Open Access
Add to BibSonomyAbstract
We experimentally demonstrate a 4 × 4 microring modulator matrix integrated with the asymmetrical directional couplers based mode multiplexer and de-multiplexer photonic circuit for on-chip optical interconnect. The inter-mode optical crosstalk of the device is less than −20 dB in the wavelength range from 1525 nm to 1565 nm. Data transmission with a throughput capacity of 4 × 4 × 32 Gbps is achieved by utilizing four wavelengths and four spatial modes multiplexing. We envision this structure as a potential solution to increase the communication capacity for on-chip interconnect within limited chip area.
© 2017 Optical Society of America
1. Introduction
With the development of multi-core processor, there is an urgent requirement on the communication capacity for on-chip interconnect [1]. Mass data transmission through traditional electrical wires is limited by both bandwidth density and power consumption. On-chip optical interconnect is widely regarded as a potential solution to overcome the limitations of its electrical counterpart [2]. To continuously increase the capacity of optical communication system, several multiplexing techniques have been developed for both optical fiber communication systems [3–6] and on-chip interconnect systems [7–16], such as wavelength multiplexing, polarization multiplexing and spatial multiplexing. Silicon photonic multiplexing devices are developing rapidly in these years due to their compact size and potential to be integrated with the driving and control circuits [17]. It’s widely believed that silicon photonics has the potential to meet the increasing communication bandwidth requirement of future multi-core processor [18].
Lots of efforts have been made to achieve higher communication capacity by combining more than one multiplexing technique [13–16]. In order to encode data on the multiplexed lights, electro-optic modulators are also required [19]. In this paper, we integrate a 4 × 4 microring modulator matrix with mode multiplexer and de-multiplexer based on asymmetrical directional couplers (ADCs) to demonstrate high-speed modulation with channel spacing tunable wavelength division multiplexing (WDM) and mode division multiplexing (MDM). Data transmission with a throughput capacity of 4 × 4 × 32 Gbps is achieved with such a structure.
2. Principle and design
Figure 1 shows the schematic of the 4 × 4 microring modulator matrix integrated with the mode multiplexer and de-multiplexer. We denote the four input ports as I1 to I4 and the four output ports as O1 to O4 respectively. Four continuous-wave (CW) lights with different wavelengths are multiplexed and coupled into each input port of the device. The input light with a specific wavelength is modulated by the corresponding microring modulator. The modulated light will not be influenced by other microring modulators with different resonance wavelengths due to the wavelength selective characteristic of the microring modulator. The resonance wavelength spacing among the 4 microring modulators in each row of the matrix can be tuned by the micro-heaters integrated with them. All the four groups of the wavelength multiplexing optical signals are fundamental transverse electric (TE0) mode. In the mode multiplexer, three of them are converted to the higher-order (TE1, TE2 and TE3) modes. As a result, there are four spatial modes in the output multimode waveguide of the mode multiplexer (bus waveguide in Fig. 1). In the mode de-multiplexer, the mode multiplexing optical signals are de-multiplexed into four TE0 modes and output at the 4 output ports (O1 to O4). In real application, the bus waveguide can be long enough to support the global communication among different processor cores on the same chip. The wavelength multiplexing CW lights can be generated either by the laser diodes and the wavelength multiplexer or by a group of microring lasers [20]. The wavelength multiplexing optical signals at each output port is finally divided into 4 individual optical signals with different wavelengths by a wavelength de-multiplexer and fed into four photodetectors. The wavelength multiplexing optical signals also can be directly fed into a group of wavelength selective photodetectors [21].
Fig. 1 Schematic of the 4 × 4 microring modulator matrix integrated with mode multiplexer and de-multiplexer for on-chip optical interconnect.
2.1. Microring modulator matrix
The silicon microring modulator consists of a ring waveguide and a straight waveguide, which are 400 nm in width, 220 nm in height and 70 nm in slab thickness. Such a waveguide only supports fundamental quasi-TE mode. When the microring modulator is on-resonance at a specific wavelength, the input light will dissipate in the ring waveguide, resulting in a minimum optical power at the output port. When the microring modulator is off-resonance at a specific wavelength, the input light will not experience much loss in the ring waveguide, maintain almost all the optical power at the output port. By tuning the effective index of the ring waveguide, the resonance wavelength is shifted, which induces a modulation of the transmitted light. To achieve high modulation speed, a PN junction is embedded around the ring waveguide to modulate its effective index by the plasma dispersion effect [22,23]. The P-doping concentration and the N-doping concentration are 2.3 × 1017 cm−3 and 1.4 × 1017 cm−3 respectively. 2.0 × 1020 cm−3 P-doping concentration and N-doping concentration are formed for Ohmic-contact. The microring modulator with a smaller radius has a larger free spectral range (FSR) which can accommodate more channels while is more challenging for fabrication, as a tradeoff the radii of the microring modulators in each row of the matrix are designed to be 9.955 μm, 9.985 μm, 10.015 μm and 10.045 μm. Minor differences make their resonance wavelengths uniformly distributed in the range of FSR. The “Ω” shaped micro-heaters are integrated on top of the microring modulators to shift the resonance wavelengths. Compared with the thermo-optic effect, the plasma dispersion effect is quite weak. So higher Q-factor is required to achieve a relatively large dynamic extinction ratio (ER) with a moderate driving voltage. While higher Q-factor means a larger photon lifetime, which affects the modulation speed of the microring modulator. The Q-factor is mainly decided by the propagation loss of the ring waveguide and the coupling coefficient between the ring waveguide and the straight waveguide. In order to achieve moderate Q-factors, distance from the heavily-doped regions to the side of rib waveguide and the gap between the straight waveguide edge and the ring waveguide edge are designed as shown in Table 1.
Table 1. Structural parameters of the microring modulators.
2.2. Mode multiplexer and de-multiplexer
Several structures can be utilized to realize mode multiplexing, such as multimode interference (MMI) couplers, adiabatic couplers and asymmetric Y-junctions [7–11]. We choose asymmetrical directional coupler (ADC) to construct the mode multiplexer due to its simple structure and good scalability [13].
According to the coupled mode theory [24], when two waveguides are put close enough, there will be periodic energy exchange between them via the evanescent fields. The used ADC consists of a single-mode waveguide and a multimode one, by which mode conversion between a fundamental mode and a high-order mode can be achieved. To achieve higher mode conversion efficiency, the phase matching condition should be obeyed. When the width of the single-mode waveguide is fixed, the fundamental mode can be transformed to the desired high-order mode by selecting a multimode waveguide with an appropriate width and controlling the coupling length. Considering the compatible waveguide structure with microring modulators, rib waveguide with 70 nm in slab thickness is utilized to construct the ADCs. Compared with the channel waveguides, the existence of the slab regions in the rib waveguides leads to a larger coupling coefficient for the same gap condition. Therefore, the coupling length is shorter, resulting in a more compact device. Also, as there are always partial lights in the slab regions which contribute to the coupling, the variation of the coupling length with the gap deviation due to the fabrication imperfection of the rib waveguides is smaller than that of the channel waveguides. While one challenge for the ADC based on the rib waveguides is the critical control of the slab thickness.
Figure 2(a) shows the effective indices of different spatial modes in the rib waveguides and the channel waveguides with different widths at the wavelength of 1.55 μm calculated by finite element method. The width of the rib waveguide carrying the fundamental mode (TE0) is chosen to be 400 nm, which is denoted by the first circle in the horizontal dotted line. The widths of the waveguides carrying high-order modes are then determined according to the phase matching condition, which are denoted by the other four circles in the horizontal dotted line. The widths of the rib waveguides carrying the TE1, TE2 and TE3 modes are chosen to be 916 nm, 1416 nm and 1916 nm, respectively. From Fig. 2 we can also see that the effective indices of the rib waveguides vary slower with the dimensional deviations than those of the channel waveguides, which means that the ADC based on the rib waveguides has a larger tolerance to fabrication imperfection than that based on the channel waveguides.
Fig. 2 (a) Effective indices for different modes in rib waveguides and channel waveguides with different widths. (b) Simulated mode conversion process from the TE0 mode to the TE1, TE2 and TE3 modes and then back to the TE0 mode. (c) Schematic of the mode multiplexer and de-multiplexer.
The finite-different time-domain (FDTD) method is used to optimize the coupling length in consideration of the contribution of the two 90 degree bent waveguides. The optimized coupling lengths for the TE1, TE2 and TE3 modes are 13 μm, 15 μm and 19 μm, respectively. In the simulation, the TE0 mode is firstly transformed to the high-order modes and then transformed back to the TE0 mode. The simulation results are shown in Fig. 2(b), with the right part showing the cross-section mode distributions in the multimode waveguides to verify the mode conversion. We also calculate the coupling length of the ADCs based on the channel waveguides. The width of the channel waveguide carrying the TE0 mode is also set to be 0.4 μm. The coupling lengths for the TE1, TE2 and TE3 modes are 23 μm, 31 μm and 33 μm, respectively, which are longer than those of the ADCs based on the rib waveguides, as analyzed before.
We also evaluate the wavelength dependences of the ADCs. The insertion loss uniformities are less than 1 dB in the wavelength range of 1525-1565 nm, which makes the ADCs suitable for WDM application. The bus waveguides of the mode converters in Fig. 2(b) have different widths. We use adiabatic tapers with the lengths of 10 μm to connect different multimode waveguides to construct the four-mode multiplexer and de-multiplexer, as shown in Fig. 2(c).
3. Fabrication and experimental characterization
3.1. Fabrication of the device
Figure 3 shows the micrograph of the device, which is fabricated on an 8-inch SOI wafer with a 220-nm-thick top silicon layer and a 2-μm-thick buried silicon dioxide layer at the Institute of Microelectronics (IME), Singapore. 248-nm deep ultraviolet photolithography and inductively coupled plasma etching are employed to form the silicon waveguides. A PN junction is embedded around the ring waveguide to modulate its effective index by the plasma dispersion effect. The P-doping and N-doping regions are formed by ion implanting. A 1500-nm-thick silica layer is deposited on the silicon layer by plasma-enhanced chemical vapor deposition (PECVD), which is used to prevent the absorption of the optical field by metal. Titanium nitride (TiN) micro-heaters are made on the separate layer to tune the resonance wavelengths of the microring modulators and slightly compensate the coupling length deviations of the ADCs due to the fabrication imperfection. Via holes are etched after the deposition of a 300-nm-thick silica layer by PECVD. Finally, aluminum wires and pads are fabricated.
Fig. 3 (a) Micrograph image of the fabricated device and magnified images of microring modulator (b), mode multiplexer (c) and butt-coupling region (d).
3.2 Experimental characterization
The experimental setup for characterizing the device is shown in Fig. 4. The spectrum response of the device is measured by an amplified spontaneous emission (ASE) source and an optical spectrum analyzer (OSA). DC power supply is utilized to characterize the thermal tuning efficiencies of the micro-heaters and the modulation efficiencies of the microring modulators.
Fig. 4 Experimental setup (ASE: amplified spontaneous emission, TL: tunable laser, PC: polarization controller, DCPS: direct-current power supply, DUT: device under test, PPG: Pulse pattern generator, OSA: optical spectrum analyzer, DCA: digital communication analyzer, EDFA: Erbium-doped fiber amplifier)
200-μm-long linearly inverse taper with 180 nm in tip is used to couple the light into and out of the device. The coupling loss between the device and the lensed fiber with the spot size of 5.0 μm is about 3 dB. Based on the measured spectral responses, the Q-factors of the microring modulators listed in Table 1 are 3016, 5614, 5707 and 6480 respectively. With the voltage applied to the “Ω” shaped micro-heater, the resonance wavelength of the microring modulator can be tuned (Fig. 5(a)). The thermal tuning efficiency can be calculated by linear fitting to the dependence of the resonance wavelength shift on the tuning power (Fig. 5(b), which is 11.9 mW/nm. With the voltage applied to the PN junction of the microring modulator, its spectral responses under different voltages can be achieved (Fig. 5(c)). The of the microring modulator can be calculated by the following equation [25]:
where V is the applied voltage, is the resonance wavelength shift, FSR is the free spectral range and R is the radius of the microring modulator. The of the microring modulator increases from 2.09 Vcm to 4.39 Vcm with the increase of the applied voltage.Fig. 5 (a) Resonance wavelength shift of the microring modulator by thermal tuning. (b) Calculation of the thermal tuning efficiency. (c) Spectral responses under different applied voltages. (d) under different applied voltages.
The performances of the mode multiplexer and de-multiplexer are also characterized. Figure 6 shows the static transmission spectra measured at the output ports O1, O2, O3 and O4 when a broadband light is coupled into the input ports I1, I2, I3 and I4, respectively. The TE0 modes from the input ports I1, I2, I3 and I4 are firstly transformed to the TE3, TE1, TE0 and TE2 modes respectively by the mode multiplexer then transformed back to the TE0 modes at the 4 output ports by the mode de-multiplexer. It can be seen that the light is mainly guided from the i-th input port to the i-th output port (i = 1, 2, 3, 4) as desired. The insertion losses for I1 to O1, I2 to O2, I3 to O3 and I4 to O4 are 8.2 dB, 7.8 dB, 7.1 dB and 8.5 dB from fiber to fiber. While because of the non-ideality of the mode multiplexer and de-multiplexer, a fraction of light is leaked to other output ports and becomes the crosstalk to the corresponding mode. From these figures we can see that the inter-mode optical crosstalk of the device is less than −20 dB in the wavelength range from 1525 nm to 1565 nm. The sharp dips in these spectra are the filtering spectral responses of the four microring modulators, which have an FSR of about 10 nm. The different extinction ratios come from the different structural parameters of the microring modulators.
Fig. 6 Transmission spectra measured at the output ports O1 (a), O2 (b), O3 (c) and O4 (d) when a broadband light input from the input ports I1, I2, I3 and I4 respectively.
High-speed optical modulation and transmission of the device is also characterized. Experiment setup is shown in lower part of Fig. 4. CW light at the corresponding wavelength is generated by a tunable laser and coupled into the device. 32 Gbps pseudorandom binary sequence electrical signal with the pattern length of 231-1 is applied to the microring modulators on by one through RF probe. Based on the DC response, the Vp-p of the RF driving signal is set to be 5 V with a −2.5 V DC bias voltage. The maximum optical power of the tunable laser is 3.0 dBm. The optical power before the digital communication analyzer (Agilent 86100D) is about −9.0 dBm, which exceed its sensitivity (−5 dBm) at 40 Gbps, so the output optical signal is amplified by erbium-doped fiber amplifier and filtered by tunable filter before it is fed into the digital communication analyzer for eye diagram observation. Figure 7 shows the eye diagrams for the data transmission at 4 different wavelengths and by 4 different modes. Clear and open eye diagrams verify the On-Off Key modulation of 32 Gbps data transmission and the (de)multiplexing functionalities. The capacity of the device can achieve is 4 × 4 × 32 Gbps.
4. Conclusion
We demonstrate a 4 × 4 microring modulator matrix integrated with the mode multiplexer and de-multiplexer based on ADCs. The inter-mode optical crosstalk of the device is less than −20 dB in the wavelength range from 1525 nm to 1565 nm. Data transmission with the capacity of 4 × 4 × 32 Gbps is demonstrated by adopting 4 wavelengths and 4 modes multiplexing.
Funding
National Natural Science Foundation of China (NSFC) (61235001, 61575187, 61535002); Program 863 (2015AA010103, 2015AA017001); the Beijing Science and Technology Plan (Z151100003615005).
References and links
1. S. Bigo, “Coherent optical long-haul system design,” in Optical Fiber Communication Conference, OSA Technical Digest Series (Optical Society of America, 2012), paper OTh3A.1. [CrossRef]
2. D. A. B. Miller, “Device Requirements for Optical Interconnects to Silicon Chips,” in Proceedings of the IEEE (IEEE, 2009), pp. 1166–1185. [CrossRef]
3. P. J. Winzer, “Scaling optical fiber networks: challenges and solutions,” Opt. Photonics News 26(3), 28–35 (2015). [CrossRef]
4. R. G. H. Van Uden, R. A. Correa, E. A. Lopez, F. M. Huijskens, C. Xia, G. Li, A. Schülzgen, H. de Waardt, A. M. J. Koonen, and C. M. Okonkwo, “Ultra-high-density spatial division multiplexing with a few-mode multicore fibre,” Nat. Photonics 8(11), 865–870 (2014). [CrossRef]
5. P. Sillard, M. Bigot-Astruc, and D. Molin, “Few-mode fibers for mode-division-multiplexed systems,” J. Lightwave Technol. 32(16), 2824–2829 (2014). [CrossRef]
6. S. Randel, R. Ryf, A. Sierra, P. J. Winzer, A. H. Gnauck, C. A. Bolle, R. J. Essiambre, D. W. Peckham, A. McCurdy, and R. Lingle Jr., “6×56-Gb/s mode-division multiplexed transmission over 33-km few-mode fiber enabled by 6×6 MIMO equalization,” Opt. Express 19(17), 16697–16707 (2011). [CrossRef] [PubMed]
7. Y. Kawaguchi and K. Tsutsumi, “Mode multiplexing and demultiplexing devices using multimode interference couplers,” Electron. Lett. 38(25), 1701 (2002). [CrossRef]
8. M. Greenberg and M. Orenstein, “Multimode add-drop multiplexing by adiabatic linearly tapered coupling,” Opt. Express 13(23), 9781–9787 (2005). [CrossRef] [PubMed]
9. M. Greenberg and M. Orenstein, “Simultaneous dual mode add/drop multiplexers for optical interconnects buses,” Opt. Commun. 266(2), 527–531 (2006). [CrossRef]
10. T. Uematsu, Y. Ishizaka, Y. Kawaguchi, K. Saitoh, and M. Koshiba, “Design of a compact two-mode multi/demultiplexer consisting of multimode interference waveguides and a wavelength-insensitive phase shifter for mode-division multiplexing transmission,” J. Lightwave Technol. 30(15), 2421–2426 (2012). [CrossRef]
11. J. B. Driscoll, R. R. Grote, B. Souhan, J. I. Dadap, M. Lu, and R. M. Osgood, “Asymmetric Y junctions in silicon waveguides for on-chip mode-division multiplexing,” Opt. Lett. 38(11), 1854–1856 (2013). [CrossRef] [PubMed]
12. H. Qiu, H. Yu, T. Hu, G. Jiang, H. Shao, P. Yu, J. Yang, and X. Jiang, “Silicon mode multi/demultiplexer based on multimode grating-assisted couplers,” Opt. Express 21(15), 17904–17911 (2013). [CrossRef] [PubMed]
13. D. Dai and J. Wang, “Multi-channel silicon mode (de) multiplexer based on asymmetrical directional couplers for on-chip optical interconnects,” IEEE Photonics Soc. News 28(2), 8–14 (2014).
14. J. Wang, S. He, and D. Dai, “On‐chip silicon 8‐channel hybrid (de) multiplexer enabling simultaneous mode‐and polarization‐division‐multiplexing,” Laser Photonics Rev. 8(2), 18–22 (2014). [CrossRef]
15. L. W. Luo, N. Ophir, C. P. Chen, L. H. Gabrielli, C. B. Poitras, K. Bergmen, and M. Lipson, “WDM-compatible mode-division multiplexing on a silicon chip,” Nat. Commun. 5, 3069 (2014). [CrossRef] [PubMed]
16. M. Ye, Y. Yu, G. Chen, Y. Luo, and X. Zhang, “On-chip WDM mode-division multiplexing interconnection with optional demodulation function,” Opt. Express 23(25), 32130–32138 (2015). [CrossRef] [PubMed]
17. R. Soref, “The past, present, and future of silicon photonics,” IEEE J. Sel. Top. Quantum Electron. 12(6), 1678–1687 (2006). [CrossRef]
18. C. Sun, M. T. Wade, Y. Lee, J. S. Orcutt, L. Alloatti, M. S. Georgas, A. S. Waterman, J. M. Shainline, R. R. Avizienis, S. Lin, B. R. Moss, R. Kumar, F. Pavanello, A. H. Atabaki, H. M. Cook, A. J. Ou, J. C. Leu, Y. H. Chen, K. Asanović, R. J. Ram, M. A. Popović, and V. M. Stojanović, “Single-chip microprocessor that communicates directly using light,” Nature 528(7583), 534–538 (2015). [CrossRef] [PubMed]
19. B. Stern and M. Lipson, “High-bandwidth Link with Single Laser Input Using Silicon Modulators and Mode Multiplexing,” in Proc. Conference on Lasers and Electro-Optics (CLEO) (2016), paper STu4G.5. [CrossRef]
20. D. Liang, M. Fiorentino, T. Okumura, H. H. Chang, D. T. Spencer, Y. H. Kuo, A. W. Fang, D. Dai, R. G. Beausoleil, and J. E. Bowers, “Electrically-pumped compact hybrid silicon microring lasers for optical interconnects,” Opt. Express 17(22), 20355–20364 (2009). [CrossRef] [PubMed]
21. K. Preston, Y. H. Lee, M. Zhang, and M. Lipson, “Waveguide-integrated telecom-wavelength photodiode in deposited silicon,” Opt. Lett. 36(1), 52–54 (2011). [CrossRef] [PubMed]
22. E. Timurdogan, C. M. Sorace-Agaskar, J. Sun, E. Shah Hosseini, A. Biberman, and M. R. Watts, “An ultralow power athermal silicon modulator,” Nat. Commun. 5, 4008 (2014). [CrossRef] [PubMed]
23. Q. Xu, B. Schmidt, S. Pradhan, and M. Lipson, “Micrometre-scale silicon electro-optic modulator,” Nature 435(7040), 325–327 (2005). [CrossRef] [PubMed]
24. W. Huang, “Coupled-mode theory for optical waveguides: an overview,” J. Opt. Soc. Am. A 11(3), 963–983 (1994). [CrossRef]
25. X. Xiao, H. Xu, X. Li, Y. Hu, K. Xiong, Z. Li, T. Chu, Y. Yu, and J. Yu, “25 Gbit/s silicon microring modulator based on misalignment-tolerant interleaved PN junctions,” Opt. Express 20(3), 2507–2515 (2012). [CrossRef] [PubMed]
