We demonstrate U-shaped silicon PN junctions for energy efficient Mach-Zehnder modulators and ring modulators in the O-band. This type of junction has an improved modulation efficiency compared to existing PN junction geometries, has low losses, and supports high-speed operation. The U-shaped junctions were fabricated in an 8” silicon photonics platform, and they were incorporated in travelling-wave Mach-Zehnder modulators and microring modulators. For the high-bandwidth Mach-Zehnder modulator, the DC VπL at −0.5 V bias was 4.6 V·mm. It exhibited a 3dB bandwidth of 13 GHz, and eye patterns at up to 24 Gb/s were observed. A VπL as low as ~2.6 V·mm at a −0.5 V bias was measured in another device. The ring modulator tuning efficiency was 40 pm·V−1 between 0 V and −0.5 V bias. It had a 3-dB bandwidth of 13.5 GHz and open eye patterns at up to 13 Gb/s were measured. This type of PN junctions can be easily fabricated without extra masks and can be incorporated into generic silicon photonics platforms.
© 2017 Optical Society of America
CorrectionsZheng Yong, Wesley D. Sacher, Ying Huang, Jared C. Mikkelsen, Yisu Yang, Xianshu Luo, Patrick Dumais, Dominic Goodwill, Hadi Bahrami, Patrick Guo-Qiang Lo, Eric Bernier, and Joyce K. S. Poon, "U-shaped PN junctions for efficient silicon Mach-Zehnder and microring modulators in the O-band: erratum," Opt. Express 26, 32757-32757 (2018)
Silicon (Si) optical modulators, in the form of Mach-Zehnder modulators (MZMs) and ring modulators, are an attractive solution for high-bandwidth electrical-to-optical conversion, since they can be fabricated at the wafer-scale in foundry processes [1–9]. These devices typically use carrier accumulation and depletion via the plasma dispersion effect, which is weak and is lower by a factor of about 0.7 in the O-band than the C-band . An increased modulation efficiency reduces the device length and drive voltage, which reduces the power consumption if the insertion loss (IL) is not compromised .
Si modulation phase-shifters can be formed using PN junction or silicon-insulator-silicon capacitor (SISCAP) structures. Lateral PN junctions are the most common, and they have a relatively high VπL around 2.5 V·cm, but low optical loss of around 10 dB/cm [2,3]. Travelling-wave electrode designs have been demonstrated for lateral junction to support high-speed operation [11–16], up to 41 GHz . Vertical junctions [18–21] have VπL of 0.75 V·cm in the C-band, but they tend to cause high waveguide losses of about 31 dB/cm and their high capacitance limits the bit rates in MZMs to 16 Gb/s so far . Ring modulators with vertical junctions have been demonstrated to support 25Gb/s operation . Interdigitated junctions [22–24] have VπL of 1.12 V·cm in the C-band, but they also have high losses of 25 dB/cm and have worked at up to 10 Gb/s . A higher bit rate (40 Gb/s) has also been demonstrated with interdigitated junctions for which the VπL is 1.5-2.0 V·cm and the loss is around 10 dB/cm . 25Gb/s operation has been shown for the ring modulators with interdigitated junction . Biasing the modulator such that the phase-shifter swings between forward and reverse bias regimes during modulation effectively reduces the VπL, at the cost of a reduced bandwidth and increased propagation loss . In the O-band, to date, the highest efficiency Si MZMs use carrier accumulation in the SISCAP geometry and have a VπL of 2 V·mm, but the propagation loss is very high at 65 dB/cm .
In this article, we present the design and measurements of U-shaped PN junctions for efficient Si MZMs and ring modulators operating in the O-band. Proposed theoretically in [27,28] for the C-band, the U-shaped junction provides an efficiency similar to the SISCAP geometry because of its wide depletion region under reverse bias and large overlap between the change in the depletion region and the waveguide mode. The DC VπL of the junction is between 2.6 and 4.6 V·mm in the O-band, and the propagation loss is 12.5 dB/cm at 0 V bias. The junction is used to demonstrate a 24 Gb/s Si MZM with a 3 dB electro-optic (EO) bandwidth of 13 GHz and a 13 Gb/s microring modulator with a 3 dB EO bandwidth of 13.5 GHz. These results compare favorably against state-of-the-art carrier accumulation MZMs , with the additional advantage that this junction has low loss. This modulation junction was fabricated on an 8” silicon-on-insulator (SOI) wafer as part of the multilayer silicon nitride (SiN)-on-Si platform in .
2. U-shaped junction design and fabrication
The U-shaped junction has a two-dimensional (2D) doping concentration profile in the waveguide cross-section, in which the P-type region extends into the N-type region near the vertical centre. The junction combines features of a lateral junction with a vertical junction. Figures 1(a) and (b) show the doping concentration of the designed U-shaped junction for 0 V and −1 V bias computed using Sentaurus TCAD for the implant conditions and waveguide geometry to be discussed in this section. The blue regions are P-type and yellow/orange/red regions are N-type. The red lines near the center of the rib waveguide mark the edges of the depletion region of the junction. For optimal efficiency, the P-type region, rather than the N-type, should be in the center of the rib since the refractive index change for a change in the density of holes is higher than that for electrons .
The junction works by carrier depletion. A high modulation efficiency requires the spatial change of the depletion region under an applied voltage to be large and maximally overlapping with the optical mode. Figure 1(a) shows the depletion region covers only a small portion of the rib at a bias of 0 V, whereas in Fig. 1(b), with a bias of −1 V, the depletion region expands to cover most of the waveguide core. Figures 1(c) and 1(d) show the vertical profiles of the electron and hole concentrations of the junction at the waveguide centre (x = 0 µm) under 0 V and −1 V bias voltages. At 0 V bias, the peak hole concentration in the middle of the waveguide is close to 1018 cm−3, and it is depleted to 4.5 x 1012 cm−3 at −1 V bias, indicating a large change in the carrier concentration under reverse bias voltage. The carrier concentration profiles also show the U-shaped junction can still be formed with slight ( ± 10 nm) shifts in the vertical position of the junction.
Due to the larger surface area of the depletion layer, the U-shaped junction has a high capacitance per length compared to lateral and vertical junctions without requiring high dopant concentrations (see Table 2 in Section 5). This contributes to the high modulation efficiency and low waveguide propagation loss. Thus, the U-shaped junction is efficient in producing an effective index change without a severe loss penalty.
2.2 Junction fabrication
The junction was defined in Si rib waveguides that were part of the photonic platform in . The platform was realized on 8” SOI, and the Si handle wafer had high resistivity > 750 Ω-cm. The Si rib height was 150 nm and the slab height was 65 nm. The implantation steps to form the junction were specified and designed using Sentaurus Process. Default settings for the implant damage and point defects were used. Table 1 summarizes the implantation steps. Boron and phosphorous were used for P- and N-dopants, respectively. The coordinates of the outer edges of each implantation window along the x-axis as labelled in Fig. 1 are given. To simplify the fabrication, the number of implantation steps was minimized, the implantation tilt angle was kept to 0° for all steps, and the implantation was designed to be alignment tolerant. A 0° tilt angle also allowed the PN diodes to be formed on curved waveguides, as in microring modulators.
To form the U-shaped junction, the Si was initially covered with a 10 nm thick layer of SiO2 to reduce the channeling effect from the subsequent implantation steps. The key parameter in forming the junction is the implantation energies. In Table 1, the first step formed the P-type region in the middle of the waveguide using an intermediary energy of 20 keV. The second step with a high energy of 90 keV formed the N-type region at the bottom of the waveguide. The third step, with a low energy of 15 keV, formed the N-type region near the top of the waveguide. Conveniently, the two phosphorous implantation steps shared the same window to reduce the number of masks, and overall, the total number of implantation masks (i.e., four) is the same as required for a simple lateral PN diode. Together, these three implantation steps formed the U-shaped junction. Steps 4 and 5 formed the P + + and N + + regions for the contacts. Lastly, a rapid thermal anneal at 1030°C for 5 seconds activated the dopants.
To improve the tolerance of the junction capacitance and phase-shifter efficiency to alignment errors in the implantation windows relative to the rib waveguide, the nominal Si rib width was 700 nm, but 500 nm wide ribs were also used (see Section 4). The separation between the implantation windows and the edges of rib was 90 nm nominally. For the carrier profile cross-section in Fig. 1, the simulated optical loss is 14 dB/cm at 0 V bias, and the VπL is 0.29 V·cm at a bias of −1 V for λ = 1310 nm. TCAD and electromagnetic mode simulations show that for up to ± 60 nm misalignment of the P and N implantation windows (Steps 1 to 3 in Table 1), the worst case VπL and propagation loss are 0.35 V·cm and 18 dB/cm, respectively (occurring when the boron implantation window is shifted to the left by 60 nm, and the phosphorous implantation window is shifted to the right by 60 nm). The results show that the designed junction is tolerant to mask alignment errors to within about 20% the VπL and propagation loss.
2.3 DC characteristics
The junctions were fabricated on 8” SOI as part of . Figure 2 shows the measured (using an LCR meter [Agilent E4980A]) and simulated diode capacitance as a function of the reverse bias. The measured phase-shifter was in a test structure with the same doping window positions as the bottom arm of the highest bandwidth MZM described in Section 3.3. In general, the measured capacitance and series resistance are in good agreement with the TCAD simulations. Compared to a lateral PN junction , the change in the capacitance of the U-shaped junction is greater for reverse bias between 0 V and 1 V, indicating the potential for higher efficiency. The simulated junction capacitance varies from 2.2 pF/mm to 0.34 pF/mm between 0 V and −1 V, and remains relatively constant beyond −1 V. However, at −1 V, the measured capacitance was 1.03 pF/mm, about 3 times higher than the designed value. The measured series resistance was 8.4 Ω·mm in agreement with the simulated value of 8.0 Ω·mm. The increased capacitance may have been caused by discrepancies between the simulated and fabricated doping profiles, and this affects the bandwidth of the MZM.
3. Mach-Zehnder modulator
3.1 MZM travelling-wave electrode design
We designed traveling-wave electrodes in the single-drive push-pull geometry for MZMs with a target electro-optic (EO) bandwidth beyond 30 GHz [15,16]. A 100 kΩ on-chip resistor (implemented in N-doped Si) and inductor (with a length of 6 mm and width of 2 µm wide implemented in the M1 layer) were added to the center DC line to isolate the DC bias from the RF signal . The phase-shifter length in the MZM was 2 mm. Figure 3 shows the modulator cross-section with the dimensions of the designed electrodes. The grey regions are the doped Si rib waveguides of Fig. 1. M1 and M2 are aluminum layers. Metal vias connect the Si layer with the two metal layers. The RF drive signal is applied to the two outermost lines connected to the N + + regions, and the DC bias is applied at the center P + + region. This configuration (P + + in the center) reduces the contact resistance, since the contact resistance for the N + + doped region is less than that for the P + + doped region.
Travelling-wave electrode design should consider the velocity match between the RF and optical signals, the impedance match between the electrode characteristic impedance and termination resistor, and the RF loss [11–13]. The design of the electrodes was carried out using ANSYS HFSS. First, we computed the S parameters of single-drive push-pull electrodes. Then, the S parameters were simplified into an effective RLGC model. By loading the capacitance and resistance of the TCAD simulated U-shaped junction at −1 V bias (which were, respectively, 8.0 Ω·mm and 0.34 pF/mm), we calculated the complex propagation constant and characteristic impedance of the electrodes. The widths and separations of the electrodes and metal vias were varied to converge to a design.
Figure 4 shows the calculated RF refractive index, characteristic impedance, RF loss, and the EO S21 frequency response. The black curves show the results with the designed U-shaped junction capacitance and the red curves are the results with the measured capacitance at −1 V. The expected EO S21 of the Si MZM is calculated using equations in [30,31] as the fraction of the average RF voltage across the junction, Vdep, relative to the average voltage at the electrode, Vavg. Thus, the EO modulation response, m(ωm), as a function of the modulation frequency, ωm, is given by
In Eqs. (1)-(3), ZSi is the PN junction series resistance, Cdep is the PN junction capacitance, ω0 is a reference modulation frequency; Vg is the voltage amplitude of the driving signal; ZT and ZS are the impedances of the load and source respectively; Z0 is the characteristic impedance of MZM; γm is the complex microwave propagation constant; βo = ωmno/c is the propagation constant in the optical waveguide with no as the optical group refractive index . Both Z0 and γm can be calculated in the effective circuit model of the electrodes. L is the device length, which is 2 mm. The EO 3 dB bandwidth is defined as the frequency when m(ωm) is reduced by 50%.
With the designed capacitance, the RF refractive index at 30 GHz matches well with the optical group index of 3.7; the characteristic impedance is around 50 Ω; and the RF loss is < 3.3 dB/mm at 30 GHz. The calculated EO S21 shows the expected 3 dB bandwidth of 38 GHz. The intrinsic, RC limited bandwidth of the designed U-shaped junction is about 56 GHz and is not the main limitation to the bandwidth. However, the higher than expected junction capacitance compromises the bandwidth. The red curves show that with the measured junction capacitance, for modulation frequencies < 30 GHz, the expected RF refractive index is > 4, the characteristic impedance is < 40 Ω, and the RF loss is < 8.9 dB/mm. Figure 4(d) shows that with the simulated capacitance, the expected EO 3dB is reduced to 11 GHz with the measured junction capacitance at −1 V.
3.2 MZM device
Figure 5 shows the optical micrograph of a fabricated Si MZM. It consisted of two 3 dB multimode interference (MMI) couplers separated by the waveguides with the U-shaped junctions. Single-mode Si channel waveguides routed light to and from the MMIs and adiabatic linear tapers were used to couple light between the routing waveguides and the 700 nm wide, multimode U-shaped junction rib waveguides. A path length difference of 40 µm between the Mach-Zhender arms was implemented to enable measurements by varying the input wavelength. The measured device free spectral range (FSR) was 10 nm near a wavelength of 1310 nm. The RF drive signal was applied using a GS probe, and the travelling-wave electrodes were terminated off-chip using an SG probe with a 50 Ω impedance. The DC bias was set by the DC pad. The device used inverse tapered edge couplers, and light was launched into and collected from the MZM using lensed fibers with a spot diameter of 2.5 μm. The IL of the test setup was normalized to the transmission on a straight waveguide on chip, and the device IL was found to be 2.7 dB. The phase-shifter loss was 12.5 dB/cm at 0 V for the highest bandwidth MZM to be described in Section 3.3.
3.3 DC characterization
To measure the DC tuning of the MZM, both the DC pad and the lower arm of the MZM were connected to ground, and a DC bias was applied to the top arm. The phase-shift was calculated from the shift of the transmission spectrum as a function of the applied bias.
Figure 6 shows the MZM transmission spectrum at different bias voltages for two devices with deliberately designed offsets in their doping window positions. Figure 6(a) shows the result for the device with the highest bandwidth, and Fig. 6(b) shows the result for device with the highest efficiency. The extracted phase-shifts from the spectra are shown in Fig. 6(c). The device with the highest bandwidth had a boron implantation window designed with a + 60 nm shift in the mask along the x-axis in Fig. 1 for the top phase shift arm. The device with the highest efficiency had a designed boron implantation window shifted + 60 nm and phosphorous implantation window shifted −60 nm in the mask for the top phase shift arm. These results translate to a DC VπL of 0.26 V·cm at −0.5 V (red line) for the most efficient phase-shifter and a DC VπL of 0.46 V·cm at −0.5 V (black line) for the phase-shifter with the highest bandwidth. At a bias of −1 V, the average VπL of the two arms of the highest bandwidth MZM was about 0.61 V·cm; at a bias of −2 V, VπL was about 0.94 V·cm. At higher reverse biases, the efficiency diminished since the increase in the area of the depletion region reduced.
3.4 High-frequency characterization
We carried out S parameter and eye pattern measurements of the device with the highest bandwidth. The S parameters were measured using a vector network analyzer (VNA) (Agilent N5227A) and a 50 GHz photodetector (Finisar XPDV2320R). The RF cables and signal probe were de-embedded from the measurement (Agilent N4694-60001 E-cal kit, GGB Industries CS-8 substrate with SOLT calibration). Figures 7(a) and 7(b) show the measured electrical S11 and EO S21 of the MZM at 0 V and −2 V bias. The wavelength was at the quadrature point of the MZM (−3dB optical transmission point) and input RF power was 0 dBm. The S11 was < −14 dB over a 30GHz bandwidth, showing the RF reflection was low. The ripples in the S11 were likely due to the external 50Ω termination, for which its RF probe was not calibrated. The EO S21 in Fig. 6(b) shows a 3dB bandwidth of 4 GHz at 0 V bias and extending to ~13 GHz at −2 V bias. The EO S21 is plotted as 10log(Popt/Vin), where Popt is the small-signal optical power modulation and Vin is the small-signal input modulation voltage to the device, to compare with the eye diagrams, which were measured for optical (not electrical) powers. For this device, the EO bandwidth at −2 V bias was similar to that at −1V (curve not shown), which may be caused by a higher than expected capacitance in the phase-shifter of the top arm of the MZI at −2 V bias (the result in Fig. 2 was for a test structure matching the bottom arm of the MZI).
Figure 8 shows the measured eye diagrams of the device for 231-1 pseudo-random binary sequence (PRBS) patterns. Drive signals from the pattern generator (SHF 78210D, 12104A) were amplified to 7.5 Vpp (Microsemi OA4MVM3) and then attenuated. The drive signal amplitude was calibrated to be 2.88 Vpp considering the attenuator and connector loss. Light from a tunable laser was amplified using an O-band semiconductor optical amplifier and bandpass filtered (passband bandwidth of 1 nm) prior to input to the MZM to overcome the IL of the device and setup. The eye patterns were captured using a sampling oscilloscope (Agilent 86100C, 86106B). For these measurements, the ER was sacrificed for the EO bandwidth as the junctions were biased at a strong reverse bias of −2.4 V to reduce the capacitance. Figure 8(a) shows the eye pattern at 16 Gb/s with a RF voltage of 2.88 Vpp. An extinction ratio (ER) of 2.6 dB was achieved with the input wavelength set at the MZM quadrature point. Figures 8(b) and 8(c) show the eye patterns at 20 and 24 Gb/s with the same bias conditions. The ERs were 2.4 dB and 2.2 dB at 20 Gb/s and 24 Gb/s, respectively.
4. Microring modulator
Because of the normal incident dopant implantation, we were also able to realize microring modulators using the U-shaped junction. The microring modulators, due to their compact sizes, did not require travelling-wave electrodes, which enable high-speed modulation tests of the U-shaped junction phase-shifters. Figure 9(a) shows an optical micrograph of the ring modulator, which contained the U-shaped PN junction. The ring had a diameter of 65 µm and rib waveguide width of 500 nm. The electrical probe pads were in a GSG configuration, and each had a size of 69 μm x 69 μm. The N region was in the center of the ring. The DC tuning efficiency was measured using the shift of the microring transmission spectrum as a reverse bias is applied to the modulation junction. Figure 9(b) shows the spectral tuning at a few bias voltages. The microring linewidth was about 50 pm (8.6 GHz) at 0 V bias. The tuning efficiency was 40 pm·V−1 (7 GHz·V−1) between 0V and −0.5V near a wavelength of 1310 nm.
Figure 10(a) shows the EO S21 taken near the −2.55 dB transmission point using a VNA (Agilent N5232A) and a 38 GHz photoreceiver (Newport 1474-A). The 3dB EO bandwidth was roughly 9.8 GHz at 0V bias, extending to 13.5 GHz at a bias of −1V. Figure 10(b) shows a 13 Gb/s eye pattern for a PRBS 231-1 pattern at a wavelength of 1310.58 nm captured using the sampling oscilloscope. The eye pattern is wide open. The ER was 10 dB at an IL of 2.5 dB (matching the EO S21 measurements) and was measured for a drive voltage of only 1.6Vpp at 0V bias applied to a GSG probe. The measured bit rate was limited by the available pattern generator at the time of the experiment.
Table 2 compares the results of this work with other phase-shifters and their implementations in MZMs. The emphasis of the comparison is the junction and phase-shifter, rather than system level performance , which depends on, for example, the type of electrodes (e.g., travelling wave or segmented ) and driver electronics.
To attain the most efficient Si modulation phase-shifter, since the refractive index change is caused by a carrier density change, the strategy is to increase the capacitance per length, while reducing the series resistance to maintain a high junction modulation bandwidth resistance . In addition to the capacitance, another figure of merit to compare the modulation sections is the loss-efficiency product, LEP, which is given by the product of the VπL (in V·cm) and waveguide loss (in dB/cm) as
Compared to the other modulation phase-shifters, the U-shaped PN junction has amongst the highest capacitance per length and the most optimal, i.e., the lowest, LEP between 3.25 and 5.75 V·dB. The LEP for the other phase-shifters range between 13 and 29.0 V·dB. The high efficiency of our demonstrated phase-shifter is the closest to the SISCAP carrier accumulation type. The major advantages are that the U-shaped junction has low optical losses similar to that of lateral PN diode phase-shifters and can be integrated in modulator geometries involving waveguide bends, such as microrings.
As a comparison of the high-speed modulation efficiency at the MZM device level, we define the modulation depth per volt, MD, as
Table 2 shows this work achieved a low MD despite the best LEP. This is due to the higher than expected capacitance realized, and the need to bias the U-shaped junction at a high reverse bias to better match the designed electrodes. Achieving a higher dynamic modulation efficiency would require reducing the reverse bias to use the high capacitance regime of the junction, which would lead to a high RF loss and impedance and velocity mismatch, as discussed in Section 3.1. In the future, electrodes that drive lumped elements of short phase-shifters can be used instead as in .
Table 3 compares our microring with other reported microring modulators. The microring modulator allows for a comparison of the phase-shifters without the limitations of the travelling-wave electrodes. We define a figure of merit (FOM) asEq. (6) is described in the Appendix. The FOM shows that the modulation efficiency of the U-shaped junction is superior to the vertical , interdigitated , and lateral PN junctions [38,39]. Vertical and interdigitated PN junctions have similar efficiencies, while lateral PN junctions have lower efficiencies. The U-shaped junction microring also exhibits amongst the highest MD amongst ring modulators. The MD demonstrated in this work is superior to lateral and interdigitated PN junctions, and is similar to that of the vertical PN junction. The absolute MD can be improved by increasing the absolute tuning efficiency, η, by reducing the size of the microring to increase the finesse (see Appendix). Simultaneously, the increased FSR enables a broader EO bandwidth since the optical linewidth can be higher.
In summary, we have demonstrated efficient and low-loss U-shaped PN junction phase-shifters for Si MZMs and microring modulators fabricated on 8” SOI as part of an integrated photonics platform. The fabrication of the junction is simple and does not require extra mask steps. The highest bandwidth Si MZM had a DC VπL of 0.46 V·cm at −0.5 V bias in the O-band and eye patterns at up to 24 Gb/s were measured. A DC VπL as low as 0.26 V·cm was measured. The microring modulator achieved a tuning efficiency is of 40 pm·V−1 and open eye patterns at 13 Gb/s were obtained. Future improvements include re-designing the electrodes to suit the experimentally realized junction capacitance and resistance to increase the EO bandwidth and to boost the microring finesse to increase the tuning efficiency. The U-shaped PN junction has superior metrics compared to demonstrated modulator phase-shifters in Si. This work shows the promise of this type of PN junction for power efficient Si modulators. Such high modulation efficiency junctions can make possible very short MZMs that obviate the need for traveling-wave electrodes.
In this section, we show that the defined FOM in Eq. (6) represents the phase-shifter efficiency. We begin by assuming a ring with circumference L and a non-resonant phase-shifter with the same length L. The phase-shift per volt incurred by propagating in the ring on resonance, , relative to the waveguide phase-shifter, is proportional to the finesse, F :Eq. (7) and using Eq. (8), we obtain1]. Therefore, Eq. (10) shows that to increase η for a given kind of junction, the finesse should be increased.
The financial support of the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged.
We thank CMC Microsystems for the loan of an LCR meter for the junction DC characterization, as well as loan of a VNA. Access to ANSYS HFSS and Synopsys Sentaurus was also subsidized by CMC Microsystems.
References and links
1. G. T. Reed, G. Mashanovich, F. Y. Gardes, and D. J. Thomson, “Silicon optical modulators,” Nat. Photonics 4(8), 518–526 (2010). [CrossRef]
2. T.-Y. Liow, K.-W. Ang, Q. Fang, J.-F. Song, Y.-Z. Xiong, M.-B. Yu, G.-Q. Lo, and D.-L. Kwong, “Silicon modulators and germanium photodetectors on SOI: monolithic integration, compatibility, and performance optimization,” IEEE J. Sel. Top. Quantum Electron. 16(1), 307–315 (2010). [CrossRef]
3. M. Streshinsky, R. Ding, Y. Liu, A. Novack, Y. Yang, Y. Ma, X. Tu, E. K. S. Chee, A. E. J. Lim, P. G. Q. Lo, T. Baehr-Jones, and M. Hochberg, “Low power 50 Gb/s silicon traveling wave Mach-Zehnder modulator near 1300 nm,” Opt. Express 21(25), 30350–30357 (2013). [CrossRef] [PubMed]
4. T. Baehr-Jones, R. Ding, Y. Liu, A. Ayazi, T. Pinguet, N. C. Harris, M. Streshinsky, P. Lee, Y. Zhang, A. E. J. Lim, T. Y. Liow, S. H.-G. Teo, G. Q. Lo, and M. Hochberg, “Ultralow drive voltage silicon traveling-wave modulator,” Opt. Express 20(11), 12014–12020 (2012). [CrossRef] [PubMed]
6. D. Thomson, F. Gardes, J. Fedeli, S. Zlatanovic, Y. Hu, B. Kuo, E. Myslivets, N. Alic, S. Radic, G. Z. Mashanovich, and G. T. Reed, “50Gbit/s silicon optical modulator,” IEEE Photonics Technol. Lett. 24(4), 234–236 (2012). [CrossRef]
7. P. Dong, S. Liao, D. Feng, H. Liang, D. Zheng, R. Shafiiha, C. C. Kung, W. Qian, G. Li, X. Zheng, A. V. Krishnamoorthy, and M. Asghari, “Low Vpp, ultralow-energy, compact, high-speed silicon electro-optic modulator,” Opt. Express 17(25), 22484–22490 (2009). [CrossRef] [PubMed]
8. M. Webster, C. Appel, P. Gothoskar, S. Sunder, B. Dama, and K. Shastri, “Silicon photonic modulator based on a MOS-capacitor and a CMOS driver,” in CSICs (2014), pp. 1–4.
9. G. T. Reed, G. Z. Mashanovich, F. Y. Gardes, M. Nedeljkovic, Y. Hu, D. J. Thomson, K. Li, P. R. Wilson, S. W. Chen, and S. S. Hsu, “Recent breakthroughs in carrier depletion based silicon optical modulators,” Nanophotonics 3(4–5), 229–245 (2014).
10. D. Miller, “Device requirements for optical interconnects to silicon chips,” Proc. IEEE 97(7), 1166–1185 (2009). [CrossRef]
11. R. Ding, Y. Liu, Y. Ma, Y. Yang, Q. Li, A. E.-J. Lim, G.-Q. Lo, K. Bergman, T. Baehr-Jones, and M. Hochberg, “High-speed silicon modulator with slow-wave electrodes and fully independent differential drive,” J. Lightwave Technol. 32(12), 2240–2247 (2014). [CrossRef]
12. R. Ding, Y. Liu, Q. Li, Y. Yang, Y. Ma, K. Padmaraju, A. E. J. Lim, G. Q. Lo, K. Bergman, T. Baehr-Jones, and M. Hochberg, “Design and characterization of a 30-GHz bandwidth low-power silicon traveling-wave modulator,” Opt. Commun. 321, 124–133 (2014). [CrossRef]
13. J. Ding, H. Chen, L. Yang, L. Zhang, R. Ji, Y. Tian, W. Zhu, Y. Lu, P. Zhou, and R. Min, “Low-voltage, high-extinction-ratio, Mach-Zehnder silicon optical modulator for CMOS-compatible integration,” Opt. Express 20(3), 3209–3218 (2012). [CrossRef] [PubMed]
14. J. Ding, R. Ji, L. Zhang, and L. Yang, “Electro-optical response analysis of a 40 Gb/s silicon Mach-Zehnder optical modulator,” J. Lightwave Technol. 31(14), 2434–2440 (2013). [CrossRef]
15. L. Chen, P. Dong, and Y.-K. Chen, “Chirp and dispersion tolerance of a single-drive push–pull silicon modulator at 28 Gb/s,” IEEE Photonics Technol. Lett. 24(11), 936–938 (2012). [CrossRef]
17. D. Patel, S. Ghosh, M. Chagnon, A. Samani, V. Veerasubramanian, M. Osman, and D. V. Plant, “Design, analysis, and transmission system performance of a 41 GHz silicon photonic modulator,” Opt. Express 23(11), 14263–14287 (2015). [CrossRef] [PubMed]
18. M. R. Watts, W. A. Zortman, D. C. Trotter, R. W. Young, and A. L. Lentine, “Low-voltage, compact, depletion-mode, silicon Mach-Zehnder modulator,” IEEE J. Sel. Top. Quantum Electron. 16(1), 159–164 (2010). [CrossRef]
19. M. R. Watts, W. A. Zortman, D. C. Trotter, R. W. Young, and A. L. Lentine, “Vertical junction silicon microdisk modulators and switches,” Opt. Express 19(22), 21989–22003 (2011). [CrossRef] [PubMed]
21. T. Latchu, M. Pochet, N. G. Usechak, C. DeRose, A. Lentine, D. C. Trotter, and W. Zortman, “Power-penalty comparison of push-pull and traveling-wave electrode Silicon Mach-Zehnder modulators,” in Optical Interconnects Conference (2014), pp. 25–26. [CrossRef]
22. Z. Y. Li, D. X. Xu, W. R. McKinnon, S. Janz, J. H. Schmid, P. Cheben, and J. Z. Yu, “Silicon waveguide modulator based on carrier depletion in periodically interleaved PN junctions,” Opt. Express 17(18), 15947–15958 (2009). [CrossRef] [PubMed]
23. H. Yu, M. Pantouvaki, J. Van Campenhout, D. Korn, K. Komorowska, P. Dumon, Y. Li, P. Verheyen, P. Absil, L. Alloatti, D. Hillerkuss, J. Leuthold, R. Baets, and W. Bogaerts, “Performance tradeoff between lateral and interdigitated doping patterns for high speed carrier-depletion based silicon modulators,” Opt. Express 20(12), 12926–12938 (2012). [CrossRef] [PubMed]
24. H. Xu, X. Xiao, X. Li, Y. Hu, Z. Li, T. Chu, Y. Yu, and J. Yu, “High speed silicon Mach-Zehnder modulator based on interleaved PN junctions,” Opt. Express 20(14), 15093–15099 (2012). [CrossRef] [PubMed]
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]
28. Y. Liu, S. Dunham, T. Baehr-Jones, A. E. J. Lim, G. Q. Lo, and M. Hochberg, “Ultra-responsive phase shifters for depletion mode silicon modulators,” J. Lightwave Technol. 31(23), 3787–3793 (2013). [CrossRef]
29. W. D. Sacher, Z. Yong, J. C. Mikkelsen, A. Bois, Y. Yang, J. C. Mak, P. Dumais, D. Goodwill, C. Ma, J. Jeong, E. Bernier, and J. K. Poon, “Multilayer silicon nitride-on-silicon integrated photonic platform for 3D photonic circuits,” in CLEO (2016), paper JTh4C.3.
30. H. Yu and W. Bogaerts, “An equivalent circuit model of the traveling wave electrode for carrier-depletion-based silicon optical modulators,” J. Lightwave Technol. 30(11), 1602–1609 (2012). [CrossRef]
32. G. Ghione, Semiconductor Devices for High-speed Optoelectronics (Cambridge University Press, 2009), Chap. 6.
33. S. Sharif Azadeh, J. Müller, F. Merget, S. Romero-García, B. Shen, and J. Witzens, “Advances in silicon photonics segmented electrode Mach-Zehnder modulators and peaking enhanced resonant devices,” in Photonics North (ISOP, 2014).
34. H. Jayatilleka, W. D. Sacher, and J. K. S. Poon, “Analytical model and fringing-field parasitics of carrier-depletion silicon-on-insulator optical modulation diodes,” IEEE Photonics J. 5(1), 2200211 (2013). [CrossRef]
35. F. Boeuf, S. Crémer, E. Temporiti, M. Ferè, M. Shaw, C. Baudot, N. Vulliet, T. Pinguet, A. Mekis, G. Masini, H. Petiton, P. Le Maitre, M. Traldi, and L. Maggi, “Silicon photonics R&D and manufacturing on 300-mm wafer platform,” J. Lightwave Technol. 34(2), 286–295 (2016). [CrossRef]
36. IMEC’s Integrated Si-Photonics Platform, (ISIPP50G), www2.imec.be/content/user/File/Brochures/leaflets 2016/IMEC'S INTEGRATED SILICON PHOTONICS PLATFORM ISIPP50G + .pdf.
37. C. T. DeRose, D. C. Trotter, W. A. Zortman, and M. R. Watts, “High speed travelling wave carrier depletion silicon Mach-Zehnder modulator,” in Optical Interconnects Conference (2012), pp. 135–136. [CrossRef]
38. Z. Xuan, Y. Ma, Y. Liu, R. Ding, Y. Li, N. Ophir, A. E. J. Lim, G.-Q. Lo, P. Magill, K. Bergman, T. Baehr-Jones, and M. Hochberg, “Silicon microring modulator for 40 Gb/s NRZ-OOK metro networks in O-band,” Opt. Express 22(23), 28284–28291 (2014). [CrossRef] [PubMed]
39. A. Malacarne, F. Gambini, S. Faralli, J. Klamkin, and L. Potì, “High-speed silicon electro-optic microring modulator for optical interconnects,” IEEE Photonics Technol. Lett. 26(10), 1042–1044 (2014). [CrossRef]
40. E. Timurdogan, C. M. Sorace-Agaskar, E. S. Hosseini, and M. R. Watts, “An interior-ridge silicon microring modulator,” J. Lightwave Technol. 31(24), 3907–3914 (2013). [CrossRef]
41. W. D. Sacher, W. M. J. Green, S. Assefa, T. Barwicz, H. Pan, S. M. Shank, Y. A. Vlasov, and J. K. S. Poon, “Coupling modulation of microrings at rates beyond the linewidth limit,” Opt. Express 21(8), 9722–9733 (2013). [CrossRef] [PubMed]