学术文献库

Recent studies have established that the anti-Stokes Raman signal from plasmonic metal nanostructures can be used to determine the two separate temperatures that characterize carriers inside the metal -- the temperature of photoexcited "hot carriers" and carriers that are thermalized with the metal lattice. However, the related signal in the Stokes spectral region has historically impeded surface enhanced Raman spectroscopy (SERS), as the vibrational peaks of adsorbed molecules are always accompanied by the broad background of the metal substrate. The fundamental source of the metal signal, and hence its contribution to the spectrum, has been unclear. Here, we outline a unified theoretical model that describes both the temperature-dependent behavior and the broad spectral distribution. We suggest that the majority of the Raman signal is from inelastic scattering directly with non-thermal carriers that have been excited via damping of the surface plasmon. In addition, a significant spectral component (~ 1%) is due to a sub-population of hot carriers in an elevated thermal distribution. We have performed temperature and power-dependent Raman experiments to show how a simple fitting procedure reveals the plasmon dephasing time, as well as the temperatures of the hot carriers and the metal lattice, in order to correlate these parameters with quantitative Raman analysis of chemical species adsorbed on metal surface.