1. Introduction

Nanostructured noble metals enabling surface plasmon confinements at specific wavelengths have been extensively studied to apply their enhanced light-matter interaction in biological imaging probes, in photocatalysts, and as plasmonic sensors [1–10]. Concurrently, periodic arrays of uniform nanostructures such as one-dimensional grating or two-dimensional mesh have become of increasing interest in the use of optical and biological filters [11–14]. In this letter, we apply a one-dimensional array of double-bent gold strips as a localized surface plasmon resonance (LSPR) sensor and also as a size-filtering device to achieve size-dependent detection sensitivity of spherical particles. Sizing and detection of spherical analytes with sizes of a few hundred nanometers, such as influenza virus or extracellular vesicles, are vital to protect and understand our immune systems [15,16].

Recently, we developed plasmonic nanoarchitectures comprised of a double-bent Au strip (DAS) array integrated on a one-dimensional transparent nanograting framework [17]. The DAS array was fabricated by two simple steps of roll-to-roll nanoimprint lithography and vacuum deposition of Au. By evaporating Au with an oblique angle on an imprinted nanograting surface, it is possible to shape a planar Au film into double-bent Au strips. The double-bent nanostructure enables the lengthening of the oscillation path of surface plasmons in the limited space of a 200 nm-period grating. A longer oscillation length under a longer excitation wavelength is advantageous to enhance refractive index sensitivity of LSPR sensor. The refractive index sensitivities of Au nanostructures is known to increase linearly with their resonance wavelengths as long as the real part of the dielectric function of Au changes linearly with the incident light wavelength [18,19]. For the DAS array, the refractive index sensitivity and its figure of merit (refractive index sensitivity per spectral width of the absorbance peak) were measured to be about 210 nm RIU⁻¹ and 4.2, respectively [17]. By utilizing the DAS array, here, a plasmonic sensing platform with a size-dendent detection sensitivity has been demonstrated with 100-nm or 200-nm polystyrene (PS) beads as model analytes.

2. Experimental

2.1 Fabrication of double-bent Au strip

By nanoimprint lithography, one-dimensional nanograting pattern (200 nm-period, 100-nm linewidth, 100-nm height) on polydimethylsiloxane (PDMS) mold was transferred onto the polyurethane acrylate (PUA)-coated polyethylene terephthalate (PET) substrate. Then, Au film of 30-nm thickness was evaporated with a tilted angle of 35° from the surface normal direction.

2.2 Electromagnetic simulation

Electromagnetic simulations were performed using three-dimensional finite-difference time-domain software (Lumerical FDTD solution 8.9). Simulation mesh was set to a 1 nm cubic grid. Perfectly matched layer and periodic boundaries were used as boundary conditions for vertical (z) and lateral directions (x, y), respectively. The periodicity of the modeled structure was set to 400 nm for both x and y directions. The direction of incident light is from top to bottom with the electric field polarized in the x direction. Refractive indices of the bead and substrate were assumed as 1.6 and 1.5, respectively.

3. Results and discussion

Figure 1 shows representative plan-view SEM images of the DAS array, upon which PS beads were drop-casted. Aqueous solutions of PS beads purchased from Thermo Fisher Scientific Inc. (3000 Series Nanosphere Size Standard) were dried in a covered petri dish to slow the evaporation of water. It is known that a capillary force, which is induced at three-phase (liquid-vapor-substrate) contact line, can push the nanoparticles in a colloidal solution into nanostructured grooves during solvent evaporation [20,21]. Likewise, both 100-nm and 200-nm beads are well-assembled at the valleys between neighboring DAS structures as shown in Fig. 1. Although the two types of beads are placed at the same lateral position, their vertical positions are quite different due to their size difference. The smaller and larger beads are respectively positioned inside and on the upper side of the valleys of the DAS array. It should be noted that the different vertical positions of the beads differentiate their influence on the local plasmonic fields developed around the DAS array.

 

Fig. 1 Plan-view SEM images of the (a, c) 100-nm and (b, d) 200-nm PS beads which were drop-casted on the surface of the double-bent Au strip array. Figure 1(c) and 1(d) are magnified images of Fig. 1(a) and 1(b), respectively.

Download Full Size | PDF

To assess the effects of the bead sizes on plasmonic sensing response, three-dimensional finite-difference time-domain simulation was used. The color contours of |E/E_o|² shown in Fig. 2(a–c), where E_o and E represent the amplitudes of the incident and enhanced electric fields, respectively, show the locally enhanced electric fields around the DAS structures. Because of the near-field coupling between neighboring DASs, the enhanced fields reach out to the inner space between the DASs as confirmed in Fig. 2(a). Therefore, 100-nm PS beads positioned inside the valley can be sensed by the local plasmonic fields of the DAS array more efficiently than 200-nm beads positioned on the upper side of the valley as shown in Fig. 2(b,c). In other words, the refractive index inside the valley, where the plasmonic field is locally enhanced, is effectively modified when the beads are smaller than the valley width. This change of local refractive index makes the change of resonance condition of the DASs, and can be observed by the red-shift in extinction spectra. Calculated extinction cross sections in Fig. 2(d) demonstrate that 100-nm beads cause a shift of the LSPR peak approximately five times larger than 200-nm beads, under the assumption that the PS beads are close packed along the valley. As shown in the inset graph in Fig. 2(d), the peak shifts calculated from four-different sized PS beads (60, 80, 100, and 200 nm) demonstrate that the refractive index modification in a valley by filling the beads is critical to determine the detection sensitivity. To affect the resonance condition of the DASs, 60-nm and 200-nm beads are too small and too large, respectively.

 

Fig. 2 (a–c) Squared amplitudes of the local electric fields around (a) the DAS structures, (b) 100-nm PS beads located inside of a valley between the DAS structures and (c) 200-nm PS beads located at an the upper side of a valley between the DAS structures. Scale bars are 100 nm. (d) Calculated extinction cross sections of (black) the DAS structures, (gray) the DAS structures with 200-nm PS beads and (red) the DAS structures with 100-nm PS beads. Inset shows the peak shifts calculated from four-different sized PS beads sitting on the DAS structure. For the simulation of the peak shifts, the PS beads were assumed to be close packed along the valley (i.e. 60, 80, 100, and 200 nm single bead per 60, 80, 100, and 200 nm periodicity, respectively).

Download Full Size | PDF

Size-dependent detection sensitivity based on the DAS arrays was experimentally confirmed with 100-nm and 200-nm PS beads. A 1 μL drop of PS bead solution with a concentration of 10⁻³, 10⁻², 10⁻¹, 1, or 10² g L⁻¹ was evaporated on a 5-mm square sensor surface. Figure 3(a,b) shows absorbance curves measured by UV-Vis spectrophotometer (Shimadzu, UV-2600). For the case of 100-nm beads, the LSPR peaks of the DAS sensor are clearly red-shifted with an increase in mass of applied beads. As confirmed from the enlarged view (inset of Fig. 3(a)), the peak shift is visible even at 1 ng. 10 μg of 100-nm beads induce a peak shift of approximately 30 nm. In contrast, in the case of 200-nm beads, it is difficult for to see any peak shift (Fig. 3(b)). As summarized in Fig. 3(c), the current design of the DAS sensor is much more sensitive to the 100-nm-sized analytes than the larger ones. Furthermore, it is possible to adjust the geometries of the DAS sensor such as width or depth of the valley for the characteristic shape of target analytes.

 

Fig. 3 (a,b) Absorbance curves measured from the DAS sensors treated with different concentrations of (a) 100-nm and (b) 200-nm PS beads. A 1 μL drop of PS bead solution with a concentration of 10⁻³, 10⁻², 10⁻¹, 1, or 10² g L⁻¹ was evaporated on the sensor surface. The inset graphs are enlarged views of the LSPR peaks. (c) Average LSPR peak shifts measured from the DAS sensors treated with different concentrations of (red) 100-nm and (black) 200-nm PS beads.

Download Full Size | PDF

4. Conclusion

In summary, the results in this letter demonstrate that an array of double-bent Au strips, which can be fabricated on a large-scale transparent substrate by two simple steps of nanoimprinting and vacuum deposition of Au, serves as a plasmonic sensor providing size-dependent detection sensitivity. The enhanced plasmonic field developed inside the valley between adjacent DAS structures is effectively disturbed by analytes having a size similar to the valley width, and consequently causes a large shift of the LSPR peak. Because the design of the DAS array can be customized to target analytes such as viruses and exosomes, it is thought that the proposed plasmonic sensor can be a useful platform for detecting particles of specific geometries.