<p>Surface enhanced Raman spectroscopy (SERS) is undoubtedly a powerful tool as it allows one to overcome the major disadvantage of Raman spectroscopy: the weakness of its signal. Enhancement factors (EF) of up to 1010 make it even possible to detect single molecules. However, using it as an analytical tool to make reproducible, quantitative measurements has so far been difficult as the enhancement of the signal is "bought" at the expense of reproducibility: The larger the EF the more the reproducibility of the substrate suffers. This has been dubbed informally the "SERS uncertainty principle" by Natan [1]. While currently a lot of research effort is taking place at the high-EF-side of the spectrum and ever more sophisticated SERS substrates are being explored, in this thesis we would like to make a shift in paradigm and revisit SERS on flat metallic surfaces, which arguably constitute the simplest substrates available. To this end we will show their usefulness in making quantitative measurements and how they are an ideal platform for a new hybrid technique that combines reproducibility and extreme sensitivity with substantial EFs. For making quantitativemeasurements two examples are explored in a systematic way: in the first example (Chapter 2) the determination of an unknown, resonant Raman cross-section is demonstrated on flat metallic films (possibly with some surface roughness) and confirmed with measurements done on more commonly used SERS substrates. Here the quantitative measurement is made possible by introducing a reference molecule as a standard and having statistics as our main ally: even though we do not know the exact EF that the individual molecules experience on the various substrates, we know that on average both, the unknown sample and the known reference, experience the same. In the second example (Chapter 3) we use commercially available flat films for which we verify experimentally that surface roughness is irrelevant. By themselves these substrates yield no enhancement – in fact they even quench the Raman signal. Yet they allow us to calculate and control the electric field on the surface which enables us to determine the orientation of adsorbed molecules by using surface selection rules (SSR). While the first example is mostly empirical, the second one allows us to test our theoretical understanding of plasmonic systems with proper numerical calculations that are in excellent agreement with the observed data. Finally, in Chapter 4, we use those flat films in a special configuration (called the Kretschmann configuration) to excite Surface Plasmon-Polaritons (SPP). This not only allows us to combine the spatial homogeneity of a flat surface with useful EFs easily predicted fromtheory but also to combine the extreme sensitivity of surface plasmon resonance spectroscopy (SPRS) with the analytical power of SERS. It is not our intention to claim that the work presented here is the first attempt to do analytical work with SERS. Rather the newmethods presented in this thesis will add new strategies and tools to the current research effort while the detailed analysis will provide the means to understand them theoretically and in their historical context.</p>
Hybrid Sol-Gel Surface-Enhanced Raman Sensor for Xylene Detection in Solution