14172
Langmuir 2008, 24, 14172-14175
Gold-Coated Nanorod Arrays as Highly Sensitive Substrates for Surface-Enhanced Raman Spectroscopy J.-G. Fan* and Y.-P. Zhao Department of Physics and Astronomy, Nanoscale Science and Engineering Center, UniVersity of Georgia, Athens, Georgia 30602 ReceiVed July 14, 2008. ReVised Manuscript ReceiVed October 7, 2008 Substrates for surface-enhanced Raman spectroscopy (SERS) have suffered from low enhancement, short shelf time, and poor uniformity or reproducibility. We report an effective method to produce SERS substrates that can potentially overcome these shortcomings. The SERS substrate consists of a layer of 100 nm Ag film deposited on a Si or glass substrate, a Si nanorod array fabricated by glancing angle deposition, and an Au coating fabricated by sputtering. The effects of the height and separation of the Si nanorods and the thickness of the Au layer on the SERS enhancement factor are investigated. Optimal substrates are capable of detecting attomolar quantities of trans-1,2-bis(4-pyridyl)ethylene.
Surface enhanced Raman spectroscopy (SERS) has tremendous potential in chemical and biological sensing applications.1-4 The development of a practical SERS-based sensor requires an efficient SERS substrate that can not only provide strong enhancement factors but also is robust, stable, uniform (with large area), and reproducible and is easy and relatively inexpensive to fabricate and store. Many different fabrication techniques have been explored to obtain such a substrate, but there are only limited successes.5-7 Structurally, most SERS substrates are made from pure metallic nanostructures such as metal nanoparticles or particle arrays, metal core-shell nanoparticles, roughened metal surfaces, metal nanorod arrays, or a combination of a variety of nanostructures and thin films.5-7 Recently, template-based fabrication methods have been used for SERS substrate production and have exhibited promising potentials.8-15 In particular, nanoporous matrixes such as porous alumina and porous Si have been used as templates, and an Ag thin film was either coated on top of the surface through vacuum deposition9 or inside the pores by chemical plating.10,11 The porosity of the matrix as well as the thickness of the Ag film has a significant effect on the SERS enhancement factor (SEF).9-11 In this case, most fabrication methods are dominated by wet chemical processes, from porous template fabrication to metal plating. Compared to the random porous matrix, aligned nanorod arrays have relatively simple geometries and have also been used as templates for preparing * Corresponding author. E-mail:
[email protected]. (1) Willets, K. A.; Van Duyne, R. P. Annu. ReV. Phys. Chem. 2007, 58, 267. (2) Kneipp, K.; Kneipp, H.; Kneipp, J. Acc. Chem. Res. 2006, 39, 443. (3) Grow, A. E.; Wood, L. L.; Claycomb, J. L.; Thompson, P. A. J. Microbiol. Methods 2003, 53, 221. (4) Nabiev, I.; Chourpa, I.; Manfait, M. J. Raman Spectrosc. 1994, 25, 13. (5) Baker, G. A.; Moore, D. S. Anal. Bioanal. Chem. 2005, 382, 1751. (6) Tian, Z. Q.; Ren, B.; Wu, D. Y. J. Phys. Chem. B 2002, 106, 9463. (7) Vo-Dinh, T. Trends Anal. Chem. 1998, 17, 557. (8) Jung, D. S.; Lee, Y. M.; Lee, Y.; Kim, N. H.; Kim, K.; Lee, J. K. J. Mater. Chem. 2006, 16, 3145. (9) Walsh, R. J.; Chumanov, G. Appl. Spectrosc. 2001, 55, 1695. (10) Chan, S.; Kwon, S.; Koo, T. W.; Lee, L. P.; Berlin, A. A. AdV. Mater. 2003, 15, 1595. (11) Lin, H. H.; Mock, J.; Smith, D.; Gao, T.; Sailor, M. J. J. Phys. Chem. B 2004, 108, 11654. (12) Henley, S. J.; Carey, J. D.; Silva, S. R. P. Appl. Phys. Lett. 2006, 89. (13) Chattopadhyay, S.; Lo, H. C.; Hsu, C. H.; Chen, L. C.; Chen, K. H. Chem. Mater. 2005, 17, 553. (14) Suzuki, M.; Maekita, W.; Wada, Y.; Nakajima, K.; Kimura, K.; Fukuoka, T.; Mori, Y. Appl. Phys. Lett. 2006, 88, 203121. (15) Suzuki, M.; Nakajima, K.; Kimura, K.; Fukuoka, T.; Mori, Y. Anal. Sci. 2007, 23, 829.
SERS substrates.13-15 Chattopadhyay et al. fabricated crystalline Si nanotip arrays using a so-called self-masked dry etching process and sputtered a layer of Ag or Au nanoparticles on the side walls of the Si tips.13 They observed that such a substrate could exhibit an SEF in the range of 106-108. Suzuki et al. used a so-called dynamic oblique angle deposition technique to deposit an elongated SiO2 nanorod array, and after coating with a thin layer of Ag or Au, a horizontally elongated and aligned array of Ag or Au nanorods was formed on the surface.14,15 This substrate not only showed high SERS enhancement but also exhibited polarization-dependent SERS signals. The advantages of the template fabrication techniques are that the template can be fabricated over a relatively large area with statistically uniform structures and the required noble metal coating is very thin, which makes the process relatively inexpensive. In this letter, we describe an alternative templating method for fabricating highly sensitive, uniform, durable SERS substrates. The optimal SERS substrate structure consists of a layer of a highly reflective coating, an aligned nanorod array, and a thin layer of a Au nanoparticle coating on the nanorod array. To fabricate the SERS substrates, we used three different substrates: a clean bare Si(100) substrate, a cleaned glass substrate, and a 100 nm Ag-film-coated Si(100) substrate. A Si nanorod array was deposited onto these substrates by the glancing angle deposition (GLAD) method. GLAD is an extension of the commonly named oblique angle deposition (OAD) with programmed substrate rotations.16-18 The deposition flux arrives on a substrate at a large angle θ (>70°) with respect to the substrate normal, and the main mechanism for forming the nanorod array is a so-called self-shadowing effect. A detailed description and review of this deposition method can be found in the literature.16-18 During the Si nanorod array fabrication, the base pressure in the vacuum chamber was below 10-6 Torr, and the deposition angles θ were set to 80, 82.7, 84, 86, 87, and 88°. The substrates were prepared both with an azimuthal rotation at 0.5 rev/s and without rotation so that both a vertically aligned nanorod array and a tilted nanorod array were fabricated. The nominal deposition rate of Si, which was monitored by a quartz crystal microbalance, was 0.2 nm/s, and the nominal Si nanorod 59.
(16) Robbie, K.; Brett, M. J. J. Vac. Sci. Technol., A 1997, 15, 1460. (17) Zhao, Y.-P.; Ye, D.-X.; Wang, G.-C.; Lu, T.-M. SPIE Proc. 2003, 5219, (18) Steele, J. J.; Brett, M. J. J. Mater. Sci: Mater. Electron. 2007, 18, 367.
10.1021/la802248t CCC: $40.75 2008 American Chemical Society Published on Web 11/21/2008
Gold-Coated Nanorod Arrays
Langmuir, Vol. 24, No. 24, 2008 14173
Figure 2. UV-vis spectra of Si nanorods sputtered with Au thicknesses d of 2.55, 7.65, 15.30, 22.95, and 38.25 nm, respectively, (a) with and (b) without Si nanorod background subtraction. The Si nanorod arrays were deposited on glass substrates.
Figure 1. (a) SEM top-view image of the Si nanorod array deposited on a Si substrate at a vapor incident angle of 84°. The inset is the cross-sectional-view SEM. (b) Distribution of the size of Au particles sputtered on a Si nanorod. The inset is a TEM of the Au-coated Si nanorod. Scale bars are (a) 1 µm and (b) 100 nm.
thickness varied from 1 to 3 µm. The depositions occurred at room temperature, and the Si nanorods were all amorphous. A thin layer of Au nanoparticles was sputtered onto the Si nanorod array by an SPI-module sputter coater (SPI Supplies), with thickness d ranging from 2.55 to 76.5 nm. The film thickness d was calibrated with a quartz crystal microbalance. For comparison, similar Au layers were also sputtered onto the flat Si substrates. The UV-vis spectra were obtained from the nanorods on the glass substrate with a double-beam spectrophotometer (Jacso V-570) in transmission mode. Figure 1a shows the top view and cross-sectional view SEM images of the resulting Au/Si nanostructures with a vapor incident angle of θ ) 84° and a sputtered Au layer of 15.30 nm. The Si nanorods formed through the GLAD process were randomly distributed on the substrate, and they all aligned vertically as shown in the SEM images. The nanorod height h was 440 ( 10 nm, and the average rod-rod separation and rod diameter at the top were 100 ( 30 and 80 ( 20 nm, respectively. Figure 1b shows the TEM image of Au coated on the Si nanorods. The nanorods were not uniform across the entire length but rather were needle-shaped. From the bottom to the top, the diameter of the nanorod changed from 35 to 130 nm. In addition, the black areas in the TEM image are an indication of a Au coating on the Si nanorod resulting from its large electron scattering crosssection. The Au particles were coated all around the nanorod. On top of the nanorod, Au formed a continuous film, and around the side wall of the nanorod, Au formed small islands with diameters ranging from ∼20 nm near the top to ∼12 nm near the bottom as shown in Figure 1b. The UV-vis spectra of Aucoated Si nanorods (θ ) 84°, and h ≈ 440 nm) with various Au thicknesses on the glass substrates are shown in Figure 2: both
spectra with (Figure 2a) and without (Figure 2b) Si nanorod background subtraction are shown. There is a broad absorption peak near the wavelength of 520 nm that appears for all of the Au-coated samples, with absorbance increasing monotonically with Au thickness. This broad absorbance corresponds to the Au surface plasmon resonance and suggests a wide distribution of Au nanoparticle sizes. The SERS activity of these Au-coated Si nanorod array substrates was tested by a Raman probe molecule, trans-1,2bis(4-pyridyl)ethylene (BPE, Aldrich, 99.9+%). The Raman system used was a commercial fiber Raman system HRC-10HT (Enwave Optronics, Inc.) (A comparison of the spectra from this fiber Raman system and a micro-Raman system is available in Supporting Information.) It consists of a diode laser, a spectrometer, an integrated Raman probe head used for both excitation and collection, and separate delivery and collection fibers. The wavelength of the diode laser is 785 nm, and the laser spot size at focus has a diameter of 0.1 mm. The nanorod samples were cut to a rectangle of 20 × 7.5 mm2. A 2 µL droplet of 10-5 M BPE/methanol solution was dispensed onto the nanorod surface. It spread out quickly and covered an area of ∼15 × 7.5 mm2. Within 1 min, the methanol solvent totally evaporated. The estimated BPE molecular coverage on the nominal surface was ∼0.015 monolayer (assuming 7 × 1014 molecules/cm2 in a monolayer19). Approximately 1.4 × 10-15 moles of BPE was excited in the laser spot. Figure 3a shows two typical Raman spectra taken by the fiber Raman system at the center of the BPE-covered area from Au-coated Si nanorod arrays (θ ) 84°, Si rod height h ) 440 nm, and Au coating thickness d ) 15.3 nm) on the bare Si substrate and a Ag-coated Si substrate. The laser power used was 35 mW, and the sampling period was 10 s. Both spectra in Figure 3a show the characteristic peaks of BPE at around 1200, 1610, and 1640 cm-1 corresponding to the pyridyl ring C-N in-plane bending and C-C stretching, pyridyl C-C stretching, C-N and C-H in-plane bending, and ethylenic CdC stretching, respectively.20,21 These strong Raman peaks indicate that there was a huge SERS enhancement from the Au-coated Si nanorod substrates. However, there are no detectable BPE Raman peaks on the direct Au-coated Si wafer as shown in Figure 3a. For further comparison, in Figure 3a a SERS spectrum of BPE on the optimal Ag nanorod substrate prepared by the (19) Norrod, K. L.; Sudnik, L. M.; Rousell, D.; Rowlen, K. L. Appl. Spectrosc. 1997, 51, 994. (20) Zhuang, Z.; Cheng, J.; Jia, H.; Zeng, J.; Han, X.; Zhao, B.; Zhang, H.; Zhang, G.; Zhao, W. Vib. Spectrosc. 2007, 43, 306. (21) Yang, W.-H.; Hulteen, J.; Schatz, G. C.; Van Duyne, R. P. J. Chem. Phys. 1996, 104, 4313.
14174 Langmuir, Vol. 24, No. 24, 2008
Figure 3. (a) Comparison of the SERS of BPE molecules adsorbed on four types of surfaces: (I) a 15.30 nm Au-coated Si nanorod array on a Si substrate; (II) a 15.30 nm Au-coated Si nanorod array on a Si substrate coated with a 100 nm Ag film; (III) a 7.65 nm Au-coated Si wafer; and (IV) a Ag nanorod array on a glass substrate coated with a 500 nm Ag film. The concentration of the BPE/methanol solution was 10-5 M except for sample III, in which it was 10-4 M. The SERS intensity of BPE on sample III was amplified 10 times. (b) SERS of BPE molecules on sample II with a BPE concentration of 10-7 M and the corresponding background signal (without BPE).
OAD method is also plotted.22-27 The measurements were carried out by the same instrument under the same detection conditions. After subtracting the background, the SERS peak intensities of the 1200 cm-1 band are 5 × 104, 4.3 × 104, and 2.5 × 104 counts for the Ag nanorod substrate, Au-coated Si nanorods on an Ag film, and the Au-coated Si nanorods on the Si substrate, respectively. The band intensity of Au-coated Si nanorods on Ag film substrates are comparable to those on the Ag nanorod substrates that have SEFs greater than 108 (refs 22, 23, and 26) and are about twice as great as those of the Au-coated Si nanorods on bare Si substrates. The SEF can be calculated by the equation SEF ) (Isurf/Nsurf)/(Ibulk/Nbulk),22 where I and N stand for the intensity of a BPE band (e.g., 1200 cm-1) and the number of molecules probed by the laser, respectively, surf means surface, and bulk means BPE in a solvent (e.g., methanol). Unfortunately, with our fiber Raman system we were able to detect only weak BPE Raman signals (
