lable at ScienceDirect

Materials Chemistry and Physics 143 (2014) 1215e1221

Contents lists avai

Materials Chemistry and Physics

journal homepage: www.elsevier .com/locate/matchemphys

One-step synthesis of star-like gold nanoparticles for surfaceenhanced Raman spectroscopy

Carlo Morasso a,*, Dora Mehn a, Renzo Vanna a, Marzia Bedoni a, Elena Forvi a,Miriam Colombo b, Davide Prosperi a,b, Furio Gramatica a

a Laboratory of Nanomedicine and Clinical Biophotonics, Fondazione Don Carlo Gnocchi ONLUS, Via Capecelatro 66, 20148 Milan, ItalybNanoBioLab, Dipartimento di Biotecnologie e Bioscienze, Università degli Studi di Milano Bicocca, Piazza della Scienza 2, 20126 Milan, Italy

h i g h l i g h t s

� Star shaped gold nanoparticles are produced with a single step procedure.� Raman enhancement of nanostars is 50 times higher than the one given by nanospheres.� Micromolar concentrations of dyes can be detected directly in liquid.� Apomorphine was detected at micromolar concentration directly in liquid.� Duoscan drastically decrease the variability of SERS signals from solid substrates.

a r t i c l e i n f o

Article history:Received 16 February 2013Received in revised form4 October 2013Accepted 13 November 2013

Keywords:NanostructuresChemical synthesisRaman spectroscopy and scatteringSurfaces

* Corresponding author.E-mail address: cmorasso@dongnocchi.it (C. Mora

0254-0584/$ e see front matter � 2013 Elsevier B.V.http://dx.doi.org/10.1016/j.matchemphys.2013.11.024

a b s t r a c t

In this paper we present a new protocol for the synthesis of Star-Like Gold Nanoparticles (SGNs) by asimple one-step, room temperature procedure not involving the use of seeds or surfactants, that can beperformed in seconds in any laboratory without the need of special technologies. These particlesexhibited excellent properties for Surface Enhanced Raman Spectroscopy (SERS) and, when comparedwith spherical nanoparticles with similar size and concentration, showed enhancing factors from 10 to50 times higher depending on the dye and on the wavelength employed. SGNs could be used directly insuspension as single, non-aggregating particles and were shown to be active in a remarkably broad rangeof the light spectrum from green to near infrared. Moreover, SGNs were adsorbed on the surface of asilicon slide to prepare SERS active solid substrate. Despite the fact that the surface of the solid substratewas not perfectly homogeneous, the signals recorded from different positions acquired through DuoScanaveraging mode show excellent reproducibility, demonstrating how this simple and cheap protocol canbe applied in order to generate reliable and homogeneous SERS substrates.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

Surface Enhanced Raman Spectroscopy (SERS) is a populartechnique in bioanalytical chemistry and a potentially powerfulenabling technology for in vitro diagnostics. In fact, SERS combinesthe excellent chemical specificity of Raman spectroscopy with thegood sensitivity provided by enhancement of the signal that isobserved when the analysed molecule lies over (or very close to)the surface of metal nanoparticles [1].

A number of SERS-based bioanalytical assays have been re-ported in literature; however, most of them are based on solid

sso).

All rights reserved.

surfaces produced by a top-down approach [2,3] or by the gener-ation of “hot spots” randomly distributed through the passivationof positively charged surfaces with negatively charged nano-particles in order to generate the required vicinity of metallicnanostructures [4,5]. Instead, the use of nanoparticles in suspen-sion offers several advantages as collected SERS signals registeredare more stable over time required for the analysis, and spatiallyhomogeneous. Traditionally, this approach has been mainly basedon silver nanoparticles as gold colloids usually provide just a weakenhancement of Raman signal. However, gold nanoparticles aresuperior to silver in terms of chemical stability and suitability foruse in biological media [6,7]. Recently, multibranched or Star-LikeGold Nanoparticles (SGNs) have been proposed as a reliablenanostructure for SERS experiments as they shown peculiar plas-monic properties and are generally considered as the most efficient

Table 1Resume of the dimension (DLS) of particles obtained varying the amount of thedifferent solutions: HAuCl4$3H2O, (10 mg mL�1) and hydroquinone (11 mg mL�1)mixed. Particles were prepared on a 10 mL scale except for the one with 105 nmdiameter.

Diameter (DLS) Water HAuCl4 Hydroquinone

70 10 mL 50 mL 100 mL85 10 mL 100 mL 100 mL140 10 mL 200 mL 100 mL105 50 mL 500 mL 1000 mL

C. Morasso et al. / Materials Chemistry and Physics 143 (2014) 1215e12211216

nanostructure for this kind of experiments [8e11]. As a conse-quence the synthesis of this kind of nanostructure attracted manyefforts and several protocols are present in literature. Despite ofthese efforts, the synthesis of SGNs is still complex and requiresmultiple steps or the use of surfactant or polymers that are difficultto remove from the surface [12,13]. Recently a single step easyprotocol has been reported; yet this protocol still requires a tightcontrol of the temperature at which the process is performed [14].

In this paper, we present a novel, simple, room temperature andsurfactant-free method for the synthesis of SGNs with excellentSERS properties. The present method is based on the reduction ofgold ions (Au3þ) by hydroquinone, a weak reducing agent, whichacts preferentially on the [111] surface of gold colloids and whichhas been previously proven able to create branched nanoparticlesyet only in presence of a co-reducing agent or in presence of seeds[15,16].

According to the stoichiometry of the reaction (Fig. 1) reportedby a previous study [17] we decided to use an excess of hydroqui-none to change the reaction kinetics. This would shift the goldreduction reaction to a more kinetically controlled regime andcreate a preference for the growth of kinetically favourable particlesmorphology, as star-like particles, on the more thermodynamicallyfavourable nanospheres [18].

Our procedure allowed us to obtain SGNs with enhancingfactors up to 5 � 103 in one-step and at room temperature with agood control of the dimension. SERS behaviour of SGNs wascompared directly with the enhancing capacity of conventionalspherical gold nanoparticles with hydrodynamic diameter in thesame range at equivalent particle concentrations. In order to studythe potential of SGNs to be functionalized with various organicmolecules for bioanalytical applications, we also tested the abilityof SGNs to be isolated and redispersed without the formation ofaggregates. The possibility to modify their surface with thiols hasbeen investigated as well.

2. Materials and methods

2.1. Materials and synthesis methods

All chemicals were purchased from SigmaeAldrich (St. Louis,MO) and used as received. Water was deionized and ultrafiltered bya MilliQ apparatus from Millipore Corporation (Billerica, MA) justbefore use.

SGNs were produced as follows. A solution of tetrachloroauricacid (99% HAuCl4$3H2O, 50 mL, 10 mg mL�1) was diluted in a 50 mLbacker containing 10 mL of freshly prepared ultrapure MilliQ waterunder magnetic stirring at room temperature. Next, 100 mL of11 mg mL�1 hydroquinone solution in water was rapidly injected.The solution turned from yellow (Au3þ) to light blue almostinstantaneously because of the formation of gold nanostars. Parti-cles produced by this method were stable for a few hours. However,this time could be extended up to fewmonths by simple addition of20 mL of sodium citrate (10 mg mL�1) within 20 min from thepreparation. The size of the particles prepared through this methodcan be controlled using different amount of the reactant used.

Fig. 1. Stoichiometry of the reaction between hydroquinone and gold ions.

Moreover, some preliminary results suggest that this protocol issuitable for scale up (Table 1).

Spherical nanoparticles with a diameter of 70 nm were pro-duced by a two-step approach as previously described [19]. Briefly,23 nm gold seeds produced by the standard Frens method based onreduction by sodium citrate [20] were diluted in 10 mL of a solutioncontaining a 25mM concentration of Au3þ ions followed by 20 mL ofa 1% citrate solution and, immediately after, by 30 mL of 30 mMhydroquinone. After addition of hydroquinone, the solutionbecame immediately red because of the formation of 70 nmspherical nanoparticles.

2.2. Characterisation methods

Particles produced following these protocols have been char-acterized by dynamic light scattering (DLS), scanning electronmicroscopy (SEM), Raman and UVeVis spectroscopies.

DLS measurements were performed at 90� with a 90Plus Par-ticle Size Analyzer from Brookhaven Instruments Corporation(Holtsville, NY), working at 15 mW of a solid state laser(l ¼ 661 nm). Nanoparticle Tracking Analysis (NTA) was performedon a Nanosight (Amesbury, UK) apparatus equipped with a solidstate green laser (l ¼ 533 nm) [21]. All samples used for thisanalysis were previously 1:100 diluted with water.

SEM images of gold nanoparticles were obtained by a FEI-NovaNanolab 600I microscope operating at 5 kV, available at the“Nanobiosciences Unit, IHCP, Joint Research CentreeIspra, Italy”.UVeVis spectra were acquired with a Nanodrop 2000c spectrom-eter in standard quartz cuvette.

Raman spectrawere recordedwith an Aramis Horiba Jobin-Yvonmicro-Raman spectrometer equipped with solid state lasers oper-ating at 532 nm, 633 nm and 785 nm and with a DuoScan mappingmode configuration. Raman signal of malachite-green and congo-red were recorded directly in liquid in presence of star-like orspherical particles using the three different, green, red and nearinfrared (NIR) laser light sources. In order to acquire the spectra,200 mL of nanoparticles dispersion and dyes weremixed in a plasticholder and analysed with a 10� objective. The same experimentalconditions were used to compare SERS spectra obtained in pres-ence of nanostars and nanospheres. A fivefold concentrated sus-pension of SGNs was prepared by filtration of the original liquidusing 50 mL of Vivaspin 30 kDa MWCO PES filter centrifuge tubes,centrifuging the particles at 3000 rpm in order to remove excesscitrate solution. Solid substrates have been dipped in a 1 mMsolution of malachite green for 20 s and then dried at room tem-perature. Signals have been recorded using 533 nm, 633 nm and785 nm lasers using a 50� objective. DuoScan averaging mode wasused with a 50� objective to measure areas of different size withdifferent sizes with a 633 nm laser line. XRD patterns were recor-ded using a Thermo Scientific (Thermo ARL X’tra) diffractometerusing a Bragg-Brentano thetaetheta configuration with amaximum excursion ranging from�8� and 180�. X-Ray source is CuKa (l¼ 1.542�A) and the accelerating voltage can be set in the range

Fig. 2. UVeVIS spectra and visual image (insets) of (a) red spherical and (b) blue SGNnanoparticles at a concentration of 5 � 1010 mL�1. (For interpretation of the referencesto colour in this figure legend, the reader is referred to the web version of this article.)

C. Morasso et al. / Materials Chemistry and Physics 143 (2014) 1215e1221 1217

20 O 40 KV. Diffracted rays are collected through a solid state Si:Lidetector cooled by Peltier element.

2.3. Coating of SGN with undecanethiol

SGNs were centrifuged at 4000 rpm for 15 min; then the su-pernatant was carefully removed and the particles redispersed in20 mL acetonitrile. After 5 min of sonication, in order to facilitatethe redispersion, undecanethiol (200 mL) was added to the disper-sion. Nanoparticles were stored in dark at room temperature for36 h, then centrifuged and redispersed in chloroform.

2.4. Preparation of SGN-based flat solid SERS substrate

Silicon slides were washed with “Piranha” solution for 30 min,then rinsed with abundant water. Cleaned slides were immersed ina 0.01% poly-L-lysine solution (Sigma Aldrich) for 1 h, quicklycleaned with water and dried on a hot stage. SGNs were thenadsorbed on the slide by simple overnight immersion in the sus-pension. The slides were dried and used as solid SERS substrate.

3. Results and discussions

3.1. Characterisation

In order to explore the effect of the shape on the enhancementof Raman signals we compared SGNs of 70 nm of diameter withgold nanospheres having a similar size and concentration.

Spherical 70 nm gold nanoparticles showed an intensive redcolour with a plasmonic peak at around 560 nm. As reported also byother authors, SGNs exhibit a plasmonic band strongly shifted to-wards the infrared region of the UVevisible spectra and have acharacteristic blue colour with an absorption peak at around620 nm [22]. Fig. 2 shows the absorption spectra and visualappearance of the spherical and star-shaped nanoparticles,respectively, at a 5 � 1010 mL�1 concentration. Based on the SEMimages (Fig. 3) the SGNs are characterized by the presence of z20spikes of 5e10 nm, rising from a spherical core having a diameter of50 nm. Dimensions and morphology of the spikes suggest that thespecific surface area of these particles do not differ significantlyfrom the surface area of 70 nm diameter spheres. XRD analysis ofSGNs confirms the typical fcc lattice structure of gold nanoparticles(Fig. S8). Measurements of hydrodynamic diameter by DLS resultedin 70 nm average value for both particle types. The similarity in sizedistribution is supported also by the Nanosight (NanoparticlesTracking System) analysis showing a mean diameter of 80 nmwitha mode of 70 nm for both particles suspensions. In order tocompare the SERS enhancement given by the same number ofparticles, the concentration of spherical and star-like gold nano-particles has been also measured by particle tracking analysis andhas been used to prepare dispersions with the same concentration.Dilution of the original samples to the same final particle concen-trations (5 � 1010 mL�1) allowed direct comparison of the resultingsuspensions in terms of plasmonic enhancement properties.

Moreover, as both synthesis methods are based on the use ofhydroquinone and citrate, the particles presented a similar chem-ical surface. The stability of the particle water suspensions in bothcases was provided by the negatively charged layer of adsorbedcitrate molecules.

3.2. SERS experiments in liquid

Star-like or multi-branched gold nanoparticles have beenreported previously as possible highly active substrates for SERS[23]. In order to study the importance of morphological and optical

effects on the enhancement of Raman signal, we tested thedifferent colloids with three laser light sources, namely green(532 nm), red (633 nm) and NIR (785 nm) using two different dyes,malachite green (absorption peak at 617 nm) and congo red(absorption peak at 498 nm).

Fig. 4 summarizes the results comparing Raman spectra of thetwo reporter dyes in the presence of the SGNs (top) or spherical(bottom) particles at each parameter set. While intensity ratiopatterns of the spectra of a certain molecule is shown to stronglydepend also on the exciting wavelength (Fig. 4aef), the patterndepends less on the particle shape (Fig. 4b). In all the cases studied,SGNs exhibited higher enhancement. It should be noticed that thesuspension of spherical gold nanoparticles absorbsmore than twicethe intensity of green light compared with SGNs. In fact theirplasmonic band almost matches the incident laser line and thecontribution to the extinction coefficient given by the interbandelectronic transition for nanoparticles of 70 nm is known to be lessimportant at this wavelength (around 10% of the SPR contribution)[24] and thus a better result could be expected when a green laserline is used for illumination.

However, even if the difference between these particle systemsis less pronounced at this wavelength, experimental data haveshown that SGNs provide a 5e50 times more intense Raman signal,suggesting that advantages provided by morphological factors,especially the enhancement at sharp edges and tips, local “hotspots” generated between the branches of a certain particle, aremore important than consideration on the use of particles perfectlymatching the incident light with their plasmonic absorption.

Fig. 3. Scanning Electron Micrographs of (a) spherical gold and (b) SGN particles. Scalebars represent 100 nm in each image.

Fig. 4. SERS spectra of malachite green 1 mM (a, b, c) and congo red 10 mM (d, e, f) acquirnanoparticles (lower part) using a 533 nm (left); 633 nm (middle) and 785 nm (right) lase

Fig. 5. Hydrodynamic diameter (measured by DLS) of SGNs after the dye addition tothe suspension in case of (a) malachite green and (b) congo red (t < 0 correspond tothe hydrodynamic diameter recorded before the addiction of the dye).

C. Morasso et al. / Materials Chemistry and Physics 143 (2014) 1215e12211218

DLS experiments confirmed that the particles do not aggregatein presence of these dyes and thus the enhancement is given by freefloating particles without the generation of “hot spots” given by thevicinity between more individual metallic nanostructures (Fig. 5).

The effects of the good matching of the excitation wavelength,Raman reporter dye and SGN absorption bands have been studiedusing various concentrations (2.5e50 mM) of malachite green,633 nm (HeeNe) laser light, and a fivefold serial dilution of SGNsuspensions. The concentration of the particle suspension gener-ated following the recipe described above has been determined byNanoparticles Tracking System to be about 5 � 1010 particles mL�1.In our study this concentration has been demonstrated to be theoptimal concentration for SERS studies using the nanoparticlesuspension and a 10� objective, as attempts with higher concen-trations did not result in higher enhancement. As shown in Fig. 6,the intensity of a specific peak (1170 cm�1, due to the in-plane

ed under the same conditions using SGNs (upper part of each section) and sphericalr light sources.

Fig. 6. Dependence of the SERS signal of malachite green (measured through thecalculation of the area of the peak at 1170 cm�1) on the concentration of SGNs. Whenthe concentration of nanoparticles is higher than 5 � 1010 used in this study, only asmall increment is observed independently from the concentration of malachite green(100 mM -; 50 mM C; 10 mM :; 5 mM ; and 2.5 mM ◄).

Fig. 8. SEM picture of SGNs adsorbed on a polylysine-coated silicon slide. Scalebarrepresents 5 mm.

C. Morasso et al. / Materials Chemistry and Physics 143 (2014) 1215e1221 1219

mode of CeH bending) in the malachite green spectrum exhibitssaturation not only with the increasing dye concentration, butalso with an increasing concentration of nanoparticles. Sucha dependence on concentration looks to be linear below5 � 1010 particles mL�1, while a negligible further signal increasewas observed at higher particle concentrations independently fromthe concentration of malachite green used. This quite surprisingresult can be partially explained with a re-absorbance of emittedlight by other SGNs as their absorption peak matches the wave-length of the laser used for the study.

Our data show that the excellent match between absorptionbands of the malachite green as reporter dye and SGNs of 70 nmresults in great signal enhancement at 633 nm laser wavelength(Fig. 4b), which is limited only at high particle concentrations bythe competition of particles and die molecules for the exciting and/or reflected light.

In order to demonstrate the effectiveness of SGNs in detectinganalytes different from standard Raman dyes, we tested our systemon apomorphine, a well-known drug used for the management ofParkinson disease. Apomorphine has been diluted in a suspensionof SGNs to a final concentration of 5 mM and Raman spectra hasbeen collected using a HeeNe laser line operating at 633 nm. Asshown in Fig. 7 it has been possible to obtain a clear spectra char-acterized by presence of bands relative to the oxidation of

Fig. 7. SERS spectra of apomorphine 5 mM obtained by the use of s

apomorphine on the surface of the metal structure as reported in aprevious study [25].

3.3. Functionalisation of SGNs

SGNs coated by undecanethiol using the method describedabove were successfully transferred to chloroform where theyproved to be perfectly dispersible and stable for months demon-strating the exchange of the solubility properties given by theconjugationwith hydrophobic molecules. The hydrophobic surface,and the resulting “solubility” in organic media opens the possibilityfor further modifications (polymer coating, coreeshell particlesynthesis, etc.) running in nonpolar organic solvents. (A picture ofSGNs dispersed in chloroform is available in the Supplementaryinformation.)

3.4. SGN-based flat solid substrates

Star-like gold nanoparticles can be easily applied for the prep-aration of solid substrates. The surface prepared by simple over-night immersion of a poly-lysine treated silicon slide in the particlesuspension at 4 �C appear to be an active and easy-to-handle SERSsubstrate and able to enhance the signal of a 1 mMdrop of malachitegreen with all the different laser lines tested in this study (SeeSupplementary information).

Polylysine is a highly positively charged polymer that adsorbson the surface of silicon slides effectively charging them andworking as an electrostatic glue attracting and firmly binding the

tar-like gold nanoparticles and assignment of the main peaks.

C. Morasso et al. / Materials Chemistry and Physics 143 (2014) 1215e12211220

negatively charged gold nanoparticles on the surface. Assembly ofgold nanoparticles on positively charged surface have been previ-ously reported in literature as good solid substrate for SERSbecause of the presence of “hot spots” at the junction betweenadsorbed particles [26]. However, this protocol suffers from a lackof homogeneity of the surface. As clearly visible from SEM pictures(Fig. 8), the surface is characterized by the presence of areas withlarge aggregates of particles and areas with a more uniformcoating. Thus the intensity of Raman signal recorded acquiredthought the standard point by point acquisition at different posi-tions of the surface is highly variable. We solved this problemusing the DuoScan averaging mode [27]. Using this configurationof the Raman microscope the laser spot is continuously scannedacross a user-defined square surface. Since the energy of the laserbeam is spread out, instead of being concentrated in a small laserpoint, an average measurement of the signal coming from theanalysed surface is obtained effectively spatially averaging theresults and reducing the differences of intensity given by a localnanometric inhomogeneity of the surface. This technique wasoriginally development for a more efficient Raman imaging anal-ysis of cells and biological tissues characterized by the presence ofvery small features that would not have been observed other waysor for the analysis of photosensitive materials [28], but hasdemonstrated useful even to improve the reliability of SERSsignals.

Using this configuration of the instrument, we collect SERSsignal of malachite green from six different points of the surfacescanning square areas. The measurement was repeated usingsquared area of different dimensions in order to find the smallestscanned areawith still good relative error (standard deviation/peakintensity) value. Our data suggest that, once that squares with 5 mmside are scanned, basically the surface becomes homogeneous(standard deviation/peak intensity ratio of less than 0.05) for the1174 cm�1 peak (see Supplementary information) independentlyfrom the position on the surface. As this data did not fit with thescale at which the surface looks homogeneous at SEM pictures, weconfirmed our results measuring on purpose at areas where thesurface was coated with large aggregates, with dispersed particlesor where an intermediate situation was present. On each of thesesituations, we repeated 10 different measures at discrete locations,recording signals with similar intensity (with a standard deviationof about 6%) demonstrating the advancement given by DuoScan inSERS measurement (Fig. 9).

Fig. 9. SERS spectra of malachite green acquired on the surface of an SGN-coatedsilicon slide by DuoScan acquisition mode on area with no aggregates (a), aggregates(b) and intermediate (c). Each graph is given by the superposition of 10 differentmeasures on different position of the substrate. All spectra have been obtained using a633 nm laser line.

4. Conclusions

In this work, we present a new protocol for the preparation ofstar-like gold nanoparticles and we prove that they are moreeffective in enhancing Raman signal of known reporter dyes whencompared with spherical nanoparticles with similar dimensionsand concentration. This protocol is a facile one step protocol basedon commercially available reactants that acts in seconds. We alsodemonstrate that these nanoparticles can be easily functionalizedusing the classical thiol chemistry, resulting in hydrophobic SGNsthat are transferable in organic solvents for further surface modi-fications. Our particles proved as well to be excellent candidates forpreparation of solid substrates by simple self-assembling procedureon a positively charged surface. Moreover, we demonstrate howDuoScan averaging mode can be successfully used in order todrastically decrease the variability of the intensity of SERS signalscollected by solid substrates that until now has been one of themain limitation to the diffusion of this technique in bio-analyticalchemistry. Based on these data and on their good physicochem-ical characteristics, we expect that SGNs prepared in this way couldbe used for further development of highly specific and sensitiveSERS-based assays for the detection of various biomarkers.

Acknowledgements

Funding for this research was provided by Fondazione Cariplo(International Recruitment Call 2011 Project: An innovative,nanostructured biosensor for early diagnosis and minimal residualdisease assessment of cancer, using Surface Enhanced RamanSpectroscopy), and was also supported by the Italian Ministry ofHealth (Conto Capitale 2010: Realizzazione e validazione di unacore facility di biofotonica clinica per diagnosi precoce e monitor-aggio di minimal residual disease in patologie tumorali). Theauthors would like to thank César Pasqual García for taking SEMimages of nanoparticles at the Joint Research Centre, IHCP, Nano-biosciences Unit, Ispra, Italy and Prof. Andrea Zappettini for theXRD characterization of SGNs at CNR-IMEM, Parma, Italy.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.matchemphys.2013.11.024.

References

[1] K.C. Bantz, A.F. Meyer, N.J. Wittenberg, H. Im, O. Kurtulus, S.H. Lee,N.C. Lindquist, S.-H. Oh, C. Haynes, Phys. Chem. Chem. Phys. 13 (2011) 11551e11567.

[2] F. De Angelis, F. Gentile, F. Mecarini, G. Das, M. Moretti, P. Candeloro,M.L. Coluccio, G. Cojoc, A. Accardo, C. Liberale, R.P. Zaccaria, G. Perozziello,L. Tirinato, A. Toma, G. Cuda, R. Cingolani, E. Di Fabrizio, Nat. Photon. 5 (2012)682e687.

[3] R. Stokes, A. Macaskill, J.A. Dougan, P.H. Hargreaves, H.M. Stanford,W.E. Smith, K. Faulds, D. Graham, Chem. Commun. 27 (2007) 2811e2813.

[4] A.M. Schwartzberg, C.D. Grant, A. Wolcott, C.A. Talley, T.R. Huser,R. Bogomolni, J.Z. Zhang, J. Phys. Chem. B 108 (2004) 19191e19197.

[5] K.W. Kho, Z.X. Shen, H.C. Zeng, K.C. Soo, M. Olivo, Anal. Chem. 77 (2005)7462e7471.

[6] L.R. Allain, T. Vo-Dinh, Anal. Chim. Acta 469 (2002) 149e154.[7] K.K. Strelau, A. Brinker, C. Schnee, K. Weber, R. Moller, J. Popp, J. Raman

Spectrosc. 42 (2011) 243e250.[8] L. Rodriguez-Lorenzo, R.A. Alvarez-Puebla, F. Garcia de Abajo, L.M. Liz-Marzan,

J. Phys. Chem. C 114 (16) (2010) 7336e7340.[9] Q. Su, X. Ma, J. Dong, C. Jiang, W. Qian, ACS Appl. Mater. Interfaces 3 (2011)

1873e1879.[10] M. Pradhan, J. Chowdhury, S. Sarkar, A. Kumar Sinha, T. Pal, J. Phys. Chem. C

116 (2012) 24301e24313.[11] E.N. Esenturk, A.R.H. Walker, J. Raman Spectrosc. 40 (2009) 86e89.[12] G. Maiorano, L. Rizzello, M.A. Malvindi, S.S. Shankar, L. Martiradonna,

A. Falqui, R. Cingolani, P. Pompa, Nanoscale 3 (2011) 2227e2232.[13] T.K. Sau, C.J. Murphy, J. Am. Chem. Soc. 126 (2004) 8648e8649.

C. Morasso et al. / Materials Chemistry and Physics 143 (2014) 1215e1221 1221

[14] S. Boca, D. Rugina, A. Pintea, L. Barbu-Tudoran, S. Astilean, Nanotechnology 22(2011) 055702.

[15] H. Peng, H. Dong-xue, N. Li, L. Hai Bo, Chem. Res. Chin. Univ. 22 (2006) 493e499.[16] J. Li, J. Wu, X. Zhang, Y. Liu, D. Zhou, H. Sun, H. Zhang, B. Yang, J. Phys. Chem. C

115 (2011) 3630e3637.[17] Sirajuddina, A. Mechler, A. Torriero, A. Nafady, C.-Y. Lee, A. Bond, A. Mullane,

S. Bhargava, Colloid Surf. A 370 (2010) 35e41.[18] M. Langille, M. Personick, J. Zhang, C.A. Mirkin, J. Am. Chem. Soc. 134 (2012)

14542e14554.[19] S.D. Perrault, W.C.W. Chan, J. Am. Chem. Soc. 131 (2009) 17042e17043.[20] G. Frens, Nature 241 (1973) 20e22.[21] K.E. Sapsford, K.M. Tyner, B.J. Dair, J.R. Deschamps, I.L. Mendintz, Anal. Chem.

83 (2011) 4453e4488.

[22] G.H. Jeong, Y.W. Lee,M. Kim, S.W.Han, J. Colloid Interface Sci. 329 (2009) 97e102.[23] M. Schütz, D. Steinigeweg, M. Salehi, K. Kömpe, S. Schlücker, Chem. Commun.

47 (2011) 4216e4218.[24] M. Losurdo, M. Giangregorio, G. Bianco, A. Suvorova, C. Kong, S. Rubanov,

P. Capezzuto, J. Humlicek, G. Bruno, Phys. Rev. B 82 (2010) 155451.[25] A. Lucotti, M. Tommasini, M. Casella, A. Morganti, F. Gramatica, G. Zerbi, Vib.

Spectrosc. 62 (2012) 286e291.[26] R. Tanra, R.J.C. Brown, M.J.T. Milton, D.D. Gohin, Appl. Spectrosc. 62 (2008)

992e1000.[27] G. Falgayrac, B. Cortet, O. Devos, J. Barbillat, V. Pansini, A. Cotten, G. Pasquier,

H. Migaud, G. Penel, Anal. Chem. 84 (2012) 9116e9123.[28] Horiba Technical Note: RA-TN04, http://www.horiba.com.