We report experimental results of label-free anti-bovine serum albumin (anti-BSA) antibody detection using a SOI planar photonic crystal waveguide previously bio-functionalized with complementary BSA antigen probes. Sharp fringes appearing in the slow-light regime near the edge of the guided band are used to perform the sensing. We have modeled the presence of these band edge fringes and demonstrated the possibility of using them for sensing purposes by performing refractive index variations detection, achieving a sensitivity of 174.8 nm/RIU. Then, label-free anti-BSA biosensing experiments have been carried out, estimating a surface mass density detection limit below 2.1 pg/mm2 and a total mass detection limit below 0.2 fg.
©2010 Optical Society of America
The use of integrated planar photonic devices for biosensing purposes has been attracting an increasing interest in the last few years. First demonstrations of the possibility of using these devices for sensing applications were reported in the middle of 90s of past century, as for example in  and , but it has not been until 4-5 years ago that many researchers have re-oriented their investigations to the application of already developed integrated planar photonic devices for biosensing applications. The transduction principle of these devices is based on the dependence of their response to changes in the refractive index (RI) of the surrounding medium. Some of the main advantages of these devices for sensing applications are their high sensitivity due to the high confinement of the electromagnetic field in the photonic structure which enhances the interaction with the target analyte, and their reduced size, which makes it possible both to detect very small analytes and to integrate many of these devices on a single chip to perform a multi-parameter detection. Particular interest is focused on integrated planar photonic devices fabricated on silicon-on-insulator (SOI), also given the important advantage of their compatibility with CMOS fabrication technologies, opening the door to a low cost and high volume production of these devices.
Many different integrated planar photonic devices are widely used for sensing applications, amongst the most popular are ring resonators [3–8], photonic crystal based and periodic structures [9–14], and Mach-Zehnder Interferometers [15,16]. Reported results range from basic characterization of the sensing device to refractive index variation to more complex bio-sensing experiments where the presence of a specific analyte such as proteins, bacteria or even DNA strands is detected and quantified without the need of any marker. This capability to perform label-free detection together with previously commented advantages (i.e., high sensitivity, small size, multi-parameter detection) make them a serious candidate to be the technological base for lab-on-a-chip devices of the future.
In this work, we report experimental results of both refractive index sensing and label-free detection of antibodies using SOI planar photonic crystal waveguides (PCW). Planar photonic crystal based sensing devices have already been used for sensing [9–13], but in our case sharp fringes appearing in the slow-light regime near the edge of the guided band are used to perform the sensing. These sharp fringes are easier to track than the edge of either the photonic bandgap or the guided band, and their use ensures us that we are working in the slow-light regime of the PCW, what makes the interaction with the target substance to be stronger. We have also modeled the origin of these band edge fringes.
2. Fabrication and characterization of the photonic crystal waveguide
Planar PCWs are fabricated in a SOI wafer with a 250nm-thick silicon layer on a 3μm-thick buried oxide layer. Designs are patterned into a PMMA positive resist layer using e-beam lithography with an acceleration voltage of 2keV and a dose of ~23 μC/cm2, and then, after developing the resist, the pattern is transferred to the silicon layer by inductively coupled plasma (ICP) etching.
PCWs used in our experiments are W1-type, where one row of holes is removed in the Γ-K direction to create the waveguide, as shown in the scanning electron microscopy (SEM) picture in Fig. 1 . This type of PCW presents a guided mode inside the photonic bandgap of the photonic crystal for TE polarization. Nominal values for the lattice constant and the hole radius are 390 nm and 110 nm respectively in order to have the upper edge (in terms of wavelength) of the guided band of the PCW located around λ = 1550 nm. The length of the PCW is set to 20 μm (≈52 periods) to have enough periods for achieving a strong photonic bandgap effect while keeping the structure size as compact as possible. Light is coupled/collected to/from the PCW using 450nm-wide single-mode access waveguides, which are tapered to a width of 3 μm to ease the butt-coupling of input light from a lensed fiber. The interface between the PCW and the single-mode access waveguides is that shown in Fig. 1, where the structure is finished besides the inner row of holes of the PCW. This interface was selected because previous experiments showed us that it presents a sharper band edge than other interfaces.
For the optical characterization of the PCWs, light from a tunable laser is TE-polarized using a polarization controller and coupled to the chip using a lensed fiber. Output light is then collected using an objective and detected using an optical powermeter after passing through a free-space polarizer configured for TE polarization. We obtained the transmission spectrum of one of the fabricated PCWs when having deionized water (DIW, n = 1.3173 ) as upper-cladding, which is shown in Fig. 1. Sharp peaks are observed at the edge of the guided band (located around 1563 nm), which will later be used for sensing purposes, but, which is the origin of these fringes?
3. Theoretical analysis of the photonic crystal waveguide
Fringes appearing near the band edge of the transmission spectrum of the PCW are Fabry-Perot (FP) fringes of the cavity defined by the interface between the PCW and the access waveguides , as schematically depicted in Fig. 2 . A good power coupling between modes in the access waveguides and the PCW is achieved for wavelengths in the transmission band, making the FP cavity effect almost negligible. However, mode mismatching increases between the two waveguides as we get closer to the edge of the guided band, thus reducing the coupling efficiency and increasing the reflected power, so higher amplitude FP fringes begin to appear. Not only is there an increase of the reflection coefficient in the interfaces (and thus a reduction of the transmission coefficient), but also a reduction of the group velocity of the guided mode of the PCW as we get closer to the edge of the Brillouin zone. The reduction of the group velocity makes the optical length of the FP cavity longer, thus increasing the frequency of the FP fringes of the transmission spectrum in the region of the band edge. The main point of using these fringes to perform the biosensing is that we are working in the slow-light regime of the PCW, so we will have a higher interaction of the electromagnetic field with the target analyte.
From expressions shown in Fig. 2(b), it can be seen that the transmission spectrum is dependent on both the transmission and reflection coefficients at the interfaces and the propagation constant of the guided mode in the PCW, which determines the optical length of the structure. We have obtained these three parameters (transmission and reflection coefficients between the access waveguides and the PCW, and propagation constant of the PCW guided mode) using the numerical tool CAMFR, which is based in the eigenmode expansion (EME) method , for the PCW with its interface with the access waveguides as shown in Fig. 2(a). Note that only the reflection coefficient for the incidence from the PCW to the access waveguide (r 21, r 23) is needed for the calculations and that the transmission coefficient is the same for the incidence from the PCW to the access waveguide and vice versa due to reciprocity properties (t ij = t ji) . Since this software does not allow 3D calculations, we have made a generic 2D modelling (i.e., infinite height of the structure) using an effective refractive index of 2.8 for the silicon slab and 1.33 for the cladding surrounding the structure; these are not the exact refractive index values for our real 3D structure but they are close enough to allow us to realistically predict the appearance of FP fringes in the transmission spectrum. Parameters are shown in Fig. 3(a) . It can be seen that the transmission coefficient decreases to zero as we get closer to the band edge (the opposite for the reflection coefficient, which tends to one), thus the amount of power coupled to the PCW decreases and a stronger cavity effect is created. Concerning the propagation constant, it can be seen that it gets flatter as the edge of the guided band is approached, what means a reduction on the group velocity (which tends to zero) and makes the fringes to be narrower and closer between them.
When we combine these three parameters using the expression shown in Fig. 2(b), the transmission response shown in Fig. 3(b) for a 20μm-long 2D-PCW is obtained. It can be seen how FP resonances appear because of the cavity created inside the PCW, and how these FP fringes get stronger and narrower as we approach the edge of the guided band because of the increase of the reflection coefficient and the decrease of the group velocity.
To corroborate the appearance of these FP fringes in the edge of the guided mode of the PCW and analyze the potential for using them for biosensing purposes, 3D-FDTD simulations of the PCW with the parameters of the fabricated structure (height = 250 nm, hole radius = 110 nm, period = 390 nm, PCW length = 20 μm, and access waveguides width = 450 nm) and two different uppercladdings (n 1 = 1.3173 and n 2 = 1.3200; as it will be described in section 4, these are the refractive indices for DIW and a ethanol-DIW 4% dilution) have also been performed using the RSoft’s FullWAVE software. A grid size of 25 nm is used in each axis. Simulation results for each uppercladding are shown in Fig. 4 , where the FP fringes at the band edge are observed again and where the position of the guided band is dependent of the refractive index of the uppercladding. We have measured the wavelength shift and obtained the sensitivity in terms of refractive index units (RIU) for each fringe near the band edge, which are shown in Table 1 . As theoretically predicted, an increase in the sensitivity is obtained as we get closer to the guided band edge because of the reduction in the group velocity of the guided mode, obtaining a sensitivity ~40% higher for the fringe closer to the band edge compared to the first of the selected fringes (78.70 nm/RIU vs. 57.15 nm/RIU).
4. Refractive index sensing using band edge FP fringes
In order to experimentally test the feasibility of using these band edge fringes to perform biosensing, we have first performed a refractive index sensing experiment using several dilutions of ethanol in DIW (in mass %): pure DIW, ethanol 2% in DIW, and ethanol 4% in DIW. Since accurate values for the RI of ethanol-DIW dilutions for our working conditions (λ ≈1550 nm and T = 25°C) were not found, we estimated them from other reported values of these liquids and dilutions at different conditions. In , the RI of an ethanol-DIW dilution with different concentrations is given for λ = 589 nm and T = 20°C. For mass concentrations below 6%, we can linearly approximate those values as:22]) and nDIW = 1.3330 (from ).
To estimate the RI of the ethanol-DIW dilutions for our working conditions (λ ≈1550 nm and T = 25°C), we used the linear relation from Eq. (1) where this time: nethanol = 1.3515 (obtained using Cauchy dispersion coefficients and approximating the RI variation for the temperature change between 20°C and 25°C from data in ) and nDIW = 1.3173. Therefore, the estimated RI for our dilutions at λ ≈1550 nm and T = 25°C are: pure DIW (n = 1.3173), ethanol 2% in DIW (n = 1.3186), and ethanol 4% in DIW (n = 1.3200).
For the RI sensing experiments, the chip is placed on our characterization setup and a 2-port flow cell with a fluidic cavity of size 5.5mm x 2mm is placed on the top of it. A syringe is connected to one of the ports of the flow cell using silicone tubing (inner diameter of 0.5 mm) and placed in a syringe pump in withdrawal mode to control the flow rate. Tubing from the second port of the flow cell is placed into a vial with the liquid to be flowed over the chip. With this configuration, liquid from the vial is sucked; this enables an easier handling of the tubes when manual change between the vials is required. The TE transmission spectrum of the PCW in the vicinity of the guided band edge was continuously acquired using a tunable laser with a sweep resolution of 10 pm, and then a cubic interpolation was used to increase the wavelength accuracy on the determination of the position of the peak’s maximum. We selected the FP peak located around 1563.3 nm (see Fig. 1) to be tracked during RI sensing experiments. The temporal evolution of the position of the maximum of the peak for the different ethanol-DIW dilutions flowed (a flow rate of 15 μl/min is used) is shown in Fig. 5 and the peak shape for different time positions is shown in Fig. 6(a) .
Since a slow negative drift is present in the peak position (around −42.1 pm/hour, attributed to the lack of a thermal control in our characterization setup), we have worked with relative wavelength shifts for changes in the ethanol-DIW concentration instead of absolute peak position in order to determine the sensor sensitivity and the detection limit. This relative peak shift is depicted in Fig. 6(b) for the three ethanol-DIW concentrations used in the experiments. A mean wavelength shift of 0.2269 nm was observed for changes between DIW and ethanol 2% (with a standard deviation of 19.3 pm in the peak position between measurements) and of 0.2452 nm between ethanol 2% and ethanol 4% (with a standard deviation of 27.1 pm in the peak position between measurements). Linear interpolation has been used to fit these data, giving a sensitivity with respect to the ethanol concentration of 118 pm/%.
Considering the almost linear RI variation for this ethanol concentrations range (Δn ≈6.75x10−4/%), a sensitivity to RI variations of 174.8 nm/RIU is obtained, what is 2.2x the value obtained in 3D-FDTD simulations. We attribute this difference to the discretization of the PCW in the simulations (a 25 nm grid size is used to model the structure), what may lead to a not proper modelling of the optical field which sense the variation of the refractive index of the cladding.
The sensitivity obtained in our experiments is significantly higher than previously reported results for other holes-on-silicon photonic crystal sensors. For example, in  the shift of the guided band edge of a PCW was used for RI sensing, achieving sensitivities around 70 nm/RIU. Moreover, in our case the position of the features to be tracked (i.e., the FP fringes) can be determined in a more accurate manner than when the pure band edge is selected. Our value is also higher than reported results for RI sensing using photonic crystal cavities , where sensitivity values around 100 nm/RIU are obtained.
Thus, the reason for our higher sensitivity value is the higher shift caused by the reduction of the group velocity in the band edge, where the FP fringes used to sense are located. Many works have been focused on exploiting this slow light effect for other applications such as optical switching by increasing the interaction of the mode field with nonlinear materials (some results are reviewed in ), but this is to our knowledge the first time that the benefits of slow light for sensing applications are clearly shown for photonic crystal based structures. Higher sensitivity values using photonic crystal based structures have been recently reported in  (350 nm/RIU), where a less common photonic crystal configuration based on silicon pillars on air instead of air holes on silicon is used. This configuration has the advantage of having a higher percentage of air (or the target substance) in contact with the silicon pillars, but more challenging fabrication techniques are required in order to have these thicker silicon pillars with a good aspect ratio.
If the maximum standard deviation for the peak position between measurements is considered (i.e., 27.1 pm for measurements of ethanol 4%), a detection limit of 1.55x10−4 RIU is obtained. This is not an extremely low value for the detection limit, since it is limited by the repeatability in the peak position when we change the dilutions. Much lower detection limits could have been reported if had considered the noise in the peak position for the continuous flow of a certain dilution (as usually made), which has a standard deviation around 0.6 pm, thus giving a detection limit around 3.5x10−6 RIU. We think that these deviations are due to fact of using a macrofluidic flow cell to flow the dilutions over the PCW; we expect to significantly reduce them (and thus the detection limit) by using microfluidic channels to flow the samples.
Finally, it is worth noting the sharp negative shift of the peak position shown at 44 minutes in Fig. 5, which is caused because the tracked peak becomes double and the tracked maximum changes its position suddenly, as depicted in the corresponding peak of Fig. 6(a).
5. Antibody sensing using photonic crystal waveguides
Once the origin of band edges fringes is modeled and the possibility of using them for sensing purposes is checked by performing RI sensing experiments, we have used a second PCW (with the same nominal parameters) from another fabrication run to perform label-free detection of antibodies. For this antibody sensing experiment, the PCW is first bio-functionalized with bovine serum albumin (BSA) antigen probes. This bio-functionalization process begins by activating the surface of the chip by exposing it to pure 3-isocyanatepropyl triethoxysilane (ICPTES) vapour for 30 minutes. Then, a drop of BSA antigen 10 μg/ml in PBS 0.1x was deposited on the sensing area and incubated in a humid chamber overnight at room temperature. Finally, a solution of ovoalbumin protein (OVA) 1% in PBS 0.1x was spread over the whole chip surface and incubated for 30 minutes to block the remaining active sites.
After the bio-functionalization process, we placed the chip on our characterization setup with the flow cell on the top of it. Both the flow cell and the tubing used for the experiments are previously blocked with OVA to avoid the absorption of the flowed molecules. Transmission spectrum for the TE polarization is obtained for the bio-functionalized PCW when having phosphate buffer saline (PBS) 0.1x as upper-cladding (n ≈1.335 , ), which is shown in Fig. 7 , where the band edge is now located around 1542 nm. A closer look to the band edge is given in Fig. 7(b), where sharp peaks appearing in this region are observed again.
Several solutions are then flowed over the chip (a flow rate of 15 μl/min is used again) to carry out the sensing experiment. The initial baseline is obtained by continuously flowing the PBS 0.1x buffer. Once a stable baseline is achieved, we switched to an anti-BSA antibody 10 μg/ml in PBS 0.1x solution, which will bind to the BSA probes attached in the PCW surface thus inducing a shift in the peak position. The anti-BSA is flowed enough time to achieve a monolayer on the top of the BSA-functionalized chip, what is indicated by the saturation in the shift of the PCW response. Then, the flow is switched again to the PBS 0.1x buffer to remove any anti-BSA which has not specifically bound to probes on the surface of the PCW, thus only the net shift due to the binding of the anti-BSA to the BSA probes is obtained. Finally, we flow an anti-digoxigenin (anti-DIG) antibody 15 μg/ml in PBS 0.1x dilution to check that the shift previously obtained for the anti-BSA antibody flow is only due to a specific binding with the BSA probes and not due to absorption or any other mechanism. Because of the low affinity between the anti-DIG and the BSA probes, no peak shift due to their binding should be observed, although a slight shift due to the change in the refractive index of the surrounding medium might be observed. Later, PBS 0.1x is flowed again to finish the experiment.
Figure 8 shows the temporal evolution of the peak shift for the first four peaks depicted in Fig. 7(b) (peaks at 1542.1nm, 1543.5nm, 1544.2nm and 1544.7nm); the fifth peak (at 1545.4nm) is not considered because it went out of the wavelength sweeping range during the flow of the anti-BSA. It can be seen that the temporal evolution is almost the same for all the peaks tracked, reaching a plateau for the anti-BSA flow what indicates the formation of a monoloayer. Concerning the total shift when the anti-BSA is flowed, it is slightly different for each peak, as shown in Table 2 . It can be seen that the shift gets higher as we move to peaks closer to the band edge (i.e., from peak #1 to peak #4), what as observed in the 3D-FDTD simulations for RI sensing, is due to the reduction of the group velocity of the guided mode. However, peak #3 does not follow this trend and shows a wavelength shift smaller than peak #2. The surface density for a close-pack anti-BSA monolayer when considering a 100% binding efficiency is ρanti-BSA = 1.7 ng/mm2, which is calculated from the molecular mass and the size of the anti-BSA molecule (as previously done in ), and will give us an upper limit for the detection limit of our device (in the real situation we don’t have a close-pack monolayer and binding efficiency is below 100%). The sensitivity for the anti-BSA detection is given by Santi-BSA = Δλanti-BSA/ρanti-BSA; calculated values for each peak are shown in Table 2.
We have obtained the noise level in the measurements for each peak by calculating the standard deviation of the peak position (σ) for the stable PBS 0.1x cycle flowed after the anti-BSA. They are shown in Table 2. It can be seen that the noise level is very low (σ = 1.5pm), except for peak #3, where the noise level is twice this value (σ = 3.1pm). With these noise values and the sensitivities previously calculated for each peak, we have obtained the surface mass density detection limit (given by DLanti-BSA = σPBS/Santi-BSA), which is shown in Table 2 for each peak. A surface mass density detection limit of <2.1 pg/mm2 is obtained from the calculations.
The total mass detection limit has also been obtained from the surface mass density detection limit considering a total surface of our PCW of ~320 μm2 (the area of the PCW is 20μm x 7μm, and the internal surface of the holes has also been considered). It is shown in Table 2 for each peak. Due to the reduced size of our PCW based sensor, a total mass detection limit of ~0.7 fg has been calculated. However, if only the region surrounding the linear defect forming the PCW is used for the calculations, since it is where the optical field is confined and where the interaction with the target anti-BSA is actually taking place (around two-three rows of holes at each side of the linear defect, what is around 100 μm2 considering the internal surface of the holes too), a total mass detection limit of ~0.2 fg is obtained.
Concerning the improvement because of the reduction of the group velocity, we can see in Table 2 how calculated values for the sensitivity and the detection limit slightly improve as we move closer to the band edge (peak #3 is the only which does not follow this trend and also has a higher noise level, what suggest that it has a poorer quality than the other peaks used in the experiments), although only a 10% improvement is obtained when moving from peak #1 to #4. This is the same increase that we obtained for the 3D-FDTD simulations when only considering the four peaks closer to the band edge. However, in the simulations we saw that the increase is higher if we consider peaks located far away of the guided band edge (around 40%). Therefore, a higher increase in the sensitivity might have been obtained in the experiments if more FP fringes were acquired (but longer sweep times might have been required to acquire each spectrum). However, it is worth noting that we have achieved high sensitivity values when working in this wavelength region for a device with a small footprint, what makes it suitable to detect very low amounts of analyte.
Concerning the flow of anti-DIG antibody, a slight shift of the peaks is also observed, which indicates that molecular absorption and/or unspecific binding has taken place. Nevertheless, this effect seems to be quite low, since there is only a shift of around 0.1nm for an anti-DIG concentration of 15 μg/ml, which is 10 times lower than the shift observed when flowing anti-BSA.
We have presented experimental results of label-free detection of anti-BSA antibody using a BSA-functionalized SOI planar photonic crystal waveguide. We have estimated a surface mass density detection limit below 2.1 pg/mm2 what is to our knowledge the lowest surface mass density detection limit reported for any planar photonic crystal based sensing structure [9,12] and close to current lowest values reported for other planar sensor technologies [7,16]. Moreover, due to the reduced size of our PCW, a total mass detection limit below 0.7 fg has been estimated, what can be even considered to be below 0.2 fg if only the active area of the PCW is considered in the calculations, what is below other previously reported results for any planar photonic sensor [4,10,16]. A lower total mass detection limit may be obtained using a shorter PCW with a smaller total surface.
The use of band edge FP fringes located in the slow-light regime of the transmission band allows having a higher interaction with the target analyte for this wavelength range. Moreover, FP fringes obtained in the band edge have a reduced bandwidth, what makes easier and more accurate the determination of their position with low uncertainty. Refractive index sensing experiments have also been performed, achieving a sensitivity of 174.8 nm/RIU. Finally, the presence of multiple band edge FP fringes can open the door to a more complex analysis of their evolution, both individual and collective, what could enhance the performance of the sensing device in terms of sensitivity and detection limit.
This work was funded by the European Commission, carried out within the FP7-ICT-223932-INTOPSENS project, and by Spanish MICINN under contracts TEC2008-06333 and CTQ2007-64735/BQU. Support from the Universidad Politécnica de Valencia through programs PAID-05-09 and PAID-06-09 and the Conselleria d’Educació through program GV-2010-031 is also acknowledged.
References and links
2. Th. Schubert, N. Haase, H. Kück, and R. Gottfried-Gottfried, “Refractive-index measurements using an integrated Mach-Zehnder interferometer,” Sens. Actuators A Phys. 60(1-3), 108–112 (1997). [CrossRef]
3. K. De Vos, I. Bartolozzi, E. Schacht, P. Bienstman, and R. Baets, “Silicon-on-Insulator microring resonator for sensitive and label-free biosensing,” Opt. Express 15(12), 7610–7615 (2007), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-15-12-7610. [CrossRef] [PubMed]
4. K. De Vos, J. Girones, S. Popelka, E. Schacht, R. Baets, and P. Bienstman, “SOI optical microring resonator with poly(ethylene glycol) polymer brush for label-free biosensor applications,” Biosens. Bioelectron. 24(8), 2528–2533 (2009). [CrossRef] [PubMed]
5. 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]
6. C. A. Barrios, M. J. Bañuls, V. González-Pedro, K. B. Gylfason, B. Sánchez, A. Griol, A. Maquieira, H. Sohlström, M. Holgado, and R. Casquel, “Label-free optical biosensing with slot-waveguides,” Opt. Lett. 33(7), 708–710 (2008). [CrossRef] [PubMed]
7. C. F. Carlborg, K. B. Gylfason, A. Kaźmierczak, F. Dortu, M. J. Bañuls Polo, A. Maquieira Catala, G. M. Kresbach, H. Sohlström, T. Moh, L. Vivien, J. Popplewell, G. Ronan, C. A. Barrios, G. Stemme, and W. van der Wijngaart, “A packaged optical slot-waveguide ring resonator sensor array for multiplex label-free assays in labs-on-chips,” Lab Chip 10(3), 281–290 (2010). [CrossRef] [PubMed]
8. M. Iqbal, M. A. Gleeson, B. Spaugh, F. Tybor, W. G. Gunn, M. Hochberg, T. Baehr-Jones, R. C. Bailey, and L. C. Gunn, “Label-Free Biosensor Arrays Based on Silicon Ring Resonators and High-Speed Optical Scanning Instrumentation,” IEEE J. Sel. Top. Quantum Electron. 16(3), 654–661 (2010). [CrossRef]
9. N. Skivesen, A. Têtu, M. Kristensen, J. Kjems, L. H. Frandsen, and P. I. Borel, “Photonic-crystal waveguide biosensor,” Opt. Express 15(6), 3169–3176 (2007), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-15-12-7610. [CrossRef] [PubMed]
11. S. Zlatanovic, L. W. Mirkarimi, M. M. Sigalas, M. A. Bynum, E. Chow, K. M. Robotti, G. W. Burr, S. Esener, and A. Grot, “Photonic crystal microcavity sensor for ultracompact monitoring of reaction kinetics and protein concentration,” Sens. Actuators B Chem. 141(1), 13–19 (2009). [CrossRef]
12. D. Dorfner, T. Zabel, T. Hürlimann, N. Hauke, L. Frandsen, U. Rant, G. Abstreiter, and J. Finley, “Photonic crystal nanostructures for optical biosensing applications,” Biosens. Bioelectron. 24(12), 3688–3692 (2009). [CrossRef] [PubMed]
13. T. Xu, N. Zhu, M. Y.-C. Xu, L. Wosinski, J. S. Aitchison, and H. E. Ruda, “Pillar-array based optical sensor,” Opt. Express 18(6), 5420–5425 (2010), http://www.opticsinfobase.org/abstract.cfm?URI=oe-18-6-5420. [CrossRef] [PubMed]
14. L. J. Kauppinen, H. J. W. M. Hoekstra, and R. M. de Ridder, “A compact refractometric sensor based on grated silicon photonic wires,” Sens. Actuators B Chem. 139(1), 194–198 (2009). [CrossRef]
15. B. Sepúlveda, J. S. Río, M. Moreno, F. J. Blanco, K. Mayora, C. Domínguez, and L. M. Lechuga, “Optical biosensor microsystems based on the integration of highly sensitive Mach–Zehnder interferometer devices,” J. Opt. A, Pure Appl. Opt. 8(7), S561–S566 (2006). [CrossRef]
16. A. Densmore, M. Vachon, D.-X. Xu, S. Janz, R. Ma, Y.-H. Li, G. Lopinski, A. Delâge, J. Lapointe, C. C. Luebbert, Q. Y. Liu, P. Cheben, and J. H. Schmid, “Silicon photonic wire biosensor array for multiplexed real-time and label-free molecular detection,” Opt. Lett. 34(23), 3598–3600 (2009). [CrossRef] [PubMed]
17. P. Schiebener, J. Straub, J. M. H. Levelt Sengers, and J. S. Gallagher, “Refractive index of water and steam as function of wavelength, temperature and density,” J. Phys. Chem. Ref. Data 19(3), 677–717 (1990). [CrossRef]
18. J. García, P. Sanchis, and J. Martí, “Detailed analysis of the influence of structure length on pulse propagation through finite-size photonic crystal waveguides,” Opt. Express 14(15), 6879–6893 (2006), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-14-15-6879. [CrossRef] [PubMed]
19. This software can be encountered at http://camfr.sourceforge.net/
20. P. Sanchis, J. Martí, P. Bienstman, and R. Baets, “Semi-analytic approach for analyzing coupling issues in photonic crystal structures,” Appl. Phys. Lett. 87(20), 203107 (2005). [CrossRef]
21. D. R. Lide, Handbook of Chemistry and Physics (CRC Press, 2008).
22. J. Rheims, J. Köser, and T. Wriedt, “Refractive-index measurements in the near-IR using an Abbe refractometer,” Meas. Sci. Technol. 8(6), 601–605 (1997). [CrossRef]
23. T. F. Krauss, “Why do we need slow light?” Nat. Photonics 2(8), 448–450 (2008). [CrossRef]
24. G. Gupta, M. Sugimoto, Y. Matsui, and J. Kondoh, “Use of a low refractive index prism in surface plasmon resonance biosensing,” Sens. Actuators B Chem. 130(2), 689–695 (2008). [CrossRef]