Open Access
Add to BibSonomyAbstract
We experimentally compare output performance between laser beam (erbium fiber ring laser) and amplified spontaneous emission (ASE) beam (erbium fiber ASE) driven supercontinuums (SCs) in terms of seed beam temporal coherence. We control the degree of temporal coherence of the seed beams by using an optical filter to change their spectral linewidth. The random phase ASE driven SC is found to have better performance than the phase-correlated laser driven SC in terms of spectral smoothness and output power. Significantly high relative-intensity-noise in the output SCs is observed for both cases, i.e. the laser driven SC and the ASE driven SC irrespective of the seed beam temporal coherence due to the nonlinear amplification of quantum fluctuations both in the input pump beam and in the Raman scattering process.
©2006 Optical Society of America
1. Introduction
Optical fiber based supercontinuum light sources have attracted huge research attention in recent years. Such interest in the type of light sources originates from their potential usefulness in a variety of fields such as communication [1], optical coherent tomography [2], metrology [3], and optical sensing [4]. Moreover, the recent technological advance in fabricating high-quality nonlinear optical fibers such as holey fiber [5] and highly nonlinear dispersion-shifted fiber (HNL-DSF) [6] allows us to readily implement stable and practical supercontinuum sources. The rapid progress of the high power fiber laser technology is also another key contributing factor for the development of high performance, compact, optical fiber based supercontinuum sources.
Typical fiber based supercontinuum sources have been based on a high-power pump laser whose output beam is injected into a nonlinear fiber, where the supercontinuum is generated. Unlike the conventional schemes using a pulse laser as a pump beam, there was a novel proposition by Prabu et al. that a high power continuous-wave (CW) laser could also be used as a pump beam for supercontinuum generation [7]. In the CW supercontinuum case, output average power is usually extremely high (watt level) due to its tight requirement of high pump power for generation of nonlinear effects in optical fiber.
The physical mechanism for CW supercontinuum evolution in optical fiber is different from the pulse-mode supercontinuum evolution [8, 9]. Both modulation instability (MI) and stimulated Raman scattering (SRS) were found to play key roles in transforming a narrowband CW laser beam into a broadband spectral continuum. MI leads to soliton like structure formation that experiences subsequent self Raman interaction. The large numbers of noisy solitonic structures, randomly distributed in space and time give rise to the broad and flat continuum. In such a supercontinuum generation one of key factors is to use a high power pump beam with a low coherence since the Raman soliton formation is induced by the random phase and intensity fluctuations of the pump beam propagating through an anomalous dispersion optical fiber [10]. A range of CW supercontinuum demonstrations have been successfully performed with various types of pump sources; for example, a high power Raman fiber laser [9,11], a rare earth-doped fiber amplified spontaneous emission source (ASE) [12], a low-coherent semiconductor laser diode [13], and a rare earth-doped fiber laser [14].
In this paper, we experimentally carry out a comparative study on the impact of input seed beam coherence on output CW supercontinuum performance. More specifically, we compare output optical properties of CW SCs generated from a highly nonlinear fiber based single pass structure by use of two different types of seed beams such as an erbium-doped fiber (EDF) ASE and an EDF based ring laser. Here, the ASE is a depolarized light with random phase fluctuations whilst the EDF ring laser is a polarized light with a phase correlation. We control the degree of temporal coherence of the sources by tuning their spectral linewidth through use of an optical bandpass filter and observe the output optical property variations. Output SC performance is evaluated and compared in terms of spectral evolution, output power, and relative-intensity-noise (RIN).
As a matter of fact, a similar comparative study between SCs driven by a Ytterbium-doped fiber ASE and a Ytterbium-doped fiber laser seed beams was performed by C. J . S. de Matos et al. [12]. Interestingly, our investigation shows a very much different result from that in Ref. [12], particularly in terms of output RIN characteristics.
2. Experimental setup
Figure 1 shows the experimental schematics for the CW SC generation with the two different types of seed beams. For the ASE driven SC, a 5-m EDF with a peak absorption of 6 dB/m at 1530 nm was initially pumped with a 980-nm pump LD and the generated ASE was spectrally sliced by a bandpass filter at a center wavelength of 1550 nm. The spectrum sliced ASE component was amplified with another 5-m EDF with a peak absorption of 6 dB/m at 1530 nm to ensure a sufficient power level for the next power amplification. Three isolators were placed at proper locations to suppress undesirable internal lasing. For the ring laser driven SC, we constructed a fiber ring laser with the same optical components as the ASE. In this case, the laser output power was extracted from the ring cavity using an 80:20 fiber coupler, with which 80 % of the oscillated light was fed back into the EDF. An optical isolator was placed before the coupler within the cavity to ensure the directional light oscillation. Since the temporal coherence of a light is known to be inversely proportional to its spectral linewidth [15], we were able to control temporal coherence of the seed beams by changing the 3-dB bandwidth of the inserted bandpass filter.
Since the optical powers of the seed beams were not strong enough to generate SC phenomenon in a HNL-DSF, we amplified the seed beams again with a power amplifier that consisted of a 5-m EDF with a peak absorption of 16.7 dB/m at 1550 nm and a high power 1480 nm fiber laser pump. 2-km length of silica-based HNL-DSF was then employed at the output end of the power amplifier. The 2-km HNL-DSF was composed of two different pieces of 1-km HNL-DSF that have different zero dispersion wavelengths of 1554 nm and 1532 nm, each but had a same nonlinearity parameter of ~15.5 W-1.km-1. The output was characterized by use of an optical spectrum analyzer (OSA) with a maximum measurable wavelength of 1750 nm, a thermal power meter, and a RIN measurement setup composed of a low-noise photodetector with a bandwidth of 150 MHz and an electrical spectrum analyzer.
Fig. 1. Experimental schematic for continuous wave supercontinuum generation with two different seed light sources: erbium doped fiber based ring laser and spectrum sliced erbium ASE.
3. Experimental results and discussion
First, we measured spectral linewidth of the two types of seed beams while we changed the 3-dB bandwidth of the inserted bandpass filter and then calculated the corresponding temporal coherence time (τc) using the following simple equation [15].
where S(ν) is the power spectral density of a beam and Δνc is its spectral linewidth. Note that the spectral linewidth defined in Eq. (1) is not the full width at half maximum.
Figure 2 shows the output optical spectra of the ring laser seed beam for various filter bandwidths. The spectral linewidths of the ring laser seed beam were 0.039, 0.19, 0.424, and 0.92 nm for the filter 3-dB bandwidths of 0.3, 1, and 3 nm, and no filter, respectively. The same measurements were performed for the ASE seed beam as shown in Fig. 3 and the measured spectral linewidths were 0.42, 1.37, 4.56, and 26 nm for the filter bandwidths of 0.3, 1, and 3 nm, and no filter, respectively. The estimated temporal coherence times for the measured linewidths are summarized in Table 2.
Fig. 2. Measured output optical spectra of the erbium fiber ring laser for various filter bandwidths and that for no filter inserted.
Fig. 3. Measured output optical spectra of the erbium fiber ASE for various filter bandwidths and that for no filter inserted.
Table 1. Temporal Coherence Time (τc) of the Seed Sources
Second, we performed the measurement of output spectral evolution as a function of the 1480-nm pump power of the power amplifier for both the ring laser driven and the ASE driven SCs. Figure 4 shows the measured SC spectra versus the 1480-nm pump power for various seed beam linewidths in the ring laser driven SC. Note that the spectral components displayed in the spectral range from 1675 to 1750 nm at a 1480-nm pump power of 0.42 W in the spectra of Fig. 4 are due to the receiver noise of the OSA used in this experiment. Among the four different temporal coherence seed laser beams, the lowest coherence beam that corresponded to the ring laser with no filter was found to produce the flattest and the smoothest SC. In case of the highest coherence pump beam that was generated with a 0.3 nm bandpass filter, it was very hard to observe a spectral continuum although MI peaks and the first order Raman Stokes peak was observed as the pump power was enlarged. There existed a huge spectral valley between the longer wavelength MI peak and the Raman Stokes peak even at a maximum pump power of ~2.1 W. When we increased the laser linewidth, the depth of the spectral valley was found to decrease. However, the strong seed wavelength peak and the Raman Stokes peak always remained unless the seed beam generated with no filter was employed. This means that noisy Raman soliton generation and their self Raman interaction were not properly developed due to lack of sufficient phase and intensity fluctuations in the narrow linewidth, highly coherent seed lasers even if their peak powers were strong enough to generate both MI and Raman Stokes.
Figure 5 shows the measured SC spectra versus the 1480-nm pump power for the ASE driven SC. The smoothest SC was found when the ASE seed beam with the lowest temporal coherence was used. One interesting finding is that the ASE seed beam generates a better SC than the laser seed beam at a same temporal coherence condition; however, a good quality of CW SC can still be obtained from the phase-correlated laser seed beam by degrading its temporal coherence property.
Fig. 4. Measured output spectra versus the 1480-nm pump power of the power amplifier for various temporal coherence times in the ring laser driven SC.
Fig. 5. Measured output spectra versus the 1480-nm pump power of the power amplifier for various temporal coherence times in the ASE driven SC.
Fig. 6. Measured output power of (a) the ring laser driven SC and (b) the ASE driven SC versus the 1480-nm pump power of the power amplifier for various temporal coherence times.
Third, we characterized both the ring laser driven and the ASE driven SCs in terms of output power. Figure 6(a) shows the measured output power of the ring laser driven SC versus the 1480-nm pump power for various temporal coherence times while the results for the ASE driven SC are shown in Fig. 6(b). Stimulated Brillouin scattering induced SC efficiency reduction was not observed due to broad seed beam linewidths. It is noticeable that larger output power for both cases, i.e. the ring laser driven SC and the ASE driven SC was found when the seed beam of lower temporal coherence was used. This phenomenon in the ring laser driven SC is believed to be associated with the degree of balance among MI induced Raman soliton generation, self Raman interaction, and Raman Stokes generation. In case of the seed beams of high temporal coherence the power spectral density of the beams is high enough to cause Raman Stokes generation at the higher-loss wavelength bands larger than 1700 nm before the MI induced Raman soliton generation is fully developed. This phenomenon was clearly observed when the ring laser seed beam with a coherence time (τc) of 205 ps was used for SC generation as shown in Fig. 4(a). Note that a critical pump power turning from the output increasing regime into the decreasing regime was observed to exist at a 1480-nm pump power of 1.23 W for both cases of τc = 42.1 and 18.9 ps as shown in Fig. 6(a) and this turning power corresponded to the point at which the second order Raman Stokes was generated. However, in case of the ring laser seed beam with τc = 8.94 ps the improved balance among the three critical nonlinear processes reduced the amount of the pump beam energy that would be transferred into the Raman Stokes at the higher-loss wavelength bands and thus wasted.
Interestingly, the largest output power SC (output power of 475 mW, pump conversion efficiency of ~22.5 %) among the 8 cases was found when the random phase ASE seed beam of the lowest temporal coherence was used. This phenomenon can be readily understood by considering the fact that a random phase ASE of lower temporal coherence allows for both more effective suppression of the Raman Stokes generation caused by its lower power spectral density and better development of noisy Raman soliton structure.
Finally, we performed the RIN measurements of the SC outputs. The SC output beams were passed through a 15-nm bandpass filter centered at 1570 nm, and subsequently coupled onto a low-noise photodetector with a bandwidth of 150 MHz. The detected electrical signals were then ac-coupled into an electrical spectrum analyzer. Figure 7(a) shows the measured RIN spectra for the ring laser driven SC while the results for the ASE driven SC are shown in Fig. 7(b). The multiple frequency peaks displayed in the graphs of Fig. 7(a) indicate resonance mode peaks of the ring lasers and their harmonics. Those peaks could be detrimental to some applications such as optical communication systems
All of the SCs were found to have significantly high RIN values irrespective of the seed beam types. In case of the ring laser driven SC, RIN values of the output SC were observed to be between -120 and -110 dBc/Hz for the four temporal coherence times and to decrease with the increase in the seed laser temporal coherence. On the other hand, RIN values of the ASE driven SC were between -110 and -100 dBc/Hz and also decreased with the increase in the seed ASE temporal coherence.
It is noticeable that these RIN measurement results are contradictory to those in Ref. [12] in which the Ytterbium fiber ASE driven SC was observed to have a better RIN than the Ytterbium fiber laser driven SC and even to exhibit an extremely low RIN of -150 dBm/Hz comparable to a detector noise floor. According to our measurements in Fig. 7 our finding is that CW SCs based on single propagation of a high power beam through a highly nonlinear fiber should have significantly high RIN values whether the seed beam has random phase fluctuations or not. This can be readily understood by use of the basic mechanism of CW SC evolution in an anomalous dispersion, highly nonlinear optical fiber. As mentioned in Introduction, the balanced interplay among MI induced noisy Raman soliton generation, self Raman interaction, and Raman Stokes generation is a vital factor to broad and flat CW SC generation. This means that considerable output intensity noise in the time domain always occurs due to the nonlinear amplification of quantum fluctuations both in the input pump beam and in the Raman scattering process [16, 17].
Another noticeable point from Fig. 7 is that the ring laser driven SC exhibits slightly better RIN values than the ASE driven SCs as opposite to the observation in Ref. [12]. This observation can also be attributed to the nonlinear amplification of quantum fluctuations in the input pump beam [16, 17]. Note that the ASE seed beam has much higher RIN values than the ring laser as shown in Fig. 7.
4. Conclusion
We have experimentally performed a study on CW supercontinuum performance depending on input seed beam coherence. For this study we prepared two types of seed beams: an erbium fiber ring laser and an erbium fiber ASE. Then we controlled the degree of temporal coherence of the seed beams by tuning their spectral linewidth through use of an optical bandpass filter. Output SC was generated by use of a highly nonlinear fiber based single pass SC generating structure. The random phase ASE driven SC was found to have a better performance than the phase-correlated laser driven SC in terms of spectral smoothness and output power. The generated SCs were also compared one another from a perspective of RIN that is another key parameter for practical application of the SCs to real photonic systems. A significantly high RIN was observed for both cases, i.e., the phase-correlated laser driven SC and the random phase ASE driven SC irrespective of the seed beam temporal coherence.
In terms of RIN characteristics we obtained a contradictory result to that in Ref. [12]; however, we believe that our observation should not be wrong according to the recent detailed studies on the mechanism of CW SC evolution [8, 9, 10]. One concern in this study is that the considerable RIN might be an inevitable feature that always accompanies the CW SC generation process relying on seed beam phase/intensity fluctuations since the nonlinear amplification of quantum fluctuations both in the input seed beam and Raman scattering process should always occur in the Raman scattering process.
In order to cope with the high RIN problem of CW SCs, more study needs to be performed for the future. Note that the high RIN problem might be a serious performance limiting factor that prevents the use of CW SCs in some applications like high speed telecommunication systems due to low spectral efficiency caused by the larger optical bandwidth requirement proportional to the electrical bandwidth to secure a high electrical signal-to-noise ratio [18]
References and links
1. K. Mori, K. Sato, H. Takara, and T. Ohara, “Supercontinuum lightwave source generating 50 GHz spaced optical ITU grid seamlessly over S-, C- and L-bands,” Electron. Lett. 39, 544–546 (2003). [CrossRef]
2. I. Hartl, X. D. Li, C. Chudoba, R. K. Ghanta, T. H. Ko, J. G. Fujimoto, J. K. Ranka, and R. S. Windeler, “Ultrahigh-resolution optical coherence tomography using continuum generation in an air-silica microstructure optical fiber,” Opt. Lett. 26, 608–610 (2001). [CrossRef]
3. K. Kim, B. R. Washburn, G. Wilpers, C. W. Oates, L. Hollberg, N. R. Newbury, S. A. Diddams, J. W. Nicholson, and M. F. Yan, “Stabilized frequency comb with a self-referenced femtosecond Cr:forsterite laser,” Opt. Lett. 30, 932–934 (2005) [CrossRef] [PubMed]
4. H. Kano and H. Hamaguchi, “Characterization of a supercontinuum generated from a photonic crystal fiber and its application to coherent Raman spectroscopy,” Opt. Lett. 28, 2360–2362 (2003). [CrossRef] [PubMed]
5. S. Coen, A. H. L. Chau, R. Leonhardt, J. D. Harvey, J. C. Knight, W. J. Wadsworth, and P. St. J. Russell, “White-light supercontinuum generation with 60-ps pump pulses in a photonic crystal fiber,” Opt. Lett. 26, 1356–1358 (2001). [CrossRef]
6. P. S. Westbrook, J. W. Nicholson, K. S. Feder, and A. D. Yablon, “Improved supercontinuum generation through UV processing of highly nonlinear fibers,” J. Lightwave Technol. 23, 13–18 (2005). [CrossRef]
7. M. Prabhu, N. S. Kim, and K. Ueda, “Ultra-broadband CW supercontinuum generation centered at 1483.4 nm from Brillouin/Raman fiber laser,” Jpn. J. Appl. Phys. 39, L291–L293 (2000). [CrossRef]
8. A. V. Avdokhin, S. V. Popov, and J. R. Taylor, “Continuous-wave, high-power, Raman continuum generation in holey fibers,” Opt. Lett. 28, 1353–1355 (2003). [CrossRef] [PubMed]
9. A. K. Abeeluck, C. Headley, and C. G. JØrgensen, “High-power supercontinuum generation in highly nonlinear dispersion-shifted fibers by use of a continuous-wave Raman fiber laser,” Opt. Lett. 29, 2163–2165 (2004). [CrossRef] [PubMed]
10. S. M. Kobtsev and S. V. Smirnov, “Modelling of high-power supercontinuum generation in highly nonlinear, dispersion shifted fibers at CW pump,” Opt. Express 13, 6912–6918 (2005). [CrossRef] [PubMed]
11. J. H. Lee, Y. Takushima, and K. Kikuchi, “Continuous-wave supercontinuum laser based on an erbium-doped fiber ring cavity incorporating a highly nonlinear fiber,” Opt. Lett. 30, 2599–2602 (2005). [CrossRef] [PubMed]
12. C. J. S. de Matos, S. V. Popov, and J. R. Taylor, “Temporal and noise characterisitcs of continuous-wave-pumped continumm generation in holey fibers around 1300 nm,” Appl. Phys. Lett. 85, 2706–2708 (2004). [CrossRef]
13. A. K. Abeeluck and C. Headley, “Supercontiuum growth in a highly nonlinear fiber with a low-coherence semiconductor laser diode,” Appl. Phys. Lett. 85, 4863–4865 (2004). [CrossRef]
14. P. A. Champert, V. Couderc, and A. Barthelemy, “1.5-2.0 μm multiwatt continuum generation in dispersion-shifted fiber by use of high-power continuous-wave fiber source,” IEEE Photon. Technol. Lett. 16, pp.2445–2447 (2004). [CrossRef]
15. B. E. A. Saleh and M. C. Teich, Fundamentals of Photonics (Wiley Interscience,1991), 344–359.
16. M. Nakazawa, H. Kubota, and K. Tamura, “Random evolution and coherence degradation of a high-order optical soliton train in the presence of noise,” Opt. Lett. 24, 318–320 (1999). [CrossRef]
17. K. L. Corwin, N. R. Newbury, J. M. Dudley, S. Coen, S. A. Diddams, K. Weber, and R. S. Windeler, “Fundamental noise limitations to supercontinuum generation in microstructure fiber,” Phys. Rev. Lett. 90, 113904 (2003). [CrossRef] [PubMed]
18. G. J. Pendock and D. D. Sampson, “Transmission performance of high bit rate spectrum-sliced WDM systems,” J. Lightwave Technol. 14, 2141–2148 (1996) [CrossRef]
