Silver Nanoparticle-Decorated Silica Nanospheres and Arrays as Potential Substrates for Surface-Enhanced Raman Scattering.

Poly(vinylpyrrolidone) (PVP) was used as both a modifier and reductant to in situ deposit silver nanoparticles (denoted Ag NPs) on the surface of silica nanospheres (nanosilica or nano-SiO2), affording Ag-decorated nanosilica (denoted SiO2@Ag). The as-obtained SiO2@Ag composite can form silver nanoparticle-decorated silica nanosphere arrays (denoted SiO2@Ag arrays) via evaporation-induced self-assembly. The as-prepared SiO2@Ag composite and SiO2@Ag array were used as the SERS substrates to measure the Raman signals of the dilute solutions of rhodamine 6G (denoted R6G), an organic dye that is a potential pollutant to the environment. The findings indicate that the as-prepared SiO2@Ag composite and SiO2@Ag array as potential SERS substrates simultaneously exhibit a high degree of metal coverage and small size of Ag NPs as well as good stability and abundant "hot spots", which contributes to their desired Raman enhancement capacities. For the detection of trace R6G, they provide a limit of detection of as low as 10-9-10-11 M as well as good reproducibility, showing promising potential for monitoring chemical and biological molecules.

Introduction

The surface-enhanced Raman scattering (SERS) technique has been applied in various fields, such as biomedicine,[1,2] chemistry,[3,4] environmental monitoring,[5,6] and food safety,[7,8] because of its high sensitivity, low detection limit, real-time detection without time-consuming pretreatment of samples, and good access to the information in the molecular fingerprint area.[9,10] To date, the electromagnetic mechanism (EM) and chemical mechanism (CM) are generally recognized for SERS enhancement.[11,12] The former is connected with “hot spots” that are dependent on the roughness of metal nanoparticles and the nanogaps between neighboring nanoparticles, and the latter is attributed to the charge transfer between the target molecule and the substrate surface. Generally speaking, EM is significant for the detection of target molecules. In other words, the SERS intensity highly relies on the microstructure and composition of the SERS substrate on a nanometer scale.[13,14] Therefore, it is imperative to fabricate SERS substrates with high Raman scattering sensitivity, good reproducibility, and low cost to obtain efficient Raman signals and achieve the widespread application of the SERS technique.

Noble metals (Au, Ag, etc.), especially Ag nanostructures, are ideal SERS substrates[15−17] because of their unique plasmonic properties in visible and near-infrared spectral regions as well as low cost as compared with Au. This could well explain why many researchers made considerable efforts to design and fabricate a variety of Ag nanostructures such as nanoparticles, nanorods, nanoprisms, nanoflowers, and nanowires as the SERS substrates with sharp edges.[18−22] For example, Xu et al. prepared triangular Ag nanoplates and obtained a detection limit of 3.125 × 10–7 M with SERS for rhodamine 6G.[20] Zhang et al. found that Ag nanoparticles and nanowires can facilitate SERS in the presence of rhodamine B since they can expose more active crystal faces.[23] The noble metal SERS-active substrates, especially their nanoparticles in a colloid solution, have the advantages of easy preparation and tunable optical properties. However, it still remains a challenge to get rid of the aggregation of nanoparticles during random growth.[24,25] Such aggregation of metallic nanoparticles is often inevitable, and the “hot spots” generated by the aggregates often exhibit poor substrate reproducibility and stability. To solve this problem, some researchers incorporate Ag nanoparticles into or onto various stable matrices to form composites.[26−30] This strategy with stable matrices as carriers has two advantages. On the one hand, stable matrices can increase the reaction sites of metal nuclei to prevent randomized aggregation. On the other hand, the stable matrices can create more “hot spots”, including not only metal nanoparticles but also the nanogaps between neighboring metal nanoparticles. With this strategy, it could be feasible to generate abundant “hot spots” and increase the density of “hot spots” in the substrates. Particularly, SERS substrates with regular and ordered array structures are advantageous over those made of ordinary metal composites.[31,32]

In terms of the study or application of SERS with noble nanoparticles, spherical silica could be an excellent core or support, mainly because of its remarkable features such as monodispersibility, high optical transparency and thermal stability, good biocompatibility, and simple synthesis by the standard Stöber method.[33−37] Furthermore, SiO2 can be easily surface-functionalized by organosilanes, including polydiallylodimethyl ammonium chloride,[38] 3-mercaptopropyltrimethoxysilane,[39,40] 3-aminopropyl trimethoxysilane,[41] polyethyleneimine, and other organic compounds such as dopamine[42] and vinylpyrrolidone.[43] The as-grafted functional groups on the silica surface can create clinical sites for the attachment and growth of Ag nanoparticles through the terminal groups,[44] which could contribute to enhancing SERS. Le Beulze et al. decorated Ag nanoparticles (diameter > 10 nm) onto SiO2 microspheres surface-functionalized with organosilanes and positively charged polymers to obtain a raspberry-like SiO2@Ag nanostructure that can function as a new type of SERS substrate for the detection of analytical targets.[45] Wang et al. utilized a thin polyethyleneimine interlayer to adsorb dense Au nanoparticles as seeds and prepared a silver shell/silica core nanostructure as the SERS substrate with a limit of detection as low as 10–9 M for the pesticide thiram.[46] Chen et al. deposited Ag onto a SiO2 nanosphere template and fabricated broccoli-like Ag arrays; the as-obtained Ag arrays as the SERS for endocrine-disrupting chemicals have a limit of detection of 2.2 × 10–8 M.[47] These research studies on Ag nanoparticle-decorated SiO2 nanosphere for SERS, nevertheless, utilize complicated and time-consuming procedures as well as strong reducing agents or external energy sources for reducing the Ag precursor. This would provide polydispersed Ag nanoparticles with good access to the surface of silica nanospheres, thereby damaging the reproducibility and stability of the as-prepared SERS substrate.

Bearing those perspectives in mind, here we adopt a facile and versatile method to prepare a Ag nanoparticle-decorated raspberry-like SiO2 composite (SiO2@Ag) as a novel SERS substrate. In the presence of poly(vinylpyrrolidone) (PVP) as both a modifier and reductant, Ag nanoparticles (Ag NPs) tend to in situ deposit on the surface of SiO2 but not aggregate thereon, while SiO2 nanospheres function to prevent the aggregation of Ag nanoparticles and facilitate the formation of a large amount of accessible “hot spots” in the nanogap region for high SERS enhancement. Moreover, Ag nanoparticle-decorated SiO2 nanosphere arrays (SiO2@Ag arrays) are assembled via the evaporation-induced self-assembly of SiO2@Ag. This article reports the preparation of the SiO2@Ag composite and SiO2@Ag arrays as well as their use as the SERS substrates for the detection of rhodamine 6G (denoted R6G), a commonly used organic dye that is a potential pollutant to the environment.

Results and Discussion

Figure a–c shows the TEM images of SiO2 nanoparticles. It is seen that the SiO2 nanoparticles obtained by the modified Stöber method under different proportions of reactants (mass fraction of TEOS, ethanol, and ammonium is 1:25:1, 2:22:1, and 2:25:1) have an average diameter of 50, 150, and 300 nm, respectively. Figure d shows the FTIR spectra of pure SiO2 nanoparticles and PVP-SiO2. The absorbance bands near 1070 and 793 cm–1 are attributed to the asymmetric stretching vibration and the symmetric stretching vibration of Si–O–Si. The stretching vibration peak of Si–OH emerges at 973 cm–1; the peaks at 1629 and 3419 cm–1 correspond to the bending vibration of H–O–H and the stretching vibration of O–H, respectively.[48] The absorbance peak of the PVP-modified silica at 1629 cm–1 is weaker than that of the pure silica, which demonstrates that the content of −OH on the silica surface decreases after surface-capping by PVP.[49] In addition, the −CH2 stretching vibration peaks are located at 2927 and 2980 cm–1, which proves that PVP is modified onto the surface of the nanosilica. Figure e shows the TEM image of PVP-capped SiO2. It can be seen that the surface of PVP-capped SiO2 is clean and relatively smooth, but the surface of the SiO2@Ag composite is relatively rough because of the formation of uniform and abundant Ag NPs.

Figure 1

TEM images of SiO2 nanoparticles with a size of (a) 50 nm, (b) 150 nm, and (c) 300 nm. (d) FTIR spectra of the pure SiO2 and PVP-modified SiO2. (e) TEM image of PVP-capped SiO2.

It is essential that SERS substrates have abundant “hot spots” (e.g., metal nanoparticles, nanogaps of closely spaced metal nanoparticles) for acquiring good performance. Figure a–c shows the TEM images of various SiO2@Ag composites. The Ag NPs in SiO2@Ag-50 exhibit a larger size and more uneven distribution than those in other composites, and the particle size of Ag NPs in SiO2@Ag-150 and SiO2@Ag-300 is about 2–3 nm. As the size of SiO2 increases from 50 to 150 and 300 nm, the plasmon peak of corresponding SiO2@Ag composites blue-shifts from 476 to 442 and 442 nm, and the absorbance spectrum of SiO2@Ag-50 is broadened in association with the appearance of a new shoulder peak (Figure d). The nonuniform particle size and diverse shape of the Ag NPs in the SiO2@Ag-50 composite could be attributed to the fact that a less amount of PVP is grafted onto the surface of a smaller nanosilica; when the size of nano-SiO2 is about 50 nm, the steric hindrance would be enhanced to hinder the nucleation of Ag NPs and allow the full growth of Ag crystals. As a result, the amount of the surface-grafted Ag nanoparticles would decrease and their average radius would increase, thereby affording Ag NPs with nonuniform particle sizes and diverse shapes. The SERS spectra of 10–4 M R6G on various SiO2@Ag composites are shown in Figure e, where the intensity of the peak at 1645 cm–1 is quantified to compare the signal enhancement ability of the nanosilica-based composites with different sizes of PVP-SiO2 (Figure f). It is seen that all the three kinds of SiO2@Ag composites can enhance the Raman signal intensity, and composites SiO2@Ag-150 and SiO2@Ag-300 provide a much stronger Raman signal intensity than SiO2@Ag-50. This is because the average radius of Ag nanoparticles decreases and the amount of Ag nanoparticles grafted on the nanosilica surface increases with increasing size of the nanosilica, which contributes to facilitating the creation of “hot spots” under enhanced plasmon coupling and local electric field and adding to Raman signal enhancement capability.

Figure 2

TEM images of (a) SiO2@Ag-50, (b) SiO2@Ag-150, and (c) SiO2@Ag-300 composites; insets are images with a higher magnification; their (d) ultraviolet–visible light (UV–vis) spectra and (e) Raman spectra; and (f) corresponding intensities of peaks at 1198 and 1645 cm–1.

Figure a–d shows the TEM images of various SiO2@Ag composites prepared under different pH values (the numeral suffixes 7, 8, 9, and 10 refer to the pH values). Sodium hydroxide not only acts as a pH regulator but also plays an essential role during the reaction process. When the pH of the reaction system is 10, excessive Ag2O will be generated, as evidenced by the relevant XRD analysis shown in Figure e, and the presence of Ag2O is harmful to SERS.[50] The UV–vis absorbance spectra of various SiO2@Ag composites prepared under different pH values are shown in Figure f. It is seen that SiO2 does not have a distinct UV–vis absorbance signal. As the pH value of the reaction system increases from 7 to 9, the plasmon peaks of corresponding SiO2@Ag composites slightly blue-shift from 468 to 462 and 456 nm, and their maximum absorbance intensity gradually increases therewith, which means that the size of Ag NPs decreases and their amount increases with increasing pH. This could be because the increase of pH from 7 to 9, i.e., the increase of −OH concentration, contributes to promoting the nucleation of Ag NPs and slowing down the growth of Ag NPs. Furthermore, the SERS spectra of R6G in Figure g and the signal intensities of the peak at 1645 cm–1 are quantified to compare the Raman signal enhancement ability of the as-prepared SiO2@Ag composites (Figure h). It can be seen that the SiO2@Ag composite prepared at a pH value of 9 exhibits stronger Raman signal enhancement capability than those composites prepared at pH values of 7 and 8. The reason lies in that the increase in the pH value of the reaction solutions results in the increase of the amount of Ag NPs and the decrease of their size, which is favorable for the formation of more closely spaced Ag NPs with dense inter-nanogaps as the “hot spots”, thereby contributing to Raman signal enhancement.

Figure 3

(a–d) TEM images of SiO2@Ag composites obtained under a reaction solution pH of 7, 8, 9, and 10 and their (e) XRD patterns, (f) UV–vis spectra, and (g) Raman spectra as well as (h) corresponding signal intensities of the absorbance peaks at 1198 and 1645 cm–1.

The Ag nanoparticles decorated on the surface of SiO2 are in direct contact with the analyte molecules and could significantly affect the SERS activity. The SiO2@Ag prepared at an initial Ag(NH3)2OH concentration of 0.15 M and a pH of 9 under room temperature is used as the example and analyzed by TEM in association with energy-dispersive X-ray (EDX) analysis to clarify that supposition. Figure a,b shows the SEM images and EDX mapping of Ag. It can be seen that the nanosilica spheres retain nearly unchanged shape after the surface-capping by PVP and the decoration by Ag nanoparticles. The modifier PVP as both a modifier and mild reductant can cover the entire surface of the nanosilica and reduce Ag+ to form Ag nanoparticles with highly uniform size and abundant nanogaps as “hot spots”, which contributes to facilitating SERS. Figure c shows the SERS spectra of R6G solutions with concentrations ranging from 10–9 to 10–2 M in the presence of the as-prepared SiO2@Ag composite as the substrate. It is seen that the Raman characteristic peak at 1645 cm–1 can be detected even under a very low R6G solution concentration of 10–9 M. Figure d shows the natural logarithms of the Raman signal intensity of R6G solutions versus the negative logarithms of their concentrations from 1.0 × 10–9 to 1.0 × 10–4 M. It is seen that there is a good linear relationship between the two parameters. The linear equation is written as y = 5.5711 – 0.3402x, and the linear regression coefficient is described as R2 = 0.9444, where y is the SERS intensity of the Raman spectra and x is the concentration of the R6G solution. The limit of detection (LOD) for R6G is about 10–9 M. The SERS enhancement factor (EF) is calculated as EF = (ISERS × cblank)/(Iblank × cSERS), where ISERS and Iblank are the intensities of the sample under SERS and non-SERS conditions, and cSERS and cblank are the concentrations of R6G solutions under SERS and non-SERS conditions. The EF value of the peak at 1645 cm–1 is estimated to be 2.76 × 106. In addition, the SERS detection toward 4-mercaptobenzoic acid (4-MBA) and 4-mercaptophenol (4-MPh) was also performed. As shown in Figure S1, their SERS signals are greatly improved.

Figure 4

FESEM images of (a) SiO2@Ag composites; (b) associated element mappings of Ag; (c) SERS spectra of R6G solutions with concentrations ranging from 10–9 to 10–4 M; (d) logarithm of the SERS intensity of R6G at 1645 cm–1 as a function of negative logarithm of R6G concentration.

It is known that the average SERS intensity of an ordered nanostructure is 2–4 orders of magnitude larger than that of a single nanostructure.[36] This is because the ordered nanostructure can provide an ultrastrong coupled plasmon and greatly enhance the local electric field, thereby greatly affecting the SERS activity. Figure a,b shows the SEM images of the SiO2@Ag array samples prepared by the evaporation-induced self-assembly of SiO2@Ag composites. The abundant small-sized Ag NPs with small interseparations or gaps as well as small gaps between Ag NPs and SiO2@Ag composites contribute to providing much more “hot spots”, thereby largely enhancing SERS. The SERS spectra of the R6G solutions with concentrations from 10–11 to 10–2 M in the presence of SiO2@Ag arrays as the substrate are shown in Figure c. The Raman characteristic peak at 1645 cm–1 can be detected even when the concentration of the R6G solution is as low as 10–11 M, and there is a linear relationship between the natural logarithms of the Raman signal intensity of the R6G solutions and the negative logarithms of their concentrations from 1.0 × 10–10 to 1.0 × 10–2 M (Figure d). The linear equation is written as y = 4.7787 – 0.1530x, and the linear regression coefficient is described as R2 = 0.9899. In this case, the LOD for R6G is 10–11 M, and the SERS enhancement factor is about 6.29 × 108.

Figure 5

(a, b) FESEM images of SiO2@Ag arrays. (c) SERS spectra of R6G solutions with concentrations ranging from 10–11 to 10–2 M. (d) Logarithm of the SERS intensity of R6G at 1645 cm–1 as a function of the negative logarithm of R6G concentration. (e) Raman spectra of SiO2@Ag arrays and (f) the corresponding peak intensities of eight randomly measured sites at 1645 cm–1.

Reproducibility is a vital factor for the SERS substrates. Figure e shows the Raman spectra of the R6G solution (10–5 M) measured at eight sites with SiO2@Ag arrays as the SERS substrates. (Here, the to-be-measured sites are randomly selected on the SERS substrate covered with R6G.) The intensities of typical Raman peaks at 1645 cm–1 are chosen to investigate the reproducibility of the SERS enhancement (Figure f). The relative standard deviation of the eight measured Raman signals is 8.58%, which confirms that the SiO2@Ag array SERS substrate is reliable and reproducible.

Conclusions

Ag nanoparticles are in situ deposited on the surface of silica nanospheres in the presence of PVP as both a modifier and reductant to afford the SiO2@Ag composite. The SiO2 nanospheres as a matrix can prevent the aggregation of Ag nanoparticles and facilitate the formation of a large amount of accessible “hot spots” in the nanogap region for high SERS enhancement. Furthermore, SiO2@Ag arrays are assembled via the evaporation-induced self-assembly of SiO2@Ag. The as-prepared SiO2@Ag composite and SiO2@Ag arrays exhibiting a high metal coverage and good stability as well as abundant “hot spots” could be potential SERS substrates with desired Raman signal enhancement capability. Particularly, the SiO2@Ag arrays as the SERS substrate can even well detect R6G at a concentration of as low as 10–11 M with good reliability and reproducibility, showing promising prospects for the detection of chemical and biological molecules.

Experiments and Methods

Materials and Chemicals

Tetraethylorthosilicate (TEOS, ≥98%, AR) was purchased from Tianjin Kemiou Chemical Reagent Company. PVP (K15, Mw: 10 000, analytical reagent), silver nitrate (AgNO3, ≥99.8%, analytical reagent), and R6G (analytical reagent) were purchased from Sigma-Aldrich. Sodium hydroxide (NaOH), ammonium hydroxide (28%), and ethanol (≥95%) were commercially obtained. Distilled water was prepared at our laboratory and used as the solvent.

Experimental Procedure

PVP-modified SiO2 particles, raspberry-like SiO2@Ag composite, and SiO2@Ag arrays were prepared through the procedure schematically shown in Scheme .

Scheme 1

Schematic Diagram of the Preparation of the SiO2@Ag Composite and SiO2@Ag Arrays

Preparation of Monodispersed SiO2 Nanoparticles via the Stöber Method

SiO2 nanoparticles were prepared by the modified Stöber method. Briefly, ethanol and ammonium were mixed in a three-neck flask under gentle stirring while a certain amount of TEOS was added dropwise into the mixed solution (the proportion (mass fraction) of TEOS, ethanol, and ammonium was adjusted to be 1:25:1, 2:22:1, and 2:25:1 to obtain SiO2 nanoparticles with different sizes). The mixed solution becomes opaque in 20 min due to the formation of SiO2 nanoparticles. No sedimentation was observed over a period of several hours, which suggests that the as-generated SiO2 nanoparticles were colloidally stable. The product was collected through filtration and washed with deionized water and ethanol three times.

Preparation of PVP-Modified SiO2 Nanoparticles

SiO2 nanoparticles (2 g) and 100 g of water were charged into a 250 mL three-neck round-bottom flask. The mixture was vigorously stirred until SiO2 was uniformly dispersed in the reaction system. Then, an aqueous PVP solution (10 mL, 5 mM) was added into the colloid solution of SiO2 nanoparticles. The resultant dispersion was allowed to react for 2 h. The product was collected by filtration and rinsed with deionized water and ethanol three times. The as-obtained modified SiO2 particles were denoted PVP-SiO2.

Preparation of the SiO2@Ag Composite

PVP-SiO2 (0.4 g) was added in 30 mL of water under vigorous stirring, while a certain amount of NaOH (1 M) was added and stirred for 10 min to adjust the pH of the mixed solution to 7–11. Then, freshly prepared Ag(NH3)2OH (0.075–0.6 M) was quickly added to the PVP-SiO2 dispersion. The color of the mixture changes from white to yellow or brown immediately due to the formation of Ag nanoparticles. The PVP-SiO2 dispersion was allowed to react for 2 h so as to achieve the sufficient deposition of Ag nanoparticles on the PVP-SiO2 surface, yielding more active sites.

Preparation of the SiO2@Ag Arrays

The SiO2@Ag arrays were prepared by the evaporation-induced self-assembly. In brief, a glass substrate was placed in a 10 mL vessel containing 6 mL of ethanolic suspension of the SiO2@Ag composite (0.8 wt %). Then, the vessel was placed vertically in a constant temperature oven to evaporate ethanol and afford SiO2@Ag arrays via the orderly arrangement of the SiO2@Ag composite on the glass substrate under capillary force.

Characterization and SERS Measurement

The as-prepared nanosilica was characterized by Fourier transform infrared spectroscopy (FTIR, Nicolet 170sx, Thermo Fisher Scientific Inc.). The morphology of the samples was observed with a transmission electron microscope (TEM; JEM-2010, JEOL, Japan) and a scanning electron microscope (SEM; Japan Electronics, Japan). The X-ray powder diffraction (XRD) patterns of the products were obtained with an X-Pert Pro diffractometer (Cu Kα radiation; λ = 0.15418 nm; voltage: 40 kV; current: 40 mA). The UV–visible absorbance spectra of SiO2@Ag composites were recorded with a Jinghua 723PC spectrometer (Shanghai, China). The Raman spectra were collected with a laser confocal Raman spectrometer (Renishaw RM-1000, U.K.) at an excitation wavelength of 532 nm.

For the SERS measurements, standard R6G solutions at a concentration from 10–2 to 10–11 M were prepared. In addition, a proper amount of concentrated SiO2@Ag dispersion (10–2 M) was placed onto a glass substrate and dried at room temperature. Onto the as-dried glass substrate was dropped the R6G solution and dried at an ambient condition to immobilize the organic dye on the surface of the substrate. Upon completion of the immobilization of the R6G dye, the glass substrate was used for SERS measurements to detect R6G.