Open Access
Add to BibSonomy
Abstract
We experimentally demonstrate an integrated visible light modulator at 532 nm on the thin-film lithium niobate platform. The waveguides on such platform feature a propagation loss of 2.2 dB/mm while a grating for fiber interface has a coupling loss of 5 dB. Our fabricated modulator demonstrates a low voltage-length product of 1.1 V·cm and a large electro-optic bandwidth with a roll-off of -1.59 dB at 25 GHz for a length of 3.3 mm. This device offers a compact and large bandwidth solution to the challenge of integrated visible wavelength modulation in lithium niobate and paves the way for future small-form-factor integrated systems at visible wavelengths.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Photonic integration offers a route to reduce the size and power consumption of the device as well as to improve chip-scale manufacturing. Over the last several decades, photonic integrated circuits have seen a rapid development that has opened up a broad variety of applications. Visible-light integrated photonics has not progressed as much as integrated platforms and devices at telecom wavelengths. Integrated photonics systems at visible wavelengths have numerous applications including optical phased array (OPA) used in the light detection and ranging (LiDAR) of high-speed operations [1], optogenetics [2–3], quantum information processing [4], quantum state preparation [5] and underwater wireless optical communication systems (UWOC) [6–7]. Needless to say, visible-light modulators are important for these applications.
Generally, visible-light integrated devices can be implemented on several material platforms such as silicon nitride (SiN) [8–10], silicon dioxide [11–12], aluminum nitride (AlN) [8,13], nematic liquid crystal (LC) [14], lead zirconate titanate (PZT) [15–16], or thin-film lithium niobate (LN) [17–19]. However, silica and silicon nitride have low thermo-optical coefficients and lack electro-optical (EO) properties, which makes integrated modulation at visible wavelengths a challenge. In addition, AlN and LC are not strictly CMOS compatible. Besides of these, PZT is a challenging to use because of significant hysteresis and an extremely large dielectric constant, limiting modulation speed via increases in RC time constants [20]. LN is one of the most promising materials for electro-optic devices due to its many desirable properties, including a wide transmission spectrum ranging from the visible to the mid-infrared, strong electro-optic effect, stable physical and chemical characteristics, etc., which makes it an excellent candidate for EO modulators at visible light wavelength. Currently, high-performance EO modulators on lithium niobate-on-insulator (LNOI) have been demonstrated for C-band with a low loss, low drive voltage, and large bandwidth [20–23]. It can be predicted that the LN platform can also achieve relatively good performance in the visible wavelengths.
In this paper, we propose and demonstrate a thin-film LN visible light modulator at 532 nm. The fabricated device has a modulation section length of 3.3 mm, an insertion loss of 15.9 dB, a half-wave voltage (Vπ) of about 3.3V, and a 3-dB EO bandwidth of more than 25 GHz. It is a compact, low Vπ, and large bandwidth solution to the challenge of integrated visible wavelengths modulators and paves the way for future high-speed, low-power, and small-form-factor integrated systems at visible wavelengths.
2. Design and fabrication
Figure 1 shows the schematic diagram of the present modulator. It is basically a Mach Zehnder interferometer (MZI) modulator with a push-pull configuration, and was fabricated on a commercial x-cut lithium niobate on silicon (LNOI) wafer (NanoLN, Jinan, China) with a 200-nm thick top LN film. The LN waveguide has a ridge height of 100 nm, i.e., half of the total thickness of the LN film. Between the thin-film LN and the silicon substrate is a 2-µm thick thermal oxide, and the thin-film LN is covered by a 300-nm thick silicon dioxide (SiO2). A grating coupler (GC) with the same etching depth was used as the fiber-to-chip interface. A multimode interferometer (MMI) was used as the 3 dB splitter and combiner for the MZI structure. The MMI can provide a relaxed fabrication tolerance as well as a large wavelength bandwidth. Traveling wave (TW) electrodes based on a classic coplanar waveguide (CPW), in which electric and optical waves propagate collinearly, were adopted in the proposed MZI. The electrodes support a quasi-transverse electromagnetic (quasi-TEM) mode for the electrical signal with low dispersion and therefore can offer a high modulation bandwidth. The electron beam lithography and dry etching technologies were used to fabricate the LN structures, and the electron beam evaporation and lift-off processes were used for the metal electrode. Pictures of one finished sample are shown in Fig. 1(b). The light is coupled into the waveguide by a GC and transmitted through a taper tapering from 5 µm to 0.3 µm to filter out higher-order modes, and then into a 90° Euler bend with a width of 0.3 µm, a maximum radius of 150 µm, and a minimum radius of 50 µm to retain TE fundamental mode propagation. It is common knowledge that fundamental mode propagation can be better preserved using narrow waveguides, especially for visible light operation at 532 nm. However, due to limitations of the manufacturing process, the narrower the waveguide, the larger the loss. Therefore, a waveguide width of 0.3 µm was used to balance the fundamental mode and loss. Next, the waveguide width is progressively modified from 0.3 µm to 1.5 µm and coupled into the 1 × 2 MMI. After beam splitting, the width of the waveguide at the output is progressively adjusted from 1.5 µm to 0.4 µm for additional filtering and then progressively changed back to 1.5 µm in the modulation region.
Fig. 1. (a) 3D structure of the proposed device; (b) Optical microscope of the overall sketch and partially enlarged view of the waveguide.
Figure 2(a) shows the cross-sectional view for the GC structure. The fiber was placed with a tilted angle of θ=-10° with respect to the normal direction of the chip surface. The sidewall angle α of all etched LN patterns was set to 60°, a typical value from the current fabrication technology in our laboratory. We optimized the coupling efficiency of the GC with a Gaussian beam with a waist diameter of 3.6 µm (the mode field diameter of the SM450 fiber for 532 nm light) by scanning the duty cycle ($\rho $) and period (p) as shown in Fig. 2(b). Accordingly, p and $\rho $ were chosen to be 288 nm and 0.6, respectively, which gives a maximal coupling efficiency of 33.6% was obtained at 532 nm. Its wavelength response is shown in Fig. 2(c), with 1-dB bandwidth of 12 nm, which is relatively narrow compared with GCs at near-infrared wavelength [24] due to the much shorter wavelength discussed here. Fortunately, even if a uniform grating is used as the diffractive element, a cavity-assisted grating structure and a top metal mirror allow for increasing the bandwidth of GCs [25]. To avoid mode converting, we remove the SiO2 on the input and output port of GCs by wet etching. In addition, a scan-electron microscope (SEM) picture of a fabricated GC is displayed in Fig. 2(d), showing that the structural parameters of the GC are well controlled using the current technology.
Fig. 2. (a) Cross-section of the grating structure; (b) simulated coupling efficiency with respect to the period and duty cycle; (c) simulated spectra of the designed GC using the optimal parameters (p = 288 nm, ρ=0.6); (d) SEM picture of the fabricated GC.
Fig. 3. (a) Top view schematic of the 3-dB MMI coupler; (b) calculated two-port transmission of the designed MMI splitter. Inset: simulated light propagation in the designed MMI at 532 nm.
The 3-dB coupler used in the MZI was realized by a MMI coupler, as shown in Fig. 3(a). The width WMMI and length LMMI of the MMI region were designed to be 4 µm and 36.1 µm, respectively. The width w of the input and output waveguide was 1.5 µm. The separation between the output ports YMMI was 2.08 µm to ensure low crosstalk between the two output ports. The simulated transmission spectra at the output Port 1 and Port 2 are shown in Fig. 3(b). The peak coupling efficiency is 49.8% at 532 nm.
The performance of a traveling wave MZI modulator is mainly determined by the cross-section of the modulation section as shown in Fig. 4. Au electrode with a thickness of 1.1 µm was used in the design, which ensures a low microwave loss. The other parameters of the electrodes, i.e., the gap between the electrode (g), the width of the signal electrode (ws), and the width of the ground electrode (wg) were designed as 5.9 µm, 13.8 µm, 60 µm, respectively. The height of over-cladding hupsio2 was 300 nm and the width w of the waveguide was 1.5 µm. Accordingly, the effective RF refractive index nm, loss αm, and the characteristic impedance Z0 can be obtained via simulation, as shown in Figs. 4(b)–4(d), respectively. From Fig. 4(b), we can see that the refractive index of the driving RF signal (nm) is still not matched with the group index of the optical mode (no) shown by the red dashed line in the current configuration (hSiO2 = 2 µm). This problem can be solved by using thinner buried oxide, which will make the silicon substrate closer to the electrode, thus increasing the nm, as shown in Fig. 4(b). Note that the thickness of the thermal oxide cannot be too thin, or it will bring light absorption from the silicon substrate. Another method is to increase the width of the electrode ws, as well as the gap g for keeping the impedance matching. However, a larger gap will result in a larger Vπ. Finally, the design parameters were set based on the wafer at hand, compromising between the EO bandwidth and Vπ. The characteristic impedance (Z0) of the electrode was designed to be slightly large than 50 Ω, as shown in Fig. 4(d). From Fig. 4(c), one can see that the RF loss αm is less than 9 dB/cm within the range of 120 GHz. The simulated electro-optic efficiency ($\Delta {\textrm{n}_{eff}}/V$) was also calculated by sweeping the DC voltage, as shown in Fig. 4(e). Obviously, we can have
Fig. 4. (a) Cross-section of the visible light modulator with traditional TW electrode; (b) RF effective index nm with different hsio2; (c) RF loss αm; (d) Impedance Z0; (e) Change of effective index under different DC voltages; (f) Calculated modulation response m[ω] of modulator with the capacitively loaded electrode/Traveling wave electrode of 3.3 mm; (g) Cross-section and top view of the visible light modulator with capacitively loaded electrode; (h) Microwave refractive index and Optical wave refractive index of the design about capacitively loaded electrode. Here the design of traveling wave electrode is with g = 5.5 µm, t = 1.1 µm, ws = 13.8 µm, and wg = 60 µm; the design of capacitively loaded electrode is with g = 18 µm, t = 1 µm, ws = 45 µm, wg = 100 µm, Tcw = Tw = 1 µm, Tgap = 3 µm, Tl = 45 µm, L = 50 µm, and hupsio2= 0.5 µm.
3. Measurements
Optical loss measurement. A single-longitudinal-mode 532-nm laser (Changchun Laser Optoelectronics Technology Co., Ltd.) with a spectral linewidth of < 0.01 pm and an optical power meter (Newport Model 1918-R) were used for characterization of the insertion loss of the fabricated device. The loss of the GC was first measured using a cascaded GC structure with a back-to-back configuration. One can see from Fig. 5(b) that the measured insertion loss of the GC is 5 dB, which is less than the insertion loss of the end-face coupling proposed by Ref.18 (6 - 10 dB at 635 nm). The reduced coupling loss by using GCs is the primary reason why we chose to use grating coupling instead of end-face coupling. It should also be noted that GCs are compatible with traditional E-beam lithography, with a larger fabrication tolerance compared with end-face coupling. We then measured the loss of the MMI by cascading five MMIs. The measured total loss (2 GCs + 5 MMIs) was 29 dB, corresponding to an excess loss for each MMI of about 0.8 dB. We also measured the propagation loss of the LN ridge waveguide by three different lengths. The propagation loss calculated via the slope of the measured loss curve was about 2.2 dB/mm. The main source of the waveguide loss may come from scattering due to the rough sidewalls, and it is expected to be much more significant at the visible wavelengths considered here since Rayleigh scattering is proportional to λ−4, where λ is the wavelength of light. Furthermore, the output optical power may be unstable due to the photorefractive effect. This problem was mitigated by a 1-h annealing process at 500°C in a nitrogen environment operation [28]. The total loss of the modulator is 15.9 dB, including 1.6 dB from MMI and 7.3 dB from the modulated waveguide. The metal electrode absorption loss is around 7 dB. The electrode absorption loss may be attributed to the misalignment of the electrodes, and can be largely mitigated by refining fabrication processes. Also, setting the electrodes further apart from the LN slab by adding a SiO2 layer between the optical waveguides and electrodes may help decrease the absorption loss, at the cost of an increased VπL. To balance the trade-off between loss and half-wave voltage, we implemented SiO2 layers with a practically measured thickness of 315 nm in this work.
Fig. 5. (a) Experimental set-up for measuring the loss; (b) GCs loss and MMI loss; (c) The schematic of the structures fabricated for characterizing GCs and MMI insertion loss; (d) Waveguide propagation loss. PC, polarization controller; PM, Power meter.
Electro-optic (EO) performance. To characterize the bandwidth of the MZI modulator, the electrical performance of TW electrode was first measured using a vector network analyzer (VNA), and the S21, as well as S11 responses, are shown in Fig. 6(a). The 6.4 dB bandwidth of S21 clearly surpasses 67 GHz and the overall S11 curve is nearly below -20 dB over the whole frequency range. The 6.4 dB bandwidth of the electrical-electrical (EE) response determines the 3-dB EO bandwidth of the modulator when index and impedance match perfectly. The half-wave voltage (Vπ) of the fabricated modulator was measured using sawtooth waves with a frequency of 100 kHz and 20 Vpp. In addition, the devices are driven in a single-drive push-pull configuration, so that applied voltage induces a positive phase shift in one arm and a negative phase shift in the other. Figure 6(b) indicates a measured Vπ of ∼3.3 V, which has a certain deviation from the simulation value (2.73 V) due to fabrication defects, such as lithography alignment problems and the metal electrode sidewall being not vertical. The extinction ratio of the modulator is 23 dB as shown in Fig. 6(c). The EO response of the fabricated MZI modulators is shown in Fig. 6(d), where the responses of the RF cables, RF probes, photodetector (Newport model 1414, 25 GHz), and microwave amplifier (Talent Microwave TLLA50K20G-30-30, 20 GHz) were subtracted for normalization. We can see clearly that the measured 3-dB EO bandwidth (S21) is beyond 25 GHz, limited by the bandwidth of the photodetector. The simulated EO response in Fig. 6(d) deduced from the measured electrode parameters also matches well with the measured EO one at frequencies < 25 GHz.
Fig. 6. (a) Simulated and (b) measured electrical reflection S11 and transmission S12 of the device; (c) Normalized optical transmission as a function of applied voltage of a 100-kHz sawtooth signal, with a measured Vπ of 3.3 V; (d) Transmission of the modulator as a function of the DC bias voltage; (e) Measured and simulated EO S12 of the modulator.
4. Conclusion
In summary, we have proposed a high-performance visible light modulator based on the thin-film LN platform. For a modulation length of 3.3 mm, the device can achieve a Vπ of 3.3 V and a large electro-optic bandwidth with a roll-off of -1.59 dB at 25 GHz. The fabricated modulator has a loss of 15.9 dB. A smooth sidewall can further reduce the transmission loss greatly, e.g., to 6 dB/m at 635 nm wavelength [18]. Furthermore, we can further decrease the loss through suppressing the fabrication imperfection during thin-film production and nano-structuring by chemical mechanical polishing (CMP) [29]. In addition, the EO bandwidth can be further improved theoretically by careful design for the index matching between the optical and RF waves. This device is a compact and large EO bandwidth solution to the challenge of integrated modulation in LN. This work paves the way for future small-form-factor integrated systems at visible wavelengths for important application areas, such as OPA used in LiDAR of high-speed operations [1], optogenetics [2–3], quantum information processing [4], quantum state preparation [5] and UWOC [6–7].
Funding
Science and Technology Planning Project of Guangdong Province (Grant No. 2019A050510039); Basic and Applied Basic Research Foundation of Guangdong Province (Grant No. 2021A1515012215).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. M. C. Shin, A. Mohanty, K. Watson, G. R. Bhatt, C. T. Phare, S. A. Miller, M. Zadka, B. S. Lee, X. Ji, and I. Datta, “Chip-scale blue light phased array,” Opt. Lett. 45(7), 1934–1937 (2020). [CrossRef]
2. A. Mohanty, Q. Li, M. A. Tadayon, S. P. Roberts, G. R. Bhatt, E. Shim, X. Ji, J. Cardenas, S. A. Miller, and A. Kepecs, “Reconfigurable nanophotonic silicon probes for sub-millisecond deep-brain optical stimulation,” Nat. Biomed. Eng. 4(2), 223–231 (2020). [CrossRef]
3. E. Segev, J. Reimer, L. C. Moreaux, T. M. Fowler, D. Chi, W. D. Sacher, M. Lo, K. Deisseroth, A. S. Tolias, and A. Faraon, “Patterned photostimulation via visible-wavelength photonic probes for deep brain optogenetics,” Neurophotonics 4(01), 1 (2016). [CrossRef]
4. A. I. Lvovsky, B. C. Sanders, and W. Tittel, “Optical quantum memory,” Nat. Photonics 3(12), 706–714 (2009). [CrossRef]
5. I. Aharonovich, D. Englund, and M. Toth, “Solid-state single-photon emitters,” Nat. Photonics 10(10), 631–641 (2016). [CrossRef]
6. T.-C. Wu, Y.-C. Chi, H.-Y. Wang, C.-T. Tsai, and G.-R. Lin, “Blue laser diode enables underwater communication at 12.4 Gbps,” Sci. Rep. 7(1), 1–10 (2017). [CrossRef]
7. Y. Zhou, X. Zhu, F. Hu, J. Shi, F. Wang, P. Zou, J. Liu, F. Jiang, and N. Chi, “Common-anode LED on a Si substrate for beyond 15 Gbit/s underwater visible light communication,” Photonics Res. 7(9), 1019–1029 (2019). [CrossRef]
8. M. Dong, G. Clark, A. J. Leenheer, M. Zimmermann, D. Dominguez, A. J. Menssen, D. Heim, G. Gilbert, D. Englund, and M. Eichenfield, “High-speed programmable photonic circuits in a cryogenically compatible, visible–near-infrared 200 mm CMOS architecture,” Nat. Photonics 16(1), 59–65 (2022). [CrossRef]
9. P. Muñoz, G. Micó, L. A. Bru, D. Pastor, D. Pérez, J. D. Doménech, J. Fernández, R. Baños, B. Gargallo, and R. Alemany, “Silicon nitride photonic integration platforms for visible, near-infrared and mid-infrared applications,” Sensors 17(9), 2088 (2017). [CrossRef]
10. M. A. Porcel, A. Hinojosa, H. Jans, A. Stassen, J. Goyvaerts, D. Geuzebroek, M. Geiselmann, C. Dominguez, and I. Artundo, “Silicon nitride photonic integration for visible light applications,” Opt. Laser Technol. 112, 299–306 (2019). [CrossRef]
11. Y. Gong and J. Vučković, “Photonic crystal cavities in silicon dioxide,” Appl. Phys. Lett. 96(3), 031107 (2010). [CrossRef]
12. S. H. Lee, D. Y. Oh, Q.-F. Yang, B. Shen, H. Wang, K. Y. Yang, Y.-H. Lai, X. Yi, X. Li, and K. Vahala, “Towards visible soliton microcomb generation,” Nat. Commun. 8(1), 1–8 (2017). [CrossRef]
13. M. Dong, D. Heim, A. Witte, G. Clark, A. J. Leenheer, D. Dominguez, M. Zimmermann, Y. H. Wen, G. Gilbert, and D. Englund, “Piezo-optomechanical cantilever modulators for VLSI visible photonics,” APL Photonics 7(5), 051304 (2022). [CrossRef]
14. M. Notaros, T. Dyer, M. Raval, C. Baiocco, J. Notaros, and M. R. Watts, “Integrated visible-light liquid-crystal-based phase modulators,” Opt. Express 30(8), 13790–13801 (2022). [CrossRef]
15. W. Jin, R. G. Polcawich, P. A. Morton, and J. E. Bowers, “Piezoelectrically tuned silicon nitride ring resonator,” Opt. Express 26(3), 3174–3187 (2018). [CrossRef]
16. N. Hosseini, R. Dekker, M. Hoekman, M. Dekkers, J. Bos, A. Leinse, and R. Heideman, “Stress-optic modulator in TriPleX platform using a piezoelectric lead zirconate titanate (PZT) thin film,” Opt. Express 23(11), 14018–14026 (2015). [CrossRef]
17. C. Li, P. Chen, J. Li, K. Chen, C. Guo, and L. Liu, “Modeling of thin-film lithium niobate modulator for visible light,” Opt. Eng. 61, 057101 (2022). [CrossRef]
18. B. Desiatov, A. Shams-Ansari, M. Zhang, C. Wang, and M. Lončar, “Ultra-low-loss integrated visible photonics using thin-film lithium niobate,” Optica 6(3), 380–384 (2019). [CrossRef]
19. O. T. Celik, C. J. Sarabalis, F. M. Mayor, H. S. Stokowski, J. F. Herrmann, T. P. McKenna, N. R. Lee, W. Jiang, K. K. Multani, and A. H. Safavi-Naeini, “High-bandwidth CMOS-voltage-level electro-optic modulation of 780 nm light in thin-film lithium niobate,” Opt. Express 30(13), 23177–23186 (2022). [CrossRef]
20. P. Stanfield, A. Leenheer, C. Michael, R. Sims, and M. Eichenfield, “CMOS-compatible, piezo-optomechanically tunable photonics for visible wavelengths and cryogenic temperatures,” Opt. Express 27(20), 28588–28605 (2019). [CrossRef]
21. G. Chen, K. Chen, R. Gan, Z. Ruan, Z. Wang, P. Huang, C. Lu, A. P. T. Lau, D. Dai, and C. Guo, “High performance thin-film lithium niobate modulator on a silicon substrate using periodic capacitively loaded traveling-wave electrode,” APL Photonics 7(2), 026103 (2022). [CrossRef]
22. M. He, M. Xu, Y. Ren, J. Jian, Z. Ruan, Y. Xu, S. Gao, S. Sun, X. Wen, and L. Zhou, “High-performance hybrid silicon and lithium niobate Mach–Zehnder modulators for 100 Gbit s− 1 and beyond,” Nat. Photonics 13(5), 359–364 (2019). [CrossRef]
23. C. Wang, M. Zhang, X. Chen, M. Bertrand, A. Shams-Ansari, S. Chandrasekhar, P. Winzer, and M. Lončar, “Integrated lithium niobate electro-optic modulators operating at CMOS-compatible voltages,” Nature 562(7725), 101–104 (2018). [CrossRef]
24. J. Liu, Z. Zheng, B. Chen, Z. Wang, C. Li, K. Chen, and L. Liu, “High-performance grating coupler array on silicon for a perfectly-vertically mounted multicore fiber,” J. Lightwave Technol. 40(16), 5654–5659 (2022). [CrossRef]
25. B. Chen, Z. Ruan, X. Fan, Z. Wang, J. Liu, C. Li, K. Chen, and L. Liu, “Low-loss fiber grating coupler on thin film lithium niobate platform,” APL Photonics 7(7), 076103 (2022). [CrossRef]
26. M. Bahadori, L. L. Goddard, and S. Gong, “Fundamental electro-optic limitations of thin-film lithium niobate microring modulators,” Opt. Express 28(9), 13731–13749 (2020). [CrossRef]
27. P. Kharel, C. Reimer, K. Luke, L. He, and M. Zhang, “Breaking voltage–bandwidth limits in integrated lithium niobate modulators using micro-structured electrodes,” Optica 8(3), 357–363 (2021). [CrossRef]
28. Y. Xu, M. Shen, J. Lu, J. B. Surya, A. Al Sayem, and H. X. Tang, “Mitigating photorefractive effect in thin-film lithium niobate microring resonators,” Opt. Express 29(4), 5497–5504 (2021). [CrossRef]
29. R. Gao, N. Yao, J. Guan, L. Deng, J. Lin, M. Wang, L. Qiao, W. Fang, and Y. Cheng, “Lithium niobate microring with ultra-high Q factor above 10 8,” Chin. Opt. Lett. 20(1), 011902 (2022). [CrossRef]
