single-molecule kinetics reveals a hidden surface reaction...

Single-Molecule Kinetics Reveals a Hidden Surface ReactionIntermediate in Single-Nanoparticle CatalysisHao Shen,‡ Xiaochun Zhou,†,‡ Ningmu Zou, and Peng Chen*

Department of Chemistry and Chemical Biology, Cornell University, Ithaca, New York 14853, United States

*S Supporting Information

ABSTRACT: Detecting and characterizing reaction inter-mediates is not only important and powerful for elucidatingreaction mechanisms but also challenging in general because ofthe low populations of intermediates in a reaction mixture.Studying surface reaction intermediates in heterogeneouscatalysis presents additional challenges, especially the ubiq-uitous structural heterogeneity among the catalyst particles and the accompanying polydispersion in reaction kinetics. Here weuse single-molecule fluorescence microscopy to study two complementary types of Au nanocatalystsmesoporous-silica-coatedAu nanorods (i.e., Au@mSiO2 nanorods) and bare 5.3 nm pseudospherical Au nanoparticlesat the single-particle, single-turnover resolution in catalyzing the oxidative deacetylation of amplex red by H2O2, a synthetically relevant and increasinglyimportant probe reaction. For both nanocatalysts, the distributions of the microscopic reaction time from a single catalyst particleclearly reveal a kinetic intermediate, which is hidden when the data are averaged over many particles or only the time-averagedturnover rates are examined for a single particle. This intermediate is further resolvable by single-turnover kinetics at thesubparticle level. Detailed single-molecule kinetic analysis leads to a quantitative reaction mechanism and supports that theintermediate is likely a surface-adsorbed one-electron-oxidized amplex red radical. The quantitation of kinetic parameters furtherallows for the evaluation of the large reactivity inhomogeneity among the individual nanorods and pseudospherical nanoparticles,and for Au@mSiO2 nanorods, it uncovers their size-dependent reactivity in catalyzing the first one-electron oxidation of amplexred to the radical. Such single-particle, single-molecule kinetic studies are expected to be broadly useful for dissecting reactionkinetics and mechanisms.

1. INTRODUCTION

Observing and characterizing reaction intermediates representsa powerful approach in elucidating the mechanism of chemicaltransformations, such as those in enzyme catalysis, homoge-neous catalysis, and heterogeneous catalysis.1−8 But trappingand detecting reaction intermediates is difficult in generalbecause of their very nature of being intermediatestheirpopulations in a reaction mixture are low, and they exist onlytransiently. Synchronization of reaction progress (e.g., via thestopped-flow technique9,10) and trapping of reactive species(e.g., via rapid-freeze quench11,12 or chemical trapping13,14) areoften needed to obtain an appreciable amount of the desiredintermediate, so as to characterize it by spectroscopictechniques.In heterogeneous catalysis, studying reaction intermediates

presents additional challenges.1,2,5−7,15,16 Besides the generallow population of surface-adsorbed species, the ubiquitousheterogeneity among catalyst particles gives rise to poly-dispersion in reaction kinetics, making synchronization moredifficult. Yet it is important to know whether and how manyintermediates are involved in a surface catalytic reaction, so onecan determine the number of rate-limiting steps in the reactionand how one could possibly modify the catalyst to acceleratethe respective reaction steps for improving the catalystperformance.

Single-molecule kinetics provides advantages in dissectingreaction steps and uncovering reaction intermediates.17−25 Bystudying individual molecules/catalysts, this approach removesensemble averaging, so dispersion in kinetics among differentmolecules can be circumvented. Synchronization of molecularactions is not needed either, as it monitors one molecule at atime. It also allows for following the actions of individualmolecules in real time, and at any time point, only onemolecular state is present even if the molecule can adoptmultiple different states; this feature is particularly useful incapturing intermediates and elucidating reaction mechanisms.The capability of single-molecule kinetics in uncovering

intermediates and elucidating mechanisms has in particularbeen shown in single-molecule fluorescence studies of bio-logical molecules. For example, Lu et al. studied the catalysis bysingle cholesterol oxidases and observed an enzyme−substratecomplex as a kinetic intermediate.26 Others have observedkinetic intermediates in the catalysis by β-galactosidase,27 α-chymotrypsin,28 and nitrite reductase,22 in the DNA unwindingby nonstructural protein 3 (NS3),29 and in virus fusion.30

Single-molecule kinetics via fluorescence microscopy has alsobeen achieved in studying catalysis (or reactions) by layered

Received: September 19, 2014Revised: October 22, 2014Published: October 23, 2014

Article

pubs.acs.org/JPCC

© 2014 American Chemical Society 26902 dx.doi.org/10.1021/jp509507u | J. Phys. Chem. C 2014, 118, 26902−26911

pubs.acs.org/JPCC

double hydroxides,31 zeolites,32,33 metal nanoparticles,34−41

semiconductor nanocrystals,42−48 carbon nanotubes,49 andsmall molecules;50−53 but no surface reaction intermediateswere resolved in these studies. Other techniques, such as opticaltweezers,54,55 magnetic tweezers,56 single-channel-recording,57

and atomic force microscopy,58 are also powerful in studyingdynamic processes.We have used single-molecule microscopy to study single Au

nanoparticles and nanoplates in catalyzing a fluorogenicreductive deoxygenation reaction at the single-turnovertemporal resolution34−36 and later at nanometer spatialresolution as well.39 In this approach, the catalyzed reactionconverts a nonfluorescent reactant to a highly fluorescentproduct on the surface of the nanoparticles, which are dispersedat a low density and spatially separated on a slide. Undercontinuous laser excitation, each catalytically produced reactionproduct emits a large number of fluorescence photons, enablingits easy detection and imaging at the single-molecule level (aswell as its position localization down to nanometer precision).By imaging this fluorescent product one molecule at a time inreal time, one can follow the catalytic reactions on a singlenanoparticle at the single-turnover resolution in situ.We have further applied this single-molecule microscopy

approach to single Pt nanoparticles37 and single mesoporous-silica-coated Au nanorods (i.e., Au@mSiO2 nanorods),

38 usinga different fluorogenic probe reaction: the oxidative deacety-lation of amplex red (i.e., AR) by H2O2 (Figure 1A). This

deacetylation reaction has been an important probe reaction forassaying the activity of horseradish peroxidase and otherenzymes that scavenge reactive oxygen species59−65 and isincreasingly adopted by others as well in studying catalysis bysemiconductor nanostructures.47,48 Despite these studies andthe study66 on noncatalyzed photooxidation of AR wherepossible mechanistic steps and intermediates were proposed, noquantitative kinetic analysis is available, including the definitionof the rate-determining steps and quantitation of their rateconstants, on this important probe reaction.In our previous study of single Au@mSiO2 nanorods in

catalyzing the oxidative deacetylation of AR,38 we measured thetime-averaged catalytic turnover rate as a function of the

reactant concentration in a spatially resolved manner anddiscovered reactivity gradients along the side facets of individualnanorods, which were attributable to an underlying defectdensity gradient. Motivated by the increasing importance of thisreaction as a probe in surface catalysis and the general relevanceof deacetylation reaction to organic synthesis,67−69 we havebeen studying this reaction more mechanistically. Here wereport that single-molecule kinetics further uncovers a hiddenkinetic intermediate during the surface catalysis of ARdeacetylation on single Au@mSiO2 nanorods. We furtherdiscover that the same catalytic intermediate is resolved as wellon single bare 5.3 nm pseudospherical Au nanoparticles. To ourknowledge, these are the first examples of single-moleculekinetics in revealing surface reaction intermediates. Byanalyzing the distributions of the microscopic reaction timeof individual catalyst particles, we were able to resolve two rate-determining steps, quantify the rate constant of each step,provide evidence for the nature of the captured intermediate,and formulate a quantitative working mechanism for thecatalytic reaction.

2. EXPERIMENTAL SECTION

2.1. Synthesis and Characterization of Au Nano-catalysts. The 5.3 nm pseudospherical Au-nanoparticles,prepared from citrate reduction of HAuCl4 and capped withtannic acid, were purchased from Ted Pella (JME1052).Mesoporous silica-coated Au nanorods (i.e., Au@mSiO2nanorods) were prepared by making the Au nanorods firstvia stepwise seeded growth,70 coating them with silica, andsubsequently etching the silica shell with base in the presence ofcetyltrimethylammonium bromide to make it mesoporous,71 asreported previously (more details in Supporting Informa-tion).38 The Au@mSiO2 nanorods were later calcinated at 500°C to remove the organic ligands for activation for surfacecatalysis, during which the Au nanorod cores maintained theirmorphology as examined by transmission electron microscopy(TEM), as we reported previously.38 All samples werecharacterized by an FEI Tecnai 12 TEM at Cornell Centerfor Materials Research.

2.2. Catalytic Reaction Conditions. The oxidativedeacetylation of AR, a nonfluorescent molecule, to resorufin,a highly fluorescent molecule, by hydrogen peroxide (H2O2),was studied in 50 mM pH 7.3 degassed sodium phosphatebuffer. Ensemble reaction measurements were done viamonitoring the fluorescence of the reaction product resorufin(excited at 532 nm and detected at 585 nm) using a VarianCary Eclipse fluorometer.

2.3. Single-Molecule Fluorescence Microscopy. Single-molecule fluorescence measurements were performed on ahome-built prism-type total internal reflection fluorescence(TIRF) microscope based on an Olympus IX71 invertedmicroscope as previously described.34−37 A 5−6 mWcontinuous-wave circularly polarized 532 nm laser (CrystaLas-er, GCL-025-L-0.5%) was focused onto an area of ∼80 × 40μm2 in a microfluidic reactor cell to directly excite thefluorescence of the product resorufin generated on immobilizednanocatalysts. The fluorescence of resorufin was collected by a60× NA 1.2 water-immersion objective (UPLSAPO60XW,Olympus), filtered by two filters (HQ550LP, HQ580m60), andprojected onto a camera (Andor iXon EMCCD) controlled bythe Andor IQ software. The time resolution of imageacquisition was 25 ms.

Figure 1. (A) Balanced chemical equation of the Au-nanoparticle-catalyzed oxidative deacetylation of amplex red by H2O2, generatingresorufin, acetate, and water. (B, C) TEM images of Au@mSiO2nanorods (B) and 5.3 nm pseudospherical Au nanoparticles (C).

The Journal of Physical Chemistry C Article

dx.doi.org/10.1021/jp509507u | J. Phys. Chem. C 2014, 118, 26902−2691126903

The microfluidic reactor, ∼100 μm (height) × 2 cm (length)× 5 mm (width), was assembled using double-sided tapessandwiched between a quartz slide (Technical Glass) and aborosilicate coverslip (Gold Seal). The quartz slide was amine-functionalized by an aminoalkylsiloxane reagent (Vectabond,Vector Laboratory) to have a positively charged surface toimmobilize the negatively charged 5.3 nm Au nanoparticles. Asfor the Au@mSiO2 nanorods, they were drop casted on thequartz slide. The reactant solution was flowed in continuouslyat 25 μL/min, maintaining nonequilibrium steady-state reactionkinetics.2.4. Scanning Electron Microscopy. The microfluidic

reactor used for the fluorescence microscopy measurementswas subsequently disassembled for characterizing the nano-particles on the slide under a LEO 1550VP FESEM (operatedat 2−5 keV). A carbon film of ∼10 nm was coated on thequartz slide before the scanning electron microscopy (SEM)measurements.

3. RESULTS AND DISCUSSION3.1. Catalytic Reaction and Nanocatalysts. We studied

the catalyzed oxidative deacetylation of AR by H2O2 to generateresorufin and acetate (Figure 1A).37 The product resorufin ishighly fluorescent, allowing for its single-molecule fluorescenceimaging.We studied two types of Au nanocatalysts: mesoporous silica-

coated Au nanorods (i.e., Au@mSiO2 nanorods) and “bare”pseudospherical 5.3 ± 0.7 nm Au nanoparticles (Figure 1B,Cand Supporting Information, Figure S1; here bare refers to thefact that particles have weak surface ligands, e.g., citrate ortannic acid, which can be readily washed off).These two types of Au nanocatalysts are complementary

model catalysts. The Au nanorods are pseudo-one-dimensionalnanocatalysts with defined surface facets (e.g., their side facetsare {110} and {100}).72,73 The Au nanorod cores havemonodisperse diameters (21.4 ± 3.2 nm), but their lengths arevariable (from ∼100 nm to 600−700 nm);38 they are largeenough to be imaged readily under SEM (SupportingInformation, Figure S2) to be correlated with opticalmicroscopy using the same sample preparation (i.e., dispersedon quartz slides). The mSiO2 shell (∼80 nm in thickness) herewas necessary for removing through calcination the cappingligands that were used in synthesizing these nanorods whilemaintaining their morphology and preventing aggregation. Themesopores here are large enough for reactants to access thegold surface, and the catalytic kinetics is not limited by masstransport of reactants, as we reported.38,39 The mSiO2 shell alsohelps temporarily trap the product resorufin, facilitating itssingle-molecule fluorescence detection before it desorbs,diffuses into surrounding solution, and gets carried away bythe solution flow.On the other hand, the 5.3 nm pseudospherical Au

nanoparticles are pseudo-zero-dimensional nanocatalysts withmany types of surface facets, which are hard to define. Theirsmall sizes also make them harder to image by SEM. But, thesepseudospherical particles can be prepared without using strongcapping ligands, removing the need for the mSiO2 encapsula-tion.3.2. Single-Turnover Kinetics of Single Au@mSiO2

Nanorods Reveal a Reaction Intermediate. Figure 2Ashows an exemplary fluorescence intensity versus timetrajectory from a single immobilized Au@mSiO2 nanorod incatalyzing the oxidative deacetylation of AR to resorufin by

H2O2. The reactants were kept at constant concentrations andsupplied continuously via a flow, giving steady-state reactionkinetics. This trajectory features stochastic fluorescenceintensity bursts over a constant background emission of theAu nanorod; each burst reports the generation of a fluorescentproduct molecule resorufin. These bursts each typically last for∼50−100 ms (Supporting Information, Figure S3), before theproduct desorbs, diffuses out of the laser excitation volume, andgets carried away by the solution flow; the times (τ) betweenthe bursts are the microscopic reaction times for catalyticproduct generation. These τ values are probabilistic in theirindividual values, but their statistical properties, such asaverages and distributions, are defined by the underlyingreaction kinetics.74,75 Although AR can also be oxidized byH2O2 spontaneously, the contribution of this spontaneousreaction to the detected catalytic events is many orders ofmagnitude lower than that from the Au-surface catalyzedreaction, as we previously showed (Supporting Materials in ref38).Figure 2B−D shows the distributions of τ from a single Au@

mSiO2 nanorod at various AR concentrations (i.e., [AR]), while

Figure 2. Single Au@mSiO2 nanorod catalysis. (A) Fluorescenceintensity versus time trajectory of a single Au@mSiO2 nanorod incatalyzing the oxidative deacetylation of AR at [AR] = 0.05 μM and[H2O2] = 60 mM. (B, C, D) Distributions of τ from a single Au@mSiO2 nanorod at [AR] = 0.05 μM (B), 0.1 μM (C), and 0.5 μM (D).H2O2 concentration was kept at 60 mM. Solid red lines are empiricalfits with y = A(e−k1τ − e−k2τ) (eq 1). (E, F) Dependences of theapparent rate constant k1 and k2 on [AR] for the same Au@mSiO2nanorod in B−D (red triangles) and when averaged over 257nanorods (blue squares). Solid lines are fits with eqs 3 and 4 at thelimiting condition of [H] → ∞ , where ((GB

1/2[H]1/2)/(1+GB

1/2[H]1/2)) → 1. (G) Dependence of ⟨τ⟩−1 on [AR] for thesame Au@mSiO2 nanorod in B−D (red triangles) and when averagedover 257 nanorods (blue squares). Solid lines are fits with eq 5 at thelimiting condition of γeff2 ≫ γeff1. The red error bars in E and Findicate standard deviation, and the other error bars in E, F, and G areall standard error of the mean. Some of the error bars are smaller thanthe symbols.



the H2O2 concentration was kept at large excess. (The catalyticactivities of these Au@mSiO2 nanorods were sufficiently stableso that each nanorod could be studied over a range of [AR]over a period of many hours.38) Strikingly, the distributions of τshow an initial rise and then decay behavior with a delayedmaximum at τ > 0 (especially clear at [AR] = 0.05 μM, Figure2B), in contrast to the typical single-exponential decay behaviorfor single-molecule kinetics that contains just one rate-determining step.18,74,76−78 This initial rise and then decaybehavior of τ distributions indicates that the catalytic kinetics informing the fluorescent product contains at least two sequentialrate-determining steps; that is, there is a hidden kineticintermediate.21,26,74 Similar behaviors of τ distributions werealso observed in the single-molecule kinetics of some enzymereactions that contain kinetic intermediates.21,26−30 It isimportant to note that this initial rise and then decay behaviorof τ distributions could not be reliably discerned when τ valuesfrom many nanorods are compiled because of the poly-dispersion in kinetics among the individual nanorods(Supporting Information, Figure S4); therefore, even at thesingle-turnover resolution, measurements at the single-particlelevel were also essential in unmasking the kinetic signature ofthis intermediate.The distributions of τ from single Au@mSiO2 nanorods can

each be fitted by an empirical equation of a combination of twoexponentials:

= −τ τ− −y A e e( )k k1 2 (1)

where k1 and k2 are the apparent rate constants for the two rate-determining steps straddling the kinetic intermediate (see laterand Supporting Information, Section 4 for derivation of thisfunctional form of the τ distribution), and A is a scaling factor.Interestingly, in examining the results across a range of [AR],we find that k1 is dependent on [AR]: it increases withincreasing [AR] and eventually saturates (Figure 2E); thisdependence indicates that the associated rate-determining stepinvolves the reactant AR. In contrast, k2 is virtually independentof [AR] (Figure 2F), indicating that the associated rate-determining step does not involve AR. Moreover, k2 is ingeneral an order of magnitude larger than k1 across different[AR]. These trends of k1 and k2 versus [AR] persist, regardlessof whether individual Au@mSiO2 nanorods are examined or k1and k2 are averaged over many nanorods (Figure 2E,F).The single-turnover kinetics also readily gives the conven-

tional rate of turnovers (v, in s−1 particle−1) for a single Au@mSiO2 nanorod, which is equivalent to ⟨τ⟩

−1, where ⟨ ⟩ denotesaveraging. With increasing [AR], the single-particle rate ofturnovers ⟨τ⟩−1 shows typical saturation kinetics, regardless ofwhether the data are from a single Au@mSiO2 nanorod or areaveraged over many nanorods (Figure 2G). These saturationkinetics of ⟨τ⟩−1 alone can be well-accounted for by the classicLangmuir−Hinshelwood mechanism for surface catalysis79

involving surface-adsorbed AR molecules, without the need ofinvoking an intermediate in the catalytic kinetics, as wepreviously showed.34,36−38 In other words, the conventionaltitration of turnover rate versus reactant concentration isinsensitive here to the presence of this intermediate. Thisinsensitivity again highlights the capability of single-turnoverresolution kinetics to produce the distribution of the micro-scopic reaction time τ to unmask kinetic intermediates.3.3. Subparticle-Level Single-Turnover Kinetics Can

Still Reveal the Reaction Intermediate on a Au@mSiO2Nanorod. In imaging the fluorescence signal of the individual

reaction products on a Au@mSiO2 nanorod, the position ofeach product can be localized to a few to tens of nanometersprecision, as we showed previously.38 Figure 3A shows the

image plot of the two-dimensional position histogram of all thereaction products detected on the particular Au@mSiO2nanorod, whose single-turnover kinetics is presented in Figure2B−D. This image clearly resolves the nanorod shape at tens ofnanometers resolution, that is, superoptical resolution. Thisspatial resolution allows for analyzing the single-moleculekinetics of a nanorod at the subparticle level, for example, bydissecting the nanorod into two halves (Figure 3A). Byextracting the real-time sequence of the catalytic productformation events for each half, we obtained the distributions ofthe respective microscopic reaction time τ for the two halvesseparately: Figure 3B−D for the left half and Figure 3E,F forthe right half at various [AR]. (Note we could also dissect thenanorod into smaller segments, but with decreasing segmentsize, the number of catalytic events per segment also decreases,limiting the statistical significance of the distributions of τ.)The average τ for each half segment is about twice longer

than that for the whole nanorod (e.g., Figure 3B−D vs Figure2B−D). This is expected, as a half segment contains about halfof the number of catalytic events over the entire observationtime. Strikingly though, the distributions of τ for each half

Figure 3. Single-turnover kinetics of Au@mSiO2 nanorods atsubparticle resolution. (A) Image plot of the 2-dimensional positionhistogram of product molecules detected on the same nanorod shownin Figure 2B−D. The nanorod is oriented horizontally. The whitedotted line is the structural contour from its SEM image (SupportingInformation, Figure S2B); the red dotted line in the middle dissectsthe rod into two halves. (B−G) Distributions of τ extracted using onlythe catalytic events occurring on the left segment (B−D) or on theright segment (E−G) of the nanorod in A. H2O2 concentration waskept at 60 mM, while [AR] = 0.05 μM (B, E), 0.1 μM (C, F), and 0.5μM (D, G) as in Figure 2B−D. Solid lines are empirical fits with y =A(e−k1τ − e−k2τ) (eq 1).



segment still show clear initial rise and then decay behaviors,indicating the presence of the reaction intermediate. Therefore,even at the subparticle level, single-turnover kinetics can stillunmask reaction intermediates in surface catalysis.3.4. Single-Turnover Kinetics of Single 5.3 nm

Pseudospherical Bare Au Nanoparticles Again Revealsa Reaction Intermediate. To probe if the observed kineticintermediate is unique to Au@mSiO2 nanorods or is somehowrelated to the presence of its mSiO2 shell, we further studiedbare pseudospherical Au nanoparticles of 5.3 nm diameter.Figure 4A shows an exemplary fluorescence intensity versus

time trajectory from a single 5.3 nm Au nanoparticle incatalyzing the deacetylation of AR by H2O2. This trajectoryexhibits the characteristic intensity bursts reporting thegeneration of the fluorescent product molecule resorufin,similarly as in Figure 2A. The distributions of the microscopicreaction time τ from a single particle also show clear initial riseand then decay behaviors (Figure 4B−C), indicating thepresence of a hidden reaction intermediate straddled by tworate-determining steps in converting AR to the productresorufin. Again, this initial rise and then decay behavior ismasked when the τ values from many particles are compiledtogether (Supporting Information, Figure S5), further stressingthe importance of having both single-particle and single-turnover level resolution in measuring kinetics here.Compared with the Au@mSiO2 nanorods, these 5.3 nm Au

nanoparticles do not have the mSiO2 shell, and the product

resorufin is detected while it is temporarily adsorbed on theparticle surface. Therefore, the persistent observation of ahidden kinetic intermediate indicates that this intermediatestems from the catalytic process on the Au surface and isunrelated to the presence or not of an mSiO2 shell.The distribution of τ for a single 5.3 nm Au nanoparticle here

again can be satisfactorily fitted by eq 1, where k1 and k2 are theapparent rate constants of the two rate-determining steps(Figure 4B,C). We further studied many 5.3 nm Aunanoparticles across a range of [AR], while [H2O2] was keptat large excess. (Note that because of particle inactivation thatbecomes significant after 2 h, we could not study the same setof 5.3 nm Au nanoparticles over the entire range of [AR];instead, a different set of particles were studied at each [AR]within a <2 h time window.) Once averaged over the many Aunanoparticles, the extracted k1 shows a clear dependence on[AR]: it increases with increasing [AR] until saturation (Figure4D), whereas k2 is essentially independent of [AR] (Figure 4E);both trends are similar to those for Au@mSiO2 nanorods(Figure 2E,F).⟨τ⟩−1, the single-particle rate of turnovers, for the 5.3 nm Au

nanoparticles shows the typical saturation kinetics withincreasing [AR] for surface-mediated catalysis (Figure 4F),similar to that of Au@mSiO2 nanorods (Figure 2G). Again, thissaturation kinetics of ⟨τ⟩−1 is insensitive to the presence of thekinetic intermediate that is unmasked by the distribution of τ inFigure 4B,C.

3.5. Mechanism of Catalysis. The single-turnover kineticsof single Au@mSiO2 nanorods and single 5.3 nm Aunanoparticles above show that the surface-catalyzed oxidativedeacetylation of AR by H2O2 involves a kinetic intermediatestraddled by two rate-determining steps (Figure 2B−D andFigure 4B,C). The kinetics of one of the two steps is dependenton [AR], whereas the other is independent (Figure 2E,F andFigure 4D−E). The catalysis also directly involves surface-adsorbed AR molecules, as evidenced by the saturationbehavior of the single-particle turnover rate ⟨τ⟩−1 versus [AR](Figure 2G and Figure 4F).Unfortunately, titrating [H2O2] in single-molecule imaging of

single-nanoparticle catalysis is unreliable: the concentration ofdilute H2O2 solutions is unstable, but the single-nanoparticlecatalysis measurements need a long observation time (>1 h) toobserve a sufficient number of catalytic turnovers per particlefor accumulating statistics on τ. Nevertheless, we performedtitration of the catalytic reaction rate versus [H2O2] at theensemble level, using 5.3 nm pseudospherical Au nanoparticlesas a representative because they are more homogeneous in size(Figure 1C) (the Au@mSiO2 nanorod sample contains amixture of pseudospherical particles, triangular particles, andvariable-length nanorods, which are more problematic forensemble-averaged measurements; Figure 1B and SupportingInformation, Figure S1,C). The reaction rate shows saturationkinetics with increasing [H2O2] (Figure 5A), supporting thatthe catalysis involves adsorption of H2O2 onto the particlesurfaces.To account mechanistically for the observed single-turnover

catalytic kinetics of Au@mSiO2 nanorods and 5.3 nm Aunanoparticles, we first considered the following reactionsequence for the surface-catalyzed oxidative deacetylation ofAR to resorufin by H2O2 (Scheme1, and SupportingInformation, Section 4). This catalyzed reaction has overall a1:1 reaction stoichiometry between AR and H2O2, as we

Figure 4. Single 5.3 nm pseudospherical Au nanoparticle catalysis. (A)Fluorescence intensity versus time trajectory of a single 5.3 nmpseudospherical Au nanoparticle under catalysis at [AR] = 2 μM and[H2O2] = 60 mM. (B, C) Distributions of τ from two 5.3 nm Aunanoparticles at [AR] = 0.8 μM (B) and 2 μM (C). H2O2concentration was kept at 60 mM. Solid lines are fits with y =A(e−k1τ − e−k2τ) (eq 1). (D, E) Dependences of the apparent rateconstants k1 and k2 on [AR]; each data point is an average from ∼90nanoparticles. Solid lines are fits as those in Figure 2E,F. (F)Dependence of ⟨τ⟩−1 on [AR]; data averaged over ∼570 nanoparticles.[H2O2] was kept at 60 mM. Solid line is a fit as those in Figure 2G.Error bars in D−F are all standard error of the mean, and some ofthem are smaller than the symbols.



determined previously (Figure 1A),37 and for AR, it is overall atwo-electron oxidation reaction.In this overall reaction sequence (Scheme 1), the two

reactants first adsorb reversibly to the Au surface (step (i)).The adsorbed H2O2 can undergo a reversible hemolytic O−Obond cleavage to generate surface-bound OH• radicals (step(ii)). Then the reaction proceeds via a one-electron oxidationof AR by a surface-bound OH• radical, generating a radicalspecies AR• (step (iii)). A subsequent second electronoxidation generates the product resorufin (step (iv)). Thedeacetylation of AR could be accompanying the first or thesecond oxidation step, or occur as a separate step that is notexplicitly included here, and both or either step can be coupledwith proton transfer as well.It is known that Au surface can catalyze H2O2 decomposition

to form OH• radicals,80−85 and this reaction is reversible.85 Totest the involvement of OH• radical in the reaction pathway, weexamined the effect of dimethyl sulfoxide (DMSO), a knowneffective scavenger of OH• radicals,64,86−90 on the catalytickinetics at the ensemble level, again using 5.3 nm Aunanoparticles as the representative. The catalytic reaction rateis strongly quenched by DMSO in the solution (Figure 5B),consistent with the involvement of OH• radical in the reactionpathway. The involvement of the AR• radical in the oxidationof AR to form resorufin has been described in the catalysis bythe enzyme horseradish peroxidase.60−62

Our previous studies on small (≤15 nm diameter) bare Aunanoparticles in catalyzing a different reductive reaction34−36

and the current results on 5.3 nm Au nanoparticles show that

we could directly detect the fluorescence of the productresorufin on their surfaces. For Au@mSiO2 nanorods, thegenerated resorufin on the Au nanorod surface could potentiallybe fluorescently quenched due to the proximity to the Ausurface, before desorbing from Au nanorod surface and gettingtemporarily trapped in the mSiO2 shell. Regardless, thisfluorescence-quenched state of the product cannot be theobserved kinetic intermediate, as the intermediate is stillpresent for the bare 5.3 nm Au nanoparticles.In Scheme 1, step (i) can be assumed to follow a fast

adsorption−desorption equilibrium, in which AR and H2O2each follow a Langmuir adsorption behavior onto differenttypes of surface sites, which would be consistent with thesaturation kinetics observed with increasing reactant concen-trations. The two rate-determining steps observed experimen-tally should come from the remaining three reaction steps (i.e.,steps (ii), (iii), and (iv)), involving two possible reactionintermediates: the surface-adsorbed OH• radical or the surface-adsorbed one-electron-oxidized AR radical AR•.The OH• radical being the observed kinetic intermediate is

significantly disfavored for two reasons (see SupportingInformation, Section 4.2 for details): (1) On the basis of themeasurements of Lunsford et al.,85 the O−O bond cleavage rateof H2O2 to generate OH• on Au surfaces is estimated to be∼0.1 μmol cm−2 s−1. For Au@mSiO2 nanorods and 5.3 nm Aunanoparticles with average surface areas of 10−10 cm2 and 10−12

cm2, respectively, this O−O bond cleavage rate wouldcorrespond to 107 and 105 s−1 particle−1, at least 6 orders ofmagnitude larger than their observed highest turnover rates at∼2 and ∼0.04 s−1 particle−1 (Figure 2G and Figure 4F). (2)The apparent rate constants of the two rate-determining stepswould both be predicted to be dependent on [AR] (SupportingInformation, Section 4.2), which is in conflict with the results inFigure 2E,F and Figure 4D,E where only one of the rate-determining steps is dependent on [AR]. Therefore, the AR•

radical, not the OH• radical, is likely the kinetic intermediateobserved, and steps (iii) and (iv) in Scheme 1 are the two rate-determining steps with the apparent rate constants kapp1 andkapp2, respectively.Assuming the O−O bond cleavage of H2O2 to generate OH

•

establishes a fast equilibrium on Au surfaces, the reactionsequence in Scheme 1 reduces to the minimal kineticmechanism in Scheme 2. The probability density function

f(τ) of the microscopic reaction time τ for this simplified kineticscheme has been derived previously21,26,74 and is described indetail in Supporting Information, Section 4.

τ =−

−τ τ− −fk k

k ke e( ) ( )k kapp1 app2

app2 app1

app1 app2

(2)

f(τ) contains two exponentials, behaving with an initialexponential rise followed by an exponential decay, as observedexperimentally (Figure 2B−D and Figure 4B,C). And kapp1 andkapp2 here correspond to k1 and k2 in the empirical eq 1,respectively.Using classic Langmuir−Hinshelwood kinetics for surface

catalysis where AR and OH• adsorb onto different types ofsurface sites, we can derive that kapp1 and kapp2 take the

Figure 5. Ensemble reaction rates catalyzed by bare 5.3 nmpseudospherical Au nanoparticles. (A) Dependence on H2O2concentration at [AR] = 1 μM. Solid line is a fit with v =((kG1/2[H]1/2)/(1+G1/2[H]1/2)), which is equivalent to the Support-ing Information, Equation S16 (Supporting Information, Section 5) atsaturating [A]. [H] is the H2O2 concentration. (B) Dependence onDMSO concentration in the reaction solution at [AR] = 10 μM and[H2O2] = 50 μM. Solid line is a spline connection for visual guide. Allexperiments here were carried out in 50 mM pH 7.3 degassedphosphate buffer with ∼0.44 nM Au nanoparticles as the catalyst.Error bars are standard error (smaller than the symbols here).

Scheme 1. Possible Reaction Steps of Amplex Red Oxidationby H2O2 Catalyzed on Au Nanoparticle Surfacesa

aM: Au nanocatalyst; A: amplex red; H: H2O2; B: OH• radical; I:

intermediate AR•; P: fluorescent product resorufin. For example,MAmB2n represents a Au nanorod/nanoparticle having m adsorbedamplex red molecules and 2n OH•.

Scheme 2. Minimal Kinetic Mechanism



following forms (see derivations in Supporting Information,Section 4).

γ γ=+ +

⎯ →⎯⎯⎯⎯⎯⎯⎯+

→∞k

GG

GG

GG

[A]1 [A]

[H]1 [H]

[A]1 [A]app1 eff1

A

A

B1/2 1/2

B1/2 1/2

[H]eff1

A

A

(3)

γ γ=+

⎯ →⎯⎯⎯⎯⎯⎯→∞

kG

G[H]

1 [H]app2 eff2B

1/2 1/2

B1/2 1/2

[H]eff2

(4)

Here GA is the adsorption equilibrium constant of AR. GB is theequilibrium constant for H2O2 adsorption to directly becometwo adsorbed OH•. γeff1 = γ1nAnB, γeff2 = γ2nB, where γ1 and γ2are two rate constants and nA and nB are, respectively, the totalnumbers of adsorption sites for AR and OH• on a particle. [A](i.e., [AR]) and [H] are the solution concentrations of AR andH2O2, respectively. Moreover, when H2O2 in the solution goesto saturation concentrations (e.g., at 60 mM in our experi-ment), ((GB

1/2[H]1/2)/(1 + GB1/2[H]1/2)) → 1. Clearly, kapp1 is

dependent on [A] (the AR concentration)it increases withincreasing [A] and eventually saturateswhereas kapp2 isindependent of [A], as observed experimentally for the twoexponents in the distributions of τ (Figure 2E,F and Figure4D,E).⟨τ⟩−1, the time-averaged single-particle rate of turnovers,

takes the following form at saturating H2O2 concentrations:

τγ γ

=+ +

⎯ →⎯⎯⎯⎯⎯⎯⎯+γ

γ

γ γ− ≫

( )G

G

G

G

[A]

1 1 [A]

[A]

1 [A]1 eff1 A

A

eff1 A

Aeff1

eff2

eff2 eff1

(5)

With increasing AR concentration, ⟨τ⟩−1 is predicted to showthe typical saturation kinetics, consistent with experimentalresults (Figure 2G and Figure 4F), and be non-informativeabout the presence of the kinetic intermediate. The limitingform of eq 5 at γeff2 ≫ γeff1 is identical to the classic Langmuir−Hinshelwood rate equation for the case there is merely onerate-determining step and no kinetic intermediate is present,which we used previously.38

3.6. Inhomogeneity and Size Dependence of Nano-catalyst Reactivity. The formulation of the effective kineticmechanism in Scheme 2 and the corresponding expressions forf(τ), kapp1, and kapp2 in eqs 2−4 allowed us to fit theexperimental results to determine the kinetic parameters foreach catalyst particle. For each Au@mSiO2 nanorod, the datacollected at all [AR] were globally fitted (e.g., fitting thedistributions of τ in Figure 2B−D, or the k1 and k2 in Figure2E,F) to obtain its γeff1 (= γ1nAnB), γeff2 (= γ2nB), and GA. Thesenanorods are sufficiently large to be easily identifiable in theirSEM image (Supporting Information, Figure S2),38 from whichthe core length of each nanorod can be obtained using knownmSiO2 shell thickness (∼80 nm, Supporting Information,Section 1.3).38 For each 5.3 nm Au nanoparticle, its distributionof τ at an [AR] was fitted to obtain its γeff1, γeff2, and GA (e.g.,Figure 4B,C).In general, Au@mSiO2 nanorods are more active on a per

particle basis than the 5.3 Au nanoparticles, reflected by their∼20 times and ∼2 times larger average γeff1 and γeff2 values,respectively (Figure 6A,B vs E,F). It is not surprising, as thenanorods are much larger and thus have more surface sites perparticle. The distributions of γeff1, γeff2, and GA for the Au@mSiO2 nanorods are all broad (Figure 6A−C), reflecting thelarge reactivity inhomogeneity among the individual nanorods.This large reactivity inhomogeneity is also observed among the

individual 5.3 nm Au nanoparticles (Figure 6E−G). For theseinhomogenieties, single-nanoparticle catalysis measurementsare uniquely capable of quantifying them, for example, here bythe distributions of the respective kinetic parameters.For the individual Au@mSiO2 nanorods, the availability of

their lengths from SEM also allowed for examining their size-dependent catalytic properties. Their γeff1 (= γ1nAnB) valuerepresents the combined reactivity of all the sites on a singlenanorod in the first one-electron oxidation of AR to AR• byOH• (the first rate-determining step in Scheme 2); γeff1 shows aclearly positive correlation with the nanorod length (i.e., L;Figure 6A, left), attributable to a larger number (i.e., nA and nB)of reactive sites on longer nanorods. However, once normalizedby the nanorod length, γeff1/L decreases with increasing length(Figure 6D), suggesting that longer Au@mSiO2 nanorods havesmaller specific reactivity (i.e., γ1) for the reaction of ARoxidation to AR•. In contrast, γeff2 (= γ2nB), which representsthe combined reactivity of all sites on a single nanorod for thesecond one-electron oxidation of AR• to P (the second rate-determining step in Scheme 2), shows no clear dependence onthe nanorod length (Figure 6B, left), even though longernanorods presumably have more reactive sites of the nB type.

Figure 6. (A−C) (left) Scatter plots of γeff1, γeff2, and GA vs the corelength L for individual Au@mSiO2 nanorods. Each black crossrepresents one nanorod (total 257 nanorods). The nanorods are thenbinned every 25 nanorods according to their core length L andaveraged in each bin to obtain the dependences of γeff1, γeff2, and GA onL (red circles). (right) The corresponding histograms of γeff1, γeff2, andGA. The average values are ⟨γeff1⟩ = 1.20 ± 0.04 s−1, ⟨γeff2⟩ = 38.4 ± 4.7s−1, and ⟨GA⟩ = 9.3 ± 0.7 μM−1 (errors here are standard error of themean). (D) Dependence of γeff1/L on nanorod length L. Each point isan average of 25 nanorods. (E−G) Distributions of γeff1, γeff2, and GAfor 5.3 nm pseudospherical Au nanoparticles. Their average values are⟨γeff1⟩ = 0.068 ± 0.003 s−1, ⟨γeff2⟩ = 15.0 ± 0.5 s−1, and ⟨GA⟩ = 8.4 ±0.7 μM−1. x-direction error bars are standard deviation; y-directionerror bars are standard error of the mean. Some error bars are smallerthan the symbols.



This independence might result from opposite dependences ofγ2 and nB on nanorod length, thus canceling each other. GA,being the adsorption equilibrium constant of AR, does notshow any clear dependence on the nanorod length, either(Figure 6C, left).

4. CONCLUSION

Using single-molecule fluorescence microscopy, we havestudied the single-molecule kinetics of individual Au@mSiO2nanorods and bare 5.3 nm pseudospherical Au nanoparticles incatalyzing the oxidative deacetylation of AR to resorufin byH2O2, a synthetically relevant and increasingly important probereaction. For both nanocatalysts, the distributions of themicroscopic reaction time from a single catalyst particle, as wellas from part of a single particle, clearly reveal a kineticintermediate straddled by two rate-determining steps. Thissurface reaction intermediate is hidden when the data areaveraged over many particles or only the time-averagedturnover rates are examined for a single particle, demonstratingthe necessity of both single-turnover and single-particleresolution in unmasking this reaction intermediate. The kinetic(in)dependence of the two rate-determining steps on thereactant concentration allows for the formulation of a reactionmechanism that can describe quantitatively the catalytickinetics, which, in combination with literature results, supportsthat the intermediate is likely a surface-adsorbed one-electron-oxidized AR radical. The quantitative kinetics also gives kineticparameters of the reaction steps for each catalyst particle,allowing the evaluation of the large reactivity inhomogeneityamong the individual nanorods and pseudospherical nano-particles. Parallel SEM imaging further enables the correlationof each Au@mSiO2 nanorod’s reactivity with its length,uncovering its size-dependent reactivity in catalyzing the firstone-electron oxidation of AR to the radical. To our knowledge,this study represents the first examples of single-moleculekinetics in revealing surface reaction intermediates. We envisionthat such single-molecule kinetic studies should be broadlyuseful for dissecting reaction kinetics and mechanism inheterogeneous catalysis, where the catalyst heterogeneity is aubiquitous challenge for which ensemble-averaging could maskthe kinetic signatures of reaction intermediates.

■ ASSOCIATED CONTENT

*S Supporting InformationDetailed procedures of catalyst synthesis and structuralcharacterization. Procedures for extracting the microscopicreaction time τ from single-molecule fluorescence movies ofnanocatalyst catalysis. Additional results on the single-moleculekinetics of Au@mSiO2 nanorods and 5.3 nm pseudosphericalAu nanoparticles. Detailed formulation of possible reactionmechanisms, and the corresponding derivations of theprobability density functions of τ. This material is availablefree of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATION

Corresponding Author*E-mail: [email protected].

Present Address†Suzhou Institute of Nano-Tech and Nano-Bionics, ChineseAcademy of Sciences, 398 Ruoshui Road, Suzhou IndustrialPark, Suzhou, Jiangsu 215123, P. R. China.

Author Contributions‡These authors contributed equally.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis research was supported by the Department of Energy(DE-FG02-10ER16199) and in part by the National ScienceFoundation (CBET-1263736), the Army Research Office(63767-CH), and the Petroleum Research Fund (50966-ND10). Part of the work was performed at the Cornell Centerfor Materials Research (DMR-1120296) and Cornell NanoscaleFacility (ECS-0335765).

■ REFERENCES(1) Ertl, G.; Knozinger, H.; Weitkamp, J. Handbook of HeterogeneousCatalysis; VCH: Weinheim, Germany, 1997.(2) Somorjai, G. A.; Park, J. Y. Molecular surface chemistry by metalsingle crystals and nanoparticles from vacuum to high pressure. Chem.Soc. Rev. 2008, 37, 2155−2162.(3) Cook, P. F.; Cleland, W. W. Enzyme Kinetics and Mechanism;Garland Science: New York, 2007.(4) Smith, J. G.; Yang, Q.; Jain, P. K. Identification of a criticalintermediate in galvanic exchange reactions by single-nanoparticle-resolved kinetics. Angew. Chem., Int. Ed. 2014, 53, 2867−2872.(5) Savara, A.; Weitz, E. Elucidation of intermediates andmechanisms in heterogeneous catalysis using infrared spectroscopy.Annu. Rev. Phys. Chem. 2014, 65, 249−273.(6) Stavitski, E.; Weckhuysen, B. M. Infrared and raman imaging ofheterogeneous catalysts. Chem. Soc. Rev. 2010, 39, 4615−4625.(7) Hemelsoet, K.; Qian, Q.; De Meyer, T.; De Wispelaere, K.; DeSterck, B.; Weckhuysen, B. M.; Waroquier, M.; Van Speybroeck, V.Identification of intermediates in zeolite-catalyzed reactions by in situuv/vis microspectroscopy and a complementary set of molecularsimulations. Chemistry 2013, 19, 16595−16606.(8) Wunder, S.; Polzer, F.; Lu, Y.; Mei, Y.; Ballauff, M. Kineticanalysis of catalytic reduction of 4-nitrophenol by metallic nano-particles immobilized in spherical polyelectrolyte brushes. J. Phys.Chem. C 2010, 114, 8814−8820.(9) Jennings, P. A.; Wright, P. E. Formation of a molten globuleintermediate early in the kinetic folding pathway of apomyoglobin.Science 1993, 262, 892−896.(10) Khalifah, R. G. Carbon dioxide hydration activity of carbonicanhydrase 0.1. Stop-flow kinetic studies on native human isoenzyme-band isoenzyme-c. J. Biol. Chem. 1971, 246, 2561−2573.(11) Bollinger, J. M., Jr.; Edmondson, D. E.; Huynh, B. H.; Filley, J.;Norton, J. R.; Stubbe, J. Mechanism of assembly of the tyrosyl radical-dinuclear iron cluster cofactor of ribonucleotide reductase. Science1991, 253, 292.(12) Licht, S.; Gerfen, G. J.; Stubbe, J. A. Thiyl radicals inribonucleotide reductases. Science 1996, 271, 477−481.(13) Cavalcanti, F. A. P.; Blackmond, D. G.; Oukaci, R.; Sayari, A.;Erdem-Senatalar, A.; Wender, I. In situ chemical trapping of coh2surface species. J. Catal. 1988, 113, 1−12.(14) Deluzarche, A.; Hindermann, J. P.; Kiennemann, A.; Kieffer, R.Application of chemical trapping to the determination of surfacespecies and to the study of their evolution under reaction conditions inheterogeneous catalysis. J. Mol. Catal. 1985, 31, 225−250.(15) Zaera, F. Probing liquid/solid interfaces at the molecular level.Chem. Rev. 2012, 112, 2920−2986.(16) Bent, B. E. Mimicking aspects of heterogeneous catalysis:Generating, isolating, and reacting proposed surface intermediates onsingle crystals in vacuum. Chem. Rev. 1996, 96, 1361−1390.(17) Tamarat, P.; Maali, A.; Lounis, B.; Orrit, M. Ten years of single-molecule spectroscopy. J. Phys. Chem. A 1999, 104, 1−16.(18) Xie, X. S.; Lu, H. P. Single-molecule enzymology. J. Biol. Chem.1999, 274, 15967−15970.



http://pubs.acs.org

mailto:[email protected]

(19) De Cremer, G.; Sels, B. F.; De Vos, D. E.; Hofkens, J.; Roeffaers,M. B. J. Fluorescence micro(spectro)scopy as a tool to study catalyticmaterials in action. Chem. Soc. Rev. 2010, 39, 4703−4717.(20) Chen, P.; Zhou, X.; Shen, H.; Andoy, N. M.; Choudhary, E.;Han, K.-S.; Liu, G.; Meng, W. Single-molecule fluorescence imaging ofnanocatalytic processes. Chem. Soc. Rev. 2010, 39, 4560.(21) Floyd, D. L.; Harrison, S. C.; van Oijen, A. M. Analysis of kineticintermediates in single-particle dwell-time distributions. Biophys. J.2010, 99, 360−366.(22) Goldsmith, R. H.; Tabares, L. C.; Kostrz, D.; Dennison, C.;Aartsma, T. J.; Canters, G. W.; Moerner, W. E. Redox cycling andkinetic analysis of single molecules of solution-phase nitrite reductase.Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 17269−17274.(23) Cordes, T.; Blum, S. A. Opportunities and challenges in single-molecule and single-particle fluorescence microscopy for mechanisticstudies of chemical reactions. Nat. Chem. 2013, 5, 993−999.(24) Janssen, K. P. F.; De Cremer, G.; Neely, R. K.; Kubarev, A. V.;Van Loon, J.; Martens, J. A.; De Vos, D. E.; Roeffaers, M. B. J.;Hofkens, J. Single molecule methods for the study of catalysis: Fromenzymes to heterogeneous catalysts. Chem. Soc. Rev. 2014, 43, 990−1006.(25) Chen, P.; Zhou, X.; Andoy, N. M.; Han, K.-S.; Choudhary, E.;Zou, N.; Chen, G.; Shen, H. Spatiotemporal catalytic dynamics withinsingle nanocatalysts revealed by single-molecule microscopy. Chem.Soc. Rev. 2014, 43, 1107−1117.(26) Lu, H. P.; Xun, L.; Xie, X. S. Single-molecule enzymaticdynamics. Science 1998, 282, 1877−1882.(27) English, B. P.; Min, W.; van Oijen, A. M.; Lee, K. T.; Luo, G.;Sun, H.; Cherayil, B. J.; Kou, S. C.; Xie, X. S. Ever-fluctuating singleenzyme molecules: Michaelis-menten equation revisited. Nat. Chem.Biol. 2006, 2, 87−94.(28) Terentyeva, T. G.; Hofkens, J.; Komatsuzaki, T.; Blank, K.; Li,C.-B. Time-resolved single molecule fluorescence spectroscopy of anα-chymotrypsin catalyzed reaction. J. Phys. Chem. B 2013, 117, 1252−1260.(29) Myong, S.; Bruno, M. M.; Pyle, A. M.; Ha, T. Spring-loadedmechanism of DNA unwinding by hepatitis c virus ns3 helicase. Science2007, 317, 513−516.(30) Floyd, D. L.; Ragains, J. R.; Skehel, J. J.; Harrison, S. C.; vanOijen, A. M. Single-particle kinetics of influenza virus membranefusion. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 15382−15387.(31) Roeffaers, M. B. J.; Sels, B. F.; Uji-i, H.; De Schryver, F. C.;Jacobs, P. A.; De Vos, D. E.; Hofkens, J. Spatially resolved observationof crystal-face-dependent catalysis by single turnover counting. Nature2006, 439, 572−575.(32) Roeffaers, M. B. J.; De Cremer, G.; Libeert, J.; Ameloot, R.;Dedecker, P.; Bons, A.-J.; Buckins, M.; Martens, J. A.; Sels, B. F.; DeVos, D. E.; et al. Super-resolution reactivity mapping of nanostructuredcatalyst particles. Angew. Chem., Int. Ed. 2009, 48, 9285−9289.(33) De Cremer, G.; Roeffaers, M. B. J.; Bartholomeeusen, E.; Lin,K.; Dedecker, P.; Pescarmona, P. P.; Jacobs, P. A.; De Vos, D. E.;Hofkens, J.; Sels, B. F. High-resolution single-turnover mappingreveals intraparticle diffusion limitation in ti-mcm-41-catalyzedepoxidation. Angew. Chem., Int. Ed. 2010, 49, 908−911.(34) Xu, W.; Kong, J. S.; Yeh, Y.-T. E.; Chen, P. Single-moleculenanocatalysis reveals heterogeneous reaction pathways and catalyticdynamics. Nat. Mater. 2008, 7, 992−996.(35) Xu, W.; Kong, J. S.; Chen, P. Probing the catalytic activity andheterogeneity of au-nanoparticles at the single-molecule level. Phys.Chem. Chem. Phys. 2009, 11, 2767.(36) Zhou, X.; Xu, W.; Liu, G.; Panda, D.; Chen, P. Size-dependentcatalytic activity and dynamics of gold nanoparticles at the single-molecule level. J. Am. Chem. Soc. 2010, 132, 138−146.(37) Han, K. S.; Liu, G.; Zhou, X.; Medina, R. E.; Chen, P. How doesa single pt nanocatalyst behave in two different reactions? A single-molecule study. Nano Lett. 2012, 12, 1253−1259.(38) Zhou, X.; Andoy, N. M.; Liu, G.; Choudhary, E.; Han, K.-S.;Shen, H.; Chen, P. Quantitative super-resolution imaging uncovers

reactivity patterns on single nanocatalysts. Nat. Nanotechnol. 2012, 7,237−241.(39) Andoy, N. M.; Zhou, X.; Choudhary, E.; Shen, H.; Liu, G.;Chen, P. Single-molecule catalysis mapping quantifies site-specificactivity and uncovers radial activity gradient on single 2d nanocrystals.J. Am. Chem. Soc. 2013, 135, 1845−1852.(40) Zhou, X.; Choudhary, E.; Andoy, N. M.; Zou, N.; Chen, P.Scalable parallel screening of catalyst activity at the single-particle leveland subdiffraction resolution. ACS Catal. 2013, 3, 1448−1453.(41) Wee, T.-L.; Schmidt, L. C.; Scaiano, J. C. Photooxidation of 9-anthraldehyde catalyzed by gold nanoparticles: Solution and singlenanoparticle studies using fluorescence lifetime imaging. J. Phys. Chem.C 2012, 116, 24373−24379.(42) Naito, K.; Tachikawa, T.; Fujitsuka, M.; Majima, T. Real-timesingle-molecule imaging of the spatial and temporal distribution ofreactive oxygen species with fluorescent probes: Applications to tio2photocatalysts. J. Phys. Chem. C 2008, 112, 1048−1059.(43) Wang, N.; Tachikawa, T.; Majima, T. Single-molecule, single-particle observation of size-dependent photocatalytic activity in au/tio2 nanocomposites. Chem. Sci. 2011, 2, 891−900.(44) Tachikawa, T.; Yamashita, S.; Majima, T. Evidence for crystal-face-dependent tio2 photocatalysis from single-molecule imaging andkinetic analysis. J. Am. Chem. Soc. 2011, 133, 7197−7204.(45) Tachikawa, T.; Yonezawa, T.; Majima, T. Super-resolutionmapping of reactive sites on titania-based nanoparticles with water-soluble fluorogenic probes. ACS Nano 2012, 7, 263−275.(46) Bian, Z.; Tachikawa, T.; Kim, W.; Choi, W.; Majima, T. Superiorelectron transport and photocatalytic abilities of metal-nanoparticle-loaded tio2 superstructures. J. Phys. Chem. C 2012, 116, 25444−25453.(47) Xu, W.; Jain, P. K.; Beberwyck, B. J.; Alivisatos, A. P. Probingredox photocatalysis of trapped electrons and holes on single sb-dopedtitania nanorod surfaces. J. Am. Chem. Soc. 2012, 134, 3946−3949.(48) Ha, J. W.; Ruberu, T. P. A.; Han, R.; Dong, B.; Vela, J.; Fang, N.Super-resolution mapping of photogenerated electron and holeseparation in single metal−semiconductor nanocatalysts. J. Am.Chem. Soc. 2014, 136, 1398−1408.(49) Xu, W.; Shen, H.; Kim, Y. J.; Zhou, X.; Liu, G.; Park, J.; Chen, P.Single-molecule electrocatalysis by single-walled carbon nanotubes.Nano Lett. 2009, 9, 3968−3973.(50) Esfandiari, N. M.; Wang, Y.; Bass, J. Y.; Cornell, T. P.; Otte, D.A. L.; Cheng, M. H.; Hemminger, J. C.; Mcintire, T. M.; Mandelshtam,V. A.; Blum, S. A. Single-molecule imaging of platinum ligandexchange reaction reveals reactivity distribution. J. Am. Chem. Soc.2010, 132, 15167−15169.(51) Esfandiari, N. M.; Wang, Y.; Mcintire, T. M.; Blum, S. A. Real-time imaging of platinum-sulfur ligand exchange reactions at thesingle-molecule level via a general chemical technique. Organometallics2010, 30, 2901−2907.(52) Esfandiari, N. M.; Wang, Y.; Bass, J. Y.; Blum, S. A.Deconvoluting subensemble chemical reaction kinetics of platinum−sulfur ligand exchange detected with single-molecule fluorescencemicroscopy. Inorg. Chem. 2011, 50, 9201−9203.(53) Zhang, Y.; Song, P.; Fu, Q.; Ruan, M.; Xu, W. Single-moleculechemical reaction reveals molecular reaction kinetics and dynamics.Nat. Commun. 2014, 5, 4238.(54) Svoboda, K.; Mitra, P. P.; Block, S. M. Fluctuation analysis ofmotor protein movement and single enzyme-kinetics. Proc. Natl. Acad.Sci. U.S.A. 1994, 91, 11782−11786.(55) Dawid, A.; Croquette, V.; Grigoriev, M.; Heslot, F. Single-molecule study of ruvab-mediated holliday-junction migration. Proc.Natl. Acad. Sci. U.S.A. 2004, 101, 11611−11616.(56) Vilfan, I. D.; Lipfert, J.; Koster, D. A.; Lemay, S. G.; Dekker, N.H. Magnetic tweezers for single-molecule experiments; Springer: NewYork, 2009.(57) Howorka, S.; Movileanu, L.; Braha, O.; Bayley, H. Kinetics ofduplex formation for individual DNA strands within a single proteinnanopore. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 12996−13001.



(58) Strunz, T.; Oroszlan, K.; Schafer, R.; Guntherodt, H. J. Dynamicforce spectroscopy of single DNA molecules. Proc. Natl. Acad. Sci.U.S.A. 1999, 96, 11277−11282.(59) Haugland, R. P. The handbook: A guide to fluorescent probes andlabeling technologies; Invitrogen Corp, 2005.(60) Gorris, H. H.; Walt, D. R. Mechanistic aspects of horseradishperoxidase elucidated through single-molecule studies. J. Am. Chem.Soc. 2009, 131, 6277−6282.(61) Mishin, V.; Gray, J. P.; Heck, D. E.; Laskin, D. L.; Laskin, J. D.Application of the amplex red/horseradish peroxidase assay to measurehydrogen peroxide generation by recombinant microsomal enzymes.Free Radical Biol. Med. 2010, 48, 1485−1491.(62) Leon, D.; Marín-García, P.; Sanchez-Nogueiro, J.; Ortega de laO, F.; García-Carmona, F.; Miras-Portugal, M. T. P2x agonist bzatpinterferes with amplex-red-coupled fluorescence assays. Anal. Biochem.2007, 367, 140−142.(63) Reszka, K. J.; Wagner, B. A.; Burns, C. P.; Britigan, B. E. Effectsof peroxidase substrates on the amplex red/peroxidase assay:Antioxidant properties of anthracyclines. Anal. Biochem. 2005, 342,327−337.(64) Gomes, A.; Fernandes, E.; Lima, J. L. F. C. Fluorescence probesused for detection of reactive oxygen species. J. Biochem. Biophys. Meth.2005, 65, 45−80.(65) Rodrigues, J. V.; Gomes, C. M. Enhanced superoxide andhydrogen peroxide detection in biological assays. Free Radical Biol.Med. 2010, 49, 61−66.(66) Zhao, B.; Summers, F. A.; Mason, R. P. Photooxidation ofamplex red to resorufin: Implications of exposing the amplex red assayto light. Free Radical Biol. Med. 2012, 53, 1080−1087.(67) Kleeblatt, D.; Siyo, B.; Hein, M.; Iaroshenko, V. O.; Iqbal, J.;Villinger, A.; Langer, P. Synthesis of n,n′-diglycosylated isoindigos.Org.. Biomol. Chem. 2013, 11, 886−895.(68) Chyba, A.; Mastihuba, V.; Mastihubova, M. Synthesis of 4-nitrophenyl caffeate and its use in assays of caffeoyl esterases. Anal.Biochem. 2014, 445, 49−53.(69) Reddy, C. R.; Krishna, G.; Reddy, M. D. Synthesis of substituted3-furanoates from mbh-acetates of acetylenic aldehydes via tandemisomerization-deacetylation-cycloisomerization: Access to elliott’salcohol. Org. Biomol. Chem. 2014, 12, 1664−1670.(70) Jana, N. R.; Gearheart, L.; Murphy, C. J. Wet chemical synthesisof high aspect ratio cylindrical gold nanorods. J. Phys. Chem. B 2001,105, 4065−4067.(71) Stober, W.; Fink, A.; Bohn, E. Controlled growth ofmonodisperse silica spheres in the micron size range. J. ColloidInterface Sci. 1968, 26, 62−69.(72) Johnson, C. J.; Dujardin, E.; Davis, S. A.; Murphy, C. J.; Mann,S. Growth and form of gold nanorods prepared by seed-mediated,surfactant-directed synthesis. J. Mater. Chem. 2002, 12, 1765−1770.(73) Gai, P. L.; Harmer, M. A. Surface atomic defect structures andgrowth of gold nanorods. Nano Lett. 2002, 2, 771−774.(74) Xu, W.; Kong, J. S.; Chen, P. Single-molecule kinetic theory ofheterogeneous and enzyme catalysis. J. Phys. Chem. C 2009, 113,2393−2404.(75) Cao, J.; Silbey, R. J. Generic schemes for single-moleculekinetics. 1: Self-consistent pathway solutions for renewal processes. J.Phys. Chem. B 2008, 112, 12867−12880.(76) Zhuang, X.; Bartley, L. E.; Babcock, H. P.; Russell, R.; Ha, T.;Herschlag, D.; Chu, S. A single-molecule study of rna catalysis andfolding. Science 2000, 288, 2048−2051.(77) Witkoskie, J. B.; Cao, J. Single molecule kinetics. I. Theoreticalanalysis of indicators. J. Chem. Phys. 2004, 121, 6361−6372.(78) Gopich, I. V.; Szabo, A. Theory of the statistics of kinetictransitions with application to single-molecule enzyme catalysis. J.Chem. Phys. 2006, 124, 154712.(79) Satterfield, C. N. Heterogeneous Catalysis in Practice; McGraw-Hill: New York, 1980.(80) Zhang, Z.; Berg, A.; Levanon, H.; Fessenden, R. W.; Meisel, D.On the interactions of free radicals with gold nanoparticles. J. Am.Chem. Soc. 2003, 125, 7959−7963.

(81) Zhang, Z.-F.; Cui, H.; Lai, C.-Z.; Liu, L.-J. Gold nanoparticle-catalyzed luminol chemiluminescence and its analytical applications.Anal. Chem. 2005, 77, 3324−3329.(82) Corma, A.; Garcia, H. Supported gold nanoparticles as catalystsfor organic reactions. Chem. Soc. Rev. 2008, 37, 2096−2126.(83) Ionita, P.; Gilbert, B. C.; Chechik, V. Radical mechanism of aplace-exchange reaction of au nanoparticles. Angew. Chem. 2005, 117,3786−3788.(84) He, W.; Zhou, Y.-T.; Wamer, W. G.; Hu, X.; Wu, X.; Zheng, Z.;Boudreau, M. D.; Yin, J.-J. Intrinsic catalytic activity of au nanoparticleswith respect to hydrogen peroxide decomposition and superoxidescavenging. Biomaterials 2013, 34, 765−773.(85) Suh, M.; Bagus, P. S.; Pak, S.; Rosynek, M. P.; Lunsford, J. H.Reactions of hydroxyl radicals on titania, silica, alumina, and goldsurfaces. J. Phys. Chem. B 2000, 104, 2736−2742.(86) Baptista, L.; Clemente da Silva, E.; Arbilla, G. Oxidationmechanism of dimethyl sulfoxide (dmso) by oh radical in liquid phase.Phys. Chem. Chem. Phys. 2008, 10, 6867−6879.(87) Rao, P. S.; Luber, J. M., Jr; Milinowicz, J.; Lalezari, P.; Mueller,H. S. Specificity of oxygen radical scavengers and assessment of freeradical scavenger efficiency using luminol enhanced chemilumines-cence. Biochem. Biophys. Res. Commun. 1988, 150, 39−44.(88) Yamamoto, K.; Inoue, S.; Yamazaki, A.; Yoshinaga, T.;Kawanishi, S. Site-specific DNA damage induced by cobalt(ii) ionand hydrogen peroxide: Role of singlet oxygen. Chem. Res. Toxicol.1989, 2, 234−239.(89) Hanasaki, Y.; Ogawa, S.; Fukui, S. The correlation betweenactive oxygens scavenging and antioxidative effects of flavonoids. FreeRadical Biol. Med. 1994, 16, 845−850.(90) Tai, C.; Gu, X.; Zou, H.; Guo, Q. A new simple and sensitivefluorometric method for the determination of hydroxyl radical and itsapplication. Talanta 2002, 58, 661−667.



single-molecule kinetics reveals a hidden surface reaction...

Documents