pubs.acs.org/NanoLett
Planar Metamaterial Analogue of Electromagnetically Induced Transparency for Plasmonic Sensing Na Liu,† Thomas Weiss,† Martin Mesch,† Lutz Langguth,§ Ulrike Eigenthaler,§ Michael Hirscher,§ Carsten So¨nnichsen,‡ and Harald Giessen*,† †
4. Physikalisches Institut, Universität Stuttgart, D-70569 Stuttgart, Germany, ‡ Max-Planck-Institut für Metallforschung, Heisenbergstrasse 3, D-70569 Stuttgart, Germany, and § Institute of Physical Chemistry, University of Mainz, Jakob-Welder-Weg 11, D-55128 Mainz, Germany
ABSTRACT We experimentally demonstrate a planar metamaterial analogue of electromagnetically induced transparency at optical frequencies. The structure consists of an optically bright dipole antenna and an optically dark quadrupole antenna, which are cut-out structures in a thin gold film. A pronounced coupling-induced reflectance peak is observed within a broad resonance spectrum. A metamaterial sensor based on these coupling effects is experimentally demonstrated and yields a sensitivity of 588 nm/RIU and a figure of merit of 3.8. KEYWORDS Metamaterials, plasmonic sensing, optical properties
T
he elimination of absorption via quantum interference in an atomic medium is known as electromagnetically induced transparency (EIT).1,2 This phenomenon allows for a very narrow transparency resonance in the absorption spectrum, which is highly desirable for sensing applications.3 Also, such a transparency feature is accompanied by extreme dispersion, which enables slowing down or even completely stopping light.4,5 In fact, EIT-like effects can occur in classical systems, such as coupled oscillators.6,7 In these classical analogues no pumping laser is necessary, significantly decreasing the complexity of experimental setups. Very recently, metamaterial analogues of EIT have attracted a lot of attention.8-10 Metamaterials are artificially structured media with unit cells much smaller than the operating wavelengths.11-14 In 2008, a plasmonic metamaterial analogue of EIT has been theoretically suggested using coupled optical resonators.8 A follow-up experiment based on the aforementioned theoretical proposal, however, showed broad Fano resonances.15 This is mainly due to the difficulties of manufacturing sharp and narrow gaps between the nanostructured resonators by electron beam lithography. Recently, this problem was circumvented by utilizing stacked metamaterials with two functional layers.16 Pronounced EIT-like features with high modulation depths were successfully demonstrated. However, the * To whom correspondence physik.uni-stuttgart.de. Received for review: 08/11/2009 Published on Web: 09/23/2009
should
© 2010 American Chemical Society
be
addressed,
manufacturing of this structure is time-consuming and involves several delicate multilayer stacking procedures, for example, surface planarization and layer registration. In this Letter, we demonstrate experimentally for the first time that a plasmonic EIT analogue can be achieved using a planar complementary metamaterial. Using Babinet’s principle,17,18 we propose a novel design composed of cut-out structures in a homogeneous gold film (see Figure 1). The cut-out design has two major advantages over the conventional one: First, it can be readily realized by focused-ion-beam writing or nanoimprint lithography.19 Second, in localized surface plasmon resonance (LSPR) sensor applications the sensing medium easily fills the voids, therefore facilitating the detection of refractive index changes. In accordance to Babinet’s principle, instead of enhanced transmission, a sharp reflectance peak within the broad spectral profile is established. In fact, accurate terminology would suggest using the expression “EIR-like effects”, as we introduce induced reflectance. In order to generate EIR-like features, coherently coupled bright and dark modes are introduced. A slot dipole antenna, which is strongly coupled to light, supports the spectrally broad bright mode. A slot quadrupole antenna, which is nonradiative in nature, supports the spectrally narrow dark mode. The quadrupole antenna has a much narrower linewidth compared to that of the dipole antenna due to the suppression of radiation damping.8,16 Due to close proximity, the two antennas are strongly coupled. As a result, destructive interference between two possible
giessen@
1103
DOI: 10.1021/nl902621d | Nano Lett. 2010, 10, 1103–1107
have a total area of 100 µm × 100 µm, consisting of about 15000 individual unit cells. Figure 1b shows an oblique incidence overview of a typical sample with a lateral displacement s ) 20 nm. An enlarged normal view is given in the inset of Figure 1b. The reflectance spectra of the samples at normal incidence were measured by a Fourier-transform infrared spectrometer with the H-field direction parallel to the dipole antenna (see Figure 1a). The experimental reflectance spectra in dependence on the lateral displacement s are characterized by black lines in Figure 2a. In the absence of lateral displacement, i.e., s ) 0 nm, only the dipole antenna can be excited. It is strongly coupled to light, suffering significant radiative losses. As a result, it leads to a broad reflectance dip as shown in the spectrum. In this case, the two antennas are not coupled due to the lack of symmetry breaking.20-22 By introduction of structural asymmetry with s ) 20 nm, a tiny reflectance peak appears within the broad resonance profile. When the lateral displacement is further enlarged, narrow and steep reflectance peaks are nicely observed in the respective spectra. Symmetry breaking allows the excitation of the nonradiative quadrupole antenna via the coupling with the dipole antenna. Particularly, the coupling strength increases by introducing more structural asymmetry. As a consequence, the two excitation pathways of the dipole antenna interfere destructively, giving rise to narrow reflectance features in the respective spectra as shown in Figure 2a. Due to nearly complete suppression of radiative losses, the linewidths of the reflectance features are solely limited by the intrinsic metal losses (Drude damping).16 The center frequencies of the two antennas are not perfectly overlapping due to a slightly too short quadrupole antenna owing to the fabrication process. In order to theoretically explore the optical properties of the samples, calculations based on scattering matrix theory23 and FDTD simulations (CST microwave studio) were carried out. The permittivity of the glass substrate is taken as 2.25. The optical constants of bulk gold in the infrared spectral regime are described by a Drude model with plasmon frequency ωpl ) 2π(2.175 × 1015) s-1 and damping constant ωc ) 2π(6.5 × 1012) s-1.24 To account for surface scattering, grain boundary effects in the thin gold film and inhomogeneous broadening, we used a three times higher damping constant than bulk.25,26 The calculated spectra are plotted as red curves in Figure 2b and show qualitative agreement with the experiment. The discrepancies are likely due to the fabrication tolerances in the experiment such as the inhomogeneity of the gap separation g. In order to better understand the underlying physics, the calculated magnetic field distribution at the reflectance peak as indicated by the red triangle in the spectrum for the case of s ) 60 nm (see Figure 2b) is presented in Figure 3. Clearly, at resonance the slot
FIGURE 1. (a) Schematic of the planar metamaterial design and the incident light polarization configuration. The geometry parameters are l1 ) 400 nm, w1 ) 80 nm, l2 ) 340 nm, w2 ) 90 nm, and g ) 45 nm. The gold film thickness is 30 nm on a glass substrate. The periods in both x and y directions are 800 nm. (b) Oblique view of the sample with lateral displacement s ) 20 nm. Inset: Enlarged view.
excitation pathways, namely, the direct excitation of the dipole antenna by the external light and the excitation by coupling with the quadrupole antenna, leads to EIR-like effects. We further experimentally demonstrate that this resonance can be carefully modified by varying the structural asymmetry. Experiments agree well with calculations. Subsequently, we show that our design has great potential for near-infrared LSPR sensing applications, a benefit substantiated by the narrow linewidths and high modulation depths of the EIR-like features. The reflectance peak position can be substantially influenced by materials which are in close vicinity of the nanostructure. Consequently, the pronounced reflectance signal can be utilized as a promising measure of mass concentration for biochemically relevant molecules, for example, proteins, glucose, and DNA. Figure 1a illustrates the geometry of the planar metamaterial design. It consists of three cut-outs in a 30 nm thick gold film on a glass substrate. The gaps between the horizontal cut-out (the slot dipole antenna) and the vertical cut-out pair (the slot quadrupole antenna) are g ) 45 nm. The lateral displacement of the dipole antenna with respect to the symmetry axis of the quadrupole antenna is defined as s. Electron micrographs of the samples fabricated by focused-ion-beam writing were obtained by field-emission scanning electron microscopy. All samples © 2010 American Chemical Society
1104
DOI: 10.1021/nl902621d | Nano Lett. 2010, 10, 1103-–1107
FIGURE 3. Calculated magnetic field distribution at the EIR-like resonance as indicated by the red triangle for the structure with s ) 60 nm in Figure 2b. The quadrupole antenna shows antisymmetric magnetic fields in the two slots, whereas the dipole antenna slot contains nearly no field.
quadrupole antenna is excited with strong magnetic fields at its ends, whereas there is nearly no magnetic field in the slot dipole antenna, which validates the destructive interference effect at the reflectance peak. Finally, we demonstrate that our metamaterial design may serve as a highly efficient LSPR sensor in the nearinfrared. This will open a new route for metamaterial applications toward sensing of chemical and biomedical molecules with different mass concentrations as well as detecting chemical reactions in a nanoenvironment. Changes by the presence of molecules close to the gold nanostructures can be detected by measuring the shift of the sharp EIR-like feature. Figure 4 shows the results of a proof-of-principle experiment, displaying the measured reflectance spectra for one of the structures (s ) 60 nm) with water (n ) 1.332) and 25% aqueous glucose solution (n ) 1.372) on the sample surface, respectively. A clear shift of the reflectance peak to longer wavelength is visible when changing water to glucose solution due to the increase of the refractive index of the liquid (see the inset of Figure 4). The sensitivity (nm/RIU) of the nanostructure ensemble, which is the shift in resonance wavelength per unit change of refractive index, is ≈588 nm/RIU. This result is remarkable even when compared with most single nanoparticle sensors. In order to provide better quantification, we also evaluated our sensor using the “figure of merit” (FOM) introduced by Sherry et al.27 It is defined as the sensitivity value divided by the resonance linewidth at half-maximum. In our case, the resonance linewidth of the EIR-like peak is ≈153 nm and it leads to a FOM ≈ 3.8. If the reflectance dip at around 180 THz were considered for sensing, the FOM would be 5.3. These values are among the highest FOMs for nanostructure LSPR sensors in the near-infrared.28 In fact, our sensor
FIGURE 2. Experimental and calculated reflectance spectra in dependence of lateral displacement s. Scanning electron microscopy images of the corresponding structures are shown in the left column. The scale bar is 200 nm. (a) Experimental reflectance spectra with air adjacent to the gold nanostructure are plotted as black curves. (b) Calculated reflectance spectra based on scattering matrix theory are plotted as red curves. The modulation depth of the EIR-like reflectance peak can be tuned by varying the lateral displacement s, which corresponds to the amount of symmetry breaking. © 2010 American Chemical Society
1105
DOI: 10.1021/nl902621d | Nano Lett. 2010, 10, 1103-–1107
dielectric surrounding. Our work has the potential to stimulate a lot of attention on metamaterial applications beyond negative refraction,12 superlensing,13 and invisibility cloaking.14 In addition, the concept of a plasmonic analogue of EIT might be a pathway toward the realization of low-loss metamaterials at optical frequencies.10 Also, our planar design could be a building block for successive stacking of multiple metamaterial slabs29,30 in order to construct “slow light” devices.4 Acknowledgment. We acknowledge S. Hein for his metamaterial visualizations. We acknowledge R. Ameling, H. Gra¨beldinger, and M. Ubl for technical assistance. This work was financially supported by Deutsche Forschungsgemeinschaft (SPP1391 and FOR557), by Landesstiftung BW, and by BMBF (13N9155 and 13N10146). REFERENCES AND NOTES (1) FIGURE 4. Experimental tuning of the EIR-like reflectance spectrum for the sample with s ) 60 nm by changing the liquid which is adjacent to the gold nanostructures from water to 25% aqueous glucose solution. The increase of the refractive index from water (n ) 1.332) to 25% glucose solution (n ) 1.372) causes a redshift of the resonance. Inset: enlarged figure of the reflectance peaks as highlighted by the dashed box.
(2) (3) (4) (5) (6) (7)
exhibits the highest values of sensitivity and FOM of any LSPR sensor that has been manufactured lithographically and not by bottom-up synthesis techniques. EIR-like effects generate narrow and steep reflectance features due to substantial reduction of radiative losses. This makes our design ideal for metamaterial sensors with high FOM. It is noteworthy that our design can also be utilized to realize a single particle sensor with a very small mode volume because the EIR-like spectral features are solely due to the resonant behavior of individual nanostructures. The theoretically predicted values (for s ) 60 nm) for sensitivity and FOM of the EIR-like feature are 725 nm/RIU and 7.4, respectively. Our experimentally realized sensor hence performs close to the theoretically predicted limit, only hampered by unaccounted losses in the metal for example due to focused ion beam patterning induced gallium impregnation and due to fabrication tolerances in the experiment. Improved fabrication methods to reduce structural errors due to inhomogeneities or elimination of gallium impregnation using nanoimprint lithography19 will result in better agreement with the theoretical simulations. In conclusion, we have investigated EIR-like phenomena using complementary planar metamaterials. Sharp reflectance peaks with narrow linewidths were observed in our experiment. The results agree well with theoretical predictions. A metamaterial LSPR sensor with high sensitivity and FOM has been experimentally demonstrated. The resonance position depends very sensitively on the © 2010 American Chemical Society
(8) (9) (10) (11) (12) (13) (14) (15) (16) (17)
(18) (19) (20) (21) (22) (23) (24)
1106
Boller, K. J.; Imamoglu, A.; Harris, S. E. Phys. Rev. Lett. 1991, 66, 2593–2596. Fleischhauer, M.; Imamoglu, A.; Marangos, J. P. Rev. Mod. Phys. 2005, 77, 633–673. Lal, S.; Link, S.; Halas, N. J. Nat. Photonics 2007, 1, 641–648. Hau, L. V.; Harris, S. E.; Dutton, Z; Behroozi, C. H. Nature 1999, 397, 594–598. Shvets, G.; Wurtele, J. S. Phys. Rev. Lett. 2002, 89, 115003. Xu, Q. F.; Sandhu, S.; Povinelli, M. L.; Shakya, J.; Lipson, M. Phys. Rev. Lett. 2006, 96, 123901. Alzar, C. L. G.; Martinez, M. A. G.; Nussenzveig, P. Am. J. Phys 2002, 70, 37–41. Zhang, S.; Genov, D. A.; Wang, Y.; Liu, M.; Zhang, X. Phys. Rev. Lett. 2008, 101, No. 047401. Papasimakis, N.; Fedotov, V. A.; Zheludev, N. I. Phys. Rev. Lett. 2008, 101, 253903. Tassin, P.; Zhang, L.; Koschny, T.; Economou, E. N.; Soukoulis, C. M. Phys. Rev. Lett. 2009, 102, No. 053901. Soukoulis, C. M.; Linden, S.; Wegener, M. Science 2007, 315, 47– 49. Shalaev, V. M. Nat. Photonics 2006, 1, 41–48. Pendry, J. B. Phys. Rev. Lett. 2000, 85, 3966–3969. Valentine, J.; Li, J.; Zentgraf, T.; Bartal, G.; Zhang, X. Nat. Mater. 2009, 8, 568–571. Verellen, N.; Sonnefraud, Y.; Sobhani, H.; Hao, F.; Moshchalkov, V. V.; Van Dorpe, P.; Nordlander, P.; Maier, S. A. Nano Lett. 2009, 9, 1663–1667. Liu, N.; Langguth, L.; Weiss, T.; Ka¨stel, J.; Fleischhauer, M.; Pfau, T.; Giessen, H. Nat. Mater. 2009, 8, 758–762. Zentgraf, T.; Meyrath, T. P.; Seidel, A.; Kaiser, S.; Giessen, H.; Rockstuhl, C.; Lederer, F. Phys. Rev. B 2007, 76, No. 033407. Babinet’s principle implies to interchange transmittance with reflectance and to rotate the polarization by 90° (interchange E and H) when going from the original to the complementary structure. Liu, N.; Kaiser, S.; Giessen, H. Adv. Mater. 2008, 20, 4521–4525. Wu, W.; Yu, Z. N.; Wang, S. Y.; Williams, R. S.; Liu, Y. M.; Sun, C.; Zhang, X.; Kim, E.; Shen, Y. R.; Fang, N. X. Appl. Phys. Lett. 2007, 90, No. 063107. Hao, F.; Nordlander, P.; Burnett, M. T.; Maier, S. A. Phys. Rev. B 2007, 76, 245417. Hao, F.; Sonnefraud, Y.; Dorpe, P. V.; Maier, S. A.; Halas, N. J.; Nordlander, P. Nano Lett. 2008, 8, 3983–3988. Hao, F.; Nordlander, P.; Sonnefraud.; Dorp, P. V.; Maier, S. A. ACS Nano 2009, 3, 643–652. Weiss, T.; Granet, G.; Gippius, N. A.; Tikhodeev, S. G.; Giessen, H. Opt. Express 2009, 17, 8051–8061. Ordal, M. A.; Long, L. L.; Bell, S. E.; Bell, R. R.; Alexander, R. W.; Ward, C. A. Appl. Opt. 1983, 22, 1099–1120. DOI: 10.1021/nl902621d | Nano Lett. 2010, 10, 1103-–1107
(25) Zhang, S.; Fan, W.; Malloy, K. J.; Brueck, S. R. J.; Panoiu, N. C.; Osgood, R. M. J. Opt. Soc. Am. B 2006, 23, 434–438. (26) Dolling, G.; Enkrich, C.; Wegener, M.; Soukoulis, C. M; Linden, S. Science 2006, 312, 892–894. (27) Sherry, L. J.; Jin, R.; Mirkin, C. A.; Schatz, G. G.; Van Duyne, R. P. Nano Lett. 2006, 6, 2060–2065.
© 2010 American Chemical Society
(28) Liao, H. W.; Nehl, C. L.; Hafner, J. H. Nanomedicine 2006, 1, 201– 208. (29) Liu, N.; Guo, H. C.; Fu, L. W.; Kaiser, S.; Schweizer, H.; Giessen, H. Nat. Mater. 2008, 7, 31–37. (30) Liu, N.; Liu, H.; Zhu, S. N.; Giessen, H. Nat. Photonics 2009, 3, 157–162.
1107
DOI: 10.1021/nl902621d | Nano Lett. 2010, 10, 1103-–1107
