New technique extends nano-Raman imaging to semiconductors
Ongoing progress in nanotechnology demands characterization techniques capable of investigating nano-scale components. Optical techniques are powerful tools for macro- and microscopic materials, but the diffraction limit of light and low sampling volumes physically limit the use of traditional optics for analysis of nanoscale materials. Raman spectroscopy, which uses laser light to probe molecular structures, is a useful technique, particularly for mapping crystal orientation and strains in silicon (Si) structures. To extend Raman spectroscopy to nanoscale imaging, however, the diffraction limit and low signal intensity must be overcome.
Two branches of nano-optics—aperture-limited and apertureless microscopy—have made significant strides toward nanoscale resolution. For Raman analysis, the aperture-limited techniques suffer from prohibitively low optical throughput for apertures less than 100nm in diameter. Apertureless microscopy, however, uses a sharp probe to disrupt the interaction between a sample and the light illuminating it. Several groups have used apertureless Raman spectroscopy, or tip-enhanced Raman spectroscopy (TERS), to achieve scanning nano-Raman resolution down to ∼15–20nm.1,2
The technique is based on strong local enhancement of an optical signal in the nanoscale vicinity of metallized tips. Top, bottom, and side illumination schemes have been developed, where bottom-illumination requires transparent substrates and samples and top-illumination does not capture the most strongly enhanced signal. Our approach, side-illumination TERS, has some disadvantages, but it appears to be the only scheme capable of working with semiconducting structures and/or samples on Si wafers.
The apertureless approach utilizes a metallized (typically gold or silver) scanning probe microscopy tip in the focal spot of a light source. Illumination of the sharp metal tip induces surface plasmons (collective electron resonances), which in turn increase the amplitude of the electric field in the vicinity of the tip. The localization of the enhanced electric field (near-field) is typically limited to within ∼10–20nm from the tip. Thus, the apertureless probe enhances and localizes the optical signal in its vicinity. For effective TERS imaging, one needs sufficient contrast between Raman signals from the tip vicinity (TERS signal) and from the rest of the illuminated area (far-field signal). Previously recorded contrast has been sufficient for the analysis of individual nano-objects such as carbon nanotubes1 and DNA molecules,2 but it is rather weak for the analysis of more complex samples, in particular Si-based structures.
To improve TERS contrast on silicon, we applied an idea recently proposed by Poborchii et al.,3 that of using depolarized scattering geometry to suppress the far-field signal. In this arrangement, the far-field Raman signal from the strongly polarized 521cm-1 mode of Si was suppressed by choosing an optimum combination of incident and scattered polarizations. We found that the near-field Raman signal was less dependent on polarization than the far-field, possibly due to depolarization of light scattered from the tip. As a result, a significant increase in contrast has been achieved in optimized polarizations.4
Quantification of depth and lateral resolution are crucial to the application of TERS. By analyzing thin films of cadmium sulfide with different thicknesses, we estimated the depth of the enhanced field to be about 20nm.5 Surface localized enhancement was also observed for a layer of strained Si on a silicon wafer.6
As shown in Figure 1(a), periodic silicon oxide (SiO) on silicon structures with 30nm thickness were used for nano-Raman imaging. The thickness of the oxide layer provides enough separation between the tip and the silicon to reduce the TERS signal of the 521cm-1 Si mode. Traditional confocal micro-Raman imaging was unable to resolve the SiOx features, but they appear clearly in a linescan using TERS: see Figure 1(b).
Figure 2 presents 2D images of the topography and the TERS signal of the 521cm-1 Si mode, which were obtained simultaneously. The lineshapes of the topographic and Raman images were nearly identical, indicating Raman lateral resolution comparable to the topography. We estimate our lateral resolution to be ∼20nm. A similar analysis was applied to poly(methyl methacrylate) structures on cadmium sulfide; the same lateral resolution of ∼20nm was estimated.4
Apertureless near-field Raman spectroscopy has made significant strides toward nanoscale resolution by overcoming the diffraction limit of light and enhancing intrinsically weak signals. This technique will greatly improve the characterization of the physical and chemical properties of nanostructures and devices in the semiconductor and communications industries. Results obtained from our side-illumination apertureless Raman spectrometer suggest that mapping semiconductor structures with a lateral resolution on the scale of 20nm is possible. We will use this exceptional resolution to investigate strain distribution in silicon nanostructures. Sharper tips should improve resolution, and efforts to increase contrast will include developing and understanding the optical properties of the tips as well as matching their plasmon resonance frequencies with the frequencies of the Raman photons.7
The authors acknowledge financial support from the National Science Foundation, the Ohio Board of Regents, and the Cooperative Center for Polymer Photonics, which is co-funded by the Air Force Research Laboratory, Air Force Office of Scientific Research, and The University of Akron. This material is based upon work partially supported under a National Science Foundation Graduate Research Fellowship. Any opinions, finding, conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Science Foundation.