Open Access
Add to BibSonomy
Abstract
Fabrication of silicon nitride (SiN) based high-Q microring resonators prepared by the hot-wire chemical vapor deposition (HWCVD) method is presented. By the virtue of low-stress HWCVD films, no special precautions against crack propagation were required for high confinement waveguide device fabrication. By using an additional annealing process, the intrinsic Q factor in excess of 5 × 105 was obtained in the telecommunication C band, and which allowed us to observe frequency comb generation. We also investigated into the anneal temperature dependence of the residual hydrogen concentration in the film as well as the optical properties of the microring resonators.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Kerr frequency combs / microcombs have attracted much attention due to their unique features such as high repetition rates and small footprints [1,2]. In particular, photonic chip-based Kerr frequency combs are expected to be useful for practical applications ranging from telecommunication [3–6], microwave generation [7], spectroscopy [8], and metrology [9]. Specifically, high repetition rates (up to ∼ 1 THz) that can uniquely be achieved by using the Kerr frequency combs implies interesting opportunities to link between photonic and terahertz carrier signals in telecommunication, which might offer seamless and enhanced signal processing capabilities expected in beyond-5 G / 6 G mobile networks.
Despite the maturity of silicon photonics, silicon waveguides are unfortunately not suitable for Kerr frequency combs when pumping at the near-infrared telecommunication wavelengths, owing to its narrow electronic bandgap [1]. Since Kerr frequency combs use four-wave-mixing (FWM), which is the third-order nonlinear optical effect, two photon absorption competes with FWM and the resultant free-carriers generated in silicon increases the propagation losses. For this reason, high-index wide-bandgap materials such as SiN [10], Hydex glass [11], poly-crystalline aluminum nitride [12], and tantalum pentoxide [13] have been proposed as suitable platforms for Kerr frequency combs. In addition, recent progress of pattern transfer or direct wafer bonding techniques offers further alternative opportunities for employing crystalline materials including diamond [14], lithium niobate [15], aluminum gallium arsenide [16], gallium phosphide [17], and silicon carbide [18].
Amongst these varieties of platforms, CMOS-compatible materials have some advantages since matured CMOS fabrication tools and methods can efficiently be utilized [19]. SiN has been one of the most prominent materials in this context and it features its high optical nonlinearity (about 10 times that of silica) [20], excellent transparency [21], and good affinity to silicon-based platforms for hybrid / heterogeneous integration [22]. However, dispersion management required for Kerr soliton comb generation demands high confinement waveguides in the SiN platform, where a substantial amount of normal material dispersion needs to be compensated for by geometric dispersion through appropriate designs of the cross-sectional dimensions of the waveguides [10]. This means that a waveguide with its thickness of more than 700 nm must be used in SiN, which imposes several fabrication challenges, qualitatively differing from extremely thin low-loss SiN waveguides with sub-100 nm thickness [23].
Since the thermal expansion coefficient (∼ 3.2 ppm/K) of SiN is greater than that of silicon substrate (∼ 2.2 ppm/K), high temperature deposition films, such as those made by well-established low-pressure chemical vapor deposition (LPCVD), tend to suffer from strong tensile stress, which often leads to spontaneous formation of catastrophic cracks. For this reason, several approaches have been proposed and successfully demonstrated including trench preparation [24,25], use of Si-rich SiN [26], pattern tiling that blocks crack propagation [27], careful control of deposition parameters [28], and wafer bow management [10,27,29]. Recent progress on the LPCVD-based microresonator device fabrication has demonstrated that a high-volume production of microresonators reaching their Q values beyond 107 is possible by sophisticating various techniques to reduce scattering losses [30] including the high temperature annealing at 1250°C [31].
A simple means to avoid detrimental tensile stress is to employ low-temperature deposition techniques such as plasma enhanced chemical vapor deposition (PECVD) method. PECVD SiN films have been intensively studied for waveguide applications because of its excellent compatibility to back-end-of-line (BEOL) processes [32–39]. However, SiN films prepared by this method contain a substantial amount of hydrogen content [40]. As a result of hydrogenation, some N-H bonds are formed in the film and their second harmonic overtone absorption coincides with telecommunication S and C bands, making the absorption loss too high to be used in Kerr comb generation when pumping at the wavelength of 1550 nm (∼2.0 dB/cm) [32,34,35]. By using deuterated precursors [36] or by employing high temperature annealing at around 1000°C, low-loss waveguides (∼0.58 dB/cm) [38] as well as high-Q ring resonators exceeding 1 million [39] were demonstrated. Recently, low-temperature reactive sputtering, a hydrogen-free deposition method, was also proposed in Ref. [41]. This approach appears very promising to significantly enhance the process compatibilities.
In this paper, we report SiN microring resonators prepared by HWCVD, also known as cat-CVD [42], which uses a pyrocatalytic filament for thermal cracking of the precursors without using any plasma and is capable of depositing SiN films at low temperatures. In contrast to PECVD and sputtering methods, there is no ion bombardment to the films during the deposition, implying an advantage for visible waveguide applications [43]. It also features highly efficient decomposition of SiH4 precursors [44] and good uniformity [45], relevant to large-scale photonic chip production as already exemplified in industrial applications such as solar cells, displays, and organic light emitting diodes [46]. While low-loss amorphous silicon waveguides based on the HWCVD method was reported in Ref. [47], there are few reports on SiN waveguide applications [48]. As described, it proves that post-annealing process is useful to reduce the hydrogen content in both LPCVD and PECVD SiN films [10,26,28,29,37,38]. Therefore, it is naturally expected that annealing is also effective to the HWCVD films. However, quantitative assessments of waveguide losses or Q-factors in microring resonators have not yet been reported.
The organization of this paper is as follows. We first describe our device fabrication method. Then, the annealing experiments are presented to quantify the hydrogen concentrations in the SiN films by using Fourier-transform infrared (FTIR) spectroscopy. The optical properties of the fabricated microring resonators are reported, including dispersion measurement. Then, the observation of frequency comb generation is presented by using a high-power pump laser source.
2. Device fabrication
We begin from a silicon wafer with a 4 um thick thermally grown BOX layer (Seiren KST). A HWCVD tool, made by ANELVA corp., was used with silane (SiH4) and ammonia (NH3) as precursors. A zigzag shaped tungsten wire filament with its diameter of 0.5 mm was placed approximately 100 mm below the substrate mounted on a heater with its face down. The flow rates were set to 2.3 for SiH4 and 80 sccm for NH3, respectively, and the deposition pressure was kept at 4.0 Pa (with the background pressure of 10−5 Pa). By applying 600 W of DC electric power to the filament, corresponding to the filament temperature of ∼2000°C, the resultant deposition rate was 6 nm/min. This recipe was chosen because it allows us to obtain near stoichiometric refractive index (∼2.005 at a wavelength of 633 nm) with a negligible amount wafer bowing. However, the UV absorption edge, characterized by the onset of the interference fringes in the UV reflectance spectra, was ∼4.9 eV, indicating a slightly Si-rich composition. This implies that the film has a lower density owing to the fast deposition rate. Atomic force microscopy for the deposition surface revealed that the root-mean-square surface roughness and the correlation length, important factors to the scattering losses [49], were 0.7 nm and 300 nm, respectively. Since the scattering loss becomes the greatest for a given roughness in the sub-micron range of the correlation length, an additional chemical-mechanical polishing (CMP) process might be useful for decreasing the scattering losses from the surface [34,39].
After depositing a 750 nm thick SiN film, we employed a standard subtractive process [10]. For the mask patterning, electron-beam lithography (JEOL EB-9500) with negative-tone resist (maN-2410) was used. The base dose was 180 uC/cm2 with a shot pitch of 2 nm. Dose corrections for the proximity effects as well as pattern fracturing were made with the aid of BEAMER software (GenIsys). The resist mask was developed by using maD-525 solution. We then used the mask reflow technique [24] in a convection oven at 113°C for 3 minutes.
For the pattern transfer, we used a two-step dry-etching process. In the first step, the ICP dry-etching tool (SAMCO ICP-101iPHJ) with CHF3/Ar chemistry was used for approximately 30 seconds to deposit a fluorocarbon protective layer. The resultant etch depth in this step was ∼50 nm. Then, we moved on to the second etching step using the cryo-etching tool (Oxford ICP180) using SF6/O2 chemistry at the temperature of -110°C. It turns out that the pattern transfer using only the first step results in roughening of the sidewalls after 600 nm of etching. This indicates that it is difficult to maintain the balance between the etching and deposition of the sidewall protective layer over a long etching duration, despite the optimization of the etching parameters. On the other hand, the selectivity in the cryo-etching alone was found to be too poor to etch the whole thickness of the SiN layer. We also observed a substantial amount of pattern thinning (∼100 nm), indicating more isotropic nature of cryo-etching for the SiN layer. Then, we combined these two recipes so that the sufficient selectivity is obtained while the pattern thinning is suppressed in the second step. As a result, the roughness of the sidewalls beyond 600 nm deep etching was improved (see Fig. 1(c)). Note that this is reminiscent of a single cycle of the Bosch process [50] although physical etching also occurs during a short period of time in the first step.
Fig. 1. (a) Optical microscope image of the ring resonator. (b) Cross sectional SEM image of the waveguides after the clad deposition near the junction region and (c) the SEM image of the etch waveguides. (d) Calculated dispersion profiles for the different waveguide height (with = 1.6 um) and an example of the modal intensity profile (inset).
After Piranha cleaning, the wafer was annealed in N2 ambient using a quartz tube furnace that can be operated up to 1050°C. The details are given in the next section. Then, the cladding with a thickness of 3 um was deposited by the trymethylorthosilane (TEOS) PECVD tool (SAMCO PDL-240) at 350°C and the pressure of 75 Pa. The refractive index of the cladding was measured to be 1.457 at 633 nm. Finally, the wafer was cut by using the dicing tool (DISCO DAD3230) to achieve the mirror quality end-facets. Note that this process compatibility exemplifies the low-stress SiN film prepared by the HWCVD method.
Regarding the layout design of microring resonators, the widths of the ring and bus waveguides were set to 1.6 and 1.2 um, respectively, with a pulley-coupler configuration [51] (see Fig. 1(a)). The gap widths were varied so that it is possible to obtain the best coupling conditions among them. A linear taper was placed at both edges of the chip to narrow down the bus waveguide width below 200 nm, in order to facilitate coupling to a lensed fiber or a tightly focused beam. Additional stub patterns were placed next to the taper edges to prevent the thin mask patterns from falling, and conveniently served as markers for dicing.
Figure 1(b) shows the cross-sectional SEM image of the ring resonator near the junction region of the directional coupler. The etch depth within the narrow gap (< 1 um) was substantially smaller than the other regions on the wafer. Therefore, we set the duration in the second etch step to fully etch the SiN in the narrowest gap region. Despite good surface coverage of the TEOS-CVD tool, some voids were formed above the narrow gap regions, which could be eliminated by using an additional planarization process or a high-density plasma CVD tool [29]. Figure 1(d) presents the dispersion profiles calculated by using the mode analyses tool (Lumerical) for different waveguide heights (i.e. film thickness). The dispersion parameter sensitively depends on the waveguide height while the anomalous dispersion can be obtained above 720 nm. For the thickness of 750 nm, the dispersion at the wavelength of 1550 nm was expected to be 25 ps/nm/km.
3. Anneal experiments
We examined the relationship between the residual concentration of the N-H bonds and the anneal temperatures by using a 300 nm thick uniform film deposited on a silicon substrate. Since a subtle amount of the curvature change after annealing could affect ellipsometry measurement results, we opted sufficiently thin film thickness for this experiment albeit that the sensitivity must be compromised. The anneal temperatures as well as the process duration were varied. The refractive index and thickness were measured by spectral ellipsometry (alpha-SE 2.0, J. A. Woollam) while the FTIR tool with a vacuum sample compartment (Vertex 80, Brucker) was used to quantify the residual concentration of the N-H bonds. An example of the absorbance spectra is presented in Fig. 2(a). These measurements were performed before and after the anneal process. To convert the absorbance data to the concentration values, we used the following expression [34,35,40]:
where [N-H] denotes the concentration of the N-H bonds, IN-H the integrated absorption across the peak around 3350 cm-1. AN-H is referred to as a proportionality factor (1.91 × 1017cm-2) and d the sample thickness. Since AN-H depends on the density as well as the composition of the SiN film [52], it is difficult to ensure the accuracy of the estimated [N-H] value. However, we used the value after Ref. [40] to gain a quantitative insight into our data trends and to provide comparisons to the previous studies in what follows. The limit of detection of [N-H] was approximately ∼1021 cm-3.Fig. 2. (a) FTIR spectra of HWCVD SiN film, (b) estimated concentrations of the N-H bonds and (c) refractive index and thickness reduction as a function of anneal temperature.
Figure 2(b) shows the [N-H] dependence on the anneal temperature. The [N-H] in the control sample (without annealing) was 1.12 × 1022 cm-3. Note that this is comparable to the low-hydrogen concentration PECVD films reported in the previous studies [34,35]. The data was fit by using Arrhenius type formula, from which the activation temperature was deduced to be 1308°C. This corresponds to the temperature, where the [N-H] reduces to 1/e of the default concentration. At a given temperature, the desorption process decays with time since there are a finite number of hydrogenated molecules that can thermally be activated depending on the local coordination. The decay time of the process was about 3 hours, after which no further reduction in the [N-H] were observed. The extrapolation of the plot suggests that the [N-H] can be reduced below 3 × 1021 cm3 when annealed at 1050°C.
Figure 2(c) illustrates the refractive indices and thickness changes observed in the ellipsometer. The correlation between them is evident and starts to become observable above 800°C. This can be attributed to the fast deposition rate used (∼ 6 nm/min.) and use of a slow deposition rate might help us reduce this effect. The refractive index would approach to the intrinsic value reflected by the composition, which ought to be intimately linked to the UV absorption edge, provided that no compositional change occurs.
We then decided to process our ring resonator samples at 1050°C for 3 hours. In this condition, a decrease in the film thickness of ∼6% is expected by extrapolating in Fig. 2(c). Note that the film undergoes compressive stress during the anneal, for which the damage is unlikely to occur (e.g. thermal oxide). However, since the bottom of the patterns are fixed on the substrate, some additional tensile stress should be imposed after cooling due to densification. Nevertheless, it was still possible to use the dicing tool after annealing.
4. Optical properties of microring resonators
We examined the ring resonators using a tunable laser with a linewidth of 400 kHz. Figure 3(a) displays the comparison of the transmittance through the ring resonators with and without annealing around the telecommunication C and L bands. For the sample without annealing, the resonators were undercoupled at shorter wavelengths and the resultant Q factors were less than 105 below the wavelength of 1550 nm due to the N-H overtone absorption. This was also noticeable from the background transmittance owing to the light propagation through the 4.7 mm of the bus waveguide. By contrast, it was rarely recognizable after annealing. Figure 3(b) shows the magnified resonance around 1547.6 nm after the anneal. By fitting, the loaded Q factor was estimated to be 3.8 × 105, corresponding to the intrinsic Q factor of 5.4 × 105. Note that, to estimate the intrinsic factor, we assumed an ideal coupling at the junction [53], solely accounting for the transmittance value at each resonance. Figure 3(c) presents the summary of the Q factor analyses. It is clearly seen that the overall Q factors are improved particularly at shorter wavelengths. The fact that even slightly better Q factors are obtained at shorter wavelengths possibly reflects tighter modal confinement at this wavelength range, and which implies that the waveguide losses are now dominated by the scattering losses caused by the sidewall roughness.
Fig. 3. (a) Comparison of the transmittance spectra of the microring resonators, (b) magnified spectra of the annealed sample for the resonance indicated by an arrow in (a), and (c) summary of the Q factor measurement for five samples.
In the unannealed HWCVD SiN waveguide with 1.6 × 0.5 um cross sectional dimensions (confinement factor∼0.76), the transmission losses at a wavelength of 1520 nm were reported to be ∼10 dB/cm [48]. Using the [N-H] estimated in the previous section, the absorption cross section due to the overtone absorption can be estimated to be ∼2.7 × 10−22 cm2. Although this is in reasonable agreement with the previous results [34,35], the loss in the annealed waveguides, calculated from the [N-H] combined with this cross section value, amounts to ∼ 3.5 dB/cm. This is significantly greater than the actual propagation loss (∼ 0.7 dB/cm) estimated from the Q factor (i.e. $\alpha \sim 2\pi {n_g}/({Q\lambda } )$ [54], where ng is the group index and λ the wavelength) even when taking the wavelength difference between 1520 and 1550 nm into account [38]. This discrepancy strongly suggests the presence of the additional background loss that was also decreased by the annealing process with a comparable magnitude to that of N-H absorption and is likely due to the scattering losses within the material itself. In other words, the cross section value extracted from the propagation losses in unannealed waveguides could result in an overestimated value.
The densification also substantially affects the free spectral range (FSR) of the microring resonator. The original FSR of 408 GHz was decreased to 399.3 GHz after the anneal, corresponding to the group index increase of 0.046, and which is in reasonable agreement with the refractive index increase expected from Fig. 2(c). The modal analyses predict that while the group index scales with the refractive index of SiN, it is only weakly dependent on the film thickness in this regime, consistent with the above observation.
We also measured the dispersion of the ring resonator by using a calibrated asymmetric Mach-Zhender interferometer with an FSR of 4 MHz as a reference (see Fig. 4(a)). Resonance frequencies of a dispersive resonator can be expressed as a Taylor expansion around an angular frequency ω0:
Fig. 4. (a) Schematic experimental setup. PD: photodiode, a-MZI: asymmetric Mach-Zhender interferometer. (b) Measured dispersion profile (FSR∼200 GHz).
The result is shown in Fig. 4(b), where D is estimated to be 23 ps/nm/km. While the decrease in the film thickness significantly reduces D, as presented in Fig. 1(b), the refractive index increase due to the densification slightly increases D. Thus, the resultant D after annealing is a combined effect of these two factors. The value reported above suggests that the actual thickness of the waveguide patterns is not significantly modified by the densification despite significant volume reduction observed in the uniform film. In fact, it was also difficult to distinguish any changes in the cross sectional dimensions in the SEM images. However, we note that since the measurement spectral range must be increased to accurately characterize a small amount of dispersion, the error in this measurement could be as large as ∼50% due to the limited tuning range of the laser source used. Improvements in this direction would help us clarify how the anneal process affects the dispersion properties of the resonators.
5. Parametric oscillations
Since the annealed waveguides still exhibit anomalous dispersion, we launched a high-power pump to one of the resonators with a 400 GHz FSR at around 1563 nm by using a high-power erbium doped amplifier (EDFA), as shown in Fig. 5(a). The coupling efficiency was about 35%. For a given pump power, the laser frequency was manually tuned from the blue side of the resonance so that the high pump power can gradually be coupled to the microresonator despite the thermally-induced resonance frequency shift. As shown in Fig. 5(b), the parametric oscillation starts at the launched pump power of ∼100 mW, resulting in primary combs separated by ∼4.42 THz. As we increased the pump power up to 525 mW, the sidebands of the primary combs gradually extended to merge with each other and a fully populated flat comb spanning from the telecommunication S to L band was eventually observed (see Fig. 5(c) and (d)). The shift of the resonance due to the thermal effect was as much as 200 GHz. Although we tried the excite soliton microcombs by various means, it turned out to be difficult to generate any soliton states. Our numerical simulation predicts that while a high power per comb line can be anticipated with a moderate Q factor and a large FSR, as reported in this paper, a greater amount anomalous dispersion (> 40 ps/nm/km) is required to support solitons [55]. This minimum dispersion requirement is relaxed when the Q factor is improved.
Fig. 5. (a) Schematic of experimental setup and (b-d) the output spectra due to parametric oscillations at different pump powers. OSA: Optical spectral analyzer.
6. Summary
We presented the frequency comb generation in the SiN based microring resonator prepared by HWCVD method. Since it allows us to deposit low-stress films, no special precautions against crack generation and propagation were required even for thick waveguides, facilitating the fabrication process compatibilities. We also detailed the anneal temperature dependence of the HWCVD SiN films, showing that the concentration of the N-H bonds monotonically decreased as we increased the anneal temperature above 600°C. By annealing at 1050°C, the intrinsic Q factor as high as 5.4 x105 was achieved, and which was likely to be limited by the scattering losses. This led us to observe frequency combs due to the spontaneous parametric oscillation by using the high-power pump source. These results highlight interesting opportunities offered by HWCVD-based SiN films for waveguide device applications and future improvements can be anticipated by sophisticating the dry etching process as well as employing the other techniques such as CMP, slow rate deposition, and annealing at higher temperatures.
Funding
Japan Society for the Promotion of Science (JP17K06410).
Acknowledgments
K. F. acknowledges Yoshimi Yamashita and Kanna Aoki for fruitful discussion.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
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. T. J. Kippenberg, A. L. Gaeta, M. Lipson, et al., “Dissipative Kerr solitons in optical microresonators,” Science 361(6402), eaan8083 (2018). [CrossRef]
2. A. L. Gaeta, M. Lipson, and T. J. Kippenberg, “Photonic-chip-based frequency Combs,” Nat. Photonics 13(3), 158–169 (2019). [CrossRef]
3. J. Pfeifle, V. Brasch, M. Lauermann, et al., “Coherent Terabit communications with microresonator Kerr frequency combs,” Nat. Photonics 8(5), 375–380 (2014). [CrossRef]
4. P. Marin-Palomo, J. N. Kemal, M. Karpov, et al., “Microresonator-based solitons for massively parallel coherent optical communications,” Nature 546(7657), 274–279 (2017). [CrossRef]
5. A. Fulop, M. Mazur, A. Lorences-Riesgo, et al., “High-order coherent communications using mode-locked dark-pulse Kerr combs from microresonators,” Nat. Commun. 9(1), 1598 (2018). [CrossRef]
6. B. Corcoran, M. Tan, X. Xu, et al., “Ultra-dense optical data transmission over standard fibre with a single chip source,” Nat. Commun. 11(1), 2568 (2020). [CrossRef]
7. J. Liu, E. Lucas, A. S. Raja, et al., “Photonic microwave generation in the X- and K-band using integrated soliton microcombs,” Nat. Photonics 14(8), 486–491 (2020). [CrossRef]
8. M. G. Suh, Q. F. Yang, K. Y. Yang, et al., “Microresonator soliton dual-comb spectroscopy,” Science 354(6312), 600–603 (2016). [CrossRef]
9. D. T. Spencer, T. Drake, T. C. Briles, et al., “An optical-frequency synthesizer using integrated photonics,” Nature 557(7703), 81–85 (2018). [CrossRef]
10. J. S. Levy, A. Gondarenko, M. A. Foster, et al., “CMOS-compatible multiple-wavelength oscillator for on-chip optical interconnects,” Nat. Photonics 4(1), 37–40 (2010). [CrossRef]
11. L. Razzari, D. Duchesne, M. Ferrera, et al., “CMOS-compatible integrated optical hyper-parametric oscillator,” Nat. Photonics 4(1), 41–45 (2010). [CrossRef]
12. H. Jung, C. Xiong, K. Y. Fong, et al., “Optical frequency comb generation from aluminum nitride microring resonator,” Opt. Lett. 38(15), 2810–2813 (2013). [CrossRef]
13. H. Jung, S. -P. Yu, D. R. Carlson, et al., “Tantala Kerr nonlinear integrated photonics,” Optica 8(6), 811–817 (2021). [CrossRef]
14. B. Hausmann, I. Bulu, V. Venkataraman, et al., “Diamond nonlinear photonics,” Nat. Photonics 8(5), 369–374 (2014). [CrossRef]
15. C. Wang, M. Zhang, M. Yu, et al., “Monolithic lithium niobate photonic circuits for Kerr frequency comb generation and modulation,” Nat. Commun. 10(1), 978 (2019). [CrossRef]
16. L. Chang, W. Xie, H. Shu, et al., “Ultra-efficient frequency comb generation in AlGaAs-on-insulator microresonators,” Nat. Commun. 11(1), 1331 (2020). [CrossRef]
17. D. J. Wilson, K. Schneider, S. Hönl, et al., “Integrated gallium phosphide nonlinear photonics,” Nat. Photonics 14(1), 57–62 (2020). [CrossRef]
18. C. Wang, J. Li, A. Yi, et al., “Soliton formation and spectral translation into visible on CMOS-compatible 4H-silicon-carbide-on-insulator platform,” Light: Sci. Appl. 11(1), 341 (2022). [CrossRef]
19. E. G. Johnson, M. J. Shaw, G. P. Nordin, et al., “Fabrication techniques for low-loss silicon nitride waveguides,” Proc. SPIE 5720, 109 (2005). [CrossRef]
20. D. T. H. Tan, K. Ikeda, P. C. Sun, et al., “Group velocity dispersion and self-phase modulation in silicon nitride waveguides,” Appl. Phys. Lett. 96(6), 061101 (2010). [CrossRef]
21. S. V. Deshpande, E. Gulari, S. W. Brown, et al., “Optical properties of silicon nitride films deposited by hot filament chemical vapor deposition,” J. Appl. Phys. 77(12), 6534–6541 (1995). [CrossRef]
22. C. Xiang, P. A. Morton, and J. E. Bowers, “Ultra-narrow linewidth laser based on a semiconductor gain chip and extended Si3N4 Bragg grating,” Opt. Lett. 44(15), 3825–3828 (2019). [CrossRef]
23. J. F. Bauters, M. J. Heck, D. John, et al., “Ultra-low-loss high-aspect-ratio Si3N4 waveguides,” Opt Express 19(4), 3163–3174 (2011). [CrossRef]
24. K. Luke, A. Dutt, C. B. Poitras, et al., “Overcoming Si3N4 film stress limitations for high quality factor ring resonators,” Opt. Express 21(19), 22829–22833 (2013). [CrossRef]
25. Y. Xuan, Y. Liu, L. T. Varghese, et al., “High-Q silicon nitride microresonators exhibiting low-power frequency comb initiation,” Optica 3(11), 1171–1180 (2016). [CrossRef]
26. C. J. Kruckel, A. Fulop, T. Klintberg, et al., “Linear and nonlinear characterization of low-stress high-confinement silicon-rich nitride waveguides,” Opt. Express 23(20), 25827–25837 (2015). [CrossRef]
27. M. H. P. Pfeiffer, C. Herkommer, J. Liu, et al., “Photonic damascene process for low-loss, high-confinement silicon nitride waveguides,” IEEE J. Sel. Top. in Quantum Electron. 24(4), 1–11 (2018). [CrossRef]
28. H. El Dirani, A. Kamel, M. Casale, et al., “Annealing-free Si3N4 frequency combs for monolithic integration with Si photonics,” Appl. Phys. Lett. 113(8), 081102 (2018). [CrossRef]
29. H. El Dirani, L. Youssef, C. Petit-Etienne, et al., “Ultralow-loss tightly confining Si3N4 waveguides and high-Q microresonators,” Opt. Express 27(21), 30726–30740 (2019). [CrossRef]
30. X. Ji, F. A. S. Barbosa, S. P. Roberts, et al., “Ultra-low-loss on-chip resonators with sub-milliwatt parametric oscillation threshold,” Optica 4(6), 619–624 (2017). [CrossRef]
31. J. Liu, G. Huang, R. N. Wang, et al., “High-yield, wafer-scale fabrication of ultralow-loss, dispersion-engineered silicon nitride photonic circuits,” Nat. Commun. 12(1), 2236 (2021). [CrossRef]
32. M. Melchiorri, N. Daldosso, F. Sbrana, et al., “Propagation losses of silicon nitride waveguides in the near-infrared range,” Appl. Phys. Lett. 86(12), 121111 (2005). [CrossRef]
33. A. Gorin, A. Jaouad, E. Grondin, et al., “Fabrication of silicon nitride waveguides for visible-light using PECVD: a study of the effect of plasma frequency on optical properties,” Opt. Express 16(18), 13509–13516 (2008). [CrossRef]
34. S. C. Mao, S. H. Tao, Y. L. Xu, et al., “Low propagation loss SiN optical waveguide prepared by optimal low-hydrogen module,” Opt. Express 16(25), 20809–20816 (2008). [CrossRef]
35. E. A. Douglas, P. Mahony, A. Starbuck, et al., “Effect of precursors on propagation loss for plasma-enhanced chemical vapor deposition of SiNX:H waveguides,” Opt. Mater. Express 6(9), 2892–2903 (2016). [CrossRef]
36. X. Cheng, J. Hong, A. M. Spring, et al., “Fabrication of a high-Q factor ring resonator using LSCVD deposited Si3N4 Film,” Opt. Mater. Express 7(7), 2182–2187 (2017). [CrossRef]
37. J. Chiles, N. Nader, D. D. Hickstein, et al., “Deuterated silicon nitride photonic devices for broadband optical frequency comb generation,” Opt. Lett. 43(7), 1527–1530 (2018). [CrossRef]
38. L. Wang, W. Xie, D. Van Thourhout, et al., “Nonlinear silicon nitride waveguides based on a PECVD deposition platform,” Opt. Express 26(8), 9645–9654 (2018). [CrossRef]
39. X. Ji, Y. Okawachi, A. Gil-Molina, et al., “Ultra-low-loss silicon nitride photonics based on deposited films compatible with foundries,” Laser Photonics Rev. 17(3), 2200544 (2023). [CrossRef]
40. W. A. Lanford and M. J. Rand, “The hydrogen content of plasma-deposited silicon nitride,” J. Appl. Phys. 49(4), 2473–2477 (1978). [CrossRef]
41. A. Frigg, A. Boes, G. Ren, et al., “Optical frequency comb generation with low temperature reactive sputtered silicon nitride waveguides,” APL Photonics 5(1), 011302 (2020). [CrossRef]
42. H. Matsumura, “Formation of silicon-based thin films prepared by catalytic chemical vapor deposition (Cat-CVD) Method,” Jpn. J. Appl. Phys. 37(6R), 3175 (1998). [CrossRef]
43. P. Neutens, M. Rutowska, W. Van Roy, et al., “Mitigation of UV-induced propagation loss in PECVD silicon nitride photonic waveguides,” ACS Photonics 5(6), 2145–2150 (2018). [CrossRef]
44. H. Umemoto, T. Morimoto, M. Yamawaki, et al., “Deposition chemistry in the cat-CVD processes of the SiH4/NH3 system,” Thin Solid Films 430(1-2), 24–27 (2003). [CrossRef]
45. A. Ledermann, U. Weber, C. Mukherjee, et al., “Influence of gas supply and filament geometry on the large-area deposition of amorphous silicon by hot-wire CVD,” Thin Solid Films 395(1-2), 61–65 (2001). [CrossRef]
46. H. Matsumura and K. Ohdaira, “New application of cat-CVD technology and recent status of industrial implementation,” Thin Solid Films 517(12), 3420–3423 (2009). [CrossRef]
47. S. Z. Oo, A. Tarazona, A. Z. Khokhar, et al., “Hot-wire chemical vapor deposition low-loss hydrogenated amorphous silicon waveguides for silicon photonic devices,” Photonics Res. 7(2), 193–200 (2019). [CrossRef]
48. K. Furusawa, Y. Yamashita, K. Aoki, et al., “Hot-wire CVD based SiN films for linear and nonlinear photonics device applications,” in Laser Congress 2018 (ASSL), OSA Technical Digest (Optica Publishing Group, 2018), paper ATu2A.11. [CrossRef]
49. J. P. R. Lacey and F. P. Payne, “A theoretical analysis of scattering loss from planar optical waveguides,” Opt. Quantum Electron. 26(10), 977–986 (1994). [CrossRef]
50. F. Laerme, A. Schilp, K. Funk, et al., “Bosch deep silicon etching: improving uniformity and etch rate for advanced MEMS applications,” Technical Digest. IEEE International MEMS 99 Conference. Twelfth IEEE International Conference on Micro Electro Mechanical Systems (Cat. No.99CH36291), Orlando, FL, USA, pp. 211–216 (1999).
51. D. T. Spencer, J. F. Bauters, M. J. R. Heck, et al., “Integrated waveguide coupled Si3N4 resonators in the ultrahigh-Q regime,” Optica 1(3), 153–157 (2014). [CrossRef]
52. V. Verlaan, C. H. M. van der Werf, W. M. Arnoldbik, et al., “Unambiguous determination of Fourier-transform infrared spectroscopy proportionality factors: the case of silicon nitride,” Phys. Rev. B 73(19), 195333 (2006). [CrossRef]
53. M. H. P. Pfeiffer, J. Liu, M. Geiselmann, et al., “Coupling Ideality of Integrated Planar High-Q Microresonators,” Phys. Rev. Appl. 7(2), 024026 (2017). [CrossRef]
54. P. Rabiei, W. H. Steier, C. Zhang, et al., “Polymer micro-ring filters and modulators,” J. Lightwave Technol. 20(11), 1968–1975 (2002). [CrossRef]
55. S. Fujii, S. Tanaka, T. Ohtsuka, et al., “Dissipative Kerr soliton microcombs for FEC-free optical communications over 100 channels,” Opt. Express 30(2), 1351–1364 (2022). [CrossRef]
