We experimentally demonstrate all-optical broadcasting through simultaneous 7×40 Gb/s base-rate wavelength conversion in RZ format based on cross absorption modulation in an electroabsorption modulator. In this experiment the original intensity-modulated information is successfully duplicated onto seven wavelengths that comply with the ITU-T proposal. The advantages of the proposed wavelength conversion scheme are also discussed.
©2004 Optical Society of America
The electroabsorption modulator (EAM)  has proven to be a versatile component in ultra fast WDM and OTDM systems with its ability to perform several different functionalities combined with advantages such as simple structure, suitability for integration, low power consumption and environmental stability. Recently, various all-optical functionalities based on cross-absorption modulation (XAM) [2,3] have been demonstrated, such as demultiplexing , wavelength conversion [5,6] and all-optical regeneration .
As to wavelength conversion, investigations have so far focused on conversion of only one wavelength. In advanced WDM networks, however, applications like network broadcasting require a multi-wavelength optical data source where more than one converted wavelength is involved. In this paper we report, for the first time to the best of our knowledge, an EAM-based multiple wavelength conversion scheme where seven WDM channels are data-encoded simultaneously by a 40 Gb/s base-rate RZ signal through XAM in a single EAM. All the wavelength-converted channels, compliant with the ITU-T proposal on the WDM wavelength grid, show clear and open eyes with an error free bit error rate (BER) performance.
2. Experimental set-up
The experimental set-up is shown in Fig. 1. The outputs of seven commercial distributed feedback lasers (DFB-LD) at a wavelength spacing of 1.6 nm are combined in the arrayed-waveguide grating (AWG)1. These lasers operate at ITU-standardized wavelengths of 1549.31 nm, 1550.90 nm, 1552.52 nm, 1554.12 nm, 1557.33 nm, 1558.96 nm and 1560.60 nm. The combined signal of the seven channels is amplified to an average power of 17 dBm, and injected into the EAM from the right port. The control signal is generated by a commercial RZ transmitter, which basically consists of two Mach-Zehnder modulators (MZM). The first MZM, biased at the peak transmission point and driven by a 20 GHz RF signal, modulates a CW light beam from a tuneable laser operating at 1555.8 nm and generates a 40 GHz stream of pulses with widths of ~8 ps (corresponding to 33 % of the bit-period). The pulse stream is then launched into the second MZM and modulated by a pseudorandom bit sequence (PRBS) of length 215-1 from a 40 Gb/s pattern generator. The generated RZ signal at 40 Gb/s is amplified to an average power of 17.8 dBm, and injected into the EAM from the left port. Due to the counter-propagation scheme employed, an isolator (right side of the EAM) and an optical circulator (left side) are used. The 3rd output of the isolator is connected to another commercial AWG2, which allows selecting out each of the seven converted channels. A pre-amplified receiver is used for subsequent BER measurements.
The EAMs used in this paper are multiple-quantum-well (MQW) devices, which are 150 µm long and consist of 15 quantum wells.
3. Experimental procedures and results
In the experiment tapered fibres are used to couple light into and out of the EAM. The positions of the tapered fibres have to be optimised to obtain as high coupling efficiency as possible, which is essential for low optical power operation and high optical modulation efficiency. We optimise the coupling between the EAM and the tapered fibres at 1550 nm; then the fibre-to-fibre loss at zero bias is about 10 dB. The static transfer curve (transmittance as function of bias) is a quasi-linear curve up to about -2.5 V with a slope of -9 dB/V. The bias of EAM2 is set to -2 V in the experiment yielding a potentially high saturation induced extinction ratio (ER). The polarizations of the seven channels are optimized individually by using seven polarization controllers (PC). The polarization of the control signal is adjusted by its own PC and remains unchanged once optimised.
Figure 2 depicts the spectrum of the seven channels in three cases: (a) at the input of EAM, (b) at the output of EAM when the control signal is injected, and (c) at the output of EAM without control signal. It is clearly seen that in the presence of the control signal all channels are modulated due to XAM and each acquires a broadened spectrum as shown in Fig. 2 (b). In case of no control signal, we find new frequency components at the output of the EAM as shown in Fig. 2(c). The new components are the fingerprint of Four Wave Mixing (FWM) in the EAM; they are 30 dB weaker than original signals, hence their impact is negligible.
As an example, we show in Fig. 3 the measured bit-error-rates for the control signal and the converted channel at 1560.6 nm. The inserts of Fig. 3 show the spectrum and the eye diagram of the converted signal, as well as the eyes of the control signal. It can be seen that AWG2 filters out one wavelength with a side mode suppression ratio larger than 30 dB. The converted signal has clear and open eyes which are comfortably error free. The BER measurement shows that the converted signal at 1560 nm has no error floor but indicates a power penalty of 8 dB compared to the control signal. This penalty is mainly due to the reduced optical signal-to-noise ratio (OSNR) resulting from the insertion loss of the EAM, but also due to the decreased ER stemming from insufficient cross absorption modulation in the EAM. The conversion efficiency, defined as the ratio of the power of the input cw signal to the peak power of the converted pulses, is found to be around 17 dB.
The eye diagrams of the remaining six channels are displayed in Fig. 4, where the receiver sensitivity of all seven channels is depicted for an overview comparison. It is found that the other six channels also have clear and open eyes and are error free. We find that the wavelength near 1555 nm has the best receiver sensitivity or lowest power penalty (6 dB) as suggested by the valley shown in Fig. 4. This can be understood by the fact that EAM2 used in this experiment has the optimum performance near 1555 nm considering the tradeoff between the static ER and the insertion loss.
Compared to a nonlinear optical loop mirror (NOLM)-based solution , an EAM offers several advantages like potential for compactness, suitability for integration as well as stability against environmental conditions such as temperature drifts and vibrations.
The receiver sensitivity of the converted signals could be improved by reducing the insertion loss of the EAM. This would increase both the XAM efficiency and the OSNR by increasing the control signal power inside the EAM as well as the absolute power level at the output.
In our experiment we used the course WDM configuration with a channel spacing of 1.6 nm; if dense WDM configuration is chosen (e.g., 0.8 nm channel spacing), it is possible to further increase the number of converted signals at different wavelengths.
According to our previous chirp measurement for EAM-based wavelength conversion , the converted signal may have lower or negative chirp compared to the control signal, which is desirable for a data source to be used in long distance transmission.
We have experimentally demonstrated all-optical broadcasting through 7×40 Gb/s wavelength conversion in the RZ modulation format using the XAM effect in an EAM. Seven wavelengths, compliant with the ITU-T proposal (1549–1560 nm), are modulated simultaneously by a single control signal at 1555.8 nm. All the converted signals show clear and open eyes with error free operation.
Our results suggest that an EAM is a promising component for all-optical processing in coming photonic networks.
This work was supported by the Danish Technical Research Council through the SCOOP project (Semiconductor COmponents for Optical signal Processing).
References and links
1. T. H. Wood, “Multiple Quantum Well (MQW) Waveguide Modulators,” J. Lightwave Technol. 6, 743–757 (1988). [CrossRef]
2. T. H. Wood, J. Z. Pastalan, C. A. Burrus JR., B. C. Johnson, B. I. Miller, J. L. Demiguel, U. Koren, and M. G. Young, “Electric field screening by photogenerated holes in multiple quantum wells: A new mechanism for absorption saturation,” Appl. Phys. Lett. 57, 1081–1083 (1990). [CrossRef]
3. M. Suzuki, H. Tanaka, and S. Akiba, “Effect of hole pile-up at heterointerface on modulation voltage in GaInAsP electroabsorption modulators,” Electron. Lett. 25, no. 2, 88–89 (1989). [CrossRef]
4. L. K. Oxenløwe, E. Hilliger, A. Tersigni, A. M. Nik, S. Højfeldt, K. Yvind, P. M. W. Skovgaard, K. Hoppe, and J. Hanberg, “All-optical Demultiplexing and Wavelength conversion in an Electroabsorption Modulator,” Proceedings of ECOC 2001, paper Th.B.2.5 (2001).
5. L. Xu, L.K. Oxenløwe, N. Chi, J. Mørk, P. Jeppesen, K. Hoppe, and J. Hanberg, “Experimental characterisation of wavelength conversion at 40Gb/s based on electroabsorption modulators,” IEEE Lasers and Electro-Optics Society 2002, paper MM3 (2002). [CrossRef]
6. N. Edagawa, M. Suzuki, and S. Yamanoto, “Novel wavelength converter using an electroabsorption modulator,” IEICE Trans. Electron. E81-C, 1251–2157 (1998).
7. E. S. Awad, P. S. Cho, C. Richardson, N. Moulton, and J. Goldhar, “Optical 3R regeneration with all-optical timing extraction and simultaneous wavelength conversion using a single Electro-Absorption Modulator,” Proceedings of ECOC 2002, paper 6.3.2 (2002).
8. J.J. Yu, X.Y. Zheng, F.H. Liu, C. Peucheret, A.T. Clausen, H.N. Poulsen, and P. Jeppsen, “8×40 Gb/s 55-km WDM Transmission over Conventional Fiber Using a New RZ Optical Source,” IEEE Photon. Technol. Lett. 12, 912–914 (2000). [CrossRef]
9. L. Xu, L.K. Oxenløwe, N. Chi, F. P. Romstad, K. Yvind, J. Mørk, P. Jeppsen, K. Hoppe, and J. Hanberg, “Bandwidth and chirp characterisation of wavelength conversion based on electroabsorption modulators,” Proceedings of ECOC 2002, paper P1.26 (2002).