We demonstrate the fine tuning capability of femtosecond laser surface modification as a permanent trimming mechanism for silicon photonic components. Silicon microring resonators with a 15µm radius were irradiated with single 400nm wavelength laser pulses at varying fluences. Below the laser ablation threshold, surface amorphization of the crystalline silicon waveguides yielded a tuning rate of 20 ± 2 nm/J·cm−2 with a minimum resonance wavelength shift of 0.10nm. Above that threshold, ablation yielded a minimum resonance shift of −1.7nm. There was some increase in waveguide loss for both trimming mechanisms. We also demonstrated the application of the method by using it to permanently correct the resonance mismatch of a second-order microring filter.
©2013 Optical Society of America
Silicon-on-insulator (SOI) is an attractive platform for achieving very large scale integration of photonic integrated circuits (PICs) . This is due to silicon’s high index of refraction, relatively low loss at telecommunication wavelengths, and compatibility with the CMOS fabrication processes. Various photonic devices based on SOI technology have been demonstrated for applications in filtering , modulation , and sensing [4,5].
The high refractive index of silicon enables strong optical mode confinement to be achieved in the SOI waveguides. This permits device structures with small bending radii to be fabricated which allows extreme miniaturization. However, the small sizes of the waveguides combined with strong modal confinement make them very sensitive to variations in the refractive index and waveguide dimensions caused by fabrication imperfections. For example, a deviation of only 1nm in a typical silicon waveguide’s height or width will affect a change of ~0.002 in the waveguide’s effective index. This deviation from designed values can significantly alter the response of phase-sensitive photonic components. For an SOI microring resonator, an index change of 0.002 will lead to over 100 GHz shift in the resonance, which is unacceptable for most PIC applications.
Although SOI fabrication technology has advanced significantly, it is still not possible to control etched dimension variations and silicon film thickness uniformity to better than a 1nm accuracy across an entire wafer [6,7]. It is therefore necessary for most applications to employ a post-fabrication trimming technique to correct for fabrication-induced variations in silicon PICs and restore their proper device functionalities. The most common approach is to modify the effective index of a silicon waveguide by means of the thermo-optic effect. This is accomplished by placing resistive micro-heaters above silicon waveguides and controlling the power flow through each heater to selectively change the waveguides’ temperatures [8,9]. While this technique is very effective, it has the disadvantages of causing thermal cross-talk and requiring a continuous power supply to the heaters in addition to the extra processing steps required to fabricate them. The development of a permanent technique for tuning silicon PICs is actively being pursued to circumvent these issues.
Many permanent techniques for trimming silicon photonic devices have been demonstrated. Silicon microring resonators have been tuned by UV  or visible light  trimming of the cladding layer. Electron beam compaction of the oxide layer  or electron beam bleaching of a polymer cladding layer  have also been shown to be effective for tuning silicon microrings. Local oxidation of silicon has been used to trim photonic crystals using a laser-assisted process  and microring resonators using an atomic force microscope tip . There have also been trimming methods demonstrated in other material platforms that could be applied to silicon devices [16–18]. Most of these techniques, however, have disadvantages in that they are slow, require complex setups, or can provide only a limited tuning range.
We have recently demonstrated a new permanent technique for trimming silicon devices based on surface modification of silicon waveguides by an out-of-plane femtosecond laser pulse . Depending on the fluence, the laser shot can modify the silicon waveguide by converting some of the crystalline silicon near the waveguide surface to amorphous silicon or by nanomilling the top of the waveguide by ablation. This technique was shown to enable permanent bi-directional tuning of a microring resonator over a wide range. The method is also fast and the setup is relatively simple and readily adaptable to wafer-scale application.
In this work we demonstrate the ability of the technique to fine tune high quality (Q) factor microring resonators. The sharp resonances of the high Q microring notch filters allowed us to accurately monitor small changes in the effective index of the silicon waveguides along with any induced loss. We observed an approximately linear tuning rate in the amorphization regime with a minimum resonance wavelength shift of 0.10nm. We also demonstrated the application of the method to permanently correct the resonance mismatch of a second-order microring filter.
2. Fabrication and measurement of silicon microring resonators
Seven all-pass silicon microring resonators were fabricated on an SOI chip consisting of a 340nm crystalline Si (c-Si) on a 1µm-thick SiO2 buffer. Each microring had a 15µm radius and was coupled to a single bus waveguide via a coupling gap of 410nm. Both the bus and microring waveguides had a nominal width of 310nm. The devices were patterned using a Raith 150-TWO electron beam lithography system with ZEP520A resist, followed by dry etching. The devices were left air cladded. An SEM image of one of the fabricated microring resonators is shown in Fig. 1(a). The chip was cleaved to expose the bus waveguide facets for measurement, leaving a total waveguide length of about 3mm. Light in the 1500-1600nm wavelength range from a continuous-wave tunable laser was adjusted to the TM polarization and butt coupled on chip using a lensed fiber aligned to the waveguide facets. The transmitted light was then collected by another lensed fiber for detection. Figure 1(b) shows a typical spectral response of the fabricated microring resonators.
To extract the device parameters, we performed curve fitting of the measured data with the theoretical response of an all-pass microring resonator. We also accounted for the effects of reflections from the waveguide facets by modeling the device as microring resonator embedded inside a Fabry-Perot cavity. The transfer function of the system is given by
In the above equation τ is the transmission coefficient of the bus-to-ring coupling junction, ϕrt is the microring roundtrip phase and art is the roundtrip field attenuation factor. The inset of Fig. 1(b) shows a typical curve fit of one of the microring resonances. Normally the round-trip attenuation coefficient, art, and the transmission coefficient, τ, of an all-pass microring cannot be independently determined from its spectral response. However, since our microring is embedded in a Fabry-Perot cavity we are able to indirectly obtain phase information from the response of the ring and therefore separately determine the roundtrip attenuation and transmission coefficient when fitting the response with Eq. (1). The average values for the transmission coefficient, τ, and roundtrip field attenuation factor, art, were 0.972 and 0.986, respectively, giving an average loaded quality factor of 25,000 and intrinsic quality factor of 75,000 for the microrings.
3. Femtosecond laser tuning setup and results
Single femtosecond laser pulses at 400nm were used in this study to modify the silicon waveguides of the microrings. The 400nm pulses were frequency doubled in a BBO nonlinear crystal from ~130 fs FWHM pulses at 800nm produced by a Ti:sapphire laser. A 1mm thick BG39 filter was then placed in the beam path to remove the remaining 800nm light. As in our previous study , we chose to use a “top hat” beam profile to ensure a constant fluence across the waveguides. A “top hat” beam of 8.8µm radius was obtained by demagnified imaging of the center part of a Gaussian beam (8 mm beam waist diameter) selected with a pinhole aperture (500µm diameter) over which the intensity distribution was nearly uniform. The experimental setup is shown in Fig. 2(a), and the spatial profile of the “top hat” beam is in Fig. 2(b). While a reasonably flat profile was achieved, some residual intensity modulation on the order of 10 to 20% was observed at the edge of the spot due to diffraction. This intensity modulation increased if the device was placed in front of the best image plane or after it as seen in Fig. 2(c). The radius of the beam was about half that of the microrings, which allowed each microring to be shot twice on fresh sections of its waveguide.
Six microrings were used for the femtosecond laser tuning experiments while one was left unmodified as a control. After the initial characterization of the devices, each microring was shot with the laser once at varying fluences. The microrings were measured and characterized again, and then shot once more at different fluences on an unmodified section of the ring waveguide. Figure 3(a) shows a plot of the resonance wavelength shift as a function of the average laser fluence. The horizontal error bars are due to uncertainty in the energy calibration and the exact size of our laser spot. The vertical error bars are mainly due to our approximately ± 2µm accuracy of aiming the center of the laser spot onto the waveguide, resulting in some uncertainty in the length of ring waveguide affected by the shot. Figure 3(b) shows the change in the effective index of the section of waveguide amorphized by the laser shot. The effective index change was computed from the resonance wavelength shift by
Figure 3(a) shows that by changing the fluence of the laser shot it is possible to control the direction and extent of the wavelength shift. The positive shifts in the resonance wavelength at fluences below 0.2 J/cm2 are due to amorphization of a small amount of the crystalline silicon at the waveguide surface. This layer of amorphous silicon has a higher index of refraction than the original crystalline silicon, leading to an increase in the effective index of the waveguide. The negative shifts in the resonance wavelength at fluences above 0.2 J/cm2 are due to ablation of a small amount of material off the top of the waveguide. This causes a decrease in the waveguide height which results in a decrease in the effective index of the waveguide. The ~0.2 J/cm2 threshold for laser ablation of silicon at the 400nm wavelength is consistent with our previous work . In the ablation regime, the minimum negative shift in the resonance wavelength was −1.7nm, corresponding to a change in the waveguide effective index of −0.013. It was observed in our previous experiments that ablation only a few hundredths of a J/cm2 above the ablation threshold could be damaging to the waveguide resulting in significantly increased loss so laser shots at higher fluences were not attempted.
In our previous work , we determined the amorphization threshold to be 0.06 J/cm2 whereas here we observed amorphization at an average fluence of 0.04 J/cm2. This discrepancy is due to small fluctuations in the beam profile which caused the local intensity to be well above 0.04 J/cm2 leading to amorphization in some small regions near the edge of the beam spot. As seen in Fig. 2(c), the intensity fluctuations were even more pronounced if the device was not placed in the proper image plane. This effect was also observed visually through a change in colour for only small regions within the spot size, as seen through an optical microscope.
In the amorphization regime (laser fluence below 0.20 J/cm2 and above 0.06 J/cm2) there appears to be a linear correlation between the resonance wavelength shift and the laser fluence. The data point at 0.04 J/cm2 was not included in Fig. 3(b) because only a small portion of the laser spot caused amorphization and the length of waveguide changed is therefore unknown. Assuming a linear relationship in Fig. 3(b) we obtain a tuning rate of 20 ± 2 nm/J·cm−2 (or Δneff = 0.16 ± 0.02 per J/cm2 for the length of waveguide shot) above the amorphization threshold of 0.06 J/cm2.
There are three possible explanations for the linear correlation between the wavelength shift and laser fluence in the amorphization regime. The first hypothesis is that as the fluence increases, the depth of the amorphized silicon layer also increases. It is reasonable to assume that a higher laser fluence will cause a larger volume of crystalline silicon near the surface to melt, resulting in a thicker layer of amorphous silicon (the waveguide width being fixed). This hypothesis is supported by J. Bonse , who optically characterized the depth of amorphous silicon from an 800nm, femtosecond Gaussian laser pulse. Assuming that the refractive index of amorphous silicon does not depend on the laser fluence, it was concluded that this depth was likely dependent on the laser fluence used. The second possibility is that the volume of the amorphous silicon stays constant with changing laser fluences, but the index of refraction of the amorphous silicon is modified instead. This hypothesis is supported by the work of Izawa et. al [21,22], where it was shown using transmission electron microscopy (TEM) that the depth of the amorphous silicon produced using 400, 800, and 1550nm laser pulses did not vary with the number of shots or laser fluence below the ablation threshold (ref . reports the depth of amorphous silicon to be 17nm when irradiating crystalline silicon with a single 400nm femtosecond laser pulse). However, in a later publication, Izawa et al  also showed that for 266nm laser pulses the depth of amorphous silicon produced does vary over a Gaussian spot size. Although it is noted in many publications that the index of amorphous silicon is highly dependent on the manner in which it is produced [24,25], to our knowledge and also noted in ref , there has been no study of the change in the refractive index of amorphous silicon produced by femtosecond laser shots of various fluences. Thus further studies are needed to confirm the second hypothesis. The third option is a combination of increased amophization depth along with increased index of refraction at higher laser fluences, which is the most probable explanation given the evidence in the literature.
For this experiment, when the entire spot size was above the amorphization threshold, we achieved a minimum positive resonance wavelength shift of 0.71nm using a fluence of 0.06 J/cm2. This corresponds to a minimum change of 5.6 × 10−3 in the waveguide effective index. However, the data point at 0.04 J/cm2 shows the current tuning resolution of the technique. Due to the shorter affected length of waveguide, a shift in the resonance of 0.10nm was observed for this shot, which is roughly 1.5 × the bandwidth of these microrings. This shift cannot be converted to a change in the effective index because the length of waveguide affected is uncertain, but it illustrates that the shift is dependent on length of the waveguide affected. The resolution of the technique will be dependent on the ratio of the length of waveguide affected to the circumference of the microring, which is evident from Eq. (3). Therefore, the technique’s resolution can be made very fine by using either a large microring and/or a very small spot size. The tuning rate obtained in Fig. 3(b) can be modified in the same manner as the tuning resolution. The spot size of the laser pulse, along with the accuracy of the alignment stage, is the limiting factor for fine tuning with this technique.
Measurements of the roundtrip loss in the microring resonators also allowed us to investigate the effect of the laser shots on the waveguide loss. The initial roundtrip loss of the microring resonators averaged over four resonance peaks between 1540 and 1560nm was found to be 0.25 ± 0.05 dB. The large variation in the roundtrip loss is most likely due to coherent backscattering from the sidewall roughness of the microring waveguides [26,27], an effect which has been reported for silicon microrings very similar to ours . In Fig. 4 the change in the average roundtrip loss is plotted versus resonance wavelength shift induced by each shot. The error bars represent the standard deviation of the initial roundtrip loss measured for four resonance peaks between 1540 and 1560nm that is then added in quadrature with the same value for the final roundtrip loss. In the amorphization regime (positive resonance shifts), it is observed that the roundtrip loss increased with every laser shot, which is expected since amorphous silicon has larger absorption than crystalline silicon . However, there appears to be no definite correlation between the loss and the laser fluence and the error bars are fairly large and vary significantly in magnitude. The large error bars cross below zero for some data points, but no single measurement indicated a decrease in loss. The lack of a definite trend combined with the large fluctuations in the loss data suggests a dominant contribution from amplification of the coherent backscattering effect that was already present in the silicon waveguides before the laser shots. When a laser shot amorphizes some of the silicon, it may roughen the top surface of the waveguide and/or produce volumetric index inhomogeneities due to mixtures of amorphous and crystalline Si. This roughness could increase the backscattering and cause large fluctuations in the measured roundtrip loss at different resonances as observed. The effect could also be amplified for the TM polarization since it is more sensitive to roughness on the top surface of the waveguide than on the sidewalls. This hypothesis is supported by the much larger increase in loss and the larger fluctuations in its measurement for shots above the ablation threshold (negative shifts in the resonance wavelength in Fig. 4) together with evidence from AFM scans of unpatterned silicon surfaces after ablation showed an increased surface roughness .
In order to quantify the loss induced by femtosecond laser surface modification, microrings where coherent backscattering is insignificant should be investigated or a separate measurement of the backscattering is required in another study. The backscattering may also be reduced by using a smoother beam profile than the one shown in Fig. 2(b) or one with a more gradual change in fluence profile, such as a Gaussian laser spot. A further reduction in backscattering may also be possible by annealing the sample at high temperature .
4. Application of the laser tuning method to a second-order microring filter
One potential application of the technique is in correcting fabrication-induced resonance misalignment in high-order silicon microring filters. These devices consist of multiple coupled microring resonators whose resonances have to be accurately controlled for the proper operation of the device .
To demonstrate the ability of the femtosecond laser tuning technique to permanently correct the resonance mismatch in this type of device, we designed and fabricated a second-order microring filter consisting of two non-identical microring resonators with unmatched resonances. An SEM image of the device is shown in Fig. 5(a). The two microrings have radii R1 = 8µm and R2 = 12µm, giving individual free spectral range (FSR) of 9.5nm and 6.33nm, respectively. When two resonances of the microrings are aligned, the FSR of the device will increase to 19nm due to the Vernier effect with suppressed resonance peaks in between. The field coupling coefficients between the bus waveguides and the microrings (κ1, κ3) and between the two microrings (κ2) were designed to give a flat-top filter response. For a 100GHz-bandwidth filter the coupling values were determined to be κ1 = κ3 = 0.60, κ2 = 0.18, and the corresponding coupling gaps were g1 = 170 nm, g2 = 300 nm, g3 = 200 nm. (The coupling gap between ring 1 and the input bus waveguide is smaller than between ring 2 and the output bus waveguide because ring 1 has a smaller radius and hence shorter interaction length with the bus).
The top trace in Fig. 5(b) shows the as-fabricated spectral response of the device with the resonance peaks of each microring labelled. The 1535nm transmission peak is split due to a ~1.1nm detuning between the resonances of the two microrings. To align the two resonances, we increased the resonance wavelength of microring 1 by applying a single laser shot at 0.06 J/cm2 to a section of the ring as indicated in Fig. 5(a). The device spectral response after the laser shot is shown by the bottom trace in Fig. 5(b). The two microring resonances at the 1535nm wavelength are seen to merge to form a second-order filter response with a bandwidth of 112GHz and an insertion loss of ~3dB. Careful modeling of the spectral response (shown by the red curve and obtained by least squares fitting) revealed that microring 1’s resonance wavelength was now greater than microring 2’s resonance wavelength by ~0.3nm, indicating that the resonances of microring 1 were moved by a total of 1.4nm by the laser shot. Note that at the same fluence of 0.06 J/cm2, the resonance shift in the 8µm-radius ring is larger than the value recorded for the 15µm microring in Fig. 3. This is because a larger fraction of the microring roundtrip length was modified by the laser shot in the smaller ring. We also observed that the adjacent resonance peaks were much more suppressed below the main filter transmission peaks than in the original as-fabricated spectrum. The roundtrip loss for microring 1 before and after the femtosecond laser shot changed only slightly according to our modelling, increasing from 0.77dB to 0.86dB, while all other fitting parameters remained the same except for the resonance wavelength of microring 1. Our model showed that this increase in the loss of ring 1 accounts for only a 0.15dB change in the insertion loss of the device.
We have shown that femtosecond laser surface modification of silicon waveguides is a promising technique for tuning silicon microring resonators and other phase sensitive devices. In the amorphization regime, we observed an approximately linear relationship between the effective index change in the silicon waveguide and the fluence of the laser shot. The physical mechanism for this relationship is either an increase in the depth of amorphous silicon and/or an increase in the index of refraction of the amorphous silicon produced by the laser shots [20–25]. The propagation loss in the silicon waveguides, induced by either the conversion of some crystalline silicon to amorphous silicon or ablation of material, increased for all shots as expected, but no definite correlation to the laser fluence was observed due to effects from coherent backscattering in the microrings [26,27]. It may be possible to reduce the coherent backscattering effects by using a smoother beam profile or by annealing the device at a high temperature in future studies.
The minimum resonance wavelength shifts achieved in the experiment were 0.10nm and −1.7nm for amorphization and ablation respectively. These shifts can be made lower by moving the beam center off the waveguide or using a smaller spot size so that a smaller section of the ring is irradiated. We also demonstrated an application of the technique in tuning a second order microring filter. The resonance mismatch of two silicon microrings was corrected by a single femtosecond laser shot, proving that this technique can be a fast and reliable method of trimming high-order filters and other advanced PICs on a silicon platform.
This research was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) and Alberta Innovates. The authors would also like to thank Ashok M. Prabhu for assisting in the fabrication of the devices.
References and links
1. W. Bogaerts, P. De Heyn, T. Van Vaerenbergh, K. De Vos, S. K. Selvaraja, T. Claes, P. Dumon, P. Bienstman, D. Van Thourhout, and R. Baets, “Silicon microring resonators,” Laser Photon. Rev. 6(1), 47–73 (2012). [CrossRef]
2. Y. Vlasov, W. M. J. Green, and F. Xia, “High-throughput silicon nanophotonic wavelength-insensitive switch for on-chip optical networks,” Nat. Photonics 2(4), 242–246 (2008). [CrossRef]
4. S. Yamatogi, Y. Amemiya, T. Ikeda, A. Kuroda, and S. Yokoyama, “Si ring optical resonators for integrated on-chip biosensing,” Jpn. J. Appl. Phys. 48(4), 04C188 (2009). [CrossRef]
7. A. V. Krishnamoorthy, X. Zheng, G. Li, J. Yao, T. Pinguet, A. Mekis, H. Thacker, I. Shubin, Y. Luo, K. Raj, and J. E. Cunningham, “Exploiting CMOS manufacturing to reduce tuning requirements for resonant optical devices,” IEEE Photon. J. 3(3), 567–579 (2011). [CrossRef]
8. P. Dong, W. Qian, H. Liang, R. Shafiiha, N. N. Feng, D. Feng, X. Zheng, A. V. Krishnamoorthy, and M. Asghari, “Low power and compact reconfigurable multiplexing devices based on silicon microring resonators,” Opt. Express 18(10), 9852–9858 (2010). [CrossRef] [PubMed]
9. H. Y. Ng, M. R. Wang, D. Li, X. Wang, J. Martinez, R. R. Panepucci, and K. Pathak, “4×4 wavelength-reconfigurable photonic switch based on thermally tuned silicon microring resonators,” Opt. Eng. 47, 044601 (2008). [CrossRef]
10. L. Zhou, K. Okamoto, and S. J. B. Yoo, “Athermalizing and trimming of slotted silicon microring resonators with UV-sensitive PMMA upper-cladding,” IEEE Photon. Technol. Lett. 21(17), 1175–1177 (2009). [CrossRef]
11. A. Canciamilla, F. Morichetti, S. Grillanda, P. Velha, M. Sorel, V. Singh, A. Agarwal, L. C. Kimerling, and A. Melloni, “Photo-induced trimming of chalcogenide-assisted silicon waveguides,” Opt. Express 20(14), 15807–15817 (2012). [CrossRef] [PubMed]
13. S. Prorok, A. Y. Petrov, M. Eich, J. Luo, and A. K. Y. Jen, “Trimming of high-Q-factor silicon ring resonators by electron beam bleaching,” Opt. Lett. 37(15), 3114–3116 (2012). [CrossRef] [PubMed]
14. C. J. Chen, J. Zheng, T. Gu, J. F. McMillan, M. Yu, G. Q. Lo, D. L. Kwong, and C. W. Wong, “Selective tuning of high-Q silicon photonic crystal nanocavities via laser-assisted local oxidation,” Opt. Express 19(13), 12480–12489 (2011). [CrossRef] [PubMed]
15. Y. Shen, I. B. Divliansky, D. N. Basov, and S. Mookherjea, “Electric-field-driven nano-oxidation trimming of silicon microrings and interferometers,” Opt. Lett. 36(14), 2668–2670 (2011). [CrossRef] [PubMed]
16. S. T. Chu, W. Pan, S. Sato, T. Kaneko, B. E. Little, and Y. Kokubun, “Wavelength trimming of a microring resonator filter by means of a UV sensitive polymer overlay,” IEEE Photon. Technol. Lett. 11(6), 688–690 (1999). [CrossRef]
17. D. K. Sparacin, C. Y. Hong, L. C. Kimerling, J. Michel, J. P. Lock, and K. K. Gleason, “Trimming of microring resonators by photooxidation of a plasma-polymerized organosilane cladding material,” Opt. Lett. 30(17), 2251–2253 (2005). [CrossRef] [PubMed]
18. U. Levy, K. Campbell, A. Groisman, S. Mookherjea, and Y. Fainman, “On-chip microfluidic tuning of an optical microring resonator,” Appl. Phys. Lett. 88(11), 111107 (2006). [CrossRef]
20. J. Bonse, “All-optical characterization of single femtosecond laser-pulse-induced amorphization in silicon,” Appl. Phys., A Mater. Sci. Process. 84(1-2), 63–66 (2006). [CrossRef]
21. Y. Izawa, Y. Izawa, Y. Setsuhara, M. Hashida, M. Fujita, R. Sasaki, H. Nagai, and M. Yoshida, “Ultrathin amorphous Si layer formation by femtosecond laser pulse irradiation,” Appl. Phys. Lett. 90(4), 044107 (2007). [CrossRef]
22. Y. Izawa, Y. Setuhara, M. Hashida, M. Fujita, and Y. Izawa, “Abaltion and amorphization of crystalline Si by femtosecond and picosecond laser irradiation,” Jpn. J. Appl. Phys. 45(7), 5791–5794 (2006). [CrossRef]
23. Y. Izawa, S. Tokita, M. Fujita, M. Nakai, T. Norimatsu, and Y. Izawa, “Ultrathin amorphization of single-crystal silicon by ultraviolet femtosecond laser pulse irradiation,” J. Appl. Phys. 105(6), 064909 (2009). [CrossRef]
24. W. J. Liu, S. Chen, H. Y. Cheng, J. D. Lin, and S. L. Fu, “Fabrication of amorphous silicon films for arrayed waveguide grating application,” Surf. Coat. Tech. 201(15), 6581–6584 (2007). [CrossRef]
25. M. J. A. de Dood, A. Polman, T. Zijlstra, and E. W. J. M. van der Drift, “Amorphous silicon waveguides for microphotonics,” J. Appl. Phys. 92(2), 649–653 (2002). [CrossRef]
27. F. Morichetti, A. Canciamilla, M. Martinelli, A. Samarelli, R. M. De La Rue, M. Sorel, and A. Melloni, “Coherent backscattering in optical microring resonators,” Appl. Phys. Lett. 96(8), 081112 (2010). [CrossRef]
29. M. S. Dahlem, C. W. Holzwarth, A. Khilo, F. X. Kärtner, H. I. Smith, and E. P. Ippen, “Reconfigurable multi-channel second-order silicon microring-resonator filterbanks for on-chip WDM systems,” Opt. Express 19(1), 306–316 (2011). [CrossRef] [PubMed]