We report experimental results demonstrating the variation of optical signal-to-noise ratio (OSNR) of laser lines in Brillouin-Raman fiber laser against Raman pump power (RPP) variation. The reduction of OSNR is attributed to the spectral broadening of laser lines depending on the RPP. The spectral broadening is owing to the effect of the interaction between laser lines and turbulent waves (nonlinear interaction between longitudinal cavity modes).In our experiment, the worst OSNR is obtained at 650 mW RPP as a result of maximum spectral broadening when the Brillouin pump wavelength is fixed at 1555 nm. On the other hand, the OSNR improvement is obtained for RPP beyond 650 mW due to the effect of red-shift, the Raman peak gain is shifted away from the laser lines generated around 1555 nm thus reduces the spectral broadening effect.
©2009 Optical Society of America
Nonlinear effects in optical fiber that are deployed in the generation of multiple wavelengths source have been extensively studied. One of the most accepted approaches is the manipulation of stimulated Brillouin scattering in laser cavities . The stimulated Brillouin scattering (SBS) effect is dependent on the length of the optical fiber used to generate the narrow-spacing Stokes lines . However, additional lengths of optical fiber induce higher cavity losses. In order to minimize this problem, the Raman amplification is employed with a short-length dispersion compensating fiber (DCF) .
For this laser structure dubbed as Brillouin-Raman fiber laser (BRFL), the benefit of distributed Raman amplification on the Brillouin Stokes lines is manifested. Flat amplitude laser lines have been demonstrated from this proposed laser structure [3–5]. This phenomenon is mainly contributed by fiber inherent property of Rayleigh scattering that acts as the virtual mirror. However, the laser lines quality in terms of optical signal-to-noise ratio (OSNR) is low owing to the weak reflectivity of virtual mirror. Recently, we have proposed a linear-cavity BRFL in which both ends are properly terminated with optical feedback components . As a result, the OSNR of the Brillouin Stokes lines improves resulting from high reflectivity of optical feedbacks that force the Stokes lines to oscillate in the laser cavity.
Owing to the high reflectivity mirrors in the laser cavity, the oscillating energy is narrowly concentrated to the Raman peak gain. In Raman fiber lasers, its Stokes spectrum suffers from the spectral broadening due to the contribution of mode competition among longitudinal cavity modes in the laser cavity. The spectral broadening is also observed even though with the presence of fiber Bragg grating to lock the lasing modes as reported in the previous papers [7–9]. The broadening mechanism is due to the nonlinear interaction of the intracavity longitudinal modes through multiple four-wave mixing (FWM) processes that leads to turbulence-induced spectral broadening [10–12]. This type of interaction creates a stochastic growth of the amplitudes and phases of the individual modes that guides to the formation of turbulent waves [9,10]. By integrating the Brillouin Stokes lines in the laser cavity, this fierce mode competition arise from the turbulent waves of Raman fiber lasers is very critical to determine the performance of Brillouin-Raman fiber laser.
In this paper, the OSNR variation of the laser lines is reported for the first time to the best of our knowledge. The experimental findings show that the OSNR degradation is due to the spectral broadening effect of the laser lines which is caused by the interaction between laser lines and turbulent waves. These turbulent waves are resulted from the nonlinear interaction between self-lasing cavity modes of Raman fiber laser induced by multiple FWM processes. In addition, the red-shift effect plays an important role to improve the OSNR of Brillouin Stokes lines by shifting the self-lasing cavity modes to longer wavelengths.
2. Fiber laser configuration
The experimental setup for the generation of multiple Brillouin Stokes lines in Raman cavity laser is depicted in Fig. 1 . The Raman gain is provided by the 14.8-km length of DCF pumped by a Raman pump unit (RPU) at 1455 nm with maximum power of 1.5 W. The DCF has a dispersion value of −1319 ps.nm−1 at 1555 nm, a dispersion slope of 0.035 ps.nm−2.km−1, an effective core area (Aeff) of 15 μm2 and the nonlinear coefficient n2/Aeff = 1.75x10−9 W−1, where n2 is the fiber nonlinear index. The Brillouin pump (BP) is injected from a tunable laser source (TLS) with maximum output power of 6 dBm from 1520 nm to 1620 nm wavelength range and a linewidth of about 200 kHz. In this laser structure, the BP is injected via a 3-dB coupler and then propagates through a wavelength selective coupler (WSC) before experiencing amplification in the DCF.
The generation of Brillouin Stokes line (BSL) occurs when the BP reaches the SBS threshold. That BSL is then amplified by Raman amplification as it propagates in the opposite direction to the BP propagation direction. During its propagation, it generates a second-order Brillouin Stokes line propagating in the same direction as the BP. This process continues as long as the particular higher-order Stokes line has a round-trip gain equal to the cavity loss. The round-trip oscillation in the laser cavity is formed by a mirror at each end of the laser structure. These mirrors are used to reflect Stokes lines as well as the residual Raman pump. The reflectivity of both mirrors is more than 95% for 1520-1570 nm range and is around 89% at 1455 nm. The laser output is taken from the 3-dB coupler and circulator (Cir), and is measured through the optical spectrum analyzer (OSA).
For the measurement of performance parameters such as peak power, signal wavelength and OSNR, the analysis function of “WDM Analysis” embedded in the OSA (Ando AQ6317B) is utilized. Throughout the experiment, all the measurements are performed with the resolution bandwidth set at 0.015 nm. The OSNR parameter is measured by taking the difference between the peak power and its noise level at each signal peak.
3. Results and discussion
In this experiment, the BP wavelength is initially set at 1555 nm which is within the Raman first peak gain (Raman-FPG). Its power is fixed to 6 dBm while the Raman pump power (RPP) is varied from its minimum value to its maximum value of 1.5 W. Figure 2 shows some selected optical spectra at different RPP values.
For the RPP of lower than 300 mW, the behavior of the multi-Stokes lines generation is similar to the conventional Brillouin-Erbium fiber laser. In this case, the number of laser lines increases in tandem with the RPP as reported previously . The behavior of BRFL is thoroughly studied afterwards for the RPP of higher than 300 mW. It has been found that the Raman second peak gain (Raman-SPG) becomes significant when the RPP is pushed to 650 mW as shown in Fig. 2(a).
Beyond this value, the Raman-SPG is observed to grow with broader spectral width and its peak value is shifted to longer wavelengths (i.e. red-shift effect). Red-shift effect on the wavelengths of laser spectra was experimentally observed when the pump power was increased, originating from thermal effects . Temperature inside the laser cavity was determined by employing the intensity ratio of two green up-conversion emission lines. The temperature rises when the pump intensity increases which leads to thermal expansion and variation of the refractive index which quantitatively explain this red-shift. The increment of cavity temperature (ΔT) affects the changes of refractive index (Δn) and the laser cavity size (Δd). Then, these changes influence the resonance condition that leads to changes of the wavelength shift (Δλ). This can be written as given by ;Eq. (1), n and d are the constant values at the room temperature. The other two parameters; ∂n/∂T and ∂d/∂T are the change rates with respect to temperature for the index of refraction and thermal expansion of fiber material respectively.
Focusing into the multiple laser lines within the 1555 nm region, the output spectra at two different RPP values i.e. at 350 mW and 650 mW are further investigated as shown in Fig. 2(b). It can be clearly seen that the OSNR degrades severely when the RPP is varied from 350 mW to 650 mW. The OSNR degradation of the Stokes lines is attributed to the spectral broadening. Even though the effect of line broadening in this study is not experimentally investigated, the main cause is expected from the turbulent waves induced by nonlinear FWM interaction between the self-lasing cavity modes [10–12]. This random behavior later interacts with the laser lines within the specified wavelength range that guides to the reduction of OSNR.
In contrast, Fig. 2(c) shows the output spectra measured for RPP at 650 mW and 900 mW in which the OSNR value is considerably improved in accordance with the RPP increment. The red-shift effect is contributed by the thermal effects resulting from an extremely high RPP delivered into the optical fiber . This trend is however more significant for the Raman-FPG. Owing to the effect of red-shift, the energy is extracted from shorter wavelengths to longer wavelengths. Thus, under this condition, the peak wavelength of the self-lasing cavity modes at Raman-FPG is also shifted to longer wavelengths. As a result, the Stokes lines face less competition with the self-lasing cavity modes within the Raman-FPG and finally, their peak powers increases. The additional energy obtained from this process is able to suppress the self-lasing cavity modes to a certain limit. Hence the OSNR improvement is obtained within this 1555 nm wavelength range (Raman-FPG).
In this analysis, the first Brillouin laser line is actually the BP line derived from the TLS, thus its characteristics such as peak-power, linewidth and central wavelength are very stable. Therefore, for a detailed analysis of the laser line characteristics, this laser line is chosen due to its strong dependence on the laser cavity profiles. Referring to Fig. 3 , the linewidth of the laser line is broadened from 0.022 nm to 0.042 nm when the RPP is varied from 400 mW to 650 mW respectively. However, the linewidth of laser line becomes narrower from 0.042 nm to 0.024 nm as the RPP is further driven to 900 mW. In order to understand this phenomenon, the interaction between laser lines and self-lasing cavity modes is elaborated.
Most laser cavities have the potential of oscillating in a large number of different modes including different axial and transverse cavity modes, these are known as self-lasing cavity modes . Different modes in general have different gains, losses and saturation parameters and thus, compete for the available pump energies in the laser. The existence of many cavity modes inside that laser line is mathematically proven using the estimation of the free spectral range, , where n is the fiber refractive index. In our case, the free spectral range of the laser cavity is found around 7 kHz which is much smaller than the BP linewidth of 200 kHz. However, owing to the limitation of OSA resolution, these modes are not individually observed as depicted in Fig. 3. The spectral information captured by the OSA is actually the envelope of the overlapping optical signals of the Stokes lines and self-lasing cavity modes. In our experiment, the effect of spectral broadening only occurs within the wavelength range of self-lasing cavity modes in BRFL. Since the BRFL utilizes a long cavity, the free spectral range modes are very closely to each other that able to interact with the laser lines. As previously reported in [10–12], these closely-spaced longitudinal cavity modes interact with each other through multiple FWM processes that form turbulent waves. Later, these turbulent waves cause the interaction with the laser lines that causes the effect of spectral broadening as shown in Fig. 3. As the pump power increases the red shift-effect has shifted the peak wavelength of self-lasing cavity modes to the longer wavelengths. As a result, the nonlinear FWM interaction between cavity modes at shorter wavelengths reduces that leads to lowering the formation of turbulent waves. Therefore, the magnitude of interaction between the laser lines and the turbulent waves is also degraded. This scenario is followed by the spectral narrowing as manifested by the OSNR improvement in this experiment.
In our further analysis, the OSNR value of the first-order Brillouin Stokes line with respect to RPP is measured for different RPP ranging from 100 mW to 1000 mW. The launching BP wavelength is varied from 1540 nm to 1570 nm and only significant results from this experiment are analyzed as shown in Fig. 4 . In general, the experimental findings show that the OSNR of laser lines become degraded as the RPP increases as previously discussed due to the spectral broadening. Further increment of the RPP, the OSNR improvement occurs for all BP wavelength except for the 1560 nm. These results show that the interaction between laser lines and turbulent waves is also influenced by the BP wavelength location. In this analysis, a critical RPP is defined as the RPP value in which the OSNR of the laser line starts to improve after experiencing degradation caused by the spectral broadening. This minimum OSNR value is the sign of the maximum interaction between laser lines and turbulent waves as indicated by the solid arrow in Fig. 4.
In this study, the critical RPP values for the BP wavelength of 1549 nm, 1552 nm, 1555 nm and 1557 nm are 300 mW, 400 mW, 650 mW and 700 mW respectively. This trend indicates that for longer BP wavelength values, higher RPP values are required to red-shift the Raman-FPG away from the first-order Brillouin Stokes line. For the BP wavelength of 1560 nm, much higher RPP values are required to shift the Raman peak gain away from the first-order Brillouin Stokes line. Briefly, this study indirectly shows that when the BP wavelength is injected within the Raman peak gain, the spectral broadening effect of Stokes lines is significantly observed. For the shorter wavelengths of BP, this maximum interaction occurs at lower critical RPP values and this value increases in tandem with the increment of BP wavelength.
In order to validate the above mentioned findings, the output spectrum of the proposed BRFL structure is recorded without applying the BP into the laser cavity. The purpose of this experiment is to investigate the characteristics of Raman peak gain in the laser cavity with respect to the RPP. The findings from this experiment are illustrated in Fig. 5 , showing the shift of Raman-FPG central wavelength from 1554.8 nm to 1561.2 nm when the RPP is varied from 450 mW to 1000 mW. This fashion is followed by the spectral width broadening from 1.4 nm to 4.9 nm respectively. It can also be seen that the Raman-SPG starts growing when the RPP is set above 650 mW, hence the gain competition with the Raman-FPG starts taking place afterwards. Since the magnitude of Raman-FPG is higher than the Raman-SPG, it is able to suppress the growth of Raman-SPG by extracting more energy available from the laser cavity. This can be seen when the RPP is pushed from 800 mW to 950 mW, which results in lowering the Raman-SPG amplitude. These findings show that the evolution of peak-wavelength of the Raman peak gain is strictly influenced by the RPP. The spectral linewidth broadening is strictly dependent on the mode competition that arises from the laser cavity, its cavity gain strongly determined by the RPP as depicted in Fig. 5. Therefore, the interaction between the turbulent waves (nonlinear FWM interaction between cavity modes) and laser lines is the main cause of the spectral linewidth broadening. In addition to this, the effect of red-shift is the origin of the spectral linewidth narrowing.
We have investigated the OSNR performances of the laser lines in Brillouin-Raman fiber laser in the presence of turbulent waves. These waves are formed by the nonlinear FWM interaction between the longitudinal cavity modes of the Raman fiber laser. With the existence of laser lines in the cavity, these turbulent waves interact with these laser lines that result in the broadening of spectral linewidth. This broadening effect leads to the OSNR deterioration of the laser lines. It is maximized when the Raman peak gain region is within the Brillouin pump wavelength region under study as indicated by the minimum OSNR values. Within the Raman peak gain, the position of the minimum OSNR value is mainly dependent on the RPP values and the wavelength of BP. In the experiment, this mode interaction is reduced by shifting away the Raman peak gain beyond their spectral location. This can be achieved by means of red-shift effect attributed to the RPP increment into the laser cavity. As a result of the red-shift effect, the OSNR performance of laser lines is improved by transferring the self-lasing cavity modes to the longer wavelengths. In conclusion, this study analyzes the OSNR performance of the Brillouin-Raman laser lines through the spectral broadening and red-shift effects.
2. G. P. Agrawal, Nonlinear Fiber Optics, 4th ed., (Academic Press, New York, 2006).
3. B. Min, P. Kim, and N. Park, “Flat amplitude equal spacing 798 channel Rayleigh assisted Brillouin/Raman multiwavelength comb generation in dispersion compensating fiber,” IEEE Photon. Technol. Lett. 13(12), 1352–1354 (2001). [CrossRef]
4. K. D. Park, B. Min, P. Kim, N. Park, J. H. Lee, and J. S. Chang, “Dynamics of cascaded Brillouin-Rayleigh scattering in a distributed fiber Raman amplifier,” Opt. Lett. 27(3), 155–157 (2002). [CrossRef]
5. K. D. Park, H. Ryu, W. K. Lee, S. K. Kim, H. S. Moon, and H. S. Suh, “Threshold features of a Brillouin Stokes comb generated in a distributed fiber Raman amplifier,” Opt. Lett. 28(15), 1311–1313 (2003). [CrossRef] [PubMed]
6. A. K. Zamzuri, M. I. Md Ali, A. Ahmad, R. Mohamad, and M. A. Mahdi, “Brillouin-Raman comb fiber laser with cooperative Rayleigh scattering in a linear cavity,” Opt. Lett. 31(7), 918–920 (2006). [CrossRef] [PubMed]
7. J. C. Bouteiller, “Spectral modeling of Raman fiber lasers,” IEEE Photon. Technol. Lett. 15(12), 1698–1700 (2003). [CrossRef]
9. J. Hagen, R. Engelbrecht, O. Welzel, A. Siekiera, and B. Schmauss, “Numerical modeling of intracavity spectral broadening of Raman fiber lasers,” IEEE Photon. Technol. Lett. 19(21), 1759–1761 (2007). [CrossRef]
10. S. A. Babin, D. V. Churkin, A. E. Ismagulov, S. I. Kablukov, and E. V. Podivilov, “Four-wave-mixng-induced turbulent spectral broadening in a long Raman fiber laser,” J. Opt. Soc. Am. B 24(8), 1729–1738 (2007). [CrossRef]
11. S. A. Babin, V. Karalekas, E. V. Podivilov, V. K. Mezentsev, P. Harper, J. D. Ania-Castanon, and S. K. Turitsyn, “Turbulent broadening of optical spectra in ultralong Raman fiber lasers,” Phys. Rev. A 77(3), 033803 (2008). [CrossRef]
12. S. A. Babin, D. V. Churkin, A. E. Ismagulov, S. I. Kablukov, and E. V. Podivilov, “Turbulence-induced square-root broadening of the Raman fiber laser output spectrum,” Opt. Lett. 33(6), 633–635 (2008). [CrossRef] [PubMed]
13. G. J. Cowle and D. Y. Stepanov, “Multiple wavelength generation with Brillouin/erbium fiber lasers,” IEEE Photon. Technol. Lett. 8(11), 1465–1467 (1996). [CrossRef]
14. Z. P. Cai, A. Chardon, H. Y. Xu, P. Feron, and G. M. Stephan, “Laser characteristics at 1535 nm and thermal effects of an Er:Yb phosphate glass microchip pumped by Ti:Sapphire laser,” Opt. Commun. 203(3-6), 301–313 (2002). [CrossRef]
15. A. E. Siegmen, Lasers, (University Science Books, California, 1986).