Electromagnetic Energy Transport in Nanoparticle Chains via Dark

Jan 31, 2012 - We show electromagnetic energy transport with gold nanoparticles that were assembled into close-packed linear chains. The small interpa...
0 downloads 8 Views 281KB Size
Letter pubs.acs.org/NanoLett

Electromagnetic Energy Transport in Nanoparticle Chains via Dark Plasmon Modes David Solis, Jr.,‡ Britain Willingham,† Scott L. Nauert,† Liane S. Slaughter,† Jana Olson,† Pattanawit Swanglap,† Aniruddha Paul,† Wei-Shun Chang,† and Stephan Link*,†,‡ †

Department of Chemistry, ‡Department of Electrical and Computer Engineering, Laboratory for Nanophotonics, Rice University, Houston, Texas 77005, United States S Supporting Information *

ABSTRACT: Using light to exchange information offers large bandwidths and high speeds, but the miniaturization of optical components is limited by diffraction. Converting light into electron waves in metals allows one to overcome this problem. However, metals are lossy at optical frequencies and large-area fabrication of nanometer-sized structures by conventional top-down methods can be cost-prohibitive. We show electromagnetic energy transport with gold nanoparticles that were assembled into closepacked linear chains. The small interparticle distances enabled strong electromagnetic coupling causing the formation of low-loss subradiant plasmons, which facilitated energy propagation over many micrometers. Electrodynamic calculations confirmed the dark nature of the propagating mode and showed that disorder in the nanoparticle arrangement enhances energy transport, demonstrating the viability of using bottom-up nanoparticle assemblies for ultracompact opto-electronic devices. KEYWORDS: Surface plasmon resonance, gold nanoparticles, plasmon waveguiding, super- and subradiance, BlIPP

M

chain geometries, straight, bent, branched, and so forth, over large areas.21,23 In addition, chemically synthesized nanostructures are single crystalline, thereby minimizing losses due to scattering of electrons at grain boundaries.26,27 Finally, this synthesis method allows much smaller interparticle spacings (∼1−5 nm) that result in stronger plasmon coupling,19,20,22,25,28,29 which is one prerequisite for increasing propagation distances in these materials. Here, we show that we can increase the propagation distance in assembled chains of closely spaced gold nanoparticles by at least an order of magnitude compared to previous studies.13 This is made possible by exploiting enhanced interparticle coupling and the resulting dark plasmon modes that support low-loss energy propagation.30−35 Using a novel imaging method, bleach imaged plasmon propagation (BlIPP),36 we determined a plasmon propagation distance of L0 = 3.9 ± 0.6 μm for these gold nanoparticle chains when exciting the dark modes, which is in contrast to no measurable propagation when exciting the bright single nanoparticle mode. This propagation distance is long enough to be comparable to other promising waveguide structures such as gold nanowires.37,38 Electrodynamic calculations furthermore showed that plasmon propagation is supported over a large bandwidth and that disorder in the nanoparticle arrangement actually enhances energy transport.

inimizing losses is a critical goal in energy and information transport applications that range from electricity transmission to fiber optic communication.1−4 At the same time, a continuing demand for improved technology requires the creation of increasingly smaller transport structures.5 When using light as the transport vehicle, this goal is hampered by the optical diffraction limit, which puts a large lower boundary on the widths of optical communication fibers.4 This problem can be overcome by exciting coherent electron waves in metals,6−10 known as surface plasmons. However, metals are lossy at optical frequencies,1−3,11,12 and large-area fabrication of nanometer-sized structures by conventional top-down methods can be expensive. Several different waveguide geometries have been explored to guide surface plasmons in subwavelength dimensions.8,13−16 Nanoparticle chains are an attractive, complementary structure to straight nanowires. Complex chain geometries within which light can be transported around corners, including sharp 90° turns,17,18 could offer flexible interconnects between more rigid waveguides. In groundbreaking work,13 Maier et al. demonstrated experimentally that electromagnetic energy transport is indeed possible in chains of lithographically prepared silver nanorods with large inter-rod gaps (50 nm), but the relatively short propagation distance (∼500 nm) proved discouraging to further exploration of nanoparticle chain waveguides. A particularly interesting alternative to structures fabricated by a serial top-down process is parallel bottom-up assembly of metal nanoparticles into chains.19−25 This approach has the advantage of enabling cost-effective fabrication of arbitrary nanoparticle © 2012 American Chemical Society

Received: November 8, 2011 Revised: January 26, 2012 Published: January 31, 2012 1349

dx.doi.org/10.1021/nl2039327 | Nano Lett. 2012, 12, 1349−1353

Nano Letters

Letter

Supporting Information Figure S1 for the corresponding SEM image). In contrast, no measurable energy propagation was observed when the chains were excited at the bright single nanoparticle mode at 514 nm (Figure 2e−h). Supporting Information Figures S2 and S3 show the scattering and extinction spectra of single 50 nm gold nanoparticles and the chain assembly, respectively, and demonstrate several important points. First, these measurements confirm that exciting at 514 nm coincides with the single gold nanoparticle plasmon mode. Next, the nanoparticle chain extinction spectrum shows that the lowest energy super-radiant plasmon mode was located at 1150 nm, confirming that dark modes were present around 785 nm.30,31,33 Furthermore, the absence of an intense extinction feature at even longer wavelengths confirms that, although the nanoparticles are closely spaced in the strong-coupling regime, the metal surfaces of the nanoparticles were not touching.39 A line section of the photobleaching intensity taken along the chain in Figure 2c yielded a dark-mode plasmon propagation distance of L0 = 4.2 μm, approximately an order of magnitude longer than measured previously13 for weakly coupled nanoparticle chains excited at a bright mode. Figure 2d shows the data together with the best fit to a previously developed model for BlIPP, which is based on the photobleaching kinetics of the dye due to direct laser excitation and indirect excitation by the plasmonic near-field that decays exponentially along the waveguide.

Figure 1. SEM characterization of nanoparticle chain assemblies. (a) SEM image of an array of chains made from 47 ± 4 nm gold nanoparticles. (b) SEM image of the nanoparticle chain highlighted by the red box in a. (c) SEM image taken for the green area indicated in b to show the close-packed arrangement of the nanoparticles. Nanoparticles were assembled in trenches formed in a polymer, which was then removed together with excess nanoparticles not assembled in the designed patterns to yield free-standing nanoparticle chains on a glass substrate. Typical chain structures measured 300 nm × 15 μm and were 5−6 nanoparticles wide, approximately 2−3 layers high, and contained approximately 1500 particles/layer. On the basis of the SEM images, the average interparticle distance is estimated to be below 5 nm with the separations likely being smaller considering the limited resolution of the SEM.

To achieve strong coupling, we assembled 50 nm gold nanoparticles into close-packed linear chains with widths of 300 nm and lengths of 15 μm as illustrated by the scanning electron microscopy (SEM) images in Figure 1. The chains were fabricated by depositing chemically synthesized gold nanoparticles into trenches that were written into a polymer resist using an electron beam. In a sense, this assembly method represents a combination of the best of top-down and bottomup methods in that small interparticle distances can be achieved because we use chemically synthesized nanoparticles but the chain itself or any other geometry can be created reproducibly using lithographic preparation methods. Importantly, the increased plasmon coupling for small interparticle distances allows the formation of subradiant plasmon modes.30−35 Subradiant plasmons correspond to collective excitations in which the phase for each nanoparticle can differ, leading to reduced overall charge separation and reduced coupling to light. The practical advantage in exploiting these dark modes for plasmon propagation is that they have decreased energy losses in comparison to the brightest, superradiant mode, in which all individual nanoparticle plasmons are in phase, or the bright single nanoparticle plasmon. The advantage of dark-mode plasmon propagation, as compared to bright-mode propagation, is illustrated in Figure 2. To measure these propagation distances, a novel technique called BlIPP was employed. BlIPP has been developed recently36 and details including the experimental setup (Scheme S1) are described in the Supporting Information. Briefly, it allows us to image the plasmon propagation by exploiting the photobleaching behavior of photoluminescent dyes. When BlIPP is performed at 785 nm, where dark modes are present, energy propagation persists over several micrometers for the gold nanoparticle chain (Figure 2a−d and

Figure 2. BlIPP of a nanoparticle chain for 785 and 514 nm circularly polarized excitation. (a) Fluorescence image of a chain coated with the dye cardiogreen and excited at 785 nm with a laser power of 66 nW. (b) Fluorescence image taken after exposure of the left end of the chain to 7.2 μW of 785 nm laser light for 20 min. Photobleaching due to plasmon propagation is most apparent in the difference image (c) and the normalized intensity line section (red) along the nanoparticle chain (d). Fitting the BlIPP data (green) yielded a plasmon propagation distance of L0 = 4.2 μm. The contribution from direct laser excitation (blue) to the photobleaching is also indicated. (e) Fluorescence image of a chain coated with rhodamine 6G and excited at 514 nm with a laser power of 31 nW. (f) Fluorescence image taken after exposure of the left end of the chain to 75 μW of 514 nm laser light for 5 min. The difference image (g) and the normalized intensity line section (red) along the nanoparticle chain (h) show no evidence of plasmon propagation as the BlIPP data was best described by photobleaching due to only direct laser excitation (blue). 1350

dx.doi.org/10.1021/nl2039327 | Nano Lett. 2012, 12, 1349−1353

Nano Letters

Letter

Because coupling strongly depends on the geometry of the waveguide tip, absolute values for the coupling efficiency of light to nanowires and nanoparticle chains are difficult to assess.40,41 Relative coupling efficiencies can be determined though by comparing the bleach intensities for the same incident power and exposure time as was done for parallel and perpendicular polarized excitation at 785 nm. We observed no significant difference in the propagation distances measured by BlIPP or relative coupling efficiency based on the bleach intensity between parallel and perpendicular excitation as shown in Supporting Information Figures S4 and S5, respectively. Interestingly, coupling to a propagating plasmon mode can also be achieved by exciting away from the waveguide end for these chains yielding again the same propagation distance (Supporting Information Figure S6). In contrast to nanowires, for which conversion of photons into plasmons only occurs at the nanowire ends, any nanoparticle within the chain can effectively act as an input coupler. In contrast, no energy propagation was observed at 514 nm within the spatial resolution of our BlIPP technique (Figure 2e−h and Supporting Information Figure S1). The difference image (Figure 2g) and the line section (Figure 2h), illustrate that photobleaching of the dye occurred only due to direct laser excitation as the bleached area was well described by only a Gaussian beam with no detectable contribution from an exponential plasmon decay. Independent photobleaching experiments without the nanoparticle chain (Supporting Information Figures S7 and S8) gave bleached areas with sizes comparable to the Gaussian laser component of the fit in Figure 2. The absence of long-range energy propagation for 514 nm excitation at the single nanoparticle resonance can be attributed to larger radiative losses in combination with interband absorptions.42 To ensure though that the results for 514 nm were not due to a difference in relative efficiency for coupling light into the waveguide, we varied the laser power and exposure time for 785 and 514 nm excitation. While for 785 nm the bleach intensity along the chain increased due to plasmon propagation (Supporting Information Figure S9), for 514 nm only direct laser excitation caused photobleaching despite using higher laser powers and longer exposure times (Supporting Information Figure S10). These results also rule out the possibility of thermally induced photobleaching. We can furthermore neglect the presence of the dye itself as the cause for the observed energy transport because destroying a section of the dye on a waveguide prior to the BlIPP measurement had no effect on the propagation distance (Supporting Information Figure S11). However, similar to other far-field measurements of surface plasmon propagation it is possible that BLIPP is more sensitive to certain modes. For example, leakage radiation is only a measure of propagating plasmons within the substrate light cone.43,44 A generalized Mie theory simulation both supports our experimental results and contributes new insights. The simulated near-field intensity for a 10 μm long nanoparticle chain excited at 785 nm indeed confirmed long plasmon propagation (Figure 3a). By fitting the intensity profile along the nanoparticle chain, an exponential decay distance of L0 = 1.7 μm was obtained (Figure 3b). The shorter theoretical propagation distance is likely because the actual nanoparticles were faceted and packed in three dimensions with separations possibly smaller than the minimum separation of 1 nm used in the calculations. The simulated charge distribution (Supporting

Figure 3. Generalized Mie theory simulations of plasmon propagation in gold nanoparticle chains. (a) Near-field intensity of the laser profile and plasmon propagation in a 1200 nanoparticle chain for excitation at 785 nm. (b) Decay of the width-averaged near-field intensity along the nanoparticle chain for 785 nm (red) compared to 514 nm (blue) excitation. The latter mimicked the laser intensity (cyan) confirming minimal energy propagation for excitation near the single nanoparticle plasmon resonance because of the strong intrinsic damping of gold. In contrast, the coupled plasmon modes excited at 785 nm sustained energy transport due to minimized losses with a propagation distance of L0 = 1.7 μm. A hexagonal close-packed (hcp) arrangement gave a shorter propagation distance of L0 = 1.1 μm (black, open symbols), illustrating that disorder enhances energy transport. (c) Section of the chain modeled in a, showing closely spaced 50 nm spherical gold nanoparticles with random positions to account for local disorder. The minimum surface to surface separation was always kept at 1 nm. (d) Propagation distances for different wavelengths demonstrating the large bandwidth that is achievable for plasmon propagation via lowloss subradiant modes. The errors were calculated by simulating different random nanoparticle arrangements.

Information Figure S12) confirmed the subradiant nature of this plasmon mode.28,30,33,45 However, 514 nm excitation resulted in minimal energy propagation. It is important to note that because of strong coupling that results from the small interparticle distances, 15 multipole modes were considered in these simulations. In contrast, it is primarily the dipole mode that contributes in assemblies with larger interparticle distances.13,46 Interestingly, the new information provided by the simulations is the discovery that including disorder of the nanoparticle arrangement in the simulations to represent the experimental structures increased the plasmon propagation distance, similar to effects observed in photonic quasicrystals.47 The nanoparticle chain was modeled as a random twodimensional assembly (Figure 3c). Comparing different random nanoparticle arrangements to an ordered hexagonal close-packed (hcp) array showed enhanced plasmon propagation if local disorder was included (Figure 3b and Supporting Information Figure S13). Furthermore, subradiant plasmon modes have the advantage of a large frequency bandwidth.30 To show this, we calculated plasmon propagation distances for excitation wavelengths in the range of 600−1050 nm (Supporting Information Figure S14). As summarized in Figure 3d, a propagation distance as long as L0 = 2.6 μm was obtained at 950 nm, while all values were larger than 1.5 μm for the entire bandwidth sampled. The long energy propagation in these nanoparticle chains was observed to be highly reproducible. The dark-mode 785 nm BlIPP experiment was repeated for 59 structures, and 1351

dx.doi.org/10.1021/nl2039327 | Nano Lett. 2012, 12, 1349−1353

Nano Letters

Letter

tionally, simulations suggest that the intrinsic heterogeneity of bottom-up nanoparticle preparations can actually increase propagation distances. Finally, the combination of large area printing methods to create ordered templates,48 together with bottom-up nanoparticle synthesis and assembly, opens the possibility of preparing inexpensive chain assemblies for novel plasmonic opto-electronic applications. These results have important implications for engineering ultracompact optoelectronic devices with nanoparticle plasmons in that they suggest the possibility for large area fabrication of nanoparticle assemblies that yield close interparticle distances, while tolerating or even exploiting the types of heterogeneity intrinsic to bottom-up assembly methods.



ASSOCIATED CONTENT

* Supporting Information S

Materials and Methods and Figures S1−S20. This material is available free of charge via the Internet at http://pubs.acs.org.



Figure 4. Energy localization at a chain defect. (a) Fluorescence image of a gold nanoparticle chain coated with cardiogreen and excited at 785 nm with a laser power of 63 nW. The white arrow highlights a region of increased fluorescence intensity. (b) Normalized photobleaching intensity along the nanoparticle chain obtained from the BlIPP analysis. The location of the area shown in a is highlighted by the black arrow and is associated with a sharp decrease of the bleach intensity indicating a break in the plasmon propagation for this waveguide. (c) SEM image of the same chain region demonstrates that energy propagation was hindered by the presence of a large defect composed of several missing nanoparticles.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was funded by the Robert A. Welch Foundation (C1664), the Office of Naval Research (N00014-10-1-0989), NSF (CHE-0955286), the ACS Petroleum Research Fund (50191DNI6), and a 3M Nontenured Faculty Grant. D.S. was supported by an NSF Graduate Research Fellowship Grant (0940902) and P.S. acknowledges support from the Royal Thai Government. We thank Dr. Christy Landes for help preparing the manuscript.

evidence was found for energy transport in ∼90% of the chains. Supporting Information Figures S15 and S16 show two additional examples of plasmon propagation along the entire length of the chain, as found for ∼40% of the waveguides yielding an average propagation distance of L0 = 3.9 ± 0.6 μm. Long range plasmon propagation also does not require broad chains (Supporting Information Figure S17), as reducing the width of the chains to about three nanoparticles had no measurable effect on the propagation distance. Further reducing the width of the chains to only one nanoparticle while avoiding larger gaps (see below) is, however, challenging with the current procedure. Simulations indicate that such a system, while offering stronger confinement, is not necessarily desirable, as the calculated propagation distance is reduced by a factor of 3 (Supporting Information Figure S18). A similar trade-off between confinement and propagation distance has also been observed for other plasmonic waveguides.44 Energy transport is strongly damped at large-scale defects along the nanoparticle chain (Figure 4 and Supporting Information Figures S19 and S20). The fluorescence image in Figure 4a, taken before photobleaching, shows an area of increased fluorescence intensity at 7.5 μm, which corresponds to an abrupt drop in the bleach intensity (Figure 4b). SEM images revealed a large hole defect in the nanoparticle arrangement at this position (Figure 4c). This explains why 50% of the chains showed energy transport but could not be fitted to an exponential decay. The larger hole defects, when examined in the SEM images, are of the same dimensions as the interparticle separations for top-down fabricated chains. This emphasizes why long-range energy transport was not observed before in such structures. In summary, we present evidence that dark-mode plasmon propagation in strongly coupled gold nanoparticles can be exploited to obtain micrometer-scale energy transport. Addi-



REFERENCES

(1) Boltasseva, A.; Atwater, H. A. Science 2011, 331, 290. (2) Brongersma, M. L.; Shalaev, V. M. Science 2010, 328, 440. (3) Soukoulis, C. M.; Wegener, M. Science 2010, 330, 1633. (4) Vlasov, Y. A.; O’Boyle, M.; Hamann, H. F.; McNab, S. J. Nature 2005, 438, 65. (5) Ozbay, E. Science 2006, 311, 189. (6) Barnes, W. L.; Dereux, A.; Ebbesen, T. W. Nature 2003, 424, 824. (7) Gather, M. C.; Meerholz, K.; Danz, N.; Leosson, K. Nat. Photonics 2010, 4, 457. (8) Gramotnev, D. K.; Bozhevolnyi, S. I. Nat. Photonics 2010, 4, 83. (9) Bozhevolnyi, S. I.; Volkov, V. S.; Devaux, E. s.; Laluet, J.-Y.; Ebbesen, T. W. Nature 2006, 440, 508. (10) De Leon, I.; Berini, P. Nat. Photonics 2010, 4, 382. (11) Stockman, M. I. Phys. Rev. Lett. 2007, 98, 177404. (12) Vakil, A.; Engheta, N. Science 2011, 332, 1291. (13) Maier, S. A.; Kik, P. G.; Atwater, H. A.; Meltzer, S.; Harel, E.; Koel, B. E.; Requicha, A. A. G. Nat. Mater. 2003, 2, 229. (14) Oulton, R. F.; Sorger, V. J.; Zentgraf, T.; Ma, R. M.; Gladden, C.; Dai, L.; Bartal, G.; Zhang, X. Nature 2009, 461, 629. (15) Walters, R. J.; van Loon, R. V. A.; Brunets, I; Schmitz, J.; Polman, A. Nat. Mater. 2010, 9, 21. (16) Koller, D. M.; Hohenau, A.; Ditlbacher, H.; Galler, N.; Reil, F.; Aussenegg, F. R.; Leitner, A.; List, E. J. W.; Krenn, J. R. Nat. Photonics 2008, 2, 684. (17) Brongersma, M. L.; Hartman, J. W.; Atwater, H. A. Phys. Rev. B 2000, 62, 16356. (18) Maier, S. A.; Brongersma, M. L.; Kik, P. G.; Meltzer, S.; Requicha, A. A. G.; Atwater, H. A. Adv. Mater. 2001, 13, 1501. 1352

dx.doi.org/10.1021/nl2039327 | Nano Lett. 2012, 12, 1349−1353

Nano Letters

Letter

(19) Barrow, S. J.; Funston, A. M.; Gomez, D. E.; Davis, T. J.; Mulvaney, P. Nano Lett. 2011, 11, 4180. (20) Braun, G. B.; Lee, S. J.; Laurence, T.; Fera, N.; Fabris, L.; Bazan, G. C.; Moskovits, M.; Reich, N. O. J. Phys. Chem. C 2009, 113, 13622. (21) Yin, Y.; Lu, Y.; Gates, B.; Xia, Y. J. Am. Chem. Soc. 2001, 123, 8718. (22) Chang, W. S.; Slaughter, L. S.; Khanal, B. P.; Manna, P.; Zubarev, E. R.; Link, S. Nano Lett. 2009, 9, 1152. (23) Cui, Y.; Bjork, M. T.; Liddle, J. A.; Sonnichsen, C.; Boussert, B.; Alivisatos, A. P. Nano Lett. 2004, 4, 1093. (24) Liu, K.; Nie, Z.; Zhao, N.; Li, W.; Rubinstein, M.; Kumacheva, E. Science 2010, 329, 197. (25) Nie, Z.; Fava, D.; Kumacheva, E.; Zou, S.; Walker, G. C.; Rubinstein, M. Nat. Mater. 2007, 6, 609. (26) Ditlbacher, H.; Hohenau, A.; Wagner, D.; Kreibig, U.; Rogers, M.; Hofer, F.; Aussenegg, F. R.; Krenn, J. R. Phys. Rev. Lett. 2005, 95, 257403. (27) Huang, J.-S.; Callegari, V.; Geisler, P.; Bruning, C.; Kern, J.; Prangsma, J. C.; Wu, X.; Feichtner, T.; Ziegler, J.; Weinmann, P.; Kamp, M.; Forchel, A.; Biagioni, P.; Sennhauser, U.; Hecht, B. Nat. Commun. 2011, 1, 150. (28) Fan, J. A.; Wu, C. H.; Bao, K.; Bao, J. M.; Bardhan, R.; Halas, N. J.; Manoharan, V. N.; Nordlander, P.; Shvets, G.; Capasso, F. Science 2010, 328, 1135. (29) Chen, J. I. L.; Chen, Y.; Ginger, D. S. J. Am. Chem. Soc. 2010, 132, 9600. (30) Willingham, B.; Link, S. Opt. Express 2011, 19, 6450. (31) Kawata, S.; Ono, A.; Verma, P. Nat. Photonics 2008, 2, 438. (32) Liu, M.; Lee, T.-W.; Gray, S. K.; Guyot-Sionnest, P.; Pelton, M. Phys. Rev. Lett. 2009, 102, 107401. (33) Nordlander, P. Nat. Photonics 2008, 2, 387. (34) Chen, H.-Y.; He, C.-L.; Wang, C.-Y.; Lin, M.-H.; Mitsui, D.; Eguchi, M.; Teranishi, T.; Gwo, S. ACS Nano 2011, 5, 8223. (35) Stockman, M. I. Nature 2010, 467, 541. (36) Solis, D.; Chang, W.-S.; Khanal, B. P.; Bao, K.; Nordlander, P.; Zubarev, E. R.; Link, S. Nano Lett. 2010, 10, 3482. (37) Krenn, J. R.; Lamprecht, B.; Ditlbacher, H.; Schider, G.; Salerno, M.; Leitner, A.; Aussenegg, F. R. Europhys. Lett. 2002, 60, 663. (38) Wild, B.; Cao, L.; Sun, Y.; Khanal, B. P.; Zubarev, E. R.; Gray, S. K.; Scherer, N. F.; Pelton, M. ACS Nano 2012, 6 (1), 472−482. (39) Berkovitch, N.; Orenstein, M. Nano Lett. 2011, 11, 2079. (40) Citrin, D. S.; Wang, Y.; Zhou, Z. P. J. Opt. Soc. Am., B 2008, 25, 937. (41) Li, Z. P.; Bao, K.; Fang, Y. R.; Huang, Y. Z.; Nordlander, P.; Xu, H. X. Nano Lett. 2010, 10, 1831. (42) Dickson, R. M.; Lyon, L. A. J. Phys. Chem. B 2000, 104, 6095. (43) Bouhelier, A.; Wiederrecht, G. P. Opt. Lett. 2005, 30, 884. (44) Maier, S. Plasmonics: Fundamentals and Applications; Springer: New York, 2007. (45) Hentschel, M.; Saliba, M.; Vogelgesang, R.; Giessen, H.; Alivisatos, A. P.; Liu, N. Nano Lett. 2010, 10, 2721. (46) Markel, V. A.; Sarychev, A. K. Phys. Rev. B 2007, 75, 085426. (47) Levi, L.; Rechtsman, M.; Freedman, B.; Schwartz, T.; Manela, O.; Segev, M. Science 2011, 332, 1541. (48) Chanda, D.; Shigeta, K.; Gupta, S.; Cain, T.; Carlson, A.; Mihi, A.; Baca, A. J.; Bogart, G. R.; Braun, P.; Rogers, J. A. Nat. Nanotechnol. 2011, 6, 402.

1353

dx.doi.org/10.1021/nl2039327 | Nano Lett. 2012, 12, 1349−1353