Near-field Raman Imaging and Spectroscopy
Nanoscale Vibrational Analysis
Raman scattering cross-sections are
usually extremely small, being approximately 14 orders of magnitude
smaller than the typical fluorescence cross-section of single molecules.
We overcome this difficulty by carefully placing a sharp metal probe
in the vicinity of a laser spot in order to enhance the amplitude
of the Raman scattering strength (surface enhanced Raman scattering,
SERS). The signal enhancement is highly localized to the apex of
the metal probe (~20nm in diameter). The metal probe is positioned
into a tightly focused laser beam and the sample is raster scanned.
A Raman scattering spectrum is acquired for each image pixel from
which different Raman images can be extracted by integrating the
intensity associated with different modes. The technique allows
us to acquire multidimensional vibrational images with spatial resolution
Fig. 1: Near-field Raman image (intensity of the G band) and simultaneously
recorded topographic image of SWNTs on glass (scan area 2 x 2 um).
But not all SWNTs are wiggly
Fig. 2: Near-field Raman image recorded over the graphite-like G
band (1590cm-1) for a CVD grown SWNT.
Studying phonon localization in single-walled carbon
We use near-field Raman scattering
to study the localization of vibrational modes along individual
single-walled carbon nanotubes (SWNTs). The well-defined size and
shape of SWNTs offers the possibility for simultaneously localizing
individual nanotubes both topographically and optically using our
tip-enhanced imaging technique. Probing the Raman scattering spectrum
of SWNTs renders a unique chemical fingerprint from which detailed
information can be extracted, i.e. tube structure (n, m) (RBM),
defects (D band), metallic or semiconducting (G band & RBM).
Fig. 3: Near-field Raman image (G' band) of a spatially
isolated SWNT using a sharp gold tip. Also shown is a Gaussian fit
to the line section shown. The FWHM was determined to be 14nm.
The spatial resolution is limited solely by the size of the
metal probe. As shown in the figure above, we achieve spatial
resolutions on the order of ~ 15nm. To date our best resolution
is ~ 10nm. This was achieved using a gold wire electrochemically
etched to a sharp apex.
Fig. 4: Raman spectrum acquired from an
individual SWNT with (green) and without
(red) a metal tip present. From the increase in the Raman signal we
determine the enhancement factor to be on the order of 10^6.
Fig. 5: Near-field spectral images of a CVD grown
clearly observe a uniform
spatial profile for both the RBM (260cm-1) and graphite-like G
band (1587cm-1). However, we observe significant localization
associated with defect-induced (1280cm-1) and IFM (835cm-1) Raman
scattering at the same location along the nanotube shown.
Studying exciton localization in single-walled carbon
In addition to studying phonon localization
in carbon nanotubes we use our tip-enhanced technique to study the
spatial extent of photoluminescence (PL) emission from SWNTs. Studies
of SWNTs deposited directly on glass substrates reveal highly confined
photoluminescence (PL) emission from short segments of about 20nm
in length. For SWNTs embedded in micelles resting on MICA substrates
we find that the PL emission typically extends over several hundreds
of nanometers. By acquiring simultaneous near-field Raman and PL
images we aim to study possible correlations between structural
defects and the PL properties of individual SWNTs.
Fig. 6: (a) Near-field PL spectra acquired at the positions 1 -6 indicated in the near-field PL
image of the SWNT in (b)
Localized Stress Analysis in Semiconductor Nanostructures
We are also exploring ways to characterize stress in silicon
devices with nanoscale resolution. Localized stress affects
device performance and also limits the lifetime of integrated
circuits. Being able to image stress with nanoscale resolution
is of great importance for improved circuit design and for further
miniaturization of semiconductor circuits.