Periodic silver nanodishes as sensitive and reproducible surface-enhanced Raman scattering substrates

Abstract

Highly ordered silver nanodishes which consist of nanorings and a film were fabricated as surface-enhanced Raman scattering (SERS) substrates. Numerical simulation reveals the bottom film can dramatically enhance the local electromagnetic (EM) field in the ring cavity, due to the plasmonic interaction between the nanorings and the film. Raman results show that the nanodishes can produce about sevenfold stronger signal than the nanorings alone, in accordance with the theoretical simulation. The detection limit for Rhodamine 6G (R6G) in the order of 10⁻¹² M and the average relative standard deviation (RSD) of less than 12% indicate the excellent sensitivity and reproducibility of silver nanodishes. The SERS enhancement factor (EF) of R6G on the nanodishes was calculated to be 6.17 × 10⁷. For practical application, the silver nanodishes were also used to detect thiram, one dithiocarbamate fungicide that has been extensively used as a pesticide in agriculture. The detection limit of thiram molecules is as low as 1 × 10⁻⁷ M, which can meet the requirements for ultra trace detection of pesticide residues. The resulting substrate with high SERS activity, stability and reproducibility makes it a perfect choice for practical SERS detection applications.

---

1. Introduction

Surface-enhanced Raman scattering (SERS) is a powerful analytical tool which can greatly enhance the Raman signal of analytes and makes the ultra trace detection of chemical and biological molecules possible.^1,2 Electromagnetic (EM) mechanism and chemical mechanism are widely accepted as two main mechanisms accounting for such an enormous Raman enhancement.^3–5 The EM mechanism is closely related to the surface plasmon resonance (SPR), which leads to a giant amplification of the local electromagnetic field. It has been reported that the small gaps between two nanoparticles (NPs) or sharp edges of nanostructures, usually called “hot spots”, are essential to exciting SPR.⁶ Normally, the EM mechanism is considered as the predominant mechanism.

Since the discovery of SERS⁷ in 1974, a large variety of SERS substrates with high EM enhancement have been extensively explored.^8,9 NPs synthesized in solution with controllable shapes and sizes usually have high enhancement because of the large number of hot spots formed in their aggregations.^10–13 However, it is very difficult to control the homogeneity of the aggregations of NPs in these systems. Consequently, the hot spots are randomly distributed and will lead to a poor reproducibility. As we know, to realize the practical application of SERS technology, some important factors such as the reproducibility, uniformity and stability also have to be taken into account except for the large Raman enhancement. Periodic nanostructures comprised of ordered metallic NPs exhibit special advantages in reproducibility, as a result of the uniform morphology. Many periodic nanostructures such as nanorings,^14–16 nanobowls,¹⁷ nanopyramids¹⁸ and nanocrescents¹⁹ have been explored. In particular, nanorings are of great interest due to their tunable localized surface plasmon and the nanocavities which can hold other small nanostructures.²⁰ E-beam lithography has been extensively applied to fabricate periodic nanorings but the low throughput and high cost have limited its further use. Alternatively, nanosphere lithography (NSL) technique is a facile and cost-effective method for fabricating ordered nanostructures on a large scale.²¹

Herein, we prepared highly ordered silver nanorings and nanodishes by NSL technique and explored their applications in SERS, considering that the silver nanomaterials have the best SERS properties. Ye and coworkers have fabricated gold nanorings–SiO₂–gold film structure and found the SERS performances can get optimized by tuning the thicknesses of the SiO₂ spacer.²² However, the nanorings in their work are dispersed and unordered on the quartz substrate. The big average spacing indicates there is no inter-particle coupling between neighboring nanorings, so they can be treated as single ones.²³ Different from their work, the NPs we fabricated are close-packed, the junctions of adjacent nanorings are expected to generate dense hot spots and improve their SERS performances greatly. Furthermore, the nanodishes could be considered as nanorings conjoined with the film, which has been neglected by Ye et al. Numerical simulation reveals the plasmonic coupling between the nanorings and the film will occur even without a dielectric spacer, the film can enhance the local EM field in the cavity of the nanorings to a large degree. Rhodamine 6G (R6G) was chosen to examine the SERS performances of nanorings and nanodishes. The results show the nanodishes produce about sevenfold stronger signal than the nanorings alone, in consistent with the numerical simulation. Using the silver nanodishes as SERS substrates, R6G with a concentration down to 5 × 10⁻¹² M can be identified.

For practical applications, the as-fabricated nanodishes can be employed to enhance the Raman signals of thiram, one of dithiocarbamate fungicides that have been extensively used as pesticides in agriculture. It is highly toxic if inhaled because thiram is irritating to the eyes, skin and respiratory tract and can lead to serious skin and eye illnesses. Experimental results indicate the SERS detection limit of thiram from the silver nanodishes is 1 × 10⁻⁷ M (about 0.03 ppm), lower than the maximal residue limit of 7 ppm in fruit prescribed by the U.S. Environmental Protection Agency. The nanodishes exhibit superior sensitivity, reproducibility and stability on SERS, which shows a very promising and practical solution to ultratrace detection of pesticide residues.

2. Experimental section

2.1 Preparation of silver nanodishes

The fabrication procedure of silver nanodishes is schematically illustrated in Fig. 1. First of all, a silver film was deposited onto the silicon slice by magnetron sputtering. Then a monolayer of hexagonally packed silica spheres were fabricated on the silicon slice by the self-assembly technique (Fig. 1A) after hydrophilic treatment of the silver film. Next, the second silver film was sputtered on the sphere templates to form semi-shells by magnetron sputtering (Fig. 1B) and this structure is usually called AgFON (Ag film over nanospheres).²¹ The silver particles can pass through the spacing between the spheres and deposited onto the silicon slice. Subsequently, the silver film together with silica spheres were partially etched by Ar⁺ ion beam (Fig. 1C), during which a ring around the sides of the silica spheres were created due to the secondary sputtering of silver NPs. Finally, the sample was immersed into HF solution to dissolve the silica spheres and a large area ordered silver nanodishes were obtained, as presented in Fig. 1D. The detailed procedures are presented as follows.

Fig. 1 Schematic illustration of the procedure for fabricating silver nanodishes.

2.1.1 Deposition of silver films. The silicon slices of 2 cm × 2 cm were immersed in boiling piranha solution (H₂SO₄/H₂O₂ 7 : 3 v/v) to obtain hydrophilism, followed by sonicate in a bath of acetone, ethanol, and deionized water for 10 min each to get them thoroughly clean and dried with nitrogen. Then the processed silicon slices were placed at the center of a magnetron sputtering chamber without any heating. Silver films with a thickness of 150 nm were directly deposited on the silicon slices and sphere templates. For sputtering, argon was used as the working gas and the airflow was 20 sccm (standard cubic centimeter per minute). The power of the magnetron was 50 W (voltage of 410 V and current of 100 mA). During deposition, the pressure in the magnetron chamber was 5 × 10⁻¹ Pa and the distance between the target and substrate was 10 cm.

2.1.2 Fabrication of the silica sphere templates. The Ag film coated silicon substrate was immersed into 0.02 M L-cysteine for 1 hour to obtain hydrophilism and then washed by deionized water and dried with nitrogen. The silica spheres with a diameter of 600 nm were synthesized by the Stober method.²⁴ The monolayer of ordered silica spheres were prepared on the hydrophilic silicon and silver coated silicon slices by a modified self-assembly process.²⁵ Briefly, the 10 wt% aqueous suspension of silica microspheres were chemically modified with 1H,1H,2H,2H-perflourodecyltrichlorosilane to obtain hydrophobicity. Then a glass slide of 2.5 cm × 7.5 cm was tilted into a glass Petri dish, which was filled with 30 mL of water, constructing the system for fabricating the silica monolayer. A 10 μL silica suspension was dispersed onto the glass substrate and floats to the surface of the water. Some domains of the silica sphere monolayer were initially formed on the water surface, at this time a 2 wt% sodium dodecyl sulfate solution was introduced to consolidate the particles. After adequately standing, the silica nanospheres will spontaneously form an ordered hexagonally close-packed monolayer which was readily transferred to the silicon and silver coated silicon slices. The slices with silica sphere monolayer could be used for the next experiment after the evaporation of water.

2.1.3 Ion beam etching. An ion beam with a plate voltage of 500 V and current of 100 mA was incident on the sample surface at an angle of 90°. Argon was used as the working gases and the airflow was 5.5 sccm. During etching process, the pressure in the chamber was 1.2 × 10⁻² Pa.

2.2. Characterization

The morphologies of the samples were observed with field emission scanning electron microscope images (FESEM) on a JEOL JSM-6300F SEM with a primary electron energy of 10 kV. The Raman scattering spectra were recorded on a confocal microprobe Raman system (Renishaw, inVia) with an excitation wavelength of 532 nm and 50× magnification. During SERS measurements, the laser light was projected vertically onto the samples with a resultant beam diameter of about 2 μm. The laser power was 1 mW and the recording time was 2 s for each spectrum. Normal incidence of light was used in all SERS measurements, which enabled stable and reproducible SERS signals. For SERS detection of R6G, a droplet of 10 μL ethanol solution of R6G was spread out on the substrates, and dried in air. For SERS detection of thiram, thiram was firstly dissolved in dimethylformamide followed by dilution with ethanol to form solution of different concentrations, then the substrates were immersed into solution of thiram overnight to make sure the binding between the analyte and substrates. Before SERS measurement of thiram, the sample was taken out and washed by ethanol, then dried in air.

3. Results and discussion

3.1 SEM images of the silver nanorings and nanodishes

The SEM images shown in Fig. 2 demonstrate the results of each step described above. Fig. 2A shows the typical SEM image of a large area of the regular AgFON substrate fabricated using 600 nm silica spheres, which are hexagonally close-packed and highly ordered on the silicon slice. Fig. 2B is an enlarged SEM image of the AgFON, from which we can see the deposited silver film is homogeneous without big silver islands. The etched AgFON is shown in Fig. 2C, demonstrating the silver rings have sharp edges. The average height of the silver rings and remained silica spheres are measured to be 268 and 560 nm respectively, indicating the silica spheres have also been partially etched. Fig. 2D shows the typical SEM images of the as-obtained nanorings, it can be seen that the thickness of the ring is about 60 nm. It is interesting that the nanorings can be readily folded into multilayers, which could be considered as three-dimensional (3D) nanonets (Fig. S1, ESI†). Fig. 2E shows the corresponding nanodishes. Compared with the nanorings, the two samples are uniform in size and the difference of them is the silver film under the nanorings. The silver film surrounded by the ring can be considered as a bottom plate, so we call them nanodishes visually. The oblique view of the nanodishes is shown in Fig. 2F.

Fig. 2 SEM images of (A) hexagonal close-packed AgFON, (B) enlarged SEM image of AgFON, (C) oblique view of etched AgFON, (D) silver nanorings and (E and F) silver nanodishes. The inset is close-up view.

3.2 Numerical simulations of the silver nanorings and nanodishes

Based on the three dimensional finite difference time domain (3D-FDTD) solution, simulations were introduced for determining EM field around silver NPs. The basic principle of FDTD is to numerically solve Maxwell's differential equations. In FDTD approach both space and time are divided into discrete segments, on the basis of which, electric fields and magnetic fields are calculated alternately step by step at sequenced time and space segments. Models in this work were constructed according to SEM images shown in Fig. 2. We need to clarify that the dimension used in the simulation is from the sectional views of the silver nanostructures. For the sake of simplicity, seven silver nanorings and nanodishes arranged hexagonally have been considered as representative. The wavelength was set to be 532 nm in all simulations and the radiation is assumed to be normal to the sample surface. Data of dielectric constants were from Kunz and Luebbers.²⁶ The Yee cell size in FDTD simulation was carefully considered to meet the accuracy needed by both wavelength and object parameters and to avoid too large memory resources and computation time required. Typical planar and side cross-sectional views of the calculated electric field intensity distributions of the nanorings and nanodishes are displayed in the Fig. 3A and B, respectively. The maximum EM field intensity is located at the tips and junctions of the nanorings in both cases. Interestingly, when the nanorings were fabricated on the silver film to form nanodishes (Fig. 3B), a homogeneous and strong electric field appeared inside the cavity of the silver nanodishes. Moreover, the EM field located in the junctions get enhanced distinctly, resulting from the LSPR of nanorings coupled with the induced surface plasmons of the silver film.²⁷

Fig. 3 Simulated EM field distribution maps of the (A) silver nanorings and (B) silver nanodishes.

3.3 SERS properties of the silver nanorings and nanodishes

To characterize the SERS performance of the proposed substrates, R6G was used as the probe molecule in our measurement because of its large Raman scattering cross-section.²⁸ The laser power was 1 mW and the recording time was 2 s for each spectrum. Fig. 4 shows the SERS spectra of R6G adsorbed on each substrate. Distinctive SERS spectra were obtained from the prepared substrates and the peak positions agree well with the literature for R6G molecules adsorbed on silver nanostructures.²⁹ Vibrations at 1186, 1309, 1360, 1509 and 1649 cm⁻¹ are assigned to C–H in-plane bending, C–O–C stretching, and C–C stretching of the aromatic ring.³⁰ Whereas the peak at 611 cm⁻¹ is due to the in-plane deformation vibration of the ring and the peak at 773 cm⁻¹ corresponds to the out-of-plane bending motion of the hydrogen atoms of the xanthene skeleton.^31,32 As shown in Fig. 4, it is clear that the nanorings (spectrum b) have a comparable SERS performance to the regular AgFON substrate (spectrum c) though the silver content is less. Previous studies about the MFON (metal film over nanospheres) nanostructures have shown that the enhanced EM field is mainly confined at the edges and crevices between adjacent metallic semi-shells.³³ The exposed metallic surface only provides a small fraction of the major part of the enhancement and measured SERS signals. When the metallic semi-shells were etched by the Ar⁺ ion beam, most of the hot spots still exist in junctions of the remaining rings, which has been demonstrated by the EM field distribution maps in Fig. 3A. Besides, the EM field located at tips of the rings may compensate the lost SERS signal contributed by the exposed metallic surface of the silver semi-shell. As expected, the nanodishes (spectrum a) show the strongest SERS signal, which is about seven times stronger than the nanorings alone. Because the nanodishes were composed of nanorings and a silver film, the same SERS detection was conducted on sputtered silver film with the same experimental parameters as a comparison. However, almost no information of the probe molecules adsorbed on the silver film can be obtained even with laser power of 10 mw and collection time of 60 s, the weak signal in spectrum d in Fig. 4 is the corresponding SERS signal after magnified by 100 times. Therefore, the stronger SERS signal must be contributed to the coupling effect of the silver nanorings with the silver film, in accordance with the numerical simulations. The strong EM field in the ring cavities of silver nanodishes is extremely beneficial to SERS detection. Because when the R6G solution was dropped on the substrates, only a small number of R6G molecules were adsorbed on the tips and junctions though the strongest local field occurred there. Most of the R6G molecules will achieve in the cavity and make the main contribution to the whole SERS signal. The SERS detection of folded nanorings was not carried out in our experiment because it is difficult to control their formation into uniform nanonets yet. Nevertheless, this novel nanonets are worthy of investigation because the 3D nanostructures can provide a larger surface area for the adsorption of analytes and potentially have a higher density of hot spots, the correlational research are underway.

Fig. 4 SERS spectra of 5 × 10⁻⁶ M R6G adsorbed on (a) silver nanodishes, (b) silver nanorings, (c) AgFON and (d) silver film.

3.4 SERS signal sensitivity and the reproducibility of the nanodishes

Subsequently, the silver nanodishes were chosen to check the SERS signal sensitivity and the reproducibility using R6G. Fig. 5A shows the Raman spectra of R6G with different concentrations (ranging from 5 × 10⁻⁷ M to 5 × 10⁻¹² M) adsorbed on the substrate. Clearly, the detection concentration for R6G on the nanodishes is in the order of 10⁻¹² M. Although the intensity of the characteristic peaks of R6G (5 × 10⁻¹² M) is weak, the signal to noise ratio is still very high. The average enhancement factor was calculated by comparing the normalized SERS intensity of the band at 611 cm⁻¹ of 5 × 10⁻⁶ M R6G solution and the normal Raman intensity of the R6G powder (Fig. S2, ESI†). According to the standard equation EF = (I_SERSN_bulk)/(I_bulkN_SERS), the typical enhancement factor for the as-fabricated substrate was estimated to be 6.17 × 10⁷ (details are given in ESI†). The reproducibility could be a significant evaluation criterion for a SERS substrate. Fig. 5B shows the SERS spectra of 5 × 10⁻⁶ M R6G obtained at 9 random positions across a 4 cm² sample for the silver nanodishes. It is evident that the Raman counts are quite consistent from place to place across the sample surface. The corresponding relative standard deviation (RSD) values at peak 611, 773, 1360, 1509 and 1649 cm⁻¹ are 8.7%, 9.9%, 12%, 14.3% and 12.1%, respectively. The average RSD value lower than 12% indicates the good SERS reproducibility of the nanodishes. Moreover, this reproducibility of the SERS substrate was also proven by 2D point-by-point SERS mapping of R6G molecules with the Raman peaks at 611 and 1360 cm⁻¹ (Fig. S3, ESI†). The integration time for each spot was 1 s. From the mapping images, we can see that the SERS substrate displays uniform SERS activity over a large area of 40 μm × 40 μm by measuring 1600 points with a regular scanning step of 1 μm.

Fig. 5 (A) SERS spectra of R6G with different concentrations adsorbed on silver nanodishes from 5 × 10⁻⁷ M to 5 × 10⁻¹² M. (B) SERS spectra of R6G (5 × 10⁻⁶ M) adsorbed on the nanodishes at 9 different spots.

3.5 SERS detection of thiram from silver nanodishes

The silver nanodishes were further applied for the detection of trace thiram, one of dithiocarbamate fungicides that are widely used as insecticidal agents on food crops. It is an irritant for the eyes, skin and respiratory tract that will cause serious illnesses. Detection of trace quantities of dithiocarbamates is necessary in order to mitigate potential human exposure via pesticide residues. Many methods have been reported for the determination of thiram, including chromatography, UV-Vis spectrophotometry, enzyme linked immunosorbent assay and chemiluminescence analysis.^34–37 However, these methods usually either involve complicated and time-consuming test procedures or require expensive devices. The analytical method based on SERS technique is promising as it can be used for rapid detection of trace pesticide.³⁸ As reported in the literature, thiram molecule has a S–S bond which will spontaneously breaks upon when exposed to the silver NPs and binds to the silver surface through the Ag–S bond, providing the best possibility for its identification and detection by SERS techniques.³⁹ To ensure the sufficient adsorption and bonding of thiram molecules on substrates, the silver nanodishes were immersed into solution of thiram overnight, followed by washing in ethanol and drying in air. Fig. 6A shows the Raman spectra of thiram in ethanol with concentrations decreasing from 1 × 10⁻⁵ to 1 × 10⁻⁷ M using the nanodishes as the substrate. The strongest peak at 1386 cm⁻¹ is attributed to the CN stretching mode and symmetric CH₃ deformation mode. The peak at 444 cm⁻¹ is due to the CH₃NC deformation and C=S stretching mode. Raman bands including 564 cm⁻¹ are attributed to the S–S stretching modes and 931 cm⁻¹ to the stretching CH₃N and C=S modes respectively. The CN stretching vibrations and rocking CH₃ mode occur at 1150 cm⁻¹ and 1514 cm⁻¹.³⁹ It can be seen that the SERS detection limit of thiram from the silver nanodishes is 1 × 10⁻⁷ M (about 0.03 ppm), lower than the maximal residue limit of 7 ppm in fruit prescribed by the U.S. Environmental Protection Agency. The calibration curve Fig. 6B reflects the relationship between SERS intensity at 1386 cm⁻¹ and the logarithm of the concentrations of thiram in ethanol, providing a diagnostic signature for the quantitative determination of thiram in solution. SERS spectra for 1 × 10⁻⁶ M thiram were collected on five points of silver nanodishes (Fig. S4, ESI†). All the vibrational bands with similar peak intensity for each measurement were observed, which validates the reliability of the substrate in SERS detection of thiram.

Fig. 6 (A) SERS spectra of thiram with various concentrations from 1 × 10⁻⁵ M to 1 × 10⁻⁷ M detected form silver nanodishes. (B) The intensities of SERS signals at 1386 cm⁻¹ as a function of the logarithm of thiram in ethanol. Each datum indicated an average from three samples, and each error bar indicated the standard deviation.

4. Conclusions

In summary, silver nanorings and nanodishes were fabricated as efficient SERS substrates based on NSL technique. The nanodishes can produce about sevenfold stronger SERS signal than the nanorings. Both theoretical and experimental studies demonstrate that this enormous enhancement is due to the plasmonic coupling between the nanorings and the underlying film. The average RSD value less than 12% from six main bands of R6G manifests the good reproducibility of the nanodishes. The detection limit for R6G and thiram were estimated to be 5 × 10⁻¹² and 1 × 10⁻⁷ M, respectively. The outstanding sensitivity was attributed to the hot spots resulting from the junctions among adjacent nanorings, the sharp tips of every single nanorings and the homogeneous local field in the dish cavity. It is believed that the silver nanodishes can serve as ideal substrates for SERS applications and provide an excellent candidate for SERS analysis, as well as in the construction of nanodevices and so forth. Further optimization of the nanodishes and other applications are under active investigation.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (no. 61378038), the National Basic Research Program of China (2011CB302103). We also acknowledge for the support of State Key Laboratories of Transducer Technology.

References

1. 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, D. Y. Wu, B. Ren, Z. L. Wang and Z. Q. Tian, Nature, 2010, 464, 392–395.
2. K. Kneipp, H. Kneipp and J. Kneipp, Acc. Chem. Res., 2006, 39, 443–450.
3. M. Moskovits, Rev. Mod. Phys., 1985, 57, 783–826.
4. A. Campion and P. Kambhampati, Chem. Soc. Rev., 1998, 27, 241–250.
5. X. T. Wang, W. S. Shi, G. W. She and L. X. Mu, J. Am. Chem. Soc., 2011, 133, 16518–16523.
6. T. Jensen, L. Kelly, A. Lazarides and G. C. Schatz, J. Cluster Sci., 1999, 10, 295–317.
7. M. Fleischm, P. J. Hendra and A. J. Mcquilla, Chem. Phys. Lett., 1974, 26, 163–166.
8. M. J. Banholzer, J. E. Millstone, L. D. Qin and C. A. Mirkin, Chem. Soc. Rev., 2008, 37, 885–897.
9. M. K. Fan, G. F. S. Andrade and A. G. Brolo, Anal. Chim. Acta, 2012, 693, 7–25.
10. N. R. Jana, L. Gearheart and C. J. Murphy, Adv. Mater., 2001, 13, 1389–1393.
11. J. L. Song, Y. Chu, Y. Liu, L. L. Li and W. D. Sun, Chem. Commun., 2008, 1223–1225.
12. M. Rycenga, X. H. Xia, C. H. Moran, F. Zhou, D. Qin, Z. Y. Li and Y. N. Xia, Angew. Chem., Int. Ed., 2011, 50, 5473–5477.
13. H. F. Lu, Z. W. Kang, H. X. Zhang, Z. L. Xie, G. H. Wang, X. Yu, H. Y. Zhang, K. T. Yong, P. Shumc and H. P. Ho, RSC Adv., 2013, 3, 966–974.
14. M. G. Banaee and K. B. Crozier, Opt. Lett., 2010, 35, 760–762.
15. Z. Q. Sun, Y. Li, J. H. Zhang, Y. F. Li, Z. H. Zhao, K. Zhang, G. Zhang, J. R. Guo and B. Yang, Adv. Funct. Mater., 2008, 18, 4036–4042.
16. F. Hao, Y. Sonnefraud, P. Van Dorpe, S. A. Maier, N. J. Halas and P. Nordlander, Nano Lett., 2008, 8, 3983–3988.
17. M. J. Xu, N. Lu, H. B. Xu, D. P. Qi, Y. D. Wang and L. F. Chi, Langmuir, 2009, 25, 11216–11220.
18. C. H. Sun, N. C. Linn and P. Jiang, Chem. Mater., 2007, 19, 4551–4556.
19. N. Vogel, J. Fischer, R. Mohammadi, M. Retsch, H. J. Butt, K. Landfester, C. K. Weiss and M. Kreiter, Nano Lett., 2011, 11, 446–454.
20. J. Aizpurua, P. Hanarp, D. S. Sutherland, M. Kall, G. W. Bryant and F. J. G. de Abajo, Phys. Rev. Lett., 2003, 90, 05740.
21. L. A. Dick, A. D. McFarland, C. L. Haynes and R. P. Van Duyne, J. Phys. Chem. B, 2002, 106, 853–860.
22. J. Ye, M. Shioi, K. Lodewijks, L. Lagae, T. Kawamura and P. Van Dorpe, Appl. Phys. Lett., 2010, 97, 163106.
23. P. H. C. Camargo, M. Rycenga, L. Au and Y. N. Xia, Angew. Chem., Int. Ed., 2009, 48, 2180–2184.
24. W. Stober, A. Fink and E. Bohn, J. Colloid Interface Sci., 1968, 26, 62–69.
25. C. Li, G. S. Hong, P. W. Wang, D. P. Yu and L. M. Qi, Chem. Mater., 2009, 21, 891–897.
26. K. Kunz and R. Luebbers, The Finite Difference Time Domain Methods for Electromagnetics, CRC press, New York, 1993.
27. K. M. Byun, S. J. Kim and D. Kim, Opt. Express, 2005, 13, 3737–3742.
28. E. C. Le Ru, E. Blackie, M. Meyer and P. G. Etchegoin, J. Phys. Chem. C, 2007, 111, 13794–13803.
29. W. C. Lin, L. S. Liao, Y. H. Chen, H. C. Chang, D. P. Tsai and H. P. Chiang, Plasmonics, 2011, 6, 201–206.
30. L. L. Sun, Y. H. Song, L. Wang, C. L. Guo, Y. J. Sun, Z. L. Liu and Z. Li, J. Phys. Chem. C, 2008, 112, 1415–1422.
31. H. Watanabe, N. Hayazawa, Y. Inouye and S. Kawata, J. Phys. Chem. B, 2005, 109, 5012–5020.
32. H. X. Xu, E. J. Bjerneld, M. Kall and L. Borjesson, Phys. Rev. Lett., 1999, 83, 4357–4360.
33. C. Farcau and S. Astilean, J. Phys. Chem. C, 2010, 114, 11717–11722.
34. B. Gupta, M. Rani and R. Kumar, Biomed. Chromatogr., 2012, 26, 69–75.
35. S. Rastegarzadeh, N. Pourreza and A. Larki, Spectrochim. Acta, Part A, 2013, 114, 46–50.
36. A. L. Queffelec, F. Boisde, J. P. Larue, J. P. Haelters, B. Corbel, D. Thouvenot and P. Nodet, J. Agric. Food Chem., 2001, 49, 1675–1680.
37. A. Waseem, M. Yaqoob and A. Nabi, Luminescence, 2010, 25, 71–75.
38. X. T. Wang, W. S. Shi, G. W. She, L. X. Mu and S. T. Lee, Appl. Phys. Lett., 2010, 96, 053104.
39. J. S. Kang, S. Y. Hwang, C. J. Lee and M. S. Lee, Bull. Korean Chem. Soc., 2002, 23, 1604–1610.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ra45935g

---