1. Introduction

Metallic nanostructures, which support surface plasmon polaritons (SPPs) or localized surface plasmons (LSPs), are commonly employed to enhance near-field intensity for strong light-matter interaction [1]. Optical signals can be significantly improved by several orders of magnitude [2] to meet the requirements of single-molecule detection by exploiting deliberately designed plasmonic nanostructures [3,4]. In the past few decades, numerous plasmonic nanostructures, which can be roughly classified into disordered [5–8] and ordered [9–12] nanostructures, have been proposed to enhance optical signals. A typical example of disordered nanostructures is nanoparticle clusters in which local electric field enhancement is achieved at the “hot spots” generated in the nanogaps between nanoparticles. However, the practical applications of such disordered nanostructures are limited by poor reproducibility and controllability. In contrast, ordered nanostructures, which are generally fabricated by using the combination of electron beam lithography (EBL) and focused ion beam (FIB) etching, exhibit great advantages in repeatability and controllability [9,10,13]. Moreover, the wavelengths for electric field enhancements can be manipulated through the coupling of SPPs and LSPs generated in ordered nanostructures [14–17], making it possible to selectively enhancing useful optical signals at specific wavelengths [18].

Basically, plasmonic nanostructures function as optical antennas capable of receiving and radiating optical signals. For this reason, plasmonic nanostructures supporting dual LSPs are highly desirable. Recently, a novel approach based on the formation of Fano resonances in plasmonic nanostructures was proposed for Raman signal enhancement [5,6]. A Fano resonance is generated by a dark mode with a narrow linewidth and a bright mode with a broad linewidth and characterized by an asymmetric spectrum [19]. So far, guided by the idea of coherent coupling of the dark mode and bright mode, Fano resonances formed in various plasmonic nanostructures, such as nanoparticles [20,21], oligomers [22–25], rings [26–28], x-type structures [29], dolmen structures [30,31], have been widely used in SERS [32], sensing [33–35], optical converters [22], and other applications. The physical mechanisms for the formation of Fano resonances in plasmonic nanostructures have been systematically investigated and summarized in previous studies [30,36–38]. It is remarkable that a much stronger Raman signal could be achieved by using a heptamer if the excitation light was chosen at the Fano dip while the Raman Stokes signal was set at the Fano peak [32]. However, when the dip and peak of a Fano resonance generated in a plasmonic nanostructure are utilized to simultaneously enhance the excitation of laser light and the radiation of optical signals, the resonances of the Fano dip and peak can hardly be tuned independently via adjusting the parameters of the plasmonic nanostructure [39]. This feature strongly limits the practical applications of the Fano resonance, especially in the detection of broadband optical signals.

Nanoantennas with double resonances have recently attracted much attention and studied intensively for convenient wavelengths tunability and their multiple boosts in excitation and detection signals. By the coupling of LSP and different photonic modes simultaneously, multiple resonances can be achieved. One typical example is to place a dimer in the center of multi-2D gratings with different periods and directions [40]. However, the superposition of multi gratings around the dimer limits their adjustments in periods and result in a small wavelength tuning range for these resonances. Recently, metasurfaces based on the coupling between LSP and SPP have been frequently utilized to realize multi-resonances [41–43]. Nevertheless, their tunning coverages of the resonances are limited or discontinuous and have not been systematically fulfilled. Limited by the commercial lasers and object, investigations on nanoantennas which simultaneously offer broadband tunable, high collection efficiency and multi-boost are still needed.

Here in this paper, we both numerically and experimentally demonstrate an optical antenna with multiple surface plasmon resonances, which simultaneously improves the near-field intensity, far-field collection efficiency and tunability at two different wavelengths. These three functions are achieved by adding a gold nanoring to enclosing the tetramer with Fano resonance. By coupling dipole resonance induced by the gold nanoring and Fano resonance from the tetramer, multi-boosts and directional radiation can be realized. In particular, the multi-surface plasmon resonances can be independently and widely tuned through geometrical parameters adjustment, making the proposed nanoantenna a flexible platform on which different commercial lasers can be employed and different light-mater interactions can be investigated. The optical properties of the nanoantennas are verified by numerical simulations and SERS experiments. Therefore, the proposed optical nanoantenna opens new horizons for designing plasmonic devices from the perspective of simultaneous enhancements in both near- and far-field functionalities, which can be applied in a wide range of applications such as surface-enhanced spectroscopy, photoluminescence, nonlinear nanomaterials/emitters and biomedicine sensing.

2. Methods

To analyze the mechanism and demonstrate the performance of the proposed nanoantenna, we carried out methods including numerical simulation, analytical model fitting, nano-fabrication and optical experiments.

2.1 Electromagnetic simulation

The numerical simulations were carried out by using commercial software based on the finite-difference time-domain technique (FDTD Solutions). The volume of the simulated region was chosen to be 2000×2000×2000 nm³ while the minimum mesh size was set to 0.25 nm. The symmetric and asymmetric boundary conditions were used, depending on the polarization direction of the incident light. The simulated region was enclosed by a perfectly matched layer. The dielectric function of gold was taken from Johnson and Christy [44]. The refractive index of the environment was set to be 1.0. The refractive index of the 185-nm-thick indium tin oxide (ITO) thin film was chosen to be 1.63, which was the average value of eleven measurements taken by a Stokes ellipsometer (L116S300 type 7109-c375) equipped with a He-Ne laser (633 nm).

2.2 Analytical formula model

The formula proposed by Gallinet and Martin based on the ab initio theory [45] was employed to analyze the scattering properties of the proposed optical nanoantennas. The measurable scattering cross-sections can be expressed as the product of symmetric and asymmetric resonances which correspond to the bright and dark modes [46]:

2.3 Fabrication

The plasmonic nanostructures studied in this work were fabricated on ITO-coated glass using electron beam lithography (EBL). The resist (PMMA 950K) was spin-coated on the substrate and then exposed in the Raith Pioneer Two. In order to increase the adhesion of the gold (Au) film to the substrate, a titanium film with a thickness of 1 nm was evaporated onto the substrate before deposition of the Au film by using electron beam evaporation (Syskey tech ASB-EPI-C6). The morphologies of the fabricated nanostructures were characterized by using scanning electron microscopy (FEI Quanta 250 FEG) and atomic force microscopy (NT-MDT).

2.4 Optical characterization

The scattering spectra of the nanostructures were measured by using a dark-field microscope (Nikon) equipped with a spectrometer and a detector (ANDOR DU970P-BVF). The white light illumination with polarization along the y-axis was incident normally on the nanostructures and the scattered light was collected by using the objective (50×, NA=0.5) of the microscope. For the measurements of the Raman spectra, a He-Ne laser at 633 nm was used as the excitation light and the Raman signals were measured by using a commercial Raman spectrometer (Renishaw inVia). Rhodamine 6G (R6G) diluted with ethanol solution was used as the testing sample.

3. Results and discussion

3.1 Tetramer supporting Fano resonance

We first examined the optical properties of the tetramer in the proposed TR nanostructure, which is schematically shown in Fig. 1(a). The tetramer consists of four identical gold nano-cuboids with a vertical gap (G) and a horizontal distance (D) of 20 nm. The width (W), length (L) and thickness (T) of nano-cuboids are 75, 110, and 30 nm, respectively. The scattering spectrum of the tetramer obtained by FDTD simulation (red dot curve) is shown in Fig. 1(b), where one can see two peaks and one dip. In order to understand the physical origins of the scattering peaks and dip, we analyzed the scattering spectrum based on Eq. (1). As shown in Fig. 1(b), the analytical spectrum agrees well with the simulated one. If we decomposed the analytical spectrum according to the Eqs. (2) and (3), we observed an asymmetric Fano mode (green curve) and a symmetric bright mode (yellow curve). It indicates that the left peak and the middle dip in the scattering spectrum arise from the Fano resonance while the right peak originates from the bright mode.

 

Fig. 1. (a) Schematic of a tetramer consisting of four Au nano-cuboids. The geometrical parameters for the tetra are length L = 110 nm, width W = 75 nm, thickness T = 30 nm, horizontal distance D = 20 nm, vertical gap G = 20 nm. (b) Scattering spectrum of the tetramer obtained from numerical simulation (red dot curve) and analytical formula (blue curve), which can be decomposed into a bright mode (yellow curve) and a Fano resonance (green curve). (c) and (d) Surface charge distributions calculated at the dip (742 nm) and peak (836 nm) of the scattering spectrum. (e) and (f) Electric field distributions calculated at the dip (742 nm) and peak (836 nm) of the scattering spectrum.

Download Full Size | PDF

In order to gain a deep insight into the optical properties of the tetramer, we calculated the electric field and charge distributions at 724 and 836 nm, which correspond to the scattering dip and peak, as shown in Figs. 1(c) and 1(d). In Fig. 1(c), it is noticed that the electric dipoles excited in two nano-cuboids at the center are out-of-phase with those excited in the two nano-cuboids on both sides, implying that the electric field enhancement achieved at this wavelength originates from the reduction of radiation loss. Thus, this mode is attributed to a dark or subradiant mode [38,47] at which a maximum electric field enhancement of 29.7 is achieved, as shown in Fig. 1(e). In contrast, the charge distribution at 836 nm, as shown in Fig. 1(d), indicates that the in-phase oscillations of the electric dipoles excited in the four nano-cuboids of the tetramer lead to the formation of a larger electric dipole with enhanced radiation energy. Such a mode is considered as a bright or superradiant mode. Accordingly, the electric field is concentrated mainly in the gap region between two nano-cuboids at the center, giving rise to an enhancement factor as large as ∼32.6 (see Fig. 1(f)). Therefore, the two plasmon modes supported by the tetramer can be employed to realize the enhancements in the near-field intensity and far-field radiation, which are pretty important in many practical applications.

3.2 Further enhancing the near-field intensity of the tetramer

Based on the tetramer with Fano resonance, a gold ring enclosing the tetramer is added, forming the proposed optical nanoantenna of tetramer with ring (TR), as shown in Fig. 2(a). The scattering spectrum simulated for a TR nanoantenna (red dot curve) is shown in Fig. 2(b). It can be seen that the addition of the gold ring does not influence so much the scattering spectrum of the tetramer except the increased scattering intensity for the left peak. The Fano dip located at ∼744 nm remains nearly unchanged. The rise of the left peak is caused mainly by the optical antenna effect of the ring. It can be seen that the analytical spectrum (blue curve) agrees well with the simulated one. According to the analytical model, another bright mode (yellow curve) appears, which accounts for the rise of the left peak of the tetramer. We also calculated the scattering spectrum of the gold ring only (green curve), which further confirms that the second bright mode originates from the gold ring. We examined again the charge distributions in the tetramer, as shown in Fig. 2(c) and Fig. 2(d). It is noticed that the phase relationships between the electric dipoles excited in the tetramer at the two wavelengths (744 and 828 nm) remain unchanged, implying that the formation of Fano resonance is not affected by the introduction of the gold ring.

 

Fig. 2. (a) Schematic of a 30-nm-thick TR nanoantenna consists of a tetramer and a ring with an inner radius of R = 290 nm. The difference between the outer radius and the inner one is C = 100 nm. (b) Scattering spectra simulated (red dotted curve) and calculated (blue curve) for the TR nanoantenna. Also shown are the bright mode (yellow curve) induced by the ring and the bright mode (green curve) involved in the formation of the Fano resonance. (c) and (d) Surface charge distributions calculated at the dip (744 nm) and peak (828 nm) of the scattering spectrum. (e) and (f) Electric field distributions calculated at the dip (744 nm) and peak (828 nm) of the scattering spectrum. (g)-(i) Electric field (Ez) distributions in the ITO layer at the bright mode 2 (622 nm), 744 nm, and 828 nm, respectively. The black dot line represents the position of gold nanoring and the white dot line the tetramer position.

Download Full Size | PDF

It is worth noting that the effect of adding a ring is not only constructing a resonance at another wavelength. On the one hand, if we consider the ITO film layer with a refractive index of 1.63 under the gold nanoring, since the top and bottom of the ITO are surrounded by a medium with a refractive index smaller than that of ITO, it can be served as a slab waveguide to support optical waveguide modes. On the other hand, the gold nanoring set on the surface of ITO can be used as a waveguide coupling import and resonant cavity at the same time under the excitation of linear polarized light. So long as the thickness of the ITO layer and the geometric size of the gold nanoring are adjusted appropriately, Fabry-Perot-like resonance can be formed in the ITO layer. For this case, Fig. 2(g) shows the distribution of the main electric field component E_Z at 622nm (bright mode 2) in the ITO layer when there is only a gold nanoring on the ITO film. It can be clearly seen that a standing wave is formed inside the ITO layer where circled by the gold nanoring. Meanwhile, it is note that the bright mode induced by the gold nanoring is a broadband dipole resonance (Fig. 2(b), green curve), as shown in Fig. 2(h) and 2(i), there are still standing waves at 744nm and 828nm. Therefore, when the position and size of the tetramer are matched with the standing wave formed in the gold nanoring, the F-P-like resonance can be coupled with the Fano resonance in a broadband range, leading to the further enhancement of the local electric field in the tetramer. Correspondingly, the electric field enhancement factor (|E_max|/|E₀|) at the scattering dip (744 nm) is increased from 29.7 to 42.2 while that at the right scattering peak (828 nm) is increased from 32.6 to 42.5, as shown in Fig. 2(e) and Fig. 2(f).

3.3 Directional radiation

As mentioned above, the second function of the addition of the gold ring is to make it work together with the tetramer and manipulates the far-field radiation of the optical antenna. This feature leads to an enhanced collection efficiency of radiation signals, making it possible to use an objective with small numerical aperture and long working distance. In Fig. 3(a)–3(c), we compare the three-dimensional radiation patterns calculated at 700 nm for the tetramer, the ring, and the TR, respectively. It can be seen that the radiation energies of the tetramer (Fig. 3(a)) and ring (Fig. 3(b)) are mainly distributed in the forward direction (along the z-axis), scattering with large side lobes (x-z plane) when there is no interaction between them. Such an energy distribution is not good for the collection of the radiation signals, which generally detect the backward scattering intensity. In contrast, it is remarkable that the side lobes are almost eliminated due to the destructive interference between the radiations from the ring and the tetramer (Fig. 3(c)). In addition, the radiation energy is highly concentrated in the backward direction, leading to a greatly enhanced collection efficiency (CE). This unique property of the TR is more clearly reflected in the energy distributions in the x-y plane (along the + z-axis) calculated for the tetramer and the TR, as shown in Fig. 3(d) and Fig. 3(e). In order to evaluate the CE of the backward far-field radiation for the tetramer and TR at different wavelengths, we used the beam efficiency formula based on the far-field projection. In this case, the CE can be expressed as:

 

Fig. 3. (a)-(c) Far-field 3D radiation patterns of the tetramer, the ring and the TR at 700 nm, respectively. (d)-(e) Far-field backward scattering energy distribution calculated for the tetramer and TR at 700 nm. The white dotted circle indicates a radiation angle of 30° while the black one indicates a radiation angle of 48.6°. (f) Collection efficiencies in the backward direction calculated for the tetramer and TR at different wavelengths. (g) Collection efficiencies in the backward direction calculated for the tetramer and TR at different radiation angles (left panel). The dependence of the collection efficiency on the radiation angle in the range of 20-55° (shaded area) is shown in the right panel. The radiation angles of 30° and 48.6° correspond to the half-angle of objectives with NA=0.5 and NA=0.75, respectively.

Download Full Size | PDF

Here, P(θ,φ) denotes the radiation power along a certain direction, θ is the angle between the radiation direction and the z-axis, φ is the azimuth angle. In Fig. 3(f), we compare the CEs of the tetramer and TR when objectives with numerical apertures of 0.5 (solid curve) and 0.75 (dot curve) are used. It is found that the CE of the TR is higher than that of the tetramer over a wide wavelength range of 634-900 nm. In Fig. 3(g), we present the CEs of the tetramer and TR at 700 nm when objectives with different NA (radiation angle) are used. Similarly, it is observed that the TR possesses a higher CE in all cases. As shown in the inset of Fig. 3(f), the CEs of the tetramer and TR increase from 30.8% to 65.7% and from 51.1% to 79.4% when the NA of the objective is changed from 0.5 (with a radiation angle of 30°) to 0.75 (with a radiation angle of 48.6°), implying the TR is more efficient than the tetramer in the collection of radiation signals in a variety of conditions.

3.4 Independent and wideband tunning of the multi-wavelength

Although Fano resonance is multi-resonance in nature, and multi-wavelength tuning is also possible, the formation of Fano resonance requires that the electric fields of the two resonances are spacially and spectrally close with each other, which makes it challenging to achieve multi-wavelength tuning by changing the geometric dimensions. In particular, when one resonance is required to remain unchanged at the excitation wavelength, it is difficult to achieve wide-spectrum tuning for the other resonance. To reduce the limitation of Fano resonance on the wavelength tuning ability, the addition of gold nanoring played a key role for the third time. The Fano resonance peak was replaced by the dipole resonance of the gold nanoring (bright mode 2), and the independent broadband tuning of multiple resonances was realized. We first examined the bright mode induced by the gold ring. As shown in Fig. 4(a), the resonant wavelength of the bright mode is redshifted from 657 to 835 nm when the inner radius (R) of the ring is increased gradually from 280 to 340 nm. Interestingly, the dark mode remains nearly unchanged, implying that the two modes can be manipulated independently. Such characteristic is beneficial for practical applications using laser as the excitation, whose wavelength is fixed while the radiation signals are broadband or vary with samples, showing broad-spectrum detection ability and wide adaptability in various applications. Apart from the tunability of the bright mode, the dark mode can also be tunned by changing other geometric parameters. As shown in Fig. 4(b), the dark mode is blueshifted from 780 to 670 nm as the thickness T of the TR varies from 20 to 50 nm. Alternatively, it can also be tuned by varying the gap G between the middle two nano-cuboids (shown in Fig. 4(c)) or the length L of nano-cuboids (shown in Fig. 4(d)). A much wider tunning of the dark mode can be realized by simultaneously adjusting these geometric parameters. From Fig. 4(c) and 4(d), it is noticed that the second bright mode induced by the ring does not shift when the dark mode is moved, verifying the independent tunning ability of the two resonances again from another way.

 

Fig. 4. (a) Normalized scattering spectra calculated for TR nanoantennas with different inner radiuses (R) of the ring. (b) Scattering spectra calculated for TR nanoantennas with different thicknesses (T). (c) Scattering spectra calculated for TR nanoantennas with different gap widths (G) between the two middle nano-cuboids. (d) Scattering spectra calculated for TR nanoantennas with different lengths (L) of the nano-cuboids.

Download Full Size | PDF

3.5 SERS application by the TR optical nanoantenn

Based on the above-mentioned analysis, we verify the enhancement performance of the proposed TR antenna by carrying out the Raman scattering comparison experiments of TR, tetramer and gold film with the same conditions. We fabricated a tetramer and a TR with the geometric parameters described in Fig. 1(a) and Fig. 2(a). Their scanning electron microscopy (SEM) images are shown in Fig. 5(a) and 5(c), respectively. The morphologies of the tetramer and TR were also characterized by using atomic force microscopy (AFM). A size deviation smaller than 7.0 nm from the designed structures was confirmed by both the SEM and AFM observations. The backward scattering spectra of the two samples were measured by using dark-field microscopy, as shown in Fig. 5(b) and 5(d). The simulated scattering spectra are also provided for comparison. Very good agreement was observed between the measured (blue curve) and simulated (red curve) spectra, indicating that the two samples were successfully fabricated based on the designed parameters. We examined the enhancement factor for the Raman scattering signal by calculating |EF_ex|²*|EF_stokes|² integrated over the surface 1nm above the antenna, as shown in Fig. 5(e). Here, EF_ex = E_ex/E₀ and EF_stokes = E_stokes/E₀ represent the electric field enhancement factors for the excitation light and the Raman stokes light. The Raman scattering spectra measured on the gold film, the tetramer and the TR are shown in Fig. 5(f). Apparently, the Raman signals from the TR (yellow curve) are significantly enhanced as compared with those from the tetramer (blue curve) and the gold film (red curve). From Fig. 5(e), it is noticed that the electric field enhancement of the TR at 612 cm^-1 is ∼3.54 times larger than that of the tetramer. However, as shown in Fig. 5(f), the Raman signal from TR at 612 cm^-1 is 3.93 times higher than that of the tetramer. The larger enhancement observed in the experiments is attributed to the improved collection efficiency. Similar behaviors are found at 1362 and 1651 cm-1, verifying the double functionalities of the gold ring in both enhancing the near-field intensity and improving the far-field radiation. Limiting by the experimental conditions, the excitation light (633 nm) for the Raman scattering measurements was not chosen at the dark mode (∼744 nm), leading to relatively small enhancements in the Raman signals. A significant improvement is expected if the dark mode is designed at the excitation light (633 nm) by adjusting the geometric parameters of the TR.

 

Fig. 5. (a) SEM image of a fabricated tetramer with the geometrical parameters shown in Fig. 1(a). (b) Simulated (red curve) and measured (blue curve) backward scattering spectra of the tetramer. (c) SEM image of a fabricated TR with the geometrical parameters shown in Fig. 2(a). (d) Simulated (red curve) and measured (blue curve) backward scattering spectra of the TR. (e) Calculated enhancement factors achieved by the TR at Stokes Raman signals with different wavelengths by using a He-Ne laser (633 nm) as the excitation source. (f) Raman scattering spectra measured for Rhodamine 6G alcohol solution dripped on the Au film (red curve), the tetramer (blue curve), and the TR (yellow curve) under the same experimental conditions.

Download Full Size | PDF

4. Conclusion

In summary, we add a gold nanoring to the tetramer that supports Fano resonance to form a new nanoantenna. Numerical calculations and experimental studies show that the proposed nanoantenna simultaneously achieves the following three functional improvements compared with the traditional Fano resonant antenna. (1). By replacing the superradiation resonance of the Fano resonance with the dipole resonance induced by the gold nanoring, the multi-wavelength tunable performance of the Fano resonance antenna is significantly improved. (2). The local field of the Fano-type antenna is further enhanced by the F-P-like resonance formed by the combination of the gold nanoring and the substrate waveguide layer. (3). Directional radiation is realized by the collaboration of the gold nanoring and the Fano-type antenna, thus improving the collection efficiency of the far-field signal. The gold nanoring served as a carrier to improve the above three functions simultaneously and always plays the corresponding critical role when it needs to act independently (i.e., maintaining the original Fano resonance and multi-wavelength independently broadband tuning) and to interact with tetramer (i.e., near-field intensity enhancement and directional radiation). Therefore, the proposed nanoantenna provides a new idea for further improving the performance of many conventional Fano-type nanoantennas and a new strategy in designing photonic devices for strong light-matter interaction and efficient signal detection. This result implies potential applications in various fields such as surface-enhanced spectroscopy, photoluminescence, nonlinear nanomaterials/emitters and biomedicine sensing.