Nanogap-based dielectric-specific colocalization for highly sensitive surface plasmon
resonance detection of biotin-streptavidin interactions
Yonghwi Kim, Kyungwha Chung, Wonju Lee, Dong Ha Kim, and Donghyun Kim
Citation: Applied Physics Letters 101, 233701 (2012); doi: 10.1063/1.4769108 View online: http://dx.doi.org/10.1063/1.4769108
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/101/23?ver=pdfcov Published by the AIP Publishing
Articles you may be interested in
Biotin-streptavidin detection with a graphene-oxide supported radio-frequency resonator Appl. Phys. Lett. 102, 193701 (2013); 10.1063/1.4802788
Silicon-based reproducible and active surface-enhanced Raman scattering substrates for sensitive, specific, and multiplex DNA detection
Appl. Phys. Lett. 100, 203104 (2012); 10.1063/1.3701731
Nanoparticle-based protein detection by optical shift of a resonant microcavity Appl. Phys. Lett. 99, 073701 (2011); 10.1063/1.3599706
Periodic nanohole arrays with shape-enhanced plasmon resonance as real-time biosensors Appl. Phys. Lett. 90, 243110 (2007); 10.1063/1.2747668
Detection of tumor markers based on extinction spectra of visible light passing through gold nanoholes Appl. Phys. Lett. 90, 073901 (2007); 10.1063/1.2535919
Nanogap-based dielectric-specific colocalization for highly sensitive surface
plasmon resonance detection of biotin-streptavidin interactions
Yonghwi Kim,1Kyungwha Chung,2Wonju Lee,1Dong Ha Kim,2and Donghyun Kim1,a)
1
School of Electrical and Electronic Engineering, Yonsei University, Seoul 120-749, South Korea 2
Department of Chemistry and Nano Science, Division of Molecular and Life Sciences, College of Natural Sciences, Ewha Womans University, Seoul 120-750, South Korea
(Received 20 July 2012; accepted 13 November 2012; published online 3 December 2012)
We have performed highly sensitive surface plasmon resonance (SPR) detection by colocalizing the evanescent near-fields and target molecular distribution. The colocalization is based on oblique metal evaporation to form nanogaps of a size under 100 nm without using electron-beam lithography. The concept was demonstrated by detecting siloxane-based biotin/streptavidin interactions. 50-nm nanogaps produced the largest amplification of optical signatures and two orders of magnitude enhancement of sensitivity over conventional thin film-based measurements. The enhancement is associated with efficient overlap of localized near-fields and target. Colocalized detection scheme is expected to provide clues to molecular sensitivity for SPR biosensing.VC 2012 American Institute of
Physics. [http://dx.doi.org/10.1063/1.4769108]
Surface plasmon (SP) is a longitudinal electron concen-tration wave that is created when incident transverse magnetic (TM) polarized light is momentum-matched to oscillating conduction electrons at a dielectric-metal interface. SP reso-nance (SPR) is sensitive to the surface condition, therefore, has been used as biosensors that measure biomolecular inter-actions label-free in real time. Because of the label-free nature, SPR biosensors suffer from moderate detection sensi-tivity typically quoted as 1 pg/mm2.1For improvement of sen-sitivity characteristics, numerous approaches have been attempted, such as nanoparticle-based amplification of optical signatures,2–4multimodal detection of molecular changes,5,6 measurement of phase variation,7–9and field amplification by nanostructures10–14and metamaterials.15
Recently, colocalization of target molecular distribution with localized near-fields was proposed16and experimentally confirmed based on inclined evaporation of a dielectric mask layer.17,18 In contrast to potentially alternative techniques such as dip-pen nanolithography, colocalization by inclined evaporation combined with standard lithographic techniques provides a simple way to reduce feature resolution and allows intrinsic self-alignment of target molecules with near-field distribution. The mask layer covers nanopatterned sur-face except for evaporation shadows. By thiolating probe biomolecules for preferential binding to the metal surface, target biointeractions can be made to be spatially selective such that they overlap field localization. The colocalization significantly increases overlap between target and field. The approach was extremely effective and produced significant sensitivity enhancement for the detection of DNA hybridiza-tion. However, one of the weaknesses of the approach was that probe biomolecules must be thiolated for binding to metal surface to maximize target-field overlap by preferen-tial metal binding. In other words, it may not be applicable if probe biomolecules cannot be thiolated.
In this work, we generalize the colocalization approach using dielectric-specific binding chemistry. The conceptual schematic is illustrated in Fig. 1. A lithographically defined dielectric surface pattern is first angle-evaporated with metal, which leaves a small fraction of the surface at the ridge exposed for biomolecules to interact. A potential concern in this configuration is that the fields localized and amplified by the evaporated metal film, or hot spots, may not fully overlap with probe molecules that adsorb to dielectric surface. In this regard, the size of dielectric surface and metal thickness are important to maximize target-field overlap. The dielectric-specific colocalization was tested and confirmed by detecting biotin/streptavidin interactions. The effectiveness of colocali-zation was compared to that of thin film-based measurements.
To ensure optimum overlap between hot spots and target molecules, we have evaluated a normalized overlap integral that is defined as Ðe(r)Ex(r)2dr. e(r) and Ex(r) denote the
spatial distribution of target permittivity and normalized tangential evanescent electric field in the lateral plane, respectively. The overlap integral was shown previously to correlate with plasmonic detection sensitivity measured in the far-field and to suggest colocalization strategies in the near-field and also qualitative trends of the structural effect
FIG. 1. Schematic structure for dielectric-specific colocalized SPR detection (K: grating period, g: nanogap size, heva: evaporation angle, da: grating
thickness, and db: evaporated metal thickness). Also shown is the direction
of light incidence (red arrow) and electric fields for TM polarization (gray arrow). Biotin and streptavidin molecules are not scaled in size.
a)Author to whom correspondence should be addressed. Electronic mail:
[email protected]. Tel.: 82-2-2123-2777. Fax: 82-2-313-2879.
0003-6951/2012/101(23)/233701/5/$30.00 101, 233701-1 VC2012 American Institute of Physics APPLIED PHYSICS LETTERS 101, 233701 (2012)
on far-field properties.19,20The near-field distribution formed by nanoscale structures has been calculated using rigorous coupled-wave analysis with periodic boundary condition. The calculation assumed uniform biotin/streptavidin distri-bution in the opening that is formed by angled evaporation. Streptavidin was modeled to form a 4.18-nm thick layer with refractive index n¼ 1.460, while an underlying biotinylated layer was assumed to be 0.84-nm thick with n¼ 1.624.21 Surface modification that underlies the biotinylated layer using (3-aminopropyl)triethoxysilane (APTES) was modeled as a 0.8-nm thick dielectric layer of n¼ 1.46 in the numerical analysis.22
The near-field distribution for g¼ 50 nm is presented in Fig.2. It is clearly observed that hot spots exist in the gap, although the secondary spot at the dielectric surface contrib-utes mainly to the target-field overlap. In principle, the pri-mary hot spot near metal surface does not contribute to the optical signature without any non-specific adsorption of binding molecules to the metal surface. A rounded metal sur-face, for example, may be more appropriate to remove or reduce the strong hot spots formed near the metal, although it is difficult to control the profile when metal layers are formed by angled evaporation. Calculation also suggests that the localized near-fields at hot spots are maximized at reso-nance condition (Supplementary Information 1 (Ref. 23)), although the plasmon momentum at maximum localized near-fields is known to differ from that of resonance condi-tion, because of increased damping associated with metallic nanostructures.24 This is associated with the nature of the current sample scheme which does not allow inherent propa-gating plasmon without underlying metal thin films.
For proof of principles, we have used two-dimensional grating structures for excitation and localization of near-fields. We have used photolithography to define long-pitch grating patterns. This is to avoid technical difficulties experi-enced when nanogaps are produced with nanoscale gratings and for potential mass-fabrication of nanopatterned samples without the need of electron-beam lithography. Initially, a cleaned SF10 glass substrate was spin-coated with AR-P 3120 positive photoresist (AllresistTM, Strausberg, Germany) at 2000 rpm and prebaked at 95C. Grating patterns were
defined at K¼ 8 lm and 50% fill factor. The grating sample was then evaporated with a 2-nm thick chrome adhesion layer and a gold film. The film thickness (da) was fixed at
da¼ 50 nm based on the numerical simulation (see
Supple-mentary Information 2) and the ease of fabrication of small nanogaps. Removal of photoresist for lift-off leaves a metal-lic grating pattern. This is followed by the angled deposition of 2-nm thick chrome and 40-nm thick gold for colocaliza-tion of target molecules. As a result, a small gap is created at the side of metallic grating ridges through which molecules can access underlying dielectric surface of the glass sub-strate. Evaporation angle (heva) was adjusted between 52
and 70 to produce various nanogap sizes (g) such that g¼ 30, 50, and 90 nm. The actual nanogap size was meas-ured to be 33 6 2.9 nm, 56 6 4.0 nm, and 96 6 7.6 nm, which is about 10% larger than designed. A SEM image of a fabri-cated nanograting structure with g¼ 50 nm is presented in Fig.3(a)and clearly shows the nanogap. Independent profile measurements using a phase-contrast AFM (Nanoscope Dimension 3100, Digital Instruments, Santa Barbara, CA, USA) confirmed the results, as shown in Fig.3(b).
For SPR detection, we have used an SPR spectrometer (RT2005, Resonant Technol. GmbH, Germany). The spec-trometer is based on conventional h-2h architecture. For measurement, TM-polarized light from a He-Ne laser (k¼ 632.8 nm, 10 mW) was incident on a nanograting sample index-matched to an SF10 prism substrate. Nominal angular resolution was 0.005. Reflected light was measured by a pho-todiode. Each measurement was repeated at least three times and averaged. The overall sensitivity of the SPR spectrometer system was estimated as 10 pg/mm2. Because of the long gra-ting pitch, higher diffraction orders may exist. Gragra-ting equa-tion stipulates that higher order diffracequa-tion angles at resonance be given by hm¼ arcsin(mk/K þ sin hSPR), where m is an
inte-ger representing diffractive orders. For the range of experi-mental resonance angles, the minimum angular difference between reflected order (m¼ 0) and higher orders was approx-imately 9, which is sufficiently large to remove the noise associated with higher-order diffraction from the photodiode.
The surface chemistry used for nanogap-based colocali-zation of biotin/streptavidin interactions is illustrated in Sup-plementary Information 3. First, nanograting samples on SF10 substrates were made hydrophilic under O2 plasma and
immersed in the APTES solution (10 vol. %, ethanol, Sigma-Aldrich, MO, USA) for 1 h. Ethanol and DI water were used to
FIG. 2. Near-field intensity distribution (jExj 2
) formed of a nanogap struc-ture at g¼ 50 nm for the lateral dimension of 1 lm. The calculation was per-formed for the whole period K¼ 8 lm. The range of x and z axis is 300 and 150 nm, respectively.
FIG. 3. (a) SEM and (b) AFM image of grating-based nanogap (g¼ 50 nm). Inset in (a) shows the surface profile following the line shown in the SEM image. Scale bar in the inset denotes 100 nm both in lateral and axial dimension.
remove unreacted APTES. After positioning the APTES-grating sample on a prism mounted on a flow cell, 1 mM biotin 3-sulfo-N-hydroxysuccinimide ester sodium salt (biotin-sulfo-NHS, Sigma-Aldrich) in phosphate buffered saline (PBS, Sigma-Aldrich) solution was injected through a flow cell for 1 h. Samples were then rinsed in PBS solution. Streptavidin so-lution with varied concentrations (fromStreptomyces avidinii, lyophilized from 10 mM potassium phosphate, Sigma-Aldrich) was injected for 1 h, followed by rinsing with PBS solution. All the protocols were carried out at room temperature. Biotin-sulfo-NHS, streptavidin, and PBS solutions were delivered by a peristaltic pump (0.5 ml/min). Rinsing with PBS buffer was held for 10 min to eliminate improperly reacted chemicals after each injection.
Control measurements were performed on a 40-nm thick gold thin film. The film was initially immersed in cysteamine 1 mM solution in ethanol for 24 h. Subsequently, rinsing was performed sufficiently with ethanol and DI water, followed by drying with nitrogen stream. After positioning the cysteamine-Au thin film on SPR spectrometer, injection and rinsing of biotin-sulfo-NHS and streptavidin solution were performed like the preceding steps.
For evaluation of the relative number of biotin/streptavidin bindings, we have also measured thin film control and nanogap-based binding of 10-nm gold nanoparticle-conjugated streptavidin on biotin. Figure4(a)shows a SEM image of ran-domly distributed nanoparticles, each of which represents an approximate single biotin/streptavidin binding event. Particle counting suggests 18 nanoparticles/lm2. In the case of nano-gap, we have used a slightly larger nanogap sample with g¼ 150 nm for facilitated observation of nanoparticles, as shown in Fig.4(b). Counting results in 85 nanoparticles/lm2in the gap area. In other words, the nanogap structure has 4.7 times (¼85/18) denser biotin/streptavidin bindings per unit area, or the density ratio (DR) between nanogap and control is equal to 4.7. The difference in the density is mainly due to the surface chemistry involved in the immobilization that is siloxane-based in the case of nanogap vs. thiolated for the control using metal thin films, because siloxane-based immobi-lization was known to show improved chemical stability due to covalent bonds in the APTES layer that produce highly or-dered and well-packed monolayers.25Also, for siloxane-based immobilization, less steric hindrance can potentially produce increased optical signatures,26 due to much longer molecular distances of APTES than those of cysteamine.27,28
Figure5presents resonance angle shifts (DhSPR) that are
produced by biotin/streptavidin interactions and measured
with nanogap-based dielectric-specific colocalization. In Fig.5(a), a larger angle shift was observed as more interac-tions occur with higher streptavidin concentration. The net angle shift was rearranged as a function of nanogap size in Fig. 5(b). From the measured data presented in Fig. 5, the largest optical signatures were obtained at g¼ 50 nm. A neg-ative control experiment using a 50-nm nanogap without bio-tin suggests that non-specific binding due to electrostatic physi-sorption between streptavidin and metal surface and also between streptavidin and APTES layer may exist within 15%, which agrees well with previous results.29If we define nominal sensitivity enhancement (NSE) as the ratio of the resonance angle shift by nanogap-based colocalization to that of thin film detection, i.e., NSE¼ DhSPR(nanogap)/
DhSPR(thin film), NSE¼ 5.0 at 1 lM of streptavidin
concen-tration for g¼ 50 nm. We emphasize that the nanogap-based colocalized measurements produced much larger resonance shifts than thin film detection without localization, even though colocalized detection involves much less interactions that contribute to the resonance shift. While it is very diffi-cult to quantify the number of interactions, the number is proportionate to the first degree to the surface area of biotin that is available for streptavidin to bind to.30In other words,
FIG. 4. SEM images of nanoparticles used to mediate biotin/streptavidin binding: (a) control image on a thin film and (b) for a nanogap at g¼ 150 nm. The nanogap image was contrast-enhanced for clear observation of nanoparticles in the nanogap.
FIG. 5. Measured resonance angle shifts: (a) target streptavidin concentra-tion is varied for various nanogap sizes (g¼ 30, 50, and 90 nm) in compari-son with the results of thin film-based control. The measured data were fit with exponential saturation functions. (b) Resonance angle shifts as gap size is varied show the existence of an optimum gap size at which the resonance shift reaches a maximum. The measured data were fit with log-normal distri-bution functions. Inset figures are the near-field distridistri-bution created by nano-gaps: A, B, and C for g¼ 30, 50, and 90 nm, respectively. Color scale coincides with that of Fig.2.
the number ratio (NR) of the bindings for colocalized detec-tion to that of non-colocalized detecdetec-tion is approximately equal to the product of DN(nanogap)/DN(thin film) g/K and the density ratio DR. Here, DN represents the number of bindings that participate in either case. For g¼ 50 nm, NR¼ 4.7/160 1/34, i.e., 34 times less biotin/streptavidin binding events participate in the resonance measured in Fig.5. The net enhancement of molecular signatures per unit number of bindings is then equal to NSE/NR¼ 170 at 1 lM of streptavidin concentration. The enhancement depends on the target streptavidin concentration, e.g., the net enhance-ment is approximately 71 times at 10 nM. The enhanceenhance-ment of optical signatures is approximately in line with what was obtained with thiolated colocalization on metal surface,16 where improvement by two orders of magnitude was observed for DNA hybridization using 3 lM target single-stranded DNA. In other words, experimental results using nanogap devices indicate that more than 100-fold sensitivity enhancement should be attainable based on colocalized detection at nanogaps for a range of molecular interactions. This exceeds what may be achieved by nanoparticle-based amplification of optical signatures.
Figure 5(b) shows that the resonance shift peaks at g¼ 50 nm for the range of target streptavidin concentrations used in the experiment. At a larger gap size, SP is less local-ized in the nanogap while secondary hot spots disappear, and detection by colocalization in this case approaches that of thin films. Nanogap with g¼ 30 nm produced hot spots simi-lar in intensity to that of g¼ 50 nm. However, a smaller area at which molecular interactions occur tends to limit the measured optical signatures. If the gap size decreases further, the nanogap weakens the hot spot in the near-field patterns significantly and reduces the overlap between hot spots and target molecules and likewise the detection sensitivity.
Also worth a note are the advantages of nanogaps asso-ciated with a long pitch (K) when nanogaps are fabricated by photolithography. For SPR detection in which SP is localized by short-pitched nanostructures, resonance char-acteristic curves tend to broaden so significantly that the broadening may degrade detection sensitivity.31In contrast, the use of long-pitched nanogaps using photolithography appears to allow the broadening to be minimal, which is related to low metal damping loss. Also observed was that as an interaction proceeds, resonance dips become slightly deeper with smaller reflectance at resonance. Figure 6
shows the curve angular widths (CAWs) of the resonance dip, which we have defined as the angular full width at the average of reflectance minimum and maximum. Measured CAW was 15.1, 16.6, and 15.6for nanogaps with g¼ 30, 50, and 90 nm. For comparison, it was 12.2 for thin film-based detection. Although the angular broadening does exist, the broadening caused by the nanogaps is much smaller than we typically see with shorter pitches. This implies that the enhancement of sensitivity we have seen by way of colocalization can be well maintained without deg-radation due to angular broadening.
In summary, we have fabricated nanogaps by angled evaporation over photographically formed metallic grating. SPR detection of biotin/streptavidin interactions colocalized to plasmonic near-fields was demonstrated to induce
signifi-cant enhancement of molecular signatures by more than two orders of magnitude. We have seen an optimum nanogap that produces the largest resonance shift in conjunction with the overlap between target molecules and near-field distribu-tion. We expect the colocalized detection scheme to be highly useful to enhancing the detection sensitivity of SPR detection on the order of fg/mm2and eventually to achieving molecular sensitivity in SPR biosensing.
This work was supported by the National Research Foundation (NRF) grants funded by the Korean Government (2009-0070732, 2010-0007993, 0017500, and 2011-0029409).
1
C. T. Campbell and G. Kim,Biomaterials28, 2380 (2007).
2
L. He, M. D. Musick, S. R. Nicewarner, F. G. Salinas, S. J. Benkovic, M. J. Natan, and C. D. Keating,J. Am. Chem. Soc.122, 9071 (2000).
3S. Moon, D. J. Kim, K. Kim, D. Kim, H. Lee, K. Lee, and S. Haam,Appl.
Opt.49, 484 (2010).
4
S. Moon, Y. Kim, Y. Oh, H. Lee, H. C. Kim, K. Lee, and D. Kim,Biosens. Bioelectron.32, 141 (2012).
5B. Sepulveda, A. Calle, L. M. Lechuga, and G. Armelles,Opt. Lett.
31, 1085 (2006).
6
J. Oh, Y. W. Chang, S. Yoo, D. J. Kim, S. Im, Y. J. Park, D. Kim, and K.-H. Yoo,Nano Lett.10, 2755 (2010).
7S. G. Nelson, K. S. Johnston, and S. S. Yee,Sens. Actuators B
35, 187 (1996).
8
S. Y. Wu, H. P. Ho, W. C. Law, C. Lin, and S. K. Kong,Opt. Lett.29, 2378 (2004).
9A. R. Halpern, Y. Chen, R. M. Corn, and D. Kim,Anal. Chem.
83, 2801 (2011).
10
K. M. Byun, S. J. Kim, and D. Kim,Opt. Express13, 3737 (2005).
11K. Kim, S. J. Yoon, and D. Kim,Opt. Express14, 12419 (2006). 12K. M. Byun, S. J. Yoon, D. Kim, and S. J. Kim,Opt. Lett.
32, 1902 (2007).
13
L. Malic, B. Cui, T. Veres, and M. Tabrizian,Opt. Lett.32, 3092 (2007).
14
X. D. Hoa, A. G. Kirk, and M. Tabrizian,Biosens. Bioelectron.24, 3043 (2009).
15A. V. Kabashin, P. Evans, S. Pastkovsky, W. Hendren, G. A. Wurtz, R.
Atkinson, R. Pollard, V. A. Podolskiy, and A. V. Zayats,Nature Mater.8, 867 (2009).
16K. M. Byun, S. M. Jang, S. J. Kim, and D. Kim,J. Opt. Soc. Am. A26,
1027 (2009).
17
K. Ma, D. J. Kim, K. Kim, S. Moon, and D. Kim, IEEE J. Sel. Top. Quantum Electron.16, 1004 (2010).
18Y. Oh, W. Lee, and D. Kim,Opt. Lett.36, 1353 (2011). 19A. Shalabney and I. Abdulhalim,Sens. Actuators A
159, 24 (2010).
20
W. Lee and D. Kim,J. Opt. Soc. Am. A29, 1367 (2012).
FIG. 6. Measured resonance characteristic curves as the biotin/streptavidin interaction proceeds with nanogap-based colocalization (g¼ 30, 50, and 90 nm) and thin film-based detection. Streptavidin concentration is 50 nM. Inset shows curve angular widths measured for each case.
21H. Morgan, D. M. Taylor, and C. D’Silva, Thin Solid Films
209, 122 (1992).
22
H. Ouyang, C. C. Striemer, and P. M. Fauchet, Appl. Phys. Lett. 88, 163108 (2006).
23See supplementary material at http://dx.doi.org/10.1063/1.4769108 for
near-field characteristics with angle of incidence and grating depth and detailed surface chemistry of biotin/streptavidin interactions.
24K. Kim, Y. Oh, K. Ma, E. Sim, and D. Kim, Opt. Lett. 34, 3905
(2009).
25
C. Haensch, S. Hoeppener, and U. S. Schubert,Chem. Soc. Rev.39, 2323 (2010).
26K. V. Gobi, K. Matsumoto, K. Toko, H. Ikezaki, and N. Miura, Anal.
Bioanal. Chem.387, 2727 (2007).
27
A. Ulman,Chem. Rev.96, 1533 (1996).
28M. Hepel and J. Dallas,Sensors8, 7224 (2008).
29E. H. Williams, A. V. Davydov, A. Motayed, S. G. Sundaresan, P. Bocchini,
L. J. Richter, G. Stan, K. Steffens, R. Zangmeister, J. A. Schreifels, and M. V. Rao,Appl. Surf. Sci.258, 6056 (2012).
30J. Korlach, P. J. Marks, R. L. Cicero, J. J. Gray, D. L. Murphy, D. B.
Roitman, T. T. Pham, G. A. Otto, M. Foquet, and S. W. Turner, Proc. Natl. Acad. Sci.105, 1176 (2008).
31
