1. Introduction

Surface plasmon resonances, which are collective oscillations of free electrons in noble metals, have been widely exploited for various applications, such as photonic sensing and manipulation [1–3]. Plasmonic waves propagate at the metal−dielectric interface under specific light illumination conditions. Strong confinement effects of plasmonic energy to the nanometric dimension effectively generate high density electromagnetic fields on metallic nanostructures. Such fields have been demonstrated to be useful for highly sensitive detection of Raman scattering signals, refractive index (RI) difference, luminescence and fluorescence [4–11]. The optical properties of plasmonic devices are determined by their structural parameters, such as size, shape, and arrangement. As these structures are firmly fixed on substrates, it is not easy to tune their optical properties once fabricated. In fact, most previously reported plasmonic devices have been limited to passive functions. Although plasmonic properties can be altered by changing illumination conditions, there is a strong limitation on the geometric configuration of optical systems. One of the most important yet challenging requirements for further applications is to introduce a means to control the optical properties of plasmonic devices. Dynamic functionality offers opportunities for further integration of other optoelectronic components. Several tuning mechanisms have been proposed so far [12–16]. In microelectromechanical systems, the resonance wavelength is in the infrared region owing to technical difficulties related to the stable positioning of electrodes with nanometric accuracy [12]. While mechanical tuning methods have been demonstrated on elastic substrates, metallic structures were easily damaged [13, 14]. In the use of liquid crystals, metal surfaces must be covered, which is not suitable for sensor applications [15].

In this study, we present electrostatically tunable plasmonic devices on three-dimensional (3D) microsprings with a novel tuning mechanism. The plasmonic properties of the devices can be controlled by inclining metallic nanostructures using electrostatic forces. The plasmonic structures were fabricated by two-photon polymerization technique followed by metal film deposition. When femtosecond laser pulses are focused into photosensitive resin, nonlinear optical phenomena such as two-photon absorption occur only near the focal volume [17–21]. This technique enables one to obtain true 3D micro-/nanostructures by translation of the laser spot inside the resin. In addition, this technique offers high spatial resolution with subdiffraction-limit [17]. The plasmonic properties of these devices can be continuouslymodulated by applying voltage. The fabrication and optical evaluation of these electrostatically tunable plasmonic devices are described in this study.

2. Device structures

Figure 1 shows a schematic of the electrostatically tunable plasmonic devices. The devices comprise a nanostructured Au microplate, two 3D microsprings, and a support part. An electrostatic force is generated by applying DC voltage between the Au microplate and metallic substrate separated by a dielectric layer, resulting in the inclination of the microplate of the device. The micro-tower adjacent to the device is used for the electrical connection of the microplate. Gold gratings with one-dimensional grooves are on the microplate. Surface plasmons can be induced by coupling with incident light via the gratings, and these plasmons then propagate along the micropate. The phase matching condition between the incident light and surface plasmons is described in Eq. (1).

 

Fig. 1 Schematic images of (a) overall view and (b) cross-sectional view of the electrostatically tunable plasmonic devices.

Download Full Size | PDF

Figure 2 shows the calculated incident angle dependence of zero-th order reflectivity when light of 632.8 nm wavelength is incident to a Au grating with a different period. The RI of Au was assumed to be 0.312 + 3.14i. A clear dip can be seen for p-polarized light in the reflection spectra of each grating, whereas no dip is observed for s-polarized light, which is parallel to the grating vector. From Eq. (1), the dips originate from optical losses due to plasmon excitation using first order diffracted light, indicating that plasmonic waves can propagate on the microplates. Therefore, the dip depths can be regarded as the plasmon excitation efficiency. The plasmonic modes are induced using optical diffraction on the gratings, which results in different dip depth and phase matching angle for the grating with each period. Note that the full-width at half-maximum (FWHM) of each dip is small less than 4°, which means that a slight angular change of the incident light can effectively modulate plasmon excitationefficiencies. For example, the efficiency changes by as much as 75% after an incident angle change of 3.0° in the case of a 600 nm period. The inclination of the micro plates is equivalent to the incident angle changes of light. Therefore, from the strong angular dependence, even a slight inclination leads to the effective modulation of the excitation efficiency. This tuning mechanism has the advantage of not requiring metal deformation or electrode nano-positioning. In addition, these devices can easily respond to various wavelength regions by changing the parameters such as period or arrangement of nanostructures on the microplate. Without microsprings, bending moment at free end of the microplate causes curvature of the plates during applying voltage. According to Eq. (1), such deformation should be avoided for homogeneous plasmon excitation on the microplate. For example, when curvature radius of the plate is 4.5 mm, which is similar to AFM cantilevers, maximum difference Δθ ( = microplate length/curvature radius) of effective incident angle is estimated to be 1.1° in the microplates. From calculation in Fig. 2, such angle variation causes difference of plasmon excitation efficiency more than 10%. The microsprings support the free end of the plate, which is effective for suppression of the curvature.

 

Fig. 2 Calculated zeroth order reflectivity of Au gratings with various periods when a He–Ne laser of 632.8 nm wavelength was incident to the gratings. Clear dips can be seen for p-polarized light, whereas no dips appeared for s-polarized light.

Download Full Size | PDF

3. Fabrication and characterization

3.1 Fabrication

Figure 3 shows fabrication processes of the electrostatically tunable plasmonic devices. First, SiO₂ thin films of 5 μm thickness were deposited on polished copper substrates as dielectric layers. Chemically amplified negative-tone photoresist SU-8 was then spin-coated. The device structures and adjacent towers were directly written by three-dimensional translation of the femtosecond laser-focus inside SU-8. The grating period on the microplate was determined to be 800 nm, and the microplate size was 90 μm × 70 μm. The diameters of the springs and wires were, respectively, 25 μm and 3−4 μm. Finally, Au thin films of 200 nm thickness were deposited by the RF sputtering method. The central wavelength, pulse duration and repetition rate of femtosecond laser were, respectively, 780 nm, 127 fs and 100 MHz. Laser pulses, which were focused by an objective lens with N.A. of 0.95, were irradiated to the samples ona computer-controlled stage with nanometer accuracy. The laser writing system allows us to obtain the devices with reproducibility. The laser writing time was approximately 40 min for each device. To obtain electrostatic forces, the Au deposition on the region just below the microplate should be suppressed during the sputtering process. Sidewalls were, therefore, formed around as shown in Fig. 3(e). Figure 4 shows scanning electron microscopy images of (a) the device, (b) the 3D microspring, (c) the nanostructured microplate and (d) 5 × 5 array, respectively. Figure 4(b) shows a 3D microspring under the microplate without sidewalls for easy observation. The spiral structures with smooth surfaces can be seen below the microplate. The groove array of 800 nm period was also confirmed in Fig. 4(c).

 

Fig. 3 The fabrication process of electrostatically tunable plasmonic devices. (a) Spin coating of SU-8 onto SiO₂-coated Cu substrates. (b) Direct writing of three-dimensional device structures inside SU-8 by femtosecond laser. (c) Baking and development treatment. (d) Au film deposition. (e) The devices with sidewalls around the microplate to suppress Au film deposition just below the region of the microplate.

Download Full Size | PDF

 

Fig. 4 Scanning electron microscope images of (a) an overview of the device and (b) three-dimensional spring below the microplate, (c) the nanostructured microplate and (d) 5 × 5 array.

Download Full Size | PDF

3.2 Mechanical property

The voltage dependence of inclination angles of the microplate is shown in Fig. 5. The inclination angles were estimated from the direction of the zero-th order reflected light when He-Ne laser light of 632.8 nm wavelength was incident to the devices. The inclination angle increased for voltage up to 60 V, and then gradually became saturated at higher voltages. The maximum inclination was measured to be 5.5 ± 0.5° at 80 V. The top views of the device before and after applying voltage are shown in inset of Fig. 5. We confirmed that the micro-plate could be inclined using an electrostatic force. The electrostatic force F between parallel plate electrodes is described by the following Eq.

 

Fig. 5 Voltage dependence of inclination angles of the microplate. The top views of the devices before and after applied voltage are shown in inset.

Download Full Size | PDF

 

Fig. 6 Calculated dependence of the electrostatic force on the displacement of the microplate at different voltages. The initial gap was assumed to be 50 μm. The measured displacement of the microplate at each voltage level was plotted (filled red circles).

Download Full Size | PDF

3.3 Optical property

Figure 7 shows changes in the normalized reflectivity as a function of applied voltage when He-Ne laser light of 2 mm diameter was incident to 5 × 5 array devices. The initial incidentangle was set to be 13°. In the case of p-polarized light, the reflectivity decreased with an applied voltage up to 50 V, followed by a sudden increase. The reduction in reflectivity was as large as approximately 35% at 50 V. In contrast, the spectrum for s-polarized light was almost flat. The inclination angle of the microplate was 4.0° at 50 V, which corresponds to effective incident angle of 17° to the microplate. From Eq. (1), the phase matching angle was estimated to be 15.2° for p-polarized light. Therefore, the dip in the reflectivity can likely be attributed to optical loss due to the coupling of incident light with plasmonic modes on the microplates. This result means that the amount of reflectivity reduction can be regarded as the plasmonon excitation efficiency in Fig. 7. Almost no change of reflectivity for s-polarized light is also consistent with this prediction because surface plasmons can be induced by p-polarized light through the gratings. The maximum plasmon excitation efficiency of approximately 35% was comparable to calculated value of 44% in Fig. 2. Note that the plasmon excitation efficiencies could be both markedly and continuously tuned from almost zero to maximum through the application of voltage alone. The difference in voltage for tuning from minimum to maximum was 50 V, which corresponds to an incident angle change of 4.0°. As shown in Fig. 2, the FWHM of the dip for the microplate of 800 nm period was calculated to be 2.9°, which is close to experimental value. To reduce the voltage required for modulation, device with a smaller FWHM is desirable. From Fig. 2, a shortening of the grating period appears to be one of the effective means of doing this. This shortening is also useful for highly efficient devices. For example, the plasmon excitation efficiency was calculated to be more than 80% for Au gratings of 400 nm period. The spatial resolution of multi-photon polymerization technique was reported to be as high as 20 nm for SU-8 [18]. Therefore, optimization of the device configuration, including the nanostructured microplates and 3D microsprings, will improve the optical properties of these devices.

 

Fig. 7 Changes in the normalized reflectivity of an arrayed plasmonic device with applied voltage when He-Ne laser light of 632.8 nm wavelength was incident to 5 × 5 array devices. The initial incident angle was set to be 13°.

Download Full Size | PDF

4. Conclusions

Electrostatically tunable plasmonic devices were fabricated using multi-photon polymerized 3D microsprings. The device comprised a nanostructured Au microplate and two 3D microsprings. The microplates were inclined by applying DC voltage between the microplates and copper substrates, leading to the modulation of plasmon excitation. The efficiencies of plasmon excitation were continuously changed from almost zero to maximum by a voltage change of 50 V. Such dynamic functionality can contribute to the realization of advanced optoelectronic devices and tunable metamaterials.