Polarization-selective microring resonators (MRRs) are designed and demonstrated to work with resonances for only one of TE and TM polarizations for the first time, which can be used for realizing the crossing-free, compact and scalable multi-wavelength coherent receiver with dual polarizations. The present polarization-selective MRRs are realized by manipulating the polarization-dependence of the bending loss of the micro-resonators and the coupling ratio of the bent directional couplers introduced here. The demonstrated TM-type MRR and the TE-type MRR work well as a wavelength-selective optical filter for TM polarization and TE polarization, respectively, as designed. On the other hand, the resonance of the undesired orthogonal polarization mode is depressed significantly with very high extinction ratio. For the realized TM-type MRR, the peak-transmission at the drop port is depressed to be lower than −50dB and the loss of the transmission at the through port is very low (<0.5dB) when TE polarization mode is input. For the realized TE-type MRR, the peak-transmission at the drop port is depressed to be ~−40dB and the loss of the transmission at the through port is very low (<0.5dB) when TM-polarization mode is input. The present polarization-selective TM- and TE-type MRRs are useful for the applications with multiple wavelengths and dual polarizations.
© 2017 Optical Society of America
There is increasing demand for high transmission capacity from data-centre interconnects to long-haul transmissions [1,2]. In order to enhance the link capacity of optical interconnects, advanced multiplexing technologies have been playing a very important role in the past decades. Currently, the most popular technologies include wavelength-division-multiplexing (WDM), polarization-division-multiplexing (PDM), mode-division-multiplexing (MDM), etc [2–5]. In order to further increase the capacity of a system with many multiplexed channels, one of the most economical means is increasing spectral efficiency with the assistance of the coherent optical detection technique [6–9]. As demonstrated in recent years, the coherent detection technique has been combined with advanced optical modulation formats, such as quadrature phase-shift keying (QPSK), quadrature amplitude modulation (QAM), and orthogonal frequency-division multiplexing (OFDM), to increase the spectral efficiency [10,11].
The fundamental concept for coherent optical detection is to measure the product of electrical fields from a modulated signal and a continuous-wave (CW) from local oscillator (LO). A coherent receiver consists of a large number of components, such as polarization beam splitters (PBSs), 90° optical hybrids and photodetectors. It is possible to integrate all these components on a single-chip monolithically to reduce the cost and footprint, which is important for coherent optical systems. In past years, people have developed some coherent receiver chips by utilizing advanced photonic integrated circuit technologies, e.g., single-quadrature coherent receiver PICs - , single-polarization dual-quadrature coherent receiver PICs [14,15], and dual-polarization dual-quadrature coherent receiver PICs . All these coherent receiver chips work for a single wavelength-channel.
In order to significantly increase the data transfer rate to 400 Gb/s and beyond, one should introduce more channels. As a result, a multi-channel coherent receiver is desired. Particularly, when multiple wavelength-channels are introduced, one needs a multiwavelength coherent receiver, which might be the next-generation coherent receiver PIC . As described in , a multiwavelength coherent receiver typically has two wavelength demultiplexers so that the signal comb and LO comb are demultiplexed and received in the coherent receivers individually. The major drawbacks include : (1) there are many waveguide crossings, which introducing some excess loss as well as crosstalk; (2) there is stringent requirement on the extinction ratio of the PBSs in a broad band. In , Doerr, et al. reported a monolithic InP multiwavelength polarization-diversity dual-quadrature coherent receiver with balanced detection, which combines all the critical functions into one arrayed-waveguide grating (AWG). With this smart design, there is no waveguide crossings or post-fabrication adjustments.
More recently silicon photonics technology has become very promising because of the capability for high integration density, low energy consumption, and the CMOS compatibility. Therefore coherent receivers on silicon is becoming very attractive [18–21]. In particular, the large birefringence of silicon-on-insulator (SOI) nanowires is also very helpful to make very compact PBSs and polarization rotators for constructing polarization-diversity circuits . This is necessary for the generation and detection of dual-polarization optical signals to double the data capacity in dual-polarization coherent optical systems. Silicon-based coherent receivers have been demonstrated successfully in recent years, including dual-polarization QPSK coherent receivers and transmitter silicon PICs from Bell Labs [23,24].
In this paper, we propose a microring-based multi-wavelength coherent receiver with SOI nanowires. Microring resonators (MRRs) have been used very popularly as wavelength-selective optical filters, and can be arrayed easily to deal with multiple wavelength-channels . It is also very well known that the resonance wavelength of an MRR is usually very polarization-sensitive because of the ultra-high birefringence of an SOI nanowire. Therefore, it is very likely that an MRR can also work as a PBS simultaneously by utilizing the polarization dependence of the resonance wavelengths [25,26]. However, one should note that a regular MRR has resonances for both TE and TM polarizations. These resonances might introduce very significant polarization-crosstalk from the orthogonal polarization at the resonance wavelengths for a WDM system with many wavelength channels. This significantly limits the wavelength range available.
In order to solve this problem, in this paper we propose and realize novel polarization-selective MRRs with bent directional couplers (DCs), which are designed to work for only one of TE and TM polarizations, for the first time. The TE-type MRR can drop the wavelength-channel of TE polarization only while the TM-type MRR can drop the wavelength-channel of TM polarization only. The polarization-selectivity of these special MRR-based optical filters is realized by manipulating the polarization-dependence of the bending loss and the coupling ratio for the bent DCs used. In this way, the critical coupling condition of a micro-resonator is satisfied for one polarization only and is broken for the other polarization automatically. Such an MRR simultaneously serves as a polarization-handling device with high extinction-ratio and an optical filter with low loss as well as low crosstalk. In this paper we present the design, fabrication and characterization of the novel polarization-selective MRRs based SOI nanowires, which are useful for the applications with multiple wavelengths and dual polarizations.
2. Structure and Design
Figure 1 shows the microring-based multi-wavelength coherent receiver proposed here. Compared to coherent receivers reported previously, the proposed configuration is crossing-free, compact and scalable for more channels. As shown in this figure, the i-th cell for the i-th wavelength-channel consists of two groups of elements. One is for TE polarization and the other is for TM polarization. Each group includes two microrings, a 90° hybrids, four photodetectors. Here balanced detection is employed at the two outputs in order to suppress dc components of the signal, which is the same as standard coherence receivers. In order to achieve complex amplitude, optical 90° hybrids are used to mix the signal and LO with four outputs . In this configuration, all the elements except the microrings have been developed successfully as demonstrated previously [28–32].
Here we introduce special novel TE-type and TM-type MRRs, which are designed to be polarization-selective. When the dual-polarization optical signals carried by the wavelengths (λ1, λ2, …, λN) arrive, the TE-type MRR drop the i-th wavelength channel (λi) of TE polarization only while all the signals of TM polarization pass the TE-type MRR with very low loss. The dropped channel of signal are then mixed with the LO by using a 90° optical hybrid. Then the in-phase components and quadrature components of the signals can be extracted and converted into electrical signals by photodetectors. For the case with TM-polarization, the operation is very similar. The TE-type MRR can drop the wavelength-channel of TE polarization only while the TM-type MRR can drop the wavelength-channel of TM polarization only. This is totally different from the polarization dependence of regular MRRs. In the following part, the structure and design of the proposed polarization-selective MRRs for TE- and TM-polarizations are described in details.
It is well known that the transmission at the drop port of an MRR is given by the following equation ,
From this equation, it can be seen that the peak power T0 approaches zero when the following condition is satisfied
Therefore, in order to achieve polarization-selective MRR-based optical filters, the key is effectively manipulating the polarization-dependence of the bending loss and the coupling ratio for the MRRs. It is possible to depress the resonance by minimizing the evanescent coupling or enhancing the bending loss of the undesired polarization. Meanwhile, for the desired polarization mode, the coupling ratio and the bending loss should be chosen optimally to satisfy the requirements of low loss and high extinction ratio. Since SOI nanowires have very high birefringence, it is possible to make the critical coupling condition satisfied for one polarization only and unsatisfied for the other polarization automatically.
For example, for a ~500nm × 220nm SOI nanowire with a SiO2 upper-cladding, which is considered in this paper, TE-polarization mode usually has much weaker evanescent coupling than TM-polarization mode due to the much stronger confinement. Therefore, it is possible to depress the resonance of TE-polarization mode by designing the coupling region between the access waveguide and the microring waveguide, so that the coupling is weak sufficiently. In this way, one can realize a TM-type MRR, which efficiently drops the wavelength-channel of TM polarization only. In contrast, the TE-type MRR can be designed by choosing a small bending radius, in which case TM-polarization mode has much higher bending loss than TE-polarization mode.
In order to enable the flexible manipulation of the bending loss and the coupling ratio, in this paper we use the MRRs with bent DCs, as shown in Fig. 2(a)-2(b). The bent DCs are introduced to achieve strongly polarization-selective coupling, which is achievable when it is designed according to the phase-matching condition, as demonstrated previously . For the TM-type MRR, the cavity is formed by using a microring with a bending radius R1, as shown in Fig. 2(a). For the TE-type MRR, we introduce a novel cavity formed by arc waveguides with bending radii R1a and R1b as shown in Fig. 2(b). In this design, the bending radius R1a should be small enough to introduce significant bending loss for TM polarization and sufficient low loss for TE polarization. Meanwhile, the bending radius R1b should be large enough to achieve sufficient coupling for TE polarization and sufficiently low loss for TM polarization (which propagates along the access waveguide with a similar bending radius R2).
2.1 TM-type MRRs
For the TM-type MRR, it is required to depress the resonance of TE-polarization. As mentioned above, the possible solution is to design the coupling region with very weak coupling for TE-polarization mode. When using a singlemode SOI nanowire with a regular ~500nm × 220nm cross section, the bending radius of the TM-type MRR should be ~10μm, which is large enough to guarantee a negligible bending loss for TM polarization. In this case, TE-polarization mode has much lower pure bending loss than TM-polarization mode because of the stronger confinement. Therefore, when TE-polarization mode propagates along the TM-type microring waveguide, the loss L is mainly from the scattering loss. According to the measured loss of the bending waveguide fabricated in our lab, we assume that the loss L of the microring waveguide is L = 5~10dB/cm for TE-polarization mode when calculating the peak power T0 of the transmission at the drop port of the MRR with a bending radius of ~10μm. Figure 3 shows the calculated result for the peak power T0 at the drop port of an MRR when the amplitude coupling ratio k varies from 0.01 to 0.03. From this figure, it can be seen that the peak power T0 is less than −30dB when the amplitude coupling coefficient is k<0.01 and k <0.015 for the cases of L = 5dB/cm and 10dB/cm, respectively.
For the design of the TM-type MRR, first we choose the bending radius R1 = 10μm to ensure that the bending loss for TM-polarization mode is negligible. The bending radius for the access waveguide is chosen as R2 = 11.1μm, so that the separation between the microring waveguide and the access waveguide is 1.1μm to achieve a reasonable coupling ratio for TM-polarization mode. The waveguide widths w1 and w2 are chosen according to the phase matching condition for TM polarization mode, i.e., neff1 R1 = neff2 R2, where neff1 and neff2 are the effective indices of the fundamental modes in the microring waveguide and the access waveguide . Figure 4 shows the calculated results of neff1R1 and neff2R2 as the waveguide core width varies. From this figure, we choose the waveguide widths as w1 = 590nm and w2 = 350nm, respectively, as an example.
Figure 5 shows the calculated amplitude coupling coefficients of the designed bent coupler for TE- and TM-polarization modes as the coupling angle θ varies from 4° to 20°. The coupling coefficients are calculated from a 3D-FDTD simulation (from Lumerical software) for light propagation in the coupling region. As this bent coupler for the TM-type MRR is designed to make the phase match condition satisfied for TM mode only, the phase match condition is not satisfied for TE polarization automatically due to the birefringence. According to the coupled mode theory, the coupling coefficient for TE polarization can be close 100% by increasing the coupling angle θ. For TE polarization, however, the maximal coupling coefficient κTE_max is less than 100% due to the phase mismatching. Therefore, when increasing the coupling angle θ, one observes a decrease of the coupling coefficient after it reaches the maximum around θ = 12°, as shown in Fig. 5. From this figure, it can be seen that the amplitude coupling ratio kTM of TM-polarization mode increases from 0.17 to 0.25, which is reasonable to achieve low-loss and low-crosstalk responses. One the other hand, the amplitude coupling ratio of TE-polarization mode kTE varies in the range from 0.00283 to 0.0122. We choose the coupling angle as θ = 16°, where the power coupling ratio of the TE polarization mode is minimal. With this design of θ = 16°, one has amplitude coupling coefficients kTM = 0.224 and kTE = 0.0028 for TM- and TE-polarization modes, respectively. Assuming the loss L = 10dB/cm, for TE polarization one has k2e−αl/2 = 8.0 × 10−5, which is much less than 1.0 and thus Eq. (3) is satisfied. As a result, it is expected that the peak power T0 approaches zero for the TM-type MRR when TE-polarization mode is launched.
Figure 6(a)-(b) shows the simulated spectral responses of the designed TM-type MRR when TM- and TE-polarization modes are launched at the input port, respectively. The spectral responses are calculated by a transfer matrix method  with the coupling ratio calculated from the 3D-FDTD simulation. Here we assume that the loss L = 10dB/cm for TE-polarization mode. It can be seen that the designed TM-type MRR works very well with low loss and low crosstalk when operating with TM polarization, which is similar to the response of regular MRRs. In contrast, when it is operating with TE polarization, the peak power T0 of the transmission at the drop port is as low as −60dB, which indicates that the resonance is depressed successfully and one has very high polarization extinction ratio (PER) at the drop port. With this design, TE-polarization mode can go through the TM-type MRR and outputs from the through port with very low loss, as shown in Fig. 6(b). Regarding that the coupling ratio is relatively sensitive to the variation Δw of the waveguide core width as well as the gap width, we analyze the sensitivity of the PER to the core width variation Δw. Our simulation shows that the PSR is kept very high (>40dB) in a broad band even when ∆w = ± 20nm, which indicates that the present design is robust.
2.2 TE-type MRRs
For TE-type MRRs, the bending radius should be chosen carefully so that TM polarization mode has huge bending loss while TE polarization mode has negligible bending loss. Figure 7 shows the calculated bending losses of the microring waveguides with different core widths (e.g., w = 450, 500, and 550nm) as the bending radius ranges from 2μm to 3μm. It can be seen that the bending loss of TE-polarization mode increases from 0.00011dB/cm to 2.49dB/cm when the bending radius decreases from 3μm to 2μm. In contrast, the bending loss of TM-polarization mode is as high as 1~7 × 103 dB/cm. For example, when one chooses R = 3μm, the bending loss of TM-polarization mode is 2470dB/cm while the bending loss of TE polarization mode is as low as ~0.01dB/cm for an SOI strip waveguide with w = 450nm. However, if we choose R1 = 3μm for the microring with bent DCs, the bending radius R2 of the bending section in the access waveguide will be too small to achieve low-loss transmission for TM-polarization mode which propagates along the access waveguide. In order to solve this problem, a novel micro-resonator is introduced to be with the arcs of different bending radii, R1a and R1b (where R1a < R1b), as shown in Fig. 2(b). Here R1a is chosen as R1a = 3μm to have sufficient bending loss for TM-polarization mode in the cavity. Meanwhile, we choose R1b = 4.8μm to be large enough to minimize the excess loss for TM-polarization mode when propagating along the access waveguide and arriving at the through port. With this design, the TE-type MRR is oval-shaped, as shown in Fig. 2(b).
According to the requirement of the fabrication processes, the gap width of the bent DC is chosen to be wgap = ~170nm. When determining the widths of the microring waveguide and the access waveguide, there are two considerations. First, the widths of the microring waveguide and the access waveguide should be chosen so that the coupling of TE-polarization mode could be sufficient by choosing the length of the coupling region appropriately; Second, the phase mismatch should be large enough for TM polarization, so that the coupling for TM polarization can be minimized. By examining the coupling coefficients of TE- and TM-polarization modes, we choose the widths of the microring waveguide and the access waveguide as w1 = 450nm and w2 = 400nm. As a result, the bending radius of the access waveguide is given as R2 = R1b + (w1/2 + wgap + w2/2) = 5.42μm.
Figure 8 shows the calculated amplitude coupling coefficients calculated from a 3D-FDTD simulation as the coupling angle of the bent coupling region varies from 10° to 60°. It can be seen that the amplitude coupling for TM-polarization mode is minimized to be around kTM = 0.071 when the coupling angle θ = 50°, which is very helpful to depress the resonance. Meanwhile, the amplitude coupling coefficient of TE-polarization mode is kTE = 0.401 which is sufficiently large for the resonance of TE-polarization mode. For TM polarization, Eq. (3) is satisfied because one has k2e−αl/2 = 0.0041. As a result, it is expected that the peak power T0 approaches zero for the TE-type MRR when TM-polarization mode is launched.
For the designed TE-type MRR with bending radii R1a = 3μm and R1b = 5μm, the bending loss are 2470dB/cm (R1a = 3μm) and 320dB/cm (R1b = 5μm) for TM-polarization mode, while the bending loss for TE-polarization mode is negligible. As it is well known, some mode-mismatch loss and reflection might be introduced at the junction between bent waveguides with different radii. For the present junction with R1a = 3μm and R1b = 4.8μm, the reflection is negligible from the simulation result. The calculated mode mismatching loss for TE polarizations is as low as 0.003dB, which is also negligible. For TM polarization, the mode mismatching loss is about 0.1dB, which is also negligible regarding that the pure bending loss is as high as 2470dB/cm. By using a transfer matrix method  with the coupling ratio calculated from the 3D-FDTD simulation, the spectral responses of the designed TE-type MRR are calculated for the cases when TE- and TM-polarization modes are launched at the input port, respectively, as shown in Fig. 9(a)-9(b). It can be seen that the designed TE-type MRR works very well with low loss and low crosstalk when operating with TE polarization, which is similar to the response of regular MRRs. In contrast, when it is operating with TM polarization, the peak power T0 of the transmission at the drop port is as low as −40dB, which indicates that the resonance is depressed successfully. With this design, TM-polarization mode can go through the TE-type MRR and outputs from the through port with very low loss, as shown in Fig. 9(b). Here the sensitivity of the PER to the core width variation Δw is also analyzed numerically for this TE-type MRR by assuming there is core-width variation ∆w = ± 20nm. The simulation shows that the PSR is kept very high (>30dB) in a broad band even when ∆w = ± 20nm. This indicates that the present design for the TE-type MRR is also robust.
3. Fabrication and Characterization
For the fabrication, we used a regular SOI wafer with a 220nm-thick top-silicon layer and a 2μm-thick buried-oxide layer. First a layer of negative photoresist MAN2403 was spin-coated on the top of the SOI wafer. An E-beam lithography process was used for photoresist patterning. After the development and the baking process of the patterned photoresist, a dry-etching process was employed to transfer the photoresist pattern to the top-silicon layer. A SiO2 layer was then deposited on the top of the chip as the upper-cladding. In order to characterize the spectral responses for the fabricated TE-type and TM-type MRRs, we included more than one devices with identical structure designs on the same chip. Some of them have TE-type grating couplers while the other ones have TM-type grating couplers, in which way one can do the measurement for TE and TM-polarizations conveniently, as shown in Fig. 10(a). Figure 10(b)-10(c) show the enlarged view for the TM-type MRRs and the TE-type MRRs fabricated on the same chip, respectively.
For the characterization of the fabricated MRRs, we used a setup with a tunable laser (Agilent 81940A) with a tunable range of wavelength from 1520nm to 1610nm and a power meter (Agilent 8163A). The polarization state of the laser source is adjusted by using a polarization controller. Figure 11(a)-11(b) show the measurement results for the TM-type MRR when TE-polarization mode and TM-polarization mode is launched, respectively. These results are normalized with respect to the transmission of the straight waveguides fabricated on the same chip. It can be seen that these experimental results agree very well with the calculation result shown in Fig. 6(a)-6(b) even though there is still some slight difference, which is attributed to the deviation of the coupling ratios and the bending losses for the practical devices. From Fig. 11(a), it can be seen that the realized TM-type MRR works as a regular filter with a low loss of <1dB and a high extinction ratio of 20~25dB when TM-polarization mode is launched at the input port. And the free spectral range (FSR) is ~10nm, which is consistent with the calculated value (show in Fig. 6(a)). From Fig. 11(b), it can be seen that the TM-type MRR allows TE-polarization mode to go through and finally output from the through port with a low loss of <0.5dB. For TE-polarization mode, the transmission at the drop port is <−50dB over the wavelength range from 1530nm to 1600nm. For this realized TM-type MRR, which is highly polarization-selective, the polarization extinction ratio (PER) at the drop port is higher than 50dB (very close to the calculation result shown in Fig. 6(b)).
Figure 12(a)-(b) show the measurement results for the fabricated TE-type MRR when TE-polarization mode and TM-polarization mode are launched, respectively. These results are normalized with respect to the transmission of the straight waveguides fabricated on the same chip. From Fig. 12(a), it can be seen that the realized TE-type MRR works as a regular filter with a low loss of ~1dB and a high extinction ratio of ~15dB when TE-polarization mode is launched at the input port. And the FSR is ~25nm. From Fig. 12(b), it can be seen that the TE-type MRR allows TM-polarization mode to go through and finally output from the through port with a low loss of <0.5dB. For TM-polarization mode, the transmission at the drop port is ~−40dB over the wavelength range from 1530nm to 1590nm. This realized TE-type MRR shows very high polarization-selectivity and the polarization extinction ratio at the drop port is ~40dB. These experimental results (including the FSR, and PER) also agree well with the calculation result shown in Fig. 9(a)-9(b). The slight difference is attributed to the deviation of the coupling ratios and the bending losses for the practical devices.
For practical applications, one should notice that that the drop wavelengths of TE-polarization should be the same as those of TM polarization. However, as it is well known, the resonance wavelengths of an MRR optical filters are very sensitive to the random variation of the waveguide dimension. The measured resonance wavelength variation is even as large as 1~5nm for the MRRs fabricated on the same chip even when they have identical structural parameters . According to the theoretical analysis, the resonance waveguide shift is ~1nm when there is a waveguide dimension variation of 1nm . Therefore, it is hard to achieve identical drop wavelengths for the TE-type and TM-type MRRs without any active tuning. In order to make the wavelengths aligned to the ITU grid for both polarizations, a potential approach is introducing micro-heaters to tune the resonance wavelengths, which has been used successfully before . We also noticed that the present TE-type MRR and the TM-type MRR have different FSRs because different bending radii are chosen for them regarding to the bending loss. Currently the FSR for the TM-type MRR is less than that for the TE-type MRR. Therefore, the number of the wavelength-channel available will be limited by the FSR of the TM-type MRR. For example, when the channel spacing Δλ = 0.8nm, there are 16 channels available. Fortunately, it is possible to have a larger FSR for the TM-type MRR by introducing multiple MRRs [39–41].
As a summary, we have proposed and realized novel polarization-selective MRRs with bent DCs. The present TM-type MRR and TE-type MRR have been realized by manipulating the polarization-dependence of the bending loss and the coupling ratio for the bent DCs. The experimental results have shown that the TM-type MRR and the TE-type MRR demonstrated in this paper work very well as a wavelength-selective optical filter for TM polarization and TE polarization, respectively. On the other hand, the resonance of the undesired orthogonal polarization is depressed with very high extinction ratio. For the TM-type MRR, when TE-polarization mode is launched, the transmission at the drop port is depressed to be lower than −50dB and the loss of the transmission at the through port is very low (<0.5dB). For the TE-type MRR, when TM polarization mode is input, the transmission at the drop port is depressed to be ~−40dB and the loss of the transmission at the through port is very low (<0.5dB). Such polarization-selective MRRs (TM-type and TE-type) can be cascaded for multiple wavelength channels, which is useful for the applications with multiple wavelengths and dual polarizations, e.g., the realization of a crossing-free, compact and scalable multi-wavelength coherent receiver.
This project was partially supported by National Key Research and Development Plan (No. 2016YFB0402502), and National Natural Science Foundation of China (NSFC) (No. 61431166001, 11374263, and 61422510).
References and links
1. W. R. Tkach, “Scaling optical communications for the next decade and beyond,” Bell Labs Tech. J. 14(4), 3–9 (2010). [CrossRef]
2. P. J. Winzer, “Making spatial multiplexing a reality,” Nat. Photonics 8(5), 345–348 (2014). [CrossRef]
3. D. J. Richardson, J. M. Fini, and L. E. Nelson, “Space-division multiplexing in optical fibres,” Nat. Photonics 7(5), 354–362 (2013). [CrossRef]
4. N. Zhao, X. Li, G. Li, and J. M. Kahn, “Capacity limits of spatially multiplexed free-space communication,” Nat. Photonics 9(12), 822–826 (2015). [CrossRef]
5. D. Dai and J. E. Bowers, “Silicon-based on-chip multiplexing technologies and devices for Peta-bit optical interconnects,” Nanophotonics 3(4), 283–311 (2014).
6. F. Derr, “Coherent optical QPSK intradyne system: concept and digital receiver realization,” J. Lightwave Technol. 10(9), 1290–1296 (1992). [CrossRef]
8. R. J. Deri, E. C. M. Pennings, A. Scherer, A. S. Gozdz, C. Caneau, N. C. Andreadakis, V. Shah, L. Curtis, R. J. Hawkins, J. B. D. Soole, and J. Song, “Ultracompact monolithic integration of balanced, polarization diversity photodetectors for coherent lightwave receivers,” IEEE Photonics Technol. Lett. 4(11), 1238–1240 (1992). [CrossRef]
9. H. G. Bach, A. Matiss, C. C. Leonhardt, R. Kunkel, D. Schmidt, M. Schell, and A. Umbach, “Monolithic 90 hybrid with balanced PIN photodiodes for 100 Gbit/s PM-QPSK receiver applications,” Optical Fiber Communication Conf. Optical Society of America, (2009).
10. J. M. Kahn and K. P. Ho, “Spectral efficiency limits and modulation/detection techniques for DWDM systems,” IEEE J. Sel. Top. Quantum Electron. 10(2), 259–272 (2004). [CrossRef]
11. W. Shieh and C. Athaudage, “Coherent optical orthogonal frequency division multiplexing,” Electron. Lett. 10(1), 42 (2006).
12. H. Takeuchi, K. Kasaya, Y. Kondo, H. Yasaka, K. Oe, and Y. Imamura, “Monolithic integrated coherent receiver on InP substrate,” IEEE Photonics Technol. Lett. 1(11), 398–400 (1989). [CrossRef]
13. T. L. Koch, U. Koren, R. P. Gnall, F. S. Choa, F. Hernandez-Gil, C. A. Burrus, M. G. Young, M. Oron, and B. I. Miller, “GaInAs/GaInAsP multiple-quantum-well integrated heterodyne receiver,” Electron. Lett. 25(24), 1621–1623 (1989). [CrossRef]
14. C. Doerr, P. Winzer, Y. Chen, S. Chandrasekhar, M. Rasras, L. Chen, T. Liow, K. Ang, and G. Lo, “Monolithic polarization and phase diversity coherent receiver in silicon,” J. Lightwave Technol. 28(4), 520–525 (2010). [CrossRef]
15. S. Faralli, K. N. Nguyen, J. D. Peters, D. T. Spencer, D. J. Blumenthal, and J. E. Bowers, “Integrated hybrid Si/InGaAs 50 Gb/s DQPSK receiver,” Opt. Express 20(18), 19726–19734 (2012). [CrossRef] [PubMed]
16. C. R. Doerr, L. Zhang, and P. J. Winzer, “Monolithic InP multiwavelength coherent receiver using a chirped arrayed waveguide grating,” J. Lightwave Technol. 29(4), 536–541 (2011). [CrossRef]
17. C. R. Doerr, L. Zhang, and P. J. Winzer, “Monolithic InP multi-wavelength coherent receiver,” National Fiber Optic Engineers Conf. Opt. Soc. Am. (2010).
18. S. Rumley, D. Nikolova, R. Hendry, Q. Li, D. Calhoun, and K. Bergman, “Silicon Photonics for Exascale Systems,” J. Lightwave Technol. 33(3), 547–562 (2015). [CrossRef]
19. D. Dai, “Silicon nanophotonic integrated devices for on-chip multiplexing and switching,” J. Lightwave Technol., in press (2017).
20. P. Dong, Y.-K. Chen, G.-H. Duan, and D. T. Neilson, “Silicon photonic devices and integrated circuits,” Nanophotonics 3(4–5), 215–228 (2014).
21. P. P. Absil, P. Verheyen, P. De Heyn, M. Pantouvaki, G. Lepage, J. De Coster, and J. Van Campenhout, “Silicon photonics integrated circuits: a manufacturing platform for high density, low power optical I/O’s,” Opt. Express 23(7), 9369–9378 (2015). [CrossRef] [PubMed]
22. D. Dai, L. Liu, S. Gao, D.-X. Xu, and S. He, “Polarization management for silicon photonic integrated circuits,” Laser Photonics Rev. 7(3), 303–328 (2013). [CrossRef]
24. P. Dong, X. Liu, C. Sethumadhavan, L. L. Buhl, R. Aroca, Y. Baeyens, and Y. K. Chen, “224-Gb/s PDM-16-QAM modulator and receiver based on silicon photonic integrated circuits,” Optical Fiber Communication Conf. Optical Society of America, (2013).
25. B. E. Little and S. T. Chu, “Theory of polarization rotation and conversion in vertically coupled microresonators,” IEEE Photonics Technol. Lett. 12(4), 401–403 (2000). [CrossRef]
27. W. Yang, M. Yin, Y. Li, X. Wang, and Z. Wang, “Ultra-compact optical 90° hybrid based on a wedge-shaped 2 × 4 MMI coupler and a 2 × 2 MMI coupler in silicon-on-insulator,” Opt. Express 21(23), 28423–28431 (2013). [CrossRef] [PubMed]
28. A. Beling and J. C. Campbell, “Heterogeneously integrated photodiodes on silicon,” IEEE J. Quantum Electron. 51(11), 1–6 (2015). [CrossRef]
29. C. R. Doerr, L. Chen, T. Nielsen, R. Aroca, L. Chen, M. Banaee, S. Azemati, G. McBrien, S. Y. Park, J. Geyer, B. Guan, B. Mikkelsen, C. Rasmussen, M. Givhechi, Z. Wang, B. Potsaid, H. Lee, E. Swanson, and J. Fujimoto, “O, E, S, C, and L Band Silicon Photonics Coherent Modulator/Receiver,” in Optical Fiber Communication Conference Postdeadline Papers, OSA Technical Digest (online) (Optical Society of America, 2016), paper Th5C.4.
30. D. Dai, M. Piels, and J. E. Bowers, “Monolithic Germanium/Silicon Photodetectors with Decoupled Structures: Resonant APDs and UTC Photodiodes,” IEEE J. Sel. Top. Quantum Electron. 20(6), 3802214 (2014).
31. D. Yin, T. He, Q. Han, Q. Lü, Y. Zhang, and X. Yang, “High-responsivity 40 Gbit/s InGaAs/InP PIN photodetectors integrated on siliconon-insulator waveguide circuits,” J. Semiconduct. 37(11), 114006 (2016). [CrossRef]
32. Z. Tu, P. Gong, Z. Zhou, and X. Wang, “Ultracompact 100 Gbps coherent receiver monolithically integrated on silicon,” Jpn. J. Appl. Phys. 55(4S), 04EC04 (2016). [CrossRef]
33. W. Bogaerts, P. D. Heyn, T. Van Vaerenbergh, K. De Vos, S. K. Selvaraja, T. Claes, P. Dumon, P. Bienstman, D. Thourhout, and R. Baets, “Silicon microring resonators,” Laser Photonics Rev. 6(1), 47–73 (2012). [CrossRef]
36. T. B. Guo, M. Zhang, Y. L. Yin, and D. Dai, “A laser-trimming-assist wavelength-alignment technique for silicon microdonut resonators,” IEEE Photonics Technol. Lett. 25(5), 419-422 (2016).
37. S. K. Selvaraja, W. Bogaerts, P. Dumon, D. Van Thourhout, and R. Baets, “Subnanometer linewidth uniformity in silicon nanophotonic waveguide devices using CMOS fabrication technology,” IEEE J. Sel. Top. Quantum Electron. 16(1), 316–324 (2010). [CrossRef]
38. B. E. Little, S. T. Chu, P. P. Absil, J. V. Hryniewicz, F. G. Johnson, F. Seiferth, D. Gill, V. Van, O. King, and M. Trakalo, “Very high-order microring resonator filters for WDM applications,” IEEE Photonics Technol. Lett. 16(10), 2263–2265 (2004). [CrossRef]
40. R. Boeck, N. A. F. Jaeger, N. Rouger, and L. Chrostowski, “Series-coupled silicon racetrack resonators and the Vernier effect: theory and measurement,” Opt. Express 18(24), 25151–25157 (2010). [CrossRef] [PubMed]
41. A. M. Matteo and G. Barbarossa, “Novel double-ring optical-guided-wave Vernier resonator,” IEEE Proc. Optoelectron. 144(4), 203-208 (1997).