One-step synthesis and applications of highly concentrated silver nanoparticles with an ultra-thin silica shell

Abstract

A one-step method for synthesizing high concentrations of 40–400 nm silver nanoparticles coated with an ultra-thin silica shell is described. The shell thickness can be further increased via post-synthetic precipitation of sodium silicate. The nanoparticles exhibit remarkable stability and can be concentrated to as much as 50% by weight in water and some organic solvents. It was also demonstrated that silica species catalyze the seeding and growth of silver nanoparticles. The herein described particles were used to develop simple surface-enhanced Raman spectroscopy substrates and chemiresistors.

---

Introduction

Metallic nanoparticles, and in particular silver nanoparticles (AgNPs), have attracted significant interest in a large variety of applications¹ including catalysis,^2,3 optics,⁴ sensing^5,6 and therapeutics.⁷ As a result, a large spectrum of synthetic methods have emerged, resulting in colloidal AgNPs with a variety of morphologies.⁸ These synthetic methods, that generally involve the reduction of silver salts in the presence of capping or stabilizing agents for the control of particle size and stability, typically give low AgNP concentrations (less than 1 mM)^9,10 and therefore are not suitable for large-scale production. Few examples exist in the literature of higher concentrations being achieved, all of which require the addition of stabilizing ligands and generally use organic solvents.^11–13

Ultra-thin silica shells (UTSS) are an effective way to improve the stability and compatibility of metallic nanoparticles without significantly attenuating their optical properties. This is of particular importance to plasmonic particles used in sensing applications such as localized surface plasmon resonance (LSPR) or surface-enhanced Raman spectroscopy (SERS).¹⁴ For AgNPs an UTSS can hinder the oxidation of Ag in both air and water and provide a scaffold for functionalization through well-established silane chemistry. In addition, the silica shell inhibits the dissolution of silver that produces potentially harmful silver ions in bio-applications. Therefore the development of a simple and reliable method to produce AgNPs with an UTSS will improve upon numerous already existing applications as well as stimulate new applications, specifically in the area of optical labeling and plasmonics.

To date, the vast majority of methods for producing silica layers around metallic nanoparticles are derived from the Stöber¹⁵ method that uses alkoxysilanes, most commonly tetraethyl orthosilicate (TEOS), as silica precursors. Condensation and nucleation reactions form a silica shell around the metallic core under the catalytic action of ammonia causing the hydrolysis of alkoxysilanes.^16–18 Another less common approach introduced by Mulvaney et al., demonstrated that gold and silver nanoparticles could be made vitreophilic using a silane coupling agent such as (3-aminopropyl)-trimethoxysilane (APTMS), which generates a monolayer on the surface of the particles. This layer can be further grown by the addition of sodium silicate and controlled precipitation by the addition of ethanol.^19,20 One of the issues with using sodium silicate as a precursor is that it results in suspensions with high ionic strength thereby leading to particle aggregation.²¹ Other methods to produce silver nanoparticles, such as those described by Lee and Meisel²² only add to the ionic strength of the resulting solution. Aggregation increases with the particle concentration making high ionic strength unacceptable when high concentrations of nanoparticles are required. Haran and Zohar were able to avoid the excess of sodium ions during the synthesis of AgNPs by using the hydrogen reduction method developed in our laboratory.²³ In a multistep process, these particles were then coated with APTMS followed by silica deposition using sodium silicate and ethanol.²⁴

Here, we report a one-step method for synthesizing 40–400 nm AgNPs with an UTSS. This method allows for the production of fairly monodisperse spherical AgNPs in concentrations, for example, up to 4.8 × 10¹⁴ AgNP (0.3% of Ag by weight) per litre for 100 nm AgNPs with 1.6 nm silica shell. The shell can be made with thickness values ranging from 1 nm to 5 nm depending on initial silica concentration and reaction time. This silica shell greatly improves the particle stability allowing for further concentration of the particles in water to 50% Ag by weight. The AgNPs exhibited shelf stability of >3 years in typical laboratory conditions under ambient light. To the best of our knowledge this is the first example of silica coated AgNPs synthesized in a one-step reaction resulting in highly concentrated, stable AgNPs without a stabilizing ligand. Very high concentrations of particles with a ultra-thin silica shell were used to realize simple and reproducible substrates for surface-enhanced Raman scattering (SERS) spectroscopy and for chemiresistors with rapid response.

Results and discussion

It was originally observed that the glass surface of the reaction vessel plays a major role in the AgNP synthesis during the reaction of hydrogen gas with silver(I) oxide.²³ A mechanism was implicated suggesting a possible catalytic role of a glass surface on the formation of initial silver seed on the surface followed by their release and further growth in the solution. This was further corroborated by comparing reaction results obtained in Pyrex®, quartz, and fluorinated-silane coated vessels.²³ A typical growth in a Pyrex® vessel results in mostly spherical AgNPs with an optical density (O.D.) of ∼1.5 for 100 nm AgNPs. When the reaction was performed under the same conditions in a high-purity quartz vessel the obtained O.D. was at least an order of magnitude lower, though the particle size and shape distribution was similar to those in the Pyrex® vessel. Using a Pyrex® vessel coated with a cross-linked fluorosilane, a particle O.D. of ∼0.4 was obtained under the same conditions and the particles were primarily square platelets.²³ These initial observations prompted more detailed studies with the goal of further understanding the role of the glass surface (silica) during the hydrogen reduction of silver(I) oxide and ultimately developing an efficient, fully controllable method for synthesis of high quality stable AgNPs.

A Teflon® reactor was used to eliminate the effect of glass walls on the reduction reaction. In this case, reduction of neither silver(I) oxide nor soluble silver salts was observed by the addition of hydrogen gas (Fig. 1A). This was a somewhat surprising result considering the fact that the standard redox potential of silver is highly positive (0.79 V) against the standard hydrogen electrode. Even though further studies are required to fully understand this behavior, the available literature suggests the preferential formation of silver hydride, over the reduction to silver metal.^25–27 In agreement with these results, Fig. 1C shows that the addition of silica species, either in the form of fumed silica or sodium silicate, to the Teflon® reactor containing 10 mM of Ag₂SO₄ and hydrogen gas at 73 °C effectively produced a nanoparticle suspension with an O.D. of 13. The results support the hypothesis that silica species could be catalyzing the reduction of silver and the subsequent formation of AgNPs. The reaction without silica produced particles with an O.D. of 0.08, most likely due to uncontrolled impurities capable of catalyzing the reduction. On the other hand, the addition of sodium silicate without hydrogen lead to the formation of insoluble silver silicate that was further reduced to silver metal by hydrogen gas, as evident from UV-Vis spectra, Fig. 1B. In the case of a Pyrex® reaction vessel, a small amount of silica dissolved from the walls forms silver silicate that can be then reduced to silver metal. It is also important to point out that the pH value of saturated solutions of silver oxide in water at 73 °C is ∼10, a value that facilitates the dissolution of the glass surface, yielding the various silica species in the solution.^28,29

Fig. 1 UV-Vis spectra of (A) 10 mM Ag₂SO₄ under H₂ for 2 hours (B) following addition of 3 mM Na₂SiO₃ (C) after reaction placed back under H₂ for 1 hour (10× dilution). All steps were performed in a Teflon® reactor.

To build upon the observation that silica species catalyze the formation of silver seeds and particle growth, the effect of both fumed silica and sodium silicate ranging in concentrations between 0.1 and 6 mM were studied. A quartz vessel was used so that the contribution from the dissolution of glass walls would be minimized, because quartz has a significantly lower solubility than Pyrex®.³⁰ Since particles with optical densities 10–100× greater than that achieved in a Pyrex® vessel were obtained, it was immediately evident that the addition of the silica species greatly increased the concentration of resulting AgNPs. It was also observed that the final concentration of AgNPs depended on the reaction time and concentration of silica species.

When a Pyrex® vessel was used without the addition of silica, approximately 5% of the AgNPs were in the form of long rods and flat platelets. This result can be attributed to the presence of a low concentration of silica, dissolved from the walls of the Pyrex® vessel. The silica species in water form insoluble silver silicate that can be reduced by hydrogen gas. In addition to the dissolved silica in solution, there is also reactive silica on the glass surface of the reaction vessel.³¹ This surface silica binds silver ions forming surface silver silicate that is also reduced during the hydrogen reaction. As a result, the walls of the vessel quickly become coated with an opaque silver film. The film consisted of particles of different sizes and shapes including rods and platelets that grow at the glass surface/silica/silver interface and break off into the suspension.

Only polyhedral AgNPs (without rods and platelets) were observed when sodium silicate or fumed silica was added to the reaction in a quartz vessel (Fig. S1†). In this case, the concentration of silica was sufficient to form large silica hydrogel particles. In addition, the surface of the quartz vessel has less reactive silica to bind silver and to form rods & platelets. The walls of the quartz vessel remain clear during the reduction reaction and only acquired a light yellow tint after long reaction times whereas the Pyrex® vessel quickly formed a dense silver film. The AgNPs grown upon the addition of silica were very uniform in size (114.53 ± 4 nm) and have polyhedral shape implying a high degree of crystallinity (Fig. 2). These particles were further concentrated via centrifugation to achieve a solution with an O.D. of 2500 corresponding to 10% Ag by weight (Fig. 2).

Fig. 2 Size distribution analysis of 50 different AgNPs in the accompanying SEM image. Optical image of AgNPs after concentrating to 10% Ag by weight and allowing to gravity precipitate.

The silica matrix serves a dual purpose; a capping agent encapsulating the AgNPs and a catalyst for the silver reduction. Common methods to produce nanoparticles incorporate capping agents such as surfactants, polymers and ligands that typically result in spherical particles.^32–34 Considering that the formation of silver hydride hinders the reduction of silver ions in aqueous solutions by hydrogen gas as was mentioned above, the catalytic role of silica could be simply to prevent the hydride formation by forming insoluble silver silicate. Even though the redox potential of silicate was measured to be approximately 0.37 V vs. SHE, a value that is less positive than that of Ag⁺ (0.79 V vs. SHE), the reduction reaction could still proceed.

Based on the model that silica species catalyze the reduction of silver and growth of AgNPs, it was assumed that a very thin silica shell may exist on the particles even when grown without the addition of silica. To investigate this assumption, AgNPs were grown in a Pyrex® vessel without adding silica to the size of 100 nm and high resolution TEM was undertaken. Indeed, as can be seen in Fig. 3, a very thin, ∼1.1 nm shell was present around the particles. This silica shell imparts stability to AgNPs in aqueous suspensions.

Fig. 3 HR-TEM images of AgNPs grown to 100 nm in Pyrex® without additional silica.

Sodium silicate or fumed silica was added to a quartz reaction vessel in concentrations ranging from 0.1 to 6 mM to determine the effect of different silica species on the growth of the AgNPs. Both species behaved similarly in that increasing the silica concentration resulted in increased concentration of AgNPs. At the same time, a longer reaction time is required to reach a desired particle size. Sodium silicate is more practical because it is easier to control its concentration as compared to fumed silica that is difficult to measure and uniformly disperse in solution. Despite this, fumed silica would be advantageous if sodium is of concern for final applications.

Separate reactions were carried out with 0.8 mM sodium silicate and 1 mM fumed silica to study the AgNPs growth over time. It was previously observed that these concentrations produced the same ∼1–2 nm silica shell on 100 nm AgNPs. As can be seen in Fig. 4, it appears that the shell thickness decreased as the particles grow larger; e.g. an initial shell of ∼5 nm on 50 nm particles decreased to ∼1.5 nm as the particles approached 100 nm. As the nanoparticles size and surface area increases the shell thickness decreases because the same amount of silica is available for covering a larger surface area.

Fig. 4 TEM of AgNPs grown with 0.8 mM Na₂SiO₃ taken at 30 minutes (A) and 4 hours (B) & AgNPs grown with 1 mM fumed silica taken at 30 minutes (C) and 4 hours (D).

While the particle size determines the shell thickness, the initial concentration of silica does not. The initial concentration of silica determines the concentration of AgNPs. In other words, growing AgNPs to the same size with two different concentrations of silica results in the same thickness of shell; the higher concentration of silica produces more AgNPs. To demonstrate this behavior, 50 nm AgNP were grown in two separate reactions using 0.8 mM and 6 mM sodium silicate. The shell thickness was ∼5 nm for both reactions, however it took 30 minutes to achieve 50 nm AgNPs with a final O.D. of 2.7 using 0.8 mM sodium silicate. When the concentration of sodium silicate was increases to 6 mM it took 5 hours to achieve the same size AgNPs with O.D. of 84 (Fig. S2†). Increasing silica concentration results in a larger number of seeds; however since the reaction kinetics are limited by the dissolution of silver(I) oxide (approximately 50 mg L⁻¹ at 73 °C) the time to achieve the same size particles increased with increasing the silica concentration.

The addition of silica species to the reaction also allows the synthesis of very large AgNPs. Particles were grown to ∼250 nm and ∼400 nm with the addition of 1.27 mM and 0.5 mM fumed silica, respectively. The final O.D. for the ∼250 nm AgNPs was 48 while the ∼400 nm AgNPs achieved an O.D. of 36 (Fig. S3†). The silica shell thickness was less than 1 nm in both cases.

When growing AgNPs with the addition of silica, the silica shell thickness ranged from <1 nm to ∼5 nm depending on the size of particle desired. If required, this thickness can be further increased. Since the AgNPs already have a layer of silica there is no need to add a silane coupling agent. A suspension of 100 nm AgNPs with ∼1.5 nm shell was diluted to an O.D. of 3 with the addition of a sodium silicate solution of the concentrations varying between 0 and 2 mM. UV-Vis measurements revealed no noticeable change in the plasmon dipole position indicating no changes of the shell thickness. Otherwise, a red spectral shift would be expected due to increasing the local refractive index with increasing the shell thickness.³⁵ The suspensions were then added to ethanol in a 1 : 4 volume ratio²⁰ to precipitate the silica onto the AgNPs and the UV-Vis was taken again. A red shift of the plasmon resonance resulted from the condensation of the additional silica on the particle surface as was further confirmed with TEM (Fig. 5). No formation of free silica particles was seen. The addition of 1 mM and 2 mM of sodium silicate increased the silica shell around AgNPs to ∼5 nm and ∼15 nm, respectively. This approach allows complete control of the silica shell thickness.

Fig. 5 UV-Vis spectra and TEM of AgNPs with ∼1.5 nm Si shell after introduction of 0 mM (A) 1 mM (B) & 2 mM (C) Na₂SiO₃ and addition to EtOH.

Sulfide, due to its high affinity to silver, was used to test the ability of the silica shell to protect the silver surface. Sodium sulfide (3 mM) was added to suspensions containing AgNPs with silica shell of different thicknesses and UV-Vis spectra of the suspensions were taken over the course of 20 hours. As can be seen in Fig. 6, the resonance of the AgNPs with a ∼1.5 nm silica shell quickly degraded and was completely lost by the 20 hour mark, at which point the spectrum represented that of Ag₂S nanoparticles.³⁶ The trend continued, although at a slower rate, as the shell thickness was increased. The optical density decreased by only 20% for ∼15 nm silica shell while the original shape of the plasmon resonance was retained. This decrease was likely due to the presence of a small fraction of particles with incomplete silica coverage, because as this solution was monitored over a period of one month with continuous exposure to the sulfide solution no further changes were observed. The 15 nm silica shell completely protected AgNPs from reacting with sulfide.

Fig. 6 UV-Vis spectra of AgNPs with ∼1.5 nm Si shell (A) and ∼15 nm Si shell (B) after exposure to 3 mM Na₂S over a period of 20 hours.

Optical densities of ∼100, corresponding to ∼4 × 10¹³ AgNP per litre (0.3% Ag by weight) can be easily achieved when synthesizing the AgNPs using this method. The particles can be further concentrated without aggregation because of their excellent stability, imparted by the thin silica shell. For example, 100 nm AgNPs were grown in a 5 L quartz vessel with the addition of 1 mM fumed silica and when centrifuged down to ∼4 mL, resulted in the final concentration of ∼1.98 × 10¹⁶ AgNP per liter (43% Ag by weight) corresponding to an O.D. of ∼11 500. While over time a large fraction of these AgNPs precipitated on the bottom of the container yielding a silver mirror, they were easily re-dispersed via gentle agitation and, in this way, were shelf stable in aqueous media for more than 3 years. The particles have also exhibited comparable stability when transferred to organic solvents, such as ethanol and 2-propanol.

The AgNPs do not produce strong SERS when used as colloids because the silica shell hinders the particles from aggregating, an important SERS prerequisite. However, when assembled into 2D arrays, an efficient SERS substrate was obtained. The ability to synthesize stable AgNP in high concentrations and assemble them into 2D arrays provides the prospect for various SERS applications.^37–41 The AgNPs were readily self-assembled onto poly(4-vinylpyridine) modified glass substrate yielding high density monolayer films of well-separated particles. The interparticle distance can be effectively controlled by adjusting the ionic strength of the AgNP suspension. These films exhibited strong SERS activity as was demonstrated using three analytes of practical importance; adenine, guanine and single strand DNA (ssDNA). The films also provided a unique opportunity for observing the unobstructed effect of the particle aggregation on the SERS signal. The films were exposed to the analyte solutions, thoroughly rinsed and SERS was measured from the films submersed in water. Next, the same films were dried, and SERS was measured again in pure water. The drying induced surface aggregation of the AgNPs, but the amount of the analyte and the number of NPs interrogated by the laser beam remained the same.

The aggregation of AgNPs was monitored by UV-Vis spectroscopy. The addition of guanine slightly decreased the intensity of the initially sharp peak that resulted from a cooperative plasmon mode that is characterized by coherent electron oscillations between adjacent particles (Fig. 7).⁴² This decrease pointed out to a weaker plasmon coupling mostly likely due to guanine induced rearrangement of the particles on the surface. Drying the films induced the surface aggregation of the AgNPs (Fig. 7, insert) resulting in the further decrease of the coupled peak intensity and the appearance of the broad peak around 750 nm due to plasmon modes of the aggregates (Fig. 7).

Fig. 7 UV-Vis of coupled 2D array (black) overnight exposure to 10 μM guanine in water before (blue) and after (red) drying. Insert: SEM image of aggregated film.

The aggregation had a dramatic effect on the SERS from guanine and adenine in that the signal increased by 3.5 and 15 fold, respectively, when compared to that from the pre-aggregated state (Fig. 8). The smaller relative enhancement of guanine SERS was attributed to its lower solubility in aqueous solution. SERS spectra was also obtained from ssDNA after the films were exposed overnight to a 1 μM solution. As in the case of the nucleotides, the nanoparticle aggregation upon drying increased the SERS signal of ssDNA by more than an order of magnitude (Fig. 8). The strongest Raman peaks at 733 cm⁻¹, 662 cm⁻¹ and 798 cm⁻¹ closely match those for adenine, guanine and cytosine, respectively.^43,44 The peak at 1161 cm⁻¹ was previously attributed to the oxygen enhanced photon driven transfer of an electron from a metal state to an unoccupied molecular orbital of the adsorbate.^45,46

Fig. 8 SERS of 2D arrays of AgNPs exposed to 10 μM guanine (A), 10 μM adenine (B), and 1 μM ssDNA (C) in water (1) then dried to induce aggregation and back in water (2).

The additional Raman enhancement resulted from the aggregation of AgNPs signifies the electromagnetic enhancement mechanism, in which the electromagnetic field is concentrated in the spaces between the aggregated particles where the analyte molecules are sandwiched.⁴⁷ The real enhancement due to the aggregation is most likely larger than that reported in Fig. 8 because not all the analyte molecules that are covering the individual AgNPs ended up in the spaces between the particles. These SERS active films can be readily fabricated within minutes using the simple self-assembly technique of highly concentrated AgNPs. Together with the drying induced surface aggregation, the films present a new approach to SERS.

Another application that takes advantage of high concentrations of AgNPs with ultra-thin silica shells relates to chemiresistor vapor sensors. The high concentration of AgNPs allows the fabrication of high density films by a simple drop-casting and drying procedure. The ultra-thin silica shell provides a semi-insulating layer such that the overall electrical conductivity of the nanocomposite films can be controlled between the fully conducting metallic and fully non-conducting states by changing the thickness of the shell. In addition, the silica shell affords the sensing medium by either changes of its intrinsic properties or via chemical modifications, imparting the sensitivity to analytes of interest.

The high density nanocomposite films were drop-casted from a 2500 O.D. (10% Ag by weight) AgNP suspension with 1–2 nm silica shell and dried between two ITO electrodes spaced ∼1 mm apart yielding the measured resistance of several hundred ohms after annealing for 1 hour at 100 °C. Electron microscopy revealed a 3D network of tightly packed AgNPs (Fig. 9) unlike the 2D arrays obtained via the self-assembly process employed for fabricating the SERS-active substrates. To demonstrate the concept that this 3D network can function as a chemiresistor, the electrical conductivity was measured across the nanocomposite films while being exposed to water or acetone vapor. The current was monitored in time at 1 V potential difference as the source of vapor was brought in and out.

Fig. 9 SEM of nanocomposite film used to fabricate chemiresistor.

As can be seen in Fig. 10, a reversible and reproducible response was observed for both water and acetone vapors on top of the steady state decline of the overall current with time. It is hypothesized that the decline resulted from voltage induced changes in the shell and/or oxidation of the silver surface. It was noted that the water and acetone vapor responses were out-of-phase in that water vapor decreased and the acetone vapor increased the conductivity, respectively. In the case of acetone vapor sensing, it was shown that sorption of oxygen plays a vital role in the electrical transport properties in that the reduction of the oxygen species by the acetone vapor would generate electrons and increase conductivity.^48,49 The conductivity decrease from the water vapor exposure can be attributed to the swelling of the silica layer as it adsorbs water. It was also noted that, in the case of either water or acetone, the time response was about 10 seconds, significantly faster than hundreds of seconds typically reported for chemiresistors.^49–51 The sensors also exhibited measurable response, albeit with a lower signal-to-noise ratio, when the applied potential difference was as low as 0.05 V. This property opens a possibility for operating the sensors in liquid environments because fewer electrochemical reactions are expected at such low potential differences.

Fig. 10 Conductivity response for air, water vapor & acetone vapor over a period of 500 seconds at a potential of +1 V. Water & acetone were introduce on (grey) and off (white) for 50 seconds starting at 50 seconds to show reversible response.

The described here chemiresistor offers several advantages such as easy fabrication via drop-casting, control over the electrical conductivity by varying the thickness of the shell, fast response and the versatility or chemical modifications that can be done to the silica scaffold. In fact, the silica semi-insulating layer on AgNPs reported here is the first example of such layer used in chemiresistors.

Conclusion

A one-step synthesis of highly concentrated silver nanoparticles ranging in size from 40 to 400 nm with an ultra-thin silica shell was developed. It was concluded that silica species catalyze the formation of metal seeds and subsequent growth of silver nanoparticles. The silica shell, which can be further grown by the addition of sodium silicate and ethanol condensation, imparts excellent stability to suspensions with ultra-high nanoparticles concentrations as well as hinders the chemical accessibility of the metal surface. The nanoparticles can be self-assembled on chemically modified surfaces and the films exhibit SERS activity that can be further enhanced by the surface aggregation of the nanoparticles via drying followed by rewetting. The high concentration of the nanoparticles with ultra-thin silica shells also enabled the preparation of a simple chemiresistor with a unique matrix and a rapid response rate, in which the conductivity through the nanocomposite film can be made reversibly responsive to different analytes.

The described results also have broader implications because they highlight the importance of considering the potential effects of small concentrations of silica dissolved from glass reaction vessels on the outcome of chemical reactions, especially reactions that are carried out at elevated pH. It is also important to recognize that dissolved silica reacts with the dissolved silver to form silver silicate that can act as centers for condensation and seed formations when synthesizing various nanoparticles in addition to the catalytic effects as was demonstrated here.

Experimental

Materials

Deionized water with a nominal resistivity of 18 MΩ cm was obtained from a Millipore Milli-Q water purification system. Silver(I) oxide (99.99%) and anhydrous sodium sulfate (99.99%) were acquired from Alfa Aesar. Poly(4-vinylpyridine) (PVP), ACS grade sodium sulfide nonahydrate, adenine (99%), guanine (98%) and zine powder (ACS grade) were purchased from Sigma-Aldrich. USP grade absolute 200 proof ethanol was obtained from Aaper Alcohol & Chemical Co. Sodium metasilicate (SiO₂ 44–47%) and fumed silica (99.8%) were purchased from Sigma-Aldrich and purified by heating at 500 °C for 5 hours under vacuum. Selex Reverse Primer (5′ TCA AGT GGT CAT GTA CTA GTC AA 3′) was obtained from Dr Christensen at Clemson University. Ultra-high-purity hydrogen and ultra-high-purity nitrogen were purchased from Air Gas. Indium tin oxide (ITO) glass (8–12 Ω sq⁻¹ inch) was received from Sigma and Delta Technologies, LTD. Unless specified, all reagents and solvents were used as received.

Instrumentation

UV-vis spectra were recorded using a Shimadzu UV-2501 PC spectrometer. Scanning electron microscopy images were taken using a Hitachi SEM-4800 and high resolution transmission electron microscopy images were obtained with a Hitachi TEM-H9500. Slides for SEM imaging were prepared by roll coating ITO slides in a 0.25% PVP in ethanol solution for 4 hours followed by roll coating in AgNP solutions with 1 mM sodium sulfate to increase packing density for 12 hours. Raman spectra was measured directly from the samples using a spectrograph (SPEX, Triplemate 1377) interfaced to a thermoelectrically cooled CCD detector (Andor Technology, Model DU420A-BV) operating at −60 °C. The spectra were excited with 514.5 nm radiation from an Innova 100 (Coherent) argon ion laser. The laser power was between 5–25 mW at the sample with a total acquisition time of 100 seconds for each measurement. The scattered light was collected by a f/1.2 camera lens in a backscattering geometry, and the instrument was calibrated using an indene and chloroform/bromoform standard. The current–time curves were measured by using an electrochemical workstation (CH Instruments, CHI440) with chronoamperometry. All Raman and UV-Vis spectra and figures prepared with Spectra-Solve for Windows software (LasTek Pty. Ltd.).

Synthesis of AgNP with UTSS

AgNP with UTSS synthesis was derived from the hydrogen reduction method, previously developed in this lab.²³ Sodium metasilicate or fumed silica was added to 250 mL of DI water in a 500 mL quartz round-bottom flask to obtain concentrations between 0.1 to 6 mM and then sonicated for 30 seconds to disperse. Following this, 0.25 to 2 grams of silver(I) oxide was added to the solution and hand shaken for 1 minute. The flask was then connected to a condenser and hydrogen line followed by heating to 73 °C with stirring under ambient conditions. Once the temperature has stabilized, the vessel was flushed with hydrogen and pressurized to 10 psi hydrogen to initiate the reaction. Typical reactions would range between 3–20 hours depending on the desired thickness of shell, concentration and size of the particles. AgNPs were washed to remove excess silica using DI water saturated with silver(I) oxide for particle stability and could be further concentrated by centrifugation.

Post-synthesis shell growth and stability study

AgNP obtained using the above method with an ∼1.6 nm silica shell were diluted from an optical density of 27.31 with varying concentrations of sodium silicate in water to achieve a solution with an optical density of 2 and a concentration of sodium silicate between 0 and 2 mM. To promote condensation of the silica onto the particle surface, 1 mL of each solution was pipetted into 3 mL of ethanol. UV-Vis was taken revealing a red shift of the dipole that increased for increasing silica concentration. UV-Vis spectra was taken before and after each step. To test the permeability of the silica shell the particles were exposed to 3 mM sodium sulfide and UV-Vis was taken in regular intervals to monitor the plasmon resonance.

Preparation of SERS substrates

Standard microscope slides were cut into 11 mm × 25 mm sized pieces and cleaned by sonication in acetone, ethanol and water baths for 15 minutes each followed by drying with nitrogen gas and plasma treating for 10 minutes. After cleaning, the slides were placed into a 0.1% PVP solution and allowed to roll coat for 4 hours. Following PVP exposure, ethanol and water rinses were performed after which the slides were dried with nitrogen gas and annealed at 120 °C for 3 hours. Next, slides were quickly cooled with a stream of nitrogen gas and rolled overnight in aqueous suspensions of ∼100 nm AgNPs with an optical density (O.D.) of 3 containing 1.5 mM sodium sulfate. After the nanoparticle adsorption, the slides were rinsed with water and one side of the slide was stripped of AgNPs using a dilute nitric acid solution. The slides were thoroughly rinsed and stored in deionized water. Slides were then exposed to aqueous solutions of adenine, guanine and ssDNA overnight. Slides were then rinsed with DI water to remove unabsorbed analyte and placed in fresh DI water in clean 1 dram vials. Raman spectra was collect on this system after which they were dried at 100 °C for 2–3 minutes before being placed back in the DI water. SERS spectra was then collected again to see the effects of aggregation on the system.

Preparation of chemiresistor vapor sensor

AgNPs were grown using the above method but in a 5 L quartz reaction vessel instead of the usual 500 mL vessel. AgNPs were then concentrated by centrifugation (∼1000 RPM or 175 G's) for 30 minutes. This was repeated until a volume of ∼2 mL of AgNPs was obtained with an optical density of ∼2500. ITO glass was cut into 1 × 2.5 cm strips. Scotch tape was used to mask all but a ∼1–2 mm strip down the middle of the glass slide. This strip was then coated with zinc powder and exposed to 2 M HCl for 15 minutes to etch the ITO followed by rinsing with DI water and removal of the mask. A multi-meter was used across this strip to confirm the ITO layer had been removed. 2.5 μL of the concentrated AgNPs were then pipetted onto this strip and allowed to dry for 1–2 minutes in the 100 °C. Current–voltage curves were performed across the device afterwards and a large hysteresis effect was observed. The slides were then annealed at 100 °C for 24 hours, current–voltage curves were then taken again and showed a linear relationship. Data was collected by positioning the sensor ∼2 mm above a 25 Erlenmeyer flask. Data was collect for 50 seconds before positioning the flask containing the various analytes under the sensor for 50 seconds followed by removal to show a reversible response. This was continued for a period of 500 seconds for each analyte. Data was collect at a potential of 1 V & −0.05 V with a sampling interval of 0.0315 seconds and a sensitivity of 1 × 10⁻³ and 5 × 10⁻⁵, respectively.

Acknowledgements

This research was supported by the United States Department of Energy, grant No. DE-FG02-06ER46342. We would also like to acknowledge the Clemson University Center for Optical Materials Science and Engineering Technologies and Yi Jin for his assistance in etching the ITO and instruction in the use of the electrochemical workstation.

References

1. M. Ahamed, M. S. AlSalhi and M. K. J. Siddiqui, Clin. Chim. Acta, 2010, 411, 1841–1848.
2. S. Sarina, E. R. Waclawik and H. Zhu, Green Chem., 2013, 15, 1814–1833.
3. B. R. Cuenya, Thin Solid Films, 2010, 518, 3127–3150.
4. D. D. Evanoff Jr and G. Chumanov, ChemPhysChem, 2005, 6, 1221–1231.
5. J. Langer, S. M. Novikov and L. M. Liz-Marzan, Nanotechnology, 2015, 26, 1–28.
6. A. J. Haes and R. P. Van Duyne, Anal. Bioanal. Chem., 2004, 379, 920–930.
7. L. Wei, J. Lu, H. Xu, A. Patel, Z.-S. Chen and G. Chen, Drug Discovery Today, 2015, 20, 595–601.
8. C. M. Cobley, S. E. Skrabalak, D. J. Campbell and Y. Xia, Plasmonics, 2009, 4, 171–179.
9. T. M. Tolaymat, A. M. El Badawy, A. Genaidy, K. G. Scheckel, T. P. Luxton and M. Suidan, Sci. Total Environ., 2010, 408, 999–1006.
10. M. Oćwieja, Z. Adamczyk, M. Morga and K. Kubiak, Adv. Colloid Interface Sci., 2015, 222, 530–563.
11. A. Shahzad, M. Chung, T. Yu and W.-S. Kim, Chem.–Asian J., 2015, 10, 2512–2517.
12. A. Shahzad, W.-S. Kim and T. Yu, RSC Adv., 2015, 5, 28652–28661.
13. I. Sondi, D. V. Goia and E. Matijevic, J. Colloid Interface Sci., 2003, 260, 75–81.
14. M. Shanthil, R. Thomas, R. S. Swathi and K. G. Thomas, J. Phys. Chem. Lett., 2012, 3, 1459–1464.
15. W. Stöber, A. Fink and E. Bohn, J. Colloid Interface Sci., 1968, 26, 62–69.
16. V. V. Hardikar and E. Matijevic, J. Colloid Interface Sci., 2000, 221, 133–136.
17. Y. Kobayashi, H. Katakami, E. Mine, D. Nagao, M. Konno and L. M. Liz-Marzan, J. Colloid Interface Sci., 2005, 283, 392–396.
18. M. Lismont, C. A. Paez and L. Dreesen, J. Colloid Interface Sci., 2015, 447, 40–49.
19. L. M. Liz-Marzan, M. Giersig and P. Mulvaney, Langmuir, 1996, 12, 4329–4335.
20. T. Ung, L. M. Liz-Marzan and P. Mulvaney, Langmuir, 1998, 14, 3740–3748.
21. R. K. Iler, US Pat., 2885366, 1959.
22. P. C. Lee and D. Meisel, J. Phys. Chem., 1982, 86, 3391–3395.
23. D. D. Evanoff and G. Chumanov, J. Phys. Chem. B, 2004, 108, 13948–13956.
24. N. Zohar and G. Haran, Langmuir, 2014, 30, 7919–7927.
25. J. Lu, C. Aydin, N. D. Browning and B. C. Gates, J. Am. Chem. Soc., 2012, 134, 5022–5025.
26. B. K. Tate, J. T. Nguyen, J. Bacsa and J. P. Sadighi, Chem.–Eur. J., 2015, 21, 10160–10169.
27. B. K. Tate, C. M. Wyss, J. Bacsa, K. Kluge, L. Gelbaum and J. P. Sadighi, Chem. Sci., 2013, 4, 3068–3074.
28. J. Chêne and P. Trocellier, J. Non-Cryst. Solids, 2004, 337, 86–96.
29. J. D. Rimstidt, Geochim. Cosmochim. Acta, 1997, 61, 2553–2558.
30. G. Perera, R. H. Doremus and W. Lanford, J. Am. Ceram. Soc., 1991, 74, 1269–1274.
31. W. Z. Robert and A. Zahid, in Mineral Scales in Biological and Industrial Systems, CRC Press, 2013, pp. 173–200.
32. N. R. Jana, L. Gearheart and C. J. Murphy, Adv. Mater., 2001, 13, 1389–1393.
33. J. H. Fendler and F. C. Meldrum, Adv. Mater., 1995, 7, 607–632.
34. G. Frens, Nature, 1973, 241, 20–22.
35. D. D. Evanoff, R. L. White and G. Chumanov, J. Phys. Chem. B, 2004, 108, 1522–1524.
36. P. Kumari, P. Chandran and S. S. Khan, J. Photochem. Photobiol., B, 2014, 141, 235–240.
37. D. R. Willett and G. Chumanov, Plasmonics, 2014, 9, 1391–1396.
38. K. Zhang, J. Ji, Y. Li and B. Liu, Anal. Chem., 2014, 86, 6660–6665.
39. M. P. Konrad, A. P. Doherty and S. E. J. Bell, Anal. Chem., 2013, 85, 6783–6789.
40. K. Sugawa, T. Akiyama, Y. Tanoue, T. Harumoto, S. Yanagida, A. Yasumori, S. Tomita and J. Otsuki, Phys. Chem. Chem. Phys., 2015, 17, 21182–21189.
41. Y. Tanoue, K. Sugawa, T. Yamamuro and T. Akiyama, Phys. Chem. Chem. Phys., 2013, 15, 15802–15805.
42. S. Malynych and G. Chumanov, J. Am. Chem. Soc., 2003, 125, 2896–2898.
43. S. E. J. Bell and N. M. S. Sirimuthu, J. Am. Chem. Soc., 2006, 128, 15580–15581.
44. C. Otto, T. J. J. Van den Tweel, F. F. M. De Mul and J. Greve, J. Raman Spectrosc., 1986, 17, 289–298.
45. E. Burstein, Y. J. Chen, C. Y. Chen, S. Lundquist and E. Tosatti, Solid State Commun., 1979, 29, 567–570.
46. P. Etchegoin, H. Liem, R. C. Maher, L. F. Cohen, R. J. C. Brown, H. Hartigan, M. J. T. Milton and J. C. Gallop, Chem. Phys. Lett., 2002, 366, 115–121.
47. G. C. Schatz, M. A. Young and R. P. Van Duyne, in Surface-Enhanced Raman Scattering: Physics and Applications, ed. K. Kneipp, M. Moskovits and H. Kneipp, Springer Berlin Heidelberg, Berlin, Heidelberg, 2006, pp. 19–45.
48. I. Halaciuga, S. LaPlante and D. V. Goia, J. Colloid Interface Sci., 2011, 354, 620–623.
49. S. Pandey, G. K. Goswami and K. K. Nanda, Sci. Rep., 2013, 3, 2082.
50. F. J. Ibanez and F. P. Zamborini, ACS Nano, 2008, 2, 1543–1552.
51. F. P. Zamborini, M. C. Leopold, J. F. Hicks, P. J. Kulesza, M. A. Malik and R. W. Murray, J. Am. Chem. Soc., 2002, 124, 8958–8964.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20669g

---