www.elsevier.com/locate/cplett

Chemical Physics Letters 447 (2007) 96–100

Suppressed blinking in single quantum dots (QDs) immobilizednear silver island films (SIFs)

Yi Fu, Jian Zhang, Joseph R. Lakowicz *

Center for Fluorescence Spectroscopy, University of Maryland School of Medicine, 725 West Lombard Street, Baltimore, MD 21201, United States

Received 30 May 2007; in final form 30 August 2007Available online 6 September 2007

Abstract

In this report, we use single-molecule spectroscopic method to study emission behaviors of streptavidin-conjugated quantum dotsimmobilized on a biotinylated BSA (bovine serum albumin) monolayer near non-continuous rough silver nanostructures. We observedgreatly reduced blinking and enhanced emission fluorescence of quantum dots next to silver island films.� 2007 Published by Elsevier B.V.

1. Introduction

Semiconducting quantum dots (QDs) have beenreported to posses numerous photophysical properties thatare superior to those of organic fluorophores. The emissionproperties of quantum dots, especially high-absorptioncross-section, exceptional photostability, wide excitationspectra and narrow emission bands, are important to livecell imaging [1–3] and FRET biosensors [4]. However, sin-gle Quantum Dot shows strong fluorescence blinking. Thefluorescence from individual QDs turns ‘on’ and ‘off’ ontimescales ranging from milliseconds to many minutes.The long ‘off’ periods are the main restriction for the useof QDots in imaging applications and limit their applicabil-ity in single-molecule detection. The origin of the fluores-cence intermittency has been studied in detail [5–7]. It hasbeen postulated that blinking arises from random ioniza-tion events that eject a carrier from the QD. The distribu-tion of ‘on’- and ‘off’-time differs from the exponential ornear-exponential distributions observed in most other typesof single fluorophores.

Many previous studies on blinking of the bright fractionand on dark fractions have been based on measurements of

0009-2614/$ - see front matter � 2007 Published by Elsevier B.V.

doi:10.1016/j.cplett.2007.08.089

* Corresponding author. Fax: +1 410 706 8408.E-mail address: Lakowicz@cfs.umbi.umd.edu (J.R. Lakowicz).

individual QDs immobilized onto dielectric surfaces, suchas glass coverslip. It has also been shown that local envi-ronments play a key role in this blinking phenomena andcan affect the emission properties of QDs [8–13]. Surround-ing the QDs with oligo ligands [8] or thiol groups [9] sup-presses the blinking. This suppression is presumably dueto the passivation of surface trap by ligand groups.Enhanced luminescence in the presence of gold colloids[10,12] have also been reported recently. A recent investiga-tion in our lab has revealed the enhanced fluorescence ofhome-synthesized QDs adsorpted directly on silver islandfilm [11]. Despite recent progress, much work needs to bedone to achieve reproducible and robust surface fuctional-ization by developing flexible bioconjugation techniques.Experimental results from various spatial configurationsin addition to QD preparation conditions differ in termsof spectral properties of QDs. In addition, as the emissionproperties of the individual adsorbed QDs vary dramati-cally with respect to the assorted experiment conditionsand rough metallic surface topology, it is difficult to evalu-ate their emission characteristics accurately. Here wedescribe a method to apply commercially available strep-tadvin-conjugated QD coupled to biotinylated BSA mono-layer to achieve structural control between the QD and themetallic nanostructure. The surface derivatization of QDswith organic ligands enables the QDs to disperse in avariety of supported matrices. The use of BSA monolayer

Y. Fu et al. / Chemical Physics Letters 447 (2007) 96–100 97

technique offers more flexibility in spatial geometry controland provides a more convenient approach for biologicalapplication.

One important technique for the elucidation opticalproperties of quantum dots is single-molecule fluorescencespectroscopy, which eliminates averaging overall membersof the ensemble and can reveal fundamental propertiesotherwise hidden in ensemble measurement. Time-depen-dent fluorescence fluctuations from a single QD revealthe presence of fluorescence intermittency. Our single-mol-ecule studies in this work show the suppressed blinking ofQuantum dots near SIF.

2. Materials and methods

Single-molecule measurements were performed using aconfocal microscopy system (MicroTime 200, Picoquant,Germany) with an excitation line at 470 nm. Images wererecorded by raster scanning (in a bidirectional fashion)the sample over the focused spot of the incident laser witha pixel integration of 0.6 ms. The excitation power intothe microscope was maintained less than 0.1 lW. Time-dependent fluorescence data were collected with a dwelltime of 50 ms. The fluorescence lifetimes of single moleculeswere measured by time-correlated single photon counting(TCSPC) with the TimeHarp 200 PCI-board (PicoQuant).The data was stored in the time-tagged-time-resolved(TTTR) mode, which allows recording every detected pho-ton with its individual timing information. In combinationwith a pulsed diode laser, Instrument Response Function(IRF) widths of about 300 ps FWHM can be obtained,which permits the recording of sub-nanosecond fluores-cence lifetimes, extendable to less than 100 ps with reconvo-lution. Lifetimes were estimated by fitting to a v2-value ofless than 1.2 and with a residuals trace that was fully sym-metrical about the zero axis. All measurements were per-formed in a dark compartment at room temperature.Silver island films (SIFs) were deposited on cleaned glasscoverslips by reduction of silver nitrate as reported previ-ously [14]. The formed silver island films are greenish andnon-continuous. Only one side of each slide was coated withSIF. The particles are typically 100–500 nm across and70 nm high covering about 20% of the surface [15]. TheQDs immobilization procedure is illustrated in Scheme 1.A solution of 1 mg/mL biotinylated BSA (Sigma) in

Scheme 1. Immobilization of streptavidin-conjugated quantum dots onbiotinylated BSA monolayer (the glass surface was coated with silverisland film).

0.1 · PBS (pH 7.2) was applied onto the surface and theslide was incubated in a humid chamber overnight at5 �C, then extensively washed with distilled water. In a fol-lowing step, a solution of streptavidin-conjugated CdSe/ZnS quantum dot (Qdot 655, Quantum Dot Corp.) wasdeposited and the slide was incubated for 1 h. The QD con-centration was adjusted at nanomolar levels (0.1–0.5 nM) togive an appropriate surface density for single-moleculestudies. After incubation, the dye-immobilized coverslipwas thoroughly rinsed with HEPEs buffer solution toremove loosely bound QD molecules and washed exten-sively with 0.1X PBS buffer.

3. Results and discussion

The blinking phenomenon is confirmed by the single-molecule fluorescence images. During the raster scan, thebright spots in the image demonstrate the characteristicfluorescence intermittency associated with single quantumdots immobilized on a BSA-coated glass coverslip. Asclearly illustrated in Fig. 1a, a single molecule stops emittingfluorescence, the ‘off’ time can be observed as a dark stripein the spot of the image and followed by reoccurrence of theflorescence (‘on’ time). On the contrary, one can clearlyobserve that whereas some molecules emit nearly continu-ously in the presence of silver island films (Fig. 1).

Fig. 1 displays typical time traces of the fluorescenceintensity from single strepavidin QDs bound to a BSA-bio-tin layer on glass substrate. In Fig. 1a, under continuousillumination, the single QD switch abruptly between a fluo-rescent ‘on’ state and a non-fluorescent ‘off’ or weak fluo-

Fig. 1. Single-molecule fluorescence images (10 lm · 10 lm) and repre-sentative fluorescence time traces and their expanded sections recorded forsingle QDs immobilized on glass substrates. Each image pixel has a 0.6 msdwell time and the fluorescence intensity is displayed in a colorized scale,ranging from dark to light.

Fig. 3. ‘Off’ time histograms compiled from around 70 single QDmolecules. The lines are approximately inverse power-law fits to thehistogram, giving rise to an exponent of �1.74 on glass (open circle) and�2.5 on SIF (solid circle).

98 Y. Fu et al. / Chemical Physics Letters 447 (2007) 96–100

rescent ‘dim’ state. Two intensity levels are clearly distin-guishable beyond those expected from shot-noise limitedfluctuations. Another representative emission pattern isillustrated in Fig. 1b while the emission fluctuates stochas-tically compared to Fig. 1a. In the presence of metallicnanostructure, a large distribution of quantum dots showsas much as more than three-fold increase in fluorescenceintensity. In addition, a large fraction of molecule studied(>70%) did not display any long off times of interest. Thetime traces observed from SIF show relatively continuousrather than discrete intensity fluctuations (Fig. 2). Theinterval of ‘on’ times increases to a point where the blink-ing events rarely appears during the illumination. As illus-trated in Fig. 2b, the fluorescence intensity is well above thebackground noise level, we do not observe dark states dur-ing the 60-s observation time. To examine blinking on thelarge timescale of many seconds, a threshold was set at alevel of three-fold standard deviation with respect to themean background during the on/off histogram analysis todistinguish between the states as described previously[16,17]. We find that the decay of probability density forsingle QDs on glass is linear on a log–log scale and canbe fit by an inverse power-law kinetics, P(s) � s�a, whichhas been intensively discussed to describe the quantumdot blinking [5–7,18,19]. The approximately linear fit yieldsa value of 1.74 (Fig. 3), which resembles previous data [6,7].Similarly, we observed a roughly fit of off-time distributionfor single QDs in the presence of SIFs with a value of 2.50.A difference in the local environment and fluctuations ofthe QD is expected to be reflected in the blinking coeffi-cients aoff [20].

Fig. 2. Single-molecule fluorescence images (10 lm · 10 lm) and repre-sentative fluorescence time traces and their expanded sections recorded forsingle QDs immobilized on SIF substrates. Each image pixel has a 0.6 msdwell time and the fluorescence intensity is displayed in a colorized scale,ranging from dark to light.

There exist several possible mechanisms to explain thenature of fluorescence intermittency in semiconductingQDs. One popular explanation for the inverse power-lawbehavior of the ‘off’ time is the presence of multiple ioniza-tion states and accordingly a distribution of recombinationrates [5–7]. As an ‘off’ state is ascribed to be caused by aphoto-induced process of the QD, the ‘off’ time intervalsare associated with the recombination rate of the ejectedcharge carrier (most likely electrons) with the ionized dot.The ‘off’ state happens when one of the electrons in thephoto-excited exciton either becomes trapped at the QDsurface with some probability to reversibly neutralize thequantum dot after some time, or tunnels into the surround-ing environment. The ionized quantum dot appears to bedark. Emission resumes once the trapped charge carrierreturns to a delocalized state in the core or a nearby elec-tron is captured from the surrounding environment. Thismodel predicts the long-lived trap state, thereby well-defined sets of on/off durations should be observed regu-larly in the emission time transients [18,21]. However thefact that a large fraction (�65%) of single QDs displayedstochastic fluctuations as depicted in Fig. 1b, indicatingthat such long-live ionized state model can not explain wellwith the observed fluorescence intermittency. Recently,Frantsuzov and Marcus [22] presented an alternative mech-anism in their discussion of the correlation between thenon-radiative relaxation and the blinking of neutral QDs.In this model, the on/off blinking arises from fluctuationsin the non-radiative relaxation rate when the excited QDreturns directly (or via a surface state) back to its groundstate. The hole-trapping is an Auger process. The extrahole energy enables the electron excitation from 1Se stateto 1Pe state. The energy gap between 1Pe and 1Se statesis a stochastic process. It acts as a slow variable and createslarge variations in the trapping rate, which can be reflectedas stochastic fluorescence intensity fluctuation clearlyobserved in this experiment.

The large disparity of off-time distributions we observedin Fig. 3 is likely caused by the changes in non-radiativedecay channels due to the effect of surrounding metallicnanostructure. The metallic structure with the sub-wave-

0 2 4 6 8 100

5

10

15

20

Fre

quen

cy

Lifetime (ns)10 20 30

0

5

10

Fre

quen

cy

Lifetime (ns)0 2 4 6 8 10

0

5

10

15

20

Fre

quen

cy

Lifetime (ns)10 20 30

0

5

10

Fre

quen

cy

Lifetime (ns)

Fig. 4. Lifetime distributions and respective Gaussian fits of single QDs (black: on SIFs; gray: on glass).

Y. Fu et al. / Chemical Physics Letters 447 (2007) 96–100 99

length size usually displays an energy resonance arisingfrom the collective oscillation of migrated electrons on itssurfaces, which is defined as plasmon resonance [23,24].An emitter is described as an oscillating dipole to radiateenergy when emission occurs. When the dipole is localizednear the metal particle, the radiating energy from the fluo-rophore is dramatically altered through coupling with themetal plasmon resonance to cause a change of the emissionproperties, which is referred as radiative decay engineering(RDE) [15]. These energy resonances express as absorbance,excitation, and scattering and induce strongly enhancedlocal fields in a near field region spatially overlap withQDs. The highly enhanced scattering fields can be adsorbedby the QDs. As a result, the plasmon interaction couldamend the energy gap between the exciton hole and trappedhole, and thus eliminates hole-trapping process, leading toreduce the frequency of blinking. However, the plasmoniceffect on energy separations between the states is varied,for instance, due to heterogeneous distributions of plasmonfield, accounting for the remaining blinking events.

The fluorescence lifetime is sensitive to the local environ-ment of the molecule. The decay time is often mono-expo-nential and given by

s ¼ ðCþ knrÞ�1 ð1Þwhere C and knr are the radiative and non-radiative decayrate, respectively. The changes in knr are typically due tochanges in an emitter’s environment, quenching or RET.The radiative decay rate C is constant and any changesare primarily due to changes in refractive index. Proximityof quantum dots to metals can result in an increase in thetotal radiative decay rate [15]. The lifetime and the quan-tum yield near the metal then become

sm ¼ ðCþ Cm þ knrÞ�1 ð2ÞQm ¼ ðCþ CmÞ=ðCþ Cm þ jnrÞ ð3Þ

The radiative rate is given by C + Cm, when Cm is the partdue to the metal. The change in radiative rate results inremarkable effects such as the increase in quantum yieldand decrease in lifetime. We implemented the time-corre-lated single photon counting (TCSPC) measurement on asingle bright spot. Fig. 4 illustrate histograms of lifetimesacquired from single QDs on the unsilvered surface (gray)

and silvered surface (black), respectively. Each distributioncontains data from approximately 70 molecules and is fit toa Gaussian function. The mean of the lifetime distributionon glass is 19.5 ns. In the presence of SIFs, the lifetime dis-tribution shows a long tail and yields a mean value of2.4 ns. The decrease in lifetime implies that the radiativerate of the quantum dot is much larger on the SIF thanon glass [14,15,25,26], which is compatible with overall in-creased intensity and reduced ‘off’ time. The wide distribu-tions of lifetimes can be viewed as the effect of thefluctuations of the nonradiative relaxation rate postulatedby Frantsuzov and Marcus [22]. The unsymmetrical distri-bution of lifetimes on silvered surface could also arise fromdiverse metal-fluorophore interactions that may be due tothe heterogeneous surface topography of metallic nano-structure [25,27,28].

In conclusion, we observed the suppressed blinking ofsingle Quantum dot immobilized near silver island films.The ‘off’ time distribution of quantum dots near SIFs dis-played power-law dependence with a dissimilar exponentcompared with that of electromagnetic inert surface. Weascribe it to the modification of fluorescence by the surfaceplasmon resonance from the metallic nanostructure. Therecombination process is dramatically altered through cou-pling with the metal plasmon resonance to cause a change inthe fluorescence properties and lead to suppression of blink-ing. The strong interaction of the excited molecules with themetal nanostructures also dramatically shortens the lifetimeof the non-emissive state. The individual QDs showed theextended ‘on’ time and overall increased intensity nearmetallic nanostructures. The findings should make the useof single QDs for a variety of applications in light-emittingdevices.

Acknowledgement

This research was supported by grants from NIH, HG-02655, EB-00682, and NCRR, RR-08119.

References

[1] M. Bruchez, M. Moronne, P. Gin, S. Weiss, A.P. Alivisatos, Science281 (1998) 2013.

100 Y. Fu et al. / Chemical Physics Letters 447 (2007) 96–100

[2] W. Chan, S. Nie, Science 281 (1998) 2016.[3] I.L. Medintz, H.T. Uyeda, E.R. Goldman, H. Mattoussi, Nat. Mater.

435–446 (2005).[4] C.-Y. Zhang, H.-C. Yeh, M.T. Kuroki, T.-H. Wang, Nat. Mater. 4

(2005) 826.[5] M. Haase, C.G. Hubner, E. Reuther, A. Herrmann, K. Mullen, T.

Basche, J. Phys. Chem. B 108 (2004).[6] M. Kuno, D.P. Fromm, H.F. Hamann, A. Gallagher, D.J. Nesbitt, J.

Appl. Phys. 112 (2000) 3117.[7] M. Kuno, D.P. Fromm, H.F. Hamann, A. Gallagher, D.J. Nesbitt, J.

Chem. Phys. 115 (2001) 1028.[8] N.I. Hammer, K.T. Early, K. Sill, M.Y. Odoi, T. Emriek, M.D.

Barnes, J. Phys. Chem. B 110 (2006) 14167.[9] S. Hohng, T. Ha, J. Am. Chem. Soc. 126 (2004) 1324.

[10] O. Kulakovich et al., Nano Lett. 2 (2002) 1449.[11] K. Ray, R. Badugu, J.R. Lakowicz, J. Am. Chem. Soc. 128 (2006)

8898.[12] K.T. Shimizu, W.K. Woo, B.R. Fisher, H.J. Eisler, M.G. Bawendi,

Phys. Rev. Lett. 89 (2002) 117401-1.[13] J.-H. Song, T. Atray, S. Shi, H. Urabe, A.V. Nurmikko, Nano Lett. 5

(2005) 1557.[14] J.R. Lakowicz, J. Malika, I. Gryczynski, Z. Gryczynski, C. Geddes, J.

Phys. D: Appl. Phys. 36 (2003) R240.[15] J.R. Lakowicz, Anal. Biochem. 298 (2001) 1.

[16] K.D. Weston, P.J. Carson, H. Metiu, S.K. Buratto, J. Chem. Phys.109 (1998) 7474.

[17] O. Panzer, W. Gohde, U.C. Fischer, H. Fuchs, K. Mullen, Adv.Mater. 10 (1998) 1469.

[18] J.R. Krogmeier, J. Hwang, Proc. SPIE 5705 (2005) 255.[19] F.D. Stefani, X. Zhong, W. Knoll, M. Han, M. Kreiter, New J. Phys.

7 (2005) 1.[20] W.G.J.H.M. van Sark, P.L.T.M. Frederix, A.A. Bol, H.C. Gerritsen,

A. Meijerink, Chem. Phys. Chem. 3 (2002).[21] A. Biebricher, M. Sauer, P. Tinnefeld, J. Phys. Chem. B 110 (2006)

5174.[22] P.A. Frantsuzov, R.A. Marcus, Phys. Rev. B 72 (2005) 155321.[23] P.A. Antunes, C.J.L. Constantino, R.F. Aroca, J. Duff, Langmuir 17

(2001) 2958.[24] A.S. Kumbhar, M.K. Kinnan, G. Chumanov, J. Am. Chem. Soc. 127

(2005) 12444.[25] J. Malicka, I. Gryczynski, J. Fang, J. Kusba, J.R. Lakowciz, J.

Fluorescence 12 (2002) 439.[26] A.L. Efros, M. Rosen, M. Kuno, M. Nirmal, D.J. Noris, M.

Bawendi, Phys. Rev. B 54 (1996) 4843.[27] F.R. Aussenegg, A. Leitner, M.E. Lippotch, H. Reinisch, M. Reigler,

Surf. Sci. 139 (1987) 935.[28] Y. Fu, J.R. Lakowicz, Anal. Chem. 78 (2006) 6238.