PDF Induced SERS activity in Ag@SiO2/Ag core-shell nanosphere arrays ... - NJU

(1)

Induced SERS activity in Ag@SiO ₂ /Ag core-shell nanosphere arrays with tunable

interior insulator

Li-Wei Liu,

^a

Qing-Wei Zhou,

^a

Zhi-Qiang Zeng,

^a

Ming-Liang Jin,

^b

Guo-Fu Zhou,

^b

Run-Ze Zhan,

^c

Huan-Jun Chen,

^c

Xing-Sen Gao,

^a

Xu-Bing Lu,

^a

Stephan Senz,

^d

Zhang Zhang

^a

* and Jun-Ming Liu

^a,e

In this work, we demonstrated a bottom-up growth of Ag@SiO2/Ag core-shell nanosphere arrays with tunable SiO2interior insulator and the optimized surface-enhanced Raman scattering (SERS) substrate based on a nanostructure performed with both high sensitivity and large-area uniformity. Their morphological, structural, and optical properties were characterized, and the induced SERS activities were investigated theoretically by the FDTD simulation and experimentally using analyte molecules. An ultrathin SiO2shell with tunable thickness can be synthesized pinhole-free by a chemical vapor deposition, working as an interior insulator between the Ag core and Ag out-layer coating. A detection limit as low as 10 ¹²M and an enhancement factor up to 3 × 10⁷were obtained, and the SERS signal was highly reproducible with small standard deviation. The method opened up a way to create a new class of SERS activity sensor with high-density‘hot spots’, and it may play an important role in device design and the corresponding biological and food safety monitoring applications. Copyright © 2016 John Wiley & Sons, Ltd.

Additional supporting information may be found in the online version of this article at the publisher’s web site.

Keywords:SERS; bottom-up method; Ag@SiO₂/Ag nanosphere; solid-state dewetting; CVD

Introduction

Surface-enhanced Raman scattering (SERS) is a powerful technique to detect and identify specific analytes. It is able to provide information about vibration modes that are associated with chemical and structural features of the targets down to the single-molecule regime.^[1,2]The sensitivity of a SERS substrate is mainly a result of the surface plasmon resonance (SPR) coupling effect, which is induced by the formation of‘hot spots’at the gaps or junctions between nanostructures of noble metals, such as Ag, Cu, and Au.^[3–5]

So far, numerous efforts have been made in the synthesis of diverse nanostructures as Raman signal enhancing agents. Noble metal nanoparticles (Ag and Au) are the most well-established SERS substrates and have been reported to have the largest enhancement factors and even single molecule detectivity.^[6,7] However, metallic colloids require stabilization. An uncontrolled aggregation of the metal nanoparticles (NPs) in the aqueous phase would promote for the formation of conglomerations, resulting in a poor reproducibility of SERS signals. To address this issue, a strategy called shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS) has been proposed to protect the metal NPs by wrap- ping them with an ultra-thin silica or alumina shell.^[8–11]Though SHINERS has greatly improved the surface generality for SERS,^[12]

the NPs are sometimes separated by complex methods, and there is still room for SHINERS to be further developed.^[13,14]Meanwhile, nanohybrid-based SERS substrates had been created by coating with noble metal films (Ag or Au) onto the surface of an ideal nanostructure. It has been found that various silicon-based nanostructures, such as nanowires,^[15,16] nanopillars,^[17,18] and

nanostars,^[19] effectively improve the electromagnetic enhancement. Furthermore, SERS substrates based on ordered silicon nanostructures possess excellent reproducibility, yielding nearly identical Raman spectra with large-area uniformity. However, the typical fabrication methods such as nanosphere lithography,^[15]

reactive ion etching,^[17]and electron-beam lithography,^[19]are still time-consuming and with a relative high fabrication cost.

Currently, there has been growing interest in developing a new class of SERS-active substrates. Novel multifunctional Fe₃O₄@Ag/

SiO2/Au core-shell microspheres (NSs)^[20]and Ag–Silica–Au hybrid

* Correspondence to: Zhang Zhang, Institute for Advanced Materials, Guangdong Provincial Key Laboratory of Quantum Engineering and Quantum Materials, South China Normal University, Guangzhou 510006, China.

E-mail: [email protected]

a Institute for Advanced Materials and Guangdong Provincial Key Laboratory of Quantum Engineering and Quantum Materials, South China Normal University, Guangzhou 510006, China

b Electronic Paper Display Institute, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, China

c State Key Laboratory of Optoelectronic Materials and Technologies, Guangdong Province Key Laboratory of Display Material and Technology, School of Physics and Engineering, SunYat-sen University, Guangzhou 510275, China

d Max Planck Institute of Microstructure Physics, Weinberg 2, Halle 06120, Germany

e Laboratory of Solid State Microstructures and Innovation Center of Advanced Microstructures, Nanjing University, Nanjing 210093, China

Received: 31 October 2015 Revised: 19 March 2016 Accepted: 28 March 2016 Published online in Wiley Online Library: 1 July 2016

(wileyonlinelibrary.com) DOI 10.1002/jrs.4941

12 00

(2)

multiple steps. In addition, immobilization of low-dimensional nanostructures onto a solid substrate usually generates an inhomo- geneous and unstable distribution, as a result of the insufficient connection between the nanostructures and the substrate.^[21]

Few attempts have been reported to realize an ultrathin silica film coated on metal cores via dry fabrication methods. Double-layer stacked Au/Al2O3@Au NSs on silicon substrate have been demonstrated to improve the stability of the metal NSs, with an ultrathin alumina cladding coated by an atomic layer deposition technique.^[25]However, for SERS substrates based on double-layer stacked metal NSs, the detection limit needs to be further optimized. In our previous work, with a chemical vapor deposition (CVD) process, a tunable Si shell could be grown homogeneously onto the surfaces of Ag NPs dewetted on Si.^[26]

In this paper, we present a bottom-up growth method to realize Ag@SiO₂/Ag core-shell nanosphere arrays on Si as a SERS substrate.

For the Ag@SiO2/Ag core-shell nanosphere arrays, the ultrathin silica shell acts as an interior insulator between the Ag core and Ag outer layer, and the interior insulator can be precisely tuned by the Si CVD growth time. Because of the localized surface plasmon resonance near the interior insulator region, each Ag@SiO2/Ag nanosphere can induce SERS activity, and the high density of nanosphere arrays can produce a remarkable enhancement of SERS signals. Through a study on their optical properties and SERS activities, the SERS-active substrate of the Ag@SiO2/Ag core-shell nanosphere arrays on Si exhibits a SERS detection limit as low as 10 ¹²M with large-area uniformity. The method opened up a way to create a new class of SERS activity sensor with a high density of‘hot spots’, and it may play an important role in device design and the corresponding biological and food safety monitoring applications.

Experimental section

Sample preparation

The n-doped Si(100) wafers were prepared by a standard RCA cleaning and then dipped into 5% hydrofluoric acid to obtain a hydrogen-terminated Si surface. A thin Ag film with 15 nm thickness was deposited by a thermal evaporation process at a growth rate of 0.1 Å/s. Subsequently, the substrate was loaded to a LP-CVD system (First Nano ET-3000 EXT, CVD Equipment Corp.) and was heated up to 540 °C for 30 min at 10 Torr H₂to form Ag nanoparticle arrays. An ultra-thin silicon shell was synthesized by a CVD process with diluted silane (5% SiH₄in H₂) as a gas precursor and 99.999% pure H2as a carrier gas. The pressure was set to be 10 Torr with a fixed mass flow of 40 sccm H₂ and 20 sccm SiH₄. The used growth times of Si layers were 1, 3, 6, and 10 min,

System (Gatan 691).CV experiments were conducted in a conven- tional three-electrode electrochemical cell on a CHI 660B electrochemical work station, using a platinum foil as an auxiliary electrode and Ag/AgCL (1 M NaCL) as a reference. Doped n-type Si substrates (2–4Ω· cm) deposited with Ag film before and after the CVD process were employed as working electrodes, respectively. CV measurements were performed in a NaOH solution (1 M), and all potentials cited in this work refer to the Ag/AgCL (1 M NaCL) electrode. The scanning range is 0.2–0.9 V, while the scanning rate is 50 mV/s. Regarding the SERS measurements, the analyte molecules p-Thiocresol (p-Tc) in ethanol solutions and R6G in aqueous solutions ranging from 1 × 10 ⁶to1 × 10 ¹²M were prepared. The substrates were dipped into a solution of dissolved molecules for 1 h, washed in deionized water to remove excess molecules, and dried in the clean air. For the measurement of SERS, all the Raman experiments were carried out with 633 nm excitation laser lines; the laser power was about 0.24 mW. The inelastically scattered radiation was collected on a microscopy Raman spec- trometer (42K864 Renishaw, inVia) with a CCD detector and an optical microscope using a 50× microscope objective with a numerical aperture value of 0.6.

Results and discussion

The fabrication procedures of Ag@SiO2/Ag nanosphere arrays on Si substrate are illustrated schematically in Fig. 1. First, a thin Ag film with a thickness of 15 nm was deposited on a Si(100) substrate via a high vacuum thermal evaporation (Fig. 1a). By a solid state dewetting at 540 °C in the high vacuum CVD system, high-density Ag nanoparticle arrays were formed on Si (Fig. 1b). After Ag dewetting, a reactive gas precursor of 5% SiH4mixed with a carrier gas of H₂were fluxed into the CVD chamber. The SiH₄decomposed on the hot surfaces, depicted in Fig. 1c, which shows an ultrathin Si layer grown on the Si substrate homogeneously with embedded Ag NPs. Because the density and size of Ag NPs changed with the dewetting time at a certain temperature,^[27] the continuously grown Si layer immobilized the Ag NPs from further agglomeration.

After the CVD process, the ultrathin Si layer was oxidized to a silica shell in the lab atmosphere (Fig. 1d). Then, an outer layer of Ag with a thickness of about 10 nm was deposited onto the substrate to form Ag@SiO2/Ag core-shell nanosphere arrays (Fig. 1e), which could work as a SERS substrate. Test molecules such as p-Tc or rhodamine 6G (R6G) were adsorbed onto the SERS substrate by immersing the substrate into diluted molecule solutions (as illustrated in Fig. 1f).

The morphologies of the Ag@SiO₂/Ag core-shell nanosphere arrays were studied by scanning electron microscopy (SEM).

12 01

(3)

Figure 2a and b are top-view and 10°-tilted side-view SEM images of the as-prepared Ag@SiO₂ core-shell nanosphere arrays on Si.

As observed in Fig. 2a, high densities of the Ag@SiO2 NSs were grown on Si substrate with large-area uniformity. The density of NSs was estimated to be about 10⁹cm ², and the large-area uniformity was further confirmed by atomic force microscope (refer to Fig. S1). Figure 2b is a tilted side-view SEM image, which shows the hemispherical nanostructure of Ag@SiO₂. The inset of Fig. 2b illustrates the size distribution of more than 900 Ag@SiO2 NSs, and the mean diameter was about 80 nm. Figure 2c is a top-view SEM image of the Ag@SiO2/Ag core-shell nanosphere arrays. By a 10-nm thick Ag coating, surfaces of both the Ag@SiO₂ NSs and the Si substrate in between are covered with a quasi-continuous Ag outer layer. The tilted side-view SEM image of Fig. 2d further shows a quasi-continuous structure of the Ag outer layer. The inset of Fig. 2d illustrates the size distribution of more than 900 Ag@SiO₂/

Ag NSs, and the mean diameter is about 100 nm, being consistent with the 10-nm thick Ag outer layer coating.

Crystallographic characterizations by transmission electron microscopy (TEM) are shown in Fig. 3. Figure 3a is the TEM image of a single Ag@SiO2 core-shell nanosphere disconnected from Si substrate, and the CVD growth time for Si was 3 min. The Ag core can be distinguished by the dark contrast wrapped around the SiO₂shell with the bright contrast. From its magnified view (from the selected area marked in Fig. 3a) shown in Fig. 3b, we observed that after the oxidization, an amorphous silica shell wrapped around the Ag core with a uniform thickness of about 3 nm.

Figure 3c was the magnified TEM image of Ag@SiO₂nanosphere with a CVD growth time of 1 min, and the silica shell thickness was confirmed to be about only 1 nm. Overall, the thickness of the SiO2shell was tunable corresponding to the different Si CVD growth times, the increase of which resulted in the increase of Figure 1. Schematic diagram to illustrate the fabrication procedures of Ag@SiO2/Ag nanosphere arrays on Si as a highly sensitive surface-enhanced Raman scattering (SERS) substrate.

Figure 2. (a) Top view and (b) 10-degree tilted side-view SEM images for the Ag@SiO2nanosphere arrays with a 1-min Si CVD growth and followed by oxidization in clean air. Inset of (b) shows the size distribution histogram of the diameters of 900 Ag@SiO2NSs. (c) Top view and (d) 10-degree tilted side- view SEM images for the Ag@SiO2Ag nanosphere arrays with 10-nm Ag outer layer coating. Inset of (d) shows the size distribution histogram of the diameters of 900 Ag@SiO2NSs.

12 02

(4)

SiO2 shell thickness (refer to Fig. S2). Figure 3d is a low- magnification cross-sectional TEM image of Ag@SiO₂/Ag nanosphere arrays on the Si(100) substrate. It was recognized that the quasi-continuous Ag outer layer covered the surfaces of both Ag@SiO2NSs and Si substrate. From a cross-sectional high-resolution-TEM image of an Ag@SiO₂/Ag nanosphere grown on Si (shown in Fig. 3e), an ultrathin SiO2 shell about 1-nm thick was distinguished as an interior insulator between the Ag core and the Ag out-layer coating. Moreover, a high-resolution-TEM image (Fig. 3f) was acquired from a region marked by the red-dotted line box plotted in Fig. 3d, which confirmed the amorphous nature of the SiO2

layer being distinguished from the lattice planes of both the Si (100) substrate and Ag core. Because the SERS activity at the interior insulator region could be induced by the inner Ag core and the outer Ag layer, the growth of high-density Ag@SiO2/Ag NSs on Si substrate (about 10⁹cm ²) would bring a remarkable enhancement of Raman signals.

Recently, it had been found that an ultra-thin SiO₂shell (e.g. less than 2-nm thick) usually brought in a pinhole effect.^[28]The exis- tence of pinholes would lead to a direct contact between Ag core and the probed molecules, which would yield SERS signals. It might introduce problems in the analysis of whether the SERS signals mainly attributed by the pinhole effect or by the SPR coupling effect between inner Ag cores and outer Ag layer. Therefore, it is essential that the ultrathin silica shell is pinhole-free. TEM observations had proven the uniformity of the 1-nm thick SiO₂ shell. However, that was insufficient to identify pinholes with small density existing in the shell. In addition, Raman spectroscopy and cyclic voltammograms (CVs) have been introduced to characterize

the pinhole effect.^[29]R6G was used as a probed molecule, and the concentration of R6G in aqueous solutions was 10 ⁶M. When R6G molecules were adsorbed onto the Ag film directly deposited on Si substrate, a representative SERS spectrum of R6G was observed (Fig. 4a, curve i) as a result of the rough metal surface.

However, when R6G molecules were adsorbed onto the dewetted Ag NPs on Si, there was not sufficient Raman signals obtained (curve ii) as a result of the relative large spacing between Ag NPs.

For the Ag@SiO2NSs with the 1-nm thick silica shell, R6G molecules would penetrate the ultra-thin SiO₂shells and be adsorbed onto the surface of Ag cores if many pinholes existed. Therefore, the observed Raman spectrum without any obvious Raman peaks of R6G (Fig. 4a, curve iii) indicated that the 1-nm thick SiO2shell had a low density of pinholes. Furthermore, Fig. 4b displays CV profiles of two kinds of nanostructure-modified working electrodes: (1) 15-nm thick Ag film coated on the Si wafer and (2) Ag@SiO₂core- shell NSs with 1-nm thick SiO2shell on the Si substrate. In the case of the Ag film directly contacting with electrolyte, an obvious electrochemical response (curve i) was observed, with anodic current peaks in the range of 0.2–0.6 V and cathodic peak at about 0.1 V.

The anodic current peaks correspond to the oxidations of Ag to Ag₂O and AgO, while the cathodic peak is assigned to the reduction of AgO back to metallic silver.^[30,31]Otherwise, to the Ag@SiO2

nanosphere arrays, the disappearance of Ag characteristic oxida- tion and reduction peaks (curve ii) demonstrated that the Ag cores have been sealed by a pinhole-free SiO₂shell.

To better understand the interactions of an electromagnetic wave with an individual Ag@SiO₂/Ag core-shell nanosphere, a three-dimensional finite difference time domain simulation was Figure 3. (a) TEM image of an individual Ag@SiO2nanosphere with a 3-min Si CVD growth time. (b) Magnified TEM image in the selected area (red dotted line) marked in (a); the thickness of SiO2shell is about 3 nm. (c) TEM image of Ag@SiO2NSs with a 1-min Si CVD growth time; the corresponding SiO2shell thickness is 1 nm. (d) Cross-sectional TEM image of Ag@SiO2/Ag nanosphere arrays grown on Si(100) substrate; a quasi-continuous Ag out-layer coating can be observed. (e) Magnified cross-sectional TEM image of one Ag@SiO2/Ag nanosphere grown on Si substrate; inset shows the ultrathin SiO2shell as an interior insulator. (f) high-resolution-TEM image of the selected area marked in (d), which confirms the amorphous SiO2interfacial layer between the crystallized Si (100) substrate and Ag core.

12 03

(5)

realized (Figs. S3 and S4). The images show that resonant coupling between the inner Ag core and outer Ag layer could propagate surface plasmons to induce SPR with enhanced electrical fields. As shown in Fig. 4c, the maximum of the enhanced electric field is about eight times higher than that of the incident wave. However, only four times enhancement was obtained from the interior insulator region (SiO2shell). It is suggested that the resonant energy would leak out to the outer surface of Ag@SiO₂/Ag NSs. Therefore, a well-controlled position of the analytic molecules adsorbed onto the Ag@SiO₂/Ag NSs is important to obtain the strongest SERS signals. To further verify the position related SERS effect, two kinds of SERS substrates with different R6G adsorbed positions were prepared: (1) Ag@SiO2/Ag NSs directly dipped into 10 ⁶M R6G solutions for 1 h and dried in the air, and (2) Ag@SiO₂NSs were treated as before, followed by 10 nm Ag coating, (as shown schematically in the inset of Fig. 4d). Strong SERS signals, with distinctive wavenumbers at 1363 and 1511 cm ¹ of R6G, arose exclusively while the probed molecules were adsorbed on the outer surface of Ag outer layer (curve i), being consistent with the three- dimensional finite difference time domain simulation. For R6G molecules positioned on the interior silica shell, relatively weak SERS signals were obtained (curve ii). Similar results were obtained in Fig. S5, with p-Tc in a solution (10 ⁶M) as an analyte. The SERS spectra using two different analytes all suggest that the SERS signals were highly dependent on the analyte position in the Ag@SiO₂/Ag NSs, and it confirmed that the probed molecules adsorbed on the outer surface of Ag@SiO2/Ag NSs possessed the best SERS effect.

For the Ag@SiO2/Ag NSs, it is also essential to control the nanostructures of both SiO₂and Ag outer layer. In order to study the direct influence of their thicknesses on the SERS performance, the

Ag outer layer was first varied with a certain thickness of SiO2. SEM images and Raman spectra of Ag@SiO₂/Ag NSs with different Ag coating thicknesses were compared (refer to Figs. S6 and S7).

The Raman spectra displayed clearly characteristic vibrations of R6G’s main peaks at 1363 and 1511 cm ¹as a result of the carbon stretching modes, being in accordance with the reported work.^[32]

The intensities of Raman peaks belonging to p-Tc (1076 cm ¹, 1593 cm ¹) were plotted as a function of the Ag outer layer thickness (as shown in Fig. 5a). Obviously, the SERS activities corresponding to each peak were improved while the Ag outer layer thickness increased from 5 nm to 10 nm. However, the intensities greatly decayed with further Ag outer layer increase from 10 nm to 20 nm. Because the Raman intensities decreased exponentially with the increasing shell thickness,^[33]a relative thicker interior silica layer should attenuate the SERS signals. To demonstrate the SERS activity dependence on the thickness of the interior insulator, the Raman spectra of the Ag@SiO2/Ag nanosphere arrays with different Si CVD growth times (1, 3, 6, and 10 min) were compared (shown in Fig. 5a), while the Ag outer layer thickness was fixed to 10 nm. The comparison of Raman signals demonstrated that, with a 1-min Si CVD growth time, the 1-nm thick SiO2layer was the optimal interior insulator for the induced SERS activity. A thicker gap would weaken the plasmonic coupling between the inner Ag core and the Ag outer layer. Therefore, by this sequence of Raman measurements, the optimized Ag@SiO2/Ag SERS substrate was obtained with a 1-nm thick SiO₂layer being as interior insulator and with a 10 nm thick Ag outer layer coating.

Moreover, optical properties of Ag@SiO₂/Ag nanosphere arrays with different thicknesses of Ag outer layer coating were characterized by an ultraviolet–visible reflection spectroscopy (shown in Fig. 5b). The spectrum of the corresponding Ag@SiO2nanosphere Figure 4. (a) Raman spectra of 1μM R6G absorbed on surfaces of a (i) 15-nm thick Ag film deposited on Si wafer, (ii) dewetted Ag NPs, and (iii) Ag@SiO2core- shell NSs. (b) CV curves of bare Ag films (black line) and Ag@SiO2nanosphere arrays (red line), with Si substrate as a working electrode in a 1 M NaOH solution.

(c) Intensity (|E|) distributions obtained from three-dimensional finite difference time domain calculations of one Ag@SiO2/Ag nanosphere at a wavelength of 633 nm. (d) Raman spectra of 1μM R6G absorbed onto the different positions: (i) outer surface of Ag@SiO2/Ag NSs and (ii) outer surface of Ag@SiO2NSs, followed by a 10 nm Ag film coating.

12 04

(6)

arrays was also given as a reference, which had a sharp reflection trough at about 440 nm and a reflectance as low as 5%. Obviously, as the Ag outer layer thickness increased, the plasmonic reflection trough was gradually red-shifted from 466 nm to 503 nm. The red shift is mainly related to the size increase of the discrete small Ag NPs composing the Ag outer layer.^[34,35]However, when further increasing the thickness of the Ag outer layer to 20 nm, not a sharp reflection trough, but a relatively broad and strong reflection appeared in the 525 nm region. Additionally, another broad reflection band in the 448 nm region appeared already above a thickness of 15 nm. This is attributed to the change from a quasi-continuous Ag outer layer film to a continuous one.

In order to verify its high sensitivity as a SERS substrate, the Raman spectra from the optimized Ag@SiO2/Ag nanosphere arrays covered with different concentrations of R6G molecules (from 10 ⁹ to 10 ¹²M) are shown in Fig. 5c. Obviously, the Raman intensity of R6G decreased with lower R6G concentrations. The characteristic Raman peaks of R6G at 612, 1363, and 1511 cm ¹could still be identified down to a R6G concentration of 10 ¹²M. As a reference sample, the detection limit of R6G (p-Tc) on a plane Si substrate coated with the identical thickness of Ag film is only 10 ⁸M (10 ⁷M) (refer to Fig. S8). Therefore, the measured high sensitivity of 10 ¹²M is mainly attributed to the nanostructure of high-density Ag@SiO2/Ag nanosphere arrays. Moreover, the absolute SERS enhancement factor (EF) of the optimized Ag@SiO₂/Ag nanosphere arrays was quantified by considering the 1511 cm ¹Raman band.

In case of R6G, the total EF was estimated to be 8.0 × 10⁶(refer to supporting information). Furthermore, to confirm the applicability of the SERS substrate for various small molecules with high sensitivity, p-Tc was also adsorbed on the optimized SERS substrate. The Raman spectra for different p-Tc concentrations from 10 ⁸ to 10 ¹¹M are shown in Fig. S9. The SERS spectral feature of p-Tc

could be still identified even at a concentration as low as 10 ¹¹M, and the corresponding EF was estimated to be 1.4 × 10⁷by considering the strongest peak in the spectra at 1076 cm ¹.

Uniformity of the SERS substrate is also essential for industrial applications. The SERS signals of the optimized Ag@SiO₂/Ag nanosphere arrays were measured spot-to-spot with a step size of 4μm.

The SERS contour was plotted after the line mapping of 100 spots as shown in Fig. 5d, exhibiting a uniform enhancement of Raman signals of the R6G molecules. The optimized SERS substrate was also used to detect 10 ⁸M p-Tc in ethanol solution, and the Raman spectrum from the p-Tc is similar with that of p-Tc powder reported in literature.^[36] The corresponding p-Tc SERS contour was also obtained (as shown in Fig. S10), in which all the 100 spots exhibit strong and uniform SERS signals. To further estimate the uniformity of the SERS signals, the relative standard deviation (RSD) of the Raman intensities of 100 curves were statistically ana- lyzed (refer to Fig. S11). The values of RSD at the strongest peaks at 1363 and 1511 cm ¹are 15.0% and 14.3%, respectively, indicat- ing the high uniformity of the SERS substrate. The RSD less than 20% further demonstrated that the nanostructure of high-density Ag@SiO₂/Ag nanosphere arrays is suitable as a highly reproducible SERS substrate.^[37]For the R6G analyte, the RSD values of vibrations at 1076 and 1593 cm ¹were 12.7% and 12.8%, respectively, which further confirmed its generally high uniformity as a SERS substrate.

Conclusions

In summary, we have developed a bottom-up method to fabricate a high-quality SERS substrate based on Ag@SiO2/Ag core-shell NSs.

The SERS activity was modified by the tunable ultrathin SiO₂shell as an interior insulator. The optimization of such a SERS substrate Figure 5. (a) Plots of the Raman peak intensities at 1076; 1593 cm ¹(belongs to 1μM p-Tc molecules) as a function of Ag out-layer coating thickness and CVD growth time. (b) Ultraviolet–visible reflection spectra of the bare Ag@SiO2nanosphere arrays (black line), and with various Ag out-layer coating thicknesses (from 5 nm to 20 nm).(c) Raman spectra of the R6G concentrations ranging from 10 ⁹to 10 ¹²M using optimized Ag@SiO2/Ag nanosphere arrays as a SERS substrate. (d) SERS contours maps of the intensities of Raman signals from 100 different locations on the optimized SERS substrate of Ag@SiO2/Ag nanosphere arrays; the step size is 4μm.

12 05

(7)

depends on various factors, such as the density of NSs, thicknesses of SiO2shell and Ag out-layer coating, and position of probed molecules. All these factors were characterized with respect to the SERS performance. CV measurements confirmed that the optimal 1-nm thick SiO₂shell as an interior insulator could be synthesized pinhole free by our CVD process. Based on FDTD simulation and Raman measurement, the probed molecules adsorbed on the surface of the Ag outer layer coating can generate stronger Raman signals.

By the adjustment of the thicknesses of both the SiO₂shell and the Ag out-layer coating, the ultimately optimized SERS substrate performed with high sensitivity and large-area uniformity. The SERS detection limit was as low as 10 ¹²M (R6G) with the corresponding EF up to 3.6 × 10⁷. The reproducibility of the uniform SERS substrate was confirmed with small RSD, e.g. 15.0% at 1363 cm ¹and 14.3%

at 1511 cm ¹. The method opened up a new way to create a class of highly sensitive SERS sensors, which may play an important role in nanodevice design and the corresponding biological and food safety monitoring applications.

Acknowledgements

We acknowledge partial support by the financial support by the Na- tional Natural Science Foundation of China (grant no. 51202072), the International Cooperation Base of Infrared Reflection Liquid Crystal Polymers and Device (grant no. 2015B050501010), the Pro- gram for Changjiang Scholars and Innovative Research Team in University (grant no. IRT13064), the Guangdong Innovative Research Team Program (grant no. 2011D039), the Science and Technology Planning Project of Guangdong Province (grant no.

2014B090914004,no. 2015B090927006), the State Key Program for Basic Researches of China (grant no. 2015CB921202), the International Science and Technology Cooperation Platform Program of Guangzhou (grant no. 2014J4500016), the Guangdong National Science Foundation (grant no. 2014A030313434), the Project for Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme (2014), and the Pearl River S&T Nova Program of Guangzhou (2015). The authors would also like to acknowledge Mr Guohui Lin for the preparation of SEM samples.

References

[1] S. Nie, S. R. Emory,Science1997,275, 1102.

[2] H. Xu, E. J. Bjerneld, M. Käll, L. Börjesson,Phys. Rev. Lett.1999,83, 4357.

[3] W. Li, P. H. Camargo, X. Lu, Y. Xia,Nano Lett.2008,9, 485.

[4] Y. Tan, J. Gu, W. Xu, Z. Chen, D. Liu, Q. Liu, D. Zhang,ACS Appl. Mater.

Inter.2013,5, 9878.

[5] D. K. Lim, K. S. Jeon, H. M. Kim, J. M. Nam, Y. D. Suh,Nat. Mater.2010,9, 60.

[6] H. Liang, Z. Li, W. Wang, Y. Wu, H. Xu,Adv. Mater.2009,21, 4614.

[7] Z. Liu, Z. Yang, B. Peng, C. Cao, C. Zhang, H. You, Q. Xiong, Z. Li, J. Fang, Adv. Mater.2014,26, 2431.

[8] J. F. Li, Y. F. Huang, Y. Ding, Z. L. Yang, S. B. Li, X. S. Zhou, F. R. Fan, W. Zhang, Z. Y. Zhou, B. Ren,Nature2010,464, 392.

[9] X. D. Lin, V. Uzayisenga, J. F. Li, P. P. Fang, D. Y. Wu, B. Ren, Z. Q. Tian, J. Raman Spectrosc.2012,43, 40.

[10] X. D. Tian, B. J. Liu, J. F. Li, Z. L. Yang, B. Ren, Z. Q. Tian,J. Raman Spectrosc.2013,44, 994.

[11] V. Uzayisenga, X. D. Lin, L. M. Li, J. R. Anema, Z. L. Yang, Y. F. Huang, H. X. Lin, S. B. Li, J. F. Li, Z. Q. Tian,Langmuir2012,28, 9140.

[12] J. R. Anema, J. F. Li, Z. L. Yang, B. Ren, Z. Q. Tian,Annu. Rev. Anal. Chem.

2011,4, 129.

[13] D. M. Dotzauer, S. Bhattacharjee, Y. Wen, M. L. Bruening,Langmuir 2009,25, 1865.

[14] B. Q. Zhang, S. B. Li, Q. Xiao, J. Li, J. J. Sun,J. Raman Spectrosc.2013,44, 1120.

[15] J. A. Huang, Y. Q. Zhao, X. J. Zhang, L. F. He, T. L. Wong, Y. S. Chui, W. J. Zhang, S. T. Lee,Nano Lett.2013,13, 5039.

[16] M. W. Shao, M. L. Zhang, N. B. Wong, D. D. Ma, H. Wang, W. Chen, S. T. Lee,Appl. Phys. Lett.2008,93, 233118.

[17] M. S. Schmidt, J. Hübner, A. Boisen,Adv. Mater.2012,24, OP11.

[18] H. Mao, W. Wu, D. She, G. Sun, P. Lv, J. Xu,Small2014,10, 127.

[19] M. Chirumamilla, A. Toma, A. Gopalakrishnan, G. Das, R. P. Zaccaria, R. Krahne, E. Rondanina, M. Leoncini, C. Liberale, F. De Angelis,Adv.

Mater.2014,26, 2353.

[20] J. Shen, Y. Zhu, X. Yang, J. Zong, C. Li,Langmuir2012,29, 690.

[21] J. J. Feng, U. Gernert, P. Hildebrandt, I. M. Weidinger,Adv. Funct. Mater.

2010,20, 1954.

[22] P. Dong, Y. Lin, J. Deng, J. Di,ACS Appl. Mater. Inter.2013,5, 2392.

[23] D. K. Lim, K. S. Jeon, J. H. Hwang, H. Kim, S. Kwon, Y. D. Suh, J. M. Nam, Nat. Nanotechnol.2011,6, 452.

[24] W. Stöber, A. Fink, E. Bohn,J. Colloid Interface Sci.1968,26, 62.

[25] Z. Hu, Z. Liu, L. Li, B. Quan, Y. Li, J. Li, C. Gu,Small2014,10, 3933.

[26] L. Liu, M. Jin, Q. Zhou, R. Zhan, H. Chen, X. Gao, S. Senz, Z. Zhang, J. Liu, RSC Adv.2015,5, 19229.

[27] C. M. Müller, F. C. F. Mornaghini, R. Spolenak,Nanotechnology2008,19, 485306.

[28] J. Gao, L. Guo, J. Wu, J. Feng, S. Wang, F. Lai, J. Xie, Z. Tian,J. Raman Spectrosc.2014,45, 619.

[29] J. F. Li, S. Y. Ding, Z. L. Yang, M. L. Bai, J. R. Anema, X. Wang, A. Wang, D. Y. Wu, B. Ren, S. M. Hou,J. Am. Chem. Soc.2011,133, 15922.

[30] T. Itoh, Y. Sasaki, T. Maeda, C. Horie,Surf. Sci.1997,389, 212.

[31] S. l. Chen, B. l. Wu, C. Cha,J. Electroanal. Chem.1997,420, 111.

[32] T. Wang, Z. Zhang, F. Liao, Q. Cai, Y. Li, S.-T. Lee, M. Shao,Sci. Rep.2014, 4, 4052.

[33] N. R. Honesty, A. A. Gewirth,J. Raman Spectrosc.2012,43, 46.

[34] P. Farzinpour, A. Sundar, K. Gilroy, Z. Eskin, R. Hughes, S. Neretina, Nanotechnology2012,23, 495604.

[35] H. Dai, M. Li, Y. Li, H. Yu, F. Bai, X. Ren,Opt. Express2012,20, A502.

[36] J. Chen, W. Yang, K. Dick, K. Deppert, H. Xu, L. Samuelson, H. Xu,Appl.

Phys. Lett.2008,92, 093110.

[37] M. J. Natan,Faraday Discuss.2006,132, 321.

Supporting information

Additional supporting information may be found in the online version of this article at the publisher’s web site.