Open Access
Add to BibSonomy
Abstract
PFCB is a semi-fluorinated polymer suitable for integrated optics due to its low absorption at the telecommunication wavelength range, compatibility with standard fabrication tools and processes, and thermomechanical robustness. In this work we report the fabrication and characterization of waveguides optimized for Kerr nonlinearity enhancement, comprised of PFCB core and optical-grade Cytop cladding having a 10% index contrast and supporting single-mode operation with an effective mode area of 4 µm2. Propagation loss of 0.9 dB/cm and 1 dB/facet coupling loss using lensed optical fibers having 2.5 µm spot size were obtained, leading to an effective length of Leff =0.82 cm. A nonlinear coefficient of γ=0.18 ± 0.03 (W·m)-1 was measured for both TE and TM polarizations using a four-wave mixing arrangement about 1.55 µm, with no observed nonlinear absorption. A relatively high Kerr coefficient of |n2|=1.9×10−15 cm2/W is derived for PFCB in the telecom wavelength range for the first time.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Perfluorocyclobutyl (PFCB) polymer is a semi-fluorinated polymer suitable for realization of optical devices at telecom wavelengths. Its polymer chain is constructed of aromatic ether repeating units for which thermomechanical robustness (i.e., Tg > 200°C) and solution processability are not compromised [1,2]. Being a semi-fluorinated polymer, PFCB exhibits low intrinsic loss of 0.26 dB/cm at 1.55µm due to the replacement of C-H bonds with C-F ones [3]. The thermal cyclopolymerization to synthesize PFCB together with the monomers’ ratio afford tunable optical properties and high temperature robustness. Together with its low absorption, this makes PFCB polymer an attractive material for optical devices over standard polymers for potential use in telecommunication regime [4]. Even in comparison to halogenated polymers, which in general show negligible transmission losses in the desired wavelength range, PFCB semi-fluorinated polymer presents lower losses [5]. Moreover, the ability to host various ligand-terminated inclusions in polymers for nanocomposite functionalization [6] has been demonstrated in PFCB for achieving optical gain with dispersion of InAs nanoparticles with an aromatic aniline capping ligand [7].
A thorough research on the properties of different PFCB monomer compositions and ratios for optical applications in the near-IR (telecommunication) and visible was previously performed [8]. The glass transition temperature, refractive index and thermal stability at 1550 nm and 630 nm wavelengths for each composition were found. PFCB has been used to construct waveguides having a PFCB core of BPVE:TVE (bistrifluorovinyloxy biphenyl: trifluorovinylaryl-ether) and a PFCB cladding of 6PVE:TVE [9], achieving an index contrast of 3%. However, to date, PFCB’s nonlinear properties are not published in the literature and have not been studied for nonlinear applications. This paper reports on the fabrication and characterization of high index contrast optical waveguides, constructed of TVE:BPVE (3:7 monomers ratio) PFCB core and Cytop polymer cladding, and optimized for maximizing third-order nonlinear applications at the resulting 10% refractive index contrast ratio.
To achieve high-performance nonlinear waveguides (NL-WG) the key requirements are strong confinement of the optical mode, low propagation and coupling losses, low chromatic dispersion, strong NL-WG parameter, γ, and lack of two photon absorption (TPA). Cytop is also a fluorinated polymer with a waveguide loss <0.1 dB/cm at 1.55 µm wavelength [10]. The use of polymers in WG fabrication allows for established clean room processes such as spin coating, plasma ashing and reactive ion etching (RIE). However, perfluorinated polymers can be challenging on account of adhesion and mechanical integrity at interfaces. These issues are addressed in detail in Section 2. Linear waveguide characteristics, i.e., propagation and coupling losses, are discussed in Section 3 and nonlinear properties of the PFCB waveguide in Section 4, where PFCB’s third-order susceptibility and lack of two-photon absorption (TPA) is revealed.
The combination of its excellent linear characteristics and relatively high third-order nonlinear properties makes PFCB-core/Cytop-cladding waveguides a good candidate for parametric nonlinear waveguiding operations.
2. Polymeric WG design and fabrication
Previously, we reported in brief on our PFCB-core/Cytop-cladding WG design and fabrication [11]. We herein review the design and fabrication process of our previously reported work and provide additional optimizations leading to a robust WG realization.
Our WG structure consists of a Cytop lower cladding (3.5 µm thickness, required to isolate the optical mode from the silicon substrate), the PFCB core layer with a waveguide rib, followed by Cytop over cladding. As we target NL-WG applications, we design the WG rib dimensions to support only the fundamental mode and aim to minimize the mode’s effective area, Aeff, which is achieved for PFCB core of dimension 1.7(W)×2.1(H) µm2, with a 600 nm thin supporting layer remaining to improve adhesion to the Cytop layer beneath. At the chosen WG dimension, we achieve effective areas of 4.3 µm2 and 4.1 µm2 for TE and TM modes, respectively. The TE/TM spatial mode profiles are very similar to each other (see Fig. 1). The rib geometry can accommodate single mode operation in many size scales, facilitating fiber coupling [12], yet here our motivation is to minimize its mode size for maximizing the nonlinearity.
Fig. 1. PFCB-core/Cytop-cladding waveguide design. (a) Spatial distribution of refractive indices, n(x,y). (b) Effective area of TE and TM modes as a function of waveguide width. (c) TE mode and (d) TM mode field distributions.
Several issues had to be resolved for reliable fabrication of the PFCB core structure: synthesis of PFCB polymer to achieve good solution viscosity and degree of polymerization for optimal spin coating and curing, adhesion of the polymer layers and more. PFCB synthesis was based on (TVE:BPVE) monomers, dissolved in Mesitylene solvent. The PFCB polymerization synthesis was carried out under an inert environment (nitrogen gas flow) in a three-neck flask immersed in an oil bath at 160°C for 8 hours while constantly stirring (see Fig. 2(a)). NMR spectroscopy was used to determine the polymerization percentage, calculated by the quantity of polymer divided by the entire material, i.e., monomers and polymer (Fig. 2(b)). An optimization of polymerization duration, temperature, ratio of TVE:BPVE monomers (1:1, 1:9 and 3:7) and concentration of monomers in the Mesitylene was carried out in order to reach 50% polymerization required for the desired viscosity which achieves good spin-coating ability (uniform layer thickness spreading and adhesion to the Cytop layer). Higher degree polymerization was reached following spin-coating by an additional curing session in an inert oven. The final PFCB polymerization percentage was measured by dissolving the core layer in chloroform and applying NMR spectroscopy to the solution (Cytop is resistant to the chloroform solvent). Of the three monomer ratios and concentrations in Mesitylene, 3:7 (TVE:BPVE) with 50 wt% showed the best process performance, achieving excellent spin-coating properties and reaching 70% final polymerization. The PFCB monomer ratio results in 1.49 refractive index, compared to Cytop’s refractive index of 1.334, yielding a ∼10% refractive index contrast. After applying a silane surface treatment to a polished silicon substrate, the lower cladding layer of 3.5 µm thick Cytop was spin-coated. This was done in two steps to achieve a uniform layer thickness. First spin coating layer was applied at 1500 RPM for 45 sec, followed by partial curing of 20 min on a hot plate set to 180°C. Second spin coating layer was performed at 2000 RPM and final curing in an oven set to 180°C overnight. The measured total Cytop thickness was 3.5 µm with 20 nm deviation across the wafer. To ensure good adhesion of PFCB to Cytop, N2 plasma ashing surface treatment was applied, as a positive termination layer. Next, the 50% polymerized PFCB solution was spin coated at 1000 RPM for 45 sec to achieve 2.1 µm layer thickness. A 50 nm thick PECVD SiO2 layer was deposited to serve as a hard mask. The SiO2 was patterned using photoresist and contact photolithography, and partially etched into the PFCB layer to a 1.5 µm depth, forming the core’s rib structure while leaving a 600 nm layer for enhanced adhesion (Fig. 2(c)). Two orders of magnitude coefficient of thermal expansion (CTE) mismatch between SiO2 and the polymers required all processing steps to be performed at low temperatures (below 100°C) to avoid cracks and deformations in each fabrication process, i.e., deposition of PECVD SiO2 hard mask, lithography and reactive ion etching (RIE). A second N2 plasma ashing surface treatment to the PFCB was applied prior to Cytop over-cladding spin coating (see Fig. 2(e) for complete process flow). To protect the fabricated device during dicing and edge polishing, we bonded a thin, fused silica glass wafer cover on top of the polymer stack using Norland optical adhesive (UV curable). PFCB is not affected by the level of UV illumination required for the adhesive, lacking acrylic groups in its structure. Singulated chips were formed by dicing and polishing the edge facets with the glass cover intact. A microscope image of rib waveguides at the edge facet can be seen in Fig. 2(d).
Fig. 2. Waveguide fabrication process: (a) PFCB synthesis setup, (b) NMR spectroscopy of partially cured PFCB, (c) SEM image of the PFCB rib waveguides, prior to application of over-cladding, (d) Microscope image of chip edge facet after polishing, and (e) Process flow of WG fabrication including a photograph of the polisher.
3. Linear characterization
The fabricated PFCB-core polymer waveguides were optically accessed at the ingress and egress edges using lensed fibers with a spot size diameter of 2.5 µm (with a small air gap due to the lensed fiber’s working distance). We measured the linear guiding properties of our polymer WG using a measurement setup based on an optical vector analyzer (OVA) instrument (LUNA Innovations, model 5100), operating as a swept laser interferometer over a spectral range of 1540-1570 nm. Light was coupled from the OVA source through the ingress lensed fiber to the WG device and out via the egress lensed fiber to the OVA detector (Fig. 3). Figures 4(a) and 4(b) show the total measured insertion loss (IL) in the frequency and time domains, respectively, with IL being lower than 6 dB inclusive of all contributions, which we break down next. In the frequency domain, we observe Fabry-Perot fringes due to Fresnel reflections at the WG facets (Fig. 4(a)). In the time domain plot this is manifest as weak echo pulses with a relative time delay of 88 psec due to the pulse’s round-trip time within the waveguide (Fig. 4(b)). To support this, we calculated the group velocity of the propagating mode using finite element method, finding vg = 0.21 m/ns which leads to 86 psec round trip delay for our 9 mm long WG, in excellent match to the observed value and WG length uncertainty. To estimate the propagation loss of the waveguide, α, we consider all loss mechanisms impacting the echo pulse strength with respect to the main pulse, which eliminates external and common contributions such as coupling in/out of the WG and the quality of the lensed fibers (which affect the main and echo pulses equally). The echo pulse additionally experiences two facet reflections and propagation loss over twice the WG length, and can be expressed by the equation:
where EA is the measured echo pulse attenuation relative to the main pulse, FrR is the Fresnel reflection at the facets, MML is the mode mismatch loss between incident waveguide mode and the reflected mode, which evolves from the spatially dependent refractive index, and L is the waveguide length.Fig. 3. Top: Linear characterization setup. Bottom: Photo of ∼9 mm long waveguide chip, accessed via lensed fibers on input and output facets.
The measured echo attenuation between the direct light path and the first echo pulse is 30.5 dB (Fig. 4(b)). The theoretical calculation of the Fresnel reflection and MML at the waveguide facets are detailed in the Appendix and contribute 14.4 dB and 0.02 dB, respectively. This yields a 14.42 dB reflection loss per facet. Using these calculated values in Eq. (1), the estimated propagation loss is 0.9 dB/cm, which is a good value for tightly-confined, high index contrast polymer WGs. Typical published loss metrics for polymer waveguides exhibit lower propagation losses (e.g., [13]), but these are reported at low refractive index contrast ratios (<1%) making a direct comparison invalid, as losses due to sidewall roughness scale with the index contrast ratio squared [14]. Note that we observe that the echo attenuation of the secondary pulse is only -26 dB, which may suggest that the WG propagation loss is lower, but we have less confidence for a signal at this level (attenuated by more than 55 dB) as it is prone to numerical roundoff errors associated with the FFT calculation for the time domain representation.
The total average IL of 5.9 dB appearing in Fig. 4(a) includes the contributions of the two lensed fibers, which contribute 3 dB to the loss as directly measured when optimally aligned to each other (back-to-back). Further subtracting the propagation loss of 0.8 dB for the 9 mm long waveguide, we arrive at an estimated coupling loss of 1 dB per facet. The theoretical coupling loss (CL) is affected by the Fresnel transmission (FrT) into the polymer waveguide and the mode mismatch loss (MML) between the fiber mode (assumed Gaussian and refracted into the polymer core/cladding) and the waveguide mode:
The Gaussian lensed fiber mode of 2.5 µm waist diameter and the waveguide mode are drawn to the same scale in Figs. 5(a) and 5(b). The computed values for FrT and MML are 0.15 dB and 0.14 dB, respectively, which result in 0.29 dB theoretical coupling loss per facet (see Appendix for more details). The coupling loss we experimentally estimate is higher than the theoretical one which might be due to facet imperfections (imperfect polished surface, silicon particles at the interface from the polishing process, as seen in Fig. 2(d), etc.) and misalignment contributions. Differences between theoretical and actual reflection losses may also impact our propagation loss estimates, with higher reflection losses implying lower propagation losses.
Fig. 5. Mode profile simulations for (a) lensed fiber after refraction into the waveguide and (b) WG mode.
4. Nonlinear characterization
Waveguides exhibiting third order nonlinearity are characterized by the NL parameter γ derived from the real part of third-order susceptibility χ(3), and βTPA derived from the imaginary part of χ(3), where [15]:
To assess whether there is any nonlinear absorption present, different input powers ranging from +10 to +30 dBm at 1.55 µm wavelength were coupled to the waveguide using a single amplified pump and measured against the output power. Figure 6 shows the IL (ratio of output power to input power) for cases of high-power excitations. The IL does not vary, suggesting there is no two-photon absorption process present and that $Im\{{{\chi^{(3 )}}} \}$ is insignificant at least up to +30 dBm input power.
To calculate the nonlinear WG parameter γ, a degenerate Four-Wave-Mixing (FWM) experiment was conducted. Two CW pumps (P1, P2) spaced 0.65 nm apart (1548.8 and 1549.5 nm) were passively combined, amplified by a high-power EDFA and coupled into the WG. Both pumps were set to the same power and polarization. Four input powers were used from 24 dBm to 28 dBm (P1 + P2). The measurements were done with and without the WG device using the same conditions (i.e., initial pump launch powers and collected power by detuning the lensed fibers in the back-to-back case), to identify the contribution of the WG device versus the output fiber patch cords. We placed a fiber Bragg grating (FBG) notch filter with 20 dB attenuation around 1550 nm right before the WG under test, into which the generated red shifted FWM harmonic falls into, to distinguish the difference of the idlers caused by the delivery fibers and by both fibers and WG device. An illustration of the optical setup is presented in Fig. 7(a). Figure 7(b) shows an exemplary output spectrum, with two strong pumps and two FWM formed peaks, one of them appearing in the Bragg grating notch (P4). To extract γWG we used the expression describing the idler output power [16]:
Fig. 7. Measurement of nonlinear FWM term. (a) Optical setup scheme. (b) Typical optical power spectrum from OSA at a resolution of 0.01 pm.
Figure 8 shows the optical power spectrum with the WG device and without at the same input and output powers. We set the same output power when measuring without the WG device to obtain the same contribution from the read-out patch fibers. This was done by deliberate misalignment of the back-to-back lensed fibers. The reference FWM signal in the case without the device (in red), i.e., the entire setup without the device such that IL=5.9 dB between the lensed fibers. An increase of 1 dB is observed for FWM harmonic with the WG device. After scanning the input powers, P1 + P2, from 24 dBm to 28 dBm at the entrance of the WG device, we measured the FWM idler and plotted the generation efficiency, P4/P1 as a function of P2. This was done with and without the WG device with same loss in between the lenses as the WG insertion loss. In this manner the contribution of the fibers to the nonlinearity is cancelled out. The efficiency expressions for the two cases, derived from Eq. (5), are as follow:
Figure 9 shows the measured efficiency metric for the TM and TE polarizations in dB scale. The blue lines correspond to the total FWM efficiency (i.e., P4/P1) for fibers + WG device vs. input power of P2 pump (recall P1 = P2). The magenta lines represent the reference measurement, i.e., FWM efficiency of fibers alone without WG device at the same input power levels. The efficiencies scale with the peak pump power squared, as expected. The fitting slopes are 2.06 ± 0.1, in agreement with the expected square-law dependence of Eqs. (6–7). The evaluated WG NL parameter for the PFCB-core WG is γ=0.18 ± 0.03 (W·m)-1 for both TE and TM polarizations. The major source of the error originates from amplified pump powers deviations of ∼0.1 dB from the laser sources themselves. Other deviations such as coupling loss are measured during alignment. From the extracted NL WG parameter, γ, we can evaluate $Re\{{{\chi^{(3 )}}} \}$ of PFCB. Using Eq. (3), we find $|{Re\{{{\chi^{(3 )}}} \}} |= 1.5 \times {10^{ - 21}}\; {m^2}/{V^2}$ in telecom wavelength (λ=1.55 µm) which yields a Kerr coefficient of |n2|=1.9×10−15 cm2/W. This value is one order of magnitude greater than fused silica (2.2×10−16 cm2/W at λ=1 µm [17]), the same order of magnitude as PAZ polymer (4.0×10−15 cm2/W at λ=1.55 µm [18]), PECVD Si3N4 (2.4×10−15 cm2/W at λ=1.55 µm [19]), but an order of magnitude less than silicon-rich nitride (1.4×10−14 cm2/W at λ=1.55 µm [20]) and GaN (3.4×10−14 cm2/W at λ=1.55 µm [16]) and two orders less than ultra-silicon-rich nitride (2.8×10−13 cm2/W at λ=1.55 µm [21]).
Fig. 9. FWM efficiency vs. pump power for (a) TM polarization: PFCB WG in blue, no WG device in magenta. (b) TE polarization: PFCB WG in blue, no WG device in magenta.
5. Conclusions and outlook
We produced a highly confined nonlinear polymer waveguide of PFCB-core/Cytop-cladding having a mode effective area of 4 µm2. The 1.7×2.1µm2 rib waveguide exhibited 0.9 dB/cm propagation loss and coupling loss of 1 dB per facet when using lensed fibers of 2.5 µm mode diameter. An absence of TPA up to 30 dBm was confirmed. A measurement of the WG nonlinearity using a degenerate FWM experiment revealed a NL coefficient of γ=0.18 ± 0.03 (W·m)-1 for TE and TM modes. From this experimental value we obtained $|{Re\{{{\chi^{(3 )}}} \}} |= 1.5 \times {10^{ - 21}}\textrm{ }{m^2}/{V^2}$ for PFCB for the first time. This value can be compared to other materials used in waveguide fabrication, being an order of magnitude higher than silica, of the same order as PAZ polymer and stoichiometric silicon nitride, and 1-2 orders of magnitude less than GaN and silicon rich nitride.
Nonlinear waveguides ideally should exhibit low coupling and propagation losses, high NL WG parameter, and be free of two-photon absorption at 1550 nm. The PFCB-core WG demonstrated here achieves many of these requirements and makes it potentially useful for parametric nonlinear-optical signal processing. Although PFCB does not have the highest intrinsic nonlinear material characteristics compared to leading platforms, such as chalcogenide glass with higher nonlinearity but with higher losses [22], it does benefit from fabrication ease and simple butt-coupling. To enhance parametric nonlinear processes, longer WG devices are required. However, as the effective length is the determining factor, waveguide losses have to be further reduced. We believe sidewall roughness is the limiting factor in the fabricated waveguides as PFCB’s intrinsic loss is lower than our waveguide loss. Also, the waveguide geometry can be further engineered for chromatic dispersion, which was not attempted in this work.
Funding
Israel Science Foundation (1359/07).
Acknowledgments
The authors graciously acknowledge the funding provided by the Israel Science Foundation, the support of the Harvey M Kruger Family Center for Nanoscience and Nanotechnology at the Hebrew University, and fruitful discussions with Profs. Roy Shenar and Meital Reches.
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.
Supplemental document
See Supplement 1 for supporting content.
References
1. D. W. Smith Jr, D. A. Babb, H. V. Shah, A. Hoeglund, R. Traiphol, D. Perahia, and M. Radler, “Perfluorocyclobutane (PFCB) polyaryl ethers: versatile coatings materials,” J. Fluorine Chem. 104(1), 109–117 (2000). [CrossRef]
2. E. B. Caldona, E. I. Borrego, K. E. Shelar, K. M. Mukeba, and D. W. Smith Jr, “Ring-Forming Polymerization toward Perfluorocyclobutyl and Ortho-Diynylarene-Derived Materials: From Synthesis to Practical Applications,” Materials 14(6), 1486 (2021). [CrossRef]
3. A. Yeniay, R. Gao, K. Takayama, R. Gao, and A. F. Garito, “Ultra-Low-Loss Polymer Waveguides,” J. Lightwave Technol. 22(1), 154–158 (2004). [CrossRef]
4. D. W. Smith Jr., S. M. Kumar, S. Chen, J. M. Ballato, E. J. Nelson, J. J. Jin, and S. H. Foulger, “Property tailored perfluorocyclobutyl (PFCB) copolymers for optical devices,” Design, Manufacturing, and Testing of Planar Optical Waveguide Devices. Vol. 4439. International Society for Optics and Photonics (2001).
5. M. Jöhncka, L. Müllera, A. Neyera, and J. W. Hofstraat, “Copolymers of halogenated acrylates and methacrylates for the application in optical telecommunication: Optical properties, thermal analysis and determination of unsaturation by quantitative FT-Raman and FT-IR spectroscopy,” Eur. Polym. J. 36(6), 1251–1264 (2000). [CrossRef]
6. C. Ingrosso, A. Panniello, R. Comparelli, M. Lucia Curri, and M. Striccoli, “Colloidal inorganic nanocrystal based nanocomposites: functional materials for micro and nanofabrication,” Materials 3(2), 1316–1352 (2010). [CrossRef]
7. G. Chen, R. Rapaport, D. T. Fuchs, L. Lucas, A. J. Lovinger, and S. Vilan, “Optical gain from InAs nanocrystal quantum dots in a polymer matrix,” Appl. Phys. Lett. 87(25), 251108 (2005). [CrossRef]
8. J. Ballato, S. H. Foulger, and D. W. Smith Jr, “Optical properties of perfluorocyclobutyl polymers. II. Theoretical and experimental attenuation,” J. Opt. Soc. Am. B 21(5), 958–967 (2004). [CrossRef]
9. J. Ballato and S. L. Pyshkin, “Advanced light emissive materials for novel optical displays, lasers, waveguides, and amplifiers,” Moldavian J. of Physical Sciences 5(2), 195–208 (2006).
10. Y. Kuwana, S. Takenobu, K. Takayama, S. Yokotsuka, and S. Kodama, “Low Loss and Highly Reliable Polymer Optical Waveguides with Perfluorinated Dopant-free Core,” in Proc. Opt. Fiber Commun. (OFC) Conf. 2006), paper OWF2.
11. M. Bin Nun and D. M. Marom, “Low loss high index contrast polymeric waveguide,” 2014 International Conference on Optical MEMS and Nanophotonics, Glasgow, pp. 197–198 (2014).
12. R. Moosburger and K. Petermann, “4 × 4 digital optical matrix switch using polymeric oversized rib waveguides,” IEEE Photon. Technol. Lett. 10(5), 684–686 (1998). [CrossRef]
13. J. Kobayashi, T. Matsuura, Y. Hida, S. Sasaki, and T. Maruno, “Fluorinated Polyimide Waveguides with Low Polarization-Dependent Loss and Their Applications to Thermooptic Switches,” J. Lightwave Technol. 16(6), 1024–1029 (1998). [CrossRef]
14. C. G. Poulton, C. Koos, M. Fujii, A. Pfrang, T. Schimmel, J. Leuthold, and W. Freude, “Radiation Modes and Roughness Loss in High Index-Contrast Waveguides,” IEEE J. Select. Topics Quantum Electron. 12(6), 1306–1321 (2006). [CrossRef]
15. C. Koos, L. Jacome, C. Poulton, J. Leuthold, and W. Freude, “Nonlinear silicon-on-insulator waveguides for all-optical signal processing,” Opt. Express 15(10), 5976–5990 (2007). [CrossRef]
16. D. Munk, M. Katzman, O. Westreich, M. Nun, Y. Lior, N. Sicron, Y. Paltiel, and A. Zadok, “Four-wave mixing and nonlinear parameter measurement in a gallium-nitride ridge waveguide,” Opt. Mater. Express 8(1), 66–72 (2018). [CrossRef]
17. P. Kabacinski, T. M. Kardas, Y. Stepanenko, and C. Radzewicz, “Nonlinear refractive index measurement by SPM-induced phase regression,” Opt. Express 27(8), 11018 (2019). [CrossRef]
18. A. Faccinetto, S. Mazzucato, D. Pedron, R. Bozio, S. Destri, and W. Porzio, “Non-Resonant z-Scan Characterization of the Third-Order Nonlinear Optical Properties of Conjugated Poly (thiophene azines),” ChemPhysChem 9(14), 2028–2034 (2008). [CrossRef]
19. K. Ikeda, R. E. Saperstein, N. Alic, and Y. Fainman, “Thermal and Kerr nonlinear properties of plasma-deposited silicon nitride/silicon dioxide waveguides,” Opt. Express 16(17), 12987–12994 (2008). [CrossRef]
20. C. J. Krückel, A. Fülöp, T. Klintberg, J. Bengtsson, P. A. Andrekson, and V. Torres-Company, “Linear and nonlinear characterization of low-stress high-confinement silicon-rich nitride waveguides,” Opt. Express 23(20), 25827–25837 (2015). [CrossRef]
21. D. T. H. Tan, D. K. T. Ng, J. W. Choi, E. Sahin, B.-U. Sohn, G. F. R. Chen, P. Xing, H. Gao, and Y. Cao, “Nonlinear optics in ultra-silicon-rich nitride devices: recent developments and future outlook,” Adv. Phys.: X 6(1), 1905544 (2021). [CrossRef]
22. R. Qi, W. Zhang, and Y. Huan, “Nonlinear optical properties of chalcogenide glass waveguides fabricated by hot melt smoothing and micro-trench filling,” Appl. Phys. Express 13(4), 042005 (2020). [CrossRef]
