Hefei 230026, China. E-mail: zhuyanwu@ubWuxi Graphene Technologies Co., Ltd, 311cJiangnan Graphene Research Institute, 6 XidNational Synchrotron Radiation Research CeDepartment of Physics, Tamkang University
† Electronic supplementary informationincluding UV-vis absorption spectra andsizes, optical image and transmissionimage of Au NPs covered with a large-arspectra of Au NPs with different size or clayers of graphene coating, Raman specconcentrations and simulated electric elfor Au NPs with or without mono10.1039/c4tc00353e
Cite this: J. Mater. Chem. C, 2014, 2,4683
Received 21st February 2014Accepted 15th April 2014
Enhanced light–matter interaction of graphene–gold nanoparticle hybrid films for high-performance SERS detection†
Yuanxin Du,a Yuan Zhao,a Yan Qu,bc Chia-Hao Chen,d Chieh-Ming Chen,e
Cheng-Hao Chuange and Yanwu Zhu*a
By simply coating graphene films on Au nanoparticles, the optical properties of the hybrid films are
investigated. It is found that the coverage of a monolayer graphene film leads to a decreased
transmittance of up to 15.8% in the visible range, much higher than the 2.3% transmittance loss for
intrinsic graphene. At the same time, the plasmonic resonance of the hybrid films experiences a red-shift
in resonance frequency and a broadening in the transmission dip. By means of finite element
simulations, these observations are attributed to strong light–matter interaction at the interface between
graphene and Au nanoparticles, as indicated by the increased absorption cross section and higher
electric field intensity. The electron transfer between graphene and Au nanoparticles is confirmed by
high resolution X-ray photoelectron spectroscopy studies. Furthermore, the enhanced electromagnetic
hot spots at the interface between graphene and Au nanoparticles make such graphene–Au nanoparticle
hybrid films cost-effective and high-performance surface-enhanced Raman scattering substrates for
detecting organic molecules such as rhodamine-6G, for which an enhancement factor of�107 is achieved.
1. Introduction
Graphene, a at monolayer of sp2 carbon atoms in a hexagonallattice conguration,1 continues to attract intense interestbecause of its excellent optical2,3 and electronic properties,4,5
which make it a promising candidate for applications inphotonic or optoelectronic devices such as photodetectors,6
phototransistors7 and optical modulators.8 The light absorptionof pristine graphene is at a level of 2.3% independent of wave-length in the visible and near infrared range, which has trig-gered a signicant amount of research on using graphene astransparent conducting lm.9 However, such a weak andwavelength-independent absorption in the visible range may
eering & CAS Key Laboratory of Materials
and Technology of China, 96 Jin Zhai Rd,
stc.edu.cn
Yanxin Rd, Wuxi 214174, China
angyun Rd, Changzhou 213149, China
enter, Hsinchu 300, Taiwan
, New Taipei City 25137, Taiwan
(ESI) available: Additional informationTEM images of Au NPs with differentspectra of monolayer graphene, SEMea monolayer graphene, transmissionover density before and aer differenttra of R6G on Au NPs with differentd distributions at the laser wavelengthlayer graphene coating. See DOI:
hemistry 2014
cause substantial restrictions and challenges for electro-opticaland all-optical applications where controlling light absorptionat specic wavelengths is needed.10,11 On the other hand, due tothe unique and tunable optical properties of localized surfaceplasmons, metallic plasmonic nanostructures have beeninvestigated for decades and are believed to be one of the mostpromising candidates for applications in photodetectors,12
photo-diagnostics and photothermal therapy,13,14 surface-enhanced Raman scattering (SERS),15 and molecular imagingand sensing.16,17 Thus, combining graphene and conventionalmetallic plasmonic nanostructures has been an effectivemethod for enhancing light–matter interaction within visiblewavelengths, which could encourage potential applications inmany elds. For example, hybrid lms of graphene and plas-monic metallic structures have been utilized in photovoltaicdevices18 or high-speed optical communications12 to simulta-neously enhance the incident light absorption and the carrier-separation efficiency, which mainly relies on the high carriermobility in graphene. Due to the excellent bio-compatibility andchemical stability of graphene, considerable efforts have beendevoted to graphene-based SERS sensors towards future tech-niques for the identication and detection of chemical andbiological species in a label-free environment.19,20
In graphene-based SERS congurations, graphene playsmultiple roles, such as a uorescence quencher, additionalchemical enhancer, molecule enricher and building block as aat-surface substrate by the virtue of its two-dimensional (2D)structure.21 It has been accepted that the enhancement of the
local electromagnetic eld in the hybrids mainly contributes toSERS signals of adsorbates, in addition to graphene-relatedweak chemical enhancement.22 In most graphene-based SERSstudies, metal nanostructures were loaded on graphene inliquid systems.19,23,24 Because the direct contact between mole-cules to be detected and bare metal nanoparticles is inevitablein liquid-based graphene–metal hybrids, SERS detection inliquids suffers from the following disadvantages: (1) ambiguousmechanism explanations such as hot-spot effects caused bynanoparticle aggregates, chemical adsorption-induced vibra-tions, and charge transfer between metal and molecules; (2)false-positive detection due to the interference of impurities incomplex solution systems; (3) unfavorable disturbances such asphoto-induced damages and metal-catalyzed side reactions.Recently, hybrid lms consisting of graphene layers andmetallic nanostructures have been developed en route to lowerdetection limits and facile detection realization. For example,Zhu et al. transferred graphene onto Au nanovoid arrays madeby sphere templates,25 and Wang et al. fabricated graphene–Aunano-pyramid hybrids via standard lithography26 for SERSdetection of rhodamine-6G (R6G) with enhancement factors of103 and 1010, respectively. A simpler and cost-effective fabrica-tion of graphene–metal hybrid SERS sensors is required forreliable and portable detection with high accuracy and goodstability.
In the present study, we fabricated graphene-coated plas-monic nanostructures by simply coating graphene layers ontoAu nanoparticles (NPs) to detect R6G with a large enhancementfactor of �107. From such graphene-nanoparticle hybrid lms,we observed a dramatic enhancement of the light–matterinteraction, enabling a decrease in the light transmittance of upto 15.8% and signicant plasmon resonance frequency red-shis. The effects of size and density of Au NPs and the numberof graphene layers were investigated. The experimental obser-vations were further explained by numerical simulations basedon the nite element method (FEM) theory. Due to theincreased electromagnetic eld and the intrinsic chemicalenhancement of graphene, the hybrid lms have been deter-mined to be an excellent SERS substrate.
2. Materials and methods2.1. Chemicals and materials
Chloroauric acid (HAuCl4$4H2O, 99.9%), sodium citrate(Na3C6H5O7$2H2O, 99.8%), hydrogen peroxide (H2O2, 30%),sulfuric acid (H2SO4, 98%), ethanol, and acetone werepurchased from Sinopharm Chemical Reagent Co., Ltd.(Shanghai, China). 3-Aminopropyltriethoxysilane (3-APTS) andrhodamine-6G (R6G) were obtained from Sigma-Aldrich. All ofthese reagents were used without further purication. Ultrapurewater (18.2 MU cm) was produced using a water puricationsystem (MERCK Millipore Direct-Q3).
2.2. Synthesis and deposition of Au nanoparticles
Au NPs were prepared by the chemical reduction of chloroauricacid with sodium citrate. The size of Au NPs was tuned by
4684 | J. Mater. Chem. C, 2014, 2, 4683–4691
changing the volume of the sodium citrate solution.27 Todeposit Au NPs, various substrates were treated with H2SO4–
H2O2 (3 : 1 v/v) at 80 �C for 30 min to derive a hydroxyl surface.The substrates were then immersed in a 10% 3-APTS ethanolsolution for 2 h and fully modied with –NH2 end groups.Unbound redundant 3-APTS monomers were removed byrinsing profusely with ethanol. The 3-APTS-modied substrateswere immersed into colloidal Au NPs, then rinsed profuselywith ultrapure water and dried in air, which resulted in theformation of a layer of Au NPs on the substrate surface. Thecover density of Au NPs was controlled by varying the immer-sion duration in the Au NP colloidal suspension.27
2.3. CVD-grown monolayer graphene transfer
Graphene was synthesized by atmospheric pressure chemicalvapor deposition (CVD) at 1000 �C on copper foils with methaneas the carbon source.28 Subsequently, the graphene lms weretransferred onto the Au NP-coated substrates by using thethermal release tape method.29
2.4. Characterization and instruments
Scanning electron microscopy (SEM, JSM-6700F) and atomicforcemicroscopy (AFM, DI Innova) were used to characterize themorphology of Au NPs deposited on substrates, before and aergraphene transfer. Transmission electron microscopy (TEM,JEOL 2010) was utilized to characterize the size and structure ofAu NPs. UV-vis absorbance and transmittance spectra wererecorded with a Shimadzu Solid 3700 spectrometer. X-rayphotoelectron spectroscopy (XPS) was utilized to determine thecore-level binding energy (BE) prole of elements. The XPSspectra of graphene/Au/ITO and Au/ITO samples weremeasured with photon energy of 380 eV in beamline 09A1 at theNational Synchrotron Radiation Research Center (NSRRC),Taiwan. The precise energy resolution (<0.10 eV) was achievedby the combination of the spherical grating monochromator(SGM) with a line density of 400 mm�1 and the PHI electronenergy analyzer (Model 1600/3057) with a pass energy of 5.85 eV.Aer calibration with an XPS spectra of standard HOPG and Ausamples, the photoelectron spectra were subtracted with thebackground intensity around the pre-edge range and werenormalized to the background intensity around the far-edgerange to avoid different photon ux effects. Raman measure-ments were conducted with a Renishaw inVia Raman Micro-scope, equipped with a CCD detector with an excitationwavelength of 532 nm. For SERS detection, 2 mL aliquots of R6Gin ethanol with different concentrations were dropped anddispersed onto as-prepared substrates and dried in air. TheRaman spectra were recorded using a 532 nm laser with 0.5 mWpower and a 50� objective for all samples. The integral time was1 s with an accumulation of 5 rounds.
2.5. Finite element numerical simulation method
To rationalize the effects of graphene on the plasmonic prop-erties of hybrid lms, FEM numerical electromagnetic simula-tions were performed using Comsol Multiphysics (COMSOL4.3a). In the simulations, Au hemispheres with a diameter of
30 nm, with or without a layer of conformed graphene, wereplaced on the surface of a semi-innite glass substrate. A planelight wave was launched perpendicular to the substrate with asingle, polarized electrical eld, Ey. The simulation area was60 nm � 60 nm in the horizontal dimension, and the compu-tational domain was considered as a single unit cell. Perfectlymatched layer (PML) absorbing boundary conditions wereadopted in the boundary. The absorption cross section wascalculated using
Pabs ¼ pc30l
ð300ðlÞjEðlÞj2dV ; (1)
where l is the incident wavelength, c is the speed of light, 30 isthe vacuum permittivity, 30 0 (l) is the imaginary part of thedielectric constant, E is the electric eld, and the integral istaken over the volume. The dielectric constant of Au wascalculated using the Drude model.30 The refractive index ofgraphene in the visible range is governed by ng ¼ 3.0 + C(l0/3)i,where the constant C z 5.446 mm�1 is obtained from theopacity measurement of Bruna et al.,31 and l0 is the vacuumwavelength. The thickness of themonolayer graphene was set as0.5 nm; multiple graphene layers were considered as homoge-nous stacking of monolayer graphene.
3. Results and discussion3.1. Fabrication and characterization of the graphene–Aunanoparticle hybrid lms
The fabrication procedure of the graphene-coated Au NPs isschematically illustrated in Fig. 1a. For the deposition of AuNPs, 3-aminopropyltriethoxysilane (3-APTS) was rst linkedonto the surface of the chemically treated substrates to form amonolayer of APTS with –NH2 end groups. Aer the substratewas immersed in the Au NPs colloid suspension, the Au NPs
Fig. 1 (a) Schematic of the fabrication of graphene–Au NP hybrid films. (per mm2. The inset shows the TEM image of the Au NPs. (c) SEM imagemonolayer graphene, as confirmed by the Raman spectrum in the inset. Aline profiles in (d) and (e) were randomly taken from each AFM scan, ind
attached to the surface by the interaction between particles and–NH2 end groups, which resulted in a uniform sub-monolayerdeposition of Au NPs on the substrate, as shown in the scanningelectron microscopy (SEM) and atomic force microscopy (AFM)images in Fig. 1b and d, respectively. The as-synthesized Au NPshave an average size of �30 nm, as conrmed by transmissionelectron microscopy (TEM; inset of Fig. 1b); two other types ofAu NPs with average sizes of �15 nm and �45 nm were alsofabricated (see TEM images in Fig. S1†). By changing theimmersion duration, the cover density for the �30 nm Au NPshas been tuned to about 69, 122, 221, 343, and 408 particles permm2 (no. per mm2) (see SEM images in Fig. S3†). The ratio of AuNPs in aggregates was calculated and shown in Table S1.† Wecan see that the ratio falls in the range 13.22–14.38% for a coverdensity from �69 to �343 no. per mm2, whereas the highestcover density of �408 no. per mm2 led to a higher aggregationratio of 19.63%.
Graphene grown by chemical vapor deposition (CVD) wastransferred onto Au NPs on various substrates. Fig. 1c shows thetypical SEM image of the graphene–Au NP hybrid lms. Thearea with graphene cracks was selected and magnied toidentify the existence of graphene by contrast. Due to enhancedconductivity, the graphene-covered region has a darker contrastin the second electron imaging mode. The SEM image inFig. S2a† further reveals the successful transfer of large-areagraphene on top of Au NPs. The Raman spectrum in the inset ofFig. 1c conrms the monolayer nature of the graphene trans-ferred onto Au NPs. The intensity ratio of 2D/G is 3.16, and thesymmetric 2D band is centered at �2680 cm�1 with a full widthat half maximum of �32 cm�1, which are the typical features ofmonolayer graphene. The absence of detectable D peakssuggests the absence of microscopic disorders in graphene aerbeing transferred onto Au NPs. The Raman signals frommonolayer graphene (Fig. S2b†) placed on the top of the Au NPs
b) SEM image of �30 nm Au NPs deposited on Si at a density of 343 no.of monolayer graphene-coated Au NPs. The dark area is covered byFM images of Au NPs (d) without and (e) with monolayer graphene. Theicating the different roughness of two samples.
have been dramatically enhanced by the plasmonic nano-structure as reported by others.25 The graphene cut from thesame growth was transferred onto 300 nm SiO2/Si or quartzsubstrates to further conrm the large-area completeness anduniformity of the monolayer graphene using an optical micro-scope or transmittance spectrum, as shown in Fig. S2c and d,†respectively.
The topography of graphene–Au NP hybrid lms waschecked with AFM (Fig. 1e), with that of bare Au NPs as refer-ence (Fig. 1d). The AFM image in Fig. 1e reveals the uniformsurface morphology aer graphene was coated onto the Au NPs,leading to a smaller roughness (arithmetic average deviation) of3.25 nm, compared with that of 8.35 nm for the bare Au NPs.Furthermore, we can see that the graphene bridges dispersedAu NPs with no signicant breakage on the scale scanned anddeforms itself to conform to the geometry of the Au NPs. Suchclose contact between graphene and the underlying Au nano-structures is expected to improve the electromagnetic hot spotsat the interface between graphene and Au, which may be majorreason for the enhanced Raman signals of adsorbates on thehybrid lms, as discussed below.
To further explore the interface between graphene and AuNPs, high-resolution X-ray photoelectron spectroscopy (XPS)with a synchrotron photon source has been performed. To
Fig. 2 (a) XPS Au 4f spectra of thick Au film on a Si substrate and�30 nm(b) XPS C 1s spectra of HOPG and graphene on ITO substrates with or wtransfer from Au NPs to ITO, from graphene to Au NPs, and from graph
4686 | J. Mater. Chem. C, 2014, 2, 4683–4691
perform XPS measurements, conductive ITO has been selectedas substrates for depositing Au NPs, with or without graphenecoating. The fabrication procedures were kept exactly the same.Fig. 2a shows the Au 4f core-level spectra of Au/ITO and G/Au/ITO, as compared with that of a thick Au lm on a Si substrate.The measured binding energies (BEs) of Au 4f7/2 and 4f5/2 statesin the thick Au/Si lm are 84.1 and 87.8 eV, respectively, whichis in good agreement with the Au metal measurement.32 Withrespect to the metallic Au0 state in thick Au lm, 30 nm Au NPson ITO exhibit an up-shi in the BEs of 4f7/2 and 4f5/2 states by0.2 eV. The shi is ascribed to the –NH2 end groups bonded tothe surface of the Au NPs (Fig. 2c), consistent with the result ofthe Au nanocrystals with amine groups.33 Aer coating withgraphene, the XPS spectrum of the graphene-coated Au NPs hasthe same peak width and BEs as those from Au/Si; the BE peaksshow a down-shi, compared with that of the bare Au NPswithout graphene. As the work functions of Au (�5.31 eV) andgraphene (�4.48 eV) are different,34 the electron transfer fromgraphene to Au NPs may occur for the Fermi energy alignment.Fig. 2b shows C 1s spectra of G/Au/ITO, G/ITO, and a standardHOPG sample. The chemical environment of HOPG is consid-ered as sp2-hybridized carbon, showing an intense peak at a BEof 284.4 eV. The C 1s XPS peak of G/ITO shows a broadenedwidth and an upshi of 0.5 eV relative to that of HOPG.
Au NPs on ITO substrates with or without monolayer graphene coating.ithout �30 nm Au NPs underneath. (c) Schematic showing the chargeene to ITO, respectively, for various XPS samples.
Although the chemical and lattice structure of graphene isanalogous to the single-layer HOPG, it is reasonable to see theelectron-transfer response when the delocalized p electrons inthe honeycomb structure of graphene are in contact with theoxygen vacancy in the ITO substrate.35,36 The nearly identical C1s peak features in G/Au/ITO can be explained by the goodcontact between graphene and Au NPs, which provides an effi-cient electron-transfer path from graphene to Au, as supportedby the good morphological conformance in the AFM results.The high BE side and the shoulders at BEs of 287.0 and 289.5 eVresult from the surface carbon adsorption on ITO; for thisreason, they are almost equal in G/Au/ITO and G/ITO while theprobing depth of X-ray beam is greater than the thickness ofmonolayer graphene. As schematically described in Fig. 2c, thep electrons of graphene layers are driven to transfer into theoxidized Au state of G/Au/ITO, and the resulting�0.5 eV upshiof the C 1s state is addressed in the upper monolayer graphene.Such an up-shi in C 1s peak, compared with the case in HOPG,is consistent with the down-shi in the Au 4f state of the gra-phene–Au NPs on the ITO, which conrms the electron transferfrom graphene to Au NPs.
3.2. Optical properties of graphene–Au nanoparticle hybridlms
The optical transmission of Au NPs on quartz substrates hasbeen signicantly modulated by the graphene layers covered. Ascan be seen from Fig. 3a, deepening, broadening, and red-shiing of the plasmonic resonance dip was observed upon thecoating of monolayer graphene, accompanied by a signicantdecrease in the absolute transmittance of about 15.8%. Such atransmittance decrease is considerably higher than the typical2.3% transmittance loss for monolayer graphene in the samewavelength range. Interestingly, once the rst layer of graphenewas transferred, the additional layers of graphene coating didnot change the transmittances as much as the rst layer. In fact,
Fig. 3 (a) Transmission spectra of �30 nm Au NPs with a density of �34numbers of graphene layers were coated. (b) Transmittance and wavelenlayers coated on �30 nm Au NPs with a density of �343 no. per mm2. (ccompared with bare Au NPs as a function of cover density for the hybrid
for the second, third, or fourth transfers, the decrease in thetransmittance linearly depends on the number of graphenelayers, and the resonance wavelength shows only a slightchange, as shown in Fig. 3b and S3.† Furthermore, when thesize of the Au NPs is xed, the transmittance at plasmonicresonance for the hybrid lms consisting of monolayer gra-phene shows a monotonous increase with the coverage densityof the Au NPs of up to �343 no. per mm2, followed by a decreasefor the higher density of�408 no. per mm2 (Fig. 3c and S3†). Thedecreased transmittance for the very high density of Au NPsmaybe attributed to the aggregation of NPs, as shown in Fig. S3dand Table S1.† The resonance wavelength shi upon graphenecoverage was studied as well. For the lowest-density Au sample,the transmission dip experiences a smaller wavelength shi(�12 nm) than the others (�17 nm; Fig. 3c and S3†), indicatingthat the interaction between light and the hybrid lms is heavilyaffected by the contact areas between graphene and plasmonicAu NPs. In addition, Fig. S4† shows that the trend of thedecrease in the transmittance and the resonance wavelengthred-shi remains similar for the graphene-coated Au NPs withthe average size of nanoparticles of �15 nm and �45 nm,respectively.
As we can see below, the modulated and suppressed opticaltransmittance of the graphene–Au NP hybrid lms can beexplained by the changes in the dielectric function of thesurroundings and refractive index of Au plasmonic nano-particles.37 Graphene effectively mimics a dielectric material inthe visible and near-infrared wavelength ranges.38 Similar to thecase where a plasmonic nanostructure is embedded in themedium with a higher refractive index,37,39 the red-shi in theplasmonic resonance and suppressed transmittance inthe graphene–Au NP hybrid lms in this work are attributed tothe presence of graphene and the graphene Ohmic loss.25 As theexperimental results show, the coupling between chargecarriers in rst layer of graphene and the surface plasmons of
3 no. per mm2 deposited on quartz substrates before and after variousgth at plasmonic resonance as a function of the number of graphene) Absolute transmittance change and resonance dip wavelength shift,films consisting of monolayer graphene and �30 nm Au NPs.
Au NPs is considerably stronger than the localized surfaceplasmon resonance (LSPR) of individual nanoparticles. Such ahybrid lm therefore provides a platform for chemical sensing,e.g. by enhanced Raman scattering as described below. As wecan see from the following discussion, the reason for thesensitivity enhancement lies in the fact that the LSPR is coupledby a diffractive wave propagating along the surface of thesample and that the wave spread is mainly affected by theconductivity and dielectric permittivity of graphene.40
Fig. 4a shows the nite element simulated absorption crosssections as the function of the incident wavelength for Au NPson a glass substrate without graphene or with one, two andthree layers of 0.5 nm-thick graphene. The calculated plasmonicresonance wavelength is approximately 530 nm for the bare AuNPs on the glass substrate, matching well with the experimentalobservations. Coating a monolayer of graphene on the Auhemisphere results in a dampening of the plasmonic resonanceand thus leads to a resonance shi towards longer wavelengthsand a broadening of the resonance peak. The increased absor-bance near the resonance wavelength is caused by enhancedlight–matter interaction, which is indicated by the increasedelectrical eld intensity of plasmonic resonance as shown inFig. 4b and in greater detail in Fig. S5.† Fig. 4b shows that themaximum electrical eld intensity around Au NP coated withmonolayer graphene is increased to approximately 4.4 from thevalue of 3.5 for the bare Au NPs, corresponding to a 19.36-foldenhancement of the optical absorption, compared with the12.25-fold enhancement for the bare Au NPs as estimated fromformula (1). Since the reection of graphene can be neglectedfor a normal incidence,31 the calculated absorption crosssection thus provides a qualitative explanation for the decreasein transmittance for graphene–Au NP hybrid lms observed inthe experiments. In addition, by comparing the electromagneticeld distribution of the Au NPs before and aer graphenecoating, we can see from Fig. 4b that the electromagnetic eld is
Fig. 4 (a) FEM-simulated absorption cross-section for�30 nm Au NPs wthree (green lines) layers of graphene transferred onto Au NPs. (b) Simulatdirection (dotted line) for Au NPs (top) without and (below) with a mono
4688 | J. Mater. Chem. C, 2014, 2, 4683–4691
more conned in a narrower region for the graphene–Au NPhybrids, due to the coupling of graphene and plasmonicAu NPs.
3.3. SERS properties of the graphene–Au nanoparticlehybrid lms
It is well known that noble metals such as Ag and Au nano-crystals can serve as excellent substrates for SERS based on theenhanced local electromagnetic eld due to the surface plas-mons. As a substrate potentially useful for detecting Ramansignals of molecules, graphene not only has the capability ofadsorbingmolecules and quenching uorescence backgrounds,but it also demonstrates chemical enhancement based oncharge transfer to enhance the Raman scattering of the mole-cules.21 In this work, R6G was chosen to probe the SERS effect ofthe graphene–Au NP hybrid lms. Fig. 5a shows the averageRaman intensities (collected from 30 spectra obtained fromrandom spots on each lm) of 2 mL R6G ethanol solution withvarious molecular concentrations immobilized on monolayergraphene–Au NP hybrid lms. The Raman peaks at 612 cm�1,774 cm�1, 1186 cm�1, 1360 cm�1, 1506 cm�1, and 1648 cm�1
are in good agreement with previous reports for R6G.26 TheRaman signal was still visible for a R6G concentration as low as10�8 M. The SERS enhancement factors (EF) for R6G on thegraphene–Au NP hybrid lms were calculated according to theequation EF¼ (ISERS/Ibulk)(Nbulk/NSERS), where ISERS and Ibulk arethe 612 cm�1 peak intensities obtained from 10�8 M R6G on thehybrid lms and from 10�2 M R6G obtained on a quartzsubstrate (Fig. 5b), respectively. NSERS and Nbulk are thenumbers of R6G molecules excited by the laser beam on thehybrid lms and quartz substrate, respectively. An EF of �107
was calculated for the graphene–Au NP hybrid lms with an AuNP cover density of �343 no. per mm2. Compared with the bareAu NPs lm (Fig. 5 and S6†), the graphene–Au NP hybrid lmshows an enhancement of 14-fold in the EF values. The
ithout graphene (black lines) or with one (red lines), two (blue lines), anded electric field distributions and corresponding line profiles along the ylayer graphene coating.
Fig. 5 (a) SERS spectra of R6Gwith varied concentrations onmonolayer graphene–Au NP hybrid films. (b) SERS spectra of 10�2 M R6G on quartz(curve 1), 10�3 M on graphene/quartz (curve 2), 10�7 M on Au NPs/quartz (curve 3), and 10�8 M on the monolayer graphene–Au NP hybrid films(curve 4). The baseline has been subtracted from each spectrum.
Paper Journal of Materials Chemistry C
Publ
ishe
d on
16
Apr
il 20
14. D
ownl
oade
d by
Uni
vers
ity o
f Sc
ienc
e an
d T
echn
olog
y of
Chi
na o
n 10
/03/
2016
06:
17:1
2.
View Article Online
enhancement factors for the bare Au NP samples and graphene-coated Au NP hybrids have been calculated for different coverdensities of Au NPs and are shown in Fig. S7.†
Such a graphene-related enhancement could be due to gra-phene-induced uorescence quenching, adsorption enrich-ment, chemical enhancement, and electromagneticenhancement from the coupling of charge carriers in grapheneand surface plasmons in the Au NPs. It has been reported thatmonolayer graphene facilitates charge transfer between gra-phene and probe molecules, resulting in a vibration-modedependent enhancement of 2–17 times.41 Based on graphene-covered Au nanovoid arrays25 and Au nano-pyramids26 fabri-cated with template technology, SERS enhancement factors of1.8 and 10 were attributed to the introduction of graphene whenbeing used to sense R6G molecules. In SERS, the enhancementfactor could be estimated from the localized electromagneticenhancement by
gzjEtj4jEij4
; (2)
where Ei is the incident eld intensity and Et is the eldintensity at the location of the molecules detected. Fig. S8†shows the simulated electrical eld distributions around AuNPs with or without monolayer graphene transferred at theRaman laser wavelength of 532 nm (slightly different from theplasmonic resonance frequency of Au). With graphene coating,the maximum enhancement of the electrical eld intensity isincreased to 4.2, corresponding to an electromagneticenhancement of about 311 for SERS. In contrast, the maximumenhancement of the electrical eld intensity and estimatedelectromagnetic enhancement are 3.7 and 187, respectively, forbare Au NPs deposited on the quartz substrate.
Additionally, we determined that graphene makes SERSdetection more stable. Due to the large-area completeness andchemical stability of graphene, the graphene coating in thehybrid lms may keep most of the R6G molecules from directcontact with the Au NPs, thus preventing R6G molecules from
photocarbonization.42,43 On the other hand, in conventional SERSexperiments especially for dyes, the photobleaching of the Ramanprobes induced by the laser may lead to uncontrollable variationsof the SERS spectra with acquisition time and laser power.21 Asshown in Fig. 6, time-resolved Raman spectroscopy was carriedout to compare the SERS detection stability of the bare Au NPsand graphene-coated Au NPs on the same substrate. It can beseen that the SERS signal intensity of R6G on the bare Au NPsdecayed quickly during a 480 s-long measurement, while thesignals of R6G on graphene–Au NP hybrids remained consider-ably more stable. This stabilization effect could be related to themetal-molecule isolation induced by the surface passivationeffect of graphene and/or the formation of a graphene–moleculecomplex (especially aromatic molecules) through p–p interac-tions.22 Furthermore, the high thermal conductivity of graphenemay also contribute to the enhanced stability by dissipating heatmore efficiently. In fact, it has been found that the morphology ofthe metal lm changes aer being exposed to a relatively higherlaser power, while it remains stable with graphene coating.43 Wealso carried out Raman studies of R6G with different concentra-tions (Fig. S9†) on the same graphene–Au NP hybrid lmsubstrate aer it was kept for half a year in our lab. The Ramanintensity of R6G only showed a slight decrease (Fig. S10†). Thus,the graphene–Au NPs hybrid lms fabricated by the simple andcost-effective self-assembly method are able to reach the stabilityrequirements in practical applications.
The high uniformity and reproducibility of SERS signals areessential for its practical application, especially when theconcentration and/or the amount of adsorbates are very low. Aswe can see from Fig. S7 and S11,† while the SERS sensitivity forbare Au NPs is similar for different cover densities, the unifor-mity is improved with the cover density until it reaches the valueof�343 no. per mm2. A higher cover density of�408 no. per mm2
leads to deteriorated uniformity due to the higher degree ofaggregation (Fig. S11†). The homogeneity of SERS signals on thehybrid lms with a cover density of�343 no. per mm2was furtherinvestigated by performing a spatially resolved Raman intensity
Fig. 6 Evolution of SERS signals from 10�4 M R6Gon (a) Au NPs films and (b) graphene–AuNPs hybrid films, respectively (the data were collectedat an interval of 20 s). Spatial resolved Raman intensity mapping of (c) graphene G peak and (d) 612 cm�1 peak of 10�7 M R6G on graphene–AuNPs hybrid films with the boundary of graphene sheet indicated by the dash line. The region below the dash line is bare Au NPs without graphenecoverage. The boundary was determined according to the optical picture recorded by Renishaw inVia Raman Microscope.
Journal of Materials Chemistry C Paper
Publ
ishe
d on
16
Apr
il 20
14. D
ownl
oade
d by
Uni
vers
ity o
f Sc
ienc
e an
d T
echn
olog
y of
Chi
na o
n 10
/03/
2016
06:
17:1
2.
View Article Online
mapping across a graphene boundary (indicated by the dashedline in Fig. 6c). The region below the dashed line is bare Au NPswithout graphene coverage. As shown in Fig. 6c and d, the SERSintensity mapping of R6G is highly consistent with the Raman Gpeak mapping of graphene. Such a result clearly indicates thatthe presence of graphene is essential to the enhanced SERSdetection ability for R6Gmolecules. In addition, the area coveredwith graphene (above the dashed line) shows a highly homoge-neous adsorption for R6G, further making the graphene-basedSERS detection attractive in practical applications.
4. Conclusions
In summary, we have demonstrated an enhanced light–gra-phene interaction by fabricating graphene–Au NP hybrid lmsand investigating their optical properties. A signicant decreasein transmittance, red-shi, and broadening of the plasmonicresonance was observed when a monolayer graphene wastransferred on top of the Au NPs. The effects of size and densityof the Au NPs and the number of graphene layers were inves-tigated. The experimental observations were explained by niteelement numerical simulations. In addition, we have demon-strated that the hybrid lms are high-performance candidatesfor SERS applications due to the combined effects of the
4690 | J. Mater. Chem. C, 2014, 2, 4683–4691
electromagnetic enhancement activity at the graphene–Auinterface, and the surface enrichment, uorescence quenching,and additional chemical enhancement of graphene. Using R6Gmolecules as probes, we obtained SERS enhancement factors ofup to �107 on graphene–Au NP hybrid lms, with a stable andhomogenous response. The simple and rapid SERS sensorreported here may open up new opportunities in developing theapplications of graphene in biomedical diagnostics, analyticalchemistry, as well as biological sensing and imaging.
Acknowledgements
The authors thank China Government 1000Plan TalentProgram, China MOE NCET Program and Natural ScienceFoundation of China (51322204) for support. C.-H. Chuangacknowledges the National Science Council of Taiwan fornancial support under Contract no. NSC 102-2112-M-032-001and NSC 102-2632-M-032-001-MY3.
Notes and references
1 K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang,S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004,306, 666.
2 A. N. Grigorenko, M. Polini and K. S. Novoselov, Nat.Photonics, 2012, 6, 749.
3 R. R. Nair, P. Blake, A. N. Grigorenko, K. S. Novoselov,T. J. Booth, T. Stauber, N. M. R. Peres and A. K. Geim,Science, 2008, 320, 1308.
4 A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselovand A. K. Geim, Rev. Mod. Phys., 2009, 81, 109.
5 F. Bonaccorso, Z. Sun, T. Hasan and A. C. Ferrari, Nat.Photonics, 2010, 4, 611.
6 A. Pospischil, M. Humer, M. M. Furchi, D. Bachmann,R. Guider, T. Fromherz and T. Mueller, Nat. Photonics,2013, 7, 892.
7 H. Chang, Z. Sun, M. Saito, Q. Yuan, H. Zhang, J. Li, Z. Wang,T. Fujita, F. Ding, Z. Zheng, F. Yan, H. Wu, M. Chen andY. Ikuhara, ACS Nano, 2013, 7, 6310.
8 E. O. Polat and C. Kocabas, Nano Lett., 2013, 13, 5851.9 J. B. Bult, R. Crisp, C. L. Perkins and J. L. Blackburn, ACSNano, 2013, 7, 7251.
10 J. Liu, N. Liu, J. Li, X. Li and J. Huang, Appl. Phys. Lett., 2012,101, 052104.
11 M. Engel, M. Steiner, A. Lombardo, A. C. Ferrari,H. V. Lohneysen, P. Avouris and R. Krupke, Nat. Commun.,2012, 3, 906.
12 T. Mueller, F. Xia and P. Avouris, Nat. Photonics, 2010, 4, 297.13 X. Huang, I. H. El-Sayed, W. Qian and M. A. El-Sayed, J. Am.
Chem. Soc., 2006, 128, 2115.14 P. K. Jain, X. Huang, I. H. El-Sayed and M. A. El-Sayed, Acc.
Chem. Res., 2008, 41, 1578.15 S. Nie and S. R. Emory, Science, 1997, 275, 1102.16 J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao and
R. P. Van Duyne, Nat. Mater., 2008, 7, 442.17 I. Chourpa, F. H. Lei, P. Dubois, M. Manfait and
G. D. Sockalingum, Chem. Soc. Rev., 2008, 37, 993.18 T. J. Echtermeyer, L. Britnell, P. K. Jasnos, A. Lombardo,
R. V. Gorbachev, A. N. Grigorenko, A. K. Geim, A. C. Ferrariand K. S. Novoselov, Nat. Commun., 2011, 2, 458.
19 W. Ren, Y. Fang and E. Wang, ACS Nano, 2011, 5, 6425.20 S. He, K. K. Liu, S. Su, J. Yan, X. Mao, D. Wang, Y. He, L. J. Li,
S. Song and C. Fan, Anal. Chem., 2012, 84, 4622.21 W. Xu, N. Mao and J. Zhang, Small, 2013, 9, 1206.22 W. Xu, J. Xiao, Y. Chen, Y. Chen, X. Ling and J. Zhang, Adv.
Mater., 2013, 25, 928.23 G. Lu, H. Li, C. Liusman, Z. Yin, S. Wu and H. Zhang, Chem.
24 Z. Zhang, F. Xu, W. Yang, M. Guo, X. Wang, B. Zhang andJ. Tang, Chem. Commun., 2011, 47, 6440.
25 X. Zhu, L. Shi, M. S. Schmidt, A. Boisen, O. Hansen, J. Zi,S. Xiao and N. A. Mortensen, Nano Lett., 2013, 13, 4690.
26 P. Wang, O. Liang, W. Zhang, T. Schroeder and Y. H. Xie,Adv. Mater., 2013, 25, 4918.
27 Y. Du, R. Liu, B. Liu, S. Wang, M.-Y. Han and Z. Zhang, Anal.Chem., 2013, 85, 3160.
28 X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner,A. Velamakanni, I. Jung, E. Tutuc, S. K. Banerjee,L. Colombo and R. S. Ruoff, Science, 2009, 324, 1312.
29 S. Bae, H. Kim, Y. Lee, X. Xu, J.-S. Park, Y. Zheng,J. Balakrishnan, T. Lei, H. R. Kim and Y. I. Song, Nat.Nanotechnol., 2010, 5, 574.
30 E. D. Palik, Handbook of optical constants of solids, Academic,New York, USA, 1985.
31 M. Bruna and S. Borini, Appl. Phys. Lett., 2009, 94, 031901.32 C. J. Powell, N. E. Erickson and T. Jach, J. Vac. Sci. Technol.,
1982, 20, 625.33 D. V. Leff, L. Brandt and R. Heath, Langmuir, 1996, 12, 4723.34 J. Li, C. Liu and Y. Liu, J. Mater. Chem., 2012, 22, 8426.35 H. Ishii, K. Sugiyama, E. Ito and K. Seki, Adv. Mater., 1999,
11, 605.36 K. L. Purvis, G. Lu, J. Schwartz and S. L. Bernasek, J. Am.
Chem. Soc., 2000, 122, 1808.37 T. Okamoto, I. Yamaguchi and T. Kobayashi, Opt. Lett., 2000,
25, 372.38 M. Jablan, H. Buljan and M. Solj�acıc, Phys. Rev. B: Condens.
Matter Mater. Phys., 2009, 80, 245435.39 X. Yu, L. Shi, D. Han, J. Zi and P. V. Braun, Adv. Funct. Mater.,
2010, 20, 1910.40 V. G. Kravets, F. Schedin, R. Jalil, L. Britnell, K. S. Novoselov
and A. N. Grigorenko, J. Phys. Chem. C, 2012, 116, 3882.41 X. Ling, L. M. Xie, Y. Fang, H. Xu, H. L. Zhang, J. Kong,
M. S. Dresselhaus, J. Zhang and Z. F. Liu, Nano Lett., 2010,10, 553.
42 J. F. Li, Y. F. Huang, Y. Ding, Z. L. Yang, S. B. Li, X. S. Zhou,F. R. Fan, W. Zhang, Z. Y. Zhou, D. Y. Wu, B. Ren, Z. L. Wangand Z. Q. Tian, Nature, 2010, 464, 392.
43 W. Xu, X. Ling, J. Xiao, M. S. Dresselhaus, J. Kong, H. Xu,Z. Liu and J. Zhang, Proc. Natl. Acad. Sci. U. S. A., 2012,109, 9281.
