Digitizing Gold Nanoparticle-Based Colorimetric Assay by Imagingand Counting Single NanoparticlesLiang Yuan, Xian Wang, Yimin Fang, Chenbin Liu, Dan Jiang, Xiang Wo, Wei Wang,*and Hong-Yuan Chen

State Key Laboratory of Analytical Chemistry for Life Science and Collaborative Innovation Center of Chemistry for Life Sciences,School of Chemistry and Chemical Engineering, Nanjing University, Nanjing, Jiangsu 210093, China

*S Supporting Information

ABSTRACT: Gold colloid changes its color when the internano-particle distance changes. On the basis of analyte-inducedaggregation or disaggregation behavior of gold nanoparticles(AuNPs), versatile colorimetric assays have been developed formeasuring various kinds of analytes including proteins, DNA, smallmolecules, and ions. Traditional read-out signals, which are usuallymeasured by a spectrometer or naked eyes, are based on theaveraged extinction properties of a bulk solution containing billionsof nanoparticles. Averaged extinction property of a large amount ofnanoparticles diminished the contribution from rare events whenthe analyte concentration was low, thus resulting in limiteddetection sensitivity. Instead of measuring the averaged optical property from bulk colloid, in the present work, we proposed adigital counterpart of the colorimetric assay by imaging and counting individual AuNPs. This method quantified the analyteconcentration with the number percentage of large-sized AuNPs aggregates, which were digitally counted with surface plasmonresonance microscopy (SPRM), a plasmonic imaging technique recently developed by us and other groups. SPRM was able toidentify rare AuNPs aggregates despite their small population and greatly improved the detection sensitivity as demonstrated bytwo model systems based on analyte-induced aggregation and disaggregation, respectively. Furthermore, besides plasmonicAuNPs, SPRM is also suitable for imaging and counting nonplasmonic nanomaterials such as silica and metal oxide with poorextinction properties. It is thus anticipated that the present digitized assay holds a great potential for expanding the colorimetricassay to broad categories of nonplasmonic nanoparticles.

Gold nanoparticles (AuNPs) exhibit a large extinctioncoefficient due to localized surface plasmon resonance

(LSPR), which is the origin of the color of gold colloids.1,2

Because LSPR is highly sensitive to the interparticle distance,gold colloids often change color in the presence of certainanalytes via analyte-induced aggregation or disaggregation(dispersion), enabling fast and convenient colorimetric assaysfor the measurement of the analyte concentration. Since thepioneer work by Mirkin and co-workers,3 AuNPs-basedcolorimetric assays have been widely used to detect broadkinds of analytes including DNA, proteins, ions, and smallmolecules.4−6

Such assays rely on the sensitive dependence of extinctionproperty, i.e., the color, of gold colloids by changing theinterparticle distance. In a typical AuNPs-based colorimetricassay, gold colloids consisting of well-dispersed gold nano-particles were red colored due to a relatively large interparticledistance. The presence of analyte resulted in the formation ofAuNPs aggregates because of the chemical or electrostaticinteractions. Consequently, the reduced interparticle distanceled to a color change from red to purple or blue, from whichthe concentration of analyte was determined.7,8 So far, opticalextinction measured by a UV−visible spectrometer or naked

eyes has been the major read-out signal in colorimetric assays.However, the optical extinction property reflects the ensembleaverage behavior of billions of individual nanoparticles in thedetection volume. Ensemble measurements show a strong biason rare events because the statistic average usually diminishesthe contribution from a minor portion of nanoparticlepopulations. For example, when the analyte concentration isextremely low, very few aggregates are formed compared to thelarge population of original individual nanoparticles, which canhardly affect the averaged optical property. Therefore, opticalextinction measurements usually exhibit limited capability todetect aggregates with low concentration, resulting in limitedsensitivity.9

Single nanoparticle counting techniques have been emergingrapidly and proven powerful in ultrasensitive detection ofanalytes.10−14 Several techniques, including dark-field imag-ing,10,11,15,16 plasmonic scattering,17−19 and single-particle massspectroscopy,20−22 have been developed to count individualAuNPs, resulting in improved detection sensitivity of the

Received: November 9, 2015Accepted: January 13, 2016Published: January 13, 2016

Article

pubs.acs.org/ac

© 2016 American Chemical Society 2321 DOI: 10.1021/acs.analchem.5b04244Anal. Chem. 2016, 88, 2321−2326

colorimetric assay. Despite their powerful capabilities, dark-fieldimaging and plasmonic scattering remain challenging forAuNPs with diameters below 20 nm, which are the mostwidely used sizes in AuNPs-based colorimetric assays. Laserillumination and flow channel have placed additional complex-ity in the sample preparation and handling.Here, we propose a digitized counterpart of the AuNPs-based

colorimetric assay by imaging and counting individual nano-particles with surface plasmon resonance microscopy (SPRM),an optical imaging technique that was recently developed by usand other groups.23−27 The principle of this SPRM imagingsystem is illustrated in Scheme 1. A p-polarized incident light (λ

= 680 nm) is directed onto a gold-coated coverslip through anobjective with high numerical aperture (N.A. = 1.45), and thereflected light is captured with a CCD camera to produce aSPRM image. Under appropriate conditions, a surface plasmonwave is excited propagating along the gold film. When externalnano-objects (AuNPs herein) are present on the surface of goldfilm, they scatter the surface plasmon wave and produce achange in the SPRM image. The location and size of eachAuNP can be determined by analyzing the SPRM images. Inthe present work, we demonstrated that SPRM was capable ofcounting the number of gold nanoparticles and quantifying thesize of each nanoparticle simultaneously. Once AuNPsunderwent the aggregation in a typical colorimetric assay,extraordinary large individuals, i.e., aggregates, appeared evenbefore the change in average extinction signal could bedetected. Consequently, a digitized colorimetric assay wasestablished by distinguishing the size distribution of AuNPs inthe absence and presence of analytes. The imaging andcounting of gold nanoparticles suppressed the optical noise in adigital way, resulting in a greatly improved sensitivity whencompared with the results from other instrument-basedmeasurements such as a UV−vis spectrometer.

Instead of measuring the plasmonic property of AuNPs,SPRM measures the scattering of nano-objects through anexisting surface plasmon wave, resulting in two importantadvantages. First, SPRM signal is proven to be sensitive to thesize of AuNPs and less sensitive to its morphology.26,28,29

Second, SPRM is suitable for imaging not only plasmonicnanoparticles (gold26 and silver30) but also nonplasmonicnanoparticles (SiO2,

31 polystyrene,26 hydrogel,27 virus,31

bacterial,32 and so on). The present work thus paves the waytoward universal “colorimetric” assays involving analyte-induced aggregation of both plasmonic and nonplasmonicnanomaterials.

■ EXPERIMENTAL SECTIONMaterials. Chloroauric acid tetrahydrate (HAuCl4·4H2O)

and tris(hydroxymethyl) aminomethane were purchased fromSinopharm Chemical Reagent Co., Ltd. Adenosine triphosphate(ATP), trisodium citrate, and poly-L-lysine (PLL) solution(0.1%, w/v in H2O) were purchased from Sigma-Aldrich(Shanghai). Melamine (99%) was bought from Aladdin. Thephosphate buffered saline (PBS, 1×) was obtained fromThermo Scientific. Other reagents were all of analytical reagentgrade and used as received without further purification. Thedeionized water (DI water, 18.2 MΩ·cm resistivity) producedby Barnstead Smart2Pure3 UF (Thermo Fisher) was usedthroughout the study.

Sample Preparation. The SPR chips were BK7 glasscoverslips coated with 2 nm of chromium and then 47 nm ofgold. Each Au chip was rinsed with DI water and ethanol andblown dry under nitrogen before use. After the chip was furthercleaned with a hydrogen flame to remove possible remainingcontamination, 2% (w/v in H2O) PLL solution was droppedonto the chip surface for 3 h under room temperature. Whenthe self-assembly of the PLL layer was fabricated, the chip wassubsequently rinsed with DI water and 0.02× PBS solution. Allthe measurements of AuNPs aggregation were carried out atroom temperature (25 °C) in a 0.02× PBS buffer solution.

SPRM Setup and SPR Imaging. The experiment of SPRimaging was performed on an inverted microscope (OlympusIX83) by using a high numerical aperture oil immersionobjective 60× (NA 1.45). A red super luminescent lightemitting diode (SLED) with wavelength of 680 nm (1 mW,Qphotonics) was used as the light source, and a polarizer wasinserted on the optical path to generate p-polarized light so asto excite the surface plasmons at the Au substrate. The incidentangle of such light was also adjustable and optimized for thepurpose of delivering the largest surface plasmon response fromAuNPs. The SPR chip was placed on the objective with index-matching liquid. The SPRM images were recorded by a CCDcamera (AVT Pike F-032B from Allied Vision Technologies) ata frame rate of 3.0 frames per second. After injecting 0.02-foldPBS buffer (pH 7.4) on top of surface-modified SRP chip, theincident angle of the laser beam was adjusted to be near theSPR resonance angle at which the reflection recorded by thecamera reached minimum intensity. Then, a given volume ofAuNPs or mixed sample solution was drop into the liquid, andthe images started to be recorded. SPRM images were recordedat a speed of 3.0 frames per seconds (fps) at pixel resolution of640 × 480.

Image Processing. Once the plasmonic images ofdispersed and/or aggregated gold nanoparticles after mixingAuNPs with target analytes were captured, the raw SPR imagesrecorded from the camera were converted to 16-bit tiff format

Scheme 1. Schematic Illustration of the DigitizedColorimetric Assay by Imaging and Counting IndividualAuNPs with SPRMa

aSPRM is able to measure the size distribution of AuNPs in a digitalway, in which the size of each nanoparticle is individually determined.Quantification of analyte concentration is achieved by analyzing thenumber percentage of large-sized AuNPs aggregates in the sizedistributions of dispersed sample (red histogram) and aggregatedsample (purple histogram), respectively.

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.5b04244Anal. Chem. 2016, 88, 2321−2326

2322

files with a Matlab program. Then, the intensity of each particlewas determined by calculating the average intensity within afixed small rectangle region-of-interest (ROI) selected toinvolve the particle. After a series of Fast Fourier Trans-formation (FFT) of individual ROI, Gaussian fittings ofhistograms of particle SPR intensities were carried out withMatlab to demonstrate intensity distribution of differentaggregation cases.Detection of Melamine. 100 μL of AuNPs suspension was

first diluted 20 times with deionized water and used as a stockliquid for the detection of melamine. Different concentrationsof melamine analyte solutions (50 μL) were added into theabove colloidal AuNPs solution (350 μL). 80 μL of this mixedsample solution was dropped into the 20 μL 0.02× PBS buffer,and then, the aggregation of dispersed gold nanoparticles wasmeasured with an SPR imaging system. As a comparison, thecolor change was also observed by the naked eyes andabsorbance spectra photographs were taken with a PanasonicDMC-LX5 digital camera and recorded by a UV−visspectrometer.Detection of ATP Based on Aptamer Target Binding.

In the ATP assay, all the DNA oligonucleotides used weresynthesized from Sangon Biotechnology Inc. (Shanghai,China), and the sequences were as follows: ATP aptamer (a-ATP) (5′-ACC TGG GGG AGT ATT GCG GAG GAA GGT-3′) and the complementary oligonucleotide (c-ATP) (5′-ACCTTC CTC CGC AAT ACT CCC CCA GGT-3′). The assayprocedure was performed mainly according to the literature.33

In brief, both the a-ATP (10 μM) and c-ATP (10 μM)solutions were mixed in 25 mM Tris-buffer containing 0.3 MNaCl (pH 8.2) for the hybridization. ATP of desirableconcentrations was added into the solution and incubated ina 37 °C warm bath for 0.5 h. Subsequently, 2 μL of the abovesolution was injected into 100 μL of AuNPs, followed bytreatment of 20 μL of 10 mM PB buffer (0.2 M NaCl, pH 7.4)for triggering the aggregation of AuNPs. Finally, 80 μL of suchreaction sample was added into the 0.02× PBS for the SPRimaging measurement.

■ RESULTS AND DISCUSSIONImaging Individual Gold Nanoparticles with SPRM.

We first evaluated the capability of SPRM to image multipleAuNPs at the same time. 20 nm AuNPs were synthesized via aclassical citrate-reduction method. The thorough character-ization results were provided in Figure S1, including theextinction spectrum, dynamic light scattering (DLS), andtransmission electron microscopy (TEM). In a typical experi-ment, as-synthesized gold colloids were first diluted 20 timeswith deionized water, and then, 80 μL of the diluted colloidswere injected into an open-well chamber containing 20 μL of0.02-fold phosphate buffered saline (PBS), which was placed ontop of the gold-coated coverslip. The solution was thoroughlymixed for a couple of seconds to reach a final dilution factor of25 and then remained undisturbed to enable the diffusion andadsorption of AuNPs onto the gold film.Because the surface plasmon wave is a near-field electro-

magnetic wave, nano-objects that are closely attached to thegold film generate the largest SPRM signals. For this reason, wemodified the gold film with positively charged poly-L-lysine(PLL), enabling the efficient electrostatic adsorption ofnegatively charged AuNPs onto the gold film. Once anindividual nanoparticle hit and stayed at the surface of thegold film, a parabolic-shaped pattern would appear at the

location of the collision due to the scattering of surface plasmonwave (Figure 1A). The whole dynamic process was

continuously recorded by SPRM (Figure S2) and was providedas Movie S1, left panel. The camera frame rate was optimized toadjust the average number of collision events in the period ofone frame, so that each of them could be separately identified.It was found that all individual nanoparticles stroke the gold

film surface and stayed there permanently. We attributed thedriving force of the hit-n-stay to be Brownian motion andelectrostatic interaction. After the collision, electrostatic forceand van de Waals force kept the gold nanoparticle staying onthe gold film. With the present surface modification of goldfilm, we did not observe any detachment event from the surfacein a recording time of over 1 h.In order to minimize the background drift during long-term

recording, we introduced the differential SPRM images asshown in the center panel of Movie S1. Each differential SPRMimage was produced by subtracting the previous frame ofSPRM image from the present one. Due to the mathematicfeature of first-order derivative, a hit-n-stay event of a singlenanoparticle appeared as a blinking pattern in the differentialSPRM movies (Movie S1, center panel). Additionally, using thedifferential SPRM images could also differentiate the movingNPs near the surface from the static NP at the surface. In a “hitand move” scenario (Movie S2), the SPRM patterns of movingNPs repeatedly appear, disappear, and reappear within a similarposition of the view, indicating the constant movements. Whilein a “hit and stay” scenario (Movie S3), the SPRM pattern of asingle nanoparticle appears only once in the video, because thestatic nanoparticle is invisible in the differential SPRM images.

Imaging and Counting of Gold Nanoparticles. Wefurther demonstrated that SPRM was capable of imaging andcounting gold nanoparticles. Differential SPRM images enabledus to distinguish two spatially overlapped AuNPs by thedifferent times of arrival (see Figure S3). Thus, we did not needto move the field of observation when collecting AuNPs. About500 collision events during 300 s were recorded. With a home-

Figure 1. (A) Typical SPRM image of multiple AuNPs. Eachnanoparticle appears as a parabolic-tailed pattern, whose centerrepresents the location of nanoparticles. An image analysis algorithm isestablished to determine the size of each nanoparticle. (B) Sizedistributions of the same AuNPs sample in three measurements. (C)Size distributions of AuNPs before (blue) and after (green) passingthrough a filtration membrane (cutoff diameter = 0.05 μm).

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.5b04244Anal. Chem. 2016, 88, 2321−2326

2323

developed image processing algorithm, the number andlocation of individual gold nanoparticles were determined byrecognizing the parabolic-shaped patterns in the differentialSPRM movies (Movie S1, right panel). The size wasdetermined by analyzing the intensity of each pattern in a 2-D Fourier domain. Fourier transform greatly enhanced theaccuracy of size determination by minimizing the fluctuation ofselecting region of interests in the space domain. This methodhad been validated in our previous work30 and was described indetail in Figure S4. After collecting the SPRM signals of ∼500gold nanoparticles, a histogram could be created todemonstrate the size distribution as shown in Figure 1B.In order to demonstrate the reproducibility of sizing, three

independent experiments were sequentially performed togenerate three histograms of the same gold colloids sample.They were found to be consistent with each other (Figure 1B).The obtained histograms exhibited similar Gaussian distribu-tion sizes as that revealed in TEM results (Figure S1C),demonstrating the capability of SPRM to count and size as-synthesized 20 nm AuNPs.It is clear that the size distribution of original AuNPs must be

as narrow as possible in order to sensitively detect the analyte-induced aggregation. Thus, a filter membrane separation (cutoffdiameter of 0.05 μm) was applied to the as-prepared goldcolloids to avoid the false positive signal by removing any large-sized dust and unexpected aggregates formed during thesynthesis and storage. The size distributions of original andfiltered gold colloids were displayed in Figure 1C. It was foundthat the filtration step obviously reduced the population oflarge-sized AuNPs. This result further demonstrated the validityof SPRM to size AuNPs.We subsequently demonstrated the capability of SPRM to

size polydisperse gold colloids, which were produced byartificially mixing two colloids with average diameters(determined by TEM) of 20 and 40 nm (Figure S5),respectively. As shown in Figure 2A,B, pure samples of 20

and 40 nm gold colloids exhibited a well-defined single peakwith average SPRM intensities of 0.5 and 2.8 × 104 (arbitraryunit), respectively. Previous results demonstrated that SPRMsignal was proportional to the volume of nanoparticles.30,31 Themeasured SPRM signal ratio (∼6) was close to the volume ratio(∼8), which could be estimated from the cubic power ofdiameter ratio. The slight difference might be attributed to thedeviation of AuNPs diameters measured by TEM. For the sizedistribution of mixed sample shown in Figure 2C, thehistogram consisted of two distinguishable peaks with expectedSPRM intensity values further validating the feasibility of SPRMin sizing mixed AuNPs with different diameters. We alsodemonstrated the capability of SPRM to count the relativenumber of AuNPs with different sizes. It was found that thenumber ratio of different sized NPs matched well with theartificial concentration ratio (Figure S6). Therefore, webelieved there was no obvious size bias in our SPRM systemwhen distinguishing the mixed populations.

Melamine-Induced Aggregation of Gold Nanopar-ticles. Melamine-induced aggregation and adenosine triphos-phate (ATP)-induced disaggregation were selected to examinethe performance of SPRM in determining analyte concentrationby digitally imaging and counting AuNPs, respectively. Themelamine34−36 and ATP33,37,38 had been intensively studied inthe AuNPs-based colorimetric assays, so that a comprehensivecomparison with our work could be established.In melamine-induced aggregation, citrate-reduced colloidal

AuNPs remained monodispersed in original states. Thepresence of positively charged melamine led to partialaggregation of AuNPs due to the decreased interparticleelectrostatic repulsion. The melamine-induced aggregation ofAuNPs under the experimental conditions was also confirmedby UV−vis, DLS, and TEM measurements, respectively (FigureS1).Size distributions of diluted gold colloids in the presence of

varied melamine concentrations were determined by SPRM asshown in Figure 3A−G. It was clear that the percentage oflarge-sized AuNPs gradually increased with the increasingmelamine concentration, which was expected in an analyte-induced aggregation process. By choosing a threshold value asdetermined by the average SPRM value of over 200 individualnanoparticles plus three times the standard derivation(magenta-colored dash lines), the number percentages ofaggregates exhibited a linear relationship with the logarithm ofmelamine concentration as shown in Figure 3H. The lowestdetectable analyte concentration was 1 nM, which was 1 orderof magnitude lower than the literature reports that measuredthe same analyte with traditional colorimetric assays,33−35 aswell as other instrument-based detections, such as electro-chemistry,39,40 chemiluminescence,41 fluorescence,42 and liquidchromatography.43

The improved sensitivity was attributed to two factors. First,this digitized assay directly counted the number of goldaggregates with high accuracy, even though its population wasextremely low. This method avoided common noise sources inthe extinction measurements, for instance, the light intensityfluctuations in the optical system as well as moving mechanicalparts like gratings. Second, since less than 1000 individuals werecollected in a typical measurement, the concentration of AuNPsrequired in the present approach was orders of magnitudelower than that of the traditional assay, so that much highermolar ratio between analyte molecules and AuNPs could beachieved. As-synthesized gold colloids were usually diluted by

Figure 2. Size distributions of AuNPs with TEM diameter of 20 nm(A) and 40 nm (B), respectively. (C) Size distribution of polydispersed AuNPs by mixing an equal amount of 20 and 40 nm AuNPs.

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.5b04244Anal. Chem. 2016, 88, 2321−2326

2324

approximately 25 times prior to use. Under such low AuNPsconcentration, the extinction measurement with a commonUV−vis spectrometer was difficult (Figure S7). In order tocompensate for the reduced collision possibility betweenAuNPs with such a low concentration, the gold colloids wereincubated in the presence of analytes at 37 °C to enhance thereaction efficiency.ATP-Induced Disaggregation of Gold Nanoparticles.

Analyte-induced disaggregation or dispersion was anothercommonly used strategy in the colorimetric assay. In such acase, preformed stable aggregates collapsed to generate AuNPsmonomers again in the presence of analyte, accompanied with acolor recovery from blue or purple to red. Surface modificationof gold nanoparticles was always required to achieve an efficientdisaggregation.We chose the ATP-induced disaggregation of aptamer-

functionalized AuNPs as a model system in the present work.The target-free ATP aptamer was first hybridized with itscomplementary DNA sequence to form a rigid duplex. In thepresence of target ATP molecule, aptamers preferentiallybound to ATP, leading to the disassembly of original duplexand the release of a single stranded DNA (ss-DNA). Thereleased ss-DNA molecules thus adsorbed to AuNP, whichsignificantly enhanced its resistance to salt-induced aggregation.Therefore, when incubated with the same amount of salt, goldcolloids with higher ATP concentration tended to be morestable and kept the original size distribution.Size distributions of gold colloids with different ATP

concentrations were displayed in Figure 4A−H. It was foundthat, as expected, the number percentage of aggregatesdecreased with the logarithm of ATP concentrations in therange of 5 nM to 10 μM as shown in Figure 4I. The lowestdetectable concentration of 5 nM was significantly lower than

the previous reports using the colorimetric method for thedetection of ATP.33,37,38 This sensitivity level was alsocomparable with other quantitative measurements of ATP.44−47

■ CONCLUSIONSIn this work, we demonstrated a SPRM technique that couldinvestigate the size distribution of AuNPs by imaging andcounting individual AuNPs. Such capability was subsequentlyutilized to determine the analyte concentration by takingadvantage of analyte-induced AuNPs aggregation or dispersion.The quantification was based on measuring the numberpercentage of gold nanoparticle aggregates, leading to a digitalread-out of aggregation status of AuNPs. The digital read-outwas free of any fluctuations from the light source andmechanical system and thus exhibited greatly improvedsensitivity in both model systems examined in the presentwork. On the other hand, because SPRM measures theinfluence of nanoparticles to a propagating surface plasmonwave excited on the surface of gold film, any nanoparticles withdifferent refractive index from the buffer medium could bedetected. SPRM is able to study the analyte-inducedaggregation of many nonplasmonic nanoparticles with poorextinction properties, which can hardly be studied with thetraditional colorimetric assays. The digital read-out thus holds agreat potential to expand the applications of the colorimetricassay involving broad kinds of colorless nanoparticles.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.anal-chem.5b04244.

Experimental details, characterization of gold nano-particles, Fourier transformation of SPRM images,mechanism of colorimetric detection of ATP, dataanalysis, and descriptions of movies. (PDF)SPRM imaging of individual AuNPs and their intensities(AVI)

Figure 3. (A−G) Size distributions of AuNPs in the presence ofdifferent concentrations of melamine from 0 to 0.001, 0.005, 0.01,0.05, 0.1, and 1 μM, respectively. Any nanoparticles with SPRMintensity higher than the threshold set by the magenta dashed line areconsidered as aggregates. (H) The linear relationship between thenumber percentage of aggregates and the logarithms of the melamineconcentration.

Figure 4. (A−H) Size distributions of AuNPs in the presence ofdifferent concentrations of ATP from 0 to 0.005, 0.01, 0.1, 0.5, 1, 5,and 10 μM, respectively. Any nanoparticles with SPRM intensityhigher than the threshold set by the magenta dashed line areconsidered as aggregates. (I) The linear relationship between thenumber percentage of aggregates and the logarithms of ATPconcentrations.

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.5b04244Anal. Chem. 2016, 88, 2321−2326

2325

Differential SPRM images of AuNPs attach to the SPRchip surface for illustrating the “hit and move” collisionevent. (AVI)Differential SPRM images of AuNPs attach to the SPRchip surface for illustrating the “hit and stay” collisionevent. (AVI)

■ AUTHOR INFORMATIONCorresponding Author*Tel/Fax: +86-25-89682185. E-mail: wei.wang@nju.edu.cn.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSWe are thankful for the financial support from the NationalNatural Science Foundation of China (NSFC, Grant Nos.21522503, 21405080, 21327902, 21327008) and the NaturalScience Foundation of Jiangsu Province (BK20150013,BK20140592).

■ REFERENCES(1) Jain, P. K.; Huang, X.; El-Sayed, I. H.; El-Sayed, M. A. Acc. Chem.Res. 2008, 41, 1578−1586.(2) Stewart, M. E.; Anderton, C. R.; Thompson, L. B.; Maria, J.;Gray, S. K.; Rogers, J. A.; Nuzzo, R. G. Chem. Rev. 2008, 108, 494−521.(3) Elghanian, R.; Storhoff, J. J.; Mucic, R. C.; Letsinger, R. L.;Mirkin, C. A. Science 1997, 277, 1078−1081.(4) Vilela, D.; Gonzalez, M. C.; Escarpa, A. Anal. Chim. Acta 2012,751, 24−43.(5) Zhao, W.; Brook, M. A.; Li, Y. F. ChemBioChem 2008, 9, 2363−2371.(6) Du, J. J.; Jiang, L.; Shao, Q.; Liu, X. G.; Marks, R. S.; Ma, J.; Chen,X. D. Small 2013, 9, 1467−1481.(7) Ghosh, S. K.; Pal, T. Chem. Rev. 2007, 107, 4797−4862.(8) Daniel, M. C.; Astruc, D. Chem. Rev. 2004, 104, 293−346.(9) Xie, X. J.; Xu, W.; Liu, X. G. Acc. Chem. Res. 2012, 45, 1511−1520.(10) Xiao, L. H.; Wei, L.; He, Y.; Yeung, E. S. Anal. Chem. 2010, 82,6308−6314.(11) Yuan, Z. Q.; Cheng, J.; Cheng, X. D.; He, Y.; Yeung, E. S.Analyst 2012, 137, 2930−2932.(12) Fedoruk, M.; Lutich, A. A.; Feldmann, J. ACS Nano 2011, 5,7377−7382.(13) Wang, W.; Tao, N. J. Anal. Chem. 2014, 86, 2−14.(14) Shang, J.; Gao, X. H. Chem. Soc. Rev. 2014, 43, 7267−7278.(15) Guo, L. H.; Ferhan, A. R.; Chen, H. L.; Li, C. M.; Chen, G. N.;Hong, S.; Kim, D. H. Small 2013, 9, 234−240.(16) Xu, X.; Li, T.; Xu, Z. X.; Wei, H. J.; Lin, R. Y.; Xia, B.; Liu, F.; Li,N. Anal. Chem. 2015, 87, 2576−2581.(17) Xu, F. G.; Dong, C. Q.; Xie, C.; Ren, J. C. Chem. - Eur. J. 2010,16, 1010−1016.(18) Lan, T.; Wang, J. J.; Dong, C. Q.; Huang, X. Y.; Ren, J. C.Talanta 2015, 132, 698−704.(19) Zhu, S. B.; Yang, L. L.; Long, Y.; Gao, M.; Huang, T. X.; Hang,W.; Yan, X. M. J. Am. Chem. Soc. 2010, 132, 12176−12178.(20) Han, G. J.; Xing, Z.; Dong, Y. H.; Zhang, S. C.; Zhang, X. R.Angew. Chem., Int. Ed. 2011, 50, 3462−3465.(21) Han, G. J.; Zhang, S. C.; Xing, Z.; Zhang, X. R. Angew. Chem.,Int. Ed. 2013, 52, 1466−1471.(22) Zhang, S. X.; Han, G. J.; Xing, Z.; Zhang, S. C.; Zhang, X. R.Anal. Chem. 2014, 86, 3541−3547.(23) Huang, B.; Yu, F.; Zare, R. N. Anal. Chem. 2007, 79, 2979−2983.(24) Shan, X. N.; Patel, U.; Wang, S. P.; Iglesias, R.; Tao, N. J. Science2010, 327, 1363−1366.

(25) Wang, W.; Yang, Y. Z.; Wang, S. P.; Nagaraj, V. J.; Liu, Q.; Wu,J.; Tao, N. J. Nat. Chem. 2012, 4, 846−853.(26) Halpern, A. R.; Wood, J. B.; Wang, Y.; Corn, R. M. ACS Nano2014, 8, 1022−1030.(27) Cho, K.; Fasoli, J. B.; Yoshimatsu, K.; Shea, K. J.; Corn, R. M.Anal. Chem. 2015, 87, 4973−4979.(28) Shan, X. N.; Diez-Perez, I.; Wang, L. J.; Wiktor, P.; Gu, Y.;Zhang, L. H.; Wang, W.; Lu, J.; Wang, S. P.; Gong, Q. H.; Li, J. H.;Tao, N. J. Nat. Nanotechnol. 2012, 7, 668−672.(29) Shan, X. N.; Fang, Y. M.; Wang, S. P.; Guan, Y.; Chen, H. Y.;Tao, N. J. Nano Lett. 2014, 14, 4151−4157.(30) Fang, Y. M.; Wang, W.; Wo, X.; Luo, Y. S.; Yin, S. W.; Wang, Y.X.; Shan, X. N.; Tao, N. J. J. Am. Chem. Soc. 2014, 136, 12584−12587.(31) Wang, S. P.; Shan, X. N.; Patel, U.; Huang, X. P.; Lu, J.; Li, J. H.;Tao, N. J. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 16028−16032.(32) Syal, K.; Wang, W.; Shan, X. N.; Wang, S. P.; Chen, H. Y.; Tao,N. J. Biosens. Bioelectron. 2015, 63, 131−137.(33) Wang, J.; Wang, L. H.; Liu, X. F.; Liang, Z. Q.; Song, S. P.; Li,W. X.; Li, G. X.; Fan, C. H. Adv. Mater. 2007, 19, 3943−3946.(34) Qi, W. J.; Wu, D.; Ling, J.; Huang, C. Z. Chem. Commun. 2010,46, 4893−4895.(35) Kumar, N.; Seth, R.; Kumar, H. Anal. Biochem. 2014, 456, 43−49.(36) Liu, Y. L.; Yang, Y. M.; Zhao, X. J.; Liu, Z. D.; Li, Y. F. Anal.Methods 2013, 5, 3242−3247.(37) Liu, J. W.; Lu, Y. Angew. Chem., Int. Ed. 2006, 45, 90−94.(38) Huang, Y. F.; Chang, H. T. Anal. Chem. 2007, 79, 4852−4859.(39) Liu, Y. T.; Deng, J.; Xiao, X. L.; Ding, L.; Yuan, Y. L.; Li, H.; Li,X. T.; et al. Electrochim. Acta 2011, 56, 4595−4602.(40) Cao, Q.; Zhao, H.; Zeng, L.; Wang, J.; Wang, R.; Qiu, X.; He, Y.Talanta 2009, 80, 484−488.(41) Zeng, H. J.; Yang, R.; Wang, Q. W.; Li, J. J.; Qu, L. B. FoodChem. 2011, 127, 842−846.(42) Zhang, M.; Cao, X.; Li, H.; Guan, F.; Guo, J.; Shen, F.; Luo, Y.;Sun, C.; Zhang, L. Food Chem. 2012, 135, 1894−1900.(43) Batista, A. D.; Nascimento, C. F.; Melchert, W. R.; Rocha, F. R.P. Microchem. J. 2014, 117, 106−110.(44) He, Y.; Wang, Z. G.; Tang, H. W.; Pang, D. W. Biosens.Bioelectron. 2011, 29, 76−81.(45) Wang, L.; Xu, M.; Han, L.; Zhou, M.; Zhu, C. Z.; Dong, S. J.Anal. Chem. 2012, 84, 7301−7307.(46) Jie, G. F.; Yuan, J. X.; Zhang, J. Biosens. Bioelectron. 2012, 31,69−76.(47) He, X. P.; Wang, G. F.; Xu, G.; Zhu, Y. H.; Chen, L.; Zhang, X.J. Langmuir 2013, 29, 14328−14334.

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.5b04244Anal. Chem. 2016, 88, 2321−2326

2326