Theory for Modeling of High Resolution Resonant and Nonresonant Raman Images

Tip-enhanced Raman imaging is capable of resolving the inner structure of a single molecule owing to the generation of highly localized nanocavity plasmon. Here we present a general theory and detailed computational methodology to fully describe resonant and nonresonant Raman scattering under the localized plasmonic field. We use an allylcarbinol molecule adsorbed on the gold surface as a model system to illustrate different effects on the Raman images. It is found that the ability of distinguishing an individual vibration mode is highly limited under the resonant condition due to the dominant contribution from the Franck-Condon term and the mode-independent component of the Herzberg-Teller term. The nonresonant Raman images of the single molecule are vibrationally distinguishable and present the vibrational motion of the corresponding vibrational modes in real space. Furthermore, the calculated results confirm that nonlinear optical effects can further improve the resolution of the images. The theoretical and computational methods presented here provide the basic tools to model high resolution Raman images at the single molecular level.