Open Access
Add to BibSonomyAbstract
A high speed 2×2 electro-optic switch using a polarization modulator (PolM) is proposed and experimentally demonstrated. In the proposed switch, two linearly polarized input lightwaves with orthogonal polarization directions are sent to the PolM which is connected to a polarization beam splitter (PBS). When a switching signal is applied to the PolM, the polarization directions of the two lightwaves at the output of PolM will exchange. Consequently, the lightwaves at the two output ports of the PBS would be switched, a 2×2 switch is thus realized. An optical switch with a crosstalk lower than -35 dB and a switching time less than 25 ps is experimentally demonstrated. The performance of the switch is also experimentally investigated by studying the bit error rates and eye diagrams. A technique to achieve polarization independent operation is also proposed and discussed.
©2007 Optical Society of America
1. Introduction
High speed optical switch is one of the key components in high speed, large capacity packet-switched optical networks, optical add-drop multiplexers (OADMs) and optical cross-connects (OXCs) [1]. Different technologies have been proposed and demonstrated to realize optical switches, such as micro-electromechanical-system- (MEMS) based switches [2], [3], acousto-optic switches [4], [5], thermal-optical switches [6], [7], electro-optic switches [8]–[11], and more recently, all-optical switches [12]–[13]. Among these technologies, the MEMS-based optical switches, thermal optical switches and the acousto-optic switches usually have a switching time longer than a few micro-seconds. In [9], an electro-optic switch was realized by using two birefringent crystals and an electrically controllable LiNbO3 crystal, which could act as a half-wave plate, to make the exchange of the two orthogonal polarization states under proper driving voltage. In [10]–[11], an electro-optic switch with similar configuration was demonstrated, in which the LiNbO3 crystal was replaced by a ferroelectric liquid-crystal, with a high driving voltage, to realize high-speed switching.
All-optical switches usually have ultra-fast switching speed, in which optical nonlinearity is usually used as the switching mechanism. However, at the current stage of development all-optical switches are usually complicated and costly. In addition, to enhance the nonlinear efficiency in an all-optical switch, the pump pulses need to be amplified by an optical amplifier, which may cause distortions. One solution to the pump pulse distortion problem is to send an inverted pulse light together with the pump to the optical amplifier [12], to ensure a constant gain of the optical amplifier. However, this method would lead to an increased complexity and decreased power conversion efficiency.
In this paper, we propose and demonstrate a new and simple high speed 2×2 electro-optic switch based on a recently developed high-speed electro-optic polarization modulator (PolM). In the proposed switch, two linearly polarized input lightwaves with orthogonal polarization directions are sent to the PolM that is connected to a polarization beam splitter (PBS). When a switching signal is applied to the PolM, the polarization directions of the two lightwaves at the output of the PolM would exchange such that the lightwaves at the two output ports of the PBS are switched, a 2×2 switch is thus realized. Our scheme shares a similar principle as that in [9]–[11], however, thanks to the use of a new and high-speed AlGaAs-GaAs PolM, the proposed switch would have a much higher switching speed. In addition, the use of the PolM provides a high potential for integration. An experiment to evaluate the proposed switch is implemented. The demonstrated switch is measured to have a crosstalk lower than -35 dB and a switching time less than 25 ps. The performance of the switch including the bit-error-rates (BERs) and eye diagrams is also experimentally investigated.
2. Principle
The schematic diagram of the proposed 2×2 electro-optic switch is shown in Fig. 1. The key component in the system is the PolM, which is an electrically tunable arbitrarily retarding, multiple-order wave plate developed by Versawave Technologies [14]. When a linearly polarized incident light is oriented with an angle of 45o to one principal axis of the PolM, the polarization state of the output lightwave would change from a linear polarization, to an orthogonal linear polarization, passing through elliptical and circular polarization states as the modulation voltage is varied by a half-wave voltage Vπ, polarization modulation is thus achieved. In Fig. 1, two linearly polarized input lightwaves, Input 1 and Input 2, are sent to the PolM, with polarization angles of 45o and 135o to one principal axis of the PolM. The polarization directions of the two input lightwaves are orthogonal to each other. When the switching signal is “off”, the two lightwaves at the output of the PolM keep the same polarization directions as those of the two input lightwaves. After the PBS, Input 1 and Input 2 are respectively directed to Output 1 and Output 2. Therefore, the switch is operating in the bar state. When the switching signal is “on”, thanks to the polarization modulation at the PolM, the polarization directions of the two output lightwaves would be rotated by 90° with respect to the two input lightwaves. Therefore, after the PBS, Input 1 is directed to Output 2 and Input 2 is directed to Output 1, a cross state of the switch is realized.
3. Experiment and Results
Fig. 2. The experimental setup of the high-speed 2×2 electro-optic switch; LD: laser diode, PC, polarization controller, PolM: polarization modulator, PMF: polarization maintaining fiber, PBS: polarization beam splitter.
The proposed 2×2 electro-optic switch is experimentally demonstrated. The experimental setup is shown in Fig. 2. Two electro-absorption-modulator- (EAM) integrated optical transmitters (TRs), TR1 and TR2, that emit two wavelengths at λ1=1542.94 nm and λ2=1544.53 nm, are used as the two input light sources with data modulation. The outputs from TR1 and TR2 are combined at an optical multiplexer (MUX) and then fed to the PolM. The polarization directions of the two lightwaves are adjusted to be 45o and 135o with respect to one principal axis of the PolM by using two fiber-based polarization controllers (PCs), PC1 and PC2. The PolM (Versawave Technologies) used in the experiment is an AlGaAs-GaAs 40-Gb/s mode-converter-based modulator that is able to operate in the wavelength range from 1530 to 1560 nm. The PolM has an electro-optic bandwidth up to 48 GHz, a polarization extinction ratio higher than 28 dB, and an insertion loss of about 4.4 dB. The output of the PolM is then sent to the PBS for polarization dependent wavelength distribution. By using a third PC, PC3, to align the principal axis of the PolM with that of the PBS, the switch is initially configured at the bar state. The total loss of the switch is measured to be about 8 dB.
In the experiment, we first investigate the dc operation of the optical switch. A dc driving voltage is applied to the PolM via its dc port. When the driving voltage is 0 V, the switch operates in the bar state, in which λ1 comes out from Output 1 and λ2 comes out from Output 2. The optical spectra measured at Output 1 and Output 2 are shown in Fig. 3(a) and 3(b), respectively. The crosstalk at Output 1 is about -37 dB and at Output 2 is about -39 dB. Here, the crosstalk is defined as the ratio of the leaked optical power to the signal optical power. Theoretically, the crosstalk should be equal to the polarization extinction ratio (PER) of the PolM with a polarization rotation angle of π on the Poincaré sphere. The PolM used in our experiment has different PER at different polarization rotation angles. At a polarization rotation angle of π, the PER is about -37 dB [14], which matches well with the experimental results. Then the switch is reconfigured to operate at the cross state by adjusting the driving voltage to be the half-wave voltage of the PolM, which is 15.9 V for the PolM working in dc state. Note that traveling wave electrodes are used in the PolM to match the velocity of the optical carrier and electrical modulation signal [14], the half-wave voltage for ac operation is much lower, which is 5.3 V for the PolM used in the experiment. At the output of the PolM, the polarization directions of the two input lightwaves are rotated by 90°, the two output lightwaves are thus switched at the two output ports of the PBS. The optical spectra measured at Output 1 and Output 2 are shown in Fig. 3(c) and 3(d), respectively. The crosstalk at Output 1 is about -38 dB and at Output 2 is about -39 dB.
Fig. 3. The output optical spectra of the optical switch; switch signal off: (a) Output 1, (b) Output 2; switch signal on: (c) Output 1, (d) Output 2.
Then, we measure the switching time of the optical switch. The measurement is performed by launching a CW lightwave into Input 1. A switching signal generated from a bit-error-rate tester (BERT) is applied to the PolM. We then measure the output signal at Output 2. The switching signal is a data sequence that has a repetition rate of 13 Gbit/s with a fixed pattern of one “1” and 15 “0”. The waveform of the switching signal is shown in Fig. 4 (dashed line). The rise time and fall time of the switching signal are measured to be 24 ps and 20 ps, respectively. The output signal at Output 2 is shown in Fig. 4 (solid line). The rise and fall time of the output pulse are measured to be 48 ps and 40 ps. Therefore, the switching time of the optical switch is calculated to be 24 ps and 20 ps at the rising and falling edges, respectively. The method to estimate the switching time may not be precise when the switch has a nonlinear response, such as a sinusoidal transfer function for the PolM used here. However, a sinusoidal transfer function can be considered linear for the range between 10% and 90% of the maximum output. Therefore, the method for switching time estimation used in this experiment would provide an acceptable accuracy.
Fig. 4. Switching time measurement; solid line: the switching signal; dashed line: the output pulse at Output 2.
For practical applications, N×N switches with N greater than 2 are usually needed. An N×N switch can be made by connecting multiple 2×2 switches with different topologies. To have a good performance, it is crucial that the individual 2×2 switch should have an excellent performance. In the experiment, the BERs and eye diagrams at the input and the output of the proposed 2×2 switch are measured. In the measurements, we apply a 1.25 Gbit/s 232-1 pseudo-random bit sequence (PRBS) to TR1 and TR2, and the output signal is measured at Output 1. Again, when the driving signal voltage to the PolM is 0 or 15.9 V, the PolM operates at the bar or cross state. The BER measurements for the back-to-back (B-to-B), bar and cross states are shown in Fig. 5. As can be seen the optical switch introduces a small power penalty of about 0.2 dB for both bar and cross states. The eye diagrams measured before and after the switch are shown in Fig. 6(a) and 6(b). No visible deterioration is observed.
4. Discussion
The proposed 2×2 electro-optic switch is polarization dependent, which is not desirable for most of the applications. The proposed switch, however, can be improved to be independent of the polarizations input lightwaves. A scheme that supports polarization-independent operation is shown in Fig. 7. As can be seen, two PBSs are used to split Input 1 and Input 2 into two orthogonal linear polarization states and the two polarization states are switched separately using two polarization-dependent switches (Switch1 and Switch2) that have the same configuration as that shown in Fig. 1. After Switch1 and Switch2, the two polarization states coming from the same input port (Input 1 or Input 2) are combined by two PBSs and sent to the two output ports. As can be seen, if the optical components (the fibers and the couplers) are all polarization maintaining components, the 2×2 electro-optic switch shown in Fig. 7 is polarization independent. The insertion loss of the polarization-independent switch is equal to the total loss of the two PBS and one 2×2 polarization dependent switch. The insertion loss of each PBS is 0.3 dB, the total insertion loss of the polarization independent switch is about 8.6 dB.
5. Conclusion
A high-speed 2×2 electro-optic switch using a polarization modulator (PolM) was proposed and experimentally demonstrated. The switch could operate at the bar or cross state by switching the modulation voltage to the PolM. The crosstalk and the switching time of the proposed switch were measured. For both the bar and cross states, the crosstalk was lower than -35 dB and the switching time was less than 25 ps. The BERs and eye-diagrams were also measured. A power penalty as low as 0.2 dB was achieved and the eye diagrams before and after the switch were widely open, no visible deterioration was observed. The proposed switch with an excellent performance was demonstrated. A scheme to improve the switch for polarization-independent operation was also proposed and discussed.
Acknowledgement
The work was supported by The Natural Sciences and Engineering Research Council of Canada (NSERC).
References and links
1. T. Shibata, M. Okuno, T. Goh, T. Watanabe, M. Yasu, M. Itoh, M. Ishii, Y. Hibino, A. Sugita, and A. Himeno, “Silica-based waveguide-type 16×16 optical switch module incorporating driving circuits,” IEEE Photon. Technol. Lett. 15, 1300–1302 (2003). [CrossRef]
2. V. A. Aksyuk, S. Arney, N. R. Basavanhally, D. J. Bishop, C. A. Bolle, C. C. Chang, R. Frahm, A. Gasparyan, J. V. Gates, R. George, C. R. Giles, J. Kim, P. R. Kolodner, T. M. Lee, D. T. Neilson, C. Nijander, C. J. Nuzman, M. Paczkowski, A. R. Papazian, R. Ryf, H. Shea, and M. E. Simon, “238×238 surface micromachined optical crossconnect with 2 dB maximum loss,” in OFC Tech. Dig. Series Anaheim, CA, 2002.
3. X. H. Ma and G. S. Kuo, “Optical switching technology comparison: optical MEMS vs. other technologies,” IEEE Commun. Mag. 41, S16–S23 (2003).
4. J. Sapriel, V. Molchanov, G. Aubin, and S. Gosselin, “Acousto-optic switch for telecommunication networks,” Proc. SPIE , 5828, 68–75 (2005). [CrossRef]
5. H. S. Park, K. Y. Song, S. H. Yun, and B. Y. Kim, “All-fiber wavelength-tunable acousto-optic switch,” in Proc. OFC’ 2001 , 3, WJ4-1–WJ4-3 (2001).
6. S. T. Feng and E. A. Irene, “Thermo-optical switching in Si based etalons,” J. Appl. Phys. 72, 3897–3903 (1992). [CrossRef]
7. R. Kasahara, M. Yanagisawa, T. Goh, A. Sugita, A. Himeno, M. Yasu, and S. Matsui, “New structure of silica-based planar lightwave circuits for low-power thermooptic switch and its application to 8×8 optical matrix switch,” J. Lightwave Technol. 20, 993–1000 (2002). [CrossRef]
8. A. Sneh, J. E. Zucker, and B. I. Miller, “Compact, low-crosstalk, and low-propagation-loss quantum-well Y-branch switches,” IEEE Photon. Technol. Lett. 8, 1644–1646 (1996). [CrossRef]
9. N. Wan, L. Liu, and Y. Yin, “Cantor network, control algorithm, two-dimensional compact structure and its optical implementation,” Appl. Opt. 34, 8176–8182 (1995) [CrossRef]
10. D. M. Maron and D. Mendlovic, “Compact all-optical bypass-exchange switch,” Appl. Opt. , 35, 248–253 (1996) [CrossRef]
11. N. Cohen, D. Mendlovic, B. Leibner, and N. Konforti, “Compact polarization-based all-optical interconnection systems with growth capability,” Appl. Opt. 37, 5479–5487 (1998) [CrossRef]
12. G. Berrettini, G. Meloni, A. Bogoni, and L. Poti, “All-optical 2×2 switch based on Kerr effect in highly nonlinear fiber for ultrafast applications,” IEEE Photon. Technol. Lett. 18, 2439–2441 (2006). [CrossRef]
13. Y. H. Kim, U. -C. Paek, and W. -T. Han, “All-optical 2×2 switching with two independent Yb3+-doped nonlinear optical fibers with a long-period fiber grating pair,” Appl. Opt. 44, 3051–3057 (2005). [CrossRef] [PubMed]
14. J. D. Bull, N. A. F. Jaeger, H. Kato, M. Fairburn, A. Reid, and P. Ghanipour, “40 GHz electro-optic polarization modulator for fiber optic communication systems,” in Proc. SPIE , 5577, 133–143 (2004). [CrossRef]
