We present the first demonstration of all optical wavelength conversion in chalcogenide glass fiber including system penalty measurements at 10 Gb/s. Our device is based on singlemode As2Se3 chalcogenide glass fiber which has the highest Kerr nonlinearity (n2) of any fiber to date for which either advanced all optical signal processing functions or system penalty measurements have been demonstrated. We achieve wavelength conversion via cross phase modulation over a 10 nm wavelength range near 1550 nm with 7 ps pulses at 2.1 W peak pump power in 1 meter of fiber, achieving only 1.4 dB excess system penalty. Analysis and comparison of the fundamental fiber parameters, including nonlinear coefficient, two-photon absorption coefficient and dispersion parameter with other nonlinear glasses shows that As2Se3 based devices show considerable promise for radically integrated nonlinear signal processing devices.
©2006 Optical Society of America
With the growing interest in high speed, agile optical networks, the need for practical, all-optical signal processing devices becomes imperative. Wavelength conversion is of particular interest in order to both efficiently utilize current network capacity by eliminating the wavelength continuity constraint  and to create reconfigurable networks.
Chalcogenide glasses have attracted significant attention as promising nonlinear materials for all-optical devices due to their large Kerr nonlinearity n 2 , which is critical for reducing the optical power requirements and footprint size of such devices. All optical switching has been demonstrated  in As2S3 chalcogenide glass fiber with an n 2 ~ 100 × silica whilst more recently higher nonlinearities have been achieved in As2S3 fiber with n 2 ~ 400 × silica [4, 5]. So far, however, all optical signal processing has not been reported in this fiber, which offers the promise of radically integrated and efficient nonlinear signal processing devices.
In this letter, we report the first demonstration of all optical wavelength conversion in any chalcogenide glass fiber. In addition, we report system bit error rate measurements at 10 Gb/s. These system measurements, in particular, are based on the highest Kerr nonlinearity (n 2) fiber to date (As2S3) for which such measurements have been reported. We demonstrate error-free wavelength conversion by cross-phase modulation (XPM) near 1550 nm over a 10 nm wavelength range with an excess system penalty of 1.4 dB in only 1 m of singlemode As2S3 fiber and 2.1 W of peak optical pump power. We find that the material parameters for an As2S3 device compares favorably against other nonlinear glasses, potentially allowing significantly smaller devices, operating with increased conversion bandwidth.
2. Device principle and experimental setup
Figure 1 summarizes the principle of a wavelength converter based on XPM . A continuous wave (CW) probe experiences XPM from co-propagating signal pump pulses, which generates optical sidebands. This is converted to amplitude modulation by using a band pass filter to select a single sideband. Importantly, the interaction length between the signal pump and the CW probe, i.e. pulse walk-off, is the limiting factor in terms of wavelength conversion range and not stringent phase matching conditions.
Figure 2 summarizes the experimental system setup used to demonstrate wavelength conversion. The 1 m length of high purity single mode As2S3 fiber used in our experiments was drawn at CorActive High-Tech Inc. The fiber had a core diameter of 6 μm, a core/cladding refractive index of 2.7 and a numerical aperture of 0.18 at 1550 nm, yielding an effective area of Aeff = 37 μm2. The nonlinear index of n 2 ~ 1.1×10-13 cm2/W (~400 × silica) results in a nonlinearity coefficient γ ~ 1200 W-1km-1. The fiber dispersion was measured using the differential phase shift method and found to be D = -560 ps nm-1km-1 at 1550 nm. The fiber ends were cleaved manually and butt-coupled to standard single mode fiber (mode field diameter = 9.5 μm) via ~5 mm of higher numerical aperture fiber (mode field diameter = 7.5 μm) to improve mode matching. Index matching (n = 1.578 at 589.3 nm) oil at the silica (n = 1.44)-As2S3 (n = 2.7) interface helped reduce, but could not completely eliminate, Fresnel losses. The total insertion loss was ~5.9 dB, of which ~1 dB was due to propagation loss, 0.6 dB due to splices to the higher numerical aperture fiber resulting in ~ 2.2 dB coupling loss per As2S3 fiber facet.
Under prolonged exposure to CW beams at a power threshold of between 150-300 mW, damage (sputtering and melting) was observed at the As2S3 fiber input facet. Damage was localized only to the facet area and  suggests that improvements in end face preparation (polishing and antireflection coatings ) would raise the power handling capability of the fiber. The relatively low damage threshold necessitated increasing the peak-to-average power ratio by using pulses equivalent to 40 Gb/s with a reduced repetition rate. The signal data was generated by encoding 40 Gb/s pseudo-random bit-stream (PRBS) data (231 - 1) from the bit error rate (BER) test bed system onto 7 ps pulses from an actively mode locked, fiber laser operating at 10 GHz. The excellent extinction ratio of the 10 GHz laser and the 40 Gb/s compatible pulses results in a simulated 40 Gb/s optical data stream where a ‘0’ is transmitted onto three out of every four bits. The receiver side calculates the BER appropriately.
The optical data was amplified and a 1.3 nm tunable band pass filter (TBF) was used to remove out-of-band amplified spontaneous emission (ASE). The pulses were combined with a CW probe from a wavelength tunable amplified laser diode and coupled into the As2S3 fiber. Pump-probe polarization alignment is necessary for efficient device operation . We utilized in-line polarization controllers with either a fiber based polarizer or the nonlinear XPM broadening itself to align the polarization state of the pump and probe. We did not observe any additional penalty beyond what is intrinsic to the process.
The output of the As2S3 fiber was then sent through a sharp 0.56 nm tunable grating filter offset to longer wavelengths by 0.55-0.70 nm to remove the pump and select a single XPM sideband. This amplified signal was filtered using a second 1.3 nm TBF to remove out of band ASE. An inline, 200 pm wide, fiber Bragg grating (FBG) notch filter was used to further suppress the residual CW carrier prior to the signal being measured with a 30 GHz optical bandwidth receiver.
3. System measurements
We demonstrated wavelength conversion over a 10 nm range using 12.3 dBm of probe and 18.6 dBm of pump average power, which corresponds to only 2.1 W peak power. Figure 3 shows the measured spectra of the As2S3 fiber output with the CW probe laser tuned to three wavelengths. The large residual CW component which is suppressed by the FBG is due to the low duty cycle (<12 dB) of the pulses, and would be greatly reduced in a full 33% return to zero (RZ) signal. The relative group velocity difference between the pump and probe (i.e. walk-off) reduces the amount of bandwidth generated through XPM as a function of wavelength offset. This is evident for the XPM sidebands for the probe at 1555.57 nm which corresponds to a walk-off length (L W = T 0/D Δλ, where T 0 is pulse width, D is the dispersion parameter, and Δλ is the pump-probe offset) of 0.35 m.
Figure 4 shows the BER vs. received optical power for the converter as well as for the back-to-back system and shows a penalty of ~1.4 dB for the three converted wavelengths at a BER = 10-9. The inset receiver traces shows a clear and open eye for the wavelength converted signals. Although no BER penalty was measured for increasing the wavelength offset, the effect of the walk-off reduced bandwidth leads to a longer converted pulse that could result in inter-symbol interference in a full 40 Gb/s 33% RZ signal.
The results reported here constitute the first system measurements for an As2S3 optical waveguide based device and the first wavelength conversion result for the entire range of chalcogenide glasses. To appreciate the potential impact of these results, it is instructive to compare this progress with other relevant instances from literature – in particular with the original XPM converter based on Silica dispersion shifted fiber [6, 9] and more recent developments  demonstrating 15 nm wavelength range at 160 Gb/s in bismuth oxide fiber. Table 1 summarizes the key fiber parameters for all three fiber compositions.
The significantly enhanced nonlinear index for As2S3 is a striking point – an order of magnitude greater than that of highly nonlinear Bi2O3. We note however that the nonlinearity coefficient γ = n 2 ω/(cAeff) for the As2S3 fiber is offset by a much larger effective core area of 37 um2 and therefore the fiber length used in both the Bi2O3 and As2S3 based device are the same (1 meter). While substantially shorter than the first demonstration in Silica DSF, 1 m is still excessive for integrated devices. Utilizing any of the three platforms of planar waveguides , fiber tapering  or micro-structured optical fiber  combined with the strong confinement provided by a large linear refractive index (2.7) would result in a greatly reduced effective core area and device length. While losses are expected to be low in fiber geometries, current propagation losses (0.2 dB/cm) in small core rib waveguides  suggest that chip based devices with interaction lengths of 10-20 cm are also conceivable. Small core silica  and Bi2O3  waveguides have already been largely demonstrated and further significant improvements in their nonlinear coefficients are doubtful.
As indicated earlier, the conversion bandwidth is governed by the necessity of ensuring the pump and probe interacts sufficiently – requiring that the device length is not much greater than the walk-off length. In fact this is broadly required for many pump-probe type devices (e.g. signal de-multiplexing and Raman modulation ). Whilst the dispersion parameter of As2S3 is high relative to Bi2O3, the potentially shorter device lengths just described make significant conversion bandwidth a possibility. For example, an effective core area of the As2S3 equivalent to the Bismuth Oxide fiber would result in γ= 11,100 W-1km-1. Only 12 cm of As2S3 fiber would be required for wavelength conversion allowing for over 40 nm of conversion bandwidth compared to only 10 nm for a 1 m Bismuth Oxide device before the walk off length became significant for 8 ps pulses. In addition, As2S3, like Bismuth Oxide, exhibits a relatively constant dispersion parameter D over the entire communications band , ensuring broadband operation of As2S3 based devices, in contrast with dispersion shifted silica fiber.
Last, the nonlinear figure of merit (FOM = n 2/λβ) of As2S3, which is a measure of the nonlinear-index relative to the two photon absorption coefficient (TPA - β), is relatively low compared to Silica and Bi2O3. While a low FOM has a strictly negative impact on nonlinear switching , recent work  has shown positive behavior for self phase modulation based pulse regenerators. Assuming that the device length is significantly shorter than the dispersion length, the threshold at which TPA affects device performance can be estimated by ignoring dispersive broadening effects and modeling the pump intensity required to generate a nonlinear phase shift cNL as a CW effect  using:
Here α is the linear absorption coefficient. We include pump-probe walk-off by setting the device length to L W. Nonlinear devices typically require φNL of order π, and Fig. 5(a) shows a plot of the required pump intensity as a function of probe-pump offset for varying FOM. As the wavelength offset increases, the length over which the phase shift must be imparted, L W decreases, requiring greater pump intensities. Eventually this results in a limit on the achievable conversion range. Figure 5(b) shows that the slope of Fig. 5(a) increases dramatically for FOM < 1, resulting in bandwidth limited devices. Similar analysis obtained by varying both the nonlinear phase shift and fiber dispersion also indicates that a FOM > 1 is needed for efficient operation. While further study is required, TPA is not expected to limit bandwidth of our device as the FOM for As2S3 is 2.3 .
We present the first demonstration of wavelength conversion in chalcogenide glass fiber as well as the first system bit error rate measurements in highly nonlinear As2S3 fiber. We achieve error-free wavelength conversion via cross phase modulation over a 10 nm with 1 m of fiber, with an excess system penalty of 1.4 dB at bit error rate (BER) = 10-9, at 10Gb/s using only 2.1 W of peak pump power. The device was based on the highest nonlinear-index coefficient n 2 fiber reported to date (As2S3) that has formed the basis of advanced signal processing functions or used for system BER measurements. Furthermore, we show that the modest levels of two-photon absorption in As2S3 is not a limitation to wavelength conversion. These results pave the way – through straightforward improvements such as mode-field size reduction – for chip based all-optical, nonlinear devices.
This work was produced with the assistance of the Australian Research Council (ARC). CUDOS (the Centre for Ultrahigh-bandwidth Devices for Optical Systems) is an ARC Centre of Excellence.
References and links
1. H. Zang, J. P. Jue, and B. Mukerjee, “A review of routing and wavelength assignment approaches for wavelength-routed Optical WDM Networks,” in Optical Networks Magazine(2000), pp. 47–60.
2. G. Lenz, J. Zimmermann, T. Katsufuji, M. E. Lines, H. Y. Hwang, S. Spalter, R. E. Slusher, S. W. Cheong, J. S. Sanghera, and I. D. Aggarwal, “Large Kerr effect in bulk Se-based chalcogenide glasses,” Opt. Lett. 25, 254–256 (2000). [CrossRef]
3. M. Asobe, “Nonlinear optical properties of chalcogenide glass fibers and their application to all-optical switching,” Opt. Fiber Technol. 3, 142–148 (1997). [CrossRef]
4. H. C. Nguyen, K. Finsterbusch, D. J. Moss, and B. J. Eggleton, “Dispersion in nonlinear figure of merit of As2S3 chalcogenide fibre,” Electron. Lett. 42, 571–572 (2006). [CrossRef]
5. L. B. Fu, M. Rochette, V. G. Ta'eed, D. J. Moss, and B. J. Eggleton, “Investigation of self-phase modulation based optical regeneration in single mode As2S3 chalcogenide glass fiber,” Opt. Express 13, 7637–7644 (2005). [CrossRef] [PubMed]
6. B. E. Olsson, P. Ohlen, L. Rau, and D. J. Blumenthal, “A simple and robust 40-Gb/s wavelength converter using fiber cross-phase modulation and optical filtering,” IEEE Photonics Technol. Lett. 12, 846–848 (2000). [CrossRef]
7. I. D. Aggarwal and J. S. Sanghera, “Development and applications of chalcogenide glass optical fibers at NRL,” J. Optoelectron. Adv. Mater. 4, 665–678 (2002).
9. P. Ohlen, B. E. Olsson, and D. J. Blumenthal, “Wavelength dependence and power requirements of a wavelength converter based on XPM in a dispersion-shifted optical fiber,” IEEE Photon. Technol. Lett. 12, 522–524 (2000). [CrossRef]
10. J. H. Lee, T. Nagashima, T. Hasegawa, S. Ohara, N. Sugimoto, and K. Kikuchi, “Wavelength conversion of 160 Gbit/s OTDM signal using bismuth oxide-based ultra-high nonlinearity fibre,” Electron. Lett. 41, 918–919 (2005). [CrossRef]
11. V. G. Ta'eed, M. Shokooh-Saremi, L. B. Fu, D. J. Moss, M. Rochette, I. C. M. Littler, B. J. Eggleton, Y. L. Ruan, and B. Luther-Davies “Integrated all-optical pulse regenerator in chalcogenide waveguides,” Opt. Lett. 30, 2900–2902 (2005). [CrossRef] [PubMed]
12. Y. K. Lize, E. C. Magi, V. G. Ta'eed, J. A. Bolger, P. Steinvurzel, and B. J. Eggleton, “Microstructured optical fiber photonic wires with subwavelength core diameter,” Opt. Express 12, 3209–3217 (2004). [CrossRef] [PubMed]
13. T. M. Monro, Y. D. West, D. W. Hewak, N. G. R. Broderick, and D. J. Richardson, “Chalcogenide holey fibres,” Electron. Lett. 36, 1998–2000 (2000). [CrossRef]
14. Y. L. Ruan, W. T. Li, R. Jarvis, N. Madsen, A. Rode, and B. Luther-Davies, “Fabrication and characterization of low loss rib chalcogenide waveguides made by dry etching,” Opt. Express 12, 5140–5145 (2004). [CrossRef] [PubMed]
15. J. H. Lee, Z. Yusoff, W. Belardi, M. Ibsen, T. M. Monro, and D. J. Richardson, “A tunable WDM wavelength converter based on cross-phase modulation effects in normal dispersion holey fiber,” IEEE Photon. Technol. Lett. 15, 437–439 (2003). [CrossRef]
16. G. Burdge, S. U. Alam, A. Grudinin, M. Durkin, M. Ibsen, I. Khrushchev, and I. White, “Ultrafast intensity modulation by Raman gain for all-optical in-fiber processing,” Opt. Lett. 23, 606–608 (1998). [CrossRef]
18. M. R. E. Lamont, M. Rochette, D. J. Moss, and B. J. Eggleton, “Two-photon absorption effects on self-phase-modulation-based 2R optical regeneration,” IEEE Photon. Technol. Lett. 18, 1185–1187 (2006). [CrossRef]