APTES Duality and Nanopore Seed Regulation in Homogeneous and Nanoscale-Controlled Reduction of Ag Shell on SiO2 Microparticle for Quantifiable Single Particle SERS.

Noble-metal nanoparticles size and packing density are critical for sensitive surface-enhanced Raman scattering (SERS) and controlled preparation of such films required to achieve reproducibility. Provided that they are made reliable, Ag shell on SiO2 microscopic particles (Ag/SiO2) are promising candidates for lab-on-a-bead analytical measurements of low analyte concentration in liquid specimen. Here, we selected nanoporous silica microparticles as a substrate for reduction of AgNO3 with 3-aminopropyltriethoxysilane (APTES). In a single preparation step, homogeneous and continuous films of Ag nanoparticles are formed on SiO2 surfaces with equimolar concentration of APTES and silver nitrate in ethanol. It is discussed that amine and silane moieties in APTES contribute first to an efficient reduction on the silica and second to capping the Ag nanoparticles. The high density and homogeneity of nanoparticle nucleation is further regulated by the nanoporosity of the silica. The Ag/SiO2 microparticles were tested for SERS using self-assembled 4-aminothiophenol monolayers, and an enhancement factor of ca. 2 × 106 is measured. Importantly, the SERS relative standard deviation is 36% when a single microparticle is considered and drops to 11% when sets of 10 microparticles are considered. As prepared, the microparticles are highly suitable for state-of-the-art quantitative lab-on-a-bead interrogation of specimens.

Introduction

Since its discovery in 1973,[1] the interest in surface-enhanced Raman scattering (SERS) has grown steadily over the years. SERS measures the vibrational spectrum of materials, and as such is also widely exploited for label-free detection of biologically relevant materials, with numerous implementations in diagnostics.[2−4] SERS is most efficient at the surface of nanostructured noble metals such as Ag and Au, where intense surface plasmon modes exist and readily couple with visible or near-infrared radiations. SERS substrates typically increase the Raman scattering cross section of the molecules that are adsorbed onto them by an enhancement factor (EF) 105–107,[5] with some demonstration of higher EF and up to values affording detection of single molecule.[6,7] In this paper, we investigate a one-step, facile preparation of microscopic SERS substrates formed on silica microparticles, taking advantage of 3-aminopropyltriethoxysilane (APTES) silane and amine duality and of nanoscale porosity on the silica surface. Beyond the silver growth mechanism, we specifically focus on the reproducibility and SERS efficiency at the level of single microparticles and of a small amount of microparticles, which are crucial properties to develop lab-on-a-bead methodology, offering a reliable and flexible analytical solution to probe small volumes of liquid specimen and microfluidic channels well beyond the scope of larger substrates.

SERS is fundamentally a nanoscale process, and it is thus understandable that poor reliability and repeatability in substrate preparation has impeded for many years the adoption of SERS in quantitative analytical applications.[8] Yet today, measurements with relative standard deviation (RSD) up to <10% can be reasonably easily achieved in laboratories without sacrifice of EF, by preparing Ag or Au nanostructures in most cases through templates or by microfabrication[9,10] and commercialized substrates such as Klarite afford EF of 103–106 and RSD < 10%.[11−15]

Another widely recognized and not yet fully overcome limitation to the deployment of SERS as an analytical technique is the rapid degradation of the substrates, especially those made with silver.[8] As such, low-cost and low-equipment/skill-requirement preparation is highly desirable so that substrate fabrication can be decentralized. Reproducible SERS have been repeatedly demonstrated, but the techniques of manufacture are costly and require a lot of time and skill, such as, for example, nanolithography, focused ion beam, holographic laser illumination, and templated substrates,[8−10] and are thus hardly transportable. In this context, an ambient, low-temperature, wet chemical process such as the one discussed in this paper offers a far more desirable route.

In the chemical reduction of nanoparticles (NPs) for SERS, Ag or Au particle size and morphology are determined by experimental parameters, such as temperature, time, and reagent concentration. Discrete particles can be easily synthesized, but greater SERS EF comes from the agglomeration of small particles in clusters with hotspots. Substrates are thus routinely prepared by addition of aggregating agents such as chlorinated salts.[16,17] Yet, the degree of aggregation is not sufficiently controllable and analytical performances are thus hindered.[18] On the other hand, on glass surfaces, in the absence of other form of spatial templating, electroless reduction has also been shown to produce highly reproducible EF.[19] The latter approach is low cost, fast, and accessible to most laboratories, and thus extremely attractive from the point of view of a cost-effective widely deployable methodology for quantitative SERS. The formation of nanoparticle films is usually carried out in several distinct ways. In the seeded growth method, metal seed particles are formed in separate reactions and are combined with the substrate in suitable solvents to maximize noncovalent interactions.[20−22] Alternatively, the substrate surface is furnished with reactive species containing amine or mercapto functional groups (usually by the hydrolysis and condensation of organosilane species) and the seed nanoparticles bind to these groups due to their high affinity for these.[22−35] Seeding is then followed by electroless plating with the reduction of salts such as silver nitrate (AgNO3) or gold chloride hydrate, utilizing various reducing agents such as citrate or ascorbic acid.[20] Other methods include the sensitization of dielectric surfaces with stannous chloride (SnCl2) and Sn particles in acidic solution onto which the Ag particles can grow. The latter method does not require a reducing agent, but produces sparse coatings with wide particle size distribution and needs successive steps to form homogenous coverage.[22] By contrast, electroless plating in ethanolic solutions of AgNO3 and butylamine leads to continuous Ag coverage on glass two-dimensional substrate without seeding.[19] This reaction system is beneficial in that there are very few reagents, it is carried out at low temperature (ca. 50 °C), and it is relatively quick (ca. 1 h). The reaction has also been applied on magnetic and smooth silica spheres,[36,37] to develop Raman-based markers in assays. Electroless reduction of AgNO3 in ethanolic solutions at reflux has also been done with aminosilanes to form discrete ca. 5 nm silica-capped soluble nanoparticles.[38,39] Noble metal and silica structures have been used for many applications, such as cancer treatment, immunoassay, and theranostics.[41−46] The simultaneous growth of silica (or of other negatively charged surface layers) on silver can also facilitate the formation of other layers on the nanoparticles. For example, negatively charged poly(acrylic acid) can activate the surface and afford the deposition of other materials.[47,48] In this paper, we focus on 3-aminopropyltriethoxysilane (APTES) as it can thus presumably be used both as a mild reducing agent and as a capping agent for growing shells of Ag nanoparticles.

For the analysis of specimen diluted in solutions, beyond solid substrate-based analysis, SERS is routinely performed by introducing Ag or Au nanoparticles directly in the liquid.[40] Yet, since it is in the aggregation of the nanoparticles that truly useful EF are achieved[16,17,48] and that these remain difficult to control, quantitative measurements are typically not possible. Within the context of liquid biopsy, where small volumes of liquid specimen and low analyte concentration are found, microfluidics offers a convenient control on the specimen and microfluidic platforms have been made to include SERS-active areas in so-called hybrid systems.[2] There, SERS substrates suffer from poor recyclability and poor molecule trapping efficiency and their design and preparation is intimately linked to the platform fabrication.[9] In keeping that microparticles can be made to mix efficiently with solutions, including within the notoriously laminar flow in microfluidics[49] and that the particles, being prepared separately from the platform, afford higher repeatability and ease of fabrication, it seems reasonable for us to focus on their exploitation. Yet, the lack of repeatability in EF has to be addressed. As such, we are considering here the case of SERS-active Ag/SiO2 microparticles. Bare silica microparticles are commercially available with high monodispersity and controlled surface porosity, and we used those as a template for electroless deposition.

According to the guidelines of Betz et al.,[18] a truly useful SERS substrate should be simple to prepare (from chemicals and equipment found readily in most laboratories) and should be easily integrated into analytical systems. Within this context, we report that electroless reduction of AgNO3 in ethanol with APTES leads to reproducible SERS-active substrates and we also quantify EF and RSD, including when the microparticles are considered as individual SERS substrate. The microspheres are easy to prepare, made from cheap, commercially available reagents, and lend themselves to analysis into already available analytical systems in microscopy and in solution (e.g., cuvettes and microfluidics). The inherent mobility of microparticle SERS substrates is particularly desirable for the analysis of low volumes of liquid specimens and should suit a wide range of applications, including microfluidics and diagnostics.[50]

Results and Discussion

AgNO3 Concentration

As can be expected, the AgNO3 concentration plays a significant role in the structure of the Ag film formed on the SiO2 microparticles. Figure a shows a photograph of a series of Ag/SiO2 microparticle dispersions in ethanol along with their associated UV–vis extinction spectra, with the concentration of AgNO3 varied from 0.1 to 2.5 mM and the APTES concentration set to 1 mM. At lower concentrations (<0.5 mM), the dispersions are very clear and we infer that a negligible amount or no Ag is formed. The UV–vis spectrum matches then with that of the bare silica that has very low absorption in the visible range. With 0.5 mM of AgNO3, the solution turns yellow-brown and a distinct plasmon band at ca. 440 nm is observed in the extinction spectrum, suggesting that Ag nanoparticles form. When the concentration is increased further, the solution darkens and the extinction maxima shifts to longer wavelengths and broadens, suggesting the formation of a more continuous Ag film.[51−54]

Figure 1

(a) Photograph under ambient light of a series of Ag/SiO2 dispersions in ethanol with AgNO3 concentration varied (left to right) from 0.1 to 2.5 mM (APTES 1 mM) and corresponding UV–vis extinction spectra. (b) Same with APTES concentration varied (left to right) from 0.1 to 2.5 mM (AgNO3 1 mM).

The scanning electron microscopy (SEM) images in Figure a–c reveal the structural evolution of Ag film. With 0.1 mM AgNO3, the image mostly shows the high roughness of the nanoporous SiO2 with very little evidence of Ag nanoparticles. On the other hand, SEM reveals a uniform coverage of Ag nanoparticles with 1 mM AgNO3. The relatively spherical Ag particles exhibit a diameter of ca. 20 nm and are densely and uniformly arranged on the nanoporous silica surface. At higher, AgNO3 concentrations (2 mM), additional, disperse, and larger Ag particles are also observed. These appear to have grown above the uniform film of smaller particles. The size of the large particles scales up to ca. 100 nm.

Figure 2

(a–c) SEM images of the surface of a silica microparticle prepared with AgNO3 concentration of 0.1, 1, and 2.5 mM, respectively (APTES 1 mM). The corresponding microparticles are shown in the inset. (d, e) SEM images highlighting clusters of Ag nanoparticles grown in the solution with 1 mM AgNO3 and 0.1 mM APTES, with little Ag growth on the silica microparticles. (f) Same highlighting excess materials formed with 1 mM AgNO3 and 2.5 mM APTES.

APTES Concentration

Figure b shows a photograph of a series of Ag/SiO2 dispersions prepared with concentration of AgNO3 fixed at 1 mM and APTES concentration varied from 0.1 to 2.5 mM. The corresponding UV–vis extinction spectra are also shown. At 0.1 mM APTES, we observe that the solution is light yellow and that the UV–vis spectra shows a weak peak at 440 nm. These data suggest that Ag nanoparticles formed, and SEM further reveals (Figure c,d) that these Ag nanoparticles are majorly reduced in the solution, away from the nanoporous silica microparticle surfaces. When the concentration of APTES in further increased, the solution progressively darkens and the UV–vis spectra also show a progressive shift and broadening commensurate with the growth of Ag nanoparticle film on the silica surface. Beyond 1.5 mM APTES, the solution gradually reverts back to a yellow color and the UV–vis spectra shows the re-emergence of an intense peak positioned at ca. 440 nm for 2.5 mM APTES. At the same concentration, the SEM image in Figure f reveals large cuboids grown mostly away from the microparticles, but with also some observed on the silica microparticles. Noteworthy at concentrations of AgNO3 around 1 mM, SEM and UV–vis data suggest that the Ag nanoparticles uniformly grow preferably on the microsphere nanoporous surfaces with little growth in the solution.

Ag Reduction and Nanoparticle Film Growth Mechanism

Thus, one observes that at constant APTES concentration (1 mM), increasing the AgNO3 concentration increases the coverage in nanoparticles formed on the nanoporous silica surfaces, with NPs size seen to reach ca. 20 nm at 1 mM and with some additional larger particles up to ca. 100 nm at 2.5 mM. On the other hand, at constant AgNO3 concentration (1 mM), increasing the APTES concentration shows a more varied behavior, with Ag NPs preferably reduced in solution at low concentration, reduced on the silica surface at intermediate concentration, and competitively reduced again in the solution at higher APTES concentration.

With respect to the silica surface, the reduction mechanism would likely follow the one described by Kim et al. with butylamine,[37] with the difference here that APTES can act both as reducing (amine moieties) and capping agent (silane moieties).[38] Prior to the addition of APTES, the microparticles are expected to attract the silver ions as the ζ-potential value suggests that a fair amount of surface OH groups are deprotonated in ethanol (see Supporting Information, Figure S1a). Upon addition of silver ions, the oxygen sites are thus bound to Ag+ ions, which then seed the Ag heterogeneous growth (Figure S1b). Moreover, we have confirmed that, with the 92 Å silica pores used in this study, Ag nanoparticles nucleate homogeneously on the silica surface whereas that is not the case with the 22 Å ones (see Supporting Information, Figure S2), suggesting that selection of appropriate porosity also affords a control on the Ag NPs’ nucleation density.

In contrast to this work, no evidence of colloid excess in the solutions was observed in an earlier study where ca. equimolar butylamine was used to reduce Ag on 200 nm SiO2 microparticles.[37] This was attributed to butylamine being a weak reducing agent. Yet, precipitation of Ag particles have also been made with APTES and other aminosilanes by others (in the absence of silica microparticles and at comparatively high concentration of APTES versus AgNO3).[38] We infer thus that in our APTES-based experiments, a balance between reduction on the nanoporous silica and reduction in the solution exists, with heterogeneous growth on the nanoporous silica preferred ca. 1.0 mM APTES. Noteworthily, ethanol alone is known to serve as a reducing agent although this is generally seen under microwave assistance or sonication,[55,56] in variance to this work. It is thus expected that at sufficient concentration, APTES is here the dominating reducing agent.

Butylamine is not necessarily advantageous when the properties of the nanostructured Ag film grown on the microsphere are considered. Indeed, the Ag films here are homogenous and fully cover the microparticle surface with suggestions of numerous nucleation sites, whereas sparse particles coalescing into larger ones were seen with butylamine.[37] Here, one can speculate on the role of the silane moieties. In line with the observations of Frattini et al. also with other aminosilanes,[38] it is suggested that the silane moieties in APTES act as capping agent when APTES and AgNO3 are approximately equimolar, with the polymerization possible from a trace amount of water. It is noted that APTES binding and polymerization is not expected directly on the silica spheres that are initially covered by Ag+ but to develop more slowly above the Ag nanoparticle film, capping thus the initially reduced nanoparticles.

SERS versus Ag Reduction

It is known that smooth Ag surfaces do not provide large EF but that the formation of nanoscaled asperities and particles, as well as gaps between these, greatly enhance the Raman spectrum of the adsorbed molecules.[57] As such, the Ag film presented above is promising and we discuss here a series of Raman spectra recorded on Ag-coated microparticles that are prepared with varied AgNO3 and APTES concentrations (Figure ). Following the Ag growth, the microparticles were exposed to a 4-aminothiophenol (4-ATP) solution, which is known to lead to the formation of complete and single monolayers.[58] Thus, our Raman measurements allow for quantifying the EF and microparticle reproducibility when used as a SERS substrate. The Raman spectra were recorded after focusing the confocal Raman onto individual microparticles, and the spectra were also normalized to a silicon peak (ca. 520 cm–1) from the substrate on which the microparticles were drop casted. Each spectrum shown is the average of 10 measurements on as many microparticles. The Raman spectra recorded on microparticles is reminiscent of SERS with two strong peaks at 1390 and 1432 cm–1 that match with normally forbidden ring modes of 4-ATP and that are appearing here due to symmetry lowering when the molecule is self-assembled on Ag surfaces.[60] Overall, a good agreement between our 4-ATP spectrum and other SERS measurements published in the literature is found.[59]

Figure 3

(a, b) Color maps showing the Raman spectra of 4-ATP on Ag-coated SiO2 microparticles as a function of AgNO3 and APTES concentrations. The data are normalized, and the lower intensities are shown in dark red and the highest in white. The spectra were averaged from 10 individual microparticles. (c) Plot of the 4-ATP 1435 cm–1 Raman mode intensity on Ag/SiO2 microparticles as a function of AgNO3 concentration (with 1 mM APTES, blue open circles, RSD for each data point: 14.3, 14.0, 13.8, 12.8, 13.5, 12.2, and 12.9%) and as a function of the APTES concentration (with 1 mM AgNO3, red dots, RSD for each data point: no signal, no signal, 21.0, 13.6, 14.7, 14.1, and 12.8%) extracted from (a) and (b). (d) 4-ATP powder spectra (red) and 4-ATP on Ag/SiO2 microparticle spectra (1 mM APTES and 1 mM AgNO3, blue). The 4-ATP powder spectrum is displayed with a factor ×50 with respect to the scale, and the microparticle spectra is the average of 10 individual microparticles.

The color maps in Figure a,b show Raman spectra for different AgNO3 and APTES concentrations, respectively. The data were recorded on the same microparticle preparations as those discussed in Figures and 2, allowing us to accurately correlate the impact of the concentration of AgNO3 and APTES on the SERS performances. The maps display the lower Raman intensities in dark red and the highest in white. The maps show that the Raman intensity is uniformly affected across all wavenumbers when the AgNO3 or APTES concentration is changed. The integrated intensity of the peak at 1435 cm–1 was extracted and is plotted in Figure c. Overall, we observe that the Raman signal of 4-ATP peaks at ca. 1 mM APTES and is weaker for APTES concentration below and above. This behavior is well in agreement with the evolution of the extinction spectra in Figure b and highlights that the Ag coverage is more suitable to SERS with sufficient APTES for the reduction but not too much that Ag is also competitively reduced in the solution. We also observed that, with 1 mM AgNO3, the Raman intensity for all peaks exhibits a strong increase and that it remains relatively unchanged afterward, although a slight increase is detected within our experimental error at 2.5 mM. The evolution of the Raman spectra is thus again in full agreement with the extinction spectra in Figure a. Altogether, a suitable nanostructured Ag film forms at ca. equimolar concentration in APTES and AgNO3. When the AgNO3 concentration is too high, the Ag nanoparticle formation is not sufficiently controlled by APTES, leading to additional broader Ag features (Figure c) that are not as efficient for SERS.

SERS Enhancement Factor

It is well recognized that establishing the EF in SERS for a given system is prone to error and often overestimation in the reported values.[60] However, the EF remains an important metric to establish the potential sensitivity of the SERS substrate and we develop below our estimation of the EF on our silvered silica microparticles. Here, we compare the Raman intensities recorded on individual microparticles on silicon with a Raman spectra recorded on a deposit of 4-ATP powder on the same silicon surface and made thicker than the depth resolution of the microscope. All data are recorded within the same experimental conditions. The 4-ATP powder spectrum is superposed to the one averaged on the microparticles and presented in Figure b.

The Raman scattering in a specimen is dependent on the amount of molecule present in the illuminated volume, and thus first, we estimate the amount of molecules measured on the powder and on a single microparticle. The Raman microscope predominantly measures the top cap of the microparticle and thus a surface area of , where R is the microparticle radius and r the cap radius. Matching the cap radius to the Raman excitation beam waist (0.24 μm) and in keeping that the surface density of adsorbed thiols is typically ca. 4.5 × 1014 molecules/cm2,[61] we estimate that ca. 8 × 105 4-ATP molecules are present on the microparticle cap. Assuming that a homogeneous cylindrical volume of 4-ATP powder is measured with a radius matching again with the Raman excitation beam waist (0.24 μm) and a depth of ca. 1.7 μm and in keeping that the calculated molecular volume of 4-ATP is 1.73 × 10–10 μm3, one estimates that ca. 2 × 109 4-ATP molecules contribute to the powder spectra. Overall, thus, there are about 2.5 × 103 times more molecules probed in the normal Raman powder spectra. Considering the mode at 1573 cm–1 that is clearly observed in both powder and silvered microparticles and that is ca. 800 times more intense on the microparticles, we estimate that the EF for 4-ATP is ca. 2 × 106. The EF here is in agreement with the values found for silvered surfaces prepared by reduction of AgNO3 by butylamine in ethanol on 250 nm silica microparticles and on glass with reported EF of ca. 106 and 2 × 105, respectively.[37,19]

SERS Reproducibility

To test the relative standard deviation in EF and therefore SERS sensitivity at the level of individual microparticle and for a discrete sample of microparticles, we have recorded the 4-ATP Raman spectra on 26 Ag/SiO2 microparticles prepared by reduction with 1 mM APTES and 1 mM AgNO3 in two batches (13 from each batch). The microparticles were dispersed on a silicon surface, and the silicon peak at 520 cm–1 was again used to normalize the individual spectra. When considering the intensity of the Raman peak at 1432 cm–1, the RSD for the series of 26 spectra is 36%. The value is thus an indication of the accuracy in an analytical context when a single microparticle is considered as a SERS substrate. Most experimental situations would involve the recovery and or in situ analysis of more than one microparticle; however, when considering the averaged spectra made of a series of randomly picked 10 microparticles, we calculate that the RSD is 11% (see Supporting Information, Figure S3).

Conclusions

This paper focuses on the reduction of AgNO3 on nanoporous silica microparticles in ethanol in the presence of APTES for the development of mobile SERS microsubstrates. The one-step preparation is simple, safe to perform, and relies on widely available and low-cost chemicals and materials. Equimolar amount of AgNO3 and APTES produces reproducible film of ca. 20 nm Ag nanoparticles densely arranged on the microparticle surface with EF ca. 2 × 106 and RSD 36% when a single Ag/SiO2 microparticle is considered (including from different preparation batches).

It is speculated that the surface roughness inherent to the 92 Å pores and the silane moieties in APTES contribute to controlling the homogeneous growth on the silica surface,[38] whereas the amine moieties participate in the reduction, as observed for butylamine by others.[19,3] Extinction spectra, SEM images, and Raman spectra measured when varying the concentration of AgNO3 and of APTES suggest that equimolar concentration better confines the growth to the silica surfaces, with uncontrolled development of >100 nm large Ag islands appearing at high AgNO3 concentration and with growth of Ag/silica cuboid in solution at high APTES concentration.

A potential application of the Ag/SiO2 microparticle is within the context of a lab-on-a-bead methodology, where single beads are sufficiently SERS efficient and reproducible to warrant quantitative analysis from fluids in which the microparticles would be introduced. Silica microparticles are also available with super-paramagnetic cores so that magnetic recovery is possible.[36] As such, one notes that the microscopic size of the beads is ideal for optical interrogation at the level of individual particles. Moreover, the surface area is sufficiently large to afford statistically relevant amount of nanostructuring and gaps in the Ag film to achieve both large EF and reproducibility. Noteworthily, other authors have used silvered surfaces when interrogating biological specimens and it is foreseen thus that our microparticles will also successfully probe biological fluids.[62,63] The microparticles have thus the flexibility of nanoparticles to sensitively probe solution whilst avoiding the need for in situ aggregation for EF increase.

Experimental Section

Materials

Monodispersed silica microspheres (ca. 1.5 μm diameter) were purchased from Glantreo (92 Å average pore diameter). Silver nitrate 99% (AgNO3), 3-aminopropyltriethoxysilane (APTES), ethanol 200 proof, 4-aminothiophenol (4-ATP), and sodium chloride (NaCl) were purchased from Sigma-Aldrich. Water was purified by a Milli-Q system and had a resistance of 18.2 MΩ/cm.

Ag Reduction

SiO2 microspheres were Ag-coated using ethanolic solutions of AgNO3, with APTES as a mild reducing agent. APTES is an amine-functionalized organosilane. The SiO2 microparticles surfaces are furnished with negatively charged hydroxyl groups, which are formed during the hydrolysis and condensation reaction of the Stöber process in water/ethanol mixtures. These provide electronic attraction for the positively charged Ag+ ions. The amines act as electron donors that reduce the Ag+ ions to elemental Ag0 nanoparticles at the silica surface and act as a stabilizing agent. These particles thus form on the silica surface without the need of pretreatment or sensitizing. Specifically, silica microspheres of an average of 1.5 μm diameter were dried for 24 h at 120 °C in a vacuum oven and then dispersed in 99.9% ethanol. This dispersion was washed three times in absolute ethanol with centrifugation at 500 × g for 10 min, before being redispersed in ethanol under sonication for 5 min. The spheres were added to AgNO3 solution in ethanol and mixed thoroughly before addition of APTES. The final concentration of silica spheres was ca. 0.1 mg/mL. The solutions were placed in polypropylene tubes and heated at 50 °C for 50 min with agitation at 200 rpm in a shaking incubator. The dispersions were then washed three times in fresh ethanol with centrifugation and redispersed in fresh ethanol under sonication. The concentrations of AgNO3 and APTES were both varied between 0.1 and 2.5 mM.

Thiol Functionalization

The organic thiol SAMs were grown according to standard procedures in the literature.[61] Fifteen milliliter polypropylene tubes with 10 mL of 0.1 mg/mL Ag/SiO2 were centrifuged at 500 × g for 10 min, and the pellets were redispersed in 5 mL of fresh ethanol under sonication for 5 min before addition of 5 mL of 2 mM 4-ATP in absolute ethanol. The tubes were vortexed and placed in a shaking incubator at 200 rpm for 12 h at 25 °C. The tubes were then centrifuged and washed three times with fresh ethanol. The pellets were finally resuspended in fresh ethanol under sonication for 10 min.

Ag/SiO2 Microparticle Characterization

UV–vis extinction spectra were acquired using a Thermo Scientific Evolution 201 spectrophotometer with Hellamax Quartz absorption cells (between 350 and 1000 nm). Scanning electron microscopy (SEM) images were acquired using a Hitachi SU-70 field emission scanning electron microscope. The samples were drop casted on a commercially available gold-coated silicon substrate (Georg Albert, GmbH) and dried overnight in a desiccator. The images were recorded at 3 keV.

Raman/SERS Microspectroscopy

For SERS measurements, the 4-ATP functionalized spheres were dispersed and dried on polished silicon substrates. Si has no Raman/fluorescence background that can interfere with the 4-ATP Raman spectra, and its Raman peak at ca. 520 cm–1 was used to normalize all Raman/SERS spectra for any quantitative analyses. SERS measurements were carried out using a high-resolution Raman spectrometer (LabRam HR Evolution) from Horiba Scientific equipped with a highly sensitive Synapse CCD detector (Horiba Scientific). SERS spectra were acquired in reflection using an infinity-corrected long working distance objective (50× NA 0.55 HCXP PL FLUOTAR, Leica), a green laser (532 nm) with a power of ca. 6 μW at the sample, and a typical acquisition time of 1 s. The spectra were accumulated 3× and averaged for a better signal-to-noise ratio and to remove any cosmic spikes. The Raman spectrum of an optically thick, compact 4-ATP powder film was also acquired to estimate the SERS EF on the microparticle.