Effects of Changing the Specific Surface Area in the Ceramic Matrix of CAC-Containing Refractory Castables on the Initial Stiffening and Setting Behaviour

3.1. Temperature of the Castables after Mixing Process The mass temperature influences the refractory castable properties, such as initial stiffening and setting. There are clear differences in the castable temperature after mixing, which are mainly due to the different specific surface areas of the matrix (and also the dispersing agent used, which influences the required water content). The warming of the castables is mainly caused by the mixing process (and also by the room, raw material and water temperature). The measured temperatures directly after mixing the refractory castables can be seen in Figure 3. The following trends can be recognised:



With a higher specific surface area, the refractory castables heat up less during mixing. This can be explained by the dilatant behaviour of the formulations with low specific surface areas. Castables with higher specific surface areas tend to exhibit rheological Newtonian or shear-thinning behaviour, requiring less mixing energy. As a result, they heat up less [39]. The temperature range extends from 22.9 to 17.4 °C for castables with high surface areas (4.7 m²/g) to 32.7 to 24.0 °C for castables with low surface areas (1.0 m²/g).



The influence of the dispersing agent system used is reflected indirectly via the temperature of the castables. Model castables dispersed with PCE (Castament FS60) show a higher temperature after mixing than with PMA (Darvan 7S in combination with citric acid). The lowest castable temperature is observed for formulations dispersed with S-TPP (N25-15) and citric acid. The greatest influence here is the different water demand, which increases in this order: 4.8 wt.-%, 6.3 wt.-% and 6.7 wt.-% (Table 2 and Table 3). The adjustments of the water content were necessary due to the reduced dispersion strength of PMA and S-TPP compared to PCE. A higher water content leads to a reduction in the interparticle forces at the high shear rates of mixing, which results in castables with a higher water content heating up less [39].



In most cases, higher concentrations of the respective dispersing agents result in lower castable temperatures. This can be explained by the lower mixing energy required when more dispersing agent is used [39]. The increased effectiveness of the dispersing agents when the concentrations are increased and thus weaker interparticle forces in the mixing process result in lower castable temperatures.



More monomodally distributed refractory castable matrix compositions (as is the case with 3.0 m²/g with a high proportion of a single particle fraction ((19 wt.-% of CTC20), Table 2 and Figure 2) tend to lead to dilatancy (although weakly pronounced) [6], which requires higher mixing energy, resulting in (slightly) higher mass temperatures [39]. Multimodal particle size distributions of the matrix (2.7 m²/g) can have a positive influence on the mixing energy [5,6,40,41,42]. This leads to lower castable temperatures. However, these assumptions still need to be verified in further tests.

3.2. PCE (Castament FS60) The results, shown in Figure 4, indicate that the first significant increase in sonic velocity (enormous increase in the rate of change of sonic signal) of the PCE dispersed refractory castables (0.10 wt.-% of Castament FS60), attributed to the initial stiffening depends on the specific surface area of the matrix composition. All castables in this series exhibit a relatively extended processing time period after mixing and moulding, during which the sonic velocity does not increase. This phenomenon can be attributed to the mode of action of this PCE, which enables the dispersion of newly formed hydrate phases [18]. For the high specific surface area mixtures, a first (significant) increase in sonic velocity can be observed at first after the start of the test. The refractory castable with a surface area of 4.7 m²/g reaches initial stiffening after 3.00 h, followed by the mixture with a surface area of 4.5 m²/g at 3.25 h. Castable 3.0 m²/g stiffens after 6.00 h and 2.7 m²/g after 8.50 h. In these cases, the initial stiffening of the refractory castables appears to depend very systematically on the specific surface area of the matrix. Formulations with higher specific surface areas stiffen earlier. This can be attributed to more rapid consumption of the dispersing agent on the high specific surface areas of the matrix particles than in castables with lower specific surface areas. At lower specific surface areas of the matrix, there is more free dispersing agent (PCE) in the pore water available so that hydrates that form can still be dispersed over a longer period of time. This is why mixtures with low specific surface areas stiffen at a later point in time. Here, the different specific surface areas (4.7, 4.5, 3.0, and 2.7 m²/g) appear to be the dominant effect, leading to different stiffening times (earlier points in time with higher specific surface areas). Zeta potential measurements on matrix suspensions correlated with this established trend. Matrix suspensions with higher specific surface areas dispersed with the same PCE stiffened earlier than those with lower specific surface areas [20]. Castables 1.0 and 1.3 m²/g exhibited stiffening after 6.25 h and 6.50 h, respectively. This is not in line with the expected trend that these castables should stiffen at later points in time (due to their low specific surface area). One explanation could be that the first stiffening is sensitive to temperature. Higher castable temperatures from 1.0 and 1.3 m²/g (30.0 and 26.4 °C) towards 2.7, 3.0, 4.5 and 4.7 m²/g (23.3, 23.5, 25.2 and 22.9 °C) were found (Figure 3, see explanation for the differences in castable temperatures under Section 3.1). For formulations with a low specific surface area (1.0 and 1.3 m²/g), the effect of the castable temperature seems to have an increased influence, which is why these castables stiffen earlier (than in line with the expected trend of a later stiffening). In mathematical terms, different inflection points can be determined in the curves of the sonic velocity. Mixtures 4.5 and 4.7 m²/g show two inflection points (one earlier, close to the first stiffening, and one later). The other compositions each show only one inflection point. For the following evaluations, only the 2nd inflection point of the sonic velocity curves of the mixtures 4.5 and 4.7 m²/g and the only inflection point of the other refractory castables are considered. This two-stage mechanism (two inflection points) in the early time range of castables 4.5 and 4.7 m²/g could possibly be cautiously explained as follows: Some research indicates the early formation of amorphous calcium aluminum hydrate phases immediately after mixing. While these early hydrates might not be the primary cause of the initial stiffening defined by rheological changes, their formation on the high surface area of the matrix particles could contribute to changes in the suspension stability and influence the effectiveness of the PCE, indirectly affecting the stiffening behaviour. The absence of a measurable temperature increase during their formation supports an amorphous nature (Figure 5). In the discussion of tabular alumina-based refractory castables from Kasper [18], the data showed that hydrate water was already present immediately after mixing and increased significantly in the first hours. This early binding of water could be attributed to the formation of poorly crystallized or gel-like hydrates (which might not be detected by X-ray diffraction) [18]. However, the nature of these early water-binding phases (e.g., amorphous hydrates or gels) requires careful consideration and potentially complementary techniques to fully understand. Analysis of the (2nd) inflection points of the sonic velocity curves reveals that, in contrast to the initial phase of stiffening, they tend to occur within a narrow time frame for all refractory castables. At this concentration of the PCE (Castament FS60, 0.10 wt.-%), the inflection points are observed to occur within a time period of 9.00 to 10.50 h. For the same dispersing agent concentration, there appears to be no dependency on the specific surface area of the matrix. The previously described characteristics of the sonic velocity progression were also observed for the other two concentrations of PCE (Castament FS60, 0.075 and 0.15 wt.-%). Figure 5 depicts that as the dispersing agent content increases, the first (significant) increase in sonic velocity, attributed to the first stiffening and the (2nd) inflection point of the sonic velocity curves, are both shifted to a later point in time. At the lowest concentration of 0.075 wt.-% the refractory castable 4.7 m²/g exhibited stiffening after 2.00 h. The castable with 0.10 wt.-% started stiffening after 3.25 h, while dispersion with 0.15 wt.-% required 6.50 h. This can be explained by the fact that the ratio of the dispersing agent to a specific surface area increases with a higher dispersing agent concentration. As a result, there is proportionally more free dispersing agent (PCE) present in the pore water. Therefore, more hydrates that form can be dispersed before the rheological system collapses. This results in a later initial stiffening. During measurements of the zeta potential matrix, suspensions with the same specific surface area but a higher content of the same PCE (Castament FS60) stiffened distinctly later than those with lower concentrations [20]. The (2nd) inflection points of the curves are at 10.00 h for 0.075 wt.-%, 10.50 h for 0.10 wt.-% and 12.50 h for 0.15 wt.‑%. At higher dispersing agent contents, the (2nd) inflection points are shifted to a later point in time approximately linearly proportional to the increase in the concentration of the dispersing agent (as with the initial stiffening). Another value measured and depicted in Figure 5 is the temperature profile. The maximum temperature in each castable corresponds to an exothermic reaction, which can be attributed to the main formation of hydrate phases. These exothermic peaks (temperature maximum) are observed at 11.25 h for a concentration of 0.075 wt.-%, 12.25 h at 0.10 wt.-%, and 14.00 h at 0.15 wt.-%. All temperature peaks coincide with the respective (2nd) inflection points of the sonic velocity, which can, therefore, be correlated with the time of main hydration. This correlation has been documented in previous literature [18,43]. The phenomenon of delayed initial stiffening of castables with increased PCE content and the subsequent delay (linearly proportional) in the (2nd) inflection point of the sonic velocity curve was also observed in all other refractory castables with different matrix compositions dispersed with the PCE. The described correlations regarding the initial stiffening and hydration are highly systematic for all refractory castables with different specific surface areas dispersed with PCE (Castament FS60) as becomes evident in Figure 6 and Figure 7. The first stiffening shows a clear dependency on the specific surface area of the matrix (Figure 6). Considering the content of the dispersing agent per specific surface area (a dashed line shows an approximate trend), a linearly proportional dependency can be recognised that the time of the first stiffening is shifted to a later time with a lower specific surface area and an increasing content of dispersing agent (for castables 2.7, 3.0, 4.5 and 4.7 m²/g). Times of an initial stiffening of 120 to 400 min (4.5; 4.7 m²/g) and 250 to 600 min (2.7; 3.0 m²/g) were established. This trend is not apparent for mixtures 1.0 and 1.3 m²/g (first stiffening after 275 to 600 min). One explanation for the earlier initial stiffening of these formulations (than in line with the expected trend) may be the significantly higher castable temperature after the mixing process (Figure 3, see explanation for the differences in castable temperatures under Section 3.1). The inflection points (2nd inflection point of the sonic velocity curves of the mixtures 4.7 and 4.5 m²/g and the only inflection points of 1.0, 1.3, 2.7 and 3.0 m²/g) of the sonic velocity curve seem with times of 8.50 to 10.00 h (0.075 wt.-%) and 11.25 to 12.50 h (0.15 wt.-%) to be independent of the matrix compositions (Figure 7). A clear dependence on the concentration of the PCE can be recognised. A higher concentration leads to a later hydration range (irrespective of the specific surface areas). The temperature maximum during setting is equally delayed and matches with the (2nd) inflection point of the sonic velocity in each case, respectively and can therefore be related to the time of main hydration. 3.3. PMA (Darvan 7S) and Citric Acid Using PMA (0.07 wt.-% of Darvan 7S) with citric acid instead of PCE (Castament FS60) leads to notable differences in the sonic velocity curve progression, although a clear dependence on the specific surface area can still be observed (Figure 8). A first significant increase in sonic velocity (enormous increase in the rate of change of sonic signal) can be determined within a couple of minutes, which is attributed to the initial stiffening of the refractory castables. This comparatively rapid initial stiffening (in contrast to PCE) can be explained by the mode of action of the PMA, as it precipitates at a critical concentration with Ca²⁺ (supersaturation of Ca-dispersant-compounds) [18,22,44]. The zoomed-in section clearly shows that formulations with a higher specific surface area of the matrix demonstrate an earlier onset of initial stiffening. Castables 4.7 and 4.5 m²/g stiffened after ca. 5 min, 3.0 m²/g within ca. 10 min and 2.7 m²/g after ca. 20 min. This is because more dispersing agent adsorbs on higher specific surface areas. This means that there is less free dispersing agent present in the pore water. When Ca²⁺ is dissolved from the cement, the free PMA (but also citric acid) is consumed first. Subsequently, adsorbed PMA (and citric acid) are removed from the surfaces, which gradually leads to coagulation by precipitation of sparingly soluble Ca-PMA-compounds (and Ca-citrate-compounds). Therefore, at high specific surface areas, initial stiffening occurs earlier, as adsorbed PMA (and citric acid) are removed from the surfaces faster (which previously maintained the dispersion). As is also the case with PCE, the different specific surface areas (4.7, 4.5, 3.0, and 2.7 m²/g) appear to be the dominant effect, which leads to different times of stiffening (earlier points in time with higher specific surface areas). However, due to the completely different dispersion mechanism at a much earlier point in time than with PCE. Zeta potential measurements on matrix suspensions correlated with this trend. Matrix suspensions with higher specific surface areas dispersed with the same PMA (combined with the same citric acid) stiffened at earlier points in time than those with lower specific surface areas [20]. Castables 1.0 and 1.3 m²/g exhibited initial stiffening after ca. 10 and ca. 15 min, respectively, and do not follow the expected trend that these castables should stiffen at later points in time (due to their low specific surface area). As the first stiffening is sensitive to temperature, higher castable temperatures from 1.0 and 1.3 m²/g with 31.7 and 26.1 °C towards 2.7, 3.0, 4.5 and 4.7 m²/g with 22.1, 21.9, 20.7 and 18.5 °C seem to have an increased influence here (as with PCE), which is why the castables with a low specific surface area of the matrix stiffen at an earlier point in time (than expected (Figure 3, explanation for the differences in castable temperatures under Section 3.1)). Similar to the initial stiffening, the 2nd inflection points of sonic velocity curves demonstrate a dependency on the specific surface areas of the matrix. In the model castables with a lower surface area, the 2nd inflection point, which also indicates the hydration range, is shifted to a later point in time. At a concentration of PMA dispersing agent (Darvan 7S) of 0.07 wt.-%, the 2nd inflection points can be detected after 12.00 and 13.25 h for refractory castables with surface areas of 4.7 and 4.5 m²/g, respectively. Castables 3.0 m²/g exhibits a 2nd inflection point at 15.00 h and 2.7 m²/g at 19.00 h. As already observed during the initial stiffening, the 2nd inflection points of the castables 1.0 m²/g at 18.25 h and 1.3 m²/g at 15.25 h do not correspond to the trend of later hydration of formulations with lower specific surface areas. This could again be explained by the higher refractory castable temperature of 1.0 and 1.3 m²/g compared to 2.7, 3.0, 4.5, and 4.7 m²/g (Figure 3). The temporal shift of the 2nd inflection point of the sonic velocity curve occurs over a much longer period of time than the relatively short temporal shift associated with the initial stiffening. The described trends in relation to the sequence of the 2nd inflection points (and also the times of the initial stiffening) could also be observed for all other model castables dispersed with PMA (Darvan 7S) in different concentrations. An extremely systematic pattern within the sonic velocity curves was observed as the dispersing agent content of PMA was increased (Figure 9). There is a recognisable trend that a first significant increase in sonic velocity (enormous increase in the rate of change of sonic signal), indicating initial stiffening and also the 2nd inflection point of the sonic velocity progression to be shifted to a later point in time. As depicted in the zoomed-in view, the refractory castable 4.5 m²/g exhibited earlier stiffening with a concentration of 0.07 wt.-% Darvan 7S (at ca. 5 min) than with 0.10 wt.-% (at ca. 10 min) and 0.13 wt.-% (at ca. 15 min) with an approximately linearly proportional dependency. Due to a higher concentration, there is more free dispersing agent present in the pore water. As Ca²⁺ is dissolved from the cement, free PMA (but also citric acid) is consumed first. Afterwards, adsorbed PMA (and citric acid) is removed from the surfaces, gradually leading to coagulation by precipitation of sparingly soluble Ca-PMA-compounds (and Ca-citrate-compounds). Therefore, at higher concentrations, initial stiffening occurs later, as more PMA (and citric acid) is in the pore water available and simultaneously adsorbed PMA (and citric acid) is removed from the surfaces more slowly (whereby the dispersion mechanism can be maintained over a longer period of time). During zeta potential measurements, matrix suspensions with the same specific surface area but a higher content of the same PMA (Darvan 7S) and the same citric acid showed a later stiffening than those with lower concentrations [20]. The 2nd inflection points of the sonic velocity curves occur at 13.25, 14.50, and 18.25 h with concentrations of Darvan 7S of 0.07 wt.-%, 0.10 wt.-% and 0.15 wt.-%, respectively, indicating a linearly proportional correlation. The temporal shift is significantly more pronounced in the hydration range than in the relatively brief temporal shift observed during the initial stiffening phase. A similar systematic can be identified when considering the temperature development. The temperature maximum values for this test series are 15.50 h at a concentration of 0.07 wt.-%, 17.25 h at 0.10 wt.-%, and 21.25 h at 0.13 wt.-%. As also observed with PCE, the time of the temperature maximum is found to correlate with the time of main hydration in each case. These described correlations regarding the initial stiffening and hydration are highly systematic for all refractory castables with different specific surface areas dispersed with PMA (Darvan 7S) and citric acid, as depicted in Figure 10 and Figure 11. The first stiffening shows a clear dependency on the specific surface area of the matrix (Figure 10). As also found with PCE, when considering the content of dispersing agent per specific surface area (a dashed line shows an approximate trend), a linearly proportional dependency can be recognised that the time of the initial stiffening is shifted to a later point in time with a lower specific surface area and increasing content of PMA dispersing agent (for castables 2.7, 3.0, 4.5 and 4.7 m²/g). Times of the initial stiffening of 5 to 15 min (4.5; 4.7 m²/g) and 10 to 25 min (2.7; 3.0 m²/g) were measured. This trend is not apparent for mixtures 1.0 and 1.3 m²/g (first stiffening after 10 to 20 min). An explanation for the earlier stiffening of these formulations (than in line with the expected trend) can be the significantly higher castable temperature after the mixing process (Figure 3, see an explanation for the differences in castable temperatures under Section 3.1). This behaviour was also observed and described with PCE. As visualised in Figure 11, the 2nd inflection points of the sonic velocity curves (marking main hydration) appear to be clearly dependent on the matrix compositions (in contrast to PCE, where independence was established). Refractory castables dispersed with PMA (Darvan 7S) with a higher specific surface area of the matrix show an earlier point of hydration compared to those with lower specific surface areas. The use of a higher concentration of PMA has been observed to result in a delayed hydration range. Hydration times of 12.00 to 19.00 h (0.07 wt.-%) and 13.50 to 21.00 h (0.15 wt.-%) were observed. The temperature maximum during the setting exhibits a highly systematic correlation with the time of main hydration (2nd inflection point of the sonic velocity) in each case. The temperature maxima also established at later points in time with lower specific surface areas and with higher PMA contents. 3.4. S-TPP (N25-15) and Citric Acid Figure 12 displays that the sonic velocity of the refractory model castables using the S-TPP dispersing agent N25-15 (0.05 wt.-%) and citric acid (0.015 wt.-%) shows a high systematicity and demonstrates a clear correlation with the specific surface area of the matrix. The sonic velocity progression exhibits a comparable profile to those observed when using PMA (Darvan 7S) in combination with citric acid as a dispersing agent. Within a few minutes at the beginning, a first significant increase in sonic velocity (enormous increase in the rate of change of sonic signal) is observed due to the initial stiffening of the refractory castables. As with PMA, this comparatively rapid initial stiffening can be explained by the mode of action of the S-TPP, which precipitates at a critical concentration with Ca²⁺ (supersaturation of Ca-dispersant-compounds) [18,45]. The zoomed-in view clarifies that castables with a higher specific surface area of the matrix exhibit an earlier onset of initial stiffening. The model castable with a specific surface area of 4.7 m²/g stiffened after ca. 5 min, 4.5 m²/g stiffened within ca. 15 min, 3.0 m²/g after ca. 20 min, and 2.7 m²/g within ca. 25 min. For the first stiffening, the same effect occurs as for PMA: More dispersing agent adsorbs on higher specific surface areas. Thus, less free dispersing agent is available in the pore water. When Ca²⁺ is dissolved from the cement, the free S-TPP but also citric acid are consumed first. Subsequently, adsorbed S-TPP and citric acid are removed from the surfaces, which gradually leads to coagulation by precipitation of sparingly soluble Ca-TPP-compounds and Ca-citrate-compounds. Therefore, initial stiffening occurs earlier at high specific surface areas, as adsorbed S-TPP and citric acid are removed from the surfaces faster (which previously maintained the dispersion). As with PCE and PMA, the different specific surface areas (4.7, 4.5, 3.0, and 2.7 m²/g) appear to be the dominant effect, leading to different stiffening times (earlier points in time with higher specific surface areas). Mixtures 1.3 and 1.0 m²/g stiffen after ca. 15 and ca. 20 min, respectively, and do not follow the expected trend that castables with lower specific surface areas stiffen at later points in time. As the first stiffening appears to be sensitive to temperature, higher castable temperatures from 1.0 and 1.3 m²/g with 25.4 and 29.3 °C towards 2.7, 3.0, 4.5, and 4.7 m²/g with 18.4, 20.2, 18.8 and 18.1 °C most probably have an increased influence here (as with PCE and PMA at 1.0 and 1.3 m²/g). This could be why the castables with a low specific surface area of the matrix stiffen at an earlier point in time (than expected (Figure 3, explanation for the differences in castable temperatures under Section 3.1)). The 2nd inflection points of the sonic velocity also correlate with the specific surface area of the refractory castables’ matrices (as observed with PMA). In the mixtures with a lower specific surface area, the 2nd inflection point, marking the hydration of CA-cement, is shifted highly systematically to a later point in time. The hydration times achieved in descending order of specific surface area from 4.7 to 1.0 m²/g are 15.25, 17.25, 21.00, 22.25, 23.00 and 25.50 h. The temporal shift in hydration occurs over a much longer period of time than the relatively short temporal shift associated with the initial stiffening. The aforementioned trends (initial stiffening and hydration) were also observed in all other model castables with S-TPP dispersing agent N25-15 and citric acid at different concentrations. Figure 13 depicts that the sonic velocity and the temperature evolution of the castable 3.0 m²/g with S-TPP dispersing agent N25-15 (0.05 wt.-%) in combination with citric acid at concentrations of 0.010, 0.015, and 0.020 wt.-% exhibit a high degree of systematic order. The first significant increase in sonic velocity (enormous increase in the rate of change of sonic signal), indicating a first stiffening, and also the 2nd inflection point of the sonic velocity curves are shifted to a later point in time with increasing dispersing agent content (in this case citric acid). In the zoomed-in section, it can be observed that the refractory castable with a concentration of 0.010 wt.-% citric acid exhibits earlier stiffening (after ca. 10 min) than that with 0.015 wt.-% (after ca. 20 min) and 0.020 wt.-% (after ca. 30 min) with a linearly proportional dependency. The same reasoning as for PMA can be used to explain this: Due to a higher concentration, there is more free dispersing agent present in the pore water. As Ca²⁺ is dissolved from the cement, free S-TPP and citric acid are consumed first. Afterwards, adsorbed S-TPP and citric acid are removed from the surfaces. This gradually leads to coagulation by precipitation of sparingly soluble Ca-TPP-compounds and Ca-citrate-compounds. At higher concentrations (of citric acid), initial stiffening occurs later, as more S-TPP and citric acid are in the pore water available and simultaneously adsorbed S-TPP and citric acid are removed from the surfaces more slowly. Thus, the dispersion mechanism can be maintained over a longer period of time. The main hydration ranges of the castables are 16.75 h (0.010 wt.-%), 21.00 h (0.015 wt.-%), and 26.50 h (0.020 wt.-%). With an increased citric acid concentration at a constant specific surface area, the times of hydration are markedly delayed. A similar trend can be identified when examining the temperature maxima. The values of the temperature maxima, indicating the predominant time of formation of hydrate phases, are determined at fairly identical points in time. It is noticeable that the height of the temperature maximum peaks decreases with increasing citric acid content. This could indicate that the exothermic reaction is less intense, and consequently, fewer hydrate phases are formed (not systematically analysed). The temporal shift of the hydration range is considerably more pronounced than that observed for the relatively brief temporal shift in the first stiffening. Figure 14 and Figure 15 underline the described highly systematic correlations regarding the initial stiffening and hydration for all refractory castables with different specific surface areas using S-TPP (N25-15) and citric acid. The first stiffening is clearly dependent on the specific surface area of the matrix (Figure 14). As with PCE and PMA, examining the content of the dispersing agent per specific surface area (a dashed line shows an approximate trend) reveals a linearly proportional dependency. The time of the first stiffening is shifted to a later time with a lower specific surface area and an increasing content of (S-TPP and) citric acid dispersing agent (for castables 2.7, 3.0, 4.5 and 4.7 m²/g). Times of the initial stiffening of < 5 to 25 min (4.5; 4.7 m²/g) and 10 to 30 min (2.7; 3.0 m²/g) were measured. This trend is not apparent for mixtures 1.0 and 1.3 m²/g (first stiffening after 15 to 20 min). The earlier initial stiffening of these formulations (than in line with the expected trend) can be explained by the significantly higher castable temperature after the mixing process (Figure 3, see explanation for the differences in castable temperatures under Section 3.1). This behaviour was also observed and described with PCE and PMA. Figure 15 shows that the 2nd inflection points of the sonic velocity curve (marking main hydration) appear to be clearly dependent on the matrix compositions (as with PMA, but in contrast to PCE, where independence was established). Castables dispersed with S-TPP (N25-15) and citric acid with a higher specific surface area of the matrix show an earlier point of hydration compared to those with lower specific surface areas. A higher concentration of citric acid results in a delayed time of hydration. Hydration times of 13.00 to 19.00 h (0.010 wt.-%), 15.25 to 25.50 h (0.015 wt.-%) and 18.00 to 30.00 h (0.020 wt.-%) were observed. A highly systematic correlation of the temperature maximum during the setting and the time of main hydration (2nd inflection point of the sonic velocity) in each case can be observed. The temperature maxima were also established at later points in time with a lower specific surface area and with higher (S-TPP and) citric acid contents. The variation of citric acid in the system N25-15 is an effective means of systematically and finely controlling both the stiffening and hydration times. Another correlation that can be recognised is that at a low concentration of citric acid, there is only a small dependency on the specific surface area, but at higher concentrations, there appears to be an increased dependency on the specific surface area. One possible explanation could be that the specific surface area has less influence at low citric acid concentrations, as S-TPP (that is kept constant) primarily influences the hydration process. In this case, S-TPP plays the primary role in delaying hydration. At higher concentrations of citric acid, its surface-mediated effect becomes more pronounced (especially for high specific surface areas), making differences in the specific surface area of the refractory castables’ matrix more significant.