Published: January 31, 2011

r 2011 American Chemical Society 1049 dx.doi.org/10.1021/nl103906f |Nano Lett. 2011, 11, 1049–1054

LETTER

pubs.acs.org/NanoLett

Manipulating Coupling between a Single Semiconductor QuantumDot and Single Gold NanoparticleDaniel Ratchford,† Farbod Shafiei,† Suenne Kim,† Stephen K. Gray,‡ and Xiaoqin Li*,†

†Department of Physics, University of Texas at Austin, Austin, Texas 78712, United States‡Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, United States

bS Supporting Information

ABSTRACT: Using atomic force microscopy nanomanipulation, weposition a single Au nanoparticle near a CdSe/ZnS quantum dot toconstruct a hybrid nanostructure with variable geometry. The couplingbetween the two particles is varied in a systematic and reversible manner.The photoluminescence lifetime and blinking of the same quantum dotare measured before and after assembly of the structure. In somehybrid structures, the total lifetime is reduced from about 30 ns to wellbelow 1 ns. This dramatic change in lifetime is accompanied by thedisappearance of blinking as the nonradiative energy transfer from the CdSe/ZnS quantum dot to the Au nanoparticle becomes thedominant decay channel. Both total lifetime and photoluminescence intensity changes are well described by simple analyticalcalculations.

KEYWORDS: Nanomanipulation, hybrid nanostructures, plasmonics, semiconductor nanocrystals, blinking, photoluminescence

Top-down and bottom-up fabrication techniques produceincreasingly sophisticated hybrid nanostructures with build-

ing blocks such as semiconductor andmetallic nanoparticles (NPs),DNA segments, and polymers.1-4 Nanoscale integration ofmetals and semiconductors is particularly interesting becausesuch hybrid structures combine the strengths of both materials.The metallic surface plasmon resonance enables concentrationand manipulation of light on a length scale well beyond the dif-fraction limit.5-8 Semiconductors function as active componentswith strong emission/absorption, optical gain, and nonlinear re-sponses. Numerous ultracompact photonic devices based onmetallic-semiconductor hybrid structures have already beendemonstrated including photodetectors,9,10 optical modulators,11

and nanoscale lasers.12,13

The properties of hybrid nanostructures depend on thephotonic and electronic coupling between the individual com-ponents, which in turn sensitively depend on composition andgeometry. While this strong dependence on structure presentsexciting opportunities to create hybrid materials with broadlytunable properties, it also poses serious challenges for fabricationand characterization. Small structural inhomogenieties in synthe-sized samples result in large variations in properties. Thus, theunderlying relationship between structure and function ismasked inensemble measurements. More elaborate experiments have demon-strated that the localized fields near metallic surfaces and NPsprofoundly alter properties of single emitters (e.g., fluorescentmolecules14-17 and nitrogen vacancy centers in diamond nano-crystals18). In particular, previous experiments have reportedmodified photoluminescence (PL) decays and blinking dynamicswhen semiconductor quantum dots (QDs) are coupled to nearby

metallic films or nanostructures.19-27 However, the experimen-tal results remain inconsistent, and a quantitative agreement withtheory has been critically lacking. Carefully designed experimentsconducted on individual hybrid structures with well-defined config-urations are necessary to obtain accurate and reliable informationon electronic/photonic couplings.

In this paper, we report the manipulation of coupling in a sim-ple model system, a single semiconductor QD near a single metallicNP, and study the resulting changes in QD PL dynamics. Weused atomic force microscopy (AFM) nanomanipulation28-31 tocontrollably push a Au NP proximal to a CdSe/ZnS QD. Weobserved gradual and reversible changes in the QD PL lifetimeand blinking dynamics. In some cases, the total lifetime reducedfrom 30 ns to well below 1 ns. The approximately 2 orders ofmagnitude change in lifetime is accompanied by the disappear-ance of blinking behavior as the nonradiative energy transferfrom the CdSe/ZnS QD to the Au NP becomes the dominantdecay channel. Both features exhibited by the hybrid structure(reduced blinking and drastically shortened total lifetimes) areadvantageous in light-emitting applications. For example, theshortened lifetimes may result in devices with higher modulationspeed and improved efficiency. In comparison to previous studies,our experiments report changes in the PL dynamics of the sameQD, therefore, eliminating the ambiguity of variable properties ofindividual QDs. These experiments enable one to extract morereliable information on electronic/photonic coupling and modified

Received: November 6, 2010Revised: January 13, 2011

1050 dx.doi.org/10.1021/nl103906f |Nano Lett. 2011, 11, 1049–1054

Nano Letters LETTER

optical properties of hybrid nanostructures. Such information isvaluable to a wide range of applications. For instance in aper-tureless near-field scanning optical microscopy, the tip-sampleinteraction often complicates the interpretation of images ob-tained and needs to be taken into account explicitly.

Individual hybrid structures were assembled and characterizedusing a combined AFM/inverted confocal microscope (Figure 1a).Samples were first prepared by spin coating colloidal CdSe/ZnSQDs (NN-laboratories) and dextran-coated Au NPs (averagediameter ∼35 nm, Nanocs) onto glass coverslips. The concen-tration of the particles was adjusted to be around 0.05 QDs/μm2

and 3 Au NPs/μm2 for experiments on single nanostructures.The sample was then excited by the frequency doubled output ofa mode-locked Ti:sapphire laser. A 100� (N.A. = 0.7) or 60�(N.A. = 0.95) objective focused the light to a diffraction-limitedspot on the sample, and the same objective collected the PL. Thecollected light was directed through a series of optical filters toone of three detectors: a camera for optical imaging and align-ment, a spectrograph, or an avalanche photodiode (APD) forconfocal imaging and time-resolved PLmeasurements. The APDwas part of a time-correlated single photon counting (TCSPC)setup with a measured instrument response function of about500 ps. Used in time-tagged time-resolved mode, the TCSPCsetup simultaneously monitored the blinking behavior and life-time of the QDs. For the measurements presented here, the QDPL was measured for durations from 180-300 s. Furthermore,all lifetime measurements were analyzed using a deconvolutiontechnique (i.e., least-squares curve fitting32) to account for theinfluence of the instrument response function.

One benefit of this multi-imaging setup is its ability to correlatestructural information with optical properties. To achieve thisgoal, the AFM cantilever tip was centered within the excitationspot of the confocal microscope. By raster scanning the samplestage, we simultaneously acquired correlated AFM and confocalimages. One example of these images is shown in Figure 1b-d

for a 4.5 μmby 4.5 μm area. Four bright QDs are identified in theconfocal optical image (Figure 1c) and can be correlated withthose present in the AFM image (Figure 1b). An enlarged sectionof the AFM image of a QD, along with a line cut, is shown inFigure 1d.

Basic spectral characterizations of ensembles of CdSe/ZnSQDs and Au NPs are shown in Figure 2a. The extinction spec-trum of the AuNPs exhibited a plasmon resonance peak at 530 nm.The ensemble PL peak of the CdSe/ZnS QDs was at 635 nm(fwhm ∼30 nm). In Figure 2a, we also display the 525 nm laserexcitation wavelength which was positioned to overlap with theAu NP plasmon resonance.

The general process used to study the individual hybrid struc-tures is illustrated in panels b-h of Figure 2. Figure 2b shows theAFM scan of a single QD located more than 300 nm away fromthe nearest Au NP. Before assembling a QD-NP structure, weinvestigated the PL decay lifetime and blinking of the QD. TheQD PL trajectory (Figure 2c) displayed common blinking behaviorwith distinct on/off-periods, which resulted in a bimodal distri-bution for the count rate histogram (right panel in Figure 2c).Previous experiments demonstrated that fluctuating nonradiativedecay processes lead to bimodal intensity variations and multi-exponential PL decays. However, when a single QD’s lumines-cence intensity is high, the QD PL decay is dominated by radiativerecombination and should be nearly a single exponential.33 Tominimize the influence of the fluctuating nonradiative relaxationprocesses, we applied a cutoff threshold34 to the PL trajectory(red line in Figure 2c). In doing so, we used only those photonscounted above this line to construct the QD PL decay. Theresulting decay shown in Figure 2d is well fit as a single exponentialwith a lifetime of 34 ns.

Coupling between the QD and Au NP was controlled bypushing the Au NP (indicated by the white arrow in Figure 2b)toward the QD. The height of this particular NPwas measured tobe 44 nm. Due to convolution of the AFM tip shape with thesample topography, we can only estimate the center-to-centerdistance between the QD and NP to be below 50 nm (AFMimage inset in Figure 2e). Before collecting the PL trajectory, theincident light polarization was aligned along the axis connectingthe QD and Au NP. This alignment was accomplished by rotatingthe incident light polarization to maximize the QD PL emissionintensity. As shown in Figure 2e, we measured a small increasein the maximum PL count rate and, more prominently, areduction of off-periods. In the histogram of the count rates(right panel in Figure 2e), this change in blinking can be seen inthe reduction of the off peak of the bimodal distribution. Again,we applied a threshold to construct the PL decay. The resultingPL decay, plotted in Figure 2f, showed approximately an order ofmagnitude decrease in lifetime to 1.7 ns (χr

2 = 2.55). To illustratethat the change in PL dynamics is reversible, the Au NP was thenpushed far away from the QD (AFM image inset of Figure 2g).The subsequent PL trajectory showed a reappearance of sig-nificantly more off-times (Figure 2g), resulting in a large off-timepeak in the count rate histogram (right panel in Figure 2g). ThePL decay lifetime (Figure 2h) was remeasured to be 27 ns. Thediscrepancy in the measured lifetimes between Figure 2d and 2hmay be attributed to possible changes in the local environmentaround the QD from leftover debris after pushing the Au NPaway.35

In these hybrid structures, the strength of the QD-NP cou-pling strongly depends on theQD-NPseparation. As an example, agradual change in the QD PL lifetime for an approaching Au NP is

Figure 1. (a) Schematic of experimental setup. (b) AFM topographyand (c) simultaneous confocal image of QDs on a glass coverslip. Anidentical white box in each image helps to illustrate the correspondencebetween the images. In the AFM image, the four dots located at thecorners of the box correspond to the four bright dots in the confocalimage. (d) Enlarged AFM image of a QD with a line cut.

1051 dx.doi.org/10.1021/nl103906f |Nano Lett. 2011, 11, 1049–1054

Nano Letters LETTER

demonstrated in Figure 3. The PL lifetime of a selected QD(Figure 3b) was initially measured to be 35 ns (Figure 3a, black).A goldNPwas then pushed close to theQD (Figure 3c) such thatthe QD remains visible. The PL lifetime was found to decrease to26 ns (Figure 3a, blue). The Au NP was pushed again such thatthe QD was no longer visible in the AFM image (Figure 3d),which resulted in a PL lifetime measurement of 22 ns (Figure 3a,green).

The strong dependence of coupling on separation distanceleads to a wide variety of optical behavior.We present one particularexample where we observed the disappearance of distinct on/offevents for a QD with a dramatically reduced lifetime. Figure 4ashows a single QD located approximately 200 nm away from twoAu NPs. The initial QD PL trajectory (Figure 4c) displayeddistinct on/off times. The corresponding PL decay shown inFigure 4e was measured to be 29 ns. After a Au NP was pushednear the QD (Figure 4b), the average photon count decreased toabout 40% of the original value (Figure 4d). At the same time, nodistinct off time events were recorded which resulted in a singlepeak in the histogram of the count rate (right panel in Figure 4d).Since there were no clear off times, we used all the counted photonsto construct the QD PL decay (Figure 4f). The apparentbiexponential decay yielded lifetimes of 0.2 and 1.6 ns (χr

2 =1.47). The fast decay component was robust regardless the details

of the analysis approach we chose. We quote this fast decay timeas the total lifetime in the following discussion (see SupportingInformation for detailed discussion).

In the following, we analyze in detail different aspects ofmodifiedPL dynamics of theQD. Specifically, we focus on various decay rateenhancements, the PL emission enhancement (or quenching),and the reduction of blinking behavior. In general terms, PL is atwo-step process consisting of the excitation of the QD followedby QD emission. The PL emission rate can be written as14,19 γPL =γexcY, where γexc is the QD excitation rate and Y is the quantumyield. The quantum yield describes the probability that theexcited QD emits a photon and is given by Y = kr /(kr þ knr),

Figure 2. (a) Ensemble optical spectra: Au NP extinction spectrum (black), QD PL spectrum (red), and excitation laser wavelength (green). (b) AFMimage of a QD and Au NPs with a white arrow denoting the path of the Au NP. Emission trajectories and corresponding PL decays are shown for (c, d)the QD alone, (e, f) the QD near the Au NP, and (g, h) the QD after Au NP was pushed away. In all emission trajectories, 50 ms time bins are used, redlines denote threshold cutoffs, and AFM images (250 nm by 250 nm) of the structures are shown as insets.

Figure 3. Gradual change in the QD PL lifetime due to an approachingAuNP: (a) PL lifetimes of 35 ns (black), 26 ns (blue), and 22 ns (green)correspond to AFM images (b), (c), and (d), respectively. The decays in(a) are displayed after the application of a smoothing filter.

Figure 4. (a) AFM image of a single QD near two 35 nm Au NPs. Thewhite circle encloses the QD. (b) AFM image after Au NP was pushednear the QD (black arrow denotes Au NP path). (c) Emission trajectoryof QD (50ms time bins, red line denotes threshold cutoff). (d) Emissiontrajectory of QD in the presence of AuNP (50ms time bins). (e) QDPLdecay with a single exponential fit of 29 ns (red line). (f) Total lifetimefor assembled structure with biexponential decay fit (red line, τfast = 0.2ns and τslow = 1.6 ns). The blue line is the measured instrument responsefunction.

1052 dx.doi.org/10.1021/nl103906f |Nano Lett. 2011, 11, 1049–1054

Nano Letters LETTER

where kr is the radiative decay rate and knr is the nonradiativedecay rate. The sum of kr and knr is the total decay rate kt that wemeasure experimentally. The total decay rate is simply related tothe total lifetime as kt = 1/τ.

In the hybrid nanostructure, the Au NPmodifies both steps ofthe PL process (i.e., γexc and Y). During excitation, the incidentlight is not only absorbed by the QD but also excites localizedsurface plasmons in the Au NP. The resulting enhanced localfield couples to the QD and increases γexc. In the subsequentemission process, the presence of the metallic NP alters thequantum yield through modification of both the radiative andnonradiative decay rates.36 Through a process commonly knownas the Purcell effect, theQD’s radiative decay rate is altered due tochanges in the local electromagnetic density of states introducedby the Au NP.37 Additionally, as the separation between the QDand metallic NP decreases, the nonradiative rate of the QDincreases due to energy transfer to the metallic NP.

To describe the process more quantitatively, the PL enhance-ment of the QD coupled to the Au NP is written as14,19

ηPL ¼γexcYγexc,0Y0

� ηexcηrηt

ð1Þ

where we define component enhancements ηexc = γexc/γexc,0,ηr = kr /kr,0, and ηt = kt /kt,0 (subscript 0 always refers to the bareQD). Within an order of magnitude, we may equate the excita-tion and radiative enhancements,19,38 ηexc ≈ ηr, which leads toηPL = ηr

2/ηt, or equivalently

ηr �ffiffiffiffiffiffiffiffiffiffiffi

ηPLηtp ð2Þ

It is easy to see that ηt = Y0ηrþ (1- Y0)ηnr, where ηnr = knr/knr,0is the analogous nonradiative enhancement. Solving for ηnr, alsoutilizing eq 2, gives19

ηnr �ηt-Y0

ffiffiffiffiffiffiffiffiffiffiffi

ηPLηtp

1-Y0ð3Þ

Using Figure 4 as an example, τt = 0.2 ns and τ0 = 29 ns yields atotal decay rate enhancement ηt = (krþ knr)/(kr,0þ knr,0) = τ0/τt = 145. Inserting the measured values of ηPL = 0.4 and ηt = 145into eq 2 gives a radiative decay rate enhancement of about 8. Thenonradative decay rate enhancement depends on the bare QDquantum yield, Y0, and is less straightforward to analyze. Thequantum yield of QDs on a substrate may deviate from thosemeasured in solution. Therefore, we calculate the nonradiativeenhancement at two different quantum yields for bare QDs. ForY0 = 50%, the nonradiative enhancement is about 300 while forY0 = 90% the nonradiative enhancement is about 1400.

The above analysis allows us to extract various decay rateenhancements from our experiments with a few assumptions.Wecan compare these results with those of a simple theoreticalmodel of a radiating point dipole near a metal sphere, for whichan exact analytical solution is possible.39 We implemented thisanalytical solution with the empirical dielectric constant for gold.40

Shown in Figure 5a are the calculated radiative, nonradiative, andtotal decay rate enhancement factors of a dipole emitter (emissionwavelength 635 nm) plotted as a function of the separation from a35 nm Au NP. These calculations are for the dipole orientedalong the QD-NP axis and assume a perfect emitter, i.e., Y0 = 1.To compare with experiment, one needs to know the QD-NPseparation. The tip convolution effect, unfortunately, makes itdifficult to measure this separation accurately from AFM images.

However, Figure 5a shows that the total decay enhancement, ηt,depends on the QD-NP separation sensitively in a monotonicfashion. Since there is a 1:1 correspondence between ηt andseparation, one can deduce the separation for a given hybridsystem from the measured total decay enhancement. The hybridstructure assembled in Figure 4 resulted in ηt = 145. Comparedto the calculations, the QD-NP separation is estimated to be3-4 nm. We can then contrast our experimental results for theindividual radiative and nonradiative enhancements with thetheoretical ones evaluated for such separations. At the 3 nmQD-NP separation, the calculated radiative decay enhancementis 7.2 and the nonradiative decay rate enhancement is 291. Bothare in excellent agreement with the extracted experimental values.

Many previous experiments have investigated the PL enhance-ment, ηPL, for an emitter near metallic NPs and reported incon-sistent numbers. We can calculate ηPL based on the dipoleemitter model and compare with our experimental measure-ments. Figure 5b shows our calculated ηexc, ηY = ηr/ηt, and ηPL =ηexcηY values as a function of the QD-NP separation. In thesecalculations we take ηexc = |E/E0|

2, where |E/E0| is the quasi-static limit field enhancement for a 35 nm diameter Au NPevaluated at the location of the emitter.41 The incident light isassumed to be at 525 nm and polarized along the QD-NP axis.Figure 5b shows that the maximum possible PL enhancement ofonly ∼1.2 occurs at a separation distance of around 20 nm.

Due to the 1:1 relation between ηt and separation notedearlier, the total PL enhancement, ηPL, and the total decay rateenhancement ηt, for a given AuNP diameter, should form a curve

Figure 5. (a) Radiative (red), nonradiative (blue), and total (black) decayrate enhancement for a point dipole emitter (emission wavelength at635 nm) near a single 35 nm AuNP plotted as a function of the separation,d (from the center of the dipole to the surface of Au NP). (b) PL emissionrate enhancement (black) and its two contribution factors: excitation rateenhancement (blue), and quantum yield enhancement (red) for the sameQD-Au NP pair as (a). (c) PL emission rate enhancement plotted as afunction of the total decay rate enhancement. The lines are calculations for apoint dipole near a Au NP with 35 nm (black) and 45 nm (red) diameter.The dots are experimental data.We verified the heights of the AuNPs usingthe AFM images for the specific NPs.

1053 dx.doi.org/10.1021/nl103906f |Nano Lett. 2011, 11, 1049–1054

Nano Letters LETTER

that does not explicitly involve the QD-NP separation. Figure 5cshows two such curves determined with the dipole emittermodel. The solid black (dashed red) curve corresponds to a 35(45) nm NP. The solid points in Figure 5c are experimentalresults for all the hybrid structures assembled and presented inthis Letter.42 Since a larger NP was used for the hybrid structurein Figure 2, its corresponding data point (highlighted in red)should be compared to the red curve. All other data pointscorrelate with the black curve. We emphasize that both the PLand total decay rate enhancements are independent, directlymeasured quantities. The excellent agreement between these twodynamical parameters and the calculations strongly supports theconclusion that our assembled hybrid systems are well character-ized experimentally and sufficiently described by the simpledipole emitter model.

The fact that the PL enhancement factor is weak in the simplehybrid structures can be explained as follows. Generally speaking,the PL enhancement is the strongest for a poor emitter (i.e., anemitter with a low quantum yield). In our case, the quantum yieldfor the bare semiconductor QD is high. The specific PL enhance-ment value depends on many additional factors such as NP size,the spectral overlap of the QD and Au NP plasmon resonance,and the QD-NP separation.43 For the particular hybrid struc-tures studied here, the excitation enhancement does not becomesignificant until the separation is less than 20 nm, as seen inFigure 5b. At such distances the nonradiative enhancement dom-inates and lowers the quantum yield. The competition betweenthe two factors leads to a moderate PL enhancement for a certainrange of QD-NP separation.

The analysis of the PL decay dynamics also provides insightinto the modified blinking dynamics. It is generally accepted thatblinking in QDs is associated with a charging event.44,45 Thebright periods of the luminescence trajectories correspond to anuncharged QD, while the dark periods correspond to a chargedQD. The quantum yield of the QD during the bright periods isdetermined by the radiative rate and other nonradiative pro-cesses which occur on the 10-100 ns time scale. For a chargedQD, however, there is an additional and dominant nonradativedecay channel, i.e., the Auger process. Instead of emitting a photon,the recombination of an electron and hole leads to the promotionof an extra charge to a higher energy state. This Auger processoccurs on times scales of∼100 ps46 and dramatically reduces thequantum yield during these off periods. Since the luminescenceintensity is directly proportional to the quantum yield, a chargedQD appears “dark”; i.e., it does not emit photons above thedetection limit.

The presence of the Au NP increases both the radiative andnonradiative decay rates. In some cases, the rates are increasedenough such that they are comparable to the Auger process.Therefore, the quantum yield remains relatively constant regard-less the charging state of the QD. We again discuss the particularexample presented in Figure 4. While the radiative rate enhance-ment of 8 is moderate, the nonradiative decay rate is enhanced byat least 2 orders of magnitude. As a result, the QD PL lifetime isdrastically reduced and becomes comparable to the time scale ofthe Auger process. Only then do we observe a complete absence(Figure 4) or a considerable reduction (Figure 2) of blinking. Insystems which the QD PL lifetime is still relatively long, weobserve no appreciable changes in blinking. Notice that in hybridstructures where the nonradiative energy transfer to the Au NPdominates, the quantum yield is greatly reduced. One may wonderthen how it is still possible to measure PL. This is because the

overall luminescence intensity is proportional to the product ofthe quantum yield and excitation rate. The increased excitationrate due to the enhanced local field in the presence of the Au NPis sufficiently high to overcome the reduction in the quantumyield. A quantitative analysis of the modified blinking behavior inhybrid structures, e.g., the power laws for the on and off periods,is desirable. However, this task is very challenging. According to arecent study, the off periods can be analyzed more reliably.47 In thehybrid structure, however, the off periods are drastically reduced asthe QD approaches the NP. A much longer measurement time(∼2000 s) may allow for a systematic and quantitative analysis.

In summary, we have assembled a simple hybrid nanostructureconsisting of a single QD and a Au NP using AFM nanomani-pulation. As the geometry of the structure is varied, couplingbetween the two changes accordingly. Both radiative and non-radiative decay rates of the QD increase near the Au NP. In somecases, the nonradative energy transfer between the QD and AuNP dominates and leads to a complete disappearance of lumi-nescence blinking. Two directly measured parameters, the totaldecay rate enhancement and PL intensity enhancement, bothsensitively depend on the size of the Au NP and the QD-NPseparation. The correlation between these two parameters for allassembled hybrid structures agrees very well with simple analy-tical calculations. Themethod of actively altering the geometry ofa hybrid nanostructure, as demonstrated here, has allowed for areliable analysis of the underlying relation between the system’sstructure and properties. These experiments present a new levelcontrol in PL dynamic studies in individual hybrid structures. Inthe near future, these experiments may be extended to investigateelectronic and photonic coupling in hybrid structures system-atically. For example, one can vary the spectral alignment betweenthe plasmonic resonance and exciton resonance by choosing eitherQDs of different sizes or metallic NPs with different shapes (rodsor cubes) and configurations (dimers or trimers). Experimentsperformed on well-defined prototypical structures may guide thedesign of more complex hybrid materials such as those withthree-dimensional structures.

’ASSOCIATED CONTENT

bS Supporting Information. Discussion of PL efficiency ofAu NP and biexponential decay observed in the hybrid structureinvestigated in Figure 4f. This material is available free of chargevia the Internet at http://pubs.acs.org.

’AUTHOR INFORMATION

Corresponding Author*Email: elaineli@physics.utexas.edu.

’ACKNOWLEDGMENT

We thank R. Bratschitsch, C. K. Shih, and Greg Sun for helpfuldiscussions. We also gratefully acknowledge financial supportfrom the following sources: NSF DMR-0747822, ONRN00014-08-1-0745, Welch Foundation F-1662, AFOSR FA9550-10-1-0022, and the Alfred P. Sloan Foundation. D.R. acknowledgesa fellowship from the NSF-IGERT program via Grant DGE-0549417. Use of the Center for Nanoscale Materials wassupported by the U.S. Department of Energy, Office of Science,Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.

1054 dx.doi.org/10.1021/nl103906f |Nano Lett. 2011, 11, 1049–1054

Nano Letters LETTER

’REFERENCES

(1) Colfen,H.;Mann, S.Angew. Chem., Int. Ed.2003, 42 (21), 2350–2365.(2) Shipway, A. N.; Katz, E.; Willner, I. ChemPhysChem 2000, 1

(1), 18–52.(3) Henzie, J.; Barton, J. E.; Stender, C. L.; Odom, T. W. Acc. Chem.

Res. 2006, 39 (4), 249–257.(4) Shevchenko, E. V.; Talapin, D. V.; Kotov, N. A.; O’Brien,

S.; Murray, C. B. Nature 2006, 439 (7072), 55–59.(5) Barnes, W. L.; Dereux, A.; Ebbesen, T. W. Nature 2003, 424

(6950), 824–830.(6) Pelton, M.; Aizpurua, J.; Bryant, G. Laser Photonics Rev. 2008, 2

(3), 136–159.(7) Schuller, J. A.; Barnard, E. S.; Cai, W.; Jun, Y. C.; White, J. S.;

Brongersma, M. L. Nat. Mater. 2010, 9 (3), 193–204.(8) Lal, S.; Link, S.; Halas, N. J. Nat. Photonics 2007, 1, 641–648.(9) White, J. S.; Veronis, G.; Yu, Z. F.; Barnard, E. S.; Chandran, A.;

Fan, S. H.; Brongersma, M. L. Opt. Lett. 2009, 34 (5), 686–688.(10) Ishi, T.; Fujikata, J.; Makita, K.; Baba, T.; Ohashi, K. Jpn. J. Appl.

Phys., Part 2 2005, 44 (12-15), L364–L366.(11) Pacifici, D.; Lezec, H. J.; Atwater, H. A. Nat. Photonics 2007

1 (7), 402–406.(12) Oulton, R. F.; Sorger, V. J.; Zentgraf, T.; Ma, R. M.; Gladden,

C.; Dai, L.; Bartal, G.; Zhang, X. Nature 2009, 461 (7264), 629–632.(13) Noginov, M. A.; Zhu, G.; Belgrave, A. M.; Bakker, R.; Shalaev,

V. M.; Narimanov, E. E.; Stout, S.; Herz, E.; Suteewong, T.; Wiesner,U. Nature 2009, 460 (7259), 1110–1112.(14) Anger, P.; Bharadwaj, P.; Novotny, L. Phys. Rev. Lett. 2006, 96

(11), No. 113002.(15) Kuhn, S.; Hakanson, U.; Rogobete, L.; Sandoghdar, V. Phys.

Rev. Lett. 2006, 97 (1), No. 017402.(16) Bek, A.; Jansen, R.; Ringler, M.; Mayilo, S.; Klar, T. A.;

Feldmann, J. Nano Lett. 2008, 8 (2), 485–490.(17) Ringler, M.; Schwemer, A.; Wunderlich, M.; Nichtl, A.;

Kurzinger, K.; Klar, T. A.; Feldmann, J. Phys. Rev. Lett. 2008, 100(20), No. 203002.(18) Schietinger, S.; Barth, M.; Alchele, T.; Benson, O. Nano Lett.

2009, 9 (4), 1694–1698.(19) Shimizu, K. T.; Woo,W. K.; Fisher, B. R.; Eisler, H. J.; Bawendi,

M. G. Phys. Rev. Lett. 2002, 89 (11), No. 117401.(20) Wu, X. H.; Sun, Y. G.; Pelton, M. Phys. Chem. Chem. Phys. 2009,

11 (28), 5867–5870.(21) Fu, Y.; Zhang, J.; Lakowicz, J. R. Chem. Commun. 2009, 3, 313–

315.(22) Masuo, S.; Naiki, H.; Machida, S.; Itaya, A. Appl. Phys. Lett.

2009, 95 (19), No. 193106.(23) Matsumoto, Y.; Kanemoto, R.; Itoh, T.; Nakanishi, S.; Ishikawa,

M.; Biju, V. J. Phys. Chem. C 2008, 112 (5), 1345–1350.(24) Ma, X. D.; Tan, H.; Kipp, T.; Mews, A. Nano Lett. 2010, 10

(10), 4166–4174.(25) Fu, Y.; Zhang, J.; Lakowicz, J. R. Chem. Phys. Lett. 2007, 447,

96–100.(26) Matsuda, K.; Ito, Y.; Kanemitsu, Y. Appl. Phys. Lett. 2008, 92

(21), 211911.(27) Yuan, C. T.; Yu, P.; Tang, J. Appl. Phys. Lett. 2009, 94 (24),

243108.(28) Junno, T.; Deppert, K.; Montelius, L.; Samuelson, L. Appl. Phys.

Lett. 1995, 66 (26), 3627–3629.(29) Requicha, A. A. G., Nanomanipulation with the atomic force

microscope. InNanotechnology; Waser, R., Ed.; Wiley-VCH: Weinheim,2008; Vol. 3, pp 239-273 and references therein.(30) Kim, S.; Ratchford, D. C.; Li, X. Q. ACS Nano 2009, 3 (10),

2989–2994.(31) Merlein, J.; Kahl, M.; Zuschlag, A.; Sell, A.; Halm, A.; Boneberg,

J.; Leiderer, P.; Leitenstorfer, A.; Bratschitsch, R. Nat. Photonics 2008, 2(4), 230–233.(32) O’Connor, D.; Phillips, D., Time-correlated Single Photon Count-

ing; Academic Press, Inc.: London, 1984.

(33) Fisher, B. R.; Eisler, H. J.; Stott, N. E.; Bawendi, M. G. J. Phys.Chem. B 2004, 108 (1), 143–148.

(34) For all PL trajectories, the threshold cutoffwas set to 50% of thedifference between the maximum and minimum emission count rate.

(35) Another possibility for the discrepancy in the lifetimes is a longtime decrease in the QD’s PL lifetimes. Assembly and characterizingthese structures may take a few hours to complete. We have measuredthe long time change in lifetime of bare QDs in air and measureddecreases in lifetime of up to 5 ns over a few hours. We can speculate thatthis may be caused by photooxidation.

(36) Chance, R.; Prock, A.; Silbey, R. Adv. Chem. Phys. 1978, 37, 1–65.(37) Purcell, E. Phys. Rev. 1946, 69 (11-12), 674.(38) Gersten, J.; Nitzan, A. J. Chem. Phys. 1980, 73 (7), 3023–3037.(39) Ruppin, R. J. Chem. Phys. 1982, 76 (4), 1681–1684.(40) Johnson, P. B.; Christy, R. W. Phys. Rev. B 1972, 6 (12), 4370–

4379.(41) Jackson, J. Classical Electrodynamics, 3rd ed.; John Wiley and

Sons, Inc.: New York, 1999.(42) A few factors contribute to the experimental uncertainties in

both total decay rate and PL enhancements. The error in fittingindividual total decay time is small, ∼0.1 ns. The decay rate, however,does fluctuate over time up to a few nanoseconds for bare QDs. Inmeasurement of PL enhancement, a QD not exactly at the center of theconfocal excitation spot leads to a smaller photon count rate. In addition,the presence of a metallic NP may change the collection efficiency of PLfrom a QD.

(43) Sun, G.; Khurgin, J. B.; Soref, R. A. Appl. Phys. Lett. 2009,94 (10), 101103.

(44) Frantsuzov, P.; Kuno, M.; Janko, B.; Marcus, R. A. Nat. Phys.2008, 4 (7), 519–522.

(45) Stefani, F. D.; Hoogenboom, J. P.; Barkai, E. Phys. Today 2009,62 (2), 34–39.

(46) Klimov, V. I.; Mikhailovsky, A. A.; McBranch, D. W.; Leatherdale,C. A.; Bawendi, M. G. Science 2000, 287 (5455), 1011–1013.

(47) Crouch, C. H.; Sauter, O.; Wu, X. H.; Purcell, R.; Querner, C.;Drndic, M.; Pelton, M. Nano Lett. 2010, 10 (5), 1692–1698.