A portable fiber SERS probe has been developed based on Ag nanorod array fabricated by oblique angle deposition. The incoming laser beam was designed to focus onto the Raman substrate at 45° incident angle in order to maximize surface enhanced Raman scattering signal. With a fiber Raman system, a detection sensitivity of 10-17 moles for trans-1, 2-bis(4-pyridyl)ethane molecules has been demonstrated. This Raman probe can also be used for in situ measurement for samples in aqueous solution. Such a fiber probe has great potential as a portable and remote sensor for on-site biological or chemical detection.
©2007 Optical Society of America
Surface-enhanced Raman spectroscopy (SERS) has been used as an analytical tool to observe trace amount of chemical and biological molecules due to its capability of giving real-time molecular vibrational information under ambient conditions [1, 2]. It has the advantages of using extremely small amount of analyte without amplification or manipulation of the samples, and the extremely short time frame for acquisition of the spectra. SERS requires minimal sample preparation and is non destructive to sample which allows real time analysis and great potential for multi components analysis. The well-known enhancement effect for SERS arises from either the adsorption or close proximity of an analyte onto a metal substrate . The morphology of the metallic structure plays a major role in determining the magnitude of signal enhancement and sensitivity of detection. Early SERS substrates included a random distribution of roughness features produced by oxidation reduction on a metal electrode  or evaporation of thin metal film on a flat substrate . Various forms of nanostructure have been explored to enhance SERS effects, such as rough metallic surfaces by chemical etching , silver films on TiO2 , colloidal silver nanoparticles , silver nanoparticle array fabricated by nanosphere lithography , electro-deposition of silver on silver films at high potential , aligned monolayer of silver nanowires . However, many of these methods are either expensive or time consuming, and it is not easy to make reproducible substrates of the correct surface morphology to provide maximum SERS enhancements. Without uniformity and good reproducibility of the metal substrates, the attainment of reproducible spectra remains a major challenge for SERS. We had previously demonstrated that silver nanorod substrate fabricated by oblique angle deposition (OAD) with length of ~868 nm, diameter of ~99nm, and tilt angle of 73° has achieved SERS enhancement factors of approximately 108 for the molecular probe trans-1,2-bis(4-pyridyl)ethane (BPE) [12,13]. The silver nanorod arrays fabricated by OAD generate large SERS response with maximum SERS intensity observed at around 45° incident angle. The maximum SERS intensity is about five times the intensity at the surface normal . Very recently, we also demonstrated that this SERS substrate can be used to distinguish between viruses and even different strains of viruses . Thus, the OAD technique offers an easy and inexpensive way for the fabrication of silver nanorod arrays for high sensitivity SERS applications. It can be easily implemented in the laboratory. The deposition procedure is straightforward and inexpensive. The SERS substrates produced by OAD have the advantages of uniformity and reproducibility.
Despite its extremely high and unique sensitivity, the application of SERS has not been incorporated into the development of practical in situ analytical tool for real-time sensing or detecting. Usually SERS measurement is carried out by a conventional Raman scattering spectrometer, which is bulky, expensive, and not easily accessible. For portable and remote applications, the ideal SERS sensor systems would be field-deployable, small and compact in size with very high sensitivity, selectivity and multiplexing ability. In addition, the need for easily fabricated SERS-active sensors that can be used in solution is essential for the application of SERS in in situ analyses. In this paper, we report our special design of a portable SERS probe with a slit sample holder which allowing the laser beam focus onto the sample with 45° incident angle, hence achieving the maximum SERS effect. Combining this custom designed SERS probe with a fiber Raman system, we demonstrated the SERS intensity is dependent on the concentration with a sensitivity of 10-17 moles on the OAD fabricated silver nanorod arrays using trans-1,2-bis(4-pyridyl)ethane (BPE) as the molecular probe. These results show the potential use of the SERS probe as a sensitive, remote and portable sensor for in situ SERS analyses.
Silver nanorod substrates were fabricated by OAD technique using a custom-designed electron beam/sputtering evaporation (E-beam) system (Torr International, New Windsor, NY) that has been described previously . Glass microscopic slides (Gold Seal® Catalog No.3010) were used as the base platform for silver nanorod arrays deposition. The glass slides were cleaned with Piranha solution before loading into the E-beam system. A base layer of 500 nm silver film was first deposited onto the glass slides before arrays of Ag nanorod (length of approximately 1 µm) was deposited on the silver base layer by OAD at a vapor incident angle of 86°. The deposition rate was 0.3 nm/s, and the deposition pressure was approximately 1×10-6 Torr. The film thickness was monitored by a quartz crystal microbalance positioned at normal incidence to the vapor source direction. Figure 1 shows a typical scanning electron microscope (SEM) image of silver nanorod surface. Average length of the silver nanorods was 868±95 nm and the average diameter was 99±29 nm. The average density of the nanorods was approximately 13±0.5 rods/µm2. The average tilting angle of the nanorods was ~73° with respect to the substrate normal.
The SERS detection system is shown in Fig. 2. The fiber Raman system used in this study was the HC-10HT Raman Analyzer (Enwave Optronics Inc., Irvine, CA). This system was made up of a diode laser, spectrometer, integrated Raman probe head for both excitation and collection, and separate excitation and collection fibers. The excitation source was a frequency stabilized, narrow linewidth near IR diode laser with a wavelength of 785nm. The excitation laser beam coupled to a 100 µm fiber was focused onto the substrate through the Raman probe head and was unpolarized at the sample. The focal length of the Raman probe was 6 mm, and the laser beam spot size was 100 µm. The Raman signal from the substrate was collected by the same Raman probe head and was coupled to a 200 µm collection fiber, which delivered the signal to the spectrometer equipped with a charge coupled device (CCD) detector [Fig. 2(a)].
We designed a portable silver nanorod substrate based SERS probe with a 45° slit to achieve the optimum experimental configuration and ensure maximum SERS scattering response from silver nanorod arrays as shown in Fig. 2(b). This cylinder shape custom-designed probe measures 7.4 cm in length and 2 cm in diameter. The front of the probe opens with a cylindrical hole that is 9.8 mm in diameter for connecting with the commercial Raman Analyzer. At 3.2 cm into the SERS probe, there is a glass sealed widow protecting the Raman head probe from liquid or vapor. Right next to the sealed window is the liquid or vapor cell with an inlet. Following the liquid cell is a slit to hold the substrate which makes the incoming laser beam focus onto the substrate at 45° incident angle to the substrate surface normal. The distance between the sealed window and the center of the silt is around 5.2 mm so that the laser beam could focus on the substrate. The silver nanorod substrates prepared by OAD technique were placed into the SERS probe through the open slit, facing the fiber optic Raman probe head. This arrangement allowed the focal length remained fixed and the investigated spot on the sample could be adjusted by moving the substrate along in the open slit. The tilting plane of the silver nanorods was parallel to the incident plane at 45° incident angle relative to surface normal [Fig. 2(c)]. Through the SERS fiber probe, SERS signal could be obtained by the Raman Analyzer.
The molecular probe used in this study was trans-1,2-bis(4-pyridyl) ethene (BPE, 99.9+%, Sigma). BPE solutions were prepared by sequential dilution in HPLC grade methanol (Aldrich). For each concentration, a 2µl drop of BPE solution was applied onto the silver nanorod substrate and allowed to dry before the acquisitions of data. SERS spectra were collected from multiple points across the substrate. The power at the sample surface was measured with a power meter (Thorlabs Inc., Newton, NJ) and was 52 mW (λ=785 nm).
3. Results and discussion
3.1 SERS probe detection limit of BPE
The SERS signals were acquired by the fiber Raman system at an excitation wavelength of 785 nm. Figure 3 shows the representative SERS spectrum of the as-grown bare silver nanorod substrate. A broad peak was seen around 1360 cm-1. Vapor deposited silver films and electrochemically reduced silver electrodes have been reported to exhibit backgrounds due to graphitic carbonaceous adsorption onto the substrate during deposition . The peak around 1360 cm-1, ascribed to disorder in the carbon chains , was probably contaminants from the fabrication process and/or storage in ambient environment. Such background signals are commonly found in SERS. Similar spectra from multiple spots on the same silver nanorods substrate or from different substrates showed no difference in peak position and the intensity remained unaffected under laser irradiation.
BPE was chosen as the molecular probe because of its high Raman scattering cross section, its ability to adsorb strongly and irreversibly onto a silver substrate, and its lack of resonant enhancement in the visible region . Besides being centrosymmetric, BPE is photo stable and exhibits very intense SERS spectra using relatively modest laser powers . In order to study the concentration of BPE in connection with SERS intensity, the SERS response of the silver nanorod substrate for nominal surface coverage of BPE was measured. Increasing amounts of BPE were consecutively applied onto the silver nanorod substrate and SERS spectra were collected for each application of BPE. Collecting multiple spectra from a single spot can cause loss of SERS intensity as a result of sample degradation or the molecules bleached on the spot. To avoid such possible artifacts in the data, spectra were obtained from 5 spots on the same substrate. One drop of 2µl BPE solutions with different concentrations, 10-7 M, 10-6 M, 10-5 M, 10-4 M and 10-3 M was subsequently applied onto the silver nanorod substrate, where it spread over a geometric sample area on the silver substrate measured a diameter of approximately 1.2 cm. The estimated BPE molecular coverage on the surface ranged from 1.5×10-4 (at 2µl of 10-7 M) to 1.5 (at 2µl of 10-3 M) monolayer, assuming 7×1014 BPE molecules per cm2 in a monolayer , thus, the amount of analyte excited in the laser spot were approximately from 1.4×10-17 to 1.4×10-13 moles. For the lowest concentration, it corresponds to ~84 BPE molecules/Ag nanorod. Figure 4 shows the SERS spectra of different amount of adsorbed BPE, from 1.4×10-17 to 1.4×10-16 moles, on the silver nanorod substrate. All spectra show the characteristic peaks of BPE at around 1200 cm- 1, 1610 cm-1 and 1640 cm-1, corresponding to ethylenic C=C stretching mode, pyridine ring C=C stretching, and the whole ring C=C stretching mode, respectively . The broad peak around 1360 cm-1 from substrate background remained unchanged during the study. Thus, it does not present as a significant interference peak in the BPE spectra. The results in Fig. 4 demonstrated that the integrated SERS probe could detect approximately 14 attomole of BPE on silver nanorod array prepared by the OAD method with good signal-to-noise ratio spectra.
Figure 5 shows the integrated band areas of the 1200 cm-1 peak plotted against the moles of BPE put down on the silver nanorod substrate in a log-log scale. The 1200 cm-1 peak of BPE was chosen for the quantification due to its relative insensitivity to molecular orientation on a silver surface . As shown in Fig. 5, the Raman intensity increases as the BPE molecules adsorbed on the Ag nanorod substrate increases. The SERS signal from BPE increase over 3 orders of magnitude (14 attomoles to 14 femtomoles), after which point the further increase in Raman intensity was barely remarkable. This reflects that the saturation of the intensity occurred between the estimated surface coverage of ~0.7 and 1.5 monolayer. It has been established that, on silver surface, maximum enhancement is observed when a monolayer of the adsorbate molecule is formed on the surface [21, 22]. According to the fast decay of the local electromagnetic field as moving away from the metallic surface, SERS spectra are optimized when a single layer of molecules is adsorbed on the substrate . Though not the dominant role, interadsorbate interactions have been found to play some role in determining the coverage dependence . Depending on the size and shape of the particles, interadsorbate interactions can either decrease or increase the intensity. Decrement in the SERS response often occurs at high surface coverage could be attributed to these intermolecular interactions. Though different model molecules and different substrate structures were used, the common observation in coverage dependence studies suggested the SERS intensity dependence can be established in a certain range of concentrations and surface coverage.
A comparison study of sensitivity on different SERS substrates reported detection limit between 270 and 0.4 femtomole for BPE. These differences depended on the substrates with the lowest detection of 0.4 femtomole found on vapor deposited annealed silver film (~5 nm) and a reduction in the SERS response occurred at coverage greater than ~0.01 monolayer . Another study reported approximately 50–200 µM detection limits for BPE on etched silver foil and vapor deposited silver film . In this study of SERS probe detection of BPE, our results showed that the portable probe has a sensitivity of 14 attomole for BPE. We also demonstrated the concentration dependent of Raman intensities of BPE on the OAD fabricated silver nanorod array. It is difficult to make direct comparisons between various substrates based on the literature since the available data were acquired with different instruments, exposure time and environmental conditions. However, unlike most SERS measurement carried out by a bulky Raman spectrometer which is not feasible for remote function, our SERS probe provides the advantage of easy implementation, mobility and compact size for real-time, on-site field application.
3.2 SERS signal of BPE as function of time
Response time and stability of the SERS signal with time are two important analytical parameters. In order to utilize our integrated SERS probe fiber Raman system for in situ measurement, the time evolution of the SERS signal of aqueous BPE on the silver nanorod substrate was studied. Distilled water was used as a solvent instead of methanol because methanol evaporates much faster than water; therefore, it is unsuitable for long observation time. A fresh silver nanorod substrate was first secured onto the SERS probe through the 45° slit. Then 50 µl of 10-6M BPE aqueous solution was applied to the substrate through the liquid inlet, the diameter of the sample area on the substrate submerged in the liquid cell was measured to be about 7 mm. The SERS spectra were collected at a 10 min interval as BPE accumulated on the surface over a period of 80 min. BPE solution was sucked out of the liquid inlet before the next higher concentration was added. Figure 6 shows the plot of the integrated band areas at 1200 cm-1 in the SERS spectra of BPE as a function of time.
The data shown in Figure 6 indicate the equilibrium of the BPE molecules between the water and the substrate was reached in 60 min after the addition of 10-6M BPE. Though a higher laser power was used to acquire higher intensity for the in situ measurement, the substrate was only exposed to the laser beam during the spectra acquisition to avoid photodecomposition of the analyte. Any substrate surface heating that may have been caused by the incident laser beam would have been dispersed instantly since the substrate was in contact with the liquid medium, which served as a heat sink. Furthermore, when any photodecomposition of the sample occurred at the substrate surface, the decomposed sample molecule could have diffused into the sample solution while being replaced by another BPE molecule from the solution. The increasing SERS signals over the first 80 min elapsed time indicate that BPE molecules were not thermally degraded with time. As expected, the Raman intensity increased when the higher concentration of BPE solution (10-5M) was added on the substrate. In 10 min after adding the BPE solution (10-5M), the SERS signal appeared to be on a slight decline.
In the study of SERS effect of BPE on gold particle arrays, Félidj et al reported that by immersing the array into 10-5M BPE solution, where single layer of absorbed molecules occurred, yielded the maximum Raman intensity . Increasing or decreasing this concentration in the solution lead to a weakening of the Raman signal. Assuming over time, BPE ultimately accumulated at the geometric surface area which the applied sample spread, the estimated BPE molecular coverage on the surface could have been as high as 1.1 monolayer (50 µl of 10-5M BPE, assuming 7×1014 BPE molecules per cm2 in a monolayer). In Fig. 6, the slight decline of the addition of BPE solution (10-5M) reflects the saturation of the Raman intensity as increasing BPE adsorbed on the surface and suggests that the single layer occurred. This result coincides with the observation from Félidj et al. The high surface coverage could explain the descent in the SERS response, considering the Raman enhancement arises mainly from the electromagnetic amplification of the local field that exponentially decreases as moving away from the surface. Another possibility for the slow decline of the SERS signal is that photodegradation of BPE molecules might have occurred at this point and caused the decay of the Raman signals. Moreover, some BPE molecules could have been lost through the dynamic equilibrium with surfaces, thus producing a smaller final concentration.
It is known that solvents induce morphology changes of silver island films. The change in surface tension at the silver surface, resulting from the adsorbed solvent, could greatly perturb the morphology. Roark et al. reported thin silver film morphology changes by dip coating solvent on the surface and found the intensity of the SERS of BPE adsorbed to the silver films increased . The mechanism by which solvent induced morphology changes appears to be associated with the mechanical aspect of dip-coating, as well as changes in the surface tension, but is not due to significant loss of metal from the substrate. Li et al found the formation of infinite regular silver rings on a thin silver island film after immersing in water for 30 min and then blow dried with nitrogen gas . As the water evaporated, capillary forces drew the nanorods together and contributed to the morphology changes on the substrate. In our study, nanorod array was immersed in the aqueous BPE solution throughout the measurement period. Sealing the liquid cell with plastic Parafilm wrap delayed the evaporation of water, therefore preventing the capillary force changing the morphology of silver nanorod substrate. We focus our attention on the in situ measurement of the SERS response of aqueous BPE on the silver nanorod substrate. Over all, our data shows the concentration dependent SERS signal of BPE on the silver nanorod substrate in an aqueous environment. Our special design of the SERS probe is an adequate sample holder for performing in situ analyses.
We have created a portable SERS probe which can be easily incorporated with a fiber Raman system. We demonstrated the concentration dependent of Raman intensities on the OAD fabricated silver nanorod array using BPE as molecular probe and observed a sensitivity of 14 attomole for BPE. Furthermore, we tested the response time and stability of the in situ SERS signal of BPE in a water solution on the silver nanorod substrate. Our present experiment indicates that SERS probe can be integrated into a fiber Raman system for in situ measurements and can act as a portable and remote sensor for accurate and rapid real-time SERS measurements. It has great potential as a multiplexing system for chemical and biological sensing, such as environmental pollution, chemical and biological warfare agent detection, virus or bacteria detection, etc. This is an important development for practical SERS applications because of the possibility of using low-power laser, inexpensive substrate and compact size sensor for field applications.
This research was supported by funding from the National Science Foundation (ECS0304340 and ECS070178) and the UGA Engineering Grant. The authors thank Ruth Ann Morrow for her feedback on the manuscript.
References and links
1. T. Vo-Dinh, “Surface-enhanced Raman spectroscopy using metallic nanostructures,” Trends Analyt. Chem. 17, 557–582 (1998). [CrossRef]
2. C. R. Yonzon, D. A. Stuart, X. Zhang, A. D. McFarland, C. L. Haynes, and R. P. Van Duyne, “Towards advanced chemical and biological nanosensors-An overview,” Talanta , 67, 438–448 (2005). [CrossRef]
3. M. Fleischmann, P. J. Hendra, and A. J. McQuillan, “Raman spectra of pyridine adsorbed at a silver electrode,” Chem. Phys. Lett. 26, 163–166(1974). [CrossRef]
4. A. A. Stacy and R. P. Van Duyne, “Surface enhanced raman and resonance raman spectroscopy in a non-aqueous electrochemical environment: tris(2,2′-bipyridine)ruthenium(II) adsorbed on silver from acetonitrile,” Chem. Phys. Lett. 102, 365–370 (1983). [CrossRef]
5. G. J. Kovacs, R. O. Loutfy, P. S. Vincett, C. Jennings, and R. Aroca, “Distance dependence of SERS enhancement factor from Langmuir-Blodgett monolayers on metal island films: evidence for the electromagnetic mechanism,” Langmuir 2, 689–694 (1986). [CrossRef]
6. K. T. Carron, X. Gi, and M. L. Lewis, “A surface enhanced Raman spectroscopy study of the corrosion-inhibiting properties of benzimidazole and benzotriazole on copper,” Langmuir 7, 2–4 (1991). [CrossRef]
7. L. M. Sudnik, K. L. Norrod, and K. L. Rowlen, “SERS-active Ag films from photoreduction of Ag+ on TiO2,” Appl. Spectrosc. 50, 422–424 (1996). [CrossRef]
9. T. R. Jensen, M. D. Malinsky, C. L. Haynes, and R. P. Van Duyne, “Nanosphere lithography: tunable localized surface plasmon resonance spectra of silver nanoparticles,” J. Phys. Chem. B 104, 10549–10556 (2000). [CrossRef]
10. G. Suer, U. Nickel, and S. Schneider, “Preparation of SERS-active silver film electrodes via electrocrystallization of silver,” J. Raman Spectrosc. 31, 359–363 (2000). [CrossRef]
11. A. Tao, F. Kim, C. Hess, J. Goldberger, R. He, Y. Sun, Y. Xia, and P. D. Yang, “Langmuir-Blodgett silver nanowire monolayers for molecular sensing using surface-enhanced Raman spectroscopy,” Nano Lett. 3, 1229–1233 (2003). [CrossRef]
12. S. B. Chaney, S. Shanmukh, R. A. Dluhy, and Y.- P. Zhao, “Aligned silver nanorod arrays produce high sensitivity surface-enhanced Raman spectroscopy substrates,” Appl. Phys. Lett. 87, 031908.1–3 (2005). [CrossRef]
13. Y.-P. Zhao, S. B. Chaney, S. Shanmukh, and R. A. Dluhy, “Polarized surface enhanced Raman and absorbance spectra of aligned silver nanorod arrays,” J. Phys. Chem. B 110, 3153–3157 (2006). [CrossRef] [PubMed]
14. Y. J. Liu, J. G. Fan, S. Shanmukh, R. A. Dluhy, and Y.-P. Zhao, “Angle dependent surface enhanced Raman scattering obtained from a Ag nanorod array substrate,” Appl. Phys. Lett. 89, 173134.1–3 (2006).
15. S. Shanmukh, L. Jones, J. Driskell, Y.-P. Zhao, R. Dluhy, and R. A. Tripp, “Rapid and sensitive detection of respiratory virus molecular signatures using a silver nanorod array SERS substrate,” Nano Lett. 6, 2630–2636 (2006). [CrossRef] [PubMed]
16. C. E. Taylor, S. D. Garvey, and J. E. Pemberton, “Carbon contamination at silver surfaces: surface preparation procedures evaluated by Raman spectroscopy and X-ray photoelectron spectroscopy,” Anal. Chem. 68, 2401–2408 (1996). [CrossRef]
17. Y. W. Alsmeyer and R. L McCreery, “Surface-enhanced Raman spectroscopy of carbon electrode surfaces following silver electrodeposition,” Anal. Chem. 63, 12891295 (1991). [CrossRef]
18. R. P. Van Duyne, J. C. Hulteen, and D. A. Treichel, “Atomic force microscopy and surface-enhanced Raman spectroscopy. I. Ag island films and Ag film over polymer nanosphere surfaces supported on glass,” J. Chem. Phys. 99, 2101–2115 (1993). [CrossRef]
19. W. -H. Yang, J. Hulteen, G. C. Schatz, and R. P. Van Duyne, “A surface-enhanced hyper-Raman and surface-enhanced Raman scattering study of trans-1,2-bis(4-pyridyl)ethylene adsorbed onto silver film over nanosphere electrodes. Vibrational assignments: experiment and theory,” J. Chem. Phys. 104, 4313–4323 (1996). [CrossRef]
20. K. L. Norrod, L. M. Sudnik, D. Rousell, and K. L. Rowlen, “Quantitative comparison of five SERS substrates: sensitivity and limit of detection,” Appl. Spectrosc. 51, 994–1001 (1997). [CrossRef]
21. P. N. Sanda, J. M. Warlaumont, J. E. Demuth, J. C. Tsang, K. Christmann, and J. A. Bradley, “Surface-enhanced Raman scattering from pyridine on Ag(111),” Phys. Rev. Lett. 45, 1519–1523 (1980). [CrossRef]
22. U. K. Sarkar, A. J. Pal, S. Chakraborti, and T. N. Misra, “Classical and chemical effects of SERS from 2,2’:5,2” terthiophene adsorbed on Ag-sols,” Chem. Phys. Lett. 190, 59–63 (1992). [CrossRef]
23. M. Moskovits, “Surface-enhanced spectroscopy,” Rev. Mod. Phys. 57,783–825 (1985). [CrossRef]
24. E. J. Zeman, K. T. Carron, G. C. Schatz, and R. P. Van Duyne, “A surface enhanced resonance Raman study of cobalt phthalocyanine on rough Ag films: theory and experiment,” J. Chem. Phys. 87, 4189–4200 (1987). [CrossRef]
25. R. J. Dijkstra, A. Gerssen, E. V. Efremov, F. Ariese, U. A. T. Brinkman, and C. Gooijer, “Substrates for the at-line coupling of capillary electrophoresis and surface-Raman spectroscopy,” Anal. Chim. Acta. 508, 127–134 (2004). [CrossRef]
26. N. Félidj, S. Lau Truong, J. Aubard, G. Lévi, J. Krenn, A. Hohenau, A. Leitner, and F. Aussenegg, “Gold particle interaction in regular arrays probed by surface enhanced Raman scattering,” J. Chem. Phys. 120, 7141–7146 (2004). [CrossRef] [PubMed]
27. S. E. Roark, D. J. Semin, A. Lo, R. Skodje, and K. L. Rowlen, “Solvent-induced morphology changes in thin silver films,” Anal. Chim. Acta. 307, 341–353 (1995). [CrossRef]
28. X. Li, W. Xu, H. Jia, X. Wang, B. Zhao, B. Li, and Y. Ozaki, “Water-induced morphology changes in an ultrathin silver film studied by ultraviolet-visble, surface-enhanced Raman scattering spectroscopy and atomic force microscopy,” Thin Solid films 474, 181–185 (2005). [CrossRef]