Published: March 18, 2011

r 2011 American Chemical Society 1556 dx.doi.org/10.1021/nl104352j |Nano Lett. 2011, 11, 1556–1560

LETTER

pubs.acs.org/NanoLett

Plexciton Dynamics: Exciton�Plasmon Coupling in a J-Aggregate�AuNanoshell Complex Provides a Mechanism for NonlinearityNche T. Fofang,†,§ Nathaniel K. Grady,‡,§ Zhiyuan Fan,|| Alexander O. Govorov,|| and Naomi J. Halas*,†,§

†Department of Electrical and Computer Engineering, ‡Department of Physics and Astronomy, and §Laboratory for Nanophotonics,Rice University, 6100 Main Street, Houston, Texas 77005, United States

)Department of Physics and Astronomy, Ohio University, Athens, Ohio 45701, United States

The optical nonlinearity of materials is a fundamental prop-erty required for the realization of active optical frequency

devices such as ultrafast switches,1�4 bistable devices,5�8

limiters,9 and modulators.10�12 Most materials found in nature,however, have a very small nonlinear optical response. Fortu-nately, this response can be enhanced by exploiting the electronicor vibrational transitions in materials for resonant enhancementat the corresponding near-resonant energies.13 This limitationcan also be addressed by designing new, nanoengineered materi-als whose resonances are tuned to a frequency region of interest,such as specially designed plasmonic structures. Interactions withdirectly adjacent media can be enhanced by coupling within theintense evanescent field of the plasmon, resulting in new, nano-composite hybrid materials whose properties are distinct fromthe properties of the individual constituent media.

Hybrid molecular�plasmonic nanostructures, where excita-tions of the respective components of the hybrid are designed tointeract, provide a promising strategy for the development ofactive, nonlinear materials at optical frequencies.14,15 Ultrafasttime-domain optical studies can reveal how these interactionsresult in new properties, and how processes like exciton�plas-mon hybridization, energy or charge transfer, may contribute tothe new hybrid material response. Interest in the combination ofmolecular media and plasmonic nanostructures results from theease in design and fabrication of hybrid media with these consti-tuents, and also from the many opportunities to exploit thenumerous interaction mechanisms associated with excitonic andplasmonic systems.

Metallic nanoparticles sustain collective electronic excitations,known as localized surface plasmons, when resonantly excited.

The local electromagnetic field associated with these plasmonscan be quite intense and well localized near the metallic nano-particle surface; so much so that it may strongly modify thenonlinear properties of any material directly within this fringingfield.13 For example, the surface plasmon may affect absorptionsaturation, nonlinear absorption line shapes,14 nonlinear refrac-tive index,16 or nonradiative decay channels17 in a directly adja-cent medium, modifying the optical nonlinear response throughany of these mechanisms.

In this Letter, we report the examination of the ultrafast opticaldynamics of a hybrid resonant plasmon�exciton system con-sisting of a J-aggregate�Au nanoshell complex. Nanoshells aretunable plasmonic core�shell nanoparticles that support a distincthierarchy of localized plasmon modes with resonant energiesdependent on the relative dimensions and composition of theircore and shell layers, the absolute size of the nanoparticle, and thepresence of other media adjacent to, or surrounding, the nano-structure.18�22 J-aggregates are organic semiconductors withlocalized excitations—excitons—that possess very high oscilla-tor strengths, suitable for interaction with the intense near field oflocalized surface plasmons.23 Strong interaction of J-aggregateexcitons with plasmons24,25 and cavity photons,26�28 resulting inthe formation of mixed states, has been reported. In these systems,the ultrafast dynamical properties of J-aggregates are stronglymodified. Exciton�photon interactions have been extensivelystudied for J-aggregates in microcavities29,30 and more recently

Received: December 14, 2010Revised: February 19, 2011

ABSTRACT: Coherently coupled plasmons and excitons give riseto newoptical excitations- plexcitons� due to the strong coupling ofthese twooscillator systems.Time-resolved studies of J-aggregate-Aunanoshell complexes when the nanoshell plasmon and J-aggregateexciton energies are degenerate probe the dynamical behavior of thiscoupled system. Transient absorption of the interacting plasmon-exciton system is observed, in dramatic contrast to the photoinducedtransmission of the pristine J-aggregate. An additional, transientFano-shapedmodulationwithin the Fano dip is also observable. Thebehavior of the J-aggregate-Au nanoshell complex is described by acombined one-exciton and two-exciton state model coupled to thenanoshell plasmon.

KEYWORDS: Plexcitons, plasmons, excitons, nanoshellsJ-aggregates, Fano resonance

1557 dx.doi.org/10.1021/nl104352j |Nano Lett. 2011, 11, 1556–1560

Nano Letters LETTER

for J-aggregates interacting with propagating surface plasmonson extended metallic structures.31 Until now, however, thedynamical response of J-aggregate excitons due to coupling withthe localized plasmons of plasmonic nanoparticles has not beendirectly examined. The ultrafast properties of metallic nanopar-ticle-J-aggregate complexes have been reported for the casewhere the transition energy of the J-aggregate is strongly detunedrelative to the resonant plasmon energy;32 however, this is amuch weaker interaction regime which is substantially differentfrom the stronger interactions in the doubly resonant regimeaddressed in the present study. In the weaker, nonresonant regime,the plasmon�molecule interaction typically leads to enhance-ment of molecular absorption and shifts in plasmon resonancefrequencies. These characteristics are replaced by entirely neweffects in the stronger interaction regime, where the coupledplasmons and excitons hybridize and form plexcitons, and wherethe behavior of this new highly coupled system is distinct fromeither of its component subsystems.23 As has been recently shownin a number of plasmonic systems, strong coupling betweenmetallic nanoparticles can lead to Fano interference and reso-nances.33�40 Such interference can also occur in metallic nano-particle/semiconductor quantum dot hybrid systems where thecoupling between excitons and plasmons can also result in anenhancement of the Fano effect.14 Further increased coupling

may lead to a transition from asymmetric Fano behavior to that ofa doubly peaked Fano resonance and finally to a bistable regime.5

In the present study, we examine these interaction regimes ex-perimentally for a hybrid J-aggregate�plasmonic nanostructureand construct a straightforward nonlinear dielectric model thatdescribes the observed interaction between the two subsystems.

Our coupled plasmon�exciton nanocomplex consists of ananoshell surrounded by an epilayer of J-aggregates (Figure 1a).The J-aggregates are formed on the surface of the nanoshellswhen a water/ethanol (50:50) solution of the dye molecules(2,20-dimethyl-8-phenyl-5,6,50,60-dibenzothiacarbocyaninechloride) is added to an aqueous solution of nanoshells. With thepeak of the nanoshell spectrum tuned to the maximum absorp-tion of the J-aggregrate (Figure 1b), the complexes exhibit coherentcoupling between the localized plasmons of the nanoshell andexcitons of the molecular J-aggregates: a well-defined Fano reso-nance is observed in the ensemble optical spectrum (Figure 1c).23

The Fano resonance created well-expressed dips in the absorptionspectra (∼10%) and the corresponding spectral peak-to-peakdistances were measured as∼100meV.23 The Fano dip structure(antiresonance) can be understood as a coherent destructivecoupling between two oscillations, an exciton and a plasmon.This coupling can be adequately described in terms of a classicaldielectric model.23

The ultrafast studies were performed using two-color, time-resolved pump�probe spectroscopy. Amode locked Ti:sapphirelaser system consisting of an oscillator (Coherent MIRA 900), aregenerative amplifier, and an OPA (Coherent) was used as theexcitation and probe source. Pulses from the OPA were approxi-mately 150 fs. A motorized stage (Newport) was used to delaythe probe beam with respect to the pump beam. The sampleconsisted of an aqueous solution of J-aggregate�Au nanoshellcomplexes in a 1 mm path length cuvette. Signals were collectedin a transmission geometry.

In Figure 2, the time-resolved transmission obtained fromJ-aggregate�Au nanoshell complexes (Figure 2a�d) is com-pared with that of J-aggregates in the absence of Au nanoshells(Figure 2e�h).The pump andprobefluences are 60 and 0.8μJ/cm2,respectively. This pump fluence is low enough such that no time-resolved signals were obtainable fromnanoshells without J-aggregatesunder identical pump and probe conditions; in other words, only thetransient dynamics of the J-aggregate and the J-aggregate complex arebeing probed. The fwhm of the laser spectrum was limited to 10 nmusing a 10 nmband-pass filter. The time-resolved relaxation dynamicsshown in Figure 2 were obtained with pump and probe pulses at thesame wavelength, for several discrete wavelengths within the spectralregion of interaction, i.e., the Fano dip. These wavelengths (680, 690,700, and 710 nm) are indicated in the nanoshell�J-aggregate extinc-tion spectrum shown in Figure 1c.

For the J-aggregate�Au nanoshell complex, a clear transitionfrom transient bleaching (pump and probe at 690 nm) totransient absorption (pump and probe at 710 nm) is observed,while no such transition can be observed for the pristine J-aggregate(Figure 2f�h). The signal at 700 nm shows a complex, mixedtransient bleaching, and absorption behavior with the bleachingoccurring at earlier times. At 680 nm, while the J-aggregate displaystransient absorption (Figure 2e),41 the J-aggregate�Au nanoshellcomplex shows transient bleaching (Figure 2a). The stark differ-ence between the behavior of the plexcitonic complex and thepristine excitonic medium indicates a strong modification of theexcitonic states of the J-aggregate due to couplingwith the localizeddipolar plasmons of the Au nanoshells.23

Figure 1. (a) Schematic of the hybrid nanocomplex consisting of a goldnanoshell coated with J-aggregate molecular film (2,20-dimethyl-8-phenyl-5,6,5060-dibenzothiacarbocyanine chloride). (b) Extinction spec-trum of the J-aggregate, obtained from a formulation consisting of amixture of 6 μL of 117 μM dye solution and 3.5 mL of aqueous poly-(vinyl alcohol).(c) Extinction spectra of pristine (black) and J-aggregatecoated (red) [r1, r2] = [41, 54] nm nanoshells. The vertical arrowsindicate the wavelengths (680, 690, 700, 710 nm)where the J-aggregatesand J-aggregate-coated nanoshells were optically interrogated.

1558 dx.doi.org/10.1021/nl104352j |Nano Lett. 2011, 11, 1556–1560

Nano Letters LETTER

We can explain the qualitative behavior of the transient signalin the vicinity of the Fano dip by considering both the nonlinearphotobleaching of the J-aggregate and the Fano effect. Whenthe plexcitonic system is excited by the pump pulse, the opticaltransition becomes partially saturated. From this we can inferthat the Fano dip itself should become transiently modulated(a transient increase in the absorption intensity within theminimum of the Fano line shape). This is due to the decreasedexciton�plasmon interaction directly following excitation by thepump pulse. This transient increase within the Fano dip meansan increase in transient absorption (and a corresponding decrease intransmission). This behavior should be limited to excitationswithin the Fano resonance since strong exciton�plasmon cou-pling is not essential for the nonresonant optical response of theJ-aggregates. Outside the Fano resonance one would expect theusual transient bleaching of the excitonic system to be observed.The above effects (a negativeΔT at the Fano dip and a positiveΔTon the left-hand side of the Fanominimum) should create a shift ofthe Fano dip in the nonequilibrium absorption spectrum in thepresence of optical pumping. Indeed, this transient shift can clearly beseen by comparing the experimental change in trans-mission ΔT/T as a function of wavelength for the plexcitoniccomplex and the J-aggregate shown in Figure 2i. This figure alsoshows a decrease in magnitude of ΔT/T for the hybrid relative tothe pristine J-aggregate (Figure 2i). A plasmon resonance can, inprinciple, enhance the optical nonlinear response of a directly adjacentnonlinear material (in this case the J-aggregates) due to its high-intensity local electromagnetic field. However, in this case thepresence of the metallic component simultaneously leads to a fasterrelaxation of the nonequilibrium exciton population, which is the

likely origin of the decrease in amplitude of ΔT/T observed(Figure 2i).

A straightforward theoretical model can be introduced thatdescribes the observed plasmon�exciton coupling in the J-ag-gregate�Au nanoshell complex. Figure 3a shows the energystructure and transitions in our model. The Fano interferenceappears in this model from the possibility for the photons tocreate a plasmon via two paths. In the resonance with the oneexciton, the paths are: |0æf |LSPæ and |0æf |1æf |LSPæ. In theresonance with the two-exciton: |1æ f |LSPæ and |1æ f |2æ f|LSPæ. Photoinduced transmission in molecules alone is theuniversal behavior originating from saturation of a two-level(excitonic) system, consistent with the behavior of the pristineJ-aggregate. Introducing the plasmon�exciton interactionrequires that we extend this initial theoretical picture of theJ-aggregate beyond a simple two-level system. We begin with adescription of the exciton dynamics using rate equations thatdescribe the populations of ground, one-exciton, and two-exciton states, n0(t), nS1(t), and nS2(t), that obey the condition:

Figure 3. (a) Energy structure and transitions in a theoretical modeldescribing the nonlinear Fano effect in our system. Vertical arrows(black) show optical transitions and the horizontal arrows (blue) depictexciton�plasmon coupling. (b) Calculated absorption of the J-aggrega-te�Au nanoshell complex at two times: t = 0 (the instant just after thepump pulse) and t =¥ (long-term equilibrium response). The nanoshelldimensions are [r1, r2] = [48, 58] nm with a 2 nm thick J-aggregate shelllayer. Parameters for the dielectric constants of the materials are takenfrom ref 28. (c) Normalized transient transmission at t = 0 for the hybridcomplex (black) and for the pristine J-aggregates (red). (d) Experi-mental pump-continuum probe absorbance spectra of the J-aggrega-te�Au nanoshell complex pumped at 690 nm at various delay times: t =0 overlap between pump pulse and continuum probe beam [black]; 80ps delay [blue]. Continuum spectrum with no pump pulse present(beam blocked) [red]. Spectra are displaced vertically for clarity. (e)Transient transmission obtained from excited state absorbance spectraat t = 0 shown in (d) by continuum background subtraction.

Figure 2. (a�d) Time-resolved differential transmission of J-aggrega-te�Au nanoshell complexes (corresponding extinction spectrum shownin Figure 1c). (e�h) Time-resolved differential transmission of pristineJ-aggregates (corresponding extinction spectrum shown in Figure 1b):(a) and (e), pump and probe at 680 nm; (b) and (f), pump and probe at690 nm; (c) and (g), pump and probe at 700 nm; (d) and (h), pump andprobe at 710 nm. Pulse width of both pump and probe is 150 fs andpump fluence is approximately 60 μJ/cm2. (i) Amplitude plots for J -aggregates (red triangles) and J-aggregate�nanoshell complexes (blacksquares).

1559 dx.doi.org/10.1021/nl104352j |Nano Lett. 2011, 11, 1556–1560

Nano Letters LETTER

n0(t) þ nS1(t) þ nS2(t) = 1. At �δtpulse < t < 0, a short pumppulse of duration δtpulse excites both one-exciton and two-exciton states which rapidly relax. In our J-aggregates, theseone-exciton and two-exciton states correspond to the statesobserved at the energies corresponding to ∼700 and 680 nm,respectively. The two-exciton state can only be excited via theone-exciton state: for the pump intensities considered, nS2(t) <nS1(t). To calculate the temporal evolution of the populations,we use a system of coupled rate equations and assume that thefrequency of the pump beam is close to the Fano-dip spectralposition (λ = 700 nm). We then calculate the transientabsorption of this system as a function of probe delay

QprobeðtÞ ¼ 12ReZVjB �ð rB, tÞ 3 EBprobeð rBÞ dV ð1Þ

where jB is the current inside the J-aggregate�Au nanoshellcomplex induced by a short, low-power probe pulse and EBprobe

is the nonuniform electric field inside the nanocomplex. Thecurrent is calculated as jB(rB,t) = �iωprobe(ε(rB,t) � 1)/4π 3EBprobe, where ε(rB,t) is the effective dielectric constant of thesystem.23 Inside the Au nanosphere, ε(rB,t) = εAu, and inside theJ-aggregate shell layer of the complex, ε(rB,t) = εJ(t). The Audielectric constant (εAu) remains a constant in this picturesince, in this experimental regime, the nonlinearity of the Ausubsystem is negligible (not observable at the fluences used).The nonlinear response enters this simple model via the time-dependent dielectric constant of the J-aggregate εJ(t) whichcan be described as

εJðω, tÞ ¼ 1� fS1ωS1

2

ωS12 �ω2 þ iγω

ðnS0ðtÞ � nS1ðtÞÞ

� fS2ωS2

2

ωS22 �ω2 þ iγω

ðnS1ðtÞ � nS2ðtÞÞ ð2Þ

where fS1 = 0.03 and fS2 = 0.05 are the reduced oscillatorstrengths, and γ = 0.052 eV. These material parameters wereobtained from the available literature on the J-aggregate system.32,41

The coefficients involving the nonequilibrium state populations ineq 2 describe the usual effect of bleaching of a molecular system.

Calculation of the electric current and the actual electric fieldinside the complex is relatively simple for a small nanocomplex inthe quasi-static regime that applies here, i.e., when c, λ, wherec and λ are the radius of the complex (c = 60 nm) and thewavelength of probe light (λ = 680�710 nm), respectively. Inthis case, one can use the Poisson equation and obtain a resultanalytically.32 Figure 3 shows the calculated absorption Qprobe justafter the probe pulse (t = 0) and at long delay times(t = ¥) when the system comes to equilibrium. The transienttransmission signal at t = 0 isΔTmax�Qprobe(t=¥)�Qprobe(t=0)and is shown in Figure 3c. In Figure 3c, we see the forma-tion of transient absorption (negative ΔTmax) in the spectralregion of the Fano dip, as already discussed. Another predictedresult from this model is the creation of a second, transient Fanodip in the vicinity of the linear Fano dip (Figure 3b) due toinvolvement of the two-exciton states (the transition |1æf |2æ inthe energy diagram of Figure 3a). Compared to the nonlinearFano model used in previous publications,15,42 a novel feature ofthe present model is the inclusion of the two-exciton states thatadds a second nonlinear Fano minimum. This theoretical modelis meant to provide a qualitative picture of the coupled system:

detailed properties of the excitonic states, their various relaxationchannels, and other effects such as spectral hole burning are notdescribed quantitatively by this simple picture.

Probing this coupled system experimentally with a short pumppulse and a continuum probe pulse whose spectral content spansthe plexciton Fano resonance allows us to experimentally verifythe theoretical predictions of our simple model (Figure 3d). Forthe case of overlapping pump and probe pulses (t = 0), we areable to observe a significant modulation within the Fano dip. Thisadditional spectral feature follows themodulation predicted fromour theoretical model remarkably well (Figure 3c). Here thepump pulse partially saturates the one-exciton transition in thecomplex. When this occurs we would anticipate that the depth ofthe Fano minimum would decrease with a transient absorptionappearing within the spectral region of the Fano dip. Simulta-neously, we would anticipate transient bleaching signals on thesides of this transient absorption region. These modulations areclearly observed (Figure 3e). Interestingly, the above resultsmean that, with the pump pulse, we are able to selectively modu-late the strength of exciton�plasmon coupling in the system.The pump pulse diminishes the one-exciton-plasmon coupling(negativeΔT for the dip region in Figures 2i and 3e) and, simul-taneously, activates the two-exciton-plasmon interaction (positiveΔT to the blue from the Fano dip in Figure 3e).

In conclusion, we have studied the ultrafast optical dynamicsof excitons in nanoshell�J-aggregate complexes where localizedplasmons and excitons are interacting strongly. The ultrafastcharacteristics indicate that these hybrid nanostructures mightpossess enhanced, tunable, on- and off-resonance nonlinearoptical properties. We believe that these results may stimulatefurther interest in the properties of nanoengineered plasmon�exciton systems that may ultimately serve as active media innew classes of nonlinear optical devices or in active, functionalmetamaterials.

’AUTHOR INFORMATION

Corresponding Author*E-mail: halas@rice.edu.

’ACKNOWLEDGMENT

We gratefully acknowledge theDepartment of DefenseNationalSecurity Science and Engineering Faculty Fellowship (N.J.H.) viaGrant N00244-09-0067, the Air Force Office of Scientific Research(F49620-03-C-0068), Robert A. Welch Foundation (C1220), andthe Center for Advanced Solar Photophysics, an Energy FrontierResearch Center funded by the U.S. Department of Energy, Officeof Science, Office of Basic Energy Sciences, for financial support. A.O.G. and Z.F. thank the NSF for support (project number CBET-0933415).

’REFERENCES

(1) Haraguchi, M.; Fukui, M.; Tamaki, Y.; Okamoto, T. J. Microsc.2003, 210, 229–233.

(2) Min, C.; Wang, P.; Chen, C.; Deng, Y.; Lu, Y.; Ming, H.; Ning,T.; Zhou, Y.; Yang, G. Opt. Lett. 2008, 33 (8), 869–871.

(3) Sato, Y.; Furuki, M.; Minquan, T.; Iwasa, I.; Pu, L. S.; Tatsuura, S.Appl. Phys. Lett. 2002, 80 (13), 2254–2256.

(4) Dintinger, J.; Klein, S.; Ebbesen, T. W. Adv. Mater. 2006, 181267–1270.

(5) Artuso, R. D.; Bryant, G. W. Nano Lett. 2008, 8 (7), 2106–2111.

1560 dx.doi.org/10.1021/nl104352j |Nano Lett. 2011, 11, 1556–1560

Nano Letters LETTER

(6) Porto, J. A.; Martin-Moreno, L.; Garcia-Vidal, F. J. Phys. Rev. B2004, 70, No. 081402.(7) Min, C.; Wang, P.; Chen, C.; Jiao, X.; Deng, Y.; Ming, H. Opt.

Express 2007, 15 (19), 12368.(8) Wurtz, G. A.; Pollard, R.; Zayats, A. V. Phys. Rev. Lett. 2006,

97, No. 057402.(9) Ispasoiu, R. G.; Balogh, L.; Varnavski, O. P.; Tomalia, D. A.; T.,

G., III J. Am. Chem. Soc. 2000, 122, 11005–11006.(10) Dintinger, J.; Robel, I.; Kamat, P. V.; Genet, C.; Ebbesen, T. W.

Adv. Mater. 2006, 18, 1645–1648.(11) Dicken, M. J.; Sweatlock, L. A.; Pacific, D.; Lezec, H. J.;

Bhattacharya, K.; Atwater, H. A. Nano Lett. 2008, 8 (11), 4048–4052.(12) Dionne, J. A.; Diest, K.; Sweatlock, L. A.; Atwater, H. A. Nano

Lett. 2009, 9 (2), 897–902.(13) Wang, Y.; Xie, X.; T., G., III Nano Lett. 2005, 5 (12), 2379–2384.(14) Lu, Z.; Zhu, K. J. Phys. B: At. Mol. Opt. Phys. 2008, 41, No.

185503.(15) Zhang, W.; Govorov, A. O.; Bryant, G. W. Phys. Rev. Lett. 2006,

97, No. 146804.(16) Kohlgraf-Owens, D. C.; Kik, P. G. Opt. Express 2008, 16 (14),

10823–10834.(17) Larkin, I. A.; Stockman, M. I.; Achermann, M.; Klimov, V. I.

Phys. Rev. B 2004, 69, No. 121403.(18) Oldenburg, S. J.; Averitt, R. D.; Westcott, S. L.; Halas, N. J.

Chem. Phys. Lett. 1998, 288, 243–247.(19) Tam, Felicia; Chen, Allen L.; Kundu, Janardan; Wang, Hui;

Halas, N. J. J. Chem. Phys. 2007, 127, 6.(20) Tam, F.;Moran, C.; Halas, N. J. J. Phys. Chem. B 2004, 108 (45),

17290–17294.(21) Bardhan, R.; Grady, N. K.; Ali, T.; Halas, N. J. ACS Nano 2010,

4 (10), 6169–6179.(22) Slocik, J.M.; Tam, F.; Halas, N. J.; Naik, R. R.Nano Lett. 2007, 7

(4), 1054–1058.(23) Fofang, N. T.; Park, T.; Neumann, O.; Mirin, N. A.; Nordlander,

P.; Halas, N. J. Nano Lett. 2008, 8 (10), 3481–3487.(24) Bellessa, J.; Bonnand, C.; Plenet, J. C.; Mugnier, J. Phys. Rev.

Lett. 2004, 93, No. 036404.(25) Dintinger, J.; Klein, S.; Bustos, F.; Barnes, W. L.; Ebbessen,

T. W. Phys. Rev. B 2005, 71, No. 035424.(26) Lidzey, D. G.; Bradley, D. D. C.; Skolnick, M. S.; Virgili, T.;

Walker, S.; Whittaker, D. M. Nature 1998, 395, 53–55.(27) Lidzey, D.; Bradley, D.; Skolnick, M.; Virgili, E.;Walker, S. Phys.

Rev. Lett. 1999, 82, 3316.(28) Lidzey, D. G.; Bradley, D. D. C.; Armitage, A.; Walker, S.;

Skolnick, M. S. Science 2000, 288, 1620.(29) Sasaki, F.; Haraichi, S.; Kobayashi, S. IEEE J. Quantum Electron.

2002, 38 (7), 943.(30) Song, J.; He, Y.; Nurmikko, A. V.; Tischler, J.; Bulovic, V. Phys.

Rev. B 2004, 69, No. 235330.(31) Vasa, P.; Pomraenke, R.; Cirmi, G.; Re, E. D.; Wang, W.;

Schwieger, S.; Leipold, D.; Runge, E.; Cerullo, G.; Lienau, C. ACS Nano2010, 4, 7559–7565.(32) Wiederrecht, G. P.; Wurtz, G. A.; Hranisavljevic, J. Nano Lett.

2004, 4, 2121–2125.(33) Luk’yanchuk, B.; Zheludev, N. I.; Maier, S. A.; Halas, N. J.;

Nordlander, P.; Giessen, H.; Chong, C. T. Nat. Mater. 2010, 9 (9),707–715.(34) Verellen, N.; Sonnefraud, Y.; Sobhani, H.; Hao, F.; Moshchalk-

ov, V. V.; Van Dorpe, P.; Nordlander, P.; Maier, S. A.Nano Lett. 2009, 9(4), 1663–1667.(35) Hao, F.; Nordlander, P.; Sonnefraud, Y.; Van Dorpe, P.; Maier,

S. A. ACS Nano 2009, 3 (3), 643–652.(36) 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. Science2010, 328 (5982), 1135–1138.(37) Fan, J. A.; Bao, K.; Wu, C. H.; Bao, J. M.; Bardhan, R.; Halas,

N. J.; Manoharan, V. N.; Shvets, G.; Nordlander, P.; Capasso, F. NanoLett. 2010, 10 (11), 4680–4685.

(38) Sonnefraud, Y.; Verellen, N.; Sobhani, H.; Vandenbosch,G. A. E.; Moshchalkov, V. V.; Van Dorpe, P.; Nordlander, P.; Maier,S. A. ACS Nano 2010, 4 (3), 1664–1670.

(39) Mukherjee, S.; Sobhani,H.; Lassiter, J. B.; Bardhan, R.;Nordlander,P.; Halas, N. J. Nano Lett. 2010, 10 (7), 2694–2701.

(40) Brown, L. V.; Sobhani, H.; Lassiter, J. B.; Nordlander, P.; Halas,N. J. ACS Nano 2010, 4 (2), 819–832.

(41) Fidder, H.; Knoester, J.; Wiersma, D. A. J. Chem. Phys. 1993, 98(8), 6564–6566.

(42) Kroner, M.; Govorov, A. O.; Remi, S.; Biedermann, B.; Seidl, S.;Badolato, A.; Petroff, P. M.; Zhang, W.; Barbour, R.; Gerardot, B. D.;Warburton, R. J.; Karrai, K. Nature 2008, 451 (7176), 311–314.