Purdue UniversityPurdue e-Pubs

Birck and NCN Publications Birck Nanotechnology Center

9-30-2011

Carbon nanowalls amplify the surface-enhancedRaman scattering from Ag nanoparticlesChandra S. RoutBirck Nanotechnology Center, Purdue University

Anurag KumarBirck Nanotechnology Center, Purdue University, kumar50@purdue.edu

Timothy FisherBirck Nanotechnology Center, Purdue University, tsfisher@purdue.edu

Follow this and additional works at: http://docs.lib.purdue.edu/nanopub

Part of the Nanoscience and Nanotechnology Commons

This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact epubs@purdue.edu foradditional information.

Rout, Chandra S.; Kumar, Anurag; and Fisher, Timothy, "Carbon nanowalls amplify the surface-enhanced Raman scattering from Agnanoparticles" (2011). Birck and NCN Publications. Paper 947.http://dx.doi.org/10.1088/0957-4484/22/39/395704

Carbon nanowalls amplify the surface-enhanced Raman scattering from Ag nanoparticles

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2011 Nanotechnology 22 395704

(http://iopscience.iop.org/0957-4484/22/39/395704)

Download details:

IP Address: 128.46.221.64

The article was downloaded on 19/07/2013 at 15:59

Please note that terms and conditions apply.

View the table of contents for this issue, or go to the journal homepage for more

Home Search Collections Journals About Contact us My IOPscience

IOP PUBLISHING NANOTECHNOLOGY

Nanotechnology 22 (2011) 395704 (8pp) doi:10.1088/0957-4484/22/39/395704

Carbon nanowalls amplify thesurface-enhanced Raman scattering fromAg nanoparticlesChandra Sekhar Rout, Anurag Kumar and Timothy S Fisher

Birck Nanotechnology Center, School of Mechanical Engineering, Purdue University,West Lafayette, IN 47907, USA

E-mail: tsfisher@purdue.edu

Received 8 June 2011, in final form 10 August 2011Published 6 September 2011Online at stacks.iop.org/Nano/22/395704

AbstractWe report surface-enhanced Raman scattering (SERS) from Ag nanoparticles decorated on thincarbon nanowalls (CNWs) grown by microwave plasma chemical vapor deposition. The Agmorphology is controlled by exposing the CNWs to oxygen plasma and through theelectrodeposition process by varying the number of deposition cycles. The SERS substrates arecapable of detecting low concentrations of rhodamine 6G and bovine serum albumin, showingmuch higher Raman enhancement than ordinary planar HOPG with Ag decoration. The majorfactors contributing to this behavior include: high density of Ag nanoparticles, large surfacearea, high surface roughness, and the underlying presence of vertically oriented CNWs. Therelatively simple procedure of substrate preparation and nanoparticle decoration suggests thatthis is a promising approach for fabricating ultrasensitive SERS substrates for biological andchemical detection at the single-molecule level, while also enabling the study of fundamentalSERS phenomena.

S Online supplementary data available from stacks.iop.org/Nano/22/395704/mmedia

1. Introduction

SERS is a powerful surface analysis tool that offers molecularspecificity and extraordinary sensitivity for single-moleculedetection [1, 2]. SERS can magnify weak Raman signalsby several orders of magnitude and can facilitate the specificdetection of biological and chemical materials in a non-destructive manner. Recent progress has produced significantinterest in designing SERS substrates in order to gainfundamental insights into this enhancement phenomenon,and a wide range of interesting applications have beenproposed from noninvasive biological detection to single-nucleotide identification in DNA sequencing [3]. The majorityof recent SERS work has been conducted on Ag or Aufilms, lithographically patterned substrates or nanoparticleaggregations, which facilitate large enhancements of Ramanscattering signals from labels that are in close proximity tonoble metal surfaces [4–6]. In the present work, we report anew underlying SERS substrate, CNWs, that provides furtherenhancement.

The fundamental SERS effect is believed to be causedby the excitation of surface plasmon resonances on the

metal surface that greatly strengthen the local electric field.To achieve ultrasensitive detection of molecules by SERS,the substrates should possess high surface area to adsorbmolecules and should have a large density of ‘hot spots’comprised of metal nanostructures to enhance the electricfield [7, 8]. Also, surface roughness effects, orientationand bonding of the detecting molecule at SERS substrate,and the immersion process significantly affect the SERSsignal. SERS-active surfaces can be produced by proceduresincluding electrochemical roughening of metal foils [9, 10],nanosphere lithography followed by metal deposition [11, 12],self-organized nanoporous templates of anodized aluminumoxide [13, 14], and self-assembled surface nanoparticles [15].

Active surfaces allow the use of SERS in microfluidicsstructures [16], time evolution studies [17] and electrochemicalSERS investigations [16]. However, most SERS substratesused in prior work give either limited sensitivity or non-reproducible and unstable Raman signals because of limitedand inconsistent surface area. Consequently, the needremains for further progress in improving the sensitivityand reproducibility of SERS substrates, and we posit that

0957-4484/11/395704+08$33.00 © 2011 IOP Publishing Ltd Printed in the UK & the USA1

Nanotechnology 22 (2011) 395704 C S Rout et al

improvements in surface composition and morphology canenable such advances.

Graphene, a single-atom-thick and two-dimensionalhoneycomb carbon lattice has attracted great attentionbecause of its remarkable electronic, mechanical and thermalproperties [18a]. Prior work has demonstrated that graphenecan be used as a substrate to suppress the fluorescencebackground by ∼103 times, and therefore can be used as partof a hybrid SERS substrate to detect fluorescent molecules atlow concentrations due to its high surface area [19]. Also, few-layer graphene can provide a useful platform for studying thechemical enhancement in SERS because its structure can beeasily modified and controlled through physical and chemicalprocesses, and it can facilitate electron transfer with adsorbedsurface molecules [20]. Upon formation of a heterojunctionbetween graphene and noble metals, several characteristicproperties of graphene can be enhanced [21]. Similarly,CNWs, which consist of layers of graphene, have attractedextensive attention due to their unique morphology andgeometric shape [22]. Recently, it has been shown that CNWshave potential application in energy storage [23], electrodesfor fuel cells [24], field emission [25], and biosensors [26],and the room-temperature ferromagnetic properties of CNWssuggests possible application in spintronics [27]. The largesurface area of aligned CNWs creates an ideal template for thefabrication of other nanostructured materials [28]. Recently,we have demonstrated that CNWs (also referred to as graphiticpetals) decorated with Au nanoparticles can act as a goodSERS substrate for detecting rhodamine 6G due to sharpnanoscale features and the high surface area [29]. Similarly,silver is a very good surface-active material with unique opticalproperties due to their localized surface plasmon resonance andcan produce extraordinarily high Raman enhancement (∼107

times) and is a lower-cost material compared to gold. Agparticles exhibit a surface plasmon band between ∼390 and490 nm, depending upon the particle size, which is a spectralregime that is distinct from that of Au (520–580 nm) [30]. Theextinction coefficient of the surface plasmon band for an Agparticle is approximately four times as large as that for an Auparticle of the same size [31]. Also, Ag nanoparticles possesslow toxicity to human cells, antibacterial properties [32], highthermal stability [33] and low volatility [34]. Also the workfunction of Ag is 4.26 eV which is lower than those of Au(5.1 eV) and graphene (4.4 eV) [18e]. Hence a charge transferprocess can take place in the case of a SERS substrate basedon few-layer graphene and Ag nanoparticles for fluorescencesuppression. In this regard, we have fabricated heterojunctionsubstrates of CNWs consisting of few-layer graphene and Agnanostructures with the aim of quantifying SERS behavior fordifferent molecules in terms of sensitivity and repeatability fora highly fluorescent molecule (rhodamine 6G, R6G), a protein(bovine serum albumin, BSA), and hexadecanethiol.

2. Experimental details

CNWs were grown on flat Ti-coated Si(100) substrates throughadaptation of a previously reported procedure used on carbonfibers and Si [27, 29]. Substrates cleaned with acetone were

placed in the chamber on an Mo puck and were kept at aheight of 6 mm from the puck surface by ceramic spacers.Uniform growth of CNWs was obtained under microwaveplasma chemical vapor deposition (MPCVD) conditions of H2

(50 sccm) and CH4 (10 sccm), as the primary feed gases, at30 Torr total pressure and 700 W plasma power with 900 ◦Csubstrate temperature. It should be noted that uniformity ofthe CNW growth significantly depends on the cleanliness andheight of the ceramic spacers. We have achieved uniformgrowth of CNWs on 2 × 2 cm2 substrates by placing thesubstrates on freshly cleaned ceramic spacers and also wewere able to grow uniform CNWs on several substrates atsame time. An optical image of a CNW sample with uniformgrowth has been included in the supporting information (S1available at stacks.iop.org/Nano/22/395704/mmedia). CNWswere functionalized prior to electrodeposition of Ag by oxygenplasma treatment using inductively coupled plasma at a radiofrequency of 13.56 MHz at 50 sccm (cubic centimeter perminute at STP) oxygen flow rate and gas pressure of 60 mTorrfor 15 s. Silver electrodeposition was performed in aBASi Epsilon three-electrode cell stand with an Ag/AgClreference electrode, Pt gauge counter electrode and a metalcontact electrode connecting the CNWs and acting as theworking electrode [35]. The samples were dipped into thesilver deposition solution, which contained 1 mM AgNO3

(99%+, Sigma-Aldrich) with 0.2 M KNO3 (Sigma-Aldrich)as the supporting electrolyte. Ag was galvanostaticallyelectrodeposited on CNWs by applying 500 ms pulses of2 mA cm−2 current between the auxiliary and workingelectrodes for different cycles. After Ag electrodeposition, thesamples were rinsed in DI water and allowed to air-dry.

R6G, BSA and hexadecanethiol were purchased fromSigma-Aldrich and were used for SERS studies without furtherpurification. Different concentrations of R6G (10−6, 10−7 and10−8 M) and BSA (10−3 and 10−6 M) were prepared in high-purity water. Soaking was used to deposit the molecules onthe surface of CNWs with and without Ag decoration. Thesoaking time was 30 min for all the cases. After soaking,the samples were repeatedly rinsed with plenty of water toremove unadsorbed molecules and were then dried for theSERS experiments. Also, we have studied SERS propertiesof Ag-coated HOPG (supporting information, S3, availableat stacks.iop.org/Nano/22/395704/mmedia) by following thesame experimental procedure of oxygen plasma treatment andelectrodeposition method as for the CNW sample.

The morphology of the carbon nanowalls was observedusing a field emission scanning electron microscope (FESEM;Hitachi S4800) and a transmission electron microscope (TEM,Titan 80–300 kV). Nanowall thicknesses were measuredfrom the TEM analysis. The XPS investigation used anAXIS ULTRA DLD system under ultra-vacuum (<10−9 Torr)conditions with a lateral resolution of 5 mm. Raman spectrawere collected by using the SENTERRA confocal Ramansystem (Bruker Optics Inc., Billerica, MA) with a 20×air objective at 633 nm excitation. The laser power andaccumulation time were 20 mW and 10 s at 633 nm excitation.

2

Nanotechnology 22 (2011) 395704 C S Rout et al

Figure 1. FESEM images of CNWs: (a) top and (b) cross-sectional views. (c) TEM image of the CNWs and (d) HRTEM image of a carbonnanowall.

3. Results and discussion

Figures 1(a) and (b) show FESEM images of CNWs grownon a Ti-coated Si(100) substrate. The typical height of thenanowall arrays is ∼500 nm and the lateral length of thenanowall varies between 200 and 1000 nm. The nanowallthicknesses were in the range of 1–10 nm as measured bycross-sectional TEM and atomic force microscopy [27, 29].Figure 1(c) confirms that the nanowalls are very thin,consisting of few-layer graphene. The lattice spacing from theTEM analysis (figure 1(d)) is approximately 0.34 nm, which isconsistent with graphite (002) planes.

XPS analysis reveals the existence of C and O peaks inthe survey-scan spectrum, and peak positions are consistentwith prior literature [36]. Figure 2 shows the XPS spectrum,highlighting the C 1s and O 1s peaks from the CNWs. TheC 1s peak can be precisely deconvoluted into three Gaussianpeaks with the main peak centered at 284.7 eV assigned tosp2 bonds; the secondary peak at 285.2 eV corresponds to sp3

hybridization. The small amount of sp3 hybridization is likelyto occur primarily at the boundary regions of the nanowalls,where many small twists and protuberances exist [36]. Thethird peak near 291 eV is broad and corresponds to carboxylate(O–C=C) [37] and confirms adsorbed oxygen species onthe surface of CNWs. The presence of oxygen is furthersubstantiated by the Gaussian fit of the O 1s peak (inset,figure 2). Quantitative analysis of the peak intensities indicatesan atomic fraction of ∼4% oxygen atoms on the CNWs.

Figure 3 shows FESEM images of the Ag-decoratedCNWs after an increasing number of electrodeposition cycles.For 50 cycles, a homogeneous distribution of discrete andspherical Ag nanoparticles with diameters near 30 nm isobserved on the nanowalls with no agglomeration (figure 3(a)).On increasing the number of electrodeposition cycles to 100,the Ag nanoparticles grow larger, with diameters in the range100–150 nm (figure 3(b)). Nanoparticle agglomerates were

Figure 2. XPS spectra of C 1s and O 1s (inset) collected from theCNWs.

(This figure is in colour only in the electronic version)

observed above 200 cycles of electrodeposition, as observedin figures 3(c) and (d).

The shapes and morphologies of nanostructured crystalsor polycrystals are strongly affected by the driving force ofcrystallization and the rate of diffusion of reactants, as evidentfrom the electrodeposition of Ag nanoparticles at higher cyclenumbers (figures 3(c) and (d)). The branched and dendriticAg structures formed above 250 cycles (see supportinginformation, S2, available at stacks.iop.org/Nano/22/395704/mmedia) are believed to be caused by the preferentialgrowth of Ag nuclei along the energetically favorable 〈111〉direction [38, 39]. We note that similar sharp morphologicalfeatures were not observed for Au particles in prior workand only Au agglomerates were formed on increasing the

3

Nanotechnology 22 (2011) 395704 C S Rout et al

Figure 3. FESEM images of Ag-electrodeposited CNWs for (a) 50 cycles, (b) 100 cycles, (c) 200 cycles and (d) 300 cycles.

deposition cycles [29]. During electrodeposition, a criticalsurface field that is a physical constant of the surface isachieved for the nucleation of a silver grain and eventuallyforms the threefold dendrite structure with longer growthduration. In the initial stage, the mechanism of depositionfollows reduction–nucleation growth phenomena at severalsites to form nanoparticles, where metals are discharged fromtheir complex under diffusion control [40]. On increasing thegrowth duration, the concentration of Ag salt greatly decreases;and during this period, the growth is mainly driven by surfacediffusion and thus dendrite morphology is favored [39]. TheSEM images taken from the Ag-decorated CNWs at 250 cycles(S2(a) and (b) available at stacks.iop.org/Nano/22/395704/mmedia) confirm that the Ag dendrite formation followsdiffusion-limited aggregation [38]. In the diffusion-limitedaggregation regime, growth takes place by the bonding ofindividual atoms to the substrate (CNWs) after the reduction ofthe metal ions. After reaching a certain grain size, the growthof the original grain begins to slow, and new grains nucleatefrom the parent grains. This process of grain size saturationand nucleation can then repeat to form higher-dimensionaldendritic structures. The absence of Au dendrite formationunder similar experimental conditions might be due to slowsurface diffusion, which is strongly dependent on the bondenergies of Au–C and Au–Au interactions [41].

Figure 4 shows Raman spectra of bare and Ag-decoratednanowalls. The Raman peaks obtained for the CNWs are ingood agreement with Raman spectra of few-layer grapheneand CNWs in prior literature (figure 4(a)) [22, 42, 43]. Theso-called disorder-induced D band appears near 1340 cm−1,and the tangential mode G is at 1580 cm−1. The second-order 2D, D + G and 2D′ bands appear at 2665, 2928 and3242 cm−1 respectively. Also, a weak band correspondingto D′ is observed at 1618 cm−1, indicating disorder in theCNWs [29]. The D band is due to the presence of defects suchas distortion, vacancies and strain in graphitic networks due

Figure 4. Raman spectra of (a) CNWs and (b) Ag nanoparticlesdecorated CNWs.

to the finite crystallite size and is associated with transverseoptical (TO) vibrations near the K point [22]. The 2D bandis the result of second-order Raman scattering involving TOphonons near the K point. The ID/IG and IG/I2D ratios are0.48 and 1.7 respectively, which are nearly same as thosereported for similar few-layer graphene [44, 45]. The majorityof nanowalls here are expected to contain at least five layerssince the intensity of the 2D mode is lower than that of the Gband [42]. Kurita et al observed that the ID/IG ratio stronglydepends on the length of the CNWs (LW) and decreaseslinearly with increasing LW [46]. They confirmed that CNWswith a narrower G band consist of small grain graphite-likecarbon with high degree of graphitization. Since the CNWsgrown by us show a similar Raman spectrum with a narrower

4

Nanotechnology 22 (2011) 395704 C S Rout et al

Figure 5. Raman spectra of 10−7 M R6G on (a) CNWs, (b) Agdeposited HOPG. Raman spectra of (c) 10−6 M, (d) 10−7 M and(e) 10−8 M R6G on Ag-decorated CNWs.

full width of half-maximum of the ‘G’ band (∼35 cm−1),our CNWs are the types of graphite-like carbon consistingof small crystallites with a higher degree of graphitization.Upon Ag deposition, the relative intensities of the D and Gbands change to produce an ID/IG ratio of 0.82. Also, weobserve a Raman band shift toward lower wavelengths uponAg decoration, but the shift is less than that observed for Au-decorated CNWs. A possible explanation for this behaviorrelates to the strengthening of the C–C bond because of theadditional electron density provided by nanoparticles [45].The ID/IG ratio of Au-decorated CNWs was ∼1.01, whichconfirms the formation of a lower number of defective sitesin the case of Ag-decorated CNWs [29].

We have selected the R6G molecule for SERS studies,since it is a strongly fluorescent xanthene derivative. R6Gis a yellowish heterocyclic compound and shows a molecularresonance Raman effect when excited into its visibleabsorption band [47, 48]. We have recorded Raman spectraof R6G molecules on CNWs, and Ag deposited on HOPG andnanowalls following the procedure documented in section 2.We did not observe any significant change in the Raman spectraof bare CNWs after R6G adsorption (figure 5(a)), whereas abroad Raman peak was observed in the case of Ag-decoratedHOPG (figure 5(a)). Figures 5(c)–(e) contain a series ofRaman spectra of varying R6G concentrations (10−6, 10−7 and10−8 M) on Ag-decorated CNWs. At lower concentrations,relatively intense SERS peaks were observed as compared tothose for the more highly concentrated R6G solutions.

The SERS results confirm that Ag-decorated nanowallscreate a hybrid SERS substrate for detecting R6G as low as10−8 M, a level that is difficult to detect through conventionalRaman spectroscopy. We have observed highly surface-enhanced and reproducible peaks of R6G molecules from theAg-decorated nanowalls at several positions on the sampleand on different samples. Hence, this result indicates strongquenching of the R6G fluorescence by the Ag nanoparticles onthe nanowall backing. The assignment of the obtained SERS

Table 1. Tentative assignment of peaks from the SERS spectra ofR6G molecules adsorbed on Ag-decorated CNWs.

Raman shift (cm−1)(this work)

Raman shift(cm−1) ([47, 51]) Assignment

608 614 C–C–C ring, in-plane bend770 774 C–H, out-of-plane bend

1122 1128 C–H, in-plane bend1182 1183 C–C stretching1302 1310 Aromatic C–C stretching1360 1362 C–N stretching1505 1508 Aromatic C–C stretching1582 1574 Aromatic C–C stretching1651 1650 Aromatic C–C stretching

peaks of R6G molecules on Ag-decorated nanowalls is shownin table 1. The band near 770 cm−1 corresponds to out-of-plane bending of hydrogen atoms from the xanthene skeleton,whereas the band near 608 cm−1 is due to plane bending of theC–C–C ring [48, 49]. The peak at 1122 cm−1 is associated withC–H in-plane bending, and the peak at 1176 cm−1 is due to C–C stretching. The peaks at 1302, 1505, 1582 and 1651 cm−1

correspond to aromatic stretching vibrations [3, 48]. Weobserved a red shift for most Raman peaks (except the peaksat 1182, 1302 and 1360 cm−1) as compared to the Au-decorated CNW case. The shift in the peak positions is causedby different surface roughness of the SERS substrate, andthe roughened surface significantly affects C–H vibrations ascompared to C–C stretching modes [50]. Also, the intensitiesof the Raman peaks obtained were ∼3 times higher in the caseof Ag-decorated CNWs compared to Au-decorated nanowalls.The enhancement can be attributed as due to higher opticalfields created near the Ag surface as compared to the Au case.Raman mapping of the SERS signal of R6G molecules onAg-decorated CNWs (S4 available at stacks.iop.org/Nano/22/395704/mmedia) and HOPG (S5 available at stacks.iop.org/Nano/22/395704/mmedia) illustrates the uniformity of goodSERS performance over the CNW substrate.

To demonstrate the broader applicability for sensitivedetection using the Ag-decorated CNWs, we measuredthe Raman spectra of two other kinds of molecules: 1-hexadecanethiol and BSA. Alkanethiols are sensitive probesfor comparing the effect of surface roughness on the bondingand ordering of a SERS-active substrate. SERS of alkanethiolscan identify changes in the sulfur head group, carbon backboneconformations, and orientation of the methylene and methylgroups [51]. We have chosen 1-hexadecanethiol, a long-chainalkanethiol, for this purpose. The SERS spectra obtained foron Ag-decorated nanowalls and HOPG are shown in figure 6.Multiple SERS spectra for 1-hexadecanethiol were obtainedwith Ag-decorated nanowalls with high reproducibility and inclose agreement with prior SERS reports [50]. In the caseof Ag-decorated HOPG, a broad, featureless Raman band wasobserved, suggesting that the nanowall structure plays a crucialrole in SERS detection (figure 6(b)). The band near 500–750 cm−1 corresponds to the C–S stretching vibration andis sensitive to the conformation, gauche or trans, of the twoadjacent carbon atoms. The peak at 654 cm−1 is assignedto a gauche conformation, and the peak at 743 cm−1 is due

5

Nanotechnology 22 (2011) 395704 C S Rout et al

Figure 6. Raman spectra of 1-hexadecanethiol onAg-electrodeposited (a) CNWs and (b) HOPG.

Table 2. Tentative assignment of peaks from the SERS spectra of1-hexadecanethiol adsorbed on Ag-decorated CNWs.

Raman shift (cm−1)(this work)

Raman shift(cm−1) [50–52] Assignment

518 520–650 C–S stretching654 655 C–Sgauche stretching699 705 CH2 rocking743 744 C–Strans stretching891 891 CH3 rocking

1059 1058 C–Cgauche stretching1096 1096 C–Ctrans stretching1123 1127 C–Ctrans stretching1294 1296 C–Htrans stretching1439 1440 C–Htrans stretching2571 2570 S–H stretching2721 2719 S–H stretching2846 2842 C–Htrans stretching2881 2876 C–Htrans stretching

to a trans conformation. In the case of Ag-decorated HOPGthe C–S stretching vibration peaks were observed at 518 and621 cm−1. The peaks at 699 and 891 cm−1 reflect CH2 andCH3 rocking vibrations respectively [52]. The bands 1000–1200 cm−1 and 2750–3050 cm−1 correspond to C–C andC–H stretching vibrations [50]. A high-intensity peak near2881 cm−1 is proposed to be due to a methyl vibration in Fermiresonance with the overtone of the ‘C–H’ bending vibration at∼1450 cm−1. The peaks at 2571 and 2721 cm−1 are due to theS–H vibrations. The assignment of the peaks obtained for onAg-decorated nanowalls is shown in table 2.

The next tested species, BSA, is an abundant water-soluble protein that protects cellular embryos and promotescell reproduction. The Raman spectra of BSA on ourSERS substrates are similar to those from previous SERSstudies [53], though some new peak positions and differencesin relative intensities are observed. Figure 7 shows the Ramanspectra of BSA on Ag-decorated HOPG and CNWs. Only

Figure 7. Raman spectra of (a) 10−3 M BSA on Ag depositedHOPG, and (b) 10−6 M and (c) 10−3 M BSA on Ag-decoratedCNWs.

Table 3. Tentative assignment of peaks from the SERS spectra ofBSA adsorbed on Ag-decorated CNWs.

Raman shift (cm−1)(this work)

Raman shift(cm−1) [53] Assignment

610 617 Phenylalanine621 625 Phenylalanine670 667 C–S vibration755 750 Tryptophan846 852 Tyrosine910 903 C–C vibration960 960 C–C vibration999 1004 Phenylalanine (aromatic ring)

1124 1128 C–N vibration1156 1159 C–N vibration1215 1207 Amide III1278 1272 Amide III1330 1320 –CH2

1380 1378 –CH2

1440 1449 –CH2

1510 1507 O–H1644 1649 Amide I

Raman peaks of graphite are apparent in the case of Ag-decorated HOPG, whereas enhanced Raman peaks of BSAadsorbed on Ag-decorated nanowalls are clearly visible at aconcentration of 10−3 M, which is a common concentrationfor the SERS enhancement [53]. We did not observe enhancedBSA peaks at concentrations below 10−3 M (figure 7(b)).The peak assignments for the adsorbed BSA on Ag-decoratednanowalls are shown in table 3. According to previous Ramanscattering studies of BSA [53], peaks at 610, 621 and 999 cm−1

derive from the phenyl ring, 670 cm−1 to the C–S bond,846 cm−1 to tyrosine, those at 910 and 960 cm−1 to C–C, 1124and 1156 cm−1 to C–N, 1215 and 1278 cm−1 to amide III,1330, 1380 and 1440 cm−1 to CH2 vibrations and 1644 cm−1 toamide I. Notably, the D band of the graphite nanowalls overlapswith the CH2 vibration arising due to the aliphatic side chains,and the G band from the nanowalls is distinct.

6

Nanotechnology 22 (2011) 395704 C S Rout et al

These results demonstrate that Ag-decorated CNWsare widely applicable as SERS substrates for identifyinga diverse set of molecules. The substrates exhibit muchstronger enhancement than ordinary planar substrates withAg nanoparticles. We have confirmed the SERS effect onseveral nanowall samples decorated with Ag nanoparticlesand observed high reproducibility and stability. SimilarRaman signals were obtained from the SERS substrateseven after several weeks. SERS signals of 10−8 M R6Gobtained from two sets of Ag-decorated CNWs after 12 and24 weeks (S6 available at stacks.iop.org/Nano/22/395704/mmedia) show similar Raman signals, with retention ofall main R6G peaks though with some changes in relativeintensities; this result indicates reasonably good long-termstability and reproducibility. We observed no structural andmorphological distortion of the CNW arrays after severalrepeated measurements, due to better adhesion of the nanowallarrays to the substrate. We observed similar SERS spectra ofR6G, BSA and 1-hexadecanethiol for different metal particlemorphologies ranging from isolated nanoparticles to dendritesconsisting of coalesced nanoparticles. Also, the observedSERS signals were uniform over a large area of the samples.

There are two major mechanisms believed to be importantin the enhancement of the Raman signal: an electromagneticeffect associated with large fields due to plasmon resonanceoccurring in the microstructures on the metal surface, anda chemical effect involving a scattering process associatedwith chemical interactions between the molecule and theSERS substrate. The electromagnetic effect is known to bea few orders of magnitude greater than that due to the chargetransfer involving chemical effects [1, 5]. The electromagneticmechanism is based upon the optical properties of the noblemetals and their ability to support plasmon resonances atvisible wavelengths. Resonant excitation of surface plasmonscreates an enhanced surface field (Es) at both incoming andoutgoing frequencies, resulting in an enhancement of theRaman signal that roughly scales as E4

s [5, 54]. Also, Es isequal to sum of the electromagnetic field on the analyte inthe absence of any roughness features and the electromagneticfield emitted from the particulate metal roughness features.

In our case, both mechanisms are expected to enhanceSERS signals from substrates based on Ag and CNWsbecause Ag nanostructures are known to exhibit plasmonresonance, whereas few-layer graphene promotes the chemicalenhancement mechanism for SERS, which has been observedrecently [20, 55]. The absence of SERS signal from thebare CNWs suggests that the electromagnetic effect of thesurface plasmons at the Ag nanoparticle surface may bedominant in the present study. However, the CNWs provide athree-dimensional support to the Ag nanoparticles, increasingexposure of the detecting molecules from all the directions.Both the size and shape of the nanoparticles greatly enhancethe SERS properties because these factors influence the ratioof adsorption and scattering events, and this effect is moresignificant when the particle sizes are in the nanometerscale regime [56]. Since the sizes of the Ag crystals onelectrodeposited HOPG substrates (∼0.5–1 mm) are muchlarger than those of the Ag nanoparticles grown on CNWs, we

did not observe any SERS signal from the Ag/HOPG substrate.The density of Ag nanoparticle decoration, high surfaceroughness, high surface area and three-dimensional structure ofthe CNWs appear to be the major reasons underlying the ultra-high enhancement as compared to Ag-decorated HOPG. FromFESEM analysis, the density of Ag nanoparticles decorationper unit base area on the CNWs is ∼103 µm−2 and is ∼102

times higher compared to the HOPG one.

4. Conclusion

In conclusion, we report a facile and versatile substrate forsurface-enhanced Raman spectroscopy with highly sensitivedetection. Advantages of this substrate include its easeof fabrication over an area of several square centimeterswithout the need for advanced microscale processes such aslithography. Further improvement and optimization of thissystem could produce superior performance as compared tothose of planar SERS substrates due to the unique combinationof 0D (nanoparticle) and 2D (graphene) base materials.

References

[1] Campion A and Kambhampati P 1998 Chem. Soc. Rev.27 241

[2] Alivisatos P 2004 Nature Biotechnol. 22 47[3] Nie S and Emory S R 1997 Science 275 1102[4] Albrecht M G and Creighton J A 1977 J. Am. Chem. Soc.

99 5215[5] Moskovits M 1985 Rev. Mod. Phys. 57 783[6] Kneipp K, Wang Y, Kneipp H, Perelman L T, Itzkan I,

Desai R R and Feld M S 1997 Phys. Rev. Lett. 78 1667[7] Garcia-Vidal F J and Pendry J B 1996 Phys. Rev. Lett.

77 1163[8] Xu H X, Azipurua J, Kall M and Apell P 2000 Phys. Rev. E

62 4318[9] Jeanmaire D L and Van Duyne R P 1977 J. Electroanal.

Chem. 84 1[10] Fleischmann M, Hendra P J and McQuillan A J 1974 Chem.

Phys. Lett. 26 163[11] Sugawara Y, Kelf T A, Baumberg J J, Abdelsalam M E and

Barlett P N 2006 Phys. Rev. Lett. 97 266808[12] Haynes C L and Van Duyne R P 2001 J. Phys. Chem. B

105 5599[13] Wang H H, Liu C Y, Wu S B, Liu N W, Peng C Y, Chan T H,

Hsu C F and Wang J K 2005 Adv. Mater. 17 222[14] Choi D, Choi Y, Hong S, Kang T and Lee L P 2010 Small 6

1741[15] Freeman R G et al 1995 Science 267 1629[16] Liu G L and Lee L P 2005 Appl. Phys. Lett. 87 074101[17] Corio P, Santos P S, Barr V W, Samsonidze G G, Chou S G

and Dresselhaus M S 2003 Chem. Phys. Lett. 370 675[18a] Wu J S, Pisula W and Mullen K 2007 Chem. Rev. 107 718[18b] Bunch J S, van der Zande A M, Parpia J M, Craighead H G

and McEuen P L 2007 Science 315 450[18c] Ling X, Xie L M, Fang Y, Xu H, Zhang H L, Kong K,

Dresselhaus M S, Zhang J and Liu Z F 2010 Nano Lett.10 553

[18d] Novoselov K S, Geim A K, Morozov S V, Jiang D, Zhang Y,Dubonos S V, Grigorieva I V and Firsov A A 2004 Science306 666

[18e] Sque S J, Jones R and Briddon P R 2007 Phys. Status Solidi a204 3078

[19] Xie L, Ling X, Fang Y, Zhang J and Liu Z 2009 J. Am. Chem.Soc. 131 9890

7

Nanotechnology 22 (2011) 395704 C S Rout et al

[20] Yu X, Cai H, Zhang W, Li X, Pan N, Liu Y, Wang X andHou J G 2011 ACS Nano 5 952

[21] Goncalves G, Marques P A A P, Granadeiro C M,Nogueira H I S, Singh M K and Gracio J 2009 Chem.Mater. 21 4796

Hu Z L, Aizawa M, Wang Z M, Yoshizawa N andHatori H 2010 Langmuir 26 6681

Yoo E, Okata T, Akita T, Kohyama M, Nakamura J andHonma I 2009 Nano Lett. 9 225

Sundaram R S, Gomez-Navarro C, Balasubramanian K,Burghard M and Kern K 2008 Adv. Mater. 20 3050

[22] Wu Y, Qiao P, Chong T and Shen Z 2002 Adv. Mater. 14 64Chuang A T H, Robertson J, Boskovic B O and Koziol K K K

2007 Appl. Phys. Lett. 90 123107[23] Hung T C, Chen C F and Whang W T 2009 Electrochem.

Solid State Lett. 12 K41Tao M, Li T, Bi H, Wang L, Zhu Y and Qian Y 2010 Eur. J.

Inorg. Chem. 2010 4314[24] Tanaike O, Kitada N, Yoshimura H, Hatori H, Kojima K and

Tachibana M 2009 Solid State Ion. 180 381[25] Wang J J, Zhu M Y, Outlaw R A, Zhao X, Manos D M,

Holloway B C and Mammana V P 2004 Appl. Phys. Lett.85 1265

Srivastava S K, Shukla A K, Vankar V D and Kumar V 2005Thin Solid Films 492 124

[26] Shang N G, Papakonstantinou P, McMullan M, Chu M,Stamboulis A, Potenza A, Dhesi S S and Marchetto H 2008Adv. Funct. Mater. 18 3506

[27] Rout C S, Kumar A, Kumar N, Sundaresan A and Fisher T S2011 Nanoscale 3 900

[28] Yang B J, Wu Y H, Zong B Y and Shen Z X 2002 Nano Lett.2 751

Wu Y H, Yang B J, Han G C, Zong B Y, Ni H Q, Luo P,Chong T C, Low T S and Shen Z X 2002 Adv. Funct.Mater. 12 489

Wu Y, Yang B, Zong B, Sun H, Shen Z and Feng Y 2004J. Mater. Chem. 14 469

[29] Rout C S, Kumar A, Xiong G, Irudayaraj J and Fisher T S2010 Appl. Phys. Lett. 97 133108

[30] Mulvaney P 1996 Langmuir 12 788Cao Y W, Jin R and Mirkin C A 2001 J. Am. Chem. Soc.

123 7961[31] Link S, Wang Z L and El-Sayed M A 1999 J. Phys. Chem. B

103 3529[32] Gong P, Li H, He X, Wang K, Hu J and Tan W 2007

Nanotechnology 18 285604[33] Duran N, Marcarto P D, De Souza G I H, Alves O L and

Esposito E 2007 J. Biomed. Nanotechnol. 3 203[34] Singh P, Thuy N T B, Aoki Y, Mott D and Maenosono S 2011

J. Appl. Phys. 109 094301[35] Claussen J C, Franklin A D, Haque A U, Porterfield D M and

Fisher T S 2009 ACS Nano 3 37[36] Jiang N, Wang H X, Zhang H, Sasaoka H and

Nishimura K 2010 J. Mater. Chem. 20 5070[37] Ago H, Kugler T, Cacialli F, Salaneck W R, Shaffer M S P,

Windle A H and Friend R H 1999 J. Phys. Chem. B38 8116

[38] Kaniyankandy S, Nuwad J, Thinaharan C, Dey G K andPillai C G 2007 Nanotechnology 18 125610

[39] Lin T-H, Lin C W, Liu H-H, Sheu J-T and Hung W-H 2011Chem. Commun. 47 2044

[40] Qiu T, Wu X L, Mei Y F, Chu P K and Siu G G 2005 Appl.Phys. A 81 669

[41] Sobri S, Roy S, Aranyi D, Nagy P M, Paap K andKalman E 2008 Surf. Interface Anal. 40 834

[42] Rao C N R, Sood A K, Subrahmanyam K S andGovindaraj A 2009 Angew. Chem. Int. Edn 48 7752

[43] Ni Z H, Fan H M, Feng Y P, Shen Z X, Yang B J and Wu Y H2006 J. Chem. Phys. 124 204703

[44] Chae S J et al 2009 Adv. Mater 21 2328[45] Yeh J-M, Huang K-Y, Lin S-Y, Wu Y-Y, Huang C-C and

Liou S J 2009 J. Nanotechnol. 2009 217469Kim Y K, Na H-K, Lee Y W, Jang H, Han S W and Min D H

2010 Chem. Commun. 46 3185[46] Kurita S, Yoshimura A, Kawamoto H, Uchida T, Kojima T,

Tachibana M, Morales P M and Nakai H 2005 J. Appl.Phys. 97 104320

[47] Hassell T, Lagonigro L, Peacock A C, Yoda S, Brown P D,Sazio P J A and Howdle S M 2008 Adv. Funct. Mater.18 1265

[48] Hildebrandt P and Stockburger M 1984 J. Phys. Chem.88 5935

[49] Sun Y, Liu K, Miao J, Wang Z, Tian B, Zhang L, Li Q,Fan S and Jiang K 2010 Nano Lett. 10 1747

[50] Sobocinski R L, Bryant M A and Pemberton J E 1990 J. Am.Chem. Soc. 112 6177

Bryant M A and Pemberton J E 1991 J. Am. Chem. Soc.113 3629

Bryant M A and Pemberton J E 1991 J. Am. Chem. Soc.113 8284

[51] Tao A, Kim F, Hess C, Goldberger J, He R, Sun Y, Xia Y andYang P 2003 Nano Lett. 3 1229

Levin C S, Janesko B G, Bardhan R, Scuseria G E,Hartgerink J D and Halas N J 2006 Nano Lett. 6 2617

Tao A R and Yang P 2005 J. Phys. Chem. B 109 15687[52] Joo T H, Kim K and Kim M S 1989 J. Phys. Chem. 90 5816[53] Xie C, Li Y Q, Tang W and Newton R J 2003 J. Appl. Phys.

94 6138David C, Guillot N, Shen H, Toury T and Chapelle M L 2010

Nanotechnology 21 475501Chen M C and Lord R C 1976 J. Am. Chem. Soc. 98 990Das G, Mecarini F, Gentile F, Angelis F D, Kumar H G M,

Candeloro P, Liberale C, Cuda G and Fabrizio E D 2009Biosens. Bioelectron. 24 1693

Shao M-W, Lu L, Wang H, Wang S, Zhang M L, Ma D D andLee S-T 2008 Chem. Commun. 2310

[54] Stiles P L, Dieringer J A, Shah N C and Duyne R P V 2008Annu. Rev. Anal. Chem. 1 601

[55] Ling X and Zhang J 2010 Small 6 2020[56] Boyack R and Ru E C L 2009 Phys. Chem. Chem. Phys.

11 7398Kelly K L, Coronado E, Zhao L L and Schatz G C 2003

J. Phys. Chem. B 107 668

8