Split Hole Resonator: A Nanoscale UV Light Source - Nano Letters

Jan 21, 2016 - Bonnie Charpentier is 2018 ACS president-elect. Bonnie Charpentier, senior vice president for regulatory, quality, and safety at Cytoki...
0 downloads 15 Views 2MB Size
Letter pubs.acs.org/NanoLett

Split Hole Resonator: A Nanoscale UV Light Source Pavel N. Melentiev,*,†,‡ Anton E. Afanasiev,† Arthur A. Kuzin,†,¶ Valeriy M. Gusev,† Oleg N. Kompanets,† Rinat O. Esenaliev,*,§ and Victor I. Balykin*,†,‡ †

Institute for Spectroscopy Russian Academy of Sciences, Phizicheskaya str., 5, Troitsk, Moscow, 142190 Russia Institute on Laser and Information Technologies of the Russian Academy of Sciences, Moscow, Troitsk, 142190, Russia ¶ Moscow Institute of Physics and Technology, Moscow reg., Dolgoprudny, 141700, Russia § The University of Texas Medical Branch, 301 University Boulevard, Galveston, Texas 77555, United States ‡

S Supporting Information *

ABSTRACT: Because of strong light absorption by metals, it is believed that plasmonic nanostructures cannot be used for generating intensive radiation harmonics in the ultraviolet (UV) spectral range. This work presents results of investigation of nonlinear optical interaction with a single gold nanostructure, the split-hole resonator (SHR) under the state-of-the-art experimentally realized conditions. To realize interaction with all spectral components of a 6 fs laser pulse several multipole plasmon resonances were simultaneously excited in the SHR nanostructure. To the best of our knowledge, this paper reports for the first time a strong nonlinear optical interaction at the frequencies of these resonances that leads to (i) the second harmonic generation, (ii) the third harmonic generation (THG), and (iii) the light generation at mixed frequencies. The THG near field amplitude reaches 0.6% of the fundamental frequency field amplitude, which enables the creation of UV radiation sources with a record high intensity. The UV THG may find many important applications including biomedical ones (such as cancer therapy). KEYWORDS: Plasmonics, nanoantennas, nonlinear optics, second-harmonic generation, third-harmonic generation, UV radiation laser fields cause another problem: because of high optical losses in plasmonic nanostructures,30 the nanostructures are heated rapidly and, as a consequence, undergo the catastrophic meltdown.31 In practice, the laser radiation intensity threshold for the nanostructure is about 1010 W/cm2, strongly limiting efficiency of harmonic generation and hence use of THG in practical applications.32 In this work, we present results of our investigations of the nonlinear optical interaction of near IR laser radiation with a gold nanostructure under the following experimental state-ofthe-art ultimate conditions: (1) the laser pulse duration is ultimately short (6 fs, which corresponds to two cycles of the laser pulse wave) to maximally reduce the thermal effects on the nanostructure; (2) the laser light intensity is ultimately high and close to the air ionization threshold, which provides maximum intensity at the fundamental frequency without optical breakdown; (3) the geometry of the nanostructure is optimal ensuring a record-high efficiency of the nonlinear optical interaction (the split-hole resonator; SHR);33 and (4) the SHR nanostructure is formed in a single crystal gold nanofilm that is flat on the atomic level. Under the chosen experimental conditions, several multipole plasmon resonances

P

lasmonic nanostructures allows for light concentration into a size considerably smaller than the light wavelength.1,2 This has found practical applications in many fields, such as photodetection,3 photovoltaics,4 optical microscopy with a nanometer resolution,5,6 biosensors,7 optical nanolithography,8 and so forth. The nonlinear optical properties of nanostructures allows one to considerably extend their applications due to the harmonics generation by the nanostructures,9−16 two-photon excited luminescence,17−20 and nonlinear four-wave mixing.21,22 Of separate practical interest is the generation of UV harmonics by gold nanoparticles inside a living tissue for the targeted action of the UV radiation on the tissue cells.23 In this approach, the UV radiation results from a strong nonlinear optical interaction of the visible and near-IR radiation (which falls into the so-called therapeutic window for tissues) with a gold nanoparticle. The mechanism of the optical action on the living cell is based on the absorption of light quanta by its chromosome. In this case, the absorption of even a single UV photon can be fatal to the living cell.24,25 The UV radiation generation by gold nanoparticles has been poorly studied. Only a few works on harmonics generation by gold nanoparticles are known26−28 and the efficiency of the nonlinear optical transformation of the visible and near-IR radiation into the UV radiation in the range 300−400 nm is low29 due to a strong UV absorption in gold. The nonlinear optical response of a nanostructure can be increased by using high-intensity laser fields. However, intense © 2016 American Chemical Society

Received: October 28, 2015 Revised: January 15, 2016 Published: January 21, 2016 1138

DOI: 10.1021/acs.nanolett.5b04373 Nano Lett. 2016, 16, 1138−1142

Letter

Nano Letters

Figure 1. (a) SEM image of a sample with an array of identical SHR nanostructures (the inset shows the SEM image of a single SHR nanostructure), (b) schematic of the experimental setup, (c) spectrum of laser radiation used to excite the SHR nanostructure, and (d) envelope of the laser pulse intensity measured with a SPIDER system.

810 nm. Sixteen nanostructures of the same size were arranged on the sample in the shape of a 4 × 4 matrix. The spacing between neighboring nanostructures in the matrix was 10 μm, which gave the possibility to focus the excitation laser radiation onto a single nanostructure and, correspondingly, to measure the generation of harmonics precisely by this particular nanostructure. The schematic of the experimental setup is presented in Figure 1b. The laser radiation with the central wavelength of 760 nm (Figure 1c) and pulse duration of 6 fs (two periods of the light wave; Figure 1d) was focused by a mirror objective (15×, NA = 0.3) such that its waist with a diameter of 4.7 μm was in the plane of a sample that was placed in the object plane of an inverted microscope. The peak intensity of the laser radiation was 8.6 × 1011 W/cm2. Excitation of plasmon resonances in the nanostructure can increase this intensity up to 1012−1013 W/cm2, which is near the intensity of air ionization (1013 −10 14 W/cm2 ). It should be specially emphasized that the precisely chosen geometry of the nanostructure allows for measurements at such a high intensity.33 The two-cycle laser pulse radiation has a broad emission spectrum that extends from 650 nm to 1 μm (Figure 1c). Such a large width of the emission spectrum requires dispersion control in the medium over the whole propagation path from the laser to the sample. Experimentally, the group velocity dispersion was controlled with an accuracy of up to 0.1 fs2 using an optical system which involved: (a) two dielectric mirrors, multiple reflection from which gathers the necessary negative group velocity dispersion (equal to −250 fs2); and (b) a system of optical wedges to produce the positive group velocity dispersion in the range between 0 and 250 fs2. The group velocity dispersion and the laser pulse duration were measured using a commercially available ultrashort pulse measurement technique, SPIDER (spectral phase interferometry for direct electric field reconstruction) system. During the measurements, the nonlinear crystal of the SPIDER system and the sample with nanostructures were arranged at the same distance from the laser as (in the object plane of the microscope with an accuracy of ±1 cm). This allowed us to guarantee with an accuracy of up to 0.2 fs that the laser pulse duration on the sample is the same as that measured by the SPIDER system. The nonlinear optical interaction of laser radiation with a single SHR nanostructure was investigated using an inverted microscope that was built based on a commercially available

can be simultaneously excited in the SHR nanostructure at fundamental frequencies. It was found for the first time that the excitation of multipole resonances considerably widens and enhances nonlinear optical interactions of the nanostructure with the laser radiation. We show that the SHR nanostructure is a source of intense UV radiation in the wavelength range of 250−300 nm. The use of SHRs in nonlinear nanoplasmonics is of special interest for the following reasons:33−35 (i) the nanohole of the SHR ensures the absence of a background from the excitation radiation that is strongly reduced by a low transmission of the nanohole; (ii) the metal film in which the SHR is made provides an efficient heat removal that ensures the stability of the nanostructure to high-intensity radiation and, therefore, highly efficient nonlinear optical transformations; (iii) the efficient heat removal by the metal film maintains the Q factor of the SHR plasmon resonance even at high intensities of the excitation field; and (iv) the polarizability of the nanorod is higher than that of the nanohole; therefore, the use of the nanorod inside the SHR ensures a high optical nonlinearity of the SHR nanostructure. A distinctive and significant optical feature of an SHR nanostructure is its multiresonant character due to the complex geometry. As was shown,33 plasmon modes in an SHR nanostructure are realized at a certain ratio of the nanohole diameter to the nanorod length. In this case, as will be shown below, several plasmon modes with different degrees of multipolarity can be excited at certain dimensions of the nanostructure. SHR nanostructures (Figure 1a) were prepared in a singlecrystal gold nanofilm (with a thickness of 200 nm) by the ionbeam lithography method using a tightly focused beam of Ga ions. The gold film (from PHASIS, Geneva, Switzerland) was obtained in the course of epitaxial growth on the surface of a mica substrate with a thickness of 100 μm. Electron microscopy of the gold film showed single-crystal flakes on the surface with a typical size of 3−5 μm. The single crystallinity of the nanofilm on the atomic level is of fundamental importance (see Supporting Information for details). In a single-crystal gold nanofilm, different SHR nanostructures with the same topology but with different dimensions were created. For all SHR nanostructures, the length of their nanorods was equal to the radius of their holes, while the width of the nanorod was roughly equal to a half of its length. In the SHR nanostructure of a minimal size, the hole diameter was 390 nm, while that in the nanostructure of a maximal size was 1139

DOI: 10.1021/acs.nanolett.5b04373 Nano Lett. 2016, 16, 1138−1142

Letter

Nano Letters Nikon Eclipse microscope (see Supporting Information for details). We did the measurements of THG both from bare mica substrate and from the SHR nanostructure. We could not detect any THG signal from the mica substrate. The measured dependence of the THG signal from the SHR nanostructure on illumination power is fitted with a cubic power law (see Supporting Information for details). Figure 2 presents the

Figure 3. Measured spectra of the second and third harmonics generated by a single SHR nanostructure. The red curve shows the spectral transmission of the filter that cuts off the fundamental frequency. The nanostructure parameters are as follows: the nanohole diameter is 700 nm and the nanorod dimensions are 350 nm × 150 nm. The SHR nanostructure was prepared in a gold film with 200 nm thickness.

Figure 4 presents results of processing of experimental data from Figure 3 that show that the laser radiation excites in the

Figure 2. Dependence of the third harmonic generation efficiency on the SHR nanostructure size. The efficiency of the third harmonic generation by a nanostructure with a maximal size was taken to be 100%. The insets show SEM images of SHR nanostructures with the corresponding efficiencies of the third harmonic generation.

radiation power measured at the third harmonic frequency from SHR nanostructures of different dimensions. These measurements showed that the third harmonic generation efficiency increases with the nanostructure diameter (Figure 2). The number of resonances in the THG spectrum (not shown) increases with the nanostructure diameter as well. As can be seen in Figure 2, the THG radiation power attains saturation with increasing SHR nanostructure dimension: as the nanostructure dimension reach 700 nm, the third harmonic generation efficiency increases considerably. However, upon further increase of the nanostructure dimension, the generation efficiency remains almost unchanged. It is likely that the measured “saturation” of the generation efficiency can be explained by the number of plasmon modes at the fundamental frequency that were excited in the SHR nanostructure, which, in turn lead to the appearance of a structure in the third harmonic generation spectrum. The number of excited modes is determined by the laser radiation spectral width and the nanostructure dimension. At small dimensions, the number of excited modes is determined only by the nanostructure dimension. On the contrary, at large dimensions of the nanostructure, the number of excited plasmon modes is determined only by the laser radiation spectral width. Figure 3 shows the emission spectrum of a nanostructure for which a maximal efficiency of the third harmonic generation was obtained. The spectrum consists of two series of spectral lines (a) the three lines in the spectral range of 380−470 nm correspond to the second harmonic generation and (b) the five lines in the range 240−300 nm are due to the third harmonic generation. The measured dependences of the harmonics generation power on the incident radiation power correspond to the degree of multiphotonicity of processes with n = 2 and 3 for the lines in the ranges of 380−470 nm and 240−300 nm, respectively. We note that the third harmonic spectrum is shown without a correction for the filter spectral transmission that cuts off the fundamental frequency (the filter transmission spectrum is shown in the figure with the red curve).

Figure 4. Generation of harmonics by an SHR nanostructure is caused by the excitation of multipole plasmon modes: (a) the spectral dependences of the field intensity of the ω1, ω2, ω3, and ω4 plasmon modes at the fundamental frequency and the calculated near-field distributions of the SHR nanostructure that correspond to them, (b) the measured emission spectrum of the second harmonic (dotted curve) and the calculated spectra of the second harmonic that correspond to the excitation of the ω1, ω2, ω3, and ω4 plasmon modes in the SHR nanostructure, and (c) the measured emission spectrum of the third harmonic and the calculated spectra of the third harmonic that correspond to the excitation of the ω1, ω2, ω3, and ω4 plasmon modes in the SHR nanostructure.

SHR nanostructure the following four plasmon modes at the fundamental frequency: ω1 is the dipole mode (the central radiation wavelength λ1 = 880 nm), ω2 is the multipole mode with n = 3 (λ2 = 800 nm), ω3 is the multipole mode with n = 4 (λ3 = 770 nm), and ω4 is the multipole mode with n = 5 (λ4 = 750 nm). The spectral dependences of the field intensity in these modes normalized to the amplitude are shown in Figure 4a. The tight focusing and the large width of the laser radiation spectrum (which corresponds to the two periods of the light 1140

DOI: 10.1021/acs.nanolett.5b04373 Nano Lett. 2016, 16, 1138−1142

Letter

Nano Letters

excitation of ω1 and ω2 plasmonic resonances by an ultrashort laser pulse with the duration of 6 fs. Therefore, the curves presented in Figure 4 show that the excitation of plasmon modes at the fundamental frequency produces resonant peaks in the second and third harmonics spectra. A small discrepancy of the width of the 3ω1 resonance with that observed in the experiment is most likely a consequence of a not quite accurate measurement of the 3ω1 resonance width, because it falls into the edge of the bandpass filter transmission spectrum (Figure 3). We also note that there is a noticeable asymmetry of the measured resonances that correspond to the resonant lines 2ω1 and 3ω1. The nature of this asymmetry can be explained by interference with corresponding fields that arise due to the generation of harmonics by the film surface.39 It is important to note that in the transmission measurements it is almost impossible to detect multipole resonances due to their fast decay in the far-fields. Because of that we could verify only a dipole resonance of the SHR nanostructure in our transmission measurements (ω1 on Figure 4a). We would like to emphasize that THG measurements (Figure 4c) are an efficient tool to investigate multipole resonances in a single nanostructure. The measured harmonics generation efficiency in the UV range (see Supporting Information for details) allowed us to evaluate the possibility of biomedical applications of the UV radiation from nanostructures. In the experiment, we found that the flux of UV photons from a single SHR nanostructure is N = 4 × 108 photon/s. This value is higher by 4 orders of magnitude than that obtained recently for THG in the UV28 and visible15,33 spectral ranges. According to the data from the literature,24,25 the flux of UV photons N = 0.01 photon/s suffices to perform the phototherapy of cells. In order to obtain such a flux of UV photons from a single SHR nanostructure, the intensity of the excitation laser radiation Iω= 1.3 × 10−12 J/ cm2 is necessary. According to the data,40 such intensity of the laser radiation is below maximum permissible exposure (MPE) for tissues; therefore, the targeted phototherapy of cells using a flux of UV photons from nanostructures seems feasible. In this work, we performed investigations of the nonlinear optical interaction of two-cycle laser pulse radiation with a single SHR plasmonic nanostructure. We showed that, due to the large spectral width of the laser radiation, several resonances of surface plasmons with different degrees of multipolarity (n = 2, 3, 4, and 5) can simultaneously be excited in the SHR nanostructure. Due to the excitation of plasmon resonances, the spectra of the second and third harmonics generated by a single SHR nanostructure have several resonance lines, which are caused by the two- and three-photon processes as well as the frequencymixing process. We showed that a single SHR nanostructure has a high optical nonlinearity that results in the high amplitude of the third harmonic radiation field of E3ω= 0.006 Eω. The harmonics radiation spectral range (extending from 240 to 470 nm) includes UV radiation and can be used for phototherapy (e.g., cancer phototherapy). Nanoparticle-mediated phototherapy of tumors and other abnormal tissues is typically based on heating of strongly absorbing (metal, carbon, or dye) nanoparticles by light (proposed in refs 41−43 for direct phototherapy without drugs and for enhancement of drug delivery in tumors). The UV radiation generation by nanoparticles studied in this work is another approach which may be used for tumor and cell

wave) allows for simultaneous excitation of the four multipole plasmon modes in the SHR nanostructure. It should be noted that the excited modes satisfy the equation λn = λα + λβ/n, where the constants are λα= 640 nm and λβ = 480 nm, and n = 1, 2, 3,... is the order of the mode multipolarity. The nature of this relation is determined by the resonant character (of the Fabry−Perot type) of excited multipoles, as it was shown in the case of a one-dimensional nanostructure in the shape of a nanorod,36 as well as in the case of a two-dimensional nanostructure in the shape of a split-ring resonator.37 The insets in Figure 4a present results of the numerical simulation of the near-field amplitude distribution for an SHR nanostructure exposed to a monochromatic depolarized Gaussian light beam with the parameters which were used in the experiment. The simulation was performed by the FDTD method. The nanostructure parameters were identical to that used in the measurements presented in Figure 3. The obtained near field amplitude distributions confirm the multipole character of excited resonances. Note that near field patterns for ω3 and ω4 modes are indistinguishable in calculations due to spectral closeness of the ω3 and ω4 modes. The difference between the numbers of multipoles presented in the insets of Figure 4a and the found multipolarity for each mode is explained by a large spectral width of plasmon modes. Excitation of plasmon modes leads to an increase in the local field at the wavelengths λ1, λ2, λ3, and λ4; therefore, emission of the second and third harmonics is realized at doubled and tripled frequencies of these resonances. Figure 4b presents the emission spectrum of the second harmonic and calculated spectra of the second harmonic that correspond to the excitation of the ω1, ω2, ω3, and ω4 plasmon modes in the SHR nanostructure. The calculated spectra amplitudes were chosen to be equal to the amplitudes of the measured resonances in the second harmonic spectrum. As follows from the figure, for each plasmon mode at the fundamental frequency there is a resonance in the second harmonics generation spectrum: (1) the 2ω1 resonance is realized at 440 nm, (2) the 2ω2 resonance occurs at 402 nm, (3) the 2ω3 resonance is at 385 nm, and (4) the 2ω4 resonance takes place at 380 nm. The small amplitude of the 2ω3 resonance peak can be explained by known high sensitivity of the second harmonic generation to the symmetry of the field distribution.38 Figure 4c shows the emission spectrum at the frequency of the third harmonic and calculated spectra of the third harmonic generation that correspond to the excitation of the ω1, ω2, ω3, and ω4 plasmon modes in the SHR nanostructure. The calculated spectra amplitudes were chosen to be equal to the amplitudes of the measured resonances in the spectrum of the third harmonic generation taking into account the spectral transmission of the bandpass filter that cuts off the radiation at the fundamental frequency. As follows from the figure, for each plasmon mode at the fundamental frequency there is a resonance in the third harmonic generation spectrum: (1) the 3ω1 resonance is realized at 293 nm, (2) the 3ω2 resonance occurs at 268 nm, (3) the 3ω3 resonance is at 257 nm, and (4) the 3ω4 resonance takes place at 250 nm. As follows from Figure 4c, the peak in the spectrum of the third harmonic generation with the central wavelength of 278 nm does not have a corresponding resonance at the fundamental frequency. We suggest that this resonant peak is caused by mixing of the ω2 and ω1 frequencies (see Supporting Information for details). It should be stressed that the realization of ω2 + 2ω1 process is the consequence of coherent 1141

DOI: 10.1021/acs.nanolett.5b04373 Nano Lett. 2016, 16, 1138−1142

Letter

Nano Letters

(16) Aouani, H.; Navarro-Cía, M.; Rahmani, M.; Maier, S. A. Adv. Opt. Mater. 2015, 3, 1059−1065. (17) Ghenuche, P.; Cherukulappurath, S.; Taminiau, T. H.; van Hulst, N. F.; Quidant, R. Phys. Rev. Lett. 2008, 101, 116805. (18) Bouhelier, A.; Beversluis, M. R.; Novotny, L. Appl. Phys. Lett. 2003, 83, 5041−5043. (19) Schuck, P.; Fromm, D. P.; Sundaramurthy, A.; Kino, G. S.; Moerner, W. E. Phys. Rev. Lett. 2005, 94, 017402. (20) Melentiev, P.; Afanasiev, A.; Kuzin, A.; Zablotskiy, A.; Balykin, V. Opt. Express 2015, 23, 11444−11452. (21) Palomba, S.; Novotny, L. Nano Lett. 2009, 9, 3801−3804. (22) Danckwerts, M.; Novotny, L. Phys. Rev. Lett. 2007, 98, 026104. (23) Sasanpour, P.; Rashidian, B.; Rashidian, B.; Vossoughi, M. Nano 2010, 5, 325−332. (24) Drew, D. IMA Preprint Series 1996, 1422−7, 1−24. (25) Sinha, R. P. Häder. Photochem. Photobiol. Sci. 2002, 1, 225−236. (26) Xu, T.; Jiao, X.; Blair, S. Opt. Express 2009, 17, 23582−23588. (27) N’Gom, M.; Ye, J. Y.; Norris, T. B.; Agarwal, A.; Kotov, N. In Optics InfoBase Conference Papers. Optical Society of America, International Quantum Electronics Conference, IQEC, 2009; Baltimore, MD, June 5−31, 2009. (28) Schwartz, O.; Oron, D. Nano Lett. 2009, 9, 4093−4097. (29) Metzger, B.; Hentschel, M.; Schumacher, T.; Lippitz, M.; Ye, X.; Murray, C. B.; Knabe, B.; Buse, K.; Giessen, H. Nano Lett. 2014, 14, 2867−2872. (30) Luk’yanchuk, B. S.; Miroshnichenko, A. E.; Tribelsky, M. I.; Kivshar, Y. S.; Khokhlov, A. R. New J. Phys. 2012, 14, 093022−14. (31) Link, S.; Wang, Z.; El-Sayed, M. J. Phys. Chem. B 2000, 104, 7867−7870. (32) Khurgin, J. B.; Sun, G. Phys. Rev. A: At., Mol., Opt. Phys. 2013, 88, 053838. (33) Melentiev, P. N.; Afanasiev, A. E.; Kuzin, A. A.; Baturin, A. S.; Balykin. Opt. Express 2013, 21, 13896−13905. (34) Melentiev, P. N.; Konstantinova, T.; Afanasiev, A.; Kuzin, A.; Baturin, A.; Tausenev, A.; Konyaschenko, A.; Balykin, V. Laser Phys. Lett. 2013, 10, 075901. (35) Melentiev, P. N.; Afanasiev, A. E.; Kuzin, A. A.; Baturin, A. S.; Balykin. Opt. Lett. 2013, 38, 2274−2276. (36) Khlebtsov, B. N.; Khlebtsov, N. G. J. Phys. Chem. C 2007, 111, 11516−11527. (37) Rockstuhl, C.; Lederer, F.; Etrich, C.; Zentgraf, T.; Kuhl, J.; Giessen, H. Opt. Express 2006, 14, 8827−8836. (38) Boyd, R. Nonlinear Optics; Academic Press: London, 2003. (39) Shcherbakov, M. R.; Neshev, D. N.; Hopkins, B.; Shorokhov, A. S.; Staude, I.; Melik-Gaykazyan, E. V.; Decker, M.; Ezhov, A. A.; Miroshnichenko, A. E.; Brener, I.; Fedyanin, A. A.; Kivshar, Y. S. Nano Lett. 2014, 14, 6488−6492. (40) Schröder, K. Handbook on Industrial Laser Safety; Technical University of Vienna: Vienna, 2000. (41) Esenaliev, R. In Laser-Tissue Interaction; SPIE Publishing: Bellingham, WA, 1999; Vol. 3601, pp 166−176. (42) Esenaliev, R. U.S. Patent 6,165,440, Dec. 26, 2000. (43) Esenaliev, R. U.S. Patent 2011/0034855 A1, Feb. 10, 2011.

therapy. An advantage of this approach is potentially lower concentration of nanoparticle and lower average intensity of laser light necessary for therapy compared to that in photothermal therapy because to provide efficient heating of tumors sufficient nanoparticle concentration and average light intensity are needed. We have shown that the SHR nanostructure under state-of-the-art experimental conditions realizes the record flux of photons at the THG frequency. We expect that our findings will be helpful for designing plasmonic nanostructures that can be used in phototherapy in future.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.5b04373. The detailed information on the used methods and technical results. (PDF)



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (P.N.M.). *E-mail: [email protected] (R.O.E.). *E-mail: [email protected] (V.I.B.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We would like to thank D. Neshev for helpful discussions. This work was partially supported by the Russian Scientific Foundation (Contract No. 14-12-00729) and by the Grant of the Government of the Russian Federation for the Support of Scientific Investigations under the Supervision of Leading Scientists (Contract No. 14.B25.31.0019).



REFERENCES

(1) Novotny, L.; van Hulst, N. Nat. Photonics 2011, 5, 83−90. (2) Giannini, V.; Fernàndez Domìnguez, A. I.; Heck, S. C.; Maier, S. A. Chem. Rev. 2011, 111, 3888−3912. (3) Tang, L.; Kocabas, S. E.; Latif, S.; Okyay, A. K.; Ly-Gagnon, D.S.; Saraswat, K. C.; Miller, D. A. B. Nat. Photonics 2008, 2, 226−229. (4) Atwater, H. A.; Polman, A. Nat. Mater. 2010, 9, 205−213. (5) Gerton, J. M.; Wade, L. A.; Lessard, G. A.; Ma, Z.; Quake, S. R. Phys. Rev. Lett. 2004, 93, 180801. (6) Frey, H. G.; Witt, S.; Felderer, K.; Guckenberger, R. Phys. Rev. Lett. 2004, 93, 200801. (7) Anker, J. N.; Hall, W. P.; Lyandres, O.; Shah, N. C.; Zhao, J.; Van Duyne, R. P. Nat. Mater. 2008, 7, 442−453. (8) Fedoruk, M.; Meixner, M.; Carretero-Palacios, S.; Lohmüller, T.; Feldmann, J. ACS Nano 2013, 7, 7648−7653. (9) Capretti, A.; Pecora, E. F.; Forestiere, C.; Negro, L. D.; Miano. Phys. Rev. B: Condens. Matter Mater. Phys. 2014, 89, 125414. (10) Czaplicki, R.; Mäkitalo, J.; Siikanen, R.; Husu, H.; Lehtolahti, J.; Kuittinen, M.; Kauranen, M. Nano Lett. 2015, 15, 530−534. (11) Bouhelier, A.; Beversluis, M.; Hartschuh, A.; Novotny, L. Phys. Rev. Lett. 2003, 90, 013903. (12) Lippitz, M.; van Dijk, M. A.; Orrit, M. Nano Lett. 2005, 5, 799− 802. (13) Kim, S.; Jin, J.; Kim, Y. J.; Park, I.-Y.; Kim, Y.; Kim, S. W. Nature 2008, 453, 757−760. (14) Bar-Elli, O.; Grinvald, E.; Meir, N.; Neeman, L.; Oron, D. ACS Nano 2015, 9, 8064−8069. (15) Aouani, H.; Rahmani, M.; Navarro-Cía, M.; Maier, S. Nat. Nanotechnol. 2014, 9, 290−294. 1142

DOI: 10.1021/acs.nanolett.5b04373 Nano Lett. 2016, 16, 1138−1142