Open Access
Add to BibSonomy
Abstract
We developed a high-speed and high-efficiency narrow-width metal-oxide-semiconductor (MOS) capacitor-type Si optical modulator (Si-MOD) by applying TM optical mode excitation. We designed and fabricated an optical-mode-converter structure from TE to TM mode. Even in the case of a 200-nm width, the Si MOS-MOD showed high-modulation efficiency in TM mode (about 0.18 Vcm), and the electrical capacitance decreased as the MOS junction width decreased. We also demonstrated high-speed operation at 32 Gbps and 40 Gbps for the 30-µm-long Si MOS-MOD in TM mode.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Silicon photonics has recently attracted much attention because it offers low cost, low power consumption, and high bandwidth for optoelectronic solutions for applications ranging from telecommunications to chip-to-chip interconnects [1]. To realize an effective photonics-electronics convergence system, it is very important to have a high-speed and high-efficiency Si-MOD with low power consumption [2,3].
Among the various Si-MODs demonstrated so far, Mach-Zehnder Si-MODs based on the free-carrier plasma dispersion effect have been reported for high-speed modulation and broad-wavelength operation [4–11]. However, carrier-depletion Si modulators need a relatively long phase-shifter or high driving voltage because of the weak plasma dispersion effect in Si [12], and they are not favorable for large-scale integration. It has been reported that ring-resonator type Si-MODs are capable of high-speed and low-power operation. However, the operation wavelength needs to be controlled to an accuracy of less than 1 nm, which makes their practical use difficult [13,14].
As for MOS capacitor-type Si-MODs, high efficiency has been achieved by accumulating free-carriers at the gate-oxide/silicon interface [15,16]. However, a more efficient and high-speed MOS type Si-MOD is a challenge to realize low power and high-bandwidth optical interconnect. It has been reported large electrical capacitance contribute to high-modulation efficiency, but there is trade-off relationship between modulation efficiency and EO (electro-optical) bandwidth. Recently many studies of enhancing the carrier plasma effect have been demonstrated [17–21]. However, heterogeneous integration is not easy in the conventional CMOS fabrication process. In addition, slot mode of silicon-insulator-silicon structure have recently been reported, but the detail analysis for low power Si-MOD has not been studied [22–24]. It is very important to design and demonstrate the low power and high-speed Si MOS-MOD with the same stack structure for TE mode, which realize the easy fabrication process and makes it possible to integrate TE and TM modes of Si MOS-MODs on the same optical circuit.
In this study, we design and fabricate a mode-converter from TE to TM mode for MOS-capacitor-type Si-MODs. It is found that by decreasing the rib-waveguide width from 400-nm to 200-nm, the resulting Si MOS-MOD shows high-modulation efficiency in TM mode (about 0.18 Vcm) and a smaller electrical capacitance. We also demonstrate high-speed operation of 32 Gbps and 40 Gbps for a 30-µm-long Si-MOS-MOD in TM mode. In this case, power consumption for the Si-MOS-MOD is estimated to be 0.69pJ/bit, which is one of the most efficient among the Si-based Mach-Zehnder Si-MODs. We propose the further improvement of electrical capacitance by using the asymmetric MOS junction structure with TM mode excitation.
2. Design and fabrication
Figures 1(a) and (b) show a schematic diagram of the Si-MOD with a MOS junction and calculated optical power-density contour map of TM0 mode for the Si-MOD at a wavelength of 1.31 µm in case of 400-nm width of MOS junction. In Fig. 1(b), the optical power-density is normalized by the maximum value. The Si-MOD consists of an asymmetric Mach-Zehnder interferometer (MZI) structure with a 20-µm long delay-line Si-WG. For TE mode, the Si MOS-MOD has no optical mode converter structure from TE0 to TM0.The fabrication process started with 4-inch silicon-on-insulator (SOI) wafers with an SOI thickness of 180 nm for the 1.31-µm wavelength. After a 3-nm-thick gate-oxide was grown by thermal oxidation, an amorphous-silicon layer was deposited by low-pressure chemical vapor deposition (LP-CVD) and recrystallized by two-step annealing [25]. Then, Si waveguides (Si-WGs) and poly-Si gate electrode layers were patterned by electron beam lithography and dry etching. After deposition of the SiO2 upper-clad layer, contact-holes were formed. Finally, stacked electrodes composed of Ti/TiN/Al layers were deposited and patterned. The dopant densities of p-Si and n-poly-Si were 1 x1018/cm3 and 2 x1018/cm3, respectively. The phase shifter lengths in the experiments were 30-120 µm.
Fig. 1. (a) Schematic diagram of MOS-capacitor-type Si modulator for 1.31-µm wavelength. (b) Calculated optical-power-density contour map of TM0 mode in case of 400-nm width of MOS junction.
Figures 2(a) and (b) show the TEM image of the fabricated MOS-capacitor-type Si-MOD with MOS junction widths of 200 nm and 400 nm. The MOS junction width was defined as the Si rib-WG width with p-type doping. According to the TEM images, an upper electrode of the n-poly-Si was smoothly stacked on the p-type doped Si rib WG via a gate-oxide layer which is the lower electrode.
Fig. 2. TEM image of MOS-capacitor-type Si-MOD with MOS junction widths of (a) 200 nm and (b) 400 nm.
The Si-MOS-MOD for TE and TM mode excitation was designed in a linked simulation consisting of carrier-transport and optical-mode analyses, and the finite difference time domain (FDTD) method. The effective refractive index of the cross-sectional Si MOS-MOD with free carrier plasma effect is calculated from refractive index change in Si with the free carrier density distribution by T-CAD. TE and TM mode converter structure was analyzed by the FDTD method, and the excited TE and TM mode power density was analyzed by commercial optical mode solver based on the finite element method [26].
3. Results and discussion
3.1 TE0 to TM0 mode converter structure, modulation efficiency, and optical loss
The TE0-to-TM0 optical mode converters are connected to the Si MOS-MOD both at the input and output ports. In this study, we applied two-step optical mode converters for TE0-TE1 and TE1-TM0 to decrease the optical loss. In the channel and the rib-WG connecting structure, optical loss for TM0 mode is relatively large due to the surface and edge roughness of the poly-silicon-based tapered spot-size converter structure. Two step optical mode converters are expected to lower the optical loss because TE1 optical mode is input into the channel and rib-WG connecting structure and TM0 optical mode is induced in the rib WG structure with a wider poly-silicon structure.
Figures 3 and 4 show a schematic diagram and an electric-field intensity contour map of the optical mode converter from TE1 to TM0 mode. TE1 mode is excited in the Si rib-WG by using an asymmetric directional coupler (ADC) structure between a single-mode rib-WG and multimode rib-WG. The ADC structure realizes close to 100% mode converting between TE1 and TE0 mode [27]. The input Si rib-WG has a width of 340 nm. The gap of the 160-µm-long directional coupler is 230-nm at the input side and 210-nm at the output side. The Si WG for TE1 mode is 700-nm wide. The TE1 mode is input into the Si rib-WG through a tapered silicon-insulator-silicon (S-I-S) junction structure, and optimization of the tapered S-I-S junction structure enables the TM0 mode of the slot mode to be excited in the MOS junction structure. The rib-WG and the slab height are 180 nm and 80 nm, respectively. From the simulation, about 90% of the optical power can be converted from TE1 to TM0 mode.
Fig. 3. Schematic diagram and optical power density contour map of optical mode converter from TE0 to TE1 mode: (a) upper view, (b) cross-section, and (c) dimensions of optical mode converter of ADC.
Fig. 4. Schematic diagram and optical power density contour map of optical mode converter from TE1 to TM0 mode.
Figure 5 shows electric-field contour maps of the optical mode converter from TE1 to TM0 mode for Ex and Ey from the FDTD simulation. The optical mode converter changes the Ex field in the TE1 mode into the Ey field in the TM0 mode in channel to rib-WG connecting structure.
Fig. 5. Normalized electric-field-intensity contour map of optical mode converter from TE1 to TM0 mode for (a) Ex and (b) Ey.
3.2 Modulation efficiency and optical loss
Figure 6, 7, and 8 show the simulated optical-power-density contour maps for TE and TM modes for junction widths for 0.4-µm, 0.2-µm, and 0.1-µm when -50dBm optical power is input. Optical-power-density contour maps are shown as a unit of W/cm2. In TE mode, the optical mode field spreads outside the rib WG and into the upper electrode layer as the rib WG width decreases. In TM mode, the optical mode field is much more confined in the rib WG even for the 100-nm-width. Therefore, the optical mode field can be maintained within the rib WG in the case of TM mode excitation, which leads to improved modulation efficiency in TM mode in comparison with TE mode. In this study, minimum line width of the rib WG is about 0.2-µm due to the relatively large EB resist thickness to etch the deep thick Si. In near future, 0.1-µm wide rib waveguide can be achieved by high-accuracy lithography technique of immersion ArF lithography [28].
Fig. 6. Simulated optical power density contour maps for junction widths of 0.4µm for (a) TE and (b) TE mode excitation.
Fig. 7. Simulated optical power density contour maps for junction widths of 0.2µm for (a) TE and (b) TE mode excitation.
Fig. 8. Simulated optical power density contour maps for junction widths of 0.1µm for (a) TE and (b) TE mode excitation.
Figure 9 shows the simulated dependence on the rib WG width for (a) VπL for TE and TM mode and (b) electrical capacitance at -1Vdc for symmetric, asymmetric electrode structures, and MOS junction structure with the same width of an upper electrode of a poly-Si layer as that of the MOS junction. For TE mode, the modulation efficiency decreases as the Si rib WG width of MOS-junction decreases, because the overlap between the optical mode and carrier-density modulation region decreases as the MOS junction width decreases for TE mode. In contrast, for TM mode, high modulation efficiency of about 0.2 Vcm is obtained for the 200-nm width, which lowers the electrical capacitance. As well, for TM mode, the optical mode is confined within the Si rib WG of the MOS-junction and high-modulation efficiency is obtained even in the case of the narrowest rib WG. In Fig. 9(b), electrical capacitance decreases linearly as the rib WG width decreases. As a relatively large stray capacitance exists between the upper and lower electrodes of the n- and p- type Si layers in the Si MOS-MOD structure, especially in carrier accumulation mode, the capacitance of the Si MOS-MOD is not proportional to the MOS junction width. Capacitance decreases by about 30% when the rib WG width is decreased from 400 nm to 200 nm at -1 Vdc. Figure 9(c) shows the simulated electric field distribution for 0.2-µm-wide MOS junction with symmetric electrode structure. The stray electric field arises around the MOS junction for the symmetric electrode structure. It is suggested that the asymmetric electrode structure with TM mode excitation would drastically decrease the electric capacitance, maintaining the high-modulation efficiency.
Fig. 9. Simulated dependence on rib WG width of (a) VπL for TE and TM mode and (b) capacitance at -1Vdc for symmetric, asymmetric electrode structures, and MOS junction structure without extending the upper electrode. (c) Simulated electric field distribution for 0.2-µm-wide MOS junction with symmetric electrode structure.
Figure 10 shows the simulated transmission losses for (a) TE and (b) TM mode for MOS junction widths from 0.1µm to 0.4µm. For TE mode excitation, optical loss decreases as the MOS junction width is decreased from 0.4µm to 0.2µm in the case of free carriers accumulating at the MOS junction, because the optical confinement factor in the Si WG of the MOS junction becomes smaller. However, the optical transmission loss drastically increases in the case of the 0.1-µm-wide MOS junction, because the optical mode spreads into the upper Si layer in Fig. 8(a), which results in a larger optical loss in the n-type highly doped Si region of the upper electrode.
Fig. 10. Simulated transmission loss for (a) TE and (b) TM mode for MOS junction widths from 0.1µm to 0.4µm.
On the other hand, as for TM mode excitation of the slot mode, the optical mode is confined at the MOS junction, and the transmission optical loss does not depend very much on the MOS junction width. In the case of the 0.1µm width, the optical mode field leaks out into the SiO2 layer adjacent to the MOS junction WG in Fig. 8(b), and the optical transmission loss decreases.
Figure 11 shows the measured and simulated phase shifts for 400-nm-wide and 200-nm-wide 120-µm long Si-MOS-MODs in case of TE and TM optical mode excitation. The phase sift was estimated from the transmission spectra shift with applied voltage for the asymmetric MZI structure of the Si MOS-MOD. For TE optical mode, the modulation efficiency of VπL decreased from 0.16 Vcm to 0.26 Vcm at -1 V to -4 V dc bias voltage (Vdc) when the junction width was decreased from 400 nm to 200 nm in Fig. 9(a). On the other hand, for TM optical mode, 0.18 Vcm was obtained for the 200-nm-wide MOS junction, which was comparable to the value for the 400-nm-wide MOS junction in Fig. 9(b). Therefore, optical TM-mode excitation is effective for high-modulation efficiency and low electrical capacitance when the MOS junction width is decreased. The measured results are almost consistent with the simulated ones in Fig. 11(c) and (d). The small difference at zero-to-one applied voltage among the experimental and simulated results would result from small fraction of phosphorus ions penetration into the surface region of the p-type Si rib WG at the MOS junction in the fabrication process, which has been examined by SIMS analysis. Penetrated phosphorus ions would neutralize the surface region of p-type doped Si a little, which result in the smaller optical phase shift at zero-to-one applied voltage than the simulated one. However, large optical phase shift could be induced with applied voltage more than flat band voltage of about 1 V, which corresponds to carrier-accumulation mode. Small deviation of estimated VπL from the simulation and experiment would come from the phosphorus ions penetration into the surface region of the p-type Si.
Fig. 11. Measured phase shift for (a) TE and (b) TM optical mode excitation of 120-µm long Si-MOS-MODs with 400-nm and 200-nm widths. Simulated phase shift for (c) TE and (d) TM optical mode excitation of 120-µm long Si-MOS-MODs with 400-nm and 200-nm widths.
Figure 12(a) shows the transmission spectra of 120-µm-long Si-MOS-MODs of an asymmetric MZI without dc bias voltage for TE and TM optical mode excitation in the case of a 400-nm-wide MOS junction. Results for a reference Si channel WG are shown for comparison. FSR in the transmission spectra of 120-µm-long Si-MOS-MOD is 18.9 nm and the difference in maximum transmission and reference Si channel WG corresponds to the optical loss of the Si MOS-MOD with input and output structure of TE0-TM1 mode converters. Figure 12(b) shows transmission of 120-µm-long phase-shifter of a Si-MOS-MOD with input and output port structures of TE0-TM1 mode converters and a reference Si channel WG. The variation of the transmission of the Si MOS-MOD dependence on wavelength would originate from the ADC dependence on wavelength and poly-Si taper structure alignment accuracy to the Si-WG. The optical loss of the mode converter including the input and output ports is estimated to be about 3.5 dB around the wavelength of 1310 nm, which means excess loss in converting the optical mode from TE to TM is about 2 dB compared with that of the simulation. The excess loss would mainly originate from the alignment accuracy and optical scattering at the tapered poly-Si structures in the mode converters.
Fig. 12. (a) Transmission spectra of 120-µm-long Si-MOS-MOD of asymmetric MZI with 400-nm-wide MOS junction in the case of TE or TM optical mode excitation and reference Si channel WG. (b) Transmission of 120-µm-long phase-shifter of Si-MOS-MOD with input and output structure of TE0-TM1 mode converters and reference Si WG.
3.3 High-speed characteristics of TM-mode Si MOS-MOD
Figure 13(a) shows the measured frequency dependence of the normalized EO response for 200-nm and 400-nm-wide 30-µm-long Si MOS-MODs at -1 Vdc, which is comparable to the flat-band voltage. The 200-nm-wide Si MOS-MOD had an 18 GHz bandwidth, while the 400-nm-wide Si MOS-MOD had a 9.7 GHz bandwidth. Therefore, the TM mode has the advantage of a wider-bandwidth with decreasing Si MOS-MOD width, while it maintains high modulation efficiency. The 120-µm-long Si MOS-MOD shows a 16 GHz bandwidth for the 200-nm wide Si MOS-MOD, while the 400-nm wide Si MOS-MOD showed a 9.1 GHz bandwidth, which is a little smaller than that of the 30-µm-long Si MOS-MOD. It is because that the intrinsic CR time constant is similar for the lumped electrode structure of the Si-MOS MOD with varying the phase-shifter length. Figure 13(b) shows the equivalent electric circuit model of the Si MOS-MOD. It consists of the pad capacitance Cpd, the series resistance Rs, the MOS junction capacitance Cj, the capacitance due to the buried oxide (BOX) layer Cox, and the resistance due to the SOI and substrate Rsi. Using the S-parameters model parameters available by Keysight Advance Design System, the capacitance of Cj and series resistance (Rs) were estimated to be 0.11 pF and 16 Ohm in case of the 200-nm-width and 0.14 pF and 20 Ohm in the case of the 400-nm-width. The capacitance of the 200-nm-wide Si MOS-MOD was about 30% smaller than that of the 400-nm-wide Si MOS-MOD in the experiment. The capacitance values for the Si MOD-MODs with 200-nm and 400-nm-width are consistent with those of the simulation in Fig. 9(b).
Fig. 13. (a) Measured frequency dependence of normalized EO response for 200-nm and 400-nm-wide 30-µm-long Si MOS-MODs at -1 Vdc. (b) Equivalent circuit model of Si MOS-MOD.
Figures 14(a) and (b) show the measured eye diagrams for 32-Gbps and 40-Gbps of non-return-to-zero (NRZ) of 231-1 pseudorandom binary sequences (PRBS) at a wavelength of 1.31 µm under the condition of 5.0 Vpp differential RF drive at -1Vdc for the 200-nm-wide 30-µm-long Si MOS-MOD. The extinction ratio was about 1.2 dB at the quadrature point. By using TM mode, high-speed optical modulation could be obtained with a broad bandwidth and high modulation efficiency. In this case, power consumption of Cj*Vpp2/4 is estimated to be 0.69 pJ/bit, which is one of the smallest consumption-power for the Si based optical modulator.
Fig. 14. Measured eye diagram of (a) 32-Gbps and (b) 40-Gbps 231-1 PRBS at 1.31-µm wavelength for 200-nm wide and 30-µm-long Si MOS-MOD with 5 Vpp at -1 Vdc.
4. Conclusion
We developed a high-speed and high-efficiency narrow-width MOS-capacitor-type Si-MOD by applying TM optical mode excitation. We designed and fabricated a two-step of optical mode-converters of TE0-toTE1 and TE1-to-TM0. Even in the case of a 200-nm width, the Si MOS-MOD showed high-modulation efficiency (about 0.18 Vcm) in TM mode, and the electrical capacitance decreased as the MOS junction width decreased. We also demonstrated 32-Gbps and 40-Gbps high-speed operation for the 30-µm-long Si MOS-MOD in TM mode. From the simulation, the Si MOS-MOD with a symmetric electrode structure has large parasitic capacitance. The power consumption for the Si-MOS-MOD is estimated to be 0.69pJ/bit, which is one of the most efficient among the Si-based Mach-Zehnder Si-MODs. It is suggested that TM-mode of the Si MOS-MOD with an asymmetric electrode structure should larger bandwidth and smaller consumption power.
Funding
New Energy and Industrial Technology Development Organization (JPNP13004).
Acknowledgements
The authors deeply thank Masashi Takahashi from AIST for his cooperation in the device fabrication.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References
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. IEEE 802.3ba 40Gb/s and 100Gb/s Ethernet Task Force, http://www.ieee802.org/3/ba/.
3. InfiniBand Trade Association, http://members.infinibandta.org/kwspub/specs/.
4. W. M. J. Green, M. J. Rooks, L. Sekaric, and Y. A. Vlasof, “Ultra-compact, low RF power, 10 Gb/s silicon Mach-Zhender modulator,” Opt. Express 15(25), 17106–17113 (2007). [CrossRef]
5. L. Liao, A. Liu, D. Rubin, J. Basak, Y. Chetrit, H. Nguen, R. Cohen, N. Izhaky, and M. Paniccia, “40 Gbit/s silicon optical modulator for highspeed applications,” Electron. Lett. 43(22), 1196–1197 (2007). [CrossRef]
6. F. Y. Gardes, D. J. Thomson, N. G. Emerson, and G. T. Reed, “40 Gb/s silicon photonics modulator for TE and TM polarisations,” Opt. Express 19(12), 11804–11814 (2011). [CrossRef]
7. D. J. Thomson, F. Y. Gardes, J. M. Fedeli, S. Zlatanovic, Y. F. Hu, B. P. P. Kuo, E. Myslivets, N. Alic, S. Radic, G. Z. Mashanovich, and G. T. Reed, “50-Gb/s Silicon Optical Modulator,” IEEE Photonics Technol. Lett. 24(4), 234–236 (2012). [CrossRef]
8. D. M. Gill, J. E. Proesel, C. Xiong, J. S. Orcutt, J. C. Rosenberg, M. H. Khater, T. Barwics, S. Assefa, S. M. Shank, C. Reinholm, J. Ellis-Monaghan, E. Kiewra, S. Kamlapurkar, C. M. Breslin, W. M. J. Green, W. Haensch, and Y. A. Vlasof, “Demonstration of a High Extinction Ratio Monolithic CMOS Integrated Nanophotonic Transmitter and 16 Gb/s Optical Link,” IEEE J. Sel. Top. Quantum Electron. 21(4), 212–222 (2015). [CrossRef]
9. 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]
10. M. Liu, X. Yin, E. Ulin-Avila, B. Geng, T. Zentgraf, L. Ju, F. Wang, and X. Zhang, “A graphene-based broadband optical modulator,” Nature 474(7349), 64–67 (2011). [CrossRef]
11. Y. Maegami, G. Cong, M. Ohno, M. Okano, K. Itoh, N. Nishiyama, S. Arai, and K. Yamada, “High-efficiency strip-loaded waveguide based silicon Mach-Zhender modulator with vertical p-n junction phase shifter,” Opt. Express 25(25), 31407–31416 (2017). [CrossRef]
12. R. Soref and B. R. Bennett, “Electrooptical Effects in Silicon,” IEEE J. Quantum Electron. 23(1), 123–129 (1987). [CrossRef]
13. 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]
14. G. Li, X. Zheng, J. Yao, H. Thacker, I. Shubin, Y. Luo, K. Raj, J. E. Cunningham, and A. V. Krishnamoorthy, “25Gb/s 1V-driving CMOS ring modulator with integrated thermal tuning,” Opt. Express 19(21), 20435–20443 (2011). [CrossRef]
15. M. Webster, P. Gothoskar, V. Patel, D. Piede, S. Anderson, R. Tummidi, D. Adams, C. Appel, P. Metz, S. Sunder, B. Dama, and K. Shastri, “An efficient MOS-capacitro based silicon modulator and CMOS drivers for optical transmitters,” in IEEE International Conference on Group IV Photonics GFP (IEEE Computer Society, 2014), pp. 1–2.
16. J. Fujikata, S. Takahashi, M. Takahashi, M. Noguchi, T. Nakamura, and Y. Arakawa, “High-performance MOS-capacitor-type Si optical modulator and surface-illumination-type Ge photodetector for optical interconnection,” Jpn. J. Appl. Phys. 55(4S), 04EC01 (2016). [CrossRef]
17. Y. Kim, J. Fujikata, S. Takahashi, M. Takenaka, and S. Takagi, “First demonstration of SiGe-based carrier-injection Mach-Zehnder modulator with enhanced plasma dispersion effect,” Opt. Express 24(3), 1979–1985 (2016). [CrossRef]
18. J. Fujikata, M. Noguchi, Y. Kim, J.-H. Han, S. Takahashi, T. Nakamura, and M. Takenaka, “High-speed and highly efficient Si optical modulator with strained SiGe layer,” Appl. Phys. Express 11(3), 032201 (2018). [CrossRef]
19. J.-H. Han, F. Boeuf, J. Fujikata, S. Takahashi, S. Takagi, and M. Takenaka, “Efficient low-loss InGaAsP/Si hybrid MOS optical modulator,” Nat. Photonics 11(8), 486–490 (2017). [CrossRef]
20. T. Hiraki, T. Aihara, K. Hasebe, K. Takeda, T. Fujii, T. Kakitsuka, T. Tsuchizawa, H. Fukuda, and S. Matsuo, “Heterogeneously integrated III-V/Si MOS capacitor Mach-Zehnder modulator,” Nat. Photonics 11(8), 482–485 (2017). [CrossRef]
21. T. Thiessen, P. Grosse, J. D. Fonseca, P. Billondeau, B. Szelag, C. Jany, J. K. S. Poon, and S. Menezo, “30 GHz heterogeneously integrated capacitive InP-on-Si Mach-Zehnder modulators,” Opt. Express 27(1), 102–109 (2019). [CrossRef]
22. W. Zhang, M. Ebert, B. Chen, J. D. Reynolds, X. Yan, H. Du, M. Banakar, D. T. Tran, K. Debnath, C. G. Littlejohns, S. Saito, and D. J. Thomson, “Integration of low loss vertical slot waveguides on SOI photonic platforms for high efficiency carrier accumulation modulators,” Opt. Express 28(16), 23143–23153 (2020). [CrossRef]
23. J. Fujikata, S. Takahashi, M. Noguchi, and T. Nakamura, “High-Efficiency of Narrow-Width MOS Capacitor Type Si Optical Modulator with TM Mode Excitation,” in IEEE International Conference on Group IV Photonics GFP (IEEE Computer Society, 2019), FC5.
24. K. Preston and M. Lipson, “Slot waveguides with polycrystalline slicon for electrical injection,” Opt. Express 17(3), 1527–1534 (2009). [CrossRef]
25. J. Fujikata, M. Takahashi, S. Takahashi, T. Horikawa, and T. Nakamura, “High-speed and high-efficiency Si optical modulator with MOS junction, using slid-phase crystallization of polycrystalline silicon,” Jpn. J. Appl. Phys. 55(4), 042202 (2016). [CrossRef]
26. FemSIM FEM Generalized Mode Solver Software - RSoft Photonic Device Tools | Synopsys Photonic Solutions.
27. T. Fujisawa, J. Takano, Y. Sawada, T. Sakamoto, T. Matsui, K. Nakajima, and K. Saitoh, “A novel Si four-wavelength multiplexer for 100/400 GbE using higher-order mode composed of (a)symmetric directional couplers and TE1-TM0 mode converter,” Opt. Express 27(25), 36286–36296 (2019). [CrossRef]
28. T. Mogami, T. Horikawa, K. Kinoshita, Y. Hagihara, J. Ushida, M. Tokushima, J. Fujikata, S. Takahashi, T. Shimizu, A. Ukita, K. Takemura, M. Kurihara, K. Yashiki, D. Okamoto, Y. Suzuki, Y. Sobu, S.-H. Jeong, Y. Tanaka, T. Nakamura, and K. Kurata, “1.2 Tbps/cm2Enabling Silicon Photonics IC Technology Based on 40-nm Generation Platform,” J. Lightwave Technol. 36(20), 4701–4712 (2018). [CrossRef]
