Direct experimental observation of salt induced aspect ratio tunable PFPT silver-nanowire formation: SERS-based ppt level Hg²⁺ sensing from ground water

Abstract

A plausible explanation based on real-time direct experimental observation has been put forward to explain the formation of common salt (NaF, NaCl, NaBr, and NaI) induced aspect ratio tunable Pentagonal Faceted Pyramidal Tipped (PFPT) silver-nanowires by considering the solubility product of the in situ generated silver halides, binding affinity of the added halides at the {111} facets on the growing front, and the free energy associated with the crystallographic planes (obtained from HRTEM study) as the driving forces. Aspect ratio dependent Raman activity was verified by using 4-mercaptobenzoic acid (4-MBA) as a Raman tag and due to the preferential binding of Hg²⁺ on the pyramidal tip of the silver-nanowire [i.e., {111} facet], as all the {100} facets are blocked by insoluble silver halides, controlled leaching of the nanowires offers the best-possible Raman platform for the ultra sensitive detection of Hg²⁺ contamination (up to 50 ppt) from ground water.

---

Introduction

One dimensional metallic nano-structure with variable aspect ratios¹ attract the interest of the scientific community because of their potential applications in nanoscale electronic device fabrication,² nano-optoelectronics,³ opto-mechanics and non-linear optics,⁴ light-harvesting,⁵ integrated circuits,⁶ and in ultra-sensitive molecular sensing devices.⁷ To find their multidimensional applications, we need to know the exact mechanism of their real-time formation along with a simple but controlled synthetic methodology to tune their physicochemical and mechanical properties. The literature is rich with methods to synthesize silver nanowires^8,4c,7a and a number of chemical approaches have been actively explored to process silver into 1D nanostructures. For example, silver nanowires have been synthesized by reducing AgNO₃ with a developer in the presence of AgBr nanocrystallites,⁹ or by arc discharge between two silver electrodes immersed in an aqueous NaNO₃ solution.¹⁰ Silver nanorods have also been produced by irradiating an aqueous AgNO₃ solution with ultraviolet light in the presence of poly-(vinyl alcohol).¹¹ The final products of all these methods, however, encounter problems such as low yields, irregular morphologies, polycrystallinity, and low aspect ratios. In contrast, template-directed synthesis offers a better-controlled route to 1D silver nano-structure. A variety of templates have been successfully demonstrated by using this process, and typical examples include channels in macroporous membranes,¹² mesoporous materials,¹³ or carbon nanotubes,¹⁴ block copolymers¹⁵ DNA chains,¹⁶ rod-shaped micelles,¹⁷ arrays of calix[4]hydroquinone nanotubes¹⁸ and steps or edges on solid substrates.¹⁹ Previously, Xia et al.^8c reported a uniform silver nanowire synthesis by a seeded (Pt seed) growth technique and later^8b modified this method to synthesize nanowires without using any seeds. In both of these methods, the authors have tried to study the growth kinetics and explored the growth mechanism, but they have not tried to vary the aspect ratio by controlling the epitaxial unidirectional growth and hence not achieved any aspect ratio tunability. Synthesized silver nanowires by Xia et al.^8c,d show a mean diameter of ∼38 nm with 6 μm in length to give an aspect ratio of ∼150. In a more recent paper, Gongxuan Lu and coworkers^8a have reported a NaCl induced silver nanowire synthesis without providing the actual dimensions. In this communication we are reporting for the first time a simplified methodology for controlled aspect ratio tunable synthesis of silver nanowires by varying just the concentration of common salt. Moreover we used ethylene glycol as the reducing agent compared to their higher glycols, which directed us to find an alternate route to synthesize them through a cost effective way as well as to make them more bio friendly. Besides that, we use different salt samples (NaF, NaCl, NaBr, and NaI) for efficiently blocking different crystal facets and depending on their solubility product of the in situ generated sparingly soluble silver halides, we can vary their aspect ratio by allowing them to grow in a single direction. The generated nanowires are extremely stable for months and show different extents of surface properties depending on their aspect ratio. Xia et al.^8c have presented a plausible mechanism of silver nanowire formation by considering homogeneous and heterogeneous nucleation of the seeds based on TEM and SEM characterization. However, real-time observations lack the explanation of how the in situ generated seeds attach to each other energetically to form a bigger nucleation centre to queue up the seeds and how the in situ generated silver atoms cover the space between nucleation centres in a streamlined arrangement (a string with several knots) to transfigure them into larger nanowires. This report explains in detail the real-time direct observation (based on HRTEM) of nanowire growth starting from the formation of the seed, inter seed interaction for the generation of the bigger nucleation centre, formation of the pentagonal nucleation site, the streamlined arrangement of the silver atoms seed generated in situ from the pentagonal nucleation site, and finally the transformation of this arrangement into bigger nanowires. We used a combined variable sonication cum fractional centrifugation technique to separate the unsymmetrical silver spheroids generated in situ from silver nanowires. Besides their aspect ratio varied synthesis, we can also effectively leach them by adopting galvanic replacement of silver with gold using HAuCl₄ to make controlled hollow bimetallic nanotubes which is a completely different result obtained by Lu et al.^8a Once we synthesized and washed the nanowire, due to high specificity of mercury (compared to other metals)²⁰ we have successfully used them for ultrasensitive Hg(II) detection from ground water. Mercury has adverse effect on living systems as it can damage DNA, cause malfunctioning of the central nervous system, kidneys, lungs, endocrine systems etc. leading to fatal diseases. The types and extent of disruption varies with the amount of mercury present in a zero oxidation state, as ions or as organomercuric compounds. The Environmental Protection Agency (EPA), United States, has prescribed a maximum uptake limit of 2 ppb for inorganic mercury in drinking water. Based on the Localized Surface Plasmon Resonance (LSPR) effect, mercuric ion sensing in aqueous medium with silver nanowires produced in polyol method has been investigated with a detection limit in the mM (∼1000 ppm) range.²¹ To the best of our knowledge, this is the only report to use silver nanowires successfully for highly specific (easy to distinguish between Au³⁺ and Hg²⁺, in the presence of other metal ions) and ultrasensitive (up to 50 ppt-level detection) mercury sensing by using a SERS-based technique.

Experimental section

Chemicals and materials

Chemicals including ethylene glycol (EG, 99%), poly(vinyl pyrrolidone) (PVP, M_W ≈ 40 000, 98%), sodium fluoride (NaF, 99%), Sodium chloride (NaCl, 99%), sodium bromide (NaBr, 99%), sodium iodide (NaI, 99%), silver nitrate (AgNO₃, BioXtra, >99% (titration)), Gold(III) chloride trihydrate (HAuCl₄·3H₂O, ≥99.9% trace metal basis), sodium sulfide (Na₂S, 99%), 4-mercaptobenzoic acid (HS–C₆H₄–COOH, 99%), DL-dithiothreitol (HS–CH₂–CH(OH)–CH(OH)–CH₂–SH, 99%), and all the metal salts (chloride) including As₂O₃ for selective sensing experiment were purchased from Sigma-Aldrich and used for the synthesis and related experiments without any further purification. All the experimental works have been performed using Milli-Q water.

Preparation of silver nanocubes

Silver nanocubes are synthesized by adopting a standard protocol reported before²² by using AgNO₃ as the building block, PVP as the surfactant, Na₂S as the facet blocking agent, and EG as the solvent as well as the reducing agent. After synthesis, the obtained aliquot was centrifuged at 6000 rpm for 1 hour to remove un-reacted reagents from the nanocubes.

Preparation of silver nanowires

The silver nanowires are synthesized in a three-neck flask. In the first step, 54 mL of EG was heated at 165 °C (below the boiling point of EG) for 1 h with continuous stirring to remove any dissolved oxygen. Then 13.5 mL of 0.18 M PVP in EG solution (considering N-vinylpyrrolidone molecular weight as 111.14 gm), 1 mL of 2.7 mM NaCl in EG solution, and 4.5 mL of 0.28 M AgNO₃ in EG solution were added sequentially. Heating was continued for another 25 min. Colour of the solution changes from colourless to yellow initially and then it changes to orange, red, and then finally to muddy off-white when the temperature again reached approximately 165 °C. The same set of synthesis was repeated with 0 mL, 0.125 mL, 0.25 mL, 0.5 mL, 1 mL, 2 mL, 3 mL, 4 mL, 8 mL, 16 mL, 32 mL, 50 mL, and 64 mL NaCl solutions (2.7 mM) to synthesize aspect ratio varied silver nanowires. It has been observed that below 1 mL NaCl there is no production of silver nanowires and only spherical nanoparticles are formed. Our synthesized silver nanowires show moderate monodispersity (≥50%) and the difficulty in producing single sized nanowires is a well described in the literature. To quantify the aspect ratio of the synthesized nanowires by varying the concentration of added salt we have considered 10 or more image frames with each frame having more than 100 nanoparticles and measured not only the number averaged length distribution (X axis: length of silver nanowire and Y axis: number of nanowires at that particular length) but also estimated their average length (ensemble average) to report their mean length or aspect ratio in Fig. 1 and 2. Quantification of the aspect ratios of the synthesized nanowires has been shown in Fig. SI1.† By keeping the salt concentration the same, we have synthesized silver nanowires using different salts (NaF, NaBr, and NaI) to explore the effect of the halogen atom in the salt. Details of the halogen atom effect and salt concentration have been discussed in detail in the results and discussion section.

Fig. 1 TEM-images of silver nanowires, synthesized using different amounts of 2.7 mM NaCl. From left to right for first, second and third row we have added 1 mL, 2 mL, 3 mL, 4 mL, 8 mL, 16 mL, 32 mL, 50 mL, and 64 mL of 2.7 mM NaCl. Along with the measurement scale, the approximate length of the obtained silver nanowires has been mentioned at the top of each image. The aspect ratios of the obtained nanowires change from 10 to 500.

Fig. 2 SEM-images of silver nanowires, synthesized using different amounts of 2.7 mM NaCl. From left to right for first, second and third row we have added 1 mL, 2 mL, 3 mL, 4 mL, 8 mL, 16 mL, 32 mL, 50 mL, and 64 mL of 2.7 mM NaCl. Along with the length scale marker at the lower right corner, the approximate length of a randomly selected silver nanowire from each of the micrographs has also been mentioned at the upper right corner of the SEM images. The aspect ratios of the obtained nanowires change from 10 to 500.

Preparation of silver tubular nano-cages

Once we synthesized silver nanowires with varied aspect ratios, we made silver tubular nanocages using the galvanic replacement reaction. We dissolved 1 mL of concentrated Ag nanowire in 50 mL nano pure water and then increased the solution temperature up to the boiling point. After that we added different volumes of 10⁻³ M HAuCl₄·3H₂O solution to prepare the silver tubular nanocages which has been discussed in detail in the results and discussion section. Generated silver tubular nanocages were purified by alternating the centrifugation three times (at 2500 rpm for 1 hour) followed by washing in water medium.

Instrumentation

1. UV-vis. The UV-visible and vis-NIR extinction spectra were taken at room temperature using Jasco V-650 and Perkin Elmer Alpha 750 spectrophotometers respectively in a quartz cuvette with a 1 cm optical path. Since the reaction occurs momentarily and it is difficult to freeze the reaction without adopting a chemical route, we have taken the absorption spectra for each sample after the completion of the reaction. Details about the UV-vis spectra will be discussed in the following sections.

2. TEM & SEM. Both for normal and high resolution transmission electron microscopic (HRTEM) measurements we used a FEI, Tecnai G2 F30, S-Twin microscope operating at 120 and 300 kV respectively. The compositional analysis was performed by energy dispersive X-ray spectroscopy (EDS, EDAX Instruments) attached to the Tecnai G2 F30. The ZEISS SUPRA 40 scanning electron microscope (SEM) fitted with a hot Schottky field-emission gun (FEG) was used to obtain secondary electron (SE) images of the nanowires. We have used simple but modified techniques for clean monolayer sample preparation both for TEM and SEM. For the TEM measurements we used a 300 mesh copper formvar/carbon grid. We used a dip-and-dry technique to make the TEM samples. In this sample preparation technique we dip a TEM grid in the concentrated nanomaterial sample solution using tweezers and the hydrophobic carbon coating allows a monolayer of the sample to stick to the copper mesh which is dried on soft tissue paper. After completely drying we use this grid for the TEM measurement. For the SEM measurement we used a piece of mirror polished silicon wafer as the sample support. We injected about 20 μL of the concentrated nanomaterial sample on the tilted silicon wafer and the hydrophobic nature of the wafer allows only a single layer of the sample to stick on the surface which quickly dries to allow immediate SEM measurement. From the TEM and SEM measurements shown in Fig. 1 and 2 respectively, it is evident that as we increase the concentration of NaCl, particles with bigger aspect ratio are developing.

3. SERS. Surface enhanced Raman scattering (SERS) experiments were done using a homemade Raman setup. We used a continuous wavelength diode-pumped solid state (DPSS) laser from Laser Glow Technology, Canada (LRS-0532-PFM-00300-03) operating at 532 nm as an excitation light source (at a fixed excitation energy of 3 mW by using a neutral density (ND) filter throughout the experiment). For efficient focusing and filtering we used an InPhotonics made 532 nm Raman fibre optics probe with a spectral range of 200–3900 cm⁻¹ (Stokes) for sample excitation and data collection. The Raman probe consists of two single fibres (105 μm excitation fibre, 200 μm collection fibre) with filtering and steering micro-optics, N.A. 0.22. The excitation fibre was connected to a fibre port to align the laser whereas the collection fibre was connected to a spectrometer. A miniaturized QE65000 scientific-grade spectrometer from Ocean Optics has been used as the Raman detector with a spectral response range of 220–3600 cm⁻¹. The Raman spectrometer is equipped with TE cooled 2048 pixel CCD and interfaced to a computer through a USB port. At the end, the Raman spectra were collected using Ocean Optics data acquisition SpectraSuite spectroscopy software. Each experiment has been performed 5–6 times and the average values are reported in this manuscript.

For the SERS measurement of the extent of mercury(II) present in the environmental samples, each Raman active analyte was made by mixing 5 μL of the concentrated silver nanowires with 5 μL pure water or 5 μL of different concentrations of Hg(II) solution in a small micro centrifuge tube cap and then finally diluted with 190 μL of 4-mercaptobenzoic acid (4-MBA, 10⁻³ M) to make a total 200 μL aliquot for the control and mercury sensing measurements by exposing the sample at 532 nm laser light and the scattering signal was collected with a 30 second acquisition time and 5 scan average. Laser power was kept at 3 mW for the entire experiment.

Results and discussion

The central theme of this report is to establish a plausible mechanism of silver nanowire formation based on the real-time direct observation of each and every intermediate step either through colorimetric or electron imaging. This is the first experimental report where we could clearly see how the bigger nucleation centres, generated seeds in situ, and that the atomic silvers queue up in a streamlined arrangement (a string with several knots) to transfigure this assembly into bigger nanowires. Though there are a few reports on real-time measurements to explain the mechanism of nanowire formation,^8b,c,f due to the lack of finer intermediates in the imaging studies, the picture of step-by-step nanowire growth was unclear and in that respect our study is unique in providing the complete picture of nanowire growth. Along with the microscopic full view of the nanowire growth mechanism, a new synthetic strategy to fabricate silver nanowires with dimensional as well as plasmonic tunability has been explained in this report to find their enormous applications in tunable Plasmon spectroscopy. A plausible explanation based on the real-time direct experimental observation has been put forward to explain the formation of salt induced nanowires by considering the solubility product of the silver halides generated in situ, the binding affinity of the added halides at the {111} facets on the growing front, and the free energy associated with the crystallographic planes (obtained from the HRTEM study) as the driving force. By virtue of their size tunability, our ultimate goal is to attune their physicochemical and mechanical properties for ultrasensitive chemical recognition. The originality of this report lies in the ability to tune the properties through a cost effective and green synthetic route and a supporting real-time direct experimental (HRTEM) observation-based growth mechanism has been illustrated in great detail. Moreover, an easy technique has been prescribed to remove the silver spheroids generated in situ from the high aspect ratio nanowires to make highly mono-dispersed silver materials. As a whole; synthesis, purification, and preservation of high aspect ratio silver materials is such an easy venture that we can easily adopt our technique for industrial levels of nanowire production. Moreover by adopting simple chemical treatment we can make improved nanowire networking for selective molecular traps or to transform them into controlled hollow bimetallic nanotubes. The following sections explain the obtained results in detail and raise a scientific discussion in support of the obtained result.

Role of common salt for the generation of silver nanowires

Though there are several reports^8b,c,f which explain the mechanism of salt-induced silver nanowire (SNW) formation, this is the sole report which describes the complete visualization of their microscopic growth and explains that the aspect ratio dependent SNW formation depends on the salt concentration as well as the halogen substituent in the salt. The first reported production of SNW was achieved via reducing AgNO₃ with ethylene glycol (EG) in the presence of platinum seeds and polyvinylpyrrolidone (PVP)^8c,d where EG serves as reducing agent while PVP serves as the surfactant. Later ​Sławiński et al.²³ reported another seeded growth technique using gold nanoparticles as seeds and cetyltrimethylammonium bromide (CTAB) as the surfactant, silver nitrate as the precursor, and ascorbic acid as the reducing agent. Silver nanowires can also be prepared by a self-seeding process without external crystal seeds, in which the initially generated silver particles serve as seeds for nanowire growth.²⁴ Scientists have reported several other self-seeded aqueous-phase methods to prepare SNW where EG has been replaced by N,N-dimethylformamide (DMF), potassium tartrate (K₂C₄H₄O₆), anionic surfactant dodecyl benzene sulfonic acid sodium (DBS), and L-cysteine.²⁵ Gongxuan Lu^8a and coworkers have synthesized rectangular-SNW by using 1,2-propylene glycol as a reducing agent followed by the same protocol as we have adopted in this manuscript. So, it is clear from the above reported methods that we can easily control their morphology^4c by adjusting several parameters such as seeding condition, surfactant, surfactant to precursor ratio, reaction temperature, reaction time, etc. Silver nanowires synthesized in this report have an average diameter of ∼50 nm (average % of SNWs with diameter ∼50 nm is >90%) associated with variable length of 500 nm to >30 μm and a pentagonal cross section (as shown in Fig. 3), with {100} facets on the side surface and {111} facets on the growing front as reported earlier.²⁶ In this report our aim is to investigate how these active crystal planes are affected by the increasing amount of added salt to produce aspect ratio varied SNWs and discuss how the achieved aspect ratio affects their sensing ability. In the case of silver nanocube synthesis, added Na₂S could act as a facet blocking agent in two different way: (i) spontaneous dissolution of Na₂S can generate sulfide anions (S²⁻) or atomic sulfur (S, by in situ oxidation) which has the ability to block all of the facets equally to form nanocubes, or (ii) it could react with AgNO₃ to generate highly insoluble Ag₂S with the solubility product, K_SP = 6.3 × 10⁻⁵⁰ at 20 °C to block selective facets. Obtained results indicate that the first path is much more favorable over the other and hence facet selective blocking is not possible with Na₂S. To avoid this situation and to allow facet selective blocking for the production of one dimensional silver nanostructures different halide salts have been used. The addition of NaCl to the AgNO₃ solution leads to the immediate formation of AgCl, which can play an important role in the shape control mechanism. It has been suggested that underpotential deposition of silver halides occurs on the crystal facets of the in situ generated nano-seed, leading to symmetry-breaking nanoparticle formation.²⁷

Fig. 3 High and low magnification SEM images of the silver nanowires taken at tilted (A and B) and non-tilted (C and D) configuration of the sample mounting stage. (E–G) are tilted TEM images. Comparison between tilted and non-tilted images clearly shows their pentagonally twinned face and pyramidal tipped structure.

It is clear from Fig. 1 and 2 that for a particular common salt (viz. NaCl), the aspect ratio of the generated silver nanowires gradually increases as we increase the concentration of the added salt with respect to a fixed concentration of AgNO₃. Since the lengths of the produced nanowires vary between 500 nm and 60 μm, recording the longitudinal plasmon mode is not possible using our NIR spectrometer as the plasmons are out of range and it only gives information about the transverse plasmon mode which has been shown in Fig. SI2.† To understand the nanowire growth phenomenon we studied the real-time HRTEM of the growth process which gives the microscopic mechanistic pathway of the nanowire formation. As is clearly visible from Fig. 4, as the reaction proceeds the mother solution changes its colour from colourless to yellow initially and then changes to orange, red, brown, gray, muddy off-white, and then finally to a silky off-white colour when the temperature again reaches approximately 165 °C. We have taken a ∼100 μL aliquot from this mother reaction mixture when it changes colour and immersed it into ice water to freeze the reaction and preserve the formed intermediate structure. In this technique we have collected several intermediate samples (100 μL each) over the reaction and made TEM grids from each sample to study their real-time growth process. The NaCl-induced silver nanowire growth process is shown in Fig. 4 which explains their colour change as well as the in situ growth as the reaction proceeds. It is clear from this real-time TEM that the appearance of the yellow color in the initial stage indicates the formation of seeds in situ. It is clear from the HRTEM study (Fig. 5) that the in situ generated seeds are about 5 nm in diameter and pentagonally twinned single crystals (step 1, Fig. 4) with all the active crystal facets growing along {111} facial planes. Immediately after the formation, seeds attach to each other along {111} facial planes both axially as well as sideways (with respect to the final SNW axis) due to the lack of crystal selectivity to generate bigger nucleation sites which has been shown in steps 2 & 3 respectively. These intermediate steps can be termed attachment steps, as mentioned in Fig. 4, and initiate the formation of pentagonal-faced and pyramidal tipped rods by epitaxial deposition of silver atoms over the {111} facial planes. Once a clear pentagonal nucleation site develops, the in situ generated seeds queue up and silver atoms cover the space in between them so that a streamline arrangement results, as shown in steps 4 & 5. A structure resembling a “string with several knots” is actually evident in the TEM pictures and those later transfigure into bigger nanowires (final step). The formation of a nanowire with greater surface area at the expense of these small nanospheres occurs following the principle of Ostwald ripening process.²⁸ Proof of this ligation process among seeds along with close views of the multiply twinned seeds themselves under the process of growth along {111} axis is also visible from the obtained HRTEM (Fig. 5) image.

Fig. 4 Real-time growth steps as anisotropic silver nanomaterials grow from seeds to high aspect ratio nanowires. Attached TEM images explain their nanoscopic material development, attachments, and subsequent streamline arrangement to finally generate nanowires with high aspect ratios.

Fig. 5 HRTEM images of the pentagonally twinned seeds along with the finally developed silver nanowires through a “string with several knots” mechanism. It is evident from the HRTEM study that all the five faces of the pentagonally twinned seeds are {111} facets. The growing front of the nanowire also has low energy {111} facets and the side face has {100} facets.

When the attachment and jacketing of seeds is completed, the newly generated {100} surface tends to smooth and be continuous. According to the literature, the free energies associated with the crystallographic planes of an fcc metal increases in the order: γ(111) < γ(100) < γ(110).²⁹ Due to the highly energetic nature of the {100} facets, produced silver halides in the medium have a higher tendency to deposit and ultimately block the {100} facets, increasing the stability of the generated nanorods. Blocking of {100} facets indirectly influences the axial growth by continuous deposition of atomic silver or in situ generated silver nanoparticles on the {111} facet in a streamline arrangement.^26,30 Diffusion of atomic silver or in situ generated silver nanoparticles is facilitated by the elevated temperature of the reaction medium. Nanowires of several micron lengths can be synthesized by this method. The silver halides have a definite role in the termination process of the {100} facet growth leading to a finite wire shape. With increasing concentrations of silver halide (when we add greater amount of sodium halide in the reaction mixture), the relative deposition of silver halides on the {100} facets and streamlining of atomic silver on the {111} facets results in the generation of silver nanowires with a very high aspect ratio. However, at very high silver halide concentrations, the selectivity of the generated silver halides on the {100} facet over the {111} reduces due to the lack of available atomic silver for axial deposition and eventually gives rise to spherical particles and the situation develops in the same way as when there is no blocking agent.

We attempted to illustrate the effect of the solubility products of silver halides on the aspect ratio of the synthesized nanowires. To do so, we have taken fluoride, chloride, bromide and iodide salts of sodium to produce AgF, AgCl, AgBr and AgI respectively in the course of the reaction. Ethylene glycol is freely miscible in water and it can be assumed that the trends in solubility in the latter will be similarly effective as that in the former. AgF is highly soluble in water and as a result it remains dissolved in the reaction medium and does not block any facet and results the formation of spherical particles only. The solubility product of the other silver halides in water at 25 °C follows the order:³¹

K^AgCl_SP (1.77 × 10⁻¹⁰) > K^AgBr_SP (5.35 × 10⁻¹³) > K^AgI_SP (8.52 × 10⁻¹⁷)

As the temperature rises, the solubility increases following the equation below:^8b

where, ΔH₀ is the standard value for the enthalpy of dissolution (enthalpy of precipitation has same value but opposite sign) and R is the gas constant = 0.008314 kJ mol⁻¹ K⁻¹. ΔH₀ for AgCl, AgBr, and AgI is 65.72, 84.83, and 111.13 kJ mol⁻¹, respectively.³² The increasing order of the solubility product derived from the above equation at 165 °C is

K^AgCl_SP@165 (8.5 × 10⁻⁷) > K^AgBr_SP@165 (3.0 × 10⁻⁸) > K^AgI_SP@165 (1.4 × 10⁻¹⁰)

In the reaction mixture (considering the case for 8 mL 2.7 × 10⁻³ M halide salt solution in 54 mL EG), the product of the concentrations of silver ion (C_Ag⁺) and the corresponding halide ion (C_X⁻) in the final volume is ∼4.3 × 10⁻⁶. Therefore precipitation in the reaction medium is expected for all of the silver halides (X = Cl, Br, and I). The extent of deposition will increase in the order AgCl < AgBr < AgI and the AgX deposition increases as the reaction medium cools down to room temperature. The preferred facial deposition will steer the anisotropic growth and this finally leads to an extended nanowire moving from chloride to bromide as shown in Fig. SI3†. In practice we could increase the aspect ratio of the generated nanowires by two times by adding NaBr instead of NaCl and the synthesized nanowires were elongated to 50–60 μm in length. The presence of the iodide ion is, however, observed to inhibit wire formation and a mixture of quasi-spherical particles and much smaller rods are the end products. The role of the iodide ion is not totally comprehensible but it is suggested that I⁻ preferentially binds to the {111} facets and thereby prevents further growth along that direction.³³

Purification, networking, leaching, and preservation of synthesized silver nanowires

Once we synthesized silver nanowires with variable aspect ratios; un-reacted NaCl, PVP, AgNO₃, and smaller silver spheroids formed in the course of reaction have been removed by repeated low speed centrifugation using EG as solvent. Depending on the dimension of the generated silver nanorods we applied variable centrifugation speed between 1000 and 3500 rpm for selective nanorod/wire precipitation and removal of the un-reacted chemicals along with smaller silver spheroids as top decant. Due to parallel arrangement as well as the cross linking nature of the nanowires in the solution, bigger silver spheroid side products got stuck in the nanowire network and it was difficult to remove them by simple fractional centrifugation. To remove the bigger spheroids, we adopted a sonication cum fractional centrifugation technique where we sonicate the purified nanowires at 40% power of 53 kHz frequency at 20 °C for 1.0 min followed by centrifugation to remove the spheroids and collected the lower portion as pure nanowires. In the sonication process, a sound wave with a high frequency (>20 kHz) was used to shake off the nanowires to remove the bigger silver spheroids and make ultra pure silver nanowires. The purity of the nanowires before and after centrifugation has been depicted in Fig. SI4† which clearly shows that we could effectively remove almost all the silver spheroids by using this sonication cum fractional centrifugation method. After this purification step we followed the leaching technique described in the experimental section and simply by controlling the extent of the gold ion (10⁻³ M) addition we could easily achieve fractional leaching to generate differently hollowed Au–Ag hybrid nanotubes. Details of this controlled generation of hollow nanotubes has been shown in Fig. SI5.† To make interlinked nanowire networking, we have used dithiothreitol (HS–CH₂–CH(OH)–CH(OH)–CH₂–SH) as a linker in the reaction mixture. Keeping the reaction conditions the same, we have used different amounts of the 10⁻⁴ M dithiothreitol (DTT) solution in the reaction mixture. Though we have achieved nanowire formation at 165 °C in the absence of DTT, nanowire networking comes at 185 °C in the presence of DTT with a creamy yellowish colour for the end product. The effect of dithiothreitol addition in the mixture is clearly shown in Fig. SI6.† The corresponding UV-vis spectra are shown in Fig. SI7.† This not only gives us the ability to form nanowire networking but can also be used as effective resources for an atomic/molecular trapping material.

Highly specific and ultrasensitive Hg(II) detection by silver nanowires

By looking at the adverse effect of mercury on living systems and the Environmental Protection Agency (EPA) regulated maximum uptake limit of 2 ppb for inorganic mercury in drinking water, society requires an efficient sensing assay for the highly selective and ultrasensitive detection of mercury in ground water. The Localized Surface Plasmon Resonance (LSPR)-based UV-vis spectral detection technique²¹ could measure the soluble mercury concentration in the range of 10⁻⁴ to 10⁻⁵ M (20–2 ppm) which is well above the EPA-limit. In a recent publication, Fan et al.³⁴ presented a colorimetric method for quantitative recognition of Hg²⁺ with excellent selectivity up to a detection limit of ∼5 ppb which is close to the EPA-limit of Hg²⁺ in drinking water. In this report we have demonstrated a SERS based sensing of Hg²⁺ ions up to a detection limit of 50 ppt (10⁻¹⁰ M in solution), a value much lower than the upper limit determined by the EPA. SERS has been studied extensively in the last few decades³⁵ and promises to be one of the most useful analytical techniques for future applications in diagnostics and sensing.³⁶ The usefulness of this technique lies in its ability to provide the chemical signature along with its signal amplification (10⁸ to 10¹⁴ order), leading to state-of-the-art highly specific and sensitive assays for diagnosis, sensing and real-time monitoring.³⁶ Though bulk SERS enhancements have been understood to be largely plasmon-based³⁷ since the first experimental demonstrations,³⁸ resonance³⁹ and chemical contributions⁴⁰ can also be quite large under certain conditions. Notably, the Raman signal from 4-mercaptobenzoic acid (4-MBA) is achieved by its surface adsorption onto silver nanowires with a high aspect ratio. The Raman enhancement factor, G, for this study is measured experimentally by direct comparison of Raman signal originating from ν(CC)_{ring stretching} at 1589.7 cm⁻¹ in the presence and absence of the nanomaterials, as described in detail elsewhere.⁴¹ All spectra are normalized for integration time. The enhancement factor estimated from the SERS signal is approximately 5 × 10⁵ which is about two orders less than the corresponding spherical silver nanoparticles obtained by adding 250 μL of 2.7 mM NaCl in the reaction mixture. This intrinsically low Raman cross section of silver nanomaterials is clearly observable in Fig. SI8.† The variation in the SERS intensity with aspect ratio (AR) for NaCl induced (the synthesis protocol remains same and only the NaCl concentration changes) silver nanowires has been depicted in Fig. SI8† with an inset which shows the effect of the halogen containing common salt for the maximum achievable Raman intensity. It is evident from Fig. SI8† that the SERS signal gradually decreases with AR and gives the maximum Raman intensity for the sample that is mostly composed of spherical particles. The lower Raman cross section of silver nanowires compared to their spherical counterpart can easily be explained²¹ by considering their relative surface-to-volume ratio (S/V). If we consider a spherical particle with the radius R_sp then the (S/V)_sp = 3/R_sp and (S/V)_wr = 2/R_wr for a nanowire with radius R_wr and length h. In the present set of experiments, the R_sp (∼6 nm) < R_wr (25 nm), (S/V) for the nanowire is ∼6 times less than the spherical particles obtained at a lower concentration of NaCl (12 nm diameter Ag nanoparticles obtained in the presence of 250 μL 2.7 mM NaCl). Moreover, the relative Raman cross section based on the S/V ratio is length independent, the obtained Raman intensity from surface adsorbed molecules (4-MBA in this work) on different length nanowires remains more or less same if their diameter is approximately constant at 50 nm and this is clearly observable from Fig. SI8.† It is logical to think that this independence of Raman cross section on the length of the nanowire is applicable only for the same kind of nano-surface and the Raman enhancement, which is highly dependent on the surface composition, should be different for nanowires synthesized using different halogen containing common salts (NaCl, NaBr, and NaI). Dependence of Raman cross section on the surface composition is clearly shown in the inset where we have observed the maximum Raman cross section for NaBr-induced nanowires and the minimum from NaI-induced one. This could be explained by the networked structure of the NaBr-induced nanowires which have an enhanced adsorption ability compared to the NaCl-induced nanowires which mostly generates non-interacting individual nanowires. In the case of NaI, the generated nanoparticles are spherical in nature with an average diameter of ∼100 nm and should produce 1.3 times less Raman cross section than a nanowire with an average diameter of 50 nm, as we have argued before and is clearly observable from the inset of Fig. SI8.† Although the SERS is greater for a smaller spherical Ag nanoparticle, nanowires are suitable for the Hg²⁺ sensing experiment due to the preferential binding of Hg²⁺ on the {111} facet as all the {100} facets are blocked by insoluble silver halides which prevent the direct access of the analyte to the nanowire surface and initiate the leaching process through tips. Here the preferential binding of any added metal on the {111} facet stems from the unavailability of the {100} facet and not due to selective binding. It is worth mentioning that the literature is rich with examples that show the higher specificity of mercury to form a silver amalgam compared to other metals.²⁰ As a result of the unavailability of the {100} facet for binding, any of the added metals will be directed solely towards the {111} facet but due to greater specificity to form silver amalgam, other metals will show negligible binding which allows highly specific and selective detection through a SERS quenching experiment. It is also well-known that the most exciting optoelectronic activities of nanomaterials stem from their nanoscale sharp tips.^42–44 The expected increase in the catalytic and plasmonic activity at sharp features (e.g., corners, crease edges, cracks, branches, tips, etc.) of the nanoparticles stem from their very low neighbouring atom density surrounding the corner atoms compared to edge atoms which offers little restoring force⁴⁵ and opens up the gate for easy access to Hg²⁺ ions for chemical reaction through the tips. Although the obtained maximum Raman signal enhancement is obtained by NaBr-induced silver nanowires, NaCl-induced nanowires give the most consistent results with a relatively high Raman signal strength and we have done all our sensing experiments using NaCl-induced silver nanowires. When Hg²⁺ is added in the solution it causes leaching of the nanowires and is expected to form a silver–mercury alloy. Unlike bulk silver which has a comparable reduction potential to that of mercury (the standard electrode potential of Ag⁺/Ag = 0.80 V vs. SHE and Hg²⁺/Hg = 0.85 V vs. SHE), at the nano level the reduction potential of silver is reduced with a decrease in size⁴⁶ and Hg²⁺ is capable of efficiently reducing Ag to Ag⁺. As a result, partial deformation of the nanowires takes place and there is a gradual decrement of the SERS signal produced as we move from lower concentrations of mercury to a higher one. As a control experiment we have performed the same experiment of nanowire deformation using Au³⁺ instead of Hg²⁺. Due to the large difference in reduction potential of the two metals (the standard electrode potential of Ag⁺/Ag = 0.80 V vs. SHE and AuCl₄⁻/Au = 0.99 V vs. SHE),⁴⁷ gold ions can easily oxidize silver nanowires to silver ions in a nano-galvanic cell reaction and our usual expectation is to achieve greater Raman intensity quenching as a result of the nanowire deformation. In reality, we have observed much less nanowire etching in the case of Au³⁺ addition compared to Hg²⁺ addition. This could possibly be due to the large lattice mismatch between Hg and Ag and is not the case for Au and Ag (Au = 4.08 Å, Ag = 4.07 Å).⁴⁸ As a result of this crystal lattice matching, gold crystallizes on the {100} facets (γ(111) < γ(100) < γ(110)²⁹) to produce hollow tubes with a comparable surface for Raman activity but Hg²⁺ leaches silver nanowires through the {111} facets to deform nanowires and results in a reduced Raman intensity. The difference in the SERS intensity due to the addition of Au³⁺ and Hg²⁺ isshown in Fig. SI9.† At lower concentrations of Au³⁺ (below 10⁻⁴ M), there is apparently no change in Raman intensity as the surface adsorbed 4-MBA does not change the effective surface area very much. At 50 ppm concentration of Hg²⁺, the deformation of silver nanowire can be vividly observed by the naked eye. At this level the signal is almost totally quenched. To verify the efficiency of the silver nanowire-induced Raman signal of 4-MBA for the ultra-sensitive quantification of Hg²⁺, we have performed the SERS experiment at different concentrations of Hg²⁺ prepared from one stock solution of 1 M. The 4-MBA and Hg²⁺ solutions were made in a buffer of pH-4 throughout the experiment so that the metal ions stay stable in the medium and can effectively bind on the nanowire surface. Concentration dependent Hg²⁺ sensing using NaCl-induced silver nanowire is shown in Fig. 6. The left inset of Fig. 6 indicates how the Raman intensity at 1589.7 cm⁻¹ [ν(CC)_{ring stretching}: 1584 cm⁻¹]⁴⁹ quenched as we increase the concentration of Hg²⁺. It is clear from the plot that at lower concentrations of Hg²⁺, Raman quenching at 1589.7 cm⁻¹ changes linearly up to a certain limit (∼1 ppb) but after that the variation is not linear any more and remains almost constant before the concentration of Hg²⁺ is too high to completely dissolute the nanostructure and results in the maximum possible Raman quenching. At the low concentration range (50–700 ppt) we have plotted SERS quenching at 1589.7 cm⁻¹ against the Hg²⁺ concentration which has been shown as the right side inset in Fig. 6. It is clear from the plot that the SERS quenching signal intensity varies linearly (a prerequisite of a sensor) with the concentration of Hg²⁺ up to 700 ppt, which clearly demonstrates that it is possible to use our SERS assay as a sensor for the quantification of Hg²⁺ from a water sample at low impurity level. The obtained Hg²⁺ detection limit of our study is 50 ppt which is of the order of 10⁻¹⁰ M in solution, a value much lower than the EPA determined upper limit of Hg²⁺ in drinking water. For real-life applications, e.g., in an environmental sample, there could be several different impurities due to other common metal ions. To check the selectivity of our assay for Hg²⁺ over other metal ions, we performed the same SERS experiment with the chloride salts of Pb²⁺, Fe³⁺, Cu²⁺, As³⁺, Cr³⁺, Cd²⁺, Zn²⁺, Co²⁺, Ni²⁺, Sr²⁺, K⁺, and Na⁺. It is observed that at the 50 ppm level (10⁻² M concentration) of the salt solutions, many of these ions seem to interfere whereas at 5 ppm level (10⁻³ M concentration) only Pb²⁺, Cu²⁺, As³⁺, and Cr³⁺ ions produce noticeable quenching interference. However, at 500 ppb (10⁻⁴ M concentration), the interferences of Pb²⁺, Cu²⁺, As³⁺, and Cr³⁺ are evidently much weaker and at 50 ppb (10⁻⁵ M concentration) or below Hg²⁺ shows selective Raman quenching, as shown in Fig. 7. However, it can be stated that the extent of quenching caused by Hg²⁺ ion always remains sufficiently high enough compared to the others for the detection of this notoriously toxic element in a sample mixture. To demonstrate the potential practical applications of our assay to measure the Hg²⁺ content in environmental sample, spiking experiments were performed and the recovery values were determined. For this purpose, we have collected water samples from Kolkata Municipality Canal, Saha Institute of Nuclear Physics tap water, and bottled purified drinking water. We have filtered out the water samples using a 0.2 μm syringe filter to remove any of the larger contaminants present in the water samples. Subsequently, different concentrations of Hg²⁺ were spiked and SERS was performed to determine the Hg²⁺ concentration. In Table 1, the supplied Hg²⁺ concentration and Hg²⁺ concentrations determined by our SERS probe are compared. Our data show that the recovery values are quite good within the experimental error. Our experimental data confirms the capability of our assay to detect a low concentration of Hg²⁺ ions (below 1 ppb) from environmental samples.

Fig. 6 Concentration dependent Hg²⁺ sensing using NaCl-induced silver nanowires. The concentration of Hg²⁺ has been varied between 50 ppt and 50 ppm. The left inset indicates how the Raman intensity at 1589.7 cm⁻¹ quenched as we increase the concentration of Hg²⁺. It shows linear variation of Raman quenching at 1589.7 cm⁻¹ with Hg²⁺ concentration up to ∼1 ppb limit which has been shown in the right inset. The experimental data in the right inset is represented by (■), and the line represents the linear fit with R = 0.98.

Fig. 7 Raman quenching efficiency of Hg(II) at different concentrations ((A): 50 ppm, (B): 5 ppm, (C): 500 ppb, and (D): 50 ppb) which shows that the silver nanowire-4MBA composite can act as an excellent sensor for selective mercury ion detection in aqueous solutions at or below a concentration of 50 ppb.

Table 1 Comparison of supplied and measured concentrations for Hg²⁺ from different sources of water samples

Water sample	Supplied Hg^2+	Determined Hg^2+
Kolkata canal water	600 ppt	670 ± 40 ppt
SINP tap water	500 ppt	480 ± 60 ppt
Bottled drinking water	400 ppt	370 ± 20 ppt

---

To check the structural effect of silver hollow wire on Hg²⁺ sensing, we have performed the same set of SERS experiments but the results were not as encouraging as the original Raman signal originating from adsorbed 4-MBA on hollow nanowires was intrinsically very low and not suitable for high throughput sensing experiments.

Conclusions

In this work we have put forward a plausible explanation based on the real-time direct experimental observation (by HRTEM) of each and every intermediate state during growth for the common salt induced aspect ratio tunable Pentagonal Faceted Pyramidal Tipped (PFPT) silver-nanowire synthesis. Real-time direct experimental observations and the subsequent nanowire growth has been explained by considering the free energies associated with the crystallographic planes, the solubility product of the in situ generated silver halides, and binding affinity of the added halides at the {111} facets on the growing front. The reported results present the first controlled aspect ratio tunable nanowires synthesis technique simply by varying the salt concentration or by changing the halogen substituent in the common salt. This research work also prescribes an easy way to purify nanowires from spheroids, a ligation technique for nanowire networking, and tubular silver nano-cages which may find several future applications. Driven by the needs, we have used these synthesized nanowires for higher specific (preferential binding of Hg²⁺ on {111} facets of silver nanowires over other metals) and ultra sensitive Hg²⁺ detection in the ppt level (50 ppt) which is 40 times lower in concentration than the EPA approved upper limit of Hg²⁺ in the drinking water. We strongly believe that this detailed study on silver nanowire will find enormous applications in future trace level sensing technology.

Acknowledgements

DS would like to express his sincere gratitude to BARD project (PIC No. 12-R&D-SIN-5.04-0103), DAE, Govt. of India; the Department of Biotechnology (No. DBT/IC2/Indo-Russia/2014-16/05); and Ministry of Science and Education of Russian Federation in the framework of Increase Competitiveness Program of NUST “MISIS” (No K3-2015-063) and Russian-Indian joint project (No RFMEFI58414X0007) for their generous funding. We wish to give special thanks to Prof. Tapas K. Chini, SPMS Division, SINP for useful discussion related to SEM measurements; Prof. P. M. G. Nambissan and Mrs Soma Roy, ANP Division, SINP for NIR measurements.

Notes and references

1. (a) F. Kim, J. H. Song and P. Yang, J. Am. Chem. Soc., 2002, 124, 14316–14317; (b) C. J. Murphy and N. R. Jana, Adv. Mater., 2002, 14, 80–82.
2. (a) Y. N. Xia, P. D. Yang, Y. G. Sun, Y. Y. Wu, B. Mayers, B. Gates, Y. D. Yin, F. Kim and Y. Q. Yan, Adv. Mater., 2003, 15, 353; (b) W. L. Barnes, A. Dereux and T. W. Ebbesen, Nature, 2003, 424, 824; (c) J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao and R. P. Van Duyne, Nat. Mater., 2008, 7, 442.
3. F. J. Beck, A. Stavrinadis, S. L. Diedenhofen, T. Lasanta and G. Konstantatos, ACS Photonics, 2014, 1, 1197–1205.
4. (a) A. K. Singh, D. Senapati, S. Wang, J. Griffin, A. Neely, P. Candice, K. M. Naylor, B. Varisli, J. R. Kalluri and P. C. Ray, ACS Nano, 2009, 3, 1906–1912; (b) Q.-Q. Wang, J.-B. Han, H.-M. Gong, D.-J. Chen, X.-J. Zhao, J.-Y. Fengand and J.-J. Ren, Adv. Funct. Mater., 2006, 16, 2405–2408; (c) X. Xiong, C.-L. Zou, X.-F. Ren, A.-P. Liu, Y. X. Ye, F.-W. Sun and G.-C. Guo, Laser Photonics Rev., 2013, 1–19.
5. (a) A. K. Rath, F. P. Garcia de Arquer, A. Stavrinadis, T. Lasanta, M. Bernechea, S. L. Diedenhofen and G. Konstantatos, Adv. Mater., 2014, 26, 4741–4747; (b) F. J. Beck, T. Lasanta and G. Konstantatos, Adv. Opt. Mater., 2014, 2, 493–500; (c) A. Mihi, F. J. Beck, T. Lasanta, A. K. Rath and G. Konstantatos, Adv. Mater., 2014, 26, 443–448.
6. (a) Y. Huang, X. Duan, Q. Wei and C. M. Lieber, Science, 2001, 291, 630–633; (b) M. S. Gudiksen, U. J. Lauhon, J. Wang, D. C. Smith and C. M. Lieber, Nature, 2002, 415, 617–620; (c) N. I. Kovtyukhova and T. E. Mallouk, Chem.–Eur. J., 2002, 8, 4354–4363; (d) N. A. Melosh, A. Boukai, F. Diana, B. Gerardot, A. Badolato, P. M. Petroff and J. R. Heath, Science, 2003, 300, 112–115.
7. (a) A. Tao, F. Kim, C. Hess, J. Goldberger, R. He, Y. Sun, Y. Xia and P. Yang, Nano Lett., 2003, 3, 1229–1233; (b) B. Zhang, H. Wang, L. Lu, K. Ai, G. Zhang and X. Cheng, Adv. Funct. Mater., 2008, 18, 2348–2355.
8. (a) Y. Bi, H. Hu and G. Lu, Chem. Commun., 2010, 46, 598–600; (b) Y. Sun and Y. Xia, J. Am. Chem.Soc., 2004, 126, 3892–3901; (c) Y. Sun, Y. Yin, B. T. Mayers, T. Herricks and Y. Xia, Chem. Mater., 2002, 14, 4736–4745; (d) Y. Sun, B. Gates, B. T. Mayers and Y. Xia, Nano Lett., 2002, 2, 165–168; (e) B. Wiley, Y. Sun and Y. Xia, Acc. Chem. Res., 2007, 40, 1067–1076; (f) Y. Xia, D. Yang, Y. Sun, Y. Wu, B. Mayers, B. Gates, Y. Yin, F. Kim and H. Yan, Adv. Mater., 2003, 15, 353–389.
9. S. Liu, J. Yue and A. Gedanken, Adv. Mater., 2001, 13, 656–658.
10. Y. Zhou, S. H. Yu, X. P. Cui, C. Y. Wang and Z. Y. Chen, Chem. Mater., 1999, 11, 545–546.
11. Y. Zhou, S. H. Yu, C. Y. Wang, X. G. Li, Y. R. Zhu and Z. Y. Chen, Adv. Mater., 1999, 11, 850–852.
12. (a) C. R. Martin, Science, 1994, 266, 1961–1966; (b) B. R. Martin, D. J. Dermody, B. D. Reiss, M. Fang, L. A. Lyon, M. J. Natan and T. E. Mallouk, Adv. Mater., 1999, 11, 1021; (c) Z. Zhang, D. Gekhtman, M. S. Dresselhaus and J. Y. Ying, Chem. Mater., 1999, 11, 1659–1665.
13. (a) Y. J. Han, J. M. Kim and G. D. Stucky, Chem. Mater., 2000, 12, 2068–2069; (b) M. H. Huang, A. Choudrey and P. Yang, Chem. Commun., 2000, 1063–1064.
14. (a) P. M. Ajayan and S. Iijima, Nature, 1993, 361, 333–334; (b) D. Ugarte, A. Chatelain and W. A. de Heer, Science, 1996, 274, 1897–1899.
15. D. Zhang, L. Qi, J. Ma and H. Cheng, Chem. Mater., 2001, 13, 2753–2755.
16. E. Braun, Y. Eichen, U. Sivan and G. Ben-Yoseph, Nature, 1998, 391, 775–778.
17. (a) N. R. Jana, L. Gearheart and C. J. Murphy, Chem. Commun., 2001, 617–618; (b) M. A. El-Sayed, Acc. Chem. Res., 2001, 34, 257–264.
18. B. H. Hong, S. C. Bae, C.-W. Lee, S. Jeong and K. S. Kim, Science, 2001, 294, 348–351.
19. (a) H. H. Song, K. M. Jones and A. A. Baski, J. Vac. Sci. Technol., A, 1999, 17, 1696; (b) M. P. Zach, K. H. Ng and R. M. Penner, Science, 2000, 290, 2120–2123.
20. (a) L. Deng, X. Ouyang, J. Jin, C. Ma, Y. Jiang, J. Zheng, J. Li, Y. Li, W. Tan and R. Yang, Anal. Chem., 2013, 85, 8594–8600; (b) G. V. Ramesh and T. P. Radhakrishnan, ACS Appl. Mater. Interfaces, 2011, 3, 988–994; (c) L. Chen, X. Fu, W. Lu and L. Chen, ACS Appl. Mater. Interfaces, 2013, 5, 284–290; (d) F. Jelen, B. Yosypchuk, A. Kourilova, L. Novotny and E. Palecek, Anal. Chem., 2002, 74, 4788–4793.
21. X. C. Jiang, S. X. Xiong, C. Y. Chen, W. M. Chen and A. B. Yu, J. Nanopart. Res., 2011, 13, 5087–5101.
22. S. A. Khan, R. Kanchanapally, Z. Fan, L. Beqa, A. K. Singh, D. Senapati and P. C. Ray, Chem. Commun., 2012, 48, 6711–6713.
23. G. W. Sławiński and F. P. Zamborini, Langmuir, 2007, 23, 10357–10365.
24. C. X. Kan, J. J. Zhu and X. G. Zhu, J. Phys. D: Appl. Phys., 2008, 41, 155304.
25. (a) M. Giersig, I. Pastoriza-Santosb and L. M. Liz-Marzán, J. Mater. Chem., 2004, 14, 607–610; (b) X. Gu, C. Nie, Y. Lai and C. Lin, Mater. Chem. Phys., 2006, 96, 217–222; (c) G. Zhou, M. Lu, Z. Yang, H. Zhang, Y. Zhou, S. Wang, S. Wang and A. Zhang, J. Cryst. Growth, 2006, 289, 255–259; (d) L. Sun, L. Wang, Y. Song, C. Guo, Y. Sun, C. Peng, Z. Liu and Z. Li, Appl. Surf. Sci., 2008, 254, 2581–2587.
26. Y. Sun, B. Mayers, T. Herricks and Y. Xia, Nano Lett., 2003, 3, 955–960.
27. (a) T. K. Sau and C. J. Murphy, J. Am. Chem. Soc., 2004, 126, 8648–8649; (b) M. Liu and P. G. Sionnest, J. Phys. Chem. B, 2005, 109, 22192–22200; (c) M. Grzelczak, J. Pérez-Juste, P. Mulvaney and L. M. Liz-Marzán, Chem. Soc. Rev., 2008, 1783–1791; (d) Z. Li, W. Li, P. H. C. Camargo and Y. Xia, Angew. Chem., Int. Ed., 2008, 47, 9653–9656; (e) H. Zhang and D. Wang, Angew. Chem., Int. Ed., 2008, 47, 3984–3987.
28. Y. Sun and Y. Xia, Adv. Mater., 2002, 14, 833–837.
29. Z. L. Wang, J. Phys. Chem. B, 2000, 104, 1153–1175.
30. B. Wiley, Y. Sun and Y. Xia, Acc. Chem. Res., 2007, 40, 1067–1076.
31. CRC Handbook of Chemistry and Physics, 90th edn, CD-ROM version, 2010.
32. D. D. Wagman and M. V. Kilday, J. Res. Natl. Bur. Stand., Sect. A 1973, 77, 569–579.
33. D. K. Smith, N. R. Miller and B. A. Korgel, Langmuir, 2009, 25, 9518–9524.
34. Y. Fan, Z. Liu, L. Wang and J. Zhan, Nanoscale Res. Lett., 2009, 4, 1230–1235.
35. (a) W. Lu, A. K. Singh, S. A. Khan, D. Senapati, H. Yu and P. C. Ray, J. Am. Chem. Soc., 2010, 132, 18103–18114; (b) C. L. Haynes, C. R. Yonzon, X. Zhang and R. P. Van Duyne, J. Raman Spectrosc., 2005, 36, 471–484; (c) L. X. Quang, C. Lim, G. H. Seong, J. Choo, K. J. Do and S. K. Yoo, Lab Chip, 2008, 8, 2214–2219; (d) T. Vo-Dinh, F. Yan and M. B. Wabuyele, J. Raman Spectrosc., 2005, 36, 640–647; (e) D. Senapati, S. S. R. Dasary, A. K. Singh, T. Senapati, H. Yu and P. C. Ray, Chem.–Eur. J., 2011, 17, 8445–8451.
36. (a) R. A. Alvarez-Puebla and L. M. Liz-Marzan, Small, 2010, 6, 604–610; (b) P. J. Vikesland and K. R. Wigginton, Environ. Sci. Technol., 2010, 44, 3656–3669.
37. (a) M. Moskovits, Rev. Mod. Phys., 1985, 57, 783–826; (b) S. Zou and G. C. Schatz, Chem. Phys. Lett., 2005, 403, 62–67.
38. M. Fleischmann, P. J. Hendra and A. J. McQuillan, Chem. Phys. Lett., 1974, 26, 163–166.
39. (a) A. B. Myers, and R. A. Mathies, in Biological applications of Raman spectroscopy, ed. T. G. Spiro, Wiley and Sons, New York, 1987, p. 1; (b) A. B. Myers, Acc. Chem. Res., 1997, 30, 519–527.
40. (a) P. S. Kumar, I. Pastoriza-Santos, B. Rodríguez-González, F. J. G. de Abajo and L. M. Liz-Marzán, Nanotechnology, 2008, 19, 015606; (b) B. N. J. Persson, Chem. Phys. Lett., 1981, 82, 561–565.
41. M. Bardhan, B. Satpati, T. Ghosh and D. Senapati, J. Mater. Chem., 2014, 2, 3795–3804.
42. G. T. Boyd, Z. H. Yu and Y. R. Shen, Phys. Rev. B: Condens. Matter Mater. Phys., 1986, 33, 7923–7936.
43. C. K. Chen, T. F. Heinz, D. Ricard and Y. R. Shen, Phys. Rev. B: Condens. Matter Mater. Phys., 1983, 27, 1965–1979.
44. G. T. Boyd, T. Rasing, J. R. R. Leite and Y. R. Shen, Phys. Rev. B: Condens. Matter Mater. Phys., 1984, 30, 519–526.
45. D. Mandal, S. Mondal, D. Senapati, B. Satpati and M. V. Sangaranarayanan, J. Phys. Chem. C, 2015, 119, 23103–23112.
46. N. Pradhan, A. Pal and T. Pal, Colloids Surf., A, 2002, 196, 247–257.
47. Y. Sun, B. T. Mayers and Y. Xia, Nano Lett., 2002, 2, 481–485.
48. W. P. Davey, Phys. Rev., 1925, 25, 753–761.
49. D. He, B. Hu, Q.-F. Yao, K. Wang and S.-H. Yu, ACS Nano, 2009, 3, 3993–4002.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra02900k

---