Photoinduced Plasmon-Driven Chemistry in trans-1,2-Bis(4-pyridyl)ethylene Gold Nanosphere OligomersEmily A. Sprague-Klein,† Bogdan Negru,‡,|| Lindsey R. Madison,‡,ξ Scott C. Coste,‡

Brandon K. Rugg,‡ Alanna M. Felts,§ Michael O. McAnally,‡,Γ Mayukh Banik,§ Vartkess A. Apkarian,§

Michael R. Wasielewski,‡ Mark A. Ratner,‡ Tamar Seideman,†,‡ George C. Schatz,†,‡

and Richard P. Van Duyne*,†,‡

†Applied Physics Graduate Program and ‡Department of Chemistry, Northwestern University, Evanston, Illinois 60208, UnitedStates§Department of Chemistry, University of California, Irvine, California 92697, United States

*S Supporting Information

ABSTRACT: Continuous wave (CW) pump−probe surface-enhanced Ramanspectroscopy (SERS) is used to examine a range of plasmon-driven chemicalbehavior in the molecular SERS signal of trans-1,2-bis(4-pyridyl)ethylene (BPE)adsorbed on individual Au nanosphere oligomers (viz., dimers, trimers, tetramers,etc.). Well-defined new transient modes are caused by high fluence CW pumping at532 nm and are monitored on the seconds time scale using a low intensity CWprobe field at 785 nm. Comparison of time-dependent density functional theory(TD-DFT) calculations with the experimental data leads to the conclusion thatthree independent chemical processes are operative: (1) plasmon-driven electrontransfer to form the BPE anion radical; (2) BPE hopping between two adsorptionsites; and (3) trans-to-cis-BPE isomerization. Resonance Raman and electronparamagnetic resonance (EPR) spectroscopy measurements provide furthersubstantiation for the observation of an anion radical species formed via aplasmon-driven electron transfer reaction. Applications of these findings will greatly impact the design of novel plasmonicdevices with the future ability to harness new and efficient energetic pathways for both chemical transformation andphotocatalysis at the nanoscale level.

■ INTRODUCTION

Noble metal nanoparticles have the unique capability ofharnessing and confining light below the diffraction limit dueto the field-enhancing capabilities created by plasmonexcitation that leads to “hot spot” junctions.1 Localized surfaceplasmon resonance (LSPR) refers to a photoinduced oscillationof the conduction electrons within a nanoparticle. The resultingelectromagnetic field that is experienced by a molecule near thesurface strongly depends on various parameters such as theexcitation wavelength and the physical properties of thenanoparticle as understood by Mie scattering theory.2

Consequently the resulting local enhancement of incident andemitted electromagnetic fields can be utilized to amplify Ramanprocesses of molecules forming the basis of plasmonicallyenhanced vibrational spectroscopic techniques such as surface-enhanced Raman spectroscopy (SERS).3,4 Within the hot spot,the SERS signal is very sensitive to molecular geometry, wheresmall changes in bond length, such as those that arise fromchanges in the population of an antibonding orbital, can lead tosignificant shifts in the vibrational spectrum. This leads toenhancement of the Raman signal that results not only from theplasmon-induced electromagnetic mechanism (EM)5,6 but alsofrom a chemical mechanism (CHEM)7−9 that depends on

charge transfer between molecule and substrate. Reports ofintensity fluctuations and few cm−1 spectral wandering arecommon in the SERS literature.10−21 Given the tensor nature ofRaman scattering, the observable spectra in the single moleculelimit can depend on location and orientation of the moleculerelative to local field and field gradients.22,23 Under strongconfinement, SERS is determined by the atomistic near field;therefore, fluctuations can be driven by atomic scale variationsin junction morphology.24−29 A majority of these intensity andspectral wandering fluctuations are explained by a combinationof EM, CHEM, and thermal contributions. However, aquantitative and comprehensive understanding of thesefluctuations in terms of molecular scale structural information,electrodynamics, and especially charge transfer remains to bedeveloped.19,30−32

Our work aims to understand a range of nonresonantphotodriven chemical transformations of the molecule locatedin optically confined near-field plasmonic nanogaps. The effectsinclude photocatalytic plasmon-driven electron transfer, photo-induced molecular absorption and binding properties, and cis−

Received: June 16, 2018Published: August 2, 2018

Article

pubs.acs.org/JACSCite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

© XXXX American Chemical Society A DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Dow

nloa

ded

via

NO

RT

HW

EST

ER

N U

NIV

on

Aug

ust 1

4, 2

018

at 1

4:38

:40

(UT

C).

Se

e ht

tps:

//pub

s.ac

s.or

g/sh

arin

ggui

delin

es f

or o

ptio

ns o

n ho

w to

legi

timat

ely

shar

e pu

blis

hed

artic

les.

trans isomerization of the molecule trans-1,2-bis(4-pyridyl)-ethylene (BPE) within 90 nm gold nanosphere oligomers.Elucidating the nature of plasmon-driven chemical behaviorcould have profound impact on LSPR-induced photochemicalcatalysis,33,34 solar cells and photovoltaics,35−37 and metal−molecule−metal junction devices.38−40 Understanding how totune charge transfer resonances can help in efficientlyengineering novel materials and devices capable of controlledexcitation and relaxation pathways.34 Moreover, examininginterfacial dynamics between molecules and plasmonic nano-particles has important applications in areas of devicefabrication,35−40 sensor design,41−45 metal−organic frame-works (MOFs),46,47 and molecular electronics.38−40

Several recent experiments have identified chemical effects,namely that of charge transfer, from the overall SERS signalusing molecular off-resonance excitation.31,32 Park and Kim48

utilized molecular off-resonance excitation of gold nano-particle−4-aminobenzenethiol (ABT)−gold film junctions toanalyze mode specific on-and-off fluctuation ratios to determinecharge transfer enhancement factors. They found that a smallpercentage of molecules within the junction exhibited chargetransfer enhancement in SERS on the order of 101−103, while

most molecules remained charge transfer inactive. Additionalcharge transfer systems have been studied in the SERS signalfluctuations of Au bowtie nanoantennas,49 finite-differencetime-domain (FDTD) modeling of the local chemical environ-ment in SERS fluctuations using nanogaps,50 and temporalfluctuations in the tip-enhanced Raman scattering (TERS)signal of adenine molecules.51 Other reports have catalogued arange of plasmon-driven chemical behavior52 in nanostructuredsystems that include light-triggered polymerization,53,54 watersplitting,55,56 and photocatalysis.57−60

In this paper, we propose several chemical processes toexplain the origins of transient nonresonant molecular spectralactivity in plasmonic nanogaps. Other polypyridine molecules,such as 4,4′-bipyridine31 and methyl viologen dichloride,32 havebeen examined previously for their role in plasmon-drivenelectron transfer reaction. However, extensive consideration ofother chemical processes in these highly complex and fluxionalnanogap systems has not been addressed. Our work thereforeseeks to reinforce the initial observation of a plasmon-drivenelectron transfer reaction through direct vibrational detection,and to introduce additional chemical processes (photoinducedmolecular absorption site-hopping and trans−cis isomerization)

Figure 1. Independent chemical processes for plasmon-driven chemical behavior. (A) 2D plot showing the appearance of transient modes (arrowsand highlighted modes) beginning in frame 183 with a prominent peak at 1553 cm−1. (B) Top panel shows neutral BPE signal at time zero, middlepanel shows an example spectrum containing the transient 1553 cm−1 mode, and bottom panel shows all frames (gray) with time-averaged spectrum(purple) C) 2D plot showing the appearance of transient modes in the 1570−1590 cm−1 region beginning in frame 322. (D) Top panel showsneutral BPE signal at time zero, middle panel shows an example spectrum containing the transient 1588 cm−1 mode, and bottom panel shows allframes (gray) with time-averaged spectrum (purple). (E) 2D plot showing the appearance of transient mode splitting in the 900−1300 cm−1 regionbeginning at frame 68. (F) Top panel shows neutral BPE signal at time zero, middle panel shows an example spectrum containing the transient peaksplitting at 1222 and 1205 cm−1 and at 1020 and 1040 cm−1, bottom panel shows all frames (gray) with time-averaged spectrum (purple).

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

B

that aid in the nuanced understanding of plasmon-drivenchemical reactions within nanogaps.

■ EXPERIMENTAL SECTIONSingle Particle Sample Preparation. Nanosphere oligomer

substrates with BPE functionalization were prepared both with andwithout an outer silica shell. Samples containing a protective silica shellare commercially available (STA Technologies Inc., Albuquerque,NM), and their spectroscopic properties have been previously studiedusing a range of techniques.61−64 Nanosphere oligomers without asilica shell were fabricated using a methodology that utilizes water-soluble polymers as an aggregation stabilizer with the BPE moleculesrandomly oriented in the nanogap.65 Nanosphere solutions werediluted to single particle concentration and drop-casted onto carbon/Formvar TEM grids (Ted Pella, Inc., Redding, CA).63,64 In this way,particles may be easily located and indexed for correlated spectroscopyand imaging studies.Microscope glass coverslips (VWR International, no.1, 25 mm

diameter) underwent piranha treatment (3:1 H2SO4:30%H2O2)followed by thorough rinsing with Milli-Q water (18.2 MΩ/cm).Coverslips were then immersed in a base treatment (5:1:1H2O:NH4OH:30%H2O2), sonicated for 60 min, and again thoroughlyrinsed with Milli-Q water. Sample grids were positioned on top of N2-dried coverslips and mounted inside a sample stage above themicroscope objective.Single Particle CW Pump−Probe SERS Measurements.

During measurements, silica-encapsulated particles were imagedusing dark-field optical microscopy with a dry condenser and a 40×/0.60 air microscope objective (Nikon Corporation) on an invertedmicroscope (Nikon Ti-U/E20L80) as shown in Figure S1. Dark-fieldimaging allows easy identification of single nanoparticle aggregates dueto their small metallic cores being great scatters of light, which wereimaged using a CCD camera (UNIQ Vision, Inc., Santa Clara, CA)attached to an auxiliary microscope port. The two-color experimentalsetup allowed for simultaneous illumination with a 785 nm CW laser(Renishaw Inc.) and 532 nm CW laser (Spectra-Physics, Santa Clara,CA). The 785 nm laser was used to monitor the BPE nanosphereassemblies by collecting low power SERS spectra (power density at thesample = 0.34 μW/μm2), while the 532 nm laser was used for highpower illumination (power density at the sample = 118 μW/μm2). A10:90 beamsplitter (ThorLabs, Newton, NJ) was used to couple thegreen and red laser beams in such a way that 90% of the green and 10%of the red photons were incident in the microscope focal plane. Theexperimental setup also features a USB shutter (Picard Industries,Albion, NY) controlled by monitoring the NOT READY output of thePIXIS 400 CCD camera mounted on a Spectra Pro 2500i spectrometer(Princeton Instruments, Trenton, NJ). A shutter allowed forsegmented illumination of the BPE nanosphere assemblies and wasincorporated into this setup to mitigate the large background causedby the 532 nm laser. The sample was exposed to green light for 10 sintervals between acquisitions with red light. A common sequence fordata acquisition consisted of (1) collecting a single frame Ramanspectrum at 785 nm for 10 s, (2) exposing the sample to the 532 nmlaser for 10 s while the camera is not accumulating a spectrum, then(3) collecting a second Raman spectrum at 785 nm for 10 s. Dataacquisition was automated with WinSpec (Princeton Instruments,Trenton, NJ) while the interface to the shutter was controlled using acustom LabVIEW virtual instrument (National Instruments, Austin,TX). Figure 1b presents a timing diagram depicting the alternationbetween 532 nm pump and CCD exposure for spectral acquisition.Single particle measurements of nanosphere oligomers without a

silica shell were carried out with a 100× oil immersion objective (NA =0.8−1.3) using an identical microscope setup. SERS spectra with highsignal-to-noise ratios were obtained using this optimized setup, whichallowed for single particle detection of in-house fabricated nanoparticleoligomers.65 A timing scheme of 1 s pump and 1 s probe wasimplemented, resulting in increased temporal information comparedto our previous work.31

Hundreds of SERS spectra were acquired from 36 BPE nanosphereassemblies. More than half of the illuminated particles showed spectracontaining unexpected fluctuations. These fluctuations are observed inthe 700 to 1800 cm−1

fingerprint region of the vibrational spectrumand resemble narrow peaks reminiscent of Raman-active lines ofvibrational modes. The fluctuations reported here are not broadbackground oscillations but rather narrow discrete peaks that are eithervisible for multiple frames or for single frames collected over tens ofseconds. The data sets shown in this report focus on four nanosphereoligomers that expressed transient vibrational modes in the 700 to1800 cm−1 spectral region that cannot be explained by the neutralnormal Raman modes of BPE. All nanosphere oligomers reported inthis study exhibited strong reporter molecule signal throughout thecourse of data acquisition with minimal neutral BPE signal decayacross time.

Transmission Electron Microscopy (TEM) Imaging. Followingspectroscopy, indexed single nanosphere oligomers were correlatedand imaged using transmission electron microscopy (TEM). Scanswere performed on a Hitachi 8100-TEM using 200 kV of thermionicemission and optimized aperture positions (condenser = 1, objective =2, selected area = 0).

Electrochemical Measurements. Cyclic voltammetry (CV) wasmeasured under a dinitrogen atmosphere at room temperature in anMBraun UniLab Pro glovebox using a CHI 760C potentiostat. A three-electrode electrochemical cell was used with a 2 mm diameter Pt diskworking electrode, a 0.5 mm thick Pt wire counter electrode, and a Agwire quasi-reference electrode (QRE) referenced to ferrocene/ferrocenium. The analyte was a 10 mM solution of BPE in 10 mL of0.1 M tetra-n-butylammonium hexafluorophosphate (TBAPF6) in dry,deoxygenated THF.

Electron Paramagnetic Resonance Spectroscopy (EPR). Thecontinuous wave electron paramagnetic resonance (EPR) spectrum ofthe chemically reduced solution-phase BPE anion radical was obtainedat X-band (∼9.8 GHz) frequency using a Bruker Elexsys E680spectrometer equipped with an ER 4118X-MS5 resonator. Scans wereperformed with a magnetic field modulation amplitude of 0.5 G andnonsaturating microwave power of 0.1 mW. The sample solution wascontained in a quartz tube with I.D. 1.50 mm and O.D. 1.80 mm. Theexperimental spectrum was fit using EasySpin66 v5.2.10 in MATLAB.

Visible-NIR Characterization of Neutral and ReducedSpecies. A spectrophotometer in double beam mode with solventand baseline/zero correction (Cary 5000, Agilent Technologies) wasused to characterize the neutral, singly reduced, and doubly reducedBPE species in solution. Samples were first prepared in a dinitrogenglovebox as described for electrochemical measurements and thentransported in sealed quartz cuvettes for spectroscopy measurements.A scanning range of 400−1000 nm was used.

Chemical Reduction of BPE and Resonance Raman ofReduced Species. BPE was selectively reduced to the anion radical(E1/2 = −2.186 V vs Fc/Fc+) by reacting an equimolar amount of BPEin THF with a 0.1 M THF solution of sodium anthracenide; thesolution was dark yellow in appearance. BPE was further reduced tothe dianion (E1/2 = −2.756 V vs Fc/Fc+) using an excess of potassiummetal in THF followed by filtration through glass wool, yielding anopaque purple solution.

An automated wavelength scanned measurement was performedusing a tunable narrow line width continuous-wave Ti:sapphire laser(M Squared, Glasgow, United Kingdom). An excitation wavelength of785 nm was determined to be the optimal wavelength for Ramancharacterization with minimal background fluorescence while stillmaintaining resonantly enhanced signal. A 785 nm CW laser(Innovative Photonics) was focused onto the quartz cuvette, andRaman scattering was collected in a backscattering geometry.Collection light was focused into a 1/2 m imaging spectrograph(Acton SP2500i, Princeton Instruments) with thermoelectricallycooled CCD detector (PIXIS 400BR, Princeton Instruments).

Theoretical Calculations. All calculations were performed usingthe Amsterdam Density Functional (ADF)67 and NWChem68

programs. The triple ς with polarization (TZP) basis set and theBecke−Perdew (BP86)69,70 exchange correlation functional were used.

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

C

The BP86 functional was used because it results in harmonicvibrational frequencies that are similar to experimental results withoutscaling.71 For nitrogen, carbon, and hydrogen, all electrons weremodeled while for gold, electrons in atomic orbitals up to the 4d werefrozen. For systems that included the gold clusters, relativistic effectswere modeled with the zero-order regular approximation (ZORA).72,73

Because the oxidized and reduced forms of BPE have an oddnumber of electrons, calculations of the real and imaginarycomponents of the polarizability tensors were performed using amethod described by Aquino and Schatz74 and implemented in theAOResponse module of the NWChem software. For comparison, theclosed shell, neutral BPE systems were calculated with theAOResponse modules in both NWChem and ADF softwarepackages,67,75 following the method described by Jensen et al.76 TheAOResponse method uses the short time approximations proposed byLee and co-workers,77−79 to the Kramers−Heisenberg−Dirac (KHD)formulation of resonance Raman scattering. The wavelength used forthese calculations was 785 nm, consistent with the wavelength of theCW laser used in the experiment.The resonance Raman scattering cross section is calculated from the

derivative of the isotropic polarizability tensor, αp′ as follows:

σ πε

ω ωπ ω ω

∂∂Ω

= −− −

hc

Shc k T

( )8

( )1

45(1 exp( / ))pp p

2

02

42 p

B

(6.1)

where ω is the frequency of the incident field, and ωp is the frequencyof the pth vibrational mode. The scattering factor, Sp, is a molecularproperty composed of α′ and γ′, which are the isotropic andanisotropic polarizability derivatives. The scattering factor is calculatedusing

α γ= ′ + ′S 45 7p p2

p2

(6.2)

where the isotropic polarizability derivative is

∑α α ′ = ′13

( )i

iip p(6.3)

and the anisotropic polarizability derivative:

∑γ α α α α′ = ′ ′ − ′ ′12

3( ) ( ) ( ) ( )i j

ij ij ii iip,

p p p p(6.4)

Following the calculation of the intensity of the Raman scatteringfor all modes, the spectrum is broadened with a Lorentzian functionwith a full-width, half-max of 20 cm−1, for easier comparison toexperiment.74,80

■ RESULTSPlasmon-Driven Chemistry. Across all four nanosphere

oligomers studied, three independent reversible chemical processesof plasmon-driven chemical behavior were identified from transientspectral patterns in three differing wavenumber regions: (1) plasmon-driven electron transfer to form the anion radical species (2) molecularhopping (3) isomerization. Figure 1 displays spectral fluctuations fromthe 900−1700 cm−1

fingerprint region. Time-dependent SERS data areshown in the 2D plots for Figures 1A, 1C, and 1E. The neutral spectraof BPE are present in all frames. Throughout the course of scanning,transient modes appear and disappear that cannot be explained by thenormal modes of neutral BPE (see Table S1). In between eachacquisition frame (10 s each) at 785 nm excitation, the nanosphereoligomers are illuminated with high intensity 532 nm light for 10 s at atime. In Figure 1A, transient modes (e.g., 1553 cm−1) are observedbeginning in frame 183. Figure 1B compares a neutral SERS spectrumobtained from a BPE nanosphere oligomer before illumination with532 nm laser light and a SERS spectrum obtained from the samenanosphere assembly after 16.5 total minutes of 532 nm laser lightexposure. The spectrum obtained after irradiation with green laser lightshows peaks consistent with the vibrational spectrum of surfaceadsorbed BPE molecules superimposed with intense peaks at 920,

1103, 1129, 1157, 1234, 1311, 1322, and 1553 cm−1. The bottompanel of Figure 1B shows all of the spectra obtained from a nanosphereoligomer superimposed with the average of all spectra. This averagespectrum closely resembles the SERS spectrum of BPE. Similarly, inFigures 1D and 1F, the time-averaged SERS spectrum resembles BPE,suggesting that plasmon-driven chemical transformation is restricted toonly a small percentage of molecules residing within the “hotspot”.31,32,81 Figure 1C and 1D show time-dependent SERS data fora nanosphere oligomer with an intense peak at 1583 cm−1 and Figures1E and 1F show a different nanosphere oligomer with peak splitting at1222 and 1206 cm−1 as well as at 1040 and 1020 cm−1.

As seen from Figure 1, the drastic and varied shifts in vibrationalenergies imply that the chemical processes taking place aretransformational, leading to the formation of entirely new chemicalspecies. The formation and consumption of a chemical product on thegold surface can explain the synchronized appearance and disappear-ance of spectral peaks as seen in Figures S2 and S3 in SupportingInformation. A change in surface adsorption geometry of the BPEmolecules could result in few wavenumber shifts, and it would mostlikely vary the relative intensity of the BPE spectral peaks,82 but it isunlikely that this would produce the large changes seen in Figure 1.The observed fluctuations are also inconsistent with thermaldegradation of BPE, because they manifest as very narrow peaks, asopposed to previously documented broad carbon backgrounds.83

Besides thermal effects, hot electrons emitted from the gold surface canalso lead to chemical transformation. In the next sections we analyzeevidence for three kinetically favorable plasmon-driven chemicalreactions.

Chemical Process 1: Plasmon-Driven Electron Transfer.Figure 2 shows the absorption spectra of (a) neutral BPE, (b) singly

reduced BPE, and (c) doubly reduced BPE. In the neutral species, noabsorption peaks are observed in the visible region whereas for thesingly reduced case an absorption peaks are observed at 487, 569, and694 nm. Absorption peaks for the doubly reduced species are at 543and 571 nm. Neutral BPE has no absorption peaks within the visiblerange. In our experiment, the pump field overlaps with the monomerlocalized surface plasmon resonance for the gold nanosphere oligomerand the probe overlaps with the multicore gold nanosphereresonance.31,62,63,65 On a molecular level, the pump overlaps withthe 543 nm doubly reduced absorption resonance while the probeoverlaps with the broad absorption feature at 694 nm for the singlyreduced species.

To characterize the energetics and Raman spectra of reduced formsof BPE, we first obtained reduction potentials from a solution-phasecyclic voltammogram of a 10 mM solution of BPE (Figure S4). Strongreducing agents were used to generate the singly reduced species andthe doubly reduced species using sodium anthracenide and potassiummetal reduction, respectively. To confirm that the anion radical wasgenerated, an EPR measurement was taken of the singly reduced

Figure 2. Absorption spectra for neutral, singly reduced, and doublyreduced BPE in THF at room temperature. Pump field at 532 nm(green) and probe field at 785 nm (red) are indicated.

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

D

species and hyperfine coupling values were obtained from simulations(shown in Figure S5 and Table S2).Figure 3A shows SERS spectra of the trans anion radical species for

the BPE molecule absorbed on a single nanosphere oligomer alongwith solution resonance Raman spectra and results from TDDFTcalculations. SERS spectra on single nanosphere oligomers are shownwith the initial neutral BPE spectrum subtracted out to highlighttransient modes. The comparisons of theory and experiment aresimilar to our previously published work for the molecule 4,4′-bipyridine (BPY),31 and they confirm the assignment of the trans anionradical. The SERS data includes results for three different nanosphereoligomers, including two with an encapsulating silica shell and onewithout a silica shell. Figure S6 shows the time-dependent data setobtained from a nanosphere oligomer (Figure 3A, trace c). Initially, in

earlier frames, a strong BPE signal was observed (Figure 1A and 1B).The intensity of the BPE signal slowly decreases, and beginning inframe 184 the appearance of new spectral peaks takes place(highlighted in red, Figure S6). The arrows in Figure S7 indicate theconcerted appearance and disappearance of some of these new peaks,as well as their persistence from frame to frame. A very intense peakcentered around 1550 cm−1 appears and disappears, taking place onthe order of tens of seconds. Automated spectral analysis techniqueshave been employed to address the highly diverse and dynamicappearance of these spectral fluctuations (as seen in Figures S2 andS3).

Spectrum d from Figure 3A shows the solution-phase resonanceRaman spectra for the BPE single reduced species, or anion radical.Additional resonance Raman spectra can be found in Figure S8. In the

Figure 3. (A) Anion radical (singly reduced) BPE spectra across three nanosphere oligomers (a−c), resonance Raman spectrum of solution-phasesingly reduced BPE (d), and gas-phase open-shell calculation of the radical BPE anion (e). (B) TEM image of nanosphere oligomer correlated to theSERS spectrum (a) shown at left. (C) Spectral variance for the 1550 cm−1 mode observed across the three nanosphere oligomers. *Indicates residualneutral BPE modes due to imperfect subtraction.

Figure 4. Schematic of additional chemical processes. (A) Varied molecular absorption from face- to vertex-binding and (B) trans−cis isomerization.

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

E

solution phase, the anion radical BPE modes are identified at 1603,1568, 1428, 1335, 1277, 1246, 1197, 1044, 908, and 872 cm−1.Likewise, for the doubly reduced species the vibrational modes occurat 1649, 1602, 1291, 1235, 1200, 1027, 993, 921, and 907 cm−1. The1409 cm−1 mode in the spectrum for the singly reduced speciescoincides with the most intense normal Raman mode of anthracene,84

which is a byproduct from the chemical reduction using sodiumanthracenide. Overall, anthracene does not contribute prominently tothe Raman signal of the singly reduced BPE species. Good agreementexists between the solution-phase anion radical spectrum (singlyreduced) species with that shown in Figure 3A for the BPE anionradical spectrum produced through a plasmon-driven electron transferprocess in the solid state. The variance in the peak intensities of anionradical modes on nanosphere oligomers when compared to normalRaman solution measurements can be attributed to differences inmolecule−surface interactions such as those that might arise from achanging local potential85 or the excitation of dark plasmon modes.63

The BPE anion radical peaks can be assigned to a range ofvibrational motions such as pyridine ring distortion modes andquinodal modes coupled to the ethylene stretch (see Tables S3 andS4). Figure 3c shows a spectral distribution for the 1550 cm−1 mode,where peak distributions have a wider variance for the silica shell case(15−20 cm−1 fwhm) when compared to without a silica shell (10 cm−1

fwhm). Changes to the local dielectric host environment for the anionradical due to the presence or lack of silica shell could account fordifferences in the spectral variance.86−88 Figure 3B displays a TEMimage correlated with the SERS spectra shown in Figure 3A (top,spectrum a).Additional Chemical Processes: Molecular Hopping and

Isomerization. Contributions from two additional processes,molecular hopping and isomerization, that are observed in BPEnanosphere oligomers can be found in Figure 4. The various specieshave been identified through corroboration with open-shell TDDFTcalculations and vibrational mode assignments can be found in TablesS5 and S6. Here we see a peak near 1580 cm−1 in a−d, which contrastswith the absence of a peak in Figure 4 e,f showing that the TDDFTspectra associated with two binding states on a Au25 gold cluster(vertex and face, as pictured) are in good agreement with theexperimental spectral intensity near 1580 cm−1. This contrasts with theunbound or symmetrically bound case (see Figure S9), where the 1580cm−1 mode is not present in TDDFT calculations. Thus, the 1580cm−1 vibrational mode appears only when symmetry breaking occurs;in other words, when the BPE molecule is bound at one end to either aface or vertex site on the Au25 cluster. Another feature of the spectranear 1580 cm−1 in Figure 4A is flipping behavior that can beunderstood as a transition between the two binding states on a goldcluster where both spectra a and d appear to show the BPE moleculebinding to a vertex site whereas spectra b and c show face-binding.Figure 4B also shows similar flipping behavior between two isomer

states, although the spectral regions involved are quite differentcompared to Figure 4A. Spectral doublets occurring at 1222 and 1205cm−1 as well as at 1022 and 1044 cm−1 are the defining features for aconfigurational change from the trans- to neutral cis-isomer. Figure S10shows comparison between experimental spectra and TDDFTcalculations for the combined effect of isomerization and plasmon-driven electron transfer to form the cis-anion. Defining features of thecis-anion include triplet peaks at 1454, 1429, and 1407 cm−1, inaddition to peaks at 1315, 1285, 1213, and 1016 cm−1. The relativecontributions of each chemical process as well as their macroscopickinetic time scales are addressed in the following section.

■ DISCUSSIONGeneration of Chemical Product and Surface Elec-

trons. The novel pathway to photochemistry on metal surfacesis provided through metal−molecule interfacial states, wherebya metal electron can be promoted to electronically excited statesof the molecule that can undergo chemistry. The spectralsignatures for the anion radical created through a plasmon-driven electron transfer reaction are reproducible, with peaks

appearing in multiple frames on the seconds time scale. Of thenew modes that appear, the 1551 and 1432 cm−1 modes arehighly reproducible while modes near 957 and 1303 cm−1

regions also appear in multiple frames. Comparisons ofexperiment and calculations of the anion radical yield excellentagreement, especially in the 1450−1550 cm−1 region, where noother source of vibrational peaks could be identified. Anegatively charged state of BPE would cause significant shiftsof the BPE Raman modes toward lower frequencies, which haspreviously been observed for other molecular systems, such asTTF,89 TCNQ,90 and most recently BPY-h8/d8

31 absorbed ongold nanosphere oligomers. Observation of the anion radical ispossible only if the charged species undergoes significantenhancement within the “hot spot” junction. It is known thatthe most enhancing sites in SERS contribute to 24% of theoverall SERS intensity while making up only 0.006% of totalsites.81 Observation of the blinking behavior of BPE anionradical signal is plausible and can be attributed to formation anddecay of the anion radicals and to diffusion into and out ofhighly enhancing sites on the nanosphere assemblies. Theintensity fluctuation of the new spectral features does notcorrelate with intensity fluctuations of the BPE signal, which isfurther indication that only a very small percentage of BPEmolecules undergo chemical change. Recent studies suggestthat plasmon-driven electron transfer is essentially a singlemolecule event, affecting only a tiny fraction of moleculeslocalized or “trapped” in a highly enhancing nanogapregion.31,32

The 532 nm generated electrons interact with surfaceadsorbed BPE molecules and activate the ensuing chemistryand reactive decay of BPE molecules. Electrons that trigger theensuing chemistry can be generated at the gold surface via twoprocesses: (1) Upon radiation at 532 nm, the LSPR mode ofgold monomers and to a lesser extent91,92 interband transitionof the gold metal are excited and result in the release of hotelectrons from the metal surface. The low energy requirementsof plasmon resonances have resulted in many recentlypublished reports that focus on plasmon-driven chemicalprocesses.93−98 Plasmon-excited nanoparticles have beenshown to be an efficient source of hot electrons92,99,100 whenthe plasmon quantum excites an electron from the Fermi levelof the metal to a nearby electron acceptor.101 In this way, hotelectrons can transfer into electronic states of adsorbed specieswith high quantum efficiencies.102 (2) Additionally, transientreflecting grating spectroscopy studies employing a 1 kHzultrafast laser determined that most chemisorbed crystal violetmolecules received charge transfer electrons from the metalupon excitation with green laser light close in energy to theinterband transition of gold.103 The charge transfer was possiblydue to the excitation of valence electrons residing close to theFermi level to adsorption energy states of the molecules.103

The theoretical resonance Raman spectra of the trans-BPEconformer of the anion radical shows significant spectraldifferences compared to the neutral form as demonstrated inFigure 5. The stretching vibration of the weaker ethylene bondin the anion radical of BPE also couples to the quinoidal-likering distortion at 1550 cm−1 which is similar in mode characterto the 1630 cm−1 in neutral BPE. A similar shift is observedgoing from the 1300 cm−1 neutral BPE mode to the 1240 cm−1

anion radical BPE Kekule mode. The 1028 cm−1 anion radicalmodes exhibits ring distortion via asymmetric bond stretching.A similar shift is observed for the anion radical of the cisconfiguration of BPE as seen previously in Figure S10. As

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

F

shown in the molecular orbital diagram in Figure 5, the LUMOof the anion radical has antibonding character for the ethyleneC = C bond and certain bonds of the pyridine rings, leading tothe reduced frequencies for many modes. Our gas-phaseelectron affinity calculations reveal that the BPE anion has astabilization energy of 1.5 eV when compared to neutral BPE,which means that the BPE anion will not undergo degradationafter electron transfer. Furthermore, while BPE molecules arenot resonantly enhanced at 785 nm, the BPE anion radical isresonantly enhanced, resulting in single molecule sensitivity andthe feasibility of detecting single reduction events. Theformation of a single BPE anion radical would not cause adetectable jump in the BPE signal intensity, but it would lead tothe addition of new peaks corresponding to the anion radical,peaks with intensities comparable to the BPE signal. The smallsize of the ”hot spot”, small number of BPE anion radicals, andsurface diffusion of these anion radical species to and from thehot spot leads to the observed intermittent detection of peakscorresponding to the BPE anion radical.Vibrational Mode Shifting and Charge Accumulation.

Systematic time-dependent spectral shifts were observed innanosphere oligomers undergoing a plasmon-driven electrontransfer reaction. Figure 6 shows data from two differentnanosphere oligomers in which the SERS spectra for the anionradical is observed, along with a systematic Stark shift of thevibrational modes. The 1550 cm−1 mode, assigned to the anionradical quinodal modes, shifts 13 cm−1 after 18 successiveintervals of pump “on” and “off”. Similarly, the 1588 cm−1

assigned to molecule−surface binding, shifts 7 cm−1 after eightsuccessive intervals of pumping time. Previous studies havedemonstrated that molecules are capable of directly measuringthe modulation of local electromagnetic fields due to electrontunneling in plasmonic nanogaps.104,105 The resulting time-dependent vibrational shifts are due to a buildup of charge dueto optical rectification in strongly coupled nanoparticle systems.A recent study on charge buildup of Au nanoparticlesconcluded that steady-state light excitation can lead tophotoelectrochemical potentials as high as 240 mV.106 With

this in mind, it is possible for the silica coating of thenanosphere assemblies to insulate against charge buildup on thegold surface. Future experiments will test this hypothesis bymonitoring the LSPR of individual nanosphere assembliesunder high power green laser light illumination.107 The BPEanion radical signal intensity flickers in and out within the timeresolution of the experiments carried within, making itimpossible to distinguish between electron recombination,photothermal desorption, surface diffusion, and irreversibledegradation. It is, however, possible that the flicker in intensityof the anion radical signal is due to the presence of a singleanion radical within the “hot spot”. This situation would resultin large intensity fluctuations depending on the location withinthe “hot spot”, as well as relative fluctuations in intensitybetween different modes of the BPE anion radical.82 Futureexperiments will aim at increasing the time resolution of theexperiment, so that we may produce kinetic models for the BPEanion radical signal loss.

Chemical Time Scales and Relative Contribution byProcess. A timeline of chemical reaction dynamics for thethree operative processes can be seen in Figure S11, whichshows the timeline of first appearance of anion radical modes inboth silica-coated and uncoated gold nanosphere oligomers.The first anion radical modes appear after 3 min of opticalpump time in particles encapsulated with a silica shell whereasthe delay time to first appearance of anion radical modes ismuch shorter in oligomers with no silica shell. However, signalstability of the anion radical is pronounceably lower and moreerratic in oligomers with no silica shell. In comparison,molecular site hopping is a delayed process which is notphotoinitiated immediately and persists for many framesafterward (3−6 min of optical pumping). Isomerizationprocesses occur on a relatively shorter time scale, after 170 sof pump time. Because each spectrum is collected on theseconds time scale, the chemical products must be long-lived.Reconciling the long-lived chemical reaction dynamics with theshort-lived hot electron dynamics is still an area requiringfurther understanding,32,108 and future studies are currentlyunderway to investigate reaction intermediates on increasinglyshorter time scales. Figure 7 shows the distribution of events bychemical process and the relative contributions of the threeeffects in explaining the spectral fluctuations in single

Figure 5. Resonant Raman scattering spectra comparing the neutraland anion radical forms of trans conformers of BPE upon excitationwith 785 nm light.

Figure 6. (A) Vibrational mode shifting in a nanosphere oligomer withcorrelated TEM image. (B) Additional shifting in a second nanosphereoligomer.

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

G

nanosphere oligomer SERS. Automated peak fitting inMATLAB was employed to extract spectral locations for allmodes in our time-dependent SERS data. In our analysis, onecount corresponds to the observance of a vibrational peak inone frame and does not refer to peak intensity information.Figure 7A shows frequency locations for peaks correspondingto the neutral BPE molecule only. Figure 7B shows all thetransient modes arising from an intense 532 nm pump field withspectral information from the neutral BPE molecule subtracted.An enlarged version of the transient mode histogram can beseen in Figure 7C. Several spectra regions can be assigned to thecoexistence of multiple chemical processes, as indicated byadditional symbols. A total of 3612 transient spectral modeswere collected across four differing nanosphere oligomers in the900−1700 cm−1

fingerprint region. Out of that total, we haveidentified and assigned a chemical process (trans anion radical,molecular binding, cis isomerization, and the cis anion radical)for spectral regions corresponding to 2075 of those transientmodes. Counts due to contribution from residual neutral BPEsignal amounts to 914 spectral counts. As a result, ourmechanistic analysis can account for 84% (2989 out of 3612)of all transient spectral features in the time-dependent SERSdata. Decomposition of BPE into moderately stable molecularspecies was also considered; however, none of the theoreticallycalculated modes can account for the transient spectral modesobserved in our SERS data set (Figure S14). The remainingspectral features could be explained by the excitation of fieldgradients, which has been addressed in Supporting InformationFigure S15.

■ CONCLUSIONIn this study, we investigate highly complex spectral fluctuationspresent in CW pump−probe SERS data on single nanosphereoligomers. Vibrational signatures for the anion radical areobserved for BPE, a molecule for which the anion radicalspectrum is previously unknown. The existence of the BPEanion radical is confirmed using TDDFT calculations of theharmonic vibrations and resonance Raman scattering crosssections. Additionally, the BPE anion radical was generated insolution using strong chemical reducing agents, and probedusing both resonance Raman and CW EPR measurements tofurther substantiate the vibrational signature for the anionradical in the solid state. In addition to the formation of ananion radical species, other chemical processes were consid-ered, allowing us to account for 84% of observed transientspectral behavior and chemical transformations occurringwithin strongly coupled nanogaps. Applications of thesefindings have the potential to greatly impact all areas of organicmolecular electronics, photocatalytic nanoreactors, and photo-voltaic design. Future studies will aim to increase the temporalresolution of the experiment to gain knowledge about chemicalintermediates as well as probe the spectral range of these light-triggered plasmon-driven chemical reactions.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/jacs.8b06347.

Experimental schematic and electronics timing diagram,automated peak fitting analysis, vibrational mode assign-ments using TDDFT, solution-phase cyclic voltammo-gram and EPR trace, hyperfine values, bimodaldistribution of anion radical modes, kinetic models ofneutral BPE signal loss, calculated photodegradationproducts, and field gradient analysis of spectral patterns(PDF)

■ AUTHOR INFORMATIONCorresponding Author*vanduyne@northwestern.eduORCIDEmily A. Sprague-Klein: 0000-0002-1179-397XBogdan Negru: 0000-0002-6387-7286Michael O. McAnally: 0000-0002-8681-2952Vartkess A. Apkarian: 0000-0002-7648-5230Michael R. Wasielewski: 0000-0003-2920-5440George C. Schatz: 0000-0001-5837-4740Richard P. Van Duyne: 0000-0001-8861-2228Present Addresses||Department of Chemistry, Sonoma State University, RohnertPark, CA 94928.ξDepartment of Chemistry, University of Washington, Seattle,WA 98195.ΓInstitute for Defense Analyses, Alexandria, VA 22311.Author ContributionsThe manuscript was written through contributions of allauthors. All authors have given approval to the final version ofthe manuscript.Author ContributionsS.C.C. and B.K.R. contributed equally.

Figure 7. (A) Neutral BPE counts from time-dependent SERSmeasurements, aggregated across four distinct nanosphere oligomers.(B) Counts from spectral fluctuations across four distinct nanosphereoligomers with neutral BPE subtracted. (C) Enlarged version ofhistogram shown in B. Green histogram bars represent molecularbinding peaks, red represents trans anion radical, blue represents cisanion radical, purple represents cis isomer, and gray representsunidentified. Symbols represent the following: *Residual neutral BPEnormal modes; α shared modes between molecular binding and thetrans- and cis-anion; β shared modes between the trans- and cis-anion;χ shared modes between the trans-anion and molecular binding; γshared modes between the cis-isomer and cis-anion.

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

H

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe authors thank Eero Hulkko, Michael Mattei, TylerUeltschi, and Yue Wu for helpful discussions. E.A.S.-K., B.N.,M.O.M., A.M.F., M.B., V.A.A., G.C.S., and R.P.V.D. weresupported by the National Science Foundation Center forChemical Innovation dedicated to Chemistry at the Space-Time Limit (CaSTL) grant CHE-1414466. E.A.S.-K., L.R.M.,and M.O.M. acknowledge support from the National ScienceFoundation Graduate Research Fellowship Program (DGE-0824162). T.S. acknowledges support from the NationalScience Foundation grant CHE-1465201. B.N. and R.P.V.D.acknowledge support from the National Science Foundation(CHE-1506683). M.R.W. acknowledges support from theNational Science foundation under grant no. CHE-1565925.

■ REFERENCES(1) Gramotnev, D. K.; Bozhevolnyi, S. I. Nat. Photonics 2010, 4, 83.(2) Willets, K. A.; Van Duyne, R. P. Annu. Rev. Phys. Chem. 2007, 58(1), 267−297.(3) Jeanmaire, D. L.; Van Duyne, R. P. J. Electroanal. Chem. InterfacialElectrochem. 1977, 84 (1), 1−20.(4) Stiles, P. L.; Dieringer, J. A.; Shah, N. C.; Van Duyne, R. P. Annu.Rev. Anal. Chem. 2008, 1 (1), 601−626.(5) Jensen, T. R.; Malinsky, M. D.; Haynes, C. L.; Van Duyne, R. P. J.Phys. Chem. B 2000, 104 (45), 10549−10556.(6) Schatz, G. C.; Van Duyne, R. P., Electromagnetic Mechanism ofSurface-Enhanced Spectroscopy. In Handbook of Vibrational Spectros-copy; John Wiley & Sons, Ltd, 2006.(7) Moskovits, M. Rev. Mod. Phys. 1985, 57 (3), 783−826.(8) Campion, A.; Kambhampati, P. Chem. Soc. Rev. 1998, 27 (4),241−250.(9) Jensen, L.; Aikens, C. M.; Schatz, G. C. Chem. Soc. Rev. 2008, 37(5), 1061−1073.(10) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.;Dasari, R. R.; Feld, M. S. Phys. Rev. Lett. 1997, 78 (9), 1667−1670.(11) Nie, S.; Emory, S. R. Science 1997, 275 (5303), 1102.(12) Titus, E. J.; Weber, M. L.; Stranahan, S. M.; Willets, K. A. NanoLett. 2012, 12 (10), 5103−5110.(13) Rusciano, G.; De Luca, A. C.; Pesce, G.; Sasso, A.; Oliviero, G.;Amato, J.; Borbone, N.; D’Errico, S.; Piccialli, V.; Piccialli, G.; Mayol,L. Anal. Chem. 2011, 83 (17), 6849−6855.(14) Bizzarri, A. R.; Cannistraro, S. Phys. Chem. Chem. Phys. 2007, 9(39), 5315−5319.(15) Zayak, A. T.; Hu, Y. S.; Choo, H.; Bokor, J.; Cabrini, S.; Schuck,P. J.; Neaton, J. B. Phys. Rev. Lett. 2011, 106 (8), 083003.(16) Weiss, A.; Haran, G. J. Phys. Chem. B 2001, 105 (49), 12348−12354.(17) Bizzarri, A. R.; Cannistraro, S. Phys. Rev. Lett. 2005, 94 (6),068303.(18) Margueritat, J.; Bouhelier, A.; Markey, L.; Colas des Francs, G.;Dereux, A.; Lau-Truong, S.; Grand, J.; Levi, G.; Felidj, N.; Aubard, J.;Finot, E. J. Phys. Chem. C 2012, 116 (51), 26919−26923.(19) Mezni, A.; Dammak, T.; Fkiri, A.; Mlayah, A.; Abid, Y.; Smiri, L.S. J. Phys. Chem. C 2014, 118 (31), 17956−17967.(20) Maruyama, Y.; Ishikawa, M.; Futamata, M. J. Phys. Chem. B2004, 108 (2), 673−678.(21) Haran, G. Isr. J. Chem. 2004, 44 (4), 385−390.(22) Banik, M.; El-Khoury, P. Z.; Nag, A.; Rodriguez-Perez, A.;Guarrottxena, N.; Bazan, G. C.; Apkarian, V. A. ACS Nano 2012, 6(11), 10343−10354.(23) Chulhai, D. V.; Jensen, L. J. Phys. Chem. C 2013, 117 (38),19622−19631.(24) Barbry, M.; Koval, P.; Marchesin, F.; Esteban, R.; Borisov, A. G.;Aizpurua, J.; Sanchez-Portal, D. Nano Lett. 2015, 15 (5), 3410−3419.

(25) 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. Science 2016, 354 (6313),726.(26) Zhu, W.; Esteban, R.; Borisov, A. G.; Baumberg, J. J.;Nordlander, P.; Lezec, H. J.; Aizpurua, J.; Crozier, K. B. Nat. Commun.2016, 7, 11495.(27) Shin, H.-H.; Yeon, G. J.; Choi, H.-K.; Park, S.-M.; Lee, K. S.;Kim, Z. H. Nano Lett. 2018, 18 (1), 262−271.(28) Urbieta, M.; Barbry, M.; Zhang, Y.; Koval, P.; Sanchez-Portal,D.; Zabala, N.; Aizpurua, J. ACS Nano 2018, 12 (1), 585−595.(29) Liu, P.; Chulhai, D. V.; Jensen, L. ACS Nano 2017, 11 (5),5094−5102.(30) Mohammed, A.; Hu, W.; Andersson, P. O.; Lundquist, M.;Landstrom, L.; Luo, Y.; Ågren, H. Chem. Phys. Lett. 2013, 581 (SupplC), 70−73.(31) Sprague-Klein, E. A.; McAnally, M. O.; Zhdanov, D. V.;Zrimsek, A. B.; Apkarian, V. A.; Seideman, T.; Schatz, G. C.; VanDuyne, R. P. J. Am. Chem. Soc. 2017, 139 (42), 15212−15221.(32) de Nijs, B.; Benz, F.; Barrow, S. J.; Sigle, D. O.; Chikkaraddy, R.;Palma, A.; Carnegie, C.; Kamp, M.; Sundararaman, R.; Narang, P.;Scherman, O. A.; Baumberg, J. J. Nat. Commun. 2017, 8 (1), 994.(33) van Schrojenstein Lantman, E. M.; de Peinder, P.; Mank, A. J.G.; Weckhuysen, B. M. ChemPhysChem 2015, 16 (3), 489−489a.(34) Linic, S.; Aslam, U.; Boerigter, C.; Morabito, M. Nat. Mater.2015, 14, 567.(35) Bach, U.; Lupo, D.; Comte, P.; Moser, J. E.; Weissortel, F.;Salbeck, J.; Spreitzer, H.; Gratzel, M. Nature 1998, 395, 583.(36) Wang, Z.-S.; Hara, K.; Dan-oh, Y.; Kasada, C.; Shinpo, A.; Suga,S.; Arakawa, H.; Sugihara, H. J. Phys. Chem. B 2005, 109 (9), 3907−3914.(37) Yang, N.; Zhai, J.; Wang, D.; Chen, Y.; Jiang, L. ACS Nano 2010,4 (2), 887−894.(38) von Wrochem, F.; Gao, D.; Scholz, F.; Nothofer, H.-G.; Nelles,G.; Wessels, J. M. Nat. Nanotechnol. 2010, 5, 618.(39) Karthauser, S. J. Phys.: Condens. Matter 2011, 23 (1), 013001.(40) Ossowski, J.; Wachter, T.; Silies, L.; Kind, M.; Noworolska, A.;Blobner, F.; Gnatek, D.; Rysz, J.; Bolte, M.; Feulner, P.; Terfort, A.;Cyganik, P.; Zharnikov, M. ACS Nano 2015, 9 (4), 4508−4526.(41) Cardinal, M. F.; Vander Ende, E.; Hackler, R. A.; McAnally, M.O.; Stair, P. C.; Schatz, G. C.; Van Duyne, R. P. Chem. Soc. Rev. 2017,46 (13), 3886−3903.(42) Bantz, K. C.; Meyer, A. F.; Wittenberg, N. J.; Im, H.; Kurtulus,O.; Lee, S. H.; Lindquist, N. C.; Oh, S.-H.; Haynes, C. L. Phys. Chem.Chem. Phys. 2011, 13 (24), 11551−11567.(43) Andreou, C.; Neuschmelting, V.; Tschaharganeh, D.-F.; Huang,C.-H.; Oseledchyk, A.; Iacono, P.; Karabeber, H.; Colen, R. R.;Mannelli, L.; Lowe, S. W.; Kircher, M. F. ACS Nano 2016, 10 (5),5015−5026.(44) Chen, J.; Wu, X.; Huang, Y.-w.; Zhao, Y. Sens. Actuators, B 2014,191 (Suppl C), 485−490.(45) Zhao, H.-Q.; Qiu, G.-H.; Liang, Z.; Li, M.-M.; Sun, B.; Qin, L.;Yang, S.-P.; Chen, W.-H.; Chen, J.-X. Anal. Chim. Acta 2016, 922, 55−63.(46) Lyndon, R.; Konstas, K.; Ladewig, B. P.; Southon, P. D.; Kepert,P. C. J.; Hill, M. R. Angew. Chem., Int. Ed. 2013, 52 (13), 3695−3698.(47) Schneemann, A.; Bon, V.; Schwedler, I.; Senkovska, I.; Kaskel,S.; Fischer, R. A. Chem. Soc. Rev. 2014, 43 (16), 6062−6096.(48) Park, W.-H.; Kim, Z. H. Nano Lett. 2010, 10 (10), 4040−4048.(49) Fromm, D. P.; Sundaramurthy, A.; Kinkhabwala, A.; Schuck, P.J.; Kino, G. S.; Moerner, W. E. J. Chem. Phys. 2006, 124 (6), 061101.(50) Ward, D. R.; Grady, N. K.; Levin, C. S.; Halas, N. J.; Wu, Y.;Nordlander, P.; Natelson, D. Nano Lett. 2007, 7 (5), 1396−1400.(51) Ichimura, T.; Watanabe, H.; Morita, Y.; Verma, P.; Kawata, S.;Inouye, Y. J. Phys. Chem. C 2007, 111 (26), 9460−9464.(52) Kamat, P. V.; Hartland, G. V. ACS Energy Lett. 2018, 3 (6),1467−1469.(53) Brandt, N. C.; Keller, E. L.; Frontiera, R. R. J. Phys. Chem. Lett.2016, 7 (16), 3179−3185.

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

I

(54) Brooks, J. L.; Frontiera, R. R. J. Phys. Chem. C 2016, 120 (37),20869−20876.(55) Hou, B.; Shen, L.; Shi, H.; Kapadia, R.; Cronin, S. B. Phys. Chem.Chem. Phys. 2017, 19 (4), 2877−2881.(56) Liu, Z.; Hou, W.; Pavaskar, P.; Aykol, M.; Cronin, S. B. NanoLett. 2011, 11 (3), 1111−1116.(57) Mukherjee, S.; Libisch, F.; Large, N.; Neumann, O.; Brown, L.V.; Cheng, J.; Lassiter, J. B.; Carter, E. A.; Nordlander, P.; Halas, N. J.Nano Lett. 2013, 13 (1), 240−247.(58) Mukherjee, S.; Zhou, L.; Goodman, A. M.; Large, N.; Ayala-Orozco, C.; Zhang, Y.; Nordlander, P.; Halas, N. J. J. Am. Chem. Soc.2014, 136 (1), 64−67.(59) Hartland, G. V.; Besteiro, L. V.; Johns, P.; Govorov, A. O. ACSEnergy Lett. 2017, 2 (7), 1641−1653.(60) Sousa-Castillo, A.; Comesana-Hermo, M.; Rodríguez-Gonzalez,B.; Perez-Lorenzo, M.; Wang, Z.; Kong, X.-T.; Govorov, A. O.; Correa-Duarte, M. A. J. Phys. Chem. C 2016, 120 (21), 11690−11699.(61) Tyler, T. P.; Henry, A.-I.; Van Duyne, R. P.; Hersam, M. C. J.Phys. Chem. Lett. 2011, 2 (3), 218−222.(62) Frontiera, R. R.; Henry, A.-I.; Gruenke, N. L.; Van Duyne, R. P.J. Phys. Chem. Lett. 2011, 2 (10), 1199−1203.(63) Kleinman, S. L.; Sharma, B.; Blaber, M. G.; Henry, A.-I.; Valley,N.; Freeman, R. G.; Natan, M. J.; Schatz, G. C.; Van Duyne, R. P. J.Am. Chem. Soc. 2013, 135 (1), 301−308.(64) Ringe, E.; Sharma, B.; Henry, A.-I.; Marks, L. D.; Van Duyne, R.P. Phys. Chem. Chem. Phys. 2013, 15 (12), 4110−4129.(65) Negru, B.; McAnally, M. O.; Mayhew, H. E.; Ueltschi, T. W.;Peng, L.; Sprague-Klein, E. A.; Schatz, G. C.; Van Duyne, R. P. J. Phys.Chem. C 2017, 121 (48), 27004−27008.(66) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178 (1), 42−55.(67) te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca Guerra,C.; van Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. J. Comput. Chem.2001, 22 (9), 931−967.(68) Valiev, M.; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma, T.P.; Van Dam, H. J. J.; Wang, D.; Nieplocha, J.; Apra, E.; Windus, T. L.;de Jong, W. A. Comput. Phys. Commun. 2010, 181 (9), 1477−1489.(69) Perdew, J. P. Phys. Rev. B: Condens. Matter Mater. Phys. 1986, 33(12), 8822−8824.(70) Becke, A. D. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38 (6),3098−3100.(71) Neugebauer, J.; Hess, B. A. J. Chem. Phys. 2003, 118 (16),7215−7225.(72) Lenthe, E. v.; Baerends, E. J.; Snijders, J. G. J. Chem. Phys. 1993,99 (6), 4597−4610.(73) van Lenthe, E.; Baerends, E. J.; Snijders, J. G. J. Chem. Phys.1994, 101 (11), 9783−9792.(74) Aquino, F. W.; Schatz, G. C. J. Phys. Chem. A 2014, 118 (2),517−525.(75) Fonseca Guerra, C.; Snijders, J. G.; te Velde, G.; Baerends, E. J.Theor. Chem. Acc. 1998, 99 (6), 391−403.(76) Jensen, L.; Zhao, L. L.; Autschbach, J.; Schatz, G. C. J. Chem.Phys. 2005, 123 (17), 174110.(77) Lee, S. Y.; Heller, E. J. J. Chem. Phys. 1979, 71 (12), 4777−4788.(78) Heller, E. J.; Sundberg, R.; Tannor, D. J. Phys. Chem. 1982, 86(10), 1822−1833.(79) Tannor, D. J.; Heller, E. J. J. Chem. Phys. 1982, 77 (1), 202−218.(80) Mullin, J.; Schatz, G. C. J. Phys. Chem. A 2012, 116 (8), 1931−1938.(81) Fang, Y.; Seong, N. H.; Dlott, D. D. Science 2008, 321 (5887),388−392.(82) Sonntag, M. D.; Chulhai, D.; Seideman, T.; Jensen, L.; VanDuyne, R. P. J. Am. Chem. Soc. 2013, 135 (45), 17187−17192.(83) Casiraghi, C.; Ferrari, A. C.; Robertson, J. Phys. Rev. B: Condens.Matter Mater. Phys. 2005, 72 (8), 14.(84) Gu, X.; Tian, S.; Zhou, Q.; Adkins, J.; Gu, Z.; Li, X.; Zheng, J.RSC Adv. 2013, 3 (48), 25989−25996.(85) Yang, W. h.; Hulteen, J.; Schatz, G. C.; Van Duyne, R. P. J. Chem.Phys. 1996, 104 (11), 4313−4323.

(86) Torngren, B.; Akitsu, K.; Ylinen, A.; Sanden, S.; Jiang, H.;Ruokolainen, J.; Komatsu, M.; Hamamura, T.; Nakazaki, J.; Kubo, T.;Segawa, H.; Osterbacka, R.; Smått, J.-H. J. Colloid Interface Sci. 2014,427, 54−61.(87) Olteanu, N. L.; Lazar, C. A.; Petcu, A. R.; Meghea, A.; Rogozea,E. A.; Mihaly, M. Arabian J. Chem. 2016, 9 (6), 854−864.(88) Schlather, A. E.; Manjavacas, A.; Lauchner, A.; Marangoni, V. S.;DeSantis, C. J.; Nordlander, P.; Halas, N. J. J. Phys. Chem. Lett. 2017, 8(9), 2060−2067.(89) Paxton, W. F.; Kleinman, S. L.; Basuray, A. N.; Stoddart, J. F.;Van Duyne, R. P. J. Phys. Chem. Lett. 2011, 2 (10), 1145−1149.(90) Rangel-Zamudio, L. I.; Rushforth, D. S.; Van Duyne, R. P. J.Phys. Chem. 1986, 90 (5), 807−811.(91) Lehmann, J.; Merschdorf, M.; Pfeiffer, W.; Thon, A.; Voll, S.;Gerber, G. Phys. Rev. Lett. 2000, 85 (14), 2921−2924.(92) Knight, M. W.; Sobhani, H.; Nordlander, P.; Halas, N. J. Science2011, 332 (6030), 702−704.(93) Hung, W. H.; Aykol, M.; Valley, D.; Hou, W. B.; Cronin, S. B.Nano Lett. 2010, 10 (4), 1314−1318.(94) Christopher, P.; Xin, H. L.; Linic, S. Nat. Chem. 2011, 3 (6),467−472.(95) Ingram, D. B.; Linic, S. J. Am. Chem. Soc. 2011, 133 (14), 5202−5205.(96) Murdoch, M.; Waterhouse, G. I. N.; Nadeem, M. A.; Metson, J.B.; Keane, M. A.; Howe, R. F.; Llorca, J.; Idriss, H. Nat. Chem. 2011, 3(6), 489−492.(97) Lee, J.; Mubeen, S.; Ji, X. L.; Stucky, G. D.; Moskovits, M. NanoLett. 2012, 12 (9), 5014−5019.(98) Silva, C. G.; Juarez, R.; Marino, T.; Molinari, R.; Garcia, H. J.Am. Chem. Soc. 2011, 133 (3), 595−602.(99) Goykhman, I.; Desiatov, B.; Khurgin, J.; Shappir, J.; Levy, U.Nano Lett. 2011, 11 (6), 2219−2224.(100) Fang, Z. Y.; Wang, Y. M.; Liu, Z.; Schlather, A.; Ajayan, P. M.;Koppens, F. H. L.; Nordlander, P.; Halas, N. J. ACS Nano 2012, 6(11), 10222−10228.(101) Mukherjee, S.; Libisch, F.; Large, N.; Neumann, O.; Brown, L.V.; Cheng, J.; Lassiter, J. B.; Carter, E. A.; Nordlander, P.; Halas, N. J.Nano Lett. 2013, 13 (1), 240−247.(102) Fang, Z. Y.; Liu, Z.; Wang, Y. M.; Ajayan, P. M.; Nordlander,P.; Halas, N. J. Nano Lett. 2012, 12 (7), 3808−3813.(103) Shibamoto, K.; Katayama, K.; Sawada, T. Chem. Phys. Lett.2007, 433 (4−6), 385−389.(104) Wang, H.; Yao, K.; Parkhill, J. A.; Schultz, Z. D. Phys. Chem.Chem. Phys. 2017, 19 (8), 5786−5796.(105) He, H. Y.; Pi, S. T.; Bai, Z. Q.; Banik, M.; Apkarian, V. A.; Wu,R. Q. J. Phys. Chem. C 2016, 120 (37), 20914−20921.(106) Kim, Y.; Torres, D. D.; Jain, P. K. Nano Lett. 2016, 16 (5),3399−3407.(107) Henglein, A. J. Phys. Chem. 1993, 97 (21), 5457−5471.(108) Zhang, Y.; He, S.; Guo, W.; Hu, Y.; Huang, J.; Mulcahy, J. R.;Wei, W. D. Chem. Rev. 2018, 118 (6), 2927−2954.

Journal of the American Chemical Society Article

DOI: 10.1021/jacs.8b06347J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

J