Using a nano light source to investigate small-scale composite materials

We are currently witnessing a strong trend toward developing artificial materials with nanoscale structures. For example, single-walled carbon nanotubes (SWNTs) doped with organic molecules are useful for controlling electrical conductivity, optical switching, and nonlinear media.^1–3 Because such materials consist of two or more different molecular components (hence the term ‘nanocomposites’), characterizing them requires a technique that combines chemical analysis with nanometer-scale spatial resolution.

A variety of methods exist for investigating and characterizing nanoscale structures, but few are suitable for molecular analysis. For instance, transmission electron microscopy enables visualization of objects inside SWNTs,⁴ but it cannot identify molecular species. One alternative is vibrational spectroscopy, including Raman spectroscopy. In addition to molecular identification, vibrational energy analysis provides detailed structural information for each species, such as intermolecular interactions, molecular orientation, and symmetry distortions.⁵ This makes Raman spectroscopy a powerful tool for studying the chemical composition of matter. By the same token, conventional confocal Raman spectroscopy techniques cannot attain spatial resolution on the nanoscale.

We have overcome this limitation through an approach we call tip-enhanced Raman spectroscopy (TERS). By introducing a sharp metal tip to the focus of a laser beam, we were able to localize Raman excitation to an area of 30nm².^6,7 The apex acts as a nanoscale light source, which couples with surface plasmon polaritons (charge density waves propagating along the metal surface) to increase both the incident light field and the Raman scattering signal. To demonstrate the technique we used samples of β−carotene encapsulated in SWNTs.⁸

Figure 1 shows a schematic experimental setup for TERS. The optical layout is a combination of Raman spectroscopy and scanning probe microscopy. TERS profiles were measured using a silver tip positioned close to the sample, and reference spectra were also obtained at the same position without the silver tip. The detailed experimental setup has been described elsewhere.⁸ Figure 2(a) shows atomic force microscopy images of the samples. Although the images appear to be of single tubes, they are actually bundles of several tens of tubes aligned parallel to each other, as shown by the 15nm average heights of the samples detected. We chose two different bundles that were completely separated. The crosses (a–g) indicate the position of the tip during the TERS, which was performed in 100nm steps along the bundles.

Figure 1. Schematic illustration of the experimental setup. The inset shows β-carotene molecules encapsulated in carbon nanotubes.

Figure 2. (a) Atomic force microscopy image of β−carotene−encapsulated carbon nanotubes. The scale bars indicate 50nm. (b) Near−field spectra of the sample measured at positions a–g, respectively.

Figure 2(b) shows near−field Raman spectra of the difference between having the silver tip in contact with the sample and without the tip. Spectra a–g correspond to the sample positions in Figure 2(a). The interesting feature of Raman spectroscopy is that we can simultaneously obtain data from the SWNTs and β−carotene. Figure 2(b) illustrates the frequency region, including both ν₁ (1523cm⁻¹ conjugated C=C mode) from β-carotene and the G-band (1592cm^-1 graphite mode) from carbon nanotubes. The most predominant feature among the spectra a–g is the absence of the encapsulated β−carotene in spectrum f. This means that the rate of encapsulation of β−carotene in the SWNTs was not uniform. The extremely low rate at position f could have occurred if the tubes were twisted during encapsulation or were filled with impurities. Another feature is that in spectrum c, the intensity of encapsulated β−carotene is higher compared with other spectra. Figure 2(a) shows the presence of another thin bundle entangled with a thicker one.

This example demonstrates that TERS is useful for avoiding the averaging of Raman spectra of nanocomposite materials that are not spatially uniform. The principle of this method is to use tip enhancement, which provides both spectral and topological information. This technique can be applied not only to surfaces but also to molecules enclosed inside nanostructures.