Single-molecule fluorescence spectroscopy in (bio)catalysisMaarten B. J. Roeffaers*, Gert De Cremer*, Hiroshi Uji-i†, Benıot Muls‡, Bert F. Sels*, Pierre A. Jacobs*,Frans C. De Schryver†, Dirk E. De Vos*§, and Johan Hofkens†§

*Department of Microbial and Molecular Systems, Centre for Surface Chemistry and Catalysis, Katholieke Universiteit Leuven,Kasteelpark Arenberg 23, B-3001 Leuven, Belgium; †Department of Chemistry, Katholieke Universiteit Leuven, Celestijnenlaan200F, B-3001 Leuven, Belgium; and ‡Department of Chemistry, Universite Catholique de Louvain, Place L. Pasteur 1,B-1348 Louvain-la-Neuve, Belgium

Edited by Robert J. Silbey, Massachusetts Institute of Technology, Cambridge, MA, and approved May 15, 2007 (received for review December 5, 2006)

The ever-improving time and space resolution and molecular detection sensitivity of fluorescence microscopy offer unique opportuni-ties to deepen our insights into the function of chemical and biological catalysts. Because single-molecule microscopy allows forcounting the turnover events one by one, one can map the distribution of the catalytic activities of different sites in solid heteroge-neous catalysts, or one can study time-dependent activity fluctuations of individual sites in enzymes or chemical catalysts. Byexperimentally monitoring individuals rather than populations, the origin of complex behavior, e.g., in kinetics or in deactivationprocesses, can be successfully elucidated. Recent progress of temporal and spatial resolution in single-molecule fluorescence micros-copy is discussed in light of its impact on catalytic assays. Key concepts are illustrated regarding the use of fluorescent reporters incatalytic reactions. Future challenges comprising the integration of other techniques, such as diffraction, scanning probe, or vibra-tional methods in single-molecule fluorescence spectroscopy are suggested.

Single-molecule fluorescence spec-troscopy (SMFS) has recentlydeveloped into a powerful toolfor studying biophysical and bio-

chemical phenomena. In studies of enzy-matic catalysis, SMFS has revealed thatthere are large differences between thecatalytic activity of individual enzymeswithin a population (‘‘static disorder’’)and that the rate constant (kcat) of anindividual enzyme may strongly fluctu-ate over time (‘‘dynamic disorder’’), thelatter resulting from conformationalchanges of the enzyme. The recent ap-plication of SMFS to catalysis by solidmaterials has shown that heterogeneitiesin kcat also exist between individual cat-alytic crystals of one powder sample andeven between the sites of an individualcrystal. In this case, heterogeneity mightarise from different chemical environ-ments within the catalyst sample. Theobservation of heterogeneity in the kcatof those different catalytic systems sug-gests that the parallel introduction andevolution of SMFS techniques in bio-and chemocatalysis will deepen our in-sights in almost any type of catalyticconversion. Indeed, the challenge to de-rive overall kinetics from the contribu-tions of individuals within a populationis essentially the same for biological,heterogeneous and even homogeneoussystems, as discussed in From Popula-tions to Individuals.

From a technical viewpoint, SMFSrequires strongly fluorescent probe mol-ecules. The concepts to use such probesand even the probes themselves can beexchanged freely between heteroge-neous, homogeneous, and biocatalysis,as discussed in Probes for SMFS in(Bio)Catalysis. If one wants to map ineven more detail the contributions ofthe individual enzymes or catalytic sites

to the overall kinetics, further improve-ments of the spatial and temporal reso-lution of SMFS will be required (seeSpatial Resolution: Micro- and Nanoscopyand Time Resolution and Dynamics). Fi-nally, to complement the informationfrom SMFS experiments with structuraldata, an important long-term aim is tointegrate SMFS with other in situ tech-niques (see Integration of Techniquesand Perspectives).

From Populations to IndividualsNonuniformity is an inherent propertyof almost all catalytic systems. Even ifheterogeneities are masked in traditional(bio)-catalyst characterization by averag-

ing over an ensemble of entities, adeeper insight in the origin and natureof nonuniformities is necessary for thedesign of optimized (bio)catalyticsystems.

Going down from bulk to the level ofindividuals has profound implications onthe interpretation of the quantitativedata. In classical ensemble experiments,a catalytic activity is evaluated by mea-suring changes in concentrations. Whenlooking at single molecules, turnoversare stochastic events, and kinetics aredescribed by probabilities rather than byconcentration changes (1). When activ-ity is monitored at the single-moleculelevel for a population of identical cata-lytic sites, the probability of observing aturnover for an individual active sitewithin a short time interval is thereforesmall (Fig. 1). If one wants to evaluatesite diversity, it is therefore necessary toobserve not only a sufficiently largenumber of individual sites but also toobserve each individual site over a suffi-ciently long time.

Until now, single-molecule catalyticresearch has focused mainly on thestudy of individual enzymes. The resultshave proven the existence of dynamicdisorder, i.e., the fluctuation of activityof an individual enzyme over time,among the individuals of a seeminglyhomogeneous population (2–4). Dy-

Author contributions: B.M. contributed new analyticaltools; and M.B.J.R., G.D.C., H.U., B.F.S., P.A.J., F.C.D.S.,D.E.D.V., and J.H. wrote the paper.

The authors declare no conflict of interest.

Abbreviations: AFM, atomic force microscopy; SMFS, single-molecule fluorescence spectroscopy.

§To whom correspondence may be addressed. E-mail:dirk.devos@biw.kuleuven.be or johan.hofkens@chem.kuleuven.be.

© 2007 by The National Academy of Sciences of the USA

Fig. 1. Stochastic nature of turnover events. Thiswide-field image was recorded during hydrolysis offluorogenic fluorescein esters on a monolayer con-taining aminopropyl groups diluted by propylgroups.Theaminogroupdensity ismuchhigherthanthe spatial resolution of the fluorescence micro-scope. However, because each group is only sporad-ically active, the individual turnover events can beobserved as isolated bright spots. (Inset) Density ofaminopropyl groups (blue) versus propyl groups(gray) and the conversion of a nonfluorescent sub-strate (black) into a fluorescent product (red).

www.pnas.org�cgi�doi�10.1073�pnas.0610755104 PNAS � July 31, 2007 � vol. 104 � no. 31 � 12603–12609

SP

EC

IAL

FE

AT

UR

E:

PE

RS

PE

CT

IVE

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

0

namic disorder is generally explained asthe result of the thermodynamic equilib-rium between different conformationalstates (Fig. 2a). In this case, the ergodicprinciple holds: the time-averaged activ-ity of one individual is equal to the av-erage activity of the whole enzyme pop-ulation at a certain point in time. Whenthe disorder is of a static nature, theindividuals no longer display the sametime-averaged activity. Static disorder inenzymes can have a variety of origins,such as a difference of posttranslationalmodifications (5), different slowly inter-changing conformational substates, ordifferent local environments in immobi-lized conditions. To account for the dis-order inferred from observations of sin-gle turnovers, the traditional Michaelis–Menten model needs to be adapted.Therefore an extended two-dimensionalMichaelis–Menten model has been pro-posed by adding a thermodynamic com-ponent that describes the interconver-sion among the several conformationalsubstates of the individual enzymes (6).The thermodynamic and kinetic compo-nents of the interconversion may affecteach other through memory effects andsubstrate imprinting (7). With imprint-ing is meant the conformational changeinduced by the binding of the sub-strate that is retained after the enzy-matic cycle.

In the field of homogeneous chemicalcatalysis, stereoselectivity and catalyststability are pursued by designing li-gands that strictly define the coordinat-ing atoms and the conformation of the

metal complex. Nevertheless, in manycases, the homogeneous catalytic mixcontains chemically distinct metal com-plexes, and each of these subpopulationsreacts with different chemo-, regio-, orenantioselectivity. In Wilkinson’s mecha-nism for Rh-catalyzed hydroformylation,HRh(CO)2L2 leads to linear aldehydes,whereas more branched products areformed with HRh(CO)2L (L � triph-enylphosphine) (8–10). In Sharpless’sligand-accelerated osmium catalysis, theenantioselectivity is controlled by dy-namic ligand association equilibria be-tween OsO4 and OsO4�L species (L �e.g., dihydroquinine) (11, 12) (Fig. 2b).Such activity variations may be causedby dynamic disorder but also by staticdisorder, for instance, when the ligandsof an enantioselective epoxidation cata-lyst are irreversibly damaged, resultingin a metal species that catalyzes the ep-oxidation with lower or even no enan-tiomeric excess (13). Dissolved singlemetal complexes, if f luorescent, can bestudied in fluorescence correlation spec-troscopy,¶ or by isolation in femtoliterreactor chambers that contain a discretenumber of complexes (0, 1, . . . ) (14).The latter approach was illustrated inthe study of inorganic redox catalysis bya discrete number of OsO4 molecules,which catalyze the formation of emissiveCe3� (15). Alternatively homogeneous

catalysts can be observed after isolationon a surface, following the concepts ofsurface organometallic chemistry (16,17). Even reversible metal ion complex-ation dynamics can now be monitoredby on/off switching of a fluorescentprobe (18), which is an excellent startingpoint for further SMFS studies in homo-geneous catalysis. However, surface immobil-ization may create additional staticdisorder.

In heterogeneous catalysts, nonunifor-mity is intrinsically generated in unitoperations such as hydrothermalsynthesis, precipitation, catalyst activa-tion, or formulation. In oxides, hydrox-ides, metals, etc., elemental composi-tions and structural properties such asporosity or lattice structures vary overdifferent length scales. Both industriallymanufactured and research samples ofzeolites contain crystals with differentdimensions, habitus, and degrees of in-tergrowth (Fig. 2c). The concept ‘‘popu-lation’’ then not only relates to the col-lection of catalytic crystals but also tothe individual active sites within onecrystal. High-resolution surface studies,even on single-crystal model catalysts,show that undercoordinated featuressuch as edges, kinks, and defects arepreferred reaction sites (19, 20). Evenon well defined single crystal surfaces,such as Pt(110) or Pd(100), catalytic re-action induces dynamic reconstructions,resulting in coexistence of different sur-face structures (21).

Probes for SMFS in (Bio)CatalysisThe potential of an analytical techniquein studying catalyst nonuniformity iscritically determined by its ability tospecifically probe molecular species andto extract information on the local envi-ronment. In this section, we will give anoverview of the different approaches onhow probe molecules can be used toexplore bio- as well as chemocatalysts.

SMFS, as such, can easily distinguishbetween reactants and products of awell chosen catalytic reaction, but thetechnique is not restricted to merely de-tecting molecules based on photon fluxor intensity. In multiparameter fluores-cence detection, other intrinsic chro-mophore properties can be probed, suchas the emission spectrum, the fluores-cence lifetime, and the fluorescenceanisotropy (22). Although counting ofstochastic turnovers is typically based onintensity f luctuations or changes, theother chromophore properties containsupporting information for the physico-chemical interpretation of these turn-over rates.

A vast amount of probes that wereoriginally designed for biological re-search can readily be integrated in dif-

¶T. Dertinger, I. Gregor, I. von der Hocht, R. Erdmann, B.Kraemer, F. Koberling, R. Hartmann, and J. Enderlein(2006) Progress in Biomedical Optics and Imaging. Pro-ceedings of SPIE 6092:609203.

a

O

OOs

O

O

O

OOs

OO

O R

O

OOs

O

OL

1

23

corner site

edge site

4

cb

Fig. 2. Disorder in catalysts. (a) Interconverting enzyme conformations with different activities/selectivities. (b) Some of the Os(VIII) species that can osmylate olefins during asymmetric dihydroxylation.(c) Within a population of crystals, one can distinguish different crystal habitus (1 versus 2), differentdegrees of intergrowth (3), and on the crystal planes of individual crystals, different sites (4).

12604 � www.pnas.org�cgi�doi�10.1073�pnas.0610755104 Roeffaers et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

0

ferent schemes in (bio)catalytic research.First, a fluorogenic substrate can beconverted to an emissive product, as inthe hydrolysis of f luorescein diacetate tofluorescein. This reaction has been ap-plied to visualizing single turnovers onisolated lipases or to visualizing varia-tions in activity between different facetsof inorganic catalytic crystals (23–25).Essentially the same approach has beenimplemented at the ensemble level forultrasensitive high-throughput catalystscreening (26, 27). A sufficient batho-chromic or hypsochromic shift can alsobe induced by formation or breakdownof conjugated systems, e.g., in homoge-neously catalyzed Heck reactions (26),or in biocatalytic lipid peroxidation (28).Catalytic cleavage of a covalently bound(tethered) quenching group also allowsselective detection of a single-productmolecule against a background of excessreactant (29) (Fig. 3a).

Such experiments can be refined be-yond simple turnover counting by add-ing a dimension of chemoselectivity.Substrate regioselectivity can be studiedby designing probes in which cleavableand quenched chromophores are incor-porated at different loci. Thus, commer-cial probes for phospholipase activityare available, from which, e.g., bodipychromophores can be released from thephospholipid backbone by cleavage at

the sn-1 or at the sn-2 position (30, 31).Product isomer distributions can beprobed locally if the chromophore prop-erties depend on the substitution pat-tern. For instance, when a monosubsti-tuted aromatic compound is vinylated,o-, m-, and p-isomers of the conjugatedproduct may be formed, each with dis-tinct chromophore properties (lifetimes,spectra, etc.).

In a second widely applicable approach,a fluorescent catalyst is designed in whichthe fluorescence of a reporter group isswitched on and off during each catalyticcycle (Fig. 3b). The early study on choles-terol oxidase by Lu et al. (7) uses the on/off cycle of fluorescent FAD and nonfluo-rescent FADH2 as an inherent reporter.In an extension of this scheme, the fluo-rescence of a reporter chromophore inclose proximity of a catalytic metal centercan be switched on and off by changes of,e.g., the metal valence state or by revers-ible coordination of a ligand to the metalcenter (see above and refs. 18, 32, and33). The latter idea is already used in thedesign of molecular switches and of fluo-rescent, cryptand-based indicators forNa�, Ca2�, etc. (34).

In a similar manner, single moleculefluorescence studies provide a toolboxfor monitoring a broad range of catalystproperties, particularly in the study ofsolid materials. For instance, local ac-

idobasicity can be followed with acidicor basic probes (35), which are commer-cially available in a large variety (36)(Fig. 3c). The acidobasicity of the com-plete catalyst particle can be mappedafter an aliquot of the probe has beenequilibrated with the sample, or thechromophore properties can be followedwhile a single probe is walking over sur-faces or through channels by diffusion(Fig. 3c). Analogous probes can be de-vised for locally reporting redox states,e.g., in a mixed-oxide catalyst, forprobing electron-donating or electron-accepting properties of solid surfaces orfor locating freely available coordinationsites for incoming reagents. Manyprobes are known to change their spec-tra, lifetime, or intensity upon changesof the polarity or because of hydrogenbond formation. Finally, use of func-tional probes with varying dimensionsallows the study of the accessibility ofenzyme pockets or the access to acid/base functions, metal nanoclusters, etc.,in heterogeneous catalysts (37).

Spatial Resolution: Micro-and NanoscopyThe characteristic lengths for catalystsand catalytic phenomena range from �1nm to the macroscopic scale. The nano-meter is the fundamental length scale toexpress the size of crystallites in welldispersed supported noble metal cata-lysts, of cavities in zeolites and metal-organic frameworks, of ligands in coor-dination complexes, or of active centersof enzymes, which implies that propercharacterization techniques should beable to monitor ultralow quantities,preferably with single-molecule sensitiv-ity, with a sufficiently high spatial reso-lution. Because optical resolution is clas-sically limited by diffraction, theresolving power is of the same order ofmagnitude as the wavelength used. Sub-nanometer wavelength techniques, likeelectron microscopy, generally operateunder ex situ conditions and lack themolecular specificity of SMFS. AlthoughSMFS can operate in condensed phaseand has the desired single-molecule sen-sitivity, it makes use of longer, visiblewavelengths, which results in a resolu-tion of �500 nm in the focal plane and�1 �m along the optical axis. The con-focal approach facilitates three-dimen-sional imaging, but it improves the reso-lution only with a factor of �2. Notethat the diffraction limit does not im-pede colocalization of spectrally distinctmolecules. However, the crucial chal-lenge in catalytic systems will be to spa-tially resolve spectrally identical probemolecules interacting with the catalyst,on length scales that are much �200nm. We will now consider technical evo-

F Q F

Q

FG F

Redox cycling

eT / ET no eT / ET

F (quenched) F (bright)

X

X X

X

M+2

X

X X

X

M+3

a b

ci

ii 0 ms 100 ms

200 ms 300 ms

x-position / nm-200 2000

200

-200

0

y-position/nm

0 650 ms

N

O

O

OH

O

Addition of PMI-COOH ∆t

5 s 90 s

2 " m

Fig. 3. Fluorescence-based visualization of catalytic sites and events. (a) A fluorescent product (F) isformed by transformation of a fluorogenic reactant (FG) or by cleavage of a covalently bound quencher(Q). (b) Cycling of a metal catalyst (M) between two redox states causes quenching and dequenching ofa fluorescent reporter (F). (c) Mapping of basic sites on a layered double hydroxide crystal with acid probes:(i) Imaging of time-dependent sorption of a perylene monoimide carboxylic acid. Even at the ensemblelevel differences between crystal faces can be distinguished (ii). At the single-molecule level, the probehops between individual basic sites.

Roeffaers et al. PNAS � July 31, 2007 � vol. 104 � no. 31 � 12605

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

0

lutions that are expected to be useful instudying structured catalysts.

Although in far-field microscopy thelight needs to be focused, this focusingprocess is no longer necessary in thenear-field variant, where ultrasharp tipsare used to scan surfaces (38), whichallows the localization of objects at sub-diffraction resolution. This approach isvaluable to study two-dimensional struc-tures, but the depth profile obtained isfar more limited than in far-fieldmethods.

A major breach through the diffrac-tion limit was achieved with a nonlinearprocess, STED (stimulated emission de-pletion microscopy). In STED, the satu-rated depletion of the excited state isused to reach macromolecular-scale res-olutions, down to 15–20 nm in the focalplane. The excitation spot is overlappedwith a doughnut-shaped beam; this sec-ond light beam serves as a deexcitationbeam forcing the excited molecules backto the ground state, resulting in the flu-orescence from a very small centralnode not covered by the depletionbeam. Oversaturating the deexcitationsqueezes the fluorescence spot to sub-diffraction dimensions (39, 40). Usingother reversible saturation processessuch as photoswitching can contribute toimproving spatial resolution, whereastime resolution is another matter ofconcern. This approach can, for exam-ple, be used to reduce the area overwhich turnovers are counted.

A recent development in wide-fieldfluorescence imaging is to combinestructured illumination with a nonlinearfluorescence response. Essentially, ahigh-intensity, sinusoidally patternedillumination results in saturation of thefluorescent molecules in the image ex-

cept for those in the small nodes of thestructured illumination. When scanningthe sample, the emission intensity onlychanges in the zero-intensity regions, soby extracting spatial high-frequencycomponents using Fourier analysis, su-perresolution is realized. As a proof ofconcept, individual 50-nm fluorescentbeads were distinguished within aggre-gates (41).

In traditional, linear wide-field fluo-rescence microscopy, the location of asingle emitter can be determined to al-most arbitrarily high accuracy if a suffi-cient signal-to-noise ratio (SNR) can beachieved. By recording the point spreadfunction (PSF) of a single emitter andlocating its center by a two-dimensionalGaussian fit, resolutions down to thenanometer scale have been shown (42,43). However, when multiple emittersare in close proximity, these resolutionscannot be obtained. Recently, a few so-lutions have been proposed. If nonre-solved probe molecules are subject tostepwise bleaching, their relative posi-tions can easily be traced back by con-secutive fits of the PSFs (44). Anotherscheme makes use of the random on/offphotoswitching of a fraction of thepresent molecules, which, combined withreconstruction, yields nanometer re-solved images (45, 46). Such a schemecan easily be implemented in the obser-vation of single fluorescent products onthe surface of a heterogeneous catalyst.Because reaction at the different sites isa ‘‘stochastic’’ process, recording consec-utive frames should allow determiningthe activity of surface areas down to afew square nanometers (Fig. 4). Therequired SNR can be obtained by usinghigh excitation powers because bleach-ing is not an issue. One can think of

visualizing differences in activity gov-erned by surface cracks or local defects,or one can now look at nanometer-sizedcrystals.

Next to direct imaging of objectsthrough their f luorescence intensity, in-direct methods have been used to studyprocesses at the subnanometer to 10-nmlength scale. Fluorescence resonanceenergy transfer (FRET) between a fluo-rescent donor and acceptor is routinelyused to quantify interactions �1–10 nm;even shorter distances, �1 nm, can beprobed by electron transfer between afluorescent molecule and donors or ac-ceptors in its proximity. By using FRET,folding of RNA during RNase activityhas been monitored (47). Electrontransfer from tyrosine residues to anisoalloxazine has been used to studyconformational dynamics in a flavinereductase at angstrom scale (48). Energyand electron transfer phenomena canalso be exploited to study the distancesbetween nanoscale metal particles andthe acid sites of the support in bifunc-tional heterogeneous catalysts. Thesephenomena could provide a direct ex-perimental verification of Weisz’s inti-macy criterion, which for hydrocrackingreactions empirically describe the maxi-mum distance between reaction sites for(de)hydrogenation and acid-catalyzedreactions such as isomerization andcracking (49, 50).

Time Resolution and DynamicsAlthough traditional characterizationtechniques mainly determine staticproperties of catalyst populations, prob-ing time-related dynamics of individualsis required for deconvoluting collectivereaction rates into contributions of spa-tially distinguishable, catalytically com-petent subpopulations. The time scalesof interest to a catalytic site can spanthe whole catalyst lifetime, in the studyof catalyst activation or deactivation, ormay be shorter than a picosecond ifshort-living reaction intermediates areto be probed. In many cases SMFS al-ready can give an experimentally justi-fied statistical basis to the kinetics ofcatalyzed processes.

Single-molecule techniques are highlyadvantageous for following the courseof a catalytic process because there is noneed to synchronize the events at indi-vidual reaction sites within the popula-tion. After statistical analysis of thewhole population, the individual compo-nents can be assigned to specific sub-populations, which allows study of cata-lyst populations in which the subgroupsare in a dynamic equilibrium with oneanother (see above). Such informationwould not be accessible by traditionalbulk experiments because of averaging.

Fig. 4. High-resolution reconstruction of the active-site distribution in a catalyst. The individual framesshow stochastic turnover events; superimposition of consecutively recorded frames yields the high-resolution image as in photoactivation high-resolution light microscopy (45).

12606 � www.pnas.org�cgi�doi�10.1073�pnas.0610755104 Roeffaers et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

0

For counting individual turnovers, thetime resolution achieved in SMFS, e.g.,10�4 s, is sufficient for most catalyticsystems. Even for the most active homo-geneous catalysts, turnover frequencies(TOFs) seldom exceed 10 s�1; onlysome enzymes, such as catalase (80,000s�1), display TOFs beyond the time res-olution of current SMFS. To followturnovers at an individual site over asufficiently long time, it is preferable tomonitor products of a fluorogenic reac-tion. Other approaches, for instance,using a chromophore with switching flu-orescence in a ligand or cofactor, mayface the problem of photobleaching,which would prevent observation of suf-ficiently long time traces. Observation ofthe dynamic disorder in single enzymesover long time periods has shown thatthe enzymatic TOF may fluctuate overtime scales between 10�3 s and 10 s (3,23, 24, 51, 52). Although turnovercounting by SMFS has not yet beenstudied extensively in heterogeneouscatalysis, it has been used to prove staticdisorder on either two-dimensionally orthree-dimensionally organized catalysts(25, 53).

Time-resolved imaging combined witha high degree of spatial resolution ful-fills the requirements for investigatingdiffusion of single molecules in pores ofindividual catalyst particles (54–56).Even if the channel dimensions in mi-cro- or mesoporous materials are wellbelow the diffraction limit, useful infor-mation can be retrieved by analyzing thediffusion trajectories of individual mole-cules or by studying correlation in a sin-gle-point fluorescence time transient(57). For a terrylenediimide dye diffus-ing in a mesostructured molecular sieve,it was possible to prove the existence oftwo subpopulations, one containing mo-bile dye molecules, and a minor fractionof stationary molecules, which could beassigned to ‘‘dead ends,’’ e.g., in col-lapsed pores, or to strong sorption sites(55). Exchanges between the subpopula-tions are evidenced by observation of asingle molecule that becomes mobileagain after a stationary period. In mi-

croporous materials such as zeolites, inwhich (sub)nanometer pores cross mi-crometer-sized crystals, information ondiffusion by SMFS will be invaluable toan understanding of the role of diffu-sion barriers, either as fault planes or aslocal obstructions of pores by synthesisdebris or collapse (Fig. 5). For hierar-chical, bimodal porous materials, whichideally combine improved transportproperties in mesopores with shape se-lectivity in micropores, the dynamictransitions of the molecules betweenboth pore systems should be easy to in-vestigate with SMFS. The results needto be confronted with data from othertechniques, such as macroscopic uptakeor pulsed-field gradient NMR experi-ments, which probe diffusion over vary-ing time and length scales (58).

Relative motions of biocatalytic cen-ters and reagents are crucial as wellwhen the enzyme is embedded in astructure, e.g., a natural or artificialmembrane, or in an inorganic or poly-meric host used for enzyme immobiliza-tion. Processive movements of enzymesover substrates have been observed onflat substrates, as for phospholipase-1action on phospholipid bilayers, or onpolymer chains, as for DNA polymeraseacting on unwound single-stranded DNA(59, 60). In the case of immobilized en-

zymes and mobile reagents, SMFS cangenerate new insights on the diffusion ofreagents and products through gel-typeor macroreticular resins or throughcross-linked enzyme crystals or aggre-gates, which is crucial in understanding,e.g., product inhibition in industrial bio-catalysis.

On top of the study of translationaldiffusion, SMFS allows us to follow ro-tational dynamics and orientation ofsingle molecules by monitoring fluctua-tions in fluorescence polarization. In anexperimental catalytic context, it is ex-tremely challenging to attempt to relatecatalytic or sorption behavior to localorientation at the single site becausedoing so requires that the orientation ofthe surrounding matrix, e.g., the proteinstructure, or the pore structure of asolid catalyst, can be controlled andmeasured as a framework of reference.Measurements of average orientationwith respect to planar surfaces are pos-sible by using, for instance, second har-monic generation, but clearly such re-sults pertain to ensemble averages (61).For fluorescent single molecules in apolymer film, heterogeneities in rota-tional dynamics have been observed(62). For adsorption and catalysis in mi-croporous solids, it is important to un-derstand not only the location of ad-

Fig. 5. Tracking of single fluorescent molecules(F) diffusing in porous materials allows finding oflocal obstructions or fault planes.

Fig. 6. Combination of techniques that can be envisioned for in situ characterization of catalytic systems.(a) SMFS and tip-enhanced Raman spectroscopy. Catalytic conversion can be followed by fluorescencewhile chemical characterization of the active site can be obtained simultaneously. (b) SMFS and (time-resolved) x-ray experiments. This combination allows for simultaneous mapping of catalytic activity andcrystallographic data. (c) Optical trapping for immobilizing small catalyst particles in solution. Thisapproach minimizes diffusion resistances. (d) SMFS combined with phase-shaped femtosecond pulses cancontrol the outcome of the catalytic reaction.

Roeffaers et al. PNAS � July 31, 2007 � vol. 104 � no. 31 � 12607

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

0

sorbed molecules but also theirrotational freedom because bulk equilib-rium adsorption data suggest that notonly enthalpy but also entropy effectsare crucial in determining shape-selec-tive uptake of guests in, for example,molecular sieves (63).

As a long-term goal, increased timeresolution is needed for detecting theelementary reactions and intermediatesof a catalytic cycle or even the transitionstates. Attempts have been made re-cently with pump-probe techniques (64).For photo-induced reactions, the pumppulse can initiate the reaction. Theprobe pulse, released with an adjustabledelay time of a few femtoseconds, cansubsequently detect intermediates. How-ever, this experimental scheme requiresaccumulation of pulse sequences overseveral milliseconds, thereby signifi-cantly lowering the real-time resolution.Because accumulation over an ensembleof molecules, as is done in traditionalfemtosecond laser spectroscopy, has tobe replaced by accumulation of pulsesequences on a single active center, pho-tostability of the probe molecule will beone of the limiting factors in this type ofexperiment.

Integration of Techniquesand PerspectivesAlthough f luorescence microscopy isable to spatially map inhomogeneousactivity in catalytic materials, oneneeds to couple such data to insightsin the nature of the active sites.When the catalytic materials containsufficiently large crystals, it is straight-forward to distinguish the crystallo-graphically different crystal planes oredges, even under an optical micro-scope. For instance, on a crystalline-layered double hydroxide sample, it iseasily recognized that the catalyticproperties of the structural OH groupson the basal plane should be differentfrom those of exchanged OH� anionson the crystal fringes (65, 66).However, many catalysts are micro- ornanocrystalline or are even amorphous.Therefore, gaining insight in the ele-mental composition or in the concen-tration of active sites with differentspectroscopic signature is required. Itis preferred to characterize the samecatalyst particle simultaneously bySMFS and a combination of othertechniques. However, when such an insitu combination is not possible, high-resolution ex situ techniques still canreveal useful information. For instance,extended x-ray absorption fine struc-ture on metal ions or wavelength-dispersive elemental analysis inscanning electron microscopy can becomplementary to f luorescence-based

activity assays, although spatially over-laying the maps obtained by differenttechniques remains a potential sourceof ambiguity.

Fortunately, experimental conditionsof SMFS and other techniques increas-ingly become compatible and show agreat potential for future applications.Atomic force microscopy (AFM) isprobably the most obvious techniquefor use in combination with SMFS(67). The possibility to perform AFMmeasurements in liquid media allows athorough in situ examination of thesurface structure, in particular the ex-act localization of the catalytically in-teresting edges, kinks, and defects,while at the same time this informationcan be coupled to the optically ob-tained catalytic data. Moreover, thepossibility of mapping the chemicalcomposition of a surface by a modifiedtip makes AFM a very unique tool. Ithas been shown that local Raman spec-tra can be recorded with silver- orgold-coated tips (68–70). IR tech-niques are frequently used to charac-terize inorganic solid catalysts, but theylack sensitivity and the optical resolu-tion that can be achieved is limited be-cause of the higher wavelengths used(71). Surface-enhanced (resonance)Raman spectroscopy [SE(R)RS] from ametal-coated AFM tip does not sufferfrom these limitations. Indeed, it hasbeen shown that fingerprint signaturesof molecules and functional groups canbe obtained with this technique, in thebest case with nanometer resolution(72, 73). As illustrated in Fig. 6a, thecombination of f luorescence andSER(R)S could give insight into thenature of the active centers responsiblefor the catalytic activity. For certaincatalysts that are conductive (e.g.,gold-based catalysts) one can even ap-ply scanning tunneling microscopy(STM) as an angstrom resolution sur-face characterization technique. Notethat besides Raman signals, f luores-cence intensity can also be locally en-hanced by proper control of the dis-tance between the tip and thef luorescing site or molecule (74, 75).

When more precise information onthe elemental composition and/or crys-tallographic properties of the inorganicphase is needed, the techniques ofchoice are x-ray diffraction and electronmicroscopy. The latter technique, how-ever, has to be applied in a vacuum orat very reduced gas pressure and thusrequires ex situ combination with fluo-rescence observation (45). In situ combi-nation of x-ray diffraction and SMFSseems now within the reach of opticalmicroscopists, thanks to recent develop-ments in benchtop x-ray sources (76–

78). Indeed, by focusing femtosecondlaser pulses of the type that is generallyused in fluorescence microscopy on, forexample, aqueous solutions of alkalimetal chlorides, femtosecond x-raypulses of well defined wavelengths aregenerated. Fiber-assisted ‘‘focusing’’ ofthese pulses on the sample allows map-ping of the crystal properties of the cat-alysts with a time resolution down to thenanosecond level and a spatial accuracyof down to 1 �m (Fig. 6b). In combina-tion with SMFS activity analysis, thesetime-resolved diffractograms can reveallocal turnover-induced structural trans-formations in the catalysts (see above)(79). Furthermore, high-resolution opti-cal images obtained by mapping stochas-tic turnover events (45, 46) can be com-pared with x-ray diffraction data of theinorganic catalyst, thus directly linkingactivity to crystal properties.

Although industrial catalysts gener-ally are micrometer-sized to facilitatefiltration or recycling, mass transferlimitations are more easily eliminatedby scaling down the crystals to submi-cron dimensions. Suspensions of suchcatalysts may be studied by combiningthe f luorescence approach with opticaltrapping (Fig. 6c). This strategy elimi-nates the artifacts that arise when acrystal is deposited on a glass slide; insuch case, the crystal may be no longeraccessible through the face makingcontact with the substrate. Using opti-cal tweezers to trap a single catalyst inthe laser focus removes this problem.Both Raman spectroscopy and f luores-cence spectroscopy can be readily com-bined with optical trapping (80–82).

Until now, possible combinations ofSMFS and other in situ characterizationtechniques have been outlined. As anew challenge, one could envision ma-nipulating the outcome of catalytic pro-cesses by applying laser light. Therefore,single-molecule microscopy could becombined with coherent control tech-niques well established today in femto-second spectroscopy (83, 84). The use oflight to control reaction pathways is oneof the most exciting prospects in chemis-try. Coherent control aims to steer amolecular system toward a desired out-come by exploiting quantum interfer-ence effects. Recent studies have shownthat quantum interference between mul-tiple excitation pathways is used to can-cel coupling to the undesired channels.To achieve this goal, complex pulse-shaping schemes are being developed. Inreactions with potential functionalizationat CH3-, -CH2-, or -CH-, such as oxida-tion reactions, one could envision selec-tively enhancing the disfavored reactionpath, i.e., the reaction at CH3. In theframework of the approach proposed

12608 � www.pnas.org�cgi�doi�10.1073�pnas.0610755104 Roeffaers et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

0

here, the potential product moleculesshould have different extents of conju-gation so that they can be discriminatedbased on their spectral properties (Fig.6d). If this scheme can be realized, thechemists’ ultimate dream of making

molecules at will, one by one, comeswithin reach.

This work was performed within the frameworkof the Interuniversity Attraction Poles-VI pro-gram Functional Supramolecular Systems ofthe Belgian Federal government and GOA

2006/2 and was supported by the KatholiekeUniversiteit Leuven in the Centre of Excellencein Catalysis, an Institute for the Promotion ofInnovation through Science and Technology inFlanders (IWT-Vlaanderen) fellowship (toM.B.J.R.), and a Fonds Wetenschappelyk Or-derroek (Vlaanderen) fellowship (to G.D.C.).

1. de Levie R (2000) J Chem Educ 77:771–774.2. Xie XS, Lu HP (1999) J Biol Chem 274:15967–15970.3. Engelkamp H, Hatzakis NS, Hofkens J, De

Schryver FC, Nolte RJM, Rowan AE (2006) ChemCommun 9:935–940.

4. Smiley RD, Hammes GG (2006) Chem Rev106:3080–3094.

5. Craig DB, Arriaga EA, Wong JCY, Lu H, DovichiNJ (1996) J Am Chem Soc 118:5245–5253.

6. Min W, English BP, Luo GB, Cherayil BJ, Kou SC,Xie XS (2005) Acc Chem Res 38:923–931.

7. Lu HP, Xun L, Xie XS (1998) Science 282:1877–1882.8. Evans D, Osborn JA, Wilkinson G (1968) J Chem

Soc 3133–3142.9. Frohning CD, Kohlpaintner CW (1996) in Applied

Homogeneous Catalysis with Organometallic Com-pounds, eds Cornils B, Herrmann WA (VCH,Weinheim), Vol 1, pp 28–104.

10. Hoegaerts D, Jacobs PA (1999) Tetrahedron-asymmetr 10:3039–3043.

11. Berrisford DJ, Bolm C, Sharpless KB (1995) An-gew Chem Int Ed 34:1059–1070.

12. Sharpless KB (2002) Angew Chem Int Ed 41:2024–2032.13. McGarrigle EM, Gilheany DG (2005) Chem Rev

105:1563–1602.14. Rondelez Y, Tresset G, Tabata KV, Arata H,

Fujita H, Takechu S, Noji H (2005) Nat Biotech-nol 23:361–365.

15. Tan W, Yeung ES (1997) Anal Chem 69:4242–4248.16. Coperet C, Chabanas M, Saint-Arroman RP, Bas-

set JM (2003) Angew Chem Int Ed 42:156–181.17. De Vos DE, Dams M, Sels BF, Jacobs PA (2002)

Chem Rev 102:3615–3640.18. Kiel A, Kovacs J, Mokhir A, Kramer R, Herten

DP (2007) Angew Chem Int Ed 46:3158.19. Over H, Kim YD, Seitsonen AP, Wendt S, Lun-

dgren E, Schmid M, Varga P, Morgante A, Ertl G(2000) Science 287:1474–1476.

20. Lauritsen JV, Vang RT, Besenbacher F (2006)Catal Today 111:34–43.

21. Hendriksen BLM, Bobaru SC, Frenken JWM(2005) Top Catal 36:43–54.

22. Tinnefeld P, Sauer M (2005) Angew Chem Int Ed44:2642–2671.

23. Flomenbom O, Velonia K, Loos D, Masuo S,Cotlet M, Engelborghs Y, Hofkens J, Rowan AE,Nolte RJM, Van der Auweraer M, et al. (2005)Proc Natl Acad Sci USA 102:2368–2372.

24. Velonia K, Flomenbom O, Loos D, Masuo S,Cotlet M, Engelborghs Y, Hofkens J, Rowan AE,Klafter J, Nolte RJM, et al. (2005) Angew ChemInt Ed 44:560–564.

25. Roeffaers MBJ, Sels BF, Uji-i H, De Schryver FC,Jacobs PA, De Vos DE, Hofkens J (2006) Nature439:572–575.

26. Shaughnessy KH, Kim P, Hartwig JF (1999) J AmChem Soc 121:2123–2132.

27. Su H, Yeung ES (2000) J Am Chem Soc 122:7422–7423.28. Pap EHW, Drummen GPC, Winter VJ, Kooij

TWA, Rijken P, Wirtz KWA, Op den Kamp JAF,Hage WJ, Post JA (1999) FEBS Lett 453:278–282.

29. Goddard JP, Reymond JL (2004) Curr Opin Bio-tech 15:314–322.

30. Meshulam T, Herscovitz H, Casavant D, Ber-nardo J, Roman R, Haugland RP, StrohmeierGS, Diamond RD, Simons ER (1992) J BiolChem 267:21465–21470.

31. Hendrickson HS, Hendrickson EK, Johnson ID,Farber SA (1999) Anal Biochem 276:27–35.

32. Fabbrizzi L, Licchelli M, Pallavicini P (1999) AccChem Res 32:846–853.

33. Rurack K (2001) Spectrochim Acta A 57:2161–2195.34. Callan JF, de Silva AP, Magri DC (2005) Tetra-

hedron 61:8551–8588.35. Roeffaers MBJ, Sels BF, Loos D, Kohl C, Mullen

K, Jacobs PA, Hofkens J, De Vos DE (2005)Chemphyschem 6:2295–2299.

36. Haugland RP (2006) in The Handbook: A Guide toFluorescent Probes and Labeling Technologies, edSpence MTZ (Invitrogen, Carlsbad, CA).

37. Calzaferri G, Huber S, Maas H, Minkowski C(2003) Angew Chem Int Ed 42:3732–3758.

38. Betzig E, Trautman JK (1992) Science 257:189–195.

39. Hell SW (2003) Nat Biotechnol 21:1347–1355.40. Donnert G, Keller J, Medda R, Andrei MA, Rizzoli

SO, Lurmann R, Jahn R, Eggeling C, Hell SW (2006)Proc Natl Acad Sci USA 103:11440–11445.

41. Gustafsson MGL (2005) Proc Natl Acad Sci USA102:13081–13086.

42. Cheezum MK, Walker WV, Guilford WH (2001)Biophys J 81:2378–2388.

43. Yildiz A, Selvin PR (2005) Acc Chem Res 38:574–582.

44. Muls B, Uji-i H, Melnikov S, Moussa A, VerheijenW, Soumillion JP, Josemon J, Mullen K, HofkensJ (2005) Chemphyschem 6:2286–2294.

45. Betzig E, Patterson GH, Sougrat R, LindwasserW, Olenych S, Bonifacino JS, Davidson MW,Lippincott-Schwartz J, Hess HF (2006) Science313:1642–1645.

46. Rust MJ, Bates M, Zhuang XW (2006) Nat Methods3:793–795.

47. Zhuang X, Kim H, Pereira MJB, Babcock HP,Walter NG, Chu S (2002) Science 296:1473–1476.

48. Yang H, Luo G, Karnchanaphanurach P, LouieTM, Rech I, Cova S, Xun LY, Xie XS (2003)Science 302:262–266.

49. Weisz PB (1962) Adv Catal 13:137–190.50. Blomsma E, Martens JA, Jacobs PA (1997) J Catal

165:241–248.51. Xie XS (2002) J Chem Phys 117:11024–11032.52. English BP, Min W, van Oijen AM, Lee KT, Luo GB,

SunHY,CherayilBJ,KouSC,XieXS(2006)NatChemBiol 2:87–94.

53. Roeffaers MBJ, Sels BF, Uji-i H, Blanpain B, L’hoestP, Jacobs PA, De Schryver FC, Hofkens J, De Vos DE(2006) Angew Chem Int Ed 46:1706–1709.

54. Schmidt T, Schutz GJ, Baumgartner W, GruberHJ, Schindler H (1996) Proc Natl Acad Sci USA93:2926–2929.

55. Seebacher C, Hellriegel C, Deeg FW, Brauchle C,Altmaier S, Behrens P, Mullen K (2002) J PhysChem B 106:5591–5595.

56. Kirstein J, Platschek B, Jung C, Brown R, Bein T,Brauchle C (2007) Nat Mater 6:303–310.

57. Fu Y, Ye F, Sanders WG, Collinson MM, HigginsDA (2006) J Phys Chem B 110:9164–9170.

58. Karger J, Ruthven DM (1989) Zeolites 9:267–281.59. Rocha S, Verheijen W, Braeckmans K, Svenson A,

Skjøt M, De Schryver FC, Uji-i H, Hofkens J(2007) in Handai Nanophotonics, Vol 3, NanoBiophotonics, eds Masuhara H, Kawata S, Toku-naga F (Elsevier, Amsterdam), pp 133–141.

60. Zhang ZQ, Spiering MM, Trakselis MA, IshmaelFT, Xi J, Benkovic SJ, Hammes GG (2005) ProcNatl Acad Sci USA 102:3254–3259.

61. Kikteva T, Star D, Zhao ZH, Baisley TL, LeachGW (1999) J Phys Chem Rev B 103:1124–1133.

62. Uji-i H, Melnikov SM, Deres A, Bergamini G, DeSchryver F, Herrmann A, Mullen K, Enderlein J,Hofkens J (2006) Polymer 47:2511–2518.

63. Denayer JFM, Ocakoglu RA, Arik IC, KirschhockCEA, Martens JA, Baron GV (2005) Angew Chem IntEd 44:400–403.

64. van Dijk EMHP, Hernando J, Garcıa-Lopez JJ, Crego-Calama M, Reinhoudt DN, Kuipers L, Garcıa-ParajoMF, van Hulst NF (2005) Phys Rev Lett 94:078302.

65. Trifiro F, Vaccari A (1996) in ComprehensiveSupramol Chem, Vol 7, Solid-State SupramolChem: Two- and Three-Dimensional Inorganic Net-works, eds Atwood JL, Davies JE, Macnicol DD,Vogtle F, Lehn JM, Alberti G, Bein T (Pergamon,Oxford), pp 251–291.

66. Sels BF, De Vos DE, Jacobs PA (2001) Catal Rev43:443–488.

67. Kassies R, Van der Werf KO, Lenferink A, HunterCN, Olsen JD, Subramaniam V, Otto C (2005) JMicrosc 217:109–116.

68. Stockle RM, Suh YD, Deckert V, Zenobi R (2000)Chem Phys Lett 318:131–136.

69. Yeo BS, Zhang WH, Vannier C, Zenobi R (2006)Appl Spectrosc 60:1142–1147.

70. Kuhn S, Håkanson U, Rogobete L, Sandoghdar V(2006) Phys Rev Lett 97:017402.

71. Lehmann E, Chmelik C, Scheidt H, Vasenkov S,Staudte B, Karger J, Kremer F, Zadrozna G, Korna-towski J (2002) J Am Chem Soc 124:8690–8692.

72. Hartschuh A, Sanchez EJ, Xie XS, Novotny L(2003) Phys Rev Lett 90:095503.

73. Anderson N, Hartschuh A, Cronin S, Novotny L(2005) J Am Chem Soc 127:2533–2537.

74. Gryczynski I, Malicka J, Gryczynski Z, LakowiczJR (2004) J Phys Chem B 108:12568 –12574.

75. Anger P, Bharadwaj P, Novotny L (2006) Phys RevLett 96:113002.

76. Hatanaka K, Miura T, Fukumura H (2002) ApplPhys Lett 80:3925–3927.

77. Hatanaka K, Miura T, Fukumura H (2004) ChemPhys 299:265–270.

78. Bargheer M, Zhavoronkov N, Woerner N, El-saesser T (2006) Chemphyschem 7:783–792.

79. Hendriksen BLM, Bobaru SC, Frenken JWM(2005) Catal Today 105:234–243.

80. Hofkens J, Hotta J, Sasaki K, Masuhara H,Taniguchi T, Miyashita T (1997) J Am Chem Soc119:2741–2742.

81. Tsuboi Y, Nishino M, Sasaki T, Kitamura N(2005) J Phys Chem B 109:7033–7039.

82. Dol GC, Tsuda K, Weener JW, Bartels MJ,Asavei T, Gensch T, Hofkens J, Latterini L,Schenning APHJ, Meijer BW et al. (2001) AngewChem Int Ed 40:1710–1714.

83. Brixner T, Damrauer NH, Niklaus P, Gerber G(2001) Nature 414:57–60.

84. Herek JL, Wohlleben W, Cogdell RJ, Zeidler D,Motzkus M (2002) Nature 417:533–535.

Roeffaers et al. PNAS � July 31, 2007 � vol. 104 � no. 31 � 12609

Dow

nloa

ded

by g

uest

on

Dec

embe

r 6,

202

0