This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn.
3.3. Sessile Drop Tests
Figure 5 shows images captured during the sessile drop tests. The images of the most important moments were isolated from the video sequence recorded.
Silicon starts to melt at the bottom of the sample, which is the area in contact with the substrates. When the silicon is totally molten, the drop is formed. Then, the contact angle decreases due to the spreading and the infiltration until silicon is completely infiltrated in the substrate (
Figure 5). For graphite, composite 1, and composite 2, the silicon drop infiltrates the substrate completely in less than 3 min. However, in the case of alumina, it can be seen that unlike the other materials, the spreading is only slight and the infiltration is completed after 28 min.
In
Figure 6, the variation of the contact angle and silicon drop height and width with time is shown for all materials. The time origin always corresponds to the moment when the drop is completely formed.
Figure 6a shows the case for the graphite substrate. The decrease in the contact angle and height and the increase in the width with time are totally linear. In addition, the process can be divided into two stages: the first one is when the silicon sample starts to melt until the drop is formed, and the second one is from the time when the drop is formed until the silicon is totally infiltrated. In both stages, infiltration and spreading happened. The spreading was managed by the reaction between silicon and graphite.
This is in agreement with many authors, for example, Israel [
17]. However, this same study asserts that the infiltration finishes when the pores of the substrate are closed, but in our case, the silicon infiltrates totally and then closure of the pores happens by the formation of silicon carbide due to the large size of the pores.
Silicon was melted on graphite under argon atmosphere using the sessile drop test. The roughness of graphite was 4.83 µm, and the initial contact angle obtained was 23.6° at 1424 °C. According to the references (
Table 3), the results are quite different. The variation in the contact angle values can be because the test conditions are not similar (roughness, temperature, and atmosphere), and the roughness and pore size show the most remarkable differences.
The behavior of the molten silicon on alumina is highly dependent on the possible formation of a passive layer of SiO
2 on the substrate. The spreading and infiltration on alumina can be divided into three stages (
Figure 6b). First, the contact angle decreases, the width increases, and the height remains almost constant (stage I). This behavior can be related to the formation of the passive layer. Then, the contact angle and the height almost linearly decrease while the width increases until a maximum (stage II). In this stage, the passive layer starts to break until it has been totally removed, and the contact angle and drop width and height dramatically decrease until the drop disappears through infiltration. Silicon was melted on alumina under argon atmosphere using the sessile drop test. The initial contact angle obtained was 86.2° at 1424 °C. According to the Yuan et al. [
18], this result is in agreement (86° at 1422 °C under argon atmosphere). Moreover, the spreading was minimum and infiltration was slower than on the other substrates, but the silicon infiltrated completely due to the high porosity of alumina. This good behavior against molten silicon can be due to the possible formation of the above-mentioned passive layer of SiO
2. The problem happened when the layer was broken and the infiltration was dramatically fast.
The variation in the contact angle and the parameters of the drop (height and width) with time for the silicon drop on composite 1 is shown in
Figure 6c. Because of the high variation in the data of the contact angle and width, a moving average filter was applied. The width remains almost constant, possibly due to the high roughness. However, two stages are shown with regard to the contact angle and height. Both magnitudes decreased much faster in the second stage than in the first one. The change in slope is remarkable. Silicon was melted on composite 1 under argon atmosphere using the sessile drop test. The initial contact angle obtained was 27.9° at 1425 °C.
On the other hand,
Figure 6d shows the variation in the contact angle of composite 2. The height and width of the drop were measured with respect to time. Similar to graphite, once the drop is formed, the width linearly increases while the height and contact angle linearly decrease. The spreading is driven by the chemical reaction between Si and C, as described by Equation (1). Silicon was melted on composite 2 under argon atmosphere using the sessile drop test. The initial contact angle obtained was 43.8° at 1422 °C. Similar to the other carbon materials, two parts are distinguished: the first one is governed mainly by infiltration and the second by spreading.
3.4. Wettability, Infiltration and Reactivity
The variation of the contact angle and relative height (mm/mm) of the drop on different substrates (graphite, alumina, and composites) with time is compared in
Figure 7.
Figure 7a compares the variation of the contact angle of all studied materials with time, once the drop was formed. First, the wettability of silicon on alumina is much lower than on the other materials. Alumina shows the highest contact angle in comparison to the other substrates. Additionally, the contact angle decreased much slower than on the other materials with time and temperature.
On the other hand, the contact angle of graphite linearly decreased in contrast to the composites where there was a change in slope. The slopes of the curves indicate that the contact angle of composite 2 decreases faster than that of the graphite and composite 1, whose slopes are very similar. This can indicate that the silicon–carbon reaction controls the triple line, and in the case of composite 2, it is possible that some scratches or irregularities of the surface improved the spreading, due to the fact the liquid move easily in the direction of the cracks.
Figure 7b shows the variation of the relative height (mm/mm) of different materials with the time. The data were collected when the silicon started to melt. The height was divided by the initial height in order to compare the values.
Since the melting starts at the bottom of the sample, infiltration and spreading are key factors in this stage, and the materials with relevant porosity will be infiltrated deeper than those whose surfaces boost the spreading.
As in the case of wettability, the infiltration in alumina was much slower than in the other materials. However, silicon infiltrated entirely in alumina, and in addition, the infiltration was faster with the increment of time.
For the carbonaceous materials, graphite and the composites show a remarkable change of slope. Before the drop, the reduction of height for composite 2 is slower than that for graphite and composite 1. However, after the drop, composite 2 and graphite are faster than composite 1. Observing the behavior of the composites, as in the case of graphite, the carbon-silicon reaction manages the reaction despite the change in the direction of the fibers that produce some modifications.
The infiltration rate Uinf and spreading rate Uspr were calculated for different substrates in each stage of the process (
Table 4). For graphite, they were calculated before and after the drop formation. The spreading was faster than the infiltration both before and after the formation of the drop. The kinetics rate agrees that the process is governed by chemical reaction at the triple line [
19].
In the case of alumina, spreading was faster than infiltration in the first stage, with both values being low. Then, in the second stage, the infiltration increased sharply while the spreading decreased slightly. These results support the formation and subsequent breakage of a passive layer of SiO2.
For composite 1, while the drop was forming, the spreading was slower than the infiltration. Then, both rates decreased; in particular, the spreading rate remained almost constant. In the second stage, the rates had the same order of magnitude, but the spreading is a bit faster than infiltration. In comparison with the behavior of graphite, some similarities were found. This indicates that spreading and infiltration is managed by the reaction at the triple line, but the roughness affected these phenomena.
For composite 2, in the first stage, spreading was faster than infiltration. Then, the spreading decreased and infiltration increased. As in the case of graphite, the results agree with the fact that the reaction (Equation (1)) manages the spreading and the infiltration of silicon on the substrate.
After the sessile drop test, the top view images were taken by an optical microscope (
Figure 8a). Then, the substrates were cut and the side view images were also taken (
Figure 8b), except for composite 1, because a high resolution was needed. For this purpose, a scanning electron microscope was used. Regarding the spreading, the effective radius
R was calculated from the wetted area
A (
A = π
R2), which was measured by image processing for each material. This technique was also used to measure the infiltration distance (
Table 5).
The wetted area can be divided into two concentric circles on the graphite and the composite materials. The smallest one matches with the area of the maximum infiltration distance and the biggest one with the maximum spreading area. On the other hand, on the alumina material, there is a single circle and the silicon has infiltrated the whole substrate.
In addition, the maximum radius was measured in situ by the sessile drop method and is included in
Table 5.
The fact that the circles are not symmetric indicates that any defect pinned the drop, decreasing the formation of the SiC layer but increasing the diffusivity of silicon in certain directions. According to
Table 5, the values of the maximum radius and the external radius are different. This fact is probably because the drop is not symmetric; thus, the size of the radius depends on the direction where it was recorded. On the other hand, another cause could be secondary wetting [
9], i.e., a parallel infiltration that takes place after the spreading time.
Regarding the infiltration, it is pointed out that the infiltration distance of composite 1 is the minimum in spite of its high grade of porosity. This fact agrees with the high rugosity and low grade of crystallinity. Both features make the infiltration very difficult. However, the infiltration distance in the alumina is the maximum because of the high grade of porosity.
On the other hand, a thin silicon carbide layer is produced on the surface (
Figure 9a). In
Figure 9b, we can see the micrographs of alumina after the sessile drop test where silicon has infiltrated inside the cracks of alumina, so that oxygen has been able to pollute the silicon, despite the X-ray results shown in
Figure 10 not revealing the formation of silicon oxide (
Figure 11). In
Figure 9c, the micrographs of composite 1 after the sessile drop test are shown where silicon has infiltrated between and has reacted with the carbon fibers. According to
Figure 6 and
Figure 12, there is no unreacted silicon. This means that the reaction in the triple line is a key factor in spite of the roughness.
Figure 9d shows the micrographs of composite 2. Silicon has infiltrated through the cracks of composite 2 and reacted with carbide, but unreacted silicon has been found (
Figure 13). In accordance with Novakovic et al. [
20], 10% of silicon remains unreacted when Si/C composites are obtained by the infiltration method. Regarding the reactivity of graphite, it can be highlighted that the graphite shows pores that are almost closed, and there is no unreacted silicon, as confirmed by the X-ray results (
Figure 10).