*3.1. Material Characterization, Dispersion, and Microstructure*

The TGA results in Figure 2, show that the fine and medium graphite materials decomposed completely (lost 100% of their weight) between 40 and 1000 ◦C, while the coarse graphite did not. This could be due to the particle size of the coarse graphite (2 mm) as the flakes were too large for all their minerals to decompose. The smaller the particle size, the faster the minerals decompose with increasing temperature, and therefore the curves shift to a lower temperature as the material size reduces. None of the materials showed any mass loss before 100 ◦C; therefore, they are all expected to be stable during cement mixing and hydration. Furthermore, the DTG curves do not show the presence of any other minerals so the graphite products are of high purity.

**Figure 2.** Characterization of the three graphite products by thermogravimetric analysis (TGA), showing the weight loss from 100 to 1000 ◦C.

SEM was used to understand the morphology of the graphite materials. In Figure 3a,b the coarse graphite and some large flakes with a size of at least 1 mm can be observed, which is expected due to the inaccuracies when sieving. For the medium graphite (Figure 3c,d), flakes are of varying sizes but are all between 150 and 200 μm. In Figure 3e,f, the fine graphite particles can be seen with a wrinkled and folded structure. As the graphite became finer, some agglomerates could be seen, and it was difficult to isolate individual flakes. SEM–EDX was also used to assess the dispersion effectiveness of the dry mixing protocol. Only a coarse graphite-cement paste specimen at 28 days of hydration was used and the coarse graphite particles are shown as green in Figure 3g, whilst the cement paste matrix is shown as pink. The EDS analysis (Figure 3h) confirms that the main elements are carbon and calcium, which was expected due to the presence of graphite and because calcium silicate (C–S–H) is the main reaction product of cement hydration. Other typical elements from cement hydration, such as silicon, oxygen, sulfur, and magnesium, were also present.

**Figure 3.** SEM images of the three graphite products, with two different magnification images for each: coarse graphite (**a**,**b**), medium graphite (**c**,**d**), fine graphite (**e**,**f**), and SEM–EDX confirming good dispersion of the coarse graphite flakes at 28 days (**g,h**).

SEM testing was also undertaken after five months of curing, to assess the interaction of the graphite particles with the hydrated cement paste and to understand the effect on microstructure. In Figure 4, the individual graphite flakes could be identified in all cases, irrespective of the graphite size. Flakes were observed in proximity, which was expected, as graphite concentrations were high (the percolation threshold samples are illustrated below), but no obvious agglomeration was present. Therefore, dispersion of graphite particles was assumed to be adequate with dry mixing, which is a technique that has also been followed in the literature [16]. Observing the interaction between the graphite particles and the cement hydration products, we can see an interfacial transition zone similar to that of aggregates for the coarse graphite (Figure 4a), which could lead to planes of weakness and a reduction in mechanical performance. These results are not directly comparable with other studies, as the microstructural interaction could also be affected by the cement type, dispersion technique and graphite morphology. Overall, none of the three natural graphite products that were used here was found to alter the microstructure of the CEMI paste.

A μCT-scan was also used to assess the dispersion of a 30 wt % coarse graphite dosage in the cement matrix with the dry mix protocol. Figure 5a,b shows the 3D reconstructed image of the 5 mm graphite-cement paste sample and Figure 5c shows a slice through the 3D image. A grayscale analysis was used to detect the different substances based on their density. Three distinct peaks were observed in the μCT-scan, with each representing the different elements including air, bulk cement paste, and graphite particles. In Figure 5, the graphite particles are represented with a pink color, whilst the bulk cement paste is shown as black. The graphite flakes were found to be well dispersed within the matrix and near each other which was due to the high graphite concentration (30 wt %). The orientation of the flakes appeared to vary, even though some clusters were oriented in the same direction. The key finding from the μCT-scan was that graphite flakes were well dispersed with the dry mix method, and even at a high graphite concentration, no agglomerates were observed.

**Figure 4.** *Cont*.

**Figure 4.** Microstructural characterization of the three graphites-cement paste after ~5 months by SEM (**a**,**b**) coarse, (**c**,**d**) medium, and (**e**,**f**) fine graphite.

**Figure 5.** X-ray computed tomography (CT-scan) of an approximately 5 mm cement paste specimen with a 30 wt % coarse graphite addition. The graphite flakes (pink) are dispersed in the matrix (gray): (**a**,**b**) 3D reconstructed image of the specimen and (**c**) a slice through the 3D image.

The microstructural characterization showed that the dry-mixing protocol that was followed here was effective, while graphite had no pronounced effect on the microstructure of cement paste, irrespective of its fineness. This means that graphite is acting as an inert filler in the mix and does not alter the chemical structure of the hydrated cement paste.
