Figure 4.5 displays HRTEM images of the four GC samples. Each of the materials contain a number of layered graphitic nanostructures. The nanostructures in the images taken of the (a) Sigradur-K, (b) Sigradur-G, and (d) AA2200 GC samples are relatively abundant and are comprised of several layers of graphitic sheets. The structure of the (c) AA1100 sample appears to be much less developed, where the clear nanostructures are less abundant and comprised of fewer graphene sheets. One particular layered graphitic nanostructure has been selected from each of these images and has been magnified and is shown in the outsets in Fig. 4.5. The graphitic layer spacings within these magnified outsets are measured to be 3.65±0.02 ˚A, 3.60±0.02 ˚
A, 3.80 ± 0.02 ˚A, and (d) 3.65 ± 0.02 ˚A for the Sigradur-K, Sigradur-G, AA1100, and AA2200 nanostructures, respectively.
Shown in Fig. 4.6 are SADPs of each of the four GC samples acquired from the same lamellae as the HRTEM images. In each of the SADPs broad symmetric reflections are observed, which confirms the isotropic structure of the materials. All peaks have been indexed to graphite. The fact that the reflections appear sharper in the Sigradur-K SADP [shown in Fig. 4.6(a)] is possibly due to the lamella being thinner than the other three.
Figure 4.5: HRTEM images showing the isotropic graphitic nanostructure of the four GC samples. (a) Sigradur-K. (b) Sigradur-G. (c) AA1100. (d) AA2200. The layer spacings quoted in this figure have been measured from the individual nanostructures shown in the outsets. All images were taken using the same beam conditions and energy.
Figure 4.6: SADPs of the four GC samples. (a) Sigradur-K. (b) Sigradur-G. (c) AA1100. (d) AA2200. All reflections have been indexed to graphite.
The average layer spacing can be determined from the location of the{002}which are indicated by the white arrows in Fig. 4.6. The average layer spacing for Sigradur- K is measured to be 3.64 ± 0.02 ˚A, for Sigradur-G it is 3.59 ± 0.02 ˚A, for AA1100 it is slightly larger at 3.81 ± 0.03 ˚A, and for AA2200 it is determined to be 3.63 ±
0.02 ˚A. These measured values are different to the layer spacings determined from the HRTEM images, however the trend is the same, with the AA1100 having the largest average layer spacing. The variations between the values determined from the SADPs and TEM images can be accounted for by assuming that there is some variation between the individual layered structures within each sample.
X-ray diffraction
An XRD spectrum of each of the four GC samples is shown in Fig. 4.7. The peaks in all four spectra have been indexed to graphite. The most intense peaks correspond to the {002}, {100}, and {110} reflections and are located at 1.8 ˚A−1, 3.0 ˚A−1, and 5.2 ˚
A−1 respectively. A broad peak corresponding to the{004}reflection is located at 3.6 ˚
{101}reflection. The turbostratic layer stacking in GC leads to a significant reduction in the intensity of the {101} reflection (relative to graphite) [59]. Further to this, any reflections that require registration between graphene layers will be significantly broadened and reduced in intensity. The turbostratic layer stacking also shifts the
{002} peak to lower Q from 1.9 ˚A−1 (in graphite) down to 1.8 ˚A−1.
Figure 4.7: XRD spectra of the four GC samples. The spectra have been scaled to the height of their respective{002}peaks, and all labelled peaks are indexed to graphite. (inset) The {002} peaks of each spectra overlayed showing the broader asymmetric shape of the AA1100 {002} peak.
The peak locations in the spectra of the Sigradur-K, Sigradur-G, and AA2200 GC samples are relatively similar, with the peaks in the Sigradur-G spectrum being slightly sharper. The {002} peak width in the Sigradur-G spectrum is 0.34 ˚A−1,
relative to 0.38 ˚A−1and 0.37 ˚A−1for the Sigradur-K and AA2200 spectra, respectively.
This can be seen in the inset. This indicates that the nanostructures in the Sigradur- G sample are slightly more ordered, which matches well with the graphene layer spacings measured from HRTEM images. However, all of the peaks in the AA1100
spectrum are much broader, especially the {002} peak, which is evident from the inset in Fig. 4.7. The asymmetry and shift to the left of the {002} peak in the AA1100 spectrum indicates that there is a larger average graphitic layer spacing and greater variability between individual layered graphitic nanostructures. This result is consistent with the Raman spectroscopy measurements, HRTEM images, and SADPs presented earlier in this section.
Neutron diffraction
Total scattering neutron diffraction spectra, I(Q), of each of the four GC samples are shown in Fig. 4.8. The first few peaks in the spectra have been indexed to graphite, as can be seen in the magnified and labelled section of the Sigradur-G spectrum shown in the inset.
Figure 4.8: The neutron diffraction spectra, I(Q), of the four GC samples plotted from Q = 1 ˚A−1 to 30 ˚A−1. (inset) A magnified region of the Sigradur-G I(Q) spectrum,
with the main peaks indexed to graphite.
AA2200 are very similar out to 15 ˚A−1, and the relative peak intensities vary only slightly. In the AA1100 spectrum the peaks are broader and less intense. This spectrum also has a large underlying background, suggesting that the AA1100 sample has a highly disordered or amorphous component.
Radial distribution functions, G(r), were generated from the I(Q) spectra of each of the GC samples to determine nearest neighbour distances. To do this an I(Q) spectrum, such as the one shown in Fig. 4.9(a), must first be normalised to create the structure factor spectrum S(Q), as shown in Fig. 4.9(b). The S(Q) is then scaled appropriately to (minimise oscillations) before being Fourier transformed to produce the G(r) which contains real space data about the sample, as shown in Fig. 4.9(c). Details of this process are given in Section 3.2.4.
Figure 4.9: (a) The Sigradur-G I(Q) from Fig.4.8. (b) The structure factor S(Q) generated from the I(Q) after it has been appropriately scaled to minimise background oscillations, but before it is Fourier transformed to generate a (c) radial distribution function, G(r).
from the I(Q) shown in Fig. 4.8 by the method described in Fig. 4.9(a-c), are presented in Fig. 4.10.
Figure 4.10: The G(r) for each of the four GC samples plotted out to 11 ˚A. The G(r) were generated from the I(Q) shown in Fig.4.8. (inset) A magnified view of the G(r) of Sigradur-G showing the first three peaks, labelled a, b, and c, which correspond to the shortest interatomic distances shown in the hexagonal schematic drawn in the top left.
The locations of the first three peaks in each G(r) spectrum, which represent the interatomic distances within one 6 membered hexagonal ring, are remarkably similar. This section of the Sigradur-G spectrum has been magnified and is shown in the inset in Fig. 4.10. The three peaks are labelled a, b, and c, whose physical origins can be seen on the labelled hexagonal schematic shown in the top left corner of Fig. 4.10. There is some broadening of the peaks in the AA1100 G(r) spectrum which could be due to the contaminant elements or is potentially a by-product of the large background in the I(Q) spectrum.
Sigradur-K, Sigradur-G, and AA2200 samples are very similar out to 10 ˚A. This distance spans ∼4-5 six-membered hexagonal rings. The AA1100 G(r) spectrum has similar peak positions and relative intensities as the other three GC samples out to 6 ˚
A. Beyond this distance the peak intensities drop and the relative peak intensities vary significantly, especially in the regions indicated by the red arrows in Fig. 4.10. These results suggest that the graphene sheets that comprise the layered nanostructures in the AA1100 sample have a higher defect density relative to the other three GC samples.