An ultra-sensitive SERS (surface-enhanced Raman scattering) substrate was fabricated orderly by depositing 360 nm Ag film on the surface of V-shaped AAO (anodized aluminum oxide), utilizing NaOH solution to remove the AAO template, the neat volcano-like Ag arrays substrate (N-V-Ag) was obtained, and then depositing 1.5, 6 and 10 nm Ag film on the surface of the N-V-Ag to obtain the AgNPs decorated volcano-like Ag arrays substrates (AgNPs-V-Ag-1.5, AgNPs-V-Ag-6 and AgNPs-V-Ag-10, respectively). Experimental results indicated that the cavity resonance mode (super-radiant bright modes) and the rim mode as well as “NP mode” (subradiant dark mode) have strong interferences, which results in a significant change in the distribution of the hot spots in the cavity compared with the N-V-Ag. The analytical enhancement factor (AEF) of AgNPs-V-Ag could reach up to 1.5 × 1011, improved by 5 orders of magnitude compared with the N-V-Ag. This novel substrate could obtain extremely low limits of detection (LOD) of 10−13 mol/L (-13 M) for Rhodamine 6G (R6G).
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Su ttering (SERS) is a non-destructive and sensitive detection technology [1,2], even at a single-molecule level . At present, the mechanism of SERS effect can be attributed to two widely accepted theories , one is the electromagnetic enhancement (EM), and another is the chemical enhancement (CE). The former holds the dominant position which provides enormous EF up to 104–107  due to localized surface plasmon resonance (LSPR) from the nanogap between the noble metallic nanostructures, called “hot spots” [6–8]. However, traditional SERS-active substrates have either high LOD but uncontrollable hot spots which decrease repeatability and uniformity , or low LOD. Hence, many advanced nanofabrication technologies have been applied in design of patterned substrates with controllable nanogaps, such as electron beam lithography (EBL) [10,11], focused-ion-beam (FIB) [12,13], etc., but these top-view methods are often high-cost, complicated and time-consuming [14–16], and more importantly, they can’t be exploited in large-areas fabrication. Hence, the fabrication of cost-efficient and large-area uniform substrate with high sensitivity is still critical for the practical SERS applications [17,18].
Porous anodized aluminum oxide (AAO) was widely used in large-area and high-ordered arrays for SERS substrate due to highly uniform porous structures, which has attracted extensive interests in recent years [19–21]. Guochao Shi et al.  developed an AAO/Al-based nanoflower-like SERS platform for the detection of goat serum, results of this substrate showed great uniformity and reproducibility, and exhibited enormous potential in biosensing. Long Liu et al.  proposed an excellent substrate of plasmonic nanopillars on plastic film fabricated by nanoimprint lithography using AAO template, the averaged EF was up to 1.4 × 108 and it showed an excellent reproducibility with a relative standard deviation (RSD) of 18%. These two representative single-pass reverse mold technologies used in SERS application are obviously more practical than double-pass AAO-based technology proved by our previous experiments.
In this work, we have successfully fabricated a novel volcano-like Ag arrays using a reverse mold technology. The edges and bottom of the volcano-like Ag nanostructure have a strong enhanced local electric field which provide characteristically enhanced Raman signals. Furthermore, in order to improve the low limits of detection (LOD) of this substrate, we utilized thermal evaporation method to add additional Ag nanoparticles into the nanogaps of the volcano-like Ag nanostructures. A finite element method (COMSOL Multiphysics) was used to simulate the distribution of the near-fields of this novel structure. Moreover, the SERS activities were investigated with Rhodamine 6G (R6G) as probe molecules.
The V-shaped AAO was purchased from Shenzhen Top Membranes Technology Co., Ltd. Rhodamine 6G (R6G) was used as a probe molecule (95%, Shanghai, Aladdin). Sodium hydroxide solid powder (AR, 96%) was purchased from Chuandong Chemical Co., Ltd. (Chongqing). Sylgard 184 (PDMS) was obtained from Dow Corning Company (2%, DC184, Shanghai). Acrylic plates were used as substrates instead of Si wafer (Shihua Acrylic Company, Chengdu). Deionized water (18.25 MΩ) was used to prepare the solutions throughout the experiment.
2.2 Fabrication of SERS substrates
The N-V-Ag was fabricated by using a V-shaped AAO template assisted thermal evaporation method, shown in Fig. 1. Firstly, PDMS solution was prepared at a ratio of 10:1 using Sylgard 184, Dow Corning, then mixed and placed in the air for 3-5 hours. Secondly, AAO template was cut into a size of 2 mm × 2 mm, and then 360 nm Ag film was deposited on the surface of AAO template by thermal evaporation. Thirdly, the acrylic plate was put on the spin coater and dripped 60 µL PDMS on the sample. The parameter of spin coater was set at 1500 r/s for 3 minutes in the first stage, 3000 r/s for 2 minutes in the second stage. Then we stuck the Ag film-coated sample face down on the surface of the acrylic plate, and the sample was placed in a vacuum oven for 10 hours. Fourthly, 8% NaOH solution was configurated: 22 g sodium hydroxide solid powders were weighted by electronic balance (Yueping instrument Co., Ltd. Shanghai) and put into the glass, and then 250 mL deionized water was added in it, the solution was mixed evenly to dissolve the hydroxide solid powder completely. Finally, the samples were put into the NaOH solution for about 11 hours. After the AAO template was removed, the samples (N-V-Ag) were washed with deionized water for 2-3 times. Moreover, based on the above steps, the AgNPs-V-Ag were prepared by AgNPs self- assembly technology, the thickness of the thermally evaporated Ag film was 1.5, 6 and 10 nm, respectively.
2.3 SERS measurements
For the test of SERS sensitivity, 47.9 mg R6G powders were weighed and dissolved in 10 mL deionized water to obtain the R6G solution with concentration of −2 M (10−2 mol/L), and the mother liquor was diluted step by step to obtain from −14 to −4 M R6G solution. Prior to detect the analyte, substrates were sealed in a vacuum bag to isolate oxygen. For the test, the corresponding concentration of 12 µL R6G solutions were dropped on the surface of the substrates using a micro-pipettor and dried at ambient environment naturally. Then the samples were analyzed using the confocal Raman spectrometer system (Horbia Jobin Yvon LabRAM HR Evolution), the laser wavelength is 532 nm, with laser power of 0.5 mW, the spot diameter of 1 µm. The objective lens was selected as × 50 with a numerical aperture of 0.75. The spectral resolution is 1 cm−1, the acquisition time was adjusted to 10, 3, 1, 0.2, 0.1 s according to the various R6G concentration of −8, −7, −6, −5, −4 M for the N-V-Ag, and the acquisition time was 10 s for the AgNPs-V-Ag-1.5, AgNPs-V-Ag-6, AgNPs-V-Ag-10. The number of repetitions for each spectrum were five, in order to eliminate random noise and gain averaged data. The baseline of spectral data was removed by the LabSpec 6 software that comes with the confocal Raman microscope.
3. Results and discussion
3.1 Characterization of the AAO template, N-V-Ag and AgNPs-V-Ag
The JSM-7800F (JEOL) field-emission scanning electron microscope (SEM) was used to characterize the surface morphology of the AAO template, N-V-Ag and AgNPs-V-Ag, and the top view, cross-section view and 60° dip view have been investigated.
Figures 2(a) and 2(c) show a top view and cross-section view morphology of the AAO template, respectively. From the top view morphology, we can observe that the large-scale and quasi-hexagonal nanoporous arrays exhibited excellent uniformity and periodicity. From the cross-section view morphology, as distinguished from the single-pass AAO template, the diameter of the top nanopore in the V-shaped AAO is larger than that of bottom nanopore, which is similar to an inverted nanocone, and the diameter of the top nanopore and bottom nanopore is 556 and 100 nm, respectively, and the height is about 1.52-1.60 µm. Figure 2(c) exhibits the highly ordered three-dimensional volcano-like Ag arrays, the mass percent of Ag is 80%. Figures 2(d1)-2(f3) exhibit the AgNPs-V-Ag-1.5, AgNPs-V-Ag-6, AgNPs-V-Ag-10, respectively. We can see quantities of Ag nanoparticles were formed when the thickness of deposited Ag film is between 1.5-10 nm, and the diameter of the Ag nanoparticles increased with the increase of the thickness of Ag film (measured by software Nano Measurer, based on SEM image). Meanwhile, the top region of the volcano-like Ag structure will gradually be filled, which can also be confirmed from the SEM images of 60° dip (as shown in Figs. 2(g1)-2(g3)). We can also see some Ag nanoparticles were absorbed on the sidewall of the volcano-like Ag nanostructures.
3.2 FEM modeling of the N-V-Ag and AgNPs-V-Ag
In order to investigate the electromagnetic field enhancement and analyze the distribution of “hot spots” around the N-V-Ag and AgNPs-V-Ag with quasi-photonic crystal nanostructures, a finite element method (COMSOL Multiphysics) was introduced. The N-V-Ag was simplified to a hollow circular table shape with a certain thickness, the top diameter was set at 216 nm, the bottom diameter was set at 472 nm, the height was set at 989.9 nm, the thickness of the bottom Ag film was set at 360 nm and the thickness of sidewall was set at 8 nm. The diameter of AgNPs formed by the thermal evaporation with different Ag film thickness (1.5, 6 and 10 nm) was corresponding to 15, 20 and 25 nm, respectively. The average gap distance between the adjacent AgNPs was set at 8.14 nm, AgNPs were placed on the top and sidewall of nanocavity, at the junction of the nanocavity and the Ag film and at the bottom of the nanocavity (Figs. 3(a1)-3(d1)). The simulation model used electromagnetic wave frequency domain interface and periodic port as excitation. Parameters mentioned above were all based on the SEM images (as shown in Fig. 2, calculated by software Nano Measurer). The incident laser is 532 nm, polarized at axis x, and transmitted at axis -z. The complex refractive index of Ag is 0.056206 - 4.2776·i . The geometric size scale of simulation figures was given in Figs. 3(a2)–3(a7), the remaining simulation figures in each line used the same size scale.
The EF values are calculated using the maximum surface electric field intensity of the N-V-Ag and AgNPs-V-Ag (Figs. 3(a1)–3(d1)), with 3.8 × 103, 2.2 × 1010, 1.4 × 1010, 8.0 × 1010 corresponding to the N-V-Ag, AgNPs-V-Ag-1.5, AgNPs-V-Ag-6, AgNPs-V-Ag-10, respectively.
Figures 3(a2)-3(d7) showed the electric field distribution of the N-V-Ag and AgNPs-V-Ag (AgNPs-V-Ag-1.5, AgNPs-V-Ag-6, AgNPs-V-Ag-10, respectively) in different xy-planes (z = 0 nm, z = 360 nm, z=989.9 nm, respectively), yz-plane and xz-plane. From which we can see that there are mainly three modes in the nanocavity, which are NP mode” , rim mode and cavity resonance mode. In order to further explain the three modes and interaction mechanisms, we simplified the physical model of AgNPs-V-Ag (as shown in Fig. 4(b4), AgNPs were placed on the top and the middle of the sidewall of nanocavity, at the junction of the nanocavity and the Ag film and at the bottom of the nanocavity) and highlighted the electric field distributions to discuss the interactions of three modes in detail.
Shown in Fig. 4(a1), a mode hybridization occurs between the cavity plasmon mode and the metallic sidewall plasmon mode [26,27], the interactions between two modes form a rim mode under normal incidence, which can be identified from Figs. 4(a2) and 4(a3). Shown in Fig. 4(b1), the enhanced local electric field is mainly distributed on the top and sidewall of the nanoarrays, as well as the junction region between the Ag film, the cavity and the bottom of cavity due to the interaction between “NP mode” , rim mode and cavity resonance mode. In general, the cavity resonance mode is considered as relatively broad resonances, called the super-radiant bright modes, the rim mode and “NP mode” is a subradiant dark mode and causes narrow resonance. When the two types of modes are close to each other, they have strong interference that causes strong Fano resonance , this effect can significantly improve the density and intensity of hotspot distribution compared with N-V-Ag. In addition, the dipole-dipole interaction is the cavity modes coupling energy back into the NP in a reciprocal process . It can be estimated by fm(λ)σ(λ), where σ(λ) is the measured scattering spectrum of the isolated volcano-like Ag arrays, and fm(λ) is the fraction of energy emitted , each NP on the sidewall of nanocavity can receive energy from its neighbors to achieve this coupling effect.
3.3 SERS sensitivity
With the abundant “hot spots” generated on the surface of the N-V-Ag and AgNPs-V-Ag, it would exhibit direct and rapid trace detection capabilities for R6G solution with different concentrations. In order to evaluate the SERS performance of our samples, we used analytical enhancement factor (AEF) as analysis method due to its form is simple and its measurement is easily reproducible, expressed by 
Figure 5(a) shows the Raman spectra of R6G with different concentrations varying from −8 M to −4 M for the N-V-Ag, it can be seen clearly that all the Raman characteristic peaks such as the peaks of 612, 773, 931, 1128, 1179, 1310, 1362, 1508, 1573 and 1649 cm−1 are consistent with standard characteristic peaks. It is difficult to identify R6G Raman peaks with concentration of −9 M, which illustrates the LOD of the N-V-Ag is −8 M, and the calculated analytical enhancement factor (AEF) is 2.2 × 106.
Figures 5(b) and 5(c) show Raman spectra of R6G with different concentrations varying from −13 M to −9 M for AgNPs-V-Ag-1.5 and AgNPs-V-Ag-6, the experimental AEF is 1.5 × 1011, 1.2 × 1011, respectively. The LOD of them are both −13 M. Obviously, the EF of the AgNPs-V-Ag-6 is lower, which would be due to the decreased SERS-sites with the diameter of AgNPs increased. Meanwhile, the local electric field of the junction region between the AgNPs and the bottom Ag film will be further weakened. The experimental EF is larger than that of simulation (see Section 3.2). The reason could be attributed to the higher roughness on the top of the N-V-Ag in real samples (as shown in Fig. 2(c)).
Figure 5(d) shows the photos of the N-V-Ag, AgNPs-V-Ag-1.5, AgNPs-V-Ag-6 and AgNPs-V-Ag-10, we can observe that the color of samples gradually changed from yellow to brown with the increase of the thickness of Ag film. It is noted that AgNPs-V-Ag-10 hasn't obtained effective Raman signal even at the concentration of −9 M, though its surface electric field intensity is higher than AgNPs-V-Ag-1.5 and AgNPs-V-Ag-6 (see Section 3.2), the reason can be attributed to the strong background fluorescence which was observed during the experimental process and the decreased effective SERS-sites with the increase of the diameter of AgNPs.
3.4 SERS signal reproducibility
The stability and reproducibility are two important properties for evaluating SERS performance. In practical substrates, the random distribution of hot spots caused by nonuniform morphology often reduces the stability and reproducibility of the SERS substrates. To further evaluate these properties of our prepared substrates, Raman Mapping characteristic intensities were measured to investigate the point-to-point reproducibility with a cover area of 10 µm × 10 µm (random choice) and a step size of 1 µm. The Raman intensity at 613 cm−1 fluctuated among 3188-7735 counts, with an averaged value of 5616 counts and calculated RSD (relative standard difference) of 8.4%, shown in Fig. 6(a) (−4 M R6G solution). The results suggested that the N-V-Ag has good uniformity and reproducibility.
Figures 6(b) and 6(c) present Raman spectra (−10 M R6G solution) of AgNPs-V-Ag-1.5 and AgNPs-V-Ag-6 (the intensity of all data has been enlarged to 8 times the original). The Raman intensity at 613 cm−1 of the substrate with Ag film thickness of 1.5 nm fluctuated among 361-1002 counts, with an averaged value of 708 counts and RSD of 23.1%. The Raman intensity of the substrate with Ag film thickness of 6 nm fluctuated among 553-1526 counts, with an averaged value of 1038 counts and RSD of 29.6%. According to the 3σ criterion of the normal distribution, all the spectral data outside the range of ± 3σ were removed.
We have developed a simple, low-cost, and high output method, based on the V-shaped porous AAO, to fabricate a highly ordered volcano-like Ag arrays, with the AEF of 2.2 × 106 and the RSD value of 8.4%, indicating that this substrate has good uniformity and reproducibility. Then the surface of sample was modified by second depositing 1.5, 6 and 10 nm Ag films, this method increased the AEF by 5 orders of magnitude, which could reach up to 1.5 × 1011 for AgNPs-V-Ag-1.5, the LOD was −13 M and RSD value was 23.1%. In addition, we have pointed out that the source of high Raman enhancement is the hybridization of different plasmon resonance modes, which are “NP mode”, rim mode and cavity resonance mode in this composite metallic volcano-like nanostructures.
National Natural Science Foundation of China (61875024); Chongqing Outstanding Youth Fund (cstc2019jcyjjqX0018); Fundamental Research Funds for the Central Universities (CQU2018CDHB1A07).
We would like to thank Dr. Gong Xiangnan at Analytical and Testing Centre of Chongqing University for his help in Raman measurement.
The authors declare no conflicts of interest.
1. M. Fleischmann, J. Hendra P, and J. McQuillan A, “Raman spectra of pyridine adsorbed at a silver electrode,” Chem. Phys. Lett. 26(2), 163–166 (1974). [CrossRef]
2. J. F. Li, Y. F. Huang, Y. Ding, Z. L. Yang, S. B. Li, X. S. Zhou, and Z. L. Wang, “Shell-isolated nanoparticle-enhanced Raman spectroscopy,” Nature 464(7287), 392–395 (2010). [CrossRef]
3. S. Nie and R. Emory S, “Probing single molecules and single nanoparticles by surface-enhanced Raman scattering,” Science 275(5303), 1102–1106 (1997). [CrossRef]
4. S. Cong, Y. Yuan, Z. Chen, J. Hou, M. Yang, Y. Su, Y. Zhang, L. Li, Q. Li, F. Geng, and Z. Zhao, “Noble metal-comparable SERS enhancement from semiconducting metal oxides by making oxygen vacancies,” Nat. Commun. 6(1), 7800–7806 (2015). [CrossRef]
5. X. Li, G. Chen, L. Yang, Z. Jin, and J. Liu, “Multifunctional Au-coated TiO2 nanotube arrays as recyclable SERS substrates for multifold organic pollutants detection,” Adv. Funct. Mater. 20(17), 2815–2824 (2010). [CrossRef]
6. H. Im, K. C. Bantz, S. H. Lee, T. W. Johnson, C. L. Haynes, and S. H. Oh, “Self-assembled plasmonic nanoring cavity arrays for SERS and LSPR biosensing,” Adv. Mater. 25(19), 2678–2685 (2013). [CrossRef]
7. K. Liu, Y. Bai, L. Zhang, Z. Yang, Q. Fan, H. Zheng, and C. Gao, “Porous Au–Ag nanospheres with high-density and highly accessible hotspots for SERS analysis,” Nano Lett. 16(6), 3675–3681 (2016). [CrossRef]
8. J. D. Caldwell, O. J. Glembocki, F. J. Bezares, M. I. Kariniemi, J. T. Niinistö, T. T. Hatanpää, and C. M. Hosten, “Large-area plasmonic hot-spot arrays: sub-2 nm interparticle separations with plasma-enhanced atomic layer deposition of Ag on periodic arrays of Si nanopillars,” Opt. Express 19(27), 26056–26064 (2011). [CrossRef]
9. Y. Zhou, X. Zhou, D. J. Park, K. Torabi, K. A. Brown, M. R. Jones, and C. A. Mirkin, “Shape-selective deposition and assembly of anisotropic nanoparticles,” Nano Lett. 14(4), 2157–2161 (2014). [CrossRef]
10. L. Petti, R. Capasso, M. Rippa, M. Pannico, P. La Manna, G. Peluso, and P. Musto, “A plasmonic nanostructure fabricated by electron beam lithography as a sensitive and highly homogeneous SERS substrate for bio-sensing applications,” Vib. Spectrosc. 82, 22–30 (2016). [CrossRef]
11. J. M. Oran, R. J. Hinde, N. Abu Hatab, S. T. Retterer, and M. J. Sepaniak, “Nanofabricated periodic arrays of silver elliptical discs as SERS substrates,” J. Raman Spectrosc. 39(12), 1811–1820 (2008). [CrossRef]
12. K. Sivashanmugan, H. Lee, C. H. Syu, B. H. C. Liu, and J. D. Liao, “Nanoplasmonic Au/Ag/Au nanorod arrays as SERS-active substrate for the detection of pesticides residue,” J. Taiwan Inst. Chem. Eng. 75, 287–291 (2017). [CrossRef]
13. K. Sivashanmugan, J.-D. Liao, and C.-K. Yao, “Nanovoids embedded in FIB-fabricated Au/Ag nanorod arrays for ultrasensitive SERS-active substrate,” Appl. Phys. Express 7(9), 092202 (2014). [CrossRef]
14. Y. Wei W and M. White I, “Inkjet-printed paper-based SERS dipsticks and swabs for trace chemical detection,” Analyst 138(4), 1020–1025 (2013). [CrossRef]
15. M. Šubr, M. Petr, O. Kylián, J. Kratochvíl, and M. Procházka, “Large-scale Ag nanoislands stabilized by a magnetron-sputtered polytetrafluoroethylene film as substrates for highly sensitive and reproducible surface-enhanced Raman scattering (SERS),” J. Mater. Chem. C 3(43), 11478–11485 (2015). [CrossRef]
16. J. Zhang, M. Irannejad, and B. Cui, “Bowtie nanoantenna with single-digit nanometer gap for surface-enhanced Raman scattering (SERS),” Plasmonics 10(4), 831–837 (2015). [CrossRef]
17. H. X. Lin, J. M. Li, B. J. Liu, D. Y. Liu, J. Liu, A. Terfort, and B. Ren, “Uniform gold spherical particles for single-particle surface-enhanced Raman spectroscopy,” Phys. Chem. Chem. Phys. 15(12), 4130–4135 (2013). [CrossRef]
18. Q. Jiwei, L. Yudong, Y. Ming, W. Qiang, C. Zongqiang, P. Jingyang, and X. Jingjun, “Fabrication of nanowire network AAO and its application in SERS,” Nanoscale Res. Lett. 8(1), 1–6 (2013). [CrossRef]
19. Z. Huang, G. Meng, Q. Huang, B. Chen, C. Zhu, and Z. Zhang, “Large-area Ag nanorod array substrates for SERS: AAO template-assisted fabrication, functionalization, and application in detection PCBs,” J. Raman Spectrosc. 44(2), 240–246 (2013). [CrossRef]
20. Z. Yi, X. Ye, J. Luo, X. Kang, Y. Yi, Y. Yi, and Y. Tang, “Ordered hexagonal nanoplasmonic au nanoparticle arrays: AAO-assisted thermal treatment synthesis and application as surface-enhanced Raman scattering substrates,” Plasmonics 12(6), 2013–2020 (2017). [CrossRef]
21. C. H. Huang, H. Y. Lin, S. Chen, C. Y. Liu, H. C. Chui, and Y. Tzeng, “Electrochemically fabricated self-aligned 2-D silver/alumina arrays as reliable SERS sensors,” Opt. Express 19(12), 11441–11450 (2011). [CrossRef]
22. G. Shi, M. Wang, Y. Zhu, X. Yan, S. Pan, and A. Zhang, “Nanoflower-like Ag/AAO SERS platform with quasi-photonic crystal nanostructure for efficient detection of goat serum,” Curr. Appl. Phys. 19(11), 1276–1285 (2019). [CrossRef]
23. L. Liu, Q. Zhang, Y. Lu, W. Du, B. Li, Y. Cui, and Y. Chen, “A high-performance and low cost SERS substrate of plasmonic nanopillars on plastic film fabricated by nanoimprint lithography with AAO template,” AIP Adv. 7(6), 065205 (2017). [CrossRef]
24. P. B. Johnson and R. W. Christy, “Optical Constants of the Noble Metals,” Phys. Rev. B 6(12), 4370–4379 (1972). [CrossRef]
25. F. M. Huang, D. Wilding, J. D. Speed, A. E. Russell, P. N. Bartlett, and J. J. Baumberg, “Dressing plasmons in particle-in-cavity architectures,” Nano Lett. 11(3), 1221–1226 (2011). [CrossRef]
26. Z. Zhu, B. Bai, O. You, Q. Li, and S. Fan, “Fano resonance boosted cascaded optical field enhancement in a plasmonic nanoparticle-in-cavity nanoantenna array and its SERS application,” Light: Sci. Appl. 4(6), e296 (2015). [CrossRef]
27. R. M. Cole, J. J. Baumberg, d. A. F. J. Garcia, S. Mahajan, M. Abdelsalam, and P. N. Bartlett, “Understanding plasmons in nanoscale voids,” Nano Lett. 7(7), 2094–2100 (2007). [CrossRef]
28. A. Christ, O. J. Martin, Y. Ekinci, N. A. Gippius, and S. G. Tikhodeev, “Symmetry breaking in a plasmonic metamaterial at optical wavelength,” Nano Lett. 8(8), 2171–2175 (2008). [CrossRef]
29. M. R. Philpott, “Effect of surface plasmons on transitions in molecules,” J. Chem. Phys. 62(5), 1812–1817 (1975). [CrossRef]
30. H. R. Stuart and D. G. Hall, “Enhanced dipole-dipole interaction between elementary radiators near a surface,” Phys. Rev. Lett. 80(25), 5663–5666 (1998). [CrossRef]
31. E. C. Le Ru, E. Blackie, M. Meyer, and P. G. Etchegoin, “Surface enhanced Raman scattering enhancement factors: a comprehensive study,” J. Phys. Chem. C 111(37), 13794–13803 (2007). [CrossRef]