Open Access
Add to BibSonomyAbstract
We demonstrate a record-high extinction-ratio of 50.4 dB in a 2 × 2 silicon Mach-Zehnder switch equipped with a variable splitter as the front 3-dB splitter. The variable splitter is adjusted to compensate for the splitting-ratio mismatch between the front and rear 3-dB splitters. The high extinction ratio does not rely on waveguide crossings and meets a strong demand in applications to multiport circuit switches. Large fabrication tolerance will make the high extinction ratio compatible with a volume production with standard complementary metal-oxide semiconductor fabrication facilities.
© 2015 Optical Society of America
1. Introduction
Optical circuit switches, in the form of matrix switches, have been successfully integrated into silica planar lightwave circuits [1, 2]. Now, it is naturally expected that the compactness of these switches should be increased by using the silicon photonics technology [3–5]. For optical circuit switches, low crosstalk is a primary requirement. Thus, the 2 × 2 (crossbar) switch, which is an element of the matrix switch, must have a sufficiently high extinction ratio.
The limitation imposed on the extinction ratio will be explained here through discussion of a directional coupler (DC)-based standard 2 × 2 Mach-Zehnder (MZ) switch, as illustrated in Fig. 1. Port 1’ (2’) is a bar port to Port 1 (2), and Port 2’ (1’) is a cross port to Port 1 (2). When ϕ, the phase-difference between the two arms, is π, the input ports are connected to their corresponding bar ports, and we call this the bar state. However, when ϕ = 0, the input ports are connected to their corresponding cross ports and we call this the cross state. For coupled-mode analysis, we introduce a parameter, θ, through which the power splitting ratio is expressed as cos2θ:sin2θ (θ = π/4 for the ideal 3-dB coupler). In the bar state, the transmittance to a cross port (1-2’ or 2-1’) is sin2(θ1 – θ2). This optical leak to a cross port becomes completely extinct if the two couplers have an equal splitting ratio, θ1 = θ2, despite the fact that they are not equal to π/4. This is likely to happen under real fabrication conditions and is consistent with experimental experience, in that good extinction is often observed at a cross port in the bar state [6]. In a typical matrix-switch configuration, however, the vast majority of the element switches are set to the cross state. It is, therefore, essential to dim the bar port in the cross state. In the cross state, the transmittance to a bar port (1-1’ or 2-2’) is cos2(θ1 + θ2), while the condition for complete extinction is θ1 + θ2 = π/2. In other words, when the front coupler has a splitting ratio of 60:40, the rear coupler must have a ratio of 40:60. However, the fabrication of such complementary couplers is far from practical, particularly in terms of silicon photonics, where the tolerable dimensional error is too small.
Fig. 1 Structure of conventional 2 × 2 Mach-Zehnder switch. DC: directional coupler. PS: phase shifter. θ1 and θ2 are products of the coupling coefficients and coupling length. ϕ is the phase shift.
Superior approaches to achieving bar-port extinction in the cross state that have been reported to date include varying the port order [7] and cascading two MZ switches [2]. Both methods use a waveguide crossing and rely on sufficiently low crosstalk in the crossing; however, waveguide crossing improvement remains an ongoing research topic in silicon photonics research [8]. In the present study, we propose a novel approach that does not use waveguide crossings. Instead, a variable splitter is used as the front 3-dB splitter. As will be described below, it is possible to retrieve a high extinction ratio even for MZ switches with imperfect 3-dB couplers. This approach is also suitable for volume production.
2. Structure and analysis
Our approach to the high extinction ratio 2 × 2 optical switch is shown in Fig. 2 [9]. The primary structure is identical to that of a conventional 2 × 2 MZ switch, except that the 3-dB coupler at the front is replaced by another MZ interferometer. The added MZ interferometer functions as a variable splitter, and its splitting ratio is optimized to maximize the extinction ratio. A similar approach, in which both the front and rear couplers of the conventional MZ switch are replaced by MZ interferometers, has been proposed in a domestic patent in Japan [10]. However, there is in fact no need to replace both couplers with variable splitters, because replacing one variable splitter allows sufficient variability to match the splitting ratios of both couplers. This report is, to the best of our knowledge, the first implementation of extinction-ratio improvement in silicon photonics.
The proposed switch was theoretically analyzed using the finite element method (FEM) and the transfer matrix method. The effective indices of the DCs and the phase shifters (PSs) were calculated using the FEM, and their wavelength dependences were taken into account. The transfer matrices of the DC and the PS were based on the structural parameters of the fabricated device described in the next section, where thermooptic PSs were used. Figure 3(a) shows the bar-port output characteristics against the electric powers applied to the front and rear PSs (PS1 and PS2, respectively) at a wavelength of 1.55 μm. The vertical axis is normalized to an input power, and we can see that the output power dip increases in response to increases in the power applied to PS1. Note that, when PS2 = 0, the bar-port output power is not at a minimum, indicating that the switch is not in the cross state. If we make the switch asymmetric in order to impose an appropriate phase shift on PS2, we can achieve a complete cross state without any electric power. This is an important characteristic in terms of the N × N matrix switch application, and is used to decrease power consumption and simplify the control mechanism [11]. Figure 3(b) is a contour plot of the bar-port output power against the powers applied to PS1 and PS2. It is found that several points with minimum output power appear, depending on the conditions of the two PSs.
Fig. 3 (a) Calculated bar-port output characteristics against power applied to PS2. (b) Calculated contour plot of bar-port output power against power applied to both phase shifters.
Next, we analyzed the influence of variations in the DC splitting ratio on the bar-port output characteristics in the cross state. The bar-port transmission spectra of the conventional and the proposed MZ switches were calculated in the same manner, with the splitting ratios of the front and rear couplers set to the same values of 60:40, 50:50, and 40:60, as shown in Figs. 4(a), 4(b), and 4(c), respectively. The splitting ratios of 60:40 and 40:60 correspond to the fabrication error of ± 100 nm. As regards the conventional MZ switch, the dip wavelength is red-shifted or blue-shifted depending on the splitting-ratio variations, and a high extinction ratio cannot be achieved at a wavelength of 1.55 μm. On the other hand, in the proposed switch, the bar-port output power can be minimized at a wavelength of 1.55 μm by adjusting the two PSs as shown in Fig. 4, even if the splitting ratio differs from the ideal value of 50:50. In principle, any splitting-ratio mismatch, apart from 100:0, can be recovered. Note that the wavelength dependence of the extinction ratio comes from the wavelength-dependent splitting ratio of the directional coupler.
Fig. 4 (top) Calculated bar-port transmission spectra of conventional Mach-Zehnder switch (grey line) and proposed switch (dark line). (bottom) Applied power for cross state in proposed switch. The applied power for the cross state in the conventional switch is zero. The splitting ratios of the two couplers are: (a) 60:40, (b) 50:50, and (c) 40:60.
3. Fabricated device
Figure 5(a) shows a microscopic image of the proposed switch, which was fabricated on silicon-on-insulator wafer with a top silicon thickness of 220 nm using e-beam lithography and reactive ion-etching. Note that Si-wire waveguides are covered by 2-μm-thick SiO2 cladding. For operation in transverse-magnetic (TM)-like mode, the DC gap and Si-wire are designed to be 430 nm in width. The scanning electron microscope image of the fabricated device shown in Fig. 5(b), however, indicates that the fabricated gap and the Si-wire width were smaller than their designed values. This difference therefore changed the DC splitting ratio at a wavelength of 1.55 μm; this would result in degradation of the extinction ratio of the bar port in the conventional MZ switch, as has been discussed in the previous section. We prepared three kinds of design, in which the DC splitting ratios were 40:60 (coupling length: 4.5 μm), 50:50 (3.5 μm), and 60:40 (2.5 μm). The thermooptic PS, which was based on a platinum heater and a gold electric-wire, was patterned using a standard lift-off method and operated using an external electric power supply applied through electric probes. The light was coupled through spot-size converters located at the edges of the device chip.
Fig. 5 (a) Microscopic image of fabricated 2 × 2 switch. The heater is 5-μm wide, 0.1-μm thick, and 50-μm long. The heater resistance is ~60 Ω. (b) Scanning electron microscope image of directional coupler.
4. Results and discussion
The bar-port output characteristics of the fabricated device were evaluated using the experimental setup shown in Fig. 6. We used two light sources: a tunable laser diode for measuring the output power against the applied heater power, and a broadband light source (λ = 1.53–1.61 μm) for the measurement of the transmission spectra. One of the light sources was selected using an optical switch; the light was then adjusted to TM-like polarization and directed into the device. Here, the sub-mode suppression ratio of the tunable laser diode was over 50 dB, and the polarization extinction ratio at the input of the fabricated switch was also estimated to be over 50 dB ( = 30 dB (polarizer) + 20 dB (on-chip polarization beam splitter)). The heaters on the device were controlled using two source meters with electric probes, while the output light from the device was coupled to a lensed fiber and then evaluated using an optical power meter or an optical spectrum analyzer.
Fig. 6 Experimental setup for measuring bar-port output characteristics. Pol. ER: polarization extinction-ratio. SMSR: sub-mode suppression-ratio. PBS: polarization beam-splitter.
Figures 7(a) and (b) compare the bar-port output characteristics of the conventional MZ switch with those of the proposed switch at a wavelength of 1.55 μm. The two switches were fabricated on the same chip. In the conventional MZ switch, the extinction-ratio was limited to 21 dB because of a coupler splitting-ratio mismatch, which was caused by a fabrication error. On the other hand, even if the same fabrication error occurred in the proposed switch, the extinction ratio could be increased by adjusting the electric power applied to PS1. The extinction ratio reached 50.4 dB here; to the best of our knowledge, this is the highest extinction ratio obtained for the silicon photonics switch. Further, the experimental result shown in Fig. 7(b) agrees well with the analytical result shown in Fig. 3(a). The on-chip loss, which is defined as the insertion loss without the fiber-to-chip coupling loss, was ~0.5 dB. The switching power was ~30 mW, and the switching time was ~10 μs. We also confirmed that the extinction ratio of the cross port reached 50 dB. From another perspective, the proposed switch can be regarded as a low-crosstalk intersection, which is a basic component of integrated optical circuits. The typical crosstalk of an experimentally demonstrated silicon intersection is ~−40 dB [12–14], while a crosstalk intersection of approximately −50 dB can be achieved in the proposed device, although applied electric power is required.
Fig. 7 Bar-port output characteristics of (a) conventional Mach-Zehnder switch and (b) proposed switch.
Figure 8 shows the transmission spectra for designed DC splitting ratios of 60:40, 50:50, and 40:60. It should be noted that these spectra are in good agreement with the analytical results shown in Fig. 4, and that the extinction ratios of the three devices reach almost 50 dB through adjustment of the splitting ratios of the front couplers. The present switch uses no crossings, which is a significant advantage over the cascaded switch [2], because the extinction ratio of the cascaded switch is limited by the crosstalk performance of the intersection. These results indicate that a 100% production yield is possible through the use of the variable coupler. For a 30-dB extinction ratio, the transmission bandwidth is 3.1 nm.
Fig. 8 (Top) Bar-port transmission spectra of proposed switch. (Bottom) Electric powers applied to phase shifters for cross state. Designed coupler splitting ratio: (a) 60:40, (b) 50:50, and (c) 40:60.
We believe that the extinction ratio of the proposed switch is limited by the polarization extinction ratio. The proposed switch is designed for use in the TM-like mode. In this case, the transverse electric (TE)-like mode propagates through the DCs without coupling, because the coupling length for the TE-like mode is ~8.6 times longer than that of the TM-like mode, which appears from the bar port. In our experimental setup, the polarization extinction ratio at the switch input is approximately 50 dB, and this value agrees well with the experimental results.
5. Conclusion
We have presented a new 2 × 2 switch design based on the conventional MZ switch, in which the front coupler is replaced by a variable coupler. The proposed switch exhibited a 50.4-dB extinction ratio at a wavelength of 1.55 μm. Recovery of the extinction ratio using the variable coupler was also demonstrated, even in cases where the coupler splitting ratios differed from the ideal splitting ratio of 50:50. The proposed switch will enable a similarly high extinction ratio even for large-volume production scenarios using standard complementary metal-oxide semiconductor fabrication facilities. All the fabricated switches passed the performance test. Thus, this proposed switch will be of considerable benefit in large-volume production scenarios, to which silicon photonics best contributes.
Polarization and wavelength independence are challenges that must still be overcome in the field of silicon photonics. For polarization-insensitive operation, we believe that a diversity circuit [15] is a realistic option, because it does not require any complex three-dimensional structures. For wavelength independence, the use of a multi-mode interferometer coupler [16] or a wavelength-insensitive coupler [17] are potential solutions that require further investigation.
Acknowledgments
This work was partly supported by the Project for Developing Innovation Systems of MEXT, Japan. The authors are grateful to Mr. K. Tashiro for his technical assistance in device fabrication.
References and links
1. S. Sohma, T. Watanabe, N. Ooba, M. Itoh, T. Shibata, and H. Takahashi, “Silica-based PLC type 32 × 32 optical matrix switch,” in European Conference and Exhibition on Optical Communication, OSA Technical Digest (online) (Optical Society of America, 2006), paper Tu.4.4.3. [CrossRef]
2. T. Goh, A. Himeno, M. Okuno, H. Takahashi, and K. Hattori, “High-extinction ratio and low-loss silica-based 8 × 8 strictly nonblocking thermooptic matrix switch,” J. Lightwave Technol. 17(7), 1192–1199 (1999). [CrossRef]
3. S. Nakamura, S. Takahashi, M. Sakauchi, T. Hino, M. Yu, and G. Lo, “Wavelength selective switching with one-chip silicon photonic circuit including 8 × 8 matrix switch,” in Optical Fiber Communication Conference, OSA Technical Digest (online) (Optical Society of America, 2011), paper OTuM2. [CrossRef]
4. K. Suzuki, K. Tanizawa, T. Matsukawa, G. Cong, S.-H. Kim, S. Suda, M. Ohno, T. Chiba, H. Tadokoro, M. Yanagihara, Y. Igarashi, M. Masahara, S. Namiki, and H. Kawashima, “Ultra-compact 8 × 8 strictly-non-blocking Si-wire PILOSS switch,” Opt. Express 22(4), 3887–3894 (2014). [CrossRef] [PubMed]
5. L. Chen and Y.-K. Chen, “Compact, low-loss and low-power 8×8 broadband silicon optical switch,” Opt. Express 20(17), 18977–18985 (2012). [CrossRef] [PubMed]
6. Y. Shoji, K. Kintaka, S. Suda, H. Kawashima, T. Hasama, and H. Ishikawa, “Low-crosstalk 2 x 2 thermo-optic switch with silicon wire waveguides,” Opt. Express 18(9), 9071–9075 (2010). [CrossRef] [PubMed]
7. M. Okuno, K. Kato, R. Nagase, A. Himeno, Y. Ohmori, and M. Kawachi, “Silica-based 8 × 8 optical matrix switch integrating new switching units with large fabrication tolerance,” J. Lightwave Technol. 17(5), 771–781 (1999). [CrossRef]
8. S.-H. Kim, G. Cong, H. Kawashima, T. Hasama, and H. Ishikawa, “Tilted MMI crossings based on silicon wire waveguide,” Opt. Express 22(3), 2545–2552 (2014). [CrossRef] [PubMed]
9. K. Suzuki, G. Cong, K. Tanizawa, S.-H. Kim, S. Namiki, and H. Kawashima, “50-dB extinction-ratio in 2×2 silicon optical switch with variable splitter,” in Conference on Lasers and Electro-Optics, OSA Technical Digest (online) (Optical Society of America, 2014), paper SM4G.2.
10. M. Kawauchi, N. Takado, and K. Jinguji, “Waveguide type optical interferometer,” Patent abstract of Japan, 62–183406, 1987.
11. T. Goh, M. Yasu, K. Hattori, A. Himeno, M. Okuno, and Y. Ohmori, “Low loss and high extinction ratio strictly nonblocking 16 × 16 thermooptic matrix switch on 6-in wafer using silica-based planar lightwave circuit technology,” J. Lightwave Technol. 19(3), 371–379 (2001). [CrossRef]
12. W. Bogaerts, S. K. Selvaraja, P. Dumon, J. Brouckaert, K. De Vos, D. Van Thourhout, and R. Baets, “Silicon-on-insulator spectral filters fabricated with CMOS technology,” IEEE J. Sel. Top. Quantum Electron. 16(1), 33–44 (2010). [CrossRef]
13. P. Sanchis, P. Villalba, F. Cuesta, A. Håkansson, A. Griol, J. V. Galán, A. Brimont, and J. Martí, “Highly efficient crossing structure for silicon-on-insulator waveguides,” Opt. Lett. 34(18), 2760–2762 (2009). [CrossRef] [PubMed]
14. Z. Yi, Y. Shuyu, A. J. Lim, L. Guo-Qiang, C. Galland, T. Baehr-Jones, and M. Hochberg, “A CMOS-compatible, low-loss, and low-crosstalk silicon waveguide crossing,” IEEE Photon. Technol. Lett. 25(5), 422–425 (2013). [CrossRef]
15. S.-H. Kim, K. Tanizawa, Y. Shoji, G. Cong, K. Suzuki, K. Ikeda, H. Ishikawa, S. Namiki, and H. Kawashima, “Compact 2 × 2 polarization-diversity Si-wire switch,” Opt. Express 22(24), 29818–29826 (2014). [CrossRef] [PubMed]
16. K. Suzuki, H. C. Nguyen, T. Tamanuki, F. Shinobu, Y. Saito, Y. Sakai, and T. Baba, “Slow-light-based variable symbol-rate silicon photonics DQPSK receiver,” Opt. Express 20(4), 4796–4804 (2012). [CrossRef] [PubMed]
17. J. Van Campenhout, W. M. Green, S. Assefa, and Y. A. Vlasov, “Low-power, 2 x 2 silicon electro-optic switch with 110-nm bandwidth for broadband reconfigurable optical networks,” Opt. Express 17(26), 24020–24029 (2009). [CrossRef] [PubMed]
