We report the first experimental demonstration of Bragg grating-based nonlinear switching in a bismuth-oxide single-mode fiber. Exploiting the strong χ(3)-nonlinearity of this fiber in a cross-phase modulation scheme, we change the transmission of a probe near the grating stop band from 90 % to 20 %, a 6.5 dB extinction ratio, at powers as low as 55 W. This is an 18-fold improvement in the switching power compared to the best demonstrations in silica. The experimental results agree well with numerical simulations.
© 2011 OSA
Nonlinear Bragg Gratings (BGs) are ideal for studying dynamic optical effects in nonlinear periodic media [1–6]. The stop band of a 1-D photonic crystal, in combination with a Kerr medium, leads to ultrafast (10s of femtoseconds), all-optical functions such as transmission (reflection) switching and power limiting [7, 8]. Nonlinear grating experiments are usually performed in silica glass fibers since the fabrication technology is well developed  and the nonlinear response is ultrafast. Though fast, the nonlinear response of silica is weak, leading to impractically high switching powers [4, 5, 8, 10]. Here we report the first nonlinear switching experiment in a BG written in Bi2O3, a glass with a strong, ultrafast nonlinear response. As a result, the switching threshold is lowered from over a 1 kW to approximately 50 W.
A transmission change from −35 dB to −15 dB in a 20 cm fiber BG at peak powers of 1.8 kW was demonstrated by Broderick et al . Although the extinction ratio was 20 dB, the absolute change in the transmissivity was 3 %. Subsequently, a 25 % change in the transmitted energy was achieved in an all-optical AND gate due to the creation of gap solitons  at peak powers above 2 kW. Exploiting a grating with a π phase-shift in a pump-probe experiment, Melloni et al demonstrated an extinction ratio of 6.5 dB with 1 kW peak power pulses . This result combines a significant extinction ratio with transmissions close to 100 %. Since phase-shifted gratings are designed to have low switching powers , the only path to additional, significant power reductions is to use highly-nonlinear media. For example, all-optical switching at powers of 100–700 W was reported in a silicon-on-isolator  and AlGaAs  waveguide gratings. However, the nonlinear threshold reduction in semiconductors comes at the expense of multi-photon absorption and/or somewhat slower response times due to free carrier generation.
Some non-silica glasses can also exhibit large, ultrafast nonlinear effects [14, 15]. For example, bismuth-oxide highly-nonlinear fiber (Bi-HNLF) has a nonlinear coefficient of γ = 1 W−1m−1 about 300 times that of silica. Additionally, negligible two-photon absorption, good mechanical, chemical and thermal durability make this glass attractive for nonlinear optical functions [15–17]. In recent years several demonstrations based on Bi-HNLF have been reported, including low-power supercontinuum generation , optical signal regeneration , time-division demultiplexing  and 2 × 2 crossbar switching in a Mach-Zehnder interferometer . Bragg gratings in Bi-HNLF would be expected to enable additional nonlinear effects at moderate powers [1, 3]. However, inscription of short-period gratings in high index bismuth-based fibers is challenging, since the traditional holographic techniques, developed for silica fibers, are not suitable. Recently, high-quality Bragg gratings were fabricated in Bi-HNLF using the phase mask technique and femtosecond infrared pulsed radiation .
In this paper we report the first experimental demonstration of nonlinear optical switching using such gratings. In our experiment the transmission of a weak CW probe is changed by sub-nanosecond strong pump pulses via cross-phase modulation (XPM). We vary the device transmissivity from 90 % to 20 % at pump powers of only 55 W. To our knowledge this represents the most efficient BG-based switch in the glass and becomes possible by the two breakthroughs mentioned earlier: the fabrication of highly-nonlinear bismuth-oxide fiber, and the development of a method to inscribe a Bragg grating into these fibers.
The paper is structured as follows: in Section 2 we briefly review the switching principle in Bragg gratings, in Section 3 we give characteristics of the Bi-HNLF and Bragg grating used in the experiment. In Section 4 we describe the experimental setup, techniques and the results we achieved. We discuss our results in Section 5 and complete this section with Conclusions.
2. Switching principle in Bragg gratings
Figure 1 illustrates the switching principle, which is similar to that of Larochelle et. al. . A BG of period Λ has a stop band centered at λ B = 2n 0Λ, where n 0 is the average refractive index. A strong pulse (a) at λ pump ≠ = λ B propagates through the BG, perturbing the refractive index in proportion to its peak power P. This leads to a fractional change Δk/k = n 2 P/n 0 A eff of the BG wave vector, and hence of the local stop band (b). Here n 2 is the nonlinear refractive index and A eff is the effective mode area. Thus, for the probe which co-propagates with the pump and is tuned close to the stop band, i.e. λ probe ≃ λ B, the gap is shifted by
where b depends on the relative polarization of pump and probe, and the factor of 2 accounts for the XPM effect. The output power of a CW probe (c) on the blue-side of the gap increases (e), since the stop band is detuned away by the pump pulse. However, for a probe on the red-side of the gap (d) the transmission drops (f) as the stop band is nonlinearly tuned towards it.
3. Bismuth-oxide nonlinear fiber and grating characteristics
In our experiment we use a Bi-HNLF (from Asahi Glass Co) which is single-mode at telecommunication wavelengths and has a step-index structure with a 1.9 μm core. The core and cladding refractive indices at 1550 nm are 2.21 and 2.11, respectively . The nonlinear co-efficient γ = 1 W−1m−1 as listed by the manufacturer, is consistent with our measurements of self-phase modulation induced spectral broadening of picosecond pulses. The nonlinear refractive index of Bi-HNLF is n 2 = 0.8 × 10−18 m2/W, originates from pure electronic polarization . Bi-HNLF has modest linear loss of −0.8 dB/m and shows no evidence of nonlinear loss according to our test measurements at the peak powers of interest.
A Gaussian apodized, 6 mm Bragg grating was inscribed in the Bi-HNLF using infrared 125 fs pulses and a silica phase mask . The mask was optimized for 800 nm radiation to produce structures with a pitch of Λ = 1.07 μm, leading to a 3rd-order resonance at λ B = 1539 nm. Figure 2 shows the transmission spectrum of the Bi-HNLF BG: the solid red line shows a JDS Uniphase Swept-Wavelength System measurement of the grating spectrum with a 3 pm resolution, whereas the dotted green line is a numerical fit based on an apodized, uniform grating. The resonance width is Δλ FWHM = 0.35 nm and the minimum transmission is T = −22 dB, corresponding to an estimated coupling strength of κ = 950 m−1. Fabry-Perot effects due to back reflections from the sample facets were suppressed by cleaving both fiber ends with a 10° angle. Nonetheless, a residual small ripple on the spectrum in Fig. 2 can still be noticed. Coupling into the Bi-HNLF was achieved by means of OzOptics tapered-end fibers with a tip-mode diameter of 2 μm, resulting in −4.5 dB loss per facet.
4. Experimental results
Figure 3 shows the experimental setup. The probe was an external cavity Photonetics laser which enabled tuning of CW with 10 pm steps in the C-band. The pump was a Nd:YAG microchip laser emitting 680 ps pulses at λ p = 1064 nm and 6.6 kHz repetition rate. The pump wavelength was far from the bandgap and hence the pump was not affected by the grating. After the probe was amplified with a Pritel erbium-doped fiber amplifier (EDFA), it was combined with the pump pulses in a 50/50 WDM coupler, and launched into the Bi-HNLF with BG. The output was filtered with highly-selective filter for the C-band and was monitored using an Agilent optical sampling oscilloscope with a resolution of 22 ps. The angle between the probe and pump polarization states was controlled using an in-fiber polarization controller (PC). In all results presented here the probe and pump were collinearly polarized to maximize the Kerr effect (b = 1 in Eq. (1)). At the pump wavelength λ p the Bi-HNLF supports several modes. By careful alignment we excited mainly the fundamental mode. Even if higher order modes carry a small fraction of the energy, it merely leads to a slight over-estimate of the pump peak power used in numerical simulations since the walk-off between the modes is negligible on the length scales of the experiment. Figure 4 shows oscilloscope traces of the normalized probe transmission (transmissivity) for the wavelengths on (a) the blue- (λ 1 = 1538.88 nm) and (b) the red-side (λ 2 = 1539.23 nm) of the gap as indicated by vertical lines in Fig. 2. Each trace corresponds to a different pump peak power coupled into the sample, in the range 17 – 55 W. We find that the switching ratio becomes larger with increasing pump peak power. This is expected, since the nonlinear shift is intensity-dependent according to Eq. (1). For the probe on the blue-side of the gap (Fig. 4 (a)), the transmission change is positive, since XPM tunes the bandgap away from the probe. The opposite happens on the red-side of the gap (Fig. 4 (b)), where the stop band is shifted towards the probe, and hence its transmission change is negative (see Fig. 1). At a power of 55 W we achieve off/on switching by +6.1 dB and on/off switching by −6.5 dB, corresponding to changes in the total transmission well over 60 % in both cases (see Fig. 5(a)). Corresponding numerical results of the probe dynamics are shown in (c) and (d). Simulations are similar to those in Ref.  and are based on solving the time-dependent CMEs. A small change in the steady-state amplitude in Fig. 4 (a) is attributed to thermal effects due to linear loss . However, these effects have a minor impact on our measurements since the introduced power variations are less than 10 %. Larger variations are expected for higher repetition rates, so that the grating then would require thermal stabilization.
Figure 5 (a) summarizes results from Fig. 4 (a)–(d) by showing the switching contrast versus coupled pump peak power: markers indicate measurement, whereas the solid and dotted line represent simulations. The switching contrast is calculated as the ratio of the grating transmissivity with and without pump pulse. Figure 5 (b) shows the measured switching contrast (dots) for different wavelengths across the bandgap at a fixed pump power of 55 W. The opposite signs of the extinction ratio on opposite sides of the gap was discussed earlier (Fig. 4), and is consistent with numerical modeling (solid red curve). Although the logarithmic transmission change is significant in the wavelengths region marked by the shadow in Fig. 5, the absolute power change is small and the instrument sensitivity precludes us from taking measurement here.
5. Discussion and conclusions
The experimental results presented in Figs. 4 and 5 are mutually consistent and are in agreement with numerical modeling. The largest measured extinction ratio is −6.5 dB (see Fig. 4 (b)), corresponding to a nonlinear shift of the Bragg resonance by approximately Δλ nl = 40 pm to longer wavelengths (Eq. (1)). The effective mode area of the Bi-HNLF at the pump wavelength λ pump = 1064 nm was calculated to be A eff = 1.68 μm2. No additional matching parameters were used. Our switching device length is only 6 mm, which is two orders of magnitude shorter than the bismuth-based nonlinear element in the interferometric setup in Ref. .
The grating in our experiment was not optimized for switching. Generally, switching off the steepest resonance minimizes the switching power. For example, longer gratings, which can be fabricated by scanning the femtosecond writing beam along the phase mask, have steeper band edges. Such devices would have correspondingly lower switching powers and higher isolation. Alternative, using phase-shifted BGs with a narrow transmission peak in the stop band, lowers the power requirements by a factor five or more .
However, the device bandwidth determines its speed and capacity, thus precluding the use of narrow resonances. Based on additional numerical simulations, not included here, we estimate the upper limit of the optical data signal to be half that of the device’s bandwidth, in our grating corresponding to 10–20 Gb/s data signals. Broader bandwidths and consequently faster signals are possible with stronger gratings. Photothermal effects  remain one of the main concerns for high-bit rate signal systems as they leads to a spectral drift of the device characteristics.
In conclusion, we have presented the first demonstration of BG-based all-optical switching in Bi-HNLF. We achieved extinction ratios of over 6 dB at pump peak powers of only ~ 50 W, at least 18-times lower compared with previous results in silica [8,10]. This was possible due to the strong intrinsic nonlinearity and high light confinement in fiber core. Thus, bismuth-based BGs are excellent geometries for studying nonlinear dynamic effects, and open new possibilities for signal processing applications.
This work was produced with the assistance of the Australian Research Council under the ARC Centers of Excellence program.
References and links
1. R. E. Slusher and B. J. Eggleton, Nonlinear photonic crystals, chapter 1 (Springer-Verlag, 2003).
3. B. J. Eggleton, C. M. de Sterke, and R. E. Slusher, “Nonlinear pulse propagation in Bragg grating,” J. Opt. Soc. Am. B 14, 2980–2993 (1997). [CrossRef]
4. D. Taverner, N. G. R. Broderick, D. J. Richardson, M. Ibsen, and R. I. Laming, “All-optical AND gate based on coupled gap-soliton formation in a fiber Bragg grating,” Opt. Lett. 23, 259–261 (1998). [CrossRef]
5. N. G. R. Broderick, D. J. Richardson, and M. Ibsen, “Nonlinear switching in a 20-cm fiber Bragg grating,” Optics Lett. 25, 536–538 (2000). [CrossRef]
6. I. V. Kabakova, B. Corcoran, J. A. Bolger, C. M. de Sterke, and B. J. Eggleton “All-optical self-switching in optimized phase-shifted fiber Bragg grating,” Opt. Express 16, 5083–5089 (2009). [CrossRef]
7. M. Scalora, J. P. Dowling, C. M. Bowden, and M. J. Bloemer “Optical limiting and switching of ultrashort pulses in nonlinear photonic band gap materials,” Phys. Rev. Lett. 73, 1368–1371 (1994). [CrossRef] [PubMed]
8. S. Larochelle, Y. Hibino, V. Mizrahi, and G.I. Stegeman, “All-optical switching of grating transmission using cross-phase modulation in optical fibers,” Electron. Lett. 26, 1459–1460 (1990). [CrossRef]
9. R. Kashyap, Fiber Bragg Gratings (San Diego, CA: Academic, 1999).
10. A. Melloni, M. Chinello, and M. Martinelli, “All-optical switching in phase-shifted fiber Bragg grating,” Photon. Tech. Lett. 12, 42–44 (2000). [CrossRef]
11. I. V. Kabakova, C. M. de Sterke, and B. J. Eggleton “Performance of field-enhanced optical switching in fiber Bragg gratings,” J. Opt. Soc. Am. B 27, 1343–1352 (2010). [CrossRef]
12. N. D. Sankey, D. F. Prelewitz, and T. G. Brown, “All-optical switching in a nonlinear periodic-waveguide structure,” Appl. Phys. Lett. 60, 1427–1429 (1992). [CrossRef]
13. P. Millar, R. M. De La Rue, T. F. Krauss, J. S. Aitchison, N. G. R. Broderick, and D. J. Richardson, “Nonlinear propagation effects in an AlGaAs Bragg grating filter,” Opt. Lett. 24, 685–687 (1999). [CrossRef]
14. H. C. Nguyen, D.-I. Yeom, E. C. Magi, B. T. Kuhlmey, C. M. de Sterke, and B. J. Eggleton, “Nonlinear switching using long-period gratings in As2Se3 chalcogenide fiber,” J. Opt. Soc. Am. B 25, 1393–1401 (2008). [CrossRef]
15. N. Sugimoto, H. Kanbara, K. Tanaka, Y. Shimizugawa, and K. Hirao, “Third-order optical nonlinearities and their ultrafast response on Bi2O3 - B2O3 - SiO2 glasses,” J. Opt. Soc. Am. B 16, 1904–1908 (1999). [CrossRef]
16. K. Kikuchi, K. Taira, and N. Sugimoto,“Highly nonlinear bismuth oxide-based glass fibers for all-optical signal processing,” Electron. Lett. 38, 166–167 (2002). [CrossRef]
17. F. Parmigiani, S. Asimakis, N. Sugimoto, F. Koizumi, P. Petropoulos, and D. J. Richardson, “2R regenerator based on a 2-m-long highly nonlinear bismuth oxide fiber,” Opt. Express 14, 5038–5044 (2006). [CrossRef] [PubMed]
18. J. T. Gopinath, H. M. Shen, H. Sotobayashl, E. P. Ippen, T. Hasegawa, T. Nagashima, and N. Sugimoto, “Highly nonlinear bismuth-oxide fiber for smooth supercontinuum generation at 1.5 μm,” Opt. Express 12, 5697 (2004). [CrossRef] [PubMed]
19. J. H. Lee, T. Tanemura, K. Kikuchi, T. Nagashima, T. Hasegawa, S. Ohara, and N. sugimoto, “Use of 1-m Bi2O3 nonlinear fiber for 160-Gbit/s optical time-division demultiplexing based on polarization rotation and a wavelength shift induced by cross-phase modulation,” Opt. Lett. 30, 1267–1269 (2005). [CrossRef] [PubMed]
20. P. Bakopoulos, O. Zouraraki, K Vyrsokinos, and H. Avramopoulos, “2x2 Echange/Bypass Switch Using 0.8 m of Highly Nonlinear Bismuth Oxide Fiber,” IEEE Photon. Tech. Lett. 19, 723–725 (2007). [CrossRef]
21. D. Grobnic, R. B. Walker, S. J. Mihailov, C. W. Smelser, and P. Lu, “Bragg Gratings Made in Highly Nonlinear Bismuth Oxide Fibers With Ultrafast IR Radiation,” IEEE Photon. Techn. Lett. 22, 124–126, (2010). [CrossRef]
22. J. Lauzon, S. Larochelle, and F. Ouelette, “Numerical analysis of all-optical switching of a fiber Bragg grating induced by a short-detuned pump pulse,” Opt. Commun. 92, 233–239 (1992). [CrossRef]