Light-Bending Nanoparticles - Nano Letters (ACS Publications)

Nikolay A. Mirin and Naomi J. Halas*. Department of Chemistry, Department of Electrical and Computer Engineering, and the Laboratory for Nanophotonics...
3 downloads 0 Views 1MB Size
NANO LETTERS

Light-Bending Nanoparticles

2009 Vol. 9, No. 3 1255-1259

Nikolay A. Mirin†,§ and Naomi J. Halas*,†,‡,§ Department of Chemistry, Department of Electrical and Computer Engineering, and the Laboratory for Nanophotonics, Rice UniVersity, 6100 Main Street, Houston, Texas 77005 Received January 20, 2009; Revised Manuscript Received February 9, 2009

ABSTRACT Metallic nanostructures with their geometry-dependent optical resonances are a topic of intense current interest due to their ability to manipulate light in ways not possible with conventional optical materials. A particularly fascinating aspect of these systems is the recently realized possibility of creating optical frequency “magnetic plasmon” responses of comparable magnitude to the “electric plasmon” response. Here we show that Au nanocups at their magnetoinductive resonance have the unique ability to redirect scattered light in a direction dependent on cup orientation, as a true three-dimensional nanoantenna.

The interaction of light with small designed particles and structures gives rise to an increasing number of phenomena of potentially dramatic technological importance, such as metamaterials,1-11 superlens focusing,11-14 and enhanced spectroscopy.15,16 Thus far, highly symmetric nanoparticles have been studied extensively, due to their relative ease of fabrication by chemical methods and straightforward theoretical models. In these structures, the far-field scattering pattern always follows the direction of incident light (e.g., the classic cos2 θ dipole scattering pattern) irrespective of particle orientation. Simple symmetry breaking, for example, deforming a sphere into an axially symmetric elongated structure17-19 to create two plasmon modes associated with the long and short axes, is insufficient to change this behavior. The split ring resonator (SRR),1-5,10,20 a conductive structure interrupted by a dielectric gap, allows magnetoinductive coupling of incident radiation by inducing a charge separation of the resonant current in the structure during each half-cycle of the incident wave. The optical frequency magnetic response of SRRs has provided a key mechanism for the realization of metamaterials with a negative index of refraction at near-IR and optical frequencies.1,2,4,9 The planar nature of SRRs studied thus far restricts their light scattering ability to the fabrication plane, limiting the possibilities for redirecting scattered resonant light. Metallic half-shell “nanocups” can be considered to be three-dimensional optical frequency analogues of SRRs.20 Their three-dimensional structure removes the geometrical restrictions of planar SRRs fabricated by current methods. Suspensions of randomly oriented nanocups21-23 and closepacked arrays of strongly interacting nanocups and nano* Corresponding author, [email protected]. † Department of Chemistry. ‡ Department of Electrical and Computer Engineering. § Laboratory for Nanophotonics. 10.1021/nl900208z CCC: $40.75 Published on Web 02/19/2009

 2009 American Chemical Society

voids24,25 do not facilitate observation of their unique light scattering properties. Here we report the fabrication and optical properties of oriented nanocups. Each nanostructure supports a special transverse plasmon mode with an asymmetric resonant electron current that controls the direction of scattered light, departing from the path of the incident wave. The resonant electron current couples to a radiating electric dipole parallel to the cup opening, emitting the scattered wave in the direction of the nanocup axis. In our measurements we observe both individual and coupled nanocup responses, where the latter results directly from the unique orientation-specific light-directing properties of these nanostructures. Nanocups were fabricated according to the procedure shown in Figure 1. Commercially available (Invitrogen) 80 nm diameter sulfate latex or polystyrene nanoparticles were randomly deposited on clean glass substrates functionalized with poly(diallydimethylammonium) (PDDA).26 A 20 nm layer of gold was evaporated (by e-beam) onto samples mounted at 0°, 30°, 45°, and 60° relative to the incident beam of evaporated metal. The top of each particle was covered with gold as well as the surface of the glass substrate around it, resulting in nanocups on the particles and submicrometer circular or elliptical holes around the particles. The orientation of the nanocups and the shape of the holes are determined by the angle of the substrate relative to the beam of evaporated metal. Scanning electron microscopy (SEM) images of a nanocup sample at two different angles are shown in Figure 1B. Elastomer liftoff was used to remove the cups from the metalized substrate and preserve their orientation.27 A two-component mixture of Dow Corning SYLGARD-184 poly(dimethylsiloxane) (PDMS) elastomer was poured on the top of each substrate, degassed at 10-2 Torr for 1 h, and cured for 36 h to form a transparent slab.

Figure 1. Fabrication procedure for randomly spaced arrays of nanocups with controlled orientation: (A) glass slide is functionalized with PDDA and patterned with polystyrene or latex colloidal particles; (B) directional Au evaporation at a specific angle and the resulting SEM images of the formed nanocup structure at different sample tilt angles; (C) elastomer deposition and curing; (D) lift-off of elastomer slab with embedded oriented nanocups.

When the slab was peeled from the substrate, it removed the nanocups and polymer cores and retained them in fixed positions while the perforated metal film remained on the substrate. The elastomer refractive index approximately matches that of the polystyrene nanocup cores. The extinction spectra were collected using Cary 5000 UV-vis-NIR spectrophotometer with baseline subtraction. A blank PDMS slab was used to measure the baseline. 1256

Finite element modeling (FEM) was performed with commercial software (COMSOL 3.4) using a three-dimensional model of scattered harmonic propagation. An Au nanocup was modeled as half of a prolate ellipsoid with a concentric spherical cavity. The plane normal to the axis of rotation of an ellipsoid divides it by two equal parts, one of which is the modeled nanocup. The cavity radius was 40 nm, which corresponds to the average radius of polystyrene cores used in the experiment. The outer ellipsoid semiaxes were 60 and 48 nm, respectively. This geometry closely matches the experimental conditions with a directional metal evaporation, which deposits the thickest Au layer on the top of the nanocup (20 nm) and the thinnest layer on the nanocup edges (8 nm). An experimentally obtained dielectric function of Au28 was used. In order to simulate the dielectric core and the suspending medium, the modeled nanocup was placed in the center of a cylindrical simulation space with 100 nm radius and 200 nm height, a dielectric permittivity of 2, and the axis aligned with the propagation direction of the incident wave. A nanocup has two principal optical modes, one axial and one transverse, as is typical for a system with axial symmetry (Figure 2). The axial plasmon resonance (Figure 2A) of a nanocup is the significantly weaker mode, and the scattering pattern is determined by the illumination direction. This resonance corresponds to the dipole-driven, electroinductive plasmon response of the structure. The transverse nanocup scattering resonance (Figure 2B) is much stronger, and for p-polarization, the scattering direction and amplitude are determined by the nanocup orientation, not the direction of the incident light. This resonance corresponds to the currentdriven, magnetoinductive plasmon response of the structure, which is confirmed by the magnetic field enhancement plots presented in Figure 2. s-Polarized excitation of the transverse resonance produces an isotropic scattering pattern. This property was examined experimentally by studying the optical response of oriented Au nanocups. Experimental extinction spectra of nanocups oriented at 0°, 30°, 45°, and 60°, for the case of normal incidence optical excitation, are shown in Figure 3, accompanied by theoretical spectra calculated using FEM. The experimental geometries are depicted in panels A and B of Figure 3 for light p-polarized and s-polarized relative to the nanocup axes, respectively. In the spectra, the axial plasmon resonance occurs near 600 nm and is observed when the E field is parallel to the cup rotational axis. The axial resonance appears most strongly in the spectra at this particular orientation and according to FEM is dominated by absorption. Figure 2A shows the scattering component of this resonance, which changes from nonresonant Rayleigh scattering (illustrated charge distribution) to a resonant plasmon scattering regime. A transverse resonance occurs near 800 nm, when the E field is normal to the cup rotational axis. The decrease in amplitude of this resonance with increasing cup angle for p-polarized light (Figure 3, parts C and E) is due to the orientation dependence of this plasmon mode as shown in Figure 2B. The small higher order resonances that occur near 570 nm (Figure 3E) and 700 nm (Figure 3F) in the theoretical simulation are Nano Lett., Vol. 9, No. 3, 2009

Figure 2. Individual nanocup plasmon resonances, magnetic field enhancement (E0 ) 1 V/m; H0 ) 2.7·10-3 A/m), and calculated far field angular scattering (red) for different nanocup orientations relative to the incident light. (A) High-energy axial electroinductive resonance with no incident light redirection. (B) Low-energy transverse magnetoinductive resonance with directional scattering and intensity dependent upon angle of incidence for the case of p-polarization. Scattered optical field intensity is normalized for each of the two modes and plotted with a magnification factor of 1 unless specified otherwise.

characteristic of an asymmetric core-shell plasmonic nanostructure with a nonuniform metallic coating.17,18 These higher order resonances contribute slightly to the experimental spectrum Figure 3D, overlapping with the 800 nm transverse plasmon resonance. The experimental plasmon linewidths are larger than those obtained in the theoretical simulation because of core size and metal thickness distribution (inhomogeneous broadening) with additional likely contributions from electron scattering at the metal interfaces.29 The spectrum changes from transverse mode dominant to axial mode dominant with increasing angle for p-polarized light (Figure 3, parts C and E). In contrast, with s-polarized light excitation, the transverse mode shows no angular dependence (Figure 3, parts D and F), consistent with Nano Lett., Vol. 9, No. 3, 2009

Figure 3. Normal incidence extinction measurement for slab samples with different internal nanocup orientations and comparison with theoretical modeling of an individual nanocup. (A) The geometry of p-polarized nanocup excitation. (B) The geometry of s-polarized nanocup excitation. (C) The experimental p-polarized extinction measurement shows a decrease of the transverse nanocup plasmon mode near 800 nm and an increase of the axial nanocup plasmon mode near 600 nm as the nanocup angle (θcup) increases. A lower energy coupled nanocup resonance is observed near 1350 nm. (D) The experimental s-polarized extinction measurement shows the transverse nanocup resonance near 800 nm, the coupled cup resonance near 1350 nm and no significant spectral change as a function of nanocup angle (θcup). (E) Theoretical modeling (FEM) for the incident plane wave polarized in the plane of nanocup rotation shows a decrease in amplitude of the transverse nanocup mode near 800 nm and an increase in the axial nanocup mode near 600 nm with increasing θcup. (F) Theoretical modeling (FEM) for the incident plane wave polarized perpendicular to the plane of nanocup rotation shows the transverse nanocup resonance near 800 nm and no axial resonance.

the angular-independent scattering amplitudes shown in Figure 2B. In addition to these individual nanocup resonant modes, a coupled nanocup resonance can be observed near 1350 nm in our experimental spectra (Figure 3, parts C and D). This feature results from the interaction, or hybridization,30 of transverse nanocup resonances on adjacent nanocups in our samples, a feature not present in our theoretical simulations, where only an individual nanocup was studied (Figure 3, parts E and F). Theoretical simulations of adjacent coupled nanocups exhibit a resonance in this wavelength range, consistent with this assignment. The presence of coupled-nanocup resonances in parts C and D of Figure 3 provides an opportunity to examine how the local interaction between magnetoinductive modes in adjacent nanostructures depends on nanocup orientation. The possible configurations of a pair of coupled nanocups are illustrated in Figure 4. The transverse coupled cup 1257

Figure 5. (A) Oblique p-polarized incidence extinction measurement geometry, where θ0 ) [-60°, -45°, -30°, 0°, 30°, 45°, 60°]. Individual (600 nm, axial; 800 nm, transverse) and coupled (1350 nm) cup resonances, where θeff is the angle between the nanocup rotational axis and the incident beam direction as determined by angle of incidence (θ0), PDMS refraction, and nanocup orientation (θcup). Extinction measurements for (B) normal internal nanocup orientation showing coupled cup resonance for all θeff. (C) 45° rotated cups, showing no coupled cup resonance at any θeff.

incident light (Figure 4D), the transverse resonances of adjacent cups do not interact for any angle of incident light. To examine this property experimentally, oblique incidence extinction for nanocups in the geometry shown in Figure 5A was studied using θ0 ) [-60°, -45°, -30°, 0°, 30°, 45°, 60°]. The effective nanocup orientation (θeff) in this case is related to the angle of incidence (θ0) and the nanocup orientation (θcup) in the slab by Snell’s law θeff ) θcup - sin-1 Figure 4. Induced nanocup dipoles and coupled cup resonance. Interacting nanocup dipoles for (A) normal incidence and normal cup orientation, (B) oblique p-polarized incidence on normally oriented cups, (C) normal incidence and cups rotated in the k-E plane, and (D) noninteracting nanocup dipoles for the case of normal incidence and cups rotated in the k-H plane.

resonance is excited when the two adjacent cups are aligned and, for nanocups in this orientation, should be independent of the orientation of input light (Figure 4, parts A and B). For s-polarization, this interaction occurs for all rotation angles (Figure 4C). For tilted cups θcup > 0° and p-polarized 1258

( ) sin θ0 nPDMS

As the angle of incidence is varied, the transverse to axial resonance transition with increasing cup angle shown in parts B and C of Figure 5 is in agreement with our observations shown in Figure 3. In this case the coupled cup plasmon appears only in Figure 5B, for the case of aligned nanocups (θcup ) 0°). For the case of θcup ) 0° (Figure 4B) the transverse dipoles are always aligned and should couple to each other regardless of incident angle of optical excitation. Indeed, the coupled-nanocup resonance is observed at all angles of incident excitation measured. Conversely, for the case of rotated cups illustrated in Figure 4D, no coupled resonance should be observed at any angle of incidence. This was also observed (Figure 5C) for the case of nanocups Nano Lett., Vol. 9, No. 3, 2009

oriented at 45° for p-polarized light excitation, clearly indicating the effect of local nanocup orientation on this hybridized resonance. In summary, we have shown that the properties of plasmonic Au nanocups possess magnetoinductive modes that redirect scattered light in a direction dependent on particle orientation, an effect that also controls the interparticle coupling of plasmons in adjacent nanocups. Our optical extinction measurements confirm that the fabricated particles exhibit these predicted modes. Acknowledgment. This work was supported by the Robert A. Welch Foundation grant C-1220 and the Multidisciplinary University Research Initiative (MURI) grant W911NF-0401-0203. The authors thank Peter Nordlander, Serge Grabtchak, Lisa Brown, J. Britt Lassiter, Jared Day, Carly Levin, and Nathaniel K. Grady for the valuable input, discussions, and proofreading of the manuscript. References (1) Shelby, R. A.; Smith, D. R.; Schultz, S. Science 2001, 292, 77–79. (2) Linden, S.; Enkrich, C.; Wegener, M.; Zhou, J.; Koschny, T.; Soukoulis, C. M. Science 2004, 306, 1351–1353. (3) Dolling, G.; Enkrich, C.; Wegener, M.; Soukoulis, C. M.; Linden, S. Science 2006, 312, 892–894. (4) Soukoulis, C. M.; Linden, S.; Wegener, M. Science 2007, 315, 47– 49. (5) Liu, N.; Guo, H.; Fu, L.; Kaiser, S.; Schweizer, H.; Giessen, H. Nat. Mater. 2008, 7, 31–37. (6) Rockstuhl, C.; Zentgraf, T.; Meyrath, T. P.; Giessen, H.; Lederer, F. Opt. Express 2008, 16, 2080–2090. (7) Urzhumov, Y. A.; Shvets, G.; Fan, J.; Capasso, F.; Brandl, D.; Nordlander, P. Opt. Express 2007, 15, 14129–14145. (8) Shalaev, V. M.; Cai, W.; Chettiar, U. K.; Yuan, H.-K.; Sarychev, A. K.; Drachev, V. P.; Kildishev, A. V. Opt. Lett. 2005, 30, 3356– 3358.

Nano Lett., Vol. 9, No. 3, 2009

(9) Driscoll, T.; Basov, D. N.; Padilla, W. J.; Mock, J. J.; Smith, D. R. Phys. ReV. B: Condens. Matter 2007, 75, 115114. (10) Schurig, D.; Mock, J. J.; Justice, B. J.; Cummer, S. A.; Pendry, J. B.; Starr, A. F.; Smith, D. R. Science 2006, 314, 977–980. (11) Taminiau, T. H.; Stefani, F. D.; van Hulst, N. F. Opt. Express 2008, 16, 10858–10866. (12) Bryant, G. W.; Garcı´a de Abajo, F. J.; Aizpurua, J. Nano Lett. 2008, 8, 631–636. (13) Mu¨hlschlegel, P.; Eisler, H.-J.; Martin, O. J. F.; Hecht, B.; Pohl, D. W. Science 2005, 308, 1607–1609. (14) Pendry, J. B. Phys. ReV. Lett. 2000, 85, 3966–3969. (15) Aizpurua, J.; Bryant, G. W.; Richter, L. J.; Garcı´a de Abajo, F. J.; Kelley, B. K.; Mallouk, T. Phys. ReV. B: Condens. Matter 2005, 71, 235420. (16) Taminiau, T. H.; Stefani, F. D.; Segerink, F. B.; van Hulst, N. F. Nat. Photonics 2008, 2, 234–237. (17) Wang, H.; Wu, Y.; Lassiter, B.; Nehl, C. L.; Hafner, J. H.; Nordlander, P.; Halas, N. J. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 10856–10860. (18) Knight, M. W.; Halas, N. J. New J. Phys. 2008, 10, 105006–105015. (19) Wang, H.; Brandl, D. W.; Le, F.; Nordlander, P.; Halas, N. J. Nano Lett. 2006, 6, 827–832. (20) Cortie, M.; Ford, M. Nanotechnology 2007, 18, 235704–235709. (21) Charnay, C.; Lee, A.; Man, S.-Q.; Moran, C. E.; Radloff, C.; Bradley, R. K.; Halas, N. J. J. Phys. Chem. B 2003, 107, 7327–7333. (22) Liu, J.; Cankurtaran, B.; McCredie, G.; Ford, M. J.; Wieczorek, L.; Cortie, M. B. Nanotechnology 2005, 16, 3023–3028. (23) Liu, J.; Maaroof, A. I.; Wieczorek, L.; Cortie, M. B. AdV. Mater. 2005, 17, 1276–1281. (24) Cole, R. M.; Sugawara, Y.; Baumberg, J. J.; Mahajan, S.; Abdelsalam, M.; Bartlett, P. N. Phys. ReV. Lett. 2006, 97, 137401–137404. (25) Liu, J.; Cankurtaran, B.; Wieczorek, L.; Ford, M. J.; Cortie, M. AdV. Funct. Mater. 2006, 16, 1457–1461. (26) Mirin, N. A.; Hainey, M.; Halas, N. J. AdV. Mater. 2008, 20, 535– 538. (27) Wang, H.; Halas, N. J. Nano Lett. 2006, 6, 2945–2948. (28) Johnson, P. B.; Christy, R. W. Phys. ReV. B: Condens. Matter 1972, 6, 4370–4379. (29) Westcott, S. L.; Jackson, J. B.; Radloff, C.; Halas, N. J. Phys. ReV. B 2002, 66, 155431–155435. (30) Nordlander, P.; Oubre, C.; Prodan, E.; Li, K.; Stockman, M. I. Nano Lett. 2004, 4, 899–903.

NL900208Z

1259