through which 70% of the power in the cavity will be kept as in-cavity feedback, while the remaining 30% will be split out of the cavity for spectrum and power measurement. When the hybrid gain provided by SBS effect and the TDFA compensates the cavity loss, the 1st Stokes wave with a Brillouin frequency shift ΩB will be excited and travel clockwise in the cavity. When the power of the 1st Stokes wave exceeds the lasing threshold, the 2nd Stokes wave will start oscillating, and so forth, to generate higher order Stokes waves.
Fig. 1 Configuration of the multiwavelength Brillouin/thulium fiber laser (MW-BTFL) with switchable (a) single-Brillouin-frequency spacing, and (b) double-Brillouin-frequency spacing outputs, respectively.
The setup for MW-BTFL with a double-Brillouin-frequency spacing Ω = 2ΩB is depicted in Fig. 1(b). It can be simply switched from the single-Brillouin-frequency spacing MW-BTFL scheme by rearranging the connection ports, which could be possibly done by an automatic optical switch. The BP goes into the HNLF and generates the 1st Stokes wave, which circulates clockwise inside the double Brillouin frequency shifter (DBFS), and is out coupled into the main laser cavity. The 1st Stokes wave excites the 2nd Stokes wave, which travels counterclockwise in the DBFS. Then the 3rd and 4th Stokes wave will be generated in a similar way. The cascaded process continues as the 1560 nm light pump power is increased. However, the odd-orders of Stokes waves are confined inside the DBFS and only even-orders of Stokes waves are allowed to oscillate in the main cavity.
3. Results and discussion
Firstly, we measured the output spectrum of the single-Brillouin-frequency spacing MW-BTFL using an optical spectrum analyzer (OSA, Yokogawa AQ6375). The resolution of the OSA was set to the highest of 0.05 nm. During the measurement, 30% of the 2 µm power after the isolator was coupled into the cavity and measured as 13.8 mW, using a power meter (Coherent FieldMax II), and working as the BP and the signal laser. As the 1560 nm pump power was increased, the cascaded SBS effect occurred in the HNLF and excited multiple Stokes waves. A group of spectra recorded at 6 different 1560 nm pump power levels, from 0.5 W to 2.3 W, is depicted in Fig. 2. The number of Stokes waves increased with the rising 1560 nm pump power, and reached a maximum of seven. As a result of the employment of HNLF instead of SMF as in [12], more lasing wavelengths were obtained at lower pump power. The output signal was measured from the 70/30 coupler, and hence it consisted of 30% of the intracavity power and the 70% of the pump power, that did not enter the cavity. As a result, a strong peak at the wavelength of 1952.7 nm (the BP wavelength) can be seen in Fig. 2. To measure the output power of the laser, we subtracted 32.2 mW (the 70% of the pump power). The rest of the signal corresponds to the power carried by the multiple Stokes waves, and it is plotted in the inset of Fig. 2.
Fig. 2 Output spectrum of the single-Brillouin-frequency spacing MW-BTFL with different 1560 nm pump powers. Inset shows the figure of Stokes power versus 1560 pump power, with linear fitting line (dashed).
Considering the resolution limit of the OSA, we sent the output laser into a high-speed photodiode (PD, Photonic Solutions, ET-5000AF) followed by a radio frequency analyzer (RFA, Agilent E4448A), to measure the Brillouin frequency shift precisely. The resolution of the RFA was set as 3 MHz. The radio frequency (RF) beat between different lasing wavelengths is shown in Fig. 3. The first peak with a center frequency of 7.62 GHz indicates the superposition of the beat signals between the BP and 1st Stokes, 1st Stokes and 2nd Stokes, 2nd Stokes and 3rd Stokes, etc. The second peak with a center frequency of 15.24 GHz indicates the superposition of the beat signals between BP and 2nd Stokes, 1st Stokes and 3rd Stokes, 2nd Stokes and 4th Stokes, etc. The second peak is actually locates outside the rated bandwidth of the PD, which is 9 GHz. The reduced responsivity of the PD explains the large amplitude difference (~30 dB) between these two peaks.
Temporal stability is an important characteristic of a laser. To evaluate it, the laser spectrum was obtained and the power was measured every 10 minutes during a total time span of an hour, with a BP power of 13.8 mW and a 1560 nm pump power of 2.3 W. The peak power fluctuations of the first seven Stokes waves are depicted in Fig. 4. The first 5 orders (1st to 5th) Stokes waves had fluctuations of less than 0.5 dB, while the 6th and 7th order Stokes waves had 0.78 dB and 1 dB fluctuation, respectively. This could be explained by assuming that the lower orders Stokes waves had reached saturation and became stable. However, the power of the higher orders of Stokes waves could be easily influenced by instability in pump power, which led to stronger power fluctuations.
Fig. 4 Temporal stability measurement for the single-Brillouin-frequency spacing MW-BTFL during an hour.
The MW-BTFL was then switched to the double-Brillouin-frequency spacing configuration. The optical and RF spectra and temporal stability were measured and calculated using the same methods and equipment as before. A group of output spectra recorded at seven different levels of the 1560 nm pump, from 0.4 W to 2.3 W, is depicted in Fig. 5. The inset to Fig. 5 shows the power of the Stokes waves versus the power of the 1560 nm pump. Several distinctions were observed compared to the single-Brillouin-frequency MW-BTFL presented before, due to the different designs. For instance, in the double-Brillouin-frequency configuration, the 1st Stokes is circulating only inside the DBFS cavity and does not reach the TmDF. With no amplification inside the TmDF, the 1st Stokes experiences the loss of HNLF and therefore is much weaker compared to the pump. The 2nd Stokes wave reaches the outer-loop where it is amplified in the TmDF. But since it is excited by the weak 1st Stokes wave its overall power is lower than that of the pump. Following this argument, it is not surprising that the power levels of the Stokes peaks in the double-spacing regime were generally lower than those in the single-spacing regime.
Fig. 5 Output spectrum of the double-Brillouin-frequency spacing MW-BTFL with different 1560 nm pump powers. Inset shows the figure of Stokes power versus 1560 pump power, with linear fitting line (dashed).
Additionally, the output spectrum of the double-spacing Brillouin laser shown in Fig. 5 has a symmetric shape (if one ignores the strong peak at the pump wavelength, which is measurement artifact as discussed before). One possible explanation of the symmetry in the laser output can be related to the effect of FWM occurring in the HNLF. This explanation is supported by the fact that the SBS process can only generate new frequencies on the long-wavelength side from the BP and thus cannot lead to a symmetric spectral output. Thanks to the small frequency difference between the spectral lines, the waves are close to being phase matched, and the FWM gain is almost at its maximum. With a nonlinear coefficient γ = 9.14 kW−1·m−1 and an effective mode area of 11.6 µm2, the value of the peak parametric gain gp = γ · Aeff is close to 1.3 × 10−13 m/W. This is 2 orders of magnitude lower than the SBS gain of 2.89 × 10−11 m/W. Despite this, powers of the 2nd, 4th, 6th and higher Stokes waves are much lower than that of the BP. Since FWM is a nonlinear process and it is power dependent, the efficiency of the degenerate FWM process initiated by the strong BP can be similar or even greater than that of the SBS processes pumped by the weak high-order Stokes waves. The influence of FWM was also pronounced in the number of output spectral lines. In the double-Brillouin-frequency configuration the number of lines was significantly expanded to eleven compared to seven in the single-Brillouin-frequency spacing configuration.
In the single-spacing output spectrum shown in Fig. 2, all Stokes waves are much stronger and comparable with the BP. This is the result of the laser cavity geometry, namely the fact that each Stokes wave circulates inside the cavity loop and is eventually amplified in the TmDF. Since the gain coefficients of FWM and SBS processes differ by two orders of magnitude, the SBS effect plays a dominant role in this configuration. The presence of FWM effect is confirmed by a small ripple on the short-wavelength side from the BP, which could be observed in the spectrum floor in Fig. 2.
The RF spectrum and the temporal stability were also measured for the double-Brillouin-frequency spacing MW-BTFL and presented in Fig. 6 and Fig. 7, respectively. In Fig. 6, we observed a main peak with a center frequency of 15.24 GHz, which corresponds to the superposition of the beat signals between adjacent lasing wavelengths (e.g. BP and 2nd Stokes, 2nd Stokes and 4th Stokes, etc.), and a weak peak with the center frequency of 7.62 GHz. Note that the peak at 7.62 GHz is much weaker (~25 dB) than the one in Fig. 3. We believe the existence of this weak peak is because of the slight odd-orders Stokes waves’ leakage from the circulators, which was reported in earlier research [21]. We found that the double-Brillouin-frequency spacing MW-BTFL manifested in excellent temporal stability, which is presented in Fig. 7. The maximum peak power fluctuation of the first seven lasing wavelengths was 0.3 dB, and the minimum was only 0.04 dB. We believe that the FWM is the major reason for this performance improvement since it has been proved useful in suppressing the mode competition and enhancing the laser stability [22,23].
Fig. 7 Temporal stability measurement for the double-Brillouin-frequency spacing MW-BTFL during an hour.
In applications such as fiber-optic communication and atmospheric spectroscopy, wavelength tunability provides system flexibility. By tuning the temperature of the 2 µm laser diode, the emitting wavelength of the diode shifts, which leads to a wavelength shift of the whole output spectrum of the MW-BTFL. The wavelength tuning process was realized for both the single- and double-Brillouin-frequency spacing MW-BTFLs. The tuning range was about 1.3 nm, from 1952.7 nm to 1954.0 nm, as shown in Fig. 8.
Fig. 8 Wavelength tuning for the (a) single-Brillouin-frequency spacing, and (b) double-Brillouin-frequency spacing MW-BTFLs.
4. Conclusion
In this paper, we propose a multiwavelength laser at 2 µm based on a hybrid gain scheme consisting of a Brillouin gain medium and a thulium-doped fiber. The laser has switchable frequency spacing, corresponding to a single and double Brillouin frequency shifts. In the 20 dB bandwidth, seven lasing channels with a frequency spacing of 0.1 nm (7.62 GHz) and eleven channels with a double-spacing of 0.2 nm (15.24 GHz) are obtained. A wavelength tunability of 1.3 nm is realized for both laser configurations by shifting the pump wavelength. Strong four wave mixing is observed in the double-spacing laser resulting in an improved performance: larger number of channels and better temporal stability. These characteristics in conjunction with the switchable wavelength spacing, extend the laser’s potential in applications, such as WDM fiber-optic communications. Spectral equalization would be possible for the double-Brillouin-spacing MW-BFTL by using long-period fiber gratings at 2 µm [24]. By employing a chalcogenide fiber or chip as the Brillouin gain medium, larger Brillouin gain could be expected, thus leading to a larger number of lasing wavelengths and more efficient operation [25,26]. In addition, the small footprint of such a Brillouin gain medium would reduce the cavity length substantially, making it possible to achieve a single-longitudinal-mode operation and phase-locking, which offers the possibility of pulse generation applications at 2 µm [27].
Acknowledgment
Funding from the Australian Research Council (ARC) through its Laureate Project (FL 120100029) is gratefully acknowledged. This research was also supported by the ARC Centre of Excellence for Ultrahigh bandwidth Devices for Optical Systems (CUDOS, Project No. CE110001018) and the China Scholarship Council (CSC) scholarship.
References and links
1. Y. G. Shee, M. H. Al-Mansoori, S. Yaakob, A. Man, A. K. Zamzuri, F. R. Adikan, and M. A. Mahdi, “Millimeter wave carrier generation based on a double-Brillouin-frequency spaced fiber laser,” Opt. Express 20(12), 13402–13408 (2012). [CrossRef] [PubMed]
2. Y. Yao, X. Chen, Y. Dai, and S. Xie, “Dual-wavelength erbium-doped fiber laser with a simple linear cavity and its application in microwave generation,” IEEE Photon. Technol. Lett. 18(1), 187–189 (2006). [CrossRef]
3. A. Bellemare, M. Rochette, and S. LaRochelle, “Room temperature multifrequency erbium-doped fiber lasers anchored on the ITU frequency grid,” J. Lightwave Technol. 18(6), 825–831 (2000). [CrossRef]
4. L. Talaverano, S. Abad, S. Jarabo, and M. Lopez-Amo, “Multiwavelength fiber laser sources with Bragg-grating sensor multiplexing capability,” J. Lightwave Technol. 19(4), 553–558 (2001). [CrossRef]
5. R. Lewicki, G. Wysocki, A. A. Kosterev, and F. K. Tittel, “Carbon dioxide and ammonia detection using 2 μm diode laser based quartz-enhanced photoacoustic spectroscopy,” Appl. Phys. B 87(1), 157–162 (2007). [CrossRef]
6. B. M. Walsh, “Review of Tm and Ho materials; spectroscopy and lasers,” Laser Phys. 19(4), 855–866 (2009). [CrossRef]
7. 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, J. O’Carroll, M. Becker, N. MacSuibhne, J. Zhao, F. C. Gunning, A. D. Ellis, P. Petropoulos, S. U. Alam, and D. J. Richardson, “Demonstration of amplified data transmission at 2 µm in a low-loss wide bandwidth hollow core photonic bandgap fiber,” Opt. Express 21(23), 28559–28569 (2013). [CrossRef] [PubMed]
8. N. M. Suibhne, Z. Li, B. Baeuerle, J. Zhao, J. Wooler, S. Alam, F. Poletti, M. Petrovich, A. Heidt, N. Wheeler, N. Baddela, E. R. N. Fokoua, I. Giles, D. Giles, R. Phelan, J. O'Carroll, B. Kelly, B. Corbett, D. Murphy, A. D. Ellis, D. J. Richardson, and F. Garcia Gunning, “WDM Transmission at 2μm over Low-Loss Hollow Core Photonic Bandgap Fiber,” in Optical Fiber Communication Conference/National Fiber Optic Engineers Conference 2013, (Optical Society of America,2013), paper OW1I.6.
9. S. Zhao, P. Lu, D. Liu, and J. Zhang, “Switchable multiwavelength thulium-doped fiber ring lasers,” Opt. Eng. 52(8), 086105 (2013). [CrossRef]
10. X. Wang, Y. Zhu, P. Zhou, X. Wang, H. Xiao, and L. Si, “Tunable, multiwavelength Tm-doped fiber laser based on polarization rotation and four-wave-mixing effect,” Opt. Express 21(22), 25977–25984 (2013). [CrossRef] [PubMed]
11. W. Peng, F. Yan, Q. Li, S. Liu, T. Feng, and S. Tan, “A 1.97 μm multiwavelength thulium-doped silica fiber laser based on a nonlinear amplifier loop mirror,” Laser Phys. Lett. 10(11), 115102 (2013). [CrossRef]
12. X. Wang, P. Zhou, H. Xiao, and L. Si, “Multiwavelength Brillouin-thulium fiber laser,” Photonics J. 6(1), 1500507 (2014).
13. Y. Luo, Y. Tang, J. Yang, Y. Wang, S. Wang, K. Tao, L. Zhan, and J. Xu, “High signal-to-noise ratio, single-frequency 2 μm Brillouin fiber laser,” Opt. Lett. 39(9), 2626–2628 (2014). [CrossRef] [PubMed]
14. S. Qin, D. Chen, Y. Tang, and S. He, “Stable and uniform multi-wavelength fiber laser based on hybrid Raman and Erbium-doped fiber gains,” Opt. Express 14(22), 10522–10527 (2006). [CrossRef] [PubMed]
15. D. N. Wang, F. W. Tong, X. Fang, W. Jin, P. K. A. Wai, and J. M. Gong, “Multiwavelength erbium-doped fiber ring laser source with a hybrid gain medium,” Opt. Commun. 228(4), 295–301 (2003). [CrossRef]
16. Y. S. Hurh, G. S. Hwang, J. Y. Jeon, K. G. Lee, K. W. Shin, S. S. Lee, K. Y. Yi, and J. S. Lee, “1-Tb/s (100× 12.4 gb/s) transmission of 12.5-GHz-spaced ultradense WDM channels over a standard single-mode fiber of 1200 km,” IEEE Photon. Technol. Lett. 17(3), 696–698 (2005). [CrossRef]
17. G. S. Hwang, U. S. Pyun, S. H. Kim, Y. Chung, J. S. Lee, and B. W. Kim, “2.5-Tb/s (256× 12.4 Gb/s) Transmission of 12.5-GHz-Spaced Ultra-dense WDM Channels over a Standard Single-Mode Fiber of 2000 km,” in Optical Fiber Communication Conference 2007, (Optical Society of America,2007), paper JWA42.
18. Z. Li, A. M. Heidt, J. M. O. Daniel, Y. Jung, S. U. Alam, and D. J. Richardson, “Thulium-doped fiber amplifier for optical communications at 2 µm,” Opt. Express 21(8), 9289–9297 (2013). [CrossRef] [PubMed]
19. L. F. Stokes, M. Chodorow, and H. J. Shaw, “All-fiber stimulated Brillouin ring laser with submilliwatt pump threshold,” Opt. Lett. 7(10), 509–511 (1982). [CrossRef] [PubMed]
20. R. W. Boyd, Nonlinear Optics, 3rd ed. (Academinc, 2003).
21. Y. G. Shee, M. H. Al-Mansoori, A. Ismail, S. Hitam, and M. A. Mahdi, “Multiwavelength Brillouin-erbium fiber laser with double-Brillouin-frequency spacing,” Opt. Express 19(3), 1699–1706 (2011). [CrossRef] [PubMed]
22. D. Chen, “Stable multi-wavelength erbium-doped fiber laser based on a photonic crystal fiber Sagnac loop filter,” Laser Phys. Lett. 4(6), 437–439 (2007). [CrossRef]
23. J. Tang, J. Sun, L. Zhao, T. Chen, T. Huang, and Y. Zhou, “Tunable multiwavelength generation based on Brillouin-erbium comb fiber laser assisted by multiple four-wave mixing processes,” Opt. Express 19(15), 14682–14689 (2011). [CrossRef] [PubMed]
24. A. M. Vengsarkar, J. R. Pedrazzani, J. B. Judkins, P. J. Lemaire, N. S. Bergano, and C. R. Davidson, “Long-period fiber-grating-based gain equalizers,” Opt. Lett. 21(5), 336–338 (1996). [CrossRef] [PubMed]
25. I. V. Kabakova, R. Pant, D.-Y. Choi, S. Debbarma, B. Luther-Davies, S. J. Madden, and B. J. Eggleton, “Narrow linewidth Brillouin laser based on chalcogenide photonic chip,” Opt. Lett. 38(17), 3208–3211 (2013). [PubMed]
26. K. Hu, I. V. Kabakova, T. F. S. Büttner, S. Lefrancois, D. D. Hudson, S. He, and B. J. Eggleton, “Low-threshold Brillouin laser at 2 μm based on suspended-core chalcogenide fiber,” Opt. Lett. 39(16), 4651–4654 (2014). [CrossRef] [PubMed]
27. T. F. S. Büttner, I. V. Kabakova, D. D. Hudson, R. Pant, C. G. Poulton, A. C. Judge, and B. J. Eggleton, “Phase-locking and Pulse Generation in Multi-Frequency Brillouin Oscillator via Four Wave Mixing,” Sci. Rep. 4, 5032 (2014). [CrossRef] [PubMed]
