Open Access
Add to BibSonomy
Abstract
We propose a thermo-optic switch based on a symmetric directional coupler formed with two parallel identical two-mode waveguides, where the two modes in one waveguide can be simultaneously switched to the corresponding modes in the other waveguide. We design and fabricate such a device with polymer materials. Our fabricated device has a total length of 22.5 mm and operates at a switching power of 128 mW. The extinction ratios measured across the C-band are higher than ∼18 dB and ∼13 dB for the fundamental mode and the higher-order mode, respectively. The switching time is ∼1 ms. The performance of the device is insensitive to the polarization state of light. Our proposed mode-independent switch could find applications in reconfigurable mode-division-multiplexing transmission systems.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Directional couplers (DCs) formed with two parallel waveguides are basic structures widely used for the construction of photonic integrated devices, such as power splitters, optical switches, wavelength filters, and polarization beam splitters (see, for example, [1–4]). For a DC formed with two identical waveguides (a symmetric DC), the modes of the same order in the two waveguides are always phase-matched, which allows periodic power transfer between the two modes along the DC. For a DC formed with two non-identical waveguides (an asymmetric DC), the modes of the same order in the two waveguides are not phase-matched, which can prohibit strong coupling between the two modes (unless the two cores of the waveguides have different refractive indices [5]). However, it is possible to design an asymmetric DC in such a way that a low-order mode in the smaller waveguide is phase-matched with a high-order mode in the larger waveguide and thus achieve strong coupling between two modes of different orders. In recent years, many mode (de)multiplexers [5–13] and switches [14,15] based on asymmetric DCs, including three-dimensional DCs [5,9–13,15], have been demonstrated for mode-division-multiplexing (MDM) optical communication where different spatial modes of a few-mode fiber carry different signal channels. MDM is emerging as a promising technology to increase the transmission capacity of conventional single-mode fiber communication systems [16–18]. There is a need to develop various kinds of mode-controlling devices for MDM systems. Apart from mode-selective devices, such as the mode (de)multiplexers and switches mentioned above, mode-independent devices are also required. There are, however, few mode-independent waveguide devices available. In this paper, we demonstrate a mode-independent optical switch based on a symmetric DC formed with two identical two-mode waveguides, where the two modes in one waveguide can be simultaneously switched to the corresponding modes in the other waveguide by means of the thermo-optic effect.
A conventional symmetric DC is formed with two single-mode waveguides and characterized by a coupling length, which is the minimum length required for complete power transfer between the two waveguides. As our proposed symmetric DC is formed with two two-mode waveguides, two coupling lengths are present, with the longer one for the fundamental mode and the shorter one for the higher-order mode. For both modes to couple completely from one waveguide to the other waveguide, the length of the DC must be equal to an odd multiple of each of the two coupling lengths. In the present study, we design and fabricate such a mode-independent DC with polymer materials. To achieve simultaneous switching of the two modes, we deposit electrode heaters on the two waveguides. When both heaters are turned off, the DC functions as a mode-independent coupler. When one of the heaters is turned on, the symmetry of the DC is broken, i.e., the DC becomes an asymmetric one, so that the two modes stay in the waveguide where they are launched. In other words, the thermo-optic effect serves to deactivate the operation of the DC for both modes. The use of polymer materials can offer a large thermo-optic effect and also make easy fiber connection [15,19,20]. Our fabricated device has a total length of 22.5 mm and operates at a switching power of 128 mW. The extinction ratios measured across the C-band (1530–1565 nm) are higher than ∼18 dB and ∼13 dB for the fundamental mode and the higher-order mode, respectively, and the switching time is ∼1 ms. The switching characteristics of the device are polarization-insensitive. The mode-dependent loss and the polarization-dependent loss of the device measured at the wavelength 1550 nm are ∼0.6 dB and ∼0.2 dB, respectively. This polarization-insensitive mode-independent switch could find applications in reconfigurable MDM transmission systems, where modes of different orders need to be switched together between two few-mode fibers.
2. Device structure and design
Figure 1(a) shows a schematic diagram of our proposed mode-independent thermo-optic switch, which is a symmetric DC formed with two parallel two-mode waveguide cores (Core 1 and Core 2). Symmetric S-bends are applied to the DC to separate the two cores at the two ends. Two identical electrode heaters (Heater 1 and Heater 2) are deposited, respectively, on the two waveguides to ensure perfect symmetry of the DC, though only one heater is actually needed for the operation of the switch. Each two-mode waveguide supports two spatial modes, the E11 and E21 modes, which correspond to the LP01 and LP11a modes of an optical fiber, respectively. The intensity patterns of the two guided spatial modes are also shown in Fig. 1(a). Each spatial mode consists of two polarizations: the quasi-transverse-electric (TE) polarization for the mode with the major electric-field component parallel to the waveguide surface and the quasi-transverse-magnetic (TM) polarization for the mode with the major electric-field component perpendicular to the waveguide surface. When the refractive-index difference between the core and the cladding is small, as in our case, the two polarization modes are almost degenerate. We analyze the modes of the device with a full-vector finite-element method (COMSOL). The refractive indices of the core and the cladding material are 1.5690 and 1.5590, respectively, which are the refractive indices of the polymer materials used in our experimental device.
Fig. 1. (a) Schematic diagram of the proposed mode-independent thermo-optic switch and (b) cross-sectional view of the device in the coupling region, which shows the waveguide dimensions.
As each waveguide supports two modes, the DC is characterized by two coupling lengths. Because the fundamental mode is more confined in the waveguide core than the higher-order mode, the evanescent field of the fundamental mode is weaker than that of the higher-order mode. As a result, the coupling length for the fundamental mode is longer than that of the higher-order mode. For a mode to couple completely from one waveguide to the other waveguide, the length of the DC must be equal to an odd number of the coupling length for that mode. Therefore, for both modes to couple completely from one waveguide to the other waveguide, the length of the DC must be equal to an odd multiple of each of the coupling length. Here, we present a specific design to demonstrate the idea.
The dimensions of the waveguides are shown in Fig. 1(b). The heights and the widths of both waveguide cores are fixed at 5.0 µm and 9.0 µm, respectively, to ensure that the waveguide supports two modes in the C-band. The distances of the cores from the substrate and the electrode heaters are both set at 13 µm, which is large enough to ignore the effects of the substrate and the electrode heaters on the mode-field distributions. The effective indices of the E11 and E21 modes supported by each waveguide calculated at the wavelength 1550 nm are 1.5646 and 1.5610, respectively.
We consider the case that both modes are launched into Core 1 and both heaters are turned off (i.e., no electric power is applied to either heater). The core separation S and the length of the parallel section of the coupler (i.e., the coupler length) are treated as variables. For a given set of coupler parameters, we calculate the normalized output powers of the two modes from both cores with a 3D finite-difference beam propagation method (3DFD-BPM, Rsoft). With intensive numerical simulation, we finally choose a value of 3.9 µm for the core separation and a value of 5962 µm for the coupler length. Figures 2(a) and 2(b) show how the normalized output powers of the two modes from the two cores vary with the core separation at 1550 nm for the TE and TM polarizations, respectively, calculated for a coupler length of 5962 µm. As shown in Fig. 2, when the core separation is equal to 3.9 µm, both modes launched into Core 1 are completely coupled to the corresponding modes in Core 2, which is insensitive to the polarization. At a core separation of 3.9 µm, the coupling lengths for the E11 and E21 modes are equal to 5962 µm and 1193 µm, respectively, so the coupler length chosen (5962 µm) is actually 1 and 5 times of the coupling lengths for the E11 and E21 modes, respectively, as expected from the operation principle of the DC.
Fig. 2. Variations of the normalized output powers of the two modes from Core 1 and Core 2 at the wavelength 1550 nm with the core separation for the (a) TE and (b) TM polarizations, when both modes are launched into Core 1 and both heaters are turned off.
When one of the heaters is turned on (by applying sufficient electric power to the heater), the refractive index of the corresponding core decreases because of the thermo-optic effect. As a result, the phase-matching conditions for both modes are no longer satisfied, i.e., the symmetric DC is turned into an asymmetric one, and both modes launched into Core 1 should mainly stay in Core 1. In other words, the thermo-optic effect deactivates the operation of the DC and thus achieves mode-independent switching. The layout of the electrode heaters is shown in Fig. 1. The two electrodes have the same width of 16 µm and the same length of 12 mm, which covers the entire coupler length and part of the S-bends. To evaluate the thermo-optic effect, we set the thermo-optic coefficient of the polymer materials at 1 × 10−4 /°C and calculate the variations of the normalized output optical powers of different modes with the electric power applied to one of the heaters with the built-in electrode-heater model in the 3DFD-BPM. The switching dynamics calculated at the wavelength 1550 nm for the TE and TM polarizations are shown in Figs. 3(a) and 3(b), respectively, where electric power is applied only to Heater 1. As shown in Fig. 3, the switching power required for the device to operate as a mode-independent switch is ∼90 mW, which is polarization-insensitive. We obtain almost the same switching dynamics (not shown) with electric power applied only to Heater 2. The electric power 90 mW can be taken as the power required for turning on a heater.
Fig. 3. Variations of the normalized output powers of the two modes from Core 1 and Core 2 at the wavelength 1550 nm with the electric power applied only to Heater 1 for the (a) TE and (b) TM polarizations, when both modes are launched into Core 1.
To evaluate the performance of the device, we assume that each mode is launched only into Core 1 and calculate the coupling ratio and the extinction ratio for that mode by the 3DFD-BPM. The coupling ratio for a mode, denoted as CR, is defined as the output optical power from Core 2, P2, over the total output power from both cores, Ptotal, obtained when both heaters are turned off:
Figure 4(a) shows the calculated coupling ratios for the E11 and E21 modes. Our design is optimized at 1550 nm, where the coupling ratios reach 100%. In the C-band, the coupling ratios for the E11 and E21 modes are higher than 99.2% (99.3%) and 90.5% (90.3%), respectively, for the TE (TM) polarization. The device functions as an effective mode-independent coupler when both heaters are turned off.Fig. 4. (a) Coupling ratios and (b) extinction ratios for the E11 and E21 modes calculated for the TE and TM polarizations.
The extinction ratio for a mode, denoted as ER, is defined as the ratio of the output powers from Core 2 obtained when one heater is turned off and on, respectively (with the other heater always turned off), which are denoted as P2 |OFF and P2 |ON, respectively:
3. Device fabrication
We followed the design parameters given in the previous section to prepare the waveguide and electrode masks and fabricated the device by using our in-house microfabrication process. We used the polymer materials EpoCore and EpoClad (Micro Resist Technology GmbH) as the core and the cladding material, respectively. The refractive indices of these two materials, measured with a prism coupler (Metricon 2010) on thin-film samples at 1536 nm wavelength, were 1.5690 and 1.5590, respectively. The refractive-index difference between the TE and TM polarizations was smaller than 0.001 for both materials, so the material birefringence was small. The fabrication process, which involves fives steps, is illustrated in Fig. 5. In the first step, EpoClad was spin-coated onto an O2-plasma treated silicon substrate and cured to form a 13-µm-thick lower cladding. In the second step, the core layer was formed by spin-coating EpoCore onto the cured lower cladding and applying photolithography with a mask that defined the core pattern in the lower layer. The height of the core was trimmed to 5.0 µm by reactive-ion etching (RIE) after curing. In the third step, EpoClad was spin-coated onto the core layer and cured to form a 13-µm-thick upper cladding. Finally, a 100-nm thick aluminum was deposited on the sample by thermal evaporation and the electrode heaters were formed by applying photolithography and wet-etching processes. The total length of the device was about 22.5 mm, which included the waveguide leads at both ends. A photograph of the fabricated device and a magnified view of the heaters are shown in Figs. 6(a) and 6(b), respectively, and a microscopic image of an end face of the device is shown in Fig. 6(c). Microscopic images of a top view and a cross-sectional view of the cores in the coupling region are shown in Fig. 6(d).
Fig. 6. (a) Photograph of a fabricated device with (b) a magnified view of the heaters; microscopic images of (c) an end face of the device and (d) a top view and a cross-sectional view of the cores in the coupling region.
4. Measurements and discussions
To characterize the performances of the fabricated device, we launched light from a pigtailed tunable laser (KEYSIGHT) into Core 1 via a lensed fiber and measured the output powers from Core 2 with a power meter (Newport 2832-C) via a 40X objective lens. We used polarization optics to control the polarization state of the input light and monitor the polarization state of the output light. We also captured the output near-field images of the device with an infrared camera. We did the measurements at the room temperature. The E11 and E21 modes were selectively launched into Core 1 by properly controlling the position and the tilt angle of the laser beam incident into the core via the lensed fiber with the help of a five-degree micro-positioner. Figures 7(a) and 7(b) show the output near-field images taken at the wavelength 1550 nm with different electric powers applied to Heater 1 (with Heater 2 turned off) for the TE and TM polarizations, respectively. As shown in Fig. 7, when the heater is off, i.e., at 0 mW, the E11 and E21 modes launched into Core 1 are almost completely coupled to the corresponding modes in Core 2. As the electric power applied to the heater increases, the amounts of the E11 and E21 modes coupled to Core 2 decrease. For the E11 mode, the coupling almost stops at an electric power larger than ∼93 mW, while for the E21 mode, the coupling almost stops at an electric power larger than ∼128 mW. Therefore, the electric power ∼128 mW can be considered to be the switching power of the device for mode-independent operation. Figure 8 shows the variations of the output powers from Core 1 normalized to the corresponding total output powers with the electric power applied to the heater, when the two modes were selectively launched into Core 1.
Fig. 7. Output near-field images taken at the wavelength 1550 nm with different electric powers applied to Heater 1, when (a) TE-polarized and (b) TM-polarized E11 and E21 modes were launched into Core 1, respectively.
Fig. 8. Variations of the normalized output powers of the E11 and E21 modes from Core 1 at the wavelength 1550 nm with the electric power applied to Heater 1 measured for the TE and TM polarizations.
To evaluate the coupling ratios and the extinction ratios of the device from Eqs. (1) and (2), we launched the E11 and E21 modes into Core 1, respectively, and measured the output powers form Core 1 and Core 2 with Heater 1 turned off and on (and Heater 2 always turned off). To characterize the performance of device, the applied electric power was set 128.2 mW. Figure 9(a) shows the measured coupling ratios for the two modes. In the C-band, the coupling ratios for the E11 and E21 modes are higher than 98.2% (98.1%) and 90.1% (89.4%) for the TE (TM) polarization, respectively. Figure 9(b) shows the measured extinction ratios for the two modes. In the C-band, the extinction ratios for the E11 and E21 modes are higher than 18.3 dB (18.4 dB) and 13.4 dB (13.6 dB) for the TE (TM) polarization. Thanks to the geometry of the cores and the small refractive-index difference between the core and the cladding material, the characteristics of the device are insensitive to the polarization. The measurement results compare well with the simulation results shown in Fig. 4. While the E11 and E21 modes can be preferentially excited to a high purity by controlling the launching conditions, as shown by the clean near-field mode patterns in Fig. 7, we cannot quantify the purity of the excited mode. Because the extinction ratio of the E11 mode is higher than that of the E21 mode, as shown in Fig. 9(b), slight impurity in the excited mode would lead to slight underestimation (overestimation) of the extinction ratio of the E11 (E21) mode. The insertion losses of the device, measured for the TE (TM) polarization at the wavelength 1550 nm, are 9.6 dB (9.8 dB) and 10.2 dB (10.4 dB) for the E11 and E21 modes, respectively, which include the fiber-coupling losses at the two ends (∼2 dB). The material loss is ∼2 dB/cm.
Fig. 9. (a) Coupling ratios and (b) extinction ratios for the E11 and E21 modes measured for the TE and TM polarizations.
To measure the switching speed of the device, we launched the E11 and E21 modes into Core 1 at the wavelength 1550 nm, respectively, and turned the switch on and off periodically with square waves. We measured the temporal response of the output optical power from Core 2 with a photodetector and an oscilloscope. Figures 10(a) and 10(b) show the output waveforms from Core 2 displayed on the oscilloscope for the E11 and E21 modes, respectively. For the E11 mode, the rise and fall times are 824 µs and 944 µs, respectively, while, for the E21 mode, the rise and fall times are 912 µs and 1.084 ms, respectively. The response times are insensitive to the polarization.
Fig. 10. Temporal responses of the device measured at the wavelength 1550 nm for the (a) E11 and (b) E21 modes.
5. Conclusions
We have proposed and designed a mode-independent thermo-optic switch based on a symmetric DC formed with two parallel identical two-mode waveguides, where two electrode heaters are placed, respectively, on the two waveguides. The device allows both modes to be switched together from one waveguide to the other waveguide by applying sufficient electric power to one of the electrode heaters. Our experimental device fabricated with polymer materials, which has a total length of 22.5 mm, shows a switching power of 128 mW, at which the extinction ratios in the C-band for the fundamental mode and the higher-order mode are higher than 18.3 dB and 13.4 dB, respectively. The switching time of the device is about 1 ms. The mode-dependent loss and the polarization-dependent loss measured at the wavelength 1550 nm are ∼0.6 dB and ∼0.2 dB, respectively. The characteristics of the device are insensitive to the polarization state of light. Our proposed device could find applications in reconfigurable MDM systems, where modes of different orders need to be switched together between two few-mode fibers. It should be possible to construct mode-independent switches for more than two modes with the same approach by using all vertical DCs [13,15]. It should also be possible to form more compact devices by using high-index-contrast material systems, such as Si and InP, but it would be difficult to achieve polarization-insensitive operation with such material systems.
Funding
Research Grants Council, University Grants Committee (CityU 11253316); Hong Kong Scholars Program (XJ2016026).
Disclosures
The authors declare no conflicts of interest.
References
1. Y. Shi, O. Eknoyan, and H. F. Taylor, “Polarization dependence of the coupling ratio in LiTaO3 directional couplers,” Appl. Opt. 44(7), 1156–1159 (2005). [CrossRef]
2. X. Wang, J. Sun, Y. Liu, J. Sun, C. Chen, X. Sun, F. Wang, and D. Zhang, “650-nm 1 × 2 polymeric thermo-optic switch with low power consumption,” Opt. Express 22(9), 11119–11128 (2014). [CrossRef]
3. M. Davanco, P. Holmström, D. J. Blumenthal, and L. Thylén, “Directional coupler wavelength filters based on waveguides exhibiting electromagnetically induced transparency,” IEEE J. Quantum Electron. 39(4), 608–613 (2003). [CrossRef]
4. F. Zhang, H. Yu, Y. Wang, Z. Lu, L. Chrostowski, and N. A. F. Jaeger, “Compact broadband polarization beam splitter using a symmetric directional coupler with sinusoidal bends,” Opt. Lett. 42(2), 235–238 (2017). [CrossRef]
5. Q. Huang and K. S. Chiang, “High-order-mode-pass mode (de)multiplexer with a hybrid-core vertical directional coupler,” J. Lightwave Technol. 37(16), 3932–3938 (2019). [CrossRef]
6. N. Hanzawa, K. Saitoh, T. Sakamoto, T. Matsui, K. Tsujikawa, M. Koshiba, and F. Yamamoto, “Two-mode PLC-based mode multi/demultiplexer for mode and wavelength division multiplexed transmission,” Opt. Express 21(22), 25752–25760 (2013). [CrossRef]
7. K. Saitoh, N. Hanzawa, T. Sakamoto, T. Fujisawa, Y. Yamashita, T. Matsui, K. Tsujikawa, and K. Nakajima, “PLC-based mode multi/demultiplexers for mode division multiplexing,” Opt. Fiber Technol. 35, 80–92 (2017). [CrossRef]
8. D. Dai and S. Wang, “Asymmetric directional couplers based on silicon nanophotonic waveguides and applications,” Front. Optoelectron. 9(3), 450–465 (2016). [CrossRef]
9. N. Riesen, S. Gross, J. D. Love, and M. J. Withford, “Femtosecond direct-written integrated mode couplers,” Opt. Express 22(24), 29855–29861 (2014). [CrossRef]
10. J. Dong, K. S. Chiang, and W. Jin, “Compact three-dimensional polymer waveguide mode multiplexer,” J. Lightwave Technol. 33(22), 4580–4588 (2015). [CrossRef]
11. T. Watanabe and Y. Kokubun, “Demonstration of mode-evolutional multiplexer for few-mode fibers using stacked polymer waveguide,” IEEE Photonics J. 7(6), 1–11 (2015). [CrossRef]
12. W. K. Zhao, K. X. Chen, J. Y. Wu, and K. S. Chiang, “Horizontal directional coupler formed with waveguides of different heights for mode-division multiplexing,” IEEE Photonics J. 9(5), 1–9 (2017). [CrossRef]
13. Q. Huang, Y. Wu, W. Jin, and K. S. Chiang, “Mode multiplexer with cascaded vertical asymmetric waveguide directional couplers,” J. Lightwave Technol. 36(14), 2903–2911 (2018). [CrossRef]
14. X. Zi, L. Wang, K. Chen, and K. S. Chiang, “Mode-selective switch based on thermo-optic asymmetric directional coupler,” IEEE Photonics Technol. Lett. 30(7), 618–621 (2018). [CrossRef]
15. Q. Huang, K. S. Chiang, and W. Jin, “Thermo-optically controlled vertical waveguide directional couplers for mode-selective switching,” IEEE Photonics J. 9(5), 1–9 (2017). [CrossRef]
16. D. J. Richardson, J. M. Fini, and L. E. Nelson, “Space-division multiplexing in optical fibres,” Nat. Photonics 7(5), 354–362 (2013). [CrossRef]
17. G. Li, N. Bai, N. Zhao, and C. Xia, “Space-division multiplexing: the next frontier in optical communication,” Adv. Opt. Photonics 6(4), 413–487 (2014). [CrossRef]
18. T. Mizuno and Y. Miyamoto, “Optical fiber technology high-capacity dense space division multiplexing transmission,” Opt. Fiber Technol. 35, 108–117 (2017). [CrossRef]
19. Q. Xu, M. Jiang, D. Niu, X. Wang, L. Wang, K. S. Chiang, and D. Zhang, “Fast and low-power thermo-optic switch based on organic-inorganic hybrid strip-loaded waveguides,” Opt. Lett. 43(20), 5102–5105 (2018). [CrossRef]
20. X. Wang, W. Jin, Z. Chang, and K. S. Chiang, “Buried graphene electrode heater for polymer-waveguide thermo-optic device,” Opt. Lett. 44(6), 1480–1483 (2019). [CrossRef]
