1. Introduction

Polarization multiplexed optical transmission techniques for next-generation optical communications have been demonstrated, and they have been shown to have a higher spectral efficiency than conventional techniques [1–4]. In a polarization multiplexing optical circuit, TE mode signals are transmitted independently of TM mode signals [5]. Silicon-photonics-based monolithically integrated transmitters and receivers that are lower cost, more compact, and more energy efficient are required as key components in coherent optical communications systems [6,7]. As such, miniaturized nano-scale optical dielectric waveguides have been implemented to realize such compact systems [8–12]. Gap plasmons, i.e., electromagnetic waves coupled to collective oscillations of free electrons at metal-insulator-metal interfaces, enable an optical waveguide to be miniaturized beyond the diffraction limit of propagating light [12–18]. However, in a metallic-gap waveguide, the excitation condition of the propagating mode orthogonal to the gap plasmons is limited by the width of the cut-off gap [15]. Gap plasmonic waveguides allow only a single polarization state and are unsuitable for polarization multiplexed transmissions. The purpose of this study was to develop a novel gap-plasmon excitation structure for dense polarization multiplexing planar lightwave circuits (PLCs) and photonic integrated circuits (PICs) that do not display any inconsistencies.

Gap plasmonic mode excitation or conversion structures interfaced with PLCs and PICs are of interest because of their potential to decrease the size of optical wires. Ono et al. demonstrated highly efficient three-dimensional mode conversion between deep-subwavelength gap plasmonic waveguides and Si stripe waveguides [13], and Choo et al. reported highly efficient nanofocusing in a Au–SiO₂–Au gap-plasmon waveguide using a carefully engineered three-dimensional taper [16]. These reports both present highly efficient excitation and conversion structures with single mode polarization. Wei et al. examined the coupling of cross-polarized far-field light to highly confined plasmonic gap modes via connected nanoantennae [19].

In this paper, we propose a highly efficient on-chip gap-plasmon excitation structure connected to dense polarization multiplexing PLCs. Here, the excited plasmons are orthogonally polarized with respect to the input lightwave (or surface plasmons). Figure 1 shows a schematic diagram of the proposed structure. The structure consists of a Au nanostripe and tapered gap for refractive index matching to a gap plasmonic waveguide; the structure was fabricated on the top surface of a dielectric-stripe-type plasmonic waveguide deposited on a Au film. Propagating surface mode plasmons that are confined into the 600-nm-wide and 500-nm-thick dielectric-stripe waveguide are localized at the corner of the Au nanostripe. Then, the localized lateral plasmons are converted to the orthogonal-polarized plasmonic gap mode by increasing the effective refractive index of the gap waveguide using the tapered gap. The effective refractive index and the mode field area of the localized lateral plasmonic mode are closer to those of the gap plasmonic mode than the dielectric stripe mode. Therefore, metallic-nanostripe-based refractive index matching was achieved using the simple and compact structure. The proposed novel excitation structure will pave the way for next-generation optical communications using higher density PLCs and PICs.

 

Fig. 1 Schematic illustration of proposed excitation structure. The width and height of the SiO_x stripe waveguide are 600 and 500 nm, respectively.

Download Full Size | PDF

2. Principle and design

We numerically designed the structure shown in Figs. 2(a) and 2(b) using the finite-difference time-domain (FDTD) method (Poynting for Optics, Fujitsu). Here, w is the Au nanostripe's width, L is the Au nanostripe's length, g is the Au-air-Au gap's width, t is the Au layer's thickness, and θ is the angle of the Au taper. The refractive indices of SiO_x and Au were set at 1.447 and 0.41 + 8.37i, respectively [20,21]. The structure was demonstrated using light with a wavelength of 1550 nm in free space. Figures 2(c)–2(e) show the cross-sectional view of the plasmonic intensity distributions corresponding to the dashed lines in Fig. 2(b). In the FDTD simulations, surface plasmons were utilized as the input wave because single mode polarization perpendicular to the substrate surface is necessary to evaluate the excitation of the orthogonal-polarized gap plasmons. Propagating surface mode plasmons, which are confined into the 600-nm-wide and 500-nm-thick dielectric-stripe waveguide, are localized at the corner of the Au nanostripe. Then, the localized lateral plasmons are converted to the orthogonal-polarized plasmonic gap mode by increasing the effective refractive index of the gap waveguide using the tapered gap.The structural parameters w and L are the critical design parameters for the excitation efficiency of the structure because the gap plasmons are confined into the waveguide (in addition to the focusing via the Au taper) via a directional-coupling-like phenomenon between the gap mode and the localized mode along the Au nanostripe. Here, the Au nanostripe length was determined to be 1500 nm by considering coupling length between the SiO_x stripe and the Au nanostripe. Figure 3(a) shows a numerical plasmonic intensity distribution 25 nm above the surface of the SiO_x layer. This efficient coupling into the plasmonic gap mode occurred periodically along the propagation direction. We investigated the dependence of the nanostripe's width on the coupling length for each thickness and gap width, as shown in Fig. 3(b). In the case of the 50-nm Au nanostripe width (w = 50 nm) and the 25-nm Au layer thickness and gap width (t = g = 25 nm), the effective refractive index of the gap plasmon mode (n_eff = 1.81) is higher than the localized mode at the nanostripe (n_eff = 1.59). Therefore, the localized lateral plasmons are converted to the orthogonal-polarized plasmonic gap mode within the Au taper end. Taper angles θ of 10, 15, 20, and 25 degrees were set by searching for the peak transmittance for each Au layer thickness and for gap widths of 25, 50, 75, and 100 nm [Fig. 3(c)]. Here, the taper length was set to 500 nm by considering fabrication accuracy. The numerical results of the input wave and the output orthogonally polarized plasmonic intensity were obtained at the distance of 100 nm from the Au nanostripe and the Au taper, respectively. In Fig. 3(c), the transmittance was calculated by comparing area-averaged optical intensity data at the dielectric waveguide [Fig. 2(c)] and the gap waveguide [Fig. 2(e)]. The optical intensity data were time-averaged over a period of the input wave. We numerically confirmed an approximate excitation ratio of 0.79 from the dielectric-stripe- surface mode to the metallic-gap mode for an Au film thickness of 100 nm and a gap width of 100 nm.

 

Fig. 2 (a) Schematic view of the simulated model. (b) Enlarged view of the Au taper region. Here, w is the Au nanostripe's width, L is the Au nanostripe's length, g is the Au-air-Au gap's width, t is the Au layer's thickness, and θ is the angle of the Au taper. (c–e) Numerical simulation results of the cross-sectional plasmonic intensity distributions corresponding to the dashed lines (c)–(e) in Fig. 2(b). The color bar is linear in intensity. The maximum and minimum values are normalized with respect to the peak and zero plasmonic intensity, respectively.

Download Full Size | PDF

 

Fig. 3 Calculated results of (a) the plasmonic intensity distribution 25 nm above the surface of the SiO_x layer (top view), (b) coupling length vs. Au stripe width, and (c) transmittance vs. tapered angle.

Download Full Size | PDF

3. Results

Based on the FDTD simulation's results for the design, an 80-nm-thick Au pattern was fabricated using focused ion beam milling on the top surface of a dielectric-stripe-type plasmonic waveguide formed on a Au film. Figure 4(a) shows a scanning electron micrograph of the fabricated structure. Here, the width and length of the Au nanostripe were 150 and 1900 nm, respectively, and the width of the plasmonic gap mode waveguide was approximately 80 nm. The coupling length was approximately 2800 nm.

 

Fig. 4 Experimental results of the fabricated devices. (a) Scanning electron micrograph of the proposed device, which was fabricated using focused ion beam milling. (b) Schematic illustration of the experimental setup. (c) Polarization angle dependence of the propagating plasmonic intensity measured from the evanescent waves at the SiO_x stripe's surface. (d) Scanning electron micrograph, (e) experimental result using the SNOM, and (f) calculated result of the near-field optical intensity distribution obtained 100 nm above the studied area. The color and gray scale bars are linear in intensity. The maximum and minimum values are normalized with respect to the peak and zero plasmonic intensity, respectively.

Download Full Size | PDF

Experiments were performed using a scanning near-field optical microscope (SNOM, Japan Spectroscopic Company, NFX-520). A schematic illustration of the experimental setup is shown in Fig. 4(b). The back side of a quartz substrate was irradiated with collimated incident light, which was chopped at a frequency of 300 Hz. The plasmonic surface mode signals were generated via the slit, and the incident light from a tunable laser output (Santec, MLS-2100) was spotted using a collimated fiber (Go Foton, C-OPCL-SMF-103/FC). The incident light was set at a wavelength of 1550 nm with a power of 1 mW. Figure 4(c) shows the polarization angle dependence of the propagating plasmonic intensity measured from the evanescent waves at the SiO_x stripe's surface. The measured intensity was normalized with respect to the maximum value. The peak intensity is shown for TM polarized incident light (0 degree), while the minimum intensity is shown for the TE polarization (90 degree). For the SNOM, the plasmonic intensity distributions were detected with a photo-multiplier tube (Hamamatsu Photonics, H10330B-75) using near-field optical probe scanning at the surface of the device. The probe was controlled by an XYZ piezoelectric stage (Nano Control, A118-01) and a controller (Nano Control, NCM7302C). The photo-multiplier output was connected to a digital lock-in amplifier (NF, LI5640), the output from which was monitored by the SNOM system and used to analyze the chopped plasmonic intensity distributions.

The plasmonic gap mode excitation was confirmed by evaluating the near-field plasmonic intensity distribution 100 nm above the surface of the fabricated structure. In Figs. 4(d)–4(f), we show the scanning electron micrograph as well as the experimental and calculated plasmonic intensity distributions obtained for the measured area. The numerical distribution was calculated in the steady state; therefore, interference fringes were observed in both the experimental and numerical results. The near-field plasmonic patterns corresponding to the gap mode excitation were observed and agreed well with the calculated results. The coupling ratio from the Au stripe to the gap waveguide was estimated to be approximately 0.51 by comparing area-averaged near-field intensity distributions at the Au nanostripe and the gap waveguide. The experimental result had a lower efficiency than expected from the simulation results as the fabrication accuracy was insufficient. In the SNOM result, the transverse shift of the plasmonic distribution occurred because the gap plasmons were confined into the waveguide via a directional-coupling-like phenomenon between the gap mode and the localized mode along the Au nanostripe. The localized mode was reflected at the Au nanostripe end because of the incomplete coupling length of the fabricated nanostripe. The results confirm the feasibility of highly efficient excitation of gap plasmons in dense polarization multiplexing circuits.

4. Conclusion

We have demonstrated an on-chip gap-plasmon excitation structure connected to a dense polarization multiplexing PLC. The excited plasmons were orthogonally polarized with respect to the single mode, perpendicularly polarized input surface plasmons. The proposed structure, which was composed of a Au nanostripe and tapered gap for refractive index matching to the gap plasmonic waveguide, could be utilized to create more compact optical communication systems with a high spectral efficiency for polarization multiplexing circuits. The gap plasmons were excited by light with a wavelength of 1550 nm in free space. Through a numerical design based on simulation results using the finite-difference time-domain method, the structural parameters were determined at each gap width and for each Au film thickness. In the Au nanostripe we controlled the coupling length by varying the Au nanostripe width and the Au film thickness. The angle of the Au taper was determined by calculating the excitation efficiency of the gap plasmons. We have numerically confirmed an excitation ratio of approximately 0.79 from the dielectric-stripe-surface mode to the metallic-gap mode for a Au film thickness of 100 nm and gap width of 100 nm. The deposited 80-nm-thick Au pattern was fabricated using focused ion beam milling on the top surface of a dielectric-stripe-type plasmonic waveguide formed on a Au film. The experimental coupling ratio was estimated to be approximately 0.51 using scanning near-field optical microscopy. We believe that this work will contribute to the development of next-generation optical communications systems with a higher spectral efficiency.