Modular Plasmonic Antennas Built of Ultrathin Silica-Shell Silver-Core

Jun 10, 2014 - Modular Plasmonic Antennas Built of Ultrathin Silica-Shell Silver-Core Nanoparticles ... Analyte molecules are attached to the silica s...
1 downloads 11 Views 7MB Size
Article pubs.acs.org/Langmuir

Modular Plasmonic Antennas Built of Ultrathin Silica-Shell SilverCore Nanoparticles Nir Zohar and Gilad Haran* Chemical Physics Department, Weizmann Institute of Science, 76100 Rehovot, Israel S Supporting Information *

ABSTRACT: Dimers of metallic nanoparticles can serve as antennas to locally enhance optical fields through plasmonic excitations. Such dimeric structures can be particularly useful for sensing applications using surface-enhanced Raman scattering (SERS). It has been challenging to devise a synthetic scheme that facilitates creating dimer antennas from different types of particles, at the same time allowing control over the size of the gap between the particles and enabling the introduction of any molecule into the gap. We describe here a method that answers to this challenge. We first introduce a recipe for the creation of a silica shell as thin as 1 nm on silver particles. Analyte molecules are attached to the silica shell, and finally, the silica-shell silver-core particles, whose surface is negatively charged, are mixed with positively charged bare silver particles to create dimers. A demonstration of SERS from individual dimers with gaps of 1.4 and 3.7 nm paves the way to systematic studies of the effect of gap size and composition on plasmonic enhancement.



INTRODUCTION Plasmonic antennas are nanoscale metallic structures that can be used for applications such as sensing 1 and light manipulation.2 They derive their utility from their ability to localize light via the concerted oscillation of conduction electrons, also known as localized surface plasmon resonance (LSPR). The optical properties of a nanoantenna are conjugated to its structure via these charge oscillations, which induce an enhanced electromagnetic field around the antenna.1 Antennas that consist of more than one nanostructure form “hot-spots”, regions of strong enhancement within the junctions between pairs of particles.3−10 This strong enahncement can be attributed to coupling between the LSPRs of adjacent structures, which strongly depends on the distance between them,5 but can be modified through other structural parameters, such as the materials,11,12 the size,13 and the symmetry14 of the particles forming the hot-spot. A successful method for the generation of two-particle plasmonic antennas for sensing applications should therefore involve the following three attributes: (1) control on the interparticle gap size over the relevant range for strong plasmonic coupling, (2) flexibility in the choice of the material and structure of the particles used, and (3) the ability to graft molecules efficiently onto the antenna, specifically into the gap region. Multiple efforts to develop methods that involve these attributes have been reported in recent years. The simplest method to generate dimer antennas is by addition of salt to a metallic colloidal solution. This is a very popular method of preparation,15−20 but it does not offer any control over the interparticle gap. The gap can be controlled, however, by introducing a spacer material to hold the particles at a predetermined distance from each other. A simple approach for © 2014 American Chemical Society

introducing such a spacer is to functionalize the surfaces of the particles. A layer of molecules, such as DNA strands,21,22 proteins,9 or long-chained organic molecules6 can be deposited or bound, replacing the molecules that stabilize the particles during their synthesis. DNA molecules might seem like convenient linkers between particles, allowing much flexibility in selecting the radii of the particles and the materials, and offering tunability of the gap through the choice of their length.11,23 However, in practice, the flexibility of DNA molecules and chemical linkers attaching them to the metallic structures may drastically limit the accuracy of gap tunability.21 Suh and co-workers deposited silver on DNA-tethered gold dimers to rigidify their structure,8,22 and Thacker et al. used DNA origami to make more rigid constructions,10 but in both cases the gap size was limited to ∼3 nm and above. Smaller gap sizes might be achieved more easily using organic molecules13 or macromolecules, such as proteins9 or polymers,24 but in this case again the gap size is difficult to control and usually cannot be varied. Further, incorporation of molecules into the gaps as spacers might lead to a strong background in SERS experiments. Metallic particles with shells of a rigid, inorganic material, such as silica, may make better precursors for nanoplasmonic structures. The potential of dielectric-shell metal-core systems as adjustable SERS amplifiers has been demonstrated in recent years using various configurations, typically with multiple particles.25−27 Different dielectrics, namely, silica, alumina, and tetrahedral amorphous carbon, were used to produce Received: May 3, 2014 Revised: June 9, 2014 Published: June 10, 2014 7919

dx.doi.org/10.1021/la5017033 | Langmuir 2014, 30, 7919−7927

Langmuir

Article

“bare” nanoparticles to the silica-coated particles. By varying the silica shell thickness, we are able to control the gap size from zero to essentially any size in the relevant range for plasmonic sensing. Finally, we show the utility of the synthesized dimers for SERS studies by attaching carboxytetramethylrhodamine (TAMRA) molecules onto the nanoantennas and studying their Raman spectra.

homogeneous shells, and their thickness was precisely adjusted. (Recently, shells as thin as 1 nm were reported on gold nanoparticles.25) Tian and co-workers28 showed that a population of core−shell particles is an efficient SERS amplifier, featuring precise control over the shell thickness from 2 to 20 nm. However, uncontrollable aggregation of a homogeneous population (as used by these authors) did not allow any regulation of the geometry of the hot-spots, and inherently limited the minimal gap size to twice the shell thickness, that is, ∼4 nm. Better control over the aggregate size was demonstrated by Van Duyne and co-workers, who encapsulated small clusters of particles together with adsorbed analyte molecules in silica shells in a process that, however, lacked control on gap size.29−31 Thomas and co-workers showed recently how the ammonium salt of pyrene can be used to create homodimers of preformed silica-shell silver-ore particles, and measured the SERS signal of the pyrene molecules.32 The gap size was varied over a broad range, but gaps smaller than 3 nm were not reported, and the authors relied on the analyte molecule to form the dimers, which limited the generality of their process. We propose here a method to generate dimer antennas that answers to all three criteria posed above. Our method involves electrostatic association of bare metal particles with particles coated by silica shells of a tunable thickness (Figure 1). A scheme for generating silica shells as thin as ∼1 nm on silver particles is introduced in this paper and is essential for the method presented.



EXPERIMENTAL SECTION

Chemicals. Distilled water (18.2 MΩ·cm) was obtained from a water-purifying system (Millipore Synergy); silver oxide (Ag(I)O, 99.99%) was purchased from AlfaAeser (42577); APTMS ((3aminopropyl)trimethoxysilane, 97%, 281778), ATC (acetylthiocholine chloride, >99%, powder, A5626), sodium silicate (26.5%, reagent grade, 338443), and TAMRA-NHS (5-carboxy-tetramethylrhodamine N-succinimidyl ester, 53048) were purchased from Sigma; ethanol (96%, 1.00983), DMF (N,N-dimethylformamide, max. 0.003% H2O, 1.02375), TEA (triethylamine for synthesis, 8.08352), and sodium nitrate (reagent grade, 1.01512) were purchased from Merck; trisodium citrate (dihydrate) was purchased from BDH Laboratory supplies (102424L); HCl (hydrochloric acid, 37%) was purchased from BioLab (84105). Methods. Synthesis of Hydroxy-Stabilized SNP (SNP-OH). These particles were synthesized according to Evanoff and Chumanov33 on a smaller scale. A 250 mL round-bottomed Schlenk flask was cleaned using concentrated nitric acid and then washed with several volumes of distilled water. Finally, distilled water was refluxed for at least 1 h and discarded. The synthesis solution was prepared by mixing 130 mg of silver oxide with 130 mL of distilled water, and was then heated to ∼80 oC in a sealed, pressure resistant reflux system connected to a pressurized hydrogen container. Silver oxide dissolves poorly in water; the purpose of the heating was to increase [Ag(I)O(aq)] before the introduction of H2, though eventually most of it remained unreacted. Hydrogen was then allowed to flow into the reflux system at 10 psi. The solution began to turn yellow after 1−2 min, and as the particles grew, it became turbid. Usually, 7−10 min of reaction yielded ∼60 nm particles. The reaction was stopped by removing hydrogen from the system. The solution was then cooled down to room-temperature and filtered using 0.2 μm nylon filters (Sigma, Z290823). Silica Shell Synthesis. The shells were synthesized on the surface of SNP-OH particles as follows: 10 mL of an as-synthesized solution of particles (pH ∼ 5) were transferred to a 50 mL polypropylene tube, and 4.5 μL of fresh 1 mM APTMS solution were added. After 15 min at room temperature, 10 μL of sodium silicate were added. After mixing, the solution pH was ∼10. A 1 nm layer of silica was deposited after 1 h. When thicker silica shells were desired, 5−10 μL of sodium silicate were added at intervals of 12 h, increasing the thickness by 1−2 nm. To separate the SNP@SiO2, the solution was first diluted in water (8−9 equiv) to let residual silica dissolve. After ∼10 min, the particles were washed by repeated cycles of centrifugation and resuspension in water. More cycles of washing were applied when sodium silicate was added more than once, to verify that the residual silicate concentration was below the solubility product. Synthesis of Citrate-Stabilized Silver Nanoparticles (SNP-cit). These particles were synthesized according to Lee and Meisel.34 We found that the glass cleaning procedure as described above increases the homogeneity for this process as well. Typically 22.5 mg of sodium nitrate were dissolved in 125 mL of distilled water and boiled in a three-necked flask connected to a reflux setup. A volume of 2.5 mL of fresh 1% (w/v) trisodium citrate solution was quickly added, and the solution was boiled for another 15 min before it was cooled down. (The color of the solution turned yellow and then turbid and greyish.) To increase the homogeneity, the heavier particles were precipitated by centrifugation (1500 rcf, 15 min), and the pellet was discarded. Thiocholine-Stabilized Silver Nanoparticles (SNP-TC). ATC (0.2 mg/mL) was dissolved in 10 mM HCl, and de-esterified by incubating the solution for 16 h at 60 oC to generate thiocholine (TC). It was then cooled down to room temperature. SNP-cit were precipitated by centrifugation (1000 rcf for 15 min), the supernatant was discarded,

Figure 1. Geometrical parameter space of dimer nanoantennas. Schematically depicted here are several possible structures that demonstrate fine-tuning of plasmonic dimer nanoantennas structures. From right to left: variation of particle size, gap size, hot-spot symmetry, or particle composition (gray, silver; yellow, gold; blue, silica; purple dot, probe molecule). All of these structures are readily attainable using the method described in the paper.

We therefore start by describing a complete synthetic recipe for the generation of silica-shell silver-core (SNP@SiO2) particles. Our procedure is relatively quick (minimum time of 1−2 h) and modular, so it can be used with colloidal solutions of different materials, shapes, and sizes (Figure 1). We describe the synthesis and demonstrate it using spherical silver colloids, onto which 1 nm (and thicker) silica shells are deposited. The effects of several reaction parameters, namely, the presence of spectrator ions, sodium silicate concentration, and pH on the formation of the silica shells, are discussed. We then introduce our method for producing dimeric nanoantennas by attaching 7920

dx.doi.org/10.1021/la5017033 | Langmuir 2014, 30, 7919−7927

Langmuir

Article

Figure 2. SNP@SiO2 preparation. A general description of the steps in the synthesis of SNP@SiO2, introduced in detail in the text. Synthesis: hydroxyl-capped silver particles are prepared to avoid the need for a large concentration of counterions in solution. Silanization: APTMS is adsorbed onto the metal surface, to form an initial silica layer. Deposition: silica concentration is tightly controlled to prevent fast formation of thick layers. Separation: dilution of the particles in water and then sedimentation stops silica deposition. The values given here were used in the synthesis of 60 nm (diameter) silver nanoparticles with a 2.5 nm silica shell. The sodium silicate concentration and deposition time depend on the desired silica thickness, and the deposition step should be repeated if a thickness larger than 2.5 nm is desired. and the particles were resuspended in the TC solution. The colloids were then incubated for 2 h at 40 °C and precipitated by centrifugation. Linking a Probe Molecule to the Surface. Any molecule can be attached to the silica surface, as long as it has a functional group. The general strategy is to first attach the probe to a bifunctional linker which consists of a silyl group, which is then polymerized to the silica. For this work, we used TAMRA-NHS. Since the linker we used was APTMS, which is sensitive to humidity, the involved reagents were pretreated (DMF was kept with sodium sulfate, TEA was kept with potassium hydroxide) and were kept in dry environment (desiccated). Mixing of the reagents was done under nitrogen. In a typical synthesis, a TAMRA-NHS powder was dissolved in dry DMF and mixed with solutions of APTMS and TEA in dry DMF, at a molar ratio of 1:1.1:1.1. This reaction was allowed to stand for at least 16 h at room temperature. Then 1 mL of aqueous SNP@SiO2 was centrifuged (1000 rcf for 15 min) and 900 μL of the supernatant were discarded. The remaining pellet was resuspended in 900 μL of ethanol. TAMRAAPTMS solution was then added such that [TAMRA-APTMS] = 50 nM, and allowed to stand at room-temperature for 2 h. The particles were then centrifuged and washed twice with water. Nanoantenna Self-Assembly. Upon mixing the SNP-TC and any negatively charged colloids (SNP@SiO2 or SNP-cit), aggregation started immediately. To control the size of the aggregates, both colloidal solutions were diluted with distilled water, mixed, and quickly sprayed onto a transmission electron microscopy (TEM) grid over a period of 90 s. The colloidal solutions were diluted in such a way that in the last mixing the concentration of each batch was ∼1 pM. The spraying setup used a nitrogen stream to spread liquid droplets through a thin nozzle made of a syringe needle. Spreading the liquid in this way covered the surface with very small droplets that dried quickly, and so further aggregation was prevented.

removal of excess sodium silicate (Figure 2). These steps were optimized to allow deposition of silica layers as thin as ∼1 nm via stringent control on the concentrations of sodium and silicate ions. Below, we introduce each step in some detail. Particle Synthesis. Alkali metal ions, and especially sodium, interfere with silica deposition in the form of homogeneous layers since they induce coagulation.38 Using sodium silicate as a precursor imposes an unavoidable minimum concentration of sodium, and this concentration is usually even larger when particles are produced using the popular Lee and Meisel method.34 Ung et al.35 attempted to overcome this problem by adding an initial step of a two-day dialysis to their procedure, which resulted in removal of 80% of the spectator ions. We essentially totally avoided the presence of sodium ions by synthesizing SNPs from silver(I) oxide and hydrogen, as described by Evanoff and Chumanov.33 In this procedure, silver(I) oxide is reduced to silver by hydrogen gas: Ag 2O(aq) + H 2(g) → 2Ag 0(s) + H 2O(l)

This reaction produces SNPs capped by water or hydroxy groups.33 Our experiments showed that using hydrogen-reduced SNPs leads to silica shells that are quite uniform and aggregation is rare. It should also be noted that this procedure allows proceeding directly from particle synthesis to deposition, without an elongated dialysis in between. Silanization. The as-synthesized SNPs were then coated with APTMS. The silane concentration was calculated to obtain a full monolayer coverage of the metallic surface. Assuming that each molecule takes an area of approximately 0.4 nm2,39 approximately 38 000 molecules are required for a 60 nm particle. With a particle concentration of ∼10 pM, estimated using electron microscopy (see Supporting Information Figure S1), a concentration of 1 mM APTMS was more than enough to achieve the above. Silica Deposition. The silica shells were formed through the reaction of silicate ions with the particle surfaces. Importantly, during the silica deposition process, the only major source of ions was the added sodium silicate, so the pH level was directly linked to silicate concentration. It is known that the polymerization of silica proceeds faster with increasing alkalinity and precursor availability,38 which may lead to faster shell thickening as well as nucleation of silica particles. The latter may decrease the shell homogeneity and lead to the creation of particles with multiple silver cores in one shell. Since homogeneous deposition occurs over the pH range of ∼9− 11,35,38 achieving a monodisperse solution of particles, homogeneously coated with thin silica, mainly relies on the



RESULTS Generation of Silica-Shell Silver-Core Particles. We developed a novel protocol for the generation of silver particles coated with a silica shell as thin as 1 nm. Our method relies on the seminal work of Liz-Marzán, Mulvaney, and co-workers,35,36 who provided a complete recipe for coating gold and silver nanoparticles (SNPs) with homogeneous silica shells in solution. They showed that the addition of sodium silicate to an aqueous colloidal silver solution can result in the formation of silica layers as thin as ∼3 nm. However, they also reported that under conditions that should allow formation of thinner shells the particles tend to coagulate and form multinuclei silica shells. Hitherto, despite the growing interest in plasmonic systems with silica shells, silica layers thinner than ∼3 nm were not synthesized on silver colloids (although such layers were synthesized on gold surfaces,37 and recently on gold colloids28). The method we developed involved the following steps: metal particle synthesis via hydrogen reduction, silanization using APTMS, silica deposition from sodium silicate, and finally 7921

dx.doi.org/10.1021/la5017033 | Langmuir 2014, 30, 7919−7927

Langmuir

Article

Figure 3. Thin silica shells on SNPs. Examples of SNP@SiO2 with silica shells of several thicknesses. These silica shells were synthesized by the procedure described in Figure 2. The silica deposition times were 1 h (1.0 nm) and 16 h (2.5 nm). The thickness of 4.3 nm was achieved by a second addition of sodium-silicate 3 h after the first one, following which the solution was allowed to stand for additional 16 h.

Figure 4. Characterization of SNP@SiO2 preparations. The procedures described in the text were used to synthesize silica shells on SNPs with deposition times of 1 h (A and B) and 16 h (C and D). The distribution of silica layer thickness in each batch is shown in the histograms (B and D), which were generated from measurements on TEM images (A and C) of at least 50 individual particles in each case. The mean silica layer thickness was found to be 1.25 ± 0.25 and 2.5 ± 0.55 nm in B and D, respectively.

However, although deposition with orthosilicate concentration of 1.5 mM is possible, it was found that since the pH in this case is below 9 the deposition rate and quality are inferior. On the other hand, silica deposition in polypropylene tubes with a higher orthosilicate concentration (4.5 mM, pH ∼ 10) proceeded fast and well enough so that homogeneous shells of 1 nm thickness were formed after 1 h at room temperature. Particle Separation. The particles were isolated from the silicate solution by centrifugation. This is a key step, since it may lead to a simultaneous increase in the concentration of both particles and silicate ions. This can result in an uncontrallable shell thickening and coagulation, especially if only an initial silica layer of 1−3 nm has formed. The particles can be stabilized to some extent prior to centrifugation by a secondary deposition of silicate using ethanol,35 but we noticed that this results in a very inhomogeneous deposition and formation of large silica particles. We found that diluting the SNP solution with 8−9 volumes of water (thereby lowering the

presence of a precise concentration of silicate ions. Ung et al. found a “critical concentration” of orthosilicate (0.02% or 3.4 mM), below which the layers were too thin to prevent multicore formation on 10 nm SNPs. However, the solubility limit of silica in water is ∼100 ppm40 or 1.3 mM, which is about a third of the above critical concentartion. Moreover, in some of our experiments a concentration of ∼4 mM was sufficient to produce 2−3 nm shells on 60 nm particles, that is, with an ∼36 times larger surface area. These observations suggest that the effective orthosilciate concentration in solution is different from the one deduced by the amount of added sodium silicate. It is known that orthosilicate deposits readily on glass,40 which may bias (reduce) the concentration of free (available) orthosilicate in the solution. Indeed, by replacing the glassware with polypropylene tubes, we were able to obtain single-core homogeneously coated SNP@SiO2 using concentrations as low as 0.009% (orthosilicate concentration = 1.5 mM). 7922

dx.doi.org/10.1021/la5017033 | Langmuir 2014, 30, 7919−7927

Langmuir

Article

interact. In this method, there is no need to add salt into the solution as in aggregation of same-charge particles. In fact, in this case, the lower the salt concentration, the faster the reaction. Moreover, since no salt is added, there is no destabilization in any of the colloidal particle types, which reduces the chance of self-aggregation. Positively charged SNPs were obtained by exchanging the citrate ions stabilizing Lee−Meisel particles with thiocholine. SNP-cit particles are usually stable in solution, with a zeta potential well below −30 mV. The citrate molecules can be relatively easily displaced by molecules that bind silver more strongly, such as thiol-containing molecules, for example, the TC used here. The thiol group of TC binds strongly to the silver surface, leaving the positively charged choline group protruding into the solution. The new SNPs (SNP-TC) were found to have a zeta potential of +30 mV. Insertion of SERS Probes. For the purpose of SERS measurements, a probe molecule can be inserted at this stage using surface chemistry (on either the silica or the metal, vide infra). For the SERS measurements described below, we used TAMRA. To attach TAMRA molecules to the antennas, we prepared conjugates of its succinimidyl ester adduct (TAMRANHS) with two different linkers. For nanoantennas with no gap, consisting of a combination of SNP-cit and SNP-TC, we reacted TAMRA-NHS with mercaptoethanolamine, the thiol group of which was then used to attach the molecule to the SNP-cit surface. In nanoantennas with a gap, consisting of a combination of SNP-TC and SNP@SiO2, TAMRA-NHS was reacted with APTMS. The methylsilane group of APTMS was then used to attach the molecule to the silica layer. Since both mercaptoethanolamine and APTMS tend to polymerize in the presence of water, in both cases they were first reacted with the NHS moiety of the dye in a semidry environment (achieved by working under a flow of dry N2). The products of these reactions could be safely used in the colloidal aqueous environments. The APTMS-containing molecules were adsorbed onto the silica layer of SNP@SiO2 in ethanol in the presence of water (∼4%), which catalyzes the polymerization of APTMS on the silica surface.41 In both cases, our purpose was to attach the probe molecules homogeneously on the particle surfaces, so as to increase the probability that one or more probe molecules would be located within the hot spot of the dimer antennas. The concentration of molecules used in this step was 30−50 nM. Self-Assembly of Dimers. While the self-assembly of SNPTC and SNP@SiO2 can be verified using TEM, we wanted to optimize the controlled aggregation process for dimers. To do that, we turned to studying the dynamics of this process using UV−vis spectroscopy. Solutions of monomeric SNPs are characterized by a distinct extinction peak at the wavelength corresponding to their plasmon resonance frequency. The formation of aggregates can be followed through a decrease in the monomer plasmon peak and the appearance of a broad, red-shifted peak due to clustered particles. Solutions of particles with negatively charged surfaces were mixed with particles of the opposite surface charge, and a complete spectrum was taken at fixed time intervals as the reaction proceeded (Figure 6A). A control experiment, which was identical to the original except that particles with the same surface charge were mixed, did not show the same trend in spectra over time. Figure 6B shows the time course of the reaction, obtained by following the intensity at 440 nm. Since we used a very dilute solution of particles (∼5 pM), it is reasonable to assume that

silicate concentration well below the solubility limit), and letting it stand for a few minutes before centrifugation, allows some of the deposited silica to redissolve, breaking slightly fused shells, smoothing small shell deformations, and dissolving small silica particles. This procedure allowed the isolation of SNPs coated with exteremley thin silica shells after only 1 h of deposition at room temperature. Furthermore, thicker silica shells were generated using the same procedure by multiple additions of sodium silicate in time intervals of at least 3 h. Figure 3 shows particles of several different shell thicknesses, while Figure 4 presents the systematic characterization of two particular batches in which the silica was deposited for 1 h (A and B) and 16 h (C and D). From each batch approximately, 50 particles were imaged using TEM and the silica shell thickness was assessed. The histograms (Figure 4B and D) show mean sizes of 1.25 and 2.5 nm with standard deviations of 0.25 and 0.55 nm, respectively. The optical spectrurm of the particle solution was found to change only minimially upon silica deposition (see Supporting Information Figure S2). Nanoantenna Labeling and Assembly. Assembly of plasmonic dimer antennas from colloidal solutions requires that two particles would attach to each other. This can be achieved by reducing the stability of the colloidal solution through the addition of salt. However, this procedure may lead to significant coagulation and precipitation of large clusters of particles. We developed the means to achieve control over such an aggregation process. The main principle behind the method described here is the employment of the negative zeta potential of the SNP@SiO2 or citrate-coated SNPs (SNP-cit) to attach them to positively charged particles, as sketched in Figure 5. Mixing solutions of the two types of particles should lead to aggregation that can be controlled through the concentration of the particles and the time over which they are allowed to

Figure 5. General method for dimer nanoantenna synthesis. The antenna comprises two SNPs with opposite zeta potentials. (A) A batch of SNPs coated with a thin silica layer is prepared. The silica layer can be further functionalized to bind molecules (e.g., SERS probes). (B) The citrate molecules stabilizing SNPs prepared by the Lee and Meisel method are displaced by thiocholine molecules, to give positively charged particles. Upon mixing solutions of A and B, controlled aggregation occurs readily and rapidly in distilled water. 7923

dx.doi.org/10.1021/la5017033 | Langmuir 2014, 30, 7919−7927

Langmuir

Article

Figure 6. Kinetic measurement of nanoantenna self-assembly. (A) Time series of extinction spectra showing a trend of decreasing intensity in the peak corresponding to monomeric SNP, and a rise in the red part of the spectrum, corresponding to a heterogeneous population of aggregates. The arrow indicates the time that passed from reaction start. Each spectrum takes ∼30 s to measure. (B) The peak (∼440 nm) intensity was measured every 0.07 s (blue curve). The data were fit assuming second-order kinetics (inset, linear fit in red).

Figure 7. Examples of dimer nanoantennas. (A,B) SNP@SiO2 with SNP-ATC. (C) SNP-cit with SNP-ATC. In this case, there is no gap between the particles.

back focal plane of a high numerical aperture objective lens (the power density on the sample was 490 W/cm2). Ramanscattered photons were dispersed using a spectrograph (SpectraPro-150 with a 1200 g/mm grating, Acton), and the spectrum was imaged and collected on a Newton CCD camera (Andor). Dark-field images of the measured aggregates were also acquired, using a dark-field condenser mounted on the Raman setup. These images were used to track the position of dimers on the grid and later image them using TEM. Indeed, we verified the structure of each dimer in the electron microscope following the Raman measurements. Examples of SERS measurements from such antennas are shown in Figure 8. The intensity of Raman scattering from an antenna with 1.4 nm gap (red) was found to be ∼2 orders of magnitude larger than the intensity from an antenna with a 3.7 nm gap. The ultralow concentrations we used and typical fluctuations of the recorded spectra13 (not shown) strongly suggested that these measurements were made close to the single-molecule level.

for a short time after mixing the monomers encountered mainly other monomers to form dimers. The kinetics of such a reaction is second-order,18 so the change of reactant (monomer) concentration with time A(t) is given by 1 1 = kt + A (t ) A0

where k is a rate constant and A0 is the initial reactant concentration. Plotted in the inset of Figure 6B are the reciprocal values of the absorption, which is directly related to the monomer concentration. The data between 0 and 70 s yielded a good linear fit (R2 > 0.99, red line), with k = 0.504 ± 0.001 s−1. At 70 s, ∼45% of the monomers have already selfassembled into dimers. The loss of linearity at later times suggests that additional reactions (formation of larger aggregates) become significant. Following these results, we hypothesized that working with ultralow concentrations of particles and quenching the aggregation within the calculated time should be an efficient way to obtain mainly dimers. Indeed, it was found that by reducing the concentrations of the two particle types to ∼2.2 pM and transferring the particles within ∼1 min from mixing time onto a TEM grid we got at least 25% heterodimers, while multiple particles still remained as monomers and the rest were already involved in larger aggregates (Supporting Information Figure S3). Several images of dimers are shown in Figure 7. SERS Measurements. The dimers were deposited onto a TEM grid as discussed above and imaged using a home-built Raman microspectrometer equipped with a DPSS laser (Samba 25, Cobolt) operating at 532 nm, which was focused on the



DISCUSSION We presented here a bottom-up synthetic method for modular production of colloidal dimer nanoantennas, in which particles are first coated with layers of silica and are then self-assembled via electrostatic interactions. We showed that the gap size, as well as other geometrical parameters, can be controlled by varying the thickness of the silica layers using an optimized procedure. Our procedure enables synthesizing layers as thin as ∼1 nm readily and reproducibly in a couple of hours at room temperature, as has been demonstrated on 60 nm spheroidal silver particles. We further presented a method for creating 7924

dx.doi.org/10.1021/la5017033 | Langmuir 2014, 30, 7919−7927

Langmuir

Article

The ability to achieve dimer antennas with gap sizes between 0 and 1 nm is crucial for studying plasmonic coupling close to the quantum limit,44,45 a regime for which experimental data is still scarce. Achieving (sub)nanometer gaps is experimentally difficult, and so only a couple of papers reported measurements on such systems,8,46 where the expected nonmonotonic behavior of the near-field enhancement44,45,47 was recreated. In a future publication, we will show how our new dimer synthesis method allows us to obtain experimental data on SERS enhancement within this so-called quantum regime. It is interesting to discuss how the optical properties of SNP@SiO2 vary compared to those of bare particles and how this might affect the properties of the dimer antennas. It seems that the thin silica shells have a negligible effect on the optical properties of the particles, as the extinction spectra indicate (see Supporting Information Figure S2). However, it is reasonable to expect that addition of thicker layers would result in a shift of the plasmon resonance frequency,25,32 which is affected by the dielectric environment of the particles.48 On the other hand, since electromagnetic fields decay more slowly with distance in higher dielectric constant materials, a probe molecule in a silica-filled interparticle gap should be subject to a stronger field than in vacuum or air. SERS measurements from single molecules within nanoantennas with gaps of ∼4 nm are not trivial, since the enhancement is expected to drop by 2−3 orders of magnitude when the gap size is increased from ∼1 to ∼4 nm.49 This was observed previously in measurements of single-molecule SERS on single nanoantennas, where the signal was practically immeasurable for gap sizes larger than ∼3 nm.8 Calculations show that due to the high refractive index of silica the electromagnetic field in the gap is significantly higher than if the gap was in vacuo or filled with water. (Not shown. A full quantitative analysis of a series of SERS measurements on dimer nanoantennas of varying gaps will be reported separately.) The strong coupling between the gap size, plasmon wavelength, and electromagnetic enhancement highlights the potential of SNP@SiO2-based hot-spots, which stems from the flexibility in changing the gap size over ∼10 nm, allowing tuning to a wide range of wavelengths while maintaining sufficiently strong enhancement. We therefore conclude that the versatile method described above may open the way to improved and more systematic studies of complex heterodimer nanoantenna structures, the production of which was hitherto experimentally difficult. Our method is expected to eventually lead to better understanding of the nature of plasmonic coupling, and simplify the design of plasmon-based sensors.

Figure 8. (A) SERS spectra of TAMRA molecules within dimer nanoantennas. Spectra were taken from nanoantennas with a gap size of 1.4 nm (B, blue spectrum) and 3.7 nm (C, pink spectrum). The inset shows a magnified version of the 3.7 nm gap antenna spectrum. Each spectrum was measured over 3 s.

dimer nanoantennas through self-assembly in solution, by modifying the zeta potential of the involved particles. TAMRA molecules were attached to the nanoantenna surfaces, and used as probes to estimate the relative SERS enhancement of antennas with gap sizes of 1.4 and 3.7 nm. The dimer yield achieved in this work, ∼25%, might be improved in the future by further studying the aggregation kinetics. It is possible that combining asymmetric functionalization of particle surfaces42 or encapsulation of dimers in a polymer shell and further separation20 would further increase the final yield of dimers. The methodology presented in this paper has all three attributes proposed in the Introduction: (1) The silica shells can be readily formed at any thickness within the range relevant to plasmonic coupling, from 1 nm and up. Therefore, and together with the electrostatic method for dimer generation, the gap size can be varied essentially at will. This gap size is welldefined and varies only weakly between antennas. The position in the gap where analyte molecule would be attached can be varied by coupling particles with shells of different thicknesses. (2) It should be easy to adapt the silica deposition method for other metals or nanoparticle structures. Indeed, the antennas can be generated from two particles of different materials or sizes. (3) There is much flexibility in the choice of analyte molecules, as these are adsorbed on the particles independently of dimer synthesis. One interesting example of a structure whose synthesis should be greatly facilitated by this method is heterogeneous dimer nanoantennas consisting of two different metals.11 Such nanoantennas are of great interest since the symmetry breaking causes scattered electric fields at different wavelengths to acquire different phases, which may lead to spatial light filtering.12 Production of such heterodimers should be quite straightforward to realize,43 and the method presented here should facilitate full control over the gap size.



ASSOCIATED CONTENT

S Supporting Information *

Three figures: Evaluation of particle concentration, the effect of thin silica shells on SNP extinction, and a transmission electron micrograph of dimers. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. 7925

dx.doi.org/10.1021/la5017033 | Langmuir 2014, 30, 7919−7927

Langmuir

Article

(19) Šašić, S.; Itoh, T.; Ozaki, Y. Detailed analysis of single-molecule surface-enhanced resonance Raman scattering spectra of Rhodamine 6G obtained from isolated nano-aggregates of colloidal silver. J. Raman Spectrosc. 2005, 36, 593−599. (20) Chen, G.; Wang, Y.; Tan, L. H.; Yang, M.; Tan, L. S.; Chen, Y.; Chen, H. High-purity separation of gold nanoparticle dimers and trimers. J. Am. Chem. Soc. 2009, 131, 4218−4219. (21) Yang, L.; Wang, H.; Yan, B.; Reinhard, B. r. M. Calibration of silver plasmon rulers in the 1− 25 nm separation range: Experimental indications of distinct plasmon coupling regimes. J. Phys. Chem. C 2010, 114, 4901−4908. (22) Lim, D.-K.; Jeon, K.-S.; Kim, H. M.; Nam, J.-M.; Suh, Y. D. Nanogap-engineerable Raman-active nanodumbbells for single-molecule detection. Nat. Mater. 2009, 9, 60−67. (23) Loweth, C. J.; Caldwell, W. B.; Peng, X.; Alivisatos, A. P.; Schultz, P. G. DNA-based assembly of gold nanocrystals. Angew. Chem., Int. Ed. Engl. 1999, 38, 1808−1812. (24) Li, W.; Camargo, P. H.; Lu, X.; Xia, Y. Dimers of silver nanospheres: facile synthesis and their use as hot spots for surfaceenhanced Raman scattering. Nano Lett. 2008, 9, 485−490. (25) Li, J. F.; Huang, Y. F.; Ding, Y.; Yang, Z. L.; Li, S. B.; Zhou, X. S.; Fan, F. R.; Zhang, W.; Zhou, Z. Y.; Wu, D. Y.; Bin, R.; Wang, Z. L.; Tian, Z. Q. Shell-isolated nanoparticle-enhanced Raman spectroscopy. Nature 2010, 464, 392−395. (26) Liu, F.; Tang, C.; Zhan, P.; Chen, Z.; Ma, H.; Wang, Z. Released plasmonic electric field of ultrathin tetrahedral-amorphous-carbon films coated Ag nanoparticles for SERS. Sci. Rep. 2014, 4, 4494. (27) Zhang, X.; Zhao, J.; Whitney, A. V.; Elam, J. W.; Van Duyne, R. P. Ultrastable substrates for surface-enhanced Raman spectroscopy: Al2O3 overlayers fabricated by atomic layer deposition yield improved anthrax biomarker detection. J. Am. Chem. Soc. 2006, 128, 10304− 10309. (28) Tian, X. D.; Liu, B. J.; Li, J. F.; Yang, Z. L.; Ren, B.; Tian, Z. Q. SHINERS and plasmonic properties of Au core SiO2 shell nanoparticles with optimal core size and shell thickness. J. Raman Spectrosc. 2013, 44, 994−998. (29) Tyler, T. P.; Henry, A.-I.; Van Duyne, R. P.; Hersam, M. C. Improved monodispersity of plasmonic nanoantennas via centrifugal processing. J. Phys. Chem. Lett. 2011, 2, 218−222. (30) Wustholz, K. L.; Henry, A.-I.; McMahon, J. M.; Freeman, R. G.; Valley, N.; Piotti, M. E.; Natan, M. J.; Schatz, G. C.; Van Duyne, R. P. Structure− Activity Relationships in Gold Nanoparticle Dimers and Trimers for Surface-Enhanced Raman Spectroscopy. J. Am. Chem. Soc. 2010, 132, 10903−10910. (31) Kleinman, S. L.; Sharma, B.; Blaber, M. G.; Henry, A.-I.; Valley, N.; Freeman, R. G.; Natan, M. J.; Schatz, G. C.; Van Duyne, R. P. Structure enhancement factor relationships in single gold nanoantennas by surface-enhanced Raman excitation spectroscopy. J. Am. Chem. Soc. 2012, 135, 301−308. (32) Shanthil, M.; Thomas, R.; Swathi, R.; Thomas, K. Ag@ SiO2 Core−Shell Nanostructures: Distance-Dependent Plasmon Coupling and SERS Investigation. J. Phys. Chem. Lett. 2012, 3, 1459−1464. (33) Evanoff, D. D.; Chumanov, G. Size-controlled synthesis of nanoparticles. 1.“Silver-only” aqueous suspensions via hydrogen reduction. J. Phys. Chem. B 2004, 108, 13948−13956. (34) Lee, P.; Meisel, D. Adsorption and surface-enhanced Raman of dyes on silver and gold sols. J. Phys. Chem. 1982, 86, 3391−3395. (35) Ung, T.; Liz-Marzán, L. M.; Mulvaney, P. Controlled method for silica coating of silver colloids. Influence of coating on the rate of chemical reactions. Langmuir 1998, 14, 3740−3748. (36) Liz-Marzán, L. M.; Giersig, M.; Mulvaney, P. Synthesis of nanosized gold-silica core-shell particles. Langmuir 1996, 12, 4329− 4335. (37) Ruach-Nir, I.; Bendikov, T. A.; Doron-Mor, I.; Barkay, Z.; Vaskevich, A.; Rubinstein, I. Silica-stabilized gold island films for transmission localized surface plasmon sensing. J. Am. Chem. Soc. 2007, 129, 84−92.

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was made possible by a grant from the Israel Science Foundation (450/10), and was supported by the NaBi CNRS-Weizmann collaboration.



REFERENCES

(1) Willets, K. A.; Van Duyne, R. P. Localized surface plasmon resonance spectroscopy and sensing. Annu. Rev. Phys. Chem. 2007, 58, 267−297. (2) Novotny, L.; van Hulst, N. Antennas for light. Nat. Photonics 2011, 5, 83−90. (3) Le Ru, E.; Etchegoin, P. Phenomenological local field enhancement factor distributions around electromagnetic hot spots. J. Chem. Phys. 2009, 130, 181101. (4) Xu, H.; Aizpurua, J.; Käll, M.; Apell, P. Electromagnetic contributions to single-molecule sensitivity in surface-enhanced Raman scattering. Phys. Rev. E 2000, 62, 4318. (5) Nordlander, P.; Oubre, C.; Prodan, E.; Li, K.; Stockman, M. Plasmon hybridization in nanoparticle dimers. Nano Lett. 2004, 4, 899−903. (6) Brown, S. D.; Nativo, P.; Smith, J.-A.; Stirling, D.; Edwards, P. R.; Venugopal, B.; Flint, D. J.; Plumb, J. A.; Graham, D.; Wheate, N. J. Gold nanoparticles for the improved anticancer drug delivery of the active component of oxaliplatin. J. Am. Chem. Soc. 2010, 132, 4678− 4684. (7) Hao, E.; Schatz, G. C. Electromagnetic fields around silver nanoparticles and dimers. J. Chem. Phys. 2004, 120, 357. (8) Lee, J.-H.; Nam, J.-M.; Jeon, K.-S.; Lim, D.-K.; Kim, H.; Kwon, S.; Lee, H.; Suh, Y. D. Tuning and maximizing the single-molecule surface-enhanced Raman scattering from DNA-tethered nanodumbbells. ACS Nano 2012, 6, 9574−9584. (9) Ringler, M.; Schwemer, A.; Wunderlich, M.; Nichtl, A.; Kürzinger, K.; Klar, T.; Feldmann, J. Shaping emission spectra of fluorescent molecules with single plasmonic nanoresonators. Phys. Rev. Lett. 2008, 100, 203002. (10) Thacker, V. V.; Herrmann, L. O.; Sigle, D. O.; Zhang, T.; Liedl, T.; Baumberg, J. J.; Keyser, U. F. DNA origami based assembly of gold nanoparticle dimers for surface-enhanced Raman scattering. Nat. Commun. 2014, 5, 3448. (11) Sheikholeslami, S.; Jun, Y.-w.; Jain, P. K.; Alivisatos, A. P. Coupling of optical resonances in a compositionally asymmetric plasmonic nanoparticle dimer. Nano Lett. 2010, 10, 2655−2660. (12) Shegai, T.; Chen, S.; Miljković, V. D.; Zengin, G.; Johansson, P.; Käll, M. A bimetallic nanoantenna for directional colour routing. Nat. Commun. 2011, 2, 481. (13) Dadosh, T.; Sperling, J.; Bryant, G.; Breslow, R.; Shegai, T.; Dyshel, M.; Haran, G.; Bar-Joseph, I. Plasmonic control of the shape of the Raman spectrum of a single molecule in a silver nanoparticle dimer. ACS Nano 2009, 3, 1988−1994. (14) Chuntonov, L.; Haran, G. Optical activity in single-molecule surface-enhanced Raman scattering: Role of symmetry. MRS Bull. 2013, 38, 642−647. (15) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Single molecule detection using surfaceenhanced Raman scattering (SERS). Phys. Rev. Lett. 1997, 78, 1667. (16) Nie, S.; Emory, S. R. Probing single molecules and single nanoparticles by surface-enhanced Raman scattering. Science 1997, 275, 1102−1106. (17) Le Ru, E.; Meyer, M.; Etchegoin, P. Proof of single-molecule sensitivity in surface enhanced Raman scattering (SERS) by means of a two-analyte technique. J. Phys. Chem. B 2006, 110, 1944−1948. (18) Meyer, M.; Le Ru, E.; Etchegoin, P. Self-limiting aggregation leads to long-lived metastable clusters in colloidal solutions. J. Phys. Chem. B 2006, 110, 6040−6047. 7926

dx.doi.org/10.1021/la5017033 | Langmuir 2014, 30, 7919−7927

Langmuir

Article

(38) Iler, R. K. Product comprising a skin of dense, hydrated amorphous silica bound upon a core of another solid material and process of making same. U.S. Patent 2885366, 1959. (39) Boerio, F.; Armogan, L.; Cheng, S. The structure of γaminopropyltriethoxysilane films on iron mirrors. J. Colloid Interface Sci. 1980, 73, 416−424. (40) Iler, R. Effect of adsorbed alumina on the solubility of amorphous silica in water. J. Colloid Interface Sci. 1973, 43, 399−408. (41) Westcott, S. L.; Oldenburg, S. J.; Lee, T. R.; Halas, N. J. Formation and adsorption of clusters of gold nanoparticles onto functionalized silica nanoparticle surfaces. Langmuir 1998, 14, 5396− 5401. (42) Romo-Herrera, J. M.; Alvarez-Puebla, R. A.; Liz-Marzán, L. M. Controlled assembly of plasmonic colloidal nanoparticle clusters. Nanoscale 2011, 3, 1304−1315. (43) Lombardi, A.; Grzelczak, M. P.; Crut, A.; Maioli, P.; PastorizaSantos, I.; Liz-Marzán, L. M.; Del Fatti, N.; Vallée, F. Optical response of individual Au−Ag@ SiO2 heterodimers. ACS Nano 2013, 7, 2522− 2531. (44) Esteban, R.; Borisov, A. G.; Nordlander, P.; Aizpurua, J. Bridging quantum and classical plasmonics with a quantum-corrected model. Nat. Commun. 2012, 3, 825. (45) Marinica, D.; Kazansky, A.; Nordlander, P.; Aizpurua, J.; Borisov, A. G. Quantum plasmonics: nonlinear effects in the field enhancement of a plasmonic nanoparticle dimer. Nano Lett. 2012, 12, 1333−1339. (46) Savage, K. J.; Hawkeye, M. M.; Esteban, R.; Borisov, A. G.; Aizpurua, J.; Baumberg, J. J. Revealing the quantum regime in tunnelling plasmonics. Nature 2012, 491, 574−577. (47) Zuloaga, J.; Prodan, E.; Nordlander, P. Quantum description of the plasmon resonances of a nanoparticle dimer. Nano Lett. 2009, 9, 887−891. (48) Kreibig, U.; Vollmer, M. Optical Properties of Metal Clusters; Springer: Berlin, Heidelberg, New York, 1995; Vol. 25, p 535. (49) Romero, I.; Aizpurua, J.; Bryant, G. W.; García De Abajo, F. J. Plasmons in nearly touching metallic nanoparticles: singular response in the limit of touching dimers. Opt. Express 2006, 14, 9988−9999.

7927

dx.doi.org/10.1021/la5017033 | Langmuir 2014, 30, 7919−7927