Introduction

Metallic nanoparticles have fascinated researchers for several decades because of their wide range of potential applications (Patolsky et al. 2004; Narayanan and El-Sayed 2005; Park et al. 2007; Aslan et al. 2007; Le et al. 2008; Ashayer et al. 2010; Nghiem et al. 2010; Pyne et al. 2011; Chu et al. 2011; Sun et al. 2014a, b). The properties of metallic nanoparticles are critically dependent on their composition, shape, geometry (Link and El-Sayed 2000; Jackson et al. 2003; Sosa et al. 2003; Cui et al. 2006; Karvianto and Chow 2012). In previous works, various morphologies have been developed, such as rods, cubes, triangles, cages, and so on (Busbee et al. 2003; Shankar et al. 2005; Chen et al. 2007; Li et al. 2010; Liu et al. 2013; Sun et al. 2014a, b). In particular, metal nanoshell is a subject with intensive experimental and theoretical studies in last two decade since the first report about Au/SiO2 nanostructures by Oldenburg et al. (1998). The silica-metal nanoshells exhibit tunable surface plasmon resonances ranging from visible to infrared region by varying the composition, core size, and shell thickness. (Oldenburg et al. 1998; Pham et al. 2002; Lu et al. 2003; Phonthammachai et al. 2008; Liang et al. 2011; Gomez et al. 2012; Brito-Silva et al. 2013; Nghiem et al. 2013) Furthermore, the metal nanoshells have been widely used for plasmon-exciton coupling (Fofang et al. 2008, 2011), catalysis (Deng et al. 2007), chemical sensing (Raschke et al. 2004), biosensor (Shen et al. 2012), drug delivery (Bikram et al. 2007), photothermal therapy (Coughlin et al. 2014), surface-enhanced Raman scattering (SERS) (Jackson and Halas 2004; Zhang and Guo 2009), and all-optical magnetic recording (Mayergoyz et al. 2009).

Surface plasmon resonance (SPR) of metal nanoshells strongly enhances the local electric field, which can be significantly improved by multishell coating. For instance, Da-Jian Wu et al. studied the plasmon coupling in a three-layered Au/SiO2/Ag nanoshell by means of Mie scattering theory and found strong near-field enhancements in infrared region (Wu et al. 2011). Zhao’s group studied the SPR properties of Au–Ag bimetallic multishell nanostructures theoretically and employed this nanostructure for SERS and surface-enhanced fluorescence (SEF) (Zhu et al. 2012). Sun’s and Yao’s groups prepared Fe3O4/Au/Ag multi-core/shell nanoparticles and used them for SERS and magnetic separation (Zeng and Sun 2008; Han et al. 2012). Furthermore, Shen et al. prepared Fe3O4@Ag/SiO2/Au core–shell microspheres that display long-range plasmon transfer of Ag to Au, the novel multifunctional microspheres have proved to be efficient SERS-active substrate (Shen et al. 2013). The nonlinear optical responses of gold nanoshells also have been studied in recent years (Gordel et al. 2014); however, the nonlinearity optical properties enhanced by the plasmonic multishells have been seldom explored.

In this article, we described a facile method to prepare water-soluble silver-coated gold nanoshells. Tunable plasmon band arise from various silver thicknesses have been achieved. SERS activities and third-order nonlinear optical properties of Au nanoshells and Ag-coated Au nanoshells were obtained and compared for investigating the local field enhancement in the gap region between the Au shells and the Ag coating. It was found that the Raman intensity of 1,4-BDT was enhanced 17 times while the maximum effective nonlinear refractive (NLR) coefficient was enhanced by 30 %.

Experimental section

Materials

All chemicals were used as obtained without further purification. Potassium carbonate (K2CO3, 99.0 %), aqueous ammonia (NH3·H2O, 25–28 %), polyvinylpyrrolidone (PVP, average Mw–55,000), anhydrous toluene (99.8 %), silver nitrate (AgNO3, 99.8 %), l-ascorbic acid (99.7 %), hexadecyltrimethylammonium (CTAC, 98.0 %), anhydrous ethanol, chloroauric acid (HAuCl4·3H2O, 99.999 %), and 1,4-benzenedithiol (1,4-BDT, 98.0 %) were purchased from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, China). Sodium borohydride (NaBH4, 98.0 %) and tetraethyl orthosilicate (TEOS, 99.0 %) were obtained from Aldrich (America). 3-aminopropyltrimethoxysilane (APTES, 99 %) was purchased from Aladdin. (Shanghai, China). Ultrapure water with a resistivity of about 18.25 MΩ cm was used as the solvent in all experiments.

Preparation of Au/SiO2 nanoshells

The Au/SiO2 nanoshells were prepared according to the method described by Alexandre and co-workers (Brito-Silva et al. 2013) with some modifications. Specifically, silica nanospheres were prepared by mixing NH3·H2O (28 %), anhydrous ethanol, and TEOS in a 50 mL plastic tube. The solution was sonicated for 2 h and the clear solution gradually turned opaque white due to the formation of silica suspension. The average diameter of silica spheres prepared by this method is about 110 nm. After centrifugation, the silica nanospheres solution was poured into a 100 mL round-bottom flask, and 2 mL of APTES was injected. The mixture was mingled under vigorous magnetic stirring at 110 °C for 12 h. Then, the amino-terminated silica nanoparticles (silica–NH2 nanoparticles) were centrifuged at 6000 rpm for 13 min for three times. To prepare aqueous solution of gold colloids (2–3 nm in diameter), fresh prepared ice-cold NaBH4 aqueous solution was poured into the mixture of PVP and HAuCl4 rapidly, and stirring vigorously for 15 min. The resulting solution was stored in the refrigerator at 4 °C.

To link gold colloids on silica nanosphere, 30 mL of gold colloids was mixed with 5 mL of the as-prepared silica–NH2 particles in a 100 mL glass beaker. The mixture was magnetic stirred for 2 h and centrifuged at 5500 rpm for 12 min. The precipitate was redispersed in 3 mL ultrapure water by sonication and mixed with 30 mL gold colloids under magnetic stirring for 2 h. The process was repeated four times to ensure the adsorption of AuNPs was completely accomplished. The precipitate was washed by water to remove the unabsorbed gold colloids and redispersed in 2 mL of ultrapure water. The resulting solution (denoted by silica-Au precursor) was stored in a dark brown glass tube at 4 °C.

To prepare the growth solution, 50 mg of K2CO3 was dissolved in 200 mL of ultrapure water in a 250 mL beaker. After shaking gently, 3 mL of 20 mM HAuCl4 was added and gently swirled. The resulting solution was stored in dark at room temperature and used within 2–5 days.

To grow Au nanoshells, different volumes of silica–Au precursor were injected into 5 mL of the growth solution. After the solution was mixed by agitation, 150 µL of 10 mM freshly prepared ice-cold NaBH4 was added dropwise. The solution changed from colorless to green within 10 min, indicating the formation of Au nanoshells. After 2 h incubation, the products (named by Au/SiO2 nanoshells) were centrifuged at 3800 rpm for 8 min and redispersed in ultrapure water for further use.

Preparation of silver-coated gold nanoshells (Ag/BDT/Au/SiO2 nanoshells)

5 mL of the as-prepared gold nanoshell aqueous solution was mixed with 5 mL of 20 µM 1,4-BDT ethanol solution. The incubation time was kept for 1 hour in order to facilitate conjugation of 1,4-BDT onto the surface of gold nanoparticles. The resulting solution was centrifuged at 3800 rpm for 8 min and redispersed in 5 mL of ultrapure water (denoted by BDT/Au/SiO2 nanoshells). The silver-coated gold nanoshells (Ag/BDT/Au/SiO2 nanoshells) were prepared via reduction of AgNO3 on gold by ascorbic acid. First of all, solution A and solution B were prepared. Solution A is formed by mixing 0.1 M ascorbic acid and 0.2 M CTAC together (the volume ratio is 3:2). Solution B is the aqueous solution of 0.01 M AgNO3. 2 mL of solution A and certain volume of solution B were injected drop by drop into the 5 mL of Au nanoshells solution simultaneously. The mixture was stirred for 3 h at room temperature. The resulting solution was centrifuged at 3400 rpm for 7 min and redispersed in 5 mL of ultrapure water.

Characterization

Transmission electron microscopy (TEM) images were taken on a JEOL 2010 HT transmission electron microscope at an accelerating voltage of 200 kV. Absorption spectra were taken on a TU-1810 UV–Vis–NIR spectrophotometer (Purkinje General Instrument Co. Ltd. Beijing, China). Dynamic light scattering (DLS) measurements were performed with a Zeta sizer (Nano ZS90, Malvern Instruments, UK) at 25 °C. Raman spectra were obtained using a Confocal Raman Microspectroscopy (RM-1000), the excitation wavelength was 514.5 nm from an Ar+ laser. Nonlinear optical properties of the obtained nanoshells were characterized using a home-made femtosecond Z-scan experiment. The excitation laser of 100 fs pulse with 76 MHz repetition rate was generated by a mode-locked Ti: sapphire laser (Coherent, Mira 900).

Results and discussion

Nanostructures and plasmon resonances of Ag/BDT/Au/SiO2 nanoshells compositions

As schematically shown in Fig. 1, the preparation of the Ag/BDT/Au/SiO 2 nanoshells can be divided into four steps: (i) the synthesis of silica–Au precursor by the modification of Au colloids outside the silica–NH2 nanoparticles; (ii) the formation of gold shell onto SiO2 core by the epitaxial growth of Au shells; (iii) the conjugation of 1,4-BDT onto Au/SiO2 nanoshells; and (iv) the coating of silver shell with different thicknesses onto BDT/Au/SiO2 nanoshells.

Fig. 1
figure 1

A schematic illustration of the fabrication process of the samples

Full size image

The formation of Au nanoshells onto SiO2 core was firstly introduced by Halas group (Oldenburg et al. 1998). The synthesis, properties, and applications of Au nanoshells have been studied for decades. The SiO2 cores were firstly modified by the small Au colloids (silica–Au). Then, the epitaxial growth of Au shells was performed by injecting the silica–Au precursors into the growth solution. For tuning the thickness and plasmon resonance band of the Au/SiO2 nanoshells, different volumes of silica-Au precursors (2.5–30 μL) was injected into 5 ml growth solution (Fig. 2). TEM images of three different growth stages illustrate the progression in metal nanoshells growth. The low-magnification TEM images demonstrate that the samples are well dispersed. The SiO2 and Au could be distinguished easily because the contrast is obvious. The gold colloids covered the silica cores partially when the precursor is 15 μL (Fig. 2b, e) and the average thickness of gold shell layer is about 12 nm. As the volume of silica–Au precursor decreases to 4 μL, the gold network becomes more densely interconnected (Fig. 2c, f) and the average thickness of gold shell layer is about 29.5 nm. The nanoshells are nearly completed and the SiO2 is hard to see except few of small pores when the silica-Au precursor is 2.5 μL and a uniform gold shell with thickness of 42 nm is formed (Fig. 2d, g). The controllable thickness of gold nanoshells results in tunable surface plasmon extinction bands in a wide range of wavelengths. As the gold nanoparticles coverage on the surface of SiO2 increases, the plasmon resonance extinction spectra broaden and the main peak red shifts. Once the grown gold reaching a complete shell, the extinction spectra red shifts to a longer wavelength (902 nm). The similar phenomenon was observed and was explained by Averitt and the coworkers (Averitt et al. 1997).

Fig. 2
figure 2

a Extinction spectra of Au nanoshells growth on silica nanospheres with various precursor volumes. b–g Corresponding TEM images of the Au/SiO2 nanoshells. The volumes of silica–Au precursor are b, e 15 μL, c, f 4 μL, and d, g 2.5 μL

Full size image

Due to the existence of –SH group, 1,4-BDT molecules could be easily bonded onto the Au/SiO2 nanoshells. The extinction spectra of Au/SiO2 nanoshells show slight red-shift after the conjugation with 1,4-BDT (the absorption of 1,4-BDT is in the range of 250–350 nm). The as-prepared BDT/Au/SiO2 nanoshells are used as seeds for the further growth of Ag shell. Due to the similar lattice constant between Au and Ag, the nucleation of Ag could be easily occured on the Au surface to form the external metallic shell. The thickness of Ag shell could be controlled by the amount of AgNO3 in the reaction mixture. Various Ag/BDT/Au/SiO2 nanoshell samples were synthesized and their extinction spectra were recorded, as presented in Fig. 3. Three representative TEM images from these samples are also shown in Fig. 3. Figure 3b and e is the TEM images of the initial BDT/Au/SiO2 nanoshells, and they present maximum extinction in the long wavelength side (832 nm). Some pores that are not covered by Au nanoparticles are apparent in the TEM images. After the coating of Ag, the plasmon band exhibits blue-shift with enhanced extinction intensity, and an extinction shoulder is appeared at around 400 nm corresponding to Ag. The dielectric function of silver is different from gold; therefore, the variation of effective dielectric function accounts for the blue-shift phenomenon (Liu and Guyot-Sionnest 2004). After the coating of silver, the Au shells are hard to distinguish but the surface morphology of nanoshells becomes smoother. As shown in the TEM images, the pores in the starting BDT/Au/SiO2 nanoshells gradually disappear and the overall diameter of nanoparticles become larger (the average diameter increase from 192.27 to 311.39 nm) and irregular. The Ag coverage increased from Fig. 3c, reaching the larger and more irregular nanoshells in Fig. 3d and g. Meanwhile, their SP extinction peak blue shifts to 600 nm and the band at 400 nm red shifts a little to 410 nm with the intensity being much stronger.

Fig. 3
figure 3

a Extinction spectra of Ag/BDT/Au/SiO2 nanoshells synthesized by different volumes of AgNO3. b–g Corresponding TEM images of Ag/BDT/Au/SiO2 nanoshells prepared by the addition of AgNO3 with different volumes: b, e 0 µL; c, f 250 µL; d, g 500 µL

Full size image

Meanwhile, DLS measurement was performed in order to determine the main particle size and the polydisperse property of the samples (see Fig. 4). From the results of DLS measurement, we can see that the average hydrodynamic diameter and polydisperse index (PDI) of Au/SiO2 nanoshells sample shown in Fig. 2g are about 210 nm and 0.095 and that of the Ag/BDT/Au/SiO2 nanoshells sample shown in Fig. 3f are about 300 nm and 0.241, which indicates that the samples have a good dispersion and the results are coincident with that measured by TEM.

Fig. 4
figure 4

a The size distribution of (a) Au/SiO2 nanoshells sample shown in Fig. 2g and b Ag/BDT/Au/SiO2 nanoshells sample shown in Fig. 3f determined by DLS

Full size image

Linear optical response with enhanced Raman signals

It is reported by Xu that the electric field could be enhanced in the gap region between a Ag core and a Ag shell (Xu 2005). In the present work, we use 1,4-BDT molecules located in the gap regions as Raman analytes to demonstrate the intense local field confined in the gaps between the Au and Ag layers. Figure 5 shows the Raman signals of 1,4-BDT modified Au/SiO2 nanoshells before and after the Ag deposition. Since the two types of nanoshells are derived from the same BDT/Au/SiO2 nanoshells sample, the coverage and density of 1,4-BDT are the same. It should be noted that all the samples labeled with 1,4-BDT are must be cleaned by centrifugation at least two times to discard the unbound molecule. The SP absorption bands of these two samples are 832 nm (BDT/Au/SiO2 nanoshell) and 600 nm (Ag/BDT/Au/SiO2 nanoshell). The Raman patterns corresponding to 1,4-BDT in Fig. 5 agree well with the previous report (Joo et al. 2001). The strongest peak at 1567 cm−1 attributed to the benzene ring mode was chosen for the calculation of enhancement factor of SERS. The Raman intensity of the Ag/BDT/Au/SiO2 nanoshells is almost 17 times against than that from the BDT/Au/SiO2 nanoshells. The large enhancement indicates that the region between Au and Ag layer may behave as a cavity to concentrate electromagnetic field. The results suggest that the Ag/BDT/Au/SiO2 nanoshells are more suitable than BDT/Au/SiO2 for using as SERS probe. In addition, it is known that the size of the nanoparticles is an important parameter of the field enhancement which has been studied in detail (Sant’Ana et al. 2009; Amendola and Meneghetti 2012; Shaw et al. 2013). Because the average size of Ag/BDT/Au/SiO2 nanoshells is larger than that of BDT/Au/SiO2 nanoshells, the size enlargement will also lead to the SERS enhancement. Furthermore, the excitation wavelength also is a possible factor for the enhanced SERS signal, because the plasmon band of Ag/BDT/Au/SiO2 nanoshells is closer to 514.5 nm than that of Au/SiO2 nanoshells.

Fig. 5
figure 5

SERS spectra of 1.0 × 10−5 mol L−1 1,4-BDT adsorbed on Au/SiO2 nanoshells nanoparticles before (black) and after (red) the addition of AgNO3. (Color figure online)

Full size image

Nonlinear optical response with enhanced nonlinear refraction

The field enhancement in these two samples used in Raman measurement can also be characterized using femtosecond Z-scan experiment. A mode-locked Ti:sapphire laser tuned to the resonance wavelength of the samples and delivering pulses of about 150 fs at a repetition rate of 76 MHz was focused into a 1-mm-thick quartz cuvette containing the sample solution. The photodegradation of the metal nanostructure under laser irradiation is a frequent problem and is more frequent for large particles. When the fluence of pulsed laser is larger than 200 mJ/cm2, the laser-induced fragmentation of colloidal nanoparticles appears (Fujiwara et al. 1999; Mafuné et al. 2002; Besner et al. 2006; Amendola et al. 2009). To avoid the photodegradation of nanoshells, the laser power density was used as low as 0.63 GW/cm2 for the Z-scan measurements. Power-dependent refractive index n of materials could be described as n(I) = n 0 + γI, where n 0 represents the linear refractive index, γ is the nonlinear refraction coefficient, and I is the excitation laser power. Figure 6 presents the normalized T CL/T OP Z-scan curves related to γ of BDT/Au/SiO2 nanoshells and Ag/BDT/Au/SiO2 nanoshells samples with different excitation wavelengths. NLR index γ can be calculated by open-aperture transmittance T OP and close-aperture transmittance T CL as following relationships though the Z-scan theory:

$$ \frac{{T_{\text{CL}} }}{{T_{\text{OP}} }} = 1 + \frac{{4\Delta \phi_{0} {\raise0.7ex\hbox{$z$} \!\mathord{\left/ {\vphantom {z {z_{0} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${z_{0} }$}}}}{{\left[ {\left( {{\raise0.7ex\hbox{$z$} \!\mathord{\left/ {\vphantom {z {z_{0} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${z_{0} }$}}} \right)^{2} + 9} \right]\left[ {\left( {{\raise0.7ex\hbox{$z$} \!\mathord{\left/ {\vphantom {z {z_{0} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${z_{0} }$}}} \right)^{2} + 1} \right]}}, $$

where Δϕ 0 = γkI 0 L eff. I 0 is the peak irradiance at the focus (z = 0). L eff is the effective thickness of the samples, k = 2π/λ is the wave vector of the laser radiation, z 0 is the Rayleigh length of the Gaussian incident beam.

Fig. 6
figure 6

Normalized T CL/T OP curves of BDT/Au/SiO2 nanoshells nanoparticles before (black) and after (red) the addition of AgNO3. The scattered dots are experimentally data while the solid lines are theoretically fitting curves. (Color figure online)

Full size image

Self-defocusing effect curves of both BDT/Au/SiO2 nanoshells and Ag/BDT/Au/SiO2 nanoshells represent the negative values of NLR index. The values of γ in BDT/Au/SiO2 nanoshells and Ag/BDT/Au/SiO2 nanoshells samples are calculated to be about −0.7 × 10−4 and −0.91 × 10−4 cm2/GW, indicating that the local field was enhanced due to the interaction between Au and Ag coating. In order to exclude the effect of the 1,4-BDT, we measured the Z-scan signals of 1,4-BDT with different concentrations and no obvious nonlinear signals were observed; moreover, the NLR index of nanoshells with or without 1,4-BDT is similar. This demonstrates that the NLR index increase as high as 30 % is mainly contributed to the enhancement of electromagnetic field. However, the field enhancement is much lower than that calculated by SERS, the reason may be that the large scattering of Ag/BDT/Au/SiO2 nanoshells decreased the effective irradiance energy and the equalization induced by the water solution.

Conclusion

In summary, we investigated enhanced linear and nonlinear optical responses of silver-coated gold nanoshells. The Au nanoshells were firstly grown on SiO2 spheres by injecting the silica-Au precursors into the K3CO3–HAuCl4 growth solution. The SPR could be tuned form 646 to 902 nm by adjusting the volume of silica-Au precursors. After the modification of 1,4-BDT molecules onto the surface of Au/SiO2 nanoshells, Ag layer was deposited by reducing AgNO3 with AA. The SPR of silver-coated gold nanoshells could be tuned from 832 to 600 nm by adjusting the amount of AgNO3. Compared with the BDT/Au/SiO2 nanoshells, the Raman intensity is enhanced by 17 times and the nonlinear refractive index is enhanced by 30 % for the Ag/BDT/Au/SiO2 nanoshells.