Article pubs.acs.org/ac
Sensing with Prism-Based Near-Infrared Surface Plasmon Resonance Spectroscopy on Nanohole Array Platforms Laurel L. Kegel, Devon Boyne, and Karl S. Booksh* Department of Chemistry and Biochemistry, University of Delaware, Newark, Delaware 19716, United States S Supporting Information *
ABSTRACT: Nanohole arrays exhibit unique surface plasmon resonance (SPR) characteristics according to hole periodicity, diameter, and excitation wavelength (λSPR). This contribution investigates the SPR characteristics and surface sensitivity of various nanohole arrays with the aim of tuning the parameters for optimal sensing capability. Both the Bragg surface plasmons (SPs) arising from diffraction by the periodic holes and the traditional propagating SPs are characterized with emphasis on sensing capability of the propagating SPs. Several trends in bulk sensitivity and penetration depth were established, and the surface sensitivity was calculated from bulk sensitivity and penetration depth of the SPs for different analyte thicknesses. Increased accuracy and precision in penetration depth values were achieved by incorporating adsorbate effects on substrate permittivity. The optimal nanohole array conditions for highest surface sensitivity were determined (820 nm periodicity, 0.27 diameter/periodicity, and λSPR = 1550 nm), which demonstrated an increase in surface sensitivity for the 10 nm analyte over continuous gold films at their optimal λSPR (1300 nm) and conventional visible λSPR (700 nm).
S
complicated analysis and optical setup.1 Thereby, this contribution seeks to combine the increased sensing capability of nanohole arrays and extension of SPR to higher excitation wavelengths within the near-infrared region. Surface plasmon resonance active nanohole arrays have demonstrated increased biosensing capability over conventional continuous gold films in the visible wavelength region,11,19 and they have other applications in surface enhanced Raman spectroscopy,20,21 optoelectronics,22 and waveguiding.23,24 This contribution investigates the sensing properties of SPR on nanohole arrays and fundamental surface plasmon (SP) characteristics, which may provide insight into other applications as well. The sensing capability arises from SP sensitivity to adsorbates adhering to the substrate surface within the shallow penetration depth (ld) region of the SP, that is, the depth at which the SP electric field intensity decays to 1/e (Figure 1). The nanohole array structures support several SP modes which exhibit variable properties. The SP modes include SPs similar to nanostructures (localized SPR or LSPR) and continuous films (propagating SPR) as well as short-range SP modes.25−29 LSPR is supported within the nanoholes, and propagating SPs are excited in the semicontinuous regions of gold between the rows.25,27,29,30 Additionally, shorter range plasmons are excited by incident light that is diffracted into the sample plane according to the Bragg resonance order, and thus they are called Bragg SPs.31,32 The SP modes may be excited in different optical modes, namely, Kretschmann (Figure 1) and transmission configurations; however,
urface plasmon resonance (SPR) spectroscopy can detect low concentrations of nanoscale analytes, such as proteins, for use in various biosensing applications. The advantage of this technique is high surface sensitivity arising from the evanescent nature of the surface plasmon extending from the sensor− sample interface. For biosensing, the analyte is bound to the sensor surface for detection. Analyte binding and selectivity is achieved by preattached biorecognition elements, which are generally antibodies or peptides. Abilities to discriminate sample refractive index (RI) changes as small as 10−6 to 10−7 refractive index units (RIU) have been reported, which is the equivalent to the RI shift from a 0.01 to 0.001 °C temperature change of water.1−3 These detection limits correspond to diagnostic protein concentration levels,4 but capabilities could improve from increased sensitivity with the ability to discriminate small changes in concentration, detection of proteins with poor binding affinity to biorecognition elements, and detection of even lower concentration biomarkers.1,5−7 Nanostructuring SPR substrates may increase SPR sensitivity, thus biosensors based on plasmonic nanostructures have been widely investigated.8−10 A comparison of nanohole arrays to thin film SPR biosensors demonstrated a marked improvement in surface sensitivity in the visible wavelength region.11 However, little difference compared to continuous gold films is observed for several other nanostructured SPR biosensors, including nanotriangles,12 nanodisks,13 and nanorods.14 Other methods have been investigated for improving sensitivity including near-infrared excitation wavelengths,15,16 longrange surface plasmons (LRSP),17,18 and phase modulation measurements.1 Of these methods, further investigation of supporting dielectrics for LRSP is necessary to overcome reduced SPR imaging capability and surface sensitivity,17,18 and phase modulation SPR spectroscopy involves interferometry with more © 2014 American Chemical Society
Received: October 31, 2013 Accepted: February 4, 2014 Published: February 5, 2014 3355
dx.doi.org/10.1021/ac4035218 | Anal. Chem. 2014, 86, 3355−3364
Analytical Chemistry
Article
where m is the bulk sensitivity. For full surface coverage, Va/Vs may be calculated by integrating over the respective distances within the exponentially decaying plasmonic field (Figure 1). The sensing volume, Vs, extends from the gold surface (d = 0 nm) to the penetration depth (d = ld),37 and the analyte volume spans from the binding distance (d = d1) to the top of the bound analyte (d = d2) such that msurf = m
the propagating SP in Kretschmann mode has exhibited greater sensitivity than Bragg SPs in transmission mode.11 Additionally, Kretschmann configuration enables real-time measurements in highly scattering samples, such as biological fluids. The high sensitivity of propagating plasmons may be further increased by interaction with the other SP modes;33,34 therefore, this study utilizes Kretschmann configuration and focuses on the sensitivity of the propagating SP and potential interacting effects of other SP modes. Nanohole array SPs may be tailored to control sensitivity, field enhancement regions, and interaction between coexcited SPs.33,34 The SPR conditions depend on the physical parameters of the nanohole array, translating to tunability of the Bragg SPs with nanohole array structure.31,35 Additionally, SPR sensitivity changes with excitation wavelength (λSPR).6,15,36 In this manner, measurements with λSPR in the near-infrared region (NIR) exhibit greater bulk sensitivity than the conventionally used visible region.6,15,16,36 However, sensitivity to bulk RI changes throughout the entire plasmonic field (i.e., bulk sensitivity) does not exactly translate to surface sensitivity. The working principle of SPR sensing is sensitivity to near surface interactions, so it is the surface sensitivity to bound analyte that is of particular interest. High surface sensitivity is achieved when a greater proportion of the sensing volume is occupied by the adsorbate layer rather than the bulk solution.37 Furthermore, it is optimal to nearly fill the sensing volume with the analyte in order to preclude contributions from bulk environmental RI fluctuation. Consequently, sensitivity to RI changes within the thin analyte adsorption layer (surface sensitivity) is a better indicator of sensor performance than is the sensitivity to RI changes throughout the entire plasmonic field (bulk sensitivity). The surface sensitivity depends on the proportion of the sensing volume that is occupied by the analyte, so surface sensitivity, msurf, to an analyte of specific volume, Va, residing within the sensing volume, Vs, may be calculated by38 Va Vs
(2)
(Details of this derivation are provided in the Supporting Information.) By evaluating the msurf merit for a specific analyte thickness, sensing parameters may be optimized. Another figure of merit for sensor capability is resolution, which depends on msurf as well as SPR dip sharpness, represented by dip height/ full width at half-maximum (fwhm), and system noise. An increase of 60% in dip height/fwhm has been previously observed for nanohole arrays compared to continuous gold in the visible region;11 however, the negative effects of peak broadening on resolution can be ameliorated by utilizing multivariate calibration techniques rather than standard minimum hunting techniques.39 Furthermore, the resolution is specific to the instrumental setup (in this case, 1 nm for continuous gold); thereby, this contribution focuses on the analysis of surface sensitivity as the primary assessment of SPR substrate sensing capability. The surface sensitivity depends on Va/Vs and bulk sensitivity, and factors which contribute to these terms affect the sensing capability and should be optimized. The SP penetration depth directly affects the sensing volume and may be tailored to specific analyte sizes.38,40 It also dictates the region of field enhancement for surface and tip enhanced spectroscopic techniques. Both penetration depth and bulk sensitivity depend on SPR wavelength. Accordingly, surface sensitivity may be optimized by tuning SPR wavelength, penetration depth, and the interplay between the two. In addition to changing wavelength, the use of different nanostructured sensing substrates offers variable SP characteristics.41 Characterization of the SP features is important to comprehensively exploit their interaction and tunability. This contribution presents the fundamental SP characterization of SPs and optimization of various nanohole array parameters for greatest surface sensitivity, and it is the first study, to our knowledge, of nanohole arrays in Kretschmann configuration in the NIR region.
Figure 1. Diagram of propagating surface plasmon on nanohole array in Kretschmann configuration with adsorbate volume (d1 to d2) and sensing volume (gold interface to ld) illustrated. The refractive index (RI) sensed by the plasmonic field is an effective RI composed of the adsorbate RI (ηa) and bulk RI (ηs).
msurf = m
exp( −2d 2/ld) − exp( −2d1/ld) exp( −2) − 1
■
EXPERIMENTAL METHODS SPR Spectroscopy Measurements. Spectroscopic measurements were acquired on a Bruker Optics (Billerica, MA) Vertex 70 FT-IR spectrometer with an IR-SPR accessory, as previously described by Menegazzo et al.16 Briefly, the halogen broadband NIR source within the spectrometer is directed through a series of lenses and mirrors to a BK-7 dove prism (10 mm H × 10 mm W × 42.40 mm L, Thorlabs, Newton, NJ) with the sample coupled to the topside. The reflected beam is then directed through a zinc selenide polarizer (Edmund Optics, Barrington, NJ) and another lens. SPR spectra are analyzed as a ratio of the reflected intensity of p-polarized to s-polarized light. The system employs a constant angle setup in which the response is measured as a shift in wavelength (ΔλSPR) at a fixed angle. The λSPR is determined by fitting a second-order polynomial to SPR dip and solving for the zero of the derivative of the curve. The shift in wavelength is relative to
(1) 3356
dx.doi.org/10.1021/ac4035218 | Anal. Chem. 2014, 86, 3355−3364
Analytical Chemistry
Article
λSPR for the bare, nonfunctionalized substrate in water. Initial λSPR for various samples ranged from 1070 to 2050 nm. Gold Nanohole Array Sensor Surface Preparation. Rectangular glass coverslips (Electron Microscopy Sciences, Ft. Washington, PA, 10.5 × 35 mm) were cleaned by immersion in boiling piranha solution (3:1 (v/v) concentrated H2SO4/30% H2O2) (Fisher Scientific, Fair Lawn, NJ) for at least 1 h, followed by triplicate rinsing and sonication in deionized water for 1 min. The slides were sonicated in a 5:1:1 (v/v) mixture of water, hydrogen peroxide, and ammonium hydroxide and then washed with deionized water. A nanosphere lithography process42−44 was utilized to produce nanohole array substrates with varying periodicity (P) and diameter (D). To investigate effects of periodicity and diameter, sets of various periodicity (490−3000 nm) with the same diameter to periodicity ratio (D/P) (0.36 ± 0.02) and sets of various D/P (0.27−0.44) with the same periodicity (820 nm) were fabricated (refer to Figure S-1). The samples are ordered over several millimeters, with no dislocations or grain boundaries over ≈20 μm2. The optical spot size is approximately 4 mm2, so measurements are representative of several azimuthal orientations. The gold was 78 ± 5 nm thick. Samples were kept under vacuum prior to metallization until use. Refer to Supporting Information for more procedural and structural details. Sensitivity Measurements. Bulk sensitivity (m = ΔλSPR/Δη) was calculated from the response to RI calibration standards of varying sucrose (≥99.5%, Sigma-Aldrich, St. Louis, MO) concentrations (0−3%) prepared in deionized water at room temperature (≈22 °C). This concentration range covered a SPR wavelength shift similar to the adsorption process investigated and remained in the linear response range. Wavelength-dependent values for RI of the sucrose solutions are not readily available in the literature. Alternatively, η of the glucose solutions determined by effective medium approximation may be used, because sucrose and glucose exhibit negligible (