Interference Localized Surface Plasmon ... - ACS Publications

DOI: 10.1021/ac902008x. Publication Date (Web): January 14, 2010. Copyright © 2010 American Chemical Society. * To whom correspondence should be addr...
0 downloads 11 Views 5MB Size
Download PDF
Anal. Chem. 2010, 82, 1221–1227

Interference Localized Surface Plasmon Resonance Nanosensor Tailored for the Detection of Specific Biomolecular Interactions Ha Minh Hiep, Hiroyuki Yoshikawa, and Eiichi Tamiya* Department of Applied Physics, Graduate School of Engineering, Osaka University, Suita, Osaka 565-0871, Japan In this paper, we present an innovative sensing nanomaterial for an interference localized surface plasmon resonance (iLSPR) sensor. The iLSPR is based on plasmonic gold nanoparticles with photonic thin-film multilayers of porous aluminum oxide (Al2O3) and aluminum (Al) on a substrate. With a controllable transparent Al2O3 layer and a highly reflective Al layer, our new nanomaterial was able to detect refractive index (RI) changes of the surrounding environment and the specific interaction of biomolecules including biotin and avidin, 5- fluorouracil (5-FU) and its antibody, anti-5-fluorouracil (anti 5-FU), when the iLSPR surfaces were biologically functionalized. Our model nanostructure will open the way to display the plasmonic properties of other noble metal nanoparticles and to develop other functionally similar nanosensors, which could then be expanded into multiarrays. The unique optical characteristics of nanoparticles from noble metals, such as gold and silver, have long fascinated scientific researchers.1-3 These characteristics are attributed to the collective oscillations of free electrons in metal nanoparticles surrounded by a dielectric media, a phenomenon known as localized surface plasmon resonance (LSPR). LSPR, which is measured through an extinction spectrum (scattering and absorption) of these nanoparticles is depended on the size, shape and refractive index (RI) of the surrounding environment.4-9 Considerable advances have been made in LSPR-based nanodevices that integrate photonic, physical, chemical, and biological properties to bring new functionality to future chemical and biological sensing * To whom correspondence should be addressed. E-mail: tamiya@ ap.eng.osaka-u.ac.jp. Telephone number: +81-6-6879-4087. Fax number: +81-66879-7840. (1) Hao, E.; Li, S. Y.; Bailey, R. C.; Zou, S. L.; Schatz, G. C.; Hupp, J. T. J. Phys. Chem. B 2004, 108 (4), 1224–1229. (2) Park, T. H.; Mirin, N.; Lassiter, J.; Hafner, J.; Halas, N. J.; Nordlander, P. ACS Nano 2008, 2 (1), 25–32. (3) Shoute, L. C.; Bergren, A. J.; Mahmoud, A. M.; Harris, K. D.; McCreery, R. L. Appl. Spectrosc. 2009, 63 (2), 133–140. (4) Anker, J. N.; Hall, W. P.; Lyandres, O.; Shah, N. C.; Zhao, J.; Van Duyne, R. P. Nat. Mater. 2008, 7 (6), 442–453. (5) Endo, T.; Kerman, K.; Nagatani, N.; Ha, M. H.; Kim, D. K.; Yonezawa, Y.; Nakano, K.; Tamiya, E. Anal. Chem. 2006, 78 (18), 6465–6475. (6) Thomas, K. G.; Barazzouk, S.; Ipe, B. I.; Joseph, S. T. S.; Kamat, P. V. J. Phys. Chem. B 2004, 108 (35), 13066–13068. (7) Ha, M. H.; Endo, T.; Saito, M.; Chikae, M.; Kim, D. K.; Yamamura, S.; Takamura, Y.; Tamiya, E. Anal. Chem. 2008, 80 (6), 1859–1864. (8) Wu, L. Y.; Ross, B. M.; Lee, L. P. Nano Lett. 2009, 9 (5), 1956–1961. (9) Zhang, J. Z.; Noguez, C. Plasmonics. 2008, 3 (4), 127–150. 10.1021/ac902008x  2010 American Chemical Society Published on Web 01/14/2010

nanoapparatuses.10-13 However, the LSPR signal from these devices is a broad spectral band with weak intensity, and thus they are useful for only for a limited number of biomolecules. To widen the sensing applications of individual plasmonic devices, we developed a three-layer interference LSPR substrate constructed with plasmonic gold nanoparticles and photonic thinfilm multilayers of silicon dioxide/silicon on a substrate.14 The term iLSPR (interference localized surface plasmon resonance) defined by our group indicates its advanced combination of plasmonic metal nanoparticles and interference thin-film multilayers. By monitoring the relative reflectance value of the plasmon band of iLSPR substrates, we were able to observe biotin-avidin binding events in real time with a detection limit of 1 µg/mL.14 Coupling the plasmon band with interference bands enhanced changes in the LSPR band that surrounds media with different refractive indexes, leading to increased sensor sensitivity. The sensitivity of the LSPR sensor depends on the adsorbed molecules on its surface and can be attributed to the intensity of the electromagnetic (EM) field localized near to the nanoparticle. It has been experimentally and theoretically shown that anisotropic nanoparticles like ellipsoids, nanorods, or nanoparticle arrays generate intense EM fields.15,16 The LSPR bands of these anisotropic nanoparticles basically split into two oscillation modes, a transverse and a longitudinal mode. Remarkably, the strength and peak wavelength of the longitudinal LSPR band undergo larger spectral changes than the LSPR band of spherical nanoparticles when the surrounding RI is altered.17-20 (10) Lee, S.; Mayer, K. M.; Hafner, J. H. Anal. Chem. 2009, 81 (11), 4450– 4455. (11) Gordon, R.; Sinton, D.; Kavanagh, K. L.; Brolo, A. G. Acc. Chem. Res. 2008, 41 (8), 1049–1057. (12) Lee, S.; Mayer, K. M.; Hafner, J. H. Anal. Chem. 2009, 81 (11), 4450– 4455. (13) Ha, M. H.; Endo, T.; Kim, D. K.; Tamiya, E. Proceedings of SPIE. 2007, 6768, 67680I1–67680I11. (14) Ha, M. H.; Yoshikawa, H.; Saito, M.; Tamiya, E. ACS Nano 2009, 3 (2), 446–452. (15) Tanaka, Y.; Yoshikawa, H.; Itoh, T.; Ishikawa, M. J. Phys. Chem. C 2009, 113 (27), 11856–11860. (16) Itoh, T.; Yoshikawa, H.; Yoshida, K.; Biju, V.; Ishikawa, M. J. Chem. Phys. 2009, 130 (21), 214706. (17) Mayer, K. M.; Lee, S.; Liao, H.; Rostro, B. C.; Fuentes, A.; Scully, P. T.; Nehl, C. L.; Hafner, J. H. ACS Nano 2008, 2 (4), 687–692. (18) Miller, M. M.; Lazarides, A. A. J. Phys. Chem. B. 2005, 109 (46), 21556– 21565. (19) Wu, C.; Xu, Q. H. Langmuir 2009, 25 (16), 9441–9446. (20) Ha, M. H.; Endo, T.; Kerman, K.; Chikae, M.; Kim, D. K.; Yamamura, S.; Takamura, Y.; Tamiya, E. Sci. Tech. Adv. Mater. 2007, 8 (4), 331–338.

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1221

In this paper, we report the creation of an innovative iLSPR sensor fabricated by immobilizing gold nanoparticles on the surface of a porous aluminum oxide (Al2O3) layer formed on an aluminum (Al) substrate. We verified that a large surface area of such nanoporous aluminum oxide templates, with wellresolved Fabry-Perot fringes in their optical properties, could be used as a highly sensitive sensor to detect the substantial change in RI that occurs upon the binding of analyte molecules.21-24 The thickness of the transparent Al2O3 layer could be controlled by adjusting electrochemical anodization time.21,25 Therefore, we consider that the reflection spectra of the iLSPR substrates are dependent upon the thickness of the Al2O3 layer. The noteworthy results in this paper show that the porous nanostructure partially trapped two nanoparticles, producing a small number of dimer nanoparticles on the substrate surface. Spectral bands attributed to the longitudinal plasmon of the dimers were found at the suitable Al2O3 thickness. These experimental reflection spectra of our iLSPR substrates were clarified by a theoretical simulation. Moreover, we validated an applicable immunosensor system based on our iLSPR substrate by examining the specific interactions of 5fluorouracil (5-FU) and anti-5-FU. EXPERIMENTAL SECTION Al2O3 Substrate Fabrication. To prepare the aluminum porous layer on the aluminum substrate (Al2O3/Al), we developed a two-step anodizing method.21 First, a 3 µm thick aluminum layer on a glass substrate was subjected to a constant voltage of 40 V in a 0.3 M aqueous oxalic acid solution for 30 min. The anodic temperature was kept constant at 0 °C, and the electrolyte was vigorously stirred during the process in order to maintain the appropriate temperature as well as electrolyte concentration. Next, the generated aluminum oxide layer was removed by immersing the sample in a solution containing a mixture of phosphoric (1.8%, w/v) and chromic (2%, w/v) acids for 15 min. The second step was then repeated under the same conditions for times ranging from 5-8 min. This resulted in aluminum oxide layers with various thicknesses, ranging from 215 to 280 nm. This process for forming aluminum oxide layers on aluminum in different thicknesses by changing the time for the second anodizing step has been previously reported.21,25 Gold Nanoparticles Immobilization. The amount of gold nanoparticles (50 nm in diameter, British Biocell, U.K.) that can be directly deposited on the Al2O3/Al substrate is very low. For that reason, the surface of Al2O3/Al substrate was first modified with cationic polymer by immersing it in an aqueous poly(allyl amine) solution (2% wt % in H2O) for 2 min to generate positive charges on the surface. Gold nanoparticles were then dispersed as a monolayer on the modified poly(allyl amine)-modified precursor substrate by absorbing for 24 h, and the monolayer was heated under vacuum conditions for 2 h. The bonding of the gold nanoparticles to the surface was strong enough to (21) Kim, D. K.; Kerman, K.; Saito, M.; Sathuluri, R. R.; Endo, T.; Yamamura, S.; Kwon, Y. S.; Tamiya, E. Anal. Chem. 2007, 79 (5), 1855–1864. (22) Pan, S.; Rothberg, L. J. Nano Lett. 2003, 3 (6), 811–814. (23) Kim, D. K.; Kerman, K.; Yamamura, S.; Kwon, Y. S.; Takamura, Y.; Tamiya, E. Jpn. J. Appl. Phys. 2008, 47 (2), 1351–1354. (24) Kim, D. K.; Kerman, K.; Ha, M. H.; Saito, M.; Yamamura, S.; Takamura, Y.; Kwon, Y. S.; Tamiya, E. Anal. Biochem. 2008, 379 (1), 1–7. (25) Masuda, H.; Satoh, M. Jpn. J. Appl. Phys. 1996, 35 (1B), 126–129.

1222

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

resist detachment during later chemical modifications. The surface of the immobilized gold nanoparticles layer on the Al2O3/Al substrate was imaged using a scanning electron microscope (SEM) (FEI Company, DB 235). Characteristic reflection spectra were observed with an optical system consisting of a spectrophotometer (USB-4000-UV-vis), a tungsten halogen light source (LS-1), and an optical fiber probe (R-400-7 UV-vis, fiber core diameter ) 200 µm), all purchased from Ocean Optics. All samples were analyzed in the wavelength range of 400-800 nm at room temperature. RESULTS AND DISCUSSIONS Immobilization of Gold Nanoparticles on Al2O3 Surface. As illustrated in Figure 1 poly(allyl amine), a water-soluble cationic polymer, was used to modify the Al2O3 surface, providing the cationic character to selectively attract negatively charged gold nanoparticles as a result of electrostatic interactions. Scanning electron microscope (SEM) image analysis revealed that treating aluminum oxide surface with poly(allyl amine) is an efficient way to fix gold nanoparticles on a tender Al2O3 layer. The dependence of nanoparticle density on various poly(allyl amine) concentrations, and SEM images illustrating these relations, are shown in Supporting Information (Figure SI1). The sufficient number of gold nanoparticles displayed on the Al2O3 surface gives us a standpoint from which to analyze further biomolecular interaction. As we can see in Figure 1b, among the single immobilized nanoparticles, some dimers of two gold nanoparticles trapped in single holes promise to generate a sensitive longitudinal LSPR band of gold nanoparticles. Dependency of Reflection Spectra of the iLSPR Substrate on the Thickness of the Al2O3 Layer. Theoretically, it should be possible to finely regulate the optical spectra of the iLSPR substrates once the plasmonic spectra of gold nanoparticles are enhanced by the interference spectra of the Al2O3/Al multilayer. Successive thicknesses of Al2O3 in the range from 215 to 280 nm on aluminum substrates were prepared. Figure 2 shows the appearance and coupling of additional plasmonic peaks of gold nanoparticles observed in the reflection spectra of six different types of iLSPR substrates. Two bands (at ∼520 nm and ∼650 nm) were observed and their relative intensities depended on the Al2O3 thickness. When the Al2O3 thickness increased, the second LSPR band appeared growing around 650 nm. These spectral shapes with two bands resemble those of gold nanoparticle dimers which show two LSPR bands as mentioned above, transverse and longitudinal modes. From SEM images, we estimated the ratio of monomers and dimers for the substrates with Al2O3 thicknesses in a range from 215 to 280 nm (Figure 2). Fascinatingly, the respective percentages of gold nanoparticle dimers were almost constant, around 10%, for all substrates. This fact seems to conflict with the increase of the relative depth of the second LSPR band with Al2O3 thickness. Reflectance spectra of the bare Al2O3/Al substrate without gold nanoparticles (Figure 2, solid lines) present a single weak dark-band due to the interference of light reflected at the top surface and bottom interface between Al2O3 and Al. Together with the increase in the Al2O3 thickness, the interference bands shift into the longer wavelength region, from 435 to 565 nm. Thus, there is a strong correlation between the LSPR and interference bands in Figure

Figure 1. (a) Optical system for measuring the reflectance spectra of our substrates. (b) Fabrication of iLSPR substrate. SEM images of iLSPR substrate surface of aluminum porous surface before and after immobilization of gold nanoparticles.

Figure 2. Experimental reflection spectra of iLSPR substrates with various Al2O3 thicknesses (a-f) and their percentages of monomer and dimer nanoparticles (insets).

2 (dotted lines). This suggests that the longitudinal LSPR band is significantly enhanced by the thin-film interference in sensors

with Al2O3 thicknesses from 270 to 280 nm (Figure 2). Reflectance spectra of the thin film multilayers were calculated Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1223

Figure 3. Dependency of reflection spectra of the iLSPR substrate (containing gold nanoparticles of 50 nm) on the thickness of the Al2O3 layer in numerical investigation (a-f). The scheme of the multilayer iLSPR substrate (g) and the dependency of percentage of absorbed light energy by the top gold nanoparticles layer on alumina thickness (h).

using the characteristic matrix method.26 The effective RI of the Au nanoparticle layer was approximated by the Maxwell-Garnett theory.27 To simplify the simulation analysis, instead of using nanoparticle dimers, a small percentage of prolate ellipsoids were mixed in the calculation model. Because both the prolate ellipsoids and dimers have similar space symmetry, their LSPRs possess the longitudinal modes that have plasmon oscillations along the major axis of nanoparticles. The effective RI of the mixed media of spherical and ellipsoidal nanoparticles in our simple calculation was found by the procedure described in references28 and.29 Figure 3 shows the reflection spectra calculated at different Al2O3 thicknesses. The aspect ratio of the semilong axis to the semishort axis and the volume fraction of ellipsoids were set (26) MacLeod, H. A. Thin-Film Optical Filters; Taylor and Francis: London, 2001. (27) Kreibig, U.; Vollmer, M. Optical Properties of Metal Clusters; Springer-Verlag: Berlin, 1995. (28) Bohren, C. F.; Huffman, D. R. The Absorption and Scattering of Light by Mall Particles; Wiley-Interscience: New York, 1998. (29) McMillan, B. G.; Berlouis, L. E. A.; Cruickshank, F. R.; Brevet, P. F. J. Electroanal. Chem. 2007, 599 (2), 177–182.

1224

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

to 3.5 and 2 × 10-4 respectively, as the wavelength and its intensity of longitudinal mode matched the experimental results (in Supporting Information). Although the LSPR band of monomers and the longitudinal band of dimers had similar peak wavelengths, the plasmon bands shown at ∼520 nm in Figure 2 and Figure 3a-f can be ascribed to the monomer particles, since the percentage of dimers was very small both in the experiment and the simulation. Except for the small difference in the reflectance and bandwidth of the plasmon band, the calculation well reproduced the spectral shape and its dependence on Al2O3 thicknesses. There is a strong correlation between two reflection spectra, the spectra of bare (Al2O3/Al) substrates and corresponding iLSPR substrates. The reflectance spectral shape of a bare substrate indicated an interference effect between two light waves reflected at the porous alumina surface and the Al2O3/Al interface. The calculated results strongly suggest that the LSPR band is enhanced by this interference effect. In the iLSPR substrate, the incident light went through the gold nanoparticles layer propagated on the porous alumina layer,

Figure 4. Experimental evaluation of changes in the LSPR band of iLSPR substrates with an Al2O3 thickness of 270 nm when they were surrounded by air (RI ) 1.000) and by water (RI ) 1.333).

reached the Al substrate, and was reflected backward. The incident and reflected light waves formed a standing wave in the porous alumina layer.30 At the particular conditions of light wavelength, optical constant, and thickness of each layer, the intensity of the standing wave was maximized in the gold nanoparticle layers, thus gold nanoparticles were exposed by the intense light field. In the thickness range of the alumina layers used for this study, the wavelength satisfying such conditions shifted from short to long, following the increase of the alumina thickness. Therefore the monomer LSPR was strongly excited in the case of a thin alumina layer, whereas the longitudinal band appeared and grew as the thickness of the alumina layer increased. Here we explained the concept of our sensor and role of interference using another simulation result. Figure 3h shows the dependency of percentage of absorbed light energy by the top gold nanoparticles layer on the alumina thickness.26 These percentage values correspond to the light intensity of standing wave in the alumina layer. Increasing the thickness of the alumina layer, the light intensity irradiated on the gold nanoparticles layer repetitively increase and decrease due to the interference of counter propagating lights in the alumina layer. On the other hand, the phase of the standing wave in Al2O3 layer also depends on the wavelength. This figure shows that the light intensity at the wavelength of the dimer band (610 nm, broken line) increases with the alumina thickness in the range of 215-280 nm, while the monomer band decreases (520 nm, solid line). Therefore, in this report, by choosing an appropriate alumina thickness around 270 nm we can focus and prominently enhance the signal from the dimer band. Refractive Index Dependency. For a better understanding of the advantages created by the longitudinal LSPR band in our iLSPR substrates, we measured the reflection spectra of iLSPR substrates with Al2O3 thicknesses of 270 nm in both air and water (Figure 4). The appearance of a red shift in the peak wavelength as a function of the RIs of the surrounding medium from 1.000 to 1.333 was consistent with the prediction obtained from Mie theory.31,32 The longitudinal LSPR band of the iLSPR (30) Shoute, L. C. T.; Bergren, A. J.; Mahmoud, A. M.; Harris, K. D.; McCreery, R. L. Appl. Spectrosc. 2009, 63 (2), 133–140. (31) Prasad, P. N. Introduction to Biophotonics; Wiley-Interscience: New York, 2003. (32) Marinakos, S. M.; Chen, S.; Chilkoti, A. Anal. Chem. 2007, 79 (14), 5278– 5283.

substrate showed an impressively high red-shift of ∼75 nm. It equates to a bulk refractive index sensitivity of approximately 225 nm/RIU, which is exactly predicted by Miller and Lazarides.18 This strongly indicates the highly sensitive capacity of these iLSPR substrates and their usefulness for further biosensing applications. iLSPR-Based Biosensor with Biotin-Avidin Interactions. Inspired by the above data and by the prospect of improving the sensitivity of the iLSPR sensor, we prepared iLSPR substrates with an Al2O3 thickness of 270 nm and tested their biosensing capacity with a biotin-avidin interactions model. 11-amino-1undecanethiol hydrochloride of 1 mM was initially formed on the substrate surface to present terminal amino groups that were subsequently able to react with a solution of sulfosuccinimidyl 1-D-diotin of 100 µg/mL via an ester link. A phosphate buffer saline solution (PBS, 50 mM, pH 7.4) containing 10 µg/ mL bovine serum albumin (BSA) was then applied to suppress nonspecific adsorption.33-35 We investigated binding interactions at this interface using two different target molecules, avidin and BSA, by which we determined the functionality of both iLSPR sensors. Figure 5b-c illustrates the spectra of the iLSPR substrates before and after incubation with avidin and BSA at concentrations of 10 µg/mL for 30 min at room temperature. Due to the increase in the local RI at the iLSPR substrate interface caused by the specific reaction of biotin and avidin, a significant change from 650 to 750 nm, especially in the second LSPR band was observed. This change was quantified by the reflection intensity differences between 2 spectra in the wavelength region of 650 and 750 nm, 4R650-750. Under similar conditions, however, a slight response was detected in the presence of 10 µg/mL BSA, with minor amplitude fluctuation. We continuously observed the 4R650-750 increments as various avidin concentrations were independently introduced onto the biotinylated surface of the iLSPR sensor. Each concentration was repeated three times. Slight 4R650-750 increments in the buffer solution (without avidin) were caused by physical bindings, not by specific avidin-biotin interactions. With higher concentrations of avidin, 4R650-750, denoting the interactions of avidin with biotinylated surface, increased constantly. That means the amount of bound avidin was directly related to the avidin concentrations. The mean increases in 4R650-750 for binding the various avidin concentrations of 0, 10, 100, 1000, and 10000 ng/mL were 0.385, 0.413, 0.650, 0.856, and 1.750, respectively, as shown in Figure 5d. The limit of detection of 100 ng/mL avidin was determined without further optimization. For comparison, iLSPR substrates with an Al2O3 thickness of 225 nm were also used to quantify the avidin concentrations with a similar procedure (Figure 5e). As expected, the higher enhancement by the coupling of thin-film interference and the longitudinal LSPR that occurred in the iLSPR substrates with a 270 nm thickness of Al2O3 resulted in higher sensitivity compared to the other substrates. Label-Free Screening of Anti 5-FU Molecules. We took a further step to propose an immunosensors based on our iLSPR nanostructure with Al2O3 thicknesses of 270 nm, for more (33) Huang, T. T.; Sturgis, J.; Gomez, R.; Geng, T.; Bashir, R.; Bhunia, A. K.; Robinson, J. P.; Ladisch, M. R. Biotechnol. Bioeng. 2003, 81 (5), 618–624. (34) Peterfi, Z.; Kocsis, B. J. Immunoassay 2000, 21 (4), 341–354. (35) Paulsson, M.; Kober, M.; Freij-Larsson, C.; Sollenwerk, M.; Wesslen, B.; Ljungh, A. Biomaterials 1993, 14 (11), 845–853.

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1225

Figure 5. Diagram of the iLSPR-based biosensor with its functionalized surface (a). The reflection spectra changes of biotin-avidin (b) biotinBSA (c) interactions and the dependence of 4R650-750 on various avidin concentrations with iLSPR sensors having Al2O3 thickness of 270 nm (d) and 225 nm (e).

reliable biomolecular interactions, specifically, 5-FU molecules and anti 5-FU. 5-FU, a pyrimidine analogue, is usually used as a drug in the treatment of cancers because it can inhibit an enzyme called thymidylate synthetase, resulting in inhibition of the DNA replication process.36 Accordingly, in the field of immunological assays of 5-FU and anti-5-FU, many scientists have attempted to develop ways to rapidly determine optimal 5-FU concentrations37,38 or to screen anti-5-FU molecules from the hybridroma cells.39,40 To our knowledge, only the ELISA (Enzyme-Linked ImmunoSorbent Assay) method, a labeling technology, is currently used to detect anti-5-FU since 5-FU molecules are fixed onto the well surface.39 However, several problems remain with this method; specifically, it requires extra time, imposes higher cost, and influences the activity of proteins.41 In this paper, we describe for the first time a labeling-free, simple, rapid detection method for the sensitive screening anti-5-FU using our iLSPR sensors. 5-Fluorouracil (36) Han, R.; Yang, Y. M.; Dietrich, J.; Luebke, A.; Mayer-Proschel, M.; Noble, M. J Biol. 2008, 7 (4), 12.112.22. (37) Falk, R. E.; Hardy, M.; Makowka, L.; Teodorczyk-Injeyan, J.; Falk, J. A. J. Clin. Invest. 1982, 70, 558–567. (38) Salamone, S. J.; Courtney, J. B.; Stocker, D. 5-Fluoro-uracil immunoassay. U.S. Patent 7205116, April 17, 2007. (39) Anti-5-Fluorouracil; Data sheet No 0812150103; Mabel Inc.: Kyoto, Japan, 2008; htttp://www.mabel.co.jp/cgi-bin/fs/wiki.cgi?action)ATTACH&page) DataSheet&file)DataSheet-MAB-PL-5FU_H3-17_Rev0902A.pdf (accessed Aug 10, 2008). (40) Honda, T.; Inagawa, H.; Fukushima, M.; Moriyama, A.; Soma, G. I. Clin. Chim. Acta 2002, 322 (1-2), 59–66. (41) Petach, H.; Gold, L. Curr. Opin. Biotechnol. 2002, 13 (4), 309–314.

1226

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

Bovine Serum Albumin complex (5-FU-BSA, 100 ng/mL, MABEL Inc., Japan) was primarily immobilized on the iLSPR substrate for 1 h. Then 2% fetus bovine serum (FBS, Sigma Aldrich, USA) was added to block the nonspecific adsorptions. The 5-FU-BSA modified surface was constructed and used to screen a series of anti 5-FU concentrations (MABEL Inc., Japan) from 0 ng/nL to 10000 ng/mL. The 4R650-750 changes were recorded depending on the anti 5-FU concentrations, as shown in Figure 6. The iLSPR sensors provided a limit of detection of 10 ng/mL, similar to the sensitivity obtained from recent ELISA results.39 In the negative samples, instead of 5-FUBSA, 100 ng/mL BSA was immobilized on the iLSPR surface, followed by the addition of 10 000 ng/mL of anti 5-FU under the same experimental conditions. The obviously smaller increments of 4R650-750 indicate that the stringent washing and the blocking solution treatment were enough to suppress the nonspecific reactions in our immunosensors (shown in Figure 6d). Without any labeling steps, our method is a promising candidate for high-throughput, low-cost and highly sensitive detection of multiple analytes in a simple and rapid format.5 CONCLUSIONS Using an Al2O3/Al substrate with gold nanoparticles and photonic thin-film multilayers, we have developed a sensitive sensing system specifically for biomolecular interactions. The key to our advanced system is the successful design of an

Figure 6. Diagram of the iLSPR based biosensor for anti- 5-FU screening with specific (in positive samples) and nonspecific (in negative sample) reactions (a). 5-FLUOROURACIL molecules (b) and the dependence of 4R650-750 on various anti 5-FU antigen concentrations with iLSPR sensor having 270 nm Al2O3 thicknesses (c). The changes of 4R650-750 values with the positive samples (specific reactions of 100 ng/mL 5-FU-BSA with 10 000 ng/mL anti-5-FU) and negative samples (nonspecific reactions of 100 ng/mL BSA with 10 000 ng/mL anti-5-FU) (d).

iLSPR nanostructure constructed with a top layer of immobilized gold nanoparticles, a middle layer of controllable transparent Al2O3, and a bottom layer of aluminum. In our iLSPR design, the multiple interactions of incident light and gold nanoparticles can exist because the combination of the high reflection Al substrate and the gold-nanoparticle layer makes up a Fabry-Perot resonator. Thus, a considerable enhancement can be occurred when a suitable matching condition between LSPR and interference spectra are established. Although gold nanoparticles were used in this report as a model for the localized surface plasmon excitation, other noble metal nanoparticles and nanostructures could also be used on the iLSPR substrate for their plasmonic properties. Owing to the controlled thickness of the Al2O3 layer and the simple use of poly(allyl amine) polymer for nanoparticle immobilization, we were able to achieve a sensitivity significantly greater than that of our original iLSPR substrate (14). The highly sensitive immunosensor of 5-FU and anti-5-FU was

used as a proof-of-concept in this study; this approach should be generalized for detecting almost all other protein markers. ACKNOWLEDGMENT The authors thank Dr. Masato Saito and Dr. Kenji Kondou from Osaka University for valuable advice during the preparations of our manuscript. H. M. Hiep expresses thanks for a postdoctoral fellowship from the Japan Society for the Promotion of Science (JSPS). NOTE ADDED AFTER ASAP PUBLICATION This paper was published on the Web on January 14, 2010, with the incorrect journal name in reference 16. The corrected version was reposted on January 22, 2010. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review December 31, 2009.

September

5,

2009.

Accepted

AC902008X

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1227