We present detailed characterization of a unique high-index-contrast integrated optical polymer waveguide platform where the index of the cladding material is closely matched to that of water. Single-mode waveguides designed to operate across a large part of the visible spectrum have been fabricated and waveguide properties, including mode size, bend loss and evanescent coupling have been modeled using effective-index approximation, finite-element and finite-difference time domain methods. Integrated components such as directional couplers for wavelength splitting and ring resonators for refractive-index or temperature sensing have been modeled, fabricated and characterized. The waveguide platform described here is applicable to a wide range of biophotonic applications relying on evanescent-wave sensing or excitation, offering a high level of integration and functionality. The technology is biocompatible and suitable for wafer-level mass production.
©2010 Optical Society of America
Integrated optical components are of particular interest for biophotonic applications due to the possibility of integrating multiple functions on a single chip . They are readily compatible with evanescent-wave sensing principles that have been employed in planar waveguides as well as stripped optical fibers and photonic crystal fibers [2–7]. Polymers are attractive for fabricating planar integrated optical circuits as they offer lower costs than devices based on inorganic materials or semiconductors while maintaining the possibility of large scale integration . Using standard microfabrication techniques, polymer optical components can be patterned and directly integrated with electrical controls. In optical waveguides, a large refractive index contrast between the waveguide core material and the cladding material is crucial for enabling strong confinement and large-scale integration of optical components. For sensing purposes, a high index contrast in single-mode waveguides also means that a higher sensing field intensity can be obtained for a given excitation power .
In the present work, we have studied a polymer optical waveguide platform with a large index contrast (as compared to most polymer waveguide systems) suitable for operation in the visible wavelength range. This platform has the additional advantage that the refractive index of the cladding material closely matches that of water, making it particularly suitable for biophotonic applications. The ability to tune the properties of polymers through surface modification or the addition of functional groups to the polymer backbone makes them appealing as waveguiding materials for biophotonics. We have selected polymethylmethacrylate (PMMA) as a waveguide core material, although a wide range of other optical polymer materials are also suitable for this purpose. PMMA is a well-studied high-resolution electron beam lithography resist  and a highly transparent optical polymer, making it ideal as a core layer material for fabricating optical components based on single-mode waveguides. Since electron-beam lithography is well suited for research and prototyping but not for mass-production of devices, we note that PMMA can also be patterned using nanoimprint lithography , which has proven useful for fabricating components at high throughput with high resolution . The low autofluorescence of PMMA  makes it well-suited for applications involving fluorescence excitation and detection. UV/ozone treatment can be used to increase the surface energy of PMMA and hence its bonding to other polymeric materials, making it particularly suitable for microfluidic applications [13,14].
The low refractive index (n ~1.34) of the fluorinated optical polymer Cytop renders it especially useful as a cladding material for evanescent-wave sensing of biological samples (typically n ~1.33 – 1.35). When used in conjunction with PMMA (n ~1.49) an index contrast of approximately 0.15 (10%) is obtained. This contrast is substantially higher than in conventional glass or polymer-based dielectric waveguides platforms used for integrated optical components . Although substantially higher index contrast can be obtained in commonly used inorganic waveguide materials like silicon-on-insulator (SOI), Si3N4, or TiO2, such materials are only compatible with hard patterning techniques (and in the case of SOI only infrared wavelengths).
Air-clad PMMA ridges on a Cytop-coated substrate have been used by Poon et al. for the fabrication of coupled resonator optical waveguides operating at telecom wavelengths around 1550 nm [15,16]. Although the present paper focuses on visible-light applications, the waveguide geometry is easily scalable to infrared wavelengths commonly used in biosensing (e.g. 785 nm or 850 nm). The symmetric Cytop/PMMA/biological sample configuration has previously been used by our group for waveguide-excitation fluorescence microscopy  and in the current paper we demonstrate how this platform can be extended to obtain full on-chip control over the excitation light – simultaneously introducing new possibilities for other evanescent-wave sensing schemes in biophotonics.
2. Fabrication and experimental methods
Cleanroom fabrication of single-mode waveguides was carried out in the following fashion. A 4″ silicon wafer was initially primed with adhesion promoter (AP3000, DOW Chemical Co.) and then spin-coated with a 4-µm layer of Cytop (CTX-809AP2, Asahi Glass Co.) to form the lower cladding layer. The Cytop surface was treated using Al deposition and removal, as described in Ref , and PMMA was subsequently spin-coated onto the Cytop to form a waveguiding layer with a typical thickness of 450-500 nm. A 50-nm layer of aluminum was deposited on the surface for charge dissipation during subsequent electron-beam writing. Single-mode waveguides were formed in the PMMA layer by exposing two 1.8-µm wide strips on each side of the 500-600 nm wide waveguide channels. All structures were written within a single writefield of 1 mm × 1 mm with a raster grid of 31 nm.
Following e-beam lithography, the charge dissipation layer was removed by wet etching and the exposed PMMA was developed in a methylisobutylketone/isopropyl alcohol (IPA) mixture (1:3 ratio) for 30 seconds and then rinsed in pure IPA for 30 seconds. Finally, the top Cytop cladding layer was spin-coated onto the wafer and baked in a cleanroom oven, producing a top cladding layer with a thickness of approximately 4 µm (see Ref . for a more detailed process description). A layer of UV photoresist was added to protect the surface of the sample during dicing. Samples were diced at the edge of the waveguide structures to enable end-fire coupling through the end facets.
Unless otherwise noted, waveguide characterization was carried out by direct end-fire excitation using light from a supercontinuum source (SuperK Versa, NKT Photonics) equipped with an acousto-optic tunable filter, as shown in Fig. 1 . A tapered lensed optical fiber (SM488nm, Nanonics Imaging) with a minimum spot size specified as 0.7 ± 0.2 µm (at 532 nm wavelength) was used for in-coupling. The fiber was taped to the optical setup to prevent movement and the polarization at the output of the fiber was adjusted using free-space polarization control of the SuperK beam. The output of the PMMA waveguides was analyzed either by coupling to a second single-mode fiber (not tapered) connected to a spectrometer (PI Acton SpectraPro SP-2356) equipped with a cooled CCD camera (PIXIS:100F, Princeton Instruments) or by imaging the output facet onto a CCD chip (Thorlabs) with a 100 × 0.9NA objective lens. Scattered light from the waveguides was also monitored from above using a microscope and recorded with a low-light level EM-CCD camera (QuantEM512C, Photometrics) or a color camera (Nikon DS-2Mv).
Numerical simulations were carried out using a mode solver based on the effective index approximation , as well as commercial finite-element (COMSOL, RF-module) and finite difference-time domain (FDTD solutions, Lumerical) software packages. Calculations were generally carried out for three wavelengths approximately equally spaced in frequency and commonly used in fluorescence applications (488 nm, 546 nm and 633 nm). The material dispersion of PMMA and Cytop was taken into account in all cases, using refractive index data provided by the respective manufacturers.
3. Basic waveguide properties
Cytop/PMMA waveguides with square cross-sectional profiles were modeled in order to determine the optimum core size with respect to range of operating wavelengths, mode confinement, evanescent field strength and bend loss.
3.1 Mode structure
Figure 2a shows the mode index (effective refractive index) of TM modes in the PMMA/Cytop waveguides, as determined by the effective index approximation (EIA). Although this method is not expected to yield highly accurate results for high index contrast waveguides  it can still provide a reasonable estimate of the range of single-mode operation (for the fundamental mode, more accurate values of neff determined by finite-element calculations using a finite cladding thickness are also shown). Similar results are obtained for TE modes (not shown).
From Fig. 2a it can be seen that single-mode operation with suitable mode confinement across the investigated wavelength range is possible (for a given polarization) for core sizes in the range 300 nm to 500 nm. Figure 2b shows how the mode field diameter (MFD, full width at 1/e2 intensity) varies with the core size, indicating a maximum confinement (minimum MFD) for core sizes of 350 nm (blue) to 450 nm (red). A waveguide core size of 450 nm to 550 nm provides a good mode match to the focused spot of the tapered input fiber. Evanescent field penetration depths (distance of 1/e2 intensity decay from the waveguide surface) are also plotted in Fig. 2b, where a uniform cladding index of 1.34 has been assumed, giving a penetration depth of about half a wavelength into the cladding material. A typical biological sample may have a different (and spatially varying) refractive index, correspondingly affecting the actual penetration depth, e.g., decreasing by 5% in pure water at n = 1.333 but increasing up to 20% for cytoplasm at n = 1.36 . A change in temperature will also affect the refractive index of the polymer and the sample solution. Within a typical operating range of a biophotonic chip (20-40°C), the refractive index change is small (0.1-0.2%) and the effect on the index contrast is negligible. A temperature change can, however, be monitored using resonant wavelength detection, as noted below, or used for thermo-optic switching or modulation of the guided signal.
Figure 3a shows the waveguide output measured at a wavelength of 570 nm. After correcting for the resolution of the imaging optics, the experimentally determined mode width is in excellent agreement with finite-element calculations. The waveguide mode is highly confined compared to, e.g., a conventional single-mode optical fiber, as shown in the figure.
The electron beam dosage used during lithography was optimized so that waveguides with smooth sidewalls were obtained (Fig. 3b). The structural quality of the waveguides ensures low scattering loss; we measured propagation losses down to approximately 0.3 dB/mm for straight waveguides at a wavelength of 550 nm.
3.2 Bend loss
A critical issue for enabling a high level of optical component integration is the reduction of bend loss. Fundamental considerations reveal that waveguide bend loss is associated with the extent of the evanescent tail of the mode in the waveguide plane  and, as demonstrated below, minimum bend loss does not necessarily coincide with maximum field confinement (as measured in terms of MFD). In order to estimate expected bend loss in single-mode Cytop/PMMA waveguides, we carried out 2D FDTD simulations of transmission through 90° circular waveguide bends of varying radii for blue, green and red wavelengths. The calculations include the loss introduced by coupling between straight and bent sections and the ends of the curved waveguide. As shown in Fig. 4a , bend loss increases rapidly for bend radii below 20 µm (for 500 nm core size). As expected, the bend loss is larger for longer wavelengths
Calculations for different core sizes (Fig. 4b) indicate that minimum bend loss occurs at core sizes of 500-600 nm, which is significantly larger than the core size corresponding to a maximum field confinement. Above 500-nm core size, however, sharp bends will cause excitation of higher order modes (cf. Figure 2a), particularly for shorter wavelengths, posing a limit to the achievable bend loss over a wide wavelength range.
The above simulations suggest that a PMMA square core size of 450-500 nm provides a good tradeoff, ensuring efficient end-fire excitation with a tapered fiber, single-mode operation for the investigated wavelength range, reasonable extent of the evanescent field, lowest possible bend loss and good structural quality ensuring low scattering loss. The experimentally fabricated waveguides were slightly asymmetric with waveguide widths ranging from 500 to 600 nm. Fabricated devices with sharp bends were, however, studied mainly at wavelengths ≥ 546 nm, where higher order modes do not play a significant role.
4. Integrated optical components
Different types of optical components were fabricated using Cytop/PMMA waveguides, ranging from basic y-junctions and s-bends to Mach-Zehnder interferometers, directional couplers (DCs) and ring resonators (RRs). Properties of fabricated DCs and RRs, as well as corresponding numerical simulations, are discussed below.
4.1 Directional couplers
Simple parallel-waveguide directional couplers can be used as precise power splitters, but only in a narrow range of wavelengths [22,23]. Conversely, the fact that the DC coupling length Lc (corresponding to full power transfer from one waveguide to another) is, in general, strongly wavelength dependent, can be used to realize compact and efficient wavelength splitters for two or more suitably spaced wavelengths.
For a Cytop/PMMA directional coupler with a 200-nm coupling gap, finite-element simulations indicate that the coupling length changes substantially (by a factor of about 7) across the visible range (400 nm to 700 nm). The coupling gap and the coupler length can be modified to realize separation of specific wavelengths λ1 and λ2 with the shortest possible DC fulfilling the condition for the interaction length L = Lc(λ1) = 2 Lc(λ2), with Lc(λ) = λ/(2ΔnDC), where ΔnDC(λ) is the difference between the effective refractive indices of the symmetric and antisymmetric modes of the coupled waveguides. A three-wavelength splitter can similarly be realized using a pair of DCs in sequence, fulfilling the conditions L1 = Lc1(λ1) = 2 Lc1(λ2) = 3 Lc1(λ3) and L2 = Lc2(λ1) = 2 Lc2(λ3), respectively.
As an example, Fig. 5a shows an SEM image of a fabricated DC with a nominal gap width of 200 nm and a parallel interaction length of 40 µm, with cosine bends on both ends (100 µm length, 30 µm offset). The output in the direct and coupled arms of the DC was measured at the sample facet for the wavelength range 450-700 nm and the sum of the outputs was normalized to 100% (Fig. 5c). A comparison of the measured data and the simulated values of ΔnDC for parallel waveguides indicated an effective coupling length of 54 µm for this particular device, suggesting that significant coupling also takes place in the initial part of the bend regions. The measured results are in good agreement with FDTD calculations which also confirm a significant deterioration of waveguide performance past 650 nm. The color image in Fig. 5b shows how the DC efficiently splits a two-color input into separate arms. The cross-talk measured for this particular device was approximately −17 dB for the shorter wavelength and −13 dB for the longer wavelength.
The possibility of simultaneously transmitting multiple wavelengths onto the chip through a single optical path, combined with on-chip filtering and signal modulation of the individual waveguide channels is ideal for, e.g., multi-wavelength time-lapse imaging of live cells or on-chip integrated waveguide excitation and multi-wavelength fluorescence monitoring as demonstrated by Dongre et al. for fluorescently labeled DNA molecules separated by capillary electrophoresis .
4.2 Ring resonators
Simple ring resonator structures are composed of one or more straight waveguides evanescently-coupled to a circular waveguide . RRs have been used in many photonic applications due to their versatility and compactness, e.g., as filters [26,27], optical switches [25,28] and laser resonators [29–31]. Light travelling in the ring experiences constructive interference when the resonance condition 2πR neff(R) = λm is met, where m is an integer, R is the ring radius and neff(R) is the mode index of the curved waveguide .
The device geometry used for RR characterization is shown in Fig. 6a , with a close-up of the coupling region of a fabricated structure shown in Fig. 6b. Different ring diameters were studied, ranging from 20 µm to 60 µm, and the separation gap between the straight waveguides and the ring was varied in order to tune the evanescent coupling. The structure shown in Fig. 6b was fabricated with a nominal coupling gap of 200 nm and a ring diameter of 40 µm. As shown in the figure, the ring and the straight waveguide are distinctly separated in the coupling region although minor proximity exposure effects associated with electron beam scattering can be observed, resulting in a slight broadening of the waveguides in this region.
The spectral characteristics of the RR were measured at both the through port and the drop port, with the in-coupling region being laterally displaced from the through port output by 180 µm to prevent residual stray light from the incoupling region from affecting the measured output (the pickup fiber has a mode field diameter much larger than the waveguide and is therefore prone to picking up scattered light propagating in the cladding). The radius of all bends in the input and output waveguides was 30 µm.
Ring resonators were characterized by combining several closely spaced AOTF channels to create a continuous input spectrum with a spectral width of about 10 nm. Typical output spectra observed in the drop and through channels of a 40-μm diameter ring are shown in Fig. 7a . The distance between the resonance peaks, corresponding to the free spectral range (FSR) in frequency units, was measured for different ring diameters, as plotted in Fig. 7b. The measured FSR accurately follows the expected 1/R dependence, assuming a constant mode index of 1.535 for the range of ring diameters investigated here. 2D FDTD calculations with no free fitting parameters reproduced the experimentally determined value of the FSR to within the experimental uncertainty.
The finesse of the measured RRs was in the range 3-4. In order to understand the origin of these values, it is instructive to consider the equation for the power transmission for the through port of a RR coupled to two waveguides, given by Fig. 4 can be used to provide a maximum theoretical value of the round-trip loss (since the straight-to-bent waveguide coupling loss included in Fig. 4 does not apply here), yielding a minimum field transmission coefficient of α ≈0.95 for a 40-µm diameter ring. The measured finesse in the resonators studied here of typically 3-4 corresponds to a Q factor of about 103. Assuming the above value for α, this requires |tj| ≈0.7 in Eq. (1) for the transmission coefficients in the coupling regions, assuming that the coefficients are similar for both regions. Increasing the coupling gap from 200 nm to 300 nm, however, did not significantly increase the measured finesse, suggesting that extra scattering loss is taking place in the ring or in the coupling regions. In order to realize high-finesse RRs, the extraneous round trip loss must be reduced and a correspondingly weak coupling realized, with |t| ≈α. In this case, it should be possible to realize Q factors in excess of 104 for RRs with 30 µm radius. For sensing purposes, however, it is also important to adjust the RR extinction ratio to the optimum value, given a certain Q-factor, as explained in detail in Ref .
A small change in the refractive index of the waveguide cladding results in a change of the effective refractive index of the circular waveguide, which may be detected as a shift in the resonance wavelength at the waveguide output. This property of ring resonators has been exploited to provide label-free detection in biosensing applications [34,35]. Similarly, embedded ring resonators may be used for on-chip temperature monitoring due to the thermo-optic effect. In the former case, cascaded or coupled non-resonant embedded RRs may be used to obtain an optical readout (on/off or position-dependent) directly on the chip, eliminating the need for off-chip spectral analysis.
We have described an integrated high-index contrast polymer waveguide platform, suitable for applications in biophotonics. The high core/cladding index contrast and small core size ensures that the guided mode is strongly confined, making light manipulation at a small scale possible. Optimization based on numerical simulations has been used to ensure single mode operation across most of the visible range and determine the core size required for optimum confinement and minimum bend loss. Electron beam lithography was used to fabricate optical waveguides with smooth sidewalls and low scattering loss. Directional couplers for wavelength separation were found to operate in good agreement with numerical modeling and ring resonators with a finesse of 3-4 and Q-factor of 103 were demonstrated. The waveguide platform can be envisioned as a particularly interesting candidate for evanescent-wave multi-wavelength live-cell time-lapse imaging with integrated temperature control and monitoring, or other applications involving spatial, spectral and temporal control of evanescent-wave excitation.
References and links
1. P. N. Prasad, Introduction to Biophotonics (Wiley-Interscience, Hoboken, NJ, 2003).
2. F. Prieto, B. Sepúlveda, A. Calle, A. Llobera, C. Domínguez, A. Abad, A. Montoya, and L. M. Lechuga, “An integrated optical interferometric nanodevice based on silicon technology for biosensor applications,” Nanotechnology 14(8), 907–912 (2003). [CrossRef]
3. A. Leung, P. M. Shankar, and R. Mutharasan, “A review of fiber-optic biosensors,” Sens. Actuators B Chem. 125(2), 688–703 (2007). [CrossRef]
4. J. B. Jensen, L. H. Pedersen, P. E. Hoiby, L. B. Nielsen, T. P. Hansen, J. R. Folkenberg, J. Riishede, D. Noordegraaf, K. Nielsen, A. Carlsen, and A. Bjarklev, “Photonic crystal fiber based evanescent-wave sensor for detection of biomolecules in aqueous solutions,” Opt. Lett. 29(17), 1974–1976 (2004). [CrossRef] [PubMed]
5. G. Pandraud, T. M. Koster, C. Gui, M. Dijkstra, A. Van Den Berg, and P. V. Lambeck, “Evanescent wave sensing: New features for detection in small volumes,” Sens. Actuators A Phys. 85(1-3), 158–162 (2000). [CrossRef]
6. M. K. Khaing Oo, Y. Han, J. Kanka, S. Sukhishvili, and H. Du, “Structure fits the purpose: photonic crystal fibers for evanescent-field surface-enhanced Raman spectroscopy,” Opt. Lett. 35(4), 466–468 (2010). [CrossRef] [PubMed]
8. H. Ma, A. K. Y. Jen, and L. R. Dalton, “Polymer-Based Optical Waveguides: Materials, Processing, and Devices,” Adv. Mater. 14(19), 1339–1365 (2002). [CrossRef]
9. C. Vieu, F. Carcenac, A. Pépin, Y. Chen, M. Mejias, A. Lebib, L. Manin-Ferlazzo, L. Couraud, and H. Launois, “Electron beam lithography: Resolution limits and applications,” Appl. Surf. Sci. 164(1-4), 111–117 (2000). [CrossRef]
10. S. Y. Chou, P. R. Krauss, and P. J. Renstrom, “Nanoimprint lithography,” J. Vac. Sci. Technol. B 14(6), 4129–4133 (1996). [CrossRef]
11. L. J. Guo, “Recent progress in nanoimprint technology and its applications,” J. Phys. D 37(11), R123–R141 (2004). [CrossRef]
12. A. Piruska, I. Nikcevic, S. H. Lee, C. Ahn, W. R. Heineman, P. A. Limbach, and C. J. Seliskar, “The autofluorescence of plastic materials and chips measured under laser irradiation,” Lab Chip 5(12), 1348–1354 (2005). [CrossRef] [PubMed]
13. J. J. Shah, J. Geist, L. E. Locascio, M. Gaitan, M. V. Rao, and W. N. Vreeland, “Surface modification of poly(methyl methacrylate) for improved adsorption of wall coating polymers for microchip electrophoresis,” Electrophoresis 27(19), 3788–3796 (2006). [CrossRef] [PubMed]
14. C. W. Tsao, L. Hromada, J. Liu, P. Kumar, and D. L. DeVoe, “Low temperature bonding of PMMA and COC microfluidic substrates using UV/ozone surface treatment,” Lab Chip 7(4), 499–505 (2007). [CrossRef] [PubMed]
16. J. K. S. Poon, L. Zhu, G. A. DeRose, and A. Yariv, “Polymer microring coupled-resonator optical waveguides,” J. Lightwave Technol. 24(4), 1843–1849 (2006). [CrossRef]
18. B. Agnarsson, J. Halldorsson, N. Arnfinnsdottir, S. Ingthorsson, T. Gudjonsson, and K. Leosson, “Fabrication of planar polymer waveguides for evanescent-wave sensing in aqueous environments,” Microelectron. Eng. 87(1), 56–61 (2010). [CrossRef]
19. M. Hammer, 2-D multilayer waveguide mode solver effective index approximation, http://wwwhome.math.utwente.nl/~hammer/eims.html.
20. L. A. Coldren, and S. W. Corzine, Diode Lasers and Photonic Integrated Circuits (Wiley-Interscience 1995).
22. E. A. J. Marcatili, “Dielectric rectangular waveguide and directional coupler for integrated optics,” Bell Syst. Tech. J. 48, 2071–2102 (1969).
23. A. Yariv, “Coupled-mode theory for guided-wave optics,” IEEE J. Quantum Electron. 9(9), 919–933 (1973). [CrossRef]
24. C. Dongre, R. Dekker, H. J. W. M. Hoekstra, M. Pollnau, R. Martinez-Vazquez, R. Osellame, G. Cerullo, R. Ramponi, R. van Weeghel, G. A. J. Besselink, and H. H. van den Vlekkert, “Fluorescence monitoring of microchip capillary electrophoresis separation with monolithically integrated waveguides,” Opt. Lett. 33(21), 2503–2505 (2008). [CrossRef] [PubMed]
25. A. Yariv, “Universal relations for coupling of optical power between microresonators and dielectric waveguides,” Electron. Lett. 36(4), 321–322 (2000). [CrossRef]
26. B. E. Little, S. T. Chu, H. A. Haus, J. Foresi, and J. P. Laine, “Microring resonator channel dropping filters,” J. Lightwave Technol. 15(6), 998–1005 (1997). [CrossRef]
27. C. J. Kaalund, “Critically coupled ring resonators for add-drop filtering,” Opt. Commun. 237(4-6), 357–362 (2004). [CrossRef]
28. A. Yariv, “Critical coupling and its control in optical waveguide-ring resonator systems,” IEEE Photon. Technol. Lett. 14(4), 483–485 (2002). [CrossRef]
29. J. C. Galas, J. Torres, M. Belotti, Q. Kou, and Y. Chen, “Microfluidic tunable dye laser with integrated mixer and ring resonator,” Appl. Phys. Lett. 86(26), 264101 (2005). [CrossRef]
30. B. Liu, A. Shakouri, and J. E. Bowers, “Wide tunable double ring resonator coupled lasers,” IEEE Photon. Technol. Lett. 14(5), 600–602 (2002). [CrossRef]
31. S. I. Shopova, H. Zhou, X. Fan, and P. Zhang, “Optofluidic ring resonator based dye laser,” Appl. Phys. Lett. 90(22), 221101 (2007). [CrossRef]
32. K. Thyagarajan, M. R. Shenoy, and A. K. Ghatak, “Accurate numerical method for the calculation of bending loss in optical waveguides using a matrix approach,” Opt. Lett. 12(4), 296–298 (1987). [CrossRef] [PubMed]
34. A. Ramachandran, S. Wang, J. Clarke, S. J. Ja, D. Goad, L. Wald, E. M. Flood, E. Knobbe, J. V. Hryniewicz, S. T. Chu, D. Gill, W. Chen, O. King, and B. E. Little, “A universal biosensing platform based on optical micro-ring resonators,” Biosens. Bioelectron. 23(7), 939–944 (2008). [CrossRef]
35. C. Y. Chao, W. Fung, and L. J. Guo, “Polymer microring resonators for biochemical sensing applications,” IEEE J. Sel. Top. Quantum Electron. 12(1), 134–142 (2006). [CrossRef]