A 1000 nm wide supercontinuum, spanning from 1470 nm in the telecom band to 2470 nm in the mid-infrared is demonstrated in a 800 nm x 220 nm 1 cm long hydrogenated amorphous silicon strip waveguide. The pump source was a picosecond Thulium doped fiber laser centered at 1950 nm. The real part of the nonlinear parameter of this waveguide at 1950 nm is measured to be 100 ± 10 W−1m−1, while the imaginary part of the nonlinear parameter is measured to be 1.2 ± 0.2 W−1m−1. The supercontinuum is stable over a period of at least several hours, as the hydrogenated amorphous silicon waveguides do not degrade when exposed to the high power picosecond pulse train.
© 2013 Optical Society of America
The mid-infrared (mid-IR) wavelength range, which is generally considered to span wavelengths in the 2-20 μm range, is of key interest for various sensing and spectroscopy applications, since many molecules have characteristic absorption bands in this wavelength range. One of the great challenges in developing mid-IR systems for various applications is the lack of practical sources and detectors. Nonlinear optics provides a good way to generate new optical frequencies spaced a predictable and controllable distance from a strong pump frequency and is thus well-placed to take advantage of commercially available sources at shorter wavelengths to generate mid-IR spatial and/or temporally coherent radiation. Supercontinuum generation (SCG) has received a lot of attention from researchers in recent years  because of the breadth of its potential applications such as in optical coherence tomography , wavelength division multiplexing in telecommunications [3,4], in optical sensing  and in spectroscopy [6,7]. A lot of the work in supercontinuum generation has focused on the use of photonic crystal fibers . The ability to tailor the dispersion profile of such fibers with high precision enables the generation of large bandwidth supercontinuum spectra. On-chip supercontinuum generation – which could make supercontinuum sources lower cost, more robust and more power efficient compared to the microstructured fiber approach – has been achieved on various waveguide platforms including chalcogenide [8–10], silicon nitride , lithium niobate  and silicon . Implementing nonlinear optical functionality in silicon photonic integrated circuits has many advantages since it provides tight confinement due to the high index contrast, resulting in high waveguide intensities for modest optical power levels. The high index contrast allows for dispersion engineering of the waveguides, a critical feature for efficient nonlinear interaction. Waveguides can be fabricated with relatively low loss (~0.5-2 dB/cm) using CMOS-compatible processes, resulting in a potential interaction length of several centimeters, which combined with the very high nonlinear parameter enables strong nonlinear interaction. However, two-photon absorption (TPA) can be a problem when working with high optical intensities since not only does TPA directly cause losses, but also the resulting free carriers lead to more nonlinear losses through free carrier absorption. Therefore, for crystalline silicon where the half bandgap wavelength lies at 2.2 μm, one has to work close to or beyond this wavelength to avoid these nonlinear losses. This requires the use of bulky optical parametric oscillator systems as a pump source, which hampers the development of compact and low-cost systems. Hydrogenated amorphous silicon (a-Si:H) waveguides on the other hand, are known to have a similar index contrast and nonlinear index as crystalline silicon and have been demonstrated to be suitable for various nonlinear optics based applications [14–20]. The material has a higher bandgap, enabling the pump to be located at shorter wavelengths where, for example, compact Thulium-based fiber sources are now commercially available. This way one can work with higher powers than is usually possible at telecom wavelengths without suffering from the adverse effects of TPA. In  we reported that hydrogenated amorphous silicon can be unstable if pumped in the 1550 nm wavelength range due to what is known as the Staebler-Wronski effect, which causes degradation of the material through breaking of the Si-Si bonds. This effect is also minimized by working at a pump wavelength of 1950 nm because the bonds purportedly break due to electron-hole pairs recombining after having been created via TPA. However, this effect seems to be related to the material deposition technique, as other groups have reported stable operation at 1550 nm [21, 22]. Supercontinuum generation in the 1.5-2.5 μm wavelength range is of high importance for the spectroscopic analysis of water-rich fluids, given the relatively low absorption of water in this wavelength range and the existence of overtone transitions of many molecular bonds in this wavelength range. From the above discussion, the use of hydrogenated amorphous silicon waveguides in combination with a Thulium-based fiber source as the pump is a promising approach for this application. In this paper we elaborate on the generation of a 1.47 μm to 2.47 μm spanning supercontinuum using a picosecond pulsed Thulium-based fiber source in hydrogenated amorphous silicon waveguides. The device performance is compared to similar devices implemented on a crystalline silicon photonics platform.
2. Supercontinuum generation
Supercontinuum generation with a picosecond pulsed pump source is mostly mediated by four-wave mixing (FWM) and the associated rise of modulation instability bands from noise . The position of these modulation instability (MI) side lobes is determined by the dispersion characteristics of the waveguide. In general, anomalous group velocity dispersion (GVD) is required for the phase matching of the MI bands which are close to the pump frequency. However, one can also generate MI bands farther away from the pump wavelength when taking into account the higher order dispersion terms. These bands grow from noise and consequently, the position of the band peak is wherever the gain is at its maximum, i.e. where perfect phase matching is obtained. The degenerate four-wave mixing mechanism in the undepleted pump approximation describes the initial experimental situation. The phase mismatch between the linear propagation constants of the pump and the generated idler and signal frequencies is compensated for by the nonlinear phase mismatch arising from self- and cross-phase modulation (SPM and XPM) as given by Eq. (1) below .Eq. (2) below where k0 is the propagation constant in vacuum, n2 the Kerr nonlinear index of the material, Aeff is the effective area of the waveguide and βTPA is the two-photon absorption coefficient.Eq. (1), the β4 term can be ignored for small frequency separations and we get the condition that the GVD has to be anomalous (β2 < 0) for materials with a positive nonlinear parameter Re(γ), resulting in a phase matched wavelength band close to the pump (labeled MI1 in this paper). At higher values of ∆ω, the β4 term has to be taken into consideration. This way, along with the condition of anomalous GVD, another phase matched wavelength band far away from the pump occurs if β4 > 0 . This second modulation instability band (labeled MI2 in this paper) is what allows for the wide breadth of the generated supercontinuum. Of course, there is self-phase modulation mediated broadening, but at picosecond pump durations, this broadening is much smaller compared to the overall experimentally observed breadth of the supercontinuum.
For the experiment, a Thulium doped modelocked fiber laser from AdValue Photonics with a pulse duration of 1.24ps (full width at half maximum) and a pulse repetition rate of 26 MHz is used. Highly nonlinear hydrogenated amorphous silicon (a-Si:H) photonic wires were defined by depositing a 220 nm a-Si:H layer on top of a 1950 nm polished silicon dioxide layer on a silicon substrate in a CMOS pilot line. The photonic wires were patterned using wafer-scale CMOS fabrication technology. The inset of Fig. 1 shows the scanning electron microscope (SEM) cross section of the waveguide used in the experiment which has average width of 792 nm and a height of 218 nm as measured by the SEM. The pump laser is coupled to the fundamental quasi-TE mode of the waveguide using a surface etched grating coupler (grating period 1040 nm, duty cycle 50%, −12 dB coupling efficiency at 1950 nm for TE polarization under a 25 degree fiber angle), while the waveguide output is collected using a lensed fiber and connected to a mid-IR optical spectrum analyzer. The source is passed through a polarizer and polarization rotator in order to ensure maximum coupling to the waveguide fundamental TE mode. As shown in Fig. 1, at a coupled peak power in the waveguide of 7.6 W, the MI1 side lobes appear on the output spectra. With an increase in power the second pair of MI2 side lobes appears farther away from the pump. The MI2 side lobe on the long wavelength side is less visible because the waveguide losses increase strongly with wavelength (from 2.2 dB/cm at 1950 nm to 7 dB/cm at 2400 nm). As power is increased further, these bands merge and create a continuum. With increasing power, the supercontinuum broadens until it is 1000 nm wide (−40 dB bandwidth) as shown in Fig. 1. Corresponding to a 0.65 octave supercontinuum (−30 dB bandwidth), the required pulse energy, peak and average powers were 21 pJ, 16.5 W and 0.54 mW respectively. Table 1 presents an overview of supercontinuum sources presented in literature, which are compared to the results obtained in this work.
While the supercontinuum presented in this work is not the broadest on-chip supercontinuum demonstration, it does highlight the possibility of using the CMOS compatible a-Si:H material platform for SCG using low pulse energies and by working at a wavelength where commercial fiber lasers are available.
3. Characterization of the a-Si waveguide parameters
In order to quantitatively compare the amorphous silicon platform to earlier reported results, the waveguide nonlinear parameter and material stability were investigated. Below, we describe the results of those characterizations.
3.1 Material stability
It has been reported previously in  that the amorphous silicon material is not stable against exposure to high optical intensity at 1550 nm, which leads to material degradation. This effect was attributed to the breaking of weak Si-Si bonds in the material mediated by the recombination of carriers created by two-photon absorption. Working at longer wavelengths should significantly decrease the TPA and consequently the material degradation. Indeed this is what was observed, as shown in Fig. 2 where the supercontinuum is maintained for several hours without any significant degradation of the spectrum, even though the peak power coupled into the waveguide was 60 ± 10W.
In , degradation in the a-Si:H waveguides was observed at 9 times lower average power and 12 times lower peak power levels and in a time frame more than 40 times shorter compared to the stability experiment shown in Fig. 2. Thus it is clear from this demonstration that the problem of material degradation is mitigated by operating at the longer wavelength of 1950 nm where commercial fiber sources are still available. This makes the use of hydrogenated amorphous silicon waveguides appealing for real-life applications.
3.2 TPA measurement for Im(γ) determination
In order to measure the two-photon absorption in the hydrogenated amorphous silicon waveguides at 1950 nm wavelength, the reciprocal of the optical transmission through such a waveguide was measured as a function of peak input power. From , we can relate the TPA coefficient to the inverse of the transmission using Eq. (3):Figure 3 below shows the results of the transmission measurement with increasing input peak powers. From the slope of the linear fit in Fig. 3(b), Im(γ) is calculated to be 1.2 ± 0.2 W−1m−1, which from Eq. (2) gives βTPA = 2.3x10−13 mW−1. This matches well with the value reported in  for the same wavelength. Compared with the value of the Im(γ) reported in  at 1550 nm wavelength (28 W−1m−1), the value obtained in the current experiment is significantly lower which accounts for the material stability observed in Fig. 2.
3.3 Determination of Re(γ)
A measurement of the self-phase modulation of the pump was made to measure the value of the real part of the nonlinear parameter γ. Figure 4 shows the SPM measurements and the comparison to the simulation of the nonlinear Schrödinger equation by the split-step Fourier transform method which included carrier effects . This gives a value for Re(γ) of 100 ± 10 W−1m−1. The wavy nature observed in the experimental output spectra is already present in the input pulse spectrum and relates to specific implementation of the Thulium doped fiber laser.
Taking this value of Re(γ) and considering the positions of the MI1 and MI2 bands in Fig. 1, we can infer the values of β2 and β4 using Eq. (1). At the pump wavelength of 1950 nm, we get β2 = −0.4 ps2/m and β4 = 1.3x10−4 ps4/m. Simulations of the dispersion of the waveguide using the finite difference method were carried out in this wavelength range to confirm this β2 value. It is known from previous work [13, 28] that β4 values are difficult to predict by simulations because of uncertainties in waveguide geometry which unfortunately have a large impact on the value of the higher order dispersion terms like β4. Thus, one can only compare the simulated and extracted GVD values. Since the material dispersion of amorphous silicon is unknown, we assumed the material dispersion of crystalline silicon as an approximation. From these simulations, a waveguide of dimensions 832 nm x 212 nm is found to match the experimental value of β2. The dimensions of the simulated waveguide lie within 5% error of the measured dimensions in Fig. 1 and the difference is thought to be due to the material dispersion of amorphous silicon being slightly different from that of the crystalline silicon.
4. Supercontinuum generation in crystalline silicon waveguides
In order to demonstrate the large performance benefits from hydrogenated amorphous silicon waveguides for this application, supercontinuum generation in a crystalline silicon waveguide with nominally identical waveguide dimensions was also investigated. As shown in Fig. 5, the positions of the MI1 and MI2 side lobes show that the dispersion in the crystalline silicon waveguide is similar to that of the amorphous waveguide. Clearly, crystalline silicon waveguides provide a much narrower supercontinuum at similar power levels, even when the waveguide dimensions, linear loss and dispersion are all comparable to the hydrogenated amorphous silicon waveguide. This is because of the higher TPA coefficient of the crystalline silicon at this wavelength (about 4 times higher), which makes it unable to support a similarly wide supercontinuum.
We have reported broadband supercontinuum generation in a hydrogenated amorphous silicon waveguide spanning from 1470 nm to 2470 nm. By pumping the waveguide at 1950 nm wavelength using a commercially available picosecond laser source, we take advantage of the very low TPA coefficient in hydrogenated amorphous silicon at this wavelength. Contrasting this with the broadening observed in a crystalline silicon waveguide of the same dimensions, similar loss and dispersion, clearly the hydrogenated amorphous silicon provides a much broader supercontinuum at similar power levels. The material is also shown to be stable when exposed for several hours continuously to the high peak powers required for the supercontinuum generation. Hydrogenated amorphous silicon is thus brought forth as a promising material for on-chip nonlinear optics.
This work was supported by the FP7-ERC-MIRACLE project. We also acknowledge the generous support extended by the suppliers of the picosecond laser source used in this work, AdValue Photonics. Bart Kuyken acknowledges a scholarship provided by the Fund for Scientific Research Flanders (FWO-Vlaanderen).
References and links
1. J. M. Dudley, G. Genty, and S. Coen, “Supercontinuum generation in photonic crystal fiber,” Rev. Mod. Phys. 78(4), 1135–1184 (2006). [CrossRef]
2. I. Hartl, X. D. Li, C. Chudoba, R. K. Ghanta, T. H. Ko, J. G. Fujimoto, J. K. Ranka, and R. S. Windeler, “Ultrahigh-resolution optical coherence tomography using continuum generation in an air-silica microstructure optical fiber,” Opt. Lett. 26(9), 608–610 (2001). [CrossRef] [PubMed]
3. T. Morioka, H. Takara, S. Kawanishi, O. Kamatani, K. Takiguchi, K. Uchiyama, M. Saruwatari, H. Takahashi, M. Yamada, T. Kanamori, and H. Ono, “1 Tbit/s (100 Gbit/s×10 channel) OTDM/WDM transmission using a single supercontinuum WDM source,” Electron. Lett. 32(10), 906–907 (1996). [CrossRef]
4. T. Nakasyotani, H. Toda, T. Kuri, and K. Kitayama, “Wavelength-division-multiplexed millimeter-waveband radio-on-fiber system using a supercontinuum light source,” J. Lightwave Technol. 24(1), 404–410 (2006). [CrossRef]
5. C. F. Kaminski, R. S. Watt, A. D. Elder, J. H. Frank, and J. Hult, “Supercontinuum radiation for applications in chemical sensing and microscopy,” Appl. Phys. B 92(3), 367–378 (2008). [CrossRef]
6. Y. Liu, M. D. King, H. Tu, Y. Zhao, and S. A. Boppart, “Broadband nonlinear vibrational spectroscopy by shaping a coherent fiber supercontinuum,” Opt. Express 21(7), 8269–8275 (2013). [CrossRef] [PubMed]
7. S. T. Sanders, “Wavelength-agile fiber laser using group-velocity dispersion of pulsed super-continua and application to broadband absorption spectroscopy,” Appl. Phys. B 75(6-7), 799–802 (2002). [CrossRef]
8. M. R. E. Lamont, B. Luther-Davies, D. Y. Choi, S. Madden, and B. J. Eggleton, “Supercontinuum generation in dispersion engineered highly nonlinear (γ = 10 /W/m) As2S3 chalcogenide planar waveguide,” Opt. Express 16(19), 14938–14944 (2008). [CrossRef] [PubMed]
9. C. W. Rudy, A. Marandi, K. L. Vodopyanov, and R. L. Byer, “Octave-spanning supercontinuum generation in in situ tapered As₂S₃ fiber pumped by a thulium-doped fiber laser,” Opt. Lett. 38(15), 2865–2868 (2013). [CrossRef] [PubMed]
10. D. D. Hudson, S. A. Dekker, E. C. Mägi, A. C. Judge, S. D. Jackson, E. Li, J. S. Sanghera, L. B. Shaw, I. D. Aggarwal, and B. J. Eggleton, “Octave spanning supercontinuum in an As2S3 taper using ultralow pump pulse energy,” Opt. Lett. 36(7), 1122–1124 (2011).
11. R. Halir, Y. Okawachi, J. S. Levy, M. A. Foster, M. Lipson, and A. L. Gaeta, “Ultrabroadband supercontinuum generation in a CMOS-compatible platform,” Opt. Lett. 37(10), 1685–1687 (2012). [CrossRef] [PubMed]
12. C. R. Phillips, C. Langrock, J. S. Pelc, M. M. Fejer, J. Jiang, M. E. Fermann, and I. Hartl, “Supercontinuum generation in quasi-phase-matched LiNbO3 waveguide pumped by a Tm-doped fiber laser system,” Opt. Lett. 36(19), 3912–3914 (2011). [CrossRef] [PubMed]
13. B. Kuyken, X. Liu, R. M. Osgood Jr, R. Baets, G. Roelkens, and W. M. J. Green, “Mid-infrared to telecom-band supercontinuum generation in highly nonlinear silicon-on-insulator wire waveguides,” Opt. Express 19(21), 20172–20181 (2011). [CrossRef] [PubMed]
14. B. Kuyken, S. Clemmen, S. K. Selvaraja, W. Bogaerts, D. Van Thourhout, P. Emplit, S. Massar, G. Roelkens, and R. Baets, “On-chip parametric amplification with 26.5 dB gain at telecommunication wavelengths using CMOS-compatible hydrogenated amorphous silicon waveguides,” Opt. Lett. 36(4), 552–554 (2011). [CrossRef] [PubMed]
17. S. Suda, K. Tanizawa, Y. Sakakibara, T. Kamei, K. Nakanishi, E. Itoga, T. Ogasawara, R. Takei, H. Kawashima, S. Namiki, M. Mori, T. Hasama, and H. Ishikawa, “Pattern-effect-free all-optical wavelength conversion using a hydrogenated amorphous silicon waveguide with ultra-fast carrier decay,” Opt. Lett. 37(8), 1382–1384 (2012). [CrossRef] [PubMed]
19. K.-Y. Wang, K. G. Petrillo, M. A. Foster, and A. C. Foster, “Ultralow-power all-optical processing of high-speed data signals in deposited silicon waveguides,” Opt. Express 20(22), 24600–24606 (2012). [CrossRef] [PubMed]
20. B. Kuyken, H. Ji, S. Clemmen, S. K. Selvaraja, H. Hu, M. Pu, M. Galili, P. Jeppesen, G. Morthier, S. Massar, L. K. Oxenløwe, G. Roelkens, and R. Baets, “Nonlinear properties of and nonlinear processing in hydrogenated amorphous silicon waveguides,” Opt. Express 19(26), B146–B153 (2011). [CrossRef] [PubMed]
21. J. Matres, G. C. Ballesteros, P. Gautier, J. M. Fédéli, J. Martí, and C. J. Oton, “High nonlinear figure-of-merit amorphous silicon waveguides,” Opt. Express 21(4), 3932–3940 (2013). [CrossRef] [PubMed]
22. C. Grillet, L. Carletti, C. Monat, P. Grosse, B. Ben Bakir, S. Menezo, J. M. Fedeli, and D. J. Moss, “Amorphous silicon nanowires combining high nonlinearity, FOM and optical stability,” Opt. Express 20(20), 22609–22615 (2012). [CrossRef] [PubMed]
23. G. P. Agrawal, Nonlinear Fiber Optics (Academic, 2001).
24. G. Roelkens, U. D. Dave, A. Gassenq, N. Hattasan, C. Hu, B. Kuyken, F. Leo, A. Malik, M. Muneeb, E. Ryckeboer, S. Uvin, Z. Hens, R. Baets, Y. Shimura, F. Gencarelli, B. Vincent, R. Loo, J. Van Campenhout, L. Cerutti, J. B. Rodriguez, E. Tournié, X. Chen, M. Nedeljkovic, G. Mashanovich, L. Shen, N. Healy, A. C. Peacock, X. Liu, R. Osgood, and W. Green, “Silicon-based heterogeneous photonic integrated circuits for the mid-infrared,” Opt. Mater. Express 3(9), 1523–1536 (2013). [CrossRef]
25. H. K. Tsang, R. V. Penty, I. H. White, R. S. Grant, W. Sibbett, J. B. D. Soole, H. P. LeBlanc, N. C. Andreadakis, R. Bhat, and M. A. Koza, “Two-photon absorption and self-phase modulation in InGaAsP/InP multi-quantumwell wave-guides,” J. Appl. Phys. 70(7), 3992–3994 (1991). [CrossRef]
26. L. Shen, N. Healy, P. Mehta, T. D. Day, J. R. Sparks, J. V. Badding, and A. C. Peacock, “Nonlinear transmission properties of hydrogenated amorphous silicon core fibers towards the mid-infrared regime,” Opt. Express 21(11), 13075–13083 (2013). [CrossRef] [PubMed]
28. F. Leo, U. D. Dave, B. Kuyken, S. Keyvaninia, and G. Roelkens, “Measurement and tuning of the chromatic dispersion around the half band gap spectral region of a silicon photonic wire,” Opt. Lett. (submitted).