IntroductionI remember the excitement when the SERS effect was discovered in the late 1970s (1,2). The Raman scattering cross section of pyridine adsorbed on a silver electrode was enhanced by 5 or 6 orders of magnitude! It was quickly understood that this represented plasmon local electromagnetic (EM) field enhancement on a rough surface, as explained below (3,4) . Actually two different EM fields are enhanced, the incoming laser field and the outgoing Stokes shifted field, making SERS especially dramatic. At that time I collaborated with Abe Nitzan on the related theoretical question: whether enhanced fields would create enhanced photochemistry(5,6). Certainly enhanced fields create an enhanced optical absorption cross section. But an excited molecule near a metal surface will undergo Forster energy transfer into the metal, and this process competes with photochemistry if the molecular excited state lifetime is long. As it turns out the energy transfer rate falls off faster than does field enhancement, and so photochemistry can be enhanced at some distance above the surface.After the initial excitement, experimental progress in SERS slowed, principally due to the extreme averaging that occurs in ensemble SERS measurements. On rough surfaces and in colloids, there is a huge distribution of molecular positions and orientations with respect to the surface, and also a wide range of different field enhancements due to different local metal topologies. I went on to work with semiconductor nanocrystals, where the issue was size-dependent development of band structure(7). Here a similar terrible ensemble averaging occurs over nanocrystal sizes, shapes and surface stiochometry. In the 1990s Betzig and Trautman created a significant advance in spectroscopic technique: observation of luminescence from spatially resolved single molecules and nanocrystals at 23 C on a surface, at first using fiber optic near field methods, and subsequently confocal far field methods (8,9,10, 11). Luminescence blinking, a signature of single molecule kinetics, was discovered in CdSe nanocrystals.In 1997 there were two unexpected reports of single molecule SERS, for dye molecules in Ag colloids(12, 13) . The combined enhancement of molecular resonance Raman and SERS was perhaps 14 orders of magnitude! The SERS single molecule Raman signal was actually at least an order of magnitude larger than a dye molecule or semiconductor nanocrystal luminescence signal in the absence of Ag. We began to explore this discovery, using the new confocal optical methods. There are two inter-related questions: where is the single molecule on the metal, and how is this molecule so strongly coupled to the excited metallic electrons. We found that the single molecule is in a junction between particles. Study of the coupling lead us to literature quantum mechanical models for photodynamics of chemisorbed molecules. We then experimentally discovered that adsorbed citrate surfactant anions, whose SERS signals are very weak compared to the dye molecule, are actually ...