We report a simple method to efficiently improve the detection limit of surface plasmon resonance in periodic metallic nanostructures by using small angle illumination and spectral integration analysis. The large-area gold nanoslit arrays were fabricated by thermal-annealing template-stripping method with a slit width of 60 nm and period of 500 nm. The small angle illumination induced a resonant coupling between surface plasmon mode and substrate mode. It increased ~2.24 times intensity sensitivity at 5.5° incident angle. The small-angle illumination also resulted in multiple resonant peaks. The spectral integration method integrated all changes near the resonant peaks and increased the signal to noise ratio about 5 times as compared to single-wavelength intensity analysis. Combining both small angle and spectral integration, the detection limit was increased to one order of magnitude. The improvement of the detection limit for antigen-antibody interactions was demonstrated.
© 2014 Optical Society of America
Surface plasmon resonance (SPR) biosensors on gold surface have attracted much attention to sensing applications in human disease, food safety, and environment pollution [1–3]. The SPR biosensors take many advantages such as high surface sensitivity, label-free and real-time detection in liquid environment. For exciting the surface plasmon wave (SPW), the momentum matching condition has to be satisfied. The widely used way to match the wave vector of SPW is to use a prism or grating coupling method [4, 5]. The detection limit is determined by the coupling method and the resolution of the equipment. For very precise measurement, the detectable refractive index unit (RIU) is on the order 10−6 RIU by angular and wavelength interrogation and 10−5 RIU by intensity measurement . Currently, the minimum RIU resolution is achieved by angular interrogation using the prism coupling method. But this measurement system is usually bulky, expensive and needs a large amount of sample solution. High throughput detection is difficult due to a large incident angle for the prism coupling. In recent years, metallic thin film with periodic nanostructures based on extraordinary transmission have shown great potentials for high-throughput, low sample-volume and chip-based detections [6–10]. Nanohole and nanoslit arrays are the commonly used nanostructures [11–14]. The typical sensitivity is about 400-1000 nm/RIU for wavelength interrogation [15, 16], and 1000-5000%/RIU for intensity measurement [17, 18]. The minimum RIU resolution is in the range of 10−4 to 10−5 RIU [19, 20]. The RIU sensitivities for these periodic nanostructures are smaller than those for prism-coupling based SPR sensors.
To improve the sensitivity and detection limit, several methods have been proposed. One of the methods is to adjust the aperture shape, size, thickness and period to optimize the nanostructures [18, 21–24]. In general, the nanostructure with a longer period and a smaller aperture size has a higher sensitivity [25, 26]. Because there is no cut-off for the metallic nanoslit with a very small slit width, it is suggested the nanoslit array is better than the nanohole array. Another way to enhance the sensitivity is to make high-quality metallic film to reduce the propagation loss of SPR. For example, by using thermally annealed process for gold film, the resonant quality of the SPR is improved. The figure of merit (FoM) of the sensor is increased because the bandwidth of the SPR is reduced. The detection limit of the SPR measurement relies both sensitivity and noise level. The signal to noise ratio (S/N) can be improved by using the data analysis methods such as multi-spectral integration analysis and integrated mean-squared analysis [27–31]. If the sensor has multiple SPR peaks sensitive to the surface change, there will be a great improvement in the S/N and the detection limit by integrating all the absolute changes near resonant peaks. In addition to above methods, many recent papers have reported that the sensitivity can be improved by producing resonant coupling mode in the nanostructures. For example, the Fano resonance is formed by the coupling between two resonant modes in the nanostructure. These two modes are the SPR mode on the sensor surface and the broad-band resonance, such as cavity mode in the nanoholes or nanoslits. The coupling between both modes produces an asymmetric resonance with a shaper resonant peak [30, 32].
Recently, we also reported a resonant coupling effect between two SPR modes . These two modes were the substrate mode at the substrate/gold interface and the SPR mode at the gold/medium interface. By tuning the incident angle, the angular momentum of the SPR mode and the substrate mode can be matched at the resonant angle. The coupling between both modes produced a sharp resonant peak and enhanced the FoM of the sensor. However, the resonant mode contained an optical field in the substrate which was insensitive to the surface refractive index change. The wavelength sensitivity was reduced under the resonant condition. In this paper, we propose a simple way to simultaneously increase the sensitivity and detection resolution by using small angle of incidence and the spectral integration method. We found the SPR mode produced the highest intensity sensitivity at an angle smaller than the resonant angle. The small-angle incidence also generated two discrete SPR modes. The spectral integration method integrated all the changes near both surface modes and efficiently increased the S/N. Compared with normal incidence without the spectral integration, we enhanced an order-of-magnitude resolution at 5.5° incident angle. This resolution is higher than prism-based SPR sensor in the intensity measurement. Since the incident angle is small, it would not affect the high throughput detections.
2. Experimental setup and sensor chip fabrication
Figure 1(a) shows the experimental setup for transmission spectrum measurement. The white light source was a 150W halogen lamp and coupled to a 80-μm-core optical fiber. It illuminated the sample with a fiber lens in order to collimate the incident angle. The linear polarizer was placed in front of the sample to control the incident polarization. All the spectra were taken by a fiber-based spectrometer (BWTEK) with another fiber lens to collect the transmission light. The sample was fixed on a rotational stage (OptoSigma). The incident angle to the sample was controlled by rotating the stage. The periodic nanostructure was fabricated by electron-beam lithography (EBL) and reactive ion etching (RIE) method. We spin coated the positive photoresist of ZEP-520 (Zeon Corp) on a silicon substrate. The ZEP-520 thickness was 180 nm. The EBL patterned nanoslit arrays on the photoresist by a nano pattern generation system (NPGS) with a field-emission scanning electron microscope (FE-SEM, Elionix ELS-F125). After development, the nanoslit arrays were transferred to the silicon substrate by using reactive ion etching. The silicon mold was rinsed and sonicated in acetone for 15 min to remove ZEP-520 photoresist.
Gold nanoslit arrays were fabricated on polycarbonate (PC) substrates by using the thermal annealing-assisted template stripping method . We deposited 50-nm-thick Au film on silicon mold by an electron gun evaporator. Next, the PC substrate was placed on silicon mold and covered with a polyethyleneterephthalate (PET) film. All the PET, PC and gold coated Si mold were put in a nitrogen chamber with two heating plates. The nitrogen produced large-area uniform pressure over the Si mold and PC through the PET cover as seen in Fig. 1(b). In the experiment, the nitrogen pressure was 2 kgw/cm2 and heating temperature was 170°C. The temperature was higher than the glass transition temperature (~150°C) of the PC. It softened the PC film for imprinting the nanostructures. Because gold film has a poor adhesion to silicon surface, the PC film was easy to be peeled off from silicon substrate. After stripping off silicon mold, PC and PET film, the flexible gold nanoslit arrays were made. Figure 1(c) shows the SEM images of the silicon mold (left) and the gold nanoslit array (right). The area of the gold nanoslit array was 5 mm × 5 mm. The slit width was 60nm and the period was 500nm. The 500-nm-period nanostructure had a SPR wavelength at ~680 nm for the water/gold interface. This wavelength coincides with the wavelength for the lowest imaginary part of the dielectric constant of gold. The lowest imaginary dielectric constant results in a narrower SPR bandwidth and higher intensity sensitivity .
3 Sensitivity of gold nanoslit arrays
Figure 2(a) explains the forward ( + Κsp) and backward (-Κsp) propagation constants of SPWs at the gold/medium interface (SPR mode) and substrate/gold interface (substrate mode) in the metallic nanostructure. Most optical fields of the SPR mode are on the outside surface. Its propagation constant is sensitive the surface refractive index change. The substrate mode is below the metal surface with most optical fields in the substrate (polycarbonate, n = 1.5). This substrate mode is not changed with the surface condition. According to the momentum matching condition of periodic structures , the + Κsp and -Κsp were varied with the angle of incident light (θ) and the period of the gold nanoslit (Λ),Figure 2(b) shows the angular measurement of transmission spectra from −20° to + 20°. There are 4 bright lines in the diagram and 4 bright spots at the crossing points. To indicate directions of the SPR modes, the theoretical curves as calculated from Eq. (1) are plotted as dashed lines. The yellow lines are the surface SPR mode and green lines are the substrate mode with the forward (m = 1, + Κsp) and backward (m = −1, -Κsp) resonances. It clearly shows that the resonant mode coupling occurs at about ± 8°, where the SPR mode and the substrate mode have the same angular momentum. The mode coupling results in bright spots in the spectra. For more details about the SPR intensity near the resonant angle, the transmission spectra of gold nanoslit arrays in water environment (εd = n2 = 1.3332) for different incident angles (0°, 2°, 4°, 6°, 8°, 10°) are shown in Fig. 2(c). There are two resonant peaks at normal incidence (0°), where the SPR mode is at 678nm and the substrate mode is at 793nm. The Bloch wave surface plasmon polariton (BW-SPP) of resonant wavelengths are close to the calculations at the water/gold interface (704 nm, εm = −29 + 2.0i for gold at 800 nm, i = ± 1, n = 1.3320, and P = 500 nm) and at the PC/gold interface (832 nm, εm = −29 + 2.0i for gold at 800 nm, i = ± 1, n = 1.584, and P = 500 nm) . When the incident light has a small angle, the SPR mode and substrate mode are separated into two modes with forward and backward propagation waves. It is noted that the backward SPR mode is coupled with forward substrate mode. A sharp resonant peak happens at 730nm wavelength when the incident angle is 8°. Before the resonant coupling angle, the SPR has a large increase of intensity. For the intensity sensitivity (SI), the relation between the angle and sensitivity can be described as . Since the is large before the resonant coupling angle, the intensity sensitivity can be greatly enhanced using a small incident angle.
In order to study the sensitivity enhancement due to coupling effect of SPR mode and substrate mode, we measured the transmission spectra in liquid environment with different refractive indexes. The refractive index solution was prepared by mixing deionized water with different volume ratio of glycerin. The refractive index varied from 1.333 to 1.361 (0 to 20% (v/v) glycerin). Figures 3(a)–3(d) show the transmission spectra of different incident angles, 0°, 3°, 5.5°, 7°, under different refractive index solutions. It shows the forward and backward SPR modes are sensitive to environmental refractive index change. The peak wavelengths of SPR modes are red shifted with the increase of refractive index. The peak intensity is also increased with the refractive index. At the angle of 5.5°, there is a large intensity change for the backward SPR mode. Such large change is attributed to the plasmonic mode coupling effect between SPR mode and substrate mode. It is noted that there is a very sharp resonance at the resonant angle. The bandwidth is only ~1/3 of the SPR mode at normal incidence. The narrower bandwidth can increase the FoM (FoM=wavelength sensitivity/bandwidth) of the sensor.
We further calculated the intensity and wavelength sensitivity of SPR modes of forward and backward modes at different incident angles. The intensity sensitivity SI and wavelength sensitivity Sλ are described as follows:Figure 4(a) shows the spectrum of absolute intensity change at 5.5°. The largest intensity change can be found at 735 nm wavelength. We then plotted the intensity change as a function of refractive index at fixed angles (0° and 5.5°) and wavelengths (688nm and 735 nm). Figure 4(b) shows the correlation between intensity change and refractive index. The intensity sensitivity is obtained by the slope and the intensity noise is determined from the fluctuations of intensity signals. In this case, the intensity sensitivities at these two angles (0° and 5.5°) are 6841%/RIU and 15556%/RIU, respectively. The intensity noise is 1.45%. Under this noise level, the detection limit for the conventional intensity method at 5.5° is 8.03 × 10−5 RIU.
Figures 4(c) and 4(d) show the results for intensity and wavelength sensitivities for different incident angles and SPR modes. For the forward SPR mode ( + Ksp), the resonant peak decreases from 678 nm to 622 nm when the incident angle increases from 0° to 10°. This mode is not coupled with substrate modes. The + Ksp mode is an independent mode without any coupling effect. The wavelength sensitivity keeps the same as seen in Fig. 4(d). On the other hand, the resonant wavelength for backward surface mode (-Ksp) increases from 678 nm to 770 nm when the incident angle increases from 0° to 10°. It has the same propagation constant with forward substrate mode at ~8° as seen in Fig. 2(b) and Fig. 2(c). At this angle, both backward SPR mode and forward substrate mode are coupled and form a resonant mode. Because optical fields of the substrate mode and -Ksp mode are combined together, the intensity in the resonant mode is increased. However, the wavelength sensitivity for -Ksp mode is decreased from 6° to 10° as seen in Fig. 4(d). The increased resonant coupling results in an increased optical field in the substrate which is insensitive to environmental changes. The increase of optical energy to the substrate reduces the wavelength sensitivity. Nevertheless, the wavelength sensitivity keeps the same before 5.5°. It indicates that the dominant mode is the surface SPR mode. Figure 4(c) shows that the maximum intensity sensitivity happens at 5.5°. The maximum value is 15556%/RIU. Compared with normal incidence (0°) where the intensity and wavelength sensitivity are 6841%/RIU and 425 nm/RIU, respectively. The intensity sensitivity near the resonant angle (5.5°) is ~2.27 times higher than at normal incidence (0°).
To compare with the nanostructure-based sensor with the prism-coupling SPR sensors, we also analyzed the angular sensitivity () of gold nanoslit arrays. Figures 5(a) and 5(c) show the transmission spectra of gold nanoslit arrays under different refractive index environment. The angular RIU sensitivity is obtained from the slope of resonant angle versus the refractive index. Figures 5(b) and 5(d) shows the fitting results. The angular RIU sensitivity was 57.86 o/ RIU at 630nm and 76.46 o/ RIU at 850nm. For the 500-nm gold nanoslit array, we have measured the angle, wavelength and intensity sensitivities. It is worthwhile to compare the results with conventional prism-based and grating-based SPR sensors. Table 1 shows the comparisons. In general, the chip-based gold nanoslit array is better than the grating-coupled based sensor in angle and intensity detection methods. The prism-based sensor is better at angle and wavelength detections. Nevertheless, through the resonant coupling effect, the chip-based gold nanoslit array can greatly enhance the intensity sensitivity. The intensity sensitivity would be higher than the prism-based sensor at the same operating wavelength. It is noted that angle for producing the resonant coupling effect is only about 5°. This small angle makes the chip-based gold nanoslit array more suitable for the application of SPR imaging and high-throughput detections.
4. Spectral integration analysis
In addition to the increase of intensity sensitivity, the small angle of incidence results in two discrete SPR resonances (forward and backward). As compared with Fig. 3(a) and Fig. 3(c), the small-angle incidence has two spectral regions sensitive to the environmental changes. Such multiple SPRs can be used to increase the S/N by using the spectral integration analysis (SIA). The SIA method has been applied to nanohole SPR sensors and Fano resonances in nansolit sensors. Due to the multiple resonant properties, the S/N has been reported to be enhanced by several times [20–22]. The simple SIA approach is to integrate all the transmission intensity changes within the wavelength range covering all SPR peaks. The integrated response R(n) of the SIA method is express as follows :
Different from the simple intensity measurement, the SIA integration method considers all the resonance peaks in the whole spectrum. In the calculation, we chose the wavelength range from 400 nm to 1000 nm. Figure 6(a) shows the integrated response of different incident angle from 0° to 10° under different refraction index solution from 1.333 to 1.361. Figure 6(b) shows the correlation between the R(n) and the refractive index. We can obtain the RIU sensitivity from linear correlation in Fig. 6(b) and the noise level in Fig. 6(a). The RIU sensitivity and averaged noise as a function of incident angle are shown in Fig. 6(c). The noise levels of different incident angles are the same and kept at 0.2 nm. It confirms that the noise level is the intrinsic property of light source and detector and is not affected by distribution of SPR peaks. The largest RIU sensitivity is 10646 nm/RIU when the incident angle is 5.5°. This angle is consistent with the condition for maximum intensity sensitivity. Comparing with normal angle incidence with SIA analysis, the RIU sensitivity is increased ~1.5 times. It is because that the surface SPR mode only has one resonant peak at 0°. There are two resonant peaks of surface SPR modes at 5.5°. With the SIA method, the additional resonant mode did increase the integrated response signal. By considering the signals and noise levels, the S/N and detection limits at different incident angle can be calculated. Figure 6(d) shows the result. The SIA method improves the sensitivity and S/N on the same nanostructure. The integrated response and noise are 10646 nm/RIU and 0.212nm at 5.5°. The S/N increases ~1.5 times than normal incidence condition. The detection limit is 1.99 × 10−5 RIU. Comparing with the simple intensity method at 5.5° (detection limit, 8.03 × 10−5 RIU), this SIA method increase 4.03 times of detection limits. Comparing with the simple intensity method at 0°, the detection limit is increased up to 9.14 times.
5. Antigen-antibody interaction experiment
The application of the small angle incidence SPR for bio-interaction study was demonstrated by measuring antigen-antibody interactions of bovine serum albumin (BSA) and anti-bovine serum albumin (anti-BSA). In the measurement, 10 mM phosphate-buffered saline (PBS) was used as the buffer solution. At first, PBS was injected into the microfluidic channel by a syringe pump. The transmission spectra were measured at θ = 0° and θ = 5.5° angles, respectively. Then 1 mg/ml BSA was injected for hours to coat the BSA on gold nanoslit surface. Next, we washed the sample surface by PBS solution to remove the unbound BSA. The BSA and PBS were injected four times. After the BSA coating, 50 μg/ml anti-BSA was injected into the channel and waited for four hours antigen-antibody interactions, the sample was washed by PBS again to remove the unbound anti-BSA. Figure 7(a) shows the BSA and anti-BSA interaction on gold nanoslit surface. When BSA and anti-BSA bound together on gold nanoslit surface, the SPR intensity and resonant wavelength were changed as shown in Fig. 7(b) (θ = 0°) and Fig. 7(c) (θ = 5.5°). It is obvious that the 5.5° measurement results in a higher response in the intensity change, especially for the backward SPR at 735 nm wavelength. Figure 7(d) shows the integrated response of BSA coating for different injection times. It is noted that BSA response kept the same after three times injections. Further BSA injection did not increase the response. It indicated that the BSA coating density had reached the saturation. After four times BSA and PBS injection, the 50 μg/ml anti-BSA resulted in a higher integrated response for 5.5° measurement. To show the improvement, we calculated the S/N values at normal and small-angle incidence with and without the spectral integration analysis by using Fig. 7(d). Figure 7(e) shows the averaged S/N values for anti-BSA binding on the nanoslit surface for different measurement conditions. Without the spectral integration, the S/N at θ = 5.5° incidence is 1.5 times higher than the ratio at normal incidence (θ = 0°). With the SIA method, the S/N at θ = 5.5° incidence is 8 times higher than the ratio S/N at θ = 0°. With the combination of small angle incidence and spectral integration analysis, the S/N is 10 times higher than the S/N at normal incidence without SIA. Based on the S/N value and the assumption of linear response, the detection limit for anti-BSA is estimated as 150 ng/ml (θ = 5.5° with SIA). This limit can be further improved if a high-stability light source and low-noise spectrometer are used.
In conclusion, we used the spectral integration method to increase the detection limit and sensitivity of gold nanoslit arrays at small angle incidence (θ = 5.5°). For 500-nm-period gold nanoslit array, the largest intensity sensitivity occurs at 5.5° due to coupling of forward substrate mode SPR mode to the backward surface SPR mode. The intensity sensitivity measured at 5.5° incident angle is 2.27 times higher than normal incidence (θ = 0°). The small-angle incidence not only increases the intensity sensitivity but also produces two distinct SPR modes in the gold nanoslit. Such multiple SPR peaks are used to enhance the S/N and detection limit by using a spectral intensity analysis method. The detection limit with spectral integration analysis is 4 times better than the SPR without spectral integration. Combining both small-angle incidence and spectral integration analysis, about one order of magnitude enhancement of detection limit is achieved. The improvement of the detection limit was further confirmed in the biosensing experiment of BSA and anti-BSA interaction on gold nanoslit arrays. Compared to prism-coupling based SPR sensors, the angular and wavelength sensitivities of the periodic nanoslits are lower. Nevertheless, the higher intensity sensitivity can be achieved by using the resonance coupling effect. The propose method is simple and efficient. It can be applied to other kinds of periodic nanostructures, such as nanohole arrays and nanogrids. Because only a small incident angle is required for the improvement, the advantages of nanostructure sensors, such as high-throughput detections and high-resolution SPR imaging are maintained.
This work was supported by National Science Council, Taipei, Taiwan, under Contract No. NSC-100-2221-E-001-010-MY3 and 101-2120-M-007-001. Technical support from NanoCore, the core facilities for nanoscience and nanotechnology at Academia Sinica in Taiwan, is acknowledged.
References and links
1. J. Homola, S. S. Yee, and G. Gauglitz, “Surface plasmon resonance sensors: review,” Sens. Actuator B-Chem. 54(1–2), 3–15 (1999). [CrossRef]
3. R. J. Green, R. A. Frazier, K. M. Shakesheff, M. C. Davies, C. J. Roberts, and S. J. Tendler, “Surface plasmon resonance analysis of dynamic biological interactions with biomaterials,” Biomaterials 21(18), 1823–1835 (2000). [CrossRef] [PubMed]
6. T. W. Ebbesen, H. Lezec, H. Ghaemi, T. Thio, and P. Wolff, “Extraordinary optical transmission through sub-wavelength hole arrays,” Nature 391(6668), 667–669 (1998). [CrossRef]
7. F. J. Garcia-Vidal, L. Martin-Moreno, T. Ebbesen, and L. Kuipers, “Light passing through subwavelength apertures,” Rev. Mod. Phys. 82(1), 729–787 (2010). [CrossRef]
9. H. J. Lezec, A. Degiron, E. Devaux, R. A. Linke, L. Martin-Moreno, F. J. Garcia-Vidal, and T. W. Ebbesen, “Beaming light from a subwavelength aperture,” Science 297(5582), 820–822 (2002). [CrossRef] [PubMed]
11. A. G. Brolo, R. Gordon, B. Leathem, and K. L. Kavanagh, “Surface plasmon sensor based on the enhanced light transmission through arrays of nanoholes in gold films,” Langmuir 20(12), 4813–4815 (2004). [CrossRef] [PubMed]
12. K. A. Tetz, L. Pang, and Y. Fainman, “High-resolution surface plasmon resonance sensor based on linewidth-optimized nanohole array transmittance,” Opt. Lett. 31(10), 1528–1530 (2006). [CrossRef] [PubMed]
13. K.-L. Lee, C.-W. Lee, W.-S. Wang, and P.-K. Wei, “Sensitive biosensor array using surface plasmon resonance on metallic nanoslits,” J. Biomed. Opt. 12, 044023 (2007).
14. Y. S. Jung, Z. Sun, J. Wuenschell, H. K. Kim, P. Kaur, L. Wang, and D. Waldeck, “High-sensitivity surface plasmon resonance spectroscopy based on a metal nanoslit array,” Appl. Phys. Lett. 88(24), 243105 (2006). [CrossRef]
15. K.-L. Lee, W.-S. Wang, and P.-K. Wei, “Comparisons of surface plasmon sensitivities in periodic gold nanostructures,” Plasmonics 3(4), 119–125 (2008). [CrossRef]
16. A. Lesuffleur, H. Im, N. C. Lindquist, and S.-H. Oh, “Periodic nanohole arrays with shape-enhanced plasmon resonance as real-time biosensors,” Appl. Phys. Lett. 90(24), 243110 (2007). [CrossRef]
18. A. P. Blanchard-Dionne, L. Guyot, S. Patskovsky, R. Gordon, and M. Meunier, “Intensity based surface plasmon resonance sensor using a nanohole rectangular array,” Opt. Express 19(16), 15041–15046 (2011). [CrossRef] [PubMed]
19. C. Escobedo, S. Vincent, A. Choudhury, J. Campbell, A. Brolo, D. Sinton, and R. Gordon, “Integrated nanohole array surface plasmon resonance sensing device using a dual-wavelength source,” J. Micromech. Microeng. 21(11), 115001 (2011). [CrossRef]
20. H. Im, S. H. Lee, N. J. Wittenberg, T. W. Johnson, N. C. Lindquist, P. Nagpal, D. J. Norris, and S.-H. Oh, “Template-stripped smooth Ag nanohole arrays with silica shells for surface plasmon resonance biosensing,” ACS Nano 5(8), 6244–6253 (2011). [CrossRef] [PubMed]
21. V. E. Bochenkov, M. Frederiksen, and D. S. Sutherland, “Enhanced refractive index sensitivity of elevated short-range ordered nanohole arrays in optically thin plasmonic Au films,” Opt. Express 21(12), 14763–14770 (2013). [CrossRef] [PubMed]
22. M. Najiminaini, F. Vasefi, B. Kaminska, and J. J. Carson, “Effect of surface plasmon energy matching on the sensing capability of metallic nano-hole arrays,” Appl. Phys. Lett. 100(6), 063110 (2012). [CrossRef]
23. J. P. Monteiro, J. Ferreira, R. G. Sabat, P. Rochon, M. J. L. Santos, and E. M. Girotto, “SPR based biosensor using surface relief grating in transmission mode,” Sens. Actuator B-Chem. 174, 270–273 (2012). [CrossRef]
24. P. Jia, H. Jiang, J. Sabarinathan, and J. Yang, “Plasmonic nanohole array sensors fabricated by template transfer with improved optical performance,” Nanotechnology 24(19), 195501 (2013). [CrossRef] [PubMed]
28. K. L. Lee, M. J. Chih, X. Shi, K. Ueno, H. Misawa, and P. K. Wei, “Improving surface plasmon detection in gold nanostructures using a multi-polarization spectral integration method,” Adv. Mater. 24(35), OP253–OP259 (2012). [CrossRef] [PubMed]
29. M. Das, D. Hohertz, R. Nirwan, A. G. Brolo, K. L. Kavanagh, and R. Gordon, “Improved performance of nanohole surface plasmon resonance sensors by the integrated response method,” IEEE Photon. J. 3(3), 441–449 (2011). [CrossRef]
30. K.-L. Lee, P.-W. Chen, S.-H. Wu, J.-B. Huang, S.-Y. Yang, and P.-K. Wei, “Enhancing surface plasmon detection using template-stripped gold nanoslit arrays on plastic films,” ACS Nano 6(4), 2931–2939 (2012). [CrossRef] [PubMed]
31. W.-S. Tsai, K.-L. Lee, M.-Y. Pan, and P.-K. Wei, “Increased detection sensitivity of surface plasmon sensors using oblique induced resonant coupling,” Opt. Lett. 38(23), 4962–4965 (2013). [CrossRef] [PubMed]
32. B. Luk’yanchuk, N. I. Zheludev, S. A. Maier, N. J. Halas, P. Nordlander, H. Giessen, and C. T. Chong, “The Fano resonance in plasmonic nanostructures and metamaterials,” Nat. Mater. 9(9), 707–715 (2010). [CrossRef] [PubMed]
33. K.-L. Lee and P.-K. Wei, “Optimization of periodic gold nanostructures for intensity-sensitive detection,” Appl. Phys. Lett. 99(8), 083108 (2011). [CrossRef]