Morphological investigations by FE-SEM and HR-TEM

Figure 3-2 shows cross-sectional images of the vertically aligned SWNT film grown on a quartz substrate acquired by a Hitachi FE-SEM S-900 at 4 kV at a tilted (upper row) and horizontal direction (lower row). The vertical alignment of SWNTs with an approximate height of 6 µm is observed along with their spatial uniformity. Typically, the film is optically uniform over the entire area of the substrate both in terms of appearance and the results of reflectance interferometry (data not shown). This alignment is believed to arise from the significantly high density of the SWNTs, as discussed below.

Fig. 3-2. Cross-sectional image of vertically aligned SWNT film with approximate thickness of 5 µm grown on a quartz substrate. The images were taken at fractured edge of the substrate by Hitachi FE-SEM S-900 at 4 kV, from tilted (upper row) and horizontal direction for the substrate (lower row).

Figure 3-3 shows higher magnification images of the vertically aligned film with an approximate thickness of 2 µm acquired by a Hitachi FE-SEM S-5200 at 2 kV for (a) the entire cross section, (b) the upper part of the film, (c) the mid cross section, and (d) the lower part of the film. Images (a), (b), and (d) are acquired at a 30° tilted angle from the horizon, and (c) is acquired from the horizontal direction. In the magnified image shown in Fig. 3-3c, most of the SWNTs are bundled and their typical diameter is ~15 nm. The degree of alignment as well as optical anisotropy in the film will be investigated and discussed in Section 4.1.

Figure 3-3d shows the morphology of the bottom of the SWNT film and clearly reveals that thinner bundles, with diameters of ~4 nm just above the substrate surface, coalesce into thicker bundles as the SWNTs grow away from the substrate. Furthermore, as shown in Fig.

3-3(d), it appears as though the roots of the SWNTs are pulled upward, presumably by an

Fig. 3-3. High-magnification images of vertically aligned SWNT film with approximate thickness of 2 µm grown on a quartz substrate, for (a) whole cross-section, (b) upper part of the film, (c) mid cross-section, and (d) lower part of the film. The images were taken at fractured edge of the substrate by Hitachi FE-SEM S-5200 at 2 kV. Images (a, b, d) are taken from 30° from the substrate plane and (c) is taken from horizontal direction to the substrate.

internal frictional force among the SWNT bundles. This internal force is considered to originate from the difference in the catalyst activity at a later stage of the growth where the activity begins to deteriorate (refer to Section 3.6 for details). This is because a difference in catalyst activity can cause a difference in the extending speed depending on the SWNT bundle. Further, it should be noted that on the surface of the quartz substrate, aggregated metal particles that would be observed if the diameter exceeded 2–3 nm by FE-SEM S-5200 were not observed. This result is consistent with the HR-TEM image of the substrate surface shown in Section 2.4.

Figure 3-4 shows the HR-TEM image of the vertically aligned SWNT film. The film was weakly sonicated in methanol for 5–10 s and dropped onto a TEM microgrid. Observations were performed on the film fragments that retained their alignment to a large extent. Two types of film edges were observed. One is shown in Fig. 3-4a in which the degree of alignment is relatively high and almost no metallic particles were found among the SWNTs.

Fig. 3-4. HR-TEM images of a fragment of a vertically aligned SWNT film deposited on a TEM grid after a weak sonication in methanol for 5 - 10 s, taken by JEOL 2000EX at 120 kV. Images are those presumably correspond to (a) the upper side of the film and (b-d) lower side of the film, as observed by FE-SEM in Fig. 3-3. Panels (c, d) are the magnification from the edge observed in the panel (b).

The other type is shown in Fig. 3-4b, in which abundant particles were found among the SWNTs and the degree of alignment is less than that shown in Fig. 3-4a. Figure 3-4c and d are magnifications of the film fragment edge shown in Fig. 3-4b. These images revealed metallic particles with diameters of 2–3 nm, and they are considered to be cobalt carbide (Co2C). However, a determination of the elemental composition of the particle using an experimental investigation technique such as EELS analysis is necessary for further discussion.

Figure 3-5 shows another HR-TEM image of the vertically aligned SWNT film. In this case, however, the film was directly rubbed onto a TEM microgrid (i.e., the film had not undergone dispersion in solvent) in order to precisely investigate the morphology and quality of the SWNTs. This observation was conducted in a manner similar for the case shown in Fig. 3-4 but at a higher magnification; the film fragment that retained alignment was found and investigated. Figure 3-5a shows the upper edge of the observed film

Fig. 3-5. HR-TEM images of a fragment of the vertically aligned SWNT film directly rubbed against a TEM grid, taken by JEOL 2000EX at 120 kV. Images are (a-c) Images are those presumably corresponds to (a-c) the top side of the film taken at various magnification, and (d) lower side of the film in which metal particles are observed among SWNTs.

fragment at a lower magnification, and the entire image of the film fragment is indicated in the inset. The morphology observed in Fig. 3-5a coincides with that of the upper side of the film observed by FE-SEM shown in Fig. 3-3b. Figure 3-5b is a magnification of the upper edge of Fig. 3-5a, where closed and empty SWNT ends are occasionally observed; these are further magnified in Fig. 3-5c. Similar to the case of Fig. 3-4, two types of edges in the film fragments were confirmed in the observation shown in Fig. 3-5, and one of these is shown in Fig. 3-5a, b, and c. The other type of edge is shown in Fig. 3-5d and is considered to be a peeled interface between the film and the substrate where the darkened area toward the right is the film, which is oriented perpendicular to the page. Particles with diameters of 2–3 nm are observed among the entangled SWNTs, as shown in Figs. 3-4c and 3-4d; these are considered to be Co₂C.

From the HR-TEM observations shown in Figs. 3-4 and 3-5, the growth of the SWNT film in the present study is ascribed to the base growth from the catalyst particles that remain on the substrate. These images, especially those observed for the film that did not undergo dispersion in the solvent, as shown in Fig. 3-5, indicate that each SWNT is sufficiently clean, i.e., virtually free from amorphous carbon and other impurities on the SWNT wall. No MWNTs were observed in repeated TEM observations whereas several double-walled carbon nanotubes (DWNTs) were found; however, their population is negligible compared with that of SWNTs. By measuring the diameter of more than 50 SWNTs, the diameter of the specimen was determined to be in the range of 0.8–3.0 nm with an average diameter d_av of ~1.9 nm and a standard deviation σ of ~0.4 nm. This diameter is larger when compared with that of typical CVD-grown SWNTs; in addition, the diameter distribution is broader.