Open Access
Add to BibSonomyAbstract
A new type of highly nonlinear fiber (HNLF) was designed and fabricated. The new HNLF was engineered to reduce dispersion shift due to transverse fluctuations while maintaining the modal confinement superior to that of the conventional fibers. The new design strategy was validated by the measurements of the global and local dispersive characteristics under considerable core and index profile deformation induced by tensile stress, which indicated that the dispersive and phase matching characteristics of the fiber did not change even under the highest tensile stress. The characteristics effectively decoupled tension-based Brillouin suppression from phase-matching impairments in parametric mixers for the first time. The new HNLF was used to demonstrate the first coherence-preserving mixer operating in the short-wavelength infrared (SWIR) band. The SWIR mixer was driven by continuous-wave near-infrared (NIR) pump and did not require pump phase dithering to suppress Brillouin scattering.
©2012 Optical Society of America
1. Introduction
While parametric mixing devices have been demonstrated with various materials and waveguide platforms [1–3], high-confinement silica fibers have been chosen in a majority of recent demonstrations in high capacity transmission [4], hybrid signal processing [5] and fast tunable emitters [6]. In contrast to the material platforms that offer considerably larger nonlinear response [7,8], silica has comparatively small Kerr index and does not represent an obvious choice for nonlinear device construction. However, silica fibers are also remarkably transparent, thus rendering kilometer-scale photon interactions possible [9]. Consequently, a highly-efficient nonlinear interaction can readily be achieved in a distributed manner, as weak silica nonlinearity can be offset by its near lossless nature.
This advantage is made even more relevant by the recent developments of high-confinement, low-loss fibers that combine precise dispersion (phase-matching) control and low propagation loss [9,10]. Indeed, the nonlinear figure of merit, defined by the product of waveguide nonlinearity, pump power and effective length [11], is nearly an order of magnitude higher in highly nonlinear fibers (HNLF) than in any other mixer platforms [9]. Strong modal confinement in HNLFs allows for simultaneously enhancement of the effective fiber nonlinearity and a significant waveguide dispersion modification [10]. While both characteristics contribute to the efficiency of a HNLF mixer device, the latter can also lead to crippling impairments in case when precise, spectrally distant four-photon mixing (FPM) interaction is sought [10,12]. In simple terms, the combination of small core size and high index contrast between core and cladding [10] induces high sensitivity to the transversal geometry fluctuations along the fiber length [13]. As a consequence, the conventional HNLF type cannot be used in distant-band convertors which require precise phase-matching [12].
Recognizing this basic limitation, we have recently described a new HNLF design that combines modal confinement with inherent resilience to fiber geometry fluctuations [12]. In contrast to the previously reported strategy in which the stochastic fluctuations are mitigated by post-fabrication processing [14], the new approach relies on an index profile that reduces dispersion shifts as the transverse geometry of the fiber is varied. Consequently, not only that the phase matching condition can be precisely controlled along the entire length of the new fiber, but the fiber can also be longitudinally strained without affecting the phase-matching condition. The latter is of significant practical importance as it allows for decoupling of Brillouin suppression and phase-matching mechanisms [15].
This paper reports on fabrication and characterization of the new HNLF type. The first part describes the measurements of the new HNLF regarding to the global and local optical characteristics under uniform strain (core compression). The second part demonstrates the ability to suppress Brillouin scattering without affecting the dispersive (phase-matching) characteristics of the parametric mixer built upon. Finally, the new HNLF were used to demonstrate the first continuous-wave (CW)-pumped short-wave infrared (SWIR) band mixer in the absence of pump-phase dithering.
2. Fluctuation-resilient HNLF characteristics
Rather than relying on a delta-like core found in conventional HNLF designs [10], the new fiber design relies on a multi-layered core structure to provide direct control of the total dispersion and optical field confinement characteristics. Figure 1(a) shows the dispersion profile of the new HNLF fabricated using the new design approach, specifically tailored for distant-band parametric generation phase-matched by negative fourth-order dispersion (β4) [16,17]. At 1550nm, the fiber was characterized by a loss of 0.4 dB/km, an effective mode area of 22µm2. The dispersion measurements showed a zero-dispersion wavelength (ZDW) of 1593 nm and a dispersion slope of 0.067 ps/nm2/km. The nonlinear coefficient at the ZDW was measured to be 5.2 W−1km−1 using self-phase modulation method [18]. The solid fiber structure of the new HNLF maintained good splicing compatibility with standard fibers, in which a splicing loss of less than 0.4 dB per splice was attained when the HNLF was spliced to a standard single-mode fiber (SMF) using a commercially available fusion splicer with a normal SMF-to-SMF splicing program. The normal dispersive characteristics in the C-band combined with the negative fourth-order dispersion term (β4 = −5 × 10−4 ps4/km) allow parametric mixing in the 2-μm short-wavelength infrared (SWIR) band using C-band laser source, as shown by the phase-matching contour in Fig. 1(b).
Fig. 1 (a) Dispersion profile of the new HNLF and (b) Phase-matching contour of the parametric mixer constructed with the new HNLF. The dashed line in (a) denotes the zero-dispersion position.
While an optimal global dispersion profile matching the desired parametric mixer response can be easily synthesized in a conventional HNLF with a single-layer core, the local dispersive characteristics are subject to considerable fluctuations due to the presence of radial geometry variation [13]. The multi-layer core structure in the new fiber was specifically designed to reduce the influence of fiber geometry fluctuations on dispersion, thereby improving dispersion stability. Figure 2 shows the calculated chromatic dispersion and zero-dispersion wavelength dependency on the core size of the new and conventional HNLFs. Indeed, the dispersion characteristics of the new HNLF are visibly less sensitive to core deformation, as depicted by the impunity to core radius deviation.
Fig. 2 Calculated core radius dependency of (a) chromatic dispersion at 1593 nm and (b) zero-dispersion wavelength of a conventional HNLF and the new HNLF.
The global dispersion stability demonstrated in the numerical model is as well exhibited in the local dispersion characteristics of the manufactured fibers. Figure 3 shows the experimental setup for characterizing the ZDW fluctuations based on the pump-scanning method [19,20]. The setup resembles a typical parametric wavelength converter pumped by a partially-coherent source with 0.2-nm spectral width, which converts a signal (probe) situated in the 1300-nm band to the 2-μm band. In the ideal case when the fiber-under-test does not possess radial (dispersive) fluctuations, the union of the idler spectra should form a narrow spectral line as a result of the highly-confined phase-matching window created by negative β4. In a conventional HNLF, however, the phase-matched window is convolved by the dispersion fluctuations, and manifests as a broadening of the idler spectra overlays.
Fig. 3 Experimental setup for ZDW fluctuation measurement using pump-wavelength scan method. ASE: amplifier spontaneous emission; TBPF: tunable band-pass filter; EDFA: erbium-doped fiber amplifier; WDMC: Wavelength-division multiplexing coupler; OSA: Optical spectrum analyzer.
Figure 4 shows the idler spectra generated by a conventional HNLF and the new fiber. Indeed, the dispersion fluctuations in the conventional HNLF that are indirectly measured in Fig. 4(a) cannot be characterized by a well-defined spectral peak; rather it is composed of an ensemble of the conversion contributed by HNLF subsections possessing widely varying dispersion. In contrast, the phase-matched conversion peak corresponding to the new HNLF type is well defined, indicating qualitatively better dispersion stability than that of the conventional fiber. While the 3-dB linewidth cannot be readily identified from the conventional HNLF measurement, its spectral width exceeded that of the new HNLF by approximately an order of magnitude, as seen in Fig. 4(b).
Fig. 4 Conversion spectra generated by (a) conventional HNLF and (b) new fiber. The probe wave was at 1298 nm in both cases.
The second set of measurements investigated the Brillouin scattering and dispersive properties of the new HNLF under tensile stress. This measurement is of particular practical importance as the efficacy of Brillouin scattering suppression using tensile straining depends on the extent of Brillouin gain spectral broadening inducible by straining, under the limit set by the mechanical resilience of the fiber. A well-behaved HNLF would allow Brillouin gain-peak shift by multiple gain bandwidth with the maximum applied strain within the elastic deformation envelope, typically below 1% [20]. The relationship between Brillouin frequency shift and tensile strain was studied using the characterization setup shown in Fig. 5 .
Fig. 5 Experimental setup for measuring Brillouin gain frequency shift induced by tensile strain. ECL: External-cavity laser; EDFA: Erbium-doped fiber amplifier; CIR: Circulator; PD: Photodetector; ESA: Electrical spectrum analyzer.
A 20m-long HNLF section was spooled between two metal poles that can be translated with precisely controlled displacement. The desired strain was applied by moving a one of the metal pole. The Brillouin gain spectrum was acquired by observing the RF-beat tone between the Rayleigh- and Brillouin-scattered waves at a circulator output port. The measurement is shown in Fig. 6(a) and depicts a linear relationship between the tensile strain and Brillouin gain peak frequency shift, with a proportionality constant of 298 MHz/% strain. The Brillouin gain bandwidth was found to be 10 MHz and demonstrated negligible broadening (less than 10%) at a strain up to 3%. Consequently, the narrow Brillouin linewidth allowed for effective Brillouin scattering suppression at a moderate level of longitudinally varied strain. For instance, a linear strain ramp can increase the Brillouin-scattering threshold by 13 dB with a maximum strain of merely 1%, as estimated by using the experimentally-calibrated model reported in [21] (Fig. 6(b)).
Fig. 6 (a) Map of Brillouin gain spectrum shift versus tensile strain. (b) Estimated Brillouin peak gain levels for a HNLF stretched with a linear ramp strain profile and with various maximum strain. The curve was calculated using the Brillouin gain characteristics of the new HNLF.
Owing to the dispersion shift due to the deformation of transversal geometry as well as the photoelastic-induced refractive index change, a conventional HNLF exhibits high dispersion fluctuations when a varying tensile strain is applied for suppressing Brillouin scattering. In contrast, the exceptional dispersion stability offered by the new fiber design effectively decouples the dispersion stability and the attainable level of Brillouin scattering suppression by straining. To verify this important proposition, the dispersion of the new HNLF under 3% strain was measured and compared against the strain-free measurement, as shown in Fig. 7 . In spite of being subjected to considerable index profile deformation (1.5% core radius constriction), the effect of the applied strain on the fiber dispersion was negligible, where the dispersion shift was below the resolution of the optical network analyzer (0.05 ps/nm/km) used for the dispersion measurement. While merely a small change in ZDW (< 0.8 nm) was observed in the new HNLF, the corresponding 1.5% core compression in a conventional HNLF can cause ZDW shift by approximately 45 nm [10].
Fig. 7 Dispersion curves of 100-m fiber sections spooled with zero strain (blue solid line) and 3% strain (red dotted line). The uncertainty is depicted by the error bars on the corresponding curves.
Consequently, the demonstrated dispersion resilience to the applied strain in the new HNLF enables Brillouin scattering management in fiber, even for distant-band parametric mixers which demand the highest level of local dispersion stability. The following section describes this important assertion.
3. Brillouin and dispersive characteristics under controllably varied tensile stress
The performance of the new HNLF with strain-based Brillouin management for wide-band parametric mixers is investigated by measuring Brillouin and dispersive characteristics of a stressed and an untreated fiber. The study compared two 50-m HNLF sections, where one of the fiber was strained to the profile shown in Fig. 8(a) during the spooling process.
Fig. 8 (a) Strain profile applied to the 50-m fiber under test. Right column shows the Brillouin scattering spectra of (b) the untreated fiber, and (c) the stretched fiber. Frequency-shifted Brillouin gain peaks are marked by red arrows in (c).
The specific staircase stress profile in Fig. 8(a) was chosen to impose the worst-case scenario expected in parametric mixer synthesis. In this scenario, the parametric interaction would experience abruptly varying phase mismatch in consecutive sections. In practical terms, the mixing process is being perturbed to the greatest extent allowable by the HNLF strain limit. In response to the applied strain profile, the narrow linewidth Brillouin gain spectrum (Fig. 8(b)) splits into five distinct peaks, spreading over an aggregate width of 715 MHz (Fig. 8(c)). The spread of Brillouin gain thus resulted in a dramatic improvement in the mixer (pump) power handling capacity.
Specifically, the Brillouin scattering threshold, which is defined by the input power level adequate to produce a 3-dB deviation in back-reflection power from the Rayleigh scattering contribution, was measured to be 26.7 dBm for the untreated fiber (Fig. 9(a) ). In contrast, the power transfer characteristics of the stressed HNLF type (Fig. 9(b)) shows no observable Brillouin scattering, at up to the power handling limit of the circulator. More importantly, in excellent agreement with the measured dispersion response to a constant strain described in the previous section, the global dispersion characteristics of the stressed and untreated HNLF were indistinguishable, as shown in Fig. 10 .
Fig. 9 Power transfer characteristics of (a) the untreated and (b) stretched fibers, showing the Brillouin scattering threshold. The dashed line in (a) shows the extrapolation of the Rayleigh-scattering contribution in the back-reflected power. The Brillouin scattering threshold is denoted as Pth in (a).
Fig. 10 Dispersion curves of the untreated (blue solid line) and stretched fibers (red dotted line). The uncertainty is depicted by the error bars on the corresponding curves.
In the final set of characterizations, the ZDW variations along the untreated and stressed HNLFs were measured using the setup shown in Fig. 3. The setup shown in Fig. 3 provided a resolution of the ZDW variations at better than 0.2 nm through a 40-THz wide pump-probe separation [19]. The conversion efficiency plots in Fig. 11 show that the dispersion shift due to fiber stretching contributed insignificantly to the overall ZDW fluctuations, amounted to 0.6 nm in both untreated and stressed fibers. Invariance of the phase-matching condition was also inferred by the similarity in the conversion efficiency levels.
Fig. 11 Conversion efficiencies against pump wavelength for (a) untreated fiber and (b) stretched fiber.
4. Continuous-wave coherent SWIR mixer synthesis
Finally, the excellent dispersion stability in the stressed new HNLF allowed for the first construction of a SWIR-band parametric mixer driven by CW pump, without using phase-dithering for Brillouin scattering suppression [22,23].
Figure 12 shows a demonstration experiment of the CW-SWIR mixer constructed with a 100-m section of the new fiber. A varying tension profile with a maximum strain of 2% was applied to the fiber to increase the Brillouin scattering threshold from 24 dBm to above 32 dBm. The mixer pump was seeded by a tunable laser with 100-kHz linewidth, and was amplified to 32 dBm at maximum. In this experiment, the pump wavelength was set to 1565 nm, which consequently produced a pair of conversion peaks centered at 1311 nm and 1941 nm. A signal wave matching the near-IR conversion peak was provided by a tunable laser with 5.6 dBm of output power. The polarization states of the pump and signal were aligned to each other as well as to one of the principal polarization states of the fiber in order to maximize the conversion efficiency. At the output, the SWIR-band idler was extracted using a fiberized band-splitter.
Fig. 12 CW coherent SWIR-band mixer setup. EDFA: erbium-doped fiber amplifier; TBPF: tunable band-pass filter; PC: polarization controller; WDMC: WDM coupler; OSA: optical spectrum analyzer.
The resultant output spectrum captured at the 10% output tap is shown in Fig. 13(a) . The benefit of adopting the negative-β4 phase-matching scheme is clearly demonstrated by the absence of parametric fluorescence noise, which was commonly observed in wide-band parametric mixers phase-matched by positive-β4 [24]. A high-resolution spectral measurement of the SWIR-band idler confirmed no spectral broadening, thus validating strict coherence preservation in the mixing process. Although the spectrum presented a conversion efficiency of −32 dB, the differential insertion loss of the NIR 10% coupler has contributed to an apparent efficiency loss of 15 dB. With the differential insertion loss taken into account, the conversion efficiency was measured against various pump power and plotted in Fig. 13(b). The efficiency plot in Fig. 13(b) depicted two distinguishable regimes of power transfer – the rate of efficiency change in the low-power regime (< 28 dBm) was steeper than the quadratic trend seen at pump powers above 28 dBm. The cause of the disparity was attributed to the reducing influence of dispersion fluctuations at increasing pump power levels, entailed by the broadening of the conversion window [16]. While the stressed HNLF allowed for Brillouin-free CW conversion at −19 dB efficiency, an untreated fiber could only provide −39-dB conversion efficiency. Considerable improvement in conversion efficiency can be expected with refinement of the fiber index profile for reaching higher waveguide nonlinearity [12].
Fig. 13 (a) Mixer output spectrum captured at 10% tap; (b) Conversion efficiency plotted against pump power levels measured at the fiber input. Inset in (a) shows 2-nm spectral slices of signal (red dashed) and idler (blue solid). OSA resolution bandwidth: 0.5 nm for full spectrum, 0.05 nm for zoom-ins.
5. Conclusion
We reported the characterization of a new type of HNLF with dispersion properties invariant to fabrication and mechanically introduced core geometry fluctuations. The new fiber was engineered to simultaneously provide high modal confinement and provide dispersion fluctuation resiliency in case when transversal profile is varied under fabrication or environmental influences. The new HNLF was measured to possess dispersive properties that do not change even under high (3% tensile) stress conditions corresponding to core compression in excess of 1.5%. This property led to a four-fold increase in Brillouin threshold by tensile straining while maintaining the phase-matching condition of a parametric mixer.
The new HNLF was subsequently used to demonstrate operation of the first CW-driven SWIR mixer capable of preserving long-wave idler coherence (linewidth). The latter was achieved without the need for pump dithering or pulsing [23,24], commonly used in conventional SWIR mixer devices.
Finally, the new design and fabrication methodology are validated and point to a clear direction toward local dispersion control that was not possible before. In addition to parametric mixers, the new approach holds particular promise in transmission fibers where management of nonlinear processes must be managed in a localized, rather than an averaged (distributed) manner.
References and links
1. S. Radic and C. J. McKinstrie, “Optical amplification and signal processing in highly nonlinear optical fiber,” IEICE Trans. Electron. 88(C), 859–869 (2005). [CrossRef]
2. M. A. Foster, A. C. Turner, J. E. Sharping, B. S. Schmidt, M. Lipson, and A. L. Gaeta, “Broad-band optical parametric gain on a silicon photonic chip,” Nature 441(7096), 960–963 (2006). [CrossRef] [PubMed]
3. M. Galili, J. Zu, H. C. Mulvadm, L. K. Oxenløwe, A. T. Clausen, P. Jeppesen, B. Luther-Davies, S. Madden, A. Rode, D.-Y. Choi, M. Pelusi, F. Luan, and B. J. Eggleton, “Breakthrough switching speed with an all-optical chalcogenide glass chip: 640 Gbits/s demultiplexing,” Opt. Express 17, 2182–2187 (2009).
4. B. P.-P. Kuo, E. Myslivets, A. O. J. Wiberg, S. Zlatanovic, C.-S. Brès, S. Moro, F. Gholami, A. Peric, N. Alic, and S. Radic, “Transmission of 640-Gb/s RZ-OOK Channel over 100-km SSMF by wavelength-transparent conjugation,” J. Lightwave Technol. 29(4), 516–523 (2011). [CrossRef]
5. A. O. J. Wiberg, C.-S. Brès, A. Danicic, E. Myslivets, and S. Radic, “Performance of self-seeded parametric multicasting of analog signal,” IEEE Photon. Technol. Lett. 23(21), 1570–1572 (2011). [CrossRef]
6. B. P.-P. Kuo, N. Alic, P. F. Wysocki, and S. Radic, “Simultaneous wavelength-swept generation in NIR and SWIR abdns over comined 329-nm band using swept-pump fiber optical parametric oscillator,” J. Lightwave Technol. 29(4), 410–416 (2011). [CrossRef]
7. R. E. Slusher, G. Lenz, J. Hodelin, J. Sanghera, L. B. Shaw, and I. D. Aggarwal, “Large Raman gain and nonlinear phase shifts in high-purity As2Se3 chalcogenide fibers,” J. Opt. Soc. Am. B 21(6), 1146–1155 (2004). [CrossRef]
8. H. Ebendorff-Heidepriem, P. Petropoulos, S. Asimakis, V. Finazzi, R. C. Moore, K. Frampton, F. Koizumi, D. J. Richardson, and T. M. Monro, “Bismuth glass holey fibers with high nonlinearity,” Opt. Express 12(21), 5082–5087 (2004). [CrossRef] [PubMed]
9. T. Okuno, T. Nakanishi, M. Hirano, and M. Onishi, “Practical considerations for the application of highly nonlinear fibers,” in Proc. OFC 2007, paper OTuJ1.
10. M. Hirano, T. Nakanishi, T. Okuno, and M. Onishi, “Silica-based highly nonlinear fiber and their application,” IEEE J. Sel. Top. Quantum Electron. 15(1), 103–113 (2009). [CrossRef]
11. S. Radic, “Parametric signal processing,” IEEE J. Sel. Top. Quantum Electron. 18(2), 670–680 (2012). [CrossRef]
12. B. P.-P. Kuo and S. Radic, “Highly nonlinear fiber with dispersive characteristic invariant to fabrication fluctuations,” Opt. Express 20(7), 7716–7725 (2012). [CrossRef] [PubMed]
13. E. Myslivets, N. Alic, J. R. Windmiller, and S. Radic, “A new class of high-resolution measurements of arbitrary-dispersion fibers: localization of four-photon mixing process,” J. Lightwave Technol. 27(3), 364–375 (2009). [CrossRef]
14. E. Myslivets, C. Lundström, J. M. Aparicio, S. Moro, A. O. J. Wiberg, C.-S. Brès, N. Alic, P. A. Andrekson, and S. Radic, “Spatial equalization of zero-dispersion wavelength profiles in nonlinear fibers,” IEEE Photon. Technol. Lett. 21(24), 1807–1809 (2009). [CrossRef]
15. M. Takahashi, M. Tadakuma, and T. Yagi, “Dispersion and Brillouin managed HNLFs by strain control techniques,” J. Lightwave Technol. 28(1), 59–64 (2010). [CrossRef]
16. M. Yu, C. J. McKinstrie, and G. P. Agrawal, “Modulational instabilities in dispersion-flattened fibers,” Phys. Rev. E Stat. Phys. Plasmas Fluids Relat. Interdiscip. Topics 52(1), 1072–1080 (1995). [CrossRef] [PubMed]
17. M. E. Marhic, K. K.-Y. Wong, and L. G. Kazovsky, “Wide-band tuning of the gain spectra of one-pump fiber optical parametric amplifiers,” IEEE J. Sel. Top. Quantum Electron. 10(5), 1133–1141 (2004). [CrossRef]
18. A. Boskovic, S. V. Chernikov, J. R. Taylor, L. Gruner-Nielsen, and O. A. Levring, “Direct continuous-wave measurement of n2 in various types of telecommincation fiber at 1.55 μm,” Opt. Lett. 21(24), 1966–1968 (1996). [CrossRef] [PubMed]
19. I. Brener, P. P. Mitra, D. D. Lee, D. J. Thomson, and D. L. Philen, “High-resolution zero-dispersion wavelength mapping in single-mode fiber,” Opt. Lett. 23(19), 1520–1522 (1998). [CrossRef] [PubMed]
20. J. M. Chávez Boggio and H. L. Fragnito, “Simple four-wave-mixing-based method for measuring the ratio between the third- and fourth-order dispersion in optical fibers,” J. Opt. Soc. Am. B 24(9), 2046–2054 (2007). [CrossRef]
21. M. Niklès, L. Thévenaz, and P. A. Robert, “Brillouin gain spectrum characterization in single-mode optical fibers,” J. Lightwave Technol. 15(10), 1842–1851 (1997). [CrossRef]
22. N. Yoshizawa and T. Imai, “Stimulated Brillouin scattering suppression by means of applying strain distribution to fiber with cabling,” J. Lightwave Technol. 11(10), 1518–1522 (1993). [CrossRef]
23. F. Gholami, S. Zlatanovic, E. Myslivets, S. Moro, B. P.-P. Kuo, C.-S. Brès, A. O. J. Wiberg, N. Alic, and S. Radic, “10 Gbps Parametric short-wave infrared transmitter,” in Proc. OFC 2011, paper OThC6.
24. J. M. Chávez-Boggio, S. Moro, B. P.-P. Kuo, N. Alic, B. Stossel, and S. Radic, “Tunable parametric all-fiber short-wavelength IR transmitter,” J. Lightwave Technol. 28(4), 443–447 (2010). [CrossRef]
