Reproducible SERRS from structured gold surfacesw Sumeet Mahajan, a Jeremy J. Baumberg, b Andrea E. Russell a and Philip N. Bartlett* a Received 7th August 2007, Accepted 24th September 2007 First published as an Advance Article on the web 10th October 2007 DOI: 10.1039/b712144j Metallic substrates with ordered spherical cavities have been shown to be very effective for surface-enhanced Raman scattering (SERS) and can be fabricated reproducibly using electrodeposition. The sensitivity of detection is increased by several orders of magnitude by using surface-enhanced resonance Raman scattering (SERRS). In this report we demonstrate SERRS for the first time on electrodeposited gold films templated with colloidal spheres and demonstrate the reproducibility of the response. We also obtain a direct comparison between SERRS and SERS by choosing two dyes, Cy5t and Cy3t, which are similar in structure but differ in their excitation maxima, such that one is resonant and the other non-resonant with our laser excitation. As expected, the resonant enhancement is found to be of the order of 10 3 over and above that for SERS. The net SERRS enhancements are shown to be of the order of 10 9 . We also find that the resonant enhancement profile of the different peaks for the chromophore follows the plasmonic resonance absorption spectrum obtained for the structured surface. Introduction After its first observation in 1974 by Fleischmann et al., 1 the phenomenon of surface-enhanced Raman scattering (SERS) was recognised by Albrecht and Creighton 2 and Jeanmarie and Van Duyne. 3 Surface-enhanced resonance Raman scatter- ing (SERRS) was demonstrated by Jeanmarie and Van Duyne 3 in the same paper. The authors noted the extremely low laser powers required for acquiring the surface Raman spectra under the conditions of molecular resonance with the laser. When the incident laser is in resonance with an electro- nic transition of a molecule in proximity to a surface-enhan- cing substrate, the cross-section of the molecule is greatly enhanced and SERRS is observed. SERS itself is a very sensitive and selective technique for detecting surface species. SERS enhancements of the order of 10 6 or greater are routinely reported. The SERS and resonance effect are multi- plicative and can give overall intensity enhancements in the range of 10 4 to 10 8 for SERRS. 4–6 Thus, use of SERRS raises the sensitivity further by several orders of magnitude, offering the possibility of single-molecule detection. 7,8 The extreme sensitivity of SERRS, giving molecule-specific information, is unparalleled amongst analytical techniques. This makes it ideal for the study of biomolecules. Thus SERRS has been used to study heme proteins, 9 to study the mechanism of electron transfer in cytochromes, 10 and more recently for trace detection of DNA 11,12 and enzyme activity. 13 The sensitivity of SERRS exceeds that of fluorescence, which is currently the most prolific technique employed for studying biomolecules, as proven theoretically 14 and experimentally. 15 SER(R)S has another immense advantage over fluorescence in that the signals have extremely small line-widths compared to the broad spectra of fluorescence emission; and hence it is possible, using SER(R)S labels, to carry out multiplexed detection. 16 Despite more than 30 years of vigorous activity, a complete understanding of the enhancement mechanism of SER(R)S and wide commercial application has remained elusive. The bane of understanding and commercial exploitation has been, primarily, the irreproducibility of the substrates for SER(R)S. The bulk of the work on SER(R)S has been carried out using electrochemically-roughened silver electrodes and colloidal silver nanoparticles. Over the last decade researchers have focused on discovering and improving methods for fabricating reproducible substrates for SER(R)S. Amongst these, vapour- deposited silver films 17 and silver colloids 18 have been the most heavily researched. There are relatively few reports 12,19 of SERRS on gold substrates. Nevertheless, gold is attractive as a substrate because, although the surface enhancements are B10 times lower than for comparable silver surfaces, gold surfaces are more stable than their silver counterparts. Contrary to claims, most substrates are inherently not reproducible. Also, in most of the techniques there is no direct control over the formation of the metallic structures. In the past few years, we have demonstrated our ability to fabricate nanostructured metallic films using colloidal tem- plated electrodeposition. 20 These have been proven to generate tunable plasmon resonances 21 for SERS. 22,23 We have recently shown the tunability of our substrates for SERS even with a near-infrared source (1064 nm laser). 22 We have also demon- strated SERS on our substrates fabricated with palladium and platinum, 24 proving the flexibility of our technique. In our a School of Chemistry, University of Southampton, Southampton, UK SO17 1BJ. E-mail: pnb@soton.ac.uk; Fax: +44 (0)2380 593781; Tel: +44 (0)2380 592373 b School of Physics & Astronomy, University of Southampton, Southampton, UK SO17 1BJ w Electronic supplementary information (ESI) available: Absorption spectrum of both the dyes (Fig. S1), a representative cyclic voltammo- gram (Fig. S2) and a typical SEM image (Fig. S3) of the structure. See DOI: 10.1039/b712144j 6016 | Phys. Chem. Chem. Phys., 2007, 9, 6016–6020 This journal is  c the Owner Societies 2007 PAPER www.rsc.org/pccp | Physical Chemistry Chemical Physics