Flexible SERS active substrates from ordered vertical Au nanorod arrays

Abstract

SERS active flexible substrates were fabricated by droplet deposition of Au nanorod solutions on rigid surfaces and subsequent stamping of assembled nanorod arrays on photographic paper and plastic surfaces. With this method plasmonic SERS substrates whose active area was constituted by Au nanorods assembled in closely spaced vertical arrays were fabricated. The SERS performances of the flexible active substrates were investigated with model molecule 4-aminobenzenethiol (4-ABT), where detection in the μM concentration range and Enhancement Factors (EF) of 10⁵ were obtained. To demonstrate field-based applications, detection of food contaminant crystal violet (CV) below the recommended 5 nM limit of detection was achieved. Also, detection of drug-marker benzocaine traces by a swabbing technique was demonstrated.

---

Introduction

Surface-Enhanced Raman Scattering (SERS) of analytes adsorbed on metal nanostructures is a powerful tool for collecting molecular chemical information. The combination of molecularly specific information provided by Raman spectroscopy and the signal enhancement provided by noble metal nanostructures has allowed detection of a number of molecular species in traces¹ and even down to single molecule identification level.² Because of this enormous potential recently large efforts have been aimed at translating the capabilities of SERS to a practical detection system that can be utilised for routine analysis or in the field.

Generally two main approaches are used to produce SERS active substrates: top-down approaches such as lithography³ and bottom-up approaches such as self-assembly.⁴ In both cases, despite the tremendous signal enhancements achieved, the high cost fabrication procedures or limited reproducibility and shell life have limited the spread use of these SERS active substrates in common analysis.

Flexible substrates can become an alternative as they have the potential to combine low cost and flexibility while still maintaining sensitivity and robustness of rigid substrates. Flexible plasmonic SERS substrates have been fabricated using plastic,^5–7 paper^8–10 and polymer materials^11,12 as surfaces in combination with metal nanostructures constituting SERS active layer. Numerous methods of deposition of nanostructures have been tried such as physical deposition¹³ and laser induced-photothermal methods.¹⁴ Although highly sensitive SERS substrates are fabricated with these techniques, the requirement for high laser power or expensive equipment has limited their production in large-scale. In contrast, solution fabrication processes such in situ synthesis on porous paper,¹⁵ printing¹⁶ and dip-coating¹⁷ are preferred, since plasmonic SERS substrates are obtained with small amount of materials and inexpensive deposition techniques.

As well as cost and ease of fabrication flexible substrates have been shown efficiency of sample collection particularly suited for real world sample analysis. For example, White et al. have achieved detection of analytes from complex samples using chromatographic separation on SERS inkjet-printed cellulose paper.¹⁸ SERS active filter paper substrates have been used for trace detection of explosives.¹⁹ Also swabbing the surface under investigation with a flexible substrate has been proposed as highly practical and efficient method to maximise sample collection. SERS dipsticks and swabs fabricated by SERS inkjet-printed paper have been proposed by White et al. for efficient detection of trace chemical detection (malathion and cocaine).²⁰

Recently our group has fabricated Au nanorod arrays of controlled geometrical order and has investigated their potential as plasmonic SERS substrates.^21,22 Enhancement factors on the order of 10⁵ were achieved together with quantitative detection of 4-aminobenzenethiol (4-ABT) down to 0.1 nM.²² The arrays formed on rigid substrates by droplet evaporation of Au nanorod solutions but had to be stamped on receiving rigid substrates prior to SERS analysis, in order to provide stability and reproducibility necessary for analytic investigation.

Taking advantage of this requirement, in this paper we used flexible surfaces as receiving substrates during droplet deposition/stamping method to fabricate flexible plasmonic SERS substrates. Large arrays of vertically aligned Au nanorods were formed by droplet evaporation on glass and were subsequently stamped on photographic paper (paper covered with a layer of polyethylene terephthalate (pPET)) and polyvinyl chloride (PVC) surfaces. The SERS activity of the fabricated plasmonic substrates was investigated with model molecule 4-ABT. PPET and PVC active substrates were robust and could be immersed in analyte solutions for up to 20 h without degradation. Enhancement factors on the order of 10⁵ were achieved for these active substrates.

As real-world applications detection of food contaminant crystal violet (CV) to sub-nM detection limits and trace detection of benzocaine powder by swabbing were achieved on pPET active substrates.

Results and discussion

Vertical Au nanorod arrays were fabricated following a method illustrated in Scheme 1a and b. The fabrication process comprised two steps: (a) deposition of a nanorod droplet solution on a rigid surface followed by evaporation under controlled conditions and (b) stamping of the obtained nanorod array on a receiving flexible surface. Specifically, a drop of Au nanorod chlorobenzene solution was deposited on a glass coverslip and left evaporate under controlled conditions (T = 20 °C, humidity = 70%, evaporation time = 3 h, Scheme 1a1). As solvent evaporated, nanorods assembled forming a gold film at the interface between solvent and air, as shown in the photograph 2 included in Scheme 1.

Scheme 1 Formation of Au nanorod vertical arrays by combined droplet evaporation/stamping technique; (a1) droplet evaporation process; (a2) photograph of evaporating droplet; (a3) nanorod vertical array deposited on glass support; (a4) photograph of dried droplet. (b1–3) Stamping and cleaning of nanorod arrays on flexible support.

Following solvent evaporation, assembled nanorods deposited on the glass surface covering the entire area of droplet. This process is illustrated in Scheme 1a3. A photograph of the evaporated droplet showing a Au nanorod layer covering the entire area of the droplet is shown in (1a4). Formed nanorod arrays were transferred onto flexible surfaces by stamping, following the method described in Scheme 1b. Specifically, a flexible surface was pressed on the original glass surface containing nanorod arrays for 20 s (1b1). Nanorod arrays transferred intact on the flexible surface together with the residual organic matter, which deposited on top of the arrays (b2). Nanorod arrays were cleaned by immersion in isopropanol, followed by multiple rinses with clean isopropanol (b3). It should be noticed that direct formation of vertical arrays on flexible supports was achieved but the concomitant deposition of excess organic surfactants arising from synthesis and phase transfer processes prevented further practical use of such substrates. In fact, the arrays lifted from the support when immersed in analyte solutions, making impossible their use as SERS active substrates. In contrast, after stamping nanorods arrays strongly attached to flexible surfaces, thus allowing removal of excess organic matter necessary to avoid interference with SERS analyte detection.

Fig. 1a shows a SEM image of vertical nanorod arrays formed on glass and stamped on pPET. Nanorods were organized in a closely-packed quasi-hexagonal monolayer structure.

Fig. 1 SEM images of flexible SERS active surfaces obtained by droplet evaporation of Au chlorobenzene solutions and stamping of the dried droplet on (a) pPET and (b) PVC surfaces.

Formed arrays displayed a high degree of pattern regularity and a high number of closely spaced nanorods, although occasional irregularities in the lattice and formation of crack patterns occurred as result of the stamping process. Analogously, vertical dense arrays were obtained from stamping of Au nanorods evaporated droplets on PVC surfaces, as shown in Fig. 1b. Photographs of stamped arrays (insets of Fig. 1c and d) show the intense gold/dark colors associated to densely packed nanorod structures.

In order to obtain SERS spectra, flexible surfaces containing nanorod arrays were immersed for 20 h in 4-ABT, methanol (MeOH) solutions.

Laser wavelength of 633 nm was used as excitation source. Fig. 2a shows SERS spectra obtained with pPET active substrates immersed in 4-ABT solutions 0.1 mM (red curve) and 1 μM (green curve). For comparison the Raman spectrum of bulk 4-ABT (powder, black curve) deposited on glass coverslip is also shown in Fig. 2a. The Raman spectrum of 4-ABT powder was characterized by a set of peaks at 1007, 1086, 1178, 1495 and 1591 cm⁻¹ associated with the a₁ vibrational, in plane, in phase modes of 4-ABT.²³ The a₁ vibrational modes observed in the Raman spectrum appeared strongly enhanced in the SERS spectra. Since it is known that a₁ modes are mainly enhanced by electromagnetic mechanism,²³ the observed enhancement was attributed to the enhancement of the electromagnetic field produced by closely spaced nanorods. In addition to a₁ peaks, SERS spectra showed a new set of peaks at 1139, 1390 and 1437 cm⁻¹. These peaks were assigned to b₂ in plane, out of phase vibrational modes. The formation of b₂ peaks has been observed in SERS spectra of 4-ABT recorded on nanostructured surfaces and has been attributed to a metal-molecule charge transfer related to a chemical enhancement process.^24,25 The high intensity of b₂ peaks has been attributed to efficient binding of 4-ATP on vertical arrays due to strong capillary forces arising in closely-spaces nanorod “forests”.²⁶ Raman and SERS peak positions and assignments are listed in Table 1. Fig. 2b shows SERS spectra recorded by immersing nanorod stamped on PVC surfaces in 4-ABT solutions (0.1 mM, red curve; 1 μM green curve). Spectra showed the same combination of a₁/b₂ peaks observed on stamped PET supports although the absolute peak intensities were lower by a factor of 2. However, detection in the range 0.1 mM to 1 μM was still achieved with such active substrates. Fig. 2c shows comparison between 4-ABT spectra measured on flexible substrates with spectra measured on stamped glass substrates (red curve) and commercially available Klarite substrates (black curve). For all samples 4-ABT concentration was 0.1 mM. Absolute scattering intensity of glass substrates was higher than flexible substrates by a factor of ca. 3, in agreement with lower detection limits demonstrated with such substrates.²² However, SERS spectra recorded with flexible substrates showed an increase of the absolute scattering intensity of the a₁ 1086 cm⁻¹ peaks of a factor of ca. 2 compared to Klarite, confirming that the electromagnetic field generated by rough Au surfaces is lower in intensity than the field generated by strongly coupled Au nanorods. An even higher enhancement factor was observed for the 1139 cm⁻¹ b₂ peak (factor 8 and 16 for PVC and pPET respectively). This enhancement is likely associated with the higher number of 4-ABT molecules adsorbed on high surface area Au vertical arrays compared to rough Au surfaces.

Fig. 2 Raman spectrum (black curve) of powder 4-ABT and SERS spectra of 4-ABT 0.1 mM (red curve) and 1 μM (green curve) deposited on (a) pPET and (b) PVC active substrates; (c) SERS spectra of 0.1 nM 4-ABT obtained in glass, pPET, PVC active substrate and Klarite commercial substrate.

Table 1 Raman band frequencies of bulk 4-ATP and SERS band frequencies of 4-ABT adsorbed on pPET active SERS substrates; [*] ν = stretching, δ, γ = bending, π = wagging. Letters in parenthesis indicate the vibrational symmetry

Peak position 4-ABT powder	Peak position 4-ABT/pPET	Assignments*
1591	1576	νCC, 8a (a1)
1569		νCC, 8b (b2)
1495	1473	νCC + δCH, 19a (a1)
	1433	νCC + δCH, 19b (b2)
	1389	δCH + νCC, 3 (b2)
	1304	νCC + δCH, 14b (b2)
1178	1187	δCH, 9a (a1)
	1140	δCH, 9b (b2)
1086	1073	νCS, 7a (a1)
1007	1005	γCC + γCCC, 18a, (a1)

---

A further evaluation of SERS performances of plasmonic flexible substrates was achieved by calculating the Enhancement Factor (EF), defined as the ratios of the intensities of the scattered radiation for SERS and normal Raman scattering per molecule

EF = (I_SERS/N_SERS)/(I_NR/N_NR)

where I_SERS and I_NR are the integrated intensities of the SERS and normal Raman scattering spectra for 4-ABT, respectively; N_SERS and N_NR are the number of molecules found in the laser excitation area adsorbed on nanorod arrays and in bulk powder form, respectively. For calculations the intensity of the band at 1086 cm⁻¹ from spectra of Fig. 2a was used. PPET and PVC active substrates gave EF values of 1.4 × 10⁵ and 9.2 × 10⁴, respectively. These values are in agreement with values reported in literature for SERS active flexible substrates.^27,28

Au nanorod vertical arrays were also stamped on TLC plates and filter paper. Such active SERS substrates were used for swabbing detection of 4-ABT down to 6 μg detection limits (see ESI†).

To demonstrate the use of these active substrates for field applications, we performed detection of crystal violet (CV), a toxic cationic dye largely used as food colouring agent and food additive, and benzocaine, a readily available marker for drug with physico-chemical similarities to cocaine. CV has been classified as recalcitrant molecule since it is poorly metabolized by microbes, is not bio-degradable and can persist in a number of environments.²⁹ For these reasons the minimum required performance limit (MPRL) was set for 2 mg L⁻¹ (ca. 4.9 nM) in European Commission and US.³⁰

The detection of CV is shown in Fig. 3 for concentrations between 0.1 μM and 1 nM. SERS spectra were recorded by immersing pPET active substrates in CV solutions (MeOH) for 4 h. As comparison the Raman spectrum of bulk CV in powder form is also shown in Fig. 3. All spectra were recorded with a 633 nm excitation source. The detection limit of CV with pPET was lower than 1 nM, because even the spectrum of the 1 nM sample had a good signal-to-noise ratio and the main peak at 1172 cm⁻¹ was still observable. A sigmoidal relationship between the SERS intensity and concentration of CV was observed (inset Fig. 3) showing concentration detection from 10 μM to 1 nM. For these calculations the change in intensity of the 1171 cm⁻¹ peak with CV concentration was used. The SERS intensities were mostly linear at low CV concentration. At high concentration a non linear response emerged, and saturation of the SERS signal occurred indicating that the adsorption of CV onto the arrays become saturated beyond this level.⁹ The demonstrated ability to detect CV below the currently accepted levels would make possible such field-based application of our plasmonic flexible substrates.

Fig. 3 SERS spectra of CV adsorbed on pPET active substrate. CV concentration was 100 nM (red curve), 10 nM (green curve), 1 nM (blue curve). Black curve is the bulk Raman spectrum of powder CV. Inset: sinusoidal change in intensity of the 1171 cm⁻¹ peak with CV concentration.

We also performed detection of benzocaine by swabbing a surface containing traces of the analyte. Specifically, trace amount of benzocaine were deposited and crushed on a glass slide. A nitrogen gun was used to remove as much benzocaine as possible, leaving only trace amounts. A drop of MeOH was added to a pPET substrate which was swabbed against the glass slide containing benzocaine traces. The pPET was subsequently dried with N₂ gun, rinsed again with MeOH and dried, to ensure that only traces amounts were left on the active substrate. Fig. 4 (red curve) shows how small amounts of benzocaine swabbed from the glass slide could be detected with pPET substrates. For comparison the Raman spectrum of bulk powder benzocaine (blue curve) is shown together with the blank featureless spectrum of pPET active substrate recorded before swabbing.

Fig. 4 SERS spectrum of benzocaine swabbed on a pPET active substrate (red curve); Raman spectrum of benzocaine powder (blue curve); Raman spectrum of pPET active substrate (black curve).

Conclusions

Plasmonic SERS substrates were fabricated by stamping on flexible surfaces vertical nanorod arrays obtained by droplet evaporation of Au nanorod solutions on glass surfaces. The SERS performances of fabricated substrates were assessed with model molecule 4-ABT. Detection down to 1 μM and EF of the order of 10⁵ were obtained for pPET and PVC active substrates. In order to show real field-based applications, detection of food contaminant CV was achieved below the recommended limit of 5 nM on pPET active substrates. Also detection of benzocaine traces was achieved by swabbing contaminated glass slides with pPET active substrates. These examples show that our fabricated substrates can be suitable for real-world applications, where flexibility on sample collection, fast responses and low limits of detection are highly desired.

Experimental

Materials

Tetrachloroauric acid, silver nitrate, sodium borohydride, ascorbic acid, hydrochloric acid, chlorobenzene, cetyltrimethylammonium bromide (CTAB), mercaptosuccinic acid, tetraoctylammoniumbromide (TOAB), 4-aminobenzethiol (4-ABT), crystal violet (CV), benzocaine were purchased from Sigma-Aldrich. All glassware was cleaned with aqua regia prior to nanorod synthesis. Milli-Q water (resistivity >18 MΩ cm⁻¹) was used throughout the experiments.

Synthesis of Au nanorods

Au nanorods (20 ± 2 × 61 ± 7 nm, AR = 3.0) were synthesized by overgrowth of seed-mediated Au nanorods (11 ± 1 × 41 ± 2 nm, AR = 3.7), according to a method described by Liz-Marzán et al.³¹ Both as-prepared nanorods were centrifuged and re-dispersed into water, in order to obtain a final CTAB concentration of 0.1–0.2 mM. Prior to droplet deposition, Au nanorods were phase-transferred from water to chlorobenzene following the method described by Chen et al.³²

Fabrication of ordered Au nanorod arrays

A small aliquot (10 μL) of Au nanorod chlorobenzene solution ([Au] = 10 nM) was deposited on a glass surface, covered with a petri dish and then left to evaporate at room temperature over a time of 3 h. Controlled solvent evaporation resulted in formation of perpendicularly ordered nanorod arrays. Fabricated arrays were transferred intact flexible surfaces by placing a receiving surface on the original glass support and pressing the two surfaces together for 20 s. Excess organic matter was removed by immersing glass-nanorod arrays in isopropanol for 1 h, followed by multiple rinses with fresh isopropanol.

Scanning electron microscopy (SEM)

Images of nanorod arrays were acquired using a field emission SEM (JSM-6700F, JEOL UK Ltd.) operating at beam voltages of 5 kV.

Raman spectra

Were obtained from a Renishaw inVia Raman system. A 632.8 nm helium–neon laser was employed as an excitation source. The laser beam was focused onto the sample through a Mitutoyo M Plan Apo 100× objective with 0.7 N.A. Measured power at the sampling level was controlled at about 5 mW. Acquisition time was usually 30 s. To obtain SERS spectra, Au nanorod arrays stamped on flexible surfaces were immersed in 4-ABT solutions (MeOH, 0.1 mM to 100 pM, 20 h). Substrates were rinsed with MeOH in order to remove unbound 4-ABT and gently dried under N₂ stream.

Acknowledgements

This work was supported by the European Union Seventh Framework Programme (project 263091 Hysens).

Notes and references

1. K. D. Osberg, M. Rycenga, G. R. Bourret, K. A. Brown and C. A. Mirkin, Adv. Mater., 2012, 24, 6065–6070.
2. P. G. Etchegoin and E. C. Le Ru, Phys. Chem. Chem. Phys., 2008, 10, 6079–6089.
3. B. B. Xu, Z. C. Ma, H. Wang, X. Q. Liu, Y. L. Zhang, X. L. Zhang, R. Zhang, H. B. Jang and H. B. Sun, Electrophoresis, 2011, 32, 3378–3384.
4. B. B. Xu, Z. C. Ma, H. Wang, R. Zhang, L. G. Niu, Z. Yang, Y. L. Zhang, H. W. Zheng, B. Zhao, Y. Xu, Q. D. Chen, H. Xia and H. B. Sun, Lab Chip, 2011, 11, 3347–3351.
5. J. P. Sing, H. Y. Chu, J. Abell, R. A. Tripp and Y. P. Zhao, Nanoscale, 2012, 4, 3410–3414.
6. G. Lu, H. Li and H. Zhang, Chem. Commun., 2011, 47, 8560–8562.
7. W. Y. Wu, Z. P. Bian, W. Wang and J. J. Zhu, Sens. Actuators, B, 2010, 147, 298–303.
8. Y. H. Ngo, W. L. Then, W. Shen and G. Garnier, J. Colloid Interface Sci., 2013, 409, 59–65.
9. Y. H. Ngo, D. Li, G. P. Simon and G. Garnier, Langmuir, 2012, 28, 8782–8790.
10. S.-C. Tseng, C.-C. Yu, D. Wan, H.-L. Chen, L. A. Wang, M.-C. Wu, W.-F. Su, H.-C. Han and L.-C. Chen, Anal. Chem., 2012, 84, 5140–5145.
11. D. He, B. Hu, Q.-F. Yao, K. Wang and S.-H. Yu, ACS Nano, 2009, 3, 3993–4002.
12. C. L. Zhang, K. P. Lu, H. P. Cong and S. H. Yu, Small, 2012, 8, 648–653.
13. R. Zhang, B. B. Xu, X. Q. Liu, Y. L. Zhang, Y. Xu, Q. D. Chen and H. B. Sun, Chem. Commun., 2012, 48, 5913–5915.
14. S. C. Tseng, C. C. Yu, D. H. Swan, H. L. Chen, L. A. Wang, M. C. Wu, W. F. Su, H. C. Han and L. C. Chen, Anal. Chem., 2012, 84, 5140–5145.
15. M. L. Cheng, B. C. Tsai and J. Yang, Anal. Chim. Acta, 2011, 708, 89–96.
16. W. W. Yu and I. M. White, Anal. Chem., 2010, 82, 9626–9630.
17. Y. H. Ngo, D. Li, G. P. Simon and G. Garnier, Langmuir, 2012, 28, 8782–8790.
18. W. W. Yu and I. M. White, Analyst, 2013, 138, 3679–3686.
19. P. M. Fierro-Mercado and S. P. Hernández-Rivera, Int. J. Spectrosc., 2012, 2012, 716527.
20. W. W. Yu and I. M. White, Analyst, 2012, 137, 1168–1173.
21. S. Diefenbach, N. Erhard, J. Schopka, A. Martin, C. Karnetzy, D. Iacopino and A. Holleitner, Phys. Status Solidi, 2014, 8, 264–268.
22. A. Martín, A. Pescaglini, C. Schopf, V. Scardaci, R. Coull, L. Byrne and D. Iacopino, J. Phys. Chem. C, submitted.
23. X. Hu, T. Wang, L. Wang and S. Dong, J. Phys. Chem. C, 2007, 111, 6962–6969.
24. M. Osawa, N. Matsuda, K. Yoshii and I. J. Uchida, Phys. Chem., 1994, 98, 12702–12707.
25. J. R. Lombardi and R. L. Birke, J. Phys. Chem. C, 2008, 112, 5605–5617.
26. B. Peng, G. Li, D. i. Li, S. Dodson, Q. Zhang, J. Zhang, Y. Hong Lee, H. V. Demir, X. Y. Ling and Q. Xiong, ACS Nano, 2013, 7, 5993–6000.
27. G. V. Bianco, M. Losurdo, M. M. Giangregorio, P. Capezzuto and G. Bruno, Plasmonics, 2013, 8, 159–165.
28. C. H. Lee, L. Tian and S. Singamaneni, ACS Appl. Mater. Interfaces, 2010, 12, 3429–3435.
29. B. Liu, W. Jiang, H. Wang, X. Yang, S. Zhang, Y. Yuan, T. Wu and Y. Du, Microchim. Acta, 2013, 180, 997–1004.
30. Y. D. Shen, X. F. Deng, Z. L. Xu, Y. Wang, H. T. Lei, H. Wang, J. Y. Yang, Z. L. Xiao and Y. M. Sun, Anal. Chim. Acta, 2011, 707, 148–154.
31. R. A. Alvarez-Puebla, A. Agarwal, P. Manna, B. P. Khanal, P. Aldeanueva-Potel, E. Carbó-Argibaya, N. Pazos-Pérez, L. Vigderman, E. R. Zubarev, N. A. Kotov and L. M. Liz-Marzán, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 8157–8161.
32. J. Yang, J.-C. Wu, J.-K. Wang and C.-C. Chen, Chem. Phys. Lett., 2005, 416, 215–219.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra01916d

---