Abstract
We numerically analyzed the modulation characteristics of an InP organic hybrid (IOH) optical modulator consisting of an InP slot waveguide and an electro-optic (EO) polymer. Since InP has a higher electron mobility and a lower electron-induced free-carrier absorption than Si, the series resistance of an InP slot waveguide can be significantly reduced with relatively smaller optical loss than an Si slot waveguide. As a result, the trade-off between optical loss and modulation bandwidth can be remarkably improved compared with a Si organic hybrid (SOH) optical modulator. When the modulation bandwidth was designed to be 100 GHz, the optical loss of the IOH modulator was 13-fold smaller than that of the SOH one. The simulation of the eye diagram revealed that the improved optical modulation amplitude enabled the clear eye opening with a 100 Gbps non return-to-zero signal using the IOH modulator. The IOH integration is promising for a high-speed modulator with low energy consumption beyond 100 Gbps.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
An efficient, high-speed, and low-loss optical modulator is essential for next-generation ultrahigh-speed optical interconnects in datacenter networks. Among advanced optical modulators [1–5], a Si organic hybrid (SOH) optical modulator consisting of a Si slot waveguide and an electro-optic (EO) polymer is promising because of its high modulation efficiency and modulation bandwidth [6]. Owing to the Pockels effect of the EO polymer filled in the gap of the Si slot waveguide, the phase modulation efficiency VπL of 0.5 Vmm has been reported [7]. Although the EO polymer itself has no optical absorption accompanying the phase shift, the free-carrier absorption in Si imposes the difficulty of reducing the parasitic resistance of the Si slot waveguide without the expense of additional optical loss, resulting in the fundamental trade-off between the optical loss and the modulation bandwidth. To overcome this fundamental trade-off, the introduction of an n-type III-V compound semiconductor into an organic hybrid modulator is effective because of its high electron mobility [8,9]. By using a III-V-on-insulator (III-V-OI) platform [10] instead of a Si-on-insulator (SOI) platform, researchers have investigated high-index-contrast ultrasmall InP-based waveguide devices [11–15]. By using 193-nm deep UV lithography, the propagation loss of an InP membrane waveguide is reported to be 1.3 dB/cm which is comparable to Si photonics [16]. Hence, this III-V-OI platform enables an InP organic hybrid (IOH) optical modulator. We can couple the light from an IOH modulator into a single-mode optical fiber through a surface grating coupler [17] or into a Si waveguide through an adiabatic coupler [18]. Recently, an IOH modulator with a VπL of 3Vmm have been reported [19]. Owing to the low parasitic resistance in the InP slot waveguide, an IOH optical modulator is expected to exhibit a larger modulation bandwidth than a SOH one [8,9]. However, the impact of an InP slot waveguide on the optical loss of an IOH optical modulator has not yet been explored sufficiently. The optical loss is essential, particularly for an ultrahigh-speed optical modulator. We have proposed the superiority of IOH to SOH [20]. In this paper, we numerically discussed the detail of the trade-off between the optical loss and the modulation bandwidth of an IOH optical modulator. We revealed that the higher electron mobility and smaller free-carrier absorption in n-InP than in n-Si enable the significant improvement of the optical loss and modulation bandwidth simultaneously, which is suitable for ultrahigh-speed optical modulator operating beyond 100 GHz.
2. Design of InP slot waveguide
A cross-sectional schematic of an IOH optical modulator is shown in Fig. 1(a). An InP slot waveguide is assumed to be formed on a SiO2 buried oxide. The total thickness of the InP layer was 250 nm to satisfy a single-mode condition in the vertical direction at a wavelength of 1.55 µ;m. Figure 1(b) shows the optical electric field distribution of a slot mode in the InP slot waveguide where the optical electric field concentrates in the gap of the slot. The gap of the slot waveguide is assumed to be filled with an EO polymer. By applying a voltage to the electrodes placed 1 µ;m away from the waveguide, the electrical field is mainly distributed in the gap as shown in Fig. 1(c), resulting in the induction of a phase shift of the slot mode through the Pockels effect of the EO polymer.
First, to obtain a maximum modulation efficiency, we optimized the rail width and the gap. Figure 2(a) shows the dependence of the optical confinement factor in the gap on the rail width and gap. It was found that the optical confinement factor became maximum when the rail width and gap were approximately 240 and 110 nm, respectively. When the rail width fluctuates by ±30 nm, the optical confinement degrades by 3%. Therefore, the impact of the fabrication errors is within the acceptable range if we can use advanced process technologies. We also analyzed the dependence of the optical confinement factor on the slab height as shown in Fig. 2(b). When the slab thickness was greater than 70 nm, the optical confinement degraded significantly. A thin slab is desirable in terms of the optical confinement, whereas a thick slab is desirable in terms of the parasitic resistance. Hence, we assumed a 60-nm-thick slab in the following discussion to achieve a good balance between the modulation efficiency and the modulation speed.
3. Simulation model
We used the finite difference Eigenmode solver and the charge transport solver provided by Lumerical to calculate the optical mode, electric field, free carrier distribution, electrical resistance, and capacitance, and calculated the figure of merits of the modulator. The operating wavelength was 1.55 µ;m and the operating temperature was 300 K. To calculate the resistance of the n-InP slab as a function of doping concentration, the electron mobility of InP was calculated using the Caugh–Thomas-like model [21] expressed by
4. Result
Figure 3(a) shows the carrier mobilities of n-InP, n-Si, and p-Si as a function of doping concentration. Owing to the Coulomb scattering, the mobilities degrade as the doping concentration increases, whereas the degradation is more moderate in n-InP than in Si. As a result, the electron mobility of InP is approximately three to ten times higher than that of Si depending on the doping concentration. The free-carrier absorptions in n-InP, n-Si, and p-Si are shown in Fig. 3(b). The free-carrier absorption is almost proportional to the doping concentration to the power of 1.2. Because of the degradation in electron mobility in n-InP, the free-carrier absorption in n-InP increases with increasing the doping concentration. However, we can achieve 1.2 to 3.5 times lower absorption in n-InP than in n-Si when the doping concentration was in the range from 1017 to 1019 cm-3. Therefore, an n-InP slot waveguide enables the considerable improvement of a series resistance without increasing optical loss. By taking into account the free-carrier absorption and hole mobility in p-Si, a p-Si slot waveguide is inferior to a n-Si slot waveguide in terms of the trade-off between the modulation bandwidth and optical loss. Therefore, we will focus on n-Si for SOH modulator in the following discussion.
Figures 4(a) and 4(b) show modulation efficiency (VπL) and optical loss as functions of doping concentration when the applied voltage was 2 V. Regardless of the use of n-InP or n-Si as a waveguide material, VπL was approximately 0.1 Vcm and almost independent of doping concentration since the phase modulation was dominated by the Pockels effect in the EO polymer, and the free-carrier effect in n-InP or n-Si had almost no contribution to the phase modulation. Owing to the structural optimization, we can expect higher modulation efficiency than that of the IOH modulator reported in [19]. The slightly higher modulation efficiency in n-Si than n-InP was attributable to the larger refractive index contrast in Si, which enabled better optical confinement in gap region. In contrast, the optical loss, which was dominated by the free-carrier absorption in n-InP or n-Si, increased with doping concentration. We can achieve 1.5 to 4.1 times lower absorption in n-InP than in n-Si when the doping concentration was in the range from 1017 to 1019 cm-3. Note that the optical loss of the n-InP slot waveguide was smaller than expected from the bulk absorption owing to the relatively lower optical confinement factor in n-InP than in n-Si. We can improve a VπL by increasing r33 of EO polymer as reported in [27], while the intrinsic advantage of the IOH modulator against the SOH modulator does not change.
To consider the dynamic modulation characteristics, we evaluated the series resistance and capacitance. Figure 5(a) shows the calculated series resistances of the n-InP and n-Si slabs as a function of doping concentration when the phase shifter length was 1 mm. Since the electron mobility of InP is greater than that of Si as shown in Fig. 3(a), the series resistance of n-InP is significantly lower than that of n-Si. The parasitic capacitance in Fig. 5(b) consists of the capacitance of the gap between two rails and the depletion capacitance. Although the depletion capacitance changed slightly when the doping concentration increased, the total capacitance was almost constant. The parasitic capacitance of the IOH modulator was 125 fF, which is slightly smaller than that of the SOH one because of the differences in the gap width and height of the slot between InP and Si.
The 3-dB modulation bandwidth was estimated using a resistor-capacitor (RC) constant as shown in Fig. 6(a). As the doping concentration increases, the RC constant decreases, but the optical loss increases as shown in Figs. 4(b) and 5(a). As a result, the modulation bandwidth and optical loss are in a trade-off relationship as shown in Fig. 6(a). Since n-InP has lower optical absorption and resistance than n-Si as shown in Fig. 4(b) and 5(a), the IOH modulation can improve this trade-off considerably. As a result, to achieve 100 GHz bandwidth, the SOH modulator requires a doping concentration of 3 × 1019 cm-3, whereas InP requires only 8 × 1017 cm-3. In this case, the optical loss was reduced from 66 dB/mm (Si) to 0.6 dB/mm (InP). The RC constant is ideally independent of the phase shifter length. However, the insertion loss depends on the phase shifter length, which can be changed by adjusting the drive voltage. Figure 6(b) shows the relationship between the drive voltage and the optical loss at Lπ, where Lπ is the phase shifter length for the π phase shift. Since VπL is almost constant with respect to the drive voltage, Lπ decreases with increasing drive voltage. As a result, the optical loss is inversely proportional to the drive voltage. If the drive voltage is 4 V, the insertion loss of the SOH modulator can be reduced to approximately 11.5 dB, which is 13 times greater than that of the IOH modulator. In the case of the IOH modulator, the insertion loss is around 1 dB even with 1 V. Therefore, the IOH modulator enables the marked reduction in energy consumption.
The eye diagram was also simulated with a Mach–Zehnder interferometer (MZI) configuration in Fig. 7(a). The wavelength of the input laser was 1550 nm. We assumed the optical loss was caused only by the optical phase modulators in the MZI. The phase shifter length was assumed to be half Lπ for a push–pull operation. A modulating voltage was applied to the center electrode between two arms, and the modulation electric field was oriented parallel to the poling direction for one arm and anti-parallel for the other arm. The voltage swing was assumed to be 2 V and both rising and falling time was assumed to be 10 ps. A photodetector (PD) with a transimpedance amplifier (TIA) was assumed as a receiver. The dark current and thermal noise of the PD were assumed to be 10 nA and 18 pA/(Hz0.5), respectively. The thermal noise of PD was determined by assuming the resistance of 50 Ω and room temperature. The transimpedance and equivalent input noise of the TIA were assumed to be 10 kΩ and 20 pA/(Hz0.5), respectively [27]. A 1st-order low pass filter was applied to an electrical signal for the modulator to take into account the effect of the RC constant discussed in Fig. 6(a). The doping concentrations of the n-InP and n-Si layers were 8 × 1017 and 3 × 1019 cm-3, respectively, as shown in Fig. 6. The Lπ values of the IOH and SOH modulators were 512 and 348 µ;m, respectively. Figure 7(b) and (c) shows the eye diagrams of two modulators with a 100 Gbps non-return-to-zero (NRZ) driving signal when the input optical power was 0 dBm. Both modulators have a bandwidth over 100 GHz. As shown in Fig. 6(b), the optical loss of the IOH modulator was significantly lower than that of the SOH modulator. As a result, the optical modulation amplitude (OMA) of the SOH modulator was −15 dBm, whereas that of the IOH modulator was −0.2 dBm. The energy consumption of the IOH modulator on this configuration was approximately 125 fJ/bit. We obtained a clear eye opening in the IOH modulator owing to the small insertion loss. In contrast, we observed a severely degraded eye in the SOH modulator because of the large optical loss induced by the free-carrier absorption in the n-Si waveguide.
Figure 8 shows the benchmark of αVπL, a product of the insertion loss α and VπL, and the modulation bandwidth for organic hybrid optical modulators with a slot-waveguide configuration. The red and blue solid lines are simulation results of the IOH and SOH modulators, respectively. The blue squares show the experimental results of the SOH modulators reported in [6,25 and 28–30]. For comparison, the experimental results of plasmonic-organic hybrid (POH) optical modulators [31–35] are also plotted as green circles. There is a trade-off relationship between αVπL and modulation bandwidth. In the case of the SOH modulator, its αVπL is greater than 10 dB·V when the modulation bandwidth is around 100 GHz. Although the POH modulator is promising in terms of its small footprint and high modulation bandwidth over 100 GHz, its trade-off relationship is similar to that of the SOH modulator owing to the large optical loss in a plasmonic waveguide. In contrast, the IOH modulator exhibits an αVπL of approximately 1 dB·V even at 100 GHz modulation bandwidth owing to the small series resistance and low optical loss in the n-InP slot waveguide.
5. Conclusion
In this study, we have clarified the impact of the IOH modulator on the trade-off between optical loss and modulation bandwidth. Owing to the high electron mobility and small free-carrier absorption in n-InP, the IOH modulator exhibited a 13-fold smaller optical loss than the SOH modulator when the modulation bandwidth was designed to be 100 GHz. As a result, the trade-off relationship between αVπL and modulation bandwidth was markedly improved by the IOH modulator, enabling αVπL of approximately 1 dB·V with 100 GHz. We also compared the simulation results with the reported modulators and clarified that the IOH modulator is superior in terms of αVπL owing to the high electron mobility of n-InP. Therefore, the IOH optical modulator is promising for an ultrahigh-speed, low-power optical interconnect beyond 100 Gbps.
Funding
The New Energy and Industrial Technology Development Organization.
Disclosures
The authors declare no conflicts of interest.
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. Y. Ogiso, “Over 67 GHz Bandwidth and 1.5 V Vπ InP-Based Optical IQ Modulator With n-i-p-n Heterostructure,” J. Lightwave Technol. 35(8), 1450–1455 (2017). [CrossRef]
3. C. Wang, “Integrated lithium niobate electro-optic modulators operating at CMOS-compatible voltages,” Nature 562(7725), 101–104 (2018). [CrossRef]
4. 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]
5. I. Datta, “Low-loss composite photonic platform based on 2D semiconductor monolayers,” Nat. Photonics 14(4), 256–262 (2020). [CrossRef]
6. L. Alloatti, R. Palmer, S. Diebold, K. P. Pahl, B. Chen, R. Dinu, M. Fournier, J.-M. Dedeli, T. Zwick, W. Dreude, C. Koos, and J. Leuthold, “100 GHz silicon-organic hybrid modulator,” Light: Sci. Appl. 3(5), e173 (2014). [CrossRef]
7. S. Koeber, R. Plamer, M. Lauermann, W. Heni, D. L. Elder, D. Korn, M. Woessner, S. Koenig, P. C. Schinindler, H. Yu, W. Bogaerts, L. R. Dalton, W. Freude, J. Leuthold, and C. Koos, “Femtojule electro-optic modulation using a silicon organic hybrid device,” Light: Sci. Appl. 4(2), e255 (2015). [CrossRef]
8. A.J. Millan Mejia, J.J.G.M. Tol van der, and M.K. Smit, “Design and simulation of a high bandwidth optical modulator for IMOS technology based on slot-waveguide with electro-optical polymer,” in the 18th Annual Symposium of the IEEE Photonics Benelux Chapter (2013), 195–198.
9. J. V. Engelen, V. Pogoretskiy, M. Smit, J. Tol, and Y. Jiao, “Towards a Fully Integrated Indium-Phosphide Membrane on Silicon Photonics Platform,” in SPIE/COS PHOTONICS ASIA (2018), 1082308.
10. M. Takenaka and Y. Nakano, “InP photonic wire waveguide using InAlAs oxide cladding layer,” Opt. Express 15(13), 8422 (2007). [CrossRef]
11. M. Takenaka, M. Yokoyama, M. Sugiyama, Y. Nakano, and S. Takagi, “InGaAsP Photonic Wire Based Ultrasmall Arrayed Waveguide Grating Multiplexer on Si Wafer,” Appl. Phys. Express 2(12), 122201 (2009). [CrossRef]
12. Y. Ikku, M. Yokoyama, O. Ichikawa, M. Hata, M. Takenaka, and S. Takagi, “Low-driving-current InGaAsP photonic-wire optical switches using III-V CMOS photonics platform,” Opt. Express 20(26), B357–B364 (2012). [CrossRef]
13. M. Takenaka, M. Yokoyama, M. Sugiyama, Y. Nakano, and S. Takagi, “InGaAsP Grating Couplers Fabricated Using Complementary-Metal–Oxide–Semiconductor-Compatible III–V-on-Insulator on Si,” Appl. Phys. Express 6(4), 042501 (2013). [CrossRef]
14. Y. Cheng, Y. Ikku, M. Takenaka, and S. Takagi, “Low-dark-current waveguide InGaAs metal–semiconductor–metal photodetector monolithically integrated with InP grating coupler on III–V CMOS photonics platform,” Jpn. J. Appl. Phys. 55(4S), 04EH01 (2016). [CrossRef]
15. J.-K. Park, S. Takagi, and M. Takenaka, “InGaAsP Mach–Zehnder interferometer optical modulator monolithically integrated with InGaAs driver MOSFET on a III-V CMOS photonics platform,” Opt. Express 26(4), 4842 (2018). [CrossRef]
16. J. van Engelen, S. Reniers, J. Bolk, K. Williams, J. van der Tol, and Y. Jiao, “Low Loss InP Membrane Photonic Integrated Circuits Enabled by 193-nm Deep UV Lithography,” in 2019 Compound Semiconductor Week (CSW), 1–2 (2019).
17. M. Takenaka, M. Yokoyama, M. Sugiyama, Y. Nakano, and S. Takagi, “InGaAsP Grating Couplers Fabricated Using Complementary-Metal–Oxide–Semiconductor-Compatible III–V-on-Insulator on Si,” Appl. Phys. Express 6(4), 042501 (2013). [CrossRef]
18. T. Hiraki, T. Aihara, T. Fujii, K. Takeda, T. I. Kakitsuka, T. Tsuchizawa, and S. Matsuo, “Membrane InGaAsP Mach–Zehnder Modulator Integrated With Optical Amplifier on Si Platform,” J. Lightwave Technol. 38(11), 3030–3036 (2020). [CrossRef]
19. A.A. Kashi, J.J.G.M. van der Tol, Y. Jiao, and K. Williams, “Electro-Optic Slot Waveguide Phase Modulator in the IMOS platform,” in proc. of the 22th European Conference on Integrated Optics (ECIO) (2020).
20. N. Sekine, S. Takagi, and M. Takenaka, “Investigation of optical loss and bandwidth of InP-organic hybrid optical modulator,” in 2019 Compound Semiconductor Week (CSW), 1–2 (2019).
21. M. Sotoodeh, A. H. Khalid, and A. A. Rezazadeh, “Empirical low-field mobility model for III–V compounds applicable in device simulation codes,” J. Appl. Phys. 87(6), 2890–2900 (2000). [CrossRef]
22. G. Masetti, M. Severi, and S. Solmi, “Modeling of carrier mobility against carrier concentration in arsenic-, phosphorus-, and boron-doped silicon,” IEEE Trans. Electron Devices 30(7), 764–769 (1983). [CrossRef]
23. S. Koeber, R. Plamer, M. Lauermann, W. Heni, D. L. Elder, D. Korn, M. Woessner, S. Koenig, P. C. Schinindler, H. Yu, W. Bogaerts, L. R. Dalton, W. Freude, J. Leuthold, and C. Koos, “Femtojule electro-optic modulation using a silicon organic hybrid device,” Light: Sci. Appl. 4(2), e255 (2015). [CrossRef]
24. R. Soref and B. Bennett, “Electrooptical effects in silicon,” IEEE J. Quantum Electron. 23(1), 123–129 (1987). [CrossRef]
25. B.R. Bennett, R.A. Soref, and J.A. Del Alamo, “Carrier-induced change in refractive index of InP, GaAs and InGaAsP,” IEEE J. Quantum Electron. 26(1), 113–122 (1990). [CrossRef]
26. C. Kieninger, C. Füllner, H. Zwickel, Y. Kutuvantavida, J. N. Kemal, C. Eschenbaum, D. L. Elder, L. R. Dalton, W. Freude, S. Randel, and C. Koos, “Silicon-organic hybrid (SOH) Mach-Zehnder modulators for 100 GBd PAM4 signaling with sub-1 dB phase-shifter loss,” Opt. Express 28(17), 24693–24707 (2020). [CrossRef]
27. V. Kushwah, A. Quazi, and N. Muchhal, “Design of CMOS based Transimpedance Amplifier for Bandwidth Enhancement with Large Gain,” Intl. J. Comput. Appl. 138(12), 32–37 (2016). [CrossRef]
28. R. Palmer, S. Koeber, D. L. Elder, M. Woessner, W. Heni, D. Korn, M. Lauermann, W. Bogaerts, L. Dalton, W. Freude, J. Leuthold, and C. Koos, “High-Speed, Low Drive-Voltage Silicon-Organic Hybrid Modulator Based on a Binary-Chromophore Electro-Optic Material,” J. Lightwave Technol. 32(16), 2726–2734 (2014). [CrossRef]
29. M. Lauermann, R. Palmer, S. Koeber, P. C. Schindler, D. Korn, T. Wahlbrink, J. Bolten, M. Waldow, D. L. Elder, L. R. Dalton, J. Leuthold, W. Freude, and C. Koos, “Low-power silicon-organic hybrid (SOH) modulators for advanced modulation formats,” Opt. Express 22(24), 29927–29936 (2014). [CrossRef]
30. S. Wolf, H. Zwickel, W. Hartmann, M. Lauermann, Y. Kutuvantavida, C. Kieninger, L. Altenhain, R. Schmid, J. Luo, A. K.-Y. Jen, S. Randel, W. Freude, and C. Koos, “Silicon-Organic Hybrid (SOH) Mach-Zehnder Modulators for 100 Gbit/s on-off Keying,” Sci. Rep. 9(2), 021051 (2019). [CrossRef]
31. C. Haffner, W. Heni, Y. Fedoryshyn, A. Josten, B. Baeuerle, C. Hoessbacher, Y. Salaimin, U. Koch, N. Đorđević, P. Mousel, R. Bonjour, A. Emboras, D. Hillerkuss, P. Leuchtmann, D. L. Elder, L. R. Dalton, C. Hafner, and J. Leuthold, “Plasmonic Organic Hybrid Modulators—Scaling Highest Speed Photonics to the Microscale,” Proc. IEEE 104(12), 2362–2379 (2016). [CrossRef]
32. A. Melikyan, K. Koehnle, M. Lauermann, R. Palmer, S. Koeber, S. Muehlbrandt, P. C. Schindler, D. L. Elder, S. Wolf, W. Heni, C. Haffner, Y. Fedoryshyn, D. Hillerkuss, M. Sommer, L. R. Dalton, D. Van Thourhout, W. Freude, M. Kohl, J. Leuthold, and C. Koos, “Plasmonic-organic hybrid (POH) modulators for OOK and BPSK signaling at 40 Gbit/s,” Opt. Express 23(8), 9938–9946 (2015). [CrossRef]
33. C. Haffner, W. Heni, Y. Fedoryshyn, J. Niegemann, A. Melikyan, D. L. Elder, B. Baeuerle, Y. Salamin, A. Josten, U. Koch, C. Hoessbacher, F. Ducry, L. Juchli, A. Emboras, D. Hillerkuss, M. Kohl, L. R. Dalton, C. Hafner, and J. Leuthold, “All-plasmonic Mach–Zehnder modulator enabling optical high-speed communication at the microscale,” Nat. Photonics 9(8), 525–528 (2015). [CrossRef]
34. M. Burla, C. Hoessbacher, W. Heni, C. Haffner, Y. Fedoryshyn, D. Werner, T. Watanabe, H. Massler, D. L. Elder, L. R. Dalton, and J. Leuthold, “500 GHz plasmonic Mach-Zehnder modulator enabling sub-THz microwave photonics,” APL Photonics 4(5), 056106 (2019). [CrossRef]
35. S. Ummethala, T. Harter, K. Koehnle, Z. Li, S. Muehlbrandt, Y. Kutuvantavida, J. Kemal, P. Marin-Palomo, J. Schaefer, A. Tessmann, S. K. Garlapati, A. Bacher, L. Hahn, M. Walther, T. Zwick, S. Randel, W. Freude, and C. Koos, “THz-to-optical conversion in wireless communications using an ultra-broadband plasmonic modulator,” Nat. Photonics 13(8), 519–524 (2019). [CrossRef]