Intensity Fluctuations in Single-Molecule Surface-Enhanced Raman

Jan 22, 2019 - https://doi.org/10.1021/acs.accounts.8b00563 ... Download Hi-Res ImageDownload to MS-PowerPointCite This:Acc. Chem. ... (1,2) Label-fre...
0 downloads 0 Views 4MB Size
Article Cite This: Acc. Chem. Res. XXXX, XXX, XXX−XXX

pubs.acs.org/accounts

Intensity Fluctuations in Single-Molecule Surface-Enhanced Raman Scattering Diego P. dos Santos,† Marcia L. A. Temperini,‡ and Alexandre G. Brolo*,§ †

Downloaded via MIDWESTERN UNIV on January 23, 2019 at 08:57:59 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Departamento de Físico-Química, Instituto de Química, Universidade Estadual de Campinas, CP 6154, CEP 13083-970, Campinas, SP, Brazil ‡ Departamento de Química Fundamental, Instituto de Química, Universidade de São Paulo, CP 26.077, CEP 05513-970, São Paulo, SP, Brazil § Department of Chemistry, University of Victoria, P.O. Box 3065, Victoria V8W 3 V6, BC Canada CONSPECTUS: Around 20 years ago, the first reports of singlemolecule surface-enhanced Raman scattering (SM-SERS) caused a revolution in nanotechnology. Several researchers were quick to recognize the importance of a technique that can provide molecular vibrational fingerprinting at the SM level. Since then, a large amount of work has been devoted to the development of nanostructures capable of SM-SERS detection. A great effort has also been geared toward elucidating the different mechanisms that contribute to the effect. The understanding of the concept of plasmonic SERS hotspots, the role of chemical effects, and the dynamics of atomic and cluster rearrangements in nanometric domains has significantly advanced, driven by new computational and experimental methods used to study SM-SERS. In particular, SERS intensity fluctuations (SIFs) are now recognized as a hallmark of SM-SERS. Interpretation of SM-SERS data must take into consideration temporal and spatial variations as a natural consequence of the extreme localization inherent to surface plasmon resonances. Further analysis of variations in spectral signature, due to either molecular reorientation or photo (or thermal) processes, pointed to a new area that combines the power of SERS fingerprinting at the SM level to modern concepts of catalysis, such as hot-electrons-driven chemistry. This large body of work on the fundamental characteristics of the SM-SERS effect paved the way to the interpretation of other related phenomena, such as tip-enhanced Raman scattering (TERS). Despite all the fundamental progress, there are still very few examples of real applications of SM-SERS. In recent years, our research group has been studying SIFs, focused on different ways to use SM-SERS. The obvious application of SM-SERS is in analytical chemistry, particularly for quantification at ultralow concentrations (below 1 nM). However, quantification using SMSERS faces a fundamental sampling problem: the analytes (adsorbed in very small amounts, i.e., low surface coverage) must find rare SERS hotspots (areas with intense electric field localization that yields SERS). This limitation leads to strong temporal and spatial variations in SERS intensities, which translates into very large error bars in an experimental calibration curve. We tackled this problem by introducing the concept of “digital SERS”. This approach provided a roadmap for SERS quantification at ultralow concentrations and a potential pathway for a better understanding of the “reproducibility problem” associated with SERS. In this Account, we discuss not only the analytical applications but also other implementations of SM-SERS demonstrated by our group. These include the use of SM-SERS as a tool to probe colloidal aggregation, to evaluate the efficiency of SERS substrates, and to characterize the energy of localized resonances. SERS involves a series of random processes: hotspots are rare; surfaces/clusters constantly reconstruct; and molecules diffuse, adsorb, and desorb. All these pathways contribute to strong fluctuations in SERS intensities. Our work indicates that a statistical view of the effect can lead to interesting insights and the potential to fulfill the promise of this SM technique for real-world applications.



INTRODUCTION

An interesting characteristic of SM-SERS is the observation of fluctuations (temporal6 and/or spatial7) of SERS intensities. The origin of SERS intensity fluctuations (SIFs) can be traced to several fundamental distinct contributions. (SIFs are also commonly known in the SERS field as “blinking”. We avoid the

Single-molecule (SM) detection capability is lauded as a significant feature of plasmonic surface-enhanced Raman scattering (SERS).1,2 Label-free molecular vibrational information at ultralow concentrations (below 1 nM) is an exciting prospect in analytical chemistry. SM detection can also provide fundamental insights into catalytic mechanisms,3 intracellular dynamics,4 and molecular charge transport in nanoelectronics.5 © XXXX American Chemical Society

Received: November 10, 2018

A

DOI: 10.1021/acs.accounts.8b00563 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 1. (A) SERS enhancement factor (F) surface distribution from a model hotspot. The F distribution (colored log scale) affects the observed shape of the SERS intensity histograms. Surface coverages of 0.1% (B), 1% (C), 10% (D), and 20% (E). The observed theoretical change in the histogram shape is confirmed experimentally by varying the R6G solution concentration: 10 nM (F), 20 nM (G), 50 nM (H), and 5 μM (I). The experimental data used in this figure are from ref 13.

use of the term here, because “blinking” is a well-established phenomenon in fluorescence with a completely different physical origin.) Ultimately, plasmonic SERS can be defined as an increase in the Raman scattering generated from molecules adsorbed on metallic nanofeatures. In those situations, the observed enhancement is mainly due to plasmonic near-field effects and may reach several orders of magnitude in certain regions known as “SERS hotspots”.8,9 It is important to emphasize that SERS signals do not originate only from hotspots regions formed by coupled nanoparticle (NP) junctions.10,11 Plasmonic SERS hotspots can be clusters, protrusions, crevices, or any nanoscopic feature at the surface that allows enough field enhancement to enable SM detection. Although plasmonic effects are in general very important, they might not be the only contribution to the enhancement. In fact, different compounds might present widely different SERS responses due to distinct chemical interactions.12 In a typical plasmonic SERS experiment, a distribution of hotspots with different characteristics is present under the laser-illuminated area.13 The enhanced electromagnetic field is not homogeneously distributed throughout the surface.14 In fact, regions (hotspots) of strong field intensities, capable of producing detectable Raman scattering from a single adsorbed molecule, occupy just a fraction (less than 1%) of the geometric illuminated area.15 This means that, when the surface concentration of the adsorbate is also small, there is a low probability of a single molecule to find a hotspot capable of

promoting a measurable Raman scattering.1,13,16 Additional thermally driven17 and photo-driven18 processes, such as molecular reorientation,19 decomposition, and surface reconstruction, among others, could all simultaneously contribute to SIFs. Moreover, the dynamics of NPs and clusters moving in and out of the laser-illuminated area also contributes to SIFs in colloidal systems. 20 The analysis of SIFs through a combination of statistical methods, chemometrics, and computational techniques provides insights into the local field magnitude at the molecular location.21 Examples of SMSERS applications will be discussed in this Account. Our goal is to inspire researchers in the SERS field to embrace the stochastic nature of SERS as a tool for new fundamental insights and applications.



SERS INTENSITY FLUCTUATIONS In a typical plasmonic SERS experiment, a laser illuminates a small fraction of the nanostructured surface that may contain a distribution of hotspots. If the number of adsorbed molecules (surface concentration) is low, then the SERS signal will show fluctuations (temporal and spatial). A robust statistical analysis of these SIFs requires at least 1000 spectra. The determination of the SERS intensity from each spectrum is done using chemometrics.1,22 Principal component analysis (PCA) or other data reduction methods allow the utilization of the whole spectral pattern. Photodecomposition products can be readily identified and removed from the data set. PCA is particularly B

DOI: 10.1021/acs.accounts.8b00563 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 2. (A) Hotspot plasmon resonance wavelength map and (B) logarithm of the maximum enhancement factor (hotspot strength), Fmax, for Ag dimer as a function of radius and gap. (C, D) Distribution of κ′ values for a simulated bi-analyte experiment (analytes A and B) involving substrates with average radius 20 and 30 nm, respectively, and average gap size 1 nm. (E) Experimental κ′ for bands at ∼200 cm−1 in a bi-analyte experiment of brilliant green (BG) and crystal violet (CV) solution (10 nM each). Quadrants I−IV describe differential κ′ for BG and CV as indicated in the figure. Adapted with permission from refs 31 and 32. Copyright 2012 and 2016 American Chemical Society.

powerful for the study of mixtures in SM-SERS.16,23,24 This procedure has been used in bi-analyte experiments with isotopologues, which is accepted as a proof of the SM detection capabilities of SERS.25 Figure 1 shows examples computational and experimentalof SERS intensity histograms calculated from SIFs. Figure 1A shows the logarithmic distribution of SERS enhancement factors (F, ratio between SERS and Raman intensities from a single-molecule) around a single hotspot. As expected, due to the hotspot’s highly localized nature, a large number of events with intensities much lower than the average are observed in Figure 1B at low surface coverages (0.1%), together with a small fraction of events with intensities much larger than the average, reflecting the rare situations when molecules access the regions of large F-values. As the coverage increases (Figure 1C−E), the SIFs decrease as a result of the higher probability for molecules to access large field enhancement.14 Similar variations in histogram shape were experimentally observed for rhodamine 6G (R6G, a typical analyte used for SIF studies due to its large Raman cross-section) solutions at different concentrations (Figure 1F−I).13 The shape of the distributions derived from the SIFs, shown in Figure 1, can provide information about the relative surface concentration under different conditions. A similar type of behavior was observed under electrochemical conditions, where the surface concentration of the adsorbate was tuned by an applied external voltage.13 An understanding of SIF statistics is fundamental for SMSERS applications, particularly at systems with low analyte concentrations and/or relatively low affinity to the metal surface. We will next demonstrate that these SIFs can actually be explored to provide fundamental insights and for chemical analysis at ultralow concentrations.

regime, however, SIFs can also be observed at the anti-Stokes side,26,27 and some of the events present unusually high antiStokes intensities.28 Vibrational pumping, where excited vibrational states become overpopulated (relative to thermal equilibrium), has been invoked to explain the large anti-Stokes signature in some SM-SERS events.26 However, the experimental observation of very large Stokes-enhanced events in a SIF trajectory cannot be explained by such a mechanism.29,30 A resonance argument that takes into consideration the evolution of the anti-Stokes-to-Stokes intensity ratios (IAS/IS) was then proposed by our group.31 The main idea is that the plasmonic resonance from a given hotspot might coincide with either the Stokes or the anti-Stokes scattering. This shifted local resonance (relative to the laser excitation) leads to an imbalance in the IAS/IS relative to the ratio expected from a thermal equilibrium (ρ). This is illustrated in Figure 2, using the parameter κ′ defined in eq 1 (κ′ > 1 for preferential antiStokes enhancements and κ′ < −1 for preferential Stokes enhancements). l IAS I o o o if AS ≥ 1 o o ISρ I o Sρ o κ′ = m o o Iρ I o o − S if AS < 1 o o o I I o Sρ n AS

(1)

Figure 2A indicates that 30 nm Ag NP dimers are expected to present strong resonances very close to 633 nm laser excitation, whereas shorter wavelength resonances, i.e., shifted to the anti-Stokes side, are expected for 20 nm Ag NP dimers. κ′ = 0 was assigned for events where SERS from only one dye was observed. Figure 2 shows intense SIFs with both enhanced Stokes and anti-Stokes signals. The changes in resonance conditions, driven by local geometry (Figures 2B), are directly reflected in the κ′ distribution. It is interesting to note events in Figure 2 where both dyes were simultaneously observed but had different κ′ values, suggesting that the molecules were probed from different hotspots. These types of events are observed even at very low coverages (0.1% in Figure



USING SIFs TO ESTIMATE THE HOTSPOT PLASMON RESONANCE WAVELENGTH Usually in Raman spectroscopy (including SERS), only the Stokes-scattered photons are recorded. In the SM-SERS C

DOI: 10.1021/acs.accounts.8b00563 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 3. Resonance contributions calculated from κ′ values and their application to estimate the hotspot plasmon resonance wavelength peak positions in a SM-SERS map. Adapted with permission from ref 31. Copyright 2012 American Chemical Society.

Figure 4. (A) Generalized Mie Theory simulations for nanoparticle clusters and analysis of the effect of the hotspot resonance on the IAS/IS (antiStokes Raman are plotted as negative shifts). (B−D) IAS/IS distributions for different [KBr]. (E) Evolution of the cluster size as a function of [KBr] obtained from the IAS/IS intensity ratios. Adapted with permission from ref 32. Copyright 2016 American Chemical Society.



USING SIFs TO MONITOR COLLOIDAL AGGREGATION The common application of IAS/IS in normal Raman is to measure temperature. There are a few examples of this application in the SERS literature.33,34 However, Figures 2 and 3 show that fluctuations in IAS/IS cannot be always uniquely assigned to temperature changes. This inability of SERS to accurately measure local temperature has been confirmed by others.28 On the other hand, IAS/IS can be correlated to the energy of the hotspot resonance,31 which is directly related to the nanoparticle’s state of aggregation. A single NP cluster will present a unique set of hotspots with resonances related to the cluster’s geometry.35 The IAS/IS calculated from the SIF statistics of a colloidal system could then be used to estimate the distribution of NP clusters (considering that each event corresponds to a SM visiting a hotspot in the cluster).32 This concept is illustrated in Figure 4A, where (calculated) examples of resonance energies of colloidal assemblies are presented. The analysis of experimental IAS/IS allied to calculations allowed the hotspot structures for single aggregates to be estimated.32 The time-dependent IAS/IS SIFs in a colloidal system should correlate to a distribution of colloidal clusters at different aggregation states. Figure 4 shows an example where the distribution of IAS/IS obtained for NPs suspended in water is

2C), especially for larger strengths (Fmax, Figure 2B). A similar profile was observed in the bi-analyte experiment (Figure 2E) involving both crystal violet (CV) and brilliant green (BG) at low concentrations (10 nM each) on a roughened Ag electrode.27,31 Figure 2 leads to the realization that, in the SM regime, the experimental κ′ distribution can be directly correlated to the distribution of hotspots’ resonance wavelengths. This concept is illustrated in Figure 3,31 which shows two experimental SIF events (from a mixture of CV and BG, Figure 3A) that yielded two very distinct κ′ values in Figure 2E. The anti-Stokes− Stokes profiles can be interpreted in terms of the local resonances visited by a SM. This means that for each SMSERS event shown in Figure 2E, a simple Lorentzian model (Figure 3A), requiring two fitting parameters (resonance peak wavelength and width), can predict the shape of the hotspot resonances. Figure 3B shows the calculated spatial distribution of the hotspot plasmon resonance wavelengths obtained from SM measurements. The evaluation of κ′ from SM-SERS is an example of how local hotspot information is embedded in the SIFs statistics. Near-field characteristics were evaluated from a careful analysis and understanding on the origins of SIFs. D

DOI: 10.1021/acs.accounts.8b00563 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 5. Electrodynamics simulations of Au nanorods (A) and Ag nanospheres (B). (C) Dependence of hotspot strength (FM) on k. Adapted with permission from ref 41. Copyright 2018 American Chemical Society.

enhancement and reproducibility). Consequently, in a direct comparison, the best performance should be assigned to the substrate that shows fewer SIFs at a given (low) concentration (for the same analyte).41 This comparative method does not require assumptions related to the number of molecules being probed. The caveat for this procedure is that it assumes the same adsorption parameters among substrates with possible different surface chemistries. The approach described above is qualitative, but a quantitative model that takes into consideration the statistics of the SIFs was also suggested. The SIFs distribution in the SM regime, e.g., Figure 1, reveals the rarity of a highly efficient hotspot. A truncated Pareto distribution model was suggested to characterize SIFs distributions from one hotspot (Figure 1A).14 For an arbitrary number of hotspots with different properties excited simultaneously, a parametrized probability density function (pdf), g(F; L M , k̅, σLM), named the truncated Pareto mixed distribution, has been suggested.41 The generalized model, eq 2, describes the probability of measuring an enhancement factor (F) from a distribution of SERS intensities in the SM regime.

clearly dependent on the halide concentration (ionic strength). The increase in cluster size led to broad, red-shifted resonances, which contributed to Stokes-enhanced events, as well as IAS/IS ratios close to the thermal equilibrium. Therefore, the evolution of IAS/IS as a function of salt concentration (Figure 4E) is a probe of the aggregation dynamics. The hatched region in Figure 4E corresponds to small colloidal aggregates (dimers for instance), while the role of large clusters becomes more prominent for [KBr] > 7.5 mM. The specific hotspot structure that generates a particular value of IAS/IS was estimated using Generalized Mie Theory.32 A direct correlation between IAS/IS and cluster structure can be well estimated for small clusters (dimers and trimers). More complex geometries are accessed as cluster sizes increases, leading to multiple possible matches between the experimental (IAS/IS) and the calculated hotspot resonance. In any case, the results from Figure 4 emphasize the idea that SIFs in the SM regime contain information that can be translated into the local structure of the hotspot in colloidal systems.32



EXPLORING SIFs TO EVALUATE THE EFFICIENCY OF A SERS SUBSTRATE The fabrication and development of different types of nanostructured surfaces correspond to one of the most sought research activity in SERS.36 Literally, several hundreds of different types of substrates have been reported.37−39 The efficiency of those substrates is usually evaluated using the average SERS enhancement factor (⟨F⟩).14,40 Although ⟨F⟩ is the common metric in the SERS field, its experimental determination can be challenging. For instance, the estimation of the number of adsorbed molecules is a major source of error. SERS substrates are generally complex and the number of species under illumination is not known a priori. Careful analysis of SIFs revealed an alternative approach for evaluation and comparison of SERS substrates.41 The amount of adsorbed molecules is related to the concentration of the species in solution; therefore, the SIFs frequencies at a particular concentration should be indicative of the density of hotspots that enables a SM response. A “good” SERS substrate should have high density of efficient hotspots (leading to good

g (F ; L M , k ̅ , σ L M ) =

∫0



wFM(FM ; L M , σLM)p(F ; k ̅ , FM) dFM

(2)

FM is the hotspot strength, defined as the maximum value of the enhancement factor in an individual hotspot. p(F;k̅,FM) is the pdf for a single hotspot, and wFM(FM; L M ,σLM) is the adsorption probability for a given hotspot. Equation 2 has three parameters that characterize a SERS substrate: L M is the average of a distribution of logarithmic hotspot strength, σLM is its standard deviation, and k̅ is related to spatial distribution of the localized field. A good SERS substrate should have high values for L M , narrow distribution of hotspot strength (given by σLM), and an enhanced field spread over a large geometric area (low k̅). This combination is equivalent to a large ⟨F⟩. However, the use of these new parameters (extracted experimentally from the SIFs histograms) provides more E

DOI: 10.1021/acs.accounts.8b00563 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

the total digital counts for each concentration.47 A proof-ofconcept of this approach is shown in Figure 6. Figure 6a,b

latitude for a critical analysis of the substrate performance. A simple example of the utility of these new metrics is illustrated from electrodynamics calculations in Figure 5, which shows hotspots formed by a nanorod dimer (panel A) and by three nanospheres (panel B). Figure 5C plots values of FM against the localization parameter for our model systems. Figure 5C shows that the nanorod-based hotspot (Figure 5A) has a highly localized hotspot with a large LM value. On the other hand, the trimer hotspot is weaker, but it spreads over a larger surface area. Calculated points for other individual clusters are also represented in Figure 5C for comparison. The point of Figure 5 is to demonstrate that, by increasing the parameter space through a rational mathematical model, it is possible to infer a more complete picture of the SERS substrate performance. For instance, strong molecular scatterers do not require a large electromagnetic boost to be detected, and having fewer spatial variations in SERS intensities could be an advantage. On the other hand, weaker Raman scatterers should be measured in substrates that provide large LM, even if the signal will ultimately originate from a very localized area.41 The examples in Figure 5 illustrate the concept without exploring the statistics of a large distribution of hotspots. Equation 2 was then applied to several random assembles of multiple hotspots.41 Most importantly, the model evaluated the experimental SIFs from a roughened electrode. The results demonstrated a large relative standard deviation of hotspot strengths (about 30%) combined with a high degree of localization of highly enhancing adsorption sites.41 These results are consistent with the stochastic electrochemical “activation” process, suggesting that only a small fraction of the surface carries hotspots capable of producing SM signals. Characterization of the roughened electrode using eq 2 is more informative than the simple calculation of ⟨F⟩. An important caveat for the application of eq 2 is that the experimental data are intensity distributions. It is not possible to recover absolute substrate metrics without converting the intensity values to enhancement factors. Future work in this area should consider previous attempts to obtain SM-SERS cross-sections42 to convert the distribution of SERS intensities to enhancement factors. This step could lead to standardization of the substrate evaluation metrics, which would be a significant advance in the SERS field.

Figure 6. (a, b) Digital SERS maps for an isotopologue mixture of ciprofloxacin. The colors are related to the detection of normal (blue) and isotopically substituted ciprofloxacin (red). The two ciprofloxacin species were detected at the same time in just one event (purple). (c) Analytical calibration curve from the digital SERS procedure. Adapted with permission from ref 47. Copyright 2018 American Chemical Society.

shown SERS mappings from a SERS-active Au surface exposed to ciprofloxacin48 (estimated 10 molecules/μm2). A 50 × 50 μm2 area of the surface was scanned with 5 μm steps, leading to 121 SERS spectra (pixels). The SIF trajectory (spatially dependent SERS spectra) was analyzed using chemometrics. The colored squares in Figure 6a,b indicate events with a measurable SERS response (attributed a value of “1”), while the white areas correspond to regions with zero SERS intensity. The SERS digital counts, obtained by counting the total number of SM events for a given concentration, were then used to generate the calibration curve in Figure 6c.47 The procedure yielded a reasonable calibration curve without the strong variations expected from an absolute SERS intensity vs concentration plot. Nonetheless, it is important to point out that the maps in this proof-of-concept have only 121 pixels. The counting fractional error in basic digital statistics decreases with the number of counts. Therefore, the digital procedure can provide quantification with a reasonable level of precision, as long as a large area is mapped. This digital method relies on the assumption that the SERS response originates from a SM. The digital SERS maps in Figure 6a,b were actually obtained from a mixture of ciprofloxacin isotopologues. The two types of ciprofloxacin present distinct vibrational signatures due to the isotopic differences and can be individually identified.47 The blue squares in Figure 6a indicate regions where the vibrational



USING SIFs FOR QUANTIFICATION AT ULTRALOW CONCENTRATIONS The most sought-after applications of SERS are in analytical chemistry.36,43,44 SM-SERS is particularly promising for quantifications at ultralow concentrations.45 However, quantification using SERS at ultralow concentrations faces a fundamental sampling problem.46 The adsorbed molecules are obviously present at low surface concentrations, and SERS hotspots are rare (Figure 1). Therefore, the probability of a SM finding a SERS hotspot capable of generating a measurable Raman signal is low. This leads to a fundamental problem, since an experimental calibration curve could present very large error bars due to SIFs. We tackled this quantification issue by introducing the concept of “digital SERS”.47 This new approach provided a roadmap for SERS quantification at ultralow concentrations and a potential pathway for a wide application of SM-SERS in analytical chemistry. The concept is relatively simple: each event in a SIF trajectory is digitized based on an intensity threshold, and a calibration curve is constructed considering F

DOI: 10.1021/acs.accounts.8b00563 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Accounts of Chemical Research



signature of the regular ciprofloxacin were found, while the red squares in Figure 6b show regions where the isotopically substituted molecules were identified. Only ∼6% of the positive events (Figure 6a,b) present SERS features from the mixture of both compounds (purple squares). These observations are strong indications that the SM regime was attained. The digital SERS method provides a unique example of quantification that explored the stochastic characteristics of SIFs, which can be easily generalized and applied for ultrasensitive analysis of biological and environmental samples.



Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Diego P. dos Santos: 0000-0001-9468-7293 Marcia L. A. Temperini: 0000-0003-4655-6891 Alexandre G. Brolo: 0000-0002-3162-0881 Notes

The authors declare no competing financial interest. Biographies

CONCLUSIONS AND OUTLOOK

Diego P. dos Santos is an Assistant Professor in the Institute of Chemistry at the University of Campinas (UNICAMP). His research interests are in Raman spectroscopy and in the application of SERS to probe metal nanoparticles’ optical properties.

SERS intensity fluctuations (SIFs) are fundamental characteristics of the response from a small number of molecules adsorbed on hotspots. Statistical analysis of the fluctuations can provide fundamental insights into the nanoenvironment visited by a SM. For instance, SIFs’ frequencies and distributions provide information about the density and efficiency of the hotspot assembly.13 Statistical analysis of IAS/IS at the SM level leads to insights into local plasmonic resonances.31,32 Analysis of SIFs can be used to evaluate the performance of SERS substrates 41 and can allow the development of a quantitative SERS-based protocol for chemical analysis at ultralow concentrations.47 The results from our group are part of growing activities in the field of plasmonics that recognizes SIFs as a central aspect of SERS. For instance, tip-enhanced Raman scattering (TERS) inherently probes a very small number of molecules.2,49 TERS provides exquisite details about the surface structure of thin films,50 and the technique is particularly powerful for studying surface reactions and catalysis.51 In all cases, an appreciation of the role of SIFs and spectral wandering in TERS is relevant. The surprising sub-molecular resolution observed in TERS52 has led to interesting experimental and computational developments in the physics of cavities within plasmonic structures.53,54 Experiments from plasmonic junctions have demonstrated interesting SIFs that have been assigned to the generation of metallic protrusions (pico-cavity).55,56 Finally, SIFs play a major role in SERS-based super-resolution imaging methods.57 Stochastic optical reconstruction microscopy (STORM) is a fluorescence-based modality for superresolution where an image is reconstructed from the intermittent emissions of SMs captured by a high-speed camera. SIFs from single emitters fulfill some of the requirements for STORM, without the necessity of labels. Narrow vibrational bandwidth, discernible molecular fingerprint, short lifetime of a vibrational excited state, and resistance to photobleaching are some advantages of SIFs-based superresolution imaging. Preliminary SERS super-resolution images obtained from SIFs have already been reported.57,58 All these recent developments demonstrate that SIFs carry unique information. However, the stochastic nature of SERS suggests that statistical methods, such as the ones implemented in our work, could potentially bring new insights that have been hidden in typical SERS. In that sense, SIFs may open the door to a variety of effects that will continue to stimulate the fields of nanotechnology and nanophotonics in the years to come.

Marcia Laudelina Arruda Temperini is a Full Professor at the Institute of Chemistry at the University of São Paulo. Her research focuses on the study of electronic and vibrational properties of conducting polymers and on the molecular SERS responses in electrochemical systems, molecular junctions, and single molecules. Alexandre G. Brolo is a Professor of Chemistry at the University of Victoria, Canada. His research interests include surface-enhanced spectroscopies and the fabrication and application of metallic nanostructures.



ACKNOWLEDGMENTS A.G.B. thanks NSERC, CFI, and BCKDF. D.P.S. gratefully acknowledges CNPq (408985/2016-0). This study was financed in part by CAPES − Finance Code 001, FAPESP (2018/15987-9 and 2016/21070-5) and CNPq 302792/20155. This work used resources of the Centro Nacional de Processamento de Alto Desempenho em São Paulo (CENAPAD-SP).



REFERENCES

(1) Le Ru, E. C.; Etchegoin, P. G.; Johnson, M. A.; Martinez, T. J. Single-Molecule Surface-Enhanced Raman Spectroscopy. Annu. Rev. Phys. Chem. 2012, 63, 65−87. (2) Zrimsek, A. B.; Chiang, N. H.; Mattei, M.; Zaleski, S.; McAnally, M. O.; Chapman, C. T.; Henry, A. I.; Schatz, G. C.; Van Duyne, R. P. Single-Molecule Chemistry with Surface- and Tip-Enhanced Raman Spectroscopy. Chem. Rev. 2017, 117, 7583−7613. (3) Joseph, V.; Engelbrekt, C.; Zhang, J. D.; Gernert, U.; Ulstrup, J.; Kneipp, J. Characterizing the Kinetics of Nanoparticle-Catalyzed Reactions by Surface-Enhanced Raman Scattering. Angew. Chem., Int. Ed. 2012, 51, 7592−7596. (4) Lane, L. A.; Qian, X. M.; Nie, S. M. SERS Nanoparticles in Medicine: From Label-Free Detection to Spectroscopic Tagging. Chem. Rev. 2015, 115, 10489−10529. (5) Li, Y.; Doak, P.; Kronik, L.; Neaton, J. B.; Natelson, D. Voltage tuning of vibrational mode energies in single-molecule junctions. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 1282−1287. (6) Andrade, G. F. S.; Min, Q.; Gordon, R.; Brolo, A. G. SurfaceEnhanced Resonance Raman Scattering on Gold Concentric Rings: Polarization Dependence and Intensity Fluctuations. J. Phys. Chem. C 2012, 116, 2672−2676. (7) Pieczonka, N. P. W.; Moula, G.; Aroca, R. F. SERRS for SingleMolecule Detection of Dye-Labeled Phospholipids in LangmuirBlodgett Monolayers. Langmuir 2009, 25, 11261−11264. (8) Nam, J. M.; Oh, J. W.; Lee, H.; Suh, Y. D. Plasmonic NanogapEnhanced Raman Scattering with Nanoparticles. Acc. Chem. Res. 2016, 49, 2746−2755.

G

DOI: 10.1021/acs.accounts.8b00563 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research (9) Le Ru, E. C.; Grand, J.; Sow, I.; Somerville, W. R. C.; Etchegoin, P. G.; Treguer-Delapierre, M.; Charron, G.; Felidj, N.; Levi, G.; Aubard, J. A Scheme for Detecting Every Single Target Molecule with Surface-Enhanced Raman Spectroscopy. Nano Lett. 2011, 11, 5013− 5019. (10) Willets, K. A.; Wilson, A. J.; Sundaresan, V.; Joshi, P. B. SuperResolution Imaging and Plasmonics. Chem. Rev. 2017, 117, 7538− 7582. (11) Asiala, S. M.; Schultz, Z. D. Label-free in situ detection of individual macromolecular assemblies by surface enhanced Raman scattering. Chem. Commun. 2013, 49, 4340−4342. (12) Fan, M. K.; Lai, F. J.; Chou, H. L.; Lu, W. T.; Hwang, B. J.; Brolo, A. G. Surface-enhanced Raman scattering (SERS) from Au:Ag bimetallic nanoparticles: the effect of the molecular probe. Chemical Science 2013, 4, 509−515. (13) dos Santos, D. P.; Andrade, G. F. S.; Temperini, M. L. A.; Brolo, A. G. Electrochemical Control of the Time-Dependent Intensity Fluctuations in Surface-Enhanced Raman Scattering (SERS). J. Phys. Chem. C 2009, 113, 17737−17744. (14) Le Ru, E. C.; Etchegoin, P. G.; Meyer, M. Enhancement factor distribution around a single surface-enhanced Raman scattering hot spot and its relation to single molecule detection. J. Chem. Phys. 2006, 125, 204701. (15) Fang, Y.; Seong, N. H.; Dlott, D. D. Measurement of the distribution of site enhancements in surface-enhanced Raman scattering. Science 2008, 321, 388−392. (16) Etchegoin, P. G.; Meyer, M.; Blackie, E.; Le Ru, E. C. Statistics of single-molecule surface enhanced Raman scattering signals: Fluctuation analysis with multiple analyte techniques. Anal. Chem. 2007, 79, 8411−8415. (17) Kudelski, A. Fluctuations of Raman spectra of hydrogenated amorphous carbon deposited on electrochemically-roughened silver. Chem. Phys. Lett. 2006, 427, 206−209. (18) Mezni, A.; Dammak, T.; Fkiri, A.; Mlayah, A.; Abid, Y.; Smiri, L. S. Photochemistry at the Surface of Gold Nanoprisms from Surface-Enhanced Raman Scattering Blinking. J. Phys. Chem. C 2014, 118, 17956−17967. (19) Marshall, A. R. L.; Stokes, J.; Viscomi, F. N.; Proctor, J. E.; Gierschner, J.; Bouillard, J. S. G.; Adawi, A. M. Determining molecular orientation via single molecule SERS in a plasmonic nano-gap. Nanoscale 2017, 9, 17415−17421. (20) Scott, B. L.; Carron, K. T. Dynamic Raman Scattering Studies of Coated Gold Nanoparticles: 4-Mercaptopyridine, 4-Mercaptophenol, and Benzenethiol. J. Phys. Chem. C 2016, 120, 20905−20913. (21) Kusch, P.; Mastel, S.; Mueller, N. S.; Azpiazu, N. M.; Heeg, S.; Gorbachev, R.; Schedin, F.; Hubner, U.; Pascual, J. I.; Reich, S.; Hillenbrand, R. Dual-Scattering Near-Field Microscope for Correlative Nanoimaging of SERS and Electromagnetic Hotspots. Nano Lett. 2017, 17, 2667−2673. (22) Bohn, J. E.; Le Ru, E. C.; Etchegoin, P. G. A Statistical Criterion for Evaluating the Single-Molecule Character of SERS Signals. J. Phys. Chem. C 2010, 114, 7330−7335. (23) Etchegoin, P. G.; Le Ru, E. C.; Meyer, M. Evidence of Natural Isotopic Distribution from Single-Molecule SERS. J. Am. Chem. Soc. 2009, 131, 2713−2716. (24) Zrimsek, A. B.; Wong, N. L.; Van Duyne, R. P. Single Molecule Surface-Enhanced Raman Spectroscopy: A Critical Analysis of the Bianalyte versus Isotopologue Proof. J. Phys. Chem. C 2016, 120, 5133−5142. (25) Kleinman, S. L.; Ringe, E.; Valley, N.; Wustholz, K. L.; Phillips, E.; Scheidt, K. A.; Schatz, G. C.; Van Duyne, R. P. Single-Molecule Surface-Enhanced Raman Spectroscopy of Crystal Violet Isotopologues: Theory and Experiment. J. Am. Chem. Soc. 2011, 133, 4115− 4122. (26) Kneipp, K.; Kneipp, H. SERS signals at the anti Stokes side of the excitation laser in extremely high local optical fields of silver and gold nanoclusters. Faraday Discuss. 2006, 132, 27−33. (27) dos Santos, D. P.; Andrade, G. F. S.; Brolo, A. G.; Temperini, M. L. A. Fluctuations of the Stokes and anti-Stokes surface-enhanced

resonance Raman scattering intensities in an electrochemical environment. Chem. Commun. 2011, 47, 7158−7160. (28) Pozzi, E. A.; Zrimsek, A. B.; Lethiec, C. M.; Schatz, G. C.; Hersam, M. C.; Van Duyne, R. P. Evaluating Single-Molecule Stokes and Anti-Stokes SERS for Nanoscale Thermometry. J. Phys. Chem. C 2015, 119, 21116−21124. (29) Galloway, C. M.; Le Ru, E. C.; Etchegoin, P. G. Single-molecule vibrational pumping in SERS. Phys. Chem. Chem. Phys. 2009, 11, 7372−7380. (30) Brolo, A. G.; Sanderson, A. C.; Smith, A. P. Ratio of the surfaceenhanced anti-Stokes scattering to the surface-enhanced StokesRaman scattering for molecules adsorbed on a silver electrode. Phys. Rev. B: Condens. Matter Mater. Phys. 2004, 69, 045424. (31) dos Santos, D. P.; Temperini, M. L. A.; Brolo, A. G. Mapping the Energy Distribution of SERRS Hot Spots from Anti-Stokes to Stokes Intensity Ratios. J. Am. Chem. Soc. 2012, 134, 13492−13500. (32) dos Santos, D. P.; Temperini, M. L. A.; Brolo, A. G. SingleMolecule Surface-Enhanced (Resonance) Raman Scattering (SE(R)RS) as a Probe for Metal Colloid Aggregation State. J. Phys. Chem. C 2016, 120, 20877−20885. (33) Ioffe, Z.; Shamai, T.; Ophir, A.; Noy, G.; Yutsis, I.; Kfir, K.; Cheshnovsky, O.; Selzer, Y. Detection of heating in current-carrying molecular junctions by Raman scattering. Nat. Nanotechnol. 2008, 3, 727−732. (34) Maher, R. C.; Cohen, L. F.; Le Ru, E. C.; Etchegoin, P. G. A study of local heating of molecules under Surface Enhanced Raman Scattering (SERS) conditions using the anti-Stokes/Stokes ratio. Faraday Discuss. 2006, 132, 77−83. (35) Kitahama, Y.; Nagahiro, T.; Tanaka, Y.; Itoh, T.; Ozaki, Y. Analysis of blinking from multicoloured SERS-active Ag colloidal nanoaggregates with poly-L-lysine via truncated power law. J. Raman Spectrosc. 2017, 48, 570−577. (36) Fan, M. K.; Andrade, G. F. S.; Brolo, A. G. A review on the fabrication of substrates for surface enhanced Raman spectroscopy and their applications in analytical chemistry. Anal. Chim. Acta 2011, 693, 7−25. (37) Luo, S. C.; Sivashanmugan, K.; Liao, J. D.; Yao, C. K.; Peng, H. C. Nanofabricated SERS-active substrates for single-molecule to virus detection in vitro: A review. Biosens. Bioelectron. 2014, 61, 232−240. (38) Gong, Z. J.; Wang, C. C.; Wang, C.; Tang, C. Y.; Cheng, F. S.; Du, H. J.; Fan, M. K.; Brolo, A. G. A silver nanoparticle embedded hydrogel as a substrate for surface contamination analysis by surfaceenhanced Raman scattering. Analyst 2014, 139, 5283−5289. (39) Qi, J.; Motwani, P.; Gheewala, M.; Brennan, C.; Wolfe, J. C.; Shih, W. C. Surface-enhanced Raman spectroscopy with monolithic nanoporous gold disk substrates. Nanoscale 2013, 5, 4105−4109. (40) Le Ru, E. C.; Blackie, E.; Meyer, M.; Etchegoin, P. G. Surface enhanced Raman scattering enhancement factors: a comprehensive study. J. Phys. Chem. C 2007, 111, 13794−13803. (41) Kiefl, E. J.; Kiefl, R. F.; dos Santos, D. P.; Brolo, A. G. Evaluation of Surface-Enhanced Raman Spectroscopy Substrates from Single-Molecule Statistics. J. Phys. Chem. C 2017, 121, 25487−25493. (42) Darby, B. L.; Etchegoin, P. G.; Le Ru, E. C. Single-molecule surface-enhanced Raman spectroscopy with nanowatt excitation. Phys. Chem. Chem. Phys. 2014, 16, 23895−23899. (43) Yang, S. K.; Dai, X. M.; Stogin, B. B.; Wong, T. S. Ultrasensitive surface-enhanced Raman scattering detection in common fluids. Proc. Natl. Acad. Sci. U. S. A. 2016, 113, 268−273. (44) Schlucker, S. Surface-Enhanced Raman Spectroscopy: Concepts and Chemical Applications. Angew. Chem., Int. Ed. 2014, 53, 4756−4795. (45) Simoncelli, S.; Roller, E. M.; Urban, P.; Schreiber, R.; Turberfield, A. J.; Liedl, T.; Lohmuller, T. Quantitative SingleMolecule Surface Enhanced Raman Scattering by Optothermal Tuning of DNA Origami-Assembled Plasmonic Nanoantennas. ACS Nano 2016, 10, 9809−9815. (46) Crawford, A. C.; Skuratovsky, A.; Porter, M. D. Sampling Error: Impact on the Quantitative Analysis of Nanoparticle-Based SurfaceH

DOI: 10.1021/acs.accounts.8b00563 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research Enhanced Raman Scattering Immunoassays. Anal. Chem. 2016, 88, 6515−6522. (47) de Albuquerque, C. D. L.; Sobral-Filho, R. G.; Poppi, R. J.; Brolo, A. G. Digital Protocol for Chemical Analysis at Ultralow Concentrations by Surface-Enhanced Raman Scattering. Anal. Chem. 2018, 90, 1248−1254. (48) Hong, K. Y.; de Albuquerque, C. D. L.; Poppi, R. J.; Brolo, A. G. Determination of aqueous antibiotic solutions using SERS nanogratings. Anal. Chim. Acta 2017, 982, 148−155. (49) Sonntag, M. D.; Klingsporn, J. M.; Garibay, L. K.; Roberts, J. M.; Dieringer, J. A.; Seideman, T.; Scheidt, K. A.; Jensen, L.; Schatz, G. C.; Van Duyne, R. P. Single-Molecule Tip-Enhanced Raman Spectroscopy. J. Phys. Chem. C 2012, 116, 478−483. (50) Zeng, Z. C.; Huang, S. C.; Wu, D. Y.; Meng, L. Y.; Li, M. H.; Huang, T. X.; Zhong, J. H.; Wang, X.; Yang, Z. L.; Ren, B. Electrochemical Tip-Enhanced Raman Spectroscopy. J. Am. Chem. Soc. 2015, 137, 11928−11931. (51) van Schrojenstein Lantman, E. M.; Deckert-Gaudig, T.; Mank, A. J. G.; Deckert, V.; Weckhuysen, B. M. Catalytic processes monitored at the nanoscale with tip-enhanced Raman spectroscopy. Nat. Nanotechnol. 2012, 7, 583−586. (52) Zhang, R.; Zhang, Y.; Dong, Z. C.; Jiang, S.; Zhang, C.; Chen, L. G.; Zhang, L.; Liao, Y.; Aizpurua, J.; Luo, Y.; Yang, J. L.; Hou, J. G. Chemical mapping of a single molecule by plasmon-enhanced Raman scattering. Nature 2013, 498, 82−86. (53) Chikkaraddy, R.; de Nijs, B.; Benz, F.; Barrow, S. J.; Scherman, O. A.; Rosta, E.; Demetriadou, A.; Fox, P.; Hess, O.; Baumberg, J. J. Single-molecule strong coupling at room temperature in plasmonic nanocavities. Nature 2016, 535, 127−130. (54) Chikkaraddy, R.; Turek, V. A.; Kongsuwan, N.; Benz, F.; Carnegie, C.; van de Goor, T.; de Nijs, B.; Demetriadou, A.; Hess, O.; Keyser, U. F.; Baumberg, J. J. Mapping Nanoscale Hotspots with Single-Molecule Emitters Assembled into Plasmonic Nanocavities Using DNA Origami. Nano Lett. 2018, 18, 405−411. (55) Roelli, P.; Galland, C.; Piro, N.; Kippenberg, T. J. Molecular cavity optomechanics as a theory of plasmon-enhanced Raman scattering. Nat. Nanotechnol. 2016, 11, 164−169. (56) Benz, F.; Schmidt, M. K.; Dreismann, A.; Chikkaraddy, R.; Zhang, Y.; Demetriadou, A.; Carnegie, C.; Ohadi, H.; de Nijs, B.; Esteban, R.; Aizpurua, J.; Baumberg, J. J. Single-molecule optomechanics in “picocavities”. Science 2016, 354, 726−729. (57) Olson, A. P.; Ertsgaard, C. T.; Elliott, S. N.; Lindquist, N. C. Super-Resolution Chemical Imaging with Plasmonic Substrates. ACS Photonics 2016, 3, 329−336. (58) Olson, A. P.; Spies, K. B.; Browning, A. C.; Soneral, P. A. G.; Lindquist, N. C. Chemically imaging bacteria with super-resolution SERS on ultra-thin silver substrates. Sci. Rep. 2017, 7, 9135.

I

DOI: 10.1021/acs.accounts.8b00563 Acc. Chem. Res. XXXX, XXX, XXX−XXX