Highly nonlinear planar glass waveguides have been shown to be useful for all optical signal processing. However, the typical SMF-28 fiber to waveguide coupling loss of ~5dB/end remains a barrier to practical implementation. Low loss coupling to a fiber using vertical tapering of the waveguide film is analyzed for rib and nanowire waveguides and experimentally demonstrated for ribs showing polarization and wavelength independence over >300nm bandwidth. Tapers with essentially zero excess loss led to total losses from the waveguide to fiber core of 1.1dB per facet comprising only material absorption (0.75dB) and mode overlap loss (0.36dB), both of which can be eliminated with improvements to processing and materials.
© 2013 OSA
Small mode area waveguides fabricated from a wide variety of nonlinear glass films have proven very effective in a plethora of nonlinear optical and high bit rate transmission system experiments, of which [1–21] are a small cross section. Similarly planar waveguide amplifier devices fabricated in a range of deposited glass films also have shown great promise but also only function efficiently with small mode area waveguides [22–26]. Despite the low waveguide losses attained in some systems [e.g. 17, 27–31], high bit rate experiments, in particular, have been hampered by the typically >10dB connector to connector insertion loss once the fiber interface is included. Given the sheer range of technologically significant demonstrations [1–21], and the variety of means by which they have been accomplished, then clearly finding an effective solution to low loss coupling between standard fibers and small mode area glass film based waveguides is a worthwhile endeavour. Taking the 2x0.85μm rib waveguides used in many nonlinear demonstrations as an example, then the large connector-to-connector loss is a combination of several factors. The lowest coupling losses are usually attained using lens tipped SMF-28 fibers e.g . with ~2.5μm mode field diameters which have an associated ~2.5dB connector to connector loss by themselves, in addition to the ~1dB/end Fresnel reflection loss and ~2.5dB/end overlap loss to the waveguide mode. The situation is of course even worse for nanowire based devices e.g .
To understand the impact of high coupling losses and what can be gained by reducing them, a simple analysis of the most commonly used nonlinear process in all-optical signal processing, four wave mixing (FWM), makes matters clear. The issue in most transmission system based experiments using highly nonlinear waveguides for FWM is insufficient output signal to noise ratio due to low overall conversion efficiency. In this context, an analytical expression for the FWM conversion efficiency derived by Batelgelj  provides insight. Neglecting pump depletion, and in the low power regime at the optimum length, the converted internal signal power in a length of fiber was derived, Eq. (1), as:Eq. (2):
Spot size convertors for improved fiber coupling to small HIC waveguides have been extensively studied in the last 20 years primarily for semiconductor lasers e.g . and more recently for Silicon nanowire waveguides. The most effective device demonstrated to date has been the so called inverse taper, where the cross section of the HIC waveguide is reduced adiabatically in either or both of width and height to form a down-taper towards the coupling fiber. This is performed in the first case to yield a “remnant” HIC waveguide with very much smaller dimensions than the untapered HIC waveguide such that the mode approaches cut off and expands to a size close to a fiber mode field yet remains guided by the remnant HIC core (e.g .). In the second option, the down taper is made to cut off and the expelled optical field captured in an overlaid waveguide structure which is mode matched to a suitable fiber [36–41]. The best results from these approaches are those of Tsuchizawa  and Bakir .
Bakir et al.  hold the record for lowest loss SSC at present with a silicon nanowire waveguide inverse taper tip device using an intermediate SiOx rib waveguide with about 10% index contrast. Coupling losses of about 0.25dB at 1550nm to an SMF-28 lensed fiber with a mode field diameter of 3μm and low polarization dependence were demonstrated. The need for a lens tipped fiber lay in the relatively high index contrast of the intermediate waveguide. However, the low loss quoted did not then include the losses of the lens tipped fiber which are significant, and showed considerable wavelength sensitivity with losses rising to >1dB below 1520nm. Tsuchizawa et al. [36, 37] followed the same route but with a 3% index contrast SiON intermediate waveguide and produced a device with 0.5dB loss from a silicon nanowire to a fiber with a 4.3μm mode field that could be spliced with low loss to SMF-28. Wide bandwidth operation was also demonstrated between 1350 and 1700nm with about 0.5dB additional loss at wavelengths below ~1400nm. No information on polarization sensitivity was provided.
Whilst these results are certainly a very significant improvement, the key issue with these down tapers has been the demanding lithography required. In fact for the remnant HIC core device, the 1dB coupling loss bandwidth corresponds to only +/−10nm variation in the core size . Further, control of line edge roughness and the consequent additional sidewall scattering induced optical losses at these small dimensions becomes an issue Even for a “simple” down taper to “zero” width, the taper tip width for polarization independent wide bandwidth operation requires sub-60nm resolution to ensure wideband performance for the TM mode [39, 42]. Clearly an approach which avoids these challenges would be very useful.
For waveguides fabricated from PVD films (e.g. sputtering, evaporation, plasma enhanced chemical vapor deposition, etc.), an elegant and simple alternative exists. With PVD, the film can be structured with vertically tapered edges simply by using a “shadow” mask during deposition. Here an aperture is spaced from the substrate surface and the material being deposited “diffuses” under the edge to produce a film edge that tapers away to zero thickness over a distance determined by the aperture to substrate distance. Importantly, zero thickness is truly reached and the taper surface is as smooth as that of the rest of the film thereby avoiding roughness induced scattering losses in the taper. Tapering by the shadow masking technique was first considered as early as 1973  and demonstrations coupled slab waveguides exist from that time. The methodology can realise a SSC by adding an overlaid low index contrast waveguide as described above, and the technique also provides for low loss coupling to rib waveguides of the type used in many glass waveguide nonlinear/EDWA demonstrations. Figure 1 illustrates the basic concept, here for a rib waveguide.
Actual demonstrations of devices using this principle in glass waveguides are, however, rather scarce, presumably due to the lack of incentive to investigate HIC glass waveguides until relatively recently because of their new found active and nonlinear properties. The first report of a 2-D waveguide vertically tapered SSC of any type was for coupling from fiber to small electro-optic polymer waveguides  via a vertically tapered intermediate glass waveguide, and after a few subsequent publications, little more seems to have been done using glasses in any form. Vertical tapers have, however, been widely applied in III-V semiconductor devices with fabrication by a wide range of methods, all somewhat awkward as tapered single crystal films have not yet been grown, (e.g . and references therein). The best results published to date are those of Soares et al. . Here a 2mm long vertical taper coupler was fabricated using grey scale lithography and dry etching to produce an SSC between an InGaAsP HIC waveguide and a high NA fiber with a 3.6μm MFD. An intermediate underlying InP rib waveguide was used to capture the light from the vertical taper, this also (unusually) having a tapered etch depth. Lowest total losses of 0.9dB were obtained, 0.4dB of which was taper excess loss and 0.5dB was fiber mode overlap loss. AR coatings were applied to the facets to reduce the reflection losses from the high index waveguide ends. There appeared little scope for further reduction of the overlap loss due to the nature of the intermediate waveguide design, and reducing the taper losses further was considered to require somewhat longer devices and so more propagation losses. No information was provided on polarization sensitivity or spectral performance.
This paper reports the design and realization of vertically tapered Chalcogenide glass HIC SSCs fabricated by shadow mask deposition, reactive ion etching, and using a polymer intermediate waveguide to mode match commercially available high NA fibers. Modeling shows the method works well for both rib and nanowire devices. Total coupling losses of 1.1dB per end from the fiber core to a Chalcogenide rib waveguide core were demonstrated, with both polarization insensitivity and operational bandwidth exceeding 300nm. Crucially the taper itself showed essentially zero excess loss with the main losses occurring in overlap due to incorrect waveguide dimensions (0.36dB) and polymer absorption (0.75dB) which can both be reduced substantially with already established processes to produce devices with potentially <0.2dB/end coupling losses.
2. Design and modeling
There are a number of design constraints and desirable properties for the taper system. Firstly the overlay waveguide should have good overlap to a suitable optical fiber and be single mode to enable future integration of other components in the low index contrast system. Secondly, the taper must be adiabatic so there is negligible excess loss. Thirdly the whole design needs to be robust to lateral alignment errors of the overlay waveguide, and be low loss (<0.5dB/end) even with misalignment factored in. Fourthly, low reflectivity is of great importance for applications where high power continuous wave pumps are used, e.g. optical phase conjugation. Fifthly, wavelength and polarization independence are highly desirable. For nonlinear waveguides it is also desirable that the intensity in the taper is no higher than that in the untapered high index waveguide, as in nonlinear experiments it is common practice to run close to the damage threshold and increasing the intensity locally in the taper would result in damaging the taper before reaching the desired power level in the HIC waveguide. Additionally it is useful if the input/output waveguide suffers no radiation loss to a silicon substrate for bottom cladding thicknesses of no more than ~5μm since this reduces costs and allows for easy hybrid integration of the high index waveguide to semiconductor devices, such as laser diodes, in the longer term. Finally it would be of benefit if the overlay waveguide system is capable of having relatively tight bend radius (~1mm radius or less) to enable passive circuit integration for more complex future hybrid applications.
The effects of these constraints pushes the design in a particular direction. The bottom-cladding objective essentially sets a lower limit on the core refractive index for the overlay waveguide. For processing convenience, SU-8 was chosen as the core material as it afforded a simple means to fabricate the overlay waveguide and modeling using R-Soft FEMSim showed it could certainly meet the requirement for substrate radiation losses. Whist the use of SU-8 is convenient, as a polymer it introduces some compromises (loss being the primary one) but it is ultimately not limiting as, for example, Tsuchizawa et al.  showed by moving from SU-8 to a low temperature processed SiOx waveguide with no absorption losses.
Addressing the desire to attain good overlap and small bend radii but retain single mode operation imposes some significant constraints on the waveguide dimensions and index contrast. To mode match SMF-28 directly typically requires a waveguide design with core dimensions ~6x6μm and attaining single mode operation for a core this size requires an index contrast <0.5% that is only compatible with > 1cm bend radii . High numerical aperture fibers optimized for coupling to planar waveguide circuits can however be spliced to SMF-28 with low loss. Fibers such as Nufern UHNA-3 or Fibercore SM-1500 [47, 48] can be directly fusion spliced with <0.2dB loss to SMF-28  with suitable splicers and splicing parameters and by thermal expansion of the core or fattening and tapering methods splice losses of <0.05 dB are attainable [50–52]. These types of fibers have much smaller mode field diameters than SMF-28 (ranging 3.2-4.2μm [47,48] for UHNA-7 to SM-1500) and thereby enable the use of index contrasts of a few present in the intermediate waveguide that allow relatively tight bending whilst retaining single mode operation .
Figure 2(a) indicates the generic waveguide design for which Fig. 2(b) shows the minimum calculated standard overlap integral to a Gaussian mode with 1/e2 full width equal to the UHNA-3 fiber mode field diameter, the dimension of the square waveguide at minimum overlap loss, and single mode operation boundary for different top cladding indices. The refractive index of the SU-8 was determined to be 1.575 at 1550nm using an SCI Filmtek 4000 dual angle spectroscopic reflectometer from measurements of 3μm thick films of SU-8 prepared as they would be under normal lithographic processes.
Evidently very low coupling losses can be obtained for any top clad index below 1.550, but the single mode operating region is relatively narrow extending down to only 1.540. Conveniently, Ormocore  has a refractive index of 1.535 at 1550nm and by slightly reducing the waveguide size to 3.3μm, single mode operation can be restored. This has negligible impact on the predicted overlap loss (increases to 0.05dB) and the 2% index contrast corresponding to the parameters at this size enabled 90 degree bends with 1.5mm radius but excess losses <0.1dB in other published experiments  thereby allowing for future integration of other components in the low index contrast waveguide system.
Having finalized a suitable intermediate waveguide design, then in order to model the projected device performance the actual taper profile had to be experimentally determined. A knife edged aperture mask was suspended 0.6mm away from the surface of a 100mm diameter silicon wafer and 850nm of As2S3 thermally evaporated though it to form a taper using the set up previously described . The wafer was mounted on a carousel 400mm above the evaporation source which could be placed either under the center of rotation of the carousel or under the circle traced out by the centers of the wafers as they rotated. A thin Aluminium layer was then sputtered onto the wafers to eliminate thin film interference effects when using a Veeco NT9100 white light interferometric optical profiler to measure the taper profiles. Figure 3 shows the profiles obtained which were highly uniform across the wafer diameter.
Either a quasi-linear or raised sine type profile was possible dependent on the evaporant source position, with both showing a smooth transition down to what appeared to be a true zero thickness at the resolution and background noise level of the instrument (<5nm).
The performance of the taper system using the SU-8 intermediate waveguide and UHNA-3 fiber was then assessed using the full vector BPM mode of R-Soft BEAMPROP at 1550nm. Linear and raised sine type tapers were evaluated for both 4μm wide x 850nm high As2S3 rib waveguides with 420nm etch depth in TM mode (low dispersion, as many nonlinear demonstration used) and 550x550nm square As2S3 nanowires in both polarizations. Modeling indicated that for adequate taper lengths there was little performance difference between the two taper types. Figure 4 (a) shows the transfer efficiency relationship for a linear taper versus taper length for the ribs in TM mode and Fig. 4 (b) for the nanowires in both polarizations. The raised sine type profile typically required a longer length for a given excess loss, for example for the rib waveguide requiring a length of 350μm to 97% versus 250μm for the linear taper. The ribs typically achieved > 97% transfer efficiency for lengths of 300μm or more, the nanowires requiring 600μm or more due to the greater film thickness at cut off caused by their much narrower width. A length of 350μm was considered a good trade-off between low loss and minimum length for rib devices and 600μm for nanowire waveguides, and these lengths were selected for further modeling and the minimum acceptable length for experimentation.
Lateral offset sensitivity of the tapers was assessed for offsets of up to 1μm (the achievable worst case overlay error for full field 1x contact lithography, though <200nm is specified for the projection or stepper based lithography tools required to actually fabricate the nanowires). For the 4μm rib devices there was <0.1dB loss increase at 1μm offset, and the results for the nanowires are shown in Fig. 5 .
There is therefore sufficient lateral offset tolerance for both the rib and nanowire designs. Note the SU-8 waveguide width for the rib has to be increased in the vertical taper region to 6.3μm to accommodate the 1μm offset and then down tapered in a 200μm distance to 3.3μm for the fiber coupling, but this does not impact the coupling efficiency.
3. Fabrication and results
To check the performance of the intermediate SU-8 waveguides, 3.3x3.3μm waveguides were fabricated on 5μm thick thermally oxidized silicon wafers using SU-8-5 photoresist and its associated standard processing techniques . The waveguide top cladding was a UV curable polysiloxane with a refractive index of 1.535 at 1550nm (RPO Pty Ltd IPG or Microresist Ormocore ) which was spun on and UV cured. The propagation losses were measured by cut-back at 1550nm with butt coupling using UHNA-3 fiber, both with a tunable laser source and polarization controller for a spot measurement and the propagation loss spectrum via a supercontinuum source and optical spectrum analyzer.
The 1550nm cut back results yielded a loss of 1.55dB/cm with negligible polarization sensitivity. Figure 6 shows the measured propagation loss curve from the supercontinuum/OSA setup which indicates the losses observed at 1550nm originate solely from material absorption as there is no short wavelength loss characteristic of scattering mechanisms. The 1550nm loss is consistent with that obtained by other researchers [56, 57]. As noted previously, whilst this is not a very low loss, for short overlay waveguides (2-3mm) it is sufficiently low and lower loss materials such as SiOx or SiON [37, 38] have been demonstrated at process temperatures compatible with Chalcogenide materials.
Given the desire to use the waveguides in nonlinear experiments, the power handling characteristics of the SU-8 waveguides were also investigated. The output from a 1W EDFA at 1550nm was launched into the waveguides through lens tipped fibers in order to ensure there were no issues with index matching media. No damage was observed at estimated intra-core powers of up to 600mW, the maximum available coupled power.
Taper rib waveguide devices were then fabricated in 850nm thick As2S3 films on 100mm Silicon wafers with 5μm thermal oxide, with overlaid SU-8 waveguides as described above. The aperture mask used for the deposition had stepped lengths of 29, 24, and 19mm on it to enable the SU-8 waveguide loss, the As2S3 waveguide loss, and the length independent losses (sum of the taper excess loss and misalignment loss plus mode overlap losses) to be estimated by least squares fitting of the insertion loss data. The chip was hand cleaved with a diamond scriber to a total length of 4cm meaning the SU-8 waveguides were 11, 16, and 21mm respectively to the As2S3 waveguide lengths above. Post fabrication the tapers were determined to be 450μm long due to an error in the shadow mask. Figure 7(a) shows an image of part of the finished device in which the overlay waveguides are visible, and Fig. 7(b) shows an optical micrograph of the actual As2S3 taper region with the overlaid SU-8 waveguide.
Insertion loss measurements were then made for 4μm wide rib waveguides using UHNA-3 fiber spliced to SMF-28 and a 1550nm laser source. Index matching fluid was also applied to the end of the chip. Insertion loss data at the lowest loss input polarization state for the three lengths allowed the SU-8 waveguide loss, the As2S3 waveguide loss, and the total length independent losses to be determined by fitting the three unknowns to the data and minimizing the least square error from the experimentally determined points. Figure 8 presents the results.
Note that the loss increases as the As2S3 waveguide length decreases as the total chip length is fixed at 4cm and shortening the As2S3 waveguide means more SU-8 waveguide which has a higher propagation loss. The best least squares fit to the data was obtained with an As2S3 waveguide loss of 0.61dB/cm, SU-8 waveguide loss of 1.54dB/cm, and a 0.72dB length independent loss. The SU-8 waveguide loss is almost exactly that measured in isolation, and the As2S3 waveguide loss, whilst a little higher than normal (0.2-0.4dB/cm [e.g 29.]), is within the expected range. This then leaves the fixed loss to be accounted for. The SU-8 waveguides were inspected under an optical microscope and measured yielding dimensions of 2.5µm wide by 3.1µm high, smaller than the design dimensions of 3.3µm square. A calculation of the mode overlap loss for these waveguide dimensions to UHNA-3 fiber produced an overlap loss of 0.36dB/end, almost exactly matching the measured length independent loss. From this and observations using an InGaAs NIR camera which showed no light radiated from the taper region, it is concluded that the taper losses were negligible.
The polarization and wavelength sensitivity of the device were also examined. Measurements were made with light launched into the devices using the pure TE and TM polarization states. It is well known that polarization mode coupling can occur in rib waveguide devices of certain dimensions sometimes causes large insertion loss variations from coupling to higher order modes [58, 59] which manifests as apparent polarization dependent loss (PDL) that varies rapidly with wavelength. This was observed in these devices as well, with 14 of 26 waveguides tested exhibiting PDLs in the 0.7-2dB range when measured at 1550nm with a tunable laser and polarization controller, apparently randomly switching from TE as lower loss to TM due to the wavelength dependence and the use of a fixed wavelength narrowband source. The remaining 12 waveguides all had PDL below 0.4dB with random distribution of the lower loss state. Thus the taper coupling can be essentially polarization independent when the basic waveguide mode coupling effects are suppressed.
Figure 9 shows the measured taper chip transmission spectrum from 800 to 1650nm measured using a supercontinuum source and optical spectrum analyzer operated with 10nm resolution bandwidth to smooth out the mode coupling effects. The spectrum was normalized by subtracting the spectrum from the two UHNA-3 fibers butt coupled to each other. The total device insertion loss at 1550nm is 3.8dB for the 4cm long chip with 2.9cm of As2S3 waveguide and 1.1cm of SU-8 coupling waveguide. Factoring out the absorption dips at 1130, 1190, 1430, and 1680nm from the SU-8 (cf Fig. 6), a curve is left which is typical of the observed wavelength dependent loss in As2S3 waveguides comprising material Rayleigh scattering, waveguide sidewall roughness scattering, and any contribution from the band edge absorption or Urbach tail absorption at shorter wavelengths . Whilst it is difficult to explicitly factor out the contribution of any wavelength dependent loss in the taper structure, then examining the losses at 1600nm and 1270nm it is clear that the taper must be performing well over at least 330nm of optical bandwidth thereby demonstrating wide bandwidth operation.
In this realization the lowest fiber to As2S3 waveguide coupling loss was 1.1dB/end (half of 3.8dB minus the As2S3 waveguide loss of 2.9cm by 0.61dB/cm) comprising 0.36dB mode mismatch due to the smaller than expected SU-8 waveguides as noted above and 0.75dB due to absorption in the 5.5mm long SU-8 waveguides each end. The mode mismatch can be completely eliminated with correct fabrication. The SU-8 losses can be reduced either by replacing it with SiOx/SiON as noted above, or by shortening the SU-8 waveguide length to 1mm or less. The latter approach can only be achieved reliably by dicing the end facets on the waveguide, and so this was trialed to determine the achievable performance. The dicing was performed using a 2” diamond in nickel blade at 40,000rpm and a feed speed of 1mm/s. Fifteen waveguides on the previously fabricated SU-8 waveguide chip were first characterized after hand cleaving using UHNA-3 fiber coupling with index matching, then diced and recharacterized. The diced insertion losses were then corrected for the reduction in SU-8 length using the previously measured SU-8 propagation loss. Figure 10 presents the results.
The variance in the insertion loss measurements was reduced from 0.8dB to 0.4dB, and the insertion loss values post dicing are almost all at the level of the best hand cleaved results. Thus it is clear that dicing can be effectively and reliably applied to the SU-8 overlay waveguides, opening the way to reduce the SU-8 waveguide length to 1mm or less thereby further reducing the total taper system insertion loss to potentially <0.2dB/end.
Low loss vertically tapered thin film rib waveguide fiber couplers using a simple low process cost method have been demonstrated for the first time. The major advantages of this method are the low loss, wide bandwidth, polarization insensitive operation, fabricational simplicity and robustness to fabrication errors. With further process improvements fiber to waveguide coupling losses of <0.2dB/end appear realizable. The modeling results show that this method can also be easily applied to nanowire waveguides.
This research was supported by an Australian Research Council Centre of Excellence grant (project CE110001018) to the Centre for Ultrahigh bandwidth Devices for Optical Systems (CUDOS).
References and links
1. J. Van Erps, F. Luan, M. Pelusi, T. Iredale, S. Madden, D. Choi, D. Bulla, B. Luther-Davies, H. Thienpont, and B. Eggleton, “High-resolution optical sampling of 640-Gb/s data using Four-Wave mixing in dispersion-engineered highly nonlinear As2S3 planar waveguides,” J. Lightwave Technol. 28(2), 209–215 (2010). [CrossRef]
2. M. Lamont, V. Ta'eed, M. Roelens, D. Moss, B. Eggleton, D. Choi, S. Madden, and B. Luther-Davies, “Error-free wavelength conversion via cross-phase modulation in 5cm of As2S3 Chalcogenide glass rib waveguide,” Elec. Lett. 43(17), 945–947 (2007). [CrossRef]
3. M. Pelusi, F. Luan, S. Madden, D. Choi, D. Bulla, B. Luther-Davies, and B. Eggleton, “Wavelength conversion of high-speed phase and intensity modulated signals using a highly nonlinear Chalcogenide glass chip,” IEEE Photon. Technol. Lett. 22(1), 3–5 (2010). [CrossRef]
4. V. G. Ta’eed, M. R. Lamont, D. J. Moss, B. J. Eggleton, D. Y. Choi, S. Madden, and B. Luther-Davies, “All optical wavelength conversion via cross phase modulation in Chalcogenide glass rib waveguides,” Opt. Express 14(23), 11242–11247 (2006). [CrossRef] [PubMed]
5. J. Van Erps, J. Schröder, T. D. Vo, M. D. Pelusi, S. Madden, D. Y. Choi, D. A. Bulla, B. Luther-Davies, and B. J. Eggleton, “Automatic dispersion compensation for 1.28Tb/s OTDM signal transmission using photonic-chip-based dispersion monitoring,” Opt. Express 18(24), 25415–25421 (2010). [CrossRef] [PubMed]
6. V. Ta'eed, M. Pelusi, B. Eggleton, D. Choi, S. Madden, D. Bulla, and B. Luther-Davies, “Broadband wavelength conversion at 40 Gb/s using long serpentine As2S3 planar waveguides,” Opt. Exp. 15(23), 15047–15052 (2007). [CrossRef]
7. M. Pelusi, F. Luan, D. Choi, S. Madden, D. Bulla, B. Luther-Davies, and B. Eggleton, “Optical phase conjugation by an As2S3 glass planar waveguide for dispersion-free transmission of WDM-DPSK signals over fiber,” Opt. Exp. 18(25), 26686–26694 (2010). [CrossRef]
8. M. D. Pelusi, T. D. Vo, F. Luan, S. J. Madden, D. Y. Choi, D. A. Bulla, B. Luther-Davies, and B. J. Eggleton, “Terahertz bandwidth RF spectrum analysis of femtosecond pulses using a Chalcogenide chip,” Opt. Express 17(11), 9314–9322 (2009). [CrossRef] [PubMed]
9. R. Pant, C. Xiong, S. Madden, B. Davies, and B. Eggleton, “Investigation of all-optical analog-to-digital quantization using a Chalcogenide waveguide: A step towards on-chip analog-to-digital conversion,” Opt. Commun. 283(10), 2258–2262 (2010). [CrossRef]
10.. C. Xiong, G. Marshall, A. Peruzzo, M. Lobino, A. Clark, D. Choi, S. Madden, C. Natarajan, M. Tanner, R. Hadfield, S. Dorenbos, T. Zijlstra, V. Zwiller, M. Thompson, J. Rarity, M. Steel, B. Luther-Davies, B. Eggleton, and J. Obrien, “Generation of correlated photon pairs in a Chalcogenide As2S3 waveguide,” App. Phys. Lett. 98, 051101–051103 (2011).
11. E. Teraoka, D. Broaddus, T. Kita, A. Tsukazaki, M. Kawasaki, A. Gaeta, and H. Yamada, “Self-phase modulation at visible wavelengths in nonlinear ZnO channel waveguides,” Appl. Phys. Lett. 97(7), 071105 (2010). [CrossRef]
12. C. C. Evans, K. Shtyrkova, J. D. Bradley, E. Ippen, and E. Mazur, “Spectral broadening of femtosecond pulses in polycrystalline anatase titanium dioxide waveguides,” paper JW4D.4. in Nonlinear Photonics, OSA Technical Digest, Optical Society of America, 2012.
13. C. Evans, J. Bradley, O. Reshef, E. Marti-Panameño, and E. Mazur, “Ultrafast all-optical switching in TiO2,” Nano-Optics for Enhancing Light-Matter Interactions on a Molecular Scale, NATO Science for Peace and Security Series B: Physics and Biophysics, ISBN 978–94–007–5312–9. Springer Science + Business Media Dordrecht, 2013, p. 377.
14. W. Pernice, C. Xiong, C. Schuck, and H. Tang, “Second harmonic generation in phase matched aluminum nitride waveguides and micro-ring resonators,” Appl. Phys. Lett. 100(22), 223501 (2012). [CrossRef]
15. M. Ferrera, L. Razzari, D. Duchesne, R. Morandotti, Z. Yang, M. Liscidini, J. Sipe, S. Chu, B. Little, and D. Moss, “Low-power continuous-wave nonlinear optics in doped silica glass integrated waveguide structures,” Nat. Photonics 2(12), 737–740 (2008). [CrossRef]
16. A. Pasquazi, M. Peccianti, Y. Park, B. Little, S. Chu, R. Morandotti, J. Azaña, and D. Moss, “Sub-picosecond phase-sensitive optical pulse characterization on a chip,” Nat. Photonics 5(10), 618–623 (2011). [CrossRef]
17. D. Duchesne, M. Peccianti, M. R. Lamont, M. Ferrera, L. Razzari, F. Légaré, R. Morandotti, S. Chu, B. E. Little, and D. J. Moss, “Supercontinuum generation in a high index doped silica glass spiral waveguide,” Opt. Express 18(2), 923–930 (2010). [CrossRef] [PubMed]
18. J. Levy, A. Gondarenko, M. Foster, A. Turner-Foster, A. Gaeta, and M. Lipson, “CMOS-compatible multiple-wavelength oscillator for on-chip optical interconnects,” Nat. Photonics 4(1), 37–40 (2010). [CrossRef]
19. Y. Okawachi, K. Saha, J. S. Levy, Y. H. Wen, M. Lipson, and A. L. Gaeta, “Octave-spanning frequency comb generation in a silicon nitride chip,” Opt. Lett. 36(17), 3398–3400 (2011). [CrossRef] [PubMed]
20. T. Ning, O. Hyvärinen, H. Pietarinen, T. Kaplas, M. Kauranen, and G. Genty, “Third-harmonic UV generation in Silicon nitride nanostructures,” Opt. Exp. 21(2), 2012–2017 (2013). [CrossRef]
21. K. Saha, Y. Okawachi, B. Shim, J. Levy, R. Salem, A. Johnson, M. Foster, M. Lamont, M. Lipson, and A. Gaeta, “Modelocking and femtosecond pulse generation in chip-based frequency combs,” Opt. Exp. 21(1), 1335–1343 (2013). [CrossRef]
22. A. Subramanian, G. Murugan, M. Zervas, and J. Wilkinson, “High index contrast Er:Ta2O5 waveguide amplifier on oxidised silicon,” Opt. Commun. 285(2), 124–127 (2012). [CrossRef]
23. J. Bradley, L. Agazzi, D. Geskus, F. Ay, K. Wörhoff, and M. Pollnau, “Gain bandwidth of 80 nm and 2 dB/cm peak gain in Al2O3:Er3+ optical amplifiers on silicon,” J. Opt. Soc. Am. B 27(2), 187–196 (2010). [CrossRef]
24. R. Schermer, W. Berglund, C. Ford, R. Ramberg, and A. Gopinath, “Optical amplification at 1534nm in Erbium-doped Zirconia waveguides,” IEEE J. Quant. Elec. 39(1), 154–159 (2003). [CrossRef]
25. K. Vu and S. Madden, “Tellurium dioxide Erbium doped planar rib waveguide amplifiers with net gain and 2.8dB/cm internal gain,” Opt. Exp. 18(18), 19192–19200 (2010). [CrossRef]
26. C. Baker, J. Heikenfeld, Z. Yu, and A. Steckl, “Optical amplification and electroluminescence at 1.54μm in Er-doped Zinc silicate Germanate on Silicon,” Appl. Phys. Lett. 84(9), 1462–1464 (2004). [CrossRef]
27. X. Gai, S. Madden, D. Choi, D. Bulla, and B. Luther-Davies, “Dispersion engineered Ge11.5As24Se64.5 nanowires with a nonlinear parameter of 136W−1m−1 at 1550nm,” Opt. Exp. 18(18), 18866–18874 (2010). [CrossRef]
28. S. J. Madden and K. T. Vu, “Very low loss reactively ion etched Tellurium dioxide planar rib waveguides for linear and non-linear optics,” Opt. Express 17(20), 17645–17651 (2009). [CrossRef] [PubMed]
29. D. Choi, S. Madden, D. Bulla, R. Wang, A. Rode, and B. Luther-Davies, “Submicrometer-thick low-loss As2S3 planar waveguides for nonlinear optical devices,” IEEE Photon. Technol. Lett. 22(7), 495–497 (2010). [CrossRef]
30. J. Hu, N. N. Feng, N. Carlie, L. Petit, J. Wang, A. Agarwal, K. Richardson, and L. Kimerling, “Low-loss high-index-contrast planar waveguides with graded-index cladding layers,” Opt. Express 15(22), 14566–14572 (2007). [CrossRef] [PubMed]
33. B. Batagelj, “Conversion efficiency of fiber wavelength converter based on degenerate FWM,” in 2nd International Conference on Transparent Optical Networks, 2000 (2000), pp. 179–182.
34. I. Moerman, P. Van Daele, and P. Demeester, “A review on fabrication technologies for the monolithic integration of tapers with III-V semiconductor devices,” IEEE J. Sel. Top. Quantum Electron. 3(6), 1308–1320 (1997). [CrossRef]
35. L. Vivien, X. Le Roux, S. Laval, E. Cassan, and D. Marris-Morini, “Design, realization, and characterization of 3-D taper for fiber/micro-waveguide coupling,” IEEE J. Sel. Top. Quantum Electron. 12(6), 1354–1358 (2006). [CrossRef]
36. T. Tsuchizawa, K. Yamada, H. Fukuda, T. Watanabe, M. Jun-ichi Takahashi, T. Takahashi, E. Shoji, S. Tamechika, Itabashi, and H. Morita, “Microphotonics devices based on silicon microfabrication technology,” IEEE J. Sel. Top. Quantum Electron. 11(1), 232–240 (2005). [CrossRef]
37. T. Tsuchizawa, K. Yamada, T. Watanabe, H. Fukuda, H. Nishi, H. Shinojima, and S. Itabashi, “Spot-size converters for rib-type Silicon photonic wire waveguides” in 2008 5th IEEE International Conference on Group IV Photonics(2008), pp. 200–202 (2008).
38. B. Ben Bakir, A. V. de Gyves, R. Orobtchouk, P. Lyan, C. Porzier, A. Roman, and J.-M. Fedeli, “Low-loss (< 1 dB) and polarization-insensitive edge fiber couplers fabricated on 200-mm Silicon-on-insulator Wafers,” IEEE Photon. Technol. Lett. 22(11), 739–741 (2010). [CrossRef]
39. T. Shoji, T. Tsuchizawa, T. Watanabe, K. Yamada, and H. Morita, “Low loss mode size converter from 0.3 μm square Si wire waveguides to singlemode fibers,” Elec. Lett. 38(25), 1669–1670 (2002). [CrossRef]
42. T. Wahlbrink, W. Tsai, M. Waldow, M. Forst, J. Bolten, T. Mollenhauer, and H. Kurz, “Fabrication of high efficiency SOI taper structures,” Microelectron. Eng. 86(4-6), 1117–1119 (2009). [CrossRef]
43. P. K. Tien, R. J. Martin, and G. Smolinsky, “Formation of light-guiding interconnections in an integrated optical circuit by composite tapered-film coupling,” Appl. Opt. 12(8), 1909–1916 (1973). [CrossRef] [PubMed]
44. S. Sottini, D. Grando, L. Palchetti, and E. Giorgetti, “Optical fiber-polymer guide coupling by a tapered graded index glass guide,” IEEE J. Quant. Elec. 31(6), 1123–1130 (1995). [CrossRef]
45. F. Soares, F. Karouta, E. Geluk, J. van Zantvoort, H. de Waardt, R. Baets, and M. Smit, “Low loss InP based spot size converter based on a vertical taper,” in 12th European Conference on Integrated Optics (Grenoble, France, 2005), pp. 104–107.
46. R. Bellman, G. Bourdon, G. Alibert, A. Beguin, E. Guiot, L. Simpson, P. Lehuede, L. Guiziou, and E. LeGuen, “Ultralow loss high delta silica germania planar waveguides,” J. Electrochem. Soc. 151(8), G541–G547 (2004). [CrossRef]
49. M. Fadel, M. Bulters, M. Niemand, E. Voges, and P. Krummrich, “Low-loss and low-birefringence high-contrast Silicon-Oxynitride waveguides for optical communication,” J. Lightwave Technol. 27(6), 698–705 (2009). [CrossRef]
50. H. Hanafusa, M. Horiguchi, and J. Noda, “Thermally diffused expanded core fibers for low loss and inexpensive photonic components,” Elec. Lett. 27(21), 1968–1969 (1991). [CrossRef]
51. A. Martinez-Rios, I. Torres-Gomez, D. Monzon-Hernandez, O. Barbosa-Garcia, and V. Duran-Ramirez, “Reduction of splice loss between dissimilar fibers by tapering & fattening,” Rev. Mex. Fis. 56, 80–84 (2010).
54. D. Bulla, R. Wang, A. Prasad, A. Rode, S. Madden, and B. Luther-Davies, “On the properties and stability of thermally evaporated Ge-As-Se thin films,” Appl. Phys., A Mater. Sci. Process. 96(3), 615–625 (2009). [CrossRef]
56. M. Nordstrom, D. Zauner, A. Boisen, and J. Hubner, “Single-mode waveguides with SU-8 polymer core and cladding for MOEMS applications,” J. Lightwave Technol. 25(5), 1284–1289 (2007). [CrossRef]
57. B. Yang, L. Yang, R. Hu, Z. Sheng, D. Dai, Q. Liu, and S. He, “Fabrication and characterization of small optical ridge waveguides based on SU-8 polymer,” J. Lightwave Technol. 27(18), 4091–4096 (2009). [CrossRef]
58. S. Madden, D. Choi, A. Rode, and B. Luther-Davies, “Low loss etched Ge33As12Se55 Chalcogenide waveguides,” in Australian Conference on Optical Fiber Technology (2006), pp. 75–78.
60. T. Han, S. Madden, S. Debbarma, and B. Luther-Davies, “Improved method for hot embossing As2S3 waveguides employing a thermally stable Chalcogenide coating,” Opt. Express 19(25), 25447–25453 (2011). [CrossRef] [PubMed]