AbstractView references

Reflection-mode tip-enhanced Raman scattering (TERS) has the advantage to characterize any sample, particularly opaque, bulk, and multilayered samples. However, the background signal in reflection-mode TERS is huge due to large focus spots associated with an objective lens that has a long working distance. Moreover, for a multilayered and bulk sample, the Raman signal from the bulk layer interferes with the Raman signal on a thin surface layer. This unwanted bulk background signal reduces the sensitivity of the measurement and makes it difficult to get a high-contrast TERS image in the reflection mode. Here, we demonstrate two techniques to suppress the far-field Raman signals coming from the focus area and bulk silicon germanium substrate. First, we reduce the far-field signal by controlling the polarization state of the incident and scattered Raman as well as manipulating the well-defined polarization of a crystalline sample, which strongly depends on the polarization and propagation of the incident light. We used Raman tensor analysis to quantify the far-field Raman intensity at different polarization states of the polarizer and analyzer at varying sample azimuth. The results of the surface-enhanced Raman spectroscopy experiments were in good agreement with the far-field Raman tensor calculation. The polarizer, analyzer, and sample azimuth combination that gives minimum far-field background signals with high contrast in SERS was utilized in the TERS experiment to obtain a high-contrast near-field Raman signal. Both the field enhancement effect and depolarized detection were considered to obtain a high signal-to-noise TERS signal. We found that, for Raman active and thin crystalline samples assembled in bulk materials, the depolarization effect outweighs the field enhancement effect in getting a high-contrast Raman signal. © 2008 American Institute of Physics.