Large Format Surface-Enhanced Raman Spectroscopy Substrate

Aug 2, 2016 - User Resources. About Us · ACS Members · Librarians · Authors & Reviewers · Website Demos · Privacy Policy · Mobile Site ...
0 downloads 0 Views 4MB Size
Large Format Surface-Enhanced Raman Spectroscopy Substrate Optimized for Enhancement and Uniformity Katherine N. Kanipe, Philip P. F. Chidester, Galen D. Stucky, and Martin Moskovits* Department of Chemistry & Biochemistry, University of California Santa Barbara, Santa Barbara, California 93106-9510, United States S Supporting Information *

ABSTRACT: Gratings have been widely investigated both theoretically and experimentally as surface-enhanced Raman spectroscopy (SERS) substrates, exhibiting, under appropriate circumstances, increased far-field extinctions and near-field intensities over those of an appropriately equivalent number of isolated particles. When the grating order transitions from evanescent to radiative, narrow resonance peaks are observed in the extinction spectrum whose properties can be manipulated by controlling the grating’s geometric parameters. Here we report the application of the architectural principles of grating fabrication using a square two-dimensional array of goldcoated nanostructures that achieves SERS enhancements of 107 uniformly over areas of square centimeters. The highperformance grating substrates were fabricated using commonly available foundry-based techniques that have been chosen for their applicability to large-scale wafer processing. Additionally, we restricted ourselves to a parametric regime that optimizes SERS performance in a repeatable and reproducible manner. KEYWORDS: Surface-enhanced Raman spectroscopy, plasmonic substrate, chemical and biochemical sensing, nanogratings, SERS substrate

S

urface-enhanced Raman spectroscopy (SERS) is a spectroscopic technique with unusually high sensitivity and molecular specificity. Efforts have been made to develop SERS-based sensors for a range of applications including the detection of analytes in biological samples,1 explosives in military and security settings,2 and pigment analysis in miniscule paint samples to help direct art conservation efforts,3 all of which can be carried out with hand-held, field compatible Raman systems.4 The success of SERS relies on the availability of highly enhancing, uniform, and reproducible SERS substrates from which spectra are collected. A variety of approaches have been proposed to achieve high enhancement factors and a high density of areas of high enhancement, so-called hot spots.5 A rich history of using bottom-up techniques to produce, for example, highly enhancing nanoparticles and nanoparticle assemblies is reported in the literature.6−8 More recently with advancements in nanofabrication technologies, top-down approaches have been developed, which create reproducible and uniform substrates but usually with enhancement factors considerably lower than those achieved using bottom-up approaches.9−12 Although two-dimensional gratings had been proposed for fabricating as SERS substrates since the early days of SERS,13 © 2016 American Chemical Society

few have exploited gratings as a SERS-substrate fabrication strategy largely because, until recently, gratings with the requisite nanostructures could only be achieved using costly and substrate size-restricted foundry nanotechnologies such as electron beam lithography.14 Nowadays, gratings can be written on large format wafers using commonly employed nanotechnologies such as interference or deep UV lithographies, enabling the fabrication of low cost, highly uniform, and highly enhancing SERS substrates, a high-performing example of which we describe herein. A helpful way to understand the manner in which the grating’s structure contributes to the SERS enhancement is to separate the contribution of the local field, which depends critically on the gap between neighboring elements of the grating, from far-field interactions resulting from the coherent superposition of light scattered by the grating’s periodic array of elements. If the gap is large compared with the wavelength (as would be the case for, say, a two-dimensional array of wellReceived: April 15, 2016 Accepted: August 2, 2016 Published: August 2, 2016 7566

DOI: 10.1021/acsnano.6b02564 ACS Nano 2016, 10, 7566−7571

Article

www.acsnano.org

Article

ACS Nano

Figure 1. Scanning electron micrographs elucidating the results of the SERS wafer processing: (a,d) lithography and etching; (b,e) plasmaenhanced chemical vapor deposition of silica; (c,f) electron-beam evaporation of gold. The periodic nature of the structures in both the x and y directions, and the modulation of the structures’ dimensions via silica deposition are clearly apparent. Images a, b, and c are top views, and images d, e, and f are taken at a 45° tilt to show the profile of the structures. (The images shown in e and f were chosen purposely because they show rarely encountered defects that illustrate the under-structure of the Si/SiO2/Au, core−shell grating elements.) The accompanying schematic evokes the fabrication state of the substrate to that point whose actual image is shown in the respective SEM images.

strongly dependent on the polarization of the excitation laser with respect to the orientation of the grating’s lines.19 The nearfield interactions also drive the grating’s resonances to longer wavelengths, making possible SERS excitation by commonly available lasers emitting in the red.

spaced particles), one need only consider the superposition of dipolar fields. The strength of the dipolar interactions for such gratings and a number of other of its properties is found to increase dramatically for grating pitches (i.e., repeat length) above a critical value at which one of its grating orders transitions from evanescent to radiative.15,16 The plasmonic properties of such gratings have been studied by Crozier and members of his group.17 When illuminated by laser light resonant with one of its orders above the aforementioned critical order, such a grating consisting of, for example, a twodimensional periodic arrangement of silver or gold nanoparticles covered with an adsorbate produces SERS spectra significantly more intense than what an equivalent number of such nanoparticles randomly arranged in solution or spaced sufficiently far apart would. If, however, the gap between neighboring elements is small compared with the wavelength of the light illuminating the grating, then the theory of its operation must include the effect of the near field in the vicinity of its features, or equivalently, the sum of many multipolar fields. Two-dimensional gratings in which the individual components are very closely spaced possess the added benefit of the near-field interactions, much as one achieves by bringing two nanoparticles covered in adsorbate together so as to form a small interstice, a so-called SERS hot spot.18 This effect was demonstrated for a onedimensional grating whose SERS enhancement, however, was

RESULTS AND DISCUSSION Here we report the fabrication of large-format SERS substrates with high enhancement (≥107), high uniformity, and high reproducibility rarely achieved simultaneously, using an entirely top-down nanofabrication approach utilizing laser interference lithography on silicon, combined with deep etching, plasmaenhanced chemical vapor deposition (PECVD), and e-beam deposition, all commonly used foundry techniques. Laser interference lithography (LIL) is currently used industrially on a whole-wafer scale to fabricate highly reproducible grating patterns, primarily for producing large-format polarizers.20 Once a grating pattern has been produced in both the x and y directions, reactive ion etching is used to transfer the pattern into the underlying silicon. The resulting posts are coated with silica deposited by PECVD and, last, gold deposited by electron-beam evaporation. Postlithography processing is used to transform these patterns into a dense array of SiO2@Au core−shell nanostructures capable of sustaining strong surface plasmons, which are essential to achieving high enhancements factors. Characterization by scanning electron microscopy 7567

DOI: 10.1021/acsnano.6b02564 ACS Nano 2016, 10, 7566−7571

Article

ACS Nano

substrates are assumed to have surface areas that are approximately equal to that of a planar single crystal gold surface, eliminating the need to estimate the number of molecules active in the SERS measurement, a common source of error in computing SERS enhancements, as well as issues of variable quenching of molecular resonances on the two samples. Within the parametric regime identified to give the best enhancements, the average observed intensities were on the order 106 to 107 counts per milliwatt per second, yielding enhancement factors ranging between 105 and 107 over single crystal gold. The enhancement of the substrate was also measured using millimolar solutions of 1,2-bis(4-pyridyl) ethylene (BPE). This molecule yielded enhancements of ∼107 (Supporting Information). BPE was chosen because it bonds to gold substantially more weakly than thionine (which is an organosulfur species) does, and unlike thionine, BPE is not resonant with the 633 nm wavelength of the exciting laser, thereby eliminating the likelihood that the large SERS enhancement observed for thionine resulted from the (unlikely) possibility that it binds differently to noncrystalline gold than to single-crystal gold, thereby creating new channels for chemical enhancement. The substrate design was optimized by exploring a geometric parameter space that makes use of the key principles relevant to gratings and plasmon excitation: a grating order that is radiative rather than evanescent, nanogaps that are sufficiently small to produce hot-spots, and a gold coating that is sufficiently thick (i.e., sufficiently metallic) to support sharp surface plasmons. These design parameters were controlled by varying the grating pitch, silica thickness, and gold thickness. The approach is performance driven, meaning that the substrates were optimized for use with a 633 nm laser. Accordingly an exhaustive examination of the resonance dependence due to parameter manipulation was not our goal. Aspects of such wavelength tunability in grating structures were investigated previously by groups such as Crozier et al.17 The grating pitch was varied between 240 and 330 nm. The transition between evanescent and radiative grating orders becomes apparent in the observed SERS intensities (Figure 3a). Furthermore, spectroscopic ellipsometry was used to obtain the reflectance spectrum at normal incidence Ro(λ) from which the normal incidence absorbance spectrum Ao(λ) was calculated as described in detail in the Supporting Information. Briefly, spectroscopic ellipsometry was used to measure

(SEM) revealed a square lattice of nanostructures that approximate a silica−gold core−shell particle supported on a silicon post (Figure 1). The measured grating pitch was 330 nm in both the x and y directions with the nanostructure mostly filling the entire 330 nm × 330 nm repeat unit. The dimensions of the nanostructure as well as proximity to neighboring structures were controlled by deposition of silica onto the lithographically produced silicon grating. Deposition of gold onto the silica-coated structures also contributed to the final core−shell nanostructure’s dimension and nearest neighbor spacing, although the thickness of the silica underlayer was the primary means we used to vary those parameters. The strongest performing substrate among those we produced had average gap sizes of less than 15 nm and demonstrated nearly defectfree patterning over the entire 15 mm × 15 mm processed surface. SERS measurements were carried out using a Horiba Raman microscope equipped with a 633 nm laser on substrates onto which thionine was allowed to adsorb for 18 to 24 h in a micromolar aqueous thionine solution, ensuring monolayer coverage. The thionine molecule forms a sulfur−Au bond with the gold substrate surfaces and produces a well-defined SERS signal due to its large Raman scattering cross-section.21 The spectra obtained are in good agreement with literature spectra. Additionally, a broad luminescence background was observed, a common feature of SERS, which was subtracted from the overall spectra to compare true SERS intensities when computing enhancement. The highest observed average intensity was 4.3 × 107 counts per milliwatt per second for the major thionine peak near 487 cm−1 when the substrate was exposed for 0.25 s with an incident laser power of 7.6 μW. A single crystal gold surface dosed with thionine in the same manner as was used above produced a SERS signal of 7.4 counts per milliwatt per second using the same illumination conditions, corresponding to an enhancement of ∼107 (Figure 2). The enhancement was calculated in accordance with the expression presented by Le Ru and Etchegoin, where I / Nsurf EF = SERS , which is commonly used to compare average I /N RS

vol

SERS enhancements across different surfaces.22 The SERS

rp̂ rŝ

= ρ = tan Ψ eiΔ at various angles of incidence (r̂s,p are

Fresnel coefficients for s and p polarization). The grating was assumed to be an effective optical medium that was sufficiently opaque so that its reflectance and absorbance at any angle summed to unity (i.e., that the transmittance was zero). The optical constants, n(λ) and k(λ), were extracted from ρ in the conventional manner23 (described in greater detail in the Supporting Information) and the Fresnel coefficient at normal incidence, r̂o(λ), was computed using those values as a function of wavelength, using which, the normal incidence reflectance (Ro = r̂*o r̂o) and normal incidence absorbance spectrum (Ao(λ) = 1 − Ro(λ)) were computed (also described in the Supporting Information). The absorption spectra at normal incidence were desired since our SERS measurements were carried out at normal incidence. The normal incidence absorption spectra computed from ellipsometric measurements carried out at various angles of incidence were similar, reinforcing the legitimacy of our approach.

Figure 2. Enhancement was determined by comparing the SERS intensity recorded for the grating substrate to the intensity measured from adsorbate-covered single crystal gold. The plotted intensity from the substrate is shown divided by a factor of 105 for clarity relative to the scale of the single crystal gold. 7568

DOI: 10.1021/acsnano.6b02564 ACS Nano 2016, 10, 7566−7571

Article

ACS Nano

Figure 3. (a) Measured SERS intensity versus grating pitch; samples shown (grating pitch): sample A, 242 nm; sample B, 282 nm; sample C, 305 nm; sample D, 330 nm. Gap size and gold thickness are held approximately constant at 16 and 105 nm, respectively. (b) Absorbance (one minus normal reflectance) versus wavelength. The transition from an evanescent grating order to a radiative one is apparent with increasing grating pitch. Samples A, B, and D are shown here; sample C is excluded for clarity. (c) SERS intensity versus gap size, which drops off rapidly and results in weak SERS intensities beyond ∼20 nm gap size; samples shown (gap size): sample E, 116 nm; sample F, 34 nm; sample G, 27 nm; sample H, 21 nm; sample I 16 nm. Grating pitch and gold thickness are held approximately constant at 330 and 120 nm, respectively. (d) SERS intensity versus gold thickness; substrates with a gold layer that demonstrates poor conductance give weak SERS intensities, while those with more metallic character give strong SERS intensities; samples shown (gold thickness): sample J, 50 nm; sample K, 100 nm; sample L, 110 nm; and sample M, 120 nm. Grating pitch and gap size were held approximately constant at 330 and 18 nm, respectively.

An absorbance feature associated with the first order grating resonance is clearly visible for gratings with appropriate grating parameters, as is the transition from evanescent to radiative modes (Figure 3b), which is signaled by a change in both peak position and quality factor of the absorption band with changing grating pitch. As expected, the absorbance feature for the most enhancing gratings showed a maximum near the wavelength of the excitation laser (633 nm). The silica layer was varied to control the gap size between vicinal elements by depositing thicknesses that varied between 50 and 284 nm. Since smaller interstitial regions tend to produce more intense hot spots,24 a study of average interstitial gap sizes between 16 nm and greater than 100 nm, holding all other parameters equal, confirmed that even small variations in gap size resulted in noticeable changes in signal intensity (Figure 3c). The smallest gap used, 16 nm, was still considerably greater than the small gaps (∼1 nm) below which quantum effects cause the SERS enhancement to decrease with decreasing gap size.25 As expected, the smallest interstitial gaps resulted in the largest observed SERS intensities. The effect of gold layer thickness was also investigated by

incrementally varying the thickness of the Au overlayer from 50 to 120 nm (Figure 3d). Since reduced conductance of the metal will result in less effective confinement of the optical field,26 the thinnest gold layer was chosen to be approximately equal to the mean free path of the conductance electrons in gold, at and below which the metal’s conductance would be reduced, broadening the surface plasmon resonance. The SERS intensity, indeed, increased monotonically with the thickness of the gold layer (keeping all other parameters constant, including the gap size). Not unexpectedly, the SERS intensity is significantly more sensitive to the interstitial gap size than to the gold layer thickness. Optimizing their nanostructural parameters yielded substrates with SERS enhancements of ∼107 uniformly across the chip using 633 nm excitation, when the grating pitch is greater than ∼300 nm, the gap size less than 20 nm, and the gold mass thickness at least 80 nm. Smaller gap sizes (but >1 nm) would produce even greater enhancements; however, such small interstices would restrict the size of the molecules that could be investigated. Many proteins and other biomolecules or polymetric species would not be able to penetrate the 7569

DOI: 10.1021/acsnano.6b02564 ACS Nano 2016, 10, 7566−7571

Article

ACS Nano

used were specifically chosen or developed to be directly applicable to large-scale wafer substrates.

interstices. Somewhat greater enhancements might be possible with a hexagonal rather than square grating and by continued fine-tuning of the structural variables of the grating. The improvements will be rather marginal, however, in a regime that is close to the highest SERS enhancements reported to date uniformly over so large a substrate area. Of course, it is possible to construct individual hot spots in which very small molecules might produce greater SERS enhancements;27,28 however, we believe what we report here are among the highest SERS enhancements uniformly measurable across an area >1 cm2, using a technology that has been demonstrated in other contexts to be scalable to large wafer formats.20 The substrates’ SERS intensity uniformity was determined by measuring the intensity distribution function obtained from a map of the SERS intensities measured over several randomly chosen 100 × 100 μm2 portions of their surface, an example of which (mean = 7.4 × 106; fwhm = 3.6 × 106) is shown in Figure 4. We find that 95% of the surface yields SERS

METHODS Fabrication. The SERS substrates were produced from a square two-dimensional grating fabricated by laser interference lithography on 15 mm × 15 mm silicon wafers. The Si wafers were washed by sonication for 3 min in acetone, 3 min in methanol, and 3 min in deionized water then baked dry at 115 °C for 10 min. The wafer was spin-coated with an antireflective coating (XHRiC-11, Brewer Science, Inc.), baked dry, then spin-coated with a positive i-line photoresist (THMR-iP3600-HP, Ohka America, Inc.). The photoresist was exposed with a 55 mJ dose from a 325 nm HeCd laser using a laser interference lithography setup. Two exposures were made at a 90° angle to one another. After postexposure processing, a square two-dimensional array of 120 nm diameter photoresist posts remain on the surface with a pitch of 330 nm in both directions. The pattern was transferred into the underlying silicon using reactive ion etching by first etching away the exposed antireflective coating in an oxygen plasma then etching the exposed silicon using a modified Bosch process (plasma consisting of SF6, C4F8, and Ar simultaneously).29 Upon removal of residual photoresist and antireflective coating in warm piranha solution (3:1 mixture of sulfuric acid and hydrogen peroxide at approximately 40 °C), a square two-dimensional array of vertical silicon posts (the parent structure) remained with diameter 120 nm and height 350 nm. The parent structure was overcoated with a layer of silica by plasmaenhanced chemical vapor deposition (PECVD) to produce more closely spaced features. The structure was completed by depositing a layer of gold over a 4 nm titanium adhesion layer. Metals were deposited at 18 MΩ (Millipore) water, then washed for 1 min in a stream of deionized water and blown dry with argon gas, to remove unadsorbed molecules. SERS spectra were collected in backscattering mode (Horiba JY Aramis Raman microscope equipped with a 633 nm HeNe laser) with an incident laser power of 7.6 μW and exposure time of 0.25 s using a 50× long working distance objective. The spectra obtained for thionine agreed with literature spectra. Uniformity and reproducibility studies were carried out using xy SERS mapping. A computercontrolled translation stage was used to collect spectra over a 100 × 100 μm2 area at 5 μm intervals in both directions (i.e., 441 spectra were used to determine statistics).

Figure 4. A typical distribution of the intensities measured over a 100 μm × 100 μm area of a substrate taken at 5 μm intervals in the x and y directions. The average SERS intensity is 7.4 × 106 ± 3.6 × 106 counts (0.25 s exposure at 7.6 μW). The inset shows the spatial distribution of intensities over the mapped area. The area shown is from sample N, which was fabricated with a grating pitch of 330 nm, a gap size of 17 nm, and a gold mass thickness of 80 nm.

intensities that vary by factors less than a factor of 2, so far the most uniform SERS substrate reported and very greatly more uniform than what is obtained for substrates comprised of aggregated noble metal nanoparticles in which the SERS-active assemblies vary from monomers to large aggregates whose SERS enhancements can vary by a factor of 103 or more.22

CONCLUSIONS Summarizing, SERS substrates with highly uniform enhancements equal to or greater than 107 were fabricated using routine silicon foundry techniques, based on two-dimensional gratings of Si nanoposts produced using interference lithography, overcoated with silica and gold to create structures with optimized grating pitch and small interstices between the gold nanofeatures. The resulting structure yields the highest SERS enhancement reported to date for a structure that is reproducibly manufacturable with high accuracy and uniformity. The fabrication techniques and engineering principles

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.6b02564. Enhancement factor calculation repeated with 1,2-bis(4pyridyl) ethylene and derivation of the quantity Ro = r̂o*r̂o 7570

DOI: 10.1021/acsnano.6b02564 ACS Nano 2016, 10, 7566−7571

Article

ACS Nano

(16) Lamprecht, B.; Schider, G.; Lechner, R. T.; Ditlbacher, H.; Krenn, J. R.; Leitner, A.; Aussenegg, F. R. Metal Nanoparticle Gratings: Influence of Dipolar Particle Interaction on the Plasmon Resonance. Phys. Rev. Lett. 2000, 84, 4721−4724. (17) Chu, Y.; Schonbrun, E.; Yang, T.; Crozier, K. B. Experimental Observation of Narrow Surface Plasmon Resonances in Gold Nanoparticle Arrays. Appl. Phys. Lett. 2008, 93, 181108. (18) García-Vidal, F. J.; Pendry, J. B. Collective Theory for Surface Enhanced Raman Scattering. Phys. Rev. Lett. 1996, 77, 1163−1166. (19) Kocabas, A.; Ertas, G.; Senlik, S. S.; Aydinli, A. Plasmonic Band Gap Structures for Surface-Enhanced Raman Scattering. Opt. Express 2008, 16, 12469−12477. (20) Wu, Q.; Liu, S.; Zhang, X.; Tai, S.; Du, X.; Tombler, T. Short Pitch Metal Gratings and Methods for Making the Same. U.S. Patent 8,506,827 B2, August 13, 2013. (21) Hutchinson, K.; Hester, R. E.; Albery, W. J.; Hillman, R. Raman Spectroscopic Studies of a Thionine-Modified Electrode. J. Chem. Soc., Faraday Trans. 1 1984, 80, 2053−2071. (22) Le Ru, E. C.; Blackie, E.; Meyer, M.; Etchegoin, P. G. Surface Enhanced Raman Scattering Enhancement Factors: A Comprehensive Study. J. Phys. Chem. C 2007, 111, 13794−13803. (23) Azzam, R. M.; Bashara, N. M. Ellipsometry and Polarized Light; North-Holland Publishing Company: New York, 1987. (24) Long, D. A. The Raman Effect. A Unified Treatment of the Theory of Raman Scattering by Molecules; John Wiley & Sons: New York, 2002. (25) Zuloaga, J.; Prodan, E.; Nordlander, P. Quantum Plasmonics: Optical Properties and Tunability of Metallic Nanorods. ACS Nano 2010, 4, 5269−5276. (26) Liver, N.; Nitzan, A.; Gersten, J. I. Local Fields in Cavity Sites of Rough Dielectric Surfaces. Chem. Phys. Lett. 1984, 111, 449−454. (27) Fang, Y.; Seong, N.-H.; Dlott, D. D. Measurement of the Distribution of Site Enhancements in Surface-Enhanced Raman Scattering. Science 2008, 321, 388−392. (28) Laurence, T. A.; Braun, G. B.; Reich, N. O.; Moskovits, M. Robust SERS Enhancement Factor Statistics Using Rotational Correlation Spectroscopy. Nano Lett. 2012, 12, 2912−2917. (29) Curtin, B. M.; Fang, E. W.; Bowers, J. E. Highly Ordered Vertical Silicon Nanowire Array Composite Thin Films for Thermoelectric Devices. J. Electron. Mater. 2012, 41, 887−894. (30) Dignam, M. J.; Moskovits, M.; Stobie, R. W. Specular Reflectance and Ellipsometric Spectroscopy of Oriented Molecular Layers. Trans. Faraday Soc. 1971, 67, 3306−3317.

as used in the analysis of spectroscopic ellipsometry measurements (PDF)

AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This work was supported by the MRSEC Program of the National Science Foundation under Award No. DMR 1121053 and by the Institute for Collaborative Biotechnologies through Grant W911NF-09-0001 from the U.S. Army Research Office. REFERENCES (1) Porter, M. D.; Lipert, R. J.; Siperko, L. M.; Wang, G.; Narayanan, R. SERS as a Bioassay Platform: Fundamentals, Design, and Applications. Chem. Soc. Rev. 2008, 37, 1001−1011. (2) Sylvia, J. M.; Janni, J. A.; Klein, J. D.; Spencer, K. M. SurfaceEnhanced Raman Detection of 2,4-Dinitrotoluene Impurity Vapor as a Marker to Locate Landmines. Anal. Chem. 2000, 72, 5834−5840. (3) Brosseau, C. L.; Casadio, F.; Van Duyne, R. P. Revealing the Invisible: Using Surface-Enhanced Raman Spectroscopy to Identify Minute Remnants of Color in Winslow Homer’s Colorless Skies. J. Raman Spectrosc. 2011, 42, 1305−1310. (4) Yan, F.; Vo-Dinh, T. Surface-Enhanced Raman Scattering Detection of Chemical and Biological Agents Using a Portable Raman Integrated Tunable Sensor. Sens. Actuators, B 2007, 121, 61− 66. (5) Banholzer, M. J.; Millstone, J. E.; Qin, L.; Mirkin, C. A. Rationally Designed Nanostructures for Surface-Enhanced Raman Spectroscopy. Chem. Soc. Rev. 2008, 37, 885−897. (6) Kneipp, K.; Kneipp, H.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Ultrasensitive Chemical Analysis by Raman Spectroscopy. Chem. Rev. 1999, 99, 2957−2976. (7) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Single Molecule Detection Using SurfaceEnhanced Raman Scattering (SERS). Phys. Rev. Lett. 1997, 78, 1667− 1670. (8) Toderas, F.; Baia, M.; Baia, L.; Astilean, S. Controlling Gold Nanoparticle Assemblies for Efficient Surface-Enhanced Raman Scattering and Localized Surface Plasmon Resonance Sensors. Nanotechnology 2007, 18, 255702−255707. (9) Jiao, Y.; Ryckman, J. D.; Koktysh, D. S.; Weiss, S. M. Controlling Surface Enhanced Raman Scattering Using Grating-Type Patterned Nanoporous Gold Substrates. Opt. Mater. Express 2013, 3, 1137−1148. (10) Dickey, M. D.; Weiss, E. A.; Smythe, E. J.; Chiechi, R. C.; Capasso, F.; Whitesides, G. M. Fabrication of Arrays of Metal and Metal Oxide Nanotubes by Shadow Evaporation. ACS Nano 2008, 2, 800−808. (11) Driskell, J. D.; Shanmukh, S.; Liu, Y.; Chaney, S. B.; Tang, X.-J.; Zhao, Y.-P.; Dluhy, R. A. The Use of Aligned Silver Nanorod Arrays Prepared by Oblique Angle Deposition as Surface Enhanced Raman scattering Substrates. J. Phys. Chem. C 2008, 112, 895−901. (12) Haynes, C. L.; Van Duyne, R. P. Nanosphere Lithography: A Versatile Nanofabrication Tool for Studies of Size-Dependent Nanoparticle Optics. J. Phys. Chem. B 2001, 105, 5599−5611. (13) Moskovits, M. Surface-Enhanced Spectroscopy. Rev. Mod. Phys. 1985, 57, 783−826. (14) Yan, B.; Thubagere, A.; Premasiri, W. R.; Ziegler, L. D.; Dal Negro, L.; Reinhard, B. M. Engineered SERS Substrates with Multiscale Signal Enhancement: Nanoparticle Cluster Arrays. ACS Nano 2009, 3, 1190−1202. (15) Meier, M.; Wokaun, A.; Liao, P. F. Enhanced Fields on Rough Surfaces: Dipolar Interactions Among Particles of Sizes Exceeding the Rayleigh Limit. J. Opt. Soc. Am. B 1985, 2, 931−949. 7571

DOI: 10.1021/acsnano.6b02564 ACS Nano 2016, 10, 7566−7571