Sensors based on surface plasmons or waveguide modes are at the focus of interest for applications in biological or environmental chemistry. Waveguide-mode spectra of 1 μm-thick pure and perforated silica films comprising isolated nanometric holes with great aspect ratio were measured before and after adhesion of streptavidin at concentrations of 500 nM. The shift of the angular position for guided modes was nine times higher in perforated films than in bulk films. Capturing of streptavidin in the nanoholes is at the origin of that largely enhanced shift in the angular position as the amplitude of the guided mode in the waveguide perfectly overlaps with the perturbation caused by the molecules. Hence, the device allows for strongly confined modes and their strong perturbation to enable ultra-sensitive sensor applications.
© 2007 Optical Society of America
Surface plasmon resonances (SPR) is a powerful phenomena for real time detection of bio-molecules in a label-free environment. For the excitation of SPRs, the Kretschmann (Fig. 1) configuration is often employed . Briefly, a TM-polarized beam is directed onto the base of a glass prism at an angle necessary for total internal reflection. On the base a suitable metallic thin film is deposited. The electric field couples with the charge density oscillation of the free electrons in the metal and excites a surface plasmon. The surface plasmon causes an evanescent field outside the film whose amplitude decays exponentially away from the metal surface. If an additional dielectric layer is incorporated onto the metal film, the incident light excites likewise weakly guided modes in this waveguiding structure. Both waveguide modes and SPR are simultaneously observed as a sudden drop of reflectivity, as shown in Fig. 2(a). Comparable to the experimental constraints, the structure here consists of a glass prism nPrism = 1.846 on which a gold film (hAu = 50 nm, nAu = 0.166+3.15 i) and a dielectric layer (h SiO2 = 500 nm , nSiO2 = 1.479) were deposited, where h and n were film thickness and refractive index, respectively. Medium on the transmission side is water (n = 1.333) and the used wavelength was λ = 633nm. Both TE and TM polarization exhibit waveguide modes up to the first or zero order, respectively. A plasmon is observed for TM polarization only. Modes are excited, if the tangential wave vector component matches to the wave number of a guided mode. As guiding losses for waveguides modes are significantly smaller, their line width is superior as compared to the SPR. Perturbing the dielectric environment changes the dispersion of the modes and causes a shift in the angle necessary for their excitation. Very high sensitivity can be expected if the overlap of the modal distribution with external stimuli is maximized. This is not easily achieved for the waveguide modes, as their field amplitudes are concentrated in the volume [Fig. 2(b)]. Nevertheless, their use for sensors is appealing as shifts could be easily detected due to their narrow line width. Full widths at half maximum of TE0, TE1 and TM0 modes are 0.04°, 0.12° and 0.33°, respectively. Thus, we chose a TE0 mode in the present work.
A higher sensitivity of such sensors is already achieved, by increasing the surface of the detector. Porous anodic alumina films are commonly employed to enlarge the surface because of their well-ordered nanostructure self-organizes when aluminum films are anodized in an acid electrolyte causing pores at sufficient diameters. These pores run straight through the film parallel to each other. The pore sizes are dictated by a combination of the applied voltage used for the anodization and the solutions employed. For example, self-ordering takes place at 25 V in sulfuric acid solution, at 40 V in oxalic acid solution, and at 195 V in phosphoric acid solution, leading to pore sizes of 65 nm, 100 nm, and 500 nm, respectively . Disadvantage of such an approach is the impossibility to obtain pores with arbitrary diameters. It has been reported that porous aluminum films may be employed simultaneously as deposition templates and as highly sensitive detectors of processes within the pores [3, 4]. Also porous silicon has been used as a bio-sensing material .
We will report here on a strategy to perforate dielectric waveguides to enhance significantly the sensitivity of the sensor as firstly, the perturbation of the modes by adsorbed molecules is enlarged and secondly, their adhesion surface is increased. Silica coat is mechanically and environmentally stable. The holes in the insulator are created in the following way: the passage of an atomic particle through an insulator results in the creation of a cylindrical, latent nuclear track that extends along a straight line corresponding to the path of the atomic particle . The presence of tracks is revealed after etching in a suitable solution. However, the shape of the etched hole is often not cylindrical, but conical, with the largest diameter being near the surface. This is a consequence of the etching rate in the latent track being comparable to the general etching rate in the unirradiated region. Musket et al. have used vapor etching of ion tracks to create high aspect ratio, isolated cylindrical holes through ~600 nm-thick films of thermally grown silica on silicon . In the present study, we have adopted the vapor etching technique to create cylindrical holes through silica waveguides. Excitation of the surface plasmon and waveguide modes are experimentally observed as very sharp dips in reflectivity when measured as a function of the incident angle. Au must be needed, otherwise we cannot observe waveguide mode. The adhesion of streptavidin molecules onto the surface and into the holes inside the waveguide causes a strong shift in the measured angular position of the modes. We found a nine times increase of the shift for the perforated waveguides as compared to bulk waveguides at the expense of only slightly larger line widths. The line width of those modes is only negligible larger due to enhanced scattering at the surface of the holes increasing the guiding losses. As the diameter of the holes was in the order of a few tens of nanometers only, the modes experience effective material parameters for the waveguide material and will not interact sensitive on the fine details of the holes. Incorporation of streptavidin causes a significant change in the optical properties of the waveguide material.
As glass substrate and prism, we used S-LAH66 glass (Ohara Co., Ltd.) with a high refractive index of 1.769 . To ensure adhesion of a 50 nm Au layer on the substrate, first a ~0.8 nm Cr layer was deposited on the surface. On top of the gold film again a ~0.8 nm Cr layer was deposited by thermal evaporation (R-DEC Co., Ltd., Japan). Finally, a 500nm thick-waveguide layer of a-SiOx film was deposited on the Cr/Au/Cr films. To improve the optical performance of the film, the structure was tempered at 600 °C for 24 h. Rutherford backscattering measurement revealed the x in a-SiOx to be in the vicinity of 2. Thus, hereinafter we call the film silica.
Ion bombardment to perforate the waveguide was performed at room temperature at a residual pressure below 1 × 10-4 Pa. The 12 UD Pelletron tandem accelerator at the University of Tsukuba was employed with 150 MeV Au14+ ions. To avoid overlap of the individual cylindrical holes, very low ion fluency was required. This was achieved by diffusing the ion beam through an aluminum foil in forward scattering geometry. 150 MeV Au14+ was reduced to 137 MeV Au30+ by passage through the Al foil. We performed vapor etching experiments with 20% hydrofluoric acid (HF). Influence of various liquid TL and sample TS temperatures was examined and TL = 19 °C and TS = 30 °C were obtained as a best condition. Changing the temperature allowed for control over the hole sizes.
The substrate glass was mounted on a prism to form an optically contiguous medium. A He-Ne laser operating at 632.8 nm was used in subsequent measurements and a polarizer selected for either TE- or TM-polarization. We also clamped a Teflon flow cell onto the exposed film to permit injection and extraction of the desired dielectric medium/reagents. To observe the optical waveguide modes, reflectivity was measured at a constant temperature of 20.8°C against incident angle.
Figure 3 illustrates the chemical surface modification procedure to ensure adhesion of molecules. At first, the surface of silica was modified by silanol groups. It has been reported that the number of surface silanol groups is increased in an alkali solution . Thus, 3-aminopropyltriethoxysilane (3APT) was bound with a silanol groups by immersing the silica waveguide substrates in 0.5 vol% 3APT ethanol solution for 3 h. The 3APT-modified substrate was rinsed with ethanol and dried in a stream of nitrogen gas and then immersed in 0.5 mM 5-[5-(N-succinimidyloxycarbonyl)pentylamido]hexyl D-biotinamide (biotin-(AC5)2-OSu) solution in 1/15 M PBS buffer (pH 7.4) for amide-coupling with the 3APT layer. Biotin-(AC5)2-OSu) was used for streptavidin capture in the presence of phosphate buffer solution (PBS) .
3. Results and discussion
SEM images of the fabricated samples provided sufficient evidence that cylindrical holes were etched into the 500nm-thick silica films (Fig. 4.) These holes with a diameter of ~ 50 nm resulted from vapor etching followed by rinsing in distilled water and drying. The hole density was estimated to be 5.0 (± 0.8) ×108 cm-2 as deduced from SEM images, which is consistent with the experimentally counted fluence of 5.0 (± 0.2) ×108 cm-2 using a Faraday cup.
Waveguide modes can be excited in the silica waveguide. The waveguide mode spectrum shifts when the silica waveguide film is exposed to a solution containing biotin of 500 μM. Figure 5(a) shows reflectivity against incident angle in the vicinity of the TE0 mode obtained on the as-prepared silica film before (black circles) and after (red circles) immersion in biotin solution. The central line shift in the as-prepared sample case was less than 0.01° and merely noticeable. The angular resolution in this experiment was 0.01°. Figure 5(b) shows the reflectivity vs. angle around the same mode obtained on porous silica film with a number of 1 × 1010 cm-2 holes, 500nm deep and 50 nm in diameter. The black and red circles indicate the spectra before and after immersion in biotin solution. The central line shift was dramatically increased up to 0.14°. It was found that the remaining dip intensity was increased by the creation of holes and the line width was likewise increased slightly. Both observations are easily attributed to an increase of the guiding losses of the modes due to additional light scattering at the etched holes and also due to the chrome on the metal film and the APT modification of silica. Reduction of refractive index with increase of holes in waveguide introduced wider dips and much smaller contrast of in off- and on-resonance reflectance. Better spectrum can be obtained with controlling thickness of waveguide and metals.
Figure 6 shows angular dependent reflectivity for a TE0 mode before and after capturing streptavidin of 500 nM. For the silica waveguide before creation of pores, dip shift was estimated to be 0.06 °. The sample having 1×1010 holes/cm2 was fragile and it was not successful to measure the spectrum after biotin-streptavidin bindings. When a number of 7 × 109 cm-2 holes were created in the silica waveguide, central position of the dip shifted dramatically by 0.53 °. Materials may not reach the bottom of the hole but the vicinity of the maximum of field distributions of the guided modes that are concentrated inside the waveguide. Additional waveguides with increased hole densities were fabricated and the TE0 spectra were measured. The relationship between dip shift due to capturing of streptavidin and hole density was examined. Change in reflectivity at a constant angle (ΔR/R) could be estimated from the TE0 spectra. For example, reflectivity at 53.56° was estimated from Fig. 6(b) to be 0.51 before and 0.78 after capturing of streptavidin. Based on those values, we evaluated the relative change in reflectivity (ΔR/R) with respect to the unexposed sample. ΔR/R values are plotted against hole-density in Fig. 7. It was found that the dip shift increases with number of holes. The ΔR/R value reached a maximum above 3 × 109 cm-2. Saturation is observed, as the line width of the unexposed compared to the exposed spectra are no more overlapping. Such a shift is hardly to achieve with conventional plasmon based sensors.
Shift of SPR peak position due to biotin-streptavidin bindings was reported as 360ng/cm2 per 0.01° . When surface was filled with streptavidin molecules of 5nmϕ, mass of streptavidin can be estimated as 400ng/cm2, which is close to the value reported in . It is impossible to compare the displacement of the angle per biotin-streptavidin bindings between SPR and TE0 mode because SPR cannot be excited in the silica waveguide. Moreover, surface area of porous silica in the laser beam spot is 7.88cm2 which is much larger than 0.03cm2 of non-porous silica. The relative change in reflectivity (ΔR/R) of SPR and TE0 mode were calculated at 0.2% and 2% per 0.01° shift, respectively. From the calculation, it was firmed up that approach of the present work was better than SPR. We expect the sensor chips to be able to detect impurities of ppb order.
In the present experiment, a genuine configuration for the detection of low levels of biomolecules was explored that employs guided modes in perforated dielectric waveguides. It was found that a concentration of 500 μM of biotin, which cannot be detected employing the bulk silica waveguide, could be detected with the perforated silica waveguide. Also by capturing streptavidin of 500nM, the angular position of the waveguide modes for the TE0 was shifted up to 0.53. The shifts exceed significantly the line width of the resonance itself, hence allowing for unprecedented change in the reflectivity for such low concentration of molecules to be detected. The high sensitivity of the present sensor derives from a maximization of the overlap of modal volume and the molecules to be detected via the nanometer sized holes that were created in the waveguides. The present approach will contribute significantly towards the development of ultra-high sensitive sensors in various fields.
C. Rockstuhl, would like to thank the Canon-Foundation Europe for financial support.
References and links
1. E. Kretschmann, “Die Bestimmung optischer Konstanten von Metallen durch Oberflächenplasmaschwingungen,” Z. Physik 241, 313–324 (1971). [CrossRef]
2. H. Masuda, K. Yada, and A. Osaka, “Self-ordering of cell configuration of anodic porous alumina with large-size pores in phosphoric acid solution,” Jpn. J. Appl. Phys. 37, L1340–L1342 (1998). [CrossRef]
3. K. H. A. Lau, L. S. Tan, K. Tamada, M. S. Sander, and W. Knoll, “Highly sensitivive detection of processes occurring inside nanoporous anodic alumina templates: A waveguide optical study,” J. Phys. Chem. B 108, 10812–108181 (2004). [CrossRef]
4. S. G. Cloutier, A. D. Lazareck, and J. Xu, “Detection of nano-confined DNA using surface-plasmon enhanced fluorescence,” Appl. Phys. Lett. 88, 013904-1–013904-3 (2006). [CrossRef]
5. J. J. Saarinen, S. M. Weiss, P. M. Fauchet, and J. E. Sipe, “Optical sensor based on resonant porous silicon structures,” Opt. Express 10, 3754–3764 (2005). [CrossRef]
6. K. Awazu, S. Ishii, K. Shima, S. Roorda, and J. L. Brebner, “Structure of latent tracks created by swift heavy ion bombardment of amorphous SiO2,” Phys. Rev. B 62, 3689–3698 (2000). [CrossRef]
7. R. G. Musket, J. M. Yoshiyama, R. J. Contolini, and J. D. Porter, “Vapor etching of ion tracks in fused silica,” J. Appl. Phys. 91, 5760–5764 (2002). [CrossRef]
10. N. Fukuda, M. Fujimaki, K. Awazu, K. Tamada, and K. Yase, “High sensitive optical detection of bio-chemicals onto a silicon oxide surface based on waveguide mode,” Mater. Res. Symp. Proc. 900, O12-39.1–6 (2006).
11. Y. Terasaka, Y. Arima, and H. Iwata, “Surface plasmon resonance-based highly sensitive immunosensing for brain natriuretic peptide using nanobeads for signal amplification,” Anal. Biochem. 357, 208–215 (2006). [CrossRef]