Plasmon-Induced Transparency in Asymmetric T-Shape Single Slit

Apr 3, 2012 - School of Science, Beijing University of Posts and Telecommunications, Beijing 100876, China. §. State Key Laboratory for Mesoscopic ...
0 downloads 0 Views 2MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Plasmon-Induced Transparency in Asymmetric T-Shape Single Slit Jianjun Chen,*,†,‡,§ Zhi Li,§ Song Yue,§ Jinghua Xiao,†,‡ and Qihuang Gong*,§ †

State Key Laboratory of Information Photonics and Optical Communications, Beijing University of Posts and Telecommunications, Beijing 100876, China ‡ School of Science, Beijing University of Posts and Telecommunications, Beijing 100876, China § State Key Laboratory for Mesoscopic Physics and Department of Physics, Peking University, Beijing 100871, China ABSTRACT: By utilizing a dielectric-film-coated asymmetric T-shape single slit, comprising two grooves of slightly detuned widths immediately contacting with a single nanoslit, the plasmon-induced transparency was experimentally demonstrated. Because of the symmetry breaking in the unit-cell structure, the scattered lights from the two grooves with slightly detuned widths interfere destructively, leading to the plasmon-induced transparency. As a result, a response spectrum with nearly the same interference contrast but a much narrower bandwidth emerges in the unit-cell structure with the footprint of only about 0.9 μm2, compared with that in the symmetric T-shape single slit. These pronounced features in the structure, such as the increased quality factor, ultracompact size, easy fabrication, and experimental observation, have significant applications in ultracompact plasmonic devices. KEYWORDS: Surface plasmon polaritons, electromagnetically induced transparency, symmetry breaking, destructive interference

E

experimentally in ultrasmall unit-cell structures with easy fabrication is very essential for highly integrated optics. In this paper, we experimentally demonstrate an EIT-like optical response in a dielectric-film-coated asymmetric T-shape single slit, which is an ultrasmall unit-cell structure and comprises two grooves of slightly detuned widths immediately contacting with a single nanoslit. Because of the deep groove in the asymmetric T-shape single slit and the tight field confinements by the dielectric film, the generated SPPs propagating along the groove bottom can be highly reflected by the metal step of the groove18 and then be scattered by the nanoslit and interfere with the directly transmitted light from the nanoslit. At destructive interference, the transmittance of the T-shape single slit is suppressed. In the asymmetric T-shape single slit, the scattered light from the two grooves are in antiphase with each other and thus interference destructively because of the symmetric breaking.9,17,19 This results in an enhanced transmission, revealing an EIT-like optical response. As a result, a response spectrum with nearly the same interference contrast but a much narrower bandwidth emerges, compared with that in the symmetric T-shape single slit. The proposed structure is quite simple and uses a bulk incident wave, which allows easy fabrication and convenient experimental observation of the EIT-like optical response.

lectromagnetically induced transparency (EIT) is a quantum interference effect with a spectrally narrow optical enhanced transmission, which results from a coherent interaction between the atomic levels and the applied optical fields.1 The enhanced transmission can lead to dramatically slowed down photons and has potential applications in enhanced nonlinearities, modulations, and sensors.2 Recently, tremendous attention has been attracted to the studies that EIT-like optical responses can be obtained in classical resonator systems,3 which are easily integrated into the chips. As we know, surface plasmon polaritons (SPPs) can be well confined by ultrasmall metal structures and break the diffractive limit.4,5 Thus, combining the EIT effect with nanoplasmonic structures would open the possibility to achieve ultrasmall sensors and modulators with high sensitivities to variations of the surroundings. Most studies on this topic mainly focused on the array metallic structures,6−13 where different plasmonic resonators coupled each other in each cell to realize the EITlike optical responses. However, these array structures are a little bulk and complicated, which is not preferred by highly integrated optics. Recently, using a unit-cell structure, different plasmonic resonators side-coupled with a waveguide, the EITlike transmission spectra were theoretically predicted.14−17 Compared with the array structures, the unit-cell structures are much compact and easy to be integrated in the optical circuits. However, these unit-cell structures utilized strongly localized guided SPPs as incident waves,14−17 which makes the experimental observation of EIT-like optical responses being a challenge. Therefore, realizing the EIT-like optical responses © 2012 American Chemical Society

Received: February 16, 2012 Revised: March 27, 2012 Published: April 3, 2012 2494

dx.doi.org/10.1021/nl300659v | Nano Lett. 2012, 12, 2494−2498

Nano Letters

Letter

160 nm, and d = 150 nm, respectively, and the thickness and the refractive index of the dielectric film were t = 160 nm and nd = 1.50, respectively. This is a two-dimensional simulation model, and these parameters of the structure are fixed throughout this paper. Above (several micrometers) the exit of the T-shape slit in the air, a detection port was placed to collect the transmitted light. The transmittance is defined as the quotient between the power flows (obtained by integrating the Poynting vector over the detection port) in the structure with and without the grooves on each side of the slit. The calculated results of the transmittance in the symmetric T-shape slit changing with the groove width (wG = wG1= wG2) at λ = 850 nm (the corresponding permittivity of the gold being εAu = −25.4 + 1.75i20) are displayed by the black line in Figure 1b. It is observed that the transmission curve exhibits oscillation behaviors with a period of 306 nm, which is exactly equal to the period of p = λspp/2 inferred from eq 1, where λspp is the SPP wavelength in the grooves. At constructive interference, the far field in the air (above the metal surface) is strong, corresponding to a transmission maximum, as shown in Figure 1c, whereas the far field in the air is weak, corresponding to a transmission minimum, as shown in Figure 1d. To show the influences of the dielectric film on the transmittance of the T-shape slit, the above calculation were carried out in the same case without the dielectric film, and the results are displayed by the red dashed line in Figure 1b. It is observed that the quality factors and the interference contrast of the dielectric-film-coated structure considerably increase due to the tight SPP confinement by the dielectric film. This greatly benefits the application of plasmonic sensors and modulators. For the asymmetric T-shape single slit, the groove widths on opposite sides of the nanoslit are slightly detuned, so the scattered lights from these two grooves have different phases for the same incident wavelength. By carefully adjusting the detuned widths in the asymmetric T-shape single slit, the scattered lights from these two grooves can be in antiphase with each other and thus interfere destructively. This could result in an enhanced transmittance, revealing an EIT-like optical response.15−17 To demonstrate this effect, the transmission spectra of the proposed T-shape single slit were calculated. Herein, the transmission spectra of the T-shape slit were obtained by changing the input wavelength. The permittivity of the gold was calculated as a function of the wavelength using interpolation and was taken from the literature.20 The transmission spectra of the T-shape single slit in dependence on the detuned groove widths are displayed in Figure 2. Without detuning the widths of these two grooves in the Tshape single slit (δ = wG1 − wG2 = 0 nm: wG1 = wG2 = 1100 nm), there exists a broadband transmission dip (at about λ = 850 nm and bandwidth of about Δλfwhm≈70 nm), as shown in Figure 2a. The transmission dip is due to the destructive interference between the scattered SPPs by the nanoslit and the directly transmitted light from the nanoslit, and the broadband spectrum is mainly owing to the large radiative loss. The corresponding field distribution of the transmitting light at the transmission dip position is shown in Figure 1d. By slightly detuning the widths of these two grooves in the T-shape single slit, a transmission peak obviously emerges in the broadband transmission spectrum due to the symmetry breaking, as show in Figure 2b−e. In this case, the destructive interference of the scattered lights from these two grooves with slightly detuned widths results in the EIT-like optical responses.15−17 Moreover, it is observed that the emerged

The proposed asymmetric T-shape single slit is schematically shown in Figure 1a, comprising a conventional nanoslit (width

Figure 1. (a) Schematic and geometric parameters of the asymmetric T-shape single slit covered with a dielectric film. (b) Dependence of the transmittances in the symmetric T-shape single slit covered with and without a dielectric film. Field distributions (Hz) of the dielectricfilm-coated symmetric T-shape single slit for (c) wG1 = wG2 = 900 nm and (d) wG1 = wG2 = 1100 nm.

of w) in immediate contacting with two grooves of slightly detuned widths (wG1 and wG2) in a gold film. The whole structure is covered by a dielectric film with a refractive index of nd. The tight field confinement by the dielectric film and the deep groove depth can ensure high reflectivity off the metal step of the grooves.18 When p-polarized light (magnetic vector parallel to the slit) illuminates the structure from the back side, a part of light can pass through the nanoslit directly, and the other can generate SPPs propagating along the bottom of the grooves. The generated SPPs can be highly reflected by the two metal steps of the grooves in the T-shape single slit18 and then be scattered by the nanoslit and interfere with the directly transmitted light from the nanoslit, as shown in Figure 1a. This interference can greatly affects the transmittance of the nanoslit, and it is easy to obtain that the phase difference between the scattered SPPs and the directly transmitted light is determined by Φ = 2k SPPwGm + φ

(m = 1, 2)

(1)

where wGm is the groove width, kspp is the wave vector of the SPPs in groove, and φ denotes the sum of the phase shifts brought by the SPP generation at the nanoslit, the reflections off the groove step, and the SPP scattering from the nanoslit. According to eq 1, constructive (or destructive) interference of the two interfering parts for the transmitted light should occur when Φ is equal to even (or odd) multiples of π. At constructive interference, the transmittance of the T-shape single slit is enhanced. On the contrary, the transmittance of the T-shape single slit is reduced at the destructive interference. To verify the above analysis, the transmittance of the symmetric T-shape slit was first calculated by using the commercial finite element (FEM) solver of COMSOL Multiphysics. In the simulations, the gold film thickness, slit width, and the groove depth were set to be tm = 250 nm, w = 2495

dx.doi.org/10.1021/nl300659v | Nano Lett. 2012, 12, 2494−2498

Nano Letters

Letter

To further test our proposal experimentally, both the symmetric and asymmetric T-shape single slits were fabricated using focused ion beams in a 250-nm-thick gold film, which was evaporated on a glass substrate with a 30-nm-thick titanium adhesion layer. Here, the symmetric T-shape single slit was used as an on-chip reference. Figure 3a shows the scanning

Figure 3. (a) SEM image of the asymmetric T-shape single slit. (b) Schematic of the operation process in the asymmetric T-shape single slit coated with the PVA film.

Figure 2. Simulation results of the transmission spectra in the dielectric-film-coated T-shape single slit with slightly detuned groove widths of (a) δ = 0 nm: wG1 = wG2 = 1100 nm, (b) δ = 50 nm: wG1 = 1100 nm and wG2 = 1050 nm, (c) δ = 100 nm: wG1 = 1100 nm and wG2 = 1000 nm, (d) δ = 150 nm: wG1 = 1100 nm and wG2 = 950 nm, and (e) δ = 200 nm: wG1 = 1100 nm and wG2 = 900 nm. (f−i) Corresponding field distributions (Hz) of the T-shape single slit at the red-arrow positions for different detuned groove widths of (b−e).

electron microscopy (SEM) image of the asymmetric T-shape single slit. In the middle of the image is the single nanoslit with the length of about 10 μm. Both sides of the nanoslit are flanked by a shallow groove with slightly detuned widths to form the asymmetric T-shape single slit. The measured geometrical parameters of the asymmetric T-shape single slit are as follows: the slit width is about w = 160 nm, the groove depth is about d = 150 nm, and the widths of the grooves are about wG1 = 1120 nm and wG2 = 970 nm. At last, the whole structure was spin-coated with a PVA (poly(vinyl alcohol), refractive index of nPVA = 1.5) film of about 160 nm thickness. This can ensure high reflectivity of SPPs off the metal step of the grooves.18 The previous experiments showed that the PVA film could settle into the groove and kept its thickness nearly unchanged, leaving an indentation on the top surface of the sample,18,21 as shown in Figure 1. In the experiment, we first measured the transmittance of the symmetric T-shape single slit using a spectrograph (Andor). The whole structure was normally illuminated from the back side using a white point-light source, as shown in Figure 3b. Here, the point-light source was first polarized to be a ppolarized light and then focused on the sample with a radius of about 100 μm, which could ensure nearly uniform incident intensities over the T-shape single slit. Noted that the incident light is a bulk wave, which is quite different from the strongly localized guided waves utilized by the previous numerical structures.14−17 This makes the experimental realization of EITlike optical response rather easier than that in the previous numerical structures.14−17 Moreover, the utilization of bulk incident waves may expand the physics of the EIT-like plasmonic resonances. The transmitted light was collected by a long working distance objective (Olympus 10×) and then coupled to a fiber, which connected the spectrograph. The measured transmission spectrum of the symmetric T-shape single slit is displayed in Figure 4a. It is observed that the transmittance reaches the minimum at about λ = 860 nm, and the bandwidth of the spectrum is about Δλfwhm ≈ 70 nm. Next, we measured the transmission spectrum of the asymmetric Tshape single slit (Figure 3a), and the results are shown in Figure 4b. It is noted that the EIT-like optical response occurs in the transmission spectrum of the asymmetric T-shape single slit. A transmission peak obviously emerges in the broadband

transmission peak is blue-shifted with decreasing the right groove width. So this EIT-like optical resonance can be carefully modified by varying the structural asymmetry. Particularly, at proper detuned groove widths of δ = 150 nm (wG1 = 1100 nm and wG2 = 950 nm), the EIT-like transmission spectrum exhibits nearly the same interference contrast (about 0.6 for δ = 150 nm and about 0.7 for δ = 0 nm) but a much narrow bandwidth (Δλ′fwhm ≈ 30 nm for δ = 150 nm and Δλfwhm ≈ 70 nm for δ = 0 nm), compared with that in the corresponding symmetric T-shape single slit, as shown in Figure 2a,d. These pronounced features are greatly favorable for the compact plasmonic devices such as sensors, modulators, and switches in highly integrated optics. Herein, the interference contrast is defined as (Tmax − Tmin)/(Tmax + Tmin), where Tmax and Tmin denote the maximum and minimum of the transmittance in the T-shape single slit, respectively. The corresponding field distributions (Hz) of the transmitted light at the transmission peaks (red-arrow positions) for different detuned groove widths are shown in Figure 2f−i. Because of the broadband transmission dip, the scattered light from the left groove is always nearly in antiphase with the directly transmitted light from the nanoslit. It is obviously noted that the SPPs (evanescent fields) in the two grooves appear in antiphase with each other, and the SPPs in the right groove are much stronger than that in the left groove, as shown in Figure 2f−i. Thus, a small part of the scattered light from the right groove can completely cancel that from the left groove, and the remaining part from the right groove interferes with the directly transmitted light constructively, resulting in the enhanced transmission, as shown by the far field distribution in Figure 2f−i. Therefore, the EIT-like transmission spectra occur in the broadband transmission spectrum. Moreover, the SPPs in the right groove at the EIT-like resonances are much stronger than that in the symmetric T-shape single slit, as shown in Figures 1c,d and 2f−i. This is also of importance for the areas of enhanced optical nonlinearity, sensors, and modulators. 2496

dx.doi.org/10.1021/nl300659v | Nano Lett. 2012, 12, 2494−2498

Nano Letters

Letter

significantly increase the modulation depths and lower the power threshold. In the sensing applications, the EIT-like transmission spectra in the asymmetric T-shape single slit also exhibit pronounced performances. Usually, the sensitivity of the sensors (nm/RIU) is defined as the shift in the resonance wavelength per unit variations of the refractive index, and the overall performance is characterized by the “figure of merit” (FOM) that represents the sensitivity value divided by the resonance spectral bandwidth.22,23 For the symmetric and asymmetric T-shape single slit with the lateral dimension of only about 2 μm, they have the same sensitivity value of about 430 nm/RIU. This is owing to the nearly equal light−matter interaction length in the sensing process. However, the FOM is quite different for the symmetric and asymmetric T-shape single slit cases because they have different spectral bandwidths. In the symmetric T-shape single slit, the FOM is about 6, whereas the FOM reaches about 14 for the asymmetric T-shape single slit. Thus, the asymmetric T-shape single slit shows higher performances (higher FOM and more compact size) for the sensing applications. Therefore, the EIT-like transmission spectra in the asymmetric T-shape single slit possess pronounced features in the sensing and modulating applications. In summary, we experimentally demonstrated an EIT-like optical response in a simple metallic structure, which comprised two grooves of slightly detuned widths in immediate contacting with a single nanoslit and then was spin-coated with a PVA film. This was a unit-cell structure with the footprint of only about 0.9 μm2. Because of the deep groove in the T-shape single slit and the tight filed confinement by the dielectric-coated film, the generated SPPs propagating along the groove bottom were highly reflected by the groove steps and then scattered by the nanoslit and interfered with the directly transmitted light from the nanoslit. In the asymmetric T-shape single slit, the scattered light from the two grooves with slightly detuned groove widths were in antiphase with each other and thus interference destructively, resulting in the plasmon-induced transparency. FEM simulation results showed that the EIT-like transmission spectra exhibit nearly the same interference contrast but much narrow bandwidth, compared with that in the corresponding symmetric T-shape single slit. The proposed structure was quite simple and used a bulk incident wave, which allows easy fabrication and convenient experimental observation of the EIT-like optical response. In the experiment, the EIT-like transmission spectrum was successfully demonstrated in the proposed unit-cell structure with the lateral dimension of only about 2 μm. Here, the utilization of bulk incident waves may expand the physics of the EIT-like plasmonic resonances. These pronounced features, such as the increased quality factor, ultracompact size, easy fabrication, and experimental observation, in the compact asymmetric T-shape single slit are greatly favorable for the ultrasmall plasmonic devices such as sensors, modulators, and switches in highly integrated optics.

Figure 4. Experimental results of the transmission spectra for (a) the symmetric and (b) asymmetric T-shape single slits. Corresponding simulation results of the transmission spectra for (c) the symmetric and (d) asymmetric T-shape single slits.

transmission spectrum. The transmittance nearly reaches the maximum at λ = 870 nm, and the bandwidth of the EIT-like spectrum becomes Δλ′fwhm ≈ 30 nm. This is quite different from that in the symmetric T-shape single slit. In order to verify the experiment results, we made the corresponding calculations using the FEM method, and the results are shown in Figure 4c,d. It is noted that the experiment results agree well with the FEM simulation results. Here, the slight deviation from the FEM simulation results is mainly owing to the fabrication defects of the sample (Figure 3a) and the dispersion of the PVA film. From the above results and analysis, we know that the asymmetric single slit have not only smaller lateral dimension but also the narrower bandwidth. These excellent features would greatly benefit the compact plasmonic sensors, modulators, and switches in highly integrated optics. To demonstrate the validity of the benefit to the performances of the compact plasmonic devices such as sensors and modulators, further calculations were made, and the results are shown in Figure 5. For the symmetric T-shape single slit, the

Figure 5. Transmission spectra in the T-shape single slit with slightly detuned groove widths of (a) δ = 0 nm (wG1 = wG2 = 1120 nm) for nd = 1.50 and nd = 1.43 and (b) δ = 150 nm (wG1 = 1120 and wG2 = 970 nm) for nd = 1.50 and nd = 1.43.

■

absolute transmittance variation becomes about ΔT = 0.77 at the dip position when the refractive index of the dielectric film changes to nd = 1.43, as shown in Figure 5a. In the case for the asymmetric T-shape single slit, the transmission peak can turn to the dip completely for the same index change of the dielectric film (Figure 5b). Moreover, the absolute variation of the transmittance reaches about ΔT = 1.63, which is more than twice that in the symmetric T-shape single slit. This can

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (J.C.), [email protected] (Q.G.). Notes

The authors declare no competing financial interest. 2497

dx.doi.org/10.1021/nl300659v | Nano Lett. 2012, 12, 2494−2498

Nano Letters

■

Letter

Subradiant LSPR Sensing and a Tunable Fano Resonance. Nano Lett. 2008, 8, 3983−3988. (20) Johnson, P. B.; Christy, R. W. Optical constants of the noble metals. Phys. Rev. B: Solid State 1972, 6, 4370−4379. (21) Chen, J. J.; Li, Z.; Yue, S.; Gong, Q. H. Ultracompact surfaceplasmon-polariton splitter based on modulations of quasicylindrical waves to the total field. J. Appl. Phys. 2011, 109, 073102. (22) Sherry, L. J.; Chang, C.; Schatz, G. C.; Duyne, R. P. V. Localized Surface Plasmon Resonance Spectroscopy of Single Silver Nanocubes. Nano Lett. 2005, 5, 2034−2038. (23) Evlyukhin, A. B.; Bozhevolnyi, S. I.; Pors, A.; Nielsen, M. G.; Radko, I. P.; Willatzen, M.; Albrektsen, O. Detuned electrical dipoles for plasmonic sensing. Nano Lett. 2010, 10, 4571−4577.

ACKNOWLEDGMENTS This work was supported by the National Basic Research Program of China (Grants 2010CB923200, 2009CB930504, and 2007CB307001) and the National Natural Science Foundation of China (Grants 10804004, 11121091, 51172030, and 90921008).

■

REFERENCES

(1) Boller, K. J.; Imamoglu, A.; Harris, S. E. Observation of electromagnetically induced transparency. Phys. Rev. Lett. 1991, 66, 2593−2596. (2) Harris, S. E.; Field, J. E.; Imamoglu, A. Nonlinear optical processes using electromagnetically induced transparency. Phys. Rev. Lett. 1990, 64, 1107−1110. (3) Xu, Q. F.; Sandhu, S.; Povinelli, M. L.; Shakya, J.; Lipson, M. Experimental Realization of an On-Chip All-Optical Analogue to Electromagnetically Induced Transparency. Phys. Rev. Lett. 2006, 96, 123901. (4) Ebbesen, T. W.; Genet, C.; Bozhevolnyi, S. I. Surface-plasmon circuitry. Phys. Today 2008, 61, 44−50. (5) Gramotnev, D. K.; Bozhevolnyi, S. I. Plasmonics beyond the diffraction limit. Nat. Photonics 2010, 4, 83−91. (6) Papasimakis, N.; Fedotov, V. A.; Zheludev, N. I.; Prosvirnin, S. L. Metamaterial analog of electromagnetically induced transparency. Phys. Rev. Lett. 2008, 101, 253903. (7) Tassin, P.; Zhang, L.; Koschny, T.; Economou, E. N.; Soukoulis, C. M. Low-Loss Metamaterials Based on Classical Electromagnetically Induced Transparency. Phys. Rev. Lett. 2009, 102, 053901. (8) Liu, N.; Langguth, L.; Weiss, T.; Kastel, J.; Fleischhauer, M.; Pfau, T.; Giessen, H. Plasmonic analogue of electromagnetically induced transparency at the Drude damping limit. Nat. Mater. 2009, 8, 758− 762. (9) Liu, N.; Weiss, T.; Mesch, M.; Langguth, L.; Eigenthaler, U.; Hirscher, M.; Sonnichsen, C.; Giessen, H. Planar metamaterial analogue of electromagnetically induced transparency for plasmonic sensing. Nano Lett. 2010, 10, 1103−1107. (10) Zhang, S.; Genov, D. A.; Wang, Y.; Liu, M.; Zhang, X. PlasmonInduced Transparency in Metamaterials. Phys. Rev. Lett. 2008, 101, 047401. (11) Artar, A.; Yanik, A. A.; Altug, H. Multispectral Plasmon Induced Transparency in Coupled Meta-Atoms. Nano Lett. 2011, 11, 1685− 1689. (12) Zhang, J.; Bai, W.; Cai, L.; Xu, Y.; Song, G.; Gan, Q. Observation of ultra-narrow band plasmon induced transparency based on large-area hybrid plasmon-waveguide systems. Appl. Phys. Lett. 2011, 99, 181120. (13) Bozhevolnyi, S. I.; Evlyukhin, A. B.; Pors, A.; Nielsen, M. G.; Willatzen, M.; Albrektsen, O. Optical transparency by detuned electrical dipoles. New J. Phys. 2011, 13, 023034. (14) Kekatpure, R. D.; Barnard, E. S.; Cai, W.; Brongersma, M. Phase-coupled plasmon-induced transparency. Phys. Rev. Lett. 2010, 104, 243902. (15) Huang, Y.; Min, C.; Veronis, G. Subwavelength slow-light waveguides based on a plasmonic analogue of electromagnetically induced transparency. Appl. Phys. Lett. 2011, 99, 143117. (16) Zhang, Y.; Darmawan, S.; Tobing, L. Y. M.; Mei, T.; Zhang, D. H. Coupled resonator-induced transparency in ring-bus-ring MachZehnder interferometer. J. Opt. Soc. Am. B 2011, 28, 28−36. (17) Han, Z.; Bozhevolnyi, S. I. Plasmon-induced transparency with detuned ultracompact Fabry-Perot resonators in integrated plasmonic devices. Opt. Express 2011, 19, 3251. (18) Chen, J. J.; Li, Z.; Lei, M.; Yue, S.; Xiao, J. H.; Gong, Q. H. Broadband unidirectional generation of surface plasmon polaritons with dielectric−film−coated asymmetric single−slit. Opt. Express 2011, 19, 26463−26469. (19) Hao, F.; Sonnefraud, Y.; Dorpe, P. V.; Maier, S. A.; Halas, N. J.; Nordlander, P. Symmetry Breaking in Plasmonic Nanocavities: 2498

dx.doi.org/10.1021/nl300659v | Nano Lett. 2012, 12, 2494−2498